Merge branch 'next' into for-linus

Prepare first round of input updates for 3.20.

792 files changed, 80314 insertions, 22295 deletions

diff --git a/drivers/gpu/drm/Kconfig b/drivers/gpu/drm/Kconfig
index b066bb3ca01a..c3413b6adb17 100644
--- a/drivers/gpu/drm/Kconfig
+++ b/drivers/gpu/drm/Kconfig
@@ -8,6 +8,7 @@ menuconfig DRM
 	tristate "Direct Rendering Manager (XFree86 4.1.0 and higher DRI support)"
 	depends on (AGP || AGP=n) && !EMULATED_CMPXCHG && MMU && HAS_DMA
 	select HDMI
+	select FB_CMDLINE
 	select I2C
 	select I2C_ALGOBIT
 	select DMA_SHARED_BUFFER
@@ -24,12 +25,6 @@ config DRM_MIPI_DSI
 	bool
 	depends on DRM
 
-config DRM_USB
-	tristate
-	depends on DRM
-	depends on USB_SUPPORT && USB_ARCH_HAS_HCD
-	select USB
-
 config DRM_KMS_HELPER
 	tristate
 	depends on DRM
@@ -115,6 +110,7 @@ config DRM_RADEON
 	select HWMON
 	select BACKLIGHT_CLASS_DEVICE
 	select INTERVAL_TREE
+	select MMU_NOTIFIER
 	help
 	  Choose this option if you have an ATI Radeon graphics card.  There
 	  are both PCI and AGP versions.  You don't need to choose this to
@@ -171,6 +167,8 @@ config DRM_SAVAGE
 
 source "drivers/gpu/drm/exynos/Kconfig"
 
+source "drivers/gpu/drm/rockchip/Kconfig"
+
 source "drivers/gpu/drm/vmwgfx/Kconfig"
 
 source "drivers/gpu/drm/gma500/Kconfig"
@@ -204,3 +202,7 @@ source "drivers/gpu/drm/tegra/Kconfig"
 source "drivers/gpu/drm/panel/Kconfig"
 
 source "drivers/gpu/drm/sti/Kconfig"
+
+source "drivers/gpu/drm/amd/amdkfd/Kconfig"
+
+source "drivers/gpu/drm/imx/Kconfig"
diff --git a/drivers/gpu/drm/Makefile b/drivers/gpu/drm/Makefile
index 4a55d59ccd22..e620807418ea 100644
--- a/drivers/gpu/drm/Makefile
+++ b/drivers/gpu/drm/Makefile
@@ -4,7 +4,7 @@
 
 ccflags-y := -Iinclude/drm
 
-drm-y       :=	drm_auth.o drm_buffer.o drm_bufs.o drm_cache.o \
+drm-y       :=	drm_auth.o drm_bufs.o drm_cache.o \
 		drm_context.o drm_dma.o \
 		drm_fops.o drm_gem.o drm_ioctl.o drm_irq.o \
 		drm_lock.o drm_memory.o drm_drv.o drm_vm.o \
@@ -14,7 +14,7 @@ drm-y       :=	drm_auth.o drm_buffer.o drm_bufs.o drm_cache.o \
 		drm_info.o drm_debugfs.o drm_encoder_slave.o \
 		drm_trace_points.o drm_global.o drm_prime.o \
 		drm_rect.o drm_vma_manager.o drm_flip_work.o \
-		drm_modeset_lock.o
+		drm_modeset_lock.o drm_atomic.o
 
 drm-$(CONFIG_COMPAT) += drm_ioc32.o
 drm-$(CONFIG_DRM_GEM_CMA_HELPER) += drm_gem_cma_helper.o
@@ -22,10 +22,8 @@ drm-$(CONFIG_PCI) += ati_pcigart.o
 drm-$(CONFIG_DRM_PANEL) += drm_panel.o
 drm-$(CONFIG_OF) += drm_of.o
 
-drm-usb-y   := drm_usb.o
-
 drm_kms_helper-y := drm_crtc_helper.o drm_dp_helper.o drm_probe_helper.o \
-		drm_plane_helper.o drm_dp_mst_topology.o
+		drm_plane_helper.o drm_dp_mst_topology.o drm_atomic_helper.o
 drm_kms_helper-$(CONFIG_DRM_LOAD_EDID_FIRMWARE) += drm_edid_load.o
 drm_kms_helper-$(CONFIG_DRM_KMS_FB_HELPER) += drm_fb_helper.o
 drm_kms_helper-$(CONFIG_DRM_KMS_CMA_HELPER) += drm_fb_cma_helper.o
@@ -36,10 +34,10 @@ CFLAGS_drm_trace_points.o := -I$(src)
 
 obj-$(CONFIG_DRM)	+= drm.o
 obj-$(CONFIG_DRM_MIPI_DSI) += drm_mipi_dsi.o
-obj-$(CONFIG_DRM_USB)   += drm_usb.o
 obj-$(CONFIG_DRM_TTM)	+= ttm/
 obj-$(CONFIG_DRM_TDFX)	+= tdfx/
 obj-$(CONFIG_DRM_R128)	+= r128/
+obj-$(CONFIG_HSA_AMD) += amd/amdkfd/
 obj-$(CONFIG_DRM_RADEON)+= radeon/
 obj-$(CONFIG_DRM_MGA)	+= mga/
 obj-$(CONFIG_DRM_I810)	+= i810/
@@ -52,6 +50,7 @@ obj-$(CONFIG_DRM_VMWGFX)+= vmwgfx/
 obj-$(CONFIG_DRM_VIA)	+=via/
 obj-$(CONFIG_DRM_NOUVEAU) +=nouveau/
 obj-$(CONFIG_DRM_EXYNOS) +=exynos/
+obj-$(CONFIG_DRM_ROCKCHIP) +=rockchip/
 obj-$(CONFIG_DRM_GMA500) += gma500/
 obj-$(CONFIG_DRM_UDL) += udl/
 obj-$(CONFIG_DRM_AST) += ast/
@@ -65,6 +64,7 @@ obj-$(CONFIG_DRM_BOCHS) += bochs/
 obj-$(CONFIG_DRM_MSM) += msm/
 obj-$(CONFIG_DRM_TEGRA) += tegra/
 obj-$(CONFIG_DRM_STI) += sti/
+obj-$(CONFIG_DRM_IMX) += imx/
 obj-y			+= i2c/
 obj-y			+= panel/
 obj-y			+= bridge/
diff --git a/drivers/gpu/drm/README.drm b/drivers/gpu/drm/README.drm
deleted file mode 100644
index b5b332722581..000000000000
--- a/drivers/gpu/drm/README.drm
+++ /dev/null
@@ -1,43 +0,0 @@
-************************************************************
-* For the very latest on DRI development, please see:      *
-*     http://dri.freedesktop.org/                          *
-************************************************************
-
-The Direct Rendering Manager (drm) is a device-independent kernel-level
-device driver that provides support for the XFree86 Direct Rendering
-Infrastructure (DRI).
-
-The DRM supports the Direct Rendering Infrastructure (DRI) in four major
-ways:
-
-    1. The DRM provides synchronized access to the graphics hardware via
-       the use of an optimized two-tiered lock.
-
-    2. The DRM enforces the DRI security policy for access to the graphics
-       hardware by only allowing authenticated X11 clients access to
-       restricted regions of memory.
-
-    3. The DRM provides a generic DMA engine, complete with multiple
-       queues and the ability to detect the need for an OpenGL context
-       switch.
-
-    4. The DRM is extensible via the use of small device-specific modules
-       that rely extensively on the API exported by the DRM module.
-
-
-Documentation on the DRI is available from:
-    http://dri.freedesktop.org/wiki/Documentation
-    http://sourceforge.net/project/showfiles.php?group_id=387
-    http://dri.sourceforge.net/doc/
-
-For specific information about kernel-level support, see:
-
-    The Direct Rendering Manager, Kernel Support for the Direct Rendering
-    Infrastructure
-    http://dri.sourceforge.net/doc/drm_low_level.html
-
-    Hardware Locking for the Direct Rendering Infrastructure
-    http://dri.sourceforge.net/doc/hardware_locking_low_level.html
-
-    A Security Analysis of the Direct Rendering Infrastructure
-    http://dri.sourceforge.net/doc/security_low_level.html
diff --git a/drivers/gpu/drm/amd/amdkfd/Kconfig b/drivers/gpu/drm/amd/amdkfd/Kconfig
new file mode 100644
index 000000000000..8dfac37ff327
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdkfd/Kconfig
@@ -0,0 +1,9 @@
+#
+# Heterogenous system architecture configuration
+#
+
+config HSA_AMD
+	tristate "HSA kernel driver for AMD GPU devices"
+	depends on DRM_RADEON && AMD_IOMMU_V2 && X86_64
+	help
+	  Enable this if you want to use HSA features on AMD GPU devices.
diff --git a/drivers/gpu/drm/amd/amdkfd/Makefile b/drivers/gpu/drm/amd/amdkfd/Makefile
new file mode 100644
index 000000000000..be6246de5091
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdkfd/Makefile
@@ -0,0 +1,14 @@
+#
+# Makefile for Heterogenous System Architecture support for AMD GPU devices
+#
+
+ccflags-y := -Iinclude/drm -Idrivers/gpu/drm/amd/include/
+
+amdkfd-y	:= kfd_module.o kfd_device.o kfd_chardev.o kfd_topology.o \
+		kfd_pasid.o kfd_doorbell.o kfd_flat_memory.o \
+		kfd_process.o kfd_queue.o kfd_mqd_manager.o \
+		kfd_kernel_queue.o kfd_packet_manager.o \
+		kfd_process_queue_manager.o kfd_device_queue_manager.o \
+		kfd_interrupt.o
+
+obj-$(CONFIG_HSA_AMD)	+= amdkfd.o
diff --git a/drivers/gpu/drm/amd/amdkfd/cik_regs.h b/drivers/gpu/drm/amd/amdkfd/cik_regs.h
new file mode 100644
index 000000000000..607fc5ceadbe
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdkfd/cik_regs.h
@@ -0,0 +1,221 @@
+/*
+ * Copyright 2014 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ */
+
+#ifndef CIK_REGS_H
+#define CIK_REGS_H
+
+#define IH_VMID_0_LUT					0x3D40u
+
+#define BIF_DOORBELL_CNTL				0x530Cu
+
+#define	SRBM_GFX_CNTL					0xE44
+#define	PIPEID(x)					((x) << 0)
+#define	MEID(x)						((x) << 2)
+#define	VMID(x)						((x) << 4)
+#define	QUEUEID(x)					((x) << 8)
+
+#define	SQ_CONFIG					0x8C00
+
+#define	SH_MEM_BASES					0x8C28
+/* if PTR32, these are the bases for scratch and lds */
+#define	PRIVATE_BASE(x)					((x) << 0) /* scratch */
+#define	SHARED_BASE(x)					((x) << 16) /* LDS */
+#define	SH_MEM_APE1_BASE				0x8C2C
+/* if PTR32, this is the base location of GPUVM */
+#define	SH_MEM_APE1_LIMIT				0x8C30
+/* if PTR32, this is the upper limit of GPUVM */
+#define	SH_MEM_CONFIG					0x8C34
+#define	PTR32						(1 << 0)
+#define PRIVATE_ATC					(1 << 1)
+#define	ALIGNMENT_MODE(x)				((x) << 2)
+#define	SH_MEM_ALIGNMENT_MODE_DWORD			0
+#define	SH_MEM_ALIGNMENT_MODE_DWORD_STRICT		1
+#define	SH_MEM_ALIGNMENT_MODE_STRICT			2
+#define	SH_MEM_ALIGNMENT_MODE_UNALIGNED			3
+#define	DEFAULT_MTYPE(x)				((x) << 4)
+#define	APE1_MTYPE(x)					((x) << 7)
+
+/* valid for both DEFAULT_MTYPE and APE1_MTYPE */
+#define	MTYPE_CACHED					0
+#define	MTYPE_NONCACHED					3
+
+
+#define SH_STATIC_MEM_CONFIG				0x9604u
+
+#define	TC_CFG_L1_LOAD_POLICY0				0xAC68
+#define	TC_CFG_L1_LOAD_POLICY1				0xAC6C
+#define	TC_CFG_L1_STORE_POLICY				0xAC70
+#define	TC_CFG_L2_LOAD_POLICY0				0xAC74
+#define	TC_CFG_L2_LOAD_POLICY1				0xAC78
+#define	TC_CFG_L2_STORE_POLICY0				0xAC7C
+#define	TC_CFG_L2_STORE_POLICY1				0xAC80
+#define	TC_CFG_L2_ATOMIC_POLICY				0xAC84
+#define	TC_CFG_L1_VOLATILE				0xAC88
+#define	TC_CFG_L2_VOLATILE				0xAC8C
+
+#define CP_PQ_WPTR_POLL_CNTL				0xC20C
+#define	WPTR_POLL_EN					(1 << 31)
+
+#define CPC_INT_CNTL					0xC2D0
+#define CP_ME1_PIPE0_INT_CNTL				0xC214
+#define CP_ME1_PIPE1_INT_CNTL				0xC218
+#define CP_ME1_PIPE2_INT_CNTL				0xC21C
+#define CP_ME1_PIPE3_INT_CNTL				0xC220
+#define CP_ME2_PIPE0_INT_CNTL				0xC224
+#define CP_ME2_PIPE1_INT_CNTL				0xC228
+#define CP_ME2_PIPE2_INT_CNTL				0xC22C
+#define CP_ME2_PIPE3_INT_CNTL				0xC230
+#define DEQUEUE_REQUEST_INT_ENABLE			(1 << 13)
+#define WRM_POLL_TIMEOUT_INT_ENABLE			(1 << 17)
+#define PRIV_REG_INT_ENABLE				(1 << 23)
+#define TIME_STAMP_INT_ENABLE				(1 << 26)
+#define GENERIC2_INT_ENABLE				(1 << 29)
+#define GENERIC1_INT_ENABLE				(1 << 30)
+#define GENERIC0_INT_ENABLE				(1 << 31)
+#define CP_ME1_PIPE0_INT_STATUS				0xC214
+#define CP_ME1_PIPE1_INT_STATUS				0xC218
+#define CP_ME1_PIPE2_INT_STATUS				0xC21C
+#define CP_ME1_PIPE3_INT_STATUS				0xC220
+#define CP_ME2_PIPE0_INT_STATUS				0xC224
+#define CP_ME2_PIPE1_INT_STATUS				0xC228
+#define CP_ME2_PIPE2_INT_STATUS				0xC22C
+#define CP_ME2_PIPE3_INT_STATUS				0xC230
+#define DEQUEUE_REQUEST_INT_STATUS			(1 << 13)
+#define WRM_POLL_TIMEOUT_INT_STATUS			(1 << 17)
+#define PRIV_REG_INT_STATUS				(1 << 23)
+#define TIME_STAMP_INT_STATUS				(1 << 26)
+#define GENERIC2_INT_STATUS				(1 << 29)
+#define GENERIC1_INT_STATUS				(1 << 30)
+#define GENERIC0_INT_STATUS				(1 << 31)
+
+#define CP_HPD_EOP_BASE_ADDR				0xC904
+#define CP_HPD_EOP_BASE_ADDR_HI				0xC908
+#define CP_HPD_EOP_VMID					0xC90C
+#define CP_HPD_EOP_CONTROL				0xC910
+#define	EOP_SIZE(x)					((x) << 0)
+#define	EOP_SIZE_MASK					(0x3f << 0)
+#define CP_MQD_BASE_ADDR				0xC914
+#define CP_MQD_BASE_ADDR_HI				0xC918
+#define CP_HQD_ACTIVE					0xC91C
+#define CP_HQD_VMID					0xC920
+
+#define CP_HQD_PERSISTENT_STATE				0xC924u
+#define	DEFAULT_CP_HQD_PERSISTENT_STATE			(0x33U << 8)
+#define	PRELOAD_REQ					(1 << 0)
+
+#define CP_HQD_PIPE_PRIORITY				0xC928u
+#define CP_HQD_QUEUE_PRIORITY				0xC92Cu
+#define CP_HQD_QUANTUM					0xC930u
+#define	QUANTUM_EN					1U
+#define	QUANTUM_SCALE_1MS				(1U << 4)
+#define	QUANTUM_DURATION(x)				((x) << 8)
+
+#define CP_HQD_PQ_BASE					0xC934
+#define CP_HQD_PQ_BASE_HI				0xC938
+#define CP_HQD_PQ_RPTR					0xC93C
+#define CP_HQD_PQ_RPTR_REPORT_ADDR			0xC940
+#define CP_HQD_PQ_RPTR_REPORT_ADDR_HI			0xC944
+#define CP_HQD_PQ_WPTR_POLL_ADDR			0xC948
+#define CP_HQD_PQ_WPTR_POLL_ADDR_HI			0xC94C
+#define CP_HQD_PQ_DOORBELL_CONTROL			0xC950
+#define	DOORBELL_OFFSET(x)				((x) << 2)
+#define	DOORBELL_OFFSET_MASK				(0x1fffff << 2)
+#define	DOORBELL_SOURCE					(1 << 28)
+#define	DOORBELL_SCHD_HIT				(1 << 29)
+#define	DOORBELL_EN					(1 << 30)
+#define	DOORBELL_HIT					(1 << 31)
+#define CP_HQD_PQ_WPTR					0xC954
+#define CP_HQD_PQ_CONTROL				0xC958
+#define	QUEUE_SIZE(x)					((x) << 0)
+#define	QUEUE_SIZE_MASK					(0x3f << 0)
+#define	RPTR_BLOCK_SIZE(x)				((x) << 8)
+#define	RPTR_BLOCK_SIZE_MASK				(0x3f << 8)
+#define	MIN_AVAIL_SIZE(x)				((x) << 20)
+#define	PQ_ATC_EN					(1 << 23)
+#define	PQ_VOLATILE					(1 << 26)
+#define	NO_UPDATE_RPTR					(1 << 27)
+#define	UNORD_DISPATCH					(1 << 28)
+#define	ROQ_PQ_IB_FLIP					(1 << 29)
+#define	PRIV_STATE					(1 << 30)
+#define	KMD_QUEUE					(1 << 31)
+
+#define	DEFAULT_RPTR_BLOCK_SIZE				RPTR_BLOCK_SIZE(5)
+#define	DEFAULT_MIN_AVAIL_SIZE				MIN_AVAIL_SIZE(3)
+
+#define CP_HQD_IB_BASE_ADDR				0xC95Cu
+#define CP_HQD_IB_BASE_ADDR_HI				0xC960u
+#define CP_HQD_IB_RPTR					0xC964u
+#define CP_HQD_IB_CONTROL				0xC968u
+#define	IB_ATC_EN					(1U << 23)
+#define	DEFAULT_MIN_IB_AVAIL_SIZE			(3U << 20)
+
+#define CP_HQD_DEQUEUE_REQUEST				0xC974
+#define	DEQUEUE_REQUEST_DRAIN				1
+#define DEQUEUE_REQUEST_RESET				2
+#define		DEQUEUE_INT					(1U << 8)
+
+#define CP_HQD_SEMA_CMD					0xC97Cu
+#define CP_HQD_MSG_TYPE					0xC980u
+#define CP_HQD_ATOMIC0_PREOP_LO				0xC984u
+#define CP_HQD_ATOMIC0_PREOP_HI				0xC988u
+#define CP_HQD_ATOMIC1_PREOP_LO				0xC98Cu
+#define CP_HQD_ATOMIC1_PREOP_HI				0xC990u
+#define CP_HQD_HQ_SCHEDULER0				0xC994u
+#define CP_HQD_HQ_SCHEDULER1				0xC998u
+
+
+#define CP_MQD_CONTROL					0xC99C
+#define	MQD_VMID(x)					((x) << 0)
+#define	MQD_VMID_MASK					(0xf << 0)
+#define	MQD_CONTROL_PRIV_STATE_EN			(1U << 8)
+
+#define GRBM_GFX_INDEX					0x30800
+#define	INSTANCE_INDEX(x)				((x) << 0)
+#define	SH_INDEX(x)					((x) << 8)
+#define	SE_INDEX(x)					((x) << 16)
+#define	SH_BROADCAST_WRITES				(1 << 29)
+#define	INSTANCE_BROADCAST_WRITES			(1 << 30)
+#define	SE_BROADCAST_WRITES				(1 << 31)
+
+#define SQC_CACHES					0x30d20
+#define SQC_POLICY					0x8C38u
+#define SQC_VOLATILE					0x8C3Cu
+
+#define CP_PERFMON_CNTL					0x36020
+
+#define ATC_VMID0_PASID_MAPPING				0x339Cu
+#define	ATC_VMID_PASID_MAPPING_UPDATE_STATUS		0x3398u
+#define	ATC_VMID_PASID_MAPPING_VALID			(1U << 31)
+
+#define ATC_VM_APERTURE0_CNTL				0x3310u
+#define	ATS_ACCESS_MODE_NEVER				0
+#define	ATS_ACCESS_MODE_ALWAYS				1
+
+#define ATC_VM_APERTURE0_CNTL2				0x3318u
+#define ATC_VM_APERTURE0_HIGH_ADDR			0x3308u
+#define ATC_VM_APERTURE0_LOW_ADDR			0x3300u
+#define ATC_VM_APERTURE1_CNTL				0x3314u
+#define ATC_VM_APERTURE1_CNTL2				0x331Cu
+#define ATC_VM_APERTURE1_HIGH_ADDR			0x330Cu
+#define ATC_VM_APERTURE1_LOW_ADDR			0x3304u
+
+#endif
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_chardev.c b/drivers/gpu/drm/amd/amdkfd/kfd_chardev.c
new file mode 100644
index 000000000000..fcfdf23e1913
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_chardev.c
@@ -0,0 +1,611 @@
+/*
+ * Copyright 2014 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ */
+
+#include <linux/device.h>
+#include <linux/export.h>
+#include <linux/err.h>
+#include <linux/fs.h>
+#include <linux/sched.h>
+#include <linux/slab.h>
+#include <linux/uaccess.h>
+#include <linux/compat.h>
+#include <uapi/linux/kfd_ioctl.h>
+#include <linux/time.h>
+#include <linux/mm.h>
+#include <uapi/asm-generic/mman-common.h>
+#include <asm/processor.h>
+#include "kfd_priv.h"
+#include "kfd_device_queue_manager.h"
+
+static long kfd_ioctl(struct file *, unsigned int, unsigned long);
+static int kfd_open(struct inode *, struct file *);
+static int kfd_mmap(struct file *, struct vm_area_struct *);
+
+static const char kfd_dev_name[] = "kfd";
+
+static const struct file_operations kfd_fops = {
+	.owner = THIS_MODULE,
+	.unlocked_ioctl = kfd_ioctl,
+	.compat_ioctl = kfd_ioctl,
+	.open = kfd_open,
+	.mmap = kfd_mmap,
+};
+
+static int kfd_char_dev_major = -1;
+static struct class *kfd_class;
+struct device *kfd_device;
+
+int kfd_chardev_init(void)
+{
+	int err = 0;
+
+	kfd_char_dev_major = register_chrdev(0, kfd_dev_name, &kfd_fops);
+	err = kfd_char_dev_major;
+	if (err < 0)
+		goto err_register_chrdev;
+
+	kfd_class = class_create(THIS_MODULE, kfd_dev_name);
+	err = PTR_ERR(kfd_class);
+	if (IS_ERR(kfd_class))
+		goto err_class_create;
+
+	kfd_device = device_create(kfd_class, NULL,
+					MKDEV(kfd_char_dev_major, 0),
+					NULL, kfd_dev_name);
+	err = PTR_ERR(kfd_device);
+	if (IS_ERR(kfd_device))
+		goto err_device_create;
+
+	return 0;
+
+err_device_create:
+	class_destroy(kfd_class);
+err_class_create:
+	unregister_chrdev(kfd_char_dev_major, kfd_dev_name);
+err_register_chrdev:
+	return err;
+}
+
+void kfd_chardev_exit(void)
+{
+	device_destroy(kfd_class, MKDEV(kfd_char_dev_major, 0));
+	class_destroy(kfd_class);
+	unregister_chrdev(kfd_char_dev_major, kfd_dev_name);
+}
+
+struct device *kfd_chardev(void)
+{
+	return kfd_device;
+}
+
+
+static int kfd_open(struct inode *inode, struct file *filep)
+{
+	struct kfd_process *process;
+	bool is_32bit_user_mode;
+
+	if (iminor(inode) != 0)
+		return -ENODEV;
+
+	is_32bit_user_mode = is_compat_task();
+
+	if (is_32bit_user_mode == true) {
+		dev_warn(kfd_device,
+			"Process %d (32-bit) failed to open /dev/kfd\n"
+			"32-bit processes are not supported by amdkfd\n",
+			current->pid);
+		return -EPERM;
+	}
+
+	process = kfd_create_process(current);
+	if (IS_ERR(process))
+		return PTR_ERR(process);
+
+	dev_dbg(kfd_device, "process %d opened, compat mode (32 bit) - %d\n",
+		process->pasid, process->is_32bit_user_mode);
+
+	return 0;
+}
+
+static int kfd_ioctl_get_version(struct file *filep, struct kfd_process *p,
+					void *data)
+{
+	struct kfd_ioctl_get_version_args *args = data;
+	int err = 0;
+
+	args->major_version = KFD_IOCTL_MAJOR_VERSION;
+	args->minor_version = KFD_IOCTL_MINOR_VERSION;
+
+	return err;
+}
+
+static int set_queue_properties_from_user(struct queue_properties *q_properties,
+				struct kfd_ioctl_create_queue_args *args)
+{
+	if (args->queue_percentage > KFD_MAX_QUEUE_PERCENTAGE) {
+		pr_err("kfd: queue percentage must be between 0 to KFD_MAX_QUEUE_PERCENTAGE\n");
+		return -EINVAL;
+	}
+
+	if (args->queue_priority > KFD_MAX_QUEUE_PRIORITY) {
+		pr_err("kfd: queue priority must be between 0 to KFD_MAX_QUEUE_PRIORITY\n");
+		return -EINVAL;
+	}
+
+	if ((args->ring_base_address) &&
+		(!access_ok(VERIFY_WRITE,
+			(const void __user *) args->ring_base_address,
+			sizeof(uint64_t)))) {
+		pr_err("kfd: can't access ring base address\n");
+		return -EFAULT;
+	}
+
+	if (!is_power_of_2(args->ring_size) && (args->ring_size != 0)) {
+		pr_err("kfd: ring size must be a power of 2 or 0\n");
+		return -EINVAL;
+	}
+
+	if (!access_ok(VERIFY_WRITE,
+			(const void __user *) args->read_pointer_address,
+			sizeof(uint32_t))) {
+		pr_err("kfd: can't access read pointer\n");
+		return -EFAULT;
+	}
+
+	if (!access_ok(VERIFY_WRITE,
+			(const void __user *) args->write_pointer_address,
+			sizeof(uint32_t))) {
+		pr_err("kfd: can't access write pointer\n");
+		return -EFAULT;
+	}
+
+	q_properties->is_interop = false;
+	q_properties->queue_percent = args->queue_percentage;
+	q_properties->priority = args->queue_priority;
+	q_properties->queue_address = args->ring_base_address;
+	q_properties->queue_size = args->ring_size;
+	q_properties->read_ptr = (uint32_t *) args->read_pointer_address;
+	q_properties->write_ptr = (uint32_t *) args->write_pointer_address;
+	if (args->queue_type == KFD_IOC_QUEUE_TYPE_COMPUTE ||
+		args->queue_type == KFD_IOC_QUEUE_TYPE_COMPUTE_AQL)
+		q_properties->type = KFD_QUEUE_TYPE_COMPUTE;
+	else
+		return -ENOTSUPP;
+
+	if (args->queue_type == KFD_IOC_QUEUE_TYPE_COMPUTE_AQL)
+		q_properties->format = KFD_QUEUE_FORMAT_AQL;
+	else
+		q_properties->format = KFD_QUEUE_FORMAT_PM4;
+
+	pr_debug("Queue Percentage (%d, %d)\n",
+			q_properties->queue_percent, args->queue_percentage);
+
+	pr_debug("Queue Priority (%d, %d)\n",
+			q_properties->priority, args->queue_priority);
+
+	pr_debug("Queue Address (0x%llX, 0x%llX)\n",
+			q_properties->queue_address, args->ring_base_address);
+
+	pr_debug("Queue Size (0x%llX, %u)\n",
+			q_properties->queue_size, args->ring_size);
+
+	pr_debug("Queue r/w Pointers (0x%llX, 0x%llX)\n",
+			(uint64_t) q_properties->read_ptr,
+			(uint64_t) q_properties->write_ptr);
+
+	pr_debug("Queue Format (%d)\n", q_properties->format);
+
+	return 0;
+}
+
+static int kfd_ioctl_create_queue(struct file *filep, struct kfd_process *p,
+					void *data)
+{
+	struct kfd_ioctl_create_queue_args *args = data;
+	struct kfd_dev *dev;
+	int err = 0;
+	unsigned int queue_id;
+	struct kfd_process_device *pdd;
+	struct queue_properties q_properties;
+
+	memset(&q_properties, 0, sizeof(struct queue_properties));
+
+	pr_debug("kfd: creating queue ioctl\n");
+
+	err = set_queue_properties_from_user(&q_properties, args);
+	if (err)
+		return err;
+
+	dev = kfd_device_by_id(args->gpu_id);
+	if (dev == NULL)
+		return -EINVAL;
+
+	mutex_lock(&p->mutex);
+
+	pdd = kfd_bind_process_to_device(dev, p);
+	if (IS_ERR(pdd)) {
+		err = -ESRCH;
+		goto err_bind_process;
+	}
+
+	pr_debug("kfd: creating queue for PASID %d on GPU 0x%x\n",
+			p->pasid,
+			dev->id);
+
+	err = pqm_create_queue(&p->pqm, dev, filep, &q_properties, 0,
+				KFD_QUEUE_TYPE_COMPUTE, &queue_id);
+	if (err != 0)
+		goto err_create_queue;
+
+	args->queue_id = queue_id;
+
+	/* Return gpu_id as doorbell offset for mmap usage */
+	args->doorbell_offset = args->gpu_id << PAGE_SHIFT;
+
+	mutex_unlock(&p->mutex);
+
+	pr_debug("kfd: queue id %d was created successfully\n", args->queue_id);
+
+	pr_debug("ring buffer address == 0x%016llX\n",
+			args->ring_base_address);
+
+	pr_debug("read ptr address    == 0x%016llX\n",
+			args->read_pointer_address);
+
+	pr_debug("write ptr address   == 0x%016llX\n",
+			args->write_pointer_address);
+
+	return 0;
+
+err_create_queue:
+err_bind_process:
+	mutex_unlock(&p->mutex);
+	return err;
+}
+
+static int kfd_ioctl_destroy_queue(struct file *filp, struct kfd_process *p,
+					void *data)
+{
+	int retval;
+	struct kfd_ioctl_destroy_queue_args *args = data;
+
+	pr_debug("kfd: destroying queue id %d for PASID %d\n",
+				args->queue_id,
+				p->pasid);
+
+	mutex_lock(&p->mutex);
+
+	retval = pqm_destroy_queue(&p->pqm, args->queue_id);
+
+	mutex_unlock(&p->mutex);
+	return retval;
+}
+
+static int kfd_ioctl_update_queue(struct file *filp, struct kfd_process *p,
+					void *data)
+{
+	int retval;
+	struct kfd_ioctl_update_queue_args *args = data;
+	struct queue_properties properties;
+
+	if (args->queue_percentage > KFD_MAX_QUEUE_PERCENTAGE) {
+		pr_err("kfd: queue percentage must be between 0 to KFD_MAX_QUEUE_PERCENTAGE\n");
+		return -EINVAL;
+	}
+
+	if (args->queue_priority > KFD_MAX_QUEUE_PRIORITY) {
+		pr_err("kfd: queue priority must be between 0 to KFD_MAX_QUEUE_PRIORITY\n");
+		return -EINVAL;
+	}
+
+	if ((args->ring_base_address) &&
+		(!access_ok(VERIFY_WRITE,
+			(const void __user *) args->ring_base_address,
+			sizeof(uint64_t)))) {
+		pr_err("kfd: can't access ring base address\n");
+		return -EFAULT;
+	}
+
+	if (!is_power_of_2(args->ring_size) && (args->ring_size != 0)) {
+		pr_err("kfd: ring size must be a power of 2 or 0\n");
+		return -EINVAL;
+	}
+
+	properties.queue_address = args->ring_base_address;
+	properties.queue_size = args->ring_size;
+	properties.queue_percent = args->queue_percentage;
+	properties.priority = args->queue_priority;
+
+	pr_debug("kfd: updating queue id %d for PASID %d\n",
+			args->queue_id, p->pasid);
+
+	mutex_lock(&p->mutex);
+
+	retval = pqm_update_queue(&p->pqm, args->queue_id, &properties);
+
+	mutex_unlock(&p->mutex);
+
+	return retval;
+}
+
+static int kfd_ioctl_set_memory_policy(struct file *filep,
+					struct kfd_process *p, void *data)
+{
+	struct kfd_ioctl_set_memory_policy_args *args = data;
+	struct kfd_dev *dev;
+	int err = 0;
+	struct kfd_process_device *pdd;
+	enum cache_policy default_policy, alternate_policy;
+
+	if (args->default_policy != KFD_IOC_CACHE_POLICY_COHERENT
+	    && args->default_policy != KFD_IOC_CACHE_POLICY_NONCOHERENT) {
+		return -EINVAL;
+	}
+
+	if (args->alternate_policy != KFD_IOC_CACHE_POLICY_COHERENT
+	    && args->alternate_policy != KFD_IOC_CACHE_POLICY_NONCOHERENT) {
+		return -EINVAL;
+	}
+
+	dev = kfd_device_by_id(args->gpu_id);
+	if (dev == NULL)
+		return -EINVAL;
+
+	mutex_lock(&p->mutex);
+
+	pdd = kfd_bind_process_to_device(dev, p);
+	if (IS_ERR(pdd)) {
+		err = -ESRCH;
+		goto out;
+	}
+
+	default_policy = (args->default_policy == KFD_IOC_CACHE_POLICY_COHERENT)
+			 ? cache_policy_coherent : cache_policy_noncoherent;
+
+	alternate_policy =
+		(args->alternate_policy == KFD_IOC_CACHE_POLICY_COHERENT)
+		   ? cache_policy_coherent : cache_policy_noncoherent;
+
+	if (!dev->dqm->set_cache_memory_policy(dev->dqm,
+				&pdd->qpd,
+				default_policy,
+				alternate_policy,
+				(void __user *)args->alternate_aperture_base,
+				args->alternate_aperture_size))
+		err = -EINVAL;
+
+out:
+	mutex_unlock(&p->mutex);
+
+	return err;
+}
+
+static int kfd_ioctl_get_clock_counters(struct file *filep,
+				struct kfd_process *p, void *data)
+{
+	struct kfd_ioctl_get_clock_counters_args *args = data;
+	struct kfd_dev *dev;
+	struct timespec time;
+
+	dev = kfd_device_by_id(args->gpu_id);
+	if (dev == NULL)
+		return -EINVAL;
+
+	/* Reading GPU clock counter from KGD */
+	args->gpu_clock_counter = kfd2kgd->get_gpu_clock_counter(dev->kgd);
+
+	/* No access to rdtsc. Using raw monotonic time */
+	getrawmonotonic(&time);
+	args->cpu_clock_counter = (uint64_t)timespec_to_ns(&time);
+
+	get_monotonic_boottime(&time);
+	args->system_clock_counter = (uint64_t)timespec_to_ns(&time);
+
+	/* Since the counter is in nano-seconds we use 1GHz frequency */
+	args->system_clock_freq = 1000000000;
+
+	return 0;
+}
+
+
+static int kfd_ioctl_get_process_apertures(struct file *filp,
+				struct kfd_process *p, void *data)
+{
+	struct kfd_ioctl_get_process_apertures_args *args = data;
+	struct kfd_process_device_apertures *pAperture;
+	struct kfd_process_device *pdd;
+
+	dev_dbg(kfd_device, "get apertures for PASID %d", p->pasid);
+
+	args->num_of_nodes = 0;
+
+	mutex_lock(&p->mutex);
+
+	/*if the process-device list isn't empty*/
+	if (kfd_has_process_device_data(p)) {
+		/* Run over all pdd of the process */
+		pdd = kfd_get_first_process_device_data(p);
+		do {
+			pAperture =
+				&args->process_apertures[args->num_of_nodes];
+			pAperture->gpu_id = pdd->dev->id;
+			pAperture->lds_base = pdd->lds_base;
+			pAperture->lds_limit = pdd->lds_limit;
+			pAperture->gpuvm_base = pdd->gpuvm_base;
+			pAperture->gpuvm_limit = pdd->gpuvm_limit;
+			pAperture->scratch_base = pdd->scratch_base;
+			pAperture->scratch_limit = pdd->scratch_limit;
+
+			dev_dbg(kfd_device,
+				"node id %u\n", args->num_of_nodes);
+			dev_dbg(kfd_device,
+				"gpu id %u\n", pdd->dev->id);
+			dev_dbg(kfd_device,
+				"lds_base %llX\n", pdd->lds_base);
+			dev_dbg(kfd_device,
+				"lds_limit %llX\n", pdd->lds_limit);
+			dev_dbg(kfd_device,
+				"gpuvm_base %llX\n", pdd->gpuvm_base);
+			dev_dbg(kfd_device,
+				"gpuvm_limit %llX\n", pdd->gpuvm_limit);
+			dev_dbg(kfd_device,
+				"scratch_base %llX\n", pdd->scratch_base);
+			dev_dbg(kfd_device,
+				"scratch_limit %llX\n", pdd->scratch_limit);
+
+			args->num_of_nodes++;
+		} while ((pdd = kfd_get_next_process_device_data(p, pdd)) != NULL &&
+				(args->num_of_nodes < NUM_OF_SUPPORTED_GPUS));
+	}
+
+	mutex_unlock(&p->mutex);
+
+	return 0;
+}
+
+#define AMDKFD_IOCTL_DEF(ioctl, _func, _flags) \
+	[_IOC_NR(ioctl)] = {.cmd = ioctl, .func = _func, .flags = _flags, .cmd_drv = 0, .name = #ioctl}
+
+/** Ioctl table */
+static const struct amdkfd_ioctl_desc amdkfd_ioctls[] = {
+	AMDKFD_IOCTL_DEF(AMDKFD_IOC_GET_VERSION,
+			kfd_ioctl_get_version, 0),
+
+	AMDKFD_IOCTL_DEF(AMDKFD_IOC_CREATE_QUEUE,
+			kfd_ioctl_create_queue, 0),
+
+	AMDKFD_IOCTL_DEF(AMDKFD_IOC_DESTROY_QUEUE,
+			kfd_ioctl_destroy_queue, 0),
+
+	AMDKFD_IOCTL_DEF(AMDKFD_IOC_SET_MEMORY_POLICY,
+			kfd_ioctl_set_memory_policy, 0),
+
+	AMDKFD_IOCTL_DEF(AMDKFD_IOC_GET_CLOCK_COUNTERS,
+			kfd_ioctl_get_clock_counters, 0),
+
+	AMDKFD_IOCTL_DEF(AMDKFD_IOC_GET_PROCESS_APERTURES,
+			kfd_ioctl_get_process_apertures, 0),
+
+	AMDKFD_IOCTL_DEF(AMDKFD_IOC_UPDATE_QUEUE,
+			kfd_ioctl_update_queue, 0),
+};
+
+#define AMDKFD_CORE_IOCTL_COUNT	ARRAY_SIZE(amdkfd_ioctls)
+
+static long kfd_ioctl(struct file *filep, unsigned int cmd, unsigned long arg)
+{
+	struct kfd_process *process;
+	amdkfd_ioctl_t *func;
+	const struct amdkfd_ioctl_desc *ioctl = NULL;
+	unsigned int nr = _IOC_NR(cmd);
+	char stack_kdata[128];
+	char *kdata = NULL;
+	unsigned int usize, asize;
+	int retcode = -EINVAL;
+
+	if (nr >= AMDKFD_CORE_IOCTL_COUNT)
+		goto err_i1;
+
+	if ((nr >= AMDKFD_COMMAND_START) && (nr < AMDKFD_COMMAND_END)) {
+		u32 amdkfd_size;
+
+		ioctl = &amdkfd_ioctls[nr];
+
+		amdkfd_size = _IOC_SIZE(ioctl->cmd);
+		usize = asize = _IOC_SIZE(cmd);
+		if (amdkfd_size > asize)
+			asize = amdkfd_size;
+
+		cmd = ioctl->cmd;
+	} else
+		goto err_i1;
+
+	dev_dbg(kfd_device, "ioctl cmd 0x%x (#%d), arg 0x%lx\n", cmd, nr, arg);
+
+	process = kfd_get_process(current);
+	if (IS_ERR(process)) {
+		dev_dbg(kfd_device, "no process\n");
+		goto err_i1;
+	}
+
+	/* Do not trust userspace, use our own definition */
+	func = ioctl->func;
+
+	if (unlikely(!func)) {
+		dev_dbg(kfd_device, "no function\n");
+		retcode = -EINVAL;
+		goto err_i1;
+	}
+
+	if (cmd & (IOC_IN | IOC_OUT)) {
+		if (asize <= sizeof(stack_kdata)) {
+			kdata = stack_kdata;
+		} else {
+			kdata = kmalloc(asize, GFP_KERNEL);
+			if (!kdata) {
+				retcode = -ENOMEM;
+				goto err_i1;
+			}
+		}
+		if (asize > usize)
+			memset(kdata + usize, 0, asize - usize);
+	}
+
+	if (cmd & IOC_IN) {
+		if (copy_from_user(kdata, (void __user *)arg, usize) != 0) {
+			retcode = -EFAULT;
+			goto err_i1;
+		}
+	} else if (cmd & IOC_OUT) {
+		memset(kdata, 0, usize);
+	}
+
+	retcode = func(filep, process, kdata);
+
+	if (cmd & IOC_OUT)
+		if (copy_to_user((void __user *)arg, kdata, usize) != 0)
+			retcode = -EFAULT;
+
+err_i1:
+	if (!ioctl)
+		dev_dbg(kfd_device, "invalid ioctl: pid=%d, cmd=0x%02x, nr=0x%02x\n",
+			  task_pid_nr(current), cmd, nr);
+
+	if (kdata != stack_kdata)
+		kfree(kdata);
+
+	if (retcode)
+		dev_dbg(kfd_device, "ret = %d\n", retcode);
+
+	return retcode;
+}
+
+static int kfd_mmap(struct file *filp, struct vm_area_struct *vma)
+{
+	struct kfd_process *process;
+
+	process = kfd_get_process(current);
+	if (IS_ERR(process))
+		return PTR_ERR(process);
+
+	return kfd_doorbell_mmap(process, vma);
+}
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_crat.h b/drivers/gpu/drm/amd/amdkfd/kfd_crat.h
new file mode 100644
index 000000000000..a374fa3d3ee6
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_crat.h
@@ -0,0 +1,294 @@
+/*
+ * Copyright 2014 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ */
+
+#ifndef KFD_CRAT_H_INCLUDED
+#define KFD_CRAT_H_INCLUDED
+
+#include <linux/types.h>
+
+#pragma pack(1)
+
+/*
+ * 4CC signature values for the CRAT and CDIT ACPI tables
+ */
+
+#define CRAT_SIGNATURE	"CRAT"
+#define CDIT_SIGNATURE	"CDIT"
+
+/*
+ * Component Resource Association Table (CRAT)
+ */
+
+#define CRAT_OEMID_LENGTH	6
+#define CRAT_OEMTABLEID_LENGTH	8
+#define CRAT_RESERVED_LENGTH	6
+
+#define CRAT_OEMID_64BIT_MASK ((1ULL << (CRAT_OEMID_LENGTH * 8)) - 1)
+
+struct crat_header {
+	uint32_t	signature;
+	uint32_t	length;
+	uint8_t		revision;
+	uint8_t		checksum;
+	uint8_t		oem_id[CRAT_OEMID_LENGTH];
+	uint8_t		oem_table_id[CRAT_OEMTABLEID_LENGTH];
+	uint32_t	oem_revision;
+	uint32_t	creator_id;
+	uint32_t	creator_revision;
+	uint32_t	total_entries;
+	uint16_t	num_domains;
+	uint8_t		reserved[CRAT_RESERVED_LENGTH];
+};
+
+/*
+ * The header structure is immediately followed by total_entries of the
+ * data definitions
+ */
+
+/*
+ * The currently defined subtype entries in the CRAT
+ */
+#define CRAT_SUBTYPE_COMPUTEUNIT_AFFINITY	0
+#define CRAT_SUBTYPE_MEMORY_AFFINITY		1
+#define CRAT_SUBTYPE_CACHE_AFFINITY		2
+#define CRAT_SUBTYPE_TLB_AFFINITY		3
+#define CRAT_SUBTYPE_CCOMPUTE_AFFINITY		4
+#define CRAT_SUBTYPE_IOLINK_AFFINITY		5
+#define CRAT_SUBTYPE_MAX			6
+
+#define CRAT_SIBLINGMAP_SIZE	32
+
+/*
+ * ComputeUnit Affinity structure and definitions
+ */
+#define CRAT_CU_FLAGS_ENABLED		0x00000001
+#define CRAT_CU_FLAGS_HOT_PLUGGABLE	0x00000002
+#define CRAT_CU_FLAGS_CPU_PRESENT	0x00000004
+#define CRAT_CU_FLAGS_GPU_PRESENT	0x00000008
+#define CRAT_CU_FLAGS_IOMMU_PRESENT	0x00000010
+#define CRAT_CU_FLAGS_RESERVED		0xffffffe0
+
+#define CRAT_COMPUTEUNIT_RESERVED_LENGTH 4
+
+struct crat_subtype_computeunit {
+	uint8_t		type;
+	uint8_t		length;
+	uint16_t	reserved;
+	uint32_t	flags;
+	uint32_t	proximity_domain;
+	uint32_t	processor_id_low;
+	uint16_t	num_cpu_cores;
+	uint16_t	num_simd_cores;
+	uint16_t	max_waves_simd;
+	uint16_t	io_count;
+	uint16_t	hsa_capability;
+	uint16_t	lds_size_in_kb;
+	uint8_t		wave_front_size;
+	uint8_t		num_banks;
+	uint16_t	micro_engine_id;
+	uint8_t		num_arrays;
+	uint8_t		num_cu_per_array;
+	uint8_t		num_simd_per_cu;
+	uint8_t		max_slots_scatch_cu;
+	uint8_t		reserved2[CRAT_COMPUTEUNIT_RESERVED_LENGTH];
+};
+
+/*
+ * HSA Memory Affinity structure and definitions
+ */
+#define CRAT_MEM_FLAGS_ENABLED		0x00000001
+#define CRAT_MEM_FLAGS_HOT_PLUGGABLE	0x00000002
+#define CRAT_MEM_FLAGS_NON_VOLATILE	0x00000004
+#define CRAT_MEM_FLAGS_RESERVED		0xfffffff8
+
+#define CRAT_MEMORY_RESERVED_LENGTH 8
+
+struct crat_subtype_memory {
+	uint8_t		type;
+	uint8_t		length;
+	uint16_t	reserved;
+	uint32_t	flags;
+	uint32_t	promixity_domain;
+	uint32_t	base_addr_low;
+	uint32_t	base_addr_high;
+	uint32_t	length_low;
+	uint32_t	length_high;
+	uint32_t	width;
+	uint8_t		reserved2[CRAT_MEMORY_RESERVED_LENGTH];
+};
+
+/*
+ * HSA Cache Affinity structure and definitions
+ */
+#define CRAT_CACHE_FLAGS_ENABLED	0x00000001
+#define CRAT_CACHE_FLAGS_DATA_CACHE	0x00000002
+#define CRAT_CACHE_FLAGS_INST_CACHE	0x00000004
+#define CRAT_CACHE_FLAGS_CPU_CACHE	0x00000008
+#define CRAT_CACHE_FLAGS_SIMD_CACHE	0x00000010
+#define CRAT_CACHE_FLAGS_RESERVED	0xffffffe0
+
+#define CRAT_CACHE_RESERVED_LENGTH 8
+
+struct crat_subtype_cache {
+	uint8_t		type;
+	uint8_t		length;
+	uint16_t	reserved;
+	uint32_t	flags;
+	uint32_t	processor_id_low;
+	uint8_t		sibling_map[CRAT_SIBLINGMAP_SIZE];
+	uint32_t	cache_size;
+	uint8_t		cache_level;
+	uint8_t		lines_per_tag;
+	uint16_t	cache_line_size;
+	uint8_t		associativity;
+	uint8_t		cache_properties;
+	uint16_t	cache_latency;
+	uint8_t		reserved2[CRAT_CACHE_RESERVED_LENGTH];
+};
+
+/*
+ * HSA TLB Affinity structure and definitions
+ */
+#define CRAT_TLB_FLAGS_ENABLED	0x00000001
+#define CRAT_TLB_FLAGS_DATA_TLB	0x00000002
+#define CRAT_TLB_FLAGS_INST_TLB	0x00000004
+#define CRAT_TLB_FLAGS_CPU_TLB	0x00000008
+#define CRAT_TLB_FLAGS_SIMD_TLB	0x00000010
+#define CRAT_TLB_FLAGS_RESERVED	0xffffffe0
+
+#define CRAT_TLB_RESERVED_LENGTH 4
+
+struct crat_subtype_tlb {
+	uint8_t		type;
+	uint8_t		length;
+	uint16_t	reserved;
+	uint32_t	flags;
+	uint32_t	processor_id_low;
+	uint8_t		sibling_map[CRAT_SIBLINGMAP_SIZE];
+	uint32_t	tlb_level;
+	uint8_t		data_tlb_associativity_2mb;
+	uint8_t		data_tlb_size_2mb;
+	uint8_t		instruction_tlb_associativity_2mb;
+	uint8_t		instruction_tlb_size_2mb;
+	uint8_t		data_tlb_associativity_4k;
+	uint8_t		data_tlb_size_4k;
+	uint8_t		instruction_tlb_associativity_4k;
+	uint8_t		instruction_tlb_size_4k;
+	uint8_t		data_tlb_associativity_1gb;
+	uint8_t		data_tlb_size_1gb;
+	uint8_t		instruction_tlb_associativity_1gb;
+	uint8_t		instruction_tlb_size_1gb;
+	uint8_t		reserved2[CRAT_TLB_RESERVED_LENGTH];
+};
+
+/*
+ * HSA CCompute/APU Affinity structure and definitions
+ */
+#define CRAT_CCOMPUTE_FLAGS_ENABLED	0x00000001
+#define CRAT_CCOMPUTE_FLAGS_RESERVED	0xfffffffe
+
+#define CRAT_CCOMPUTE_RESERVED_LENGTH 16
+
+struct crat_subtype_ccompute {
+	uint8_t		type;
+	uint8_t		length;
+	uint16_t	reserved;
+	uint32_t	flags;
+	uint32_t	processor_id_low;
+	uint8_t		sibling_map[CRAT_SIBLINGMAP_SIZE];
+	uint32_t	apu_size;
+	uint8_t		reserved2[CRAT_CCOMPUTE_RESERVED_LENGTH];
+};
+
+/*
+ * HSA IO Link Affinity structure and definitions
+ */
+#define CRAT_IOLINK_FLAGS_ENABLED	0x00000001
+#define CRAT_IOLINK_FLAGS_COHERENCY	0x00000002
+#define CRAT_IOLINK_FLAGS_RESERVED	0xfffffffc
+
+/*
+ * IO interface types
+ */
+#define CRAT_IOLINK_TYPE_UNDEFINED	0
+#define CRAT_IOLINK_TYPE_HYPERTRANSPORT	1
+#define CRAT_IOLINK_TYPE_PCIEXPRESS	2
+#define CRAT_IOLINK_TYPE_OTHER		3
+#define CRAT_IOLINK_TYPE_MAX		255
+
+#define CRAT_IOLINK_RESERVED_LENGTH 24
+
+struct crat_subtype_iolink {
+	uint8_t		type;
+	uint8_t		length;
+	uint16_t	reserved;
+	uint32_t	flags;
+	uint32_t	proximity_domain_from;
+	uint32_t	proximity_domain_to;
+	uint8_t		io_interface_type;
+	uint8_t		version_major;
+	uint16_t	version_minor;
+	uint32_t	minimum_latency;
+	uint32_t	maximum_latency;
+	uint32_t	minimum_bandwidth_mbs;
+	uint32_t	maximum_bandwidth_mbs;
+	uint32_t	recommended_transfer_size;
+	uint8_t		reserved2[CRAT_IOLINK_RESERVED_LENGTH];
+};
+
+/*
+ * HSA generic sub-type header
+ */
+
+#define CRAT_SUBTYPE_FLAGS_ENABLED 0x00000001
+
+struct crat_subtype_generic {
+	uint8_t		type;
+	uint8_t		length;
+	uint16_t	reserved;
+	uint32_t	flags;
+};
+
+/*
+ * Component Locality Distance Information Table (CDIT)
+ */
+#define CDIT_OEMID_LENGTH	6
+#define CDIT_OEMTABLEID_LENGTH	8
+
+struct cdit_header {
+	uint32_t	signature;
+	uint32_t	length;
+	uint8_t		revision;
+	uint8_t		checksum;
+	uint8_t		oem_id[CDIT_OEMID_LENGTH];
+	uint8_t		oem_table_id[CDIT_OEMTABLEID_LENGTH];
+	uint32_t	oem_revision;
+	uint32_t	creator_id;
+	uint32_t	creator_revision;
+	uint32_t	total_entries;
+	uint16_t	num_domains;
+	uint8_t		entry[1];
+};
+
+#pragma pack()
+
+#endif /* KFD_CRAT_H_INCLUDED */
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_device.c b/drivers/gpu/drm/amd/amdkfd/kfd_device.c
new file mode 100644
index 000000000000..43884ebd4303
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_device.c
@@ -0,0 +1,308 @@
+/*
+ * Copyright 2014 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ */
+
+#include <linux/amd-iommu.h>
+#include <linux/bsearch.h>
+#include <linux/pci.h>
+#include <linux/slab.h>
+#include "kfd_priv.h"
+#include "kfd_device_queue_manager.h"
+
+#define MQD_SIZE_ALIGNED 768
+
+static const struct kfd_device_info kaveri_device_info = {
+	.max_pasid_bits = 16,
+	.ih_ring_entry_size = 4 * sizeof(uint32_t),
+	.mqd_size_aligned = MQD_SIZE_ALIGNED
+};
+
+struct kfd_deviceid {
+	unsigned short did;
+	const struct kfd_device_info *device_info;
+};
+
+/* Please keep this sorted by increasing device id. */
+static const struct kfd_deviceid supported_devices[] = {
+	{ 0x1304, &kaveri_device_info },	/* Kaveri */
+	{ 0x1305, &kaveri_device_info },	/* Kaveri */
+	{ 0x1306, &kaveri_device_info },	/* Kaveri */
+	{ 0x1307, &kaveri_device_info },	/* Kaveri */
+	{ 0x1309, &kaveri_device_info },	/* Kaveri */
+	{ 0x130A, &kaveri_device_info },	/* Kaveri */
+	{ 0x130B, &kaveri_device_info },	/* Kaveri */
+	{ 0x130C, &kaveri_device_info },	/* Kaveri */
+	{ 0x130D, &kaveri_device_info },	/* Kaveri */
+	{ 0x130E, &kaveri_device_info },	/* Kaveri */
+	{ 0x130F, &kaveri_device_info },	/* Kaveri */
+	{ 0x1310, &kaveri_device_info },	/* Kaveri */
+	{ 0x1311, &kaveri_device_info },	/* Kaveri */
+	{ 0x1312, &kaveri_device_info },	/* Kaveri */
+	{ 0x1313, &kaveri_device_info },	/* Kaveri */
+	{ 0x1315, &kaveri_device_info },	/* Kaveri */
+	{ 0x1316, &kaveri_device_info },	/* Kaveri */
+	{ 0x1317, &kaveri_device_info },	/* Kaveri */
+	{ 0x1318, &kaveri_device_info },	/* Kaveri */
+	{ 0x131B, &kaveri_device_info },	/* Kaveri */
+	{ 0x131C, &kaveri_device_info },	/* Kaveri */
+	{ 0x131D, &kaveri_device_info },	/* Kaveri */
+};
+
+static const struct kfd_device_info *lookup_device_info(unsigned short did)
+{
+	size_t i;
+
+	for (i = 0; i < ARRAY_SIZE(supported_devices); i++) {
+		if (supported_devices[i].did == did) {
+			BUG_ON(supported_devices[i].device_info == NULL);
+			return supported_devices[i].device_info;
+		}
+	}
+
+	return NULL;
+}
+
+struct kfd_dev *kgd2kfd_probe(struct kgd_dev *kgd, struct pci_dev *pdev)
+{
+	struct kfd_dev *kfd;
+
+	const struct kfd_device_info *device_info =
+					lookup_device_info(pdev->device);
+
+	if (!device_info)
+		return NULL;
+
+	kfd = kzalloc(sizeof(*kfd), GFP_KERNEL);
+	if (!kfd)
+		return NULL;
+
+	kfd->kgd = kgd;
+	kfd->device_info = device_info;
+	kfd->pdev = pdev;
+	kfd->init_complete = false;
+
+	return kfd;
+}
+
+static bool device_iommu_pasid_init(struct kfd_dev *kfd)
+{
+	const u32 required_iommu_flags = AMD_IOMMU_DEVICE_FLAG_ATS_SUP |
+					AMD_IOMMU_DEVICE_FLAG_PRI_SUP |
+					AMD_IOMMU_DEVICE_FLAG_PASID_SUP;
+
+	struct amd_iommu_device_info iommu_info;
+	unsigned int pasid_limit;
+	int err;
+
+	err = amd_iommu_device_info(kfd->pdev, &iommu_info);
+	if (err < 0) {
+		dev_err(kfd_device,
+			"error getting iommu info. is the iommu enabled?\n");
+		return false;
+	}
+
+	if ((iommu_info.flags & required_iommu_flags) != required_iommu_flags) {
+		dev_err(kfd_device, "error required iommu flags ats(%i), pri(%i), pasid(%i)\n",
+		       (iommu_info.flags & AMD_IOMMU_DEVICE_FLAG_ATS_SUP) != 0,
+		       (iommu_info.flags & AMD_IOMMU_DEVICE_FLAG_PRI_SUP) != 0,
+		       (iommu_info.flags & AMD_IOMMU_DEVICE_FLAG_PASID_SUP) != 0);
+		return false;
+	}
+
+	pasid_limit = min_t(unsigned int,
+			(unsigned int)1 << kfd->device_info->max_pasid_bits,
+			iommu_info.max_pasids);
+	/*
+	 * last pasid is used for kernel queues doorbells
+	 * in the future the last pasid might be used for a kernel thread.
+	 */
+	pasid_limit = min_t(unsigned int,
+				pasid_limit,
+				kfd->doorbell_process_limit - 1);
+
+	err = amd_iommu_init_device(kfd->pdev, pasid_limit);
+	if (err < 0) {
+		dev_err(kfd_device, "error initializing iommu device\n");
+		return false;
+	}
+
+	if (!kfd_set_pasid_limit(pasid_limit)) {
+		dev_err(kfd_device, "error setting pasid limit\n");
+		amd_iommu_free_device(kfd->pdev);
+		return false;
+	}
+
+	return true;
+}
+
+static void iommu_pasid_shutdown_callback(struct pci_dev *pdev, int pasid)
+{
+	struct kfd_dev *dev = kfd_device_by_pci_dev(pdev);
+
+	if (dev)
+		kfd_unbind_process_from_device(dev, pasid);
+}
+
+bool kgd2kfd_device_init(struct kfd_dev *kfd,
+			 const struct kgd2kfd_shared_resources *gpu_resources)
+{
+	unsigned int size;
+
+	kfd->shared_resources = *gpu_resources;
+
+	/* calculate max size of mqds needed for queues */
+	size = max_num_of_processes *
+		max_num_of_queues_per_process *
+		kfd->device_info->mqd_size_aligned;
+
+	/* add another 512KB for all other allocations on gart */
+	size += 512 * 1024;
+
+	if (kfd2kgd->init_sa_manager(kfd->kgd, size)) {
+		dev_err(kfd_device,
+			"Error initializing sa manager for device (%x:%x)\n",
+			kfd->pdev->vendor, kfd->pdev->device);
+		goto out;
+	}
+
+	kfd_doorbell_init(kfd);
+
+	if (kfd_topology_add_device(kfd) != 0) {
+		dev_err(kfd_device,
+			"Error adding device (%x:%x) to topology\n",
+			kfd->pdev->vendor, kfd->pdev->device);
+		goto kfd_topology_add_device_error;
+	}
+
+	if (kfd_interrupt_init(kfd)) {
+		dev_err(kfd_device,
+			"Error initializing interrupts for device (%x:%x)\n",
+			kfd->pdev->vendor, kfd->pdev->device);
+		goto kfd_interrupt_error;
+	}
+
+	if (!device_iommu_pasid_init(kfd)) {
+		dev_err(kfd_device,
+			"Error initializing iommuv2 for device (%x:%x)\n",
+			kfd->pdev->vendor, kfd->pdev->device);
+		goto device_iommu_pasid_error;
+	}
+	amd_iommu_set_invalidate_ctx_cb(kfd->pdev,
+						iommu_pasid_shutdown_callback);
+
+	kfd->dqm = device_queue_manager_init(kfd);
+	if (!kfd->dqm) {
+		dev_err(kfd_device,
+			"Error initializing queue manager for device (%x:%x)\n",
+			kfd->pdev->vendor, kfd->pdev->device);
+		goto device_queue_manager_error;
+	}
+
+	if (kfd->dqm->start(kfd->dqm) != 0) {
+		dev_err(kfd_device,
+			"Error starting queuen manager for device (%x:%x)\n",
+			kfd->pdev->vendor, kfd->pdev->device);
+		goto dqm_start_error;
+	}
+
+	kfd->init_complete = true;
+	dev_info(kfd_device, "added device (%x:%x)\n", kfd->pdev->vendor,
+		 kfd->pdev->device);
+
+	pr_debug("kfd: Starting kfd with the following scheduling policy %d\n",
+		sched_policy);
+
+	goto out;
+
+dqm_start_error:
+	device_queue_manager_uninit(kfd->dqm);
+device_queue_manager_error:
+	amd_iommu_free_device(kfd->pdev);
+device_iommu_pasid_error:
+	kfd_interrupt_exit(kfd);
+kfd_interrupt_error:
+	kfd_topology_remove_device(kfd);
+kfd_topology_add_device_error:
+	kfd2kgd->fini_sa_manager(kfd->kgd);
+	dev_err(kfd_device,
+		"device (%x:%x) NOT added due to errors\n",
+		kfd->pdev->vendor, kfd->pdev->device);
+out:
+	return kfd->init_complete;
+}
+
+void kgd2kfd_device_exit(struct kfd_dev *kfd)
+{
+	if (kfd->init_complete) {
+		device_queue_manager_uninit(kfd->dqm);
+		amd_iommu_free_device(kfd->pdev);
+		kfd_interrupt_exit(kfd);
+		kfd_topology_remove_device(kfd);
+	}
+
+	kfree(kfd);
+}
+
+void kgd2kfd_suspend(struct kfd_dev *kfd)
+{
+	BUG_ON(kfd == NULL);
+
+	if (kfd->init_complete) {
+		kfd->dqm->stop(kfd->dqm);
+		amd_iommu_set_invalidate_ctx_cb(kfd->pdev, NULL);
+		amd_iommu_free_device(kfd->pdev);
+	}
+}
+
+int kgd2kfd_resume(struct kfd_dev *kfd)
+{
+	unsigned int pasid_limit;
+	int err;
+
+	BUG_ON(kfd == NULL);
+
+	pasid_limit = kfd_get_pasid_limit();
+
+	if (kfd->init_complete) {
+		err = amd_iommu_init_device(kfd->pdev, pasid_limit);
+		if (err < 0)
+			return -ENXIO;
+		amd_iommu_set_invalidate_ctx_cb(kfd->pdev,
+						iommu_pasid_shutdown_callback);
+		kfd->dqm->start(kfd->dqm);
+	}
+
+	return 0;
+}
+
+/* This is called directly from KGD at ISR. */
+void kgd2kfd_interrupt(struct kfd_dev *kfd, const void *ih_ring_entry)
+{
+	if (kfd->init_complete) {
+		spin_lock(&kfd->interrupt_lock);
+
+		if (kfd->interrupts_active
+		    && enqueue_ih_ring_entry(kfd, ih_ring_entry))
+			schedule_work(&kfd->interrupt_work);
+
+		spin_unlock(&kfd->interrupt_lock);
+	}
+}
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.c b/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.c
new file mode 100644
index 000000000000..9c8961d22360
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.c
@@ -0,0 +1,1086 @@
+/*
+ * Copyright 2014 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#include <linux/slab.h>
+#include <linux/list.h>
+#include <linux/types.h>
+#include <linux/printk.h>
+#include <linux/bitops.h>
+#include "kfd_priv.h"
+#include "kfd_device_queue_manager.h"
+#include "kfd_mqd_manager.h"
+#include "cik_regs.h"
+#include "kfd_kernel_queue.h"
+#include "../../radeon/cik_reg.h"
+
+/* Size of the per-pipe EOP queue */
+#define CIK_HPD_EOP_BYTES_LOG2 11
+#define CIK_HPD_EOP_BYTES (1U << CIK_HPD_EOP_BYTES_LOG2)
+
+static bool is_mem_initialized;
+
+static int init_memory(struct device_queue_manager *dqm);
+static int set_pasid_vmid_mapping(struct device_queue_manager *dqm,
+					unsigned int pasid, unsigned int vmid);
+
+static int create_compute_queue_nocpsch(struct device_queue_manager *dqm,
+					struct queue *q,
+					struct qcm_process_device *qpd);
+static int execute_queues_cpsch(struct device_queue_manager *dqm, bool lock);
+static int destroy_queues_cpsch(struct device_queue_manager *dqm, bool lock);
+
+
+static inline unsigned int get_pipes_num(struct device_queue_manager *dqm)
+{
+	BUG_ON(!dqm || !dqm->dev);
+	return dqm->dev->shared_resources.compute_pipe_count;
+}
+
+static inline unsigned int get_first_pipe(struct device_queue_manager *dqm)
+{
+	BUG_ON(!dqm);
+	return dqm->dev->shared_resources.first_compute_pipe;
+}
+
+static inline unsigned int get_pipes_num_cpsch(void)
+{
+	return PIPE_PER_ME_CP_SCHEDULING;
+}
+
+static inline unsigned int
+get_sh_mem_bases_nybble_64(struct kfd_process_device *pdd)
+{
+	uint32_t nybble;
+
+	nybble = (pdd->lds_base >> 60) & 0x0E;
+
+	return nybble;
+
+}
+
+static inline unsigned int get_sh_mem_bases_32(struct kfd_process_device *pdd)
+{
+	unsigned int shared_base;
+
+	shared_base = (pdd->lds_base >> 16) & 0xFF;
+
+	return shared_base;
+}
+
+static uint32_t compute_sh_mem_bases_64bit(unsigned int top_address_nybble);
+static void init_process_memory(struct device_queue_manager *dqm,
+				struct qcm_process_device *qpd)
+{
+	struct kfd_process_device *pdd;
+	unsigned int temp;
+
+	BUG_ON(!dqm || !qpd);
+
+	pdd = qpd_to_pdd(qpd);
+
+	/* check if sh_mem_config register already configured */
+	if (qpd->sh_mem_config == 0) {
+		qpd->sh_mem_config =
+			ALIGNMENT_MODE(SH_MEM_ALIGNMENT_MODE_UNALIGNED) |
+			DEFAULT_MTYPE(MTYPE_NONCACHED) |
+			APE1_MTYPE(MTYPE_NONCACHED);
+		qpd->sh_mem_ape1_limit = 0;
+		qpd->sh_mem_ape1_base = 0;
+	}
+
+	if (qpd->pqm->process->is_32bit_user_mode) {
+		temp = get_sh_mem_bases_32(pdd);
+		qpd->sh_mem_bases = SHARED_BASE(temp);
+		qpd->sh_mem_config |= PTR32;
+	} else {
+		temp = get_sh_mem_bases_nybble_64(pdd);
+		qpd->sh_mem_bases = compute_sh_mem_bases_64bit(temp);
+	}
+
+	pr_debug("kfd: is32bit process: %d sh_mem_bases nybble: 0x%X and register 0x%X\n",
+		qpd->pqm->process->is_32bit_user_mode, temp, qpd->sh_mem_bases);
+}
+
+static void program_sh_mem_settings(struct device_queue_manager *dqm,
+					struct qcm_process_device *qpd)
+{
+	return kfd2kgd->program_sh_mem_settings(dqm->dev->kgd, qpd->vmid,
+						qpd->sh_mem_config,
+						qpd->sh_mem_ape1_base,
+						qpd->sh_mem_ape1_limit,
+						qpd->sh_mem_bases);
+}
+
+static int allocate_vmid(struct device_queue_manager *dqm,
+			struct qcm_process_device *qpd,
+			struct queue *q)
+{
+	int bit, allocated_vmid;
+
+	if (dqm->vmid_bitmap == 0)
+		return -ENOMEM;
+
+	bit = find_first_bit((unsigned long *)&dqm->vmid_bitmap, CIK_VMID_NUM);
+	clear_bit(bit, (unsigned long *)&dqm->vmid_bitmap);
+
+	/* Kaveri kfd vmid's starts from vmid 8 */
+	allocated_vmid = bit + KFD_VMID_START_OFFSET;
+	pr_debug("kfd: vmid allocation %d\n", allocated_vmid);
+	qpd->vmid = allocated_vmid;
+	q->properties.vmid = allocated_vmid;
+
+	set_pasid_vmid_mapping(dqm, q->process->pasid, q->properties.vmid);
+	program_sh_mem_settings(dqm, qpd);
+
+	return 0;
+}
+
+static void deallocate_vmid(struct device_queue_manager *dqm,
+				struct qcm_process_device *qpd,
+				struct queue *q)
+{
+	int bit = qpd->vmid - KFD_VMID_START_OFFSET;
+
+	/* Release the vmid mapping */
+	set_pasid_vmid_mapping(dqm, 0, qpd->vmid);
+
+	set_bit(bit, (unsigned long *)&dqm->vmid_bitmap);
+	qpd->vmid = 0;
+	q->properties.vmid = 0;
+}
+
+static int create_queue_nocpsch(struct device_queue_manager *dqm,
+				struct queue *q,
+				struct qcm_process_device *qpd,
+				int *allocated_vmid)
+{
+	int retval;
+
+	BUG_ON(!dqm || !q || !qpd || !allocated_vmid);
+
+	pr_debug("kfd: In func %s\n", __func__);
+	print_queue(q);
+
+	mutex_lock(&dqm->lock);
+
+	if (list_empty(&qpd->queues_list)) {
+		retval = allocate_vmid(dqm, qpd, q);
+		if (retval != 0) {
+			mutex_unlock(&dqm->lock);
+			return retval;
+		}
+	}
+	*allocated_vmid = qpd->vmid;
+	q->properties.vmid = qpd->vmid;
+
+	retval = create_compute_queue_nocpsch(dqm, q, qpd);
+
+	if (retval != 0) {
+		if (list_empty(&qpd->queues_list)) {
+			deallocate_vmid(dqm, qpd, q);
+			*allocated_vmid = 0;
+		}
+		mutex_unlock(&dqm->lock);
+		return retval;
+	}
+
+	list_add(&q->list, &qpd->queues_list);
+	dqm->queue_count++;
+
+	mutex_unlock(&dqm->lock);
+	return 0;
+}
+
+static int allocate_hqd(struct device_queue_manager *dqm, struct queue *q)
+{
+	bool set;
+	int pipe, bit;
+
+	set = false;
+
+	for (pipe = dqm->next_pipe_to_allocate; pipe < get_pipes_num(dqm);
+			pipe = (pipe + 1) % get_pipes_num(dqm)) {
+		if (dqm->allocated_queues[pipe] != 0) {
+			bit = find_first_bit(
+				(unsigned long *)&dqm->allocated_queues[pipe],
+				QUEUES_PER_PIPE);
+
+			clear_bit(bit,
+				(unsigned long *)&dqm->allocated_queues[pipe]);
+			q->pipe = pipe;
+			q->queue = bit;
+			set = true;
+			break;
+		}
+	}
+
+	if (set == false)
+		return -EBUSY;
+
+	pr_debug("kfd: DQM %s hqd slot - pipe (%d) queue(%d)\n",
+				__func__, q->pipe, q->queue);
+	/* horizontal hqd allocation */
+	dqm->next_pipe_to_allocate = (pipe + 1) % get_pipes_num(dqm);
+
+	return 0;
+}
+
+static inline void deallocate_hqd(struct device_queue_manager *dqm,
+				struct queue *q)
+{
+	set_bit(q->queue, (unsigned long *)&dqm->allocated_queues[q->pipe]);
+}
+
+static int create_compute_queue_nocpsch(struct device_queue_manager *dqm,
+					struct queue *q,
+					struct qcm_process_device *qpd)
+{
+	int retval;
+	struct mqd_manager *mqd;
+
+	BUG_ON(!dqm || !q || !qpd);
+
+	mqd = dqm->get_mqd_manager(dqm, KFD_MQD_TYPE_CIK_COMPUTE);
+	if (mqd == NULL)
+		return -ENOMEM;
+
+	retval = allocate_hqd(dqm, q);
+	if (retval != 0)
+		return retval;
+
+	retval = mqd->init_mqd(mqd, &q->mqd, &q->mqd_mem_obj,
+				&q->gart_mqd_addr, &q->properties);
+	if (retval != 0) {
+		deallocate_hqd(dqm, q);
+		return retval;
+	}
+
+	pr_debug("kfd: loading mqd to hqd on pipe (%d) queue (%d)\n",
+			q->pipe,
+			q->queue);
+
+	retval = mqd->load_mqd(mqd, q->mqd, q->pipe,
+			q->queue, q->properties.write_ptr);
+	if (retval != 0) {
+		deallocate_hqd(dqm, q);
+		mqd->uninit_mqd(mqd, q->mqd, q->mqd_mem_obj);
+		return retval;
+	}
+
+	return 0;
+}
+
+static int destroy_queue_nocpsch(struct device_queue_manager *dqm,
+				struct qcm_process_device *qpd,
+				struct queue *q)
+{
+	int retval;
+	struct mqd_manager *mqd;
+
+	BUG_ON(!dqm || !q || !q->mqd || !qpd);
+
+	retval = 0;
+
+	pr_debug("kfd: In Func %s\n", __func__);
+
+	mutex_lock(&dqm->lock);
+	mqd = dqm->get_mqd_manager(dqm, KFD_MQD_TYPE_CIK_COMPUTE);
+	if (mqd == NULL) {
+		retval = -ENOMEM;
+		goto out;
+	}
+
+	retval = mqd->destroy_mqd(mqd, q->mqd,
+				KFD_PREEMPT_TYPE_WAVEFRONT,
+				QUEUE_PREEMPT_DEFAULT_TIMEOUT_MS,
+				q->pipe, q->queue);
+
+	if (retval != 0)
+		goto out;
+
+	deallocate_hqd(dqm, q);
+
+	mqd->uninit_mqd(mqd, q->mqd, q->mqd_mem_obj);
+
+	list_del(&q->list);
+	if (list_empty(&qpd->queues_list))
+		deallocate_vmid(dqm, qpd, q);
+	dqm->queue_count--;
+out:
+	mutex_unlock(&dqm->lock);
+	return retval;
+}
+
+static int update_queue(struct device_queue_manager *dqm, struct queue *q)
+{
+	int retval;
+	struct mqd_manager *mqd;
+	bool prev_active = false;
+
+	BUG_ON(!dqm || !q || !q->mqd);
+
+	mutex_lock(&dqm->lock);
+	mqd = dqm->get_mqd_manager(dqm, KFD_MQD_TYPE_CIK_COMPUTE);
+	if (mqd == NULL) {
+		mutex_unlock(&dqm->lock);
+		return -ENOMEM;
+	}
+
+	if (q->properties.is_active == true)
+		prev_active = true;
+
+	/*
+	 *
+	 * check active state vs. the previous state
+	 * and modify counter accordingly
+	 */
+	retval = mqd->update_mqd(mqd, q->mqd, &q->properties);
+	if ((q->properties.is_active == true) && (prev_active == false))
+		dqm->queue_count++;
+	else if ((q->properties.is_active == false) && (prev_active == true))
+		dqm->queue_count--;
+
+	if (sched_policy != KFD_SCHED_POLICY_NO_HWS)
+		retval = execute_queues_cpsch(dqm, false);
+
+	mutex_unlock(&dqm->lock);
+	return retval;
+}
+
+static struct mqd_manager *get_mqd_manager_nocpsch(
+		struct device_queue_manager *dqm, enum KFD_MQD_TYPE type)
+{
+	struct mqd_manager *mqd;
+
+	BUG_ON(!dqm || type >= KFD_MQD_TYPE_MAX);
+
+	pr_debug("kfd: In func %s mqd type %d\n", __func__, type);
+
+	mqd = dqm->mqds[type];
+	if (!mqd) {
+		mqd = mqd_manager_init(type, dqm->dev);
+		if (mqd == NULL)
+			pr_err("kfd: mqd manager is NULL");
+		dqm->mqds[type] = mqd;
+	}
+
+	return mqd;
+}
+
+static int register_process_nocpsch(struct device_queue_manager *dqm,
+					struct qcm_process_device *qpd)
+{
+	struct device_process_node *n;
+
+	BUG_ON(!dqm || !qpd);
+
+	pr_debug("kfd: In func %s\n", __func__);
+
+	n = kzalloc(sizeof(struct device_process_node), GFP_KERNEL);
+	if (!n)
+		return -ENOMEM;
+
+	n->qpd = qpd;
+
+	mutex_lock(&dqm->lock);
+	list_add(&n->list, &dqm->queues);
+
+	init_process_memory(dqm, qpd);
+	dqm->processes_count++;
+
+	mutex_unlock(&dqm->lock);
+
+	return 0;
+}
+
+static int unregister_process_nocpsch(struct device_queue_manager *dqm,
+					struct qcm_process_device *qpd)
+{
+	int retval;
+	struct device_process_node *cur, *next;
+
+	BUG_ON(!dqm || !qpd);
+
+	BUG_ON(!list_empty(&qpd->queues_list));
+
+	pr_debug("kfd: In func %s\n", __func__);
+
+	retval = 0;
+	mutex_lock(&dqm->lock);
+
+	list_for_each_entry_safe(cur, next, &dqm->queues, list) {
+		if (qpd == cur->qpd) {
+			list_del(&cur->list);
+			kfree(cur);
+			dqm->processes_count--;
+			goto out;
+		}
+	}
+	/* qpd not found in dqm list */
+	retval = 1;
+out:
+	mutex_unlock(&dqm->lock);
+	return retval;
+}
+
+static int
+set_pasid_vmid_mapping(struct device_queue_manager *dqm, unsigned int pasid,
+			unsigned int vmid)
+{
+	uint32_t pasid_mapping;
+
+	pasid_mapping = (pasid == 0) ? 0 : (uint32_t)pasid |
+						ATC_VMID_PASID_MAPPING_VALID;
+	return kfd2kgd->set_pasid_vmid_mapping(dqm->dev->kgd, pasid_mapping,
+						vmid);
+}
+
+static uint32_t compute_sh_mem_bases_64bit(unsigned int top_address_nybble)
+{
+	/* In 64-bit mode, we can only control the top 3 bits of the LDS,
+	 * scratch and GPUVM apertures.
+	 * The hardware fills in the remaining 59 bits according to the
+	 * following pattern:
+	 * LDS:		X0000000'00000000 - X0000001'00000000 (4GB)
+	 * Scratch:	X0000001'00000000 - X0000002'00000000 (4GB)
+	 * GPUVM:	Y0010000'00000000 - Y0020000'00000000 (1TB)
+	 *
+	 * (where X/Y is the configurable nybble with the low-bit 0)
+	 *
+	 * LDS and scratch will have the same top nybble programmed in the
+	 * top 3 bits of SH_MEM_BASES.PRIVATE_BASE.
+	 * GPUVM can have a different top nybble programmed in the
+	 * top 3 bits of SH_MEM_BASES.SHARED_BASE.
+	 * We don't bother to support different top nybbles
+	 * for LDS/Scratch and GPUVM.
+	 */
+
+	BUG_ON((top_address_nybble & 1) || top_address_nybble > 0xE ||
+		top_address_nybble == 0);
+
+	return PRIVATE_BASE(top_address_nybble << 12) |
+			SHARED_BASE(top_address_nybble << 12);
+}
+
+static int init_memory(struct device_queue_manager *dqm)
+{
+	int i, retval;
+
+	for (i = 8; i < 16; i++)
+		set_pasid_vmid_mapping(dqm, 0, i);
+
+	retval = kfd2kgd->init_memory(dqm->dev->kgd);
+	if (retval == 0)
+		is_mem_initialized = true;
+	return retval;
+}
+
+
+static int init_pipelines(struct device_queue_manager *dqm,
+			unsigned int pipes_num, unsigned int first_pipe)
+{
+	void *hpdptr;
+	struct mqd_manager *mqd;
+	unsigned int i, err, inx;
+	uint64_t pipe_hpd_addr;
+
+	BUG_ON(!dqm || !dqm->dev);
+
+	pr_debug("kfd: In func %s\n", __func__);
+
+	/*
+	 * Allocate memory for the HPDs. This is hardware-owned per-pipe data.
+	 * The driver never accesses this memory after zeroing it.
+	 * It doesn't even have to be saved/restored on suspend/resume
+	 * because it contains no data when there are no active queues.
+	 */
+
+	err = kfd2kgd->allocate_mem(dqm->dev->kgd,
+				CIK_HPD_EOP_BYTES * pipes_num,
+				PAGE_SIZE,
+				KFD_MEMPOOL_SYSTEM_WRITECOMBINE,
+				(struct kgd_mem **) &dqm->pipeline_mem);
+
+	if (err) {
+		pr_err("kfd: error allocate vidmem num pipes: %d\n",
+			pipes_num);
+		return -ENOMEM;
+	}
+
+	hpdptr = dqm->pipeline_mem->cpu_ptr;
+	dqm->pipelines_addr = dqm->pipeline_mem->gpu_addr;
+
+	memset(hpdptr, 0, CIK_HPD_EOP_BYTES * pipes_num);
+
+	mqd = dqm->get_mqd_manager(dqm, KFD_MQD_TYPE_CIK_COMPUTE);
+	if (mqd == NULL) {
+		kfd2kgd->free_mem(dqm->dev->kgd,
+				(struct kgd_mem *) dqm->pipeline_mem);
+		return -ENOMEM;
+	}
+
+	for (i = 0; i < pipes_num; i++) {
+		inx = i + first_pipe;
+		pipe_hpd_addr = dqm->pipelines_addr + i * CIK_HPD_EOP_BYTES;
+		pr_debug("kfd: pipeline address %llX\n", pipe_hpd_addr);
+		/* = log2(bytes/4)-1 */
+		kfd2kgd->init_pipeline(dqm->dev->kgd, i,
+				CIK_HPD_EOP_BYTES_LOG2 - 3, pipe_hpd_addr);
+	}
+
+	return 0;
+}
+
+
+static int init_scheduler(struct device_queue_manager *dqm)
+{
+	int retval;
+
+	BUG_ON(!dqm);
+
+	pr_debug("kfd: In %s\n", __func__);
+
+	retval = init_pipelines(dqm, get_pipes_num(dqm), KFD_DQM_FIRST_PIPE);
+	if (retval != 0)
+		return retval;
+
+	retval = init_memory(dqm);
+
+	return retval;
+}
+
+static int initialize_nocpsch(struct device_queue_manager *dqm)
+{
+	int i;
+
+	BUG_ON(!dqm);
+
+	pr_debug("kfd: In func %s num of pipes: %d\n",
+			__func__, get_pipes_num(dqm));
+
+	mutex_init(&dqm->lock);
+	INIT_LIST_HEAD(&dqm->queues);
+	dqm->queue_count = dqm->next_pipe_to_allocate = 0;
+	dqm->allocated_queues = kcalloc(get_pipes_num(dqm),
+					sizeof(unsigned int), GFP_KERNEL);
+	if (!dqm->allocated_queues) {
+		mutex_destroy(&dqm->lock);
+		return -ENOMEM;
+	}
+
+	for (i = 0; i < get_pipes_num(dqm); i++)
+		dqm->allocated_queues[i] = (1 << QUEUES_PER_PIPE) - 1;
+
+	dqm->vmid_bitmap = (1 << VMID_PER_DEVICE) - 1;
+
+	init_scheduler(dqm);
+	return 0;
+}
+
+static void uninitialize_nocpsch(struct device_queue_manager *dqm)
+{
+	int i;
+
+	BUG_ON(!dqm);
+
+	BUG_ON(dqm->queue_count > 0 || dqm->processes_count > 0);
+
+	kfree(dqm->allocated_queues);
+	for (i = 0 ; i < KFD_MQD_TYPE_MAX ; i++)
+		kfree(dqm->mqds[i]);
+	mutex_destroy(&dqm->lock);
+	kfd2kgd->free_mem(dqm->dev->kgd,
+			(struct kgd_mem *) dqm->pipeline_mem);
+}
+
+static int start_nocpsch(struct device_queue_manager *dqm)
+{
+	return 0;
+}
+
+static int stop_nocpsch(struct device_queue_manager *dqm)
+{
+	return 0;
+}
+
+/*
+ * Device Queue Manager implementation for cp scheduler
+ */
+
+static int set_sched_resources(struct device_queue_manager *dqm)
+{
+	struct scheduling_resources res;
+	unsigned int queue_num, queue_mask;
+
+	BUG_ON(!dqm);
+
+	pr_debug("kfd: In func %s\n", __func__);
+
+	queue_num = get_pipes_num_cpsch() * QUEUES_PER_PIPE;
+	queue_mask = (1 << queue_num) - 1;
+	res.vmid_mask = (1 << VMID_PER_DEVICE) - 1;
+	res.vmid_mask <<= KFD_VMID_START_OFFSET;
+	res.queue_mask = queue_mask << (get_first_pipe(dqm) * QUEUES_PER_PIPE);
+	res.gws_mask = res.oac_mask = res.gds_heap_base =
+						res.gds_heap_size = 0;
+
+	pr_debug("kfd: scheduling resources:\n"
+			"      vmid mask: 0x%8X\n"
+			"      queue mask: 0x%8llX\n",
+			res.vmid_mask, res.queue_mask);
+
+	return pm_send_set_resources(&dqm->packets, &res);
+}
+
+static int initialize_cpsch(struct device_queue_manager *dqm)
+{
+	int retval;
+
+	BUG_ON(!dqm);
+
+	pr_debug("kfd: In func %s num of pipes: %d\n",
+			__func__, get_pipes_num_cpsch());
+
+	mutex_init(&dqm->lock);
+	INIT_LIST_HEAD(&dqm->queues);
+	dqm->queue_count = dqm->processes_count = 0;
+	dqm->active_runlist = false;
+	retval = init_pipelines(dqm, get_pipes_num(dqm), 0);
+	if (retval != 0)
+		goto fail_init_pipelines;
+
+	return 0;
+
+fail_init_pipelines:
+	mutex_destroy(&dqm->lock);
+	return retval;
+}
+
+static int start_cpsch(struct device_queue_manager *dqm)
+{
+	struct device_process_node *node;
+	int retval;
+
+	BUG_ON(!dqm);
+
+	retval = 0;
+
+	retval = pm_init(&dqm->packets, dqm);
+	if (retval != 0)
+		goto fail_packet_manager_init;
+
+	retval = set_sched_resources(dqm);
+	if (retval != 0)
+		goto fail_set_sched_resources;
+
+	pr_debug("kfd: allocating fence memory\n");
+
+	/* allocate fence memory on the gart */
+	retval = kfd2kgd->allocate_mem(dqm->dev->kgd,
+					sizeof(*dqm->fence_addr),
+					32,
+					KFD_MEMPOOL_SYSTEM_WRITECOMBINE,
+					(struct kgd_mem **) &dqm->fence_mem);
+
+	if (retval != 0)
+		goto fail_allocate_vidmem;
+
+	dqm->fence_addr = dqm->fence_mem->cpu_ptr;
+	dqm->fence_gpu_addr = dqm->fence_mem->gpu_addr;
+
+	list_for_each_entry(node, &dqm->queues, list)
+		if (node->qpd->pqm->process && dqm->dev)
+			kfd_bind_process_to_device(dqm->dev,
+						node->qpd->pqm->process);
+
+	execute_queues_cpsch(dqm, true);
+
+	return 0;
+fail_allocate_vidmem:
+fail_set_sched_resources:
+	pm_uninit(&dqm->packets);
+fail_packet_manager_init:
+	return retval;
+}
+
+static int stop_cpsch(struct device_queue_manager *dqm)
+{
+	struct device_process_node *node;
+	struct kfd_process_device *pdd;
+
+	BUG_ON(!dqm);
+
+	destroy_queues_cpsch(dqm, true);
+
+	list_for_each_entry(node, &dqm->queues, list) {
+		pdd = qpd_to_pdd(node->qpd);
+		pdd->bound = false;
+	}
+	kfd2kgd->free_mem(dqm->dev->kgd,
+			(struct kgd_mem *) dqm->fence_mem);
+	pm_uninit(&dqm->packets);
+
+	return 0;
+}
+
+static int create_kernel_queue_cpsch(struct device_queue_manager *dqm,
+					struct kernel_queue *kq,
+					struct qcm_process_device *qpd)
+{
+	BUG_ON(!dqm || !kq || !qpd);
+
+	pr_debug("kfd: In func %s\n", __func__);
+
+	mutex_lock(&dqm->lock);
+	list_add(&kq->list, &qpd->priv_queue_list);
+	dqm->queue_count++;
+	qpd->is_debug = true;
+	execute_queues_cpsch(dqm, false);
+	mutex_unlock(&dqm->lock);
+
+	return 0;
+}
+
+static void destroy_kernel_queue_cpsch(struct device_queue_manager *dqm,
+					struct kernel_queue *kq,
+					struct qcm_process_device *qpd)
+{
+	BUG_ON(!dqm || !kq);
+
+	pr_debug("kfd: In %s\n", __func__);
+
+	mutex_lock(&dqm->lock);
+	destroy_queues_cpsch(dqm, false);
+	list_del(&kq->list);
+	dqm->queue_count--;
+	qpd->is_debug = false;
+	execute_queues_cpsch(dqm, false);
+	mutex_unlock(&dqm->lock);
+}
+
+static int create_queue_cpsch(struct device_queue_manager *dqm, struct queue *q,
+			struct qcm_process_device *qpd, int *allocate_vmid)
+{
+	int retval;
+	struct mqd_manager *mqd;
+
+	BUG_ON(!dqm || !q || !qpd);
+
+	retval = 0;
+
+	if (allocate_vmid)
+		*allocate_vmid = 0;
+
+	mutex_lock(&dqm->lock);
+
+	mqd = dqm->get_mqd_manager(dqm, KFD_MQD_TYPE_CIK_CP);
+	if (mqd == NULL) {
+		mutex_unlock(&dqm->lock);
+		return -ENOMEM;
+	}
+
+	retval = mqd->init_mqd(mqd, &q->mqd, &q->mqd_mem_obj,
+				&q->gart_mqd_addr, &q->properties);
+	if (retval != 0)
+		goto out;
+
+	list_add(&q->list, &qpd->queues_list);
+	if (q->properties.is_active) {
+		dqm->queue_count++;
+		retval = execute_queues_cpsch(dqm, false);
+	}
+
+out:
+	mutex_unlock(&dqm->lock);
+	return retval;
+}
+
+static int fence_wait_timeout(unsigned int *fence_addr,
+				unsigned int fence_value,
+				unsigned long timeout)
+{
+	BUG_ON(!fence_addr);
+	timeout += jiffies;
+
+	while (*fence_addr != fence_value) {
+		if (time_after(jiffies, timeout)) {
+			pr_err("kfd: qcm fence wait loop timeout expired\n");
+			return -ETIME;
+		}
+		cpu_relax();
+	}
+
+	return 0;
+}
+
+static int destroy_queues_cpsch(struct device_queue_manager *dqm, bool lock)
+{
+	int retval;
+
+	BUG_ON(!dqm);
+
+	retval = 0;
+
+	if (lock)
+		mutex_lock(&dqm->lock);
+	if (dqm->active_runlist == false)
+		goto out;
+	retval = pm_send_unmap_queue(&dqm->packets, KFD_QUEUE_TYPE_COMPUTE,
+			KFD_PREEMPT_TYPE_FILTER_ALL_QUEUES, 0, false, 0);
+	if (retval != 0)
+		goto out;
+
+	*dqm->fence_addr = KFD_FENCE_INIT;
+	pm_send_query_status(&dqm->packets, dqm->fence_gpu_addr,
+				KFD_FENCE_COMPLETED);
+	/* should be timed out */
+	fence_wait_timeout(dqm->fence_addr, KFD_FENCE_COMPLETED,
+				QUEUE_PREEMPT_DEFAULT_TIMEOUT_MS);
+	pm_release_ib(&dqm->packets);
+	dqm->active_runlist = false;
+
+out:
+	if (lock)
+		mutex_unlock(&dqm->lock);
+	return retval;
+}
+
+static int execute_queues_cpsch(struct device_queue_manager *dqm, bool lock)
+{
+	int retval;
+
+	BUG_ON(!dqm);
+
+	if (lock)
+		mutex_lock(&dqm->lock);
+
+	retval = destroy_queues_cpsch(dqm, false);
+	if (retval != 0) {
+		pr_err("kfd: the cp might be in an unrecoverable state due to an unsuccessful queues preemption");
+		goto out;
+	}
+
+	if (dqm->queue_count <= 0 || dqm->processes_count <= 0) {
+		retval = 0;
+		goto out;
+	}
+
+	if (dqm->active_runlist) {
+		retval = 0;
+		goto out;
+	}
+
+	retval = pm_send_runlist(&dqm->packets, &dqm->queues);
+	if (retval != 0) {
+		pr_err("kfd: failed to execute runlist");
+		goto out;
+	}
+	dqm->active_runlist = true;
+
+out:
+	if (lock)
+		mutex_unlock(&dqm->lock);
+	return retval;
+}
+
+static int destroy_queue_cpsch(struct device_queue_manager *dqm,
+				struct qcm_process_device *qpd,
+				struct queue *q)
+{
+	int retval;
+	struct mqd_manager *mqd;
+
+	BUG_ON(!dqm || !qpd || !q);
+
+	retval = 0;
+
+	/* remove queue from list to prevent rescheduling after preemption */
+	mutex_lock(&dqm->lock);
+
+	mqd = dqm->get_mqd_manager(dqm, KFD_MQD_TYPE_CIK_CP);
+	if (!mqd) {
+		retval = -ENOMEM;
+		goto failed;
+	}
+
+	list_del(&q->list);
+	dqm->queue_count--;
+
+	execute_queues_cpsch(dqm, false);
+
+	mqd->uninit_mqd(mqd, q->mqd, q->mqd_mem_obj);
+
+	mutex_unlock(&dqm->lock);
+
+	return 0;
+
+failed:
+	mutex_unlock(&dqm->lock);
+	return retval;
+}
+
+/*
+ * Low bits must be 0000/FFFF as required by HW, high bits must be 0 to
+ * stay in user mode.
+ */
+#define APE1_FIXED_BITS_MASK 0xFFFF80000000FFFFULL
+/* APE1 limit is inclusive and 64K aligned. */
+#define APE1_LIMIT_ALIGNMENT 0xFFFF
+
+static bool set_cache_memory_policy(struct device_queue_manager *dqm,
+				   struct qcm_process_device *qpd,
+				   enum cache_policy default_policy,
+				   enum cache_policy alternate_policy,
+				   void __user *alternate_aperture_base,
+				   uint64_t alternate_aperture_size)
+{
+	uint32_t default_mtype;
+	uint32_t ape1_mtype;
+
+	pr_debug("kfd: In func %s\n", __func__);
+
+	mutex_lock(&dqm->lock);
+
+	if (alternate_aperture_size == 0) {
+		/* base > limit disables APE1 */
+		qpd->sh_mem_ape1_base = 1;
+		qpd->sh_mem_ape1_limit = 0;
+	} else {
+		/*
+		 * In FSA64, APE1_Base[63:0] = { 16{SH_MEM_APE1_BASE[31]},
+		 *			SH_MEM_APE1_BASE[31:0], 0x0000 }
+		 * APE1_Limit[63:0] = { 16{SH_MEM_APE1_LIMIT[31]},
+		 *			SH_MEM_APE1_LIMIT[31:0], 0xFFFF }
+		 * Verify that the base and size parameters can be
+		 * represented in this format and convert them.
+		 * Additionally restrict APE1 to user-mode addresses.
+		 */
+
+		uint64_t base = (uintptr_t)alternate_aperture_base;
+		uint64_t limit = base + alternate_aperture_size - 1;
+
+		if (limit <= base)
+			goto out;
+
+		if ((base & APE1_FIXED_BITS_MASK) != 0)
+			goto out;
+
+		if ((limit & APE1_FIXED_BITS_MASK) != APE1_LIMIT_ALIGNMENT)
+			goto out;
+
+		qpd->sh_mem_ape1_base = base >> 16;
+		qpd->sh_mem_ape1_limit = limit >> 16;
+	}
+
+	default_mtype = (default_policy == cache_policy_coherent) ?
+			MTYPE_NONCACHED :
+			MTYPE_CACHED;
+
+	ape1_mtype = (alternate_policy == cache_policy_coherent) ?
+			MTYPE_NONCACHED :
+			MTYPE_CACHED;
+
+	qpd->sh_mem_config = (qpd->sh_mem_config & PTR32)
+			| ALIGNMENT_MODE(SH_MEM_ALIGNMENT_MODE_UNALIGNED)
+			| DEFAULT_MTYPE(default_mtype)
+			| APE1_MTYPE(ape1_mtype);
+
+	if ((sched_policy == KFD_SCHED_POLICY_NO_HWS) && (qpd->vmid != 0))
+		program_sh_mem_settings(dqm, qpd);
+
+	pr_debug("kfd: sh_mem_config: 0x%x, ape1_base: 0x%x, ape1_limit: 0x%x\n",
+		qpd->sh_mem_config, qpd->sh_mem_ape1_base,
+		qpd->sh_mem_ape1_limit);
+
+	mutex_unlock(&dqm->lock);
+	return true;
+
+out:
+	mutex_unlock(&dqm->lock);
+	return false;
+}
+
+struct device_queue_manager *device_queue_manager_init(struct kfd_dev *dev)
+{
+	struct device_queue_manager *dqm;
+
+	BUG_ON(!dev);
+
+	dqm = kzalloc(sizeof(struct device_queue_manager), GFP_KERNEL);
+	if (!dqm)
+		return NULL;
+
+	dqm->dev = dev;
+	switch (sched_policy) {
+	case KFD_SCHED_POLICY_HWS:
+	case KFD_SCHED_POLICY_HWS_NO_OVERSUBSCRIPTION:
+		/* initialize dqm for cp scheduling */
+		dqm->create_queue = create_queue_cpsch;
+		dqm->initialize = initialize_cpsch;
+		dqm->start = start_cpsch;
+		dqm->stop = stop_cpsch;
+		dqm->destroy_queue = destroy_queue_cpsch;
+		dqm->update_queue = update_queue;
+		dqm->get_mqd_manager = get_mqd_manager_nocpsch;
+		dqm->register_process = register_process_nocpsch;
+		dqm->unregister_process = unregister_process_nocpsch;
+		dqm->uninitialize = uninitialize_nocpsch;
+		dqm->create_kernel_queue = create_kernel_queue_cpsch;
+		dqm->destroy_kernel_queue = destroy_kernel_queue_cpsch;
+		dqm->set_cache_memory_policy = set_cache_memory_policy;
+		break;
+	case KFD_SCHED_POLICY_NO_HWS:
+		/* initialize dqm for no cp scheduling */
+		dqm->start = start_nocpsch;
+		dqm->stop = stop_nocpsch;
+		dqm->create_queue = create_queue_nocpsch;
+		dqm->destroy_queue = destroy_queue_nocpsch;
+		dqm->update_queue = update_queue;
+		dqm->get_mqd_manager = get_mqd_manager_nocpsch;
+		dqm->register_process = register_process_nocpsch;
+		dqm->unregister_process = unregister_process_nocpsch;
+		dqm->initialize = initialize_nocpsch;
+		dqm->uninitialize = uninitialize_nocpsch;
+		dqm->set_cache_memory_policy = set_cache_memory_policy;
+		break;
+	default:
+		BUG();
+		break;
+	}
+
+	if (dqm->initialize(dqm) != 0) {
+		kfree(dqm);
+		return NULL;
+	}
+
+	return dqm;
+}
+
+void device_queue_manager_uninit(struct device_queue_manager *dqm)
+{
+	BUG_ON(!dqm);
+
+	dqm->uninitialize(dqm);
+	kfree(dqm);
+}
+
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.h b/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.h
new file mode 100644
index 000000000000..c3f189e8ae35
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.h
@@ -0,0 +1,146 @@
+/*
+ * Copyright 2014 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#ifndef KFD_DEVICE_QUEUE_MANAGER_H_
+#define KFD_DEVICE_QUEUE_MANAGER_H_
+
+#include <linux/rwsem.h>
+#include <linux/list.h>
+#include "kfd_priv.h"
+#include "kfd_mqd_manager.h"
+
+#define QUEUE_PREEMPT_DEFAULT_TIMEOUT_MS	(500)
+#define QUEUES_PER_PIPE				(8)
+#define PIPE_PER_ME_CP_SCHEDULING		(3)
+#define CIK_VMID_NUM				(8)
+#define KFD_VMID_START_OFFSET			(8)
+#define VMID_PER_DEVICE				CIK_VMID_NUM
+#define KFD_DQM_FIRST_PIPE			(0)
+
+struct device_process_node {
+	struct qcm_process_device *qpd;
+	struct list_head list;
+};
+
+/**
+ * struct device_queue_manager
+ *
+ * @create_queue: Queue creation routine.
+ *
+ * @destroy_queue: Queue destruction routine.
+ *
+ * @update_queue: Queue update routine.
+ *
+ * @get_mqd_manager: Returns the mqd manager according to the mqd type.
+ *
+ * @exeute_queues: Dispatches the queues list to the H/W.
+ *
+ * @register_process: This routine associates a specific process with device.
+ *
+ * @unregister_process: destroys the associations between process to device.
+ *
+ * @initialize: Initializes the pipelines and memory module for that device.
+ *
+ * @start: Initializes the resources/modules the the device needs for queues
+ * execution. This function is called on device initialization and after the
+ * system woke up after suspension.
+ *
+ * @stop: This routine stops execution of all the active queue running on the
+ * H/W and basically this function called on system suspend.
+ *
+ * @uninitialize: Destroys all the device queue manager resources allocated in
+ * initialize routine.
+ *
+ * @create_kernel_queue: Creates kernel queue. Used for debug queue.
+ *
+ * @destroy_kernel_queue: Destroys kernel queue. Used for debug queue.
+ *
+ * @set_cache_memory_policy: Sets memory policy (cached/ non cached) for the
+ * memory apertures.
+ *
+ * This struct is a base class for the kfd queues scheduler in the
+ * device level. The device base class should expose the basic operations
+ * for queue creation and queue destruction. This base class hides the
+ * scheduling mode of the driver and the specific implementation of the
+ * concrete device. This class is the only class in the queues scheduler
+ * that configures the H/W.
+ */
+
+struct device_queue_manager {
+	int	(*create_queue)(struct device_queue_manager *dqm,
+				struct queue *q,
+				struct qcm_process_device *qpd,
+				int *allocate_vmid);
+	int	(*destroy_queue)(struct device_queue_manager *dqm,
+				struct qcm_process_device *qpd,
+				struct queue *q);
+	int	(*update_queue)(struct device_queue_manager *dqm,
+				struct queue *q);
+
+	struct mqd_manager * (*get_mqd_manager)
+					(struct device_queue_manager *dqm,
+					enum KFD_MQD_TYPE type);
+
+	int	(*register_process)(struct device_queue_manager *dqm,
+					struct qcm_process_device *qpd);
+	int	(*unregister_process)(struct device_queue_manager *dqm,
+					struct qcm_process_device *qpd);
+	int	(*initialize)(struct device_queue_manager *dqm);
+	int	(*start)(struct device_queue_manager *dqm);
+	int	(*stop)(struct device_queue_manager *dqm);
+	void	(*uninitialize)(struct device_queue_manager *dqm);
+	int	(*create_kernel_queue)(struct device_queue_manager *dqm,
+					struct kernel_queue *kq,
+					struct qcm_process_device *qpd);
+	void	(*destroy_kernel_queue)(struct device_queue_manager *dqm,
+					struct kernel_queue *kq,
+					struct qcm_process_device *qpd);
+	bool	(*set_cache_memory_policy)(struct device_queue_manager *dqm,
+					   struct qcm_process_device *qpd,
+					   enum cache_policy default_policy,
+					   enum cache_policy alternate_policy,
+					   void __user *alternate_aperture_base,
+					   uint64_t alternate_aperture_size);
+
+
+	struct mqd_manager	*mqds[KFD_MQD_TYPE_MAX];
+	struct packet_manager	packets;
+	struct kfd_dev		*dev;
+	struct mutex		lock;
+	struct list_head	queues;
+	unsigned int		processes_count;
+	unsigned int		queue_count;
+	unsigned int		next_pipe_to_allocate;
+	unsigned int		*allocated_queues;
+	unsigned int		vmid_bitmap;
+	uint64_t		pipelines_addr;
+	struct kfd_mem_obj	*pipeline_mem;
+	uint64_t		fence_gpu_addr;
+	unsigned int		*fence_addr;
+	struct kfd_mem_obj	*fence_mem;
+	bool			active_runlist;
+};
+
+
+
+#endif /* KFD_DEVICE_QUEUE_MANAGER_H_ */
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_doorbell.c b/drivers/gpu/drm/amd/amdkfd/kfd_doorbell.c
new file mode 100644
index 000000000000..b5791a5c7c06
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_doorbell.c
@@ -0,0 +1,256 @@
+/*
+ * Copyright 2014 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ */
+#include "kfd_priv.h"
+#include <linux/mm.h>
+#include <linux/mman.h>
+#include <linux/slab.h>
+#include <linux/io.h>
+
+/*
+ * This extension supports a kernel level doorbells management for
+ * the kernel queues.
+ * Basically the last doorbells page is devoted to kernel queues
+ * and that's assures that any user process won't get access to the
+ * kernel doorbells page
+ */
+static DEFINE_MUTEX(doorbell_mutex);
+static unsigned long doorbell_available_index[
+	DIV_ROUND_UP(KFD_MAX_NUM_OF_QUEUES_PER_PROCESS, BITS_PER_LONG)] = { 0 };
+
+#define KERNEL_DOORBELL_PASID 1
+#define KFD_SIZE_OF_DOORBELL_IN_BYTES 4
+
+/*
+ * Each device exposes a doorbell aperture, a PCI MMIO aperture that
+ * receives 32-bit writes that are passed to queues as wptr values.
+ * The doorbells are intended to be written by applications as part
+ * of queueing work on user-mode queues.
+ * We assign doorbells to applications in PAGE_SIZE-sized and aligned chunks.
+ * We map the doorbell address space into user-mode when a process creates
+ * its first queue on each device.
+ * Although the mapping is done by KFD, it is equivalent to an mmap of
+ * the /dev/kfd with the particular device encoded in the mmap offset.
+ * There will be other uses for mmap of /dev/kfd, so only a range of
+ * offsets (KFD_MMAP_DOORBELL_START-END) is used for doorbells.
+ */
+
+/* # of doorbell bytes allocated for each process. */
+static inline size_t doorbell_process_allocation(void)
+{
+	return roundup(KFD_SIZE_OF_DOORBELL_IN_BYTES *
+			KFD_MAX_NUM_OF_QUEUES_PER_PROCESS,
+			PAGE_SIZE);
+}
+
+/* Doorbell calculations for device init. */
+void kfd_doorbell_init(struct kfd_dev *kfd)
+{
+	size_t doorbell_start_offset;
+	size_t doorbell_aperture_size;
+	size_t doorbell_process_limit;
+
+	/*
+	 * We start with calculations in bytes because the input data might
+	 * only be byte-aligned.
+	 * Only after we have done the rounding can we assume any alignment.
+	 */
+
+	doorbell_start_offset =
+			roundup(kfd->shared_resources.doorbell_start_offset,
+					doorbell_process_allocation());
+
+	doorbell_aperture_size =
+			rounddown(kfd->shared_resources.doorbell_aperture_size,
+					doorbell_process_allocation());
+
+	if (doorbell_aperture_size > doorbell_start_offset)
+		doorbell_process_limit =
+			(doorbell_aperture_size - doorbell_start_offset) /
+						doorbell_process_allocation();
+	else
+		doorbell_process_limit = 0;
+
+	kfd->doorbell_base = kfd->shared_resources.doorbell_physical_address +
+				doorbell_start_offset;
+
+	kfd->doorbell_id_offset = doorbell_start_offset / sizeof(u32);
+	kfd->doorbell_process_limit = doorbell_process_limit - 1;
+
+	kfd->doorbell_kernel_ptr = ioremap(kfd->doorbell_base,
+						doorbell_process_allocation());
+
+	BUG_ON(!kfd->doorbell_kernel_ptr);
+
+	pr_debug("kfd: doorbell initialization:\n");
+	pr_debug("kfd: doorbell base           == 0x%08lX\n",
+			(uintptr_t)kfd->doorbell_base);
+
+	pr_debug("kfd: doorbell_id_offset      == 0x%08lX\n",
+			kfd->doorbell_id_offset);
+
+	pr_debug("kfd: doorbell_process_limit  == 0x%08lX\n",
+			doorbell_process_limit);
+
+	pr_debug("kfd: doorbell_kernel_offset  == 0x%08lX\n",
+			(uintptr_t)kfd->doorbell_base);
+
+	pr_debug("kfd: doorbell aperture size  == 0x%08lX\n",
+			kfd->shared_resources.doorbell_aperture_size);
+
+	pr_debug("kfd: doorbell kernel address == 0x%08lX\n",
+			(uintptr_t)kfd->doorbell_kernel_ptr);
+}
+
+int kfd_doorbell_mmap(struct kfd_process *process, struct vm_area_struct *vma)
+{
+	phys_addr_t address;
+	struct kfd_dev *dev;
+
+	/*
+	 * For simplicitly we only allow mapping of the entire doorbell
+	 * allocation of a single device & process.
+	 */
+	if (vma->vm_end - vma->vm_start != doorbell_process_allocation())
+		return -EINVAL;
+
+	/* Find kfd device according to gpu id */
+	dev = kfd_device_by_id(vma->vm_pgoff);
+	if (dev == NULL)
+		return -EINVAL;
+
+	/* Find if pdd exists for combination of process and gpu id */
+	if (!kfd_get_process_device_data(dev, process, 0))
+		return -EINVAL;
+
+	/* Calculate physical address of doorbell */
+	address = kfd_get_process_doorbells(dev, process);
+
+	vma->vm_flags |= VM_IO | VM_DONTCOPY | VM_DONTEXPAND | VM_NORESERVE |
+				VM_DONTDUMP | VM_PFNMAP;
+
+	vma->vm_page_prot = pgprot_noncached(vma->vm_page_prot);
+
+	pr_debug("kfd: mapping doorbell page in kfd_doorbell_mmap\n"
+		 "     target user address == 0x%08llX\n"
+		 "     physical address    == 0x%08llX\n"
+		 "     vm_flags            == 0x%04lX\n"
+		 "     size                == 0x%04lX\n",
+		 (unsigned long long) vma->vm_start, address, vma->vm_flags,
+		 doorbell_process_allocation());
+
+
+	return io_remap_pfn_range(vma,
+				vma->vm_start,
+				address >> PAGE_SHIFT,
+				doorbell_process_allocation(),
+				vma->vm_page_prot);
+}
+
+
+/* get kernel iomem pointer for a doorbell */
+u32 __iomem *kfd_get_kernel_doorbell(struct kfd_dev *kfd,
+					unsigned int *doorbell_off)
+{
+	u32 inx;
+
+	BUG_ON(!kfd || !doorbell_off);
+
+	mutex_lock(&doorbell_mutex);
+	inx = find_first_zero_bit(doorbell_available_index,
+					KFD_MAX_NUM_OF_QUEUES_PER_PROCESS);
+
+	__set_bit(inx, doorbell_available_index);
+	mutex_unlock(&doorbell_mutex);
+
+	if (inx >= KFD_MAX_NUM_OF_QUEUES_PER_PROCESS)
+		return NULL;
+
+	/*
+	 * Calculating the kernel doorbell offset using "faked" kernel
+	 * pasid that allocated for kernel queues only
+	 */
+	*doorbell_off = KERNEL_DOORBELL_PASID * (doorbell_process_allocation() /
+							sizeof(u32)) + inx;
+
+	pr_debug("kfd: get kernel queue doorbell\n"
+			 "     doorbell offset   == 0x%08d\n"
+			 "     kernel address    == 0x%08lX\n",
+		*doorbell_off, (uintptr_t)(kfd->doorbell_kernel_ptr + inx));
+
+	return kfd->doorbell_kernel_ptr + inx;
+}
+
+void kfd_release_kernel_doorbell(struct kfd_dev *kfd, u32 __iomem *db_addr)
+{
+	unsigned int inx;
+
+	BUG_ON(!kfd || !db_addr);
+
+	inx = (unsigned int)(db_addr - kfd->doorbell_kernel_ptr);
+
+	mutex_lock(&doorbell_mutex);
+	__clear_bit(inx, doorbell_available_index);
+	mutex_unlock(&doorbell_mutex);
+}
+
+inline void write_kernel_doorbell(u32 __iomem *db, u32 value)
+{
+	if (db) {
+		writel(value, db);
+		pr_debug("writing %d to doorbell address 0x%p\n", value, db);
+	}
+}
+
+/*
+ * queue_ids are in the range [0,MAX_PROCESS_QUEUES) and are mapped 1:1
+ * to doorbells with the process's doorbell page
+ */
+unsigned int kfd_queue_id_to_doorbell(struct kfd_dev *kfd,
+					struct kfd_process *process,
+					unsigned int queue_id)
+{
+	/*
+	 * doorbell_id_offset accounts for doorbells taken by KGD.
+	 * pasid * doorbell_process_allocation/sizeof(u32) adjusts
+	 * to the process's doorbells
+	 */
+	return kfd->doorbell_id_offset +
+		process->pasid * (doorbell_process_allocation()/sizeof(u32)) +
+		queue_id;
+}
+
+uint64_t kfd_get_number_elems(struct kfd_dev *kfd)
+{
+	uint64_t num_of_elems = (kfd->shared_resources.doorbell_aperture_size -
+				kfd->shared_resources.doorbell_start_offset) /
+					doorbell_process_allocation() + 1;
+
+	return num_of_elems;
+
+}
+
+phys_addr_t kfd_get_process_doorbells(struct kfd_dev *dev,
+					struct kfd_process *process)
+{
+	return dev->doorbell_base +
+		process->pasid * doorbell_process_allocation();
+}
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_flat_memory.c b/drivers/gpu/drm/amd/amdkfd/kfd_flat_memory.c
new file mode 100644
index 000000000000..e64aa99e5e41
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_flat_memory.c
@@ -0,0 +1,354 @@
+/*
+ * Copyright 2014 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#include <linux/device.h>
+#include <linux/export.h>
+#include <linux/err.h>
+#include <linux/fs.h>
+#include <linux/sched.h>
+#include <linux/slab.h>
+#include <linux/uaccess.h>
+#include <linux/compat.h>
+#include <uapi/linux/kfd_ioctl.h>
+#include <linux/time.h>
+#include "kfd_priv.h"
+#include <linux/mm.h>
+#include <uapi/asm-generic/mman-common.h>
+#include <asm/processor.h>
+
+/*
+ * The primary memory I/O features being added for revisions of gfxip
+ * beyond 7.0 (Kaveri) are:
+ *
+ * Access to ATC/IOMMU mapped memory w/ associated extension of VA to 48b
+ *
+ * “Flat” shader memory access – These are new shader vector memory
+ * operations that do not reference a T#/V# so a “pointer” is what is
+ * sourced from the vector gprs for direct access to memory.
+ * This pointer space has the Shared(LDS) and Private(Scratch) memory
+ * mapped into this pointer space as apertures.
+ * The hardware then determines how to direct the memory request
+ * based on what apertures the request falls in.
+ *
+ * Unaligned support and alignment check
+ *
+ *
+ * System Unified Address - SUA
+ *
+ * The standard usage for GPU virtual addresses are that they are mapped by
+ * a set of page tables we call GPUVM and these page tables are managed by
+ * a combination of vidMM/driver software components.  The current virtual
+ * address (VA) range for GPUVM is 40b.
+ *
+ * As of gfxip7.1 and beyond we’re adding the ability for compute memory
+ * clients (CP/RLC, DMA, SHADER(ifetch, scalar, and vector ops)) to access
+ * the same page tables used by host x86 processors and that are managed by
+ * the operating system. This is via a technique and hardware called ATC/IOMMU.
+ * The GPU has the capability of accessing both the GPUVM and ATC address
+ * spaces for a given VMID (process) simultaneously and we call this feature
+ * system unified address (SUA).
+ *
+ * There are three fundamental address modes of operation for a given VMID
+ * (process) on the GPU:
+ *
+ *	HSA64 – 64b pointers and the default address space is ATC
+ *	HSA32 – 32b pointers and the default address space is ATC
+ *	GPUVM – 64b pointers and the default address space is GPUVM (driver
+ *		model mode)
+ *
+ *
+ * HSA64 - ATC/IOMMU 64b
+ *
+ * A 64b pointer in the AMD64/IA64 CPU architecture is not fully utilized
+ * by the CPU so an AMD CPU can only access the high area
+ * (VA[63:47] == 0x1FFFF) and low area (VA[63:47 == 0) of the address space
+ * so the actual VA carried to translation is 48b.  There is a “hole” in
+ * the middle of the 64b VA space.
+ *
+ * The GPU not only has access to all of the CPU accessible address space via
+ * ATC/IOMMU, but it also has access to the GPUVM address space.  The “system
+ * unified address” feature (SUA) is the mapping of GPUVM and ATC address
+ * spaces into a unified pointer space.  The method we take for 64b mode is
+ * to map the full 40b GPUVM address space into the hole of the 64b address
+ * space.
+
+ * The GPUVM_Base/GPUVM_Limit defines the aperture in the 64b space where we
+ * direct requests to be translated via GPUVM page tables instead of the
+ * IOMMU path.
+ *
+ *
+ * 64b to 49b Address conversion
+ *
+ * Note that there are still significant portions of unused regions (holes)
+ * in the 64b address space even for the GPU.  There are several places in
+ * the pipeline (sw and hw), we wish to compress the 64b virtual address
+ * to a 49b address.  This 49b address is constituted of an “ATC” bit
+ * plus a 48b virtual address.  This 49b address is what is passed to the
+ * translation hardware.  ATC==0 means the 48b address is a GPUVM address
+ * (max of 2^40 – 1) intended to be translated via GPUVM page tables.
+ * ATC==1 means the 48b address is intended to be translated via IOMMU
+ * page tables.
+ *
+ * A 64b pointer is compared to the apertures that are defined (Base/Limit), in
+ * this case the GPUVM aperture (red) is defined and if a pointer falls in this
+ * aperture, we subtract the GPUVM_Base address and set the ATC bit to zero
+ * as part of the 64b to 49b conversion.
+ *
+ * Where this 64b to 49b conversion is done is a function of the usage.
+ * Most GPU memory access is via memory objects where the driver builds
+ * a descriptor which consists of a base address and a memory access by
+ * the GPU usually consists of some kind of an offset or Cartesian coordinate
+ * that references this memory descriptor.  This is the case for shader
+ * instructions that reference the T# or V# constants, or for specified
+ * locations of assets (ex. the shader program location).  In these cases
+ * the driver is what handles the 64b to 49b conversion and the base
+ * address in the descriptor (ex. V# or T# or shader program location)
+ * is defined as a 48b address w/ an ATC bit.  For this usage a given
+ * memory object cannot straddle multiple apertures in the 64b address
+ * space. For example a shader program cannot jump in/out between ATC
+ * and GPUVM space.
+ *
+ * In some cases we wish to pass a 64b pointer to the GPU hardware and
+ * the GPU hw does the 64b to 49b conversion before passing memory
+ * requests to the cache/memory system.  This is the case for the
+ * S_LOAD and FLAT_* shader memory instructions where we have 64b pointers
+ * in scalar and vector GPRs respectively.
+ *
+ * In all cases (no matter where the 64b -> 49b conversion is done), the gfxip
+ * hardware sends a 48b address along w/ an ATC bit, to the memory controller
+ * on the memory request interfaces.
+ *
+ *	<client>_MC_rdreq_atc   // read request ATC bit
+ *
+ *		0 : <client>_MC_rdreq_addr is a GPUVM VA
+ *
+ *		1 : <client>_MC_rdreq_addr is a ATC VA
+ *
+ *
+ * “Spare” aperture (APE1)
+ *
+ * We use the GPUVM aperture to differentiate ATC vs. GPUVM, but we also use
+ * apertures to set the Mtype field for S_LOAD/FLAT_* ops which is input to the
+ * config tables for setting cache policies. The “spare” (APE1) aperture is
+ * motivated by getting a different Mtype from the default.
+ * The default aperture isn’t an actual base/limit aperture; it is just the
+ * address space that doesn’t hit any defined base/limit apertures.
+ * The following diagram is a complete picture of the gfxip7.x SUA apertures.
+ * The APE1 can be placed either below or above
+ * the hole (cannot be in the hole).
+ *
+ *
+ * General Aperture definitions and rules
+ *
+ * An aperture register definition consists of a Base, Limit, Mtype, and
+ * usually an ATC bit indicating which translation tables that aperture uses.
+ * In all cases (for SUA and DUA apertures discussed later), aperture base
+ * and limit definitions are 64KB aligned.
+ *
+ *	<ape>_Base[63:0] = { <ape>_Base_register[63:16], 0x0000 }
+ *
+ *	<ape>_Limit[63:0] = { <ape>_Limit_register[63:16], 0xFFFF }
+ *
+ * The base and limit are considered inclusive to an aperture so being
+ * inside an aperture means (address >= Base) AND (address <= Limit).
+ *
+ * In no case is a payload that straddles multiple apertures expected to work.
+ * For example a load_dword_x4 that starts in one aperture and ends in another,
+ * does not work.  For the vector FLAT_* ops we have detection capability in
+ * the shader for reporting a “memory violation” back to the
+ * SQ block for use in traps.
+ * A memory violation results when an op falls into the hole,
+ * or a payload straddles multiple apertures.  The S_LOAD instruction
+ * does not have this detection.
+ *
+ * Apertures cannot overlap.
+ *
+ *
+ *
+ * HSA32 - ATC/IOMMU 32b
+ *
+ * For HSA32 mode, the pointers are interpreted as 32 bits and use a single GPR
+ * instead of two for the S_LOAD and FLAT_* ops. The entire GPUVM space of 40b
+ * will not fit so there is only partial visibility to the GPUVM
+ * space (defined by the aperture) for S_LOAD and FLAT_* ops.
+ * There is no spare (APE1) aperture for HSA32 mode.
+ *
+ *
+ * GPUVM 64b mode (driver model)
+ *
+ * This mode is related to HSA64 in that the difference really is that
+ * the default aperture is GPUVM (ATC==0) and not ATC space.
+ * We have gfxip7.x hardware that has FLAT_* and S_LOAD support for
+ * SUA GPUVM mode, but does not support HSA32/HSA64.
+ *
+ *
+ * Device Unified Address - DUA
+ *
+ * Device unified address (DUA) is the name of the feature that maps the
+ * Shared(LDS) memory and Private(Scratch) memory into the overall address
+ * space for use by the new FLAT_* vector memory ops.  The Shared and
+ * Private memories are mapped as apertures into the address space,
+ * and the hardware detects when a FLAT_* memory request is to be redirected
+ * to the LDS or Scratch memory when it falls into one of these apertures.
+ * Like the SUA apertures, the Shared/Private apertures are 64KB aligned and
+ * the base/limit is “in” the aperture. For both HSA64 and GPUVM SUA modes,
+ * the Shared/Private apertures are always placed in a limited selection of
+ * options in the hole of the 64b address space. For HSA32 mode, the
+ * Shared/Private apertures can be placed anywhere in the 32b space
+ * except at 0.
+ *
+ *
+ * HSA64 Apertures for FLAT_* vector ops
+ *
+ * For HSA64 SUA mode, the Shared and Private apertures are always placed
+ * in the hole w/ a limited selection of possible locations. The requests
+ * that fall in the private aperture are expanded as a function of the
+ * work-item id (tid) and redirected to the location of the
+ * “hidden private memory”. The hidden private can be placed in either GPUVM
+ * or ATC space. The addresses that fall in the shared aperture are
+ * re-directed to the on-chip LDS memory hardware.
+ *
+ *
+ * HSA32 Apertures for FLAT_* vector ops
+ *
+ * In HSA32 mode, the Private and Shared apertures can be placed anywhere
+ * in the 32b space except at 0 (Private or Shared Base at zero disables
+ * the apertures). If the base address of the apertures are non-zero
+ * (ie apertures exists), the size is always 64KB.
+ *
+ *
+ * GPUVM Apertures for FLAT_* vector ops
+ *
+ * In GPUVM mode, the Shared/Private apertures are specified identically
+ * to HSA64 mode where they are always in the hole at a limited selection
+ * of locations.
+ *
+ *
+ * Aperture Definitions for SUA and DUA
+ *
+ * The interpretation of the aperture register definitions for a given
+ * VMID is a function of the “SUA Mode” which is one of HSA64, HSA32, or
+ * GPUVM64 discussed in previous sections. The mode is first decoded, and
+ * then the remaining register decode is a function of the mode.
+ *
+ *
+ * SUA Mode Decode
+ *
+ * For the S_LOAD and FLAT_* shader operations, the SUA mode is decoded from
+ * the COMPUTE_DISPATCH_INITIATOR:DATA_ATC bit and
+ * the SH_MEM_CONFIG:PTR32 bits.
+ *
+ * COMPUTE_DISPATCH_INITIATOR:DATA_ATC    SH_MEM_CONFIG:PTR32        Mode
+ *
+ * 1                                              0                  HSA64
+ *
+ * 1                                              1                  HSA32
+ *
+ * 0                                              X                 GPUVM64
+ *
+ * In general the hardware will ignore the PTR32 bit and treat
+ * as “0” whenever DATA_ATC = “0”, but sw should set PTR32=0
+ * when DATA_ATC=0.
+ *
+ * The DATA_ATC bit is only set for compute dispatches.
+ * All “Draw” dispatches are hardcoded to GPUVM64 mode
+ * for FLAT_* / S_LOAD operations.
+ */
+
+#define MAKE_GPUVM_APP_BASE(gpu_num) \
+	(((uint64_t)(gpu_num) << 61) + 0x1000000000000L)
+
+#define MAKE_GPUVM_APP_LIMIT(base) \
+	(((uint64_t)(base) & \
+		0xFFFFFF0000000000UL) | 0xFFFFFFFFFFL)
+
+#define MAKE_SCRATCH_APP_BASE(gpu_num) \
+	(((uint64_t)(gpu_num) << 61) + 0x100000000L)
+
+#define MAKE_SCRATCH_APP_LIMIT(base) \
+	(((uint64_t)base & 0xFFFFFFFF00000000UL) | 0xFFFFFFFF)
+
+#define MAKE_LDS_APP_BASE(gpu_num) \
+	(((uint64_t)(gpu_num) << 61) + 0x0)
+#define MAKE_LDS_APP_LIMIT(base) \
+	(((uint64_t)(base) & 0xFFFFFFFF00000000UL) | 0xFFFFFFFF)
+
+int kfd_init_apertures(struct kfd_process *process)
+{
+	uint8_t id  = 0;
+	struct kfd_dev *dev;
+	struct kfd_process_device *pdd;
+
+	/*Iterating over all devices*/
+	while ((dev = kfd_topology_enum_kfd_devices(id)) != NULL &&
+		id < NUM_OF_SUPPORTED_GPUS) {
+
+		pdd = kfd_get_process_device_data(dev, process, 1);
+		if (!pdd)
+			return -1;
+
+		/*
+		 * For 64 bit process aperture will be statically reserved in
+		 * the x86_64 non canonical process address space
+		 * amdkfd doesn't currently support apertures for 32 bit process
+		 */
+		if (process->is_32bit_user_mode) {
+			pdd->lds_base = pdd->lds_limit = 0;
+			pdd->gpuvm_base = pdd->gpuvm_limit = 0;
+			pdd->scratch_base = pdd->scratch_limit = 0;
+		} else {
+			/*
+			 * node id couldn't be 0 - the three MSB bits of
+			 * aperture shoudn't be 0
+			 */
+			pdd->lds_base = MAKE_LDS_APP_BASE(id + 1);
+
+			pdd->lds_limit = MAKE_LDS_APP_LIMIT(pdd->lds_base);
+
+			pdd->gpuvm_base = MAKE_GPUVM_APP_BASE(id + 1);
+
+			pdd->gpuvm_limit =
+					MAKE_GPUVM_APP_LIMIT(pdd->gpuvm_base);
+
+			pdd->scratch_base = MAKE_SCRATCH_APP_BASE(id + 1);
+
+			pdd->scratch_limit =
+				MAKE_SCRATCH_APP_LIMIT(pdd->scratch_base);
+		}
+
+		dev_dbg(kfd_device, "node id %u\n", id);
+		dev_dbg(kfd_device, "gpu id %u\n", pdd->dev->id);
+		dev_dbg(kfd_device, "lds_base %llX\n", pdd->lds_base);
+		dev_dbg(kfd_device, "lds_limit %llX\n", pdd->lds_limit);
+		dev_dbg(kfd_device, "gpuvm_base %llX\n", pdd->gpuvm_base);
+		dev_dbg(kfd_device, "gpuvm_limit %llX\n", pdd->gpuvm_limit);
+		dev_dbg(kfd_device, "scratch_base %llX\n", pdd->scratch_base);
+		dev_dbg(kfd_device, "scratch_limit %llX\n", pdd->scratch_limit);
+
+		id++;
+	}
+
+	return 0;
+}
+
+
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_interrupt.c b/drivers/gpu/drm/amd/amdkfd/kfd_interrupt.c
new file mode 100644
index 000000000000..5b999095a1f7
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_interrupt.c
@@ -0,0 +1,176 @@
+/*
+ * Copyright 2014 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ */
+
+/*
+ * KFD Interrupts.
+ *
+ * AMD GPUs deliver interrupts by pushing an interrupt description onto the
+ * interrupt ring and then sending an interrupt. KGD receives the interrupt
+ * in ISR and sends us a pointer to each new entry on the interrupt ring.
+ *
+ * We generally can't process interrupt-signaled events from ISR, so we call
+ * out to each interrupt client module (currently only the scheduler) to ask if
+ * each interrupt is interesting. If they return true, then it requires further
+ * processing so we copy it to an internal interrupt ring and call each
+ * interrupt client again from a work-queue.
+ *
+ * There's no acknowledgment for the interrupts we use. The hardware simply
+ * queues a new interrupt each time without waiting.
+ *
+ * The fixed-size internal queue means that it's possible for us to lose
+ * interrupts because we have no back-pressure to the hardware.
+ */
+
+#include <linux/slab.h>
+#include <linux/device.h>
+#include "kfd_priv.h"
+
+#define KFD_INTERRUPT_RING_SIZE 256
+
+static void interrupt_wq(struct work_struct *);
+
+int kfd_interrupt_init(struct kfd_dev *kfd)
+{
+	void *interrupt_ring = kmalloc_array(KFD_INTERRUPT_RING_SIZE,
+					kfd->device_info->ih_ring_entry_size,
+					GFP_KERNEL);
+	if (!interrupt_ring)
+		return -ENOMEM;
+
+	kfd->interrupt_ring = interrupt_ring;
+	kfd->interrupt_ring_size =
+		KFD_INTERRUPT_RING_SIZE * kfd->device_info->ih_ring_entry_size;
+	atomic_set(&kfd->interrupt_ring_wptr, 0);
+	atomic_set(&kfd->interrupt_ring_rptr, 0);
+
+	spin_lock_init(&kfd->interrupt_lock);
+
+	INIT_WORK(&kfd->interrupt_work, interrupt_wq);
+
+	kfd->interrupts_active = true;
+
+	/*
+	 * After this function returns, the interrupt will be enabled. This
+	 * barrier ensures that the interrupt running on a different processor
+	 * sees all the above writes.
+	 */
+	smp_wmb();
+
+	return 0;
+}
+
+void kfd_interrupt_exit(struct kfd_dev *kfd)
+{
+	/*
+	 * Stop the interrupt handler from writing to the ring and scheduling
+	 * workqueue items. The spinlock ensures that any interrupt running
+	 * after we have unlocked sees interrupts_active = false.
+	 */
+	unsigned long flags;
+
+	spin_lock_irqsave(&kfd->interrupt_lock, flags);
+	kfd->interrupts_active = false;
+	spin_unlock_irqrestore(&kfd->interrupt_lock, flags);
+
+	/*
+	 * Flush_scheduled_work ensures that there are no outstanding
+	 * work-queue items that will access interrupt_ring. New work items
+	 * can't be created because we stopped interrupt handling above.
+	 */
+	flush_scheduled_work();
+
+	kfree(kfd->interrupt_ring);
+}
+
+/*
+ * This assumes that it can't be called concurrently with itself
+ * but only with dequeue_ih_ring_entry.
+ */
+bool enqueue_ih_ring_entry(struct kfd_dev *kfd,	const void *ih_ring_entry)
+{
+	unsigned int rptr = atomic_read(&kfd->interrupt_ring_rptr);
+	unsigned int wptr = atomic_read(&kfd->interrupt_ring_wptr);
+
+	if ((rptr - wptr) % kfd->interrupt_ring_size ==
+					kfd->device_info->ih_ring_entry_size) {
+		/* This is very bad, the system is likely to hang. */
+		dev_err_ratelimited(kfd_chardev(),
+			"Interrupt ring overflow, dropping interrupt.\n");
+		return false;
+	}
+
+	memcpy(kfd->interrupt_ring + wptr, ih_ring_entry,
+			kfd->device_info->ih_ring_entry_size);
+
+	wptr = (wptr + kfd->device_info->ih_ring_entry_size) %
+			kfd->interrupt_ring_size;
+	smp_wmb(); /* Ensure memcpy'd data is visible before wptr update. */
+	atomic_set(&kfd->interrupt_ring_wptr, wptr);
+
+	return true;
+}
+
+/*
+ * This assumes that it can't be called concurrently with itself
+ * but only with enqueue_ih_ring_entry.
+ */
+static bool dequeue_ih_ring_entry(struct kfd_dev *kfd, void *ih_ring_entry)
+{
+	/*
+	 * Assume that wait queues have an implicit barrier, i.e. anything that
+	 * happened in the ISR before it queued work is visible.
+	 */
+
+	unsigned int wptr = atomic_read(&kfd->interrupt_ring_wptr);
+	unsigned int rptr = atomic_read(&kfd->interrupt_ring_rptr);
+
+	if (rptr == wptr)
+		return false;
+
+	memcpy(ih_ring_entry, kfd->interrupt_ring + rptr,
+			kfd->device_info->ih_ring_entry_size);
+
+	rptr = (rptr + kfd->device_info->ih_ring_entry_size) %
+			kfd->interrupt_ring_size;
+
+	/*
+	 * Ensure the rptr write update is not visible until
+	 * memcpy has finished reading.
+	 */
+	smp_mb();
+	atomic_set(&kfd->interrupt_ring_rptr, rptr);
+
+	return true;
+}
+
+static void interrupt_wq(struct work_struct *work)
+{
+	struct kfd_dev *dev = container_of(work, struct kfd_dev,
+						interrupt_work);
+
+	uint32_t ih_ring_entry[DIV_ROUND_UP(
+				dev->device_info->ih_ring_entry_size,
+				sizeof(uint32_t))];
+
+	while (dequeue_ih_ring_entry(dev, ih_ring_entry))
+		;
+}
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_kernel_queue.c b/drivers/gpu/drm/amd/amdkfd/kfd_kernel_queue.c
new file mode 100644
index 000000000000..935071410724
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_kernel_queue.c
@@ -0,0 +1,353 @@
+/*
+ * Copyright 2014 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#include <linux/types.h>
+#include <linux/mutex.h>
+#include <linux/slab.h>
+#include <linux/printk.h>
+#include <linux/sched.h>
+#include "kfd_kernel_queue.h"
+#include "kfd_priv.h"
+#include "kfd_device_queue_manager.h"
+#include "kfd_pm4_headers.h"
+#include "kfd_pm4_opcodes.h"
+
+#define PM4_COUNT_ZERO (((1 << 15) - 1) << 16)
+
+static bool initialize(struct kernel_queue *kq, struct kfd_dev *dev,
+		enum kfd_queue_type type, unsigned int queue_size)
+{
+	struct queue_properties prop;
+	int retval;
+	union PM4_MES_TYPE_3_HEADER nop;
+
+	BUG_ON(!kq || !dev);
+	BUG_ON(type != KFD_QUEUE_TYPE_DIQ && type != KFD_QUEUE_TYPE_HIQ);
+
+	pr_debug("kfd: In func %s initializing queue type %d size %d\n",
+			__func__, KFD_QUEUE_TYPE_HIQ, queue_size);
+
+	nop.opcode = IT_NOP;
+	nop.type = PM4_TYPE_3;
+	nop.u32all |= PM4_COUNT_ZERO;
+
+	kq->dev = dev;
+	kq->nop_packet = nop.u32all;
+	switch (type) {
+	case KFD_QUEUE_TYPE_DIQ:
+	case KFD_QUEUE_TYPE_HIQ:
+		kq->mqd = dev->dqm->get_mqd_manager(dev->dqm,
+						KFD_MQD_TYPE_CIK_HIQ);
+		break;
+	default:
+		BUG();
+		break;
+	}
+
+	if (kq->mqd == NULL)
+		return false;
+
+	prop.doorbell_ptr = kfd_get_kernel_doorbell(dev, &prop.doorbell_off);
+
+	if (prop.doorbell_ptr == NULL)
+		goto err_get_kernel_doorbell;
+
+	retval = kfd2kgd->allocate_mem(dev->kgd,
+					queue_size,
+					PAGE_SIZE,
+					KFD_MEMPOOL_SYSTEM_WRITECOMBINE,
+					(struct kgd_mem **) &kq->pq);
+
+	if (retval != 0)
+		goto err_pq_allocate_vidmem;
+
+	kq->pq_kernel_addr = kq->pq->cpu_ptr;
+	kq->pq_gpu_addr = kq->pq->gpu_addr;
+
+	retval = kfd2kgd->allocate_mem(dev->kgd,
+					sizeof(*kq->rptr_kernel),
+					32,
+					KFD_MEMPOOL_SYSTEM_WRITECOMBINE,
+					(struct kgd_mem **) &kq->rptr_mem);
+
+	if (retval != 0)
+		goto err_rptr_allocate_vidmem;
+
+	kq->rptr_kernel = kq->rptr_mem->cpu_ptr;
+	kq->rptr_gpu_addr = kq->rptr_mem->gpu_addr;
+
+	retval = kfd2kgd->allocate_mem(dev->kgd,
+					sizeof(*kq->wptr_kernel),
+					32,
+					KFD_MEMPOOL_SYSTEM_WRITECOMBINE,
+					(struct kgd_mem **) &kq->wptr_mem);
+
+	if (retval != 0)
+		goto err_wptr_allocate_vidmem;
+
+	kq->wptr_kernel = kq->wptr_mem->cpu_ptr;
+	kq->wptr_gpu_addr = kq->wptr_mem->gpu_addr;
+
+	memset(kq->pq_kernel_addr, 0, queue_size);
+	memset(kq->rptr_kernel, 0, sizeof(*kq->rptr_kernel));
+	memset(kq->wptr_kernel, 0, sizeof(*kq->wptr_kernel));
+
+	prop.queue_size = queue_size;
+	prop.is_interop = false;
+	prop.priority = 1;
+	prop.queue_percent = 100;
+	prop.type = type;
+	prop.vmid = 0;
+	prop.queue_address = kq->pq_gpu_addr;
+	prop.read_ptr = (uint32_t *) kq->rptr_gpu_addr;
+	prop.write_ptr = (uint32_t *) kq->wptr_gpu_addr;
+
+	if (init_queue(&kq->queue, prop) != 0)
+		goto err_init_queue;
+
+	kq->queue->device = dev;
+	kq->queue->process = kfd_get_process(current);
+
+	retval = kq->mqd->init_mqd(kq->mqd, &kq->queue->mqd,
+					&kq->queue->mqd_mem_obj,
+					&kq->queue->gart_mqd_addr,
+					&kq->queue->properties);
+	if (retval != 0)
+		goto err_init_mqd;
+
+	/* assign HIQ to HQD */
+	if (type == KFD_QUEUE_TYPE_HIQ) {
+		pr_debug("assigning hiq to hqd\n");
+		kq->queue->pipe = KFD_CIK_HIQ_PIPE;
+		kq->queue->queue = KFD_CIK_HIQ_QUEUE;
+		kq->mqd->load_mqd(kq->mqd, kq->queue->mqd, kq->queue->pipe,
+					kq->queue->queue, NULL);
+	} else {
+		/* allocate fence for DIQ */
+
+		retval = kfd2kgd->allocate_mem(dev->kgd,
+					sizeof(uint32_t),
+					32,
+					KFD_MEMPOOL_SYSTEM_WRITECOMBINE,
+					(struct kgd_mem **) &kq->fence_mem_obj);
+
+		if (retval != 0)
+			goto err_alloc_fence;
+
+		kq->fence_kernel_address = kq->fence_mem_obj->cpu_ptr;
+		kq->fence_gpu_addr = kq->fence_mem_obj->gpu_addr;
+	}
+
+	print_queue(kq->queue);
+
+	return true;
+err_alloc_fence:
+err_init_mqd:
+	uninit_queue(kq->queue);
+err_init_queue:
+	kfd2kgd->free_mem(dev->kgd, (struct kgd_mem *) kq->wptr_mem);
+err_wptr_allocate_vidmem:
+	kfd2kgd->free_mem(dev->kgd, (struct kgd_mem *) kq->rptr_mem);
+err_rptr_allocate_vidmem:
+	kfd2kgd->free_mem(dev->kgd, (struct kgd_mem *) kq->pq);
+err_pq_allocate_vidmem:
+	pr_err("kfd: error init pq\n");
+	kfd_release_kernel_doorbell(dev, prop.doorbell_ptr);
+err_get_kernel_doorbell:
+	pr_err("kfd: error init doorbell");
+	return false;
+
+}
+
+static void uninitialize(struct kernel_queue *kq)
+{
+	BUG_ON(!kq);
+
+	if (kq->queue->properties.type == KFD_QUEUE_TYPE_HIQ)
+		kq->mqd->destroy_mqd(kq->mqd,
+					NULL,
+					false,
+					QUEUE_PREEMPT_DEFAULT_TIMEOUT_MS,
+					kq->queue->pipe,
+					kq->queue->queue);
+
+	kfd2kgd->free_mem(kq->dev->kgd, (struct kgd_mem *) kq->rptr_mem);
+	kfd2kgd->free_mem(kq->dev->kgd, (struct kgd_mem *) kq->wptr_mem);
+	kfd2kgd->free_mem(kq->dev->kgd, (struct kgd_mem *) kq->pq);
+	kfd_release_kernel_doorbell(kq->dev,
+					kq->queue->properties.doorbell_ptr);
+	uninit_queue(kq->queue);
+}
+
+static int acquire_packet_buffer(struct kernel_queue *kq,
+		size_t packet_size_in_dwords, unsigned int **buffer_ptr)
+{
+	size_t available_size;
+	size_t queue_size_dwords;
+	uint32_t wptr, rptr;
+	unsigned int *queue_address;
+
+	BUG_ON(!kq || !buffer_ptr);
+
+	rptr = *kq->rptr_kernel;
+	wptr = *kq->wptr_kernel;
+	queue_address = (unsigned int *)kq->pq_kernel_addr;
+	queue_size_dwords = kq->queue->properties.queue_size / sizeof(uint32_t);
+
+	pr_debug("kfd: In func %s\nrptr: %d\nwptr: %d\nqueue_address 0x%p\n",
+			__func__, rptr, wptr, queue_address);
+
+	available_size = (rptr - 1 - wptr + queue_size_dwords) %
+							queue_size_dwords;
+
+	if (packet_size_in_dwords >= queue_size_dwords ||
+			packet_size_in_dwords >= available_size) {
+		/*
+		 * make sure calling functions know
+		 * acquire_packet_buffer() failed
+		 */
+		*buffer_ptr = NULL;
+		return -ENOMEM;
+	}
+
+	if (wptr + packet_size_in_dwords >= queue_size_dwords) {
+		while (wptr > 0) {
+			queue_address[wptr] = kq->nop_packet;
+			wptr = (wptr + 1) % queue_size_dwords;
+		}
+	}
+
+	*buffer_ptr = &queue_address[wptr];
+	kq->pending_wptr = wptr + packet_size_in_dwords;
+
+	return 0;
+}
+
+static void submit_packet(struct kernel_queue *kq)
+{
+#ifdef DEBUG
+	int i;
+#endif
+
+	BUG_ON(!kq);
+
+#ifdef DEBUG
+	for (i = *kq->wptr_kernel; i < kq->pending_wptr; i++) {
+		pr_debug("0x%2X ", kq->pq_kernel_addr[i]);
+		if (i % 15 == 0)
+			pr_debug("\n");
+	}
+	pr_debug("\n");
+#endif
+
+	*kq->wptr_kernel = kq->pending_wptr;
+	write_kernel_doorbell(kq->queue->properties.doorbell_ptr,
+				kq->pending_wptr);
+}
+
+static int sync_with_hw(struct kernel_queue *kq, unsigned long timeout_ms)
+{
+	unsigned long org_timeout_ms;
+
+	BUG_ON(!kq);
+
+	org_timeout_ms = timeout_ms;
+	timeout_ms += jiffies * 1000 / HZ;
+	while (*kq->wptr_kernel != *kq->rptr_kernel) {
+		if (time_after(jiffies * 1000 / HZ, timeout_ms)) {
+			pr_err("kfd: kernel_queue %s timeout expired %lu\n",
+				__func__, org_timeout_ms);
+			pr_err("kfd: wptr: %d rptr: %d\n",
+				*kq->wptr_kernel, *kq->rptr_kernel);
+			return -ETIME;
+		}
+		schedule();
+	}
+
+	return 0;
+}
+
+static void rollback_packet(struct kernel_queue *kq)
+{
+	BUG_ON(!kq);
+	kq->pending_wptr = *kq->queue->properties.write_ptr;
+}
+
+struct kernel_queue *kernel_queue_init(struct kfd_dev *dev,
+					enum kfd_queue_type type)
+{
+	struct kernel_queue *kq;
+
+	BUG_ON(!dev);
+
+	kq = kzalloc(sizeof(struct kernel_queue), GFP_KERNEL);
+	if (!kq)
+		return NULL;
+
+	kq->initialize = initialize;
+	kq->uninitialize = uninitialize;
+	kq->acquire_packet_buffer = acquire_packet_buffer;
+	kq->submit_packet = submit_packet;
+	kq->sync_with_hw = sync_with_hw;
+	kq->rollback_packet = rollback_packet;
+
+	if (kq->initialize(kq, dev, type, KFD_KERNEL_QUEUE_SIZE) == false) {
+		pr_err("kfd: failed to init kernel queue\n");
+		kfree(kq);
+		return NULL;
+	}
+	return kq;
+}
+
+void kernel_queue_uninit(struct kernel_queue *kq)
+{
+	BUG_ON(!kq);
+
+	kq->uninitialize(kq);
+	kfree(kq);
+}
+
+static __attribute__((unused)) void test_kq(struct kfd_dev *dev)
+{
+	struct kernel_queue *kq;
+	uint32_t *buffer, i;
+	int retval;
+
+	BUG_ON(!dev);
+
+	pr_debug("kfd: starting kernel queue test\n");
+
+	kq = kernel_queue_init(dev, KFD_QUEUE_TYPE_HIQ);
+	BUG_ON(!kq);
+
+	retval = kq->acquire_packet_buffer(kq, 5, &buffer);
+	BUG_ON(retval != 0);
+	for (i = 0; i < 5; i++)
+		buffer[i] = kq->nop_packet;
+	kq->submit_packet(kq);
+	kq->sync_with_hw(kq, 1000);
+
+	pr_debug("kfd: ending kernel queue test\n");
+}
+
+
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_kernel_queue.h b/drivers/gpu/drm/amd/amdkfd/kfd_kernel_queue.h
new file mode 100644
index 000000000000..dcd2bdb68d44
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_kernel_queue.h
@@ -0,0 +1,69 @@
+/*
+ * Copyright 2014 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#ifndef KFD_KERNEL_QUEUE_H_
+#define KFD_KERNEL_QUEUE_H_
+
+#include <linux/list.h>
+#include <linux/types.h>
+#include "kfd_priv.h"
+
+struct kernel_queue {
+	/* interface */
+	bool	(*initialize)(struct kernel_queue *kq, struct kfd_dev *dev,
+			enum kfd_queue_type type, unsigned int queue_size);
+	void	(*uninitialize)(struct kernel_queue *kq);
+	int	(*acquire_packet_buffer)(struct kernel_queue *kq,
+					size_t packet_size_in_dwords,
+					unsigned int **buffer_ptr);
+
+	void	(*submit_packet)(struct kernel_queue *kq);
+	int	(*sync_with_hw)(struct kernel_queue *kq,
+				unsigned long timeout_ms);
+	void	(*rollback_packet)(struct kernel_queue *kq);
+
+	/* data */
+	struct kfd_dev		*dev;
+	struct mqd_manager	*mqd;
+	struct queue		*queue;
+	uint32_t		pending_wptr;
+	unsigned int		nop_packet;
+
+	struct kfd_mem_obj	*rptr_mem;
+	uint32_t		*rptr_kernel;
+	uint64_t		rptr_gpu_addr;
+	struct kfd_mem_obj	*wptr_mem;
+	uint32_t		*wptr_kernel;
+	uint64_t		wptr_gpu_addr;
+	struct kfd_mem_obj	*pq;
+	uint64_t		pq_gpu_addr;
+	uint32_t		*pq_kernel_addr;
+
+	struct kfd_mem_obj	*fence_mem_obj;
+	uint64_t		fence_gpu_addr;
+	void			*fence_kernel_address;
+
+	struct list_head	list;
+};
+
+#endif /* KFD_KERNEL_QUEUE_H_ */
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_module.c b/drivers/gpu/drm/amd/amdkfd/kfd_module.c
new file mode 100644
index 000000000000..95d5af138e6e
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_module.c
@@ -0,0 +1,159 @@
+/*
+ * Copyright 2014 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ */
+
+#include <linux/module.h>
+#include <linux/sched.h>
+#include <linux/moduleparam.h>
+#include <linux/device.h>
+#include "kfd_priv.h"
+
+#define KFD_DRIVER_AUTHOR	"AMD Inc. and others"
+
+#define KFD_DRIVER_DESC		"Standalone HSA driver for AMD's GPUs"
+#define KFD_DRIVER_DATE		"20141113"
+#define KFD_DRIVER_MAJOR	0
+#define KFD_DRIVER_MINOR	7
+#define KFD_DRIVER_PATCHLEVEL	0
+
+const struct kfd2kgd_calls *kfd2kgd;
+static const struct kgd2kfd_calls kgd2kfd = {
+	.exit		= kgd2kfd_exit,
+	.probe		= kgd2kfd_probe,
+	.device_init	= kgd2kfd_device_init,
+	.device_exit	= kgd2kfd_device_exit,
+	.interrupt	= kgd2kfd_interrupt,
+	.suspend	= kgd2kfd_suspend,
+	.resume		= kgd2kfd_resume,
+};
+
+int sched_policy = KFD_SCHED_POLICY_HWS;
+module_param(sched_policy, int, 0444);
+MODULE_PARM_DESC(sched_policy,
+	"Kernel cmdline parameter that defines the amdkfd scheduling policy");
+
+int max_num_of_processes = KFD_MAX_NUM_OF_PROCESSES_DEFAULT;
+module_param(max_num_of_processes, int, 0444);
+MODULE_PARM_DESC(max_num_of_processes,
+	"Kernel cmdline parameter that defines the amdkfd maximum number of supported processes");
+
+int max_num_of_queues_per_process = KFD_MAX_NUM_OF_QUEUES_PER_PROCESS_DEFAULT;
+module_param(max_num_of_queues_per_process, int, 0444);
+MODULE_PARM_DESC(max_num_of_queues_per_process,
+	"Kernel cmdline parameter that defines the amdkfd maximum number of supported queues per process");
+
+bool kgd2kfd_init(unsigned interface_version,
+		  const struct kfd2kgd_calls *f2g,
+		  const struct kgd2kfd_calls **g2f)
+{
+	/*
+	 * Only one interface version is supported,
+	 * no kfd/kgd version skew allowed.
+	 */
+	if (interface_version != KFD_INTERFACE_VERSION)
+		return false;
+
+	/* Protection against multiple amd kgd loads */
+	if (kfd2kgd)
+		return true;
+
+	kfd2kgd = f2g;
+	*g2f = &kgd2kfd;
+
+	return true;
+}
+EXPORT_SYMBOL(kgd2kfd_init);
+
+void kgd2kfd_exit(void)
+{
+}
+
+static int __init kfd_module_init(void)
+{
+	int err;
+
+	kfd2kgd = NULL;
+
+	/* Verify module parameters */
+	if ((sched_policy < KFD_SCHED_POLICY_HWS) ||
+		(sched_policy > KFD_SCHED_POLICY_NO_HWS)) {
+		pr_err("kfd: sched_policy has invalid value\n");
+		return -1;
+	}
+
+	/* Verify module parameters */
+	if ((max_num_of_processes < 0) ||
+		(max_num_of_processes > KFD_MAX_NUM_OF_PROCESSES)) {
+		pr_err("kfd: max_num_of_processes must be between 0 to KFD_MAX_NUM_OF_PROCESSES\n");
+		return -1;
+	}
+
+	if ((max_num_of_queues_per_process < 0) ||
+		(max_num_of_queues_per_process >
+			KFD_MAX_NUM_OF_QUEUES_PER_PROCESS)) {
+		pr_err("kfd: max_num_of_queues_per_process must be between 0 to KFD_MAX_NUM_OF_QUEUES_PER_PROCESS\n");
+		return -1;
+	}
+
+	err = kfd_pasid_init();
+	if (err < 0)
+		goto err_pasid;
+
+	err = kfd_chardev_init();
+	if (err < 0)
+		goto err_ioctl;
+
+	err = kfd_topology_init();
+	if (err < 0)
+		goto err_topology;
+
+	kfd_process_create_wq();
+
+	dev_info(kfd_device, "Initialized module\n");
+
+	return 0;
+
+err_topology:
+	kfd_chardev_exit();
+err_ioctl:
+	kfd_pasid_exit();
+err_pasid:
+	return err;
+}
+
+static void __exit kfd_module_exit(void)
+{
+	kfd_process_destroy_wq();
+	kfd_topology_shutdown();
+	kfd_chardev_exit();
+	kfd_pasid_exit();
+	dev_info(kfd_device, "Removed module\n");
+}
+
+module_init(kfd_module_init);
+module_exit(kfd_module_exit);
+
+MODULE_AUTHOR(KFD_DRIVER_AUTHOR);
+MODULE_DESCRIPTION(KFD_DRIVER_DESC);
+MODULE_LICENSE("GPL and additional rights");
+MODULE_VERSION(__stringify(KFD_DRIVER_MAJOR) "."
+	       __stringify(KFD_DRIVER_MINOR) "."
+	       __stringify(KFD_DRIVER_PATCHLEVEL));
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager.c b/drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager.c
new file mode 100644
index 000000000000..4c3828cf45bf
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager.c
@@ -0,0 +1,346 @@
+/*
+ * Copyright 2014 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#include <linux/printk.h>
+#include <linux/slab.h>
+#include "kfd_priv.h"
+#include "kfd_mqd_manager.h"
+#include "cik_regs.h"
+#include "../../radeon/cik_reg.h"
+
+inline void busy_wait(unsigned long ms)
+{
+	while (time_before(jiffies, ms))
+		cpu_relax();
+}
+
+static inline struct cik_mqd *get_mqd(void *mqd)
+{
+	return (struct cik_mqd *)mqd;
+}
+
+static int init_mqd(struct mqd_manager *mm, void **mqd,
+		struct kfd_mem_obj **mqd_mem_obj, uint64_t *gart_addr,
+		struct queue_properties *q)
+{
+	uint64_t addr;
+	struct cik_mqd *m;
+	int retval;
+
+	BUG_ON(!mm || !q || !mqd);
+
+	pr_debug("kfd: In func %s\n", __func__);
+
+	retval = kfd2kgd->allocate_mem(mm->dev->kgd,
+					sizeof(struct cik_mqd),
+					256,
+					KFD_MEMPOOL_SYSTEM_WRITECOMBINE,
+					(struct kgd_mem **) mqd_mem_obj);
+
+	if (retval != 0)
+		return -ENOMEM;
+
+	m = (struct cik_mqd *) (*mqd_mem_obj)->cpu_ptr;
+	addr = (*mqd_mem_obj)->gpu_addr;
+
+	memset(m, 0, ALIGN(sizeof(struct cik_mqd), 256));
+
+	m->header = 0xC0310800;
+	m->compute_pipelinestat_enable = 1;
+	m->compute_static_thread_mgmt_se0 = 0xFFFFFFFF;
+	m->compute_static_thread_mgmt_se1 = 0xFFFFFFFF;
+	m->compute_static_thread_mgmt_se2 = 0xFFFFFFFF;
+	m->compute_static_thread_mgmt_se3 = 0xFFFFFFFF;
+
+	/*
+	 * Make sure to use the last queue state saved on mqd when the cp
+	 * reassigns the queue, so when queue is switched on/off (e.g over
+	 * subscription or quantum timeout) the context will be consistent
+	 */
+	m->cp_hqd_persistent_state =
+				DEFAULT_CP_HQD_PERSISTENT_STATE | PRELOAD_REQ;
+
+	m->cp_mqd_control             = MQD_CONTROL_PRIV_STATE_EN;
+	m->cp_mqd_base_addr_lo        = lower_32_bits(addr);
+	m->cp_mqd_base_addr_hi        = upper_32_bits(addr);
+
+	m->cp_hqd_ib_control = DEFAULT_MIN_IB_AVAIL_SIZE | IB_ATC_EN;
+	/* Although WinKFD writes this, I suspect it should not be necessary */
+	m->cp_hqd_ib_control = IB_ATC_EN | DEFAULT_MIN_IB_AVAIL_SIZE;
+
+	m->cp_hqd_quantum = QUANTUM_EN | QUANTUM_SCALE_1MS |
+				QUANTUM_DURATION(10);
+
+	/*
+	 * Pipe Priority
+	 * Identifies the pipe relative priority when this queue is connected
+	 * to the pipeline. The pipe priority is against the GFX pipe and HP3D.
+	 * In KFD we are using a fixed pipe priority set to CS_MEDIUM.
+	 * 0 = CS_LOW (typically below GFX)
+	 * 1 = CS_MEDIUM (typically between HP3D and GFX
+	 * 2 = CS_HIGH (typically above HP3D)
+	 */
+	m->cp_hqd_pipe_priority = 1;
+	m->cp_hqd_queue_priority = 15;
+
+	*mqd = m;
+	if (gart_addr != NULL)
+		*gart_addr = addr;
+	retval = mm->update_mqd(mm, m, q);
+
+	return retval;
+}
+
+static void uninit_mqd(struct mqd_manager *mm, void *mqd,
+			struct kfd_mem_obj *mqd_mem_obj)
+{
+	BUG_ON(!mm || !mqd);
+	kfd2kgd->free_mem(mm->dev->kgd, (struct kgd_mem *) mqd_mem_obj);
+}
+
+static int load_mqd(struct mqd_manager *mm, void *mqd, uint32_t pipe_id,
+			uint32_t queue_id, uint32_t __user *wptr)
+{
+	return kfd2kgd->hqd_load(mm->dev->kgd, mqd, pipe_id, queue_id, wptr);
+
+}
+
+static int update_mqd(struct mqd_manager *mm, void *mqd,
+			struct queue_properties *q)
+{
+	struct cik_mqd *m;
+
+	BUG_ON(!mm || !q || !mqd);
+
+	pr_debug("kfd: In func %s\n", __func__);
+
+	m = get_mqd(mqd);
+	m->cp_hqd_pq_control = DEFAULT_RPTR_BLOCK_SIZE |
+				DEFAULT_MIN_AVAIL_SIZE | PQ_ATC_EN;
+
+	/*
+	 * Calculating queue size which is log base 2 of actual queue size -1
+	 * dwords and another -1 for ffs
+	 */
+	m->cp_hqd_pq_control |= ffs(q->queue_size / sizeof(unsigned int))
+								- 1 - 1;
+	m->cp_hqd_pq_base_lo = lower_32_bits((uint64_t)q->queue_address >> 8);
+	m->cp_hqd_pq_base_hi = upper_32_bits((uint64_t)q->queue_address >> 8);
+	m->cp_hqd_pq_rptr_report_addr_lo = lower_32_bits((uint64_t)q->read_ptr);
+	m->cp_hqd_pq_rptr_report_addr_hi = upper_32_bits((uint64_t)q->read_ptr);
+	m->cp_hqd_pq_doorbell_control = DOORBELL_EN |
+					DOORBELL_OFFSET(q->doorbell_off);
+
+	m->cp_hqd_vmid = q->vmid;
+
+	if (q->format == KFD_QUEUE_FORMAT_AQL) {
+		m->cp_hqd_iq_rptr = AQL_ENABLE;
+		m->cp_hqd_pq_control |= NO_UPDATE_RPTR;
+	}
+
+	m->cp_hqd_active = 0;
+	q->is_active = false;
+	if (q->queue_size > 0 &&
+			q->queue_address != 0 &&
+			q->queue_percent > 0) {
+		m->cp_hqd_active = 1;
+		q->is_active = true;
+	}
+
+	return 0;
+}
+
+static int destroy_mqd(struct mqd_manager *mm, void *mqd,
+			enum kfd_preempt_type type,
+			unsigned int timeout, uint32_t pipe_id,
+			uint32_t queue_id)
+{
+	return kfd2kgd->hqd_destroy(mm->dev->kgd, type, timeout,
+					pipe_id, queue_id);
+}
+
+static bool is_occupied(struct mqd_manager *mm, void *mqd,
+			uint64_t queue_address,	uint32_t pipe_id,
+			uint32_t queue_id)
+{
+
+	return kfd2kgd->hqd_is_occupied(mm->dev->kgd, queue_address,
+					pipe_id, queue_id);
+
+}
+
+/*
+ * HIQ MQD Implementation, concrete implementation for HIQ MQD implementation.
+ * The HIQ queue in Kaveri is using the same MQD structure as all the user mode
+ * queues but with different initial values.
+ */
+
+static int init_mqd_hiq(struct mqd_manager *mm, void **mqd,
+		struct kfd_mem_obj **mqd_mem_obj, uint64_t *gart_addr,
+		struct queue_properties *q)
+{
+	uint64_t addr;
+	struct cik_mqd *m;
+	int retval;
+
+	BUG_ON(!mm || !q || !mqd || !mqd_mem_obj);
+
+	pr_debug("kfd: In func %s\n", __func__);
+
+	retval = kfd2kgd->allocate_mem(mm->dev->kgd,
+					sizeof(struct cik_mqd),
+					256,
+					KFD_MEMPOOL_SYSTEM_WRITECOMBINE,
+					(struct kgd_mem **) mqd_mem_obj);
+
+	if (retval != 0)
+		return -ENOMEM;
+
+	m = (struct cik_mqd *) (*mqd_mem_obj)->cpu_ptr;
+	addr = (*mqd_mem_obj)->gpu_addr;
+
+	memset(m, 0, ALIGN(sizeof(struct cik_mqd), 256));
+
+	m->header = 0xC0310800;
+	m->compute_pipelinestat_enable = 1;
+	m->compute_static_thread_mgmt_se0 = 0xFFFFFFFF;
+	m->compute_static_thread_mgmt_se1 = 0xFFFFFFFF;
+	m->compute_static_thread_mgmt_se2 = 0xFFFFFFFF;
+	m->compute_static_thread_mgmt_se3 = 0xFFFFFFFF;
+
+	m->cp_hqd_persistent_state = DEFAULT_CP_HQD_PERSISTENT_STATE |
+					PRELOAD_REQ;
+	m->cp_hqd_quantum = QUANTUM_EN | QUANTUM_SCALE_1MS |
+				QUANTUM_DURATION(10);
+
+	m->cp_mqd_control             = MQD_CONTROL_PRIV_STATE_EN;
+	m->cp_mqd_base_addr_lo        = lower_32_bits(addr);
+	m->cp_mqd_base_addr_hi        = upper_32_bits(addr);
+
+	m->cp_hqd_ib_control = DEFAULT_MIN_IB_AVAIL_SIZE;
+
+	/*
+	 * Pipe Priority
+	 * Identifies the pipe relative priority when this queue is connected
+	 * to the pipeline. The pipe priority is against the GFX pipe and HP3D.
+	 * In KFD we are using a fixed pipe priority set to CS_MEDIUM.
+	 * 0 = CS_LOW (typically below GFX)
+	 * 1 = CS_MEDIUM (typically between HP3D and GFX
+	 * 2 = CS_HIGH (typically above HP3D)
+	 */
+	m->cp_hqd_pipe_priority = 1;
+	m->cp_hqd_queue_priority = 15;
+
+	*mqd = m;
+	if (gart_addr)
+		*gart_addr = addr;
+	retval = mm->update_mqd(mm, m, q);
+
+	return retval;
+}
+
+static int update_mqd_hiq(struct mqd_manager *mm, void *mqd,
+				struct queue_properties *q)
+{
+	struct cik_mqd *m;
+
+	BUG_ON(!mm || !q || !mqd);
+
+	pr_debug("kfd: In func %s\n", __func__);
+
+	m = get_mqd(mqd);
+	m->cp_hqd_pq_control = DEFAULT_RPTR_BLOCK_SIZE |
+				DEFAULT_MIN_AVAIL_SIZE |
+				PRIV_STATE |
+				KMD_QUEUE;
+
+	/*
+	 * Calculating queue size which is log base 2 of actual queue
+	 * size -1 dwords
+	 */
+	m->cp_hqd_pq_control |= ffs(q->queue_size / sizeof(unsigned int))
+								- 1 - 1;
+	m->cp_hqd_pq_base_lo = lower_32_bits((uint64_t)q->queue_address >> 8);
+	m->cp_hqd_pq_base_hi = upper_32_bits((uint64_t)q->queue_address >> 8);
+	m->cp_hqd_pq_rptr_report_addr_lo = lower_32_bits((uint64_t)q->read_ptr);
+	m->cp_hqd_pq_rptr_report_addr_hi = upper_32_bits((uint64_t)q->read_ptr);
+	m->cp_hqd_pq_doorbell_control = DOORBELL_EN |
+					DOORBELL_OFFSET(q->doorbell_off);
+
+	m->cp_hqd_vmid = q->vmid;
+
+	m->cp_hqd_active = 0;
+	q->is_active = false;
+	if (q->queue_size > 0 &&
+			q->queue_address != 0 &&
+			q->queue_percent > 0) {
+		m->cp_hqd_active = 1;
+		q->is_active = true;
+	}
+
+	return 0;
+}
+
+struct mqd_manager *mqd_manager_init(enum KFD_MQD_TYPE type,
+					struct kfd_dev *dev)
+{
+	struct mqd_manager *mqd;
+
+	BUG_ON(!dev);
+	BUG_ON(type >= KFD_MQD_TYPE_MAX);
+
+	pr_debug("kfd: In func %s\n", __func__);
+
+	mqd = kzalloc(sizeof(struct mqd_manager), GFP_KERNEL);
+	if (!mqd)
+		return NULL;
+
+	mqd->dev = dev;
+
+	switch (type) {
+	case KFD_MQD_TYPE_CIK_CP:
+	case KFD_MQD_TYPE_CIK_COMPUTE:
+		mqd->init_mqd = init_mqd;
+		mqd->uninit_mqd = uninit_mqd;
+		mqd->load_mqd = load_mqd;
+		mqd->update_mqd = update_mqd;
+		mqd->destroy_mqd = destroy_mqd;
+		mqd->is_occupied = is_occupied;
+		break;
+	case KFD_MQD_TYPE_CIK_HIQ:
+		mqd->init_mqd = init_mqd_hiq;
+		mqd->uninit_mqd = uninit_mqd;
+		mqd->load_mqd = load_mqd;
+		mqd->update_mqd = update_mqd_hiq;
+		mqd->destroy_mqd = destroy_mqd;
+		mqd->is_occupied = is_occupied;
+		break;
+	default:
+		kfree(mqd);
+		return NULL;
+	}
+
+	return mqd;
+}
+
+/* SDMA queues should be implemented here when the cp will supports them */
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager.h b/drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager.h
new file mode 100644
index 000000000000..213a71e0b6c7
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager.h
@@ -0,0 +1,91 @@
+/*
+ * Copyright 2014 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#ifndef KFD_MQD_MANAGER_H_
+#define KFD_MQD_MANAGER_H_
+
+#include "kfd_priv.h"
+
+/**
+ * struct mqd_manager
+ *
+ * @init_mqd: Allocates the mqd buffer on local gpu memory and initialize it.
+ *
+ * @load_mqd: Loads the mqd to a concrete hqd slot. Used only for no cp
+ * scheduling mode.
+ *
+ * @update_mqd: Handles a update call for the MQD
+ *
+ * @destroy_mqd: Destroys the HQD slot and by that preempt the relevant queue.
+ * Used only for no cp scheduling.
+ *
+ * @uninit_mqd: Releases the mqd buffer from local gpu memory.
+ *
+ * @is_occupied: Checks if the relevant HQD slot is occupied.
+ *
+ * @mqd_mutex: Mqd manager mutex.
+ *
+ * @dev: The kfd device structure coupled with this module.
+ *
+ * MQD stands for Memory Queue Descriptor which represents the current queue
+ * state in the memory and initiate the HQD (Hardware Queue Descriptor) state.
+ * This structure is actually a base class for the different types of MQDs
+ * structures for the variant ASICs that should be supported in the future.
+ * This base class is also contains all the MQD specific operations.
+ * Another important thing to mention is that each queue has a MQD that keeps
+ * his state (or context) after each preemption or reassignment.
+ * Basically there are a instances of the mqd manager class per MQD type per
+ * ASIC. Currently the kfd driver supports only Kaveri so there are instances
+ * per KFD_MQD_TYPE for each device.
+ *
+ */
+
+struct mqd_manager {
+	int	(*init_mqd)(struct mqd_manager *mm, void **mqd,
+			struct kfd_mem_obj **mqd_mem_obj, uint64_t *gart_addr,
+			struct queue_properties *q);
+
+	int	(*load_mqd)(struct mqd_manager *mm, void *mqd,
+				uint32_t pipe_id, uint32_t queue_id,
+				uint32_t __user *wptr);
+
+	int	(*update_mqd)(struct mqd_manager *mm, void *mqd,
+				struct queue_properties *q);
+
+	int	(*destroy_mqd)(struct mqd_manager *mm, void *mqd,
+				enum kfd_preempt_type type,
+				unsigned int timeout, uint32_t pipe_id,
+				uint32_t queue_id);
+
+	void	(*uninit_mqd)(struct mqd_manager *mm, void *mqd,
+				struct kfd_mem_obj *mqd_mem_obj);
+
+	bool	(*is_occupied)(struct mqd_manager *mm, void *mqd,
+				uint64_t queue_address,	uint32_t pipe_id,
+				uint32_t queue_id);
+
+	struct mutex	mqd_mutex;
+	struct kfd_dev	*dev;
+};
+
+#endif /* KFD_MQD_MANAGER_H_ */
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_packet_manager.c b/drivers/gpu/drm/amd/amdkfd/kfd_packet_manager.c
new file mode 100644
index 000000000000..5ce9233d2004
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_packet_manager.c
@@ -0,0 +1,565 @@
+/*
+ * Copyright 2014 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#include <linux/slab.h>
+#include <linux/mutex.h>
+#include "kfd_device_queue_manager.h"
+#include "kfd_kernel_queue.h"
+#include "kfd_priv.h"
+#include "kfd_pm4_headers.h"
+#include "kfd_pm4_opcodes.h"
+
+static inline void inc_wptr(unsigned int *wptr, unsigned int increment_bytes,
+				unsigned int buffer_size_bytes)
+{
+	unsigned int temp = *wptr + increment_bytes / sizeof(uint32_t);
+
+	BUG_ON((temp * sizeof(uint32_t)) > buffer_size_bytes);
+	*wptr = temp;
+}
+
+static unsigned int build_pm4_header(unsigned int opcode, size_t packet_size)
+{
+	union PM4_MES_TYPE_3_HEADER header;
+
+	header.u32all = 0;
+	header.opcode = opcode;
+	header.count = packet_size/sizeof(uint32_t) - 2;
+	header.type = PM4_TYPE_3;
+
+	return header.u32all;
+}
+
+static void pm_calc_rlib_size(struct packet_manager *pm,
+				unsigned int *rlib_size,
+				bool *over_subscription)
+{
+	unsigned int process_count, queue_count;
+
+	BUG_ON(!pm || !rlib_size || !over_subscription);
+
+	process_count = pm->dqm->processes_count;
+	queue_count = pm->dqm->queue_count;
+
+	/* check if there is over subscription*/
+	*over_subscription = false;
+	if ((process_count > 1) ||
+		queue_count > PIPE_PER_ME_CP_SCHEDULING * QUEUES_PER_PIPE) {
+		*over_subscription = true;
+		pr_debug("kfd: over subscribed runlist\n");
+	}
+
+	/* calculate run list ib allocation size */
+	*rlib_size = process_count * sizeof(struct pm4_map_process) +
+		     queue_count * sizeof(struct pm4_map_queues);
+
+	/*
+	 * Increase the allocation size in case we need a chained run list
+	 * when over subscription
+	 */
+	if (*over_subscription)
+		*rlib_size += sizeof(struct pm4_runlist);
+
+	pr_debug("kfd: runlist ib size %d\n", *rlib_size);
+}
+
+static int pm_allocate_runlist_ib(struct packet_manager *pm,
+				unsigned int **rl_buffer,
+				uint64_t *rl_gpu_buffer,
+				unsigned int *rl_buffer_size,
+				bool *is_over_subscription)
+{
+	int retval;
+
+	BUG_ON(!pm);
+	BUG_ON(pm->allocated == true);
+	BUG_ON(is_over_subscription == NULL);
+
+	pm_calc_rlib_size(pm, rl_buffer_size, is_over_subscription);
+
+	retval = kfd2kgd->allocate_mem(pm->dqm->dev->kgd,
+					*rl_buffer_size,
+					PAGE_SIZE,
+					KFD_MEMPOOL_SYSTEM_WRITECOMBINE,
+					(struct kgd_mem **) &pm->ib_buffer_obj);
+
+	if (retval != 0) {
+		pr_err("kfd: failed to allocate runlist IB\n");
+		return retval;
+	}
+
+	*(void **)rl_buffer = pm->ib_buffer_obj->cpu_ptr;
+	*rl_gpu_buffer = pm->ib_buffer_obj->gpu_addr;
+
+	memset(*rl_buffer, 0, *rl_buffer_size);
+	pm->allocated = true;
+	return retval;
+}
+
+static int pm_create_runlist(struct packet_manager *pm, uint32_t *buffer,
+			uint64_t ib, size_t ib_size_in_dwords, bool chain)
+{
+	struct pm4_runlist *packet;
+
+	BUG_ON(!pm || !buffer || !ib);
+
+	packet = (struct pm4_runlist *)buffer;
+
+	memset(buffer, 0, sizeof(struct pm4_runlist));
+	packet->header.u32all = build_pm4_header(IT_RUN_LIST,
+						sizeof(struct pm4_runlist));
+
+	packet->bitfields4.ib_size = ib_size_in_dwords;
+	packet->bitfields4.chain = chain ? 1 : 0;
+	packet->bitfields4.offload_polling = 0;
+	packet->bitfields4.valid = 1;
+	packet->ordinal2 = lower_32_bits(ib);
+	packet->bitfields3.ib_base_hi = upper_32_bits(ib);
+
+	return 0;
+}
+
+static int pm_create_map_process(struct packet_manager *pm, uint32_t *buffer,
+				struct qcm_process_device *qpd)
+{
+	struct pm4_map_process *packet;
+	struct queue *cur;
+	uint32_t num_queues;
+
+	BUG_ON(!pm || !buffer || !qpd);
+
+	packet = (struct pm4_map_process *)buffer;
+
+	pr_debug("kfd: In func %s\n", __func__);
+
+	memset(buffer, 0, sizeof(struct pm4_map_process));
+
+	packet->header.u32all = build_pm4_header(IT_MAP_PROCESS,
+					sizeof(struct pm4_map_process));
+	packet->bitfields2.diq_enable = (qpd->is_debug) ? 1 : 0;
+	packet->bitfields2.process_quantum = 1;
+	packet->bitfields2.pasid = qpd->pqm->process->pasid;
+	packet->bitfields3.page_table_base = qpd->page_table_base;
+	packet->bitfields10.gds_size = qpd->gds_size;
+	packet->bitfields10.num_gws = qpd->num_gws;
+	packet->bitfields10.num_oac = qpd->num_oac;
+	num_queues = 0;
+	list_for_each_entry(cur, &qpd->queues_list, list)
+		num_queues++;
+	packet->bitfields10.num_queues = num_queues;
+
+	packet->sh_mem_config = qpd->sh_mem_config;
+	packet->sh_mem_bases = qpd->sh_mem_bases;
+	packet->sh_mem_ape1_base = qpd->sh_mem_ape1_base;
+	packet->sh_mem_ape1_limit = qpd->sh_mem_ape1_limit;
+
+	packet->gds_addr_lo = lower_32_bits(qpd->gds_context_area);
+	packet->gds_addr_hi = upper_32_bits(qpd->gds_context_area);
+
+	return 0;
+}
+
+static int pm_create_map_queue(struct packet_manager *pm, uint32_t *buffer,
+				struct queue *q)
+{
+	struct pm4_map_queues *packet;
+
+	BUG_ON(!pm || !buffer || !q);
+
+	pr_debug("kfd: In func %s\n", __func__);
+
+	packet = (struct pm4_map_queues *)buffer;
+	memset(buffer, 0, sizeof(struct pm4_map_queues));
+
+	packet->header.u32all = build_pm4_header(IT_MAP_QUEUES,
+						sizeof(struct pm4_map_queues));
+	packet->bitfields2.alloc_format =
+				alloc_format__mes_map_queues__one_per_pipe;
+	packet->bitfields2.num_queues = 1;
+	packet->bitfields2.queue_sel =
+		queue_sel__mes_map_queues__map_to_hws_determined_queue_slots;
+
+	packet->bitfields2.vidmem = (q->properties.is_interop) ?
+			vidmem__mes_map_queues__uses_video_memory :
+			vidmem__mes_map_queues__uses_no_video_memory;
+
+	switch (q->properties.type) {
+	case KFD_QUEUE_TYPE_COMPUTE:
+	case KFD_QUEUE_TYPE_DIQ:
+		packet->bitfields2.engine_sel =
+				engine_sel__mes_map_queues__compute;
+		break;
+	case KFD_QUEUE_TYPE_SDMA:
+		packet->bitfields2.engine_sel =
+				engine_sel__mes_map_queues__sdma0;
+		break;
+	default:
+		BUG();
+		break;
+	}
+
+	packet->mes_map_queues_ordinals[0].bitfields3.doorbell_offset =
+			q->properties.doorbell_off;
+
+	packet->mes_map_queues_ordinals[0].mqd_addr_lo =
+			lower_32_bits(q->gart_mqd_addr);
+
+	packet->mes_map_queues_ordinals[0].mqd_addr_hi =
+			upper_32_bits(q->gart_mqd_addr);
+
+	packet->mes_map_queues_ordinals[0].wptr_addr_lo =
+			lower_32_bits((uint64_t)q->properties.write_ptr);
+
+	packet->mes_map_queues_ordinals[0].wptr_addr_hi =
+			upper_32_bits((uint64_t)q->properties.write_ptr);
+
+	return 0;
+}
+
+static int pm_create_runlist_ib(struct packet_manager *pm,
+				struct list_head *queues,
+				uint64_t *rl_gpu_addr,
+				size_t *rl_size_bytes)
+{
+	unsigned int alloc_size_bytes;
+	unsigned int *rl_buffer, rl_wptr, i;
+	int retval, proccesses_mapped;
+	struct device_process_node *cur;
+	struct qcm_process_device *qpd;
+	struct queue *q;
+	struct kernel_queue *kq;
+	bool is_over_subscription;
+
+	BUG_ON(!pm || !queues || !rl_size_bytes || !rl_gpu_addr);
+
+	rl_wptr = retval = proccesses_mapped = 0;
+
+	retval = pm_allocate_runlist_ib(pm, &rl_buffer, rl_gpu_addr,
+				&alloc_size_bytes, &is_over_subscription);
+	if (retval != 0)
+		return retval;
+
+	*rl_size_bytes = alloc_size_bytes;
+
+	pr_debug("kfd: In func %s\n", __func__);
+	pr_debug("kfd: building runlist ib process count: %d queues count %d\n",
+		pm->dqm->processes_count, pm->dqm->queue_count);
+
+	/* build the run list ib packet */
+	list_for_each_entry(cur, queues, list) {
+		qpd = cur->qpd;
+		/* build map process packet */
+		if (proccesses_mapped >= pm->dqm->processes_count) {
+			pr_debug("kfd: not enough space left in runlist IB\n");
+			pm_release_ib(pm);
+			return -ENOMEM;
+		}
+		retval = pm_create_map_process(pm, &rl_buffer[rl_wptr], qpd);
+		if (retval != 0)
+			return retval;
+		proccesses_mapped++;
+		inc_wptr(&rl_wptr, sizeof(struct pm4_map_process),
+				alloc_size_bytes);
+
+		list_for_each_entry(kq, &qpd->priv_queue_list, list) {
+			if (kq->queue->properties.is_active != true)
+				continue;
+			retval = pm_create_map_queue(pm, &rl_buffer[rl_wptr],
+							kq->queue);
+			if (retval != 0)
+				return retval;
+			inc_wptr(&rl_wptr, sizeof(struct pm4_map_queues),
+					alloc_size_bytes);
+		}
+
+		list_for_each_entry(q, &qpd->queues_list, list) {
+			if (q->properties.is_active != true)
+				continue;
+			retval = pm_create_map_queue(pm,
+						&rl_buffer[rl_wptr], q);
+			if (retval != 0)
+				return retval;
+			inc_wptr(&rl_wptr, sizeof(struct pm4_map_queues),
+					alloc_size_bytes);
+		}
+	}
+
+	pr_debug("kfd: finished map process and queues to runlist\n");
+
+	if (is_over_subscription)
+		pm_create_runlist(pm, &rl_buffer[rl_wptr], *rl_gpu_addr,
+				alloc_size_bytes / sizeof(uint32_t), true);
+
+	for (i = 0; i < alloc_size_bytes / sizeof(uint32_t); i++)
+		pr_debug("0x%2X ", rl_buffer[i]);
+	pr_debug("\n");
+
+	return 0;
+}
+
+int pm_init(struct packet_manager *pm, struct device_queue_manager *dqm)
+{
+	BUG_ON(!dqm);
+
+	pm->dqm = dqm;
+	mutex_init(&pm->lock);
+	pm->priv_queue = kernel_queue_init(dqm->dev, KFD_QUEUE_TYPE_HIQ);
+	if (pm->priv_queue == NULL) {
+		mutex_destroy(&pm->lock);
+		return -ENOMEM;
+	}
+	pm->allocated = false;
+
+	return 0;
+}
+
+void pm_uninit(struct packet_manager *pm)
+{
+	BUG_ON(!pm);
+
+	mutex_destroy(&pm->lock);
+	kernel_queue_uninit(pm->priv_queue);
+}
+
+int pm_send_set_resources(struct packet_manager *pm,
+				struct scheduling_resources *res)
+{
+	struct pm4_set_resources *packet;
+
+	BUG_ON(!pm || !res);
+
+	pr_debug("kfd: In func %s\n", __func__);
+
+	mutex_lock(&pm->lock);
+	pm->priv_queue->acquire_packet_buffer(pm->priv_queue,
+					sizeof(*packet) / sizeof(uint32_t),
+			(unsigned int **)&packet);
+	if (packet == NULL) {
+		mutex_unlock(&pm->lock);
+		pr_err("kfd: failed to allocate buffer on kernel queue\n");
+		return -ENOMEM;
+	}
+
+	memset(packet, 0, sizeof(struct pm4_set_resources));
+	packet->header.u32all = build_pm4_header(IT_SET_RESOURCES,
+					sizeof(struct pm4_set_resources));
+
+	packet->bitfields2.queue_type =
+			queue_type__mes_set_resources__hsa_interface_queue_hiq;
+	packet->bitfields2.vmid_mask = res->vmid_mask;
+	packet->bitfields2.unmap_latency = KFD_UNMAP_LATENCY;
+	packet->bitfields7.oac_mask = res->oac_mask;
+	packet->bitfields8.gds_heap_base = res->gds_heap_base;
+	packet->bitfields8.gds_heap_size = res->gds_heap_size;
+
+	packet->gws_mask_lo = lower_32_bits(res->gws_mask);
+	packet->gws_mask_hi = upper_32_bits(res->gws_mask);
+
+	packet->queue_mask_lo = lower_32_bits(res->queue_mask);
+	packet->queue_mask_hi = upper_32_bits(res->queue_mask);
+
+	pm->priv_queue->submit_packet(pm->priv_queue);
+	pm->priv_queue->sync_with_hw(pm->priv_queue, KFD_HIQ_TIMEOUT);
+
+	mutex_unlock(&pm->lock);
+
+	return 0;
+}
+
+int pm_send_runlist(struct packet_manager *pm, struct list_head *dqm_queues)
+{
+	uint64_t rl_gpu_ib_addr;
+	uint32_t *rl_buffer;
+	size_t rl_ib_size, packet_size_dwords;
+	int retval;
+
+	BUG_ON(!pm || !dqm_queues);
+
+	retval = pm_create_runlist_ib(pm, dqm_queues, &rl_gpu_ib_addr,
+					&rl_ib_size);
+	if (retval != 0)
+		goto fail_create_runlist_ib;
+
+	pr_debug("kfd: runlist IB address: 0x%llX\n", rl_gpu_ib_addr);
+
+	packet_size_dwords = sizeof(struct pm4_runlist) / sizeof(uint32_t);
+	mutex_lock(&pm->lock);
+
+	retval = pm->priv_queue->acquire_packet_buffer(pm->priv_queue,
+					packet_size_dwords, &rl_buffer);
+	if (retval != 0)
+		goto fail_acquire_packet_buffer;
+
+	retval = pm_create_runlist(pm, rl_buffer, rl_gpu_ib_addr,
+					rl_ib_size / sizeof(uint32_t), false);
+	if (retval != 0)
+		goto fail_create_runlist;
+
+	pm->priv_queue->submit_packet(pm->priv_queue);
+	pm->priv_queue->sync_with_hw(pm->priv_queue, KFD_HIQ_TIMEOUT);
+
+	mutex_unlock(&pm->lock);
+
+	return retval;
+
+fail_create_runlist:
+	pm->priv_queue->rollback_packet(pm->priv_queue);
+fail_acquire_packet_buffer:
+	mutex_unlock(&pm->lock);
+fail_create_runlist_ib:
+	if (pm->allocated == true)
+		pm_release_ib(pm);
+	return retval;
+}
+
+int pm_send_query_status(struct packet_manager *pm, uint64_t fence_address,
+			uint32_t fence_value)
+{
+	int retval;
+	struct pm4_query_status *packet;
+
+	BUG_ON(!pm || !fence_address);
+
+	mutex_lock(&pm->lock);
+	retval = pm->priv_queue->acquire_packet_buffer(
+			pm->priv_queue,
+			sizeof(struct pm4_query_status) / sizeof(uint32_t),
+			(unsigned int **)&packet);
+	if (retval != 0)
+		goto fail_acquire_packet_buffer;
+
+	packet->header.u32all = build_pm4_header(IT_QUERY_STATUS,
+					sizeof(struct pm4_query_status));
+
+	packet->bitfields2.context_id = 0;
+	packet->bitfields2.interrupt_sel =
+			interrupt_sel__mes_query_status__completion_status;
+	packet->bitfields2.command =
+			command__mes_query_status__fence_only_after_write_ack;
+
+	packet->addr_hi = upper_32_bits((uint64_t)fence_address);
+	packet->addr_lo = lower_32_bits((uint64_t)fence_address);
+	packet->data_hi = upper_32_bits((uint64_t)fence_value);
+	packet->data_lo = lower_32_bits((uint64_t)fence_value);
+
+	pm->priv_queue->submit_packet(pm->priv_queue);
+	pm->priv_queue->sync_with_hw(pm->priv_queue, KFD_HIQ_TIMEOUT);
+	mutex_unlock(&pm->lock);
+
+	return 0;
+
+fail_acquire_packet_buffer:
+	mutex_unlock(&pm->lock);
+	return retval;
+}
+
+int pm_send_unmap_queue(struct packet_manager *pm, enum kfd_queue_type type,
+			enum kfd_preempt_type_filter mode,
+			uint32_t filter_param, bool reset,
+			unsigned int sdma_engine)
+{
+	int retval;
+	uint32_t *buffer;
+	struct pm4_unmap_queues *packet;
+
+	BUG_ON(!pm);
+
+	mutex_lock(&pm->lock);
+	retval = pm->priv_queue->acquire_packet_buffer(
+			pm->priv_queue,
+			sizeof(struct pm4_unmap_queues) / sizeof(uint32_t),
+			&buffer);
+	if (retval != 0)
+		goto err_acquire_packet_buffer;
+
+	packet = (struct pm4_unmap_queues *)buffer;
+	memset(buffer, 0, sizeof(struct pm4_unmap_queues));
+
+	packet->header.u32all = build_pm4_header(IT_UNMAP_QUEUES,
+					sizeof(struct pm4_unmap_queues));
+	switch (type) {
+	case KFD_QUEUE_TYPE_COMPUTE:
+	case KFD_QUEUE_TYPE_DIQ:
+		packet->bitfields2.engine_sel =
+			engine_sel__mes_unmap_queues__compute;
+		break;
+	case KFD_QUEUE_TYPE_SDMA:
+		packet->bitfields2.engine_sel =
+			engine_sel__mes_unmap_queues__sdma0 + sdma_engine;
+		break;
+	default:
+		BUG();
+		break;
+	}
+
+	if (reset)
+		packet->bitfields2.action =
+				action__mes_unmap_queues__reset_queues;
+	else
+		packet->bitfields2.action =
+				action__mes_unmap_queues__preempt_queues;
+
+	switch (mode) {
+	case KFD_PREEMPT_TYPE_FILTER_SINGLE_QUEUE:
+		packet->bitfields2.queue_sel =
+				queue_sel__mes_unmap_queues__perform_request_on_specified_queues;
+		packet->bitfields2.num_queues = 1;
+		packet->bitfields3b.doorbell_offset0 = filter_param;
+		break;
+	case KFD_PREEMPT_TYPE_FILTER_BY_PASID:
+		packet->bitfields2.queue_sel =
+				queue_sel__mes_unmap_queues__perform_request_on_pasid_queues;
+		packet->bitfields3a.pasid = filter_param;
+		break;
+	case KFD_PREEMPT_TYPE_FILTER_ALL_QUEUES:
+		packet->bitfields2.queue_sel =
+				queue_sel__mes_unmap_queues__perform_request_on_all_active_queues;
+		break;
+	default:
+		BUG();
+		break;
+	};
+
+	pm->priv_queue->submit_packet(pm->priv_queue);
+	pm->priv_queue->sync_with_hw(pm->priv_queue, KFD_HIQ_TIMEOUT);
+
+	mutex_unlock(&pm->lock);
+	return 0;
+
+err_acquire_packet_buffer:
+	mutex_unlock(&pm->lock);
+	return retval;
+}
+
+void pm_release_ib(struct packet_manager *pm)
+{
+	BUG_ON(!pm);
+
+	mutex_lock(&pm->lock);
+	if (pm->allocated) {
+		kfd2kgd->free_mem(pm->dqm->dev->kgd,
+				(struct kgd_mem *) pm->ib_buffer_obj);
+		pm->allocated = false;
+	}
+	mutex_unlock(&pm->lock);
+}
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_pasid.c b/drivers/gpu/drm/amd/amdkfd/kfd_pasid.c
new file mode 100644
index 000000000000..4c25ef504f79
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_pasid.c
@@ -0,0 +1,96 @@
+/*
+ * Copyright 2014 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ */
+
+#include <linux/slab.h>
+#include <linux/types.h>
+#include "kfd_priv.h"
+
+static unsigned long *pasid_bitmap;
+static unsigned int pasid_limit;
+static DEFINE_MUTEX(pasid_mutex);
+
+int kfd_pasid_init(void)
+{
+	pasid_limit = max_num_of_processes;
+
+	pasid_bitmap = kcalloc(BITS_TO_LONGS(pasid_limit), sizeof(long), GFP_KERNEL);
+	if (!pasid_bitmap)
+		return -ENOMEM;
+
+	set_bit(0, pasid_bitmap); /* PASID 0 is reserved. */
+
+	return 0;
+}
+
+void kfd_pasid_exit(void)
+{
+	kfree(pasid_bitmap);
+}
+
+bool kfd_set_pasid_limit(unsigned int new_limit)
+{
+	if (new_limit < pasid_limit) {
+		bool ok;
+
+		mutex_lock(&pasid_mutex);
+
+		/* ensure that no pasids >= new_limit are in-use */
+		ok = (find_next_bit(pasid_bitmap, pasid_limit, new_limit) ==
+								pasid_limit);
+		if (ok)
+			pasid_limit = new_limit;
+
+		mutex_unlock(&pasid_mutex);
+
+		return ok;
+	}
+
+	return true;
+}
+
+inline unsigned int kfd_get_pasid_limit(void)
+{
+	return pasid_limit;
+}
+
+unsigned int kfd_pasid_alloc(void)
+{
+	unsigned int found;
+
+	mutex_lock(&pasid_mutex);
+
+	found = find_first_zero_bit(pasid_bitmap, pasid_limit);
+	if (found == pasid_limit)
+		found = 0;
+	else
+		set_bit(found, pasid_bitmap);
+
+	mutex_unlock(&pasid_mutex);
+
+	return found;
+}
+
+void kfd_pasid_free(unsigned int pasid)
+{
+	BUG_ON(pasid == 0 || pasid >= pasid_limit);
+	clear_bit(pasid, pasid_bitmap);
+}
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_pm4_headers.h b/drivers/gpu/drm/amd/amdkfd/kfd_pm4_headers.h
new file mode 100644
index 000000000000..071ad5724bd2
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_pm4_headers.h
@@ -0,0 +1,405 @@
+/*
+ * Copyright 2014 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#ifndef KFD_PM4_HEADERS_H_
+#define KFD_PM4_HEADERS_H_
+
+#ifndef PM4_MES_HEADER_DEFINED
+#define PM4_MES_HEADER_DEFINED
+union PM4_MES_TYPE_3_HEADER {
+	struct {
+		uint32_t reserved1:8;	/* < reserved */
+		uint32_t opcode:8;	/* < IT opcode */
+		uint32_t count:14;	/* < number of DWORDs - 1
+					 * in the information body.
+					 */
+		uint32_t type:2;	/* < packet identifier.
+					 * It should be 3 for type 3 packets
+					 */
+	};
+	uint32_t u32all;
+};
+#endif /* PM4_MES_HEADER_DEFINED */
+
+/* --------------------MES_SET_RESOURCES-------------------- */
+
+#ifndef PM4_MES_SET_RESOURCES_DEFINED
+#define PM4_MES_SET_RESOURCES_DEFINED
+enum set_resources_queue_type_enum {
+	queue_type__mes_set_resources__kernel_interface_queue_kiq = 0,
+	queue_type__mes_set_resources__hsa_interface_queue_hiq = 1,
+	queue_type__mes_set_resources__hsa_debug_interface_queue = 4
+};
+
+struct pm4_set_resources {
+	union {
+		union PM4_MES_TYPE_3_HEADER header;	/* header */
+		uint32_t ordinal1;
+	};
+
+	union {
+		struct {
+			uint32_t vmid_mask:16;
+			uint32_t unmap_latency:8;
+			uint32_t reserved1:5;
+			enum set_resources_queue_type_enum queue_type:3;
+		} bitfields2;
+		uint32_t ordinal2;
+	};
+
+	uint32_t queue_mask_lo;
+	uint32_t queue_mask_hi;
+	uint32_t gws_mask_lo;
+	uint32_t gws_mask_hi;
+
+	union {
+		struct {
+			uint32_t oac_mask:16;
+			uint32_t reserved2:16;
+		} bitfields7;
+		uint32_t ordinal7;
+	};
+
+	union {
+		struct {
+			uint32_t gds_heap_base:6;
+			uint32_t reserved3:5;
+			uint32_t gds_heap_size:6;
+			uint32_t reserved4:15;
+		} bitfields8;
+		uint32_t ordinal8;
+	};
+
+};
+#endif
+
+/*--------------------MES_RUN_LIST-------------------- */
+
+#ifndef PM4_MES_RUN_LIST_DEFINED
+#define PM4_MES_RUN_LIST_DEFINED
+
+struct pm4_runlist {
+	union {
+		union PM4_MES_TYPE_3_HEADER header;	/* header */
+		uint32_t ordinal1;
+	};
+
+	union {
+		struct {
+			uint32_t reserved1:2;
+			uint32_t ib_base_lo:30;
+		} bitfields2;
+		uint32_t ordinal2;
+	};
+
+	union {
+		struct {
+			uint32_t ib_base_hi:16;
+			uint32_t reserved2:16;
+		} bitfields3;
+		uint32_t ordinal3;
+	};
+
+	union {
+		struct {
+			uint32_t ib_size:20;
+			uint32_t chain:1;
+			uint32_t offload_polling:1;
+			uint32_t reserved3:1;
+			uint32_t valid:1;
+			uint32_t reserved4:8;
+		} bitfields4;
+		uint32_t ordinal4;
+	};
+
+};
+#endif
+
+/*--------------------MES_MAP_PROCESS-------------------- */
+
+#ifndef PM4_MES_MAP_PROCESS_DEFINED
+#define PM4_MES_MAP_PROCESS_DEFINED
+
+struct pm4_map_process {
+	union {
+		union PM4_MES_TYPE_3_HEADER header;	/* header */
+		uint32_t ordinal1;
+	};
+
+	union {
+		struct {
+			uint32_t pasid:16;
+			uint32_t reserved1:8;
+			uint32_t diq_enable:1;
+			uint32_t process_quantum:7;
+		} bitfields2;
+		uint32_t ordinal2;
+	};
+
+	union {
+		struct {
+			uint32_t page_table_base:28;
+			uint32_t reserved3:4;
+		} bitfields3;
+		uint32_t ordinal3;
+	};
+
+	uint32_t sh_mem_bases;
+	uint32_t sh_mem_ape1_base;
+	uint32_t sh_mem_ape1_limit;
+	uint32_t sh_mem_config;
+	uint32_t gds_addr_lo;
+	uint32_t gds_addr_hi;
+
+	union {
+		struct {
+			uint32_t num_gws:6;
+			uint32_t reserved4:2;
+			uint32_t num_oac:4;
+			uint32_t reserved5:4;
+			uint32_t gds_size:6;
+			uint32_t num_queues:10;
+		} bitfields10;
+		uint32_t ordinal10;
+	};
+
+};
+#endif
+
+/*--------------------MES_MAP_QUEUES--------------------*/
+
+#ifndef PM4_MES_MAP_QUEUES_DEFINED
+#define PM4_MES_MAP_QUEUES_DEFINED
+enum map_queues_queue_sel_enum {
+	queue_sel__mes_map_queues__map_to_specified_queue_slots = 0,
+	queue_sel__mes_map_queues__map_to_hws_determined_queue_slots = 1,
+	queue_sel__mes_map_queues__enable_process_queues = 2
+};
+
+enum map_queues_vidmem_enum {
+	vidmem__mes_map_queues__uses_no_video_memory = 0,
+	vidmem__mes_map_queues__uses_video_memory = 1
+};
+
+enum map_queues_alloc_format_enum {
+	alloc_format__mes_map_queues__one_per_pipe = 0,
+	alloc_format__mes_map_queues__all_on_one_pipe = 1
+};
+
+enum map_queues_engine_sel_enum {
+	engine_sel__mes_map_queues__compute = 0,
+	engine_sel__mes_map_queues__sdma0 = 2,
+	engine_sel__mes_map_queues__sdma1 = 3
+};
+
+struct pm4_map_queues {
+	union {
+		union PM4_MES_TYPE_3_HEADER header;	/* header */
+		uint32_t ordinal1;
+	};
+
+	union {
+		struct {
+			uint32_t reserved1:4;
+			enum map_queues_queue_sel_enum queue_sel:2;
+			uint32_t reserved2:2;
+			uint32_t vmid:4;
+			uint32_t reserved3:4;
+			enum map_queues_vidmem_enum vidmem:2;
+			uint32_t reserved4:6;
+			enum map_queues_alloc_format_enum alloc_format:2;
+			enum map_queues_engine_sel_enum engine_sel:3;
+			uint32_t num_queues:3;
+		} bitfields2;
+		uint32_t ordinal2;
+	};
+
+	struct {
+		union {
+			struct {
+				uint32_t reserved5:2;
+				uint32_t doorbell_offset:21;
+				uint32_t reserved6:3;
+				uint32_t queue:6;
+			} bitfields3;
+			uint32_t ordinal3;
+		};
+
+		uint32_t mqd_addr_lo;
+		uint32_t mqd_addr_hi;
+		uint32_t wptr_addr_lo;
+		uint32_t wptr_addr_hi;
+
+	} mes_map_queues_ordinals[1];	/* 1..N of these ordinal groups */
+
+};
+#endif
+
+/*--------------------MES_QUERY_STATUS--------------------*/
+
+#ifndef PM4_MES_QUERY_STATUS_DEFINED
+#define PM4_MES_QUERY_STATUS_DEFINED
+enum query_status_interrupt_sel_enum {
+	interrupt_sel__mes_query_status__completion_status = 0,
+	interrupt_sel__mes_query_status__process_status = 1,
+	interrupt_sel__mes_query_status__queue_status = 2
+};
+
+enum query_status_command_enum {
+	command__mes_query_status__interrupt_only = 0,
+	command__mes_query_status__fence_only_immediate = 1,
+	command__mes_query_status__fence_only_after_write_ack = 2,
+	command__mes_query_status__fence_wait_for_write_ack_send_interrupt = 3
+};
+
+enum query_status_engine_sel_enum {
+	engine_sel__mes_query_status__compute = 0,
+	engine_sel__mes_query_status__sdma0_queue = 2,
+	engine_sel__mes_query_status__sdma1_queue = 3
+};
+
+struct pm4_query_status {
+	union {
+		union PM4_MES_TYPE_3_HEADER header;	/* header */
+		uint32_t ordinal1;
+	};
+
+	union {
+		struct {
+			uint32_t context_id:28;
+			enum query_status_interrupt_sel_enum interrupt_sel:2;
+			enum query_status_command_enum command:2;
+		} bitfields2;
+		uint32_t ordinal2;
+	};
+
+	union {
+		struct {
+			uint32_t pasid:16;
+			uint32_t reserved1:16;
+		} bitfields3a;
+		struct {
+			uint32_t reserved2:2;
+			uint32_t doorbell_offset:21;
+			uint32_t reserved3:3;
+			enum query_status_engine_sel_enum engine_sel:3;
+			uint32_t reserved4:3;
+		} bitfields3b;
+		uint32_t ordinal3;
+	};
+
+	uint32_t addr_lo;
+	uint32_t addr_hi;
+	uint32_t data_lo;
+	uint32_t data_hi;
+};
+#endif
+
+/*--------------------MES_UNMAP_QUEUES--------------------*/
+
+#ifndef PM4_MES_UNMAP_QUEUES_DEFINED
+#define PM4_MES_UNMAP_QUEUES_DEFINED
+enum unmap_queues_action_enum {
+	action__mes_unmap_queues__preempt_queues = 0,
+	action__mes_unmap_queues__reset_queues = 1,
+	action__mes_unmap_queues__disable_process_queues = 2
+};
+
+enum unmap_queues_queue_sel_enum {
+	queue_sel__mes_unmap_queues__perform_request_on_specified_queues = 0,
+	queue_sel__mes_unmap_queues__perform_request_on_pasid_queues = 1,
+	queue_sel__mes_unmap_queues__perform_request_on_all_active_queues = 2
+};
+
+enum unmap_queues_engine_sel_enum {
+	engine_sel__mes_unmap_queues__compute = 0,
+	engine_sel__mes_unmap_queues__sdma0 = 2,
+	engine_sel__mes_unmap_queues__sdma1 = 3
+};
+
+struct pm4_unmap_queues {
+	union {
+		union PM4_MES_TYPE_3_HEADER header;	/* header */
+		uint32_t ordinal1;
+	};
+
+	union {
+		struct {
+			enum unmap_queues_action_enum action:2;
+			uint32_t reserved1:2;
+			enum unmap_queues_queue_sel_enum queue_sel:2;
+			uint32_t reserved2:20;
+			enum unmap_queues_engine_sel_enum engine_sel:3;
+			uint32_t num_queues:3;
+		} bitfields2;
+		uint32_t ordinal2;
+	};
+
+	union {
+		struct {
+			uint32_t pasid:16;
+			uint32_t reserved3:16;
+		} bitfields3a;
+		struct {
+			uint32_t reserved4:2;
+			uint32_t doorbell_offset0:21;
+			uint32_t reserved5:9;
+		} bitfields3b;
+		uint32_t ordinal3;
+	};
+
+	union {
+		struct {
+			uint32_t reserved6:2;
+			uint32_t doorbell_offset1:21;
+			uint32_t reserved7:9;
+		} bitfields4;
+		uint32_t ordinal4;
+	};
+
+	union {
+		struct {
+			uint32_t reserved8:2;
+			uint32_t doorbell_offset2:21;
+			uint32_t reserved9:9;
+		} bitfields5;
+		uint32_t ordinal5;
+	};
+
+	union {
+		struct {
+			uint32_t reserved10:2;
+			uint32_t doorbell_offset3:21;
+			uint32_t reserved11:9;
+		} bitfields6;
+		uint32_t ordinal6;
+	};
+
+};
+#endif
+
+enum {
+	CACHE_FLUSH_AND_INV_TS_EVENT = 0x00000014
+};
+
+#endif /* KFD_PM4_HEADERS_H_ */
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_pm4_opcodes.h b/drivers/gpu/drm/amd/amdkfd/kfd_pm4_opcodes.h
new file mode 100644
index 000000000000..b72fa3b8c2d4
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_pm4_opcodes.h
@@ -0,0 +1,107 @@
+/*
+ * Copyright 2014 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+
+#ifndef KFD_PM4_OPCODES_H
+#define KFD_PM4_OPCODES_H
+
+enum it_opcode_type {
+	IT_NOP                               = 0x10,
+	IT_SET_BASE                          = 0x11,
+	IT_CLEAR_STATE                       = 0x12,
+	IT_INDEX_BUFFER_SIZE                 = 0x13,
+	IT_DISPATCH_DIRECT                   = 0x15,
+	IT_DISPATCH_INDIRECT                 = 0x16,
+	IT_ATOMIC_GDS                        = 0x1D,
+	IT_OCCLUSION_QUERY                   = 0x1F,
+	IT_SET_PREDICATION                   = 0x20,
+	IT_REG_RMW                           = 0x21,
+	IT_COND_EXEC                         = 0x22,
+	IT_PRED_EXEC                         = 0x23,
+	IT_DRAW_INDIRECT                     = 0x24,
+	IT_DRAW_INDEX_INDIRECT               = 0x25,
+	IT_INDEX_BASE                        = 0x26,
+	IT_DRAW_INDEX_2                      = 0x27,
+	IT_CONTEXT_CONTROL                   = 0x28,
+	IT_INDEX_TYPE                        = 0x2A,
+	IT_DRAW_INDIRECT_MULTI               = 0x2C,
+	IT_DRAW_INDEX_AUTO                   = 0x2D,
+	IT_NUM_INSTANCES                     = 0x2F,
+	IT_DRAW_INDEX_MULTI_AUTO             = 0x30,
+	IT_INDIRECT_BUFFER_CNST              = 0x33,
+	IT_STRMOUT_BUFFER_UPDATE             = 0x34,
+	IT_DRAW_INDEX_OFFSET_2               = 0x35,
+	IT_DRAW_PREAMBLE                     = 0x36,
+	IT_WRITE_DATA                        = 0x37,
+	IT_DRAW_INDEX_INDIRECT_MULTI         = 0x38,
+	IT_MEM_SEMAPHORE                     = 0x39,
+	IT_COPY_DW                           = 0x3B,
+	IT_WAIT_REG_MEM                      = 0x3C,
+	IT_INDIRECT_BUFFER                   = 0x3F,
+	IT_COPY_DATA                         = 0x40,
+	IT_PFP_SYNC_ME                       = 0x42,
+	IT_SURFACE_SYNC                      = 0x43,
+	IT_COND_WRITE                        = 0x45,
+	IT_EVENT_WRITE                       = 0x46,
+	IT_EVENT_WRITE_EOP                   = 0x47,
+	IT_EVENT_WRITE_EOS                   = 0x48,
+	IT_RELEASE_MEM                       = 0x49,
+	IT_PREAMBLE_CNTL                     = 0x4A,
+	IT_DMA_DATA                          = 0x50,
+	IT_ACQUIRE_MEM                       = 0x58,
+	IT_REWIND                            = 0x59,
+	IT_LOAD_UCONFIG_REG                  = 0x5E,
+	IT_LOAD_SH_REG                       = 0x5F,
+	IT_LOAD_CONFIG_REG                   = 0x60,
+	IT_LOAD_CONTEXT_REG                  = 0x61,
+	IT_SET_CONFIG_REG                    = 0x68,
+	IT_SET_CONTEXT_REG                   = 0x69,
+	IT_SET_CONTEXT_REG_INDIRECT          = 0x73,
+	IT_SET_SH_REG                        = 0x76,
+	IT_SET_SH_REG_OFFSET                 = 0x77,
+	IT_SET_QUEUE_REG                     = 0x78,
+	IT_SET_UCONFIG_REG                   = 0x79,
+	IT_SCRATCH_RAM_WRITE                 = 0x7D,
+	IT_SCRATCH_RAM_READ                  = 0x7E,
+	IT_LOAD_CONST_RAM                    = 0x80,
+	IT_WRITE_CONST_RAM                   = 0x81,
+	IT_DUMP_CONST_RAM                    = 0x83,
+	IT_INCREMENT_CE_COUNTER              = 0x84,
+	IT_INCREMENT_DE_COUNTER              = 0x85,
+	IT_WAIT_ON_CE_COUNTER                = 0x86,
+	IT_WAIT_ON_DE_COUNTER_DIFF           = 0x88,
+	IT_SWITCH_BUFFER                     = 0x8B,
+	IT_SET_RESOURCES                     = 0xA0,
+	IT_MAP_PROCESS                       = 0xA1,
+	IT_MAP_QUEUES                        = 0xA2,
+	IT_UNMAP_QUEUES                      = 0xA3,
+	IT_QUERY_STATUS                      = 0xA4,
+	IT_RUN_LIST                          = 0xA5,
+};
+
+#define PM4_TYPE_0 0
+#define PM4_TYPE_2 2
+#define PM4_TYPE_3 3
+
+#endif /* KFD_PM4_OPCODES_H */
+
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_priv.h b/drivers/gpu/drm/amd/amdkfd/kfd_priv.h
new file mode 100644
index 000000000000..a5edb29507e3
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_priv.h
@@ -0,0 +1,618 @@
+/*
+ * Copyright 2014 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ */
+
+#ifndef KFD_PRIV_H_INCLUDED
+#define KFD_PRIV_H_INCLUDED
+
+#include <linux/hashtable.h>
+#include <linux/mmu_notifier.h>
+#include <linux/mutex.h>
+#include <linux/types.h>
+#include <linux/atomic.h>
+#include <linux/workqueue.h>
+#include <linux/spinlock.h>
+#include <linux/kfd_ioctl.h>
+#include <kgd_kfd_interface.h>
+
+#define KFD_SYSFS_FILE_MODE 0444
+
+/*
+ * When working with cp scheduler we should assign the HIQ manually or via
+ * the radeon driver to a fixed hqd slot, here are the fixed HIQ hqd slot
+ * definitions for Kaveri. In Kaveri only the first ME queues participates
+ * in the cp scheduling taking that in mind we set the HIQ slot in the
+ * second ME.
+ */
+#define KFD_CIK_HIQ_PIPE 4
+#define KFD_CIK_HIQ_QUEUE 0
+
+/* GPU ID hash width in bits */
+#define KFD_GPU_ID_HASH_WIDTH 16
+
+/* Macro for allocating structures */
+#define kfd_alloc_struct(ptr_to_struct)	\
+	((typeof(ptr_to_struct)) kzalloc(sizeof(*ptr_to_struct), GFP_KERNEL))
+
+/* Kernel module parameter to specify maximum number of supported processes */
+extern int max_num_of_processes;
+
+#define KFD_MAX_NUM_OF_PROCESSES_DEFAULT 32
+#define KFD_MAX_NUM_OF_PROCESSES 512
+
+/*
+ * Kernel module parameter to specify maximum number of supported queues
+ * per process
+ */
+extern int max_num_of_queues_per_process;
+
+#define KFD_MAX_NUM_OF_QUEUES_PER_PROCESS_DEFAULT 128
+#define KFD_MAX_NUM_OF_QUEUES_PER_PROCESS 1024
+
+#define KFD_KERNEL_QUEUE_SIZE 2048
+
+/* Kernel module parameter to specify the scheduling policy */
+extern int sched_policy;
+
+/**
+ * enum kfd_sched_policy
+ *
+ * @KFD_SCHED_POLICY_HWS: H/W scheduling policy known as command processor (cp)
+ * scheduling. In this scheduling mode we're using the firmware code to
+ * schedule the user mode queues and kernel queues such as HIQ and DIQ.
+ * the HIQ queue is used as a special queue that dispatches the configuration
+ * to the cp and the user mode queues list that are currently running.
+ * the DIQ queue is a debugging queue that dispatches debugging commands to the
+ * firmware.
+ * in this scheduling mode user mode queues over subscription feature is
+ * enabled.
+ *
+ * @KFD_SCHED_POLICY_HWS_NO_OVERSUBSCRIPTION: The same as above but the over
+ * subscription feature disabled.
+ *
+ * @KFD_SCHED_POLICY_NO_HWS: no H/W scheduling policy is a mode which directly
+ * set the command processor registers and sets the queues "manually". This
+ * mode is used *ONLY* for debugging proposes.
+ *
+ */
+enum kfd_sched_policy {
+	KFD_SCHED_POLICY_HWS = 0,
+	KFD_SCHED_POLICY_HWS_NO_OVERSUBSCRIPTION,
+	KFD_SCHED_POLICY_NO_HWS
+};
+
+enum cache_policy {
+	cache_policy_coherent,
+	cache_policy_noncoherent
+};
+
+struct kfd_device_info {
+	unsigned int max_pasid_bits;
+	size_t ih_ring_entry_size;
+	uint16_t mqd_size_aligned;
+};
+
+struct kfd_dev {
+	struct kgd_dev *kgd;
+
+	const struct kfd_device_info *device_info;
+	struct pci_dev *pdev;
+
+	unsigned int id;		/* topology stub index */
+
+	phys_addr_t doorbell_base;	/* Start of actual doorbells used by
+					 * KFD. It is aligned for mapping
+					 * into user mode
+					 */
+	size_t doorbell_id_offset;	/* Doorbell offset (from KFD doorbell
+					 * to HW doorbell, GFX reserved some
+					 * at the start)
+					 */
+	size_t doorbell_process_limit;	/* Number of processes we have doorbell
+					 * space for.
+					 */
+	u32 __iomem *doorbell_kernel_ptr; /* This is a pointer for a doorbells
+					   * page used by kernel queue
+					   */
+
+	struct kgd2kfd_shared_resources shared_resources;
+
+	void *interrupt_ring;
+	size_t interrupt_ring_size;
+	atomic_t interrupt_ring_rptr;
+	atomic_t interrupt_ring_wptr;
+	struct work_struct interrupt_work;
+	spinlock_t interrupt_lock;
+
+	/* QCM Device instance */
+	struct device_queue_manager *dqm;
+
+	bool init_complete;
+	/*
+	 * Interrupts of interest to KFD are copied
+	 * from the HW ring into a SW ring.
+	 */
+	bool interrupts_active;
+};
+
+/* KGD2KFD callbacks */
+void kgd2kfd_exit(void);
+struct kfd_dev *kgd2kfd_probe(struct kgd_dev *kgd, struct pci_dev *pdev);
+bool kgd2kfd_device_init(struct kfd_dev *kfd,
+			 const struct kgd2kfd_shared_resources *gpu_resources);
+void kgd2kfd_device_exit(struct kfd_dev *kfd);
+
+extern const struct kfd2kgd_calls *kfd2kgd;
+
+struct kfd_mem_obj {
+	void *bo;
+	uint64_t gpu_addr;
+	uint32_t *cpu_ptr;
+};
+
+enum kfd_mempool {
+	KFD_MEMPOOL_SYSTEM_CACHEABLE = 1,
+	KFD_MEMPOOL_SYSTEM_WRITECOMBINE = 2,
+	KFD_MEMPOOL_FRAMEBUFFER = 3,
+};
+
+/* Character device interface */
+int kfd_chardev_init(void);
+void kfd_chardev_exit(void);
+struct device *kfd_chardev(void);
+
+/**
+ * enum kfd_preempt_type_filter
+ *
+ * @KFD_PREEMPT_TYPE_FILTER_SINGLE_QUEUE: Preempts single queue.
+ *
+ * @KFD_PRERMPT_TYPE_FILTER_ALL_QUEUES: Preempts all queues in the
+ *						running queues list.
+ *
+ * @KFD_PRERMPT_TYPE_FILTER_BY_PASID: Preempts queues that belongs to
+ *						specific process.
+ *
+ */
+enum kfd_preempt_type_filter {
+	KFD_PREEMPT_TYPE_FILTER_SINGLE_QUEUE,
+	KFD_PREEMPT_TYPE_FILTER_ALL_QUEUES,
+	KFD_PREEMPT_TYPE_FILTER_BY_PASID
+};
+
+enum kfd_preempt_type {
+	KFD_PREEMPT_TYPE_WAVEFRONT,
+	KFD_PREEMPT_TYPE_WAVEFRONT_RESET
+};
+
+/**
+ * enum kfd_queue_type
+ *
+ * @KFD_QUEUE_TYPE_COMPUTE: Regular user mode queue type.
+ *
+ * @KFD_QUEUE_TYPE_SDMA: Sdma user mode queue type.
+ *
+ * @KFD_QUEUE_TYPE_HIQ: HIQ queue type.
+ *
+ * @KFD_QUEUE_TYPE_DIQ: DIQ queue type.
+ */
+enum kfd_queue_type  {
+	KFD_QUEUE_TYPE_COMPUTE,
+	KFD_QUEUE_TYPE_SDMA,
+	KFD_QUEUE_TYPE_HIQ,
+	KFD_QUEUE_TYPE_DIQ
+};
+
+enum kfd_queue_format {
+	KFD_QUEUE_FORMAT_PM4,
+	KFD_QUEUE_FORMAT_AQL
+};
+
+/**
+ * struct queue_properties
+ *
+ * @type: The queue type.
+ *
+ * @queue_id: Queue identifier.
+ *
+ * @queue_address: Queue ring buffer address.
+ *
+ * @queue_size: Queue ring buffer size.
+ *
+ * @priority: Defines the queue priority relative to other queues in the
+ * process.
+ * This is just an indication and HW scheduling may override the priority as
+ * necessary while keeping the relative prioritization.
+ * the priority granularity is from 0 to f which f is the highest priority.
+ * currently all queues are initialized with the highest priority.
+ *
+ * @queue_percent: This field is partially implemented and currently a zero in
+ * this field defines that the queue is non active.
+ *
+ * @read_ptr: User space address which points to the number of dwords the
+ * cp read from the ring buffer. This field updates automatically by the H/W.
+ *
+ * @write_ptr: Defines the number of dwords written to the ring buffer.
+ *
+ * @doorbell_ptr: This field aim is to notify the H/W of new packet written to
+ * the queue ring buffer. This field should be similar to write_ptr and the user
+ * should update this field after he updated the write_ptr.
+ *
+ * @doorbell_off: The doorbell offset in the doorbell pci-bar.
+ *
+ * @is_interop: Defines if this is a interop queue. Interop queue means that the
+ * queue can access both graphics and compute resources.
+ *
+ * @is_active: Defines if the queue is active or not.
+ *
+ * @vmid: If the scheduling mode is no cp scheduling the field defines the vmid
+ * of the queue.
+ *
+ * This structure represents the queue properties for each queue no matter if
+ * it's user mode or kernel mode queue.
+ *
+ */
+struct queue_properties {
+	enum kfd_queue_type type;
+	enum kfd_queue_format format;
+	unsigned int queue_id;
+	uint64_t queue_address;
+	uint64_t  queue_size;
+	uint32_t priority;
+	uint32_t queue_percent;
+	uint32_t *read_ptr;
+	uint32_t *write_ptr;
+	uint32_t __iomem *doorbell_ptr;
+	uint32_t doorbell_off;
+	bool is_interop;
+	bool is_active;
+	/* Not relevant for user mode queues in cp scheduling */
+	unsigned int vmid;
+};
+
+/**
+ * struct queue
+ *
+ * @list: Queue linked list.
+ *
+ * @mqd: The queue MQD.
+ *
+ * @mqd_mem_obj: The MQD local gpu memory object.
+ *
+ * @gart_mqd_addr: The MQD gart mc address.
+ *
+ * @properties: The queue properties.
+ *
+ * @mec: Used only in no cp scheduling mode and identifies to micro engine id
+ * that the queue should be execute on.
+ *
+ * @pipe: Used only in no cp scheduling mode and identifies the queue's pipe id.
+ *
+ * @queue: Used only in no cp scheduliong mode and identifies the queue's slot.
+ *
+ * @process: The kfd process that created this queue.
+ *
+ * @device: The kfd device that created this queue.
+ *
+ * This structure represents user mode compute queues.
+ * It contains all the necessary data to handle such queues.
+ *
+ */
+
+struct queue {
+	struct list_head list;
+	void *mqd;
+	struct kfd_mem_obj *mqd_mem_obj;
+	uint64_t gart_mqd_addr;
+	struct queue_properties properties;
+
+	uint32_t mec;
+	uint32_t pipe;
+	uint32_t queue;
+
+	struct kfd_process	*process;
+	struct kfd_dev		*device;
+};
+
+/*
+ * Please read the kfd_mqd_manager.h description.
+ */
+enum KFD_MQD_TYPE {
+	KFD_MQD_TYPE_CIK_COMPUTE = 0, /* for no cp scheduling */
+	KFD_MQD_TYPE_CIK_HIQ, /* for hiq */
+	KFD_MQD_TYPE_CIK_CP, /* for cp queues and diq */
+	KFD_MQD_TYPE_CIK_SDMA, /* for sdma queues */
+	KFD_MQD_TYPE_MAX
+};
+
+struct scheduling_resources {
+	unsigned int vmid_mask;
+	enum kfd_queue_type type;
+	uint64_t queue_mask;
+	uint64_t gws_mask;
+	uint32_t oac_mask;
+	uint32_t gds_heap_base;
+	uint32_t gds_heap_size;
+};
+
+struct process_queue_manager {
+	/* data */
+	struct kfd_process	*process;
+	unsigned int		num_concurrent_processes;
+	struct list_head	queues;
+	unsigned long		*queue_slot_bitmap;
+};
+
+struct qcm_process_device {
+	/* The Device Queue Manager that owns this data */
+	struct device_queue_manager *dqm;
+	struct process_queue_manager *pqm;
+	/* Device Queue Manager lock */
+	struct mutex *lock;
+	/* Queues list */
+	struct list_head queues_list;
+	struct list_head priv_queue_list;
+
+	unsigned int queue_count;
+	unsigned int vmid;
+	bool is_debug;
+	/*
+	 * All the memory management data should be here too
+	 */
+	uint64_t gds_context_area;
+	uint32_t sh_mem_config;
+	uint32_t sh_mem_bases;
+	uint32_t sh_mem_ape1_base;
+	uint32_t sh_mem_ape1_limit;
+	uint32_t page_table_base;
+	uint32_t gds_size;
+	uint32_t num_gws;
+	uint32_t num_oac;
+};
+
+/* Data that is per-process-per device. */
+struct kfd_process_device {
+	/*
+	 * List of all per-device data for a process.
+	 * Starts from kfd_process.per_device_data.
+	 */
+	struct list_head per_device_list;
+
+	/* The device that owns this data. */
+	struct kfd_dev *dev;
+
+
+	/* per-process-per device QCM data structure */
+	struct qcm_process_device qpd;
+
+	/*Apertures*/
+	uint64_t lds_base;
+	uint64_t lds_limit;
+	uint64_t gpuvm_base;
+	uint64_t gpuvm_limit;
+	uint64_t scratch_base;
+	uint64_t scratch_limit;
+
+	/* Is this process/pasid bound to this device? (amd_iommu_bind_pasid) */
+	bool bound;
+};
+
+#define qpd_to_pdd(x) container_of(x, struct kfd_process_device, qpd)
+
+/* Process data */
+struct kfd_process {
+	/*
+	 * kfd_process are stored in an mm_struct*->kfd_process*
+	 * hash table (kfd_processes in kfd_process.c)
+	 */
+	struct hlist_node kfd_processes;
+
+	struct mm_struct *mm;
+
+	struct mutex mutex;
+
+	/*
+	 * In any process, the thread that started main() is the lead
+	 * thread and outlives the rest.
+	 * It is here because amd_iommu_bind_pasid wants a task_struct.
+	 */
+	struct task_struct *lead_thread;
+
+	/* We want to receive a notification when the mm_struct is destroyed */
+	struct mmu_notifier mmu_notifier;
+
+	/* Use for delayed freeing of kfd_process structure */
+	struct rcu_head	rcu;
+
+	unsigned int pasid;
+
+	/*
+	 * List of kfd_process_device structures,
+	 * one for each device the process is using.
+	 */
+	struct list_head per_device_data;
+
+	struct process_queue_manager pqm;
+
+	/* The process's queues. */
+	size_t queue_array_size;
+
+	/* Size is queue_array_size, up to MAX_PROCESS_QUEUES. */
+	struct kfd_queue **queues;
+
+	unsigned long allocated_queue_bitmap[DIV_ROUND_UP(KFD_MAX_NUM_OF_QUEUES_PER_PROCESS, BITS_PER_LONG)];
+
+	/*Is the user space process 32 bit?*/
+	bool is_32bit_user_mode;
+};
+
+/**
+ * Ioctl function type.
+ *
+ * \param filep pointer to file structure.
+ * \param p amdkfd process pointer.
+ * \param data pointer to arg that was copied from user.
+ */
+typedef int amdkfd_ioctl_t(struct file *filep, struct kfd_process *p,
+				void *data);
+
+struct amdkfd_ioctl_desc {
+	unsigned int cmd;
+	int flags;
+	amdkfd_ioctl_t *func;
+	unsigned int cmd_drv;
+	const char *name;
+};
+
+void kfd_process_create_wq(void);
+void kfd_process_destroy_wq(void);
+struct kfd_process *kfd_create_process(const struct task_struct *);
+struct kfd_process *kfd_get_process(const struct task_struct *);
+
+struct kfd_process_device *kfd_bind_process_to_device(struct kfd_dev *dev,
+							struct kfd_process *p);
+void kfd_unbind_process_from_device(struct kfd_dev *dev, unsigned int pasid);
+struct kfd_process_device *kfd_get_process_device_data(struct kfd_dev *dev,
+							struct kfd_process *p,
+							int create_pdd);
+
+/* Process device data iterator */
+struct kfd_process_device *kfd_get_first_process_device_data(struct kfd_process *p);
+struct kfd_process_device *kfd_get_next_process_device_data(struct kfd_process *p,
+						struct kfd_process_device *pdd);
+bool kfd_has_process_device_data(struct kfd_process *p);
+
+/* PASIDs */
+int kfd_pasid_init(void);
+void kfd_pasid_exit(void);
+bool kfd_set_pasid_limit(unsigned int new_limit);
+unsigned int kfd_get_pasid_limit(void);
+unsigned int kfd_pasid_alloc(void);
+void kfd_pasid_free(unsigned int pasid);
+
+/* Doorbells */
+void kfd_doorbell_init(struct kfd_dev *kfd);
+int kfd_doorbell_mmap(struct kfd_process *process, struct vm_area_struct *vma);
+u32 __iomem *kfd_get_kernel_doorbell(struct kfd_dev *kfd,
+					unsigned int *doorbell_off);
+void kfd_release_kernel_doorbell(struct kfd_dev *kfd, u32 __iomem *db_addr);
+u32 read_kernel_doorbell(u32 __iomem *db);
+void write_kernel_doorbell(u32 __iomem *db, u32 value);
+unsigned int kfd_queue_id_to_doorbell(struct kfd_dev *kfd,
+					struct kfd_process *process,
+					unsigned int queue_id);
+
+extern struct device *kfd_device;
+
+/* Topology */
+int kfd_topology_init(void);
+void kfd_topology_shutdown(void);
+int kfd_topology_add_device(struct kfd_dev *gpu);
+int kfd_topology_remove_device(struct kfd_dev *gpu);
+struct kfd_dev *kfd_device_by_id(uint32_t gpu_id);
+struct kfd_dev *kfd_device_by_pci_dev(const struct pci_dev *pdev);
+struct kfd_dev *kfd_topology_enum_kfd_devices(uint8_t idx);
+
+/* Interrupts */
+int kfd_interrupt_init(struct kfd_dev *dev);
+void kfd_interrupt_exit(struct kfd_dev *dev);
+void kgd2kfd_interrupt(struct kfd_dev *kfd, const void *ih_ring_entry);
+bool enqueue_ih_ring_entry(struct kfd_dev *kfd,	const void *ih_ring_entry);
+
+/* Power Management */
+void kgd2kfd_suspend(struct kfd_dev *kfd);
+int kgd2kfd_resume(struct kfd_dev *kfd);
+
+/* amdkfd Apertures */
+int kfd_init_apertures(struct kfd_process *process);
+
+/* Queue Context Management */
+inline uint32_t lower_32(uint64_t x);
+inline uint32_t upper_32(uint64_t x);
+
+int init_queue(struct queue **q, struct queue_properties properties);
+void uninit_queue(struct queue *q);
+void print_queue_properties(struct queue_properties *q);
+void print_queue(struct queue *q);
+
+struct mqd_manager *mqd_manager_init(enum KFD_MQD_TYPE type,
+					struct kfd_dev *dev);
+struct device_queue_manager *device_queue_manager_init(struct kfd_dev *dev);
+void device_queue_manager_uninit(struct device_queue_manager *dqm);
+struct kernel_queue *kernel_queue_init(struct kfd_dev *dev,
+					enum kfd_queue_type type);
+void kernel_queue_uninit(struct kernel_queue *kq);
+
+/* Process Queue Manager */
+struct process_queue_node {
+	struct queue *q;
+	struct kernel_queue *kq;
+	struct list_head process_queue_list;
+};
+
+int pqm_init(struct process_queue_manager *pqm, struct kfd_process *p);
+void pqm_uninit(struct process_queue_manager *pqm);
+int pqm_create_queue(struct process_queue_manager *pqm,
+			    struct kfd_dev *dev,
+			    struct file *f,
+			    struct queue_properties *properties,
+			    unsigned int flags,
+			    enum kfd_queue_type type,
+			    unsigned int *qid);
+int pqm_destroy_queue(struct process_queue_manager *pqm, unsigned int qid);
+int pqm_update_queue(struct process_queue_manager *pqm, unsigned int qid,
+			struct queue_properties *p);
+
+/* Packet Manager */
+
+#define KFD_HIQ_TIMEOUT (500)
+
+#define KFD_FENCE_COMPLETED (100)
+#define KFD_FENCE_INIT   (10)
+#define KFD_UNMAP_LATENCY (150)
+
+struct packet_manager {
+	struct device_queue_manager *dqm;
+	struct kernel_queue *priv_queue;
+	struct mutex lock;
+	bool allocated;
+	struct kfd_mem_obj *ib_buffer_obj;
+};
+
+int pm_init(struct packet_manager *pm, struct device_queue_manager *dqm);
+void pm_uninit(struct packet_manager *pm);
+int pm_send_set_resources(struct packet_manager *pm,
+				struct scheduling_resources *res);
+int pm_send_runlist(struct packet_manager *pm, struct list_head *dqm_queues);
+int pm_send_query_status(struct packet_manager *pm, uint64_t fence_address,
+				uint32_t fence_value);
+
+int pm_send_unmap_queue(struct packet_manager *pm, enum kfd_queue_type type,
+			enum kfd_preempt_type_filter mode,
+			uint32_t filter_param, bool reset,
+			unsigned int sdma_engine);
+
+void pm_release_ib(struct packet_manager *pm);
+
+uint64_t kfd_get_number_elems(struct kfd_dev *kfd);
+phys_addr_t kfd_get_process_doorbells(struct kfd_dev *dev,
+					struct kfd_process *process);
+
+#endif
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_process.c b/drivers/gpu/drm/amd/amdkfd/kfd_process.c
new file mode 100644
index 000000000000..3c76ef05cbcf
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_process.c
@@ -0,0 +1,419 @@
+/*
+ * Copyright 2014 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ */
+
+#include <linux/mutex.h>
+#include <linux/log2.h>
+#include <linux/sched.h>
+#include <linux/slab.h>
+#include <linux/amd-iommu.h>
+#include <linux/notifier.h>
+#include <linux/compat.h>
+
+struct mm_struct;
+
+#include "kfd_priv.h"
+
+/*
+ * Initial size for the array of queues.
+ * The allocated size is doubled each time
+ * it is exceeded up to MAX_PROCESS_QUEUES.
+ */
+#define INITIAL_QUEUE_ARRAY_SIZE 16
+
+/*
+ * List of struct kfd_process (field kfd_process).
+ * Unique/indexed by mm_struct*
+ */
+#define KFD_PROCESS_TABLE_SIZE 5 /* bits: 32 entries */
+static DEFINE_HASHTABLE(kfd_processes_table, KFD_PROCESS_TABLE_SIZE);
+static DEFINE_MUTEX(kfd_processes_mutex);
+
+DEFINE_STATIC_SRCU(kfd_processes_srcu);
+
+static struct workqueue_struct *kfd_process_wq;
+
+struct kfd_process_release_work {
+	struct work_struct kfd_work;
+	struct kfd_process *p;
+};
+
+static struct kfd_process *find_process(const struct task_struct *thread);
+static struct kfd_process *create_process(const struct task_struct *thread);
+
+void kfd_process_create_wq(void)
+{
+	if (!kfd_process_wq)
+		kfd_process_wq = create_workqueue("kfd_process_wq");
+}
+
+void kfd_process_destroy_wq(void)
+{
+	if (kfd_process_wq) {
+		flush_workqueue(kfd_process_wq);
+		destroy_workqueue(kfd_process_wq);
+		kfd_process_wq = NULL;
+	}
+}
+
+struct kfd_process *kfd_create_process(const struct task_struct *thread)
+{
+	struct kfd_process *process;
+
+	BUG_ON(!kfd_process_wq);
+
+	if (thread->mm == NULL)
+		return ERR_PTR(-EINVAL);
+
+	/* Only the pthreads threading model is supported. */
+	if (thread->group_leader->mm != thread->mm)
+		return ERR_PTR(-EINVAL);
+
+	/* Take mmap_sem because we call __mmu_notifier_register inside */
+	down_write(&thread->mm->mmap_sem);
+
+	/*
+	 * take kfd processes mutex before starting of process creation
+	 * so there won't be a case where two threads of the same process
+	 * create two kfd_process structures
+	 */
+	mutex_lock(&kfd_processes_mutex);
+
+	/* A prior open of /dev/kfd could have already created the process. */
+	process = find_process(thread);
+	if (process)
+		pr_debug("kfd: process already found\n");
+
+	if (!process)
+		process = create_process(thread);
+
+	mutex_unlock(&kfd_processes_mutex);
+
+	up_write(&thread->mm->mmap_sem);
+
+	return process;
+}
+
+struct kfd_process *kfd_get_process(const struct task_struct *thread)
+{
+	struct kfd_process *process;
+
+	if (thread->mm == NULL)
+		return ERR_PTR(-EINVAL);
+
+	/* Only the pthreads threading model is supported. */
+	if (thread->group_leader->mm != thread->mm)
+		return ERR_PTR(-EINVAL);
+
+	process = find_process(thread);
+
+	return process;
+}
+
+static struct kfd_process *find_process_by_mm(const struct mm_struct *mm)
+{
+	struct kfd_process *process;
+
+	hash_for_each_possible_rcu(kfd_processes_table, process,
+					kfd_processes, (uintptr_t)mm)
+		if (process->mm == mm)
+			return process;
+
+	return NULL;
+}
+
+static struct kfd_process *find_process(const struct task_struct *thread)
+{
+	struct kfd_process *p;
+	int idx;
+
+	idx = srcu_read_lock(&kfd_processes_srcu);
+	p = find_process_by_mm(thread->mm);
+	srcu_read_unlock(&kfd_processes_srcu, idx);
+
+	return p;
+}
+
+static void kfd_process_wq_release(struct work_struct *work)
+{
+	struct kfd_process_release_work *my_work;
+	struct kfd_process_device *pdd, *temp;
+	struct kfd_process *p;
+
+	my_work = (struct kfd_process_release_work *) work;
+
+	p = my_work->p;
+
+	mutex_lock(&p->mutex);
+
+	list_for_each_entry_safe(pdd, temp, &p->per_device_data,
+							per_device_list) {
+		amd_iommu_unbind_pasid(pdd->dev->pdev, p->pasid);
+		list_del(&pdd->per_device_list);
+
+		kfree(pdd);
+	}
+
+	kfd_pasid_free(p->pasid);
+
+	mutex_unlock(&p->mutex);
+
+	mutex_destroy(&p->mutex);
+
+	kfree(p->queues);
+
+	kfree(p);
+
+	kfree((void *)work);
+}
+
+static void kfd_process_destroy_delayed(struct rcu_head *rcu)
+{
+	struct kfd_process_release_work *work;
+	struct kfd_process *p;
+
+	BUG_ON(!kfd_process_wq);
+
+	p = container_of(rcu, struct kfd_process, rcu);
+	BUG_ON(atomic_read(&p->mm->mm_count) <= 0);
+
+	mmdrop(p->mm);
+
+	work = (struct kfd_process_release_work *)
+		kmalloc(sizeof(struct kfd_process_release_work), GFP_ATOMIC);
+
+	if (work) {
+		INIT_WORK((struct work_struct *) work, kfd_process_wq_release);
+		work->p = p;
+		queue_work(kfd_process_wq, (struct work_struct *) work);
+	}
+}
+
+static void kfd_process_notifier_release(struct mmu_notifier *mn,
+					struct mm_struct *mm)
+{
+	struct kfd_process *p;
+
+	/*
+	 * The kfd_process structure can not be free because the
+	 * mmu_notifier srcu is read locked
+	 */
+	p = container_of(mn, struct kfd_process, mmu_notifier);
+	BUG_ON(p->mm != mm);
+
+	mutex_lock(&kfd_processes_mutex);
+	hash_del_rcu(&p->kfd_processes);
+	mutex_unlock(&kfd_processes_mutex);
+	synchronize_srcu(&kfd_processes_srcu);
+
+	mutex_lock(&p->mutex);
+
+	/* In case our notifier is called before IOMMU notifier */
+	pqm_uninit(&p->pqm);
+
+	mutex_unlock(&p->mutex);
+
+	/*
+	 * Because we drop mm_count inside kfd_process_destroy_delayed
+	 * and because the mmu_notifier_unregister function also drop
+	 * mm_count we need to take an extra count here.
+	 */
+	atomic_inc(&p->mm->mm_count);
+	mmu_notifier_unregister_no_release(&p->mmu_notifier, p->mm);
+	mmu_notifier_call_srcu(&p->rcu, &kfd_process_destroy_delayed);
+}
+
+static const struct mmu_notifier_ops kfd_process_mmu_notifier_ops = {
+	.release = kfd_process_notifier_release,
+};
+
+static struct kfd_process *create_process(const struct task_struct *thread)
+{
+	struct kfd_process *process;
+	int err = -ENOMEM;
+
+	process = kzalloc(sizeof(*process), GFP_KERNEL);
+
+	if (!process)
+		goto err_alloc_process;
+
+	process->queues = kmalloc_array(INITIAL_QUEUE_ARRAY_SIZE,
+					sizeof(process->queues[0]), GFP_KERNEL);
+	if (!process->queues)
+		goto err_alloc_queues;
+
+	process->pasid = kfd_pasid_alloc();
+	if (process->pasid == 0)
+		goto err_alloc_pasid;
+
+	mutex_init(&process->mutex);
+
+	process->mm = thread->mm;
+
+	/* register notifier */
+	process->mmu_notifier.ops = &kfd_process_mmu_notifier_ops;
+	err = __mmu_notifier_register(&process->mmu_notifier, process->mm);
+	if (err)
+		goto err_mmu_notifier;
+
+	hash_add_rcu(kfd_processes_table, &process->kfd_processes,
+			(uintptr_t)process->mm);
+
+	process->lead_thread = thread->group_leader;
+
+	process->queue_array_size = INITIAL_QUEUE_ARRAY_SIZE;
+
+	INIT_LIST_HEAD(&process->per_device_data);
+
+	err = pqm_init(&process->pqm, process);
+	if (err != 0)
+		goto err_process_pqm_init;
+
+	/* init process apertures*/
+	process->is_32bit_user_mode = is_compat_task();
+	if (kfd_init_apertures(process) != 0)
+		goto err_init_apretures;
+
+	return process;
+
+err_init_apretures:
+	pqm_uninit(&process->pqm);
+err_process_pqm_init:
+	hash_del_rcu(&process->kfd_processes);
+	synchronize_rcu();
+	mmu_notifier_unregister_no_release(&process->mmu_notifier, process->mm);
+err_mmu_notifier:
+	kfd_pasid_free(process->pasid);
+err_alloc_pasid:
+	kfree(process->queues);
+err_alloc_queues:
+	kfree(process);
+err_alloc_process:
+	return ERR_PTR(err);
+}
+
+struct kfd_process_device *kfd_get_process_device_data(struct kfd_dev *dev,
+							struct kfd_process *p,
+							int create_pdd)
+{
+	struct kfd_process_device *pdd = NULL;
+
+	list_for_each_entry(pdd, &p->per_device_data, per_device_list)
+		if (pdd->dev == dev)
+			return pdd;
+
+	if (create_pdd) {
+		pdd = kzalloc(sizeof(*pdd), GFP_KERNEL);
+		if (pdd != NULL) {
+			pdd->dev = dev;
+			INIT_LIST_HEAD(&pdd->qpd.queues_list);
+			INIT_LIST_HEAD(&pdd->qpd.priv_queue_list);
+			pdd->qpd.dqm = dev->dqm;
+			list_add(&pdd->per_device_list, &p->per_device_data);
+		}
+	}
+
+	return pdd;
+}
+
+/*
+ * Direct the IOMMU to bind the process (specifically the pasid->mm)
+ * to the device.
+ * Unbinding occurs when the process dies or the device is removed.
+ *
+ * Assumes that the process lock is held.
+ */
+struct kfd_process_device *kfd_bind_process_to_device(struct kfd_dev *dev,
+							struct kfd_process *p)
+{
+	struct kfd_process_device *pdd = kfd_get_process_device_data(dev, p, 1);
+	int err;
+
+	if (pdd == NULL)
+		return ERR_PTR(-ENOMEM);
+
+	if (pdd->bound)
+		return pdd;
+
+	err = amd_iommu_bind_pasid(dev->pdev, p->pasid, p->lead_thread);
+	if (err < 0)
+		return ERR_PTR(err);
+
+	pdd->bound = true;
+
+	return pdd;
+}
+
+void kfd_unbind_process_from_device(struct kfd_dev *dev, unsigned int pasid)
+{
+	struct kfd_process *p;
+	struct kfd_process_device *pdd;
+	int idx, i;
+
+	BUG_ON(dev == NULL);
+
+	idx = srcu_read_lock(&kfd_processes_srcu);
+
+	hash_for_each_rcu(kfd_processes_table, i, p, kfd_processes)
+		if (p->pasid == pasid)
+			break;
+
+	srcu_read_unlock(&kfd_processes_srcu, idx);
+
+	BUG_ON(p->pasid != pasid);
+
+	mutex_lock(&p->mutex);
+
+	pqm_uninit(&p->pqm);
+
+	pdd = kfd_get_process_device_data(dev, p, 0);
+
+	/*
+	 * Just mark pdd as unbound, because we still need it to call
+	 * amd_iommu_unbind_pasid() in when the process exits.
+	 * We don't call amd_iommu_unbind_pasid() here
+	 * because the IOMMU called us.
+	 */
+	if (pdd)
+		pdd->bound = false;
+
+	mutex_unlock(&p->mutex);
+}
+
+struct kfd_process_device *kfd_get_first_process_device_data(struct kfd_process *p)
+{
+	return list_first_entry(&p->per_device_data,
+				struct kfd_process_device,
+				per_device_list);
+}
+
+struct kfd_process_device *kfd_get_next_process_device_data(struct kfd_process *p,
+						struct kfd_process_device *pdd)
+{
+	if (list_is_last(&pdd->per_device_list, &p->per_device_data))
+		return NULL;
+	return list_next_entry(pdd, per_device_list);
+}
+
+bool kfd_has_process_device_data(struct kfd_process *p)
+{
+	return !(list_empty(&p->per_device_data));
+}
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_process_queue_manager.c b/drivers/gpu/drm/amd/amdkfd/kfd_process_queue_manager.c
new file mode 100644
index 000000000000..47526780d736
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_process_queue_manager.c
@@ -0,0 +1,343 @@
+/*
+ * Copyright 2014 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#include <linux/slab.h>
+#include <linux/list.h>
+#include "kfd_device_queue_manager.h"
+#include "kfd_priv.h"
+#include "kfd_kernel_queue.h"
+
+static inline struct process_queue_node *get_queue_by_qid(
+			struct process_queue_manager *pqm, unsigned int qid)
+{
+	struct process_queue_node *pqn;
+
+	BUG_ON(!pqm);
+
+	list_for_each_entry(pqn, &pqm->queues, process_queue_list) {
+		if (pqn->q && pqn->q->properties.queue_id == qid)
+			return pqn;
+		if (pqn->kq && pqn->kq->queue->properties.queue_id == qid)
+			return pqn;
+	}
+
+	return NULL;
+}
+
+static int find_available_queue_slot(struct process_queue_manager *pqm,
+					unsigned int *qid)
+{
+	unsigned long found;
+
+	BUG_ON(!pqm || !qid);
+
+	pr_debug("kfd: in %s\n", __func__);
+
+	found = find_first_zero_bit(pqm->queue_slot_bitmap,
+			max_num_of_queues_per_process);
+
+	pr_debug("kfd: the new slot id %lu\n", found);
+
+	if (found >= max_num_of_queues_per_process) {
+		pr_info("amdkfd: Can not open more queues for process with pasid %d\n",
+				pqm->process->pasid);
+		return -ENOMEM;
+	}
+
+	set_bit(found, pqm->queue_slot_bitmap);
+	*qid = found;
+
+	return 0;
+}
+
+int pqm_init(struct process_queue_manager *pqm, struct kfd_process *p)
+{
+	BUG_ON(!pqm);
+
+	INIT_LIST_HEAD(&pqm->queues);
+	pqm->queue_slot_bitmap =
+			kzalloc(DIV_ROUND_UP(max_num_of_queues_per_process,
+					BITS_PER_BYTE), GFP_KERNEL);
+	if (pqm->queue_slot_bitmap == NULL)
+		return -ENOMEM;
+	pqm->process = p;
+
+	return 0;
+}
+
+void pqm_uninit(struct process_queue_manager *pqm)
+{
+	int retval;
+	struct process_queue_node *pqn, *next;
+
+	BUG_ON(!pqm);
+
+	pr_debug("In func %s\n", __func__);
+
+	list_for_each_entry_safe(pqn, next, &pqm->queues, process_queue_list) {
+		retval = pqm_destroy_queue(
+				pqm,
+				(pqn->q != NULL) ?
+					pqn->q->properties.queue_id :
+					pqn->kq->queue->properties.queue_id);
+
+		if (retval != 0) {
+			pr_err("kfd: failed to destroy queue\n");
+			return;
+		}
+	}
+	kfree(pqm->queue_slot_bitmap);
+	pqm->queue_slot_bitmap = NULL;
+}
+
+static int create_cp_queue(struct process_queue_manager *pqm,
+				struct kfd_dev *dev, struct queue **q,
+				struct queue_properties *q_properties,
+				struct file *f, unsigned int qid)
+{
+	int retval;
+
+	retval = 0;
+
+	/* Doorbell initialized in user space*/
+	q_properties->doorbell_ptr = NULL;
+
+	q_properties->doorbell_off =
+			kfd_queue_id_to_doorbell(dev, pqm->process, qid);
+
+	/* let DQM handle it*/
+	q_properties->vmid = 0;
+	q_properties->queue_id = qid;
+	q_properties->type = KFD_QUEUE_TYPE_COMPUTE;
+
+	retval = init_queue(q, *q_properties);
+	if (retval != 0)
+		goto err_init_queue;
+
+	(*q)->device = dev;
+	(*q)->process = pqm->process;
+
+	pr_debug("kfd: PQM After init queue");
+
+	return retval;
+
+err_init_queue:
+	return retval;
+}
+
+int pqm_create_queue(struct process_queue_manager *pqm,
+			    struct kfd_dev *dev,
+			    struct file *f,
+			    struct queue_properties *properties,
+			    unsigned int flags,
+			    enum kfd_queue_type type,
+			    unsigned int *qid)
+{
+	int retval;
+	struct kfd_process_device *pdd;
+	struct queue_properties q_properties;
+	struct queue *q;
+	struct process_queue_node *pqn;
+	struct kernel_queue *kq;
+
+	BUG_ON(!pqm || !dev || !properties || !qid);
+
+	memset(&q_properties, 0, sizeof(struct queue_properties));
+	memcpy(&q_properties, properties, sizeof(struct queue_properties));
+	q = NULL;
+	kq = NULL;
+
+	pdd = kfd_get_process_device_data(dev, pqm->process, 1);
+	BUG_ON(!pdd);
+
+	retval = find_available_queue_slot(pqm, qid);
+	if (retval != 0)
+		return retval;
+
+	if (list_empty(&pqm->queues)) {
+		pdd->qpd.pqm = pqm;
+		dev->dqm->register_process(dev->dqm, &pdd->qpd);
+	}
+
+	pqn = kzalloc(sizeof(struct process_queue_node), GFP_KERNEL);
+	if (!pqn) {
+		retval = -ENOMEM;
+		goto err_allocate_pqn;
+	}
+
+	switch (type) {
+	case KFD_QUEUE_TYPE_COMPUTE:
+		/* check if there is over subscription */
+		if ((sched_policy == KFD_SCHED_POLICY_HWS_NO_OVERSUBSCRIPTION) &&
+		((dev->dqm->processes_count >= VMID_PER_DEVICE) ||
+		(dev->dqm->queue_count >= PIPE_PER_ME_CP_SCHEDULING * QUEUES_PER_PIPE))) {
+			pr_err("kfd: over-subscription is not allowed in radeon_kfd.sched_policy == 1\n");
+			retval = -EPERM;
+			goto err_create_queue;
+		}
+
+		retval = create_cp_queue(pqm, dev, &q, &q_properties, f, *qid);
+		if (retval != 0)
+			goto err_create_queue;
+		pqn->q = q;
+		pqn->kq = NULL;
+		retval = dev->dqm->create_queue(dev->dqm, q, &pdd->qpd,
+						&q->properties.vmid);
+		print_queue(q);
+		break;
+	case KFD_QUEUE_TYPE_DIQ:
+		kq = kernel_queue_init(dev, KFD_QUEUE_TYPE_DIQ);
+		if (kq == NULL) {
+			retval = -ENOMEM;
+			goto err_create_queue;
+		}
+		kq->queue->properties.queue_id = *qid;
+		pqn->kq = kq;
+		pqn->q = NULL;
+		retval = dev->dqm->create_kernel_queue(dev->dqm, kq, &pdd->qpd);
+		break;
+	default:
+		BUG();
+		break;
+	}
+
+	if (retval != 0) {
+		pr_err("kfd: error dqm create queue\n");
+		goto err_create_queue;
+	}
+
+	pr_debug("kfd: PQM After DQM create queue\n");
+
+	list_add(&pqn->process_queue_list, &pqm->queues);
+
+	if (q) {
+		*properties = q->properties;
+		pr_debug("kfd: PQM done creating queue\n");
+		print_queue_properties(properties);
+	}
+
+	return retval;
+
+err_create_queue:
+	kfree(pqn);
+err_allocate_pqn:
+	clear_bit(*qid, pqm->queue_slot_bitmap);
+	return retval;
+}
+
+int pqm_destroy_queue(struct process_queue_manager *pqm, unsigned int qid)
+{
+	struct process_queue_node *pqn;
+	struct kfd_process_device *pdd;
+	struct device_queue_manager *dqm;
+	struct kfd_dev *dev;
+	int retval;
+
+	dqm = NULL;
+
+	BUG_ON(!pqm);
+	retval = 0;
+
+	pr_debug("kfd: In Func %s\n", __func__);
+
+	pqn = get_queue_by_qid(pqm, qid);
+	if (pqn == NULL) {
+		pr_err("kfd: queue id does not match any known queue\n");
+		return -EINVAL;
+	}
+
+	dev = NULL;
+	if (pqn->kq)
+		dev = pqn->kq->dev;
+	if (pqn->q)
+		dev = pqn->q->device;
+	BUG_ON(!dev);
+
+	pdd = kfd_get_process_device_data(dev, pqm->process, 1);
+	BUG_ON(!pdd);
+
+	if (pqn->kq) {
+		/* destroy kernel queue (DIQ) */
+		dqm = pqn->kq->dev->dqm;
+		dqm->destroy_kernel_queue(dqm, pqn->kq, &pdd->qpd);
+		kernel_queue_uninit(pqn->kq);
+	}
+
+	if (pqn->q) {
+		dqm = pqn->q->device->dqm;
+		retval = dqm->destroy_queue(dqm, &pdd->qpd, pqn->q);
+		if (retval != 0)
+			return retval;
+
+		uninit_queue(pqn->q);
+	}
+
+	list_del(&pqn->process_queue_list);
+	kfree(pqn);
+	clear_bit(qid, pqm->queue_slot_bitmap);
+
+	if (list_empty(&pqm->queues))
+		dqm->unregister_process(dqm, &pdd->qpd);
+
+	return retval;
+}
+
+int pqm_update_queue(struct process_queue_manager *pqm, unsigned int qid,
+			struct queue_properties *p)
+{
+	int retval;
+	struct process_queue_node *pqn;
+
+	BUG_ON(!pqm);
+
+	pqn = get_queue_by_qid(pqm, qid);
+	BUG_ON(!pqn);
+
+	pqn->q->properties.queue_address = p->queue_address;
+	pqn->q->properties.queue_size = p->queue_size;
+	pqn->q->properties.queue_percent = p->queue_percent;
+	pqn->q->properties.priority = p->priority;
+
+	retval = pqn->q->device->dqm->update_queue(pqn->q->device->dqm, pqn->q);
+	if (retval != 0)
+		return retval;
+
+	return 0;
+}
+
+static __attribute__((unused)) struct kernel_queue *pqm_get_kernel_queue(
+					struct process_queue_manager *pqm,
+					unsigned int qid)
+{
+	struct process_queue_node *pqn;
+
+	BUG_ON(!pqm);
+
+	pqn = get_queue_by_qid(pqm, qid);
+	if (pqn && pqn->kq)
+		return pqn->kq;
+
+	return NULL;
+}
+
+
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_queue.c b/drivers/gpu/drm/amd/amdkfd/kfd_queue.c
new file mode 100644
index 000000000000..9a0c90b0702e
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_queue.c
@@ -0,0 +1,85 @@
+/*
+ * Copyright 2014 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#include <linux/slab.h>
+#include "kfd_priv.h"
+
+void print_queue_properties(struct queue_properties *q)
+{
+	if (!q)
+		return;
+
+	pr_debug("Printing queue properties:\n");
+	pr_debug("Queue Type: %u\n", q->type);
+	pr_debug("Queue Size: %llu\n", q->queue_size);
+	pr_debug("Queue percent: %u\n", q->queue_percent);
+	pr_debug("Queue Address: 0x%llX\n", q->queue_address);
+	pr_debug("Queue Id: %u\n", q->queue_id);
+	pr_debug("Queue Process Vmid: %u\n", q->vmid);
+	pr_debug("Queue Read Pointer: 0x%p\n", q->read_ptr);
+	pr_debug("Queue Write Pointer: 0x%p\n", q->write_ptr);
+	pr_debug("Queue Doorbell Pointer: 0x%p\n", q->doorbell_ptr);
+	pr_debug("Queue Doorbell Offset: %u\n", q->doorbell_off);
+}
+
+void print_queue(struct queue *q)
+{
+	if (!q)
+		return;
+	pr_debug("Printing queue:\n");
+	pr_debug("Queue Type: %u\n", q->properties.type);
+	pr_debug("Queue Size: %llu\n", q->properties.queue_size);
+	pr_debug("Queue percent: %u\n", q->properties.queue_percent);
+	pr_debug("Queue Address: 0x%llX\n", q->properties.queue_address);
+	pr_debug("Queue Id: %u\n", q->properties.queue_id);
+	pr_debug("Queue Process Vmid: %u\n", q->properties.vmid);
+	pr_debug("Queue Read Pointer: 0x%p\n", q->properties.read_ptr);
+	pr_debug("Queue Write Pointer: 0x%p\n", q->properties.write_ptr);
+	pr_debug("Queue Doorbell Pointer: 0x%p\n", q->properties.doorbell_ptr);
+	pr_debug("Queue Doorbell Offset: %u\n", q->properties.doorbell_off);
+	pr_debug("Queue MQD Address: 0x%p\n", q->mqd);
+	pr_debug("Queue MQD Gart: 0x%llX\n", q->gart_mqd_addr);
+	pr_debug("Queue Process Address: 0x%p\n", q->process);
+	pr_debug("Queue Device Address: 0x%p\n", q->device);
+}
+
+int init_queue(struct queue **q, struct queue_properties properties)
+{
+	struct queue *tmp;
+
+	BUG_ON(!q);
+
+	tmp = kzalloc(sizeof(struct queue), GFP_KERNEL);
+	if (!tmp)
+		return -ENOMEM;
+
+	memcpy(&tmp->properties, &properties, sizeof(struct queue_properties));
+
+	*q = tmp;
+	return 0;
+}
+
+void uninit_queue(struct queue *q)
+{
+	kfree(q);
+}
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_topology.c b/drivers/gpu/drm/amd/amdkfd/kfd_topology.c
new file mode 100644
index 000000000000..cca1708fd811
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_topology.c
@@ -0,0 +1,1239 @@
+/*
+ * Copyright 2014 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ */
+
+#include <linux/types.h>
+#include <linux/kernel.h>
+#include <linux/pci.h>
+#include <linux/errno.h>
+#include <linux/acpi.h>
+#include <linux/hash.h>
+#include <linux/cpufreq.h>
+
+#include "kfd_priv.h"
+#include "kfd_crat.h"
+#include "kfd_topology.h"
+
+static struct list_head topology_device_list;
+static int topology_crat_parsed;
+static struct kfd_system_properties sys_props;
+
+static DECLARE_RWSEM(topology_lock);
+
+struct kfd_dev *kfd_device_by_id(uint32_t gpu_id)
+{
+	struct kfd_topology_device *top_dev;
+	struct kfd_dev *device = NULL;
+
+	down_read(&topology_lock);
+
+	list_for_each_entry(top_dev, &topology_device_list, list)
+		if (top_dev->gpu_id == gpu_id) {
+			device = top_dev->gpu;
+			break;
+		}
+
+	up_read(&topology_lock);
+
+	return device;
+}
+
+struct kfd_dev *kfd_device_by_pci_dev(const struct pci_dev *pdev)
+{
+	struct kfd_topology_device *top_dev;
+	struct kfd_dev *device = NULL;
+
+	down_read(&topology_lock);
+
+	list_for_each_entry(top_dev, &topology_device_list, list)
+		if (top_dev->gpu->pdev == pdev) {
+			device = top_dev->gpu;
+			break;
+		}
+
+	up_read(&topology_lock);
+
+	return device;
+}
+
+static int kfd_topology_get_crat_acpi(void *crat_image, size_t *size)
+{
+	struct acpi_table_header *crat_table;
+	acpi_status status;
+
+	if (!size)
+		return -EINVAL;
+
+	/*
+	 * Fetch the CRAT table from ACPI
+	 */
+	status = acpi_get_table(CRAT_SIGNATURE, 0, &crat_table);
+	if (status == AE_NOT_FOUND) {
+		pr_warn("CRAT table not found\n");
+		return -ENODATA;
+	} else if (ACPI_FAILURE(status)) {
+		const char *err = acpi_format_exception(status);
+
+		pr_err("CRAT table error: %s\n", err);
+		return -EINVAL;
+	}
+
+	if (*size >= crat_table->length && crat_image != NULL)
+		memcpy(crat_image, crat_table, crat_table->length);
+
+	*size = crat_table->length;
+
+	return 0;
+}
+
+static void kfd_populated_cu_info_cpu(struct kfd_topology_device *dev,
+		struct crat_subtype_computeunit *cu)
+{
+	BUG_ON(!dev);
+	BUG_ON(!cu);
+
+	dev->node_props.cpu_cores_count = cu->num_cpu_cores;
+	dev->node_props.cpu_core_id_base = cu->processor_id_low;
+	if (cu->hsa_capability & CRAT_CU_FLAGS_IOMMU_PRESENT)
+		dev->node_props.capability |= HSA_CAP_ATS_PRESENT;
+
+	pr_info("CU CPU: cores=%d id_base=%d\n", cu->num_cpu_cores,
+			cu->processor_id_low);
+}
+
+static void kfd_populated_cu_info_gpu(struct kfd_topology_device *dev,
+		struct crat_subtype_computeunit *cu)
+{
+	BUG_ON(!dev);
+	BUG_ON(!cu);
+
+	dev->node_props.simd_id_base = cu->processor_id_low;
+	dev->node_props.simd_count = cu->num_simd_cores;
+	dev->node_props.lds_size_in_kb = cu->lds_size_in_kb;
+	dev->node_props.max_waves_per_simd = cu->max_waves_simd;
+	dev->node_props.wave_front_size = cu->wave_front_size;
+	dev->node_props.mem_banks_count = cu->num_banks;
+	dev->node_props.array_count = cu->num_arrays;
+	dev->node_props.cu_per_simd_array = cu->num_cu_per_array;
+	dev->node_props.simd_per_cu = cu->num_simd_per_cu;
+	dev->node_props.max_slots_scratch_cu = cu->max_slots_scatch_cu;
+	if (cu->hsa_capability & CRAT_CU_FLAGS_HOT_PLUGGABLE)
+		dev->node_props.capability |= HSA_CAP_HOT_PLUGGABLE;
+	pr_info("CU GPU: simds=%d id_base=%d\n", cu->num_simd_cores,
+				cu->processor_id_low);
+}
+
+/* kfd_parse_subtype_cu is called when the topology mutex is already acquired */
+static int kfd_parse_subtype_cu(struct crat_subtype_computeunit *cu)
+{
+	struct kfd_topology_device *dev;
+	int i = 0;
+
+	BUG_ON(!cu);
+
+	pr_info("Found CU entry in CRAT table with proximity_domain=%d caps=%x\n",
+			cu->proximity_domain, cu->hsa_capability);
+	list_for_each_entry(dev, &topology_device_list, list) {
+		if (cu->proximity_domain == i) {
+			if (cu->flags & CRAT_CU_FLAGS_CPU_PRESENT)
+				kfd_populated_cu_info_cpu(dev, cu);
+
+			if (cu->flags & CRAT_CU_FLAGS_GPU_PRESENT)
+				kfd_populated_cu_info_gpu(dev, cu);
+			break;
+		}
+		i++;
+	}
+
+	return 0;
+}
+
+/*
+ * kfd_parse_subtype_mem is called when the topology mutex is
+ * already acquired
+ */
+static int kfd_parse_subtype_mem(struct crat_subtype_memory *mem)
+{
+	struct kfd_mem_properties *props;
+	struct kfd_topology_device *dev;
+	int i = 0;
+
+	BUG_ON(!mem);
+
+	pr_info("Found memory entry in CRAT table with proximity_domain=%d\n",
+			mem->promixity_domain);
+	list_for_each_entry(dev, &topology_device_list, list) {
+		if (mem->promixity_domain == i) {
+			props = kfd_alloc_struct(props);
+			if (props == NULL)
+				return -ENOMEM;
+
+			if (dev->node_props.cpu_cores_count == 0)
+				props->heap_type = HSA_MEM_HEAP_TYPE_FB_PRIVATE;
+			else
+				props->heap_type = HSA_MEM_HEAP_TYPE_SYSTEM;
+
+			if (mem->flags & CRAT_MEM_FLAGS_HOT_PLUGGABLE)
+				props->flags |= HSA_MEM_FLAGS_HOT_PLUGGABLE;
+			if (mem->flags & CRAT_MEM_FLAGS_NON_VOLATILE)
+				props->flags |= HSA_MEM_FLAGS_NON_VOLATILE;
+
+			props->size_in_bytes =
+				((uint64_t)mem->length_high << 32) +
+							mem->length_low;
+			props->width = mem->width;
+
+			dev->mem_bank_count++;
+			list_add_tail(&props->list, &dev->mem_props);
+
+			break;
+		}
+		i++;
+	}
+
+	return 0;
+}
+
+/*
+ * kfd_parse_subtype_cache is called when the topology mutex
+ * is already acquired
+ */
+static int kfd_parse_subtype_cache(struct crat_subtype_cache *cache)
+{
+	struct kfd_cache_properties *props;
+	struct kfd_topology_device *dev;
+	uint32_t id;
+
+	BUG_ON(!cache);
+
+	id = cache->processor_id_low;
+
+	pr_info("Found cache entry in CRAT table with processor_id=%d\n", id);
+	list_for_each_entry(dev, &topology_device_list, list)
+		if (id == dev->node_props.cpu_core_id_base ||
+		    id == dev->node_props.simd_id_base) {
+			props = kfd_alloc_struct(props);
+			if (props == NULL)
+				return -ENOMEM;
+
+			props->processor_id_low = id;
+			props->cache_level = cache->cache_level;
+			props->cache_size = cache->cache_size;
+			props->cacheline_size = cache->cache_line_size;
+			props->cachelines_per_tag = cache->lines_per_tag;
+			props->cache_assoc = cache->associativity;
+			props->cache_latency = cache->cache_latency;
+
+			if (cache->flags & CRAT_CACHE_FLAGS_DATA_CACHE)
+				props->cache_type |= HSA_CACHE_TYPE_DATA;
+			if (cache->flags & CRAT_CACHE_FLAGS_INST_CACHE)
+				props->cache_type |= HSA_CACHE_TYPE_INSTRUCTION;
+			if (cache->flags & CRAT_CACHE_FLAGS_CPU_CACHE)
+				props->cache_type |= HSA_CACHE_TYPE_CPU;
+			if (cache->flags & CRAT_CACHE_FLAGS_SIMD_CACHE)
+				props->cache_type |= HSA_CACHE_TYPE_HSACU;
+
+			dev->cache_count++;
+			dev->node_props.caches_count++;
+			list_add_tail(&props->list, &dev->cache_props);
+
+			break;
+		}
+
+	return 0;
+}
+
+/*
+ * kfd_parse_subtype_iolink is called when the topology mutex
+ * is already acquired
+ */
+static int kfd_parse_subtype_iolink(struct crat_subtype_iolink *iolink)
+{
+	struct kfd_iolink_properties *props;
+	struct kfd_topology_device *dev;
+	uint32_t i = 0;
+	uint32_t id_from;
+	uint32_t id_to;
+
+	BUG_ON(!iolink);
+
+	id_from = iolink->proximity_domain_from;
+	id_to = iolink->proximity_domain_to;
+
+	pr_info("Found IO link entry in CRAT table with id_from=%d\n", id_from);
+	list_for_each_entry(dev, &topology_device_list, list) {
+		if (id_from == i) {
+			props = kfd_alloc_struct(props);
+			if (props == NULL)
+				return -ENOMEM;
+
+			props->node_from = id_from;
+			props->node_to = id_to;
+			props->ver_maj = iolink->version_major;
+			props->ver_min = iolink->version_minor;
+
+			/*
+			 * weight factor (derived from CDIR), currently always 1
+			 */
+			props->weight = 1;
+
+			props->min_latency = iolink->minimum_latency;
+			props->max_latency = iolink->maximum_latency;
+			props->min_bandwidth = iolink->minimum_bandwidth_mbs;
+			props->max_bandwidth = iolink->maximum_bandwidth_mbs;
+			props->rec_transfer_size =
+					iolink->recommended_transfer_size;
+
+			dev->io_link_count++;
+			dev->node_props.io_links_count++;
+			list_add_tail(&props->list, &dev->io_link_props);
+
+			break;
+		}
+		i++;
+	}
+
+	return 0;
+}
+
+static int kfd_parse_subtype(struct crat_subtype_generic *sub_type_hdr)
+{
+	struct crat_subtype_computeunit *cu;
+	struct crat_subtype_memory *mem;
+	struct crat_subtype_cache *cache;
+	struct crat_subtype_iolink *iolink;
+	int ret = 0;
+
+	BUG_ON(!sub_type_hdr);
+
+	switch (sub_type_hdr->type) {
+	case CRAT_SUBTYPE_COMPUTEUNIT_AFFINITY:
+		cu = (struct crat_subtype_computeunit *)sub_type_hdr;
+		ret = kfd_parse_subtype_cu(cu);
+		break;
+	case CRAT_SUBTYPE_MEMORY_AFFINITY:
+		mem = (struct crat_subtype_memory *)sub_type_hdr;
+		ret = kfd_parse_subtype_mem(mem);
+		break;
+	case CRAT_SUBTYPE_CACHE_AFFINITY:
+		cache = (struct crat_subtype_cache *)sub_type_hdr;
+		ret = kfd_parse_subtype_cache(cache);
+		break;
+	case CRAT_SUBTYPE_TLB_AFFINITY:
+		/*
+		 * For now, nothing to do here
+		 */
+		pr_info("Found TLB entry in CRAT table (not processing)\n");
+		break;
+	case CRAT_SUBTYPE_CCOMPUTE_AFFINITY:
+		/*
+		 * For now, nothing to do here
+		 */
+		pr_info("Found CCOMPUTE entry in CRAT table (not processing)\n");
+		break;
+	case CRAT_SUBTYPE_IOLINK_AFFINITY:
+		iolink = (struct crat_subtype_iolink *)sub_type_hdr;
+		ret = kfd_parse_subtype_iolink(iolink);
+		break;
+	default:
+		pr_warn("Unknown subtype (%d) in CRAT\n",
+				sub_type_hdr->type);
+	}
+
+	return ret;
+}
+
+static void kfd_release_topology_device(struct kfd_topology_device *dev)
+{
+	struct kfd_mem_properties *mem;
+	struct kfd_cache_properties *cache;
+	struct kfd_iolink_properties *iolink;
+
+	BUG_ON(!dev);
+
+	list_del(&dev->list);
+
+	while (dev->mem_props.next != &dev->mem_props) {
+		mem = container_of(dev->mem_props.next,
+				struct kfd_mem_properties, list);
+		list_del(&mem->list);
+		kfree(mem);
+	}
+
+	while (dev->cache_props.next != &dev->cache_props) {
+		cache = container_of(dev->cache_props.next,
+				struct kfd_cache_properties, list);
+		list_del(&cache->list);
+		kfree(cache);
+	}
+
+	while (dev->io_link_props.next != &dev->io_link_props) {
+		iolink = container_of(dev->io_link_props.next,
+				struct kfd_iolink_properties, list);
+		list_del(&iolink->list);
+		kfree(iolink);
+	}
+
+	kfree(dev);
+
+	sys_props.num_devices--;
+}
+
+static void kfd_release_live_view(void)
+{
+	struct kfd_topology_device *dev;
+
+	while (topology_device_list.next != &topology_device_list) {
+		dev = container_of(topology_device_list.next,
+				 struct kfd_topology_device, list);
+		kfd_release_topology_device(dev);
+}
+
+	memset(&sys_props, 0, sizeof(sys_props));
+}
+
+static struct kfd_topology_device *kfd_create_topology_device(void)
+{
+	struct kfd_topology_device *dev;
+
+	dev = kfd_alloc_struct(dev);
+	if (dev == NULL) {
+		pr_err("No memory to allocate a topology device");
+		return NULL;
+	}
+
+	INIT_LIST_HEAD(&dev->mem_props);
+	INIT_LIST_HEAD(&dev->cache_props);
+	INIT_LIST_HEAD(&dev->io_link_props);
+
+	list_add_tail(&dev->list, &topology_device_list);
+	sys_props.num_devices++;
+
+	return dev;
+}
+
+static int kfd_parse_crat_table(void *crat_image)
+{
+	struct kfd_topology_device *top_dev;
+	struct crat_subtype_generic *sub_type_hdr;
+	uint16_t node_id;
+	int ret;
+	struct crat_header *crat_table = (struct crat_header *)crat_image;
+	uint16_t num_nodes;
+	uint32_t image_len;
+
+	if (!crat_image)
+		return -EINVAL;
+
+	num_nodes = crat_table->num_domains;
+	image_len = crat_table->length;
+
+	pr_info("Parsing CRAT table with %d nodes\n", num_nodes);
+
+	for (node_id = 0; node_id < num_nodes; node_id++) {
+		top_dev = kfd_create_topology_device();
+		if (!top_dev) {
+			kfd_release_live_view();
+			return -ENOMEM;
+		}
+	}
+
+	sys_props.platform_id =
+		(*((uint64_t *)crat_table->oem_id)) & CRAT_OEMID_64BIT_MASK;
+	sys_props.platform_oem = *((uint64_t *)crat_table->oem_table_id);
+	sys_props.platform_rev = crat_table->revision;
+
+	sub_type_hdr = (struct crat_subtype_generic *)(crat_table+1);
+	while ((char *)sub_type_hdr + sizeof(struct crat_subtype_generic) <
+			((char *)crat_image) + image_len) {
+		if (sub_type_hdr->flags & CRAT_SUBTYPE_FLAGS_ENABLED) {
+			ret = kfd_parse_subtype(sub_type_hdr);
+			if (ret != 0) {
+				kfd_release_live_view();
+				return ret;
+			}
+		}
+
+		sub_type_hdr = (typeof(sub_type_hdr))((char *)sub_type_hdr +
+				sub_type_hdr->length);
+	}
+
+	sys_props.generation_count++;
+	topology_crat_parsed = 1;
+
+	return 0;
+}
+
+
+#define sysfs_show_gen_prop(buffer, fmt, ...) \
+		snprintf(buffer, PAGE_SIZE, "%s"fmt, buffer, __VA_ARGS__)
+#define sysfs_show_32bit_prop(buffer, name, value) \
+		sysfs_show_gen_prop(buffer, "%s %u\n", name, value)
+#define sysfs_show_64bit_prop(buffer, name, value) \
+		sysfs_show_gen_prop(buffer, "%s %llu\n", name, value)
+#define sysfs_show_32bit_val(buffer, value) \
+		sysfs_show_gen_prop(buffer, "%u\n", value)
+#define sysfs_show_str_val(buffer, value) \
+		sysfs_show_gen_prop(buffer, "%s\n", value)
+
+static ssize_t sysprops_show(struct kobject *kobj, struct attribute *attr,
+		char *buffer)
+{
+	ssize_t ret;
+
+	/* Making sure that the buffer is an empty string */
+	buffer[0] = 0;
+
+	if (attr == &sys_props.attr_genid) {
+		ret = sysfs_show_32bit_val(buffer, sys_props.generation_count);
+	} else if (attr == &sys_props.attr_props) {
+		sysfs_show_64bit_prop(buffer, "platform_oem",
+				sys_props.platform_oem);
+		sysfs_show_64bit_prop(buffer, "platform_id",
+				sys_props.platform_id);
+		ret = sysfs_show_64bit_prop(buffer, "platform_rev",
+				sys_props.platform_rev);
+	} else {
+		ret = -EINVAL;
+	}
+
+	return ret;
+}
+
+static const struct sysfs_ops sysprops_ops = {
+	.show = sysprops_show,
+};
+
+static struct kobj_type sysprops_type = {
+	.sysfs_ops = &sysprops_ops,
+};
+
+static ssize_t iolink_show(struct kobject *kobj, struct attribute *attr,
+		char *buffer)
+{
+	ssize_t ret;
+	struct kfd_iolink_properties *iolink;
+
+	/* Making sure that the buffer is an empty string */
+	buffer[0] = 0;
+
+	iolink = container_of(attr, struct kfd_iolink_properties, attr);
+	sysfs_show_32bit_prop(buffer, "type", iolink->iolink_type);
+	sysfs_show_32bit_prop(buffer, "version_major", iolink->ver_maj);
+	sysfs_show_32bit_prop(buffer, "version_minor", iolink->ver_min);
+	sysfs_show_32bit_prop(buffer, "node_from", iolink->node_from);
+	sysfs_show_32bit_prop(buffer, "node_to", iolink->node_to);
+	sysfs_show_32bit_prop(buffer, "weight", iolink->weight);
+	sysfs_show_32bit_prop(buffer, "min_latency", iolink->min_latency);
+	sysfs_show_32bit_prop(buffer, "max_latency", iolink->max_latency);
+	sysfs_show_32bit_prop(buffer, "min_bandwidth", iolink->min_bandwidth);
+	sysfs_show_32bit_prop(buffer, "max_bandwidth", iolink->max_bandwidth);
+	sysfs_show_32bit_prop(buffer, "recommended_transfer_size",
+			iolink->rec_transfer_size);
+	ret = sysfs_show_32bit_prop(buffer, "flags", iolink->flags);
+
+	return ret;
+}
+
+static const struct sysfs_ops iolink_ops = {
+	.show = iolink_show,
+};
+
+static struct kobj_type iolink_type = {
+	.sysfs_ops = &iolink_ops,
+};
+
+static ssize_t mem_show(struct kobject *kobj, struct attribute *attr,
+		char *buffer)
+{
+	ssize_t ret;
+	struct kfd_mem_properties *mem;
+
+	/* Making sure that the buffer is an empty string */
+	buffer[0] = 0;
+
+	mem = container_of(attr, struct kfd_mem_properties, attr);
+	sysfs_show_32bit_prop(buffer, "heap_type", mem->heap_type);
+	sysfs_show_64bit_prop(buffer, "size_in_bytes", mem->size_in_bytes);
+	sysfs_show_32bit_prop(buffer, "flags", mem->flags);
+	sysfs_show_32bit_prop(buffer, "width", mem->width);
+	ret = sysfs_show_32bit_prop(buffer, "mem_clk_max", mem->mem_clk_max);
+
+	return ret;
+}
+
+static const struct sysfs_ops mem_ops = {
+	.show = mem_show,
+};
+
+static struct kobj_type mem_type = {
+	.sysfs_ops = &mem_ops,
+};
+
+static ssize_t kfd_cache_show(struct kobject *kobj, struct attribute *attr,
+		char *buffer)
+{
+	ssize_t ret;
+	uint32_t i;
+	struct kfd_cache_properties *cache;
+
+	/* Making sure that the buffer is an empty string */
+	buffer[0] = 0;
+
+	cache = container_of(attr, struct kfd_cache_properties, attr);
+	sysfs_show_32bit_prop(buffer, "processor_id_low",
+			cache->processor_id_low);
+	sysfs_show_32bit_prop(buffer, "level", cache->cache_level);
+	sysfs_show_32bit_prop(buffer, "size", cache->cache_size);
+	sysfs_show_32bit_prop(buffer, "cache_line_size", cache->cacheline_size);
+	sysfs_show_32bit_prop(buffer, "cache_lines_per_tag",
+			cache->cachelines_per_tag);
+	sysfs_show_32bit_prop(buffer, "association", cache->cache_assoc);
+	sysfs_show_32bit_prop(buffer, "latency", cache->cache_latency);
+	sysfs_show_32bit_prop(buffer, "type", cache->cache_type);
+	snprintf(buffer, PAGE_SIZE, "%ssibling_map ", buffer);
+	for (i = 0; i < KFD_TOPOLOGY_CPU_SIBLINGS; i++)
+		ret = snprintf(buffer, PAGE_SIZE, "%s%d%s",
+				buffer, cache->sibling_map[i],
+				(i == KFD_TOPOLOGY_CPU_SIBLINGS-1) ?
+						"\n" : ",");
+
+	return ret;
+}
+
+static const struct sysfs_ops cache_ops = {
+	.show = kfd_cache_show,
+};
+
+static struct kobj_type cache_type = {
+	.sysfs_ops = &cache_ops,
+};
+
+static ssize_t node_show(struct kobject *kobj, struct attribute *attr,
+		char *buffer)
+{
+	ssize_t ret;
+	struct kfd_topology_device *dev;
+	char public_name[KFD_TOPOLOGY_PUBLIC_NAME_SIZE];
+	uint32_t i;
+
+	/* Making sure that the buffer is an empty string */
+	buffer[0] = 0;
+
+	if (strcmp(attr->name, "gpu_id") == 0) {
+		dev = container_of(attr, struct kfd_topology_device,
+				attr_gpuid);
+		ret = sysfs_show_32bit_val(buffer, dev->gpu_id);
+	} else if (strcmp(attr->name, "name") == 0) {
+		dev = container_of(attr, struct kfd_topology_device,
+				attr_name);
+		for (i = 0; i < KFD_TOPOLOGY_PUBLIC_NAME_SIZE; i++) {
+			public_name[i] =
+					(char)dev->node_props.marketing_name[i];
+			if (dev->node_props.marketing_name[i] == 0)
+				break;
+		}
+		public_name[KFD_TOPOLOGY_PUBLIC_NAME_SIZE-1] = 0x0;
+		ret = sysfs_show_str_val(buffer, public_name);
+	} else {
+		dev = container_of(attr, struct kfd_topology_device,
+				attr_props);
+		sysfs_show_32bit_prop(buffer, "cpu_cores_count",
+				dev->node_props.cpu_cores_count);
+		sysfs_show_32bit_prop(buffer, "simd_count",
+				dev->node_props.simd_count);
+
+		if (dev->mem_bank_count < dev->node_props.mem_banks_count) {
+			pr_warn("kfd: mem_banks_count truncated from %d to %d\n",
+					dev->node_props.mem_banks_count,
+					dev->mem_bank_count);
+			sysfs_show_32bit_prop(buffer, "mem_banks_count",
+					dev->mem_bank_count);
+		} else {
+			sysfs_show_32bit_prop(buffer, "mem_banks_count",
+					dev->node_props.mem_banks_count);
+		}
+
+		sysfs_show_32bit_prop(buffer, "caches_count",
+				dev->node_props.caches_count);
+		sysfs_show_32bit_prop(buffer, "io_links_count",
+				dev->node_props.io_links_count);
+		sysfs_show_32bit_prop(buffer, "cpu_core_id_base",
+				dev->node_props.cpu_core_id_base);
+		sysfs_show_32bit_prop(buffer, "simd_id_base",
+				dev->node_props.simd_id_base);
+		sysfs_show_32bit_prop(buffer, "capability",
+				dev->node_props.capability);
+		sysfs_show_32bit_prop(buffer, "max_waves_per_simd",
+				dev->node_props.max_waves_per_simd);
+		sysfs_show_32bit_prop(buffer, "lds_size_in_kb",
+				dev->node_props.lds_size_in_kb);
+		sysfs_show_32bit_prop(buffer, "gds_size_in_kb",
+				dev->node_props.gds_size_in_kb);
+		sysfs_show_32bit_prop(buffer, "wave_front_size",
+				dev->node_props.wave_front_size);
+		sysfs_show_32bit_prop(buffer, "array_count",
+				dev->node_props.array_count);
+		sysfs_show_32bit_prop(buffer, "simd_arrays_per_engine",
+				dev->node_props.simd_arrays_per_engine);
+		sysfs_show_32bit_prop(buffer, "cu_per_simd_array",
+				dev->node_props.cu_per_simd_array);
+		sysfs_show_32bit_prop(buffer, "simd_per_cu",
+				dev->node_props.simd_per_cu);
+		sysfs_show_32bit_prop(buffer, "max_slots_scratch_cu",
+				dev->node_props.max_slots_scratch_cu);
+		sysfs_show_32bit_prop(buffer, "vendor_id",
+				dev->node_props.vendor_id);
+		sysfs_show_32bit_prop(buffer, "device_id",
+				dev->node_props.device_id);
+		sysfs_show_32bit_prop(buffer, "location_id",
+				dev->node_props.location_id);
+
+		if (dev->gpu) {
+			sysfs_show_32bit_prop(buffer, "max_engine_clk_fcompute",
+					kfd2kgd->get_max_engine_clock_in_mhz(
+						dev->gpu->kgd));
+			sysfs_show_64bit_prop(buffer, "local_mem_size",
+					kfd2kgd->get_vmem_size(dev->gpu->kgd));
+
+			sysfs_show_32bit_prop(buffer, "fw_version",
+					kfd2kgd->get_fw_version(
+							dev->gpu->kgd,
+							KGD_ENGINE_MEC1));
+
+		}
+
+		ret = sysfs_show_32bit_prop(buffer, "max_engine_clk_ccompute",
+				cpufreq_quick_get_max(0)/1000);
+	}
+
+	return ret;
+}
+
+static const struct sysfs_ops node_ops = {
+	.show = node_show,
+};
+
+static struct kobj_type node_type = {
+	.sysfs_ops = &node_ops,
+};
+
+static void kfd_remove_sysfs_file(struct kobject *kobj, struct attribute *attr)
+{
+	sysfs_remove_file(kobj, attr);
+	kobject_del(kobj);
+	kobject_put(kobj);
+}
+
+static void kfd_remove_sysfs_node_entry(struct kfd_topology_device *dev)
+{
+	struct kfd_iolink_properties *iolink;
+	struct kfd_cache_properties *cache;
+	struct kfd_mem_properties *mem;
+
+	BUG_ON(!dev);
+
+	if (dev->kobj_iolink) {
+		list_for_each_entry(iolink, &dev->io_link_props, list)
+			if (iolink->kobj) {
+				kfd_remove_sysfs_file(iolink->kobj,
+							&iolink->attr);
+				iolink->kobj = NULL;
+			}
+		kobject_del(dev->kobj_iolink);
+		kobject_put(dev->kobj_iolink);
+		dev->kobj_iolink = NULL;
+	}
+
+	if (dev->kobj_cache) {
+		list_for_each_entry(cache, &dev->cache_props, list)
+			if (cache->kobj) {
+				kfd_remove_sysfs_file(cache->kobj,
+							&cache->attr);
+				cache->kobj = NULL;
+			}
+		kobject_del(dev->kobj_cache);
+		kobject_put(dev->kobj_cache);
+		dev->kobj_cache = NULL;
+	}
+
+	if (dev->kobj_mem) {
+		list_for_each_entry(mem, &dev->mem_props, list)
+			if (mem->kobj) {
+				kfd_remove_sysfs_file(mem->kobj, &mem->attr);
+				mem->kobj = NULL;
+			}
+		kobject_del(dev->kobj_mem);
+		kobject_put(dev->kobj_mem);
+		dev->kobj_mem = NULL;
+	}
+
+	if (dev->kobj_node) {
+		sysfs_remove_file(dev->kobj_node, &dev->attr_gpuid);
+		sysfs_remove_file(dev->kobj_node, &dev->attr_name);
+		sysfs_remove_file(dev->kobj_node, &dev->attr_props);
+		kobject_del(dev->kobj_node);
+		kobject_put(dev->kobj_node);
+		dev->kobj_node = NULL;
+	}
+}
+
+static int kfd_build_sysfs_node_entry(struct kfd_topology_device *dev,
+		uint32_t id)
+{
+	struct kfd_iolink_properties *iolink;
+	struct kfd_cache_properties *cache;
+	struct kfd_mem_properties *mem;
+	int ret;
+	uint32_t i;
+
+	BUG_ON(!dev);
+
+	/*
+	 * Creating the sysfs folders
+	 */
+	BUG_ON(dev->kobj_node);
+	dev->kobj_node = kfd_alloc_struct(dev->kobj_node);
+	if (!dev->kobj_node)
+		return -ENOMEM;
+
+	ret = kobject_init_and_add(dev->kobj_node, &node_type,
+			sys_props.kobj_nodes, "%d", id);
+	if (ret < 0)
+		return ret;
+
+	dev->kobj_mem = kobject_create_and_add("mem_banks", dev->kobj_node);
+	if (!dev->kobj_mem)
+		return -ENOMEM;
+
+	dev->kobj_cache = kobject_create_and_add("caches", dev->kobj_node);
+	if (!dev->kobj_cache)
+		return -ENOMEM;
+
+	dev->kobj_iolink = kobject_create_and_add("io_links", dev->kobj_node);
+	if (!dev->kobj_iolink)
+		return -ENOMEM;
+
+	/*
+	 * Creating sysfs files for node properties
+	 */
+	dev->attr_gpuid.name = "gpu_id";
+	dev->attr_gpuid.mode = KFD_SYSFS_FILE_MODE;
+	sysfs_attr_init(&dev->attr_gpuid);
+	dev->attr_name.name = "name";
+	dev->attr_name.mode = KFD_SYSFS_FILE_MODE;
+	sysfs_attr_init(&dev->attr_name);
+	dev->attr_props.name = "properties";
+	dev->attr_props.mode = KFD_SYSFS_FILE_MODE;
+	sysfs_attr_init(&dev->attr_props);
+	ret = sysfs_create_file(dev->kobj_node, &dev->attr_gpuid);
+	if (ret < 0)
+		return ret;
+	ret = sysfs_create_file(dev->kobj_node, &dev->attr_name);
+	if (ret < 0)
+		return ret;
+	ret = sysfs_create_file(dev->kobj_node, &dev->attr_props);
+	if (ret < 0)
+		return ret;
+
+	i = 0;
+	list_for_each_entry(mem, &dev->mem_props, list) {
+		mem->kobj = kzalloc(sizeof(struct kobject), GFP_KERNEL);
+		if (!mem->kobj)
+			return -ENOMEM;
+		ret = kobject_init_and_add(mem->kobj, &mem_type,
+				dev->kobj_mem, "%d", i);
+		if (ret < 0)
+			return ret;
+
+		mem->attr.name = "properties";
+		mem->attr.mode = KFD_SYSFS_FILE_MODE;
+		sysfs_attr_init(&mem->attr);
+		ret = sysfs_create_file(mem->kobj, &mem->attr);
+		if (ret < 0)
+			return ret;
+		i++;
+	}
+
+	i = 0;
+	list_for_each_entry(cache, &dev->cache_props, list) {
+		cache->kobj = kzalloc(sizeof(struct kobject), GFP_KERNEL);
+		if (!cache->kobj)
+			return -ENOMEM;
+		ret = kobject_init_and_add(cache->kobj, &cache_type,
+				dev->kobj_cache, "%d", i);
+		if (ret < 0)
+			return ret;
+
+		cache->attr.name = "properties";
+		cache->attr.mode = KFD_SYSFS_FILE_MODE;
+		sysfs_attr_init(&cache->attr);
+		ret = sysfs_create_file(cache->kobj, &cache->attr);
+		if (ret < 0)
+			return ret;
+		i++;
+	}
+
+	i = 0;
+	list_for_each_entry(iolink, &dev->io_link_props, list) {
+		iolink->kobj = kzalloc(sizeof(struct kobject), GFP_KERNEL);
+		if (!iolink->kobj)
+			return -ENOMEM;
+		ret = kobject_init_and_add(iolink->kobj, &iolink_type,
+				dev->kobj_iolink, "%d", i);
+		if (ret < 0)
+			return ret;
+
+		iolink->attr.name = "properties";
+		iolink->attr.mode = KFD_SYSFS_FILE_MODE;
+		sysfs_attr_init(&iolink->attr);
+		ret = sysfs_create_file(iolink->kobj, &iolink->attr);
+		if (ret < 0)
+			return ret;
+		i++;
+}
+
+	return 0;
+}
+
+static int kfd_build_sysfs_node_tree(void)
+{
+	struct kfd_topology_device *dev;
+	int ret;
+	uint32_t i = 0;
+
+	list_for_each_entry(dev, &topology_device_list, list) {
+		ret = kfd_build_sysfs_node_entry(dev, i);
+		if (ret < 0)
+			return ret;
+		i++;
+	}
+
+	return 0;
+}
+
+static void kfd_remove_sysfs_node_tree(void)
+{
+	struct kfd_topology_device *dev;
+
+	list_for_each_entry(dev, &topology_device_list, list)
+		kfd_remove_sysfs_node_entry(dev);
+}
+
+static int kfd_topology_update_sysfs(void)
+{
+	int ret;
+
+	pr_info("Creating topology SYSFS entries\n");
+	if (sys_props.kobj_topology == NULL) {
+		sys_props.kobj_topology =
+				kfd_alloc_struct(sys_props.kobj_topology);
+		if (!sys_props.kobj_topology)
+			return -ENOMEM;
+
+		ret = kobject_init_and_add(sys_props.kobj_topology,
+				&sysprops_type,  &kfd_device->kobj,
+				"topology");
+		if (ret < 0)
+			return ret;
+
+		sys_props.kobj_nodes = kobject_create_and_add("nodes",
+				sys_props.kobj_topology);
+		if (!sys_props.kobj_nodes)
+			return -ENOMEM;
+
+		sys_props.attr_genid.name = "generation_id";
+		sys_props.attr_genid.mode = KFD_SYSFS_FILE_MODE;
+		sysfs_attr_init(&sys_props.attr_genid);
+		ret = sysfs_create_file(sys_props.kobj_topology,
+				&sys_props.attr_genid);
+		if (ret < 0)
+			return ret;
+
+		sys_props.attr_props.name = "system_properties";
+		sys_props.attr_props.mode = KFD_SYSFS_FILE_MODE;
+		sysfs_attr_init(&sys_props.attr_props);
+		ret = sysfs_create_file(sys_props.kobj_topology,
+				&sys_props.attr_props);
+		if (ret < 0)
+			return ret;
+	}
+
+	kfd_remove_sysfs_node_tree();
+
+	return kfd_build_sysfs_node_tree();
+}
+
+static void kfd_topology_release_sysfs(void)
+{
+	kfd_remove_sysfs_node_tree();
+	if (sys_props.kobj_topology) {
+		sysfs_remove_file(sys_props.kobj_topology,
+				&sys_props.attr_genid);
+		sysfs_remove_file(sys_props.kobj_topology,
+				&sys_props.attr_props);
+		if (sys_props.kobj_nodes) {
+			kobject_del(sys_props.kobj_nodes);
+			kobject_put(sys_props.kobj_nodes);
+			sys_props.kobj_nodes = NULL;
+		}
+		kobject_del(sys_props.kobj_topology);
+		kobject_put(sys_props.kobj_topology);
+		sys_props.kobj_topology = NULL;
+	}
+}
+
+int kfd_topology_init(void)
+{
+	void *crat_image = NULL;
+	size_t image_size = 0;
+	int ret;
+
+	/*
+	 * Initialize the head for the topology device list
+	 */
+	INIT_LIST_HEAD(&topology_device_list);
+	init_rwsem(&topology_lock);
+	topology_crat_parsed = 0;
+
+	memset(&sys_props, 0, sizeof(sys_props));
+
+	/*
+	 * Get the CRAT image from the ACPI
+	 */
+	ret = kfd_topology_get_crat_acpi(crat_image, &image_size);
+	if (ret == 0 && image_size > 0) {
+		pr_info("Found CRAT image with size=%zd\n", image_size);
+		crat_image = kmalloc(image_size, GFP_KERNEL);
+		if (!crat_image) {
+			ret = -ENOMEM;
+			pr_err("No memory for allocating CRAT image\n");
+			goto err;
+		}
+		ret = kfd_topology_get_crat_acpi(crat_image, &image_size);
+
+		if (ret == 0) {
+			down_write(&topology_lock);
+			ret = kfd_parse_crat_table(crat_image);
+			if (ret == 0)
+				ret = kfd_topology_update_sysfs();
+			up_write(&topology_lock);
+		} else {
+			pr_err("Couldn't get CRAT table size from ACPI\n");
+		}
+		kfree(crat_image);
+	} else if (ret == -ENODATA) {
+		ret = 0;
+	} else {
+		pr_err("Couldn't get CRAT table size from ACPI\n");
+	}
+
+err:
+	pr_info("Finished initializing topology ret=%d\n", ret);
+	return ret;
+}
+
+void kfd_topology_shutdown(void)
+{
+	kfd_topology_release_sysfs();
+	kfd_release_live_view();
+}
+
+static void kfd_debug_print_topology(void)
+{
+	struct kfd_topology_device *dev;
+	uint32_t i = 0;
+
+	pr_info("DEBUG PRINT OF TOPOLOGY:");
+	list_for_each_entry(dev, &topology_device_list, list) {
+		pr_info("Node: %d\n", i);
+		pr_info("\tGPU assigned: %s\n", (dev->gpu ? "yes" : "no"));
+		pr_info("\tCPU count: %d\n", dev->node_props.cpu_cores_count);
+		pr_info("\tSIMD count: %d", dev->node_props.simd_count);
+		i++;
+	}
+}
+
+static uint32_t kfd_generate_gpu_id(struct kfd_dev *gpu)
+{
+	uint32_t hashout;
+	uint32_t buf[7];
+	int i;
+
+	if (!gpu)
+		return 0;
+
+	buf[0] = gpu->pdev->devfn;
+	buf[1] = gpu->pdev->subsystem_vendor;
+	buf[2] = gpu->pdev->subsystem_device;
+	buf[3] = gpu->pdev->device;
+	buf[4] = gpu->pdev->bus->number;
+	buf[5] = (uint32_t)(kfd2kgd->get_vmem_size(gpu->kgd) & 0xffffffff);
+	buf[6] = (uint32_t)(kfd2kgd->get_vmem_size(gpu->kgd) >> 32);
+
+	for (i = 0, hashout = 0; i < 7; i++)
+		hashout ^= hash_32(buf[i], KFD_GPU_ID_HASH_WIDTH);
+
+	return hashout;
+}
+
+static struct kfd_topology_device *kfd_assign_gpu(struct kfd_dev *gpu)
+{
+	struct kfd_topology_device *dev;
+	struct kfd_topology_device *out_dev = NULL;
+
+	BUG_ON(!gpu);
+
+	list_for_each_entry(dev, &topology_device_list, list)
+		if (dev->gpu == NULL && dev->node_props.simd_count > 0) {
+			dev->gpu = gpu;
+			out_dev = dev;
+			break;
+		}
+
+	return out_dev;
+}
+
+static void kfd_notify_gpu_change(uint32_t gpu_id, int arrival)
+{
+	/*
+	 * TODO: Generate an event for thunk about the arrival/removal
+	 * of the GPU
+	 */
+}
+
+int kfd_topology_add_device(struct kfd_dev *gpu)
+{
+	uint32_t gpu_id;
+	struct kfd_topology_device *dev;
+	int res;
+
+	BUG_ON(!gpu);
+
+	gpu_id = kfd_generate_gpu_id(gpu);
+
+	pr_debug("kfd: Adding new GPU (ID: 0x%x) to topology\n", gpu_id);
+
+	down_write(&topology_lock);
+	/*
+	 * Try to assign the GPU to existing topology device (generated from
+	 * CRAT table
+	 */
+	dev = kfd_assign_gpu(gpu);
+	if (!dev) {
+		pr_info("GPU was not found in the current topology. Extending.\n");
+		kfd_debug_print_topology();
+		dev = kfd_create_topology_device();
+		if (!dev) {
+			res = -ENOMEM;
+			goto err;
+		}
+		dev->gpu = gpu;
+
+		/*
+		 * TODO: Make a call to retrieve topology information from the
+		 * GPU vBIOS
+		 */
+
+		/*
+		 * Update the SYSFS tree, since we added another topology device
+		 */
+		if (kfd_topology_update_sysfs() < 0)
+			kfd_topology_release_sysfs();
+
+	}
+
+	dev->gpu_id = gpu_id;
+	gpu->id = gpu_id;
+	dev->node_props.vendor_id = gpu->pdev->vendor;
+	dev->node_props.device_id = gpu->pdev->device;
+	dev->node_props.location_id = (gpu->pdev->bus->number << 24) +
+			(gpu->pdev->devfn & 0xffffff);
+	/*
+	 * TODO: Retrieve max engine clock values from KGD
+	 */
+
+	res = 0;
+
+err:
+	up_write(&topology_lock);
+
+	if (res == 0)
+		kfd_notify_gpu_change(gpu_id, 1);
+
+	return res;
+}
+
+int kfd_topology_remove_device(struct kfd_dev *gpu)
+{
+	struct kfd_topology_device *dev;
+	uint32_t gpu_id;
+	int res = -ENODEV;
+
+	BUG_ON(!gpu);
+
+	down_write(&topology_lock);
+
+	list_for_each_entry(dev, &topology_device_list, list)
+		if (dev->gpu == gpu) {
+			gpu_id = dev->gpu_id;
+			kfd_remove_sysfs_node_entry(dev);
+			kfd_release_topology_device(dev);
+			res = 0;
+			if (kfd_topology_update_sysfs() < 0)
+				kfd_topology_release_sysfs();
+			break;
+		}
+
+	up_write(&topology_lock);
+
+	if (res == 0)
+		kfd_notify_gpu_change(gpu_id, 0);
+
+	return res;
+}
+
+/*
+ * When idx is out of bounds, the function will return NULL
+ */
+struct kfd_dev *kfd_topology_enum_kfd_devices(uint8_t idx)
+{
+
+	struct kfd_topology_device *top_dev;
+	struct kfd_dev *device = NULL;
+	uint8_t device_idx = 0;
+
+	down_read(&topology_lock);
+
+	list_for_each_entry(top_dev, &topology_device_list, list) {
+		if (device_idx == idx) {
+			device = top_dev->gpu;
+			break;
+		}
+
+		device_idx++;
+	}
+
+	up_read(&topology_lock);
+
+	return device;
+
+}
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_topology.h b/drivers/gpu/drm/amd/amdkfd/kfd_topology.h
new file mode 100644
index 000000000000..989624b3cd14
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_topology.h
@@ -0,0 +1,168 @@
+/*
+ * Copyright 2014 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ */
+
+#ifndef __KFD_TOPOLOGY_H__
+#define __KFD_TOPOLOGY_H__
+
+#include <linux/types.h>
+#include <linux/list.h>
+#include "kfd_priv.h"
+
+#define KFD_TOPOLOGY_PUBLIC_NAME_SIZE 128
+
+#define HSA_CAP_HOT_PLUGGABLE			0x00000001
+#define HSA_CAP_ATS_PRESENT			0x00000002
+#define HSA_CAP_SHARED_WITH_GRAPHICS		0x00000004
+#define HSA_CAP_QUEUE_SIZE_POW2			0x00000008
+#define HSA_CAP_QUEUE_SIZE_32BIT		0x00000010
+#define HSA_CAP_QUEUE_IDLE_EVENT		0x00000020
+#define HSA_CAP_VA_LIMIT			0x00000040
+#define HSA_CAP_WATCH_POINTS_SUPPORTED		0x00000080
+#define HSA_CAP_WATCH_POINTS_TOTALBITS_MASK	0x00000f00
+#define HSA_CAP_WATCH_POINTS_TOTALBITS_SHIFT	8
+#define HSA_CAP_RESERVED			0xfffff000
+
+struct kfd_node_properties {
+	uint32_t cpu_cores_count;
+	uint32_t simd_count;
+	uint32_t mem_banks_count;
+	uint32_t caches_count;
+	uint32_t io_links_count;
+	uint32_t cpu_core_id_base;
+	uint32_t simd_id_base;
+	uint32_t capability;
+	uint32_t max_waves_per_simd;
+	uint32_t lds_size_in_kb;
+	uint32_t gds_size_in_kb;
+	uint32_t wave_front_size;
+	uint32_t array_count;
+	uint32_t simd_arrays_per_engine;
+	uint32_t cu_per_simd_array;
+	uint32_t simd_per_cu;
+	uint32_t max_slots_scratch_cu;
+	uint32_t engine_id;
+	uint32_t vendor_id;
+	uint32_t device_id;
+	uint32_t location_id;
+	uint32_t max_engine_clk_fcompute;
+	uint32_t max_engine_clk_ccompute;
+	uint16_t marketing_name[KFD_TOPOLOGY_PUBLIC_NAME_SIZE];
+};
+
+#define HSA_MEM_HEAP_TYPE_SYSTEM	0
+#define HSA_MEM_HEAP_TYPE_FB_PUBLIC	1
+#define HSA_MEM_HEAP_TYPE_FB_PRIVATE	2
+#define HSA_MEM_HEAP_TYPE_GPU_GDS	3
+#define HSA_MEM_HEAP_TYPE_GPU_LDS	4
+#define HSA_MEM_HEAP_TYPE_GPU_SCRATCH	5
+
+#define HSA_MEM_FLAGS_HOT_PLUGGABLE	0x00000001
+#define HSA_MEM_FLAGS_NON_VOLATILE	0x00000002
+#define HSA_MEM_FLAGS_RESERVED		0xfffffffc
+
+struct kfd_mem_properties {
+	struct list_head	list;
+	uint32_t		heap_type;
+	uint64_t		size_in_bytes;
+	uint32_t		flags;
+	uint32_t		width;
+	uint32_t		mem_clk_max;
+	struct kobject		*kobj;
+	struct attribute	attr;
+};
+
+#define KFD_TOPOLOGY_CPU_SIBLINGS 256
+
+#define HSA_CACHE_TYPE_DATA		0x00000001
+#define HSA_CACHE_TYPE_INSTRUCTION	0x00000002
+#define HSA_CACHE_TYPE_CPU		0x00000004
+#define HSA_CACHE_TYPE_HSACU		0x00000008
+#define HSA_CACHE_TYPE_RESERVED		0xfffffff0
+
+struct kfd_cache_properties {
+	struct list_head	list;
+	uint32_t		processor_id_low;
+	uint32_t		cache_level;
+	uint32_t		cache_size;
+	uint32_t		cacheline_size;
+	uint32_t		cachelines_per_tag;
+	uint32_t		cache_assoc;
+	uint32_t		cache_latency;
+	uint32_t		cache_type;
+	uint8_t			sibling_map[KFD_TOPOLOGY_CPU_SIBLINGS];
+	struct kobject		*kobj;
+	struct attribute	attr;
+};
+
+struct kfd_iolink_properties {
+	struct list_head	list;
+	uint32_t		iolink_type;
+	uint32_t		ver_maj;
+	uint32_t		ver_min;
+	uint32_t		node_from;
+	uint32_t		node_to;
+	uint32_t		weight;
+	uint32_t		min_latency;
+	uint32_t		max_latency;
+	uint32_t		min_bandwidth;
+	uint32_t		max_bandwidth;
+	uint32_t		rec_transfer_size;
+	uint32_t		flags;
+	struct kobject		*kobj;
+	struct attribute	attr;
+};
+
+struct kfd_topology_device {
+	struct list_head		list;
+	uint32_t			gpu_id;
+	struct kfd_node_properties	node_props;
+	uint32_t			mem_bank_count;
+	struct list_head		mem_props;
+	uint32_t			cache_count;
+	struct list_head		cache_props;
+	uint32_t			io_link_count;
+	struct list_head		io_link_props;
+	struct kfd_dev			*gpu;
+	struct kobject			*kobj_node;
+	struct kobject			*kobj_mem;
+	struct kobject			*kobj_cache;
+	struct kobject			*kobj_iolink;
+	struct attribute		attr_gpuid;
+	struct attribute		attr_name;
+	struct attribute		attr_props;
+};
+
+struct kfd_system_properties {
+	uint32_t		num_devices;     /* Number of H-NUMA nodes */
+	uint32_t		generation_count;
+	uint64_t		platform_oem;
+	uint64_t		platform_id;
+	uint64_t		platform_rev;
+	struct kobject		*kobj_topology;
+	struct kobject		*kobj_nodes;
+	struct attribute	attr_genid;
+	struct attribute	attr_props;
+};
+
+
+
+#endif /* __KFD_TOPOLOGY_H__ */
diff --git a/drivers/gpu/drm/amd/include/kgd_kfd_interface.h b/drivers/gpu/drm/amd/include/kgd_kfd_interface.h
new file mode 100644
index 000000000000..96a512208fad
--- /dev/null
+++ b/drivers/gpu/drm/amd/include/kgd_kfd_interface.h
@@ -0,0 +1,200 @@
+/*
+ * Copyright 2014 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ */
+
+/*
+ * This file defines the private interface between the
+ * AMD kernel graphics drivers and the AMD KFD.
+ */
+
+#ifndef KGD_KFD_INTERFACE_H_INCLUDED
+#define KGD_KFD_INTERFACE_H_INCLUDED
+
+#include <linux/types.h>
+
+struct pci_dev;
+
+#define KFD_INTERFACE_VERSION 1
+
+struct kfd_dev;
+struct kgd_dev;
+
+struct kgd_mem;
+
+enum kgd_memory_pool {
+	KGD_POOL_SYSTEM_CACHEABLE = 1,
+	KGD_POOL_SYSTEM_WRITECOMBINE = 2,
+	KGD_POOL_FRAMEBUFFER = 3,
+};
+
+enum kgd_engine_type {
+	KGD_ENGINE_PFP = 1,
+	KGD_ENGINE_ME,
+	KGD_ENGINE_CE,
+	KGD_ENGINE_MEC1,
+	KGD_ENGINE_MEC2,
+	KGD_ENGINE_RLC,
+	KGD_ENGINE_SDMA,
+	KGD_ENGINE_MAX
+};
+
+struct kgd2kfd_shared_resources {
+	/* Bit n == 1 means VMID n is available for KFD. */
+	unsigned int compute_vmid_bitmap;
+
+	/* Compute pipes are counted starting from MEC0/pipe0 as 0. */
+	unsigned int first_compute_pipe;
+
+	/* Number of MEC pipes available for KFD. */
+	unsigned int compute_pipe_count;
+
+	/* Base address of doorbell aperture. */
+	phys_addr_t doorbell_physical_address;
+
+	/* Size in bytes of doorbell aperture. */
+	size_t doorbell_aperture_size;
+
+	/* Number of bytes at start of aperture reserved for KGD. */
+	size_t doorbell_start_offset;
+};
+
+/**
+ * struct kgd2kfd_calls
+ *
+ * @exit: Notifies amdkfd that kgd module is unloaded
+ *
+ * @probe: Notifies amdkfd about a probe done on a device in the kgd driver.
+ *
+ * @device_init: Initialize the newly probed device (if it is a device that
+ * amdkfd supports)
+ *
+ * @device_exit: Notifies amdkfd about a removal of a kgd device
+ *
+ * @suspend: Notifies amdkfd about a suspend action done to a kgd device
+ *
+ * @resume: Notifies amdkfd about a resume action done to a kgd device
+ *
+ * This structure contains function callback pointers so the kgd driver
+ * will notify to the amdkfd about certain status changes.
+ *
+ */
+struct kgd2kfd_calls {
+	void (*exit)(void);
+	struct kfd_dev* (*probe)(struct kgd_dev *kgd, struct pci_dev *pdev);
+	bool (*device_init)(struct kfd_dev *kfd,
+			const struct kgd2kfd_shared_resources *gpu_resources);
+	void (*device_exit)(struct kfd_dev *kfd);
+	void (*interrupt)(struct kfd_dev *kfd, const void *ih_ring_entry);
+	void (*suspend)(struct kfd_dev *kfd);
+	int (*resume)(struct kfd_dev *kfd);
+};
+
+/**
+ * struct kfd2kgd_calls
+ *
+ * @init_sa_manager: Initialize an instance of the sa manager, used by
+ * amdkfd for all system memory allocations that are mapped to the GART
+ * address space
+ *
+ * @fini_sa_manager: Releases all memory allocations for amdkfd that are
+ * handled by kgd sa manager
+ *
+ * @allocate_mem: Allocate a buffer from amdkfd's sa manager. The buffer can
+ * be used for mqds, hpds, kernel queue, fence and runlists
+ *
+ * @free_mem: Frees a buffer that was allocated by amdkfd's sa manager
+ *
+ * @get_vmem_size: Retrieves (physical) size of VRAM
+ *
+ * @get_gpu_clock_counter: Retrieves GPU clock counter
+ *
+ * @get_max_engine_clock_in_mhz: Retrieves maximum GPU clock in MHz
+ *
+ * @program_sh_mem_settings: A function that should initiate the memory
+ * properties such as main aperture memory type (cache / non cached) and
+ * secondary aperture base address, size and memory type.
+ * This function is used only for no cp scheduling mode.
+ *
+ * @set_pasid_vmid_mapping: Exposes pasid/vmid pair to the H/W for no cp
+ * scheduling mode. Only used for no cp scheduling mode.
+ *
+ * @init_memory: Initializes memory apertures to fixed base/limit address
+ * and non cached memory types.
+ *
+ * @init_pipeline: Initialized the compute pipelines.
+ *
+ * @hqd_load: Loads the mqd structure to a H/W hqd slot. used only for no cp
+ * sceduling mode.
+ *
+ * @hqd_is_occupies: Checks if a hqd slot is occupied.
+ *
+ * @hqd_destroy: Destructs and preempts the queue assigned to that hqd slot.
+ *
+ * @get_fw_version: Returns FW versions from the header
+ *
+ * This structure contains function pointers to services that the kgd driver
+ * provides to amdkfd driver.
+ *
+ */
+struct kfd2kgd_calls {
+	/* Memory management. */
+	int (*init_sa_manager)(struct kgd_dev *kgd, unsigned int size);
+	void (*fini_sa_manager)(struct kgd_dev *kgd);
+	int (*allocate_mem)(struct kgd_dev *kgd, size_t size, size_t alignment,
+			enum kgd_memory_pool pool, struct kgd_mem **mem);
+
+	void (*free_mem)(struct kgd_dev *kgd, struct kgd_mem *mem);
+
+	uint64_t (*get_vmem_size)(struct kgd_dev *kgd);
+	uint64_t (*get_gpu_clock_counter)(struct kgd_dev *kgd);
+
+	uint32_t (*get_max_engine_clock_in_mhz)(struct kgd_dev *kgd);
+
+	/* Register access functions */
+	void (*program_sh_mem_settings)(struct kgd_dev *kgd, uint32_t vmid,
+			uint32_t sh_mem_config,	uint32_t sh_mem_ape1_base,
+			uint32_t sh_mem_ape1_limit, uint32_t sh_mem_bases);
+
+	int (*set_pasid_vmid_mapping)(struct kgd_dev *kgd, unsigned int pasid,
+					unsigned int vmid);
+
+	int (*init_memory)(struct kgd_dev *kgd);
+	int (*init_pipeline)(struct kgd_dev *kgd, uint32_t pipe_id,
+				uint32_t hpd_size, uint64_t hpd_gpu_addr);
+
+	int (*hqd_load)(struct kgd_dev *kgd, void *mqd, uint32_t pipe_id,
+			uint32_t queue_id, uint32_t __user *wptr);
+
+	bool (*hqd_is_occupied)(struct kgd_dev *kgd, uint64_t queue_address,
+				uint32_t pipe_id, uint32_t queue_id);
+
+	int (*hqd_destroy)(struct kgd_dev *kgd, uint32_t reset_type,
+				unsigned int timeout, uint32_t pipe_id,
+				uint32_t queue_id);
+	uint16_t (*get_fw_version)(struct kgd_dev *kgd,
+				enum kgd_engine_type type);
+};
+
+bool kgd2kfd_init(unsigned interface_version,
+		  const struct kfd2kgd_calls *f2g,
+		  const struct kgd2kfd_calls **g2f);
+
+#endif /* KGD_KFD_INTERFACE_H_INCLUDED */
diff --git a/drivers/gpu/drm/armada/armada_crtc.c b/drivers/gpu/drm/armada/armada_crtc.c
index 9a0cc09e6653..e3a7a5078e5c 100644
--- a/drivers/gpu/drm/armada/armada_crtc.c
+++ b/drivers/gpu/drm/armada/armada_crtc.c
@@ -12,6 +12,7 @@
 #include <linux/platform_device.h>
 #include <drm/drmP.h>
 #include <drm/drm_crtc_helper.h>
+#include <drm/drm_plane_helper.h>
 #include "armada_crtc.h"
 #include "armada_drm.h"
 #include "armada_fb.h"
@@ -260,7 +261,7 @@ static void armada_drm_vblank_off(struct armada_crtc *dcrtc)
 	 * Tell the DRM core that vblank IRQs aren't going to happen for
 	 * a while.  This cleans up any pending vblank events for us.
 	 */
-	drm_vblank_off(dev, dcrtc->num);
+	drm_crtc_vblank_off(&dcrtc->crtc);
 
 	/* Handle any pending flip event. */
 	spin_lock_irq(&dev->event_lock);
@@ -289,6 +290,8 @@ static void armada_drm_crtc_dpms(struct drm_crtc *crtc, int dpms)
 		armada_drm_crtc_update(dcrtc);
 		if (dpms_blanked(dpms))
 			armada_drm_vblank_off(dcrtc);
+		else
+			drm_crtc_vblank_on(&dcrtc->crtc);
 	}
 }
 
@@ -526,7 +529,7 @@ static int armada_drm_crtc_mode_set(struct drm_crtc *crtc,
 	/* Wait for pending flips to complete */
 	wait_event(dcrtc->frame_wait, !dcrtc->frame_work);
 
-	drm_vblank_pre_modeset(crtc->dev, dcrtc->num);
+	drm_crtc_vblank_off(crtc);
 
 	crtc->mode = *adj;
 
@@ -617,7 +620,7 @@ static int armada_drm_crtc_mode_set(struct drm_crtc *crtc,
 
 	armada_drm_crtc_update(dcrtc);
 
-	drm_vblank_post_modeset(crtc->dev, dcrtc->num);
+	drm_crtc_vblank_on(crtc);
 	armada_drm_crtc_finish_fb(dcrtc, old_fb, dpms_blanked(dcrtc->dpms));
 
 	return 0;
@@ -945,18 +948,15 @@ static int armada_drm_crtc_page_flip(struct drm_crtc *crtc,
 	armada_reg_queue_end(work->regs, i);
 
 	/*
-	 * Hold the old framebuffer for the work - DRM appears to drop our
-	 * reference to the old framebuffer in drm_mode_page_flip_ioctl().
+	 * Ensure that we hold a reference on the new framebuffer.
+	 * This has to match the behaviour in mode_set.
 	 */
-	drm_framebuffer_reference(work->old_fb);
+	drm_framebuffer_reference(fb);
 
 	ret = armada_drm_crtc_queue_frame_work(dcrtc, work);
 	if (ret) {
-		/*
-		 * Undo our reference above; DRM does not drop the reference
-		 * to this object on error, so that's okay.
-		 */
-		drm_framebuffer_unreference(work->old_fb);
+		/* Undo our reference above */
+		drm_framebuffer_unreference(fb);
 		kfree(work);
 		return ret;
 	}
diff --git a/drivers/gpu/drm/armada/armada_drv.c b/drivers/gpu/drm/armada/armada_drv.c
index e2d5792b140f..b01420c84864 100644
--- a/drivers/gpu/drm/armada/armada_drv.c
+++ b/drivers/gpu/drm/armada/armada_drv.c
@@ -190,6 +190,7 @@ static int armada_drm_load(struct drm_device *dev, unsigned long flags)
 	if (ret)
 		goto err_comp;
 
+	dev->irq_enabled = true;
 	dev->vblank_disable_allowed = 1;
 
 	ret = armada_fbdev_init(dev);
@@ -308,6 +309,7 @@ static struct drm_driver armada_drm_driver = {
 	.postclose		= NULL,
 	.lastclose		= armada_drm_lastclose,
 	.unload			= armada_drm_unload,
+	.set_busid		= drm_platform_set_busid,
 	.get_vblank_counter	= drm_vblank_count,
 	.enable_vblank		= armada_drm_enable_vblank,
 	.disable_vblank		= armada_drm_disable_vblank,
@@ -330,7 +332,7 @@ static struct drm_driver armada_drm_driver = {
 	.desc			= "Armada SoC DRM",
 	.date			= "20120730",
 	.driver_features	= DRIVER_GEM | DRIVER_MODESET |
-				  DRIVER_PRIME,
+				  DRIVER_HAVE_IRQ | DRIVER_PRIME,
 	.ioctls			= armada_ioctls,
 	.fops			= &armada_drm_fops,
 };
@@ -484,7 +486,6 @@ static struct platform_driver armada_drm_platform_driver = {
 	.remove	= armada_drm_remove,
 	.driver	= {
 		.name	= "armada-drm",
-		.owner	= THIS_MODULE,
 	},
 	.id_table = armada_drm_platform_ids,
 };
diff --git a/drivers/gpu/drm/armada/armada_gem.c b/drivers/gpu/drm/armada/armada_gem.c
index 7496f55611a5..ef5feeecec84 100644
--- a/drivers/gpu/drm/armada/armada_gem.c
+++ b/drivers/gpu/drm/armada/armada_gem.c
@@ -226,7 +226,7 @@ struct armada_gem_object *armada_gem_alloc_object(struct drm_device *dev,
 
 	obj->dev_addr = DMA_ERROR_CODE;
 
-	mapping = obj->obj.filp->f_path.dentry->d_inode->i_mapping;
+	mapping = file_inode(obj->obj.filp)->i_mapping;
 	mapping_set_gfp_mask(mapping, GFP_HIGHUSER | __GFP_RECLAIMABLE);
 
 	DRM_DEBUG_DRIVER("alloc obj %p size %zu\n", obj, size);
diff --git a/drivers/gpu/drm/armada/armada_gem.h b/drivers/gpu/drm/armada/armada_gem.h
index 00b6cd461a03..b000ea3a829a 100644
--- a/drivers/gpu/drm/armada/armada_gem.h
+++ b/drivers/gpu/drm/armada/armada_gem.h
@@ -8,6 +8,8 @@
 #ifndef ARMADA_GEM_H
 #define ARMADA_GEM_H
 
+#include <drm/drm_gem.h>
+
 /* GEM */
 struct armada_gem_object {
 	struct drm_gem_object	obj;
diff --git a/drivers/gpu/drm/ast/ast_dp501.c b/drivers/gpu/drm/ast/ast_dp501.c
index 5da4b62285fa..76f07f38b941 100644
--- a/drivers/gpu/drm/ast/ast_dp501.c
+++ b/drivers/gpu/drm/ast/ast_dp501.c
@@ -379,11 +379,39 @@ static bool ast_init_dvo(struct drm_device *dev)
 	return true;
 }
 
+
+static void ast_init_analog(struct drm_device *dev)
+{
+	struct ast_private *ast = dev->dev_private;
+	u32 data;
+
+	/*
+	 * Set DAC source to VGA mode in SCU2C via the P2A
+	 * bridge. First configure the P2U to target the SCU
+	 * in case it isn't at this stage.
+	 */
+	ast_write32(ast, 0xf004, 0x1e6e0000);
+	ast_write32(ast, 0xf000, 0x1);
+
+	/* Then unlock the SCU with the magic password */
+	ast_write32(ast, 0x12000, 0x1688a8a8);
+	ast_write32(ast, 0x12000, 0x1688a8a8);
+	ast_write32(ast, 0x12000, 0x1688a8a8);
+
+	/* Finally, clear bits [17:16] of SCU2c */
+	data = ast_read32(ast, 0x1202c);
+	data &= 0xfffcffff;
+	ast_write32(ast, 0, data);
+
+	/* Disable DVO */
+	ast_set_index_reg_mask(ast, AST_IO_CRTC_PORT, 0xa3, 0xcf, 0x00);
+}
+
 void ast_init_3rdtx(struct drm_device *dev)
 {
 	struct ast_private *ast = dev->dev_private;
 	u8 jreg;
-	u32 data;
+
 	if (ast->chip == AST2300 || ast->chip == AST2400) {
 		jreg = ast_get_index_reg_mask(ast, AST_IO_CRTC_PORT, 0xd1, 0xff);
 		switch (jreg & 0x0e) {
@@ -399,12 +427,8 @@ void ast_init_3rdtx(struct drm_device *dev)
 		default:
 			if (ast->tx_chip_type == AST_TX_SIL164)
 				ast_init_dvo(dev);
-			else {
-				ast_write32(ast, 0x12000, 0x1688a8a8);
-				data = ast_read32(ast, 0x1202c);
-				data &= 0xfffcffff;
-				ast_write32(ast, 0, data);
-			}
+			else
+				ast_init_analog(dev);
 		}
 	}
 }
diff --git a/drivers/gpu/drm/ast/ast_drv.c b/drivers/gpu/drm/ast/ast_drv.c
index f19682a93c24..9a32d9dfdd26 100644
--- a/drivers/gpu/drm/ast/ast_drv.c
+++ b/drivers/gpu/drm/ast/ast_drv.c
@@ -199,6 +199,7 @@ static struct drm_driver driver = {
 
 	.load = ast_driver_load,
 	.unload = ast_driver_unload,
+	.set_busid = drm_pci_set_busid,
 
 	.fops = &ast_fops,
 	.name = DRIVER_NAME,
diff --git a/drivers/gpu/drm/ast/ast_drv.h b/drivers/gpu/drm/ast/ast_drv.h
index 957d4fabf1e1..86205a28e56b 100644
--- a/drivers/gpu/drm/ast/ast_drv.h
+++ b/drivers/gpu/drm/ast/ast_drv.h
@@ -36,6 +36,8 @@
 #include <drm/ttm/ttm_memory.h>
 #include <drm/ttm/ttm_module.h>
 
+#include <drm/drm_gem.h>
+
 #include <linux/i2c.h>
 #include <linux/i2c-algo-bit.h>
 
@@ -125,8 +127,9 @@ struct ast_gem_object;
 
 #define AST_IO_AR_PORT_WRITE		(0x40)
 #define AST_IO_MISC_PORT_WRITE		(0x42)
+#define AST_IO_VGA_ENABLE_PORT		(0x43)
 #define AST_IO_SEQ_PORT			(0x44)
-#define AST_DAC_INDEX_READ		(0x3c7)
+#define AST_IO_DAC_INDEX_READ		(0x47)
 #define AST_IO_DAC_INDEX_WRITE		(0x48)
 #define AST_IO_DAC_DATA		        (0x49)
 #define AST_IO_GR_PORT			(0x4E)
@@ -134,6 +137,8 @@ struct ast_gem_object;
 #define AST_IO_INPUT_STATUS1_READ	(0x5A)
 #define AST_IO_MISC_PORT_READ		(0x4C)
 
+#define AST_IO_MM_OFFSET		(0x380)
+
 #define __ast_read(x) \
 static inline u##x ast_read##x(struct ast_private *ast, u32 reg) { \
 u##x val = 0;\
@@ -316,7 +321,7 @@ struct ast_bo {
 	struct ttm_placement placement;
 	struct ttm_bo_kmap_obj kmap;
 	struct drm_gem_object gem;
-	u32 placements[3];
+	struct ttm_place placements[3];
 	int pin_count;
 };
 #define gem_to_ast_bo(gobj) container_of((gobj), struct ast_bo, gem)
@@ -381,6 +386,9 @@ int ast_bo_push_sysram(struct ast_bo *bo);
 int ast_mmap(struct file *filp, struct vm_area_struct *vma);
 
 /* ast post */
+void ast_enable_vga(struct drm_device *dev);
+void ast_enable_mmio(struct drm_device *dev);
+bool ast_is_vga_enabled(struct drm_device *dev);
 void ast_post_gpu(struct drm_device *dev);
 u32 ast_mindwm(struct ast_private *ast, u32 r);
 void ast_moutdwm(struct ast_private *ast, u32 r, u32 v);
diff --git a/drivers/gpu/drm/ast/ast_fb.c b/drivers/gpu/drm/ast/ast_fb.c
index cba45c774552..5c60ae524c45 100644
--- a/drivers/gpu/drm/ast/ast_fb.c
+++ b/drivers/gpu/drm/ast/ast_fb.c
@@ -186,7 +186,8 @@ static int astfb_create_object(struct ast_fbdev *afbdev,
 static int astfb_create(struct drm_fb_helper *helper,
 			struct drm_fb_helper_surface_size *sizes)
 {
-	struct ast_fbdev *afbdev = (struct ast_fbdev *)helper;
+	struct ast_fbdev *afbdev =
+		container_of(helper, struct ast_fbdev, helper);
 	struct drm_device *dev = afbdev->helper.dev;
 	struct drm_mode_fb_cmd2 mode_cmd;
 	struct drm_framebuffer *fb;
diff --git a/drivers/gpu/drm/ast/ast_main.c b/drivers/gpu/drm/ast/ast_main.c
index b792194e0d9c..035dacc93382 100644
--- a/drivers/gpu/drm/ast/ast_main.c
+++ b/drivers/gpu/drm/ast/ast_main.c
@@ -63,7 +63,7 @@ uint8_t ast_get_index_reg_mask(struct ast_private *ast,
 }
 
 
-static int ast_detect_chip(struct drm_device *dev)
+static int ast_detect_chip(struct drm_device *dev, bool *need_post)
 {
 	struct ast_private *ast = dev->dev_private;
 	uint32_t data, jreg;
@@ -110,6 +110,21 @@ static int ast_detect_chip(struct drm_device *dev)
 		}
 	}
 
+	/*
+	 * If VGA isn't enabled, we need to enable now or subsequent
+	 * access to the scratch registers will fail. We also inform
+	 * our caller that it needs to POST the chip
+	 * (Assumption: VGA not enabled -> need to POST)
+	 */
+	if (!ast_is_vga_enabled(dev)) {
+		ast_enable_vga(dev);
+		ast_enable_mmio(dev);
+		DRM_INFO("VGA not enabled on entry, requesting chip POST\n");
+		*need_post = true;
+	} else
+		*need_post = false;
+
+	/* Check if we support wide screen */
 	switch (ast->chip) {
 	case AST1180:
 		ast->support_wide_screen = true;
@@ -125,6 +140,7 @@ static int ast_detect_chip(struct drm_device *dev)
 			ast->support_wide_screen = true;
 		else {
 			ast->support_wide_screen = false;
+			/* Read SCU7c (silicon revision register) */
 			ast_write32(ast, 0xf004, 0x1e6e0000);
 			ast_write32(ast, 0xf000, 0x1);
 			data = ast_read32(ast, 0x1207c);
@@ -137,11 +153,29 @@ static int ast_detect_chip(struct drm_device *dev)
 		break;
 	}
 
+	/* Check 3rd Tx option (digital output afaik) */
 	ast->tx_chip_type = AST_TX_NONE;
-	jreg = ast_get_index_reg_mask(ast, AST_IO_CRTC_PORT, 0xa3, 0xff);
-	if (jreg & 0x80)
-		ast->tx_chip_type = AST_TX_SIL164;
+
+	/*
+	 * VGACRA3 Enhanced Color Mode Register, check if DVO is already
+	 * enabled, in that case, assume we have a SIL164 TMDS transmitter
+	 *
+	 * Don't make that assumption if we the chip wasn't enabled and
+	 * is at power-on reset, otherwise we'll incorrectly "detect" a
+	 * SIL164 when there is none.
+	 */
+	if (!*need_post) {
+		jreg = ast_get_index_reg_mask(ast, AST_IO_CRTC_PORT, 0xa3, 0xff);
+		if (jreg & 0x80)
+			ast->tx_chip_type = AST_TX_SIL164;
+	}
+
 	if ((ast->chip == AST2300) || (ast->chip == AST2400)) {
+		/*
+		 * On AST2300 and 2400, look the configuration set by the SoC in
+		 * the SOC scratch register #1 bits 11:8 (interestingly marked
+		 * as "reserved" in the spec)
+		 */
 		jreg = ast_get_index_reg_mask(ast, AST_IO_CRTC_PORT, 0xd1, 0xff);
 		switch (jreg) {
 		case 0x04:
@@ -162,6 +196,17 @@ static int ast_detect_chip(struct drm_device *dev)
 		}
 	}
 
+	/* Print stuff for diagnostic purposes */
+	switch(ast->tx_chip_type) {
+	case AST_TX_SIL164:
+		DRM_INFO("Using Sil164 TMDS transmitter\n");
+		break;
+	case AST_TX_DP501:
+		DRM_INFO("Using DP501 DisplayPort transmitter\n");
+		break;
+	default:
+		DRM_INFO("Analog VGA only\n");
+	}
 	return 0;
 }
 
@@ -346,6 +391,7 @@ static u32 ast_get_vram_info(struct drm_device *dev)
 int ast_driver_load(struct drm_device *dev, unsigned long flags)
 {
 	struct ast_private *ast;
+	bool need_post;
 	int ret = 0;
 
 	ast = kzalloc(sizeof(struct ast_private), GFP_KERNEL);
@@ -360,13 +406,27 @@ int ast_driver_load(struct drm_device *dev, unsigned long flags)
 		ret = -EIO;
 		goto out_free;
 	}
-	ast->ioregs = pci_iomap(dev->pdev, 2, 0);
+
+	/*
+	 * If we don't have IO space at all, use MMIO now and
+	 * assume the chip has MMIO enabled by default (rev 0x20
+	 * and higher).
+	 */
+	if (!(pci_resource_flags(dev->pdev, 2) & IORESOURCE_IO)) {
+		DRM_INFO("platform has no IO space, trying MMIO\n");
+		ast->ioregs = ast->regs + AST_IO_MM_OFFSET;
+	}
+
+	/* "map" IO regs if the above hasn't done so already */
 	if (!ast->ioregs) {
-		ret = -EIO;
-		goto out_free;
+		ast->ioregs = pci_iomap(dev->pdev, 2, 0);
+		if (!ast->ioregs) {
+			ret = -EIO;
+			goto out_free;
+		}
 	}
 
-	ast_detect_chip(dev);
+	ast_detect_chip(dev, &need_post);
 
 	if (ast->chip != AST1180) {
 		ast_get_dram_info(dev);
@@ -374,6 +434,9 @@ int ast_driver_load(struct drm_device *dev, unsigned long flags)
 		DRM_INFO("dram %d %d %d %08x\n", ast->mclk, ast->dram_type, ast->dram_bus_width, ast->vram_size);
 	}
 
+	if (need_post)
+		ast_post_gpu(dev);
+
 	ret = ast_mm_init(ast);
 	if (ret)
 		goto out_free;
diff --git a/drivers/gpu/drm/ast/ast_mode.c b/drivers/gpu/drm/ast/ast_mode.c
index 5389350244f2..b7ee2634e47c 100644
--- a/drivers/gpu/drm/ast/ast_mode.c
+++ b/drivers/gpu/drm/ast/ast_mode.c
@@ -31,6 +31,7 @@
 #include <drm/drmP.h>
 #include <drm/drm_crtc.h>
 #include <drm/drm_crtc_helper.h>
+#include <drm/drm_plane_helper.h>
 #include "ast_drv.h"
 
 #include "ast_tables.h"
@@ -80,6 +81,8 @@ static bool ast_get_vbios_mode_info(struct drm_crtc *crtc, struct drm_display_mo
 	struct ast_private *ast = crtc->dev->dev_private;
 	u32 refresh_rate_index = 0, mode_id, color_index, refresh_rate;
 	u32 hborder, vborder;
+	bool check_sync;
+	struct ast_vbios_enhtable *best = NULL;
 
 	switch (crtc->primary->fb->bits_per_pixel) {
 	case 8:
@@ -141,14 +144,34 @@ static bool ast_get_vbios_mode_info(struct drm_crtc *crtc, struct drm_display_mo
 	}
 
 	refresh_rate = drm_mode_vrefresh(mode);
-	while (vbios_mode->enh_table->refresh_rate < refresh_rate) {
-		vbios_mode->enh_table++;
-		if ((vbios_mode->enh_table->refresh_rate > refresh_rate) ||
-		    (vbios_mode->enh_table->refresh_rate == 0xff)) {
-			vbios_mode->enh_table--;
-			break;
+	check_sync = vbios_mode->enh_table->flags & WideScreenMode;
+	do {
+		struct ast_vbios_enhtable *loop = vbios_mode->enh_table;
+
+		while (loop->refresh_rate != 0xff) {
+			if ((check_sync) &&
+			    (((mode->flags & DRM_MODE_FLAG_NVSYNC)  &&
+			      (loop->flags & PVSync))  ||
+			     ((mode->flags & DRM_MODE_FLAG_PVSYNC)  &&
+			      (loop->flags & NVSync))  ||
+			     ((mode->flags & DRM_MODE_FLAG_NHSYNC)  &&
+			      (loop->flags & PHSync))  ||
+			     ((mode->flags & DRM_MODE_FLAG_PHSYNC)  &&
+			      (loop->flags & NHSync)))) {
+				loop++;
+				continue;
+			}
+			if (loop->refresh_rate <= refresh_rate
+			    && (!best || loop->refresh_rate > best->refresh_rate))
+				best = loop;
+			loop++;
 		}
-	}
+		if (best || !check_sync)
+			break;
+		check_sync = 0;
+	} while (1);
+	if (best)
+		vbios_mode->enh_table = best;
 
 	hborder = (vbios_mode->enh_table->flags & HBorder) ? 8 : 0;
 	vborder = (vbios_mode->enh_table->flags & VBorder) ? 8 : 0;
@@ -419,8 +442,10 @@ static void ast_set_sync_reg(struct drm_device *dev, struct drm_display_mode *mo
 	struct ast_private *ast = dev->dev_private;
 	u8 jreg;
 
-	jreg = ast_io_read8(ast, AST_IO_MISC_PORT_READ);
-	jreg |= (vbios_mode->enh_table->flags & SyncNN);
+	jreg  = ast_io_read8(ast, AST_IO_MISC_PORT_READ);
+	jreg &= ~0xC0;
+	if (vbios_mode->enh_table->flags & NVSync) jreg |= 0x80;
+	if (vbios_mode->enh_table->flags & NHSync) jreg |= 0x40;
 	ast_io_write8(ast, AST_IO_MISC_PORT_WRITE, jreg);
 }
 
@@ -1080,8 +1105,8 @@ static u32 copy_cursor_image(u8 *src, u8 *dst, int width, int height)
 			srcdata32[1].ul = *((u32 *)(srcxor + 4)) & 0xf0f0f0f0;
 			data32.b[0] = srcdata32[0].b[1] | (srcdata32[0].b[0] >> 4);
 			data32.b[1] = srcdata32[0].b[3] | (srcdata32[0].b[2] >> 4);
-			data32.b[2] = srcdata32[0].b[1] | (srcdata32[1].b[0] >> 4);
-			data32.b[3] = srcdata32[0].b[3] | (srcdata32[1].b[2] >> 4);
+			data32.b[2] = srcdata32[1].b[1] | (srcdata32[1].b[0] >> 4);
+			data32.b[3] = srcdata32[1].b[3] | (srcdata32[1].b[2] >> 4);
 
 			writel(data32.ul, dstxor);
 			csum += data32.ul;
diff --git a/drivers/gpu/drm/ast/ast_post.c b/drivers/gpu/drm/ast/ast_post.c
index 38d437f3a267..810c51d92b99 100644
--- a/drivers/gpu/drm/ast/ast_post.c
+++ b/drivers/gpu/drm/ast/ast_post.c
@@ -33,18 +33,23 @@
 
 static void ast_init_dram_2300(struct drm_device *dev);
 
-static void
-ast_enable_vga(struct drm_device *dev)
+void ast_enable_vga(struct drm_device *dev)
+{
+	struct ast_private *ast = dev->dev_private;
+
+	ast_io_write8(ast, AST_IO_VGA_ENABLE_PORT, 0x01);
+	ast_io_write8(ast, AST_IO_MISC_PORT_WRITE, 0x01);
+}
+
+void ast_enable_mmio(struct drm_device *dev)
 {
 	struct ast_private *ast = dev->dev_private;
 
-	ast_io_write8(ast, 0x43, 0x01);
-	ast_io_write8(ast, 0x42, 0x01);
+	ast_set_index_reg_mask(ast, AST_IO_CRTC_PORT, 0xa1, 0xff, 0x04);
 }
 
-#if 0 /* will use later */
-static bool
-ast_is_vga_enabled(struct drm_device *dev)
+
+bool ast_is_vga_enabled(struct drm_device *dev)
 {
 	struct ast_private *ast = dev->dev_private;
 	u8 ch;
@@ -52,7 +57,7 @@ ast_is_vga_enabled(struct drm_device *dev)
 	if (ast->chip == AST1180) {
 		/* TODO 1180 */
 	} else {
-		ch = ast_io_read8(ast, 0x43);
+		ch = ast_io_read8(ast, AST_IO_VGA_ENABLE_PORT);
 		if (ch) {
 			ast_open_key(ast);
 			ch = ast_get_index_reg_mask(ast, AST_IO_CRTC_PORT, 0xb6, 0xff);
@@ -61,7 +66,6 @@ ast_is_vga_enabled(struct drm_device *dev)
 	}
 	return 0;
 }
-#endif
 
 static const u8 extreginfo[] = { 0x0f, 0x04, 0x1c, 0xff };
 static const u8 extreginfo_ast2300a0[] = { 0x0f, 0x04, 0x1c, 0xff };
@@ -371,6 +375,7 @@ void ast_post_gpu(struct drm_device *dev)
 	pci_write_config_dword(ast->dev->pdev, 0x04, reg);
 
 	ast_enable_vga(dev);
+	ast_enable_mmio(dev);
 	ast_open_key(ast);
 	ast_set_def_ext_reg(dev);
 
diff --git a/drivers/gpu/drm/ast/ast_tables.h b/drivers/gpu/drm/ast/ast_tables.h
index 05c01ea85294..3608d5aa7451 100644
--- a/drivers/gpu/drm/ast/ast_tables.h
+++ b/drivers/gpu/drm/ast/ast_tables.h
@@ -35,14 +35,18 @@
 #define HalfDCLK                0x00000002
 #define DoubleScanMode          0x00000004
 #define LineCompareOff          0x00000008
-#define SyncPP                  0x00000000
-#define SyncPN                  0x00000040
-#define SyncNP                  0x00000080
-#define SyncNN                  0x000000C0
 #define HBorder                 0x00000020
 #define VBorder                 0x00000010
 #define WideScreenMode		0x00000100
 #define NewModeInfo		0x00000200
+#define NHSync			0x00000400
+#define PHSync			0x00000800
+#define NVSync			0x00001000
+#define PVSync			0x00002000
+#define SyncPP			(PVSync | PHSync)
+#define SyncPN			(PVSync | NHSync)
+#define SyncNP			(NVSync | PHSync)
+#define SyncNN			(NVSync | NHSync)
 
 /* DCLK Index */
 #define VCLK25_175     		0x00
@@ -72,6 +76,7 @@
 #define VCLK119     		0x17
 #define VCLK85_5     		0x18
 #define VCLK97_75     		0x19
+#define VCLK118_25			0x1A
 
 static struct ast_vbios_dclk_info dclk_table[] = {
 	{0x2C, 0xE7, 0x03},					/* 00: VCLK25_175	*/
@@ -100,6 +105,7 @@ static struct ast_vbios_dclk_info dclk_table[] = {
 	{0x77, 0x58, 0x80},					/* 17: VCLK119      */
 	{0x32, 0x67, 0x80},				    /* 18: VCLK85_5     */
 	{0x6a, 0x6d, 0x80},					/* 19: VCLK97_75	*/
+	{0x3b, 0x2c, 0x81},					/* 1A: VCLK118_25	*/
 };
 
 static struct ast_vbios_stdtable vbios_stdtable[] = {
@@ -246,8 +252,10 @@ static struct ast_vbios_enhtable res_1360x768[] = {
 static struct ast_vbios_enhtable res_1600x900[] = {
 	{1760, 1600, 48, 32, 926,  900, 3, 5, VCLK97_75,	/* 60Hz CVT RB */
 	 (SyncNP | Charx8Dot | LineCompareOff | WideScreenMode | NewModeInfo), 60, 1, 0x3A },
-	{1760, 1600, 48, 32, 926,  900, 3, 5, VCLK97_75,	/* end */
-	 (SyncNP | Charx8Dot | LineCompareOff | WideScreenMode | NewModeInfo), 0xFF, 1, 0x3A }
+	{2112, 1600, 88,168, 934,  900, 3, 5, VCLK118_25,	/* 60Hz CVT */
+	 (SyncPN | Charx8Dot | LineCompareOff | WideScreenMode | NewModeInfo), 60, 2, 0x3A },
+	{2112, 1600, 88,168, 934,  900, 3, 5, VCLK118_25,	/* 60Hz CVT */
+	 (SyncPN | Charx8Dot | LineCompareOff | WideScreenMode | NewModeInfo), 0xFF, 2, 0x3A },
 };
 
 static struct ast_vbios_enhtable res_1920x1080[] = {
@@ -261,11 +269,11 @@ static struct ast_vbios_enhtable res_1920x1080[] = {
 /* 16:10 */
 static struct ast_vbios_enhtable res_1280x800[] = {
 	{1440, 1280, 48, 32,  823,  800, 3, 6, VCLK71,	/* 60Hz RB */
-	 (SyncNP | Charx8Dot | LineCompareOff | WideScreenMode | NewModeInfo), 60, 1, 35 },
+	 (SyncNP | Charx8Dot | LineCompareOff | WideScreenMode | NewModeInfo), 60, 1, 0x35 },
 	{1680, 1280, 72,128,  831,  800, 3, 6, VCLK83_5,	/* 60Hz */
-	 (SyncPN | Charx8Dot | LineCompareOff | WideScreenMode | NewModeInfo), 60, 1, 0x35 },
+	 (SyncPN | Charx8Dot | LineCompareOff | WideScreenMode | NewModeInfo), 60, 2, 0x35 },
 	{1680, 1280, 72,128,  831,  800, 3, 6, VCLK83_5,	/* 60Hz */
-	 (SyncPN | Charx8Dot | LineCompareOff | WideScreenMode | NewModeInfo), 0xFF, 1, 0x35 },
+	 (SyncPN | Charx8Dot | LineCompareOff | WideScreenMode | NewModeInfo), 0xFF, 2, 0x35 },
 
 };
 
@@ -273,24 +281,24 @@ static struct ast_vbios_enhtable res_1440x900[] = {
 	{1600, 1440, 48, 32,  926,  900, 3, 6, VCLK88_75,	/* 60Hz RB */
 	 (SyncNP | Charx8Dot | LineCompareOff | WideScreenMode | NewModeInfo), 60, 1, 0x36 },
 	{1904, 1440, 80,152,  934,  900, 3, 6, VCLK106_5,	/* 60Hz */
-	 (SyncPN | Charx8Dot | LineCompareOff | WideScreenMode | NewModeInfo), 60, 1, 0x36 },
+	 (SyncPN | Charx8Dot | LineCompareOff | WideScreenMode | NewModeInfo), 60, 2, 0x36 },
 	{1904, 1440, 80,152,  934,  900, 3, 6, VCLK106_5,	/* 60Hz */
-	 (SyncPN | Charx8Dot | LineCompareOff | WideScreenMode | NewModeInfo), 0xFF, 1, 0x36 },
+	 (SyncPN | Charx8Dot | LineCompareOff | WideScreenMode | NewModeInfo), 0xFF, 2, 0x36 },
 };
 
 static struct ast_vbios_enhtable res_1680x1050[] = {
 	{1840, 1680, 48, 32, 1080, 1050, 3, 6, VCLK119,	/* 60Hz RB */
 	 (SyncNP | Charx8Dot | LineCompareOff | WideScreenMode | NewModeInfo), 60, 1, 0x37 },
 	{2240, 1680,104,176, 1089, 1050, 3, 6, VCLK146_25,	/* 60Hz */
-	 (SyncPN | Charx8Dot | LineCompareOff | WideScreenMode | NewModeInfo), 60, 1, 0x37 },
+	 (SyncPN | Charx8Dot | LineCompareOff | WideScreenMode | NewModeInfo), 60, 2, 0x37 },
 	{2240, 1680,104,176, 1089, 1050, 3, 6, VCLK146_25,	/* 60Hz */
-	 (SyncPN | Charx8Dot | LineCompareOff | WideScreenMode | NewModeInfo), 0xFF, 1, 0x37 },
+	 (SyncPN | Charx8Dot | LineCompareOff | WideScreenMode | NewModeInfo), 0xFF, 2, 0x37 },
 };
 
 static struct ast_vbios_enhtable res_1920x1200[] = {
-	{2080, 1920, 48, 32, 1235, 1200, 3, 6, VCLK154,	/* 60Hz */
+	{2080, 1920, 48, 32, 1235, 1200, 3, 6, VCLK154,	/* 60Hz RB*/
 	 (SyncNP | Charx8Dot | LineCompareOff | WideScreenMode | NewModeInfo), 60, 1, 0x34 },
-	{2080, 1920, 48, 32, 1235, 1200, 3, 6, VCLK154,	/* 60Hz */
+	{2080, 1920, 48, 32, 1235, 1200, 3, 6, VCLK154,	/* 60Hz RB */
 	 (SyncNP | Charx8Dot | LineCompareOff | WideScreenMode | NewModeInfo), 0xFF, 1, 0x34 },
 };
 
diff --git a/drivers/gpu/drm/ast/ast_ttm.c b/drivers/gpu/drm/ast/ast_ttm.c
index b8246227bab0..08f82eae6939 100644
--- a/drivers/gpu/drm/ast/ast_ttm.c
+++ b/drivers/gpu/drm/ast/ast_ttm.c
@@ -293,18 +293,22 @@ void ast_mm_fini(struct ast_private *ast)
 void ast_ttm_placement(struct ast_bo *bo, int domain)
 {
 	u32 c = 0;
-	bo->placement.fpfn = 0;
-	bo->placement.lpfn = 0;
+	unsigned i;
+
 	bo->placement.placement = bo->placements;
 	bo->placement.busy_placement = bo->placements;
 	if (domain & TTM_PL_FLAG_VRAM)
-		bo->placements[c++] = TTM_PL_FLAG_WC | TTM_PL_FLAG_UNCACHED | TTM_PL_FLAG_VRAM;
+		bo->placements[c++].flags = TTM_PL_FLAG_WC | TTM_PL_FLAG_UNCACHED | TTM_PL_FLAG_VRAM;
 	if (domain & TTM_PL_FLAG_SYSTEM)
-		bo->placements[c++] = TTM_PL_MASK_CACHING | TTM_PL_FLAG_SYSTEM;
+		bo->placements[c++].flags = TTM_PL_FLAG_CACHED | TTM_PL_FLAG_SYSTEM;
 	if (!c)
-		bo->placements[c++] = TTM_PL_MASK_CACHING | TTM_PL_FLAG_SYSTEM;
+		bo->placements[c++].flags = TTM_PL_FLAG_CACHED | TTM_PL_FLAG_SYSTEM;
 	bo->placement.num_placement = c;
 	bo->placement.num_busy_placement = c;
+	for (i = 0; i < c; ++i) {
+		bo->placements[i].fpfn = 0;
+		bo->placements[i].lpfn = 0;
+	}
 }
 
 int ast_bo_create(struct drm_device *dev, int size, int align,
@@ -335,7 +339,7 @@ int ast_bo_create(struct drm_device *dev, int size, int align,
 	ret = ttm_bo_init(&ast->ttm.bdev, &astbo->bo, size,
 			  ttm_bo_type_device, &astbo->placement,
 			  align >> PAGE_SHIFT, false, NULL, acc_size,
-			  NULL, ast_bo_ttm_destroy);
+			  NULL, NULL, ast_bo_ttm_destroy);
 	if (ret)
 		return ret;
 
@@ -360,7 +364,7 @@ int ast_bo_pin(struct ast_bo *bo, u32 pl_flag, u64 *gpu_addr)
 
 	ast_ttm_placement(bo, pl_flag);
 	for (i = 0; i < bo->placement.num_placement; i++)
-		bo->placements[i] |= TTM_PL_FLAG_NO_EVICT;
+		bo->placements[i].flags |= TTM_PL_FLAG_NO_EVICT;
 	ret = ttm_bo_validate(&bo->bo, &bo->placement, false, false);
 	if (ret)
 		return ret;
@@ -383,7 +387,7 @@ int ast_bo_unpin(struct ast_bo *bo)
 		return 0;
 
 	for (i = 0; i < bo->placement.num_placement ; i++)
-		bo->placements[i] &= ~TTM_PL_FLAG_NO_EVICT;
+		bo->placements[i].flags &= ~TTM_PL_FLAG_NO_EVICT;
 	ret = ttm_bo_validate(&bo->bo, &bo->placement, false, false);
 	if (ret)
 		return ret;
@@ -407,7 +411,7 @@ int ast_bo_push_sysram(struct ast_bo *bo)
 
 	ast_ttm_placement(bo, TTM_PL_FLAG_SYSTEM);
 	for (i = 0; i < bo->placement.num_placement ; i++)
-		bo->placements[i] |= TTM_PL_FLAG_NO_EVICT;
+		bo->placements[i].flags |= TTM_PL_FLAG_NO_EVICT;
 
 	ret = ttm_bo_validate(&bo->bo, &bo->placement, false, false);
 	if (ret) {
@@ -423,7 +427,7 @@ int ast_mmap(struct file *filp, struct vm_area_struct *vma)
 	struct ast_private *ast;
 
 	if (unlikely(vma->vm_pgoff < DRM_FILE_PAGE_OFFSET))
-		return drm_mmap(filp, vma);
+		return -EINVAL;
 
 	file_priv = filp->private_data;
 	ast = file_priv->minor->dev->dev_private;
diff --git a/drivers/gpu/drm/ati_pcigart.c b/drivers/gpu/drm/ati_pcigart.c
index c399dea27a3b..6c4d4b6eba80 100644
--- a/drivers/gpu/drm/ati_pcigart.c
+++ b/drivers/gpu/drm/ati_pcigart.c
@@ -34,6 +34,8 @@
 #include <linux/export.h>
 #include <drm/drmP.h>
 
+#include <drm/ati_pcigart.h>
+
 # define ATI_PCIGART_PAGE_SIZE		4096	/**< PCI GART page size */
 
 static int drm_ati_alloc_pcigart_table(struct drm_device *dev,
diff --git a/drivers/gpu/drm/bochs/bochs.h b/drivers/gpu/drm/bochs/bochs.h
index 7eb52dd44b01..71f2687fc3cc 100644
--- a/drivers/gpu/drm/bochs/bochs.h
+++ b/drivers/gpu/drm/bochs/bochs.h
@@ -7,6 +7,8 @@
 #include <drm/drm_crtc_helper.h>
 #include <drm/drm_fb_helper.h>
 
+#include <drm/drm_gem.h>
+
 #include <ttm/ttm_bo_driver.h>
 #include <ttm/ttm_page_alloc.h>
 
@@ -99,7 +101,7 @@ struct bochs_bo {
 	struct ttm_placement placement;
 	struct ttm_bo_kmap_obj kmap;
 	struct drm_gem_object gem;
-	u32 placements[3];
+	struct ttm_place placements[3];
 	int pin_count;
 };
 
diff --git a/drivers/gpu/drm/bochs/bochs_drv.c b/drivers/gpu/drm/bochs/bochs_drv.c
index 9738e9b14708..98837bde2d25 100644
--- a/drivers/gpu/drm/bochs/bochs_drv.c
+++ b/drivers/gpu/drm/bochs/bochs_drv.c
@@ -82,6 +82,7 @@ static struct drm_driver bochs_driver = {
 	.driver_features	= DRIVER_GEM | DRIVER_MODESET,
 	.load			= bochs_load,
 	.unload			= bochs_unload,
+	.set_busid		= drm_pci_set_busid,
 	.fops			= &bochs_fops,
 	.name			= "bochs-drm",
 	.desc			= "bochs dispi vga interface (qemu stdvga)",
diff --git a/drivers/gpu/drm/bochs/bochs_fbdev.c b/drivers/gpu/drm/bochs/bochs_fbdev.c
index fe95d31cd110..61dbf09dff5d 100644
--- a/drivers/gpu/drm/bochs/bochs_fbdev.c
+++ b/drivers/gpu/drm/bochs/bochs_fbdev.c
@@ -9,6 +9,17 @@
 
 /* ---------------------------------------------------------------------- */
 
+static int bochsfb_mmap(struct fb_info *info,
+			struct vm_area_struct *vma)
+{
+	struct drm_fb_helper *fb_helper = info->par;
+	struct bochs_device *bochs =
+		container_of(fb_helper, struct bochs_device, fb.helper);
+	struct bochs_bo *bo = gem_to_bochs_bo(bochs->fb.gfb.obj);
+
+	return ttm_fbdev_mmap(vma, &bo->bo);
+}
+
 static struct fb_ops bochsfb_ops = {
 	.owner = THIS_MODULE,
 	.fb_check_var = drm_fb_helper_check_var,
@@ -19,6 +30,7 @@ static struct fb_ops bochsfb_ops = {
 	.fb_pan_display = drm_fb_helper_pan_display,
 	.fb_blank = drm_fb_helper_blank,
 	.fb_setcmap = drm_fb_helper_setcmap,
+	.fb_mmap = bochsfb_mmap,
 };
 
 static int bochsfb_create_object(struct bochs_device *bochs,
@@ -123,11 +135,9 @@ static int bochsfb_create(struct drm_fb_helper *helper,
 	info->screen_base = bo->kmap.virtual;
 	info->screen_size = size;
 
-#if 0
-	/* FIXME: get this right for mmap(/dev/fb0) */
-	info->fix.smem_start = bochs_bo_mmap_offset(bo);
+	drm_vma_offset_remove(&bo->bo.bdev->vma_manager, &bo->bo.vma_node);
+	info->fix.smem_start = 0;
 	info->fix.smem_len = size;
-#endif
 
 	ret = fb_alloc_cmap(&info->cmap, 256, 0);
 	if (ret) {
diff --git a/drivers/gpu/drm/bochs/bochs_hw.c b/drivers/gpu/drm/bochs/bochs_hw.c
index dbe619e6aab4..460389702d31 100644
--- a/drivers/gpu/drm/bochs/bochs_hw.c
+++ b/drivers/gpu/drm/bochs/bochs_hw.c
@@ -51,11 +51,10 @@ int bochs_hw_init(struct drm_device *dev, uint32_t flags)
 {
 	struct bochs_device *bochs = dev->dev_private;
 	struct pci_dev *pdev = dev->pdev;
-	unsigned long addr, size, mem, ioaddr, iosize;
+	unsigned long addr, size, mem, ioaddr, iosize, qext_size;
 	u16 id;
 
-	if (/* (ent->driver_data == BOCHS_QEMU_STDVGA) && */
-	    (pdev->resource[2].flags & IORESOURCE_MEM)) {
+	if (pdev->resource[2].flags & IORESOURCE_MEM) {
 		/* mmio bar with vga and bochs registers present */
 		if (pci_request_region(pdev, 2, "bochs-drm") != 0) {
 			DRM_ERROR("Cannot request mmio region\n");
@@ -116,6 +115,24 @@ int bochs_hw_init(struct drm_device *dev, uint32_t flags)
 		 size / 1024, addr,
 		 bochs->ioports ? "ioports" : "mmio",
 		 ioaddr);
+
+	if (bochs->mmio && pdev->revision >= 2) {
+		qext_size = readl(bochs->mmio + 0x600);
+		if (qext_size < 4 || qext_size > iosize)
+			goto noext;
+		DRM_DEBUG("Found qemu ext regs, size %ld\n", qext_size);
+		if (qext_size >= 8) {
+#ifdef __BIG_ENDIAN
+			writel(0xbebebebe, bochs->mmio + 0x604);
+#else
+			writel(0x1e1e1e1e, bochs->mmio + 0x604);
+#endif
+			DRM_DEBUG("  qext endian: 0x%x\n",
+				  readl(bochs->mmio + 0x604));
+		}
+	}
+
+noext:
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/bochs/bochs_kms.c b/drivers/gpu/drm/bochs/bochs_kms.c
index 6b7efcf363d6..85f0f8cf1fb8 100644
--- a/drivers/gpu/drm/bochs/bochs_kms.c
+++ b/drivers/gpu/drm/bochs/bochs_kms.c
@@ -6,6 +6,7 @@
  */
 
 #include "bochs.h"
+#include <drm/drm_plane_helper.h>
 
 static int defx = 1024;
 static int defy = 768;
@@ -108,11 +109,32 @@ static void bochs_crtc_gamma_set(struct drm_crtc *crtc, u16 *red, u16 *green,
 {
 }
 
+static int bochs_crtc_page_flip(struct drm_crtc *crtc,
+				struct drm_framebuffer *fb,
+				struct drm_pending_vblank_event *event,
+				uint32_t page_flip_flags)
+{
+	struct bochs_device *bochs =
+		container_of(crtc, struct bochs_device, crtc);
+	struct drm_framebuffer *old_fb = crtc->primary->fb;
+	unsigned long irqflags;
+
+	crtc->primary->fb = fb;
+	bochs_crtc_mode_set_base(crtc, 0, 0, old_fb);
+	if (event) {
+		spin_lock_irqsave(&bochs->dev->event_lock, irqflags);
+		drm_send_vblank_event(bochs->dev, -1, event);
+		spin_unlock_irqrestore(&bochs->dev->event_lock, irqflags);
+	}
+	return 0;
+}
+
 /* These provide the minimum set of functions required to handle a CRTC */
 static const struct drm_crtc_funcs bochs_crtc_funcs = {
 	.gamma_set = bochs_crtc_gamma_set,
 	.set_config = drm_crtc_helper_set_config,
 	.destroy = drm_crtc_cleanup,
+	.page_flip = bochs_crtc_page_flip,
 };
 
 static const struct drm_crtc_helper_funcs bochs_helper_funcs = {
diff --git a/drivers/gpu/drm/bochs/bochs_mm.c b/drivers/gpu/drm/bochs/bochs_mm.c
index 1728a1b0b813..66286ff518d4 100644
--- a/drivers/gpu/drm/bochs/bochs_mm.c
+++ b/drivers/gpu/drm/bochs/bochs_mm.c
@@ -257,20 +257,26 @@ void bochs_mm_fini(struct bochs_device *bochs)
 
 static void bochs_ttm_placement(struct bochs_bo *bo, int domain)
 {
+	unsigned i;
 	u32 c = 0;
-	bo->placement.fpfn = 0;
-	bo->placement.lpfn = 0;
 	bo->placement.placement = bo->placements;
 	bo->placement.busy_placement = bo->placements;
 	if (domain & TTM_PL_FLAG_VRAM) {
-		bo->placements[c++] = TTM_PL_FLAG_WC | TTM_PL_FLAG_UNCACHED
+		bo->placements[c++].flags = TTM_PL_FLAG_WC
+			| TTM_PL_FLAG_UNCACHED
 			| TTM_PL_FLAG_VRAM;
 	}
 	if (domain & TTM_PL_FLAG_SYSTEM) {
-		bo->placements[c++] = TTM_PL_MASK_CACHING | TTM_PL_FLAG_SYSTEM;
+		bo->placements[c++].flags = TTM_PL_MASK_CACHING
+			| TTM_PL_FLAG_SYSTEM;
 	}
 	if (!c) {
-		bo->placements[c++] = TTM_PL_MASK_CACHING | TTM_PL_FLAG_SYSTEM;
+		bo->placements[c++].flags = TTM_PL_MASK_CACHING
+			| TTM_PL_FLAG_SYSTEM;
+	}
+	for (i = 0; i < c; ++i) {
+		bo->placements[i].fpfn = 0;
+		bo->placements[i].lpfn = 0;
 	}
 	bo->placement.num_placement = c;
 	bo->placement.num_busy_placement = c;
@@ -294,7 +300,7 @@ int bochs_bo_pin(struct bochs_bo *bo, u32 pl_flag, u64 *gpu_addr)
 
 	bochs_ttm_placement(bo, pl_flag);
 	for (i = 0; i < bo->placement.num_placement; i++)
-		bo->placements[i] |= TTM_PL_FLAG_NO_EVICT;
+		bo->placements[i].flags |= TTM_PL_FLAG_NO_EVICT;
 	ret = ttm_bo_validate(&bo->bo, &bo->placement, false, false);
 	if (ret)
 		return ret;
@@ -319,7 +325,7 @@ int bochs_bo_unpin(struct bochs_bo *bo)
 		return 0;
 
 	for (i = 0; i < bo->placement.num_placement; i++)
-		bo->placements[i] &= ~TTM_PL_FLAG_NO_EVICT;
+		bo->placements[i].flags &= ~TTM_PL_FLAG_NO_EVICT;
 	ret = ttm_bo_validate(&bo->bo, &bo->placement, false, false);
 	if (ret)
 		return ret;
@@ -333,7 +339,7 @@ int bochs_mmap(struct file *filp, struct vm_area_struct *vma)
 	struct bochs_device *bochs;
 
 	if (unlikely(vma->vm_pgoff < DRM_FILE_PAGE_OFFSET))
-		return drm_mmap(filp, vma);
+		return -EINVAL;
 
 	file_priv = filp->private_data;
 	bochs = file_priv->minor->dev->dev_private;
@@ -371,7 +377,7 @@ static int bochs_bo_create(struct drm_device *dev, int size, int align,
 	ret = ttm_bo_init(&bochs->ttm.bdev, &bochsbo->bo, size,
 			  ttm_bo_type_device, &bochsbo->placement,
 			  align >> PAGE_SHIFT, false, NULL, acc_size,
-			  NULL, bochs_bo_ttm_destroy);
+			  NULL, NULL, bochs_bo_ttm_destroy);
 	if (ret)
 		return ret;
 
diff --git a/drivers/gpu/drm/cirrus/cirrus_drv.c b/drivers/gpu/drm/cirrus/cirrus_drv.c
index 919c73b94447..c2a1cba1e984 100644
--- a/drivers/gpu/drm/cirrus/cirrus_drv.c
+++ b/drivers/gpu/drm/cirrus/cirrus_drv.c
@@ -32,6 +32,8 @@ static struct drm_driver driver;
 static const struct pci_device_id pciidlist[] = {
 	{ PCI_VENDOR_ID_CIRRUS, PCI_DEVICE_ID_CIRRUS_5446, 0x1af4, 0x1100, 0,
 	  0, 0 },
+	{ PCI_VENDOR_ID_CIRRUS, PCI_DEVICE_ID_CIRRUS_5446, PCI_VENDOR_ID_XEN,
+	  0x0001, 0, 0, 0 },
 	{0,}
 };
 
@@ -128,6 +130,7 @@ static struct drm_driver driver = {
 	.driver_features = DRIVER_MODESET | DRIVER_GEM,
 	.load = cirrus_driver_load,
 	.unload = cirrus_driver_unload,
+	.set_busid = drm_pci_set_busid,
 	.fops = &cirrus_driver_fops,
 	.name = DRIVER_NAME,
 	.desc = DRIVER_DESC,
diff --git a/drivers/gpu/drm/cirrus/cirrus_drv.h b/drivers/gpu/drm/cirrus/cirrus_drv.h
index 401c890b6c6a..693a4565c4ff 100644
--- a/drivers/gpu/drm/cirrus/cirrus_drv.h
+++ b/drivers/gpu/drm/cirrus/cirrus_drv.h
@@ -21,6 +21,8 @@
 #include <drm/ttm/ttm_memory.h>
 #include <drm/ttm/ttm_module.h>
 
+#include <drm/drm_gem.h>
+
 #define DRIVER_AUTHOR		"Matthew Garrett"
 
 #define DRIVER_NAME		"cirrus"
@@ -163,7 +165,7 @@ struct cirrus_bo {
 	struct ttm_placement placement;
 	struct ttm_bo_kmap_obj kmap;
 	struct drm_gem_object gem;
-	u32 placements[3];
+	struct ttm_place placements[3];
 	int pin_count;
 };
 #define gem_to_cirrus_bo(gobj) container_of((gobj), struct cirrus_bo, gem)
@@ -208,6 +210,9 @@ int cirrus_framebuffer_init(struct drm_device *dev,
 			    struct drm_mode_fb_cmd2 *mode_cmd,
 			    struct drm_gem_object *obj);
 
+bool cirrus_check_framebuffer(struct cirrus_device *cdev, int width, int height,
+			      int bpp, int pitch);
+
 				/* cirrus_display.c */
 int cirrus_modeset_init(struct cirrus_device *cdev);
 void cirrus_modeset_fini(struct cirrus_device *cdev);
diff --git a/drivers/gpu/drm/cirrus/cirrus_fbdev.c b/drivers/gpu/drm/cirrus/cirrus_fbdev.c
index 2a135f253e29..502a89eb54b5 100644
--- a/drivers/gpu/drm/cirrus/cirrus_fbdev.c
+++ b/drivers/gpu/drm/cirrus/cirrus_fbdev.c
@@ -139,6 +139,7 @@ static int cirrusfb_create_object(struct cirrus_fbdev *afbdev,
 			       struct drm_gem_object **gobj_p)
 {
 	struct drm_device *dev = afbdev->helper.dev;
+	struct cirrus_device *cdev = dev->dev_private;
 	u32 bpp, depth;
 	u32 size;
 	struct drm_gem_object *gobj;
@@ -146,8 +147,10 @@ static int cirrusfb_create_object(struct cirrus_fbdev *afbdev,
 	int ret = 0;
 	drm_fb_get_bpp_depth(mode_cmd->pixel_format, &depth, &bpp);
 
-	if (bpp > 24)
+	if (!cirrus_check_framebuffer(cdev, mode_cmd->width, mode_cmd->height,
+				      bpp, mode_cmd->pitches[0]))
 		return -EINVAL;
+
 	size = mode_cmd->pitches[0] * mode_cmd->height;
 	ret = cirrus_gem_create(dev, size, true, &gobj);
 	if (ret)
@@ -160,7 +163,8 @@ static int cirrusfb_create_object(struct cirrus_fbdev *afbdev,
 static int cirrusfb_create(struct drm_fb_helper *helper,
 			   struct drm_fb_helper_surface_size *sizes)
 {
-	struct cirrus_fbdev *gfbdev = (struct cirrus_fbdev *)helper;
+	struct cirrus_fbdev *gfbdev =
+		container_of(helper, struct cirrus_fbdev, helper);
 	struct drm_device *dev = gfbdev->helper.dev;
 	struct cirrus_device *cdev = gfbdev->helper.dev->dev_private;
 	struct fb_info *info;
diff --git a/drivers/gpu/drm/cirrus/cirrus_main.c b/drivers/gpu/drm/cirrus/cirrus_main.c
index 99c1983f99d2..4c2d68e9102d 100644
--- a/drivers/gpu/drm/cirrus/cirrus_main.c
+++ b/drivers/gpu/drm/cirrus/cirrus_main.c
@@ -49,14 +49,16 @@ cirrus_user_framebuffer_create(struct drm_device *dev,
 			       struct drm_file *filp,
 			       struct drm_mode_fb_cmd2 *mode_cmd)
 {
+	struct cirrus_device *cdev = dev->dev_private;
 	struct drm_gem_object *obj;
 	struct cirrus_framebuffer *cirrus_fb;
 	int ret;
 	u32 bpp, depth;
 
 	drm_fb_get_bpp_depth(mode_cmd->pixel_format, &depth, &bpp);
-	/* cirrus can't handle > 24bpp framebuffers at all */
-	if (bpp > 24)
+
+	if (!cirrus_check_framebuffer(cdev, mode_cmd->width, mode_cmd->height,
+				      bpp, mode_cmd->pitches[0]))
 		return ERR_PTR(-EINVAL);
 
 	obj = drm_gem_object_lookup(dev, filp, mode_cmd->handles[0]);
@@ -96,8 +98,7 @@ static int cirrus_vram_init(struct cirrus_device *cdev)
 {
 	/* BAR 0 is VRAM */
 	cdev->mc.vram_base = pci_resource_start(cdev->dev->pdev, 0);
-	/* We have 4MB of VRAM */
-	cdev->mc.vram_size = 4 * 1024 * 1024;
+	cdev->mc.vram_size = pci_resource_len(cdev->dev->pdev, 0);
 
 	if (!request_mem_region(cdev->mc.vram_base, cdev->mc.vram_size,
 				"cirrusdrmfb_vram")) {
@@ -179,17 +180,22 @@ int cirrus_driver_load(struct drm_device *dev, unsigned long flags)
 	}
 
 	r = cirrus_mm_init(cdev);
-	if (r)
+	if (r) {
 		dev_err(&dev->pdev->dev, "fatal err on mm init\n");
+		goto out;
+	}
 
 	r = cirrus_modeset_init(cdev);
-	if (r)
+	if (r) {
 		dev_err(&dev->pdev->dev, "Fatal error during modeset init: %d\n", r);
+		goto out;
+	}
 
 	dev->mode_config.funcs = (void *)&cirrus_mode_funcs;
+
+	return 0;
 out:
-	if (r)
-		cirrus_driver_unload(dev);
+	cirrus_driver_unload(dev);
 	return r;
 }
 
@@ -307,3 +313,21 @@ out_unlock:
 	return ret;
 
 }
+
+bool cirrus_check_framebuffer(struct cirrus_device *cdev, int width, int height,
+			      int bpp, int pitch)
+{
+	const int max_pitch = 0x1FF << 3; /* (4096 - 1) & ~111b bytes */
+	const int max_size = cdev->mc.vram_size;
+
+	if (bpp > 32)
+		return false;
+
+	if (pitch > max_pitch)
+		return false;
+
+	if (pitch * height > max_size)
+		return false;
+
+	return true;
+}
diff --git a/drivers/gpu/drm/cirrus/cirrus_mode.c b/drivers/gpu/drm/cirrus/cirrus_mode.c
index c7c5a9d91fa0..99d4a74ffeaf 100644
--- a/drivers/gpu/drm/cirrus/cirrus_mode.c
+++ b/drivers/gpu/drm/cirrus/cirrus_mode.c
@@ -16,6 +16,7 @@
  */
 #include <drm/drmP.h>
 #include <drm/drm_crtc_helper.h>
+#include <drm/drm_plane_helper.h>
 
 #include <video/cirrus.h>
 
diff --git a/drivers/gpu/drm/cirrus/cirrus_ttm.c b/drivers/gpu/drm/cirrus/cirrus_ttm.c
index 92e6b7786097..dfffd528517a 100644
--- a/drivers/gpu/drm/cirrus/cirrus_ttm.c
+++ b/drivers/gpu/drm/cirrus/cirrus_ttm.c
@@ -298,18 +298,21 @@ void cirrus_mm_fini(struct cirrus_device *cirrus)
 void cirrus_ttm_placement(struct cirrus_bo *bo, int domain)
 {
 	u32 c = 0;
-	bo->placement.fpfn = 0;
-	bo->placement.lpfn = 0;
+	unsigned i;
 	bo->placement.placement = bo->placements;
 	bo->placement.busy_placement = bo->placements;
 	if (domain & TTM_PL_FLAG_VRAM)
-		bo->placements[c++] = TTM_PL_FLAG_WC | TTM_PL_FLAG_UNCACHED | TTM_PL_FLAG_VRAM;
+		bo->placements[c++].flags = TTM_PL_FLAG_WC | TTM_PL_FLAG_UNCACHED | TTM_PL_FLAG_VRAM;
 	if (domain & TTM_PL_FLAG_SYSTEM)
-		bo->placements[c++] = TTM_PL_MASK_CACHING | TTM_PL_FLAG_SYSTEM;
+		bo->placements[c++].flags = TTM_PL_MASK_CACHING | TTM_PL_FLAG_SYSTEM;
 	if (!c)
-		bo->placements[c++] = TTM_PL_MASK_CACHING | TTM_PL_FLAG_SYSTEM;
+		bo->placements[c++].flags = TTM_PL_MASK_CACHING | TTM_PL_FLAG_SYSTEM;
 	bo->placement.num_placement = c;
 	bo->placement.num_busy_placement = c;
+	for (i = 0; i < c; ++i) {
+		bo->placements[i].fpfn = 0;
+		bo->placements[i].lpfn = 0;
+	}
 }
 
 int cirrus_bo_create(struct drm_device *dev, int size, int align,
@@ -340,7 +343,7 @@ int cirrus_bo_create(struct drm_device *dev, int size, int align,
 	ret = ttm_bo_init(&cirrus->ttm.bdev, &cirrusbo->bo, size,
 			  ttm_bo_type_device, &cirrusbo->placement,
 			  align >> PAGE_SHIFT, false, NULL, acc_size,
-			  NULL, cirrus_bo_ttm_destroy);
+			  NULL, NULL, cirrus_bo_ttm_destroy);
 	if (ret)
 		return ret;
 
@@ -365,7 +368,7 @@ int cirrus_bo_pin(struct cirrus_bo *bo, u32 pl_flag, u64 *gpu_addr)
 
 	cirrus_ttm_placement(bo, pl_flag);
 	for (i = 0; i < bo->placement.num_placement; i++)
-		bo->placements[i] |= TTM_PL_FLAG_NO_EVICT;
+		bo->placements[i].flags |= TTM_PL_FLAG_NO_EVICT;
 	ret = ttm_bo_validate(&bo->bo, &bo->placement, false, false);
 	if (ret)
 		return ret;
@@ -392,7 +395,7 @@ int cirrus_bo_push_sysram(struct cirrus_bo *bo)
 
 	cirrus_ttm_placement(bo, TTM_PL_FLAG_SYSTEM);
 	for (i = 0; i < bo->placement.num_placement ; i++)
-		bo->placements[i] |= TTM_PL_FLAG_NO_EVICT;
+		bo->placements[i].flags |= TTM_PL_FLAG_NO_EVICT;
 
 	ret = ttm_bo_validate(&bo->bo, &bo->placement, false, false);
 	if (ret) {
@@ -408,7 +411,7 @@ int cirrus_mmap(struct file *filp, struct vm_area_struct *vma)
 	struct cirrus_device *cirrus;
 
 	if (unlikely(vma->vm_pgoff < DRM_FILE_PAGE_OFFSET))
-		return drm_mmap(filp, vma);
+		return -EINVAL;
 
 	file_priv = filp->private_data;
 	cirrus = file_priv->minor->dev->dev_private;
diff --git a/drivers/gpu/drm/drm_agpsupport.c b/drivers/gpu/drm/drm_agpsupport.c
index dde205cef384..4b2b4aa5033b 100644
--- a/drivers/gpu/drm/drm_agpsupport.c
+++ b/drivers/gpu/drm/drm_agpsupport.c
@@ -34,6 +34,7 @@
 #include <drm/drmP.h>
 #include <linux/module.h>
 #include <linux/slab.h>
+#include "drm_legacy.h"
 
 #if __OS_HAS_AGP
 
diff --git a/drivers/gpu/drm/drm_atomic.c b/drivers/gpu/drm/drm_atomic.c
new file mode 100644
index 000000000000..ff5f034cc405
--- /dev/null
+++ b/drivers/gpu/drm/drm_atomic.c
@@ -0,0 +1,657 @@
+/*
+ * Copyright (C) 2014 Red Hat
+ * Copyright (C) 2014 Intel Corp.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors:
+ * Rob Clark <robdclark@gmail.com>
+ * Daniel Vetter <daniel.vetter@ffwll.ch>
+ */
+
+
+#include <drm/drmP.h>
+#include <drm/drm_atomic.h>
+#include <drm/drm_plane_helper.h>
+
+static void kfree_state(struct drm_atomic_state *state)
+{
+	kfree(state->connectors);
+	kfree(state->connector_states);
+	kfree(state->crtcs);
+	kfree(state->crtc_states);
+	kfree(state->planes);
+	kfree(state->plane_states);
+	kfree(state);
+}
+
+/**
+ * drm_atomic_state_alloc - allocate atomic state
+ * @dev: DRM device
+ *
+ * This allocates an empty atomic state to track updates.
+ */
+struct drm_atomic_state *
+drm_atomic_state_alloc(struct drm_device *dev)
+{
+	struct drm_atomic_state *state;
+
+	state = kzalloc(sizeof(*state), GFP_KERNEL);
+	if (!state)
+		return NULL;
+
+	state->num_connector = ACCESS_ONCE(dev->mode_config.num_connector);
+
+	state->crtcs = kcalloc(dev->mode_config.num_crtc,
+			       sizeof(*state->crtcs), GFP_KERNEL);
+	if (!state->crtcs)
+		goto fail;
+	state->crtc_states = kcalloc(dev->mode_config.num_crtc,
+				     sizeof(*state->crtc_states), GFP_KERNEL);
+	if (!state->crtc_states)
+		goto fail;
+	state->planes = kcalloc(dev->mode_config.num_total_plane,
+				sizeof(*state->planes), GFP_KERNEL);
+	if (!state->planes)
+		goto fail;
+	state->plane_states = kcalloc(dev->mode_config.num_total_plane,
+				      sizeof(*state->plane_states), GFP_KERNEL);
+	if (!state->plane_states)
+		goto fail;
+	state->connectors = kcalloc(state->num_connector,
+				    sizeof(*state->connectors),
+				    GFP_KERNEL);
+	if (!state->connectors)
+		goto fail;
+	state->connector_states = kcalloc(state->num_connector,
+					  sizeof(*state->connector_states),
+					  GFP_KERNEL);
+	if (!state->connector_states)
+		goto fail;
+
+	state->dev = dev;
+
+	DRM_DEBUG_KMS("Allocate atomic state %p\n", state);
+
+	return state;
+fail:
+	kfree_state(state);
+
+	return NULL;
+}
+EXPORT_SYMBOL(drm_atomic_state_alloc);
+
+/**
+ * drm_atomic_state_clear - clear state object
+ * @state: atomic state
+ *
+ * When the w/w mutex algorithm detects a deadlock we need to back off and drop
+ * all locks. So someone else could sneak in and change the current modeset
+ * configuration. Which means that all the state assembled in @state is no
+ * longer an atomic update to the current state, but to some arbitrary earlier
+ * state. Which could break assumptions the driver's ->atomic_check likely
+ * relies on.
+ *
+ * Hence we must clear all cached state and completely start over, using this
+ * function.
+ */
+void drm_atomic_state_clear(struct drm_atomic_state *state)
+{
+	struct drm_device *dev = state->dev;
+	struct drm_mode_config *config = &dev->mode_config;
+	int i;
+
+	DRM_DEBUG_KMS("Clearing atomic state %p\n", state);
+
+	for (i = 0; i < state->num_connector; i++) {
+		struct drm_connector *connector = state->connectors[i];
+
+		if (!connector)
+			continue;
+
+		WARN_ON(!drm_modeset_is_locked(&config->connection_mutex));
+
+		connector->funcs->atomic_destroy_state(connector,
+						       state->connector_states[i]);
+	}
+
+	for (i = 0; i < config->num_crtc; i++) {
+		struct drm_crtc *crtc = state->crtcs[i];
+
+		if (!crtc)
+			continue;
+
+		crtc->funcs->atomic_destroy_state(crtc,
+						  state->crtc_states[i]);
+	}
+
+	for (i = 0; i < config->num_total_plane; i++) {
+		struct drm_plane *plane = state->planes[i];
+
+		if (!plane)
+			continue;
+
+		plane->funcs->atomic_destroy_state(plane,
+						   state->plane_states[i]);
+	}
+}
+EXPORT_SYMBOL(drm_atomic_state_clear);
+
+/**
+ * drm_atomic_state_free - free all memory for an atomic state
+ * @state: atomic state to deallocate
+ *
+ * This frees all memory associated with an atomic state, including all the
+ * per-object state for planes, crtcs and connectors.
+ */
+void drm_atomic_state_free(struct drm_atomic_state *state)
+{
+	drm_atomic_state_clear(state);
+
+	DRM_DEBUG_KMS("Freeing atomic state %p\n", state);
+
+	kfree_state(state);
+}
+EXPORT_SYMBOL(drm_atomic_state_free);
+
+/**
+ * drm_atomic_get_crtc_state - get crtc state
+ * @state: global atomic state object
+ * @crtc: crtc to get state object for
+ *
+ * This function returns the crtc state for the given crtc, allocating it if
+ * needed. It will also grab the relevant crtc lock to make sure that the state
+ * is consistent.
+ *
+ * Returns:
+ *
+ * Either the allocated state or the error code encoded into the pointer. When
+ * the error is EDEADLK then the w/w mutex code has detected a deadlock and the
+ * entire atomic sequence must be restarted. All other errors are fatal.
+ */
+struct drm_crtc_state *
+drm_atomic_get_crtc_state(struct drm_atomic_state *state,
+			  struct drm_crtc *crtc)
+{
+	int ret, index;
+	struct drm_crtc_state *crtc_state;
+
+	index = drm_crtc_index(crtc);
+
+	if (state->crtc_states[index])
+		return state->crtc_states[index];
+
+	ret = drm_modeset_lock(&crtc->mutex, state->acquire_ctx);
+	if (ret)
+		return ERR_PTR(ret);
+
+	crtc_state = crtc->funcs->atomic_duplicate_state(crtc);
+	if (!crtc_state)
+		return ERR_PTR(-ENOMEM);
+
+	state->crtc_states[index] = crtc_state;
+	state->crtcs[index] = crtc;
+	crtc_state->state = state;
+
+	DRM_DEBUG_KMS("Added [CRTC:%d] %p state to %p\n",
+		      crtc->base.id, crtc_state, state);
+
+	return crtc_state;
+}
+EXPORT_SYMBOL(drm_atomic_get_crtc_state);
+
+/**
+ * drm_atomic_get_plane_state - get plane state
+ * @state: global atomic state object
+ * @plane: plane to get state object for
+ *
+ * This function returns the plane state for the given plane, allocating it if
+ * needed. It will also grab the relevant plane lock to make sure that the state
+ * is consistent.
+ *
+ * Returns:
+ *
+ * Either the allocated state or the error code encoded into the pointer. When
+ * the error is EDEADLK then the w/w mutex code has detected a deadlock and the
+ * entire atomic sequence must be restarted. All other errors are fatal.
+ */
+struct drm_plane_state *
+drm_atomic_get_plane_state(struct drm_atomic_state *state,
+			  struct drm_plane *plane)
+{
+	int ret, index;
+	struct drm_plane_state *plane_state;
+
+	index = drm_plane_index(plane);
+
+	if (state->plane_states[index])
+		return state->plane_states[index];
+
+	ret = drm_modeset_lock(&plane->mutex, state->acquire_ctx);
+	if (ret)
+		return ERR_PTR(ret);
+
+	plane_state = plane->funcs->atomic_duplicate_state(plane);
+	if (!plane_state)
+		return ERR_PTR(-ENOMEM);
+
+	state->plane_states[index] = plane_state;
+	state->planes[index] = plane;
+	plane_state->state = state;
+
+	DRM_DEBUG_KMS("Added [PLANE:%d] %p state to %p\n",
+		      plane->base.id, plane_state, state);
+
+	if (plane_state->crtc) {
+		struct drm_crtc_state *crtc_state;
+
+		crtc_state = drm_atomic_get_crtc_state(state,
+						       plane_state->crtc);
+		if (IS_ERR(crtc_state))
+			return ERR_CAST(crtc_state);
+	}
+
+	return plane_state;
+}
+EXPORT_SYMBOL(drm_atomic_get_plane_state);
+
+/**
+ * drm_atomic_get_connector_state - get connector state
+ * @state: global atomic state object
+ * @connector: connector to get state object for
+ *
+ * This function returns the connector state for the given connector,
+ * allocating it if needed. It will also grab the relevant connector lock to
+ * make sure that the state is consistent.
+ *
+ * Returns:
+ *
+ * Either the allocated state or the error code encoded into the pointer. When
+ * the error is EDEADLK then the w/w mutex code has detected a deadlock and the
+ * entire atomic sequence must be restarted. All other errors are fatal.
+ */
+struct drm_connector_state *
+drm_atomic_get_connector_state(struct drm_atomic_state *state,
+			  struct drm_connector *connector)
+{
+	int ret, index;
+	struct drm_mode_config *config = &connector->dev->mode_config;
+	struct drm_connector_state *connector_state;
+
+	ret = drm_modeset_lock(&config->connection_mutex, state->acquire_ctx);
+	if (ret)
+		return ERR_PTR(ret);
+
+	index = drm_connector_index(connector);
+
+	/*
+	 * Construction of atomic state updates can race with a connector
+	 * hot-add which might overflow. In this case flip the table and just
+	 * restart the entire ioctl - no one is fast enough to livelock a cpu
+	 * with physical hotplug events anyway.
+	 *
+	 * Note that we only grab the indexes once we have the right lock to
+	 * prevent hotplug/unplugging of connectors. So removal is no problem,
+	 * at most the array is a bit too large.
+	 */
+	if (index >= state->num_connector) {
+		DRM_DEBUG_KMS("Hot-added connector would overflow state array, restarting\n");
+		return ERR_PTR(-EAGAIN);
+	}
+
+	if (state->connector_states[index])
+		return state->connector_states[index];
+
+	connector_state = connector->funcs->atomic_duplicate_state(connector);
+	if (!connector_state)
+		return ERR_PTR(-ENOMEM);
+
+	state->connector_states[index] = connector_state;
+	state->connectors[index] = connector;
+	connector_state->state = state;
+
+	DRM_DEBUG_KMS("Added [CONNECTOR:%d] %p state to %p\n",
+		      connector->base.id, connector_state, state);
+
+	if (connector_state->crtc) {
+		struct drm_crtc_state *crtc_state;
+
+		crtc_state = drm_atomic_get_crtc_state(state,
+						       connector_state->crtc);
+		if (IS_ERR(crtc_state))
+			return ERR_CAST(crtc_state);
+	}
+
+	return connector_state;
+}
+EXPORT_SYMBOL(drm_atomic_get_connector_state);
+
+/**
+ * drm_atomic_set_crtc_for_plane - set crtc for plane
+ * @state: the incoming atomic state
+ * @plane: the plane whose incoming state to update
+ * @crtc: crtc to use for the plane
+ *
+ * Changing the assigned crtc for a plane requires us to grab the lock and state
+ * for the new crtc, as needed. This function takes care of all these details
+ * besides updating the pointer in the state object itself.
+ *
+ * Returns:
+ * 0 on success or can fail with -EDEADLK or -ENOMEM. When the error is EDEADLK
+ * then the w/w mutex code has detected a deadlock and the entire atomic
+ * sequence must be restarted. All other errors are fatal.
+ */
+int
+drm_atomic_set_crtc_for_plane(struct drm_atomic_state *state,
+			      struct drm_plane *plane, struct drm_crtc *crtc)
+{
+	struct drm_plane_state *plane_state =
+			drm_atomic_get_plane_state(state, plane);
+	struct drm_crtc_state *crtc_state;
+
+	if (WARN_ON(IS_ERR(plane_state)))
+		return PTR_ERR(plane_state);
+
+	if (plane_state->crtc) {
+		crtc_state = drm_atomic_get_crtc_state(plane_state->state,
+						       plane_state->crtc);
+		if (WARN_ON(IS_ERR(crtc_state)))
+			return PTR_ERR(crtc_state);
+
+		crtc_state->plane_mask &= ~(1 << drm_plane_index(plane));
+	}
+
+	plane_state->crtc = crtc;
+
+	if (crtc) {
+		crtc_state = drm_atomic_get_crtc_state(plane_state->state,
+						       crtc);
+		if (IS_ERR(crtc_state))
+			return PTR_ERR(crtc_state);
+		crtc_state->plane_mask |= (1 << drm_plane_index(plane));
+	}
+
+	if (crtc)
+		DRM_DEBUG_KMS("Link plane state %p to [CRTC:%d]\n",
+			      plane_state, crtc->base.id);
+	else
+		DRM_DEBUG_KMS("Link plane state %p to [NOCRTC]\n", plane_state);
+
+	return 0;
+}
+EXPORT_SYMBOL(drm_atomic_set_crtc_for_plane);
+
+/**
+ * drm_atomic_set_fb_for_plane - set crtc for plane
+ * @plane_state: atomic state object for the plane
+ * @fb: fb to use for the plane
+ *
+ * Changing the assigned framebuffer for a plane requires us to grab a reference
+ * to the new fb and drop the reference to the old fb, if there is one. This
+ * function takes care of all these details besides updating the pointer in the
+ * state object itself.
+ */
+void
+drm_atomic_set_fb_for_plane(struct drm_plane_state *plane_state,
+			    struct drm_framebuffer *fb)
+{
+	if (plane_state->fb)
+		drm_framebuffer_unreference(plane_state->fb);
+	if (fb)
+		drm_framebuffer_reference(fb);
+	plane_state->fb = fb;
+
+	if (fb)
+		DRM_DEBUG_KMS("Set [FB:%d] for plane state %p\n",
+			      fb->base.id, plane_state);
+	else
+		DRM_DEBUG_KMS("Set [NOFB] for plane state %p\n", plane_state);
+}
+EXPORT_SYMBOL(drm_atomic_set_fb_for_plane);
+
+/**
+ * drm_atomic_set_crtc_for_connector - set crtc for connector
+ * @conn_state: atomic state object for the connector
+ * @crtc: crtc to use for the connector
+ *
+ * Changing the assigned crtc for a connector requires us to grab the lock and
+ * state for the new crtc, as needed. This function takes care of all these
+ * details besides updating the pointer in the state object itself.
+ *
+ * Returns:
+ * 0 on success or can fail with -EDEADLK or -ENOMEM. When the error is EDEADLK
+ * then the w/w mutex code has detected a deadlock and the entire atomic
+ * sequence must be restarted. All other errors are fatal.
+ */
+int
+drm_atomic_set_crtc_for_connector(struct drm_connector_state *conn_state,
+				  struct drm_crtc *crtc)
+{
+	struct drm_crtc_state *crtc_state;
+
+	if (crtc) {
+		crtc_state = drm_atomic_get_crtc_state(conn_state->state, crtc);
+		if (IS_ERR(crtc_state))
+			return PTR_ERR(crtc_state);
+	}
+
+	conn_state->crtc = crtc;
+
+	if (crtc)
+		DRM_DEBUG_KMS("Link connector state %p to [CRTC:%d]\n",
+			      conn_state, crtc->base.id);
+	else
+		DRM_DEBUG_KMS("Link connector state %p to [NOCRTC]\n",
+			      conn_state);
+
+	return 0;
+}
+EXPORT_SYMBOL(drm_atomic_set_crtc_for_connector);
+
+/**
+ * drm_atomic_add_affected_connectors - add connectors for crtc
+ * @state: atomic state
+ * @crtc: DRM crtc
+ *
+ * This function walks the current configuration and adds all connectors
+ * currently using @crtc to the atomic configuration @state. Note that this
+ * function must acquire the connection mutex. This can potentially cause
+ * unneeded seralization if the update is just for the planes on one crtc. Hence
+ * drivers and helpers should only call this when really needed (e.g. when a
+ * full modeset needs to happen due to some change).
+ *
+ * Returns:
+ * 0 on success or can fail with -EDEADLK or -ENOMEM. When the error is EDEADLK
+ * then the w/w mutex code has detected a deadlock and the entire atomic
+ * sequence must be restarted. All other errors are fatal.
+ */
+int
+drm_atomic_add_affected_connectors(struct drm_atomic_state *state,
+				   struct drm_crtc *crtc)
+{
+	struct drm_mode_config *config = &state->dev->mode_config;
+	struct drm_connector *connector;
+	struct drm_connector_state *conn_state;
+	int ret;
+
+	ret = drm_modeset_lock(&config->connection_mutex, state->acquire_ctx);
+	if (ret)
+		return ret;
+
+	DRM_DEBUG_KMS("Adding all current connectors for [CRTC:%d] to %p\n",
+		      crtc->base.id, state);
+
+	/*
+	 * Changed connectors are already in @state, so only need to look at the
+	 * current configuration.
+	 */
+	list_for_each_entry(connector, &config->connector_list, head) {
+		if (connector->state->crtc != crtc)
+			continue;
+
+		conn_state = drm_atomic_get_connector_state(state, connector);
+		if (IS_ERR(conn_state))
+			return PTR_ERR(conn_state);
+	}
+
+	return 0;
+}
+EXPORT_SYMBOL(drm_atomic_add_affected_connectors);
+
+/**
+ * drm_atomic_connectors_for_crtc - count number of connected outputs
+ * @state: atomic state
+ * @crtc: DRM crtc
+ *
+ * This function counts all connectors which will be connected to @crtc
+ * according to @state. Useful to recompute the enable state for @crtc.
+ */
+int
+drm_atomic_connectors_for_crtc(struct drm_atomic_state *state,
+			       struct drm_crtc *crtc)
+{
+	int i, num_connected_connectors = 0;
+
+	for (i = 0; i < state->num_connector; i++) {
+		struct drm_connector_state *conn_state;
+
+		conn_state = state->connector_states[i];
+
+		if (conn_state && conn_state->crtc == crtc)
+			num_connected_connectors++;
+	}
+
+	DRM_DEBUG_KMS("State %p has %i connectors for [CRTC:%d]\n",
+		      state, num_connected_connectors, crtc->base.id);
+
+	return num_connected_connectors;
+}
+EXPORT_SYMBOL(drm_atomic_connectors_for_crtc);
+
+/**
+ * drm_atomic_legacy_backoff - locking backoff for legacy ioctls
+ * @state: atomic state
+ *
+ * This function should be used by legacy entry points which don't understand
+ * -EDEADLK semantics. For simplicity this one will grab all modeset locks after
+ *  the slowpath completed.
+ */
+void drm_atomic_legacy_backoff(struct drm_atomic_state *state)
+{
+	int ret;
+
+retry:
+	drm_modeset_backoff(state->acquire_ctx);
+
+	ret = drm_modeset_lock(&state->dev->mode_config.connection_mutex,
+			       state->acquire_ctx);
+	if (ret)
+		goto retry;
+	ret = drm_modeset_lock_all_crtcs(state->dev,
+					 state->acquire_ctx);
+	if (ret)
+		goto retry;
+}
+EXPORT_SYMBOL(drm_atomic_legacy_backoff);
+
+/**
+ * drm_atomic_check_only - check whether a given config would work
+ * @state: atomic configuration to check
+ *
+ * Note that this function can return -EDEADLK if the driver needed to acquire
+ * more locks but encountered a deadlock. The caller must then do the usual w/w
+ * backoff dance and restart. All other errors are fatal.
+ *
+ * Returns:
+ * 0 on success, negative error code on failure.
+ */
+int drm_atomic_check_only(struct drm_atomic_state *state)
+{
+	struct drm_mode_config *config = &state->dev->mode_config;
+
+	DRM_DEBUG_KMS("checking %p\n", state);
+
+	if (config->funcs->atomic_check)
+		return config->funcs->atomic_check(state->dev, state);
+	else
+		return 0;
+}
+EXPORT_SYMBOL(drm_atomic_check_only);
+
+/**
+ * drm_atomic_commit - commit configuration atomically
+ * @state: atomic configuration to check
+ *
+ * Note that this function can return -EDEADLK if the driver needed to acquire
+ * more locks but encountered a deadlock. The caller must then do the usual w/w
+ * backoff dance and restart. All other errors are fatal.
+ *
+ * Also note that on successful execution ownership of @state is transferred
+ * from the caller of this function to the function itself. The caller must not
+ * free or in any other way access @state. If the function fails then the caller
+ * must clean up @state itself.
+ *
+ * Returns:
+ * 0 on success, negative error code on failure.
+ */
+int drm_atomic_commit(struct drm_atomic_state *state)
+{
+	struct drm_mode_config *config = &state->dev->mode_config;
+	int ret;
+
+	ret = drm_atomic_check_only(state);
+	if (ret)
+		return ret;
+
+	DRM_DEBUG_KMS("commiting %p\n", state);
+
+	return config->funcs->atomic_commit(state->dev, state, false);
+}
+EXPORT_SYMBOL(drm_atomic_commit);
+
+/**
+ * drm_atomic_async_commit - atomic&async configuration commit
+ * @state: atomic configuration to check
+ *
+ * Note that this function can return -EDEADLK if the driver needed to acquire
+ * more locks but encountered a deadlock. The caller must then do the usual w/w
+ * backoff dance and restart. All other errors are fatal.
+ *
+ * Also note that on successful execution ownership of @state is transferred
+ * from the caller of this function to the function itself. The caller must not
+ * free or in any other way access @state. If the function fails then the caller
+ * must clean up @state itself.
+ *
+ * Returns:
+ * 0 on success, negative error code on failure.
+ */
+int drm_atomic_async_commit(struct drm_atomic_state *state)
+{
+	struct drm_mode_config *config = &state->dev->mode_config;
+	int ret;
+
+	ret = drm_atomic_check_only(state);
+	if (ret)
+		return ret;
+
+	DRM_DEBUG_KMS("commiting %p asynchronously\n", state);
+
+	return config->funcs->atomic_commit(state->dev, state, true);
+}
+EXPORT_SYMBOL(drm_atomic_async_commit);
diff --git a/drivers/gpu/drm/drm_atomic_helper.c b/drivers/gpu/drm/drm_atomic_helper.c
new file mode 100644
index 000000000000..bbdbe4721573
--- /dev/null
+++ b/drivers/gpu/drm/drm_atomic_helper.c
@@ -0,0 +1,1966 @@
+/*
+ * Copyright (C) 2014 Red Hat
+ * Copyright (C) 2014 Intel Corp.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors:
+ * Rob Clark <robdclark@gmail.com>
+ * Daniel Vetter <daniel.vetter@ffwll.ch>
+ */
+
+#include <drm/drmP.h>
+#include <drm/drm_atomic.h>
+#include <drm/drm_plane_helper.h>
+#include <drm/drm_crtc_helper.h>
+#include <drm/drm_atomic_helper.h>
+#include <linux/fence.h>
+
+/**
+ * DOC: overview
+ *
+ * This helper library provides implementations of check and commit functions on
+ * top of the CRTC modeset helper callbacks and the plane helper callbacks. It
+ * also provides convenience implementations for the atomic state handling
+ * callbacks for drivers which don't need to subclass the drm core structures to
+ * add their own additional internal state.
+ *
+ * This library also provides default implementations for the check callback in
+ * drm_atomic_helper_check and for the commit callback with
+ * drm_atomic_helper_commit. But the individual stages and callbacks are expose
+ * to allow drivers to mix and match and e.g. use the plane helpers only
+ * together with a driver private modeset implementation.
+ *
+ * This library also provides implementations for all the legacy driver
+ * interfaces on top of the atomic interface. See drm_atomic_helper_set_config,
+ * drm_atomic_helper_disable_plane, drm_atomic_helper_disable_plane and the
+ * various functions to implement set_property callbacks. New drivers must not
+ * implement these functions themselves but must use the provided helpers.
+ */
+static void
+drm_atomic_helper_plane_changed(struct drm_atomic_state *state,
+				struct drm_plane_state *plane_state,
+				struct drm_plane *plane)
+{
+	struct drm_crtc_state *crtc_state;
+
+	if (plane->state->crtc) {
+		crtc_state = state->crtc_states[drm_crtc_index(plane->state->crtc)];
+
+		if (WARN_ON(!crtc_state))
+			return;
+
+		crtc_state->planes_changed = true;
+	}
+
+	if (plane_state->crtc) {
+		crtc_state =
+			state->crtc_states[drm_crtc_index(plane_state->crtc)];
+
+		if (WARN_ON(!crtc_state))
+			return;
+
+		crtc_state->planes_changed = true;
+	}
+}
+
+static struct drm_crtc *
+get_current_crtc_for_encoder(struct drm_device *dev,
+			     struct drm_encoder *encoder)
+{
+	struct drm_mode_config *config = &dev->mode_config;
+	struct drm_connector *connector;
+
+	WARN_ON(!drm_modeset_is_locked(&config->connection_mutex));
+
+	list_for_each_entry(connector, &config->connector_list, head) {
+		if (connector->state->best_encoder != encoder)
+			continue;
+
+		return connector->state->crtc;
+	}
+
+	return NULL;
+}
+
+static int
+steal_encoder(struct drm_atomic_state *state,
+	      struct drm_encoder *encoder,
+	      struct drm_crtc *encoder_crtc)
+{
+	struct drm_mode_config *config = &state->dev->mode_config;
+	struct drm_crtc_state *crtc_state;
+	struct drm_connector *connector;
+	struct drm_connector_state *connector_state;
+	int ret;
+
+	/*
+	 * We can only steal an encoder coming from a connector, which means we
+	 * must already hold the connection_mutex.
+	 */
+	WARN_ON(!drm_modeset_is_locked(&config->connection_mutex));
+
+	DRM_DEBUG_KMS("[ENCODER:%d:%s] in use on [CRTC:%d], stealing it\n",
+		      encoder->base.id, encoder->name,
+		      encoder_crtc->base.id);
+
+	crtc_state = drm_atomic_get_crtc_state(state, encoder_crtc);
+	if (IS_ERR(crtc_state))
+		return PTR_ERR(crtc_state);
+
+	crtc_state->mode_changed = true;
+
+	list_for_each_entry(connector, &config->connector_list, head) {
+		if (connector->state->best_encoder != encoder)
+			continue;
+
+		DRM_DEBUG_KMS("Stealing encoder from [CONNECTOR:%d:%s]\n",
+			      connector->base.id,
+			      connector->name);
+
+		connector_state = drm_atomic_get_connector_state(state,
+								 connector);
+		if (IS_ERR(connector_state))
+			return PTR_ERR(connector_state);
+
+		ret = drm_atomic_set_crtc_for_connector(connector_state, NULL);
+		if (ret)
+			return ret;
+		connector_state->best_encoder = NULL;
+	}
+
+	return 0;
+}
+
+static int
+update_connector_routing(struct drm_atomic_state *state, int conn_idx)
+{
+	struct drm_connector_helper_funcs *funcs;
+	struct drm_encoder *new_encoder;
+	struct drm_crtc *encoder_crtc;
+	struct drm_connector *connector;
+	struct drm_connector_state *connector_state;
+	struct drm_crtc_state *crtc_state;
+	int idx, ret;
+
+	connector = state->connectors[conn_idx];
+	connector_state = state->connector_states[conn_idx];
+
+	if (!connector)
+		return 0;
+
+	DRM_DEBUG_KMS("Updating routing for [CONNECTOR:%d:%s]\n",
+			connector->base.id,
+			connector->name);
+
+	if (connector->state->crtc != connector_state->crtc) {
+		if (connector->state->crtc) {
+			idx = drm_crtc_index(connector->state->crtc);
+
+			crtc_state = state->crtc_states[idx];
+			crtc_state->mode_changed = true;
+		}
+
+		if (connector_state->crtc) {
+			idx = drm_crtc_index(connector_state->crtc);
+
+			crtc_state = state->crtc_states[idx];
+			crtc_state->mode_changed = true;
+		}
+	}
+
+	if (!connector_state->crtc) {
+		DRM_DEBUG_KMS("Disabling [CONNECTOR:%d:%s]\n",
+				connector->base.id,
+				connector->name);
+
+		connector_state->best_encoder = NULL;
+
+		return 0;
+	}
+
+	funcs = connector->helper_private;
+	new_encoder = funcs->best_encoder(connector);
+
+	if (!new_encoder) {
+		DRM_DEBUG_KMS("No suitable encoder found for [CONNECTOR:%d:%s]\n",
+			      connector->base.id,
+			      connector->name);
+		return -EINVAL;
+	}
+
+	if (new_encoder == connector_state->best_encoder) {
+		DRM_DEBUG_KMS("[CONNECTOR:%d:%s] keeps [ENCODER:%d:%s], now on [CRTC:%d]\n",
+			      connector->base.id,
+			      connector->name,
+			      new_encoder->base.id,
+			      new_encoder->name,
+			      connector_state->crtc->base.id);
+
+		return 0;
+	}
+
+	encoder_crtc = get_current_crtc_for_encoder(state->dev,
+						    new_encoder);
+
+	if (encoder_crtc) {
+		ret = steal_encoder(state, new_encoder, encoder_crtc);
+		if (ret) {
+			DRM_DEBUG_KMS("Encoder stealing failed for [CONNECTOR:%d:%s]\n",
+				      connector->base.id,
+				      connector->name);
+			return ret;
+		}
+	}
+
+	connector_state->best_encoder = new_encoder;
+	idx = drm_crtc_index(connector_state->crtc);
+
+	crtc_state = state->crtc_states[idx];
+	crtc_state->mode_changed = true;
+
+	DRM_DEBUG_KMS("[CONNECTOR:%d:%s] using [ENCODER:%d:%s] on [CRTC:%d]\n",
+		      connector->base.id,
+		      connector->name,
+		      new_encoder->base.id,
+		      new_encoder->name,
+		      connector_state->crtc->base.id);
+
+	return 0;
+}
+
+static int
+mode_fixup(struct drm_atomic_state *state)
+{
+	int ncrtcs = state->dev->mode_config.num_crtc;
+	struct drm_crtc_state *crtc_state;
+	struct drm_connector_state *conn_state;
+	int i;
+	bool ret;
+
+	for (i = 0; i < ncrtcs; i++) {
+		crtc_state = state->crtc_states[i];
+
+		if (!crtc_state || !crtc_state->mode_changed)
+			continue;
+
+		drm_mode_copy(&crtc_state->adjusted_mode, &crtc_state->mode);
+	}
+
+	for (i = 0; i < state->num_connector; i++) {
+		struct drm_encoder_helper_funcs *funcs;
+		struct drm_encoder *encoder;
+
+		conn_state = state->connector_states[i];
+
+		if (!conn_state)
+			continue;
+
+		WARN_ON(!!conn_state->best_encoder != !!conn_state->crtc);
+
+		if (!conn_state->crtc || !conn_state->best_encoder)
+			continue;
+
+		crtc_state =
+			state->crtc_states[drm_crtc_index(conn_state->crtc)];
+
+		/*
+		 * Each encoder has at most one connector (since we always steal
+		 * it away), so we won't call ->mode_fixup twice.
+		 */
+		encoder = conn_state->best_encoder;
+		funcs = encoder->helper_private;
+
+		if (encoder->bridge && encoder->bridge->funcs->mode_fixup) {
+			ret = encoder->bridge->funcs->mode_fixup(
+					encoder->bridge, &crtc_state->mode,
+					&crtc_state->adjusted_mode);
+			if (!ret) {
+				DRM_DEBUG_KMS("Bridge fixup failed\n");
+				return -EINVAL;
+			}
+		}
+
+
+		ret = funcs->mode_fixup(encoder, &crtc_state->mode,
+					&crtc_state->adjusted_mode);
+		if (!ret) {
+			DRM_DEBUG_KMS("[ENCODER:%d:%s] fixup failed\n",
+				      encoder->base.id, encoder->name);
+			return -EINVAL;
+		}
+	}
+
+	for (i = 0; i < ncrtcs; i++) {
+		struct drm_crtc_helper_funcs *funcs;
+		struct drm_crtc *crtc;
+
+		crtc_state = state->crtc_states[i];
+		crtc = state->crtcs[i];
+
+		if (!crtc_state || !crtc_state->mode_changed)
+			continue;
+
+		funcs = crtc->helper_private;
+		ret = funcs->mode_fixup(crtc, &crtc_state->mode,
+					&crtc_state->adjusted_mode);
+		if (!ret) {
+			DRM_DEBUG_KMS("[CRTC:%d] fixup failed\n",
+				      crtc->base.id);
+			return -EINVAL;
+		}
+	}
+
+	return 0;
+}
+
+static int
+drm_atomic_helper_check_modeset(struct drm_device *dev,
+				struct drm_atomic_state *state)
+{
+	int ncrtcs = dev->mode_config.num_crtc;
+	struct drm_crtc *crtc;
+	struct drm_crtc_state *crtc_state;
+	int i, ret;
+
+	for (i = 0; i < ncrtcs; i++) {
+		crtc = state->crtcs[i];
+		crtc_state = state->crtc_states[i];
+
+		if (!crtc)
+			continue;
+
+		if (!drm_mode_equal(&crtc->state->mode, &crtc_state->mode)) {
+			DRM_DEBUG_KMS("[CRTC:%d] mode changed\n",
+				      crtc->base.id);
+			crtc_state->mode_changed = true;
+		}
+
+		if (crtc->state->enable != crtc_state->enable) {
+			DRM_DEBUG_KMS("[CRTC:%d] enable changed\n",
+				      crtc->base.id);
+			crtc_state->mode_changed = true;
+		}
+	}
+
+	for (i = 0; i < state->num_connector; i++) {
+		/*
+		 * This only sets crtc->mode_changed for routing changes,
+		 * drivers must set crtc->mode_changed themselves when connector
+		 * properties need to be updated.
+		 */
+		ret = update_connector_routing(state, i);
+		if (ret)
+			return ret;
+	}
+
+	/*
+	 * After all the routing has been prepared we need to add in any
+	 * connector which is itself unchanged, but who's crtc changes it's
+	 * configuration. This must be done before calling mode_fixup in case a
+	 * crtc only changed its mode but has the same set of connectors.
+	 */
+	for (i = 0; i < ncrtcs; i++) {
+		int num_connectors;
+
+		crtc = state->crtcs[i];
+		crtc_state = state->crtc_states[i];
+
+		if (!crtc || !crtc_state->mode_changed)
+			continue;
+
+		DRM_DEBUG_KMS("[CRTC:%d] needs full modeset, enable: %c\n",
+			      crtc->base.id,
+			      crtc_state->enable ? 'y' : 'n');
+
+		ret = drm_atomic_add_affected_connectors(state, crtc);
+		if (ret != 0)
+			return ret;
+
+		num_connectors = drm_atomic_connectors_for_crtc(state,
+								crtc);
+
+		if (crtc_state->enable != !!num_connectors) {
+			DRM_DEBUG_KMS("[CRTC:%d] enabled/connectors mismatch\n",
+				      crtc->base.id);
+
+			return -EINVAL;
+		}
+	}
+
+	return mode_fixup(state);
+}
+
+/**
+ * drm_atomic_helper_check - validate state object
+ * @dev: DRM device
+ * @state: the driver state object
+ *
+ * Check the state object to see if the requested state is physically possible.
+ * Only crtcs and planes have check callbacks, so for any additional (global)
+ * checking that a driver needs it can simply wrap that around this function.
+ * Drivers without such needs can directly use this as their ->atomic_check()
+ * callback.
+ *
+ * RETURNS
+ * Zero for success or -errno
+ */
+int drm_atomic_helper_check(struct drm_device *dev,
+			    struct drm_atomic_state *state)
+{
+	int nplanes = dev->mode_config.num_total_plane;
+	int ncrtcs = dev->mode_config.num_crtc;
+	int i, ret = 0;
+
+	for (i = 0; i < nplanes; i++) {
+		struct drm_plane_helper_funcs *funcs;
+		struct drm_plane *plane = state->planes[i];
+		struct drm_plane_state *plane_state = state->plane_states[i];
+
+		if (!plane)
+			continue;
+
+		funcs = plane->helper_private;
+
+		drm_atomic_helper_plane_changed(state, plane_state, plane);
+
+		if (!funcs || !funcs->atomic_check)
+			continue;
+
+		ret = funcs->atomic_check(plane, plane_state);
+		if (ret) {
+			DRM_DEBUG_KMS("[PLANE:%d] atomic check failed\n",
+				      plane->base.id);
+			return ret;
+		}
+	}
+
+	for (i = 0; i < ncrtcs; i++) {
+		struct drm_crtc_helper_funcs *funcs;
+		struct drm_crtc *crtc = state->crtcs[i];
+
+		if (!crtc)
+			continue;
+
+		funcs = crtc->helper_private;
+
+		if (!funcs || !funcs->atomic_check)
+			continue;
+
+		ret = funcs->atomic_check(crtc, state->crtc_states[i]);
+		if (ret) {
+			DRM_DEBUG_KMS("[CRTC:%d] atomic check failed\n",
+				      crtc->base.id);
+			return ret;
+		}
+	}
+
+	ret = drm_atomic_helper_check_modeset(dev, state);
+	if (ret)
+		return ret;
+
+	return ret;
+}
+EXPORT_SYMBOL(drm_atomic_helper_check);
+
+static void
+disable_outputs(struct drm_device *dev, struct drm_atomic_state *old_state)
+{
+	int ncrtcs = old_state->dev->mode_config.num_crtc;
+	int i;
+
+	for (i = 0; i < old_state->num_connector; i++) {
+		struct drm_connector_state *old_conn_state;
+		struct drm_connector *connector;
+		struct drm_encoder_helper_funcs *funcs;
+		struct drm_encoder *encoder;
+
+		old_conn_state = old_state->connector_states[i];
+		connector = old_state->connectors[i];
+
+		/* Shut down everything that's in the changeset and currently
+		 * still on. So need to check the old, saved state. */
+		if (!old_conn_state || !old_conn_state->crtc)
+			continue;
+
+		encoder = old_conn_state->best_encoder;
+
+		/* We shouldn't get this far if we didn't previously have
+		 * an encoder.. but WARN_ON() rather than explode.
+		 */
+		if (WARN_ON(!encoder))
+			continue;
+
+		funcs = encoder->helper_private;
+
+		/*
+		 * Each encoder has at most one connector (since we always steal
+		 * it away), so we won't call call disable hooks twice.
+		 */
+		if (encoder->bridge)
+			encoder->bridge->funcs->disable(encoder->bridge);
+
+		/* Right function depends upon target state. */
+		if (connector->state->crtc)
+			funcs->prepare(encoder);
+		else if (funcs->disable)
+			funcs->disable(encoder);
+		else
+			funcs->dpms(encoder, DRM_MODE_DPMS_OFF);
+
+		if (encoder->bridge)
+			encoder->bridge->funcs->post_disable(encoder->bridge);
+	}
+
+	for (i = 0; i < ncrtcs; i++) {
+		struct drm_crtc_helper_funcs *funcs;
+		struct drm_crtc *crtc;
+
+		crtc = old_state->crtcs[i];
+
+		/* Shut down everything that needs a full modeset. */
+		if (!crtc || !crtc->state->mode_changed)
+			continue;
+
+		funcs = crtc->helper_private;
+
+		/* Right function depends upon target state. */
+		if (crtc->state->enable)
+			funcs->prepare(crtc);
+		else if (funcs->disable)
+			funcs->disable(crtc);
+		else
+			funcs->dpms(crtc, DRM_MODE_DPMS_OFF);
+	}
+}
+
+static void
+set_routing_links(struct drm_device *dev, struct drm_atomic_state *old_state)
+{
+	int ncrtcs = old_state->dev->mode_config.num_crtc;
+	int i;
+
+	/* clear out existing links */
+	for (i = 0; i < old_state->num_connector; i++) {
+		struct drm_connector *connector;
+
+		connector = old_state->connectors[i];
+
+		if (!connector || !connector->encoder)
+			continue;
+
+		WARN_ON(!connector->encoder->crtc);
+
+		connector->encoder->crtc = NULL;
+		connector->encoder = NULL;
+	}
+
+	/* set new links */
+	for (i = 0; i < old_state->num_connector; i++) {
+		struct drm_connector *connector;
+
+		connector = old_state->connectors[i];
+
+		if (!connector || !connector->state->crtc)
+			continue;
+
+		if (WARN_ON(!connector->state->best_encoder))
+			continue;
+
+		connector->encoder = connector->state->best_encoder;
+		connector->encoder->crtc = connector->state->crtc;
+	}
+
+	/* set legacy state in the crtc structure */
+	for (i = 0; i < ncrtcs; i++) {
+		struct drm_crtc *crtc;
+
+		crtc = old_state->crtcs[i];
+
+		if (!crtc)
+			continue;
+
+		crtc->mode = crtc->state->mode;
+		crtc->enabled = crtc->state->enable;
+		crtc->x = crtc->primary->state->src_x >> 16;
+		crtc->y = crtc->primary->state->src_y >> 16;
+	}
+}
+
+static void
+crtc_set_mode(struct drm_device *dev, struct drm_atomic_state *old_state)
+{
+	int ncrtcs = old_state->dev->mode_config.num_crtc;
+	int i;
+
+	for (i = 0; i < ncrtcs; i++) {
+		struct drm_crtc_helper_funcs *funcs;
+		struct drm_crtc *crtc;
+
+		crtc = old_state->crtcs[i];
+
+		if (!crtc || !crtc->state->mode_changed)
+			continue;
+
+		funcs = crtc->helper_private;
+
+		if (crtc->state->enable)
+			funcs->mode_set_nofb(crtc);
+	}
+
+	for (i = 0; i < old_state->num_connector; i++) {
+		struct drm_connector *connector;
+		struct drm_crtc_state *new_crtc_state;
+		struct drm_encoder_helper_funcs *funcs;
+		struct drm_encoder *encoder;
+		struct drm_display_mode *mode, *adjusted_mode;
+
+		connector = old_state->connectors[i];
+
+		if (!connector || !connector->state->best_encoder)
+			continue;
+
+		encoder = connector->state->best_encoder;
+		funcs = encoder->helper_private;
+		new_crtc_state = connector->state->crtc->state;
+		mode = &new_crtc_state->mode;
+		adjusted_mode = &new_crtc_state->adjusted_mode;
+
+		/*
+		 * Each encoder has at most one connector (since we always steal
+		 * it away), so we won't call call mode_set hooks twice.
+		 */
+		funcs->mode_set(encoder, mode, adjusted_mode);
+
+		if (encoder->bridge && encoder->bridge->funcs->mode_set)
+			encoder->bridge->funcs->mode_set(encoder->bridge,
+							 mode, adjusted_mode);
+	}
+}
+
+/**
+ * drm_atomic_helper_commit_pre_planes - modeset commit before plane updates
+ * @dev: DRM device
+ * @state: atomic state
+ *
+ * This function commits the modeset changes that need to be committed before
+ * updating planes. It shuts down all the outputs that need to be shut down and
+ * prepares them (if required) with the new mode.
+ */
+void drm_atomic_helper_commit_pre_planes(struct drm_device *dev,
+					 struct drm_atomic_state *state)
+{
+	disable_outputs(dev, state);
+	set_routing_links(dev, state);
+	crtc_set_mode(dev, state);
+}
+EXPORT_SYMBOL(drm_atomic_helper_commit_pre_planes);
+
+/**
+ * drm_atomic_helper_commit_post_planes - modeset commit after plane updates
+ * @dev: DRM device
+ * @old_state: atomic state object with old state structures
+ *
+ * This function commits the modeset changes that need to be committed after
+ * updating planes: It enables all the outputs with the new configuration which
+ * had to be turned off for the update.
+ */
+void drm_atomic_helper_commit_post_planes(struct drm_device *dev,
+					  struct drm_atomic_state *old_state)
+{
+	int ncrtcs = old_state->dev->mode_config.num_crtc;
+	int i;
+
+	for (i = 0; i < ncrtcs; i++) {
+		struct drm_crtc_helper_funcs *funcs;
+		struct drm_crtc *crtc;
+
+		crtc = old_state->crtcs[i];
+
+		/* Need to filter out CRTCs where only planes change. */
+		if (!crtc || !crtc->state->mode_changed)
+			continue;
+
+		funcs = crtc->helper_private;
+
+		if (crtc->state->enable)
+			funcs->commit(crtc);
+	}
+
+	for (i = 0; i < old_state->num_connector; i++) {
+		struct drm_connector *connector;
+		struct drm_encoder_helper_funcs *funcs;
+		struct drm_encoder *encoder;
+
+		connector = old_state->connectors[i];
+
+		if (!connector || !connector->state->best_encoder)
+			continue;
+
+		encoder = connector->state->best_encoder;
+		funcs = encoder->helper_private;
+
+		/*
+		 * Each encoder has at most one connector (since we always steal
+		 * it away), so we won't call call enable hooks twice.
+		 */
+		if (encoder->bridge)
+			encoder->bridge->funcs->pre_enable(encoder->bridge);
+
+		funcs->commit(encoder);
+
+		if (encoder->bridge)
+			encoder->bridge->funcs->enable(encoder->bridge);
+	}
+}
+EXPORT_SYMBOL(drm_atomic_helper_commit_post_planes);
+
+static void wait_for_fences(struct drm_device *dev,
+			    struct drm_atomic_state *state)
+{
+	int nplanes = dev->mode_config.num_total_plane;
+	int i;
+
+	for (i = 0; i < nplanes; i++) {
+		struct drm_plane *plane = state->planes[i];
+
+		if (!plane || !plane->state->fence)
+			continue;
+
+		WARN_ON(!plane->state->fb);
+
+		fence_wait(plane->state->fence, false);
+		fence_put(plane->state->fence);
+		plane->state->fence = NULL;
+	}
+}
+
+static bool framebuffer_changed(struct drm_device *dev,
+				struct drm_atomic_state *old_state,
+				struct drm_crtc *crtc)
+{
+	struct drm_plane *plane;
+	struct drm_plane_state *old_plane_state;
+	int nplanes = old_state->dev->mode_config.num_total_plane;
+	int i;
+
+	for (i = 0; i < nplanes; i++) {
+		plane = old_state->planes[i];
+		old_plane_state = old_state->plane_states[i];
+
+		if (!plane)
+			continue;
+
+		if (plane->state->crtc != crtc &&
+		    old_plane_state->crtc != crtc)
+			continue;
+
+		if (plane->state->fb != old_plane_state->fb)
+			return true;
+	}
+
+	return false;
+}
+
+/**
+ * drm_atomic_helper_wait_for_vblanks - wait for vblank on crtcs
+ * @dev: DRM device
+ * @old_state: atomic state object with old state structures
+ *
+ * Helper to, after atomic commit, wait for vblanks on all effected
+ * crtcs (ie. before cleaning up old framebuffers using
+ * drm_atomic_helper_cleanup_planes()). It will only wait on crtcs where the
+ * framebuffers have actually changed to optimize for the legacy cursor and
+ * plane update use-case.
+ */
+void
+drm_atomic_helper_wait_for_vblanks(struct drm_device *dev,
+		struct drm_atomic_state *old_state)
+{
+	struct drm_crtc *crtc;
+	struct drm_crtc_state *old_crtc_state;
+	int ncrtcs = old_state->dev->mode_config.num_crtc;
+	int i, ret;
+
+	for (i = 0; i < ncrtcs; i++) {
+		crtc = old_state->crtcs[i];
+		old_crtc_state = old_state->crtc_states[i];
+
+		if (!crtc)
+			continue;
+
+		/* No one cares about the old state, so abuse it for tracking
+		 * and store whether we hold a vblank reference (and should do a
+		 * vblank wait) in the ->enable boolean. */
+		old_crtc_state->enable = false;
+
+		if (!crtc->state->enable)
+			continue;
+
+		if (!framebuffer_changed(dev, old_state, crtc))
+			continue;
+
+		ret = drm_crtc_vblank_get(crtc);
+		if (ret != 0)
+			continue;
+
+		old_crtc_state->enable = true;
+		old_crtc_state->last_vblank_count = drm_vblank_count(dev, i);
+	}
+
+	for (i = 0; i < ncrtcs; i++) {
+		crtc = old_state->crtcs[i];
+		old_crtc_state = old_state->crtc_states[i];
+
+		if (!crtc || !old_crtc_state->enable)
+			continue;
+
+		ret = wait_event_timeout(dev->vblank[i].queue,
+				old_crtc_state->last_vblank_count !=
+					drm_vblank_count(dev, i),
+				msecs_to_jiffies(50));
+
+		drm_crtc_vblank_put(crtc);
+	}
+}
+EXPORT_SYMBOL(drm_atomic_helper_wait_for_vblanks);
+
+/**
+ * drm_atomic_helper_commit - commit validated state object
+ * @dev: DRM device
+ * @state: the driver state object
+ * @async: asynchronous commit
+ *
+ * This function commits a with drm_atomic_helper_check() pre-validated state
+ * object. This can still fail when e.g. the framebuffer reservation fails. For
+ * now this doesn't implement asynchronous commits.
+ *
+ * RETURNS
+ * Zero for success or -errno.
+ */
+int drm_atomic_helper_commit(struct drm_device *dev,
+			     struct drm_atomic_state *state,
+			     bool async)
+{
+	int ret;
+
+	if (async)
+		return -EBUSY;
+
+	ret = drm_atomic_helper_prepare_planes(dev, state);
+	if (ret)
+		return ret;
+
+	/*
+	 * This is the point of no return - everything below never fails except
+	 * when the hw goes bonghits. Which means we can commit the new state on
+	 * the software side now.
+	 */
+
+	drm_atomic_helper_swap_state(dev, state);
+
+	/*
+	 * Everything below can be run asynchronously without the need to grab
+	 * any modeset locks at all under one conditions: It must be guaranteed
+	 * that the asynchronous work has either been cancelled (if the driver
+	 * supports it, which at least requires that the framebuffers get
+	 * cleaned up with drm_atomic_helper_cleanup_planes()) or completed
+	 * before the new state gets committed on the software side with
+	 * drm_atomic_helper_swap_state().
+	 *
+	 * This scheme allows new atomic state updates to be prepared and
+	 * checked in parallel to the asynchronous completion of the previous
+	 * update. Which is important since compositors need to figure out the
+	 * composition of the next frame right after having submitted the
+	 * current layout.
+	 */
+
+	wait_for_fences(dev, state);
+
+	drm_atomic_helper_commit_pre_planes(dev, state);
+
+	drm_atomic_helper_commit_planes(dev, state);
+
+	drm_atomic_helper_commit_post_planes(dev, state);
+
+	drm_atomic_helper_wait_for_vblanks(dev, state);
+
+	drm_atomic_helper_cleanup_planes(dev, state);
+
+	drm_atomic_state_free(state);
+
+	return 0;
+}
+EXPORT_SYMBOL(drm_atomic_helper_commit);
+
+/**
+ * DOC: implementing async commit
+ *
+ * For now the atomic helpers don't support async commit directly. If there is
+ * real need it could be added though, using the dma-buf fence infrastructure
+ * for generic synchronization with outstanding rendering.
+ *
+ * For now drivers have to implement async commit themselves, with the following
+ * sequence being the recommended one:
+ *
+ * 1. Run drm_atomic_helper_prepare_planes() first. This is the only function
+ * which commit needs to call which can fail, so we want to run it first and
+ * synchronously.
+ *
+ * 2. Synchronize with any outstanding asynchronous commit worker threads which
+ * might be affected the new state update. This can be done by either cancelling
+ * or flushing the work items, depending upon whether the driver can deal with
+ * cancelled updates. Note that it is important to ensure that the framebuffer
+ * cleanup is still done when cancelling.
+ *
+ * For sufficient parallelism it is recommended to have a work item per crtc
+ * (for updates which don't touch global state) and a global one. Then we only
+ * need to synchronize with the crtc work items for changed crtcs and the global
+ * work item, which allows nice concurrent updates on disjoint sets of crtcs.
+ *
+ * 3. The software state is updated synchronously with
+ * drm_atomic_helper_swap_state. Doing this under the protection of all modeset
+ * locks means concurrent callers never see inconsistent state. And doing this
+ * while it's guaranteed that no relevant async worker runs means that async
+ * workers do not need grab any locks. Actually they must not grab locks, for
+ * otherwise the work flushing will deadlock.
+ *
+ * 4. Schedule a work item to do all subsequent steps, using the split-out
+ * commit helpers: a) pre-plane commit b) plane commit c) post-plane commit and
+ * then cleaning up the framebuffers after the old framebuffer is no longer
+ * being displayed.
+ */
+
+/**
+ * drm_atomic_helper_prepare_planes - prepare plane resources after commit
+ * @dev: DRM device
+ * @state: atomic state object with old state structures
+ *
+ * This function prepares plane state, specifically framebuffers, for the new
+ * configuration. If any failure is encountered this function will call
+ * ->cleanup_fb on any already successfully prepared framebuffer.
+ *
+ * Returns:
+ * 0 on success, negative error code on failure.
+ */
+int drm_atomic_helper_prepare_planes(struct drm_device *dev,
+				     struct drm_atomic_state *state)
+{
+	int nplanes = dev->mode_config.num_total_plane;
+	int ret, i;
+
+	for (i = 0; i < nplanes; i++) {
+		struct drm_plane_helper_funcs *funcs;
+		struct drm_plane *plane = state->planes[i];
+		struct drm_framebuffer *fb;
+
+		if (!plane)
+			continue;
+
+		funcs = plane->helper_private;
+
+		fb = state->plane_states[i]->fb;
+
+		if (fb && funcs->prepare_fb) {
+			ret = funcs->prepare_fb(plane, fb);
+			if (ret)
+				goto fail;
+		}
+	}
+
+	return 0;
+
+fail:
+	for (i--; i >= 0; i--) {
+		struct drm_plane_helper_funcs *funcs;
+		struct drm_plane *plane = state->planes[i];
+		struct drm_framebuffer *fb;
+
+		if (!plane)
+			continue;
+
+		funcs = plane->helper_private;
+
+		fb = state->plane_states[i]->fb;
+
+		if (fb && funcs->cleanup_fb)
+			funcs->cleanup_fb(plane, fb);
+
+	}
+
+	return ret;
+}
+EXPORT_SYMBOL(drm_atomic_helper_prepare_planes);
+
+/**
+ * drm_atomic_helper_commit_planes - commit plane state
+ * @dev: DRM device
+ * @old_state: atomic state object with old state structures
+ *
+ * This function commits the new plane state using the plane and atomic helper
+ * functions for planes and crtcs. It assumes that the atomic state has already
+ * been pushed into the relevant object state pointers, since this step can no
+ * longer fail.
+ *
+ * It still requires the global state object @old_state to know which planes and
+ * crtcs need to be updated though.
+ */
+void drm_atomic_helper_commit_planes(struct drm_device *dev,
+				     struct drm_atomic_state *old_state)
+{
+	int nplanes = dev->mode_config.num_total_plane;
+	int ncrtcs = dev->mode_config.num_crtc;
+	int i;
+
+	for (i = 0; i < ncrtcs; i++) {
+		struct drm_crtc_helper_funcs *funcs;
+		struct drm_crtc *crtc = old_state->crtcs[i];
+
+		if (!crtc)
+			continue;
+
+		funcs = crtc->helper_private;
+
+		if (!funcs || !funcs->atomic_begin)
+			continue;
+
+		funcs->atomic_begin(crtc);
+	}
+
+	for (i = 0; i < nplanes; i++) {
+		struct drm_plane_helper_funcs *funcs;
+		struct drm_plane *plane = old_state->planes[i];
+		struct drm_plane_state *old_plane_state;
+
+		if (!plane)
+			continue;
+
+		funcs = plane->helper_private;
+
+		if (!funcs || !funcs->atomic_update)
+			continue;
+
+		old_plane_state = old_state->plane_states[i];
+
+		funcs->atomic_update(plane, old_plane_state);
+	}
+
+	for (i = 0; i < ncrtcs; i++) {
+		struct drm_crtc_helper_funcs *funcs;
+		struct drm_crtc *crtc = old_state->crtcs[i];
+
+		if (!crtc)
+			continue;
+
+		funcs = crtc->helper_private;
+
+		if (!funcs || !funcs->atomic_flush)
+			continue;
+
+		funcs->atomic_flush(crtc);
+	}
+}
+EXPORT_SYMBOL(drm_atomic_helper_commit_planes);
+
+/**
+ * drm_atomic_helper_cleanup_planes - cleanup plane resources after commit
+ * @dev: DRM device
+ * @old_state: atomic state object with old state structures
+ *
+ * This function cleans up plane state, specifically framebuffers, from the old
+ * configuration. Hence the old configuration must be perserved in @old_state to
+ * be able to call this function.
+ *
+ * This function must also be called on the new state when the atomic update
+ * fails at any point after calling drm_atomic_helper_prepare_planes().
+ */
+void drm_atomic_helper_cleanup_planes(struct drm_device *dev,
+				      struct drm_atomic_state *old_state)
+{
+	int nplanes = dev->mode_config.num_total_plane;
+	int i;
+
+	for (i = 0; i < nplanes; i++) {
+		struct drm_plane_helper_funcs *funcs;
+		struct drm_plane *plane = old_state->planes[i];
+		struct drm_framebuffer *old_fb;
+
+		if (!plane)
+			continue;
+
+		funcs = plane->helper_private;
+
+		old_fb = old_state->plane_states[i]->fb;
+
+		if (old_fb && funcs->cleanup_fb)
+			funcs->cleanup_fb(plane, old_fb);
+	}
+}
+EXPORT_SYMBOL(drm_atomic_helper_cleanup_planes);
+
+/**
+ * drm_atomic_helper_swap_state - store atomic state into current sw state
+ * @dev: DRM device
+ * @state: atomic state
+ *
+ * This function stores the atomic state into the current state pointers in all
+ * driver objects. It should be called after all failing steps have been done
+ * and succeeded, but before the actual hardware state is committed.
+ *
+ * For cleanup and error recovery the current state for all changed objects will
+ * be swaped into @state.
+ *
+ * With that sequence it fits perfectly into the plane prepare/cleanup sequence:
+ *
+ * 1. Call drm_atomic_helper_prepare_planes() with the staged atomic state.
+ *
+ * 2. Do any other steps that might fail.
+ *
+ * 3. Put the staged state into the current state pointers with this function.
+ *
+ * 4. Actually commit the hardware state.
+ *
+ * 5. Call drm_atomic_helper_cleanup_planes with @state, which since step 3
+ * contains the old state. Also do any other cleanup required with that state.
+ */
+void drm_atomic_helper_swap_state(struct drm_device *dev,
+				  struct drm_atomic_state *state)
+{
+	int i;
+
+	for (i = 0; i < dev->mode_config.num_connector; i++) {
+		struct drm_connector *connector = state->connectors[i];
+
+		if (!connector)
+			continue;
+
+		connector->state->state = state;
+		swap(state->connector_states[i], connector->state);
+		connector->state->state = NULL;
+	}
+
+	for (i = 0; i < dev->mode_config.num_crtc; i++) {
+		struct drm_crtc *crtc = state->crtcs[i];
+
+		if (!crtc)
+			continue;
+
+		crtc->state->state = state;
+		swap(state->crtc_states[i], crtc->state);
+		crtc->state->state = NULL;
+	}
+
+	for (i = 0; i < dev->mode_config.num_total_plane; i++) {
+		struct drm_plane *plane = state->planes[i];
+
+		if (!plane)
+			continue;
+
+		plane->state->state = state;
+		swap(state->plane_states[i], plane->state);
+		plane->state->state = NULL;
+	}
+}
+EXPORT_SYMBOL(drm_atomic_helper_swap_state);
+
+/**
+ * drm_atomic_helper_update_plane - Helper for primary plane update using atomic
+ * @plane: plane object to update
+ * @crtc: owning CRTC of owning plane
+ * @fb: framebuffer to flip onto plane
+ * @crtc_x: x offset of primary plane on crtc
+ * @crtc_y: y offset of primary plane on crtc
+ * @crtc_w: width of primary plane rectangle on crtc
+ * @crtc_h: height of primary plane rectangle on crtc
+ * @src_x: x offset of @fb for panning
+ * @src_y: y offset of @fb for panning
+ * @src_w: width of source rectangle in @fb
+ * @src_h: height of source rectangle in @fb
+ *
+ * Provides a default plane update handler using the atomic driver interface.
+ *
+ * RETURNS:
+ * Zero on success, error code on failure
+ */
+int drm_atomic_helper_update_plane(struct drm_plane *plane,
+				   struct drm_crtc *crtc,
+				   struct drm_framebuffer *fb,
+				   int crtc_x, int crtc_y,
+				   unsigned int crtc_w, unsigned int crtc_h,
+				   uint32_t src_x, uint32_t src_y,
+				   uint32_t src_w, uint32_t src_h)
+{
+	struct drm_atomic_state *state;
+	struct drm_plane_state *plane_state;
+	int ret = 0;
+
+	state = drm_atomic_state_alloc(plane->dev);
+	if (!state)
+		return -ENOMEM;
+
+	state->acquire_ctx = drm_modeset_legacy_acquire_ctx(crtc);
+retry:
+	plane_state = drm_atomic_get_plane_state(state, plane);
+	if (IS_ERR(plane_state)) {
+		ret = PTR_ERR(plane_state);
+		goto fail;
+	}
+
+	ret = drm_atomic_set_crtc_for_plane(state, plane, crtc);
+	if (ret != 0)
+		goto fail;
+	drm_atomic_set_fb_for_plane(plane_state, fb);
+	plane_state->crtc_x = crtc_x;
+	plane_state->crtc_y = crtc_y;
+	plane_state->crtc_h = crtc_h;
+	plane_state->crtc_w = crtc_w;
+	plane_state->src_x = src_x;
+	plane_state->src_y = src_y;
+	plane_state->src_h = src_h;
+	plane_state->src_w = src_w;
+
+	ret = drm_atomic_commit(state);
+	if (ret != 0)
+		goto fail;
+
+	/* Driver takes ownership of state on successful commit. */
+	return 0;
+fail:
+	if (ret == -EDEADLK)
+		goto backoff;
+
+	drm_atomic_state_free(state);
+
+	return ret;
+backoff:
+	drm_atomic_state_clear(state);
+	drm_atomic_legacy_backoff(state);
+
+	/*
+	 * Someone might have exchanged the framebuffer while we dropped locks
+	 * in the backoff code. We need to fix up the fb refcount tracking the
+	 * core does for us.
+	 */
+	plane->old_fb = plane->fb;
+
+	goto retry;
+}
+EXPORT_SYMBOL(drm_atomic_helper_update_plane);
+
+/**
+ * drm_atomic_helper_disable_plane - Helper for primary plane disable using * atomic
+ * @plane: plane to disable
+ *
+ * Provides a default plane disable handler using the atomic driver interface.
+ *
+ * RETURNS:
+ * Zero on success, error code on failure
+ */
+int drm_atomic_helper_disable_plane(struct drm_plane *plane)
+{
+	struct drm_atomic_state *state;
+	struct drm_plane_state *plane_state;
+	int ret = 0;
+
+	/*
+	 * FIXME: Without plane->crtc set we can't get at the implicit legacy
+	 * acquire context. The real fix will be to wire the acquire ctx through
+	 * everywhere we need it, but meanwhile prevent chaos by just skipping
+	 * this noop. The critical case is the cursor ioctls which a) only grab
+	 * crtc/cursor-plane locks (so we need the crtc to get at the right
+	 * acquire context) and b) can try to disable the plane multiple times.
+	 */
+	if (!plane->crtc)
+		return 0;
+
+	state = drm_atomic_state_alloc(plane->dev);
+	if (!state)
+		return -ENOMEM;
+
+	state->acquire_ctx = drm_modeset_legacy_acquire_ctx(plane->crtc);
+retry:
+	plane_state = drm_atomic_get_plane_state(state, plane);
+	if (IS_ERR(plane_state)) {
+		ret = PTR_ERR(plane_state);
+		goto fail;
+	}
+
+	ret = drm_atomic_set_crtc_for_plane(state, plane, NULL);
+	if (ret != 0)
+		goto fail;
+	drm_atomic_set_fb_for_plane(plane_state, NULL);
+	plane_state->crtc_x = 0;
+	plane_state->crtc_y = 0;
+	plane_state->crtc_h = 0;
+	plane_state->crtc_w = 0;
+	plane_state->src_x = 0;
+	plane_state->src_y = 0;
+	plane_state->src_h = 0;
+	plane_state->src_w = 0;
+
+	ret = drm_atomic_commit(state);
+	if (ret != 0)
+		goto fail;
+
+	/* Driver takes ownership of state on successful commit. */
+	return 0;
+fail:
+	if (ret == -EDEADLK)
+		goto backoff;
+
+	drm_atomic_state_free(state);
+
+	return ret;
+backoff:
+	drm_atomic_state_clear(state);
+	drm_atomic_legacy_backoff(state);
+
+	/*
+	 * Someone might have exchanged the framebuffer while we dropped locks
+	 * in the backoff code. We need to fix up the fb refcount tracking the
+	 * core does for us.
+	 */
+	plane->old_fb = plane->fb;
+
+	goto retry;
+}
+EXPORT_SYMBOL(drm_atomic_helper_disable_plane);
+
+static int update_output_state(struct drm_atomic_state *state,
+			       struct drm_mode_set *set)
+{
+	struct drm_device *dev = set->crtc->dev;
+	struct drm_connector_state *conn_state;
+	int ncrtcs = state->dev->mode_config.num_crtc;
+	int ret, i, j;
+
+	ret = drm_modeset_lock(&dev->mode_config.connection_mutex,
+			       state->acquire_ctx);
+	if (ret)
+		return ret;
+
+	/* First grab all affected connector/crtc states. */
+	for (i = 0; i < set->num_connectors; i++) {
+		conn_state = drm_atomic_get_connector_state(state,
+							    set->connectors[i]);
+		if (IS_ERR(conn_state))
+			return PTR_ERR(conn_state);
+	}
+
+	for (i = 0; i < ncrtcs; i++) {
+		struct drm_crtc *crtc = state->crtcs[i];
+
+		if (!crtc)
+			continue;
+
+		ret = drm_atomic_add_affected_connectors(state, crtc);
+		if (ret)
+			return ret;
+	}
+
+	/* Then recompute connector->crtc links and crtc enabling state. */
+	for (i = 0; i < state->num_connector; i++) {
+		struct drm_connector *connector;
+
+		connector = state->connectors[i];
+		conn_state = state->connector_states[i];
+
+		if (!connector)
+			continue;
+
+		if (conn_state->crtc == set->crtc) {
+			ret = drm_atomic_set_crtc_for_connector(conn_state,
+								NULL);
+			if (ret)
+				return ret;
+		}
+
+		for (j = 0; j < set->num_connectors; j++) {
+			if (set->connectors[j] == connector) {
+				ret = drm_atomic_set_crtc_for_connector(conn_state,
+									set->crtc);
+				if (ret)
+					return ret;
+				break;
+			}
+		}
+	}
+
+	for (i = 0; i < ncrtcs; i++) {
+		struct drm_crtc *crtc = state->crtcs[i];
+		struct drm_crtc_state *crtc_state = state->crtc_states[i];
+
+		if (!crtc)
+			continue;
+
+		/* Don't update ->enable for the CRTC in the set_config request,
+		 * since a mismatch would indicate a bug in the upper layers.
+		 * The actual modeset code later on will catch any
+		 * inconsistencies here. */
+		if (crtc == set->crtc)
+			continue;
+
+		crtc_state->enable =
+			drm_atomic_connectors_for_crtc(state, crtc);
+	}
+
+	return 0;
+}
+
+/**
+ * drm_atomic_helper_set_config - set a new config from userspace
+ * @set: mode set configuration
+ *
+ * Provides a default crtc set_config handler using the atomic driver interface.
+ *
+ * Returns:
+ * Returns 0 on success, negative errno numbers on failure.
+ */
+int drm_atomic_helper_set_config(struct drm_mode_set *set)
+{
+	struct drm_atomic_state *state;
+	struct drm_crtc *crtc = set->crtc;
+	struct drm_crtc_state *crtc_state;
+	struct drm_plane_state *primary_state;
+	int ret = 0;
+
+	state = drm_atomic_state_alloc(crtc->dev);
+	if (!state)
+		return -ENOMEM;
+
+	state->acquire_ctx = drm_modeset_legacy_acquire_ctx(crtc);
+retry:
+	crtc_state = drm_atomic_get_crtc_state(state, crtc);
+	if (IS_ERR(crtc_state)) {
+		ret = PTR_ERR(crtc_state);
+		goto fail;
+	}
+
+	primary_state = drm_atomic_get_plane_state(state, crtc->primary);
+	if (IS_ERR(primary_state)) {
+		ret = PTR_ERR(primary_state);
+		goto fail;
+	}
+
+	if (!set->mode) {
+		WARN_ON(set->fb);
+		WARN_ON(set->num_connectors);
+
+		crtc_state->enable = false;
+
+		ret = drm_atomic_set_crtc_for_plane(state, crtc->primary, NULL);
+		if (ret != 0)
+			goto fail;
+
+		drm_atomic_set_fb_for_plane(primary_state, NULL);
+
+		goto commit;
+	}
+
+	WARN_ON(!set->fb);
+	WARN_ON(!set->num_connectors);
+
+	crtc_state->enable = true;
+	drm_mode_copy(&crtc_state->mode, set->mode);
+
+	ret = drm_atomic_set_crtc_for_plane(state, crtc->primary, crtc);
+	if (ret != 0)
+		goto fail;
+	drm_atomic_set_fb_for_plane(primary_state, set->fb);
+	primary_state->crtc_x = 0;
+	primary_state->crtc_y = 0;
+	primary_state->crtc_h = set->mode->vdisplay;
+	primary_state->crtc_w = set->mode->hdisplay;
+	primary_state->src_x = set->x << 16;
+	primary_state->src_y = set->y << 16;
+	primary_state->src_h = set->mode->vdisplay << 16;
+	primary_state->src_w = set->mode->hdisplay << 16;
+
+commit:
+	ret = update_output_state(state, set);
+	if (ret)
+		goto fail;
+
+	ret = drm_atomic_commit(state);
+	if (ret != 0)
+		goto fail;
+
+	/* Driver takes ownership of state on successful commit. */
+	return 0;
+fail:
+	if (ret == -EDEADLK)
+		goto backoff;
+
+	drm_atomic_state_free(state);
+
+	return ret;
+backoff:
+	drm_atomic_state_clear(state);
+	drm_atomic_legacy_backoff(state);
+
+	/*
+	 * Someone might have exchanged the framebuffer while we dropped locks
+	 * in the backoff code. We need to fix up the fb refcount tracking the
+	 * core does for us.
+	 */
+	crtc->primary->old_fb = crtc->primary->fb;
+
+	goto retry;
+}
+EXPORT_SYMBOL(drm_atomic_helper_set_config);
+
+/**
+ * drm_atomic_helper_crtc_set_property - helper for crtc prorties
+ * @crtc: DRM crtc
+ * @property: DRM property
+ * @val: value of property
+ *
+ * Provides a default plane disablle handler using the atomic driver interface.
+ *
+ * RETURNS:
+ * Zero on success, error code on failure
+ */
+int
+drm_atomic_helper_crtc_set_property(struct drm_crtc *crtc,
+				    struct drm_property *property,
+				    uint64_t val)
+{
+	struct drm_atomic_state *state;
+	struct drm_crtc_state *crtc_state;
+	int ret = 0;
+
+	state = drm_atomic_state_alloc(crtc->dev);
+	if (!state)
+		return -ENOMEM;
+
+	/* ->set_property is always called with all locks held. */
+	state->acquire_ctx = crtc->dev->mode_config.acquire_ctx;
+retry:
+	crtc_state = drm_atomic_get_crtc_state(state, crtc);
+	if (IS_ERR(crtc_state)) {
+		ret = PTR_ERR(crtc_state);
+		goto fail;
+	}
+
+	ret = crtc->funcs->atomic_set_property(crtc, crtc_state,
+					       property, val);
+	if (ret)
+		goto fail;
+
+	ret = drm_atomic_commit(state);
+	if (ret != 0)
+		goto fail;
+
+	/* Driver takes ownership of state on successful commit. */
+	return 0;
+fail:
+	if (ret == -EDEADLK)
+		goto backoff;
+
+	drm_atomic_state_free(state);
+
+	return ret;
+backoff:
+	drm_atomic_state_clear(state);
+	drm_atomic_legacy_backoff(state);
+
+	goto retry;
+}
+EXPORT_SYMBOL(drm_atomic_helper_crtc_set_property);
+
+/**
+ * drm_atomic_helper_plane_set_property - helper for plane prorties
+ * @plane: DRM plane
+ * @property: DRM property
+ * @val: value of property
+ *
+ * Provides a default plane disable handler using the atomic driver interface.
+ *
+ * RETURNS:
+ * Zero on success, error code on failure
+ */
+int
+drm_atomic_helper_plane_set_property(struct drm_plane *plane,
+				    struct drm_property *property,
+				    uint64_t val)
+{
+	struct drm_atomic_state *state;
+	struct drm_plane_state *plane_state;
+	int ret = 0;
+
+	state = drm_atomic_state_alloc(plane->dev);
+	if (!state)
+		return -ENOMEM;
+
+	/* ->set_property is always called with all locks held. */
+	state->acquire_ctx = plane->dev->mode_config.acquire_ctx;
+retry:
+	plane_state = drm_atomic_get_plane_state(state, plane);
+	if (IS_ERR(plane_state)) {
+		ret = PTR_ERR(plane_state);
+		goto fail;
+	}
+
+	ret = plane->funcs->atomic_set_property(plane, plane_state,
+					       property, val);
+	if (ret)
+		goto fail;
+
+	ret = drm_atomic_commit(state);
+	if (ret != 0)
+		goto fail;
+
+	/* Driver takes ownership of state on successful commit. */
+	return 0;
+fail:
+	if (ret == -EDEADLK)
+		goto backoff;
+
+	drm_atomic_state_free(state);
+
+	return ret;
+backoff:
+	drm_atomic_state_clear(state);
+	drm_atomic_legacy_backoff(state);
+
+	goto retry;
+}
+EXPORT_SYMBOL(drm_atomic_helper_plane_set_property);
+
+/**
+ * drm_atomic_helper_connector_set_property - helper for connector prorties
+ * @connector: DRM connector
+ * @property: DRM property
+ * @val: value of property
+ *
+ * Provides a default plane disablle handler using the atomic driver interface.
+ *
+ * RETURNS:
+ * Zero on success, error code on failure
+ */
+int
+drm_atomic_helper_connector_set_property(struct drm_connector *connector,
+				    struct drm_property *property,
+				    uint64_t val)
+{
+	struct drm_atomic_state *state;
+	struct drm_connector_state *connector_state;
+	int ret = 0;
+
+	state = drm_atomic_state_alloc(connector->dev);
+	if (!state)
+		return -ENOMEM;
+
+	/* ->set_property is always called with all locks held. */
+	state->acquire_ctx = connector->dev->mode_config.acquire_ctx;
+retry:
+	connector_state = drm_atomic_get_connector_state(state, connector);
+	if (IS_ERR(connector_state)) {
+		ret = PTR_ERR(connector_state);
+		goto fail;
+	}
+
+	ret = connector->funcs->atomic_set_property(connector, connector_state,
+					       property, val);
+	if (ret)
+		goto fail;
+
+	ret = drm_atomic_commit(state);
+	if (ret != 0)
+		goto fail;
+
+	/* Driver takes ownership of state on successful commit. */
+	return 0;
+fail:
+	if (ret == -EDEADLK)
+		goto backoff;
+
+	drm_atomic_state_free(state);
+
+	return ret;
+backoff:
+	drm_atomic_state_clear(state);
+	drm_atomic_legacy_backoff(state);
+
+	goto retry;
+}
+EXPORT_SYMBOL(drm_atomic_helper_connector_set_property);
+
+/**
+ * drm_atomic_helper_page_flip - execute a legacy page flip
+ * @crtc: DRM crtc
+ * @fb: DRM framebuffer
+ * @event: optional DRM event to signal upon completion
+ * @flags: flip flags for non-vblank sync'ed updates
+ *
+ * Provides a default page flip implementation using the atomic driver interface.
+ *
+ * Note that for now so called async page flips (i.e. updates which are not
+ * synchronized to vblank) are not supported, since the atomic interfaces have
+ * no provisions for this yet.
+ *
+ * Returns:
+ * Returns 0 on success, negative errno numbers on failure.
+ */
+int drm_atomic_helper_page_flip(struct drm_crtc *crtc,
+				struct drm_framebuffer *fb,
+				struct drm_pending_vblank_event *event,
+				uint32_t flags)
+{
+	struct drm_plane *plane = crtc->primary;
+	struct drm_atomic_state *state;
+	struct drm_plane_state *plane_state;
+	struct drm_crtc_state *crtc_state;
+	int ret = 0;
+
+	if (flags & DRM_MODE_PAGE_FLIP_ASYNC)
+		return -EINVAL;
+
+	state = drm_atomic_state_alloc(plane->dev);
+	if (!state)
+		return -ENOMEM;
+
+	state->acquire_ctx = drm_modeset_legacy_acquire_ctx(crtc);
+retry:
+	crtc_state = drm_atomic_get_crtc_state(state, crtc);
+	if (IS_ERR(crtc_state)) {
+		ret = PTR_ERR(crtc_state);
+		goto fail;
+	}
+	crtc_state->event = event;
+
+	plane_state = drm_atomic_get_plane_state(state, plane);
+	if (IS_ERR(plane_state)) {
+		ret = PTR_ERR(plane_state);
+		goto fail;
+	}
+
+	ret = drm_atomic_set_crtc_for_plane(state, plane, crtc);
+	if (ret != 0)
+		goto fail;
+	drm_atomic_set_fb_for_plane(plane_state, fb);
+
+	ret = drm_atomic_async_commit(state);
+	if (ret != 0)
+		goto fail;
+
+	/* TODO: ->page_flip is the only driver callback where the core
+	 * doesn't update plane->fb. For now patch it up here. */
+	plane->fb = plane->state->fb;
+
+	/* Driver takes ownership of state on successful async commit. */
+	return 0;
+fail:
+	if (ret == -EDEADLK)
+		goto backoff;
+
+	drm_atomic_state_free(state);
+
+	return ret;
+backoff:
+	drm_atomic_state_clear(state);
+	drm_atomic_legacy_backoff(state);
+
+	/*
+	 * Someone might have exchanged the framebuffer while we dropped locks
+	 * in the backoff code. We need to fix up the fb refcount tracking the
+	 * core does for us.
+	 */
+	plane->old_fb = plane->fb;
+
+	goto retry;
+}
+EXPORT_SYMBOL(drm_atomic_helper_page_flip);
+
+/**
+ * DOC: atomic state reset and initialization
+ *
+ * Both the drm core and the atomic helpers assume that there is always the full
+ * and correct atomic software state for all connectors, CRTCs and planes
+ * available. Which is a bit a problem on driver load and also after system
+ * suspend. One way to solve this is to have a hardware state read-out
+ * infrastructure which reconstructs the full software state (e.g. the i915
+ * driver).
+ *
+ * The simpler solution is to just reset the software state to everything off,
+ * which is easiest to do by calling drm_mode_config_reset(). To facilitate this
+ * the atomic helpers provide default reset implementations for all hooks.
+ */
+
+/**
+ * drm_atomic_helper_crtc_reset - default ->reset hook for CRTCs
+ * @crtc: drm CRTC
+ *
+ * Resets the atomic state for @crtc by freeing the state pointer (which might
+ * be NULL, e.g. at driver load time) and allocating a new empty state object.
+ */
+void drm_atomic_helper_crtc_reset(struct drm_crtc *crtc)
+{
+	kfree(crtc->state);
+	crtc->state = kzalloc(sizeof(*crtc->state), GFP_KERNEL);
+}
+EXPORT_SYMBOL(drm_atomic_helper_crtc_reset);
+
+/**
+ * drm_atomic_helper_crtc_duplicate_state - default state duplicate hook
+ * @crtc: drm CRTC
+ *
+ * Default CRTC state duplicate hook for drivers which don't have their own
+ * subclassed CRTC state structure.
+ */
+struct drm_crtc_state *
+drm_atomic_helper_crtc_duplicate_state(struct drm_crtc *crtc)
+{
+	struct drm_crtc_state *state;
+
+	if (WARN_ON(!crtc->state))
+		return NULL;
+
+	state = kmemdup(crtc->state, sizeof(*crtc->state), GFP_KERNEL);
+
+	if (state) {
+		state->mode_changed = false;
+		state->planes_changed = false;
+		state->event = NULL;
+	}
+
+	return state;
+}
+EXPORT_SYMBOL(drm_atomic_helper_crtc_duplicate_state);
+
+/**
+ * drm_atomic_helper_crtc_destroy_state - default state destroy hook
+ * @crtc: drm CRTC
+ * @state: CRTC state object to release
+ *
+ * Default CRTC state destroy hook for drivers which don't have their own
+ * subclassed CRTC state structure.
+ */
+void drm_atomic_helper_crtc_destroy_state(struct drm_crtc *crtc,
+					  struct drm_crtc_state *state)
+{
+	kfree(state);
+}
+EXPORT_SYMBOL(drm_atomic_helper_crtc_destroy_state);
+
+/**
+ * drm_atomic_helper_plane_reset - default ->reset hook for planes
+ * @plane: drm plane
+ *
+ * Resets the atomic state for @plane by freeing the state pointer (which might
+ * be NULL, e.g. at driver load time) and allocating a new empty state object.
+ */
+void drm_atomic_helper_plane_reset(struct drm_plane *plane)
+{
+	if (plane->state && plane->state->fb)
+		drm_framebuffer_unreference(plane->state->fb);
+
+	kfree(plane->state);
+	plane->state = kzalloc(sizeof(*plane->state), GFP_KERNEL);
+}
+EXPORT_SYMBOL(drm_atomic_helper_plane_reset);
+
+/**
+ * drm_atomic_helper_plane_duplicate_state - default state duplicate hook
+ * @plane: drm plane
+ *
+ * Default plane state duplicate hook for drivers which don't have their own
+ * subclassed plane state structure.
+ */
+struct drm_plane_state *
+drm_atomic_helper_plane_duplicate_state(struct drm_plane *plane)
+{
+	struct drm_plane_state *state;
+
+	if (WARN_ON(!plane->state))
+		return NULL;
+
+	state = kmemdup(plane->state, sizeof(*plane->state), GFP_KERNEL);
+
+	if (state && state->fb)
+		drm_framebuffer_reference(state->fb);
+
+	return state;
+}
+EXPORT_SYMBOL(drm_atomic_helper_plane_duplicate_state);
+
+/**
+ * drm_atomic_helper_plane_destroy_state - default state destroy hook
+ * @plane: drm plane
+ * @state: plane state object to release
+ *
+ * Default plane state destroy hook for drivers which don't have their own
+ * subclassed plane state structure.
+ */
+void drm_atomic_helper_plane_destroy_state(struct drm_plane *plane,
+					   struct drm_plane_state *state)
+{
+	if (state->fb)
+		drm_framebuffer_unreference(state->fb);
+
+	kfree(state);
+}
+EXPORT_SYMBOL(drm_atomic_helper_plane_destroy_state);
+
+/**
+ * drm_atomic_helper_connector_reset - default ->reset hook for connectors
+ * @connector: drm connector
+ *
+ * Resets the atomic state for @connector by freeing the state pointer (which
+ * might be NULL, e.g. at driver load time) and allocating a new empty state
+ * object.
+ */
+void drm_atomic_helper_connector_reset(struct drm_connector *connector)
+{
+	kfree(connector->state);
+	connector->state = kzalloc(sizeof(*connector->state), GFP_KERNEL);
+}
+EXPORT_SYMBOL(drm_atomic_helper_connector_reset);
+
+/**
+ * drm_atomic_helper_connector_duplicate_state - default state duplicate hook
+ * @connector: drm connector
+ *
+ * Default connector state duplicate hook for drivers which don't have their own
+ * subclassed connector state structure.
+ */
+struct drm_connector_state *
+drm_atomic_helper_connector_duplicate_state(struct drm_connector *connector)
+{
+	if (WARN_ON(!connector->state))
+		return NULL;
+
+	return kmemdup(connector->state, sizeof(*connector->state), GFP_KERNEL);
+}
+EXPORT_SYMBOL(drm_atomic_helper_connector_duplicate_state);
+
+/**
+ * drm_atomic_helper_connector_destroy_state - default state destroy hook
+ * @connector: drm connector
+ * @state: connector state object to release
+ *
+ * Default connector state destroy hook for drivers which don't have their own
+ * subclassed connector state structure.
+ */
+void drm_atomic_helper_connector_destroy_state(struct drm_connector *connector,
+					  struct drm_connector_state *state)
+{
+	kfree(state);
+}
+EXPORT_SYMBOL(drm_atomic_helper_connector_destroy_state);
diff --git a/drivers/gpu/drm/drm_auth.c b/drivers/gpu/drm/drm_auth.c
index 3cedae12b3c1..fc8e8aaa34fb 100644
--- a/drivers/gpu/drm/drm_auth.c
+++ b/drivers/gpu/drm/drm_auth.c
@@ -34,6 +34,13 @@
  */
 
 #include <drm/drmP.h>
+#include "drm_internal.h"
+
+struct drm_magic_entry {
+	struct list_head head;
+	struct drm_hash_item hash_item;
+	struct drm_file *priv;
+};
 
 /**
  * Find the file with the given magic number.
diff --git a/drivers/gpu/drm/drm_bufs.c b/drivers/gpu/drm/drm_bufs.c
index 61acb8f6756d..569064a00693 100644
--- a/drivers/gpu/drm/drm_bufs.c
+++ b/drivers/gpu/drm/drm_bufs.c
@@ -1,18 +1,13 @@
-/**
- * \file drm_bufs.c
- * Generic buffer template
- *
- * \author Rickard E. (Rik) Faith <faith@valinux.com>
- * \author Gareth Hughes <gareth@valinux.com>
- */
-
 /*
- * Created: Thu Nov 23 03:10:50 2000 by gareth@valinux.com
+ * Legacy: Generic DRM Buffer Management
  *
  * Copyright 1999, 2000 Precision Insight, Inc., Cedar Park, Texas.
  * Copyright 2000 VA Linux Systems, Inc., Sunnyvale, California.
  * All Rights Reserved.
  *
+ * Author: Rickard E. (Rik) Faith <faith@valinux.com>
+ * Author: Gareth Hughes <gareth@valinux.com>
+ *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
  * to deal in the Software without restriction, including without limitation
@@ -39,6 +34,7 @@
 #include <linux/export.h>
 #include <asm/shmparam.h>
 #include <drm/drmP.h>
+#include "drm_legacy.h"
 
 static struct drm_map_list *drm_find_matching_map(struct drm_device *dev,
 						  struct drm_local_map *map)
@@ -365,9 +361,9 @@ static int drm_addmap_core(struct drm_device * dev, resource_size_t offset,
 	return 0;
 }
 
-int drm_addmap(struct drm_device * dev, resource_size_t offset,
-	       unsigned int size, enum drm_map_type type,
-	       enum drm_map_flags flags, struct drm_local_map ** map_ptr)
+int drm_legacy_addmap(struct drm_device * dev, resource_size_t offset,
+		      unsigned int size, enum drm_map_type type,
+		      enum drm_map_flags flags, struct drm_local_map **map_ptr)
 {
 	struct drm_map_list *list;
 	int rc;
@@ -377,8 +373,7 @@ int drm_addmap(struct drm_device * dev, resource_size_t offset,
 		*map_ptr = list->map;
 	return rc;
 }
-
-EXPORT_SYMBOL(drm_addmap);
+EXPORT_SYMBOL(drm_legacy_addmap);
 
 /**
  * Ioctl to specify a range of memory that is available for mapping by a
@@ -391,8 +386,8 @@ EXPORT_SYMBOL(drm_addmap);
  * \return zero on success or a negative value on error.
  *
  */
-int drm_addmap_ioctl(struct drm_device *dev, void *data,
-		     struct drm_file *file_priv)
+int drm_legacy_addmap_ioctl(struct drm_device *dev, void *data,
+			    struct drm_file *file_priv)
 {
 	struct drm_map *map = data;
 	struct drm_map_list *maplist;
@@ -429,9 +424,9 @@ int drm_addmap_ioctl(struct drm_device *dev, void *data,
  * its being used, and free any associate resource (such as MTRR's) if it's not
  * being on use.
  *
- * \sa drm_addmap
+ * \sa drm_legacy_addmap
  */
-int drm_rmmap_locked(struct drm_device *dev, struct drm_local_map *map)
+int drm_legacy_rmmap_locked(struct drm_device *dev, struct drm_local_map *map)
 {
 	struct drm_map_list *r_list = NULL, *list_t;
 	drm_dma_handle_t dmah;
@@ -478,26 +473,26 @@ int drm_rmmap_locked(struct drm_device *dev, struct drm_local_map *map)
 		dmah.vaddr = map->handle;
 		dmah.busaddr = map->offset;
 		dmah.size = map->size;
-		__drm_pci_free(dev, &dmah);
+		__drm_legacy_pci_free(dev, &dmah);
 		break;
 	}
 	kfree(map);
 
 	return 0;
 }
-EXPORT_SYMBOL(drm_rmmap_locked);
+EXPORT_SYMBOL(drm_legacy_rmmap_locked);
 
-int drm_rmmap(struct drm_device *dev, struct drm_local_map *map)
+int drm_legacy_rmmap(struct drm_device *dev, struct drm_local_map *map)
 {
 	int ret;
 
 	mutex_lock(&dev->struct_mutex);
-	ret = drm_rmmap_locked(dev, map);
+	ret = drm_legacy_rmmap_locked(dev, map);
 	mutex_unlock(&dev->struct_mutex);
 
 	return ret;
 }
-EXPORT_SYMBOL(drm_rmmap);
+EXPORT_SYMBOL(drm_legacy_rmmap);
 
 /* The rmmap ioctl appears to be unnecessary.  All mappings are torn down on
  * the last close of the device, and this is necessary for cleanup when things
@@ -514,8 +509,8 @@ EXPORT_SYMBOL(drm_rmmap);
  * \param arg pointer to a struct drm_map structure.
  * \return zero on success or a negative value on error.
  */
-int drm_rmmap_ioctl(struct drm_device *dev, void *data,
-		    struct drm_file *file_priv)
+int drm_legacy_rmmap_ioctl(struct drm_device *dev, void *data,
+			   struct drm_file *file_priv)
 {
 	struct drm_map *request = data;
 	struct drm_local_map *map = NULL;
@@ -546,7 +541,7 @@ int drm_rmmap_ioctl(struct drm_device *dev, void *data,
 		return 0;
 	}
 
-	ret = drm_rmmap_locked(dev, map);
+	ret = drm_legacy_rmmap_locked(dev, map);
 
 	mutex_unlock(&dev->struct_mutex);
 
@@ -599,7 +594,8 @@ static void drm_cleanup_buf_error(struct drm_device * dev,
  * reallocates the buffer list of the same size order to accommodate the new
  * buffers.
  */
-int drm_addbufs_agp(struct drm_device * dev, struct drm_buf_desc * request)
+int drm_legacy_addbufs_agp(struct drm_device *dev,
+			   struct drm_buf_desc *request)
 {
 	struct drm_device_dma *dma = dev->dma;
 	struct drm_buf_entry *entry;
@@ -759,10 +755,11 @@ int drm_addbufs_agp(struct drm_device * dev, struct drm_buf_desc * request)
 	atomic_dec(&dev->buf_alloc);
 	return 0;
 }
-EXPORT_SYMBOL(drm_addbufs_agp);
+EXPORT_SYMBOL(drm_legacy_addbufs_agp);
 #endif				/* __OS_HAS_AGP */
 
-int drm_addbufs_pci(struct drm_device * dev, struct drm_buf_desc * request)
+int drm_legacy_addbufs_pci(struct drm_device *dev,
+			   struct drm_buf_desc *request)
 {
 	struct drm_device_dma *dma = dev->dma;
 	int count;
@@ -964,9 +961,10 @@ int drm_addbufs_pci(struct drm_device * dev, struct drm_buf_desc * request)
 	return 0;
 
 }
-EXPORT_SYMBOL(drm_addbufs_pci);
+EXPORT_SYMBOL(drm_legacy_addbufs_pci);
 
-static int drm_addbufs_sg(struct drm_device * dev, struct drm_buf_desc * request)
+static int drm_legacy_addbufs_sg(struct drm_device *dev,
+				 struct drm_buf_desc *request)
 {
 	struct drm_device_dma *dma = dev->dma;
 	struct drm_buf_entry *entry;
@@ -1135,8 +1133,8 @@ static int drm_addbufs_sg(struct drm_device * dev, struct drm_buf_desc * request
  * addbufs_sg() or addbufs_pci() for AGP, scatter-gather or consistent
  * PCI memory respectively.
  */
-int drm_addbufs(struct drm_device *dev, void *data,
-		struct drm_file *file_priv)
+int drm_legacy_addbufs(struct drm_device *dev, void *data,
+		       struct drm_file *file_priv)
 {
 	struct drm_buf_desc *request = data;
 	int ret;
@@ -1149,15 +1147,15 @@ int drm_addbufs(struct drm_device *dev, void *data,
 
 #if __OS_HAS_AGP
 	if (request->flags & _DRM_AGP_BUFFER)
-		ret = drm_addbufs_agp(dev, request);
+		ret = drm_legacy_addbufs_agp(dev, request);
 	else
 #endif
 	if (request->flags & _DRM_SG_BUFFER)
-		ret = drm_addbufs_sg(dev, request);
+		ret = drm_legacy_addbufs_sg(dev, request);
 	else if (request->flags & _DRM_FB_BUFFER)
 		ret = -EINVAL;
 	else
-		ret = drm_addbufs_pci(dev, request);
+		ret = drm_legacy_addbufs_pci(dev, request);
 
 	return ret;
 }
@@ -1179,8 +1177,8 @@ int drm_addbufs(struct drm_device *dev, void *data,
  * lock, preventing of allocating more buffers after this call. Information
  * about each requested buffer is then copied into user space.
  */
-int drm_infobufs(struct drm_device *dev, void *data,
-		 struct drm_file *file_priv)
+int drm_legacy_infobufs(struct drm_device *dev, void *data,
+			struct drm_file *file_priv)
 {
 	struct drm_device_dma *dma = dev->dma;
 	struct drm_buf_info *request = data;
@@ -1260,8 +1258,8 @@ int drm_infobufs(struct drm_device *dev, void *data,
  *
  * \note This ioctl is deprecated and mostly never used.
  */
-int drm_markbufs(struct drm_device *dev, void *data,
-		 struct drm_file *file_priv)
+int drm_legacy_markbufs(struct drm_device *dev, void *data,
+			struct drm_file *file_priv)
 {
 	struct drm_device_dma *dma = dev->dma;
 	struct drm_buf_desc *request = data;
@@ -1307,8 +1305,8 @@ int drm_markbufs(struct drm_device *dev, void *data,
  * Calls free_buffer() for each used buffer.
  * This function is primarily used for debugging.
  */
-int drm_freebufs(struct drm_device *dev, void *data,
-		 struct drm_file *file_priv)
+int drm_legacy_freebufs(struct drm_device *dev, void *data,
+			struct drm_file *file_priv)
 {
 	struct drm_device_dma *dma = dev->dma;
 	struct drm_buf_free *request = data;
@@ -1340,7 +1338,7 @@ int drm_freebufs(struct drm_device *dev, void *data,
 				  task_pid_nr(current));
 			return -EINVAL;
 		}
-		drm_free_buffer(dev, buf);
+		drm_legacy_free_buffer(dev, buf);
 	}
 
 	return 0;
@@ -1360,8 +1358,8 @@ int drm_freebufs(struct drm_device *dev, void *data,
  * offset equal to 0, which drm_mmap() interpretes as PCI buffers and calls
  * drm_mmap_dma().
  */
-int drm_mapbufs(struct drm_device *dev, void *data,
-	        struct drm_file *file_priv)
+int drm_legacy_mapbufs(struct drm_device *dev, void *data,
+		       struct drm_file *file_priv)
 {
 	struct drm_device_dma *dma = dev->dma;
 	int retcode = 0;
@@ -1448,7 +1446,7 @@ int drm_mapbufs(struct drm_device *dev, void *data,
 	return retcode;
 }
 
-int drm_dma_ioctl(struct drm_device *dev, void *data,
+int drm_legacy_dma_ioctl(struct drm_device *dev, void *data,
 		  struct drm_file *file_priv)
 {
 	if (drm_core_check_feature(dev, DRIVER_MODESET))
@@ -1460,7 +1458,7 @@ int drm_dma_ioctl(struct drm_device *dev, void *data,
 		return -EINVAL;
 }
 
-struct drm_local_map *drm_getsarea(struct drm_device *dev)
+struct drm_local_map *drm_legacy_getsarea(struct drm_device *dev)
 {
 	struct drm_map_list *entry;
 
@@ -1472,4 +1470,4 @@ struct drm_local_map *drm_getsarea(struct drm_device *dev)
 	}
 	return NULL;
 }
-EXPORT_SYMBOL(drm_getsarea);
+EXPORT_SYMBOL(drm_legacy_getsarea);
diff --git a/drivers/gpu/drm/drm_crtc.c b/drivers/gpu/drm/drm_crtc.c
index 90e773019eac..5213da499d39 100644
--- a/drivers/gpu/drm/drm_crtc.c
+++ b/drivers/gpu/drm/drm_crtc.c
@@ -40,106 +40,12 @@
 #include <drm/drm_modeset_lock.h>
 
 #include "drm_crtc_internal.h"
+#include "drm_internal.h"
 
 static struct drm_framebuffer *add_framebuffer_internal(struct drm_device *dev,
 							struct drm_mode_fb_cmd2 *r,
 							struct drm_file *file_priv);
 
-/**
- * drm_modeset_lock_all - take all modeset locks
- * @dev: drm device
- *
- * This function takes all modeset locks, suitable where a more fine-grained
- * scheme isn't (yet) implemented. Locks must be dropped with
- * drm_modeset_unlock_all.
- */
-void drm_modeset_lock_all(struct drm_device *dev)
-{
-	struct drm_mode_config *config = &dev->mode_config;
-	struct drm_modeset_acquire_ctx *ctx;
-	int ret;
-
-	ctx = kzalloc(sizeof(*ctx), GFP_KERNEL);
-	if (WARN_ON(!ctx))
-		return;
-
-	mutex_lock(&config->mutex);
-
-	drm_modeset_acquire_init(ctx, 0);
-
-retry:
-	ret = drm_modeset_lock(&config->connection_mutex, ctx);
-	if (ret)
-		goto fail;
-	ret = drm_modeset_lock_all_crtcs(dev, ctx);
-	if (ret)
-		goto fail;
-
-	WARN_ON(config->acquire_ctx);
-
-	/* now we hold the locks, so now that it is safe, stash the
-	 * ctx for drm_modeset_unlock_all():
-	 */
-	config->acquire_ctx = ctx;
-
-	drm_warn_on_modeset_not_all_locked(dev);
-
-	return;
-
-fail:
-	if (ret == -EDEADLK) {
-		drm_modeset_backoff(ctx);
-		goto retry;
-	}
-}
-EXPORT_SYMBOL(drm_modeset_lock_all);
-
-/**
- * drm_modeset_unlock_all - drop all modeset locks
- * @dev: device
- *
- * This function drop all modeset locks taken by drm_modeset_lock_all.
- */
-void drm_modeset_unlock_all(struct drm_device *dev)
-{
-	struct drm_mode_config *config = &dev->mode_config;
-	struct drm_modeset_acquire_ctx *ctx = config->acquire_ctx;
-
-	if (WARN_ON(!ctx))
-		return;
-
-	config->acquire_ctx = NULL;
-	drm_modeset_drop_locks(ctx);
-	drm_modeset_acquire_fini(ctx);
-
-	kfree(ctx);
-
-	mutex_unlock(&dev->mode_config.mutex);
-}
-EXPORT_SYMBOL(drm_modeset_unlock_all);
-
-/**
- * drm_warn_on_modeset_not_all_locked - check that all modeset locks are locked
- * @dev: device
- *
- * Useful as a debug assert.
- */
-void drm_warn_on_modeset_not_all_locked(struct drm_device *dev)
-{
-	struct drm_crtc *crtc;
-
-	/* Locking is currently fubar in the panic handler. */
-	if (oops_in_progress)
-		return;
-
-	list_for_each_entry(crtc, &dev->mode_config.crtc_list, head)
-		WARN_ON(!drm_modeset_is_locked(&crtc->mutex));
-
-	WARN_ON(!drm_modeset_is_locked(&dev->mode_config.connection_mutex));
-	WARN_ON(!mutex_is_locked(&dev->mode_config.mutex));
-}
-EXPORT_SYMBOL(drm_warn_on_modeset_not_all_locked);
-
 /* Avoid boilerplate.  I'm tired of typing. */
 #define DRM_ENUM_NAME_FN(fnname, list)				\
 	const char *fnname(int val)				\
@@ -515,9 +421,6 @@ int drm_framebuffer_init(struct drm_device *dev, struct drm_framebuffer *fb,
 	if (ret)
 		goto out;
 
-	/* Grab the idr reference. */
-	drm_framebuffer_reference(fb);
-
 	dev->mode_config.num_fb++;
 	list_add(&fb->head, &dev->mode_config.fb_list);
 out:
@@ -527,10 +430,34 @@ out:
 }
 EXPORT_SYMBOL(drm_framebuffer_init);
 
+/* dev->mode_config.fb_lock must be held! */
+static void __drm_framebuffer_unregister(struct drm_device *dev,
+					 struct drm_framebuffer *fb)
+{
+	mutex_lock(&dev->mode_config.idr_mutex);
+	idr_remove(&dev->mode_config.crtc_idr, fb->base.id);
+	mutex_unlock(&dev->mode_config.idr_mutex);
+
+	fb->base.id = 0;
+}
+
 static void drm_framebuffer_free(struct kref *kref)
 {
 	struct drm_framebuffer *fb =
 			container_of(kref, struct drm_framebuffer, refcount);
+	struct drm_device *dev = fb->dev;
+
+	/*
+	 * The lookup idr holds a weak reference, which has not necessarily been
+	 * removed at this point. Check for that.
+	 */
+	mutex_lock(&dev->mode_config.fb_lock);
+	if (fb->base.id) {
+		/* Mark fb as reaped and drop idr ref. */
+		__drm_framebuffer_unregister(dev, fb);
+	}
+	mutex_unlock(&dev->mode_config.fb_lock);
+
 	fb->funcs->destroy(fb);
 }
 
@@ -567,8 +494,10 @@ struct drm_framebuffer *drm_framebuffer_lookup(struct drm_device *dev,
 
 	mutex_lock(&dev->mode_config.fb_lock);
 	fb = __drm_framebuffer_lookup(dev, id);
-	if (fb)
-		drm_framebuffer_reference(fb);
+	if (fb) {
+		if (!kref_get_unless_zero(&fb->refcount))
+			fb = NULL;
+	}
 	mutex_unlock(&dev->mode_config.fb_lock);
 
 	return fb;
@@ -612,19 +541,6 @@ static void __drm_framebuffer_unreference(struct drm_framebuffer *fb)
 	kref_put(&fb->refcount, drm_framebuffer_free_bug);
 }
 
-/* dev->mode_config.fb_lock must be held! */
-static void __drm_framebuffer_unregister(struct drm_device *dev,
-					 struct drm_framebuffer *fb)
-{
-	mutex_lock(&dev->mode_config.idr_mutex);
-	idr_remove(&dev->mode_config.crtc_idr, fb->base.id);
-	mutex_unlock(&dev->mode_config.idr_mutex);
-
-	fb->base.id = 0;
-
-	__drm_framebuffer_unreference(fb);
-}
-
 /**
  * drm_framebuffer_unregister_private - unregister a private fb from the lookup idr
  * @fb: fb to unregister
@@ -764,14 +680,10 @@ int drm_crtc_init_with_planes(struct drm_device *dev, struct drm_crtc *crtc,
 	crtc->funcs = funcs;
 	crtc->invert_dimensions = false;
 
-	drm_modeset_lock_all(dev);
 	drm_modeset_lock_init(&crtc->mutex);
-	/* dropped by _unlock_all(): */
-	drm_modeset_lock(&crtc->mutex, config->acquire_ctx);
-
 	ret = drm_mode_object_get(dev, &crtc->base, DRM_MODE_OBJECT_CRTC);
 	if (ret)
-		goto out;
+		return ret;
 
 	crtc->base.properties = &crtc->properties;
 
@@ -785,10 +697,7 @@ int drm_crtc_init_with_planes(struct drm_device *dev, struct drm_crtc *crtc,
 	if (cursor)
 		cursor->possible_crtcs = 1 << drm_crtc_index(crtc);
 
- out:
-	drm_modeset_unlock_all(dev);
-
-	return ret;
+	return 0;
 }
 EXPORT_SYMBOL(drm_crtc_init_with_planes);
 
@@ -812,6 +721,12 @@ void drm_crtc_cleanup(struct drm_crtc *crtc)
 	drm_mode_object_put(dev, &crtc->base);
 	list_del(&crtc->head);
 	dev->mode_config.num_crtc--;
+
+	WARN_ON(crtc->state && !crtc->funcs->atomic_destroy_state);
+	if (crtc->state && crtc->funcs->atomic_destroy_state)
+		crtc->funcs->atomic_destroy_state(crtc, crtc->state);
+
+	memset(crtc, 0, sizeof(*crtc));
 }
 EXPORT_SYMBOL(drm_crtc_cleanup);
 
@@ -853,6 +768,58 @@ static void drm_mode_remove(struct drm_connector *connector,
 }
 
 /**
+ * drm_connector_get_cmdline_mode - reads the user's cmdline mode
+ * @connector: connector to quwery
+ *
+ * The kernel supports per-connector configration of its consoles through
+ * use of the video= parameter. This function parses that option and
+ * extracts the user's specified mode (or enable/disable status) for a
+ * particular connector. This is typically only used during the early fbdev
+ * setup.
+ */
+static void drm_connector_get_cmdline_mode(struct drm_connector *connector)
+{
+	struct drm_cmdline_mode *mode = &connector->cmdline_mode;
+	char *option = NULL;
+
+	if (fb_get_options(connector->name, &option))
+		return;
+
+	if (!drm_mode_parse_command_line_for_connector(option,
+						       connector,
+						       mode))
+		return;
+
+	if (mode->force) {
+		const char *s;
+
+		switch (mode->force) {
+		case DRM_FORCE_OFF:
+			s = "OFF";
+			break;
+		case DRM_FORCE_ON_DIGITAL:
+			s = "ON - dig";
+			break;
+		default:
+		case DRM_FORCE_ON:
+			s = "ON";
+			break;
+		}
+
+		DRM_INFO("forcing %s connector %s\n", connector->name, s);
+		connector->force = mode->force;
+	}
+
+	DRM_DEBUG_KMS("cmdline mode for connector %s %dx%d@%dHz%s%s%s\n",
+		      connector->name,
+		      mode->xres, mode->yres,
+		      mode->refresh_specified ? mode->refresh : 60,
+		      mode->rb ? " reduced blanking" : "",
+		      mode->margins ? " with margins" : "",
+		      mode->interlace ?  " interlaced" : "");
+}
+
+/**
  * drm_connector_init - Init a preallocated connector
  * @dev: DRM device
  * @connector: the connector to init
@@ -904,6 +871,10 @@ int drm_connector_init(struct drm_device *dev,
 	connector->edid_blob_ptr = NULL;
 	connector->status = connector_status_unknown;
 
+	drm_connector_get_cmdline_mode(connector);
+
+	/* We should add connectors at the end to avoid upsetting the connector
+	 * index too much. */
 	list_add_tail(&connector->head, &dev->mode_config.connector_list);
 	dev->mode_config.num_connector++;
 
@@ -939,6 +910,11 @@ void drm_connector_cleanup(struct drm_connector *connector)
 	struct drm_device *dev = connector->dev;
 	struct drm_display_mode *mode, *t;
 
+	if (connector->tile_group) {
+		drm_mode_put_tile_group(dev, connector->tile_group);
+		connector->tile_group = NULL;
+	}
+
 	list_for_each_entry_safe(mode, t, &connector->probed_modes, head)
 		drm_mode_remove(connector, mode);
 
@@ -953,10 +929,43 @@ void drm_connector_cleanup(struct drm_connector *connector)
 	connector->name = NULL;
 	list_del(&connector->head);
 	dev->mode_config.num_connector--;
+
+	WARN_ON(connector->state && !connector->funcs->atomic_destroy_state);
+	if (connector->state && connector->funcs->atomic_destroy_state)
+		connector->funcs->atomic_destroy_state(connector,
+						       connector->state);
+
+	memset(connector, 0, sizeof(*connector));
 }
 EXPORT_SYMBOL(drm_connector_cleanup);
 
 /**
+ * drm_connector_index - find the index of a registered connector
+ * @connector: connector to find index for
+ *
+ * Given a registered connector, return the index of that connector within a DRM
+ * device's list of connectors.
+ */
+unsigned int drm_connector_index(struct drm_connector *connector)
+{
+	unsigned int index = 0;
+	struct drm_connector *tmp;
+	struct drm_mode_config *config = &connector->dev->mode_config;
+
+	WARN_ON(!drm_modeset_is_locked(&config->connection_mutex));
+
+	list_for_each_entry(tmp, &connector->dev->mode_config.connector_list, head) {
+		if (tmp == connector)
+			return index;
+
+		index++;
+	}
+
+	BUG();
+}
+EXPORT_SYMBOL(drm_connector_index);
+
+/**
  * drm_connector_register - register a connector
  * @connector: the connector to register
  *
@@ -1068,6 +1077,8 @@ void drm_bridge_cleanup(struct drm_bridge *bridge)
 	list_del(&bridge->head);
 	dev->mode_config.num_bridge--;
 	drm_modeset_unlock_all(dev);
+
+	memset(bridge, 0, sizeof(*bridge));
 }
 EXPORT_SYMBOL(drm_bridge_cleanup);
 
@@ -1134,10 +1145,11 @@ void drm_encoder_cleanup(struct drm_encoder *encoder)
 	drm_modeset_lock_all(dev);
 	drm_mode_object_put(dev, &encoder->base);
 	kfree(encoder->name);
-	encoder->name = NULL;
 	list_del(&encoder->head);
 	dev->mode_config.num_encoder--;
 	drm_modeset_unlock_all(dev);
+
+	memset(encoder, 0, sizeof(*encoder));
 }
 EXPORT_SYMBOL(drm_encoder_cleanup);
 
@@ -1164,11 +1176,11 @@ int drm_universal_plane_init(struct drm_device *dev, struct drm_plane *plane,
 {
 	int ret;
 
-	drm_modeset_lock_all(dev);
-
 	ret = drm_mode_object_get(dev, &plane->base, DRM_MODE_OBJECT_PLANE);
 	if (ret)
-		goto out;
+		return ret;
+
+	drm_modeset_lock_init(&plane->mutex);
 
 	plane->base.properties = &plane->properties;
 	plane->dev = dev;
@@ -1178,8 +1190,7 @@ int drm_universal_plane_init(struct drm_device *dev, struct drm_plane *plane,
 	if (!plane->format_types) {
 		DRM_DEBUG_KMS("out of memory when allocating plane\n");
 		drm_mode_object_put(dev, &plane->base);
-		ret = -ENOMEM;
-		goto out;
+		return -ENOMEM;
 	}
 
 	memcpy(plane->format_types, formats, format_count * sizeof(uint32_t));
@@ -1196,10 +1207,7 @@ int drm_universal_plane_init(struct drm_device *dev, struct drm_plane *plane,
 				   dev->mode_config.plane_type_property,
 				   plane->type);
 
- out:
-	drm_modeset_unlock_all(dev);
-
-	return ret;
+	return 0;
 }
 EXPORT_SYMBOL(drm_universal_plane_init);
 
@@ -1257,10 +1265,39 @@ void drm_plane_cleanup(struct drm_plane *plane)
 	if (plane->type == DRM_PLANE_TYPE_OVERLAY)
 		dev->mode_config.num_overlay_plane--;
 	drm_modeset_unlock_all(dev);
+
+	WARN_ON(plane->state && !plane->funcs->atomic_destroy_state);
+	if (plane->state && plane->funcs->atomic_destroy_state)
+		plane->funcs->atomic_destroy_state(plane, plane->state);
+
+	memset(plane, 0, sizeof(*plane));
 }
 EXPORT_SYMBOL(drm_plane_cleanup);
 
 /**
+ * drm_plane_index - find the index of a registered plane
+ * @plane: plane to find index for
+ *
+ * Given a registered plane, return the index of that CRTC within a DRM
+ * device's list of planes.
+ */
+unsigned int drm_plane_index(struct drm_plane *plane)
+{
+	unsigned int index = 0;
+	struct drm_plane *tmp;
+
+	list_for_each_entry(tmp, &plane->dev->mode_config.plane_list, head) {
+		if (tmp == plane)
+			return index;
+
+		index++;
+	}
+
+	BUG();
+}
+EXPORT_SYMBOL(drm_plane_index);
+
+/**
  * drm_plane_force_disable - Forcibly disable a plane
  * @plane: plane to disable
  *
@@ -1271,19 +1308,21 @@ EXPORT_SYMBOL(drm_plane_cleanup);
  */
 void drm_plane_force_disable(struct drm_plane *plane)
 {
-	struct drm_framebuffer *old_fb = plane->fb;
 	int ret;
 
-	if (!old_fb)
+	if (!plane->fb)
 		return;
 
+	plane->old_fb = plane->fb;
 	ret = plane->funcs->disable_plane(plane);
 	if (ret) {
 		DRM_ERROR("failed to disable plane with busy fb\n");
+		plane->old_fb = NULL;
 		return;
 	}
 	/* disconnect the plane from the fb and crtc: */
-	__drm_framebuffer_unreference(old_fb);
+	__drm_framebuffer_unreference(plane->old_fb);
+	plane->old_fb = NULL;
 	plane->fb = NULL;
 	plane->crtc = NULL;
 }
@@ -1314,6 +1353,11 @@ static int drm_mode_create_standard_connector_properties(struct drm_device *dev)
 				       "PATH", 0);
 	dev->mode_config.path_property = dev_path;
 
+	dev->mode_config.tile_property = drm_property_create(dev,
+							     DRM_MODE_PROP_BLOB |
+							     DRM_MODE_PROP_IMMUTABLE,
+							     "TILE", 0);
+
 	return 0;
 }
 
@@ -1374,12 +1418,13 @@ EXPORT_SYMBOL(drm_mode_create_dvi_i_properties);
  * responsible for allocating a list of format names and passing them to
  * this routine.
  */
-int drm_mode_create_tv_properties(struct drm_device *dev, int num_modes,
+int drm_mode_create_tv_properties(struct drm_device *dev,
+				  unsigned int num_modes,
 				  char *modes[])
 {
 	struct drm_property *tv_selector;
 	struct drm_property *tv_subconnector;
-	int i;
+	unsigned int i;
 
 	if (dev->mode_config.tv_select_subconnector_property)
 		return 0;
@@ -1477,7 +1522,7 @@ EXPORT_SYMBOL(drm_mode_create_scaling_mode_property);
  * connectors.
  *
  * Returns:
- * Zero on success, errno on failure.
+ * Zero on success, negative errno on failure.
  */
 int drm_mode_create_aspect_ratio_property(struct drm_device *dev)
 {
@@ -1521,6 +1566,30 @@ int drm_mode_create_dirty_info_property(struct drm_device *dev)
 }
 EXPORT_SYMBOL(drm_mode_create_dirty_info_property);
 
+/**
+ * drm_mode_create_suggested_offset_properties - create suggests offset properties
+ * @dev: DRM device
+ *
+ * Create the the suggested x/y offset property for connectors.
+ */
+int drm_mode_create_suggested_offset_properties(struct drm_device *dev)
+{
+	if (dev->mode_config.suggested_x_property && dev->mode_config.suggested_y_property)
+		return 0;
+
+	dev->mode_config.suggested_x_property =
+		drm_property_create_range(dev, DRM_MODE_PROP_IMMUTABLE, "suggested X", 0, 0xffffffff);
+
+	dev->mode_config.suggested_y_property =
+		drm_property_create_range(dev, DRM_MODE_PROP_IMMUTABLE, "suggested Y", 0, 0xffffffff);
+
+	if (dev->mode_config.suggested_x_property == NULL ||
+	    dev->mode_config.suggested_y_property == NULL)
+		return -ENOMEM;
+	return 0;
+}
+EXPORT_SYMBOL(drm_mode_create_suggested_offset_properties);
+
 static int drm_mode_group_init(struct drm_device *dev, struct drm_mode_group *group)
 {
 	uint32_t total_objects = 0;
@@ -1637,7 +1706,7 @@ static void drm_crtc_convert_to_umode(struct drm_mode_modeinfo *out,
  * the caller.
  *
  * Returns:
- * Zero on success, errno on failure.
+ * Zero on success, negative errno on failure.
  */
 static int drm_crtc_convert_umode(struct drm_display_mode *out,
 				  const struct drm_mode_modeinfo *in)
@@ -1680,7 +1749,7 @@ static int drm_crtc_convert_umode(struct drm_display_mode *out,
  * Called by the user via ioctl.
  *
  * Returns:
- * Zero on success, errno on failure.
+ * Zero on success, negative errno on failure.
  */
 int drm_mode_getresources(struct drm_device *dev, void *data,
 			  struct drm_file *file_priv)
@@ -1731,7 +1800,9 @@ int drm_mode_getresources(struct drm_device *dev, void *data,
 	card_res->count_fbs = fb_count;
 	mutex_unlock(&file_priv->fbs_lock);
 
-	drm_modeset_lock_all(dev);
+	/* mode_config.mutex protects the connector list against e.g. DP MST
+	 * connector hot-adding. CRTC/Plane lists are invariant. */
+	mutex_lock(&dev->mode_config.mutex);
 	if (!drm_is_primary_client(file_priv)) {
 
 		mode_group = NULL;
@@ -1851,7 +1922,7 @@ int drm_mode_getresources(struct drm_device *dev, void *data,
 		  card_res->count_connectors, card_res->count_encoders);
 
 out:
-	drm_modeset_unlock_all(dev);
+	mutex_unlock(&dev->mode_config.mutex);
 	return ret;
 }
 
@@ -1866,26 +1937,22 @@ out:
  * Called by the user via ioctl.
  *
  * Returns:
- * Zero on success, errno on failure.
+ * Zero on success, negative errno on failure.
  */
 int drm_mode_getcrtc(struct drm_device *dev,
 		     void *data, struct drm_file *file_priv)
 {
 	struct drm_mode_crtc *crtc_resp = data;
 	struct drm_crtc *crtc;
-	int ret = 0;
 
 	if (!drm_core_check_feature(dev, DRIVER_MODESET))
 		return -EINVAL;
 
-	drm_modeset_lock_all(dev);
-
 	crtc = drm_crtc_find(dev, crtc_resp->crtc_id);
-	if (!crtc) {
-		ret = -ENOENT;
-		goto out;
-	}
+	if (!crtc)
+		return -ENOENT;
 
+	drm_modeset_lock_crtc(crtc, crtc->primary);
 	crtc_resp->x = crtc->x;
 	crtc_resp->y = crtc->y;
 	crtc_resp->gamma_size = crtc->gamma_size;
@@ -1902,10 +1969,9 @@ int drm_mode_getcrtc(struct drm_device *dev,
 	} else {
 		crtc_resp->mode_valid = 0;
 	}
+	drm_modeset_unlock_crtc(crtc);
 
-out:
-	drm_modeset_unlock_all(dev);
-	return ret;
+	return 0;
 }
 
 static bool drm_mode_expose_to_userspace(const struct drm_display_mode *mode,
@@ -1921,6 +1987,15 @@ static bool drm_mode_expose_to_userspace(const struct drm_display_mode *mode,
 	return true;
 }
 
+static struct drm_encoder *drm_connector_get_encoder(struct drm_connector *connector)
+{
+	/* For atomic drivers only state objects are synchronously updated and
+	 * protected by modeset locks, so check those first. */
+	if (connector->state)
+		return connector->state->best_encoder;
+	return connector->encoder;
+}
+
 /**
  * drm_mode_getconnector - get connector configuration
  * @dev: drm device for the ioctl
@@ -1932,13 +2007,14 @@ static bool drm_mode_expose_to_userspace(const struct drm_display_mode *mode,
  * Called by the user via ioctl.
  *
  * Returns:
- * Zero on success, errno on failure.
+ * Zero on success, negative errno on failure.
  */
 int drm_mode_getconnector(struct drm_device *dev, void *data,
 			  struct drm_file *file_priv)
 {
 	struct drm_mode_get_connector *out_resp = data;
 	struct drm_connector *connector;
+	struct drm_encoder *encoder;
 	struct drm_display_mode *mode;
 	int mode_count = 0;
 	int props_count = 0;
@@ -1994,8 +2070,10 @@ int drm_mode_getconnector(struct drm_device *dev, void *data,
 	out_resp->subpixel = connector->display_info.subpixel_order;
 	out_resp->connection = connector->status;
 	drm_modeset_lock(&dev->mode_config.connection_mutex, NULL);
-	if (connector->encoder)
-		out_resp->encoder_id = connector->encoder->base.id;
+
+	encoder = drm_connector_get_encoder(connector);
+	if (encoder)
+		out_resp->encoder_id = encoder->base.id;
 	else
 		out_resp->encoder_id = 0;
 	drm_modeset_unlock(&dev->mode_config.connection_mutex);
@@ -2065,6 +2143,33 @@ out:
 	return ret;
 }
 
+static struct drm_crtc *drm_encoder_get_crtc(struct drm_encoder *encoder)
+{
+	struct drm_connector *connector;
+	struct drm_device *dev = encoder->dev;
+	bool uses_atomic = false;
+
+	/* For atomic drivers only state objects are synchronously updated and
+	 * protected by modeset locks, so check those first. */
+	list_for_each_entry(connector, &dev->mode_config.connector_list, head) {
+		if (!connector->state)
+			continue;
+
+		uses_atomic = true;
+
+		if (connector->state->best_encoder != encoder)
+			continue;
+
+		return connector->state->crtc;
+	}
+
+	/* Don't return stale data (e.g. pending async disable). */
+	if (uses_atomic)
+		return NULL;
+
+	return encoder->crtc;
+}
+
 /**
  * drm_mode_getencoder - get encoder configuration
  * @dev: drm device for the ioctl
@@ -2076,37 +2181,38 @@ out:
  * Called by the user via ioctl.
  *
  * Returns:
- * Zero on success, errno on failure.
+ * Zero on success, negative errno on failure.
  */
 int drm_mode_getencoder(struct drm_device *dev, void *data,
 			struct drm_file *file_priv)
 {
 	struct drm_mode_get_encoder *enc_resp = data;
 	struct drm_encoder *encoder;
-	int ret = 0;
+	struct drm_crtc *crtc;
 
 	if (!drm_core_check_feature(dev, DRIVER_MODESET))
 		return -EINVAL;
 
-	drm_modeset_lock_all(dev);
 	encoder = drm_encoder_find(dev, enc_resp->encoder_id);
-	if (!encoder) {
-		ret = -ENOENT;
-		goto out;
-	}
+	if (!encoder)
+		return -ENOENT;
 
-	if (encoder->crtc)
+	drm_modeset_lock(&dev->mode_config.connection_mutex, NULL);
+	crtc = drm_encoder_get_crtc(encoder);
+	if (crtc)
+		enc_resp->crtc_id = crtc->base.id;
+	else if (encoder->crtc)
 		enc_resp->crtc_id = encoder->crtc->base.id;
 	else
 		enc_resp->crtc_id = 0;
+	drm_modeset_unlock(&dev->mode_config.connection_mutex);
+
 	enc_resp->encoder_type = encoder->encoder_type;
 	enc_resp->encoder_id = encoder->base.id;
 	enc_resp->possible_crtcs = encoder->possible_crtcs;
 	enc_resp->possible_clones = encoder->possible_clones;
 
-out:
-	drm_modeset_unlock_all(dev);
-	return ret;
+	return 0;
 }
 
 /**
@@ -2120,7 +2226,7 @@ out:
  * Called by the user via ioctl.
  *
  * Returns:
- * Zero on success, errno on failure.
+ * Zero on success, negative errno on failure.
  */
 int drm_mode_getplane_res(struct drm_device *dev, void *data,
 			  struct drm_file *file_priv)
@@ -2129,13 +2235,12 @@ int drm_mode_getplane_res(struct drm_device *dev, void *data,
 	struct drm_mode_config *config;
 	struct drm_plane *plane;
 	uint32_t __user *plane_ptr;
-	int copied = 0, ret = 0;
+	int copied = 0;
 	unsigned num_planes;
 
 	if (!drm_core_check_feature(dev, DRIVER_MODESET))
 		return -EINVAL;
 
-	drm_modeset_lock_all(dev);
 	config = &dev->mode_config;
 
 	if (file_priv->universal_planes)
@@ -2151,6 +2256,7 @@ int drm_mode_getplane_res(struct drm_device *dev, void *data,
 	    (plane_resp->count_planes >= num_planes)) {
 		plane_ptr = (uint32_t __user *)(unsigned long)plane_resp->plane_id_ptr;
 
+		/* Plane lists are invariant, no locking needed. */
 		list_for_each_entry(plane, &config->plane_list, head) {
 			/*
 			 * Unless userspace set the 'universal planes'
@@ -2160,18 +2266,14 @@ int drm_mode_getplane_res(struct drm_device *dev, void *data,
 			    !file_priv->universal_planes)
 				continue;
 
-			if (put_user(plane->base.id, plane_ptr + copied)) {
-				ret = -EFAULT;
-				goto out;
-			}
+			if (put_user(plane->base.id, plane_ptr + copied))
+				return -EFAULT;
 			copied++;
 		}
 	}
 	plane_resp->count_planes = num_planes;
 
-out:
-	drm_modeset_unlock_all(dev);
-	return ret;
+	return 0;
 }
 
 /**
@@ -2185,7 +2287,7 @@ out:
  * Called by the user via ioctl.
  *
  * Returns:
- * Zero on success, errno on failure.
+ * Zero on success, negative errno on failure.
  */
 int drm_mode_getplane(struct drm_device *dev, void *data,
 		      struct drm_file *file_priv)
@@ -2193,18 +2295,15 @@ int drm_mode_getplane(struct drm_device *dev, void *data,
 	struct drm_mode_get_plane *plane_resp = data;
 	struct drm_plane *plane;
 	uint32_t __user *format_ptr;
-	int ret = 0;
 
 	if (!drm_core_check_feature(dev, DRIVER_MODESET))
 		return -EINVAL;
 
-	drm_modeset_lock_all(dev);
 	plane = drm_plane_find(dev, plane_resp->plane_id);
-	if (!plane) {
-		ret = -ENOENT;
-		goto out;
-	}
+	if (!plane)
+		return -ENOENT;
 
+	drm_modeset_lock(&plane->mutex, NULL);
 	if (plane->crtc)
 		plane_resp->crtc_id = plane->crtc->base.id;
 	else
@@ -2214,6 +2313,7 @@ int drm_mode_getplane(struct drm_device *dev, void *data,
 		plane_resp->fb_id = plane->fb->base.id;
 	else
 		plane_resp->fb_id = 0;
+	drm_modeset_unlock(&plane->mutex);
 
 	plane_resp->plane_id = plane->base.id;
 	plane_resp->possible_crtcs = plane->possible_crtcs;
@@ -2229,15 +2329,12 @@ int drm_mode_getplane(struct drm_device *dev, void *data,
 		if (copy_to_user(format_ptr,
 				 plane->format_types,
 				 sizeof(uint32_t) * plane->format_count)) {
-			ret = -EFAULT;
-			goto out;
+			return -EFAULT;
 		}
 	}
 	plane_resp->count_format_types = plane->format_count;
 
-out:
-	drm_modeset_unlock_all(dev);
-	return ret;
+	return 0;
 }
 
 /*
@@ -2249,33 +2346,29 @@ out:
  *
  * src_{x,y,w,h} are provided in 16.16 fixed point format
  */
-static int setplane_internal(struct drm_plane *plane,
-			     struct drm_crtc *crtc,
-			     struct drm_framebuffer *fb,
-			     int32_t crtc_x, int32_t crtc_y,
-			     uint32_t crtc_w, uint32_t crtc_h,
-			     /* src_{x,y,w,h} values are 16.16 fixed point */
-			     uint32_t src_x, uint32_t src_y,
-			     uint32_t src_w, uint32_t src_h)
+static int __setplane_internal(struct drm_plane *plane,
+			       struct drm_crtc *crtc,
+			       struct drm_framebuffer *fb,
+			       int32_t crtc_x, int32_t crtc_y,
+			       uint32_t crtc_w, uint32_t crtc_h,
+			       /* src_{x,y,w,h} values are 16.16 fixed point */
+			       uint32_t src_x, uint32_t src_y,
+			       uint32_t src_w, uint32_t src_h)
 {
-	struct drm_device *dev = plane->dev;
-	struct drm_framebuffer *old_fb = NULL;
 	int ret = 0;
 	unsigned int fb_width, fb_height;
-	int i;
+	unsigned int i;
 
 	/* No fb means shut it down */
 	if (!fb) {
-		drm_modeset_lock_all(dev);
-		old_fb = plane->fb;
+		plane->old_fb = plane->fb;
 		ret = plane->funcs->disable_plane(plane);
 		if (!ret) {
 			plane->crtc = NULL;
 			plane->fb = NULL;
 		} else {
-			old_fb = NULL;
+			plane->old_fb = NULL;
 		}
-		drm_modeset_unlock_all(dev);
 		goto out;
 	}
 
@@ -2315,8 +2408,7 @@ static int setplane_internal(struct drm_plane *plane,
 		goto out;
 	}
 
-	drm_modeset_lock_all(dev);
-	old_fb = plane->fb;
+	plane->old_fb = plane->fb;
 	ret = plane->funcs->update_plane(plane, crtc, fb,
 					 crtc_x, crtc_y, crtc_w, crtc_h,
 					 src_x, src_y, src_w, src_h);
@@ -2325,18 +2417,37 @@ static int setplane_internal(struct drm_plane *plane,
 		plane->fb = fb;
 		fb = NULL;
 	} else {
-		old_fb = NULL;
+		plane->old_fb = NULL;
 	}
-	drm_modeset_unlock_all(dev);
 
 out:
 	if (fb)
 		drm_framebuffer_unreference(fb);
-	if (old_fb)
-		drm_framebuffer_unreference(old_fb);
+	if (plane->old_fb)
+		drm_framebuffer_unreference(plane->old_fb);
+	plane->old_fb = NULL;
 
 	return ret;
+}
+
+static int setplane_internal(struct drm_plane *plane,
+			     struct drm_crtc *crtc,
+			     struct drm_framebuffer *fb,
+			     int32_t crtc_x, int32_t crtc_y,
+			     uint32_t crtc_w, uint32_t crtc_h,
+			     /* src_{x,y,w,h} values are 16.16 fixed point */
+			     uint32_t src_x, uint32_t src_y,
+			     uint32_t src_w, uint32_t src_h)
+{
+	int ret;
 
+	drm_modeset_lock_all(plane->dev);
+	ret = __setplane_internal(plane, crtc, fb,
+				  crtc_x, crtc_y, crtc_w, crtc_h,
+				  src_x, src_y, src_w, src_h);
+	drm_modeset_unlock_all(plane->dev);
+
+	return ret;
 }
 
 /**
@@ -2350,13 +2461,12 @@ out:
  * valid crtc).
  *
  * Returns:
- * Zero on success, errno on failure.
+ * Zero on success, negative errno on failure.
  */
 int drm_mode_setplane(struct drm_device *dev, void *data,
 		      struct drm_file *file_priv)
 {
 	struct drm_mode_set_plane *plane_req = data;
-	struct drm_mode_object *obj;
 	struct drm_plane *plane;
 	struct drm_crtc *crtc = NULL;
 	struct drm_framebuffer *fb = NULL;
@@ -2379,14 +2489,12 @@ int drm_mode_setplane(struct drm_device *dev, void *data,
 	 * First, find the plane, crtc, and fb objects.  If not available,
 	 * we don't bother to call the driver.
 	 */
-	obj = drm_mode_object_find(dev, plane_req->plane_id,
-				   DRM_MODE_OBJECT_PLANE);
-	if (!obj) {
+	plane = drm_plane_find(dev, plane_req->plane_id);
+	if (!plane) {
 		DRM_DEBUG_KMS("Unknown plane ID %d\n",
 			      plane_req->plane_id);
 		return -ENOENT;
 	}
-	plane = obj_to_plane(obj);
 
 	if (plane_req->fb_id) {
 		fb = drm_framebuffer_lookup(dev, plane_req->fb_id);
@@ -2396,14 +2504,12 @@ int drm_mode_setplane(struct drm_device *dev, void *data,
 			return -ENOENT;
 		}
 
-		obj = drm_mode_object_find(dev, plane_req->crtc_id,
-					   DRM_MODE_OBJECT_CRTC);
-		if (!obj) {
+		crtc = drm_crtc_find(dev, plane_req->crtc_id);
+		if (!crtc) {
 			DRM_DEBUG_KMS("Unknown crtc ID %d\n",
 				      plane_req->crtc_id);
 			return -ENOENT;
 		}
-		crtc = obj_to_crtc(obj);
 	}
 
 	/*
@@ -2425,7 +2531,7 @@ int drm_mode_setplane(struct drm_device *dev, void *data,
  * interface. The only thing it adds is correct refcounting dance.
  * 
  * Returns:
- * Zero on success, errno on failure.
+ * Zero on success, negative errno on failure.
  */
 int drm_mode_set_config_internal(struct drm_mode_set *set)
 {
@@ -2440,7 +2546,7 @@ int drm_mode_set_config_internal(struct drm_mode_set *set)
 	 * crtcs. Atomic modeset will have saner semantics ...
 	 */
 	list_for_each_entry(tmp, &crtc->dev->mode_config.crtc_list, head)
-		tmp->old_fb = tmp->primary->fb;
+		tmp->primary->old_fb = tmp->primary->fb;
 
 	fb = set->fb;
 
@@ -2453,8 +2559,9 @@ int drm_mode_set_config_internal(struct drm_mode_set *set)
 	list_for_each_entry(tmp, &crtc->dev->mode_config.crtc_list, head) {
 		if (tmp->primary->fb)
 			drm_framebuffer_reference(tmp->primary->fb);
-		if (tmp->old_fb)
-			drm_framebuffer_unreference(tmp->old_fb);
+		if (tmp->primary->old_fb)
+			drm_framebuffer_unreference(tmp->primary->old_fb);
+		tmp->primary->old_fb = NULL;
 	}
 
 	return ret;
@@ -2517,7 +2624,7 @@ EXPORT_SYMBOL(drm_crtc_check_viewport);
  * Called by the user via ioctl.
  *
  * Returns:
- * Zero on success, errno on failure.
+ * Zero on success, negative errno on failure.
  */
 int drm_mode_setcrtc(struct drm_device *dev, void *data,
 		     struct drm_file *file_priv)
@@ -2680,7 +2787,7 @@ out:
  * userspace wants to make use of these capabilities.
  *
  * Returns:
- * Zero on success, errno on failure.
+ * Zero on success, negative errno on failure.
  */
 static int drm_mode_cursor_universal(struct drm_crtc *crtc,
 				     struct drm_mode_cursor2 *req,
@@ -2701,6 +2808,7 @@ static int drm_mode_cursor_universal(struct drm_crtc *crtc,
 	int ret = 0;
 
 	BUG_ON(!crtc->cursor);
+	WARN_ON(crtc->cursor->crtc != crtc && crtc->cursor->crtc != NULL);
 
 	/*
 	 * Obtain fb we'll be using (either new or existing) and take an extra
@@ -2720,11 +2828,9 @@ static int drm_mode_cursor_universal(struct drm_crtc *crtc,
 			fb = NULL;
 		}
 	} else {
-		mutex_lock(&dev->mode_config.mutex);
 		fb = crtc->cursor->fb;
 		if (fb)
 			drm_framebuffer_reference(fb);
-		mutex_unlock(&dev->mode_config.mutex);
 	}
 
 	if (req->flags & DRM_MODE_CURSOR_MOVE) {
@@ -2746,7 +2852,7 @@ static int drm_mode_cursor_universal(struct drm_crtc *crtc,
 	 * setplane_internal will take care of deref'ing either the old or new
 	 * framebuffer depending on success.
 	 */
-	ret = setplane_internal(crtc->cursor, crtc, fb,
+	ret = __setplane_internal(crtc->cursor, crtc, fb,
 				crtc_x, crtc_y, crtc_w, crtc_h,
 				0, 0, src_w, src_h);
 
@@ -2782,10 +2888,12 @@ static int drm_mode_cursor_common(struct drm_device *dev,
 	 * If this crtc has a universal cursor plane, call that plane's update
 	 * handler rather than using legacy cursor handlers.
 	 */
-	if (crtc->cursor)
-		return drm_mode_cursor_universal(crtc, req, file_priv);
+	drm_modeset_lock_crtc(crtc, crtc->cursor);
+	if (crtc->cursor) {
+		ret = drm_mode_cursor_universal(crtc, req, file_priv);
+		goto out;
+	}
 
-	drm_modeset_lock(&crtc->mutex, NULL);
 	if (req->flags & DRM_MODE_CURSOR_BO) {
 		if (!crtc->funcs->cursor_set && !crtc->funcs->cursor_set2) {
 			ret = -ENXIO;
@@ -2809,7 +2917,7 @@ static int drm_mode_cursor_common(struct drm_device *dev,
 		}
 	}
 out:
-	drm_modeset_unlock(&crtc->mutex);
+	drm_modeset_unlock_crtc(crtc);
 
 	return ret;
 
@@ -2827,7 +2935,7 @@ out:
  * Called by the user via ioctl.
  *
  * Returns:
- * Zero on success, errno on failure.
+ * Zero on success, negative errno on failure.
  */
 int drm_mode_cursor_ioctl(struct drm_device *dev,
 			  void *data, struct drm_file *file_priv)
@@ -2854,7 +2962,7 @@ int drm_mode_cursor_ioctl(struct drm_device *dev,
  * Called by the user via ioctl.
  *
  * Returns:
- * Zero on success, errno on failure.
+ * Zero on success, negative errno on failure.
  */
 int drm_mode_cursor2_ioctl(struct drm_device *dev,
 			   void *data, struct drm_file *file_priv)
@@ -2913,23 +3021,21 @@ EXPORT_SYMBOL(drm_mode_legacy_fb_format);
  * @file_priv: drm file for the ioctl call
  *
  * Add a new FB to the specified CRTC, given a user request. This is the
- * original addfb ioclt which only supported RGB formats.
+ * original addfb ioctl which only supported RGB formats.
  *
  * Called by the user via ioctl.
  *
  * Returns:
- * Zero on success, errno on failure.
+ * Zero on success, negative errno on failure.
  */
 int drm_mode_addfb(struct drm_device *dev,
 		   void *data, struct drm_file *file_priv)
 {
 	struct drm_mode_fb_cmd *or = data;
 	struct drm_mode_fb_cmd2 r = {};
-	struct drm_mode_config *config = &dev->mode_config;
-	struct drm_framebuffer *fb;
-	int ret = 0;
+	int ret;
 
-	/* Use new struct with format internally */
+	/* convert to new format and call new ioctl */
 	r.fb_id = or->fb_id;
 	r.width = or->width;
 	r.height = or->height;
@@ -2937,28 +3043,13 @@ int drm_mode_addfb(struct drm_device *dev,
 	r.pixel_format = drm_mode_legacy_fb_format(or->bpp, or->depth);
 	r.handles[0] = or->handle;
 
-	if (!drm_core_check_feature(dev, DRIVER_MODESET))
-		return -EINVAL;
-
-	if ((config->min_width > r.width) || (r.width > config->max_width))
-		return -EINVAL;
-
-	if ((config->min_height > r.height) || (r.height > config->max_height))
-		return -EINVAL;
-
-	fb = dev->mode_config.funcs->fb_create(dev, file_priv, &r);
-	if (IS_ERR(fb)) {
-		DRM_DEBUG_KMS("could not create framebuffer\n");
-		return PTR_ERR(fb);
-	}
+	ret = drm_mode_addfb2(dev, &r, file_priv);
+	if (ret)
+		return ret;
 
-	mutex_lock(&file_priv->fbs_lock);
-	or->fb_id = fb->base.id;
-	list_add(&fb->filp_head, &file_priv->fbs);
-	DRM_DEBUG_KMS("[FB:%d]\n", fb->base.id);
-	mutex_unlock(&file_priv->fbs_lock);
+	or->fb_id = r.fb_id;
 
-	return ret;
+	return 0;
 }
 
 static int format_check(const struct drm_mode_fb_cmd2 *r)
@@ -3050,7 +3141,7 @@ static int framebuffer_check(const struct drm_mode_fb_cmd2 *r)
 	num_planes = drm_format_num_planes(r->pixel_format);
 
 	if (r->width == 0 || r->width % hsub) {
-		DRM_DEBUG_KMS("bad framebuffer width %u\n", r->height);
+		DRM_DEBUG_KMS("bad framebuffer width %u\n", r->width);
 		return -EINVAL;
 	}
 
@@ -3140,7 +3231,7 @@ static struct drm_framebuffer *add_framebuffer_internal(struct drm_device *dev,
  * Called by the user via ioctl.
  *
  * Returns:
- * Zero on success, errno on failure.
+ * Zero on success, negative errno on failure.
  */
 int drm_mode_addfb2(struct drm_device *dev,
 		    void *data, struct drm_file *file_priv)
@@ -3168,7 +3259,7 @@ int drm_mode_addfb2(struct drm_device *dev,
  * Called by the user via ioctl.
  *
  * Returns:
- * Zero on success, errno on failure.
+ * Zero on success, negative errno on failure.
  */
 int drm_mode_rmfb(struct drm_device *dev,
 		   void *data, struct drm_file *file_priv)
@@ -3222,7 +3313,7 @@ fail_lookup:
  * Called by the user via ioctl.
  *
  * Returns:
- * Zero on success, errno on failure.
+ * Zero on success, negative errno on failure.
  */
 int drm_mode_getfb(struct drm_device *dev,
 		   void *data, struct drm_file *file_priv)
@@ -3283,7 +3374,7 @@ int drm_mode_getfb(struct drm_device *dev,
  * Called by the user via ioctl.
  *
  * Returns:
- * Zero on success, errno on failure.
+ * Zero on success, negative errno on failure.
  */
 int drm_mode_dirtyfb_ioctl(struct drm_device *dev,
 			   void *data, struct drm_file *file_priv)
@@ -3363,14 +3454,23 @@ out_err1:
  * Called by the user via ioctl.
  *
  * Returns:
- * Zero on success, errno on failure.
+ * Zero on success, negative errno on failure.
  */
 void drm_fb_release(struct drm_file *priv)
 {
 	struct drm_device *dev = priv->minor->dev;
 	struct drm_framebuffer *fb, *tfb;
 
-	mutex_lock(&priv->fbs_lock);
+	/*
+	 * When the file gets released that means no one else can access the fb
+	 * list any more, so no need to grab fpriv->fbs_lock. And we need to
+	 * avoid upsetting lockdep since the universal cursor code adds a
+	 * framebuffer while holding mutex locks.
+	 *
+	 * Note that a real deadlock between fpriv->fbs_lock and the modeset
+	 * locks is impossible here since no one else but this function can get
+	 * at it any more.
+	 */
 	list_for_each_entry_safe(fb, tfb, &priv->fbs, filp_head) {
 
 		mutex_lock(&dev->mode_config.fb_lock);
@@ -3383,7 +3483,6 @@ void drm_fb_release(struct drm_file *priv)
 		/* This will also drop the fpriv->fbs reference. */
 		drm_framebuffer_remove(fb);
 	}
-	mutex_unlock(&priv->fbs_lock);
 }
 
 /**
@@ -3397,6 +3496,10 @@ void drm_fb_release(struct drm_file *priv)
  * object with drm_object_attach_property. The returned property object must be
  * freed with drm_property_destroy.
  *
+ * Note that the DRM core keeps a per-device list of properties and that, if
+ * drm_mode_config_cleanup() is called, it will destroy all properties created
+ * by the driver.
+ *
  * Returns:
  * A pointer to the newly created property on success, NULL on failure.
  */
@@ -3424,7 +3527,7 @@ struct drm_property *drm_property_create(struct drm_device *dev, int flags,
 
 	property->flags = flags;
 	property->num_values = num_values;
-	INIT_LIST_HEAD(&property->enum_blob_list);
+	INIT_LIST_HEAD(&property->enum_list);
 
 	if (name) {
 		strncpy(property->name, name, DRM_PROP_NAME_LEN);
@@ -3495,9 +3598,10 @@ EXPORT_SYMBOL(drm_property_create_enum);
  * @flags: flags specifying the property type
  * @name: name of the property
  * @props: enumeration lists with property bitflags
- * @num_values: number of pre-defined values
+ * @num_props: size of the @props array
+ * @supported_bits: bitmask of all supported enumeration values
  *
- * This creates a new generic drm property which can then be attached to a drm
+ * This creates a new bitmask drm property which can then be attached to a drm
  * object with drm_object_attach_property. The returned property object must be
  * freed with drm_property_destroy.
  *
@@ -3572,7 +3676,7 @@ static struct drm_property *property_create_range(struct drm_device *dev,
  * object with drm_object_attach_property. The returned property object must be
  * freed with drm_property_destroy.
  *
- * Userspace is allowed to set any interger value in the (min, max) range
+ * Userspace is allowed to set any integer value in the (min, max) range
  * inclusive.
  *
  * Returns:
@@ -3645,8 +3749,8 @@ int drm_property_add_enum(struct drm_property *property, int index,
 			(value > 63))
 		return -EINVAL;
 
-	if (!list_empty(&property->enum_blob_list)) {
-		list_for_each_entry(prop_enum, &property->enum_blob_list, head) {
+	if (!list_empty(&property->enum_list)) {
+		list_for_each_entry(prop_enum, &property->enum_list, head) {
 			if (prop_enum->value == value) {
 				strncpy(prop_enum->name, name, DRM_PROP_NAME_LEN);
 				prop_enum->name[DRM_PROP_NAME_LEN-1] = '\0';
@@ -3664,7 +3768,7 @@ int drm_property_add_enum(struct drm_property *property, int index,
 	prop_enum->value = value;
 
 	property->values[index] = value;
-	list_add_tail(&prop_enum->head, &property->enum_blob_list);
+	list_add_tail(&prop_enum->head, &property->enum_list);
 	return 0;
 }
 EXPORT_SYMBOL(drm_property_add_enum);
@@ -3681,7 +3785,7 @@ void drm_property_destroy(struct drm_device *dev, struct drm_property *property)
 {
 	struct drm_property_enum *prop_enum, *pt;
 
-	list_for_each_entry_safe(prop_enum, pt, &property->enum_blob_list, head) {
+	list_for_each_entry_safe(prop_enum, pt, &property->enum_list, head) {
 		list_del(&prop_enum->head);
 		kfree(prop_enum);
 	}
@@ -3784,17 +3888,20 @@ int drm_object_property_get_value(struct drm_mode_object *obj,
 EXPORT_SYMBOL(drm_object_property_get_value);
 
 /**
- * drm_mode_getproperty_ioctl - get the current value of a connector's property
+ * drm_mode_getproperty_ioctl - get the property metadata
  * @dev: DRM device
  * @data: ioctl data
  * @file_priv: DRM file info
  *
- * This function retrieves the current value for an connectors's property.
+ * This function retrieves the metadata for a given property, like the different
+ * possible values for an enum property or the limits for a range property.
+ *
+ * Blob properties are special
  *
  * Called by the user via ioctl.
  *
  * Returns:
- * Zero on success, errno on failure.
+ * Zero on success, negative errno on failure.
  */
 int drm_mode_getproperty_ioctl(struct drm_device *dev,
 			       void *data, struct drm_file *file_priv)
@@ -3802,16 +3909,12 @@ int drm_mode_getproperty_ioctl(struct drm_device *dev,
 	struct drm_mode_get_property *out_resp = data;
 	struct drm_property *property;
 	int enum_count = 0;
-	int blob_count = 0;
 	int value_count = 0;
 	int ret = 0, i;
 	int copied;
 	struct drm_property_enum *prop_enum;
 	struct drm_mode_property_enum __user *enum_ptr;
-	struct drm_property_blob *prop_blob;
-	uint32_t __user *blob_id_ptr;
 	uint64_t __user *values_ptr;
-	uint32_t __user *blob_length_ptr;
 
 	if (!drm_core_check_feature(dev, DRIVER_MODESET))
 		return -EINVAL;
@@ -3825,11 +3928,8 @@ int drm_mode_getproperty_ioctl(struct drm_device *dev,
 
 	if (drm_property_type_is(property, DRM_MODE_PROP_ENUM) ||
 			drm_property_type_is(property, DRM_MODE_PROP_BITMASK)) {
-		list_for_each_entry(prop_enum, &property->enum_blob_list, head)
+		list_for_each_entry(prop_enum, &property->enum_list, head)
 			enum_count++;
-	} else if (drm_property_type_is(property, DRM_MODE_PROP_BLOB)) {
-		list_for_each_entry(prop_blob, &property->enum_blob_list, head)
-			blob_count++;
 	}
 
 	value_count = property->num_values;
@@ -3854,7 +3954,7 @@ int drm_mode_getproperty_ioctl(struct drm_device *dev,
 		if ((out_resp->count_enum_blobs >= enum_count) && enum_count) {
 			copied = 0;
 			enum_ptr = (struct drm_mode_property_enum __user *)(unsigned long)out_resp->enum_blob_ptr;
-			list_for_each_entry(prop_enum, &property->enum_blob_list, head) {
+			list_for_each_entry(prop_enum, &property->enum_list, head) {
 
 				if (copy_to_user(&enum_ptr[copied].value, &prop_enum->value, sizeof(uint64_t))) {
 					ret = -EFAULT;
@@ -3872,35 +3972,24 @@ int drm_mode_getproperty_ioctl(struct drm_device *dev,
 		out_resp->count_enum_blobs = enum_count;
 	}
 
-	if (drm_property_type_is(property, DRM_MODE_PROP_BLOB)) {
-		if ((out_resp->count_enum_blobs >= blob_count) && blob_count) {
-			copied = 0;
-			blob_id_ptr = (uint32_t __user *)(unsigned long)out_resp->enum_blob_ptr;
-			blob_length_ptr = (uint32_t __user *)(unsigned long)out_resp->values_ptr;
-
-			list_for_each_entry(prop_blob, &property->enum_blob_list, head) {
-				if (put_user(prop_blob->base.id, blob_id_ptr + copied)) {
-					ret = -EFAULT;
-					goto done;
-				}
-
-				if (put_user(prop_blob->length, blob_length_ptr + copied)) {
-					ret = -EFAULT;
-					goto done;
-				}
-
-				copied++;
-			}
-		}
-		out_resp->count_enum_blobs = blob_count;
-	}
+	/*
+	 * NOTE: The idea seems to have been to use this to read all the blob
+	 * property values. But nothing ever added them to the corresponding
+	 * list, userspace always used the special-purpose get_blob ioctl to
+	 * read the value for a blob property. It also doesn't make a lot of
+	 * sense to return values here when everything else is just metadata for
+	 * the property itself.
+	 */
+	if (drm_property_type_is(property, DRM_MODE_PROP_BLOB))
+		out_resp->count_enum_blobs = 0;
 done:
 	drm_modeset_unlock_all(dev);
 	return ret;
 }
 
-static struct drm_property_blob *drm_property_create_blob(struct drm_device *dev, int length,
-							  void *data)
+static struct drm_property_blob *
+drm_property_create_blob(struct drm_device *dev, size_t length,
+			 const void *data)
 {
 	struct drm_property_blob *blob;
 	int ret;
@@ -3946,7 +4035,7 @@ static void drm_property_destroy_blob(struct drm_device *dev,
  * Called by the user via ioctl.
  *
  * Returns:
- * Zero on success, errno on failure.
+ * Zero on success, negative errno on failure.
  */
 int drm_mode_getblob_ioctl(struct drm_device *dev,
 			   void *data, struct drm_file *file_priv)
@@ -3980,12 +4069,25 @@ done:
 	return ret;
 }
 
+/**
+ * drm_mode_connector_set_path_property - set tile property on connector
+ * @connector: connector to set property on.
+ * @path: path to use for property.
+ *
+ * This creates a property to expose to userspace to specify a
+ * connector path. This is mainly used for DisplayPort MST where
+ * connectors have a topology and we want to allow userspace to give
+ * them more meaningful names.
+ *
+ * Returns:
+ * Zero on success, negative errno on failure.
+ */
 int drm_mode_connector_set_path_property(struct drm_connector *connector,
-					 char *path)
+					 const char *path)
 {
 	struct drm_device *dev = connector->dev;
-	int ret, size;
-	size = strlen(path) + 1;
+	size_t size = strlen(path) + 1;
+	int ret;
 
 	connector->path_blob_ptr = drm_property_create_blob(connector->dev,
 							    size, path);
@@ -4000,6 +4102,52 @@ int drm_mode_connector_set_path_property(struct drm_connector *connector,
 EXPORT_SYMBOL(drm_mode_connector_set_path_property);
 
 /**
+ * drm_mode_connector_set_tile_property - set tile property on connector
+ * @connector: connector to set property on.
+ *
+ * This looks up the tile information for a connector, and creates a
+ * property for userspace to parse if it exists. The property is of
+ * the form of 8 integers using ':' as a separator.
+ *
+ * Returns:
+ * Zero on success, errno on failure.
+ */
+int drm_mode_connector_set_tile_property(struct drm_connector *connector)
+{
+	struct drm_device *dev = connector->dev;
+	int ret, size;
+	char tile[256];
+
+	if (connector->tile_blob_ptr)
+		drm_property_destroy_blob(dev, connector->tile_blob_ptr);
+
+	if (!connector->has_tile) {
+		connector->tile_blob_ptr = NULL;
+		ret = drm_object_property_set_value(&connector->base,
+						    dev->mode_config.tile_property, 0);
+		return ret;
+	}
+
+	snprintf(tile, 256, "%d:%d:%d:%d:%d:%d:%d:%d",
+		 connector->tile_group->id, connector->tile_is_single_monitor,
+		 connector->num_h_tile, connector->num_v_tile,
+		 connector->tile_h_loc, connector->tile_v_loc,
+		 connector->tile_h_size, connector->tile_v_size);
+	size = strlen(tile) + 1;
+
+	connector->tile_blob_ptr = drm_property_create_blob(connector->dev,
+							    size, tile);
+	if (!connector->tile_blob_ptr)
+		return -EINVAL;
+
+	ret = drm_object_property_set_value(&connector->base,
+					    dev->mode_config.tile_property,
+					    connector->tile_blob_ptr->base.id);
+	return ret;
+}
+EXPORT_SYMBOL(drm_mode_connector_set_tile_property);
+
+/**
  * drm_mode_connector_update_edid_property - update the edid property of a connector
  * @connector: drm connector
  * @edid: new value of the edid property
@@ -4008,13 +4156,14 @@ EXPORT_SYMBOL(drm_mode_connector_set_path_property);
  * connector's edid property.
  *
  * Returns:
- * Zero on success, errno on failure.
+ * Zero on success, negative errno on failure.
  */
 int drm_mode_connector_update_edid_property(struct drm_connector *connector,
-					    struct edid *edid)
+					    const struct edid *edid)
 {
 	struct drm_device *dev = connector->dev;
-	int ret, size;
+	size_t size;
+	int ret;
 
 	/* ignore requests to set edid when overridden */
 	if (connector->override_edid)
@@ -4104,7 +4253,7 @@ static bool drm_property_change_is_valid(struct drm_property *property,
  * Called by the user via ioctl.
  *
  * Returns:
- * Zero on success, errno on failure.
+ * Zero on success, negative errno on failure.
  */
 int drm_mode_connector_property_set_ioctl(struct drm_device *dev,
 				       void *data, struct drm_file *file_priv)
@@ -4157,12 +4306,25 @@ static int drm_mode_crtc_set_obj_prop(struct drm_mode_object *obj,
 	return ret;
 }
 
-static int drm_mode_plane_set_obj_prop(struct drm_mode_object *obj,
-				      struct drm_property *property,
-				      uint64_t value)
+/**
+ * drm_mode_plane_set_obj_prop - set the value of a property
+ * @plane: drm plane object to set property value for
+ * @property: property to set
+ * @value: value the property should be set to
+ *
+ * This functions sets a given property on a given plane object. This function
+ * calls the driver's ->set_property callback and changes the software state of
+ * the property if the callback succeeds.
+ *
+ * Returns:
+ * Zero on success, error code on failure.
+ */
+int drm_mode_plane_set_obj_prop(struct drm_plane *plane,
+				struct drm_property *property,
+				uint64_t value)
 {
 	int ret = -EINVAL;
-	struct drm_plane *plane = obj_to_plane(obj);
+	struct drm_mode_object *obj = &plane->base;
 
 	if (plane->funcs->set_property)
 		ret = plane->funcs->set_property(plane, property, value);
@@ -4171,9 +4333,10 @@ static int drm_mode_plane_set_obj_prop(struct drm_mode_object *obj,
 
 	return ret;
 }
+EXPORT_SYMBOL(drm_mode_plane_set_obj_prop);
 
 /**
- * drm_mode_getproperty_ioctl - get the current value of a object's property
+ * drm_mode_obj_get_properties_ioctl - get the current value of a object's property
  * @dev: DRM device
  * @data: ioctl data
  * @file_priv: DRM file info
@@ -4185,7 +4348,7 @@ static int drm_mode_plane_set_obj_prop(struct drm_mode_object *obj,
  * Called by the user via ioctl.
  *
  * Returns:
- * Zero on success, errno on failure.
+ * Zero on success, negative errno on failure.
  */
 int drm_mode_obj_get_properties_ioctl(struct drm_device *dev, void *data,
 				      struct drm_file *file_priv)
@@ -4257,7 +4420,7 @@ out:
  * Called by the user via ioctl.
  *
  * Returns:
- * Zero on success, errno on failure.
+ * Zero on success, negative errno on failure.
  */
 int drm_mode_obj_set_property_ioctl(struct drm_device *dev, void *data,
 				    struct drm_file *file_priv)
@@ -4309,7 +4472,8 @@ int drm_mode_obj_set_property_ioctl(struct drm_device *dev, void *data,
 		ret = drm_mode_crtc_set_obj_prop(arg_obj, property, arg->value);
 		break;
 	case DRM_MODE_OBJECT_PLANE:
-		ret = drm_mode_plane_set_obj_prop(arg_obj, property, arg->value);
+		ret = drm_mode_plane_set_obj_prop(obj_to_plane(arg_obj),
+						  property, arg->value);
 		break;
 	}
 
@@ -4328,7 +4492,7 @@ out:
  * possible_clones and possible_crtcs bitmasks.
  *
  * Returns:
- * Zero on success, errno on failure.
+ * Zero on success, negative errno on failure.
  */
 int drm_mode_connector_attach_encoder(struct drm_connector *connector,
 				      struct drm_encoder *encoder)
@@ -4355,7 +4519,7 @@ EXPORT_SYMBOL(drm_mode_connector_attach_encoder);
  * fixed gamma table size.
  *
  * Returns:
- * Zero on success, errno on failure.
+ * Zero on success, negative errno on failure.
  */
 int drm_mode_crtc_set_gamma_size(struct drm_crtc *crtc,
 				 int gamma_size)
@@ -4384,7 +4548,7 @@ EXPORT_SYMBOL(drm_mode_crtc_set_gamma_size);
  * Called by the user via ioctl.
  *
  * Returns:
- * Zero on success, errno on failure.
+ * Zero on success, negative errno on failure.
  */
 int drm_mode_gamma_set_ioctl(struct drm_device *dev,
 			     void *data, struct drm_file *file_priv)
@@ -4456,7 +4620,7 @@ out:
  * Called by the user via ioctl.
  *
  * Returns:
- * Zero on success, errno on failure.
+ * Zero on success, negative errno on failure.
  */
 int drm_mode_gamma_get_ioctl(struct drm_device *dev,
 			     void *data, struct drm_file *file_priv)
@@ -4522,14 +4686,14 @@ out:
  * Called by the user via ioctl.
  *
  * Returns:
- * Zero on success, errno on failure.
+ * Zero on success, negative errno on failure.
  */
 int drm_mode_page_flip_ioctl(struct drm_device *dev,
 			     void *data, struct drm_file *file_priv)
 {
 	struct drm_mode_crtc_page_flip *page_flip = data;
 	struct drm_crtc *crtc;
-	struct drm_framebuffer *fb = NULL, *old_fb = NULL;
+	struct drm_framebuffer *fb = NULL;
 	struct drm_pending_vblank_event *e = NULL;
 	unsigned long flags;
 	int ret = -EINVAL;
@@ -4545,7 +4709,7 @@ int drm_mode_page_flip_ioctl(struct drm_device *dev,
 	if (!crtc)
 		return -ENOENT;
 
-	drm_modeset_lock(&crtc->mutex, NULL);
+	drm_modeset_lock_crtc(crtc, crtc->primary);
 	if (crtc->primary->fb == NULL) {
 		/* The framebuffer is currently unbound, presumably
 		 * due to a hotplug event, that userspace has not
@@ -4601,7 +4765,7 @@ int drm_mode_page_flip_ioctl(struct drm_device *dev,
 			(void (*) (struct drm_pending_event *)) kfree;
 	}
 
-	old_fb = crtc->primary->fb;
+	crtc->primary->old_fb = crtc->primary->fb;
 	ret = crtc->funcs->page_flip(crtc, fb, e, page_flip->flags);
 	if (ret) {
 		if (page_flip->flags & DRM_MODE_PAGE_FLIP_EVENT) {
@@ -4611,7 +4775,7 @@ int drm_mode_page_flip_ioctl(struct drm_device *dev,
 			kfree(e);
 		}
 		/* Keep the old fb, don't unref it. */
-		old_fb = NULL;
+		crtc->primary->old_fb = NULL;
 	} else {
 		/*
 		 * Warn if the driver hasn't properly updated the crtc->fb
@@ -4627,9 +4791,10 @@ int drm_mode_page_flip_ioctl(struct drm_device *dev,
 out:
 	if (fb)
 		drm_framebuffer_unreference(fb);
-	if (old_fb)
-		drm_framebuffer_unreference(old_fb);
-	drm_modeset_unlock(&crtc->mutex);
+	if (crtc->primary->old_fb)
+		drm_framebuffer_unreference(crtc->primary->old_fb);
+	crtc->primary->old_fb = NULL;
+	drm_modeset_unlock_crtc(crtc);
 
 	return ret;
 }
@@ -4645,9 +4810,14 @@ out:
 void drm_mode_config_reset(struct drm_device *dev)
 {
 	struct drm_crtc *crtc;
+	struct drm_plane *plane;
 	struct drm_encoder *encoder;
 	struct drm_connector *connector;
 
+	list_for_each_entry(plane, &dev->mode_config.plane_list, head)
+		if (plane->funcs->reset)
+			plane->funcs->reset(plane);
+
 	list_for_each_entry(crtc, &dev->mode_config.crtc_list, head)
 		if (crtc->funcs->reset)
 			crtc->funcs->reset(crtc);
@@ -4682,7 +4852,7 @@ EXPORT_SYMBOL(drm_mode_config_reset);
  * Called by the user via ioctl.
  *
  * Returns:
- * Zero on success, errno on failure.
+ * Zero on success, negative errno on failure.
  */
 int drm_mode_create_dumb_ioctl(struct drm_device *dev,
 			       void *data, struct drm_file *file_priv)
@@ -4709,6 +4879,16 @@ int drm_mode_create_dumb_ioctl(struct drm_device *dev,
 	if (PAGE_ALIGN(size) == 0)
 		return -EINVAL;
 
+	/*
+	 * handle, pitch and size are output parameters. Zero them out to
+	 * prevent drivers from accidentally using uninitialized data. Since
+	 * not all existing userspace is clearing these fields properly we
+	 * cannot reject IOCTL with garbage in them.
+	 */
+	args->handle = 0;
+	args->pitch = 0;
+	args->size = 0;
+
 	return dev->driver->dumb_create(file_priv, dev, args);
 }
 
@@ -4724,7 +4904,7 @@ int drm_mode_create_dumb_ioctl(struct drm_device *dev,
  * Called by the user via ioctl.
  *
  * Returns:
- * Zero on success, errno on failure.
+ * Zero on success, negative errno on failure.
  */
 int drm_mode_mmap_dumb_ioctl(struct drm_device *dev,
 			     void *data, struct drm_file *file_priv)
@@ -4751,7 +4931,7 @@ int drm_mode_mmap_dumb_ioctl(struct drm_device *dev,
  * Called by the user via ioctl.
  *
  * Returns:
- * Zero on success, errno on failure.
+ * Zero on success, negative errno on failure.
  */
 int drm_mode_destroy_dumb_ioctl(struct drm_device *dev,
 				void *data, struct drm_file *file_priv)
@@ -5037,6 +5217,7 @@ void drm_mode_config_init(struct drm_device *dev)
 	INIT_LIST_HEAD(&dev->mode_config.property_blob_list);
 	INIT_LIST_HEAD(&dev->mode_config.plane_list);
 	idr_init(&dev->mode_config.crtc_idr);
+	idr_init(&dev->mode_config.tile_idr);
 
 	drm_modeset_lock_all(dev);
 	drm_mode_create_standard_connector_properties(dev);
@@ -5124,6 +5305,7 @@ void drm_mode_config_cleanup(struct drm_device *dev)
 		crtc->funcs->destroy(crtc);
 	}
 
+	idr_destroy(&dev->mode_config.tile_idr);
 	idr_destroy(&dev->mode_config.crtc_idr);
 	drm_modeset_lock_fini(&dev->mode_config.connection_mutex);
 }
@@ -5146,3 +5328,100 @@ struct drm_property *drm_mode_create_rotation_property(struct drm_device *dev,
 					   supported_rotations);
 }
 EXPORT_SYMBOL(drm_mode_create_rotation_property);
+
+/**
+ * DOC: Tile group
+ *
+ * Tile groups are used to represent tiled monitors with a unique
+ * integer identifier. Tiled monitors using DisplayID v1.3 have
+ * a unique 8-byte handle, we store this in a tile group, so we
+ * have a common identifier for all tiles in a monitor group.
+ */
+static void drm_tile_group_free(struct kref *kref)
+{
+	struct drm_tile_group *tg = container_of(kref, struct drm_tile_group, refcount);
+	struct drm_device *dev = tg->dev;
+	mutex_lock(&dev->mode_config.idr_mutex);
+	idr_remove(&dev->mode_config.tile_idr, tg->id);
+	mutex_unlock(&dev->mode_config.idr_mutex);
+	kfree(tg);
+}
+
+/**
+ * drm_mode_put_tile_group - drop a reference to a tile group.
+ * @dev: DRM device
+ * @tg: tile group to drop reference to.
+ *
+ * drop reference to tile group and free if 0.
+ */
+void drm_mode_put_tile_group(struct drm_device *dev,
+			     struct drm_tile_group *tg)
+{
+	kref_put(&tg->refcount, drm_tile_group_free);
+}
+
+/**
+ * drm_mode_get_tile_group - get a reference to an existing tile group
+ * @dev: DRM device
+ * @topology: 8-bytes unique per monitor.
+ *
+ * Use the unique bytes to get a reference to an existing tile group.
+ *
+ * RETURNS:
+ * tile group or NULL if not found.
+ */
+struct drm_tile_group *drm_mode_get_tile_group(struct drm_device *dev,
+					       char topology[8])
+{
+	struct drm_tile_group *tg;
+	int id;
+	mutex_lock(&dev->mode_config.idr_mutex);
+	idr_for_each_entry(&dev->mode_config.tile_idr, tg, id) {
+		if (!memcmp(tg->group_data, topology, 8)) {
+			if (!kref_get_unless_zero(&tg->refcount))
+				tg = NULL;
+			mutex_unlock(&dev->mode_config.idr_mutex);
+			return tg;
+		}
+	}
+	mutex_unlock(&dev->mode_config.idr_mutex);
+	return NULL;
+}
+
+/**
+ * drm_mode_create_tile_group - create a tile group from a displayid description
+ * @dev: DRM device
+ * @topology: 8-bytes unique per monitor.
+ *
+ * Create a tile group for the unique monitor, and get a unique
+ * identifier for the tile group.
+ *
+ * RETURNS:
+ * new tile group or error.
+ */
+struct drm_tile_group *drm_mode_create_tile_group(struct drm_device *dev,
+						  char topology[8])
+{
+	struct drm_tile_group *tg;
+	int ret;
+
+	tg = kzalloc(sizeof(*tg), GFP_KERNEL);
+	if (!tg)
+		return ERR_PTR(-ENOMEM);
+
+	kref_init(&tg->refcount);
+	memcpy(tg->group_data, topology, 8);
+	tg->dev = dev;
+
+	mutex_lock(&dev->mode_config.idr_mutex);
+	ret = idr_alloc(&dev->mode_config.tile_idr, tg, 1, 0, GFP_KERNEL);
+	if (ret >= 0) {
+		tg->id = ret;
+	} else {
+		kfree(tg);
+		tg = ERR_PTR(ret);
+	}
+
+	mutex_unlock(&dev->mode_config.idr_mutex);
+	return tg;
+}
diff --git a/drivers/gpu/drm/drm_crtc_helper.c b/drivers/gpu/drm/drm_crtc_helper.c
index 6c65a0a28fbd..d552708409de 100644
--- a/drivers/gpu/drm/drm_crtc_helper.c
+++ b/drivers/gpu/drm/drm_crtc_helper.c
@@ -34,12 +34,35 @@
 #include <linux/moduleparam.h>
 
 #include <drm/drmP.h>
+#include <drm/drm_atomic.h>
 #include <drm/drm_crtc.h>
 #include <drm/drm_fourcc.h>
 #include <drm/drm_crtc_helper.h>
 #include <drm/drm_fb_helper.h>
+#include <drm/drm_plane_helper.h>
+#include <drm/drm_atomic_helper.h>
 #include <drm/drm_edid.h>
 
+/**
+ * DOC: overview
+ *
+ * The CRTC modeset helper library provides a default set_config implementation
+ * in drm_crtc_helper_set_config(). Plus a few other convenience functions using
+ * the same callbacks which drivers can use to e.g. restore the modeset
+ * configuration on resume with drm_helper_resume_force_mode().
+ *
+ * The driver callbacks are mostly compatible with the atomic modeset helpers,
+ * except for the handling of the primary plane: Atomic helpers require that the
+ * primary plane is implemented as a real standalone plane and not directly tied
+ * to the CRTC state. For easier transition this library provides functions to
+ * implement the old semantics required by the CRTC helpers using the new plane
+ * and atomic helper callbacks.
+ *
+ * Drivers are strongly urged to convert to the atomic helpers (by way of first
+ * converting to the plane helpers). New drivers must not use these functions
+ * but need to implement the atomic interface instead, potentially using the
+ * atomic helpers for that.
+ */
 MODULE_AUTHOR("David Airlie, Jesse Barnes");
 MODULE_DESCRIPTION("DRM KMS helper");
 MODULE_LICENSE("GPL and additional rights");
@@ -888,3 +911,112 @@ void drm_helper_resume_force_mode(struct drm_device *dev)
 	drm_modeset_unlock_all(dev);
 }
 EXPORT_SYMBOL(drm_helper_resume_force_mode);
+
+/**
+ * drm_helper_crtc_mode_set - mode_set implementation for atomic plane helpers
+ * @crtc: DRM CRTC
+ * @mode: DRM display mode which userspace requested
+ * @adjusted_mode: DRM display mode adjusted by ->mode_fixup callbacks
+ * @x: x offset of the CRTC scanout area on the underlying framebuffer
+ * @y: y offset of the CRTC scanout area on the underlying framebuffer
+ * @old_fb: previous framebuffer
+ *
+ * This function implements a callback useable as the ->mode_set callback
+ * required by the crtc helpers. Besides the atomic plane helper functions for
+ * the primary plane the driver must also provide the ->mode_set_nofb callback
+ * to set up the crtc.
+ *
+ * This is a transitional helper useful for converting drivers to the atomic
+ * interfaces.
+ */
+int drm_helper_crtc_mode_set(struct drm_crtc *crtc, struct drm_display_mode *mode,
+			     struct drm_display_mode *adjusted_mode, int x, int y,
+			     struct drm_framebuffer *old_fb)
+{
+	struct drm_crtc_state *crtc_state;
+	struct drm_crtc_helper_funcs *crtc_funcs = crtc->helper_private;
+	int ret;
+
+	if (crtc->funcs->atomic_duplicate_state)
+		crtc_state = crtc->funcs->atomic_duplicate_state(crtc);
+	else if (crtc->state)
+		crtc_state = kmemdup(crtc->state, sizeof(*crtc_state),
+				     GFP_KERNEL);
+	else
+		crtc_state = kzalloc(sizeof(*crtc_state), GFP_KERNEL);
+	if (!crtc_state)
+		return -ENOMEM;
+
+	crtc_state->enable = true;
+	crtc_state->planes_changed = true;
+	crtc_state->mode_changed = true;
+	drm_mode_copy(&crtc_state->mode, mode);
+	drm_mode_copy(&crtc_state->adjusted_mode, adjusted_mode);
+
+	if (crtc_funcs->atomic_check) {
+		ret = crtc_funcs->atomic_check(crtc, crtc_state);
+		if (ret) {
+			kfree(crtc_state);
+
+			return ret;
+		}
+	}
+
+	swap(crtc->state, crtc_state);
+
+	crtc_funcs->mode_set_nofb(crtc);
+
+	if (crtc_state) {
+		if (crtc->funcs->atomic_destroy_state)
+			crtc->funcs->atomic_destroy_state(crtc, crtc_state);
+		else
+			kfree(crtc_state);
+	}
+
+	return drm_helper_crtc_mode_set_base(crtc, x, y, old_fb);
+}
+EXPORT_SYMBOL(drm_helper_crtc_mode_set);
+
+/**
+ * drm_helper_crtc_mode_set_base - mode_set_base implementation for atomic plane helpers
+ * @crtc: DRM CRTC
+ * @x: x offset of the CRTC scanout area on the underlying framebuffer
+ * @y: y offset of the CRTC scanout area on the underlying framebuffer
+ * @old_fb: previous framebuffer
+ *
+ * This function implements a callback useable as the ->mode_set_base used
+ * required by the crtc helpers. The driver must provide the atomic plane helper
+ * functions for the primary plane.
+ *
+ * This is a transitional helper useful for converting drivers to the atomic
+ * interfaces.
+ */
+int drm_helper_crtc_mode_set_base(struct drm_crtc *crtc, int x, int y,
+				  struct drm_framebuffer *old_fb)
+{
+	struct drm_plane_state *plane_state;
+	struct drm_plane *plane = crtc->primary;
+
+	if (plane->funcs->atomic_duplicate_state)
+		plane_state = plane->funcs->atomic_duplicate_state(plane);
+	else if (plane->state)
+		plane_state = drm_atomic_helper_plane_duplicate_state(plane);
+	else
+		plane_state = kzalloc(sizeof(*plane_state), GFP_KERNEL);
+	if (!plane_state)
+		return -ENOMEM;
+
+	plane_state->crtc = crtc;
+	drm_atomic_set_fb_for_plane(plane_state, crtc->primary->fb);
+	plane_state->crtc_x = 0;
+	plane_state->crtc_y = 0;
+	plane_state->crtc_h = crtc->mode.vdisplay;
+	plane_state->crtc_w = crtc->mode.hdisplay;
+	plane_state->src_x = x << 16;
+	plane_state->src_y = y << 16;
+	plane_state->src_h = crtc->mode.vdisplay << 16;
+	plane_state->src_w = crtc->mode.hdisplay << 16;
+
+	return drm_plane_helper_commit(plane, plane_state, old_fb);
+}
+EXPORT_SYMBOL(drm_helper_crtc_mode_set_base);
diff --git a/drivers/gpu/drm/drm_debugfs.c b/drivers/gpu/drm/drm_debugfs.c
index 13bd42923dd4..3bcf8e6a85b3 100644
--- a/drivers/gpu/drm/drm_debugfs.c
+++ b/drivers/gpu/drm/drm_debugfs.c
@@ -36,6 +36,7 @@
 #include <linux/export.h>
 #include <drm/drmP.h>
 #include <drm/drm_edid.h>
+#include "drm_internal.h"
 
 #if defined(CONFIG_DEBUG_FS)
 
@@ -49,9 +50,7 @@ static const struct drm_info_list drm_debugfs_list[] = {
 	{"clients", drm_clients_info, 0},
 	{"bufs", drm_bufs_info, 0},
 	{"gem_names", drm_gem_name_info, DRIVER_GEM},
-#if DRM_DEBUG_CODE
 	{"vma", drm_vma_info, 0},
-#endif
 };
 #define DRM_DEBUGFS_ENTRIES ARRAY_SIZE(drm_debugfs_list)
 
diff --git a/drivers/gpu/drm/drm_dma.c b/drivers/gpu/drm/drm_dma.c
index 8a140a953754..ea481800ef56 100644
--- a/drivers/gpu/drm/drm_dma.c
+++ b/drivers/gpu/drm/drm_dma.c
@@ -35,6 +35,7 @@
 
 #include <linux/export.h>
 #include <drm/drmP.h>
+#include "drm_legacy.h"
 
 /**
  * Initialize the DMA data.
@@ -124,7 +125,7 @@ void drm_legacy_dma_takedown(struct drm_device *dev)
  *
  * Resets the fields of \p buf.
  */
-void drm_free_buffer(struct drm_device *dev, struct drm_buf * buf)
+void drm_legacy_free_buffer(struct drm_device *dev, struct drm_buf * buf)
 {
 	if (!buf)
 		return;
@@ -142,8 +143,8 @@ void drm_free_buffer(struct drm_device *dev, struct drm_buf * buf)
  *
  * Frees each buffer associated with \p file_priv not already on the hardware.
  */
-void drm_core_reclaim_buffers(struct drm_device *dev,
-			      struct drm_file *file_priv)
+void drm_legacy_reclaim_buffers(struct drm_device *dev,
+				struct drm_file *file_priv)
 {
 	struct drm_device_dma *dma = dev->dma;
 	int i;
@@ -154,7 +155,7 @@ void drm_core_reclaim_buffers(struct drm_device *dev,
 		if (dma->buflist[i]->file_priv == file_priv) {
 			switch (dma->buflist[i]->list) {
 			case DRM_LIST_NONE:
-				drm_free_buffer(dev, dma->buflist[i]);
+				drm_legacy_free_buffer(dev, dma->buflist[i]);
 				break;
 			case DRM_LIST_WAIT:
 				dma->buflist[i]->list = DRM_LIST_RECLAIM;
@@ -166,5 +167,3 @@ void drm_core_reclaim_buffers(struct drm_device *dev,
 		}
 	}
 }
-
-EXPORT_SYMBOL(drm_core_reclaim_buffers);
diff --git a/drivers/gpu/drm/drm_dp_helper.c b/drivers/gpu/drm/drm_dp_helper.c
index 08e33b8b13a4..79968e39c8d0 100644
--- a/drivers/gpu/drm/drm_dp_helper.c
+++ b/drivers/gpu/drm/drm_dp_helper.c
@@ -39,198 +39,6 @@
  * blocks, ...
  */
 
-/* Run a single AUX_CH I2C transaction, writing/reading data as necessary */
-static int
-i2c_algo_dp_aux_transaction(struct i2c_adapter *adapter, int mode,
-			    uint8_t write_byte, uint8_t *read_byte)
-{
-	struct i2c_algo_dp_aux_data *algo_data = adapter->algo_data;
-	int ret;
-
-	ret = (*algo_data->aux_ch)(adapter, mode,
-				   write_byte, read_byte);
-	return ret;
-}
-
-/*
- * I2C over AUX CH
- */
-
-/*
- * Send the address. If the I2C link is running, this 'restarts'
- * the connection with the new address, this is used for doing
- * a write followed by a read (as needed for DDC)
- */
-static int
-i2c_algo_dp_aux_address(struct i2c_adapter *adapter, u16 address, bool reading)
-{
-	struct i2c_algo_dp_aux_data *algo_data = adapter->algo_data;
-	int mode = MODE_I2C_START;
-	int ret;
-
-	if (reading)
-		mode |= MODE_I2C_READ;
-	else
-		mode |= MODE_I2C_WRITE;
-	algo_data->address = address;
-	algo_data->running = true;
-	ret = i2c_algo_dp_aux_transaction(adapter, mode, 0, NULL);
-	return ret;
-}
-
-/*
- * Stop the I2C transaction. This closes out the link, sending
- * a bare address packet with the MOT bit turned off
- */
-static void
-i2c_algo_dp_aux_stop(struct i2c_adapter *adapter, bool reading)
-{
-	struct i2c_algo_dp_aux_data *algo_data = adapter->algo_data;
-	int mode = MODE_I2C_STOP;
-
-	if (reading)
-		mode |= MODE_I2C_READ;
-	else
-		mode |= MODE_I2C_WRITE;
-	if (algo_data->running) {
-		(void) i2c_algo_dp_aux_transaction(adapter, mode, 0, NULL);
-		algo_data->running = false;
-	}
-}
-
-/*
- * Write a single byte to the current I2C address, the
- * the I2C link must be running or this returns -EIO
- */
-static int
-i2c_algo_dp_aux_put_byte(struct i2c_adapter *adapter, u8 byte)
-{
-	struct i2c_algo_dp_aux_data *algo_data = adapter->algo_data;
-	int ret;
-
-	if (!algo_data->running)
-		return -EIO;
-
-	ret = i2c_algo_dp_aux_transaction(adapter, MODE_I2C_WRITE, byte, NULL);
-	return ret;
-}
-
-/*
- * Read a single byte from the current I2C address, the
- * I2C link must be running or this returns -EIO
- */
-static int
-i2c_algo_dp_aux_get_byte(struct i2c_adapter *adapter, u8 *byte_ret)
-{
-	struct i2c_algo_dp_aux_data *algo_data = adapter->algo_data;
-	int ret;
-
-	if (!algo_data->running)
-		return -EIO;
-
-	ret = i2c_algo_dp_aux_transaction(adapter, MODE_I2C_READ, 0, byte_ret);
-	return ret;
-}
-
-static int
-i2c_algo_dp_aux_xfer(struct i2c_adapter *adapter,
-		     struct i2c_msg *msgs,
-		     int num)
-{
-	int ret = 0;
-	bool reading = false;
-	int m;
-	int b;
-
-	for (m = 0; m < num; m++) {
-		u16 len = msgs[m].len;
-		u8 *buf = msgs[m].buf;
-		reading = (msgs[m].flags & I2C_M_RD) != 0;
-		ret = i2c_algo_dp_aux_address(adapter, msgs[m].addr, reading);
-		if (ret < 0)
-			break;
-		if (reading) {
-			for (b = 0; b < len; b++) {
-				ret = i2c_algo_dp_aux_get_byte(adapter, &buf[b]);
-				if (ret < 0)
-					break;
-			}
-		} else {
-			for (b = 0; b < len; b++) {
-				ret = i2c_algo_dp_aux_put_byte(adapter, buf[b]);
-				if (ret < 0)
-					break;
-			}
-		}
-		if (ret < 0)
-			break;
-	}
-	if (ret >= 0)
-		ret = num;
-	i2c_algo_dp_aux_stop(adapter, reading);
-	DRM_DEBUG_KMS("dp_aux_xfer return %d\n", ret);
-	return ret;
-}
-
-static u32
-i2c_algo_dp_aux_functionality(struct i2c_adapter *adapter)
-{
-	return I2C_FUNC_I2C | I2C_FUNC_SMBUS_EMUL |
-	       I2C_FUNC_SMBUS_READ_BLOCK_DATA |
-	       I2C_FUNC_SMBUS_BLOCK_PROC_CALL |
-	       I2C_FUNC_10BIT_ADDR;
-}
-
-static const struct i2c_algorithm i2c_dp_aux_algo = {
-	.master_xfer	= i2c_algo_dp_aux_xfer,
-	.functionality	= i2c_algo_dp_aux_functionality,
-};
-
-static void
-i2c_dp_aux_reset_bus(struct i2c_adapter *adapter)
-{
-	(void) i2c_algo_dp_aux_address(adapter, 0, false);
-	(void) i2c_algo_dp_aux_stop(adapter, false);
-}
-
-static int
-i2c_dp_aux_prepare_bus(struct i2c_adapter *adapter)
-{
-	adapter->algo = &i2c_dp_aux_algo;
-	adapter->retries = 3;
-	i2c_dp_aux_reset_bus(adapter);
-	return 0;
-}
-
-/**
- * i2c_dp_aux_add_bus() - register an i2c adapter using the aux ch helper
- * @adapter: i2c adapter to register
- *
- * This registers an i2c adapter that uses dp aux channel as it's underlaying
- * transport. The driver needs to fill out the &i2c_algo_dp_aux_data structure
- * and store it in the algo_data member of the @adapter argument. This will be
- * used by the i2c over dp aux algorithm to drive the hardware.
- *
- * RETURNS:
- * 0 on success, -ERRNO on failure.
- *
- * IMPORTANT:
- * This interface is deprecated, please switch to the new dp aux helpers and
- * drm_dp_aux_register().
- */
-int
-i2c_dp_aux_add_bus(struct i2c_adapter *adapter)
-{
-	int error;
-
-	error = i2c_dp_aux_prepare_bus(adapter);
-	if (error)
-		return error;
-	error = i2c_add_adapter(adapter);
-	return error;
-}
-EXPORT_SYMBOL(i2c_dp_aux_add_bus);
-
 /* Helpers for DP link training */
 static u8 dp_link_status(const u8 link_status[DP_LINK_STATUS_SIZE], int r)
 {
@@ -378,10 +186,11 @@ static int drm_dp_dpcd_access(struct drm_dp_aux *aux, u8 request,
 
 	/*
 	 * The specification doesn't give any recommendation on how often to
-	 * retry native transactions, so retry 7 times like for I2C-over-AUX
-	 * transactions.
+	 * retry native transactions. We used to retry 7 times like for
+	 * aux i2c transactions but real world devices this wasn't
+	 * sufficient, bump to 32 which makes Dell 4k monitors happier.
 	 */
-	for (retry = 0; retry < 7; retry++) {
+	for (retry = 0; retry < 32; retry++) {
 
 		mutex_lock(&aux->hw_mutex);
 		err = aux->transfer(aux, &msg);
@@ -654,10 +463,12 @@ static int drm_dp_i2c_do_msg(struct drm_dp_aux *aux, struct drm_dp_aux_msg *msg)
 
 		case DP_AUX_I2C_REPLY_NACK:
 			DRM_DEBUG_KMS("I2C nack\n");
+			aux->i2c_nack_count++;
 			return -EREMOTEIO;
 
 		case DP_AUX_I2C_REPLY_DEFER:
 			DRM_DEBUG_KMS("I2C defer\n");
+			aux->i2c_defer_count++;
 			usleep_range(400, 500);
 			continue;
 
diff --git a/drivers/gpu/drm/drm_dp_mst_topology.c b/drivers/gpu/drm/drm_dp_mst_topology.c
index ac3c2738db94..9a5b68717ec8 100644
--- a/drivers/gpu/drm/drm_dp_mst_topology.c
+++ b/drivers/gpu/drm/drm_dp_mst_topology.c
@@ -682,7 +682,7 @@ static int build_allocate_payload(struct drm_dp_sideband_msg_tx *msg, int port_n
 static int drm_dp_mst_assign_payload_id(struct drm_dp_mst_topology_mgr *mgr,
 					struct drm_dp_vcpi *vcpi)
 {
-	int ret;
+	int ret, vcpi_ret;
 
 	mutex_lock(&mgr->payload_lock);
 	ret = find_first_zero_bit(&mgr->payload_mask, mgr->max_payloads + 1);
@@ -692,8 +692,16 @@ static int drm_dp_mst_assign_payload_id(struct drm_dp_mst_topology_mgr *mgr,
 		goto out_unlock;
 	}
 
+	vcpi_ret = find_first_zero_bit(&mgr->vcpi_mask, mgr->max_payloads + 1);
+	if (vcpi_ret > mgr->max_payloads) {
+		ret = -EINVAL;
+		DRM_DEBUG_KMS("out of vcpi ids %d\n", ret);
+		goto out_unlock;
+	}
+
 	set_bit(ret, &mgr->payload_mask);
-	vcpi->vcpi = ret;
+	set_bit(vcpi_ret, &mgr->vcpi_mask);
+	vcpi->vcpi = vcpi_ret + 1;
 	mgr->proposed_vcpis[ret - 1] = vcpi;
 out_unlock:
 	mutex_unlock(&mgr->payload_lock);
@@ -701,15 +709,23 @@ out_unlock:
 }
 
 static void drm_dp_mst_put_payload_id(struct drm_dp_mst_topology_mgr *mgr,
-				      int id)
+				      int vcpi)
 {
-	if (id == 0)
+	int i;
+	if (vcpi == 0)
 		return;
 
 	mutex_lock(&mgr->payload_lock);
-	DRM_DEBUG_KMS("putting payload %d\n", id);
-	clear_bit(id, &mgr->payload_mask);
-	mgr->proposed_vcpis[id - 1] = NULL;
+	DRM_DEBUG_KMS("putting payload %d\n", vcpi);
+	clear_bit(vcpi - 1, &mgr->vcpi_mask);
+
+	for (i = 0; i < mgr->max_payloads; i++) {
+		if (mgr->proposed_vcpis[i])
+			if (mgr->proposed_vcpis[i]->vcpi == vcpi) {
+				mgr->proposed_vcpis[i] = NULL;
+				clear_bit(i + 1, &mgr->payload_mask);
+			}
+	}
 	mutex_unlock(&mgr->payload_lock);
 }
 
@@ -823,6 +839,8 @@ static void drm_dp_put_mst_branch_device(struct drm_dp_mst_branch *mstb)
 
 static void drm_dp_port_teardown_pdt(struct drm_dp_mst_port *port, int old_pdt)
 {
+	struct drm_dp_mst_branch *mstb;
+
 	switch (old_pdt) {
 	case DP_PEER_DEVICE_DP_LEGACY_CONV:
 	case DP_PEER_DEVICE_SST_SINK:
@@ -830,8 +848,9 @@ static void drm_dp_port_teardown_pdt(struct drm_dp_mst_port *port, int old_pdt)
 		drm_dp_mst_unregister_i2c_bus(&port->aux);
 		break;
 	case DP_PEER_DEVICE_MST_BRANCHING:
-		drm_dp_put_mst_branch_device(port->mstb);
+		mstb = port->mstb;
 		port->mstb = NULL;
+		drm_dp_put_mst_branch_device(mstb);
 		break;
 	}
 }
@@ -842,6 +861,8 @@ static void drm_dp_destroy_port(struct kref *kref)
 	struct drm_dp_mst_topology_mgr *mgr = port->mgr;
 	if (!port->input) {
 		port->vcpi.num_slots = 0;
+
+		kfree(port->cached_edid);
 		if (port->connector)
 			(*port->mgr->cbs->destroy_connector)(mgr, port->connector);
 		drm_dp_port_teardown_pdt(port, port->pdt);
@@ -995,19 +1016,20 @@ static void drm_dp_check_port_guid(struct drm_dp_mst_branch *mstb,
 
 static void build_mst_prop_path(struct drm_dp_mst_port *port,
 				struct drm_dp_mst_branch *mstb,
-				char *proppath)
+				char *proppath,
+				size_t proppath_size)
 {
 	int i;
 	char temp[8];
-	snprintf(proppath, 255, "mst:%d", mstb->mgr->conn_base_id);
+	snprintf(proppath, proppath_size, "mst:%d", mstb->mgr->conn_base_id);
 	for (i = 0; i < (mstb->lct - 1); i++) {
 		int shift = (i % 2) ? 0 : 4;
 		int port_num = mstb->rad[i / 2] >> shift;
-		snprintf(temp, 8, "-%d", port_num);
-		strncat(proppath, temp, 255);
+		snprintf(temp, sizeof(temp), "-%d", port_num);
+		strlcat(proppath, temp, proppath_size);
 	}
-	snprintf(temp, 8, "-%d", port->port_num);
-	strncat(proppath, temp, 255);
+	snprintf(temp, sizeof(temp), "-%d", port->port_num);
+	strlcat(proppath, temp, proppath_size);
 }
 
 static void drm_dp_add_port(struct drm_dp_mst_branch *mstb,
@@ -1078,8 +1100,12 @@ static void drm_dp_add_port(struct drm_dp_mst_branch *mstb,
 
 	if (created && !port->input) {
 		char proppath[255];
-		build_mst_prop_path(port, mstb, proppath);
+		build_mst_prop_path(port, mstb, proppath, sizeof(proppath));
 		port->connector = (*mstb->mgr->cbs->add_connector)(mstb->mgr, port, proppath);
+
+		if (port->port_num >= 8) {
+			port->cached_edid = drm_get_edid(port->connector, &port->aux.ddc);
+		}
 	}
 
 	/* put reference to this port */
@@ -1563,7 +1589,7 @@ static int drm_dp_destroy_payload_step1(struct drm_dp_mst_topology_mgr *mgr,
 	}
 
 	drm_dp_dpcd_write_payload(mgr, id, payload);
-	payload->payload_state = 0;
+	payload->payload_state = DP_PAYLOAD_DELETE_LOCAL;
 	return 0;
 }
 
@@ -1590,7 +1616,7 @@ static int drm_dp_destroy_payload_step2(struct drm_dp_mst_topology_mgr *mgr,
  */
 int drm_dp_update_payload_part1(struct drm_dp_mst_topology_mgr *mgr)
 {
-	int i;
+	int i, j;
 	int cur_slots = 1;
 	struct drm_dp_payload req_payload;
 	struct drm_dp_mst_port *port;
@@ -1607,26 +1633,46 @@ int drm_dp_update_payload_part1(struct drm_dp_mst_topology_mgr *mgr)
 			port = NULL;
 			req_payload.num_slots = 0;
 		}
+
+		if (mgr->payloads[i].start_slot != req_payload.start_slot) {
+			mgr->payloads[i].start_slot = req_payload.start_slot;
+		}
 		/* work out what is required to happen with this payload */
-		if (mgr->payloads[i].start_slot != req_payload.start_slot ||
-		    mgr->payloads[i].num_slots != req_payload.num_slots) {
+		if (mgr->payloads[i].num_slots != req_payload.num_slots) {
 
 			/* need to push an update for this payload */
 			if (req_payload.num_slots) {
-				drm_dp_create_payload_step1(mgr, i + 1, &req_payload);
+				drm_dp_create_payload_step1(mgr, mgr->proposed_vcpis[i]->vcpi, &req_payload);
 				mgr->payloads[i].num_slots = req_payload.num_slots;
 			} else if (mgr->payloads[i].num_slots) {
 				mgr->payloads[i].num_slots = 0;
-				drm_dp_destroy_payload_step1(mgr, port, i + 1, &mgr->payloads[i]);
+				drm_dp_destroy_payload_step1(mgr, port, port->vcpi.vcpi, &mgr->payloads[i]);
 				req_payload.payload_state = mgr->payloads[i].payload_state;
-			} else
-				req_payload.payload_state = 0;
-
-			mgr->payloads[i].start_slot = req_payload.start_slot;
+				mgr->payloads[i].start_slot = 0;
+			}
 			mgr->payloads[i].payload_state = req_payload.payload_state;
 		}
 		cur_slots += req_payload.num_slots;
 	}
+
+	for (i = 0; i < mgr->max_payloads; i++) {
+		if (mgr->payloads[i].payload_state == DP_PAYLOAD_DELETE_LOCAL) {
+			DRM_DEBUG_KMS("removing payload %d\n", i);
+			for (j = i; j < mgr->max_payloads - 1; j++) {
+				memcpy(&mgr->payloads[j], &mgr->payloads[j + 1], sizeof(struct drm_dp_payload));
+				mgr->proposed_vcpis[j] = mgr->proposed_vcpis[j + 1];
+				if (mgr->proposed_vcpis[j] && mgr->proposed_vcpis[j]->num_slots) {
+					set_bit(j + 1, &mgr->payload_mask);
+				} else {
+					clear_bit(j + 1, &mgr->payload_mask);
+				}
+			}
+			memset(&mgr->payloads[mgr->max_payloads - 1], 0, sizeof(struct drm_dp_payload));
+			mgr->proposed_vcpis[mgr->max_payloads - 1] = NULL;
+			clear_bit(mgr->max_payloads, &mgr->payload_mask);
+
+		}
+	}
 	mutex_unlock(&mgr->payload_lock);
 
 	return 0;
@@ -1657,9 +1703,9 @@ int drm_dp_update_payload_part2(struct drm_dp_mst_topology_mgr *mgr)
 
 		DRM_DEBUG_KMS("payload %d %d\n", i, mgr->payloads[i].payload_state);
 		if (mgr->payloads[i].payload_state == DP_PAYLOAD_LOCAL) {
-			ret = drm_dp_create_payload_step2(mgr, port, i + 1, &mgr->payloads[i]);
+			ret = drm_dp_create_payload_step2(mgr, port, mgr->proposed_vcpis[i]->vcpi, &mgr->payloads[i]);
 		} else if (mgr->payloads[i].payload_state == DP_PAYLOAD_DELETE_LOCAL) {
-			ret = drm_dp_destroy_payload_step2(mgr, i + 1, &mgr->payloads[i]);
+			ret = drm_dp_destroy_payload_step2(mgr, mgr->proposed_vcpis[i]->vcpi, &mgr->payloads[i]);
 		}
 		if (ret) {
 			mutex_unlock(&mgr->payload_lock);
@@ -1762,17 +1808,27 @@ static int drm_dp_send_up_ack_reply(struct drm_dp_mst_topology_mgr *mgr,
 	return 0;
 }
 
-static int drm_dp_get_vc_payload_bw(int dp_link_bw, int dp_link_count)
+static bool drm_dp_get_vc_payload_bw(int dp_link_bw,
+				     int dp_link_count,
+				     int *out)
 {
 	switch (dp_link_bw) {
+	default:
+		DRM_DEBUG_KMS("invalid link bandwidth in DPCD: %x (link count: %d)\n",
+			      dp_link_bw, dp_link_count);
+		return false;
+
 	case DP_LINK_BW_1_62:
-		return 3 * dp_link_count;
+		*out = 3 * dp_link_count;
+		break;
 	case DP_LINK_BW_2_7:
-		return 5 * dp_link_count;
+		*out = 5 * dp_link_count;
+		break;
 	case DP_LINK_BW_5_4:
-		return 10 * dp_link_count;
+		*out = 10 * dp_link_count;
+		break;
 	}
-	return 0;
+	return true;
 }
 
 /**
@@ -1804,7 +1860,13 @@ int drm_dp_mst_topology_mgr_set_mst(struct drm_dp_mst_topology_mgr *mgr, bool ms
 			goto out_unlock;
 		}
 
-		mgr->pbn_div = drm_dp_get_vc_payload_bw(mgr->dpcd[1], mgr->dpcd[2] & DP_MAX_LANE_COUNT_MASK);
+		if (!drm_dp_get_vc_payload_bw(mgr->dpcd[1],
+					      mgr->dpcd[2] & DP_MAX_LANE_COUNT_MASK,
+					      &mgr->pbn_div)) {
+			ret = -EINVAL;
+			goto out_unlock;
+		}
+
 		mgr->total_pbn = 2560;
 		mgr->total_slots = DIV_ROUND_UP(mgr->total_pbn, mgr->pbn_div);
 		mgr->avail_slots = mgr->total_slots;
@@ -1861,6 +1923,7 @@ int drm_dp_mst_topology_mgr_set_mst(struct drm_dp_mst_topology_mgr *mgr, bool ms
 		memset(mgr->payloads, 0, mgr->max_payloads * sizeof(struct drm_dp_payload));
 		mgr->payload_mask = 0;
 		set_bit(0, &mgr->payload_mask);
+		mgr->vcpi_mask = 0;
 	}
 
 out_unlock:
@@ -2071,6 +2134,7 @@ static int drm_dp_mst_handle_up_req(struct drm_dp_mst_topology_mgr *mgr)
  * drm_dp_mst_hpd_irq() - MST hotplug IRQ notify
  * @mgr: manager to notify irq for.
  * @esi: 4 bytes from SINK_COUNT_ESI
+ * @handled: whether the hpd interrupt was consumed or not
  *
  * This should be called from the driver when it detects a short IRQ,
  * along with the value of the DEVICE_SERVICE_IRQ_VECTOR_ESI0. The
@@ -2112,7 +2176,8 @@ EXPORT_SYMBOL(drm_dp_mst_hpd_irq);
  * This returns the current connection state for a port. It validates the
  * port pointer still exists so the caller doesn't require a reference
  */
-enum drm_connector_status drm_dp_mst_detect_port(struct drm_dp_mst_topology_mgr *mgr, struct drm_dp_mst_port *port)
+enum drm_connector_status drm_dp_mst_detect_port(struct drm_connector *connector,
+						 struct drm_dp_mst_topology_mgr *mgr, struct drm_dp_mst_port *port)
 {
 	enum drm_connector_status status = connector_status_disconnected;
 
@@ -2131,6 +2196,10 @@ enum drm_connector_status drm_dp_mst_detect_port(struct drm_dp_mst_topology_mgr
 
 	case DP_PEER_DEVICE_SST_SINK:
 		status = connector_status_connected;
+		/* for logical ports - cache the EDID */
+		if (port->port_num >= 8 && !port->cached_edid) {
+			port->cached_edid = drm_get_edid(connector, &port->aux.ddc);
+		}
 		break;
 	case DP_PEER_DEVICE_DP_LEGACY_CONV:
 		if (port->ldps)
@@ -2162,7 +2231,12 @@ struct edid *drm_dp_mst_get_edid(struct drm_connector *connector, struct drm_dp_
 	if (!port)
 		return NULL;
 
-	edid = drm_get_edid(connector, &port->aux.ddc);
+	if (port->cached_edid)
+		edid = drm_edid_duplicate(port->cached_edid);
+	else
+		edid = drm_get_edid(connector, &port->aux.ddc);
+
+	drm_mode_connector_set_tile_property(connector);
 	drm_dp_put_port(port);
 	return edid;
 }
@@ -2474,7 +2548,7 @@ void drm_dp_mst_dump_topology(struct seq_file *m,
 	mutex_unlock(&mgr->lock);
 
 	mutex_lock(&mgr->payload_lock);
-	seq_printf(m, "vcpi: %lx\n", mgr->payload_mask);
+	seq_printf(m, "vcpi: %lx %lx\n", mgr->payload_mask, mgr->vcpi_mask);
 
 	for (i = 0; i < mgr->max_payloads; i++) {
 		if (mgr->proposed_vcpis[i]) {
diff --git a/drivers/gpu/drm/drm_drv.c b/drivers/gpu/drm/drm_drv.c
index 3242e208c0d0..4f41377b0b80 100644
--- a/drivers/gpu/drm/drm_drv.c
+++ b/drivers/gpu/drm/drm_drv.c
@@ -35,32 +35,20 @@
 #include <drm/drmP.h>
 #include <drm/drm_core.h>
 #include "drm_legacy.h"
+#include "drm_internal.h"
 
 unsigned int drm_debug = 0;	/* 1 to enable debug output */
 EXPORT_SYMBOL(drm_debug);
 
-unsigned int drm_vblank_offdelay = 5000;    /* Default to 5000 msecs. */
-
-unsigned int drm_timestamp_precision = 20;  /* Default to 20 usecs. */
-
-/*
- * Default to use monotonic timestamps for wait-for-vblank and page-flip
- * complete events.
- */
-unsigned int drm_timestamp_monotonic = 1;
-
 MODULE_AUTHOR(CORE_AUTHOR);
 MODULE_DESCRIPTION(CORE_DESC);
 MODULE_LICENSE("GPL and additional rights");
 MODULE_PARM_DESC(debug, "Enable debug output");
-MODULE_PARM_DESC(vblankoffdelay, "Delay until vblank irq auto-disable [msecs]");
+MODULE_PARM_DESC(vblankoffdelay, "Delay until vblank irq auto-disable [msecs] (0: never disable, <0: disable immediately)");
 MODULE_PARM_DESC(timestamp_precision_usec, "Max. error on timestamps [usecs]");
 MODULE_PARM_DESC(timestamp_monotonic, "Use monotonic timestamps");
 
 module_param_named(debug, drm_debug, int, 0600);
-module_param_named(vblankoffdelay, drm_vblank_offdelay, int, 0600);
-module_param_named(timestamp_precision_usec, drm_timestamp_precision, int, 0600);
-module_param_named(timestamp_monotonic, drm_timestamp_monotonic, int, 0600);
 
 static DEFINE_SPINLOCK(drm_minor_lock);
 static struct idr drm_minors_idr;
@@ -68,22 +56,20 @@ static struct idr drm_minors_idr;
 struct class *drm_class;
 static struct dentry *drm_debugfs_root;
 
-int drm_err(const char *func, const char *format, ...)
+void drm_err(const char *format, ...)
 {
 	struct va_format vaf;
 	va_list args;
-	int r;
 
 	va_start(args, format);
 
 	vaf.fmt = format;
 	vaf.va = &args;
 
-	r = printk(KERN_ERR "[" DRM_NAME ":%s] *ERROR* %pV", func, &vaf);
+	printk(KERN_ERR "[" DRM_NAME ":%pf] *ERROR* %pV",
+	       __builtin_return_address(0), &vaf);
 
 	va_end(args);
-
-	return r;
 }
 EXPORT_SYMBOL(drm_err);
 
@@ -102,6 +88,8 @@ void drm_ut_debug_printk(const char *function_name, const char *format, ...)
 }
 EXPORT_SYMBOL(drm_ut_debug_printk);
 
+#define DRM_MAGIC_HASH_ORDER  4  /**< Size of key hash table. Must be power of 2. */
+
 struct drm_master *drm_master_create(struct drm_minor *minor)
 {
 	struct drm_master *master;
@@ -133,7 +121,6 @@ EXPORT_SYMBOL(drm_master_get);
 static void drm_master_destroy(struct kref *kref)
 {
 	struct drm_master *master = container_of(kref, struct drm_master, refcount);
-	struct drm_magic_entry *pt, *next;
 	struct drm_device *dev = master->minor->dev;
 	struct drm_map_list *r_list, *list_temp;
 
@@ -143,7 +130,7 @@ static void drm_master_destroy(struct kref *kref)
 
 	list_for_each_entry_safe(r_list, list_temp, &dev->maplist, head) {
 		if (r_list->master == master) {
-			drm_rmmap_locked(dev, r_list->map);
+			drm_legacy_rmmap_locked(dev, r_list->map);
 			r_list = NULL;
 		}
 	}
@@ -154,12 +141,6 @@ static void drm_master_destroy(struct kref *kref)
 		master->unique_len = 0;
 	}
 
-	list_for_each_entry_safe(pt, next, &master->magicfree, head) {
-		list_del(&pt->head);
-		drm_ht_remove_item(&master->magiclist, &pt->hash_item);
-		kfree(pt);
-	}
-
 	drm_ht_remove(&master->magiclist);
 
 	mutex_unlock(&dev->struct_mutex);
@@ -554,6 +535,8 @@ static void drm_fs_inode_free(struct inode *inode)
  * The initial ref-count of the object is 1. Use drm_dev_ref() and
  * drm_dev_unref() to take and drop further ref-counts.
  *
+ * Note that for purely virtual devices @parent can be NULL.
+ *
  * RETURNS:
  * Pointer to new DRM device, or NULL if out of memory.
  */
@@ -615,7 +598,7 @@ struct drm_device *drm_dev_alloc(struct drm_driver *driver,
 		goto err_ht;
 	}
 
-	if (driver->driver_features & DRIVER_GEM) {
+	if (drm_core_check_feature(dev, DRIVER_GEM)) {
 		ret = drm_gem_init(dev);
 		if (ret) {
 			DRM_ERROR("Cannot initialize graphics execution manager (GEM)\n");
@@ -645,7 +628,7 @@ static void drm_dev_release(struct kref *ref)
 {
 	struct drm_device *dev = container_of(ref, struct drm_device, ref);
 
-	if (dev->driver->driver_features & DRIVER_GEM)
+	if (drm_core_check_feature(dev, DRIVER_GEM))
 		drm_gem_destroy(dev);
 
 	drm_legacy_ctxbitmap_cleanup(dev);
@@ -779,7 +762,7 @@ void drm_dev_unregister(struct drm_device *dev)
 	drm_vblank_cleanup(dev);
 
 	list_for_each_entry_safe(r_list, list_temp, &dev->maplist, head)
-		drm_rmmap(dev, r_list->map);
+		drm_legacy_rmmap(dev, r_list->map);
 
 	drm_minor_unregister(dev, DRM_MINOR_LEGACY);
 	drm_minor_unregister(dev, DRM_MINOR_RENDER);
diff --git a/drivers/gpu/drm/drm_edid.c b/drivers/gpu/drm/drm_edid.c
index 1dbf3bc4c6a3..53bc7a628909 100644
--- a/drivers/gpu/drm/drm_edid.c
+++ b/drivers/gpu/drm/drm_edid.c
@@ -34,6 +34,7 @@
 #include <linux/module.h>
 #include <drm/drmP.h>
 #include <drm/drm_edid.h>
+#include <drm/drm_displayid.h>
 
 #define version_greater(edid, maj, min) \
 	(((edid)->version > (maj)) || \
@@ -632,27 +633,27 @@ static const struct drm_display_mode edid_cea_modes[] = {
 		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC |
 			DRM_MODE_FLAG_INTERLACE),
 	  .vrefresh = 60, .picture_aspect_ratio = HDMI_PICTURE_ASPECT_16_9, },
-	/* 6 - 1440x480i@60Hz */
-	{ DRM_MODE("1440x480i", DRM_MODE_TYPE_DRIVER, 27000, 1440, 1478,
-		   1602, 1716, 0, 480, 488, 494, 525, 0,
+	/* 6 - 720(1440)x480i@60Hz */
+	{ DRM_MODE("720x480i", DRM_MODE_TYPE_DRIVER, 13500, 720, 739,
+		   801, 858, 0, 480, 488, 494, 525, 0,
 		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC |
 			DRM_MODE_FLAG_INTERLACE | DRM_MODE_FLAG_DBLCLK),
 	  .vrefresh = 60, .picture_aspect_ratio = HDMI_PICTURE_ASPECT_4_3, },
-	/* 7 - 1440x480i@60Hz */
-	{ DRM_MODE("1440x480i", DRM_MODE_TYPE_DRIVER, 27000, 1440, 1478,
-		   1602, 1716, 0, 480, 488, 494, 525, 0,
+	/* 7 - 720(1440)x480i@60Hz */
+	{ DRM_MODE("720x480i", DRM_MODE_TYPE_DRIVER, 13500, 720, 739,
+		   801, 858, 0, 480, 488, 494, 525, 0,
 		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC |
 			DRM_MODE_FLAG_INTERLACE | DRM_MODE_FLAG_DBLCLK),
 	  .vrefresh = 60, .picture_aspect_ratio = HDMI_PICTURE_ASPECT_16_9, },
-	/* 8 - 1440x240@60Hz */
-	{ DRM_MODE("1440x240", DRM_MODE_TYPE_DRIVER, 27000, 1440, 1478,
-		   1602, 1716, 0, 240, 244, 247, 262, 0,
+	/* 8 - 720(1440)x240@60Hz */
+	{ DRM_MODE("720x240", DRM_MODE_TYPE_DRIVER, 13500, 720, 739,
+		   801, 858, 0, 240, 244, 247, 262, 0,
 		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC |
 			DRM_MODE_FLAG_DBLCLK),
 	  .vrefresh = 60, .picture_aspect_ratio = HDMI_PICTURE_ASPECT_4_3, },
-	/* 9 - 1440x240@60Hz */
-	{ DRM_MODE("1440x240", DRM_MODE_TYPE_DRIVER, 27000, 1440, 1478,
-		   1602, 1716, 0, 240, 244, 247, 262, 0,
+	/* 9 - 720(1440)x240@60Hz */
+	{ DRM_MODE("720x240", DRM_MODE_TYPE_DRIVER, 13500, 720, 739,
+		   801, 858, 0, 240, 244, 247, 262, 0,
 		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC |
 			DRM_MODE_FLAG_DBLCLK),
 	  .vrefresh = 60, .picture_aspect_ratio = HDMI_PICTURE_ASPECT_16_9, },
@@ -714,27 +715,27 @@ static const struct drm_display_mode edid_cea_modes[] = {
 		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC |
 			DRM_MODE_FLAG_INTERLACE),
 	  .vrefresh = 50, .picture_aspect_ratio = HDMI_PICTURE_ASPECT_16_9, },
-	/* 21 - 1440x576i@50Hz */
-	{ DRM_MODE("1440x576i", DRM_MODE_TYPE_DRIVER, 27000, 1440, 1464,
-		   1590, 1728, 0, 576, 580, 586, 625, 0,
+	/* 21 - 720(1440)x576i@50Hz */
+	{ DRM_MODE("720x576i", DRM_MODE_TYPE_DRIVER, 13500, 720, 732,
+		   795, 864, 0, 576, 580, 586, 625, 0,
 		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC |
 			DRM_MODE_FLAG_INTERLACE | DRM_MODE_FLAG_DBLCLK),
 	  .vrefresh = 50, .picture_aspect_ratio = HDMI_PICTURE_ASPECT_4_3, },
-	/* 22 - 1440x576i@50Hz */
-	{ DRM_MODE("1440x576i", DRM_MODE_TYPE_DRIVER, 27000, 1440, 1464,
-		   1590, 1728, 0, 576, 580, 586, 625, 0,
+	/* 22 - 720(1440)x576i@50Hz */
+	{ DRM_MODE("720x576i", DRM_MODE_TYPE_DRIVER, 13500, 720, 732,
+		   795, 864, 0, 576, 580, 586, 625, 0,
 		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC |
 			DRM_MODE_FLAG_INTERLACE | DRM_MODE_FLAG_DBLCLK),
 	  .vrefresh = 50, .picture_aspect_ratio = HDMI_PICTURE_ASPECT_16_9, },
-	/* 23 - 1440x288@50Hz */
-	{ DRM_MODE("1440x288", DRM_MODE_TYPE_DRIVER, 27000, 1440, 1464,
-		   1590, 1728, 0, 288, 290, 293, 312, 0,
+	/* 23 - 720(1440)x288@50Hz */
+	{ DRM_MODE("720x288", DRM_MODE_TYPE_DRIVER, 13500, 720, 732,
+		   795, 864, 0, 288, 290, 293, 312, 0,
 		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC |
 			DRM_MODE_FLAG_DBLCLK),
 	  .vrefresh = 50, .picture_aspect_ratio = HDMI_PICTURE_ASPECT_4_3, },
-	/* 24 - 1440x288@50Hz */
-	{ DRM_MODE("1440x288", DRM_MODE_TYPE_DRIVER, 27000, 1440, 1464,
-		   1590, 1728, 0, 288, 290, 293, 312, 0,
+	/* 24 - 720(1440)x288@50Hz */
+	{ DRM_MODE("720x288", DRM_MODE_TYPE_DRIVER, 13500, 720, 732,
+		   795, 864, 0, 288, 290, 293, 312, 0,
 		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC |
 			DRM_MODE_FLAG_DBLCLK),
 	  .vrefresh = 50, .picture_aspect_ratio = HDMI_PICTURE_ASPECT_16_9, },
@@ -837,17 +838,17 @@ static const struct drm_display_mode edid_cea_modes[] = {
 		   796, 864, 0, 576, 581, 586, 625, 0,
 		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC),
 	  .vrefresh = 100, .picture_aspect_ratio = HDMI_PICTURE_ASPECT_16_9, },
-	/* 44 - 1440x576i@100Hz */
-	{ DRM_MODE("1440x576", DRM_MODE_TYPE_DRIVER, 54000, 1440, 1464,
-		   1590, 1728, 0, 576, 580, 586, 625, 0,
+	/* 44 - 720(1440)x576i@100Hz */
+	{ DRM_MODE("720x576i", DRM_MODE_TYPE_DRIVER, 27000, 720, 732,
+		   795, 864, 0, 576, 580, 586, 625, 0,
 		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC |
-			DRM_MODE_FLAG_DBLCLK),
+			DRM_MODE_FLAG_INTERLACE | DRM_MODE_FLAG_DBLCLK),
 	  .vrefresh = 100, .picture_aspect_ratio = HDMI_PICTURE_ASPECT_4_3, },
-	/* 45 - 1440x576i@100Hz */
-	{ DRM_MODE("1440x576", DRM_MODE_TYPE_DRIVER, 54000, 1440, 1464,
-		   1590, 1728, 0, 576, 580, 586, 625, 0,
+	/* 45 - 720(1440)x576i@100Hz */
+	{ DRM_MODE("720x576i", DRM_MODE_TYPE_DRIVER, 27000, 720, 732,
+		   795, 864, 0, 576, 580, 586, 625, 0,
 		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC |
-			DRM_MODE_FLAG_DBLCLK),
+			DRM_MODE_FLAG_INTERLACE | DRM_MODE_FLAG_DBLCLK),
 	  .vrefresh = 100, .picture_aspect_ratio = HDMI_PICTURE_ASPECT_16_9, },
 	/* 46 - 1920x1080i@120Hz */
 	{ DRM_MODE("1920x1080i", DRM_MODE_TYPE_DRIVER, 148500, 1920, 2008,
@@ -870,15 +871,15 @@ static const struct drm_display_mode edid_cea_modes[] = {
 		   798, 858, 0, 480, 489, 495, 525, 0,
 		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC),
 	  .vrefresh = 120, .picture_aspect_ratio = HDMI_PICTURE_ASPECT_16_9, },
-	/* 50 - 1440x480i@120Hz */
-	{ DRM_MODE("1440x480i", DRM_MODE_TYPE_DRIVER, 54000, 1440, 1478,
-		   1602, 1716, 0, 480, 488, 494, 525, 0,
+	/* 50 - 720(1440)x480i@120Hz */
+	{ DRM_MODE("720x480i", DRM_MODE_TYPE_DRIVER, 27000, 720, 739,
+		   801, 858, 0, 480, 488, 494, 525, 0,
 		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC |
 			DRM_MODE_FLAG_INTERLACE | DRM_MODE_FLAG_DBLCLK),
 	  .vrefresh = 120, .picture_aspect_ratio = HDMI_PICTURE_ASPECT_4_3, },
-	/* 51 - 1440x480i@120Hz */
-	{ DRM_MODE("1440x480i", DRM_MODE_TYPE_DRIVER, 54000, 1440, 1478,
-		   1602, 1716, 0, 480, 488, 494, 525, 0,
+	/* 51 - 720(1440)x480i@120Hz */
+	{ DRM_MODE("720x480i", DRM_MODE_TYPE_DRIVER, 27000, 720, 739,
+		   801, 858, 0, 480, 488, 494, 525, 0,
 		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC |
 			DRM_MODE_FLAG_INTERLACE | DRM_MODE_FLAG_DBLCLK),
 	  .vrefresh = 120, .picture_aspect_ratio = HDMI_PICTURE_ASPECT_16_9, },
@@ -892,15 +893,15 @@ static const struct drm_display_mode edid_cea_modes[] = {
 		   796, 864, 0, 576, 581, 586, 625, 0,
 		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC),
 	  .vrefresh = 200, .picture_aspect_ratio = HDMI_PICTURE_ASPECT_16_9, },
-	/* 54 - 1440x576i@200Hz */
-	{ DRM_MODE("1440x576i", DRM_MODE_TYPE_DRIVER, 108000, 1440, 1464,
-		   1590, 1728, 0, 576, 580, 586, 625, 0,
+	/* 54 - 720(1440)x576i@200Hz */
+	{ DRM_MODE("720x576i", DRM_MODE_TYPE_DRIVER, 54000, 720, 732,
+		   795, 864, 0, 576, 580, 586, 625, 0,
 		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC |
 			DRM_MODE_FLAG_INTERLACE | DRM_MODE_FLAG_DBLCLK),
 	  .vrefresh = 200, .picture_aspect_ratio = HDMI_PICTURE_ASPECT_4_3, },
-	/* 55 - 1440x576i@200Hz */
-	{ DRM_MODE("1440x576i", DRM_MODE_TYPE_DRIVER, 108000, 1440, 1464,
-		   1590, 1728, 0, 576, 580, 586, 625, 0,
+	/* 55 - 720(1440)x576i@200Hz */
+	{ DRM_MODE("720x576i", DRM_MODE_TYPE_DRIVER, 54000, 720, 732,
+		   795, 864, 0, 576, 580, 586, 625, 0,
 		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC |
 			DRM_MODE_FLAG_INTERLACE | DRM_MODE_FLAG_DBLCLK),
 	  .vrefresh = 200, .picture_aspect_ratio = HDMI_PICTURE_ASPECT_16_9, },
@@ -914,15 +915,15 @@ static const struct drm_display_mode edid_cea_modes[] = {
 		   798, 858, 0, 480, 489, 495, 525, 0,
 		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC),
 	  .vrefresh = 240, .picture_aspect_ratio = HDMI_PICTURE_ASPECT_16_9, },
-	/* 58 - 1440x480i@240 */
-	{ DRM_MODE("1440x480i", DRM_MODE_TYPE_DRIVER, 108000, 1440, 1478,
-		   1602, 1716, 0, 480, 488, 494, 525, 0,
+	/* 58 - 720(1440)x480i@240 */
+	{ DRM_MODE("720x480i", DRM_MODE_TYPE_DRIVER, 54000, 720, 739,
+		   801, 858, 0, 480, 488, 494, 525, 0,
 		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC |
 			DRM_MODE_FLAG_INTERLACE | DRM_MODE_FLAG_DBLCLK),
 	  .vrefresh = 240, .picture_aspect_ratio = HDMI_PICTURE_ASPECT_4_3, },
-	/* 59 - 1440x480i@240 */
-	{ DRM_MODE("1440x480i", DRM_MODE_TYPE_DRIVER, 108000, 1440, 1478,
-		   1602, 1716, 0, 480, 488, 494, 525, 0,
+	/* 59 - 720(1440)x480i@240 */
+	{ DRM_MODE("720x480i", DRM_MODE_TYPE_DRIVER, 54000, 720, 739,
+		   801, 858, 0, 480, 488, 494, 525, 0,
 		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC |
 			DRM_MODE_FLAG_INTERLACE | DRM_MODE_FLAG_DBLCLK),
 	  .vrefresh = 240, .picture_aspect_ratio = HDMI_PICTURE_ASPECT_16_9, },
@@ -1014,6 +1015,27 @@ module_param_named(edid_fixup, edid_fixup, int, 0400);
 MODULE_PARM_DESC(edid_fixup,
 		 "Minimum number of valid EDID header bytes (0-8, default 6)");
 
+static void drm_get_displayid(struct drm_connector *connector,
+			      struct edid *edid);
+
+static int drm_edid_block_checksum(const u8 *raw_edid)
+{
+	int i;
+	u8 csum = 0;
+	for (i = 0; i < EDID_LENGTH; i++)
+		csum += raw_edid[i];
+
+	return csum;
+}
+
+static bool drm_edid_is_zero(const u8 *in_edid, int length)
+{
+	if (memchr_inv(in_edid, 0, length))
+		return false;
+
+	return true;
+}
+
 /**
  * drm_edid_block_valid - Sanity check the EDID block (base or extension)
  * @raw_edid: pointer to raw EDID block
@@ -1027,8 +1049,7 @@ MODULE_PARM_DESC(edid_fixup,
  */
 bool drm_edid_block_valid(u8 *raw_edid, int block, bool print_bad_edid)
 {
-	int i;
-	u8 csum = 0;
+	u8 csum;
 	struct edid *edid = (struct edid *)raw_edid;
 
 	if (WARN_ON(!raw_edid))
@@ -1048,8 +1069,7 @@ bool drm_edid_block_valid(u8 *raw_edid, int block, bool print_bad_edid)
 		}
 	}
 
-	for (i = 0; i < EDID_LENGTH; i++)
-		csum += raw_edid[i];
+	csum = drm_edid_block_checksum(raw_edid);
 	if (csum) {
 		if (print_bad_edid) {
 			DRM_ERROR("EDID checksum is invalid, remainder is %d\n", csum);
@@ -1080,9 +1100,13 @@ bool drm_edid_block_valid(u8 *raw_edid, int block, bool print_bad_edid)
 
 bad:
 	if (print_bad_edid) {
-		printk(KERN_ERR "Raw EDID:\n");
-		print_hex_dump(KERN_ERR, " \t", DUMP_PREFIX_NONE, 16, 1,
+		if (drm_edid_is_zero(raw_edid, EDID_LENGTH)) {
+			printk(KERN_ERR "EDID block is all zeroes\n");
+		} else {
+			printk(KERN_ERR "Raw EDID:\n");
+			print_hex_dump(KERN_ERR, " \t", DUMP_PREFIX_NONE, 16, 1,
 			       raw_edid, EDID_LENGTH, false);
+		}
 	}
 	return false;
 }
@@ -1115,7 +1139,7 @@ EXPORT_SYMBOL(drm_edid_is_valid);
 #define DDC_SEGMENT_ADDR 0x30
 /**
  * drm_do_probe_ddc_edid() - get EDID information via I2C
- * @adapter: I2C device adaptor
+ * @data: I2C device adapter
  * @buf: EDID data buffer to be filled
  * @block: 128 byte EDID block to start fetching from
  * @len: EDID data buffer length to fetch
@@ -1125,9 +1149,9 @@ EXPORT_SYMBOL(drm_edid_is_valid);
  * Return: 0 on success or -1 on failure.
  */
 static int
-drm_do_probe_ddc_edid(struct i2c_adapter *adapter, unsigned char *buf,
-		      int block, int len)
+drm_do_probe_ddc_edid(void *data, u8 *buf, unsigned int block, size_t len)
 {
+	struct i2c_adapter *adapter = data;
 	unsigned char start = block * EDID_LENGTH;
 	unsigned char segment = block >> 1;
 	unsigned char xfers = segment ? 3 : 2;
@@ -1176,16 +1200,26 @@ drm_do_probe_ddc_edid(struct i2c_adapter *adapter, unsigned char *buf,
 	return ret == xfers ? 0 : -1;
 }
 
-static bool drm_edid_is_zero(u8 *in_edid, int length)
-{
-	if (memchr_inv(in_edid, 0, length))
-		return false;
-
-	return true;
-}
-
-static u8 *
-drm_do_get_edid(struct drm_connector *connector, struct i2c_adapter *adapter)
+/**
+ * drm_do_get_edid - get EDID data using a custom EDID block read function
+ * @connector: connector we're probing
+ * @get_edid_block: EDID block read function
+ * @data: private data passed to the block read function
+ *
+ * When the I2C adapter connected to the DDC bus is hidden behind a device that
+ * exposes a different interface to read EDID blocks this function can be used
+ * to get EDID data using a custom block read function.
+ *
+ * As in the general case the DDC bus is accessible by the kernel at the I2C
+ * level, drivers must make all reasonable efforts to expose it as an I2C
+ * adapter and use drm_get_edid() instead of abusing this function.
+ *
+ * Return: Pointer to valid EDID or NULL if we couldn't find any.
+ */
+struct edid *drm_do_get_edid(struct drm_connector *connector,
+	int (*get_edid_block)(void *data, u8 *buf, unsigned int block,
+			      size_t len),
+	void *data)
 {
 	int i, j = 0, valid_extensions = 0;
 	u8 *block, *new;
@@ -1196,7 +1230,7 @@ drm_do_get_edid(struct drm_connector *connector, struct i2c_adapter *adapter)
 
 	/* base block fetch */
 	for (i = 0; i < 4; i++) {
-		if (drm_do_probe_ddc_edid(adapter, block, 0, EDID_LENGTH))
+		if (get_edid_block(data, block, 0, EDID_LENGTH))
 			goto out;
 		if (drm_edid_block_valid(block, 0, print_bad_edid))
 			break;
@@ -1210,7 +1244,7 @@ drm_do_get_edid(struct drm_connector *connector, struct i2c_adapter *adapter)
 
 	/* if there's no extensions, we're done */
 	if (block[0x7e] == 0)
-		return block;
+		return (struct edid *)block;
 
 	new = krealloc(block, (block[0x7e] + 1) * EDID_LENGTH, GFP_KERNEL);
 	if (!new)
@@ -1219,7 +1253,7 @@ drm_do_get_edid(struct drm_connector *connector, struct i2c_adapter *adapter)
 
 	for (j = 1; j <= block[0x7e]; j++) {
 		for (i = 0; i < 4; i++) {
-			if (drm_do_probe_ddc_edid(adapter,
+			if (get_edid_block(data,
 				  block + (valid_extensions + 1) * EDID_LENGTH,
 				  j, EDID_LENGTH))
 				goto out;
@@ -1247,7 +1281,7 @@ drm_do_get_edid(struct drm_connector *connector, struct i2c_adapter *adapter)
 		block = new;
 	}
 
-	return block;
+	return (struct edid *)block;
 
 carp:
 	if (print_bad_edid) {
@@ -1260,6 +1294,7 @@ out:
 	kfree(block);
 	return NULL;
 }
+EXPORT_SYMBOL_GPL(drm_do_get_edid);
 
 /**
  * drm_probe_ddc() - probe DDC presence
@@ -1289,11 +1324,14 @@ EXPORT_SYMBOL(drm_probe_ddc);
 struct edid *drm_get_edid(struct drm_connector *connector,
 			  struct i2c_adapter *adapter)
 {
-	struct edid *edid = NULL;
+	struct edid *edid;
 
-	if (drm_probe_ddc(adapter))
-		edid = (struct edid *)drm_do_get_edid(connector, adapter);
+	if (!drm_probe_ddc(adapter))
+		return NULL;
 
+	edid = drm_do_get_edid(connector, drm_do_probe_ddc_edid, adapter);
+	if (edid)
+		drm_get_displayid(connector, edid);
 	return edid;
 }
 EXPORT_SYMBOL(drm_get_edid);
@@ -2103,7 +2141,8 @@ static int
 add_inferred_modes(struct drm_connector *connector, struct edid *edid)
 {
 	struct detailed_mode_closure closure = {
-		connector, edid, 0, 0, 0
+		.connector = connector,
+		.edid = edid,
 	};
 
 	if (version_greater(edid, 1, 0))
@@ -2169,7 +2208,8 @@ add_established_modes(struct drm_connector *connector, struct edid *edid)
 		((edid->established_timings.mfg_rsvd & 0x80) << 9);
 	int i, modes = 0;
 	struct detailed_mode_closure closure = {
-		connector, edid, 0, 0, 0
+		.connector = connector,
+		.edid = edid,
 	};
 
 	for (i = 0; i <= EDID_EST_TIMINGS; i++) {
@@ -2227,7 +2267,8 @@ add_standard_modes(struct drm_connector *connector, struct edid *edid)
 {
 	int i, modes = 0;
 	struct detailed_mode_closure closure = {
-		connector, edid, 0, 0, 0
+		.connector = connector,
+		.edid = edid,
 	};
 
 	for (i = 0; i < EDID_STD_TIMINGS; i++) {
@@ -2313,7 +2354,8 @@ static int
 add_cvt_modes(struct drm_connector *connector, struct edid *edid)
 {	
 	struct detailed_mode_closure closure = {
-		connector, edid, 0, 0, 0
+		.connector = connector,
+		.edid = edid,
 	};
 
 	if (version_greater(edid, 1, 2))
@@ -2357,11 +2399,10 @@ add_detailed_modes(struct drm_connector *connector, struct edid *edid,
 		   u32 quirks)
 {
 	struct detailed_mode_closure closure = {
-		connector,
-		edid,
-		1,
-		quirks,
-		0
+		.connector = connector,
+		.edid = edid,
+		.preferred = 1,
+		.quirks = quirks,
 	};
 
 	if (closure.preferred && !version_greater(edid, 1, 3))
@@ -2386,7 +2427,7 @@ add_detailed_modes(struct drm_connector *connector, struct edid *edid,
 /*
  * Search EDID for CEA extension block.
  */
-static u8 *drm_find_cea_extension(struct edid *edid)
+static u8 *drm_find_edid_extension(struct edid *edid, int ext_id)
 {
 	u8 *edid_ext = NULL;
 	int i;
@@ -2398,7 +2439,7 @@ static u8 *drm_find_cea_extension(struct edid *edid)
 	/* Find CEA extension */
 	for (i = 0; i < edid->extensions; i++) {
 		edid_ext = (u8 *)edid + EDID_LENGTH * (i + 1);
-		if (edid_ext[0] == CEA_EXT)
+		if (edid_ext[0] == ext_id)
 			break;
 	}
 
@@ -2408,6 +2449,16 @@ static u8 *drm_find_cea_extension(struct edid *edid)
 	return edid_ext;
 }
 
+static u8 *drm_find_cea_extension(struct edid *edid)
+{
+	return drm_find_edid_extension(edid, CEA_EXT);
+}
+
+static u8 *drm_find_displayid_extension(struct edid *edid)
+{
+	return drm_find_edid_extension(edid, DISPLAYID_EXT);
+}
+
 /*
  * Calculate the alternate clock for the CEA mode
  * (60Hz vs. 59.94Hz etc.)
@@ -3125,9 +3176,12 @@ void drm_edid_to_eld(struct drm_connector *connector, struct edid *edid)
 		}
 	}
 	eld[5] |= sad_count << 4;
-	eld[2] = (20 + mnl + sad_count * 3 + 3) / 4;
 
-	DRM_DEBUG_KMS("ELD size %d, SAD count %d\n", (int)eld[2], sad_count);
+	eld[DRM_ELD_BASELINE_ELD_LEN] =
+		DIV_ROUND_UP(drm_eld_calc_baseline_block_size(eld), 4);
+
+	DRM_DEBUG_KMS("ELD size %d, SAD count %d\n",
+		      drm_eld_size(eld), sad_count);
 }
 EXPORT_SYMBOL(drm_edid_to_eld);
 
@@ -3433,10 +3487,10 @@ EXPORT_SYMBOL(drm_rgb_quant_range_selectable);
 /**
  * drm_assign_hdmi_deep_color_info - detect whether monitor supports
  * hdmi deep color modes and update drm_display_info if so.
- *
  * @edid: monitor EDID information
  * @info: Updated with maximum supported deep color bpc and color format
  *        if deep color supported.
+ * @connector: DRM connector, used only for debug output
  *
  * Parse the CEA extension according to CEA-861-B.
  * Return true if HDMI deep color supported, false if not or unknown.
@@ -3865,3 +3919,123 @@ drm_hdmi_vendor_infoframe_from_display_mode(struct hdmi_vendor_infoframe *frame,
 	return 0;
 }
 EXPORT_SYMBOL(drm_hdmi_vendor_infoframe_from_display_mode);
+
+static int drm_parse_display_id(struct drm_connector *connector,
+				u8 *displayid, int length,
+				bool is_edid_extension)
+{
+	/* if this is an EDID extension the first byte will be 0x70 */
+	int idx = 0;
+	struct displayid_hdr *base;
+	struct displayid_block *block;
+	u8 csum = 0;
+	int i;
+
+	if (is_edid_extension)
+		idx = 1;
+
+	base = (struct displayid_hdr *)&displayid[idx];
+
+	DRM_DEBUG_KMS("base revision 0x%x, length %d, %d %d\n",
+		      base->rev, base->bytes, base->prod_id, base->ext_count);
+
+	if (base->bytes + 5 > length - idx)
+		return -EINVAL;
+
+	for (i = idx; i <= base->bytes + 5; i++) {
+		csum += displayid[i];
+	}
+	if (csum) {
+		DRM_ERROR("DisplayID checksum invalid, remainder is %d\n", csum);
+		return -EINVAL;
+	}
+
+	block = (struct displayid_block *)&displayid[idx + 4];
+	DRM_DEBUG_KMS("block id %d, rev %d, len %d\n",
+		      block->tag, block->rev, block->num_bytes);
+
+	switch (block->tag) {
+	case DATA_BLOCK_TILED_DISPLAY: {
+		struct displayid_tiled_block *tile = (struct displayid_tiled_block *)block;
+
+		u16 w, h;
+		u8 tile_v_loc, tile_h_loc;
+		u8 num_v_tile, num_h_tile;
+		struct drm_tile_group *tg;
+
+		w = tile->tile_size[0] | tile->tile_size[1] << 8;
+		h = tile->tile_size[2] | tile->tile_size[3] << 8;
+
+		num_v_tile = (tile->topo[0] & 0xf) | (tile->topo[2] & 0x30);
+		num_h_tile = (tile->topo[0] >> 4) | ((tile->topo[2] >> 2) & 0x30);
+		tile_v_loc = (tile->topo[1] & 0xf) | ((tile->topo[2] & 0x3) << 4);
+		tile_h_loc = (tile->topo[1] >> 4) | (((tile->topo[2] >> 2) & 0x3) << 4);
+
+		connector->has_tile = true;
+		if (tile->tile_cap & 0x80)
+			connector->tile_is_single_monitor = true;
+
+		connector->num_h_tile = num_h_tile + 1;
+		connector->num_v_tile = num_v_tile + 1;
+		connector->tile_h_loc = tile_h_loc;
+		connector->tile_v_loc = tile_v_loc;
+		connector->tile_h_size = w + 1;
+		connector->tile_v_size = h + 1;
+
+		DRM_DEBUG_KMS("tile cap 0x%x\n", tile->tile_cap);
+		DRM_DEBUG_KMS("tile_size %d x %d\n", w + 1, h + 1);
+		DRM_DEBUG_KMS("topo num tiles %dx%d, location %dx%d\n",
+		       num_h_tile + 1, num_v_tile + 1, tile_h_loc, tile_v_loc);
+		DRM_DEBUG_KMS("vend %c%c%c\n", tile->topology_id[0], tile->topology_id[1], tile->topology_id[2]);
+
+		tg = drm_mode_get_tile_group(connector->dev, tile->topology_id);
+		if (!tg) {
+			tg = drm_mode_create_tile_group(connector->dev, tile->topology_id);
+		}
+		if (!tg)
+			return -ENOMEM;
+
+		if (connector->tile_group != tg) {
+			/* if we haven't got a pointer,
+			   take the reference, drop ref to old tile group */
+			if (connector->tile_group) {
+				drm_mode_put_tile_group(connector->dev, connector->tile_group);
+			}
+			connector->tile_group = tg;
+		} else
+			/* if same tile group, then release the ref we just took. */
+			drm_mode_put_tile_group(connector->dev, tg);
+	}
+		break;
+	default:
+		printk("unknown displayid tag %d\n", block->tag);
+		break;
+	}
+	return 0;
+}
+
+static void drm_get_displayid(struct drm_connector *connector,
+			      struct edid *edid)
+{
+	void *displayid = NULL;
+	int ret;
+	connector->has_tile = false;
+	displayid = drm_find_displayid_extension(edid);
+	if (!displayid) {
+		/* drop reference to any tile group we had */
+		goto out_drop_ref;
+	}
+
+	ret = drm_parse_display_id(connector, displayid, EDID_LENGTH, true);
+	if (ret < 0)
+		goto out_drop_ref;
+	if (!connector->has_tile)
+		goto out_drop_ref;
+	return;
+out_drop_ref:
+	if (connector->tile_group) {
+		drm_mode_put_tile_group(connector->dev, connector->tile_group);
+		connector->tile_group = NULL;
+	}
+	return;
+}
diff --git a/drivers/gpu/drm/drm_edid_load.c b/drivers/gpu/drm/drm_edid_load.c
index 0a235fe61c9b..732cb6f8e653 100644
--- a/drivers/gpu/drm/drm_edid_load.c
+++ b/drivers/gpu/drm/drm_edid_load.c
@@ -254,8 +254,7 @@ static void *edid_load(struct drm_connector *connector, const char *name,
 	    name, connector_name);
 
 out:
-	if (fw)
-		release_firmware(fw);
+	release_firmware(fw);
 	return edid;
 }
 
diff --git a/drivers/gpu/drm/drm_fb_helper.c b/drivers/gpu/drm/drm_fb_helper.c
index 3144db9dc0f1..52ce26d6b4fb 100644
--- a/drivers/gpu/drm/drm_fb_helper.c
+++ b/drivers/gpu/drm/drm_fb_helper.c
@@ -126,7 +126,7 @@ int drm_fb_helper_add_one_connector(struct drm_fb_helper *fb_helper, struct drm_
 
 	WARN_ON(!mutex_is_locked(&fb_helper->dev->mode_config.mutex));
 	if (fb_helper->connector_count + 1 > fb_helper->connector_info_alloc_count) {
-		temp = krealloc(fb_helper->connector_info, sizeof(struct drm_fb_helper_connector) * (fb_helper->connector_count + 1), GFP_KERNEL);
+		temp = krealloc(fb_helper->connector_info, sizeof(struct drm_fb_helper_connector *) * (fb_helper->connector_count + 1), GFP_KERNEL);
 		if (!temp)
 			return -ENOMEM;
 
@@ -171,60 +171,6 @@ int drm_fb_helper_remove_one_connector(struct drm_fb_helper *fb_helper,
 }
 EXPORT_SYMBOL(drm_fb_helper_remove_one_connector);
 
-static int drm_fb_helper_parse_command_line(struct drm_fb_helper *fb_helper)
-{
-	struct drm_fb_helper_connector *fb_helper_conn;
-	int i;
-
-	for (i = 0; i < fb_helper->connector_count; i++) {
-		struct drm_cmdline_mode *mode;
-		struct drm_connector *connector;
-		char *option = NULL;
-
-		fb_helper_conn = fb_helper->connector_info[i];
-		connector = fb_helper_conn->connector;
-		mode = &fb_helper_conn->cmdline_mode;
-
-		/* do something on return - turn off connector maybe */
-		if (fb_get_options(connector->name, &option))
-			continue;
-
-		if (drm_mode_parse_command_line_for_connector(option,
-							      connector,
-							      mode)) {
-			if (mode->force) {
-				const char *s;
-				switch (mode->force) {
-				case DRM_FORCE_OFF:
-					s = "OFF";
-					break;
-				case DRM_FORCE_ON_DIGITAL:
-					s = "ON - dig";
-					break;
-				default:
-				case DRM_FORCE_ON:
-					s = "ON";
-					break;
-				}
-
-				DRM_INFO("forcing %s connector %s\n",
-					 connector->name, s);
-				connector->force = mode->force;
-			}
-
-			DRM_DEBUG_KMS("cmdline mode for connector %s %dx%d@%dHz%s%s%s\n",
-				      connector->name,
-				      mode->xres, mode->yres,
-				      mode->refresh_specified ? mode->refresh : 60,
-				      mode->rb ? " reduced blanking" : "",
-				      mode->margins ? " with margins" : "",
-				      mode->interlace ?  " interlaced" : "");
-		}
-
-	}
-	return 0;
-}
-
 static void drm_fb_helper_save_lut_atomic(struct drm_crtc *crtc, struct drm_fb_helper *helper)
 {
 	uint16_t *r_base, *g_base, *b_base;
@@ -345,10 +291,17 @@ static bool restore_fbdev_mode(struct drm_fb_helper *fb_helper)
 
 	drm_warn_on_modeset_not_all_locked(dev);
 
-	list_for_each_entry(plane, &dev->mode_config.plane_list, head)
+	list_for_each_entry(plane, &dev->mode_config.plane_list, head) {
 		if (plane->type != DRM_PLANE_TYPE_PRIMARY)
 			drm_plane_force_disable(plane);
 
+		if (dev->mode_config.rotation_property) {
+			drm_mode_plane_set_obj_prop(plane,
+						    dev->mode_config.rotation_property,
+						    BIT(DRM_ROTATE_0));
+		}
+	}
+
 	for (i = 0; i < fb_helper->crtc_count; i++) {
 		struct drm_mode_set *mode_set = &fb_helper->crtc_info[i].mode_set;
 		struct drm_crtc *crtc = mode_set->crtc;
@@ -394,9 +347,18 @@ bool drm_fb_helper_restore_fbdev_mode_unlocked(struct drm_fb_helper *fb_helper)
 {
 	struct drm_device *dev = fb_helper->dev;
 	bool ret;
+	bool do_delayed = false;
+
 	drm_modeset_lock_all(dev);
 	ret = restore_fbdev_mode(fb_helper);
+
+	do_delayed = fb_helper->delayed_hotplug;
+	if (do_delayed)
+		fb_helper->delayed_hotplug = false;
 	drm_modeset_unlock_all(dev);
+
+	if (do_delayed)
+		drm_fb_helper_hotplug_event(fb_helper);
 	return ret;
 }
 EXPORT_SYMBOL(drm_fb_helper_restore_fbdev_mode_unlocked);
@@ -419,11 +381,11 @@ static bool drm_fb_helper_force_kernel_mode(void)
 		if (dev->switch_power_state == DRM_SWITCH_POWER_OFF)
 			continue;
 
-		/* NOTE: we use lockless flag below to avoid grabbing other
-		 * modeset locks.  So just trylock the underlying mutex
-		 * directly:
+		/*
+		 * NOTE: Use trylock mode to avoid deadlocks and sleeping in
+		 * panic context.
 		 */
-		if (!mutex_trylock(&dev->mode_config.mutex)) {
+		if (__drm_modeset_lock_all(dev, true) != 0) {
 			error = true;
 			continue;
 		}
@@ -432,7 +394,7 @@ static bool drm_fb_helper_force_kernel_mode(void)
 		if (ret)
 			error = true;
 
-		mutex_unlock(&dev->mode_config.mutex);
+		drm_modeset_unlock_all(dev);
 	}
 	return error;
 }
@@ -935,10 +897,6 @@ int drm_fb_helper_set_par(struct fb_info *info)
 
 	drm_fb_helper_restore_fbdev_mode_unlocked(fb_helper);
 
-	if (fb_helper->delayed_hotplug) {
-		fb_helper->delayed_hotplug = false;
-		drm_fb_helper_hotplug_event(fb_helper);
-	}
 	return 0;
 }
 EXPORT_SYMBOL(drm_fb_helper_set_par);
@@ -1013,7 +971,7 @@ static int drm_fb_helper_single_fb_probe(struct drm_fb_helper *fb_helper,
 		struct drm_fb_helper_connector *fb_helper_conn = fb_helper->connector_info[i];
 		struct drm_cmdline_mode *cmdline_mode;
 
-		cmdline_mode = &fb_helper_conn->cmdline_mode;
+		cmdline_mode = &fb_helper_conn->connector->cmdline_mode;
 
 		if (cmdline_mode->bpp_specified) {
 			switch (cmdline_mode->bpp) {
@@ -1042,19 +1000,21 @@ static int drm_fb_helper_single_fb_probe(struct drm_fb_helper *fb_helper,
 	crtc_count = 0;
 	for (i = 0; i < fb_helper->crtc_count; i++) {
 		struct drm_display_mode *desired_mode;
+		int x, y;
 		desired_mode = fb_helper->crtc_info[i].desired_mode;
-
+		x = fb_helper->crtc_info[i].x;
+		y = fb_helper->crtc_info[i].y;
 		if (desired_mode) {
 			if (gamma_size == 0)
 				gamma_size = fb_helper->crtc_info[i].mode_set.crtc->gamma_size;
-			if (desired_mode->hdisplay < sizes.fb_width)
-				sizes.fb_width = desired_mode->hdisplay;
-			if (desired_mode->vdisplay < sizes.fb_height)
-				sizes.fb_height = desired_mode->vdisplay;
-			if (desired_mode->hdisplay > sizes.surface_width)
-				sizes.surface_width = desired_mode->hdisplay;
-			if (desired_mode->vdisplay > sizes.surface_height)
-				sizes.surface_height = desired_mode->vdisplay;
+			if (desired_mode->hdisplay + x < sizes.fb_width)
+				sizes.fb_width = desired_mode->hdisplay + x;
+			if (desired_mode->vdisplay + y < sizes.fb_height)
+				sizes.fb_height = desired_mode->vdisplay + y;
+			if (desired_mode->hdisplay + x > sizes.surface_width)
+				sizes.surface_width = desired_mode->hdisplay + x;
+			if (desired_mode->vdisplay + y > sizes.surface_height)
+				sizes.surface_height = desired_mode->vdisplay + y;
 			crtc_count++;
 		}
 	}
@@ -1260,9 +1220,7 @@ EXPORT_SYMBOL(drm_has_preferred_mode);
 
 static bool drm_has_cmdline_mode(struct drm_fb_helper_connector *fb_connector)
 {
-	struct drm_cmdline_mode *cmdline_mode;
-	cmdline_mode = &fb_connector->cmdline_mode;
-	return cmdline_mode->specified;
+	return fb_connector->connector->cmdline_mode.specified;
 }
 
 struct drm_display_mode *drm_pick_cmdline_mode(struct drm_fb_helper_connector *fb_helper_conn,
@@ -1272,7 +1230,7 @@ struct drm_display_mode *drm_pick_cmdline_mode(struct drm_fb_helper_connector *f
 	struct drm_display_mode *mode = NULL;
 	bool prefer_non_interlace;
 
-	cmdline_mode = &fb_helper_conn->cmdline_mode;
+	cmdline_mode = &fb_helper_conn->connector->cmdline_mode;
 	if (cmdline_mode->specified == false)
 		return mode;
 
@@ -1356,6 +1314,7 @@ static void drm_enable_connectors(struct drm_fb_helper *fb_helper,
 
 static bool drm_target_cloned(struct drm_fb_helper *fb_helper,
 			      struct drm_display_mode **modes,
+			      struct drm_fb_offset *offsets,
 			      bool *enabled, int width, int height)
 {
 	int count, i, j;
@@ -1427,27 +1386,88 @@ static bool drm_target_cloned(struct drm_fb_helper *fb_helper,
 	return false;
 }
 
+static int drm_get_tile_offsets(struct drm_fb_helper *fb_helper,
+				struct drm_display_mode **modes,
+				struct drm_fb_offset *offsets,
+				int idx,
+				int h_idx, int v_idx)
+{
+	struct drm_fb_helper_connector *fb_helper_conn;
+	int i;
+	int hoffset = 0, voffset = 0;
+
+	for (i = 0; i < fb_helper->connector_count; i++) {
+		fb_helper_conn = fb_helper->connector_info[i];
+		if (!fb_helper_conn->connector->has_tile)
+			continue;
+
+		if (!modes[i] && (h_idx || v_idx)) {
+			DRM_DEBUG_KMS("no modes for connector tiled %d %d\n", i,
+				      fb_helper_conn->connector->base.id);
+			continue;
+		}
+		if (fb_helper_conn->connector->tile_h_loc < h_idx)
+			hoffset += modes[i]->hdisplay;
+
+		if (fb_helper_conn->connector->tile_v_loc < v_idx)
+			voffset += modes[i]->vdisplay;
+	}
+	offsets[idx].x = hoffset;
+	offsets[idx].y = voffset;
+	DRM_DEBUG_KMS("returned %d %d for %d %d\n", hoffset, voffset, h_idx, v_idx);
+	return 0;
+}
+
 static bool drm_target_preferred(struct drm_fb_helper *fb_helper,
 				 struct drm_display_mode **modes,
+				 struct drm_fb_offset *offsets,
 				 bool *enabled, int width, int height)
 {
 	struct drm_fb_helper_connector *fb_helper_conn;
 	int i;
-
+	uint64_t conn_configured = 0, mask;
+	int tile_pass = 0;
+	mask = (1 << fb_helper->connector_count) - 1;
+retry:
 	for (i = 0; i < fb_helper->connector_count; i++) {
 		fb_helper_conn = fb_helper->connector_info[i];
 
-		if (enabled[i] == false)
+		if (conn_configured & (1 << i))
+			continue;
+
+		if (enabled[i] == false) {
+			conn_configured |= (1 << i);
+			continue;
+		}
+
+		/* first pass over all the untiled connectors */
+		if (tile_pass == 0 && fb_helper_conn->connector->has_tile)
 			continue;
 
+		if (tile_pass == 1) {
+			if (fb_helper_conn->connector->tile_h_loc != 0 ||
+			    fb_helper_conn->connector->tile_v_loc != 0)
+				continue;
+
+		} else {
+			if (fb_helper_conn->connector->tile_h_loc != tile_pass -1 &&
+			    fb_helper_conn->connector->tile_v_loc != tile_pass - 1)
+			/* if this tile_pass doesn't cover any of the tiles - keep going */
+				continue;
+
+			/* find the tile offsets for this pass - need
+			   to find all tiles left and above */
+			drm_get_tile_offsets(fb_helper, modes, offsets,
+					     i, fb_helper_conn->connector->tile_h_loc, fb_helper_conn->connector->tile_v_loc);
+		}
 		DRM_DEBUG_KMS("looking for cmdline mode on connector %d\n",
 			      fb_helper_conn->connector->base.id);
 
 		/* got for command line mode first */
 		modes[i] = drm_pick_cmdline_mode(fb_helper_conn, width, height);
 		if (!modes[i]) {
-			DRM_DEBUG_KMS("looking for preferred mode on connector %d\n",
-				      fb_helper_conn->connector->base.id);
+			DRM_DEBUG_KMS("looking for preferred mode on connector %d %d\n",
+				      fb_helper_conn->connector->base.id, fb_helper_conn->connector->tile_group ? fb_helper_conn->connector->tile_group->id : 0);
 			modes[i] = drm_has_preferred_mode(fb_helper_conn, width, height);
 		}
 		/* No preferred modes, pick one off the list */
@@ -1457,6 +1477,12 @@ static bool drm_target_preferred(struct drm_fb_helper *fb_helper,
 		}
 		DRM_DEBUG_KMS("found mode %s\n", modes[i] ? modes[i]->name :
 			  "none");
+		conn_configured |= (1 << i);
+	}
+
+	if ((conn_configured & mask) != mask) {
+		tile_pass++;
+		goto retry;
 	}
 	return true;
 }
@@ -1546,6 +1572,7 @@ static void drm_setup_crtcs(struct drm_fb_helper *fb_helper)
 	struct drm_device *dev = fb_helper->dev;
 	struct drm_fb_helper_crtc **crtcs;
 	struct drm_display_mode **modes;
+	struct drm_fb_offset *offsets;
 	struct drm_mode_set *modeset;
 	bool *enabled;
 	int width, height;
@@ -1560,9 +1587,11 @@ static void drm_setup_crtcs(struct drm_fb_helper *fb_helper)
 			sizeof(struct drm_fb_helper_crtc *), GFP_KERNEL);
 	modes = kcalloc(dev->mode_config.num_connector,
 			sizeof(struct drm_display_mode *), GFP_KERNEL);
+	offsets = kcalloc(dev->mode_config.num_connector,
+			  sizeof(struct drm_fb_offset), GFP_KERNEL);
 	enabled = kcalloc(dev->mode_config.num_connector,
 			  sizeof(bool), GFP_KERNEL);
-	if (!crtcs || !modes || !enabled) {
+	if (!crtcs || !modes || !enabled || !offsets) {
 		DRM_ERROR("Memory allocation failed\n");
 		goto out;
 	}
@@ -1572,14 +1601,16 @@ static void drm_setup_crtcs(struct drm_fb_helper *fb_helper)
 
 	if (!(fb_helper->funcs->initial_config &&
 	      fb_helper->funcs->initial_config(fb_helper, crtcs, modes,
+					       offsets,
 					       enabled, width, height))) {
 		memset(modes, 0, dev->mode_config.num_connector*sizeof(modes[0]));
 		memset(crtcs, 0, dev->mode_config.num_connector*sizeof(crtcs[0]));
+		memset(offsets, 0, dev->mode_config.num_connector*sizeof(offsets[0]));
 
-		if (!drm_target_cloned(fb_helper,
-				       modes, enabled, width, height) &&
-		    !drm_target_preferred(fb_helper,
-					  modes, enabled, width, height))
+		if (!drm_target_cloned(fb_helper, modes, offsets,
+				       enabled, width, height) &&
+		    !drm_target_preferred(fb_helper, modes, offsets,
+					  enabled, width, height))
 			DRM_ERROR("Unable to find initial modes\n");
 
 		DRM_DEBUG_KMS("picking CRTCs for %dx%d config\n",
@@ -1599,18 +1630,23 @@ static void drm_setup_crtcs(struct drm_fb_helper *fb_helper)
 	for (i = 0; i < fb_helper->connector_count; i++) {
 		struct drm_display_mode *mode = modes[i];
 		struct drm_fb_helper_crtc *fb_crtc = crtcs[i];
+		struct drm_fb_offset *offset = &offsets[i];
 		modeset = &fb_crtc->mode_set;
 
 		if (mode && fb_crtc) {
-			DRM_DEBUG_KMS("desired mode %s set on crtc %d\n",
-				      mode->name, fb_crtc->mode_set.crtc->base.id);
+			DRM_DEBUG_KMS("desired mode %s set on crtc %d (%d,%d)\n",
+				      mode->name, fb_crtc->mode_set.crtc->base.id, offset->x, offset->y);
 			fb_crtc->desired_mode = mode;
+			fb_crtc->x = offset->x;
+			fb_crtc->y = offset->y;
 			if (modeset->mode)
 				drm_mode_destroy(dev, modeset->mode);
 			modeset->mode = drm_mode_duplicate(dev,
 							   fb_crtc->desired_mode);
 			modeset->connectors[modeset->num_connectors++] = fb_helper->connector_info[i]->connector;
 			modeset->fb = fb_helper->fb;
+			modeset->x = offset->x;
+			modeset->y = offset->y;
 		}
 	}
 
@@ -1619,7 +1655,6 @@ static void drm_setup_crtcs(struct drm_fb_helper *fb_helper)
 		modeset = &fb_helper->crtc_info[i].mode_set;
 		if (modeset->num_connectors == 0) {
 			BUG_ON(modeset->fb);
-			BUG_ON(modeset->num_connectors);
 			if (modeset->mode)
 				drm_mode_destroy(dev, modeset->mode);
 			modeset->mode = NULL;
@@ -1628,6 +1663,7 @@ static void drm_setup_crtcs(struct drm_fb_helper *fb_helper)
 out:
 	kfree(crtcs);
 	kfree(modes);
+	kfree(offsets);
 	kfree(enabled);
 }
 
@@ -1657,8 +1693,6 @@ bool drm_fb_helper_initial_config(struct drm_fb_helper *fb_helper, int bpp_sel)
 	struct drm_device *dev = fb_helper->dev;
 	int count = 0;
 
-	drm_fb_helper_parse_command_line(fb_helper);
-
 	mutex_lock(&dev->mode_config.mutex);
 	count = drm_fb_helper_probe_connector_modes(fb_helper,
 						    dev->mode_config.max_width,
diff --git a/drivers/gpu/drm/drm_flip_work.c b/drivers/gpu/drm/drm_flip_work.c
index f9c7fa3d0012..43d9b950ef9f 100644
--- a/drivers/gpu/drm/drm_flip_work.c
+++ b/drivers/gpu/drm/drm_flip_work.c
@@ -25,6 +25,44 @@
 #include "drm_flip_work.h"
 
 /**
+ * drm_flip_work_allocate_task - allocate a flip-work task
+ * @data: data associated to the task
+ * @flags: allocator flags
+ *
+ * Allocate a drm_flip_task object and attach private data to it.
+ */
+struct drm_flip_task *drm_flip_work_allocate_task(void *data, gfp_t flags)
+{
+	struct drm_flip_task *task;
+
+	task = kzalloc(sizeof(*task), flags);
+	if (task)
+		task->data = data;
+
+	return task;
+}
+EXPORT_SYMBOL(drm_flip_work_allocate_task);
+
+/**
+ * drm_flip_work_queue_task - queue a specific task
+ * @work: the flip-work
+ * @task: the task to handle
+ *
+ * Queues task, that will later be run (passed back to drm_flip_func_t
+ * func) on a work queue after drm_flip_work_commit() is called.
+ */
+void drm_flip_work_queue_task(struct drm_flip_work *work,
+			      struct drm_flip_task *task)
+{
+	unsigned long flags;
+
+	spin_lock_irqsave(&work->lock, flags);
+	list_add_tail(&task->node, &work->queued);
+	spin_unlock_irqrestore(&work->lock, flags);
+}
+EXPORT_SYMBOL(drm_flip_work_queue_task);
+
+/**
  * drm_flip_work_queue - queue work
  * @work: the flip-work
  * @val: the value to queue
@@ -34,10 +72,14 @@
  */
 void drm_flip_work_queue(struct drm_flip_work *work, void *val)
 {
-	if (kfifo_put(&work->fifo, val)) {
-		atomic_inc(&work->pending);
+	struct drm_flip_task *task;
+
+	task = drm_flip_work_allocate_task(val,
+				drm_can_sleep() ? GFP_KERNEL : GFP_ATOMIC);
+	if (task) {
+		drm_flip_work_queue_task(work, task);
 	} else {
-		DRM_ERROR("%s fifo full!\n", work->name);
+		DRM_ERROR("%s could not allocate task!\n", work->name);
 		work->func(work, val);
 	}
 }
@@ -56,9 +98,12 @@ EXPORT_SYMBOL(drm_flip_work_queue);
 void drm_flip_work_commit(struct drm_flip_work *work,
 		struct workqueue_struct *wq)
 {
-	uint32_t pending = atomic_read(&work->pending);
-	atomic_add(pending, &work->count);
-	atomic_sub(pending, &work->pending);
+	unsigned long flags;
+
+	spin_lock_irqsave(&work->lock, flags);
+	list_splice_tail(&work->queued, &work->commited);
+	INIT_LIST_HEAD(&work->queued);
+	spin_unlock_irqrestore(&work->lock, flags);
 	queue_work(wq, &work->worker);
 }
 EXPORT_SYMBOL(drm_flip_work_commit);
@@ -66,47 +111,46 @@ EXPORT_SYMBOL(drm_flip_work_commit);
 static void flip_worker(struct work_struct *w)
 {
 	struct drm_flip_work *work = container_of(w, struct drm_flip_work, worker);
-	uint32_t count = atomic_read(&work->count);
-	void *val = NULL;
+	struct list_head tasks;
+	unsigned long flags;
+
+	while (1) {
+		struct drm_flip_task *task, *tmp;
+
+		INIT_LIST_HEAD(&tasks);
+		spin_lock_irqsave(&work->lock, flags);
+		list_splice_tail(&work->commited, &tasks);
+		INIT_LIST_HEAD(&work->commited);
+		spin_unlock_irqrestore(&work->lock, flags);
 
-	atomic_sub(count, &work->count);
+		if (list_empty(&tasks))
+			break;
 
-	while(count--)
-		if (!WARN_ON(!kfifo_get(&work->fifo, &val)))
-			work->func(work, val);
+		list_for_each_entry_safe(task, tmp, &tasks, node) {
+			work->func(work, task->data);
+			kfree(task);
+		}
+	}
 }
 
 /**
  * drm_flip_work_init - initialize flip-work
  * @work: the flip-work to initialize
- * @size: the max queue depth
  * @name: debug name
  * @func: the callback work function
  *
  * Initializes/allocates resources for the flip-work
- *
- * RETURNS:
- * Zero on success, error code on failure.
  */
-int drm_flip_work_init(struct drm_flip_work *work, int size,
+void drm_flip_work_init(struct drm_flip_work *work,
 		const char *name, drm_flip_func_t func)
 {
-	int ret;
-
 	work->name = name;
-	atomic_set(&work->count, 0);
-	atomic_set(&work->pending, 0);
+	INIT_LIST_HEAD(&work->queued);
+	INIT_LIST_HEAD(&work->commited);
+	spin_lock_init(&work->lock);
 	work->func = func;
 
-	ret = kfifo_alloc(&work->fifo, size, GFP_KERNEL);
-	if (ret) {
-		DRM_ERROR("could not allocate %s fifo\n", name);
-		return ret;
-	}
-
 	INIT_WORK(&work->worker, flip_worker);
-
-	return 0;
 }
 EXPORT_SYMBOL(drm_flip_work_init);
 
@@ -118,7 +162,6 @@ EXPORT_SYMBOL(drm_flip_work_init);
  */
 void drm_flip_work_cleanup(struct drm_flip_work *work)
 {
-	WARN_ON(!kfifo_is_empty(&work->fifo));
-	kfifo_free(&work->fifo);
+	WARN_ON(!list_empty(&work->queued) || !list_empty(&work->commited));
 }
 EXPORT_SYMBOL(drm_flip_work_cleanup);
diff --git a/drivers/gpu/drm/drm_fops.c b/drivers/gpu/drm/drm_fops.c
index 79d5221c6e41..0b9514b6cd64 100644
--- a/drivers/gpu/drm/drm_fops.c
+++ b/drivers/gpu/drm/drm_fops.c
@@ -39,10 +39,10 @@
 #include <linux/slab.h>
 #include <linux/module.h>
 #include "drm_legacy.h"
+#include "drm_internal.h"
 
 /* from BKL pushdown */
 DEFINE_MUTEX(drm_global_mutex);
-EXPORT_SYMBOL(drm_global_mutex);
 
 static int drm_open_helper(struct file *filp, struct drm_minor *minor);
 
@@ -171,7 +171,7 @@ static int drm_open_helper(struct file *filp, struct drm_minor *minor)
 	init_waitqueue_head(&priv->event_wait);
 	priv->event_space = 4096; /* set aside 4k for event buffer */
 
-	if (dev->driver->driver_features & DRIVER_GEM)
+	if (drm_core_check_feature(dev, DRIVER_GEM))
 		drm_gem_open(dev, priv);
 
 	if (drm_core_check_feature(dev, DRIVER_PRIME))
@@ -256,7 +256,7 @@ out_close:
 out_prime_destroy:
 	if (drm_core_check_feature(dev, DRIVER_PRIME))
 		drm_prime_destroy_file_private(&priv->prime);
-	if (dev->driver->driver_features & DRIVER_GEM)
+	if (drm_core_check_feature(dev, DRIVER_GEM))
 		drm_gem_release(dev, priv);
 	put_pid(priv->pid);
 	kfree(priv);
@@ -268,11 +268,11 @@ static void drm_master_release(struct drm_device *dev, struct file *filp)
 {
 	struct drm_file *file_priv = filp->private_data;
 
-	if (drm_i_have_hw_lock(dev, file_priv)) {
+	if (drm_legacy_i_have_hw_lock(dev, file_priv)) {
 		DRM_DEBUG("File %p released, freeing lock for context %d\n",
 			  filp, _DRM_LOCKING_CONTEXT(file_priv->master->lock.hw_lock->lock));
-		drm_lock_free(&file_priv->master->lock,
-			      _DRM_LOCKING_CONTEXT(file_priv->master->lock.hw_lock->lock));
+		drm_legacy_lock_free(&file_priv->master->lock,
+				     _DRM_LOCKING_CONTEXT(file_priv->master->lock.hw_lock->lock));
 	}
 }
 
@@ -330,8 +330,6 @@ static void drm_legacy_dev_reinit(struct drm_device *dev)
  */
 int drm_lastclose(struct drm_device * dev)
 {
-	struct drm_vma_entry *vma, *vma_temp;
-
 	DRM_DEBUG("\n");
 
 	if (dev->driver->lastclose)
@@ -346,13 +344,7 @@ int drm_lastclose(struct drm_device * dev)
 	drm_agp_clear(dev);
 
 	drm_legacy_sg_cleanup(dev);
-
-	/* Clear vma list (only built for debugging) */
-	list_for_each_entry_safe(vma, vma_temp, &dev->vmalist, head) {
-		list_del(&vma->head);
-		kfree(vma);
-	}
-
+	drm_legacy_vma_flush(dev);
 	drm_legacy_dma_takedown(dev);
 
 	mutex_unlock(&dev->struct_mutex);
@@ -412,14 +404,14 @@ int drm_release(struct inode *inode, struct file *filp)
 		drm_master_release(dev, filp);
 
 	if (drm_core_check_feature(dev, DRIVER_HAVE_DMA))
-		drm_core_reclaim_buffers(dev, file_priv);
+		drm_legacy_reclaim_buffers(dev, file_priv);
 
 	drm_events_release(file_priv);
 
-	if (dev->driver->driver_features & DRIVER_MODESET)
+	if (drm_core_check_feature(dev, DRIVER_MODESET))
 		drm_fb_release(file_priv);
 
-	if (dev->driver->driver_features & DRIVER_GEM)
+	if (drm_core_check_feature(dev, DRIVER_GEM))
 		drm_gem_release(dev, file_priv);
 
 	drm_legacy_ctxbitmap_flush(dev, file_priv);
@@ -464,6 +456,8 @@ int drm_release(struct inode *inode, struct file *filp)
 	if (drm_core_check_feature(dev, DRIVER_PRIME))
 		drm_prime_destroy_file_private(&file_priv->prime);
 
+	WARN_ON(!list_empty(&file_priv->event_list));
+
 	put_pid(file_priv->pid);
 	kfree(file_priv);
 
@@ -521,16 +515,19 @@ ssize_t drm_read(struct file *filp, char __user *buffer,
 	size_t total;
 	ssize_t ret;
 
-	ret = wait_event_interruptible(file_priv->event_wait,
-				       !list_empty(&file_priv->event_list));
-	if (ret < 0)
-		return ret;
+	if ((filp->f_flags & O_NONBLOCK) == 0) {
+		ret = wait_event_interruptible(file_priv->event_wait,
+					       !list_empty(&file_priv->event_list));
+		if (ret < 0)
+			return ret;
+	}
 
 	total = 0;
 	while (drm_dequeue_event(file_priv, total, count, &e)) {
 		if (copy_to_user(buffer + total,
 				 e->event, e->event->length)) {
 			total = -EFAULT;
+			e->destroy(e);
 			break;
 		}
 
@@ -538,7 +535,7 @@ ssize_t drm_read(struct file *filp, char __user *buffer,
 		e->destroy(e);
 	}
 
-	return total;
+	return total ?: -EAGAIN;
 }
 EXPORT_SYMBOL(drm_read);
 
diff --git a/drivers/gpu/drm/drm_gem.c b/drivers/gpu/drm/drm_gem.c
index 6adee4c2afc0..16a164770713 100644
--- a/drivers/gpu/drm/drm_gem.c
+++ b/drivers/gpu/drm/drm_gem.c
@@ -38,6 +38,8 @@
 #include <linux/dma-buf.h>
 #include <drm/drmP.h>
 #include <drm/drm_vma_manager.h>
+#include <drm/drm_gem.h>
+#include "drm_internal.h"
 
 /** @file drm_gem.c
  *
@@ -146,7 +148,7 @@ int drm_gem_object_init(struct drm_device *dev,
 EXPORT_SYMBOL(drm_gem_object_init);
 
 /**
- * drm_gem_object_init - initialize an allocated private GEM object
+ * drm_gem_private_object_init - initialize an allocated private GEM object
  * @dev: drm_device the object should be initialized for
  * @obj: drm_gem_object to initialize
  * @size: object size
@@ -186,7 +188,7 @@ drm_gem_remove_prime_handles(struct drm_gem_object *obj, struct drm_file *filp)
 }
 
 /**
- * drm_gem_object_free - release resources bound to userspace handles
+ * drm_gem_object_handle_free - release resources bound to userspace handles
  * @obj: GEM object to clean up.
  *
  * Called after the last handle to the object has been closed
@@ -307,7 +309,7 @@ EXPORT_SYMBOL(drm_gem_dumb_destroy);
  * drm_gem_handle_create_tail - internal functions to create a handle
  * @file_priv: drm file-private structure to register the handle for
  * @obj: object to register
- * @handlep: pionter to return the created handle to the caller
+ * @handlep: pointer to return the created handle to the caller
  * 
  * This expects the dev->object_name_lock to be held already and will drop it
  * before returning. Used to avoid races in establishing new handles when
@@ -360,7 +362,7 @@ drm_gem_handle_create_tail(struct drm_file *file_priv,
 }
 
 /**
- * gem_handle_create - create a gem handle for an object
+ * drm_gem_handle_create - create a gem handle for an object
  * @file_priv: drm file-private structure to register the handle for
  * @obj: object to register
  * @handlep: pionter to return the created handle to the caller
@@ -369,10 +371,9 @@ drm_gem_handle_create_tail(struct drm_file *file_priv,
  * to the object, which includes a regular reference count. Callers
  * will likely want to dereference the object afterwards.
  */
-int
-drm_gem_handle_create(struct drm_file *file_priv,
-		       struct drm_gem_object *obj,
-		       u32 *handlep)
+int drm_gem_handle_create(struct drm_file *file_priv,
+			  struct drm_gem_object *obj,
+			  u32 *handlep)
 {
 	mutex_lock(&obj->dev->object_name_lock);
 
@@ -579,7 +580,7 @@ drm_gem_close_ioctl(struct drm_device *dev, void *data,
 	struct drm_gem_close *args = data;
 	int ret;
 
-	if (!(dev->driver->driver_features & DRIVER_GEM))
+	if (!drm_core_check_feature(dev, DRIVER_GEM))
 		return -ENODEV;
 
 	ret = drm_gem_handle_delete(file_priv, args->handle);
@@ -606,7 +607,7 @@ drm_gem_flink_ioctl(struct drm_device *dev, void *data,
 	struct drm_gem_object *obj;
 	int ret;
 
-	if (!(dev->driver->driver_features & DRIVER_GEM))
+	if (!drm_core_check_feature(dev, DRIVER_GEM))
 		return -ENODEV;
 
 	obj = drm_gem_object_lookup(dev, file_priv, args->handle);
@@ -659,7 +660,7 @@ drm_gem_open_ioctl(struct drm_device *dev, void *data,
 	int ret;
 	u32 handle;
 
-	if (!(dev->driver->driver_features & DRIVER_GEM))
+	if (!drm_core_check_feature(dev, DRIVER_GEM))
 		return -ENODEV;
 
 	mutex_lock(&dev->object_name_lock);
@@ -887,7 +888,7 @@ int drm_gem_mmap(struct file *filp, struct vm_area_struct *vma)
 					   vma_pages(vma));
 	if (!node) {
 		mutex_unlock(&dev->struct_mutex);
-		return drm_mmap(filp, vma);
+		return -EINVAL;
 	} else if (!drm_vma_node_is_allowed(node, filp)) {
 		mutex_unlock(&dev->struct_mutex);
 		return -EACCES;
diff --git a/drivers/gpu/drm/drm_gem_cma_helper.c b/drivers/gpu/drm/drm_gem_cma_helper.c
index e467e67af6e7..e419eedf751d 100644
--- a/drivers/gpu/drm/drm_gem_cma_helper.c
+++ b/drivers/gpu/drm/drm_gem_cma_helper.c
@@ -29,18 +29,31 @@
 #include <drm/drm_gem_cma_helper.h>
 #include <drm/drm_vma_manager.h>
 
-/*
+/**
+ * DOC: cma helpers
+ *
+ * The Contiguous Memory Allocator reserves a pool of memory at early boot
+ * that is used to service requests for large blocks of contiguous memory.
+ *
+ * The DRM GEM/CMA helpers use this allocator as a means to provide buffer
+ * objects that are physically contiguous in memory. This is useful for
+ * display drivers that are unable to map scattered buffers via an IOMMU.
+ */
+
+/**
  * __drm_gem_cma_create - Create a GEM CMA object without allocating memory
- * @drm: The drm device
- * @size: The GEM object size
+ * @drm: DRM device
+ * @size: size of the object to allocate
  *
- * This function creates and initializes a GEM CMA object of the given size, but
- * doesn't allocate any memory to back the object.
+ * This function creates and initializes a GEM CMA object of the given size,
+ * but doesn't allocate any memory to back the object.
  *
- * Return a struct drm_gem_cma_object* on success or ERR_PTR values on failure.
+ * Returns:
+ * A struct drm_gem_cma_object * on success or an ERR_PTR()-encoded negative
+ * error code on failure.
  */
 static struct drm_gem_cma_object *
-__drm_gem_cma_create(struct drm_device *drm, unsigned int size)
+__drm_gem_cma_create(struct drm_device *drm, size_t size)
 {
 	struct drm_gem_cma_object *cma_obj;
 	struct drm_gem_object *gem_obj;
@@ -69,14 +82,21 @@ error:
 	return ERR_PTR(ret);
 }
 
-/*
+/**
  * drm_gem_cma_create - allocate an object with the given size
+ * @drm: DRM device
+ * @size: size of the object to allocate
+ *
+ * This function creates a CMA GEM object and allocates a contiguous chunk of
+ * memory as backing store. The backing memory has the writecombine attribute
+ * set.
  *
- * returns a struct drm_gem_cma_object* on success or ERR_PTR values
- * on failure.
+ * Returns:
+ * A struct drm_gem_cma_object * on success or an ERR_PTR()-encoded negative
+ * error code on failure.
  */
 struct drm_gem_cma_object *drm_gem_cma_create(struct drm_device *drm,
-		unsigned int size)
+					      size_t size)
 {
 	struct drm_gem_cma_object *cma_obj;
 	int ret;
@@ -104,17 +124,26 @@ error:
 }
 EXPORT_SYMBOL_GPL(drm_gem_cma_create);
 
-/*
- * drm_gem_cma_create_with_handle - allocate an object with the given
- * size and create a gem handle on it
+/**
+ * drm_gem_cma_create_with_handle - allocate an object with the given size and
+ *     return a GEM handle to it
+ * @file_priv: DRM file-private structure to register the handle for
+ * @drm: DRM device
+ * @size: size of the object to allocate
+ * @handle: return location for the GEM handle
+ *
+ * This function creates a CMA GEM object, allocating a physically contiguous
+ * chunk of memory as backing store. The GEM object is then added to the list
+ * of object associated with the given file and a handle to it is returned.
  *
- * returns a struct drm_gem_cma_object* on success or ERR_PTR values
- * on failure.
+ * Returns:
+ * A struct drm_gem_cma_object * on success or an ERR_PTR()-encoded negative
+ * error code on failure.
  */
-static struct drm_gem_cma_object *drm_gem_cma_create_with_handle(
-		struct drm_file *file_priv,
-		struct drm_device *drm, unsigned int size,
-		unsigned int *handle)
+static struct drm_gem_cma_object *
+drm_gem_cma_create_with_handle(struct drm_file *file_priv,
+			       struct drm_device *drm, size_t size,
+			       uint32_t *handle)
 {
 	struct drm_gem_cma_object *cma_obj;
 	struct drm_gem_object *gem_obj;
@@ -145,16 +174,19 @@ err_handle_create:
 	return ERR_PTR(ret);
 }
 
-/*
- * drm_gem_cma_free_object - (struct drm_driver)->gem_free_object callback
- * function
+/**
+ * drm_gem_cma_free_object - free resources associated with a CMA GEM object
+ * @gem_obj: GEM object to free
+ *
+ * This function frees the backing memory of the CMA GEM object, cleans up the
+ * GEM object state and frees the memory used to store the object itself.
+ * Drivers using the CMA helpers should set this as their DRM driver's
+ * ->gem_free_object() callback.
  */
 void drm_gem_cma_free_object(struct drm_gem_object *gem_obj)
 {
 	struct drm_gem_cma_object *cma_obj;
 
-	drm_gem_free_mmap_offset(gem_obj);
-
 	cma_obj = to_drm_gem_cma_obj(gem_obj);
 
 	if (cma_obj->vaddr) {
@@ -170,18 +202,26 @@ void drm_gem_cma_free_object(struct drm_gem_object *gem_obj)
 }
 EXPORT_SYMBOL_GPL(drm_gem_cma_free_object);
 
-/*
- * drm_gem_cma_dumb_create - (struct drm_driver)->dumb_create callback
- * function
+/**
+ * drm_gem_cma_dumb_create_internal - create a dumb buffer object
+ * @file_priv: DRM file-private structure to create the dumb buffer for
+ * @drm: DRM device
+ * @args: IOCTL data
+ *
+ * This aligns the pitch and size arguments to the minimum required. This is
+ * an internal helper that can be wrapped by a driver to account for hardware
+ * with more specific alignment requirements. It should not be used directly
+ * as the ->dumb_create() callback in a DRM driver.
  *
- * This aligns the pitch and size arguments to the minimum required. wrap
- * this into your own function if you need bigger alignment.
+ * Returns:
+ * 0 on success or a negative error code on failure.
  */
-int drm_gem_cma_dumb_create(struct drm_file *file_priv,
-		struct drm_device *dev, struct drm_mode_create_dumb *args)
+int drm_gem_cma_dumb_create_internal(struct drm_file *file_priv,
+				     struct drm_device *drm,
+				     struct drm_mode_create_dumb *args)
 {
+	unsigned int min_pitch = DIV_ROUND_UP(args->width * args->bpp, 8);
 	struct drm_gem_cma_object *cma_obj;
-	int min_pitch = DIV_ROUND_UP(args->width * args->bpp, 8);
 
 	if (args->pitch < min_pitch)
 		args->pitch = min_pitch;
@@ -189,18 +229,63 @@ int drm_gem_cma_dumb_create(struct drm_file *file_priv,
 	if (args->size < args->pitch * args->height)
 		args->size = args->pitch * args->height;
 
-	cma_obj = drm_gem_cma_create_with_handle(file_priv, dev,
-			args->size, &args->handle);
+	cma_obj = drm_gem_cma_create_with_handle(file_priv, drm, args->size,
+						 &args->handle);
+	return PTR_ERR_OR_ZERO(cma_obj);
+}
+EXPORT_SYMBOL_GPL(drm_gem_cma_dumb_create_internal);
+
+/**
+ * drm_gem_cma_dumb_create - create a dumb buffer object
+ * @file_priv: DRM file-private structure to create the dumb buffer for
+ * @drm: DRM device
+ * @args: IOCTL data
+ *
+ * This function computes the pitch of the dumb buffer and rounds it up to an
+ * integer number of bytes per pixel. Drivers for hardware that doesn't have
+ * any additional restrictions on the pitch can directly use this function as
+ * their ->dumb_create() callback.
+ *
+ * For hardware with additional restrictions, drivers can adjust the fields
+ * set up by userspace and pass the IOCTL data along to the
+ * drm_gem_cma_dumb_create_internal() function.
+ *
+ * Returns:
+ * 0 on success or a negative error code on failure.
+ */
+int drm_gem_cma_dumb_create(struct drm_file *file_priv,
+			    struct drm_device *drm,
+			    struct drm_mode_create_dumb *args)
+{
+	struct drm_gem_cma_object *cma_obj;
+
+	args->pitch = DIV_ROUND_UP(args->width * args->bpp, 8);
+	args->size = args->pitch * args->height;
+
+	cma_obj = drm_gem_cma_create_with_handle(file_priv, drm, args->size,
+						 &args->handle);
 	return PTR_ERR_OR_ZERO(cma_obj);
 }
 EXPORT_SYMBOL_GPL(drm_gem_cma_dumb_create);
 
-/*
- * drm_gem_cma_dumb_map_offset - (struct drm_driver)->dumb_map_offset callback
- * function
+/**
+ * drm_gem_cma_dumb_map_offset - return the fake mmap offset for a CMA GEM
+ *     object
+ * @file_priv: DRM file-private structure containing the GEM object
+ * @drm: DRM device
+ * @handle: GEM object handle
+ * @offset: return location for the fake mmap offset
+ *
+ * This function look up an object by its handle and returns the fake mmap
+ * offset associated with it. Drivers using the CMA helpers should set this
+ * as their DRM driver's ->dumb_map_offset() callback.
+ *
+ * Returns:
+ * 0 on success or a negative error code on failure.
  */
 int drm_gem_cma_dumb_map_offset(struct drm_file *file_priv,
-		struct drm_device *drm, uint32_t handle, uint64_t *offset)
+				struct drm_device *drm, u32 handle,
+				u64 *offset)
 {
 	struct drm_gem_object *gem_obj;
 
@@ -208,7 +293,7 @@ int drm_gem_cma_dumb_map_offset(struct drm_file *file_priv,
 
 	gem_obj = drm_gem_object_lookup(drm, file_priv, handle);
 	if (!gem_obj) {
-		dev_err(drm->dev, "failed to lookup gem object\n");
+		dev_err(drm->dev, "failed to lookup GEM object\n");
 		mutex_unlock(&drm->struct_mutex);
 		return -EINVAL;
 	}
@@ -251,8 +336,20 @@ static int drm_gem_cma_mmap_obj(struct drm_gem_cma_object *cma_obj,
 	return ret;
 }
 
-/*
- * drm_gem_cma_mmap - (struct file_operation)->mmap callback function
+/**
+ * drm_gem_cma_mmap - memory-map a CMA GEM object
+ * @filp: file object
+ * @vma: VMA for the area to be mapped
+ *
+ * This function implements an augmented version of the GEM DRM file mmap
+ * operation for CMA objects: In addition to the usual GEM VMA setup it
+ * immediately faults in the entire object instead of using on-demaind
+ * faulting. Drivers which employ the CMA helpers should use this function
+ * as their ->mmap() handler in the DRM device file's file_operations
+ * structure.
+ *
+ * Returns:
+ * 0 on success or a negative error code on failure.
  */
 int drm_gem_cma_mmap(struct file *filp, struct vm_area_struct *vma)
 {
@@ -272,7 +369,16 @@ int drm_gem_cma_mmap(struct file *filp, struct vm_area_struct *vma)
 EXPORT_SYMBOL_GPL(drm_gem_cma_mmap);
 
 #ifdef CONFIG_DEBUG_FS
-void drm_gem_cma_describe(struct drm_gem_cma_object *cma_obj, struct seq_file *m)
+/**
+ * drm_gem_cma_describe - describe a CMA GEM object for debugfs
+ * @cma_obj: CMA GEM object
+ * @m: debugfs file handle
+ *
+ * This function can be used to dump a human-readable representation of the
+ * CMA GEM object into a synthetic file.
+ */
+void drm_gem_cma_describe(struct drm_gem_cma_object *cma_obj,
+			  struct seq_file *m)
 {
 	struct drm_gem_object *obj = &cma_obj->base;
 	struct drm_device *dev = obj->dev;
@@ -291,7 +397,18 @@ void drm_gem_cma_describe(struct drm_gem_cma_object *cma_obj, struct seq_file *m
 EXPORT_SYMBOL_GPL(drm_gem_cma_describe);
 #endif
 
-/* low-level interface prime helpers */
+/**
+ * drm_gem_cma_prime_get_sg_table - provide a scatter/gather table of pinned
+ *     pages for a CMA GEM object
+ * @obj: GEM object
+ *
+ * This function exports a scatter/gather table suitable for PRIME usage by
+ * calling the standard DMA mapping API. Drivers using the CMA helpers should
+ * set this as their DRM driver's ->gem_prime_get_sg_table() callback.
+ *
+ * Returns:
+ * A pointer to the scatter/gather table of pinned pages or NULL on failure.
+ */
 struct sg_table *drm_gem_cma_prime_get_sg_table(struct drm_gem_object *obj)
 {
 	struct drm_gem_cma_object *cma_obj = to_drm_gem_cma_obj(obj);
@@ -315,8 +432,26 @@ out:
 }
 EXPORT_SYMBOL_GPL(drm_gem_cma_prime_get_sg_table);
 
+/**
+ * drm_gem_cma_prime_import_sg_table - produce a CMA GEM object from another
+ *     driver's scatter/gather table of pinned pages
+ * @dev: device to import into
+ * @attach: DMA-BUF attachment
+ * @sgt: scatter/gather table of pinned pages
+ *
+ * This function imports a scatter/gather table exported via DMA-BUF by
+ * another driver. Imported buffers must be physically contiguous in memory
+ * (i.e. the scatter/gather table must contain a single entry). Drivers that
+ * use the CMA helpers should set this as their DRM driver's
+ * ->gem_prime_import_sg_table() callback.
+ *
+ * Returns:
+ * A pointer to a newly created GEM object or an ERR_PTR-encoded negative
+ * error code on failure.
+ */
 struct drm_gem_object *
-drm_gem_cma_prime_import_sg_table(struct drm_device *dev, size_t size,
+drm_gem_cma_prime_import_sg_table(struct drm_device *dev,
+				  struct dma_buf_attachment *attach,
 				  struct sg_table *sgt)
 {
 	struct drm_gem_cma_object *cma_obj;
@@ -325,19 +460,31 @@ drm_gem_cma_prime_import_sg_table(struct drm_device *dev, size_t size,
 		return ERR_PTR(-EINVAL);
 
 	/* Create a CMA GEM buffer. */
-	cma_obj = __drm_gem_cma_create(dev, size);
+	cma_obj = __drm_gem_cma_create(dev, attach->dmabuf->size);
 	if (IS_ERR(cma_obj))
 		return ERR_CAST(cma_obj);
 
 	cma_obj->paddr = sg_dma_address(sgt->sgl);
 	cma_obj->sgt = sgt;
 
-	DRM_DEBUG_PRIME("dma_addr = %pad, size = %zu\n", &cma_obj->paddr, size);
+	DRM_DEBUG_PRIME("dma_addr = %pad, size = %zu\n", &cma_obj->paddr, attach->dmabuf->size);
 
 	return &cma_obj->base;
 }
 EXPORT_SYMBOL_GPL(drm_gem_cma_prime_import_sg_table);
 
+/**
+ * drm_gem_cma_prime_mmap - memory-map an exported CMA GEM object
+ * @obj: GEM object
+ * @vma: VMA for the area to be mapped
+ *
+ * This function maps a buffer imported via DRM PRIME into a userspace
+ * process's address space. Drivers that use the CMA helpers should set this
+ * as their DRM driver's ->gem_prime_mmap() callback.
+ *
+ * Returns:
+ * 0 on success or a negative error code on failure.
+ */
 int drm_gem_cma_prime_mmap(struct drm_gem_object *obj,
 			   struct vm_area_struct *vma)
 {
@@ -356,6 +503,20 @@ int drm_gem_cma_prime_mmap(struct drm_gem_object *obj,
 }
 EXPORT_SYMBOL_GPL(drm_gem_cma_prime_mmap);
 
+/**
+ * drm_gem_cma_prime_vmap - map a CMA GEM object into the kernel's virtual
+ *     address space
+ * @obj: GEM object
+ *
+ * This function maps a buffer exported via DRM PRIME into the kernel's
+ * virtual address space. Since the CMA buffers are already mapped into the
+ * kernel virtual address space this simply returns the cached virtual
+ * address. Drivers using the CMA helpers should set this as their DRM
+ * driver's ->gem_prime_vmap() callback.
+ *
+ * Returns:
+ * The kernel virtual address of the CMA GEM object's backing store.
+ */
 void *drm_gem_cma_prime_vmap(struct drm_gem_object *obj)
 {
 	struct drm_gem_cma_object *cma_obj = to_drm_gem_cma_obj(obj);
@@ -364,6 +525,17 @@ void *drm_gem_cma_prime_vmap(struct drm_gem_object *obj)
 }
 EXPORT_SYMBOL_GPL(drm_gem_cma_prime_vmap);
 
+/**
+ * drm_gem_cma_prime_vunmap - unmap a CMA GEM object from the kernel's virtual
+ *     address space
+ * @obj: GEM object
+ * @vaddr: kernel virtual address where the CMA GEM object was mapped
+ *
+ * This function removes a buffer exported via DRM PRIME from the kernel's
+ * virtual address space. This is a no-op because CMA buffers cannot be
+ * unmapped from kernel space. Drivers using the CMA helpers should set this
+ * as their DRM driver's ->gem_prime_vunmap() callback.
+ */
 void drm_gem_cma_prime_vunmap(struct drm_gem_object *obj, void *vaddr)
 {
 	/* Nothing to do */
diff --git a/drivers/gpu/drm/drm_info.c b/drivers/gpu/drm/drm_info.c
index ecaf0fa2eec8..51efebd434f3 100644
--- a/drivers/gpu/drm/drm_info.c
+++ b/drivers/gpu/drm/drm_info.c
@@ -35,6 +35,9 @@
 
 #include <linux/seq_file.h>
 #include <drm/drmP.h>
+#include <drm/drm_gem.h>
+
+#include "drm_legacy.h"
 
 /**
  * Called when "/proc/dri/.../name" is read.
@@ -183,15 +186,32 @@ int drm_clients_info(struct seq_file *m, void *data)
 	struct drm_device *dev = node->minor->dev;
 	struct drm_file *priv;
 
+	seq_printf(m,
+		   "%20s %5s %3s master a %5s %10s\n",
+		   "command",
+		   "pid",
+		   "dev",
+		   "uid",
+		   "magic");
+
+	/* dev->filelist is sorted youngest first, but we want to present
+	 * oldest first (i.e. kernel, servers, clients), so walk backwardss.
+	 */
 	mutex_lock(&dev->struct_mutex);
-	seq_printf(m, "a dev	pid    uid	magic\n\n");
-	list_for_each_entry(priv, &dev->filelist, lhead) {
-		seq_printf(m, "%c %3d %5d %5d %10u\n",
-			   priv->authenticated ? 'y' : 'n',
-			   priv->minor->index,
+	list_for_each_entry_reverse(priv, &dev->filelist, lhead) {
+		struct task_struct *task;
+
+		rcu_read_lock(); /* locks pid_task()->comm */
+		task = pid_task(priv->pid, PIDTYPE_PID);
+		seq_printf(m, "%20s %5d %3d   %c    %c %5d %10u\n",
+			   task ? task->comm : "<unknown>",
 			   pid_vnr(priv->pid),
+			   priv->minor->index,
+			   priv->is_master ? 'y' : 'n',
+			   priv->authenticated ? 'y' : 'n',
 			   from_kuid_munged(seq_user_ns(m), priv->uid),
 			   priv->magic);
+		rcu_read_unlock();
 	}
 	mutex_unlock(&dev->struct_mutex);
 	return 0;
@@ -223,62 +243,3 @@ int drm_gem_name_info(struct seq_file *m, void *data)
 
 	return 0;
 }
-
-#if DRM_DEBUG_CODE
-
-int drm_vma_info(struct seq_file *m, void *data)
-{
-	struct drm_info_node *node = (struct drm_info_node *) m->private;
-	struct drm_device *dev = node->minor->dev;
-	struct drm_vma_entry *pt;
-	struct vm_area_struct *vma;
-	unsigned long vma_count = 0;
-#if defined(__i386__)
-	unsigned int pgprot;
-#endif
-
-	mutex_lock(&dev->struct_mutex);
-	list_for_each_entry(pt, &dev->vmalist, head)
-		vma_count++;
-
-	seq_printf(m, "vma use count: %lu, high_memory = %pK, 0x%pK\n",
-		   vma_count, high_memory,
-		   (void *)(unsigned long)virt_to_phys(high_memory));
-
-	list_for_each_entry(pt, &dev->vmalist, head) {
-		vma = pt->vma;
-		if (!vma)
-			continue;
-		seq_printf(m,
-			   "\n%5d 0x%pK-0x%pK %c%c%c%c%c%c 0x%08lx000",
-			   pt->pid,
-			   (void *)vma->vm_start, (void *)vma->vm_end,
-			   vma->vm_flags & VM_READ ? 'r' : '-',
-			   vma->vm_flags & VM_WRITE ? 'w' : '-',
-			   vma->vm_flags & VM_EXEC ? 'x' : '-',
-			   vma->vm_flags & VM_MAYSHARE ? 's' : 'p',
-			   vma->vm_flags & VM_LOCKED ? 'l' : '-',
-			   vma->vm_flags & VM_IO ? 'i' : '-',
-			   vma->vm_pgoff);
-
-#if defined(__i386__)
-		pgprot = pgprot_val(vma->vm_page_prot);
-		seq_printf(m, " %c%c%c%c%c%c%c%c%c",
-			   pgprot & _PAGE_PRESENT ? 'p' : '-',
-			   pgprot & _PAGE_RW ? 'w' : 'r',
-			   pgprot & _PAGE_USER ? 'u' : 's',
-			   pgprot & _PAGE_PWT ? 't' : 'b',
-			   pgprot & _PAGE_PCD ? 'u' : 'c',
-			   pgprot & _PAGE_ACCESSED ? 'a' : '-',
-			   pgprot & _PAGE_DIRTY ? 'd' : '-',
-			   pgprot & _PAGE_PSE ? 'm' : 'k',
-			   pgprot & _PAGE_GLOBAL ? 'g' : 'l');
-#endif
-		seq_printf(m, "\n");
-	}
-	mutex_unlock(&dev->struct_mutex);
-	return 0;
-}
-
-#endif
-
diff --git a/drivers/gpu/drm/drm_internal.h b/drivers/gpu/drm/drm_internal.h
new file mode 100644
index 000000000000..7cc0a3516871
--- /dev/null
+++ b/drivers/gpu/drm/drm_internal.h
@@ -0,0 +1,132 @@
+/*
+ * Copyright © 2014 Intel Corporation
+ *   Daniel Vetter <daniel.vetter@ffwll.ch>
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ */
+
+/* drm_irq.c */
+extern unsigned int drm_timestamp_monotonic;
+
+/* drm_fops.c */
+extern struct mutex drm_global_mutex;
+int drm_lastclose(struct drm_device *dev);
+
+/* drm_pci.c */
+int drm_pci_set_unique(struct drm_device *dev,
+		       struct drm_master *master,
+		       struct drm_unique *u);
+int drm_irq_by_busid(struct drm_device *dev, void *data,
+		     struct drm_file *file_priv);
+
+/* drm_vm.c */
+int drm_vma_info(struct seq_file *m, void *data);
+void drm_vm_open_locked(struct drm_device *dev, struct vm_area_struct *vma);
+void drm_vm_close_locked(struct drm_device *dev, struct vm_area_struct *vma);
+
+/* drm_prime.c */
+int drm_prime_handle_to_fd_ioctl(struct drm_device *dev, void *data,
+				 struct drm_file *file_priv);
+int drm_prime_fd_to_handle_ioctl(struct drm_device *dev, void *data,
+				 struct drm_file *file_priv);
+
+void drm_prime_init_file_private(struct drm_prime_file_private *prime_fpriv);
+void drm_prime_destroy_file_private(struct drm_prime_file_private *prime_fpriv);
+void drm_prime_remove_buf_handle_locked(struct drm_prime_file_private *prime_fpriv,
+					struct dma_buf *dma_buf);
+
+/* drm_info.c */
+int drm_name_info(struct seq_file *m, void *data);
+int drm_vm_info(struct seq_file *m, void *data);
+int drm_bufs_info(struct seq_file *m, void *data);
+int drm_vblank_info(struct seq_file *m, void *data);
+int drm_clients_info(struct seq_file *m, void* data);
+int drm_gem_name_info(struct seq_file *m, void *data);
+
+/* drm_irq.c */
+int drm_control(struct drm_device *dev, void *data,
+		struct drm_file *file_priv);
+int drm_modeset_ctl(struct drm_device *dev, void *data,
+		    struct drm_file *file_priv);
+
+/* drm_auth.c */
+int drm_getmagic(struct drm_device *dev, void *data,
+		 struct drm_file *file_priv);
+int drm_authmagic(struct drm_device *dev, void *data,
+		  struct drm_file *file_priv);
+int drm_remove_magic(struct drm_master *master, drm_magic_t magic);
+
+/* drm_sysfs.c */
+extern struct class *drm_class;
+
+struct class *drm_sysfs_create(struct module *owner, char *name);
+void drm_sysfs_destroy(void);
+struct device *drm_sysfs_minor_alloc(struct drm_minor *minor);
+int drm_sysfs_connector_add(struct drm_connector *connector);
+void drm_sysfs_connector_remove(struct drm_connector *connector);
+
+/* drm_gem.c */
+int drm_gem_init(struct drm_device *dev);
+void drm_gem_destroy(struct drm_device *dev);
+int drm_gem_handle_create_tail(struct drm_file *file_priv,
+			       struct drm_gem_object *obj,
+			       u32 *handlep);
+int drm_gem_close_ioctl(struct drm_device *dev, void *data,
+			struct drm_file *file_priv);
+int drm_gem_flink_ioctl(struct drm_device *dev, void *data,
+			struct drm_file *file_priv);
+int drm_gem_open_ioctl(struct drm_device *dev, void *data,
+		       struct drm_file *file_priv);
+void drm_gem_open(struct drm_device *dev, struct drm_file *file_private);
+void drm_gem_release(struct drm_device *dev, struct drm_file *file_private);
+
+/* drm_drv.c */
+int drm_setmaster_ioctl(struct drm_device *dev, void *data,
+			struct drm_file *file_priv);
+int drm_dropmaster_ioctl(struct drm_device *dev, void *data,
+			 struct drm_file *file_priv);
+struct drm_master *drm_master_create(struct drm_minor *minor);
+
+/* drm_debugfs.c */
+#if defined(CONFIG_DEBUG_FS)
+int drm_debugfs_init(struct drm_minor *minor, int minor_id,
+		     struct dentry *root);
+int drm_debugfs_cleanup(struct drm_minor *minor);
+int drm_debugfs_connector_add(struct drm_connector *connector);
+void drm_debugfs_connector_remove(struct drm_connector *connector);
+#else
+static inline int drm_debugfs_init(struct drm_minor *minor, int minor_id,
+				   struct dentry *root)
+{
+	return 0;
+}
+
+static inline int drm_debugfs_cleanup(struct drm_minor *minor)
+{
+	return 0;
+}
+
+static inline int drm_debugfs_connector_add(struct drm_connector *connector)
+{
+	return 0;
+}
+static inline void drm_debugfs_connector_remove(struct drm_connector *connector)
+{
+}
+#endif
diff --git a/drivers/gpu/drm/drm_ioctl.c b/drivers/gpu/drm/drm_ioctl.c
index 40be746b7e68..00587a1e3c83 100644
--- a/drivers/gpu/drm/drm_ioctl.c
+++ b/drivers/gpu/drm/drm_ioctl.c
@@ -31,6 +31,7 @@
 #include <drm/drmP.h>
 #include <drm/drm_core.h>
 #include "drm_legacy.h"
+#include "drm_internal.h"
 
 #include <linux/pci.h>
 #include <linux/export.h>
@@ -41,121 +42,6 @@
 static int drm_version(struct drm_device *dev, void *data,
 		       struct drm_file *file_priv);
 
-#define DRM_IOCTL_DEF(ioctl, _func, _flags) \
-	[DRM_IOCTL_NR(ioctl)] = {.cmd = ioctl, .func = _func, .flags = _flags, .cmd_drv = 0, .name = #ioctl}
-
-/** Ioctl table */
-static const struct drm_ioctl_desc drm_ioctls[] = {
-	DRM_IOCTL_DEF(DRM_IOCTL_VERSION, drm_version, DRM_UNLOCKED|DRM_RENDER_ALLOW),
-	DRM_IOCTL_DEF(DRM_IOCTL_GET_UNIQUE, drm_getunique, 0),
-	DRM_IOCTL_DEF(DRM_IOCTL_GET_MAGIC, drm_getmagic, 0),
-	DRM_IOCTL_DEF(DRM_IOCTL_IRQ_BUSID, drm_irq_by_busid, DRM_MASTER|DRM_ROOT_ONLY),
-	DRM_IOCTL_DEF(DRM_IOCTL_GET_MAP, drm_getmap, DRM_UNLOCKED),
-	DRM_IOCTL_DEF(DRM_IOCTL_GET_CLIENT, drm_getclient, DRM_UNLOCKED),
-	DRM_IOCTL_DEF(DRM_IOCTL_GET_STATS, drm_getstats, DRM_UNLOCKED),
-	DRM_IOCTL_DEF(DRM_IOCTL_GET_CAP, drm_getcap, DRM_UNLOCKED|DRM_RENDER_ALLOW),
-	DRM_IOCTL_DEF(DRM_IOCTL_SET_CLIENT_CAP, drm_setclientcap, 0),
-	DRM_IOCTL_DEF(DRM_IOCTL_SET_VERSION, drm_setversion, DRM_MASTER),
-
-	DRM_IOCTL_DEF(DRM_IOCTL_SET_UNIQUE, drm_setunique, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
-	DRM_IOCTL_DEF(DRM_IOCTL_BLOCK, drm_noop, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
-	DRM_IOCTL_DEF(DRM_IOCTL_UNBLOCK, drm_noop, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
-	DRM_IOCTL_DEF(DRM_IOCTL_AUTH_MAGIC, drm_authmagic, DRM_AUTH|DRM_MASTER),
-
-	DRM_IOCTL_DEF(DRM_IOCTL_ADD_MAP, drm_addmap_ioctl, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
-	DRM_IOCTL_DEF(DRM_IOCTL_RM_MAP, drm_rmmap_ioctl, DRM_AUTH),
-
-	DRM_IOCTL_DEF(DRM_IOCTL_SET_SAREA_CTX, drm_legacy_setsareactx, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
-	DRM_IOCTL_DEF(DRM_IOCTL_GET_SAREA_CTX, drm_legacy_getsareactx, DRM_AUTH),
-
-	DRM_IOCTL_DEF(DRM_IOCTL_SET_MASTER, drm_setmaster_ioctl, DRM_ROOT_ONLY),
-	DRM_IOCTL_DEF(DRM_IOCTL_DROP_MASTER, drm_dropmaster_ioctl, DRM_ROOT_ONLY),
-
-	DRM_IOCTL_DEF(DRM_IOCTL_ADD_CTX, drm_legacy_addctx, DRM_AUTH|DRM_ROOT_ONLY),
-	DRM_IOCTL_DEF(DRM_IOCTL_RM_CTX, drm_legacy_rmctx, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
-	DRM_IOCTL_DEF(DRM_IOCTL_MOD_CTX, drm_noop, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
-	DRM_IOCTL_DEF(DRM_IOCTL_GET_CTX, drm_legacy_getctx, DRM_AUTH),
-	DRM_IOCTL_DEF(DRM_IOCTL_SWITCH_CTX, drm_legacy_switchctx, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
-	DRM_IOCTL_DEF(DRM_IOCTL_NEW_CTX, drm_legacy_newctx, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
-	DRM_IOCTL_DEF(DRM_IOCTL_RES_CTX, drm_legacy_resctx, DRM_AUTH),
-
-	DRM_IOCTL_DEF(DRM_IOCTL_ADD_DRAW, drm_noop, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
-	DRM_IOCTL_DEF(DRM_IOCTL_RM_DRAW, drm_noop, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
-
-	DRM_IOCTL_DEF(DRM_IOCTL_LOCK, drm_lock, DRM_AUTH),
-	DRM_IOCTL_DEF(DRM_IOCTL_UNLOCK, drm_unlock, DRM_AUTH),
-
-	DRM_IOCTL_DEF(DRM_IOCTL_FINISH, drm_noop, DRM_AUTH),
-
-	DRM_IOCTL_DEF(DRM_IOCTL_ADD_BUFS, drm_addbufs, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
-	DRM_IOCTL_DEF(DRM_IOCTL_MARK_BUFS, drm_markbufs, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
-	DRM_IOCTL_DEF(DRM_IOCTL_INFO_BUFS, drm_infobufs, DRM_AUTH),
-	DRM_IOCTL_DEF(DRM_IOCTL_MAP_BUFS, drm_mapbufs, DRM_AUTH),
-	DRM_IOCTL_DEF(DRM_IOCTL_FREE_BUFS, drm_freebufs, DRM_AUTH),
-	DRM_IOCTL_DEF(DRM_IOCTL_DMA, drm_dma_ioctl, DRM_AUTH),
-
-	DRM_IOCTL_DEF(DRM_IOCTL_CONTROL, drm_control, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
-
-#if __OS_HAS_AGP
-	DRM_IOCTL_DEF(DRM_IOCTL_AGP_ACQUIRE, drm_agp_acquire_ioctl, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
-	DRM_IOCTL_DEF(DRM_IOCTL_AGP_RELEASE, drm_agp_release_ioctl, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
-	DRM_IOCTL_DEF(DRM_IOCTL_AGP_ENABLE, drm_agp_enable_ioctl, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
-	DRM_IOCTL_DEF(DRM_IOCTL_AGP_INFO, drm_agp_info_ioctl, DRM_AUTH),
-	DRM_IOCTL_DEF(DRM_IOCTL_AGP_ALLOC, drm_agp_alloc_ioctl, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
-	DRM_IOCTL_DEF(DRM_IOCTL_AGP_FREE, drm_agp_free_ioctl, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
-	DRM_IOCTL_DEF(DRM_IOCTL_AGP_BIND, drm_agp_bind_ioctl, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
-	DRM_IOCTL_DEF(DRM_IOCTL_AGP_UNBIND, drm_agp_unbind_ioctl, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
-#endif
-
-	DRM_IOCTL_DEF(DRM_IOCTL_SG_ALLOC, drm_sg_alloc, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
-	DRM_IOCTL_DEF(DRM_IOCTL_SG_FREE, drm_sg_free, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
-
-	DRM_IOCTL_DEF(DRM_IOCTL_WAIT_VBLANK, drm_wait_vblank, DRM_UNLOCKED),
-
-	DRM_IOCTL_DEF(DRM_IOCTL_MODESET_CTL, drm_modeset_ctl, 0),
-
-	DRM_IOCTL_DEF(DRM_IOCTL_UPDATE_DRAW, drm_noop, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
-
-	DRM_IOCTL_DEF(DRM_IOCTL_GEM_CLOSE, drm_gem_close_ioctl, DRM_UNLOCKED|DRM_RENDER_ALLOW),
-	DRM_IOCTL_DEF(DRM_IOCTL_GEM_FLINK, drm_gem_flink_ioctl, DRM_AUTH|DRM_UNLOCKED),
-	DRM_IOCTL_DEF(DRM_IOCTL_GEM_OPEN, drm_gem_open_ioctl, DRM_AUTH|DRM_UNLOCKED),
-
-	DRM_IOCTL_DEF(DRM_IOCTL_MODE_GETRESOURCES, drm_mode_getresources, DRM_CONTROL_ALLOW|DRM_UNLOCKED),
-
-	DRM_IOCTL_DEF(DRM_IOCTL_PRIME_HANDLE_TO_FD, drm_prime_handle_to_fd_ioctl, DRM_AUTH|DRM_UNLOCKED|DRM_RENDER_ALLOW),
-	DRM_IOCTL_DEF(DRM_IOCTL_PRIME_FD_TO_HANDLE, drm_prime_fd_to_handle_ioctl, DRM_AUTH|DRM_UNLOCKED|DRM_RENDER_ALLOW),
-
-	DRM_IOCTL_DEF(DRM_IOCTL_MODE_GETPLANERESOURCES, drm_mode_getplane_res, DRM_CONTROL_ALLOW|DRM_UNLOCKED),
-	DRM_IOCTL_DEF(DRM_IOCTL_MODE_GETCRTC, drm_mode_getcrtc, DRM_CONTROL_ALLOW|DRM_UNLOCKED),
-	DRM_IOCTL_DEF(DRM_IOCTL_MODE_SETCRTC, drm_mode_setcrtc, DRM_MASTER|DRM_CONTROL_ALLOW|DRM_UNLOCKED),
-	DRM_IOCTL_DEF(DRM_IOCTL_MODE_GETPLANE, drm_mode_getplane, DRM_CONTROL_ALLOW|DRM_UNLOCKED),
-	DRM_IOCTL_DEF(DRM_IOCTL_MODE_SETPLANE, drm_mode_setplane, DRM_MASTER|DRM_CONTROL_ALLOW|DRM_UNLOCKED),
-	DRM_IOCTL_DEF(DRM_IOCTL_MODE_CURSOR, drm_mode_cursor_ioctl, DRM_MASTER|DRM_CONTROL_ALLOW|DRM_UNLOCKED),
-	DRM_IOCTL_DEF(DRM_IOCTL_MODE_GETGAMMA, drm_mode_gamma_get_ioctl, DRM_UNLOCKED),
-	DRM_IOCTL_DEF(DRM_IOCTL_MODE_SETGAMMA, drm_mode_gamma_set_ioctl, DRM_MASTER|DRM_UNLOCKED),
-	DRM_IOCTL_DEF(DRM_IOCTL_MODE_GETENCODER, drm_mode_getencoder, DRM_CONTROL_ALLOW|DRM_UNLOCKED),
-	DRM_IOCTL_DEF(DRM_IOCTL_MODE_GETCONNECTOR, drm_mode_getconnector, DRM_CONTROL_ALLOW|DRM_UNLOCKED),
-	DRM_IOCTL_DEF(DRM_IOCTL_MODE_ATTACHMODE, drm_noop, DRM_MASTER|DRM_CONTROL_ALLOW|DRM_UNLOCKED),
-	DRM_IOCTL_DEF(DRM_IOCTL_MODE_DETACHMODE, drm_noop, DRM_MASTER|DRM_CONTROL_ALLOW|DRM_UNLOCKED),
-	DRM_IOCTL_DEF(DRM_IOCTL_MODE_GETPROPERTY, drm_mode_getproperty_ioctl, DRM_CONTROL_ALLOW|DRM_UNLOCKED),
-	DRM_IOCTL_DEF(DRM_IOCTL_MODE_SETPROPERTY, drm_mode_connector_property_set_ioctl, DRM_MASTER|DRM_CONTROL_ALLOW|DRM_UNLOCKED),
-	DRM_IOCTL_DEF(DRM_IOCTL_MODE_GETPROPBLOB, drm_mode_getblob_ioctl, DRM_CONTROL_ALLOW|DRM_UNLOCKED),
-	DRM_IOCTL_DEF(DRM_IOCTL_MODE_GETFB, drm_mode_getfb, DRM_CONTROL_ALLOW|DRM_UNLOCKED),
-	DRM_IOCTL_DEF(DRM_IOCTL_MODE_ADDFB, drm_mode_addfb, DRM_CONTROL_ALLOW|DRM_UNLOCKED),
-	DRM_IOCTL_DEF(DRM_IOCTL_MODE_ADDFB2, drm_mode_addfb2, DRM_CONTROL_ALLOW|DRM_UNLOCKED),
-	DRM_IOCTL_DEF(DRM_IOCTL_MODE_RMFB, drm_mode_rmfb, DRM_CONTROL_ALLOW|DRM_UNLOCKED),
-	DRM_IOCTL_DEF(DRM_IOCTL_MODE_PAGE_FLIP, drm_mode_page_flip_ioctl, DRM_MASTER|DRM_CONTROL_ALLOW|DRM_UNLOCKED),
-	DRM_IOCTL_DEF(DRM_IOCTL_MODE_DIRTYFB, drm_mode_dirtyfb_ioctl, DRM_MASTER|DRM_CONTROL_ALLOW|DRM_UNLOCKED),
-	DRM_IOCTL_DEF(DRM_IOCTL_MODE_CREATE_DUMB, drm_mode_create_dumb_ioctl, DRM_CONTROL_ALLOW|DRM_UNLOCKED),
-	DRM_IOCTL_DEF(DRM_IOCTL_MODE_MAP_DUMB, drm_mode_mmap_dumb_ioctl, DRM_CONTROL_ALLOW|DRM_UNLOCKED),
-	DRM_IOCTL_DEF(DRM_IOCTL_MODE_DESTROY_DUMB, drm_mode_destroy_dumb_ioctl, DRM_CONTROL_ALLOW|DRM_UNLOCKED),
-	DRM_IOCTL_DEF(DRM_IOCTL_MODE_OBJ_GETPROPERTIES, drm_mode_obj_get_properties_ioctl, DRM_CONTROL_ALLOW|DRM_UNLOCKED),
-	DRM_IOCTL_DEF(DRM_IOCTL_MODE_OBJ_SETPROPERTY, drm_mode_obj_set_property_ioctl, DRM_MASTER|DRM_CONTROL_ALLOW|DRM_UNLOCKED),
-	DRM_IOCTL_DEF(DRM_IOCTL_MODE_CURSOR2, drm_mode_cursor2_ioctl, DRM_MASTER|DRM_CONTROL_ALLOW|DRM_UNLOCKED),
-};
-
-#define DRM_CORE_IOCTL_COUNT	ARRAY_SIZE( drm_ioctls )
-
 /**
  * Get the bus id.
  *
@@ -167,7 +53,7 @@ static const struct drm_ioctl_desc drm_ioctls[] = {
  *
  * Copies the bus id from drm_device::unique into user space.
  */
-int drm_getunique(struct drm_device *dev, void *data,
+static int drm_getunique(struct drm_device *dev, void *data,
 		  struct drm_file *file_priv)
 {
 	struct drm_unique *u = data;
@@ -189,7 +75,6 @@ drm_unset_busid(struct drm_device *dev,
 	kfree(master->unique);
 	master->unique = NULL;
 	master->unique_len = 0;
-	master->unique_size = 0;
 }
 
 /**
@@ -207,7 +92,7 @@ drm_unset_busid(struct drm_device *dev,
  * version 1.1 or greater. Also note that KMS is all version 1.1 and later and
  * UMS was only ever supported on pci devices.
  */
-int drm_setunique(struct drm_device *dev, void *data,
+static int drm_setunique(struct drm_device *dev, void *data,
 		  struct drm_file *file_priv)
 {
 	struct drm_unique *u = data;
@@ -245,15 +130,15 @@ static int drm_set_busid(struct drm_device *dev, struct drm_file *file_priv)
 	if (master->unique != NULL)
 		drm_unset_busid(dev, master);
 
-	if (dev->driver->bus && dev->driver->bus->set_busid) {
-		ret = dev->driver->bus->set_busid(dev, master);
+	if (dev->driver->set_busid) {
+		ret = dev->driver->set_busid(dev, master);
 		if (ret) {
 			drm_unset_busid(dev, master);
 			return ret;
 		}
 	} else {
 		if (WARN(dev->unique == NULL,
-			 "No drm_bus.set_busid() implementation provided by "
+			 "No drm_driver.set_busid() implementation provided by "
 			 "%ps. Use drm_dev_set_unique() to set the unique "
 			 "name explicitly.", dev->driver))
 			return -EINVAL;
@@ -279,7 +164,7 @@ static int drm_set_busid(struct drm_device *dev, struct drm_file *file_priv)
  * Searches for the mapping with the specified offset and copies its information
  * into userspace
  */
-int drm_getmap(struct drm_device *dev, void *data,
+static int drm_getmap(struct drm_device *dev, void *data,
 	       struct drm_file *file_priv)
 {
 	struct drm_map *map = data;
@@ -340,7 +225,7 @@ int drm_getmap(struct drm_device *dev, void *data,
  * Searches for the client with the specified index and copies its information
  * into userspace
  */
-int drm_getclient(struct drm_device *dev, void *data,
+static int drm_getclient(struct drm_device *dev, void *data,
 		  struct drm_file *file_priv)
 {
 	struct drm_client *client = data;
@@ -380,7 +265,7 @@ int drm_getclient(struct drm_device *dev, void *data,
  *
  * \return zero on success or a negative number on failure.
  */
-int drm_getstats(struct drm_device *dev, void *data,
+static int drm_getstats(struct drm_device *dev, void *data,
 		 struct drm_file *file_priv)
 {
 	struct drm_stats *stats = data;
@@ -394,7 +279,7 @@ int drm_getstats(struct drm_device *dev, void *data,
 /**
  * Get device/driver capabilities
  */
-int drm_getcap(struct drm_device *dev, void *data, struct drm_file *file_priv)
+static int drm_getcap(struct drm_device *dev, void *data, struct drm_file *file_priv)
 {
 	struct drm_get_cap *req = data;
 
@@ -444,7 +329,7 @@ int drm_getcap(struct drm_device *dev, void *data, struct drm_file *file_priv)
 /**
  * Set device/driver capabilities
  */
-int
+static int
 drm_setclientcap(struct drm_device *dev, void *data, struct drm_file *file_priv)
 {
 	struct drm_set_client_cap *req = data;
@@ -478,7 +363,7 @@ drm_setclientcap(struct drm_device *dev, void *data, struct drm_file *file_priv)
  *
  * Sets the requested interface version
  */
-int drm_setversion(struct drm_device *dev, void *data, struct drm_file *file_priv)
+static int drm_setversion(struct drm_device *dev, void *data, struct drm_file *file_priv)
 {
 	struct drm_set_version *sv = data;
 	int if_version, retcode = 0;
@@ -624,6 +509,121 @@ static int drm_ioctl_permit(u32 flags, struct drm_file *file_priv)
 	return 0;
 }
 
+#define DRM_IOCTL_DEF(ioctl, _func, _flags) \
+	[DRM_IOCTL_NR(ioctl)] = {.cmd = ioctl, .func = _func, .flags = _flags, .cmd_drv = 0, .name = #ioctl}
+
+/** Ioctl table */
+static const struct drm_ioctl_desc drm_ioctls[] = {
+	DRM_IOCTL_DEF(DRM_IOCTL_VERSION, drm_version, DRM_UNLOCKED|DRM_RENDER_ALLOW),
+	DRM_IOCTL_DEF(DRM_IOCTL_GET_UNIQUE, drm_getunique, 0),
+	DRM_IOCTL_DEF(DRM_IOCTL_GET_MAGIC, drm_getmagic, 0),
+	DRM_IOCTL_DEF(DRM_IOCTL_IRQ_BUSID, drm_irq_by_busid, DRM_MASTER|DRM_ROOT_ONLY),
+	DRM_IOCTL_DEF(DRM_IOCTL_GET_MAP, drm_getmap, DRM_UNLOCKED),
+	DRM_IOCTL_DEF(DRM_IOCTL_GET_CLIENT, drm_getclient, DRM_UNLOCKED),
+	DRM_IOCTL_DEF(DRM_IOCTL_GET_STATS, drm_getstats, DRM_UNLOCKED),
+	DRM_IOCTL_DEF(DRM_IOCTL_GET_CAP, drm_getcap, DRM_UNLOCKED|DRM_RENDER_ALLOW),
+	DRM_IOCTL_DEF(DRM_IOCTL_SET_CLIENT_CAP, drm_setclientcap, 0),
+	DRM_IOCTL_DEF(DRM_IOCTL_SET_VERSION, drm_setversion, DRM_MASTER),
+
+	DRM_IOCTL_DEF(DRM_IOCTL_SET_UNIQUE, drm_setunique, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
+	DRM_IOCTL_DEF(DRM_IOCTL_BLOCK, drm_noop, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
+	DRM_IOCTL_DEF(DRM_IOCTL_UNBLOCK, drm_noop, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
+	DRM_IOCTL_DEF(DRM_IOCTL_AUTH_MAGIC, drm_authmagic, DRM_AUTH|DRM_MASTER),
+
+	DRM_IOCTL_DEF(DRM_IOCTL_ADD_MAP, drm_legacy_addmap_ioctl, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
+	DRM_IOCTL_DEF(DRM_IOCTL_RM_MAP, drm_legacy_rmmap_ioctl, DRM_AUTH),
+
+	DRM_IOCTL_DEF(DRM_IOCTL_SET_SAREA_CTX, drm_legacy_setsareactx, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
+	DRM_IOCTL_DEF(DRM_IOCTL_GET_SAREA_CTX, drm_legacy_getsareactx, DRM_AUTH),
+
+	DRM_IOCTL_DEF(DRM_IOCTL_SET_MASTER, drm_setmaster_ioctl, DRM_ROOT_ONLY),
+	DRM_IOCTL_DEF(DRM_IOCTL_DROP_MASTER, drm_dropmaster_ioctl, DRM_ROOT_ONLY),
+
+	DRM_IOCTL_DEF(DRM_IOCTL_ADD_CTX, drm_legacy_addctx, DRM_AUTH|DRM_ROOT_ONLY),
+	DRM_IOCTL_DEF(DRM_IOCTL_RM_CTX, drm_legacy_rmctx, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
+	DRM_IOCTL_DEF(DRM_IOCTL_MOD_CTX, drm_noop, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
+	DRM_IOCTL_DEF(DRM_IOCTL_GET_CTX, drm_legacy_getctx, DRM_AUTH),
+	DRM_IOCTL_DEF(DRM_IOCTL_SWITCH_CTX, drm_legacy_switchctx, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
+	DRM_IOCTL_DEF(DRM_IOCTL_NEW_CTX, drm_legacy_newctx, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
+	DRM_IOCTL_DEF(DRM_IOCTL_RES_CTX, drm_legacy_resctx, DRM_AUTH),
+
+	DRM_IOCTL_DEF(DRM_IOCTL_ADD_DRAW, drm_noop, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
+	DRM_IOCTL_DEF(DRM_IOCTL_RM_DRAW, drm_noop, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
+
+	DRM_IOCTL_DEF(DRM_IOCTL_LOCK, drm_legacy_lock, DRM_AUTH),
+	DRM_IOCTL_DEF(DRM_IOCTL_UNLOCK, drm_legacy_unlock, DRM_AUTH),
+
+	DRM_IOCTL_DEF(DRM_IOCTL_FINISH, drm_noop, DRM_AUTH),
+
+	DRM_IOCTL_DEF(DRM_IOCTL_ADD_BUFS, drm_legacy_addbufs, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
+	DRM_IOCTL_DEF(DRM_IOCTL_MARK_BUFS, drm_legacy_markbufs, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
+	DRM_IOCTL_DEF(DRM_IOCTL_INFO_BUFS, drm_legacy_infobufs, DRM_AUTH),
+	DRM_IOCTL_DEF(DRM_IOCTL_MAP_BUFS, drm_legacy_mapbufs, DRM_AUTH),
+	DRM_IOCTL_DEF(DRM_IOCTL_FREE_BUFS, drm_legacy_freebufs, DRM_AUTH),
+	DRM_IOCTL_DEF(DRM_IOCTL_DMA, drm_legacy_dma_ioctl, DRM_AUTH),
+
+	DRM_IOCTL_DEF(DRM_IOCTL_CONTROL, drm_control, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
+
+#if __OS_HAS_AGP
+	DRM_IOCTL_DEF(DRM_IOCTL_AGP_ACQUIRE, drm_agp_acquire_ioctl, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
+	DRM_IOCTL_DEF(DRM_IOCTL_AGP_RELEASE, drm_agp_release_ioctl, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
+	DRM_IOCTL_DEF(DRM_IOCTL_AGP_ENABLE, drm_agp_enable_ioctl, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
+	DRM_IOCTL_DEF(DRM_IOCTL_AGP_INFO, drm_agp_info_ioctl, DRM_AUTH),
+	DRM_IOCTL_DEF(DRM_IOCTL_AGP_ALLOC, drm_agp_alloc_ioctl, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
+	DRM_IOCTL_DEF(DRM_IOCTL_AGP_FREE, drm_agp_free_ioctl, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
+	DRM_IOCTL_DEF(DRM_IOCTL_AGP_BIND, drm_agp_bind_ioctl, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
+	DRM_IOCTL_DEF(DRM_IOCTL_AGP_UNBIND, drm_agp_unbind_ioctl, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
+#endif
+
+	DRM_IOCTL_DEF(DRM_IOCTL_SG_ALLOC, drm_legacy_sg_alloc, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
+	DRM_IOCTL_DEF(DRM_IOCTL_SG_FREE, drm_legacy_sg_free, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
+
+	DRM_IOCTL_DEF(DRM_IOCTL_WAIT_VBLANK, drm_wait_vblank, DRM_UNLOCKED),
+
+	DRM_IOCTL_DEF(DRM_IOCTL_MODESET_CTL, drm_modeset_ctl, 0),
+
+	DRM_IOCTL_DEF(DRM_IOCTL_UPDATE_DRAW, drm_noop, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
+
+	DRM_IOCTL_DEF(DRM_IOCTL_GEM_CLOSE, drm_gem_close_ioctl, DRM_UNLOCKED|DRM_RENDER_ALLOW),
+	DRM_IOCTL_DEF(DRM_IOCTL_GEM_FLINK, drm_gem_flink_ioctl, DRM_AUTH|DRM_UNLOCKED),
+	DRM_IOCTL_DEF(DRM_IOCTL_GEM_OPEN, drm_gem_open_ioctl, DRM_AUTH|DRM_UNLOCKED),
+
+	DRM_IOCTL_DEF(DRM_IOCTL_MODE_GETRESOURCES, drm_mode_getresources, DRM_CONTROL_ALLOW|DRM_UNLOCKED),
+
+	DRM_IOCTL_DEF(DRM_IOCTL_PRIME_HANDLE_TO_FD, drm_prime_handle_to_fd_ioctl, DRM_AUTH|DRM_UNLOCKED|DRM_RENDER_ALLOW),
+	DRM_IOCTL_DEF(DRM_IOCTL_PRIME_FD_TO_HANDLE, drm_prime_fd_to_handle_ioctl, DRM_AUTH|DRM_UNLOCKED|DRM_RENDER_ALLOW),
+
+	DRM_IOCTL_DEF(DRM_IOCTL_MODE_GETPLANERESOURCES, drm_mode_getplane_res, DRM_CONTROL_ALLOW|DRM_UNLOCKED),
+	DRM_IOCTL_DEF(DRM_IOCTL_MODE_GETCRTC, drm_mode_getcrtc, DRM_CONTROL_ALLOW|DRM_UNLOCKED),
+	DRM_IOCTL_DEF(DRM_IOCTL_MODE_SETCRTC, drm_mode_setcrtc, DRM_MASTER|DRM_CONTROL_ALLOW|DRM_UNLOCKED),
+	DRM_IOCTL_DEF(DRM_IOCTL_MODE_GETPLANE, drm_mode_getplane, DRM_CONTROL_ALLOW|DRM_UNLOCKED),
+	DRM_IOCTL_DEF(DRM_IOCTL_MODE_SETPLANE, drm_mode_setplane, DRM_MASTER|DRM_CONTROL_ALLOW|DRM_UNLOCKED),
+	DRM_IOCTL_DEF(DRM_IOCTL_MODE_CURSOR, drm_mode_cursor_ioctl, DRM_MASTER|DRM_CONTROL_ALLOW|DRM_UNLOCKED),
+	DRM_IOCTL_DEF(DRM_IOCTL_MODE_GETGAMMA, drm_mode_gamma_get_ioctl, DRM_UNLOCKED),
+	DRM_IOCTL_DEF(DRM_IOCTL_MODE_SETGAMMA, drm_mode_gamma_set_ioctl, DRM_MASTER|DRM_UNLOCKED),
+	DRM_IOCTL_DEF(DRM_IOCTL_MODE_GETENCODER, drm_mode_getencoder, DRM_CONTROL_ALLOW|DRM_UNLOCKED),
+	DRM_IOCTL_DEF(DRM_IOCTL_MODE_GETCONNECTOR, drm_mode_getconnector, DRM_CONTROL_ALLOW|DRM_UNLOCKED),
+	DRM_IOCTL_DEF(DRM_IOCTL_MODE_ATTACHMODE, drm_noop, DRM_MASTER|DRM_CONTROL_ALLOW|DRM_UNLOCKED),
+	DRM_IOCTL_DEF(DRM_IOCTL_MODE_DETACHMODE, drm_noop, DRM_MASTER|DRM_CONTROL_ALLOW|DRM_UNLOCKED),
+	DRM_IOCTL_DEF(DRM_IOCTL_MODE_GETPROPERTY, drm_mode_getproperty_ioctl, DRM_CONTROL_ALLOW|DRM_UNLOCKED),
+	DRM_IOCTL_DEF(DRM_IOCTL_MODE_SETPROPERTY, drm_mode_connector_property_set_ioctl, DRM_MASTER|DRM_CONTROL_ALLOW|DRM_UNLOCKED),
+	DRM_IOCTL_DEF(DRM_IOCTL_MODE_GETPROPBLOB, drm_mode_getblob_ioctl, DRM_CONTROL_ALLOW|DRM_UNLOCKED),
+	DRM_IOCTL_DEF(DRM_IOCTL_MODE_GETFB, drm_mode_getfb, DRM_CONTROL_ALLOW|DRM_UNLOCKED),
+	DRM_IOCTL_DEF(DRM_IOCTL_MODE_ADDFB, drm_mode_addfb, DRM_CONTROL_ALLOW|DRM_UNLOCKED),
+	DRM_IOCTL_DEF(DRM_IOCTL_MODE_ADDFB2, drm_mode_addfb2, DRM_CONTROL_ALLOW|DRM_UNLOCKED),
+	DRM_IOCTL_DEF(DRM_IOCTL_MODE_RMFB, drm_mode_rmfb, DRM_CONTROL_ALLOW|DRM_UNLOCKED),
+	DRM_IOCTL_DEF(DRM_IOCTL_MODE_PAGE_FLIP, drm_mode_page_flip_ioctl, DRM_MASTER|DRM_CONTROL_ALLOW|DRM_UNLOCKED),
+	DRM_IOCTL_DEF(DRM_IOCTL_MODE_DIRTYFB, drm_mode_dirtyfb_ioctl, DRM_MASTER|DRM_CONTROL_ALLOW|DRM_UNLOCKED),
+	DRM_IOCTL_DEF(DRM_IOCTL_MODE_CREATE_DUMB, drm_mode_create_dumb_ioctl, DRM_CONTROL_ALLOW|DRM_UNLOCKED),
+	DRM_IOCTL_DEF(DRM_IOCTL_MODE_MAP_DUMB, drm_mode_mmap_dumb_ioctl, DRM_CONTROL_ALLOW|DRM_UNLOCKED),
+	DRM_IOCTL_DEF(DRM_IOCTL_MODE_DESTROY_DUMB, drm_mode_destroy_dumb_ioctl, DRM_CONTROL_ALLOW|DRM_UNLOCKED),
+	DRM_IOCTL_DEF(DRM_IOCTL_MODE_OBJ_GETPROPERTIES, drm_mode_obj_get_properties_ioctl, DRM_CONTROL_ALLOW|DRM_UNLOCKED),
+	DRM_IOCTL_DEF(DRM_IOCTL_MODE_OBJ_SETPROPERTY, drm_mode_obj_set_property_ioctl, DRM_MASTER|DRM_CONTROL_ALLOW|DRM_UNLOCKED),
+	DRM_IOCTL_DEF(DRM_IOCTL_MODE_CURSOR2, drm_mode_cursor2_ioctl, DRM_MASTER|DRM_CONTROL_ALLOW|DRM_UNLOCKED),
+};
+
+#define DRM_CORE_IOCTL_COUNT	ARRAY_SIZE( drm_ioctls )
+
 /**
  * Called whenever a process performs an ioctl on /dev/drm.
  *
diff --git a/drivers/gpu/drm/drm_irq.c b/drivers/gpu/drm/drm_irq.c
index 08ba1209228e..4d79dad9d44f 100644
--- a/drivers/gpu/drm/drm_irq.c
+++ b/drivers/gpu/drm/drm_irq.c
@@ -34,6 +34,7 @@
 
 #include <drm/drmP.h>
 #include "drm_trace.h"
+#include "drm_internal.h"
 
 #include <linux/interrupt.h>	/* For task queue support */
 #include <linux/slab.h>
@@ -55,12 +56,91 @@
  */
 #define DRM_REDUNDANT_VBLIRQ_THRESH_NS 1000000
 
+static bool
+drm_get_last_vbltimestamp(struct drm_device *dev, int crtc,
+			  struct timeval *tvblank, unsigned flags);
+
+static unsigned int drm_timestamp_precision = 20;  /* Default to 20 usecs. */
+
 /*
- * Clear vblank timestamp buffer for a crtc.
+ * Default to use monotonic timestamps for wait-for-vblank and page-flip
+ * complete events.
  */
-static void clear_vblank_timestamps(struct drm_device *dev, int crtc)
+unsigned int drm_timestamp_monotonic = 1;
+
+static int drm_vblank_offdelay = 5000;    /* Default to 5000 msecs. */
+
+module_param_named(vblankoffdelay, drm_vblank_offdelay, int, 0600);
+module_param_named(timestamp_precision_usec, drm_timestamp_precision, int, 0600);
+module_param_named(timestamp_monotonic, drm_timestamp_monotonic, int, 0600);
+
+/**
+ * drm_update_vblank_count - update the master vblank counter
+ * @dev: DRM device
+ * @crtc: counter to update
+ *
+ * Call back into the driver to update the appropriate vblank counter
+ * (specified by @crtc).  Deal with wraparound, if it occurred, and
+ * update the last read value so we can deal with wraparound on the next
+ * call if necessary.
+ *
+ * Only necessary when going from off->on, to account for frames we
+ * didn't get an interrupt for.
+ *
+ * Note: caller must hold dev->vbl_lock since this reads & writes
+ * device vblank fields.
+ */
+static void drm_update_vblank_count(struct drm_device *dev, int crtc)
 {
-	memset(dev->vblank[crtc].time, 0, sizeof(dev->vblank[crtc].time));
+	struct drm_vblank_crtc *vblank = &dev->vblank[crtc];
+	u32 cur_vblank, diff, tslot;
+	bool rc;
+	struct timeval t_vblank;
+
+	/*
+	 * Interrupts were disabled prior to this call, so deal with counter
+	 * wrap if needed.
+	 * NOTE!  It's possible we lost a full dev->max_vblank_count events
+	 * here if the register is small or we had vblank interrupts off for
+	 * a long time.
+	 *
+	 * We repeat the hardware vblank counter & timestamp query until
+	 * we get consistent results. This to prevent races between gpu
+	 * updating its hardware counter while we are retrieving the
+	 * corresponding vblank timestamp.
+	 */
+	do {
+		cur_vblank = dev->driver->get_vblank_counter(dev, crtc);
+		rc = drm_get_last_vbltimestamp(dev, crtc, &t_vblank, 0);
+	} while (cur_vblank != dev->driver->get_vblank_counter(dev, crtc));
+
+	/* Deal with counter wrap */
+	diff = cur_vblank - vblank->last;
+	if (cur_vblank < vblank->last) {
+		diff += dev->max_vblank_count;
+
+		DRM_DEBUG("last_vblank[%d]=0x%x, cur_vblank=0x%x => diff=0x%x\n",
+			  crtc, vblank->last, cur_vblank, diff);
+	}
+
+	DRM_DEBUG("updating vblank count on crtc %d, missed %d\n",
+		  crtc, diff);
+
+	if (diff == 0)
+		return;
+
+	/* Reinitialize corresponding vblank timestamp if high-precision query
+	 * available. Skip this step if query unsupported or failed. Will
+	 * reinitialize delayed at next vblank interrupt in that case.
+	 */
+	if (rc) {
+		tslot = atomic_read(&vblank->count) + diff;
+		vblanktimestamp(dev, crtc, tslot) = t_vblank;
+	}
+
+	smp_mb__before_atomic();
+	atomic_add(diff, &vblank->count);
+	smp_mb__after_atomic();
 }
 
 /*
@@ -71,10 +151,11 @@ static void clear_vblank_timestamps(struct drm_device *dev, int crtc)
  */
 static void vblank_disable_and_save(struct drm_device *dev, int crtc)
 {
+	struct drm_vblank_crtc *vblank = &dev->vblank[crtc];
 	unsigned long irqflags;
 	u32 vblcount;
 	s64 diff_ns;
-	int vblrc;
+	bool vblrc;
 	struct timeval tvblank;
 	int count = DRM_TIMESTAMP_MAXRETRIES;
 
@@ -84,8 +165,28 @@ static void vblank_disable_and_save(struct drm_device *dev, int crtc)
 	 */
 	spin_lock_irqsave(&dev->vblank_time_lock, irqflags);
 
+	/*
+	 * If the vblank interrupt was already disabled update the count
+	 * and timestamp to maintain the appearance that the counter
+	 * has been ticking all along until this time. This makes the
+	 * count account for the entire time between drm_vblank_on() and
+	 * drm_vblank_off().
+	 *
+	 * But only do this if precise vblank timestamps are available.
+	 * Otherwise we might read a totally bogus timestamp since drivers
+	 * lacking precise timestamp support rely upon sampling the system clock
+	 * at vblank interrupt time. Which obviously won't work out well if the
+	 * vblank interrupt is disabled.
+	 */
+	if (!vblank->enabled &&
+	    drm_get_last_vbltimestamp(dev, crtc, &tvblank, 0)) {
+		drm_update_vblank_count(dev, crtc);
+		spin_unlock_irqrestore(&dev->vblank_time_lock, irqflags);
+		return;
+	}
+
 	dev->driver->disable_vblank(dev, crtc);
-	dev->vblank[crtc].enabled = false;
+	vblank->enabled = false;
 
 	/* No further vblank irq's will be processed after
 	 * this point. Get current hardware vblank count and
@@ -100,9 +201,9 @@ static void vblank_disable_and_save(struct drm_device *dev, int crtc)
 	 * delayed gpu counter increment.
 	 */
 	do {
-		dev->vblank[crtc].last = dev->driver->get_vblank_counter(dev, crtc);
+		vblank->last = dev->driver->get_vblank_counter(dev, crtc);
 		vblrc = drm_get_last_vbltimestamp(dev, crtc, &tvblank, 0);
-	} while (dev->vblank[crtc].last != dev->driver->get_vblank_counter(dev, crtc) && (--count) && vblrc);
+	} while (vblank->last != dev->driver->get_vblank_counter(dev, crtc) && (--count) && vblrc);
 
 	if (!count)
 		vblrc = 0;
@@ -110,7 +211,7 @@ static void vblank_disable_and_save(struct drm_device *dev, int crtc)
 	/* Compute time difference to stored timestamp of last vblank
 	 * as updated by last invocation of drm_handle_vblank() in vblank irq.
 	 */
-	vblcount = atomic_read(&dev->vblank[crtc].count);
+	vblcount = atomic_read(&vblank->count);
 	diff_ns = timeval_to_ns(&tvblank) -
 		  timeval_to_ns(&vblanktimestamp(dev, crtc, vblcount));
 
@@ -126,14 +227,18 @@ static void vblank_disable_and_save(struct drm_device *dev, int crtc)
 	 * available. In that case we can't account for this and just
 	 * hope for the best.
 	 */
-	if ((vblrc > 0) && (abs64(diff_ns) > 1000000)) {
-		atomic_inc(&dev->vblank[crtc].count);
+	if (vblrc && (abs64(diff_ns) > 1000000)) {
+		/* Store new timestamp in ringbuffer. */
+		vblanktimestamp(dev, crtc, vblcount + 1) = tvblank;
+
+		/* Increment cooked vblank count. This also atomically commits
+		 * the timestamp computed above.
+		 */
+		smp_mb__before_atomic();
+		atomic_inc(&vblank->count);
 		smp_mb__after_atomic();
 	}
 
-	/* Invalidate all timestamps while vblank irq's are off. */
-	clear_vblank_timestamps(dev, crtc);
-
 	spin_unlock_irqrestore(&dev->vblank_time_lock, irqflags);
 }
 
@@ -164,14 +269,20 @@ static void vblank_disable_fn(unsigned long arg)
 void drm_vblank_cleanup(struct drm_device *dev)
 {
 	int crtc;
+	unsigned long irqflags;
 
 	/* Bail if the driver didn't call drm_vblank_init() */
 	if (dev->num_crtcs == 0)
 		return;
 
 	for (crtc = 0; crtc < dev->num_crtcs; crtc++) {
-		del_timer_sync(&dev->vblank[crtc].disable_timer);
-		vblank_disable_fn((unsigned long)&dev->vblank[crtc]);
+		struct drm_vblank_crtc *vblank = &dev->vblank[crtc];
+
+		del_timer_sync(&vblank->disable_timer);
+
+		spin_lock_irqsave(&dev->vbl_lock, irqflags);
+		vblank_disable_and_save(dev, crtc);
+		spin_unlock_irqrestore(&dev->vbl_lock, irqflags);
 	}
 
 	kfree(dev->vblank);
@@ -204,11 +315,13 @@ int drm_vblank_init(struct drm_device *dev, int num_crtcs)
 		goto err;
 
 	for (i = 0; i < num_crtcs; i++) {
-		dev->vblank[i].dev = dev;
-		dev->vblank[i].crtc = i;
-		init_waitqueue_head(&dev->vblank[i].queue);
-		setup_timer(&dev->vblank[i].disable_timer, vblank_disable_fn,
-			    (unsigned long)&dev->vblank[i]);
+		struct drm_vblank_crtc *vblank = &dev->vblank[i];
+
+		vblank->dev = dev;
+		vblank->crtc = i;
+		init_waitqueue_head(&vblank->queue);
+		setup_timer(&vblank->disable_timer, vblank_disable_fn,
+			    (unsigned long)vblank);
 	}
 
 	DRM_INFO("Supports vblank timestamp caching Rev 2 (21.10.2013).\n");
@@ -224,7 +337,7 @@ int drm_vblank_init(struct drm_device *dev, int num_crtcs)
 	return 0;
 
 err:
-	drm_vblank_cleanup(dev);
+	dev->num_crtcs = 0;
 	return ret;
 }
 EXPORT_SYMBOL(drm_vblank_init);
@@ -360,9 +473,11 @@ int drm_irq_uninstall(struct drm_device *dev)
 	if (dev->num_crtcs) {
 		spin_lock_irqsave(&dev->vbl_lock, irqflags);
 		for (i = 0; i < dev->num_crtcs; i++) {
-			wake_up(&dev->vblank[i].queue);
-			dev->vblank[i].enabled = false;
-			dev->vblank[i].last =
+			struct drm_vblank_crtc *vblank = &dev->vblank[i];
+
+			wake_up(&vblank->queue);
+			vblank->enabled = false;
+			vblank->last =
 				dev->driver->get_vblank_counter(dev, i);
 		}
 		spin_unlock_irqrestore(&dev->vbl_lock, irqflags);
@@ -617,7 +732,7 @@ int drm_calc_vbltimestamp_from_scanoutpos(struct drm_device *dev, int crtc,
 	 * within vblank area, counting down the number of lines until
 	 * start of scanout.
 	 */
-	invbl = vbl_status & DRM_SCANOUTPOS_INVBL;
+	invbl = vbl_status & DRM_SCANOUTPOS_IN_VBLANK;
 
 	/* Convert scanout position into elapsed time at raw_time query
 	 * since start of scanout at first display scanline. delta_ns
@@ -647,7 +762,7 @@ int drm_calc_vbltimestamp_from_scanoutpos(struct drm_device *dev, int crtc,
 
 	vbl_status = DRM_VBLANKTIME_SCANOUTPOS_METHOD;
 	if (invbl)
-		vbl_status |= DRM_VBLANKTIME_INVBL;
+		vbl_status |= DRM_VBLANKTIME_IN_VBLANK;
 
 	return vbl_status;
 }
@@ -679,10 +794,11 @@ static struct timeval get_drm_timestamp(void)
  * call, i.e., it isn't very precisely locked to the true vblank.
  *
  * Returns:
- * Non-zero if timestamp is considered to be very precise, zero otherwise.
+ * True if timestamp is considered to be very precise, false otherwise.
  */
-u32 drm_get_last_vbltimestamp(struct drm_device *dev, int crtc,
-			      struct timeval *tvblank, unsigned flags)
+static bool
+drm_get_last_vbltimestamp(struct drm_device *dev, int crtc,
+			  struct timeval *tvblank, unsigned flags)
 {
 	int ret;
 
@@ -694,7 +810,7 @@ u32 drm_get_last_vbltimestamp(struct drm_device *dev, int crtc,
 		ret = dev->driver->get_vblank_timestamp(dev, crtc, &max_error,
 							tvblank, flags);
 		if (ret > 0)
-			return (u32) ret;
+			return true;
 	}
 
 	/* GPU high precision timestamp query unsupported or failed.
@@ -702,9 +818,8 @@ u32 drm_get_last_vbltimestamp(struct drm_device *dev, int crtc,
 	 */
 	*tvblank = get_drm_timestamp();
 
-	return 0;
+	return false;
 }
-EXPORT_SYMBOL(drm_get_last_vbltimestamp);
 
 /**
  * drm_vblank_count - retrieve "cooked" vblank counter value
@@ -715,16 +830,41 @@ EXPORT_SYMBOL(drm_get_last_vbltimestamp);
  * vblank events since the system was booted, including lost events due to
  * modesetting activity.
  *
+ * This is the legacy version of drm_crtc_vblank_count().
+ *
  * Returns:
  * The software vblank counter.
  */
 u32 drm_vblank_count(struct drm_device *dev, int crtc)
 {
-	return atomic_read(&dev->vblank[crtc].count);
+	struct drm_vblank_crtc *vblank = &dev->vblank[crtc];
+
+	if (WARN_ON(crtc >= dev->num_crtcs))
+		return 0;
+	return atomic_read(&vblank->count);
 }
 EXPORT_SYMBOL(drm_vblank_count);
 
 /**
+ * drm_crtc_vblank_count - retrieve "cooked" vblank counter value
+ * @crtc: which counter to retrieve
+ *
+ * Fetches the "cooked" vblank count value that represents the number of
+ * vblank events since the system was booted, including lost events due to
+ * modesetting activity.
+ *
+ * This is the native KMS version of drm_vblank_count().
+ *
+ * Returns:
+ * The software vblank counter.
+ */
+u32 drm_crtc_vblank_count(struct drm_crtc *crtc)
+{
+	return drm_vblank_count(crtc->dev, drm_crtc_index(crtc));
+}
+EXPORT_SYMBOL(drm_crtc_vblank_count);
+
+/**
  * drm_vblank_count_and_time - retrieve "cooked" vblank counter value
  * and the system timestamp corresponding to that vblank counter value.
  *
@@ -740,18 +880,22 @@ EXPORT_SYMBOL(drm_vblank_count);
 u32 drm_vblank_count_and_time(struct drm_device *dev, int crtc,
 			      struct timeval *vblanktime)
 {
+	struct drm_vblank_crtc *vblank = &dev->vblank[crtc];
 	u32 cur_vblank;
 
+	if (WARN_ON(crtc >= dev->num_crtcs))
+		return 0;
+
 	/* Read timestamp from slot of _vblank_time ringbuffer
 	 * that corresponds to current vblank count. Retry if
 	 * count has incremented during readout. This works like
 	 * a seqlock.
 	 */
 	do {
-		cur_vblank = atomic_read(&dev->vblank[crtc].count);
+		cur_vblank = atomic_read(&vblank->count);
 		*vblanktime = vblanktimestamp(dev, crtc, cur_vblank);
 		smp_rmb();
-	} while (cur_vblank != atomic_read(&dev->vblank[crtc].count));
+	} while (cur_vblank != atomic_read(&vblank->count));
 
 	return cur_vblank;
 }
@@ -781,6 +925,8 @@ static void send_vblank_event(struct drm_device *dev,
  *
  * Updates sequence # and timestamp on event, and sends it to userspace.
  * Caller must hold event lock.
+ *
+ * This is the legacy version of drm_crtc_send_vblank_event().
  */
 void drm_send_vblank_event(struct drm_device *dev, int crtc,
 		struct drm_pending_vblank_event *e)
@@ -800,68 +946,21 @@ void drm_send_vblank_event(struct drm_device *dev, int crtc,
 EXPORT_SYMBOL(drm_send_vblank_event);
 
 /**
- * drm_update_vblank_count - update the master vblank counter
- * @dev: DRM device
- * @crtc: counter to update
- *
- * Call back into the driver to update the appropriate vblank counter
- * (specified by @crtc).  Deal with wraparound, if it occurred, and
- * update the last read value so we can deal with wraparound on the next
- * call if necessary.
+ * drm_crtc_send_vblank_event - helper to send vblank event after pageflip
+ * @crtc: the source CRTC of the vblank event
+ * @e: the event to send
  *
- * Only necessary when going from off->on, to account for frames we
- * didn't get an interrupt for.
+ * Updates sequence # and timestamp on event, and sends it to userspace.
+ * Caller must hold event lock.
  *
- * Note: caller must hold dev->vbl_lock since this reads & writes
- * device vblank fields.
+ * This is the native KMS version of drm_send_vblank_event().
  */
-static void drm_update_vblank_count(struct drm_device *dev, int crtc)
+void drm_crtc_send_vblank_event(struct drm_crtc *crtc,
+				struct drm_pending_vblank_event *e)
 {
-	u32 cur_vblank, diff, tslot, rc;
-	struct timeval t_vblank;
-
-	/*
-	 * Interrupts were disabled prior to this call, so deal with counter
-	 * wrap if needed.
-	 * NOTE!  It's possible we lost a full dev->max_vblank_count events
-	 * here if the register is small or we had vblank interrupts off for
-	 * a long time.
-	 *
-	 * We repeat the hardware vblank counter & timestamp query until
-	 * we get consistent results. This to prevent races between gpu
-	 * updating its hardware counter while we are retrieving the
-	 * corresponding vblank timestamp.
-	 */
-	do {
-		cur_vblank = dev->driver->get_vblank_counter(dev, crtc);
-		rc = drm_get_last_vbltimestamp(dev, crtc, &t_vblank, 0);
-	} while (cur_vblank != dev->driver->get_vblank_counter(dev, crtc));
-
-	/* Deal with counter wrap */
-	diff = cur_vblank - dev->vblank[crtc].last;
-	if (cur_vblank < dev->vblank[crtc].last) {
-		diff += dev->max_vblank_count;
-
-		DRM_DEBUG("last_vblank[%d]=0x%x, cur_vblank=0x%x => diff=0x%x\n",
-			  crtc, dev->vblank[crtc].last, cur_vblank, diff);
-	}
-
-	DRM_DEBUG("enabling vblank interrupts on crtc %d, missed %d\n",
-		  crtc, diff);
-
-	/* Reinitialize corresponding vblank timestamp if high-precision query
-	 * available. Skip this step if query unsupported or failed. Will
-	 * reinitialize delayed at next vblank interrupt in that case.
-	 */
-	if (rc) {
-		tslot = atomic_read(&dev->vblank[crtc].count) + diff;
-		vblanktimestamp(dev, crtc, tslot) = t_vblank;
-	}
-
-	smp_mb__before_atomic();
-	atomic_add(diff, &dev->vblank[crtc].count);
-	smp_mb__after_atomic();
+	drm_send_vblank_event(crtc->dev, drm_crtc_index(crtc), e);
 }
+EXPORT_SYMBOL(drm_crtc_send_vblank_event);
 
 /**
  * drm_vblank_enable - enable the vblank interrupt on a CRTC
@@ -870,13 +969,14 @@ static void drm_update_vblank_count(struct drm_device *dev, int crtc)
  */
 static int drm_vblank_enable(struct drm_device *dev, int crtc)
 {
+	struct drm_vblank_crtc *vblank = &dev->vblank[crtc];
 	int ret = 0;
 
 	assert_spin_locked(&dev->vbl_lock);
 
 	spin_lock(&dev->vblank_time_lock);
 
-	if (!dev->vblank[crtc].enabled) {
+	if (!vblank->enabled) {
 		/*
 		 * Enable vblank irqs under vblank_time_lock protection.
 		 * All vblank count & timestamp updates are held off
@@ -887,9 +987,9 @@ static int drm_vblank_enable(struct drm_device *dev, int crtc)
 		ret = dev->driver->enable_vblank(dev, crtc);
 		DRM_DEBUG("enabling vblank on crtc %d, ret: %d\n", crtc, ret);
 		if (ret)
-			atomic_dec(&dev->vblank[crtc].refcount);
+			atomic_dec(&vblank->refcount);
 		else {
-			dev->vblank[crtc].enabled = true;
+			vblank->enabled = true;
 			drm_update_vblank_count(dev, crtc);
 		}
 	}
@@ -914,16 +1014,20 @@ static int drm_vblank_enable(struct drm_device *dev, int crtc)
  */
 int drm_vblank_get(struct drm_device *dev, int crtc)
 {
+	struct drm_vblank_crtc *vblank = &dev->vblank[crtc];
 	unsigned long irqflags;
 	int ret = 0;
 
+	if (WARN_ON(crtc >= dev->num_crtcs))
+		return -EINVAL;
+
 	spin_lock_irqsave(&dev->vbl_lock, irqflags);
 	/* Going from 0->1 means we have to enable interrupts again */
-	if (atomic_add_return(1, &dev->vblank[crtc].refcount) == 1) {
+	if (atomic_add_return(1, &vblank->refcount) == 1) {
 		ret = drm_vblank_enable(dev, crtc);
 	} else {
-		if (!dev->vblank[crtc].enabled) {
-			atomic_dec(&dev->vblank[crtc].refcount);
+		if (!vblank->enabled) {
+			atomic_dec(&vblank->refcount);
 			ret = -EINVAL;
 		}
 	}
@@ -963,13 +1067,24 @@ EXPORT_SYMBOL(drm_crtc_vblank_get);
  */
 void drm_vblank_put(struct drm_device *dev, int crtc)
 {
-	BUG_ON(atomic_read(&dev->vblank[crtc].refcount) == 0);
+	struct drm_vblank_crtc *vblank = &dev->vblank[crtc];
+
+	if (WARN_ON(atomic_read(&vblank->refcount) == 0))
+		return;
+
+	if (WARN_ON(crtc >= dev->num_crtcs))
+		return;
 
 	/* Last user schedules interrupt disable */
-	if (atomic_dec_and_test(&dev->vblank[crtc].refcount) &&
-	    (drm_vblank_offdelay > 0))
-		mod_timer(&dev->vblank[crtc].disable_timer,
-			  jiffies + ((drm_vblank_offdelay * HZ)/1000));
+	if (atomic_dec_and_test(&vblank->refcount)) {
+		if (drm_vblank_offdelay == 0)
+			return;
+		else if (dev->vblank_disable_immediate || drm_vblank_offdelay < 0)
+			vblank_disable_fn((unsigned long)vblank);
+		else
+			mod_timer(&vblank->disable_timer,
+				  jiffies + ((drm_vblank_offdelay * HZ)/1000));
+	}
 }
 EXPORT_SYMBOL(drm_vblank_put);
 
@@ -989,6 +1104,50 @@ void drm_crtc_vblank_put(struct drm_crtc *crtc)
 EXPORT_SYMBOL(drm_crtc_vblank_put);
 
 /**
+ * drm_wait_one_vblank - wait for one vblank
+ * @dev: DRM device
+ * @crtc: crtc index
+ *
+ * This waits for one vblank to pass on @crtc, using the irq driver interfaces.
+ * It is a failure to call this when the vblank irq for @crtc is disabled, e.g.
+ * due to lack of driver support or because the crtc is off.
+ */
+void drm_wait_one_vblank(struct drm_device *dev, int crtc)
+{
+	int ret;
+	u32 last;
+
+	ret = drm_vblank_get(dev, crtc);
+	if (WARN(ret, "vblank not available on crtc %i, ret=%i\n", crtc, ret))
+		return;
+
+	last = drm_vblank_count(dev, crtc);
+
+	ret = wait_event_timeout(dev->vblank[crtc].queue,
+				 last != drm_vblank_count(dev, crtc),
+				 msecs_to_jiffies(100));
+
+	WARN(ret == 0, "vblank wait timed out on crtc %i\n", crtc);
+
+	drm_vblank_put(dev, crtc);
+}
+EXPORT_SYMBOL(drm_wait_one_vblank);
+
+/**
+ * drm_crtc_wait_one_vblank - wait for one vblank
+ * @crtc: DRM crtc
+ *
+ * This waits for one vblank to pass on @crtc, using the irq driver interfaces.
+ * It is a failure to call this when the vblank irq for @crtc is disabled, e.g.
+ * due to lack of driver support or because the crtc is off.
+ */
+void drm_crtc_wait_one_vblank(struct drm_crtc *crtc)
+{
+	drm_wait_one_vblank(crtc->dev, drm_crtc_index(crtc));
+}
+EXPORT_SYMBOL(drm_crtc_wait_one_vblank);
+
+/**
  * drm_vblank_off - disable vblank events on a CRTC
  * @dev: DRM device
  * @crtc: CRTC in question
@@ -1004,19 +1163,34 @@ EXPORT_SYMBOL(drm_crtc_vblank_put);
  */
 void drm_vblank_off(struct drm_device *dev, int crtc)
 {
+	struct drm_vblank_crtc *vblank = &dev->vblank[crtc];
 	struct drm_pending_vblank_event *e, *t;
 	struct timeval now;
 	unsigned long irqflags;
 	unsigned int seq;
 
-	spin_lock_irqsave(&dev->vbl_lock, irqflags);
+	if (WARN_ON(crtc >= dev->num_crtcs))
+		return;
+
+	spin_lock_irqsave(&dev->event_lock, irqflags);
+
+	spin_lock(&dev->vbl_lock);
 	vblank_disable_and_save(dev, crtc);
-	wake_up(&dev->vblank[crtc].queue);
+	wake_up(&vblank->queue);
+
+	/*
+	 * Prevent subsequent drm_vblank_get() from re-enabling
+	 * the vblank interrupt by bumping the refcount.
+	 */
+	if (!vblank->inmodeset) {
+		atomic_inc(&vblank->refcount);
+		vblank->inmodeset = 1;
+	}
+	spin_unlock(&dev->vbl_lock);
 
 	/* Send any queued vblank events, lest the natives grow disquiet */
 	seq = drm_vblank_count_and_time(dev, crtc, &now);
 
-	spin_lock(&dev->event_lock);
 	list_for_each_entry_safe(e, t, &dev->vblank_event_list, base.link) {
 		if (e->pipe != crtc)
 			continue;
@@ -1027,9 +1201,7 @@ void drm_vblank_off(struct drm_device *dev, int crtc)
 		drm_vblank_put(dev, e->pipe);
 		send_vblank_event(dev, e, seq, &now);
 	}
-	spin_unlock(&dev->event_lock);
-
-	spin_unlock_irqrestore(&dev->vbl_lock, irqflags);
+	spin_unlock_irqrestore(&dev->event_lock, irqflags);
 }
 EXPORT_SYMBOL(drm_vblank_off);
 
@@ -1059,18 +1231,42 @@ EXPORT_SYMBOL(drm_crtc_vblank_off);
  *
  * This functions restores the vblank interrupt state captured with
  * drm_vblank_off() again. Note that calls to drm_vblank_on() and
- * drm_vblank_off() can be unbalanced and so can also be unconditionaly called
+ * drm_vblank_off() can be unbalanced and so can also be unconditionally called
  * in driver load code to reflect the current hardware state of the crtc.
  *
  * This is the legacy version of drm_crtc_vblank_on().
  */
 void drm_vblank_on(struct drm_device *dev, int crtc)
 {
+	struct drm_vblank_crtc *vblank = &dev->vblank[crtc];
 	unsigned long irqflags;
 
+	if (WARN_ON(crtc >= dev->num_crtcs))
+		return;
+
 	spin_lock_irqsave(&dev->vbl_lock, irqflags);
-	/* re-enable interrupts if there's are users left */
-	if (atomic_read(&dev->vblank[crtc].refcount) != 0)
+	/* Drop our private "prevent drm_vblank_get" refcount */
+	if (vblank->inmodeset) {
+		atomic_dec(&vblank->refcount);
+		vblank->inmodeset = 0;
+	}
+
+	/*
+	 * sample the current counter to avoid random jumps
+	 * when drm_vblank_enable() applies the diff
+	 *
+	 * -1 to make sure user will never see the same
+	 * vblank counter value before and after a modeset
+	 */
+	vblank->last =
+		(dev->driver->get_vblank_counter(dev, crtc) - 1) &
+		dev->max_vblank_count;
+	/*
+	 * re-enable interrupts if there are users left, or the
+	 * user wishes vblank interrupts to be enabled all the time.
+	 */
+	if (atomic_read(&vblank->refcount) != 0 ||
+	    (!dev->vblank_disable_immediate && drm_vblank_offdelay == 0))
 		WARN_ON(drm_vblank_enable(dev, crtc));
 	spin_unlock_irqrestore(&dev->vbl_lock, irqflags);
 }
@@ -1082,7 +1278,7 @@ EXPORT_SYMBOL(drm_vblank_on);
  *
  * This functions restores the vblank interrupt state captured with
  * drm_vblank_off() again. Note that calls to drm_vblank_on() and
- * drm_vblank_off() can be unbalanced and so can also be unconditionaly called
+ * drm_vblank_off() can be unbalanced and so can also be unconditionally called
  * in driver load code to reflect the current hardware state of the crtc.
  *
  * This is the native kms version of drm_vblank_on().
@@ -1118,9 +1314,15 @@ EXPORT_SYMBOL(drm_crtc_vblank_on);
  */
 void drm_vblank_pre_modeset(struct drm_device *dev, int crtc)
 {
+	struct drm_vblank_crtc *vblank = &dev->vblank[crtc];
+
 	/* vblank is not initialized (IRQ not installed ?), or has been freed */
 	if (!dev->num_crtcs)
 		return;
+
+	if (WARN_ON(crtc >= dev->num_crtcs))
+		return;
+
 	/*
 	 * To avoid all the problems that might happen if interrupts
 	 * were enabled/disabled around or between these calls, we just
@@ -1128,10 +1330,10 @@ void drm_vblank_pre_modeset(struct drm_device *dev, int crtc)
 	 * to avoid corrupting the count if multiple, mismatch calls occur),
 	 * so that interrupts remain enabled in the interim.
 	 */
-	if (!dev->vblank[crtc].inmodeset) {
-		dev->vblank[crtc].inmodeset = 0x1;
+	if (!vblank->inmodeset) {
+		vblank->inmodeset = 0x1;
 		if (drm_vblank_get(dev, crtc) == 0)
-			dev->vblank[crtc].inmodeset |= 0x2;
+			vblank->inmodeset |= 0x2;
 	}
 }
 EXPORT_SYMBOL(drm_vblank_pre_modeset);
@@ -1146,21 +1348,22 @@ EXPORT_SYMBOL(drm_vblank_pre_modeset);
  */
 void drm_vblank_post_modeset(struct drm_device *dev, int crtc)
 {
+	struct drm_vblank_crtc *vblank = &dev->vblank[crtc];
 	unsigned long irqflags;
 
 	/* vblank is not initialized (IRQ not installed ?), or has been freed */
 	if (!dev->num_crtcs)
 		return;
 
-	if (dev->vblank[crtc].inmodeset) {
+	if (vblank->inmodeset) {
 		spin_lock_irqsave(&dev->vbl_lock, irqflags);
 		dev->vblank_disable_allowed = true;
 		spin_unlock_irqrestore(&dev->vbl_lock, irqflags);
 
-		if (dev->vblank[crtc].inmodeset & 0x2)
+		if (vblank->inmodeset & 0x2)
 			drm_vblank_put(dev, crtc);
 
-		dev->vblank[crtc].inmodeset = 0;
+		vblank->inmodeset = 0;
 	}
 }
 EXPORT_SYMBOL(drm_vblank_post_modeset);
@@ -1212,6 +1415,7 @@ static int drm_queue_vblank_event(struct drm_device *dev, int pipe,
 				  union drm_wait_vblank *vblwait,
 				  struct drm_file *file_priv)
 {
+	struct drm_vblank_crtc *vblank = &dev->vblank[pipe];
 	struct drm_pending_vblank_event *e;
 	struct timeval now;
 	unsigned long flags;
@@ -1235,6 +1439,18 @@ static int drm_queue_vblank_event(struct drm_device *dev, int pipe,
 
 	spin_lock_irqsave(&dev->event_lock, flags);
 
+	/*
+	 * drm_vblank_off() might have been called after we called
+	 * drm_vblank_get(). drm_vblank_off() holds event_lock
+	 * around the vblank disable, so no need for further locking.
+	 * The reference from drm_vblank_get() protects against
+	 * vblank disable from another source.
+	 */
+	if (!vblank->enabled) {
+		ret = -EINVAL;
+		goto err_unlock;
+	}
+
 	if (file_priv->event_space < sizeof e->event) {
 		ret = -EBUSY;
 		goto err_unlock;
@@ -1295,6 +1511,7 @@ err_put:
 int drm_wait_vblank(struct drm_device *dev, void *data,
 		    struct drm_file *file_priv)
 {
+	struct drm_vblank_crtc *vblank;
 	union drm_wait_vblank *vblwait = data;
 	int ret;
 	unsigned int flags, seq, crtc, high_crtc;
@@ -1324,6 +1541,8 @@ int drm_wait_vblank(struct drm_device *dev, void *data,
 	if (crtc >= dev->num_crtcs)
 		return -EINVAL;
 
+	vblank = &dev->vblank[crtc];
+
 	ret = drm_vblank_get(dev, crtc);
 	if (ret) {
 		DRM_DEBUG("failed to acquire vblank counter, %d\n", ret);
@@ -1356,11 +1575,11 @@ int drm_wait_vblank(struct drm_device *dev, void *data,
 
 	DRM_DEBUG("waiting on vblank count %d, crtc %d\n",
 		  vblwait->request.sequence, crtc);
-	dev->vblank[crtc].last_wait = vblwait->request.sequence;
-	DRM_WAIT_ON(ret, dev->vblank[crtc].queue, 3 * HZ,
+	vblank->last_wait = vblwait->request.sequence;
+	DRM_WAIT_ON(ret, vblank->queue, 3 * HZ,
 		    (((drm_vblank_count(dev, crtc) -
 		       vblwait->request.sequence) <= (1 << 23)) ||
-		     !dev->vblank[crtc].enabled ||
+		     !vblank->enabled ||
 		     !dev->irq_enabled));
 
 	if (ret != -EINTR) {
@@ -1385,12 +1604,11 @@ static void drm_handle_vblank_events(struct drm_device *dev, int crtc)
 {
 	struct drm_pending_vblank_event *e, *t;
 	struct timeval now;
-	unsigned long flags;
 	unsigned int seq;
 
-	seq = drm_vblank_count_and_time(dev, crtc, &now);
+	assert_spin_locked(&dev->event_lock);
 
-	spin_lock_irqsave(&dev->event_lock, flags);
+	seq = drm_vblank_count_and_time(dev, crtc, &now);
 
 	list_for_each_entry_safe(e, t, &dev->vblank_event_list, base.link) {
 		if (e->pipe != crtc)
@@ -1406,8 +1624,6 @@ static void drm_handle_vblank_events(struct drm_device *dev, int crtc)
 		send_vblank_event(dev, e, seq, &now);
 	}
 
-	spin_unlock_irqrestore(&dev->event_lock, flags);
-
 	trace_drm_vblank_event(crtc, seq);
 }
 
@@ -1418,9 +1634,12 @@ static void drm_handle_vblank_events(struct drm_device *dev, int crtc)
  *
  * Drivers should call this routine in their vblank interrupt handlers to
  * update the vblank counter and send any signals that may be pending.
+ *
+ * This is the legacy version of drm_crtc_handle_vblank().
  */
 bool drm_handle_vblank(struct drm_device *dev, int crtc)
 {
+	struct drm_vblank_crtc *vblank = &dev->vblank[crtc];
 	u32 vblcount;
 	s64 diff_ns;
 	struct timeval tvblank;
@@ -1429,15 +1648,21 @@ bool drm_handle_vblank(struct drm_device *dev, int crtc)
 	if (!dev->num_crtcs)
 		return false;
 
+	if (WARN_ON(crtc >= dev->num_crtcs))
+		return false;
+
+	spin_lock_irqsave(&dev->event_lock, irqflags);
+
 	/* Need timestamp lock to prevent concurrent execution with
 	 * vblank enable/disable, as this would cause inconsistent
 	 * or corrupted timestamps and vblank counts.
 	 */
-	spin_lock_irqsave(&dev->vblank_time_lock, irqflags);
+	spin_lock(&dev->vblank_time_lock);
 
 	/* Vblank irq handling disabled. Nothing to do. */
-	if (!dev->vblank[crtc].enabled) {
-		spin_unlock_irqrestore(&dev->vblank_time_lock, irqflags);
+	if (!vblank->enabled) {
+		spin_unlock(&dev->vblank_time_lock);
+		spin_unlock_irqrestore(&dev->event_lock, irqflags);
 		return false;
 	}
 
@@ -1446,7 +1671,7 @@ bool drm_handle_vblank(struct drm_device *dev, int crtc)
 	 */
 
 	/* Get current timestamp and count. */
-	vblcount = atomic_read(&dev->vblank[crtc].count);
+	vblcount = atomic_read(&vblank->count);
 	drm_get_last_vbltimestamp(dev, crtc, &tvblank, DRM_CALLED_FROM_VBLIRQ);
 
 	/* Compute time difference to timestamp of last vblank */
@@ -1470,17 +1695,38 @@ bool drm_handle_vblank(struct drm_device *dev, int crtc)
 		 * the timestamp computed above.
 		 */
 		smp_mb__before_atomic();
-		atomic_inc(&dev->vblank[crtc].count);
+		atomic_inc(&vblank->count);
 		smp_mb__after_atomic();
 	} else {
 		DRM_DEBUG("crtc %d: Redundant vblirq ignored. diff_ns = %d\n",
 			  crtc, (int) diff_ns);
 	}
 
-	wake_up(&dev->vblank[crtc].queue);
+	spin_unlock(&dev->vblank_time_lock);
+
+	wake_up(&vblank->queue);
 	drm_handle_vblank_events(dev, crtc);
 
-	spin_unlock_irqrestore(&dev->vblank_time_lock, irqflags);
+	spin_unlock_irqrestore(&dev->event_lock, irqflags);
+
 	return true;
 }
 EXPORT_SYMBOL(drm_handle_vblank);
+
+/**
+ * drm_crtc_handle_vblank - handle a vblank event
+ * @crtc: where this event occurred
+ *
+ * Drivers should call this routine in their vblank interrupt handlers to
+ * update the vblank counter and send any signals that may be pending.
+ *
+ * This is the native KMS version of drm_handle_vblank().
+ *
+ * Returns:
+ * True if the event was successfully handled, false on failure.
+ */
+bool drm_crtc_handle_vblank(struct drm_crtc *crtc)
+{
+	return drm_handle_vblank(crtc->dev, drm_crtc_index(crtc));
+}
+EXPORT_SYMBOL(drm_crtc_handle_vblank);
diff --git a/drivers/gpu/drm/drm_legacy.h b/drivers/gpu/drm/drm_legacy.h
index d34f20a79b7c..c1dc61473db5 100644
--- a/drivers/gpu/drm/drm_legacy.h
+++ b/drivers/gpu/drm/drm_legacy.h
@@ -23,6 +23,15 @@
  * OTHER DEALINGS IN THE SOFTWARE.
  */
 
+/*
+ * This file contains legacy interfaces that modern drm drivers
+ * should no longer be using. They cannot be removed as legacy
+ * drivers use them, and removing them are API breaks.
+ */
+#include <linux/list.h>
+#include <drm/drm_legacy.h>
+
+struct agp_memory;
 struct drm_device;
 struct drm_file;
 
@@ -48,4 +57,57 @@ int drm_legacy_rmctx(struct drm_device *d, void *v, struct drm_file *f);
 int drm_legacy_setsareactx(struct drm_device *d, void *v, struct drm_file *f);
 int drm_legacy_getsareactx(struct drm_device *d, void *v, struct drm_file *f);
 
+/*
+ * Generic Buffer Management
+ */
+
+#define DRM_MAP_HASH_OFFSET 0x10000000
+
+int drm_legacy_addmap_ioctl(struct drm_device *d, void *v, struct drm_file *f);
+int drm_legacy_rmmap_ioctl(struct drm_device *d, void *v, struct drm_file *f);
+int drm_legacy_addbufs(struct drm_device *d, void *v, struct drm_file *f);
+int drm_legacy_infobufs(struct drm_device *d, void *v, struct drm_file *f);
+int drm_legacy_markbufs(struct drm_device *d, void *v, struct drm_file *f);
+int drm_legacy_freebufs(struct drm_device *d, void *v, struct drm_file *f);
+int drm_legacy_mapbufs(struct drm_device *d, void *v, struct drm_file *f);
+int drm_legacy_dma_ioctl(struct drm_device *d, void *v, struct drm_file *f);
+
+void drm_legacy_vma_flush(struct drm_device *d);
+
+/*
+ * AGP Support
+ */
+
+struct drm_agp_mem {
+	unsigned long handle;
+	struct agp_memory *memory;
+	unsigned long bound;
+	int pages;
+	struct list_head head;
+};
+
+/*
+ * Generic Userspace Locking-API
+ */
+
+int drm_legacy_i_have_hw_lock(struct drm_device *d, struct drm_file *f);
+int drm_legacy_lock(struct drm_device *d, void *v, struct drm_file *f);
+int drm_legacy_unlock(struct drm_device *d, void *v, struct drm_file *f);
+int drm_legacy_lock_free(struct drm_lock_data *lock, unsigned int ctx);
+
+/* DMA support */
+int drm_legacy_dma_setup(struct drm_device *dev);
+void drm_legacy_dma_takedown(struct drm_device *dev);
+void drm_legacy_free_buffer(struct drm_device *dev,
+			    struct drm_buf * buf);
+void drm_legacy_reclaim_buffers(struct drm_device *dev,
+				struct drm_file *filp);
+
+/* Scatter Gather Support */
+void drm_legacy_sg_cleanup(struct drm_device *dev);
+int drm_legacy_sg_alloc(struct drm_device *dev, void *data,
+			struct drm_file *file_priv);
+int drm_legacy_sg_free(struct drm_device *dev, void *data,
+		       struct drm_file *file_priv);
+
 #endif /* __DRM_LEGACY_H__ */
diff --git a/drivers/gpu/drm/drm_lock.c b/drivers/gpu/drm/drm_lock.c
index e26b59e385ff..f861361a635e 100644
--- a/drivers/gpu/drm/drm_lock.c
+++ b/drivers/gpu/drm/drm_lock.c
@@ -36,6 +36,7 @@
 #include <linux/export.h>
 #include <drm/drmP.h>
 #include "drm_legacy.h"
+#include "drm_internal.h"
 
 static int drm_notifier(void *priv);
 
@@ -52,7 +53,8 @@ static int drm_lock_take(struct drm_lock_data *lock_data, unsigned int context);
  *
  * Add the current task to the lock wait queue, and attempt to take to lock.
  */
-int drm_lock(struct drm_device *dev, void *data, struct drm_file *file_priv)
+int drm_legacy_lock(struct drm_device *dev, void *data,
+		    struct drm_file *file_priv)
 {
 	DECLARE_WAITQUEUE(entry, current);
 	struct drm_lock *lock = data;
@@ -120,7 +122,7 @@ int drm_lock(struct drm_device *dev, void *data, struct drm_file *file_priv)
 		sigaddset(&dev->sigmask, SIGTTOU);
 		dev->sigdata.context = lock->context;
 		dev->sigdata.lock = master->lock.hw_lock;
-		block_all_signals(drm_notifier, &dev->sigdata, &dev->sigmask);
+		block_all_signals(drm_notifier, dev, &dev->sigmask);
 	}
 
 	if (dev->driver->dma_quiescent && (lock->flags & _DRM_LOCK_QUIESCENT))
@@ -146,7 +148,7 @@ int drm_lock(struct drm_device *dev, void *data, struct drm_file *file_priv)
  *
  * Transfer and free the lock.
  */
-int drm_unlock(struct drm_device *dev, void *data, struct drm_file *file_priv)
+int drm_legacy_unlock(struct drm_device *dev, void *data, struct drm_file *file_priv)
 {
 	struct drm_lock *lock = data;
 	struct drm_master *master = file_priv->master;
@@ -157,7 +159,7 @@ int drm_unlock(struct drm_device *dev, void *data, struct drm_file *file_priv)
 		return -EINVAL;
 	}
 
-	if (drm_lock_free(&master->lock, lock->context)) {
+	if (drm_legacy_lock_free(&master->lock, lock->context)) {
 		/* FIXME: Should really bail out here. */
 	}
 
@@ -250,7 +252,7 @@ static int drm_lock_transfer(struct drm_lock_data *lock_data,
  * Marks the lock as not held, via the \p cmpxchg instruction. Wakes any task
  * waiting on the lock queue.
  */
-int drm_lock_free(struct drm_lock_data *lock_data, unsigned int context)
+int drm_legacy_lock_free(struct drm_lock_data *lock_data, unsigned int context)
 {
 	unsigned int old, new, prev;
 	volatile unsigned int *lock = &lock_data->hw_lock->lock;
@@ -286,26 +288,27 @@ int drm_lock_free(struct drm_lock_data *lock_data, unsigned int context)
  * If the lock is not held, then let the signal proceed as usual.  If the lock
  * is held, then set the contended flag and keep the signal blocked.
  *
- * \param priv pointer to a drm_sigdata structure.
+ * \param priv pointer to a drm_device structure.
  * \return one if the signal should be delivered normally, or zero if the
  * signal should be blocked.
  */
 static int drm_notifier(void *priv)
 {
-	struct drm_sigdata *s = (struct drm_sigdata *) priv;
+	struct drm_device *dev = priv;
+	struct drm_hw_lock *lock = dev->sigdata.lock;
 	unsigned int old, new, prev;
 
 	/* Allow signal delivery if lock isn't held */
-	if (!s->lock || !_DRM_LOCK_IS_HELD(s->lock->lock)
-	    || _DRM_LOCKING_CONTEXT(s->lock->lock) != s->context)
+	if (!lock || !_DRM_LOCK_IS_HELD(lock->lock)
+	    || _DRM_LOCKING_CONTEXT(lock->lock) != dev->sigdata.context)
 		return 1;
 
 	/* Otherwise, set flag to force call to
 	   drmUnlock */
 	do {
-		old = s->lock->lock;
+		old = lock->lock;
 		new = old | _DRM_LOCK_CONT;
-		prev = cmpxchg(&s->lock->lock, old, new);
+		prev = cmpxchg(&lock->lock, old, new);
 	} while (prev != old);
 	return 0;
 }
@@ -323,7 +326,7 @@ static int drm_notifier(void *priv)
  * having to worry about starvation.
  */
 
-void drm_idlelock_take(struct drm_lock_data *lock_data)
+void drm_legacy_idlelock_take(struct drm_lock_data *lock_data)
 {
 	int ret;
 
@@ -340,9 +343,9 @@ void drm_idlelock_take(struct drm_lock_data *lock_data)
 	}
 	spin_unlock_bh(&lock_data->spinlock);
 }
-EXPORT_SYMBOL(drm_idlelock_take);
+EXPORT_SYMBOL(drm_legacy_idlelock_take);
 
-void drm_idlelock_release(struct drm_lock_data *lock_data)
+void drm_legacy_idlelock_release(struct drm_lock_data *lock_data)
 {
 	unsigned int old, prev;
 	volatile unsigned int *lock = &lock_data->hw_lock->lock;
@@ -360,9 +363,10 @@ void drm_idlelock_release(struct drm_lock_data *lock_data)
 	}
 	spin_unlock_bh(&lock_data->spinlock);
 }
-EXPORT_SYMBOL(drm_idlelock_release);
+EXPORT_SYMBOL(drm_legacy_idlelock_release);
 
-int drm_i_have_hw_lock(struct drm_device *dev, struct drm_file *file_priv)
+int drm_legacy_i_have_hw_lock(struct drm_device *dev,
+			      struct drm_file *file_priv)
 {
 	struct drm_master *master = file_priv->master;
 	return (file_priv->lock_count && master->lock.hw_lock &&
diff --git a/drivers/gpu/drm/drm_memory.c b/drivers/gpu/drm/drm_memory.c
index 00c67c0f2381..a521ef6ff807 100644
--- a/drivers/gpu/drm/drm_memory.c
+++ b/drivers/gpu/drm/drm_memory.c
@@ -36,8 +36,20 @@
 #include <linux/highmem.h>
 #include <linux/export.h>
 #include <drm/drmP.h>
+#include "drm_legacy.h"
 
 #if __OS_HAS_AGP
+
+#ifdef HAVE_PAGE_AGP
+# include <asm/agp.h>
+#else
+# ifdef __powerpc__
+#  define PAGE_AGP	__pgprot(_PAGE_KERNEL | _PAGE_NO_CACHE)
+# else
+#  define PAGE_AGP	PAGE_KERNEL
+# endif
+#endif
+
 static void *agp_remap(unsigned long offset, unsigned long size,
 		       struct drm_device * dev)
 {
@@ -108,25 +120,25 @@ static inline void *agp_remap(unsigned long offset, unsigned long size,
 
 #endif				/* agp */
 
-void drm_core_ioremap(struct drm_local_map *map, struct drm_device *dev)
+void drm_legacy_ioremap(struct drm_local_map *map, struct drm_device *dev)
 {
 	if (dev->agp && dev->agp->cant_use_aperture && map->type == _DRM_AGP)
 		map->handle = agp_remap(map->offset, map->size, dev);
 	else
 		map->handle = ioremap(map->offset, map->size);
 }
-EXPORT_SYMBOL(drm_core_ioremap);
+EXPORT_SYMBOL(drm_legacy_ioremap);
 
-void drm_core_ioremap_wc(struct drm_local_map *map, struct drm_device *dev)
+void drm_legacy_ioremap_wc(struct drm_local_map *map, struct drm_device *dev)
 {
 	if (dev->agp && dev->agp->cant_use_aperture && map->type == _DRM_AGP)
 		map->handle = agp_remap(map->offset, map->size, dev);
 	else
 		map->handle = ioremap_wc(map->offset, map->size);
 }
-EXPORT_SYMBOL(drm_core_ioremap_wc);
+EXPORT_SYMBOL(drm_legacy_ioremap_wc);
 
-void drm_core_ioremapfree(struct drm_local_map *map, struct drm_device *dev)
+void drm_legacy_ioremapfree(struct drm_local_map *map, struct drm_device *dev)
 {
 	if (!map->handle || !map->size)
 		return;
@@ -136,4 +148,4 @@ void drm_core_ioremapfree(struct drm_local_map *map, struct drm_device *dev)
 	else
 		iounmap(map->handle);
 }
-EXPORT_SYMBOL(drm_core_ioremapfree);
+EXPORT_SYMBOL(drm_legacy_ioremapfree);
diff --git a/drivers/gpu/drm/drm_mipi_dsi.c b/drivers/gpu/drm/drm_mipi_dsi.c
index 6aa6a9e95570..c0644bb865f2 100644
--- a/drivers/gpu/drm/drm_mipi_dsi.c
+++ b/drivers/gpu/drm/drm_mipi_dsi.c
@@ -35,6 +35,16 @@
 
 #include <video/mipi_display.h>
 
+/**
+ * DOC: dsi helpers
+ *
+ * These functions contain some common logic and helpers to deal with MIPI DSI
+ * peripherals.
+ *
+ * Helpers are provided for a number of standard MIPI DSI command as well as a
+ * subset of the MIPI DCS command set.
+ */
+
 static int mipi_dsi_device_match(struct device *dev, struct device_driver *drv)
 {
 	return of_driver_match_device(dev, drv);
@@ -57,6 +67,29 @@ static struct bus_type mipi_dsi_bus_type = {
 	.pm = &mipi_dsi_device_pm_ops,
 };
 
+static int of_device_match(struct device *dev, void *data)
+{
+	return dev->of_node == data;
+}
+
+/**
+ * of_find_mipi_dsi_device_by_node() - find the MIPI DSI device matching a
+ *    device tree node
+ * @np: device tree node
+ *
+ * Return: A pointer to the MIPI DSI device corresponding to @np or NULL if no
+ *    such device exists (or has not been registered yet).
+ */
+struct mipi_dsi_device *of_find_mipi_dsi_device_by_node(struct device_node *np)
+{
+	struct device *dev;
+
+	dev = bus_find_device(&mipi_dsi_bus_type, NULL, np, of_device_match);
+
+	return dev ? to_mipi_dsi_device(dev) : NULL;
+}
+EXPORT_SYMBOL(of_find_mipi_dsi_device_by_node);
+
 static void mipi_dsi_dev_release(struct device *dev)
 {
 	struct mipi_dsi_device *dsi = to_mipi_dsi_device(dev);
@@ -198,56 +231,351 @@ int mipi_dsi_detach(struct mipi_dsi_device *dsi)
 }
 EXPORT_SYMBOL(mipi_dsi_detach);
 
+static ssize_t mipi_dsi_device_transfer(struct mipi_dsi_device *dsi,
+					struct mipi_dsi_msg *msg)
+{
+	const struct mipi_dsi_host_ops *ops = dsi->host->ops;
+
+	if (!ops || !ops->transfer)
+		return -ENOSYS;
+
+	if (dsi->mode_flags & MIPI_DSI_MODE_LPM)
+		msg->flags |= MIPI_DSI_MSG_USE_LPM;
+
+	return ops->transfer(dsi->host, msg);
+}
+
 /**
- * mipi_dsi_dcs_write - send DCS write command
- * @dsi: DSI device
- * @data: pointer to the command followed by parameters
- * @len: length of @data
+ * mipi_dsi_packet_format_is_short - check if a packet is of the short format
+ * @type: MIPI DSI data type of the packet
+ *
+ * Return: true if the packet for the given data type is a short packet, false
+ * otherwise.
  */
-ssize_t mipi_dsi_dcs_write(struct mipi_dsi_device *dsi, const void *data,
-			    size_t len)
+bool mipi_dsi_packet_format_is_short(u8 type)
+{
+	switch (type) {
+	case MIPI_DSI_V_SYNC_START:
+	case MIPI_DSI_V_SYNC_END:
+	case MIPI_DSI_H_SYNC_START:
+	case MIPI_DSI_H_SYNC_END:
+	case MIPI_DSI_END_OF_TRANSMISSION:
+	case MIPI_DSI_COLOR_MODE_OFF:
+	case MIPI_DSI_COLOR_MODE_ON:
+	case MIPI_DSI_SHUTDOWN_PERIPHERAL:
+	case MIPI_DSI_TURN_ON_PERIPHERAL:
+	case MIPI_DSI_GENERIC_SHORT_WRITE_0_PARAM:
+	case MIPI_DSI_GENERIC_SHORT_WRITE_1_PARAM:
+	case MIPI_DSI_GENERIC_SHORT_WRITE_2_PARAM:
+	case MIPI_DSI_GENERIC_READ_REQUEST_0_PARAM:
+	case MIPI_DSI_GENERIC_READ_REQUEST_1_PARAM:
+	case MIPI_DSI_GENERIC_READ_REQUEST_2_PARAM:
+	case MIPI_DSI_DCS_SHORT_WRITE:
+	case MIPI_DSI_DCS_SHORT_WRITE_PARAM:
+	case MIPI_DSI_DCS_READ:
+	case MIPI_DSI_SET_MAXIMUM_RETURN_PACKET_SIZE:
+		return true;
+	}
+
+	return false;
+}
+EXPORT_SYMBOL(mipi_dsi_packet_format_is_short);
+
+/**
+ * mipi_dsi_packet_format_is_long - check if a packet is of the long format
+ * @type: MIPI DSI data type of the packet
+ *
+ * Return: true if the packet for the given data type is a long packet, false
+ * otherwise.
+ */
+bool mipi_dsi_packet_format_is_long(u8 type)
+{
+	switch (type) {
+	case MIPI_DSI_NULL_PACKET:
+	case MIPI_DSI_BLANKING_PACKET:
+	case MIPI_DSI_GENERIC_LONG_WRITE:
+	case MIPI_DSI_DCS_LONG_WRITE:
+	case MIPI_DSI_LOOSELY_PACKED_PIXEL_STREAM_YCBCR20:
+	case MIPI_DSI_PACKED_PIXEL_STREAM_YCBCR24:
+	case MIPI_DSI_PACKED_PIXEL_STREAM_YCBCR16:
+	case MIPI_DSI_PACKED_PIXEL_STREAM_30:
+	case MIPI_DSI_PACKED_PIXEL_STREAM_36:
+	case MIPI_DSI_PACKED_PIXEL_STREAM_YCBCR12:
+	case MIPI_DSI_PACKED_PIXEL_STREAM_16:
+	case MIPI_DSI_PACKED_PIXEL_STREAM_18:
+	case MIPI_DSI_PIXEL_STREAM_3BYTE_18:
+	case MIPI_DSI_PACKED_PIXEL_STREAM_24:
+		return true;
+	}
+
+	return false;
+}
+EXPORT_SYMBOL(mipi_dsi_packet_format_is_long);
+
+/**
+ * mipi_dsi_create_packet - create a packet from a message according to the
+ *     DSI protocol
+ * @packet: pointer to a DSI packet structure
+ * @msg: message to translate into a packet
+ *
+ * Return: 0 on success or a negative error code on failure.
+ */
+int mipi_dsi_create_packet(struct mipi_dsi_packet *packet,
+			   const struct mipi_dsi_msg *msg)
+{
+	const u8 *tx = msg->tx_buf;
+
+	if (!packet || !msg)
+		return -EINVAL;
+
+	/* do some minimum sanity checking */
+	if (!mipi_dsi_packet_format_is_short(msg->type) &&
+	    !mipi_dsi_packet_format_is_long(msg->type))
+		return -EINVAL;
+
+	if (msg->channel > 3)
+		return -EINVAL;
+
+	memset(packet, 0, sizeof(*packet));
+	packet->header[0] = ((msg->channel & 0x3) << 6) | (msg->type & 0x3f);
+
+	/* TODO: compute ECC if hardware support is not available */
+
+	/*
+	 * Long write packets contain the word count in header bytes 1 and 2.
+	 * The payload follows the header and is word count bytes long.
+	 *
+	 * Short write packets encode up to two parameters in header bytes 1
+	 * and 2.
+	 */
+	if (mipi_dsi_packet_format_is_long(msg->type)) {
+		packet->header[1] = (msg->tx_len >> 0) & 0xff;
+		packet->header[2] = (msg->tx_len >> 8) & 0xff;
+
+		packet->payload_length = msg->tx_len;
+		packet->payload = tx;
+	} else {
+		packet->header[1] = (msg->tx_len > 0) ? tx[0] : 0;
+		packet->header[2] = (msg->tx_len > 1) ? tx[1] : 0;
+	}
+
+	packet->size = sizeof(packet->header) + packet->payload_length;
+
+	return 0;
+}
+EXPORT_SYMBOL(mipi_dsi_create_packet);
+
+/*
+ * mipi_dsi_set_maximum_return_packet_size() - specify the maximum size of the
+ *    the payload in a long packet transmitted from the peripheral back to the
+ *    host processor
+ * @dsi: DSI peripheral device
+ * @value: the maximum size of the payload
+ *
+ * Return: 0 on success or a negative error code on failure.
+ */
+int mipi_dsi_set_maximum_return_packet_size(struct mipi_dsi_device *dsi,
+					    u16 value)
+{
+	u8 tx[2] = { value & 0xff, value >> 8 };
+	struct mipi_dsi_msg msg = {
+		.channel = dsi->channel,
+		.type = MIPI_DSI_SET_MAXIMUM_RETURN_PACKET_SIZE,
+		.tx_len = sizeof(tx),
+		.tx_buf = tx,
+	};
+
+	return mipi_dsi_device_transfer(dsi, &msg);
+}
+EXPORT_SYMBOL(mipi_dsi_set_maximum_return_packet_size);
+
+/**
+ * mipi_dsi_generic_write() - transmit data using a generic write packet
+ * @dsi: DSI peripheral device
+ * @payload: buffer containing the payload
+ * @size: size of payload buffer
+ *
+ * This function will automatically choose the right data type depending on
+ * the payload length.
+ *
+ * Return: The number of bytes transmitted on success or a negative error code
+ * on failure.
+ */
+ssize_t mipi_dsi_generic_write(struct mipi_dsi_device *dsi, const void *payload,
+			       size_t size)
+{
+	struct mipi_dsi_msg msg = {
+		.channel = dsi->channel,
+		.tx_buf = payload,
+		.tx_len = size
+	};
+
+	switch (size) {
+	case 0:
+		msg.type = MIPI_DSI_GENERIC_SHORT_WRITE_0_PARAM;
+		break;
+
+	case 1:
+		msg.type = MIPI_DSI_GENERIC_SHORT_WRITE_1_PARAM;
+		break;
+
+	case 2:
+		msg.type = MIPI_DSI_GENERIC_SHORT_WRITE_2_PARAM;
+		break;
+
+	default:
+		msg.type = MIPI_DSI_GENERIC_LONG_WRITE;
+		break;
+	}
+
+	return mipi_dsi_device_transfer(dsi, &msg);
+}
+EXPORT_SYMBOL(mipi_dsi_generic_write);
+
+/**
+ * mipi_dsi_generic_read() - receive data using a generic read packet
+ * @dsi: DSI peripheral device
+ * @params: buffer containing the request parameters
+ * @num_params: number of request parameters
+ * @data: buffer in which to return the received data
+ * @size: size of receive buffer
+ *
+ * This function will automatically choose the right data type depending on
+ * the number of parameters passed in.
+ *
+ * Return: The number of bytes successfully read or a negative error code on
+ * failure.
+ */
+ssize_t mipi_dsi_generic_read(struct mipi_dsi_device *dsi, const void *params,
+			      size_t num_params, void *data, size_t size)
+{
+	struct mipi_dsi_msg msg = {
+		.channel = dsi->channel,
+		.tx_len = num_params,
+		.tx_buf = params,
+		.rx_len = size,
+		.rx_buf = data
+	};
+
+	switch (num_params) {
+	case 0:
+		msg.type = MIPI_DSI_GENERIC_READ_REQUEST_0_PARAM;
+		break;
+
+	case 1:
+		msg.type = MIPI_DSI_GENERIC_READ_REQUEST_1_PARAM;
+		break;
+
+	case 2:
+		msg.type = MIPI_DSI_GENERIC_READ_REQUEST_2_PARAM;
+		break;
+
+	default:
+		return -EINVAL;
+	}
+
+	return mipi_dsi_device_transfer(dsi, &msg);
+}
+EXPORT_SYMBOL(mipi_dsi_generic_read);
+
+/**
+ * mipi_dsi_dcs_write_buffer() - transmit a DCS command with payload
+ * @dsi: DSI peripheral device
+ * @data: buffer containing data to be transmitted
+ * @len: size of transmission buffer
+ *
+ * This function will automatically choose the right data type depending on
+ * the command payload length.
+ *
+ * Return: The number of bytes successfully transmitted or a negative error
+ * code on failure.
+ */
+ssize_t mipi_dsi_dcs_write_buffer(struct mipi_dsi_device *dsi,
+				  const void *data, size_t len)
 {
-	const struct mipi_dsi_host_ops *ops = dsi->host->ops;
 	struct mipi_dsi_msg msg = {
 		.channel = dsi->channel,
 		.tx_buf = data,
 		.tx_len = len
 	};
 
-	if (!ops || !ops->transfer)
-		return -ENOSYS;
-
 	switch (len) {
 	case 0:
 		return -EINVAL;
+
 	case 1:
 		msg.type = MIPI_DSI_DCS_SHORT_WRITE;
 		break;
+
 	case 2:
 		msg.type = MIPI_DSI_DCS_SHORT_WRITE_PARAM;
 		break;
+
 	default:
 		msg.type = MIPI_DSI_DCS_LONG_WRITE;
 		break;
 	}
 
-	return ops->transfer(dsi->host, &msg);
+	return mipi_dsi_device_transfer(dsi, &msg);
+}
+EXPORT_SYMBOL(mipi_dsi_dcs_write_buffer);
+
+/**
+ * mipi_dsi_dcs_write() - send DCS write command
+ * @dsi: DSI peripheral device
+ * @cmd: DCS command
+ * @data: buffer containing the command payload
+ * @len: command payload length
+ *
+ * This function will automatically choose the right data type depending on
+ * the command payload length.
+ *
+ * Return: The number of bytes successfully transmitted or a negative error
+ * code on failure.
+ */
+ssize_t mipi_dsi_dcs_write(struct mipi_dsi_device *dsi, u8 cmd,
+			   const void *data, size_t len)
+{
+	ssize_t err;
+	size_t size;
+	u8 *tx;
+
+	if (len > 0) {
+		size = 1 + len;
+
+		tx = kmalloc(size, GFP_KERNEL);
+		if (!tx)
+			return -ENOMEM;
+
+		/* concatenate the DCS command byte and the payload */
+		tx[0] = cmd;
+		memcpy(&tx[1], data, len);
+	} else {
+		tx = &cmd;
+		size = 1;
+	}
+
+	err = mipi_dsi_dcs_write_buffer(dsi, tx, size);
+
+	if (len > 0)
+		kfree(tx);
+
+	return err;
 }
 EXPORT_SYMBOL(mipi_dsi_dcs_write);
 
 /**
- * mipi_dsi_dcs_read - send DCS read request command
- * @dsi: DSI device
- * @cmd: DCS read command
- * @data: pointer to read buffer
- * @len: length of @data
+ * mipi_dsi_dcs_read() - send DCS read request command
+ * @dsi: DSI peripheral device
+ * @cmd: DCS command
+ * @data: buffer in which to receive data
+ * @len: size of receive buffer
  *
- * Function returns number of read bytes or error code.
+ * Return: The number of bytes read or a negative error code on failure.
  */
 ssize_t mipi_dsi_dcs_read(struct mipi_dsi_device *dsi, u8 cmd, void *data,
 			  size_t len)
 {
-	const struct mipi_dsi_host_ops *ops = dsi->host->ops;
 	struct mipi_dsi_msg msg = {
 		.channel = dsi->channel,
 		.type = MIPI_DSI_DCS_READ,
@@ -257,13 +585,283 @@ ssize_t mipi_dsi_dcs_read(struct mipi_dsi_device *dsi, u8 cmd, void *data,
 		.rx_len = len
 	};
 
-	if (!ops || !ops->transfer)
-		return -ENOSYS;
-
-	return ops->transfer(dsi->host, &msg);
+	return mipi_dsi_device_transfer(dsi, &msg);
 }
 EXPORT_SYMBOL(mipi_dsi_dcs_read);
 
+/**
+ * mipi_dsi_dcs_nop() - send DCS nop packet
+ * @dsi: DSI peripheral device
+ *
+ * Return: 0 on success or a negative error code on failure.
+ */
+int mipi_dsi_dcs_nop(struct mipi_dsi_device *dsi)
+{
+	ssize_t err;
+
+	err = mipi_dsi_dcs_write(dsi, MIPI_DCS_NOP, NULL, 0);
+	if (err < 0)
+		return err;
+
+	return 0;
+}
+EXPORT_SYMBOL(mipi_dsi_dcs_nop);
+
+/**
+ * mipi_dsi_dcs_soft_reset() - perform a software reset of the display module
+ * @dsi: DSI peripheral device
+ *
+ * Return: 0 on success or a negative error code on failure.
+ */
+int mipi_dsi_dcs_soft_reset(struct mipi_dsi_device *dsi)
+{
+	ssize_t err;
+
+	err = mipi_dsi_dcs_write(dsi, MIPI_DCS_SOFT_RESET, NULL, 0);
+	if (err < 0)
+		return err;
+
+	return 0;
+}
+EXPORT_SYMBOL(mipi_dsi_dcs_soft_reset);
+
+/**
+ * mipi_dsi_dcs_get_power_mode() - query the display module's current power
+ *    mode
+ * @dsi: DSI peripheral device
+ * @mode: return location for the current power mode
+ *
+ * Return: 0 on success or a negative error code on failure.
+ */
+int mipi_dsi_dcs_get_power_mode(struct mipi_dsi_device *dsi, u8 *mode)
+{
+	ssize_t err;
+
+	err = mipi_dsi_dcs_read(dsi, MIPI_DCS_GET_POWER_MODE, mode,
+				sizeof(*mode));
+	if (err <= 0) {
+		if (err == 0)
+			err = -ENODATA;
+
+		return err;
+	}
+
+	return 0;
+}
+EXPORT_SYMBOL(mipi_dsi_dcs_get_power_mode);
+
+/**
+ * mipi_dsi_dcs_get_pixel_format() - gets the pixel format for the RGB image
+ *    data used by the interface
+ * @dsi: DSI peripheral device
+ * @format: return location for the pixel format
+ *
+ * Return: 0 on success or a negative error code on failure.
+ */
+int mipi_dsi_dcs_get_pixel_format(struct mipi_dsi_device *dsi, u8 *format)
+{
+	ssize_t err;
+
+	err = mipi_dsi_dcs_read(dsi, MIPI_DCS_GET_PIXEL_FORMAT, format,
+				sizeof(*format));
+	if (err <= 0) {
+		if (err == 0)
+			err = -ENODATA;
+
+		return err;
+	}
+
+	return 0;
+}
+EXPORT_SYMBOL(mipi_dsi_dcs_get_pixel_format);
+
+/**
+ * mipi_dsi_dcs_enter_sleep_mode() - disable all unnecessary blocks inside the
+ *    display module except interface communication
+ * @dsi: DSI peripheral device
+ *
+ * Return: 0 on success or a negative error code on failure.
+ */
+int mipi_dsi_dcs_enter_sleep_mode(struct mipi_dsi_device *dsi)
+{
+	ssize_t err;
+
+	err = mipi_dsi_dcs_write(dsi, MIPI_DCS_ENTER_SLEEP_MODE, NULL, 0);
+	if (err < 0)
+		return err;
+
+	return 0;
+}
+EXPORT_SYMBOL(mipi_dsi_dcs_enter_sleep_mode);
+
+/**
+ * mipi_dsi_dcs_exit_sleep_mode() - enable all blocks inside the display
+ *    module
+ * @dsi: DSI peripheral device
+ *
+ * Return: 0 on success or a negative error code on failure.
+ */
+int mipi_dsi_dcs_exit_sleep_mode(struct mipi_dsi_device *dsi)
+{
+	ssize_t err;
+
+	err = mipi_dsi_dcs_write(dsi, MIPI_DCS_EXIT_SLEEP_MODE, NULL, 0);
+	if (err < 0)
+		return err;
+
+	return 0;
+}
+EXPORT_SYMBOL(mipi_dsi_dcs_exit_sleep_mode);
+
+/**
+ * mipi_dsi_dcs_set_display_off() - stop displaying the image data on the
+ *    display device
+ * @dsi: DSI peripheral device
+ *
+ * Return: 0 on success or a negative error code on failure.
+ */
+int mipi_dsi_dcs_set_display_off(struct mipi_dsi_device *dsi)
+{
+	ssize_t err;
+
+	err = mipi_dsi_dcs_write(dsi, MIPI_DCS_SET_DISPLAY_OFF, NULL, 0);
+	if (err < 0)
+		return err;
+
+	return 0;
+}
+EXPORT_SYMBOL(mipi_dsi_dcs_set_display_off);
+
+/**
+ * mipi_dsi_dcs_set_display_on() - start displaying the image data on the
+ *    display device
+ * @dsi: DSI peripheral device
+ *
+ * Return: 0 on success or a negative error code on failure
+ */
+int mipi_dsi_dcs_set_display_on(struct mipi_dsi_device *dsi)
+{
+	ssize_t err;
+
+	err = mipi_dsi_dcs_write(dsi, MIPI_DCS_SET_DISPLAY_ON, NULL, 0);
+	if (err < 0)
+		return err;
+
+	return 0;
+}
+EXPORT_SYMBOL(mipi_dsi_dcs_set_display_on);
+
+/**
+ * mipi_dsi_dcs_set_column_address() - define the column extent of the frame
+ *    memory accessed by the host processor
+ * @dsi: DSI peripheral device
+ * @start: first column of frame memory
+ * @end: last column of frame memory
+ *
+ * Return: 0 on success or a negative error code on failure.
+ */
+int mipi_dsi_dcs_set_column_address(struct mipi_dsi_device *dsi, u16 start,
+				    u16 end)
+{
+	u8 payload[4] = { start >> 8, start & 0xff, end >> 8, end & 0xff };
+	ssize_t err;
+
+	err = mipi_dsi_dcs_write(dsi, MIPI_DCS_SET_COLUMN_ADDRESS, payload,
+				 sizeof(payload));
+	if (err < 0)
+		return err;
+
+	return 0;
+}
+EXPORT_SYMBOL(mipi_dsi_dcs_set_column_address);
+
+/**
+ * mipi_dsi_dcs_set_page_address() - define the page extent of the frame
+ *    memory accessed by the host processor
+ * @dsi: DSI peripheral device
+ * @start: first page of frame memory
+ * @end: last page of frame memory
+ *
+ * Return: 0 on success or a negative error code on failure.
+ */
+int mipi_dsi_dcs_set_page_address(struct mipi_dsi_device *dsi, u16 start,
+				  u16 end)
+{
+	u8 payload[4] = { start >> 8, start & 0xff, end >> 8, end & 0xff };
+	ssize_t err;
+
+	err = mipi_dsi_dcs_write(dsi, MIPI_DCS_SET_PAGE_ADDRESS, payload,
+				 sizeof(payload));
+	if (err < 0)
+		return err;
+
+	return 0;
+}
+EXPORT_SYMBOL(mipi_dsi_dcs_set_page_address);
+
+/**
+ * mipi_dsi_dcs_set_tear_off() - turn off the display module's Tearing Effect
+ *    output signal on the TE signal line
+ * @dsi: DSI peripheral device
+ *
+ * Return: 0 on success or a negative error code on failure
+ */
+int mipi_dsi_dcs_set_tear_off(struct mipi_dsi_device *dsi)
+{
+	ssize_t err;
+
+	err = mipi_dsi_dcs_write(dsi, MIPI_DCS_SET_TEAR_OFF, NULL, 0);
+	if (err < 0)
+		return err;
+
+	return 0;
+}
+EXPORT_SYMBOL(mipi_dsi_dcs_set_tear_off);
+
+/**
+ * mipi_dsi_dcs_set_tear_on() - turn on the display module's Tearing Effect
+ *    output signal on the TE signal line.
+ * @dsi: DSI peripheral device
+ * @mode: the Tearing Effect Output Line mode
+ *
+ * Return: 0 on success or a negative error code on failure
+ */
+int mipi_dsi_dcs_set_tear_on(struct mipi_dsi_device *dsi,
+			     enum mipi_dsi_dcs_tear_mode mode)
+{
+	u8 value = mode;
+	ssize_t err;
+
+	err = mipi_dsi_dcs_write(dsi, MIPI_DCS_SET_TEAR_ON, &value,
+				 sizeof(value));
+	if (err < 0)
+		return err;
+
+	return 0;
+}
+EXPORT_SYMBOL(mipi_dsi_dcs_set_tear_on);
+
+/**
+ * mipi_dsi_dcs_set_pixel_format() - sets the pixel format for the RGB image
+ *    data used by the interface
+ * @dsi: DSI peripheral device
+ * @format: pixel format
+ *
+ * Return: 0 on success or a negative error code on failure.
+ */
+int mipi_dsi_dcs_set_pixel_format(struct mipi_dsi_device *dsi, u8 format)
+{
+	ssize_t err;
+
+	err = mipi_dsi_dcs_write(dsi, MIPI_DCS_SET_PIXEL_FORMAT, &format,
+				 sizeof(format));
+	if (err < 0)
+		return err;
+
+	return 0;
+}
+EXPORT_SYMBOL(mipi_dsi_dcs_set_pixel_format);
+
 static int mipi_dsi_drv_probe(struct device *dev)
 {
 	struct mipi_dsi_driver *drv = to_mipi_dsi_driver(dev->driver);
@@ -289,12 +887,18 @@ static void mipi_dsi_drv_shutdown(struct device *dev)
 }
 
 /**
- * mipi_dsi_driver_register - register a driver for DSI devices
+ * mipi_dsi_driver_register_full() - register a driver for DSI devices
  * @drv: DSI driver structure
+ * @owner: owner module
+ *
+ * Return: 0 on success or a negative error code on failure.
  */
-int mipi_dsi_driver_register(struct mipi_dsi_driver *drv)
+int mipi_dsi_driver_register_full(struct mipi_dsi_driver *drv,
+				  struct module *owner)
 {
 	drv->driver.bus = &mipi_dsi_bus_type;
+	drv->driver.owner = owner;
+
 	if (drv->probe)
 		drv->driver.probe = mipi_dsi_drv_probe;
 	if (drv->remove)
@@ -304,11 +908,13 @@ int mipi_dsi_driver_register(struct mipi_dsi_driver *drv)
 
 	return driver_register(&drv->driver);
 }
-EXPORT_SYMBOL(mipi_dsi_driver_register);
+EXPORT_SYMBOL(mipi_dsi_driver_register_full);
 
 /**
- * mipi_dsi_driver_unregister - unregister a driver for DSI devices
+ * mipi_dsi_driver_unregister() - unregister a driver for DSI devices
  * @drv: DSI driver structure
+ *
+ * Return: 0 on success or a negative error code on failure.
  */
 void mipi_dsi_driver_unregister(struct mipi_dsi_driver *drv)
 {
diff --git a/drivers/gpu/drm/drm_modes.c b/drivers/gpu/drm/drm_modes.c
index bedf1894e17e..6d8b941c8200 100644
--- a/drivers/gpu/drm/drm_modes.c
+++ b/drivers/gpu/drm/drm_modes.c
@@ -914,7 +914,7 @@ EXPORT_SYMBOL(drm_mode_equal_no_clocks_no_stereo);
  *
  * This function is a helper which can be used to validate modes against size
  * limitations of the DRM device/connector. If a mode is too big its status
- * memeber is updated with the appropriate validation failure code. The list
+ * member is updated with the appropriate validation failure code. The list
  * itself is not changed.
  */
 void drm_mode_validate_size(struct drm_device *dev,
@@ -1259,6 +1259,7 @@ drm_mode_create_from_cmdline_mode(struct drm_device *dev,
 	if (!mode)
 		return NULL;
 
+	mode->type |= DRM_MODE_TYPE_USERDEF;
 	drm_mode_set_crtcinfo(mode, CRTC_INTERLACE_HALVE_V);
 	return mode;
 }
diff --git a/drivers/gpu/drm/drm_modeset_lock.c b/drivers/gpu/drm/drm_modeset_lock.c
index 0dc57d5ecd10..51cc47d827d8 100644
--- a/drivers/gpu/drm/drm_modeset_lock.c
+++ b/drivers/gpu/drm/drm_modeset_lock.c
@@ -35,7 +35,7 @@
  * of extra utility/tracking out of our acquire-ctx.  This is provided
  * by drm_modeset_lock / drm_modeset_acquire_ctx.
  *
- * For basic principles of ww_mutex, see: Documentation/ww-mutex-design.txt
+ * For basic principles of ww_mutex, see: Documentation/locking/ww-mutex-design.txt
  *
  * The basic usage pattern is to:
  *
@@ -57,6 +57,230 @@
 
 
 /**
+ * __drm_modeset_lock_all - internal helper to grab all modeset locks
+ * @dev: DRM device
+ * @trylock: trylock mode for atomic contexts
+ *
+ * This is a special version of drm_modeset_lock_all() which can also be used in
+ * atomic contexts. Then @trylock must be set to true.
+ *
+ * Returns:
+ * 0 on success or negative error code on failure.
+ */
+int __drm_modeset_lock_all(struct drm_device *dev,
+			   bool trylock)
+{
+	struct drm_mode_config *config = &dev->mode_config;
+	struct drm_modeset_acquire_ctx *ctx;
+	int ret;
+
+	ctx = kzalloc(sizeof(*ctx),
+		      trylock ? GFP_ATOMIC : GFP_KERNEL);
+	if (!ctx)
+		return -ENOMEM;
+
+	if (trylock) {
+		if (!mutex_trylock(&config->mutex))
+			return -EBUSY;
+	} else {
+		mutex_lock(&config->mutex);
+	}
+
+	drm_modeset_acquire_init(ctx, 0);
+	ctx->trylock_only = trylock;
+
+retry:
+	ret = drm_modeset_lock(&config->connection_mutex, ctx);
+	if (ret)
+		goto fail;
+	ret = drm_modeset_lock_all_crtcs(dev, ctx);
+	if (ret)
+		goto fail;
+
+	WARN_ON(config->acquire_ctx);
+
+	/* now we hold the locks, so now that it is safe, stash the
+	 * ctx for drm_modeset_unlock_all():
+	 */
+	config->acquire_ctx = ctx;
+
+	drm_warn_on_modeset_not_all_locked(dev);
+
+	return 0;
+
+fail:
+	if (ret == -EDEADLK) {
+		drm_modeset_backoff(ctx);
+		goto retry;
+	}
+
+	return ret;
+}
+EXPORT_SYMBOL(__drm_modeset_lock_all);
+
+/**
+ * drm_modeset_lock_all - take all modeset locks
+ * @dev: drm device
+ *
+ * This function takes all modeset locks, suitable where a more fine-grained
+ * scheme isn't (yet) implemented. Locks must be dropped with
+ * drm_modeset_unlock_all.
+ */
+void drm_modeset_lock_all(struct drm_device *dev)
+{
+	WARN_ON(__drm_modeset_lock_all(dev, false) != 0);
+}
+EXPORT_SYMBOL(drm_modeset_lock_all);
+
+/**
+ * drm_modeset_unlock_all - drop all modeset locks
+ * @dev: device
+ *
+ * This function drop all modeset locks taken by drm_modeset_lock_all.
+ */
+void drm_modeset_unlock_all(struct drm_device *dev)
+{
+	struct drm_mode_config *config = &dev->mode_config;
+	struct drm_modeset_acquire_ctx *ctx = config->acquire_ctx;
+
+	if (WARN_ON(!ctx))
+		return;
+
+	config->acquire_ctx = NULL;
+	drm_modeset_drop_locks(ctx);
+	drm_modeset_acquire_fini(ctx);
+
+	kfree(ctx);
+
+	mutex_unlock(&dev->mode_config.mutex);
+}
+EXPORT_SYMBOL(drm_modeset_unlock_all);
+
+/**
+ * drm_modeset_lock_crtc - lock crtc with hidden acquire ctx for a plane update
+ * @crtc: DRM CRTC
+ * @plane: DRM plane to be updated on @crtc
+ *
+ * This function locks the given crtc and plane (which should be either the
+ * primary or cursor plane) using a hidden acquire context. This is necessary so
+ * that drivers internally using the atomic interfaces can grab further locks
+ * with the lock acquire context.
+ *
+ * Note that @plane can be NULL, e.g. when the cursor support hasn't yet been
+ * converted to universal planes yet.
+ */
+void drm_modeset_lock_crtc(struct drm_crtc *crtc,
+			   struct drm_plane *plane)
+{
+	struct drm_modeset_acquire_ctx *ctx;
+	int ret;
+
+	ctx = kzalloc(sizeof(*ctx), GFP_KERNEL);
+	if (WARN_ON(!ctx))
+		return;
+
+	drm_modeset_acquire_init(ctx, 0);
+
+retry:
+	ret = drm_modeset_lock(&crtc->mutex, ctx);
+	if (ret)
+		goto fail;
+
+	if (plane) {
+		ret = drm_modeset_lock(&plane->mutex, ctx);
+		if (ret)
+			goto fail;
+
+		if (plane->crtc) {
+			ret = drm_modeset_lock(&plane->crtc->mutex, ctx);
+			if (ret)
+				goto fail;
+		}
+	}
+
+	WARN_ON(crtc->acquire_ctx);
+
+	/* now we hold the locks, so now that it is safe, stash the
+	 * ctx for drm_modeset_unlock_crtc():
+	 */
+	crtc->acquire_ctx = ctx;
+
+	return;
+
+fail:
+	if (ret == -EDEADLK) {
+		drm_modeset_backoff(ctx);
+		goto retry;
+	}
+}
+EXPORT_SYMBOL(drm_modeset_lock_crtc);
+
+/**
+ * drm_modeset_legacy_acquire_ctx - find acquire ctx for legacy ioctls
+ * @crtc: drm crtc
+ *
+ * Legacy ioctl operations like cursor updates or page flips only have per-crtc
+ * locking, and store the acquire ctx in the corresponding crtc. All other
+ * legacy operations take all locks and use a global acquire context. This
+ * function grabs the right one.
+ */
+struct drm_modeset_acquire_ctx *
+drm_modeset_legacy_acquire_ctx(struct drm_crtc *crtc)
+{
+	if (crtc->acquire_ctx)
+		return crtc->acquire_ctx;
+
+	WARN_ON(!crtc->dev->mode_config.acquire_ctx);
+
+	return crtc->dev->mode_config.acquire_ctx;
+}
+EXPORT_SYMBOL(drm_modeset_legacy_acquire_ctx);
+
+/**
+ * drm_modeset_unlock_crtc - drop crtc lock
+ * @crtc: drm crtc
+ *
+ * This drops the crtc lock acquire with drm_modeset_lock_crtc() and all other
+ * locks acquired through the hidden context.
+ */
+void drm_modeset_unlock_crtc(struct drm_crtc *crtc)
+{
+	struct drm_modeset_acquire_ctx *ctx = crtc->acquire_ctx;
+
+	if (WARN_ON(!ctx))
+		return;
+
+	crtc->acquire_ctx = NULL;
+	drm_modeset_drop_locks(ctx);
+	drm_modeset_acquire_fini(ctx);
+
+	kfree(ctx);
+}
+EXPORT_SYMBOL(drm_modeset_unlock_crtc);
+
+/**
+ * drm_warn_on_modeset_not_all_locked - check that all modeset locks are locked
+ * @dev: device
+ *
+ * Useful as a debug assert.
+ */
+void drm_warn_on_modeset_not_all_locked(struct drm_device *dev)
+{
+	struct drm_crtc *crtc;
+
+	/* Locking is currently fubar in the panic handler. */
+	if (oops_in_progress)
+		return;
+
+	list_for_each_entry(crtc, &dev->mode_config.crtc_list, head)
+		WARN_ON(!drm_modeset_is_locked(&crtc->mutex));
+
+	WARN_ON(!drm_modeset_is_locked(&dev->mode_config.connection_mutex));
+	WARN_ON(!mutex_is_locked(&dev->mode_config.mutex));
+}
+EXPORT_SYMBOL(drm_warn_on_modeset_not_all_locked);
+
+/**
  * drm_modeset_acquire_init - initialize acquire context
  * @ctx: the acquire context
  * @flags: for future
@@ -108,7 +332,12 @@ static inline int modeset_lock(struct drm_modeset_lock *lock,
 
 	WARN_ON(ctx->contended);
 
-	if (interruptible && slow) {
+	if (ctx->trylock_only) {
+		if (!ww_mutex_trylock(&lock->mutex))
+			return -EBUSY;
+		else
+			return 0;
+	} else if (interruptible && slow) {
 		ret = ww_mutex_lock_slow_interruptible(&lock->mutex, &ctx->ww_ctx);
 	} else if (interruptible) {
 		ret = ww_mutex_lock_interruptible(&lock->mutex, &ctx->ww_ctx);
@@ -226,15 +455,14 @@ void drm_modeset_unlock(struct drm_modeset_lock *lock)
 }
 EXPORT_SYMBOL(drm_modeset_unlock);
 
-/* Temporary.. until we have sufficiently fine grained locking, there
- * are a couple scenarios where it is convenient to grab all crtc locks.
- * It is planned to remove this:
- */
+/* In some legacy codepaths it's convenient to just grab all the crtc and plane
+ * related locks. */
 int drm_modeset_lock_all_crtcs(struct drm_device *dev,
 		struct drm_modeset_acquire_ctx *ctx)
 {
 	struct drm_mode_config *config = &dev->mode_config;
 	struct drm_crtc *crtc;
+	struct drm_plane *plane;
 	int ret = 0;
 
 	list_for_each_entry(crtc, &config->crtc_list, head) {
@@ -243,6 +471,12 @@ int drm_modeset_lock_all_crtcs(struct drm_device *dev,
 			return ret;
 	}
 
+	list_for_each_entry(plane, &config->plane_list, head) {
+		ret = drm_modeset_lock(&plane->mutex, ctx);
+		if (ret)
+			return ret;
+	}
+
 	return 0;
 }
 EXPORT_SYMBOL(drm_modeset_lock_all_crtcs);
diff --git a/drivers/gpu/drm/drm_pci.c b/drivers/gpu/drm/drm_pci.c
index 020cfd934854..fd29f03645b8 100644
--- a/drivers/gpu/drm/drm_pci.c
+++ b/drivers/gpu/drm/drm_pci.c
@@ -27,6 +27,7 @@
 #include <linux/dma-mapping.h>
 #include <linux/export.h>
 #include <drm/drmP.h>
+#include "drm_legacy.h"
 
 /**
  * drm_pci_alloc - Allocate a PCI consistent memory block, for DMA.
@@ -81,7 +82,7 @@ EXPORT_SYMBOL(drm_pci_alloc);
  *
  * This function is for internal use in the Linux-specific DRM core code.
  */
-void __drm_pci_free(struct drm_device * dev, drm_dma_handle_t * dmah)
+void __drm_legacy_pci_free(struct drm_device * dev, drm_dma_handle_t * dmah)
 {
 	unsigned long addr;
 	size_t sz;
@@ -105,7 +106,7 @@ void __drm_pci_free(struct drm_device * dev, drm_dma_handle_t * dmah)
  */
 void drm_pci_free(struct drm_device * dev, drm_dma_handle_t * dmah)
 {
-	__drm_pci_free(dev, dmah);
+	__drm_legacy_pci_free(dev, dmah);
 	kfree(dmah);
 }
 
@@ -127,34 +128,20 @@ static int drm_get_pci_domain(struct drm_device *dev)
 	return pci_domain_nr(dev->pdev->bus);
 }
 
-static int drm_pci_set_busid(struct drm_device *dev, struct drm_master *master)
+int drm_pci_set_busid(struct drm_device *dev, struct drm_master *master)
 {
-	int len, ret;
-	master->unique_len = 40;
-	master->unique_size = master->unique_len;
-	master->unique = kmalloc(master->unique_size, GFP_KERNEL);
-	if (master->unique == NULL)
+	master->unique = kasprintf(GFP_KERNEL, "pci:%04x:%02x:%02x.%d",
+					drm_get_pci_domain(dev),
+					dev->pdev->bus->number,
+					PCI_SLOT(dev->pdev->devfn),
+					PCI_FUNC(dev->pdev->devfn));
+	if (!master->unique)
 		return -ENOMEM;
 
-
-	len = snprintf(master->unique, master->unique_len,
-		       "pci:%04x:%02x:%02x.%d",
-		       drm_get_pci_domain(dev),
-		       dev->pdev->bus->number,
-		       PCI_SLOT(dev->pdev->devfn),
-		       PCI_FUNC(dev->pdev->devfn));
-
-	if (len >= master->unique_len) {
-		DRM_ERROR("buffer overflow");
-		ret = -EINVAL;
-		goto err;
-	} else
-		master->unique_len = len;
-
+	master->unique_len = strlen(master->unique);
 	return 0;
-err:
-	return ret;
 }
+EXPORT_SYMBOL(drm_pci_set_busid);
 
 int drm_pci_set_unique(struct drm_device *dev,
 		       struct drm_master *master,
@@ -163,8 +150,7 @@ int drm_pci_set_unique(struct drm_device *dev,
 	int domain, bus, slot, func, ret;
 
 	master->unique_len = u->unique_len;
-	master->unique_size = u->unique_len + 1;
-	master->unique = kmalloc(master->unique_size, GFP_KERNEL);
+	master->unique = kmalloc(master->unique_len + 1, GFP_KERNEL);
 	if (!master->unique) {
 		ret = -ENOMEM;
 		goto err;
@@ -269,10 +255,6 @@ void drm_pci_agp_destroy(struct drm_device *dev)
 	}
 }
 
-static struct drm_bus drm_pci_bus = {
-	.set_busid = drm_pci_set_busid,
-};
-
 /**
  * drm_get_pci_dev - Register a PCI device with the DRM subsystem
  * @pdev: PCI device
@@ -353,8 +335,6 @@ int drm_pci_init(struct drm_driver *driver, struct pci_driver *pdriver)
 
 	DRM_DEBUG("\n");
 
-	driver->bus = &drm_pci_bus;
-
 	if (driver->driver_features & DRIVER_MODESET)
 		return pci_register_driver(pdriver);
 
diff --git a/drivers/gpu/drm/drm_plane_helper.c b/drivers/gpu/drm/drm_plane_helper.c
index 827ec1a3040b..18a1ac6ac22f 100644
--- a/drivers/gpu/drm/drm_plane_helper.c
+++ b/drivers/gpu/drm/drm_plane_helper.c
@@ -27,10 +27,38 @@
 #include <drm/drmP.h>
 #include <drm/drm_plane_helper.h>
 #include <drm/drm_rect.h>
-#include <drm/drm_plane_helper.h>
+#include <drm/drm_atomic.h>
+#include <drm/drm_crtc_helper.h>
+#include <drm/drm_atomic_helper.h>
 
 #define SUBPIXEL_MASK 0xffff
 
+/**
+ * DOC: overview
+ *
+ * This helper library has two parts. The first part has support to implement
+ * primary plane support on top of the normal CRTC configuration interface.
+ * Since the legacy ->set_config interface ties the primary plane together with
+ * the CRTC state this does not allow userspace to disable the primary plane
+ * itself.  To avoid too much duplicated code use
+ * drm_plane_helper_check_update() which can be used to enforce the same
+ * restrictions as primary planes had thus. The default primary plane only
+ * expose XRBG8888 and ARGB8888 as valid pixel formats for the attached
+ * framebuffer.
+ *
+ * Drivers are highly recommended to implement proper support for primary
+ * planes, and newly merged drivers must not rely upon these transitional
+ * helpers.
+ *
+ * The second part also implements transitional helpers which allow drivers to
+ * gradually switch to the atomic helper infrastructure for plane updates. Once
+ * that switch is complete drivers shouldn't use these any longer, instead using
+ * the proper legacy implementations for update and disable plane hooks provided
+ * by the atomic helpers.
+ *
+ * Again drivers are strongly urged to switch to the new interfaces.
+ */
+
 /*
  * This is the minimal list of formats that seem to be safe for modeset use
  * with all current DRM drivers.  Most hardware can actually support more
@@ -127,6 +155,11 @@ int drm_plane_helper_check_update(struct drm_plane *plane,
 		return -ERANGE;
 	}
 
+	if (!fb) {
+		*visible = false;
+		return 0;
+	}
+
 	*visible = drm_rect_clip_scaled(src, dest, clip, hscale, vscale);
 	if (!*visible)
 		/*
@@ -369,3 +402,171 @@ int drm_crtc_init(struct drm_device *dev, struct drm_crtc *crtc,
 	return drm_crtc_init_with_planes(dev, crtc, primary, NULL, funcs);
 }
 EXPORT_SYMBOL(drm_crtc_init);
+
+int drm_plane_helper_commit(struct drm_plane *plane,
+			    struct drm_plane_state *plane_state,
+			    struct drm_framebuffer *old_fb)
+{
+	struct drm_plane_helper_funcs *plane_funcs;
+	struct drm_crtc *crtc[2];
+	struct drm_crtc_helper_funcs *crtc_funcs[2];
+	int i, ret = 0;
+
+	plane_funcs = plane->helper_private;
+
+	/* Since this is a transitional helper we can't assume that plane->state
+	 * is always valid. Hence we need to use plane->crtc instead of
+	 * plane->state->crtc as the old crtc. */
+	crtc[0] = plane->crtc;
+	crtc[1] = crtc[0] != plane_state->crtc ? plane_state->crtc : NULL;
+
+	for (i = 0; i < 2; i++)
+		crtc_funcs[i] = crtc[i] ? crtc[i]->helper_private : NULL;
+
+	if (plane_funcs->atomic_check) {
+		ret = plane_funcs->atomic_check(plane, plane_state);
+		if (ret)
+			goto out;
+	}
+
+	if (plane_funcs->prepare_fb && plane_state->fb) {
+		ret = plane_funcs->prepare_fb(plane, plane_state->fb);
+		if (ret)
+			goto out;
+	}
+
+	/* Point of no return, commit sw state. */
+	swap(plane->state, plane_state);
+
+	for (i = 0; i < 2; i++) {
+		if (crtc_funcs[i] && crtc_funcs[i]->atomic_begin)
+			crtc_funcs[i]->atomic_begin(crtc[i]);
+	}
+
+	plane_funcs->atomic_update(plane, plane_state);
+
+	for (i = 0; i < 2; i++) {
+		if (crtc_funcs[i] && crtc_funcs[i]->atomic_flush)
+			crtc_funcs[i]->atomic_flush(crtc[i]);
+	}
+
+	for (i = 0; i < 2; i++) {
+		if (!crtc[i])
+			continue;
+
+		/* There's no other way to figure out whether the crtc is running. */
+		ret = drm_crtc_vblank_get(crtc[i]);
+		if (ret == 0) {
+			drm_crtc_wait_one_vblank(crtc[i]);
+			drm_crtc_vblank_put(crtc[i]);
+		}
+
+		ret = 0;
+	}
+
+	if (plane_funcs->cleanup_fb && old_fb)
+		plane_funcs->cleanup_fb(plane, old_fb);
+out:
+	if (plane_state) {
+		if (plane->funcs->atomic_destroy_state)
+			plane->funcs->atomic_destroy_state(plane, plane_state);
+		else
+			drm_atomic_helper_plane_destroy_state(plane, plane_state);
+	}
+
+	return ret;
+}
+
+/**
+ * drm_plane_helper_update() - Helper for primary plane update
+ * @plane: plane object to update
+ * @crtc: owning CRTC of owning plane
+ * @fb: framebuffer to flip onto plane
+ * @crtc_x: x offset of primary plane on crtc
+ * @crtc_y: y offset of primary plane on crtc
+ * @crtc_w: width of primary plane rectangle on crtc
+ * @crtc_h: height of primary plane rectangle on crtc
+ * @src_x: x offset of @fb for panning
+ * @src_y: y offset of @fb for panning
+ * @src_w: width of source rectangle in @fb
+ * @src_h: height of source rectangle in @fb
+ *
+ * Provides a default plane update handler using the atomic plane update
+ * functions. It is fully left to the driver to check plane constraints and
+ * handle corner-cases like a fully occluded or otherwise invisible plane.
+ *
+ * This is useful for piecewise transitioning of a driver to the atomic helpers.
+ *
+ * RETURNS:
+ * Zero on success, error code on failure
+ */
+int drm_plane_helper_update(struct drm_plane *plane, struct drm_crtc *crtc,
+			    struct drm_framebuffer *fb,
+			    int crtc_x, int crtc_y,
+			    unsigned int crtc_w, unsigned int crtc_h,
+			    uint32_t src_x, uint32_t src_y,
+			    uint32_t src_w, uint32_t src_h)
+{
+	struct drm_plane_state *plane_state;
+
+	if (plane->funcs->atomic_duplicate_state)
+		plane_state = plane->funcs->atomic_duplicate_state(plane);
+	else if (plane->state)
+		plane_state = drm_atomic_helper_plane_duplicate_state(plane);
+	else
+		plane_state = kzalloc(sizeof(*plane_state), GFP_KERNEL);
+	if (!plane_state)
+		return -ENOMEM;
+
+	plane_state->crtc = crtc;
+	drm_atomic_set_fb_for_plane(plane_state, fb);
+	plane_state->crtc_x = crtc_x;
+	plane_state->crtc_y = crtc_y;
+	plane_state->crtc_h = crtc_h;
+	plane_state->crtc_w = crtc_w;
+	plane_state->src_x = src_x;
+	plane_state->src_y = src_y;
+	plane_state->src_h = src_h;
+	plane_state->src_w = src_w;
+
+	return drm_plane_helper_commit(plane, plane_state, plane->fb);
+}
+EXPORT_SYMBOL(drm_plane_helper_update);
+
+/**
+ * drm_plane_helper_disable() - Helper for primary plane disable
+ * @plane: plane to disable
+ *
+ * Provides a default plane disable handler using the atomic plane update
+ * functions. It is fully left to the driver to check plane constraints and
+ * handle corner-cases like a fully occluded or otherwise invisible plane.
+ *
+ * This is useful for piecewise transitioning of a driver to the atomic helpers.
+ *
+ * RETURNS:
+ * Zero on success, error code on failure
+ */
+int drm_plane_helper_disable(struct drm_plane *plane)
+{
+	struct drm_plane_state *plane_state;
+
+	/* crtc helpers love to call disable functions for already disabled hw
+	 * functions. So cope with that. */
+	if (!plane->crtc)
+		return 0;
+
+	if (plane->funcs->atomic_duplicate_state)
+		plane_state = plane->funcs->atomic_duplicate_state(plane);
+	else if (plane->state)
+		plane_state = drm_atomic_helper_plane_duplicate_state(plane);
+	else
+		plane_state = kzalloc(sizeof(*plane_state), GFP_KERNEL);
+	if (!plane_state)
+		return -ENOMEM;
+
+	plane_state->crtc = NULL;
+	drm_atomic_set_fb_for_plane(plane_state, NULL);
+
+	return drm_plane_helper_commit(plane, plane_state, plane->fb);
+}
+EXPORT_SYMBOL(drm_plane_helper_disable);
diff --git a/drivers/gpu/drm/drm_platform.c b/drivers/gpu/drm/drm_platform.c
index d5b76f148c12..5314c9d5fef4 100644
--- a/drivers/gpu/drm/drm_platform.c
+++ b/drivers/gpu/drm/drm_platform.c
@@ -68,42 +68,23 @@ err_free:
 	return ret;
 }
 
-static int drm_platform_set_busid(struct drm_device *dev, struct drm_master *master)
+int drm_platform_set_busid(struct drm_device *dev, struct drm_master *master)
 {
-	int len, ret, id;
-
-	master->unique_len = 13 + strlen(dev->platformdev->name);
-	master->unique_size = master->unique_len;
-	master->unique = kmalloc(master->unique_len + 1, GFP_KERNEL);
-
-	if (master->unique == NULL)
-		return -ENOMEM;
+	int id;
 
 	id = dev->platformdev->id;
-
-	/* if only a single instance of the platform device, id will be
-	 * set to -1.. use 0 instead to avoid a funny looking bus-id:
-	 */
-	if (id == -1)
+	if (id < 0)
 		id = 0;
 
-	len = snprintf(master->unique, master->unique_len,
-			"platform:%s:%02d", dev->platformdev->name, id);
-
-	if (len > master->unique_len) {
-		DRM_ERROR("Unique buffer overflowed\n");
-		ret = -EINVAL;
-		goto err;
-	}
+	master->unique = kasprintf(GFP_KERNEL, "platform:%s:%02d",
+						dev->platformdev->name, id);
+	if (!master->unique)
+		return -ENOMEM;
 
+	master->unique_len = strlen(master->unique);
 	return 0;
-err:
-	return ret;
 }
-
-static struct drm_bus drm_platform_bus = {
-	.set_busid = drm_platform_set_busid,
-};
+EXPORT_SYMBOL(drm_platform_set_busid);
 
 /**
  * drm_platform_init - Register a platform device with the DRM subsystem
@@ -120,7 +101,6 @@ int drm_platform_init(struct drm_driver *driver, struct platform_device *platfor
 {
 	DRM_DEBUG("\n");
 
-	driver->bus = &drm_platform_bus;
 	return drm_get_platform_dev(platform_device, driver);
 }
 EXPORT_SYMBOL(drm_platform_init);
diff --git a/drivers/gpu/drm/drm_prime.c b/drivers/gpu/drm/drm_prime.c
index 99d578bad17e..7482b06cd08f 100644
--- a/drivers/gpu/drm/drm_prime.c
+++ b/drivers/gpu/drm/drm_prime.c
@@ -29,6 +29,9 @@
 #include <linux/export.h>
 #include <linux/dma-buf.h>
 #include <drm/drmP.h>
+#include <drm/drm_gem.h>
+
+#include "drm_internal.h"
 
 /*
  * DMA-BUF/GEM Object references and lifetime overview:
@@ -325,7 +328,7 @@ static const struct dma_buf_ops drm_gem_prime_dmabuf_ops =  {
  */
 
 /**
- * drm_gem_prime_export - helper library implemention of the export callback
+ * drm_gem_prime_export - helper library implementation of the export callback
  * @dev: drm_device to export from
  * @obj: GEM object to export
  * @flags: flags like DRM_CLOEXEC
@@ -480,7 +483,7 @@ out_unlock:
 EXPORT_SYMBOL(drm_gem_prime_handle_to_fd);
 
 /**
- * drm_gem_prime_import - helper library implemention of the import callback
+ * drm_gem_prime_import - helper library implementation of the import callback
  * @dev: drm_device to import into
  * @dma_buf: dma-buf object to import
  *
@@ -522,7 +525,7 @@ struct drm_gem_object *drm_gem_prime_import(struct drm_device *dev,
 		goto fail_detach;
 	}
 
-	obj = dev->driver->gem_prime_import_sg_table(dev, dma_buf->size, sgt);
+	obj = dev->driver->gem_prime_import_sg_table(dev, attach, sgt);
 	if (IS_ERR(obj)) {
 		ret = PTR_ERR(obj);
 		goto fail_unmap;
@@ -666,7 +669,7 @@ int drm_prime_fd_to_handle_ioctl(struct drm_device *dev, void *data,
  * the driver is responsible for mapping the pages into the
  * importers address space for use with dma_buf itself.
  */
-struct sg_table *drm_prime_pages_to_sg(struct page **pages, int nr_pages)
+struct sg_table *drm_prime_pages_to_sg(struct page **pages, unsigned int nr_pages)
 {
 	struct sg_table *sg = NULL;
 	int ret;
diff --git a/drivers/gpu/drm/drm_probe_helper.c b/drivers/gpu/drm/drm_probe_helper.c
index db7d250f7ac7..7483a47de8e4 100644
--- a/drivers/gpu/drm/drm_probe_helper.c
+++ b/drivers/gpu/drm/drm_probe_helper.c
@@ -82,6 +82,22 @@ static void drm_mode_validate_flag(struct drm_connector *connector,
 	return;
 }
 
+static int drm_helper_probe_add_cmdline_mode(struct drm_connector *connector)
+{
+	struct drm_display_mode *mode;
+
+	if (!connector->cmdline_mode.specified)
+		return 0;
+
+	mode = drm_mode_create_from_cmdline_mode(connector->dev,
+						 &connector->cmdline_mode);
+	if (mode == NULL)
+		return 0;
+
+	drm_mode_probed_add(connector, mode);
+	return 1;
+}
+
 static int drm_helper_probe_single_connector_modes_merge_bits(struct drm_connector *connector,
 							      uint32_t maxX, uint32_t maxY, bool merge_type_bits)
 {
@@ -102,7 +118,8 @@ static int drm_helper_probe_single_connector_modes_merge_bits(struct drm_connect
 		mode->status = MODE_UNVERIFIED;
 
 	if (connector->force) {
-		if (connector->force == DRM_FORCE_ON)
+		if (connector->force == DRM_FORCE_ON ||
+		    connector->force == DRM_FORCE_ON_DIGITAL)
 			connector->status = connector_status_connected;
 		else
 			connector->status = connector_status_disconnected;
@@ -141,6 +158,7 @@ static int drm_helper_probe_single_connector_modes_merge_bits(struct drm_connect
 
 	if (count == 0 && connector->status == connector_status_connected)
 		count = drm_add_modes_noedid(connector, 1024, 768);
+	count += drm_helper_probe_add_cmdline_mode(connector);
 	if (count == 0)
 		goto prune;
 
diff --git a/drivers/gpu/drm/drm_scatter.c b/drivers/gpu/drm/drm_scatter.c
index 1c78406f6e71..4f0f3b36d537 100644
--- a/drivers/gpu/drm/drm_scatter.c
+++ b/drivers/gpu/drm/drm_scatter.c
@@ -34,6 +34,7 @@
 #include <linux/vmalloc.h>
 #include <linux/slab.h>
 #include <drm/drmP.h>
+#include "drm_legacy.h"
 
 #define DEBUG_SCATTER 0
 
@@ -78,8 +79,8 @@ void drm_legacy_sg_cleanup(struct drm_device *dev)
 # define ScatterHandle(x) (unsigned int)(x)
 #endif
 
-int drm_sg_alloc(struct drm_device *dev, void *data,
-		 struct drm_file *file_priv)
+int drm_legacy_sg_alloc(struct drm_device *dev, void *data,
+			struct drm_file *file_priv)
 {
 	struct drm_scatter_gather *request = data;
 	struct drm_sg_mem *entry;
@@ -194,8 +195,8 @@ int drm_sg_alloc(struct drm_device *dev, void *data,
 	return -ENOMEM;
 }
 
-int drm_sg_free(struct drm_device *dev, void *data,
-		struct drm_file *file_priv)
+int drm_legacy_sg_free(struct drm_device *dev, void *data,
+		       struct drm_file *file_priv)
 {
 	struct drm_scatter_gather *request = data;
 	struct drm_sg_mem *entry;
diff --git a/drivers/gpu/drm/drm_sysfs.c b/drivers/gpu/drm/drm_sysfs.c
index ab1a5f6dde8a..cc3d6d6d67e0 100644
--- a/drivers/gpu/drm/drm_sysfs.c
+++ b/drivers/gpu/drm/drm_sysfs.c
@@ -21,6 +21,7 @@
 #include <drm/drm_sysfs.h>
 #include <drm/drm_core.h>
 #include <drm/drmP.h>
+#include "drm_internal.h"
 
 #define to_drm_minor(d) dev_get_drvdata(d)
 #define to_drm_connector(d) dev_get_drvdata(d)
diff --git a/drivers/gpu/drm/drm_usb.c b/drivers/gpu/drm/drm_usb.c
deleted file mode 100644
index f2fe94aab901..000000000000
--- a/drivers/gpu/drm/drm_usb.c
+++ /dev/null
@@ -1,88 +0,0 @@
-#include <drm/drmP.h>
-#include <drm/drm_usb.h>
-#include <linux/usb.h>
-#include <linux/module.h>
-
-int drm_get_usb_dev(struct usb_interface *interface,
-		    const struct usb_device_id *id,
-		    struct drm_driver *driver)
-{
-	struct drm_device *dev;
-	int ret;
-
-	DRM_DEBUG("\n");
-
-	dev = drm_dev_alloc(driver, &interface->dev);
-	if (!dev)
-		return -ENOMEM;
-
-	dev->usbdev = interface_to_usbdev(interface);
-	usb_set_intfdata(interface, dev);
-
-	ret = drm_dev_register(dev, 0);
-	if (ret)
-		goto err_free;
-
-	DRM_INFO("Initialized %s %d.%d.%d %s on minor %d\n",
-		 driver->name, driver->major, driver->minor, driver->patchlevel,
-		 driver->date, dev->primary->index);
-
-	return 0;
-
-err_free:
-	drm_dev_unref(dev);
-	return ret;
-
-}
-EXPORT_SYMBOL(drm_get_usb_dev);
-
-static int drm_usb_set_busid(struct drm_device *dev,
-			       struct drm_master *master)
-{
-	return 0;
-}
-
-static struct drm_bus drm_usb_bus = {
-	.set_busid = drm_usb_set_busid,
-};
-
-/**
- * drm_usb_init - Register matching USB devices with the DRM subsystem
- * @driver: DRM device driver
- * @udriver: USB device driver
- *
- * Registers one or more devices matched by a USB driver with the DRM
- * subsystem.
- *
- * Return: 0 on success or a negative error code on failure.
- */
-int drm_usb_init(struct drm_driver *driver, struct usb_driver *udriver)
-{
-	int res;
-	DRM_DEBUG("\n");
-
-	driver->bus = &drm_usb_bus;
-
-	res = usb_register(udriver);
-	return res;
-}
-EXPORT_SYMBOL(drm_usb_init);
-
-/**
- * drm_usb_exit - Unregister matching USB devices from the DRM subsystem
- * @driver: DRM device driver
- * @udriver: USB device driver
- *
- * Unregisters one or more devices matched by a USB driver from the DRM
- * subsystem.
- */
-void drm_usb_exit(struct drm_driver *driver,
-		  struct usb_driver *udriver)
-{
-	usb_deregister(udriver);
-}
-EXPORT_SYMBOL(drm_usb_exit);
-
-MODULE_AUTHOR("David Airlie");
-MODULE_DESCRIPTION("USB DRM support");
-MODULE_LICENSE("GPL and additional rights");
diff --git a/drivers/gpu/drm/drm_vm.c b/drivers/gpu/drm/drm_vm.c
index 24e045c4f531..4a2c328959e5 100644
--- a/drivers/gpu/drm/drm_vm.c
+++ b/drivers/gpu/drm/drm_vm.c
@@ -35,10 +35,19 @@
 
 #include <drm/drmP.h>
 #include <linux/export.h>
+#include <linux/seq_file.h>
 #if defined(__ia64__)
 #include <linux/efi.h>
 #include <linux/slab.h>
 #endif
+#include <asm/pgtable.h>
+#include "drm_legacy.h"
+
+struct drm_vma_entry {
+	struct list_head head;
+	struct vm_area_struct *vma;
+	pid_t pid;
+};
 
 static void drm_vm_open(struct vm_area_struct *vma);
 static void drm_vm_close(struct vm_area_struct *vma);
@@ -48,15 +57,11 @@ static pgprot_t drm_io_prot(struct drm_local_map *map,
 {
 	pgprot_t tmp = vm_get_page_prot(vma->vm_flags);
 
-#if defined(__i386__) || defined(__x86_64__)
+#if defined(__i386__) || defined(__x86_64__) || defined(__powerpc__)
 	if (map->type == _DRM_REGISTERS && !(map->flags & _DRM_WRITE_COMBINING))
 		tmp = pgprot_noncached(tmp);
 	else
 		tmp = pgprot_writecombine(tmp);
-#elif defined(__powerpc__)
-	pgprot_val(tmp) |= _PAGE_NO_CACHE;
-	if (map->type == _DRM_REGISTERS)
-		pgprot_val(tmp) |= _PAGE_GUARDED;
 #elif defined(__ia64__)
 	if (efi_range_is_wc(vma->vm_start, vma->vm_end -
 				    vma->vm_start))
@@ -263,7 +268,7 @@ static void drm_vm_shm_close(struct vm_area_struct *vma)
 				dmah.vaddr = map->handle;
 				dmah.busaddr = map->offset;
 				dmah.size = map->size;
-				__drm_pci_free(dev, &dmah);
+				__drm_legacy_pci_free(dev, &dmah);
 				break;
 			}
 			kfree(map);
@@ -412,7 +417,6 @@ void drm_vm_open_locked(struct drm_device *dev,
 		list_add(&vma_entry->head, &dev->vmalist);
 	}
 }
-EXPORT_SYMBOL_GPL(drm_vm_open_locked);
 
 static void drm_vm_open(struct vm_area_struct *vma)
 {
@@ -532,7 +536,7 @@ static resource_size_t drm_core_get_reg_ofs(struct drm_device *dev)
  * according to the mapping type and remaps the pages. Finally sets the file
  * pointer and calls vm_open().
  */
-int drm_mmap_locked(struct file *filp, struct vm_area_struct *vma)
+static int drm_mmap_locked(struct file *filp, struct vm_area_struct *vma)
 {
 	struct drm_file *priv = filp->private_data;
 	struct drm_device *dev = priv->minor->dev;
@@ -646,7 +650,7 @@ int drm_mmap_locked(struct file *filp, struct vm_area_struct *vma)
 	return 0;
 }
 
-int drm_mmap(struct file *filp, struct vm_area_struct *vma)
+int drm_legacy_mmap(struct file *filp, struct vm_area_struct *vma)
 {
 	struct drm_file *priv = filp->private_data;
 	struct drm_device *dev = priv->minor->dev;
@@ -661,4 +665,69 @@ int drm_mmap(struct file *filp, struct vm_area_struct *vma)
 
 	return ret;
 }
-EXPORT_SYMBOL(drm_mmap);
+EXPORT_SYMBOL(drm_legacy_mmap);
+
+void drm_legacy_vma_flush(struct drm_device *dev)
+{
+	struct drm_vma_entry *vma, *vma_temp;
+
+	/* Clear vma list (only needed for legacy drivers) */
+	list_for_each_entry_safe(vma, vma_temp, &dev->vmalist, head) {
+		list_del(&vma->head);
+		kfree(vma);
+	}
+}
+
+int drm_vma_info(struct seq_file *m, void *data)
+{
+	struct drm_info_node *node = (struct drm_info_node *) m->private;
+	struct drm_device *dev = node->minor->dev;
+	struct drm_vma_entry *pt;
+	struct vm_area_struct *vma;
+	unsigned long vma_count = 0;
+#if defined(__i386__)
+	unsigned int pgprot;
+#endif
+
+	mutex_lock(&dev->struct_mutex);
+	list_for_each_entry(pt, &dev->vmalist, head)
+		vma_count++;
+
+	seq_printf(m, "vma use count: %lu, high_memory = %pK, 0x%pK\n",
+		   vma_count, high_memory,
+		   (void *)(unsigned long)virt_to_phys(high_memory));
+
+	list_for_each_entry(pt, &dev->vmalist, head) {
+		vma = pt->vma;
+		if (!vma)
+			continue;
+		seq_printf(m,
+			   "\n%5d 0x%pK-0x%pK %c%c%c%c%c%c 0x%08lx000",
+			   pt->pid,
+			   (void *)vma->vm_start, (void *)vma->vm_end,
+			   vma->vm_flags & VM_READ ? 'r' : '-',
+			   vma->vm_flags & VM_WRITE ? 'w' : '-',
+			   vma->vm_flags & VM_EXEC ? 'x' : '-',
+			   vma->vm_flags & VM_MAYSHARE ? 's' : 'p',
+			   vma->vm_flags & VM_LOCKED ? 'l' : '-',
+			   vma->vm_flags & VM_IO ? 'i' : '-',
+			   vma->vm_pgoff);
+
+#if defined(__i386__)
+		pgprot = pgprot_val(vma->vm_page_prot);
+		seq_printf(m, " %c%c%c%c%c%c%c%c%c",
+			   pgprot & _PAGE_PRESENT ? 'p' : '-',
+			   pgprot & _PAGE_RW ? 'w' : 'r',
+			   pgprot & _PAGE_USER ? 'u' : 's',
+			   pgprot & _PAGE_PWT ? 't' : 'b',
+			   pgprot & _PAGE_PCD ? 'u' : 'c',
+			   pgprot & _PAGE_ACCESSED ? 'a' : '-',
+			   pgprot & _PAGE_DIRTY ? 'd' : '-',
+			   pgprot & _PAGE_PSE ? 'm' : 'k',
+			   pgprot & _PAGE_GLOBAL ? 'g' : 'l');
+#endif
+		seq_printf(m, "\n");
+	}
+	mutex_unlock(&dev->struct_mutex);
+	return 0;
+}
diff --git a/drivers/gpu/drm/exynos/exynos_dp_core.c b/drivers/gpu/drm/exynos/exynos_dp_core.c
index 4f3c7eb2d37d..34d46aa75416 100644
--- a/drivers/gpu/drm/exynos/exynos_dp_core.c
+++ b/drivers/gpu/drm/exynos/exynos_dp_core.c
@@ -30,12 +30,17 @@
 #include <drm/drm_panel.h>
 #include <drm/bridge/ptn3460.h>
 
-#include "exynos_drm_drv.h"
 #include "exynos_dp_core.h"
 
 #define ctx_from_connector(c)	container_of(c, struct exynos_dp_device, \
 					connector)
 
+static inline struct exynos_dp_device *
+display_to_dp(struct exynos_drm_display *d)
+{
+	return container_of(d, struct exynos_dp_device, display);
+}
+
 struct bridge_init {
 	struct i2c_client *client;
 	struct device_node *node;
@@ -329,8 +334,8 @@ static int exynos_dp_link_start(struct exynos_dp_device *dp)
 		return retval;
 
 	for (lane = 0; lane < lane_count; lane++)
-		buf[lane] = DP_TRAIN_PRE_EMPHASIS_0 |
-			    DP_TRAIN_VOLTAGE_SWING_400;
+		buf[lane] = DP_TRAIN_PRE_EMPH_LEVEL_0 |
+			    DP_TRAIN_VOLTAGE_SWING_LEVEL_0;
 
 	retval = exynos_dp_write_bytes_to_dpcd(dp, DP_TRAINING_LANE0_SET,
 			lane_count, buf);
@@ -882,7 +887,7 @@ static void exynos_dp_hotplug(struct work_struct *work)
 
 static void exynos_dp_commit(struct exynos_drm_display *display)
 {
-	struct exynos_dp_device *dp = display->ctx;
+	struct exynos_dp_device *dp = display_to_dp(display);
 	int ret;
 
 	/* Keep the panel disabled while we configure video */
@@ -937,6 +942,8 @@ static enum drm_connector_status exynos_dp_detect(
 
 static void exynos_dp_connector_destroy(struct drm_connector *connector)
 {
+	drm_connector_unregister(connector);
+	drm_connector_cleanup(connector);
 }
 
 static struct drm_connector_funcs exynos_dp_connector_funcs = {
@@ -1018,7 +1025,7 @@ static int exynos_drm_attach_lcd_bridge(struct drm_device *dev,
 static int exynos_dp_create_connector(struct exynos_drm_display *display,
 				struct drm_encoder *encoder)
 {
-	struct exynos_dp_device *dp = display->ctx;
+	struct exynos_dp_device *dp = display_to_dp(display);
 	struct drm_connector *connector = &dp->connector;
 	int ret;
 
@@ -1050,33 +1057,19 @@ static int exynos_dp_create_connector(struct exynos_drm_display *display,
 
 static void exynos_dp_phy_init(struct exynos_dp_device *dp)
 {
-	if (dp->phy) {
+	if (dp->phy)
 		phy_power_on(dp->phy);
-	} else if (dp->phy_addr) {
-		u32 reg;
-
-		reg = __raw_readl(dp->phy_addr);
-		reg |= dp->enable_mask;
-		__raw_writel(reg, dp->phy_addr);
-	}
 }
 
 static void exynos_dp_phy_exit(struct exynos_dp_device *dp)
 {
-	if (dp->phy) {
+	if (dp->phy)
 		phy_power_off(dp->phy);
-	} else if (dp->phy_addr) {
-		u32 reg;
-
-		reg = __raw_readl(dp->phy_addr);
-		reg &= ~(dp->enable_mask);
-		__raw_writel(reg, dp->phy_addr);
-	}
 }
 
 static void exynos_dp_poweron(struct exynos_drm_display *display)
 {
-	struct exynos_dp_device *dp = display->ctx;
+	struct exynos_dp_device *dp = display_to_dp(display);
 
 	if (dp->dpms_mode == DRM_MODE_DPMS_ON)
 		return;
@@ -1097,7 +1090,7 @@ static void exynos_dp_poweron(struct exynos_drm_display *display)
 
 static void exynos_dp_poweroff(struct exynos_drm_display *display)
 {
-	struct exynos_dp_device *dp = display->ctx;
+	struct exynos_dp_device *dp = display_to_dp(display);
 
 	if (dp->dpms_mode != DRM_MODE_DPMS_ON)
 		return;
@@ -1122,7 +1115,7 @@ static void exynos_dp_poweroff(struct exynos_drm_display *display)
 
 static void exynos_dp_dpms(struct exynos_drm_display *display, int mode)
 {
-	struct exynos_dp_device *dp = display->ctx;
+	struct exynos_dp_device *dp = display_to_dp(display);
 
 	switch (mode) {
 	case DRM_MODE_DPMS_ON:
@@ -1145,11 +1138,6 @@ static struct exynos_drm_display_ops exynos_dp_display_ops = {
 	.commit = exynos_dp_commit,
 };
 
-static struct exynos_drm_display exynos_dp_display = {
-	.type = EXYNOS_DISPLAY_TYPE_LCD,
-	.ops = &exynos_dp_display_ops,
-};
-
 static struct video_info *exynos_dp_dt_parse_pdata(struct device *dev)
 {
 	struct device_node *dp_node = dev->of_node;
@@ -1208,44 +1196,6 @@ static struct video_info *exynos_dp_dt_parse_pdata(struct device *dev)
 	return dp_video_config;
 }
 
-static int exynos_dp_dt_parse_phydata(struct exynos_dp_device *dp)
-{
-	struct device_node *dp_phy_node = of_node_get(dp->dev->of_node);
-	u32 phy_base;
-	int ret = 0;
-
-	dp_phy_node = of_find_node_by_name(dp_phy_node, "dptx-phy");
-	if (!dp_phy_node) {
-		dp->phy = devm_phy_get(dp->dev, "dp");
-		return PTR_ERR_OR_ZERO(dp->phy);
-	}
-
-	if (of_property_read_u32(dp_phy_node, "reg", &phy_base)) {
-		dev_err(dp->dev, "failed to get reg for dptx-phy\n");
-		ret = -EINVAL;
-		goto err;
-	}
-
-	if (of_property_read_u32(dp_phy_node, "samsung,enable-mask",
-				&dp->enable_mask)) {
-		dev_err(dp->dev, "failed to get enable-mask for dptx-phy\n");
-		ret = -EINVAL;
-		goto err;
-	}
-
-	dp->phy_addr = ioremap(phy_base, SZ_4);
-	if (!dp->phy_addr) {
-		dev_err(dp->dev, "failed to ioremap dp-phy\n");
-		ret = -ENOMEM;
-		goto err;
-	}
-
-err:
-	of_node_put(dp_phy_node);
-
-	return ret;
-}
-
 static int exynos_dp_dt_parse_panel(struct exynos_dp_device *dp)
 {
 	int ret;
@@ -1261,10 +1211,10 @@ static int exynos_dp_dt_parse_panel(struct exynos_dp_device *dp)
 
 static int exynos_dp_bind(struct device *dev, struct device *master, void *data)
 {
+	struct exynos_dp_device *dp = dev_get_drvdata(dev);
 	struct platform_device *pdev = to_platform_device(dev);
 	struct drm_device *drm_dev = data;
 	struct resource *res;
-	struct exynos_dp_device *dp = exynos_dp_display.ctx;
 	unsigned int irq_flags;
 	int ret = 0;
 
@@ -1275,9 +1225,21 @@ static int exynos_dp_bind(struct device *dev, struct device *master, void *data)
 	if (IS_ERR(dp->video_info))
 		return PTR_ERR(dp->video_info);
 
-	ret = exynos_dp_dt_parse_phydata(dp);
-	if (ret)
-		return ret;
+	dp->phy = devm_phy_get(dp->dev, "dp");
+	if (IS_ERR(dp->phy)) {
+		dev_err(dp->dev, "no DP phy configured\n");
+		ret = PTR_ERR(dp->phy);
+		if (ret) {
+			/*
+			 * phy itself is not enabled, so we can move forward
+			 * assigning NULL to phy pointer.
+			 */
+			if (ret == -ENOSYS || ret == -ENODEV)
+				dp->phy = NULL;
+			else
+				return ret;
+		}
+	}
 
 	if (!dp->panel) {
 		ret = exynos_dp_dt_parse_panel(dp);
@@ -1344,22 +1306,15 @@ static int exynos_dp_bind(struct device *dev, struct device *master, void *data)
 
 	dp->drm_dev = drm_dev;
 
-	platform_set_drvdata(pdev, &exynos_dp_display);
-
-	return exynos_drm_create_enc_conn(drm_dev, &exynos_dp_display);
+	return exynos_drm_create_enc_conn(drm_dev, &dp->display);
 }
 
 static void exynos_dp_unbind(struct device *dev, struct device *master,
 				void *data)
 {
-	struct exynos_drm_display *display = dev_get_drvdata(dev);
-	struct exynos_dp_device *dp = display->ctx;
-	struct drm_encoder *encoder = dp->encoder;
-
-	exynos_dp_dpms(display, DRM_MODE_DPMS_OFF);
+	struct exynos_dp_device *dp = dev_get_drvdata(dev);
 
-	encoder->funcs->destroy(encoder);
-	drm_connector_cleanup(&dp->connector);
+	exynos_dp_dpms(&dp->display, DRM_MODE_DPMS_OFF);
 }
 
 static const struct component_ops exynos_dp_ops = {
@@ -1374,16 +1329,20 @@ static int exynos_dp_probe(struct platform_device *pdev)
 	struct exynos_dp_device *dp;
 	int ret;
 
-	ret = exynos_drm_component_add(&pdev->dev, EXYNOS_DEVICE_TYPE_CONNECTOR,
-					exynos_dp_display.type);
-	if (ret)
-		return ret;
-
 	dp = devm_kzalloc(&pdev->dev, sizeof(struct exynos_dp_device),
 				GFP_KERNEL);
 	if (!dp)
 		return -ENOMEM;
 
+	dp->display.type = EXYNOS_DISPLAY_TYPE_LCD;
+	dp->display.ops = &exynos_dp_display_ops;
+	platform_set_drvdata(pdev, dp);
+
+	ret = exynos_drm_component_add(&pdev->dev, EXYNOS_DEVICE_TYPE_CONNECTOR,
+					dp->display.type);
+	if (ret)
+		return ret;
+
 	panel_node = of_parse_phandle(dev->of_node, "panel", 0);
 	if (panel_node) {
 		dp->panel = of_drm_find_panel(panel_node);
@@ -1392,8 +1351,6 @@ static int exynos_dp_probe(struct platform_device *pdev)
 			return -EPROBE_DEFER;
 	}
 
-	exynos_dp_display.ctx = dp;
-
 	ret = component_add(&pdev->dev, &exynos_dp_ops);
 	if (ret)
 		exynos_drm_component_del(&pdev->dev,
@@ -1413,19 +1370,17 @@ static int exynos_dp_remove(struct platform_device *pdev)
 #ifdef CONFIG_PM_SLEEP
 static int exynos_dp_suspend(struct device *dev)
 {
-	struct platform_device *pdev = to_platform_device(dev);
-	struct exynos_drm_display *display = platform_get_drvdata(pdev);
+	struct exynos_dp_device *dp = dev_get_drvdata(dev);
 
-	exynos_dp_dpms(display, DRM_MODE_DPMS_OFF);
+	exynos_dp_dpms(&dp->display, DRM_MODE_DPMS_OFF);
 	return 0;
 }
 
 static int exynos_dp_resume(struct device *dev)
 {
-	struct platform_device *pdev = to_platform_device(dev);
-	struct exynos_drm_display *display = platform_get_drvdata(pdev);
+	struct exynos_dp_device *dp = dev_get_drvdata(dev);
 
-	exynos_dp_dpms(display, DRM_MODE_DPMS_ON);
+	exynos_dp_dpms(&dp->display, DRM_MODE_DPMS_ON);
 	return 0;
 }
 #endif
diff --git a/drivers/gpu/drm/exynos/exynos_dp_core.h b/drivers/gpu/drm/exynos/exynos_dp_core.h
index a1aee6931bd7..164f171168e7 100644
--- a/drivers/gpu/drm/exynos/exynos_dp_core.h
+++ b/drivers/gpu/drm/exynos/exynos_dp_core.h
@@ -17,6 +17,8 @@
 #include <drm/drm_dp_helper.h>
 #include <drm/exynos_drm.h>
 
+#include "exynos_drm_drv.h"
+
 #define DP_TIMEOUT_LOOP_COUNT 100
 #define MAX_CR_LOOP 5
 #define MAX_EQ_LOOP 5
@@ -145,6 +147,7 @@ struct link_train {
 };
 
 struct exynos_dp_device {
+	struct exynos_drm_display display;
 	struct device		*dev;
 	struct drm_device	*drm_dev;
 	struct drm_connector	connector;
@@ -153,8 +156,6 @@ struct exynos_dp_device {
 	struct clk		*clock;
 	unsigned int		irq;
 	void __iomem		*reg_base;
-	void __iomem		*phy_addr;
-	unsigned int		enable_mask;
 
 	struct video_info	*video_info;
 	struct link_train	link_train;
diff --git a/drivers/gpu/drm/exynos/exynos_drm_crtc.c b/drivers/gpu/drm/exynos/exynos_drm_crtc.c
index b68e58f78cd1..45026e693225 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_crtc.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_crtc.c
@@ -32,7 +32,6 @@ enum exynos_crtc_mode {
  * Exynos specific crtc structure.
  *
  * @drm_crtc: crtc object.
- * @drm_plane: pointer of private plane object for this crtc
  * @manager: the manager associated with this crtc
  * @pipe: a crtc index created at load() with a new crtc object creation
  *	and the crtc object would be set to private->crtc array
@@ -46,7 +45,6 @@ enum exynos_crtc_mode {
  */
 struct exynos_drm_crtc {
 	struct drm_crtc			drm_crtc;
-	struct drm_plane		*plane;
 	struct exynos_drm_manager	*manager;
 	unsigned int			pipe;
 	unsigned int			dpms;
@@ -73,13 +71,16 @@ static void exynos_drm_crtc_dpms(struct drm_crtc *crtc, int mode)
 				!atomic_read(&exynos_crtc->pending_flip),
 				HZ/20))
 			atomic_set(&exynos_crtc->pending_flip, 0);
-		drm_vblank_off(crtc->dev, exynos_crtc->pipe);
+		drm_crtc_vblank_off(crtc);
 	}
 
 	if (manager->ops->dpms)
 		manager->ops->dpms(manager, mode);
 
 	exynos_crtc->dpms = mode;
+
+	if (mode == DRM_MODE_DPMS_ON)
+		drm_crtc_vblank_on(crtc);
 }
 
 static void exynos_drm_crtc_prepare(struct drm_crtc *crtc)
@@ -94,12 +95,12 @@ static void exynos_drm_crtc_commit(struct drm_crtc *crtc)
 
 	exynos_drm_crtc_dpms(crtc, DRM_MODE_DPMS_ON);
 
-	exynos_plane_commit(exynos_crtc->plane);
+	exynos_plane_commit(crtc->primary);
 
 	if (manager->ops->commit)
 		manager->ops->commit(manager);
 
-	exynos_plane_dpms(exynos_crtc->plane, DRM_MODE_DPMS_ON);
+	exynos_plane_dpms(crtc->primary, DRM_MODE_DPMS_ON);
 }
 
 static bool
@@ -123,10 +124,9 @@ exynos_drm_crtc_mode_set(struct drm_crtc *crtc, struct drm_display_mode *mode,
 {
 	struct exynos_drm_crtc *exynos_crtc = to_exynos_crtc(crtc);
 	struct exynos_drm_manager *manager = exynos_crtc->manager;
-	struct drm_plane *plane = exynos_crtc->plane;
+	struct drm_framebuffer *fb = crtc->primary->fb;
 	unsigned int crtc_w;
 	unsigned int crtc_h;
-	int ret;
 
 	/*
 	 * copy the mode data adjusted by mode_fixup() into crtc->mode
@@ -134,29 +134,21 @@ exynos_drm_crtc_mode_set(struct drm_crtc *crtc, struct drm_display_mode *mode,
 	 */
 	memcpy(&crtc->mode, adjusted_mode, sizeof(*adjusted_mode));
 
-	crtc_w = crtc->primary->fb->width - x;
-	crtc_h = crtc->primary->fb->height - y;
+	crtc_w = fb->width - x;
+	crtc_h = fb->height - y;
 
 	if (manager->ops->mode_set)
 		manager->ops->mode_set(manager, &crtc->mode);
 
-	ret = exynos_plane_mode_set(plane, crtc, crtc->primary->fb, 0, 0, crtc_w, crtc_h,
-				    x, y, crtc_w, crtc_h);
-	if (ret)
-		return ret;
-
-	plane->crtc = crtc;
-	plane->fb = crtc->primary->fb;
-	drm_framebuffer_reference(plane->fb);
-
-	return 0;
+	return exynos_plane_mode_set(crtc->primary, crtc, fb, 0, 0,
+				     crtc_w, crtc_h, x, y, crtc_w, crtc_h);
 }
 
 static int exynos_drm_crtc_mode_set_commit(struct drm_crtc *crtc, int x, int y,
 					  struct drm_framebuffer *old_fb)
 {
 	struct exynos_drm_crtc *exynos_crtc = to_exynos_crtc(crtc);
-	struct drm_plane *plane = exynos_crtc->plane;
+	struct drm_framebuffer *fb = crtc->primary->fb;
 	unsigned int crtc_w;
 	unsigned int crtc_h;
 	int ret;
@@ -167,11 +159,11 @@ static int exynos_drm_crtc_mode_set_commit(struct drm_crtc *crtc, int x, int y,
 		return -EPERM;
 	}
 
-	crtc_w = crtc->primary->fb->width - x;
-	crtc_h = crtc->primary->fb->height - y;
+	crtc_w = fb->width - x;
+	crtc_h = fb->height - y;
 
-	ret = exynos_plane_mode_set(plane, crtc, crtc->primary->fb, 0, 0, crtc_w, crtc_h,
-				    x, y, crtc_w, crtc_h);
+	ret = exynos_plane_mode_set(crtc->primary, crtc, fb, 0, 0,
+				    crtc_w, crtc_h, x, y, crtc_w, crtc_h);
 	if (ret)
 		return ret;
 
@@ -304,8 +296,7 @@ static int exynos_drm_crtc_set_property(struct drm_crtc *crtc,
 			exynos_drm_crtc_commit(crtc);
 			break;
 		case CRTC_MODE_BLANK:
-			exynos_plane_dpms(exynos_crtc->plane,
-					  DRM_MODE_DPMS_OFF);
+			exynos_plane_dpms(crtc->primary, DRM_MODE_DPMS_OFF);
 			break;
 		default:
 			break;
@@ -351,8 +342,10 @@ static void exynos_drm_crtc_attach_mode_property(struct drm_crtc *crtc)
 int exynos_drm_crtc_create(struct exynos_drm_manager *manager)
 {
 	struct exynos_drm_crtc *exynos_crtc;
+	struct drm_plane *plane;
 	struct exynos_drm_private *private = manager->drm_dev->dev_private;
 	struct drm_crtc *crtc;
+	int ret;
 
 	exynos_crtc = kzalloc(sizeof(*exynos_crtc), GFP_KERNEL);
 	if (!exynos_crtc)
@@ -364,11 +357,11 @@ int exynos_drm_crtc_create(struct exynos_drm_manager *manager)
 	exynos_crtc->dpms = DRM_MODE_DPMS_OFF;
 	exynos_crtc->manager = manager;
 	exynos_crtc->pipe = manager->pipe;
-	exynos_crtc->plane = exynos_plane_init(manager->drm_dev,
-				1 << manager->pipe, true);
-	if (!exynos_crtc->plane) {
-		kfree(exynos_crtc);
-		return -ENOMEM;
+	plane = exynos_plane_init(manager->drm_dev, 1 << manager->pipe,
+				  DRM_PLANE_TYPE_PRIMARY);
+	if (IS_ERR(plane)) {
+		ret = PTR_ERR(plane);
+		goto err_plane;
 	}
 
 	manager->crtc = &exynos_crtc->drm_crtc;
@@ -376,12 +369,22 @@ int exynos_drm_crtc_create(struct exynos_drm_manager *manager)
 
 	private->crtc[manager->pipe] = crtc;
 
-	drm_crtc_init(manager->drm_dev, crtc, &exynos_crtc_funcs);
+	ret = drm_crtc_init_with_planes(manager->drm_dev, crtc, plane, NULL,
+					&exynos_crtc_funcs);
+	if (ret < 0)
+		goto err_crtc;
+
 	drm_crtc_helper_add(crtc, &exynos_crtc_helper_funcs);
 
 	exynos_drm_crtc_attach_mode_property(crtc);
 
 	return 0;
+
+err_crtc:
+	plane->funcs->destroy(plane);
+err_plane:
+	kfree(exynos_crtc);
+	return ret;
 }
 
 int exynos_drm_crtc_enable_vblank(struct drm_device *dev, int pipe)
diff --git a/drivers/gpu/drm/exynos/exynos_drm_crtc.h b/drivers/gpu/drm/exynos/exynos_drm_crtc.h
index 690dcddab725..e353d353836f 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_crtc.h
+++ b/drivers/gpu/drm/exynos/exynos_drm_crtc.h
@@ -15,10 +15,7 @@
 #ifndef _EXYNOS_DRM_CRTC_H_
 #define _EXYNOS_DRM_CRTC_H_
 
-struct drm_device;
-struct drm_crtc;
-struct exynos_drm_manager;
-struct exynos_drm_overlay;
+#include "exynos_drm_drv.h"
 
 int exynos_drm_crtc_create(struct exynos_drm_manager *manager);
 int exynos_drm_crtc_enable_vblank(struct drm_device *dev, int pipe);
diff --git a/drivers/gpu/drm/exynos/exynos_drm_dpi.c b/drivers/gpu/drm/exynos/exynos_drm_dpi.c
index fa08f05e3e34..37678cf4425a 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_dpi.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_dpi.c
@@ -22,6 +22,7 @@
 #include "exynos_drm_drv.h"
 
 struct exynos_dpi {
+	struct exynos_drm_display display;
 	struct device *dev;
 	struct device_node *panel_node;
 
@@ -35,6 +36,11 @@ struct exynos_dpi {
 
 #define connector_to_dpi(c) container_of(c, struct exynos_dpi, connector)
 
+static inline struct exynos_dpi *display_to_dpi(struct exynos_drm_display *d)
+{
+	return container_of(d, struct exynos_dpi, display);
+}
+
 static enum drm_connector_status
 exynos_dpi_detect(struct drm_connector *connector, bool force)
 {
@@ -100,7 +106,7 @@ static struct drm_connector_helper_funcs exynos_dpi_connector_helper_funcs = {
 static int exynos_dpi_create_connector(struct exynos_drm_display *display,
 				       struct drm_encoder *encoder)
 {
-	struct exynos_dpi *ctx = display->ctx;
+	struct exynos_dpi *ctx = display_to_dpi(display);
 	struct drm_connector *connector = &ctx->connector;
 	int ret;
 
@@ -141,7 +147,7 @@ static void exynos_dpi_poweroff(struct exynos_dpi *ctx)
 
 static void exynos_dpi_dpms(struct exynos_drm_display *display, int mode)
 {
-	struct exynos_dpi *ctx = display->ctx;
+	struct exynos_dpi *ctx = display_to_dpi(display);
 
 	switch (mode) {
 	case DRM_MODE_DPMS_ON:
@@ -165,11 +171,6 @@ static struct exynos_drm_display_ops exynos_dpi_display_ops = {
 	.dpms = exynos_dpi_dpms
 };
 
-static struct exynos_drm_display exynos_dpi_display = {
-	.type = EXYNOS_DISPLAY_TYPE_LCD,
-	.ops = &exynos_dpi_display_ops,
-};
-
 /* of_* functions will be removed after merge of of_graph patches */
 static struct device_node *
 of_get_child_by_name_reg(struct device_node *parent, const char *name, u32 reg)
@@ -299,20 +300,21 @@ struct exynos_drm_display *exynos_dpi_probe(struct device *dev)
 	struct exynos_dpi *ctx;
 	int ret;
 
-	ret = exynos_drm_component_add(dev,
-					EXYNOS_DEVICE_TYPE_CONNECTOR,
-					exynos_dpi_display.type);
-	if (ret)
-		return ERR_PTR(ret);
-
 	ctx = devm_kzalloc(dev, sizeof(*ctx), GFP_KERNEL);
 	if (!ctx)
-		goto err_del_component;
+		return ERR_PTR(-ENOMEM);
 
+	ctx->display.type = EXYNOS_DISPLAY_TYPE_LCD;
+	ctx->display.ops = &exynos_dpi_display_ops;
 	ctx->dev = dev;
-	exynos_dpi_display.ctx = ctx;
 	ctx->dpms_mode = DRM_MODE_DPMS_OFF;
 
+	ret = exynos_drm_component_add(dev,
+					EXYNOS_DEVICE_TYPE_CONNECTOR,
+					ctx->display.type);
+	if (ret)
+		return ERR_PTR(ret);
+
 	ret = exynos_dpi_parse_dt(ctx);
 	if (ret < 0) {
 		devm_kfree(dev, ctx);
@@ -328,7 +330,7 @@ struct exynos_drm_display *exynos_dpi_probe(struct device *dev)
 		}
 	}
 
-	return &exynos_dpi_display;
+	return &ctx->display;
 
 err_del_component:
 	exynos_drm_component_del(dev, EXYNOS_DEVICE_TYPE_CONNECTOR);
@@ -336,16 +338,16 @@ err_del_component:
 	return NULL;
 }
 
-int exynos_dpi_remove(struct device *dev)
+int exynos_dpi_remove(struct exynos_drm_display *display)
 {
-	struct drm_encoder *encoder = exynos_dpi_display.encoder;
-	struct exynos_dpi *ctx = exynos_dpi_display.ctx;
+	struct exynos_dpi *ctx = display_to_dpi(display);
 
-	exynos_dpi_dpms(&exynos_dpi_display, DRM_MODE_DPMS_OFF);
-	encoder->funcs->destroy(encoder);
-	drm_connector_cleanup(&ctx->connector);
+	exynos_dpi_dpms(&ctx->display, DRM_MODE_DPMS_OFF);
 
-	exynos_drm_component_del(dev, EXYNOS_DEVICE_TYPE_CONNECTOR);
+	if (ctx->panel)
+		drm_panel_detach(ctx->panel);
+
+	exynos_drm_component_del(ctx->dev, EXYNOS_DEVICE_TYPE_CONNECTOR);
 
 	return 0;
 }
diff --git a/drivers/gpu/drm/exynos/exynos_drm_drv.c b/drivers/gpu/drm/exynos/exynos_drm_drv.c
index 0d74e9b99c4e..121470a83d1a 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_drv.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_drv.c
@@ -15,7 +15,6 @@
 #include <drm/drmP.h>
 #include <drm/drm_crtc_helper.h>
 
-#include <linux/anon_inodes.h>
 #include <linux/component.h>
 
 #include <drm/exynos_drm.h>
@@ -86,17 +85,14 @@ static int exynos_drm_load(struct drm_device *dev, unsigned long flags)
 		struct drm_plane *plane;
 		unsigned long possible_crtcs = (1 << MAX_CRTC) - 1;
 
-		plane = exynos_plane_init(dev, possible_crtcs, false);
-		if (!plane)
-			goto err_mode_config_cleanup;
-	}
-
-	/* init kms poll for handling hpd */
-	drm_kms_helper_poll_init(dev);
+		plane = exynos_plane_init(dev, possible_crtcs,
+					  DRM_PLANE_TYPE_OVERLAY);
+		if (!IS_ERR(plane))
+			continue;
 
-	ret = drm_vblank_init(dev, MAX_CRTC);
-	if (ret)
+		ret = PTR_ERR(plane);
 		goto err_mode_config_cleanup;
+	}
 
 	/* setup possible_clones. */
 	exynos_drm_encoder_setup(dev);
@@ -106,22 +102,46 @@ static int exynos_drm_load(struct drm_device *dev, unsigned long flags)
 	/* Try to bind all sub drivers. */
 	ret = component_bind_all(dev->dev, dev);
 	if (ret)
-		goto err_cleanup_vblank;
+		goto err_mode_config_cleanup;
+
+	ret = drm_vblank_init(dev, dev->mode_config.num_crtc);
+	if (ret)
+		goto err_unbind_all;
 
 	/* Probe non kms sub drivers and virtual display driver. */
 	ret = exynos_drm_device_subdrv_probe(dev);
 	if (ret)
-		goto err_unbind_all;
+		goto err_cleanup_vblank;
+
+	/*
+	 * enable drm irq mode.
+	 * - with irq_enabled = true, we can use the vblank feature.
+	 *
+	 * P.S. note that we wouldn't use drm irq handler but
+	 *	just specific driver own one instead because
+	 *	drm framework supports only one irq handler.
+	 */
+	dev->irq_enabled = true;
+
+	/*
+	 * with vblank_disable_allowed = true, vblank interrupt will be disabled
+	 * by drm timer once a current process gives up ownership of
+	 * vblank event.(after drm_vblank_put function is called)
+	 */
+	dev->vblank_disable_allowed = true;
+
+	/* init kms poll for handling hpd */
+	drm_kms_helper_poll_init(dev);
 
 	/* force connectors detection */
 	drm_helper_hpd_irq_event(dev);
 
 	return 0;
 
-err_unbind_all:
-	component_unbind_all(dev->dev, dev);
 err_cleanup_vblank:
 	drm_vblank_cleanup(dev);
+err_unbind_all:
+	component_unbind_all(dev->dev, dev);
 err_mode_config_cleanup:
 	drm_mode_config_cleanup(dev);
 	drm_release_iommu_mapping(dev);
@@ -136,23 +156,19 @@ static int exynos_drm_unload(struct drm_device *dev)
 	exynos_drm_device_subdrv_remove(dev);
 
 	exynos_drm_fbdev_fini(dev);
-	drm_vblank_cleanup(dev);
 	drm_kms_helper_poll_fini(dev);
-	drm_mode_config_cleanup(dev);
 
+	drm_vblank_cleanup(dev);
+	component_unbind_all(dev->dev, dev);
+	drm_mode_config_cleanup(dev);
 	drm_release_iommu_mapping(dev);
-	kfree(dev->dev_private);
 
-	component_unbind_all(dev->dev, dev);
+	kfree(dev->dev_private);
 	dev->dev_private = NULL;
 
 	return 0;
 }
 
-static const struct file_operations exynos_drm_gem_fops = {
-	.mmap = exynos_drm_gem_mmap_buffer,
-};
-
 static int exynos_drm_suspend(struct drm_device *dev, pm_message_t state)
 {
 	struct drm_connector *connector;
@@ -178,20 +194,21 @@ static int exynos_drm_resume(struct drm_device *dev)
 
 	drm_modeset_lock_all(dev);
 	list_for_each_entry(connector, &dev->mode_config.connector_list, head) {
-		if (connector->funcs->dpms)
-			connector->funcs->dpms(connector, connector->dpms);
+		if (connector->funcs->dpms) {
+			int dpms = connector->dpms;
+
+			connector->dpms = DRM_MODE_DPMS_OFF;
+			connector->funcs->dpms(connector, dpms);
+		}
 	}
 	drm_modeset_unlock_all(dev);
 
-	drm_helper_resume_force_mode(dev);
-
 	return 0;
 }
 
 static int exynos_drm_open(struct drm_device *dev, struct drm_file *file)
 {
 	struct drm_exynos_file_private *file_priv;
-	struct file *anon_filp;
 	int ret;
 
 	file_priv = kzalloc(sizeof(*file_priv), GFP_KERNEL);
@@ -204,21 +221,8 @@ static int exynos_drm_open(struct drm_device *dev, struct drm_file *file)
 	if (ret)
 		goto err_file_priv_free;
 
-	anon_filp = anon_inode_getfile("exynos_gem", &exynos_drm_gem_fops,
-					NULL, 0);
-	if (IS_ERR(anon_filp)) {
-		ret = PTR_ERR(anon_filp);
-		goto err_subdrv_close;
-	}
-
-	anon_filp->f_mode = FMODE_READ | FMODE_WRITE;
-	file_priv->anon_filp = anon_filp;
-
 	return ret;
 
-err_subdrv_close:
-	exynos_drm_subdrv_close(dev, file);
-
 err_file_priv_free:
 	kfree(file_priv);
 	file->driver_priv = NULL;
@@ -234,7 +238,6 @@ static void exynos_drm_preclose(struct drm_device *dev,
 static void exynos_drm_postclose(struct drm_device *dev, struct drm_file *file)
 {
 	struct exynos_drm_private *private = dev->dev_private;
-	struct drm_exynos_file_private *file_priv;
 	struct drm_pending_vblank_event *v, *vt;
 	struct drm_pending_event *e, *et;
 	unsigned long flags;
@@ -260,10 +263,6 @@ static void exynos_drm_postclose(struct drm_device *dev, struct drm_file *file)
 	}
 	spin_unlock_irqrestore(&dev->event_lock, flags);
 
-	file_priv = file->driver_priv;
-	if (file_priv->anon_filp)
-		fput(file_priv->anon_filp);
-
 	kfree(file->driver_priv);
 	file->driver_priv = NULL;
 }
@@ -282,11 +281,6 @@ static const struct vm_operations_struct exynos_drm_gem_vm_ops = {
 static const struct drm_ioctl_desc exynos_ioctls[] = {
 	DRM_IOCTL_DEF_DRV(EXYNOS_GEM_CREATE, exynos_drm_gem_create_ioctl,
 			DRM_UNLOCKED | DRM_AUTH),
-	DRM_IOCTL_DEF_DRV(EXYNOS_GEM_MAP_OFFSET,
-			exynos_drm_gem_map_offset_ioctl, DRM_UNLOCKED |
-			DRM_AUTH),
-	DRM_IOCTL_DEF_DRV(EXYNOS_GEM_MMAP,
-			exynos_drm_gem_mmap_ioctl, DRM_UNLOCKED | DRM_AUTH),
 	DRM_IOCTL_DEF_DRV(EXYNOS_GEM_GET,
 			exynos_drm_gem_get_ioctl, DRM_UNLOCKED),
 	DRM_IOCTL_DEF_DRV(EXYNOS_VIDI_CONNECTION,
@@ -330,6 +324,7 @@ static struct drm_driver exynos_drm_driver = {
 	.preclose		= exynos_drm_preclose,
 	.lastclose		= exynos_drm_lastclose,
 	.postclose		= exynos_drm_postclose,
+	.set_busid		= drm_platform_set_busid,
 	.get_vblank_counter	= drm_vblank_count,
 	.enable_vblank		= exynos_drm_crtc_enable_vblank,
 	.disable_vblank		= exynos_drm_crtc_disable_vblank,
@@ -478,28 +473,31 @@ void exynos_drm_component_del(struct device *dev,
 			list_del(&cdev->list);
 			kfree(cdev);
 		}
-
-		break;
 	}
 
 	mutex_unlock(&drm_component_lock);
 }
 
-static int compare_of(struct device *dev, void *data)
+static int compare_dev(struct device *dev, void *data)
 {
 	return dev == (struct device *)data;
 }
 
-static int exynos_drm_add_components(struct device *dev, struct master *m)
+static struct component_match *exynos_drm_match_add(struct device *dev)
 {
+	struct component_match *match = NULL;
 	struct component_dev *cdev;
 	unsigned int attach_cnt = 0;
 
 	mutex_lock(&drm_component_lock);
 
-	list_for_each_entry(cdev, &drm_component_list, list) {
-		int ret;
+	/* Do not retry to probe if there is no any kms driver regitered. */
+	if (list_empty(&drm_component_list)) {
+		mutex_unlock(&drm_component_lock);
+		return ERR_PTR(-ENODEV);
+	}
 
+	list_for_each_entry(cdev, &drm_component_list, list) {
 		/*
 		 * Add components to master only in case that crtc and
 		 * encoder/connector device objects exist.
@@ -514,16 +512,10 @@ static int exynos_drm_add_components(struct device *dev, struct master *m)
 		/*
 		 * fimd and dpi modules have same device object so add
 		 * only crtc device object in this case.
-		 *
-		 * TODO. if dpi module follows driver-model driver then
-		 * below codes can be removed.
 		 */
 		if (cdev->crtc_dev == cdev->conn_dev) {
-			ret = component_master_add_child(m, compare_of,
-					cdev->crtc_dev);
-			if (ret < 0)
-				return ret;
-
+			component_match_add(dev, &match, compare_dev,
+						cdev->crtc_dev);
 			goto out_lock;
 		}
 
@@ -533,11 +525,8 @@ static int exynos_drm_add_components(struct device *dev, struct master *m)
 		 * connector/encoder need pipe number of crtc when they
 		 * are created.
 		 */
-		ret = component_master_add_child(m, compare_of, cdev->crtc_dev);
-		ret |= component_master_add_child(m, compare_of,
-							cdev->conn_dev);
-		if (ret < 0)
-			return ret;
+		component_match_add(dev, &match, compare_dev, cdev->crtc_dev);
+		component_match_add(dev, &match, compare_dev, cdev->conn_dev);
 
 out_lock:
 		mutex_lock(&drm_component_lock);
@@ -545,7 +534,7 @@ out_lock:
 
 	mutex_unlock(&drm_component_lock);
 
-	return attach_cnt ? 0 : -ENODEV;
+	return attach_cnt ? match : ERR_PTR(-EPROBE_DEFER);
 }
 
 static int exynos_drm_bind(struct device *dev)
@@ -559,182 +548,76 @@ static void exynos_drm_unbind(struct device *dev)
 }
 
 static const struct component_master_ops exynos_drm_ops = {
-	.add_components = exynos_drm_add_components,
 	.bind		= exynos_drm_bind,
 	.unbind		= exynos_drm_unbind,
 };
 
-static int exynos_drm_platform_probe(struct platform_device *pdev)
-{
-	int ret;
-
-	pdev->dev.coherent_dma_mask = DMA_BIT_MASK(32);
-	exynos_drm_driver.num_ioctls = ARRAY_SIZE(exynos_ioctls);
-
+static struct platform_driver *const exynos_drm_kms_drivers[] = {
 #ifdef CONFIG_DRM_EXYNOS_FIMD
-	ret = platform_driver_register(&fimd_driver);
-	if (ret < 0)
-		return ret;
+	&fimd_driver,
 #endif
-
 #ifdef CONFIG_DRM_EXYNOS_DP
-	ret = platform_driver_register(&dp_driver);
-	if (ret < 0)
-		goto err_unregister_fimd_drv;
+	&dp_driver,
 #endif
-
 #ifdef CONFIG_DRM_EXYNOS_DSI
-	ret = platform_driver_register(&dsi_driver);
-	if (ret < 0)
-		goto err_unregister_dp_drv;
+	&dsi_driver,
 #endif
-
 #ifdef CONFIG_DRM_EXYNOS_HDMI
-	ret = platform_driver_register(&mixer_driver);
-	if (ret < 0)
-		goto err_unregister_dsi_drv;
-	ret = platform_driver_register(&hdmi_driver);
-	if (ret < 0)
-		goto err_unregister_mixer_drv;
+	&mixer_driver,
+	&hdmi_driver,
 #endif
+};
 
+static struct platform_driver *const exynos_drm_non_kms_drivers[] = {
 #ifdef CONFIG_DRM_EXYNOS_G2D
-	ret = platform_driver_register(&g2d_driver);
-	if (ret < 0)
-		goto err_unregister_hdmi_drv;
+	&g2d_driver,
 #endif
-
 #ifdef CONFIG_DRM_EXYNOS_FIMC
-	ret = platform_driver_register(&fimc_driver);
-	if (ret < 0)
-		goto err_unregister_g2d_drv;
+	&fimc_driver,
 #endif
-
 #ifdef CONFIG_DRM_EXYNOS_ROTATOR
-	ret = platform_driver_register(&rotator_driver);
-	if (ret < 0)
-		goto err_unregister_fimc_drv;
+	&rotator_driver,
 #endif
-
 #ifdef CONFIG_DRM_EXYNOS_GSC
-	ret = platform_driver_register(&gsc_driver);
-	if (ret < 0)
-		goto err_unregister_rotator_drv;
+	&gsc_driver,
 #endif
-
 #ifdef CONFIG_DRM_EXYNOS_IPP
-	ret = platform_driver_register(&ipp_driver);
-	if (ret < 0)
-		goto err_unregister_gsc_drv;
-
-	ret = exynos_platform_device_ipp_register();
-	if (ret < 0)
-		goto err_unregister_ipp_drv;
-#endif
-
-	ret = component_master_add(&pdev->dev, &exynos_drm_ops);
-	if (ret < 0)
-		DRM_DEBUG_KMS("re-tried by last sub driver probed later.\n");
-
-	return 0;
-
-#ifdef CONFIG_DRM_EXYNOS_IPP
-err_unregister_ipp_drv:
-	platform_driver_unregister(&ipp_driver);
-err_unregister_gsc_drv:
-#endif
-
-#ifdef CONFIG_DRM_EXYNOS_GSC
-	platform_driver_unregister(&gsc_driver);
-err_unregister_rotator_drv:
-#endif
-
-#ifdef CONFIG_DRM_EXYNOS_ROTATOR
-	platform_driver_unregister(&rotator_driver);
-err_unregister_fimc_drv:
-#endif
-
-#ifdef CONFIG_DRM_EXYNOS_FIMC
-	platform_driver_unregister(&fimc_driver);
-err_unregister_g2d_drv:
-#endif
-
-#ifdef CONFIG_DRM_EXYNOS_G2D
-	platform_driver_unregister(&g2d_driver);
-err_unregister_hdmi_drv:
+	&ipp_driver,
 #endif
+};
 
-#ifdef CONFIG_DRM_EXYNOS_HDMI
-	platform_driver_unregister(&hdmi_driver);
-err_unregister_mixer_drv:
-	platform_driver_unregister(&mixer_driver);
-err_unregister_dsi_drv:
-#endif
+static int exynos_drm_platform_probe(struct platform_device *pdev)
+{
+	struct component_match *match;
 
-#ifdef CONFIG_DRM_EXYNOS_DSI
-	platform_driver_unregister(&dsi_driver);
-err_unregister_dp_drv:
-#endif
+	pdev->dev.coherent_dma_mask = DMA_BIT_MASK(32);
+	exynos_drm_driver.num_ioctls = ARRAY_SIZE(exynos_ioctls);
 
-#ifdef CONFIG_DRM_EXYNOS_DP
-	platform_driver_unregister(&dp_driver);
-err_unregister_fimd_drv:
-#endif
+	match = exynos_drm_match_add(&pdev->dev);
+	if (IS_ERR(match)) {
+		return PTR_ERR(match);
+	}
 
-#ifdef CONFIG_DRM_EXYNOS_FIMD
-	platform_driver_unregister(&fimd_driver);
-#endif
-	return ret;
+	return component_master_add_with_match(&pdev->dev, &exynos_drm_ops,
+					       match);
 }
 
 static int exynos_drm_platform_remove(struct platform_device *pdev)
 {
-#ifdef CONFIG_DRM_EXYNOS_IPP
-	exynos_platform_device_ipp_unregister();
-	platform_driver_unregister(&ipp_driver);
-#endif
-
-#ifdef CONFIG_DRM_EXYNOS_GSC
-	platform_driver_unregister(&gsc_driver);
-#endif
-
-#ifdef CONFIG_DRM_EXYNOS_ROTATOR
-	platform_driver_unregister(&rotator_driver);
-#endif
-
-#ifdef CONFIG_DRM_EXYNOS_FIMC
-	platform_driver_unregister(&fimc_driver);
-#endif
-
-#ifdef CONFIG_DRM_EXYNOS_G2D
-	platform_driver_unregister(&g2d_driver);
-#endif
-
-#ifdef CONFIG_DRM_EXYNOS_HDMI
-	platform_driver_unregister(&mixer_driver);
-	platform_driver_unregister(&hdmi_driver);
-#endif
-
-#ifdef CONFIG_DRM_EXYNOS_FIMD
-	platform_driver_unregister(&fimd_driver);
-#endif
-
-#ifdef CONFIG_DRM_EXYNOS_DSI
-	platform_driver_unregister(&dsi_driver);
-#endif
-
-#ifdef CONFIG_DRM_EXYNOS_DP
-	platform_driver_unregister(&dp_driver);
-#endif
 	component_master_del(&pdev->dev, &exynos_drm_ops);
 	return 0;
 }
 
+static const char * const strings[] = {
+	"samsung,exynos3",
+	"samsung,exynos4",
+	"samsung,exynos5",
+};
+
 static struct platform_driver exynos_drm_platform_driver = {
 	.probe	= exynos_drm_platform_probe,
 	.remove	= exynos_drm_platform_remove,
 	.driver	= {
-		.owner	= THIS_MODULE,
 		.name	= "exynos-drm",
 		.pm	= &exynos_drm_pm_ops,
 	},
@@ -742,31 +625,87 @@ static struct platform_driver exynos_drm_platform_driver = {
 
 static int exynos_drm_init(void)
 {
-	int ret;
+	bool is_exynos = false;
+	int ret, i, j;
+
+	/*
+	 * Register device object only in case of Exynos SoC.
+	 *
+	 * Below codes resolves temporarily infinite loop issue incurred
+	 * by Exynos drm driver when using multi-platform kernel.
+	 * So these codes will be replaced with more generic way later.
+	 */
+	for (i = 0; i < ARRAY_SIZE(strings); i++) {
+		if (of_machine_is_compatible(strings[i])) {
+			is_exynos = true;
+			break;
+		}
+	}
+
+	if (!is_exynos)
+		return -ENODEV;
+
+	/*
+	 * Register device object only in case of Exynos SoC.
+	 *
+	 * Below codes resolves temporarily infinite loop issue incurred
+	 * by Exynos drm driver when using multi-platform kernel.
+	 * So these codes will be replaced with more generic way later.
+	 */
+	if (!of_machine_is_compatible("samsung,exynos3") &&
+			!of_machine_is_compatible("samsung,exynos4") &&
+			!of_machine_is_compatible("samsung,exynos5"))
+		return -ENODEV;
 
 	exynos_drm_pdev = platform_device_register_simple("exynos-drm", -1,
 								NULL, 0);
 	if (IS_ERR(exynos_drm_pdev))
 		return PTR_ERR(exynos_drm_pdev);
 
-#ifdef CONFIG_DRM_EXYNOS_VIDI
 	ret = exynos_drm_probe_vidi();
 	if (ret < 0)
 		goto err_unregister_pd;
+
+	for (i = 0; i < ARRAY_SIZE(exynos_drm_kms_drivers); ++i) {
+		ret = platform_driver_register(exynos_drm_kms_drivers[i]);
+		if (ret < 0)
+			goto err_unregister_kms_drivers;
+	}
+
+	for (j = 0; j < ARRAY_SIZE(exynos_drm_non_kms_drivers); ++j) {
+		ret = platform_driver_register(exynos_drm_non_kms_drivers[j]);
+		if (ret < 0)
+			goto err_unregister_non_kms_drivers;
+	}
+
+#ifdef CONFIG_DRM_EXYNOS_IPP
+	ret = exynos_platform_device_ipp_register();
+	if (ret < 0)
+		goto err_unregister_non_kms_drivers;
 #endif
 
 	ret = platform_driver_register(&exynos_drm_platform_driver);
 	if (ret)
-		goto err_remove_vidi;
+		goto err_unregister_resources;
 
 	return 0;
 
-err_remove_vidi:
-#ifdef CONFIG_DRM_EXYNOS_VIDI
+err_unregister_resources:
+#ifdef CONFIG_DRM_EXYNOS_IPP
+	exynos_platform_device_ipp_unregister();
+#endif
+
+err_unregister_non_kms_drivers:
+	while (--j >= 0)
+		platform_driver_unregister(exynos_drm_non_kms_drivers[j]);
+
+err_unregister_kms_drivers:
+	while (--i >= 0)
+		platform_driver_unregister(exynos_drm_kms_drivers[i]);
+
 	exynos_drm_remove_vidi();
 
 err_unregister_pd:
-#endif
 	platform_device_unregister(exynos_drm_pdev);
 
 	return ret;
@@ -774,10 +713,22 @@ err_unregister_pd:
 
 static void exynos_drm_exit(void)
 {
+	int i;
+
+#ifdef CONFIG_DRM_EXYNOS_IPP
+	exynos_platform_device_ipp_unregister();
+#endif
+
+	for (i = ARRAY_SIZE(exynos_drm_non_kms_drivers) - 1; i >= 0; --i)
+		platform_driver_unregister(exynos_drm_non_kms_drivers[i]);
+
+	for (i = ARRAY_SIZE(exynos_drm_kms_drivers) - 1; i >= 0; --i)
+		platform_driver_unregister(exynos_drm_kms_drivers[i]);
+
 	platform_driver_unregister(&exynos_drm_platform_driver);
-#ifdef CONFIG_DRM_EXYNOS_VIDI
+
 	exynos_drm_remove_vidi();
-#endif
+
 	platform_device_unregister(exynos_drm_pdev);
 }
 
diff --git a/drivers/gpu/drm/exynos/exynos_drm_drv.h b/drivers/gpu/drm/exynos/exynos_drm_drv.h
index 69a6fa397d75..2e5063488c50 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_drv.h
+++ b/drivers/gpu/drm/exynos/exynos_drm_drv.h
@@ -15,6 +15,7 @@
 #ifndef _EXYNOS_DRM_DRV_H_
 #define _EXYNOS_DRM_DRV_H_
 
+#include <drm/drmP.h>
 #include <linux/module.h>
 
 #define MAX_CRTC	3
@@ -22,24 +23,6 @@
 #define MAX_FB_BUFFER	4
 #define DEFAULT_ZPOS	-1
 
-#define _wait_for(COND, MS) ({ \
-	unsigned long timeout__ = jiffies + msecs_to_jiffies(MS);	\
-	int ret__ = 0;							\
-	while (!(COND)) {						\
-		if (time_after(jiffies, timeout__)) {			\
-			ret__ = -ETIMEDOUT;				\
-			break;						\
-		}							\
-	}								\
-	ret__;								\
-})
-
-#define wait_for(COND, MS) _wait_for(COND, MS)
-
-struct drm_device;
-struct exynos_drm_overlay;
-struct drm_connector;
-
 /* This enumerates device type. */
 enum exynos_drm_device_type {
 	EXYNOS_DEVICE_TYPE_NONE,
@@ -83,10 +66,10 @@ enum exynos_drm_output_type {
  * @dma_addr: array of bus(accessed by dma) address to the memory region
  *	      allocated for a overlay.
  * @zpos: order of overlay layer(z position).
- * @default_win: a window to be enabled.
- * @color_key: color key on or off.
  * @index_color: if using color key feature then this value would be used
  *			as index color.
+ * @default_win: a window to be enabled.
+ * @color_key: color key on or off.
  * @local_path: in case of lcd type, local path mode on or off.
  * @transparency: transparency on or off.
  * @activated: activated or not.
@@ -114,19 +97,20 @@ struct exynos_drm_overlay {
 	uint32_t pixel_format;
 	dma_addr_t dma_addr[MAX_FB_BUFFER];
 	int zpos;
-
-	bool default_win;
-	bool color_key;
 	unsigned int index_color;
-	bool local_path;
-	bool transparency;
-	bool activated;
+
+	bool default_win:1;
+	bool color_key:1;
+	bool local_path:1;
+	bool transparency:1;
+	bool activated:1;
 };
 
 /*
  * Exynos DRM Display Structure.
  *	- this structure is common to analog tv, digital tv and lcd panel.
  *
+ * @create_connector: initialize and register a new connector
  * @remove: cleans up the display for removal
  * @mode_fixup: fix mode data comparing to hw specific display mode.
  * @mode_set: convert drm_display_mode to hw specific display mode and
@@ -168,7 +152,6 @@ struct exynos_drm_display {
 	struct drm_encoder *encoder;
 	struct drm_connector *connector;
 	struct exynos_drm_display_ops *ops;
-	void *ctx;
 };
 
 /*
@@ -227,7 +210,6 @@ struct exynos_drm_manager {
 	struct drm_crtc *crtc;
 	int pipe;
 	struct exynos_drm_manager_ops *ops;
-	void *ctx;
 };
 
 struct exynos_drm_g2d_private {
@@ -240,7 +222,6 @@ struct exynos_drm_g2d_private {
 struct drm_exynos_file_private {
 	struct exynos_drm_g2d_private	*g2d_priv;
 	struct device			*ipp_dev;
-	struct file			*anon_filp;
 };
 
 /*
@@ -280,8 +261,6 @@ struct exynos_drm_private {
  * @dev: pointer to device object for subdrv device driver.
  * @drm_dev: pointer to drm_device and this pointer would be set
  *	when sub driver calls exynos_drm_subdrv_register().
- * @manager: subdrv has its own manager to control a hardware appropriately
- *     and we can access a hardware drawing on this manager.
  * @probe: this callback would be called by exynos drm driver after
  *     subdrv is registered to it.
  * @remove: this callback is used to release resources created
@@ -313,45 +292,34 @@ int exynos_drm_device_subdrv_remove(struct drm_device *dev);
 int exynos_drm_subdrv_open(struct drm_device *dev, struct drm_file *file);
 void exynos_drm_subdrv_close(struct drm_device *dev, struct drm_file *file);
 
-/*
- * this function registers exynos drm hdmi platform device. It ensures only one
- * instance of the device is created.
- */
-int exynos_platform_device_hdmi_register(void);
-
-/*
- * this function unregisters exynos drm hdmi platform device if it exists.
- */
-void exynos_platform_device_hdmi_unregister(void);
-
-/*
- * this function registers exynos drm ipp platform device.
- */
+#ifdef CONFIG_DRM_EXYNOS_IPP
 int exynos_platform_device_ipp_register(void);
-
-/*
- * this function unregisters exynos drm ipp platform device if it exists.
- */
 void exynos_platform_device_ipp_unregister(void);
+#else
+static inline int exynos_platform_device_ipp_register(void) { return 0; }
+static inline void exynos_platform_device_ipp_unregister(void) {}
+#endif
+
 
 #ifdef CONFIG_DRM_EXYNOS_DPI
 struct exynos_drm_display * exynos_dpi_probe(struct device *dev);
-int exynos_dpi_remove(struct device *dev);
+int exynos_dpi_remove(struct exynos_drm_display *display);
 #else
 static inline struct exynos_drm_display *
 exynos_dpi_probe(struct device *dev) { return NULL; }
-static inline int exynos_dpi_remove(struct device *dev) { return 0; }
+static inline int exynos_dpi_remove(struct exynos_drm_display *display)
+{
+	return 0;
+}
 #endif
 
-/*
- * this function registers exynos drm vidi platform device/driver.
- */
+#ifdef CONFIG_DRM_EXYNOS_VIDI
 int exynos_drm_probe_vidi(void);
-
-/*
- * this function unregister exynos drm vidi platform device/driver.
- */
 void exynos_drm_remove_vidi(void);
+#else
+static inline int exynos_drm_probe_vidi(void) { return 0; }
+static inline void exynos_drm_remove_vidi(void) {}
+#endif
 
 /* This function creates a encoder and a connector, and initializes them. */
 int exynos_drm_create_enc_conn(struct drm_device *dev,
diff --git a/drivers/gpu/drm/exynos/exynos_drm_dsi.c b/drivers/gpu/drm/exynos/exynos_drm_dsi.c
index 442aa2d00132..05fe93dc57a8 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_dsi.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_dsi.c
@@ -114,6 +114,8 @@
 #define DSIM_SYNC_INFORM		(1 << 27)
 #define DSIM_EOT_DISABLE		(1 << 28)
 #define DSIM_MFLUSH_VS			(1 << 29)
+/* This flag is valid only for exynos3250/3472/4415/5260/5430 */
+#define DSIM_CLKLANE_STOP		(1 << 30)
 
 /* DSIM_ESCMODE */
 #define DSIM_TX_TRIGGER_RST		(1 << 4)
@@ -262,12 +264,13 @@ struct exynos_dsi_driver_data {
 	unsigned int plltmr_reg;
 
 	unsigned int has_freqband:1;
+	unsigned int has_clklane_stop:1;
 };
 
 struct exynos_dsi {
+	struct exynos_drm_display display;
 	struct mipi_dsi_host dsi_host;
 	struct drm_connector connector;
-	struct drm_encoder *encoder;
 	struct device_node *panel_node;
 	struct drm_panel *panel;
 	struct device *dev;
@@ -301,9 +304,26 @@ struct exynos_dsi {
 #define host_to_dsi(host) container_of(host, struct exynos_dsi, dsi_host)
 #define connector_to_dsi(c) container_of(c, struct exynos_dsi, connector)
 
+static inline struct exynos_dsi *display_to_dsi(struct exynos_drm_display *d)
+{
+	return container_of(d, struct exynos_dsi, display);
+}
+
+static struct exynos_dsi_driver_data exynos3_dsi_driver_data = {
+	.plltmr_reg = 0x50,
+	.has_freqband = 1,
+	.has_clklane_stop = 1,
+};
+
 static struct exynos_dsi_driver_data exynos4_dsi_driver_data = {
 	.plltmr_reg = 0x50,
 	.has_freqband = 1,
+	.has_clklane_stop = 1,
+};
+
+static struct exynos_dsi_driver_data exynos4415_dsi_driver_data = {
+	.plltmr_reg = 0x58,
+	.has_clklane_stop = 1,
 };
 
 static struct exynos_dsi_driver_data exynos5_dsi_driver_data = {
@@ -311,8 +331,12 @@ static struct exynos_dsi_driver_data exynos5_dsi_driver_data = {
 };
 
 static struct of_device_id exynos_dsi_of_match[] = {
+	{ .compatible = "samsung,exynos3250-mipi-dsi",
+	  .data = &exynos3_dsi_driver_data },
 	{ .compatible = "samsung,exynos4210-mipi-dsi",
 	  .data = &exynos4_dsi_driver_data },
+	{ .compatible = "samsung,exynos4415-mipi-dsi",
+	  .data = &exynos4415_dsi_driver_data },
 	{ .compatible = "samsung,exynos5410-mipi-dsi",
 	  .data = &exynos5_dsi_driver_data },
 	{ }
@@ -421,7 +445,7 @@ static unsigned long exynos_dsi_set_pll(struct exynos_dsi *dsi,
 	if (!fout) {
 		dev_err(dsi->dev,
 			"failed to find PLL PMS for requested frequency\n");
-		return -EFAULT;
+		return 0;
 	}
 	dev_dbg(dsi->dev, "PLL freq %lu, (p %d, m %d, s %d)\n", fout, p, m, s);
 
@@ -453,7 +477,7 @@ static unsigned long exynos_dsi_set_pll(struct exynos_dsi *dsi,
 	do {
 		if (timeout-- == 0) {
 			dev_err(dsi->dev, "PLL failed to stabilize\n");
-			return -EFAULT;
+			return 0;
 		}
 		reg = readl(dsi->reg_base + DSIM_STATUS_REG);
 	} while ((reg & DSIM_PLL_STABLE) == 0);
@@ -569,6 +593,7 @@ static void exynos_dsi_disable_clock(struct exynos_dsi *dsi)
 
 static int exynos_dsi_init_link(struct exynos_dsi *dsi)
 {
+	struct exynos_dsi_driver_data *driver_data = dsi->driver_data;
 	int timeout;
 	u32 reg;
 	u32 lanes_mask;
@@ -650,6 +675,20 @@ static int exynos_dsi_init_link(struct exynos_dsi *dsi)
 	reg |= DSIM_LANE_EN(lanes_mask);
 	writel(reg, dsi->reg_base + DSIM_CONFIG_REG);
 
+	/*
+	 * Use non-continuous clock mode if the periparal wants and
+	 * host controller supports
+	 *
+	 * In non-continous clock mode, host controller will turn off
+	 * the HS clock between high-speed transmissions to reduce
+	 * power consumption.
+	 */
+	if (driver_data->has_clklane_stop &&
+			dsi->mode_flags & MIPI_DSI_CLOCK_NON_CONTINUOUS) {
+		reg |= DSIM_CLKLANE_STOP;
+		writel(reg, dsi->reg_base + DSIM_CONFIG_REG);
+	}
+
 	/* Check clock and data lane state are stop state */
 	timeout = 100;
 	do {
@@ -1077,7 +1116,7 @@ static irqreturn_t exynos_dsi_irq(int irq, void *dev_id)
 static irqreturn_t exynos_dsi_te_irq_handler(int irq, void *dev_id)
 {
 	struct exynos_dsi *dsi = (struct exynos_dsi *)dev_id;
-	struct drm_encoder *encoder = dsi->encoder;
+	struct drm_encoder *encoder = dsi->display.encoder;
 
 	if (dsi->state & DSIM_STATE_ENABLED)
 		exynos_drm_crtc_te_handler(encoder->crtc);
@@ -1116,6 +1155,7 @@ static int exynos_dsi_init(struct exynos_dsi *dsi)
 static int exynos_dsi_register_te_irq(struct exynos_dsi *dsi)
 {
 	int ret;
+	int te_gpio_irq;
 
 	dsi->te_gpio = of_get_named_gpio(dsi->panel_node, "te-gpios", 0);
 	if (!gpio_is_valid(dsi->te_gpio)) {
@@ -1130,14 +1170,10 @@ static int exynos_dsi_register_te_irq(struct exynos_dsi *dsi)
 		goto out;
 	}
 
-	/*
-	 * This TE GPIO IRQ should not be set to IRQ_NOAUTOEN, because panel
-	 * calls drm_panel_init() first then calls mipi_dsi_attach() in probe().
-	 * It means that te_gpio is invalid when exynos_dsi_enable_irq() is
-	 * called by drm_panel_init() before panel is attached.
-	 */
-	ret = request_threaded_irq(gpio_to_irq(dsi->te_gpio),
-					exynos_dsi_te_irq_handler, NULL,
+	te_gpio_irq = gpio_to_irq(dsi->te_gpio);
+
+	irq_set_status_flags(te_gpio_irq, IRQ_NOAUTOEN);
+	ret = request_threaded_irq(te_gpio_irq, exynos_dsi_te_irq_handler, NULL,
 					IRQF_TRIGGER_RISING, "TE", dsi);
 	if (ret) {
 		dev_err(dsi->dev, "request interrupt failed with %d\n", ret);
@@ -1168,9 +1204,6 @@ static int exynos_dsi_host_attach(struct mipi_dsi_host *host,
 	dsi->mode_flags = device->mode_flags;
 	dsi->panel_node = device->dev.of_node;
 
-	if (dsi->connector.dev)
-		drm_helper_hpd_irq_event(dsi->connector.dev);
-
 	/*
 	 * This is a temporary solution and should be made by more generic way.
 	 *
@@ -1184,6 +1217,9 @@ static int exynos_dsi_host_attach(struct mipi_dsi_host *host,
 			return ret;
 	}
 
+	if (dsi->connector.dev)
+		drm_helper_hpd_irq_event(dsi->connector.dev);
+
 	return 0;
 }
 
@@ -1209,7 +1245,7 @@ static bool exynos_dsi_is_short_dsi_type(u8 type)
 }
 
 static ssize_t exynos_dsi_host_transfer(struct mipi_dsi_host *host,
-				       struct mipi_dsi_msg *msg)
+				        const struct mipi_dsi_msg *msg)
 {
 	struct exynos_dsi *dsi = host_to_dsi(host);
 	struct exynos_dsi_transfer xfer;
@@ -1342,16 +1378,17 @@ static int exynos_dsi_enable(struct exynos_dsi *dsi)
 	exynos_dsi_set_display_mode(dsi);
 	exynos_dsi_set_display_enable(dsi, true);
 
+	dsi->state |= DSIM_STATE_ENABLED;
+
 	ret = drm_panel_enable(dsi->panel);
 	if (ret < 0) {
+		dsi->state &= ~DSIM_STATE_ENABLED;
 		exynos_dsi_set_display_enable(dsi, false);
 		drm_panel_unprepare(dsi->panel);
 		exynos_dsi_poweroff(dsi);
 		return ret;
 	}
 
-	dsi->state |= DSIM_STATE_ENABLED;
-
 	return 0;
 }
 
@@ -1370,7 +1407,7 @@ static void exynos_dsi_disable(struct exynos_dsi *dsi)
 
 static void exynos_dsi_dpms(struct exynos_drm_display *display, int mode)
 {
-	struct exynos_dsi *dsi = display->ctx;
+	struct exynos_dsi *dsi = display_to_dsi(display);
 
 	if (dsi->panel) {
 		switch (mode) {
@@ -1414,6 +1451,9 @@ exynos_dsi_detect(struct drm_connector *connector, bool force)
 
 static void exynos_dsi_connector_destroy(struct drm_connector *connector)
 {
+	drm_connector_unregister(connector);
+	drm_connector_cleanup(connector);
+	connector->dev = NULL;
 }
 
 static struct drm_connector_funcs exynos_dsi_connector_funcs = {
@@ -1444,7 +1484,7 @@ exynos_dsi_best_encoder(struct drm_connector *connector)
 {
 	struct exynos_dsi *dsi = connector_to_dsi(connector);
 
-	return dsi->encoder;
+	return dsi->display.encoder;
 }
 
 static struct drm_connector_helper_funcs exynos_dsi_connector_helper_funcs = {
@@ -1456,12 +1496,10 @@ static struct drm_connector_helper_funcs exynos_dsi_connector_helper_funcs = {
 static int exynos_dsi_create_connector(struct exynos_drm_display *display,
 				       struct drm_encoder *encoder)
 {
-	struct exynos_dsi *dsi = display->ctx;
+	struct exynos_dsi *dsi = display_to_dsi(display);
 	struct drm_connector *connector = &dsi->connector;
 	int ret;
 
-	dsi->encoder = encoder;
-
 	connector->polled = DRM_CONNECTOR_POLL_HPD;
 
 	ret = drm_connector_init(encoder->dev, connector,
@@ -1482,7 +1520,7 @@ static int exynos_dsi_create_connector(struct exynos_drm_display *display,
 static void exynos_dsi_mode_set(struct exynos_drm_display *display,
 			 struct drm_display_mode *mode)
 {
-	struct exynos_dsi *dsi = display->ctx;
+	struct exynos_dsi *dsi = display_to_dsi(display);
 	struct videomode *vm = &dsi->vm;
 
 	vm->hactive = mode->hdisplay;
@@ -1501,10 +1539,6 @@ static struct exynos_drm_display_ops exynos_dsi_display_ops = {
 	.dpms = exynos_dsi_dpms
 };
 
-static struct exynos_drm_display exynos_dsi_display = {
-	.type = EXYNOS_DISPLAY_TYPE_LCD,
-	.ops = &exynos_dsi_display_ops,
-};
 MODULE_DEVICE_TABLE(of, exynos_dsi_of_match);
 
 /* of_* functions will be removed after merge of of_graph patches */
@@ -1610,34 +1644,30 @@ end:
 static int exynos_dsi_bind(struct device *dev, struct device *master,
 				void *data)
 {
+	struct exynos_drm_display *display = dev_get_drvdata(dev);
+	struct exynos_dsi *dsi = display_to_dsi(display);
 	struct drm_device *drm_dev = data;
-	struct exynos_dsi *dsi;
 	int ret;
 
-	ret = exynos_drm_create_enc_conn(drm_dev, &exynos_dsi_display);
+	ret = exynos_drm_create_enc_conn(drm_dev, display);
 	if (ret) {
 		DRM_ERROR("Encoder create [%d] failed with %d\n",
-				exynos_dsi_display.type, ret);
+			  display->type, ret);
 		return ret;
 	}
 
-	dsi = exynos_dsi_display.ctx;
-
 	return mipi_dsi_host_register(&dsi->dsi_host);
 }
 
 static void exynos_dsi_unbind(struct device *dev, struct device *master,
 				void *data)
 {
-	struct exynos_dsi *dsi = exynos_dsi_display.ctx;
-	struct drm_encoder *encoder = dsi->encoder;
+	struct exynos_drm_display *display = dev_get_drvdata(dev);
+	struct exynos_dsi *dsi = display_to_dsi(display);
 
-	exynos_dsi_dpms(&exynos_dsi_display, DRM_MODE_DPMS_OFF);
+	exynos_dsi_dpms(display, DRM_MODE_DPMS_OFF);
 
 	mipi_dsi_host_unregister(&dsi->dsi_host);
-
-	encoder->funcs->destroy(encoder);
-	drm_connector_cleanup(&dsi->connector);
 }
 
 static const struct component_ops exynos_dsi_component_ops = {
@@ -1647,22 +1677,23 @@ static const struct component_ops exynos_dsi_component_ops = {
 
 static int exynos_dsi_probe(struct platform_device *pdev)
 {
+	struct device *dev = &pdev->dev;
 	struct resource *res;
 	struct exynos_dsi *dsi;
 	int ret;
 
-	ret = exynos_drm_component_add(&pdev->dev, EXYNOS_DEVICE_TYPE_CONNECTOR,
-					exynos_dsi_display.type);
+	dsi = devm_kzalloc(dev, sizeof(*dsi), GFP_KERNEL);
+	if (!dsi)
+		return -ENOMEM;
+
+	dsi->display.type = EXYNOS_DISPLAY_TYPE_LCD;
+	dsi->display.ops = &exynos_dsi_display_ops;
+
+	ret = exynos_drm_component_add(dev, EXYNOS_DEVICE_TYPE_CONNECTOR,
+				       dsi->display.type);
 	if (ret)
 		return ret;
 
-	dsi = devm_kzalloc(&pdev->dev, sizeof(*dsi), GFP_KERNEL);
-	if (!dsi) {
-		dev_err(&pdev->dev, "failed to allocate dsi object.\n");
-		ret = -ENOMEM;
-		goto err_del_component;
-	}
-
 	/* To be checked as invalid one */
 	dsi->te_gpio = -ENOENT;
 
@@ -1671,9 +1702,9 @@ static int exynos_dsi_probe(struct platform_device *pdev)
 	INIT_LIST_HEAD(&dsi->transfer_list);
 
 	dsi->dsi_host.ops = &exynos_dsi_ops;
-	dsi->dsi_host.dev = &pdev->dev;
+	dsi->dsi_host.dev = dev;
 
-	dsi->dev = &pdev->dev;
+	dsi->dev = dev;
 	dsi->driver_data = exynos_dsi_get_driver_data(pdev);
 
 	ret = exynos_dsi_parse_dt(dsi);
@@ -1682,70 +1713,68 @@ static int exynos_dsi_probe(struct platform_device *pdev)
 
 	dsi->supplies[0].supply = "vddcore";
 	dsi->supplies[1].supply = "vddio";
-	ret = devm_regulator_bulk_get(&pdev->dev, ARRAY_SIZE(dsi->supplies),
+	ret = devm_regulator_bulk_get(dev, ARRAY_SIZE(dsi->supplies),
 				      dsi->supplies);
 	if (ret) {
-		dev_info(&pdev->dev, "failed to get regulators: %d\n", ret);
+		dev_info(dev, "failed to get regulators: %d\n", ret);
 		return -EPROBE_DEFER;
 	}
 
-	dsi->pll_clk = devm_clk_get(&pdev->dev, "pll_clk");
+	dsi->pll_clk = devm_clk_get(dev, "pll_clk");
 	if (IS_ERR(dsi->pll_clk)) {
-		dev_info(&pdev->dev, "failed to get dsi pll input clock\n");
+		dev_info(dev, "failed to get dsi pll input clock\n");
 		ret = PTR_ERR(dsi->pll_clk);
 		goto err_del_component;
 	}
 
-	dsi->bus_clk = devm_clk_get(&pdev->dev, "bus_clk");
+	dsi->bus_clk = devm_clk_get(dev, "bus_clk");
 	if (IS_ERR(dsi->bus_clk)) {
-		dev_info(&pdev->dev, "failed to get dsi bus clock\n");
+		dev_info(dev, "failed to get dsi bus clock\n");
 		ret = PTR_ERR(dsi->bus_clk);
 		goto err_del_component;
 	}
 
 	res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
-	dsi->reg_base = devm_ioremap_resource(&pdev->dev, res);
+	dsi->reg_base = devm_ioremap_resource(dev, res);
 	if (IS_ERR(dsi->reg_base)) {
-		dev_err(&pdev->dev, "failed to remap io region\n");
+		dev_err(dev, "failed to remap io region\n");
 		ret = PTR_ERR(dsi->reg_base);
 		goto err_del_component;
 	}
 
-	dsi->phy = devm_phy_get(&pdev->dev, "dsim");
+	dsi->phy = devm_phy_get(dev, "dsim");
 	if (IS_ERR(dsi->phy)) {
-		dev_info(&pdev->dev, "failed to get dsim phy\n");
+		dev_info(dev, "failed to get dsim phy\n");
 		ret = PTR_ERR(dsi->phy);
 		goto err_del_component;
 	}
 
 	dsi->irq = platform_get_irq(pdev, 0);
 	if (dsi->irq < 0) {
-		dev_err(&pdev->dev, "failed to request dsi irq resource\n");
+		dev_err(dev, "failed to request dsi irq resource\n");
 		ret = dsi->irq;
 		goto err_del_component;
 	}
 
 	irq_set_status_flags(dsi->irq, IRQ_NOAUTOEN);
-	ret = devm_request_threaded_irq(&pdev->dev, dsi->irq, NULL,
+	ret = devm_request_threaded_irq(dev, dsi->irq, NULL,
 					exynos_dsi_irq, IRQF_ONESHOT,
-					dev_name(&pdev->dev), dsi);
+					dev_name(dev), dsi);
 	if (ret) {
-		dev_err(&pdev->dev, "failed to request dsi irq\n");
+		dev_err(dev, "failed to request dsi irq\n");
 		goto err_del_component;
 	}
 
-	exynos_dsi_display.ctx = dsi;
+	platform_set_drvdata(pdev, &dsi->display);
 
-	platform_set_drvdata(pdev, &exynos_dsi_display);
-
-	ret = component_add(&pdev->dev, &exynos_dsi_component_ops);
+	ret = component_add(dev, &exynos_dsi_component_ops);
 	if (ret)
 		goto err_del_component;
 
 	return ret;
 
 err_del_component:
-	exynos_drm_component_del(&pdev->dev, EXYNOS_DEVICE_TYPE_CONNECTOR);
+	exynos_drm_component_del(dev, EXYNOS_DEVICE_TYPE_CONNECTOR);
 	return ret;
 }
 
diff --git a/drivers/gpu/drm/exynos/exynos_drm_encoder.h b/drivers/gpu/drm/exynos/exynos_drm_encoder.h
index b7a1620a7e79..26305d8dd93a 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_encoder.h
+++ b/drivers/gpu/drm/exynos/exynos_drm_encoder.h
@@ -14,8 +14,6 @@
 #ifndef _EXYNOS_DRM_ENCODER_H_
 #define _EXYNOS_DRM_ENCODER_H_
 
-struct exynos_drm_manager;
-
 void exynos_drm_encoder_setup(struct drm_device *dev);
 struct drm_encoder *exynos_drm_encoder_create(struct drm_device *dev,
 			struct exynos_drm_display *mgr,
diff --git a/drivers/gpu/drm/exynos/exynos_drm_fb.c b/drivers/gpu/drm/exynos/exynos_drm_fb.c
index 65a22cad7b36..d346d1e6eda0 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_fb.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_fb.c
@@ -165,6 +165,7 @@ exynos_drm_framebuffer_init(struct drm_device *dev,
 
 	ret = drm_framebuffer_init(dev, &exynos_fb->fb, &exynos_drm_fb_funcs);
 	if (ret) {
+		kfree(exynos_fb);
 		DRM_ERROR("failed to initialize framebuffer\n");
 		return ERR_PTR(ret);
 	}
diff --git a/drivers/gpu/drm/exynos/exynos_drm_fbdev.c b/drivers/gpu/drm/exynos/exynos_drm_fbdev.c
index 32e63f60e1d1..e12ea90c6237 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_fbdev.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_fbdev.c
@@ -123,6 +123,7 @@ static int exynos_drm_fbdev_update(struct drm_fb_helper *helper,
 
 	fbi->screen_base = buffer->kvaddr + offset;
 	fbi->screen_size = size;
+	fbi->fix.smem_len = size;
 
 	return 0;
 }
@@ -353,9 +354,6 @@ void exynos_drm_fbdev_fini(struct drm_device *dev)
 
 	fbdev = to_exynos_fbdev(private->fb_helper);
 
-	if (fbdev->exynos_gem_obj)
-		exynos_drm_gem_destroy(fbdev->exynos_gem_obj);
-
 	exynos_drm_fbdev_destroy(dev, private->fb_helper);
 	kfree(fbdev);
 	private->fb_helper = NULL;
diff --git a/drivers/gpu/drm/exynos/exynos_drm_fimc.c b/drivers/gpu/drm/exynos/exynos_drm_fimc.c
index ec7cc9ea50df..835b6af00970 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_fimc.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_fimc.c
@@ -336,9 +336,6 @@ static bool fimc_check_ovf(struct fimc_context *ctx)
 		fimc_set_bits(ctx, EXYNOS_CIWDOFST,
 			EXYNOS_CIWDOFST_CLROVFIY | EXYNOS_CIWDOFST_CLROVFICB |
 			EXYNOS_CIWDOFST_CLROVFICR);
-		fimc_clear_bits(ctx, EXYNOS_CIWDOFST,
-			EXYNOS_CIWDOFST_CLROVFIY | EXYNOS_CIWDOFST_CLROVFICB |
-			EXYNOS_CIWDOFST_CLROVFICR);
 
 		dev_err(ippdrv->dev, "occurred overflow at %d, status 0x%x.\n",
 			ctx->id, status);
@@ -718,24 +715,24 @@ static int fimc_src_set_addr(struct device *dev,
 	case IPP_BUF_ENQUEUE:
 		config = &property->config[EXYNOS_DRM_OPS_SRC];
 		fimc_write(ctx, buf_info->base[EXYNOS_DRM_PLANAR_Y],
-			EXYNOS_CIIYSA(buf_id));
+			EXYNOS_CIIYSA0);
 
 		if (config->fmt == DRM_FORMAT_YVU420) {
 			fimc_write(ctx, buf_info->base[EXYNOS_DRM_PLANAR_CR],
-				EXYNOS_CIICBSA(buf_id));
+				EXYNOS_CIICBSA0);
 			fimc_write(ctx, buf_info->base[EXYNOS_DRM_PLANAR_CB],
-				EXYNOS_CIICRSA(buf_id));
+				EXYNOS_CIICRSA0);
 		} else {
 			fimc_write(ctx, buf_info->base[EXYNOS_DRM_PLANAR_CB],
-				EXYNOS_CIICBSA(buf_id));
+				EXYNOS_CIICBSA0);
 			fimc_write(ctx, buf_info->base[EXYNOS_DRM_PLANAR_CR],
-				EXYNOS_CIICRSA(buf_id));
+				EXYNOS_CIICRSA0);
 		}
 		break;
 	case IPP_BUF_DEQUEUE:
-		fimc_write(ctx, 0x0, EXYNOS_CIIYSA(buf_id));
-		fimc_write(ctx, 0x0, EXYNOS_CIICBSA(buf_id));
-		fimc_write(ctx, 0x0, EXYNOS_CIICRSA(buf_id));
+		fimc_write(ctx, 0x0, EXYNOS_CIIYSA0);
+		fimc_write(ctx, 0x0, EXYNOS_CIICBSA0);
+		fimc_write(ctx, 0x0, EXYNOS_CIICRSA0);
 		break;
 	default:
 		/* bypass */
@@ -1122,67 +1119,34 @@ static int fimc_dst_set_size(struct device *dev, int swap,
 	return 0;
 }
 
-static int fimc_dst_get_buf_count(struct fimc_context *ctx)
-{
-	u32 cfg, buf_num;
-
-	cfg = fimc_read(ctx, EXYNOS_CIFCNTSEQ);
-
-	buf_num = hweight32(cfg);
-
-	DRM_DEBUG_KMS("buf_num[%d]\n", buf_num);
-
-	return buf_num;
-}
-
-static int fimc_dst_set_buf_seq(struct fimc_context *ctx, u32 buf_id,
+static void fimc_dst_set_buf_seq(struct fimc_context *ctx, u32 buf_id,
 		enum drm_exynos_ipp_buf_type buf_type)
 {
-	struct exynos_drm_ippdrv *ippdrv = &ctx->ippdrv;
-	bool enable;
-	u32 cfg;
-	u32 mask = 0x00000001 << buf_id;
-	int ret = 0;
 	unsigned long flags;
+	u32 buf_num;
+	u32 cfg;
 
 	DRM_DEBUG_KMS("buf_id[%d]buf_type[%d]\n", buf_id, buf_type);
 
 	spin_lock_irqsave(&ctx->lock, flags);
 
-	/* mask register set */
 	cfg = fimc_read(ctx, EXYNOS_CIFCNTSEQ);
 
-	switch (buf_type) {
-	case IPP_BUF_ENQUEUE:
-		enable = true;
-		break;
-	case IPP_BUF_DEQUEUE:
-		enable = false;
-		break;
-	default:
-		dev_err(ippdrv->dev, "invalid buf ctrl parameter.\n");
-		ret =  -EINVAL;
-		goto err_unlock;
-	}
+	if (buf_type == IPP_BUF_ENQUEUE)
+		cfg |= (1 << buf_id);
+	else
+		cfg &= ~(1 << buf_id);
 
-	/* sequence id */
-	cfg &= ~mask;
-	cfg |= (enable << buf_id);
 	fimc_write(ctx, cfg, EXYNOS_CIFCNTSEQ);
 
-	/* interrupt enable */
-	if (buf_type == IPP_BUF_ENQUEUE &&
-	    fimc_dst_get_buf_count(ctx) >= FIMC_BUF_START)
-		fimc_mask_irq(ctx, true);
+	buf_num = hweight32(cfg);
 
-	/* interrupt disable */
-	if (buf_type == IPP_BUF_DEQUEUE &&
-	    fimc_dst_get_buf_count(ctx) <= FIMC_BUF_STOP)
+	if (buf_type == IPP_BUF_ENQUEUE && buf_num >= FIMC_BUF_START)
+		fimc_mask_irq(ctx, true);
+	else if (buf_type == IPP_BUF_DEQUEUE && buf_num <= FIMC_BUF_STOP)
 		fimc_mask_irq(ctx, false);
 
-err_unlock:
 	spin_unlock_irqrestore(&ctx->lock, flags);
-	return ret;
 }
 
 static int fimc_dst_set_addr(struct device *dev,
@@ -1240,7 +1204,9 @@ static int fimc_dst_set_addr(struct device *dev,
 		break;
 	}
 
-	return fimc_dst_set_buf_seq(ctx, buf_id, buf_type);
+	fimc_dst_set_buf_seq(ctx, buf_id, buf_type);
+
+	return 0;
 }
 
 static struct exynos_drm_ipp_ops fimc_dst_ops = {
@@ -1291,14 +1257,11 @@ static irqreturn_t fimc_irq_handler(int irq, void *dev_id)
 
 	DRM_DEBUG_KMS("buf_id[%d]\n", buf_id);
 
-	if (fimc_dst_set_buf_seq(ctx, buf_id, IPP_BUF_DEQUEUE) < 0) {
-		DRM_ERROR("failed to dequeue.\n");
-		return IRQ_HANDLED;
-	}
+	fimc_dst_set_buf_seq(ctx, buf_id, IPP_BUF_DEQUEUE);
 
 	event_work->ippdrv = ippdrv;
 	event_work->buf_id[EXYNOS_DRM_OPS_DST] = buf_id;
-	queue_work(ippdrv->event_workq, (struct work_struct *)event_work);
+	queue_work(ippdrv->event_workq, &event_work->work);
 
 	return IRQ_HANDLED;
 }
@@ -1590,11 +1553,8 @@ static int fimc_ippdrv_start(struct device *dev, enum drm_exynos_ipp_cmd cmd)
 
 	fimc_clear_bits(ctx, EXYNOS_CIOCTRL, EXYNOS_CIOCTRL_WEAVE_MASK);
 
-	if (cmd == IPP_CMD_M2M) {
-		fimc_set_bits(ctx, EXYNOS_MSCTRL, EXYNOS_MSCTRL_ENVID);
-
+	if (cmd == IPP_CMD_M2M)
 		fimc_set_bits(ctx, EXYNOS_MSCTRL, EXYNOS_MSCTRL_ENVID);
-	}
 
 	return 0;
 }
@@ -1857,7 +1817,7 @@ static int fimc_resume(struct device *dev)
 }
 #endif
 
-#ifdef CONFIG_PM_RUNTIME
+#ifdef CONFIG_PM
 static int fimc_runtime_suspend(struct device *dev)
 {
 	struct fimc_context *ctx = get_fimc_context(dev);
diff --git a/drivers/gpu/drm/exynos/exynos_drm_fimd.c b/drivers/gpu/drm/exynos/exynos_drm_fimd.c
index 5d09e33fef87..e5810d13bf9c 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_fimd.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_fimd.c
@@ -84,8 +84,6 @@
 /* FIMD has totally five hardware windows. */
 #define WINDOWS_NR	5
 
-#define get_fimd_manager(mgr)	platform_get_drvdata(to_platform_device(dev))
-
 struct fimd_driver_data {
 	unsigned int timing_base;
 	unsigned int lcdblk_offset;
@@ -96,6 +94,7 @@ struct fimd_driver_data {
 	unsigned int has_clksel:1;
 	unsigned int has_limited_fmt:1;
 	unsigned int has_vidoutcon:1;
+	unsigned int has_vtsel:1;
 };
 
 static struct fimd_driver_data s3c64xx_fimd_driver_data = {
@@ -104,12 +103,31 @@ static struct fimd_driver_data s3c64xx_fimd_driver_data = {
 	.has_limited_fmt = 1,
 };
 
+static struct fimd_driver_data exynos3_fimd_driver_data = {
+	.timing_base = 0x20000,
+	.lcdblk_offset = 0x210,
+	.lcdblk_bypass_shift = 1,
+	.has_shadowcon = 1,
+	.has_vidoutcon = 1,
+};
+
 static struct fimd_driver_data exynos4_fimd_driver_data = {
 	.timing_base = 0x0,
 	.lcdblk_offset = 0x210,
 	.lcdblk_vt_shift = 10,
 	.lcdblk_bypass_shift = 1,
 	.has_shadowcon = 1,
+	.has_vtsel = 1,
+};
+
+static struct fimd_driver_data exynos4415_fimd_driver_data = {
+	.timing_base = 0x20000,
+	.lcdblk_offset = 0x210,
+	.lcdblk_vt_shift = 10,
+	.lcdblk_bypass_shift = 1,
+	.has_shadowcon = 1,
+	.has_vidoutcon = 1,
+	.has_vtsel = 1,
 };
 
 static struct fimd_driver_data exynos5_fimd_driver_data = {
@@ -119,6 +137,7 @@ static struct fimd_driver_data exynos5_fimd_driver_data = {
 	.lcdblk_bypass_shift = 15,
 	.has_shadowcon = 1,
 	.has_vidoutcon = 1,
+	.has_vtsel = 1,
 };
 
 struct fimd_win_data {
@@ -138,6 +157,7 @@ struct fimd_win_data {
 };
 
 struct fimd_context {
+	struct exynos_drm_manager	manager;
 	struct device			*dev;
 	struct drm_device		*drm_dev;
 	struct clk			*bus_clk;
@@ -165,11 +185,20 @@ struct fimd_context {
 	struct exynos_drm_display *display;
 };
 
+static inline struct fimd_context *mgr_to_fimd(struct exynos_drm_manager *mgr)
+{
+	return container_of(mgr, struct fimd_context, manager);
+}
+
 static const struct of_device_id fimd_driver_dt_match[] = {
 	{ .compatible = "samsung,s3c6400-fimd",
 	  .data = &s3c64xx_fimd_driver_data },
+	{ .compatible = "samsung,exynos3250-fimd",
+	  .data = &exynos3_fimd_driver_data },
 	{ .compatible = "samsung,exynos4210-fimd",
 	  .data = &exynos4_fimd_driver_data },
+	{ .compatible = "samsung,exynos4415-fimd",
+	  .data = &exynos4415_fimd_driver_data },
 	{ .compatible = "samsung,exynos5250-fimd",
 	  .data = &exynos5_fimd_driver_data },
 	{},
@@ -187,7 +216,7 @@ static inline struct fimd_driver_data *drm_fimd_get_driver_data(
 
 static void fimd_wait_for_vblank(struct exynos_drm_manager *mgr)
 {
-	struct fimd_context *ctx = mgr->ctx;
+	struct fimd_context *ctx = mgr_to_fimd(mgr);
 
 	if (ctx->suspended)
 		return;
@@ -204,56 +233,74 @@ static void fimd_wait_for_vblank(struct exynos_drm_manager *mgr)
 		DRM_DEBUG_KMS("vblank wait timed out.\n");
 }
 
+static void fimd_enable_video_output(struct fimd_context *ctx, int win,
+					bool enable)
+{
+	u32 val = readl(ctx->regs + WINCON(win));
+
+	if (enable)
+		val |= WINCONx_ENWIN;
+	else
+		val &= ~WINCONx_ENWIN;
+
+	writel(val, ctx->regs + WINCON(win));
+}
+
+static void fimd_enable_shadow_channel_path(struct fimd_context *ctx, int win,
+						bool enable)
+{
+	u32 val = readl(ctx->regs + SHADOWCON);
+
+	if (enable)
+		val |= SHADOWCON_CHx_ENABLE(win);
+	else
+		val &= ~SHADOWCON_CHx_ENABLE(win);
+
+	writel(val, ctx->regs + SHADOWCON);
+}
 
 static void fimd_clear_channel(struct exynos_drm_manager *mgr)
 {
-	struct fimd_context *ctx = mgr->ctx;
+	struct fimd_context *ctx = mgr_to_fimd(mgr);
 	int win, ch_enabled = 0;
 
 	DRM_DEBUG_KMS("%s\n", __FILE__);
 
 	/* Check if any channel is enabled. */
 	for (win = 0; win < WINDOWS_NR; win++) {
-		u32 val = readl(ctx->regs + SHADOWCON);
-		if (val & SHADOWCON_CHx_ENABLE(win)) {
-			val &= ~SHADOWCON_CHx_ENABLE(win);
-			writel(val, ctx->regs + SHADOWCON);
+		u32 val = readl(ctx->regs + WINCON(win));
+
+		if (val & WINCONx_ENWIN) {
+			fimd_enable_video_output(ctx, win, false);
+
+			if (ctx->driver_data->has_shadowcon)
+				fimd_enable_shadow_channel_path(ctx, win,
+								false);
+
 			ch_enabled = 1;
 		}
 	}
 
 	/* Wait for vsync, as disable channel takes effect at next vsync */
-	if (ch_enabled)
+	if (ch_enabled) {
+		unsigned int state = ctx->suspended;
+
+		ctx->suspended = 0;
 		fimd_wait_for_vblank(mgr);
+		ctx->suspended = state;
+	}
 }
 
 static int fimd_mgr_initialize(struct exynos_drm_manager *mgr,
 			struct drm_device *drm_dev)
 {
-	struct fimd_context *ctx = mgr->ctx;
+	struct fimd_context *ctx = mgr_to_fimd(mgr);
 	struct exynos_drm_private *priv;
 	priv = drm_dev->dev_private;
 
 	mgr->drm_dev = ctx->drm_dev = drm_dev;
 	mgr->pipe = ctx->pipe = priv->pipe++;
 
-	/*
-	 * enable drm irq mode.
-	 * - with irq_enabled = true, we can use the vblank feature.
-	 *
-	 * P.S. note that we wouldn't use drm irq handler but
-	 *	just specific driver own one instead because
-	 *	drm framework supports only one irq handler.
-	 */
-	drm_dev->irq_enabled = true;
-
-	/*
-	 * with vblank_disable_allowed = true, vblank interrupt will be disabled
-	 * by drm timer once a current process gives up ownership of
-	 * vblank event.(after drm_vblank_put function is called)
-	 */
-	drm_dev->vblank_disable_allowed = true;
-
 	/* attach this sub driver to iommu mapping if supported. */
 	if (is_drm_iommu_supported(ctx->drm_dev)) {
 		/*
@@ -269,7 +316,7 @@ static int fimd_mgr_initialize(struct exynos_drm_manager *mgr,
 
 static void fimd_mgr_remove(struct exynos_drm_manager *mgr)
 {
-	struct fimd_context *ctx = mgr->ctx;
+	struct fimd_context *ctx = mgr_to_fimd(mgr);
 
 	/* detach this sub driver from iommu mapping if supported. */
 	if (is_drm_iommu_supported(ctx->drm_dev))
@@ -309,14 +356,14 @@ static bool fimd_mode_fixup(struct exynos_drm_manager *mgr,
 static void fimd_mode_set(struct exynos_drm_manager *mgr,
 		const struct drm_display_mode *in_mode)
 {
-	struct fimd_context *ctx = mgr->ctx;
+	struct fimd_context *ctx = mgr_to_fimd(mgr);
 
 	drm_mode_copy(&ctx->mode, in_mode);
 }
 
 static void fimd_commit(struct exynos_drm_manager *mgr)
 {
-	struct fimd_context *ctx = mgr->ctx;
+	struct fimd_context *ctx = mgr_to_fimd(mgr);
 	struct drm_display_mode *mode = &ctx->mode;
 	struct fimd_driver_data *driver_data = ctx->driver_data;
 	void *timing_base = ctx->regs + driver_data->timing_base;
@@ -337,7 +384,8 @@ static void fimd_commit(struct exynos_drm_manager *mgr)
 		writel(0, timing_base + I80IFCONFBx(0));
 
 		/* set video type selection to I80 interface */
-		if (ctx->sysreg && regmap_update_bits(ctx->sysreg,
+		if (driver_data->has_vtsel && ctx->sysreg &&
+				regmap_update_bits(ctx->sysreg,
 					driver_data->lcdblk_offset,
 					0x3 << driver_data->lcdblk_vt_shift,
 					0x1 << driver_data->lcdblk_vt_shift)) {
@@ -415,7 +463,7 @@ static void fimd_commit(struct exynos_drm_manager *mgr)
 
 static int fimd_enable_vblank(struct exynos_drm_manager *mgr)
 {
-	struct fimd_context *ctx = mgr->ctx;
+	struct fimd_context *ctx = mgr_to_fimd(mgr);
 	u32 val;
 
 	if (ctx->suspended)
@@ -425,12 +473,19 @@ static int fimd_enable_vblank(struct exynos_drm_manager *mgr)
 		val = readl(ctx->regs + VIDINTCON0);
 
 		val |= VIDINTCON0_INT_ENABLE;
-		val |= VIDINTCON0_INT_FRAME;
 
-		val &= ~VIDINTCON0_FRAMESEL0_MASK;
-		val |= VIDINTCON0_FRAMESEL0_VSYNC;
-		val &= ~VIDINTCON0_FRAMESEL1_MASK;
-		val |= VIDINTCON0_FRAMESEL1_NONE;
+		if (ctx->i80_if) {
+			val |= VIDINTCON0_INT_I80IFDONE;
+			val |= VIDINTCON0_INT_SYSMAINCON;
+			val &= ~VIDINTCON0_INT_SYSSUBCON;
+		} else {
+			val |= VIDINTCON0_INT_FRAME;
+
+			val &= ~VIDINTCON0_FRAMESEL0_MASK;
+			val |= VIDINTCON0_FRAMESEL0_VSYNC;
+			val &= ~VIDINTCON0_FRAMESEL1_MASK;
+			val |= VIDINTCON0_FRAMESEL1_NONE;
+		}
 
 		writel(val, ctx->regs + VIDINTCON0);
 	}
@@ -440,7 +495,7 @@ static int fimd_enable_vblank(struct exynos_drm_manager *mgr)
 
 static void fimd_disable_vblank(struct exynos_drm_manager *mgr)
 {
-	struct fimd_context *ctx = mgr->ctx;
+	struct fimd_context *ctx = mgr_to_fimd(mgr);
 	u32 val;
 
 	if (ctx->suspended)
@@ -449,9 +504,15 @@ static void fimd_disable_vblank(struct exynos_drm_manager *mgr)
 	if (test_and_clear_bit(0, &ctx->irq_flags)) {
 		val = readl(ctx->regs + VIDINTCON0);
 
-		val &= ~VIDINTCON0_INT_FRAME;
 		val &= ~VIDINTCON0_INT_ENABLE;
 
+		if (ctx->i80_if) {
+			val &= ~VIDINTCON0_INT_I80IFDONE;
+			val &= ~VIDINTCON0_INT_SYSMAINCON;
+			val &= ~VIDINTCON0_INT_SYSSUBCON;
+		} else
+			val &= ~VIDINTCON0_INT_FRAME;
+
 		writel(val, ctx->regs + VIDINTCON0);
 	}
 }
@@ -459,7 +520,7 @@ static void fimd_disable_vblank(struct exynos_drm_manager *mgr)
 static void fimd_win_mode_set(struct exynos_drm_manager *mgr,
 			struct exynos_drm_overlay *overlay)
 {
-	struct fimd_context *ctx = mgr->ctx;
+	struct fimd_context *ctx = mgr_to_fimd(mgr);
 	struct fimd_win_data *win_data;
 	int win;
 	unsigned long offset;
@@ -617,7 +678,7 @@ static void fimd_shadow_protect_win(struct fimd_context *ctx,
 
 static void fimd_win_commit(struct exynos_drm_manager *mgr, int zpos)
 {
-	struct fimd_context *ctx = mgr->ctx;
+	struct fimd_context *ctx = mgr_to_fimd(mgr);
 	struct fimd_win_data *win_data;
 	int win = zpos;
 	unsigned long val, alpha, size;
@@ -724,20 +785,14 @@ static void fimd_win_commit(struct exynos_drm_manager *mgr, int zpos)
 	if (win != 0)
 		fimd_win_set_colkey(ctx, win);
 
-	/* wincon */
-	val = readl(ctx->regs + WINCON(win));
-	val |= WINCONx_ENWIN;
-	writel(val, ctx->regs + WINCON(win));
+	fimd_enable_video_output(ctx, win, true);
+
+	if (ctx->driver_data->has_shadowcon)
+		fimd_enable_shadow_channel_path(ctx, win, true);
 
 	/* Enable DMA channel and unprotect windows */
 	fimd_shadow_protect_win(ctx, win, false);
 
-	if (ctx->driver_data->has_shadowcon) {
-		val = readl(ctx->regs + SHADOWCON);
-		val |= SHADOWCON_CHx_ENABLE(win);
-		writel(val, ctx->regs + SHADOWCON);
-	}
-
 	win_data->enabled = true;
 
 	if (ctx->i80_if)
@@ -746,10 +801,9 @@ static void fimd_win_commit(struct exynos_drm_manager *mgr, int zpos)
 
 static void fimd_win_disable(struct exynos_drm_manager *mgr, int zpos)
 {
-	struct fimd_context *ctx = mgr->ctx;
+	struct fimd_context *ctx = mgr_to_fimd(mgr);
 	struct fimd_win_data *win_data;
 	int win = zpos;
-	u32 val;
 
 	if (win == DEFAULT_ZPOS)
 		win = ctx->default_win;
@@ -768,18 +822,12 @@ static void fimd_win_disable(struct exynos_drm_manager *mgr, int zpos)
 	/* protect windows */
 	fimd_shadow_protect_win(ctx, win, true);
 
-	/* wincon */
-	val = readl(ctx->regs + WINCON(win));
-	val &= ~WINCONx_ENWIN;
-	writel(val, ctx->regs + WINCON(win));
+	fimd_enable_video_output(ctx, win, false);
 
-	/* unprotect windows */
-	if (ctx->driver_data->has_shadowcon) {
-		val = readl(ctx->regs + SHADOWCON);
-		val &= ~SHADOWCON_CHx_ENABLE(win);
-		writel(val, ctx->regs + SHADOWCON);
-	}
+	if (ctx->driver_data->has_shadowcon)
+		fimd_enable_shadow_channel_path(ctx, win, false);
 
+	/* unprotect windows */
 	fimd_shadow_protect_win(ctx, win, false);
 
 	win_data->enabled = false;
@@ -787,7 +835,7 @@ static void fimd_win_disable(struct exynos_drm_manager *mgr, int zpos)
 
 static void fimd_window_suspend(struct exynos_drm_manager *mgr)
 {
-	struct fimd_context *ctx = mgr->ctx;
+	struct fimd_context *ctx = mgr_to_fimd(mgr);
 	struct fimd_win_data *win_data;
 	int i;
 
@@ -797,12 +845,11 @@ static void fimd_window_suspend(struct exynos_drm_manager *mgr)
 		if (win_data->enabled)
 			fimd_win_disable(mgr, i);
 	}
-	fimd_wait_for_vblank(mgr);
 }
 
 static void fimd_window_resume(struct exynos_drm_manager *mgr)
 {
-	struct fimd_context *ctx = mgr->ctx;
+	struct fimd_context *ctx = mgr_to_fimd(mgr);
 	struct fimd_win_data *win_data;
 	int i;
 
@@ -815,7 +862,7 @@ static void fimd_window_resume(struct exynos_drm_manager *mgr)
 
 static void fimd_apply(struct exynos_drm_manager *mgr)
 {
-	struct fimd_context *ctx = mgr->ctx;
+	struct fimd_context *ctx = mgr_to_fimd(mgr);
 	struct fimd_win_data *win_data;
 	int i;
 
@@ -832,7 +879,7 @@ static void fimd_apply(struct exynos_drm_manager *mgr)
 
 static int fimd_poweron(struct exynos_drm_manager *mgr)
 {
-	struct fimd_context *ctx = mgr->ctx;
+	struct fimd_context *ctx = mgr_to_fimd(mgr);
 	int ret;
 
 	if (!ctx->suspended)
@@ -880,7 +927,7 @@ bus_clk_err:
 
 static int fimd_poweroff(struct exynos_drm_manager *mgr)
 {
-	struct fimd_context *ctx = mgr->ctx;
+	struct fimd_context *ctx = mgr_to_fimd(mgr);
 
 	if (ctx->suspended)
 		return 0;
@@ -922,39 +969,41 @@ static void fimd_dpms(struct exynos_drm_manager *mgr, int mode)
 
 static void fimd_trigger(struct device *dev)
 {
-	struct exynos_drm_manager *mgr = get_fimd_manager(dev);
-	struct fimd_context *ctx = mgr->ctx;
+	struct fimd_context *ctx = dev_get_drvdata(dev);
 	struct fimd_driver_data *driver_data = ctx->driver_data;
 	void *timing_base = ctx->regs + driver_data->timing_base;
 	u32 reg;
 
-	atomic_set(&ctx->triggering, 1);
+	 /*
+	  * Skips triggering if in triggering state, because multiple triggering
+	  * requests can cause panel reset.
+	  */
+	if (atomic_read(&ctx->triggering))
+		return;
 
-	reg = readl(ctx->regs + VIDINTCON0);
-	reg |= (VIDINTCON0_INT_ENABLE | VIDINTCON0_INT_I80IFDONE |
-						VIDINTCON0_INT_SYSMAINCON);
-	writel(reg, ctx->regs + VIDINTCON0);
+	/* Enters triggering mode */
+	atomic_set(&ctx->triggering, 1);
 
 	reg = readl(timing_base + TRIGCON);
 	reg |= (TRGMODE_I80_RGB_ENABLE_I80 | SWTRGCMD_I80_RGB_ENABLE);
 	writel(reg, timing_base + TRIGCON);
+
+	/*
+	 * Exits triggering mode if vblank is not enabled yet, because when the
+	 * VIDINTCON0 register is not set, it can not exit from triggering mode.
+	 */
+	if (!test_bit(0, &ctx->irq_flags))
+		atomic_set(&ctx->triggering, 0);
 }
 
 static void fimd_te_handler(struct exynos_drm_manager *mgr)
 {
-	struct fimd_context *ctx = mgr->ctx;
+	struct fimd_context *ctx = mgr_to_fimd(mgr);
 
 	/* Checks the crtc is detached already from encoder */
 	if (ctx->pipe < 0 || !ctx->drm_dev)
 		return;
 
-	 /*
-	 * Skips to trigger if in triggering state, because multiple triggering
-	 * requests can cause panel reset.
-	 */
-	if (atomic_read(&ctx->triggering))
-		return;
-
 	/*
 	 * If there is a page flip request, triggers and handles the page flip
 	 * event so that current fb can be updated into panel GRAM.
@@ -966,10 +1015,10 @@ static void fimd_te_handler(struct exynos_drm_manager *mgr)
 	if (atomic_read(&ctx->wait_vsync_event)) {
 		atomic_set(&ctx->wait_vsync_event, 0);
 		wake_up(&ctx->wait_vsync_queue);
-
-		if (!atomic_read(&ctx->triggering))
-			drm_handle_vblank(ctx->drm_dev, ctx->pipe);
 	}
+
+	if (test_bit(0, &ctx->irq_flags))
+		drm_handle_vblank(ctx->drm_dev, ctx->pipe);
 }
 
 static struct exynos_drm_manager_ops fimd_manager_ops = {
@@ -986,11 +1035,6 @@ static struct exynos_drm_manager_ops fimd_manager_ops = {
 	.te_handler = fimd_te_handler,
 };
 
-static struct exynos_drm_manager fimd_manager = {
-	.type = EXYNOS_DISPLAY_TYPE_LCD,
-	.ops = &fimd_manager_ops,
-};
-
 static irqreturn_t fimd_irq_handler(int irq, void *dev_id)
 {
 	struct fimd_context *ctx = (struct fimd_context *)dev_id;
@@ -1007,16 +1051,10 @@ static irqreturn_t fimd_irq_handler(int irq, void *dev_id)
 		goto out;
 
 	if (ctx->i80_if) {
-		/* unset I80 frame done interrupt */
-		val = readl(ctx->regs + VIDINTCON0);
-		val &= ~(VIDINTCON0_INT_I80IFDONE | VIDINTCON0_INT_SYSMAINCON);
-		writel(val, ctx->regs + VIDINTCON0);
+		exynos_drm_crtc_finish_pageflip(ctx->drm_dev, ctx->pipe);
 
-		/* exit triggering mode */
+		/* Exits triggering mode */
 		atomic_set(&ctx->triggering, 0);
-
-		drm_handle_vblank(ctx->drm_dev, ctx->pipe);
-		exynos_drm_crtc_finish_pageflip(ctx->drm_dev, ctx->pipe);
 	} else {
 		drm_handle_vblank(ctx->drm_dev, ctx->pipe);
 		exynos_drm_crtc_finish_pageflip(ctx->drm_dev, ctx->pipe);
@@ -1034,11 +1072,11 @@ out:
 
 static int fimd_bind(struct device *dev, struct device *master, void *data)
 {
-	struct fimd_context *ctx = fimd_manager.ctx;
+	struct fimd_context *ctx = dev_get_drvdata(dev);
 	struct drm_device *drm_dev = data;
 
-	fimd_mgr_initialize(&fimd_manager, drm_dev);
-	exynos_drm_crtc_create(&fimd_manager);
+	fimd_mgr_initialize(&ctx->manager, drm_dev);
+	exynos_drm_crtc_create(&ctx->manager);
 	if (ctx->display)
 		exynos_drm_create_enc_conn(drm_dev, ctx->display);
 
@@ -1049,18 +1087,14 @@ static int fimd_bind(struct device *dev, struct device *master, void *data)
 static void fimd_unbind(struct device *dev, struct device *master,
 			void *data)
 {
-	struct exynos_drm_manager *mgr = dev_get_drvdata(dev);
-	struct fimd_context *ctx = fimd_manager.ctx;
-	struct drm_crtc *crtc = mgr->crtc;
+	struct fimd_context *ctx = dev_get_drvdata(dev);
 
-	fimd_dpms(mgr, DRM_MODE_DPMS_OFF);
+	fimd_dpms(&ctx->manager, DRM_MODE_DPMS_OFF);
 
 	if (ctx->display)
-		exynos_dpi_remove(dev);
-
-	fimd_mgr_remove(mgr);
+		exynos_dpi_remove(ctx->display);
 
-	crtc->funcs->destroy(crtc);
+	fimd_mgr_remove(&ctx->manager);
 }
 
 static const struct component_ops fimd_component_ops = {
@@ -1076,21 +1110,20 @@ static int fimd_probe(struct platform_device *pdev)
 	struct resource *res;
 	int ret = -EINVAL;
 
-	ret = exynos_drm_component_add(&pdev->dev, EXYNOS_DEVICE_TYPE_CRTC,
-					fimd_manager.type);
-	if (ret)
-		return ret;
-
-	if (!dev->of_node) {
-		ret = -ENODEV;
-		goto err_del_component;
-	}
+	if (!dev->of_node)
+		return -ENODEV;
 
 	ctx = devm_kzalloc(dev, sizeof(*ctx), GFP_KERNEL);
-	if (!ctx) {
-		ret = -ENOMEM;
-		goto err_del_component;
-	}
+	if (!ctx)
+		return -ENOMEM;
+
+	ctx->manager.type = EXYNOS_DISPLAY_TYPE_LCD;
+	ctx->manager.ops = &fimd_manager_ops;
+
+	ret = exynos_drm_component_add(dev, EXYNOS_DEVICE_TYPE_CRTC,
+				       ctx->manager.type);
+	if (ret)
+		return ret;
 
 	ctx->dev = dev;
 	ctx->suspended = true;
@@ -1179,27 +1212,27 @@ static int fimd_probe(struct platform_device *pdev)
 	init_waitqueue_head(&ctx->wait_vsync_queue);
 	atomic_set(&ctx->wait_vsync_event, 0);
 
-	platform_set_drvdata(pdev, &fimd_manager);
-
-	fimd_manager.ctx = ctx;
+	platform_set_drvdata(pdev, ctx);
 
 	ctx->display = exynos_dpi_probe(dev);
-	if (IS_ERR(ctx->display))
-		return PTR_ERR(ctx->display);
+	if (IS_ERR(ctx->display)) {
+		ret = PTR_ERR(ctx->display);
+		goto err_del_component;
+	}
 
-	pm_runtime_enable(&pdev->dev);
+	pm_runtime_enable(dev);
 
-	ret = component_add(&pdev->dev, &fimd_component_ops);
+	ret = component_add(dev, &fimd_component_ops);
 	if (ret)
 		goto err_disable_pm_runtime;
 
 	return ret;
 
 err_disable_pm_runtime:
-	pm_runtime_disable(&pdev->dev);
+	pm_runtime_disable(dev);
 
 err_del_component:
-	exynos_drm_component_del(&pdev->dev, EXYNOS_DEVICE_TYPE_CRTC);
+	exynos_drm_component_del(dev, EXYNOS_DEVICE_TYPE_CRTC);
 	return ret;
 }
 
diff --git a/drivers/gpu/drm/exynos/exynos_drm_g2d.c b/drivers/gpu/drm/exynos/exynos_drm_g2d.c
index df7a77d3eff8..81a250830808 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_g2d.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_g2d.c
@@ -302,9 +302,12 @@ static void g2d_fini_cmdlist(struct g2d_data *g2d)
 	struct exynos_drm_subdrv *subdrv = &g2d->subdrv;
 
 	kfree(g2d->cmdlist_node);
-	dma_free_attrs(subdrv->drm_dev->dev, G2D_CMDLIST_POOL_SIZE,
-			g2d->cmdlist_pool_virt,
-			g2d->cmdlist_pool, &g2d->cmdlist_dma_attrs);
+
+	if (g2d->cmdlist_pool_virt && g2d->cmdlist_pool) {
+		dma_free_attrs(subdrv->drm_dev->dev, G2D_CMDLIST_POOL_SIZE,
+				g2d->cmdlist_pool_virt,
+				g2d->cmdlist_pool, &g2d->cmdlist_dma_attrs);
+	}
 }
 
 static struct g2d_cmdlist_node *g2d_get_cmdlist(struct g2d_data *g2d)
@@ -1537,7 +1540,7 @@ static int g2d_resume(struct device *dev)
 }
 #endif
 
-#ifdef CONFIG_PM_RUNTIME
+#ifdef CONFIG_PM
 static int g2d_runtime_suspend(struct device *dev)
 {
 	struct g2d_data *g2d = dev_get_drvdata(dev);
diff --git a/drivers/gpu/drm/exynos/exynos_drm_gem.c b/drivers/gpu/drm/exynos/exynos_drm_gem.c
index 15db80138382..0d5b9698d384 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_gem.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_gem.c
@@ -318,40 +318,16 @@ void exynos_drm_gem_put_dma_addr(struct drm_device *dev,
 	drm_gem_object_unreference_unlocked(obj);
 }
 
-int exynos_drm_gem_map_offset_ioctl(struct drm_device *dev, void *data,
-				    struct drm_file *file_priv)
-{
-	struct drm_exynos_gem_map_off *args = data;
-
-	DRM_DEBUG_KMS("handle = 0x%x, offset = 0x%lx\n",
-			args->handle, (unsigned long)args->offset);
-
-	if (!(dev->driver->driver_features & DRIVER_GEM)) {
-		DRM_ERROR("does not support GEM.\n");
-		return -ENODEV;
-	}
-
-	return exynos_drm_gem_dumb_map_offset(file_priv, dev, args->handle,
-			&args->offset);
-}
-
-int exynos_drm_gem_mmap_buffer(struct file *filp,
+int exynos_drm_gem_mmap_buffer(struct exynos_drm_gem_obj *exynos_gem_obj,
 				      struct vm_area_struct *vma)
 {
-	struct drm_gem_object *obj = filp->private_data;
-	struct exynos_drm_gem_obj *exynos_gem_obj = to_exynos_gem_obj(obj);
-	struct drm_device *drm_dev = obj->dev;
+	struct drm_device *drm_dev = exynos_gem_obj->base.dev;
 	struct exynos_drm_gem_buf *buffer;
 	unsigned long vm_size;
 	int ret;
 
-	WARN_ON(!mutex_is_locked(&obj->dev->struct_mutex));
-
-	vma->vm_flags |= VM_IO | VM_DONTEXPAND | VM_DONTDUMP;
-	vma->vm_private_data = obj;
-	vma->vm_ops = drm_dev->driver->gem_vm_ops;
-
-	update_vm_cache_attr(exynos_gem_obj, vma);
+	vma->vm_flags &= ~VM_PFNMAP;
+	vma->vm_pgoff = 0;
 
 	vm_size = vma->vm_end - vma->vm_start;
 
@@ -373,60 +349,6 @@ int exynos_drm_gem_mmap_buffer(struct file *filp,
 		return ret;
 	}
 
-	/*
-	 * take a reference to this mapping of the object. And this reference
-	 * is unreferenced by the corresponding vm_close call.
-	 */
-	drm_gem_object_reference(obj);
-
-	drm_vm_open_locked(drm_dev, vma);
-
-	return 0;
-}
-
-int exynos_drm_gem_mmap_ioctl(struct drm_device *dev, void *data,
-			      struct drm_file *file_priv)
-{
-	struct drm_exynos_file_private *exynos_file_priv;
-	struct drm_exynos_gem_mmap *args = data;
-	struct drm_gem_object *obj;
-	struct file *anon_filp;
-	unsigned long addr;
-
-	if (!(dev->driver->driver_features & DRIVER_GEM)) {
-		DRM_ERROR("does not support GEM.\n");
-		return -ENODEV;
-	}
-
-	mutex_lock(&dev->struct_mutex);
-
-	obj = drm_gem_object_lookup(dev, file_priv, args->handle);
-	if (!obj) {
-		DRM_ERROR("failed to lookup gem object.\n");
-		mutex_unlock(&dev->struct_mutex);
-		return -EINVAL;
-	}
-
-	exynos_file_priv = file_priv->driver_priv;
-	anon_filp = exynos_file_priv->anon_filp;
-	anon_filp->private_data = obj;
-
-	addr = vm_mmap(anon_filp, 0, args->size, PROT_READ | PROT_WRITE,
-			MAP_SHARED, 0);
-
-	drm_gem_object_unreference(obj);
-
-	if (IS_ERR_VALUE(addr)) {
-		mutex_unlock(&dev->struct_mutex);
-		return (int)addr;
-	}
-
-	mutex_unlock(&dev->struct_mutex);
-
-	args->mapped = addr;
-
-	DRM_DEBUG_KMS("mapped = 0x%lx\n", (unsigned long)args->mapped);
-
 	return 0;
 }
 
@@ -710,16 +632,20 @@ int exynos_drm_gem_mmap(struct file *filp, struct vm_area_struct *vma)
 	exynos_gem_obj = to_exynos_gem_obj(obj);
 
 	ret = check_gem_flags(exynos_gem_obj->flags);
-	if (ret) {
-		drm_gem_vm_close(vma);
-		drm_gem_free_mmap_offset(obj);
-		return ret;
-	}
-
-	vma->vm_flags &= ~VM_PFNMAP;
-	vma->vm_flags |= VM_MIXEDMAP;
+	if (ret)
+		goto err_close_vm;
 
 	update_vm_cache_attr(exynos_gem_obj, vma);
 
+	ret = exynos_drm_gem_mmap_buffer(exynos_gem_obj, vma);
+	if (ret)
+		goto err_close_vm;
+
+	return ret;
+
+err_close_vm:
+	drm_gem_vm_close(vma);
+	drm_gem_free_mmap_offset(obj);
+
 	return ret;
 }
diff --git a/drivers/gpu/drm/exynos/exynos_drm_gem.h b/drivers/gpu/drm/exynos/exynos_drm_gem.h
index 1592c0ba7de8..ec58fe9c40df 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_gem.h
+++ b/drivers/gpu/drm/exynos/exynos_drm_gem.h
@@ -12,6 +12,8 @@
 #ifndef _EXYNOS_DRM_GEM_H_
 #define _EXYNOS_DRM_GEM_H_
 
+#include <drm/drm_gem.h>
+
 #define to_exynos_gem_obj(x)	container_of(x,\
 			struct exynos_drm_gem_obj, base)
 
@@ -111,20 +113,6 @@ void exynos_drm_gem_put_dma_addr(struct drm_device *dev,
 					unsigned int gem_handle,
 					struct drm_file *filp);
 
-/* get buffer offset to map to user space. */
-int exynos_drm_gem_map_offset_ioctl(struct drm_device *dev, void *data,
-				    struct drm_file *file_priv);
-
-/*
- * mmap the physically continuous memory that a gem object contains
- * to user space.
- */
-int exynos_drm_gem_mmap_ioctl(struct drm_device *dev, void *data,
-			      struct drm_file *file_priv);
-
-int exynos_drm_gem_mmap_buffer(struct file *filp,
-				      struct vm_area_struct *vma);
-
 /* map user space allocated by malloc to pages. */
 int exynos_drm_gem_userptr_ioctl(struct drm_device *dev, void *data,
 				      struct drm_file *file_priv);
diff --git a/drivers/gpu/drm/exynos/exynos_drm_gsc.c b/drivers/gpu/drm/exynos/exynos_drm_gsc.c
index 9e3ff1672965..0261468c8019 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_gsc.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_gsc.c
@@ -1326,8 +1326,7 @@ static irqreturn_t gsc_irq_handler(int irq, void *dev_id)
 			buf_id[EXYNOS_DRM_OPS_SRC];
 		event_work->buf_id[EXYNOS_DRM_OPS_DST] =
 			buf_id[EXYNOS_DRM_OPS_DST];
-		queue_work(ippdrv->event_workq,
-			(struct work_struct *)event_work);
+		queue_work(ippdrv->event_workq, &event_work->work);
 	}
 
 	return IRQ_HANDLED;
@@ -1765,7 +1764,7 @@ static int gsc_resume(struct device *dev)
 }
 #endif
 
-#ifdef CONFIG_PM_RUNTIME
+#ifdef CONFIG_PM
 static int gsc_runtime_suspend(struct device *dev)
 {
 	struct gsc_context *ctx = get_gsc_context(dev);
diff --git a/drivers/gpu/drm/exynos/exynos_drm_iommu.h b/drivers/gpu/drm/exynos/exynos_drm_iommu.h
index 72376d41c512..35d25889b476 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_iommu.h
+++ b/drivers/gpu/drm/exynos/exynos_drm_iommu.h
@@ -40,7 +40,6 @@ static inline bool is_drm_iommu_supported(struct drm_device *drm_dev)
 
 #else
 
-struct dma_iommu_mapping;
 static inline int drm_create_iommu_mapping(struct drm_device *drm_dev)
 {
 	return 0;
diff --git a/drivers/gpu/drm/exynos/exynos_drm_ipp.c b/drivers/gpu/drm/exynos/exynos_drm_ipp.c
index c411399070d6..d5ad17dfc24d 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_ipp.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_ipp.c
@@ -75,7 +75,6 @@ struct drm_exynos_ipp_mem_node {
 	u32	prop_id;
 	u32	buf_id;
 	struct drm_exynos_ipp_buf_info	buf_info;
-	struct drm_file		*filp;
 };
 
 /*
@@ -319,44 +318,6 @@ static void ipp_print_property(struct drm_exynos_ipp_property *property,
 		sz->hsize, sz->vsize, config->flip, config->degree);
 }
 
-static int ipp_find_and_set_property(struct drm_exynos_ipp_property *property)
-{
-	struct exynos_drm_ippdrv *ippdrv;
-	struct drm_exynos_ipp_cmd_node *c_node;
-	u32 prop_id = property->prop_id;
-
-	DRM_DEBUG_KMS("prop_id[%d]\n", prop_id);
-
-	ippdrv = ipp_find_drv_by_handle(prop_id);
-	if (IS_ERR(ippdrv)) {
-		DRM_ERROR("failed to get ipp driver.\n");
-		return -EINVAL;
-	}
-
-	/*
-	 * Find command node using command list in ippdrv.
-	 * when we find this command no using prop_id.
-	 * return property information set in this command node.
-	 */
-	mutex_lock(&ippdrv->cmd_lock);
-	list_for_each_entry(c_node, &ippdrv->cmd_list, list) {
-		if ((c_node->property.prop_id == prop_id) &&
-		    (c_node->state == IPP_STATE_STOP)) {
-			mutex_unlock(&ippdrv->cmd_lock);
-			DRM_DEBUG_KMS("found cmd[%d]ippdrv[0x%x]\n",
-				property->cmd, (int)ippdrv);
-
-			c_node->property = *property;
-			return 0;
-		}
-	}
-	mutex_unlock(&ippdrv->cmd_lock);
-
-	DRM_ERROR("failed to search property.\n");
-
-	return -EINVAL;
-}
-
 static struct drm_exynos_ipp_cmd_work *ipp_create_cmd_work(void)
 {
 	struct drm_exynos_ipp_cmd_work *cmd_work;
@@ -392,6 +353,7 @@ int exynos_drm_ipp_set_property(struct drm_device *drm_dev, void *data,
 	struct drm_exynos_ipp_property *property = data;
 	struct exynos_drm_ippdrv *ippdrv;
 	struct drm_exynos_ipp_cmd_node *c_node;
+	u32 prop_id;
 	int ret, i;
 
 	if (!ctx) {
@@ -404,6 +366,8 @@ int exynos_drm_ipp_set_property(struct drm_device *drm_dev, void *data,
 		return -EINVAL;
 	}
 
+	prop_id = property->prop_id;
+
 	/*
 	 * This is log print for user application property.
 	 * user application set various property.
@@ -412,14 +376,24 @@ int exynos_drm_ipp_set_property(struct drm_device *drm_dev, void *data,
 		ipp_print_property(property, i);
 
 	/*
-	 * set property ioctl generated new prop_id.
-	 * but in this case already asigned prop_id using old set property.
-	 * e.g PAUSE state. this case supports find current prop_id and use it
-	 * instead of allocation.
+	 * In case prop_id is not zero try to set existing property.
 	 */
-	if (property->prop_id) {
-		DRM_DEBUG_KMS("prop_id[%d]\n", property->prop_id);
-		return ipp_find_and_set_property(property);
+	if (prop_id) {
+		c_node = ipp_find_obj(&ctx->prop_idr, &ctx->prop_lock, prop_id);
+
+		if (!c_node || c_node->filp != file) {
+			DRM_DEBUG_KMS("prop_id[%d] not found\n", prop_id);
+			return -EINVAL;
+		}
+
+		if (c_node->state != IPP_STATE_STOP) {
+			DRM_DEBUG_KMS("prop_id[%d] not stopped\n", prop_id);
+			return -EINVAL;
+		}
+
+		c_node->property = *property;
+
+		return 0;
 	}
 
 	/* find ipp driver using ipp id */
@@ -445,25 +419,28 @@ int exynos_drm_ipp_set_property(struct drm_device *drm_dev, void *data,
 		property->prop_id, property->cmd, (int)ippdrv);
 
 	/* stored property information and ippdrv in private data */
-	c_node->dev = dev;
 	c_node->property = *property;
 	c_node->state = IPP_STATE_IDLE;
+	c_node->filp = file;
 
 	c_node->start_work = ipp_create_cmd_work();
 	if (IS_ERR(c_node->start_work)) {
 		DRM_ERROR("failed to create start work.\n");
+		ret = PTR_ERR(c_node->start_work);
 		goto err_remove_id;
 	}
 
 	c_node->stop_work = ipp_create_cmd_work();
 	if (IS_ERR(c_node->stop_work)) {
 		DRM_ERROR("failed to create stop work.\n");
+		ret = PTR_ERR(c_node->stop_work);
 		goto err_free_start;
 	}
 
 	c_node->event_work = ipp_create_event_work();
 	if (IS_ERR(c_node->event_work)) {
 		DRM_ERROR("failed to create event work.\n");
+		ret = PTR_ERR(c_node->event_work);
 		goto err_free_stop;
 	}
 
@@ -499,105 +476,37 @@ err_clear:
 	return ret;
 }
 
-static void ipp_clean_cmd_node(struct ipp_context *ctx,
-				struct drm_exynos_ipp_cmd_node *c_node)
-{
-	/* delete list */
-	list_del(&c_node->list);
-
-	ipp_remove_id(&ctx->prop_idr, &ctx->prop_lock,
-			c_node->property.prop_id);
-
-	/* destroy mutex */
-	mutex_destroy(&c_node->lock);
-	mutex_destroy(&c_node->mem_lock);
-	mutex_destroy(&c_node->event_lock);
-
-	/* free command node */
-	kfree(c_node->start_work);
-	kfree(c_node->stop_work);
-	kfree(c_node->event_work);
-	kfree(c_node);
-}
-
-static bool ipp_check_mem_list(struct drm_exynos_ipp_cmd_node *c_node)
-{
-	switch (c_node->property.cmd) {
-	case IPP_CMD_WB:
-		return !list_empty(&c_node->mem_list[EXYNOS_DRM_OPS_DST]);
-	case IPP_CMD_OUTPUT:
-		return !list_empty(&c_node->mem_list[EXYNOS_DRM_OPS_SRC]);
-	case IPP_CMD_M2M:
-	default:
-		return !list_empty(&c_node->mem_list[EXYNOS_DRM_OPS_SRC]) &&
-		       !list_empty(&c_node->mem_list[EXYNOS_DRM_OPS_DST]);
-	}
-}
-
-static struct drm_exynos_ipp_mem_node
-		*ipp_find_mem_node(struct drm_exynos_ipp_cmd_node *c_node,
-		struct drm_exynos_ipp_queue_buf *qbuf)
-{
-	struct drm_exynos_ipp_mem_node *m_node;
-	struct list_head *head;
-	int count = 0;
-
-	DRM_DEBUG_KMS("buf_id[%d]\n", qbuf->buf_id);
-
-	/* source/destination memory list */
-	head = &c_node->mem_list[qbuf->ops_id];
-
-	/* find memory node from memory list */
-	list_for_each_entry(m_node, head, list) {
-		DRM_DEBUG_KMS("count[%d]m_node[0x%x]\n", count++, (int)m_node);
-
-		/* compare buffer id */
-		if (m_node->buf_id == qbuf->buf_id)
-			return m_node;
-	}
-
-	return NULL;
-}
-
-static int ipp_set_mem_node(struct exynos_drm_ippdrv *ippdrv,
+static int ipp_put_mem_node(struct drm_device *drm_dev,
 		struct drm_exynos_ipp_cmd_node *c_node,
 		struct drm_exynos_ipp_mem_node *m_node)
 {
-	struct exynos_drm_ipp_ops *ops = NULL;
-	int ret = 0;
+	int i;
 
 	DRM_DEBUG_KMS("node[0x%x]\n", (int)m_node);
 
 	if (!m_node) {
-		DRM_ERROR("invalid queue node.\n");
+		DRM_ERROR("invalid dequeue node.\n");
 		return -EFAULT;
 	}
 
 	DRM_DEBUG_KMS("ops_id[%d]\n", m_node->ops_id);
 
-	/* get operations callback */
-	ops = ippdrv->ops[m_node->ops_id];
-	if (!ops) {
-		DRM_ERROR("not support ops.\n");
-		return -EFAULT;
+	/* put gem buffer */
+	for_each_ipp_planar(i) {
+		unsigned long handle = m_node->buf_info.handles[i];
+		if (handle)
+			exynos_drm_gem_put_dma_addr(drm_dev, handle,
+							c_node->filp);
 	}
 
-	/* set address and enable irq */
-	if (ops->set_addr) {
-		ret = ops->set_addr(ippdrv->dev, &m_node->buf_info,
-			m_node->buf_id, IPP_BUF_ENQUEUE);
-		if (ret) {
-			DRM_ERROR("failed to set addr.\n");
-			return ret;
-		}
-	}
+	list_del(&m_node->list);
+	kfree(m_node);
 
-	return ret;
+	return 0;
 }
 
 static struct drm_exynos_ipp_mem_node
 		*ipp_get_mem_node(struct drm_device *drm_dev,
-		struct drm_file *file,
 		struct drm_exynos_ipp_cmd_node *c_node,
 		struct drm_exynos_ipp_queue_buf *qbuf)
 {
@@ -615,6 +524,7 @@ static struct drm_exynos_ipp_mem_node
 	m_node->ops_id = qbuf->ops_id;
 	m_node->prop_id = qbuf->prop_id;
 	m_node->buf_id = qbuf->buf_id;
+	INIT_LIST_HEAD(&m_node->list);
 
 	DRM_DEBUG_KMS("m_node[0x%x]ops_id[%d]\n", (int)m_node, qbuf->ops_id);
 	DRM_DEBUG_KMS("prop_id[%d]buf_id[%d]\n", qbuf->prop_id, m_node->buf_id);
@@ -627,10 +537,11 @@ static struct drm_exynos_ipp_mem_node
 			dma_addr_t *addr;
 
 			addr = exynos_drm_gem_get_dma_addr(drm_dev,
-					qbuf->handle[i], file);
+					qbuf->handle[i], c_node->filp);
 			if (IS_ERR(addr)) {
 				DRM_ERROR("failed to get addr.\n");
-				goto err_clear;
+				ipp_put_mem_node(drm_dev, c_node, m_node);
+				return ERR_PTR(-EFAULT);
 			}
 
 			buf_info->handles[i] = qbuf->handle[i];
@@ -640,46 +551,30 @@ static struct drm_exynos_ipp_mem_node
 		}
 	}
 
-	m_node->filp = file;
 	mutex_lock(&c_node->mem_lock);
 	list_add_tail(&m_node->list, &c_node->mem_list[qbuf->ops_id]);
 	mutex_unlock(&c_node->mem_lock);
 
 	return m_node;
-
-err_clear:
-	kfree(m_node);
-	return ERR_PTR(-EFAULT);
 }
 
-static int ipp_put_mem_node(struct drm_device *drm_dev,
-		struct drm_exynos_ipp_cmd_node *c_node,
-		struct drm_exynos_ipp_mem_node *m_node)
+static void ipp_clean_mem_nodes(struct drm_device *drm_dev,
+			       struct drm_exynos_ipp_cmd_node *c_node, int ops)
 {
-	int i;
-
-	DRM_DEBUG_KMS("node[0x%x]\n", (int)m_node);
+	struct drm_exynos_ipp_mem_node *m_node, *tm_node;
+	struct list_head *head = &c_node->mem_list[ops];
 
-	if (!m_node) {
-		DRM_ERROR("invalid dequeue node.\n");
-		return -EFAULT;
-	}
+	mutex_lock(&c_node->mem_lock);
 
-	DRM_DEBUG_KMS("ops_id[%d]\n", m_node->ops_id);
+	list_for_each_entry_safe(m_node, tm_node, head, list) {
+		int ret;
 
-	/* put gem buffer */
-	for_each_ipp_planar(i) {
-		unsigned long handle = m_node->buf_info.handles[i];
-		if (handle)
-			exynos_drm_gem_put_dma_addr(drm_dev, handle,
-							m_node->filp);
+		ret = ipp_put_mem_node(drm_dev, c_node, m_node);
+		if (ret)
+			DRM_ERROR("failed to put m_node.\n");
 	}
 
-	/* delete list in queue */
-	list_del(&m_node->list);
-	kfree(m_node);
-
-	return 0;
+	mutex_unlock(&c_node->mem_lock);
 }
 
 static void ipp_free_event(struct drm_pending_event *event)
@@ -688,7 +583,6 @@ static void ipp_free_event(struct drm_pending_event *event)
 }
 
 static int ipp_get_event(struct drm_device *drm_dev,
-		struct drm_file *file,
 		struct drm_exynos_ipp_cmd_node *c_node,
 		struct drm_exynos_ipp_queue_buf *qbuf)
 {
@@ -700,7 +594,7 @@ static int ipp_get_event(struct drm_device *drm_dev,
 	e = kzalloc(sizeof(*e), GFP_KERNEL);
 	if (!e) {
 		spin_lock_irqsave(&drm_dev->event_lock, flags);
-		file->event_space += sizeof(e->event);
+		c_node->filp->event_space += sizeof(e->event);
 		spin_unlock_irqrestore(&drm_dev->event_lock, flags);
 		return -ENOMEM;
 	}
@@ -712,7 +606,7 @@ static int ipp_get_event(struct drm_device *drm_dev,
 	e->event.prop_id = qbuf->prop_id;
 	e->event.buf_id[EXYNOS_DRM_OPS_DST] = qbuf->buf_id;
 	e->base.event = &e->event.base;
-	e->base.file_priv = file;
+	e->base.file_priv = c_node->filp;
 	e->base.destroy = ipp_free_event;
 	mutex_lock(&c_node->event_lock);
 	list_add_tail(&e->base.link, &c_node->event_list);
@@ -757,6 +651,115 @@ out_unlock:
 	return;
 }
 
+static void ipp_clean_cmd_node(struct ipp_context *ctx,
+				struct drm_exynos_ipp_cmd_node *c_node)
+{
+	int i;
+
+	/* cancel works */
+	cancel_work_sync(&c_node->start_work->work);
+	cancel_work_sync(&c_node->stop_work->work);
+	cancel_work_sync(&c_node->event_work->work);
+
+	/* put event */
+	ipp_put_event(c_node, NULL);
+
+	for_each_ipp_ops(i)
+		ipp_clean_mem_nodes(ctx->subdrv.drm_dev, c_node, i);
+
+	/* delete list */
+	list_del(&c_node->list);
+
+	ipp_remove_id(&ctx->prop_idr, &ctx->prop_lock,
+			c_node->property.prop_id);
+
+	/* destroy mutex */
+	mutex_destroy(&c_node->lock);
+	mutex_destroy(&c_node->mem_lock);
+	mutex_destroy(&c_node->event_lock);
+
+	/* free command node */
+	kfree(c_node->start_work);
+	kfree(c_node->stop_work);
+	kfree(c_node->event_work);
+	kfree(c_node);
+}
+
+static bool ipp_check_mem_list(struct drm_exynos_ipp_cmd_node *c_node)
+{
+	switch (c_node->property.cmd) {
+	case IPP_CMD_WB:
+		return !list_empty(&c_node->mem_list[EXYNOS_DRM_OPS_DST]);
+	case IPP_CMD_OUTPUT:
+		return !list_empty(&c_node->mem_list[EXYNOS_DRM_OPS_SRC]);
+	case IPP_CMD_M2M:
+	default:
+		return !list_empty(&c_node->mem_list[EXYNOS_DRM_OPS_SRC]) &&
+		       !list_empty(&c_node->mem_list[EXYNOS_DRM_OPS_DST]);
+	}
+}
+
+static struct drm_exynos_ipp_mem_node
+		*ipp_find_mem_node(struct drm_exynos_ipp_cmd_node *c_node,
+		struct drm_exynos_ipp_queue_buf *qbuf)
+{
+	struct drm_exynos_ipp_mem_node *m_node;
+	struct list_head *head;
+	int count = 0;
+
+	DRM_DEBUG_KMS("buf_id[%d]\n", qbuf->buf_id);
+
+	/* source/destination memory list */
+	head = &c_node->mem_list[qbuf->ops_id];
+
+	/* find memory node from memory list */
+	list_for_each_entry(m_node, head, list) {
+		DRM_DEBUG_KMS("count[%d]m_node[0x%x]\n", count++, (int)m_node);
+
+		/* compare buffer id */
+		if (m_node->buf_id == qbuf->buf_id)
+			return m_node;
+	}
+
+	return NULL;
+}
+
+static int ipp_set_mem_node(struct exynos_drm_ippdrv *ippdrv,
+		struct drm_exynos_ipp_cmd_node *c_node,
+		struct drm_exynos_ipp_mem_node *m_node)
+{
+	struct exynos_drm_ipp_ops *ops = NULL;
+	int ret = 0;
+
+	DRM_DEBUG_KMS("node[0x%x]\n", (int)m_node);
+
+	if (!m_node) {
+		DRM_ERROR("invalid queue node.\n");
+		return -EFAULT;
+	}
+
+	DRM_DEBUG_KMS("ops_id[%d]\n", m_node->ops_id);
+
+	/* get operations callback */
+	ops = ippdrv->ops[m_node->ops_id];
+	if (!ops) {
+		DRM_ERROR("not support ops.\n");
+		return -EFAULT;
+	}
+
+	/* set address and enable irq */
+	if (ops->set_addr) {
+		ret = ops->set_addr(ippdrv->dev, &m_node->buf_info,
+			m_node->buf_id, IPP_BUF_ENQUEUE);
+		if (ret) {
+			DRM_ERROR("failed to set addr.\n");
+			return ret;
+		}
+	}
+
+	return ret;
+}
+
 static void ipp_handle_cmd_work(struct device *dev,
 		struct exynos_drm_ippdrv *ippdrv,
 		struct drm_exynos_ipp_cmd_work *cmd_work,
@@ -766,7 +769,7 @@ static void ipp_handle_cmd_work(struct device *dev,
 
 	cmd_work->ippdrv = ippdrv;
 	cmd_work->c_node = c_node;
-	queue_work(ctx->cmd_workq, (struct work_struct *)cmd_work);
+	queue_work(ctx->cmd_workq, &cmd_work->work);
 }
 
 static int ipp_queue_buf_with_run(struct device *dev,
@@ -872,7 +875,7 @@ int exynos_drm_ipp_queue_buf(struct drm_device *drm_dev, void *data,
 	/* find command node */
 	c_node = ipp_find_obj(&ctx->prop_idr, &ctx->prop_lock,
 		qbuf->prop_id);
-	if (!c_node) {
+	if (!c_node || c_node->filp != file) {
 		DRM_ERROR("failed to get command node.\n");
 		return -ENODEV;
 	}
@@ -881,7 +884,7 @@ int exynos_drm_ipp_queue_buf(struct drm_device *drm_dev, void *data,
 	switch (qbuf->buf_type) {
 	case IPP_BUF_ENQUEUE:
 		/* get memory node */
-		m_node = ipp_get_mem_node(drm_dev, file, c_node, qbuf);
+		m_node = ipp_get_mem_node(drm_dev, c_node, qbuf);
 		if (IS_ERR(m_node)) {
 			DRM_ERROR("failed to get m_node.\n");
 			return PTR_ERR(m_node);
@@ -894,7 +897,7 @@ int exynos_drm_ipp_queue_buf(struct drm_device *drm_dev, void *data,
 		 */
 		if (qbuf->ops_id == EXYNOS_DRM_OPS_DST) {
 			/* get event for destination buffer */
-			ret = ipp_get_event(drm_dev, file, c_node, qbuf);
+			ret = ipp_get_event(drm_dev, c_node, qbuf);
 			if (ret) {
 				DRM_ERROR("failed to get event.\n");
 				goto err_clean_node;
@@ -1007,7 +1010,7 @@ int exynos_drm_ipp_cmd_ctrl(struct drm_device *drm_dev, void *data,
 
 	c_node = ipp_find_obj(&ctx->prop_idr, &ctx->prop_lock,
 		cmd_ctrl->prop_id);
-	if (!c_node) {
+	if (!c_node || c_node->filp != file) {
 		DRM_ERROR("invalid command node list.\n");
 		return -ENODEV;
 	}
@@ -1257,80 +1260,39 @@ static int ipp_stop_property(struct drm_device *drm_dev,
 		struct exynos_drm_ippdrv *ippdrv,
 		struct drm_exynos_ipp_cmd_node *c_node)
 {
-	struct drm_exynos_ipp_mem_node *m_node, *tm_node;
 	struct drm_exynos_ipp_property *property = &c_node->property;
-	struct list_head *head;
-	int ret = 0, i;
+	int i;
 
 	DRM_DEBUG_KMS("prop_id[%d]\n", property->prop_id);
 
-	/* put event */
-	ipp_put_event(c_node, NULL);
-
-	mutex_lock(&c_node->mem_lock);
+	/* stop operations */
+	if (ippdrv->stop)
+		ippdrv->stop(ippdrv->dev, property->cmd);
 
 	/* check command */
 	switch (property->cmd) {
 	case IPP_CMD_M2M:
-		for_each_ipp_ops(i) {
-			/* source/destination memory list */
-			head = &c_node->mem_list[i];
-
-			list_for_each_entry_safe(m_node, tm_node,
-				head, list) {
-				ret = ipp_put_mem_node(drm_dev, c_node,
-					m_node);
-				if (ret) {
-					DRM_ERROR("failed to put m_node.\n");
-					goto err_clear;
-				}
-			}
-		}
+		for_each_ipp_ops(i)
+			ipp_clean_mem_nodes(drm_dev, c_node, i);
 		break;
 	case IPP_CMD_WB:
-		/* destination memory list */
-		head = &c_node->mem_list[EXYNOS_DRM_OPS_DST];
-
-		list_for_each_entry_safe(m_node, tm_node, head, list) {
-			ret = ipp_put_mem_node(drm_dev, c_node, m_node);
-			if (ret) {
-				DRM_ERROR("failed to put m_node.\n");
-				goto err_clear;
-			}
-		}
+		ipp_clean_mem_nodes(drm_dev, c_node, EXYNOS_DRM_OPS_DST);
 		break;
 	case IPP_CMD_OUTPUT:
-		/* source memory list */
-		head = &c_node->mem_list[EXYNOS_DRM_OPS_SRC];
-
-		list_for_each_entry_safe(m_node, tm_node, head, list) {
-			ret = ipp_put_mem_node(drm_dev, c_node, m_node);
-			if (ret) {
-				DRM_ERROR("failed to put m_node.\n");
-				goto err_clear;
-			}
-		}
+		ipp_clean_mem_nodes(drm_dev, c_node, EXYNOS_DRM_OPS_SRC);
 		break;
 	default:
 		DRM_ERROR("invalid operations.\n");
-		ret = -EINVAL;
-		goto err_clear;
+		return -EINVAL;
 	}
 
-err_clear:
-	mutex_unlock(&c_node->mem_lock);
-
-	/* stop operations */
-	if (ippdrv->stop)
-		ippdrv->stop(ippdrv->dev, property->cmd);
-
-	return ret;
+	return 0;
 }
 
 void ipp_sched_cmd(struct work_struct *work)
 {
 	struct drm_exynos_ipp_cmd_work *cmd_work =
-		(struct drm_exynos_ipp_cmd_work *)work;
+		container_of(work, struct drm_exynos_ipp_cmd_work, work);
 	struct exynos_drm_ippdrv *ippdrv;
 	struct drm_exynos_ipp_cmd_node *c_node;
 	struct drm_exynos_ipp_property *property;
@@ -1543,7 +1505,7 @@ err_event_unlock:
 void ipp_sched_event(struct work_struct *work)
 {
 	struct drm_exynos_ipp_event_work *event_work =
-		(struct drm_exynos_ipp_event_work *)work;
+		container_of(work, struct drm_exynos_ipp_event_work, work);
 	struct exynos_drm_ippdrv *ippdrv;
 	struct drm_exynos_ipp_cmd_node *c_node;
 	int ret;
@@ -1646,11 +1608,11 @@ err:
 
 static void ipp_subdrv_remove(struct drm_device *drm_dev, struct device *dev)
 {
-	struct exynos_drm_ippdrv *ippdrv;
+	struct exynos_drm_ippdrv *ippdrv, *t;
 	struct ipp_context *ctx = get_ipp_context(dev);
 
 	/* get ipp driver entry */
-	list_for_each_entry(ippdrv, &exynos_drm_ippdrv_list, drv_list) {
+	list_for_each_entry_safe(ippdrv, t, &exynos_drm_ippdrv_list, drv_list) {
 		if (is_drm_iommu_supported(drm_dev))
 			drm_iommu_detach_device(drm_dev, ippdrv->dev);
 
@@ -1677,14 +1639,11 @@ static int ipp_subdrv_open(struct drm_device *drm_dev, struct device *dev,
 static void ipp_subdrv_close(struct drm_device *drm_dev, struct device *dev,
 		struct drm_file *file)
 {
-	struct drm_exynos_file_private *file_priv = file->driver_priv;
 	struct exynos_drm_ippdrv *ippdrv = NULL;
 	struct ipp_context *ctx = get_ipp_context(dev);
 	struct drm_exynos_ipp_cmd_node *c_node, *tc_node;
 	int count = 0;
 
-	DRM_DEBUG_KMS("for priv[0x%x]\n", (int)file_priv->ipp_dev);
-
 	list_for_each_entry(ippdrv, &exynos_drm_ippdrv_list, drv_list) {
 		mutex_lock(&ippdrv->cmd_lock);
 		list_for_each_entry_safe(c_node, tc_node,
@@ -1692,7 +1651,7 @@ static void ipp_subdrv_close(struct drm_device *drm_dev, struct device *dev,
 			DRM_DEBUG_KMS("count[%d]ippdrv[0x%x]\n",
 				count++, (int)ippdrv);
 
-			if (c_node->dev == file_priv->ipp_dev) {
+			if (c_node->filp == file) {
 				/*
 				 * userland goto unnormal state. process killed.
 				 * and close the file.
@@ -1808,63 +1767,12 @@ static int ipp_remove(struct platform_device *pdev)
 	return 0;
 }
 
-static int ipp_power_ctrl(struct ipp_context *ctx, bool enable)
-{
-	DRM_DEBUG_KMS("enable[%d]\n", enable);
-
-	return 0;
-}
-
-#ifdef CONFIG_PM_SLEEP
-static int ipp_suspend(struct device *dev)
-{
-	struct ipp_context *ctx = get_ipp_context(dev);
-
-	if (pm_runtime_suspended(dev))
-		return 0;
-
-	return ipp_power_ctrl(ctx, false);
-}
-
-static int ipp_resume(struct device *dev)
-{
-	struct ipp_context *ctx = get_ipp_context(dev);
-
-	if (!pm_runtime_suspended(dev))
-		return ipp_power_ctrl(ctx, true);
-
-	return 0;
-}
-#endif
-
-#ifdef CONFIG_PM_RUNTIME
-static int ipp_runtime_suspend(struct device *dev)
-{
-	struct ipp_context *ctx = get_ipp_context(dev);
-
-	return ipp_power_ctrl(ctx, false);
-}
-
-static int ipp_runtime_resume(struct device *dev)
-{
-	struct ipp_context *ctx = get_ipp_context(dev);
-
-	return ipp_power_ctrl(ctx, true);
-}
-#endif
-
-static const struct dev_pm_ops ipp_pm_ops = {
-	SET_SYSTEM_SLEEP_PM_OPS(ipp_suspend, ipp_resume)
-	SET_RUNTIME_PM_OPS(ipp_runtime_suspend, ipp_runtime_resume, NULL)
-};
-
 struct platform_driver ipp_driver = {
 	.probe		= ipp_probe,
 	.remove		= ipp_remove,
 	.driver		= {
 		.name	= "exynos-drm-ipp",
 		.owner	= THIS_MODULE,
-		.pm	= &ipp_pm_ops,
 	},
 };
 
diff --git a/drivers/gpu/drm/exynos/exynos_drm_ipp.h b/drivers/gpu/drm/exynos/exynos_drm_ipp.h
index 6f48d62aeb30..2a61547a39d0 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_ipp.h
+++ b/drivers/gpu/drm/exynos/exynos_drm_ipp.h
@@ -48,7 +48,6 @@ struct drm_exynos_ipp_cmd_work {
 /*
  * A structure of command node.
  *
- * @dev: IPP device.
  * @list: list head to command queue information.
  * @event_list: list head of event.
  * @mem_list: list head to source,destination memory queue information.
@@ -62,9 +61,9 @@ struct drm_exynos_ipp_cmd_work {
  * @stop_work: stop command work structure.
  * @event_work: event work structure.
  * @state: state of command node.
+ * @filp: associated file pointer.
  */
 struct drm_exynos_ipp_cmd_node {
-	struct device		*dev;
 	struct list_head	list;
 	struct list_head	event_list;
 	struct list_head	mem_list[EXYNOS_DRM_OPS_MAX];
@@ -78,6 +77,7 @@ struct drm_exynos_ipp_cmd_node {
 	struct drm_exynos_ipp_cmd_work *stop_work;
 	struct drm_exynos_ipp_event_work *event_work;
 	enum drm_exynos_ipp_state	state;
+	struct drm_file	*filp;
 };
 
 /*
diff --git a/drivers/gpu/drm/exynos/exynos_drm_plane.c b/drivers/gpu/drm/exynos/exynos_drm_plane.c
index 8371cbd7631d..c7045a663763 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_plane.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_plane.c
@@ -139,6 +139,8 @@ int exynos_plane_mode_set(struct drm_plane *plane, struct drm_crtc *crtc,
 			overlay->crtc_x, overlay->crtc_y,
 			overlay->crtc_width, overlay->crtc_height);
 
+	plane->crtc = crtc;
+
 	exynos_drm_crtc_plane_mode_set(crtc, overlay);
 
 	return 0;
@@ -187,8 +189,6 @@ exynos_update_plane(struct drm_plane *plane, struct drm_crtc *crtc,
 	if (ret < 0)
 		return ret;
 
-	plane->crtc = crtc;
-
 	exynos_plane_commit(plane);
 	exynos_plane_dpms(plane, DRM_MODE_DPMS_ON);
 
@@ -254,25 +254,26 @@ static void exynos_plane_attach_zpos_property(struct drm_plane *plane)
 }
 
 struct drm_plane *exynos_plane_init(struct drm_device *dev,
-				    unsigned long possible_crtcs, bool priv)
+				    unsigned long possible_crtcs,
+				    enum drm_plane_type type)
 {
 	struct exynos_plane *exynos_plane;
 	int err;
 
 	exynos_plane = kzalloc(sizeof(struct exynos_plane), GFP_KERNEL);
 	if (!exynos_plane)
-		return NULL;
+		return ERR_PTR(-ENOMEM);
 
-	err = drm_plane_init(dev, &exynos_plane->base, possible_crtcs,
-			      &exynos_plane_funcs, formats, ARRAY_SIZE(formats),
-			      priv);
+	err = drm_universal_plane_init(dev, &exynos_plane->base, possible_crtcs,
+				       &exynos_plane_funcs, formats,
+				       ARRAY_SIZE(formats), type);
 	if (err) {
 		DRM_ERROR("failed to initialize plane\n");
 		kfree(exynos_plane);
-		return NULL;
+		return ERR_PTR(err);
 	}
 
-	if (priv)
+	if (type == DRM_PLANE_TYPE_PRIMARY)
 		exynos_plane->overlay.zpos = DEFAULT_ZPOS;
 	else
 		exynos_plane_attach_zpos_property(&exynos_plane->base);
diff --git a/drivers/gpu/drm/exynos/exynos_drm_plane.h b/drivers/gpu/drm/exynos/exynos_drm_plane.h
index 84d464c90d3d..0d1986b115f8 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_plane.h
+++ b/drivers/gpu/drm/exynos/exynos_drm_plane.h
@@ -17,4 +17,5 @@ int exynos_plane_mode_set(struct drm_plane *plane, struct drm_crtc *crtc,
 void exynos_plane_commit(struct drm_plane *plane);
 void exynos_plane_dpms(struct drm_plane *plane, int mode);
 struct drm_plane *exynos_plane_init(struct drm_device *dev,
-				    unsigned long possible_crtcs, bool priv);
+				    unsigned long possible_crtcs,
+				    enum drm_plane_type type);
diff --git a/drivers/gpu/drm/exynos/exynos_drm_rotator.c b/drivers/gpu/drm/exynos/exynos_drm_rotator.c
index 55af6b41c1df..425e70625388 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_rotator.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_rotator.c
@@ -156,8 +156,7 @@ static irqreturn_t rotator_irq_handler(int irq, void *arg)
 		event_work->ippdrv = ippdrv;
 		event_work->buf_id[EXYNOS_DRM_OPS_DST] =
 			rot->cur_buf_id[EXYNOS_DRM_OPS_DST];
-		queue_work(ippdrv->event_workq,
-			(struct work_struct *)event_work);
+		queue_work(ippdrv->event_workq, &event_work->work);
 	} else {
 		DRM_ERROR("the SFR is set illegally\n");
 	}
@@ -823,7 +822,7 @@ static int rotator_resume(struct device *dev)
 }
 #endif
 
-#ifdef CONFIG_PM_RUNTIME
+#ifdef CONFIG_PM
 static int rotator_runtime_suspend(struct device *dev)
 {
 	struct rot_context *rot = dev_get_drvdata(dev);
diff --git a/drivers/gpu/drm/exynos/exynos_drm_vidi.c b/drivers/gpu/drm/exynos/exynos_drm_vidi.c
index 9528d81d8004..45899fb63272 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_vidi.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_vidi.c
@@ -14,6 +14,7 @@
 
 #include <linux/kernel.h>
 #include <linux/platform_device.h>
+#include <linux/component.h>
 
 #include <drm/exynos_drm.h>
 
@@ -28,7 +29,6 @@
 /* vidi has totally three virtual windows. */
 #define WINDOWS_NR		3
 
-#define get_vidi_mgr(dev)	platform_get_drvdata(to_platform_device(dev))
 #define ctx_from_connector(c)	container_of(c, struct vidi_context, \
 					connector)
 
@@ -47,11 +47,13 @@ struct vidi_win_data {
 };
 
 struct vidi_context {
+	struct exynos_drm_manager	manager;
+	struct exynos_drm_display	display;
+	struct platform_device		*pdev;
 	struct drm_device		*drm_dev;
 	struct drm_crtc			*crtc;
 	struct drm_encoder		*encoder;
 	struct drm_connector		connector;
-	struct exynos_drm_subdrv	subdrv;
 	struct vidi_win_data		win_data[WINDOWS_NR];
 	struct edid			*raw_edid;
 	unsigned int			clkdiv;
@@ -66,6 +68,16 @@ struct vidi_context {
 	int				pipe;
 };
 
+static inline struct vidi_context *manager_to_vidi(struct exynos_drm_manager *m)
+{
+	return container_of(m, struct vidi_context, manager);
+}
+
+static inline struct vidi_context *display_to_vidi(struct exynos_drm_display *d)
+{
+	return container_of(d, struct vidi_context, display);
+}
+
 static const char fake_edid_info[] = {
 	0x00, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0x00, 0x4c, 0x2d, 0x05, 0x05,
 	0x00, 0x00, 0x00, 0x00, 0x30, 0x12, 0x01, 0x03, 0x80, 0x10, 0x09, 0x78,
@@ -93,7 +105,7 @@ static const char fake_edid_info[] = {
 
 static void vidi_apply(struct exynos_drm_manager *mgr)
 {
-	struct vidi_context *ctx = mgr->ctx;
+	struct vidi_context *ctx = manager_to_vidi(mgr);
 	struct exynos_drm_manager_ops *mgr_ops = mgr->ops;
 	struct vidi_win_data *win_data;
 	int i;
@@ -110,7 +122,7 @@ static void vidi_apply(struct exynos_drm_manager *mgr)
 
 static void vidi_commit(struct exynos_drm_manager *mgr)
 {
-	struct vidi_context *ctx = mgr->ctx;
+	struct vidi_context *ctx = manager_to_vidi(mgr);
 
 	if (ctx->suspended)
 		return;
@@ -118,7 +130,7 @@ static void vidi_commit(struct exynos_drm_manager *mgr)
 
 static int vidi_enable_vblank(struct exynos_drm_manager *mgr)
 {
-	struct vidi_context *ctx = mgr->ctx;
+	struct vidi_context *ctx = manager_to_vidi(mgr);
 
 	if (ctx->suspended)
 		return -EPERM;
@@ -140,7 +152,7 @@ static int vidi_enable_vblank(struct exynos_drm_manager *mgr)
 
 static void vidi_disable_vblank(struct exynos_drm_manager *mgr)
 {
-	struct vidi_context *ctx = mgr->ctx;
+	struct vidi_context *ctx = manager_to_vidi(mgr);
 
 	if (ctx->suspended)
 		return;
@@ -152,7 +164,7 @@ static void vidi_disable_vblank(struct exynos_drm_manager *mgr)
 static void vidi_win_mode_set(struct exynos_drm_manager *mgr,
 			struct exynos_drm_overlay *overlay)
 {
-	struct vidi_context *ctx = mgr->ctx;
+	struct vidi_context *ctx = manager_to_vidi(mgr);
 	struct vidi_win_data *win_data;
 	int win;
 	unsigned long offset;
@@ -204,7 +216,7 @@ static void vidi_win_mode_set(struct exynos_drm_manager *mgr,
 
 static void vidi_win_commit(struct exynos_drm_manager *mgr, int zpos)
 {
-	struct vidi_context *ctx = mgr->ctx;
+	struct vidi_context *ctx = manager_to_vidi(mgr);
 	struct vidi_win_data *win_data;
 	int win = zpos;
 
@@ -229,7 +241,7 @@ static void vidi_win_commit(struct exynos_drm_manager *mgr, int zpos)
 
 static void vidi_win_disable(struct exynos_drm_manager *mgr, int zpos)
 {
-	struct vidi_context *ctx = mgr->ctx;
+	struct vidi_context *ctx = manager_to_vidi(mgr);
 	struct vidi_win_data *win_data;
 	int win = zpos;
 
@@ -247,7 +259,7 @@ static void vidi_win_disable(struct exynos_drm_manager *mgr, int zpos)
 
 static int vidi_power_on(struct exynos_drm_manager *mgr, bool enable)
 {
-	struct vidi_context *ctx = mgr->ctx;
+	struct vidi_context *ctx = manager_to_vidi(mgr);
 
 	DRM_DEBUG_KMS("%s\n", __FILE__);
 
@@ -271,7 +283,7 @@ static int vidi_power_on(struct exynos_drm_manager *mgr, bool enable)
 
 static void vidi_dpms(struct exynos_drm_manager *mgr, int mode)
 {
-	struct vidi_context *ctx = mgr->ctx;
+	struct vidi_context *ctx = manager_to_vidi(mgr);
 
 	DRM_DEBUG_KMS("%d\n", mode);
 
@@ -297,29 +309,12 @@ static void vidi_dpms(struct exynos_drm_manager *mgr, int mode)
 static int vidi_mgr_initialize(struct exynos_drm_manager *mgr,
 			struct drm_device *drm_dev)
 {
-	struct vidi_context *ctx = mgr->ctx;
+	struct vidi_context *ctx = manager_to_vidi(mgr);
 	struct exynos_drm_private *priv = drm_dev->dev_private;
 
 	mgr->drm_dev = ctx->drm_dev = drm_dev;
 	mgr->pipe = ctx->pipe = priv->pipe++;
 
-	/*
-	 * enable drm irq mode.
-	 * - with irq_enabled = 1, we can use the vblank feature.
-	 *
-	 * P.S. note that we wouldn't use drm irq handler but
-	 *	just specific driver own one instead because
-	 *	drm framework supports only one irq handler.
-	 */
-	drm_dev->irq_enabled = 1;
-
-	/*
-	 * with vblank_disable_allowed = 1, vblank interrupt will be disabled
-	 * by drm timer once a current process gives up ownership of
-	 * vblank event.(after drm_vblank_put function is called)
-	 */
-	drm_dev->vblank_disable_allowed = 1;
-
 	return 0;
 }
 
@@ -333,11 +328,6 @@ static struct exynos_drm_manager_ops vidi_manager_ops = {
 	.win_disable = vidi_win_disable,
 };
 
-static struct exynos_drm_manager vidi_manager = {
-	.type = EXYNOS_DISPLAY_TYPE_VIDI,
-	.ops = &vidi_manager_ops,
-};
-
 static void vidi_fake_vblank_handler(struct work_struct *work)
 {
 	struct vidi_context *ctx = container_of(work, struct vidi_context,
@@ -366,9 +356,8 @@ static void vidi_fake_vblank_handler(struct work_struct *work)
 static int vidi_show_connection(struct device *dev,
 				struct device_attribute *attr, char *buf)
 {
+	struct vidi_context *ctx = dev_get_drvdata(dev);
 	int rc;
-	struct exynos_drm_manager *mgr = get_vidi_mgr(dev);
-	struct vidi_context *ctx = mgr->ctx;
 
 	mutex_lock(&ctx->lock);
 
@@ -383,8 +372,7 @@ static int vidi_store_connection(struct device *dev,
 				struct device_attribute *attr,
 				const char *buf, size_t len)
 {
-	struct exynos_drm_manager *mgr = get_vidi_mgr(dev);
-	struct vidi_context *ctx = mgr->ctx;
+	struct vidi_context *ctx = dev_get_drvdata(dev);
 	int ret;
 
 	ret = kstrtoint(buf, 0, &ctx->connected);
@@ -437,7 +425,7 @@ int vidi_connection_ioctl(struct drm_device *drm_dev, void *data,
 		display = exynos_drm_get_display(encoder);
 
 		if (display->type == EXYNOS_DISPLAY_TYPE_VIDI) {
-			ctx = display->ctx;
+			ctx = display_to_vidi(display);
 			break;
 		}
 	}
@@ -547,7 +535,7 @@ static struct drm_connector_helper_funcs vidi_connector_helper_funcs = {
 static int vidi_create_connector(struct exynos_drm_display *display,
 				struct drm_encoder *encoder)
 {
-	struct vidi_context *ctx = display->ctx;
+	struct vidi_context *ctx = display_to_vidi(display);
 	struct drm_connector *connector = &ctx->connector;
 	int ret;
 
@@ -573,27 +561,22 @@ static struct exynos_drm_display_ops vidi_display_ops = {
 	.create_connector = vidi_create_connector,
 };
 
-static struct exynos_drm_display vidi_display = {
-	.type = EXYNOS_DISPLAY_TYPE_VIDI,
-	.ops = &vidi_display_ops,
-};
-
-static int vidi_subdrv_probe(struct drm_device *drm_dev, struct device *dev)
+static int vidi_bind(struct device *dev, struct device *master, void *data)
 {
-	struct exynos_drm_manager *mgr = get_vidi_mgr(dev);
-	struct vidi_context *ctx = mgr->ctx;
+	struct vidi_context *ctx = dev_get_drvdata(dev);
+	struct drm_device *drm_dev = data;
 	struct drm_crtc *crtc = ctx->crtc;
 	int ret;
 
-	vidi_mgr_initialize(mgr, drm_dev);
+	vidi_mgr_initialize(&ctx->manager, drm_dev);
 
-	ret = exynos_drm_crtc_create(&vidi_manager);
+	ret = exynos_drm_crtc_create(&ctx->manager);
 	if (ret) {
 		DRM_ERROR("failed to create crtc.\n");
 		return ret;
 	}
 
-	ret = exynos_drm_create_enc_conn(drm_dev, &vidi_display);
+	ret = exynos_drm_create_enc_conn(drm_dev, &ctx->display);
 	if (ret) {
 		crtc->funcs->destroy(crtc);
 		DRM_ERROR("failed to create encoder and connector.\n");
@@ -603,9 +586,18 @@ static int vidi_subdrv_probe(struct drm_device *drm_dev, struct device *dev)
 	return 0;
 }
 
+
+static void vidi_unbind(struct device *dev, struct device *master, void *data)
+{
+}
+
+static const struct component_ops vidi_component_ops = {
+	.bind	= vidi_bind,
+	.unbind = vidi_unbind,
+};
+
 static int vidi_probe(struct platform_device *pdev)
 {
-	struct exynos_drm_subdrv *subdrv;
 	struct vidi_context *ctx;
 	int ret;
 
@@ -613,42 +605,54 @@ static int vidi_probe(struct platform_device *pdev)
 	if (!ctx)
 		return -ENOMEM;
 
+	ctx->manager.type = EXYNOS_DISPLAY_TYPE_VIDI;
+	ctx->manager.ops = &vidi_manager_ops;
+	ctx->display.type = EXYNOS_DISPLAY_TYPE_VIDI;
+	ctx->display.ops = &vidi_display_ops;
 	ctx->default_win = 0;
+	ctx->pdev = pdev;
 
-	INIT_WORK(&ctx->work, vidi_fake_vblank_handler);
-
-	vidi_manager.ctx = ctx;
-	vidi_display.ctx = ctx;
+	ret = exynos_drm_component_add(&pdev->dev, EXYNOS_DEVICE_TYPE_CRTC,
+					ctx->manager.type);
+	if (ret)
+		return ret;
 
-	mutex_init(&ctx->lock);
+	ret = exynos_drm_component_add(&pdev->dev, EXYNOS_DEVICE_TYPE_CONNECTOR,
+					ctx->display.type);
+	if (ret)
+		goto err_del_crtc_component;
 
-	platform_set_drvdata(pdev, &vidi_manager);
+	INIT_WORK(&ctx->work, vidi_fake_vblank_handler);
 
-	subdrv = &ctx->subdrv;
-	subdrv->dev = &pdev->dev;
-	subdrv->probe = vidi_subdrv_probe;
+	mutex_init(&ctx->lock);
 
-	ret = exynos_drm_subdrv_register(subdrv);
-	if (ret < 0) {
-		dev_err(&pdev->dev, "failed to register drm vidi device\n");
-		return ret;
-	}
+	platform_set_drvdata(pdev, ctx);
 
 	ret = device_create_file(&pdev->dev, &dev_attr_connection);
 	if (ret < 0) {
-		exynos_drm_subdrv_unregister(subdrv);
-		DRM_INFO("failed to create connection sysfs.\n");
+		DRM_ERROR("failed to create connection sysfs.\n");
+		goto err_del_conn_component;
 	}
 
-	return 0;
+	ret = component_add(&pdev->dev, &vidi_component_ops);
+	if (ret)
+		goto err_remove_file;
+
+	return ret;
+
+err_remove_file:
+	device_remove_file(&pdev->dev, &dev_attr_connection);
+err_del_conn_component:
+	exynos_drm_component_del(&pdev->dev, EXYNOS_DEVICE_TYPE_CONNECTOR);
+err_del_crtc_component:
+	exynos_drm_component_del(&pdev->dev, EXYNOS_DEVICE_TYPE_CRTC);
+
+	return ret;
 }
 
 static int vidi_remove(struct platform_device *pdev)
 {
-	struct exynos_drm_manager *mgr = platform_get_drvdata(pdev);
-	struct vidi_context *ctx = mgr->ctx;
-	struct drm_encoder *encoder = ctx->encoder;
-	struct drm_crtc *crtc = mgr->crtc;
+	struct vidi_context *ctx = platform_get_drvdata(pdev);
 
 	if (ctx->raw_edid != (struct edid *)fake_edid_info) {
 		kfree(ctx->raw_edid);
@@ -657,9 +661,9 @@ static int vidi_remove(struct platform_device *pdev)
 		return -EINVAL;
 	}
 
-	crtc->funcs->destroy(crtc);
-	encoder->funcs->destroy(encoder);
-	drm_connector_cleanup(&ctx->connector);
+	component_del(&pdev->dev, &vidi_component_ops);
+	exynos_drm_component_del(&pdev->dev, EXYNOS_DEVICE_TYPE_CONNECTOR);
+	exynos_drm_component_del(&pdev->dev, EXYNOS_DEVICE_TYPE_CRTC);
 
 	return 0;
 }
@@ -691,12 +695,19 @@ int exynos_drm_probe_vidi(void)
 	return ret;
 }
 
+static int exynos_drm_remove_vidi_device(struct device *dev, void *data)
+{
+	platform_device_unregister(to_platform_device(dev));
+
+	return 0;
+}
+
 void exynos_drm_remove_vidi(void)
 {
-	struct vidi_context *ctx = vidi_manager.ctx;
-	struct exynos_drm_subdrv *subdrv = &ctx->subdrv;
-	struct platform_device *pdev = to_platform_device(subdrv->dev);
+	int ret = driver_for_each_device(&vidi_driver.driver, NULL, NULL,
+					 exynos_drm_remove_vidi_device);
+	/* silence compiler warning */
+	(void)ret;
 
 	platform_driver_unregister(&vidi_driver);
-	platform_device_unregister(pdev);
 }
diff --git a/drivers/gpu/drm/exynos/exynos_hdmi.c b/drivers/gpu/drm/exynos/exynos_hdmi.c
index 562966db2aa1..5765a161abdd 100644
--- a/drivers/gpu/drm/exynos/exynos_hdmi.c
+++ b/drivers/gpu/drm/exynos/exynos_hdmi.c
@@ -49,7 +49,6 @@
 #include <linux/gpio.h>
 #include <media/s5p_hdmi.h>
 
-#define get_hdmi_display(dev)	platform_get_drvdata(to_platform_device(dev))
 #define ctx_from_connector(c)	container_of(c, struct hdmi_context, connector)
 
 #define HOTPLUG_DEBOUNCE_MS		1100
@@ -182,6 +181,7 @@ struct hdmi_conf_regs {
 };
 
 struct hdmi_context {
+	struct exynos_drm_display	display;
 	struct device			*dev;
 	struct drm_device		*drm_dev;
 	struct drm_connector		connector;
@@ -213,6 +213,11 @@ struct hdmi_context {
 	enum hdmi_type			type;
 };
 
+static inline struct hdmi_context *display_to_hdmi(struct exynos_drm_display *d)
+{
+	return container_of(d, struct hdmi_context, display);
+}
+
 struct hdmiphy_config {
 	int pixel_clock;
 	u8 conf[32];
@@ -1040,6 +1045,8 @@ static enum drm_connector_status hdmi_detect(struct drm_connector *connector,
 
 static void hdmi_connector_destroy(struct drm_connector *connector)
 {
+	drm_connector_unregister(connector);
+	drm_connector_cleanup(connector);
 }
 
 static struct drm_connector_funcs hdmi_connector_funcs = {
@@ -1121,7 +1128,7 @@ static struct drm_connector_helper_funcs hdmi_connector_helper_funcs = {
 static int hdmi_create_connector(struct exynos_drm_display *display,
 			struct drm_encoder *encoder)
 {
-	struct hdmi_context *hdata = display->ctx;
+	struct hdmi_context *hdata = display_to_hdmi(display);
 	struct drm_connector *connector = &hdata->connector;
 	int ret;
 
@@ -1998,7 +2005,7 @@ static void hdmi_v14_mode_set(struct hdmi_context *hdata,
 static void hdmi_mode_set(struct exynos_drm_display *display,
 			struct drm_display_mode *mode)
 {
-	struct hdmi_context *hdata = display->ctx;
+	struct hdmi_context *hdata = display_to_hdmi(display);
 	struct drm_display_mode *m = mode;
 
 	DRM_DEBUG_KMS("xres=%d, yres=%d, refresh=%d, intl=%s\n",
@@ -2017,7 +2024,7 @@ static void hdmi_mode_set(struct exynos_drm_display *display,
 
 static void hdmi_commit(struct exynos_drm_display *display)
 {
-	struct hdmi_context *hdata = display->ctx;
+	struct hdmi_context *hdata = display_to_hdmi(display);
 
 	mutex_lock(&hdata->hdmi_mutex);
 	if (!hdata->powered) {
@@ -2031,7 +2038,7 @@ static void hdmi_commit(struct exynos_drm_display *display)
 
 static void hdmi_poweron(struct exynos_drm_display *display)
 {
-	struct hdmi_context *hdata = display->ctx;
+	struct hdmi_context *hdata = display_to_hdmi(display);
 	struct hdmi_resources *res = &hdata->res;
 
 	mutex_lock(&hdata->hdmi_mutex);
@@ -2062,7 +2069,7 @@ static void hdmi_poweron(struct exynos_drm_display *display)
 
 static void hdmi_poweroff(struct exynos_drm_display *display)
 {
-	struct hdmi_context *hdata = display->ctx;
+	struct hdmi_context *hdata = display_to_hdmi(display);
 	struct hdmi_resources *res = &hdata->res;
 
 	mutex_lock(&hdata->hdmi_mutex);
@@ -2097,7 +2104,7 @@ out:
 
 static void hdmi_dpms(struct exynos_drm_display *display, int mode)
 {
-	struct hdmi_context *hdata = display->ctx;
+	struct hdmi_context *hdata = display_to_hdmi(display);
 	struct drm_encoder *encoder = hdata->encoder;
 	struct drm_crtc *crtc = encoder->crtc;
 	struct drm_crtc_helper_funcs *funcs = NULL;
@@ -2141,11 +2148,6 @@ static struct exynos_drm_display_ops hdmi_display_ops = {
 	.commit		= hdmi_commit,
 };
 
-static struct exynos_drm_display hdmi_display = {
-	.type = EXYNOS_DISPLAY_TYPE_HDMI,
-	.ops = &hdmi_display_ops,
-};
-
 static void hdmi_hotplug_work_func(struct work_struct *work)
 {
 	struct hdmi_context *hdata;
@@ -2300,22 +2302,15 @@ MODULE_DEVICE_TABLE (of, hdmi_match_types);
 static int hdmi_bind(struct device *dev, struct device *master, void *data)
 {
 	struct drm_device *drm_dev = data;
-	struct hdmi_context *hdata;
+	struct hdmi_context *hdata = dev_get_drvdata(dev);
 
-	hdata = hdmi_display.ctx;
 	hdata->drm_dev = drm_dev;
 
-	return exynos_drm_create_enc_conn(drm_dev, &hdmi_display);
+	return exynos_drm_create_enc_conn(drm_dev, &hdata->display);
 }
 
 static void hdmi_unbind(struct device *dev, struct device *master, void *data)
 {
-	struct exynos_drm_display *display = get_hdmi_display(dev);
-	struct drm_encoder *encoder = display->encoder;
-	struct hdmi_context *hdata = display->ctx;
-
-	encoder->funcs->destroy(encoder);
-	drm_connector_cleanup(&hdata->connector);
 }
 
 static const struct component_ops hdmi_component_ops = {
@@ -2353,31 +2348,28 @@ static int hdmi_probe(struct platform_device *pdev)
 	struct resource *res;
 	int ret;
 
-	ret = exynos_drm_component_add(&pdev->dev, EXYNOS_DEVICE_TYPE_CONNECTOR,
-					hdmi_display.type);
-	if (ret)
-		return ret;
-
-	if (!dev->of_node) {
-		ret = -ENODEV;
-		goto err_del_component;
-	}
+	if (!dev->of_node)
+		return -ENODEV;
 
 	pdata = drm_hdmi_dt_parse_pdata(dev);
-	if (!pdata) {
-		ret = -EINVAL;
-		goto err_del_component;
-	}
+	if (!pdata)
+		return -EINVAL;
 
 	hdata = devm_kzalloc(dev, sizeof(struct hdmi_context), GFP_KERNEL);
-	if (!hdata) {
-		ret = -ENOMEM;
-		goto err_del_component;
-	}
+	if (!hdata)
+		return -ENOMEM;
+
+	hdata->display.type = EXYNOS_DISPLAY_TYPE_HDMI;
+	hdata->display.ops = &hdmi_display_ops;
+
+	ret = exynos_drm_component_add(&pdev->dev, EXYNOS_DEVICE_TYPE_CONNECTOR,
+					hdata->display.type);
+	if (ret)
+		return ret;
 
 	mutex_init(&hdata->hdmi_mutex);
 
-	platform_set_drvdata(pdev, &hdmi_display);
+	platform_set_drvdata(pdev, hdata);
 
 	match = of_match_node(hdmi_match_types, dev->of_node);
 	if (!match) {
@@ -2489,7 +2481,6 @@ out_get_phy_port:
 	}
 
 	pm_runtime_enable(dev);
-	hdmi_display.ctx = hdata;
 
 	ret = component_add(&pdev->dev, &hdmi_component_ops);
 	if (ret)
@@ -2514,7 +2505,7 @@ err_del_component:
 
 static int hdmi_remove(struct platform_device *pdev)
 {
-	struct hdmi_context *hdata = hdmi_display.ctx;
+	struct hdmi_context *hdata = platform_get_drvdata(pdev);
 
 	cancel_delayed_work_sync(&hdata->hotplug_work);
 
diff --git a/drivers/gpu/drm/exynos/exynos_mixer.c b/drivers/gpu/drm/exynos/exynos_mixer.c
index e8b4ec84b312..820b76234ef4 100644
--- a/drivers/gpu/drm/exynos/exynos_mixer.c
+++ b/drivers/gpu/drm/exynos/exynos_mixer.c
@@ -40,8 +40,6 @@
 #include "exynos_drm_iommu.h"
 #include "exynos_mixer.h"
 
-#define get_mixer_manager(dev)	platform_get_drvdata(to_platform_device(dev))
-
 #define MIXER_WIN_NR		3
 #define MIXER_DEFAULT_WIN	0
 
@@ -86,6 +84,7 @@ enum mixer_version_id {
 };
 
 struct mixer_context {
+	struct exynos_drm_manager manager;
 	struct platform_device *pdev;
 	struct device		*dev;
 	struct drm_device	*drm_dev;
@@ -104,6 +103,11 @@ struct mixer_context {
 	atomic_t		wait_vsync_event;
 };
 
+static inline struct mixer_context *mgr_to_mixer(struct exynos_drm_manager *mgr)
+{
+	return container_of(mgr, struct mixer_context, manager);
+}
+
 struct mixer_drv_data {
 	enum mixer_version_id	version;
 	bool					is_vp_enabled;
@@ -854,7 +858,7 @@ static int mixer_initialize(struct exynos_drm_manager *mgr,
 			struct drm_device *drm_dev)
 {
 	int ret;
-	struct mixer_context *mixer_ctx = mgr->ctx;
+	struct mixer_context *mixer_ctx = mgr_to_mixer(mgr);
 	struct exynos_drm_private *priv;
 	priv = drm_dev->dev_private;
 
@@ -885,7 +889,7 @@ static int mixer_initialize(struct exynos_drm_manager *mgr,
 
 static void mixer_mgr_remove(struct exynos_drm_manager *mgr)
 {
-	struct mixer_context *mixer_ctx = mgr->ctx;
+	struct mixer_context *mixer_ctx = mgr_to_mixer(mgr);
 
 	if (is_drm_iommu_supported(mixer_ctx->drm_dev))
 		drm_iommu_detach_device(mixer_ctx->drm_dev, mixer_ctx->dev);
@@ -893,7 +897,7 @@ static void mixer_mgr_remove(struct exynos_drm_manager *mgr)
 
 static int mixer_enable_vblank(struct exynos_drm_manager *mgr)
 {
-	struct mixer_context *mixer_ctx = mgr->ctx;
+	struct mixer_context *mixer_ctx = mgr_to_mixer(mgr);
 	struct mixer_resources *res = &mixer_ctx->mixer_res;
 
 	if (!mixer_ctx->powered) {
@@ -910,7 +914,7 @@ static int mixer_enable_vblank(struct exynos_drm_manager *mgr)
 
 static void mixer_disable_vblank(struct exynos_drm_manager *mgr)
 {
-	struct mixer_context *mixer_ctx = mgr->ctx;
+	struct mixer_context *mixer_ctx = mgr_to_mixer(mgr);
 	struct mixer_resources *res = &mixer_ctx->mixer_res;
 
 	/* disable vsync interrupt */
@@ -920,7 +924,7 @@ static void mixer_disable_vblank(struct exynos_drm_manager *mgr)
 static void mixer_win_mode_set(struct exynos_drm_manager *mgr,
 			struct exynos_drm_overlay *overlay)
 {
-	struct mixer_context *mixer_ctx = mgr->ctx;
+	struct mixer_context *mixer_ctx = mgr_to_mixer(mgr);
 	struct hdmi_win_data *win_data;
 	int win;
 
@@ -971,7 +975,7 @@ static void mixer_win_mode_set(struct exynos_drm_manager *mgr,
 
 static void mixer_win_commit(struct exynos_drm_manager *mgr, int zpos)
 {
-	struct mixer_context *mixer_ctx = mgr->ctx;
+	struct mixer_context *mixer_ctx = mgr_to_mixer(mgr);
 	int win = zpos == DEFAULT_ZPOS ? MIXER_DEFAULT_WIN : zpos;
 
 	DRM_DEBUG_KMS("win: %d\n", win);
@@ -993,7 +997,7 @@ static void mixer_win_commit(struct exynos_drm_manager *mgr, int zpos)
 
 static void mixer_win_disable(struct exynos_drm_manager *mgr, int zpos)
 {
-	struct mixer_context *mixer_ctx = mgr->ctx;
+	struct mixer_context *mixer_ctx = mgr_to_mixer(mgr);
 	struct mixer_resources *res = &mixer_ctx->mixer_res;
 	int win = zpos == DEFAULT_ZPOS ? MIXER_DEFAULT_WIN : zpos;
 	unsigned long flags;
@@ -1021,7 +1025,7 @@ static void mixer_win_disable(struct exynos_drm_manager *mgr, int zpos)
 
 static void mixer_wait_for_vblank(struct exynos_drm_manager *mgr)
 {
-	struct mixer_context *mixer_ctx = mgr->ctx;
+	struct mixer_context *mixer_ctx = mgr_to_mixer(mgr);
 
 	mutex_lock(&mixer_ctx->mixer_mutex);
 	if (!mixer_ctx->powered) {
@@ -1048,7 +1052,7 @@ static void mixer_wait_for_vblank(struct exynos_drm_manager *mgr)
 
 static void mixer_window_suspend(struct exynos_drm_manager *mgr)
 {
-	struct mixer_context *ctx = mgr->ctx;
+	struct mixer_context *ctx = mgr_to_mixer(mgr);
 	struct hdmi_win_data *win_data;
 	int i;
 
@@ -1062,7 +1066,7 @@ static void mixer_window_suspend(struct exynos_drm_manager *mgr)
 
 static void mixer_window_resume(struct exynos_drm_manager *mgr)
 {
-	struct mixer_context *ctx = mgr->ctx;
+	struct mixer_context *ctx = mgr_to_mixer(mgr);
 	struct hdmi_win_data *win_data;
 	int i;
 
@@ -1077,7 +1081,7 @@ static void mixer_window_resume(struct exynos_drm_manager *mgr)
 
 static void mixer_poweron(struct exynos_drm_manager *mgr)
 {
-	struct mixer_context *ctx = mgr->ctx;
+	struct mixer_context *ctx = mgr_to_mixer(mgr);
 	struct mixer_resources *res = &ctx->mixer_res;
 
 	mutex_lock(&ctx->mixer_mutex);
@@ -1111,7 +1115,7 @@ static void mixer_poweron(struct exynos_drm_manager *mgr)
 
 static void mixer_poweroff(struct exynos_drm_manager *mgr)
 {
-	struct mixer_context *ctx = mgr->ctx;
+	struct mixer_context *ctx = mgr_to_mixer(mgr);
 	struct mixer_resources *res = &ctx->mixer_res;
 
 	mutex_lock(&ctx->mixer_mutex);
@@ -1187,11 +1191,6 @@ static struct exynos_drm_manager_ops mixer_manager_ops = {
 	.win_disable		= mixer_win_disable,
 };
 
-static struct exynos_drm_manager mixer_manager = {
-	.type			= EXYNOS_DISPLAY_TYPE_HDMI,
-	.ops			= &mixer_manager_ops,
-};
-
 static struct mixer_drv_data exynos5420_mxr_drv_data = {
 	.version = MXR_VER_128_0_0_184,
 	.is_vp_enabled = 0,
@@ -1249,48 +1248,17 @@ MODULE_DEVICE_TABLE(of, mixer_match_types);
 
 static int mixer_bind(struct device *dev, struct device *manager, void *data)
 {
-	struct platform_device *pdev = to_platform_device(dev);
+	struct mixer_context *ctx = dev_get_drvdata(dev);
 	struct drm_device *drm_dev = data;
-	struct mixer_context *ctx;
-	struct mixer_drv_data *drv;
 	int ret;
 
-	dev_info(dev, "probe start\n");
-
-	ctx = devm_kzalloc(&pdev->dev, sizeof(*ctx), GFP_KERNEL);
-	if (!ctx) {
-		DRM_ERROR("failed to alloc mixer context.\n");
-		return -ENOMEM;
-	}
-
-	mutex_init(&ctx->mixer_mutex);
-
-	if (dev->of_node) {
-		const struct of_device_id *match;
-		match = of_match_node(mixer_match_types, dev->of_node);
-		drv = (struct mixer_drv_data *)match->data;
-	} else {
-		drv = (struct mixer_drv_data *)
-			platform_get_device_id(pdev)->driver_data;
-	}
-
-	ctx->pdev = pdev;
-	ctx->dev = dev;
-	ctx->vp_enabled = drv->is_vp_enabled;
-	ctx->has_sclk = drv->has_sclk;
-	ctx->mxr_ver = drv->version;
-	init_waitqueue_head(&ctx->wait_vsync_queue);
-	atomic_set(&ctx->wait_vsync_event, 0);
-
-	mixer_manager.ctx = ctx;
-	ret = mixer_initialize(&mixer_manager, drm_dev);
+	ret = mixer_initialize(&ctx->manager, drm_dev);
 	if (ret)
 		return ret;
 
-	platform_set_drvdata(pdev, &mixer_manager);
-	ret = exynos_drm_crtc_create(&mixer_manager);
+	ret = exynos_drm_crtc_create(&ctx->manager);
 	if (ret) {
-		mixer_mgr_remove(&mixer_manager);
+		mixer_mgr_remove(&ctx->manager);
 		return ret;
 	}
 
@@ -1301,16 +1269,11 @@ static int mixer_bind(struct device *dev, struct device *manager, void *data)
 
 static void mixer_unbind(struct device *dev, struct device *master, void *data)
 {
-	struct exynos_drm_manager *mgr = dev_get_drvdata(dev);
-	struct drm_crtc *crtc = mgr->crtc;
-
-	dev_info(dev, "remove successful\n");
+	struct mixer_context *ctx = dev_get_drvdata(dev);
 
-	mixer_mgr_remove(mgr);
+	mixer_mgr_remove(&ctx->manager);
 
 	pm_runtime_disable(dev);
-
-	crtc->funcs->destroy(crtc);
 }
 
 static const struct component_ops mixer_component_ops = {
@@ -1320,22 +1283,62 @@ static const struct component_ops mixer_component_ops = {
 
 static int mixer_probe(struct platform_device *pdev)
 {
+	struct device *dev = &pdev->dev;
+	struct mixer_drv_data *drv;
+	struct mixer_context *ctx;
 	int ret;
 
+	ctx = devm_kzalloc(&pdev->dev, sizeof(*ctx), GFP_KERNEL);
+	if (!ctx) {
+		DRM_ERROR("failed to alloc mixer context.\n");
+		return -ENOMEM;
+	}
+
+	mutex_init(&ctx->mixer_mutex);
+
+	ctx->manager.type = EXYNOS_DISPLAY_TYPE_HDMI;
+	ctx->manager.ops = &mixer_manager_ops;
+
+	if (dev->of_node) {
+		const struct of_device_id *match;
+
+		match = of_match_node(mixer_match_types, dev->of_node);
+		drv = (struct mixer_drv_data *)match->data;
+	} else {
+		drv = (struct mixer_drv_data *)
+			platform_get_device_id(pdev)->driver_data;
+	}
+
+	ctx->pdev = pdev;
+	ctx->dev = dev;
+	ctx->vp_enabled = drv->is_vp_enabled;
+	ctx->has_sclk = drv->has_sclk;
+	ctx->mxr_ver = drv->version;
+	init_waitqueue_head(&ctx->wait_vsync_queue);
+	atomic_set(&ctx->wait_vsync_event, 0);
+
+	platform_set_drvdata(pdev, ctx);
+
 	ret = exynos_drm_component_add(&pdev->dev, EXYNOS_DEVICE_TYPE_CRTC,
-					mixer_manager.type);
+					ctx->manager.type);
 	if (ret)
 		return ret;
 
 	ret = component_add(&pdev->dev, &mixer_component_ops);
-	if (ret)
+	if (ret) {
 		exynos_drm_component_del(&pdev->dev, EXYNOS_DEVICE_TYPE_CRTC);
+		return ret;
+	}
+
+	pm_runtime_enable(dev);
 
 	return ret;
 }
 
 static int mixer_remove(struct platform_device *pdev)
 {
+	pm_runtime_disable(&pdev->dev);
+
 	component_del(&pdev->dev, &mixer_component_ops);
 	exynos_drm_component_del(&pdev->dev, EXYNOS_DEVICE_TYPE_CRTC);
 
diff --git a/drivers/gpu/drm/gma500/Makefile b/drivers/gpu/drm/gma500/Makefile
index b15315576376..190e55f2f891 100644
--- a/drivers/gpu/drm/gma500/Makefile
+++ b/drivers/gpu/drm/gma500/Makefile
@@ -39,6 +39,7 @@ gma500_gfx-$(CONFIG_DRM_GMA3600) +=  cdv_device.o \
 gma500_gfx-$(CONFIG_DRM_GMA600) += oaktrail_device.o \
 	  oaktrail_crtc.o \
 	  oaktrail_lvds.o \
+	  oaktrail_lvds_i2c.o \
 	  oaktrail_hdmi.o \
 	  oaktrail_hdmi_i2c.o
 
diff --git a/drivers/gpu/drm/gma500/cdv_intel_dp.c b/drivers/gpu/drm/gma500/cdv_intel_dp.c
index a4cc0e60a1be..0fafb8e2483a 100644
--- a/drivers/gpu/drm/gma500/cdv_intel_dp.c
+++ b/drivers/gpu/drm/gma500/cdv_intel_dp.c
@@ -37,6 +37,201 @@
 #include "gma_display.h"
 #include <drm/drm_dp_helper.h>
 
+/**
+ * struct i2c_algo_dp_aux_data - driver interface structure for i2c over dp
+ * 				 aux algorithm
+ * @running: set by the algo indicating whether an i2c is ongoing or whether
+ * 	     the i2c bus is quiescent
+ * @address: i2c target address for the currently ongoing transfer
+ * @aux_ch: driver callback to transfer a single byte of the i2c payload
+ */
+struct i2c_algo_dp_aux_data {
+	bool running;
+	u16 address;
+	int (*aux_ch) (struct i2c_adapter *adapter,
+		       int mode, uint8_t write_byte,
+		       uint8_t *read_byte);
+};
+
+/* Run a single AUX_CH I2C transaction, writing/reading data as necessary */
+static int
+i2c_algo_dp_aux_transaction(struct i2c_adapter *adapter, int mode,
+			    uint8_t write_byte, uint8_t *read_byte)
+{
+	struct i2c_algo_dp_aux_data *algo_data = adapter->algo_data;
+	int ret;
+
+	ret = (*algo_data->aux_ch)(adapter, mode,
+				   write_byte, read_byte);
+	return ret;
+}
+
+/*
+ * I2C over AUX CH
+ */
+
+/*
+ * Send the address. If the I2C link is running, this 'restarts'
+ * the connection with the new address, this is used for doing
+ * a write followed by a read (as needed for DDC)
+ */
+static int
+i2c_algo_dp_aux_address(struct i2c_adapter *adapter, u16 address, bool reading)
+{
+	struct i2c_algo_dp_aux_data *algo_data = adapter->algo_data;
+	int mode = MODE_I2C_START;
+	int ret;
+
+	if (reading)
+		mode |= MODE_I2C_READ;
+	else
+		mode |= MODE_I2C_WRITE;
+	algo_data->address = address;
+	algo_data->running = true;
+	ret = i2c_algo_dp_aux_transaction(adapter, mode, 0, NULL);
+	return ret;
+}
+
+/*
+ * Stop the I2C transaction. This closes out the link, sending
+ * a bare address packet with the MOT bit turned off
+ */
+static void
+i2c_algo_dp_aux_stop(struct i2c_adapter *adapter, bool reading)
+{
+	struct i2c_algo_dp_aux_data *algo_data = adapter->algo_data;
+	int mode = MODE_I2C_STOP;
+
+	if (reading)
+		mode |= MODE_I2C_READ;
+	else
+		mode |= MODE_I2C_WRITE;
+	if (algo_data->running) {
+		(void) i2c_algo_dp_aux_transaction(adapter, mode, 0, NULL);
+		algo_data->running = false;
+	}
+}
+
+/*
+ * Write a single byte to the current I2C address, the
+ * the I2C link must be running or this returns -EIO
+ */
+static int
+i2c_algo_dp_aux_put_byte(struct i2c_adapter *adapter, u8 byte)
+{
+	struct i2c_algo_dp_aux_data *algo_data = adapter->algo_data;
+	int ret;
+
+	if (!algo_data->running)
+		return -EIO;
+
+	ret = i2c_algo_dp_aux_transaction(adapter, MODE_I2C_WRITE, byte, NULL);
+	return ret;
+}
+
+/*
+ * Read a single byte from the current I2C address, the
+ * I2C link must be running or this returns -EIO
+ */
+static int
+i2c_algo_dp_aux_get_byte(struct i2c_adapter *adapter, u8 *byte_ret)
+{
+	struct i2c_algo_dp_aux_data *algo_data = adapter->algo_data;
+	int ret;
+
+	if (!algo_data->running)
+		return -EIO;
+
+	ret = i2c_algo_dp_aux_transaction(adapter, MODE_I2C_READ, 0, byte_ret);
+	return ret;
+}
+
+static int
+i2c_algo_dp_aux_xfer(struct i2c_adapter *adapter,
+		     struct i2c_msg *msgs,
+		     int num)
+{
+	int ret = 0;
+	bool reading = false;
+	int m;
+	int b;
+
+	for (m = 0; m < num; m++) {
+		u16 len = msgs[m].len;
+		u8 *buf = msgs[m].buf;
+		reading = (msgs[m].flags & I2C_M_RD) != 0;
+		ret = i2c_algo_dp_aux_address(adapter, msgs[m].addr, reading);
+		if (ret < 0)
+			break;
+		if (reading) {
+			for (b = 0; b < len; b++) {
+				ret = i2c_algo_dp_aux_get_byte(adapter, &buf[b]);
+				if (ret < 0)
+					break;
+			}
+		} else {
+			for (b = 0; b < len; b++) {
+				ret = i2c_algo_dp_aux_put_byte(adapter, buf[b]);
+				if (ret < 0)
+					break;
+			}
+		}
+		if (ret < 0)
+			break;
+	}
+	if (ret >= 0)
+		ret = num;
+	i2c_algo_dp_aux_stop(adapter, reading);
+	DRM_DEBUG_KMS("dp_aux_xfer return %d\n", ret);
+	return ret;
+}
+
+static u32
+i2c_algo_dp_aux_functionality(struct i2c_adapter *adapter)
+{
+	return I2C_FUNC_I2C | I2C_FUNC_SMBUS_EMUL |
+	       I2C_FUNC_SMBUS_READ_BLOCK_DATA |
+	       I2C_FUNC_SMBUS_BLOCK_PROC_CALL |
+	       I2C_FUNC_10BIT_ADDR;
+}
+
+static const struct i2c_algorithm i2c_dp_aux_algo = {
+	.master_xfer	= i2c_algo_dp_aux_xfer,
+	.functionality	= i2c_algo_dp_aux_functionality,
+};
+
+static void
+i2c_dp_aux_reset_bus(struct i2c_adapter *adapter)
+{
+	(void) i2c_algo_dp_aux_address(adapter, 0, false);
+	(void) i2c_algo_dp_aux_stop(adapter, false);
+}
+
+static int
+i2c_dp_aux_prepare_bus(struct i2c_adapter *adapter)
+{
+	adapter->algo = &i2c_dp_aux_algo;
+	adapter->retries = 3;
+	i2c_dp_aux_reset_bus(adapter);
+	return 0;
+}
+
+/*
+ * FIXME: This is the old dp aux helper, gma500 is the last driver that needs to
+ * be ported over to the new helper code in drm_dp_helper.c like i915 or radeon.
+ */
+static int __deprecated
+i2c_dp_aux_add_bus(struct i2c_adapter *adapter)
+{
+	int error;
+
+	error = i2c_dp_aux_prepare_bus(adapter);
+	if (error)
+		return error;
+	error = i2c_add_adapter(adapter);
+	return error;
+}
+
 #define _wait_for(COND, MS, W) ({ \
         unsigned long timeout__ = jiffies + msecs_to_jiffies(MS);       \
         int ret__ = 0;                                                  \
@@ -1089,7 +1284,7 @@ static char	*link_train_names[] = {
 };
 #endif
 
-#define CDV_DP_VOLTAGE_MAX	    DP_TRAIN_VOLTAGE_SWING_1200
+#define CDV_DP_VOLTAGE_MAX	    DP_TRAIN_VOLTAGE_SWING_LEVEL_3
 /*
 static uint8_t
 cdv_intel_dp_pre_emphasis_max(uint8_t voltage_swing)
@@ -1276,7 +1471,7 @@ cdv_intel_dp_set_vswing_premph(struct gma_encoder *encoder, uint8_t signal_level
 		cdv_sb_write(dev, ddi_reg->VSwing2, dp_vswing_premph_table[index]);
 
 	/* ;gfx_dpio_set_reg(0x814c, 0x40802040) */
-	if ((vswing + premph) == DP_TRAIN_VOLTAGE_SWING_1200)
+	if ((vswing + premph) == DP_TRAIN_VOLTAGE_SWING_LEVEL_3)
 		cdv_sb_write(dev, ddi_reg->VSwing3, 0x70802040);
 	else
 		cdv_sb_write(dev, ddi_reg->VSwing3, 0x40802040);
diff --git a/drivers/gpu/drm/gma500/framebuffer.c b/drivers/gpu/drm/gma500/framebuffer.c
index d0dd3bea8aa5..ddd90ddbc200 100644
--- a/drivers/gpu/drm/gma500/framebuffer.c
+++ b/drivers/gpu/drm/gma500/framebuffer.c
@@ -540,7 +540,8 @@ static void psbfb_gamma_get(struct drm_crtc *crtc, u16 *red,
 static int psbfb_probe(struct drm_fb_helper *helper,
 				struct drm_fb_helper_surface_size *sizes)
 {
-	struct psb_fbdev *psb_fbdev = (struct psb_fbdev *)helper;
+	struct psb_fbdev *psb_fbdev =
+		container_of(helper, struct psb_fbdev, psb_fb_helper);
 	struct drm_device *dev = psb_fbdev->psb_fb_helper.dev;
 	struct drm_psb_private *dev_priv = dev->dev_private;
 	int bytespp;
diff --git a/drivers/gpu/drm/gma500/gtt.h b/drivers/gpu/drm/gma500/gtt.h
index f5860a739bd8..cdbb350c9d5d 100644
--- a/drivers/gpu/drm/gma500/gtt.h
+++ b/drivers/gpu/drm/gma500/gtt.h
@@ -21,6 +21,7 @@
 #define _PSB_GTT_H_
 
 #include <drm/drmP.h>
+#include <drm/drm_gem.h>
 
 /* This wants cleaning up with respect to the psb_dev and un-needed stuff */
 struct psb_gtt {
diff --git a/drivers/gpu/drm/gma500/intel_bios.c b/drivers/gpu/drm/gma500/intel_bios.c
index d3497348c4d5..63bde4e86c6a 100644
--- a/drivers/gpu/drm/gma500/intel_bios.c
+++ b/drivers/gpu/drm/gma500/intel_bios.c
@@ -116,30 +116,30 @@ parse_edp(struct drm_psb_private *dev_priv, struct bdb_header *bdb)
 
 	switch (edp_link_params->preemphasis) {
 	case 0:
-		dev_priv->edp.preemphasis = DP_TRAIN_PRE_EMPHASIS_0;
+		dev_priv->edp.preemphasis = DP_TRAIN_PRE_EMPH_LEVEL_0;
 		break;
 	case 1:
-		dev_priv->edp.preemphasis = DP_TRAIN_PRE_EMPHASIS_3_5;
+		dev_priv->edp.preemphasis = DP_TRAIN_PRE_EMPH_LEVEL_1;
 		break;
 	case 2:
-		dev_priv->edp.preemphasis = DP_TRAIN_PRE_EMPHASIS_6;
+		dev_priv->edp.preemphasis = DP_TRAIN_PRE_EMPH_LEVEL_2;
 		break;
 	case 3:
-		dev_priv->edp.preemphasis = DP_TRAIN_PRE_EMPHASIS_9_5;
+		dev_priv->edp.preemphasis = DP_TRAIN_PRE_EMPH_LEVEL_3;
 		break;
 	}
 	switch (edp_link_params->vswing) {
 	case 0:
-		dev_priv->edp.vswing = DP_TRAIN_VOLTAGE_SWING_400;
+		dev_priv->edp.vswing = DP_TRAIN_VOLTAGE_SWING_LEVEL_0;
 		break;
 	case 1:
-		dev_priv->edp.vswing = DP_TRAIN_VOLTAGE_SWING_600;
+		dev_priv->edp.vswing = DP_TRAIN_VOLTAGE_SWING_LEVEL_1;
 		break;
 	case 2:
-		dev_priv->edp.vswing = DP_TRAIN_VOLTAGE_SWING_800;
+		dev_priv->edp.vswing = DP_TRAIN_VOLTAGE_SWING_LEVEL_2;
 		break;
 	case 3:
-		dev_priv->edp.vswing = DP_TRAIN_VOLTAGE_SWING_1200;
+		dev_priv->edp.vswing = DP_TRAIN_VOLTAGE_SWING_LEVEL_3;
 		break;
 	}
 	DRM_DEBUG_KMS("VBT reports EDP: VSwing  %d, Preemph %d\n",
diff --git a/drivers/gpu/drm/gma500/mdfld_dsi_pkg_sender.c b/drivers/gpu/drm/gma500/mdfld_dsi_pkg_sender.c
index 87885d8c06e8..6b43ae3ffd73 100644
--- a/drivers/gpu/drm/gma500/mdfld_dsi_pkg_sender.c
+++ b/drivers/gpu/drm/gma500/mdfld_dsi_pkg_sender.c
@@ -25,6 +25,7 @@
  */
 
 #include <linux/freezer.h>
+#include <video/mipi_display.h>
 
 #include "mdfld_dsi_output.h"
 #include "mdfld_dsi_pkg_sender.h"
@@ -32,20 +33,6 @@
 
 #define MDFLD_DSI_READ_MAX_COUNT		5000
 
-enum data_type {
-	DSI_DT_GENERIC_SHORT_WRITE_0	= 0x03,
-	DSI_DT_GENERIC_SHORT_WRITE_1	= 0x13,
-	DSI_DT_GENERIC_SHORT_WRITE_2	= 0x23,
-	DSI_DT_GENERIC_READ_0		= 0x04,
-	DSI_DT_GENERIC_READ_1		= 0x14,
-	DSI_DT_GENERIC_READ_2		= 0x24,
-	DSI_DT_GENERIC_LONG_WRITE	= 0x29,
-	DSI_DT_DCS_SHORT_WRITE_0	= 0x05,
-	DSI_DT_DCS_SHORT_WRITE_1	= 0x15,
-	DSI_DT_DCS_READ			= 0x06,
-	DSI_DT_DCS_LONG_WRITE		= 0x39,
-};
-
 enum {
 	MDFLD_DSI_PANEL_MODE_SLEEP = 0x1,
 };
@@ -321,9 +308,9 @@ static int send_pkg_prepare(struct mdfld_dsi_pkg_sender *sender, u8 data_type,
 	u8 cmd;
 
 	switch (data_type) {
-	case DSI_DT_DCS_SHORT_WRITE_0:
-	case DSI_DT_DCS_SHORT_WRITE_1:
-	case DSI_DT_DCS_LONG_WRITE:
+	case MIPI_DSI_DCS_SHORT_WRITE:
+	case MIPI_DSI_DCS_SHORT_WRITE_PARAM:
+	case MIPI_DSI_DCS_LONG_WRITE:
 		cmd = *data;
 		break;
 	default:
@@ -334,12 +321,12 @@ static int send_pkg_prepare(struct mdfld_dsi_pkg_sender *sender, u8 data_type,
 	sender->status = MDFLD_DSI_PKG_SENDER_BUSY;
 
 	/*wait for 120 milliseconds in case exit_sleep_mode just be sent*/
-	if (unlikely(cmd == DCS_ENTER_SLEEP_MODE)) {
+	if (unlikely(cmd == MIPI_DCS_ENTER_SLEEP_MODE)) {
 		/*TODO: replace it with msleep later*/
 		mdelay(120);
 	}
 
-	if (unlikely(cmd == DCS_EXIT_SLEEP_MODE)) {
+	if (unlikely(cmd == MIPI_DCS_EXIT_SLEEP_MODE)) {
 		/*TODO: replace it with msleep later*/
 		mdelay(120);
 	}
@@ -352,9 +339,9 @@ static int send_pkg_done(struct mdfld_dsi_pkg_sender *sender, u8 data_type,
 	u8 cmd;
 
 	switch (data_type) {
-	case DSI_DT_DCS_SHORT_WRITE_0:
-	case DSI_DT_DCS_SHORT_WRITE_1:
-	case DSI_DT_DCS_LONG_WRITE:
+	case MIPI_DSI_DCS_SHORT_WRITE:
+	case MIPI_DSI_DCS_SHORT_WRITE_PARAM:
+	case MIPI_DSI_DCS_LONG_WRITE:
 		cmd = *data;
 		break;
 	default:
@@ -362,15 +349,15 @@ static int send_pkg_done(struct mdfld_dsi_pkg_sender *sender, u8 data_type,
 	}
 
 	/*update panel status*/
-	if (unlikely(cmd == DCS_ENTER_SLEEP_MODE)) {
+	if (unlikely(cmd == MIPI_DCS_ENTER_SLEEP_MODE)) {
 		sender->panel_mode |= MDFLD_DSI_PANEL_MODE_SLEEP;
 		/*TODO: replace it with msleep later*/
 		mdelay(120);
-	} else if (unlikely(cmd == DCS_EXIT_SLEEP_MODE)) {
+	} else if (unlikely(cmd == MIPI_DCS_EXIT_SLEEP_MODE)) {
 		sender->panel_mode &= ~MDFLD_DSI_PANEL_MODE_SLEEP;
 		/*TODO: replace it with msleep later*/
 		mdelay(120);
-	} else if (unlikely(cmd == DCS_SOFT_RESET)) {
+	} else if (unlikely(cmd == MIPI_DCS_SOFT_RESET)) {
 		/*TODO: replace it with msleep later*/
 		mdelay(5);
 	}
@@ -405,19 +392,19 @@ static int send_pkg(struct mdfld_dsi_pkg_sender *sender, u8 data_type,
 	}
 
 	switch (data_type) {
-	case DSI_DT_GENERIC_SHORT_WRITE_0:
-	case DSI_DT_GENERIC_SHORT_WRITE_1:
-	case DSI_DT_GENERIC_SHORT_WRITE_2:
-	case DSI_DT_GENERIC_READ_0:
-	case DSI_DT_GENERIC_READ_1:
-	case DSI_DT_GENERIC_READ_2:
-	case DSI_DT_DCS_SHORT_WRITE_0:
-	case DSI_DT_DCS_SHORT_WRITE_1:
-	case DSI_DT_DCS_READ:
+	case MIPI_DSI_GENERIC_SHORT_WRITE_0_PARAM:
+	case MIPI_DSI_GENERIC_SHORT_WRITE_1_PARAM:
+	case MIPI_DSI_GENERIC_SHORT_WRITE_2_PARAM:
+	case MIPI_DSI_GENERIC_READ_REQUEST_0_PARAM:
+	case MIPI_DSI_GENERIC_READ_REQUEST_1_PARAM:
+	case MIPI_DSI_GENERIC_READ_REQUEST_2_PARAM:
+	case MIPI_DSI_DCS_SHORT_WRITE:
+	case MIPI_DSI_DCS_SHORT_WRITE_PARAM:
+	case MIPI_DSI_DCS_READ:
 		ret = send_short_pkg(sender, data_type, data[0], data[1], hs);
 		break;
-	case DSI_DT_GENERIC_LONG_WRITE:
-	case DSI_DT_DCS_LONG_WRITE:
+	case MIPI_DSI_GENERIC_LONG_WRITE:
+	case MIPI_DSI_DCS_LONG_WRITE:
 		ret = send_long_pkg(sender, data_type, data, len, hs);
 		break;
 	}
@@ -440,7 +427,7 @@ int mdfld_dsi_send_mcs_long(struct mdfld_dsi_pkg_sender *sender, u8 *data,
 	}
 
 	spin_lock_irqsave(&sender->lock, flags);
-	send_pkg(sender, DSI_DT_DCS_LONG_WRITE, data, len, hs);
+	send_pkg(sender, MIPI_DSI_DCS_LONG_WRITE, data, len, hs);
 	spin_unlock_irqrestore(&sender->lock, flags);
 
 	return 0;
@@ -461,10 +448,10 @@ int mdfld_dsi_send_mcs_short(struct mdfld_dsi_pkg_sender *sender, u8 cmd,
 	data[0] = cmd;
 
 	if (param_num) {
-		data_type = DSI_DT_DCS_SHORT_WRITE_1;
+		data_type = MIPI_DSI_DCS_SHORT_WRITE_PARAM;
 		data[1] = param;
 	} else {
-		data_type = DSI_DT_DCS_SHORT_WRITE_0;
+		data_type = MIPI_DSI_DCS_SHORT_WRITE;
 		data[1] = 0;
 	}
 
@@ -489,17 +476,17 @@ int mdfld_dsi_send_gen_short(struct mdfld_dsi_pkg_sender *sender, u8 param0,
 
 	switch (param_num) {
 	case 0:
-		data_type = DSI_DT_GENERIC_SHORT_WRITE_0;
+		data_type = MIPI_DSI_GENERIC_SHORT_WRITE_0_PARAM;
 		data[0] = 0;
 		data[1] = 0;
 		break;
 	case 1:
-		data_type = DSI_DT_GENERIC_SHORT_WRITE_1;
+		data_type = MIPI_DSI_GENERIC_SHORT_WRITE_1_PARAM;
 		data[0] = param0;
 		data[1] = 0;
 		break;
 	case 2:
-		data_type = DSI_DT_GENERIC_SHORT_WRITE_2;
+		data_type = MIPI_DSI_GENERIC_SHORT_WRITE_2_PARAM;
 		data[0] = param0;
 		data[1] = param1;
 		break;
@@ -523,7 +510,7 @@ int mdfld_dsi_send_gen_long(struct mdfld_dsi_pkg_sender *sender, u8 *data,
 	}
 
 	spin_lock_irqsave(&sender->lock, flags);
-	send_pkg(sender, DSI_DT_GENERIC_LONG_WRITE, data, len, hs);
+	send_pkg(sender, MIPI_DSI_GENERIC_LONG_WRITE, data, len, hs);
 	spin_unlock_irqrestore(&sender->lock, flags);
 
 	return 0;
@@ -594,7 +581,7 @@ int mdfld_dsi_read_mcs(struct mdfld_dsi_pkg_sender *sender, u8 cmd,
 		return -EINVAL;
 	}
 
-	return __read_panel_data(sender, DSI_DT_DCS_READ, &cmd, 1,
+	return __read_panel_data(sender, MIPI_DSI_DCS_READ, &cmd, 1,
 				data, len, hs);
 }
 
diff --git a/drivers/gpu/drm/gma500/mdfld_dsi_pkg_sender.h b/drivers/gpu/drm/gma500/mdfld_dsi_pkg_sender.h
index 459cd7ea8b81..0478a21c15d5 100644
--- a/drivers/gpu/drm/gma500/mdfld_dsi_pkg_sender.h
+++ b/drivers/gpu/drm/gma500/mdfld_dsi_pkg_sender.h
@@ -62,18 +62,6 @@ struct mdfld_dsi_pkg_sender {
 	u32 mipi_cmd_len_reg;
 };
 
-/* DCS definitions */
-#define DCS_SOFT_RESET			0x01
-#define DCS_ENTER_SLEEP_MODE		0x10
-#define DCS_EXIT_SLEEP_MODE		0x11
-#define DCS_SET_DISPLAY_OFF		0x28
-#define DCS_SET_DISPLAY_ON		0x29
-#define DCS_SET_COLUMN_ADDRESS		0x2a
-#define DCS_SET_PAGE_ADDRESS		0x2b
-#define DCS_WRITE_MEM_START		0x2c
-#define DCS_SET_TEAR_OFF		0x34
-#define DCS_SET_TEAR_ON			0x35
-
 extern int mdfld_dsi_pkg_sender_init(struct mdfld_dsi_connector *dsi_connector,
 					int pipe);
 extern void mdfld_dsi_pkg_sender_destroy(struct mdfld_dsi_pkg_sender *sender);
diff --git a/drivers/gpu/drm/gma500/oaktrail_lvds.c b/drivers/gpu/drm/gma500/oaktrail_lvds.c
index 0d39da6e8b7a..83bbc271bcfb 100644
--- a/drivers/gpu/drm/gma500/oaktrail_lvds.c
+++ b/drivers/gpu/drm/gma500/oaktrail_lvds.c
@@ -359,22 +359,26 @@ void oaktrail_lvds_init(struct drm_device *dev,
 	 *    if closed, act like it's not there for now
 	 */
 
+	edid = NULL;
 	mutex_lock(&dev->mode_config.mutex);
 	i2c_adap = i2c_get_adapter(dev_priv->ops->i2c_bus);
-	if (i2c_adap == NULL)
-		dev_err(dev->dev, "No ddc adapter available!\n");
+	if (i2c_adap)
+		edid = drm_get_edid(connector, i2c_adap);
+	if (edid == NULL && dev_priv->lpc_gpio_base) {
+		oaktrail_lvds_i2c_init(encoder);
+		if (gma_encoder->ddc_bus != NULL) {
+			i2c_adap = &gma_encoder->ddc_bus->adapter;
+			edid = drm_get_edid(connector, i2c_adap);
+		}
+	}
 	/*
 	 * Attempt to get the fixed panel mode from DDC.  Assume that the
 	 * preferred mode is the right one.
 	 */
-	if (i2c_adap) {
-		edid = drm_get_edid(connector, i2c_adap);
-		if (edid) {
-			drm_mode_connector_update_edid_property(connector,
-									edid);
-			drm_add_edid_modes(connector, edid);
-			kfree(edid);
-		}
+	if (edid) {
+		drm_mode_connector_update_edid_property(connector, edid);
+		drm_add_edid_modes(connector, edid);
+		kfree(edid);
 
 		list_for_each_entry(scan, &connector->probed_modes, head) {
 			if (scan->type & DRM_MODE_TYPE_PREFERRED) {
@@ -383,7 +387,8 @@ void oaktrail_lvds_init(struct drm_device *dev,
 				goto out;	/* FIXME: check for quirks */
 			}
 		}
-	}
+	} else
+		dev_err(dev->dev, "No ddc adapter available!\n");
 	/*
 	 * If we didn't get EDID, try geting panel timing
 	 * from configuration data
@@ -411,8 +416,10 @@ failed_find:
 	mutex_unlock(&dev->mode_config.mutex);
 
 	dev_dbg(dev->dev, "No LVDS modes found, disabling.\n");
-	if (gma_encoder->ddc_bus)
+	if (gma_encoder->ddc_bus) {
 		psb_intel_i2c_destroy(gma_encoder->ddc_bus);
+		gma_encoder->ddc_bus = NULL;
+	}
 
 /* failed_ddc: */
 
diff --git a/drivers/gpu/drm/gma500/oaktrail_lvds_i2c.c b/drivers/gpu/drm/gma500/oaktrail_lvds_i2c.c
new file mode 100644
index 000000000000..f913a62eee5f
--- /dev/null
+++ b/drivers/gpu/drm/gma500/oaktrail_lvds_i2c.c
@@ -0,0 +1,170 @@
+/*
+ * Copyright (c) 2002-2010, Intel Corporation.
+ * Copyright (c) 2014 ATRON electronic GmbH
+ *   Author: Jan Safrata <jan.nikitenko@gmail.com>
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ *
+ */
+
+#include <linux/kernel.h>
+#include <linux/module.h>
+#include <linux/pci.h>
+#include <linux/types.h>
+#include <linux/i2c.h>
+#include <linux/i2c-algo-bit.h>
+#include <linux/init.h>
+#include <linux/io.h>
+#include <linux/delay.h>
+
+#include <drm/drmP.h>
+#include "psb_drv.h"
+#include "psb_intel_reg.h"
+
+
+/*
+ * LPC GPIO based I2C bus for LVDS of Atom E6xx
+ */
+
+/*-----------------------------------------------------------------------------
+ * LPC Register Offsets. Used for LVDS GPIO Bit Bashing. Registers are part
+ * Atom E6xx [D31:F0]
+ ----------------------------------------------------------------------------*/
+#define RGEN    0x20
+#define RGIO    0x24
+#define RGLVL   0x28
+#define RGTPE   0x2C
+#define RGTNE   0x30
+#define RGGPE   0x34
+#define RGSMI   0x38
+#define RGTS    0x3C
+
+/* The LVDS GPIO clock lines are GPIOSUS[3]
+ * The LVDS GPIO data lines are GPIOSUS[4]
+ */
+#define GPIO_CLOCK	0x08
+#define GPIO_DATA	0x10
+
+#define LPC_READ_REG(chan, r) inl((chan)->reg + (r))
+#define LPC_WRITE_REG(chan, r, val) outl((val), (chan)->reg + (r))
+
+static int get_clock(void *data)
+{
+	struct psb_intel_i2c_chan *chan = data;
+	u32 val, tmp;
+
+	val = LPC_READ_REG(chan, RGIO);
+	val |= GPIO_CLOCK;
+	LPC_WRITE_REG(chan, RGIO, val);
+	tmp = LPC_READ_REG(chan, RGLVL);
+	val = (LPC_READ_REG(chan, RGLVL) & GPIO_CLOCK) ? 1 : 0;
+
+	return val;
+}
+
+static int get_data(void *data)
+{
+	struct psb_intel_i2c_chan *chan = data;
+	u32 val, tmp;
+
+	val = LPC_READ_REG(chan, RGIO);
+	val |= GPIO_DATA;
+	LPC_WRITE_REG(chan, RGIO, val);
+	tmp = LPC_READ_REG(chan, RGLVL);
+	val = (LPC_READ_REG(chan, RGLVL) & GPIO_DATA) ? 1 : 0;
+
+	return val;
+}
+
+static void set_clock(void *data, int state_high)
+{
+	struct psb_intel_i2c_chan *chan = data;
+	u32 val;
+
+	if (state_high) {
+		val = LPC_READ_REG(chan, RGIO);
+		val |= GPIO_CLOCK;
+		LPC_WRITE_REG(chan, RGIO, val);
+	} else {
+		val = LPC_READ_REG(chan, RGIO);
+		val &= ~GPIO_CLOCK;
+		LPC_WRITE_REG(chan, RGIO, val);
+		val = LPC_READ_REG(chan, RGLVL);
+		val &= ~GPIO_CLOCK;
+		LPC_WRITE_REG(chan, RGLVL, val);
+	}
+}
+
+static void set_data(void *data, int state_high)
+{
+	struct psb_intel_i2c_chan *chan = data;
+	u32 val;
+
+	if (state_high) {
+		val = LPC_READ_REG(chan, RGIO);
+		val |= GPIO_DATA;
+		LPC_WRITE_REG(chan, RGIO, val);
+	} else {
+		val = LPC_READ_REG(chan, RGIO);
+		val &= ~GPIO_DATA;
+		LPC_WRITE_REG(chan, RGIO, val);
+		val = LPC_READ_REG(chan, RGLVL);
+		val &= ~GPIO_DATA;
+		LPC_WRITE_REG(chan, RGLVL, val);
+	}
+}
+
+void oaktrail_lvds_i2c_init(struct drm_encoder *encoder)
+{
+	struct drm_device *dev = encoder->dev;
+	struct gma_encoder *gma_encoder = to_gma_encoder(encoder);
+	struct drm_psb_private *dev_priv = dev->dev_private;
+	struct psb_intel_i2c_chan *chan;
+
+	chan = kzalloc(sizeof(struct psb_intel_i2c_chan), GFP_KERNEL);
+	if (!chan)
+		return;
+
+	chan->drm_dev = dev;
+	chan->reg = dev_priv->lpc_gpio_base;
+	strncpy(chan->adapter.name, "gma500 LPC",  I2C_NAME_SIZE - 1);
+	chan->adapter.owner = THIS_MODULE;
+	chan->adapter.algo_data = &chan->algo;
+	chan->adapter.dev.parent = &dev->pdev->dev;
+	chan->algo.setsda = set_data;
+	chan->algo.setscl = set_clock;
+	chan->algo.getsda = get_data;
+	chan->algo.getscl = get_clock;
+	chan->algo.udelay = 100;
+	chan->algo.timeout = usecs_to_jiffies(2200);
+	chan->algo.data = chan;
+
+	i2c_set_adapdata(&chan->adapter, chan);
+
+	set_data(chan, 1);
+	set_clock(chan, 1);
+	udelay(50);
+
+	if (i2c_bit_add_bus(&chan->adapter)) {
+		kfree(chan);
+		return;
+	}
+
+	gma_encoder->ddc_bus = chan;
+}
diff --git a/drivers/gpu/drm/gma500/psb_drv.c b/drivers/gpu/drm/gma500/psb_drv.c
index eec993f93b1a..92e7e5795398 100644
--- a/drivers/gpu/drm/gma500/psb_drv.c
+++ b/drivers/gpu/drm/gma500/psb_drv.c
@@ -212,6 +212,8 @@ static int psb_driver_unload(struct drm_device *dev)
 		}
 		if (dev_priv->aux_pdev)
 			pci_dev_put(dev_priv->aux_pdev);
+		if (dev_priv->lpc_pdev)
+			pci_dev_put(dev_priv->lpc_pdev);
 
 		/* Destroy VBT data */
 		psb_intel_destroy_bios(dev);
@@ -280,6 +282,24 @@ static int psb_driver_load(struct drm_device *dev, unsigned long flags)
 			DRM_DEBUG_KMS("Couldn't find aux pci device");
 		}
 		dev_priv->gmbus_reg = dev_priv->aux_reg;
+
+		dev_priv->lpc_pdev = pci_get_bus_and_slot(0, PCI_DEVFN(31, 0));
+		if (dev_priv->lpc_pdev) {
+			pci_read_config_word(dev_priv->lpc_pdev, PSB_LPC_GBA,
+				&dev_priv->lpc_gpio_base);
+			pci_write_config_dword(dev_priv->lpc_pdev, PSB_LPC_GBA,
+				(u32)dev_priv->lpc_gpio_base | (1L<<31));
+			pci_read_config_word(dev_priv->lpc_pdev, PSB_LPC_GBA,
+				&dev_priv->lpc_gpio_base);
+			dev_priv->lpc_gpio_base &= 0xffc0;
+			if (dev_priv->lpc_gpio_base)
+				DRM_DEBUG_KMS("Found LPC GPIO at 0x%04x\n",
+						dev_priv->lpc_gpio_base);
+			else {
+				pci_dev_put(dev_priv->lpc_pdev);
+				dev_priv->lpc_pdev = NULL;
+			}
+		}
 	} else {
 		dev_priv->gmbus_reg = dev_priv->vdc_reg;
 	}
@@ -476,6 +496,7 @@ static struct drm_driver driver = {
 	.unload = psb_driver_unload,
 	.lastclose = psb_driver_lastclose,
 	.preclose = psb_driver_preclose,
+	.set_busid = drm_pci_set_busid,
 
 	.num_ioctls = ARRAY_SIZE(psb_ioctls),
 	.device_is_agp = psb_driver_device_is_agp,
diff --git a/drivers/gpu/drm/gma500/psb_drv.h b/drivers/gpu/drm/gma500/psb_drv.h
index 55ebe2bd88dd..e38057b91865 100644
--- a/drivers/gpu/drm/gma500/psb_drv.h
+++ b/drivers/gpu/drm/gma500/psb_drv.h
@@ -83,6 +83,7 @@ enum {
 #define PSB_PGETBL_CTL		 0x2020
 #define _PSB_PGETBL_ENABLED	 0x00000001
 #define PSB_SGX_2D_SLAVE_PORT	 0x4000
+#define PSB_LPC_GBA		 0x44
 
 /* TODO: To get rid of */
 #define PSB_TT_PRIV0_LIMIT	 (256*1024*1024)
@@ -441,6 +442,7 @@ struct psb_ops;
 struct drm_psb_private {
 	struct drm_device *dev;
 	struct pci_dev *aux_pdev; /* Currently only used by mrst */
+	struct pci_dev *lpc_pdev; /* Currently only used by mrst */
 	const struct psb_ops *ops;
 	const struct psb_offset *regmap;
 	
@@ -470,6 +472,7 @@ struct drm_psb_private {
 	uint8_t __iomem *sgx_reg;
 	uint8_t __iomem *vdc_reg;
 	uint8_t __iomem *aux_reg; /* Auxillary vdc pipe regs */
+	uint16_t lpc_gpio_base;
 	uint32_t gatt_free_offset;
 
 	/* Fencing / irq */
diff --git a/drivers/gpu/drm/gma500/psb_intel_display.c b/drivers/gpu/drm/gma500/psb_intel_display.c
index 87b50ba64ed4..b21a09451d1d 100644
--- a/drivers/gpu/drm/gma500/psb_intel_display.c
+++ b/drivers/gpu/drm/gma500/psb_intel_display.c
@@ -21,6 +21,7 @@
 #include <linux/i2c.h>
 
 #include <drm/drmP.h>
+#include <drm/drm_plane_helper.h>
 #include "framebuffer.h"
 #include "psb_drv.h"
 #include "psb_intel_drv.h"
diff --git a/drivers/gpu/drm/gma500/psb_intel_drv.h b/drivers/gpu/drm/gma500/psb_intel_drv.h
index 336bd3aa1a06..860dd2177ca1 100644
--- a/drivers/gpu/drm/gma500/psb_intel_drv.h
+++ b/drivers/gpu/drm/gma500/psb_intel_drv.h
@@ -223,6 +223,7 @@ extern void oaktrail_lvds_init(struct drm_device *dev,
 extern void oaktrail_wait_for_INTR_PKT_SENT(struct drm_device *dev);
 extern void oaktrail_dsi_init(struct drm_device *dev,
 			   struct psb_intel_mode_device *mode_dev);
+extern void oaktrail_lvds_i2c_init(struct drm_encoder *encoder);
 extern void mid_dsi_init(struct drm_device *dev,
 		    struct psb_intel_mode_device *mode_dev, int dsi_num);
 
diff --git a/drivers/gpu/drm/gma500/psb_intel_sdvo.c b/drivers/gpu/drm/gma500/psb_intel_sdvo.c
index 0be96fdb5e28..58529cea575d 100644
--- a/drivers/gpu/drm/gma500/psb_intel_sdvo.c
+++ b/drivers/gpu/drm/gma500/psb_intel_sdvo.c
@@ -1631,57 +1631,8 @@ static int psb_intel_sdvo_get_modes(struct drm_connector *connector)
 	return !list_empty(&connector->probed_modes);
 }
 
-static void
-psb_intel_sdvo_destroy_enhance_property(struct drm_connector *connector)
-{
-	struct psb_intel_sdvo_connector *psb_intel_sdvo_connector = to_psb_intel_sdvo_connector(connector);
-	struct drm_device *dev = connector->dev;
-
-	if (psb_intel_sdvo_connector->left)
-		drm_property_destroy(dev, psb_intel_sdvo_connector->left);
-	if (psb_intel_sdvo_connector->right)
-		drm_property_destroy(dev, psb_intel_sdvo_connector->right);
-	if (psb_intel_sdvo_connector->top)
-		drm_property_destroy(dev, psb_intel_sdvo_connector->top);
-	if (psb_intel_sdvo_connector->bottom)
-		drm_property_destroy(dev, psb_intel_sdvo_connector->bottom);
-	if (psb_intel_sdvo_connector->hpos)
-		drm_property_destroy(dev, psb_intel_sdvo_connector->hpos);
-	if (psb_intel_sdvo_connector->vpos)
-		drm_property_destroy(dev, psb_intel_sdvo_connector->vpos);
-	if (psb_intel_sdvo_connector->saturation)
-		drm_property_destroy(dev, psb_intel_sdvo_connector->saturation);
-	if (psb_intel_sdvo_connector->contrast)
-		drm_property_destroy(dev, psb_intel_sdvo_connector->contrast);
-	if (psb_intel_sdvo_connector->hue)
-		drm_property_destroy(dev, psb_intel_sdvo_connector->hue);
-	if (psb_intel_sdvo_connector->sharpness)
-		drm_property_destroy(dev, psb_intel_sdvo_connector->sharpness);
-	if (psb_intel_sdvo_connector->flicker_filter)
-		drm_property_destroy(dev, psb_intel_sdvo_connector->flicker_filter);
-	if (psb_intel_sdvo_connector->flicker_filter_2d)
-		drm_property_destroy(dev, psb_intel_sdvo_connector->flicker_filter_2d);
-	if (psb_intel_sdvo_connector->flicker_filter_adaptive)
-		drm_property_destroy(dev, psb_intel_sdvo_connector->flicker_filter_adaptive);
-	if (psb_intel_sdvo_connector->tv_luma_filter)
-		drm_property_destroy(dev, psb_intel_sdvo_connector->tv_luma_filter);
-	if (psb_intel_sdvo_connector->tv_chroma_filter)
-		drm_property_destroy(dev, psb_intel_sdvo_connector->tv_chroma_filter);
-	if (psb_intel_sdvo_connector->dot_crawl)
-		drm_property_destroy(dev, psb_intel_sdvo_connector->dot_crawl);
-	if (psb_intel_sdvo_connector->brightness)
-		drm_property_destroy(dev, psb_intel_sdvo_connector->brightness);
-}
-
 static void psb_intel_sdvo_destroy(struct drm_connector *connector)
 {
-	struct psb_intel_sdvo_connector *psb_intel_sdvo_connector = to_psb_intel_sdvo_connector(connector);
-
-	if (psb_intel_sdvo_connector->tv_format)
-		drm_property_destroy(connector->dev,
-				     psb_intel_sdvo_connector->tv_format);
-
-	psb_intel_sdvo_destroy_enhance_property(connector);
 	drm_connector_unregister(connector);
 	drm_connector_cleanup(connector);
 	kfree(connector);
diff --git a/drivers/gpu/drm/i2c/Kconfig b/drivers/gpu/drm/i2c/Kconfig
index 4d341db462a2..22c7ed63a001 100644
--- a/drivers/gpu/drm/i2c/Kconfig
+++ b/drivers/gpu/drm/i2c/Kconfig
@@ -1,6 +1,12 @@
 menu "I2C encoder or helper chips"
      depends on DRM && DRM_KMS_HELPER && I2C
 
+config DRM_I2C_ADV7511
+	tristate "AV7511 encoder"
+	select REGMAP_I2C
+	help
+	  Support for the Analog Device ADV7511(W) and ADV7513 HDMI encoders.
+
 config DRM_I2C_CH7006
 	tristate "Chrontel ch7006 TV encoder"
 	default m if DRM_NOUVEAU
diff --git a/drivers/gpu/drm/i2c/Makefile b/drivers/gpu/drm/i2c/Makefile
index 43aa33baebed..2c72eb584ab7 100644
--- a/drivers/gpu/drm/i2c/Makefile
+++ b/drivers/gpu/drm/i2c/Makefile
@@ -1,5 +1,7 @@
 ccflags-y := -Iinclude/drm
 
+obj-$(CONFIG_DRM_I2C_ADV7511) += adv7511.o
+
 ch7006-y := ch7006_drv.o ch7006_mode.o
 obj-$(CONFIG_DRM_I2C_CH7006) += ch7006.o
 
diff --git a/drivers/gpu/drm/i2c/adv7511.c b/drivers/gpu/drm/i2c/adv7511.c
new file mode 100644
index 000000000000..faf1c0c5ab2e
--- /dev/null
+++ b/drivers/gpu/drm/i2c/adv7511.c
@@ -0,0 +1,1010 @@
+/*
+ * Analog Devices ADV7511 HDMI transmitter driver
+ *
+ * Copyright 2012 Analog Devices Inc.
+ *
+ * Licensed under the GPL-2.
+ */
+
+#include <linux/device.h>
+#include <linux/gpio/consumer.h>
+#include <linux/i2c.h>
+#include <linux/module.h>
+#include <linux/regmap.h>
+#include <linux/slab.h>
+
+#include <drm/drmP.h>
+#include <drm/drm_crtc_helper.h>
+#include <drm/drm_edid.h>
+#include <drm/drm_encoder_slave.h>
+
+#include "adv7511.h"
+
+struct adv7511 {
+	struct i2c_client *i2c_main;
+	struct i2c_client *i2c_edid;
+
+	struct regmap *regmap;
+	struct regmap *packet_memory_regmap;
+	enum drm_connector_status status;
+	int dpms_mode;
+
+	unsigned int f_tmds;
+
+	unsigned int current_edid_segment;
+	uint8_t edid_buf[256];
+
+	wait_queue_head_t wq;
+	struct drm_encoder *encoder;
+
+	bool embedded_sync;
+	enum adv7511_sync_polarity vsync_polarity;
+	enum adv7511_sync_polarity hsync_polarity;
+	bool rgb;
+
+	struct edid *edid;
+
+	struct gpio_desc *gpio_pd;
+};
+
+static struct adv7511 *encoder_to_adv7511(struct drm_encoder *encoder)
+{
+	return to_encoder_slave(encoder)->slave_priv;
+}
+
+/* ADI recommended values for proper operation. */
+static const struct reg_default adv7511_fixed_registers[] = {
+	{ 0x98, 0x03 },
+	{ 0x9a, 0xe0 },
+	{ 0x9c, 0x30 },
+	{ 0x9d, 0x61 },
+	{ 0xa2, 0xa4 },
+	{ 0xa3, 0xa4 },
+	{ 0xe0, 0xd0 },
+	{ 0xf9, 0x00 },
+	{ 0x55, 0x02 },
+};
+
+/* -----------------------------------------------------------------------------
+ * Register access
+ */
+
+static const uint8_t adv7511_register_defaults[] = {
+	0x12, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 00 */
+	0x00, 0x00, 0x01, 0x0e, 0xbc, 0x18, 0x01, 0x13,
+	0x25, 0x37, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 10 */
+	0x46, 0x62, 0x04, 0xa8, 0x00, 0x00, 0x1c, 0x84,
+	0x1c, 0xbf, 0x04, 0xa8, 0x1e, 0x70, 0x02, 0x1e, /* 20 */
+	0x00, 0x00, 0x04, 0xa8, 0x08, 0x12, 0x1b, 0xac,
+	0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 30 */
+	0x00, 0x00, 0x00, 0x80, 0x00, 0x00, 0x00, 0xb0,
+	0x00, 0x50, 0x90, 0x7e, 0x79, 0x70, 0x00, 0x00, /* 40 */
+	0x00, 0xa8, 0x80, 0x00, 0x00, 0x00, 0x00, 0x00,
+	0x00, 0x00, 0x02, 0x0d, 0x00, 0x00, 0x00, 0x00, /* 50 */
+	0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+	0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 60 */
+	0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+	0x01, 0x0a, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 70 */
+	0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+	0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 80 */
+	0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+	0x00, 0x00, 0x00, 0x00, 0xc0, 0x00, 0x00, 0x00, /* 90 */
+	0x0b, 0x02, 0x00, 0x18, 0x5a, 0x60, 0x00, 0x00,
+	0x00, 0x00, 0x80, 0x80, 0x08, 0x04, 0x00, 0x00, /* a0 */
+	0x00, 0x00, 0x00, 0x40, 0x00, 0x00, 0x40, 0x14,
+	0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* b0 */
+	0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+	0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* c0 */
+	0x00, 0x03, 0x00, 0x00, 0x02, 0x00, 0x01, 0x04,
+	0x30, 0xff, 0x80, 0x80, 0x80, 0x00, 0x00, 0x00, /* d0 */
+	0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x10, 0x01,
+	0x80, 0x75, 0x00, 0x00, 0x60, 0x00, 0x00, 0x00, /* e0 */
+	0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+	0x00, 0x00, 0x00, 0x00, 0x00, 0x75, 0x11, 0x00, /* f0 */
+	0x00, 0x7c, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+};
+
+static bool adv7511_register_volatile(struct device *dev, unsigned int reg)
+{
+	switch (reg) {
+	case ADV7511_REG_CHIP_REVISION:
+	case ADV7511_REG_SPDIF_FREQ:
+	case ADV7511_REG_CTS_AUTOMATIC1:
+	case ADV7511_REG_CTS_AUTOMATIC2:
+	case ADV7511_REG_VIC_DETECTED:
+	case ADV7511_REG_VIC_SEND:
+	case ADV7511_REG_AUX_VIC_DETECTED:
+	case ADV7511_REG_STATUS:
+	case ADV7511_REG_GC(1):
+	case ADV7511_REG_INT(0):
+	case ADV7511_REG_INT(1):
+	case ADV7511_REG_PLL_STATUS:
+	case ADV7511_REG_AN(0):
+	case ADV7511_REG_AN(1):
+	case ADV7511_REG_AN(2):
+	case ADV7511_REG_AN(3):
+	case ADV7511_REG_AN(4):
+	case ADV7511_REG_AN(5):
+	case ADV7511_REG_AN(6):
+	case ADV7511_REG_AN(7):
+	case ADV7511_REG_HDCP_STATUS:
+	case ADV7511_REG_BCAPS:
+	case ADV7511_REG_BKSV(0):
+	case ADV7511_REG_BKSV(1):
+	case ADV7511_REG_BKSV(2):
+	case ADV7511_REG_BKSV(3):
+	case ADV7511_REG_BKSV(4):
+	case ADV7511_REG_DDC_STATUS:
+	case ADV7511_REG_BSTATUS(0):
+	case ADV7511_REG_BSTATUS(1):
+	case ADV7511_REG_CHIP_ID_HIGH:
+	case ADV7511_REG_CHIP_ID_LOW:
+		return true;
+	}
+
+	return false;
+}
+
+static const struct regmap_config adv7511_regmap_config = {
+	.reg_bits = 8,
+	.val_bits = 8,
+
+	.max_register = 0xff,
+	.cache_type = REGCACHE_RBTREE,
+	.reg_defaults_raw = adv7511_register_defaults,
+	.num_reg_defaults_raw = ARRAY_SIZE(adv7511_register_defaults),
+
+	.volatile_reg = adv7511_register_volatile,
+};
+
+/* -----------------------------------------------------------------------------
+ * Hardware configuration
+ */
+
+static void adv7511_set_colormap(struct adv7511 *adv7511, bool enable,
+				 const uint16_t *coeff,
+				 unsigned int scaling_factor)
+{
+	unsigned int i;
+
+	regmap_update_bits(adv7511->regmap, ADV7511_REG_CSC_UPPER(1),
+			   ADV7511_CSC_UPDATE_MODE, ADV7511_CSC_UPDATE_MODE);
+
+	if (enable) {
+		for (i = 0; i < 12; ++i) {
+			regmap_update_bits(adv7511->regmap,
+					   ADV7511_REG_CSC_UPPER(i),
+					   0x1f, coeff[i] >> 8);
+			regmap_write(adv7511->regmap,
+				     ADV7511_REG_CSC_LOWER(i),
+				     coeff[i] & 0xff);
+		}
+	}
+
+	if (enable)
+		regmap_update_bits(adv7511->regmap, ADV7511_REG_CSC_UPPER(0),
+				   0xe0, 0x80 | (scaling_factor << 5));
+	else
+		regmap_update_bits(adv7511->regmap, ADV7511_REG_CSC_UPPER(0),
+				   0x80, 0x00);
+
+	regmap_update_bits(adv7511->regmap, ADV7511_REG_CSC_UPPER(1),
+			   ADV7511_CSC_UPDATE_MODE, 0);
+}
+
+static int adv7511_packet_enable(struct adv7511 *adv7511, unsigned int packet)
+{
+	if (packet & 0xff)
+		regmap_update_bits(adv7511->regmap, ADV7511_REG_PACKET_ENABLE0,
+				   packet, 0xff);
+
+	if (packet & 0xff00) {
+		packet >>= 8;
+		regmap_update_bits(adv7511->regmap, ADV7511_REG_PACKET_ENABLE1,
+				   packet, 0xff);
+	}
+
+	return 0;
+}
+
+static int adv7511_packet_disable(struct adv7511 *adv7511, unsigned int packet)
+{
+	if (packet & 0xff)
+		regmap_update_bits(adv7511->regmap, ADV7511_REG_PACKET_ENABLE0,
+				   packet, 0x00);
+
+	if (packet & 0xff00) {
+		packet >>= 8;
+		regmap_update_bits(adv7511->regmap, ADV7511_REG_PACKET_ENABLE1,
+				   packet, 0x00);
+	}
+
+	return 0;
+}
+
+/* Coefficients for adv7511 color space conversion */
+static const uint16_t adv7511_csc_ycbcr_to_rgb[] = {
+	0x0734, 0x04ad, 0x0000, 0x1c1b,
+	0x1ddc, 0x04ad, 0x1f24, 0x0135,
+	0x0000, 0x04ad, 0x087c, 0x1b77,
+};
+
+static void adv7511_set_config_csc(struct adv7511 *adv7511,
+				   struct drm_connector *connector,
+				   bool rgb)
+{
+	struct adv7511_video_config config;
+	bool output_format_422, output_format_ycbcr;
+	unsigned int mode;
+	uint8_t infoframe[17];
+
+	if (adv7511->edid)
+		config.hdmi_mode = drm_detect_hdmi_monitor(adv7511->edid);
+	else
+		config.hdmi_mode = false;
+
+	hdmi_avi_infoframe_init(&config.avi_infoframe);
+
+	config.avi_infoframe.scan_mode = HDMI_SCAN_MODE_UNDERSCAN;
+
+	if (rgb) {
+		config.csc_enable = false;
+		config.avi_infoframe.colorspace = HDMI_COLORSPACE_RGB;
+	} else {
+		config.csc_scaling_factor = ADV7511_CSC_SCALING_4;
+		config.csc_coefficents = adv7511_csc_ycbcr_to_rgb;
+
+		if ((connector->display_info.color_formats &
+		     DRM_COLOR_FORMAT_YCRCB422) &&
+		    config.hdmi_mode) {
+			config.csc_enable = false;
+			config.avi_infoframe.colorspace =
+				HDMI_COLORSPACE_YUV422;
+		} else {
+			config.csc_enable = true;
+			config.avi_infoframe.colorspace = HDMI_COLORSPACE_RGB;
+		}
+	}
+
+	if (config.hdmi_mode) {
+		mode = ADV7511_HDMI_CFG_MODE_HDMI;
+
+		switch (config.avi_infoframe.colorspace) {
+		case HDMI_COLORSPACE_YUV444:
+			output_format_422 = false;
+			output_format_ycbcr = true;
+			break;
+		case HDMI_COLORSPACE_YUV422:
+			output_format_422 = true;
+			output_format_ycbcr = true;
+			break;
+		default:
+			output_format_422 = false;
+			output_format_ycbcr = false;
+			break;
+		}
+	} else {
+		mode = ADV7511_HDMI_CFG_MODE_DVI;
+		output_format_422 = false;
+		output_format_ycbcr = false;
+	}
+
+	adv7511_packet_disable(adv7511, ADV7511_PACKET_ENABLE_AVI_INFOFRAME);
+
+	adv7511_set_colormap(adv7511, config.csc_enable,
+			     config.csc_coefficents,
+			     config.csc_scaling_factor);
+
+	regmap_update_bits(adv7511->regmap, ADV7511_REG_VIDEO_INPUT_CFG1, 0x81,
+			   (output_format_422 << 7) | output_format_ycbcr);
+
+	regmap_update_bits(adv7511->regmap, ADV7511_REG_HDCP_HDMI_CFG,
+			   ADV7511_HDMI_CFG_MODE_MASK, mode);
+
+	hdmi_avi_infoframe_pack(&config.avi_infoframe, infoframe,
+				sizeof(infoframe));
+
+	/* The AVI infoframe id is not configurable */
+	regmap_bulk_write(adv7511->regmap, ADV7511_REG_AVI_INFOFRAME_VERSION,
+			  infoframe + 1, sizeof(infoframe) - 1);
+
+	adv7511_packet_enable(adv7511, ADV7511_PACKET_ENABLE_AVI_INFOFRAME);
+}
+
+static void adv7511_set_link_config(struct adv7511 *adv7511,
+				    const struct adv7511_link_config *config)
+{
+	/*
+	 * The input style values documented in the datasheet don't match the
+	 * hardware register field values :-(
+	 */
+	static const unsigned int input_styles[4] = { 0, 2, 1, 3 };
+
+	unsigned int clock_delay;
+	unsigned int color_depth;
+	unsigned int input_id;
+
+	clock_delay = (config->clock_delay + 1200) / 400;
+	color_depth = config->input_color_depth == 8 ? 3
+		    : (config->input_color_depth == 10 ? 1 : 2);
+
+	/* TODO Support input ID 6 */
+	if (config->input_colorspace != HDMI_COLORSPACE_YUV422)
+		input_id = config->input_clock == ADV7511_INPUT_CLOCK_DDR
+			 ? 5 : 0;
+	else if (config->input_clock == ADV7511_INPUT_CLOCK_DDR)
+		input_id = config->embedded_sync ? 8 : 7;
+	else if (config->input_clock == ADV7511_INPUT_CLOCK_2X)
+		input_id = config->embedded_sync ? 4 : 3;
+	else
+		input_id = config->embedded_sync ? 2 : 1;
+
+	regmap_update_bits(adv7511->regmap, ADV7511_REG_I2C_FREQ_ID_CFG, 0xf,
+			   input_id);
+	regmap_update_bits(adv7511->regmap, ADV7511_REG_VIDEO_INPUT_CFG1, 0x7e,
+			   (color_depth << 4) |
+			   (input_styles[config->input_style] << 2));
+	regmap_write(adv7511->regmap, ADV7511_REG_VIDEO_INPUT_CFG2,
+		     config->input_justification << 3);
+	regmap_write(adv7511->regmap, ADV7511_REG_TIMING_GEN_SEQ,
+		     config->sync_pulse << 2);
+
+	regmap_write(adv7511->regmap, 0xba, clock_delay << 5);
+
+	adv7511->embedded_sync = config->embedded_sync;
+	adv7511->hsync_polarity = config->hsync_polarity;
+	adv7511->vsync_polarity = config->vsync_polarity;
+	adv7511->rgb = config->input_colorspace == HDMI_COLORSPACE_RGB;
+}
+
+/* -----------------------------------------------------------------------------
+ * Interrupt and hotplug detection
+ */
+
+static bool adv7511_hpd(struct adv7511 *adv7511)
+{
+	unsigned int irq0;
+	int ret;
+
+	ret = regmap_read(adv7511->regmap, ADV7511_REG_INT(0), &irq0);
+	if (ret < 0)
+		return false;
+
+	if (irq0 & ADV7511_INT0_HDP) {
+		regmap_write(adv7511->regmap, ADV7511_REG_INT(0),
+			     ADV7511_INT0_HDP);
+		return true;
+	}
+
+	return false;
+}
+
+static irqreturn_t adv7511_irq_handler(int irq, void *devid)
+{
+	struct adv7511 *adv7511 = devid;
+
+	if (adv7511_hpd(adv7511))
+		drm_helper_hpd_irq_event(adv7511->encoder->dev);
+
+	wake_up_all(&adv7511->wq);
+
+	return IRQ_HANDLED;
+}
+
+static unsigned int adv7511_is_interrupt_pending(struct adv7511 *adv7511,
+						 unsigned int irq)
+{
+	unsigned int irq0, irq1;
+	unsigned int pending;
+	int ret;
+
+	ret = regmap_read(adv7511->regmap, ADV7511_REG_INT(0), &irq0);
+	if (ret < 0)
+		return 0;
+	ret = regmap_read(adv7511->regmap, ADV7511_REG_INT(1), &irq1);
+	if (ret < 0)
+		return 0;
+
+	pending = (irq1 << 8) | irq0;
+
+	return pending & irq;
+}
+
+static int adv7511_wait_for_interrupt(struct adv7511 *adv7511, int irq,
+				      int timeout)
+{
+	unsigned int pending;
+	int ret;
+
+	if (adv7511->i2c_main->irq) {
+		ret = wait_event_interruptible_timeout(adv7511->wq,
+				adv7511_is_interrupt_pending(adv7511, irq),
+				msecs_to_jiffies(timeout));
+		if (ret <= 0)
+			return 0;
+		pending = adv7511_is_interrupt_pending(adv7511, irq);
+	} else {
+		if (timeout < 25)
+			timeout = 25;
+		do {
+			pending = adv7511_is_interrupt_pending(adv7511, irq);
+			if (pending)
+				break;
+			msleep(25);
+			timeout -= 25;
+		} while (timeout >= 25);
+	}
+
+	return pending;
+}
+
+/* -----------------------------------------------------------------------------
+ * EDID retrieval
+ */
+
+static int adv7511_get_edid_block(void *data, u8 *buf, unsigned int block,
+				  size_t len)
+{
+	struct adv7511 *adv7511 = data;
+	struct i2c_msg xfer[2];
+	uint8_t offset;
+	unsigned int i;
+	int ret;
+
+	if (len > 128)
+		return -EINVAL;
+
+	if (adv7511->current_edid_segment != block / 2) {
+		unsigned int status;
+
+		ret = regmap_read(adv7511->regmap, ADV7511_REG_DDC_STATUS,
+				  &status);
+		if (ret < 0)
+			return ret;
+
+		if (status != 2) {
+			regmap_write(adv7511->regmap, ADV7511_REG_EDID_SEGMENT,
+				     block);
+			ret = adv7511_wait_for_interrupt(adv7511,
+					ADV7511_INT0_EDID_READY |
+					ADV7511_INT1_DDC_ERROR, 200);
+
+			if (!(ret & ADV7511_INT0_EDID_READY))
+				return -EIO;
+		}
+
+		regmap_write(adv7511->regmap, ADV7511_REG_INT(0),
+			     ADV7511_INT0_EDID_READY | ADV7511_INT1_DDC_ERROR);
+
+		/* Break this apart, hopefully more I2C controllers will
+		 * support 64 byte transfers than 256 byte transfers
+		 */
+
+		xfer[0].addr = adv7511->i2c_edid->addr;
+		xfer[0].flags = 0;
+		xfer[0].len = 1;
+		xfer[0].buf = &offset;
+		xfer[1].addr = adv7511->i2c_edid->addr;
+		xfer[1].flags = I2C_M_RD;
+		xfer[1].len = 64;
+		xfer[1].buf = adv7511->edid_buf;
+
+		offset = 0;
+
+		for (i = 0; i < 4; ++i) {
+			ret = i2c_transfer(adv7511->i2c_edid->adapter, xfer,
+					   ARRAY_SIZE(xfer));
+			if (ret < 0)
+				return ret;
+			else if (ret != 2)
+				return -EIO;
+
+			xfer[1].buf += 64;
+			offset += 64;
+		}
+
+		adv7511->current_edid_segment = block / 2;
+	}
+
+	if (block % 2 == 0)
+		memcpy(buf, adv7511->edid_buf, len);
+	else
+		memcpy(buf, adv7511->edid_buf + 128, len);
+
+	return 0;
+}
+
+/* -----------------------------------------------------------------------------
+ * Encoder operations
+ */
+
+static int adv7511_get_modes(struct drm_encoder *encoder,
+			     struct drm_connector *connector)
+{
+	struct adv7511 *adv7511 = encoder_to_adv7511(encoder);
+	struct edid *edid;
+	unsigned int count;
+
+	/* Reading the EDID only works if the device is powered */
+	if (adv7511->dpms_mode != DRM_MODE_DPMS_ON) {
+		regmap_write(adv7511->regmap, ADV7511_REG_INT(0),
+			     ADV7511_INT0_EDID_READY | ADV7511_INT1_DDC_ERROR);
+		regmap_update_bits(adv7511->regmap, ADV7511_REG_POWER,
+				   ADV7511_POWER_POWER_DOWN, 0);
+		adv7511->current_edid_segment = -1;
+	}
+
+	edid = drm_do_get_edid(connector, adv7511_get_edid_block, adv7511);
+
+	if (adv7511->dpms_mode != DRM_MODE_DPMS_ON)
+		regmap_update_bits(adv7511->regmap, ADV7511_REG_POWER,
+				   ADV7511_POWER_POWER_DOWN,
+				   ADV7511_POWER_POWER_DOWN);
+
+	kfree(adv7511->edid);
+	adv7511->edid = edid;
+	if (!edid)
+		return 0;
+
+	drm_mode_connector_update_edid_property(connector, edid);
+	count = drm_add_edid_modes(connector, edid);
+
+	adv7511_set_config_csc(adv7511, connector, adv7511->rgb);
+
+	return count;
+}
+
+static void adv7511_encoder_dpms(struct drm_encoder *encoder, int mode)
+{
+	struct adv7511 *adv7511 = encoder_to_adv7511(encoder);
+
+	switch (mode) {
+	case DRM_MODE_DPMS_ON:
+		adv7511->current_edid_segment = -1;
+
+		regmap_write(adv7511->regmap, ADV7511_REG_INT(0),
+			     ADV7511_INT0_EDID_READY | ADV7511_INT1_DDC_ERROR);
+		regmap_update_bits(adv7511->regmap, ADV7511_REG_POWER,
+				   ADV7511_POWER_POWER_DOWN, 0);
+		/*
+		 * Per spec it is allowed to pulse the HDP signal to indicate
+		 * that the EDID information has changed. Some monitors do this
+		 * when they wakeup from standby or are enabled. When the HDP
+		 * goes low the adv7511 is reset and the outputs are disabled
+		 * which might cause the monitor to go to standby again. To
+		 * avoid this we ignore the HDP pin for the first few seconds
+		 * after enabeling the output.
+		 */
+		regmap_update_bits(adv7511->regmap, ADV7511_REG_POWER2,
+				   ADV7511_REG_POWER2_HDP_SRC_MASK,
+				   ADV7511_REG_POWER2_HDP_SRC_NONE);
+		/* Most of the registers are reset during power down or
+		 * when HPD is low
+		 */
+		regcache_sync(adv7511->regmap);
+		break;
+	default:
+		/* TODO: setup additional power down modes */
+		regmap_update_bits(adv7511->regmap, ADV7511_REG_POWER,
+				   ADV7511_POWER_POWER_DOWN,
+				   ADV7511_POWER_POWER_DOWN);
+		regcache_mark_dirty(adv7511->regmap);
+		break;
+	}
+
+	adv7511->dpms_mode = mode;
+}
+
+static enum drm_connector_status
+adv7511_encoder_detect(struct drm_encoder *encoder,
+		       struct drm_connector *connector)
+{
+	struct adv7511 *adv7511 = encoder_to_adv7511(encoder);
+	enum drm_connector_status status;
+	unsigned int val;
+	bool hpd;
+	int ret;
+
+	ret = regmap_read(adv7511->regmap, ADV7511_REG_STATUS, &val);
+	if (ret < 0)
+		return connector_status_disconnected;
+
+	if (val & ADV7511_STATUS_HPD)
+		status = connector_status_connected;
+	else
+		status = connector_status_disconnected;
+
+	hpd = adv7511_hpd(adv7511);
+
+	/* The chip resets itself when the cable is disconnected, so in case
+	 * there is a pending HPD interrupt and the cable is connected there was
+	 * at least one transition from disconnected to connected and the chip
+	 * has to be reinitialized. */
+	if (status == connector_status_connected && hpd &&
+	    adv7511->dpms_mode == DRM_MODE_DPMS_ON) {
+		regcache_mark_dirty(adv7511->regmap);
+		adv7511_encoder_dpms(encoder, adv7511->dpms_mode);
+		adv7511_get_modes(encoder, connector);
+		if (adv7511->status == connector_status_connected)
+			status = connector_status_disconnected;
+	} else {
+		/* Renable HDP sensing */
+		regmap_update_bits(adv7511->regmap, ADV7511_REG_POWER2,
+				   ADV7511_REG_POWER2_HDP_SRC_MASK,
+				   ADV7511_REG_POWER2_HDP_SRC_BOTH);
+	}
+
+	adv7511->status = status;
+	return status;
+}
+
+static int adv7511_encoder_mode_valid(struct drm_encoder *encoder,
+				      struct drm_display_mode *mode)
+{
+	if (mode->clock > 165000)
+		return MODE_CLOCK_HIGH;
+
+	if (mode->flags & DRM_MODE_FLAG_INTERLACE)
+		return MODE_NO_INTERLACE;
+
+	return MODE_OK;
+}
+
+static void adv7511_encoder_mode_set(struct drm_encoder *encoder,
+				     struct drm_display_mode *mode,
+				     struct drm_display_mode *adj_mode)
+{
+	struct adv7511 *adv7511 = encoder_to_adv7511(encoder);
+	unsigned int low_refresh_rate;
+	unsigned int hsync_polarity = 0;
+	unsigned int vsync_polarity = 0;
+
+	if (adv7511->embedded_sync) {
+		unsigned int hsync_offset, hsync_len;
+		unsigned int vsync_offset, vsync_len;
+
+		hsync_offset = adj_mode->crtc_hsync_start -
+			       adj_mode->crtc_hdisplay;
+		vsync_offset = adj_mode->crtc_vsync_start -
+			       adj_mode->crtc_vdisplay;
+		hsync_len = adj_mode->crtc_hsync_end -
+			    adj_mode->crtc_hsync_start;
+		vsync_len = adj_mode->crtc_vsync_end -
+			    adj_mode->crtc_vsync_start;
+
+		/* The hardware vsync generator has a off-by-one bug */
+		vsync_offset += 1;
+
+		regmap_write(adv7511->regmap, ADV7511_REG_HSYNC_PLACEMENT_MSB,
+			     ((hsync_offset >> 10) & 0x7) << 5);
+		regmap_write(adv7511->regmap, ADV7511_REG_SYNC_DECODER(0),
+			     (hsync_offset >> 2) & 0xff);
+		regmap_write(adv7511->regmap, ADV7511_REG_SYNC_DECODER(1),
+			     ((hsync_offset & 0x3) << 6) |
+			     ((hsync_len >> 4) & 0x3f));
+		regmap_write(adv7511->regmap, ADV7511_REG_SYNC_DECODER(2),
+			     ((hsync_len & 0xf) << 4) |
+			     ((vsync_offset >> 6) & 0xf));
+		regmap_write(adv7511->regmap, ADV7511_REG_SYNC_DECODER(3),
+			     ((vsync_offset & 0x3f) << 2) |
+			     ((vsync_len >> 8) & 0x3));
+		regmap_write(adv7511->regmap, ADV7511_REG_SYNC_DECODER(4),
+			     vsync_len & 0xff);
+
+		hsync_polarity = !(adj_mode->flags & DRM_MODE_FLAG_PHSYNC);
+		vsync_polarity = !(adj_mode->flags & DRM_MODE_FLAG_PVSYNC);
+	} else {
+		enum adv7511_sync_polarity mode_hsync_polarity;
+		enum adv7511_sync_polarity mode_vsync_polarity;
+
+		/**
+		 * If the input signal is always low or always high we want to
+		 * invert or let it passthrough depending on the polarity of the
+		 * current mode.
+		 **/
+		if (adj_mode->flags & DRM_MODE_FLAG_NHSYNC)
+			mode_hsync_polarity = ADV7511_SYNC_POLARITY_LOW;
+		else
+			mode_hsync_polarity = ADV7511_SYNC_POLARITY_HIGH;
+
+		if (adj_mode->flags & DRM_MODE_FLAG_NVSYNC)
+			mode_vsync_polarity = ADV7511_SYNC_POLARITY_LOW;
+		else
+			mode_vsync_polarity = ADV7511_SYNC_POLARITY_HIGH;
+
+		if (adv7511->hsync_polarity != mode_hsync_polarity &&
+		    adv7511->hsync_polarity !=
+		    ADV7511_SYNC_POLARITY_PASSTHROUGH)
+			hsync_polarity = 1;
+
+		if (adv7511->vsync_polarity != mode_vsync_polarity &&
+		    adv7511->vsync_polarity !=
+		    ADV7511_SYNC_POLARITY_PASSTHROUGH)
+			vsync_polarity = 1;
+	}
+
+	if (mode->vrefresh <= 24000)
+		low_refresh_rate = ADV7511_LOW_REFRESH_RATE_24HZ;
+	else if (mode->vrefresh <= 25000)
+		low_refresh_rate = ADV7511_LOW_REFRESH_RATE_25HZ;
+	else if (mode->vrefresh <= 30000)
+		low_refresh_rate = ADV7511_LOW_REFRESH_RATE_30HZ;
+	else
+		low_refresh_rate = ADV7511_LOW_REFRESH_RATE_NONE;
+
+	regmap_update_bits(adv7511->regmap, 0xfb,
+		0x6, low_refresh_rate << 1);
+	regmap_update_bits(adv7511->regmap, 0x17,
+		0x60, (vsync_polarity << 6) | (hsync_polarity << 5));
+
+	/*
+	 * TODO Test first order 4:2:2 to 4:4:4 up conversion method, which is
+	 * supposed to give better results.
+	 */
+
+	adv7511->f_tmds = mode->clock;
+}
+
+static struct drm_encoder_slave_funcs adv7511_encoder_funcs = {
+	.dpms = adv7511_encoder_dpms,
+	.mode_valid = adv7511_encoder_mode_valid,
+	.mode_set = adv7511_encoder_mode_set,
+	.detect = adv7511_encoder_detect,
+	.get_modes = adv7511_get_modes,
+};
+
+/* -----------------------------------------------------------------------------
+ * Probe & remove
+ */
+
+static int adv7511_parse_dt(struct device_node *np,
+			    struct adv7511_link_config *config)
+{
+	const char *str;
+	int ret;
+
+	memset(config, 0, sizeof(*config));
+
+	of_property_read_u32(np, "adi,input-depth", &config->input_color_depth);
+	if (config->input_color_depth != 8 && config->input_color_depth != 10 &&
+	    config->input_color_depth != 12)
+		return -EINVAL;
+
+	ret = of_property_read_string(np, "adi,input-colorspace", &str);
+	if (ret < 0)
+		return ret;
+
+	if (!strcmp(str, "rgb"))
+		config->input_colorspace = HDMI_COLORSPACE_RGB;
+	else if (!strcmp(str, "yuv422"))
+		config->input_colorspace = HDMI_COLORSPACE_YUV422;
+	else if (!strcmp(str, "yuv444"))
+		config->input_colorspace = HDMI_COLORSPACE_YUV444;
+	else
+		return -EINVAL;
+
+	ret = of_property_read_string(np, "adi,input-clock", &str);
+	if (ret < 0)
+		return ret;
+
+	if (!strcmp(str, "1x"))
+		config->input_clock = ADV7511_INPUT_CLOCK_1X;
+	else if (!strcmp(str, "2x"))
+		config->input_clock = ADV7511_INPUT_CLOCK_2X;
+	else if (!strcmp(str, "ddr"))
+		config->input_clock = ADV7511_INPUT_CLOCK_DDR;
+	else
+		return -EINVAL;
+
+	if (config->input_colorspace == HDMI_COLORSPACE_YUV422 ||
+	    config->input_clock != ADV7511_INPUT_CLOCK_1X) {
+		ret = of_property_read_u32(np, "adi,input-style",
+					   &config->input_style);
+		if (ret)
+			return ret;
+
+		if (config->input_style < 1 || config->input_style > 3)
+			return -EINVAL;
+
+		ret = of_property_read_string(np, "adi,input-justification",
+					      &str);
+		if (ret < 0)
+			return ret;
+
+		if (!strcmp(str, "left"))
+			config->input_justification =
+				ADV7511_INPUT_JUSTIFICATION_LEFT;
+		else if (!strcmp(str, "evenly"))
+			config->input_justification =
+				ADV7511_INPUT_JUSTIFICATION_EVENLY;
+		else if (!strcmp(str, "right"))
+			config->input_justification =
+				ADV7511_INPUT_JUSTIFICATION_RIGHT;
+		else
+			return -EINVAL;
+
+	} else {
+		config->input_style = 1;
+		config->input_justification = ADV7511_INPUT_JUSTIFICATION_LEFT;
+	}
+
+	of_property_read_u32(np, "adi,clock-delay", &config->clock_delay);
+	if (config->clock_delay < -1200 || config->clock_delay > 1600)
+		return -EINVAL;
+
+	config->embedded_sync = of_property_read_bool(np, "adi,embedded-sync");
+
+	/* Hardcode the sync pulse configurations for now. */
+	config->sync_pulse = ADV7511_INPUT_SYNC_PULSE_NONE;
+	config->vsync_polarity = ADV7511_SYNC_POLARITY_PASSTHROUGH;
+	config->hsync_polarity = ADV7511_SYNC_POLARITY_PASSTHROUGH;
+
+	return 0;
+}
+
+static const int edid_i2c_addr = 0x7e;
+static const int packet_i2c_addr = 0x70;
+static const int cec_i2c_addr = 0x78;
+
+static int adv7511_probe(struct i2c_client *i2c, const struct i2c_device_id *id)
+{
+	struct adv7511_link_config link_config;
+	struct adv7511 *adv7511;
+	struct device *dev = &i2c->dev;
+	unsigned int val;
+	int ret;
+
+	if (!dev->of_node)
+		return -EINVAL;
+
+	adv7511 = devm_kzalloc(dev, sizeof(*adv7511), GFP_KERNEL);
+	if (!adv7511)
+		return -ENOMEM;
+
+	adv7511->dpms_mode = DRM_MODE_DPMS_OFF;
+	adv7511->status = connector_status_disconnected;
+
+	ret = adv7511_parse_dt(dev->of_node, &link_config);
+	if (ret)
+		return ret;
+
+	/*
+	 * The power down GPIO is optional. If present, toggle it from active to
+	 * inactive to wake up the encoder.
+	 */
+	adv7511->gpio_pd = devm_gpiod_get_optional(dev, "pd", GPIOD_OUT_HIGH);
+	if (IS_ERR(adv7511->gpio_pd))
+		return PTR_ERR(adv7511->gpio_pd);
+
+	if (adv7511->gpio_pd) {
+		mdelay(5);
+		gpiod_set_value_cansleep(adv7511->gpio_pd, 0);
+	}
+
+	adv7511->regmap = devm_regmap_init_i2c(i2c, &adv7511_regmap_config);
+	if (IS_ERR(adv7511->regmap))
+		return PTR_ERR(adv7511->regmap);
+
+	ret = regmap_read(adv7511->regmap, ADV7511_REG_CHIP_REVISION, &val);
+	if (ret)
+		return ret;
+	dev_dbg(dev, "Rev. %d\n", val);
+
+	ret = regmap_register_patch(adv7511->regmap, adv7511_fixed_registers,
+				    ARRAY_SIZE(adv7511_fixed_registers));
+	if (ret)
+		return ret;
+
+	regmap_write(adv7511->regmap, ADV7511_REG_EDID_I2C_ADDR, edid_i2c_addr);
+	regmap_write(adv7511->regmap, ADV7511_REG_PACKET_I2C_ADDR,
+		     packet_i2c_addr);
+	regmap_write(adv7511->regmap, ADV7511_REG_CEC_I2C_ADDR, cec_i2c_addr);
+	adv7511_packet_disable(adv7511, 0xffff);
+
+	adv7511->i2c_main = i2c;
+	adv7511->i2c_edid = i2c_new_dummy(i2c->adapter, edid_i2c_addr >> 1);
+	if (!adv7511->i2c_edid)
+		return -ENOMEM;
+
+	if (i2c->irq) {
+		init_waitqueue_head(&adv7511->wq);
+
+		ret = devm_request_threaded_irq(dev, i2c->irq, NULL,
+						adv7511_irq_handler,
+						IRQF_ONESHOT, dev_name(dev),
+						adv7511);
+		if (ret)
+			goto err_i2c_unregister_device;
+	}
+
+	/* CEC is unused for now */
+	regmap_write(adv7511->regmap, ADV7511_REG_CEC_CTRL,
+		     ADV7511_CEC_CTRL_POWER_DOWN);
+
+	regmap_update_bits(adv7511->regmap, ADV7511_REG_POWER,
+			   ADV7511_POWER_POWER_DOWN, ADV7511_POWER_POWER_DOWN);
+
+	adv7511->current_edid_segment = -1;
+
+	i2c_set_clientdata(i2c, adv7511);
+
+	adv7511_set_link_config(adv7511, &link_config);
+
+	return 0;
+
+err_i2c_unregister_device:
+	i2c_unregister_device(adv7511->i2c_edid);
+
+	return ret;
+}
+
+static int adv7511_remove(struct i2c_client *i2c)
+{
+	struct adv7511 *adv7511 = i2c_get_clientdata(i2c);
+
+	i2c_unregister_device(adv7511->i2c_edid);
+
+	kfree(adv7511->edid);
+
+	return 0;
+}
+
+static int adv7511_encoder_init(struct i2c_client *i2c, struct drm_device *dev,
+				struct drm_encoder_slave *encoder)
+{
+
+	struct adv7511 *adv7511 = i2c_get_clientdata(i2c);
+
+	encoder->slave_priv = adv7511;
+	encoder->slave_funcs = &adv7511_encoder_funcs;
+
+	adv7511->encoder = &encoder->base;
+
+	return 0;
+}
+
+static const struct i2c_device_id adv7511_i2c_ids[] = {
+	{ "adv7511", 0 },
+	{ "adv7511w", 0 },
+	{ "adv7513", 0 },
+	{ }
+};
+MODULE_DEVICE_TABLE(i2c, adv7511_i2c_ids);
+
+static const struct of_device_id adv7511_of_ids[] = {
+	{ .compatible = "adi,adv7511", },
+	{ .compatible = "adi,adv7511w", },
+	{ .compatible = "adi,adv7513", },
+	{ }
+};
+MODULE_DEVICE_TABLE(of, adv7511_of_ids);
+
+static struct drm_i2c_encoder_driver adv7511_driver = {
+	.i2c_driver = {
+		.driver = {
+			.name = "adv7511",
+			.of_match_table = adv7511_of_ids,
+		},
+		.id_table = adv7511_i2c_ids,
+		.probe = adv7511_probe,
+		.remove = adv7511_remove,
+	},
+
+	.encoder_init = adv7511_encoder_init,
+};
+
+static int __init adv7511_init(void)
+{
+	return drm_i2c_encoder_register(THIS_MODULE, &adv7511_driver);
+}
+module_init(adv7511_init);
+
+static void __exit adv7511_exit(void)
+{
+	drm_i2c_encoder_unregister(&adv7511_driver);
+}
+module_exit(adv7511_exit);
+
+MODULE_AUTHOR("Lars-Peter Clausen <lars@metafoo.de>");
+MODULE_DESCRIPTION("ADV7511 HDMI transmitter driver");
+MODULE_LICENSE("GPL");
diff --git a/drivers/gpu/drm/i2c/adv7511.h b/drivers/gpu/drm/i2c/adv7511.h
new file mode 100644
index 000000000000..6599ed538426
--- /dev/null
+++ b/drivers/gpu/drm/i2c/adv7511.h
@@ -0,0 +1,289 @@
+/*
+ * Analog Devices ADV7511 HDMI transmitter driver
+ *
+ * Copyright 2012 Analog Devices Inc.
+ *
+ * Licensed under the GPL-2.
+ */
+
+#ifndef __DRM_I2C_ADV7511_H__
+#define __DRM_I2C_ADV7511_H__
+
+#include <linux/hdmi.h>
+
+#define ADV7511_REG_CHIP_REVISION		0x00
+#define ADV7511_REG_N0				0x01
+#define ADV7511_REG_N1				0x02
+#define ADV7511_REG_N2				0x03
+#define ADV7511_REG_SPDIF_FREQ			0x04
+#define ADV7511_REG_CTS_AUTOMATIC1		0x05
+#define ADV7511_REG_CTS_AUTOMATIC2		0x06
+#define ADV7511_REG_CTS_MANUAL0			0x07
+#define ADV7511_REG_CTS_MANUAL1			0x08
+#define ADV7511_REG_CTS_MANUAL2			0x09
+#define ADV7511_REG_AUDIO_SOURCE		0x0a
+#define ADV7511_REG_AUDIO_CONFIG		0x0b
+#define ADV7511_REG_I2S_CONFIG			0x0c
+#define ADV7511_REG_I2S_WIDTH			0x0d
+#define ADV7511_REG_AUDIO_SUB_SRC0		0x0e
+#define ADV7511_REG_AUDIO_SUB_SRC1		0x0f
+#define ADV7511_REG_AUDIO_SUB_SRC2		0x10
+#define ADV7511_REG_AUDIO_SUB_SRC3		0x11
+#define ADV7511_REG_AUDIO_CFG1			0x12
+#define ADV7511_REG_AUDIO_CFG2			0x13
+#define ADV7511_REG_AUDIO_CFG3			0x14
+#define ADV7511_REG_I2C_FREQ_ID_CFG		0x15
+#define ADV7511_REG_VIDEO_INPUT_CFG1		0x16
+#define ADV7511_REG_CSC_UPPER(x)		(0x18 + (x) * 2)
+#define ADV7511_REG_CSC_LOWER(x)		(0x19 + (x) * 2)
+#define ADV7511_REG_SYNC_DECODER(x)		(0x30 + (x))
+#define ADV7511_REG_DE_GENERATOR		(0x35 + (x))
+#define ADV7511_REG_PIXEL_REPETITION		0x3b
+#define ADV7511_REG_VIC_MANUAL			0x3c
+#define ADV7511_REG_VIC_SEND			0x3d
+#define ADV7511_REG_VIC_DETECTED		0x3e
+#define ADV7511_REG_AUX_VIC_DETECTED		0x3f
+#define ADV7511_REG_PACKET_ENABLE0		0x40
+#define ADV7511_REG_POWER			0x41
+#define ADV7511_REG_STATUS			0x42
+#define ADV7511_REG_EDID_I2C_ADDR		0x43
+#define ADV7511_REG_PACKET_ENABLE1		0x44
+#define ADV7511_REG_PACKET_I2C_ADDR		0x45
+#define ADV7511_REG_DSD_ENABLE			0x46
+#define ADV7511_REG_VIDEO_INPUT_CFG2		0x48
+#define ADV7511_REG_INFOFRAME_UPDATE		0x4a
+#define ADV7511_REG_GC(x)			(0x4b + (x)) /* 0x4b - 0x51 */
+#define ADV7511_REG_AVI_INFOFRAME_VERSION	0x52
+#define ADV7511_REG_AVI_INFOFRAME_LENGTH	0x53
+#define ADV7511_REG_AVI_INFOFRAME_CHECKSUM	0x54
+#define ADV7511_REG_AVI_INFOFRAME(x)		(0x55 + (x)) /* 0x55 - 0x6f */
+#define ADV7511_REG_AUDIO_INFOFRAME_VERSION	0x70
+#define ADV7511_REG_AUDIO_INFOFRAME_LENGTH	0x71
+#define ADV7511_REG_AUDIO_INFOFRAME_CHECKSUM	0x72
+#define ADV7511_REG_AUDIO_INFOFRAME(x)		(0x73 + (x)) /* 0x73 - 0x7c */
+#define ADV7511_REG_INT_ENABLE(x)		(0x94 + (x))
+#define ADV7511_REG_INT(x)			(0x96 + (x))
+#define ADV7511_REG_INPUT_CLK_DIV		0x9d
+#define ADV7511_REG_PLL_STATUS			0x9e
+#define ADV7511_REG_HDMI_POWER			0xa1
+#define ADV7511_REG_HDCP_HDMI_CFG		0xaf
+#define ADV7511_REG_AN(x)			(0xb0 + (x)) /* 0xb0 - 0xb7 */
+#define ADV7511_REG_HDCP_STATUS			0xb8
+#define ADV7511_REG_BCAPS			0xbe
+#define ADV7511_REG_BKSV(x)			(0xc0 + (x)) /* 0xc0 - 0xc3 */
+#define ADV7511_REG_EDID_SEGMENT		0xc4
+#define ADV7511_REG_DDC_STATUS			0xc8
+#define ADV7511_REG_EDID_READ_CTRL		0xc9
+#define ADV7511_REG_BSTATUS(x)			(0xca + (x)) /* 0xca - 0xcb */
+#define ADV7511_REG_TIMING_GEN_SEQ		0xd0
+#define ADV7511_REG_POWER2			0xd6
+#define ADV7511_REG_HSYNC_PLACEMENT_MSB		0xfa
+
+#define ADV7511_REG_SYNC_ADJUSTMENT(x)		(0xd7 + (x)) /* 0xd7 - 0xdc */
+#define ADV7511_REG_TMDS_CLOCK_INV		0xde
+#define ADV7511_REG_ARC_CTRL			0xdf
+#define ADV7511_REG_CEC_I2C_ADDR		0xe1
+#define ADV7511_REG_CEC_CTRL			0xe2
+#define ADV7511_REG_CHIP_ID_HIGH		0xf5
+#define ADV7511_REG_CHIP_ID_LOW			0xf6
+
+#define ADV7511_CSC_ENABLE			BIT(7)
+#define ADV7511_CSC_UPDATE_MODE			BIT(5)
+
+#define ADV7511_INT0_HDP			BIT(7)
+#define ADV7511_INT0_VSYNC			BIT(5)
+#define ADV7511_INT0_AUDIO_FIFO_FULL		BIT(4)
+#define ADV7511_INT0_EDID_READY			BIT(2)
+#define ADV7511_INT0_HDCP_AUTHENTICATED		BIT(1)
+
+#define ADV7511_INT1_DDC_ERROR			BIT(7)
+#define ADV7511_INT1_BKSV			BIT(6)
+#define ADV7511_INT1_CEC_TX_READY		BIT(5)
+#define ADV7511_INT1_CEC_TX_ARBIT_LOST		BIT(4)
+#define ADV7511_INT1_CEC_TX_RETRY_TIMEOUT	BIT(3)
+#define ADV7511_INT1_CEC_RX_READY3		BIT(2)
+#define ADV7511_INT1_CEC_RX_READY2		BIT(1)
+#define ADV7511_INT1_CEC_RX_READY1		BIT(0)
+
+#define ADV7511_ARC_CTRL_POWER_DOWN		BIT(0)
+
+#define ADV7511_CEC_CTRL_POWER_DOWN		BIT(0)
+
+#define ADV7511_POWER_POWER_DOWN		BIT(6)
+
+#define ADV7511_HDMI_CFG_MODE_MASK		0x2
+#define ADV7511_HDMI_CFG_MODE_DVI		0x0
+#define ADV7511_HDMI_CFG_MODE_HDMI		0x2
+
+#define ADV7511_AUDIO_SELECT_I2C		0x0
+#define ADV7511_AUDIO_SELECT_SPDIF		0x1
+#define ADV7511_AUDIO_SELECT_DSD		0x2
+#define ADV7511_AUDIO_SELECT_HBR		0x3
+#define ADV7511_AUDIO_SELECT_DST		0x4
+
+#define ADV7511_I2S_SAMPLE_LEN_16		0x2
+#define ADV7511_I2S_SAMPLE_LEN_20		0x3
+#define ADV7511_I2S_SAMPLE_LEN_18		0x4
+#define ADV7511_I2S_SAMPLE_LEN_22		0x5
+#define ADV7511_I2S_SAMPLE_LEN_19		0x8
+#define ADV7511_I2S_SAMPLE_LEN_23		0x9
+#define ADV7511_I2S_SAMPLE_LEN_24		0xb
+#define ADV7511_I2S_SAMPLE_LEN_17		0xc
+#define ADV7511_I2S_SAMPLE_LEN_21		0xd
+
+#define ADV7511_SAMPLE_FREQ_44100		0x0
+#define ADV7511_SAMPLE_FREQ_48000		0x2
+#define ADV7511_SAMPLE_FREQ_32000		0x3
+#define ADV7511_SAMPLE_FREQ_88200		0x8
+#define ADV7511_SAMPLE_FREQ_96000		0xa
+#define ADV7511_SAMPLE_FREQ_176400		0xc
+#define ADV7511_SAMPLE_FREQ_192000		0xe
+
+#define ADV7511_STATUS_POWER_DOWN_POLARITY	BIT(7)
+#define ADV7511_STATUS_HPD			BIT(6)
+#define ADV7511_STATUS_MONITOR_SENSE		BIT(5)
+#define ADV7511_STATUS_I2S_32BIT_MODE		BIT(3)
+
+#define ADV7511_PACKET_ENABLE_N_CTS		BIT(8+6)
+#define ADV7511_PACKET_ENABLE_AUDIO_SAMPLE	BIT(8+5)
+#define ADV7511_PACKET_ENABLE_AVI_INFOFRAME	BIT(8+4)
+#define ADV7511_PACKET_ENABLE_AUDIO_INFOFRAME	BIT(8+3)
+#define ADV7511_PACKET_ENABLE_GC		BIT(7)
+#define ADV7511_PACKET_ENABLE_SPD		BIT(6)
+#define ADV7511_PACKET_ENABLE_MPEG		BIT(5)
+#define ADV7511_PACKET_ENABLE_ACP		BIT(4)
+#define ADV7511_PACKET_ENABLE_ISRC		BIT(3)
+#define ADV7511_PACKET_ENABLE_GM		BIT(2)
+#define ADV7511_PACKET_ENABLE_SPARE2		BIT(1)
+#define ADV7511_PACKET_ENABLE_SPARE1		BIT(0)
+
+#define ADV7511_REG_POWER2_HDP_SRC_MASK		0xc0
+#define ADV7511_REG_POWER2_HDP_SRC_BOTH		0x00
+#define ADV7511_REG_POWER2_HDP_SRC_HDP		0x40
+#define ADV7511_REG_POWER2_HDP_SRC_CEC		0x80
+#define ADV7511_REG_POWER2_HDP_SRC_NONE		0xc0
+#define ADV7511_REG_POWER2_TDMS_ENABLE		BIT(4)
+#define ADV7511_REG_POWER2_GATE_INPUT_CLK	BIT(0)
+
+#define ADV7511_LOW_REFRESH_RATE_NONE		0x0
+#define ADV7511_LOW_REFRESH_RATE_24HZ		0x1
+#define ADV7511_LOW_REFRESH_RATE_25HZ		0x2
+#define ADV7511_LOW_REFRESH_RATE_30HZ		0x3
+
+#define ADV7511_AUDIO_CFG3_LEN_MASK		0x0f
+#define ADV7511_I2C_FREQ_ID_CFG_RATE_MASK	0xf0
+
+#define ADV7511_AUDIO_SOURCE_I2S		0
+#define ADV7511_AUDIO_SOURCE_SPDIF		1
+
+#define ADV7511_I2S_FORMAT_I2S			0
+#define ADV7511_I2S_FORMAT_RIGHT_J		1
+#define ADV7511_I2S_FORMAT_LEFT_J		2
+
+#define ADV7511_PACKET(p, x)	    ((p) * 0x20 + (x))
+#define ADV7511_PACKET_SDP(x)	    ADV7511_PACKET(0, x)
+#define ADV7511_PACKET_MPEG(x)	    ADV7511_PACKET(1, x)
+#define ADV7511_PACKET_ACP(x)	    ADV7511_PACKET(2, x)
+#define ADV7511_PACKET_ISRC1(x)	    ADV7511_PACKET(3, x)
+#define ADV7511_PACKET_ISRC2(x)	    ADV7511_PACKET(4, x)
+#define ADV7511_PACKET_GM(x)	    ADV7511_PACKET(5, x)
+#define ADV7511_PACKET_SPARE(x)	    ADV7511_PACKET(6, x)
+
+enum adv7511_input_clock {
+	ADV7511_INPUT_CLOCK_1X,
+	ADV7511_INPUT_CLOCK_2X,
+	ADV7511_INPUT_CLOCK_DDR,
+};
+
+enum adv7511_input_justification {
+	ADV7511_INPUT_JUSTIFICATION_EVENLY = 0,
+	ADV7511_INPUT_JUSTIFICATION_RIGHT = 1,
+	ADV7511_INPUT_JUSTIFICATION_LEFT = 2,
+};
+
+enum adv7511_input_sync_pulse {
+	ADV7511_INPUT_SYNC_PULSE_DE = 0,
+	ADV7511_INPUT_SYNC_PULSE_HSYNC = 1,
+	ADV7511_INPUT_SYNC_PULSE_VSYNC = 2,
+	ADV7511_INPUT_SYNC_PULSE_NONE = 3,
+};
+
+/**
+ * enum adv7511_sync_polarity - Polarity for the input sync signals
+ * @ADV7511_SYNC_POLARITY_PASSTHROUGH:  Sync polarity matches that of
+ *				       the currently configured mode.
+ * @ADV7511_SYNC_POLARITY_LOW:	    Sync polarity is low
+ * @ADV7511_SYNC_POLARITY_HIGH:	    Sync polarity is high
+ *
+ * If the polarity is set to either LOW or HIGH the driver will configure the
+ * ADV7511 to internally invert the sync signal if required to match the sync
+ * polarity setting for the currently selected output mode.
+ *
+ * If the polarity is set to PASSTHROUGH, the ADV7511 will route the signal
+ * unchanged. This is used when the upstream graphics core already generates
+ * the sync signals with the correct polarity.
+ */
+enum adv7511_sync_polarity {
+	ADV7511_SYNC_POLARITY_PASSTHROUGH,
+	ADV7511_SYNC_POLARITY_LOW,
+	ADV7511_SYNC_POLARITY_HIGH,
+};
+
+/**
+ * struct adv7511_link_config - Describes adv7511 hardware configuration
+ * @input_color_depth:		Number of bits per color component (8, 10 or 12)
+ * @input_colorspace:		The input colorspace (RGB, YUV444, YUV422)
+ * @input_clock:		The input video clock style (1x, 2x, DDR)
+ * @input_style:		The input component arrangement variant
+ * @input_justification:	Video input format bit justification
+ * @clock_delay:		Clock delay for the input clock (in ps)
+ * @embedded_sync:		Video input uses BT.656-style embedded sync
+ * @sync_pulse:			Select the sync pulse
+ * @vsync_polarity:		vsync input signal configuration
+ * @hsync_polarity:		hsync input signal configuration
+ */
+struct adv7511_link_config {
+	unsigned int input_color_depth;
+	enum hdmi_colorspace input_colorspace;
+	enum adv7511_input_clock input_clock;
+	unsigned int input_style;
+	enum adv7511_input_justification input_justification;
+
+	int clock_delay;
+
+	bool embedded_sync;
+	enum adv7511_input_sync_pulse sync_pulse;
+	enum adv7511_sync_polarity vsync_polarity;
+	enum adv7511_sync_polarity hsync_polarity;
+};
+
+/**
+ * enum adv7511_csc_scaling - Scaling factor for the ADV7511 CSC
+ * @ADV7511_CSC_SCALING_1: CSC results are not scaled
+ * @ADV7511_CSC_SCALING_2: CSC results are scaled by a factor of two
+ * @ADV7511_CSC_SCALING_4: CSC results are scalled by a factor of four
+ */
+enum adv7511_csc_scaling {
+	ADV7511_CSC_SCALING_1 = 0,
+	ADV7511_CSC_SCALING_2 = 1,
+	ADV7511_CSC_SCALING_4 = 2,
+};
+
+/**
+ * struct adv7511_video_config - Describes adv7511 hardware configuration
+ * @csc_enable:			Whether to enable color space conversion
+ * @csc_scaling_factor:		Color space conversion scaling factor
+ * @csc_coefficents:		Color space conversion coefficents
+ * @hdmi_mode:			Whether to use HDMI or DVI output mode
+ * @avi_infoframe:		HDMI infoframe
+ */
+struct adv7511_video_config {
+	bool csc_enable;
+	enum adv7511_csc_scaling csc_scaling_factor;
+	const uint16_t *csc_coefficents;
+
+	bool hdmi_mode;
+	struct hdmi_avi_infoframe avi_infoframe;
+};
+
+#endif /* __DRM_I2C_ADV7511_H__ */
diff --git a/drivers/gpu/drm/i810/i810_dma.c b/drivers/gpu/drm/i810/i810_dma.c
index bae897de9468..d91856779beb 100644
--- a/drivers/gpu/drm/i810/i810_dma.c
+++ b/drivers/gpu/drm/i810/i810_dma.c
@@ -213,7 +213,7 @@ static int i810_dma_cleanup(struct drm_device *dev)
 		    (drm_i810_private_t *) dev->dev_private;
 
 		if (dev_priv->ring.virtual_start)
-			drm_core_ioremapfree(&dev_priv->ring.map, dev);
+			drm_legacy_ioremapfree(&dev_priv->ring.map, dev);
 		if (dev_priv->hw_status_page) {
 			pci_free_consistent(dev->pdev, PAGE_SIZE,
 					    dev_priv->hw_status_page,
@@ -227,7 +227,7 @@ static int i810_dma_cleanup(struct drm_device *dev)
 			drm_i810_buf_priv_t *buf_priv = buf->dev_private;
 
 			if (buf_priv->kernel_virtual && buf->total)
-				drm_core_ioremapfree(&buf_priv->map, dev);
+				drm_legacy_ioremapfree(&buf_priv->map, dev);
 		}
 	}
 	return 0;
@@ -306,7 +306,7 @@ static int i810_freelist_init(struct drm_device *dev, drm_i810_private_t *dev_pr
 		buf_priv->map.flags = 0;
 		buf_priv->map.mtrr = 0;
 
-		drm_core_ioremap(&buf_priv->map, dev);
+		drm_legacy_ioremap(&buf_priv->map, dev);
 		buf_priv->kernel_virtual = buf_priv->map.handle;
 
 	}
@@ -334,7 +334,7 @@ static int i810_dma_initialize(struct drm_device *dev,
 		DRM_ERROR("can not find sarea!\n");
 		return -EINVAL;
 	}
-	dev_priv->mmio_map = drm_core_findmap(dev, init->mmio_offset);
+	dev_priv->mmio_map = drm_legacy_findmap(dev, init->mmio_offset);
 	if (!dev_priv->mmio_map) {
 		dev->dev_private = (void *)dev_priv;
 		i810_dma_cleanup(dev);
@@ -342,7 +342,7 @@ static int i810_dma_initialize(struct drm_device *dev,
 		return -EINVAL;
 	}
 	dev->agp_buffer_token = init->buffers_offset;
-	dev->agp_buffer_map = drm_core_findmap(dev, init->buffers_offset);
+	dev->agp_buffer_map = drm_legacy_findmap(dev, init->buffers_offset);
 	if (!dev->agp_buffer_map) {
 		dev->dev_private = (void *)dev_priv;
 		i810_dma_cleanup(dev);
@@ -363,7 +363,7 @@ static int i810_dma_initialize(struct drm_device *dev,
 	dev_priv->ring.map.flags = 0;
 	dev_priv->ring.map.mtrr = 0;
 
-	drm_core_ioremap(&dev_priv->ring.map, dev);
+	drm_legacy_ioremap(&dev_priv->ring.map, dev);
 
 	if (dev_priv->ring.map.handle == NULL) {
 		dev->dev_private = (void *)dev_priv;
@@ -1215,9 +1215,9 @@ void i810_driver_preclose(struct drm_device *dev, struct drm_file *file_priv)
 	}
 
 	if (file_priv->master && file_priv->master->lock.hw_lock) {
-		drm_idlelock_take(&file_priv->master->lock);
+		drm_legacy_idlelock_take(&file_priv->master->lock);
 		i810_driver_reclaim_buffers(dev, file_priv);
-		drm_idlelock_release(&file_priv->master->lock);
+		drm_legacy_idlelock_release(&file_priv->master->lock);
 	} else {
 		/* master disappeared, clean up stuff anyway and hope nothing
 		 * goes wrong */
diff --git a/drivers/gpu/drm/i810/i810_drv.c b/drivers/gpu/drm/i810/i810_drv.c
index 441ccf8f5bdc..44f4a131c8dd 100644
--- a/drivers/gpu/drm/i810/i810_drv.c
+++ b/drivers/gpu/drm/i810/i810_drv.c
@@ -47,7 +47,7 @@ static const struct file_operations i810_driver_fops = {
 	.open = drm_open,
 	.release = drm_release,
 	.unlocked_ioctl = drm_ioctl,
-	.mmap = drm_mmap,
+	.mmap = drm_legacy_mmap,
 	.poll = drm_poll,
 #ifdef CONFIG_COMPAT
 	.compat_ioctl = drm_compat_ioctl,
@@ -63,6 +63,7 @@ static struct drm_driver driver = {
 	.load = i810_driver_load,
 	.lastclose = i810_driver_lastclose,
 	.preclose = i810_driver_preclose,
+	.set_busid = drm_pci_set_busid,
 	.device_is_agp = i810_driver_device_is_agp,
 	.dma_quiescent = i810_driver_dma_quiescent,
 	.ioctls = i810_ioctls,
diff --git a/drivers/gpu/drm/i810/i810_drv.h b/drivers/gpu/drm/i810/i810_drv.h
index d4d16eddd651..93ec5dc4e7d3 100644
--- a/drivers/gpu/drm/i810/i810_drv.h
+++ b/drivers/gpu/drm/i810/i810_drv.h
@@ -32,6 +32,8 @@
 #ifndef _I810_DRV_H_
 #define _I810_DRV_H_
 
+#include <drm/drm_legacy.h>
+
 /* General customization:
  */
 
diff --git a/drivers/gpu/drm/i915/Makefile b/drivers/gpu/drm/i915/Makefile
index 91bd167e1cb7..e4083e41a600 100644
--- a/drivers/gpu/drm/i915/Makefile
+++ b/drivers/gpu/drm/i915/Makefile
@@ -11,7 +11,9 @@ i915-y := i915_drv.o \
 	  i915_params.o \
           i915_suspend.o \
 	  i915_sysfs.o \
-	  intel_pm.o
+	  intel_pm.o \
+	  intel_runtime_pm.o
+
 i915-$(CONFIG_COMPAT)   += i915_ioc32.o
 i915-$(CONFIG_DEBUG_FS) += i915_debugfs.o
 
@@ -31,19 +33,25 @@ i915-y += i915_cmd_parser.o \
 	  i915_gpu_error.o \
 	  i915_irq.o \
 	  i915_trace_points.o \
+	  intel_lrc.o \
 	  intel_ringbuffer.o \
 	  intel_uncore.o
 
 # autogenerated null render state
 i915-y += intel_renderstate_gen6.o \
 	  intel_renderstate_gen7.o \
-	  intel_renderstate_gen8.o
+	  intel_renderstate_gen8.o \
+	  intel_renderstate_gen9.o
 
 # modesetting core code
-i915-y += intel_bios.o \
+i915-y += intel_audio.o \
+	  intel_bios.o \
 	  intel_display.o \
+	  intel_fifo_underrun.o \
+	  intel_frontbuffer.o \
 	  intel_modes.o \
 	  intel_overlay.o \
+	  intel_psr.o \
 	  intel_sideband.o \
 	  intel_sprite.o
 i915-$(CONFIG_ACPI)		+= intel_acpi.o intel_opregion.o
diff --git a/drivers/gpu/drm/i915/dvo_ns2501.c b/drivers/gpu/drm/i915/dvo_ns2501.c
index 74f2af7c2d3e..441630434d34 100644
--- a/drivers/gpu/drm/i915/dvo_ns2501.c
+++ b/drivers/gpu/drm/i915/dvo_ns2501.c
@@ -60,16 +60,297 @@
 
 #define NS2501_REGC 0x0c
 
+enum {
+	MODE_640x480,
+	MODE_800x600,
+	MODE_1024x768,
+};
+
+struct ns2501_reg {
+	 uint8_t offset;
+	 uint8_t value;
+};
+
+/*
+ * Magic values based on what the BIOS on
+ * Fujitsu-Siemens Lifebook S6010 programs (1024x768 panel).
+ */
+static const struct ns2501_reg regs_1024x768[][86] = {
+	[MODE_640x480] = {
+		[0] = { .offset = 0x0a, .value = 0x81, },
+		[1] = { .offset = 0x18, .value = 0x07, },
+		[2] = { .offset = 0x19, .value = 0x00, },
+		[3] = { .offset = 0x1a, .value = 0x00, },
+		[4] = { .offset = 0x1b, .value = 0x11, },
+		[5] = { .offset = 0x1c, .value = 0x54, },
+		[6] = { .offset = 0x1d, .value = 0x03, },
+		[7] = { .offset = 0x1e, .value = 0x02, },
+		[8] = { .offset = 0xf3, .value = 0x90, },
+		[9] = { .offset = 0xf9, .value = 0x00, },
+		[10] = { .offset = 0xc1, .value = 0x90, },
+		[11] = { .offset = 0xc2, .value = 0x00, },
+		[12] = { .offset = 0xc3, .value = 0x0f, },
+		[13] = { .offset = 0xc4, .value = 0x03, },
+		[14] = { .offset = 0xc5, .value = 0x16, },
+		[15] = { .offset = 0xc6, .value = 0x00, },
+		[16] = { .offset = 0xc7, .value = 0x02, },
+		[17] = { .offset = 0xc8, .value = 0x02, },
+		[18] = { .offset = 0xf4, .value = 0x00, },
+		[19] = { .offset = 0x80, .value = 0xff, },
+		[20] = { .offset = 0x81, .value = 0x07, },
+		[21] = { .offset = 0x82, .value = 0x3d, },
+		[22] = { .offset = 0x83, .value = 0x05, },
+		[23] = { .offset = 0x94, .value = 0x00, },
+		[24] = { .offset = 0x95, .value = 0x00, },
+		[25] = { .offset = 0x96, .value = 0x05, },
+		[26] = { .offset = 0x97, .value = 0x00, },
+		[27] = { .offset = 0x9a, .value = 0x88, },
+		[28] = { .offset = 0x9b, .value = 0x00, },
+		[29] = { .offset = 0x98, .value = 0x00, },
+		[30] = { .offset = 0x99, .value = 0x00, },
+		[31] = { .offset = 0xf7, .value = 0x88, },
+		[32] = { .offset = 0xf8, .value = 0x0a, },
+		[33] = { .offset = 0x9c, .value = 0x24, },
+		[34] = { .offset = 0x9d, .value = 0x00, },
+		[35] = { .offset = 0x9e, .value = 0x25, },
+		[36] = { .offset = 0x9f, .value = 0x03, },
+		[37] = { .offset = 0xa0, .value = 0x28, },
+		[38] = { .offset = 0xa1, .value = 0x01, },
+		[39] = { .offset = 0xa2, .value = 0x28, },
+		[40] = { .offset = 0xa3, .value = 0x05, },
+		[41] = { .offset = 0xb6, .value = 0x09, },
+		[42] = { .offset = 0xb8, .value = 0x00, },
+		[43] = { .offset = 0xb9, .value = 0xa0, },
+		[44] = { .offset = 0xba, .value = 0x00, },
+		[45] = { .offset = 0xbb, .value = 0x20, },
+		[46] = { .offset = 0x10, .value = 0x00, },
+		[47] = { .offset = 0x11, .value = 0xa0, },
+		[48] = { .offset = 0x12, .value = 0x02, },
+		[49] = { .offset = 0x20, .value = 0x00, },
+		[50] = { .offset = 0x22, .value = 0x00, },
+		[51] = { .offset = 0x23, .value = 0x00, },
+		[52] = { .offset = 0x24, .value = 0x00, },
+		[53] = { .offset = 0x25, .value = 0x00, },
+		[54] = { .offset = 0x8c, .value = 0x10, },
+		[55] = { .offset = 0x8d, .value = 0x02, },
+		[56] = { .offset = 0x8e, .value = 0x10, },
+		[57] = { .offset = 0x8f, .value = 0x00, },
+		[58] = { .offset = 0x90, .value = 0xff, },
+		[59] = { .offset = 0x91, .value = 0x07, },
+		[60] = { .offset = 0x92, .value = 0xa0, },
+		[61] = { .offset = 0x93, .value = 0x02, },
+		[62] = { .offset = 0xa5, .value = 0x00, },
+		[63] = { .offset = 0xa6, .value = 0x00, },
+		[64] = { .offset = 0xa7, .value = 0x00, },
+		[65] = { .offset = 0xa8, .value = 0x00, },
+		[66] = { .offset = 0xa9, .value = 0x04, },
+		[67] = { .offset = 0xaa, .value = 0x70, },
+		[68] = { .offset = 0xab, .value = 0x4f, },
+		[69] = { .offset = 0xac, .value = 0x00, },
+		[70] = { .offset = 0xa4, .value = 0x84, },
+		[71] = { .offset = 0x7e, .value = 0x18, },
+		[72] = { .offset = 0x84, .value = 0x00, },
+		[73] = { .offset = 0x85, .value = 0x00, },
+		[74] = { .offset = 0x86, .value = 0x00, },
+		[75] = { .offset = 0x87, .value = 0x00, },
+		[76] = { .offset = 0x88, .value = 0x00, },
+		[77] = { .offset = 0x89, .value = 0x00, },
+		[78] = { .offset = 0x8a, .value = 0x00, },
+		[79] = { .offset = 0x8b, .value = 0x00, },
+		[80] = { .offset = 0x26, .value = 0x00, },
+		[81] = { .offset = 0x27, .value = 0x00, },
+		[82] = { .offset = 0xad, .value = 0x00, },
+		[83] = { .offset = 0x08, .value = 0x30, }, /* 0x31 */
+		[84] = { .offset = 0x41, .value = 0x00, },
+		[85] = { .offset = 0xc0, .value = 0x05, },
+	},
+	[MODE_800x600] = {
+		[0] = { .offset = 0x0a, .value = 0x81, },
+		[1] = { .offset = 0x18, .value = 0x07, },
+		[2] = { .offset = 0x19, .value = 0x00, },
+		[3] = { .offset = 0x1a, .value = 0x00, },
+		[4] = { .offset = 0x1b, .value = 0x19, },
+		[5] = { .offset = 0x1c, .value = 0x64, },
+		[6] = { .offset = 0x1d, .value = 0x02, },
+		[7] = { .offset = 0x1e, .value = 0x02, },
+		[8] = { .offset = 0xf3, .value = 0x90, },
+		[9] = { .offset = 0xf9, .value = 0x00, },
+		[10] = { .offset = 0xc1, .value = 0xd7, },
+		[11] = { .offset = 0xc2, .value = 0x00, },
+		[12] = { .offset = 0xc3, .value = 0xf8, },
+		[13] = { .offset = 0xc4, .value = 0x03, },
+		[14] = { .offset = 0xc5, .value = 0x1a, },
+		[15] = { .offset = 0xc6, .value = 0x00, },
+		[16] = { .offset = 0xc7, .value = 0x73, },
+		[17] = { .offset = 0xc8, .value = 0x02, },
+		[18] = { .offset = 0xf4, .value = 0x00, },
+		[19] = { .offset = 0x80, .value = 0x27, },
+		[20] = { .offset = 0x81, .value = 0x03, },
+		[21] = { .offset = 0x82, .value = 0x41, },
+		[22] = { .offset = 0x83, .value = 0x05, },
+		[23] = { .offset = 0x94, .value = 0x00, },
+		[24] = { .offset = 0x95, .value = 0x00, },
+		[25] = { .offset = 0x96, .value = 0x05, },
+		[26] = { .offset = 0x97, .value = 0x00, },
+		[27] = { .offset = 0x9a, .value = 0x88, },
+		[28] = { .offset = 0x9b, .value = 0x00, },
+		[29] = { .offset = 0x98, .value = 0x00, },
+		[30] = { .offset = 0x99, .value = 0x00, },
+		[31] = { .offset = 0xf7, .value = 0x88, },
+		[32] = { .offset = 0xf8, .value = 0x06, },
+		[33] = { .offset = 0x9c, .value = 0x23, },
+		[34] = { .offset = 0x9d, .value = 0x00, },
+		[35] = { .offset = 0x9e, .value = 0x25, },
+		[36] = { .offset = 0x9f, .value = 0x03, },
+		[37] = { .offset = 0xa0, .value = 0x28, },
+		[38] = { .offset = 0xa1, .value = 0x01, },
+		[39] = { .offset = 0xa2, .value = 0x28, },
+		[40] = { .offset = 0xa3, .value = 0x05, },
+		[41] = { .offset = 0xb6, .value = 0x09, },
+		[42] = { .offset = 0xb8, .value = 0x30, },
+		[43] = { .offset = 0xb9, .value = 0xc8, },
+		[44] = { .offset = 0xba, .value = 0x00, },
+		[45] = { .offset = 0xbb, .value = 0x20, },
+		[46] = { .offset = 0x10, .value = 0x20, },
+		[47] = { .offset = 0x11, .value = 0xc8, },
+		[48] = { .offset = 0x12, .value = 0x02, },
+		[49] = { .offset = 0x20, .value = 0x00, },
+		[50] = { .offset = 0x22, .value = 0x00, },
+		[51] = { .offset = 0x23, .value = 0x00, },
+		[52] = { .offset = 0x24, .value = 0x00, },
+		[53] = { .offset = 0x25, .value = 0x00, },
+		[54] = { .offset = 0x8c, .value = 0x10, },
+		[55] = { .offset = 0x8d, .value = 0x02, },
+		[56] = { .offset = 0x8e, .value = 0x04, },
+		[57] = { .offset = 0x8f, .value = 0x00, },
+		[58] = { .offset = 0x90, .value = 0xff, },
+		[59] = { .offset = 0x91, .value = 0x07, },
+		[60] = { .offset = 0x92, .value = 0xa0, },
+		[61] = { .offset = 0x93, .value = 0x02, },
+		[62] = { .offset = 0xa5, .value = 0x00, },
+		[63] = { .offset = 0xa6, .value = 0x00, },
+		[64] = { .offset = 0xa7, .value = 0x00, },
+		[65] = { .offset = 0xa8, .value = 0x00, },
+		[66] = { .offset = 0xa9, .value = 0x83, },
+		[67] = { .offset = 0xaa, .value = 0x40, },
+		[68] = { .offset = 0xab, .value = 0x32, },
+		[69] = { .offset = 0xac, .value = 0x00, },
+		[70] = { .offset = 0xa4, .value = 0x80, },
+		[71] = { .offset = 0x7e, .value = 0x18, },
+		[72] = { .offset = 0x84, .value = 0x00, },
+		[73] = { .offset = 0x85, .value = 0x00, },
+		[74] = { .offset = 0x86, .value = 0x00, },
+		[75] = { .offset = 0x87, .value = 0x00, },
+		[76] = { .offset = 0x88, .value = 0x00, },
+		[77] = { .offset = 0x89, .value = 0x00, },
+		[78] = { .offset = 0x8a, .value = 0x00, },
+		[79] = { .offset = 0x8b, .value = 0x00, },
+		[80] = { .offset = 0x26, .value = 0x00, },
+		[81] = { .offset = 0x27, .value = 0x00, },
+		[82] = { .offset = 0xad, .value = 0x00, },
+		[83] = { .offset = 0x08, .value = 0x30, }, /* 0x31 */
+		[84] = { .offset = 0x41, .value = 0x00, },
+		[85] = { .offset = 0xc0, .value = 0x07, },
+	},
+	[MODE_1024x768] = {
+		[0] = { .offset = 0x0a, .value = 0x81, },
+		[1] = { .offset = 0x18, .value = 0x07, },
+		[2] = { .offset = 0x19, .value = 0x00, },
+		[3] = { .offset = 0x1a, .value = 0x00, },
+		[4] = { .offset = 0x1b, .value = 0x11, },
+		[5] = { .offset = 0x1c, .value = 0x54, },
+		[6] = { .offset = 0x1d, .value = 0x03, },
+		[7] = { .offset = 0x1e, .value = 0x02, },
+		[8] = { .offset = 0xf3, .value = 0x90, },
+		[9] = { .offset = 0xf9, .value = 0x00, },
+		[10] = { .offset = 0xc1, .value = 0x90, },
+		[11] = { .offset = 0xc2, .value = 0x00, },
+		[12] = { .offset = 0xc3, .value = 0x0f, },
+		[13] = { .offset = 0xc4, .value = 0x03, },
+		[14] = { .offset = 0xc5, .value = 0x16, },
+		[15] = { .offset = 0xc6, .value = 0x00, },
+		[16] = { .offset = 0xc7, .value = 0x02, },
+		[17] = { .offset = 0xc8, .value = 0x02, },
+		[18] = { .offset = 0xf4, .value = 0x00, },
+		[19] = { .offset = 0x80, .value = 0xff, },
+		[20] = { .offset = 0x81, .value = 0x07, },
+		[21] = { .offset = 0x82, .value = 0x3d, },
+		[22] = { .offset = 0x83, .value = 0x05, },
+		[23] = { .offset = 0x94, .value = 0x00, },
+		[24] = { .offset = 0x95, .value = 0x00, },
+		[25] = { .offset = 0x96, .value = 0x05, },
+		[26] = { .offset = 0x97, .value = 0x00, },
+		[27] = { .offset = 0x9a, .value = 0x88, },
+		[28] = { .offset = 0x9b, .value = 0x00, },
+		[29] = { .offset = 0x98, .value = 0x00, },
+		[30] = { .offset = 0x99, .value = 0x00, },
+		[31] = { .offset = 0xf7, .value = 0x88, },
+		[32] = { .offset = 0xf8, .value = 0x0a, },
+		[33] = { .offset = 0x9c, .value = 0x24, },
+		[34] = { .offset = 0x9d, .value = 0x00, },
+		[35] = { .offset = 0x9e, .value = 0x25, },
+		[36] = { .offset = 0x9f, .value = 0x03, },
+		[37] = { .offset = 0xa0, .value = 0x28, },
+		[38] = { .offset = 0xa1, .value = 0x01, },
+		[39] = { .offset = 0xa2, .value = 0x28, },
+		[40] = { .offset = 0xa3, .value = 0x05, },
+		[41] = { .offset = 0xb6, .value = 0x09, },
+		[42] = { .offset = 0xb8, .value = 0x00, },
+		[43] = { .offset = 0xb9, .value = 0xa0, },
+		[44] = { .offset = 0xba, .value = 0x00, },
+		[45] = { .offset = 0xbb, .value = 0x20, },
+		[46] = { .offset = 0x10, .value = 0x00, },
+		[47] = { .offset = 0x11, .value = 0xa0, },
+		[48] = { .offset = 0x12, .value = 0x02, },
+		[49] = { .offset = 0x20, .value = 0x00, },
+		[50] = { .offset = 0x22, .value = 0x00, },
+		[51] = { .offset = 0x23, .value = 0x00, },
+		[52] = { .offset = 0x24, .value = 0x00, },
+		[53] = { .offset = 0x25, .value = 0x00, },
+		[54] = { .offset = 0x8c, .value = 0x10, },
+		[55] = { .offset = 0x8d, .value = 0x02, },
+		[56] = { .offset = 0x8e, .value = 0x10, },
+		[57] = { .offset = 0x8f, .value = 0x00, },
+		[58] = { .offset = 0x90, .value = 0xff, },
+		[59] = { .offset = 0x91, .value = 0x07, },
+		[60] = { .offset = 0x92, .value = 0xa0, },
+		[61] = { .offset = 0x93, .value = 0x02, },
+		[62] = { .offset = 0xa5, .value = 0x00, },
+		[63] = { .offset = 0xa6, .value = 0x00, },
+		[64] = { .offset = 0xa7, .value = 0x00, },
+		[65] = { .offset = 0xa8, .value = 0x00, },
+		[66] = { .offset = 0xa9, .value = 0x04, },
+		[67] = { .offset = 0xaa, .value = 0x70, },
+		[68] = { .offset = 0xab, .value = 0x4f, },
+		[69] = { .offset = 0xac, .value = 0x00, },
+		[70] = { .offset = 0xa4, .value = 0x84, },
+		[71] = { .offset = 0x7e, .value = 0x18, },
+		[72] = { .offset = 0x84, .value = 0x00, },
+		[73] = { .offset = 0x85, .value = 0x00, },
+		[74] = { .offset = 0x86, .value = 0x00, },
+		[75] = { .offset = 0x87, .value = 0x00, },
+		[76] = { .offset = 0x88, .value = 0x00, },
+		[77] = { .offset = 0x89, .value = 0x00, },
+		[78] = { .offset = 0x8a, .value = 0x00, },
+		[79] = { .offset = 0x8b, .value = 0x00, },
+		[80] = { .offset = 0x26, .value = 0x00, },
+		[81] = { .offset = 0x27, .value = 0x00, },
+		[82] = { .offset = 0xad, .value = 0x00, },
+		[83] = { .offset = 0x08, .value = 0x34, }, /* 0x35 */
+		[84] = { .offset = 0x41, .value = 0x00, },
+		[85] = { .offset = 0xc0, .value = 0x01, },
+	},
+};
+
+static const struct ns2501_reg regs_init[] = {
+	[0] = { .offset = 0x35, .value = 0xff, },
+	[1] = { .offset = 0x34, .value = 0x00, },
+	[2] = { .offset = 0x08, .value = 0x30, },
+};
+
 struct ns2501_priv {
-	//I2CDevRec d;
 	bool quiet;
-	int reg_8_shadow;
-	int reg_8_set;
-	// Shadow registers for i915
-	int dvoc;
-	int pll_a;
-	int srcdim;
-	int fw_blc;
+	const struct ns2501_reg *regs;
 };
 
 #define NSPTR(d) ((NS2501Ptr)(d->DriverPrivate.ptr))
@@ -205,11 +486,9 @@ static bool ns2501_init(struct intel_dvo_device *dvo,
 		goto out;
 	}
 	ns->quiet = false;
-	ns->reg_8_set = 0;
-	ns->reg_8_shadow =
-	    NS2501_8_PD | NS2501_8_BPAS | NS2501_8_VEN | NS2501_8_HEN;
 
 	DRM_DEBUG_KMS("init ns2501 dvo controller successfully!\n");
+
 	return true;
 
 out:
@@ -242,9 +521,9 @@ static enum drm_mode_status ns2501_mode_valid(struct intel_dvo_device *dvo,
 	 * of the panel in here so we could always accept it
 	 * by disabling the scaler.
 	 */
-	if ((mode->hdisplay == 800 && mode->vdisplay == 600) ||
-	    (mode->hdisplay == 640 && mode->vdisplay == 480) ||
-	    (mode->hdisplay == 1024 && mode->vdisplay == 768)) {
+	if ((mode->hdisplay == 640 && mode->vdisplay == 480 && mode->clock == 25175) ||
+	    (mode->hdisplay == 800 && mode->vdisplay == 600 && mode->clock == 40000) ||
+	    (mode->hdisplay == 1024 && mode->vdisplay == 768 && mode->clock == 65000)) {
 		return MODE_OK;
 	} else {
 		return MODE_ONE_SIZE;	/* Is this a reasonable error? */
@@ -255,180 +534,30 @@ static void ns2501_mode_set(struct intel_dvo_device *dvo,
 			    struct drm_display_mode *mode,
 			    struct drm_display_mode *adjusted_mode)
 {
-	bool ok;
-	int retries = 10;
 	struct ns2501_priv *ns = (struct ns2501_priv *)(dvo->dev_priv);
+	int mode_idx, i;
 
 	DRM_DEBUG_KMS
 	    ("set mode (hdisplay=%d,htotal=%d,vdisplay=%d,vtotal=%d).\n",
 	     mode->hdisplay, mode->htotal, mode->vdisplay, mode->vtotal);
 
-	/*
-	 * Where do I find the native resolution for which scaling is not required???
-	 *
-	 * First trigger the DVO on as otherwise the chip does not appear on the i2c
-	 * bus.
-	 */
-	do {
-		ok = true;
-
-		if (mode->hdisplay == 800 && mode->vdisplay == 600) {
-			/* mode 277 */
-			ns->reg_8_shadow &= ~NS2501_8_BPAS;
-			DRM_DEBUG_KMS("switching to 800x600\n");
-
-			/*
-			 * No, I do not know where this data comes from.
-			 * It is just what the video bios left in the DVO, so
-			 * I'm just copying it here over.
-			 * This also means that I cannot support any other modes
-			 * except the ones supported by the bios.
-			 */
-			ok &= ns2501_writeb(dvo, 0x11, 0xc8);	// 0xc7 also works.
-			ok &= ns2501_writeb(dvo, 0x1b, 0x19);
-			ok &= ns2501_writeb(dvo, 0x1c, 0x62);	// VBIOS left 0x64 here, but 0x62 works nicer
-			ok &= ns2501_writeb(dvo, 0x1d, 0x02);
-
-			ok &= ns2501_writeb(dvo, 0x34, 0x03);
-			ok &= ns2501_writeb(dvo, 0x35, 0xff);
+	if (mode->hdisplay == 640 && mode->vdisplay == 480)
+		mode_idx = MODE_640x480;
+	else if (mode->hdisplay == 800 && mode->vdisplay == 600)
+		mode_idx = MODE_800x600;
+	else if (mode->hdisplay == 1024 && mode->vdisplay == 768)
+		mode_idx = MODE_1024x768;
+	else
+		return;
 
-			ok &= ns2501_writeb(dvo, 0x80, 0x27);
-			ok &= ns2501_writeb(dvo, 0x81, 0x03);
-			ok &= ns2501_writeb(dvo, 0x82, 0x41);
-			ok &= ns2501_writeb(dvo, 0x83, 0x05);
+	/* Hopefully doing it every time won't hurt... */
+	for (i = 0; i < ARRAY_SIZE(regs_init); i++)
+		ns2501_writeb(dvo, regs_init[i].offset, regs_init[i].value);
 
-			ok &= ns2501_writeb(dvo, 0x8d, 0x02);
-			ok &= ns2501_writeb(dvo, 0x8e, 0x04);
-			ok &= ns2501_writeb(dvo, 0x8f, 0x00);
+	ns->regs = regs_1024x768[mode_idx];
 
-			ok &= ns2501_writeb(dvo, 0x90, 0xfe);	/* vertical. VBIOS left 0xff here, but 0xfe works better */
-			ok &= ns2501_writeb(dvo, 0x91, 0x07);
-			ok &= ns2501_writeb(dvo, 0x94, 0x00);
-			ok &= ns2501_writeb(dvo, 0x95, 0x00);
-
-			ok &= ns2501_writeb(dvo, 0x96, 0x00);
-
-			ok &= ns2501_writeb(dvo, 0x99, 0x00);
-			ok &= ns2501_writeb(dvo, 0x9a, 0x88);
-
-			ok &= ns2501_writeb(dvo, 0x9c, 0x23);	/* Looks like first and last line of the image. */
-			ok &= ns2501_writeb(dvo, 0x9d, 0x00);
-			ok &= ns2501_writeb(dvo, 0x9e, 0x25);
-			ok &= ns2501_writeb(dvo, 0x9f, 0x03);
-
-			ok &= ns2501_writeb(dvo, 0xa4, 0x80);
-
-			ok &= ns2501_writeb(dvo, 0xb6, 0x00);
-
-			ok &= ns2501_writeb(dvo, 0xb9, 0xc8);	/* horizontal? */
-			ok &= ns2501_writeb(dvo, 0xba, 0x00);	/* horizontal? */
-
-			ok &= ns2501_writeb(dvo, 0xc0, 0x05);	/* horizontal? */
-			ok &= ns2501_writeb(dvo, 0xc1, 0xd7);
-
-			ok &= ns2501_writeb(dvo, 0xc2, 0x00);
-			ok &= ns2501_writeb(dvo, 0xc3, 0xf8);
-
-			ok &= ns2501_writeb(dvo, 0xc4, 0x03);
-			ok &= ns2501_writeb(dvo, 0xc5, 0x1a);
-
-			ok &= ns2501_writeb(dvo, 0xc6, 0x00);
-			ok &= ns2501_writeb(dvo, 0xc7, 0x73);
-			ok &= ns2501_writeb(dvo, 0xc8, 0x02);
-
-		} else if (mode->hdisplay == 640 && mode->vdisplay == 480) {
-			/* mode 274 */
-			DRM_DEBUG_KMS("switching to 640x480\n");
-			/*
-			 * No, I do not know where this data comes from.
-			 * It is just what the video bios left in the DVO, so
-			 * I'm just copying it here over.
-			 * This also means that I cannot support any other modes
-			 * except the ones supported by the bios.
-			 */
-			ns->reg_8_shadow &= ~NS2501_8_BPAS;
-
-			ok &= ns2501_writeb(dvo, 0x11, 0xa0);
-			ok &= ns2501_writeb(dvo, 0x1b, 0x11);
-			ok &= ns2501_writeb(dvo, 0x1c, 0x54);
-			ok &= ns2501_writeb(dvo, 0x1d, 0x03);
-
-			ok &= ns2501_writeb(dvo, 0x34, 0x03);
-			ok &= ns2501_writeb(dvo, 0x35, 0xff);
-
-			ok &= ns2501_writeb(dvo, 0x80, 0xff);
-			ok &= ns2501_writeb(dvo, 0x81, 0x07);
-			ok &= ns2501_writeb(dvo, 0x82, 0x3d);
-			ok &= ns2501_writeb(dvo, 0x83, 0x05);
-
-			ok &= ns2501_writeb(dvo, 0x8d, 0x02);
-			ok &= ns2501_writeb(dvo, 0x8e, 0x10);
-			ok &= ns2501_writeb(dvo, 0x8f, 0x00);
-
-			ok &= ns2501_writeb(dvo, 0x90, 0xff);	/* vertical */
-			ok &= ns2501_writeb(dvo, 0x91, 0x07);
-			ok &= ns2501_writeb(dvo, 0x94, 0x00);
-			ok &= ns2501_writeb(dvo, 0x95, 0x00);
-
-			ok &= ns2501_writeb(dvo, 0x96, 0x05);
-
-			ok &= ns2501_writeb(dvo, 0x99, 0x00);
-			ok &= ns2501_writeb(dvo, 0x9a, 0x88);
-
-			ok &= ns2501_writeb(dvo, 0x9c, 0x24);
-			ok &= ns2501_writeb(dvo, 0x9d, 0x00);
-			ok &= ns2501_writeb(dvo, 0x9e, 0x25);
-			ok &= ns2501_writeb(dvo, 0x9f, 0x03);
-
-			ok &= ns2501_writeb(dvo, 0xa4, 0x84);
-
-			ok &= ns2501_writeb(dvo, 0xb6, 0x09);
-
-			ok &= ns2501_writeb(dvo, 0xb9, 0xa0);	/* horizontal? */
-			ok &= ns2501_writeb(dvo, 0xba, 0x00);	/* horizontal? */
-
-			ok &= ns2501_writeb(dvo, 0xc0, 0x05);	/* horizontal? */
-			ok &= ns2501_writeb(dvo, 0xc1, 0x90);
-
-			ok &= ns2501_writeb(dvo, 0xc2, 0x00);
-			ok &= ns2501_writeb(dvo, 0xc3, 0x0f);
-
-			ok &= ns2501_writeb(dvo, 0xc4, 0x03);
-			ok &= ns2501_writeb(dvo, 0xc5, 0x16);
-
-			ok &= ns2501_writeb(dvo, 0xc6, 0x00);
-			ok &= ns2501_writeb(dvo, 0xc7, 0x02);
-			ok &= ns2501_writeb(dvo, 0xc8, 0x02);
-
-		} else if (mode->hdisplay == 1024 && mode->vdisplay == 768) {
-			/* mode 280 */
-			DRM_DEBUG_KMS("switching to 1024x768\n");
-			/*
-			 * This might or might not work, actually. I'm silently
-			 * assuming here that the native panel resolution is
-			 * 1024x768. If not, then this leaves the scaler disabled
-			 * generating a picture that is likely not the expected.
-			 *
-			 * Problem is that I do not know where to take the panel
-			 * dimensions from.
-			 *
-			 * Enable the bypass, scaling not required.
-			 *
-			 * The scaler registers are irrelevant here....
-			 *
-			 */
-			ns->reg_8_shadow |= NS2501_8_BPAS;
-			ok &= ns2501_writeb(dvo, 0x37, 0x44);
-		} else {
-			/*
-			 * Data not known. Bummer!
-			 * Hopefully, the code should not go here
-			 * as mode_OK delivered no other modes.
-			 */
-			ns->reg_8_shadow |= NS2501_8_BPAS;
-		}
-		ok &= ns2501_writeb(dvo, NS2501_REG8, ns->reg_8_shadow);
-	} while (!ok && retries--);
+	for (i = 0; i < 84; i++)
+		ns2501_writeb(dvo, ns->regs[i].offset, ns->regs[i].value);
 }
 
 /* set the NS2501 power state */
@@ -439,60 +568,46 @@ static bool ns2501_get_hw_state(struct intel_dvo_device *dvo)
 	if (!ns2501_readb(dvo, NS2501_REG8, &ch))
 		return false;
 
-	if (ch & NS2501_8_PD)
-		return true;
-	else
-		return false;
+	return ch & NS2501_8_PD;
 }
 
 /* set the NS2501 power state */
 static void ns2501_dpms(struct intel_dvo_device *dvo, bool enable)
 {
-	bool ok;
-	int retries = 10;
 	struct ns2501_priv *ns = (struct ns2501_priv *)(dvo->dev_priv);
-	unsigned char ch;
 
 	DRM_DEBUG_KMS("Trying set the dpms of the DVO to %i\n", enable);
 
-	ch = ns->reg_8_shadow;
+	if (enable) {
+		if (WARN_ON(ns->regs[83].offset != 0x08 ||
+			    ns->regs[84].offset != 0x41 ||
+			    ns->regs[85].offset != 0xc0))
+			return;
 
-	if (enable)
-		ch |= NS2501_8_PD;
-	else
-		ch &= ~NS2501_8_PD;
-
-	if (ns->reg_8_set == 0 || ns->reg_8_shadow != ch) {
-		ns->reg_8_set = 1;
-		ns->reg_8_shadow = ch;
-
-		do {
-			ok = true;
-			ok &= ns2501_writeb(dvo, NS2501_REG8, ch);
-			ok &=
-			    ns2501_writeb(dvo, 0x34,
-					  enable ? 0x03 : 0x00);
-			ok &=
-			    ns2501_writeb(dvo, 0x35,
-					  enable ? 0xff : 0x00);
-		} while (!ok && retries--);
-	}
-}
+		ns2501_writeb(dvo, 0xc0, ns->regs[85].value | 0x08);
 
-static void ns2501_dump_regs(struct intel_dvo_device *dvo)
-{
-	uint8_t val;
-
-	ns2501_readb(dvo, NS2501_FREQ_LO, &val);
-	DRM_DEBUG_KMS("NS2501_FREQ_LO: 0x%02x\n", val);
-	ns2501_readb(dvo, NS2501_FREQ_HI, &val);
-	DRM_DEBUG_KMS("NS2501_FREQ_HI: 0x%02x\n", val);
-	ns2501_readb(dvo, NS2501_REG8, &val);
-	DRM_DEBUG_KMS("NS2501_REG8: 0x%02x\n", val);
-	ns2501_readb(dvo, NS2501_REG9, &val);
-	DRM_DEBUG_KMS("NS2501_REG9: 0x%02x\n", val);
-	ns2501_readb(dvo, NS2501_REGC, &val);
-	DRM_DEBUG_KMS("NS2501_REGC: 0x%02x\n", val);
+		ns2501_writeb(dvo, 0x41, ns->regs[84].value);
+
+		ns2501_writeb(dvo, 0x34, 0x01);
+		msleep(15);
+
+		ns2501_writeb(dvo, 0x08, 0x35);
+		if (!(ns->regs[83].value & NS2501_8_BPAS))
+			ns2501_writeb(dvo, 0x08, 0x31);
+		msleep(200);
+
+		ns2501_writeb(dvo, 0x34, 0x03);
+
+		ns2501_writeb(dvo, 0xc0, ns->regs[85].value);
+	} else {
+		ns2501_writeb(dvo, 0x34, 0x01);
+		msleep(200);
+
+		ns2501_writeb(dvo, 0x08, 0x34);
+		msleep(15);
+
+		ns2501_writeb(dvo, 0x34, 0x00);
+	}
 }
 
 static void ns2501_destroy(struct intel_dvo_device *dvo)
@@ -512,6 +627,5 @@ struct intel_dvo_dev_ops ns2501_ops = {
 	.mode_set = ns2501_mode_set,
 	.dpms = ns2501_dpms,
 	.get_hw_state = ns2501_get_hw_state,
-	.dump_regs = ns2501_dump_regs,
 	.destroy = ns2501_destroy,
 };
diff --git a/drivers/gpu/drm/i915/i915_cmd_parser.c b/drivers/gpu/drm/i915/i915_cmd_parser.c
index 4b7ed5289217..22c992a78ac6 100644
--- a/drivers/gpu/drm/i915/i915_cmd_parser.c
+++ b/drivers/gpu/drm/i915/i915_cmd_parser.c
@@ -73,7 +73,7 @@
  * those commands required by the parser. This generally works because command
  * opcode ranges have standard command length encodings. So for commands that
  * the parser does not need to check, it can easily skip them. This is
- * implementated via a per-ring length decoding vfunc.
+ * implemented via a per-ring length decoding vfunc.
  *
  * Unfortunately, there are a number of commands that do not follow the standard
  * length encoding for their opcode range, primarily amongst the MI_* commands.
@@ -138,6 +138,11 @@ static const struct drm_i915_cmd_descriptor common_cmds[] = {
 			.mask = MI_GLOBAL_GTT,
 			.expected = 0,
 	      }},						       ),
+	/*
+	 * MI_BATCH_BUFFER_START requires some special handling. It's not
+	 * really a 'skip' action but it doesn't seem like it's worth adding
+	 * a new action. See i915_parse_cmds().
+	 */
 	CMD(  MI_BATCH_BUFFER_START,            SMI,   !F,  0xFF,   S  ),
 };
 
@@ -408,6 +413,8 @@ static const u32 gen7_render_regs[] = {
 	REG64(PS_INVOCATION_COUNT),
 	REG64(PS_DEPTH_COUNT),
 	OACONTROL, /* Only allowed for LRI and SRM. See below. */
+	REG64(MI_PREDICATE_SRC0),
+	REG64(MI_PREDICATE_SRC1),
 	GEN7_3DPRIM_END_OFFSET,
 	GEN7_3DPRIM_START_VERTEX,
 	GEN7_3DPRIM_VERTEX_COUNT,
@@ -838,23 +845,16 @@ finish:
  * @ring: the ring in question
  *
  * Only certain platforms require software batch buffer command parsing, and
- * only when enabled via module paramter.
+ * only when enabled via module parameter.
  *
  * Return: true if the ring requires software command parsing
  */
 bool i915_needs_cmd_parser(struct intel_engine_cs *ring)
 {
-	struct drm_i915_private *dev_priv = ring->dev->dev_private;
-
 	if (!ring->needs_cmd_parser)
 		return false;
 
-	/*
-	 * XXX: VLV is Gen7 and therefore has cmd_tables, but has PPGTT
-	 * disabled. That will cause all of the parser's PPGTT checks to
-	 * fail. For now, disable parsing when PPGTT is off.
-	 */
-	if (!dev_priv->mm.aliasing_ppgtt)
+	if (!USES_PPGTT(ring->dev))
 		return false;
 
 	return (i915.enable_cmd_parser == 1);
@@ -890,8 +890,10 @@ static bool check_cmd(const struct intel_engine_cs *ring,
 		 * OACONTROL writes to only MI_LOAD_REGISTER_IMM commands.
 		 */
 		if (reg_addr == OACONTROL) {
-			if (desc->cmd.value == MI_LOAD_REGISTER_MEM)
+			if (desc->cmd.value == MI_LOAD_REGISTER_MEM) {
+				DRM_DEBUG_DRIVER("CMD: Rejected LRM to OACONTROL\n");
 				return false;
+			}
 
 			if (desc->cmd.value == MI_LOAD_REGISTER_IMM(1))
 				*oacontrol_set = (cmd[2] != 0);
@@ -960,7 +962,8 @@ static bool check_cmd(const struct intel_engine_cs *ring,
  * Parses the specified batch buffer looking for privilege violations as
  * described in the overview.
  *
- * Return: non-zero if the parser finds violations or otherwise fails
+ * Return: non-zero if the parser finds violations or otherwise fails; -EACCES
+ * if the batch appears legal but should use hardware parsing
  */
 int i915_parse_cmds(struct intel_engine_cs *ring,
 		    struct drm_i915_gem_object *batch_obj,
@@ -1007,6 +1010,16 @@ int i915_parse_cmds(struct intel_engine_cs *ring,
 			break;
 		}
 
+		/*
+		 * If the batch buffer contains a chained batch, return an
+		 * error that tells the caller to abort and dispatch the
+		 * workload as a non-secure batch.
+		 */
+		if (desc->cmd.value == MI_BATCH_BUFFER_START) {
+			ret = -EACCES;
+			break;
+		}
+
 		if (desc->flags & CMD_DESC_FIXED)
 			length = desc->length.fixed;
 		else
@@ -1061,6 +1074,8 @@ int i915_cmd_parser_get_version(void)
 	 *
 	 * 1. Initial version. Checks batches and reports violations, but leaves
 	 *    hardware parsing enabled (so does not allow new use cases).
+	 * 2. Allow access to the MI_PREDICATE_SRC0 and
+	 *    MI_PREDICATE_SRC1 registers.
 	 */
-	return 1;
+	return 2;
 }
diff --git a/drivers/gpu/drm/i915/i915_debugfs.c b/drivers/gpu/drm/i915/i915_debugfs.c
index 9e737b771c40..779a275eb1fd 100644
--- a/drivers/gpu/drm/i915/i915_debugfs.c
+++ b/drivers/gpu/drm/i915/i915_debugfs.c
@@ -116,7 +116,7 @@ static const char *get_tiling_flag(struct drm_i915_gem_object *obj)
 
 static inline const char *get_global_flag(struct drm_i915_gem_object *obj)
 {
-	return obj->has_global_gtt_mapping ? "g" : " ";
+	return i915_gem_obj_to_ggtt(obj) ? "g" : " ";
 }
 
 static void
@@ -136,7 +136,7 @@ describe_obj(struct seq_file *m, struct drm_i915_gem_object *obj)
 		   obj->last_read_seqno,
 		   obj->last_write_seqno,
 		   obj->last_fenced_seqno,
-		   i915_cache_level_str(obj->cache_level),
+		   i915_cache_level_str(to_i915(obj->base.dev), obj->cache_level),
 		   obj->dirty ? " dirty" : "",
 		   obj->madv == I915_MADV_DONTNEED ? " purgeable" : "");
 	if (obj->base.name)
@@ -333,7 +333,7 @@ static int per_file_stats(int id, void *ptr, void *data)
 			}
 
 			ppgtt = container_of(vma->vm, struct i915_hw_ppgtt, base);
-			if (ppgtt->ctx && ppgtt->ctx->file_priv != stats->file_priv)
+			if (ppgtt->file_priv != stats->file_priv)
 				continue;
 
 			if (obj->ring) /* XXX per-vma statistic */
@@ -515,7 +515,7 @@ static int i915_gem_pageflip_info(struct seq_file *m, void *data)
 {
 	struct drm_info_node *node = m->private;
 	struct drm_device *dev = node->minor->dev;
-	unsigned long flags;
+	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_crtc *crtc;
 	int ret;
 
@@ -528,12 +528,14 @@ static int i915_gem_pageflip_info(struct seq_file *m, void *data)
 		const char plane = plane_name(crtc->plane);
 		struct intel_unpin_work *work;
 
-		spin_lock_irqsave(&dev->event_lock, flags);
+		spin_lock_irq(&dev->event_lock);
 		work = crtc->unpin_work;
 		if (work == NULL) {
 			seq_printf(m, "No flip due on pipe %c (plane %c)\n",
 				   pipe, plane);
 		} else {
+			u32 addr;
+
 			if (atomic_read(&work->pending) < INTEL_FLIP_COMPLETE) {
 				seq_printf(m, "Flip queued on pipe %c (plane %c)\n",
 					   pipe, plane);
@@ -541,26 +543,38 @@ static int i915_gem_pageflip_info(struct seq_file *m, void *data)
 				seq_printf(m, "Flip pending (waiting for vsync) on pipe %c (plane %c)\n",
 					   pipe, plane);
 			}
+			if (work->flip_queued_ring) {
+				seq_printf(m, "Flip queued on %s at seqno %u, next seqno %u [current breadcrumb %u], completed? %d\n",
+					   work->flip_queued_ring->name,
+					   work->flip_queued_seqno,
+					   dev_priv->next_seqno,
+					   work->flip_queued_ring->get_seqno(work->flip_queued_ring, true),
+					   i915_seqno_passed(work->flip_queued_ring->get_seqno(work->flip_queued_ring, true),
+							     work->flip_queued_seqno));
+			} else
+				seq_printf(m, "Flip not associated with any ring\n");
+			seq_printf(m, "Flip queued on frame %d, (was ready on frame %d), now %d\n",
+				   work->flip_queued_vblank,
+				   work->flip_ready_vblank,
+				   drm_vblank_count(dev, crtc->pipe));
 			if (work->enable_stall_check)
 				seq_puts(m, "Stall check enabled, ");
 			else
 				seq_puts(m, "Stall check waiting for page flip ioctl, ");
 			seq_printf(m, "%d prepares\n", atomic_read(&work->pending));
 
-			if (work->old_fb_obj) {
-				struct drm_i915_gem_object *obj = work->old_fb_obj;
-				if (obj)
-					seq_printf(m, "Old framebuffer gtt_offset 0x%08lx\n",
-						   i915_gem_obj_ggtt_offset(obj));
-			}
+			if (INTEL_INFO(dev)->gen >= 4)
+				addr = I915_HI_DISPBASE(I915_READ(DSPSURF(crtc->plane)));
+			else
+				addr = I915_READ(DSPADDR(crtc->plane));
+			seq_printf(m, "Current scanout address 0x%08x\n", addr);
+
 			if (work->pending_flip_obj) {
-				struct drm_i915_gem_object *obj = work->pending_flip_obj;
-				if (obj)
-					seq_printf(m, "New framebuffer gtt_offset 0x%08lx\n",
-						   i915_gem_obj_ggtt_offset(obj));
+				seq_printf(m, "New framebuffer address 0x%08lx\n", (long)work->gtt_offset);
+				seq_printf(m, "MMIO update completed? %d\n",  addr == work->gtt_offset);
 			}
 		}
-		spin_unlock_irqrestore(&dev->event_lock, flags);
+		spin_unlock_irq(&dev->event_lock);
 	}
 
 	mutex_unlock(&dev->struct_mutex);
@@ -650,7 +664,6 @@ static int i915_interrupt_info(struct seq_file *m, void *data)
 	intel_runtime_pm_get(dev_priv);
 
 	if (IS_CHERRYVIEW(dev)) {
-		int i;
 		seq_printf(m, "Master Interrupt Control:\t%08x\n",
 			   I915_READ(GEN8_MASTER_IRQ));
 
@@ -662,7 +675,7 @@ static int i915_interrupt_info(struct seq_file *m, void *data)
 			   I915_READ(VLV_IIR_RW));
 		seq_printf(m, "Display IMR:\t%08x\n",
 			   I915_READ(VLV_IMR));
-		for_each_pipe(pipe)
+		for_each_pipe(dev_priv, pipe)
 			seq_printf(m, "Pipe %c stat:\t%08x\n",
 				   pipe_name(pipe),
 				   I915_READ(PIPESTAT(pipe)));
@@ -702,7 +715,13 @@ static int i915_interrupt_info(struct seq_file *m, void *data)
 				   i, I915_READ(GEN8_GT_IER(i)));
 		}
 
-		for_each_pipe(pipe) {
+		for_each_pipe(dev_priv, pipe) {
+			if (!intel_display_power_is_enabled(dev_priv,
+						POWER_DOMAIN_PIPE(pipe))) {
+				seq_printf(m, "Pipe %c power disabled\n",
+					   pipe_name(pipe));
+				continue;
+			}
 			seq_printf(m, "Pipe %c IMR:\t%08x\n",
 				   pipe_name(pipe),
 				   I915_READ(GEN8_DE_PIPE_IMR(pipe)));
@@ -743,7 +762,7 @@ static int i915_interrupt_info(struct seq_file *m, void *data)
 			   I915_READ(VLV_IIR_RW));
 		seq_printf(m, "Display IMR:\t%08x\n",
 			   I915_READ(VLV_IMR));
-		for_each_pipe(pipe)
+		for_each_pipe(dev_priv, pipe)
 			seq_printf(m, "Pipe %c stat:\t%08x\n",
 				   pipe_name(pipe),
 				   I915_READ(PIPESTAT(pipe)));
@@ -779,7 +798,7 @@ static int i915_interrupt_info(struct seq_file *m, void *data)
 			   I915_READ(IIR));
 		seq_printf(m, "Interrupt mask:      %08x\n",
 			   I915_READ(IMR));
-		for_each_pipe(pipe)
+		for_each_pipe(dev_priv, pipe)
 			seq_printf(m, "Pipe %c stat:         %08x\n",
 				   pipe_name(pipe),
 				   I915_READ(PIPESTAT(pipe)));
@@ -927,7 +946,7 @@ static ssize_t i915_error_state_read(struct file *file, char __user *userbuf,
 	ssize_t ret_count = 0;
 	int ret;
 
-	ret = i915_error_state_buf_init(&error_str, count, *pos);
+	ret = i915_error_state_buf_init(&error_str, to_i915(error_priv->dev), count, *pos);
 	if (ret)
 		return ret;
 
@@ -1024,6 +1043,7 @@ static int i915_frequency_info(struct seq_file *m, void *unused)
 		u32 rpstat, cagf, reqf;
 		u32 rpupei, rpcurup, rpprevup;
 		u32 rpdownei, rpcurdown, rpprevdown;
+		u32 pm_ier, pm_imr, pm_isr, pm_iir, pm_mask;
 		int max_freq;
 
 		/* RPSTAT1 is in the GT power well */
@@ -1061,12 +1081,21 @@ static int i915_frequency_info(struct seq_file *m, void *unused)
 		gen6_gt_force_wake_put(dev_priv, FORCEWAKE_ALL);
 		mutex_unlock(&dev->struct_mutex);
 
+		if (IS_GEN6(dev) || IS_GEN7(dev)) {
+			pm_ier = I915_READ(GEN6_PMIER);
+			pm_imr = I915_READ(GEN6_PMIMR);
+			pm_isr = I915_READ(GEN6_PMISR);
+			pm_iir = I915_READ(GEN6_PMIIR);
+			pm_mask = I915_READ(GEN6_PMINTRMSK);
+		} else {
+			pm_ier = I915_READ(GEN8_GT_IER(2));
+			pm_imr = I915_READ(GEN8_GT_IMR(2));
+			pm_isr = I915_READ(GEN8_GT_ISR(2));
+			pm_iir = I915_READ(GEN8_GT_IIR(2));
+			pm_mask = I915_READ(GEN6_PMINTRMSK);
+		}
 		seq_printf(m, "PM IER=0x%08x IMR=0x%08x ISR=0x%08x IIR=0x%08x, MASK=0x%08x\n",
-			   I915_READ(GEN6_PMIER),
-			   I915_READ(GEN6_PMIMR),
-			   I915_READ(GEN6_PMISR),
-			   I915_READ(GEN6_PMIIR),
-			   I915_READ(GEN6_PMINTRMSK));
+			   pm_ier, pm_imr, pm_isr, pm_iir, pm_mask);
 		seq_printf(m, "GT_PERF_STATUS: 0x%08x\n", gt_perf_status);
 		seq_printf(m, "Render p-state ratio: %d\n",
 			   (gt_perf_status & 0xff00) >> 8);
@@ -1211,11 +1240,12 @@ static int vlv_drpc_info(struct seq_file *m)
 	struct drm_info_node *node = m->private;
 	struct drm_device *dev = node->minor->dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	u32 rpmodectl1, rcctl1;
+	u32 rpmodectl1, rcctl1, pw_status;
 	unsigned fw_rendercount = 0, fw_mediacount = 0;
 
 	intel_runtime_pm_get(dev_priv);
 
+	pw_status = I915_READ(VLV_GTLC_PW_STATUS);
 	rpmodectl1 = I915_READ(GEN6_RP_CONTROL);
 	rcctl1 = I915_READ(GEN6_RC_CONTROL);
 
@@ -1234,11 +1264,9 @@ static int vlv_drpc_info(struct seq_file *m)
 		   yesno(rcctl1 & (GEN7_RC_CTL_TO_MODE |
 					GEN6_RC_CTL_EI_MODE(1))));
 	seq_printf(m, "Render Power Well: %s\n",
-			(I915_READ(VLV_GTLC_PW_STATUS) &
-				VLV_GTLC_PW_RENDER_STATUS_MASK) ? "Up" : "Down");
+		   (pw_status & VLV_GTLC_PW_RENDER_STATUS_MASK) ? "Up" : "Down");
 	seq_printf(m, "Media Power Well: %s\n",
-			(I915_READ(VLV_GTLC_PW_STATUS) &
-				VLV_GTLC_PW_MEDIA_STATUS_MASK) ? "Up" : "Down");
+		   (pw_status & VLV_GTLC_PW_MEDIA_STATUS_MASK) ? "Up" : "Down");
 
 	seq_printf(m, "Render RC6 residency since boot: %u\n",
 		   I915_READ(VLV_GT_RENDER_RC6));
@@ -1365,7 +1393,7 @@ static int i915_drpc_info(struct seq_file *m, void *unused)
 
 	if (IS_VALLEYVIEW(dev))
 		return vlv_drpc_info(m);
-	else if (IS_GEN6(dev) || IS_GEN7(dev))
+	else if (INTEL_INFO(dev)->gen >= 6)
 		return gen6_drpc_info(m);
 	else
 		return ironlake_drpc_info(m);
@@ -1433,6 +1461,47 @@ static int i915_fbc_status(struct seq_file *m, void *unused)
 	return 0;
 }
 
+static int i915_fbc_fc_get(void *data, u64 *val)
+{
+	struct drm_device *dev = data;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+
+	if (INTEL_INFO(dev)->gen < 7 || !HAS_FBC(dev))
+		return -ENODEV;
+
+	drm_modeset_lock_all(dev);
+	*val = dev_priv->fbc.false_color;
+	drm_modeset_unlock_all(dev);
+
+	return 0;
+}
+
+static int i915_fbc_fc_set(void *data, u64 val)
+{
+	struct drm_device *dev = data;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	u32 reg;
+
+	if (INTEL_INFO(dev)->gen < 7 || !HAS_FBC(dev))
+		return -ENODEV;
+
+	drm_modeset_lock_all(dev);
+
+	reg = I915_READ(ILK_DPFC_CONTROL);
+	dev_priv->fbc.false_color = val;
+
+	I915_WRITE(ILK_DPFC_CONTROL, val ?
+		   (reg | FBC_CTL_FALSE_COLOR) :
+		   (reg & ~FBC_CTL_FALSE_COLOR));
+
+	drm_modeset_unlock_all(dev);
+	return 0;
+}
+
+DEFINE_SIMPLE_ATTRIBUTE(i915_fbc_fc_fops,
+			i915_fbc_fc_get, i915_fbc_fc_set,
+			"%llu\n");
+
 static int i915_ips_status(struct seq_file *m, void *unused)
 {
 	struct drm_info_node *node = m->private;
@@ -1630,6 +1699,14 @@ static int i915_gem_framebuffer_info(struct seq_file *m, void *data)
 	return 0;
 }
 
+static void describe_ctx_ringbuf(struct seq_file *m,
+				 struct intel_ringbuffer *ringbuf)
+{
+	seq_printf(m, " (ringbuffer, space: %d, head: %u, tail: %u, last head: %d)",
+		   ringbuf->space, ringbuf->head, ringbuf->tail,
+		   ringbuf->last_retired_head);
+}
+
 static int i915_context_status(struct seq_file *m, void *unused)
 {
 	struct drm_info_node *node = m->private;
@@ -1656,19 +1733,198 @@ static int i915_context_status(struct seq_file *m, void *unused)
 	}
 
 	list_for_each_entry(ctx, &dev_priv->context_list, link) {
-		if (ctx->legacy_hw_ctx.rcs_state == NULL)
+		if (!i915.enable_execlists &&
+		    ctx->legacy_hw_ctx.rcs_state == NULL)
 			continue;
 
 		seq_puts(m, "HW context ");
 		describe_ctx(m, ctx);
-		for_each_ring(ring, dev_priv, i)
+		for_each_ring(ring, dev_priv, i) {
 			if (ring->default_context == ctx)
-				seq_printf(m, "(default context %s) ", ring->name);
+				seq_printf(m, "(default context %s) ",
+					   ring->name);
+		}
+
+		if (i915.enable_execlists) {
+			seq_putc(m, '\n');
+			for_each_ring(ring, dev_priv, i) {
+				struct drm_i915_gem_object *ctx_obj =
+					ctx->engine[i].state;
+				struct intel_ringbuffer *ringbuf =
+					ctx->engine[i].ringbuf;
+
+				seq_printf(m, "%s: ", ring->name);
+				if (ctx_obj)
+					describe_obj(m, ctx_obj);
+				if (ringbuf)
+					describe_ctx_ringbuf(m, ringbuf);
+				seq_putc(m, '\n');
+			}
+		} else {
+			describe_obj(m, ctx->legacy_hw_ctx.rcs_state);
+		}
+
+		seq_putc(m, '\n');
+	}
+
+	mutex_unlock(&dev->struct_mutex);
+
+	return 0;
+}
+
+static void i915_dump_lrc_obj(struct seq_file *m,
+			      struct intel_engine_cs *ring,
+			      struct drm_i915_gem_object *ctx_obj)
+{
+	struct page *page;
+	uint32_t *reg_state;
+	int j;
+	unsigned long ggtt_offset = 0;
+
+	if (ctx_obj == NULL) {
+		seq_printf(m, "Context on %s with no gem object\n",
+			   ring->name);
+		return;
+	}
+
+	seq_printf(m, "CONTEXT: %s %u\n", ring->name,
+		   intel_execlists_ctx_id(ctx_obj));
+
+	if (!i915_gem_obj_ggtt_bound(ctx_obj))
+		seq_puts(m, "\tNot bound in GGTT\n");
+	else
+		ggtt_offset = i915_gem_obj_ggtt_offset(ctx_obj);
+
+	if (i915_gem_object_get_pages(ctx_obj)) {
+		seq_puts(m, "\tFailed to get pages for context object\n");
+		return;
+	}
+
+	page = i915_gem_object_get_page(ctx_obj, 1);
+	if (!WARN_ON(page == NULL)) {
+		reg_state = kmap_atomic(page);
+
+		for (j = 0; j < 0x600 / sizeof(u32) / 4; j += 4) {
+			seq_printf(m, "\t[0x%08lx] 0x%08x 0x%08x 0x%08x 0x%08x\n",
+				   ggtt_offset + 4096 + (j * 4),
+				   reg_state[j], reg_state[j + 1],
+				   reg_state[j + 2], reg_state[j + 3]);
+		}
+		kunmap_atomic(reg_state);
+	}
+
+	seq_putc(m, '\n');
+}
+
+static int i915_dump_lrc(struct seq_file *m, void *unused)
+{
+	struct drm_info_node *node = (struct drm_info_node *) m->private;
+	struct drm_device *dev = node->minor->dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct intel_engine_cs *ring;
+	struct intel_context *ctx;
+	int ret, i;
+
+	if (!i915.enable_execlists) {
+		seq_printf(m, "Logical Ring Contexts are disabled\n");
+		return 0;
+	}
+
+	ret = mutex_lock_interruptible(&dev->struct_mutex);
+	if (ret)
+		return ret;
+
+	list_for_each_entry(ctx, &dev_priv->context_list, link) {
+		for_each_ring(ring, dev_priv, i) {
+			if (ring->default_context != ctx)
+				i915_dump_lrc_obj(m, ring,
+						  ctx->engine[i].state);
+		}
+	}
+
+	mutex_unlock(&dev->struct_mutex);
+
+	return 0;
+}
+
+static int i915_execlists(struct seq_file *m, void *data)
+{
+	struct drm_info_node *node = (struct drm_info_node *)m->private;
+	struct drm_device *dev = node->minor->dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct intel_engine_cs *ring;
+	u32 status_pointer;
+	u8 read_pointer;
+	u8 write_pointer;
+	u32 status;
+	u32 ctx_id;
+	struct list_head *cursor;
+	int ring_id, i;
+	int ret;
+
+	if (!i915.enable_execlists) {
+		seq_puts(m, "Logical Ring Contexts are disabled\n");
+		return 0;
+	}
+
+	ret = mutex_lock_interruptible(&dev->struct_mutex);
+	if (ret)
+		return ret;
+
+	intel_runtime_pm_get(dev_priv);
+
+	for_each_ring(ring, dev_priv, ring_id) {
+		struct intel_ctx_submit_request *head_req = NULL;
+		int count = 0;
+		unsigned long flags;
+
+		seq_printf(m, "%s\n", ring->name);
+
+		status = I915_READ(RING_EXECLIST_STATUS(ring));
+		ctx_id = I915_READ(RING_EXECLIST_STATUS(ring) + 4);
+		seq_printf(m, "\tExeclist status: 0x%08X, context: %u\n",
+			   status, ctx_id);
+
+		status_pointer = I915_READ(RING_CONTEXT_STATUS_PTR(ring));
+		seq_printf(m, "\tStatus pointer: 0x%08X\n", status_pointer);
+
+		read_pointer = ring->next_context_status_buffer;
+		write_pointer = status_pointer & 0x07;
+		if (read_pointer > write_pointer)
+			write_pointer += 6;
+		seq_printf(m, "\tRead pointer: 0x%08X, write pointer 0x%08X\n",
+			   read_pointer, write_pointer);
+
+		for (i = 0; i < 6; i++) {
+			status = I915_READ(RING_CONTEXT_STATUS_BUF(ring) + 8*i);
+			ctx_id = I915_READ(RING_CONTEXT_STATUS_BUF(ring) + 8*i + 4);
+
+			seq_printf(m, "\tStatus buffer %d: 0x%08X, context: %u\n",
+				   i, status, ctx_id);
+		}
+
+		spin_lock_irqsave(&ring->execlist_lock, flags);
+		list_for_each(cursor, &ring->execlist_queue)
+			count++;
+		head_req = list_first_entry_or_null(&ring->execlist_queue,
+				struct intel_ctx_submit_request, execlist_link);
+		spin_unlock_irqrestore(&ring->execlist_lock, flags);
+
+		seq_printf(m, "\t%d requests in queue\n", count);
+		if (head_req) {
+			struct drm_i915_gem_object *ctx_obj;
+
+			ctx_obj = head_req->ctx->engine[ring_id].state;
+			seq_printf(m, "\tHead request id: %u\n",
+				   intel_execlists_ctx_id(ctx_obj));
+			seq_printf(m, "\tHead request tail: %u\n",
+				   head_req->tail);
+		}
 
-		describe_obj(m, ctx->legacy_hw_ctx.rcs_state);
 		seq_putc(m, '\n');
 	}
 
+	intel_runtime_pm_put(dev_priv);
 	mutex_unlock(&dev->struct_mutex);
 
 	return 0;
@@ -1742,6 +1998,8 @@ static int i915_swizzle_info(struct seq_file *m, void *data)
 	if (IS_GEN3(dev) || IS_GEN4(dev)) {
 		seq_printf(m, "DDC = 0x%08x\n",
 			   I915_READ(DCC));
+		seq_printf(m, "DDC2 = 0x%08x\n",
+			   I915_READ(DCC2));
 		seq_printf(m, "C0DRB3 = 0x%04x\n",
 			   I915_READ16(C0DRB3));
 		seq_printf(m, "C1DRB3 = 0x%04x\n",
@@ -1755,7 +2013,7 @@ static int i915_swizzle_info(struct seq_file *m, void *data)
 			   I915_READ(MAD_DIMM_C2));
 		seq_printf(m, "TILECTL = 0x%08x\n",
 			   I915_READ(TILECTL));
-		if (IS_GEN8(dev))
+		if (INTEL_INFO(dev)->gen >= 8)
 			seq_printf(m, "GAMTARBMODE = 0x%08x\n",
 				   I915_READ(GAMTARBMODE));
 		else
@@ -1764,6 +2022,10 @@ static int i915_swizzle_info(struct seq_file *m, void *data)
 		seq_printf(m, "DISP_ARB_CTL = 0x%08x\n",
 			   I915_READ(DISP_ARB_CTL));
 	}
+
+	if (dev_priv->quirks & QUIRK_PIN_SWIZZLED_PAGES)
+		seq_puts(m, "L-shaped memory detected\n");
+
 	intel_runtime_pm_put(dev_priv);
 	mutex_unlock(&dev->struct_mutex);
 
@@ -1774,7 +2036,13 @@ static int per_file_ctx(int id, void *ptr, void *data)
 {
 	struct intel_context *ctx = ptr;
 	struct seq_file *m = data;
-	struct i915_hw_ppgtt *ppgtt = ctx_to_ppgtt(ctx);
+	struct i915_hw_ppgtt *ppgtt = ctx->ppgtt;
+
+	if (!ppgtt) {
+		seq_printf(m, "  no ppgtt for context %d\n",
+			   ctx->user_handle);
+		return 0;
+	}
 
 	if (i915_gem_context_is_default(ctx))
 		seq_puts(m, "  default context:\n");
@@ -1834,8 +2102,7 @@ static void gen6_ppgtt_info(struct seq_file *m, struct drm_device *dev)
 		seq_printf(m, "pd gtt offset: 0x%08x\n", ppgtt->pd_offset);
 
 		ppgtt->debug_dump(ppgtt, m);
-	} else
-		return;
+	}
 
 	list_for_each_entry_reverse(file, &dev->filelist, lhead) {
 		struct drm_i915_file_private *file_priv = file->driver_priv;
@@ -2392,20 +2659,91 @@ static int i915_shared_dplls_info(struct seq_file *m, void *unused)
 		struct intel_shared_dpll *pll = &dev_priv->shared_dplls[i];
 
 		seq_printf(m, "DPLL%i: %s, id: %i\n", i, pll->name, pll->id);
-		seq_printf(m, " refcount: %i, active: %i, on: %s\n", pll->refcount,
-			   pll->active, yesno(pll->on));
+		seq_printf(m, " crtc_mask: 0x%08x, active: %d, on: %s\n",
+			   pll->config.crtc_mask, pll->active, yesno(pll->on));
 		seq_printf(m, " tracked hardware state:\n");
-		seq_printf(m, " dpll:    0x%08x\n", pll->hw_state.dpll);
-		seq_printf(m, " dpll_md: 0x%08x\n", pll->hw_state.dpll_md);
-		seq_printf(m, " fp0:     0x%08x\n", pll->hw_state.fp0);
-		seq_printf(m, " fp1:     0x%08x\n", pll->hw_state.fp1);
-		seq_printf(m, " wrpll:   0x%08x\n", pll->hw_state.wrpll);
+		seq_printf(m, " dpll:    0x%08x\n", pll->config.hw_state.dpll);
+		seq_printf(m, " dpll_md: 0x%08x\n",
+			   pll->config.hw_state.dpll_md);
+		seq_printf(m, " fp0:     0x%08x\n", pll->config.hw_state.fp0);
+		seq_printf(m, " fp1:     0x%08x\n", pll->config.hw_state.fp1);
+		seq_printf(m, " wrpll:   0x%08x\n", pll->config.hw_state.wrpll);
 	}
 	drm_modeset_unlock_all(dev);
 
 	return 0;
 }
 
+static int i915_wa_registers(struct seq_file *m, void *unused)
+{
+	int i;
+	int ret;
+	struct drm_info_node *node = (struct drm_info_node *) m->private;
+	struct drm_device *dev = node->minor->dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+
+	ret = mutex_lock_interruptible(&dev->struct_mutex);
+	if (ret)
+		return ret;
+
+	intel_runtime_pm_get(dev_priv);
+
+	seq_printf(m, "Workarounds applied: %d\n", dev_priv->workarounds.count);
+	for (i = 0; i < dev_priv->workarounds.count; ++i) {
+		u32 addr, mask, value, read;
+		bool ok;
+
+		addr = dev_priv->workarounds.reg[i].addr;
+		mask = dev_priv->workarounds.reg[i].mask;
+		value = dev_priv->workarounds.reg[i].value;
+		read = I915_READ(addr);
+		ok = (value & mask) == (read & mask);
+		seq_printf(m, "0x%X: 0x%08X, mask: 0x%08X, read: 0x%08x, status: %s\n",
+			   addr, value, mask, read, ok ? "OK" : "FAIL");
+	}
+
+	intel_runtime_pm_put(dev_priv);
+	mutex_unlock(&dev->struct_mutex);
+
+	return 0;
+}
+
+static int i915_ddb_info(struct seq_file *m, void *unused)
+{
+	struct drm_info_node *node = m->private;
+	struct drm_device *dev = node->minor->dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct skl_ddb_allocation *ddb;
+	struct skl_ddb_entry *entry;
+	enum pipe pipe;
+	int plane;
+
+	drm_modeset_lock_all(dev);
+
+	ddb = &dev_priv->wm.skl_hw.ddb;
+
+	seq_printf(m, "%-15s%8s%8s%8s\n", "", "Start", "End", "Size");
+
+	for_each_pipe(dev_priv, pipe) {
+		seq_printf(m, "Pipe %c\n", pipe_name(pipe));
+
+		for_each_plane(pipe, plane) {
+			entry = &ddb->plane[pipe][plane];
+			seq_printf(m, "  Plane%-8d%8u%8u%8u\n", plane + 1,
+				   entry->start, entry->end,
+				   skl_ddb_entry_size(entry));
+		}
+
+		entry = &ddb->cursor[pipe];
+		seq_printf(m, "  %-13s%8u%8u%8u\n", "Cursor", entry->start,
+			   entry->end, skl_ddb_entry_size(entry));
+	}
+
+	drm_modeset_unlock_all(dev);
+
+	return 0;
+}
+
 struct pipe_crc_info {
 	const char *name;
 	struct drm_device *dev;
@@ -2667,8 +3005,7 @@ static int i9xx_pipe_crc_auto_source(struct drm_device *dev, enum pipe pipe,
 	*source = INTEL_PIPE_CRC_SOURCE_PIPE;
 
 	drm_modeset_lock_all(dev);
-	list_for_each_entry(encoder, &dev->mode_config.encoder_list,
-			    base.head) {
+	for_each_intel_encoder(dev, encoder) {
 		if (!encoder->base.crtc)
 			continue;
 
@@ -2700,6 +3037,8 @@ static int i9xx_pipe_crc_auto_source(struct drm_device *dev, enum pipe pipe,
 				break;
 			}
 			break;
+		default:
+			break;
 		}
 	}
 	drm_modeset_unlock_all(dev);
@@ -2987,6 +3326,8 @@ static int pipe_crc_set_source(struct drm_device *dev, enum pipe pipe,
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_pipe_crc *pipe_crc = &dev_priv->pipe_crc[pipe];
+	struct intel_crtc *crtc = to_intel_crtc(intel_get_crtc_for_pipe(dev,
+									pipe));
 	u32 val = 0; /* shut up gcc */
 	int ret;
 
@@ -2997,6 +3338,11 @@ static int pipe_crc_set_source(struct drm_device *dev, enum pipe pipe,
 	if (pipe_crc->source && source)
 		return -EINVAL;
 
+	if (!intel_display_power_is_enabled(dev_priv, POWER_DOMAIN_PIPE(pipe))) {
+		DRM_DEBUG_KMS("Trying to capture CRC while pipe is off\n");
+		return -EIO;
+	}
+
 	if (IS_GEN2(dev))
 		ret = i8xx_pipe_crc_ctl_reg(&source, &val);
 	else if (INTEL_INFO(dev)->gen < 5)
@@ -3022,6 +3368,14 @@ static int pipe_crc_set_source(struct drm_device *dev, enum pipe pipe,
 		if (!pipe_crc->entries)
 			return -ENOMEM;
 
+		/*
+		 * When IPS gets enabled, the pipe CRC changes. Since IPS gets
+		 * enabled and disabled dynamically based on package C states,
+		 * user space can't make reliable use of the CRCs, so let's just
+		 * completely disable it.
+		 */
+		hsw_disable_ips(crtc);
+
 		spin_lock_irq(&pipe_crc->lock);
 		pipe_crc->head = 0;
 		pipe_crc->tail = 0;
@@ -3060,6 +3414,8 @@ static int pipe_crc_set_source(struct drm_device *dev, enum pipe pipe,
 			vlv_undo_pipe_scramble_reset(dev, pipe);
 		else if (IS_HASWELL(dev) && pipe == PIPE_A)
 			hsw_undo_trans_edp_pipe_A_crc_wa(dev);
+
+		hsw_enable_ips(crtc);
 	}
 
 	return 0;
@@ -3237,7 +3593,7 @@ static const struct file_operations i915_display_crc_ctl_fops = {
 	.write = display_crc_ctl_write
 };
 
-static void wm_latency_show(struct seq_file *m, const uint16_t wm[5])
+static void wm_latency_show(struct seq_file *m, const uint16_t wm[8])
 {
 	struct drm_device *dev = m->private;
 	int num_levels = ilk_wm_max_level(dev) + 1;
@@ -3248,13 +3604,17 @@ static void wm_latency_show(struct seq_file *m, const uint16_t wm[5])
 	for (level = 0; level < num_levels; level++) {
 		unsigned int latency = wm[level];
 
-		/* WM1+ latency values in 0.5us units */
-		if (level > 0)
+		/*
+		 * - WM1+ latency values in 0.5us units
+		 * - latencies are in us on gen9
+		 */
+		if (INTEL_INFO(dev)->gen >= 9)
+			latency *= 10;
+		else if (level > 0)
 			latency *= 5;
 
 		seq_printf(m, "WM%d %u (%u.%u usec)\n",
-			   level, wm[level],
-			   latency / 10, latency % 10);
+			   level, wm[level], latency / 10, latency % 10);
 	}
 
 	drm_modeset_unlock_all(dev);
@@ -3263,8 +3623,15 @@ static void wm_latency_show(struct seq_file *m, const uint16_t wm[5])
 static int pri_wm_latency_show(struct seq_file *m, void *data)
 {
 	struct drm_device *dev = m->private;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	const uint16_t *latencies;
 
-	wm_latency_show(m, to_i915(dev)->wm.pri_latency);
+	if (INTEL_INFO(dev)->gen >= 9)
+		latencies = dev_priv->wm.skl_latency;
+	else
+		latencies = to_i915(dev)->wm.pri_latency;
+
+	wm_latency_show(m, latencies);
 
 	return 0;
 }
@@ -3272,8 +3639,15 @@ static int pri_wm_latency_show(struct seq_file *m, void *data)
 static int spr_wm_latency_show(struct seq_file *m, void *data)
 {
 	struct drm_device *dev = m->private;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	const uint16_t *latencies;
 
-	wm_latency_show(m, to_i915(dev)->wm.spr_latency);
+	if (INTEL_INFO(dev)->gen >= 9)
+		latencies = dev_priv->wm.skl_latency;
+	else
+		latencies = to_i915(dev)->wm.spr_latency;
+
+	wm_latency_show(m, latencies);
 
 	return 0;
 }
@@ -3281,8 +3655,15 @@ static int spr_wm_latency_show(struct seq_file *m, void *data)
 static int cur_wm_latency_show(struct seq_file *m, void *data)
 {
 	struct drm_device *dev = m->private;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	const uint16_t *latencies;
+
+	if (INTEL_INFO(dev)->gen >= 9)
+		latencies = dev_priv->wm.skl_latency;
+	else
+		latencies = to_i915(dev)->wm.cur_latency;
 
-	wm_latency_show(m, to_i915(dev)->wm.cur_latency);
+	wm_latency_show(m, latencies);
 
 	return 0;
 }
@@ -3318,11 +3699,11 @@ static int cur_wm_latency_open(struct inode *inode, struct file *file)
 }
 
 static ssize_t wm_latency_write(struct file *file, const char __user *ubuf,
-				size_t len, loff_t *offp, uint16_t wm[5])
+				size_t len, loff_t *offp, uint16_t wm[8])
 {
 	struct seq_file *m = file->private_data;
 	struct drm_device *dev = m->private;
-	uint16_t new[5] = { 0 };
+	uint16_t new[8] = { 0 };
 	int num_levels = ilk_wm_max_level(dev) + 1;
 	int level;
 	int ret;
@@ -3336,7 +3717,9 @@ static ssize_t wm_latency_write(struct file *file, const char __user *ubuf,
 
 	tmp[len] = '\0';
 
-	ret = sscanf(tmp, "%hu %hu %hu %hu %hu", &new[0], &new[1], &new[2], &new[3], &new[4]);
+	ret = sscanf(tmp, "%hu %hu %hu %hu %hu %hu %hu %hu",
+		     &new[0], &new[1], &new[2], &new[3],
+		     &new[4], &new[5], &new[6], &new[7]);
 	if (ret != num_levels)
 		return -EINVAL;
 
@@ -3356,8 +3739,15 @@ static ssize_t pri_wm_latency_write(struct file *file, const char __user *ubuf,
 {
 	struct seq_file *m = file->private_data;
 	struct drm_device *dev = m->private;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	uint16_t *latencies;
+
+	if (INTEL_INFO(dev)->gen >= 9)
+		latencies = dev_priv->wm.skl_latency;
+	else
+		latencies = to_i915(dev)->wm.pri_latency;
 
-	return wm_latency_write(file, ubuf, len, offp, to_i915(dev)->wm.pri_latency);
+	return wm_latency_write(file, ubuf, len, offp, latencies);
 }
 
 static ssize_t spr_wm_latency_write(struct file *file, const char __user *ubuf,
@@ -3365,8 +3755,15 @@ static ssize_t spr_wm_latency_write(struct file *file, const char __user *ubuf,
 {
 	struct seq_file *m = file->private_data;
 	struct drm_device *dev = m->private;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	uint16_t *latencies;
+
+	if (INTEL_INFO(dev)->gen >= 9)
+		latencies = dev_priv->wm.skl_latency;
+	else
+		latencies = to_i915(dev)->wm.spr_latency;
 
-	return wm_latency_write(file, ubuf, len, offp, to_i915(dev)->wm.spr_latency);
+	return wm_latency_write(file, ubuf, len, offp, latencies);
 }
 
 static ssize_t cur_wm_latency_write(struct file *file, const char __user *ubuf,
@@ -3374,8 +3771,15 @@ static ssize_t cur_wm_latency_write(struct file *file, const char __user *ubuf,
 {
 	struct seq_file *m = file->private_data;
 	struct drm_device *dev = m->private;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	uint16_t *latencies;
+
+	if (INTEL_INFO(dev)->gen >= 9)
+		latencies = dev_priv->wm.skl_latency;
+	else
+		latencies = to_i915(dev)->wm.cur_latency;
 
-	return wm_latency_write(file, ubuf, len, offp, to_i915(dev)->wm.cur_latency);
+	return wm_latency_write(file, ubuf, len, offp, latencies);
 }
 
 static const struct file_operations i915_pri_wm_latency_fops = {
@@ -3557,9 +3961,6 @@ i915_drop_caches_set(void *data, u64 val)
 {
 	struct drm_device *dev = data;
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct drm_i915_gem_object *obj, *next;
-	struct i915_address_space *vm;
-	struct i915_vma *vma, *x;
 	int ret;
 
 	DRM_DEBUG("Dropping caches: 0x%08llx\n", val);
@@ -3579,29 +3980,11 @@ i915_drop_caches_set(void *data, u64 val)
 	if (val & (DROP_RETIRE | DROP_ACTIVE))
 		i915_gem_retire_requests(dev);
 
-	if (val & DROP_BOUND) {
-		list_for_each_entry(vm, &dev_priv->vm_list, global_link) {
-			list_for_each_entry_safe(vma, x, &vm->inactive_list,
-						 mm_list) {
-				if (vma->pin_count)
-					continue;
+	if (val & DROP_BOUND)
+		i915_gem_shrink(dev_priv, LONG_MAX, I915_SHRINK_BOUND);
 
-				ret = i915_vma_unbind(vma);
-				if (ret)
-					goto unlock;
-			}
-		}
-	}
-
-	if (val & DROP_UNBOUND) {
-		list_for_each_entry_safe(obj, next, &dev_priv->mm.unbound_list,
-					 global_list)
-			if (obj->pages_pin_count == 0) {
-				ret = i915_gem_object_put_pages(obj);
-				if (ret)
-					goto unlock;
-			}
-	}
+	if (val & DROP_UNBOUND)
+		i915_gem_shrink(dev_priv, LONG_MAX, I915_SHRINK_UNBOUND);
 
 unlock:
 	mutex_unlock(&dev->struct_mutex);
@@ -3923,6 +4306,8 @@ static const struct drm_info_list i915_debugfs_list[] = {
 	{"i915_opregion", i915_opregion, 0},
 	{"i915_gem_framebuffer", i915_gem_framebuffer_info, 0},
 	{"i915_context_status", i915_context_status, 0},
+	{"i915_dump_lrc", i915_dump_lrc, 0},
+	{"i915_execlists", i915_execlists, 0},
 	{"i915_gen6_forcewake_count", i915_gen6_forcewake_count_info, 0},
 	{"i915_swizzle_info", i915_swizzle_info, 0},
 	{"i915_ppgtt_info", i915_ppgtt_info, 0},
@@ -3936,6 +4321,8 @@ static const struct drm_info_list i915_debugfs_list[] = {
 	{"i915_semaphore_status", i915_semaphore_status, 0},
 	{"i915_shared_dplls_info", i915_shared_dplls_info, 0},
 	{"i915_dp_mst_info", i915_dp_mst_info, 0},
+	{"i915_wa_registers", i915_wa_registers, 0},
+	{"i915_ddb_info", i915_ddb_info, 0},
 };
 #define I915_DEBUGFS_ENTRIES ARRAY_SIZE(i915_debugfs_list)
 
@@ -3957,6 +4344,7 @@ static const struct i915_debugfs_files {
 	{"i915_pri_wm_latency", &i915_pri_wm_latency_fops},
 	{"i915_spr_wm_latency", &i915_spr_wm_latency_fops},
 	{"i915_cur_wm_latency", &i915_cur_wm_latency_fops},
+	{"i915_fbc_false_color", &i915_fbc_fc_fops},
 };
 
 void intel_display_crc_init(struct drm_device *dev)
@@ -3964,7 +4352,7 @@ void intel_display_crc_init(struct drm_device *dev)
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	enum pipe pipe;
 
-	for_each_pipe(pipe) {
+	for_each_pipe(dev_priv, pipe) {
 		struct intel_pipe_crc *pipe_crc = &dev_priv->pipe_crc[pipe];
 
 		pipe_crc->opened = false;
diff --git a/drivers/gpu/drm/i915/i915_dma.c b/drivers/gpu/drm/i915/i915_dma.c
index 9933c26017ed..ecee3bcc8772 100644
--- a/drivers/gpu/drm/i915/i915_dma.c
+++ b/drivers/gpu/drm/i915/i915_dma.c
@@ -28,9 +28,11 @@
 
 #define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
 
+#include <linux/async.h>
 #include <drm/drmP.h>
 #include <drm/drm_crtc_helper.h>
 #include <drm/drm_fb_helper.h>
+#include <drm/drm_legacy.h>
 #include "intel_drv.h"
 #include <drm/i915_drm.h>
 #include "i915_drv.h"
@@ -48,884 +50,6 @@
 #include <linux/pm_runtime.h>
 #include <linux/oom.h>
 
-#define LP_RING(d) (&((struct drm_i915_private *)(d))->ring[RCS])
-
-#define BEGIN_LP_RING(n) \
-	intel_ring_begin(LP_RING(dev_priv), (n))
-
-#define OUT_RING(x) \
-	intel_ring_emit(LP_RING(dev_priv), x)
-
-#define ADVANCE_LP_RING() \
-	__intel_ring_advance(LP_RING(dev_priv))
-
-/**
- * Lock test for when it's just for synchronization of ring access.
- *
- * In that case, we don't need to do it when GEM is initialized as nobody else
- * has access to the ring.
- */
-#define RING_LOCK_TEST_WITH_RETURN(dev, file) do {			\
-	if (LP_RING(dev->dev_private)->buffer->obj == NULL)			\
-		LOCK_TEST_WITH_RETURN(dev, file);			\
-} while (0)
-
-static inline u32
-intel_read_legacy_status_page(struct drm_i915_private *dev_priv, int reg)
-{
-	if (I915_NEED_GFX_HWS(dev_priv->dev))
-		return ioread32(dev_priv->dri1.gfx_hws_cpu_addr + reg);
-	else
-		return intel_read_status_page(LP_RING(dev_priv), reg);
-}
-
-#define READ_HWSP(dev_priv, reg) intel_read_legacy_status_page(dev_priv, reg)
-#define READ_BREADCRUMB(dev_priv) READ_HWSP(dev_priv, I915_BREADCRUMB_INDEX)
-#define I915_BREADCRUMB_INDEX		0x21
-
-void i915_update_dri1_breadcrumb(struct drm_device *dev)
-{
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct drm_i915_master_private *master_priv;
-
-	/*
-	 * The dri breadcrumb update races against the drm master disappearing.
-	 * Instead of trying to fix this (this is by far not the only ums issue)
-	 * just don't do the update in kms mode.
-	 */
-	if (drm_core_check_feature(dev, DRIVER_MODESET))
-		return;
-
-	if (dev->primary->master) {
-		master_priv = dev->primary->master->driver_priv;
-		if (master_priv->sarea_priv)
-			master_priv->sarea_priv->last_dispatch =
-				READ_BREADCRUMB(dev_priv);
-	}
-}
-
-static void i915_write_hws_pga(struct drm_device *dev)
-{
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	u32 addr;
-
-	addr = dev_priv->status_page_dmah->busaddr;
-	if (INTEL_INFO(dev)->gen >= 4)
-		addr |= (dev_priv->status_page_dmah->busaddr >> 28) & 0xf0;
-	I915_WRITE(HWS_PGA, addr);
-}
-
-/**
- * Frees the hardware status page, whether it's a physical address or a virtual
- * address set up by the X Server.
- */
-static void i915_free_hws(struct drm_device *dev)
-{
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct intel_engine_cs *ring = LP_RING(dev_priv);
-
-	if (dev_priv->status_page_dmah) {
-		drm_pci_free(dev, dev_priv->status_page_dmah);
-		dev_priv->status_page_dmah = NULL;
-	}
-
-	if (ring->status_page.gfx_addr) {
-		ring->status_page.gfx_addr = 0;
-		iounmap(dev_priv->dri1.gfx_hws_cpu_addr);
-	}
-
-	/* Need to rewrite hardware status page */
-	I915_WRITE(HWS_PGA, 0x1ffff000);
-}
-
-void i915_kernel_lost_context(struct drm_device *dev)
-{
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct drm_i915_master_private *master_priv;
-	struct intel_engine_cs *ring = LP_RING(dev_priv);
-	struct intel_ringbuffer *ringbuf = ring->buffer;
-
-	/*
-	 * We should never lose context on the ring with modesetting
-	 * as we don't expose it to userspace
-	 */
-	if (drm_core_check_feature(dev, DRIVER_MODESET))
-		return;
-
-	ringbuf->head = I915_READ_HEAD(ring) & HEAD_ADDR;
-	ringbuf->tail = I915_READ_TAIL(ring) & TAIL_ADDR;
-	ringbuf->space = ringbuf->head - (ringbuf->tail + I915_RING_FREE_SPACE);
-	if (ringbuf->space < 0)
-		ringbuf->space += ringbuf->size;
-
-	if (!dev->primary->master)
-		return;
-
-	master_priv = dev->primary->master->driver_priv;
-	if (ringbuf->head == ringbuf->tail && master_priv->sarea_priv)
-		master_priv->sarea_priv->perf_boxes |= I915_BOX_RING_EMPTY;
-}
-
-static int i915_dma_cleanup(struct drm_device *dev)
-{
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	int i;
-
-	/* Make sure interrupts are disabled here because the uninstall ioctl
-	 * may not have been called from userspace and after dev_private
-	 * is freed, it's too late.
-	 */
-	if (dev->irq_enabled)
-		drm_irq_uninstall(dev);
-
-	mutex_lock(&dev->struct_mutex);
-	for (i = 0; i < I915_NUM_RINGS; i++)
-		intel_cleanup_ring_buffer(&dev_priv->ring[i]);
-	mutex_unlock(&dev->struct_mutex);
-
-	/* Clear the HWS virtual address at teardown */
-	if (I915_NEED_GFX_HWS(dev))
-		i915_free_hws(dev);
-
-	return 0;
-}
-
-static int i915_initialize(struct drm_device *dev, drm_i915_init_t *init)
-{
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct drm_i915_master_private *master_priv = dev->primary->master->driver_priv;
-	int ret;
-
-	master_priv->sarea = drm_getsarea(dev);
-	if (master_priv->sarea) {
-		master_priv->sarea_priv = (drm_i915_sarea_t *)
-			((u8 *)master_priv->sarea->handle + init->sarea_priv_offset);
-	} else {
-		DRM_DEBUG_DRIVER("sarea not found assuming DRI2 userspace\n");
-	}
-
-	if (init->ring_size != 0) {
-		if (LP_RING(dev_priv)->buffer->obj != NULL) {
-			i915_dma_cleanup(dev);
-			DRM_ERROR("Client tried to initialize ringbuffer in "
-				  "GEM mode\n");
-			return -EINVAL;
-		}
-
-		ret = intel_render_ring_init_dri(dev,
-						 init->ring_start,
-						 init->ring_size);
-		if (ret) {
-			i915_dma_cleanup(dev);
-			return ret;
-		}
-	}
-
-	dev_priv->dri1.cpp = init->cpp;
-	dev_priv->dri1.back_offset = init->back_offset;
-	dev_priv->dri1.front_offset = init->front_offset;
-	dev_priv->dri1.current_page = 0;
-	if (master_priv->sarea_priv)
-		master_priv->sarea_priv->pf_current_page = 0;
-
-	/* Allow hardware batchbuffers unless told otherwise.
-	 */
-	dev_priv->dri1.allow_batchbuffer = 1;
-
-	return 0;
-}
-
-static int i915_dma_resume(struct drm_device *dev)
-{
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct intel_engine_cs *ring = LP_RING(dev_priv);
-
-	DRM_DEBUG_DRIVER("%s\n", __func__);
-
-	if (ring->buffer->virtual_start == NULL) {
-		DRM_ERROR("can not ioremap virtual address for"
-			  " ring buffer\n");
-		return -ENOMEM;
-	}
-
-	/* Program Hardware Status Page */
-	if (!ring->status_page.page_addr) {
-		DRM_ERROR("Can not find hardware status page\n");
-		return -EINVAL;
-	}
-	DRM_DEBUG_DRIVER("hw status page @ %p\n",
-				ring->status_page.page_addr);
-	if (ring->status_page.gfx_addr != 0)
-		intel_ring_setup_status_page(ring);
-	else
-		i915_write_hws_pga(dev);
-
-	DRM_DEBUG_DRIVER("Enabled hardware status page\n");
-
-	return 0;
-}
-
-static int i915_dma_init(struct drm_device *dev, void *data,
-			 struct drm_file *file_priv)
-{
-	drm_i915_init_t *init = data;
-	int retcode = 0;
-
-	if (drm_core_check_feature(dev, DRIVER_MODESET))
-		return -ENODEV;
-
-	switch (init->func) {
-	case I915_INIT_DMA:
-		retcode = i915_initialize(dev, init);
-		break;
-	case I915_CLEANUP_DMA:
-		retcode = i915_dma_cleanup(dev);
-		break;
-	case I915_RESUME_DMA:
-		retcode = i915_dma_resume(dev);
-		break;
-	default:
-		retcode = -EINVAL;
-		break;
-	}
-
-	return retcode;
-}
-
-/* Implement basically the same security restrictions as hardware does
- * for MI_BATCH_NON_SECURE.  These can be made stricter at any time.
- *
- * Most of the calculations below involve calculating the size of a
- * particular instruction.  It's important to get the size right as
- * that tells us where the next instruction to check is.  Any illegal
- * instruction detected will be given a size of zero, which is a
- * signal to abort the rest of the buffer.
- */
-static int validate_cmd(int cmd)
-{
-	switch (((cmd >> 29) & 0x7)) {
-	case 0x0:
-		switch ((cmd >> 23) & 0x3f) {
-		case 0x0:
-			return 1;	/* MI_NOOP */
-		case 0x4:
-			return 1;	/* MI_FLUSH */
-		default:
-			return 0;	/* disallow everything else */
-		}
-		break;
-	case 0x1:
-		return 0;	/* reserved */
-	case 0x2:
-		return (cmd & 0xff) + 2;	/* 2d commands */
-	case 0x3:
-		if (((cmd >> 24) & 0x1f) <= 0x18)
-			return 1;
-
-		switch ((cmd >> 24) & 0x1f) {
-		case 0x1c:
-			return 1;
-		case 0x1d:
-			switch ((cmd >> 16) & 0xff) {
-			case 0x3:
-				return (cmd & 0x1f) + 2;
-			case 0x4:
-				return (cmd & 0xf) + 2;
-			default:
-				return (cmd & 0xffff) + 2;
-			}
-		case 0x1e:
-			if (cmd & (1 << 23))
-				return (cmd & 0xffff) + 1;
-			else
-				return 1;
-		case 0x1f:
-			if ((cmd & (1 << 23)) == 0)	/* inline vertices */
-				return (cmd & 0x1ffff) + 2;
-			else if (cmd & (1 << 17))	/* indirect random */
-				if ((cmd & 0xffff) == 0)
-					return 0;	/* unknown length, too hard */
-				else
-					return (((cmd & 0xffff) + 1) / 2) + 1;
-			else
-				return 2;	/* indirect sequential */
-		default:
-			return 0;
-		}
-	default:
-		return 0;
-	}
-
-	return 0;
-}
-
-static int i915_emit_cmds(struct drm_device *dev, int *buffer, int dwords)
-{
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	int i, ret;
-
-	if ((dwords+1) * sizeof(int) >= LP_RING(dev_priv)->buffer->size - 8)
-		return -EINVAL;
-
-	for (i = 0; i < dwords;) {
-		int sz = validate_cmd(buffer[i]);
-
-		if (sz == 0 || i + sz > dwords)
-			return -EINVAL;
-		i += sz;
-	}
-
-	ret = BEGIN_LP_RING((dwords+1)&~1);
-	if (ret)
-		return ret;
-
-	for (i = 0; i < dwords; i++)
-		OUT_RING(buffer[i]);
-	if (dwords & 1)
-		OUT_RING(0);
-
-	ADVANCE_LP_RING();
-
-	return 0;
-}
-
-int
-i915_emit_box(struct drm_device *dev,
-	      struct drm_clip_rect *box,
-	      int DR1, int DR4)
-{
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	int ret;
-
-	if (box->y2 <= box->y1 || box->x2 <= box->x1 ||
-	    box->y2 <= 0 || box->x2 <= 0) {
-		DRM_ERROR("Bad box %d,%d..%d,%d\n",
-			  box->x1, box->y1, box->x2, box->y2);
-		return -EINVAL;
-	}
-
-	if (INTEL_INFO(dev)->gen >= 4) {
-		ret = BEGIN_LP_RING(4);
-		if (ret)
-			return ret;
-
-		OUT_RING(GFX_OP_DRAWRECT_INFO_I965);
-		OUT_RING((box->x1 & 0xffff) | (box->y1 << 16));
-		OUT_RING(((box->x2 - 1) & 0xffff) | ((box->y2 - 1) << 16));
-		OUT_RING(DR4);
-	} else {
-		ret = BEGIN_LP_RING(6);
-		if (ret)
-			return ret;
-
-		OUT_RING(GFX_OP_DRAWRECT_INFO);
-		OUT_RING(DR1);
-		OUT_RING((box->x1 & 0xffff) | (box->y1 << 16));
-		OUT_RING(((box->x2 - 1) & 0xffff) | ((box->y2 - 1) << 16));
-		OUT_RING(DR4);
-		OUT_RING(0);
-	}
-	ADVANCE_LP_RING();
-
-	return 0;
-}
-
-/* XXX: Emitting the counter should really be moved to part of the IRQ
- * emit. For now, do it in both places:
- */
-
-static void i915_emit_breadcrumb(struct drm_device *dev)
-{
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct drm_i915_master_private *master_priv = dev->primary->master->driver_priv;
-
-	dev_priv->dri1.counter++;
-	if (dev_priv->dri1.counter > 0x7FFFFFFFUL)
-		dev_priv->dri1.counter = 0;
-	if (master_priv->sarea_priv)
-		master_priv->sarea_priv->last_enqueue = dev_priv->dri1.counter;
-
-	if (BEGIN_LP_RING(4) == 0) {
-		OUT_RING(MI_STORE_DWORD_INDEX);
-		OUT_RING(I915_BREADCRUMB_INDEX << MI_STORE_DWORD_INDEX_SHIFT);
-		OUT_RING(dev_priv->dri1.counter);
-		OUT_RING(0);
-		ADVANCE_LP_RING();
-	}
-}
-
-static int i915_dispatch_cmdbuffer(struct drm_device *dev,
-				   drm_i915_cmdbuffer_t *cmd,
-				   struct drm_clip_rect *cliprects,
-				   void *cmdbuf)
-{
-	int nbox = cmd->num_cliprects;
-	int i = 0, count, ret;
-
-	if (cmd->sz & 0x3) {
-		DRM_ERROR("alignment");
-		return -EINVAL;
-	}
-
-	i915_kernel_lost_context(dev);
-
-	count = nbox ? nbox : 1;
-
-	for (i = 0; i < count; i++) {
-		if (i < nbox) {
-			ret = i915_emit_box(dev, &cliprects[i],
-					    cmd->DR1, cmd->DR4);
-			if (ret)
-				return ret;
-		}
-
-		ret = i915_emit_cmds(dev, cmdbuf, cmd->sz / 4);
-		if (ret)
-			return ret;
-	}
-
-	i915_emit_breadcrumb(dev);
-	return 0;
-}
-
-static int i915_dispatch_batchbuffer(struct drm_device *dev,
-				     drm_i915_batchbuffer_t *batch,
-				     struct drm_clip_rect *cliprects)
-{
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	int nbox = batch->num_cliprects;
-	int i, count, ret;
-
-	if ((batch->start | batch->used) & 0x7) {
-		DRM_ERROR("alignment");
-		return -EINVAL;
-	}
-
-	i915_kernel_lost_context(dev);
-
-	count = nbox ? nbox : 1;
-	for (i = 0; i < count; i++) {
-		if (i < nbox) {
-			ret = i915_emit_box(dev, &cliprects[i],
-					    batch->DR1, batch->DR4);
-			if (ret)
-				return ret;
-		}
-
-		if (!IS_I830(dev) && !IS_845G(dev)) {
-			ret = BEGIN_LP_RING(2);
-			if (ret)
-				return ret;
-
-			if (INTEL_INFO(dev)->gen >= 4) {
-				OUT_RING(MI_BATCH_BUFFER_START | (2 << 6) | MI_BATCH_NON_SECURE_I965);
-				OUT_RING(batch->start);
-			} else {
-				OUT_RING(MI_BATCH_BUFFER_START | (2 << 6));
-				OUT_RING(batch->start | MI_BATCH_NON_SECURE);
-			}
-		} else {
-			ret = BEGIN_LP_RING(4);
-			if (ret)
-				return ret;
-
-			OUT_RING(MI_BATCH_BUFFER);
-			OUT_RING(batch->start | MI_BATCH_NON_SECURE);
-			OUT_RING(batch->start + batch->used - 4);
-			OUT_RING(0);
-		}
-		ADVANCE_LP_RING();
-	}
-
-
-	if (IS_G4X(dev) || IS_GEN5(dev)) {
-		if (BEGIN_LP_RING(2) == 0) {
-			OUT_RING(MI_FLUSH | MI_NO_WRITE_FLUSH | MI_INVALIDATE_ISP);
-			OUT_RING(MI_NOOP);
-			ADVANCE_LP_RING();
-		}
-	}
-
-	i915_emit_breadcrumb(dev);
-	return 0;
-}
-
-static int i915_dispatch_flip(struct drm_device *dev)
-{
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct drm_i915_master_private *master_priv =
-		dev->primary->master->driver_priv;
-	int ret;
-
-	if (!master_priv->sarea_priv)
-		return -EINVAL;
-
-	DRM_DEBUG_DRIVER("%s: page=%d pfCurrentPage=%d\n",
-			  __func__,
-			 dev_priv->dri1.current_page,
-			 master_priv->sarea_priv->pf_current_page);
-
-	i915_kernel_lost_context(dev);
-
-	ret = BEGIN_LP_RING(10);
-	if (ret)
-		return ret;
-
-	OUT_RING(MI_FLUSH | MI_READ_FLUSH);
-	OUT_RING(0);
-
-	OUT_RING(CMD_OP_DISPLAYBUFFER_INFO | ASYNC_FLIP);
-	OUT_RING(0);
-	if (dev_priv->dri1.current_page == 0) {
-		OUT_RING(dev_priv->dri1.back_offset);
-		dev_priv->dri1.current_page = 1;
-	} else {
-		OUT_RING(dev_priv->dri1.front_offset);
-		dev_priv->dri1.current_page = 0;
-	}
-	OUT_RING(0);
-
-	OUT_RING(MI_WAIT_FOR_EVENT | MI_WAIT_FOR_PLANE_A_FLIP);
-	OUT_RING(0);
-
-	ADVANCE_LP_RING();
-
-	master_priv->sarea_priv->last_enqueue = dev_priv->dri1.counter++;
-
-	if (BEGIN_LP_RING(4) == 0) {
-		OUT_RING(MI_STORE_DWORD_INDEX);
-		OUT_RING(I915_BREADCRUMB_INDEX << MI_STORE_DWORD_INDEX_SHIFT);
-		OUT_RING(dev_priv->dri1.counter);
-		OUT_RING(0);
-		ADVANCE_LP_RING();
-	}
-
-	master_priv->sarea_priv->pf_current_page = dev_priv->dri1.current_page;
-	return 0;
-}
-
-static int i915_quiescent(struct drm_device *dev)
-{
-	i915_kernel_lost_context(dev);
-	return intel_ring_idle(LP_RING(dev->dev_private));
-}
-
-static int i915_flush_ioctl(struct drm_device *dev, void *data,
-			    struct drm_file *file_priv)
-{
-	int ret;
-
-	if (drm_core_check_feature(dev, DRIVER_MODESET))
-		return -ENODEV;
-
-	RING_LOCK_TEST_WITH_RETURN(dev, file_priv);
-
-	mutex_lock(&dev->struct_mutex);
-	ret = i915_quiescent(dev);
-	mutex_unlock(&dev->struct_mutex);
-
-	return ret;
-}
-
-static int i915_batchbuffer(struct drm_device *dev, void *data,
-			    struct drm_file *file_priv)
-{
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct drm_i915_master_private *master_priv;
-	drm_i915_sarea_t *sarea_priv;
-	drm_i915_batchbuffer_t *batch = data;
-	int ret;
-	struct drm_clip_rect *cliprects = NULL;
-
-	if (drm_core_check_feature(dev, DRIVER_MODESET))
-		return -ENODEV;
-
-	master_priv = dev->primary->master->driver_priv;
-	sarea_priv = (drm_i915_sarea_t *) master_priv->sarea_priv;
-
-	if (!dev_priv->dri1.allow_batchbuffer) {
-		DRM_ERROR("Batchbuffer ioctl disabled\n");
-		return -EINVAL;
-	}
-
-	DRM_DEBUG_DRIVER("i915 batchbuffer, start %x used %d cliprects %d\n",
-			batch->start, batch->used, batch->num_cliprects);
-
-	RING_LOCK_TEST_WITH_RETURN(dev, file_priv);
-
-	if (batch->num_cliprects < 0)
-		return -EINVAL;
-
-	if (batch->num_cliprects) {
-		cliprects = kcalloc(batch->num_cliprects,
-				    sizeof(*cliprects),
-				    GFP_KERNEL);
-		if (cliprects == NULL)
-			return -ENOMEM;
-
-		ret = copy_from_user(cliprects, batch->cliprects,
-				     batch->num_cliprects *
-				     sizeof(struct drm_clip_rect));
-		if (ret != 0) {
-			ret = -EFAULT;
-			goto fail_free;
-		}
-	}
-
-	mutex_lock(&dev->struct_mutex);
-	ret = i915_dispatch_batchbuffer(dev, batch, cliprects);
-	mutex_unlock(&dev->struct_mutex);
-
-	if (sarea_priv)
-		sarea_priv->last_dispatch = READ_BREADCRUMB(dev_priv);
-
-fail_free:
-	kfree(cliprects);
-
-	return ret;
-}
-
-static int i915_cmdbuffer(struct drm_device *dev, void *data,
-			  struct drm_file *file_priv)
-{
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct drm_i915_master_private *master_priv;
-	drm_i915_sarea_t *sarea_priv;
-	drm_i915_cmdbuffer_t *cmdbuf = data;
-	struct drm_clip_rect *cliprects = NULL;
-	void *batch_data;
-	int ret;
-
-	DRM_DEBUG_DRIVER("i915 cmdbuffer, buf %p sz %d cliprects %d\n",
-			cmdbuf->buf, cmdbuf->sz, cmdbuf->num_cliprects);
-
-	if (drm_core_check_feature(dev, DRIVER_MODESET))
-		return -ENODEV;
-
-	master_priv = dev->primary->master->driver_priv;
-	sarea_priv = (drm_i915_sarea_t *) master_priv->sarea_priv;
-
-	RING_LOCK_TEST_WITH_RETURN(dev, file_priv);
-
-	if (cmdbuf->num_cliprects < 0)
-		return -EINVAL;
-
-	batch_data = kmalloc(cmdbuf->sz, GFP_KERNEL);
-	if (batch_data == NULL)
-		return -ENOMEM;
-
-	ret = copy_from_user(batch_data, cmdbuf->buf, cmdbuf->sz);
-	if (ret != 0) {
-		ret = -EFAULT;
-		goto fail_batch_free;
-	}
-
-	if (cmdbuf->num_cliprects) {
-		cliprects = kcalloc(cmdbuf->num_cliprects,
-				    sizeof(*cliprects), GFP_KERNEL);
-		if (cliprects == NULL) {
-			ret = -ENOMEM;
-			goto fail_batch_free;
-		}
-
-		ret = copy_from_user(cliprects, cmdbuf->cliprects,
-				     cmdbuf->num_cliprects *
-				     sizeof(struct drm_clip_rect));
-		if (ret != 0) {
-			ret = -EFAULT;
-			goto fail_clip_free;
-		}
-	}
-
-	mutex_lock(&dev->struct_mutex);
-	ret = i915_dispatch_cmdbuffer(dev, cmdbuf, cliprects, batch_data);
-	mutex_unlock(&dev->struct_mutex);
-	if (ret) {
-		DRM_ERROR("i915_dispatch_cmdbuffer failed\n");
-		goto fail_clip_free;
-	}
-
-	if (sarea_priv)
-		sarea_priv->last_dispatch = READ_BREADCRUMB(dev_priv);
-
-fail_clip_free:
-	kfree(cliprects);
-fail_batch_free:
-	kfree(batch_data);
-
-	return ret;
-}
-
-static int i915_emit_irq(struct drm_device *dev)
-{
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct drm_i915_master_private *master_priv = dev->primary->master->driver_priv;
-
-	i915_kernel_lost_context(dev);
-
-	DRM_DEBUG_DRIVER("\n");
-
-	dev_priv->dri1.counter++;
-	if (dev_priv->dri1.counter > 0x7FFFFFFFUL)
-		dev_priv->dri1.counter = 1;
-	if (master_priv->sarea_priv)
-		master_priv->sarea_priv->last_enqueue = dev_priv->dri1.counter;
-
-	if (BEGIN_LP_RING(4) == 0) {
-		OUT_RING(MI_STORE_DWORD_INDEX);
-		OUT_RING(I915_BREADCRUMB_INDEX << MI_STORE_DWORD_INDEX_SHIFT);
-		OUT_RING(dev_priv->dri1.counter);
-		OUT_RING(MI_USER_INTERRUPT);
-		ADVANCE_LP_RING();
-	}
-
-	return dev_priv->dri1.counter;
-}
-
-static int i915_wait_irq(struct drm_device *dev, int irq_nr)
-{
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct drm_i915_master_private *master_priv = dev->primary->master->driver_priv;
-	int ret = 0;
-	struct intel_engine_cs *ring = LP_RING(dev_priv);
-
-	DRM_DEBUG_DRIVER("irq_nr=%d breadcrumb=%d\n", irq_nr,
-		  READ_BREADCRUMB(dev_priv));
-
-	if (READ_BREADCRUMB(dev_priv) >= irq_nr) {
-		if (master_priv->sarea_priv)
-			master_priv->sarea_priv->last_dispatch = READ_BREADCRUMB(dev_priv);
-		return 0;
-	}
-
-	if (master_priv->sarea_priv)
-		master_priv->sarea_priv->perf_boxes |= I915_BOX_WAIT;
-
-	if (ring->irq_get(ring)) {
-		DRM_WAIT_ON(ret, ring->irq_queue, 3 * HZ,
-			    READ_BREADCRUMB(dev_priv) >= irq_nr);
-		ring->irq_put(ring);
-	} else if (wait_for(READ_BREADCRUMB(dev_priv) >= irq_nr, 3000))
-		ret = -EBUSY;
-
-	if (ret == -EBUSY) {
-		DRM_ERROR("EBUSY -- rec: %d emitted: %d\n",
-			  READ_BREADCRUMB(dev_priv), (int)dev_priv->dri1.counter);
-	}
-
-	return ret;
-}
-
-/* Needs the lock as it touches the ring.
- */
-static int i915_irq_emit(struct drm_device *dev, void *data,
-			 struct drm_file *file_priv)
-{
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	drm_i915_irq_emit_t *emit = data;
-	int result;
-
-	if (drm_core_check_feature(dev, DRIVER_MODESET))
-		return -ENODEV;
-
-	if (!dev_priv || !LP_RING(dev_priv)->buffer->virtual_start) {
-		DRM_ERROR("called with no initialization\n");
-		return -EINVAL;
-	}
-
-	RING_LOCK_TEST_WITH_RETURN(dev, file_priv);
-
-	mutex_lock(&dev->struct_mutex);
-	result = i915_emit_irq(dev);
-	mutex_unlock(&dev->struct_mutex);
-
-	if (copy_to_user(emit->irq_seq, &result, sizeof(int))) {
-		DRM_ERROR("copy_to_user\n");
-		return -EFAULT;
-	}
-
-	return 0;
-}
-
-/* Doesn't need the hardware lock.
- */
-static int i915_irq_wait(struct drm_device *dev, void *data,
-			 struct drm_file *file_priv)
-{
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	drm_i915_irq_wait_t *irqwait = data;
-
-	if (drm_core_check_feature(dev, DRIVER_MODESET))
-		return -ENODEV;
-
-	if (!dev_priv) {
-		DRM_ERROR("called with no initialization\n");
-		return -EINVAL;
-	}
-
-	return i915_wait_irq(dev, irqwait->irq_seq);
-}
-
-static int i915_vblank_pipe_get(struct drm_device *dev, void *data,
-			 struct drm_file *file_priv)
-{
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	drm_i915_vblank_pipe_t *pipe = data;
-
-	if (drm_core_check_feature(dev, DRIVER_MODESET))
-		return -ENODEV;
-
-	if (!dev_priv) {
-		DRM_ERROR("called with no initialization\n");
-		return -EINVAL;
-	}
-
-	pipe->pipe = DRM_I915_VBLANK_PIPE_A | DRM_I915_VBLANK_PIPE_B;
-
-	return 0;
-}
-
-/**
- * Schedule buffer swap at given vertical blank.
- */
-static int i915_vblank_swap(struct drm_device *dev, void *data,
-		     struct drm_file *file_priv)
-{
-	/* The delayed swap mechanism was fundamentally racy, and has been
-	 * removed.  The model was that the client requested a delayed flip/swap
-	 * from the kernel, then waited for vblank before continuing to perform
-	 * rendering.  The problem was that the kernel might wake the client
-	 * up before it dispatched the vblank swap (since the lock has to be
-	 * held while touching the ringbuffer), in which case the client would
-	 * clear and start the next frame before the swap occurred, and
-	 * flicker would occur in addition to likely missing the vblank.
-	 *
-	 * In the absence of this ioctl, userland falls back to a correct path
-	 * of waiting for a vblank, then dispatching the swap on its own.
-	 * Context switching to userland and back is plenty fast enough for
-	 * meeting the requirements of vblank swapping.
-	 */
-	return -EINVAL;
-}
-
-static int i915_flip_bufs(struct drm_device *dev, void *data,
-			  struct drm_file *file_priv)
-{
-	int ret;
-
-	if (drm_core_check_feature(dev, DRIVER_MODESET))
-		return -ENODEV;
-
-	DRM_DEBUG_DRIVER("%s\n", __func__);
-
-	RING_LOCK_TEST_WITH_RETURN(dev, file_priv);
-
-	mutex_lock(&dev->struct_mutex);
-	ret = i915_dispatch_flip(dev);
-	mutex_unlock(&dev->struct_mutex);
-
-	return ret;
-}
 
 static int i915_getparam(struct drm_device *dev, void *data,
 			 struct drm_file *file_priv)
@@ -934,21 +58,12 @@ static int i915_getparam(struct drm_device *dev, void *data,
 	drm_i915_getparam_t *param = data;
 	int value;
 
-	if (!dev_priv) {
-		DRM_ERROR("called with no initialization\n");
-		return -EINVAL;
-	}
-
 	switch (param->param) {
 	case I915_PARAM_IRQ_ACTIVE:
-		value = dev->pdev->irq ? 1 : 0;
-		break;
 	case I915_PARAM_ALLOW_BATCHBUFFER:
-		value = dev_priv->dri1.allow_batchbuffer ? 1 : 0;
-		break;
 	case I915_PARAM_LAST_DISPATCH:
-		value = READ_BREADCRUMB(dev_priv);
-		break;
+		/* Reject all old ums/dri params. */
+		return -ENODEV;
 	case I915_PARAM_CHIPSET_ID:
 		value = dev->pdev->device;
 		break;
@@ -999,7 +114,7 @@ static int i915_getparam(struct drm_device *dev, void *data,
 		value = HAS_WT(dev);
 		break;
 	case I915_PARAM_HAS_ALIASING_PPGTT:
-		value = dev_priv->mm.aliasing_ppgtt || USES_FULL_PPGTT(dev);
+		value = USES_PPGTT(dev);
 		break;
 	case I915_PARAM_HAS_WAIT_TIMEOUT:
 		value = 1;
@@ -1025,6 +140,9 @@ static int i915_getparam(struct drm_device *dev, void *data,
 	case I915_PARAM_CMD_PARSER_VERSION:
 		value = i915_cmd_parser_get_version();
 		break;
+	case I915_PARAM_HAS_COHERENT_PHYS_GTT:
+		value = 1;
+		break;
 	default:
 		DRM_DEBUG("Unknown parameter %d\n", param->param);
 		return -EINVAL;
@@ -1044,19 +162,13 @@ static int i915_setparam(struct drm_device *dev, void *data,
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	drm_i915_setparam_t *param = data;
 
-	if (!dev_priv) {
-		DRM_ERROR("called with no initialization\n");
-		return -EINVAL;
-	}
-
 	switch (param->param) {
 	case I915_SETPARAM_USE_MI_BATCHBUFFER_START:
-		break;
 	case I915_SETPARAM_TEX_LRU_LOG_GRANULARITY:
-		break;
 	case I915_SETPARAM_ALLOW_BATCHBUFFER:
-		dev_priv->dri1.allow_batchbuffer = param->value ? 1 : 0;
-		break;
+		/* Reject all old ums/dri params. */
+		return -ENODEV;
+
 	case I915_SETPARAM_NUM_USED_FENCES:
 		if (param->value > dev_priv->num_fence_regs ||
 		    param->value < 0)
@@ -1073,54 +185,6 @@ static int i915_setparam(struct drm_device *dev, void *data,
 	return 0;
 }
 
-static int i915_set_status_page(struct drm_device *dev, void *data,
-				struct drm_file *file_priv)
-{
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	drm_i915_hws_addr_t *hws = data;
-	struct intel_engine_cs *ring;
-
-	if (drm_core_check_feature(dev, DRIVER_MODESET))
-		return -ENODEV;
-
-	if (!I915_NEED_GFX_HWS(dev))
-		return -EINVAL;
-
-	if (!dev_priv) {
-		DRM_ERROR("called with no initialization\n");
-		return -EINVAL;
-	}
-
-	if (drm_core_check_feature(dev, DRIVER_MODESET)) {
-		WARN(1, "tried to set status page when mode setting active\n");
-		return 0;
-	}
-
-	DRM_DEBUG_DRIVER("set status page addr 0x%08x\n", (u32)hws->addr);
-
-	ring = LP_RING(dev_priv);
-	ring->status_page.gfx_addr = hws->addr & (0x1ffff<<12);
-
-	dev_priv->dri1.gfx_hws_cpu_addr =
-		ioremap_wc(dev_priv->gtt.mappable_base + hws->addr, 4096);
-	if (dev_priv->dri1.gfx_hws_cpu_addr == NULL) {
-		i915_dma_cleanup(dev);
-		ring->status_page.gfx_addr = 0;
-		DRM_ERROR("can not ioremap virtual address for"
-				" G33 hw status page\n");
-		return -ENOMEM;
-	}
-
-	memset_io(dev_priv->dri1.gfx_hws_cpu_addr, 0, PAGE_SIZE);
-	I915_WRITE(HWS_PGA, ring->status_page.gfx_addr);
-
-	DRM_DEBUG_DRIVER("load hws HWS_PGA with gfx mem 0x%x\n",
-			 ring->status_page.gfx_addr);
-	DRM_DEBUG_DRIVER("load hws at %p\n",
-			 ring->status_page.page_addr);
-	return 0;
-}
-
 static int i915_get_bridge_dev(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
@@ -1273,12 +337,12 @@ static void i915_switcheroo_set_state(struct pci_dev *pdev, enum vga_switcheroo_
 		dev->switch_power_state = DRM_SWITCH_POWER_CHANGING;
 		/* i915 resume handler doesn't set to D0 */
 		pci_set_power_state(dev->pdev, PCI_D0);
-		i915_resume(dev);
+		i915_resume_legacy(dev);
 		dev->switch_power_state = DRM_SWITCH_POWER_ON;
 	} else {
 		pr_err("switched off\n");
 		dev->switch_power_state = DRM_SWITCH_POWER_CHANGING;
-		i915_suspend(dev, pmm);
+		i915_suspend_legacy(dev, pmm);
 		dev->switch_power_state = DRM_SWITCH_POWER_OFF;
 	}
 }
@@ -1336,14 +400,7 @@ static int i915_load_modeset_init(struct drm_device *dev)
 
 	intel_power_domains_init_hw(dev_priv);
 
-	/*
-	 * We enable some interrupt sources in our postinstall hooks, so mark
-	 * interrupts as enabled _before_ actually enabling them to avoid
-	 * special cases in our ordering checks.
-	 */
-	dev_priv->pm._irqs_disabled = false;
-
-	ret = drm_irq_install(dev, dev->pdev->irq);
+	ret = intel_irq_install(dev_priv);
 	if (ret)
 		goto cleanup_gem_stolen;
 
@@ -1355,8 +412,6 @@ static int i915_load_modeset_init(struct drm_device *dev)
 	if (ret)
 		goto cleanup_irq;
 
-	INIT_WORK(&dev_priv->console_resume_work, intel_console_resume);
-
 	intel_modeset_gem_init(dev);
 
 	/* Always safe in the mode setting case. */
@@ -1370,7 +425,7 @@ static int i915_load_modeset_init(struct drm_device *dev)
 		goto cleanup_gem;
 
 	/* Only enable hotplug handling once the fbdev is fully set up. */
-	intel_hpd_init(dev);
+	intel_hpd_init(dev_priv);
 
 	/*
 	 * Some ports require correctly set-up hpd registers for detection to
@@ -1382,7 +437,7 @@ static int i915_load_modeset_init(struct drm_device *dev)
 	 * scanning against hotplug events. Hence do this first and ignore the
 	 * tiny window where we will loose hotplug notifactions.
 	 */
-	intel_fbdev_initial_config(dev);
+	async_schedule(intel_fbdev_initial_config, dev_priv);
 
 	drm_kms_helper_poll_init(dev);
 
@@ -1393,7 +448,6 @@ cleanup_gem:
 	i915_gem_cleanup_ringbuffer(dev);
 	i915_gem_context_fini(dev);
 	mutex_unlock(&dev->struct_mutex);
-	WARN_ON(dev_priv->mm.aliasing_ppgtt);
 cleanup_irq:
 	drm_irq_uninstall(dev);
 cleanup_gem_stolen:
@@ -1406,30 +460,6 @@ out:
 	return ret;
 }
 
-int i915_master_create(struct drm_device *dev, struct drm_master *master)
-{
-	struct drm_i915_master_private *master_priv;
-
-	master_priv = kzalloc(sizeof(*master_priv), GFP_KERNEL);
-	if (!master_priv)
-		return -ENOMEM;
-
-	master->driver_priv = master_priv;
-	return 0;
-}
-
-void i915_master_destroy(struct drm_device *dev, struct drm_master *master)
-{
-	struct drm_i915_master_private *master_priv = master->driver_priv;
-
-	if (!master_priv)
-		return;
-
-	kfree(master_priv);
-
-	master->driver_priv = NULL;
-}
-
 #if IS_ENABLED(CONFIG_FB)
 static int i915_kick_out_firmware_fb(struct drm_i915_private *dev_priv)
 {
@@ -1535,11 +565,11 @@ static void intel_device_info_runtime_init(struct drm_device *dev)
 
 	info = (struct intel_device_info *)&dev_priv->info;
 
-	if (IS_VALLEYVIEW(dev))
-		for_each_pipe(pipe)
+	if (IS_VALLEYVIEW(dev) || INTEL_INFO(dev)->gen == 9)
+		for_each_pipe(dev_priv, pipe)
 			info->num_sprites[pipe] = 2;
 	else
-		for_each_pipe(pipe)
+		for_each_pipe(dev_priv, pipe)
 			info->num_sprites[pipe] = 1;
 
 	if (i915.disable_display) {
@@ -1608,13 +638,14 @@ int i915_driver_load(struct drm_device *dev, unsigned long flags)
 	dev->dev_private = dev_priv;
 	dev_priv->dev = dev;
 
-	/* copy initial configuration to dev_priv->info */
+	/* Setup the write-once "constant" device info */
 	device_info = (struct intel_device_info *)&dev_priv->info;
-	*device_info = *info;
+	memcpy(device_info, info, sizeof(dev_priv->info));
+	device_info->device_id = dev->pdev->device;
 
 	spin_lock_init(&dev_priv->irq_lock);
 	spin_lock_init(&dev_priv->gpu_error.lock);
-	spin_lock_init(&dev_priv->backlight_lock);
+	mutex_init(&dev_priv->backlight_lock);
 	spin_lock_init(&dev_priv->uncore.lock);
 	spin_lock_init(&dev_priv->mm.object_stat_lock);
 	spin_lock_init(&dev_priv->mmio_flip_lock);
@@ -1670,15 +701,17 @@ int i915_driver_load(struct drm_device *dev, unsigned long flags)
 		goto out_regs;
 
 	if (drm_core_check_feature(dev, DRIVER_MODESET)) {
-		ret = i915_kick_out_vgacon(dev_priv);
+		/* WARNING: Apparently we must kick fbdev drivers before vgacon,
+		 * otherwise the vga fbdev driver falls over. */
+		ret = i915_kick_out_firmware_fb(dev_priv);
 		if (ret) {
-			DRM_ERROR("failed to remove conflicting VGA console\n");
+			DRM_ERROR("failed to remove conflicting framebuffer drivers\n");
 			goto out_gtt;
 		}
 
-		ret = i915_kick_out_firmware_fb(dev_priv);
+		ret = i915_kick_out_vgacon(dev_priv);
 		if (ret) {
-			DRM_ERROR("failed to remove conflicting framebuffer drivers\n");
+			DRM_ERROR("failed to remove conflicting VGA console\n");
 			goto out_gtt;
 		}
 	}
@@ -1740,7 +773,7 @@ int i915_driver_load(struct drm_device *dev, unsigned long flags)
 		goto out_freewq;
 	}
 
-	intel_irq_init(dev);
+	intel_irq_init(dev_priv);
 	intel_uncore_sanitize(dev);
 
 	/* Try to make sure MCHBAR is enabled before poking at it */
@@ -1782,9 +815,6 @@ int i915_driver_load(struct drm_device *dev, unsigned long flags)
 			DRM_ERROR("failed to init modeset\n");
 			goto out_power_well;
 		}
-	} else {
-		/* Start out suspended in ums mode. */
-		dev_priv->ums.mm_suspended = 1;
 	}
 
 	i915_setup_sysfs(dev);
@@ -1798,12 +828,12 @@ int i915_driver_load(struct drm_device *dev, unsigned long flags)
 	if (IS_GEN5(dev))
 		intel_gpu_ips_init(dev_priv);
 
-	intel_init_runtime_pm(dev_priv);
+	intel_runtime_pm_enable(dev_priv);
 
 	return 0;
 
 out_power_well:
-	intel_power_domains_remove(dev_priv);
+	intel_power_domains_fini(dev_priv);
 	drm_vblank_cleanup(dev);
 out_gem_unload:
 	WARN_ON(unregister_oom_notifier(&dev_priv->mm.oom_notifier));
@@ -1822,7 +852,7 @@ out_mtrrfree:
 	arch_phys_wc_del(dev_priv->gtt.mtrr);
 	io_mapping_free(dev_priv->gtt.mappable);
 out_gtt:
-	dev_priv->gtt.base.cleanup(&dev_priv->gtt.base);
+	i915_global_gtt_cleanup(dev);
 out_regs:
 	intel_uncore_fini(dev);
 	pci_iounmap(dev->pdev, dev_priv->regs);
@@ -1846,16 +876,10 @@ int i915_driver_unload(struct drm_device *dev)
 		return ret;
 	}
 
-	intel_fini_runtime_pm(dev_priv);
+	intel_power_domains_fini(dev_priv);
 
 	intel_gpu_ips_teardown();
 
-	/* The i915.ko module is still not prepared to be loaded when
-	 * the power well is not enabled, so just enable it in case
-	 * we're going to unload/reload. */
-	intel_display_set_init_power(dev_priv, true);
-	intel_power_domains_remove(dev_priv);
-
 	i915_teardown_sysfs(dev);
 
 	WARN_ON(unregister_oom_notifier(&dev_priv->mm.oom_notifier));
@@ -1866,10 +890,13 @@ int i915_driver_unload(struct drm_device *dev)
 
 	acpi_video_unregister();
 
-	if (drm_core_check_feature(dev, DRIVER_MODESET)) {
+	if (drm_core_check_feature(dev, DRIVER_MODESET))
 		intel_fbdev_fini(dev);
+
+	drm_vblank_cleanup(dev);
+
+	if (drm_core_check_feature(dev, DRIVER_MODESET)) {
 		intel_modeset_cleanup(dev);
-		cancel_work_sync(&dev_priv->console_resume_work);
 
 		/*
 		 * free the memory space allocated for the child device
@@ -1902,18 +929,10 @@ int i915_driver_unload(struct drm_device *dev)
 		mutex_lock(&dev->struct_mutex);
 		i915_gem_cleanup_ringbuffer(dev);
 		i915_gem_context_fini(dev);
-		WARN_ON(dev_priv->mm.aliasing_ppgtt);
 		mutex_unlock(&dev->struct_mutex);
 		i915_gem_cleanup_stolen(dev);
-
-		if (!I915_NEED_GFX_HWS(dev))
-			i915_free_hws(dev);
 	}
 
-	WARN_ON(!list_empty(&dev_priv->vm_list));
-
-	drm_vblank_cleanup(dev);
-
 	intel_teardown_gmbus(dev);
 	intel_teardown_mchbar(dev);
 
@@ -1921,7 +940,7 @@ int i915_driver_unload(struct drm_device *dev)
 	destroy_workqueue(dev_priv->wq);
 	pm_qos_remove_request(&dev_priv->pm_qos);
 
-	dev_priv->gtt.base.cleanup(&dev_priv->gtt.base);
+	i915_global_gtt_cleanup(dev);
 
 	intel_uncore_fini(dev);
 	if (dev_priv->regs != NULL)
@@ -1961,23 +980,8 @@ int i915_driver_open(struct drm_device *dev, struct drm_file *file)
  */
 void i915_driver_lastclose(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
-
-	/* On gen6+ we refuse to init without kms enabled, but then the drm core
-	 * goes right around and calls lastclose. Check for this and don't clean
-	 * up anything. */
-	if (!dev_priv)
-		return;
-
-	if (drm_core_check_feature(dev, DRIVER_MODESET)) {
-		intel_fbdev_restore_mode(dev);
-		vga_switcheroo_process_delayed_switch();
-		return;
-	}
-
-	i915_gem_lastclose(dev);
-
-	i915_dma_cleanup(dev);
+	intel_fbdev_restore_mode(dev);
+	vga_switcheroo_process_delayed_switch();
 }
 
 void i915_driver_preclose(struct drm_device *dev, struct drm_file *file)
@@ -1986,6 +990,9 @@ void i915_driver_preclose(struct drm_device *dev, struct drm_file *file)
 	i915_gem_context_close(dev, file);
 	i915_gem_release(dev, file);
 	mutex_unlock(&dev->struct_mutex);
+
+	if (drm_core_check_feature(dev, DRIVER_MODESET))
+		intel_modeset_preclose(dev, file);
 }
 
 void i915_driver_postclose(struct drm_device *dev, struct drm_file *file)
@@ -1998,24 +1005,24 @@ void i915_driver_postclose(struct drm_device *dev, struct drm_file *file)
 }
 
 const struct drm_ioctl_desc i915_ioctls[] = {
-	DRM_IOCTL_DEF_DRV(I915_INIT, i915_dma_init, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
-	DRM_IOCTL_DEF_DRV(I915_FLUSH, i915_flush_ioctl, DRM_AUTH),
-	DRM_IOCTL_DEF_DRV(I915_FLIP, i915_flip_bufs, DRM_AUTH),
-	DRM_IOCTL_DEF_DRV(I915_BATCHBUFFER, i915_batchbuffer, DRM_AUTH),
-	DRM_IOCTL_DEF_DRV(I915_IRQ_EMIT, i915_irq_emit, DRM_AUTH),
-	DRM_IOCTL_DEF_DRV(I915_IRQ_WAIT, i915_irq_wait, DRM_AUTH),
+	DRM_IOCTL_DEF_DRV(I915_INIT, drm_noop, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
+	DRM_IOCTL_DEF_DRV(I915_FLUSH, drm_noop, DRM_AUTH),
+	DRM_IOCTL_DEF_DRV(I915_FLIP, drm_noop, DRM_AUTH),
+	DRM_IOCTL_DEF_DRV(I915_BATCHBUFFER, drm_noop, DRM_AUTH),
+	DRM_IOCTL_DEF_DRV(I915_IRQ_EMIT, drm_noop, DRM_AUTH),
+	DRM_IOCTL_DEF_DRV(I915_IRQ_WAIT, drm_noop, DRM_AUTH),
 	DRM_IOCTL_DEF_DRV(I915_GETPARAM, i915_getparam, DRM_AUTH|DRM_RENDER_ALLOW),
 	DRM_IOCTL_DEF_DRV(I915_SETPARAM, i915_setparam, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
 	DRM_IOCTL_DEF_DRV(I915_ALLOC, drm_noop, DRM_AUTH),
 	DRM_IOCTL_DEF_DRV(I915_FREE, drm_noop, DRM_AUTH),
 	DRM_IOCTL_DEF_DRV(I915_INIT_HEAP, drm_noop, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
-	DRM_IOCTL_DEF_DRV(I915_CMDBUFFER, i915_cmdbuffer, DRM_AUTH),
+	DRM_IOCTL_DEF_DRV(I915_CMDBUFFER, drm_noop, DRM_AUTH),
 	DRM_IOCTL_DEF_DRV(I915_DESTROY_HEAP,  drm_noop, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
 	DRM_IOCTL_DEF_DRV(I915_SET_VBLANK_PIPE,  drm_noop, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
-	DRM_IOCTL_DEF_DRV(I915_GET_VBLANK_PIPE,  i915_vblank_pipe_get, DRM_AUTH),
-	DRM_IOCTL_DEF_DRV(I915_VBLANK_SWAP, i915_vblank_swap, DRM_AUTH),
-	DRM_IOCTL_DEF_DRV(I915_HWS_ADDR, i915_set_status_page, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
-	DRM_IOCTL_DEF_DRV(I915_GEM_INIT, i915_gem_init_ioctl, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY|DRM_UNLOCKED),
+	DRM_IOCTL_DEF_DRV(I915_GET_VBLANK_PIPE,  drm_noop, DRM_AUTH),
+	DRM_IOCTL_DEF_DRV(I915_VBLANK_SWAP, drm_noop, DRM_AUTH),
+	DRM_IOCTL_DEF_DRV(I915_HWS_ADDR, drm_noop, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
+	DRM_IOCTL_DEF_DRV(I915_GEM_INIT, drm_noop, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY|DRM_UNLOCKED),
 	DRM_IOCTL_DEF_DRV(I915_GEM_EXECBUFFER, i915_gem_execbuffer, DRM_AUTH|DRM_UNLOCKED),
 	DRM_IOCTL_DEF_DRV(I915_GEM_EXECBUFFER2, i915_gem_execbuffer2, DRM_AUTH|DRM_UNLOCKED|DRM_RENDER_ALLOW),
 	DRM_IOCTL_DEF_DRV(I915_GEM_PIN, i915_gem_pin_ioctl, DRM_AUTH|DRM_ROOT_ONLY|DRM_UNLOCKED),
@@ -2024,8 +1031,8 @@ const struct drm_ioctl_desc i915_ioctls[] = {
 	DRM_IOCTL_DEF_DRV(I915_GEM_SET_CACHING, i915_gem_set_caching_ioctl, DRM_UNLOCKED|DRM_RENDER_ALLOW),
 	DRM_IOCTL_DEF_DRV(I915_GEM_GET_CACHING, i915_gem_get_caching_ioctl, DRM_UNLOCKED|DRM_RENDER_ALLOW),
 	DRM_IOCTL_DEF_DRV(I915_GEM_THROTTLE, i915_gem_throttle_ioctl, DRM_AUTH|DRM_UNLOCKED|DRM_RENDER_ALLOW),
-	DRM_IOCTL_DEF_DRV(I915_GEM_ENTERVT, i915_gem_entervt_ioctl, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY|DRM_UNLOCKED),
-	DRM_IOCTL_DEF_DRV(I915_GEM_LEAVEVT, i915_gem_leavevt_ioctl, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY|DRM_UNLOCKED),
+	DRM_IOCTL_DEF_DRV(I915_GEM_ENTERVT, drm_noop, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY|DRM_UNLOCKED),
+	DRM_IOCTL_DEF_DRV(I915_GEM_LEAVEVT, drm_noop, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY|DRM_UNLOCKED),
 	DRM_IOCTL_DEF_DRV(I915_GEM_CREATE, i915_gem_create_ioctl, DRM_UNLOCKED|DRM_RENDER_ALLOW),
 	DRM_IOCTL_DEF_DRV(I915_GEM_PREAD, i915_gem_pread_ioctl, DRM_UNLOCKED|DRM_RENDER_ALLOW),
 	DRM_IOCTL_DEF_DRV(I915_GEM_PWRITE, i915_gem_pwrite_ioctl, DRM_UNLOCKED|DRM_RENDER_ALLOW),
diff --git a/drivers/gpu/drm/i915/i915_drv.c b/drivers/gpu/drm/i915/i915_drv.c
index e27cdbe9d524..574057cd1d09 100644
--- a/drivers/gpu/drm/i915/i915_drv.c
+++ b/drivers/gpu/drm/i915/i915_drv.c
@@ -356,6 +356,19 @@ static const struct intel_device_info intel_cherryview_info = {
 	CURSOR_OFFSETS,
 };
 
+static const struct intel_device_info intel_skylake_info = {
+	.is_preliminary = 1,
+	.is_skylake = 1,
+	.gen = 9, .num_pipes = 3,
+	.need_gfx_hws = 1, .has_hotplug = 1,
+	.ring_mask = RENDER_RING | BSD_RING | BLT_RING | VEBOX_RING,
+	.has_llc = 1,
+	.has_ddi = 1,
+	.has_fbc = 1,
+	GEN_DEFAULT_PIPEOFFSETS,
+	IVB_CURSOR_OFFSETS,
+};
+
 /*
  * Make sure any device matches here are from most specific to most
  * general.  For example, since the Quanta match is based on the subsystem
@@ -392,7 +405,8 @@ static const struct intel_device_info intel_cherryview_info = {
 	INTEL_BDW_GT12D_IDS(&intel_broadwell_d_info),	\
 	INTEL_BDW_GT3M_IDS(&intel_broadwell_gt3m_info),	\
 	INTEL_BDW_GT3D_IDS(&intel_broadwell_gt3d_info), \
-	INTEL_CHV_IDS(&intel_cherryview_info)
+	INTEL_CHV_IDS(&intel_cherryview_info),	\
+	INTEL_SKL_IDS(&intel_skylake_info)
 
 static const struct pci_device_id pciidlist[] = {		/* aka */
 	INTEL_PCI_IDS,
@@ -449,7 +463,7 @@ void intel_detect_pch(struct drm_device *dev)
 				dev_priv->pch_type = PCH_LPT;
 				DRM_DEBUG_KMS("Found LynxPoint PCH\n");
 				WARN_ON(!IS_HASWELL(dev));
-				WARN_ON(IS_ULT(dev));
+				WARN_ON(IS_HSW_ULT(dev));
 			} else if (IS_BROADWELL(dev)) {
 				dev_priv->pch_type = PCH_LPT;
 				dev_priv->pch_id =
@@ -460,7 +474,15 @@ void intel_detect_pch(struct drm_device *dev)
 				dev_priv->pch_type = PCH_LPT;
 				DRM_DEBUG_KMS("Found LynxPoint LP PCH\n");
 				WARN_ON(!IS_HASWELL(dev));
-				WARN_ON(!IS_ULT(dev));
+				WARN_ON(!IS_HSW_ULT(dev));
+			} else if (id == INTEL_PCH_SPT_DEVICE_ID_TYPE) {
+				dev_priv->pch_type = PCH_SPT;
+				DRM_DEBUG_KMS("Found SunrisePoint PCH\n");
+				WARN_ON(!IS_SKYLAKE(dev));
+			} else if (id == INTEL_PCH_SPT_LP_DEVICE_ID_TYPE) {
+				dev_priv->pch_type = PCH_SPT;
+				DRM_DEBUG_KMS("Found SunrisePoint LP PCH\n");
+				WARN_ON(!IS_SKYLAKE(dev));
 			} else
 				continue;
 
@@ -481,6 +503,10 @@ bool i915_semaphore_is_enabled(struct drm_device *dev)
 	if (i915.semaphores >= 0)
 		return i915.semaphores;
 
+	/* TODO: make semaphores and Execlists play nicely together */
+	if (i915.enable_execlists)
+		return false;
+
 	/* Until we get further testing... */
 	if (IS_GEN8(dev))
 		return false;
@@ -524,7 +550,11 @@ static void intel_suspend_encoders(struct drm_i915_private *dev_priv)
 	drm_modeset_unlock_all(dev);
 }
 
-static int i915_drm_freeze(struct drm_device *dev)
+static int intel_suspend_complete(struct drm_i915_private *dev_priv);
+static int vlv_resume_prepare(struct drm_i915_private *dev_priv,
+			      bool rpm_resume);
+
+static int i915_drm_suspend(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct drm_crtc *crtc;
@@ -554,6 +584,8 @@ static int i915_drm_freeze(struct drm_device *dev)
 			return error;
 		}
 
+		intel_suspend_gt_powersave(dev);
+
 		/*
 		 * Disable CRTCs directly since we want to preserve sw state
 		 * for _thaw. Also, power gate the CRTC power wells.
@@ -565,16 +597,12 @@ static int i915_drm_freeze(struct drm_device *dev)
 
 		intel_dp_mst_suspend(dev);
 
-		flush_delayed_work(&dev_priv->rps.delayed_resume_work);
-
-		intel_runtime_pm_disable_interrupts(dev);
+		intel_runtime_pm_disable_interrupts(dev_priv);
 		intel_hpd_cancel_work(dev_priv);
 
 		intel_suspend_encoders(dev_priv);
 
-		intel_suspend_gt_powersave(dev);
-
-		intel_modeset_suspend_hw(dev);
+		intel_suspend_hw(dev);
 	}
 
 	i915_gem_suspend_gtt_mappings(dev);
@@ -591,9 +619,7 @@ static int i915_drm_freeze(struct drm_device *dev)
 	intel_uncore_forcewake_reset(dev, false);
 	intel_opregion_fini(dev);
 
-	console_lock();
-	intel_fbdev_set_suspend(dev, FBINFO_STATE_SUSPENDED);
-	console_unlock();
+	intel_fbdev_set_suspend(dev, FBINFO_STATE_SUSPENDED, true);
 
 	dev_priv->suspend_count++;
 
@@ -602,7 +628,26 @@ static int i915_drm_freeze(struct drm_device *dev)
 	return 0;
 }
 
-int i915_suspend(struct drm_device *dev, pm_message_t state)
+static int i915_drm_suspend_late(struct drm_device *drm_dev)
+{
+	struct drm_i915_private *dev_priv = drm_dev->dev_private;
+	int ret;
+
+	ret = intel_suspend_complete(dev_priv);
+
+	if (ret) {
+		DRM_ERROR("Suspend complete failed: %d\n", ret);
+
+		return ret;
+	}
+
+	pci_disable_device(drm_dev->pdev);
+	pci_set_power_state(drm_dev->pdev, PCI_D3hot);
+
+	return 0;
+}
+
+int i915_suspend_legacy(struct drm_device *dev, pm_message_t state)
 {
 	int error;
 
@@ -612,58 +657,25 @@ int i915_suspend(struct drm_device *dev, pm_message_t state)
 		return -ENODEV;
 	}
 
-	if (state.event == PM_EVENT_PRETHAW)
-		return 0;
-
+	if (WARN_ON_ONCE(state.event != PM_EVENT_SUSPEND &&
+			 state.event != PM_EVENT_FREEZE))
+		return -EINVAL;
 
 	if (dev->switch_power_state == DRM_SWITCH_POWER_OFF)
 		return 0;
 
-	error = i915_drm_freeze(dev);
+	error = i915_drm_suspend(dev);
 	if (error)
 		return error;
 
-	if (state.event == PM_EVENT_SUSPEND) {
-		/* Shut down the device */
-		pci_disable_device(dev->pdev);
-		pci_set_power_state(dev->pdev, PCI_D3hot);
-	}
-
-	return 0;
-}
-
-void intel_console_resume(struct work_struct *work)
-{
-	struct drm_i915_private *dev_priv =
-		container_of(work, struct drm_i915_private,
-			     console_resume_work);
-	struct drm_device *dev = dev_priv->dev;
-
-	console_lock();
-	intel_fbdev_set_suspend(dev, FBINFO_STATE_RUNNING);
-	console_unlock();
+	return i915_drm_suspend_late(dev);
 }
 
-static int i915_drm_thaw_early(struct drm_device *dev)
+static int i915_drm_resume(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 
-	if (IS_HASWELL(dev) || IS_BROADWELL(dev))
-		hsw_disable_pc8(dev_priv);
-
-	intel_uncore_early_sanitize(dev, true);
-	intel_uncore_sanitize(dev);
-	intel_power_domains_init_hw(dev_priv);
-
-	return 0;
-}
-
-static int __i915_drm_thaw(struct drm_device *dev, bool restore_gtt_mappings)
-{
-	struct drm_i915_private *dev_priv = dev->dev_private;
-
-	if (drm_core_check_feature(dev, DRIVER_MODESET) &&
-	    restore_gtt_mappings) {
+	if (drm_core_check_feature(dev, DRIVER_MODESET)) {
 		mutex_lock(&dev->struct_mutex);
 		i915_gem_restore_gtt_mappings(dev);
 		mutex_unlock(&dev->struct_mutex);
@@ -684,47 +696,36 @@ static int __i915_drm_thaw(struct drm_device *dev, bool restore_gtt_mappings)
 		}
 		mutex_unlock(&dev->struct_mutex);
 
-		intel_runtime_pm_restore_interrupts(dev);
+		/* We need working interrupts for modeset enabling ... */
+		intel_runtime_pm_enable_interrupts(dev_priv);
 
 		intel_modeset_init_hw(dev);
 
-		{
-			unsigned long irqflags;
-			spin_lock_irqsave(&dev_priv->irq_lock, irqflags);
-			if (dev_priv->display.hpd_irq_setup)
-				dev_priv->display.hpd_irq_setup(dev);
-			spin_unlock_irqrestore(&dev_priv->irq_lock, irqflags);
-		}
+		spin_lock_irq(&dev_priv->irq_lock);
+		if (dev_priv->display.hpd_irq_setup)
+			dev_priv->display.hpd_irq_setup(dev);
+		spin_unlock_irq(&dev_priv->irq_lock);
 
-		intel_dp_mst_resume(dev);
 		drm_modeset_lock_all(dev);
 		intel_modeset_setup_hw_state(dev, true);
 		drm_modeset_unlock_all(dev);
 
+		intel_dp_mst_resume(dev);
+
 		/*
 		 * ... but also need to make sure that hotplug processing
 		 * doesn't cause havoc. Like in the driver load code we don't
 		 * bother with the tiny race here where we might loose hotplug
 		 * notifications.
 		 * */
-		intel_hpd_init(dev);
+		intel_hpd_init(dev_priv);
 		/* Config may have changed between suspend and resume */
 		drm_helper_hpd_irq_event(dev);
 	}
 
 	intel_opregion_init(dev);
 
-	/*
-	 * The console lock can be pretty contented on resume due
-	 * to all the printk activity.  Try to keep it out of the hot
-	 * path of resume if possible.
-	 */
-	if (console_trylock()) {
-		intel_fbdev_set_suspend(dev, FBINFO_STATE_RUNNING);
-		console_unlock();
-	} else {
-		schedule_work(&dev_priv->console_resume_work);
-	}
+	intel_fbdev_set_suspend(dev, FBINFO_STATE_RUNNING, false);
 
 	mutex_lock(&dev_priv->modeset_restore_lock);
 	dev_priv->modeset_restore = MODESET_DONE;
@@ -732,21 +733,15 @@ static int __i915_drm_thaw(struct drm_device *dev, bool restore_gtt_mappings)
 
 	intel_opregion_notify_adapter(dev, PCI_D0);
 
-	return 0;
-}
-
-static int i915_drm_thaw(struct drm_device *dev)
-{
-	if (drm_core_check_feature(dev, DRIVER_MODESET))
-		i915_check_and_clear_faults(dev);
+	drm_kms_helper_poll_enable(dev);
 
-	return __i915_drm_thaw(dev, true);
+	return 0;
 }
 
-static int i915_resume_early(struct drm_device *dev)
+static int i915_drm_resume_early(struct drm_device *dev)
 {
-	if (dev->switch_power_state == DRM_SWITCH_POWER_OFF)
-		return 0;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	int ret = 0;
 
 	/*
 	 * We have a resume ordering issue with the snd-hda driver also
@@ -762,33 +757,34 @@ static int i915_resume_early(struct drm_device *dev)
 
 	pci_set_master(dev->pdev);
 
-	return i915_drm_thaw_early(dev);
+	if (IS_VALLEYVIEW(dev_priv))
+		ret = vlv_resume_prepare(dev_priv, false);
+	if (ret)
+		DRM_ERROR("Resume prepare failed: %d,Continuing resume\n", ret);
+
+	intel_uncore_early_sanitize(dev, true);
+
+	if (IS_HASWELL(dev_priv) || IS_BROADWELL(dev_priv))
+		hsw_disable_pc8(dev_priv);
+
+	intel_uncore_sanitize(dev);
+	intel_power_domains_init_hw(dev_priv);
+
+	return ret;
 }
 
-int i915_resume(struct drm_device *dev)
+int i915_resume_legacy(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
 	int ret;
 
-	/*
-	 * Platforms with opregion should have sane BIOS, older ones (gen3 and
-	 * earlier) need to restore the GTT mappings since the BIOS might clear
-	 * all our scratch PTEs.
-	 */
-	ret = __i915_drm_thaw(dev, !dev_priv->opregion.header);
+	if (dev->switch_power_state == DRM_SWITCH_POWER_OFF)
+		return 0;
+
+	ret = i915_drm_resume_early(dev);
 	if (ret)
 		return ret;
 
-	drm_kms_helper_poll_enable(dev);
-	return 0;
-}
-
-static int i915_resume_legacy(struct drm_device *dev)
-{
-	i915_resume_early(dev);
-	i915_resume(dev);
-
-	return 0;
+	return i915_drm_resume(dev);
 }
 
 /**
@@ -815,6 +811,8 @@ int i915_reset(struct drm_device *dev)
 	if (!i915.reset)
 		return 0;
 
+	intel_reset_gt_powersave(dev);
+
 	mutex_lock(&dev->struct_mutex);
 
 	i915_gem_reset(dev);
@@ -834,6 +832,9 @@ int i915_reset(struct drm_device *dev)
 		}
 	}
 
+	if (i915_stop_ring_allow_warn(dev_priv))
+		pr_notice("drm/i915: Resetting chip after gpu hang\n");
+
 	if (ret) {
 		DRM_ERROR("Failed to reset chip: %i\n", ret);
 		mutex_unlock(&dev->struct_mutex);
@@ -854,11 +855,14 @@ int i915_reset(struct drm_device *dev)
 	 * was running at the time of the reset (i.e. we weren't VT
 	 * switched away).
 	 */
-	if (drm_core_check_feature(dev, DRIVER_MODESET) ||
-			!dev_priv->ums.mm_suspended) {
-		dev_priv->ums.mm_suspended = 0;
+	if (drm_core_check_feature(dev, DRIVER_MODESET)) {
+		/* Used to prevent gem_check_wedged returning -EAGAIN during gpu reset */
+		dev_priv->gpu_error.reload_in_reset = true;
 
 		ret = i915_gem_init_hw(dev);
+
+		dev_priv->gpu_error.reload_in_reset = false;
+
 		mutex_unlock(&dev->struct_mutex);
 		if (ret) {
 			DRM_ERROR("Failed hw init on reset %d\n", ret);
@@ -878,9 +882,7 @@ int i915_reset(struct drm_device *dev)
 		 * of re-init after reset.
 		 */
 		if (INTEL_INFO(dev)->gen > 5)
-			intel_reset_gt_powersave(dev);
-
-		intel_hpd_init(dev);
+			intel_enable_gt_powersave(dev);
 	} else {
 		mutex_unlock(&dev->struct_mutex);
 	}
@@ -933,14 +935,13 @@ static int i915_pm_suspend(struct device *dev)
 	if (drm_dev->switch_power_state == DRM_SWITCH_POWER_OFF)
 		return 0;
 
-	return i915_drm_freeze(drm_dev);
+	return i915_drm_suspend(drm_dev);
 }
 
 static int i915_pm_suspend_late(struct device *dev)
 {
 	struct pci_dev *pdev = to_pci_dev(dev);
 	struct drm_device *drm_dev = pci_get_drvdata(pdev);
-	struct drm_i915_private *dev_priv = drm_dev->dev_private;
 
 	/*
 	 * We have a suspedn ordering issue with the snd-hda driver also
@@ -954,13 +955,7 @@ static int i915_pm_suspend_late(struct device *dev)
 	if (drm_dev->switch_power_state == DRM_SWITCH_POWER_OFF)
 		return 0;
 
-	if (IS_HASWELL(drm_dev) || IS_BROADWELL(drm_dev))
-		hsw_enable_pc8(dev_priv);
-
-	pci_disable_device(pdev);
-	pci_set_power_state(pdev, PCI_D3hot);
-
-	return 0;
+	return i915_drm_suspend_late(drm_dev);
 }
 
 static int i915_pm_resume_early(struct device *dev)
@@ -968,77 +963,30 @@ static int i915_pm_resume_early(struct device *dev)
 	struct pci_dev *pdev = to_pci_dev(dev);
 	struct drm_device *drm_dev = pci_get_drvdata(pdev);
 
-	return i915_resume_early(drm_dev);
-}
-
-static int i915_pm_resume(struct device *dev)
-{
-	struct pci_dev *pdev = to_pci_dev(dev);
-	struct drm_device *drm_dev = pci_get_drvdata(pdev);
-
-	return i915_resume(drm_dev);
-}
-
-static int i915_pm_freeze(struct device *dev)
-{
-	struct pci_dev *pdev = to_pci_dev(dev);
-	struct drm_device *drm_dev = pci_get_drvdata(pdev);
-
-	if (!drm_dev || !drm_dev->dev_private) {
-		dev_err(dev, "DRM not initialized, aborting suspend.\n");
-		return -ENODEV;
-	}
-
-	return i915_drm_freeze(drm_dev);
-}
-
-static int i915_pm_thaw_early(struct device *dev)
-{
-	struct pci_dev *pdev = to_pci_dev(dev);
-	struct drm_device *drm_dev = pci_get_drvdata(pdev);
+	if (drm_dev->switch_power_state == DRM_SWITCH_POWER_OFF)
+		return 0;
 
-	return i915_drm_thaw_early(drm_dev);
+	return i915_drm_resume_early(drm_dev);
 }
 
-static int i915_pm_thaw(struct device *dev)
+static int i915_pm_resume(struct device *dev)
 {
 	struct pci_dev *pdev = to_pci_dev(dev);
 	struct drm_device *drm_dev = pci_get_drvdata(pdev);
 
-	return i915_drm_thaw(drm_dev);
-}
-
-static int i915_pm_poweroff(struct device *dev)
-{
-	struct pci_dev *pdev = to_pci_dev(dev);
-	struct drm_device *drm_dev = pci_get_drvdata(pdev);
+	if (drm_dev->switch_power_state == DRM_SWITCH_POWER_OFF)
+		return 0;
 
-	return i915_drm_freeze(drm_dev);
+	return i915_drm_resume(drm_dev);
 }
 
-static int hsw_runtime_suspend(struct drm_i915_private *dev_priv)
+static int hsw_suspend_complete(struct drm_i915_private *dev_priv)
 {
 	hsw_enable_pc8(dev_priv);
 
 	return 0;
 }
 
-static int snb_runtime_resume(struct drm_i915_private *dev_priv)
-{
-	struct drm_device *dev = dev_priv->dev;
-
-	intel_init_pch_refclk(dev);
-
-	return 0;
-}
-
-static int hsw_runtime_resume(struct drm_i915_private *dev_priv)
-{
-	hsw_disable_pc8(dev_priv);
-
-	return 0;
-}
-
 /*
  * Save all Gunit registers that may be lost after a D3 and a subsequent
  * S0i[R123] transition. The list of registers needing a save/restore is
@@ -1328,7 +1276,7 @@ static void vlv_check_no_gt_access(struct drm_i915_private *dev_priv)
 	I915_WRITE(VLV_GTLC_PW_STATUS, VLV_GTLC_ALLOWWAKEERR);
 }
 
-static int vlv_runtime_suspend(struct drm_i915_private *dev_priv)
+static int vlv_suspend_complete(struct drm_i915_private *dev_priv)
 {
 	u32 mask;
 	int err;
@@ -1368,7 +1316,8 @@ err1:
 	return err;
 }
 
-static int vlv_runtime_resume(struct drm_i915_private *dev_priv)
+static int vlv_resume_prepare(struct drm_i915_private *dev_priv,
+				bool rpm_resume)
 {
 	struct drm_device *dev = dev_priv->dev;
 	int err;
@@ -1393,8 +1342,10 @@ static int vlv_runtime_resume(struct drm_i915_private *dev_priv)
 
 	vlv_check_no_gt_access(dev_priv);
 
-	intel_init_clock_gating(dev);
-	i915_gem_restore_fences(dev);
+	if (rpm_resume) {
+		intel_init_clock_gating(dev);
+		i915_gem_restore_fences(dev);
+	}
 
 	return ret;
 }
@@ -1409,7 +1360,9 @@ static int intel_runtime_suspend(struct device *device)
 	if (WARN_ON_ONCE(!(dev_priv->rps.enabled && intel_enable_rc6(dev))))
 		return -ENODEV;
 
-	WARN_ON(!HAS_RUNTIME_PM(dev));
+	if (WARN_ON_ONCE(!HAS_RUNTIME_PM(dev)))
+		return -ENODEV;
+
 	assert_force_wake_inactive(dev_priv);
 
 	DRM_DEBUG_KMS("Suspending device\n");
@@ -1438,28 +1391,13 @@ static int intel_runtime_suspend(struct device *device)
 	i915_gem_release_all_mmaps(dev_priv);
 	mutex_unlock(&dev->struct_mutex);
 
-	/*
-	 * rps.work can't be rearmed here, since we get here only after making
-	 * sure the GPU is idle and the RPS freq is set to the minimum. See
-	 * intel_mark_idle().
-	 */
-	cancel_work_sync(&dev_priv->rps.work);
-	intel_runtime_pm_disable_interrupts(dev);
-
-	if (IS_GEN6(dev)) {
-		ret = 0;
-	} else if (IS_HASWELL(dev) || IS_BROADWELL(dev)) {
-		ret = hsw_runtime_suspend(dev_priv);
-	} else if (IS_VALLEYVIEW(dev)) {
-		ret = vlv_runtime_suspend(dev_priv);
-	} else {
-		ret = -ENODEV;
-		WARN_ON(1);
-	}
+	intel_suspend_gt_powersave(dev);
+	intel_runtime_pm_disable_interrupts(dev_priv);
 
+	ret = intel_suspend_complete(dev_priv);
 	if (ret) {
 		DRM_ERROR("Runtime suspend failed, disabling it (%d)\n", ret);
-		intel_runtime_pm_restore_interrupts(dev);
+		intel_runtime_pm_enable_interrupts(dev_priv);
 
 		return ret;
 	}
@@ -1468,13 +1406,29 @@ static int intel_runtime_suspend(struct device *device)
 	dev_priv->pm.suspended = true;
 
 	/*
-	 * current versions of firmware which depend on this opregion
-	 * notification have repurposed the D1 definition to mean
-	 * "runtime suspended" vs. what you would normally expect (D3)
-	 * to distinguish it from notifications that might be sent
-	 * via the suspend path.
+	 * FIXME: We really should find a document that references the arguments
+	 * used below!
 	 */
-	intel_opregion_notify_adapter(dev, PCI_D1);
+	if (IS_HASWELL(dev)) {
+		/*
+		 * current versions of firmware which depend on this opregion
+		 * notification have repurposed the D1 definition to mean
+		 * "runtime suspended" vs. what you would normally expect (D3)
+		 * to distinguish it from notifications that might be sent via
+		 * the suspend path.
+		 */
+		intel_opregion_notify_adapter(dev, PCI_D1);
+	} else {
+		/*
+		 * On Broadwell, if we use PCI_D1 the PCH DDI ports will stop
+		 * being detected, and the call we do at intel_runtime_resume()
+		 * won't be able to restore them. Since PCI_D3hot matches the
+		 * actual specification and appears to be working, use it. Let's
+		 * assume the other non-Haswell platforms will stay the same as
+		 * Broadwell.
+		 */
+		intel_opregion_notify_adapter(dev, PCI_D3hot);
+	}
 
 	DRM_DEBUG_KMS("Device suspended\n");
 	return 0;
@@ -1485,25 +1439,22 @@ static int intel_runtime_resume(struct device *device)
 	struct pci_dev *pdev = to_pci_dev(device);
 	struct drm_device *dev = pci_get_drvdata(pdev);
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	int ret;
+	int ret = 0;
 
-	WARN_ON(!HAS_RUNTIME_PM(dev));
+	if (WARN_ON_ONCE(!HAS_RUNTIME_PM(dev)))
+		return -ENODEV;
 
 	DRM_DEBUG_KMS("Resuming device\n");
 
 	intel_opregion_notify_adapter(dev, PCI_D0);
 	dev_priv->pm.suspended = false;
 
-	if (IS_GEN6(dev)) {
-		ret = snb_runtime_resume(dev_priv);
-	} else if (IS_HASWELL(dev) || IS_BROADWELL(dev)) {
-		ret = hsw_runtime_resume(dev_priv);
-	} else if (IS_VALLEYVIEW(dev)) {
-		ret = vlv_runtime_resume(dev_priv);
-	} else {
-		WARN_ON(1);
-		ret = -ENODEV;
-	}
+	if (IS_GEN6(dev_priv))
+		intel_init_pch_refclk(dev);
+	else if (IS_HASWELL(dev_priv) || IS_BROADWELL(dev_priv))
+		hsw_disable_pc8(dev_priv);
+	else if (IS_VALLEYVIEW(dev_priv))
+		ret = vlv_resume_prepare(dev_priv, true);
 
 	/*
 	 * No point of rolling back things in case of an error, as the best
@@ -1512,8 +1463,8 @@ static int intel_runtime_resume(struct device *device)
 	i915_gem_init_swizzling(dev);
 	gen6_update_ring_freq(dev);
 
-	intel_runtime_pm_restore_interrupts(dev);
-	intel_reset_gt_powersave(dev);
+	intel_runtime_pm_enable_interrupts(dev_priv);
+	intel_enable_gt_powersave(dev);
 
 	if (ret)
 		DRM_ERROR("Runtime resume failed, disabling it (%d)\n", ret);
@@ -1523,17 +1474,60 @@ static int intel_runtime_resume(struct device *device)
 	return ret;
 }
 
+/*
+ * This function implements common functionality of runtime and system
+ * suspend sequence.
+ */
+static int intel_suspend_complete(struct drm_i915_private *dev_priv)
+{
+	struct drm_device *dev = dev_priv->dev;
+	int ret;
+
+	if (IS_HASWELL(dev) || IS_BROADWELL(dev))
+		ret = hsw_suspend_complete(dev_priv);
+	else if (IS_VALLEYVIEW(dev))
+		ret = vlv_suspend_complete(dev_priv);
+	else
+		ret = 0;
+
+	return ret;
+}
+
 static const struct dev_pm_ops i915_pm_ops = {
+	/*
+	 * S0ix (via system suspend) and S3 event handlers [PMSG_SUSPEND,
+	 * PMSG_RESUME]
+	 */
 	.suspend = i915_pm_suspend,
 	.suspend_late = i915_pm_suspend_late,
 	.resume_early = i915_pm_resume_early,
 	.resume = i915_pm_resume,
-	.freeze = i915_pm_freeze,
-	.thaw_early = i915_pm_thaw_early,
-	.thaw = i915_pm_thaw,
-	.poweroff = i915_pm_poweroff,
+
+	/*
+	 * S4 event handlers
+	 * @freeze, @freeze_late    : called (1) before creating the
+	 *                            hibernation image [PMSG_FREEZE] and
+	 *                            (2) after rebooting, before restoring
+	 *                            the image [PMSG_QUIESCE]
+	 * @thaw, @thaw_early       : called (1) after creating the hibernation
+	 *                            image, before writing it [PMSG_THAW]
+	 *                            and (2) after failing to create or
+	 *                            restore the image [PMSG_RECOVER]
+	 * @poweroff, @poweroff_late: called after writing the hibernation
+	 *                            image, before rebooting [PMSG_HIBERNATE]
+	 * @restore, @restore_early : called after rebooting and restoring the
+	 *                            hibernation image [PMSG_RESTORE]
+	 */
+	.freeze = i915_pm_suspend,
+	.freeze_late = i915_pm_suspend_late,
+	.thaw_early = i915_pm_resume_early,
+	.thaw = i915_pm_resume,
+	.poweroff = i915_pm_suspend,
+	.poweroff_late = i915_pm_suspend_late,
 	.restore_early = i915_pm_resume_early,
 	.restore = i915_pm_resume,
+
+	/* S0ix (via runtime suspend) event handlers */
 	.runtime_suspend = intel_runtime_suspend,
 	.runtime_resume = intel_runtime_resume,
 };
@@ -1572,14 +1566,13 @@ static struct drm_driver driver = {
 	.lastclose = i915_driver_lastclose,
 	.preclose = i915_driver_preclose,
 	.postclose = i915_driver_postclose,
+	.set_busid = drm_pci_set_busid,
 
 	/* Used in place of i915_pm_ops for non-DRIVER_MODESET */
-	.suspend = i915_suspend,
+	.suspend = i915_suspend_legacy,
 	.resume = i915_resume_legacy,
 
 	.device_is_agp = i915_driver_device_is_agp,
-	.master_create = i915_master_create,
-	.master_destroy = i915_master_destroy,
 #if defined(CONFIG_DEBUG_FS)
 	.debugfs_init = i915_debugfs_init,
 	.debugfs_cleanup = i915_debugfs_cleanup,
@@ -1663,6 +1656,8 @@ static void __exit i915_exit(void)
 module_init(i915_init);
 module_exit(i915_exit);
 
-MODULE_AUTHOR(DRIVER_AUTHOR);
+MODULE_AUTHOR("Tungsten Graphics, Inc.");
+MODULE_AUTHOR("Intel Corporation");
+
 MODULE_DESCRIPTION(DRIVER_DESC);
 MODULE_LICENSE("GPL and additional rights");
diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h
index 3524306d8cfb..e9f891c432f8 100644
--- a/drivers/gpu/drm/i915/i915_drv.h
+++ b/drivers/gpu/drm/i915/i915_drv.h
@@ -35,11 +35,15 @@
 #include "i915_reg.h"
 #include "intel_bios.h"
 #include "intel_ringbuffer.h"
+#include "intel_lrc.h"
 #include "i915_gem_gtt.h"
+#include "i915_gem_render_state.h"
 #include <linux/io-mapping.h>
 #include <linux/i2c.h>
 #include <linux/i2c-algo-bit.h>
 #include <drm/intel-gtt.h>
+#include <drm/drm_legacy.h> /* for struct drm_dma_handle */
+#include <drm/drm_gem.h>
 #include <linux/backlight.h>
 #include <linux/hashtable.h>
 #include <linux/intel-iommu.h>
@@ -49,11 +53,12 @@
 /* General customization:
  */
 
-#define DRIVER_AUTHOR		"Tungsten Graphics, Inc."
-
 #define DRIVER_NAME		"i915"
 #define DRIVER_DESC		"Intel Graphics"
-#define DRIVER_DATE		"20140725"
+#define DRIVER_DATE		"20141121"
+
+#undef WARN_ON
+#define WARN_ON(x)		WARN(x, "WARN_ON(" #x ")")
 
 enum pipe {
 	INVALID_PIPE = -1,
@@ -74,6 +79,14 @@ enum transcoder {
 };
 #define transcoder_name(t) ((t) + 'A')
 
+/*
+ * This is the maximum (across all platforms) number of planes (primary +
+ * sprites) that can be active at the same time on one pipe.
+ *
+ * This value doesn't count the cursor plane.
+ */
+#define I915_MAX_PLANES	3
+
 enum plane {
 	PLANE_A = 0,
 	PLANE_B,
@@ -162,7 +175,10 @@ enum hpd_pin {
 	 I915_GEM_DOMAIN_INSTRUCTION | \
 	 I915_GEM_DOMAIN_VERTEX)
 
-#define for_each_pipe(p) for ((p) = 0; (p) < INTEL_INFO(dev)->num_pipes; (p)++)
+#define for_each_pipe(__dev_priv, __p) \
+	for ((__p) = 0; (__p) < INTEL_INFO(__dev_priv)->num_pipes; (__p)++)
+#define for_each_plane(pipe, p) \
+	for ((p) = 0; (p) < INTEL_INFO(dev)->num_sprites[(pipe)] + 1; (p)++)
 #define for_each_sprite(p, s) for ((s) = 0; (s) < INTEL_INFO(dev)->num_sprites[(p)]; (s)++)
 
 #define for_each_crtc(dev, crtc) \
@@ -171,6 +187,11 @@ enum hpd_pin {
 #define for_each_intel_crtc(dev, intel_crtc) \
 	list_for_each_entry(intel_crtc, &dev->mode_config.crtc_list, base.head)
 
+#define for_each_intel_encoder(dev, intel_encoder)		\
+	list_for_each_entry(intel_encoder,			\
+			    &(dev)->mode_config.encoder_list,	\
+			    base.head)
+
 #define for_each_encoder_on_crtc(dev, __crtc, intel_encoder) \
 	list_for_each_entry((intel_encoder), &(dev)->mode_config.encoder_list, base.head) \
 		if ((intel_encoder)->base.crtc == (__crtc))
@@ -192,27 +213,52 @@ enum intel_dpll_id {
 	/* real shared dpll ids must be >= 0 */
 	DPLL_ID_PCH_PLL_A = 0,
 	DPLL_ID_PCH_PLL_B = 1,
+	/* hsw/bdw */
 	DPLL_ID_WRPLL1 = 0,
 	DPLL_ID_WRPLL2 = 1,
+	/* skl */
+	DPLL_ID_SKL_DPLL1 = 0,
+	DPLL_ID_SKL_DPLL2 = 1,
+	DPLL_ID_SKL_DPLL3 = 2,
 };
-#define I915_NUM_PLLS 2
+#define I915_NUM_PLLS 3
 
 struct intel_dpll_hw_state {
+	/* i9xx, pch plls */
 	uint32_t dpll;
 	uint32_t dpll_md;
 	uint32_t fp0;
 	uint32_t fp1;
+
+	/* hsw, bdw */
 	uint32_t wrpll;
+
+	/* skl */
+	/*
+	 * DPLL_CTRL1 has 6 bits for each each this DPLL. We store those in
+	 * lower part of crtl1 and they get shifted into position when writing
+	 * the register.  This allows us to easily compare the state to share
+	 * the DPLL.
+	 */
+	uint32_t ctrl1;
+	/* HDMI only, 0 when used for DP */
+	uint32_t cfgcr1, cfgcr2;
+};
+
+struct intel_shared_dpll_config {
+	unsigned crtc_mask; /* mask of CRTCs sharing this PLL */
+	struct intel_dpll_hw_state hw_state;
 };
 
 struct intel_shared_dpll {
-	int refcount; /* count of number of CRTCs sharing this PLL */
+	struct intel_shared_dpll_config config;
+	struct intel_shared_dpll_config *new_config;
+
 	int active; /* count of number of active CRTCs (i.e. DPMS on) */
 	bool on; /* is the PLL actually active? Disabled during modeset */
 	const char *name;
 	/* should match the index in the dev_priv->shared_dplls array */
 	enum intel_dpll_id id;
-	struct intel_dpll_hw_state hw_state;
 	/* The mode_set hook is optional and should be used together with the
 	 * intel_prepare_shared_dpll function. */
 	void (*mode_set)(struct drm_i915_private *dev_priv,
@@ -226,6 +272,11 @@ struct intel_shared_dpll {
 			     struct intel_dpll_hw_state *hw_state);
 };
 
+#define SKL_DPLL0 0
+#define SKL_DPLL1 1
+#define SKL_DPLL2 2
+#define SKL_DPLL3 3
+
 /* Used by dp and fdi links */
 struct intel_link_m_n {
 	uint32_t	tu;
@@ -254,7 +305,6 @@ void intel_link_compute_m_n(int bpp, int nlanes,
 #define DRIVER_PATCHLEVEL	0
 
 #define WATCH_LISTS	0
-#define WATCH_GTT	0
 
 struct opregion_header;
 struct opregion_acpi;
@@ -277,10 +327,6 @@ struct intel_opregion {
 struct intel_overlay;
 struct intel_overlay_error_state;
 
-struct drm_i915_master_private {
-	drm_local_map_t *sarea;
-	struct _drm_i915_sarea *sarea_priv;
-};
 #define I915_FENCE_REG_NONE -1
 #define I915_MAX_NUM_FENCES 32
 /* 32 fences + sign bit for FENCE_REG_NONE */
@@ -388,6 +434,7 @@ struct drm_i915_error_state {
 		pid_t pid;
 		char comm[TASK_COMM_LEN];
 	} ring[I915_NUM_RINGS];
+
 	struct drm_i915_error_buffer {
 		u32 size;
 		u32 name;
@@ -406,9 +453,11 @@ struct drm_i915_error_state {
 	} **active_bo, **pinned_bo;
 
 	u32 *active_bo_count, *pinned_bo_count;
+	u32 vm_count;
 };
 
 struct intel_connector;
+struct intel_encoder;
 struct intel_crtc_config;
 struct intel_plane_config;
 struct intel_crtc;
@@ -435,7 +484,7 @@ struct drm_i915_display_funcs {
 	 * Returns true on success, false on failure.
 	 */
 	bool (*find_dpll)(const struct intel_limit *limit,
-			  struct drm_crtc *crtc,
+			  struct intel_crtc *crtc,
 			  int target, int refclk,
 			  struct dpll *match_clock,
 			  struct dpll *best_clock);
@@ -451,15 +500,14 @@ struct drm_i915_display_funcs {
 				struct intel_crtc_config *);
 	void (*get_plane_config)(struct intel_crtc *,
 				 struct intel_plane_config *);
-	int (*crtc_mode_set)(struct drm_crtc *crtc,
-			     int x, int y,
-			     struct drm_framebuffer *old_fb);
+	int (*crtc_compute_clock)(struct intel_crtc *crtc);
 	void (*crtc_enable)(struct drm_crtc *crtc);
 	void (*crtc_disable)(struct drm_crtc *crtc);
 	void (*off)(struct drm_crtc *crtc);
-	void (*write_eld)(struct drm_connector *connector,
-			  struct drm_crtc *crtc,
-			  struct drm_display_mode *mode);
+	void (*audio_codec_enable)(struct drm_connector *connector,
+				   struct intel_encoder *encoder,
+				   struct drm_display_mode *mode);
+	void (*audio_codec_disable)(struct intel_encoder *encoder);
 	void (*fdi_link_train)(struct drm_crtc *crtc);
 	void (*init_clock_gating)(struct drm_device *dev);
 	int (*queue_flip)(struct drm_device *dev, struct drm_crtc *crtc,
@@ -477,7 +525,7 @@ struct drm_i915_display_funcs {
 	/* display clock increase/decrease */
 	/* pll clock increase/decrease */
 
-	int (*setup_backlight)(struct intel_connector *connector);
+	int (*setup_backlight)(struct intel_connector *connector, enum pipe pipe);
 	uint32_t (*get_backlight)(struct intel_connector *connector);
 	void (*set_backlight)(struct intel_connector *connector,
 			      uint32_t level);
@@ -516,6 +564,7 @@ struct intel_uncore {
 
 	unsigned fw_rendercount;
 	unsigned fw_mediacount;
+	unsigned fw_blittercount;
 
 	struct timer_list force_wake_timer;
 };
@@ -534,6 +583,7 @@ struct intel_uncore {
 	func(is_ivybridge) sep \
 	func(is_valleyview) sep \
 	func(is_haswell) sep \
+	func(is_skylake) sep \
 	func(is_preliminary) sep \
 	func(has_fbc) sep \
 	func(has_pipe_cxsr) sep \
@@ -551,6 +601,7 @@ struct intel_uncore {
 
 struct intel_device_info {
 	u32 display_mmio_offset;
+	u16 device_id;
 	u8 num_pipes:3;
 	u8 num_sprites[I915_MAX_PIPES];
 	u8 gen;
@@ -615,13 +666,22 @@ struct intel_context {
 	uint8_t remap_slice;
 	struct drm_i915_file_private *file_priv;
 	struct i915_ctx_hang_stats hang_stats;
-	struct i915_address_space *vm;
+	struct i915_hw_ppgtt *ppgtt;
 
+	/* Legacy ring buffer submission */
 	struct {
 		struct drm_i915_gem_object *rcs_state;
 		bool initialized;
 	} legacy_hw_ctx;
 
+	/* Execlists */
+	bool rcs_initialized;
+	struct {
+		struct drm_i915_gem_object *state;
+		struct intel_ringbuffer *ringbuf;
+		int unpin_count;
+	} engine[I915_NUM_RINGS];
+
 	struct list_head link;
 };
 
@@ -635,6 +695,20 @@ struct i915_fbc {
 	struct drm_mm_node compressed_fb;
 	struct drm_mm_node *compressed_llb;
 
+	bool false_color;
+
+	/* Tracks whether the HW is actually enabled, not whether the feature is
+	 * possible. */
+	bool enabled;
+
+	/* On gen8 some rings cannont perform fbc clean operation so for now
+	 * we are doing this on SW with mmio.
+	 * This variable works in the opposite information direction
+	 * of ring->fbc_dirty telling software on frontbuffer tracking
+	 * to perform the cache clean on sw side.
+	 */
+	bool need_sw_cache_clean;
+
 	struct intel_fbc_work {
 		struct delayed_work work;
 		struct drm_crtc *crtc;
@@ -676,6 +750,7 @@ enum intel_pch {
 	PCH_IBX,	/* Ibexpeak PCH */
 	PCH_CPT,	/* Cougarpoint PCH */
 	PCH_LPT,	/* Lynxpoint PCH */
+	PCH_SPT,        /* Sunrisepoint PCH */
 	PCH_NOP,
 };
 
@@ -688,6 +763,8 @@ enum intel_sbi_destination {
 #define QUIRK_LVDS_SSC_DISABLE (1<<1)
 #define QUIRK_INVERT_BRIGHTNESS (1<<2)
 #define QUIRK_BACKLIGHT_PRESENT (1<<3)
+#define QUIRK_PIPEB_FORCE (1<<4)
+#define QUIRK_PIN_SWIZZLED_PAGES (1<<5)
 
 struct intel_fbdev;
 struct intel_fbc_work;
@@ -739,7 +816,6 @@ struct i915_suspend_saved_registers {
 	u32 saveBLC_HIST_CTL;
 	u32 saveBLC_PWM_CTL;
 	u32 saveBLC_PWM_CTL2;
-	u32 saveBLC_HIST_CTL_B;
 	u32 saveBLC_CPU_PWM_CTL;
 	u32 saveBLC_CPU_PWM_CTL2;
 	u32 saveFPB0;
@@ -848,6 +924,7 @@ struct i915_suspend_saved_registers {
 	u32 savePIPEB_LINK_N1;
 	u32 saveMCHBAR_RENDER_STANDBY;
 	u32 savePCH_PORT_HOTPLUG;
+	u16 saveGCDGMBUS;
 };
 
 struct vlv_s0ix_state {
@@ -918,8 +995,12 @@ struct intel_rps_ei {
 };
 
 struct intel_gen6_power_mgmt {
-	/* work and pm_iir are protected by dev_priv->irq_lock */
+	/*
+	 * work, interrupts_enabled and pm_iir are protected by
+	 * dev_priv->irq_lock
+	 */
 	struct work_struct work;
+	bool interrupts_enabled;
 	u32 pm_iir;
 
 	/* Frequencies are stored in potentially platform dependent multiples.
@@ -1042,31 +1123,6 @@ struct i915_power_domains {
 	struct i915_power_well *power_wells;
 };
 
-struct i915_dri1_state {
-	unsigned allow_batchbuffer : 1;
-	u32 __iomem *gfx_hws_cpu_addr;
-
-	unsigned int cpp;
-	int back_offset;
-	int front_offset;
-	int current_page;
-	int page_flipping;
-
-	uint32_t counter;
-};
-
-struct i915_ums_state {
-	/**
-	 * Flag if the X Server, and thus DRM, is not currently in
-	 * control of the device.
-	 *
-	 * This is set between LeaveVT and EnterVT.  It needs to be
-	 * replaced with a semaphore.  It also needs to be
-	 * transitioned away from for kernel modesetting.
-	 */
-	int mm_suspended;
-};
-
 #define MAX_L3_SLICES 2
 struct intel_l3_parity {
 	u32 *remap_info[MAX_L3_SLICES];
@@ -1147,6 +1203,7 @@ struct i915_gem_mm {
 };
 
 struct drm_i915_error_state_buf {
+	struct drm_i915_private *i915;
 	unsigned bytes;
 	unsigned size;
 	int err;
@@ -1219,6 +1276,9 @@ struct i915_gpu_error {
 
 	/* For missed irq/seqno simulation. */
 	unsigned int test_irq_rings;
+
+	/* Used to prevent gem_check_wedged returning -EAGAIN during gpu reset   */
+	bool reload_in_reset;
 };
 
 enum modeset_restore {
@@ -1228,6 +1288,12 @@ enum modeset_restore {
 };
 
 struct ddi_vbt_port_info {
+	/*
+	 * This is an index in the HDMI/DVI DDI buffer translation table.
+	 * The special value HDMI_LEVEL_SHIFT_UNKNOWN means the VBT didn't
+	 * populate this field.
+	 */
+#define HDMI_LEVEL_SHIFT_UNKNOWN	0xff
 	uint8_t hdmi_level_shift;
 
 	uint8_t supports_dvi:1;
@@ -1318,6 +1384,49 @@ struct ilk_wm_values {
 	enum intel_ddb_partitioning partitioning;
 };
 
+struct skl_ddb_entry {
+	uint16_t start, end;	/* in number of blocks, 'end' is exclusive */
+};
+
+static inline uint16_t skl_ddb_entry_size(const struct skl_ddb_entry *entry)
+{
+	return entry->end - entry->start;
+}
+
+static inline bool skl_ddb_entry_equal(const struct skl_ddb_entry *e1,
+				       const struct skl_ddb_entry *e2)
+{
+	if (e1->start == e2->start && e1->end == e2->end)
+		return true;
+
+	return false;
+}
+
+struct skl_ddb_allocation {
+	struct skl_ddb_entry pipe[I915_MAX_PIPES];
+	struct skl_ddb_entry plane[I915_MAX_PIPES][I915_MAX_PLANES];
+	struct skl_ddb_entry cursor[I915_MAX_PIPES];
+};
+
+struct skl_wm_values {
+	bool dirty[I915_MAX_PIPES];
+	struct skl_ddb_allocation ddb;
+	uint32_t wm_linetime[I915_MAX_PIPES];
+	uint32_t plane[I915_MAX_PIPES][I915_MAX_PLANES][8];
+	uint32_t cursor[I915_MAX_PIPES][8];
+	uint32_t plane_trans[I915_MAX_PIPES][I915_MAX_PLANES];
+	uint32_t cursor_trans[I915_MAX_PIPES];
+};
+
+struct skl_wm_level {
+	bool plane_en[I915_MAX_PLANES];
+	bool cursor_en;
+	uint16_t plane_res_b[I915_MAX_PLANES];
+	uint8_t plane_res_l[I915_MAX_PLANES];
+	uint16_t cursor_res_b;
+	uint8_t cursor_res_l;
+};
+
 /*
  * This struct helps tracking the state needed for runtime PM, which puts the
  * device in PCI D3 state. Notice that when this happens, nothing on the
@@ -1330,7 +1439,7 @@ struct ilk_wm_values {
  *
  * Our driver uses the autosuspend delay feature, which means we'll only really
  * suspend if we stay with zero refcount for a certain amount of time. The
- * default value is currently very conservative (see intel_init_runtime_pm), but
+ * default value is currently very conservative (see intel_runtime_pm_enable), but
  * it can be changed with the standard runtime PM files from sysfs.
  *
  * The irqs_disabled variable becomes true exactly after we disable the IRQs and
@@ -1343,7 +1452,7 @@ struct ilk_wm_values {
  */
 struct i915_runtime_pm {
 	bool suspended;
-	bool _irqs_disabled;
+	bool irqs_enabled;
 };
 
 enum intel_pipe_crc_source {
@@ -1387,6 +1496,20 @@ struct i915_frontbuffer_tracking {
 	unsigned flip_bits;
 };
 
+struct i915_wa_reg {
+	u32 addr;
+	u32 value;
+	/* bitmask representing WA bits */
+	u32 mask;
+};
+
+#define I915_MAX_WA_REGS 16
+
+struct i915_workarounds {
+	struct i915_wa_reg reg[I915_MAX_WA_REGS];
+	u32 count;
+};
+
 struct drm_i915_private {
 	struct drm_device *dev;
 	struct kmem_cache *slab;
@@ -1421,7 +1544,7 @@ struct drm_i915_private {
 	struct drm_i915_gem_object *semaphore_obj;
 	uint32_t last_seqno, next_seqno;
 
-	drm_dma_handle_t *status_page_dmah;
+	struct drm_dma_handle *status_page_dmah;
 	struct resource mch_res;
 
 	/* protects the irq masks */
@@ -1466,21 +1589,27 @@ struct drm_i915_private {
 	struct intel_opregion opregion;
 	struct intel_vbt_data vbt;
 
+	bool preserve_bios_swizzle;
+
 	/* overlay */
 	struct intel_overlay *overlay;
 
 	/* backlight registers and fields in struct intel_panel */
-	spinlock_t backlight_lock;
+	struct mutex backlight_lock;
 
 	/* LVDS info */
 	bool no_aux_handshake;
 
+	/* protects panel power sequencer state */
+	struct mutex pps_mutex;
+
 	struct drm_i915_fence_reg fence_regs[I915_MAX_NUM_FENCES]; /* assume 965 */
 	int fence_reg_start; /* 4 if userland hasn't ioctl'd us yet */
 	int num_fence_regs; /* 8 on pre-965, 16 otherwise */
 
 	unsigned int fsb_freq, mem_freq, is_ddr3;
 	unsigned int vlv_cdclk_freq;
+	unsigned int hpll_freq;
 
 	/**
 	 * wq - Driver workqueue for GEM.
@@ -1526,6 +1655,8 @@ struct drm_i915_private {
 	struct intel_shared_dpll shared_dplls[I915_NUM_PLLS];
 	int dpio_phy_iosf_port[I915_NUM_PHYS_VLV];
 
+	struct i915_workarounds workarounds;
+
 	/* Reclocking support */
 	bool render_reclock_avail;
 	bool lvds_downclock_avail;
@@ -1561,14 +1692,9 @@ struct drm_i915_private {
 #ifdef CONFIG_DRM_I915_FBDEV
 	/* list of fbdev register on this device */
 	struct intel_fbdev *fbdev;
+	struct work_struct fbdev_suspend_work;
 #endif
 
-	/*
-	 * The console may be contended at resume, but we don't
-	 * want it to block on it.
-	 */
-	struct work_struct console_resume_work;
-
 	struct drm_property *broadcast_rgb_property;
 	struct drm_property *force_audio_property;
 
@@ -1593,9 +1719,25 @@ struct drm_i915_private {
 		uint16_t spr_latency[5];
 		/* cursor */
 		uint16_t cur_latency[5];
+		/*
+		 * Raw watermark memory latency values
+		 * for SKL for all 8 levels
+		 * in 1us units.
+		 */
+		uint16_t skl_latency[8];
+
+		/*
+		 * The skl_wm_values structure is a bit too big for stack
+		 * allocation, so we keep the staging struct where we store
+		 * intermediate results here instead.
+		 */
+		struct skl_wm_values skl_results;
 
 		/* current hardware state */
-		struct ilk_wm_values hw;
+		union {
+			struct ilk_wm_values hw;
+			struct skl_wm_values skl_hw;
+		};
 	} wm;
 
 	struct i915_runtime_pm pm;
@@ -1614,11 +1756,19 @@ struct drm_i915_private {
 	 */
 	struct workqueue_struct *dp_wq;
 
-	/* Old dri1 support infrastructure, beware the dragons ya fools entering
-	 * here! */
-	struct i915_dri1_state dri1;
-	/* Old ums support infrastructure, same warning applies. */
-	struct i915_ums_state ums;
+	/* Abstract the submission mechanism (legacy ringbuffer or execlists) away */
+	struct {
+		int (*do_execbuf)(struct drm_device *dev, struct drm_file *file,
+				  struct intel_engine_cs *ring,
+				  struct intel_context *ctx,
+				  struct drm_i915_gem_execbuffer2 *args,
+				  struct list_head *vmas,
+				  struct drm_i915_gem_object *batch_obj,
+				  u64 exec_start, u32 flags);
+		int (*init_rings)(struct drm_device *dev);
+		void (*cleanup_ring)(struct intel_engine_cs *ring);
+		void (*stop_ring)(struct intel_engine_cs *ring);
+	} gt;
 
 	/*
 	 * NOTE: This is the dri1/ums dungeon, don't add stuff here. Your patch
@@ -1761,17 +1911,8 @@ struct drm_i915_gem_object {
 	 * Only honoured if hardware has relevant pte bit
 	 */
 	unsigned long gt_ro:1;
-
-	/*
-	 * Is the GPU currently using a fence to access this buffer,
-	 */
-	unsigned int pending_fenced_gpu_access:1;
-	unsigned int fenced_gpu_access:1;
-
 	unsigned int cache_level:3;
 
-	unsigned int has_aliasing_ppgtt_mapping:1;
-	unsigned int has_global_gtt_mapping:1;
 	unsigned int has_dma_mapping:1;
 
 	unsigned int frontbuffer_bits:INTEL_FRONTBUFFER_BITS;
@@ -1804,10 +1945,10 @@ struct drm_i915_gem_object {
 	unsigned long user_pin_count;
 	struct drm_file *pin_filp;
 
-	/** for phy allocated objects */
-	drm_dma_handle_t *phys_handle;
-
 	union {
+		/** for phy allocated objects */
+		struct drm_dma_handle *phys_handle;
+
 		struct i915_gem_userptr {
 			uintptr_t ptr;
 			unsigned read_only :1;
@@ -1971,51 +2112,65 @@ struct drm_i915_cmd_table {
 	int count;
 };
 
-#define INTEL_INFO(dev)	(&to_i915(dev)->info)
-
-#define IS_I830(dev)		((dev)->pdev->device == 0x3577)
-#define IS_845G(dev)		((dev)->pdev->device == 0x2562)
+/* Note that the (struct drm_i915_private *) cast is just to shut up gcc. */
+#define __I915__(p) ({ \
+	struct drm_i915_private *__p; \
+	if (__builtin_types_compatible_p(typeof(*p), struct drm_i915_private)) \
+		__p = (struct drm_i915_private *)p; \
+	else if (__builtin_types_compatible_p(typeof(*p), struct drm_device)) \
+		__p = to_i915((struct drm_device *)p); \
+	else \
+		BUILD_BUG(); \
+	__p; \
+})
+#define INTEL_INFO(p) 	(&__I915__(p)->info)
+#define INTEL_DEVID(p)	(INTEL_INFO(p)->device_id)
+
+#define IS_I830(dev)		(INTEL_DEVID(dev) == 0x3577)
+#define IS_845G(dev)		(INTEL_DEVID(dev) == 0x2562)
 #define IS_I85X(dev)		(INTEL_INFO(dev)->is_i85x)
-#define IS_I865G(dev)		((dev)->pdev->device == 0x2572)
+#define IS_I865G(dev)		(INTEL_DEVID(dev) == 0x2572)
 #define IS_I915G(dev)		(INTEL_INFO(dev)->is_i915g)
-#define IS_I915GM(dev)		((dev)->pdev->device == 0x2592)
-#define IS_I945G(dev)		((dev)->pdev->device == 0x2772)
+#define IS_I915GM(dev)		(INTEL_DEVID(dev) == 0x2592)
+#define IS_I945G(dev)		(INTEL_DEVID(dev) == 0x2772)
 #define IS_I945GM(dev)		(INTEL_INFO(dev)->is_i945gm)
 #define IS_BROADWATER(dev)	(INTEL_INFO(dev)->is_broadwater)
 #define IS_CRESTLINE(dev)	(INTEL_INFO(dev)->is_crestline)
-#define IS_GM45(dev)		((dev)->pdev->device == 0x2A42)
+#define IS_GM45(dev)		(INTEL_DEVID(dev) == 0x2A42)
 #define IS_G4X(dev)		(INTEL_INFO(dev)->is_g4x)
-#define IS_PINEVIEW_G(dev)	((dev)->pdev->device == 0xa001)
-#define IS_PINEVIEW_M(dev)	((dev)->pdev->device == 0xa011)
+#define IS_PINEVIEW_G(dev)	(INTEL_DEVID(dev) == 0xa001)
+#define IS_PINEVIEW_M(dev)	(INTEL_DEVID(dev) == 0xa011)
 #define IS_PINEVIEW(dev)	(INTEL_INFO(dev)->is_pineview)
 #define IS_G33(dev)		(INTEL_INFO(dev)->is_g33)
-#define IS_IRONLAKE_M(dev)	((dev)->pdev->device == 0x0046)
+#define IS_IRONLAKE_M(dev)	(INTEL_DEVID(dev) == 0x0046)
 #define IS_IVYBRIDGE(dev)	(INTEL_INFO(dev)->is_ivybridge)
-#define IS_IVB_GT1(dev)		((dev)->pdev->device == 0x0156 || \
-				 (dev)->pdev->device == 0x0152 || \
-				 (dev)->pdev->device == 0x015a)
-#define IS_SNB_GT1(dev)		((dev)->pdev->device == 0x0102 || \
-				 (dev)->pdev->device == 0x0106 || \
-				 (dev)->pdev->device == 0x010A)
+#define IS_IVB_GT1(dev)		(INTEL_DEVID(dev) == 0x0156 || \
+				 INTEL_DEVID(dev) == 0x0152 || \
+				 INTEL_DEVID(dev) == 0x015a)
+#define IS_SNB_GT1(dev)		(INTEL_DEVID(dev) == 0x0102 || \
+				 INTEL_DEVID(dev) == 0x0106 || \
+				 INTEL_DEVID(dev) == 0x010A)
 #define IS_VALLEYVIEW(dev)	(INTEL_INFO(dev)->is_valleyview)
 #define IS_CHERRYVIEW(dev)	(INTEL_INFO(dev)->is_valleyview && IS_GEN8(dev))
 #define IS_HASWELL(dev)	(INTEL_INFO(dev)->is_haswell)
 #define IS_BROADWELL(dev)	(!INTEL_INFO(dev)->is_valleyview && IS_GEN8(dev))
+#define IS_SKYLAKE(dev)	(INTEL_INFO(dev)->is_skylake)
 #define IS_MOBILE(dev)		(INTEL_INFO(dev)->is_mobile)
 #define IS_HSW_EARLY_SDV(dev)	(IS_HASWELL(dev) && \
-				 ((dev)->pdev->device & 0xFF00) == 0x0C00)
+				 (INTEL_DEVID(dev) & 0xFF00) == 0x0C00)
 #define IS_BDW_ULT(dev)		(IS_BROADWELL(dev) && \
-				 (((dev)->pdev->device & 0xf) == 0x2  || \
-				 ((dev)->pdev->device & 0xf) == 0x6 || \
-				 ((dev)->pdev->device & 0xf) == 0xe))
+				 ((INTEL_DEVID(dev) & 0xf) == 0x2  || \
+				 (INTEL_DEVID(dev) & 0xf) == 0x6 || \
+				 (INTEL_DEVID(dev) & 0xf) == 0xe))
+#define IS_BDW_GT3(dev)		(IS_BROADWELL(dev) && \
+				 (INTEL_DEVID(dev) & 0x00F0) == 0x0020)
 #define IS_HSW_ULT(dev)		(IS_HASWELL(dev) && \
-				 ((dev)->pdev->device & 0xFF00) == 0x0A00)
-#define IS_ULT(dev)		(IS_HSW_ULT(dev) || IS_BDW_ULT(dev))
+				 (INTEL_DEVID(dev) & 0xFF00) == 0x0A00)
 #define IS_HSW_GT3(dev)		(IS_HASWELL(dev) && \
-				 ((dev)->pdev->device & 0x00F0) == 0x0020)
+				 (INTEL_DEVID(dev) & 0x00F0) == 0x0020)
 /* ULX machines are also considered ULT. */
-#define IS_HSW_ULX(dev)		((dev)->pdev->device == 0x0A0E || \
-				 (dev)->pdev->device == 0x0A1E)
+#define IS_HSW_ULX(dev)		(INTEL_DEVID(dev) == 0x0A0E || \
+				 INTEL_DEVID(dev) == 0x0A1E)
 #define IS_PRELIMINARY_HW(intel_info) ((intel_info)->is_preliminary)
 
 /*
@@ -2031,6 +2186,7 @@ struct drm_i915_cmd_table {
 #define IS_GEN6(dev)	(INTEL_INFO(dev)->gen == 6)
 #define IS_GEN7(dev)	(INTEL_INFO(dev)->gen == 7)
 #define IS_GEN8(dev)	(INTEL_INFO(dev)->gen == 8)
+#define IS_GEN9(dev)	(INTEL_INFO(dev)->gen == 9)
 
 #define RENDER_RING		(1<<RCS)
 #define BSD_RING		(1<<VCS)
@@ -2043,14 +2199,13 @@ struct drm_i915_cmd_table {
 #define HAS_VEBOX(dev)		(INTEL_INFO(dev)->ring_mask & VEBOX_RING)
 #define HAS_LLC(dev)		(INTEL_INFO(dev)->has_llc)
 #define HAS_WT(dev)		((IS_HASWELL(dev) || IS_BROADWELL(dev)) && \
-				 to_i915(dev)->ellc_size)
+				 __I915__(dev)->ellc_size)
 #define I915_NEED_GFX_HWS(dev)	(INTEL_INFO(dev)->need_gfx_hws)
 
 #define HAS_HW_CONTEXTS(dev)	(INTEL_INFO(dev)->gen >= 6)
-#define HAS_ALIASING_PPGTT(dev)	(INTEL_INFO(dev)->gen >= 6)
-#define HAS_PPGTT(dev)		(INTEL_INFO(dev)->gen >= 7 && !IS_GEN8(dev))
-#define USES_PPGTT(dev)		intel_enable_ppgtt(dev, false)
-#define USES_FULL_PPGTT(dev)	intel_enable_ppgtt(dev, true)
+#define HAS_LOGICAL_RING_CONTEXTS(dev)	(INTEL_INFO(dev)->gen >= 8)
+#define USES_PPGTT(dev)		(i915.enable_ppgtt)
+#define USES_FULL_PPGTT(dev)	(i915.enable_ppgtt == 2)
 
 #define HAS_OVERLAY(dev)		(INTEL_INFO(dev)->has_overlay)
 #define OVERLAY_NEEDS_PHYSICAL(dev)	(INTEL_INFO(dev)->overlay_needs_physical)
@@ -2081,13 +2236,15 @@ struct drm_i915_cmd_table {
 #define HAS_PIPE_CXSR(dev) (INTEL_INFO(dev)->has_pipe_cxsr)
 #define HAS_FBC(dev) (INTEL_INFO(dev)->has_fbc)
 
-#define HAS_IPS(dev)		(IS_ULT(dev) || IS_BROADWELL(dev))
+#define HAS_IPS(dev)		(IS_HSW_ULT(dev) || IS_BROADWELL(dev))
 
 #define HAS_DDI(dev)		(INTEL_INFO(dev)->has_ddi)
 #define HAS_FPGA_DBG_UNCLAIMED(dev)	(INTEL_INFO(dev)->has_fpga_dbg)
 #define HAS_PSR(dev)		(IS_HASWELL(dev) || IS_BROADWELL(dev))
 #define HAS_RUNTIME_PM(dev)	(IS_GEN6(dev) || IS_HASWELL(dev) || \
 				 IS_BROADWELL(dev) || IS_VALLEYVIEW(dev))
+#define HAS_RC6(dev)		(INTEL_INFO(dev)->gen >= 6)
+#define HAS_RC6p(dev)		(INTEL_INFO(dev)->gen == 6 || IS_IVYBRIDGE(dev))
 
 #define INTEL_PCH_DEVICE_ID_MASK		0xff00
 #define INTEL_PCH_IBX_DEVICE_ID_TYPE		0x3b00
@@ -2095,8 +2252,11 @@ struct drm_i915_cmd_table {
 #define INTEL_PCH_PPT_DEVICE_ID_TYPE		0x1e00
 #define INTEL_PCH_LPT_DEVICE_ID_TYPE		0x8c00
 #define INTEL_PCH_LPT_LP_DEVICE_ID_TYPE		0x9c00
+#define INTEL_PCH_SPT_DEVICE_ID_TYPE		0xA100
+#define INTEL_PCH_SPT_LP_DEVICE_ID_TYPE		0x9D00
 
-#define INTEL_PCH_TYPE(dev) (to_i915(dev)->pch_type)
+#define INTEL_PCH_TYPE(dev) (__I915__(dev)->pch_type)
+#define HAS_PCH_SPT(dev) (INTEL_PCH_TYPE(dev) == PCH_SPT)
 #define HAS_PCH_LPT(dev) (INTEL_PCH_TYPE(dev) == PCH_LPT)
 #define HAS_PCH_CPT(dev) (INTEL_PCH_TYPE(dev) == PCH_CPT)
 #define HAS_PCH_IBX(dev) (INTEL_PCH_TYPE(dev) == PCH_IBX)
@@ -2116,8 +2276,8 @@ struct drm_i915_cmd_table {
 extern const struct drm_ioctl_desc i915_ioctls[];
 extern int i915_max_ioctl;
 
-extern int i915_suspend(struct drm_device *dev, pm_message_t state);
-extern int i915_resume(struct drm_device *dev);
+extern int i915_suspend_legacy(struct drm_device *dev, pm_message_t state);
+extern int i915_resume_legacy(struct drm_device *dev);
 extern int i915_master_create(struct drm_device *dev, struct drm_master *master);
 extern void i915_master_destroy(struct drm_device *dev, struct drm_master *master);
 
@@ -2134,6 +2294,7 @@ struct i915_params {
 	int enable_rc6;
 	int enable_fbc;
 	int enable_ppgtt;
+	int enable_execlists;
 	int enable_psr;
 	unsigned int preliminary_hw_support;
 	int disable_power_well;
@@ -2153,8 +2314,6 @@ struct i915_params {
 extern struct i915_params i915 __read_mostly;
 
 				/* i915_dma.c */
-void i915_update_dri1_breadcrumb(struct drm_device *dev);
-extern void i915_kernel_lost_context(struct drm_device * dev);
 extern int i915_driver_load(struct drm_device *, unsigned long flags);
 extern int i915_driver_unload(struct drm_device *);
 extern int i915_driver_open(struct drm_device *dev, struct drm_file *file);
@@ -2168,9 +2327,6 @@ extern int i915_driver_device_is_agp(struct drm_device * dev);
 extern long i915_compat_ioctl(struct file *filp, unsigned int cmd,
 			      unsigned long arg);
 #endif
-extern int i915_emit_box(struct drm_device *dev,
-			 struct drm_clip_rect *box,
-			 int DR1, int DR4);
 extern int intel_gpu_reset(struct drm_device *dev);
 extern int i915_reset(struct drm_device *dev);
 extern unsigned long i915_chipset_val(struct drm_i915_private *dev_priv);
@@ -2180,18 +2336,16 @@ extern void i915_update_gfx_val(struct drm_i915_private *dev_priv);
 int vlv_force_gfx_clock(struct drm_i915_private *dev_priv, bool on);
 void intel_hpd_cancel_work(struct drm_i915_private *dev_priv);
 
-extern void intel_console_resume(struct work_struct *work);
-
 /* i915_irq.c */
 void i915_queue_hangcheck(struct drm_device *dev);
 __printf(3, 4)
 void i915_handle_error(struct drm_device *dev, bool wedged,
 		       const char *fmt, ...);
 
-void gen6_set_pm_mask(struct drm_i915_private *dev_priv, u32 pm_iir,
-							int new_delay);
-extern void intel_irq_init(struct drm_device *dev);
-extern void intel_hpd_init(struct drm_device *dev);
+extern void intel_irq_init(struct drm_i915_private *dev_priv);
+extern void intel_hpd_init(struct drm_i915_private *dev_priv);
+int intel_irq_install(struct drm_i915_private *dev_priv);
+void intel_irq_uninstall(struct drm_i915_private *dev_priv);
 
 extern void intel_uncore_sanitize(struct drm_device *dev);
 extern void intel_uncore_early_sanitize(struct drm_device *dev,
@@ -2211,10 +2365,19 @@ i915_disable_pipestat(struct drm_i915_private *dev_priv, enum pipe pipe,
 
 void valleyview_enable_display_irqs(struct drm_i915_private *dev_priv);
 void valleyview_disable_display_irqs(struct drm_i915_private *dev_priv);
+void
+ironlake_enable_display_irq(struct drm_i915_private *dev_priv, u32 mask);
+void
+ironlake_disable_display_irq(struct drm_i915_private *dev_priv, u32 mask);
+void ibx_display_interrupt_update(struct drm_i915_private *dev_priv,
+				  uint32_t interrupt_mask,
+				  uint32_t enabled_irq_mask);
+#define ibx_enable_display_interrupt(dev_priv, bits) \
+	ibx_display_interrupt_update((dev_priv), (bits), (bits))
+#define ibx_disable_display_interrupt(dev_priv, bits) \
+	ibx_display_interrupt_update((dev_priv), (bits), 0)
 
 /* i915_gem.c */
-int i915_gem_init_ioctl(struct drm_device *dev, void *data,
-			struct drm_file *file_priv);
 int i915_gem_create_ioctl(struct drm_device *dev, void *data,
 			  struct drm_file *file_priv);
 int i915_gem_pread_ioctl(struct drm_device *dev, void *data,
@@ -2229,6 +2392,20 @@ int i915_gem_set_domain_ioctl(struct drm_device *dev, void *data,
 			      struct drm_file *file_priv);
 int i915_gem_sw_finish_ioctl(struct drm_device *dev, void *data,
 			     struct drm_file *file_priv);
+void i915_gem_execbuffer_move_to_active(struct list_head *vmas,
+					struct intel_engine_cs *ring);
+void i915_gem_execbuffer_retire_commands(struct drm_device *dev,
+					 struct drm_file *file,
+					 struct intel_engine_cs *ring,
+					 struct drm_i915_gem_object *obj);
+int i915_gem_ringbuffer_submission(struct drm_device *dev,
+				   struct drm_file *file,
+				   struct intel_engine_cs *ring,
+				   struct intel_context *ctx,
+				   struct drm_i915_gem_execbuffer2 *args,
+				   struct list_head *vmas,
+				   struct drm_i915_gem_object *batch_obj,
+				   u64 exec_start, u32 flags);
 int i915_gem_execbuffer(struct drm_device *dev, void *data,
 			struct drm_file *file_priv);
 int i915_gem_execbuffer2(struct drm_device *dev, void *data,
@@ -2247,10 +2424,6 @@ int i915_gem_throttle_ioctl(struct drm_device *dev, void *data,
 			    struct drm_file *file_priv);
 int i915_gem_madvise_ioctl(struct drm_device *dev, void *data,
 			   struct drm_file *file_priv);
-int i915_gem_entervt_ioctl(struct drm_device *dev, void *data,
-			   struct drm_file *file_priv);
-int i915_gem_leavevt_ioctl(struct drm_device *dev, void *data,
-			   struct drm_file *file_priv);
 int i915_gem_set_tiling(struct drm_device *dev, void *data,
 			struct drm_file *file_priv);
 int i915_gem_get_tiling(struct drm_device *dev, void *data,
@@ -2263,6 +2436,12 @@ int i915_gem_get_aperture_ioctl(struct drm_device *dev, void *data,
 int i915_gem_wait_ioctl(struct drm_device *dev, void *data,
 			struct drm_file *file_priv);
 void i915_gem_load(struct drm_device *dev);
+unsigned long i915_gem_shrink(struct drm_i915_private *dev_priv,
+			      long target,
+			      unsigned flags);
+#define I915_SHRINK_PURGEABLE 0x1
+#define I915_SHRINK_UNBOUND 0x2
+#define I915_SHRINK_BOUND 0x4
 void *i915_gem_object_alloc(struct drm_device *dev);
 void i915_gem_object_free(struct drm_i915_gem_object *obj);
 void i915_gem_object_init(struct drm_i915_gem_object *obj,
@@ -2287,7 +2466,6 @@ int __must_check i915_vma_unbind(struct i915_vma *vma);
 int i915_gem_object_put_pages(struct drm_i915_gem_object *obj);
 void i915_gem_release_all_mmaps(struct drm_i915_private *dev_priv);
 void i915_gem_release_mmap(struct drm_i915_gem_object *obj);
-void i915_gem_lastclose(struct drm_device *dev);
 
 int i915_gem_obj_prepare_shmem_read(struct drm_i915_gem_object *obj,
 				    int *needs_clflush);
@@ -2381,6 +2559,7 @@ void i915_gem_reset(struct drm_device *dev);
 bool i915_gem_clflush_object(struct drm_i915_gem_object *obj, bool force);
 int __must_check i915_gem_object_finish_gpu(struct drm_i915_gem_object *obj);
 int __must_check i915_gem_init(struct drm_device *dev);
+int i915_gem_init_rings(struct drm_device *dev);
 int __must_check i915_gem_init_hw(struct drm_device *dev);
 int i915_gem_l3_remap(struct intel_engine_cs *ring, int slice);
 void i915_gem_init_swizzling(struct drm_device *dev);
@@ -2393,6 +2572,11 @@ int __i915_add_request(struct intel_engine_cs *ring,
 		       u32 *seqno);
 #define i915_add_request(ring, seqno) \
 	__i915_add_request(ring, NULL, NULL, seqno)
+int __i915_wait_seqno(struct intel_engine_cs *ring, u32 seqno,
+			unsigned reset_counter,
+			bool interruptible,
+			s64 *timeout,
+			struct drm_i915_file_private *file_priv);
 int __must_check i915_wait_seqno(struct intel_engine_cs *ring,
 				 uint32_t seqno);
 int i915_gem_fault(struct vm_area_struct *vma, struct vm_fault *vmf);
@@ -2451,7 +2635,7 @@ static inline bool i915_gem_obj_is_pinned(struct drm_i915_gem_object *obj) {
 }
 
 /* Some GGTT VM helpers */
-#define obj_to_ggtt(obj) \
+#define i915_obj_to_ggtt(obj) \
 	(&((struct drm_i915_private *)(obj)->base.dev->dev_private)->gtt.base)
 static inline bool i915_is_ggtt(struct i915_address_space *vm)
 {
@@ -2460,21 +2644,30 @@ static inline bool i915_is_ggtt(struct i915_address_space *vm)
 	return vm == ggtt;
 }
 
+static inline struct i915_hw_ppgtt *
+i915_vm_to_ppgtt(struct i915_address_space *vm)
+{
+	WARN_ON(i915_is_ggtt(vm));
+
+	return container_of(vm, struct i915_hw_ppgtt, base);
+}
+
+
 static inline bool i915_gem_obj_ggtt_bound(struct drm_i915_gem_object *obj)
 {
-	return i915_gem_obj_bound(obj, obj_to_ggtt(obj));
+	return i915_gem_obj_bound(obj, i915_obj_to_ggtt(obj));
 }
 
 static inline unsigned long
 i915_gem_obj_ggtt_offset(struct drm_i915_gem_object *obj)
 {
-	return i915_gem_obj_offset(obj, obj_to_ggtt(obj));
+	return i915_gem_obj_offset(obj, i915_obj_to_ggtt(obj));
 }
 
 static inline unsigned long
 i915_gem_obj_ggtt_size(struct drm_i915_gem_object *obj)
 {
-	return i915_gem_obj_size(obj, obj_to_ggtt(obj));
+	return i915_gem_obj_size(obj, i915_obj_to_ggtt(obj));
 }
 
 static inline int __must_check
@@ -2482,7 +2675,8 @@ i915_gem_obj_ggtt_pin(struct drm_i915_gem_object *obj,
 		      uint32_t alignment,
 		      unsigned flags)
 {
-	return i915_gem_object_pin(obj, obj_to_ggtt(obj), alignment, flags | PIN_GLOBAL);
+	return i915_gem_object_pin(obj, i915_obj_to_ggtt(obj),
+				   alignment, flags | PIN_GLOBAL);
 }
 
 static inline int
@@ -2494,7 +2688,6 @@ i915_gem_object_ggtt_unbind(struct drm_i915_gem_object *obj)
 void i915_gem_object_ggtt_unpin(struct drm_i915_gem_object *obj);
 
 /* i915_gem_context.c */
-#define ctx_to_ppgtt(ctx) container_of((ctx)->vm, struct i915_hw_ppgtt, base)
 int __must_check i915_gem_context_init(struct drm_device *dev);
 void i915_gem_context_fini(struct drm_device *dev);
 void i915_gem_context_reset(struct drm_device *dev);
@@ -2506,6 +2699,8 @@ int i915_switch_context(struct intel_engine_cs *ring,
 struct intel_context *
 i915_gem_context_get(struct drm_i915_file_private *file_priv, u32 id);
 void i915_gem_context_free(struct kref *ctx_ref);
+struct drm_i915_gem_object *
+i915_gem_alloc_context_obj(struct drm_device *dev, size_t size);
 static inline void i915_gem_context_reference(struct intel_context *ctx)
 {
 	kref_get(&ctx->ref);
@@ -2526,8 +2721,6 @@ int i915_gem_context_create_ioctl(struct drm_device *dev, void *data,
 int i915_gem_context_destroy_ioctl(struct drm_device *dev, void *data,
 				   struct drm_file *file);
 
-/* i915_gem_render_state.c */
-int i915_gem_render_state_init(struct intel_engine_cs *ring);
 /* i915_gem_evict.c */
 int __must_check i915_gem_evict_something(struct drm_device *dev,
 					  struct i915_address_space *vm,
@@ -2595,6 +2788,7 @@ void i915_error_printf(struct drm_i915_error_state_buf *e, const char *f, ...);
 int i915_error_state_to_str(struct drm_i915_error_state_buf *estr,
 			    const struct i915_error_state_file_priv *error);
 int i915_error_state_buf_init(struct drm_i915_error_state_buf *eb,
+			      struct drm_i915_private *i915,
 			      size_t count, loff_t pos);
 static inline void i915_error_state_buf_release(
 	struct drm_i915_error_state_buf *eb)
@@ -2609,7 +2803,7 @@ void i915_error_state_put(struct i915_error_state_file_priv *error_priv);
 void i915_destroy_error_state(struct drm_device *dev);
 
 void i915_get_extra_instdone(struct drm_device *dev, uint32_t *instdone);
-const char *i915_cache_level_str(int type);
+const char *i915_cache_level_str(struct drm_i915_private *i915, int type);
 
 /* i915_cmd_parser.c */
 int i915_cmd_parser_get_version(void);
@@ -2652,7 +2846,6 @@ static inline bool intel_gmbus_is_forced_bit(struct i2c_adapter *adapter)
 extern void intel_i2c_reset(struct drm_device *dev);
 
 /* intel_opregion.c */
-struct intel_encoder;
 #ifdef CONFIG_ACPI
 extern int intel_opregion_setup(struct drm_device *dev);
 extern void intel_opregion_init(struct drm_device *dev);
@@ -2690,7 +2883,6 @@ static inline void intel_unregister_dsm_handler(void) { return; }
 
 /* modesetting */
 extern void intel_modeset_init_hw(struct drm_device *dev);
-extern void intel_modeset_suspend_hw(struct drm_device *dev);
 extern void intel_modeset_init(struct drm_device *dev);
 extern void intel_modeset_gem_init(struct drm_device *dev);
 extern void intel_modeset_cleanup(struct drm_device *dev);
@@ -2701,6 +2893,7 @@ extern void intel_modeset_setup_hw_state(struct drm_device *dev,
 extern void i915_redisable_vga(struct drm_device *dev);
 extern void i915_redisable_vga_power_on(struct drm_device *dev);
 extern bool intel_fbc_enabled(struct drm_device *dev);
+extern void bdw_fbc_sw_flush(struct drm_device *dev, u32 value);
 extern void intel_disable_fbc(struct drm_device *dev);
 extern bool ironlake_set_drps(struct drm_device *dev, u8 val);
 extern void intel_init_pch_refclk(struct drm_device *dev);
@@ -2738,8 +2931,8 @@ void gen6_gt_force_wake_get(struct drm_i915_private *dev_priv, int fw_engine);
 void gen6_gt_force_wake_put(struct drm_i915_private *dev_priv, int fw_engine);
 void assert_force_wake_inactive(struct drm_i915_private *dev_priv);
 
-int sandybridge_pcode_read(struct drm_i915_private *dev_priv, u8 mbox, u32 *val);
-int sandybridge_pcode_write(struct drm_i915_private *dev_priv, u8 mbox, u32 val);
+int sandybridge_pcode_read(struct drm_i915_private *dev_priv, u32 mbox, u32 *val);
+int sandybridge_pcode_write(struct drm_i915_private *dev_priv, u32 mbox, u32 val);
 
 /* intel_sideband.c */
 u32 vlv_punit_read(struct drm_i915_private *dev_priv, u8 addr);
@@ -2769,7 +2962,9 @@ int vlv_freq_opcode(struct drm_i915_private *dev_priv, int val);
 
 #define FORCEWAKE_RENDER	(1 << 0)
 #define FORCEWAKE_MEDIA		(1 << 1)
-#define FORCEWAKE_ALL		(FORCEWAKE_RENDER | FORCEWAKE_MEDIA)
+#define FORCEWAKE_BLITTER	(1 << 2)
+#define FORCEWAKE_ALL		(FORCEWAKE_RENDER | FORCEWAKE_MEDIA | \
+					FORCEWAKE_BLITTER)
 
 
 #define I915_READ8(reg)		dev_priv->uncore.funcs.mmio_readb(dev_priv, (reg), true)
@@ -2835,6 +3030,11 @@ static inline unsigned long msecs_to_jiffies_timeout(const unsigned int m)
 	return min_t(unsigned long, MAX_JIFFY_OFFSET, j + 1);
 }
 
+static inline unsigned long nsecs_to_jiffies_timeout(const u64 n)
+{
+        return min_t(u64, MAX_JIFFY_OFFSET, nsecs_to_jiffies64(n) + 1);
+}
+
 static inline unsigned long
 timespec_to_jiffies_timeout(const struct timespec *value)
 {
diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c
index ad55b06a3cb1..c11603b4cf1d 100644
--- a/drivers/gpu/drm/i915/i915_gem.c
+++ b/drivers/gpu/drm/i915/i915_gem.c
@@ -60,7 +60,6 @@ static unsigned long i915_gem_shrinker_scan(struct shrinker *shrinker,
 static int i915_gem_shrinker_oom(struct notifier_block *nb,
 				 unsigned long event,
 				 void *ptr);
-static unsigned long i915_gem_purge(struct drm_i915_private *dev_priv, long target);
 static unsigned long i915_gem_shrink_all(struct drm_i915_private *dev_priv);
 
 static bool cpu_cache_is_coherent(struct drm_device *dev,
@@ -161,33 +160,6 @@ i915_gem_object_is_inactive(struct drm_i915_gem_object *obj)
 }
 
 int
-i915_gem_init_ioctl(struct drm_device *dev, void *data,
-		    struct drm_file *file)
-{
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct drm_i915_gem_init *args = data;
-
-	if (drm_core_check_feature(dev, DRIVER_MODESET))
-		return -ENODEV;
-
-	if (args->gtt_start >= args->gtt_end ||
-	    (args->gtt_end | args->gtt_start) & (PAGE_SIZE - 1))
-		return -EINVAL;
-
-	/* GEM with user mode setting was never supported on ilk and later. */
-	if (INTEL_INFO(dev)->gen >= 5)
-		return -ENODEV;
-
-	mutex_lock(&dev->struct_mutex);
-	i915_gem_setup_global_gtt(dev, args->gtt_start, args->gtt_end,
-				  args->gtt_end);
-	dev_priv->gtt.mappable_end = args->gtt_end;
-	mutex_unlock(&dev->struct_mutex);
-
-	return 0;
-}
-
-int
 i915_gem_get_aperture_ioctl(struct drm_device *dev, void *data,
 			    struct drm_file *file)
 {
@@ -209,40 +181,137 @@ i915_gem_get_aperture_ioctl(struct drm_device *dev, void *data,
 	return 0;
 }
 
-static void i915_gem_object_detach_phys(struct drm_i915_gem_object *obj)
+static int
+i915_gem_object_get_pages_phys(struct drm_i915_gem_object *obj)
 {
-	drm_dma_handle_t *phys = obj->phys_handle;
+	struct address_space *mapping = file_inode(obj->base.filp)->i_mapping;
+	char *vaddr = obj->phys_handle->vaddr;
+	struct sg_table *st;
+	struct scatterlist *sg;
+	int i;
 
-	if (!phys)
-		return;
+	if (WARN_ON(i915_gem_object_needs_bit17_swizzle(obj)))
+		return -EINVAL;
+
+	for (i = 0; i < obj->base.size / PAGE_SIZE; i++) {
+		struct page *page;
+		char *src;
+
+		page = shmem_read_mapping_page(mapping, i);
+		if (IS_ERR(page))
+			return PTR_ERR(page);
+
+		src = kmap_atomic(page);
+		memcpy(vaddr, src, PAGE_SIZE);
+		drm_clflush_virt_range(vaddr, PAGE_SIZE);
+		kunmap_atomic(src);
+
+		page_cache_release(page);
+		vaddr += PAGE_SIZE;
+	}
+
+	i915_gem_chipset_flush(obj->base.dev);
+
+	st = kmalloc(sizeof(*st), GFP_KERNEL);
+	if (st == NULL)
+		return -ENOMEM;
+
+	if (sg_alloc_table(st, 1, GFP_KERNEL)) {
+		kfree(st);
+		return -ENOMEM;
+	}
+
+	sg = st->sgl;
+	sg->offset = 0;
+	sg->length = obj->base.size;
+
+	sg_dma_address(sg) = obj->phys_handle->busaddr;
+	sg_dma_len(sg) = obj->base.size;
 
-	if (obj->madv == I915_MADV_WILLNEED) {
+	obj->pages = st;
+	obj->has_dma_mapping = true;
+	return 0;
+}
+
+static void
+i915_gem_object_put_pages_phys(struct drm_i915_gem_object *obj)
+{
+	int ret;
+
+	BUG_ON(obj->madv == __I915_MADV_PURGED);
+
+	ret = i915_gem_object_set_to_cpu_domain(obj, true);
+	if (ret) {
+		/* In the event of a disaster, abandon all caches and
+		 * hope for the best.
+		 */
+		WARN_ON(ret != -EIO);
+		obj->base.read_domains = obj->base.write_domain = I915_GEM_DOMAIN_CPU;
+	}
+
+	if (obj->madv == I915_MADV_DONTNEED)
+		obj->dirty = 0;
+
+	if (obj->dirty) {
 		struct address_space *mapping = file_inode(obj->base.filp)->i_mapping;
-		char *vaddr = phys->vaddr;
+		char *vaddr = obj->phys_handle->vaddr;
 		int i;
 
 		for (i = 0; i < obj->base.size / PAGE_SIZE; i++) {
-			struct page *page = shmem_read_mapping_page(mapping, i);
-			if (!IS_ERR(page)) {
-				char *dst = kmap_atomic(page);
-				memcpy(dst, vaddr, PAGE_SIZE);
-				drm_clflush_virt_range(dst, PAGE_SIZE);
-				kunmap_atomic(dst);
-
-				set_page_dirty(page);
+			struct page *page;
+			char *dst;
+
+			page = shmem_read_mapping_page(mapping, i);
+			if (IS_ERR(page))
+				continue;
+
+			dst = kmap_atomic(page);
+			drm_clflush_virt_range(vaddr, PAGE_SIZE);
+			memcpy(dst, vaddr, PAGE_SIZE);
+			kunmap_atomic(dst);
+
+			set_page_dirty(page);
+			if (obj->madv == I915_MADV_WILLNEED)
 				mark_page_accessed(page);
-				page_cache_release(page);
-			}
+			page_cache_release(page);
 			vaddr += PAGE_SIZE;
 		}
-		i915_gem_chipset_flush(obj->base.dev);
+		obj->dirty = 0;
 	}
 
-#ifdef CONFIG_X86
-	set_memory_wb((unsigned long)phys->vaddr, phys->size / PAGE_SIZE);
-#endif
-	drm_pci_free(obj->base.dev, phys);
-	obj->phys_handle = NULL;
+	sg_free_table(obj->pages);
+	kfree(obj->pages);
+
+	obj->has_dma_mapping = false;
+}
+
+static void
+i915_gem_object_release_phys(struct drm_i915_gem_object *obj)
+{
+	drm_pci_free(obj->base.dev, obj->phys_handle);
+}
+
+static const struct drm_i915_gem_object_ops i915_gem_phys_ops = {
+	.get_pages = i915_gem_object_get_pages_phys,
+	.put_pages = i915_gem_object_put_pages_phys,
+	.release = i915_gem_object_release_phys,
+};
+
+static int
+drop_pages(struct drm_i915_gem_object *obj)
+{
+	struct i915_vma *vma, *next;
+	int ret;
+
+	drm_gem_object_reference(&obj->base);
+	list_for_each_entry_safe(vma, next, &obj->vma_list, vma_link)
+		if (i915_vma_unbind(vma))
+			break;
+
+	ret = i915_gem_object_put_pages(obj);
+	drm_gem_object_unreference(&obj->base);
+
+	return ret;
 }
 
 int
@@ -250,9 +319,7 @@ i915_gem_object_attach_phys(struct drm_i915_gem_object *obj,
 			    int align)
 {
 	drm_dma_handle_t *phys;
-	struct address_space *mapping;
-	char *vaddr;
-	int i;
+	int ret;
 
 	if (obj->phys_handle) {
 		if ((unsigned long)obj->phys_handle->vaddr & (align -1))
@@ -267,41 +334,19 @@ i915_gem_object_attach_phys(struct drm_i915_gem_object *obj,
 	if (obj->base.filp == NULL)
 		return -EINVAL;
 
+	ret = drop_pages(obj);
+	if (ret)
+		return ret;
+
 	/* create a new object */
 	phys = drm_pci_alloc(obj->base.dev, obj->base.size, align);
 	if (!phys)
 		return -ENOMEM;
 
-	vaddr = phys->vaddr;
-#ifdef CONFIG_X86
-	set_memory_wc((unsigned long)vaddr, phys->size / PAGE_SIZE);
-#endif
-	mapping = file_inode(obj->base.filp)->i_mapping;
-	for (i = 0; i < obj->base.size / PAGE_SIZE; i++) {
-		struct page *page;
-		char *src;
-
-		page = shmem_read_mapping_page(mapping, i);
-		if (IS_ERR(page)) {
-#ifdef CONFIG_X86
-			set_memory_wb((unsigned long)phys->vaddr, phys->size / PAGE_SIZE);
-#endif
-			drm_pci_free(obj->base.dev, phys);
-			return PTR_ERR(page);
-		}
-
-		src = kmap_atomic(page);
-		memcpy(vaddr, src, PAGE_SIZE);
-		kunmap_atomic(src);
-
-		mark_page_accessed(page);
-		page_cache_release(page);
-
-		vaddr += PAGE_SIZE;
-	}
-
 	obj->phys_handle = phys;
-	return 0;
+	obj->ops = &i915_gem_phys_ops;
+
+	return i915_gem_object_get_pages(obj);
 }
 
 static int
@@ -312,6 +357,14 @@ i915_gem_phys_pwrite(struct drm_i915_gem_object *obj,
 	struct drm_device *dev = obj->base.dev;
 	void *vaddr = obj->phys_handle->vaddr + args->offset;
 	char __user *user_data = to_user_ptr(args->data_ptr);
+	int ret;
+
+	/* We manually control the domain here and pretend that it
+	 * remains coherent i.e. in the GTT domain, like shmem_pwrite.
+	 */
+	ret = i915_gem_object_wait_rendering(obj, false);
+	if (ret)
+		return ret;
 
 	if (__copy_from_user_inatomic_nocache(vaddr, user_data, args->size)) {
 		unsigned long unwritten;
@@ -327,6 +380,7 @@ i915_gem_phys_pwrite(struct drm_i915_gem_object *obj,
 			return -EFAULT;
 	}
 
+	drm_clflush_virt_range(vaddr, args->size);
 	i915_gem_chipset_flush(dev);
 	return 0;
 }
@@ -994,6 +1048,7 @@ int
 i915_gem_pwrite_ioctl(struct drm_device *dev, void *data,
 		      struct drm_file *file)
 {
+	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct drm_i915_gem_pwrite *args = data;
 	struct drm_i915_gem_object *obj;
 	int ret;
@@ -1013,9 +1068,11 @@ i915_gem_pwrite_ioctl(struct drm_device *dev, void *data,
 			return -EFAULT;
 	}
 
+	intel_runtime_pm_get(dev_priv);
+
 	ret = i915_mutex_lock_interruptible(dev);
 	if (ret)
-		return ret;
+		goto put_rpm;
 
 	obj = to_intel_bo(drm_gem_object_lookup(dev, file, args->handle));
 	if (&obj->base == NULL) {
@@ -1047,11 +1104,6 @@ i915_gem_pwrite_ioctl(struct drm_device *dev, void *data,
 	 * pread/pwrite currently are reading and writing from the CPU
 	 * perspective, requiring manual detiling by the client.
 	 */
-	if (obj->phys_handle) {
-		ret = i915_gem_phys_pwrite(obj, args, file);
-		goto out;
-	}
-
 	if (obj->tiling_mode == I915_TILING_NONE &&
 	    obj->base.write_domain != I915_GEM_DOMAIN_CPU &&
 	    cpu_write_needs_clflush(obj)) {
@@ -1061,13 +1113,20 @@ i915_gem_pwrite_ioctl(struct drm_device *dev, void *data,
 		 * textures). Fallback to the shmem path in that case. */
 	}
 
-	if (ret == -EFAULT || ret == -ENOSPC)
-		ret = i915_gem_shmem_pwrite(dev, obj, args, file);
+	if (ret == -EFAULT || ret == -ENOSPC) {
+		if (obj->phys_handle)
+			ret = i915_gem_phys_pwrite(obj, args, file);
+		else
+			ret = i915_gem_shmem_pwrite(dev, obj, args, file);
+	}
 
 out:
 	drm_gem_object_unreference(&obj->base);
 unlock:
 	mutex_unlock(&dev->struct_mutex);
+put_rpm:
+	intel_runtime_pm_put(dev_priv);
+
 	return ret;
 }
 
@@ -1085,7 +1144,13 @@ i915_gem_check_wedge(struct i915_gpu_error *error,
 		if (i915_terminally_wedged(error))
 			return -EIO;
 
-		return -EAGAIN;
+		/*
+		 * Check if GPU Reset is in progress - we need intel_ring_begin
+		 * to work properly to reinit the hw state while the gpu is
+		 * still marked as reset-in-progress. Handle this with a flag.
+		 */
+		if (!error->reload_in_reset)
+			return -EAGAIN;
 	}
 
 	return 0;
@@ -1129,7 +1194,7 @@ static bool can_wait_boost(struct drm_i915_file_private *file_priv)
 }
 
 /**
- * __wait_seqno - wait until execution of seqno has finished
+ * __i915_wait_seqno - wait until execution of seqno has finished
  * @ring: the ring expected to report seqno
  * @seqno: duh!
  * @reset_counter: reset sequence associated with the given seqno
@@ -1146,7 +1211,7 @@ static bool can_wait_boost(struct drm_i915_file_private *file_priv)
  * Returns 0 if the seqno was found within the alloted time. Else returns the
  * errno with remaining time filled in timeout argument.
  */
-static int __wait_seqno(struct intel_engine_cs *ring, u32 seqno,
+int __i915_wait_seqno(struct intel_engine_cs *ring, u32 seqno,
 			unsigned reset_counter,
 			bool interruptible,
 			s64 *timeout,
@@ -1166,7 +1231,8 @@ static int __wait_seqno(struct intel_engine_cs *ring, u32 seqno,
 	if (i915_seqno_passed(ring->get_seqno(ring, true), seqno))
 		return 0;
 
-	timeout_expire = timeout ? jiffies + nsecs_to_jiffies((u64)*timeout) : 0;
+	timeout_expire = timeout ?
+		jiffies + nsecs_to_jiffies_timeout((u64)*timeout) : 0;
 
 	if (INTEL_INFO(dev)->gen >= 6 && ring->id == RCS && can_wait_boost(file_priv)) {
 		gen6_rps_boost(dev_priv);
@@ -1242,6 +1308,16 @@ static int __wait_seqno(struct intel_engine_cs *ring, u32 seqno,
 		s64 tres = *timeout - (now - before);
 
 		*timeout = tres < 0 ? 0 : tres;
+
+		/*
+		 * Apparently ktime isn't accurate enough and occasionally has a
+		 * bit of mismatch in the jiffies<->nsecs<->ktime loop. So patch
+		 * things up to make the test happy. We allow up to 1 jiffy.
+		 *
+		 * This is a regrssion from the timespec->ktime conversion.
+		 */
+		if (ret == -ETIME && *timeout < jiffies_to_usecs(1)*1000)
+			*timeout = 0;
 	}
 
 	return ret;
@@ -1257,6 +1333,7 @@ i915_wait_seqno(struct intel_engine_cs *ring, uint32_t seqno)
 	struct drm_device *dev = ring->dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	bool interruptible = dev_priv->mm.interruptible;
+	unsigned reset_counter;
 	int ret;
 
 	BUG_ON(!mutex_is_locked(&dev->struct_mutex));
@@ -1270,14 +1347,13 @@ i915_wait_seqno(struct intel_engine_cs *ring, uint32_t seqno)
 	if (ret)
 		return ret;
 
-	return __wait_seqno(ring, seqno,
-			    atomic_read(&dev_priv->gpu_error.reset_counter),
-			    interruptible, NULL, NULL);
+	reset_counter = atomic_read(&dev_priv->gpu_error.reset_counter);
+	return __i915_wait_seqno(ring, seqno, reset_counter, interruptible,
+				 NULL, NULL);
 }
 
 static int
-i915_gem_object_wait_rendering__tail(struct drm_i915_gem_object *obj,
-				     struct intel_engine_cs *ring)
+i915_gem_object_wait_rendering__tail(struct drm_i915_gem_object *obj)
 {
 	if (!obj->active)
 		return 0;
@@ -1314,7 +1390,7 @@ i915_gem_object_wait_rendering(struct drm_i915_gem_object *obj,
 	if (ret)
 		return ret;
 
-	return i915_gem_object_wait_rendering__tail(obj, ring);
+	return i915_gem_object_wait_rendering__tail(obj);
 }
 
 /* A nonblocking variant of the above wait. This is a highly dangerous routine
@@ -1349,12 +1425,13 @@ i915_gem_object_wait_rendering__nonblocking(struct drm_i915_gem_object *obj,
 
 	reset_counter = atomic_read(&dev_priv->gpu_error.reset_counter);
 	mutex_unlock(&dev->struct_mutex);
-	ret = __wait_seqno(ring, seqno, reset_counter, true, NULL, file_priv);
+	ret = __i915_wait_seqno(ring, seqno, reset_counter, true, NULL,
+				file_priv);
 	mutex_lock(&dev->struct_mutex);
 	if (ret)
 		return ret;
 
-	return i915_gem_object_wait_rendering__tail(obj, ring);
+	return i915_gem_object_wait_rendering__tail(obj);
 }
 
 /**
@@ -1461,6 +1538,16 @@ unlock:
  *
  * While the mapping holds a reference on the contents of the object, it doesn't
  * imply a ref on the object itself.
+ *
+ * IMPORTANT:
+ *
+ * DRM driver writers who look a this function as an example for how to do GEM
+ * mmap support, please don't implement mmap support like here. The modern way
+ * to implement DRM mmap support is with an mmap offset ioctl (like
+ * i915_gem_mmap_gtt) and then using the mmap syscall on the DRM fd directly.
+ * That way debug tooling like valgrind will understand what's going on, hiding
+ * the mmap call in a driver private ioctl will break that. The i915 driver only
+ * does cpu mmaps this way because we didn't know better.
  */
 int
 i915_gem_mmap_ioctl(struct drm_device *dev, void *data,
@@ -1735,7 +1822,11 @@ static int i915_gem_object_create_mmap_offset(struct drm_i915_gem_object *obj)
 	 * offsets on purgeable objects by truncating it and marking it purged,
 	 * which prevents userspace from ever using that object again.
 	 */
-	i915_gem_purge(dev_priv, obj->base.size >> PAGE_SHIFT);
+	i915_gem_shrink(dev_priv,
+			obj->base.size >> PAGE_SHIFT,
+			I915_SHRINK_BOUND |
+			I915_SHRINK_UNBOUND |
+			I915_SHRINK_PURGEABLE);
 	ret = drm_gem_create_mmap_offset(&obj->base);
 	if (ret != -ENOSPC)
 		goto out;
@@ -1932,12 +2023,18 @@ i915_gem_object_put_pages(struct drm_i915_gem_object *obj)
 	return 0;
 }
 
-static unsigned long
-__i915_gem_shrink(struct drm_i915_private *dev_priv, long target,
-		  bool purgeable_only)
+unsigned long
+i915_gem_shrink(struct drm_i915_private *dev_priv,
+		long target, unsigned flags)
 {
-	struct list_head still_in_list;
-	struct drm_i915_gem_object *obj;
+	const struct {
+		struct list_head *list;
+		unsigned int bit;
+	} phases[] = {
+		{ &dev_priv->mm.unbound_list, I915_SHRINK_UNBOUND },
+		{ &dev_priv->mm.bound_list, I915_SHRINK_BOUND },
+		{ NULL, 0 },
+	}, *phase;
 	unsigned long count = 0;
 
 	/*
@@ -1959,62 +2056,50 @@ __i915_gem_shrink(struct drm_i915_private *dev_priv, long target,
 	 * dev->struct_mutex and so we won't ever be able to observe an
 	 * object on the bound_list with a reference count equals 0.
 	 */
-	INIT_LIST_HEAD(&still_in_list);
-	while (count < target && !list_empty(&dev_priv->mm.unbound_list)) {
-		obj = list_first_entry(&dev_priv->mm.unbound_list,
-				       typeof(*obj), global_list);
-		list_move_tail(&obj->global_list, &still_in_list);
+	for (phase = phases; phase->list; phase++) {
+		struct list_head still_in_list;
 
-		if (!i915_gem_object_is_purgeable(obj) && purgeable_only)
+		if ((flags & phase->bit) == 0)
 			continue;
 
-		drm_gem_object_reference(&obj->base);
-
-		if (i915_gem_object_put_pages(obj) == 0)
-			count += obj->base.size >> PAGE_SHIFT;
-
-		drm_gem_object_unreference(&obj->base);
-	}
-	list_splice(&still_in_list, &dev_priv->mm.unbound_list);
-
-	INIT_LIST_HEAD(&still_in_list);
-	while (count < target && !list_empty(&dev_priv->mm.bound_list)) {
-		struct i915_vma *vma, *v;
+		INIT_LIST_HEAD(&still_in_list);
+		while (count < target && !list_empty(phase->list)) {
+			struct drm_i915_gem_object *obj;
+			struct i915_vma *vma, *v;
 
-		obj = list_first_entry(&dev_priv->mm.bound_list,
-				       typeof(*obj), global_list);
-		list_move_tail(&obj->global_list, &still_in_list);
+			obj = list_first_entry(phase->list,
+					       typeof(*obj), global_list);
+			list_move_tail(&obj->global_list, &still_in_list);
 
-		if (!i915_gem_object_is_purgeable(obj) && purgeable_only)
-			continue;
+			if (flags & I915_SHRINK_PURGEABLE &&
+			    !i915_gem_object_is_purgeable(obj))
+				continue;
 
-		drm_gem_object_reference(&obj->base);
+			drm_gem_object_reference(&obj->base);
 
-		list_for_each_entry_safe(vma, v, &obj->vma_list, vma_link)
-			if (i915_vma_unbind(vma))
-				break;
+			/* For the unbound phase, this should be a no-op! */
+			list_for_each_entry_safe(vma, v,
+						 &obj->vma_list, vma_link)
+				if (i915_vma_unbind(vma))
+					break;
 
-		if (i915_gem_object_put_pages(obj) == 0)
-			count += obj->base.size >> PAGE_SHIFT;
+			if (i915_gem_object_put_pages(obj) == 0)
+				count += obj->base.size >> PAGE_SHIFT;
 
-		drm_gem_object_unreference(&obj->base);
+			drm_gem_object_unreference(&obj->base);
+		}
+		list_splice(&still_in_list, phase->list);
 	}
-	list_splice(&still_in_list, &dev_priv->mm.bound_list);
 
 	return count;
 }
 
 static unsigned long
-i915_gem_purge(struct drm_i915_private *dev_priv, long target)
-{
-	return __i915_gem_shrink(dev_priv, target, true);
-}
-
-static unsigned long
 i915_gem_shrink_all(struct drm_i915_private *dev_priv)
 {
 	i915_gem_evict_everything(dev_priv->dev);
-	return __i915_gem_shrink(dev_priv, LONG_MAX, false);
+	return i915_gem_shrink(dev_priv, LONG_MAX,
+			       I915_SHRINK_BOUND | I915_SHRINK_UNBOUND);
 }
 
 static int
@@ -2061,7 +2146,11 @@ i915_gem_object_get_pages_gtt(struct drm_i915_gem_object *obj)
 	for (i = 0; i < page_count; i++) {
 		page = shmem_read_mapping_page_gfp(mapping, i, gfp);
 		if (IS_ERR(page)) {
-			i915_gem_purge(dev_priv, page_count);
+			i915_gem_shrink(dev_priv,
+					page_count,
+					I915_SHRINK_BOUND |
+					I915_SHRINK_UNBOUND |
+					I915_SHRINK_PURGEABLE);
 			page = shmem_read_mapping_page_gfp(mapping, i, gfp);
 		}
 		if (IS_ERR(page)) {
@@ -2104,6 +2193,10 @@ i915_gem_object_get_pages_gtt(struct drm_i915_gem_object *obj)
 	if (i915_gem_object_needs_bit17_swizzle(obj))
 		i915_gem_object_do_bit_17_swizzle(obj);
 
+	if (obj->tiling_mode != I915_TILING_NONE &&
+	    dev_priv->quirks & QUIRK_PIN_SWIZZLED_PAGES)
+		i915_gem_object_pin_pages(obj);
+
 	return 0;
 
 err_pages:
@@ -2163,8 +2256,6 @@ static void
 i915_gem_object_move_to_active(struct drm_i915_gem_object *obj,
 			       struct intel_engine_cs *ring)
 {
-	struct drm_device *dev = obj->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
 	u32 seqno = intel_ring_get_seqno(ring);
 
 	BUG_ON(ring == NULL);
@@ -2183,19 +2274,6 @@ i915_gem_object_move_to_active(struct drm_i915_gem_object *obj,
 	list_move_tail(&obj->ring_list, &ring->active_list);
 
 	obj->last_read_seqno = seqno;
-
-	if (obj->fenced_gpu_access) {
-		obj->last_fenced_seqno = seqno;
-
-		/* Bump MRU to take account of the delayed flush */
-		if (obj->fence_reg != I915_FENCE_REG_NONE) {
-			struct drm_i915_fence_reg *reg;
-
-			reg = &dev_priv->fence_regs[obj->fence_reg];
-			list_move_tail(&reg->lru_list,
-				       &dev_priv->mm.fence_list);
-		}
-	}
 }
 
 void i915_vma_move_to_active(struct i915_vma *vma,
@@ -2231,7 +2309,6 @@ i915_gem_object_move_to_inactive(struct drm_i915_gem_object *obj)
 	obj->base.write_domain = 0;
 
 	obj->last_fenced_seqno = 0;
-	obj->fenced_gpu_access = false;
 
 	obj->active = 0;
 	drm_gem_object_unreference(&obj->base);
@@ -2329,10 +2406,21 @@ int __i915_add_request(struct intel_engine_cs *ring,
 {
 	struct drm_i915_private *dev_priv = ring->dev->dev_private;
 	struct drm_i915_gem_request *request;
+	struct intel_ringbuffer *ringbuf;
 	u32 request_ring_position, request_start;
 	int ret;
 
-	request_start = intel_ring_get_tail(ring->buffer);
+	request = ring->preallocated_lazy_request;
+	if (WARN_ON(request == NULL))
+		return -ENOMEM;
+
+	if (i915.enable_execlists) {
+		struct intel_context *ctx = request->ctx;
+		ringbuf = ctx->engine[ring->id].ringbuf;
+	} else
+		ringbuf = ring->buffer;
+
+	request_start = intel_ring_get_tail(ringbuf);
 	/*
 	 * Emit any outstanding flushes - execbuf can fail to emit the flush
 	 * after having emitted the batchbuffer command. Hence we need to fix
@@ -2340,24 +2428,32 @@ int __i915_add_request(struct intel_engine_cs *ring,
 	 * is that the flush _must_ happen before the next request, no matter
 	 * what.
 	 */
-	ret = intel_ring_flush_all_caches(ring);
-	if (ret)
-		return ret;
-
-	request = ring->preallocated_lazy_request;
-	if (WARN_ON(request == NULL))
-		return -ENOMEM;
+	if (i915.enable_execlists) {
+		ret = logical_ring_flush_all_caches(ringbuf);
+		if (ret)
+			return ret;
+	} else {
+		ret = intel_ring_flush_all_caches(ring);
+		if (ret)
+			return ret;
+	}
 
 	/* Record the position of the start of the request so that
 	 * should we detect the updated seqno part-way through the
 	 * GPU processing the request, we never over-estimate the
 	 * position of the head.
 	 */
-	request_ring_position = intel_ring_get_tail(ring->buffer);
+	request_ring_position = intel_ring_get_tail(ringbuf);
 
-	ret = ring->add_request(ring);
-	if (ret)
-		return ret;
+	if (i915.enable_execlists) {
+		ret = ring->emit_request(ringbuf);
+		if (ret)
+			return ret;
+	} else {
+		ret = ring->add_request(ring);
+		if (ret)
+			return ret;
+	}
 
 	request->seqno = intel_ring_get_seqno(ring);
 	request->ring = ring;
@@ -2372,12 +2468,14 @@ int __i915_add_request(struct intel_engine_cs *ring,
 	 */
 	request->batch_obj = obj;
 
-	/* Hold a reference to the current context so that we can inspect
-	 * it later in case a hangcheck error event fires.
-	 */
-	request->ctx = ring->last_context;
-	if (request->ctx)
-		i915_gem_context_reference(request->ctx);
+	if (!i915.enable_execlists) {
+		/* Hold a reference to the current context so that we can inspect
+		 * it later in case a hangcheck error event fires.
+		 */
+		request->ctx = ring->last_context;
+		if (request->ctx)
+			i915_gem_context_reference(request->ctx);
+	}
 
 	request->emitted_jiffies = jiffies;
 	list_add_tail(&request->list, &ring->request_list);
@@ -2397,15 +2495,13 @@ int __i915_add_request(struct intel_engine_cs *ring,
 	ring->outstanding_lazy_seqno = 0;
 	ring->preallocated_lazy_request = NULL;
 
-	if (!dev_priv->ums.mm_suspended) {
-		i915_queue_hangcheck(ring->dev);
+	i915_queue_hangcheck(ring->dev);
 
-		cancel_delayed_work_sync(&dev_priv->mm.idle_work);
-		queue_delayed_work(dev_priv->wq,
-				   &dev_priv->mm.retire_work,
-				   round_jiffies_up_relative(HZ));
-		intel_mark_busy(dev_priv->dev);
-	}
+	cancel_delayed_work_sync(&dev_priv->mm.idle_work);
+	queue_delayed_work(dev_priv->wq,
+			   &dev_priv->mm.retire_work,
+			   round_jiffies_up_relative(HZ));
+	intel_mark_busy(dev_priv->dev);
 
 	if (out_seqno)
 		*out_seqno = request->seqno;
@@ -2472,12 +2568,20 @@ static void i915_set_reset_status(struct drm_i915_private *dev_priv,
 
 static void i915_gem_free_request(struct drm_i915_gem_request *request)
 {
+	struct intel_context *ctx = request->ctx;
+
 	list_del(&request->list);
 	i915_gem_request_remove_from_client(request);
 
-	if (request->ctx)
-		i915_gem_context_unreference(request->ctx);
+	if (ctx) {
+		if (i915.enable_execlists) {
+			struct intel_engine_cs *ring = request->ring;
 
+			if (ctx != ring->default_context)
+				intel_lr_context_unpin(ring, ctx);
+		}
+		i915_gem_context_unreference(ctx);
+	}
 	kfree(request);
 }
 
@@ -2532,6 +2636,23 @@ static void i915_gem_reset_ring_cleanup(struct drm_i915_private *dev_priv,
 	}
 
 	/*
+	 * Clear the execlists queue up before freeing the requests, as those
+	 * are the ones that keep the context and ringbuffer backing objects
+	 * pinned in place.
+	 */
+	while (!list_empty(&ring->execlist_queue)) {
+		struct intel_ctx_submit_request *submit_req;
+
+		submit_req = list_first_entry(&ring->execlist_queue,
+				struct intel_ctx_submit_request,
+				execlist_link);
+		list_del(&submit_req->execlist_link);
+		intel_runtime_pm_put(dev_priv);
+		i915_gem_context_unreference(submit_req->ctx);
+		kfree(submit_req);
+	}
+
+	/*
 	 * We must free the requests after all the corresponding objects have
 	 * been moved off active lists. Which is the same order as the normal
 	 * retire_requests function does. This is important if object hold
@@ -2632,6 +2753,7 @@ i915_gem_retire_requests_ring(struct intel_engine_cs *ring)
 
 	while (!list_empty(&ring->request_list)) {
 		struct drm_i915_gem_request *request;
+		struct intel_ringbuffer *ringbuf;
 
 		request = list_first_entry(&ring->request_list,
 					   struct drm_i915_gem_request,
@@ -2641,12 +2763,24 @@ i915_gem_retire_requests_ring(struct intel_engine_cs *ring)
 			break;
 
 		trace_i915_gem_request_retire(ring, request->seqno);
+
+		/* This is one of the few common intersection points
+		 * between legacy ringbuffer submission and execlists:
+		 * we need to tell them apart in order to find the correct
+		 * ringbuffer to which the request belongs to.
+		 */
+		if (i915.enable_execlists) {
+			struct intel_context *ctx = request->ctx;
+			ringbuf = ctx->engine[ring->id].ringbuf;
+		} else
+			ringbuf = ring->buffer;
+
 		/* We know the GPU must have read the request to have
 		 * sent us the seqno + interrupt, so use the position
 		 * of tail of the request to update the last known position
 		 * of the GPU head.
 		 */
-		ring->buffer->last_retired_head = request->tail;
+		ringbuf->last_retired_head = request->tail;
 
 		i915_gem_free_request(request);
 	}
@@ -2671,6 +2805,15 @@ i915_gem_retire_requests(struct drm_device *dev)
 	for_each_ring(ring, dev_priv, i) {
 		i915_gem_retire_requests_ring(ring);
 		idle &= list_empty(&ring->request_list);
+		if (i915.enable_execlists) {
+			unsigned long flags;
+
+			spin_lock_irqsave(&ring->execlist_lock, flags);
+			idle &= list_empty(&ring->execlist_queue);
+			spin_unlock_irqrestore(&ring->execlist_lock, flags);
+
+			intel_execlists_retire_requests(ring);
+		}
 	}
 
 	if (idle)
@@ -2763,6 +2906,9 @@ i915_gem_wait_ioctl(struct drm_device *dev, void *data, struct drm_file *file)
 	u32 seqno = 0;
 	int ret = 0;
 
+	if (args->flags != 0)
+		return -EINVAL;
+
 	ret = i915_mutex_lock_interruptible(dev);
 	if (ret)
 		return ret;
@@ -2798,8 +2944,8 @@ i915_gem_wait_ioctl(struct drm_device *dev, void *data, struct drm_file *file)
 	reset_counter = atomic_read(&dev_priv->gpu_error.reset_counter);
 	mutex_unlock(&dev->struct_mutex);
 
-	return __wait_seqno(ring, seqno, reset_counter, true, &args->timeout_ns,
-			    file->driver_priv);
+	return __i915_wait_seqno(ring, seqno, reset_counter, true,
+				 &args->timeout_ns, file->driver_priv);
 
 out:
 	drm_gem_object_unreference(&obj->base);
@@ -2908,6 +3054,9 @@ int i915_vma_unbind(struct i915_vma *vma)
 	 * cause memory corruption through use-after-free.
 	 */
 
+	/* Throw away the active reference before moving to the unbound list */
+	i915_gem_object_retire(obj);
+
 	if (i915_is_ggtt(vma->vm)) {
 		i915_gem_object_finish_gtt(obj);
 
@@ -2922,9 +3071,8 @@ int i915_vma_unbind(struct i915_vma *vma)
 	vma->unbind_vma(vma);
 
 	list_del_init(&vma->mm_list);
-	/* Avoid an unnecessary call to unbind on rebind. */
 	if (i915_is_ggtt(vma->vm))
-		obj->map_and_fenceable = true;
+		obj->map_and_fenceable = false;
 
 	drm_mm_remove_node(&vma->node);
 	i915_gem_vma_destroy(vma);
@@ -2953,9 +3101,11 @@ int i915_gpu_idle(struct drm_device *dev)
 
 	/* Flush everything onto the inactive list. */
 	for_each_ring(ring, dev_priv, i) {
-		ret = i915_switch_context(ring, ring->default_context);
-		if (ret)
-			return ret;
+		if (!i915.enable_execlists) {
+			ret = i915_switch_context(ring, ring->default_context);
+			if (ret)
+				return ret;
+		}
 
 		ret = intel_ring_idle(ring);
 		if (ret)
@@ -3114,6 +3264,7 @@ static void i915_gem_write_fence(struct drm_device *dev, int reg,
 	     obj->stride, obj->tiling_mode);
 
 	switch (INTEL_INFO(dev)->gen) {
+	case 9:
 	case 8:
 	case 7:
 	case 6:
@@ -3169,7 +3320,6 @@ i915_gem_object_wait_fence(struct drm_i915_gem_object *obj)
 		obj->last_fenced_seqno = 0;
 	}
 
-	obj->fenced_gpu_access = false;
 	return 0;
 }
 
@@ -3276,6 +3426,9 @@ i915_gem_object_get_fence(struct drm_i915_gem_object *obj)
 			return 0;
 		}
 	} else if (enable) {
+		if (WARN_ON(!obj->map_and_fenceable))
+			return -EINVAL;
+
 		reg = i915_find_fence_reg(dev);
 		if (IS_ERR(reg))
 			return PTR_ERR(reg);
@@ -3297,17 +3450,20 @@ i915_gem_object_get_fence(struct drm_i915_gem_object *obj)
 	return 0;
 }
 
-static bool i915_gem_valid_gtt_space(struct drm_device *dev,
-				     struct drm_mm_node *gtt_space,
+static bool i915_gem_valid_gtt_space(struct i915_vma *vma,
 				     unsigned long cache_level)
 {
+	struct drm_mm_node *gtt_space = &vma->node;
 	struct drm_mm_node *other;
 
-	/* On non-LLC machines we have to be careful when putting differing
-	 * types of snoopable memory together to avoid the prefetcher
-	 * crossing memory domains and dying.
+	/*
+	 * On some machines we have to be careful when putting differing types
+	 * of snoopable memory together to avoid the prefetcher crossing memory
+	 * domains and dying. During vm initialisation, we decide whether or not
+	 * these constraints apply and set the drm_mm.color_adjust
+	 * appropriately.
 	 */
-	if (HAS_LLC(dev))
+	if (vma->vm->mm.color_adjust == NULL)
 		return true;
 
 	if (!drm_mm_node_allocated(gtt_space))
@@ -3327,46 +3483,6 @@ static bool i915_gem_valid_gtt_space(struct drm_device *dev,
 	return true;
 }
 
-static void i915_gem_verify_gtt(struct drm_device *dev)
-{
-#if WATCH_GTT
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct drm_i915_gem_object *obj;
-	int err = 0;
-
-	list_for_each_entry(obj, &dev_priv->mm.gtt_list, global_list) {
-		if (obj->gtt_space == NULL) {
-			printk(KERN_ERR "object found on GTT list with no space reserved\n");
-			err++;
-			continue;
-		}
-
-		if (obj->cache_level != obj->gtt_space->color) {
-			printk(KERN_ERR "object reserved space [%08lx, %08lx] with wrong color, cache_level=%x, color=%lx\n",
-			       i915_gem_obj_ggtt_offset(obj),
-			       i915_gem_obj_ggtt_offset(obj) + i915_gem_obj_ggtt_size(obj),
-			       obj->cache_level,
-			       obj->gtt_space->color);
-			err++;
-			continue;
-		}
-
-		if (!i915_gem_valid_gtt_space(dev,
-					      obj->gtt_space,
-					      obj->cache_level)) {
-			printk(KERN_ERR "invalid GTT space found at [%08lx, %08lx] - color=%x\n",
-			       i915_gem_obj_ggtt_offset(obj),
-			       i915_gem_obj_ggtt_offset(obj) + i915_gem_obj_ggtt_size(obj),
-			       obj->cache_level);
-			err++;
-			continue;
-		}
-	}
-
-	WARN_ON(err);
-#endif
-}
-
 /**
  * Finds free space in the GTT aperture and binds the object there.
  */
@@ -3445,8 +3561,7 @@ search_free:
 
 		goto err_free_vma;
 	}
-	if (WARN_ON(!i915_gem_valid_gtt_space(dev, &vma->node,
-					      obj->cache_level))) {
+	if (WARN_ON(!i915_gem_valid_gtt_space(vma, obj->cache_level))) {
 		ret = -EINVAL;
 		goto err_remove_node;
 	}
@@ -3458,25 +3573,10 @@ search_free:
 	list_move_tail(&obj->global_list, &dev_priv->mm.bound_list);
 	list_add_tail(&vma->mm_list, &vm->inactive_list);
 
-	if (i915_is_ggtt(vm)) {
-		bool mappable, fenceable;
-
-		fenceable = (vma->node.size == fence_size &&
-			     (vma->node.start & (fence_alignment - 1)) == 0);
-
-		mappable = (vma->node.start + obj->base.size <=
-			    dev_priv->gtt.mappable_end);
-
-		obj->map_and_fenceable = mappable && fenceable;
-	}
-
-	WARN_ON(flags & PIN_MAPPABLE && !obj->map_and_fenceable);
-
 	trace_i915_vma_bind(vma, flags);
 	vma->bind_vma(vma, obj->cache_level,
-		      flags & (PIN_MAPPABLE | PIN_GLOBAL) ? GLOBAL_BIND : 0);
+		      flags & PIN_GLOBAL ? GLOBAL_BIND : 0);
 
-	i915_gem_verify_gtt(dev);
 	return vma;
 
 err_remove_node:
@@ -3504,7 +3604,7 @@ i915_gem_clflush_object(struct drm_i915_gem_object *obj,
 	 * Stolen memory is always coherent with the GPU as it is explicitly
 	 * marked as wc by the system, or the system is cache-coherent.
 	 */
-	if (obj->stolen)
+	if (obj->stolen || obj->phys_handle)
 		return false;
 
 	/* If the GPU is snooping the contents of the CPU cache,
@@ -3586,11 +3686,12 @@ int
 i915_gem_object_set_to_gtt_domain(struct drm_i915_gem_object *obj, bool write)
 {
 	struct drm_i915_private *dev_priv = obj->base.dev->dev_private;
+	struct i915_vma *vma = i915_gem_obj_to_ggtt(obj);
 	uint32_t old_write_domain, old_read_domains;
 	int ret;
 
 	/* Not valid to be called on unbound objects. */
-	if (!i915_gem_obj_bound_any(obj))
+	if (vma == NULL)
 		return -EINVAL;
 
 	if (obj->base.write_domain == I915_GEM_DOMAIN_GTT)
@@ -3632,13 +3733,9 @@ i915_gem_object_set_to_gtt_domain(struct drm_i915_gem_object *obj, bool write)
 					    old_write_domain);
 
 	/* And bump the LRU for this access */
-	if (i915_gem_object_is_inactive(obj)) {
-		struct i915_vma *vma = i915_gem_obj_to_ggtt(obj);
-		if (vma)
-			list_move_tail(&vma->mm_list,
-				       &dev_priv->gtt.base.inactive_list);
-
-	}
+	if (i915_gem_object_is_inactive(obj))
+		list_move_tail(&vma->mm_list,
+			       &dev_priv->gtt.base.inactive_list);
 
 	return 0;
 }
@@ -3659,7 +3756,7 @@ int i915_gem_object_set_cache_level(struct drm_i915_gem_object *obj,
 	}
 
 	list_for_each_entry_safe(vma, next, &obj->vma_list, vma_link) {
-		if (!i915_gem_valid_gtt_space(dev, &vma->node, cache_level)) {
+		if (!i915_gem_valid_gtt_space(vma, cache_level)) {
 			ret = i915_vma_unbind(vma);
 			if (ret)
 				return ret;
@@ -3686,7 +3783,7 @@ int i915_gem_object_set_cache_level(struct drm_i915_gem_object *obj,
 		list_for_each_entry(vma, &obj->vma_list, vma_link)
 			if (drm_mm_node_allocated(&vma->node))
 				vma->bind_vma(vma, cache_level,
-					      obj->has_global_gtt_mapping ? GLOBAL_BIND : 0);
+						vma->bound & GLOBAL_BIND);
 	}
 
 	list_for_each_entry(vma, &obj->vma_list, vma_link)
@@ -3716,7 +3813,6 @@ int i915_gem_object_set_cache_level(struct drm_i915_gem_object *obj,
 						    old_write_domain);
 	}
 
-	i915_gem_verify_gtt(dev);
 	return 0;
 }
 
@@ -3802,9 +3898,6 @@ static bool is_pin_display(struct drm_i915_gem_object *obj)
 {
 	struct i915_vma *vma;
 
-	if (list_empty(&obj->vma_list))
-		return false;
-
 	vma = i915_gem_obj_to_ggtt(obj);
 	if (!vma)
 		return false;
@@ -4017,7 +4110,7 @@ i915_gem_ring_throttle(struct drm_device *dev, struct drm_file *file)
 	if (seqno == 0)
 		return 0;
 
-	ret = __wait_seqno(ring, seqno, reset_counter, true, NULL, NULL);
+	ret = __i915_wait_seqno(ring, seqno, reset_counter, true, NULL, NULL);
 	if (ret == 0)
 		queue_delayed_work(dev_priv->wq, &dev_priv->mm.retire_work, 0);
 
@@ -4051,6 +4144,7 @@ i915_gem_object_pin(struct drm_i915_gem_object *obj,
 {
 	struct drm_i915_private *dev_priv = obj->base.dev->dev_private;
 	struct i915_vma *vma;
+	unsigned bound;
 	int ret;
 
 	if (WARN_ON(vm == &dev_priv->mm.aliasing_ppgtt->base))
@@ -4059,6 +4153,9 @@ i915_gem_object_pin(struct drm_i915_gem_object *obj,
 	if (WARN_ON(flags & (PIN_GLOBAL | PIN_MAPPABLE) && !i915_is_ggtt(vm)))
 		return -EINVAL;
 
+	if (WARN_ON((flags & (PIN_MAPPABLE | PIN_GLOBAL)) == PIN_MAPPABLE))
+		return -EINVAL;
+
 	vma = i915_gem_obj_to_vma(obj, vm);
 	if (vma) {
 		if (WARN_ON(vma->pin_count == DRM_I915_GEM_OBJECT_MAX_PIN_COUNT))
@@ -4080,15 +4177,39 @@ i915_gem_object_pin(struct drm_i915_gem_object *obj,
 		}
 	}
 
+	bound = vma ? vma->bound : 0;
 	if (vma == NULL || !drm_mm_node_allocated(&vma->node)) {
 		vma = i915_gem_object_bind_to_vm(obj, vm, alignment, flags);
 		if (IS_ERR(vma))
 			return PTR_ERR(vma);
 	}
 
-	if (flags & PIN_GLOBAL && !obj->has_global_gtt_mapping)
+	if (flags & PIN_GLOBAL && !(vma->bound & GLOBAL_BIND))
 		vma->bind_vma(vma, obj->cache_level, GLOBAL_BIND);
 
+	if ((bound ^ vma->bound) & GLOBAL_BIND) {
+		bool mappable, fenceable;
+		u32 fence_size, fence_alignment;
+
+		fence_size = i915_gem_get_gtt_size(obj->base.dev,
+						   obj->base.size,
+						   obj->tiling_mode);
+		fence_alignment = i915_gem_get_gtt_alignment(obj->base.dev,
+							     obj->base.size,
+							     obj->tiling_mode,
+							     true);
+
+		fenceable = (vma->node.size == fence_size &&
+			     (vma->node.start & (fence_alignment - 1)) == 0);
+
+		mappable = (vma->node.start + obj->base.size <=
+			    dev_priv->gtt.mappable_end);
+
+		obj->map_and_fenceable = mappable && fenceable;
+	}
+
+	WARN_ON(flags & PIN_MAPPABLE && !obj->map_and_fenceable);
+
 	vma->pin_count++;
 	if (flags & PIN_MAPPABLE)
 		obj->pin_mappable |= true;
@@ -4143,7 +4264,7 @@ i915_gem_pin_ioctl(struct drm_device *dev, void *data,
 	struct drm_i915_gem_object *obj;
 	int ret;
 
-	if (INTEL_INFO(dev)->gen >= 6)
+	if (drm_core_check_feature(dev, DRIVER_MODESET))
 		return -ENODEV;
 
 	ret = i915_mutex_lock_interruptible(dev);
@@ -4199,6 +4320,9 @@ i915_gem_unpin_ioctl(struct drm_device *dev, void *data,
 	struct drm_i915_gem_object *obj;
 	int ret;
 
+	if (drm_core_check_feature(dev, DRIVER_MODESET))
+		return -ENODEV;
+
 	ret = i915_mutex_lock_interruptible(dev);
 	if (ret)
 		return ret;
@@ -4276,6 +4400,7 @@ int
 i915_gem_madvise_ioctl(struct drm_device *dev, void *data,
 		       struct drm_file *file_priv)
 {
+	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct drm_i915_gem_madvise *args = data;
 	struct drm_i915_gem_object *obj;
 	int ret;
@@ -4303,6 +4428,15 @@ i915_gem_madvise_ioctl(struct drm_device *dev, void *data,
 		goto out;
 	}
 
+	if (obj->pages &&
+	    obj->tiling_mode != I915_TILING_NONE &&
+	    dev_priv->quirks & QUIRK_PIN_SWIZZLED_PAGES) {
+		if (obj->madv == I915_MADV_WILLNEED)
+			i915_gem_object_unpin_pages(obj);
+		if (args->madv == I915_MADV_WILLNEED)
+			i915_gem_object_pin_pages(obj);
+	}
+
 	if (obj->madv != __I915_MADV_PURGED)
 		obj->madv = args->madv;
 
@@ -4331,8 +4465,6 @@ void i915_gem_object_init(struct drm_i915_gem_object *obj,
 
 	obj->fence_reg = I915_FENCE_REG_NONE;
 	obj->madv = I915_MADV_WILLNEED;
-	/* Avoid an unnecessary call to unbind on the first bind. */
-	obj->map_and_fenceable = true;
 
 	i915_gem_info_add_obj(obj->base.dev->dev_private, obj->base.size);
 }
@@ -4447,8 +4579,6 @@ void i915_gem_free_object(struct drm_gem_object *gem_obj)
 		}
 	}
 
-	i915_gem_object_detach_phys(obj);
-
 	/* Stolen objects don't hold a ref, but do hold pin count. Fix that up
 	 * before progressing. */
 	if (obj->stolen)
@@ -4456,6 +4586,11 @@ void i915_gem_free_object(struct drm_gem_object *gem_obj)
 
 	WARN_ON(obj->frontbuffer_bits);
 
+	if (obj->pages && obj->madv == I915_MADV_WILLNEED &&
+	    dev_priv->quirks & QUIRK_PIN_SWIZZLED_PAGES &&
+	    obj->tiling_mode != I915_TILING_NONE)
+		i915_gem_object_unpin_pages(obj);
+
 	if (WARN_ON(obj->pages_pin_count))
 		obj->pages_pin_count = 0;
 	if (discard_backing_storage(obj))
@@ -4493,12 +4628,18 @@ struct i915_vma *i915_gem_obj_to_vma(struct drm_i915_gem_object *obj,
 
 void i915_gem_vma_destroy(struct i915_vma *vma)
 {
+	struct i915_address_space *vm = NULL;
 	WARN_ON(vma->node.allocated);
 
 	/* Keep the vma as a placeholder in the execbuffer reservation lists */
 	if (!list_empty(&vma->exec_list))
 		return;
 
+	vm = vma->vm;
+
+	if (!i915_is_ggtt(vm))
+		i915_ppgtt_put(i915_vm_to_ppgtt(vm));
+
 	list_del(&vma->vma_link);
 
 	kfree(vma);
@@ -4512,7 +4653,7 @@ i915_gem_stop_ringbuffers(struct drm_device *dev)
 	int i;
 
 	for_each_ring(ring, dev_priv, i)
-		intel_stop_ring_buffer(ring);
+		dev_priv->gt.stop_ring(ring);
 }
 
 int
@@ -4522,9 +4663,6 @@ i915_gem_suspend(struct drm_device *dev)
 	int ret = 0;
 
 	mutex_lock(&dev->struct_mutex);
-	if (dev_priv->ums.mm_suspended)
-		goto err;
-
 	ret = i915_gpu_idle(dev);
 	if (ret)
 		goto err;
@@ -4535,15 +4673,7 @@ i915_gem_suspend(struct drm_device *dev)
 	if (!drm_core_check_feature(dev, DRIVER_MODESET))
 		i915_gem_evict_everything(dev);
 
-	i915_kernel_lost_context(dev);
 	i915_gem_stop_ringbuffers(dev);
-
-	/* Hack!  Don't let anybody do execbuf while we don't control the chip.
-	 * We need to replace this with a semaphore, or something.
-	 * And not confound ums.mm_suspended!
-	 */
-	dev_priv->ums.mm_suspended = !drm_core_check_feature(dev,
-							     DRIVER_MODESET);
 	mutex_unlock(&dev->struct_mutex);
 
 	del_timer_sync(&dev_priv->gpu_error.hangcheck_timer);
@@ -4629,11 +4759,46 @@ intel_enable_blt(struct drm_device *dev)
 	return true;
 }
 
-static int i915_gem_init_rings(struct drm_device *dev)
+static void init_unused_ring(struct drm_device *dev, u32 base)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+
+	I915_WRITE(RING_CTL(base), 0);
+	I915_WRITE(RING_HEAD(base), 0);
+	I915_WRITE(RING_TAIL(base), 0);
+	I915_WRITE(RING_START(base), 0);
+}
+
+static void init_unused_rings(struct drm_device *dev)
+{
+	if (IS_I830(dev)) {
+		init_unused_ring(dev, PRB1_BASE);
+		init_unused_ring(dev, SRB0_BASE);
+		init_unused_ring(dev, SRB1_BASE);
+		init_unused_ring(dev, SRB2_BASE);
+		init_unused_ring(dev, SRB3_BASE);
+	} else if (IS_GEN2(dev)) {
+		init_unused_ring(dev, SRB0_BASE);
+		init_unused_ring(dev, SRB1_BASE);
+	} else if (IS_GEN3(dev)) {
+		init_unused_ring(dev, PRB1_BASE);
+		init_unused_ring(dev, PRB2_BASE);
+	}
+}
+
+int i915_gem_init_rings(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	int ret;
 
+	/*
+	 * At least 830 can leave some of the unused rings
+	 * "active" (ie. head != tail) after resume which
+	 * will prevent c3 entry. Makes sure all unused rings
+	 * are totally idle.
+	 */
+	init_unused_rings(dev);
+
 	ret = intel_init_render_ring_buffer(dev);
 	if (ret)
 		return ret;
@@ -4712,7 +4877,7 @@ i915_gem_init_hw(struct drm_device *dev)
 
 	i915_gem_init_swizzling(dev);
 
-	ret = i915_gem_init_rings(dev);
+	ret = dev_priv->gt.init_rings(dev);
 	if (ret)
 		return ret;
 
@@ -4730,6 +4895,14 @@ i915_gem_init_hw(struct drm_device *dev)
 	if (ret && ret != -EIO) {
 		DRM_ERROR("Context enable failed %d\n", ret);
 		i915_gem_cleanup_ringbuffer(dev);
+
+		return ret;
+	}
+
+	ret = i915_ppgtt_init_hw(dev);
+	if (ret && ret != -EIO) {
+		DRM_ERROR("PPGTT enable failed %d\n", ret);
+		i915_gem_cleanup_ringbuffer(dev);
 	}
 
 	return ret;
@@ -4740,6 +4913,9 @@ int i915_gem_init(struct drm_device *dev)
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	int ret;
 
+	i915.enable_execlists = intel_sanitize_enable_execlists(dev,
+			i915.enable_execlists);
+
 	mutex_lock(&dev->struct_mutex);
 
 	if (IS_VALLEYVIEW(dev)) {
@@ -4750,7 +4926,24 @@ int i915_gem_init(struct drm_device *dev)
 			DRM_DEBUG_DRIVER("allow wake ack timed out\n");
 	}
 
-	i915_gem_init_userptr(dev);
+	if (!i915.enable_execlists) {
+		dev_priv->gt.do_execbuf = i915_gem_ringbuffer_submission;
+		dev_priv->gt.init_rings = i915_gem_init_rings;
+		dev_priv->gt.cleanup_ring = intel_cleanup_ring_buffer;
+		dev_priv->gt.stop_ring = intel_stop_ring_buffer;
+	} else {
+		dev_priv->gt.do_execbuf = intel_execlists_submission;
+		dev_priv->gt.init_rings = intel_logical_rings_init;
+		dev_priv->gt.cleanup_ring = intel_logical_ring_cleanup;
+		dev_priv->gt.stop_ring = intel_logical_ring_stop;
+	}
+
+	ret = i915_gem_init_userptr(dev);
+	if (ret) {
+		mutex_unlock(&dev->struct_mutex);
+		return ret;
+	}
+
 	i915_gem_init_global_gtt(dev);
 
 	ret = i915_gem_context_init(dev);
@@ -4771,9 +4964,6 @@ int i915_gem_init(struct drm_device *dev)
 	}
 	mutex_unlock(&dev->struct_mutex);
 
-	/* Allow hardware batchbuffers unless told otherwise, but not for KMS. */
-	if (!drm_core_check_feature(dev, DRIVER_MODESET))
-		dev_priv->dri1.allow_batchbuffer = 1;
 	return ret;
 }
 
@@ -4785,75 +4975,7 @@ i915_gem_cleanup_ringbuffer(struct drm_device *dev)
 	int i;
 
 	for_each_ring(ring, dev_priv, i)
-		intel_cleanup_ring_buffer(ring);
-}
-
-int
-i915_gem_entervt_ioctl(struct drm_device *dev, void *data,
-		       struct drm_file *file_priv)
-{
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	int ret;
-
-	if (drm_core_check_feature(dev, DRIVER_MODESET))
-		return 0;
-
-	if (i915_reset_in_progress(&dev_priv->gpu_error)) {
-		DRM_ERROR("Reenabling wedged hardware, good luck\n");
-		atomic_set(&dev_priv->gpu_error.reset_counter, 0);
-	}
-
-	mutex_lock(&dev->struct_mutex);
-	dev_priv->ums.mm_suspended = 0;
-
-	ret = i915_gem_init_hw(dev);
-	if (ret != 0) {
-		mutex_unlock(&dev->struct_mutex);
-		return ret;
-	}
-
-	BUG_ON(!list_empty(&dev_priv->gtt.base.active_list));
-
-	ret = drm_irq_install(dev, dev->pdev->irq);
-	if (ret)
-		goto cleanup_ringbuffer;
-	mutex_unlock(&dev->struct_mutex);
-
-	return 0;
-
-cleanup_ringbuffer:
-	i915_gem_cleanup_ringbuffer(dev);
-	dev_priv->ums.mm_suspended = 1;
-	mutex_unlock(&dev->struct_mutex);
-
-	return ret;
-}
-
-int
-i915_gem_leavevt_ioctl(struct drm_device *dev, void *data,
-		       struct drm_file *file_priv)
-{
-	if (drm_core_check_feature(dev, DRIVER_MODESET))
-		return 0;
-
-	mutex_lock(&dev->struct_mutex);
-	drm_irq_uninstall(dev);
-	mutex_unlock(&dev->struct_mutex);
-
-	return i915_gem_suspend(dev);
-}
-
-void
-i915_gem_lastclose(struct drm_device *dev)
-{
-	int ret;
-
-	if (drm_core_check_feature(dev, DRIVER_MODESET))
-		return;
-
-	ret = i915_gem_suspend(dev);
-	if (ret)
-		DRM_ERROR("failed to idle hardware: %d\n", ret);
+		dev_priv->gt.cleanup_ring(ring);
 }
 
 static void
@@ -5002,6 +5124,15 @@ int i915_gem_open(struct drm_device *dev, struct drm_file *file)
 	return ret;
 }
 
+/**
+ * i915_gem_track_fb - update frontbuffer tracking
+ * old: current GEM buffer for the frontbuffer slots
+ * new: new GEM buffer for the frontbuffer slots
+ * frontbuffer_bits: bitmask of frontbuffer slots
+ *
+ * This updates the frontbuffer tracking bits @frontbuffer_bits by clearing them
+ * from @old and setting them in @new. Both @old and @new can be NULL.
+ */
 void i915_gem_track_fb(struct drm_i915_gem_object *old,
 		       struct drm_i915_gem_object *new,
 		       unsigned frontbuffer_bits)
@@ -5097,9 +5228,7 @@ unsigned long i915_gem_obj_offset(struct drm_i915_gem_object *o,
 	struct drm_i915_private *dev_priv = o->base.dev->dev_private;
 	struct i915_vma *vma;
 
-	if (!dev_priv->mm.aliasing_ppgtt ||
-	    vm == &dev_priv->mm.aliasing_ppgtt->base)
-		vm = &dev_priv->gtt.base;
+	WARN_ON(vm == &dev_priv->mm.aliasing_ppgtt->base);
 
 	list_for_each_entry(vma, &o->vma_list, vma_link) {
 		if (vma->vm == vm)
@@ -5140,9 +5269,7 @@ unsigned long i915_gem_obj_size(struct drm_i915_gem_object *o,
 	struct drm_i915_private *dev_priv = o->base.dev->dev_private;
 	struct i915_vma *vma;
 
-	if (!dev_priv->mm.aliasing_ppgtt ||
-	    vm == &dev_priv->mm.aliasing_ppgtt->base)
-		vm = &dev_priv->gtt.base;
+	WARN_ON(vm == &dev_priv->mm.aliasing_ppgtt->base);
 
 	BUG_ON(list_empty(&o->vma_list));
 
@@ -5165,11 +5292,16 @@ i915_gem_shrinker_scan(struct shrinker *shrinker, struct shrink_control *sc)
 	if (!i915_gem_shrinker_lock(dev, &unlock))
 		return SHRINK_STOP;
 
-	freed = i915_gem_purge(dev_priv, sc->nr_to_scan);
+	freed = i915_gem_shrink(dev_priv,
+				sc->nr_to_scan,
+				I915_SHRINK_BOUND |
+				I915_SHRINK_UNBOUND |
+				I915_SHRINK_PURGEABLE);
 	if (freed < sc->nr_to_scan)
-		freed += __i915_gem_shrink(dev_priv,
-					   sc->nr_to_scan - freed,
-					   false);
+		freed += i915_gem_shrink(dev_priv,
+					 sc->nr_to_scan - freed,
+					 I915_SHRINK_BOUND |
+					 I915_SHRINK_UNBOUND);
 	if (unlock)
 		mutex_unlock(&dev->struct_mutex);
 
@@ -5184,7 +5316,7 @@ i915_gem_shrinker_oom(struct notifier_block *nb, unsigned long event, void *ptr)
 	struct drm_device *dev = dev_priv->dev;
 	struct drm_i915_gem_object *obj;
 	unsigned long timeout = msecs_to_jiffies(5000) + 1;
-	unsigned long pinned, bound, unbound, freed;
+	unsigned long pinned, bound, unbound, freed_pages;
 	bool was_interruptible;
 	bool unlock;
 
@@ -5201,7 +5333,7 @@ i915_gem_shrinker_oom(struct notifier_block *nb, unsigned long event, void *ptr)
 	was_interruptible = dev_priv->mm.interruptible;
 	dev_priv->mm.interruptible = false;
 
-	freed = i915_gem_shrink_all(dev_priv);
+	freed_pages = i915_gem_shrink_all(dev_priv);
 
 	dev_priv->mm.interruptible = was_interruptible;
 
@@ -5232,14 +5364,15 @@ i915_gem_shrinker_oom(struct notifier_block *nb, unsigned long event, void *ptr)
 	if (unlock)
 		mutex_unlock(&dev->struct_mutex);
 
-	pr_info("Purging GPU memory, %lu bytes freed, %lu bytes still pinned.\n",
-		freed, pinned);
+	if (freed_pages || unbound || bound)
+		pr_info("Purging GPU memory, %lu bytes freed, %lu bytes still pinned.\n",
+			freed_pages << PAGE_SHIFT, pinned);
 	if (unbound || bound)
 		pr_err("%lu and %lu bytes still available in the "
 		       "bound and unbound GPU page lists.\n",
 		       bound, unbound);
 
-	*(unsigned long *)ptr += freed;
+	*(unsigned long *)ptr += freed_pages;
 	return NOTIFY_DONE;
 }
 
@@ -5247,14 +5380,8 @@ struct i915_vma *i915_gem_obj_to_ggtt(struct drm_i915_gem_object *obj)
 {
 	struct i915_vma *vma;
 
-	/* This WARN has probably outlived its usefulness (callers already
-	 * WARN if they don't find the GGTT vma they expect). When removing,
-	 * remember to remove the pre-check in is_pin_display() as well */
-	if (WARN_ON(list_empty(&obj->vma_list)))
-		return NULL;
-
 	vma = list_first_entry(&obj->vma_list, typeof(*vma), vma_link);
-	if (vma->vm != obj_to_ggtt(obj))
+	if (vma->vm != i915_obj_to_ggtt(obj))
 		return NULL;
 
 	return vma;
diff --git a/drivers/gpu/drm/i915/i915_gem_context.c b/drivers/gpu/drm/i915/i915_gem_context.c
index 3b99390e467a..d011ec82ef1e 100644
--- a/drivers/gpu/drm/i915/i915_gem_context.c
+++ b/drivers/gpu/drm/i915/i915_gem_context.c
@@ -88,6 +88,7 @@
 #include <drm/drmP.h>
 #include <drm/i915_drm.h>
 #include "i915_drv.h"
+#include "i915_trace.h"
 
 /* This is a HW constraint. The value below is the largest known requirement
  * I've seen in a spec to date, and that was a workaround for a non-shipping
@@ -96,50 +97,6 @@
 #define GEN6_CONTEXT_ALIGN (64<<10)
 #define GEN7_CONTEXT_ALIGN 4096
 
-static void do_ppgtt_cleanup(struct i915_hw_ppgtt *ppgtt)
-{
-	struct drm_device *dev = ppgtt->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct i915_address_space *vm = &ppgtt->base;
-
-	if (ppgtt == dev_priv->mm.aliasing_ppgtt ||
-	    (list_empty(&vm->active_list) && list_empty(&vm->inactive_list))) {
-		ppgtt->base.cleanup(&ppgtt->base);
-		return;
-	}
-
-	/*
-	 * Make sure vmas are unbound before we take down the drm_mm
-	 *
-	 * FIXME: Proper refcounting should take care of this, this shouldn't be
-	 * needed at all.
-	 */
-	if (!list_empty(&vm->active_list)) {
-		struct i915_vma *vma;
-
-		list_for_each_entry(vma, &vm->active_list, mm_list)
-			if (WARN_ON(list_empty(&vma->vma_link) ||
-				    list_is_singular(&vma->vma_link)))
-				break;
-
-		i915_gem_evict_vm(&ppgtt->base, true);
-	} else {
-		i915_gem_retire_requests(dev);
-		i915_gem_evict_vm(&ppgtt->base, false);
-	}
-
-	ppgtt->base.cleanup(&ppgtt->base);
-}
-
-static void ppgtt_release(struct kref *kref)
-{
-	struct i915_hw_ppgtt *ppgtt =
-		container_of(kref, struct i915_hw_ppgtt, ref);
-
-	do_ppgtt_cleanup(ppgtt);
-	kfree(ppgtt);
-}
-
 static size_t get_context_alignment(struct drm_device *dev)
 {
 	if (IS_GEN6(dev))
@@ -179,24 +136,22 @@ static int get_context_size(struct drm_device *dev)
 void i915_gem_context_free(struct kref *ctx_ref)
 {
 	struct intel_context *ctx = container_of(ctx_ref,
-						   typeof(*ctx), ref);
-	struct i915_hw_ppgtt *ppgtt = NULL;
+						 typeof(*ctx), ref);
 
-	if (ctx->legacy_hw_ctx.rcs_state) {
-		/* We refcount even the aliasing PPGTT to keep the code symmetric */
-		if (USES_PPGTT(ctx->legacy_hw_ctx.rcs_state->base.dev))
-			ppgtt = ctx_to_ppgtt(ctx);
-	}
+	trace_i915_context_free(ctx);
+
+	if (i915.enable_execlists)
+		intel_lr_context_free(ctx);
+
+	i915_ppgtt_put(ctx->ppgtt);
 
-	if (ppgtt)
-		kref_put(&ppgtt->ref, ppgtt_release);
 	if (ctx->legacy_hw_ctx.rcs_state)
 		drm_gem_object_unreference(&ctx->legacy_hw_ctx.rcs_state->base);
 	list_del(&ctx->link);
 	kfree(ctx);
 }
 
-static struct drm_i915_gem_object *
+struct drm_i915_gem_object *
 i915_gem_alloc_context_obj(struct drm_device *dev, size_t size)
 {
 	struct drm_i915_gem_object *obj;
@@ -226,29 +181,9 @@ i915_gem_alloc_context_obj(struct drm_device *dev, size_t size)
 	return obj;
 }
 
-static struct i915_hw_ppgtt *
-create_vm_for_ctx(struct drm_device *dev, struct intel_context *ctx)
-{
-	struct i915_hw_ppgtt *ppgtt;
-	int ret;
-
-	ppgtt = kzalloc(sizeof(*ppgtt), GFP_KERNEL);
-	if (!ppgtt)
-		return ERR_PTR(-ENOMEM);
-
-	ret = i915_gem_init_ppgtt(dev, ppgtt);
-	if (ret) {
-		kfree(ppgtt);
-		return ERR_PTR(ret);
-	}
-
-	ppgtt->ctx = ctx;
-	return ppgtt;
-}
-
 static struct intel_context *
 __create_hw_context(struct drm_device *dev,
-		  struct drm_i915_file_private *file_priv)
+		    struct drm_i915_file_private *file_priv)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_context *ctx;
@@ -301,11 +236,9 @@ err_out:
  */
 static struct intel_context *
 i915_gem_create_context(struct drm_device *dev,
-			struct drm_i915_file_private *file_priv,
-			bool create_vm)
+			struct drm_i915_file_private *file_priv)
 {
 	const bool is_global_default_ctx = file_priv == NULL;
-	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_context *ctx;
 	int ret = 0;
 
@@ -331,34 +264,20 @@ i915_gem_create_context(struct drm_device *dev,
 		}
 	}
 
-	if (create_vm) {
-		struct i915_hw_ppgtt *ppgtt = create_vm_for_ctx(dev, ctx);
+	if (USES_FULL_PPGTT(dev)) {
+		struct i915_hw_ppgtt *ppgtt = i915_ppgtt_create(dev, file_priv);
 
 		if (IS_ERR_OR_NULL(ppgtt)) {
 			DRM_DEBUG_DRIVER("PPGTT setup failed (%ld)\n",
 					 PTR_ERR(ppgtt));
 			ret = PTR_ERR(ppgtt);
 			goto err_unpin;
-		} else
-			ctx->vm = &ppgtt->base;
-
-		/* This case is reserved for the global default context and
-		 * should only happen once. */
-		if (is_global_default_ctx) {
-			if (WARN_ON(dev_priv->mm.aliasing_ppgtt)) {
-				ret = -EEXIST;
-				goto err_unpin;
-			}
-
-			dev_priv->mm.aliasing_ppgtt = ppgtt;
 		}
-	} else if (USES_PPGTT(dev)) {
-		/* For platforms which only have aliasing PPGTT, we fake the
-		 * address space and refcounting. */
-		ctx->vm = &dev_priv->mm.aliasing_ppgtt->base;
-		kref_get(&dev_priv->mm.aliasing_ppgtt->ref);
-	} else
-		ctx->vm = &dev_priv->gtt.base;
+
+		ctx->ppgtt = ppgtt;
+	}
+
+	trace_i915_context_create(ctx);
 
 	return ctx;
 
@@ -375,34 +294,23 @@ void i915_gem_context_reset(struct drm_device *dev)
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	int i;
 
-	/* Prevent the hardware from restoring the last context (which hung) on
-	 * the next switch */
+	/* In execlists mode we will unreference the context when the execlist
+	 * queue is cleared and the requests destroyed.
+	 */
+	if (i915.enable_execlists)
+		return;
+
 	for (i = 0; i < I915_NUM_RINGS; i++) {
 		struct intel_engine_cs *ring = &dev_priv->ring[i];
-		struct intel_context *dctx = ring->default_context;
 		struct intel_context *lctx = ring->last_context;
 
-		/* Do a fake switch to the default context */
-		if (lctx == dctx)
-			continue;
-
-		if (!lctx)
-			continue;
+		if (lctx) {
+			if (lctx->legacy_hw_ctx.rcs_state && i == RCS)
+				i915_gem_object_ggtt_unpin(lctx->legacy_hw_ctx.rcs_state);
 
-		if (dctx->legacy_hw_ctx.rcs_state && i == RCS) {
-			WARN_ON(i915_gem_obj_ggtt_pin(dctx->legacy_hw_ctx.rcs_state,
-						      get_context_alignment(dev), 0));
-			/* Fake a finish/inactive */
-			dctx->legacy_hw_ctx.rcs_state->base.write_domain = 0;
-			dctx->legacy_hw_ctx.rcs_state->active = 0;
+			i915_gem_context_unreference(lctx);
+			ring->last_context = NULL;
 		}
-
-		if (lctx->legacy_hw_ctx.rcs_state && i == RCS)
-			i915_gem_object_ggtt_unpin(lctx->legacy_hw_ctx.rcs_state);
-
-		i915_gem_context_unreference(lctx);
-		i915_gem_context_reference(dctx);
-		ring->last_context = dctx;
 	}
 }
 
@@ -417,7 +325,11 @@ int i915_gem_context_init(struct drm_device *dev)
 	if (WARN_ON(dev_priv->ring[RCS].default_context))
 		return 0;
 
-	if (HAS_HW_CONTEXTS(dev)) {
+	if (i915.enable_execlists) {
+		/* NB: intentionally left blank. We will allocate our own
+		 * backing objects as we need them, thank you very much */
+		dev_priv->hw_context_size = 0;
+	} else if (HAS_HW_CONTEXTS(dev)) {
 		dev_priv->hw_context_size = round_up(get_context_size(dev), 4096);
 		if (dev_priv->hw_context_size > (1<<20)) {
 			DRM_DEBUG_DRIVER("Disabling HW Contexts; invalid size %d\n",
@@ -426,18 +338,23 @@ int i915_gem_context_init(struct drm_device *dev)
 		}
 	}
 
-	ctx = i915_gem_create_context(dev, NULL, USES_PPGTT(dev));
+	ctx = i915_gem_create_context(dev, NULL);
 	if (IS_ERR(ctx)) {
 		DRM_ERROR("Failed to create default global context (error %ld)\n",
 			  PTR_ERR(ctx));
 		return PTR_ERR(ctx);
 	}
 
-	/* NB: RCS will hold a ref for all rings */
-	for (i = 0; i < I915_NUM_RINGS; i++)
-		dev_priv->ring[i].default_context = ctx;
+	for (i = 0; i < I915_NUM_RINGS; i++) {
+		struct intel_engine_cs *ring = &dev_priv->ring[i];
+
+		/* NB: RCS will hold a ref for all rings */
+		ring->default_context = ctx;
+	}
 
-	DRM_DEBUG_DRIVER("%s context support initialized\n", dev_priv->hw_context_size ? "HW" : "fake");
+	DRM_DEBUG_DRIVER("%s context support initialized\n",
+			i915.enable_execlists ? "LR" :
+			dev_priv->hw_context_size ? "HW" : "fake");
 	return 0;
 }
 
@@ -489,19 +406,11 @@ int i915_gem_context_enable(struct drm_i915_private *dev_priv)
 	struct intel_engine_cs *ring;
 	int ret, i;
 
-	/* This is the only place the aliasing PPGTT gets enabled, which means
-	 * it has to happen before we bail on reset */
-	if (dev_priv->mm.aliasing_ppgtt) {
-		struct i915_hw_ppgtt *ppgtt = dev_priv->mm.aliasing_ppgtt;
-		ppgtt->enable(ppgtt);
-	}
+	BUG_ON(!dev_priv->ring[RCS].default_context);
 
-	/* FIXME: We should make this work, even in reset */
-	if (i915_reset_in_progress(&dev_priv->gpu_error))
+	if (i915.enable_execlists)
 		return 0;
 
-	BUG_ON(!dev_priv->ring[RCS].default_context);
-
 	for_each_ring(ring, dev_priv, i) {
 		ret = i915_switch_context(ring, ring->default_context);
 		if (ret)
@@ -527,7 +436,7 @@ int i915_gem_context_open(struct drm_device *dev, struct drm_file *file)
 	idr_init(&file_priv->context_idr);
 
 	mutex_lock(&dev->struct_mutex);
-	ctx = i915_gem_create_context(dev, file_priv, USES_FULL_PPGTT(dev));
+	ctx = i915_gem_create_context(dev, file_priv);
 	mutex_unlock(&dev->struct_mutex);
 
 	if (IS_ERR(ctx)) {
@@ -563,7 +472,13 @@ mi_set_context(struct intel_engine_cs *ring,
 	       struct intel_context *new_context,
 	       u32 hw_flags)
 {
-	int ret;
+	u32 flags = hw_flags | MI_MM_SPACE_GTT;
+	const int num_rings =
+		/* Use an extended w/a on ivb+ if signalling from other rings */
+		i915_semaphore_is_enabled(ring->dev) ?
+		hweight32(INTEL_INFO(ring->dev)->ring_mask) - 1 :
+		0;
+	int len, i, ret;
 
 	/* w/a: If Flush TLB Invalidation Mode is enabled, driver must do a TLB
 	 * invalidation prior to MI_SET_CONTEXT. On GEN6 we don't set the value
@@ -576,33 +491,61 @@ mi_set_context(struct intel_engine_cs *ring,
 			return ret;
 	}
 
-	ret = intel_ring_begin(ring, 6);
+	/* These flags are for resource streamer on HSW+ */
+	if (!IS_HASWELL(ring->dev) && INTEL_INFO(ring->dev)->gen < 8)
+		flags |= (MI_SAVE_EXT_STATE_EN | MI_RESTORE_EXT_STATE_EN);
+
+
+	len = 4;
+	if (INTEL_INFO(ring->dev)->gen >= 7)
+		len += 2 + (num_rings ? 4*num_rings + 2 : 0);
+
+	ret = intel_ring_begin(ring, len);
 	if (ret)
 		return ret;
 
 	/* WaProgramMiArbOnOffAroundMiSetContext:ivb,vlv,hsw,bdw,chv */
-	if (INTEL_INFO(ring->dev)->gen >= 7)
+	if (INTEL_INFO(ring->dev)->gen >= 7) {
 		intel_ring_emit(ring, MI_ARB_ON_OFF | MI_ARB_DISABLE);
-	else
-		intel_ring_emit(ring, MI_NOOP);
+		if (num_rings) {
+			struct intel_engine_cs *signaller;
+
+			intel_ring_emit(ring, MI_LOAD_REGISTER_IMM(num_rings));
+			for_each_ring(signaller, to_i915(ring->dev), i) {
+				if (signaller == ring)
+					continue;
+
+				intel_ring_emit(ring, RING_PSMI_CTL(signaller->mmio_base));
+				intel_ring_emit(ring, _MASKED_BIT_ENABLE(GEN6_PSMI_SLEEP_MSG_DISABLE));
+			}
+		}
+	}
 
 	intel_ring_emit(ring, MI_NOOP);
 	intel_ring_emit(ring, MI_SET_CONTEXT);
 	intel_ring_emit(ring, i915_gem_obj_ggtt_offset(new_context->legacy_hw_ctx.rcs_state) |
-			MI_MM_SPACE_GTT |
-			MI_SAVE_EXT_STATE_EN |
-			MI_RESTORE_EXT_STATE_EN |
-			hw_flags);
+			flags);
 	/*
 	 * w/a: MI_SET_CONTEXT must always be followed by MI_NOOP
 	 * WaMiSetContext_Hang:snb,ivb,vlv
 	 */
 	intel_ring_emit(ring, MI_NOOP);
 
-	if (INTEL_INFO(ring->dev)->gen >= 7)
+	if (INTEL_INFO(ring->dev)->gen >= 7) {
+		if (num_rings) {
+			struct intel_engine_cs *signaller;
+
+			intel_ring_emit(ring, MI_LOAD_REGISTER_IMM(num_rings));
+			for_each_ring(signaller, to_i915(ring->dev), i) {
+				if (signaller == ring)
+					continue;
+
+				intel_ring_emit(ring, RING_PSMI_CTL(signaller->mmio_base));
+				intel_ring_emit(ring, _MASKED_BIT_DISABLE(GEN6_PSMI_SLEEP_MSG_DISABLE));
+			}
+		}
 		intel_ring_emit(ring, MI_ARB_ON_OFF | MI_ARB_ENABLE);
-	else
-		intel_ring_emit(ring, MI_NOOP);
+	}
 
 	intel_ring_advance(ring);
 
@@ -614,9 +557,9 @@ static int do_switch(struct intel_engine_cs *ring,
 {
 	struct drm_i915_private *dev_priv = ring->dev->dev_private;
 	struct intel_context *from = ring->last_context;
-	struct i915_hw_ppgtt *ppgtt = ctx_to_ppgtt(to);
 	u32 hw_flags = 0;
 	bool uninitialized = false;
+	struct i915_vma *vma;
 	int ret, i;
 
 	if (from != NULL && ring == &dev_priv->ring[RCS]) {
@@ -642,8 +585,9 @@ static int do_switch(struct intel_engine_cs *ring,
 	 */
 	from = ring->last_context;
 
-	if (USES_FULL_PPGTT(ring->dev)) {
-		ret = ppgtt->switch_mm(ppgtt, ring, false);
+	if (to->ppgtt) {
+		trace_switch_mm(ring, to);
+		ret = to->ppgtt->switch_mm(to->ppgtt, ring);
 		if (ret)
 			goto unpin_out;
 	}
@@ -666,11 +610,10 @@ static int do_switch(struct intel_engine_cs *ring,
 	if (ret)
 		goto unpin_out;
 
-	if (!to->legacy_hw_ctx.rcs_state->has_global_gtt_mapping) {
-		struct i915_vma *vma = i915_gem_obj_to_vma(to->legacy_hw_ctx.rcs_state,
-							   &dev_priv->gtt.base);
-		vma->bind_vma(vma, to->legacy_hw_ctx.rcs_state->cache_level, GLOBAL_BIND);
-	}
+	vma = i915_gem_obj_to_ggtt(to->legacy_hw_ctx.rcs_state);
+	if (!(vma->bound & GLOBAL_BIND))
+		vma->bind_vma(vma, to->legacy_hw_ctx.rcs_state->cache_level,
+				GLOBAL_BIND);
 
 	if (!to->legacy_hw_ctx.initialized || i915_gem_context_is_default(to))
 		hw_flags |= MI_RESTORE_INHIBIT;
@@ -723,6 +666,12 @@ done:
 	ring->last_context = to;
 
 	if (uninitialized) {
+		if (ring->init_context) {
+			ret = ring->init_context(ring, to);
+			if (ret)
+				DRM_ERROR("ring init context: %d\n", ret);
+		}
+
 		ret = i915_gem_render_state_init(ring);
 		if (ret)
 			DRM_ERROR("init render state: %d\n", ret);
@@ -743,14 +692,19 @@ unpin_out:
  *
  * The context life cycle is simple. The context refcount is incremented and
  * decremented by 1 and create and destroy. If the context is in use by the GPU,
- * it will have a refoucnt > 1. This allows us to destroy the context abstract
+ * it will have a refcount > 1. This allows us to destroy the context abstract
  * object while letting the normal object tracking destroy the backing BO.
+ *
+ * This function should not be used in execlists mode.  Instead the context is
+ * switched by writing to the ELSP and requests keep a reference to their
+ * context.
  */
 int i915_switch_context(struct intel_engine_cs *ring,
 			struct intel_context *to)
 {
 	struct drm_i915_private *dev_priv = ring->dev->dev_private;
 
+	WARN_ON(i915.enable_execlists);
 	WARN_ON(!mutex_is_locked(&dev_priv->dev->struct_mutex));
 
 	if (to->legacy_hw_ctx.rcs_state == NULL) { /* We have the fake context */
@@ -766,9 +720,9 @@ int i915_switch_context(struct intel_engine_cs *ring,
 	return do_switch(ring, to);
 }
 
-static bool hw_context_enabled(struct drm_device *dev)
+static bool contexts_enabled(struct drm_device *dev)
 {
-	return to_i915(dev)->hw_context_size;
+	return i915.enable_execlists || to_i915(dev)->hw_context_size;
 }
 
 int i915_gem_context_create_ioctl(struct drm_device *dev, void *data,
@@ -779,14 +733,14 @@ int i915_gem_context_create_ioctl(struct drm_device *dev, void *data,
 	struct intel_context *ctx;
 	int ret;
 
-	if (!hw_context_enabled(dev))
+	if (!contexts_enabled(dev))
 		return -ENODEV;
 
 	ret = i915_mutex_lock_interruptible(dev);
 	if (ret)
 		return ret;
 
-	ctx = i915_gem_create_context(dev, file_priv, USES_FULL_PPGTT(dev));
+	ctx = i915_gem_create_context(dev, file_priv);
 	mutex_unlock(&dev->struct_mutex);
 	if (IS_ERR(ctx))
 		return PTR_ERR(ctx);
diff --git a/drivers/gpu/drm/i915/i915_gem_evict.c b/drivers/gpu/drm/i915/i915_gem_evict.c
index bbf4b12d842e..886ff2ee7a28 100644
--- a/drivers/gpu/drm/i915/i915_gem_evict.c
+++ b/drivers/gpu/drm/i915/i915_gem_evict.c
@@ -243,7 +243,7 @@ int
 i915_gem_evict_everything(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct i915_address_space *vm;
+	struct i915_address_space *vm, *v;
 	bool lists_empty = true;
 	int ret;
 
@@ -270,7 +270,7 @@ i915_gem_evict_everything(struct drm_device *dev)
 	i915_gem_retire_requests(dev);
 
 	/* Having flushed everything, unbind() should never raise an error */
-	list_for_each_entry(vm, &dev_priv->vm_list, global_link)
+	list_for_each_entry_safe(vm, v, &dev_priv->vm_list, global_link)
 		WARN_ON(i915_gem_evict_vm(vm, false));
 
 	return 0;
diff --git a/drivers/gpu/drm/i915/i915_gem_execbuffer.c b/drivers/gpu/drm/i915/i915_gem_execbuffer.c
index 60998fc4e5b2..11738316394a 100644
--- a/drivers/gpu/drm/i915/i915_gem_execbuffer.c
+++ b/drivers/gpu/drm/i915/i915_gem_execbuffer.c
@@ -35,6 +35,7 @@
 
 #define  __EXEC_OBJECT_HAS_PIN (1<<31)
 #define  __EXEC_OBJECT_HAS_FENCE (1<<30)
+#define  __EXEC_OBJECT_NEEDS_MAP (1<<29)
 #define  __EXEC_OBJECT_NEEDS_BIAS (1<<28)
 
 #define BATCH_OFFSET_BIAS (256*1024)
@@ -94,7 +95,6 @@ eb_lookup_vmas(struct eb_vmas *eb,
 	       struct i915_address_space *vm,
 	       struct drm_file *file)
 {
-	struct drm_i915_private *dev_priv = vm->dev->dev_private;
 	struct drm_i915_gem_object *obj;
 	struct list_head objects;
 	int i, ret;
@@ -129,20 +129,6 @@ eb_lookup_vmas(struct eb_vmas *eb,
 	i = 0;
 	while (!list_empty(&objects)) {
 		struct i915_vma *vma;
-		struct i915_address_space *bind_vm = vm;
-
-		if (exec[i].flags & EXEC_OBJECT_NEEDS_GTT &&
-		    USES_FULL_PPGTT(vm->dev)) {
-			ret = -EINVAL;
-			goto err;
-		}
-
-		/* If we have secure dispatch, or the userspace assures us that
-		 * they know what they're doing, use the GGTT VM.
-		 */
-		if (((args->flags & I915_EXEC_SECURE) &&
-		    (i == (args->buffer_count - 1))))
-			bind_vm = &dev_priv->gtt.base;
 
 		obj = list_first_entry(&objects,
 				       struct drm_i915_gem_object,
@@ -156,7 +142,7 @@ eb_lookup_vmas(struct eb_vmas *eb,
 		 * from the (obj, vm) we don't run the risk of creating
 		 * duplicated vmas for the same vm.
 		 */
-		vma = i915_gem_obj_lookup_or_create_vma(obj, bind_vm);
+		vma = i915_gem_obj_lookup_or_create_vma(obj, vm);
 		if (IS_ERR(vma)) {
 			DRM_DEBUG("Failed to lookup VMA\n");
 			ret = PTR_ERR(vma);
@@ -307,7 +293,7 @@ relocate_entry_gtt(struct drm_i915_gem_object *obj,
 	struct drm_device *dev = obj->base.dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	uint64_t delta = reloc->delta + target_offset;
-	uint32_t __iomem *reloc_entry;
+	uint64_t offset;
 	void __iomem *reloc_page;
 	int ret;
 
@@ -320,25 +306,24 @@ relocate_entry_gtt(struct drm_i915_gem_object *obj,
 		return ret;
 
 	/* Map the page containing the relocation we're going to perform.  */
-	reloc->offset += i915_gem_obj_ggtt_offset(obj);
+	offset = i915_gem_obj_ggtt_offset(obj);
+	offset += reloc->offset;
 	reloc_page = io_mapping_map_atomic_wc(dev_priv->gtt.mappable,
-			reloc->offset & PAGE_MASK);
-	reloc_entry = (uint32_t __iomem *)
-		(reloc_page + offset_in_page(reloc->offset));
-	iowrite32(lower_32_bits(delta), reloc_entry);
+					      offset & PAGE_MASK);
+	iowrite32(lower_32_bits(delta), reloc_page + offset_in_page(offset));
 
 	if (INTEL_INFO(dev)->gen >= 8) {
-		reloc_entry += 1;
+		offset += sizeof(uint32_t);
 
-		if (offset_in_page(reloc->offset + sizeof(uint32_t)) == 0) {
+		if (offset_in_page(offset) == 0) {
 			io_mapping_unmap_atomic(reloc_page);
-			reloc_page = io_mapping_map_atomic_wc(
-					dev_priv->gtt.mappable,
-					reloc->offset + sizeof(uint32_t));
-			reloc_entry = reloc_page;
+			reloc_page =
+				io_mapping_map_atomic_wc(dev_priv->gtt.mappable,
+							 offset);
 		}
 
-		iowrite32(upper_32_bits(delta), reloc_entry);
+		iowrite32(upper_32_bits(delta),
+			  reloc_page + offset_in_page(offset));
 	}
 
 	io_mapping_unmap_atomic(reloc_page);
@@ -372,12 +357,9 @@ i915_gem_execbuffer_relocate_entry(struct drm_i915_gem_object *obj,
 	 * through the ppgtt for non_secure batchbuffers. */
 	if (unlikely(IS_GEN6(dev) &&
 	    reloc->write_domain == I915_GEM_DOMAIN_INSTRUCTION &&
-	    !target_i915_obj->has_global_gtt_mapping)) {
-		struct i915_vma *vma =
-			list_first_entry(&target_i915_obj->vma_list,
-					 typeof(*vma), vma_link);
-		vma->bind_vma(vma, target_i915_obj->cache_level, GLOBAL_BIND);
-	}
+	    !(target_vma->bound & GLOBAL_BIND)))
+		target_vma->bind_vma(target_vma, target_i915_obj->cache_level,
+				GLOBAL_BIND);
 
 	/* Validate that the target is in a valid r/w GPU domain */
 	if (unlikely(reloc->write_domain & (reloc->write_domain - 1))) {
@@ -535,34 +517,18 @@ i915_gem_execbuffer_relocate(struct eb_vmas *eb)
 }
 
 static int
-need_reloc_mappable(struct i915_vma *vma)
-{
-	struct drm_i915_gem_exec_object2 *entry = vma->exec_entry;
-	return entry->relocation_count && !use_cpu_reloc(vma->obj) &&
-		i915_is_ggtt(vma->vm);
-}
-
-static int
 i915_gem_execbuffer_reserve_vma(struct i915_vma *vma,
 				struct intel_engine_cs *ring,
 				bool *need_reloc)
 {
 	struct drm_i915_gem_object *obj = vma->obj;
 	struct drm_i915_gem_exec_object2 *entry = vma->exec_entry;
-	bool has_fenced_gpu_access = INTEL_INFO(ring->dev)->gen < 4;
-	bool need_fence;
 	uint64_t flags;
 	int ret;
 
 	flags = 0;
-
-	need_fence =
-		has_fenced_gpu_access &&
-		entry->flags & EXEC_OBJECT_NEEDS_FENCE &&
-		obj->tiling_mode != I915_TILING_NONE;
-	if (need_fence || need_reloc_mappable(vma))
-		flags |= PIN_MAPPABLE;
-
+	if (entry->flags & __EXEC_OBJECT_NEEDS_MAP)
+		flags |= PIN_GLOBAL | PIN_MAPPABLE;
 	if (entry->flags & EXEC_OBJECT_NEEDS_GTT)
 		flags |= PIN_GLOBAL;
 	if (entry->flags & __EXEC_OBJECT_NEEDS_BIAS)
@@ -574,17 +540,13 @@ i915_gem_execbuffer_reserve_vma(struct i915_vma *vma,
 
 	entry->flags |= __EXEC_OBJECT_HAS_PIN;
 
-	if (has_fenced_gpu_access) {
-		if (entry->flags & EXEC_OBJECT_NEEDS_FENCE) {
-			ret = i915_gem_object_get_fence(obj);
-			if (ret)
-				return ret;
-
-			if (i915_gem_object_pin_fence(obj))
-				entry->flags |= __EXEC_OBJECT_HAS_FENCE;
+	if (entry->flags & EXEC_OBJECT_NEEDS_FENCE) {
+		ret = i915_gem_object_get_fence(obj);
+		if (ret)
+			return ret;
 
-			obj->pending_fenced_gpu_access = true;
-		}
+		if (i915_gem_object_pin_fence(obj))
+			entry->flags |= __EXEC_OBJECT_HAS_FENCE;
 	}
 
 	if (entry->offset != vma->node.start) {
@@ -601,26 +563,40 @@ i915_gem_execbuffer_reserve_vma(struct i915_vma *vma,
 }
 
 static bool
-eb_vma_misplaced(struct i915_vma *vma, bool has_fenced_gpu_access)
+need_reloc_mappable(struct i915_vma *vma)
 {
 	struct drm_i915_gem_exec_object2 *entry = vma->exec_entry;
-	struct drm_i915_gem_object *obj = vma->obj;
-	bool need_fence, need_mappable;
 
-	need_fence =
-		has_fenced_gpu_access &&
-		entry->flags & EXEC_OBJECT_NEEDS_FENCE &&
-		obj->tiling_mode != I915_TILING_NONE;
-	need_mappable = need_fence || need_reloc_mappable(vma);
+	if (entry->relocation_count == 0)
+		return false;
 
-	WARN_ON((need_mappable || need_fence) &&
+	if (!i915_is_ggtt(vma->vm))
+		return false;
+
+	/* See also use_cpu_reloc() */
+	if (HAS_LLC(vma->obj->base.dev))
+		return false;
+
+	if (vma->obj->base.write_domain == I915_GEM_DOMAIN_CPU)
+		return false;
+
+	return true;
+}
+
+static bool
+eb_vma_misplaced(struct i915_vma *vma)
+{
+	struct drm_i915_gem_exec_object2 *entry = vma->exec_entry;
+	struct drm_i915_gem_object *obj = vma->obj;
+
+	WARN_ON(entry->flags & __EXEC_OBJECT_NEEDS_MAP &&
 	       !i915_is_ggtt(vma->vm));
 
 	if (entry->alignment &&
 	    vma->node.start & (entry->alignment - 1))
 		return true;
 
-	if (need_mappable && !obj->map_and_fenceable)
+	if (entry->flags & __EXEC_OBJECT_NEEDS_MAP && !obj->map_and_fenceable)
 		return true;
 
 	if (entry->flags & __EXEC_OBJECT_NEEDS_BIAS &&
@@ -642,9 +618,6 @@ i915_gem_execbuffer_reserve(struct intel_engine_cs *ring,
 	bool has_fenced_gpu_access = INTEL_INFO(ring->dev)->gen < 4;
 	int retry;
 
-	if (list_empty(vmas))
-		return 0;
-
 	i915_gem_retire_requests_ring(ring);
 
 	vm = list_first_entry(vmas, struct i915_vma, exec_list)->vm;
@@ -658,20 +631,21 @@ i915_gem_execbuffer_reserve(struct intel_engine_cs *ring,
 		obj = vma->obj;
 		entry = vma->exec_entry;
 
+		if (!has_fenced_gpu_access)
+			entry->flags &= ~EXEC_OBJECT_NEEDS_FENCE;
 		need_fence =
-			has_fenced_gpu_access &&
 			entry->flags & EXEC_OBJECT_NEEDS_FENCE &&
 			obj->tiling_mode != I915_TILING_NONE;
 		need_mappable = need_fence || need_reloc_mappable(vma);
 
-		if (need_mappable)
+		if (need_mappable) {
+			entry->flags |= __EXEC_OBJECT_NEEDS_MAP;
 			list_move(&vma->exec_list, &ordered_vmas);
-		else
+		} else
 			list_move_tail(&vma->exec_list, &ordered_vmas);
 
 		obj->base.pending_read_domains = I915_GEM_GPU_DOMAINS & ~I915_GEM_DOMAIN_COMMAND;
 		obj->base.pending_write_domain = 0;
-		obj->pending_fenced_gpu_access = false;
 	}
 	list_splice(&ordered_vmas, vmas);
 
@@ -696,7 +670,7 @@ i915_gem_execbuffer_reserve(struct intel_engine_cs *ring,
 			if (!drm_mm_node_allocated(&vma->node))
 				continue;
 
-			if (eb_vma_misplaced(vma, has_fenced_gpu_access))
+			if (eb_vma_misplaced(vma))
 				ret = i915_vma_unbind(vma);
 			else
 				ret = i915_gem_execbuffer_reserve_vma(vma, ring, need_relocs);
@@ -744,9 +718,6 @@ i915_gem_execbuffer_relocate_slow(struct drm_device *dev,
 	int i, total, ret;
 	unsigned count = args->buffer_count;
 
-	if (WARN_ON(list_empty(&eb->vmas)))
-		return 0;
-
 	vm = list_first_entry(&eb->vmas, struct i915_vma, exec_list)->vm;
 
 	/* We may process another execbuffer during the unlock... */
@@ -890,18 +861,24 @@ i915_gem_check_execbuffer(struct drm_i915_gem_execbuffer2 *exec)
 }
 
 static int
-validate_exec_list(struct drm_i915_gem_exec_object2 *exec,
+validate_exec_list(struct drm_device *dev,
+		   struct drm_i915_gem_exec_object2 *exec,
 		   int count)
 {
-	int i;
 	unsigned relocs_total = 0;
 	unsigned relocs_max = UINT_MAX / sizeof(struct drm_i915_gem_relocation_entry);
+	unsigned invalid_flags;
+	int i;
+
+	invalid_flags = __EXEC_OBJECT_UNKNOWN_FLAGS;
+	if (USES_FULL_PPGTT(dev))
+		invalid_flags |= EXEC_OBJECT_NEEDS_GTT;
 
 	for (i = 0; i < count; i++) {
 		char __user *ptr = to_user_ptr(exec[i].relocs_ptr);
 		int length; /* limited by fault_in_pages_readable() */
 
-		if (exec[i].flags & __EXEC_OBJECT_UNKNOWN_FLAGS)
+		if (exec[i].flags & invalid_flags)
 			return -EINVAL;
 
 		/* First check for malicious input causing overflow in
@@ -951,16 +928,26 @@ i915_gem_validate_context(struct drm_device *dev, struct drm_file *file,
 		return ERR_PTR(-EIO);
 	}
 
+	if (i915.enable_execlists && !ctx->engine[ring->id].state) {
+		int ret = intel_lr_context_deferred_create(ctx, ring);
+		if (ret) {
+			DRM_DEBUG("Could not create LRC %u: %d\n", ctx_id, ret);
+			return ERR_PTR(ret);
+		}
+	}
+
 	return ctx;
 }
 
-static void
+void
 i915_gem_execbuffer_move_to_active(struct list_head *vmas,
 				   struct intel_engine_cs *ring)
 {
+	u32 seqno = intel_ring_get_seqno(ring);
 	struct i915_vma *vma;
 
 	list_for_each_entry(vma, vmas, exec_list) {
+		struct drm_i915_gem_exec_object2 *entry = vma->exec_entry;
 		struct drm_i915_gem_object *obj = vma->obj;
 		u32 old_read = obj->base.read_domains;
 		u32 old_write = obj->base.write_domain;
@@ -969,24 +956,31 @@ i915_gem_execbuffer_move_to_active(struct list_head *vmas,
 		if (obj->base.write_domain == 0)
 			obj->base.pending_read_domains |= obj->base.read_domains;
 		obj->base.read_domains = obj->base.pending_read_domains;
-		obj->fenced_gpu_access = obj->pending_fenced_gpu_access;
 
 		i915_vma_move_to_active(vma, ring);
 		if (obj->base.write_domain) {
 			obj->dirty = 1;
-			obj->last_write_seqno = intel_ring_get_seqno(ring);
+			obj->last_write_seqno = seqno;
 
 			intel_fb_obj_invalidate(obj, ring);
 
 			/* update for the implicit flush after a batch */
 			obj->base.write_domain &= ~I915_GEM_GPU_DOMAINS;
 		}
+		if (entry->flags & EXEC_OBJECT_NEEDS_FENCE) {
+			obj->last_fenced_seqno = seqno;
+			if (entry->flags & __EXEC_OBJECT_HAS_FENCE) {
+				struct drm_i915_private *dev_priv = to_i915(ring->dev);
+				list_move_tail(&dev_priv->fence_regs[obj->fence_reg].lru_list,
+					       &dev_priv->mm.fence_list);
+			}
+		}
 
 		trace_i915_gem_object_change_domain(obj, old_read, old_write);
 	}
 }
 
-static void
+void
 i915_gem_execbuffer_retire_commands(struct drm_device *dev,
 				    struct drm_file *file,
 				    struct intel_engine_cs *ring,
@@ -1027,13 +1021,54 @@ i915_reset_gen7_sol_offsets(struct drm_device *dev,
 }
 
 static int
-legacy_ringbuffer_submission(struct drm_device *dev, struct drm_file *file,
-			     struct intel_engine_cs *ring,
-			     struct intel_context *ctx,
-			     struct drm_i915_gem_execbuffer2 *args,
-			     struct list_head *vmas,
-			     struct drm_i915_gem_object *batch_obj,
-			     u64 exec_start, u32 flags)
+i915_emit_box(struct intel_engine_cs *ring,
+	      struct drm_clip_rect *box,
+	      int DR1, int DR4)
+{
+	int ret;
+
+	if (box->y2 <= box->y1 || box->x2 <= box->x1 ||
+	    box->y2 <= 0 || box->x2 <= 0) {
+		DRM_ERROR("Bad box %d,%d..%d,%d\n",
+			  box->x1, box->y1, box->x2, box->y2);
+		return -EINVAL;
+	}
+
+	if (INTEL_INFO(ring->dev)->gen >= 4) {
+		ret = intel_ring_begin(ring, 4);
+		if (ret)
+			return ret;
+
+		intel_ring_emit(ring, GFX_OP_DRAWRECT_INFO_I965);
+		intel_ring_emit(ring, (box->x1 & 0xffff) | box->y1 << 16);
+		intel_ring_emit(ring, ((box->x2 - 1) & 0xffff) | (box->y2 - 1) << 16);
+		intel_ring_emit(ring, DR4);
+	} else {
+		ret = intel_ring_begin(ring, 6);
+		if (ret)
+			return ret;
+
+		intel_ring_emit(ring, GFX_OP_DRAWRECT_INFO);
+		intel_ring_emit(ring, DR1);
+		intel_ring_emit(ring, (box->x1 & 0xffff) | box->y1 << 16);
+		intel_ring_emit(ring, ((box->x2 - 1) & 0xffff) | (box->y2 - 1) << 16);
+		intel_ring_emit(ring, DR4);
+		intel_ring_emit(ring, 0);
+	}
+	intel_ring_advance(ring);
+
+	return 0;
+}
+
+
+int
+i915_gem_ringbuffer_submission(struct drm_device *dev, struct drm_file *file,
+			       struct intel_engine_cs *ring,
+			       struct intel_context *ctx,
+			       struct drm_i915_gem_execbuffer2 *args,
+			       struct list_head *vmas,
+			       struct drm_i915_gem_object *batch_obj,
+			       u64 exec_start, u32 flags)
 {
 	struct drm_clip_rect *cliprects = NULL;
 	struct drm_i915_private *dev_priv = dev->dev_private;
@@ -1154,7 +1189,7 @@ legacy_ringbuffer_submission(struct drm_device *dev, struct drm_file *file,
 	exec_len = args->batch_len;
 	if (cliprects) {
 		for (i = 0; i < args->num_cliprects; i++) {
-			ret = i915_emit_box(dev, &cliprects[i],
+			ret = i915_emit_box(ring, &cliprects[i],
 					    args->DR1, args->DR4);
 			if (ret)
 				goto error;
@@ -1254,7 +1289,7 @@ i915_gem_do_execbuffer(struct drm_device *dev, void *data,
 	if (!i915_gem_check_execbuffer(args))
 		return -EINVAL;
 
-	ret = validate_exec_list(exec, args->buffer_count);
+	ret = validate_exec_list(dev, exec, args->buffer_count);
 	if (ret)
 		return ret;
 
@@ -1303,12 +1338,6 @@ i915_gem_do_execbuffer(struct drm_device *dev, void *data,
 	if (ret)
 		goto pre_mutex_err;
 
-	if (dev_priv->ums.mm_suspended) {
-		mutex_unlock(&dev->struct_mutex);
-		ret = -EBUSY;
-		goto pre_mutex_err;
-	}
-
 	ctx = i915_gem_validate_context(dev, file, ring, ctx_id);
 	if (IS_ERR(ctx)) {
 		mutex_unlock(&dev->struct_mutex);
@@ -1318,8 +1347,9 @@ i915_gem_do_execbuffer(struct drm_device *dev, void *data,
 
 	i915_gem_context_reference(ctx);
 
-	vm = ctx->vm;
-	if (!USES_FULL_PPGTT(dev))
+	if (ctx->ppgtt)
+		vm = &ctx->ppgtt->base;
+	else
 		vm = &dev_priv->gtt.base;
 
 	eb = eb_create(args);
@@ -1370,41 +1400,54 @@ i915_gem_do_execbuffer(struct drm_device *dev, void *data,
 				      batch_obj,
 				      args->batch_start_offset,
 				      file->is_master);
-		if (ret)
-			goto err;
-
-		/*
-		 * XXX: Actually do this when enabling batch copy...
-		 *
-		 * Set the DISPATCH_SECURE bit to remove the NON_SECURE bit
-		 * from MI_BATCH_BUFFER_START commands issued in the
-		 * dispatch_execbuffer implementations. We specifically don't
-		 * want that set when the command parser is enabled.
-		 */
+		if (ret) {
+			if (ret != -EACCES)
+				goto err;
+		} else {
+			/*
+			 * XXX: Actually do this when enabling batch copy...
+			 *
+			 * Set the DISPATCH_SECURE bit to remove the NON_SECURE bit
+			 * from MI_BATCH_BUFFER_START commands issued in the
+			 * dispatch_execbuffer implementations. We specifically don't
+			 * want that set when the command parser is enabled.
+			 */
+		}
 	}
 
 	/* snb/ivb/vlv conflate the "batch in ppgtt" bit with the "non-secure
 	 * batch" bit. Hence we need to pin secure batches into the global gtt.
 	 * hsw should have this fixed, but bdw mucks it up again. */
-	if (flags & I915_DISPATCH_SECURE &&
-	    !batch_obj->has_global_gtt_mapping) {
-		/* When we have multiple VMs, we'll need to make sure that we
-		 * allocate space first */
-		struct i915_vma *vma = i915_gem_obj_to_ggtt(batch_obj);
-		BUG_ON(!vma);
-		vma->bind_vma(vma, batch_obj->cache_level, GLOBAL_BIND);
-	}
+	if (flags & I915_DISPATCH_SECURE) {
+		/*
+		 * So on first glance it looks freaky that we pin the batch here
+		 * outside of the reservation loop. But:
+		 * - The batch is already pinned into the relevant ppgtt, so we
+		 *   already have the backing storage fully allocated.
+		 * - No other BO uses the global gtt (well contexts, but meh),
+		 *   so we don't really have issues with mutliple objects not
+		 *   fitting due to fragmentation.
+		 * So this is actually safe.
+		 */
+		ret = i915_gem_obj_ggtt_pin(batch_obj, 0, 0);
+		if (ret)
+			goto err;
 
-	if (flags & I915_DISPATCH_SECURE)
 		exec_start += i915_gem_obj_ggtt_offset(batch_obj);
-	else
+	} else
 		exec_start += i915_gem_obj_offset(batch_obj, vm);
 
-	ret = legacy_ringbuffer_submission(dev, file, ring, ctx,
-			args, &eb->vmas, batch_obj, exec_start, flags);
-	if (ret)
-		goto err;
+	ret = dev_priv->gt.do_execbuf(dev, file, ring, ctx, args,
+				      &eb->vmas, batch_obj, exec_start, flags);
 
+	/*
+	 * FIXME: We crucially rely upon the active tracking for the (ppgtt)
+	 * batch vma for correctness. For less ugly and less fragility this
+	 * needs to be adjusted to also track the ggtt batch vma properly as
+	 * active.
+	 */
+	if (flags & I915_DISPATCH_SECURE)
+		i915_gem_object_ggtt_unpin(batch_obj);
 err:
 	/* the request owns the ref now */
 	i915_gem_context_unreference(ctx);
diff --git a/drivers/gpu/drm/i915/i915_gem_gtt.c b/drivers/gpu/drm/i915/i915_gem_gtt.c
index e42925f76b4b..171f6eafdeee 100644
--- a/drivers/gpu/drm/i915/i915_gem_gtt.c
+++ b/drivers/gpu/drm/i915/i915_gem_gtt.c
@@ -33,26 +33,28 @@
 static void bdw_setup_private_ppat(struct drm_i915_private *dev_priv);
 static void chv_setup_private_ppat(struct drm_i915_private *dev_priv);
 
-bool intel_enable_ppgtt(struct drm_device *dev, bool full)
+static int sanitize_enable_ppgtt(struct drm_device *dev, int enable_ppgtt)
 {
-	if (i915.enable_ppgtt == 0)
-		return false;
-
-	if (i915.enable_ppgtt == 1 && full)
-		return false;
+	bool has_aliasing_ppgtt;
+	bool has_full_ppgtt;
 
-	return true;
-}
+	has_aliasing_ppgtt = INTEL_INFO(dev)->gen >= 6;
+	has_full_ppgtt = INTEL_INFO(dev)->gen >= 7;
+	if (IS_GEN8(dev))
+		has_full_ppgtt = false; /* XXX why? */
 
-static int sanitize_enable_ppgtt(struct drm_device *dev, int enable_ppgtt)
-{
-	if (enable_ppgtt == 0 || !HAS_ALIASING_PPGTT(dev))
+	/*
+	 * We don't allow disabling PPGTT for gen9+ as it's a requirement for
+	 * execlists, the sole mechanism available to submit work.
+	 */
+	if (INTEL_INFO(dev)->gen < 9 &&
+	    (enable_ppgtt == 0 || !has_aliasing_ppgtt))
 		return 0;
 
 	if (enable_ppgtt == 1)
 		return 1;
 
-	if (enable_ppgtt == 2 && HAS_PPGTT(dev))
+	if (enable_ppgtt == 2 && has_full_ppgtt)
 		return 2;
 
 #ifdef CONFIG_INTEL_IOMMU
@@ -70,7 +72,7 @@ static int sanitize_enable_ppgtt(struct drm_device *dev, int enable_ppgtt)
 		return 0;
 	}
 
-	return HAS_ALIASING_PPGTT(dev) ? 1 : 0;
+	return has_aliasing_ppgtt ? 1 : 0;
 }
 
 
@@ -78,7 +80,6 @@ static void ppgtt_bind_vma(struct i915_vma *vma,
 			   enum i915_cache_level cache_level,
 			   u32 flags);
 static void ppgtt_unbind_vma(struct i915_vma *vma);
-static int gen8_ppgtt_enable(struct i915_hw_ppgtt *ppgtt);
 
 static inline gen8_gtt_pte_t gen8_pte_encode(dma_addr_t addr,
 					     enum i915_cache_level level,
@@ -168,9 +169,6 @@ static gen6_gtt_pte_t byt_pte_encode(dma_addr_t addr,
 	gen6_gtt_pte_t pte = valid ? GEN6_PTE_VALID : 0;
 	pte |= GEN6_PTE_ADDR_ENCODE(addr);
 
-	/* Mark the page as writeable.  Other platforms don't have a
-	 * setting for read-only/writable, so this matches that behavior.
-	 */
 	if (!(flags & PTE_READ_ONLY))
 		pte |= BYT_PTE_WRITEABLE;
 
@@ -216,19 +214,12 @@ static gen6_gtt_pte_t iris_pte_encode(dma_addr_t addr,
 
 /* Broadwell Page Directory Pointer Descriptors */
 static int gen8_write_pdp(struct intel_engine_cs *ring, unsigned entry,
-			   uint64_t val, bool synchronous)
+			   uint64_t val)
 {
-	struct drm_i915_private *dev_priv = ring->dev->dev_private;
 	int ret;
 
 	BUG_ON(entry >= 4);
 
-	if (synchronous) {
-		I915_WRITE(GEN8_RING_PDP_UDW(ring, entry), val >> 32);
-		I915_WRITE(GEN8_RING_PDP_LDW(ring, entry), (u32)val);
-		return 0;
-	}
-
 	ret = intel_ring_begin(ring, 6);
 	if (ret)
 		return ret;
@@ -245,8 +236,7 @@ static int gen8_write_pdp(struct intel_engine_cs *ring, unsigned entry,
 }
 
 static int gen8_mm_switch(struct i915_hw_ppgtt *ppgtt,
-			  struct intel_engine_cs *ring,
-			  bool synchronous)
+			  struct intel_engine_cs *ring)
 {
 	int i, ret;
 
@@ -255,7 +245,7 @@ static int gen8_mm_switch(struct i915_hw_ppgtt *ppgtt,
 
 	for (i = used_pd - 1; i >= 0; i--) {
 		dma_addr_t addr = ppgtt->pd_dma_addr[i];
-		ret = gen8_write_pdp(ring, i, addr, synchronous);
+		ret = gen8_write_pdp(ring, i, addr);
 		if (ret)
 			return ret;
 	}
@@ -403,9 +393,6 @@ static void gen8_ppgtt_cleanup(struct i915_address_space *vm)
 	struct i915_hw_ppgtt *ppgtt =
 		container_of(vm, struct i915_hw_ppgtt, base);
 
-	list_del(&vm->global_link);
-	drm_mm_takedown(&vm->mm);
-
 	gen8_ppgtt_unmap_pages(ppgtt);
 	gen8_ppgtt_free(ppgtt);
 }
@@ -615,7 +602,6 @@ static int gen8_ppgtt_init(struct i915_hw_ppgtt *ppgtt, uint64_t size)
 		kunmap_atomic(pd_vaddr);
 	}
 
-	ppgtt->enable = gen8_ppgtt_enable;
 	ppgtt->switch_mm = gen8_mm_switch;
 	ppgtt->base.clear_range = gen8_ppgtt_clear_range;
 	ppgtt->base.insert_entries = gen8_ppgtt_insert_entries;
@@ -724,29 +710,10 @@ static uint32_t get_pd_offset(struct i915_hw_ppgtt *ppgtt)
 }
 
 static int hsw_mm_switch(struct i915_hw_ppgtt *ppgtt,
-			 struct intel_engine_cs *ring,
-			 bool synchronous)
+			 struct intel_engine_cs *ring)
 {
-	struct drm_device *dev = ppgtt->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
 	int ret;
 
-	/* If we're in reset, we can assume the GPU is sufficiently idle to
-	 * manually frob these bits. Ideally we could use the ring functions,
-	 * except our error handling makes it quite difficult (can't use
-	 * intel_ring_begin, ring->flush, or intel_ring_advance)
-	 *
-	 * FIXME: We should try not to special case reset
-	 */
-	if (synchronous ||
-	    i915_reset_in_progress(&dev_priv->gpu_error)) {
-		WARN_ON(ppgtt != dev_priv->mm.aliasing_ppgtt);
-		I915_WRITE(RING_PP_DIR_DCLV(ring), PP_DIR_DCLV_2G);
-		I915_WRITE(RING_PP_DIR_BASE(ring), get_pd_offset(ppgtt));
-		POSTING_READ(RING_PP_DIR_BASE(ring));
-		return 0;
-	}
-
 	/* NB: TLBs must be flushed and invalidated before a switch */
 	ret = ring->flush(ring, I915_GEM_GPU_DOMAINS, I915_GEM_GPU_DOMAINS);
 	if (ret)
@@ -768,29 +735,10 @@ static int hsw_mm_switch(struct i915_hw_ppgtt *ppgtt,
 }
 
 static int gen7_mm_switch(struct i915_hw_ppgtt *ppgtt,
-			  struct intel_engine_cs *ring,
-			  bool synchronous)
+			  struct intel_engine_cs *ring)
 {
-	struct drm_device *dev = ppgtt->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
 	int ret;
 
-	/* If we're in reset, we can assume the GPU is sufficiently idle to
-	 * manually frob these bits. Ideally we could use the ring functions,
-	 * except our error handling makes it quite difficult (can't use
-	 * intel_ring_begin, ring->flush, or intel_ring_advance)
-	 *
-	 * FIXME: We should try not to special case reset
-	 */
-	if (synchronous ||
-	    i915_reset_in_progress(&dev_priv->gpu_error)) {
-		WARN_ON(ppgtt != dev_priv->mm.aliasing_ppgtt);
-		I915_WRITE(RING_PP_DIR_DCLV(ring), PP_DIR_DCLV_2G);
-		I915_WRITE(RING_PP_DIR_BASE(ring), get_pd_offset(ppgtt));
-		POSTING_READ(RING_PP_DIR_BASE(ring));
-		return 0;
-	}
-
 	/* NB: TLBs must be flushed and invalidated before a switch */
 	ret = ring->flush(ring, I915_GEM_GPU_DOMAINS, I915_GEM_GPU_DOMAINS);
 	if (ret)
@@ -819,14 +767,11 @@ static int gen7_mm_switch(struct i915_hw_ppgtt *ppgtt,
 }
 
 static int gen6_mm_switch(struct i915_hw_ppgtt *ppgtt,
-			  struct intel_engine_cs *ring,
-			  bool synchronous)
+			  struct intel_engine_cs *ring)
 {
 	struct drm_device *dev = ppgtt->base.dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 
-	if (!synchronous)
-		return 0;
 
 	I915_WRITE(RING_PP_DIR_DCLV(ring), PP_DIR_DCLV_2G);
 	I915_WRITE(RING_PP_DIR_BASE(ring), get_pd_offset(ppgtt));
@@ -836,39 +781,20 @@ static int gen6_mm_switch(struct i915_hw_ppgtt *ppgtt,
 	return 0;
 }
 
-static int gen8_ppgtt_enable(struct i915_hw_ppgtt *ppgtt)
+static void gen8_ppgtt_enable(struct drm_device *dev)
 {
-	struct drm_device *dev = ppgtt->base.dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_engine_cs *ring;
-	int j, ret;
+	int j;
 
 	for_each_ring(ring, dev_priv, j) {
 		I915_WRITE(RING_MODE_GEN7(ring),
 			   _MASKED_BIT_ENABLE(GFX_PPGTT_ENABLE));
-
-		/* We promise to do a switch later with FULL PPGTT. If this is
-		 * aliasing, this is the one and only switch we'll do */
-		if (USES_FULL_PPGTT(dev))
-			continue;
-
-		ret = ppgtt->switch_mm(ppgtt, ring, true);
-		if (ret)
-			goto err_out;
 	}
-
-	return 0;
-
-err_out:
-	for_each_ring(ring, dev_priv, j)
-		I915_WRITE(RING_MODE_GEN7(ring),
-			   _MASKED_BIT_DISABLE(GFX_PPGTT_ENABLE));
-	return ret;
 }
 
-static int gen7_ppgtt_enable(struct i915_hw_ppgtt *ppgtt)
+static void gen7_ppgtt_enable(struct drm_device *dev)
 {
-	struct drm_device *dev = ppgtt->base.dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_engine_cs *ring;
 	uint32_t ecochk, ecobits;
@@ -887,31 +813,16 @@ static int gen7_ppgtt_enable(struct i915_hw_ppgtt *ppgtt)
 	I915_WRITE(GAM_ECOCHK, ecochk);
 
 	for_each_ring(ring, dev_priv, i) {
-		int ret;
 		/* GFX_MODE is per-ring on gen7+ */
 		I915_WRITE(RING_MODE_GEN7(ring),
 			   _MASKED_BIT_ENABLE(GFX_PPGTT_ENABLE));
-
-		/* We promise to do a switch later with FULL PPGTT. If this is
-		 * aliasing, this is the one and only switch we'll do */
-		if (USES_FULL_PPGTT(dev))
-			continue;
-
-		ret = ppgtt->switch_mm(ppgtt, ring, true);
-		if (ret)
-			return ret;
 	}
-
-	return 0;
 }
 
-static int gen6_ppgtt_enable(struct i915_hw_ppgtt *ppgtt)
+static void gen6_ppgtt_enable(struct drm_device *dev)
 {
-	struct drm_device *dev = ppgtt->base.dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct intel_engine_cs *ring;
 	uint32_t ecochk, gab_ctl, ecobits;
-	int i;
 
 	ecobits = I915_READ(GAC_ECO_BITS);
 	I915_WRITE(GAC_ECO_BITS, ecobits | ECOBITS_SNB_BIT |
@@ -924,14 +835,6 @@ static int gen6_ppgtt_enable(struct i915_hw_ppgtt *ppgtt)
 	I915_WRITE(GAM_ECOCHK, ecochk | ECOCHK_SNB_BIT | ECOCHK_PPGTT_CACHE64B);
 
 	I915_WRITE(GFX_MODE, _MASKED_BIT_ENABLE(GFX_PPGTT_ENABLE));
-
-	for_each_ring(ring, dev_priv, i) {
-		int ret = ppgtt->switch_mm(ppgtt, ring, true);
-		if (ret)
-			return ret;
-	}
-
-	return 0;
 }
 
 /* PPGTT support for Sandybdrige/Gen6 and later */
@@ -1029,8 +932,6 @@ static void gen6_ppgtt_cleanup(struct i915_address_space *vm)
 	struct i915_hw_ppgtt *ppgtt =
 		container_of(vm, struct i915_hw_ppgtt, base);
 
-	list_del(&vm->global_link);
-	drm_mm_takedown(&ppgtt->base.mm);
 	drm_mm_remove_node(&ppgtt->node);
 
 	gen6_ppgtt_unmap_pages(ppgtt);
@@ -1151,13 +1052,10 @@ static int gen6_ppgtt_init(struct i915_hw_ppgtt *ppgtt)
 
 	ppgtt->base.pte_encode = dev_priv->gtt.base.pte_encode;
 	if (IS_GEN6(dev)) {
-		ppgtt->enable = gen6_ppgtt_enable;
 		ppgtt->switch_mm = gen6_mm_switch;
 	} else if (IS_HASWELL(dev)) {
-		ppgtt->enable = gen7_ppgtt_enable;
 		ppgtt->switch_mm = hsw_mm_switch;
 	} else if (IS_GEN7(dev)) {
-		ppgtt->enable = gen7_ppgtt_enable;
 		ppgtt->switch_mm = gen7_mm_switch;
 	} else
 		BUG();
@@ -1188,39 +1086,118 @@ static int gen6_ppgtt_init(struct i915_hw_ppgtt *ppgtt)
 			 ppgtt->node.size >> 20,
 			 ppgtt->node.start / PAGE_SIZE);
 
+	gen6_write_pdes(ppgtt);
+	DRM_DEBUG("Adding PPGTT at offset %x\n",
+		  ppgtt->pd_offset << 10);
+
 	return 0;
 }
 
-int i915_gem_init_ppgtt(struct drm_device *dev, struct i915_hw_ppgtt *ppgtt)
+static int __hw_ppgtt_init(struct drm_device *dev, struct i915_hw_ppgtt *ppgtt)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	int ret = 0;
 
 	ppgtt->base.dev = dev;
 	ppgtt->base.scratch = dev_priv->gtt.base.scratch;
 
 	if (INTEL_INFO(dev)->gen < 8)
-		ret = gen6_ppgtt_init(ppgtt);
-	else if (IS_GEN8(dev))
-		ret = gen8_ppgtt_init(ppgtt, dev_priv->gtt.base.total);
+		return gen6_ppgtt_init(ppgtt);
+	else if (IS_GEN8(dev) || IS_GEN9(dev))
+		return gen8_ppgtt_init(ppgtt, dev_priv->gtt.base.total);
 	else
 		BUG();
+}
+int i915_ppgtt_init(struct drm_device *dev, struct i915_hw_ppgtt *ppgtt)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	int ret = 0;
 
-	if (!ret) {
-		struct drm_i915_private *dev_priv = dev->dev_private;
+	ret = __hw_ppgtt_init(dev, ppgtt);
+	if (ret == 0) {
 		kref_init(&ppgtt->ref);
 		drm_mm_init(&ppgtt->base.mm, ppgtt->base.start,
 			    ppgtt->base.total);
 		i915_init_vm(dev_priv, &ppgtt->base);
-		if (INTEL_INFO(dev)->gen < 8) {
-			gen6_write_pdes(ppgtt);
-			DRM_DEBUG("Adding PPGTT at offset %x\n",
-				  ppgtt->pd_offset << 10);
+	}
+
+	return ret;
+}
+
+int i915_ppgtt_init_hw(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct intel_engine_cs *ring;
+	struct i915_hw_ppgtt *ppgtt = dev_priv->mm.aliasing_ppgtt;
+	int i, ret = 0;
+
+	/* In the case of execlists, PPGTT is enabled by the context descriptor
+	 * and the PDPs are contained within the context itself.  We don't
+	 * need to do anything here. */
+	if (i915.enable_execlists)
+		return 0;
+
+	if (!USES_PPGTT(dev))
+		return 0;
+
+	if (IS_GEN6(dev))
+		gen6_ppgtt_enable(dev);
+	else if (IS_GEN7(dev))
+		gen7_ppgtt_enable(dev);
+	else if (INTEL_INFO(dev)->gen >= 8)
+		gen8_ppgtt_enable(dev);
+	else
+		WARN_ON(1);
+
+	if (ppgtt) {
+		for_each_ring(ring, dev_priv, i) {
+			ret = ppgtt->switch_mm(ppgtt, ring);
+			if (ret != 0)
+				return ret;
 		}
 	}
 
 	return ret;
 }
+struct i915_hw_ppgtt *
+i915_ppgtt_create(struct drm_device *dev, struct drm_i915_file_private *fpriv)
+{
+	struct i915_hw_ppgtt *ppgtt;
+	int ret;
+
+	ppgtt = kzalloc(sizeof(*ppgtt), GFP_KERNEL);
+	if (!ppgtt)
+		return ERR_PTR(-ENOMEM);
+
+	ret = i915_ppgtt_init(dev, ppgtt);
+	if (ret) {
+		kfree(ppgtt);
+		return ERR_PTR(ret);
+	}
+
+	ppgtt->file_priv = fpriv;
+
+	trace_i915_ppgtt_create(&ppgtt->base);
+
+	return ppgtt;
+}
+
+void  i915_ppgtt_release(struct kref *kref)
+{
+	struct i915_hw_ppgtt *ppgtt =
+		container_of(kref, struct i915_hw_ppgtt, ref);
+
+	trace_i915_ppgtt_release(&ppgtt->base);
+
+	/* vmas should already be unbound */
+	WARN_ON(!list_empty(&ppgtt->base.active_list));
+	WARN_ON(!list_empty(&ppgtt->base.inactive_list));
+
+	list_del(&ppgtt->base.global_link);
+	drm_mm_takedown(&ppgtt->base.mm);
+
+	ppgtt->base.cleanup(&ppgtt->base);
+	kfree(ppgtt);
+}
 
 static void
 ppgtt_bind_vma(struct i915_vma *vma,
@@ -1295,7 +1272,7 @@ void i915_check_and_clear_faults(struct drm_device *dev)
 		fault_reg = I915_READ(RING_FAULT_REG(ring));
 		if (fault_reg & RING_FAULT_VALID) {
 			DRM_DEBUG_DRIVER("Unexpected fault\n"
-					 "\tAddr: 0x%08lx\\n"
+					 "\tAddr: 0x%08lx\n"
 					 "\tAddress space: %s\n"
 					 "\tSource ID: %d\n"
 					 "\tType: %d\n",
@@ -1365,7 +1342,7 @@ void i915_gem_restore_gtt_mappings(struct drm_device *dev)
 		 * Unfortunately above, we've just wiped out the mappings
 		 * without telling our object about it. So we need to fake it.
 		 */
-		obj->has_global_gtt_mapping = 0;
+		vma->bound &= ~GLOBAL_BIND;
 		vma->bind_vma(vma, obj->cache_level, GLOBAL_BIND);
 	}
 
@@ -1562,7 +1539,7 @@ static void i915_ggtt_bind_vma(struct i915_vma *vma,
 
 	BUG_ON(!i915_is_ggtt(vma->vm));
 	intel_gtt_insert_sg_entries(vma->obj->pages, entry, flags);
-	vma->obj->has_global_gtt_mapping = 1;
+	vma->bound = GLOBAL_BIND;
 }
 
 static void i915_ggtt_clear_range(struct i915_address_space *vm,
@@ -1581,7 +1558,7 @@ static void i915_ggtt_unbind_vma(struct i915_vma *vma)
 	const unsigned int size = vma->obj->base.size >> PAGE_SHIFT;
 
 	BUG_ON(!i915_is_ggtt(vma->vm));
-	vma->obj->has_global_gtt_mapping = 0;
+	vma->bound = 0;
 	intel_gtt_clear_range(first, size);
 }
 
@@ -1609,24 +1586,24 @@ static void ggtt_bind_vma(struct i915_vma *vma,
 	 * flags. At all other times, the GPU will use the aliasing PPGTT.
 	 */
 	if (!dev_priv->mm.aliasing_ppgtt || flags & GLOBAL_BIND) {
-		if (!obj->has_global_gtt_mapping ||
+		if (!(vma->bound & GLOBAL_BIND) ||
 		    (cache_level != obj->cache_level)) {
 			vma->vm->insert_entries(vma->vm, obj->pages,
 						vma->node.start,
 						cache_level, flags);
-			obj->has_global_gtt_mapping = 1;
+			vma->bound |= GLOBAL_BIND;
 		}
 	}
 
 	if (dev_priv->mm.aliasing_ppgtt &&
-	    (!obj->has_aliasing_ppgtt_mapping ||
+	    (!(vma->bound & LOCAL_BIND) ||
 	     (cache_level != obj->cache_level))) {
 		struct i915_hw_ppgtt *appgtt = dev_priv->mm.aliasing_ppgtt;
 		appgtt->base.insert_entries(&appgtt->base,
 					    vma->obj->pages,
 					    vma->node.start,
 					    cache_level, flags);
-		vma->obj->has_aliasing_ppgtt_mapping = 1;
+		vma->bound |= LOCAL_BIND;
 	}
 }
 
@@ -1636,21 +1613,21 @@ static void ggtt_unbind_vma(struct i915_vma *vma)
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct drm_i915_gem_object *obj = vma->obj;
 
-	if (obj->has_global_gtt_mapping) {
+	if (vma->bound & GLOBAL_BIND) {
 		vma->vm->clear_range(vma->vm,
 				     vma->node.start,
 				     obj->base.size,
 				     true);
-		obj->has_global_gtt_mapping = 0;
+		vma->bound &= ~GLOBAL_BIND;
 	}
 
-	if (obj->has_aliasing_ppgtt_mapping) {
+	if (vma->bound & LOCAL_BIND) {
 		struct i915_hw_ppgtt *appgtt = dev_priv->mm.aliasing_ppgtt;
 		appgtt->base.clear_range(&appgtt->base,
 					 vma->node.start,
 					 obj->base.size,
 					 true);
-		obj->has_aliasing_ppgtt_mapping = 0;
+		vma->bound &= ~LOCAL_BIND;
 	}
 }
 
@@ -1687,10 +1664,10 @@ static void i915_gtt_color_adjust(struct drm_mm_node *node,
 	}
 }
 
-void i915_gem_setup_global_gtt(struct drm_device *dev,
-			       unsigned long start,
-			       unsigned long mappable_end,
-			       unsigned long end)
+static int i915_gem_setup_global_gtt(struct drm_device *dev,
+				     unsigned long start,
+				     unsigned long mappable_end,
+				     unsigned long end)
 {
 	/* Let GEM Manage all of the aperture.
 	 *
@@ -1706,6 +1683,7 @@ void i915_gem_setup_global_gtt(struct drm_device *dev,
 	struct drm_mm_node *entry;
 	struct drm_i915_gem_object *obj;
 	unsigned long hole_start, hole_end;
+	int ret;
 
 	BUG_ON(mappable_end > end);
 
@@ -1717,15 +1695,17 @@ void i915_gem_setup_global_gtt(struct drm_device *dev,
 	/* Mark any preallocated objects as occupied */
 	list_for_each_entry(obj, &dev_priv->mm.bound_list, global_list) {
 		struct i915_vma *vma = i915_gem_obj_to_vma(obj, ggtt_vm);
-		int ret;
+
 		DRM_DEBUG_KMS("reserving preallocated space: %lx + %zx\n",
 			      i915_gem_obj_ggtt_offset(obj), obj->base.size);
 
 		WARN_ON(i915_gem_obj_ggtt_bound(obj));
 		ret = drm_mm_reserve_node(&ggtt_vm->mm, &vma->node);
-		if (ret)
-			DRM_DEBUG_KMS("Reservation failed\n");
-		obj->has_global_gtt_mapping = 1;
+		if (ret) {
+			DRM_DEBUG_KMS("Reservation failed: %i\n", ret);
+			return ret;
+		}
+		vma->bound |= GLOBAL_BIND;
 	}
 
 	dev_priv->gtt.base.start = start;
@@ -1741,6 +1721,22 @@ void i915_gem_setup_global_gtt(struct drm_device *dev,
 
 	/* And finally clear the reserved guard page */
 	ggtt_vm->clear_range(ggtt_vm, end - PAGE_SIZE, PAGE_SIZE, true);
+
+	if (USES_PPGTT(dev) && !USES_FULL_PPGTT(dev)) {
+		struct i915_hw_ppgtt *ppgtt;
+
+		ppgtt = kzalloc(sizeof(*ppgtt), GFP_KERNEL);
+		if (!ppgtt)
+			return -ENOMEM;
+
+		ret = __hw_ppgtt_init(dev, ppgtt);
+		if (ret != 0)
+			return ret;
+
+		dev_priv->mm.aliasing_ppgtt = ppgtt;
+	}
+
+	return 0;
 }
 
 void i915_gem_init_global_gtt(struct drm_device *dev)
@@ -1754,6 +1750,25 @@ void i915_gem_init_global_gtt(struct drm_device *dev)
 	i915_gem_setup_global_gtt(dev, 0, mappable_size, gtt_size);
 }
 
+void i915_global_gtt_cleanup(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct i915_address_space *vm = &dev_priv->gtt.base;
+
+	if (dev_priv->mm.aliasing_ppgtt) {
+		struct i915_hw_ppgtt *ppgtt = dev_priv->mm.aliasing_ppgtt;
+
+		ppgtt->base.cleanup(&ppgtt->base);
+	}
+
+	if (drm_mm_initialized(&vm->mm)) {
+		drm_mm_takedown(&vm->mm);
+		list_del(&vm->global_link);
+	}
+
+	vm->cleanup(vm);
+}
+
 static int setup_scratch_page(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
@@ -1763,7 +1778,6 @@ static int setup_scratch_page(struct drm_device *dev)
 	page = alloc_page(GFP_KERNEL | GFP_DMA32 | __GFP_ZERO);
 	if (page == NULL)
 		return -ENOMEM;
-	get_page(page);
 	set_pages_uc(page, 1);
 
 #ifdef CONFIG_INTEL_IOMMU
@@ -1788,7 +1802,6 @@ static void teardown_scratch_page(struct drm_device *dev)
 	set_pages_wb(page, 1);
 	pci_unmap_page(dev->pdev, dev_priv->gtt.base.scratch.addr,
 		       PAGE_SIZE, PCI_DMA_BIDIRECTIONAL);
-	put_page(page);
 	__free_page(page);
 }
 
@@ -1858,6 +1871,18 @@ static size_t chv_get_stolen_size(u16 gmch_ctrl)
 		return (gmch_ctrl - 0x17 + 9) << 22;
 }
 
+static size_t gen9_get_stolen_size(u16 gen9_gmch_ctl)
+{
+	gen9_gmch_ctl >>= BDW_GMCH_GMS_SHIFT;
+	gen9_gmch_ctl &= BDW_GMCH_GMS_MASK;
+
+	if (gen9_gmch_ctl < 0xf0)
+		return gen9_gmch_ctl << 25; /* 32 MB units */
+	else
+		/* 4MB increments starting at 0xf0 for 4MB */
+		return (gen9_gmch_ctl - 0xf0 + 1) << 22;
+}
+
 static int ggtt_probe_common(struct drm_device *dev,
 			     size_t gtt_size)
 {
@@ -1901,6 +1926,22 @@ static void bdw_setup_private_ppat(struct drm_i915_private *dev_priv)
 	      GEN8_PPAT(6, GEN8_PPAT_WB | GEN8_PPAT_LLCELLC | GEN8_PPAT_AGE(2)) |
 	      GEN8_PPAT(7, GEN8_PPAT_WB | GEN8_PPAT_LLCELLC | GEN8_PPAT_AGE(3));
 
+	if (!USES_PPGTT(dev_priv->dev))
+		/* Spec: "For GGTT, there is NO pat_sel[2:0] from the entry,
+		 * so RTL will always use the value corresponding to
+		 * pat_sel = 000".
+		 * So let's disable cache for GGTT to avoid screen corruptions.
+		 * MOCS still can be used though.
+		 * - System agent ggtt writes (i.e. cpu gtt mmaps) already work
+		 * before this patch, i.e. the same uncached + snooping access
+		 * like on gen6/7 seems to be in effect.
+		 * - So this just fixes blitter/render access. Again it looks
+		 * like it's not just uncached access, but uncached + snooping.
+		 * So we can still hold onto all our assumptions wrt cpu
+		 * clflushing on LLC machines.
+		 */
+		pat = GEN8_PPAT(0, GEN8_PPAT_UC);
+
 	/* XXX: spec defines this as 2 distinct registers. It's unclear if a 64b
 	 * write would work. */
 	I915_WRITE(GEN8_PRIVATE_PAT, pat);
@@ -1917,9 +1958,17 @@ static void chv_setup_private_ppat(struct drm_i915_private *dev_priv)
 	 * Only the snoop bit has meaning for CHV, the rest is
 	 * ignored.
 	 *
-	 * Note that the harware enforces snooping for all page
-	 * table accesses. The snoop bit is actually ignored for
-	 * PDEs.
+	 * The hardware will never snoop for certain types of accesses:
+	 * - CPU GTT (GMADR->GGTT->no snoop->memory)
+	 * - PPGTT page tables
+	 * - some other special cycles
+	 *
+	 * As with BDW, we also need to consider the following for GT accesses:
+	 * "For GGTT, there is NO pat_sel[2:0] from the entry,
+	 * so RTL will always use the value corresponding to
+	 * pat_sel = 000".
+	 * Which means we must set the snoop bit in PAT entry 0
+	 * in order to keep the global status page working.
 	 */
 	pat = GEN8_PPAT(0, CHV_PPAT_SNOOP) |
 	      GEN8_PPAT(1, 0) |
@@ -1954,7 +2003,10 @@ static int gen8_gmch_probe(struct drm_device *dev,
 
 	pci_read_config_word(dev->pdev, SNB_GMCH_CTRL, &snb_gmch_ctl);
 
-	if (IS_CHERRYVIEW(dev)) {
+	if (INTEL_INFO(dev)->gen >= 9) {
+		*stolen = gen9_get_stolen_size(snb_gmch_ctl);
+		gtt_size = gen8_get_total_gtt_size(snb_gmch_ctl);
+	} else if (IS_CHERRYVIEW(dev)) {
 		*stolen = chv_get_stolen_size(snb_gmch_ctl);
 		gtt_size = chv_get_total_gtt_size(snb_gmch_ctl);
 	} else {
@@ -2022,10 +2074,6 @@ static void gen6_gmch_remove(struct i915_address_space *vm)
 
 	struct i915_gtt *gtt = container_of(vm, struct i915_gtt, base);
 
-	if (drm_mm_initialized(&vm->mm)) {
-		drm_mm_takedown(&vm->mm);
-		list_del(&vm->global_link);
-	}
 	iounmap(gtt->gsm);
 	teardown_scratch_page(vm->dev);
 }
@@ -2058,10 +2106,6 @@ static int i915_gmch_probe(struct drm_device *dev,
 
 static void i915_gmch_remove(struct i915_address_space *vm)
 {
-	if (drm_mm_initialized(&vm->mm)) {
-		drm_mm_takedown(&vm->mm);
-		list_del(&vm->global_link);
-	}
 	intel_gmch_remove();
 }
 
@@ -2134,6 +2178,7 @@ static struct i915_vma *__i915_gem_vma_create(struct drm_i915_gem_object *obj,
 	vma->obj = obj;
 
 	switch (INTEL_INFO(vm->dev)->gen) {
+	case 9:
 	case 8:
 	case 7:
 	case 6:
@@ -2160,8 +2205,10 @@ static struct i915_vma *__i915_gem_vma_create(struct drm_i915_gem_object *obj,
 	/* Keep GGTT vmas first to make debug easier */
 	if (i915_is_ggtt(vm))
 		list_add(&vma->vma_link, &obj->vma_list);
-	else
+	else {
 		list_add_tail(&vma->vma_link, &obj->vma_list);
+		i915_ppgtt_get(i915_vm_to_ppgtt(vm));
+	}
 
 	return vma;
 }
diff --git a/drivers/gpu/drm/i915/i915_gem_gtt.h b/drivers/gpu/drm/i915/i915_gem_gtt.h
index 8d6f7c18c404..beaf4bcfdac8 100644
--- a/drivers/gpu/drm/i915/i915_gem_gtt.h
+++ b/drivers/gpu/drm/i915/i915_gem_gtt.h
@@ -34,6 +34,8 @@
 #ifndef __I915_GEM_GTT_H__
 #define __I915_GEM_GTT_H__
 
+struct drm_i915_file_private;
+
 typedef uint32_t gen6_gtt_pte_t;
 typedef uint64_t gen8_gtt_pte_t;
 typedef gen8_gtt_pte_t gen8_ppgtt_pde_t;
@@ -121,6 +123,12 @@ struct i915_vma {
 	struct drm_i915_gem_object *obj;
 	struct i915_address_space *vm;
 
+	/** Flags and address space this VMA is bound to */
+#define GLOBAL_BIND	(1<<0)
+#define LOCAL_BIND	(1<<1)
+#define PTE_READ_ONLY	(1<<2)
+	unsigned int bound : 4;
+
 	/** This object's place on the active/inactive lists */
 	struct list_head mm_list;
 
@@ -153,8 +161,6 @@ struct i915_vma {
 	 * setting the valid PTE entries to a reserved scratch page. */
 	void (*unbind_vma)(struct i915_vma *vma);
 	/* Map an object into an address space with the given cache flags. */
-#define GLOBAL_BIND (1<<0)
-#define PTE_READ_ONLY (1<<1)
 	void (*bind_vma)(struct i915_vma *vma,
 			 enum i915_cache_level cache_level,
 			 u32 flags);
@@ -258,22 +264,34 @@ struct i915_hw_ppgtt {
 		dma_addr_t *gen8_pt_dma_addr[4];
 	};
 
-	struct intel_context *ctx;
+	struct drm_i915_file_private *file_priv;
 
 	int (*enable)(struct i915_hw_ppgtt *ppgtt);
 	int (*switch_mm)(struct i915_hw_ppgtt *ppgtt,
-			 struct intel_engine_cs *ring,
-			 bool synchronous);
+			 struct intel_engine_cs *ring);
 	void (*debug_dump)(struct i915_hw_ppgtt *ppgtt, struct seq_file *m);
 };
 
 int i915_gem_gtt_init(struct drm_device *dev);
 void i915_gem_init_global_gtt(struct drm_device *dev);
-void i915_gem_setup_global_gtt(struct drm_device *dev, unsigned long start,
-			       unsigned long mappable_end, unsigned long end);
-
-bool intel_enable_ppgtt(struct drm_device *dev, bool full);
-int i915_gem_init_ppgtt(struct drm_device *dev, struct i915_hw_ppgtt *ppgtt);
+void i915_global_gtt_cleanup(struct drm_device *dev);
+
+
+int i915_ppgtt_init(struct drm_device *dev, struct i915_hw_ppgtt *ppgtt);
+int i915_ppgtt_init_hw(struct drm_device *dev);
+void i915_ppgtt_release(struct kref *kref);
+struct i915_hw_ppgtt *i915_ppgtt_create(struct drm_device *dev,
+					struct drm_i915_file_private *fpriv);
+static inline void i915_ppgtt_get(struct i915_hw_ppgtt *ppgtt)
+{
+	if (ppgtt)
+		kref_get(&ppgtt->ref);
+}
+static inline void i915_ppgtt_put(struct i915_hw_ppgtt *ppgtt)
+{
+	if (ppgtt)
+		kref_put(&ppgtt->ref, i915_ppgtt_release);
+}
 
 void i915_check_and_clear_faults(struct drm_device *dev);
 void i915_gem_suspend_gtt_mappings(struct drm_device *dev);
diff --git a/drivers/gpu/drm/i915/i915_gem_render_state.c b/drivers/gpu/drm/i915/i915_gem_render_state.c
index e60be3f552a6..98dcd94acba8 100644
--- a/drivers/gpu/drm/i915/i915_gem_render_state.c
+++ b/drivers/gpu/drm/i915/i915_gem_render_state.c
@@ -28,13 +28,6 @@
 #include "i915_drv.h"
 #include "intel_renderstate.h"
 
-struct render_state {
-	const struct intel_renderstate_rodata *rodata;
-	struct drm_i915_gem_object *obj;
-	u64 ggtt_offset;
-	int gen;
-};
-
 static const struct intel_renderstate_rodata *
 render_state_get_rodata(struct drm_device *dev, const int gen)
 {
@@ -45,6 +38,8 @@ render_state_get_rodata(struct drm_device *dev, const int gen)
 		return &gen7_null_state;
 	case 8:
 		return &gen8_null_state;
+	case 9:
+		return &gen9_null_state;
 	}
 
 	return NULL;
@@ -127,30 +122,47 @@ static int render_state_setup(struct render_state *so)
 	return 0;
 }
 
-static void render_state_fini(struct render_state *so)
+void i915_gem_render_state_fini(struct render_state *so)
 {
 	i915_gem_object_ggtt_unpin(so->obj);
 	drm_gem_object_unreference(&so->obj->base);
 }
 
-int i915_gem_render_state_init(struct intel_engine_cs *ring)
+int i915_gem_render_state_prepare(struct intel_engine_cs *ring,
+				  struct render_state *so)
 {
-	struct render_state so;
 	int ret;
 
 	if (WARN_ON(ring->id != RCS))
 		return -ENOENT;
 
-	ret = render_state_init(&so, ring->dev);
+	ret = render_state_init(so, ring->dev);
 	if (ret)
 		return ret;
 
-	if (so.rodata == NULL)
+	if (so->rodata == NULL)
 		return 0;
 
-	ret = render_state_setup(&so);
+	ret = render_state_setup(so);
+	if (ret) {
+		i915_gem_render_state_fini(so);
+		return ret;
+	}
+
+	return 0;
+}
+
+int i915_gem_render_state_init(struct intel_engine_cs *ring)
+{
+	struct render_state so;
+	int ret;
+
+	ret = i915_gem_render_state_prepare(ring, &so);
 	if (ret)
-		goto out;
+		return ret;
+
+	if (so.rodata == NULL)
+		return 0;
 
 	ret = ring->dispatch_execbuffer(ring,
 					so.ggtt_offset,
@@ -164,6 +176,6 @@ int i915_gem_render_state_init(struct intel_engine_cs *ring)
 	ret = __i915_add_request(ring, NULL, so.obj, NULL);
 	/* __i915_add_request moves object to inactive if it fails */
 out:
-	render_state_fini(&so);
+	i915_gem_render_state_fini(&so);
 	return ret;
 }
diff --git a/drivers/gpu/drm/i915/i915_gem_render_state.h b/drivers/gpu/drm/i915/i915_gem_render_state.h
new file mode 100644
index 000000000000..c44961ed3fad
--- /dev/null
+++ b/drivers/gpu/drm/i915/i915_gem_render_state.h
@@ -0,0 +1,47 @@
+/*
+ * Copyright © 2014 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ */
+
+#ifndef _I915_GEM_RENDER_STATE_H_
+#define _I915_GEM_RENDER_STATE_H_
+
+#include <linux/types.h>
+
+struct intel_renderstate_rodata {
+	const u32 *reloc;
+	const u32 *batch;
+	const u32 batch_items;
+};
+
+struct render_state {
+	const struct intel_renderstate_rodata *rodata;
+	struct drm_i915_gem_object *obj;
+	u64 ggtt_offset;
+	int gen;
+};
+
+int i915_gem_render_state_init(struct intel_engine_cs *ring);
+void i915_gem_render_state_fini(struct render_state *so);
+int i915_gem_render_state_prepare(struct intel_engine_cs *ring,
+				  struct render_state *so);
+
+#endif /* _I915_GEM_RENDER_STATE_H_ */
diff --git a/drivers/gpu/drm/i915/i915_gem_stolen.c b/drivers/gpu/drm/i915/i915_gem_stolen.c
index 21c025a209c0..a2045848bd1a 100644
--- a/drivers/gpu/drm/i915/i915_gem_stolen.c
+++ b/drivers/gpu/drm/i915/i915_gem_stolen.c
@@ -137,7 +137,11 @@ static unsigned long i915_stolen_to_physical(struct drm_device *dev)
 		r = devm_request_mem_region(dev->dev, base + 1,
 					    dev_priv->gtt.stolen_size - 1,
 					    "Graphics Stolen Memory");
-		if (r == NULL) {
+		/*
+		 * GEN3 firmware likes to smash pci bridges into the stolen
+		 * range. Apparently this works.
+		 */
+		if (r == NULL && !IS_GEN3(dev)) {
 			DRM_ERROR("conflict detected with stolen region: [0x%08x - 0x%08x]\n",
 				  base, base + (uint32_t)dev_priv->gtt.stolen_size);
 			base = 0;
@@ -289,6 +293,7 @@ void i915_gem_cleanup_stolen(struct drm_device *dev)
 int i915_gem_init_stolen(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
+	u32 tmp;
 	int bios_reserved = 0;
 
 #ifdef CONFIG_INTEL_IOMMU
@@ -308,8 +313,16 @@ int i915_gem_init_stolen(struct drm_device *dev)
 	DRM_DEBUG_KMS("found %zd bytes of stolen memory at %08lx\n",
 		      dev_priv->gtt.stolen_size, dev_priv->mm.stolen_base);
 
-	if (IS_VALLEYVIEW(dev))
-		bios_reserved = 1024*1024; /* top 1M on VLV/BYT */
+	if (INTEL_INFO(dev)->gen >= 8) {
+		tmp = I915_READ(GEN7_BIOS_RESERVED);
+		tmp >>= GEN8_BIOS_RESERVED_SHIFT;
+		tmp &= GEN8_BIOS_RESERVED_MASK;
+		bios_reserved = (1024*1024) << tmp;
+	} else if (IS_GEN7(dev)) {
+		tmp = I915_READ(GEN7_BIOS_RESERVED);
+		bios_reserved = tmp & GEN7_BIOS_RESERVED_256K ?
+			256*1024 : 1024*1024;
+	}
 
 	if (WARN_ON(bios_reserved > dev_priv->gtt.stolen_size))
 		return 0;
@@ -524,7 +537,7 @@ i915_gem_object_create_stolen_for_preallocated(struct drm_device *dev,
 		}
 	}
 
-	obj->has_global_gtt_mapping = 1;
+	vma->bound |= GLOBAL_BIND;
 
 	list_add_tail(&obj->global_list, &dev_priv->mm.bound_list);
 	list_add_tail(&vma->mm_list, &ggtt->inactive_list);
diff --git a/drivers/gpu/drm/i915/i915_gem_tiling.c b/drivers/gpu/drm/i915/i915_gem_tiling.c
index cb150e8b4336..4727a4e2c87c 100644
--- a/drivers/gpu/drm/i915/i915_gem_tiling.c
+++ b/drivers/gpu/drm/i915/i915_gem_tiling.c
@@ -91,26 +91,44 @@ i915_gem_detect_bit_6_swizzle(struct drm_device *dev)
 	uint32_t swizzle_x = I915_BIT_6_SWIZZLE_UNKNOWN;
 	uint32_t swizzle_y = I915_BIT_6_SWIZZLE_UNKNOWN;
 
-	if (IS_VALLEYVIEW(dev)) {
+	if (INTEL_INFO(dev)->gen >= 8 || IS_VALLEYVIEW(dev)) {
+		/*
+		 * On BDW+, swizzling is not used. We leave the CPU memory
+		 * controller in charge of optimizing memory accesses without
+		 * the extra address manipulation GPU side.
+		 *
+		 * VLV and CHV don't have GPU swizzling.
+		 */
 		swizzle_x = I915_BIT_6_SWIZZLE_NONE;
 		swizzle_y = I915_BIT_6_SWIZZLE_NONE;
 	} else if (INTEL_INFO(dev)->gen >= 6) {
-		uint32_t dimm_c0, dimm_c1;
-		dimm_c0 = I915_READ(MAD_DIMM_C0);
-		dimm_c1 = I915_READ(MAD_DIMM_C1);
-		dimm_c0 &= MAD_DIMM_A_SIZE_MASK | MAD_DIMM_B_SIZE_MASK;
-		dimm_c1 &= MAD_DIMM_A_SIZE_MASK | MAD_DIMM_B_SIZE_MASK;
-		/* Enable swizzling when the channels are populated with
-		 * identically sized dimms. We don't need to check the 3rd
-		 * channel because no cpu with gpu attached ships in that
-		 * configuration. Also, swizzling only makes sense for 2
-		 * channels anyway. */
-		if (dimm_c0 == dimm_c1) {
-			swizzle_x = I915_BIT_6_SWIZZLE_9_10;
-			swizzle_y = I915_BIT_6_SWIZZLE_9;
+		if (dev_priv->preserve_bios_swizzle) {
+			if (I915_READ(DISP_ARB_CTL) &
+			    DISP_TILE_SURFACE_SWIZZLING) {
+				swizzle_x = I915_BIT_6_SWIZZLE_9_10;
+				swizzle_y = I915_BIT_6_SWIZZLE_9;
+			} else {
+				swizzle_x = I915_BIT_6_SWIZZLE_NONE;
+				swizzle_y = I915_BIT_6_SWIZZLE_NONE;
+			}
 		} else {
-			swizzle_x = I915_BIT_6_SWIZZLE_NONE;
-			swizzle_y = I915_BIT_6_SWIZZLE_NONE;
+			uint32_t dimm_c0, dimm_c1;
+			dimm_c0 = I915_READ(MAD_DIMM_C0);
+			dimm_c1 = I915_READ(MAD_DIMM_C1);
+			dimm_c0 &= MAD_DIMM_A_SIZE_MASK | MAD_DIMM_B_SIZE_MASK;
+			dimm_c1 &= MAD_DIMM_A_SIZE_MASK | MAD_DIMM_B_SIZE_MASK;
+			/* Enable swizzling when the channels are populated
+			 * with identically sized dimms. We don't need to check
+			 * the 3rd channel because no cpu with gpu attached
+			 * ships in that configuration. Also, swizzling only
+			 * makes sense for 2 channels anyway. */
+			if (dimm_c0 == dimm_c1) {
+				swizzle_x = I915_BIT_6_SWIZZLE_9_10;
+				swizzle_y = I915_BIT_6_SWIZZLE_9;
+			} else {
+				swizzle_x = I915_BIT_6_SWIZZLE_NONE;
+				swizzle_y = I915_BIT_6_SWIZZLE_NONE;
+			}
 		}
 	} else if (IS_GEN5(dev)) {
 		/* On Ironlake whatever DRAM config, GPU always do
@@ -160,6 +178,15 @@ i915_gem_detect_bit_6_swizzle(struct drm_device *dev)
 			}
 			break;
 		}
+
+		/* check for L-shaped memory aka modified enhanced addressing */
+		if (IS_GEN4(dev)) {
+			uint32_t ddc2 = I915_READ(DCC2);
+
+			if (!(ddc2 & DCC2_MODIFIED_ENHANCED_DISABLE))
+				dev_priv->quirks |= QUIRK_PIN_SWIZZLED_PAGES;
+		}
+
 		if (dcc == 0xffffffff) {
 			DRM_ERROR("Couldn't read from MCHBAR.  "
 				  "Disabling tiling.\n");
@@ -357,26 +384,22 @@ i915_gem_set_tiling(struct drm_device *dev, void *data,
 		 * has to also include the unfenced register the GPU uses
 		 * whilst executing a fenced command for an untiled object.
 		 */
-
-		obj->map_and_fenceable =
-			!i915_gem_obj_ggtt_bound(obj) ||
-			(i915_gem_obj_ggtt_offset(obj) +
-			 obj->base.size <= dev_priv->gtt.mappable_end &&
-			 i915_gem_object_fence_ok(obj, args->tiling_mode));
-
-		/* Rebind if we need a change of alignment */
-		if (!obj->map_and_fenceable) {
-			u32 unfenced_align =
-				i915_gem_get_gtt_alignment(dev, obj->base.size,
-							    args->tiling_mode,
-							    false);
-			if (i915_gem_obj_ggtt_offset(obj) & (unfenced_align - 1))
-				ret = i915_gem_object_ggtt_unbind(obj);
-		}
+		if (obj->map_and_fenceable &&
+		    !i915_gem_object_fence_ok(obj, args->tiling_mode))
+			ret = i915_gem_object_ggtt_unbind(obj);
 
 		if (ret == 0) {
+			if (obj->pages &&
+			    obj->madv == I915_MADV_WILLNEED &&
+			    dev_priv->quirks & QUIRK_PIN_SWIZZLED_PAGES) {
+				if (args->tiling_mode == I915_TILING_NONE)
+					i915_gem_object_unpin_pages(obj);
+				if (obj->tiling_mode == I915_TILING_NONE)
+					i915_gem_object_pin_pages(obj);
+			}
+
 			obj->fence_dirty =
-				obj->fenced_gpu_access ||
+				obj->last_fenced_seqno ||
 				obj->fence_reg != I915_FENCE_REG_NONE;
 
 			obj->tiling_mode = args->tiling_mode;
@@ -440,6 +463,7 @@ i915_gem_get_tiling(struct drm_device *dev, void *data,
 	}
 
 	/* Hide bit 17 from the user -- see comment in i915_gem_set_tiling */
+	args->phys_swizzle_mode = args->swizzle_mode;
 	if (args->swizzle_mode == I915_BIT_6_SWIZZLE_9_17)
 		args->swizzle_mode = I915_BIT_6_SWIZZLE_9;
 	if (args->swizzle_mode == I915_BIT_6_SWIZZLE_9_10_17)
diff --git a/drivers/gpu/drm/i915/i915_gem_userptr.c b/drivers/gpu/drm/i915/i915_gem_userptr.c
index d38413997379..d182058383a9 100644
--- a/drivers/gpu/drm/i915/i915_gem_userptr.c
+++ b/drivers/gpu/drm/i915/i915_gem_userptr.c
@@ -293,15 +293,23 @@ i915_gem_userptr_release__mmu_notifier(struct drm_i915_gem_object *obj)
 static struct i915_mmu_notifier *
 i915_mmu_notifier_find(struct i915_mm_struct *mm)
 {
-	if (mm->mn == NULL) {
-		down_write(&mm->mm->mmap_sem);
-		mutex_lock(&to_i915(mm->dev)->mm_lock);
-		if (mm->mn == NULL)
-			mm->mn = i915_mmu_notifier_create(mm->mm);
-		mutex_unlock(&to_i915(mm->dev)->mm_lock);
-		up_write(&mm->mm->mmap_sem);
+	struct i915_mmu_notifier *mn = mm->mn;
+
+	mn = mm->mn;
+	if (mn)
+		return mn;
+
+	down_write(&mm->mm->mmap_sem);
+	mutex_lock(&to_i915(mm->dev)->mm_lock);
+	if ((mn = mm->mn) == NULL) {
+		mn = i915_mmu_notifier_create(mm->mm);
+		if (!IS_ERR(mn))
+			mm->mn = mn;
 	}
-	return mm->mn;
+	mutex_unlock(&to_i915(mm->dev)->mm_lock);
+	up_write(&mm->mm->mmap_sem);
+
+	return mn;
 }
 
 static int
@@ -681,16 +689,15 @@ i915_gem_userptr_get_pages(struct drm_i915_gem_object *obj)
 static void
 i915_gem_userptr_put_pages(struct drm_i915_gem_object *obj)
 {
-	struct scatterlist *sg;
-	int i;
+	struct sg_page_iter sg_iter;
 
 	BUG_ON(obj->userptr.work != NULL);
 
 	if (obj->madv != I915_MADV_WILLNEED)
 		obj->dirty = 0;
 
-	for_each_sg(obj->pages->sgl, sg, obj->pages->nents, i) {
-		struct page *page = sg_page(sg);
+	for_each_sg_page(obj->pages->sgl, &sg_iter, obj->pages->nents, 0) {
+		struct page *page = sg_page_iter_page(&sg_iter);
 
 		if (obj->dirty)
 			set_page_dirty(page);
diff --git a/drivers/gpu/drm/i915/i915_gpu_error.c b/drivers/gpu/drm/i915/i915_gpu_error.c
index eab41f9390f8..cdaee6ce05f8 100644
--- a/drivers/gpu/drm/i915/i915_gpu_error.c
+++ b/drivers/gpu/drm/i915/i915_gpu_error.c
@@ -192,10 +192,10 @@ static void print_error_buffers(struct drm_i915_error_state_buf *m,
 				struct drm_i915_error_buffer *err,
 				int count)
 {
-	err_printf(m, "%s [%d]:\n", name, count);
+	err_printf(m, "  %s [%d]:\n", name, count);
 
 	while (count--) {
-		err_printf(m, "  %08x %8u %02x %02x %x %x",
+		err_printf(m, "    %08x %8u %02x %02x %x %x",
 			   err->gtt_offset,
 			   err->size,
 			   err->read_domains,
@@ -208,7 +208,7 @@ static void print_error_buffers(struct drm_i915_error_state_buf *m,
 		err_puts(m, err->userptr ? " userptr" : "");
 		err_puts(m, err->ring != -1 ? " " : "");
 		err_puts(m, ring_str(err->ring));
-		err_puts(m, i915_cache_level_str(err->cache_level));
+		err_puts(m, i915_cache_level_str(m->i915, err->cache_level));
 
 		if (err->name)
 			err_printf(m, " (name: %d)", err->name);
@@ -242,11 +242,15 @@ static const char *hangcheck_action_to_str(enum intel_ring_hangcheck_action a)
 
 static void i915_ring_error_state(struct drm_i915_error_state_buf *m,
 				  struct drm_device *dev,
-				  struct drm_i915_error_ring *ring)
+				  struct drm_i915_error_state *error,
+				  int ring_idx)
 {
+	struct drm_i915_error_ring *ring = &error->ring[ring_idx];
+
 	if (!ring->valid)
 		return;
 
+	err_printf(m, "%s command stream:\n", ring_str(ring_idx));
 	err_printf(m, "  HEAD: 0x%08x\n", ring->head);
 	err_printf(m, "  TAIL: 0x%08x\n", ring->tail);
 	err_printf(m, "  CTL: 0x%08x\n", ring->ctl);
@@ -388,20 +392,20 @@ int i915_error_state_to_str(struct drm_i915_error_state_buf *m,
 	if (INTEL_INFO(dev)->gen == 7)
 		err_printf(m, "ERR_INT: 0x%08x\n", error->err_int);
 
-	for (i = 0; i < ARRAY_SIZE(error->ring); i++) {
-		err_printf(m, "%s command stream:\n", ring_str(i));
-		i915_ring_error_state(m, dev, &error->ring[i]);
-	}
+	for (i = 0; i < ARRAY_SIZE(error->ring); i++)
+		i915_ring_error_state(m, dev, error, i);
+
+	for (i = 0; i < error->vm_count; i++) {
+		err_printf(m, "vm[%d]\n", i);
 
-	if (error->active_bo)
 		print_error_buffers(m, "Active",
-				    error->active_bo[0],
-				    error->active_bo_count[0]);
+				    error->active_bo[i],
+				    error->active_bo_count[i]);
 
-	if (error->pinned_bo)
 		print_error_buffers(m, "Pinned",
-				    error->pinned_bo[0],
-				    error->pinned_bo_count[0]);
+				    error->pinned_bo[i],
+				    error->pinned_bo_count[i]);
+	}
 
 	for (i = 0; i < ARRAY_SIZE(error->ring); i++) {
 		obj = error->ring[i].batchbuffer;
@@ -492,9 +496,11 @@ out:
 }
 
 int i915_error_state_buf_init(struct drm_i915_error_state_buf *ebuf,
+			      struct drm_i915_private *i915,
 			      size_t count, loff_t pos)
 {
 	memset(ebuf, 0, sizeof(*ebuf));
+	ebuf->i915 = i915;
 
 	/* We need to have enough room to store any i915_error_state printf
 	 * so that we can move it to start position.
@@ -556,24 +562,56 @@ static void i915_error_state_free(struct kref *error_ref)
 }
 
 static struct drm_i915_error_object *
-i915_error_object_create_sized(struct drm_i915_private *dev_priv,
-			       struct drm_i915_gem_object *src,
-			       struct i915_address_space *vm,
-			       const int num_pages)
+i915_error_object_create(struct drm_i915_private *dev_priv,
+			 struct drm_i915_gem_object *src,
+			 struct i915_address_space *vm)
 {
 	struct drm_i915_error_object *dst;
-	int i;
+	struct i915_vma *vma = NULL;
+	int num_pages;
+	bool use_ggtt;
+	int i = 0;
 	u32 reloc_offset;
 
 	if (src == NULL || src->pages == NULL)
 		return NULL;
 
+	num_pages = src->base.size >> PAGE_SHIFT;
+
 	dst = kmalloc(sizeof(*dst) + num_pages * sizeof(u32 *), GFP_ATOMIC);
 	if (dst == NULL)
 		return NULL;
 
-	reloc_offset = dst->gtt_offset = i915_gem_obj_offset(src, vm);
-	for (i = 0; i < num_pages; i++) {
+	if (i915_gem_obj_bound(src, vm))
+		dst->gtt_offset = i915_gem_obj_offset(src, vm);
+	else
+		dst->gtt_offset = -1;
+
+	reloc_offset = dst->gtt_offset;
+	if (i915_is_ggtt(vm))
+		vma = i915_gem_obj_to_ggtt(src);
+	use_ggtt = (src->cache_level == I915_CACHE_NONE &&
+		   vma && (vma->bound & GLOBAL_BIND) &&
+		   reloc_offset + num_pages * PAGE_SIZE <= dev_priv->gtt.mappable_end);
+
+	/* Cannot access stolen address directly, try to use the aperture */
+	if (src->stolen) {
+		use_ggtt = true;
+
+		if (!(vma && vma->bound & GLOBAL_BIND))
+			goto unwind;
+
+		reloc_offset = i915_gem_obj_ggtt_offset(src);
+		if (reloc_offset + num_pages * PAGE_SIZE > dev_priv->gtt.mappable_end)
+			goto unwind;
+	}
+
+	/* Cannot access snooped pages through the aperture */
+	if (use_ggtt && src->cache_level != I915_CACHE_NONE && !HAS_LLC(dev_priv->dev))
+		goto unwind;
+
+	dst->page_count = num_pages;
+	while (num_pages--) {
 		unsigned long flags;
 		void *d;
 
@@ -582,10 +620,7 @@ i915_error_object_create_sized(struct drm_i915_private *dev_priv,
 			goto unwind;
 
 		local_irq_save(flags);
-		if (src->cache_level == I915_CACHE_NONE &&
-		    reloc_offset < dev_priv->gtt.mappable_end &&
-		    src->has_global_gtt_mapping &&
-		    i915_is_ggtt(vm)) {
+		if (use_ggtt) {
 			void __iomem *s;
 
 			/* Simply ignore tiling or any overlapping fence.
@@ -597,14 +632,6 @@ i915_error_object_create_sized(struct drm_i915_private *dev_priv,
 						     reloc_offset);
 			memcpy_fromio(d, s, PAGE_SIZE);
 			io_mapping_unmap_atomic(s);
-		} else if (src->stolen) {
-			unsigned long offset;
-
-			offset = dev_priv->mm.stolen_base;
-			offset += src->stolen->start;
-			offset += i << PAGE_SHIFT;
-
-			memcpy_fromio(d, (void __iomem *) offset, PAGE_SIZE);
 		} else {
 			struct page *page;
 			void *s;
@@ -621,11 +648,9 @@ i915_error_object_create_sized(struct drm_i915_private *dev_priv,
 		}
 		local_irq_restore(flags);
 
-		dst->pages[i] = d;
-
+		dst->pages[i++] = d;
 		reloc_offset += PAGE_SIZE;
 	}
-	dst->page_count = num_pages;
 
 	return dst;
 
@@ -635,22 +660,19 @@ unwind:
 	kfree(dst);
 	return NULL;
 }
-#define i915_error_object_create(dev_priv, src, vm) \
-	i915_error_object_create_sized((dev_priv), (src), (vm), \
-				       (src)->base.size>>PAGE_SHIFT)
-
 #define i915_error_ggtt_object_create(dev_priv, src) \
-	i915_error_object_create_sized((dev_priv), (src), &(dev_priv)->gtt.base, \
-				       (src)->base.size>>PAGE_SHIFT)
+	i915_error_object_create((dev_priv), (src), &(dev_priv)->gtt.base)
 
 static void capture_bo(struct drm_i915_error_buffer *err,
-		       struct drm_i915_gem_object *obj)
+		       struct i915_vma *vma)
 {
+	struct drm_i915_gem_object *obj = vma->obj;
+
 	err->size = obj->base.size;
 	err->name = obj->base.name;
 	err->rseqno = obj->last_read_seqno;
 	err->wseqno = obj->last_write_seqno;
-	err->gtt_offset = i915_gem_obj_ggtt_offset(obj);
+	err->gtt_offset = vma->node.start;
 	err->read_domains = obj->base.read_domains;
 	err->write_domain = obj->base.write_domain;
 	err->fence_reg = obj->fence_reg;
@@ -674,7 +696,7 @@ static u32 capture_active_bo(struct drm_i915_error_buffer *err,
 	int i = 0;
 
 	list_for_each_entry(vma, head, mm_list) {
-		capture_bo(err++, vma->obj);
+		capture_bo(err++, vma);
 		if (++i == count)
 			break;
 	}
@@ -683,21 +705,27 @@ static u32 capture_active_bo(struct drm_i915_error_buffer *err,
 }
 
 static u32 capture_pinned_bo(struct drm_i915_error_buffer *err,
-			     int count, struct list_head *head)
+			     int count, struct list_head *head,
+			     struct i915_address_space *vm)
 {
 	struct drm_i915_gem_object *obj;
-	int i = 0;
+	struct drm_i915_error_buffer * const first = err;
+	struct drm_i915_error_buffer * const last = err + count;
 
 	list_for_each_entry(obj, head, global_list) {
-		if (!i915_gem_obj_is_pinned(obj))
-			continue;
+		struct i915_vma *vma;
 
-		capture_bo(err++, obj);
-		if (++i == count)
+		if (err == last)
 			break;
+
+		list_for_each_entry(vma, &obj->vma_list, vma_link)
+			if (vma->vm == vm && vma->pin_count > 0) {
+				capture_bo(err++, vma);
+				break;
+			}
 	}
 
-	return i;
+	return err - first;
 }
 
 /* Generate a semi-unique error code. The code is not meant to have meaning, The
@@ -741,6 +769,7 @@ static void i915_gem_record_fences(struct drm_device *dev,
 
 	/* Fences */
 	switch (INTEL_INFO(dev)->gen) {
+	case 9:
 	case 8:
 	case 7:
 	case 6:
@@ -780,9 +809,8 @@ static void gen8_record_semaphore_state(struct drm_i915_private *dev_priv,
 
 	if (!error->semaphore_obj)
 		error->semaphore_obj =
-			i915_error_object_create(dev_priv,
-						 dev_priv->semaphore_obj,
-						 &dev_priv->gtt.base);
+			i915_error_ggtt_object_create(dev_priv,
+						      dev_priv->semaphore_obj);
 
 	for_each_ring(to, dev_priv, i) {
 		int idx;
@@ -890,9 +918,6 @@ static void i915_record_ring_state(struct drm_device *dev,
 		ering->hws = I915_READ(mmio);
 	}
 
-	ering->cpu_ring_head = ring->buffer->head;
-	ering->cpu_ring_tail = ring->buffer->tail;
-
 	ering->hangcheck_score = ring->hangcheck.score;
 	ering->hangcheck_action = ring->hangcheck.action;
 
@@ -902,6 +927,7 @@ static void i915_record_ring_state(struct drm_device *dev,
 		ering->vm_info.gfx_mode = I915_READ(RING_MODE_GEN7(ring));
 
 		switch (INTEL_INFO(dev)->gen) {
+		case 9:
 		case 8:
 			for (i = 0; i < 4; i++) {
 				ering->vm_info.pdp[i] =
@@ -955,6 +981,7 @@ static void i915_gem_record_rings(struct drm_device *dev,
 
 	for (i = 0; i < I915_NUM_RINGS; i++) {
 		struct intel_engine_cs *ring = &dev_priv->ring[i];
+		struct intel_ringbuffer *rbuf;
 
 		error->ring[i].pid = -1;
 
@@ -967,6 +994,12 @@ static void i915_gem_record_rings(struct drm_device *dev,
 
 		request = i915_gem_find_active_request(ring);
 		if (request) {
+			struct i915_address_space *vm;
+
+			vm = request->ctx && request->ctx->ppgtt ?
+				&request->ctx->ppgtt->base :
+				&dev_priv->gtt.base;
+
 			/* We need to copy these to an anonymous buffer
 			 * as the simplest method to avoid being overwritten
 			 * by userspace.
@@ -974,12 +1007,9 @@ static void i915_gem_record_rings(struct drm_device *dev,
 			error->ring[i].batchbuffer =
 				i915_error_object_create(dev_priv,
 							 request->batch_obj,
-							 request->ctx ?
-							 request->ctx->vm :
-							 &dev_priv->gtt.base);
+							 vm);
 
-			if (HAS_BROKEN_CS_TLB(dev_priv->dev) &&
-			    ring->scratch.obj)
+			if (HAS_BROKEN_CS_TLB(dev_priv->dev))
 				error->ring[i].wa_batchbuffer =
 					i915_error_ggtt_object_create(dev_priv,
 							     ring->scratch.obj);
@@ -998,12 +1028,27 @@ static void i915_gem_record_rings(struct drm_device *dev,
 			}
 		}
 
+		if (i915.enable_execlists) {
+			/* TODO: This is only a small fix to keep basic error
+			 * capture working, but we need to add more information
+			 * for it to be useful (e.g. dump the context being
+			 * executed).
+			 */
+			if (request)
+				rbuf = request->ctx->engine[ring->id].ringbuf;
+			else
+				rbuf = ring->default_context->engine[ring->id].ringbuf;
+		} else
+			rbuf = ring->buffer;
+
+		error->ring[i].cpu_ring_head = rbuf->head;
+		error->ring[i].cpu_ring_tail = rbuf->tail;
+
 		error->ring[i].ringbuffer =
-			i915_error_ggtt_object_create(dev_priv, ring->buffer->obj);
+			i915_error_ggtt_object_create(dev_priv, rbuf->obj);
 
-		if (ring->status_page.obj)
-			error->ring[i].hws_page =
-				i915_error_ggtt_object_create(dev_priv, ring->status_page.obj);
+		error->ring[i].hws_page =
+			i915_error_ggtt_object_create(dev_priv, ring->status_page.obj);
 
 		i915_gem_record_active_context(ring, error, &error->ring[i]);
 
@@ -1049,9 +1094,14 @@ static void i915_gem_capture_vm(struct drm_i915_private *dev_priv,
 	list_for_each_entry(vma, &vm->active_list, mm_list)
 		i++;
 	error->active_bo_count[ndx] = i;
-	list_for_each_entry(obj, &dev_priv->mm.bound_list, global_list)
-		if (i915_gem_obj_is_pinned(obj))
-			i++;
+
+	list_for_each_entry(obj, &dev_priv->mm.bound_list, global_list) {
+		list_for_each_entry(vma, &obj->vma_list, vma_link)
+			if (vma->vm == vm && vma->pin_count > 0) {
+				i++;
+				break;
+			}
+	}
 	error->pinned_bo_count[ndx] = i - error->active_bo_count[ndx];
 
 	if (i) {
@@ -1070,7 +1120,7 @@ static void i915_gem_capture_vm(struct drm_i915_private *dev_priv,
 		error->pinned_bo_count[ndx] =
 			capture_pinned_bo(pinned_bo,
 					  error->pinned_bo_count[ndx],
-					  &dev_priv->mm.bound_list);
+					  &dev_priv->mm.bound_list, vm);
 	error->active_bo[ndx] = active_bo;
 	error->pinned_bo[ndx] = pinned_bo;
 }
@@ -1091,8 +1141,25 @@ static void i915_gem_capture_buffers(struct drm_i915_private *dev_priv,
 	error->pinned_bo_count = kcalloc(cnt, sizeof(*error->pinned_bo_count),
 					 GFP_ATOMIC);
 
-	list_for_each_entry(vm, &dev_priv->vm_list, global_link)
-		i915_gem_capture_vm(dev_priv, error, vm, i++);
+	if (error->active_bo == NULL ||
+	    error->pinned_bo == NULL ||
+	    error->active_bo_count == NULL ||
+	    error->pinned_bo_count == NULL) {
+		kfree(error->active_bo);
+		kfree(error->active_bo_count);
+		kfree(error->pinned_bo);
+		kfree(error->pinned_bo_count);
+
+		error->active_bo = NULL;
+		error->active_bo_count = NULL;
+		error->pinned_bo = NULL;
+		error->pinned_bo_count = NULL;
+	} else {
+		list_for_each_entry(vm, &dev_priv->vm_list, global_link)
+			i915_gem_capture_vm(dev_priv, error, vm, i++);
+
+		error->vm_count = cnt;
+	}
 }
 
 /* Capture all registers which don't fit into another category. */
@@ -1176,7 +1243,8 @@ static void i915_error_capture_msg(struct drm_device *dev,
 	ecode = i915_error_generate_code(dev_priv, error, &ring_id);
 
 	len = scnprintf(error->error_msg, sizeof(error->error_msg),
-			"GPU HANG: ecode %d:0x%08x", ring_id, ecode);
+			"GPU HANG: ecode %d:%d:0x%08x",
+			INTEL_INFO(dev)->gen, ring_id, ecode);
 
 	if (ring_id != -1 && error->ring[ring_id].pid != -1)
 		len += scnprintf(error->error_msg + len,
@@ -1264,13 +1332,12 @@ void i915_error_state_get(struct drm_device *dev,
 			  struct i915_error_state_file_priv *error_priv)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	unsigned long flags;
 
-	spin_lock_irqsave(&dev_priv->gpu_error.lock, flags);
+	spin_lock_irq(&dev_priv->gpu_error.lock);
 	error_priv->error = dev_priv->gpu_error.first_error;
 	if (error_priv->error)
 		kref_get(&error_priv->error->ref);
-	spin_unlock_irqrestore(&dev_priv->gpu_error.lock, flags);
+	spin_unlock_irq(&dev_priv->gpu_error.lock);
 
 }
 
@@ -1284,22 +1351,21 @@ void i915_destroy_error_state(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct drm_i915_error_state *error;
-	unsigned long flags;
 
-	spin_lock_irqsave(&dev_priv->gpu_error.lock, flags);
+	spin_lock_irq(&dev_priv->gpu_error.lock);
 	error = dev_priv->gpu_error.first_error;
 	dev_priv->gpu_error.first_error = NULL;
-	spin_unlock_irqrestore(&dev_priv->gpu_error.lock, flags);
+	spin_unlock_irq(&dev_priv->gpu_error.lock);
 
 	if (error)
 		kref_put(&error->ref, i915_error_state_free);
 }
 
-const char *i915_cache_level_str(int type)
+const char *i915_cache_level_str(struct drm_i915_private *i915, int type)
 {
 	switch (type) {
 	case I915_CACHE_NONE: return " uncached";
-	case I915_CACHE_LLC: return " snooped or LLC";
+	case I915_CACHE_LLC: return HAS_LLC(i915) ? " LLC" : " snooped";
 	case I915_CACHE_L3_LLC: return " L3+LLC";
 	case I915_CACHE_WT: return " WT";
 	default: return "";
@@ -1327,6 +1393,7 @@ void i915_get_extra_instdone(struct drm_device *dev, uint32_t *instdone)
 		WARN_ONCE(1, "Unsupported platform\n");
 	case 7:
 	case 8:
+	case 9:
 		instdone[0] = I915_READ(GEN7_INSTDONE_1);
 		instdone[1] = I915_READ(GEN7_SC_INSTDONE);
 		instdone[2] = I915_READ(GEN7_SAMPLER_INSTDONE);
diff --git a/drivers/gpu/drm/i915/i915_ioc32.c b/drivers/gpu/drm/i915/i915_ioc32.c
index 2e0613e26251..176de6322e4d 100644
--- a/drivers/gpu/drm/i915/i915_ioc32.c
+++ b/drivers/gpu/drm/i915/i915_ioc32.c
@@ -189,7 +189,6 @@ static drm_ioctl_compat_t *i915_compat_ioctls[] = {
 	[DRM_I915_ALLOC] = compat_i915_alloc
 };
 
-#ifdef CONFIG_COMPAT
 /**
  * Called whenever a 32-bit process running under a 64-bit kernel
  * performs an ioctl on /dev/dri/card<n>.
@@ -218,4 +217,3 @@ long i915_compat_ioctl(struct file *filp, unsigned int cmd, unsigned long arg)
 
 	return ret;
 }
-#endif
diff --git a/drivers/gpu/drm/i915/i915_irq.c b/drivers/gpu/drm/i915/i915_irq.c
index 0050ee9470f1..d0d3dfbe6d2a 100644
--- a/drivers/gpu/drm/i915/i915_irq.c
+++ b/drivers/gpu/drm/i915/i915_irq.c
@@ -37,6 +37,14 @@
 #include "i915_trace.h"
 #include "intel_drv.h"
 
+/**
+ * DOC: interrupt handling
+ *
+ * These functions provide the basic support for enabling and disabling the
+ * interrupt handling support. There's a lot more functionality in i915_irq.c
+ * and related files, but that will be described in separate chapters.
+ */
+
 static const u32 hpd_ibx[] = {
 	[HPD_CRT] = SDE_CRT_HOTPLUG,
 	[HPD_SDVO_B] = SDE_SDVOB_HOTPLUG,
@@ -118,20 +126,22 @@ static const u32 hpd_status_i915[] = { /* i915 and valleyview are the same */
 
 #define GEN8_IRQ_INIT_NDX(type, which, imr_val, ier_val) do { \
 	GEN5_ASSERT_IIR_IS_ZERO(GEN8_##type##_IIR(which)); \
-	I915_WRITE(GEN8_##type##_IMR(which), (imr_val)); \
 	I915_WRITE(GEN8_##type##_IER(which), (ier_val)); \
-	POSTING_READ(GEN8_##type##_IER(which)); \
+	I915_WRITE(GEN8_##type##_IMR(which), (imr_val)); \
+	POSTING_READ(GEN8_##type##_IMR(which)); \
 } while (0)
 
 #define GEN5_IRQ_INIT(type, imr_val, ier_val) do { \
 	GEN5_ASSERT_IIR_IS_ZERO(type##IIR); \
-	I915_WRITE(type##IMR, (imr_val)); \
 	I915_WRITE(type##IER, (ier_val)); \
-	POSTING_READ(type##IER); \
+	I915_WRITE(type##IMR, (imr_val)); \
+	POSTING_READ(type##IMR); \
 } while (0)
 
+static void gen6_rps_irq_handler(struct drm_i915_private *dev_priv, u32 pm_iir);
+
 /* For display hotplug interrupt */
-static void
+void
 ironlake_enable_display_irq(struct drm_i915_private *dev_priv, u32 mask)
 {
 	assert_spin_locked(&dev_priv->irq_lock);
@@ -146,12 +156,12 @@ ironlake_enable_display_irq(struct drm_i915_private *dev_priv, u32 mask)
 	}
 }
 
-static void
+void
 ironlake_disable_display_irq(struct drm_i915_private *dev_priv, u32 mask)
 {
 	assert_spin_locked(&dev_priv->irq_lock);
 
-	if (!intel_irqs_enabled(dev_priv))
+	if (WARN_ON(!intel_irqs_enabled(dev_priv)))
 		return;
 
 	if ((dev_priv->irq_mask & mask) != mask) {
@@ -192,71 +202,28 @@ void gen5_disable_gt_irq(struct drm_i915_private *dev_priv, uint32_t mask)
 	ilk_update_gt_irq(dev_priv, mask, 0);
 }
 
-/**
-  * snb_update_pm_irq - update GEN6_PMIMR
-  * @dev_priv: driver private
-  * @interrupt_mask: mask of interrupt bits to update
-  * @enabled_irq_mask: mask of interrupt bits to enable
-  */
-static void snb_update_pm_irq(struct drm_i915_private *dev_priv,
-			      uint32_t interrupt_mask,
-			      uint32_t enabled_irq_mask)
-{
-	uint32_t new_val;
-
-	assert_spin_locked(&dev_priv->irq_lock);
-
-	if (WARN_ON(!intel_irqs_enabled(dev_priv)))
-		return;
-
-	new_val = dev_priv->pm_irq_mask;
-	new_val &= ~interrupt_mask;
-	new_val |= (~enabled_irq_mask & interrupt_mask);
-
-	if (new_val != dev_priv->pm_irq_mask) {
-		dev_priv->pm_irq_mask = new_val;
-		I915_WRITE(GEN6_PMIMR, dev_priv->pm_irq_mask);
-		POSTING_READ(GEN6_PMIMR);
-	}
-}
-
-void gen6_enable_pm_irq(struct drm_i915_private *dev_priv, uint32_t mask)
+static u32 gen6_pm_iir(struct drm_i915_private *dev_priv)
 {
-	snb_update_pm_irq(dev_priv, mask, mask);
+	return INTEL_INFO(dev_priv)->gen >= 8 ? GEN8_GT_IIR(2) : GEN6_PMIIR;
 }
 
-void gen6_disable_pm_irq(struct drm_i915_private *dev_priv, uint32_t mask)
+static u32 gen6_pm_imr(struct drm_i915_private *dev_priv)
 {
-	snb_update_pm_irq(dev_priv, mask, 0);
+	return INTEL_INFO(dev_priv)->gen >= 8 ? GEN8_GT_IMR(2) : GEN6_PMIMR;
 }
 
-static bool ivb_can_enable_err_int(struct drm_device *dev)
+static u32 gen6_pm_ier(struct drm_i915_private *dev_priv)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct intel_crtc *crtc;
-	enum pipe pipe;
-
-	assert_spin_locked(&dev_priv->irq_lock);
-
-	for_each_pipe(pipe) {
-		crtc = to_intel_crtc(dev_priv->pipe_to_crtc_mapping[pipe]);
-
-		if (crtc->cpu_fifo_underrun_disabled)
-			return false;
-	}
-
-	return true;
+	return INTEL_INFO(dev_priv)->gen >= 8 ? GEN8_GT_IER(2) : GEN6_PMIER;
 }
 
 /**
-  * bdw_update_pm_irq - update GT interrupt 2
+  * snb_update_pm_irq - update GEN6_PMIMR
   * @dev_priv: driver private
   * @interrupt_mask: mask of interrupt bits to update
   * @enabled_irq_mask: mask of interrupt bits to enable
-  *
-  * Copied from the snb function, updated with relevant register offsets
   */
-static void bdw_update_pm_irq(struct drm_i915_private *dev_priv,
+static void snb_update_pm_irq(struct drm_i915_private *dev_priv,
 			      uint32_t interrupt_mask,
 			      uint32_t enabled_irq_mask)
 {
@@ -264,144 +231,91 @@ static void bdw_update_pm_irq(struct drm_i915_private *dev_priv,
 
 	assert_spin_locked(&dev_priv->irq_lock);
 
-	if (WARN_ON(!intel_irqs_enabled(dev_priv)))
-		return;
-
 	new_val = dev_priv->pm_irq_mask;
 	new_val &= ~interrupt_mask;
 	new_val |= (~enabled_irq_mask & interrupt_mask);
 
 	if (new_val != dev_priv->pm_irq_mask) {
 		dev_priv->pm_irq_mask = new_val;
-		I915_WRITE(GEN8_GT_IMR(2), dev_priv->pm_irq_mask);
-		POSTING_READ(GEN8_GT_IMR(2));
+		I915_WRITE(gen6_pm_imr(dev_priv), dev_priv->pm_irq_mask);
+		POSTING_READ(gen6_pm_imr(dev_priv));
 	}
 }
 
-void gen8_enable_pm_irq(struct drm_i915_private *dev_priv, uint32_t mask)
+void gen6_enable_pm_irq(struct drm_i915_private *dev_priv, uint32_t mask)
 {
-	bdw_update_pm_irq(dev_priv, mask, mask);
+	if (WARN_ON(!intel_irqs_enabled(dev_priv)))
+		return;
+
+	snb_update_pm_irq(dev_priv, mask, mask);
 }
 
-void gen8_disable_pm_irq(struct drm_i915_private *dev_priv, uint32_t mask)
+static void __gen6_disable_pm_irq(struct drm_i915_private *dev_priv,
+				  uint32_t mask)
 {
-	bdw_update_pm_irq(dev_priv, mask, 0);
+	snb_update_pm_irq(dev_priv, mask, 0);
 }
 
-static bool cpt_can_enable_serr_int(struct drm_device *dev)
+void gen6_disable_pm_irq(struct drm_i915_private *dev_priv, uint32_t mask)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	enum pipe pipe;
-	struct intel_crtc *crtc;
-
-	assert_spin_locked(&dev_priv->irq_lock);
-
-	for_each_pipe(pipe) {
-		crtc = to_intel_crtc(dev_priv->pipe_to_crtc_mapping[pipe]);
-
-		if (crtc->pch_fifo_underrun_disabled)
-			return false;
-	}
+	if (WARN_ON(!intel_irqs_enabled(dev_priv)))
+		return;
 
-	return true;
+	__gen6_disable_pm_irq(dev_priv, mask);
 }
 
-void i9xx_check_fifo_underruns(struct drm_device *dev)
+void gen6_reset_rps_interrupts(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct intel_crtc *crtc;
-	unsigned long flags;
-
-	spin_lock_irqsave(&dev_priv->irq_lock, flags);
+	uint32_t reg = gen6_pm_iir(dev_priv);
 
-	for_each_intel_crtc(dev, crtc) {
-		u32 reg = PIPESTAT(crtc->pipe);
-		u32 pipestat;
-
-		if (crtc->cpu_fifo_underrun_disabled)
-			continue;
-
-		pipestat = I915_READ(reg) & 0xffff0000;
-		if ((pipestat & PIPE_FIFO_UNDERRUN_STATUS) == 0)
-			continue;
-
-		I915_WRITE(reg, pipestat | PIPE_FIFO_UNDERRUN_STATUS);
-		POSTING_READ(reg);
-
-		DRM_ERROR("pipe %c underrun\n", pipe_name(crtc->pipe));
-	}
-
-	spin_unlock_irqrestore(&dev_priv->irq_lock, flags);
+	spin_lock_irq(&dev_priv->irq_lock);
+	I915_WRITE(reg, dev_priv->pm_rps_events);
+	I915_WRITE(reg, dev_priv->pm_rps_events);
+	POSTING_READ(reg);
+	spin_unlock_irq(&dev_priv->irq_lock);
 }
 
-static void i9xx_set_fifo_underrun_reporting(struct drm_device *dev,
-					     enum pipe pipe,
-					     bool enable, bool old)
+void gen6_enable_rps_interrupts(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	u32 reg = PIPESTAT(pipe);
-	u32 pipestat = I915_READ(reg) & 0xffff0000;
-
-	assert_spin_locked(&dev_priv->irq_lock);
 
-	if (enable) {
-		I915_WRITE(reg, pipestat | PIPE_FIFO_UNDERRUN_STATUS);
-		POSTING_READ(reg);
-	} else {
-		if (old && pipestat & PIPE_FIFO_UNDERRUN_STATUS)
-			DRM_ERROR("pipe %c underrun\n", pipe_name(pipe));
-	}
-}
+	spin_lock_irq(&dev_priv->irq_lock);
 
-static void ironlake_set_fifo_underrun_reporting(struct drm_device *dev,
-						 enum pipe pipe, bool enable)
-{
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	uint32_t bit = (pipe == PIPE_A) ? DE_PIPEA_FIFO_UNDERRUN :
-					  DE_PIPEB_FIFO_UNDERRUN;
+	WARN_ON(dev_priv->rps.pm_iir);
+	WARN_ON(I915_READ(gen6_pm_iir(dev_priv)) & dev_priv->pm_rps_events);
+	dev_priv->rps.interrupts_enabled = true;
+	I915_WRITE(gen6_pm_ier(dev_priv), I915_READ(gen6_pm_ier(dev_priv)) |
+				dev_priv->pm_rps_events);
+	gen6_enable_pm_irq(dev_priv, dev_priv->pm_rps_events);
 
-	if (enable)
-		ironlake_enable_display_irq(dev_priv, bit);
-	else
-		ironlake_disable_display_irq(dev_priv, bit);
+	spin_unlock_irq(&dev_priv->irq_lock);
 }
 
-static void ivybridge_set_fifo_underrun_reporting(struct drm_device *dev,
-						  enum pipe pipe,
-						  bool enable, bool old)
+void gen6_disable_rps_interrupts(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	if (enable) {
-		I915_WRITE(GEN7_ERR_INT, ERR_INT_FIFO_UNDERRUN(pipe));
 
-		if (!ivb_can_enable_err_int(dev))
-			return;
+	spin_lock_irq(&dev_priv->irq_lock);
+	dev_priv->rps.interrupts_enabled = false;
+	spin_unlock_irq(&dev_priv->irq_lock);
 
-		ironlake_enable_display_irq(dev_priv, DE_ERR_INT_IVB);
-	} else {
-		ironlake_disable_display_irq(dev_priv, DE_ERR_INT_IVB);
+	cancel_work_sync(&dev_priv->rps.work);
 
-		if (old &&
-		    I915_READ(GEN7_ERR_INT) & ERR_INT_FIFO_UNDERRUN(pipe)) {
-			DRM_ERROR("uncleared fifo underrun on pipe %c\n",
-				  pipe_name(pipe));
-		}
-	}
-}
+	spin_lock_irq(&dev_priv->irq_lock);
 
-static void broadwell_set_fifo_underrun_reporting(struct drm_device *dev,
-						  enum pipe pipe, bool enable)
-{
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	I915_WRITE(GEN6_PMINTRMSK, INTEL_INFO(dev_priv)->gen >= 8 ?
+		   ~GEN8_PMINTR_REDIRECT_TO_NON_DISP : ~0);
 
-	assert_spin_locked(&dev_priv->irq_lock);
+	__gen6_disable_pm_irq(dev_priv, dev_priv->pm_rps_events);
+	I915_WRITE(gen6_pm_ier(dev_priv), I915_READ(gen6_pm_ier(dev_priv)) &
+				~dev_priv->pm_rps_events);
+	I915_WRITE(gen6_pm_iir(dev_priv), dev_priv->pm_rps_events);
+	I915_WRITE(gen6_pm_iir(dev_priv), dev_priv->pm_rps_events);
 
-	if (enable)
-		dev_priv->de_irq_mask[pipe] &= ~GEN8_PIPE_FIFO_UNDERRUN;
-	else
-		dev_priv->de_irq_mask[pipe] |= GEN8_PIPE_FIFO_UNDERRUN;
-	I915_WRITE(GEN8_DE_PIPE_IMR(pipe), dev_priv->de_irq_mask[pipe]);
-	POSTING_READ(GEN8_DE_PIPE_IMR(pipe));
+	dev_priv->rps.pm_iir = 0;
+
+	spin_unlock_irq(&dev_priv->irq_lock);
 }
 
 /**
@@ -410,9 +324,9 @@ static void broadwell_set_fifo_underrun_reporting(struct drm_device *dev,
  * @interrupt_mask: mask of interrupt bits to update
  * @enabled_irq_mask: mask of interrupt bits to enable
  */
-static void ibx_display_interrupt_update(struct drm_i915_private *dev_priv,
-					 uint32_t interrupt_mask,
-					 uint32_t enabled_irq_mask)
+void ibx_display_interrupt_update(struct drm_i915_private *dev_priv,
+				  uint32_t interrupt_mask,
+				  uint32_t enabled_irq_mask)
 {
 	uint32_t sdeimr = I915_READ(SDEIMR);
 	sdeimr &= ~interrupt_mask;
@@ -426,160 +340,6 @@ static void ibx_display_interrupt_update(struct drm_i915_private *dev_priv,
 	I915_WRITE(SDEIMR, sdeimr);
 	POSTING_READ(SDEIMR);
 }
-#define ibx_enable_display_interrupt(dev_priv, bits) \
-	ibx_display_interrupt_update((dev_priv), (bits), (bits))
-#define ibx_disable_display_interrupt(dev_priv, bits) \
-	ibx_display_interrupt_update((dev_priv), (bits), 0)
-
-static void ibx_set_fifo_underrun_reporting(struct drm_device *dev,
-					    enum transcoder pch_transcoder,
-					    bool enable)
-{
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	uint32_t bit = (pch_transcoder == TRANSCODER_A) ?
-		       SDE_TRANSA_FIFO_UNDER : SDE_TRANSB_FIFO_UNDER;
-
-	if (enable)
-		ibx_enable_display_interrupt(dev_priv, bit);
-	else
-		ibx_disable_display_interrupt(dev_priv, bit);
-}
-
-static void cpt_set_fifo_underrun_reporting(struct drm_device *dev,
-					    enum transcoder pch_transcoder,
-					    bool enable, bool old)
-{
-	struct drm_i915_private *dev_priv = dev->dev_private;
-
-	if (enable) {
-		I915_WRITE(SERR_INT,
-			   SERR_INT_TRANS_FIFO_UNDERRUN(pch_transcoder));
-
-		if (!cpt_can_enable_serr_int(dev))
-			return;
-
-		ibx_enable_display_interrupt(dev_priv, SDE_ERROR_CPT);
-	} else {
-		ibx_disable_display_interrupt(dev_priv, SDE_ERROR_CPT);
-
-		if (old && I915_READ(SERR_INT) &
-		    SERR_INT_TRANS_FIFO_UNDERRUN(pch_transcoder)) {
-			DRM_ERROR("uncleared pch fifo underrun on pch transcoder %c\n",
-				  transcoder_name(pch_transcoder));
-		}
-	}
-}
-
-/**
- * intel_set_cpu_fifo_underrun_reporting - enable/disable FIFO underrun messages
- * @dev: drm device
- * @pipe: pipe
- * @enable: true if we want to report FIFO underrun errors, false otherwise
- *
- * This function makes us disable or enable CPU fifo underruns for a specific
- * pipe. Notice that on some Gens (e.g. IVB, HSW), disabling FIFO underrun
- * reporting for one pipe may also disable all the other CPU error interruts for
- * the other pipes, due to the fact that there's just one interrupt mask/enable
- * bit for all the pipes.
- *
- * Returns the previous state of underrun reporting.
- */
-static bool __intel_set_cpu_fifo_underrun_reporting(struct drm_device *dev,
-						    enum pipe pipe, bool enable)
-{
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct drm_crtc *crtc = dev_priv->pipe_to_crtc_mapping[pipe];
-	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
-	bool old;
-
-	assert_spin_locked(&dev_priv->irq_lock);
-
-	old = !intel_crtc->cpu_fifo_underrun_disabled;
-	intel_crtc->cpu_fifo_underrun_disabled = !enable;
-
-	if (INTEL_INFO(dev)->gen < 5 || IS_VALLEYVIEW(dev))
-		i9xx_set_fifo_underrun_reporting(dev, pipe, enable, old);
-	else if (IS_GEN5(dev) || IS_GEN6(dev))
-		ironlake_set_fifo_underrun_reporting(dev, pipe, enable);
-	else if (IS_GEN7(dev))
-		ivybridge_set_fifo_underrun_reporting(dev, pipe, enable, old);
-	else if (IS_GEN8(dev))
-		broadwell_set_fifo_underrun_reporting(dev, pipe, enable);
-
-	return old;
-}
-
-bool intel_set_cpu_fifo_underrun_reporting(struct drm_device *dev,
-					   enum pipe pipe, bool enable)
-{
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	unsigned long flags;
-	bool ret;
-
-	spin_lock_irqsave(&dev_priv->irq_lock, flags);
-	ret = __intel_set_cpu_fifo_underrun_reporting(dev, pipe, enable);
-	spin_unlock_irqrestore(&dev_priv->irq_lock, flags);
-
-	return ret;
-}
-
-static bool __cpu_fifo_underrun_reporting_enabled(struct drm_device *dev,
-						  enum pipe pipe)
-{
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct drm_crtc *crtc = dev_priv->pipe_to_crtc_mapping[pipe];
-	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
-
-	return !intel_crtc->cpu_fifo_underrun_disabled;
-}
-
-/**
- * intel_set_pch_fifo_underrun_reporting - enable/disable FIFO underrun messages
- * @dev: drm device
- * @pch_transcoder: the PCH transcoder (same as pipe on IVB and older)
- * @enable: true if we want to report FIFO underrun errors, false otherwise
- *
- * This function makes us disable or enable PCH fifo underruns for a specific
- * PCH transcoder. Notice that on some PCHs (e.g. CPT/PPT), disabling FIFO
- * underrun reporting for one transcoder may also disable all the other PCH
- * error interruts for the other transcoders, due to the fact that there's just
- * one interrupt mask/enable bit for all the transcoders.
- *
- * Returns the previous state of underrun reporting.
- */
-bool intel_set_pch_fifo_underrun_reporting(struct drm_device *dev,
-					   enum transcoder pch_transcoder,
-					   bool enable)
-{
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct drm_crtc *crtc = dev_priv->pipe_to_crtc_mapping[pch_transcoder];
-	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
-	unsigned long flags;
-	bool old;
-
-	/*
-	 * NOTE: Pre-LPT has a fixed cpu pipe -> pch transcoder mapping, but LPT
-	 * has only one pch transcoder A that all pipes can use. To avoid racy
-	 * pch transcoder -> pipe lookups from interrupt code simply store the
-	 * underrun statistics in crtc A. Since we never expose this anywhere
-	 * nor use it outside of the fifo underrun code here using the "wrong"
-	 * crtc on LPT won't cause issues.
-	 */
-
-	spin_lock_irqsave(&dev_priv->irq_lock, flags);
-
-	old = !intel_crtc->pch_fifo_underrun_disabled;
-	intel_crtc->pch_fifo_underrun_disabled = !enable;
-
-	if (HAS_PCH_IBX(dev))
-		ibx_set_fifo_underrun_reporting(dev, pch_transcoder, enable);
-	else
-		cpt_set_fifo_underrun_reporting(dev, pch_transcoder, enable, old);
-
-	spin_unlock_irqrestore(&dev_priv->irq_lock, flags);
-	return old;
-}
-
 
 static void
 __i915_enable_pipestat(struct drm_i915_private *dev_priv, enum pipe pipe,
@@ -589,6 +349,7 @@ __i915_enable_pipestat(struct drm_i915_private *dev_priv, enum pipe pipe,
 	u32 pipestat = I915_READ(reg) & PIPESTAT_INT_ENABLE_MASK;
 
 	assert_spin_locked(&dev_priv->irq_lock);
+	WARN_ON(!intel_irqs_enabled(dev_priv));
 
 	if (WARN_ONCE(enable_mask & ~PIPESTAT_INT_ENABLE_MASK ||
 		      status_mask & ~PIPESTAT_INT_STATUS_MASK,
@@ -615,6 +376,7 @@ __i915_disable_pipestat(struct drm_i915_private *dev_priv, enum pipe pipe,
 	u32 pipestat = I915_READ(reg) & PIPESTAT_INT_ENABLE_MASK;
 
 	assert_spin_locked(&dev_priv->irq_lock);
+	WARN_ON(!intel_irqs_enabled(dev_priv));
 
 	if (WARN_ONCE(enable_mask & ~PIPESTAT_INT_ENABLE_MASK ||
 		      status_mask & ~PIPESTAT_INT_STATUS_MASK,
@@ -694,19 +456,18 @@ i915_disable_pipestat(struct drm_i915_private *dev_priv, enum pipe pipe,
 static void i915_enable_asle_pipestat(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	unsigned long irqflags;
 
 	if (!dev_priv->opregion.asle || !IS_MOBILE(dev))
 		return;
 
-	spin_lock_irqsave(&dev_priv->irq_lock, irqflags);
+	spin_lock_irq(&dev_priv->irq_lock);
 
 	i915_enable_pipestat(dev_priv, PIPE_B, PIPE_LEGACY_BLC_EVENT_STATUS);
 	if (INTEL_INFO(dev)->gen >= 4)
 		i915_enable_pipestat(dev_priv, PIPE_A,
 				     PIPE_LEGACY_BLC_EVENT_STATUS);
 
-	spin_unlock_irqrestore(&dev_priv->irq_lock, irqflags);
+	spin_unlock_irq(&dev_priv->irq_lock);
 }
 
 /**
@@ -1020,7 +781,7 @@ static int i915_get_crtc_scanoutpos(struct drm_device *dev, int pipe,
 
 	/* In vblank? */
 	if (in_vbl)
-		ret |= DRM_SCANOUTPOS_INVBL;
+		ret |= DRM_SCANOUTPOS_IN_VBLANK;
 
 	return ret;
 }
@@ -1094,18 +855,17 @@ static void i915_digport_work_func(struct work_struct *work)
 {
 	struct drm_i915_private *dev_priv =
 		container_of(work, struct drm_i915_private, dig_port_work);
-	unsigned long irqflags;
 	u32 long_port_mask, short_port_mask;
 	struct intel_digital_port *intel_dig_port;
 	int i, ret;
 	u32 old_bits = 0;
 
-	spin_lock_irqsave(&dev_priv->irq_lock, irqflags);
+	spin_lock_irq(&dev_priv->irq_lock);
 	long_port_mask = dev_priv->long_hpd_port_mask;
 	dev_priv->long_hpd_port_mask = 0;
 	short_port_mask = dev_priv->short_hpd_port_mask;
 	dev_priv->short_hpd_port_mask = 0;
-	spin_unlock_irqrestore(&dev_priv->irq_lock, irqflags);
+	spin_unlock_irq(&dev_priv->irq_lock);
 
 	for (i = 0; i < I915_MAX_PORTS; i++) {
 		bool valid = false;
@@ -1130,9 +890,9 @@ static void i915_digport_work_func(struct work_struct *work)
 	}
 
 	if (old_bits) {
-		spin_lock_irqsave(&dev_priv->irq_lock, irqflags);
+		spin_lock_irq(&dev_priv->irq_lock);
 		dev_priv->hpd_event_bits |= old_bits;
-		spin_unlock_irqrestore(&dev_priv->irq_lock, irqflags);
+		spin_unlock_irq(&dev_priv->irq_lock);
 		schedule_work(&dev_priv->hotplug_work);
 	}
 }
@@ -1151,7 +911,6 @@ static void i915_hotplug_work_func(struct work_struct *work)
 	struct intel_connector *intel_connector;
 	struct intel_encoder *intel_encoder;
 	struct drm_connector *connector;
-	unsigned long irqflags;
 	bool hpd_disabled = false;
 	bool changed = false;
 	u32 hpd_event_bits;
@@ -1159,7 +918,7 @@ static void i915_hotplug_work_func(struct work_struct *work)
 	mutex_lock(&mode_config->mutex);
 	DRM_DEBUG_KMS("running encoder hotplug functions\n");
 
-	spin_lock_irqsave(&dev_priv->irq_lock, irqflags);
+	spin_lock_irq(&dev_priv->irq_lock);
 
 	hpd_event_bits = dev_priv->hpd_event_bits;
 	dev_priv->hpd_event_bits = 0;
@@ -1193,7 +952,7 @@ static void i915_hotplug_work_func(struct work_struct *work)
 				 msecs_to_jiffies(I915_REENABLE_HOTPLUG_DELAY));
 	}
 
-	spin_unlock_irqrestore(&dev_priv->irq_lock, irqflags);
+	spin_unlock_irq(&dev_priv->irq_lock);
 
 	list_for_each_entry(connector, &mode_config->connector_list, head) {
 		intel_connector = to_intel_connector(connector);
@@ -1260,11 +1019,7 @@ static void notify_ring(struct drm_device *dev,
 
 	trace_i915_gem_request_complete(ring);
 
-	if (drm_core_check_feature(dev, DRIVER_MODESET))
-		intel_notify_mmio_flip(ring);
-
 	wake_up_all(&ring->irq_queue);
-	i915_queue_hangcheck(dev);
 }
 
 static u32 vlv_c0_residency(struct drm_i915_private *dev_priv,
@@ -1322,10 +1077,10 @@ static u32 vlv_c0_residency(struct drm_i915_private *dev_priv,
  * @dev_priv: DRM device private
  *
  */
-static u32 vlv_calc_delay_from_C0_counters(struct drm_i915_private *dev_priv)
+static int vlv_calc_delay_from_C0_counters(struct drm_i915_private *dev_priv)
 {
 	u32 residency_C0_up = 0, residency_C0_down = 0;
-	u8 new_delay, adj;
+	int new_delay, adj;
 
 	dev_priv->rps.ei_interrupt_count++;
 
@@ -1400,14 +1155,15 @@ static void gen6_pm_rps_work(struct work_struct *work)
 	int new_delay, adj;
 
 	spin_lock_irq(&dev_priv->irq_lock);
+	/* Speed up work cancelation during disabling rps interrupts. */
+	if (!dev_priv->rps.interrupts_enabled) {
+		spin_unlock_irq(&dev_priv->irq_lock);
+		return;
+	}
 	pm_iir = dev_priv->rps.pm_iir;
 	dev_priv->rps.pm_iir = 0;
-	if (INTEL_INFO(dev_priv->dev)->gen >= 8)
-		gen8_enable_pm_irq(dev_priv, dev_priv->pm_rps_events);
-	else {
-		/* Make sure not to corrupt PMIMR state used by ringbuffer */
-		gen6_enable_pm_irq(dev_priv, dev_priv->pm_rps_events);
-	}
+	/* Make sure not to corrupt PMIMR state used by ringbuffer on GEN6 */
+	gen6_enable_pm_irq(dev_priv, dev_priv->pm_rps_events);
 	spin_unlock_irq(&dev_priv->irq_lock);
 
 	/* Make sure we didn't queue anything we're not going to process. */
@@ -1488,7 +1244,6 @@ static void ivybridge_parity_work(struct work_struct *work)
 	u32 error_status, row, bank, subbank;
 	char *parity_event[6];
 	uint32_t misccpctl;
-	unsigned long flags;
 	uint8_t slice = 0;
 
 	/* We must turn off DOP level clock gating to access the L3 registers.
@@ -1547,9 +1302,9 @@ static void ivybridge_parity_work(struct work_struct *work)
 
 out:
 	WARN_ON(dev_priv->l3_parity.which_slice);
-	spin_lock_irqsave(&dev_priv->irq_lock, flags);
+	spin_lock_irq(&dev_priv->irq_lock);
 	gen5_enable_gt_irq(dev_priv, GT_PARITY_ERROR(dev_priv->dev));
-	spin_unlock_irqrestore(&dev_priv->irq_lock, flags);
+	spin_unlock_irq(&dev_priv->irq_lock);
 
 	mutex_unlock(&dev_priv->dev->struct_mutex);
 }
@@ -1601,32 +1356,18 @@ static void snb_gt_irq_handler(struct drm_device *dev,
 
 	if (gt_iir & (GT_BLT_CS_ERROR_INTERRUPT |
 		      GT_BSD_CS_ERROR_INTERRUPT |
-		      GT_RENDER_CS_MASTER_ERROR_INTERRUPT)) {
-		i915_handle_error(dev, false, "GT error interrupt 0x%08x",
-				  gt_iir);
-	}
+		      GT_RENDER_CS_MASTER_ERROR_INTERRUPT))
+		DRM_DEBUG("Command parser error, gt_iir 0x%08x\n", gt_iir);
 
 	if (gt_iir & GT_PARITY_ERROR(dev))
 		ivybridge_parity_error_irq_handler(dev, gt_iir);
 }
 
-static void gen8_rps_irq_handler(struct drm_i915_private *dev_priv, u32 pm_iir)
-{
-	if ((pm_iir & dev_priv->pm_rps_events) == 0)
-		return;
-
-	spin_lock(&dev_priv->irq_lock);
-	dev_priv->rps.pm_iir |= pm_iir & dev_priv->pm_rps_events;
-	gen8_disable_pm_irq(dev_priv, pm_iir & dev_priv->pm_rps_events);
-	spin_unlock(&dev_priv->irq_lock);
-
-	queue_work(dev_priv->wq, &dev_priv->rps.work);
-}
-
 static irqreturn_t gen8_gt_irq_handler(struct drm_device *dev,
 				       struct drm_i915_private *dev_priv,
 				       u32 master_ctl)
 {
+	struct intel_engine_cs *ring;
 	u32 rcs, bcs, vcs;
 	uint32_t tmp = 0;
 	irqreturn_t ret = IRQ_NONE;
@@ -1636,12 +1377,20 @@ static irqreturn_t gen8_gt_irq_handler(struct drm_device *dev,
 		if (tmp) {
 			I915_WRITE(GEN8_GT_IIR(0), tmp);
 			ret = IRQ_HANDLED;
+
 			rcs = tmp >> GEN8_RCS_IRQ_SHIFT;
-			bcs = tmp >> GEN8_BCS_IRQ_SHIFT;
+			ring = &dev_priv->ring[RCS];
 			if (rcs & GT_RENDER_USER_INTERRUPT)
-				notify_ring(dev, &dev_priv->ring[RCS]);
+				notify_ring(dev, ring);
+			if (rcs & GT_CONTEXT_SWITCH_INTERRUPT)
+				intel_execlists_handle_ctx_events(ring);
+
+			bcs = tmp >> GEN8_BCS_IRQ_SHIFT;
+			ring = &dev_priv->ring[BCS];
 			if (bcs & GT_RENDER_USER_INTERRUPT)
-				notify_ring(dev, &dev_priv->ring[BCS]);
+				notify_ring(dev, ring);
+			if (bcs & GT_CONTEXT_SWITCH_INTERRUPT)
+				intel_execlists_handle_ctx_events(ring);
 		} else
 			DRM_ERROR("The master control interrupt lied (GT0)!\n");
 	}
@@ -1651,12 +1400,20 @@ static irqreturn_t gen8_gt_irq_handler(struct drm_device *dev,
 		if (tmp) {
 			I915_WRITE(GEN8_GT_IIR(1), tmp);
 			ret = IRQ_HANDLED;
+
 			vcs = tmp >> GEN8_VCS1_IRQ_SHIFT;
+			ring = &dev_priv->ring[VCS];
 			if (vcs & GT_RENDER_USER_INTERRUPT)
-				notify_ring(dev, &dev_priv->ring[VCS]);
+				notify_ring(dev, ring);
+			if (vcs & GT_CONTEXT_SWITCH_INTERRUPT)
+				intel_execlists_handle_ctx_events(ring);
+
 			vcs = tmp >> GEN8_VCS2_IRQ_SHIFT;
+			ring = &dev_priv->ring[VCS2];
 			if (vcs & GT_RENDER_USER_INTERRUPT)
-				notify_ring(dev, &dev_priv->ring[VCS2]);
+				notify_ring(dev, ring);
+			if (vcs & GT_CONTEXT_SWITCH_INTERRUPT)
+				intel_execlists_handle_ctx_events(ring);
 		} else
 			DRM_ERROR("The master control interrupt lied (GT1)!\n");
 	}
@@ -1667,7 +1424,7 @@ static irqreturn_t gen8_gt_irq_handler(struct drm_device *dev,
 			I915_WRITE(GEN8_GT_IIR(2),
 				   tmp & dev_priv->pm_rps_events);
 			ret = IRQ_HANDLED;
-			gen8_rps_irq_handler(dev_priv, tmp);
+			gen6_rps_irq_handler(dev_priv, tmp);
 		} else
 			DRM_ERROR("The master control interrupt lied (PM)!\n");
 	}
@@ -1677,9 +1434,13 @@ static irqreturn_t gen8_gt_irq_handler(struct drm_device *dev,
 		if (tmp) {
 			I915_WRITE(GEN8_GT_IIR(3), tmp);
 			ret = IRQ_HANDLED;
+
 			vcs = tmp >> GEN8_VECS_IRQ_SHIFT;
+			ring = &dev_priv->ring[VECS];
 			if (vcs & GT_RENDER_USER_INTERRUPT)
-				notify_ring(dev, &dev_priv->ring[VECS]);
+				notify_ring(dev, ring);
+			if (vcs & GT_CONTEXT_SWITCH_INTERRUPT)
+				intel_execlists_handle_ctx_events(ring);
 		} else
 			DRM_ERROR("The master control interrupt lied (GT3)!\n");
 	}
@@ -1690,7 +1451,7 @@ static irqreturn_t gen8_gt_irq_handler(struct drm_device *dev,
 #define HPD_STORM_DETECT_PERIOD 1000
 #define HPD_STORM_THRESHOLD 5
 
-static int ilk_port_to_hotplug_shift(enum port port)
+static int pch_port_to_hotplug_shift(enum port port)
 {
 	switch (port) {
 	case PORT_A:
@@ -1706,7 +1467,7 @@ static int ilk_port_to_hotplug_shift(enum port port)
 	}
 }
 
-static int g4x_port_to_hotplug_shift(enum port port)
+static int i915_port_to_hotplug_shift(enum port port)
 {
 	switch (port) {
 	case PORT_A:
@@ -1764,15 +1525,17 @@ static inline void intel_hpd_irq_handler(struct drm_device *dev,
 		if (port && dev_priv->hpd_irq_port[port]) {
 			bool long_hpd;
 
-			if (IS_G4X(dev)) {
-				dig_shift = g4x_port_to_hotplug_shift(port);
-				long_hpd = (hotplug_trigger >> dig_shift) & PORTB_HOTPLUG_LONG_DETECT;
-			} else {
-				dig_shift = ilk_port_to_hotplug_shift(port);
+			if (HAS_PCH_SPLIT(dev)) {
+				dig_shift = pch_port_to_hotplug_shift(port);
 				long_hpd = (dig_hotplug_reg >> dig_shift) & PORTB_HOTPLUG_LONG_DETECT;
+			} else {
+				dig_shift = i915_port_to_hotplug_shift(port);
+				long_hpd = (hotplug_trigger >> dig_shift) & PORTB_HOTPLUG_LONG_DETECT;
 			}
 
-			DRM_DEBUG_DRIVER("digital hpd port %d %d\n", port, long_hpd);
+			DRM_DEBUG_DRIVER("digital hpd port %c - %s\n",
+					 port_name(port),
+					 long_hpd ? "long" : "short");
 			/* for long HPD pulses we want to have the digital queue happen,
 			   but we still want HPD storm detection to function. */
 			if (long_hpd) {
@@ -1875,7 +1638,7 @@ static void display_pipe_crc_irq_handler(struct drm_device *dev, enum pipe pipe,
 
 	if (!pipe_crc->entries) {
 		spin_unlock(&pipe_crc->lock);
-		DRM_ERROR("spurious interrupt\n");
+		DRM_DEBUG_KMS("spurious interrupt\n");
 		return;
 	}
 
@@ -1961,37 +1724,38 @@ static void i9xx_pipe_crc_irq_handler(struct drm_device *dev, enum pipe pipe)
  * the work queue. */
 static void gen6_rps_irq_handler(struct drm_i915_private *dev_priv, u32 pm_iir)
 {
+	/* TODO: RPS on GEN9+ is not supported yet. */
+	if (WARN_ONCE(INTEL_INFO(dev_priv)->gen >= 9,
+		      "GEN9+: unexpected RPS IRQ\n"))
+		return;
+
 	if (pm_iir & dev_priv->pm_rps_events) {
 		spin_lock(&dev_priv->irq_lock);
-		dev_priv->rps.pm_iir |= pm_iir & dev_priv->pm_rps_events;
 		gen6_disable_pm_irq(dev_priv, pm_iir & dev_priv->pm_rps_events);
+		if (dev_priv->rps.interrupts_enabled) {
+			dev_priv->rps.pm_iir |= pm_iir & dev_priv->pm_rps_events;
+			queue_work(dev_priv->wq, &dev_priv->rps.work);
+		}
 		spin_unlock(&dev_priv->irq_lock);
-
-		queue_work(dev_priv->wq, &dev_priv->rps.work);
 	}
 
+	if (INTEL_INFO(dev_priv)->gen >= 8)
+		return;
+
 	if (HAS_VEBOX(dev_priv->dev)) {
 		if (pm_iir & PM_VEBOX_USER_INTERRUPT)
 			notify_ring(dev_priv->dev, &dev_priv->ring[VECS]);
 
-		if (pm_iir & PM_VEBOX_CS_ERROR_INTERRUPT) {
-			i915_handle_error(dev_priv->dev, false,
-					  "VEBOX CS error interrupt 0x%08x",
-					  pm_iir);
-		}
+		if (pm_iir & PM_VEBOX_CS_ERROR_INTERRUPT)
+			DRM_DEBUG("Command parser error, pm_iir 0x%08x\n", pm_iir);
 	}
 }
 
 static bool intel_pipe_handle_vblank(struct drm_device *dev, enum pipe pipe)
 {
-	struct intel_crtc *crtc;
-
 	if (!drm_handle_vblank(dev, pipe))
 		return false;
 
-	crtc = to_intel_crtc(intel_get_crtc_for_pipe(dev, pipe));
-	wake_up(&crtc->vbl_wait);
-
 	return true;
 }
 
@@ -2002,7 +1766,7 @@ static void valleyview_pipestat_irq_handler(struct drm_device *dev, u32 iir)
 	int pipe;
 
 	spin_lock(&dev_priv->irq_lock);
-	for_each_pipe(pipe) {
+	for_each_pipe(dev_priv, pipe) {
 		int reg;
 		u32 mask, iir_bit = 0;
 
@@ -2013,9 +1777,9 @@ static void valleyview_pipestat_irq_handler(struct drm_device *dev, u32 iir)
 		 * we need to be careful that we only handle what we want to
 		 * handle.
 		 */
-		mask = 0;
-		if (__cpu_fifo_underrun_reporting_enabled(dev, pipe))
-			mask |= PIPE_FIFO_UNDERRUN_STATUS;
+
+		/* fifo underruns are filterered in the underrun handler. */
+		mask = PIPE_FIFO_UNDERRUN_STATUS;
 
 		switch (pipe) {
 		case PIPE_A:
@@ -2047,9 +1811,10 @@ static void valleyview_pipestat_irq_handler(struct drm_device *dev, u32 iir)
 	}
 	spin_unlock(&dev_priv->irq_lock);
 
-	for_each_pipe(pipe) {
-		if (pipe_stats[pipe] & PIPE_START_VBLANK_INTERRUPT_STATUS)
-			intel_pipe_handle_vblank(dev, pipe);
+	for_each_pipe(dev_priv, pipe) {
+		if (pipe_stats[pipe] & PIPE_START_VBLANK_INTERRUPT_STATUS &&
+		    intel_pipe_handle_vblank(dev, pipe))
+			intel_check_page_flip(dev, pipe);
 
 		if (pipe_stats[pipe] & PLANE_FLIP_DONE_INT_STATUS_VLV) {
 			intel_prepare_page_flip(dev, pipe);
@@ -2059,9 +1824,8 @@ static void valleyview_pipestat_irq_handler(struct drm_device *dev, u32 iir)
 		if (pipe_stats[pipe] & PIPE_CRC_DONE_INTERRUPT_STATUS)
 			i9xx_pipe_crc_irq_handler(dev, pipe);
 
-		if (pipe_stats[pipe] & PIPE_FIFO_UNDERRUN_STATUS &&
-		    intel_set_cpu_fifo_underrun_reporting(dev, pipe, false))
-			DRM_ERROR("pipe %c underrun\n", pipe_name(pipe));
+		if (pipe_stats[pipe] & PIPE_FIFO_UNDERRUN_STATUS)
+			intel_cpu_fifo_underrun_irq_handler(dev_priv, pipe);
 	}
 
 	if (pipe_stats[0] & PIPE_GMBUS_INTERRUPT_STATUS)
@@ -2216,7 +1980,7 @@ static void ibx_irq_handler(struct drm_device *dev, u32 pch_iir)
 		DRM_ERROR("PCH poison interrupt\n");
 
 	if (pch_iir & SDE_FDI_MASK)
-		for_each_pipe(pipe)
+		for_each_pipe(dev_priv, pipe)
 			DRM_DEBUG_DRIVER("  pipe %c FDI IIR: 0x%08x\n",
 					 pipe_name(pipe),
 					 I915_READ(FDI_RX_IIR(pipe)));
@@ -2228,14 +1992,10 @@ static void ibx_irq_handler(struct drm_device *dev, u32 pch_iir)
 		DRM_DEBUG_DRIVER("PCH transcoder CRC error interrupt\n");
 
 	if (pch_iir & SDE_TRANSA_FIFO_UNDER)
-		if (intel_set_pch_fifo_underrun_reporting(dev, TRANSCODER_A,
-							  false))
-			DRM_ERROR("PCH transcoder A FIFO underrun\n");
+		intel_pch_fifo_underrun_irq_handler(dev_priv, TRANSCODER_A);
 
 	if (pch_iir & SDE_TRANSB_FIFO_UNDER)
-		if (intel_set_pch_fifo_underrun_reporting(dev, TRANSCODER_B,
-							  false))
-			DRM_ERROR("PCH transcoder B FIFO underrun\n");
+		intel_pch_fifo_underrun_irq_handler(dev_priv, TRANSCODER_B);
 }
 
 static void ivb_err_int_handler(struct drm_device *dev)
@@ -2247,13 +2007,9 @@ static void ivb_err_int_handler(struct drm_device *dev)
 	if (err_int & ERR_INT_POISON)
 		DRM_ERROR("Poison interrupt\n");
 
-	for_each_pipe(pipe) {
-		if (err_int & ERR_INT_FIFO_UNDERRUN(pipe)) {
-			if (intel_set_cpu_fifo_underrun_reporting(dev, pipe,
-								  false))
-				DRM_ERROR("Pipe %c FIFO underrun\n",
-					  pipe_name(pipe));
-		}
+	for_each_pipe(dev_priv, pipe) {
+		if (err_int & ERR_INT_FIFO_UNDERRUN(pipe))
+			intel_cpu_fifo_underrun_irq_handler(dev_priv, pipe);
 
 		if (err_int & ERR_INT_PIPE_CRC_DONE(pipe)) {
 			if (IS_IVYBRIDGE(dev))
@@ -2275,19 +2031,13 @@ static void cpt_serr_int_handler(struct drm_device *dev)
 		DRM_ERROR("PCH poison interrupt\n");
 
 	if (serr_int & SERR_INT_TRANS_A_FIFO_UNDERRUN)
-		if (intel_set_pch_fifo_underrun_reporting(dev, TRANSCODER_A,
-							  false))
-			DRM_ERROR("PCH transcoder A FIFO underrun\n");
+		intel_pch_fifo_underrun_irq_handler(dev_priv, TRANSCODER_A);
 
 	if (serr_int & SERR_INT_TRANS_B_FIFO_UNDERRUN)
-		if (intel_set_pch_fifo_underrun_reporting(dev, TRANSCODER_B,
-							  false))
-			DRM_ERROR("PCH transcoder B FIFO underrun\n");
+		intel_pch_fifo_underrun_irq_handler(dev_priv, TRANSCODER_B);
 
 	if (serr_int & SERR_INT_TRANS_C_FIFO_UNDERRUN)
-		if (intel_set_pch_fifo_underrun_reporting(dev, TRANSCODER_C,
-							  false))
-			DRM_ERROR("PCH transcoder C FIFO underrun\n");
+		intel_pch_fifo_underrun_irq_handler(dev_priv, TRANSCODER_C);
 
 	I915_WRITE(SERR_INT, serr_int);
 }
@@ -2324,7 +2074,7 @@ static void cpt_irq_handler(struct drm_device *dev, u32 pch_iir)
 		DRM_DEBUG_DRIVER("Audio CP change interrupt\n");
 
 	if (pch_iir & SDE_FDI_MASK_CPT)
-		for_each_pipe(pipe)
+		for_each_pipe(dev_priv, pipe)
 			DRM_DEBUG_DRIVER("  pipe %c FDI IIR: 0x%08x\n",
 					 pipe_name(pipe),
 					 I915_READ(FDI_RX_IIR(pipe)));
@@ -2347,14 +2097,13 @@ static void ilk_display_irq_handler(struct drm_device *dev, u32 de_iir)
 	if (de_iir & DE_POISON)
 		DRM_ERROR("Poison interrupt\n");
 
-	for_each_pipe(pipe) {
-		if (de_iir & DE_PIPE_VBLANK(pipe))
-			intel_pipe_handle_vblank(dev, pipe);
+	for_each_pipe(dev_priv, pipe) {
+		if (de_iir & DE_PIPE_VBLANK(pipe) &&
+		    intel_pipe_handle_vblank(dev, pipe))
+			intel_check_page_flip(dev, pipe);
 
 		if (de_iir & DE_PIPE_FIFO_UNDERRUN(pipe))
-			if (intel_set_cpu_fifo_underrun_reporting(dev, pipe, false))
-				DRM_ERROR("Pipe %c FIFO underrun\n",
-					  pipe_name(pipe));
+			intel_cpu_fifo_underrun_irq_handler(dev_priv, pipe);
 
 		if (de_iir & DE_PIPE_CRC_DONE(pipe))
 			i9xx_pipe_crc_irq_handler(dev, pipe);
@@ -2397,9 +2146,10 @@ static void ivb_display_irq_handler(struct drm_device *dev, u32 de_iir)
 	if (de_iir & DE_GSE_IVB)
 		intel_opregion_asle_intr(dev);
 
-	for_each_pipe(pipe) {
-		if (de_iir & (DE_PIPE_VBLANK_IVB(pipe)))
-			intel_pipe_handle_vblank(dev, pipe);
+	for_each_pipe(dev_priv, pipe) {
+		if (de_iir & (DE_PIPE_VBLANK_IVB(pipe)) &&
+		    intel_pipe_handle_vblank(dev, pipe))
+			intel_check_page_flip(dev, pipe);
 
 		/* plane/pipes map 1:1 on ilk+ */
 		if (de_iir & DE_PLANE_FLIP_DONE_IVB(pipe)) {
@@ -2503,6 +2253,11 @@ static irqreturn_t gen8_irq_handler(int irq, void *arg)
 	irqreturn_t ret = IRQ_NONE;
 	uint32_t tmp = 0;
 	enum pipe pipe;
+	u32 aux_mask = GEN8_AUX_CHANNEL_A;
+
+	if (IS_GEN9(dev))
+		aux_mask |=  GEN9_AUX_CHANNEL_B | GEN9_AUX_CHANNEL_C |
+			GEN9_AUX_CHANNEL_D;
 
 	master_ctl = I915_READ(GEN8_MASTER_IRQ);
 	master_ctl &= ~GEN8_MASTER_IRQ_CONTROL;
@@ -2535,7 +2290,8 @@ static irqreturn_t gen8_irq_handler(int irq, void *arg)
 		if (tmp) {
 			I915_WRITE(GEN8_DE_PORT_IIR, tmp);
 			ret = IRQ_HANDLED;
-			if (tmp & GEN8_AUX_CHANNEL_A)
+
+			if (tmp & aux_mask)
 				dp_aux_irq_handler(dev);
 			else
 				DRM_ERROR("Unexpected DE Port interrupt\n");
@@ -2544,8 +2300,8 @@ static irqreturn_t gen8_irq_handler(int irq, void *arg)
 			DRM_ERROR("The master control interrupt lied (DE PORT)!\n");
 	}
 
-	for_each_pipe(pipe) {
-		uint32_t pipe_iir;
+	for_each_pipe(dev_priv, pipe) {
+		uint32_t pipe_iir, flip_done = 0, fault_errors = 0;
 
 		if (!(master_ctl & GEN8_DE_PIPE_IRQ(pipe)))
 			continue;
@@ -2554,10 +2310,17 @@ static irqreturn_t gen8_irq_handler(int irq, void *arg)
 		if (pipe_iir) {
 			ret = IRQ_HANDLED;
 			I915_WRITE(GEN8_DE_PIPE_IIR(pipe), pipe_iir);
-			if (pipe_iir & GEN8_PIPE_VBLANK)
-				intel_pipe_handle_vblank(dev, pipe);
 
-			if (pipe_iir & GEN8_PIPE_PRIMARY_FLIP_DONE) {
+			if (pipe_iir & GEN8_PIPE_VBLANK &&
+			    intel_pipe_handle_vblank(dev, pipe))
+				intel_check_page_flip(dev, pipe);
+
+			if (IS_GEN9(dev))
+				flip_done = pipe_iir & GEN9_PIPE_PLANE1_FLIP_DONE;
+			else
+				flip_done = pipe_iir & GEN8_PIPE_PRIMARY_FLIP_DONE;
+
+			if (flip_done) {
 				intel_prepare_page_flip(dev, pipe);
 				intel_finish_page_flip_plane(dev, pipe);
 			}
@@ -2565,18 +2328,20 @@ static irqreturn_t gen8_irq_handler(int irq, void *arg)
 			if (pipe_iir & GEN8_PIPE_CDCLK_CRC_DONE)
 				hsw_pipe_crc_irq_handler(dev, pipe);
 
-			if (pipe_iir & GEN8_PIPE_FIFO_UNDERRUN) {
-				if (intel_set_cpu_fifo_underrun_reporting(dev, pipe,
-									  false))
-					DRM_ERROR("Pipe %c FIFO underrun\n",
-						  pipe_name(pipe));
-			}
+			if (pipe_iir & GEN8_PIPE_FIFO_UNDERRUN)
+				intel_cpu_fifo_underrun_irq_handler(dev_priv,
+								    pipe);
 
-			if (pipe_iir & GEN8_DE_PIPE_IRQ_FAULT_ERRORS) {
+
+			if (IS_GEN9(dev))
+				fault_errors = pipe_iir & GEN9_DE_PIPE_IRQ_FAULT_ERRORS;
+			else
+				fault_errors = pipe_iir & GEN8_DE_PIPE_IRQ_FAULT_ERRORS;
+
+			if (fault_errors)
 				DRM_ERROR("Fault errors on pipe %c\n: 0x%08x",
 					  pipe_name(pipe),
 					  pipe_iir & GEN8_DE_PIPE_IRQ_FAULT_ERRORS);
-			}
 		} else
 			DRM_ERROR("The master control interrupt lied (DE PIPE)!\n");
 	}
@@ -2675,6 +2440,9 @@ static void i915_error_work_func(struct work_struct *work)
 		 * simulated reset via debugs, so get an RPM reference.
 		 */
 		intel_runtime_pm_get(dev_priv);
+
+		intel_prepare_reset(dev);
+
 		/*
 		 * All state reset _must_ be completed before we update the
 		 * reset counter, for otherwise waiters might miss the reset
@@ -2683,7 +2451,7 @@ static void i915_error_work_func(struct work_struct *work)
 		 */
 		ret = i915_reset(dev);
 
-		intel_display_handle_reset(dev);
+		intel_finish_reset(dev);
 
 		intel_runtime_pm_put(dev_priv);
 
@@ -2763,7 +2531,7 @@ static void i915_report_and_clear_eir(struct drm_device *dev)
 
 	if (eir & I915_ERROR_MEMORY_REFRESH) {
 		pr_err("memory refresh error:\n");
-		for_each_pipe(pipe)
+		for_each_pipe(dev_priv, pipe)
 			pr_err("pipe %c stat: 0x%08x\n",
 			       pipe_name(pipe), I915_READ(PIPESTAT(pipe)));
 		/* pipestat has already been acked */
@@ -2860,52 +2628,6 @@ void i915_handle_error(struct drm_device *dev, bool wedged,
 	schedule_work(&dev_priv->gpu_error.work);
 }
 
-static void __always_unused i915_pageflip_stall_check(struct drm_device *dev, int pipe)
-{
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct drm_crtc *crtc = dev_priv->pipe_to_crtc_mapping[pipe];
-	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
-	struct drm_i915_gem_object *obj;
-	struct intel_unpin_work *work;
-	unsigned long flags;
-	bool stall_detected;
-
-	/* Ignore early vblank irqs */
-	if (intel_crtc == NULL)
-		return;
-
-	spin_lock_irqsave(&dev->event_lock, flags);
-	work = intel_crtc->unpin_work;
-
-	if (work == NULL ||
-	    atomic_read(&work->pending) >= INTEL_FLIP_COMPLETE ||
-	    !work->enable_stall_check) {
-		/* Either the pending flip IRQ arrived, or we're too early. Don't check */
-		spin_unlock_irqrestore(&dev->event_lock, flags);
-		return;
-	}
-
-	/* Potential stall - if we see that the flip has happened, assume a missed interrupt */
-	obj = work->pending_flip_obj;
-	if (INTEL_INFO(dev)->gen >= 4) {
-		int dspsurf = DSPSURF(intel_crtc->plane);
-		stall_detected = I915_HI_DISPBASE(I915_READ(dspsurf)) ==
-					i915_gem_obj_ggtt_offset(obj);
-	} else {
-		int dspaddr = DSPADDR(intel_crtc->plane);
-		stall_detected = I915_READ(dspaddr) == (i915_gem_obj_ggtt_offset(obj) +
-							crtc->y * crtc->primary->fb->pitches[0] +
-							crtc->x * crtc->primary->fb->bits_per_pixel/8);
-	}
-
-	spin_unlock_irqrestore(&dev->event_lock, flags);
-
-	if (stall_detected) {
-		DRM_DEBUG_DRIVER("Pageflip stall detected\n");
-		intel_prepare_page_flip(dev, intel_crtc->plane);
-	}
-}
-
 /* Called from drm generic code, passed 'crtc' which
  * we use as a pipe index
  */
@@ -3354,10 +3076,15 @@ static void i915_hangcheck_elapsed(unsigned long data)
 void i915_queue_hangcheck(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct timer_list *timer = &dev_priv->gpu_error.hangcheck_timer;
+
 	if (!i915.enable_hangcheck)
 		return;
 
-	mod_timer(&dev_priv->gpu_error.hangcheck_timer,
+	/* Don't continually defer the hangcheck, but make sure it is active */
+	if (timer_pending(timer))
+		return;
+	mod_timer(timer,
 		  round_jiffies_up(jiffies + DRM_I915_HANGCHECK_JIFFIES));
 }
 
@@ -3420,10 +3147,22 @@ static void ironlake_irq_reset(struct drm_device *dev)
 	ibx_irq_reset(dev);
 }
 
+static void vlv_display_irq_reset(struct drm_i915_private *dev_priv)
+{
+	enum pipe pipe;
+
+	I915_WRITE(PORT_HOTPLUG_EN, 0);
+	I915_WRITE(PORT_HOTPLUG_STAT, I915_READ(PORT_HOTPLUG_STAT));
+
+	for_each_pipe(dev_priv, pipe)
+		I915_WRITE(PIPESTAT(pipe), 0xffff);
+
+	GEN5_IRQ_RESET(VLV_);
+}
+
 static void valleyview_irq_preinstall(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	int pipe;
 
 	/* VLV magic */
 	I915_WRITE(VLV_IMR, 0);
@@ -3431,22 +3170,11 @@ static void valleyview_irq_preinstall(struct drm_device *dev)
 	I915_WRITE(RING_IMR(GEN6_BSD_RING_BASE), 0);
 	I915_WRITE(RING_IMR(BLT_RING_BASE), 0);
 
-	/* and GT */
-	I915_WRITE(GTIIR, I915_READ(GTIIR));
-	I915_WRITE(GTIIR, I915_READ(GTIIR));
-
 	gen5_gt_irq_reset(dev);
 
-	I915_WRITE(DPINVGTT, 0xff);
+	I915_WRITE(DPINVGTT, DPINVGTT_STATUS_MASK);
 
-	I915_WRITE(PORT_HOTPLUG_EN, 0);
-	I915_WRITE(PORT_HOTPLUG_STAT, I915_READ(PORT_HOTPLUG_STAT));
-	for_each_pipe(pipe)
-		I915_WRITE(PIPESTAT(pipe), 0xffff);
-	I915_WRITE(VLV_IIR, 0xffffffff);
-	I915_WRITE(VLV_IMR, 0xffffffff);
-	I915_WRITE(VLV_IER, 0x0);
-	POSTING_READ(VLV_IER);
+	vlv_display_irq_reset(dev_priv);
 }
 
 static void gen8_gt_irq_reset(struct drm_i915_private *dev_priv)
@@ -3467,9 +3195,9 @@ static void gen8_irq_reset(struct drm_device *dev)
 
 	gen8_gt_irq_reset(dev_priv);
 
-	for_each_pipe(pipe)
-		if (intel_display_power_enabled(dev_priv,
-						POWER_DOMAIN_PIPE(pipe)))
+	for_each_pipe(dev_priv, pipe)
+		if (intel_display_power_is_enabled(dev_priv,
+						   POWER_DOMAIN_PIPE(pipe)))
 			GEN8_IRQ_RESET_NDX(DE_PIPE, pipe);
 
 	GEN5_IRQ_RESET(GEN8_DE_PORT_);
@@ -3481,20 +3209,19 @@ static void gen8_irq_reset(struct drm_device *dev)
 
 void gen8_irq_power_well_post_enable(struct drm_i915_private *dev_priv)
 {
-	unsigned long irqflags;
+	uint32_t extra_ier = GEN8_PIPE_VBLANK | GEN8_PIPE_FIFO_UNDERRUN;
 
-	spin_lock_irqsave(&dev_priv->irq_lock, irqflags);
+	spin_lock_irq(&dev_priv->irq_lock);
 	GEN8_IRQ_INIT_NDX(DE_PIPE, PIPE_B, dev_priv->de_irq_mask[PIPE_B],
-			  ~dev_priv->de_irq_mask[PIPE_B]);
+			  ~dev_priv->de_irq_mask[PIPE_B] | extra_ier);
 	GEN8_IRQ_INIT_NDX(DE_PIPE, PIPE_C, dev_priv->de_irq_mask[PIPE_C],
-			  ~dev_priv->de_irq_mask[PIPE_C]);
-	spin_unlock_irqrestore(&dev_priv->irq_lock, irqflags);
+			  ~dev_priv->de_irq_mask[PIPE_C] | extra_ier);
+	spin_unlock_irq(&dev_priv->irq_lock);
 }
 
 static void cherryview_irq_preinstall(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	int pipe;
 
 	I915_WRITE(GEN8_MASTER_IRQ, 0);
 	POSTING_READ(GEN8_MASTER_IRQ);
@@ -3503,37 +3230,25 @@ static void cherryview_irq_preinstall(struct drm_device *dev)
 
 	GEN5_IRQ_RESET(GEN8_PCU_);
 
-	POSTING_READ(GEN8_PCU_IIR);
-
 	I915_WRITE(DPINVGTT, DPINVGTT_STATUS_MASK_CHV);
 
-	I915_WRITE(PORT_HOTPLUG_EN, 0);
-	I915_WRITE(PORT_HOTPLUG_STAT, I915_READ(PORT_HOTPLUG_STAT));
-
-	for_each_pipe(pipe)
-		I915_WRITE(PIPESTAT(pipe), 0xffff);
-
-	I915_WRITE(VLV_IMR, 0xffffffff);
-	I915_WRITE(VLV_IER, 0x0);
-	I915_WRITE(VLV_IIR, 0xffffffff);
-	POSTING_READ(VLV_IIR);
+	vlv_display_irq_reset(dev_priv);
 }
 
 static void ibx_hpd_irq_setup(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct drm_mode_config *mode_config = &dev->mode_config;
 	struct intel_encoder *intel_encoder;
 	u32 hotplug_irqs, hotplug, enabled_irqs = 0;
 
 	if (HAS_PCH_IBX(dev)) {
 		hotplug_irqs = SDE_HOTPLUG_MASK;
-		list_for_each_entry(intel_encoder, &mode_config->encoder_list, base.head)
+		for_each_intel_encoder(dev, intel_encoder)
 			if (dev_priv->hpd_stats[intel_encoder->hpd_pin].hpd_mark == HPD_ENABLED)
 				enabled_irqs |= hpd_ibx[intel_encoder->hpd_pin];
 	} else {
 		hotplug_irqs = SDE_HOTPLUG_MASK_CPT;
-		list_for_each_entry(intel_encoder, &mode_config->encoder_list, base.head)
+		for_each_intel_encoder(dev, intel_encoder)
 			if (dev_priv->hpd_stats[intel_encoder->hpd_pin].hpd_mark == HPD_ENABLED)
 				enabled_irqs |= hpd_cpt[intel_encoder->hpd_pin];
 	}
@@ -3596,8 +3311,10 @@ static void gen5_gt_irq_postinstall(struct drm_device *dev)
 	GEN5_IRQ_INIT(GT, dev_priv->gt_irq_mask, gt_irqs);
 
 	if (INTEL_INFO(dev)->gen >= 6) {
-		pm_irqs |= dev_priv->pm_rps_events;
-
+		/*
+		 * RPS interrupts will get enabled/disabled on demand when RPS
+		 * itself is enabled/disabled.
+		 */
 		if (HAS_VEBOX(dev))
 			pm_irqs |= PM_VEBOX_USER_INTERRUPT;
 
@@ -3608,7 +3325,6 @@ static void gen5_gt_irq_postinstall(struct drm_device *dev)
 
 static int ironlake_irq_postinstall(struct drm_device *dev)
 {
-	unsigned long irqflags;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	u32 display_mask, extra_mask;
 
@@ -3647,9 +3363,9 @@ static int ironlake_irq_postinstall(struct drm_device *dev)
 		 * spinlocking not required here for correctness since interrupt
 		 * setup is guaranteed to run in single-threaded context. But we
 		 * need it to make the assert_spin_locked happy. */
-		spin_lock_irqsave(&dev_priv->irq_lock, irqflags);
+		spin_lock_irq(&dev_priv->irq_lock);
 		ironlake_enable_display_irq(dev_priv, DE_PCU_EVENT);
-		spin_unlock_irqrestore(&dev_priv->irq_lock, irqflags);
+		spin_unlock_irq(&dev_priv->irq_lock);
 	}
 
 	return 0;
@@ -3659,45 +3375,51 @@ static void valleyview_display_irqs_install(struct drm_i915_private *dev_priv)
 {
 	u32 pipestat_mask;
 	u32 iir_mask;
+	enum pipe pipe;
 
 	pipestat_mask = PIPESTAT_INT_STATUS_MASK |
 			PIPE_FIFO_UNDERRUN_STATUS;
 
-	I915_WRITE(PIPESTAT(PIPE_A), pipestat_mask);
-	I915_WRITE(PIPESTAT(PIPE_B), pipestat_mask);
+	for_each_pipe(dev_priv, pipe)
+		I915_WRITE(PIPESTAT(pipe), pipestat_mask);
 	POSTING_READ(PIPESTAT(PIPE_A));
 
 	pipestat_mask = PLANE_FLIP_DONE_INT_STATUS_VLV |
 			PIPE_CRC_DONE_INTERRUPT_STATUS;
 
-	i915_enable_pipestat(dev_priv, PIPE_A, pipestat_mask |
-					       PIPE_GMBUS_INTERRUPT_STATUS);
-	i915_enable_pipestat(dev_priv, PIPE_B, pipestat_mask);
+	i915_enable_pipestat(dev_priv, PIPE_A, PIPE_GMBUS_INTERRUPT_STATUS);
+	for_each_pipe(dev_priv, pipe)
+		      i915_enable_pipestat(dev_priv, pipe, pipestat_mask);
 
 	iir_mask = I915_DISPLAY_PORT_INTERRUPT |
 		   I915_DISPLAY_PIPE_A_EVENT_INTERRUPT |
 		   I915_DISPLAY_PIPE_B_EVENT_INTERRUPT;
+	if (IS_CHERRYVIEW(dev_priv))
+		iir_mask |= I915_DISPLAY_PIPE_C_EVENT_INTERRUPT;
 	dev_priv->irq_mask &= ~iir_mask;
 
 	I915_WRITE(VLV_IIR, iir_mask);
 	I915_WRITE(VLV_IIR, iir_mask);
-	I915_WRITE(VLV_IMR, dev_priv->irq_mask);
 	I915_WRITE(VLV_IER, ~dev_priv->irq_mask);
-	POSTING_READ(VLV_IER);
+	I915_WRITE(VLV_IMR, dev_priv->irq_mask);
+	POSTING_READ(VLV_IMR);
 }
 
 static void valleyview_display_irqs_uninstall(struct drm_i915_private *dev_priv)
 {
 	u32 pipestat_mask;
 	u32 iir_mask;
+	enum pipe pipe;
 
 	iir_mask = I915_DISPLAY_PORT_INTERRUPT |
 		   I915_DISPLAY_PIPE_A_EVENT_INTERRUPT |
 		   I915_DISPLAY_PIPE_B_EVENT_INTERRUPT;
+	if (IS_CHERRYVIEW(dev_priv))
+		iir_mask |= I915_DISPLAY_PIPE_C_EVENT_INTERRUPT;
 
 	dev_priv->irq_mask |= iir_mask;
-	I915_WRITE(VLV_IER, ~dev_priv->irq_mask);
 	I915_WRITE(VLV_IMR, dev_priv->irq_mask);
+	I915_WRITE(VLV_IER, ~dev_priv->irq_mask);
 	I915_WRITE(VLV_IIR, iir_mask);
 	I915_WRITE(VLV_IIR, iir_mask);
 	POSTING_READ(VLV_IIR);
@@ -3705,14 +3427,15 @@ static void valleyview_display_irqs_uninstall(struct drm_i915_private *dev_priv)
 	pipestat_mask = PLANE_FLIP_DONE_INT_STATUS_VLV |
 			PIPE_CRC_DONE_INTERRUPT_STATUS;
 
-	i915_disable_pipestat(dev_priv, PIPE_A, pipestat_mask |
-					        PIPE_GMBUS_INTERRUPT_STATUS);
-	i915_disable_pipestat(dev_priv, PIPE_B, pipestat_mask);
+	i915_disable_pipestat(dev_priv, PIPE_A, PIPE_GMBUS_INTERRUPT_STATUS);
+	for_each_pipe(dev_priv, pipe)
+		i915_disable_pipestat(dev_priv, pipe, pipestat_mask);
 
 	pipestat_mask = PIPESTAT_INT_STATUS_MASK |
 			PIPE_FIFO_UNDERRUN_STATUS;
-	I915_WRITE(PIPESTAT(PIPE_A), pipestat_mask);
-	I915_WRITE(PIPESTAT(PIPE_B), pipestat_mask);
+
+	for_each_pipe(dev_priv, pipe)
+		I915_WRITE(PIPESTAT(pipe), pipestat_mask);
 	POSTING_READ(PIPESTAT(PIPE_A));
 }
 
@@ -3725,7 +3448,7 @@ void valleyview_enable_display_irqs(struct drm_i915_private *dev_priv)
 
 	dev_priv->display_irqs_enabled = true;
 
-	if (dev_priv->dev->irq_enabled)
+	if (intel_irqs_enabled(dev_priv))
 		valleyview_display_irqs_install(dev_priv);
 }
 
@@ -3738,34 +3461,36 @@ void valleyview_disable_display_irqs(struct drm_i915_private *dev_priv)
 
 	dev_priv->display_irqs_enabled = false;
 
-	if (dev_priv->dev->irq_enabled)
+	if (intel_irqs_enabled(dev_priv))
 		valleyview_display_irqs_uninstall(dev_priv);
 }
 
-static int valleyview_irq_postinstall(struct drm_device *dev)
+static void vlv_display_irq_postinstall(struct drm_i915_private *dev_priv)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	unsigned long irqflags;
-
 	dev_priv->irq_mask = ~0;
 
 	I915_WRITE(PORT_HOTPLUG_EN, 0);
 	POSTING_READ(PORT_HOTPLUG_EN);
 
-	I915_WRITE(VLV_IMR, dev_priv->irq_mask);
-	I915_WRITE(VLV_IER, ~dev_priv->irq_mask);
 	I915_WRITE(VLV_IIR, 0xffffffff);
-	POSTING_READ(VLV_IER);
+	I915_WRITE(VLV_IIR, 0xffffffff);
+	I915_WRITE(VLV_IER, ~dev_priv->irq_mask);
+	I915_WRITE(VLV_IMR, dev_priv->irq_mask);
+	POSTING_READ(VLV_IMR);
 
 	/* Interrupt setup is already guaranteed to be single-threaded, this is
 	 * just to make the assert_spin_locked check happy. */
-	spin_lock_irqsave(&dev_priv->irq_lock, irqflags);
+	spin_lock_irq(&dev_priv->irq_lock);
 	if (dev_priv->display_irqs_enabled)
 		valleyview_display_irqs_install(dev_priv);
-	spin_unlock_irqrestore(&dev_priv->irq_lock, irqflags);
+	spin_unlock_irq(&dev_priv->irq_lock);
+}
 
-	I915_WRITE(VLV_IIR, 0xffffffff);
-	I915_WRITE(VLV_IIR, 0xffffffff);
+static int valleyview_irq_postinstall(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+
+	vlv_display_irq_postinstall(dev_priv);
 
 	gen5_gt_irq_postinstall(dev);
 
@@ -3782,46 +3507,64 @@ static int valleyview_irq_postinstall(struct drm_device *dev)
 
 static void gen8_gt_irq_postinstall(struct drm_i915_private *dev_priv)
 {
-	int i;
-
 	/* These are interrupts we'll toggle with the ring mask register */
 	uint32_t gt_interrupts[] = {
 		GT_RENDER_USER_INTERRUPT << GEN8_RCS_IRQ_SHIFT |
+			GT_CONTEXT_SWITCH_INTERRUPT << GEN8_RCS_IRQ_SHIFT |
 			GT_RENDER_L3_PARITY_ERROR_INTERRUPT |
-			GT_RENDER_USER_INTERRUPT << GEN8_BCS_IRQ_SHIFT,
+			GT_RENDER_USER_INTERRUPT << GEN8_BCS_IRQ_SHIFT |
+			GT_CONTEXT_SWITCH_INTERRUPT << GEN8_BCS_IRQ_SHIFT,
 		GT_RENDER_USER_INTERRUPT << GEN8_VCS1_IRQ_SHIFT |
-			GT_RENDER_USER_INTERRUPT << GEN8_VCS2_IRQ_SHIFT,
+			GT_CONTEXT_SWITCH_INTERRUPT << GEN8_VCS1_IRQ_SHIFT |
+			GT_RENDER_USER_INTERRUPT << GEN8_VCS2_IRQ_SHIFT |
+			GT_CONTEXT_SWITCH_INTERRUPT << GEN8_VCS2_IRQ_SHIFT,
 		0,
-		GT_RENDER_USER_INTERRUPT << GEN8_VECS_IRQ_SHIFT
+		GT_RENDER_USER_INTERRUPT << GEN8_VECS_IRQ_SHIFT |
+			GT_CONTEXT_SWITCH_INTERRUPT << GEN8_VECS_IRQ_SHIFT
 		};
 
-	for (i = 0; i < ARRAY_SIZE(gt_interrupts); i++)
-		GEN8_IRQ_INIT_NDX(GT, i, ~gt_interrupts[i], gt_interrupts[i]);
-
 	dev_priv->pm_irq_mask = 0xffffffff;
+	GEN8_IRQ_INIT_NDX(GT, 0, ~gt_interrupts[0], gt_interrupts[0]);
+	GEN8_IRQ_INIT_NDX(GT, 1, ~gt_interrupts[1], gt_interrupts[1]);
+	/*
+	 * RPS interrupts will get enabled/disabled on demand when RPS itself
+	 * is enabled/disabled.
+	 */
+	GEN8_IRQ_INIT_NDX(GT, 2, dev_priv->pm_irq_mask, 0);
+	GEN8_IRQ_INIT_NDX(GT, 3, ~gt_interrupts[3], gt_interrupts[3]);
 }
 
 static void gen8_de_irq_postinstall(struct drm_i915_private *dev_priv)
 {
-	struct drm_device *dev = dev_priv->dev;
-	uint32_t de_pipe_masked = GEN8_PIPE_PRIMARY_FLIP_DONE |
-		GEN8_PIPE_CDCLK_CRC_DONE |
-		GEN8_DE_PIPE_IRQ_FAULT_ERRORS;
-	uint32_t de_pipe_enables = de_pipe_masked | GEN8_PIPE_VBLANK |
-		GEN8_PIPE_FIFO_UNDERRUN;
+	uint32_t de_pipe_masked = GEN8_PIPE_CDCLK_CRC_DONE;
+	uint32_t de_pipe_enables;
 	int pipe;
+	u32 aux_en = GEN8_AUX_CHANNEL_A;
+
+	if (IS_GEN9(dev_priv)) {
+		de_pipe_masked |= GEN9_PIPE_PLANE1_FLIP_DONE |
+				  GEN9_DE_PIPE_IRQ_FAULT_ERRORS;
+		aux_en |= GEN9_AUX_CHANNEL_B | GEN9_AUX_CHANNEL_C |
+			GEN9_AUX_CHANNEL_D;
+	} else
+		de_pipe_masked |= GEN8_PIPE_PRIMARY_FLIP_DONE |
+				  GEN8_DE_PIPE_IRQ_FAULT_ERRORS;
+
+	de_pipe_enables = de_pipe_masked | GEN8_PIPE_VBLANK |
+					   GEN8_PIPE_FIFO_UNDERRUN;
+
 	dev_priv->de_irq_mask[PIPE_A] = ~de_pipe_masked;
 	dev_priv->de_irq_mask[PIPE_B] = ~de_pipe_masked;
 	dev_priv->de_irq_mask[PIPE_C] = ~de_pipe_masked;
 
-	for_each_pipe(pipe)
-		if (intel_display_power_enabled(dev_priv,
+	for_each_pipe(dev_priv, pipe)
+		if (intel_display_power_is_enabled(dev_priv,
 				POWER_DOMAIN_PIPE(pipe)))
 			GEN8_IRQ_INIT_NDX(DE_PIPE, pipe,
 					  dev_priv->de_irq_mask[pipe],
 					  de_pipe_enables);
 
-	GEN5_IRQ_INIT(GEN8_DE_PORT_, ~GEN8_AUX_CHANNEL_A, GEN8_AUX_CHANNEL_A);
+	GEN5_IRQ_INIT(GEN8_DE_PORT_, ~aux_en, aux_en);
 }
 
 static int gen8_irq_postinstall(struct drm_device *dev)
@@ -3844,33 +3587,8 @@ static int gen8_irq_postinstall(struct drm_device *dev)
 static int cherryview_irq_postinstall(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	u32 enable_mask = I915_DISPLAY_PORT_INTERRUPT |
-		I915_DISPLAY_PIPE_A_EVENT_INTERRUPT |
-		I915_DISPLAY_PIPE_B_EVENT_INTERRUPT |
-		I915_DISPLAY_PIPE_C_EVENT_INTERRUPT;
-	u32 pipestat_enable = PLANE_FLIP_DONE_INT_STATUS_VLV |
-		PIPE_CRC_DONE_INTERRUPT_STATUS;
-	unsigned long irqflags;
-	int pipe;
-
-	/*
-	 * Leave vblank interrupts masked initially.  enable/disable will
-	 * toggle them based on usage.
-	 */
-	dev_priv->irq_mask = ~enable_mask;
-
-	for_each_pipe(pipe)
-		I915_WRITE(PIPESTAT(pipe), 0xffff);
 
-	spin_lock_irqsave(&dev_priv->irq_lock, irqflags);
-	i915_enable_pipestat(dev_priv, PIPE_A, PIPE_GMBUS_INTERRUPT_STATUS);
-	for_each_pipe(pipe)
-		i915_enable_pipestat(dev_priv, pipe, pipestat_enable);
-	spin_unlock_irqrestore(&dev_priv->irq_lock, irqflags);
-
-	I915_WRITE(VLV_IIR, 0xffffffff);
-	I915_WRITE(VLV_IMR, dev_priv->irq_mask);
-	I915_WRITE(VLV_IER, enable_mask);
+	vlv_display_irq_postinstall(dev_priv);
 
 	gen8_gt_irq_postinstall(dev_priv);
 
@@ -3890,41 +3608,39 @@ static void gen8_irq_uninstall(struct drm_device *dev)
 	gen8_irq_reset(dev);
 }
 
+static void vlv_display_irq_uninstall(struct drm_i915_private *dev_priv)
+{
+	/* Interrupt setup is already guaranteed to be single-threaded, this is
+	 * just to make the assert_spin_locked check happy. */
+	spin_lock_irq(&dev_priv->irq_lock);
+	if (dev_priv->display_irqs_enabled)
+		valleyview_display_irqs_uninstall(dev_priv);
+	spin_unlock_irq(&dev_priv->irq_lock);
+
+	vlv_display_irq_reset(dev_priv);
+
+	dev_priv->irq_mask = ~0;
+}
+
 static void valleyview_irq_uninstall(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	unsigned long irqflags;
-	int pipe;
 
 	if (!dev_priv)
 		return;
 
 	I915_WRITE(VLV_MASTER_IER, 0);
 
-	for_each_pipe(pipe)
-		I915_WRITE(PIPESTAT(pipe), 0xffff);
+	gen5_gt_irq_reset(dev);
 
 	I915_WRITE(HWSTAM, 0xffffffff);
-	I915_WRITE(PORT_HOTPLUG_EN, 0);
-	I915_WRITE(PORT_HOTPLUG_STAT, I915_READ(PORT_HOTPLUG_STAT));
-
-	spin_lock_irqsave(&dev_priv->irq_lock, irqflags);
-	if (dev_priv->display_irqs_enabled)
-		valleyview_display_irqs_uninstall(dev_priv);
-	spin_unlock_irqrestore(&dev_priv->irq_lock, irqflags);
 
-	dev_priv->irq_mask = 0;
-
-	I915_WRITE(VLV_IIR, 0xffffffff);
-	I915_WRITE(VLV_IMR, 0xffffffff);
-	I915_WRITE(VLV_IER, 0x0);
-	POSTING_READ(VLV_IER);
+	vlv_display_irq_uninstall(dev_priv);
 }
 
 static void cherryview_irq_uninstall(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	int pipe;
 
 	if (!dev_priv)
 		return;
@@ -3932,44 +3648,11 @@ static void cherryview_irq_uninstall(struct drm_device *dev)
 	I915_WRITE(GEN8_MASTER_IRQ, 0);
 	POSTING_READ(GEN8_MASTER_IRQ);
 
-#define GEN8_IRQ_FINI_NDX(type, which)				\
-do {								\
-	I915_WRITE(GEN8_##type##_IMR(which), 0xffffffff);	\
-	I915_WRITE(GEN8_##type##_IER(which), 0);		\
-	I915_WRITE(GEN8_##type##_IIR(which), 0xffffffff);	\
-	POSTING_READ(GEN8_##type##_IIR(which));			\
-	I915_WRITE(GEN8_##type##_IIR(which), 0xffffffff);	\
-} while (0)
-
-#define GEN8_IRQ_FINI(type)				\
-do {							\
-	I915_WRITE(GEN8_##type##_IMR, 0xffffffff);	\
-	I915_WRITE(GEN8_##type##_IER, 0);		\
-	I915_WRITE(GEN8_##type##_IIR, 0xffffffff);	\
-	POSTING_READ(GEN8_##type##_IIR);		\
-	I915_WRITE(GEN8_##type##_IIR, 0xffffffff);	\
-} while (0)
-
-	GEN8_IRQ_FINI_NDX(GT, 0);
-	GEN8_IRQ_FINI_NDX(GT, 1);
-	GEN8_IRQ_FINI_NDX(GT, 2);
-	GEN8_IRQ_FINI_NDX(GT, 3);
-
-	GEN8_IRQ_FINI(PCU);
-
-#undef GEN8_IRQ_FINI
-#undef GEN8_IRQ_FINI_NDX
-
-	I915_WRITE(PORT_HOTPLUG_EN, 0);
-	I915_WRITE(PORT_HOTPLUG_STAT, I915_READ(PORT_HOTPLUG_STAT));
+	gen8_gt_irq_reset(dev_priv);
 
-	for_each_pipe(pipe)
-		I915_WRITE(PIPESTAT(pipe), 0xffff);
+	GEN5_IRQ_RESET(GEN8_PCU_);
 
-	I915_WRITE(VLV_IMR, 0xffffffff);
-	I915_WRITE(VLV_IER, 0x0);
-	I915_WRITE(VLV_IIR, 0xffffffff);
-	POSTING_READ(VLV_IIR);
+	vlv_display_irq_uninstall(dev_priv);
 }
 
 static void ironlake_irq_uninstall(struct drm_device *dev)
@@ -3987,7 +3670,7 @@ static void i8xx_irq_preinstall(struct drm_device * dev)
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	int pipe;
 
-	for_each_pipe(pipe)
+	for_each_pipe(dev_priv, pipe)
 		I915_WRITE(PIPESTAT(pipe), 0);
 	I915_WRITE16(IMR, 0xffff);
 	I915_WRITE16(IER, 0x0);
@@ -3997,7 +3680,6 @@ static void i8xx_irq_preinstall(struct drm_device * dev)
 static int i8xx_irq_postinstall(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	unsigned long irqflags;
 
 	I915_WRITE16(EMR,
 		     ~(I915_ERROR_PAGE_TABLE | I915_ERROR_MEMORY_REFRESH));
@@ -4020,10 +3702,10 @@ static int i8xx_irq_postinstall(struct drm_device *dev)
 
 	/* Interrupt setup is already guaranteed to be single-threaded, this is
 	 * just to make the assert_spin_locked check happy. */
-	spin_lock_irqsave(&dev_priv->irq_lock, irqflags);
+	spin_lock_irq(&dev_priv->irq_lock);
 	i915_enable_pipestat(dev_priv, PIPE_A, PIPE_CRC_DONE_INTERRUPT_STATUS);
 	i915_enable_pipestat(dev_priv, PIPE_B, PIPE_CRC_DONE_INTERRUPT_STATUS);
-	spin_unlock_irqrestore(&dev_priv->irq_lock, irqflags);
+	spin_unlock_irq(&dev_priv->irq_lock);
 
 	return 0;
 }
@@ -4041,9 +3723,7 @@ static bool i8xx_handle_vblank(struct drm_device *dev,
 		return false;
 
 	if ((iir & flip_pending) == 0)
-		return false;
-
-	intel_prepare_page_flip(dev, plane);
+		goto check_page_flip;
 
 	/* We detect FlipDone by looking for the change in PendingFlip from '1'
 	 * to '0' on the following vblank, i.e. IIR has the Pendingflip
@@ -4052,11 +3732,15 @@ static bool i8xx_handle_vblank(struct drm_device *dev,
 	 * an interrupt per se, we watch for the change at vblank.
 	 */
 	if (I915_READ16(ISR) & flip_pending)
-		return false;
+		goto check_page_flip;
 
+	intel_prepare_page_flip(dev, plane);
 	intel_finish_page_flip(dev, pipe);
-
 	return true;
+
+check_page_flip:
+	intel_check_page_flip(dev, pipe);
+	return false;
 }
 
 static irqreturn_t i8xx_irq_handler(int irq, void *arg)
@@ -4065,7 +3749,6 @@ static irqreturn_t i8xx_irq_handler(int irq, void *arg)
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	u16 iir, new_iir;
 	u32 pipe_stats[2];
-	unsigned long irqflags;
 	int pipe;
 	u16 flip_mask =
 		I915_DISPLAY_PLANE_A_FLIP_PENDING_INTERRUPT |
@@ -4081,13 +3764,11 @@ static irqreturn_t i8xx_irq_handler(int irq, void *arg)
 		 * It doesn't set the bit in iir again, but it still produces
 		 * interrupts (for non-MSI).
 		 */
-		spin_lock_irqsave(&dev_priv->irq_lock, irqflags);
+		spin_lock(&dev_priv->irq_lock);
 		if (iir & I915_RENDER_COMMAND_PARSER_ERROR_INTERRUPT)
-			i915_handle_error(dev, false,
-					  "Command parser error, iir 0x%08x",
-					  iir);
+			DRM_DEBUG("Command parser error, iir 0x%08x\n", iir);
 
-		for_each_pipe(pipe) {
+		for_each_pipe(dev_priv, pipe) {
 			int reg = PIPESTAT(pipe);
 			pipe_stats[pipe] = I915_READ(reg);
 
@@ -4097,17 +3778,15 @@ static irqreturn_t i8xx_irq_handler(int irq, void *arg)
 			if (pipe_stats[pipe] & 0x8000ffff)
 				I915_WRITE(reg, pipe_stats[pipe]);
 		}
-		spin_unlock_irqrestore(&dev_priv->irq_lock, irqflags);
+		spin_unlock(&dev_priv->irq_lock);
 
 		I915_WRITE16(IIR, iir & ~flip_mask);
 		new_iir = I915_READ16(IIR); /* Flush posted writes */
 
-		i915_update_dri1_breadcrumb(dev);
-
 		if (iir & I915_USER_INTERRUPT)
 			notify_ring(dev, &dev_priv->ring[RCS]);
 
-		for_each_pipe(pipe) {
+		for_each_pipe(dev_priv, pipe) {
 			int plane = pipe;
 			if (HAS_FBC(dev))
 				plane = !plane;
@@ -4119,9 +3798,9 @@ static irqreturn_t i8xx_irq_handler(int irq, void *arg)
 			if (pipe_stats[pipe] & PIPE_CRC_DONE_INTERRUPT_STATUS)
 				i9xx_pipe_crc_irq_handler(dev, pipe);
 
-			if (pipe_stats[pipe] & PIPE_FIFO_UNDERRUN_STATUS &&
-			    intel_set_cpu_fifo_underrun_reporting(dev, pipe, false))
-				DRM_ERROR("pipe %c underrun\n", pipe_name(pipe));
+			if (pipe_stats[pipe] & PIPE_FIFO_UNDERRUN_STATUS)
+				intel_cpu_fifo_underrun_irq_handler(dev_priv,
+								    pipe);
 		}
 
 		iir = new_iir;
@@ -4135,7 +3814,7 @@ static void i8xx_irq_uninstall(struct drm_device * dev)
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	int pipe;
 
-	for_each_pipe(pipe) {
+	for_each_pipe(dev_priv, pipe) {
 		/* Clear enable bits; then clear status bits */
 		I915_WRITE(PIPESTAT(pipe), 0);
 		I915_WRITE(PIPESTAT(pipe), I915_READ(PIPESTAT(pipe)));
@@ -4156,7 +3835,7 @@ static void i915_irq_preinstall(struct drm_device * dev)
 	}
 
 	I915_WRITE16(HWSTAM, 0xeffe);
-	for_each_pipe(pipe)
+	for_each_pipe(dev_priv, pipe)
 		I915_WRITE(PIPESTAT(pipe), 0);
 	I915_WRITE(IMR, 0xffffffff);
 	I915_WRITE(IER, 0x0);
@@ -4167,7 +3846,6 @@ static int i915_irq_postinstall(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	u32 enable_mask;
-	unsigned long irqflags;
 
 	I915_WRITE(EMR, ~(I915_ERROR_PAGE_TABLE | I915_ERROR_MEMORY_REFRESH));
 
@@ -4205,10 +3883,10 @@ static int i915_irq_postinstall(struct drm_device *dev)
 
 	/* Interrupt setup is already guaranteed to be single-threaded, this is
 	 * just to make the assert_spin_locked check happy. */
-	spin_lock_irqsave(&dev_priv->irq_lock, irqflags);
+	spin_lock_irq(&dev_priv->irq_lock);
 	i915_enable_pipestat(dev_priv, PIPE_A, PIPE_CRC_DONE_INTERRUPT_STATUS);
 	i915_enable_pipestat(dev_priv, PIPE_B, PIPE_CRC_DONE_INTERRUPT_STATUS);
-	spin_unlock_irqrestore(&dev_priv->irq_lock, irqflags);
+	spin_unlock_irq(&dev_priv->irq_lock);
 
 	return 0;
 }
@@ -4226,9 +3904,7 @@ static bool i915_handle_vblank(struct drm_device *dev,
 		return false;
 
 	if ((iir & flip_pending) == 0)
-		return false;
-
-	intel_prepare_page_flip(dev, plane);
+		goto check_page_flip;
 
 	/* We detect FlipDone by looking for the change in PendingFlip from '1'
 	 * to '0' on the following vblank, i.e. IIR has the Pendingflip
@@ -4237,11 +3913,15 @@ static bool i915_handle_vblank(struct drm_device *dev,
 	 * an interrupt per se, we watch for the change at vblank.
 	 */
 	if (I915_READ(ISR) & flip_pending)
-		return false;
+		goto check_page_flip;
 
+	intel_prepare_page_flip(dev, plane);
 	intel_finish_page_flip(dev, pipe);
-
 	return true;
+
+check_page_flip:
+	intel_check_page_flip(dev, pipe);
+	return false;
 }
 
 static irqreturn_t i915_irq_handler(int irq, void *arg)
@@ -4249,7 +3929,6 @@ static irqreturn_t i915_irq_handler(int irq, void *arg)
 	struct drm_device *dev = arg;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	u32 iir, new_iir, pipe_stats[I915_MAX_PIPES];
-	unsigned long irqflags;
 	u32 flip_mask =
 		I915_DISPLAY_PLANE_A_FLIP_PENDING_INTERRUPT |
 		I915_DISPLAY_PLANE_B_FLIP_PENDING_INTERRUPT;
@@ -4265,13 +3944,11 @@ static irqreturn_t i915_irq_handler(int irq, void *arg)
 		 * It doesn't set the bit in iir again, but it still produces
 		 * interrupts (for non-MSI).
 		 */
-		spin_lock_irqsave(&dev_priv->irq_lock, irqflags);
+		spin_lock(&dev_priv->irq_lock);
 		if (iir & I915_RENDER_COMMAND_PARSER_ERROR_INTERRUPT)
-			i915_handle_error(dev, false,
-					  "Command parser error, iir 0x%08x",
-					  iir);
+			DRM_DEBUG("Command parser error, iir 0x%08x\n", iir);
 
-		for_each_pipe(pipe) {
+		for_each_pipe(dev_priv, pipe) {
 			int reg = PIPESTAT(pipe);
 			pipe_stats[pipe] = I915_READ(reg);
 
@@ -4281,7 +3958,7 @@ static irqreturn_t i915_irq_handler(int irq, void *arg)
 				irq_received = true;
 			}
 		}
-		spin_unlock_irqrestore(&dev_priv->irq_lock, irqflags);
+		spin_unlock(&dev_priv->irq_lock);
 
 		if (!irq_received)
 			break;
@@ -4297,7 +3974,7 @@ static irqreturn_t i915_irq_handler(int irq, void *arg)
 		if (iir & I915_USER_INTERRUPT)
 			notify_ring(dev, &dev_priv->ring[RCS]);
 
-		for_each_pipe(pipe) {
+		for_each_pipe(dev_priv, pipe) {
 			int plane = pipe;
 			if (HAS_FBC(dev))
 				plane = !plane;
@@ -4312,9 +3989,9 @@ static irqreturn_t i915_irq_handler(int irq, void *arg)
 			if (pipe_stats[pipe] & PIPE_CRC_DONE_INTERRUPT_STATUS)
 				i9xx_pipe_crc_irq_handler(dev, pipe);
 
-			if (pipe_stats[pipe] & PIPE_FIFO_UNDERRUN_STATUS &&
-			    intel_set_cpu_fifo_underrun_reporting(dev, pipe, false))
-				DRM_ERROR("pipe %c underrun\n", pipe_name(pipe));
+			if (pipe_stats[pipe] & PIPE_FIFO_UNDERRUN_STATUS)
+				intel_cpu_fifo_underrun_irq_handler(dev_priv,
+								    pipe);
 		}
 
 		if (blc_event || (iir & I915_ASLE_INTERRUPT))
@@ -4339,8 +4016,6 @@ static irqreturn_t i915_irq_handler(int irq, void *arg)
 		iir = new_iir;
 	} while (iir & ~flip_mask);
 
-	i915_update_dri1_breadcrumb(dev);
-
 	return ret;
 }
 
@@ -4355,7 +4030,7 @@ static void i915_irq_uninstall(struct drm_device * dev)
 	}
 
 	I915_WRITE16(HWSTAM, 0xffff);
-	for_each_pipe(pipe) {
+	for_each_pipe(dev_priv, pipe) {
 		/* Clear enable bits; then clear status bits */
 		I915_WRITE(PIPESTAT(pipe), 0);
 		I915_WRITE(PIPESTAT(pipe), I915_READ(PIPESTAT(pipe)));
@@ -4375,7 +4050,7 @@ static void i965_irq_preinstall(struct drm_device * dev)
 	I915_WRITE(PORT_HOTPLUG_STAT, I915_READ(PORT_HOTPLUG_STAT));
 
 	I915_WRITE(HWSTAM, 0xeffe);
-	for_each_pipe(pipe)
+	for_each_pipe(dev_priv, pipe)
 		I915_WRITE(PIPESTAT(pipe), 0);
 	I915_WRITE(IMR, 0xffffffff);
 	I915_WRITE(IER, 0x0);
@@ -4387,7 +4062,6 @@ static int i965_irq_postinstall(struct drm_device *dev)
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	u32 enable_mask;
 	u32 error_mask;
-	unsigned long irqflags;
 
 	/* Unmask the interrupts that we always want on. */
 	dev_priv->irq_mask = ~(I915_ASLE_INTERRUPT |
@@ -4408,11 +4082,11 @@ static int i965_irq_postinstall(struct drm_device *dev)
 
 	/* Interrupt setup is already guaranteed to be single-threaded, this is
 	 * just to make the assert_spin_locked check happy. */
-	spin_lock_irqsave(&dev_priv->irq_lock, irqflags);
+	spin_lock_irq(&dev_priv->irq_lock);
 	i915_enable_pipestat(dev_priv, PIPE_A, PIPE_GMBUS_INTERRUPT_STATUS);
 	i915_enable_pipestat(dev_priv, PIPE_A, PIPE_CRC_DONE_INTERRUPT_STATUS);
 	i915_enable_pipestat(dev_priv, PIPE_B, PIPE_CRC_DONE_INTERRUPT_STATUS);
-	spin_unlock_irqrestore(&dev_priv->irq_lock, irqflags);
+	spin_unlock_irq(&dev_priv->irq_lock);
 
 	/*
 	 * Enable some error detection, note the instruction error mask
@@ -4444,7 +4118,6 @@ static int i965_irq_postinstall(struct drm_device *dev)
 static void i915_hpd_irq_setup(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct drm_mode_config *mode_config = &dev->mode_config;
 	struct intel_encoder *intel_encoder;
 	u32 hotplug_en;
 
@@ -4455,7 +4128,7 @@ static void i915_hpd_irq_setup(struct drm_device *dev)
 		hotplug_en &= ~HOTPLUG_INT_EN_MASK;
 		/* Note HDMI and DP share hotplug bits */
 		/* enable bits are the same for all generations */
-		list_for_each_entry(intel_encoder, &mode_config->encoder_list, base.head)
+		for_each_intel_encoder(dev, intel_encoder)
 			if (dev_priv->hpd_stats[intel_encoder->hpd_pin].hpd_mark == HPD_ENABLED)
 				hotplug_en |= hpd_mask_i915[intel_encoder->hpd_pin];
 		/* Programming the CRT detection parameters tends
@@ -4478,7 +4151,6 @@ static irqreturn_t i965_irq_handler(int irq, void *arg)
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	u32 iir, new_iir;
 	u32 pipe_stats[I915_MAX_PIPES];
-	unsigned long irqflags;
 	int ret = IRQ_NONE, pipe;
 	u32 flip_mask =
 		I915_DISPLAY_PLANE_A_FLIP_PENDING_INTERRUPT |
@@ -4495,13 +4167,11 @@ static irqreturn_t i965_irq_handler(int irq, void *arg)
 		 * It doesn't set the bit in iir again, but it still produces
 		 * interrupts (for non-MSI).
 		 */
-		spin_lock_irqsave(&dev_priv->irq_lock, irqflags);
+		spin_lock(&dev_priv->irq_lock);
 		if (iir & I915_RENDER_COMMAND_PARSER_ERROR_INTERRUPT)
-			i915_handle_error(dev, false,
-					  "Command parser error, iir 0x%08x",
-					  iir);
+			DRM_DEBUG("Command parser error, iir 0x%08x\n", iir);
 
-		for_each_pipe(pipe) {
+		for_each_pipe(dev_priv, pipe) {
 			int reg = PIPESTAT(pipe);
 			pipe_stats[pipe] = I915_READ(reg);
 
@@ -4513,7 +4183,7 @@ static irqreturn_t i965_irq_handler(int irq, void *arg)
 				irq_received = true;
 			}
 		}
-		spin_unlock_irqrestore(&dev_priv->irq_lock, irqflags);
+		spin_unlock(&dev_priv->irq_lock);
 
 		if (!irq_received)
 			break;
@@ -4532,7 +4202,7 @@ static irqreturn_t i965_irq_handler(int irq, void *arg)
 		if (iir & I915_BSD_USER_INTERRUPT)
 			notify_ring(dev, &dev_priv->ring[VCS]);
 
-		for_each_pipe(pipe) {
+		for_each_pipe(dev_priv, pipe) {
 			if (pipe_stats[pipe] & PIPE_START_VBLANK_INTERRUPT_STATUS &&
 			    i915_handle_vblank(dev, pipe, pipe, iir))
 				flip_mask &= ~DISPLAY_PLANE_FLIP_PENDING(pipe);
@@ -4543,9 +4213,8 @@ static irqreturn_t i965_irq_handler(int irq, void *arg)
 			if (pipe_stats[pipe] & PIPE_CRC_DONE_INTERRUPT_STATUS)
 				i9xx_pipe_crc_irq_handler(dev, pipe);
 
-			if (pipe_stats[pipe] & PIPE_FIFO_UNDERRUN_STATUS &&
-			    intel_set_cpu_fifo_underrun_reporting(dev, pipe, false))
-				DRM_ERROR("pipe %c underrun\n", pipe_name(pipe));
+			if (pipe_stats[pipe] & PIPE_FIFO_UNDERRUN_STATUS)
+				intel_cpu_fifo_underrun_irq_handler(dev_priv, pipe);
 		}
 
 		if (blc_event || (iir & I915_ASLE_INTERRUPT))
@@ -4572,8 +4241,6 @@ static irqreturn_t i965_irq_handler(int irq, void *arg)
 		iir = new_iir;
 	}
 
-	i915_update_dri1_breadcrumb(dev);
-
 	return ret;
 }
 
@@ -4589,30 +4256,29 @@ static void i965_irq_uninstall(struct drm_device * dev)
 	I915_WRITE(PORT_HOTPLUG_STAT, I915_READ(PORT_HOTPLUG_STAT));
 
 	I915_WRITE(HWSTAM, 0xffffffff);
-	for_each_pipe(pipe)
+	for_each_pipe(dev_priv, pipe)
 		I915_WRITE(PIPESTAT(pipe), 0);
 	I915_WRITE(IMR, 0xffffffff);
 	I915_WRITE(IER, 0x0);
 
-	for_each_pipe(pipe)
+	for_each_pipe(dev_priv, pipe)
 		I915_WRITE(PIPESTAT(pipe),
 			   I915_READ(PIPESTAT(pipe)) & 0x8000ffff);
 	I915_WRITE(IIR, I915_READ(IIR));
 }
 
-static void intel_hpd_irq_reenable(struct work_struct *work)
+static void intel_hpd_irq_reenable_work(struct work_struct *work)
 {
 	struct drm_i915_private *dev_priv =
 		container_of(work, typeof(*dev_priv),
 			     hotplug_reenable_work.work);
 	struct drm_device *dev = dev_priv->dev;
 	struct drm_mode_config *mode_config = &dev->mode_config;
-	unsigned long irqflags;
 	int i;
 
 	intel_runtime_pm_get(dev_priv);
 
-	spin_lock_irqsave(&dev_priv->irq_lock, irqflags);
+	spin_lock_irq(&dev_priv->irq_lock);
 	for (i = (HPD_NONE + 1); i < HPD_NUM_PINS; i++) {
 		struct drm_connector *connector;
 
@@ -4636,14 +4302,21 @@ static void intel_hpd_irq_reenable(struct work_struct *work)
 	}
 	if (dev_priv->display.hpd_irq_setup)
 		dev_priv->display.hpd_irq_setup(dev);
-	spin_unlock_irqrestore(&dev_priv->irq_lock, irqflags);
+	spin_unlock_irq(&dev_priv->irq_lock);
 
 	intel_runtime_pm_put(dev_priv);
 }
 
-void intel_irq_init(struct drm_device *dev)
+/**
+ * intel_irq_init - initializes irq support
+ * @dev_priv: i915 device instance
+ *
+ * This function initializes all the irq support including work items, timers
+ * and all the vtables. It does not setup the interrupt itself though.
+ */
+void intel_irq_init(struct drm_i915_private *dev_priv)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_device *dev = dev_priv->dev;
 
 	INIT_WORK(&dev_priv->hotplug_work, i915_hotplug_work_func);
 	INIT_WORK(&dev_priv->dig_port_work, i915_digport_work_func);
@@ -4652,8 +4325,8 @@ void intel_irq_init(struct drm_device *dev)
 	INIT_WORK(&dev_priv->l3_parity.error_work, ivybridge_parity_work);
 
 	/* Let's track the enabled rps events */
-	if (IS_VALLEYVIEW(dev))
-		/* WaGsvRC0ResidenncyMethod:VLV */
+	if (IS_VALLEYVIEW(dev_priv) && !IS_CHERRYVIEW(dev_priv))
+		/* WaGsvRC0ResidencyMethod:vlv */
 		dev_priv->pm_rps_events = GEN6_PM_RP_UP_EI_EXPIRED;
 	else
 		dev_priv->pm_rps_events = GEN6_PM_RPS_EVENTS;
@@ -4662,17 +4335,14 @@ void intel_irq_init(struct drm_device *dev)
 		    i915_hangcheck_elapsed,
 		    (unsigned long) dev);
 	INIT_DELAYED_WORK(&dev_priv->hotplug_reenable_work,
-			  intel_hpd_irq_reenable);
+			  intel_hpd_irq_reenable_work);
 
 	pm_qos_add_request(&dev_priv->pm_qos, PM_QOS_CPU_DMA_LATENCY, PM_QOS_DEFAULT_VALUE);
 
-	/* Haven't installed the IRQ handler yet */
-	dev_priv->pm._irqs_disabled = true;
-
-	if (IS_GEN2(dev)) {
+	if (IS_GEN2(dev_priv)) {
 		dev->max_vblank_count = 0;
 		dev->driver->get_vblank_counter = i8xx_get_vblank_counter;
-	} else if (IS_G4X(dev) || INTEL_INFO(dev)->gen >= 5) {
+	} else if (IS_G4X(dev_priv) || INTEL_INFO(dev_priv)->gen >= 5) {
 		dev->max_vblank_count = 0xffffffff; /* full 32 bit counter */
 		dev->driver->get_vblank_counter = gm45_get_vblank_counter;
 	} else {
@@ -4680,12 +4350,20 @@ void intel_irq_init(struct drm_device *dev)
 		dev->max_vblank_count = 0xffffff; /* only 24 bits of frame count */
 	}
 
+	/*
+	 * Opt out of the vblank disable timer on everything except gen2.
+	 * Gen2 doesn't have a hardware frame counter and so depends on
+	 * vblank interrupts to produce sane vblank seuquence numbers.
+	 */
+	if (!IS_GEN2(dev_priv))
+		dev->vblank_disable_immediate = true;
+
 	if (drm_core_check_feature(dev, DRIVER_MODESET)) {
 		dev->driver->get_vblank_timestamp = i915_get_vblank_timestamp;
 		dev->driver->get_scanout_position = i915_get_crtc_scanoutpos;
 	}
 
-	if (IS_CHERRYVIEW(dev)) {
+	if (IS_CHERRYVIEW(dev_priv)) {
 		dev->driver->irq_handler = cherryview_irq_handler;
 		dev->driver->irq_preinstall = cherryview_irq_preinstall;
 		dev->driver->irq_postinstall = cherryview_irq_postinstall;
@@ -4693,7 +4371,7 @@ void intel_irq_init(struct drm_device *dev)
 		dev->driver->enable_vblank = valleyview_enable_vblank;
 		dev->driver->disable_vblank = valleyview_disable_vblank;
 		dev_priv->display.hpd_irq_setup = i915_hpd_irq_setup;
-	} else if (IS_VALLEYVIEW(dev)) {
+	} else if (IS_VALLEYVIEW(dev_priv)) {
 		dev->driver->irq_handler = valleyview_irq_handler;
 		dev->driver->irq_preinstall = valleyview_irq_preinstall;
 		dev->driver->irq_postinstall = valleyview_irq_postinstall;
@@ -4701,7 +4379,7 @@ void intel_irq_init(struct drm_device *dev)
 		dev->driver->enable_vblank = valleyview_enable_vblank;
 		dev->driver->disable_vblank = valleyview_disable_vblank;
 		dev_priv->display.hpd_irq_setup = i915_hpd_irq_setup;
-	} else if (IS_GEN8(dev)) {
+	} else if (INTEL_INFO(dev_priv)->gen >= 8) {
 		dev->driver->irq_handler = gen8_irq_handler;
 		dev->driver->irq_preinstall = gen8_irq_reset;
 		dev->driver->irq_postinstall = gen8_irq_postinstall;
@@ -4718,12 +4396,12 @@ void intel_irq_init(struct drm_device *dev)
 		dev->driver->disable_vblank = ironlake_disable_vblank;
 		dev_priv->display.hpd_irq_setup = ibx_hpd_irq_setup;
 	} else {
-		if (INTEL_INFO(dev)->gen == 2) {
+		if (INTEL_INFO(dev_priv)->gen == 2) {
 			dev->driver->irq_preinstall = i8xx_irq_preinstall;
 			dev->driver->irq_postinstall = i8xx_irq_postinstall;
 			dev->driver->irq_handler = i8xx_irq_handler;
 			dev->driver->irq_uninstall = i8xx_irq_uninstall;
-		} else if (INTEL_INFO(dev)->gen == 3) {
+		} else if (INTEL_INFO(dev_priv)->gen == 3) {
 			dev->driver->irq_preinstall = i915_irq_preinstall;
 			dev->driver->irq_postinstall = i915_irq_postinstall;
 			dev->driver->irq_uninstall = i915_irq_uninstall;
@@ -4741,12 +4419,23 @@ void intel_irq_init(struct drm_device *dev)
 	}
 }
 
-void intel_hpd_init(struct drm_device *dev)
+/**
+ * intel_hpd_init - initializes and enables hpd support
+ * @dev_priv: i915 device instance
+ *
+ * This function enables the hotplug support. It requires that interrupts have
+ * already been enabled with intel_irq_init_hw(). From this point on hotplug and
+ * poll request can run concurrently to other code, so locking rules must be
+ * obeyed.
+ *
+ * This is a separate step from interrupt enabling to simplify the locking rules
+ * in the driver load and resume code.
+ */
+void intel_hpd_init(struct drm_i915_private *dev_priv)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_device *dev = dev_priv->dev;
 	struct drm_mode_config *mode_config = &dev->mode_config;
 	struct drm_connector *connector;
-	unsigned long irqflags;
 	int i;
 
 	for (i = 1; i < HPD_NUM_PINS; i++) {
@@ -4764,27 +4453,72 @@ void intel_hpd_init(struct drm_device *dev)
 
 	/* Interrupt setup is already guaranteed to be single-threaded, this is
 	 * just to make the assert_spin_locked checks happy. */
-	spin_lock_irqsave(&dev_priv->irq_lock, irqflags);
+	spin_lock_irq(&dev_priv->irq_lock);
 	if (dev_priv->display.hpd_irq_setup)
 		dev_priv->display.hpd_irq_setup(dev);
-	spin_unlock_irqrestore(&dev_priv->irq_lock, irqflags);
+	spin_unlock_irq(&dev_priv->irq_lock);
 }
 
-/* Disable interrupts so we can allow runtime PM. */
-void intel_runtime_pm_disable_interrupts(struct drm_device *dev)
+/**
+ * intel_irq_install - enables the hardware interrupt
+ * @dev_priv: i915 device instance
+ *
+ * This function enables the hardware interrupt handling, but leaves the hotplug
+ * handling still disabled. It is called after intel_irq_init().
+ *
+ * In the driver load and resume code we need working interrupts in a few places
+ * but don't want to deal with the hassle of concurrent probe and hotplug
+ * workers. Hence the split into this two-stage approach.
+ */
+int intel_irq_install(struct drm_i915_private *dev_priv)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	/*
+	 * We enable some interrupt sources in our postinstall hooks, so mark
+	 * interrupts as enabled _before_ actually enabling them to avoid
+	 * special cases in our ordering checks.
+	 */
+	dev_priv->pm.irqs_enabled = true;
 
-	dev->driver->irq_uninstall(dev);
-	dev_priv->pm._irqs_disabled = true;
+	return drm_irq_install(dev_priv->dev, dev_priv->dev->pdev->irq);
 }
 
-/* Restore interrupts so we can recover from runtime PM. */
-void intel_runtime_pm_restore_interrupts(struct drm_device *dev)
+/**
+ * intel_irq_uninstall - finilizes all irq handling
+ * @dev_priv: i915 device instance
+ *
+ * This stops interrupt and hotplug handling and unregisters and frees all
+ * resources acquired in the init functions.
+ */
+void intel_irq_uninstall(struct drm_i915_private *dev_priv)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	drm_irq_uninstall(dev_priv->dev);
+	intel_hpd_cancel_work(dev_priv);
+	dev_priv->pm.irqs_enabled = false;
+}
 
-	dev_priv->pm._irqs_disabled = false;
-	dev->driver->irq_preinstall(dev);
-	dev->driver->irq_postinstall(dev);
+/**
+ * intel_runtime_pm_disable_interrupts - runtime interrupt disabling
+ * @dev_priv: i915 device instance
+ *
+ * This function is used to disable interrupts at runtime, both in the runtime
+ * pm and the system suspend/resume code.
+ */
+void intel_runtime_pm_disable_interrupts(struct drm_i915_private *dev_priv)
+{
+	dev_priv->dev->driver->irq_uninstall(dev_priv->dev);
+	dev_priv->pm.irqs_enabled = false;
+}
+
+/**
+ * intel_runtime_pm_enable_interrupts - runtime interrupt enabling
+ * @dev_priv: i915 device instance
+ *
+ * This function is used to enable interrupts at runtime, both in the runtime
+ * pm and the system suspend/resume code.
+ */
+void intel_runtime_pm_enable_interrupts(struct drm_i915_private *dev_priv)
+{
+	dev_priv->pm.irqs_enabled = true;
+	dev_priv->dev->driver->irq_preinstall(dev_priv->dev);
+	dev_priv->dev->driver->irq_postinstall(dev_priv->dev);
 }
diff --git a/drivers/gpu/drm/i915/i915_params.c b/drivers/gpu/drm/i915/i915_params.c
index 7f84dd263ee8..c91cb2033cc5 100644
--- a/drivers/gpu/drm/i915/i915_params.c
+++ b/drivers/gpu/drm/i915/i915_params.c
@@ -35,6 +35,7 @@ struct i915_params i915 __read_mostly = {
 	.vbt_sdvo_panel_type = -1,
 	.enable_rc6 = -1,
 	.enable_fbc = -1,
+	.enable_execlists = 0,
 	.enable_hangcheck = true,
 	.enable_ppgtt = -1,
 	.enable_psr = 0,
@@ -66,12 +67,12 @@ module_param_named(powersave, i915.powersave, int, 0600);
 MODULE_PARM_DESC(powersave,
 	"Enable powersavings, fbc, downclocking, etc. (default: true)");
 
-module_param_named(semaphores, i915.semaphores, int, 0400);
+module_param_named_unsafe(semaphores, i915.semaphores, int, 0400);
 MODULE_PARM_DESC(semaphores,
 	"Use semaphores for inter-ring sync "
 	"(default: -1 (use per-chip defaults))");
 
-module_param_named(enable_rc6, i915.enable_rc6, int, 0400);
+module_param_named_unsafe(enable_rc6, i915.enable_rc6, int, 0400);
 MODULE_PARM_DESC(enable_rc6,
 	"Enable power-saving render C-state 6. "
 	"Different stages can be selected via bitmask values "
@@ -79,7 +80,7 @@ MODULE_PARM_DESC(enable_rc6,
 	"For example, 3 would enable rc6 and deep rc6, and 7 would enable everything. "
 	"default: -1 (use per-chip default)");
 
-module_param_named(enable_fbc, i915.enable_fbc, int, 0600);
+module_param_named_unsafe(enable_fbc, i915.enable_fbc, int, 0600);
 MODULE_PARM_DESC(enable_fbc,
 	"Enable frame buffer compression for power savings "
 	"(default: -1 (use per-chip default))");
@@ -113,11 +114,16 @@ MODULE_PARM_DESC(enable_hangcheck,
 	"WARNING: Disabling this can cause system wide hangs. "
 	"(default: true)");
 
-module_param_named(enable_ppgtt, i915.enable_ppgtt, int, 0400);
+module_param_named_unsafe(enable_ppgtt, i915.enable_ppgtt, int, 0400);
 MODULE_PARM_DESC(enable_ppgtt,
 	"Override PPGTT usage. "
 	"(-1=auto [default], 0=disabled, 1=aliasing, 2=full)");
 
+module_param_named(enable_execlists, i915.enable_execlists, int, 0400);
+MODULE_PARM_DESC(enable_execlists,
+	"Override execlists usage. "
+	"(-1=auto, 0=disabled [default], 1=enabled)");
+
 module_param_named(enable_psr, i915.enable_psr, int, 0600);
 MODULE_PARM_DESC(enable_psr, "Enable PSR (default: false)");
 
diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h
index f29b44c86a2f..172de3b3433b 100644
--- a/drivers/gpu/drm/i915/i915_reg.h
+++ b/drivers/gpu/drm/i915/i915_reg.h
@@ -26,14 +26,25 @@
 #define _I915_REG_H_
 
 #define _PIPE(pipe, a, b) ((a) + (pipe)*((b)-(a)))
+#define _PLANE(plane, a, b) _PIPE(plane, a, b)
 #define _TRANSCODER(tran, a, b) ((a) + (tran)*((b)-(a)))
-
 #define _PORT(port, a, b) ((a) + (port)*((b)-(a)))
 #define _PIPE3(pipe, a, b, c) ((pipe) == PIPE_A ? (a) : \
 			       (pipe) == PIPE_B ? (b) : (c))
 
-#define _MASKED_BIT_ENABLE(a) (((a) << 16) | (a))
-#define _MASKED_BIT_DISABLE(a) ((a) << 16)
+#define _MASKED_FIELD(mask, value) ({					   \
+	if (__builtin_constant_p(mask))					   \
+		BUILD_BUG_ON_MSG(((mask) & 0xffff0000), "Incorrect mask"); \
+	if (__builtin_constant_p(value))				   \
+		BUILD_BUG_ON_MSG((value) & 0xffff0000, "Incorrect value"); \
+	if (__builtin_constant_p(mask) && __builtin_constant_p(value))	   \
+		BUILD_BUG_ON_MSG((value) & ~(mask),			   \
+				 "Incorrect value for mask");		   \
+	(mask) << 16 | (value); })
+#define _MASKED_BIT_ENABLE(a)	({ typeof(a) _a = (a); _MASKED_FIELD(_a, _a); })
+#define _MASKED_BIT_DISABLE(a)	(_MASKED_FIELD((a), 0))
+
+
 
 /* PCI config space */
 
@@ -74,15 +85,17 @@
 #define   I915_GC_RENDER_CLOCK_166_MHZ	(0 << 0)
 #define   I915_GC_RENDER_CLOCK_200_MHZ	(1 << 0)
 #define   I915_GC_RENDER_CLOCK_333_MHZ	(4 << 0)
+#define GCDGMBUS 0xcc
 #define PCI_LBPC 0xf4 /* legacy/combination backlight modes, also called LBB */
 
 
 /* Graphics reset regs */
-#define I965_GDRST 0xc0 /* PCI config register */
+#define I915_GDRST 0xc0 /* PCI config register */
 #define  GRDOM_FULL	(0<<2)
 #define  GRDOM_RENDER	(1<<2)
 #define  GRDOM_MEDIA	(3<<2)
 #define  GRDOM_MASK	(3<<2)
+#define  GRDOM_RESET_STATUS (1<<1)
 #define  GRDOM_RESET_ENABLE (1<<0)
 
 #define ILK_GDSR 0x2ca4 /* MCHBAR offset */
@@ -143,6 +156,14 @@
 #define GAB_CTL				0x24000
 #define   GAB_CTL_CONT_AFTER_PAGEFAULT	(1<<8)
 
+#define GEN7_BIOS_RESERVED		0x1082C0
+#define GEN7_BIOS_RESERVED_1M		(0 << 5)
+#define GEN7_BIOS_RESERVED_256K		(1 << 5)
+#define GEN8_BIOS_RESERVED_SHIFT       7
+#define GEN7_BIOS_RESERVED_MASK        0x1
+#define GEN8_BIOS_RESERVED_MASK        0x3
+
+
 /* VGA stuff */
 
 #define VGA_ST01_MDA 0x3ba
@@ -240,6 +261,16 @@
 #define   MI_DISPLAY_FLIP_IVB_SPRITE_B (3 << 19)
 #define   MI_DISPLAY_FLIP_IVB_PLANE_C  (4 << 19)
 #define   MI_DISPLAY_FLIP_IVB_SPRITE_C (5 << 19)
+/* SKL ones */
+#define   MI_DISPLAY_FLIP_SKL_PLANE_1_A	(0 << 8)
+#define   MI_DISPLAY_FLIP_SKL_PLANE_1_B	(1 << 8)
+#define   MI_DISPLAY_FLIP_SKL_PLANE_1_C	(2 << 8)
+#define   MI_DISPLAY_FLIP_SKL_PLANE_2_A	(4 << 8)
+#define   MI_DISPLAY_FLIP_SKL_PLANE_2_B	(5 << 8)
+#define   MI_DISPLAY_FLIP_SKL_PLANE_2_C	(6 << 8)
+#define   MI_DISPLAY_FLIP_SKL_PLANE_3_A	(7 << 8)
+#define   MI_DISPLAY_FLIP_SKL_PLANE_3_B	(8 << 8)
+#define   MI_DISPLAY_FLIP_SKL_PLANE_3_C	(9 << 8)
 #define MI_SEMAPHORE_MBOX	MI_INSTR(0x16, 1) /* gen6, gen7 */
 #define   MI_SEMAPHORE_GLOBAL_GTT    (1<<22)
 #define   MI_SEMAPHORE_UPDATE	    (1<<21)
@@ -272,6 +303,7 @@
 #define   MI_SEMAPHORE_POLL		(1<<15)
 #define   MI_SEMAPHORE_SAD_GTE_SDD	(1<<12)
 #define MI_STORE_DWORD_IMM	MI_INSTR(0x20, 1)
+#define MI_STORE_DWORD_IMM_GEN8	MI_INSTR(0x20, 2)
 #define   MI_MEM_VIRTUAL	(1 << 22) /* 965+ only */
 #define MI_STORE_DWORD_INDEX	MI_INSTR(0x21, 1)
 #define   MI_STORE_DWORD_INDEX_SHIFT 2
@@ -282,6 +314,7 @@
  *   address/value pairs. Don't overdue it, though, x <= 2^4 must hold!
  */
 #define MI_LOAD_REGISTER_IMM(x)	MI_INSTR(0x22, 2*(x)-1)
+#define   MI_LRI_FORCE_POSTED		(1<<12)
 #define MI_STORE_REGISTER_MEM(x) MI_INSTR(0x24, 2*(x)-1)
 #define MI_STORE_REGISTER_MEM_GEN8(x) MI_INSTR(0x24, 3*(x)-1)
 #define   MI_SRM_LRM_GLOBAL_GTT		(1<<22)
@@ -304,6 +337,8 @@
 #define   MI_BATCH_GTT		    (2<<6) /* aliased with (1<<7) on gen4 */
 #define MI_BATCH_BUFFER_START_GEN8	MI_INSTR(0x31, 1)
 
+#define MI_PREDICATE_SRC0	(0x2400)
+#define MI_PREDICATE_SRC1	(0x2408)
 
 #define MI_PREDICATE_RESULT_2	(0x2214)
 #define  LOWER_SLICE_ENABLED	(1<<0)
@@ -360,6 +395,7 @@
 #define   PIPE_CONTROL_STORE_DATA_INDEX			(1<<21)
 #define   PIPE_CONTROL_CS_STALL				(1<<20)
 #define   PIPE_CONTROL_TLB_INVALIDATE			(1<<18)
+#define   PIPE_CONTROL_MEDIA_STATE_CLEAR		(1<<16)
 #define   PIPE_CONTROL_QW_WRITE				(1<<14)
 #define   PIPE_CONTROL_POST_SYNC_OP_MASK                (3<<14)
 #define   PIPE_CONTROL_DEPTH_STALL			(1<<13)
@@ -501,10 +537,26 @@
 #define BUNIT_REG_BISOC				0x11
 
 #define PUNIT_REG_DSPFREQ			0x36
+#define   DSPFREQSTAT_SHIFT_CHV			24
+#define   DSPFREQSTAT_MASK_CHV			(0x1f << DSPFREQSTAT_SHIFT_CHV)
+#define   DSPFREQGUAR_SHIFT_CHV			8
+#define   DSPFREQGUAR_MASK_CHV			(0x1f << DSPFREQGUAR_SHIFT_CHV)
 #define   DSPFREQSTAT_SHIFT			30
 #define   DSPFREQSTAT_MASK			(0x3 << DSPFREQSTAT_SHIFT)
 #define   DSPFREQGUAR_SHIFT			14
 #define   DSPFREQGUAR_MASK			(0x3 << DSPFREQGUAR_SHIFT)
+#define   _DP_SSC(val, pipe)			((val) << (2 * (pipe)))
+#define   DP_SSC_MASK(pipe)			_DP_SSC(0x3, (pipe))
+#define   DP_SSC_PWR_ON(pipe)			_DP_SSC(0x0, (pipe))
+#define   DP_SSC_CLK_GATE(pipe)			_DP_SSC(0x1, (pipe))
+#define   DP_SSC_RESET(pipe)			_DP_SSC(0x2, (pipe))
+#define   DP_SSC_PWR_GATE(pipe)			_DP_SSC(0x3, (pipe))
+#define   _DP_SSS(val, pipe)			((val) << (2 * (pipe) + 16))
+#define   DP_SSS_MASK(pipe)			_DP_SSS(0x3, (pipe))
+#define   DP_SSS_PWR_ON(pipe)			_DP_SSS(0x0, (pipe))
+#define   DP_SSS_CLK_GATE(pipe)			_DP_SSS(0x1, (pipe))
+#define   DP_SSS_RESET(pipe)			_DP_SSS(0x2, (pipe))
+#define   DP_SSS_PWR_GATE(pipe)			_DP_SSS(0x3, (pipe))
 
 /* See the PUNIT HAS v0.8 for the below bits */
 enum punit_power_well {
@@ -518,6 +570,11 @@ enum punit_power_well {
 	PUNIT_POWER_WELL_DPIO_TX_C_LANES_23	= 9,
 	PUNIT_POWER_WELL_DPIO_RX0		= 10,
 	PUNIT_POWER_WELL_DPIO_RX1		= 11,
+	PUNIT_POWER_WELL_DPIO_CMN_D		= 12,
+	/* FIXME: guesswork below */
+	PUNIT_POWER_WELL_DPIO_TX_D_LANES_01	= 13,
+	PUNIT_POWER_WELL_DPIO_TX_D_LANES_23	= 14,
+	PUNIT_POWER_WELL_DPIO_RX2		= 15,
 
 	PUNIT_POWER_WELL_NUM,
 };
@@ -533,6 +590,7 @@ enum punit_power_well {
 #define PUNIT_REG_GPU_LFM			0xd3
 #define PUNIT_REG_GPU_FREQ_REQ			0xd4
 #define PUNIT_REG_GPU_FREQ_STS			0xd8
+#define   GPLLENABLE				(1<<4)
 #define   GENFREQSTATUS				(1<<0)
 #define PUNIT_REG_MEDIA_TURBO_FREQ_REQ		0xdc
 #define PUNIT_REG_CZ_TIMESTAMP			0xce
@@ -641,7 +699,7 @@ enum punit_power_well {
  * need to be accessed during AUX communication,
  *
  * Generally the common lane corresponds to the pipe and
- * the spline (PCS/TX) correponds to the port.
+ * the spline (PCS/TX) corresponds to the port.
  *
  * For dual channel PHY (VLV/CHV):
  *
@@ -765,6 +823,8 @@ enum punit_power_well {
 #define _VLV_PCS_DW0_CH1		0x8400
 #define   DPIO_PCS_TX_LANE2_RESET	(1<<16)
 #define   DPIO_PCS_TX_LANE1_RESET	(1<<7)
+#define   DPIO_LEFT_TXFIFO_RST_MASTER2	(1<<4)
+#define   DPIO_RIGHT_TXFIFO_RST_MASTER2	(1<<3)
 #define VLV_PCS_DW0(ch) _PORT(ch, _VLV_PCS_DW0_CH0, _VLV_PCS_DW0_CH1)
 
 #define _VLV_PCS01_DW0_CH0		0x200
@@ -805,12 +865,31 @@ enum punit_power_well {
 
 #define _VLV_PCS_DW9_CH0		0x8224
 #define _VLV_PCS_DW9_CH1		0x8424
+#define   DPIO_PCS_TX2MARGIN_MASK	(0x7<<13)
+#define   DPIO_PCS_TX2MARGIN_000	(0<<13)
+#define   DPIO_PCS_TX2MARGIN_101	(1<<13)
+#define   DPIO_PCS_TX1MARGIN_MASK	(0x7<<10)
+#define   DPIO_PCS_TX1MARGIN_000	(0<<10)
+#define   DPIO_PCS_TX1MARGIN_101	(1<<10)
 #define	VLV_PCS_DW9(ch) _PORT(ch, _VLV_PCS_DW9_CH0, _VLV_PCS_DW9_CH1)
 
+#define _VLV_PCS01_DW9_CH0		0x224
+#define _VLV_PCS23_DW9_CH0		0x424
+#define _VLV_PCS01_DW9_CH1		0x2624
+#define _VLV_PCS23_DW9_CH1		0x2824
+#define VLV_PCS01_DW9(ch) _PORT(ch, _VLV_PCS01_DW9_CH0, _VLV_PCS01_DW9_CH1)
+#define VLV_PCS23_DW9(ch) _PORT(ch, _VLV_PCS23_DW9_CH0, _VLV_PCS23_DW9_CH1)
+
 #define _CHV_PCS_DW10_CH0		0x8228
 #define _CHV_PCS_DW10_CH1		0x8428
 #define   DPIO_PCS_SWING_CALC_TX0_TX2	(1<<30)
 #define   DPIO_PCS_SWING_CALC_TX1_TX3	(1<<31)
+#define   DPIO_PCS_TX2DEEMP_MASK	(0xf<<24)
+#define   DPIO_PCS_TX2DEEMP_9P5		(0<<24)
+#define   DPIO_PCS_TX2DEEMP_6P0		(2<<24)
+#define   DPIO_PCS_TX1DEEMP_MASK	(0xf<<16)
+#define   DPIO_PCS_TX1DEEMP_9P5		(0<<16)
+#define   DPIO_PCS_TX1DEEMP_6P0		(2<<16)
 #define CHV_PCS_DW10(ch) _PORT(ch, _CHV_PCS_DW10_CH0, _CHV_PCS_DW10_CH1)
 
 #define _VLV_PCS01_DW10_CH0		0x0228
@@ -822,8 +901,18 @@ enum punit_power_well {
 
 #define _VLV_PCS_DW11_CH0		0x822c
 #define _VLV_PCS_DW11_CH1		0x842c
+#define   DPIO_LANEDESKEW_STRAP_OVRD	(1<<3)
+#define   DPIO_LEFT_TXFIFO_RST_MASTER	(1<<1)
+#define   DPIO_RIGHT_TXFIFO_RST_MASTER	(1<<0)
 #define VLV_PCS_DW11(ch) _PORT(ch, _VLV_PCS_DW11_CH0, _VLV_PCS_DW11_CH1)
 
+#define _VLV_PCS01_DW11_CH0		0x022c
+#define _VLV_PCS23_DW11_CH0		0x042c
+#define _VLV_PCS01_DW11_CH1		0x262c
+#define _VLV_PCS23_DW11_CH1		0x282c
+#define VLV_PCS01_DW11(ch) _PORT(ch, _VLV_PCS01_DW11_CH0, _VLV_PCS01_DW11_CH1)
+#define VLV_PCS23_DW11(ch) _PORT(ch, _VLV_PCS23_DW11_CH0, _VLV_PCS23_DW11_CH1)
+
 #define _VLV_PCS_DW12_CH0		0x8230
 #define _VLV_PCS_DW12_CH1		0x8430
 #define VLV_PCS_DW12(ch) _PORT(ch, _VLV_PCS_DW12_CH0, _VLV_PCS_DW12_CH1)
@@ -838,8 +927,8 @@ enum punit_power_well {
 
 #define _VLV_TX_DW2_CH0			0x8288
 #define _VLV_TX_DW2_CH1			0x8488
-#define   DPIO_SWING_MARGIN_SHIFT	16
-#define   DPIO_SWING_MARGIN_MASK	(0xff << DPIO_SWING_MARGIN_SHIFT)
+#define   DPIO_SWING_MARGIN000_SHIFT	16
+#define   DPIO_SWING_MARGIN000_MASK	(0xff << DPIO_SWING_MARGIN000_SHIFT)
 #define   DPIO_UNIQ_TRANS_SCALE_SHIFT	8
 #define VLV_TX_DW2(ch) _PORT(ch, _VLV_TX_DW2_CH0, _VLV_TX_DW2_CH1)
 
@@ -847,12 +936,16 @@ enum punit_power_well {
 #define _VLV_TX_DW3_CH1			0x848c
 /* The following bit for CHV phy */
 #define   DPIO_TX_UNIQ_TRANS_SCALE_EN	(1<<27)
+#define   DPIO_SWING_MARGIN101_SHIFT	16
+#define   DPIO_SWING_MARGIN101_MASK	(0xff << DPIO_SWING_MARGIN101_SHIFT)
 #define VLV_TX_DW3(ch) _PORT(ch, _VLV_TX_DW3_CH0, _VLV_TX_DW3_CH1)
 
 #define _VLV_TX_DW4_CH0			0x8290
 #define _VLV_TX_DW4_CH1			0x8490
 #define   DPIO_SWING_DEEMPH9P5_SHIFT	24
 #define   DPIO_SWING_DEEMPH9P5_MASK	(0xff << DPIO_SWING_DEEMPH9P5_SHIFT)
+#define   DPIO_SWING_DEEMPH6P0_SHIFT	16
+#define   DPIO_SWING_DEEMPH6P0_MASK	(0xff << DPIO_SWING_DEEMPH6P0_SHIFT)
 #define VLV_TX_DW4(ch) _PORT(ch, _VLV_TX_DW4_CH0, _VLV_TX_DW4_CH1)
 
 #define _VLV_TX3_DW4_CH0		0x690
@@ -1003,6 +1096,13 @@ enum punit_power_well {
 #define   PGTBL_ADDRESS_LO_MASK	0xfffff000 /* bits [31:12] */
 #define   PGTBL_ADDRESS_HI_MASK	0x000000f0 /* bits [35:32] (gen4) */
 #define PGTBL_ER	0x02024
+#define PRB0_BASE (0x2030-0x30)
+#define PRB1_BASE (0x2040-0x30) /* 830,gen3 */
+#define PRB2_BASE (0x2050-0x30) /* gen3 */
+#define SRB0_BASE (0x2100-0x30) /* gen2 */
+#define SRB1_BASE (0x2110-0x30) /* gen2 */
+#define SRB2_BASE (0x2120-0x30) /* 830 */
+#define SRB3_BASE (0x2130-0x30) /* 830 */
 #define RENDER_RING_BASE	0x02000
 #define BSD_RING_BASE		0x04000
 #define GEN6_BSD_RING_BASE	0x12000
@@ -1029,6 +1129,7 @@ enum punit_power_well {
 #define GEN6_VERSYNC	(RING_SYNC_1(VEBOX_RING_BASE))
 #define GEN6_VEVSYNC	(RING_SYNC_2(VEBOX_RING_BASE))
 #define GEN6_NOSYNC 0
+#define RING_PSMI_CTL(base)	((base)+0x50)
 #define RING_MAX_IDLE(base)	((base)+0x54)
 #define RING_HWS_PGA(base)	((base)+0x80)
 #define RING_HWS_PGA_GEN6(base)	((base)+0x2080)
@@ -1064,6 +1165,7 @@ enum punit_power_well {
 #define RING_ACTHD_UDW(base)	((base)+0x5c)
 #define RING_NOPID(base)	((base)+0x94)
 #define RING_IMR(base)		((base)+0xa8)
+#define RING_HWSTAM(base)	((base)+0x98)
 #define RING_TIMESTAMP(base)	((base)+0x358)
 #define   TAIL_ADDR		0x001FFFF8
 #define   HEAD_WRAP_COUNT	0xFFE00000
@@ -1194,7 +1296,7 @@ enum punit_power_well {
 #define   GEN6_WIZ_HASHING_8x8				GEN6_WIZ_HASHING(0, 0)
 #define   GEN6_WIZ_HASHING_8x4				GEN6_WIZ_HASHING(0, 1)
 #define   GEN6_WIZ_HASHING_16x4				GEN6_WIZ_HASHING(1, 0)
-#define   GEN6_WIZ_HASHING_MASK				(GEN6_WIZ_HASHING(1, 1) << 16)
+#define   GEN6_WIZ_HASHING_MASK				GEN6_WIZ_HASHING(1, 1)
 #define   GEN6_TD_FOUR_ROW_DISPATCH_DISABLE		(1 << 5)
 
 #define GFX_MODE	0x02520
@@ -1248,6 +1350,10 @@ enum punit_power_well {
 #define   INSTPM_TLB_INVALIDATE	(1<<9)
 #define   INSTPM_SYNC_FLUSH	(1<<5)
 #define ACTHD	        0x020c8
+#define MEM_MODE	0x020cc
+#define   MEM_DISPLAY_B_TRICKLE_FEED_DISABLE (1<<3) /* 830 only */
+#define   MEM_DISPLAY_A_TRICKLE_FEED_DISABLE (1<<2) /* 830/845 only */
+#define   MEM_DISPLAY_TRICKLE_FEED_DISABLE (1<<2) /* 85x only */
 #define FW_BLC		0x020d8
 #define FW_BLC2		0x020dc
 #define FW_BLC_SELF	0x020e0 /* 915+ only */
@@ -1354,6 +1460,7 @@ enum punit_power_well {
 #define   GEN6_BLITTER_FBC_NOTIFY			(1<<3)
 
 #define GEN6_RC_SLEEP_PSMI_CONTROL	0x2050
+#define   GEN6_PSMI_SLEEP_MSG_DISABLE	(1 << 0)
 #define   GEN8_RC_SEMA_IDLE_MSG_DISABLE	(1 << 12)
 #define   GEN8_FF_DOP_CLOCK_GATE_DISABLE	(1<<10)
 
@@ -1380,6 +1487,7 @@ enum punit_power_well {
 #define GT_BSD_CS_ERROR_INTERRUPT		(1 << 15)
 #define GT_BSD_USER_INTERRUPT			(1 << 12)
 #define GT_RENDER_L3_PARITY_ERROR_INTERRUPT_S1	(1 << 11) /* hsw+; rsvd on snb, ivb, vlv */
+#define GT_CONTEXT_SWITCH_INTERRUPT		(1 <<  8)
 #define GT_RENDER_L3_PARITY_ERROR_INTERRUPT	(1 <<  5) /* !snb */
 #define GT_RENDER_PIPECTL_NOTIFY_INTERRUPT	(1 <<  4)
 #define GT_RENDER_CS_MASTER_ERROR_INTERRUPT	(1 <<  3)
@@ -1519,6 +1627,7 @@ enum punit_power_well {
 /* Framebuffer compression for Ironlake */
 #define ILK_DPFC_CB_BASE	0x43200
 #define ILK_DPFC_CONTROL	0x43208
+#define   FBC_CTL_FALSE_COLOR	(1<<10)
 /* The bit 28-8 is reserved */
 #define   DPFC_RESERVED		(0x1FFFFF00)
 #define ILK_DPFC_RECOMP_CTL	0x4320c
@@ -1675,12 +1784,9 @@ enum punit_power_well {
 #define DPIO_PHY_STATUS			(VLV_DISPLAY_BASE + 0x6240)
 #define   DPLL_PORTD_READY_MASK		(0xf)
 #define DISPLAY_PHY_CONTROL (VLV_DISPLAY_BASE + 0x60100)
-#define   PHY_COM_LANE_RESET_DEASSERT(phy, val) \
-				((phy == DPIO_PHY0) ? (val | 1) : (val | 2))
-#define   PHY_COM_LANE_RESET_ASSERT(phy, val) \
-				((phy == DPIO_PHY0) ? (val & ~1) : (val & ~2))
+#define   PHY_COM_LANE_RESET_DEASSERT(phy) (1 << (phy))
 #define DISPLAY_PHY_STATUS (VLV_DISPLAY_BASE + 0x60104)
-#define   PHY_POWERGOOD(phy)	((phy == DPIO_PHY0) ? (1<<31) : (1<<30))
+#define   PHY_POWERGOOD(phy)	(((phy) == DPIO_PHY0) ? (1<<31) : (1<<30))
 
 /*
  * The i830 generation, in LVDS mode, defines P1 as the bit number set within
@@ -1953,6 +2059,8 @@ enum punit_power_well {
 #define DCC_ADDRESSING_MODE_MASK			(3 << 0)
 #define DCC_CHANNEL_XOR_DISABLE				(1 << 10)
 #define DCC_CHANNEL_XOR_BIT_17				(1 << 9)
+#define DCC2			0x10204
+#define DCC2_MODIFIED_ENHANCED_DISABLE			(1 << 20)
 
 /* Pineview MCH register contains DDR3 setting */
 #define CSHRDDR3CTL            0x101a8
@@ -2236,7 +2344,6 @@ enum punit_power_well {
 
 #define GEN6_GT_THREAD_STATUS_REG 0x13805c
 #define GEN6_GT_THREAD_STATUS_CORE_MASK 0x7
-#define GEN6_GT_THREAD_STATUS_CORE_MASK_HSW (0x7 | (0x07 << 16))
 
 #define GEN6_GT_PERF_STATUS	(MCHBAR_MIRROR_BASE_SNB + 0x5948)
 #define GEN6_RP_STATE_LIMITS	(MCHBAR_MIRROR_BASE_SNB + 0x5994)
@@ -2397,6 +2504,7 @@ enum punit_power_well {
 #define _PIPEASRC	0x6001c
 #define _BCLRPAT_A	0x60020
 #define _VSYNCSHIFT_A	0x60028
+#define _PIPE_MULT_A	0x6002c
 
 /* Pipe B timing regs */
 #define _HTOTAL_B	0x61000
@@ -2408,6 +2516,7 @@ enum punit_power_well {
 #define _PIPEBSRC	0x6101c
 #define _BCLRPAT_B	0x61020
 #define _VSYNCSHIFT_B	0x61028
+#define _PIPE_MULT_B	0x6102c
 
 #define TRANSCODER_A_OFFSET 0x60000
 #define TRANSCODER_B_OFFSET 0x61000
@@ -2428,6 +2537,7 @@ enum punit_power_well {
 #define BCLRPAT(trans) _TRANSCODER2(trans, _BCLRPAT_A)
 #define VSYNCSHIFT(trans) _TRANSCODER2(trans, _VSYNCSHIFT_A)
 #define PIPESRC(trans) _TRANSCODER2(trans, _PIPEASRC)
+#define PIPE_MULT(trans) _TRANSCODER2(trans, _PIPE_MULT_A)
 
 /* HSW+ eDP PSR registers */
 #define EDP_PSR_BASE(dev)                       (IS_HASWELL(dev) ? 0x64800 : 0x6f800)
@@ -2457,9 +2567,7 @@ enum punit_power_well {
 
 #define EDP_PSR_AUX_CTL(dev)			(EDP_PSR_BASE(dev) + 0x10)
 #define EDP_PSR_AUX_DATA1(dev)			(EDP_PSR_BASE(dev) + 0x14)
-#define   EDP_PSR_DPCD_COMMAND		0x80060000
 #define EDP_PSR_AUX_DATA2(dev)			(EDP_PSR_BASE(dev) + 0x18)
-#define   EDP_PSR_DPCD_NORMAL_OPERATION	(1<<24)
 #define EDP_PSR_AUX_DATA3(dev)			(EDP_PSR_BASE(dev) + 0x1c)
 #define EDP_PSR_AUX_DATA4(dev)			(EDP_PSR_BASE(dev) + 0x20)
 #define EDP_PSR_AUX_DATA5(dev)			(EDP_PSR_BASE(dev) + 0x24)
@@ -3476,6 +3584,8 @@ enum punit_power_well {
 #define   DP_LINK_TRAIN_OFF		(3 << 28)
 #define   DP_LINK_TRAIN_MASK		(3 << 28)
 #define   DP_LINK_TRAIN_SHIFT		28
+#define   DP_LINK_TRAIN_PAT_3_CHV	(1 << 14)
+#define   DP_LINK_TRAIN_MASK_CHV	((3 << 28)|(1<<14))
 
 /* CPT Link training mode */
 #define   DP_LINK_TRAIN_PAT_1_CPT	(0 << 8)
@@ -3594,6 +3704,7 @@ enum punit_power_well {
 #define   DP_AUX_CH_CTL_PRECHARGE_TEST	    (1 << 11)
 #define   DP_AUX_CH_CTL_BIT_CLOCK_2X_MASK    (0x7ff)
 #define   DP_AUX_CH_CTL_BIT_CLOCK_2X_SHIFT   0
+#define   DP_AUX_CH_CTL_SYNC_PULSE_SKL(c)   ((c) - 1)
 
 /*
  * Computing GMCH M and N values for the Display Port link
@@ -3732,7 +3843,6 @@ enum punit_power_well {
 #define   PIPE_VSYNC_INTERRUPT_STATUS		(1UL<<9)
 #define   PIPE_DISPLAY_LINE_COMPARE_STATUS	(1UL<<8)
 #define   PIPE_DPST_EVENT_STATUS		(1UL<<7)
-#define   PIPE_LEGACY_BLC_EVENT_STATUS		(1UL<<6)
 #define   PIPE_A_PSR_STATUS_VLV			(1UL<<6)
 #define   PIPE_LEGACY_BLC_EVENT_STATUS		(1UL<<6)
 #define   PIPE_ODD_FIELD_INTERRUPT_STATUS	(1UL<<5)
@@ -3842,73 +3952,152 @@ enum punit_power_well {
 #define   DSPARB_BEND_SHIFT	9 /* on 855 */
 #define   DSPARB_AEND_SHIFT	0
 
+/* pnv/gen4/g4x/vlv/chv */
 #define DSPFW1			(dev_priv->info.display_mmio_offset + 0x70034)
-#define   DSPFW_SR_SHIFT	23
-#define   DSPFW_SR_MASK		(0x1ff<<23)
-#define   DSPFW_CURSORB_SHIFT	16
-#define   DSPFW_CURSORB_MASK	(0x3f<<16)
-#define   DSPFW_PLANEB_SHIFT	8
-#define   DSPFW_PLANEB_MASK	(0x7f<<8)
-#define   DSPFW_PLANEA_MASK	(0x7f)
+#define   DSPFW_SR_SHIFT		23
+#define   DSPFW_SR_MASK			(0x1ff<<23)
+#define   DSPFW_CURSORB_SHIFT		16
+#define   DSPFW_CURSORB_MASK		(0x3f<<16)
+#define   DSPFW_PLANEB_SHIFT		8
+#define   DSPFW_PLANEB_MASK		(0x7f<<8)
+#define   DSPFW_PLANEB_MASK_VLV		(0xff<<8) /* vlv/chv */
+#define   DSPFW_PLANEA_SHIFT		0
+#define   DSPFW_PLANEA_MASK		(0x7f<<0)
+#define   DSPFW_PLANEA_MASK_VLV		(0xff<<0) /* vlv/chv */
 #define DSPFW2			(dev_priv->info.display_mmio_offset + 0x70038)
-#define   DSPFW_CURSORA_MASK	0x00003f00
-#define   DSPFW_CURSORA_SHIFT	8
-#define   DSPFW_PLANEC_MASK	(0x7f)
+#define   DSPFW_FBC_SR_EN		(1<<31)	  /* g4x */
+#define   DSPFW_FBC_SR_SHIFT		28
+#define   DSPFW_FBC_SR_MASK		(0x7<<28) /* g4x */
+#define   DSPFW_FBC_HPLL_SR_SHIFT	24
+#define   DSPFW_FBC_HPLL_SR_MASK	(0xf<<24) /* g4x */
+#define   DSPFW_SPRITEB_SHIFT		(16)
+#define   DSPFW_SPRITEB_MASK		(0x7f<<16) /* g4x */
+#define   DSPFW_SPRITEB_MASK_VLV	(0xff<<16) /* vlv/chv */
+#define   DSPFW_CURSORA_SHIFT		8
+#define   DSPFW_CURSORA_MASK		(0x3f<<8)
+#define   DSPFW_PLANEC_SHIFT_OLD	0
+#define   DSPFW_PLANEC_MASK_OLD		(0x7f<<0) /* pre-gen4 sprite C */
+#define   DSPFW_SPRITEA_SHIFT		0
+#define   DSPFW_SPRITEA_MASK		(0x7f<<0) /* g4x */
+#define   DSPFW_SPRITEA_MASK_VLV	(0xff<<0) /* vlv/chv */
 #define DSPFW3			(dev_priv->info.display_mmio_offset + 0x7003c)
-#define   DSPFW_HPLL_SR_EN	(1<<31)
-#define   DSPFW_CURSOR_SR_SHIFT	24
+#define   DSPFW_HPLL_SR_EN		(1<<31)
 #define   PINEVIEW_SELF_REFRESH_EN	(1<<30)
+#define   DSPFW_CURSOR_SR_SHIFT		24
 #define   DSPFW_CURSOR_SR_MASK		(0x3f<<24)
 #define   DSPFW_HPLL_CURSOR_SHIFT	16
 #define   DSPFW_HPLL_CURSOR_MASK	(0x3f<<16)
-#define   DSPFW_HPLL_SR_MASK		(0x1ff)
-#define DSPFW4			(dev_priv->info.display_mmio_offset + 0x70070)
-#define DSPFW7			(dev_priv->info.display_mmio_offset + 0x7007c)
+#define   DSPFW_HPLL_SR_SHIFT		0
+#define   DSPFW_HPLL_SR_MASK		(0x1ff<<0)
+
+/* vlv/chv */
+#define DSPFW4			(VLV_DISPLAY_BASE + 0x70070)
+#define   DSPFW_SPRITEB_WM1_SHIFT	16
+#define   DSPFW_SPRITEB_WM1_MASK	(0xff<<16)
+#define   DSPFW_CURSORA_WM1_SHIFT	8
+#define   DSPFW_CURSORA_WM1_MASK	(0x3f<<8)
+#define   DSPFW_SPRITEA_WM1_SHIFT	0
+#define   DSPFW_SPRITEA_WM1_MASK	(0xff<<0)
+#define DSPFW5			(VLV_DISPLAY_BASE + 0x70074)
+#define   DSPFW_PLANEB_WM1_SHIFT	24
+#define   DSPFW_PLANEB_WM1_MASK		(0xff<<24)
+#define   DSPFW_PLANEA_WM1_SHIFT	16
+#define   DSPFW_PLANEA_WM1_MASK		(0xff<<16)
+#define   DSPFW_CURSORB_WM1_SHIFT	8
+#define   DSPFW_CURSORB_WM1_MASK	(0x3f<<8)
+#define   DSPFW_CURSOR_SR_WM1_SHIFT	0
+#define   DSPFW_CURSOR_SR_WM1_MASK	(0x3f<<0)
+#define DSPFW6			(VLV_DISPLAY_BASE + 0x70078)
+#define   DSPFW_SR_WM1_SHIFT		0
+#define   DSPFW_SR_WM1_MASK		(0x1ff<<0)
+#define DSPFW7			(VLV_DISPLAY_BASE + 0x7007c)
+#define DSPFW7_CHV		(VLV_DISPLAY_BASE + 0x700b4) /* wtf #1? */
+#define   DSPFW_SPRITED_WM1_SHIFT	24
+#define   DSPFW_SPRITED_WM1_MASK	(0xff<<24)
+#define   DSPFW_SPRITED_SHIFT		16
+#define   DSPFW_SPRITED_MASK		(0xff<<16)
+#define   DSPFW_SPRITEC_WM1_SHIFT	8
+#define   DSPFW_SPRITEC_WM1_MASK	(0xff<<8)
+#define   DSPFW_SPRITEC_SHIFT		0
+#define   DSPFW_SPRITEC_MASK		(0xff<<0)
+#define DSPFW8_CHV		(VLV_DISPLAY_BASE + 0x700b8)
+#define   DSPFW_SPRITEF_WM1_SHIFT	24
+#define   DSPFW_SPRITEF_WM1_MASK	(0xff<<24)
+#define   DSPFW_SPRITEF_SHIFT		16
+#define   DSPFW_SPRITEF_MASK		(0xff<<16)
+#define   DSPFW_SPRITEE_WM1_SHIFT	8
+#define   DSPFW_SPRITEE_WM1_MASK	(0xff<<8)
+#define   DSPFW_SPRITEE_SHIFT		0
+#define   DSPFW_SPRITEE_MASK		(0xff<<0)
+#define DSPFW9_CHV		(VLV_DISPLAY_BASE + 0x7007c) /* wtf #2? */
+#define   DSPFW_PLANEC_WM1_SHIFT	24
+#define   DSPFW_PLANEC_WM1_MASK		(0xff<<24)
+#define   DSPFW_PLANEC_SHIFT		16
+#define   DSPFW_PLANEC_MASK		(0xff<<16)
+#define   DSPFW_CURSORC_WM1_SHIFT	8
+#define   DSPFW_CURSORC_WM1_MASK	(0x3f<<16)
+#define   DSPFW_CURSORC_SHIFT		0
+#define   DSPFW_CURSORC_MASK		(0x3f<<0)
+
+/* vlv/chv high order bits */
+#define DSPHOWM			(VLV_DISPLAY_BASE + 0x70064)
+#define   DSPFW_SR_HI_SHIFT		24
+#define   DSPFW_SR_HI_MASK		(1<<24)
+#define   DSPFW_SPRITEF_HI_SHIFT	23
+#define   DSPFW_SPRITEF_HI_MASK		(1<<23)
+#define   DSPFW_SPRITEE_HI_SHIFT	22
+#define   DSPFW_SPRITEE_HI_MASK		(1<<22)
+#define   DSPFW_PLANEC_HI_SHIFT		21
+#define   DSPFW_PLANEC_HI_MASK		(1<<21)
+#define   DSPFW_SPRITED_HI_SHIFT	20
+#define   DSPFW_SPRITED_HI_MASK		(1<<20)
+#define   DSPFW_SPRITEC_HI_SHIFT	16
+#define   DSPFW_SPRITEC_HI_MASK		(1<<16)
+#define   DSPFW_PLANEB_HI_SHIFT		12
+#define   DSPFW_PLANEB_HI_MASK		(1<<12)
+#define   DSPFW_SPRITEB_HI_SHIFT	8
+#define   DSPFW_SPRITEB_HI_MASK		(1<<8)
+#define   DSPFW_SPRITEA_HI_SHIFT	4
+#define   DSPFW_SPRITEA_HI_MASK		(1<<4)
+#define   DSPFW_PLANEA_HI_SHIFT		0
+#define   DSPFW_PLANEA_HI_MASK		(1<<0)
+#define DSPHOWM1		(VLV_DISPLAY_BASE + 0x70068)
+#define   DSPFW_SR_WM1_HI_SHIFT		24
+#define   DSPFW_SR_WM1_HI_MASK		(1<<24)
+#define   DSPFW_SPRITEF_WM1_HI_SHIFT	23
+#define   DSPFW_SPRITEF_WM1_HI_MASK	(1<<23)
+#define   DSPFW_SPRITEE_WM1_HI_SHIFT	22
+#define   DSPFW_SPRITEE_WM1_HI_MASK	(1<<22)
+#define   DSPFW_PLANEC_WM1_HI_SHIFT	21
+#define   DSPFW_PLANEC_WM1_HI_MASK	(1<<21)
+#define   DSPFW_SPRITED_WM1_HI_SHIFT	20
+#define   DSPFW_SPRITED_WM1_HI_MASK	(1<<20)
+#define   DSPFW_SPRITEC_WM1_HI_SHIFT	16
+#define   DSPFW_SPRITEC_WM1_HI_MASK	(1<<16)
+#define   DSPFW_PLANEB_WM1_HI_SHIFT	12
+#define   DSPFW_PLANEB_WM1_HI_MASK	(1<<12)
+#define   DSPFW_SPRITEB_WM1_HI_SHIFT	8
+#define   DSPFW_SPRITEB_WM1_HI_MASK	(1<<8)
+#define   DSPFW_SPRITEA_WM1_HI_SHIFT	4
+#define   DSPFW_SPRITEA_WM1_HI_MASK	(1<<4)
+#define   DSPFW_PLANEA_WM1_HI_SHIFT	0
+#define   DSPFW_PLANEA_WM1_HI_MASK	(1<<0)
 
 /* drain latency register values*/
+#define DRAIN_LATENCY_PRECISION_16	16
 #define DRAIN_LATENCY_PRECISION_32	32
 #define DRAIN_LATENCY_PRECISION_64	64
-#define VLV_DDL1			(VLV_DISPLAY_BASE + 0x70050)
-#define DDL_CURSORA_PRECISION_64	(1<<31)
-#define DDL_CURSORA_PRECISION_32	(0<<31)
-#define DDL_CURSORA_SHIFT		24
-#define DDL_SPRITEB_PRECISION_64	(1<<23)
-#define DDL_SPRITEB_PRECISION_32	(0<<23)
-#define DDL_SPRITEB_SHIFT		16
-#define DDL_SPRITEA_PRECISION_64	(1<<15)
-#define DDL_SPRITEA_PRECISION_32	(0<<15)
-#define DDL_SPRITEA_SHIFT		8
-#define DDL_PLANEA_PRECISION_64		(1<<7)
-#define DDL_PLANEA_PRECISION_32		(0<<7)
-#define DDL_PLANEA_SHIFT		0
-
-#define VLV_DDL2			(VLV_DISPLAY_BASE + 0x70054)
-#define DDL_CURSORB_PRECISION_64	(1<<31)
-#define DDL_CURSORB_PRECISION_32	(0<<31)
-#define DDL_CURSORB_SHIFT		24
-#define DDL_SPRITED_PRECISION_64	(1<<23)
-#define DDL_SPRITED_PRECISION_32	(0<<23)
-#define DDL_SPRITED_SHIFT		16
-#define DDL_SPRITEC_PRECISION_64	(1<<15)
-#define DDL_SPRITEC_PRECISION_32	(0<<15)
-#define DDL_SPRITEC_SHIFT		8
-#define DDL_PLANEB_PRECISION_64		(1<<7)
-#define DDL_PLANEB_PRECISION_32		(0<<7)
-#define DDL_PLANEB_SHIFT		0
-
-#define VLV_DDL3			(VLV_DISPLAY_BASE + 0x70058)
-#define DDL_CURSORC_PRECISION_64	(1<<31)
-#define DDL_CURSORC_PRECISION_32	(0<<31)
-#define DDL_CURSORC_SHIFT		24
-#define DDL_SPRITEF_PRECISION_64	(1<<23)
-#define DDL_SPRITEF_PRECISION_32	(0<<23)
-#define DDL_SPRITEF_SHIFT		16
-#define DDL_SPRITEE_PRECISION_64	(1<<15)
-#define DDL_SPRITEE_PRECISION_32	(0<<15)
-#define DDL_SPRITEE_SHIFT		8
-#define DDL_PLANEC_PRECISION_64		(1<<7)
-#define DDL_PLANEC_PRECISION_32		(0<<7)
-#define DDL_PLANEC_SHIFT		0
+#define VLV_DDL(pipe)			(VLV_DISPLAY_BASE + 0x70050 + 4 * (pipe))
+#define DDL_CURSOR_PRECISION_HIGH	(1<<31)
+#define DDL_CURSOR_PRECISION_LOW	(0<<31)
+#define DDL_CURSOR_SHIFT		24
+#define DDL_SPRITE_PRECISION_HIGH(sprite)	(1<<(15+8*(sprite)))
+#define DDL_SPRITE_PRECISION_LOW(sprite)	(0<<(15+8*(sprite)))
+#define DDL_SPRITE_SHIFT(sprite)	(8+8*(sprite))
+#define DDL_PLANE_PRECISION_HIGH	(1<<7)
+#define DDL_PLANE_PRECISION_LOW		(0<<7)
+#define DDL_PLANE_SHIFT			0
+#define DRAIN_LATENCY_MASK		0x7f
 
 /* FIFO watermark sizes etc */
 #define G4X_FIFO_LINE_SIZE	64
@@ -3943,6 +4132,41 @@ enum punit_power_well {
 #define I965_CURSOR_MAX_WM	32
 #define I965_CURSOR_DFT_WM	8
 
+/* Watermark register definitions for SKL */
+#define CUR_WM_A_0		0x70140
+#define CUR_WM_B_0		0x71140
+#define PLANE_WM_1_A_0		0x70240
+#define PLANE_WM_1_B_0		0x71240
+#define PLANE_WM_2_A_0		0x70340
+#define PLANE_WM_2_B_0		0x71340
+#define PLANE_WM_TRANS_1_A_0	0x70268
+#define PLANE_WM_TRANS_1_B_0	0x71268
+#define PLANE_WM_TRANS_2_A_0	0x70368
+#define PLANE_WM_TRANS_2_B_0	0x71368
+#define CUR_WM_TRANS_A_0	0x70168
+#define CUR_WM_TRANS_B_0	0x71168
+#define   PLANE_WM_EN		(1 << 31)
+#define   PLANE_WM_LINES_SHIFT	14
+#define   PLANE_WM_LINES_MASK	0x1f
+#define   PLANE_WM_BLOCKS_MASK	0x3ff
+
+#define CUR_WM_0(pipe) _PIPE(pipe, CUR_WM_A_0, CUR_WM_B_0)
+#define CUR_WM(pipe, level) (CUR_WM_0(pipe) + ((4) * (level)))
+#define CUR_WM_TRANS(pipe) _PIPE(pipe, CUR_WM_TRANS_A_0, CUR_WM_TRANS_B_0)
+
+#define _PLANE_WM_1(pipe) _PIPE(pipe, PLANE_WM_1_A_0, PLANE_WM_1_B_0)
+#define _PLANE_WM_2(pipe) _PIPE(pipe, PLANE_WM_2_A_0, PLANE_WM_2_B_0)
+#define _PLANE_WM_BASE(pipe, plane)	\
+			_PLANE(plane, _PLANE_WM_1(pipe), _PLANE_WM_2(pipe))
+#define PLANE_WM(pipe, plane, level)	\
+			(_PLANE_WM_BASE(pipe, plane) + ((4) * (level)))
+#define _PLANE_WM_TRANS_1(pipe)	\
+			_PIPE(pipe, PLANE_WM_TRANS_1_A_0, PLANE_WM_TRANS_1_B_0)
+#define _PLANE_WM_TRANS_2(pipe)	\
+			_PIPE(pipe, PLANE_WM_TRANS_2_A_0, PLANE_WM_TRANS_2_B_0)
+#define PLANE_WM_TRANS(pipe, plane)	\
+		_PLANE(plane, _PLANE_WM_TRANS_1(pipe), _PLANE_WM_TRANS_2(pipe))
+
 /* define the Watermark register on Ironlake */
 #define WM0_PIPEA_ILK		0x45100
 #define  WM0_PIPE_PLANE_MASK	(0xffff<<16)
@@ -4026,7 +4250,8 @@ enum punit_power_well {
 /* Old style CUR*CNTR flags (desktop 8xx) */
 #define   CURSOR_ENABLE		0x80000000
 #define   CURSOR_GAMMA_ENABLE	0x40000000
-#define   CURSOR_STRIDE_MASK	0x30000000
+#define   CURSOR_STRIDE_SHIFT	28
+#define   CURSOR_STRIDE(x)	((ffs(x)-9) << CURSOR_STRIDE_SHIFT) /* 256,512,1k,2k */
 #define   CURSOR_PIPE_CSC_ENABLE (1<<24)
 #define   CURSOR_FORMAT_SHIFT	24
 #define   CURSOR_FORMAT_MASK	(0x07 << CURSOR_FORMAT_SHIFT)
@@ -4048,6 +4273,7 @@ enum punit_power_well {
 #define   MCURSOR_PIPE_A	0x00
 #define   MCURSOR_PIPE_B	(1 << 28)
 #define   MCURSOR_GAMMA_ENABLE  (1 << 26)
+#define   CURSOR_ROTATE_180	(1<<15)
 #define   CURSOR_TRICKLE_FEED_DISABLE	(1 << 14)
 #define _CURABASE		0x70084
 #define _CURAPOS		0x70088
@@ -4111,8 +4337,11 @@ enum punit_power_well {
 #define   DISPPLANE_NO_LINE_DOUBLE		0
 #define   DISPPLANE_STEREO_POLARITY_FIRST	0
 #define   DISPPLANE_STEREO_POLARITY_SECOND	(1<<18)
+#define   DISPPLANE_ALPHA_PREMULTIPLY		(1<<16) /* CHV pipe B */
+#define   DISPPLANE_ROTATE_180			(1<<15)
 #define   DISPPLANE_TRICKLE_FEED_DISABLE	(1<<14) /* Ironlake */
 #define   DISPPLANE_TILED			(1<<10)
+#define   DISPPLANE_MIRROR			(1<<8) /* CHV pipe B */
 #define _DSPAADDR				0x70184
 #define _DSPASTRIDE				0x70188
 #define _DSPAPOS				0x7018C /* reserved */
@@ -4133,6 +4362,24 @@ enum punit_power_well {
 #define DSPOFFSET(plane) _PIPE2(plane, _DSPAOFFSET)
 #define DSPSURFLIVE(plane) _PIPE2(plane, _DSPASURFLIVE)
 
+/* CHV pipe B blender and primary plane */
+#define _CHV_BLEND_A		0x60a00
+#define   CHV_BLEND_LEGACY		(0<<30)
+#define   CHV_BLEND_ANDROID		(1<<30)
+#define   CHV_BLEND_MPO			(2<<30)
+#define   CHV_BLEND_MASK		(3<<30)
+#define _CHV_CANVAS_A		0x60a04
+#define _PRIMPOS_A		0x60a08
+#define _PRIMSIZE_A		0x60a0c
+#define _PRIMCNSTALPHA_A	0x60a10
+#define   PRIM_CONST_ALPHA_ENABLE	(1<<31)
+
+#define CHV_BLEND(pipe) _TRANSCODER2(pipe, _CHV_BLEND_A)
+#define CHV_CANVAS(pipe) _TRANSCODER2(pipe, _CHV_CANVAS_A)
+#define PRIMPOS(plane) _TRANSCODER2(plane, _PRIMPOS_A)
+#define PRIMSIZE(plane) _TRANSCODER2(plane, _PRIMSIZE_A)
+#define PRIMCNSTALPHA(plane) _TRANSCODER2(plane, _PRIMCNSTALPHA_A)
+
 /* Display/Sprite base address macros */
 #define DISP_BASEADDR_MASK	(0xfffff000)
 #define I915_LO_DISPBASE(val)	(val & ~DISP_BASEADDR_MASK)
@@ -4195,6 +4442,7 @@ enum punit_power_well {
 #define   DVS_YUV_ORDER_UYVY	(1<<16)
 #define   DVS_YUV_ORDER_YVYU	(2<<16)
 #define   DVS_YUV_ORDER_VYUY	(3<<16)
+#define   DVS_ROTATE_180	(1<<15)
 #define   DVS_DEST_KEY		(1<<2)
 #define   DVS_TRICKLE_FEED_DISABLE (1<<14)
 #define   DVS_TILED		(1<<10)
@@ -4265,6 +4513,7 @@ enum punit_power_well {
 #define   SPRITE_YUV_ORDER_UYVY		(1<<16)
 #define   SPRITE_YUV_ORDER_YVYU		(2<<16)
 #define   SPRITE_YUV_ORDER_VYUY		(3<<16)
+#define   SPRITE_ROTATE_180		(1<<15)
 #define   SPRITE_TRICKLE_FEED_DISABLE	(1<<14)
 #define   SPRITE_INT_GAMMA_ENABLE	(1<<13)
 #define   SPRITE_TILED			(1<<10)
@@ -4332,13 +4581,16 @@ enum punit_power_well {
 #define   SP_FORMAT_RGBA1010102		(9<<26)
 #define   SP_FORMAT_RGBX8888		(0xe<<26)
 #define   SP_FORMAT_RGBA8888		(0xf<<26)
+#define   SP_ALPHA_PREMULTIPLY		(1<<23) /* CHV pipe B */
 #define   SP_SOURCE_KEY			(1<<22)
 #define   SP_YUV_BYTE_ORDER_MASK	(3<<16)
 #define   SP_YUV_ORDER_YUYV		(0<<16)
 #define   SP_YUV_ORDER_UYVY		(1<<16)
 #define   SP_YUV_ORDER_YVYU		(2<<16)
 #define   SP_YUV_ORDER_VYUY		(3<<16)
+#define   SP_ROTATE_180			(1<<15)
 #define   SP_TILED			(1<<10)
+#define   SP_MIRROR			(1<<8) /* CHV pipe B */
 #define _SPALINOFF		(VLV_DISPLAY_BASE + 0x72184)
 #define _SPASTRIDE		(VLV_DISPLAY_BASE + 0x72188)
 #define _SPAPOS			(VLV_DISPLAY_BASE + 0x7218c)
@@ -4349,6 +4601,7 @@ enum punit_power_well {
 #define _SPAKEYMAXVAL		(VLV_DISPLAY_BASE + 0x721a0)
 #define _SPATILEOFF		(VLV_DISPLAY_BASE + 0x721a4)
 #define _SPACONSTALPHA		(VLV_DISPLAY_BASE + 0x721a8)
+#define   SP_CONST_ALPHA_ENABLE		(1<<31)
 #define _SPAGAMC		(VLV_DISPLAY_BASE + 0x721f4)
 
 #define _SPBCNTR		(VLV_DISPLAY_BASE + 0x72280)
@@ -4377,6 +4630,195 @@ enum punit_power_well {
 #define SPCONSTALPHA(pipe, plane) _PIPE(pipe * 2 + plane, _SPACONSTALPHA, _SPBCONSTALPHA)
 #define SPGAMC(pipe, plane) _PIPE(pipe * 2 + plane, _SPAGAMC, _SPBGAMC)
 
+/*
+ * CHV pipe B sprite CSC
+ *
+ * |cr|   |c0 c1 c2|   |cr + cr_ioff|   |cr_ooff|
+ * |yg| = |c3 c4 c5| x |yg + yg_ioff| + |yg_ooff|
+ * |cb|   |c6 c7 c8|   |cb + cr_ioff|   |cb_ooff|
+ */
+#define SPCSCYGOFF(sprite)	(VLV_DISPLAY_BASE + 0x6d900 + (sprite) * 0x1000)
+#define SPCSCCBOFF(sprite)	(VLV_DISPLAY_BASE + 0x6d904 + (sprite) * 0x1000)
+#define SPCSCCROFF(sprite)	(VLV_DISPLAY_BASE + 0x6d908 + (sprite) * 0x1000)
+#define  SPCSC_OOFF(x)		(((x) & 0x7ff) << 16) /* s11 */
+#define  SPCSC_IOFF(x)		(((x) & 0x7ff) << 0) /* s11 */
+
+#define SPCSCC01(sprite)	(VLV_DISPLAY_BASE + 0x6d90c + (sprite) * 0x1000)
+#define SPCSCC23(sprite)	(VLV_DISPLAY_BASE + 0x6d910 + (sprite) * 0x1000)
+#define SPCSCC45(sprite)	(VLV_DISPLAY_BASE + 0x6d914 + (sprite) * 0x1000)
+#define SPCSCC67(sprite)	(VLV_DISPLAY_BASE + 0x6d918 + (sprite) * 0x1000)
+#define SPCSCC8(sprite)		(VLV_DISPLAY_BASE + 0x6d91c + (sprite) * 0x1000)
+#define  SPCSC_C1(x)		(((x) & 0x7fff) << 16) /* s3.12 */
+#define  SPCSC_C0(x)		(((x) & 0x7fff) << 0) /* s3.12 */
+
+#define SPCSCYGICLAMP(sprite)	(VLV_DISPLAY_BASE + 0x6d920 + (sprite) * 0x1000)
+#define SPCSCCBICLAMP(sprite)	(VLV_DISPLAY_BASE + 0x6d924 + (sprite) * 0x1000)
+#define SPCSCCRICLAMP(sprite)	(VLV_DISPLAY_BASE + 0x6d928 + (sprite) * 0x1000)
+#define  SPCSC_IMAX(x)		(((x) & 0x7ff) << 16) /* s11 */
+#define  SPCSC_IMIN(x)		(((x) & 0x7ff) << 0) /* s11 */
+
+#define SPCSCYGOCLAMP(sprite)	(VLV_DISPLAY_BASE + 0x6d92c + (sprite) * 0x1000)
+#define SPCSCCBOCLAMP(sprite)	(VLV_DISPLAY_BASE + 0x6d930 + (sprite) * 0x1000)
+#define SPCSCCROCLAMP(sprite)	(VLV_DISPLAY_BASE + 0x6d934 + (sprite) * 0x1000)
+#define  SPCSC_OMAX(x)		((x) << 16) /* u10 */
+#define  SPCSC_OMIN(x)		((x) << 0) /* u10 */
+
+/* Skylake plane registers */
+
+#define _PLANE_CTL_1_A				0x70180
+#define _PLANE_CTL_2_A				0x70280
+#define _PLANE_CTL_3_A				0x70380
+#define   PLANE_CTL_ENABLE			(1 << 31)
+#define   PLANE_CTL_PIPE_GAMMA_ENABLE		(1 << 30)
+#define   PLANE_CTL_FORMAT_MASK			(0xf << 24)
+#define   PLANE_CTL_FORMAT_YUV422		(  0 << 24)
+#define   PLANE_CTL_FORMAT_NV12			(  1 << 24)
+#define   PLANE_CTL_FORMAT_XRGB_2101010		(  2 << 24)
+#define   PLANE_CTL_FORMAT_XRGB_8888		(  4 << 24)
+#define   PLANE_CTL_FORMAT_XRGB_16161616F	(  6 << 24)
+#define   PLANE_CTL_FORMAT_AYUV			(  8 << 24)
+#define   PLANE_CTL_FORMAT_INDEXED		( 12 << 24)
+#define   PLANE_CTL_FORMAT_RGB_565		( 14 << 24)
+#define   PLANE_CTL_PIPE_CSC_ENABLE		(1 << 23)
+#define   PLANE_CTL_KEY_ENABLE_MASK		(0x3 << 21)
+#define   PLANE_CTL_KEY_ENABLE_SOURCE		(  1 << 21)
+#define   PLANE_CTL_KEY_ENABLE_DESTINATION	(  2 << 21)
+#define   PLANE_CTL_ORDER_BGRX			(0 << 20)
+#define   PLANE_CTL_ORDER_RGBX			(1 << 20)
+#define   PLANE_CTL_YUV422_ORDER_MASK		(0x3 << 16)
+#define   PLANE_CTL_YUV422_YUYV			(  0 << 16)
+#define   PLANE_CTL_YUV422_UYVY			(  1 << 16)
+#define   PLANE_CTL_YUV422_YVYU			(  2 << 16)
+#define   PLANE_CTL_YUV422_VYUY			(  3 << 16)
+#define   PLANE_CTL_DECOMPRESSION_ENABLE	(1 << 15)
+#define   PLANE_CTL_TRICKLE_FEED_DISABLE	(1 << 14)
+#define   PLANE_CTL_PLANE_GAMMA_DISABLE		(1 << 13)
+#define   PLANE_CTL_TILED_MASK			(0x7 << 10)
+#define   PLANE_CTL_TILED_LINEAR		(  0 << 10)
+#define   PLANE_CTL_TILED_X			(  1 << 10)
+#define   PLANE_CTL_TILED_Y			(  4 << 10)
+#define   PLANE_CTL_TILED_YF			(  5 << 10)
+#define   PLANE_CTL_ALPHA_MASK			(0x3 << 4)
+#define   PLANE_CTL_ALPHA_DISABLE		(  0 << 4)
+#define   PLANE_CTL_ALPHA_SW_PREMULTIPLY	(  2 << 4)
+#define   PLANE_CTL_ALPHA_HW_PREMULTIPLY	(  3 << 4)
+#define   PLANE_CTL_ROTATE_MASK			0x3
+#define   PLANE_CTL_ROTATE_0			0x0
+#define   PLANE_CTL_ROTATE_180			0x2
+#define _PLANE_STRIDE_1_A			0x70188
+#define _PLANE_STRIDE_2_A			0x70288
+#define _PLANE_STRIDE_3_A			0x70388
+#define _PLANE_POS_1_A				0x7018c
+#define _PLANE_POS_2_A				0x7028c
+#define _PLANE_POS_3_A				0x7038c
+#define _PLANE_SIZE_1_A				0x70190
+#define _PLANE_SIZE_2_A				0x70290
+#define _PLANE_SIZE_3_A				0x70390
+#define _PLANE_SURF_1_A				0x7019c
+#define _PLANE_SURF_2_A				0x7029c
+#define _PLANE_SURF_3_A				0x7039c
+#define _PLANE_OFFSET_1_A			0x701a4
+#define _PLANE_OFFSET_2_A			0x702a4
+#define _PLANE_OFFSET_3_A			0x703a4
+#define _PLANE_KEYVAL_1_A			0x70194
+#define _PLANE_KEYVAL_2_A			0x70294
+#define _PLANE_KEYMSK_1_A			0x70198
+#define _PLANE_KEYMSK_2_A			0x70298
+#define _PLANE_KEYMAX_1_A			0x701a0
+#define _PLANE_KEYMAX_2_A			0x702a0
+#define _PLANE_BUF_CFG_1_A			0x7027c
+#define _PLANE_BUF_CFG_2_A			0x7037c
+
+#define _PLANE_CTL_1_B				0x71180
+#define _PLANE_CTL_2_B				0x71280
+#define _PLANE_CTL_3_B				0x71380
+#define _PLANE_CTL_1(pipe)	_PIPE(pipe, _PLANE_CTL_1_A, _PLANE_CTL_1_B)
+#define _PLANE_CTL_2(pipe)	_PIPE(pipe, _PLANE_CTL_2_A, _PLANE_CTL_2_B)
+#define _PLANE_CTL_3(pipe)	_PIPE(pipe, _PLANE_CTL_3_A, _PLANE_CTL_3_B)
+#define PLANE_CTL(pipe, plane)	\
+	_PLANE(plane, _PLANE_CTL_1(pipe), _PLANE_CTL_2(pipe))
+
+#define _PLANE_STRIDE_1_B			0x71188
+#define _PLANE_STRIDE_2_B			0x71288
+#define _PLANE_STRIDE_3_B			0x71388
+#define _PLANE_STRIDE_1(pipe)	\
+	_PIPE(pipe, _PLANE_STRIDE_1_A, _PLANE_STRIDE_1_B)
+#define _PLANE_STRIDE_2(pipe)	\
+	_PIPE(pipe, _PLANE_STRIDE_2_A, _PLANE_STRIDE_2_B)
+#define _PLANE_STRIDE_3(pipe)	\
+	_PIPE(pipe, _PLANE_STRIDE_3_A, _PLANE_STRIDE_3_B)
+#define PLANE_STRIDE(pipe, plane)	\
+	_PLANE(plane, _PLANE_STRIDE_1(pipe), _PLANE_STRIDE_2(pipe))
+
+#define _PLANE_POS_1_B				0x7118c
+#define _PLANE_POS_2_B				0x7128c
+#define _PLANE_POS_3_B				0x7138c
+#define _PLANE_POS_1(pipe)	_PIPE(pipe, _PLANE_POS_1_A, _PLANE_POS_1_B)
+#define _PLANE_POS_2(pipe)	_PIPE(pipe, _PLANE_POS_2_A, _PLANE_POS_2_B)
+#define _PLANE_POS_3(pipe)	_PIPE(pipe, _PLANE_POS_3_A, _PLANE_POS_3_B)
+#define PLANE_POS(pipe, plane)	\
+	_PLANE(plane, _PLANE_POS_1(pipe), _PLANE_POS_2(pipe))
+
+#define _PLANE_SIZE_1_B				0x71190
+#define _PLANE_SIZE_2_B				0x71290
+#define _PLANE_SIZE_3_B				0x71390
+#define _PLANE_SIZE_1(pipe)	_PIPE(pipe, _PLANE_SIZE_1_A, _PLANE_SIZE_1_B)
+#define _PLANE_SIZE_2(pipe)	_PIPE(pipe, _PLANE_SIZE_2_A, _PLANE_SIZE_2_B)
+#define _PLANE_SIZE_3(pipe)	_PIPE(pipe, _PLANE_SIZE_3_A, _PLANE_SIZE_3_B)
+#define PLANE_SIZE(pipe, plane)	\
+	_PLANE(plane, _PLANE_SIZE_1(pipe), _PLANE_SIZE_2(pipe))
+
+#define _PLANE_SURF_1_B				0x7119c
+#define _PLANE_SURF_2_B				0x7129c
+#define _PLANE_SURF_3_B				0x7139c
+#define _PLANE_SURF_1(pipe)	_PIPE(pipe, _PLANE_SURF_1_A, _PLANE_SURF_1_B)
+#define _PLANE_SURF_2(pipe)	_PIPE(pipe, _PLANE_SURF_2_A, _PLANE_SURF_2_B)
+#define _PLANE_SURF_3(pipe)	_PIPE(pipe, _PLANE_SURF_3_A, _PLANE_SURF_3_B)
+#define PLANE_SURF(pipe, plane)	\
+	_PLANE(plane, _PLANE_SURF_1(pipe), _PLANE_SURF_2(pipe))
+
+#define _PLANE_OFFSET_1_B			0x711a4
+#define _PLANE_OFFSET_2_B			0x712a4
+#define _PLANE_OFFSET_1(pipe) _PIPE(pipe, _PLANE_OFFSET_1_A, _PLANE_OFFSET_1_B)
+#define _PLANE_OFFSET_2(pipe) _PIPE(pipe, _PLANE_OFFSET_2_A, _PLANE_OFFSET_2_B)
+#define PLANE_OFFSET(pipe, plane)	\
+	_PLANE(plane, _PLANE_OFFSET_1(pipe), _PLANE_OFFSET_2(pipe))
+
+#define _PLANE_KEYVAL_1_B			0x71194
+#define _PLANE_KEYVAL_2_B			0x71294
+#define _PLANE_KEYVAL_1(pipe) _PIPE(pipe, _PLANE_KEYVAL_1_A, _PLANE_KEYVAL_1_B)
+#define _PLANE_KEYVAL_2(pipe) _PIPE(pipe, _PLANE_KEYVAL_2_A, _PLANE_KEYVAL_2_B)
+#define PLANE_KEYVAL(pipe, plane)	\
+	_PLANE(plane, _PLANE_KEYVAL_1(pipe), _PLANE_KEYVAL_2(pipe))
+
+#define _PLANE_KEYMSK_1_B			0x71198
+#define _PLANE_KEYMSK_2_B			0x71298
+#define _PLANE_KEYMSK_1(pipe) _PIPE(pipe, _PLANE_KEYMSK_1_A, _PLANE_KEYMSK_1_B)
+#define _PLANE_KEYMSK_2(pipe) _PIPE(pipe, _PLANE_KEYMSK_2_A, _PLANE_KEYMSK_2_B)
+#define PLANE_KEYMSK(pipe, plane)	\
+	_PLANE(plane, _PLANE_KEYMSK_1(pipe), _PLANE_KEYMSK_2(pipe))
+
+#define _PLANE_KEYMAX_1_B			0x711a0
+#define _PLANE_KEYMAX_2_B			0x712a0
+#define _PLANE_KEYMAX_1(pipe) _PIPE(pipe, _PLANE_KEYMAX_1_A, _PLANE_KEYMAX_1_B)
+#define _PLANE_KEYMAX_2(pipe) _PIPE(pipe, _PLANE_KEYMAX_2_A, _PLANE_KEYMAX_2_B)
+#define PLANE_KEYMAX(pipe, plane)	\
+	_PLANE(plane, _PLANE_KEYMAX_1(pipe), _PLANE_KEYMAX_2(pipe))
+
+#define _PLANE_BUF_CFG_1_B			0x7127c
+#define _PLANE_BUF_CFG_2_B			0x7137c
+#define _PLANE_BUF_CFG_1(pipe)	\
+	_PIPE(pipe, _PLANE_BUF_CFG_1_A, _PLANE_BUF_CFG_1_B)
+#define _PLANE_BUF_CFG_2(pipe)	\
+	_PIPE(pipe, _PLANE_BUF_CFG_2_A, _PLANE_BUF_CFG_2_B)
+#define PLANE_BUF_CFG(pipe, plane)	\
+	_PLANE(plane, _PLANE_BUF_CFG_1(pipe), _PLANE_BUF_CFG_2(pipe))
+
+/* SKL new cursor registers */
+#define _CUR_BUF_CFG_A				0x7017c
+#define _CUR_BUF_CFG_B				0x7117c
+#define CUR_BUF_CFG(pipe)	_PIPE(pipe, _CUR_BUF_CFG_A, _CUR_BUF_CFG_B)
+
 /* VBIOS regs */
 #define VGACNTRL		0x71400
 # define VGA_DISP_DISABLE			(1 << 31)
@@ -4492,6 +4934,18 @@ enum punit_power_well {
 #define PF_VSCALE(pipe)		_PIPE(pipe, _PFA_VSCALE, _PFB_VSCALE)
 #define PF_HSCALE(pipe)		_PIPE(pipe, _PFA_HSCALE, _PFB_HSCALE)
 
+#define _PSA_CTL		0x68180
+#define _PSB_CTL		0x68980
+#define PS_ENABLE		(1<<31)
+#define _PSA_WIN_SZ		0x68174
+#define _PSB_WIN_SZ		0x68974
+#define _PSA_WIN_POS		0x68170
+#define _PSB_WIN_POS		0x68970
+
+#define PS_CTL(pipe)		_PIPE(pipe, _PSA_CTL, _PSB_CTL)
+#define PS_WIN_SZ(pipe)		_PIPE(pipe, _PSA_WIN_SZ, _PSB_WIN_SZ)
+#define PS_WIN_POS(pipe)	_PIPE(pipe, _PSA_WIN_POS, _PSB_WIN_POS)
+
 /* legacy palette */
 #define _LGC_PALETTE_A           0x4a000
 #define _LGC_PALETTE_B           0x4a800
@@ -4613,16 +5067,32 @@ enum punit_power_well {
 #define  GEN8_PIPE_SCAN_LINE_EVENT	(1 << 2)
 #define  GEN8_PIPE_VSYNC		(1 << 1)
 #define  GEN8_PIPE_VBLANK		(1 << 0)
+#define  GEN9_PIPE_CURSOR_FAULT		(1 << 11)
+#define  GEN9_PIPE_PLANE3_FAULT		(1 << 9)
+#define  GEN9_PIPE_PLANE2_FAULT		(1 << 8)
+#define  GEN9_PIPE_PLANE1_FAULT		(1 << 7)
+#define  GEN9_PIPE_PLANE3_FLIP_DONE	(1 << 5)
+#define  GEN9_PIPE_PLANE2_FLIP_DONE	(1 << 4)
+#define  GEN9_PIPE_PLANE1_FLIP_DONE	(1 << 3)
+#define  GEN9_PIPE_PLANE_FLIP_DONE(p)	(1 << (3 + p))
 #define GEN8_DE_PIPE_IRQ_FAULT_ERRORS \
 	(GEN8_PIPE_CURSOR_FAULT | \
 	 GEN8_PIPE_SPRITE_FAULT | \
 	 GEN8_PIPE_PRIMARY_FAULT)
+#define GEN9_DE_PIPE_IRQ_FAULT_ERRORS \
+	(GEN9_PIPE_CURSOR_FAULT | \
+	 GEN9_PIPE_PLANE3_FAULT | \
+	 GEN9_PIPE_PLANE2_FAULT | \
+	 GEN9_PIPE_PLANE1_FAULT)
 
 #define GEN8_DE_PORT_ISR 0x44440
 #define GEN8_DE_PORT_IMR 0x44444
 #define GEN8_DE_PORT_IIR 0x44448
 #define GEN8_DE_PORT_IER 0x4444c
 #define  GEN8_PORT_DP_A_HOTPLUG		(1 << 3)
+#define  GEN9_AUX_CHANNEL_D		(1 << 27)
+#define  GEN9_AUX_CHANNEL_C		(1 << 26)
+#define  GEN9_AUX_CHANNEL_B		(1 << 25)
 #define  GEN8_AUX_CHANNEL_A		(1 << 0)
 
 #define GEN8_DE_MISC_ISR 0x44460
@@ -4706,6 +5176,8 @@ enum punit_power_well {
 /* GEN8 chicken */
 #define HDC_CHICKEN0				0x7300
 #define  HDC_FORCE_NON_COHERENT			(1<<4)
+#define  HDC_DONOT_FETCH_MEM_WHEN_MASKED	(1<<11)
+#define  HDC_FENCE_DEST_SLM_DISABLE		(1<<14)
 
 /* WaCatErrorRejectionIssue */
 #define GEN7_SQ_CHICKEN_MBCUNIT_CONFIG		0x9030
@@ -5246,8 +5718,7 @@ enum punit_power_well {
 #define PIPEA_PP_STATUS         (VLV_DISPLAY_BASE + 0x61200)
 #define PIPEA_PP_CONTROL        (VLV_DISPLAY_BASE + 0x61204)
 #define PIPEA_PP_ON_DELAYS      (VLV_DISPLAY_BASE + 0x61208)
-#define  PANEL_PORT_SELECT_DPB_VLV	(1 << 30)
-#define  PANEL_PORT_SELECT_DPC_VLV	(2 << 30)
+#define  PANEL_PORT_SELECT_VLV(port)	((port) << 30)
 #define PIPEA_PP_OFF_DELAYS     (VLV_DISPLAY_BASE + 0x6120c)
 #define PIPEA_PP_DIVISOR        (VLV_DISPLAY_BASE + 0x61210)
 
@@ -5407,8 +5878,13 @@ enum punit_power_well {
 #define   VLV_GTLC_ALLOWWAKEERR			(1 << 1)
 #define   VLV_GTLC_PW_MEDIA_STATUS_MASK		(1 << 5)
 #define   VLV_GTLC_PW_RENDER_STATUS_MASK	(1 << 7)
-#define VLV_GTLC_SURVIVABILITY_REG              0x130098
 #define  FORCEWAKE_MT				0xa188 /* multi-threaded */
+#define  FORCEWAKE_MEDIA_GEN9			0xa270
+#define  FORCEWAKE_RENDER_GEN9			0xa278
+#define  FORCEWAKE_BLITTER_GEN9			0xa188
+#define  FORCEWAKE_ACK_MEDIA_GEN9		0x0D88
+#define  FORCEWAKE_ACK_RENDER_GEN9		0x0D84
+#define  FORCEWAKE_ACK_BLITTER_GEN9		0x130044
 #define   FORCEWAKE_KERNEL			0x1
 #define   FORCEWAKE_USER			0x2
 #define  FORCEWAKE_MT_ACK			0x130040
@@ -5545,12 +6021,6 @@ enum punit_power_well {
 						 GEN6_PM_RP_DOWN_THRESHOLD | \
 						 GEN6_PM_RP_DOWN_TIMEOUT)
 
-#define CHV_CZ_CLOCK_FREQ_MODE_200			200
-#define CHV_CZ_CLOCK_FREQ_MODE_267			267
-#define CHV_CZ_CLOCK_FREQ_MODE_320			320
-#define CHV_CZ_CLOCK_FREQ_MODE_333			333
-#define CHV_CZ_CLOCK_FREQ_MODE_400			400
-
 #define GEN7_GT_SCRATCH_BASE			0x4F100
 #define GEN7_GT_SCRATCH_REG_NUM			8
 
@@ -5586,9 +6056,17 @@ enum punit_power_well {
 #define   GEN6_ENCODE_RC6_VID(mv)		(((mv) - 245) / 5)
 #define   GEN6_DECODE_RC6_VID(vids)		(((vids) * 5) + 245)
 #define   DISPLAY_IPS_CONTROL			0x19
+#define	  HSW_PCODE_DYNAMIC_DUTY_CYCLE_CONTROL	0x1A
 #define GEN6_PCODE_DATA				0x138128
 #define   GEN6_PCODE_FREQ_IA_RATIO_SHIFT	8
 #define   GEN6_PCODE_FREQ_RING_RATIO_SHIFT	16
+#define GEN6_PCODE_DATA1			0x13812C
+
+#define   GEN9_PCODE_READ_MEM_LATENCY		0x6
+#define   GEN9_MEM_LATENCY_LEVEL_MASK		0xFF
+#define   GEN9_MEM_LATENCY_LEVEL_1_5_SHIFT	8
+#define   GEN9_MEM_LATENCY_LEVEL_2_6_SHIFT	16
+#define   GEN9_MEM_LATENCY_LEVEL_3_7_SHIFT	24
 
 #define GEN6_GT_CORE_STATUS		0x138060
 #define   GEN6_CORE_CPD_STATE_MASK	(7<<4)
@@ -5626,6 +6104,9 @@ enum punit_power_well {
 #define   GEN7_SINGLE_SUBSCAN_DISPATCH_ENABLE	(1<<10)
 #define   GEN7_PSD_SINGLE_PORT_DISPATCH_ENABLE	(1<<3)
 
+#define GEN9_HALF_SLICE_CHICKEN5	0xe188
+#define   GEN9_DG_MIRROR_FIX_ENABLE	(1<<5)
+
 #define GEN8_ROW_CHICKEN		0xe4f0
 #define   PARTIAL_INSTRUCTION_SHOOTDOWN_DISABLE	(1<<8)
 #define   STALL_DOP_GATING_DISABLE		(1<<5)
@@ -5641,57 +6122,58 @@ enum punit_power_well {
 #define   GEN8_CENTROID_PIXEL_OPT_DIS	(1<<8)
 #define   GEN8_SAMPLER_POWER_BYPASS_DIS	(1<<1)
 
+/* Audio */
 #define G4X_AUD_VID_DID			(dev_priv->info.display_mmio_offset + 0x62020)
-#define INTEL_AUDIO_DEVCL		0x808629FB
-#define INTEL_AUDIO_DEVBLC		0x80862801
-#define INTEL_AUDIO_DEVCTG		0x80862802
+#define   INTEL_AUDIO_DEVCL		0x808629FB
+#define   INTEL_AUDIO_DEVBLC		0x80862801
+#define   INTEL_AUDIO_DEVCTG		0x80862802
 
 #define G4X_AUD_CNTL_ST			0x620B4
-#define G4X_ELDV_DEVCL_DEVBLC		(1 << 13)
-#define G4X_ELDV_DEVCTG			(1 << 14)
-#define G4X_ELD_ADDR			(0xf << 5)
-#define G4X_ELD_ACK			(1 << 4)
+#define   G4X_ELDV_DEVCL_DEVBLC		(1 << 13)
+#define   G4X_ELDV_DEVCTG		(1 << 14)
+#define   G4X_ELD_ADDR_MASK		(0xf << 5)
+#define   G4X_ELD_ACK			(1 << 4)
 #define G4X_HDMIW_HDMIEDID		0x6210C
 
-#define IBX_HDMIW_HDMIEDID_A		0xE2050
-#define IBX_HDMIW_HDMIEDID_B		0xE2150
+#define _IBX_HDMIW_HDMIEDID_A		0xE2050
+#define _IBX_HDMIW_HDMIEDID_B		0xE2150
 #define IBX_HDMIW_HDMIEDID(pipe) _PIPE(pipe, \
-					IBX_HDMIW_HDMIEDID_A, \
-					IBX_HDMIW_HDMIEDID_B)
-#define IBX_AUD_CNTL_ST_A		0xE20B4
-#define IBX_AUD_CNTL_ST_B		0xE21B4
+					_IBX_HDMIW_HDMIEDID_A, \
+					_IBX_HDMIW_HDMIEDID_B)
+#define _IBX_AUD_CNTL_ST_A		0xE20B4
+#define _IBX_AUD_CNTL_ST_B		0xE21B4
 #define IBX_AUD_CNTL_ST(pipe) _PIPE(pipe, \
-					IBX_AUD_CNTL_ST_A, \
-					IBX_AUD_CNTL_ST_B)
-#define IBX_ELD_BUFFER_SIZE		(0x1f << 10)
-#define IBX_ELD_ADDRESS			(0x1f << 5)
-#define IBX_ELD_ACK			(1 << 4)
+					_IBX_AUD_CNTL_ST_A, \
+					_IBX_AUD_CNTL_ST_B)
+#define   IBX_ELD_BUFFER_SIZE_MASK	(0x1f << 10)
+#define   IBX_ELD_ADDRESS_MASK		(0x1f << 5)
+#define   IBX_ELD_ACK			(1 << 4)
 #define IBX_AUD_CNTL_ST2		0xE20C0
-#define IBX_ELD_VALIDB			(1 << 0)
-#define IBX_CP_READYB			(1 << 1)
+#define   IBX_CP_READY(port)		((1 << 1) << (((port) - 1) * 4))
+#define   IBX_ELD_VALID(port)		((1 << 0) << (((port) - 1) * 4))
 
-#define CPT_HDMIW_HDMIEDID_A		0xE5050
-#define CPT_HDMIW_HDMIEDID_B		0xE5150
+#define _CPT_HDMIW_HDMIEDID_A		0xE5050
+#define _CPT_HDMIW_HDMIEDID_B		0xE5150
 #define CPT_HDMIW_HDMIEDID(pipe) _PIPE(pipe, \
-					CPT_HDMIW_HDMIEDID_A, \
-					CPT_HDMIW_HDMIEDID_B)
-#define CPT_AUD_CNTL_ST_A		0xE50B4
-#define CPT_AUD_CNTL_ST_B		0xE51B4
+					_CPT_HDMIW_HDMIEDID_A, \
+					_CPT_HDMIW_HDMIEDID_B)
+#define _CPT_AUD_CNTL_ST_A		0xE50B4
+#define _CPT_AUD_CNTL_ST_B		0xE51B4
 #define CPT_AUD_CNTL_ST(pipe) _PIPE(pipe, \
-					CPT_AUD_CNTL_ST_A, \
-					CPT_AUD_CNTL_ST_B)
+					_CPT_AUD_CNTL_ST_A, \
+					_CPT_AUD_CNTL_ST_B)
 #define CPT_AUD_CNTRL_ST2		0xE50C0
 
-#define VLV_HDMIW_HDMIEDID_A		(VLV_DISPLAY_BASE + 0x62050)
-#define VLV_HDMIW_HDMIEDID_B		(VLV_DISPLAY_BASE + 0x62150)
+#define _VLV_HDMIW_HDMIEDID_A		(VLV_DISPLAY_BASE + 0x62050)
+#define _VLV_HDMIW_HDMIEDID_B		(VLV_DISPLAY_BASE + 0x62150)
 #define VLV_HDMIW_HDMIEDID(pipe) _PIPE(pipe, \
-					VLV_HDMIW_HDMIEDID_A, \
-					VLV_HDMIW_HDMIEDID_B)
-#define VLV_AUD_CNTL_ST_A		(VLV_DISPLAY_BASE + 0x620B4)
-#define VLV_AUD_CNTL_ST_B		(VLV_DISPLAY_BASE + 0x621B4)
+					_VLV_HDMIW_HDMIEDID_A, \
+					_VLV_HDMIW_HDMIEDID_B)
+#define _VLV_AUD_CNTL_ST_A		(VLV_DISPLAY_BASE + 0x620B4)
+#define _VLV_AUD_CNTL_ST_B		(VLV_DISPLAY_BASE + 0x621B4)
 #define VLV_AUD_CNTL_ST(pipe) _PIPE(pipe, \
-					VLV_AUD_CNTL_ST_A, \
-					VLV_AUD_CNTL_ST_B)
+					_VLV_AUD_CNTL_ST_A, \
+					_VLV_AUD_CNTL_ST_B)
 #define VLV_AUD_CNTL_ST2		(VLV_DISPLAY_BASE + 0x620C0)
 
 /* These are the 4 32-bit write offset registers for each stream
@@ -5700,28 +6182,28 @@ enum punit_power_well {
  */
 #define GEN7_SO_WRITE_OFFSET(n)		(0x5280 + (n) * 4)
 
-#define IBX_AUD_CONFIG_A			0xe2000
-#define IBX_AUD_CONFIG_B			0xe2100
+#define _IBX_AUD_CONFIG_A		0xe2000
+#define _IBX_AUD_CONFIG_B		0xe2100
 #define IBX_AUD_CFG(pipe) _PIPE(pipe, \
-					IBX_AUD_CONFIG_A, \
-					IBX_AUD_CONFIG_B)
-#define CPT_AUD_CONFIG_A			0xe5000
-#define CPT_AUD_CONFIG_B			0xe5100
+					_IBX_AUD_CONFIG_A, \
+					_IBX_AUD_CONFIG_B)
+#define _CPT_AUD_CONFIG_A		0xe5000
+#define _CPT_AUD_CONFIG_B		0xe5100
 #define CPT_AUD_CFG(pipe) _PIPE(pipe, \
-					CPT_AUD_CONFIG_A, \
-					CPT_AUD_CONFIG_B)
-#define VLV_AUD_CONFIG_A		(VLV_DISPLAY_BASE + 0x62000)
-#define VLV_AUD_CONFIG_B		(VLV_DISPLAY_BASE + 0x62100)
+					_CPT_AUD_CONFIG_A, \
+					_CPT_AUD_CONFIG_B)
+#define _VLV_AUD_CONFIG_A		(VLV_DISPLAY_BASE + 0x62000)
+#define _VLV_AUD_CONFIG_B		(VLV_DISPLAY_BASE + 0x62100)
 #define VLV_AUD_CFG(pipe) _PIPE(pipe, \
-					VLV_AUD_CONFIG_A, \
-					VLV_AUD_CONFIG_B)
+					_VLV_AUD_CONFIG_A, \
+					_VLV_AUD_CONFIG_B)
 
 #define   AUD_CONFIG_N_VALUE_INDEX		(1 << 29)
 #define   AUD_CONFIG_N_PROG_ENABLE		(1 << 28)
 #define   AUD_CONFIG_UPPER_N_SHIFT		20
-#define   AUD_CONFIG_UPPER_N_VALUE		(0xff << 20)
+#define   AUD_CONFIG_UPPER_N_MASK		(0xff << 20)
 #define   AUD_CONFIG_LOWER_N_SHIFT		4
-#define   AUD_CONFIG_LOWER_N_VALUE		(0xfff << 4)
+#define   AUD_CONFIG_LOWER_N_MASK		(0xfff << 4)
 #define   AUD_CONFIG_PIXEL_CLOCK_HDMI_SHIFT	16
 #define   AUD_CONFIG_PIXEL_CLOCK_HDMI_MASK	(0xf << 16)
 #define   AUD_CONFIG_PIXEL_CLOCK_HDMI_25175	(0 << 16)
@@ -5737,52 +6219,44 @@ enum punit_power_well {
 #define   AUD_CONFIG_DISABLE_NCTS		(1 << 3)
 
 /* HSW Audio */
-#define   HSW_AUD_CONFIG_A		0x65000 /* Audio Configuration Transcoder A */
-#define   HSW_AUD_CONFIG_B		0x65100 /* Audio Configuration Transcoder B */
-#define   HSW_AUD_CFG(pipe) _PIPE(pipe, \
-					HSW_AUD_CONFIG_A, \
-					HSW_AUD_CONFIG_B)
-
-#define   HSW_AUD_MISC_CTRL_A		0x65010 /* Audio Misc Control Convert 1 */
-#define   HSW_AUD_MISC_CTRL_B		0x65110 /* Audio Misc Control Convert 2 */
-#define   HSW_AUD_MISC_CTRL(pipe) _PIPE(pipe, \
-					HSW_AUD_MISC_CTRL_A, \
-					HSW_AUD_MISC_CTRL_B)
-
-#define   HSW_AUD_DIP_ELD_CTRL_ST_A	0x650b4 /* Audio DIP and ELD Control State Transcoder A */
-#define   HSW_AUD_DIP_ELD_CTRL_ST_B	0x651b4 /* Audio DIP and ELD Control State Transcoder B */
-#define   HSW_AUD_DIP_ELD_CTRL(pipe) _PIPE(pipe, \
-					HSW_AUD_DIP_ELD_CTRL_ST_A, \
-					HSW_AUD_DIP_ELD_CTRL_ST_B)
+#define _HSW_AUD_CONFIG_A		0x65000
+#define _HSW_AUD_CONFIG_B		0x65100
+#define HSW_AUD_CFG(pipe) _PIPE(pipe, \
+					_HSW_AUD_CONFIG_A, \
+					_HSW_AUD_CONFIG_B)
+
+#define _HSW_AUD_MISC_CTRL_A		0x65010
+#define _HSW_AUD_MISC_CTRL_B		0x65110
+#define HSW_AUD_MISC_CTRL(pipe) _PIPE(pipe, \
+					_HSW_AUD_MISC_CTRL_A, \
+					_HSW_AUD_MISC_CTRL_B)
+
+#define _HSW_AUD_DIP_ELD_CTRL_ST_A	0x650b4
+#define _HSW_AUD_DIP_ELD_CTRL_ST_B	0x651b4
+#define HSW_AUD_DIP_ELD_CTRL(pipe) _PIPE(pipe, \
+					_HSW_AUD_DIP_ELD_CTRL_ST_A, \
+					_HSW_AUD_DIP_ELD_CTRL_ST_B)
 
 /* Audio Digital Converter */
-#define   HSW_AUD_DIG_CNVT_1		0x65080 /* Audio Converter 1 */
-#define   HSW_AUD_DIG_CNVT_2		0x65180 /* Audio Converter 1 */
-#define   AUD_DIG_CNVT(pipe) _PIPE(pipe, \
-					HSW_AUD_DIG_CNVT_1, \
-					HSW_AUD_DIG_CNVT_2)
-#define   DIP_PORT_SEL_MASK		0x3
-
-#define   HSW_AUD_EDID_DATA_A		0x65050
-#define   HSW_AUD_EDID_DATA_B		0x65150
-#define   HSW_AUD_EDID_DATA(pipe) _PIPE(pipe, \
-					HSW_AUD_EDID_DATA_A, \
-					HSW_AUD_EDID_DATA_B)
-
-#define   HSW_AUD_PIPE_CONV_CFG		0x6507c /* Audio pipe and converter configs */
-#define   HSW_AUD_PIN_ELD_CP_VLD	0x650c0 /* Audio ELD and CP Ready Status */
-#define   AUDIO_INACTIVE_C		(1<<11)
-#define   AUDIO_INACTIVE_B		(1<<7)
-#define   AUDIO_INACTIVE_A		(1<<3)
-#define   AUDIO_OUTPUT_ENABLE_A		(1<<2)
-#define   AUDIO_OUTPUT_ENABLE_B		(1<<6)
-#define   AUDIO_OUTPUT_ENABLE_C		(1<<10)
-#define   AUDIO_ELD_VALID_A		(1<<0)
-#define   AUDIO_ELD_VALID_B		(1<<4)
-#define   AUDIO_ELD_VALID_C		(1<<8)
-#define   AUDIO_CP_READY_A		(1<<1)
-#define   AUDIO_CP_READY_B		(1<<5)
-#define   AUDIO_CP_READY_C		(1<<9)
+#define _HSW_AUD_DIG_CNVT_1		0x65080
+#define _HSW_AUD_DIG_CNVT_2		0x65180
+#define AUD_DIG_CNVT(pipe) _PIPE(pipe, \
+					_HSW_AUD_DIG_CNVT_1, \
+					_HSW_AUD_DIG_CNVT_2)
+#define DIP_PORT_SEL_MASK		0x3
+
+#define _HSW_AUD_EDID_DATA_A		0x65050
+#define _HSW_AUD_EDID_DATA_B		0x65150
+#define HSW_AUD_EDID_DATA(pipe) _PIPE(pipe, \
+					_HSW_AUD_EDID_DATA_A, \
+					_HSW_AUD_EDID_DATA_B)
+
+#define HSW_AUD_PIPE_CONV_CFG		0x6507c
+#define HSW_AUD_PIN_ELD_CP_VLD		0x650c0
+#define   AUDIO_INACTIVE(trans)		((1 << 3) << ((trans) * 4))
+#define   AUDIO_OUTPUT_ENABLE(trans)	((1 << 2) << ((trans) * 4))
+#define   AUDIO_CP_READY(trans)		((1 << 1) << ((trans) * 4))
+#define   AUDIO_ELD_VALID(trans)	((1 << 0) << ((trans) * 4))
 
 /* HSW Power Wells */
 #define HSW_PWR_WELL_BIOS			0x45400 /* CTL1 */
@@ -5866,15 +6340,7 @@ enum punit_power_well {
 #define DDI_BUF_CTL_B				0x64100
 #define DDI_BUF_CTL(port) _PORT(port, DDI_BUF_CTL_A, DDI_BUF_CTL_B)
 #define  DDI_BUF_CTL_ENABLE			(1<<31)
-#define  DDI_BUF_EMP_400MV_0DB_HSW		(0<<24)   /* Sel0 */
-#define  DDI_BUF_EMP_400MV_3_5DB_HSW		(1<<24)   /* Sel1 */
-#define  DDI_BUF_EMP_400MV_6DB_HSW		(2<<24)   /* Sel2 */
-#define  DDI_BUF_EMP_400MV_9_5DB_HSW		(3<<24)   /* Sel3 */
-#define  DDI_BUF_EMP_600MV_0DB_HSW		(4<<24)   /* Sel4 */
-#define  DDI_BUF_EMP_600MV_3_5DB_HSW		(5<<24)   /* Sel5 */
-#define  DDI_BUF_EMP_600MV_6DB_HSW		(6<<24)   /* Sel6 */
-#define  DDI_BUF_EMP_800MV_0DB_HSW		(7<<24)   /* Sel7 */
-#define  DDI_BUF_EMP_800MV_3_5DB_HSW		(8<<24)   /* Sel8 */
+#define  DDI_BUF_TRANS_SELECT(n)	((n) << 24)
 #define  DDI_BUF_EMP_MASK			(0xf<<24)
 #define  DDI_BUF_PORT_REVERSAL			(1<<16)
 #define  DDI_BUF_IS_IDLE			(1<<7)
@@ -6008,6 +6474,83 @@ enum punit_power_well {
 #define  LCPLL_CD_SOURCE_FCLK		(1<<21)
 #define  LCPLL_CD_SOURCE_FCLK_DONE	(1<<19)
 
+/*
+ * SKL Clocks
+ */
+
+/* CDCLK_CTL */
+#define CDCLK_CTL			0x46000
+#define  CDCLK_FREQ_SEL_MASK		(3<<26)
+#define  CDCLK_FREQ_450_432		(0<<26)
+#define  CDCLK_FREQ_540			(1<<26)
+#define  CDCLK_FREQ_337_308		(2<<26)
+#define  CDCLK_FREQ_675_617		(3<<26)
+#define  CDCLK_FREQ_DECIMAL_MASK	(0x7ff)
+
+/* LCPLL_CTL */
+#define LCPLL1_CTL		0x46010
+#define LCPLL2_CTL		0x46014
+#define  LCPLL_PLL_ENABLE	(1<<31)
+
+/* DPLL control1 */
+#define DPLL_CTRL1		0x6C058
+#define  DPLL_CTRL1_HDMI_MODE(id)		(1<<((id)*6+5))
+#define  DPLL_CTRL1_SSC(id)			(1<<((id)*6+4))
+#define  DPLL_CRTL1_LINK_RATE_MASK(id)		(7<<((id)*6+1))
+#define  DPLL_CRTL1_LINK_RATE_SHIFT(id)		((id)*6+1)
+#define  DPLL_CRTL1_LINK_RATE(linkrate, id)	((linkrate)<<((id)*6+1))
+#define  DPLL_CTRL1_OVERRIDE(id)		(1<<((id)*6))
+#define  DPLL_CRTL1_LINK_RATE_2700		0
+#define  DPLL_CRTL1_LINK_RATE_1350		1
+#define  DPLL_CRTL1_LINK_RATE_810		2
+#define  DPLL_CRTL1_LINK_RATE_1620		3
+#define  DPLL_CRTL1_LINK_RATE_1080		4
+#define  DPLL_CRTL1_LINK_RATE_2160		5
+
+/* DPLL control2 */
+#define DPLL_CTRL2				0x6C05C
+#define  DPLL_CTRL2_DDI_CLK_OFF(port)		(1<<(port+15))
+#define  DPLL_CTRL2_DDI_CLK_SEL_MASK(port)	(3<<((port)*3+1))
+#define  DPLL_CTRL2_DDI_CLK_SEL_SHIFT(port)    ((port)*3+1)
+#define  DPLL_CTRL2_DDI_CLK_SEL(clk, port)	(clk<<((port)*3+1))
+#define  DPLL_CTRL2_DDI_SEL_OVERRIDE(port)     (1<<((port)*3))
+
+/* DPLL Status */
+#define DPLL_STATUS	0x6C060
+#define  DPLL_LOCK(id) (1<<((id)*8))
+
+/* DPLL cfg */
+#define DPLL1_CFGCR1	0x6C040
+#define DPLL2_CFGCR1	0x6C048
+#define DPLL3_CFGCR1	0x6C050
+#define  DPLL_CFGCR1_FREQ_ENABLE	(1<<31)
+#define  DPLL_CFGCR1_DCO_FRACTION_MASK	(0x7fff<<9)
+#define  DPLL_CFGCR1_DCO_FRACTION(x)	(x<<9)
+#define  DPLL_CFGCR1_DCO_INTEGER_MASK	(0x1ff)
+
+#define DPLL1_CFGCR2	0x6C044
+#define DPLL2_CFGCR2	0x6C04C
+#define DPLL3_CFGCR2	0x6C054
+#define  DPLL_CFGCR2_QDIV_RATIO_MASK	(0xff<<8)
+#define  DPLL_CFGCR2_QDIV_RATIO(x)	(x<<8)
+#define  DPLL_CFGCR2_QDIV_MODE(x)	(x<<7)
+#define  DPLL_CFGCR2_KDIV_MASK		(3<<5)
+#define  DPLL_CFGCR2_KDIV(x)		(x<<5)
+#define  DPLL_CFGCR2_KDIV_5 (0<<5)
+#define  DPLL_CFGCR2_KDIV_2 (1<<5)
+#define  DPLL_CFGCR2_KDIV_3 (2<<5)
+#define  DPLL_CFGCR2_KDIV_1 (3<<5)
+#define  DPLL_CFGCR2_PDIV_MASK		(7<<2)
+#define  DPLL_CFGCR2_PDIV(x)		(x<<2)
+#define  DPLL_CFGCR2_PDIV_1 (0<<2)
+#define  DPLL_CFGCR2_PDIV_2 (1<<2)
+#define  DPLL_CFGCR2_PDIV_3 (2<<2)
+#define  DPLL_CFGCR2_PDIV_7 (4<<2)
+#define  DPLL_CFGCR2_CENTRAL_FREQ_MASK	(3)
+
+#define GET_CFG_CR1_REG(id) (DPLL1_CFGCR1 + (id - SKL_DPLL1) * 8)
+#define GET_CFG_CR2_REG(id) (DPLL1_CFGCR2 + (id - SKL_DPLL1) * 8)
+
 /* Please see hsw_read_dcomp() and hsw_write_dcomp() before using this register,
  * since on HSW we can't write to it using I915_WRITE. */
 #define D_COMP_HSW			(MCHBAR_MIRROR_BASE_SNB + 0x5F0C)
diff --git a/drivers/gpu/drm/i915/i915_suspend.c b/drivers/gpu/drm/i915/i915_suspend.c
index 043123c77a1f..26368822a33f 100644
--- a/drivers/gpu/drm/i915/i915_suspend.c
+++ b/drivers/gpu/drm/i915/i915_suspend.c
@@ -203,34 +203,19 @@ static void i915_save_display(struct drm_device *dev)
 		i915_save_display_reg(dev);
 
 	/* LVDS state */
-	if (HAS_PCH_SPLIT(dev)) {
-		dev_priv->regfile.savePP_CONTROL = I915_READ(PCH_PP_CONTROL);
-		if (HAS_PCH_IBX(dev) || HAS_PCH_CPT(dev))
-			dev_priv->regfile.saveLVDS = I915_READ(PCH_LVDS);
-	} else if (IS_VALLEYVIEW(dev)) {
-		dev_priv->regfile.savePP_CONTROL = I915_READ(PP_CONTROL);
-		dev_priv->regfile.savePFIT_PGM_RATIOS = I915_READ(PFIT_PGM_RATIOS);
-
-		dev_priv->regfile.saveBLC_HIST_CTL =
-			I915_READ(VLV_BLC_HIST_CTL(PIPE_A));
-		dev_priv->regfile.saveBLC_HIST_CTL_B =
-			I915_READ(VLV_BLC_HIST_CTL(PIPE_B));
-	} else {
-		dev_priv->regfile.savePP_CONTROL = I915_READ(PP_CONTROL);
-		dev_priv->regfile.savePFIT_PGM_RATIOS = I915_READ(PFIT_PGM_RATIOS);
-		dev_priv->regfile.saveBLC_HIST_CTL = I915_READ(BLC_HIST_CTL);
-		if (IS_MOBILE(dev) && !IS_I830(dev))
-			dev_priv->regfile.saveLVDS = I915_READ(LVDS);
-	}
-
-	if (!IS_I830(dev) && !IS_845G(dev) && !HAS_PCH_SPLIT(dev))
-		dev_priv->regfile.savePFIT_CONTROL = I915_READ(PFIT_CONTROL);
+	if (HAS_PCH_IBX(dev) || HAS_PCH_CPT(dev))
+		dev_priv->regfile.saveLVDS = I915_READ(PCH_LVDS);
+	else if (INTEL_INFO(dev)->gen <= 4 && IS_MOBILE(dev) && !IS_I830(dev))
+		dev_priv->regfile.saveLVDS = I915_READ(LVDS);
 
+	/* Panel power sequencer */
 	if (HAS_PCH_SPLIT(dev)) {
+		dev_priv->regfile.savePP_CONTROL = I915_READ(PCH_PP_CONTROL);
 		dev_priv->regfile.savePP_ON_DELAYS = I915_READ(PCH_PP_ON_DELAYS);
 		dev_priv->regfile.savePP_OFF_DELAYS = I915_READ(PCH_PP_OFF_DELAYS);
 		dev_priv->regfile.savePP_DIVISOR = I915_READ(PCH_PP_DIVISOR);
-	} else {
+	} else if (!IS_VALLEYVIEW(dev)) {
+		dev_priv->regfile.savePP_CONTROL = I915_READ(PP_CONTROL);
 		dev_priv->regfile.savePP_ON_DELAYS = I915_READ(PP_ON_DELAYS);
 		dev_priv->regfile.savePP_OFF_DELAYS = I915_READ(PP_OFF_DELAYS);
 		dev_priv->regfile.savePP_DIVISOR = I915_READ(PP_DIVISOR);
@@ -259,29 +244,19 @@ static void i915_restore_display(struct drm_device *dev)
 	if (drm_core_check_feature(dev, DRIVER_MODESET))
 		mask = ~LVDS_PORT_EN;
 
+	/* LVDS state */
 	if (HAS_PCH_IBX(dev) || HAS_PCH_CPT(dev))
 		I915_WRITE(PCH_LVDS, dev_priv->regfile.saveLVDS & mask);
 	else if (INTEL_INFO(dev)->gen <= 4 && IS_MOBILE(dev) && !IS_I830(dev))
 		I915_WRITE(LVDS, dev_priv->regfile.saveLVDS & mask);
 
-	if (!IS_I830(dev) && !IS_845G(dev) && !HAS_PCH_SPLIT(dev))
-		I915_WRITE(PFIT_CONTROL, dev_priv->regfile.savePFIT_CONTROL);
-
+	/* Panel power sequencer */
 	if (HAS_PCH_SPLIT(dev)) {
 		I915_WRITE(PCH_PP_ON_DELAYS, dev_priv->regfile.savePP_ON_DELAYS);
 		I915_WRITE(PCH_PP_OFF_DELAYS, dev_priv->regfile.savePP_OFF_DELAYS);
 		I915_WRITE(PCH_PP_DIVISOR, dev_priv->regfile.savePP_DIVISOR);
 		I915_WRITE(PCH_PP_CONTROL, dev_priv->regfile.savePP_CONTROL);
-		I915_WRITE(RSTDBYCTL,
-			   dev_priv->regfile.saveMCHBAR_RENDER_STANDBY);
-	} else if (IS_VALLEYVIEW(dev)) {
-		I915_WRITE(VLV_BLC_HIST_CTL(PIPE_A),
-			   dev_priv->regfile.saveBLC_HIST_CTL);
-		I915_WRITE(VLV_BLC_HIST_CTL(PIPE_B),
-			   dev_priv->regfile.saveBLC_HIST_CTL);
-	} else {
-		I915_WRITE(PFIT_PGM_RATIOS, dev_priv->regfile.savePFIT_PGM_RATIOS);
-		I915_WRITE(BLC_HIST_CTL, dev_priv->regfile.saveBLC_HIST_CTL);
+	} else if (!IS_VALLEYVIEW(dev)) {
 		I915_WRITE(PP_ON_DELAYS, dev_priv->regfile.savePP_ON_DELAYS);
 		I915_WRITE(PP_OFF_DELAYS, dev_priv->regfile.savePP_OFF_DELAYS);
 		I915_WRITE(PP_DIVISOR, dev_priv->regfile.savePP_DIVISOR);
@@ -328,6 +303,10 @@ int i915_save_state(struct drm_device *dev)
 		}
 	}
 
+	if (IS_GEN4(dev))
+		pci_read_config_word(dev->pdev, GCDGMBUS,
+				     &dev_priv->regfile.saveGCDGMBUS);
+
 	/* Cache mode state */
 	if (INTEL_INFO(dev)->gen < 7)
 		dev_priv->regfile.saveCACHE_MODE_0 = I915_READ(CACHE_MODE_0);
@@ -356,6 +335,10 @@ int i915_restore_state(struct drm_device *dev)
 	mutex_lock(&dev->struct_mutex);
 
 	i915_gem_restore_fences(dev);
+
+	if (IS_GEN4(dev))
+		pci_write_config_word(dev->pdev, GCDGMBUS,
+				      dev_priv->regfile.saveGCDGMBUS);
 	i915_restore_display(dev);
 
 	if (!drm_core_check_feature(dev, DRIVER_MODESET)) {
@@ -368,6 +351,8 @@ int i915_restore_state(struct drm_device *dev)
 			I915_WRITE(_FDI_RXA_IMR, dev_priv->regfile.saveFDI_RXA_IMR);
 			I915_WRITE(_FDI_RXB_IMR, dev_priv->regfile.saveFDI_RXB_IMR);
 			I915_WRITE(PCH_PORT_HOTPLUG, dev_priv->regfile.savePCH_PORT_HOTPLUG);
+			I915_WRITE(RSTDBYCTL,
+				   dev_priv->regfile.saveMCHBAR_RENDER_STANDBY);
 		} else {
 			I915_WRITE(IER, dev_priv->regfile.saveIER);
 			I915_WRITE(IMR, dev_priv->regfile.saveIMR);
diff --git a/drivers/gpu/drm/i915/i915_sysfs.c b/drivers/gpu/drm/i915/i915_sysfs.c
index ae7fd8fc27f0..4a5af695307e 100644
--- a/drivers/gpu/drm/i915/i915_sysfs.c
+++ b/drivers/gpu/drm/i915/i915_sysfs.c
@@ -139,8 +139,6 @@ static DEVICE_ATTR(rc6pp_residency_ms, S_IRUGO, show_rc6pp_ms, NULL);
 static struct attribute *rc6_attrs[] = {
 	&dev_attr_rc6_enable.attr,
 	&dev_attr_rc6_residency_ms.attr,
-	&dev_attr_rc6p_residency_ms.attr,
-	&dev_attr_rc6pp_residency_ms.attr,
 	NULL
 };
 
@@ -148,6 +146,17 @@ static struct attribute_group rc6_attr_group = {
 	.name = power_group_name,
 	.attrs =  rc6_attrs
 };
+
+static struct attribute *rc6p_attrs[] = {
+	&dev_attr_rc6p_residency_ms.attr,
+	&dev_attr_rc6pp_residency_ms.attr,
+	NULL
+};
+
+static struct attribute_group rc6p_attr_group = {
+	.name = power_group_name,
+	.attrs =  rc6p_attrs
+};
 #endif
 
 static int l3_access_valid(struct drm_device *dev, loff_t offset)
@@ -540,7 +549,7 @@ static ssize_t error_state_read(struct file *filp, struct kobject *kobj,
 
 	memset(&error_priv, 0, sizeof(error_priv));
 
-	ret = i915_error_state_buf_init(&error_str, count, off);
+	ret = i915_error_state_buf_init(&error_str, to_i915(dev), count, off);
 	if (ret)
 		return ret;
 
@@ -595,12 +604,18 @@ void i915_setup_sysfs(struct drm_device *dev)
 	int ret;
 
 #ifdef CONFIG_PM
-	if (INTEL_INFO(dev)->gen >= 6) {
+	if (HAS_RC6(dev)) {
 		ret = sysfs_merge_group(&dev->primary->kdev->kobj,
 					&rc6_attr_group);
 		if (ret)
 			DRM_ERROR("RC6 residency sysfs setup failed\n");
 	}
+	if (HAS_RC6p(dev)) {
+		ret = sysfs_merge_group(&dev->primary->kdev->kobj,
+					&rc6p_attr_group);
+		if (ret)
+			DRM_ERROR("RC6p residency sysfs setup failed\n");
+	}
 #endif
 	if (HAS_L3_DPF(dev)) {
 		ret = device_create_bin_file(dev->primary->kdev, &dpf_attrs);
@@ -640,5 +655,6 @@ void i915_teardown_sysfs(struct drm_device *dev)
 	device_remove_bin_file(dev->primary->kdev,  &dpf_attrs);
 #ifdef CONFIG_PM
 	sysfs_unmerge_group(&dev->primary->kdev->kobj, &rc6_attr_group);
+	sysfs_unmerge_group(&dev->primary->kdev->kobj, &rc6p_attr_group);
 #endif
 }
diff --git a/drivers/gpu/drm/i915/i915_trace.h b/drivers/gpu/drm/i915/i915_trace.h
index f5aa0067755a..751d4ad14d62 100644
--- a/drivers/gpu/drm/i915/i915_trace.h
+++ b/drivers/gpu/drm/i915/i915_trace.h
@@ -587,6 +587,110 @@ TRACE_EVENT(intel_gpu_freq_change,
 	    TP_printk("new_freq=%u", __entry->freq)
 );
 
+/**
+ * DOC: i915_ppgtt_create and i915_ppgtt_release tracepoints
+ *
+ * With full ppgtt enabled each process using drm will allocate at least one
+ * translation table. With these traces it is possible to keep track of the
+ * allocation and of the lifetime of the tables; this can be used during
+ * testing/debug to verify that we are not leaking ppgtts.
+ * These traces identify the ppgtt through the vm pointer, which is also printed
+ * by the i915_vma_bind and i915_vma_unbind tracepoints.
+ */
+DECLARE_EVENT_CLASS(i915_ppgtt,
+	TP_PROTO(struct i915_address_space *vm),
+	TP_ARGS(vm),
+
+	TP_STRUCT__entry(
+			__field(struct i915_address_space *, vm)
+			__field(u32, dev)
+	),
+
+	TP_fast_assign(
+			__entry->vm = vm;
+			__entry->dev = vm->dev->primary->index;
+	),
+
+	TP_printk("dev=%u, vm=%p", __entry->dev, __entry->vm)
+)
+
+DEFINE_EVENT(i915_ppgtt, i915_ppgtt_create,
+	TP_PROTO(struct i915_address_space *vm),
+	TP_ARGS(vm)
+);
+
+DEFINE_EVENT(i915_ppgtt, i915_ppgtt_release,
+	TP_PROTO(struct i915_address_space *vm),
+	TP_ARGS(vm)
+);
+
+/**
+ * DOC: i915_context_create and i915_context_free tracepoints
+ *
+ * These tracepoints are used to track creation and deletion of contexts.
+ * If full ppgtt is enabled, they also print the address of the vm assigned to
+ * the context.
+ */
+DECLARE_EVENT_CLASS(i915_context,
+	TP_PROTO(struct intel_context *ctx),
+	TP_ARGS(ctx),
+
+	TP_STRUCT__entry(
+			__field(u32, dev)
+			__field(struct intel_context *, ctx)
+			__field(struct i915_address_space *, vm)
+	),
+
+	TP_fast_assign(
+			__entry->ctx = ctx;
+			__entry->vm = ctx->ppgtt ? &ctx->ppgtt->base : NULL;
+			__entry->dev = ctx->file_priv->dev_priv->dev->primary->index;
+	),
+
+	TP_printk("dev=%u, ctx=%p, ctx_vm=%p",
+		  __entry->dev, __entry->ctx, __entry->vm)
+)
+
+DEFINE_EVENT(i915_context, i915_context_create,
+	TP_PROTO(struct intel_context *ctx),
+	TP_ARGS(ctx)
+);
+
+DEFINE_EVENT(i915_context, i915_context_free,
+	TP_PROTO(struct intel_context *ctx),
+	TP_ARGS(ctx)
+);
+
+/**
+ * DOC: switch_mm tracepoint
+ *
+ * This tracepoint allows tracking of the mm switch, which is an important point
+ * in the lifetime of the vm in the legacy submission path. This tracepoint is
+ * called only if full ppgtt is enabled.
+ */
+TRACE_EVENT(switch_mm,
+	TP_PROTO(struct intel_engine_cs *ring, struct intel_context *to),
+
+	TP_ARGS(ring, to),
+
+	TP_STRUCT__entry(
+			__field(u32, ring)
+			__field(struct intel_context *, to)
+			__field(struct i915_address_space *, vm)
+			__field(u32, dev)
+	),
+
+	TP_fast_assign(
+			__entry->ring = ring->id;
+			__entry->to = to;
+			__entry->vm = to->ppgtt? &to->ppgtt->base : NULL;
+			__entry->dev = ring->dev->primary->index;
+	),
+
+	TP_printk("dev=%u, ring=%u, ctx=%p, ctx_vm=%p",
+		  __entry->dev, __entry->ring, __entry->to, __entry->vm)
+);
+
 #endif /* _I915_TRACE_H_ */
 
 /* This part must be outside protection */
diff --git a/drivers/gpu/drm/i915/i915_ums.c b/drivers/gpu/drm/i915/i915_ums.c
index 480da593e6c0..d10fe3e9c49f 100644
--- a/drivers/gpu/drm/i915/i915_ums.c
+++ b/drivers/gpu/drm/i915/i915_ums.c
@@ -270,6 +270,12 @@ void i915_save_display_reg(struct drm_device *dev)
 	}
 	/* FIXME: regfile.save TV & SDVO state */
 
+	/* Panel fitter */
+	if (!IS_I830(dev) && !IS_845G(dev) && !HAS_PCH_SPLIT(dev)) {
+		dev_priv->regfile.savePFIT_CONTROL = I915_READ(PFIT_CONTROL);
+		dev_priv->regfile.savePFIT_PGM_RATIOS = I915_READ(PFIT_PGM_RATIOS);
+	}
+
 	/* Backlight */
 	if (INTEL_INFO(dev)->gen <= 4)
 		pci_read_config_byte(dev->pdev, PCI_LBPC,
@@ -284,6 +290,7 @@ void i915_save_display_reg(struct drm_device *dev)
 		dev_priv->regfile.saveBLC_PWM_CTL = I915_READ(BLC_PWM_CTL);
 		if (INTEL_INFO(dev)->gen >= 4)
 			dev_priv->regfile.saveBLC_PWM_CTL2 = I915_READ(BLC_PWM_CTL2);
+		dev_priv->regfile.saveBLC_HIST_CTL = I915_READ(BLC_HIST_CTL);
 	}
 
 	return;
@@ -313,6 +320,13 @@ void i915_restore_display_reg(struct drm_device *dev)
 		if (INTEL_INFO(dev)->gen >= 4)
 			I915_WRITE(BLC_PWM_CTL2, dev_priv->regfile.saveBLC_PWM_CTL2);
 		I915_WRITE(BLC_PWM_CTL, dev_priv->regfile.saveBLC_PWM_CTL);
+		I915_WRITE(BLC_HIST_CTL, dev_priv->regfile.saveBLC_HIST_CTL);
+	}
+
+	/* Panel fitter */
+	if (!IS_I830(dev) && !IS_845G(dev) && !HAS_PCH_SPLIT(dev)) {
+		I915_WRITE(PFIT_PGM_RATIOS, dev_priv->regfile.savePFIT_PGM_RATIOS);
+		I915_WRITE(PFIT_CONTROL, dev_priv->regfile.savePFIT_CONTROL);
 	}
 
 	/* Display port ratios (must be done before clock is set) */
diff --git a/drivers/gpu/drm/i915/intel_audio.c b/drivers/gpu/drm/i915/intel_audio.c
new file mode 100644
index 000000000000..2c7ed5cb29c0
--- /dev/null
+++ b/drivers/gpu/drm/i915/intel_audio.c
@@ -0,0 +1,463 @@
+/*
+ * Copyright © 2014 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ */
+
+#include <linux/kernel.h>
+
+#include <drm/drmP.h>
+#include <drm/drm_edid.h>
+#include "intel_drv.h"
+#include "i915_drv.h"
+
+/**
+ * DOC: High Definition Audio over HDMI and Display Port
+ *
+ * The graphics and audio drivers together support High Definition Audio over
+ * HDMI and Display Port. The audio programming sequences are divided into audio
+ * codec and controller enable and disable sequences. The graphics driver
+ * handles the audio codec sequences, while the audio driver handles the audio
+ * controller sequences.
+ *
+ * The disable sequences must be performed before disabling the transcoder or
+ * port. The enable sequences may only be performed after enabling the
+ * transcoder and port, and after completed link training.
+ *
+ * The codec and controller sequences could be done either parallel or serial,
+ * but generally the ELDV/PD change in the codec sequence indicates to the audio
+ * driver that the controller sequence should start. Indeed, most of the
+ * co-operation between the graphics and audio drivers is handled via audio
+ * related registers. (The notable exception is the power management, not
+ * covered here.)
+ */
+
+static const struct {
+	int clock;
+	u32 config;
+} hdmi_audio_clock[] = {
+	{ DIV_ROUND_UP(25200 * 1000, 1001), AUD_CONFIG_PIXEL_CLOCK_HDMI_25175 },
+	{ 25200, AUD_CONFIG_PIXEL_CLOCK_HDMI_25200 }, /* default per bspec */
+	{ 27000, AUD_CONFIG_PIXEL_CLOCK_HDMI_27000 },
+	{ 27000 * 1001 / 1000, AUD_CONFIG_PIXEL_CLOCK_HDMI_27027 },
+	{ 54000, AUD_CONFIG_PIXEL_CLOCK_HDMI_54000 },
+	{ 54000 * 1001 / 1000, AUD_CONFIG_PIXEL_CLOCK_HDMI_54054 },
+	{ DIV_ROUND_UP(74250 * 1000, 1001), AUD_CONFIG_PIXEL_CLOCK_HDMI_74176 },
+	{ 74250, AUD_CONFIG_PIXEL_CLOCK_HDMI_74250 },
+	{ DIV_ROUND_UP(148500 * 1000, 1001), AUD_CONFIG_PIXEL_CLOCK_HDMI_148352 },
+	{ 148500, AUD_CONFIG_PIXEL_CLOCK_HDMI_148500 },
+};
+
+/* get AUD_CONFIG_PIXEL_CLOCK_HDMI_* value for mode */
+static u32 audio_config_hdmi_pixel_clock(struct drm_display_mode *mode)
+{
+	int i;
+
+	for (i = 0; i < ARRAY_SIZE(hdmi_audio_clock); i++) {
+		if (mode->clock == hdmi_audio_clock[i].clock)
+			break;
+	}
+
+	if (i == ARRAY_SIZE(hdmi_audio_clock)) {
+		DRM_DEBUG_KMS("HDMI audio pixel clock setting for %d not found, falling back to defaults\n", mode->clock);
+		i = 1;
+	}
+
+	DRM_DEBUG_KMS("Configuring HDMI audio for pixel clock %d (0x%08x)\n",
+		      hdmi_audio_clock[i].clock,
+		      hdmi_audio_clock[i].config);
+
+	return hdmi_audio_clock[i].config;
+}
+
+static bool intel_eld_uptodate(struct drm_connector *connector,
+			       int reg_eldv, uint32_t bits_eldv,
+			       int reg_elda, uint32_t bits_elda,
+			       int reg_edid)
+{
+	struct drm_i915_private *dev_priv = connector->dev->dev_private;
+	uint8_t *eld = connector->eld;
+	uint32_t tmp;
+	int i;
+
+	tmp = I915_READ(reg_eldv);
+	tmp &= bits_eldv;
+
+	if (!tmp)
+		return false;
+
+	tmp = I915_READ(reg_elda);
+	tmp &= ~bits_elda;
+	I915_WRITE(reg_elda, tmp);
+
+	for (i = 0; i < drm_eld_size(eld) / 4; i++)
+		if (I915_READ(reg_edid) != *((uint32_t *)eld + i))
+			return false;
+
+	return true;
+}
+
+static void g4x_audio_codec_disable(struct intel_encoder *encoder)
+{
+	struct drm_i915_private *dev_priv = encoder->base.dev->dev_private;
+	uint32_t eldv, tmp;
+
+	DRM_DEBUG_KMS("Disable audio codec\n");
+
+	tmp = I915_READ(G4X_AUD_VID_DID);
+	if (tmp == INTEL_AUDIO_DEVBLC || tmp == INTEL_AUDIO_DEVCL)
+		eldv = G4X_ELDV_DEVCL_DEVBLC;
+	else
+		eldv = G4X_ELDV_DEVCTG;
+
+	/* Invalidate ELD */
+	tmp = I915_READ(G4X_AUD_CNTL_ST);
+	tmp &= ~eldv;
+	I915_WRITE(G4X_AUD_CNTL_ST, tmp);
+}
+
+static void g4x_audio_codec_enable(struct drm_connector *connector,
+				   struct intel_encoder *encoder,
+				   struct drm_display_mode *mode)
+{
+	struct drm_i915_private *dev_priv = connector->dev->dev_private;
+	uint8_t *eld = connector->eld;
+	uint32_t eldv;
+	uint32_t tmp;
+	int len, i;
+
+	DRM_DEBUG_KMS("Enable audio codec, %u bytes ELD\n", eld[2]);
+
+	tmp = I915_READ(G4X_AUD_VID_DID);
+	if (tmp == INTEL_AUDIO_DEVBLC || tmp == INTEL_AUDIO_DEVCL)
+		eldv = G4X_ELDV_DEVCL_DEVBLC;
+	else
+		eldv = G4X_ELDV_DEVCTG;
+
+	if (intel_eld_uptodate(connector,
+			       G4X_AUD_CNTL_ST, eldv,
+			       G4X_AUD_CNTL_ST, G4X_ELD_ADDR_MASK,
+			       G4X_HDMIW_HDMIEDID))
+		return;
+
+	tmp = I915_READ(G4X_AUD_CNTL_ST);
+	tmp &= ~(eldv | G4X_ELD_ADDR_MASK);
+	len = (tmp >> 9) & 0x1f;		/* ELD buffer size */
+	I915_WRITE(G4X_AUD_CNTL_ST, tmp);
+
+	len = min(drm_eld_size(eld) / 4, len);
+	DRM_DEBUG_DRIVER("ELD size %d\n", len);
+	for (i = 0; i < len; i++)
+		I915_WRITE(G4X_HDMIW_HDMIEDID, *((uint32_t *)eld + i));
+
+	tmp = I915_READ(G4X_AUD_CNTL_ST);
+	tmp |= eldv;
+	I915_WRITE(G4X_AUD_CNTL_ST, tmp);
+}
+
+static void hsw_audio_codec_disable(struct intel_encoder *encoder)
+{
+	struct drm_i915_private *dev_priv = encoder->base.dev->dev_private;
+	struct intel_crtc *intel_crtc = to_intel_crtc(encoder->base.crtc);
+	enum pipe pipe = intel_crtc->pipe;
+	uint32_t tmp;
+
+	DRM_DEBUG_KMS("Disable audio codec on pipe %c\n", pipe_name(pipe));
+
+	/* Disable timestamps */
+	tmp = I915_READ(HSW_AUD_CFG(pipe));
+	tmp &= ~AUD_CONFIG_N_VALUE_INDEX;
+	tmp |= AUD_CONFIG_N_PROG_ENABLE;
+	tmp &= ~AUD_CONFIG_UPPER_N_MASK;
+	tmp &= ~AUD_CONFIG_LOWER_N_MASK;
+	if (intel_pipe_has_type(intel_crtc, INTEL_OUTPUT_DISPLAYPORT))
+		tmp |= AUD_CONFIG_N_VALUE_INDEX;
+	I915_WRITE(HSW_AUD_CFG(pipe), tmp);
+
+	/* Invalidate ELD */
+	tmp = I915_READ(HSW_AUD_PIN_ELD_CP_VLD);
+	tmp &= ~AUDIO_ELD_VALID(pipe);
+	tmp &= ~AUDIO_OUTPUT_ENABLE(pipe);
+	I915_WRITE(HSW_AUD_PIN_ELD_CP_VLD, tmp);
+}
+
+static void hsw_audio_codec_enable(struct drm_connector *connector,
+				   struct intel_encoder *encoder,
+				   struct drm_display_mode *mode)
+{
+	struct drm_i915_private *dev_priv = connector->dev->dev_private;
+	struct intel_crtc *intel_crtc = to_intel_crtc(encoder->base.crtc);
+	enum pipe pipe = intel_crtc->pipe;
+	const uint8_t *eld = connector->eld;
+	uint32_t tmp;
+	int len, i;
+
+	DRM_DEBUG_KMS("Enable audio codec on pipe %c, %u bytes ELD\n",
+		      pipe_name(pipe), drm_eld_size(eld));
+
+	/* Enable audio presence detect, invalidate ELD */
+	tmp = I915_READ(HSW_AUD_PIN_ELD_CP_VLD);
+	tmp |= AUDIO_OUTPUT_ENABLE(pipe);
+	tmp &= ~AUDIO_ELD_VALID(pipe);
+	I915_WRITE(HSW_AUD_PIN_ELD_CP_VLD, tmp);
+
+	/*
+	 * FIXME: We're supposed to wait for vblank here, but we have vblanks
+	 * disabled during the mode set. The proper fix would be to push the
+	 * rest of the setup into a vblank work item, queued here, but the
+	 * infrastructure is not there yet.
+	 */
+
+	/* Reset ELD write address */
+	tmp = I915_READ(HSW_AUD_DIP_ELD_CTRL(pipe));
+	tmp &= ~IBX_ELD_ADDRESS_MASK;
+	I915_WRITE(HSW_AUD_DIP_ELD_CTRL(pipe), tmp);
+
+	/* Up to 84 bytes of hw ELD buffer */
+	len = min(drm_eld_size(eld), 84);
+	for (i = 0; i < len / 4; i++)
+		I915_WRITE(HSW_AUD_EDID_DATA(pipe), *((uint32_t *)eld + i));
+
+	/* ELD valid */
+	tmp = I915_READ(HSW_AUD_PIN_ELD_CP_VLD);
+	tmp |= AUDIO_ELD_VALID(pipe);
+	I915_WRITE(HSW_AUD_PIN_ELD_CP_VLD, tmp);
+
+	/* Enable timestamps */
+	tmp = I915_READ(HSW_AUD_CFG(pipe));
+	tmp &= ~AUD_CONFIG_N_VALUE_INDEX;
+	tmp &= ~AUD_CONFIG_N_PROG_ENABLE;
+	tmp &= ~AUD_CONFIG_PIXEL_CLOCK_HDMI_MASK;
+	if (intel_pipe_has_type(intel_crtc, INTEL_OUTPUT_DISPLAYPORT))
+		tmp |= AUD_CONFIG_N_VALUE_INDEX;
+	else
+		tmp |= audio_config_hdmi_pixel_clock(mode);
+	I915_WRITE(HSW_AUD_CFG(pipe), tmp);
+}
+
+static void ilk_audio_codec_disable(struct intel_encoder *encoder)
+{
+	struct drm_i915_private *dev_priv = encoder->base.dev->dev_private;
+	struct intel_crtc *intel_crtc = to_intel_crtc(encoder->base.crtc);
+	struct intel_digital_port *intel_dig_port =
+		enc_to_dig_port(&encoder->base);
+	enum port port = intel_dig_port->port;
+	enum pipe pipe = intel_crtc->pipe;
+	uint32_t tmp, eldv;
+	int aud_config;
+	int aud_cntrl_st2;
+
+	DRM_DEBUG_KMS("Disable audio codec on port %c, pipe %c\n",
+		      port_name(port), pipe_name(pipe));
+
+	if (HAS_PCH_IBX(dev_priv->dev)) {
+		aud_config = IBX_AUD_CFG(pipe);
+		aud_cntrl_st2 = IBX_AUD_CNTL_ST2;
+	} else if (IS_VALLEYVIEW(dev_priv)) {
+		aud_config = VLV_AUD_CFG(pipe);
+		aud_cntrl_st2 = VLV_AUD_CNTL_ST2;
+	} else {
+		aud_config = CPT_AUD_CFG(pipe);
+		aud_cntrl_st2 = CPT_AUD_CNTRL_ST2;
+	}
+
+	/* Disable timestamps */
+	tmp = I915_READ(aud_config);
+	tmp &= ~AUD_CONFIG_N_VALUE_INDEX;
+	tmp |= AUD_CONFIG_N_PROG_ENABLE;
+	tmp &= ~AUD_CONFIG_UPPER_N_MASK;
+	tmp &= ~AUD_CONFIG_LOWER_N_MASK;
+	if (intel_pipe_has_type(intel_crtc, INTEL_OUTPUT_DISPLAYPORT))
+		tmp |= AUD_CONFIG_N_VALUE_INDEX;
+	I915_WRITE(aud_config, tmp);
+
+	if (WARN_ON(!port)) {
+		eldv = IBX_ELD_VALID(PORT_B) | IBX_ELD_VALID(PORT_C) |
+			IBX_ELD_VALID(PORT_D);
+	} else {
+		eldv = IBX_ELD_VALID(port);
+	}
+
+	/* Invalidate ELD */
+	tmp = I915_READ(aud_cntrl_st2);
+	tmp &= ~eldv;
+	I915_WRITE(aud_cntrl_st2, tmp);
+}
+
+static void ilk_audio_codec_enable(struct drm_connector *connector,
+				   struct intel_encoder *encoder,
+				   struct drm_display_mode *mode)
+{
+	struct drm_i915_private *dev_priv = connector->dev->dev_private;
+	struct intel_crtc *intel_crtc = to_intel_crtc(encoder->base.crtc);
+	struct intel_digital_port *intel_dig_port =
+		enc_to_dig_port(&encoder->base);
+	enum port port = intel_dig_port->port;
+	enum pipe pipe = intel_crtc->pipe;
+	uint8_t *eld = connector->eld;
+	uint32_t eldv;
+	uint32_t tmp;
+	int len, i;
+	int hdmiw_hdmiedid;
+	int aud_config;
+	int aud_cntl_st;
+	int aud_cntrl_st2;
+
+	DRM_DEBUG_KMS("Enable audio codec on port %c, pipe %c, %u bytes ELD\n",
+		      port_name(port), pipe_name(pipe), drm_eld_size(eld));
+
+	/*
+	 * FIXME: We're supposed to wait for vblank here, but we have vblanks
+	 * disabled during the mode set. The proper fix would be to push the
+	 * rest of the setup into a vblank work item, queued here, but the
+	 * infrastructure is not there yet.
+	 */
+
+	if (HAS_PCH_IBX(connector->dev)) {
+		hdmiw_hdmiedid = IBX_HDMIW_HDMIEDID(pipe);
+		aud_config = IBX_AUD_CFG(pipe);
+		aud_cntl_st = IBX_AUD_CNTL_ST(pipe);
+		aud_cntrl_st2 = IBX_AUD_CNTL_ST2;
+	} else if (IS_VALLEYVIEW(connector->dev)) {
+		hdmiw_hdmiedid = VLV_HDMIW_HDMIEDID(pipe);
+		aud_config = VLV_AUD_CFG(pipe);
+		aud_cntl_st = VLV_AUD_CNTL_ST(pipe);
+		aud_cntrl_st2 = VLV_AUD_CNTL_ST2;
+	} else {
+		hdmiw_hdmiedid = CPT_HDMIW_HDMIEDID(pipe);
+		aud_config = CPT_AUD_CFG(pipe);
+		aud_cntl_st = CPT_AUD_CNTL_ST(pipe);
+		aud_cntrl_st2 = CPT_AUD_CNTRL_ST2;
+	}
+
+	if (WARN_ON(!port)) {
+		eldv = IBX_ELD_VALID(PORT_B) | IBX_ELD_VALID(PORT_C) |
+			IBX_ELD_VALID(PORT_D);
+	} else {
+		eldv = IBX_ELD_VALID(port);
+	}
+
+	/* Invalidate ELD */
+	tmp = I915_READ(aud_cntrl_st2);
+	tmp &= ~eldv;
+	I915_WRITE(aud_cntrl_st2, tmp);
+
+	/* Reset ELD write address */
+	tmp = I915_READ(aud_cntl_st);
+	tmp &= ~IBX_ELD_ADDRESS_MASK;
+	I915_WRITE(aud_cntl_st, tmp);
+
+	/* Up to 84 bytes of hw ELD buffer */
+	len = min(drm_eld_size(eld), 84);
+	for (i = 0; i < len / 4; i++)
+		I915_WRITE(hdmiw_hdmiedid, *((uint32_t *)eld + i));
+
+	/* ELD valid */
+	tmp = I915_READ(aud_cntrl_st2);
+	tmp |= eldv;
+	I915_WRITE(aud_cntrl_st2, tmp);
+
+	/* Enable timestamps */
+	tmp = I915_READ(aud_config);
+	tmp &= ~AUD_CONFIG_N_VALUE_INDEX;
+	tmp &= ~AUD_CONFIG_N_PROG_ENABLE;
+	tmp &= ~AUD_CONFIG_PIXEL_CLOCK_HDMI_MASK;
+	if (intel_pipe_has_type(intel_crtc, INTEL_OUTPUT_DISPLAYPORT))
+		tmp |= AUD_CONFIG_N_VALUE_INDEX;
+	else
+		tmp |= audio_config_hdmi_pixel_clock(mode);
+	I915_WRITE(aud_config, tmp);
+}
+
+/**
+ * intel_audio_codec_enable - Enable the audio codec for HD audio
+ * @intel_encoder: encoder on which to enable audio
+ *
+ * The enable sequences may only be performed after enabling the transcoder and
+ * port, and after completed link training.
+ */
+void intel_audio_codec_enable(struct intel_encoder *intel_encoder)
+{
+	struct drm_encoder *encoder = &intel_encoder->base;
+	struct intel_crtc *crtc = to_intel_crtc(encoder->crtc);
+	struct drm_display_mode *mode = &crtc->config.adjusted_mode;
+	struct drm_connector *connector;
+	struct drm_device *dev = encoder->dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+
+	connector = drm_select_eld(encoder, mode);
+	if (!connector)
+		return;
+
+	DRM_DEBUG_DRIVER("ELD on [CONNECTOR:%d:%s], [ENCODER:%d:%s]\n",
+			 connector->base.id,
+			 connector->name,
+			 connector->encoder->base.id,
+			 connector->encoder->name);
+
+	/* ELD Conn_Type */
+	connector->eld[5] &= ~(3 << 2);
+	if (intel_pipe_has_type(crtc, INTEL_OUTPUT_DISPLAYPORT))
+		connector->eld[5] |= (1 << 2);
+
+	connector->eld[6] = drm_av_sync_delay(connector, mode) / 2;
+
+	if (dev_priv->display.audio_codec_enable)
+		dev_priv->display.audio_codec_enable(connector, intel_encoder, mode);
+}
+
+/**
+ * intel_audio_codec_disable - Disable the audio codec for HD audio
+ * @encoder: encoder on which to disable audio
+ *
+ * The disable sequences must be performed before disabling the transcoder or
+ * port.
+ */
+void intel_audio_codec_disable(struct intel_encoder *encoder)
+{
+	struct drm_device *dev = encoder->base.dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+
+	if (dev_priv->display.audio_codec_disable)
+		dev_priv->display.audio_codec_disable(encoder);
+}
+
+/**
+ * intel_init_audio - Set up chip specific audio functions
+ * @dev: drm device
+ */
+void intel_init_audio(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+
+	if (IS_G4X(dev)) {
+		dev_priv->display.audio_codec_enable = g4x_audio_codec_enable;
+		dev_priv->display.audio_codec_disable = g4x_audio_codec_disable;
+	} else if (IS_VALLEYVIEW(dev)) {
+		dev_priv->display.audio_codec_enable = ilk_audio_codec_enable;
+		dev_priv->display.audio_codec_disable = ilk_audio_codec_disable;
+	} else if (IS_HASWELL(dev) || INTEL_INFO(dev)->gen >= 8) {
+		dev_priv->display.audio_codec_enable = hsw_audio_codec_enable;
+		dev_priv->display.audio_codec_disable = hsw_audio_codec_disable;
+	} else if (HAS_PCH_SPLIT(dev)) {
+		dev_priv->display.audio_codec_enable = ilk_audio_codec_enable;
+		dev_priv->display.audio_codec_disable = ilk_audio_codec_disable;
+	}
+}
diff --git a/drivers/gpu/drm/i915/intel_bios.c b/drivers/gpu/drm/i915/intel_bios.c
index eee79e1c3222..a4bd90f36a03 100644
--- a/drivers/gpu/drm/i915/intel_bios.c
+++ b/drivers/gpu/drm/i915/intel_bios.c
@@ -627,16 +627,16 @@ parse_edp(struct drm_i915_private *dev_priv, struct bdb_header *bdb)
 
 	switch (edp_link_params->preemphasis) {
 	case EDP_PREEMPHASIS_NONE:
-		dev_priv->vbt.edp_preemphasis = DP_TRAIN_PRE_EMPHASIS_0;
+		dev_priv->vbt.edp_preemphasis = DP_TRAIN_PRE_EMPH_LEVEL_0;
 		break;
 	case EDP_PREEMPHASIS_3_5dB:
-		dev_priv->vbt.edp_preemphasis = DP_TRAIN_PRE_EMPHASIS_3_5;
+		dev_priv->vbt.edp_preemphasis = DP_TRAIN_PRE_EMPH_LEVEL_1;
 		break;
 	case EDP_PREEMPHASIS_6dB:
-		dev_priv->vbt.edp_preemphasis = DP_TRAIN_PRE_EMPHASIS_6;
+		dev_priv->vbt.edp_preemphasis = DP_TRAIN_PRE_EMPH_LEVEL_2;
 		break;
 	case EDP_PREEMPHASIS_9_5dB:
-		dev_priv->vbt.edp_preemphasis = DP_TRAIN_PRE_EMPHASIS_9_5;
+		dev_priv->vbt.edp_preemphasis = DP_TRAIN_PRE_EMPH_LEVEL_3;
 		break;
 	default:
 		DRM_DEBUG_KMS("VBT has unknown eDP pre-emphasis value %u\n",
@@ -646,16 +646,16 @@ parse_edp(struct drm_i915_private *dev_priv, struct bdb_header *bdb)
 
 	switch (edp_link_params->vswing) {
 	case EDP_VSWING_0_4V:
-		dev_priv->vbt.edp_vswing = DP_TRAIN_VOLTAGE_SWING_400;
+		dev_priv->vbt.edp_vswing = DP_TRAIN_VOLTAGE_SWING_LEVEL_0;
 		break;
 	case EDP_VSWING_0_6V:
-		dev_priv->vbt.edp_vswing = DP_TRAIN_VOLTAGE_SWING_600;
+		dev_priv->vbt.edp_vswing = DP_TRAIN_VOLTAGE_SWING_LEVEL_1;
 		break;
 	case EDP_VSWING_0_8V:
-		dev_priv->vbt.edp_vswing = DP_TRAIN_VOLTAGE_SWING_800;
+		dev_priv->vbt.edp_vswing = DP_TRAIN_VOLTAGE_SWING_LEVEL_2;
 		break;
 	case EDP_VSWING_1_2V:
-		dev_priv->vbt.edp_vswing = DP_TRAIN_VOLTAGE_SWING_1200;
+		dev_priv->vbt.edp_vswing = DP_TRAIN_VOLTAGE_SWING_LEVEL_3;
 		break;
 	default:
 		DRM_DEBUG_KMS("VBT has unknown eDP voltage swing value %u\n",
@@ -946,7 +946,7 @@ static void parse_ddi_port(struct drm_i915_private *dev_priv, enum port port,
 		DRM_DEBUG_KMS("Analog port %c is also DP or TMDS compatible\n",
 			      port_name(port));
 	if (is_dvi && (port == PORT_A || port == PORT_E))
-		DRM_DEBUG_KMS("Port %c is TMDS compabile\n", port_name(port));
+		DRM_DEBUG_KMS("Port %c is TMDS compatible\n", port_name(port));
 	if (!is_dvi && !is_dp && !is_crt)
 		DRM_DEBUG_KMS("Port %c is not DP/TMDS/CRT compatible\n",
 			      port_name(port));
@@ -976,12 +976,10 @@ static void parse_ddi_port(struct drm_i915_private *dev_priv, enum port port,
 	if (bdb->version >= 158) {
 		/* The VBT HDMI level shift values match the table we have. */
 		hdmi_level_shift = child->raw[7] & 0xF;
-		if (hdmi_level_shift < 0xC) {
-			DRM_DEBUG_KMS("VBT HDMI level shift for port %c: %d\n",
-				      port_name(port),
-				      hdmi_level_shift);
-			info->hdmi_level_shift = hdmi_level_shift;
-		}
+		DRM_DEBUG_KMS("VBT HDMI level shift for port %c: %d\n",
+			      port_name(port),
+			      hdmi_level_shift);
+		info->hdmi_level_shift = hdmi_level_shift;
 	}
 }
 
@@ -1114,8 +1112,7 @@ init_vbt_defaults(struct drm_i915_private *dev_priv)
 		struct ddi_vbt_port_info *info =
 			&dev_priv->vbt.ddi_port_info[port];
 
-		/* Recommended BSpec default: 800mV 0dB. */
-		info->hdmi_level_shift = 6;
+		info->hdmi_level_shift = HDMI_LEVEL_SHIFT_UNKNOWN;
 
 		info->supports_dvi = (port != PORT_A && port != PORT_E);
 		info->supports_hdmi = info->supports_dvi;
diff --git a/drivers/gpu/drm/i915/intel_bios.h b/drivers/gpu/drm/i915/intel_bios.h
index b98667796337..7603765c91fc 100644
--- a/drivers/gpu/drm/i915/intel_bios.h
+++ b/drivers/gpu/drm/i915/intel_bios.h
@@ -46,7 +46,7 @@ struct bdb_header {
 	u16 version;			/**< decimal */
 	u16 header_size;		/**< in bytes */
 	u16 bdb_size;			/**< in bytes */
-};
+} __packed;
 
 /* strictly speaking, this is a "skip" block, but it has interesting info */
 struct vbios_data {
@@ -252,7 +252,7 @@ union child_device_config {
 	/* This one should also be safe to use anywhere, even without version
 	 * checks. */
 	struct common_child_dev_config common;
-};
+} __packed;
 
 struct bdb_general_definitions {
 	/* DDC GPIO */
@@ -802,7 +802,8 @@ struct mipi_config {
 
 	u16 rsvd4;
 
-	u8 rsvd5[5];
+	u8 rsvd5;
+	u32 target_burst_mode_freq;
 	u32 dsi_ddr_clk;
 	u32 bridge_ref_clk;
 
@@ -887,12 +888,12 @@ struct mipi_pps_data {
 	u16 bl_disable_delay;
 	u16 panel_off_delay;
 	u16 panel_power_cycle_delay;
-};
+} __packed;
 
 struct bdb_mipi_config {
 	struct mipi_config config[MAX_MIPI_CONFIGURATIONS];
 	struct mipi_pps_data pps[MAX_MIPI_CONFIGURATIONS];
-};
+} __packed;
 
 /* Block 53 contains MIPI sequences as needed by the panel
  * for enabling it. This block can be variable in size and
@@ -901,7 +902,7 @@ struct bdb_mipi_config {
 struct bdb_mipi_sequence {
 	u8 version;
 	u8 data[0];
-};
+} __packed;
 
 /* MIPI Sequnece Block definitions */
 enum mipi_seq {
diff --git a/drivers/gpu/drm/i915/intel_crt.c b/drivers/gpu/drm/i915/intel_crt.c
index 9212e6504e0f..a9af9a4866db 100644
--- a/drivers/gpu/drm/i915/intel_crt.c
+++ b/drivers/gpu/drm/i915/intel_crt.c
@@ -72,7 +72,7 @@ static bool intel_crt_get_hw_state(struct intel_encoder *encoder,
 	u32 tmp;
 
 	power_domain = intel_display_port_power_domain(encoder);
-	if (!intel_display_power_enabled(dev_priv, power_domain))
+	if (!intel_display_power_is_enabled(dev_priv, power_domain))
 		return false;
 
 	tmp = I915_READ(crt->adpa_reg);
@@ -775,7 +775,7 @@ static void intel_crt_reset(struct drm_connector *connector)
 		I915_WRITE(crt->adpa_reg, adpa);
 		POSTING_READ(crt->adpa_reg);
 
-		DRM_DEBUG_KMS("pch crt adpa set to 0x%x\n", adpa);
+		DRM_DEBUG_KMS("crt adpa set to 0x%x\n", adpa);
 		crt->force_hotplug_required = 1;
 	}
 
diff --git a/drivers/gpu/drm/i915/intel_ddi.c b/drivers/gpu/drm/i915/intel_ddi.c
index 5db0b5552e39..e6b45cd150d3 100644
--- a/drivers/gpu/drm/i915/intel_ddi.c
+++ b/drivers/gpu/drm/i915/intel_ddi.c
@@ -28,87 +28,129 @@
 #include "i915_drv.h"
 #include "intel_drv.h"
 
+struct ddi_buf_trans {
+	u32 trans1;	/* balance leg enable, de-emph level */
+	u32 trans2;	/* vref sel, vswing */
+};
+
 /* HDMI/DVI modes ignore everything but the last 2 items. So we share
  * them for both DP and FDI transports, allowing those ports to
  * automatically adapt to HDMI connections as well
  */
-static const u32 hsw_ddi_translations_dp[] = {
-	0x00FFFFFF, 0x0006000E,		/* DP parameters */
-	0x00D75FFF, 0x0005000A,
-	0x00C30FFF, 0x00040006,
-	0x80AAAFFF, 0x000B0000,
-	0x00FFFFFF, 0x0005000A,
-	0x00D75FFF, 0x000C0004,
-	0x80C30FFF, 0x000B0000,
-	0x00FFFFFF, 0x00040006,
-	0x80D75FFF, 0x000B0000,
+static const struct ddi_buf_trans hsw_ddi_translations_dp[] = {
+	{ 0x00FFFFFF, 0x0006000E },
+	{ 0x00D75FFF, 0x0005000A },
+	{ 0x00C30FFF, 0x00040006 },
+	{ 0x80AAAFFF, 0x000B0000 },
+	{ 0x00FFFFFF, 0x0005000A },
+	{ 0x00D75FFF, 0x000C0004 },
+	{ 0x80C30FFF, 0x000B0000 },
+	{ 0x00FFFFFF, 0x00040006 },
+	{ 0x80D75FFF, 0x000B0000 },
+};
+
+static const struct ddi_buf_trans hsw_ddi_translations_fdi[] = {
+	{ 0x00FFFFFF, 0x0007000E },
+	{ 0x00D75FFF, 0x000F000A },
+	{ 0x00C30FFF, 0x00060006 },
+	{ 0x00AAAFFF, 0x001E0000 },
+	{ 0x00FFFFFF, 0x000F000A },
+	{ 0x00D75FFF, 0x00160004 },
+	{ 0x00C30FFF, 0x001E0000 },
+	{ 0x00FFFFFF, 0x00060006 },
+	{ 0x00D75FFF, 0x001E0000 },
+};
+
+static const struct ddi_buf_trans hsw_ddi_translations_hdmi[] = {
+					/* Idx	NT mV d	T mV d	db	*/
+	{ 0x00FFFFFF, 0x0006000E },	/* 0:	400	400	0	*/
+	{ 0x00E79FFF, 0x000E000C },	/* 1:	400	500	2	*/
+	{ 0x00D75FFF, 0x0005000A },	/* 2:	400	600	3.5	*/
+	{ 0x00FFFFFF, 0x0005000A },	/* 3:	600	600	0	*/
+	{ 0x00E79FFF, 0x001D0007 },	/* 4:	600	750	2	*/
+	{ 0x00D75FFF, 0x000C0004 },	/* 5:	600	900	3.5	*/
+	{ 0x00FFFFFF, 0x00040006 },	/* 6:	800	800	0	*/
+	{ 0x80E79FFF, 0x00030002 },	/* 7:	800	1000	2	*/
+	{ 0x00FFFFFF, 0x00140005 },	/* 8:	850	850	0	*/
+	{ 0x00FFFFFF, 0x000C0004 },	/* 9:	900	900	0	*/
+	{ 0x00FFFFFF, 0x001C0003 },	/* 10:	950	950	0	*/
+	{ 0x80FFFFFF, 0x00030002 },	/* 11:	1000	1000	0	*/
+};
+
+static const struct ddi_buf_trans bdw_ddi_translations_edp[] = {
+	{ 0x00FFFFFF, 0x00000012 },
+	{ 0x00EBAFFF, 0x00020011 },
+	{ 0x00C71FFF, 0x0006000F },
+	{ 0x00AAAFFF, 0x000E000A },
+	{ 0x00FFFFFF, 0x00020011 },
+	{ 0x00DB6FFF, 0x0005000F },
+	{ 0x00BEEFFF, 0x000A000C },
+	{ 0x00FFFFFF, 0x0005000F },
+	{ 0x00DB6FFF, 0x000A000C },
 };
 
-static const u32 hsw_ddi_translations_fdi[] = {
-	0x00FFFFFF, 0x0007000E,		/* FDI parameters */
-	0x00D75FFF, 0x000F000A,
-	0x00C30FFF, 0x00060006,
-	0x00AAAFFF, 0x001E0000,
-	0x00FFFFFF, 0x000F000A,
-	0x00D75FFF, 0x00160004,
-	0x00C30FFF, 0x001E0000,
-	0x00FFFFFF, 0x00060006,
-	0x00D75FFF, 0x001E0000,
+static const struct ddi_buf_trans bdw_ddi_translations_dp[] = {
+	{ 0x00FFFFFF, 0x0007000E },
+	{ 0x00D75FFF, 0x000E000A },
+	{ 0x00BEFFFF, 0x00140006 },
+	{ 0x80B2CFFF, 0x001B0002 },
+	{ 0x00FFFFFF, 0x000E000A },
+	{ 0x00DB6FFF, 0x00160005 },
+	{ 0x80C71FFF, 0x001A0002 },
+	{ 0x00F7DFFF, 0x00180004 },
+	{ 0x80D75FFF, 0x001B0002 },
 };
 
-static const u32 hsw_ddi_translations_hdmi[] = {
-				/* Idx	NT mV diff	T mV diff	db  */
-	0x00FFFFFF, 0x0006000E, /* 0:	400		400		0   */
-	0x00E79FFF, 0x000E000C, /* 1:	400		500		2   */
-	0x00D75FFF, 0x0005000A, /* 2:	400		600		3.5 */
-	0x00FFFFFF, 0x0005000A, /* 3:	600		600		0   */
-	0x00E79FFF, 0x001D0007, /* 4:	600		750		2   */
-	0x00D75FFF, 0x000C0004, /* 5:	600		900		3.5 */
-	0x00FFFFFF, 0x00040006, /* 6:	800		800		0   */
-	0x80E79FFF, 0x00030002, /* 7:	800		1000		2   */
-	0x00FFFFFF, 0x00140005, /* 8:	850		850		0   */
-	0x00FFFFFF, 0x000C0004, /* 9:	900		900		0   */
-	0x00FFFFFF, 0x001C0003, /* 10:	950		950		0   */
-	0x80FFFFFF, 0x00030002, /* 11:	1000		1000		0   */
+static const struct ddi_buf_trans bdw_ddi_translations_fdi[] = {
+	{ 0x00FFFFFF, 0x0001000E },
+	{ 0x00D75FFF, 0x0004000A },
+	{ 0x00C30FFF, 0x00070006 },
+	{ 0x00AAAFFF, 0x000C0000 },
+	{ 0x00FFFFFF, 0x0004000A },
+	{ 0x00D75FFF, 0x00090004 },
+	{ 0x00C30FFF, 0x000C0000 },
+	{ 0x00FFFFFF, 0x00070006 },
+	{ 0x00D75FFF, 0x000C0000 },
 };
 
-static const u32 bdw_ddi_translations_edp[] = {
-	0x00FFFFFF, 0x00000012,		/* eDP parameters */
-	0x00EBAFFF, 0x00020011,
-	0x00C71FFF, 0x0006000F,
-	0x00AAAFFF, 0x000E000A,
-	0x00FFFFFF, 0x00020011,
-	0x00DB6FFF, 0x0005000F,
-	0x00BEEFFF, 0x000A000C,
-	0x00FFFFFF, 0x0005000F,
-	0x00DB6FFF, 0x000A000C,
-	0x00FFFFFF, 0x00140006		/* HDMI parameters 800mV 0dB*/
+static const struct ddi_buf_trans bdw_ddi_translations_hdmi[] = {
+					/* Idx	NT mV d	T mV df	db	*/
+	{ 0x00FFFFFF, 0x0007000E },	/* 0:	400	400	0	*/
+	{ 0x00D75FFF, 0x000E000A },	/* 1:	400	600	3.5	*/
+	{ 0x00BEFFFF, 0x00140006 },	/* 2:	400	800	6	*/
+	{ 0x00FFFFFF, 0x0009000D },	/* 3:	450	450	0	*/
+	{ 0x00FFFFFF, 0x000E000A },	/* 4:	600	600	0	*/
+	{ 0x00D7FFFF, 0x00140006 },	/* 5:	600	800	2.5	*/
+	{ 0x80CB2FFF, 0x001B0002 },	/* 6:	600	1000	4.5	*/
+	{ 0x00FFFFFF, 0x00140006 },	/* 7:	800	800	0	*/
+	{ 0x80E79FFF, 0x001B0002 },	/* 8:	800	1000	2	*/
+	{ 0x80FFFFFF, 0x001B0002 },	/* 9:	1000	1000	0	*/
 };
 
-static const u32 bdw_ddi_translations_dp[] = {
-	0x00FFFFFF, 0x0007000E,		/* DP parameters */
-	0x00D75FFF, 0x000E000A,
-	0x00BEFFFF, 0x00140006,
-	0x80B2CFFF, 0x001B0002,
-	0x00FFFFFF, 0x000E000A,
-	0x00D75FFF, 0x00180004,
-	0x80CB2FFF, 0x001B0002,
-	0x00F7DFFF, 0x00180004,
-	0x80D75FFF, 0x001B0002,
-	0x00FFFFFF, 0x00140006		/* HDMI parameters 800mV 0dB*/
+static const struct ddi_buf_trans skl_ddi_translations_dp[] = {
+	{ 0x00000018, 0x000000a0 },
+	{ 0x00004014, 0x00000098 },
+	{ 0x00006012, 0x00000088 },
+	{ 0x00008010, 0x00000080 },
+	{ 0x00000018, 0x00000098 },
+	{ 0x00004014, 0x00000088 },
+	{ 0x00006012, 0x00000080 },
+	{ 0x00000018, 0x00000088 },
+	{ 0x00004014, 0x00000080 },
 };
 
-static const u32 bdw_ddi_translations_fdi[] = {
-	0x00FFFFFF, 0x0001000E,		/* FDI parameters */
-	0x00D75FFF, 0x0004000A,
-	0x00C30FFF, 0x00070006,
-	0x00AAAFFF, 0x000C0000,
-	0x00FFFFFF, 0x0004000A,
-	0x00D75FFF, 0x00090004,
-	0x00C30FFF, 0x000C0000,
-	0x00FFFFFF, 0x00070006,
-	0x00D75FFF, 0x000C0000,
-	0x00FFFFFF, 0x00140006		/* HDMI parameters 800mV 0dB*/
+static const struct ddi_buf_trans skl_ddi_translations_hdmi[] = {
+					/* Idx	NT mV   T mV    db  */
+	{ 0x00000018, 0x000000a0 },	/* 0:	400	400	0   */
+	{ 0x00004014, 0x00000098 },	/* 1:	400	600	3.5 */
+	{ 0x00006012, 0x00000088 },	/* 2:	400	800	6   */
+	{ 0x00000018, 0x0000003c },	/* 3:	450	450	0   */
+	{ 0x00000018, 0x00000098 },	/* 4:	600	600	0   */
+	{ 0x00003015, 0x00000088 },	/* 5:	600	800	2.5 */
+	{ 0x00005013, 0x00000080 },	/* 6:	600	1000	4.5 */
+	{ 0x00000018, 0x00000088 },	/* 7:	800	800	0   */
+	{ 0x00000096, 0x00000080 },	/* 8:	800	1000	2   */
+	{ 0x00000018, 0x00000080 },	/* 9:	1200	1200	0   */
 };
 
 enum port intel_ddi_get_encoder_port(struct intel_encoder *intel_encoder)
@@ -145,26 +187,43 @@ static void intel_prepare_ddi_buffers(struct drm_device *dev, enum port port)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	u32 reg;
-	int i;
+	int i, n_hdmi_entries, hdmi_800mV_0dB;
 	int hdmi_level = dev_priv->vbt.ddi_port_info[port].hdmi_level_shift;
-	const u32 *ddi_translations_fdi;
-	const u32 *ddi_translations_dp;
-	const u32 *ddi_translations_edp;
-	const u32 *ddi_translations;
-
-	if (IS_BROADWELL(dev)) {
+	const struct ddi_buf_trans *ddi_translations_fdi;
+	const struct ddi_buf_trans *ddi_translations_dp;
+	const struct ddi_buf_trans *ddi_translations_edp;
+	const struct ddi_buf_trans *ddi_translations_hdmi;
+	const struct ddi_buf_trans *ddi_translations;
+
+	if (IS_SKYLAKE(dev)) {
+		ddi_translations_fdi = NULL;
+		ddi_translations_dp = skl_ddi_translations_dp;
+		ddi_translations_edp = skl_ddi_translations_dp;
+		ddi_translations_hdmi = skl_ddi_translations_hdmi;
+		n_hdmi_entries = ARRAY_SIZE(skl_ddi_translations_hdmi);
+		hdmi_800mV_0dB = 7;
+	} else if (IS_BROADWELL(dev)) {
 		ddi_translations_fdi = bdw_ddi_translations_fdi;
 		ddi_translations_dp = bdw_ddi_translations_dp;
 		ddi_translations_edp = bdw_ddi_translations_edp;
+		ddi_translations_hdmi = bdw_ddi_translations_hdmi;
+		n_hdmi_entries = ARRAY_SIZE(bdw_ddi_translations_hdmi);
+		hdmi_800mV_0dB = 7;
 	} else if (IS_HASWELL(dev)) {
 		ddi_translations_fdi = hsw_ddi_translations_fdi;
 		ddi_translations_dp = hsw_ddi_translations_dp;
 		ddi_translations_edp = hsw_ddi_translations_dp;
+		ddi_translations_hdmi = hsw_ddi_translations_hdmi;
+		n_hdmi_entries = ARRAY_SIZE(hsw_ddi_translations_hdmi);
+		hdmi_800mV_0dB = 6;
 	} else {
 		WARN(1, "ddi translation table missing\n");
 		ddi_translations_edp = bdw_ddi_translations_dp;
 		ddi_translations_fdi = bdw_ddi_translations_fdi;
 		ddi_translations_dp = bdw_ddi_translations_dp;
+		ddi_translations_hdmi = bdw_ddi_translations_hdmi;
+		n_hdmi_entries = ARRAY_SIZE(bdw_ddi_translations_hdmi);
+		hdmi_800mV_0dB = 7;
 	}
 
 	switch (port) {
@@ -182,7 +241,10 @@ static void intel_prepare_ddi_buffers(struct drm_device *dev, enum port port)
 			ddi_translations = ddi_translations_dp;
 		break;
 	case PORT_E:
-		ddi_translations = ddi_translations_fdi;
+		if (ddi_translations_fdi)
+			ddi_translations = ddi_translations_fdi;
+		else
+			ddi_translations = ddi_translations_dp;
 		break;
 	default:
 		BUG();
@@ -190,14 +252,22 @@ static void intel_prepare_ddi_buffers(struct drm_device *dev, enum port port)
 
 	for (i = 0, reg = DDI_BUF_TRANS(port);
 	     i < ARRAY_SIZE(hsw_ddi_translations_fdi); i++) {
-		I915_WRITE(reg, ddi_translations[i]);
+		I915_WRITE(reg, ddi_translations[i].trans1);
 		reg += 4;
-	}
-	/* Entry 9 is for HDMI: */
-	for (i = 0; i < 2; i++) {
-		I915_WRITE(reg, hsw_ddi_translations_hdmi[hdmi_level * 2 + i]);
+		I915_WRITE(reg, ddi_translations[i].trans2);
 		reg += 4;
 	}
+
+	/* Choose a good default if VBT is badly populated */
+	if (hdmi_level == HDMI_LEVEL_SHIFT_UNKNOWN ||
+	    hdmi_level >= n_hdmi_entries)
+		hdmi_level = hdmi_800mV_0dB;
+
+	/* Entry 9 is for HDMI: */
+	I915_WRITE(reg, ddi_translations_hdmi[hdmi_level].trans1);
+	reg += 4;
+	I915_WRITE(reg, ddi_translations_hdmi[hdmi_level].trans2);
+	reg += 4;
 }
 
 /* Program DDI buffers translations for DP. By default, program ports A-D in DP
@@ -214,18 +284,6 @@ void intel_prepare_ddi(struct drm_device *dev)
 		intel_prepare_ddi_buffers(dev, port);
 }
 
-static const long hsw_ddi_buf_ctl_values[] = {
-	DDI_BUF_EMP_400MV_0DB_HSW,
-	DDI_BUF_EMP_400MV_3_5DB_HSW,
-	DDI_BUF_EMP_400MV_6DB_HSW,
-	DDI_BUF_EMP_400MV_9_5DB_HSW,
-	DDI_BUF_EMP_600MV_0DB_HSW,
-	DDI_BUF_EMP_600MV_3_5DB_HSW,
-	DDI_BUF_EMP_600MV_6DB_HSW,
-	DDI_BUF_EMP_800MV_0DB_HSW,
-	DDI_BUF_EMP_800MV_3_5DB_HSW
-};
-
 static void intel_wait_ddi_buf_idle(struct drm_i915_private *dev_priv,
 				    enum port port)
 {
@@ -285,7 +343,7 @@ void hsw_fdi_link_train(struct drm_crtc *crtc)
 
 	/* Start the training iterating through available voltages and emphasis,
 	 * testing each value twice. */
-	for (i = 0; i < ARRAY_SIZE(hsw_ddi_buf_ctl_values) * 2; i++) {
+	for (i = 0; i < ARRAY_SIZE(hsw_ddi_translations_fdi) * 2; i++) {
 		/* Configure DP_TP_CTL with auto-training */
 		I915_WRITE(DP_TP_CTL(PORT_E),
 					DP_TP_CTL_FDI_AUTOTRAIN |
@@ -300,7 +358,7 @@ void hsw_fdi_link_train(struct drm_crtc *crtc)
 		I915_WRITE(DDI_BUF_CTL(PORT_E),
 			   DDI_BUF_CTL_ENABLE |
 			   ((intel_crtc->config.fdi_lanes - 1) << 1) |
-			   hsw_ddi_buf_ctl_values[i / 2]);
+			   DDI_BUF_TRANS_SELECT(i / 2));
 		POSTING_READ(DDI_BUF_CTL(PORT_E));
 
 		udelay(600);
@@ -375,7 +433,7 @@ void intel_ddi_init_dp_buf_reg(struct intel_encoder *encoder)
 		enc_to_dig_port(&encoder->base);
 
 	intel_dp->DP = intel_dig_port->saved_port_bits |
-		DDI_BUF_CTL_ENABLE | DDI_BUF_EMP_400MV_0DB_HSW;
+		DDI_BUF_CTL_ENABLE | DDI_BUF_TRANS_SELECT(0);
 	intel_dp->DP |= DDI_PORT_WIDTH(intel_dp->lane_count);
 
 }
@@ -401,8 +459,29 @@ intel_ddi_get_crtc_encoder(struct drm_crtc *crtc)
 	return ret;
 }
 
+static struct intel_encoder *
+intel_ddi_get_crtc_new_encoder(struct intel_crtc *crtc)
+{
+	struct drm_device *dev = crtc->base.dev;
+	struct intel_encoder *intel_encoder, *ret = NULL;
+	int num_encoders = 0;
+
+	for_each_intel_encoder(dev, intel_encoder) {
+		if (intel_encoder->new_crtc == crtc) {
+			ret = intel_encoder;
+			num_encoders++;
+		}
+	}
+
+	WARN(num_encoders != 1, "%d encoders on crtc for pipe %c\n", num_encoders,
+	     pipe_name(crtc->pipe));
+
+	BUG_ON(ret == NULL);
+	return ret;
+}
+
 #define LC_FREQ 2700
-#define LC_FREQ_2K (LC_FREQ * 2000)
+#define LC_FREQ_2K U64_C(LC_FREQ * 2000)
 
 #define P_MIN 2
 #define P_MAX 64
@@ -414,7 +493,11 @@ intel_ddi_get_crtc_encoder(struct drm_crtc *crtc)
 #define VCO_MIN 2400
 #define VCO_MAX 4800
 
-#define ABS_DIFF(a, b) ((a > b) ? (a - b) : (b - a))
+#define abs_diff(a, b) ({			\
+	typeof(a) __a = (a);			\
+	typeof(b) __b = (b);			\
+	(void) (&__a == &__b);			\
+	__a > __b ? (__a - __b) : (__b - __a); })
 
 struct wrpll_rnp {
 	unsigned p, n2, r2;
@@ -524,9 +607,9 @@ static void wrpll_update_rnp(uint64_t freq2k, unsigned budget,
 	 */
 	a = freq2k * budget * p * r2;
 	b = freq2k * budget * best->p * best->r2;
-	diff = ABS_DIFF((freq2k * p * r2), (LC_FREQ_2K * n2));
-	diff_best = ABS_DIFF((freq2k * best->p * best->r2),
-			     (LC_FREQ_2K * best->n2));
+	diff = abs_diff(freq2k * p * r2, LC_FREQ_2K * n2);
+	diff_best = abs_diff(freq2k * best->p * best->r2,
+			     LC_FREQ_2K * best->n2);
 	c = 1000000 * diff;
 	d = 1000000 * diff_best;
 
@@ -587,8 +670,113 @@ static int intel_ddi_calc_wrpll_link(struct drm_i915_private *dev_priv,
 	return (refclk * n * 100) / (p * r);
 }
 
-void intel_ddi_clock_get(struct intel_encoder *encoder,
-			 struct intel_crtc_config *pipe_config)
+static int skl_calc_wrpll_link(struct drm_i915_private *dev_priv,
+			       uint32_t dpll)
+{
+	uint32_t cfgcr1_reg, cfgcr2_reg;
+	uint32_t cfgcr1_val, cfgcr2_val;
+	uint32_t p0, p1, p2, dco_freq;
+
+	cfgcr1_reg = GET_CFG_CR1_REG(dpll);
+	cfgcr2_reg = GET_CFG_CR2_REG(dpll);
+
+	cfgcr1_val = I915_READ(cfgcr1_reg);
+	cfgcr2_val = I915_READ(cfgcr2_reg);
+
+	p0 = cfgcr2_val & DPLL_CFGCR2_PDIV_MASK;
+	p2 = cfgcr2_val & DPLL_CFGCR2_KDIV_MASK;
+
+	if (cfgcr2_val &  DPLL_CFGCR2_QDIV_MODE(1))
+		p1 = (cfgcr2_val & DPLL_CFGCR2_QDIV_RATIO_MASK) >> 8;
+	else
+		p1 = 1;
+
+
+	switch (p0) {
+	case DPLL_CFGCR2_PDIV_1:
+		p0 = 1;
+		break;
+	case DPLL_CFGCR2_PDIV_2:
+		p0 = 2;
+		break;
+	case DPLL_CFGCR2_PDIV_3:
+		p0 = 3;
+		break;
+	case DPLL_CFGCR2_PDIV_7:
+		p0 = 7;
+		break;
+	}
+
+	switch (p2) {
+	case DPLL_CFGCR2_KDIV_5:
+		p2 = 5;
+		break;
+	case DPLL_CFGCR2_KDIV_2:
+		p2 = 2;
+		break;
+	case DPLL_CFGCR2_KDIV_3:
+		p2 = 3;
+		break;
+	case DPLL_CFGCR2_KDIV_1:
+		p2 = 1;
+		break;
+	}
+
+	dco_freq = (cfgcr1_val & DPLL_CFGCR1_DCO_INTEGER_MASK) * 24 * 1000;
+
+	dco_freq += (((cfgcr1_val & DPLL_CFGCR1_DCO_FRACTION_MASK) >> 9) * 24 *
+		1000) / 0x8000;
+
+	return dco_freq / (p0 * p1 * p2 * 5);
+}
+
+
+static void skl_ddi_clock_get(struct intel_encoder *encoder,
+				struct intel_crtc_config *pipe_config)
+{
+	struct drm_i915_private *dev_priv = encoder->base.dev->dev_private;
+	int link_clock = 0;
+	uint32_t dpll_ctl1, dpll;
+
+	dpll = pipe_config->ddi_pll_sel;
+
+	dpll_ctl1 = I915_READ(DPLL_CTRL1);
+
+	if (dpll_ctl1 & DPLL_CTRL1_HDMI_MODE(dpll)) {
+		link_clock = skl_calc_wrpll_link(dev_priv, dpll);
+	} else {
+		link_clock = dpll_ctl1 & DPLL_CRTL1_LINK_RATE_MASK(dpll);
+		link_clock >>= DPLL_CRTL1_LINK_RATE_SHIFT(dpll);
+
+		switch (link_clock) {
+		case DPLL_CRTL1_LINK_RATE_810:
+			link_clock = 81000;
+			break;
+		case DPLL_CRTL1_LINK_RATE_1350:
+			link_clock = 135000;
+			break;
+		case DPLL_CRTL1_LINK_RATE_2700:
+			link_clock = 270000;
+			break;
+		default:
+			WARN(1, "Unsupported link rate\n");
+			break;
+		}
+		link_clock *= 2;
+	}
+
+	pipe_config->port_clock = link_clock;
+
+	if (pipe_config->has_dp_encoder)
+		pipe_config->adjusted_mode.crtc_clock =
+			intel_dotclock_calculate(pipe_config->port_clock,
+						 &pipe_config->dp_m_n);
+	else
+		pipe_config->adjusted_mode.crtc_clock = pipe_config->port_clock;
+}
+
+static void hsw_ddi_clock_get(struct intel_encoder *encoder,
+			      struct intel_crtc_config *pipe_config)
 {
 	struct drm_i915_private *dev_priv = encoder->base.dev->dev_private;
 	int link_clock = 0;
@@ -643,9 +831,15 @@ void intel_ddi_clock_get(struct intel_encoder *encoder,
 		pipe_config->adjusted_mode.crtc_clock = pipe_config->port_clock;
 }
 
+void intel_ddi_clock_get(struct intel_encoder *encoder,
+			 struct intel_crtc_config *pipe_config)
+{
+	hsw_ddi_clock_get(encoder, pipe_config);
+}
+
 static void
-intel_ddi_calculate_wrpll(int clock /* in Hz */,
-			  unsigned *r2_out, unsigned *n2_out, unsigned *p_out)
+hsw_ddi_calculate_wrpll(int clock /* in Hz */,
+			unsigned *r2_out, unsigned *n2_out, unsigned *p_out)
 {
 	uint64_t freq2k;
 	unsigned p, n2, r2;
@@ -708,33 +902,23 @@ intel_ddi_calculate_wrpll(int clock /* in Hz */,
 	*r2_out = best.r2;
 }
 
-/*
- * Tries to find a PLL for the CRTC. If it finds, it increases the refcount and
- * stores it in intel_crtc->ddi_pll_sel, so other mode sets won't be able to
- * steal the selected PLL. You need to call intel_ddi_pll_enable to actually
- * enable the PLL.
- */
-bool intel_ddi_pll_select(struct intel_crtc *intel_crtc)
+static bool
+hsw_ddi_pll_select(struct intel_crtc *intel_crtc,
+		   struct intel_encoder *intel_encoder,
+		   int clock)
 {
-	struct drm_crtc *crtc = &intel_crtc->base;
-	struct intel_encoder *intel_encoder = intel_ddi_get_crtc_encoder(crtc);
-	int type = intel_encoder->type;
-	int clock = intel_crtc->config.port_clock;
-
-	intel_put_shared_dpll(intel_crtc);
-
-	if (type == INTEL_OUTPUT_HDMI) {
+	if (intel_encoder->type == INTEL_OUTPUT_HDMI) {
 		struct intel_shared_dpll *pll;
 		uint32_t val;
 		unsigned p, n2, r2;
 
-		intel_ddi_calculate_wrpll(clock * 1000, &r2, &n2, &p);
+		hsw_ddi_calculate_wrpll(clock * 1000, &r2, &n2, &p);
 
 		val = WRPLL_PLL_ENABLE | WRPLL_PLL_LCPLL |
 		      WRPLL_DIVIDER_REFERENCE(r2) | WRPLL_DIVIDER_FEEDBACK(n2) |
 		      WRPLL_DIVIDER_POST(p);
 
-		intel_crtc->config.dpll_hw_state.wrpll = val;
+		intel_crtc->new_config->dpll_hw_state.wrpll = val;
 
 		pll = intel_get_shared_dpll(intel_crtc);
 		if (pll == NULL) {
@@ -743,12 +927,255 @@ bool intel_ddi_pll_select(struct intel_crtc *intel_crtc)
 			return false;
 		}
 
-		intel_crtc->config.ddi_pll_sel = PORT_CLK_SEL_WRPLL(pll->id);
+		intel_crtc->new_config->ddi_pll_sel = PORT_CLK_SEL_WRPLL(pll->id);
+	}
+
+	return true;
+}
+
+struct skl_wrpll_params {
+	uint32_t        dco_fraction;
+	uint32_t        dco_integer;
+	uint32_t        qdiv_ratio;
+	uint32_t        qdiv_mode;
+	uint32_t        kdiv;
+	uint32_t        pdiv;
+	uint32_t        central_freq;
+};
+
+static void
+skl_ddi_calculate_wrpll(int clock /* in Hz */,
+			struct skl_wrpll_params *wrpll_params)
+{
+	uint64_t afe_clock = clock * 5; /* AFE Clock is 5x Pixel clock */
+	uint64_t dco_central_freq[3] = {8400000000ULL,
+					9000000000ULL,
+					9600000000ULL};
+	uint32_t min_dco_deviation = 400;
+	uint32_t min_dco_index = 3;
+	uint32_t P0[4] = {1, 2, 3, 7};
+	uint32_t P2[4] = {1, 2, 3, 5};
+	bool found = false;
+	uint32_t candidate_p = 0;
+	uint32_t candidate_p0[3] = {0}, candidate_p1[3] = {0};
+	uint32_t candidate_p2[3] = {0};
+	uint32_t dco_central_freq_deviation[3];
+	uint32_t i, P1, k, dco_count;
+	bool retry_with_odd = false;
+	uint64_t dco_freq;
+
+	/* Determine P0, P1 or P2 */
+	for (dco_count = 0; dco_count < 3; dco_count++) {
+		found = false;
+		candidate_p =
+			div64_u64(dco_central_freq[dco_count], afe_clock);
+		if (retry_with_odd == false)
+			candidate_p = (candidate_p % 2 == 0 ?
+				candidate_p : candidate_p + 1);
+
+		for (P1 = 1; P1 < candidate_p; P1++) {
+			for (i = 0; i < 4; i++) {
+				if (!(P0[i] != 1 || P1 == 1))
+					continue;
+
+				for (k = 0; k < 4; k++) {
+					if (P1 != 1 && P2[k] != 2)
+						continue;
+
+					if (candidate_p == P0[i] * P1 * P2[k]) {
+						/* Found possible P0, P1, P2 */
+						found = true;
+						candidate_p0[dco_count] = P0[i];
+						candidate_p1[dco_count] = P1;
+						candidate_p2[dco_count] = P2[k];
+						goto found;
+					}
+
+				}
+			}
+		}
+
+found:
+		if (found) {
+			dco_central_freq_deviation[dco_count] =
+				div64_u64(10000 *
+					  abs_diff((candidate_p * afe_clock),
+						   dco_central_freq[dco_count]),
+					  dco_central_freq[dco_count]);
+
+			if (dco_central_freq_deviation[dco_count] <
+				min_dco_deviation) {
+				min_dco_deviation =
+					dco_central_freq_deviation[dco_count];
+				min_dco_index = dco_count;
+			}
+		}
+
+		if (min_dco_index > 2 && dco_count == 2) {
+			retry_with_odd = true;
+			dco_count = 0;
+		}
+	}
+
+	if (min_dco_index > 2) {
+		WARN(1, "No valid values found for the given pixel clock\n");
+	} else {
+		 wrpll_params->central_freq = dco_central_freq[min_dco_index];
+
+		 switch (dco_central_freq[min_dco_index]) {
+		 case 9600000000ULL:
+			wrpll_params->central_freq = 0;
+			break;
+		 case 9000000000ULL:
+			wrpll_params->central_freq = 1;
+			break;
+		 case 8400000000ULL:
+			wrpll_params->central_freq = 3;
+		 }
+
+		 switch (candidate_p0[min_dco_index]) {
+		 case 1:
+			wrpll_params->pdiv = 0;
+			break;
+		 case 2:
+			wrpll_params->pdiv = 1;
+			break;
+		 case 3:
+			wrpll_params->pdiv = 2;
+			break;
+		 case 7:
+			wrpll_params->pdiv = 4;
+			break;
+		 default:
+			WARN(1, "Incorrect PDiv\n");
+		 }
+
+		 switch (candidate_p2[min_dco_index]) {
+		 case 5:
+			wrpll_params->kdiv = 0;
+			break;
+		 case 2:
+			wrpll_params->kdiv = 1;
+			break;
+		 case 3:
+			wrpll_params->kdiv = 2;
+			break;
+		 case 1:
+			wrpll_params->kdiv = 3;
+			break;
+		 default:
+			WARN(1, "Incorrect KDiv\n");
+		 }
+
+		 wrpll_params->qdiv_ratio = candidate_p1[min_dco_index];
+		 wrpll_params->qdiv_mode =
+			(wrpll_params->qdiv_ratio == 1) ? 0 : 1;
+
+		 dco_freq = candidate_p0[min_dco_index] *
+			 candidate_p1[min_dco_index] *
+			 candidate_p2[min_dco_index] * afe_clock;
+
+		/*
+		* Intermediate values are in Hz.
+		* Divide by MHz to match bsepc
+		*/
+		 wrpll_params->dco_integer = div_u64(dco_freq, (24 * MHz(1)));
+		 wrpll_params->dco_fraction =
+			 div_u64(((div_u64(dco_freq, 24) -
+				   wrpll_params->dco_integer * MHz(1)) * 0x8000), MHz(1));
+
+	}
+}
+
+
+static bool
+skl_ddi_pll_select(struct intel_crtc *intel_crtc,
+		   struct intel_encoder *intel_encoder,
+		   int clock)
+{
+	struct intel_shared_dpll *pll;
+	uint32_t ctrl1, cfgcr1, cfgcr2;
+
+	/*
+	 * See comment in intel_dpll_hw_state to understand why we always use 0
+	 * as the DPLL id in this function.
+	 */
+
+	ctrl1 = DPLL_CTRL1_OVERRIDE(0);
+
+	if (intel_encoder->type == INTEL_OUTPUT_HDMI) {
+		struct skl_wrpll_params wrpll_params = { 0, };
+
+		ctrl1 |= DPLL_CTRL1_HDMI_MODE(0);
+
+		skl_ddi_calculate_wrpll(clock * 1000, &wrpll_params);
+
+		cfgcr1 = DPLL_CFGCR1_FREQ_ENABLE |
+			 DPLL_CFGCR1_DCO_FRACTION(wrpll_params.dco_fraction) |
+			 wrpll_params.dco_integer;
+
+		cfgcr2 = DPLL_CFGCR2_QDIV_RATIO(wrpll_params.qdiv_ratio) |
+			 DPLL_CFGCR2_QDIV_MODE(wrpll_params.qdiv_mode) |
+			 DPLL_CFGCR2_KDIV(wrpll_params.kdiv) |
+			 DPLL_CFGCR2_PDIV(wrpll_params.pdiv) |
+			 wrpll_params.central_freq;
+	} else if (intel_encoder->type == INTEL_OUTPUT_DISPLAYPORT) {
+		struct drm_encoder *encoder = &intel_encoder->base;
+		struct intel_dp *intel_dp = enc_to_intel_dp(encoder);
+
+		switch (intel_dp->link_bw) {
+		case DP_LINK_BW_1_62:
+			ctrl1 |= DPLL_CRTL1_LINK_RATE(DPLL_CRTL1_LINK_RATE_810, 0);
+			break;
+		case DP_LINK_BW_2_7:
+			ctrl1 |= DPLL_CRTL1_LINK_RATE(DPLL_CRTL1_LINK_RATE_1350, 0);
+			break;
+		case DP_LINK_BW_5_4:
+			ctrl1 |= DPLL_CRTL1_LINK_RATE(DPLL_CRTL1_LINK_RATE_2700, 0);
+			break;
+		}
+
+		cfgcr1 = cfgcr2 = 0;
+	} else /* eDP */
+		return true;
+
+	intel_crtc->new_config->dpll_hw_state.ctrl1 = ctrl1;
+	intel_crtc->new_config->dpll_hw_state.cfgcr1 = cfgcr1;
+	intel_crtc->new_config->dpll_hw_state.cfgcr2 = cfgcr2;
+
+	pll = intel_get_shared_dpll(intel_crtc);
+	if (pll == NULL) {
+		DRM_DEBUG_DRIVER("failed to find PLL for pipe %c\n",
+				 pipe_name(intel_crtc->pipe));
+		return false;
 	}
 
+	/* shared DPLL id 0 is DPLL 1 */
+	intel_crtc->new_config->ddi_pll_sel = pll->id + 1;
+
 	return true;
 }
 
+/*
+ * Tries to find a *shared* PLL for the CRTC and store it in
+ * intel_crtc->ddi_pll_sel.
+ *
+ * For private DPLLs, compute_config() should do the selection for us. This
+ * function should be folded into compute_config() eventually.
+ */
+bool intel_ddi_pll_select(struct intel_crtc *intel_crtc)
+{
+	struct drm_device *dev = intel_crtc->base.dev;
+	struct intel_encoder *intel_encoder =
+		intel_ddi_get_crtc_new_encoder(intel_crtc);
+	int clock = intel_crtc->new_config->port_clock;
+
+	if (IS_SKYLAKE(dev))
+		return skl_ddi_pll_select(intel_crtc, intel_encoder, clock);
+	else
+		return hsw_ddi_pll_select(intel_crtc, intel_encoder, clock);
+}
+
 void intel_ddi_set_pipe_settings(struct drm_crtc *crtc)
 {
 	struct drm_i915_private *dev_priv = crtc->dev->dev_private;
@@ -921,7 +1348,7 @@ bool intel_ddi_connector_get_hw_state(struct intel_connector *intel_connector)
 	uint32_t tmp;
 
 	power_domain = intel_display_port_power_domain(intel_encoder);
-	if (!intel_display_power_enabled(dev_priv, power_domain))
+	if (!intel_display_power_is_enabled(dev_priv, power_domain))
 		return false;
 
 	if (!intel_encoder->get_hw_state(intel_encoder, &pipe))
@@ -967,7 +1394,7 @@ bool intel_ddi_get_hw_state(struct intel_encoder *encoder,
 	int i;
 
 	power_domain = intel_display_port_power_domain(encoder);
-	if (!intel_display_power_enabled(dev_priv, power_domain))
+	if (!intel_display_power_is_enabled(dev_priv, power_domain))
 		return false;
 
 	tmp = I915_READ(DDI_BUF_CTL(port));
@@ -1038,27 +1465,53 @@ void intel_ddi_disable_pipe_clock(struct intel_crtc *intel_crtc)
 static void intel_ddi_pre_enable(struct intel_encoder *intel_encoder)
 {
 	struct drm_encoder *encoder = &intel_encoder->base;
-	struct drm_i915_private *dev_priv = encoder->dev->dev_private;
+	struct drm_device *dev = encoder->dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_crtc *crtc = to_intel_crtc(encoder->crtc);
 	enum port port = intel_ddi_get_encoder_port(intel_encoder);
 	int type = intel_encoder->type;
 
-	if (crtc->config.has_audio) {
-		DRM_DEBUG_DRIVER("Audio on pipe %c on DDI\n",
-				 pipe_name(crtc->pipe));
-
-		/* write eld */
-		DRM_DEBUG_DRIVER("DDI audio: write eld information\n");
-		intel_write_eld(encoder, &crtc->config.adjusted_mode);
-	}
-
 	if (type == INTEL_OUTPUT_EDP) {
 		struct intel_dp *intel_dp = enc_to_intel_dp(encoder);
 		intel_edp_panel_on(intel_dp);
 	}
 
-	WARN_ON(crtc->config.ddi_pll_sel == PORT_CLK_SEL_NONE);
-	I915_WRITE(PORT_CLK_SEL(port), crtc->config.ddi_pll_sel);
+	if (IS_SKYLAKE(dev)) {
+		uint32_t dpll = crtc->config.ddi_pll_sel;
+		uint32_t val;
+
+		/*
+		 * DPLL0 is used for eDP and is the only "private" DPLL (as
+		 * opposed to shared) on SKL
+		 */
+		if (type == INTEL_OUTPUT_EDP) {
+			WARN_ON(dpll != SKL_DPLL0);
+
+			val = I915_READ(DPLL_CTRL1);
+
+			val &= ~(DPLL_CTRL1_HDMI_MODE(dpll) |
+				 DPLL_CTRL1_SSC(dpll) |
+				 DPLL_CRTL1_LINK_RATE_MASK(dpll));
+			val |= crtc->config.dpll_hw_state.ctrl1 << (dpll * 6);
+
+			I915_WRITE(DPLL_CTRL1, val);
+			POSTING_READ(DPLL_CTRL1);
+		}
+
+		/* DDI -> PLL mapping  */
+		val = I915_READ(DPLL_CTRL2);
+
+		val &= ~(DPLL_CTRL2_DDI_CLK_OFF(port) |
+			DPLL_CTRL2_DDI_CLK_SEL_MASK(port));
+		val |= (DPLL_CTRL2_DDI_CLK_SEL(dpll, port) |
+			DPLL_CTRL2_DDI_SEL_OVERRIDE(port));
+
+		I915_WRITE(DPLL_CTRL2, val);
+
+	} else {
+		WARN_ON(crtc->config.ddi_pll_sel == PORT_CLK_SEL_NONE);
+		I915_WRITE(PORT_CLK_SEL(port), crtc->config.ddi_pll_sel);
+	}
 
 	if (type == INTEL_OUTPUT_DISPLAYPORT || type == INTEL_OUTPUT_EDP) {
 		struct intel_dp *intel_dp = enc_to_intel_dp(encoder);
@@ -1068,7 +1521,7 @@ static void intel_ddi_pre_enable(struct intel_encoder *intel_encoder)
 		intel_dp_sink_dpms(intel_dp, DRM_MODE_DPMS_ON);
 		intel_dp_start_link_train(intel_dp);
 		intel_dp_complete_link_train(intel_dp);
-		if (port != PORT_A)
+		if (port != PORT_A || INTEL_INFO(dev)->gen >= 9)
 			intel_dp_stop_link_train(intel_dp);
 	} else if (type == INTEL_OUTPUT_HDMI) {
 		struct intel_hdmi *intel_hdmi = enc_to_intel_hdmi(encoder);
@@ -1082,7 +1535,8 @@ static void intel_ddi_pre_enable(struct intel_encoder *intel_encoder)
 static void intel_ddi_post_disable(struct intel_encoder *intel_encoder)
 {
 	struct drm_encoder *encoder = &intel_encoder->base;
-	struct drm_i915_private *dev_priv = encoder->dev->dev_private;
+	struct drm_device *dev = encoder->dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
 	enum port port = intel_ddi_get_encoder_port(intel_encoder);
 	int type = intel_encoder->type;
 	uint32_t val;
@@ -1110,7 +1564,11 @@ static void intel_ddi_post_disable(struct intel_encoder *intel_encoder)
 		intel_edp_panel_off(intel_dp);
 	}
 
-	I915_WRITE(PORT_CLK_SEL(port), PORT_CLK_SEL_NONE);
+	if (IS_SKYLAKE(dev))
+		I915_WRITE(DPLL_CTRL2, (I915_READ(DPLL_CTRL2) |
+					DPLL_CTRL2_DDI_CLK_OFF(port)));
+	else
+		I915_WRITE(PORT_CLK_SEL(port), PORT_CLK_SEL_NONE);
 }
 
 static void intel_enable_ddi(struct intel_encoder *intel_encoder)
@@ -1118,12 +1576,10 @@ static void intel_enable_ddi(struct intel_encoder *intel_encoder)
 	struct drm_encoder *encoder = &intel_encoder->base;
 	struct drm_crtc *crtc = encoder->crtc;
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
-	int pipe = intel_crtc->pipe;
 	struct drm_device *dev = encoder->dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	enum port port = intel_ddi_get_encoder_port(intel_encoder);
 	int type = intel_encoder->type;
-	uint32_t tmp;
 
 	if (type == INTEL_OUTPUT_HDMI) {
 		struct intel_digital_port *intel_dig_port =
@@ -1139,18 +1595,16 @@ static void intel_enable_ddi(struct intel_encoder *intel_encoder)
 	} else if (type == INTEL_OUTPUT_EDP) {
 		struct intel_dp *intel_dp = enc_to_intel_dp(encoder);
 
-		if (port == PORT_A)
+		if (port == PORT_A && INTEL_INFO(dev)->gen < 9)
 			intel_dp_stop_link_train(intel_dp);
 
 		intel_edp_backlight_on(intel_dp);
-		intel_edp_psr_enable(intel_dp);
+		intel_psr_enable(intel_dp);
 	}
 
 	if (intel_crtc->config.has_audio) {
 		intel_display_power_get(dev_priv, POWER_DOMAIN_AUDIO);
-		tmp = I915_READ(HSW_AUD_PIN_ELD_CP_VLD);
-		tmp |= ((AUDIO_OUTPUT_ENABLE_A | AUDIO_ELD_VALID_A) << (pipe * 4));
-		I915_WRITE(HSW_AUD_PIN_ELD_CP_VLD, tmp);
+		intel_audio_codec_enable(intel_encoder);
 	}
 }
 
@@ -1159,61 +1613,126 @@ static void intel_disable_ddi(struct intel_encoder *intel_encoder)
 	struct drm_encoder *encoder = &intel_encoder->base;
 	struct drm_crtc *crtc = encoder->crtc;
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
-	int pipe = intel_crtc->pipe;
 	int type = intel_encoder->type;
 	struct drm_device *dev = encoder->dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	uint32_t tmp;
 
-	/* We can't touch HSW_AUD_PIN_ELD_CP_VLD uncionditionally because this
-	 * register is part of the power well on Haswell. */
 	if (intel_crtc->config.has_audio) {
-		tmp = I915_READ(HSW_AUD_PIN_ELD_CP_VLD);
-		tmp &= ~((AUDIO_OUTPUT_ENABLE_A | AUDIO_ELD_VALID_A) <<
-			 (pipe * 4));
-		I915_WRITE(HSW_AUD_PIN_ELD_CP_VLD, tmp);
+		intel_audio_codec_disable(intel_encoder);
 		intel_display_power_put(dev_priv, POWER_DOMAIN_AUDIO);
 	}
 
 	if (type == INTEL_OUTPUT_EDP) {
 		struct intel_dp *intel_dp = enc_to_intel_dp(encoder);
 
-		intel_edp_psr_disable(intel_dp);
+		intel_psr_disable(intel_dp);
 		intel_edp_backlight_off(intel_dp);
 	}
 }
 
-int intel_ddi_get_cdclk_freq(struct drm_i915_private *dev_priv)
+static int skl_get_cdclk_freq(struct drm_i915_private *dev_priv)
+{
+	uint32_t lcpll1 = I915_READ(LCPLL1_CTL);
+	uint32_t cdctl = I915_READ(CDCLK_CTL);
+	uint32_t linkrate;
+
+	if (!(lcpll1 & LCPLL_PLL_ENABLE)) {
+		WARN(1, "LCPLL1 not enabled\n");
+		return 24000; /* 24MHz is the cd freq with NSSC ref */
+	}
+
+	if ((cdctl & CDCLK_FREQ_SEL_MASK) == CDCLK_FREQ_540)
+		return 540000;
+
+	linkrate = (I915_READ(DPLL_CTRL1) &
+		    DPLL_CRTL1_LINK_RATE_MASK(SKL_DPLL0)) >> 1;
+
+	if (linkrate == DPLL_CRTL1_LINK_RATE_2160 ||
+	    linkrate == DPLL_CRTL1_LINK_RATE_1080) {
+		/* vco 8640 */
+		switch (cdctl & CDCLK_FREQ_SEL_MASK) {
+		case CDCLK_FREQ_450_432:
+			return 432000;
+		case CDCLK_FREQ_337_308:
+			return 308570;
+		case CDCLK_FREQ_675_617:
+			return 617140;
+		default:
+			WARN(1, "Unknown cd freq selection\n");
+		}
+	} else {
+		/* vco 8100 */
+		switch (cdctl & CDCLK_FREQ_SEL_MASK) {
+		case CDCLK_FREQ_450_432:
+			return 450000;
+		case CDCLK_FREQ_337_308:
+			return 337500;
+		case CDCLK_FREQ_675_617:
+			return 675000;
+		default:
+			WARN(1, "Unknown cd freq selection\n");
+		}
+	}
+
+	/* error case, do as if DPLL0 isn't enabled */
+	return 24000;
+}
+
+static int bdw_get_cdclk_freq(struct drm_i915_private *dev_priv)
+{
+	uint32_t lcpll = I915_READ(LCPLL_CTL);
+	uint32_t freq = lcpll & LCPLL_CLK_FREQ_MASK;
+
+	if (lcpll & LCPLL_CD_SOURCE_FCLK)
+		return 800000;
+	else if (I915_READ(FUSE_STRAP) & HSW_CDCLK_LIMIT)
+		return 450000;
+	else if (freq == LCPLL_CLK_FREQ_450)
+		return 450000;
+	else if (freq == LCPLL_CLK_FREQ_54O_BDW)
+		return 540000;
+	else if (freq == LCPLL_CLK_FREQ_337_5_BDW)
+		return 337500;
+	else
+		return 675000;
+}
+
+static int hsw_get_cdclk_freq(struct drm_i915_private *dev_priv)
 {
 	struct drm_device *dev = dev_priv->dev;
 	uint32_t lcpll = I915_READ(LCPLL_CTL);
 	uint32_t freq = lcpll & LCPLL_CLK_FREQ_MASK;
 
-	if (lcpll & LCPLL_CD_SOURCE_FCLK) {
+	if (lcpll & LCPLL_CD_SOURCE_FCLK)
 		return 800000;
-	} else if (I915_READ(FUSE_STRAP) & HSW_CDCLK_LIMIT) {
+	else if (I915_READ(FUSE_STRAP) & HSW_CDCLK_LIMIT)
 		return 450000;
-	} else if (freq == LCPLL_CLK_FREQ_450) {
+	else if (freq == LCPLL_CLK_FREQ_450)
 		return 450000;
-	} else if (IS_HASWELL(dev)) {
-		if (IS_ULT(dev))
-			return 337500;
-		else
-			return 540000;
-	} else {
-		if (freq == LCPLL_CLK_FREQ_54O_BDW)
-			return 540000;
-		else if (freq == LCPLL_CLK_FREQ_337_5_BDW)
-			return 337500;
-		else
-			return 675000;
-	}
+	else if (IS_HSW_ULT(dev))
+		return 337500;
+	else
+		return 540000;
+}
+
+int intel_ddi_get_cdclk_freq(struct drm_i915_private *dev_priv)
+{
+	struct drm_device *dev = dev_priv->dev;
+
+	if (IS_SKYLAKE(dev))
+		return skl_get_cdclk_freq(dev_priv);
+
+	if (IS_BROADWELL(dev))
+		return bdw_get_cdclk_freq(dev_priv);
+
+	/* Haswell */
+	return hsw_get_cdclk_freq(dev_priv);
 }
 
 static void hsw_ddi_pll_enable(struct drm_i915_private *dev_priv,
 			       struct intel_shared_dpll *pll)
 {
-	I915_WRITE(WRPLL_CTL(pll->id), pll->hw_state.wrpll);
+	I915_WRITE(WRPLL_CTL(pll->id), pll->config.hw_state.wrpll);
 	POSTING_READ(WRPLL_CTL(pll->id));
 	udelay(20);
 }
@@ -1234,7 +1753,7 @@ static bool hsw_ddi_pll_get_hw_state(struct drm_i915_private *dev_priv,
 {
 	uint32_t val;
 
-	if (!intel_display_power_enabled(dev_priv, POWER_DOMAIN_PLLS))
+	if (!intel_display_power_is_enabled(dev_priv, POWER_DOMAIN_PLLS))
 		return false;
 
 	val = I915_READ(WRPLL_CTL(pll->id));
@@ -1248,10 +1767,8 @@ static const char * const hsw_ddi_pll_names[] = {
 	"WRPLL 2",
 };
 
-void intel_ddi_pll_init(struct drm_device *dev)
+static void hsw_shared_dplls_init(struct drm_i915_private *dev_priv)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	uint32_t val = I915_READ(LCPLL_CTL);
 	int i;
 
 	dev_priv->num_shared_dpll = 2;
@@ -1264,20 +1781,158 @@ void intel_ddi_pll_init(struct drm_device *dev)
 		dev_priv->shared_dplls[i].get_hw_state =
 			hsw_ddi_pll_get_hw_state;
 	}
+}
 
-	/* The LCPLL register should be turned on by the BIOS. For now let's
-	 * just check its state and print errors in case something is wrong.
-	 * Don't even try to turn it on.
-	 */
+static const char * const skl_ddi_pll_names[] = {
+	"DPLL 1",
+	"DPLL 2",
+	"DPLL 3",
+};
+
+struct skl_dpll_regs {
+	u32 ctl, cfgcr1, cfgcr2;
+};
+
+/* this array is indexed by the *shared* pll id */
+static const struct skl_dpll_regs skl_dpll_regs[3] = {
+	{
+		/* DPLL 1 */
+		.ctl = LCPLL2_CTL,
+		.cfgcr1 = DPLL1_CFGCR1,
+		.cfgcr2 = DPLL1_CFGCR2,
+	},
+	{
+		/* DPLL 2 */
+		.ctl = WRPLL_CTL1,
+		.cfgcr1 = DPLL2_CFGCR1,
+		.cfgcr2 = DPLL2_CFGCR2,
+	},
+	{
+		/* DPLL 3 */
+		.ctl = WRPLL_CTL2,
+		.cfgcr1 = DPLL3_CFGCR1,
+		.cfgcr2 = DPLL3_CFGCR2,
+	},
+};
+
+static void skl_ddi_pll_enable(struct drm_i915_private *dev_priv,
+			       struct intel_shared_dpll *pll)
+{
+	uint32_t val;
+	unsigned int dpll;
+	const struct skl_dpll_regs *regs = skl_dpll_regs;
+
+	/* DPLL0 is not part of the shared DPLLs, so pll->id is 0 for DPLL1 */
+	dpll = pll->id + 1;
+
+	val = I915_READ(DPLL_CTRL1);
+
+	val &= ~(DPLL_CTRL1_HDMI_MODE(dpll) | DPLL_CTRL1_SSC(dpll) |
+		 DPLL_CRTL1_LINK_RATE_MASK(dpll));
+	val |= pll->config.hw_state.ctrl1 << (dpll * 6);
+
+	I915_WRITE(DPLL_CTRL1, val);
+	POSTING_READ(DPLL_CTRL1);
+
+	I915_WRITE(regs[pll->id].cfgcr1, pll->config.hw_state.cfgcr1);
+	I915_WRITE(regs[pll->id].cfgcr2, pll->config.hw_state.cfgcr2);
+	POSTING_READ(regs[pll->id].cfgcr1);
+	POSTING_READ(regs[pll->id].cfgcr2);
+
+	/* the enable bit is always bit 31 */
+	I915_WRITE(regs[pll->id].ctl,
+		   I915_READ(regs[pll->id].ctl) | LCPLL_PLL_ENABLE);
+
+	if (wait_for(I915_READ(DPLL_STATUS) & DPLL_LOCK(dpll), 5))
+		DRM_ERROR("DPLL %d not locked\n", dpll);
+}
+
+static void skl_ddi_pll_disable(struct drm_i915_private *dev_priv,
+				struct intel_shared_dpll *pll)
+{
+	const struct skl_dpll_regs *regs = skl_dpll_regs;
+
+	/* the enable bit is always bit 31 */
+	I915_WRITE(regs[pll->id].ctl,
+		   I915_READ(regs[pll->id].ctl) & ~LCPLL_PLL_ENABLE);
+	POSTING_READ(regs[pll->id].ctl);
+}
+
+static bool skl_ddi_pll_get_hw_state(struct drm_i915_private *dev_priv,
+				     struct intel_shared_dpll *pll,
+				     struct intel_dpll_hw_state *hw_state)
+{
+	uint32_t val;
+	unsigned int dpll;
+	const struct skl_dpll_regs *regs = skl_dpll_regs;
+
+	if (!intel_display_power_is_enabled(dev_priv, POWER_DOMAIN_PLLS))
+		return false;
+
+	/* DPLL0 is not part of the shared DPLLs, so pll->id is 0 for DPLL1 */
+	dpll = pll->id + 1;
+
+	val = I915_READ(regs[pll->id].ctl);
+	if (!(val & LCPLL_PLL_ENABLE))
+		return false;
+
+	val = I915_READ(DPLL_CTRL1);
+	hw_state->ctrl1 = (val >> (dpll * 6)) & 0x3f;
+
+	/* avoid reading back stale values if HDMI mode is not enabled */
+	if (val & DPLL_CTRL1_HDMI_MODE(dpll)) {
+		hw_state->cfgcr1 = I915_READ(regs[pll->id].cfgcr1);
+		hw_state->cfgcr2 = I915_READ(regs[pll->id].cfgcr2);
+	}
+
+	return true;
+}
+
+static void skl_shared_dplls_init(struct drm_i915_private *dev_priv)
+{
+	int i;
+
+	dev_priv->num_shared_dpll = 3;
+
+	for (i = 0; i < dev_priv->num_shared_dpll; i++) {
+		dev_priv->shared_dplls[i].id = i;
+		dev_priv->shared_dplls[i].name = skl_ddi_pll_names[i];
+		dev_priv->shared_dplls[i].disable = skl_ddi_pll_disable;
+		dev_priv->shared_dplls[i].enable = skl_ddi_pll_enable;
+		dev_priv->shared_dplls[i].get_hw_state =
+			skl_ddi_pll_get_hw_state;
+	}
+}
+
+void intel_ddi_pll_init(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	uint32_t val = I915_READ(LCPLL_CTL);
+
+	if (IS_SKYLAKE(dev))
+		skl_shared_dplls_init(dev_priv);
+	else
+		hsw_shared_dplls_init(dev_priv);
 
 	DRM_DEBUG_KMS("CDCLK running at %dKHz\n",
 		      intel_ddi_get_cdclk_freq(dev_priv));
 
-	if (val & LCPLL_CD_SOURCE_FCLK)
-		DRM_ERROR("CDCLK source is not LCPLL\n");
+	if (IS_SKYLAKE(dev)) {
+		if (!(I915_READ(LCPLL1_CTL) & LCPLL_PLL_ENABLE))
+			DRM_ERROR("LCPLL1 is disabled\n");
+	} else {
+		/*
+		 * The LCPLL register should be turned on by the BIOS. For now
+		 * let's just check its state and print errors in case
+		 * something is wrong.  Don't even try to turn it on.
+		 */
 
-	if (val & LCPLL_PLL_DISABLE)
-		DRM_ERROR("LCPLL is disabled\n");
+		if (val & LCPLL_CD_SOURCE_FCLK)
+			DRM_ERROR("CDCLK source is not LCPLL\n");
+
+		if (val & LCPLL_PLL_DISABLE)
+			DRM_ERROR("LCPLL is disabled\n");
+	}
 }
 
 void intel_ddi_prepare_link_retrain(struct drm_encoder *encoder)
@@ -1372,7 +2027,9 @@ void intel_ddi_get_config(struct intel_encoder *encoder,
 	struct drm_i915_private *dev_priv = encoder->base.dev->dev_private;
 	struct intel_crtc *intel_crtc = to_intel_crtc(encoder->base.crtc);
 	enum transcoder cpu_transcoder = intel_crtc->config.cpu_transcoder;
+	struct intel_hdmi *intel_hdmi;
 	u32 temp, flags = 0;
+	struct drm_device *dev = dev_priv->dev;
 
 	temp = I915_READ(TRANS_DDI_FUNC_CTL(cpu_transcoder));
 	if (temp & TRANS_DDI_PHSYNC)
@@ -1406,6 +2063,11 @@ void intel_ddi_get_config(struct intel_encoder *encoder,
 	switch (temp & TRANS_DDI_MODE_SELECT_MASK) {
 	case TRANS_DDI_MODE_SELECT_HDMI:
 		pipe_config->has_hdmi_sink = true;
+		intel_hdmi = enc_to_intel_hdmi(&encoder->base);
+
+		if (intel_hdmi->infoframe_enabled(&encoder->base))
+			pipe_config->has_infoframe = true;
+		break;
 	case TRANS_DDI_MODE_SELECT_DVI:
 	case TRANS_DDI_MODE_SELECT_FDI:
 		break;
@@ -1418,9 +2080,9 @@ void intel_ddi_get_config(struct intel_encoder *encoder,
 		break;
 	}
 
-	if (intel_display_power_enabled(dev_priv, POWER_DOMAIN_AUDIO)) {
+	if (intel_display_power_is_enabled(dev_priv, POWER_DOMAIN_AUDIO)) {
 		temp = I915_READ(HSW_AUD_PIN_ELD_CP_VLD);
-		if (temp & (AUDIO_OUTPUT_ENABLE_A << (intel_crtc->pipe * 4)))
+		if (temp & AUDIO_OUTPUT_ENABLE(intel_crtc->pipe))
 			pipe_config->has_audio = true;
 	}
 
@@ -1444,7 +2106,10 @@ void intel_ddi_get_config(struct intel_encoder *encoder,
 		dev_priv->vbt.edp_bpp = pipe_config->pipe_bpp;
 	}
 
-	intel_ddi_clock_get(encoder, pipe_config);
+	if (INTEL_INFO(dev)->gen <= 8)
+		hsw_ddi_clock_get(encoder, pipe_config);
+	else
+		skl_ddi_clock_get(encoder, pipe_config);
 }
 
 static void intel_ddi_destroy(struct drm_encoder *encoder)
diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
index d8324c69fa86..e2af1383b179 100644
--- a/drivers/gpu/drm/i915/intel_display.c
+++ b/drivers/gpu/drm/i915/intel_display.c
@@ -73,11 +73,6 @@ static const uint32_t intel_cursor_formats[] = {
 	DRM_FORMAT_ARGB8888,
 };
 
-#define DIV_ROUND_CLOSEST_ULL(ll, d)	\
-({ unsigned long long _tmp = (ll)+(d)/2; do_div(_tmp, d); _tmp; })
-
-static void intel_increase_pllclock(struct drm_device *dev,
-				    enum pipe pipe);
 static void intel_crtc_update_cursor(struct drm_crtc *crtc, bool on);
 
 static void i9xx_crtc_clock_get(struct intel_crtc *crtc,
@@ -91,15 +86,18 @@ static int intel_framebuffer_init(struct drm_device *dev,
 				  struct intel_framebuffer *ifb,
 				  struct drm_mode_fb_cmd2 *mode_cmd,
 				  struct drm_i915_gem_object *obj);
-static void intel_dp_set_m_n(struct intel_crtc *crtc);
 static void i9xx_set_pipeconf(struct intel_crtc *intel_crtc);
 static void intel_set_pipe_timings(struct intel_crtc *intel_crtc);
 static void intel_cpu_transcoder_set_m_n(struct intel_crtc *crtc,
-					 struct intel_link_m_n *m_n);
+					 struct intel_link_m_n *m_n,
+					 struct intel_link_m_n *m2_n2);
 static void ironlake_set_pipeconf(struct drm_crtc *crtc);
 static void haswell_set_pipeconf(struct drm_crtc *crtc);
 static void intel_set_pipe_csc(struct drm_crtc *crtc);
-static void vlv_prepare_pll(struct intel_crtc *crtc);
+static void vlv_prepare_pll(struct intel_crtc *crtc,
+			    const struct intel_crtc_config *pipe_config);
+static void chv_prepare_pll(struct intel_crtc *crtc,
+			    const struct intel_crtc_config *pipe_config);
 
 static struct intel_encoder *intel_find_encoder(struct intel_connector *connector, int pipe)
 {
@@ -410,25 +408,43 @@ static void vlv_clock(int refclk, intel_clock_t *clock)
 /**
  * Returns whether any output on the specified pipe is of the specified type
  */
-static bool intel_pipe_has_type(struct drm_crtc *crtc, int type)
+bool intel_pipe_has_type(struct intel_crtc *crtc, enum intel_output_type type)
 {
-	struct drm_device *dev = crtc->dev;
+	struct drm_device *dev = crtc->base.dev;
 	struct intel_encoder *encoder;
 
-	for_each_encoder_on_crtc(dev, crtc, encoder)
+	for_each_encoder_on_crtc(dev, &crtc->base, encoder)
 		if (encoder->type == type)
 			return true;
 
 	return false;
 }
 
-static const intel_limit_t *intel_ironlake_limit(struct drm_crtc *crtc,
+/**
+ * Returns whether any output on the specified pipe will have the specified
+ * type after a staged modeset is complete, i.e., the same as
+ * intel_pipe_has_type() but looking at encoder->new_crtc instead of
+ * encoder->crtc.
+ */
+static bool intel_pipe_will_have_type(struct intel_crtc *crtc, int type)
+{
+	struct drm_device *dev = crtc->base.dev;
+	struct intel_encoder *encoder;
+
+	for_each_intel_encoder(dev, encoder)
+		if (encoder->new_crtc == crtc && encoder->type == type)
+			return true;
+
+	return false;
+}
+
+static const intel_limit_t *intel_ironlake_limit(struct intel_crtc *crtc,
 						int refclk)
 {
-	struct drm_device *dev = crtc->dev;
+	struct drm_device *dev = crtc->base.dev;
 	const intel_limit_t *limit;
 
-	if (intel_pipe_has_type(crtc, INTEL_OUTPUT_LVDS)) {
+	if (intel_pipe_will_have_type(crtc, INTEL_OUTPUT_LVDS)) {
 		if (intel_is_dual_link_lvds(dev)) {
 			if (refclk == 100000)
 				limit = &intel_limits_ironlake_dual_lvds_100m;
@@ -446,20 +462,20 @@ static const intel_limit_t *intel_ironlake_limit(struct drm_crtc *crtc,
 	return limit;
 }
 
-static const intel_limit_t *intel_g4x_limit(struct drm_crtc *crtc)
+static const intel_limit_t *intel_g4x_limit(struct intel_crtc *crtc)
 {
-	struct drm_device *dev = crtc->dev;
+	struct drm_device *dev = crtc->base.dev;
 	const intel_limit_t *limit;
 
-	if (intel_pipe_has_type(crtc, INTEL_OUTPUT_LVDS)) {
+	if (intel_pipe_will_have_type(crtc, INTEL_OUTPUT_LVDS)) {
 		if (intel_is_dual_link_lvds(dev))
 			limit = &intel_limits_g4x_dual_channel_lvds;
 		else
 			limit = &intel_limits_g4x_single_channel_lvds;
-	} else if (intel_pipe_has_type(crtc, INTEL_OUTPUT_HDMI) ||
-		   intel_pipe_has_type(crtc, INTEL_OUTPUT_ANALOG)) {
+	} else if (intel_pipe_will_have_type(crtc, INTEL_OUTPUT_HDMI) ||
+		   intel_pipe_will_have_type(crtc, INTEL_OUTPUT_ANALOG)) {
 		limit = &intel_limits_g4x_hdmi;
-	} else if (intel_pipe_has_type(crtc, INTEL_OUTPUT_SDVO)) {
+	} else if (intel_pipe_will_have_type(crtc, INTEL_OUTPUT_SDVO)) {
 		limit = &intel_limits_g4x_sdvo;
 	} else /* The option is for other outputs */
 		limit = &intel_limits_i9xx_sdvo;
@@ -467,9 +483,9 @@ static const intel_limit_t *intel_g4x_limit(struct drm_crtc *crtc)
 	return limit;
 }
 
-static const intel_limit_t *intel_limit(struct drm_crtc *crtc, int refclk)
+static const intel_limit_t *intel_limit(struct intel_crtc *crtc, int refclk)
 {
-	struct drm_device *dev = crtc->dev;
+	struct drm_device *dev = crtc->base.dev;
 	const intel_limit_t *limit;
 
 	if (HAS_PCH_SPLIT(dev))
@@ -477,7 +493,7 @@ static const intel_limit_t *intel_limit(struct drm_crtc *crtc, int refclk)
 	else if (IS_G4X(dev)) {
 		limit = intel_g4x_limit(crtc);
 	} else if (IS_PINEVIEW(dev)) {
-		if (intel_pipe_has_type(crtc, INTEL_OUTPUT_LVDS))
+		if (intel_pipe_will_have_type(crtc, INTEL_OUTPUT_LVDS))
 			limit = &intel_limits_pineview_lvds;
 		else
 			limit = &intel_limits_pineview_sdvo;
@@ -486,14 +502,14 @@ static const intel_limit_t *intel_limit(struct drm_crtc *crtc, int refclk)
 	} else if (IS_VALLEYVIEW(dev)) {
 		limit = &intel_limits_vlv;
 	} else if (!IS_GEN2(dev)) {
-		if (intel_pipe_has_type(crtc, INTEL_OUTPUT_LVDS))
+		if (intel_pipe_will_have_type(crtc, INTEL_OUTPUT_LVDS))
 			limit = &intel_limits_i9xx_lvds;
 		else
 			limit = &intel_limits_i9xx_sdvo;
 	} else {
-		if (intel_pipe_has_type(crtc, INTEL_OUTPUT_LVDS))
+		if (intel_pipe_will_have_type(crtc, INTEL_OUTPUT_LVDS))
 			limit = &intel_limits_i8xx_lvds;
-		else if (intel_pipe_has_type(crtc, INTEL_OUTPUT_DVO))
+		else if (intel_pipe_will_have_type(crtc, INTEL_OUTPUT_DVO))
 			limit = &intel_limits_i8xx_dvo;
 		else
 			limit = &intel_limits_i8xx_dac;
@@ -580,15 +596,15 @@ static bool intel_PLL_is_valid(struct drm_device *dev,
 }
 
 static bool
-i9xx_find_best_dpll(const intel_limit_t *limit, struct drm_crtc *crtc,
+i9xx_find_best_dpll(const intel_limit_t *limit, struct intel_crtc *crtc,
 		    int target, int refclk, intel_clock_t *match_clock,
 		    intel_clock_t *best_clock)
 {
-	struct drm_device *dev = crtc->dev;
+	struct drm_device *dev = crtc->base.dev;
 	intel_clock_t clock;
 	int err = target;
 
-	if (intel_pipe_has_type(crtc, INTEL_OUTPUT_LVDS)) {
+	if (intel_pipe_will_have_type(crtc, INTEL_OUTPUT_LVDS)) {
 		/*
 		 * For LVDS just rely on its current settings for dual-channel.
 		 * We haven't figured out how to reliably set up different
@@ -641,15 +657,15 @@ i9xx_find_best_dpll(const intel_limit_t *limit, struct drm_crtc *crtc,
 }
 
 static bool
-pnv_find_best_dpll(const intel_limit_t *limit, struct drm_crtc *crtc,
+pnv_find_best_dpll(const intel_limit_t *limit, struct intel_crtc *crtc,
 		   int target, int refclk, intel_clock_t *match_clock,
 		   intel_clock_t *best_clock)
 {
-	struct drm_device *dev = crtc->dev;
+	struct drm_device *dev = crtc->base.dev;
 	intel_clock_t clock;
 	int err = target;
 
-	if (intel_pipe_has_type(crtc, INTEL_OUTPUT_LVDS)) {
+	if (intel_pipe_will_have_type(crtc, INTEL_OUTPUT_LVDS)) {
 		/*
 		 * For LVDS just rely on its current settings for dual-channel.
 		 * We haven't figured out how to reliably set up different
@@ -700,11 +716,11 @@ pnv_find_best_dpll(const intel_limit_t *limit, struct drm_crtc *crtc,
 }
 
 static bool
-g4x_find_best_dpll(const intel_limit_t *limit, struct drm_crtc *crtc,
+g4x_find_best_dpll(const intel_limit_t *limit, struct intel_crtc *crtc,
 		   int target, int refclk, intel_clock_t *match_clock,
 		   intel_clock_t *best_clock)
 {
-	struct drm_device *dev = crtc->dev;
+	struct drm_device *dev = crtc->base.dev;
 	intel_clock_t clock;
 	int max_n;
 	bool found;
@@ -712,7 +728,7 @@ g4x_find_best_dpll(const intel_limit_t *limit, struct drm_crtc *crtc,
 	int err_most = (target >> 8) + (target >> 9);
 	found = false;
 
-	if (intel_pipe_has_type(crtc, INTEL_OUTPUT_LVDS)) {
+	if (intel_pipe_will_have_type(crtc, INTEL_OUTPUT_LVDS)) {
 		if (intel_is_dual_link_lvds(dev))
 			clock.p2 = limit->p2.p2_fast;
 		else
@@ -757,11 +773,11 @@ g4x_find_best_dpll(const intel_limit_t *limit, struct drm_crtc *crtc,
 }
 
 static bool
-vlv_find_best_dpll(const intel_limit_t *limit, struct drm_crtc *crtc,
+vlv_find_best_dpll(const intel_limit_t *limit, struct intel_crtc *crtc,
 		   int target, int refclk, intel_clock_t *match_clock,
 		   intel_clock_t *best_clock)
 {
-	struct drm_device *dev = crtc->dev;
+	struct drm_device *dev = crtc->base.dev;
 	intel_clock_t clock;
 	unsigned int bestppm = 1000000;
 	/* min update 19.2 MHz */
@@ -814,11 +830,11 @@ vlv_find_best_dpll(const intel_limit_t *limit, struct drm_crtc *crtc,
 }
 
 static bool
-chv_find_best_dpll(const intel_limit_t *limit, struct drm_crtc *crtc,
+chv_find_best_dpll(const intel_limit_t *limit, struct intel_crtc *crtc,
 		   int target, int refclk, intel_clock_t *match_clock,
 		   intel_clock_t *best_clock)
 {
-	struct drm_device *dev = crtc->dev;
+	struct drm_device *dev = crtc->base.dev;
 	intel_clock_t clock;
 	uint64_t m2;
 	int found = false;
@@ -891,58 +907,6 @@ enum transcoder intel_pipe_to_cpu_transcoder(struct drm_i915_private *dev_priv,
 	return intel_crtc->config.cpu_transcoder;
 }
 
-static void g4x_wait_for_vblank(struct drm_device *dev, int pipe)
-{
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	u32 frame, frame_reg = PIPE_FRMCOUNT_GM45(pipe);
-
-	frame = I915_READ(frame_reg);
-
-	if (wait_for(I915_READ_NOTRACE(frame_reg) != frame, 50))
-		WARN(1, "vblank wait timed out\n");
-}
-
-/**
- * intel_wait_for_vblank - wait for vblank on a given pipe
- * @dev: drm device
- * @pipe: pipe to wait for
- *
- * Wait for vblank to occur on a given pipe.  Needed for various bits of
- * mode setting code.
- */
-void intel_wait_for_vblank(struct drm_device *dev, int pipe)
-{
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	int pipestat_reg = PIPESTAT(pipe);
-
-	if (IS_G4X(dev) || INTEL_INFO(dev)->gen >= 5) {
-		g4x_wait_for_vblank(dev, pipe);
-		return;
-	}
-
-	/* Clear existing vblank status. Note this will clear any other
-	 * sticky status fields as well.
-	 *
-	 * This races with i915_driver_irq_handler() with the result
-	 * that either function could miss a vblank event.  Here it is not
-	 * fatal, as we will either wait upon the next vblank interrupt or
-	 * timeout.  Generally speaking intel_wait_for_vblank() is only
-	 * called during modeset at which time the GPU should be idle and
-	 * should *not* be performing page flips and thus not waiting on
-	 * vblanks...
-	 * Currently, the result of us stealing a vblank from the irq
-	 * handler is that a single frame will be skipped during swapbuffers.
-	 */
-	I915_WRITE(pipestat_reg,
-		   I915_READ(pipestat_reg) | PIPE_VBLANK_INTERRUPT_STATUS);
-
-	/* Wait for vblank interrupt bit to set */
-	if (wait_for(I915_READ(pipestat_reg) &
-		     PIPE_VBLANK_INTERRUPT_STATUS,
-		     50))
-		DRM_DEBUG_KMS("vblank wait timed out\n");
-}
-
 static bool pipe_dsl_stopped(struct drm_device *dev, enum pipe pipe)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
@@ -964,8 +928,7 @@ static bool pipe_dsl_stopped(struct drm_device *dev, enum pipe pipe)
 
 /*
  * intel_wait_for_pipe_off - wait for pipe to turn off
- * @dev: drm device
- * @pipe: pipe to wait for
+ * @crtc: crtc whose pipe to wait for
  *
  * After disabling a pipe, we can't wait for vblank in the usual way,
  * spinning on the vblank interrupt status bit, since we won't actually
@@ -979,11 +942,12 @@ static bool pipe_dsl_stopped(struct drm_device *dev, enum pipe pipe)
  *   ends up stopping at the start of the next frame).
  *
  */
-void intel_wait_for_pipe_off(struct drm_device *dev, int pipe)
+static void intel_wait_for_pipe_off(struct intel_crtc *crtc)
 {
+	struct drm_device *dev = crtc->base.dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	enum transcoder cpu_transcoder = intel_pipe_to_cpu_transcoder(dev_priv,
-								      pipe);
+	enum transcoder cpu_transcoder = crtc->config.cpu_transcoder;
+	enum pipe pipe = crtc->pipe;
 
 	if (INTEL_INFO(dev)->gen >= 4) {
 		int reg = PIPECONF(cpu_transcoder);
@@ -1189,30 +1153,43 @@ void assert_fdi_rx_pll(struct drm_i915_private *dev_priv,
 	     state_string(state), state_string(cur_state));
 }
 
-static void assert_panel_unlocked(struct drm_i915_private *dev_priv,
-				  enum pipe pipe)
+void assert_panel_unlocked(struct drm_i915_private *dev_priv,
+			   enum pipe pipe)
 {
-	int pp_reg, lvds_reg;
+	struct drm_device *dev = dev_priv->dev;
+	int pp_reg;
 	u32 val;
 	enum pipe panel_pipe = PIPE_A;
 	bool locked = true;
 
-	if (HAS_PCH_SPLIT(dev_priv->dev)) {
+	if (WARN_ON(HAS_DDI(dev)))
+		return;
+
+	if (HAS_PCH_SPLIT(dev)) {
+		u32 port_sel;
+
 		pp_reg = PCH_PP_CONTROL;
-		lvds_reg = PCH_LVDS;
+		port_sel = I915_READ(PCH_PP_ON_DELAYS) & PANEL_PORT_SELECT_MASK;
+
+		if (port_sel == PANEL_PORT_SELECT_LVDS &&
+		    I915_READ(PCH_LVDS) & LVDS_PIPEB_SELECT)
+			panel_pipe = PIPE_B;
+		/* XXX: else fix for eDP */
+	} else if (IS_VALLEYVIEW(dev)) {
+		/* presumably write lock depends on pipe, not port select */
+		pp_reg = VLV_PIPE_PP_CONTROL(pipe);
+		panel_pipe = pipe;
 	} else {
 		pp_reg = PP_CONTROL;
-		lvds_reg = LVDS;
+		if (I915_READ(LVDS) & LVDS_PIPEB_SELECT)
+			panel_pipe = PIPE_B;
 	}
 
 	val = I915_READ(pp_reg);
 	if (!(val & PANEL_POWER_ON) ||
-	    ((val & PANEL_UNLOCK_REGS) == PANEL_UNLOCK_REGS))
+	    ((val & PANEL_UNLOCK_MASK) == PANEL_UNLOCK_REGS))
 		locked = false;
 
-	if (I915_READ(lvds_reg) & LVDS_PIPEB_SELECT)
-		panel_pipe = PIPE_B;
-
 	WARN(panel_pipe == pipe && locked,
 	     "panel assertion failure, pipe %c regs locked\n",
 	     pipe_name(pipe));
@@ -1245,11 +1222,12 @@ void assert_pipe(struct drm_i915_private *dev_priv,
 	enum transcoder cpu_transcoder = intel_pipe_to_cpu_transcoder(dev_priv,
 								      pipe);
 
-	/* if we need the pipe A quirk it must be always on */
-	if (pipe == PIPE_A && dev_priv->quirks & QUIRK_PIPEA_FORCE)
+	/* if we need the pipe quirk it must be always on */
+	if ((pipe == PIPE_A && dev_priv->quirks & QUIRK_PIPEA_FORCE) ||
+	    (pipe == PIPE_B && dev_priv->quirks & QUIRK_PIPEB_FORCE))
 		state = true;
 
-	if (!intel_display_power_enabled(dev_priv,
+	if (!intel_display_power_is_enabled(dev_priv,
 				POWER_DOMAIN_TRANSCODER(cpu_transcoder))) {
 		cur_state = false;
 	} else {
@@ -1300,7 +1278,7 @@ static void assert_planes_disabled(struct drm_i915_private *dev_priv,
 	}
 
 	/* Need to check both planes against the pipe */
-	for_each_pipe(i) {
+	for_each_pipe(dev_priv, i) {
 		reg = DSPCNTR(i);
 		val = I915_READ(reg);
 		cur_pipe = (val & DISPPLANE_SEL_PIPE_MASK) >>
@@ -1318,7 +1296,14 @@ static void assert_sprites_disabled(struct drm_i915_private *dev_priv,
 	int reg, sprite;
 	u32 val;
 
-	if (IS_VALLEYVIEW(dev)) {
+	if (INTEL_INFO(dev)->gen >= 9) {
+		for_each_sprite(pipe, sprite) {
+			val = I915_READ(PLANE_CTL(pipe, sprite));
+			WARN(val & PLANE_CTL_ENABLE,
+			     "plane %d assertion failure, should be off on pipe %c but is still active\n",
+			     sprite, pipe_name(pipe));
+		}
+	} else if (IS_VALLEYVIEW(dev)) {
 		for_each_sprite(pipe, sprite) {
 			reg = SPCNTR(pipe, sprite);
 			val = I915_READ(reg);
@@ -1341,6 +1326,12 @@ static void assert_sprites_disabled(struct drm_i915_private *dev_priv,
 	}
 }
 
+static void assert_vblank_disabled(struct drm_crtc *crtc)
+{
+	if (WARN_ON(drm_crtc_vblank_get(crtc) == 0))
+		drm_crtc_vblank_put(crtc);
+}
+
 static void ibx_assert_pch_refclk_enabled(struct drm_i915_private *dev_priv)
 {
 	u32 val;
@@ -1513,40 +1504,13 @@ static void intel_init_dpio(struct drm_device *dev)
 	}
 }
 
-static void intel_reset_dpio(struct drm_device *dev)
-{
-	struct drm_i915_private *dev_priv = dev->dev_private;
-
-	if (IS_CHERRYVIEW(dev)) {
-		enum dpio_phy phy;
-		u32 val;
-
-		for (phy = DPIO_PHY0; phy < I915_NUM_PHYS_VLV; phy++) {
-			/* Poll for phypwrgood signal */
-			if (wait_for(I915_READ(DISPLAY_PHY_STATUS) &
-						PHY_POWERGOOD(phy), 1))
-				DRM_ERROR("Display PHY %d is not power up\n", phy);
-
-			/*
-			 * Deassert common lane reset for PHY.
-			 *
-			 * This should only be done on init and resume from S3
-			 * with both PLLs disabled, or we risk losing DPIO and
-			 * PLL synchronization.
-			 */
-			val = I915_READ(DISPLAY_PHY_CONTROL);
-			I915_WRITE(DISPLAY_PHY_CONTROL,
-				PHY_COM_LANE_RESET_DEASSERT(phy, val));
-		}
-	}
-}
-
-static void vlv_enable_pll(struct intel_crtc *crtc)
+static void vlv_enable_pll(struct intel_crtc *crtc,
+			   const struct intel_crtc_config *pipe_config)
 {
 	struct drm_device *dev = crtc->base.dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	int reg = DPLL(crtc->pipe);
-	u32 dpll = crtc->config.dpll_hw_state.dpll;
+	u32 dpll = pipe_config->dpll_hw_state.dpll;
 
 	assert_pipe_disabled(dev_priv, crtc->pipe);
 
@@ -1554,7 +1518,7 @@ static void vlv_enable_pll(struct intel_crtc *crtc)
 	BUG_ON(!IS_VALLEYVIEW(dev_priv->dev));
 
 	/* PLL is protected by panel, make sure we can write it */
-	if (IS_MOBILE(dev_priv->dev) && !IS_I830(dev_priv->dev))
+	if (IS_MOBILE(dev_priv->dev))
 		assert_panel_unlocked(dev_priv, crtc->pipe);
 
 	I915_WRITE(reg, dpll);
@@ -1564,7 +1528,7 @@ static void vlv_enable_pll(struct intel_crtc *crtc)
 	if (wait_for(((I915_READ(reg) & DPLL_LOCK_VLV) == DPLL_LOCK_VLV), 1))
 		DRM_ERROR("DPLL %d failed to lock\n", crtc->pipe);
 
-	I915_WRITE(DPLL_MD(crtc->pipe), crtc->config.dpll_hw_state.dpll_md);
+	I915_WRITE(DPLL_MD(crtc->pipe), pipe_config->dpll_hw_state.dpll_md);
 	POSTING_READ(DPLL_MD(crtc->pipe));
 
 	/* We do this three times for luck */
@@ -1579,7 +1543,8 @@ static void vlv_enable_pll(struct intel_crtc *crtc)
 	udelay(150); /* wait for warmup */
 }
 
-static void chv_enable_pll(struct intel_crtc *crtc)
+static void chv_enable_pll(struct intel_crtc *crtc,
+			   const struct intel_crtc_config *pipe_config)
 {
 	struct drm_device *dev = crtc->base.dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
@@ -1604,19 +1569,31 @@ static void chv_enable_pll(struct intel_crtc *crtc)
 	udelay(1);
 
 	/* Enable PLL */
-	I915_WRITE(DPLL(pipe), crtc->config.dpll_hw_state.dpll);
+	I915_WRITE(DPLL(pipe), pipe_config->dpll_hw_state.dpll);
 
 	/* Check PLL is locked */
 	if (wait_for(((I915_READ(DPLL(pipe)) & DPLL_LOCK_VLV) == DPLL_LOCK_VLV), 1))
 		DRM_ERROR("PLL %d failed to lock\n", pipe);
 
 	/* not sure when this should be written */
-	I915_WRITE(DPLL_MD(pipe), crtc->config.dpll_hw_state.dpll_md);
+	I915_WRITE(DPLL_MD(pipe), pipe_config->dpll_hw_state.dpll_md);
 	POSTING_READ(DPLL_MD(pipe));
 
 	mutex_unlock(&dev_priv->dpio_lock);
 }
 
+static int intel_num_dvo_pipes(struct drm_device *dev)
+{
+	struct intel_crtc *crtc;
+	int count = 0;
+
+	for_each_intel_crtc(dev, crtc)
+		count += crtc->active &&
+			intel_pipe_has_type(crtc, INTEL_OUTPUT_DVO);
+
+	return count;
+}
+
 static void i9xx_enable_pll(struct intel_crtc *crtc)
 {
 	struct drm_device *dev = crtc->base.dev;
@@ -1633,7 +1610,18 @@ static void i9xx_enable_pll(struct intel_crtc *crtc)
 	if (IS_MOBILE(dev) && !IS_I830(dev))
 		assert_panel_unlocked(dev_priv, crtc->pipe);
 
-	I915_WRITE(reg, dpll);
+	/* Enable DVO 2x clock on both PLLs if necessary */
+	if (IS_I830(dev) && intel_num_dvo_pipes(dev) > 0) {
+		/*
+		 * It appears to be important that we don't enable this
+		 * for the current pipe before otherwise configuring the
+		 * PLL. No idea how this should be handled if multiple
+		 * DVO outputs are enabled simultaneosly.
+		 */
+		dpll |= DPLL_DVO_2X_MODE;
+		I915_WRITE(DPLL(!crtc->pipe),
+			   I915_READ(DPLL(!crtc->pipe)) | DPLL_DVO_2X_MODE);
+	}
 
 	/* Wait for the clocks to stabilize. */
 	POSTING_READ(reg);
@@ -1672,10 +1660,25 @@ static void i9xx_enable_pll(struct intel_crtc *crtc)
  *
  * Note!  This is for pre-ILK only.
  */
-static void i9xx_disable_pll(struct drm_i915_private *dev_priv, enum pipe pipe)
+static void i9xx_disable_pll(struct intel_crtc *crtc)
 {
-	/* Don't disable pipe A or pipe A PLLs if needed */
-	if (pipe == PIPE_A && (dev_priv->quirks & QUIRK_PIPEA_FORCE))
+	struct drm_device *dev = crtc->base.dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	enum pipe pipe = crtc->pipe;
+
+	/* Disable DVO 2x clock on both PLLs if necessary */
+	if (IS_I830(dev) &&
+	    intel_pipe_has_type(crtc, INTEL_OUTPUT_DVO) &&
+	    intel_num_dvo_pipes(dev) == 1) {
+		I915_WRITE(DPLL(PIPE_B),
+			   I915_READ(DPLL(PIPE_B)) & ~DPLL_DVO_2X_MODE);
+		I915_WRITE(DPLL(PIPE_A),
+			   I915_READ(DPLL(PIPE_A)) & ~DPLL_DVO_2X_MODE);
+	}
+
+	/* Don't disable pipe or pipe PLLs if needed */
+	if ((pipe == PIPE_A && dev_priv->quirks & QUIRK_PIPEA_FORCE) ||
+	    (pipe == PIPE_B && dev_priv->quirks & QUIRK_PIPEB_FORCE))
 		return;
 
 	/* Make sure the pipe isn't still relying on us */
@@ -1712,7 +1715,7 @@ static void chv_disable_pll(struct drm_i915_private *dev_priv, enum pipe pipe)
 	assert_pipe_disabled(dev_priv, pipe);
 
 	/* Set PLL en = 0 */
-	val = DPLL_SSC_REF_CLOCK_CHV;
+	val = DPLL_SSC_REF_CLOCK_CHV | DPLL_REFA_CLK_ENABLE_VLV;
 	if (pipe != PIPE_A)
 		val |= DPLL_INTEGRATED_CRI_CLK_VLV;
 	I915_WRITE(DPLL(pipe), val);
@@ -1776,7 +1779,7 @@ static void intel_prepare_shared_dpll(struct intel_crtc *crtc)
 	if (WARN_ON(pll == NULL))
 		return;
 
-	WARN_ON(!pll->refcount);
+	WARN_ON(!pll->config.crtc_mask);
 	if (pll->active == 0) {
 		DRM_DEBUG_DRIVER("setting up %s\n", pll->name);
 		WARN_ON(pll->on);
@@ -1803,10 +1806,10 @@ static void intel_enable_shared_dpll(struct intel_crtc *crtc)
 	if (WARN_ON(pll == NULL))
 		return;
 
-	if (WARN_ON(pll->refcount == 0))
+	if (WARN_ON(pll->config.crtc_mask == 0))
 		return;
 
-	DRM_DEBUG_KMS("enable %s (active %d, on? %d)for crtc %d\n",
+	DRM_DEBUG_KMS("enable %s (active %d, on? %d) for crtc %d\n",
 		      pll->name, pll->active, pll->on,
 		      crtc->base.base.id);
 
@@ -1824,7 +1827,7 @@ static void intel_enable_shared_dpll(struct intel_crtc *crtc)
 	pll->on = true;
 }
 
-void intel_disable_shared_dpll(struct intel_crtc *crtc)
+static void intel_disable_shared_dpll(struct intel_crtc *crtc)
 {
 	struct drm_device *dev = crtc->base.dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
@@ -1835,7 +1838,7 @@ void intel_disable_shared_dpll(struct intel_crtc *crtc)
 	if (WARN_ON(pll == NULL))
 	       return;
 
-	if (WARN_ON(pll->refcount == 0))
+	if (WARN_ON(pll->config.crtc_mask == 0))
 		return;
 
 	DRM_DEBUG_KMS("disable %s (active %d, on? %d) for crtc %d\n",
@@ -1868,7 +1871,7 @@ static void ironlake_enable_pch_transcoder(struct drm_i915_private *dev_priv,
 	uint32_t reg, val, pipeconf_val;
 
 	/* PCH only available on ILK+ */
-	BUG_ON(INTEL_INFO(dev)->gen < 5);
+	BUG_ON(!HAS_PCH_SPLIT(dev));
 
 	/* Make sure PCH DPLL is enabled */
 	assert_shared_dpll_enabled(dev_priv,
@@ -1903,7 +1906,7 @@ static void ironlake_enable_pch_transcoder(struct drm_i915_private *dev_priv,
 	val &= ~TRANS_INTERLACE_MASK;
 	if ((pipeconf_val & PIPECONF_INTERLACE_MASK) == PIPECONF_INTERLACED_ILK)
 		if (HAS_PCH_IBX(dev_priv->dev) &&
-		    intel_pipe_has_type(crtc, INTEL_OUTPUT_SDVO))
+		    intel_pipe_has_type(intel_crtc, INTEL_OUTPUT_SDVO))
 			val |= TRANS_LEGACY_INTERLACED_ILK;
 		else
 			val |= TRANS_INTERLACED;
@@ -1921,7 +1924,7 @@ static void lpt_enable_pch_transcoder(struct drm_i915_private *dev_priv,
 	u32 val, pipeconf_val;
 
 	/* PCH only available on ILK+ */
-	BUG_ON(INTEL_INFO(dev_priv->dev)->gen < 5);
+	BUG_ON(!HAS_PCH_SPLIT(dev_priv->dev));
 
 	/* FDI must be feeding us bits for PCH ports */
 	assert_fdi_tx_enabled(dev_priv, (enum pipe) cpu_transcoder);
@@ -2026,7 +2029,7 @@ static void intel_enable_pipe(struct intel_crtc *crtc)
 	 * need the check.
 	 */
 	if (!HAS_PCH_SPLIT(dev_priv->dev))
-		if (intel_pipe_has_type(&crtc->base, INTEL_OUTPUT_DSI))
+		if (intel_pipe_has_type(crtc, INTEL_OUTPUT_DSI))
 			assert_dsi_pll_enabled(dev_priv);
 		else
 			assert_pll_enabled(dev_priv, pipe);
@@ -2043,8 +2046,8 @@ static void intel_enable_pipe(struct intel_crtc *crtc)
 	reg = PIPECONF(cpu_transcoder);
 	val = I915_READ(reg);
 	if (val & PIPECONF_ENABLE) {
-		WARN_ON(!(pipe == PIPE_A &&
-			  dev_priv->quirks & QUIRK_PIPEA_FORCE));
+		WARN_ON(!((pipe == PIPE_A && dev_priv->quirks & QUIRK_PIPEA_FORCE) ||
+			  (pipe == PIPE_B && dev_priv->quirks & QUIRK_PIPEB_FORCE)));
 		return;
 	}
 
@@ -2054,21 +2057,19 @@ static void intel_enable_pipe(struct intel_crtc *crtc)
 
 /**
  * intel_disable_pipe - disable a pipe, asserting requirements
- * @dev_priv: i915 private structure
- * @pipe: pipe to disable
- *
- * Disable @pipe, making sure that various hardware specific requirements
- * are met, if applicable, e.g. plane disabled, panel fitter off, etc.
+ * @crtc: crtc whose pipes is to be disabled
  *
- * @pipe should be %PIPE_A or %PIPE_B.
+ * Disable the pipe of @crtc, making sure that various hardware
+ * specific requirements are met, if applicable, e.g. plane
+ * disabled, panel fitter off, etc.
  *
  * Will wait until the pipe has shut down before returning.
  */
-static void intel_disable_pipe(struct drm_i915_private *dev_priv,
-			       enum pipe pipe)
+static void intel_disable_pipe(struct intel_crtc *crtc)
 {
-	enum transcoder cpu_transcoder = intel_pipe_to_cpu_transcoder(dev_priv,
-								      pipe);
+	struct drm_i915_private *dev_priv = crtc->base.dev->dev_private;
+	enum transcoder cpu_transcoder = crtc->config.cpu_transcoder;
+	enum pipe pipe = crtc->pipe;
 	int reg;
 	u32 val;
 
@@ -2080,17 +2081,26 @@ static void intel_disable_pipe(struct drm_i915_private *dev_priv,
 	assert_cursor_disabled(dev_priv, pipe);
 	assert_sprites_disabled(dev_priv, pipe);
 
-	/* Don't disable pipe A or pipe A PLLs if needed */
-	if (pipe == PIPE_A && (dev_priv->quirks & QUIRK_PIPEA_FORCE))
-		return;
-
 	reg = PIPECONF(cpu_transcoder);
 	val = I915_READ(reg);
 	if ((val & PIPECONF_ENABLE) == 0)
 		return;
 
-	I915_WRITE(reg, val & ~PIPECONF_ENABLE);
-	intel_wait_for_pipe_off(dev_priv->dev, pipe);
+	/*
+	 * Double wide has implications for planes
+	 * so best keep it disabled when not needed.
+	 */
+	if (crtc->config.double_wide)
+		val &= ~PIPECONF_DOUBLE_WIDE;
+
+	/* Don't disable pipe or pipe PLLs if needed */
+	if (!(pipe == PIPE_A && dev_priv->quirks & QUIRK_PIPEA_FORCE) &&
+	    !(pipe == PIPE_B && dev_priv->quirks & QUIRK_PIPEB_FORCE))
+		val &= ~PIPECONF_ENABLE;
+
+	I915_WRITE(reg, val);
+	if ((val & PIPECONF_ENABLE) == 0)
+		intel_wait_for_pipe_off(crtc);
 }
 
 /*
@@ -2109,35 +2119,28 @@ void intel_flush_primary_plane(struct drm_i915_private *dev_priv,
 
 /**
  * intel_enable_primary_hw_plane - enable the primary plane on a given pipe
- * @dev_priv: i915 private structure
- * @plane: plane to enable
- * @pipe: pipe being fed
+ * @plane:  plane to be enabled
+ * @crtc: crtc for the plane
  *
- * Enable @plane on @pipe, making sure that @pipe is running first.
+ * Enable @plane on @crtc, making sure that the pipe is running first.
  */
-static void intel_enable_primary_hw_plane(struct drm_i915_private *dev_priv,
-					  enum plane plane, enum pipe pipe)
+static void intel_enable_primary_hw_plane(struct drm_plane *plane,
+					  struct drm_crtc *crtc)
 {
-	struct drm_device *dev = dev_priv->dev;
-	struct intel_crtc *intel_crtc =
-		to_intel_crtc(dev_priv->pipe_to_crtc_mapping[pipe]);
-	int reg;
-	u32 val;
+	struct drm_device *dev = plane->dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 
 	/* If the pipe isn't enabled, we can't pump pixels and may hang */
-	assert_pipe_enabled(dev_priv, pipe);
+	assert_pipe_enabled(dev_priv, intel_crtc->pipe);
 
 	if (intel_crtc->primary_enabled)
 		return;
 
 	intel_crtc->primary_enabled = true;
 
-	reg = DSPCNTR(plane);
-	val = I915_READ(reg);
-	WARN_ON(val & DISPLAY_PLANE_ENABLE);
-
-	I915_WRITE(reg, val | DISPLAY_PLANE_ENABLE);
-	intel_flush_primary_plane(dev_priv, plane);
+	dev_priv->display.update_primary_plane(crtc, plane->fb,
+					       crtc->x, crtc->y);
 
 	/*
 	 * BDW signals flip done immediately if the plane
@@ -2150,31 +2153,27 @@ static void intel_enable_primary_hw_plane(struct drm_i915_private *dev_priv,
 
 /**
  * intel_disable_primary_hw_plane - disable the primary hardware plane
- * @dev_priv: i915 private structure
- * @plane: plane to disable
- * @pipe: pipe consuming the data
+ * @plane: plane to be disabled
+ * @crtc: crtc for the plane
  *
- * Disable @plane; should be an independent operation.
+ * Disable @plane on @crtc, making sure that the pipe is running first.
  */
-static void intel_disable_primary_hw_plane(struct drm_i915_private *dev_priv,
-					   enum plane plane, enum pipe pipe)
+static void intel_disable_primary_hw_plane(struct drm_plane *plane,
+					   struct drm_crtc *crtc)
 {
-	struct intel_crtc *intel_crtc =
-		to_intel_crtc(dev_priv->pipe_to_crtc_mapping[pipe]);
-	int reg;
-	u32 val;
+	struct drm_device *dev = plane->dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
+
+	assert_pipe_enabled(dev_priv, intel_crtc->pipe);
 
 	if (!intel_crtc->primary_enabled)
 		return;
 
 	intel_crtc->primary_enabled = false;
 
-	reg = DSPCNTR(plane);
-	val = I915_READ(reg);
-	WARN_ON((val & DISPLAY_PLANE_ENABLE) == 0);
-
-	I915_WRITE(reg, val & ~DISPLAY_PLANE_ENABLE);
-	intel_flush_primary_plane(dev_priv, plane);
+	dev_priv->display.update_primary_plane(crtc, plane->fb,
+					       crtc->x, crtc->y);
 }
 
 static bool need_vtd_wa(struct drm_device *dev)
@@ -2195,11 +2194,13 @@ static int intel_align_height(struct drm_device *dev, int height, bool tiled)
 }
 
 int
-intel_pin_and_fence_fb_obj(struct drm_device *dev,
-			   struct drm_i915_gem_object *obj,
+intel_pin_and_fence_fb_obj(struct drm_plane *plane,
+			   struct drm_framebuffer *fb,
 			   struct intel_engine_cs *pipelined)
 {
+	struct drm_device *dev = fb->dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_gem_object *obj = intel_fb_obj(fb);
 	u32 alignment;
 	int ret;
 
@@ -2207,7 +2208,9 @@ intel_pin_and_fence_fb_obj(struct drm_device *dev,
 
 	switch (obj->tiling_mode) {
 	case I915_TILING_NONE:
-		if (IS_BROADWATER(dev) || IS_CRESTLINE(dev))
+		if (INTEL_INFO(dev)->gen >= 9)
+			alignment = 256 * 1024;
+		else if (IS_BROADWATER(dev) || IS_CRESTLINE(dev))
 			alignment = 128 * 1024;
 		else if (INTEL_INFO(dev)->gen >= 4)
 			alignment = 4 * 1024;
@@ -2215,8 +2218,12 @@ intel_pin_and_fence_fb_obj(struct drm_device *dev,
 			alignment = 64 * 1024;
 		break;
 	case I915_TILING_X:
-		/* pin() will align the object as required by fence */
-		alignment = 0;
+		if (INTEL_INFO(dev)->gen >= 9)
+			alignment = 256 * 1024;
+		else {
+			/* pin() will align the object as required by fence */
+			alignment = 0;
+		}
 		break;
 	case I915_TILING_Y:
 		WARN(1, "Y tiled bo slipped through, driver bug!\n");
@@ -2376,6 +2383,7 @@ static void intel_find_plane_obj(struct intel_crtc *intel_crtc,
 				 struct intel_plane_config *plane_config)
 {
 	struct drm_device *dev = intel_crtc->base.dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct drm_crtc *c;
 	struct intel_crtc *i;
 	struct drm_i915_gem_object *obj;
@@ -2407,6 +2415,9 @@ static void intel_find_plane_obj(struct intel_crtc *intel_crtc,
 			continue;
 
 		if (i915_gem_obj_ggtt_offset(obj) == plane_config->base) {
+			if (obj->tiling_mode != I915_TILING_NONE)
+				dev_priv->preserve_bios_swizzle = true;
+
 			drm_framebuffer_reference(c->primary->fb);
 			intel_crtc->base.primary->fb = c->primary->fb;
 			obj->frontbuffer_bits |= INTEL_FRONTBUFFER_PRIMARY(intel_crtc->pipe);
@@ -2422,16 +2433,52 @@ static void i9xx_update_primary_plane(struct drm_crtc *crtc,
 	struct drm_device *dev = crtc->dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
-	struct drm_i915_gem_object *obj = intel_fb_obj(fb);
+	struct drm_i915_gem_object *obj;
 	int plane = intel_crtc->plane;
 	unsigned long linear_offset;
 	u32 dspcntr;
-	u32 reg;
+	u32 reg = DSPCNTR(plane);
+	int pixel_size;
+
+	if (!intel_crtc->primary_enabled) {
+		I915_WRITE(reg, 0);
+		if (INTEL_INFO(dev)->gen >= 4)
+			I915_WRITE(DSPSURF(plane), 0);
+		else
+			I915_WRITE(DSPADDR(plane), 0);
+		POSTING_READ(reg);
+		return;
+	}
+
+	obj = intel_fb_obj(fb);
+	if (WARN_ON(obj == NULL))
+		return;
+
+	pixel_size = drm_format_plane_cpp(fb->pixel_format, 0);
+
+	dspcntr = DISPPLANE_GAMMA_ENABLE;
+
+	dspcntr |= DISPLAY_PLANE_ENABLE;
+
+	if (INTEL_INFO(dev)->gen < 4) {
+		if (intel_crtc->pipe == PIPE_B)
+			dspcntr |= DISPPLANE_SEL_PIPE_B;
+
+		/* pipesrc and dspsize control the size that is scaled from,
+		 * which should always be the user's requested size.
+		 */
+		I915_WRITE(DSPSIZE(plane),
+			   ((intel_crtc->config.pipe_src_h - 1) << 16) |
+			   (intel_crtc->config.pipe_src_w - 1));
+		I915_WRITE(DSPPOS(plane), 0);
+	} else if (IS_CHERRYVIEW(dev) && plane == PLANE_B) {
+		I915_WRITE(PRIMSIZE(plane),
+			   ((intel_crtc->config.pipe_src_h - 1) << 16) |
+			   (intel_crtc->config.pipe_src_w - 1));
+		I915_WRITE(PRIMPOS(plane), 0);
+		I915_WRITE(PRIMCNSTALPHA(plane), 0);
+	}
 
-	reg = DSPCNTR(plane);
-	dspcntr = I915_READ(reg);
-	/* Mask out pixel format bits in case we change it */
-	dspcntr &= ~DISPPLANE_PIXFORMAT_MASK;
 	switch (fb->pixel_format) {
 	case DRM_FORMAT_C8:
 		dspcntr |= DISPPLANE_8BPP;
@@ -2463,30 +2510,40 @@ static void i9xx_update_primary_plane(struct drm_crtc *crtc,
 		BUG();
 	}
 
-	if (INTEL_INFO(dev)->gen >= 4) {
-		if (obj->tiling_mode != I915_TILING_NONE)
-			dspcntr |= DISPPLANE_TILED;
-		else
-			dspcntr &= ~DISPPLANE_TILED;
-	}
+	if (INTEL_INFO(dev)->gen >= 4 &&
+	    obj->tiling_mode != I915_TILING_NONE)
+		dspcntr |= DISPPLANE_TILED;
 
 	if (IS_G4X(dev))
 		dspcntr |= DISPPLANE_TRICKLE_FEED_DISABLE;
 
-	I915_WRITE(reg, dspcntr);
-
-	linear_offset = y * fb->pitches[0] + x * (fb->bits_per_pixel / 8);
+	linear_offset = y * fb->pitches[0] + x * pixel_size;
 
 	if (INTEL_INFO(dev)->gen >= 4) {
 		intel_crtc->dspaddr_offset =
 			intel_gen4_compute_page_offset(&x, &y, obj->tiling_mode,
-						       fb->bits_per_pixel / 8,
+						       pixel_size,
 						       fb->pitches[0]);
 		linear_offset -= intel_crtc->dspaddr_offset;
 	} else {
 		intel_crtc->dspaddr_offset = linear_offset;
 	}
 
+	if (to_intel_plane(crtc->primary)->rotation == BIT(DRM_ROTATE_180)) {
+		dspcntr |= DISPPLANE_ROTATE_180;
+
+		x += (intel_crtc->config.pipe_src_w - 1);
+		y += (intel_crtc->config.pipe_src_h - 1);
+
+		/* Finding the last pixel of the last line of the display
+		data and adding to linear_offset*/
+		linear_offset +=
+			(intel_crtc->config.pipe_src_h - 1) * fb->pitches[0] +
+			(intel_crtc->config.pipe_src_w - 1) * pixel_size;
+	}
+
+	I915_WRITE(reg, dspcntr);
+
 	DRM_DEBUG_KMS("Writing base %08lX %08lX %d %d %d\n",
 		      i915_gem_obj_ggtt_offset(obj), linear_offset, x, y,
 		      fb->pitches[0]);
@@ -2508,16 +2565,33 @@ static void ironlake_update_primary_plane(struct drm_crtc *crtc,
 	struct drm_device *dev = crtc->dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
-	struct drm_i915_gem_object *obj = intel_fb_obj(fb);
+	struct drm_i915_gem_object *obj;
 	int plane = intel_crtc->plane;
 	unsigned long linear_offset;
 	u32 dspcntr;
-	u32 reg;
+	u32 reg = DSPCNTR(plane);
+	int pixel_size;
+
+	if (!intel_crtc->primary_enabled) {
+		I915_WRITE(reg, 0);
+		I915_WRITE(DSPSURF(plane), 0);
+		POSTING_READ(reg);
+		return;
+	}
+
+	obj = intel_fb_obj(fb);
+	if (WARN_ON(obj == NULL))
+		return;
+
+	pixel_size = drm_format_plane_cpp(fb->pixel_format, 0);
+
+	dspcntr = DISPPLANE_GAMMA_ENABLE;
+
+	dspcntr |= DISPLAY_PLANE_ENABLE;
+
+	if (IS_HASWELL(dev) || IS_BROADWELL(dev))
+		dspcntr |= DISPPLANE_PIPE_CSC_ENABLE;
 
-	reg = DSPCNTR(plane);
-	dspcntr = I915_READ(reg);
-	/* Mask out pixel format bits in case we change it */
-	dspcntr &= ~DISPPLANE_PIXFORMAT_MASK;
 	switch (fb->pixel_format) {
 	case DRM_FORMAT_C8:
 		dspcntr |= DISPPLANE_8BPP;
@@ -2547,22 +2621,32 @@ static void ironlake_update_primary_plane(struct drm_crtc *crtc,
 
 	if (obj->tiling_mode != I915_TILING_NONE)
 		dspcntr |= DISPPLANE_TILED;
-	else
-		dspcntr &= ~DISPPLANE_TILED;
 
-	if (IS_HASWELL(dev) || IS_BROADWELL(dev))
-		dspcntr &= ~DISPPLANE_TRICKLE_FEED_DISABLE;
-	else
+	if (!IS_HASWELL(dev) && !IS_BROADWELL(dev))
 		dspcntr |= DISPPLANE_TRICKLE_FEED_DISABLE;
 
-	I915_WRITE(reg, dspcntr);
-
-	linear_offset = y * fb->pitches[0] + x * (fb->bits_per_pixel / 8);
+	linear_offset = y * fb->pitches[0] + x * pixel_size;
 	intel_crtc->dspaddr_offset =
 		intel_gen4_compute_page_offset(&x, &y, obj->tiling_mode,
-					       fb->bits_per_pixel / 8,
+					       pixel_size,
 					       fb->pitches[0]);
 	linear_offset -= intel_crtc->dspaddr_offset;
+	if (to_intel_plane(crtc->primary)->rotation == BIT(DRM_ROTATE_180)) {
+		dspcntr |= DISPPLANE_ROTATE_180;
+
+		if (!IS_HASWELL(dev) && !IS_BROADWELL(dev)) {
+			x += (intel_crtc->config.pipe_src_w - 1);
+			y += (intel_crtc->config.pipe_src_h - 1);
+
+			/* Finding the last pixel of the last line of the display
+			data and adding to linear_offset*/
+			linear_offset +=
+				(intel_crtc->config.pipe_src_h - 1) * fb->pitches[0] +
+				(intel_crtc->config.pipe_src_w - 1) * pixel_size;
+		}
+	}
+
+	I915_WRITE(reg, dspcntr);
 
 	DRM_DEBUG_KMS("Writing base %08lX %08lX %d %d %d\n",
 		      i915_gem_obj_ggtt_offset(obj), linear_offset, x, y,
@@ -2579,6 +2663,92 @@ static void ironlake_update_primary_plane(struct drm_crtc *crtc,
 	POSTING_READ(reg);
 }
 
+static void skylake_update_primary_plane(struct drm_crtc *crtc,
+					 struct drm_framebuffer *fb,
+					 int x, int y)
+{
+	struct drm_device *dev = crtc->dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
+	struct intel_framebuffer *intel_fb;
+	struct drm_i915_gem_object *obj;
+	int pipe = intel_crtc->pipe;
+	u32 plane_ctl, stride;
+
+	if (!intel_crtc->primary_enabled) {
+		I915_WRITE(PLANE_CTL(pipe, 0), 0);
+		I915_WRITE(PLANE_SURF(pipe, 0), 0);
+		POSTING_READ(PLANE_CTL(pipe, 0));
+		return;
+	}
+
+	plane_ctl = PLANE_CTL_ENABLE |
+		    PLANE_CTL_PIPE_GAMMA_ENABLE |
+		    PLANE_CTL_PIPE_CSC_ENABLE;
+
+	switch (fb->pixel_format) {
+	case DRM_FORMAT_RGB565:
+		plane_ctl |= PLANE_CTL_FORMAT_RGB_565;
+		break;
+	case DRM_FORMAT_XRGB8888:
+		plane_ctl |= PLANE_CTL_FORMAT_XRGB_8888;
+		break;
+	case DRM_FORMAT_XBGR8888:
+		plane_ctl |= PLANE_CTL_ORDER_RGBX;
+		plane_ctl |= PLANE_CTL_FORMAT_XRGB_8888;
+		break;
+	case DRM_FORMAT_XRGB2101010:
+		plane_ctl |= PLANE_CTL_FORMAT_XRGB_2101010;
+		break;
+	case DRM_FORMAT_XBGR2101010:
+		plane_ctl |= PLANE_CTL_ORDER_RGBX;
+		plane_ctl |= PLANE_CTL_FORMAT_XRGB_2101010;
+		break;
+	default:
+		BUG();
+	}
+
+	intel_fb = to_intel_framebuffer(fb);
+	obj = intel_fb->obj;
+
+	/*
+	 * The stride is either expressed as a multiple of 64 bytes chunks for
+	 * linear buffers or in number of tiles for tiled buffers.
+	 */
+	switch (obj->tiling_mode) {
+	case I915_TILING_NONE:
+		stride = fb->pitches[0] >> 6;
+		break;
+	case I915_TILING_X:
+		plane_ctl |= PLANE_CTL_TILED_X;
+		stride = fb->pitches[0] >> 9;
+		break;
+	default:
+		BUG();
+	}
+
+	plane_ctl |= PLANE_CTL_PLANE_GAMMA_DISABLE;
+	if (to_intel_plane(crtc->primary)->rotation == BIT(DRM_ROTATE_180))
+		plane_ctl |= PLANE_CTL_ROTATE_180;
+
+	I915_WRITE(PLANE_CTL(pipe, 0), plane_ctl);
+
+	DRM_DEBUG_KMS("Writing base %08lX %d,%d,%d,%d pitch=%d\n",
+		      i915_gem_obj_ggtt_offset(obj),
+		      x, y, fb->width, fb->height,
+		      fb->pitches[0]);
+
+	I915_WRITE(PLANE_POS(pipe, 0), 0);
+	I915_WRITE(PLANE_OFFSET(pipe, 0), (y << 16) | x);
+	I915_WRITE(PLANE_SIZE(pipe, 0),
+		   (intel_crtc->config.pipe_src_h - 1) << 16 |
+		   (intel_crtc->config.pipe_src_w - 1));
+	I915_WRITE(PLANE_STRIDE(pipe, 0), stride);
+	I915_WRITE(PLANE_SURF(pipe, 0), i915_gem_obj_ggtt_offset(obj));
+
+	POSTING_READ(PLANE_SURF(pipe, 0));
+}
+
 /* Assume fb object is pinned & idle & fenced and just update base pointers */
 static int
 intel_pipe_set_base_atomic(struct drm_crtc *crtc, struct drm_framebuffer *fb,
@@ -2589,32 +2759,16 @@ intel_pipe_set_base_atomic(struct drm_crtc *crtc, struct drm_framebuffer *fb,
 
 	if (dev_priv->display.disable_fbc)
 		dev_priv->display.disable_fbc(dev);
-	intel_increase_pllclock(dev, to_intel_crtc(crtc)->pipe);
 
 	dev_priv->display.update_primary_plane(crtc, fb, x, y);
 
 	return 0;
 }
 
-void intel_display_handle_reset(struct drm_device *dev)
+static void intel_complete_page_flips(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct drm_crtc *crtc;
 
-	/*
-	 * Flips in the rings have been nuked by the reset,
-	 * so complete all pending flips so that user space
-	 * will get its events and not get stuck.
-	 *
-	 * Also update the base address of all primary
-	 * planes to the the last fb to make sure we're
-	 * showing the correct fb after a reset.
-	 *
-	 * Need to make two loops over the crtcs so that we
-	 * don't try to grab a crtc mutex before the
-	 * pending_flip_queue really got woken up.
-	 */
-
 	for_each_crtc(dev, crtc) {
 		struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 		enum plane plane = intel_crtc->plane;
@@ -2622,6 +2776,12 @@ void intel_display_handle_reset(struct drm_device *dev)
 		intel_prepare_page_flip(dev, plane);
 		intel_finish_page_flip_plane(dev, plane);
 	}
+}
+
+static void intel_update_primary_planes(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_crtc *crtc;
 
 	for_each_crtc(dev, crtc) {
 		struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
@@ -2641,6 +2801,79 @@ void intel_display_handle_reset(struct drm_device *dev)
 	}
 }
 
+void intel_prepare_reset(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	struct intel_crtc *crtc;
+
+	/* no reset support for gen2 */
+	if (IS_GEN2(dev))
+		return;
+
+	/* reset doesn't touch the display */
+	if (INTEL_INFO(dev)->gen >= 5 || IS_G4X(dev))
+		return;
+
+	drm_modeset_lock_all(dev);
+
+	/*
+	 * Disabling the crtcs gracefully seems nicer. Also the
+	 * g33 docs say we should at least disable all the planes.
+	 */
+	for_each_intel_crtc(dev, crtc) {
+		if (crtc->active)
+			dev_priv->display.crtc_disable(&crtc->base);
+	}
+}
+
+void intel_finish_reset(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = to_i915(dev);
+
+	/*
+	 * Flips in the rings will be nuked by the reset,
+	 * so complete all pending flips so that user space
+	 * will get its events and not get stuck.
+	 */
+	intel_complete_page_flips(dev);
+
+	/* no reset support for gen2 */
+	if (IS_GEN2(dev))
+		return;
+
+	/* reset doesn't touch the display */
+	if (INTEL_INFO(dev)->gen >= 5 || IS_G4X(dev)) {
+		/*
+		 * Flips in the rings have been nuked by the reset,
+		 * so update the base address of all primary
+		 * planes to the the last fb to make sure we're
+		 * showing the correct fb after a reset.
+		 */
+		intel_update_primary_planes(dev);
+		return;
+	}
+
+	/*
+	 * The display has been reset as well,
+	 * so need a full re-initialization.
+	 */
+	intel_runtime_pm_disable_interrupts(dev_priv);
+	intel_runtime_pm_enable_interrupts(dev_priv);
+
+	intel_modeset_init_hw(dev);
+
+	spin_lock_irq(&dev_priv->irq_lock);
+	if (dev_priv->display.hpd_irq_setup)
+		dev_priv->display.hpd_irq_setup(dev);
+	spin_unlock_irq(&dev_priv->irq_lock);
+
+	intel_modeset_setup_hw_state(dev, true);
+
+	intel_hpd_init(dev_priv);
+
+	drm_modeset_unlock_all(dev);
+}
+
 static int
 intel_finish_fb(struct drm_framebuffer *old_fb)
 {
@@ -2669,20 +2902,58 @@ static bool intel_crtc_has_pending_flip(struct drm_crtc *crtc)
 	struct drm_device *dev = crtc->dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
-	unsigned long flags;
 	bool pending;
 
 	if (i915_reset_in_progress(&dev_priv->gpu_error) ||
 	    intel_crtc->reset_counter != atomic_read(&dev_priv->gpu_error.reset_counter))
 		return false;
 
-	spin_lock_irqsave(&dev->event_lock, flags);
+	spin_lock_irq(&dev->event_lock);
 	pending = to_intel_crtc(crtc)->unpin_work != NULL;
-	spin_unlock_irqrestore(&dev->event_lock, flags);
+	spin_unlock_irq(&dev->event_lock);
 
 	return pending;
 }
 
+static void intel_update_pipe_size(struct intel_crtc *crtc)
+{
+	struct drm_device *dev = crtc->base.dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	const struct drm_display_mode *adjusted_mode;
+
+	if (!i915.fastboot)
+		return;
+
+	/*
+	 * Update pipe size and adjust fitter if needed: the reason for this is
+	 * that in compute_mode_changes we check the native mode (not the pfit
+	 * mode) to see if we can flip rather than do a full mode set. In the
+	 * fastboot case, we'll flip, but if we don't update the pipesrc and
+	 * pfit state, we'll end up with a big fb scanned out into the wrong
+	 * sized surface.
+	 *
+	 * To fix this properly, we need to hoist the checks up into
+	 * compute_mode_changes (or above), check the actual pfit state and
+	 * whether the platform allows pfit disable with pipe active, and only
+	 * then update the pipesrc and pfit state, even on the flip path.
+	 */
+
+	adjusted_mode = &crtc->config.adjusted_mode;
+
+	I915_WRITE(PIPESRC(crtc->pipe),
+		   ((adjusted_mode->crtc_hdisplay - 1) << 16) |
+		   (adjusted_mode->crtc_vdisplay - 1));
+	if (!crtc->config.pch_pfit.enabled &&
+	    (intel_pipe_has_type(crtc, INTEL_OUTPUT_LVDS) ||
+	     intel_pipe_has_type(crtc, INTEL_OUTPUT_EDP))) {
+		I915_WRITE(PF_CTL(crtc->pipe), 0);
+		I915_WRITE(PF_WIN_POS(crtc->pipe), 0);
+		I915_WRITE(PF_WIN_SZ(crtc->pipe), 0);
+	}
+	crtc->config.pipe_src_w = adjusted_mode->crtc_hdisplay;
+	crtc->config.pipe_src_h = adjusted_mode->crtc_vdisplay;
+}
+
 static int
 intel_pipe_set_base(struct drm_crtc *crtc, int x, int y,
 		    struct drm_framebuffer *fb)
@@ -2692,7 +2963,6 @@ intel_pipe_set_base(struct drm_crtc *crtc, int x, int y,
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 	enum pipe pipe = intel_crtc->pipe;
 	struct drm_framebuffer *old_fb = crtc->primary->fb;
-	struct drm_i915_gem_object *obj = intel_fb_obj(fb);
 	struct drm_i915_gem_object *old_obj = intel_fb_obj(old_fb);
 	int ret;
 
@@ -2715,9 +2985,9 @@ intel_pipe_set_base(struct drm_crtc *crtc, int x, int y,
 	}
 
 	mutex_lock(&dev->struct_mutex);
-	ret = intel_pin_and_fence_fb_obj(dev, obj, NULL);
+	ret = intel_pin_and_fence_fb_obj(crtc->primary, fb, NULL);
 	if (ret == 0)
-		i915_gem_track_fb(old_obj, obj,
+		i915_gem_track_fb(old_obj, intel_fb_obj(fb),
 				  INTEL_FRONTBUFFER_PRIMARY(pipe));
 	mutex_unlock(&dev->struct_mutex);
 	if (ret != 0) {
@@ -2725,37 +2995,6 @@ intel_pipe_set_base(struct drm_crtc *crtc, int x, int y,
 		return ret;
 	}
 
-	/*
-	 * Update pipe size and adjust fitter if needed: the reason for this is
-	 * that in compute_mode_changes we check the native mode (not the pfit
-	 * mode) to see if we can flip rather than do a full mode set. In the
-	 * fastboot case, we'll flip, but if we don't update the pipesrc and
-	 * pfit state, we'll end up with a big fb scanned out into the wrong
-	 * sized surface.
-	 *
-	 * To fix this properly, we need to hoist the checks up into
-	 * compute_mode_changes (or above), check the actual pfit state and
-	 * whether the platform allows pfit disable with pipe active, and only
-	 * then update the pipesrc and pfit state, even on the flip path.
-	 */
-	if (i915.fastboot) {
-		const struct drm_display_mode *adjusted_mode =
-			&intel_crtc->config.adjusted_mode;
-
-		I915_WRITE(PIPESRC(intel_crtc->pipe),
-			   ((adjusted_mode->crtc_hdisplay - 1) << 16) |
-			   (adjusted_mode->crtc_vdisplay - 1));
-		if (!intel_crtc->config.pch_pfit.enabled &&
-		    (intel_pipe_has_type(crtc, INTEL_OUTPUT_LVDS) ||
-		     intel_pipe_has_type(crtc, INTEL_OUTPUT_EDP))) {
-			I915_WRITE(PF_CTL(intel_crtc->pipe), 0);
-			I915_WRITE(PF_WIN_POS(intel_crtc->pipe), 0);
-			I915_WRITE(PF_WIN_SZ(intel_crtc->pipe), 0);
-		}
-		intel_crtc->config.pipe_src_w = adjusted_mode->crtc_hdisplay;
-		intel_crtc->config.pipe_src_h = adjusted_mode->crtc_vdisplay;
-	}
-
 	dev_priv->display.update_primary_plane(crtc, fb, x, y);
 
 	if (intel_crtc->active)
@@ -3346,23 +3585,53 @@ bool intel_has_pending_fb_unpin(struct drm_device *dev)
 	return false;
 }
 
+static void page_flip_completed(struct intel_crtc *intel_crtc)
+{
+	struct drm_i915_private *dev_priv = to_i915(intel_crtc->base.dev);
+	struct intel_unpin_work *work = intel_crtc->unpin_work;
+
+	/* ensure that the unpin work is consistent wrt ->pending. */
+	smp_rmb();
+	intel_crtc->unpin_work = NULL;
+
+	if (work->event)
+		drm_send_vblank_event(intel_crtc->base.dev,
+				      intel_crtc->pipe,
+				      work->event);
+
+	drm_crtc_vblank_put(&intel_crtc->base);
+
+	wake_up_all(&dev_priv->pending_flip_queue);
+	queue_work(dev_priv->wq, &work->work);
+
+	trace_i915_flip_complete(intel_crtc->plane,
+				 work->pending_flip_obj);
+}
+
 void intel_crtc_wait_for_pending_flips(struct drm_crtc *crtc)
 {
 	struct drm_device *dev = crtc->dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 
-	if (crtc->primary->fb == NULL)
-		return;
-
 	WARN_ON(waitqueue_active(&dev_priv->pending_flip_queue));
+	if (WARN_ON(wait_event_timeout(dev_priv->pending_flip_queue,
+				       !intel_crtc_has_pending_flip(crtc),
+				       60*HZ) == 0)) {
+		struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 
-	WARN_ON(wait_event_timeout(dev_priv->pending_flip_queue,
-				   !intel_crtc_has_pending_flip(crtc),
-				   60*HZ) == 0);
+		spin_lock_irq(&dev->event_lock);
+		if (intel_crtc->unpin_work) {
+			WARN_ONCE(1, "Removing stuck page flip\n");
+			page_flip_completed(intel_crtc);
+		}
+		spin_unlock_irq(&dev->event_lock);
+	}
 
-	mutex_lock(&dev->struct_mutex);
-	intel_finish_fb(crtc->primary->fb);
-	mutex_unlock(&dev->struct_mutex);
+	if (crtc->primary->fb) {
+		mutex_lock(&dev->struct_mutex);
+		intel_finish_fb(crtc->primary->fb);
+		mutex_unlock(&dev->struct_mutex);
+	}
 }
 
 /* Program iCLKIP clock to the desired frequency */
@@ -3580,9 +3849,7 @@ static void ironlake_pch_enable(struct drm_crtc *crtc)
 	intel_fdi_normal_train(crtc);
 
 	/* For PCH DP, enable TRANS_DP_CTL */
-	if (HAS_PCH_CPT(dev) &&
-	    (intel_pipe_has_type(crtc, INTEL_OUTPUT_DISPLAYPORT) ||
-	     intel_pipe_has_type(crtc, INTEL_OUTPUT_EDP))) {
+	if (HAS_PCH_CPT(dev) && intel_crtc->config.has_dp_encoder) {
 		u32 bpc = (I915_READ(PIPECONF(pipe)) & PIPECONF_BPC_MASK) >> 5;
 		reg = TRANS_DP_CTL(pipe);
 		temp = I915_READ(reg);
@@ -3642,12 +3909,13 @@ void intel_put_shared_dpll(struct intel_crtc *crtc)
 	if (pll == NULL)
 		return;
 
-	if (pll->refcount == 0) {
-		WARN(1, "bad %s refcount\n", pll->name);
+	if (!(pll->config.crtc_mask & (1 << crtc->pipe))) {
+		WARN(1, "bad %s crtc mask\n", pll->name);
 		return;
 	}
 
-	if (--pll->refcount == 0) {
+	pll->config.crtc_mask &= ~(1 << crtc->pipe);
+	if (pll->config.crtc_mask == 0) {
 		WARN_ON(pll->on);
 		WARN_ON(pll->active);
 	}
@@ -3658,15 +3926,9 @@ void intel_put_shared_dpll(struct intel_crtc *crtc)
 struct intel_shared_dpll *intel_get_shared_dpll(struct intel_crtc *crtc)
 {
 	struct drm_i915_private *dev_priv = crtc->base.dev->dev_private;
-	struct intel_shared_dpll *pll = intel_crtc_to_shared_dpll(crtc);
+	struct intel_shared_dpll *pll;
 	enum intel_dpll_id i;
 
-	if (pll) {
-		DRM_DEBUG_KMS("CRTC:%d dropping existing %s\n",
-			      crtc->base.base.id, pll->name);
-		intel_put_shared_dpll(crtc);
-	}
-
 	if (HAS_PCH_IBX(dev_priv->dev)) {
 		/* Ironlake PCH has a fixed PLL->PCH pipe mapping. */
 		i = (enum intel_dpll_id) crtc->pipe;
@@ -3675,7 +3937,7 @@ struct intel_shared_dpll *intel_get_shared_dpll(struct intel_crtc *crtc)
 		DRM_DEBUG_KMS("CRTC:%d using pre-allocated %s\n",
 			      crtc->base.base.id, pll->name);
 
-		WARN_ON(pll->refcount);
+		WARN_ON(pll->new_config->crtc_mask);
 
 		goto found;
 	}
@@ -3684,15 +3946,16 @@ struct intel_shared_dpll *intel_get_shared_dpll(struct intel_crtc *crtc)
 		pll = &dev_priv->shared_dplls[i];
 
 		/* Only want to check enabled timings first */
-		if (pll->refcount == 0)
+		if (pll->new_config->crtc_mask == 0)
 			continue;
 
-		if (memcmp(&crtc->config.dpll_hw_state, &pll->hw_state,
-			   sizeof(pll->hw_state)) == 0) {
-			DRM_DEBUG_KMS("CRTC:%d sharing existing %s (refcount %d, ative %d)\n",
-				      crtc->base.base.id,
-				      pll->name, pll->refcount, pll->active);
-
+		if (memcmp(&crtc->new_config->dpll_hw_state,
+			   &pll->new_config->hw_state,
+			   sizeof(pll->new_config->hw_state)) == 0) {
+			DRM_DEBUG_KMS("CRTC:%d sharing existing %s (crtc mask 0x%08x, ative %d)\n",
+				      crtc->base.base.id, pll->name,
+				      pll->new_config->crtc_mask,
+				      pll->active);
 			goto found;
 		}
 	}
@@ -3700,7 +3963,7 @@ struct intel_shared_dpll *intel_get_shared_dpll(struct intel_crtc *crtc)
 	/* Ok no matching timings, maybe there's a free one? */
 	for (i = 0; i < dev_priv->num_shared_dpll; i++) {
 		pll = &dev_priv->shared_dplls[i];
-		if (pll->refcount == 0) {
+		if (pll->new_config->crtc_mask == 0) {
 			DRM_DEBUG_KMS("CRTC:%d allocated %s\n",
 				      crtc->base.base.id, pll->name);
 			goto found;
@@ -3710,18 +3973,86 @@ struct intel_shared_dpll *intel_get_shared_dpll(struct intel_crtc *crtc)
 	return NULL;
 
 found:
-	if (pll->refcount == 0)
-		pll->hw_state = crtc->config.dpll_hw_state;
+	if (pll->new_config->crtc_mask == 0)
+		pll->new_config->hw_state = crtc->new_config->dpll_hw_state;
 
-	crtc->config.shared_dpll = i;
+	crtc->new_config->shared_dpll = i;
 	DRM_DEBUG_DRIVER("using %s for pipe %c\n", pll->name,
 			 pipe_name(crtc->pipe));
 
-	pll->refcount++;
+	pll->new_config->crtc_mask |= 1 << crtc->pipe;
 
 	return pll;
 }
 
+/**
+ * intel_shared_dpll_start_config - start a new PLL staged config
+ * @dev_priv: DRM device
+ * @clear_pipes: mask of pipes that will have their PLLs freed
+ *
+ * Starts a new PLL staged config, copying the current config but
+ * releasing the references of pipes specified in clear_pipes.
+ */
+static int intel_shared_dpll_start_config(struct drm_i915_private *dev_priv,
+					  unsigned clear_pipes)
+{
+	struct intel_shared_dpll *pll;
+	enum intel_dpll_id i;
+
+	for (i = 0; i < dev_priv->num_shared_dpll; i++) {
+		pll = &dev_priv->shared_dplls[i];
+
+		pll->new_config = kmemdup(&pll->config, sizeof pll->config,
+					  GFP_KERNEL);
+		if (!pll->new_config)
+			goto cleanup;
+
+		pll->new_config->crtc_mask &= ~clear_pipes;
+	}
+
+	return 0;
+
+cleanup:
+	while (--i >= 0) {
+		pll = &dev_priv->shared_dplls[i];
+		kfree(pll->new_config);
+		pll->new_config = NULL;
+	}
+
+	return -ENOMEM;
+}
+
+static void intel_shared_dpll_commit(struct drm_i915_private *dev_priv)
+{
+	struct intel_shared_dpll *pll;
+	enum intel_dpll_id i;
+
+	for (i = 0; i < dev_priv->num_shared_dpll; i++) {
+		pll = &dev_priv->shared_dplls[i];
+
+		WARN_ON(pll->new_config == &pll->config);
+
+		pll->config = *pll->new_config;
+		kfree(pll->new_config);
+		pll->new_config = NULL;
+	}
+}
+
+static void intel_shared_dpll_abort_config(struct drm_i915_private *dev_priv)
+{
+	struct intel_shared_dpll *pll;
+	enum intel_dpll_id i;
+
+	for (i = 0; i < dev_priv->num_shared_dpll; i++) {
+		pll = &dev_priv->shared_dplls[i];
+
+		WARN_ON(pll->new_config == &pll->config);
+
+		kfree(pll->new_config);
+		pll->new_config = NULL;
+	}
+}
+
 static void cpt_verify_modeset(struct drm_device *dev, int pipe)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
@@ -3736,6 +4067,19 @@ static void cpt_verify_modeset(struct drm_device *dev, int pipe)
 	}
 }
 
+static void skylake_pfit_enable(struct intel_crtc *crtc)
+{
+	struct drm_device *dev = crtc->base.dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	int pipe = crtc->pipe;
+
+	if (crtc->config.pch_pfit.enabled) {
+		I915_WRITE(PS_CTL(pipe), PS_ENABLE);
+		I915_WRITE(PS_WIN_POS(pipe), crtc->config.pch_pfit.pos);
+		I915_WRITE(PS_WIN_SZ(pipe), crtc->config.pch_pfit.size);
+	}
+}
+
 static void ironlake_pfit_enable(struct intel_crtc *crtc)
 {
 	struct drm_device *dev = crtc->base.dev;
@@ -3859,7 +4203,7 @@ static void intel_crtc_load_lut(struct drm_crtc *crtc)
 		return;
 
 	if (!HAS_PCH_SPLIT(dev_priv->dev)) {
-		if (intel_pipe_has_type(crtc, INTEL_OUTPUT_DSI))
+		if (intel_pipe_has_type(intel_crtc, INTEL_OUTPUT_DSI))
 			assert_dsi_pll_enabled(dev_priv);
 		else
 			assert_pll_enabled(dev_priv, pipe);
@@ -3911,14 +4255,10 @@ static void intel_crtc_dpms_overlay(struct intel_crtc *intel_crtc, bool enable)
 static void intel_crtc_enable_planes(struct drm_crtc *crtc)
 {
 	struct drm_device *dev = crtc->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 	int pipe = intel_crtc->pipe;
-	int plane = intel_crtc->plane;
-
-	drm_vblank_on(dev, pipe);
 
-	intel_enable_primary_hw_plane(dev_priv, plane, pipe);
+	intel_enable_primary_hw_plane(crtc->primary, crtc);
 	intel_enable_planes(crtc);
 	intel_crtc_update_cursor(crtc, true);
 	intel_crtc_dpms_overlay(intel_crtc, true);
@@ -3955,7 +4295,7 @@ static void intel_crtc_disable_planes(struct drm_crtc *crtc)
 	intel_crtc_dpms_overlay(intel_crtc, false);
 	intel_crtc_update_cursor(crtc, false);
 	intel_disable_planes(crtc);
-	intel_disable_primary_hw_plane(dev_priv, plane, pipe);
+	intel_disable_primary_hw_plane(crtc->primary, crtc);
 
 	/*
 	 * FIXME: Once we grow proper nuclear flip support out of this we need
@@ -3963,8 +4303,6 @@ static void intel_crtc_disable_planes(struct drm_crtc *crtc)
 	 * consider this a flip to a NULL plane.
 	 */
 	intel_frontbuffer_flip(dev, INTEL_FRONTBUFFER_ALL_MASK(pipe));
-
-	drm_vblank_off(dev, pipe);
 }
 
 static void ironlake_crtc_enable(struct drm_crtc *crtc)
@@ -3974,7 +4312,6 @@ static void ironlake_crtc_enable(struct drm_crtc *crtc)
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 	struct intel_encoder *encoder;
 	int pipe = intel_crtc->pipe;
-	enum plane plane = intel_crtc->plane;
 
 	WARN_ON(!crtc->enabled);
 
@@ -3991,22 +4328,15 @@ static void ironlake_crtc_enable(struct drm_crtc *crtc)
 
 	if (intel_crtc->config.has_pch_encoder) {
 		intel_cpu_transcoder_set_m_n(intel_crtc,
-					     &intel_crtc->config.fdi_m_n);
+				     &intel_crtc->config.fdi_m_n, NULL);
 	}
 
 	ironlake_set_pipeconf(crtc);
 
-	/* Set up the display plane register */
-	I915_WRITE(DSPCNTR(plane), DISPPLANE_GAMMA_ENABLE);
-	POSTING_READ(DSPCNTR(plane));
-
-	dev_priv->display.update_primary_plane(crtc, crtc->primary->fb,
-					       crtc->x, crtc->y);
-
 	intel_crtc->active = true;
 
-	intel_set_cpu_fifo_underrun_reporting(dev, pipe, true);
-	intel_set_pch_fifo_underrun_reporting(dev, pipe, true);
+	intel_set_cpu_fifo_underrun_reporting(dev_priv, pipe, true);
+	intel_set_pch_fifo_underrun_reporting(dev_priv, pipe, true);
 
 	for_each_encoder_on_crtc(dev, crtc, encoder)
 		if (encoder->pre_enable)
@@ -4042,6 +4372,9 @@ static void ironlake_crtc_enable(struct drm_crtc *crtc)
 	if (HAS_PCH_CPT(dev))
 		cpt_verify_modeset(dev, intel_crtc->pipe);
 
+	assert_vblank_disabled(crtc);
+	drm_crtc_vblank_on(crtc);
+
 	intel_crtc_enable_planes(crtc);
 }
 
@@ -4087,7 +4420,6 @@ static void haswell_crtc_enable(struct drm_crtc *crtc)
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 	struct intel_encoder *encoder;
 	int pipe = intel_crtc->pipe;
-	enum plane plane = intel_crtc->plane;
 
 	WARN_ON(!crtc->enabled);
 
@@ -4102,37 +4434,39 @@ static void haswell_crtc_enable(struct drm_crtc *crtc)
 
 	intel_set_pipe_timings(intel_crtc);
 
+	if (intel_crtc->config.cpu_transcoder != TRANSCODER_EDP) {
+		I915_WRITE(PIPE_MULT(intel_crtc->config.cpu_transcoder),
+			   intel_crtc->config.pixel_multiplier - 1);
+	}
+
 	if (intel_crtc->config.has_pch_encoder) {
 		intel_cpu_transcoder_set_m_n(intel_crtc,
-					     &intel_crtc->config.fdi_m_n);
+				     &intel_crtc->config.fdi_m_n, NULL);
 	}
 
 	haswell_set_pipeconf(crtc);
 
 	intel_set_pipe_csc(crtc);
 
-	/* Set up the display plane register */
-	I915_WRITE(DSPCNTR(plane), DISPPLANE_GAMMA_ENABLE | DISPPLANE_PIPE_CSC_ENABLE);
-	POSTING_READ(DSPCNTR(plane));
-
-	dev_priv->display.update_primary_plane(crtc, crtc->primary->fb,
-					       crtc->x, crtc->y);
-
 	intel_crtc->active = true;
 
-	intel_set_cpu_fifo_underrun_reporting(dev, pipe, true);
+	intel_set_cpu_fifo_underrun_reporting(dev_priv, pipe, true);
 	for_each_encoder_on_crtc(dev, crtc, encoder)
 		if (encoder->pre_enable)
 			encoder->pre_enable(encoder);
 
 	if (intel_crtc->config.has_pch_encoder) {
-		intel_set_pch_fifo_underrun_reporting(dev, TRANSCODER_A, true);
+		intel_set_pch_fifo_underrun_reporting(dev_priv, TRANSCODER_A,
+						      true);
 		dev_priv->display.fdi_link_train(crtc);
 	}
 
 	intel_ddi_enable_pipe_clock(intel_crtc);
 
-	ironlake_pfit_enable(intel_crtc);
+	if (IS_SKYLAKE(dev))
+		skylake_pfit_enable(intel_crtc);
+	else
+		ironlake_pfit_enable(intel_crtc);
 
 	/*
 	 * On ILK+ LUT must be loaded before the pipe is running but with
@@ -4157,12 +4491,30 @@ static void haswell_crtc_enable(struct drm_crtc *crtc)
 		intel_opregion_notify_encoder(encoder, true);
 	}
 
+	assert_vblank_disabled(crtc);
+	drm_crtc_vblank_on(crtc);
+
 	/* If we change the relative order between pipe/planes enabling, we need
 	 * to change the workaround. */
 	haswell_mode_set_planes_workaround(intel_crtc);
 	intel_crtc_enable_planes(crtc);
 }
 
+static void skylake_pfit_disable(struct intel_crtc *crtc)
+{
+	struct drm_device *dev = crtc->base.dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	int pipe = crtc->pipe;
+
+	/* To avoid upsetting the power well on haswell only disable the pfit if
+	 * it's in use. The hw state code will make sure we get this right. */
+	if (crtc->config.pch_pfit.enabled) {
+		I915_WRITE(PS_CTL(pipe), 0);
+		I915_WRITE(PS_WIN_POS(pipe), 0);
+		I915_WRITE(PS_WIN_SZ(pipe), 0);
+	}
+}
+
 static void ironlake_pfit_disable(struct intel_crtc *crtc)
 {
 	struct drm_device *dev = crtc->base.dev;
@@ -4192,13 +4544,17 @@ static void ironlake_crtc_disable(struct drm_crtc *crtc)
 
 	intel_crtc_disable_planes(crtc);
 
+	drm_crtc_vblank_off(crtc);
+	assert_vblank_disabled(crtc);
+
 	for_each_encoder_on_crtc(dev, crtc, encoder)
 		encoder->disable(encoder);
 
 	if (intel_crtc->config.has_pch_encoder)
-		intel_set_pch_fifo_underrun_reporting(dev, pipe, false);
+		intel_set_pch_fifo_underrun_reporting(dev_priv, pipe, false);
+
+	intel_disable_pipe(intel_crtc);
 
-	intel_disable_pipe(dev_priv, pipe);
 	ironlake_pfit_disable(intel_crtc);
 
 	for_each_encoder_on_crtc(dev, crtc, encoder)
@@ -4209,7 +4565,6 @@ static void ironlake_crtc_disable(struct drm_crtc *crtc)
 		ironlake_fdi_disable(crtc);
 
 		ironlake_disable_pch_transcoder(dev_priv, pipe);
-		intel_set_pch_fifo_underrun_reporting(dev, pipe, true);
 
 		if (HAS_PCH_CPT(dev)) {
 			/* disable TRANS_DP_CTL */
@@ -4246,7 +4601,6 @@ static void haswell_crtc_disable(struct drm_crtc *crtc)
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 	struct intel_encoder *encoder;
-	int pipe = intel_crtc->pipe;
 	enum transcoder cpu_transcoder = intel_crtc->config.cpu_transcoder;
 
 	if (!intel_crtc->active)
@@ -4254,27 +4608,33 @@ static void haswell_crtc_disable(struct drm_crtc *crtc)
 
 	intel_crtc_disable_planes(crtc);
 
+	drm_crtc_vblank_off(crtc);
+	assert_vblank_disabled(crtc);
+
 	for_each_encoder_on_crtc(dev, crtc, encoder) {
 		intel_opregion_notify_encoder(encoder, false);
 		encoder->disable(encoder);
 	}
 
 	if (intel_crtc->config.has_pch_encoder)
-		intel_set_pch_fifo_underrun_reporting(dev, TRANSCODER_A, false);
-	intel_disable_pipe(dev_priv, pipe);
+		intel_set_pch_fifo_underrun_reporting(dev_priv, TRANSCODER_A,
+						      false);
+	intel_disable_pipe(intel_crtc);
 
 	if (intel_crtc->config.dp_encoder_is_mst)
 		intel_ddi_set_vc_payload_alloc(crtc, false);
 
 	intel_ddi_disable_transcoder_func(dev_priv, cpu_transcoder);
 
-	ironlake_pfit_disable(intel_crtc);
+	if (IS_SKYLAKE(dev))
+		skylake_pfit_disable(intel_crtc);
+	else
+		ironlake_pfit_disable(intel_crtc);
 
 	intel_ddi_disable_pipe_clock(intel_crtc);
 
 	if (intel_crtc->config.has_pch_encoder) {
 		lpt_disable_pch_transcoder(dev_priv);
-		intel_set_pch_fifo_underrun_reporting(dev, TRANSCODER_A, true);
 		intel_ddi_fdi_disable(crtc);
 	}
 
@@ -4395,20 +4755,6 @@ static unsigned long get_crtc_power_domains(struct drm_crtc *crtc)
 	return mask;
 }
 
-void intel_display_set_init_power(struct drm_i915_private *dev_priv,
-				  bool enable)
-{
-	if (dev_priv->power_domains.init_power_on == enable)
-		return;
-
-	if (enable)
-		intel_display_power_get(dev_priv, POWER_DOMAIN_INIT);
-	else
-		intel_display_power_put(dev_priv, POWER_DOMAIN_INIT);
-
-	dev_priv->power_domains.init_power_on = enable;
-}
-
 static void modeset_update_crtc_power_domains(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
@@ -4431,6 +4777,9 @@ static void modeset_update_crtc_power_domains(struct drm_device *dev)
 			intel_display_power_get(dev_priv, domain);
 	}
 
+	if (dev_priv->display.modeset_global_resources)
+		dev_priv->display.modeset_global_resources(dev);
+
 	for_each_intel_crtc(dev, crtc) {
 		enum intel_display_power_domain domain;
 
@@ -4462,7 +4811,7 @@ static void vlv_update_cdclk(struct drm_device *dev)
 	struct drm_i915_private *dev_priv = dev->dev_private;
 
 	dev_priv->vlv_cdclk_freq = dev_priv->display.get_display_clock_speed(dev);
-	DRM_DEBUG_DRIVER("Current CD clock rate: %d kHz",
+	DRM_DEBUG_DRIVER("Current CD clock rate: %d kHz\n",
 			 dev_priv->vlv_cdclk_freq);
 
 	/*
@@ -4470,7 +4819,7 @@ static void vlv_update_cdclk(struct drm_device *dev)
 	 * BSpec erroneously claims we should aim for 4MHz, but
 	 * in fact 1MHz is the correct frequency.
 	 */
-	I915_WRITE(GMBUSFREQ_VLV, dev_priv->vlv_cdclk_freq);
+	I915_WRITE(GMBUSFREQ_VLV, DIV_ROUND_UP(dev_priv->vlv_cdclk_freq, 1000));
 }
 
 /* Adjust CDclk dividers to allow high res or save power if possible */
@@ -4501,10 +4850,9 @@ static void valleyview_set_cdclk(struct drm_device *dev, int cdclk)
 	mutex_unlock(&dev_priv->rps.hw_lock);
 
 	if (cdclk == 400000) {
-		u32 divider, vco;
+		u32 divider;
 
-		vco = valleyview_get_vco(dev_priv);
-		divider = DIV_ROUND_CLOSEST(vco << 1, cdclk) - 1;
+		divider = DIV_ROUND_CLOSEST(dev_priv->hpll_freq << 1, cdclk) - 1;
 
 		mutex_lock(&dev_priv->dpio_lock);
 		/* adjust cdclk divider */
@@ -4539,11 +4887,55 @@ static void valleyview_set_cdclk(struct drm_device *dev, int cdclk)
 	vlv_update_cdclk(dev);
 }
 
+static void cherryview_set_cdclk(struct drm_device *dev, int cdclk)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	u32 val, cmd;
+
+	WARN_ON(dev_priv->display.get_display_clock_speed(dev) != dev_priv->vlv_cdclk_freq);
+
+	switch (cdclk) {
+	case 400000:
+		cmd = 3;
+		break;
+	case 333333:
+	case 320000:
+		cmd = 2;
+		break;
+	case 266667:
+		cmd = 1;
+		break;
+	case 200000:
+		cmd = 0;
+		break;
+	default:
+		WARN_ON(1);
+		return;
+	}
+
+	mutex_lock(&dev_priv->rps.hw_lock);
+	val = vlv_punit_read(dev_priv, PUNIT_REG_DSPFREQ);
+	val &= ~DSPFREQGUAR_MASK_CHV;
+	val |= (cmd << DSPFREQGUAR_SHIFT_CHV);
+	vlv_punit_write(dev_priv, PUNIT_REG_DSPFREQ, val);
+	if (wait_for((vlv_punit_read(dev_priv, PUNIT_REG_DSPFREQ) &
+		      DSPFREQSTAT_MASK_CHV) == (cmd << DSPFREQSTAT_SHIFT_CHV),
+		     50)) {
+		DRM_ERROR("timed out waiting for CDclk change\n");
+	}
+	mutex_unlock(&dev_priv->rps.hw_lock);
+
+	vlv_update_cdclk(dev);
+}
+
 static int valleyview_calc_cdclk(struct drm_i915_private *dev_priv,
 				 int max_pixclk)
 {
-	int vco = valleyview_get_vco(dev_priv);
-	int freq_320 = (vco <<  1) % 320000 != 0 ? 333333 : 320000;
+	int freq_320 = (dev_priv->hpll_freq <<  1) % 320000 != 0 ? 333333 : 320000;
+
+	/* FIXME: Punit isn't quite ready yet */
+	if (IS_CHERRYVIEW(dev_priv->dev))
+		return 400000;
 
 	/*
 	 * Really only a few cases to deal with, as only 4 CDclks are supported:
@@ -4607,59 +4999,67 @@ static void valleyview_modeset_global_resources(struct drm_device *dev)
 	int max_pixclk = intel_mode_max_pixclk(dev_priv);
 	int req_cdclk = valleyview_calc_cdclk(dev_priv, max_pixclk);
 
-	if (req_cdclk != dev_priv->vlv_cdclk_freq)
-		valleyview_set_cdclk(dev, req_cdclk);
-	modeset_update_crtc_power_domains(dev);
+	if (req_cdclk != dev_priv->vlv_cdclk_freq) {
+		/*
+		 * FIXME: We can end up here with all power domains off, yet
+		 * with a CDCLK frequency other than the minimum. To account
+		 * for this take the PIPE-A power domain, which covers the HW
+		 * blocks needed for the following programming. This can be
+		 * removed once it's guaranteed that we get here either with
+		 * the minimum CDCLK set, or the required power domains
+		 * enabled.
+		 */
+		intel_display_power_get(dev_priv, POWER_DOMAIN_PIPE_A);
+
+		if (IS_CHERRYVIEW(dev))
+			cherryview_set_cdclk(dev, req_cdclk);
+		else
+			valleyview_set_cdclk(dev, req_cdclk);
+
+		intel_display_power_put(dev_priv, POWER_DOMAIN_PIPE_A);
+	}
 }
 
 static void valleyview_crtc_enable(struct drm_crtc *crtc)
 {
 	struct drm_device *dev = crtc->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 	struct intel_encoder *encoder;
 	int pipe = intel_crtc->pipe;
-	int plane = intel_crtc->plane;
 	bool is_dsi;
-	u32 dspcntr;
 
 	WARN_ON(!crtc->enabled);
 
 	if (intel_crtc->active)
 		return;
 
-	is_dsi = intel_pipe_has_type(crtc, INTEL_OUTPUT_DSI);
-
-	if (!is_dsi && !IS_CHERRYVIEW(dev))
-		vlv_prepare_pll(intel_crtc);
+	is_dsi = intel_pipe_has_type(intel_crtc, INTEL_OUTPUT_DSI);
 
-	/* Set up the display plane register */
-	dspcntr = DISPPLANE_GAMMA_ENABLE;
+	if (!is_dsi) {
+		if (IS_CHERRYVIEW(dev))
+			chv_prepare_pll(intel_crtc, &intel_crtc->config);
+		else
+			vlv_prepare_pll(intel_crtc, &intel_crtc->config);
+	}
 
 	if (intel_crtc->config.has_dp_encoder)
 		intel_dp_set_m_n(intel_crtc);
 
 	intel_set_pipe_timings(intel_crtc);
 
-	/* pipesrc and dspsize control the size that is scaled from,
-	 * which should always be the user's requested size.
-	 */
-	I915_WRITE(DSPSIZE(plane),
-		   ((intel_crtc->config.pipe_src_h - 1) << 16) |
-		   (intel_crtc->config.pipe_src_w - 1));
-	I915_WRITE(DSPPOS(plane), 0);
-
-	i9xx_set_pipeconf(intel_crtc);
+	if (IS_CHERRYVIEW(dev) && pipe == PIPE_B) {
+		struct drm_i915_private *dev_priv = dev->dev_private;
 
-	I915_WRITE(DSPCNTR(plane), dspcntr);
-	POSTING_READ(DSPCNTR(plane));
+		I915_WRITE(CHV_BLEND(pipe), CHV_BLEND_LEGACY);
+		I915_WRITE(CHV_CANVAS(pipe), 0);
+	}
 
-	dev_priv->display.update_primary_plane(crtc, crtc->primary->fb,
-					       crtc->x, crtc->y);
+	i9xx_set_pipeconf(intel_crtc);
 
 	intel_crtc->active = true;
 
-	intel_set_cpu_fifo_underrun_reporting(dev, pipe, true);
+	intel_set_cpu_fifo_underrun_reporting(dev_priv, pipe, true);
 
 	for_each_encoder_on_crtc(dev, crtc, encoder)
 		if (encoder->pre_pll_enable)
@@ -4667,9 +5067,9 @@ static void valleyview_crtc_enable(struct drm_crtc *crtc)
 
 	if (!is_dsi) {
 		if (IS_CHERRYVIEW(dev))
-			chv_enable_pll(intel_crtc);
+			chv_enable_pll(intel_crtc, &intel_crtc->config);
 		else
-			vlv_enable_pll(intel_crtc);
+			vlv_enable_pll(intel_crtc, &intel_crtc->config);
 	}
 
 	for_each_encoder_on_crtc(dev, crtc, encoder)
@@ -4686,10 +5086,13 @@ static void valleyview_crtc_enable(struct drm_crtc *crtc)
 	for_each_encoder_on_crtc(dev, crtc, encoder)
 		encoder->enable(encoder);
 
+	assert_vblank_disabled(crtc);
+	drm_crtc_vblank_on(crtc);
+
 	intel_crtc_enable_planes(crtc);
 
 	/* Underruns don't raise interrupts, so check manually. */
-	i9xx_check_fifo_underruns(dev);
+	i9xx_check_fifo_underruns(dev_priv);
 }
 
 static void i9xx_set_pll_dividers(struct intel_crtc *crtc)
@@ -4704,12 +5107,10 @@ static void i9xx_set_pll_dividers(struct intel_crtc *crtc)
 static void i9xx_crtc_enable(struct drm_crtc *crtc)
 {
 	struct drm_device *dev = crtc->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 	struct intel_encoder *encoder;
 	int pipe = intel_crtc->pipe;
-	int plane = intel_crtc->plane;
-	u32 dspcntr;
 
 	WARN_ON(!crtc->enabled);
 
@@ -4718,39 +5119,17 @@ static void i9xx_crtc_enable(struct drm_crtc *crtc)
 
 	i9xx_set_pll_dividers(intel_crtc);
 
-	/* Set up the display plane register */
-	dspcntr = DISPPLANE_GAMMA_ENABLE;
-
-	if (pipe == 0)
-		dspcntr &= ~DISPPLANE_SEL_PIPE_MASK;
-	else
-		dspcntr |= DISPPLANE_SEL_PIPE_B;
-
 	if (intel_crtc->config.has_dp_encoder)
 		intel_dp_set_m_n(intel_crtc);
 
 	intel_set_pipe_timings(intel_crtc);
 
-	/* pipesrc and dspsize control the size that is scaled from,
-	 * which should always be the user's requested size.
-	 */
-	I915_WRITE(DSPSIZE(plane),
-		   ((intel_crtc->config.pipe_src_h - 1) << 16) |
-		   (intel_crtc->config.pipe_src_w - 1));
-	I915_WRITE(DSPPOS(plane), 0);
-
 	i9xx_set_pipeconf(intel_crtc);
 
-	I915_WRITE(DSPCNTR(plane), dspcntr);
-	POSTING_READ(DSPCNTR(plane));
-
-	dev_priv->display.update_primary_plane(crtc, crtc->primary->fb,
-					       crtc->x, crtc->y);
-
 	intel_crtc->active = true;
 
 	if (!IS_GEN2(dev))
-		intel_set_cpu_fifo_underrun_reporting(dev, pipe, true);
+		intel_set_cpu_fifo_underrun_reporting(dev_priv, pipe, true);
 
 	for_each_encoder_on_crtc(dev, crtc, encoder)
 		if (encoder->pre_enable)
@@ -4768,6 +5147,9 @@ static void i9xx_crtc_enable(struct drm_crtc *crtc)
 	for_each_encoder_on_crtc(dev, crtc, encoder)
 		encoder->enable(encoder);
 
+	assert_vblank_disabled(crtc);
+	drm_crtc_vblank_on(crtc);
+
 	intel_crtc_enable_planes(crtc);
 
 	/*
@@ -4778,10 +5160,10 @@ static void i9xx_crtc_enable(struct drm_crtc *crtc)
 	 * but leave the pipe running.
 	 */
 	if (IS_GEN2(dev))
-		intel_set_cpu_fifo_underrun_reporting(dev, pipe, true);
+		intel_set_cpu_fifo_underrun_reporting(dev_priv, pipe, true);
 
 	/* Underruns don't raise interrupts, so check manually. */
-	i9xx_check_fifo_underruns(dev);
+	i9xx_check_fifo_underruns(dev_priv);
 }
 
 static void i9xx_pfit_disable(struct intel_crtc *crtc)
@@ -4817,7 +5199,7 @@ static void i9xx_crtc_disable(struct drm_crtc *crtc)
 	 * but leave the pipe running.
 	 */
 	if (IS_GEN2(dev))
-		intel_set_cpu_fifo_underrun_reporting(dev, pipe, false);
+		intel_set_cpu_fifo_underrun_reporting(dev_priv, pipe, false);
 
 	/*
 	 * Vblank time updates from the shadow to live plane control register
@@ -4831,9 +5213,6 @@ static void i9xx_crtc_disable(struct drm_crtc *crtc)
 	intel_set_memory_cxsr(dev_priv, false);
 	intel_crtc_disable_planes(crtc);
 
-	for_each_encoder_on_crtc(dev, crtc, encoder)
-		encoder->disable(encoder);
-
 	/*
 	 * On gen2 planes are double buffered but the pipe isn't, so we must
 	 * wait for planes to fully turn off before disabling the pipe.
@@ -4842,7 +5221,13 @@ static void i9xx_crtc_disable(struct drm_crtc *crtc)
 	 */
 	intel_wait_for_vblank(dev, pipe);
 
-	intel_disable_pipe(dev_priv, pipe);
+	drm_crtc_vblank_off(crtc);
+	assert_vblank_disabled(crtc);
+
+	for_each_encoder_on_crtc(dev, crtc, encoder)
+		encoder->disable(encoder);
+
+	intel_disable_pipe(intel_crtc);
 
 	i9xx_pfit_disable(intel_crtc);
 
@@ -4850,17 +5235,17 @@ static void i9xx_crtc_disable(struct drm_crtc *crtc)
 		if (encoder->post_disable)
 			encoder->post_disable(encoder);
 
-	if (!intel_pipe_has_type(crtc, INTEL_OUTPUT_DSI)) {
+	if (!intel_pipe_has_type(intel_crtc, INTEL_OUTPUT_DSI)) {
 		if (IS_CHERRYVIEW(dev))
 			chv_disable_pll(dev_priv, pipe);
 		else if (IS_VALLEYVIEW(dev))
 			vlv_disable_pll(dev_priv, pipe);
 		else
-			i9xx_disable_pll(dev_priv, pipe);
+			i9xx_disable_pll(intel_crtc);
 	}
 
 	if (!IS_GEN2(dev))
-		intel_set_cpu_fifo_underrun_reporting(dev, pipe, false);
+		intel_set_cpu_fifo_underrun_reporting(dev_priv, pipe, false);
 
 	intel_crtc->active = false;
 	intel_update_watermarks(crtc);
@@ -4874,36 +5259,6 @@ static void i9xx_crtc_off(struct drm_crtc *crtc)
 {
 }
 
-static void intel_crtc_update_sarea(struct drm_crtc *crtc,
-				    bool enabled)
-{
-	struct drm_device *dev = crtc->dev;
-	struct drm_i915_master_private *master_priv;
-	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
-	int pipe = intel_crtc->pipe;
-
-	if (!dev->primary->master)
-		return;
-
-	master_priv = dev->primary->master->driver_priv;
-	if (!master_priv->sarea_priv)
-		return;
-
-	switch (pipe) {
-	case 0:
-		master_priv->sarea_priv->pipeA_w = enabled ? crtc->mode.hdisplay : 0;
-		master_priv->sarea_priv->pipeA_h = enabled ? crtc->mode.vdisplay : 0;
-		break;
-	case 1:
-		master_priv->sarea_priv->pipeB_w = enabled ? crtc->mode.hdisplay : 0;
-		master_priv->sarea_priv->pipeB_h = enabled ? crtc->mode.vdisplay : 0;
-		break;
-	default:
-		DRM_ERROR("Can't update pipe %c in SAREA\n", pipe_name(pipe));
-		break;
-	}
-}
-
 /* Master function to enable/disable CRTC and corresponding power wells */
 void intel_crtc_control(struct drm_crtc *crtc, bool enable)
 {
@@ -4947,8 +5302,6 @@ void intel_crtc_update_dpms(struct drm_crtc *crtc)
 		enable |= intel_encoder->connectors_active;
 
 	intel_crtc_control(crtc, enable);
-
-	intel_crtc_update_sarea(crtc, enable);
 }
 
 static void intel_crtc_disable(struct drm_crtc *crtc)
@@ -4963,7 +5316,6 @@ static void intel_crtc_disable(struct drm_crtc *crtc)
 	WARN_ON(!crtc->enabled);
 
 	dev_priv->display.crtc_disable(crtc);
-	intel_crtc_update_sarea(crtc, false);
 	dev_priv->display.off(crtc);
 
 	if (crtc->primary->fb) {
@@ -5202,11 +5554,11 @@ static int intel_crtc_compute_config(struct intel_crtc *crtc,
 				     struct intel_crtc_config *pipe_config)
 {
 	struct drm_device *dev = crtc->base.dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct drm_display_mode *adjusted_mode = &pipe_config->adjusted_mode;
 
 	/* FIXME should check pixel clock limits on all platforms */
 	if (INTEL_INFO(dev)->gen < 4) {
-		struct drm_i915_private *dev_priv = dev->dev_private;
 		int clock_limit =
 			dev_priv->display.get_display_clock_speed(dev);
 
@@ -5233,7 +5585,7 @@ static int intel_crtc_compute_config(struct intel_crtc *crtc,
 	 * - LVDS dual channel mode
 	 * - Double wide pipe
 	 */
-	if ((intel_pipe_has_type(&crtc->base, INTEL_OUTPUT_LVDS) &&
+	if ((intel_pipe_has_type(crtc, INTEL_OUTPUT_LVDS) &&
 	     intel_is_dual_link_lvds(dev)) || pipe_config->double_wide)
 		pipe_config->pipe_src_w &= ~1;
 
@@ -5255,13 +5607,6 @@ static int intel_crtc_compute_config(struct intel_crtc *crtc,
 	if (HAS_IPS(dev))
 		hsw_compute_ips_config(crtc, pipe_config);
 
-	/*
-	 * XXX: PCH/WRPLL clock sharing is done in ->mode_set, so make sure the
-	 * old clock survives for now.
-	 */
-	if (HAS_PCH_IBX(dev) || HAS_PCH_CPT(dev) || HAS_DDI(dev))
-		pipe_config->shared_dpll = crtc->config.shared_dpll;
-
 	if (pipe_config->has_pch_encoder)
 		return ironlake_fdi_compute_config(crtc, pipe_config);
 
@@ -5271,10 +5616,16 @@ static int intel_crtc_compute_config(struct intel_crtc *crtc,
 static int valleyview_get_display_clock_speed(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	int vco = valleyview_get_vco(dev_priv);
 	u32 val;
 	int divider;
 
+	/* FIXME: Punit isn't quite ready yet */
+	if (IS_CHERRYVIEW(dev))
+		return 400000;
+
+	if (dev_priv->hpll_freq == 0)
+		dev_priv->hpll_freq = valleyview_get_vco(dev_priv);
+
 	mutex_lock(&dev_priv->dpio_lock);
 	val = vlv_cck_read(dev_priv, CCK_DISPLAY_CLOCK_CONTROL);
 	mutex_unlock(&dev_priv->dpio_lock);
@@ -5285,7 +5636,7 @@ static int valleyview_get_display_clock_speed(struct drm_device *dev)
 	     (divider << DISPLAY_FREQUENCY_STATUS_SHIFT),
 	     "cdclk change in progress\n");
 
-	return DIV_ROUND_CLOSEST(vco << 1, divider + 1);
+	return DIV_ROUND_CLOSEST(dev_priv->hpll_freq << 1, divider + 1);
 }
 
 static int i945_get_display_clock_speed(struct drm_device *dev)
@@ -5417,15 +5768,15 @@ static inline bool intel_panel_use_ssc(struct drm_i915_private *dev_priv)
 		&& !(dev_priv->quirks & QUIRK_LVDS_SSC_DISABLE);
 }
 
-static int i9xx_get_refclk(struct drm_crtc *crtc, int num_connectors)
+static int i9xx_get_refclk(struct intel_crtc *crtc, int num_connectors)
 {
-	struct drm_device *dev = crtc->dev;
+	struct drm_device *dev = crtc->base.dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	int refclk;
 
 	if (IS_VALLEYVIEW(dev)) {
 		refclk = 100000;
-	} else if (intel_pipe_has_type(crtc, INTEL_OUTPUT_LVDS) &&
+	} else if (intel_pipe_will_have_type(crtc, INTEL_OUTPUT_LVDS) &&
 	    intel_panel_use_ssc(dev_priv) && num_connectors < 2) {
 		refclk = dev_priv->vbt.lvds_ssc_freq;
 		DRM_DEBUG_KMS("using SSC reference clock of %d kHz\n", refclk);
@@ -5455,24 +5806,24 @@ static void i9xx_update_pll_dividers(struct intel_crtc *crtc,
 	u32 fp, fp2 = 0;
 
 	if (IS_PINEVIEW(dev)) {
-		fp = pnv_dpll_compute_fp(&crtc->config.dpll);
+		fp = pnv_dpll_compute_fp(&crtc->new_config->dpll);
 		if (reduced_clock)
 			fp2 = pnv_dpll_compute_fp(reduced_clock);
 	} else {
-		fp = i9xx_dpll_compute_fp(&crtc->config.dpll);
+		fp = i9xx_dpll_compute_fp(&crtc->new_config->dpll);
 		if (reduced_clock)
 			fp2 = i9xx_dpll_compute_fp(reduced_clock);
 	}
 
-	crtc->config.dpll_hw_state.fp0 = fp;
+	crtc->new_config->dpll_hw_state.fp0 = fp;
 
 	crtc->lowfreq_avail = false;
-	if (intel_pipe_has_type(&crtc->base, INTEL_OUTPUT_LVDS) &&
+	if (intel_pipe_will_have_type(crtc, INTEL_OUTPUT_LVDS) &&
 	    reduced_clock && i915.powersave) {
-		crtc->config.dpll_hw_state.fp1 = fp2;
+		crtc->new_config->dpll_hw_state.fp1 = fp2;
 		crtc->lowfreq_avail = true;
 	} else {
-		crtc->config.dpll_hw_state.fp1 = fp;
+		crtc->new_config->dpll_hw_state.fp1 = fp;
 	}
 }
 
@@ -5519,7 +5870,8 @@ static void intel_pch_transcoder_set_m_n(struct intel_crtc *crtc,
 }
 
 static void intel_cpu_transcoder_set_m_n(struct intel_crtc *crtc,
-					 struct intel_link_m_n *m_n)
+					 struct intel_link_m_n *m_n,
+					 struct intel_link_m_n *m2_n2)
 {
 	struct drm_device *dev = crtc->base.dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
@@ -5531,6 +5883,18 @@ static void intel_cpu_transcoder_set_m_n(struct intel_crtc *crtc,
 		I915_WRITE(PIPE_DATA_N1(transcoder), m_n->gmch_n);
 		I915_WRITE(PIPE_LINK_M1(transcoder), m_n->link_m);
 		I915_WRITE(PIPE_LINK_N1(transcoder), m_n->link_n);
+		/* M2_N2 registers to be set only for gen < 8 (M2_N2 available
+		 * for gen < 8) and if DRRS is supported (to make sure the
+		 * registers are not unnecessarily accessed).
+		 */
+		if (m2_n2 && INTEL_INFO(dev)->gen < 8 &&
+			crtc->config.has_drrs) {
+			I915_WRITE(PIPE_DATA_M2(transcoder),
+					TU_SIZE(m2_n2->tu) | m2_n2->gmch_m);
+			I915_WRITE(PIPE_DATA_N2(transcoder), m2_n2->gmch_n);
+			I915_WRITE(PIPE_LINK_M2(transcoder), m2_n2->link_m);
+			I915_WRITE(PIPE_LINK_N2(transcoder), m2_n2->link_n);
+		}
 	} else {
 		I915_WRITE(PIPE_DATA_M_G4X(pipe), TU_SIZE(m_n->tu) | m_n->gmch_m);
 		I915_WRITE(PIPE_DATA_N_G4X(pipe), m_n->gmch_n);
@@ -5539,15 +5903,17 @@ static void intel_cpu_transcoder_set_m_n(struct intel_crtc *crtc,
 	}
 }
 
-static void intel_dp_set_m_n(struct intel_crtc *crtc)
+void intel_dp_set_m_n(struct intel_crtc *crtc)
 {
 	if (crtc->config.has_pch_encoder)
 		intel_pch_transcoder_set_m_n(crtc, &crtc->config.dp_m_n);
 	else
-		intel_cpu_transcoder_set_m_n(crtc, &crtc->config.dp_m_n);
+		intel_cpu_transcoder_set_m_n(crtc, &crtc->config.dp_m_n,
+						   &crtc->config.dp_m2_n2);
 }
 
-static void vlv_update_pll(struct intel_crtc *crtc)
+static void vlv_update_pll(struct intel_crtc *crtc,
+			   struct intel_crtc_config *pipe_config)
 {
 	u32 dpll, dpll_md;
 
@@ -5562,14 +5928,15 @@ static void vlv_update_pll(struct intel_crtc *crtc)
 	if (crtc->pipe == PIPE_B)
 		dpll |= DPLL_INTEGRATED_CRI_CLK_VLV;
 	dpll |= DPLL_VCO_ENABLE;
-	crtc->config.dpll_hw_state.dpll = dpll;
+	pipe_config->dpll_hw_state.dpll = dpll;
 
-	dpll_md = (crtc->config.pixel_multiplier - 1)
+	dpll_md = (pipe_config->pixel_multiplier - 1)
 		<< DPLL_MD_UDI_MULTIPLIER_SHIFT;
-	crtc->config.dpll_hw_state.dpll_md = dpll_md;
+	pipe_config->dpll_hw_state.dpll_md = dpll_md;
 }
 
-static void vlv_prepare_pll(struct intel_crtc *crtc)
+static void vlv_prepare_pll(struct intel_crtc *crtc,
+			    const struct intel_crtc_config *pipe_config)
 {
 	struct drm_device *dev = crtc->base.dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
@@ -5580,11 +5947,11 @@ static void vlv_prepare_pll(struct intel_crtc *crtc)
 
 	mutex_lock(&dev_priv->dpio_lock);
 
-	bestn = crtc->config.dpll.n;
-	bestm1 = crtc->config.dpll.m1;
-	bestm2 = crtc->config.dpll.m2;
-	bestp1 = crtc->config.dpll.p1;
-	bestp2 = crtc->config.dpll.p2;
+	bestn = pipe_config->dpll.n;
+	bestm1 = pipe_config->dpll.m1;
+	bestm2 = pipe_config->dpll.m2;
+	bestp1 = pipe_config->dpll.p1;
+	bestp2 = pipe_config->dpll.p2;
 
 	/* See eDP HDMI DPIO driver vbios notes doc */
 
@@ -5621,17 +5988,16 @@ static void vlv_prepare_pll(struct intel_crtc *crtc)
 	vlv_dpio_write(dev_priv, pipe, VLV_PLL_DW3(pipe), mdiv);
 
 	/* Set HBR and RBR LPF coefficients */
-	if (crtc->config.port_clock == 162000 ||
-	    intel_pipe_has_type(&crtc->base, INTEL_OUTPUT_ANALOG) ||
-	    intel_pipe_has_type(&crtc->base, INTEL_OUTPUT_HDMI))
+	if (pipe_config->port_clock == 162000 ||
+	    intel_pipe_has_type(crtc, INTEL_OUTPUT_ANALOG) ||
+	    intel_pipe_has_type(crtc, INTEL_OUTPUT_HDMI))
 		vlv_dpio_write(dev_priv, pipe, VLV_PLL_DW10(pipe),
 				 0x009f0003);
 	else
 		vlv_dpio_write(dev_priv, pipe, VLV_PLL_DW10(pipe),
 				 0x00d0000f);
 
-	if (intel_pipe_has_type(&crtc->base, INTEL_OUTPUT_EDP) ||
-	    intel_pipe_has_type(&crtc->base, INTEL_OUTPUT_DISPLAYPORT)) {
+	if (crtc->config.has_dp_encoder) {
 		/* Use SSC source */
 		if (pipe == PIPE_A)
 			vlv_dpio_write(dev_priv, pipe, VLV_PLL_DW5(pipe),
@@ -5651,8 +6017,8 @@ static void vlv_prepare_pll(struct intel_crtc *crtc)
 
 	coreclk = vlv_dpio_read(dev_priv, pipe, VLV_PLL_DW7(pipe));
 	coreclk = (coreclk & 0x0000ff00) | 0x01c00000;
-	if (intel_pipe_has_type(&crtc->base, INTEL_OUTPUT_DISPLAYPORT) ||
-	    intel_pipe_has_type(&crtc->base, INTEL_OUTPUT_EDP))
+	if (intel_pipe_has_type(crtc, INTEL_OUTPUT_DISPLAYPORT) ||
+	    intel_pipe_has_type(crtc, INTEL_OUTPUT_EDP))
 		coreclk |= 0x01000000;
 	vlv_dpio_write(dev_priv, pipe, VLV_PLL_DW7(pipe), coreclk);
 
@@ -5660,7 +6026,21 @@ static void vlv_prepare_pll(struct intel_crtc *crtc)
 	mutex_unlock(&dev_priv->dpio_lock);
 }
 
-static void chv_update_pll(struct intel_crtc *crtc)
+static void chv_update_pll(struct intel_crtc *crtc,
+			   struct intel_crtc_config *pipe_config)
+{
+	pipe_config->dpll_hw_state.dpll = DPLL_SSC_REF_CLOCK_CHV |
+		DPLL_REFA_CLK_ENABLE_VLV | DPLL_VGA_MODE_DIS |
+		DPLL_VCO_ENABLE;
+	if (crtc->pipe != PIPE_A)
+		pipe_config->dpll_hw_state.dpll |= DPLL_INTEGRATED_CRI_CLK_VLV;
+
+	pipe_config->dpll_hw_state.dpll_md =
+		(pipe_config->pixel_multiplier - 1) << DPLL_MD_UDI_MULTIPLIER_SHIFT;
+}
+
+static void chv_prepare_pll(struct intel_crtc *crtc,
+			    const struct intel_crtc_config *pipe_config)
 {
 	struct drm_device *dev = crtc->base.dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
@@ -5671,27 +6051,18 @@ static void chv_update_pll(struct intel_crtc *crtc)
 	u32 bestn, bestm1, bestm2, bestp1, bestp2, bestm2_frac;
 	int refclk;
 
-	crtc->config.dpll_hw_state.dpll = DPLL_SSC_REF_CLOCK_CHV |
-		DPLL_REFA_CLK_ENABLE_VLV | DPLL_VGA_MODE_DIS |
-		DPLL_VCO_ENABLE;
-	if (pipe != PIPE_A)
-		crtc->config.dpll_hw_state.dpll |= DPLL_INTEGRATED_CRI_CLK_VLV;
-
-	crtc->config.dpll_hw_state.dpll_md =
-		(crtc->config.pixel_multiplier - 1) << DPLL_MD_UDI_MULTIPLIER_SHIFT;
-
-	bestn = crtc->config.dpll.n;
-	bestm2_frac = crtc->config.dpll.m2 & 0x3fffff;
-	bestm1 = crtc->config.dpll.m1;
-	bestm2 = crtc->config.dpll.m2 >> 22;
-	bestp1 = crtc->config.dpll.p1;
-	bestp2 = crtc->config.dpll.p2;
+	bestn = pipe_config->dpll.n;
+	bestm2_frac = pipe_config->dpll.m2 & 0x3fffff;
+	bestm1 = pipe_config->dpll.m1;
+	bestm2 = pipe_config->dpll.m2 >> 22;
+	bestp1 = pipe_config->dpll.p1;
+	bestp2 = pipe_config->dpll.p2;
 
 	/*
 	 * Enable Refclk and SSC
 	 */
 	I915_WRITE(dpll_reg,
-		   crtc->config.dpll_hw_state.dpll & ~DPLL_VCO_ENABLE);
+		   pipe_config->dpll_hw_state.dpll & ~DPLL_VCO_ENABLE);
 
 	mutex_lock(&dev_priv->dpio_lock);
 
@@ -5719,7 +6090,7 @@ static void chv_update_pll(struct intel_crtc *crtc)
 		       (2 << DPIO_CHV_FEEDFWD_GAIN_SHIFT));
 
 	/* Loop filter */
-	refclk = i9xx_get_refclk(&crtc->base, 0);
+	refclk = i9xx_get_refclk(crtc, 0);
 	loopfilter = 5 << DPIO_CHV_PROP_COEFF_SHIFT |
 		2 << DPIO_CHV_GAIN_CTRL_SHIFT;
 	if (refclk == 100000)
@@ -5739,6 +6110,53 @@ static void chv_update_pll(struct intel_crtc *crtc)
 	mutex_unlock(&dev_priv->dpio_lock);
 }
 
+/**
+ * vlv_force_pll_on - forcibly enable just the PLL
+ * @dev_priv: i915 private structure
+ * @pipe: pipe PLL to enable
+ * @dpll: PLL configuration
+ *
+ * Enable the PLL for @pipe using the supplied @dpll config. To be used
+ * in cases where we need the PLL enabled even when @pipe is not going to
+ * be enabled.
+ */
+void vlv_force_pll_on(struct drm_device *dev, enum pipe pipe,
+		      const struct dpll *dpll)
+{
+	struct intel_crtc *crtc =
+		to_intel_crtc(intel_get_crtc_for_pipe(dev, pipe));
+	struct intel_crtc_config pipe_config = {
+		.pixel_multiplier = 1,
+		.dpll = *dpll,
+	};
+
+	if (IS_CHERRYVIEW(dev)) {
+		chv_update_pll(crtc, &pipe_config);
+		chv_prepare_pll(crtc, &pipe_config);
+		chv_enable_pll(crtc, &pipe_config);
+	} else {
+		vlv_update_pll(crtc, &pipe_config);
+		vlv_prepare_pll(crtc, &pipe_config);
+		vlv_enable_pll(crtc, &pipe_config);
+	}
+}
+
+/**
+ * vlv_force_pll_off - forcibly disable just the PLL
+ * @dev_priv: i915 private structure
+ * @pipe: pipe PLL to disable
+ *
+ * Disable the PLL for @pipe. To be used in cases where we need
+ * the PLL enabled even when @pipe is not going to be enabled.
+ */
+void vlv_force_pll_off(struct drm_device *dev, enum pipe pipe)
+{
+	if (IS_CHERRYVIEW(dev))
+		chv_disable_pll(to_i915(dev), pipe);
+	else
+		vlv_disable_pll(to_i915(dev), pipe);
+}
+
 static void i9xx_update_pll(struct intel_crtc *crtc,
 			    intel_clock_t *reduced_clock,
 			    int num_connectors)
@@ -5747,29 +6165,29 @@ static void i9xx_update_pll(struct intel_crtc *crtc,
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	u32 dpll;
 	bool is_sdvo;
-	struct dpll *clock = &crtc->config.dpll;
+	struct dpll *clock = &crtc->new_config->dpll;
 
 	i9xx_update_pll_dividers(crtc, reduced_clock);
 
-	is_sdvo = intel_pipe_has_type(&crtc->base, INTEL_OUTPUT_SDVO) ||
-		intel_pipe_has_type(&crtc->base, INTEL_OUTPUT_HDMI);
+	is_sdvo = intel_pipe_will_have_type(crtc, INTEL_OUTPUT_SDVO) ||
+		intel_pipe_will_have_type(crtc, INTEL_OUTPUT_HDMI);
 
 	dpll = DPLL_VGA_MODE_DIS;
 
-	if (intel_pipe_has_type(&crtc->base, INTEL_OUTPUT_LVDS))
+	if (intel_pipe_will_have_type(crtc, INTEL_OUTPUT_LVDS))
 		dpll |= DPLLB_MODE_LVDS;
 	else
 		dpll |= DPLLB_MODE_DAC_SERIAL;
 
 	if (IS_I945G(dev) || IS_I945GM(dev) || IS_G33(dev)) {
-		dpll |= (crtc->config.pixel_multiplier - 1)
+		dpll |= (crtc->new_config->pixel_multiplier - 1)
 			<< SDVO_MULTIPLIER_SHIFT_HIRES;
 	}
 
 	if (is_sdvo)
 		dpll |= DPLL_SDVO_HIGH_SPEED;
 
-	if (intel_pipe_has_type(&crtc->base, INTEL_OUTPUT_DISPLAYPORT))
+	if (crtc->new_config->has_dp_encoder)
 		dpll |= DPLL_SDVO_HIGH_SPEED;
 
 	/* compute bitmask from p1 value */
@@ -5797,21 +6215,21 @@ static void i9xx_update_pll(struct intel_crtc *crtc,
 	if (INTEL_INFO(dev)->gen >= 4)
 		dpll |= (6 << PLL_LOAD_PULSE_PHASE_SHIFT);
 
-	if (crtc->config.sdvo_tv_clock)
+	if (crtc->new_config->sdvo_tv_clock)
 		dpll |= PLL_REF_INPUT_TVCLKINBC;
-	else if (intel_pipe_has_type(&crtc->base, INTEL_OUTPUT_LVDS) &&
+	else if (intel_pipe_will_have_type(crtc, INTEL_OUTPUT_LVDS) &&
 		 intel_panel_use_ssc(dev_priv) && num_connectors < 2)
 		dpll |= PLLB_REF_INPUT_SPREADSPECTRUMIN;
 	else
 		dpll |= PLL_REF_INPUT_DREFCLK;
 
 	dpll |= DPLL_VCO_ENABLE;
-	crtc->config.dpll_hw_state.dpll = dpll;
+	crtc->new_config->dpll_hw_state.dpll = dpll;
 
 	if (INTEL_INFO(dev)->gen >= 4) {
-		u32 dpll_md = (crtc->config.pixel_multiplier - 1)
+		u32 dpll_md = (crtc->new_config->pixel_multiplier - 1)
 			<< DPLL_MD_UDI_MULTIPLIER_SHIFT;
-		crtc->config.dpll_hw_state.dpll_md = dpll_md;
+		crtc->new_config->dpll_hw_state.dpll_md = dpll_md;
 	}
 }
 
@@ -5822,13 +6240,13 @@ static void i8xx_update_pll(struct intel_crtc *crtc,
 	struct drm_device *dev = crtc->base.dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	u32 dpll;
-	struct dpll *clock = &crtc->config.dpll;
+	struct dpll *clock = &crtc->new_config->dpll;
 
 	i9xx_update_pll_dividers(crtc, reduced_clock);
 
 	dpll = DPLL_VGA_MODE_DIS;
 
-	if (intel_pipe_has_type(&crtc->base, INTEL_OUTPUT_LVDS)) {
+	if (intel_pipe_will_have_type(crtc, INTEL_OUTPUT_LVDS)) {
 		dpll |= (1 << (clock->p1 - 1)) << DPLL_FPA01_P1_POST_DIV_SHIFT;
 	} else {
 		if (clock->p1 == 2)
@@ -5839,17 +6257,17 @@ static void i8xx_update_pll(struct intel_crtc *crtc,
 			dpll |= PLL_P2_DIVIDE_BY_4;
 	}
 
-	if (intel_pipe_has_type(&crtc->base, INTEL_OUTPUT_DVO))
+	if (!IS_I830(dev) && intel_pipe_will_have_type(crtc, INTEL_OUTPUT_DVO))
 		dpll |= DPLL_DVO_2X_MODE;
 
-	if (intel_pipe_has_type(&crtc->base, INTEL_OUTPUT_LVDS) &&
+	if (intel_pipe_will_have_type(crtc, INTEL_OUTPUT_LVDS) &&
 		 intel_panel_use_ssc(dev_priv) && num_connectors < 2)
 		dpll |= PLLB_REF_INPUT_SPREADSPECTRUMIN;
 	else
 		dpll |= PLL_REF_INPUT_DREFCLK;
 
 	dpll |= DPLL_VCO_ENABLE;
-	crtc->config.dpll_hw_state.dpll = dpll;
+	crtc->new_config->dpll_hw_state.dpll = dpll;
 }
 
 static void intel_set_pipe_timings(struct intel_crtc *intel_crtc)
@@ -5873,7 +6291,7 @@ static void intel_set_pipe_timings(struct intel_crtc *intel_crtc)
 		crtc_vtotal -= 1;
 		crtc_vblank_end -= 1;
 
-		if (intel_pipe_has_type(&intel_crtc->base, INTEL_OUTPUT_SDVO))
+		if (intel_pipe_has_type(intel_crtc, INTEL_OUTPUT_SDVO))
 			vsyncshift = (adjusted_mode->crtc_htotal - 1) / 2;
 		else
 			vsyncshift = adjusted_mode->crtc_hsync_start -
@@ -5990,9 +6408,9 @@ static void i9xx_set_pipeconf(struct intel_crtc *intel_crtc)
 
 	pipeconf = 0;
 
-	if (dev_priv->quirks & QUIRK_PIPEA_FORCE &&
-	    I915_READ(PIPECONF(intel_crtc->pipe)) & PIPECONF_ENABLE)
-		pipeconf |= PIPECONF_ENABLE;
+	if ((intel_crtc->pipe == PIPE_A && dev_priv->quirks & QUIRK_PIPEA_FORCE) ||
+	    (intel_crtc->pipe == PIPE_B && dev_priv->quirks & QUIRK_PIPEB_FORCE))
+		pipeconf |= I915_READ(PIPECONF(intel_crtc->pipe)) & PIPECONF_ENABLE;
 
 	if (intel_crtc->config.double_wide)
 		pipeconf |= PIPECONF_DOUBLE_WIDE;
@@ -6031,7 +6449,7 @@ static void i9xx_set_pipeconf(struct intel_crtc *intel_crtc)
 
 	if (intel_crtc->config.adjusted_mode.flags & DRM_MODE_FLAG_INTERLACE) {
 		if (INTEL_INFO(dev)->gen < 4 ||
-		    intel_pipe_has_type(&intel_crtc->base, INTEL_OUTPUT_SDVO))
+		    intel_pipe_has_type(intel_crtc, INTEL_OUTPUT_SDVO))
 			pipeconf |= PIPECONF_INTERLACE_W_FIELD_INDICATION;
 		else
 			pipeconf |= PIPECONF_INTERLACE_W_SYNC_SHIFT;
@@ -6045,13 +6463,10 @@ static void i9xx_set_pipeconf(struct intel_crtc *intel_crtc)
 	POSTING_READ(PIPECONF(intel_crtc->pipe));
 }
 
-static int i9xx_crtc_mode_set(struct drm_crtc *crtc,
-			      int x, int y,
-			      struct drm_framebuffer *fb)
+static int i9xx_crtc_compute_clock(struct intel_crtc *crtc)
 {
-	struct drm_device *dev = crtc->dev;
+	struct drm_device *dev = crtc->base.dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 	int refclk, num_connectors = 0;
 	intel_clock_t clock, reduced_clock;
 	bool ok, has_reduced_clock = false;
@@ -6059,7 +6474,10 @@ static int i9xx_crtc_mode_set(struct drm_crtc *crtc,
 	struct intel_encoder *encoder;
 	const intel_limit_t *limit;
 
-	for_each_encoder_on_crtc(dev, crtc, encoder) {
+	for_each_intel_encoder(dev, encoder) {
+		if (encoder->new_crtc != crtc)
+			continue;
+
 		switch (encoder->type) {
 		case INTEL_OUTPUT_LVDS:
 			is_lvds = true;
@@ -6067,6 +6485,8 @@ static int i9xx_crtc_mode_set(struct drm_crtc *crtc,
 		case INTEL_OUTPUT_DSI:
 			is_dsi = true;
 			break;
+		default:
+			break;
 		}
 
 		num_connectors++;
@@ -6075,7 +6495,7 @@ static int i9xx_crtc_mode_set(struct drm_crtc *crtc,
 	if (is_dsi)
 		return 0;
 
-	if (!intel_crtc->config.clock_set) {
+	if (!crtc->new_config->clock_set) {
 		refclk = i9xx_get_refclk(crtc, num_connectors);
 
 		/*
@@ -6086,7 +6506,7 @@ static int i9xx_crtc_mode_set(struct drm_crtc *crtc,
 		 */
 		limit = intel_limit(crtc, refclk);
 		ok = dev_priv->display.find_dpll(limit, crtc,
-						 intel_crtc->config.port_clock,
+						 crtc->new_config->port_clock,
 						 refclk, NULL, &clock);
 		if (!ok) {
 			DRM_ERROR("Couldn't find PLL settings for mode!\n");
@@ -6107,23 +6527,23 @@ static int i9xx_crtc_mode_set(struct drm_crtc *crtc,
 							    &reduced_clock);
 		}
 		/* Compat-code for transition, will disappear. */
-		intel_crtc->config.dpll.n = clock.n;
-		intel_crtc->config.dpll.m1 = clock.m1;
-		intel_crtc->config.dpll.m2 = clock.m2;
-		intel_crtc->config.dpll.p1 = clock.p1;
-		intel_crtc->config.dpll.p2 = clock.p2;
+		crtc->new_config->dpll.n = clock.n;
+		crtc->new_config->dpll.m1 = clock.m1;
+		crtc->new_config->dpll.m2 = clock.m2;
+		crtc->new_config->dpll.p1 = clock.p1;
+		crtc->new_config->dpll.p2 = clock.p2;
 	}
 
 	if (IS_GEN2(dev)) {
-		i8xx_update_pll(intel_crtc,
+		i8xx_update_pll(crtc,
 				has_reduced_clock ? &reduced_clock : NULL,
 				num_connectors);
 	} else if (IS_CHERRYVIEW(dev)) {
-		chv_update_pll(intel_crtc);
+		chv_update_pll(crtc, crtc->new_config);
 	} else if (IS_VALLEYVIEW(dev)) {
-		vlv_update_pll(intel_crtc);
+		vlv_update_pll(crtc, crtc->new_config);
 	} else {
-		i9xx_update_pll(intel_crtc,
+		i9xx_update_pll(crtc,
 				has_reduced_clock ? &reduced_clock : NULL,
 				num_connectors);
 	}
@@ -6235,7 +6655,7 @@ static void i9xx_get_plane_config(struct intel_crtc *crtc,
 	crtc->base.primary->fb->height = ((val >> 0) & 0xfff) + 1;
 
 	val = I915_READ(DSPSTRIDE(pipe));
-	crtc->base.primary->fb->pitches[0] = val & 0xffffff80;
+	crtc->base.primary->fb->pitches[0] = val & 0xffffffc0;
 
 	aligned_height = intel_align_height(dev, crtc->base.primary->fb->height,
 					    plane_config->tiled);
@@ -6289,8 +6709,8 @@ static bool i9xx_get_pipe_config(struct intel_crtc *crtc,
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	uint32_t tmp;
 
-	if (!intel_display_power_enabled(dev_priv,
-					 POWER_DOMAIN_PIPE(crtc->pipe)))
+	if (!intel_display_power_is_enabled(dev_priv,
+					    POWER_DOMAIN_PIPE(crtc->pipe)))
 		return false;
 
 	pipe_config->cpu_transcoder = (enum transcoder) crtc->pipe;
@@ -6345,6 +6765,14 @@ static bool i9xx_get_pipe_config(struct intel_crtc *crtc,
 	}
 	pipe_config->dpll_hw_state.dpll = I915_READ(DPLL(crtc->pipe));
 	if (!IS_VALLEYVIEW(dev)) {
+		/*
+		 * DPLL_DVO_2X_MODE must be enabled for both DPLLs
+		 * on 830. Filter it out here so that we don't
+		 * report errors due to that.
+		 */
+		if (IS_I830(dev))
+			pipe_config->dpll_hw_state.dpll &= ~DPLL_DVO_2X_MODE;
+
 		pipe_config->dpll_hw_state.fp0 = I915_READ(FP0(crtc->pipe));
 		pipe_config->dpll_hw_state.fp1 = I915_READ(FP1(crtc->pipe));
 	} else {
@@ -6367,7 +6795,6 @@ static bool i9xx_get_pipe_config(struct intel_crtc *crtc,
 static void ironlake_init_pch_refclk(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct drm_mode_config *mode_config = &dev->mode_config;
 	struct intel_encoder *encoder;
 	u32 val, final;
 	bool has_lvds = false;
@@ -6377,8 +6804,7 @@ static void ironlake_init_pch_refclk(struct drm_device *dev)
 	bool can_ssc = false;
 
 	/* We need to take the global config into account */
-	list_for_each_entry(encoder, &mode_config->encoder_list,
-			    base.head) {
+	for_each_intel_encoder(dev, encoder) {
 		switch (encoder->type) {
 		case INTEL_OUTPUT_LVDS:
 			has_panel = true;
@@ -6389,6 +6815,8 @@ static void ironlake_init_pch_refclk(struct drm_device *dev)
 			if (enc_to_dig_port(&encoder->base)->port == PORT_A)
 				has_cpu_edp = true;
 			break;
+		default:
+			break;
 		}
 	}
 
@@ -6685,15 +7113,16 @@ static void lpt_disable_clkout_dp(struct drm_device *dev)
 
 static void lpt_init_pch_refclk(struct drm_device *dev)
 {
-	struct drm_mode_config *mode_config = &dev->mode_config;
 	struct intel_encoder *encoder;
 	bool has_vga = false;
 
-	list_for_each_entry(encoder, &mode_config->encoder_list, base.head) {
+	for_each_intel_encoder(dev, encoder) {
 		switch (encoder->type) {
 		case INTEL_OUTPUT_ANALOG:
 			has_vga = true;
 			break;
+		default:
+			break;
 		}
 	}
 
@@ -6722,11 +7151,16 @@ static int ironlake_get_refclk(struct drm_crtc *crtc)
 	int num_connectors = 0;
 	bool is_lvds = false;
 
-	for_each_encoder_on_crtc(dev, crtc, encoder) {
+	for_each_intel_encoder(dev, encoder) {
+		if (encoder->new_crtc != to_intel_crtc(crtc))
+			continue;
+
 		switch (encoder->type) {
 		case INTEL_OUTPUT_LVDS:
 			is_lvds = true;
 			break;
+		default:
+			break;
 		}
 		num_connectors++;
 	}
@@ -6871,7 +7305,7 @@ static void haswell_set_pipeconf(struct drm_crtc *crtc)
 	I915_WRITE(GAMMA_MODE(intel_crtc->pipe), GAMMA_MODE_MODE_8BIT);
 	POSTING_READ(GAMMA_MODE(intel_crtc->pipe));
 
-	if (IS_BROADWELL(dev)) {
+	if (IS_BROADWELL(dev) || INTEL_INFO(dev)->gen >= 9) {
 		val = 0;
 
 		switch (intel_crtc->config.pipe_bpp) {
@@ -6906,18 +7340,12 @@ static bool ironlake_compute_clocks(struct drm_crtc *crtc,
 {
 	struct drm_device *dev = crtc->dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct intel_encoder *intel_encoder;
+	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 	int refclk;
 	const intel_limit_t *limit;
 	bool ret, is_lvds = false;
 
-	for_each_encoder_on_crtc(dev, crtc, intel_encoder) {
-		switch (intel_encoder->type) {
-		case INTEL_OUTPUT_LVDS:
-			is_lvds = true;
-			break;
-		}
-	}
+	is_lvds = intel_pipe_will_have_type(intel_crtc, INTEL_OUTPUT_LVDS);
 
 	refclk = ironlake_get_refclk(crtc);
 
@@ -6926,9 +7354,9 @@ static bool ironlake_compute_clocks(struct drm_crtc *crtc,
 	 * refclk, or FALSE.  The returned values represent the clock equation:
 	 * reflck * (5 * (m1 + 2) + (m2 + 2)) / (n + 2) / p1 / p2.
 	 */
-	limit = intel_limit(crtc, refclk);
-	ret = dev_priv->display.find_dpll(limit, crtc,
-					  to_intel_crtc(crtc)->config.port_clock,
+	limit = intel_limit(intel_crtc, refclk);
+	ret = dev_priv->display.find_dpll(limit, intel_crtc,
+					  intel_crtc->new_config->port_clock,
 					  refclk, NULL, clock);
 	if (!ret)
 		return false;
@@ -6941,7 +7369,7 @@ static bool ironlake_compute_clocks(struct drm_crtc *crtc,
 		 * downclock feature.
 		*/
 		*has_reduced_clock =
-			dev_priv->display.find_dpll(limit, crtc,
+			dev_priv->display.find_dpll(limit, intel_crtc,
 						    dev_priv->lvds_downclock,
 						    refclk, clock,
 						    reduced_clock);
@@ -6978,7 +7406,10 @@ static uint32_t ironlake_compute_dpll(struct intel_crtc *intel_crtc,
 	int factor, num_connectors = 0;
 	bool is_lvds = false, is_sdvo = false;
 
-	for_each_encoder_on_crtc(dev, crtc, intel_encoder) {
+	for_each_intel_encoder(dev, intel_encoder) {
+		if (intel_encoder->new_crtc != to_intel_crtc(crtc))
+			continue;
+
 		switch (intel_encoder->type) {
 		case INTEL_OUTPUT_LVDS:
 			is_lvds = true;
@@ -6987,6 +7418,8 @@ static uint32_t ironlake_compute_dpll(struct intel_crtc *intel_crtc,
 		case INTEL_OUTPUT_HDMI:
 			is_sdvo = true;
 			break;
+		default:
+			break;
 		}
 
 		num_connectors++;
@@ -6999,10 +7432,10 @@ static uint32_t ironlake_compute_dpll(struct intel_crtc *intel_crtc,
 		     dev_priv->vbt.lvds_ssc_freq == 100000) ||
 		    (HAS_PCH_IBX(dev) && intel_is_dual_link_lvds(dev)))
 			factor = 25;
-	} else if (intel_crtc->config.sdvo_tv_clock)
+	} else if (intel_crtc->new_config->sdvo_tv_clock)
 		factor = 20;
 
-	if (ironlake_needs_fb_cb_tune(&intel_crtc->config.dpll, factor))
+	if (ironlake_needs_fb_cb_tune(&intel_crtc->new_config->dpll, factor))
 		*fp |= FP_CB_TUNE;
 
 	if (fp2 && (reduced_clock->m < factor * reduced_clock->n))
@@ -7015,20 +7448,20 @@ static uint32_t ironlake_compute_dpll(struct intel_crtc *intel_crtc,
 	else
 		dpll |= DPLLB_MODE_DAC_SERIAL;
 
-	dpll |= (intel_crtc->config.pixel_multiplier - 1)
+	dpll |= (intel_crtc->new_config->pixel_multiplier - 1)
 		<< PLL_REF_SDVO_HDMI_MULTIPLIER_SHIFT;
 
 	if (is_sdvo)
 		dpll |= DPLL_SDVO_HIGH_SPEED;
-	if (intel_crtc->config.has_dp_encoder)
+	if (intel_crtc->new_config->has_dp_encoder)
 		dpll |= DPLL_SDVO_HIGH_SPEED;
 
 	/* compute bitmask from p1 value */
-	dpll |= (1 << (intel_crtc->config.dpll.p1 - 1)) << DPLL_FPA01_P1_POST_DIV_SHIFT;
+	dpll |= (1 << (intel_crtc->new_config->dpll.p1 - 1)) << DPLL_FPA01_P1_POST_DIV_SHIFT;
 	/* also FPA1 */
-	dpll |= (1 << (intel_crtc->config.dpll.p1 - 1)) << DPLL_FPA1_P1_POST_DIV_SHIFT;
+	dpll |= (1 << (intel_crtc->new_config->dpll.p1 - 1)) << DPLL_FPA1_P1_POST_DIV_SHIFT;
 
-	switch (intel_crtc->config.dpll.p2) {
+	switch (intel_crtc->new_config->dpll.p2) {
 	case 5:
 		dpll |= DPLL_DAC_SERIAL_P2_CLOCK_DIV_5;
 		break;
@@ -7051,78 +7484,64 @@ static uint32_t ironlake_compute_dpll(struct intel_crtc *intel_crtc,
 	return dpll | DPLL_VCO_ENABLE;
 }
 
-static int ironlake_crtc_mode_set(struct drm_crtc *crtc,
-				  int x, int y,
-				  struct drm_framebuffer *fb)
+static int ironlake_crtc_compute_clock(struct intel_crtc *crtc)
 {
-	struct drm_device *dev = crtc->dev;
-	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
-	int num_connectors = 0;
+	struct drm_device *dev = crtc->base.dev;
 	intel_clock_t clock, reduced_clock;
 	u32 dpll = 0, fp = 0, fp2 = 0;
 	bool ok, has_reduced_clock = false;
 	bool is_lvds = false;
-	struct intel_encoder *encoder;
 	struct intel_shared_dpll *pll;
 
-	for_each_encoder_on_crtc(dev, crtc, encoder) {
-		switch (encoder->type) {
-		case INTEL_OUTPUT_LVDS:
-			is_lvds = true;
-			break;
-		}
-
-		num_connectors++;
-	}
+	is_lvds = intel_pipe_has_type(crtc, INTEL_OUTPUT_LVDS);
 
 	WARN(!(HAS_PCH_IBX(dev) || HAS_PCH_CPT(dev)),
 	     "Unexpected PCH type %d\n", INTEL_PCH_TYPE(dev));
 
-	ok = ironlake_compute_clocks(crtc, &clock,
+	ok = ironlake_compute_clocks(&crtc->base, &clock,
 				     &has_reduced_clock, &reduced_clock);
-	if (!ok && !intel_crtc->config.clock_set) {
+	if (!ok && !crtc->new_config->clock_set) {
 		DRM_ERROR("Couldn't find PLL settings for mode!\n");
 		return -EINVAL;
 	}
 	/* Compat-code for transition, will disappear. */
-	if (!intel_crtc->config.clock_set) {
-		intel_crtc->config.dpll.n = clock.n;
-		intel_crtc->config.dpll.m1 = clock.m1;
-		intel_crtc->config.dpll.m2 = clock.m2;
-		intel_crtc->config.dpll.p1 = clock.p1;
-		intel_crtc->config.dpll.p2 = clock.p2;
+	if (!crtc->new_config->clock_set) {
+		crtc->new_config->dpll.n = clock.n;
+		crtc->new_config->dpll.m1 = clock.m1;
+		crtc->new_config->dpll.m2 = clock.m2;
+		crtc->new_config->dpll.p1 = clock.p1;
+		crtc->new_config->dpll.p2 = clock.p2;
 	}
 
 	/* CPU eDP is the only output that doesn't need a PCH PLL of its own. */
-	if (intel_crtc->config.has_pch_encoder) {
-		fp = i9xx_dpll_compute_fp(&intel_crtc->config.dpll);
+	if (crtc->new_config->has_pch_encoder) {
+		fp = i9xx_dpll_compute_fp(&crtc->new_config->dpll);
 		if (has_reduced_clock)
 			fp2 = i9xx_dpll_compute_fp(&reduced_clock);
 
-		dpll = ironlake_compute_dpll(intel_crtc,
+		dpll = ironlake_compute_dpll(crtc,
 					     &fp, &reduced_clock,
 					     has_reduced_clock ? &fp2 : NULL);
 
-		intel_crtc->config.dpll_hw_state.dpll = dpll;
-		intel_crtc->config.dpll_hw_state.fp0 = fp;
+		crtc->new_config->dpll_hw_state.dpll = dpll;
+		crtc->new_config->dpll_hw_state.fp0 = fp;
 		if (has_reduced_clock)
-			intel_crtc->config.dpll_hw_state.fp1 = fp2;
+			crtc->new_config->dpll_hw_state.fp1 = fp2;
 		else
-			intel_crtc->config.dpll_hw_state.fp1 = fp;
+			crtc->new_config->dpll_hw_state.fp1 = fp;
 
-		pll = intel_get_shared_dpll(intel_crtc);
+		pll = intel_get_shared_dpll(crtc);
 		if (pll == NULL) {
 			DRM_DEBUG_DRIVER("failed to find PLL for pipe %c\n",
-					 pipe_name(intel_crtc->pipe));
+					 pipe_name(crtc->pipe));
 			return -EINVAL;
 		}
-	} else
-		intel_put_shared_dpll(intel_crtc);
+	}
 
 	if (is_lvds && has_reduced_clock && i915.powersave)
-		intel_crtc->lowfreq_avail = true;
+		crtc->lowfreq_avail = true;
 	else
-		intel_crtc->lowfreq_avail = false;
+		crtc->lowfreq_avail = false;
 
 	return 0;
 }
@@ -7145,7 +7564,8 @@ static void intel_pch_transcoder_get_m_n(struct intel_crtc *crtc,
 
 static void intel_cpu_transcoder_get_m_n(struct intel_crtc *crtc,
 					 enum transcoder transcoder,
-					 struct intel_link_m_n *m_n)
+					 struct intel_link_m_n *m_n,
+					 struct intel_link_m_n *m2_n2)
 {
 	struct drm_device *dev = crtc->base.dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
@@ -7159,6 +7579,20 @@ static void intel_cpu_transcoder_get_m_n(struct intel_crtc *crtc,
 		m_n->gmch_n = I915_READ(PIPE_DATA_N1(transcoder));
 		m_n->tu = ((I915_READ(PIPE_DATA_M1(transcoder))
 			    & TU_SIZE_MASK) >> TU_SIZE_SHIFT) + 1;
+		/* Read M2_N2 registers only for gen < 8 (M2_N2 available for
+		 * gen < 8) and if DRRS is supported (to make sure the
+		 * registers are not unnecessarily read).
+		 */
+		if (m2_n2 && INTEL_INFO(dev)->gen < 8 &&
+			crtc->config.has_drrs) {
+			m2_n2->link_m = I915_READ(PIPE_LINK_M2(transcoder));
+			m2_n2->link_n =	I915_READ(PIPE_LINK_N2(transcoder));
+			m2_n2->gmch_m =	I915_READ(PIPE_DATA_M2(transcoder))
+					& ~TU_SIZE_MASK;
+			m2_n2->gmch_n =	I915_READ(PIPE_DATA_N2(transcoder));
+			m2_n2->tu = ((I915_READ(PIPE_DATA_M2(transcoder))
+					& TU_SIZE_MASK) >> TU_SIZE_SHIFT) + 1;
+		}
 	} else {
 		m_n->link_m = I915_READ(PIPE_LINK_M_G4X(pipe));
 		m_n->link_n = I915_READ(PIPE_LINK_N_G4X(pipe));
@@ -7177,14 +7611,31 @@ void intel_dp_get_m_n(struct intel_crtc *crtc,
 		intel_pch_transcoder_get_m_n(crtc, &pipe_config->dp_m_n);
 	else
 		intel_cpu_transcoder_get_m_n(crtc, pipe_config->cpu_transcoder,
-					     &pipe_config->dp_m_n);
+					     &pipe_config->dp_m_n,
+					     &pipe_config->dp_m2_n2);
 }
 
 static void ironlake_get_fdi_m_n_config(struct intel_crtc *crtc,
 					struct intel_crtc_config *pipe_config)
 {
 	intel_cpu_transcoder_get_m_n(crtc, pipe_config->cpu_transcoder,
-				     &pipe_config->fdi_m_n);
+				     &pipe_config->fdi_m_n, NULL);
+}
+
+static void skylake_get_pfit_config(struct intel_crtc *crtc,
+				    struct intel_crtc_config *pipe_config)
+{
+	struct drm_device *dev = crtc->base.dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	uint32_t tmp;
+
+	tmp = I915_READ(PS_CTL(crtc->pipe));
+
+	if (tmp & PS_ENABLE) {
+		pipe_config->pch_pfit.enabled = true;
+		pipe_config->pch_pfit.pos = I915_READ(PS_WIN_POS(crtc->pipe));
+		pipe_config->pch_pfit.size = I915_READ(PS_WIN_SZ(crtc->pipe));
+	}
 }
 
 static void ironlake_get_pfit_config(struct intel_crtc *crtc,
@@ -7255,7 +7706,7 @@ static void ironlake_get_plane_config(struct intel_crtc *crtc,
 	crtc->base.primary->fb->height = ((val >> 0) & 0xfff) + 1;
 
 	val = I915_READ(DSPSTRIDE(pipe));
-	crtc->base.primary->fb->pitches[0] = val & 0xffffff80;
+	crtc->base.primary->fb->pitches[0] = val & 0xffffffc0;
 
 	aligned_height = intel_align_height(dev, crtc->base.primary->fb->height,
 					    plane_config->tiled);
@@ -7278,8 +7729,8 @@ static bool ironlake_get_pipe_config(struct intel_crtc *crtc,
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	uint32_t tmp;
 
-	if (!intel_display_power_enabled(dev_priv,
-					 POWER_DOMAIN_PIPE(crtc->pipe)))
+	if (!intel_display_power_is_enabled(dev_priv,
+					    POWER_DOMAIN_PIPE(crtc->pipe)))
 		return false;
 
 	pipe_config->cpu_transcoder = (enum transcoder) crtc->pipe;
@@ -7472,7 +7923,6 @@ static void hsw_disable_lcpll(struct drm_i915_private *dev_priv,
 static void hsw_restore_lcpll(struct drm_i915_private *dev_priv)
 {
 	uint32_t val;
-	unsigned long irqflags;
 
 	val = I915_READ(LCPLL_CTL);
 
@@ -7492,10 +7942,10 @@ static void hsw_restore_lcpll(struct drm_i915_private *dev_priv)
 	 * to call special forcewake code that doesn't touch runtime PM and
 	 * doesn't enable the forcewake delayed work.
 	 */
-	spin_lock_irqsave(&dev_priv->uncore.lock, irqflags);
+	spin_lock_irq(&dev_priv->uncore.lock);
 	if (dev_priv->uncore.forcewake_count++ == 0)
 		dev_priv->uncore.funcs.force_wake_get(dev_priv, FORCEWAKE_ALL);
-	spin_unlock_irqrestore(&dev_priv->uncore.lock, irqflags);
+	spin_unlock_irq(&dev_priv->uncore.lock);
 
 	if (val & LCPLL_POWER_DOWN_ALLOW) {
 		val &= ~LCPLL_POWER_DOWN_ALLOW;
@@ -7526,10 +7976,10 @@ static void hsw_restore_lcpll(struct drm_i915_private *dev_priv)
 	}
 
 	/* See the big comment above. */
-	spin_lock_irqsave(&dev_priv->uncore.lock, irqflags);
+	spin_lock_irq(&dev_priv->uncore.lock);
 	if (--dev_priv->uncore.forcewake_count == 0)
 		dev_priv->uncore.funcs.force_wake_put(dev_priv, FORCEWAKE_ALL);
-	spin_unlock_irqrestore(&dev_priv->uncore.lock, irqflags);
+	spin_unlock_irq(&dev_priv->uncore.lock);
 }
 
 /*
@@ -7591,28 +8041,52 @@ void hsw_disable_pc8(struct drm_i915_private *dev_priv)
 	intel_prepare_ddi(dev);
 }
 
-static void snb_modeset_global_resources(struct drm_device *dev)
+static int haswell_crtc_compute_clock(struct intel_crtc *crtc)
 {
-	modeset_update_crtc_power_domains(dev);
-}
+	if (!intel_ddi_pll_select(crtc))
+		return -EINVAL;
 
-static void haswell_modeset_global_resources(struct drm_device *dev)
-{
-	modeset_update_crtc_power_domains(dev);
+	crtc->lowfreq_avail = false;
+
+	return 0;
 }
 
-static int haswell_crtc_mode_set(struct drm_crtc *crtc,
-				 int x, int y,
-				 struct drm_framebuffer *fb)
+static void skylake_get_ddi_pll(struct drm_i915_private *dev_priv,
+				enum port port,
+				struct intel_crtc_config *pipe_config)
 {
-	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
+	u32 temp;
 
-	if (!intel_ddi_pll_select(intel_crtc))
-		return -EINVAL;
+	temp = I915_READ(DPLL_CTRL2) & DPLL_CTRL2_DDI_CLK_SEL_MASK(port);
+	pipe_config->ddi_pll_sel = temp >> (port * 3 + 1);
+
+	switch (pipe_config->ddi_pll_sel) {
+	case SKL_DPLL1:
+		pipe_config->shared_dpll = DPLL_ID_SKL_DPLL1;
+		break;
+	case SKL_DPLL2:
+		pipe_config->shared_dpll = DPLL_ID_SKL_DPLL2;
+		break;
+	case SKL_DPLL3:
+		pipe_config->shared_dpll = DPLL_ID_SKL_DPLL3;
+		break;
+	}
+}
 
-	intel_crtc->lowfreq_avail = false;
+static void haswell_get_ddi_pll(struct drm_i915_private *dev_priv,
+				enum port port,
+				struct intel_crtc_config *pipe_config)
+{
+	pipe_config->ddi_pll_sel = I915_READ(PORT_CLK_SEL(port));
 
-	return 0;
+	switch (pipe_config->ddi_pll_sel) {
+	case PORT_CLK_SEL_WRPLL1:
+		pipe_config->shared_dpll = DPLL_ID_WRPLL1;
+		break;
+	case PORT_CLK_SEL_WRPLL2:
+		pipe_config->shared_dpll = DPLL_ID_WRPLL2;
+		break;
+	}
 }
 
 static void haswell_get_ddi_port_state(struct intel_crtc *crtc,
@@ -7628,16 +8102,10 @@ static void haswell_get_ddi_port_state(struct intel_crtc *crtc,
 
 	port = (tmp & TRANS_DDI_PORT_MASK) >> TRANS_DDI_PORT_SHIFT;
 
-	pipe_config->ddi_pll_sel = I915_READ(PORT_CLK_SEL(port));
-
-	switch (pipe_config->ddi_pll_sel) {
-	case PORT_CLK_SEL_WRPLL1:
-		pipe_config->shared_dpll = DPLL_ID_WRPLL1;
-		break;
-	case PORT_CLK_SEL_WRPLL2:
-		pipe_config->shared_dpll = DPLL_ID_WRPLL2;
-		break;
-	}
+	if (IS_SKYLAKE(dev))
+		skylake_get_ddi_pll(dev_priv, port, pipe_config);
+	else
+		haswell_get_ddi_pll(dev_priv, port, pipe_config);
 
 	if (pipe_config->shared_dpll >= 0) {
 		pll = &dev_priv->shared_dplls[pipe_config->shared_dpll];
@@ -7651,7 +8119,8 @@ static void haswell_get_ddi_port_state(struct intel_crtc *crtc,
 	 * DDI E. So just check whether this pipe is wired to DDI E and whether
 	 * the PCH transcoder is on.
 	 */
-	if ((port == PORT_E) && I915_READ(LPT_TRANSCONF) & TRANS_ENABLE) {
+	if (INTEL_INFO(dev)->gen < 9 &&
+	    (port == PORT_E) && I915_READ(LPT_TRANSCONF) & TRANS_ENABLE) {
 		pipe_config->has_pch_encoder = true;
 
 		tmp = I915_READ(FDI_RX_CTL(PIPE_A));
@@ -7670,7 +8139,7 @@ static bool haswell_get_pipe_config(struct intel_crtc *crtc,
 	enum intel_display_power_domain pfit_domain;
 	uint32_t tmp;
 
-	if (!intel_display_power_enabled(dev_priv,
+	if (!intel_display_power_is_enabled(dev_priv,
 					 POWER_DOMAIN_PIPE(crtc->pipe)))
 		return false;
 
@@ -7699,7 +8168,7 @@ static bool haswell_get_pipe_config(struct intel_crtc *crtc,
 			pipe_config->cpu_transcoder = TRANSCODER_EDP;
 	}
 
-	if (!intel_display_power_enabled(dev_priv,
+	if (!intel_display_power_is_enabled(dev_priv,
 			POWER_DOMAIN_TRANSCODER(pipe_config->cpu_transcoder)))
 		return false;
 
@@ -7712,353 +8181,82 @@ static bool haswell_get_pipe_config(struct intel_crtc *crtc,
 	intel_get_pipe_timings(crtc, pipe_config);
 
 	pfit_domain = POWER_DOMAIN_PIPE_PANEL_FITTER(crtc->pipe);
-	if (intel_display_power_enabled(dev_priv, pfit_domain))
-		ironlake_get_pfit_config(crtc, pipe_config);
+	if (intel_display_power_is_enabled(dev_priv, pfit_domain)) {
+		if (IS_SKYLAKE(dev))
+			skylake_get_pfit_config(crtc, pipe_config);
+		else
+			ironlake_get_pfit_config(crtc, pipe_config);
+	}
 
 	if (IS_HASWELL(dev))
 		pipe_config->ips_enabled = hsw_crtc_supports_ips(crtc) &&
 			(I915_READ(IPS_CTL) & IPS_ENABLE);
 
-	pipe_config->pixel_multiplier = 1;
-
-	return true;
-}
-
-static struct {
-	int clock;
-	u32 config;
-} hdmi_audio_clock[] = {
-	{ DIV_ROUND_UP(25200 * 1000, 1001), AUD_CONFIG_PIXEL_CLOCK_HDMI_25175 },
-	{ 25200, AUD_CONFIG_PIXEL_CLOCK_HDMI_25200 }, /* default per bspec */
-	{ 27000, AUD_CONFIG_PIXEL_CLOCK_HDMI_27000 },
-	{ 27000 * 1001 / 1000, AUD_CONFIG_PIXEL_CLOCK_HDMI_27027 },
-	{ 54000, AUD_CONFIG_PIXEL_CLOCK_HDMI_54000 },
-	{ 54000 * 1001 / 1000, AUD_CONFIG_PIXEL_CLOCK_HDMI_54054 },
-	{ DIV_ROUND_UP(74250 * 1000, 1001), AUD_CONFIG_PIXEL_CLOCK_HDMI_74176 },
-	{ 74250, AUD_CONFIG_PIXEL_CLOCK_HDMI_74250 },
-	{ DIV_ROUND_UP(148500 * 1000, 1001), AUD_CONFIG_PIXEL_CLOCK_HDMI_148352 },
-	{ 148500, AUD_CONFIG_PIXEL_CLOCK_HDMI_148500 },
-};
-
-/* get AUD_CONFIG_PIXEL_CLOCK_HDMI_* value for mode */
-static u32 audio_config_hdmi_pixel_clock(struct drm_display_mode *mode)
-{
-	int i;
-
-	for (i = 0; i < ARRAY_SIZE(hdmi_audio_clock); i++) {
-		if (mode->clock == hdmi_audio_clock[i].clock)
-			break;
-	}
-
-	if (i == ARRAY_SIZE(hdmi_audio_clock)) {
-		DRM_DEBUG_KMS("HDMI audio pixel clock setting for %d not found, falling back to defaults\n", mode->clock);
-		i = 1;
+	if (pipe_config->cpu_transcoder != TRANSCODER_EDP) {
+		pipe_config->pixel_multiplier =
+			I915_READ(PIPE_MULT(pipe_config->cpu_transcoder)) + 1;
+	} else {
+		pipe_config->pixel_multiplier = 1;
 	}
 
-	DRM_DEBUG_KMS("Configuring HDMI audio for pixel clock %d (0x%08x)\n",
-		      hdmi_audio_clock[i].clock,
-		      hdmi_audio_clock[i].config);
-
-	return hdmi_audio_clock[i].config;
-}
-
-static bool intel_eld_uptodate(struct drm_connector *connector,
-			       int reg_eldv, uint32_t bits_eldv,
-			       int reg_elda, uint32_t bits_elda,
-			       int reg_edid)
-{
-	struct drm_i915_private *dev_priv = connector->dev->dev_private;
-	uint8_t *eld = connector->eld;
-	uint32_t i;
-
-	i = I915_READ(reg_eldv);
-	i &= bits_eldv;
-
-	if (!eld[0])
-		return !i;
-
-	if (!i)
-		return false;
-
-	i = I915_READ(reg_elda);
-	i &= ~bits_elda;
-	I915_WRITE(reg_elda, i);
-
-	for (i = 0; i < eld[2]; i++)
-		if (I915_READ(reg_edid) != *((uint32_t *)eld + i))
-			return false;
-
 	return true;
 }
 
-static void g4x_write_eld(struct drm_connector *connector,
-			  struct drm_crtc *crtc,
-			  struct drm_display_mode *mode)
+static void i845_update_cursor(struct drm_crtc *crtc, u32 base)
 {
-	struct drm_i915_private *dev_priv = connector->dev->dev_private;
-	uint8_t *eld = connector->eld;
-	uint32_t eldv;
-	uint32_t len;
-	uint32_t i;
-
-	i = I915_READ(G4X_AUD_VID_DID);
-
-	if (i == INTEL_AUDIO_DEVBLC || i == INTEL_AUDIO_DEVCL)
-		eldv = G4X_ELDV_DEVCL_DEVBLC;
-	else
-		eldv = G4X_ELDV_DEVCTG;
-
-	if (intel_eld_uptodate(connector,
-			       G4X_AUD_CNTL_ST, eldv,
-			       G4X_AUD_CNTL_ST, G4X_ELD_ADDR,
-			       G4X_HDMIW_HDMIEDID))
-		return;
-
-	i = I915_READ(G4X_AUD_CNTL_ST);
-	i &= ~(eldv | G4X_ELD_ADDR);
-	len = (i >> 9) & 0x1f;		/* ELD buffer size */
-	I915_WRITE(G4X_AUD_CNTL_ST, i);
-
-	if (!eld[0])
-		return;
-
-	len = min_t(uint8_t, eld[2], len);
-	DRM_DEBUG_DRIVER("ELD size %d\n", len);
-	for (i = 0; i < len; i++)
-		I915_WRITE(G4X_HDMIW_HDMIEDID, *((uint32_t *)eld + i));
-
-	i = I915_READ(G4X_AUD_CNTL_ST);
-	i |= eldv;
-	I915_WRITE(G4X_AUD_CNTL_ST, i);
-}
-
-static void haswell_write_eld(struct drm_connector *connector,
-			      struct drm_crtc *crtc,
-			      struct drm_display_mode *mode)
-{
-	struct drm_i915_private *dev_priv = connector->dev->dev_private;
-	uint8_t *eld = connector->eld;
-	uint32_t eldv;
-	uint32_t i;
-	int len;
-	int pipe = to_intel_crtc(crtc)->pipe;
-	int tmp;
-
-	int hdmiw_hdmiedid = HSW_AUD_EDID_DATA(pipe);
-	int aud_cntl_st = HSW_AUD_DIP_ELD_CTRL(pipe);
-	int aud_config = HSW_AUD_CFG(pipe);
-	int aud_cntrl_st2 = HSW_AUD_PIN_ELD_CP_VLD;
-
-	/* Audio output enable */
-	DRM_DEBUG_DRIVER("HDMI audio: enable codec\n");
-	tmp = I915_READ(aud_cntrl_st2);
-	tmp |= (AUDIO_OUTPUT_ENABLE_A << (pipe * 4));
-	I915_WRITE(aud_cntrl_st2, tmp);
-	POSTING_READ(aud_cntrl_st2);
-
-	assert_pipe_disabled(dev_priv, to_intel_crtc(crtc)->pipe);
-
-	/* Set ELD valid state */
-	tmp = I915_READ(aud_cntrl_st2);
-	DRM_DEBUG_DRIVER("HDMI audio: pin eld vld status=0x%08x\n", tmp);
-	tmp |= (AUDIO_ELD_VALID_A << (pipe * 4));
-	I915_WRITE(aud_cntrl_st2, tmp);
-	tmp = I915_READ(aud_cntrl_st2);
-	DRM_DEBUG_DRIVER("HDMI audio: eld vld status=0x%08x\n", tmp);
-
-	/* Enable HDMI mode */
-	tmp = I915_READ(aud_config);
-	DRM_DEBUG_DRIVER("HDMI audio: audio conf: 0x%08x\n", tmp);
-	/* clear N_programing_enable and N_value_index */
-	tmp &= ~(AUD_CONFIG_N_VALUE_INDEX | AUD_CONFIG_N_PROG_ENABLE);
-	I915_WRITE(aud_config, tmp);
-
-	DRM_DEBUG_DRIVER("ELD on pipe %c\n", pipe_name(pipe));
-
-	eldv = AUDIO_ELD_VALID_A << (pipe * 4);
-
-	if (intel_pipe_has_type(crtc, INTEL_OUTPUT_DISPLAYPORT)) {
-		DRM_DEBUG_DRIVER("ELD: DisplayPort detected\n");
-		eld[5] |= (1 << 2);	/* Conn_Type, 0x1 = DisplayPort */
-		I915_WRITE(aud_config, AUD_CONFIG_N_VALUE_INDEX); /* 0x1 = DP */
-	} else {
-		I915_WRITE(aud_config, audio_config_hdmi_pixel_clock(mode));
-	}
-
-	if (intel_eld_uptodate(connector,
-			       aud_cntrl_st2, eldv,
-			       aud_cntl_st, IBX_ELD_ADDRESS,
-			       hdmiw_hdmiedid))
-		return;
-
-	i = I915_READ(aud_cntrl_st2);
-	i &= ~eldv;
-	I915_WRITE(aud_cntrl_st2, i);
-
-	if (!eld[0])
-		return;
+	struct drm_device *dev = crtc->dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
+	uint32_t cntl = 0, size = 0;
 
-	i = I915_READ(aud_cntl_st);
-	i &= ~IBX_ELD_ADDRESS;
-	I915_WRITE(aud_cntl_st, i);
-	i = (i >> 29) & DIP_PORT_SEL_MASK;		/* DIP_Port_Select, 0x1 = PortB */
-	DRM_DEBUG_DRIVER("port num:%d\n", i);
-
-	len = min_t(uint8_t, eld[2], 21);	/* 84 bytes of hw ELD buffer */
-	DRM_DEBUG_DRIVER("ELD size %d\n", len);
-	for (i = 0; i < len; i++)
-		I915_WRITE(hdmiw_hdmiedid, *((uint32_t *)eld + i));
-
-	i = I915_READ(aud_cntrl_st2);
-	i |= eldv;
-	I915_WRITE(aud_cntrl_st2, i);
-
-}
-
-static void ironlake_write_eld(struct drm_connector *connector,
-			       struct drm_crtc *crtc,
-			       struct drm_display_mode *mode)
-{
-	struct drm_i915_private *dev_priv = connector->dev->dev_private;
-	uint8_t *eld = connector->eld;
-	uint32_t eldv;
-	uint32_t i;
-	int len;
-	int hdmiw_hdmiedid;
-	int aud_config;
-	int aud_cntl_st;
-	int aud_cntrl_st2;
-	int pipe = to_intel_crtc(crtc)->pipe;
-
-	if (HAS_PCH_IBX(connector->dev)) {
-		hdmiw_hdmiedid = IBX_HDMIW_HDMIEDID(pipe);
-		aud_config = IBX_AUD_CFG(pipe);
-		aud_cntl_st = IBX_AUD_CNTL_ST(pipe);
-		aud_cntrl_st2 = IBX_AUD_CNTL_ST2;
-	} else if (IS_VALLEYVIEW(connector->dev)) {
-		hdmiw_hdmiedid = VLV_HDMIW_HDMIEDID(pipe);
-		aud_config = VLV_AUD_CFG(pipe);
-		aud_cntl_st = VLV_AUD_CNTL_ST(pipe);
-		aud_cntrl_st2 = VLV_AUD_CNTL_ST2;
-	} else {
-		hdmiw_hdmiedid = CPT_HDMIW_HDMIEDID(pipe);
-		aud_config = CPT_AUD_CFG(pipe);
-		aud_cntl_st = CPT_AUD_CNTL_ST(pipe);
-		aud_cntrl_st2 = CPT_AUD_CNTRL_ST2;
-	}
+	if (base) {
+		unsigned int width = intel_crtc->cursor_width;
+		unsigned int height = intel_crtc->cursor_height;
+		unsigned int stride = roundup_pow_of_two(width) * 4;
 
-	DRM_DEBUG_DRIVER("ELD on pipe %c\n", pipe_name(pipe));
+		switch (stride) {
+		default:
+			WARN_ONCE(1, "Invalid cursor width/stride, width=%u, stride=%u\n",
+				  width, stride);
+			stride = 256;
+			/* fallthrough */
+		case 256:
+		case 512:
+		case 1024:
+		case 2048:
+			break;
+		}
 
-	if (IS_VALLEYVIEW(connector->dev))  {
-		struct intel_encoder *intel_encoder;
-		struct intel_digital_port *intel_dig_port;
+		cntl |= CURSOR_ENABLE |
+			CURSOR_GAMMA_ENABLE |
+			CURSOR_FORMAT_ARGB |
+			CURSOR_STRIDE(stride);
 
-		intel_encoder = intel_attached_encoder(connector);
-		intel_dig_port = enc_to_dig_port(&intel_encoder->base);
-		i = intel_dig_port->port;
-	} else {
-		i = I915_READ(aud_cntl_st);
-		i = (i >> 29) & DIP_PORT_SEL_MASK;
-		/* DIP_Port_Select, 0x1 = PortB */
+		size = (height << 12) | width;
 	}
 
-	if (!i) {
-		DRM_DEBUG_DRIVER("Audio directed to unknown port\n");
-		/* operate blindly on all ports */
-		eldv = IBX_ELD_VALIDB;
-		eldv |= IBX_ELD_VALIDB << 4;
-		eldv |= IBX_ELD_VALIDB << 8;
-	} else {
-		DRM_DEBUG_DRIVER("ELD on port %c\n", port_name(i));
-		eldv = IBX_ELD_VALIDB << ((i - 1) * 4);
+	if (intel_crtc->cursor_cntl != 0 &&
+	    (intel_crtc->cursor_base != base ||
+	     intel_crtc->cursor_size != size ||
+	     intel_crtc->cursor_cntl != cntl)) {
+		/* On these chipsets we can only modify the base/size/stride
+		 * whilst the cursor is disabled.
+		 */
+		I915_WRITE(_CURACNTR, 0);
+		POSTING_READ(_CURACNTR);
+		intel_crtc->cursor_cntl = 0;
 	}
 
-	if (intel_pipe_has_type(crtc, INTEL_OUTPUT_DISPLAYPORT)) {
-		DRM_DEBUG_DRIVER("ELD: DisplayPort detected\n");
-		eld[5] |= (1 << 2);	/* Conn_Type, 0x1 = DisplayPort */
-		I915_WRITE(aud_config, AUD_CONFIG_N_VALUE_INDEX); /* 0x1 = DP */
-	} else {
-		I915_WRITE(aud_config, audio_config_hdmi_pixel_clock(mode));
+	if (intel_crtc->cursor_base != base) {
+		I915_WRITE(_CURABASE, base);
+		intel_crtc->cursor_base = base;
 	}
 
-	if (intel_eld_uptodate(connector,
-			       aud_cntrl_st2, eldv,
-			       aud_cntl_st, IBX_ELD_ADDRESS,
-			       hdmiw_hdmiedid))
-		return;
-
-	i = I915_READ(aud_cntrl_st2);
-	i &= ~eldv;
-	I915_WRITE(aud_cntrl_st2, i);
-
-	if (!eld[0])
-		return;
-
-	i = I915_READ(aud_cntl_st);
-	i &= ~IBX_ELD_ADDRESS;
-	I915_WRITE(aud_cntl_st, i);
-
-	len = min_t(uint8_t, eld[2], 21);	/* 84 bytes of hw ELD buffer */
-	DRM_DEBUG_DRIVER("ELD size %d\n", len);
-	for (i = 0; i < len; i++)
-		I915_WRITE(hdmiw_hdmiedid, *((uint32_t *)eld + i));
-
-	i = I915_READ(aud_cntrl_st2);
-	i |= eldv;
-	I915_WRITE(aud_cntrl_st2, i);
-}
-
-void intel_write_eld(struct drm_encoder *encoder,
-		     struct drm_display_mode *mode)
-{
-	struct drm_crtc *crtc = encoder->crtc;
-	struct drm_connector *connector;
-	struct drm_device *dev = encoder->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
-
-	connector = drm_select_eld(encoder, mode);
-	if (!connector)
-		return;
-
-	DRM_DEBUG_DRIVER("ELD on [CONNECTOR:%d:%s], [ENCODER:%d:%s]\n",
-			 connector->base.id,
-			 connector->name,
-			 connector->encoder->base.id,
-			 connector->encoder->name);
-
-	connector->eld[6] = drm_av_sync_delay(connector, mode) / 2;
-
-	if (dev_priv->display.write_eld)
-		dev_priv->display.write_eld(connector, crtc, mode);
-}
-
-static void i845_update_cursor(struct drm_crtc *crtc, u32 base)
-{
-	struct drm_device *dev = crtc->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
-	uint32_t cntl;
-
-	if (base != intel_crtc->cursor_base) {
-		/* On these chipsets we can only modify the base whilst
-		 * the cursor is disabled.
-		 */
-		if (intel_crtc->cursor_cntl) {
-			I915_WRITE(_CURACNTR, 0);
-			POSTING_READ(_CURACNTR);
-			intel_crtc->cursor_cntl = 0;
-		}
-
-		I915_WRITE(_CURABASE, base);
-		POSTING_READ(_CURABASE);
+	if (intel_crtc->cursor_size != size) {
+		I915_WRITE(CURSIZE, size);
+		intel_crtc->cursor_size = size;
 	}
 
-	/* XXX width must be 64, stride 256 => 0x00 << 28 */
-	cntl = 0;
-	if (base)
-		cntl = (CURSOR_ENABLE |
-			CURSOR_GAMMA_ENABLE |
-			CURSOR_FORMAT_ARGB);
 	if (intel_crtc->cursor_cntl != cntl) {
 		I915_WRITE(_CURACNTR, cntl);
 		POSTING_READ(_CURACNTR);
@@ -8092,46 +8290,13 @@ static void i9xx_update_cursor(struct drm_crtc *crtc, u32 base)
 				return;
 		}
 		cntl |= pipe << 28; /* Connect to correct pipe */
-	}
-	if (intel_crtc->cursor_cntl != cntl) {
-		I915_WRITE(CURCNTR(pipe), cntl);
-		POSTING_READ(CURCNTR(pipe));
-		intel_crtc->cursor_cntl = cntl;
-	}
-
-	/* and commit changes on next vblank */
-	I915_WRITE(CURBASE(pipe), base);
-	POSTING_READ(CURBASE(pipe));
-}
 
-static void ivb_update_cursor(struct drm_crtc *crtc, u32 base)
-{
-	struct drm_device *dev = crtc->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
-	int pipe = intel_crtc->pipe;
-	uint32_t cntl;
-
-	cntl = 0;
-	if (base) {
-		cntl = MCURSOR_GAMMA_ENABLE;
-		switch (intel_crtc->cursor_width) {
-			case 64:
-				cntl |= CURSOR_MODE_64_ARGB_AX;
-				break;
-			case 128:
-				cntl |= CURSOR_MODE_128_ARGB_AX;
-				break;
-			case 256:
-				cntl |= CURSOR_MODE_256_ARGB_AX;
-				break;
-			default:
-				WARN_ON(1);
-				return;
-		}
+		if (IS_HASWELL(dev) || IS_BROADWELL(dev))
+			cntl |= CURSOR_PIPE_CSC_ENABLE;
 	}
-	if (IS_HASWELL(dev) || IS_BROADWELL(dev))
-		cntl |= CURSOR_PIPE_CSC_ENABLE;
+
+	if (to_intel_plane(crtc->cursor)->rotation == BIT(DRM_ROTATE_180))
+		cntl |= CURSOR_ROTATE_180;
 
 	if (intel_crtc->cursor_cntl != cntl) {
 		I915_WRITE(CURCNTR(pipe), cntl);
@@ -8142,6 +8307,8 @@ static void ivb_update_cursor(struct drm_crtc *crtc, u32 base)
 	/* and commit changes on next vblank */
 	I915_WRITE(CURBASE(pipe), base);
 	POSTING_READ(CURBASE(pipe));
+
+	intel_crtc->cursor_base = base;
 }
 
 /* If no-part of the cursor is visible on the framebuffer, then the GPU may hang... */
@@ -8188,28 +8355,62 @@ static void intel_crtc_update_cursor(struct drm_crtc *crtc,
 
 	I915_WRITE(CURPOS(pipe), pos);
 
-	if (IS_IVYBRIDGE(dev) || IS_HASWELL(dev) || IS_BROADWELL(dev))
-		ivb_update_cursor(crtc, base);
-	else if (IS_845G(dev) || IS_I865G(dev))
+	/* ILK+ do this automagically */
+	if (HAS_GMCH_DISPLAY(dev) &&
+		to_intel_plane(crtc->cursor)->rotation == BIT(DRM_ROTATE_180)) {
+		base += (intel_crtc->cursor_height *
+			intel_crtc->cursor_width - 1) * 4;
+	}
+
+	if (IS_845G(dev) || IS_I865G(dev))
 		i845_update_cursor(crtc, base);
 	else
 		i9xx_update_cursor(crtc, base);
-	intel_crtc->cursor_base = base;
 }
 
-/*
- * intel_crtc_cursor_set_obj - Set cursor to specified GEM object
- *
- * Note that the object's reference will be consumed if the update fails.  If
- * the update succeeds, the reference of the old object (if any) will be
- * consumed.
- */
+static bool cursor_size_ok(struct drm_device *dev,
+			   uint32_t width, uint32_t height)
+{
+	if (width == 0 || height == 0)
+		return false;
+
+	/*
+	 * 845g/865g are special in that they are only limited by
+	 * the width of their cursors, the height is arbitrary up to
+	 * the precision of the register. Everything else requires
+	 * square cursors, limited to a few power-of-two sizes.
+	 */
+	if (IS_845G(dev) || IS_I865G(dev)) {
+		if ((width & 63) != 0)
+			return false;
+
+		if (width > (IS_845G(dev) ? 64 : 512))
+			return false;
+
+		if (height > 1023)
+			return false;
+	} else {
+		switch (width | height) {
+		case 256:
+		case 128:
+			if (IS_GEN2(dev))
+				return false;
+		case 64:
+			break;
+		default:
+			return false;
+		}
+	}
+
+	return true;
+}
+
 static int intel_crtc_cursor_set_obj(struct drm_crtc *crtc,
 				     struct drm_i915_gem_object *obj,
 				     uint32_t width, uint32_t height)
 {
 	struct drm_device *dev = crtc->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 	enum pipe pipe = intel_crtc->pipe;
 	unsigned old_width;
@@ -8220,36 +8421,15 @@ static int intel_crtc_cursor_set_obj(struct drm_crtc *crtc,
 	if (!obj) {
 		DRM_DEBUG_KMS("cursor off\n");
 		addr = 0;
-		obj = NULL;
 		mutex_lock(&dev->struct_mutex);
 		goto finish;
 	}
 
-	/* Check for which cursor types we support */
-	if (!((width == 64 && height == 64) ||
-			(width == 128 && height == 128 && !IS_GEN2(dev)) ||
-			(width == 256 && height == 256 && !IS_GEN2(dev)))) {
-		DRM_DEBUG("Cursor dimension not supported\n");
-		return -EINVAL;
-	}
-
-	if (obj->base.size < width * height * 4) {
-		DRM_DEBUG_KMS("buffer is too small\n");
-		ret = -ENOMEM;
-		goto fail;
-	}
-
 	/* we only need to pin inside GTT if cursor is non-phy */
 	mutex_lock(&dev->struct_mutex);
 	if (!INTEL_INFO(dev)->cursor_needs_physical) {
 		unsigned alignment;
 
-		if (obj->tiling_mode) {
-			DRM_DEBUG_KMS("cursor cannot be tiled\n");
-			ret = -EINVAL;
-			goto fail_locked;
-		}
-
 		/*
 		 * Global gtt pte registers are special registers which actually
 		 * forward writes to a chunk of system memory. Which means that
@@ -8295,9 +8475,6 @@ static int intel_crtc_cursor_set_obj(struct drm_crtc *crtc,
 		addr = obj->phys_handle->busaddr;
 	}
 
-	if (IS_GEN2(dev))
-		I915_WRITE(CURSIZE, (height << 12) | width);
-
  finish:
 	if (intel_crtc->cursor_bo) {
 		if (!INTEL_INFO(dev)->cursor_needs_physical)
@@ -8319,17 +8496,15 @@ static int intel_crtc_cursor_set_obj(struct drm_crtc *crtc,
 		if (old_width != width)
 			intel_update_watermarks(crtc);
 		intel_crtc_update_cursor(crtc, intel_crtc->cursor_bo != NULL);
-	}
 
-	intel_frontbuffer_flip(dev, INTEL_FRONTBUFFER_CURSOR(pipe));
+		intel_frontbuffer_flip(dev, INTEL_FRONTBUFFER_CURSOR(pipe));
+	}
 
 	return 0;
 fail_unpin:
 	i915_gem_object_unpin_from_display_plane(obj);
 fail_locked:
 	mutex_unlock(&dev->struct_mutex);
-fail:
-	drm_gem_object_unreference_unlocked(&obj->base);
 	return ret;
 }
 
@@ -8364,7 +8539,7 @@ __intel_framebuffer_create(struct drm_device *dev,
 
 	intel_fb = kzalloc(sizeof(*intel_fb), GFP_KERNEL);
 	if (!intel_fb) {
-		drm_gem_object_unreference_unlocked(&obj->base);
+		drm_gem_object_unreference(&obj->base);
 		return ERR_PTR(-ENOMEM);
 	}
 
@@ -8374,7 +8549,7 @@ __intel_framebuffer_create(struct drm_device *dev,
 
 	return &intel_fb->base;
 err:
-	drm_gem_object_unreference_unlocked(&obj->base);
+	drm_gem_object_unreference(&obj->base);
 	kfree(intel_fb);
 
 	return ERR_PTR(ret);
@@ -8507,6 +8682,9 @@ retry:
 		ret = drm_modeset_lock(&crtc->mutex, ctx);
 		if (ret)
 			goto fail_unlock;
+		ret = drm_modeset_lock(&crtc->primary->mutex, ctx);
+		if (ret)
+			goto fail_unlock;
 
 		old->dpms_mode = connector->dpms;
 		old->load_detect_temp = false;
@@ -8544,6 +8722,9 @@ retry:
 	ret = drm_modeset_lock(&crtc->mutex, ctx);
 	if (ret)
 		goto fail_unlock;
+	ret = drm_modeset_lock(&crtc->primary->mutex, ctx);
+	if (ret)
+		goto fail_unlock;
 	intel_encoder->new_crtc = to_intel_crtc(crtc);
 	to_intel_connector(connector)->new_encoder = intel_encoder;
 
@@ -8826,35 +9007,6 @@ struct drm_display_mode *intel_crtc_mode_get(struct drm_device *dev,
 	return mode;
 }
 
-static void intel_increase_pllclock(struct drm_device *dev,
-				    enum pipe pipe)
-{
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	int dpll_reg = DPLL(pipe);
-	int dpll;
-
-	if (!HAS_GMCH_DISPLAY(dev))
-		return;
-
-	if (!dev_priv->lvds_downclock_avail)
-		return;
-
-	dpll = I915_READ(dpll_reg);
-	if (!HAS_PIPE_CXSR(dev) && (dpll & DISPLAY_RATE_SELECT_FPA1)) {
-		DRM_DEBUG_DRIVER("upclocking LVDS\n");
-
-		assert_panel_unlocked(dev_priv, pipe);
-
-		dpll &= ~DISPLAY_RATE_SELECT_FPA1;
-		I915_WRITE(dpll_reg, dpll);
-		intel_wait_for_vblank(dev, pipe);
-
-		dpll = I915_READ(dpll_reg);
-		if (dpll & DISPLAY_RATE_SELECT_FPA1)
-			DRM_DEBUG_DRIVER("failed to upclock LVDS!\n");
-	}
-}
-
 static void intel_decrease_pllclock(struct drm_crtc *crtc)
 {
 	struct drm_device *dev = crtc->dev;
@@ -8930,190 +9082,16 @@ out:
 	intel_runtime_pm_put(dev_priv);
 }
 
-
-/**
- * intel_mark_fb_busy - mark given planes as busy
- * @dev: DRM device
- * @frontbuffer_bits: bits for the affected planes
- * @ring: optional ring for asynchronous commands
- *
- * This function gets called every time the screen contents change. It can be
- * used to keep e.g. the update rate at the nominal refresh rate with DRRS.
- */
-static void intel_mark_fb_busy(struct drm_device *dev,
-			       unsigned frontbuffer_bits,
-			       struct intel_engine_cs *ring)
-{
-	enum pipe pipe;
-
-	if (!i915.powersave)
-		return;
-
-	for_each_pipe(pipe) {
-		if (!(frontbuffer_bits & INTEL_FRONTBUFFER_ALL_MASK(pipe)))
-			continue;
-
-		intel_increase_pllclock(dev, pipe);
-		if (ring && intel_fbc_enabled(dev))
-			ring->fbc_dirty = true;
-	}
-}
-
-/**
- * intel_fb_obj_invalidate - invalidate frontbuffer object
- * @obj: GEM object to invalidate
- * @ring: set for asynchronous rendering
- *
- * This function gets called every time rendering on the given object starts and
- * frontbuffer caching (fbc, low refresh rate for DRRS, panel self refresh) must
- * be invalidated. If @ring is non-NULL any subsequent invalidation will be delayed
- * until the rendering completes or a flip on this frontbuffer plane is
- * scheduled.
- */
-void intel_fb_obj_invalidate(struct drm_i915_gem_object *obj,
-			     struct intel_engine_cs *ring)
-{
-	struct drm_device *dev = obj->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
-
-	WARN_ON(!mutex_is_locked(&dev->struct_mutex));
-
-	if (!obj->frontbuffer_bits)
-		return;
-
-	if (ring) {
-		mutex_lock(&dev_priv->fb_tracking.lock);
-		dev_priv->fb_tracking.busy_bits
-			|= obj->frontbuffer_bits;
-		dev_priv->fb_tracking.flip_bits
-			&= ~obj->frontbuffer_bits;
-		mutex_unlock(&dev_priv->fb_tracking.lock);
-	}
-
-	intel_mark_fb_busy(dev, obj->frontbuffer_bits, ring);
-
-	intel_edp_psr_invalidate(dev, obj->frontbuffer_bits);
-}
-
-/**
- * intel_frontbuffer_flush - flush frontbuffer
- * @dev: DRM device
- * @frontbuffer_bits: frontbuffer plane tracking bits
- *
- * This function gets called every time rendering on the given planes has
- * completed and frontbuffer caching can be started again. Flushes will get
- * delayed if they're blocked by some oustanding asynchronous rendering.
- *
- * Can be called without any locks held.
- */
-void intel_frontbuffer_flush(struct drm_device *dev,
-			     unsigned frontbuffer_bits)
-{
-	struct drm_i915_private *dev_priv = dev->dev_private;
-
-	/* Delay flushing when rings are still busy.*/
-	mutex_lock(&dev_priv->fb_tracking.lock);
-	frontbuffer_bits &= ~dev_priv->fb_tracking.busy_bits;
-	mutex_unlock(&dev_priv->fb_tracking.lock);
-
-	intel_mark_fb_busy(dev, frontbuffer_bits, NULL);
-
-	intel_edp_psr_flush(dev, frontbuffer_bits);
-}
-
-/**
- * intel_fb_obj_flush - flush frontbuffer object
- * @obj: GEM object to flush
- * @retire: set when retiring asynchronous rendering
- *
- * This function gets called every time rendering on the given object has
- * completed and frontbuffer caching can be started again. If @retire is true
- * then any delayed flushes will be unblocked.
- */
-void intel_fb_obj_flush(struct drm_i915_gem_object *obj,
-			bool retire)
-{
-	struct drm_device *dev = obj->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	unsigned frontbuffer_bits;
-
-	WARN_ON(!mutex_is_locked(&dev->struct_mutex));
-
-	if (!obj->frontbuffer_bits)
-		return;
-
-	frontbuffer_bits = obj->frontbuffer_bits;
-
-	if (retire) {
-		mutex_lock(&dev_priv->fb_tracking.lock);
-		/* Filter out new bits since rendering started. */
-		frontbuffer_bits &= dev_priv->fb_tracking.busy_bits;
-
-		dev_priv->fb_tracking.busy_bits &= ~frontbuffer_bits;
-		mutex_unlock(&dev_priv->fb_tracking.lock);
-	}
-
-	intel_frontbuffer_flush(dev, frontbuffer_bits);
-}
-
-/**
- * intel_frontbuffer_flip_prepare - prepare asnychronous frontbuffer flip
- * @dev: DRM device
- * @frontbuffer_bits: frontbuffer plane tracking bits
- *
- * This function gets called after scheduling a flip on @obj. The actual
- * frontbuffer flushing will be delayed until completion is signalled with
- * intel_frontbuffer_flip_complete. If an invalidate happens in between this
- * flush will be cancelled.
- *
- * Can be called without any locks held.
- */
-void intel_frontbuffer_flip_prepare(struct drm_device *dev,
-				    unsigned frontbuffer_bits)
-{
-	struct drm_i915_private *dev_priv = dev->dev_private;
-
-	mutex_lock(&dev_priv->fb_tracking.lock);
-	dev_priv->fb_tracking.flip_bits
-		|= frontbuffer_bits;
-	mutex_unlock(&dev_priv->fb_tracking.lock);
-}
-
-/**
- * intel_frontbuffer_flip_complete - complete asynchronous frontbuffer flush
- * @dev: DRM device
- * @frontbuffer_bits: frontbuffer plane tracking bits
- *
- * This function gets called after the flip has been latched and will complete
- * on the next vblank. It will execute the fush if it hasn't been cancalled yet.
- *
- * Can be called without any locks held.
- */
-void intel_frontbuffer_flip_complete(struct drm_device *dev,
-				     unsigned frontbuffer_bits)
-{
-	struct drm_i915_private *dev_priv = dev->dev_private;
-
-	mutex_lock(&dev_priv->fb_tracking.lock);
-	/* Mask any cancelled flips. */
-	frontbuffer_bits &= dev_priv->fb_tracking.flip_bits;
-	dev_priv->fb_tracking.flip_bits &= ~frontbuffer_bits;
-	mutex_unlock(&dev_priv->fb_tracking.lock);
-
-	intel_frontbuffer_flush(dev, frontbuffer_bits);
-}
-
 static void intel_crtc_destroy(struct drm_crtc *crtc)
 {
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 	struct drm_device *dev = crtc->dev;
 	struct intel_unpin_work *work;
-	unsigned long flags;
 
-	spin_lock_irqsave(&dev->event_lock, flags);
+	spin_lock_irq(&dev->event_lock);
 	work = intel_crtc->unpin_work;
 	intel_crtc->unpin_work = NULL;
-	spin_unlock_irqrestore(&dev->event_lock, flags);
+	spin_unlock_irq(&dev->event_lock);
 
 	if (work) {
 		cancel_work_sync(&work->work);
@@ -9151,7 +9129,6 @@ static void intel_unpin_work_fn(struct work_struct *__work)
 static void do_intel_finish_page_flip(struct drm_device *dev,
 				      struct drm_crtc *crtc)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 	struct intel_unpin_work *work;
 	unsigned long flags;
@@ -9160,6 +9137,10 @@ static void do_intel_finish_page_flip(struct drm_device *dev,
 	if (intel_crtc == NULL)
 		return;
 
+	/*
+	 * This is called both by irq handlers and the reset code (to complete
+	 * lost pageflips) so needs the full irqsave spinlocks.
+	 */
 	spin_lock_irqsave(&dev->event_lock, flags);
 	work = intel_crtc->unpin_work;
 
@@ -9171,23 +9152,9 @@ static void do_intel_finish_page_flip(struct drm_device *dev,
 		return;
 	}
 
-	/* and that the unpin work is consistent wrt ->pending. */
-	smp_rmb();
-
-	intel_crtc->unpin_work = NULL;
-
-	if (work->event)
-		drm_send_vblank_event(dev, intel_crtc->pipe, work->event);
-
-	drm_crtc_vblank_put(crtc);
+	page_flip_completed(intel_crtc);
 
 	spin_unlock_irqrestore(&dev->event_lock, flags);
-
-	wake_up_all(&dev_priv->pending_flip_queue);
-
-	queue_work(dev_priv->wq, &work->work);
-
-	trace_i915_flip_complete(intel_crtc->plane, work->pending_flip_obj);
 }
 
 void intel_finish_page_flip(struct drm_device *dev, int pipe)
@@ -9217,6 +9184,10 @@ static bool page_flip_finished(struct intel_crtc *crtc)
 	struct drm_device *dev = crtc->base.dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 
+	if (i915_reset_in_progress(&dev_priv->gpu_error) ||
+	    crtc->reset_counter != atomic_read(&dev_priv->gpu_error.reset_counter))
+		return true;
+
 	/*
 	 * The relevant registers doen't exist on pre-ctg.
 	 * As the flip done interrupt doesn't trigger for mmio
@@ -9255,7 +9226,12 @@ void intel_prepare_page_flip(struct drm_device *dev, int plane)
 		to_intel_crtc(dev_priv->plane_to_crtc_mapping[plane]);
 	unsigned long flags;
 
-	/* NB: An MMIO update of the plane base pointer will also
+
+	/*
+	 * This is called both by irq handlers and the reset code (to complete
+	 * lost pageflips) so needs the full irqsave spinlocks.
+	 *
+	 * NB: An MMIO update of the plane base pointer will also
 	 * generate a page-flip completion irq, i.e. every modeset
 	 * is also accompanied by a spurious intel_prepare_page_flip().
 	 */
@@ -9532,6 +9508,8 @@ static bool use_mmio_flip(struct intel_engine_cs *ring,
 		return false;
 	else if (i915.use_mmio_flip > 0)
 		return true;
+	else if (i915.enable_execlists)
+		return true;
 	else
 		return ring != obj->ring;
 }
@@ -9543,115 +9521,128 @@ static void intel_do_mmio_flip(struct intel_crtc *intel_crtc)
 	struct intel_framebuffer *intel_fb =
 		to_intel_framebuffer(intel_crtc->base.primary->fb);
 	struct drm_i915_gem_object *obj = intel_fb->obj;
+	bool atomic_update;
+	u32 start_vbl_count;
 	u32 dspcntr;
 	u32 reg;
 
 	intel_mark_page_flip_active(intel_crtc);
 
+	atomic_update = intel_pipe_update_start(intel_crtc, &start_vbl_count);
+
 	reg = DSPCNTR(intel_crtc->plane);
 	dspcntr = I915_READ(reg);
 
-	if (INTEL_INFO(dev)->gen >= 4) {
-		if (obj->tiling_mode != I915_TILING_NONE)
-			dspcntr |= DISPPLANE_TILED;
-		else
-			dspcntr &= ~DISPPLANE_TILED;
-	}
+	if (obj->tiling_mode != I915_TILING_NONE)
+		dspcntr |= DISPPLANE_TILED;
+	else
+		dspcntr &= ~DISPPLANE_TILED;
+
 	I915_WRITE(reg, dspcntr);
 
 	I915_WRITE(DSPSURF(intel_crtc->plane),
 		   intel_crtc->unpin_work->gtt_offset);
 	POSTING_READ(DSPSURF(intel_crtc->plane));
+
+	if (atomic_update)
+		intel_pipe_update_end(intel_crtc, start_vbl_count);
 }
 
-static int intel_postpone_flip(struct drm_i915_gem_object *obj)
+static void intel_mmio_flip_work_func(struct work_struct *work)
 {
+	struct intel_crtc *intel_crtc =
+		container_of(work, struct intel_crtc, mmio_flip.work);
 	struct intel_engine_cs *ring;
-	int ret;
-
-	lockdep_assert_held(&obj->base.dev->struct_mutex);
+	uint32_t seqno;
 
-	if (!obj->last_write_seqno)
-		return 0;
-
-	ring = obj->ring;
-
-	if (i915_seqno_passed(ring->get_seqno(ring, true),
-			      obj->last_write_seqno))
-		return 0;
-
-	ret = i915_gem_check_olr(ring, obj->last_write_seqno);
-	if (ret)
-		return ret;
+	seqno = intel_crtc->mmio_flip.seqno;
+	ring = intel_crtc->mmio_flip.ring;
 
-	if (WARN_ON(!ring->irq_get(ring)))
-		return 0;
+	if (seqno)
+		WARN_ON(__i915_wait_seqno(ring, seqno,
+					  intel_crtc->reset_counter,
+					  false, NULL, NULL) != 0);
 
-	return 1;
+	intel_do_mmio_flip(intel_crtc);
 }
 
-void intel_notify_mmio_flip(struct intel_engine_cs *ring)
+static int intel_queue_mmio_flip(struct drm_device *dev,
+				 struct drm_crtc *crtc,
+				 struct drm_framebuffer *fb,
+				 struct drm_i915_gem_object *obj,
+				 struct intel_engine_cs *ring,
+				 uint32_t flags)
 {
-	struct drm_i915_private *dev_priv = to_i915(ring->dev);
-	struct intel_crtc *intel_crtc;
-	unsigned long irq_flags;
-	u32 seqno;
-
-	seqno = ring->get_seqno(ring, false);
-
-	spin_lock_irqsave(&dev_priv->mmio_flip_lock, irq_flags);
-	for_each_intel_crtc(ring->dev, intel_crtc) {
-		struct intel_mmio_flip *mmio_flip;
+	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 
-		mmio_flip = &intel_crtc->mmio_flip;
-		if (mmio_flip->seqno == 0)
-			continue;
+	intel_crtc->mmio_flip.seqno = obj->last_write_seqno;
+	intel_crtc->mmio_flip.ring = obj->ring;
 
-		if (ring->id != mmio_flip->ring_id)
-			continue;
+	schedule_work(&intel_crtc->mmio_flip.work);
 
-		if (i915_seqno_passed(seqno, mmio_flip->seqno)) {
-			intel_do_mmio_flip(intel_crtc);
-			mmio_flip->seqno = 0;
-			ring->irq_put(ring);
-		}
-	}
-	spin_unlock_irqrestore(&dev_priv->mmio_flip_lock, irq_flags);
+	return 0;
 }
 
-static int intel_queue_mmio_flip(struct drm_device *dev,
+static int intel_gen9_queue_flip(struct drm_device *dev,
 				 struct drm_crtc *crtc,
 				 struct drm_framebuffer *fb,
 				 struct drm_i915_gem_object *obj,
 				 struct intel_engine_cs *ring,
 				 uint32_t flags)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
-	unsigned long irq_flags;
+	uint32_t plane = 0, stride;
 	int ret;
 
-	if (WARN_ON(intel_crtc->mmio_flip.seqno))
-		return -EBUSY;
+	switch(intel_crtc->pipe) {
+	case PIPE_A:
+		plane = MI_DISPLAY_FLIP_SKL_PLANE_1_A;
+		break;
+	case PIPE_B:
+		plane = MI_DISPLAY_FLIP_SKL_PLANE_1_B;
+		break;
+	case PIPE_C:
+		plane = MI_DISPLAY_FLIP_SKL_PLANE_1_C;
+		break;
+	default:
+		WARN_ONCE(1, "unknown plane in flip command\n");
+		return -ENODEV;
+	}
 
-	ret = intel_postpone_flip(obj);
-	if (ret < 0)
-		return ret;
-	if (ret == 0) {
-		intel_do_mmio_flip(intel_crtc);
-		return 0;
+	switch (obj->tiling_mode) {
+	case I915_TILING_NONE:
+		stride = fb->pitches[0] >> 6;
+		break;
+	case I915_TILING_X:
+		stride = fb->pitches[0] >> 9;
+		break;
+	default:
+		WARN_ONCE(1, "unknown tiling in flip command\n");
+		return -ENODEV;
 	}
 
-	spin_lock_irqsave(&dev_priv->mmio_flip_lock, irq_flags);
-	intel_crtc->mmio_flip.seqno = obj->last_write_seqno;
-	intel_crtc->mmio_flip.ring_id = obj->ring->id;
-	spin_unlock_irqrestore(&dev_priv->mmio_flip_lock, irq_flags);
+	ret = intel_ring_begin(ring, 10);
+	if (ret)
+		return ret;
+
+	intel_ring_emit(ring, MI_LOAD_REGISTER_IMM(1));
+	intel_ring_emit(ring, DERRMR);
+	intel_ring_emit(ring, ~(DERRMR_PIPEA_PRI_FLIP_DONE |
+				DERRMR_PIPEB_PRI_FLIP_DONE |
+				DERRMR_PIPEC_PRI_FLIP_DONE));
+	intel_ring_emit(ring, MI_STORE_REGISTER_MEM_GEN8(1) |
+			      MI_SRM_LRM_GLOBAL_GTT);
+	intel_ring_emit(ring, DERRMR);
+	intel_ring_emit(ring, ring->scratch.gtt_offset + 256);
+	intel_ring_emit(ring, 0);
+
+	intel_ring_emit(ring, MI_DISPLAY_FLIP_I915 | plane);
+	intel_ring_emit(ring, stride << 6 | obj->tiling_mode);
+	intel_ring_emit(ring, intel_crtc->unpin_work->gtt_offset);
+
+	intel_mark_page_flip_active(intel_crtc);
+	__intel_ring_advance(ring);
 
-	/*
-	 * Double check to catch cases where irq fired before
-	 * mmio flip data was ready
-	 */
-	intel_notify_mmio_flip(obj->ring);
 	return 0;
 }
 
@@ -9665,6 +9656,66 @@ static int intel_default_queue_flip(struct drm_device *dev,
 	return -ENODEV;
 }
 
+static bool __intel_pageflip_stall_check(struct drm_device *dev,
+					 struct drm_crtc *crtc)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
+	struct intel_unpin_work *work = intel_crtc->unpin_work;
+	u32 addr;
+
+	if (atomic_read(&work->pending) >= INTEL_FLIP_COMPLETE)
+		return true;
+
+	if (!work->enable_stall_check)
+		return false;
+
+	if (work->flip_ready_vblank == 0) {
+		if (work->flip_queued_ring &&
+		    !i915_seqno_passed(work->flip_queued_ring->get_seqno(work->flip_queued_ring, true),
+				       work->flip_queued_seqno))
+			return false;
+
+		work->flip_ready_vblank = drm_vblank_count(dev, intel_crtc->pipe);
+	}
+
+	if (drm_vblank_count(dev, intel_crtc->pipe) - work->flip_ready_vblank < 3)
+		return false;
+
+	/* Potential stall - if we see that the flip has happened,
+	 * assume a missed interrupt. */
+	if (INTEL_INFO(dev)->gen >= 4)
+		addr = I915_HI_DISPBASE(I915_READ(DSPSURF(intel_crtc->plane)));
+	else
+		addr = I915_READ(DSPADDR(intel_crtc->plane));
+
+	/* There is a potential issue here with a false positive after a flip
+	 * to the same address. We could address this by checking for a
+	 * non-incrementing frame counter.
+	 */
+	return addr == work->gtt_offset;
+}
+
+void intel_check_page_flip(struct drm_device *dev, int pipe)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_crtc *crtc = dev_priv->pipe_to_crtc_mapping[pipe];
+	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
+
+	WARN_ON(!in_irq());
+
+	if (crtc == NULL)
+		return;
+
+	spin_lock(&dev->event_lock);
+	if (intel_crtc->unpin_work && __intel_pageflip_stall_check(dev, crtc)) {
+		WARN_ONCE(1, "Kicking stuck page flip: queued at %d, now %d\n",
+			 intel_crtc->unpin_work->flip_queued_vblank, drm_vblank_count(dev, pipe));
+		page_flip_completed(intel_crtc);
+	}
+	spin_unlock(&dev->event_lock);
+}
+
 static int intel_crtc_page_flip(struct drm_crtc *crtc,
 				struct drm_framebuffer *fb,
 				struct drm_pending_vblank_event *event,
@@ -9678,7 +9729,6 @@ static int intel_crtc_page_flip(struct drm_crtc *crtc,
 	enum pipe pipe = intel_crtc->pipe;
 	struct intel_unpin_work *work;
 	struct intel_engine_cs *ring;
-	unsigned long flags;
 	int ret;
 
 	/*
@@ -9719,17 +9769,25 @@ static int intel_crtc_page_flip(struct drm_crtc *crtc,
 		goto free_work;
 
 	/* We borrow the event spin lock for protecting unpin_work */
-	spin_lock_irqsave(&dev->event_lock, flags);
+	spin_lock_irq(&dev->event_lock);
 	if (intel_crtc->unpin_work) {
-		spin_unlock_irqrestore(&dev->event_lock, flags);
-		kfree(work);
-		drm_crtc_vblank_put(crtc);
+		/* Before declaring the flip queue wedged, check if
+		 * the hardware completed the operation behind our backs.
+		 */
+		if (__intel_pageflip_stall_check(dev, crtc)) {
+			DRM_DEBUG_DRIVER("flip queue: previous flip completed, continuing\n");
+			page_flip_completed(intel_crtc);
+		} else {
+			DRM_DEBUG_DRIVER("flip queue: crtc already busy\n");
+			spin_unlock_irq(&dev->event_lock);
 
-		DRM_DEBUG_DRIVER("flip queue: crtc already busy\n");
-		return -EBUSY;
+			drm_crtc_vblank_put(crtc);
+			kfree(work);
+			return -EBUSY;
+		}
 	}
 	intel_crtc->unpin_work = work;
-	spin_unlock_irqrestore(&dev->event_lock, flags);
+	spin_unlock_irq(&dev->event_lock);
 
 	if (atomic_read(&intel_crtc->unpin_work_count) >= 2)
 		flush_workqueue(dev_priv->wq);
@@ -9746,8 +9804,6 @@ static int intel_crtc_page_flip(struct drm_crtc *crtc,
 
 	work->pending_flip_obj = obj;
 
-	work->enable_stall_check = true;
-
 	atomic_inc(&intel_crtc->unpin_work_count);
 	intel_crtc->reset_counter = atomic_read(&dev_priv->gpu_error.reset_counter);
 
@@ -9769,21 +9825,33 @@ static int intel_crtc_page_flip(struct drm_crtc *crtc,
 		ring = &dev_priv->ring[RCS];
 	}
 
-	ret = intel_pin_and_fence_fb_obj(dev, obj, ring);
+	ret = intel_pin_and_fence_fb_obj(crtc->primary, fb, ring);
 	if (ret)
 		goto cleanup_pending;
 
 	work->gtt_offset =
 		i915_gem_obj_ggtt_offset(obj) + intel_crtc->dspaddr_offset;
 
-	if (use_mmio_flip(ring, obj))
+	if (use_mmio_flip(ring, obj)) {
 		ret = intel_queue_mmio_flip(dev, crtc, fb, obj, ring,
 					    page_flip_flags);
-	else
+		if (ret)
+			goto cleanup_unpin;
+
+		work->flip_queued_seqno = obj->last_write_seqno;
+		work->flip_queued_ring = obj->ring;
+	} else {
 		ret = dev_priv->display.queue_flip(dev, crtc, fb, obj, ring,
-				page_flip_flags);
-	if (ret)
-		goto cleanup_unpin;
+						   page_flip_flags);
+		if (ret)
+			goto cleanup_unpin;
+
+		work->flip_queued_seqno = intel_ring_get_seqno(ring);
+		work->flip_queued_ring = ring;
+	}
+
+	work->flip_queued_vblank = drm_vblank_count(dev, intel_crtc->pipe);
+	work->enable_stall_check = true;
 
 	i915_gem_track_fb(work->old_fb_obj, obj,
 			  INTEL_FRONTBUFFER_PRIMARY(pipe));
@@ -9806,9 +9874,9 @@ cleanup_pending:
 	mutex_unlock(&dev->struct_mutex);
 
 cleanup:
-	spin_lock_irqsave(&dev->event_lock, flags);
+	spin_lock_irq(&dev->event_lock);
 	intel_crtc->unpin_work = NULL;
-	spin_unlock_irqrestore(&dev->event_lock, flags);
+	spin_unlock_irq(&dev->event_lock);
 
 	drm_crtc_vblank_put(crtc);
 free_work:
@@ -9818,8 +9886,11 @@ free_work:
 out_hang:
 		intel_crtc_wait_for_pending_flips(crtc);
 		ret = intel_pipe_set_base(crtc, crtc->x, crtc->y, fb);
-		if (ret == 0 && event)
+		if (ret == 0 && event) {
+			spin_lock_irq(&dev->event_lock);
 			drm_send_vblank_event(dev, pipe, event);
+			spin_unlock_irq(&dev->event_lock);
+		}
 	}
 	return ret;
 }
@@ -9847,8 +9918,7 @@ static void intel_modeset_update_staged_output_state(struct drm_device *dev)
 			to_intel_encoder(connector->base.encoder);
 	}
 
-	list_for_each_entry(encoder, &dev->mode_config.encoder_list,
-			    base.head) {
+	for_each_intel_encoder(dev, encoder) {
 		encoder->new_crtc =
 			to_intel_crtc(encoder->base.crtc);
 	}
@@ -9879,8 +9949,7 @@ static void intel_modeset_commit_output_state(struct drm_device *dev)
 		connector->base.encoder = &connector->new_encoder->base;
 	}
 
-	list_for_each_entry(encoder, &dev->mode_config.encoder_list,
-			    base.head) {
+	for_each_intel_encoder(dev, encoder) {
 		encoder->base.crtc = &encoder->new_crtc->base;
 	}
 
@@ -10007,6 +10076,19 @@ static void intel_dump_pipe_config(struct intel_crtc *crtc,
 		      pipe_config->dp_m_n.gmch_m, pipe_config->dp_m_n.gmch_n,
 		      pipe_config->dp_m_n.link_m, pipe_config->dp_m_n.link_n,
 		      pipe_config->dp_m_n.tu);
+
+	DRM_DEBUG_KMS("dp: %i, gmch_m2: %u, gmch_n2: %u, link_m2: %u, link_n2: %u, tu2: %u\n",
+		      pipe_config->has_dp_encoder,
+		      pipe_config->dp_m2_n2.gmch_m,
+		      pipe_config->dp_m2_n2.gmch_n,
+		      pipe_config->dp_m2_n2.link_m,
+		      pipe_config->dp_m2_n2.link_n,
+		      pipe_config->dp_m2_n2.tu);
+
+	DRM_DEBUG_KMS("audio: %i, infoframes: %i\n",
+		      pipe_config->has_audio,
+		      pipe_config->has_infoframe);
+
 	DRM_DEBUG_KMS("requested mode:\n");
 	drm_mode_debug_printmodeline(&pipe_config->requested_mode);
 	DRM_DEBUG_KMS("adjusted mode:\n");
@@ -10041,8 +10123,7 @@ static bool check_single_encoder_cloning(struct intel_crtc *crtc,
 	struct drm_device *dev = crtc->base.dev;
 	struct intel_encoder *source_encoder;
 
-	list_for_each_entry(source_encoder,
-			    &dev->mode_config.encoder_list, base.head) {
+	for_each_intel_encoder(dev, source_encoder) {
 		if (source_encoder->new_crtc != crtc)
 			continue;
 
@@ -10058,8 +10139,7 @@ static bool check_encoder_cloning(struct intel_crtc *crtc)
 	struct drm_device *dev = crtc->base.dev;
 	struct intel_encoder *encoder;
 
-	list_for_each_entry(encoder,
-			    &dev->mode_config.encoder_list, base.head) {
+	for_each_intel_encoder(dev, encoder) {
 		if (encoder->new_crtc != crtc)
 			continue;
 
@@ -10070,6 +10150,48 @@ static bool check_encoder_cloning(struct intel_crtc *crtc)
 	return true;
 }
 
+static bool check_digital_port_conflicts(struct drm_device *dev)
+{
+	struct intel_connector *connector;
+	unsigned int used_ports = 0;
+
+	/*
+	 * Walk the connector list instead of the encoder
+	 * list to detect the problem on ddi platforms
+	 * where there's just one encoder per digital port.
+	 */
+	list_for_each_entry(connector,
+			    &dev->mode_config.connector_list, base.head) {
+		struct intel_encoder *encoder = connector->new_encoder;
+
+		if (!encoder)
+			continue;
+
+		WARN_ON(!encoder->new_crtc);
+
+		switch (encoder->type) {
+			unsigned int port_mask;
+		case INTEL_OUTPUT_UNKNOWN:
+			if (WARN_ON(!HAS_DDI(dev)))
+				break;
+		case INTEL_OUTPUT_DISPLAYPORT:
+		case INTEL_OUTPUT_HDMI:
+		case INTEL_OUTPUT_EDP:
+			port_mask = 1 << enc_to_dig_port(&encoder->base)->port;
+
+			/* the same port mustn't appear more than once */
+			if (used_ports & port_mask)
+				return false;
+
+			used_ports |= port_mask;
+		default:
+			break;
+		}
+	}
+
+	return true;
+}
+
 static struct intel_crtc_config *
 intel_modeset_pipe_config(struct drm_crtc *crtc,
 			  struct drm_framebuffer *fb,
@@ -10086,6 +10208,11 @@ intel_modeset_pipe_config(struct drm_crtc *crtc,
 		return ERR_PTR(-EINVAL);
 	}
 
+	if (!check_digital_port_conflicts(dev)) {
+		DRM_DEBUG_KMS("rejecting conflicting digital port configuration\n");
+		return ERR_PTR(-EINVAL);
+	}
+
 	pipe_config = kzalloc(sizeof(*pipe_config), GFP_KERNEL);
 	if (!pipe_config)
 		return ERR_PTR(-ENOMEM);
@@ -10143,8 +10270,7 @@ encoder_retry:
 	 * adjust it according to limitations or connector properties, and also
 	 * a chance to reject the mode entirely.
 	 */
-	list_for_each_entry(encoder, &dev->mode_config.encoder_list,
-			    base.head) {
+	for_each_intel_encoder(dev, encoder) {
 
 		if (&encoder->new_crtc->base != crtc)
 			continue;
@@ -10222,8 +10348,7 @@ intel_modeset_affected_pipes(struct drm_crtc *crtc, unsigned *modeset_pipes,
 				1 << connector->new_encoder->new_crtc->pipe;
 	}
 
-	list_for_each_entry(encoder, &dev->mode_config.encoder_list,
-			    base.head) {
+	for_each_intel_encoder(dev, encoder) {
 		if (encoder->base.crtc == &encoder->new_crtc->base)
 			continue;
 
@@ -10293,12 +10418,14 @@ static bool intel_crtc_in_use(struct drm_crtc *crtc)
 static void
 intel_modeset_update_state(struct drm_device *dev, unsigned prepare_pipes)
 {
+	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_encoder *intel_encoder;
 	struct intel_crtc *intel_crtc;
 	struct drm_connector *connector;
 
-	list_for_each_entry(intel_encoder, &dev->mode_config.encoder_list,
-			    base.head) {
+	intel_shared_dpll_commit(dev_priv);
+
+	for_each_intel_encoder(dev, intel_encoder) {
 		if (!intel_encoder->base.crtc)
 			continue;
 
@@ -10387,6 +10514,22 @@ intel_pipe_config_compare(struct drm_device *dev,
 		return false; \
 	}
 
+/* This is required for BDW+ where there is only one set of registers for
+ * switching between high and low RR.
+ * This macro can be used whenever a comparison has to be made between one
+ * hw state and multiple sw state variables.
+ */
+#define PIPE_CONF_CHECK_I_ALT(name, alt_name) \
+	if ((current_config->name != pipe_config->name) && \
+		(current_config->alt_name != pipe_config->name)) { \
+			DRM_ERROR("mismatch in " #name " " \
+				  "(expected %i or %i, found %i)\n", \
+				  current_config->name, \
+				  current_config->alt_name, \
+				  pipe_config->name); \
+			return false; \
+	}
+
 #define PIPE_CONF_CHECK_FLAGS(name, mask)	\
 	if ((current_config->name ^ pipe_config->name) & (mask)) { \
 		DRM_ERROR("mismatch in " #name "(" #mask ") "	   \
@@ -10419,11 +10562,28 @@ intel_pipe_config_compare(struct drm_device *dev,
 	PIPE_CONF_CHECK_I(fdi_m_n.tu);
 
 	PIPE_CONF_CHECK_I(has_dp_encoder);
-	PIPE_CONF_CHECK_I(dp_m_n.gmch_m);
-	PIPE_CONF_CHECK_I(dp_m_n.gmch_n);
-	PIPE_CONF_CHECK_I(dp_m_n.link_m);
-	PIPE_CONF_CHECK_I(dp_m_n.link_n);
-	PIPE_CONF_CHECK_I(dp_m_n.tu);
+
+	if (INTEL_INFO(dev)->gen < 8) {
+		PIPE_CONF_CHECK_I(dp_m_n.gmch_m);
+		PIPE_CONF_CHECK_I(dp_m_n.gmch_n);
+		PIPE_CONF_CHECK_I(dp_m_n.link_m);
+		PIPE_CONF_CHECK_I(dp_m_n.link_n);
+		PIPE_CONF_CHECK_I(dp_m_n.tu);
+
+		if (current_config->has_drrs) {
+			PIPE_CONF_CHECK_I(dp_m2_n2.gmch_m);
+			PIPE_CONF_CHECK_I(dp_m2_n2.gmch_n);
+			PIPE_CONF_CHECK_I(dp_m2_n2.link_m);
+			PIPE_CONF_CHECK_I(dp_m2_n2.link_n);
+			PIPE_CONF_CHECK_I(dp_m2_n2.tu);
+		}
+	} else {
+		PIPE_CONF_CHECK_I_ALT(dp_m_n.gmch_m, dp_m2_n2.gmch_m);
+		PIPE_CONF_CHECK_I_ALT(dp_m_n.gmch_n, dp_m2_n2.gmch_n);
+		PIPE_CONF_CHECK_I_ALT(dp_m_n.link_m, dp_m2_n2.link_m);
+		PIPE_CONF_CHECK_I_ALT(dp_m_n.link_n, dp_m2_n2.link_n);
+		PIPE_CONF_CHECK_I_ALT(dp_m_n.tu, dp_m2_n2.tu);
+	}
 
 	PIPE_CONF_CHECK_I(adjusted_mode.crtc_hdisplay);
 	PIPE_CONF_CHECK_I(adjusted_mode.crtc_htotal);
@@ -10444,6 +10604,7 @@ intel_pipe_config_compare(struct drm_device *dev,
 	if ((INTEL_INFO(dev)->gen < 8 && !IS_HASWELL(dev)) ||
 	    IS_VALLEYVIEW(dev))
 		PIPE_CONF_CHECK_I(limited_color_range);
+	PIPE_CONF_CHECK_I(has_infoframe);
 
 	PIPE_CONF_CHECK_I(has_audio);
 
@@ -10500,6 +10661,9 @@ intel_pipe_config_compare(struct drm_device *dev,
 	PIPE_CONF_CHECK_X(dpll_hw_state.fp0);
 	PIPE_CONF_CHECK_X(dpll_hw_state.fp1);
 	PIPE_CONF_CHECK_X(dpll_hw_state.wrpll);
+	PIPE_CONF_CHECK_X(dpll_hw_state.ctrl1);
+	PIPE_CONF_CHECK_X(dpll_hw_state.cfgcr1);
+	PIPE_CONF_CHECK_X(dpll_hw_state.cfgcr2);
 
 	if (IS_G4X(dev) || INTEL_INFO(dev)->gen >= 5)
 		PIPE_CONF_CHECK_I(pipe_bpp);
@@ -10509,6 +10673,7 @@ intel_pipe_config_compare(struct drm_device *dev,
 
 #undef PIPE_CONF_CHECK_X
 #undef PIPE_CONF_CHECK_I
+#undef PIPE_CONF_CHECK_I_ALT
 #undef PIPE_CONF_CHECK_FLAGS
 #undef PIPE_CONF_CHECK_CLOCK_FUZZY
 #undef PIPE_CONF_QUIRK
@@ -10516,6 +10681,56 @@ intel_pipe_config_compare(struct drm_device *dev,
 	return true;
 }
 
+static void check_wm_state(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct skl_ddb_allocation hw_ddb, *sw_ddb;
+	struct intel_crtc *intel_crtc;
+	int plane;
+
+	if (INTEL_INFO(dev)->gen < 9)
+		return;
+
+	skl_ddb_get_hw_state(dev_priv, &hw_ddb);
+	sw_ddb = &dev_priv->wm.skl_hw.ddb;
+
+	for_each_intel_crtc(dev, intel_crtc) {
+		struct skl_ddb_entry *hw_entry, *sw_entry;
+		const enum pipe pipe = intel_crtc->pipe;
+
+		if (!intel_crtc->active)
+			continue;
+
+		/* planes */
+		for_each_plane(pipe, plane) {
+			hw_entry = &hw_ddb.plane[pipe][plane];
+			sw_entry = &sw_ddb->plane[pipe][plane];
+
+			if (skl_ddb_entry_equal(hw_entry, sw_entry))
+				continue;
+
+			DRM_ERROR("mismatch in DDB state pipe %c plane %d "
+				  "(expected (%u,%u), found (%u,%u))\n",
+				  pipe_name(pipe), plane + 1,
+				  sw_entry->start, sw_entry->end,
+				  hw_entry->start, hw_entry->end);
+		}
+
+		/* cursor */
+		hw_entry = &hw_ddb.cursor[pipe];
+		sw_entry = &sw_ddb->cursor[pipe];
+
+		if (skl_ddb_entry_equal(hw_entry, sw_entry))
+			continue;
+
+		DRM_ERROR("mismatch in DDB state pipe %c cursor "
+			  "(expected (%u,%u), found (%u,%u))\n",
+			  pipe_name(pipe),
+			  sw_entry->start, sw_entry->end,
+			  hw_entry->start, hw_entry->end);
+	}
+}
+
 static void
 check_connector_state(struct drm_device *dev)
 {
@@ -10538,8 +10753,7 @@ check_encoder_state(struct drm_device *dev)
 	struct intel_encoder *encoder;
 	struct intel_connector *connector;
 
-	list_for_each_entry(encoder, &dev->mode_config.encoder_list,
-			    base.head) {
+	for_each_intel_encoder(dev, encoder) {
 		bool enabled = false;
 		bool active = false;
 		enum pipe pipe, tracked_pipe;
@@ -10618,8 +10832,7 @@ check_crtc_state(struct drm_device *dev)
 		WARN(crtc->active && !crtc->base.enabled,
 		     "active crtc, but not enabled in sw tracking\n");
 
-		list_for_each_entry(encoder, &dev->mode_config.encoder_list,
-				    base.head) {
+		for_each_intel_encoder(dev, encoder) {
 			if (encoder->base.crtc != &crtc->base)
 				continue;
 			enabled = true;
@@ -10637,12 +10850,12 @@ check_crtc_state(struct drm_device *dev)
 		active = dev_priv->display.get_pipe_config(crtc,
 							   &pipe_config);
 
-		/* hw state is inconsistent with the pipe A quirk */
-		if (crtc->pipe == PIPE_A && dev_priv->quirks & QUIRK_PIPEA_FORCE)
+		/* hw state is inconsistent with the pipe quirk */
+		if ((crtc->pipe == PIPE_A && dev_priv->quirks & QUIRK_PIPEA_FORCE) ||
+		    (crtc->pipe == PIPE_B && dev_priv->quirks & QUIRK_PIPEB_FORCE))
 			active = crtc->active;
 
-		list_for_each_entry(encoder, &dev->mode_config.encoder_list,
-				    base.head) {
+		for_each_intel_encoder(dev, encoder) {
 			enum pipe pipe;
 			if (encoder->base.crtc != &crtc->base)
 				continue;
@@ -10684,9 +10897,9 @@ check_shared_dpll_state(struct drm_device *dev)
 
 		active = pll->get_hw_state(dev_priv, pll, &dpll_hw_state);
 
-		WARN(pll->active > pll->refcount,
+		WARN(pll->active > hweight32(pll->config.crtc_mask),
 		     "more active pll users than references: %i vs %i\n",
-		     pll->active, pll->refcount);
+		     pll->active, hweight32(pll->config.crtc_mask));
 		WARN(pll->active && !pll->on,
 		     "pll in active use but not on in sw tracking\n");
 		WARN(pll->on && !pll->active,
@@ -10704,11 +10917,11 @@ check_shared_dpll_state(struct drm_device *dev)
 		WARN(pll->active != active_crtcs,
 		     "pll active crtcs mismatch (expected %i, found %i)\n",
 		     pll->active, active_crtcs);
-		WARN(pll->refcount != enabled_crtcs,
+		WARN(hweight32(pll->config.crtc_mask) != enabled_crtcs,
 		     "pll enabled crtcs mismatch (expected %i, found %i)\n",
-		     pll->refcount, enabled_crtcs);
+		     hweight32(pll->config.crtc_mask), enabled_crtcs);
 
-		WARN(pll->on && memcmp(&pll->hw_state, &dpll_hw_state,
+		WARN(pll->on && memcmp(&pll->config.hw_state, &dpll_hw_state,
 				       sizeof(dpll_hw_state)),
 		     "pll hw state mismatch\n");
 	}
@@ -10717,6 +10930,7 @@ check_shared_dpll_state(struct drm_device *dev)
 void
 intel_modeset_check_state(struct drm_device *dev)
 {
+	check_wm_state(dev);
 	check_connector_state(dev);
 	check_encoder_state(dev);
 	check_crtc_state(dev);
@@ -10767,50 +10981,67 @@ static void update_scanline_offset(struct intel_crtc *crtc)
 
 		crtc->scanline_offset = vtotal - 1;
 	} else if (HAS_DDI(dev) &&
-		   intel_pipe_has_type(&crtc->base, INTEL_OUTPUT_HDMI)) {
+		   intel_pipe_has_type(crtc, INTEL_OUTPUT_HDMI)) {
 		crtc->scanline_offset = 2;
 	} else
 		crtc->scanline_offset = 1;
 }
 
+static struct intel_crtc_config *
+intel_modeset_compute_config(struct drm_crtc *crtc,
+			     struct drm_display_mode *mode,
+			     struct drm_framebuffer *fb,
+			     unsigned *modeset_pipes,
+			     unsigned *prepare_pipes,
+			     unsigned *disable_pipes)
+{
+	struct intel_crtc_config *pipe_config = NULL;
+
+	intel_modeset_affected_pipes(crtc, modeset_pipes,
+				     prepare_pipes, disable_pipes);
+
+	if ((*modeset_pipes) == 0)
+		goto out;
+
+	/*
+	 * Note this needs changes when we start tracking multiple modes
+	 * and crtcs.  At that point we'll need to compute the whole config
+	 * (i.e. one pipe_config for each crtc) rather than just the one
+	 * for this crtc.
+	 */
+	pipe_config = intel_modeset_pipe_config(crtc, fb, mode);
+	if (IS_ERR(pipe_config)) {
+		goto out;
+	}
+	intel_dump_pipe_config(to_intel_crtc(crtc), pipe_config,
+			       "[modeset]");
+
+out:
+	return pipe_config;
+}
+
 static int __intel_set_mode(struct drm_crtc *crtc,
 			    struct drm_display_mode *mode,
-			    int x, int y, struct drm_framebuffer *fb)
+			    int x, int y, struct drm_framebuffer *fb,
+			    struct intel_crtc_config *pipe_config,
+			    unsigned modeset_pipes,
+			    unsigned prepare_pipes,
+			    unsigned disable_pipes)
 {
 	struct drm_device *dev = crtc->dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct drm_display_mode *saved_mode;
-	struct intel_crtc_config *pipe_config = NULL;
 	struct intel_crtc *intel_crtc;
-	unsigned disable_pipes, prepare_pipes, modeset_pipes;
 	int ret = 0;
 
 	saved_mode = kmalloc(sizeof(*saved_mode), GFP_KERNEL);
 	if (!saved_mode)
 		return -ENOMEM;
 
-	intel_modeset_affected_pipes(crtc, &modeset_pipes,
-				     &prepare_pipes, &disable_pipes);
-
 	*saved_mode = crtc->mode;
 
-	/* Hack: Because we don't (yet) support global modeset on multiple
-	 * crtcs, we don't keep track of the new mode for more than one crtc.
-	 * Hence simply check whether any bit is set in modeset_pipes in all the
-	 * pieces of code that are not yet converted to deal with mutliple crtcs
-	 * changing their mode at the same time. */
-	if (modeset_pipes) {
-		pipe_config = intel_modeset_pipe_config(crtc, fb, mode);
-		if (IS_ERR(pipe_config)) {
-			ret = PTR_ERR(pipe_config);
-			pipe_config = NULL;
-
-			goto out;
-		}
-		intel_dump_pipe_config(to_intel_crtc(crtc), pipe_config,
-				       "[modeset]");
+	if (modeset_pipes)
 		to_intel_crtc(crtc)->new_config = pipe_config;
-	}
 
 	/*
 	 * See if the config requires any additional preparation, e.g.
@@ -10826,6 +11057,22 @@ static int __intel_set_mode(struct drm_crtc *crtc,
 		prepare_pipes &= ~disable_pipes;
 	}
 
+	if (dev_priv->display.crtc_compute_clock) {
+		unsigned clear_pipes = modeset_pipes | disable_pipes;
+
+		ret = intel_shared_dpll_start_config(dev_priv, clear_pipes);
+		if (ret)
+			goto done;
+
+		for_each_intel_crtc_masked(dev, modeset_pipes, intel_crtc) {
+			ret = dev_priv->display.crtc_compute_clock(intel_crtc);
+			if (ret) {
+				intel_shared_dpll_abort_config(dev_priv);
+				goto done;
+			}
+		}
+	}
+
 	for_each_intel_crtc_masked(dev, disable_pipes, intel_crtc)
 		intel_crtc_disable(&intel_crtc->base);
 
@@ -10836,6 +11083,10 @@ static int __intel_set_mode(struct drm_crtc *crtc,
 
 	/* crtc->mode is already used by the ->mode_set callbacks, hence we need
 	 * to set it here already despite that we pass it down the callchain.
+	 *
+	 * Note we'll need to fix this up when we start tracking multiple
+	 * pipes; here we assume a single modeset_pipe and only track the
+	 * single crtc and mode.
 	 */
 	if (modeset_pipes) {
 		crtc->mode = *mode;
@@ -10857,8 +11108,7 @@ static int __intel_set_mode(struct drm_crtc *crtc,
 	 * update the the output configuration. */
 	intel_modeset_update_state(dev, prepare_pipes);
 
-	if (dev_priv->display.modeset_global_resources)
-		dev_priv->display.modeset_global_resources(dev);
+	modeset_update_crtc_power_domains(dev);
 
 	/* Set up the DPLL and any encoders state that needs to adjust or depend
 	 * on the DPLL.
@@ -10869,9 +11119,7 @@ static int __intel_set_mode(struct drm_crtc *crtc,
 		struct drm_i915_gem_object *obj = intel_fb_obj(fb);
 
 		mutex_lock(&dev->struct_mutex);
-		ret = intel_pin_and_fence_fb_obj(dev,
-						 obj,
-						 NULL);
+		ret = intel_pin_and_fence_fb_obj(crtc->primary, fb, NULL);
 		if (ret != 0) {
 			DRM_ERROR("pin & fence failed\n");
 			mutex_unlock(&dev->struct_mutex);
@@ -10886,11 +11134,6 @@ static int __intel_set_mode(struct drm_crtc *crtc,
 		crtc->primary->fb = fb;
 		crtc->x = x;
 		crtc->y = y;
-
-		ret = dev_priv->display.crtc_mode_set(&intel_crtc->base,
-						      x, y, fb);
-		if (ret)
-			goto done;
 	}
 
 	/* Now enable the clocks, plane, pipe, and connectors that we set up. */
@@ -10905,19 +11148,23 @@ done:
 	if (ret && crtc->enabled)
 		crtc->mode = *saved_mode;
 
-out:
 	kfree(pipe_config);
 	kfree(saved_mode);
 	return ret;
 }
 
-static int intel_set_mode(struct drm_crtc *crtc,
-			  struct drm_display_mode *mode,
-			  int x, int y, struct drm_framebuffer *fb)
+static int intel_set_mode_pipes(struct drm_crtc *crtc,
+				struct drm_display_mode *mode,
+				int x, int y, struct drm_framebuffer *fb,
+				struct intel_crtc_config *pipe_config,
+				unsigned modeset_pipes,
+				unsigned prepare_pipes,
+				unsigned disable_pipes)
 {
 	int ret;
 
-	ret = __intel_set_mode(crtc, mode, x, y, fb);
+	ret = __intel_set_mode(crtc, mode, x, y, fb, pipe_config, modeset_pipes,
+			       prepare_pipes, disable_pipes);
 
 	if (ret == 0)
 		intel_modeset_check_state(crtc->dev);
@@ -10925,6 +11172,26 @@ static int intel_set_mode(struct drm_crtc *crtc,
 	return ret;
 }
 
+static int intel_set_mode(struct drm_crtc *crtc,
+			  struct drm_display_mode *mode,
+			  int x, int y, struct drm_framebuffer *fb)
+{
+	struct intel_crtc_config *pipe_config;
+	unsigned modeset_pipes, prepare_pipes, disable_pipes;
+
+	pipe_config = intel_modeset_compute_config(crtc, mode, fb,
+						   &modeset_pipes,
+						   &prepare_pipes,
+						   &disable_pipes);
+
+	if (IS_ERR(pipe_config))
+		return PTR_ERR(pipe_config);
+
+	return intel_set_mode_pipes(crtc, mode, x, y, fb, pipe_config,
+				    modeset_pipes, prepare_pipes,
+				    disable_pipes);
+}
+
 void intel_crtc_restore_mode(struct drm_crtc *crtc)
 {
 	intel_set_mode(crtc, &crtc->mode, crtc->x, crtc->y, crtc->primary->fb);
@@ -11010,7 +11277,7 @@ static void intel_set_config_restore_state(struct drm_device *dev,
 	}
 
 	count = 0;
-	list_for_each_entry(encoder, &dev->mode_config.encoder_list, base.head) {
+	for_each_intel_encoder(dev, encoder) {
 		encoder->new_crtc =
 			to_intel_crtc(config->save_encoder_crtcs[count++]);
 	}
@@ -11169,8 +11436,7 @@ intel_modeset_stage_output_state(struct drm_device *dev,
 	}
 
 	/* Check for any encoders that needs to be disabled. */
-	list_for_each_entry(encoder, &dev->mode_config.encoder_list,
-			    base.head) {
+	for_each_intel_encoder(dev, encoder) {
 		int num_connectors = 0;
 		list_for_each_entry(connector,
 				    &dev->mode_config.connector_list,
@@ -11203,9 +11469,7 @@ intel_modeset_stage_output_state(struct drm_device *dev,
 	for_each_intel_crtc(dev, crtc) {
 		crtc->new_enabled = false;
 
-		list_for_each_entry(encoder,
-				    &dev->mode_config.encoder_list,
-				    base.head) {
+		for_each_intel_encoder(dev, encoder) {
 			if (encoder->new_crtc == crtc) {
 				crtc->new_enabled = true;
 				break;
@@ -11242,7 +11506,7 @@ static void disable_crtc_nofb(struct intel_crtc *crtc)
 			connector->new_encoder = NULL;
 	}
 
-	list_for_each_entry(encoder, &dev->mode_config.encoder_list, base.head) {
+	for_each_intel_encoder(dev, encoder) {
 		if (encoder->new_crtc == crtc)
 			encoder->new_crtc = NULL;
 	}
@@ -11256,6 +11520,8 @@ static int intel_crtc_set_config(struct drm_mode_set *set)
 	struct drm_device *dev;
 	struct drm_mode_set save_set;
 	struct intel_set_config *config;
+	struct intel_crtc_config *pipe_config;
+	unsigned modeset_pipes, prepare_pipes, disable_pipes;
 	int ret;
 
 	BUG_ON(!set);
@@ -11301,11 +11567,39 @@ static int intel_crtc_set_config(struct drm_mode_set *set)
 	if (ret)
 		goto fail;
 
+	pipe_config = intel_modeset_compute_config(set->crtc, set->mode,
+						   set->fb,
+						   &modeset_pipes,
+						   &prepare_pipes,
+						   &disable_pipes);
+	if (IS_ERR(pipe_config)) {
+		ret = PTR_ERR(pipe_config);
+		goto fail;
+	} else if (pipe_config) {
+		if (pipe_config->has_audio !=
+		    to_intel_crtc(set->crtc)->config.has_audio)
+			config->mode_changed = true;
+
+		/*
+		 * Note we have an issue here with infoframes: current code
+		 * only updates them on the full mode set path per hw
+		 * requirements.  So here we should be checking for any
+		 * required changes and forcing a mode set.
+		 */
+	}
+
+	/* set_mode will free it in the mode_changed case */
+	if (!config->mode_changed)
+		kfree(pipe_config);
+
+	intel_update_pipe_size(to_intel_crtc(set->crtc));
+
 	if (config->mode_changed) {
-		ret = intel_set_mode(set->crtc, set->mode,
-				     set->x, set->y, set->fb);
+		ret = intel_set_mode_pipes(set->crtc, set->mode,
+					   set->x, set->y, set->fb, pipe_config,
+					   modeset_pipes, prepare_pipes,
+					   disable_pipes);
 	} else if (config->fb_changed) {
-		struct drm_i915_private *dev_priv = dev->dev_private;
 		struct intel_crtc *intel_crtc = to_intel_crtc(set->crtc);
 
 		intel_crtc_wait_for_pending_flips(set->crtc);
@@ -11319,8 +11613,7 @@ static int intel_crtc_set_config(struct drm_mode_set *set)
 		 */
 		if (!intel_crtc->primary_enabled && ret == 0) {
 			WARN_ON(!intel_crtc->active);
-			intel_enable_primary_hw_plane(dev_priv, intel_crtc->plane,
-						      intel_crtc->pipe);
+			intel_enable_primary_hw_plane(set->crtc->primary, set->crtc);
 		}
 
 		/*
@@ -11375,7 +11668,7 @@ static bool ibx_pch_dpll_get_hw_state(struct drm_i915_private *dev_priv,
 {
 	uint32_t val;
 
-	if (!intel_display_power_enabled(dev_priv, POWER_DOMAIN_PLLS))
+	if (!intel_display_power_is_enabled(dev_priv, POWER_DOMAIN_PLLS))
 		return false;
 
 	val = I915_READ(PCH_DPLL(pll->id));
@@ -11389,8 +11682,8 @@ static bool ibx_pch_dpll_get_hw_state(struct drm_i915_private *dev_priv,
 static void ibx_pch_dpll_mode_set(struct drm_i915_private *dev_priv,
 				  struct intel_shared_dpll *pll)
 {
-	I915_WRITE(PCH_FP0(pll->id), pll->hw_state.fp0);
-	I915_WRITE(PCH_FP1(pll->id), pll->hw_state.fp1);
+	I915_WRITE(PCH_FP0(pll->id), pll->config.hw_state.fp0);
+	I915_WRITE(PCH_FP1(pll->id), pll->config.hw_state.fp1);
 }
 
 static void ibx_pch_dpll_enable(struct drm_i915_private *dev_priv,
@@ -11399,7 +11692,7 @@ static void ibx_pch_dpll_enable(struct drm_i915_private *dev_priv,
 	/* PCH refclock must be enabled first */
 	ibx_assert_pch_refclk_enabled(dev_priv);
 
-	I915_WRITE(PCH_DPLL(pll->id), pll->hw_state.dpll);
+	I915_WRITE(PCH_DPLL(pll->id), pll->config.hw_state.dpll);
 
 	/* Wait for the clocks to stabilize. */
 	POSTING_READ(PCH_DPLL(pll->id));
@@ -11410,7 +11703,7 @@ static void ibx_pch_dpll_enable(struct drm_i915_private *dev_priv,
 	 *
 	 * So write it again.
 	 */
-	I915_WRITE(PCH_DPLL(pll->id), pll->hw_state.dpll);
+	I915_WRITE(PCH_DPLL(pll->id), pll->config.hw_state.dpll);
 	POSTING_READ(PCH_DPLL(pll->id));
 	udelay(200);
 }
@@ -11473,8 +11766,6 @@ static int
 intel_primary_plane_disable(struct drm_plane *plane)
 {
 	struct drm_device *dev = plane->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct intel_plane *intel_plane = to_intel_plane(plane);
 	struct intel_crtc *intel_crtc;
 
 	if (!plane->fb)
@@ -11497,8 +11788,8 @@ intel_primary_plane_disable(struct drm_plane *plane)
 		goto disable_unpin;
 
 	intel_crtc_wait_for_pending_flips(plane->crtc);
-	intel_disable_primary_hw_plane(dev_priv, intel_plane->plane,
-				       intel_plane->pipe);
+	intel_disable_primary_hw_plane(plane, plane->crtc);
+
 disable_unpin:
 	mutex_lock(&dev->struct_mutex);
 	i915_gem_track_fb(intel_fb_obj(plane->fb), NULL,
@@ -11511,123 +11802,195 @@ disable_unpin:
 }
 
 static int
-intel_primary_plane_setplane(struct drm_plane *plane, struct drm_crtc *crtc,
-			     struct drm_framebuffer *fb, int crtc_x, int crtc_y,
-			     unsigned int crtc_w, unsigned int crtc_h,
-			     uint32_t src_x, uint32_t src_y,
-			     uint32_t src_w, uint32_t src_h)
+intel_check_primary_plane(struct drm_plane *plane,
+			  struct intel_plane_state *state)
 {
+	struct drm_crtc *crtc = state->crtc;
+	struct drm_framebuffer *fb = state->fb;
+	struct drm_rect *dest = &state->dst;
+	struct drm_rect *src = &state->src;
+	const struct drm_rect *clip = &state->clip;
+
+	return drm_plane_helper_check_update(plane, crtc, fb,
+					     src, dest, clip,
+					     DRM_PLANE_HELPER_NO_SCALING,
+					     DRM_PLANE_HELPER_NO_SCALING,
+					     false, true, &state->visible);
+}
+
+static int
+intel_prepare_primary_plane(struct drm_plane *plane,
+			    struct intel_plane_state *state)
+{
+	struct drm_crtc *crtc = state->crtc;
+	struct drm_framebuffer *fb = state->fb;
 	struct drm_device *dev = crtc->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
-	struct intel_plane *intel_plane = to_intel_plane(plane);
+	enum pipe pipe = intel_crtc->pipe;
 	struct drm_i915_gem_object *obj = intel_fb_obj(fb);
 	struct drm_i915_gem_object *old_obj = intel_fb_obj(plane->fb);
-	struct drm_rect dest = {
-		/* integer pixels */
-		.x1 = crtc_x,
-		.y1 = crtc_y,
-		.x2 = crtc_x + crtc_w,
-		.y2 = crtc_y + crtc_h,
-	};
-	struct drm_rect src = {
-		/* 16.16 fixed point */
-		.x1 = src_x,
-		.y1 = src_y,
-		.x2 = src_x + src_w,
-		.y2 = src_y + src_h,
-	};
-	const struct drm_rect clip = {
-		/* integer pixels */
-		.x2 = intel_crtc->active ? intel_crtc->config.pipe_src_w : 0,
-		.y2 = intel_crtc->active ? intel_crtc->config.pipe_src_h : 0,
-	};
-	bool visible;
 	int ret;
 
-	ret = drm_plane_helper_check_update(plane, crtc, fb,
-					    &src, &dest, &clip,
-					    DRM_PLANE_HELPER_NO_SCALING,
-					    DRM_PLANE_HELPER_NO_SCALING,
-					    false, true, &visible);
+	intel_crtc_wait_for_pending_flips(crtc);
 
-	if (ret)
-		return ret;
+	if (intel_crtc_has_pending_flip(crtc)) {
+		DRM_ERROR("pipe is still busy with an old pageflip\n");
+		return -EBUSY;
+	}
 
-	/*
-	 * If the CRTC isn't enabled, we're just pinning the framebuffer,
-	 * updating the fb pointer, and returning without touching the
-	 * hardware.  This allows us to later do a drmModeSetCrtc with fb=-1 to
-	 * turn on the display with all planes setup as desired.
-	 */
-	if (!crtc->enabled) {
+	if (old_obj != obj) {
 		mutex_lock(&dev->struct_mutex);
+		ret = intel_pin_and_fence_fb_obj(plane, fb, NULL);
+		if (ret == 0)
+			i915_gem_track_fb(old_obj, obj,
+					  INTEL_FRONTBUFFER_PRIMARY(pipe));
+		mutex_unlock(&dev->struct_mutex);
+		if (ret != 0) {
+			DRM_DEBUG_KMS("pin & fence failed\n");
+			return ret;
+		}
+	}
+
+	return 0;
+}
+
+static void
+intel_commit_primary_plane(struct drm_plane *plane,
+			   struct intel_plane_state *state)
+{
+	struct drm_crtc *crtc = state->crtc;
+	struct drm_framebuffer *fb = state->fb;
+	struct drm_device *dev = crtc->dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
+	enum pipe pipe = intel_crtc->pipe;
+	struct drm_framebuffer *old_fb = plane->fb;
+	struct drm_i915_gem_object *obj = intel_fb_obj(fb);
+	struct drm_i915_gem_object *old_obj = intel_fb_obj(plane->fb);
+	struct intel_plane *intel_plane = to_intel_plane(plane);
+	struct drm_rect *src = &state->src;
+
+	crtc->primary->fb = fb;
+	crtc->x = src->x1 >> 16;
+	crtc->y = src->y1 >> 16;
+
+	intel_plane->crtc_x = state->orig_dst.x1;
+	intel_plane->crtc_y = state->orig_dst.y1;
+	intel_plane->crtc_w = drm_rect_width(&state->orig_dst);
+	intel_plane->crtc_h = drm_rect_height(&state->orig_dst);
+	intel_plane->src_x = state->orig_src.x1;
+	intel_plane->src_y = state->orig_src.y1;
+	intel_plane->src_w = drm_rect_width(&state->orig_src);
+	intel_plane->src_h = drm_rect_height(&state->orig_src);
+	intel_plane->obj = obj;
 
+	if (intel_crtc->active) {
 		/*
-		 * If we already called setplane while the crtc was disabled,
-		 * we may have an fb pinned; unpin it.
+		 * FBC does not work on some platforms for rotated
+		 * planes, so disable it when rotation is not 0 and
+		 * update it when rotation is set back to 0.
+		 *
+		 * FIXME: This is redundant with the fbc update done in
+		 * the primary plane enable function except that that
+		 * one is done too late. We eventually need to unify
+		 * this.
 		 */
-		if (plane->fb)
-			intel_unpin_fb_obj(old_obj);
+		if (intel_crtc->primary_enabled &&
+		    INTEL_INFO(dev)->gen <= 4 && !IS_G4X(dev) &&
+		    dev_priv->fbc.plane == intel_crtc->plane &&
+		    intel_plane->rotation != BIT(DRM_ROTATE_0)) {
+			intel_disable_fbc(dev);
+		}
 
-		i915_gem_track_fb(old_obj, obj,
-				  INTEL_FRONTBUFFER_PRIMARY(intel_crtc->pipe));
+		if (state->visible) {
+			bool was_enabled = intel_crtc->primary_enabled;
 
-		/* Pin and return without programming hardware */
-		ret = intel_pin_and_fence_fb_obj(dev, obj, NULL);
-		mutex_unlock(&dev->struct_mutex);
+			/* FIXME: kill this fastboot hack */
+			intel_update_pipe_size(intel_crtc);
 
-		return ret;
+			intel_crtc->primary_enabled = true;
+
+			dev_priv->display.update_primary_plane(crtc, plane->fb,
+					crtc->x, crtc->y);
+
+			/*
+			 * BDW signals flip done immediately if the plane
+			 * is disabled, even if the plane enable is already
+			 * armed to occur at the next vblank :(
+			 */
+			if (IS_BROADWELL(dev) && !was_enabled)
+				intel_wait_for_vblank(dev, intel_crtc->pipe);
+		} else {
+			/*
+			 * If clipping results in a non-visible primary plane,
+			 * we'll disable the primary plane.  Note that this is
+			 * a bit different than what happens if userspace
+			 * explicitly disables the plane by passing fb=0
+			 * because plane->fb still gets set and pinned.
+			 */
+			intel_disable_primary_hw_plane(plane, crtc);
+		}
+
+		intel_frontbuffer_flip(dev, INTEL_FRONTBUFFER_PRIMARY(pipe));
+
+		mutex_lock(&dev->struct_mutex);
+		intel_update_fbc(dev);
+		mutex_unlock(&dev->struct_mutex);
 	}
 
-	intel_crtc_wait_for_pending_flips(crtc);
+	if (old_fb && old_fb != fb) {
+		if (intel_crtc->active)
+			intel_wait_for_vblank(dev, intel_crtc->pipe);
 
-	/*
-	 * If clipping results in a non-visible primary plane, we'll disable
-	 * the primary plane.  Note that this is a bit different than what
-	 * happens if userspace explicitly disables the plane by passing fb=0
-	 * because plane->fb still gets set and pinned.
-	 */
-	if (!visible) {
 		mutex_lock(&dev->struct_mutex);
+		intel_unpin_fb_obj(old_obj);
+		mutex_unlock(&dev->struct_mutex);
+	}
+}
 
-		/*
-		 * Try to pin the new fb first so that we can bail out if we
-		 * fail.
-		 */
-		if (plane->fb != fb) {
-			ret = intel_pin_and_fence_fb_obj(dev, obj, NULL);
-			if (ret) {
-				mutex_unlock(&dev->struct_mutex);
-				return ret;
-			}
-		}
+static int
+intel_primary_plane_setplane(struct drm_plane *plane, struct drm_crtc *crtc,
+			     struct drm_framebuffer *fb, int crtc_x, int crtc_y,
+			     unsigned int crtc_w, unsigned int crtc_h,
+			     uint32_t src_x, uint32_t src_y,
+			     uint32_t src_w, uint32_t src_h)
+{
+	struct intel_plane_state state;
+	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
+	int ret;
 
-		i915_gem_track_fb(old_obj, obj,
-				  INTEL_FRONTBUFFER_PRIMARY(intel_crtc->pipe));
+	state.crtc = crtc;
+	state.fb = fb;
 
-		if (intel_crtc->primary_enabled)
-			intel_disable_primary_hw_plane(dev_priv,
-						       intel_plane->plane,
-						       intel_plane->pipe);
+	/* sample coordinates in 16.16 fixed point */
+	state.src.x1 = src_x;
+	state.src.x2 = src_x + src_w;
+	state.src.y1 = src_y;
+	state.src.y2 = src_y + src_h;
 
+	/* integer pixels */
+	state.dst.x1 = crtc_x;
+	state.dst.x2 = crtc_x + crtc_w;
+	state.dst.y1 = crtc_y;
+	state.dst.y2 = crtc_y + crtc_h;
 
-		if (plane->fb != fb)
-			if (plane->fb)
-				intel_unpin_fb_obj(old_obj);
+	state.clip.x1 = 0;
+	state.clip.y1 = 0;
+	state.clip.x2 = intel_crtc->active ? intel_crtc->config.pipe_src_w : 0;
+	state.clip.y2 = intel_crtc->active ? intel_crtc->config.pipe_src_h : 0;
 
-		mutex_unlock(&dev->struct_mutex);
+	state.orig_src = state.src;
+	state.orig_dst = state.dst;
 
-		return 0;
-	}
+	ret = intel_check_primary_plane(plane, &state);
+	if (ret)
+		return ret;
 
-	ret = intel_pipe_set_base(crtc, src.x1, src.y1, fb);
+	ret = intel_prepare_primary_plane(plane, &state);
 	if (ret)
 		return ret;
 
-	if (!intel_crtc->primary_enabled)
-		intel_enable_primary_hw_plane(dev_priv, intel_crtc->plane,
-					      intel_crtc->pipe);
+	intel_commit_primary_plane(plane, &state);
 
 	return 0;
 }
@@ -11644,6 +12007,7 @@ static const struct drm_plane_funcs intel_primary_plane_funcs = {
 	.update_plane = intel_primary_plane_setplane,
 	.disable_plane = intel_primary_plane_disable,
 	.destroy = intel_plane_destroy,
+	.set_property = intel_plane_set_property
 };
 
 static struct drm_plane *intel_primary_plane_create(struct drm_device *dev,
@@ -11661,6 +12025,7 @@ static struct drm_plane *intel_primary_plane_create(struct drm_device *dev,
 	primary->max_downscale = 1;
 	primary->pipe = pipe;
 	primary->plane = pipe;
+	primary->rotation = BIT(DRM_ROTATE_0);
 	if (HAS_FBC(dev) && INTEL_INFO(dev)->gen < 4)
 		primary->plane = !pipe;
 
@@ -11676,6 +12041,19 @@ static struct drm_plane *intel_primary_plane_create(struct drm_device *dev,
 				 &intel_primary_plane_funcs,
 				 intel_primary_formats, num_formats,
 				 DRM_PLANE_TYPE_PRIMARY);
+
+	if (INTEL_INFO(dev)->gen >= 4) {
+		if (!dev->mode_config.rotation_property)
+			dev->mode_config.rotation_property =
+				drm_mode_create_rotation_property(dev,
+							BIT(DRM_ROTATE_0) |
+							BIT(DRM_ROTATE_180));
+		if (dev->mode_config.rotation_property)
+			drm_object_attach_property(&primary->base.base,
+				dev->mode_config.rotation_property,
+				primary->rotation);
+	}
+
 	return &primary->base;
 }
 
@@ -11691,58 +12069,146 @@ intel_cursor_plane_disable(struct drm_plane *plane)
 }
 
 static int
-intel_cursor_plane_update(struct drm_plane *plane, struct drm_crtc *crtc,
-			  struct drm_framebuffer *fb, int crtc_x, int crtc_y,
-			  unsigned int crtc_w, unsigned int crtc_h,
-			  uint32_t src_x, uint32_t src_y,
-			  uint32_t src_w, uint32_t src_h)
+intel_check_cursor_plane(struct drm_plane *plane,
+			 struct intel_plane_state *state)
 {
-	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
-	struct intel_framebuffer *intel_fb = to_intel_framebuffer(fb);
-	struct drm_i915_gem_object *obj = intel_fb->obj;
-	struct drm_rect dest = {
-		/* integer pixels */
-		.x1 = crtc_x,
-		.y1 = crtc_y,
-		.x2 = crtc_x + crtc_w,
-		.y2 = crtc_y + crtc_h,
-	};
-	struct drm_rect src = {
-		/* 16.16 fixed point */
-		.x1 = src_x,
-		.y1 = src_y,
-		.x2 = src_x + src_w,
-		.y2 = src_y + src_h,
-	};
-	const struct drm_rect clip = {
-		/* integer pixels */
-		.x2 = intel_crtc->active ? intel_crtc->config.pipe_src_w : 0,
-		.y2 = intel_crtc->active ? intel_crtc->config.pipe_src_h : 0,
-	};
-	bool visible;
+	struct drm_crtc *crtc = state->crtc;
+	struct drm_device *dev = crtc->dev;
+	struct drm_framebuffer *fb = state->fb;
+	struct drm_rect *dest = &state->dst;
+	struct drm_rect *src = &state->src;
+	const struct drm_rect *clip = &state->clip;
+	struct drm_i915_gem_object *obj = intel_fb_obj(fb);
+	int crtc_w, crtc_h;
+	unsigned stride;
 	int ret;
 
 	ret = drm_plane_helper_check_update(plane, crtc, fb,
-					    &src, &dest, &clip,
+					    src, dest, clip,
 					    DRM_PLANE_HELPER_NO_SCALING,
 					    DRM_PLANE_HELPER_NO_SCALING,
-					    true, true, &visible);
+					    true, true, &state->visible);
 	if (ret)
 		return ret;
 
-	crtc->cursor_x = crtc_x;
-	crtc->cursor_y = crtc_y;
+
+	/* if we want to turn off the cursor ignore width and height */
+	if (!obj)
+		return 0;
+
+	/* Check for which cursor types we support */
+	crtc_w = drm_rect_width(&state->orig_dst);
+	crtc_h = drm_rect_height(&state->orig_dst);
+	if (!cursor_size_ok(dev, crtc_w, crtc_h)) {
+		DRM_DEBUG("Cursor dimension not supported\n");
+		return -EINVAL;
+	}
+
+	stride = roundup_pow_of_two(crtc_w) * 4;
+	if (obj->base.size < stride * crtc_h) {
+		DRM_DEBUG_KMS("buffer is too small\n");
+		return -ENOMEM;
+	}
+
+	if (fb == crtc->cursor->fb)
+		return 0;
+
+	/* we only need to pin inside GTT if cursor is non-phy */
+	mutex_lock(&dev->struct_mutex);
+	if (!INTEL_INFO(dev)->cursor_needs_physical && obj->tiling_mode) {
+		DRM_DEBUG_KMS("cursor cannot be tiled\n");
+		ret = -EINVAL;
+	}
+	mutex_unlock(&dev->struct_mutex);
+
+	return ret;
+}
+
+static int
+intel_commit_cursor_plane(struct drm_plane *plane,
+			  struct intel_plane_state *state)
+{
+	struct drm_crtc *crtc = state->crtc;
+	struct drm_framebuffer *fb = state->fb;
+	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
+	struct intel_plane *intel_plane = to_intel_plane(plane);
+	struct intel_framebuffer *intel_fb = to_intel_framebuffer(fb);
+	struct drm_i915_gem_object *obj = intel_fb->obj;
+	int crtc_w, crtc_h;
+
+	crtc->cursor_x = state->orig_dst.x1;
+	crtc->cursor_y = state->orig_dst.y1;
+
+	intel_plane->crtc_x = state->orig_dst.x1;
+	intel_plane->crtc_y = state->orig_dst.y1;
+	intel_plane->crtc_w = drm_rect_width(&state->orig_dst);
+	intel_plane->crtc_h = drm_rect_height(&state->orig_dst);
+	intel_plane->src_x = state->orig_src.x1;
+	intel_plane->src_y = state->orig_src.y1;
+	intel_plane->src_w = drm_rect_width(&state->orig_src);
+	intel_plane->src_h = drm_rect_height(&state->orig_src);
+	intel_plane->obj = obj;
+
 	if (fb != crtc->cursor->fb) {
+		crtc_w = drm_rect_width(&state->orig_dst);
+		crtc_h = drm_rect_height(&state->orig_dst);
 		return intel_crtc_cursor_set_obj(crtc, obj, crtc_w, crtc_h);
 	} else {
-		intel_crtc_update_cursor(crtc, visible);
+		intel_crtc_update_cursor(crtc, state->visible);
+
+		intel_frontbuffer_flip(crtc->dev,
+				       INTEL_FRONTBUFFER_CURSOR(intel_crtc->pipe));
+
 		return 0;
 	}
 }
+
+static int
+intel_cursor_plane_update(struct drm_plane *plane, struct drm_crtc *crtc,
+			  struct drm_framebuffer *fb, int crtc_x, int crtc_y,
+			  unsigned int crtc_w, unsigned int crtc_h,
+			  uint32_t src_x, uint32_t src_y,
+			  uint32_t src_w, uint32_t src_h)
+{
+	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
+	struct intel_plane_state state;
+	int ret;
+
+	state.crtc = crtc;
+	state.fb = fb;
+
+	/* sample coordinates in 16.16 fixed point */
+	state.src.x1 = src_x;
+	state.src.x2 = src_x + src_w;
+	state.src.y1 = src_y;
+	state.src.y2 = src_y + src_h;
+
+	/* integer pixels */
+	state.dst.x1 = crtc_x;
+	state.dst.x2 = crtc_x + crtc_w;
+	state.dst.y1 = crtc_y;
+	state.dst.y2 = crtc_y + crtc_h;
+
+	state.clip.x1 = 0;
+	state.clip.y1 = 0;
+	state.clip.x2 = intel_crtc->active ? intel_crtc->config.pipe_src_w : 0;
+	state.clip.y2 = intel_crtc->active ? intel_crtc->config.pipe_src_h : 0;
+
+	state.orig_src = state.src;
+	state.orig_dst = state.dst;
+
+	ret = intel_check_cursor_plane(plane, &state);
+	if (ret)
+		return ret;
+
+	return intel_commit_cursor_plane(plane, &state);
+}
+
 static const struct drm_plane_funcs intel_cursor_plane_funcs = {
 	.update_plane = intel_cursor_plane_update,
 	.disable_plane = intel_cursor_plane_disable,
 	.destroy = intel_plane_destroy,
+	.set_property = intel_plane_set_property,
 };
 
 static struct drm_plane *intel_cursor_plane_create(struct drm_device *dev,
@@ -11758,12 +12224,26 @@ static struct drm_plane *intel_cursor_plane_create(struct drm_device *dev,
 	cursor->max_downscale = 1;
 	cursor->pipe = pipe;
 	cursor->plane = pipe;
+	cursor->rotation = BIT(DRM_ROTATE_0);
 
 	drm_universal_plane_init(dev, &cursor->base, 0,
 				 &intel_cursor_plane_funcs,
 				 intel_cursor_formats,
 				 ARRAY_SIZE(intel_cursor_formats),
 				 DRM_PLANE_TYPE_CURSOR);
+
+	if (INTEL_INFO(dev)->gen >= 4) {
+		if (!dev->mode_config.rotation_property)
+			dev->mode_config.rotation_property =
+				drm_mode_create_rotation_property(dev,
+							BIT(DRM_ROTATE_0) |
+							BIT(DRM_ROTATE_180));
+		if (dev->mode_config.rotation_property)
+			drm_object_attach_property(&cursor->base.base,
+				dev->mode_config.rotation_property,
+				cursor->rotation);
+	}
+
 	return &cursor->base;
 }
 
@@ -11812,14 +12292,15 @@ static void intel_crtc_init(struct drm_device *dev, int pipe)
 
 	intel_crtc->cursor_base = ~0;
 	intel_crtc->cursor_cntl = ~0;
-
-	init_waitqueue_head(&intel_crtc->vbl_wait);
+	intel_crtc->cursor_size = ~0;
 
 	BUG_ON(pipe >= ARRAY_SIZE(dev_priv->plane_to_crtc_mapping) ||
 	       dev_priv->plane_to_crtc_mapping[intel_crtc->plane] != NULL);
 	dev_priv->plane_to_crtc_mapping[intel_crtc->plane] = &intel_crtc->base;
 	dev_priv->pipe_to_crtc_mapping[intel_crtc->pipe] = &intel_crtc->base;
 
+	INIT_WORK(&intel_crtc->mmio_flip.work, intel_mmio_flip_work_func);
+
 	drm_crtc_helper_add(&intel_crtc->base, &intel_helper_funcs);
 
 	WARN_ON(drm_crtc_index(&intel_crtc->base) != intel_crtc->pipe);
@@ -11840,7 +12321,7 @@ enum pipe intel_get_pipe_from_connector(struct intel_connector *connector)
 
 	WARN_ON(!drm_modeset_is_locked(&dev->mode_config.connection_mutex));
 
-	if (!encoder)
+	if (!encoder || WARN_ON(!encoder->crtc))
 		return INVALID_PIPE;
 
 	return to_intel_crtc(encoder->crtc)->pipe;
@@ -11876,8 +12357,7 @@ static int intel_encoder_clones(struct intel_encoder *encoder)
 	int index_mask = 0;
 	int entry = 0;
 
-	list_for_each_entry(source_encoder,
-			    &dev->mode_config.encoder_list, base.head) {
+	for_each_intel_encoder(dev, source_encoder) {
 		if (encoders_cloneable(encoder, source_encoder))
 			index_mask |= (1 << entry);
 
@@ -11929,7 +12409,10 @@ static bool intel_crt_present(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 
-	if (IS_ULT(dev))
+	if (INTEL_INFO(dev)->gen >= 9)
+		return false;
+
+	if (IS_HSW_ULT(dev) || IS_BDW_ULT(dev))
 		return false;
 
 	if (IS_CHERRYVIEW(dev))
@@ -11999,27 +12482,36 @@ static void intel_setup_outputs(struct drm_device *dev)
 		if (I915_READ(PCH_DP_D) & DP_DETECTED)
 			intel_dp_init(dev, PCH_DP_D, PORT_D);
 	} else if (IS_VALLEYVIEW(dev)) {
-		if (I915_READ(VLV_DISPLAY_BASE + GEN4_HDMIB) & SDVO_DETECTED) {
+		/*
+		 * The DP_DETECTED bit is the latched state of the DDC
+		 * SDA pin at boot. However since eDP doesn't require DDC
+		 * (no way to plug in a DP->HDMI dongle) the DDC pins for
+		 * eDP ports may have been muxed to an alternate function.
+		 * Thus we can't rely on the DP_DETECTED bit alone to detect
+		 * eDP ports. Consult the VBT as well as DP_DETECTED to
+		 * detect eDP ports.
+		 */
+		if (I915_READ(VLV_DISPLAY_BASE + GEN4_HDMIB) & SDVO_DETECTED)
 			intel_hdmi_init(dev, VLV_DISPLAY_BASE + GEN4_HDMIB,
 					PORT_B);
-			if (I915_READ(VLV_DISPLAY_BASE + DP_B) & DP_DETECTED)
-				intel_dp_init(dev, VLV_DISPLAY_BASE + DP_B, PORT_B);
-		}
+		if (I915_READ(VLV_DISPLAY_BASE + DP_B) & DP_DETECTED ||
+		    intel_dp_is_edp(dev, PORT_B))
+			intel_dp_init(dev, VLV_DISPLAY_BASE + DP_B, PORT_B);
 
-		if (I915_READ(VLV_DISPLAY_BASE + GEN4_HDMIC) & SDVO_DETECTED) {
+		if (I915_READ(VLV_DISPLAY_BASE + GEN4_HDMIC) & SDVO_DETECTED)
 			intel_hdmi_init(dev, VLV_DISPLAY_BASE + GEN4_HDMIC,
 					PORT_C);
-			if (I915_READ(VLV_DISPLAY_BASE + DP_C) & DP_DETECTED)
-				intel_dp_init(dev, VLV_DISPLAY_BASE + DP_C, PORT_C);
-		}
+		if (I915_READ(VLV_DISPLAY_BASE + DP_C) & DP_DETECTED ||
+		    intel_dp_is_edp(dev, PORT_C))
+			intel_dp_init(dev, VLV_DISPLAY_BASE + DP_C, PORT_C);
 
 		if (IS_CHERRYVIEW(dev)) {
-			if (I915_READ(VLV_DISPLAY_BASE + CHV_HDMID) & SDVO_DETECTED) {
+			if (I915_READ(VLV_DISPLAY_BASE + CHV_HDMID) & SDVO_DETECTED)
 				intel_hdmi_init(dev, VLV_DISPLAY_BASE + CHV_HDMID,
 						PORT_D);
-				if (I915_READ(VLV_DISPLAY_BASE + DP_D) & DP_DETECTED)
-					intel_dp_init(dev, VLV_DISPLAY_BASE + DP_D, PORT_D);
-			}
+			/* eDP not supported on port D, so don't check VBT */
+			if (I915_READ(VLV_DISPLAY_BASE + DP_D) & DP_DETECTED)
+				intel_dp_init(dev, VLV_DISPLAY_BASE + DP_D, PORT_D);
 		}
 
 		intel_dsi_init(dev);
@@ -12064,9 +12556,9 @@ static void intel_setup_outputs(struct drm_device *dev)
 	if (SUPPORTS_TV(dev))
 		intel_tv_init(dev);
 
-	intel_edp_psr_init(dev);
+	intel_psr_init(dev);
 
-	list_for_each_entry(encoder, &dev->mode_config.encoder_list, base.head) {
+	for_each_intel_encoder(dev, encoder) {
 		encoder->base.possible_crtcs = encoder->crtc_mask;
 		encoder->base.possible_clones =
 			intel_encoder_clones(encoder);
@@ -12268,16 +12760,22 @@ static void intel_init_display(struct drm_device *dev)
 	if (HAS_DDI(dev)) {
 		dev_priv->display.get_pipe_config = haswell_get_pipe_config;
 		dev_priv->display.get_plane_config = ironlake_get_plane_config;
-		dev_priv->display.crtc_mode_set = haswell_crtc_mode_set;
+		dev_priv->display.crtc_compute_clock =
+			haswell_crtc_compute_clock;
 		dev_priv->display.crtc_enable = haswell_crtc_enable;
 		dev_priv->display.crtc_disable = haswell_crtc_disable;
 		dev_priv->display.off = ironlake_crtc_off;
-		dev_priv->display.update_primary_plane =
-			ironlake_update_primary_plane;
+		if (INTEL_INFO(dev)->gen >= 9)
+			dev_priv->display.update_primary_plane =
+				skylake_update_primary_plane;
+		else
+			dev_priv->display.update_primary_plane =
+				ironlake_update_primary_plane;
 	} else if (HAS_PCH_SPLIT(dev)) {
 		dev_priv->display.get_pipe_config = ironlake_get_pipe_config;
 		dev_priv->display.get_plane_config = ironlake_get_plane_config;
-		dev_priv->display.crtc_mode_set = ironlake_crtc_mode_set;
+		dev_priv->display.crtc_compute_clock =
+			ironlake_crtc_compute_clock;
 		dev_priv->display.crtc_enable = ironlake_crtc_enable;
 		dev_priv->display.crtc_disable = ironlake_crtc_disable;
 		dev_priv->display.off = ironlake_crtc_off;
@@ -12286,7 +12784,7 @@ static void intel_init_display(struct drm_device *dev)
 	} else if (IS_VALLEYVIEW(dev)) {
 		dev_priv->display.get_pipe_config = i9xx_get_pipe_config;
 		dev_priv->display.get_plane_config = i9xx_get_plane_config;
-		dev_priv->display.crtc_mode_set = i9xx_crtc_mode_set;
+		dev_priv->display.crtc_compute_clock = i9xx_crtc_compute_clock;
 		dev_priv->display.crtc_enable = valleyview_crtc_enable;
 		dev_priv->display.crtc_disable = i9xx_crtc_disable;
 		dev_priv->display.off = i9xx_crtc_off;
@@ -12295,7 +12793,7 @@ static void intel_init_display(struct drm_device *dev)
 	} else {
 		dev_priv->display.get_pipe_config = i9xx_get_pipe_config;
 		dev_priv->display.get_plane_config = i9xx_get_plane_config;
-		dev_priv->display.crtc_mode_set = i9xx_crtc_mode_set;
+		dev_priv->display.crtc_compute_clock = i9xx_crtc_compute_clock;
 		dev_priv->display.crtc_enable = i9xx_crtc_enable;
 		dev_priv->display.crtc_disable = i9xx_crtc_disable;
 		dev_priv->display.off = i9xx_crtc_off;
@@ -12332,33 +12830,20 @@ static void intel_init_display(struct drm_device *dev)
 		dev_priv->display.get_display_clock_speed =
 			i830_get_display_clock_speed;
 
-	if (HAS_PCH_SPLIT(dev)) {
-		if (IS_GEN5(dev)) {
-			dev_priv->display.fdi_link_train = ironlake_fdi_link_train;
-			dev_priv->display.write_eld = ironlake_write_eld;
-		} else if (IS_GEN6(dev)) {
-			dev_priv->display.fdi_link_train = gen6_fdi_link_train;
-			dev_priv->display.write_eld = ironlake_write_eld;
-			dev_priv->display.modeset_global_resources =
-				snb_modeset_global_resources;
-		} else if (IS_IVYBRIDGE(dev)) {
-			/* FIXME: detect B0+ stepping and use auto training */
-			dev_priv->display.fdi_link_train = ivb_manual_fdi_link_train;
-			dev_priv->display.write_eld = ironlake_write_eld;
-			dev_priv->display.modeset_global_resources =
-				ivb_modeset_global_resources;
-		} else if (IS_HASWELL(dev) || IS_GEN8(dev)) {
-			dev_priv->display.fdi_link_train = hsw_fdi_link_train;
-			dev_priv->display.write_eld = haswell_write_eld;
-			dev_priv->display.modeset_global_resources =
-				haswell_modeset_global_resources;
-		}
-	} else if (IS_G4X(dev)) {
-		dev_priv->display.write_eld = g4x_write_eld;
+	if (IS_GEN5(dev)) {
+		dev_priv->display.fdi_link_train = ironlake_fdi_link_train;
+	} else if (IS_GEN6(dev)) {
+		dev_priv->display.fdi_link_train = gen6_fdi_link_train;
+	} else if (IS_IVYBRIDGE(dev)) {
+		/* FIXME: detect B0+ stepping and use auto training */
+		dev_priv->display.fdi_link_train = ivb_manual_fdi_link_train;
+		dev_priv->display.modeset_global_resources =
+			ivb_modeset_global_resources;
+	} else if (IS_HASWELL(dev) || IS_BROADWELL(dev)) {
+		dev_priv->display.fdi_link_train = hsw_fdi_link_train;
 	} else if (IS_VALLEYVIEW(dev)) {
 		dev_priv->display.modeset_global_resources =
 			valleyview_modeset_global_resources;
-		dev_priv->display.write_eld = ironlake_write_eld;
 	}
 
 	/* Default just returns -ENODEV to indicate unsupported */
@@ -12385,9 +12870,14 @@ static void intel_init_display(struct drm_device *dev)
 	case 8: /* FIXME(BDW): Check that the gen8 RCS flip works. */
 		dev_priv->display.queue_flip = intel_gen7_queue_flip;
 		break;
+	case 9:
+		dev_priv->display.queue_flip = intel_gen9_queue_flip;
+		break;
 	}
 
 	intel_panel_init_backlight_funcs(dev);
+
+	mutex_init(&dev_priv->pps_mutex);
 }
 
 /*
@@ -12403,6 +12893,14 @@ static void quirk_pipea_force(struct drm_device *dev)
 	DRM_INFO("applying pipe a force quirk\n");
 }
 
+static void quirk_pipeb_force(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+
+	dev_priv->quirks |= QUIRK_PIPEB_FORCE;
+	DRM_INFO("applying pipe b force quirk\n");
+}
+
 /*
  * Some machines (Lenovo U160) do not work with SSC on LVDS for some reason
  */
@@ -12477,6 +12975,12 @@ static struct intel_quirk intel_quirks[] = {
 	/* ThinkPad T60 needs pipe A force quirk (bug #16494) */
 	{ 0x2782, 0x17aa, 0x201a, quirk_pipea_force },
 
+	/* 830 needs to leave pipe A & dpll A up */
+	{ 0x3577, PCI_ANY_ID, PCI_ANY_ID, quirk_pipea_force },
+
+	/* 830 needs to leave pipe B & dpll B up */
+	{ 0x3577, PCI_ANY_ID, PCI_ANY_ID, quirk_pipeb_force },
+
 	/* Lenovo U160 cannot use SSC on LVDS */
 	{ 0x0046, 0x17aa, 0x3920, quirk_ssc_force_disable },
 
@@ -12507,6 +13011,9 @@ static struct intel_quirk intel_quirks[] = {
 	/* Acer C720 Chromebook (Core i3 4005U) */
 	{ 0x0a16, 0x1025, 0x0a11, quirk_backlight_present },
 
+	/* Apple Macbook 2,1 (Core 2 T7400) */
+	{ 0x27a2, 0x8086, 0x7270, quirk_backlight_present },
+
 	/* Toshiba CB35 Chromebook (Celeron 2955U) */
 	{ 0x0a06, 0x1179, 0x0a88, quirk_backlight_present },
 
@@ -12563,16 +13070,9 @@ void intel_modeset_init_hw(struct drm_device *dev)
 
 	intel_init_clock_gating(dev);
 
-	intel_reset_dpio(dev);
-
 	intel_enable_gt_powersave(dev);
 }
 
-void intel_modeset_suspend_hw(struct drm_device *dev)
-{
-	intel_suspend_hw(dev);
-}
-
 void intel_modeset_init(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
@@ -12598,6 +13098,7 @@ void intel_modeset_init(struct drm_device *dev)
 		return;
 
 	intel_init_display(dev);
+	intel_init_audio(dev);
 
 	if (IS_GEN2(dev)) {
 		dev->mode_config.max_width = 2048;
@@ -12610,7 +13111,10 @@ void intel_modeset_init(struct drm_device *dev)
 		dev->mode_config.max_height = 8192;
 	}
 
-	if (IS_GEN2(dev)) {
+	if (IS_845G(dev) || IS_I865G(dev)) {
+		dev->mode_config.cursor_width = IS_845G(dev) ? 64 : 512;
+		dev->mode_config.cursor_height = 1023;
+	} else if (IS_GEN2(dev)) {
 		dev->mode_config.cursor_width = GEN2_CURSOR_WIDTH;
 		dev->mode_config.cursor_height = GEN2_CURSOR_HEIGHT;
 	} else {
@@ -12624,7 +13128,7 @@ void intel_modeset_init(struct drm_device *dev)
 		      INTEL_INFO(dev)->num_pipes,
 		      INTEL_INFO(dev)->num_pipes > 1 ? "s" : "");
 
-	for_each_pipe(pipe) {
+	for_each_pipe(dev_priv, pipe) {
 		intel_crtc_init(dev, pipe);
 		for_each_sprite(pipe, sprite) {
 			ret = intel_plane_init(dev, pipe, sprite);
@@ -12635,7 +13139,6 @@ void intel_modeset_init(struct drm_device *dev)
 	}
 
 	intel_init_dpio(dev);
-	intel_reset_dpio(dev);
 
 	intel_shared_dpll_init(dev);
 
@@ -12730,9 +13233,10 @@ static void intel_sanitize_crtc(struct intel_crtc *crtc)
 	I915_WRITE(reg, I915_READ(reg) & ~PIPECONF_FRAME_START_DELAY_MASK);
 
 	/* restore vblank interrupts to correct state */
-	if (crtc->active)
+	if (crtc->active) {
+		update_scanline_offset(crtc);
 		drm_vblank_on(dev, crtc->pipe);
-	else
+	} else
 		drm_vblank_off(dev, crtc->pipe);
 
 	/* We need to sanitize the plane -> pipe mapping first because this will
@@ -12815,7 +13319,7 @@ static void intel_sanitize_crtc(struct intel_crtc *crtc)
 		}
 	}
 
-	if (crtc->active || IS_VALLEYVIEW(dev) || INTEL_INFO(dev)->gen < 5) {
+	if (crtc->active || HAS_GMCH_DISPLAY(dev)) {
 		/*
 		 * We start out with underrun reporting disabled to avoid races.
 		 * For correct bookkeeping mark this on active crtcs.
@@ -12831,8 +13335,6 @@ static void intel_sanitize_crtc(struct intel_crtc *crtc)
 		 */
 		crtc->cpu_fifo_underrun_disabled = true;
 		crtc->pch_fifo_underrun_disabled = true;
-
-		update_scanline_offset(crtc);
 	}
 }
 
@@ -12905,7 +13407,7 @@ void i915_redisable_vga(struct drm_device *dev)
 	 * level, just check if the power well is enabled instead of trying to
 	 * follow the "don't touch the power well if we don't need it" policy
 	 * the rest of the driver uses. */
-	if (!intel_display_power_enabled(dev_priv, POWER_DOMAIN_VGA))
+	if (!intel_display_power_is_enabled(dev_priv, POWER_DOMAIN_VGA))
 		return;
 
 	i915_redisable_vga_power_on(dev);
@@ -12949,23 +13451,25 @@ static void intel_modeset_readout_hw_state(struct drm_device *dev)
 	for (i = 0; i < dev_priv->num_shared_dpll; i++) {
 		struct intel_shared_dpll *pll = &dev_priv->shared_dplls[i];
 
-		pll->on = pll->get_hw_state(dev_priv, pll, &pll->hw_state);
+		pll->on = pll->get_hw_state(dev_priv, pll,
+					    &pll->config.hw_state);
 		pll->active = 0;
+		pll->config.crtc_mask = 0;
 		for_each_intel_crtc(dev, crtc) {
-			if (crtc->active && intel_crtc_to_shared_dpll(crtc) == pll)
+			if (crtc->active && intel_crtc_to_shared_dpll(crtc) == pll) {
 				pll->active++;
+				pll->config.crtc_mask |= 1 << crtc->pipe;
+			}
 		}
-		pll->refcount = pll->active;
 
-		DRM_DEBUG_KMS("%s hw state readout: refcount %i, on %i\n",
-			      pll->name, pll->refcount, pll->on);
+		DRM_DEBUG_KMS("%s hw state readout: crtc_mask 0x%08x, on %i\n",
+			      pll->name, pll->config.crtc_mask, pll->on);
 
-		if (pll->refcount)
+		if (pll->config.crtc_mask)
 			intel_display_power_get(dev_priv, POWER_DOMAIN_PLLS);
 	}
 
-	list_for_each_entry(encoder, &dev->mode_config.encoder_list,
-			    base.head) {
+	for_each_intel_encoder(dev, encoder) {
 		pipe = 0;
 
 		if (encoder->get_hw_state(encoder, &pipe)) {
@@ -13029,12 +13533,11 @@ void intel_modeset_setup_hw_state(struct drm_device *dev,
 	}
 
 	/* HW state is read out, now we need to sanitize this mess. */
-	list_for_each_entry(encoder, &dev->mode_config.encoder_list,
-			    base.head) {
+	for_each_intel_encoder(dev, encoder) {
 		intel_sanitize_encoder(encoder);
 	}
 
-	for_each_pipe(pipe) {
+	for_each_pipe(dev_priv, pipe) {
 		crtc = to_intel_crtc(dev_priv->pipe_to_crtc_mapping[pipe]);
 		intel_sanitize_crtc(crtc);
 		intel_dump_pipe_config(crtc, &crtc->config, "[setup_hw_state]");
@@ -13052,7 +13555,9 @@ void intel_modeset_setup_hw_state(struct drm_device *dev,
 		pll->on = false;
 	}
 
-	if (HAS_PCH_SPLIT(dev))
+	if (IS_GEN9(dev))
+		skl_wm_get_hw_state(dev);
+	else if (HAS_PCH_SPLIT(dev))
 		ilk_wm_get_hw_state(dev);
 
 	if (force_restore) {
@@ -13062,12 +13567,12 @@ void intel_modeset_setup_hw_state(struct drm_device *dev,
 		 * We need to use raw interfaces for restoring state to avoid
 		 * checking (bogus) intermediate states.
 		 */
-		for_each_pipe(pipe) {
+		for_each_pipe(dev_priv, pipe) {
 			struct drm_crtc *crtc =
 				dev_priv->pipe_to_crtc_mapping[pipe];
 
-			__intel_set_mode(crtc, &crtc->mode, crtc->x, crtc->y,
-					 crtc->primary->fb);
+			intel_set_mode(crtc, &crtc->mode, crtc->x, crtc->y,
+				       crtc->primary->fb);
 		}
 	} else {
 		intel_modeset_update_staged_output_state(dev);
@@ -13078,6 +13583,7 @@ void intel_modeset_setup_hw_state(struct drm_device *dev,
 
 void intel_modeset_gem_init(struct drm_device *dev)
 {
+	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct drm_crtc *c;
 	struct drm_i915_gem_object *obj;
 
@@ -13085,6 +13591,16 @@ void intel_modeset_gem_init(struct drm_device *dev)
 	intel_init_gt_powersave(dev);
 	mutex_unlock(&dev->struct_mutex);
 
+	/*
+	 * There may be no VBT; and if the BIOS enabled SSC we can
+	 * just keep using it to avoid unnecessary flicker.  Whereas if the
+	 * BIOS isn't using it, don't assume it will work even if the VBT
+	 * indicates as much.
+	 */
+	if (HAS_PCH_IBX(dev) || HAS_PCH_CPT(dev))
+		dev_priv->vbt.lvds_use_ssc = !!(I915_READ(PCH_DREF_CONTROL) &
+						DREF_SSC1_ENABLE);
+
 	intel_modeset_init_hw(dev);
 
 	intel_setup_overlay(dev);
@@ -13100,7 +13616,9 @@ void intel_modeset_gem_init(struct drm_device *dev)
 		if (obj == NULL)
 			continue;
 
-		if (intel_pin_and_fence_fb_obj(dev, obj, NULL)) {
+		if (intel_pin_and_fence_fb_obj(c->primary,
+					       c->primary->fb,
+					       NULL)) {
 			DRM_ERROR("failed to pin boot fb on pipe %d\n",
 				  to_intel_crtc(c)->pipe);
 			drm_framebuffer_unreference(c->primary->fb);
@@ -13108,6 +13626,8 @@ void intel_modeset_gem_init(struct drm_device *dev)
 		}
 	}
 	mutex_unlock(&dev->struct_mutex);
+
+	intel_backlight_register(dev);
 }
 
 void intel_connector_unregister(struct intel_connector *intel_connector)
@@ -13123,14 +13643,16 @@ void intel_modeset_cleanup(struct drm_device *dev)
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct drm_connector *connector;
 
+	intel_disable_gt_powersave(dev);
+
+	intel_backlight_unregister(dev);
+
 	/*
 	 * Interrupts and polling as the first thing to avoid creating havoc.
-	 * Too much stuff here (turning of rps, connectors, ...) would
+	 * Too much stuff here (turning of connectors, ...) would
 	 * experience fancy races otherwise.
 	 */
-	drm_irq_uninstall(dev);
-	intel_hpd_cancel_work(dev_priv);
-	dev_priv->pm._irqs_disabled = true;
+	intel_irq_uninstall(dev_priv);
 
 	/*
 	 * Due to the hpd irq storm handling the hotplug work can re-arm the
@@ -13144,8 +13666,6 @@ void intel_modeset_cleanup(struct drm_device *dev)
 
 	intel_disable_fbc(dev);
 
-	intel_disable_gt_powersave(dev);
-
 	ironlake_teardown_rc6(dev);
 
 	mutex_unlock(&dev->struct_mutex);
@@ -13283,10 +13803,10 @@ intel_display_capture_error_state(struct drm_device *dev)
 	if (IS_HASWELL(dev) || IS_BROADWELL(dev))
 		error->power_well_driver = I915_READ(HSW_PWR_WELL_DRIVER);
 
-	for_each_pipe(i) {
+	for_each_pipe(dev_priv, i) {
 		error->pipe[i].power_domain_on =
-			intel_display_power_enabled_unlocked(dev_priv,
-							   POWER_DOMAIN_PIPE(i));
+			__intel_display_power_is_enabled(dev_priv,
+							 POWER_DOMAIN_PIPE(i));
 		if (!error->pipe[i].power_domain_on)
 			continue;
 
@@ -13321,7 +13841,7 @@ intel_display_capture_error_state(struct drm_device *dev)
 		enum transcoder cpu_transcoder = transcoders[i];
 
 		error->transcoder[i].power_domain_on =
-			intel_display_power_enabled_unlocked(dev_priv,
+			__intel_display_power_is_enabled(dev_priv,
 				POWER_DOMAIN_TRANSCODER(cpu_transcoder));
 		if (!error->transcoder[i].power_domain_on)
 			continue;
@@ -13347,6 +13867,7 @@ intel_display_print_error_state(struct drm_i915_error_state_buf *m,
 				struct drm_device *dev,
 				struct intel_display_error_state *error)
 {
+	struct drm_i915_private *dev_priv = dev->dev_private;
 	int i;
 
 	if (!error)
@@ -13356,7 +13877,7 @@ intel_display_print_error_state(struct drm_i915_error_state_buf *m,
 	if (IS_HASWELL(dev) || IS_BROADWELL(dev))
 		err_printf(m, "PWR_WELL_CTL2: %08x\n",
 			   error->power_well_driver);
-	for_each_pipe(i) {
+	for_each_pipe(dev_priv, i) {
 		err_printf(m, "Pipe [%d]:\n", i);
 		err_printf(m, "  Power: %s\n",
 			   error->pipe[i].power_domain_on ? "on" : "off");
@@ -13397,3 +13918,24 @@ intel_display_print_error_state(struct drm_i915_error_state_buf *m,
 		err_printf(m, "  VSYNC: %08x\n", error->transcoder[i].vsync);
 	}
 }
+
+void intel_modeset_preclose(struct drm_device *dev, struct drm_file *file)
+{
+	struct intel_crtc *crtc;
+
+	for_each_intel_crtc(dev, crtc) {
+		struct intel_unpin_work *work;
+
+		spin_lock_irq(&dev->event_lock);
+
+		work = crtc->unpin_work;
+
+		if (work && work->event &&
+		    work->event->base.file_priv == file) {
+			kfree(work->event);
+			work->event = NULL;
+		}
+
+		spin_unlock_irq(&dev->event_lock);
+	}
+}
diff --git a/drivers/gpu/drm/i915/intel_dp.c b/drivers/gpu/drm/i915/intel_dp.c
index fdff1d420c14..5cecc20efa71 100644
--- a/drivers/gpu/drm/i915/intel_dp.c
+++ b/drivers/gpu/drm/i915/intel_dp.c
@@ -111,8 +111,11 @@ static struct intel_dp *intel_attached_dp(struct drm_connector *connector)
 }
 
 static void intel_dp_link_down(struct intel_dp *intel_dp);
-static bool _edp_panel_vdd_on(struct intel_dp *intel_dp);
+static bool edp_panel_vdd_on(struct intel_dp *intel_dp);
 static void edp_panel_vdd_off(struct intel_dp *intel_dp, bool sync);
+static void vlv_init_panel_power_sequencer(struct intel_dp *intel_dp);
+static void vlv_steal_power_sequencer(struct drm_device *dev,
+				      enum pipe pipe);
 
 int
 intel_dp_max_link_bw(struct intel_dp *intel_dp)
@@ -224,8 +227,7 @@ intel_dp_mode_valid(struct drm_connector *connector,
 	return MODE_OK;
 }
 
-static uint32_t
-pack_aux(uint8_t *src, int src_bytes)
+uint32_t intel_dp_pack_aux(const uint8_t *src, int src_bytes)
 {
 	int	i;
 	uint32_t v = 0;
@@ -237,8 +239,7 @@ pack_aux(uint8_t *src, int src_bytes)
 	return v;
 }
 
-static void
-unpack_aux(uint32_t src, uint8_t *dst, int dst_bytes)
+void intel_dp_unpack_aux(uint32_t src, uint8_t *dst, int dst_bytes)
 {
 	int i;
 	if (dst_bytes > 4)
@@ -283,39 +284,275 @@ intel_hrawclk(struct drm_device *dev)
 
 static void
 intel_dp_init_panel_power_sequencer(struct drm_device *dev,
-				    struct intel_dp *intel_dp,
-				    struct edp_power_seq *out);
+				    struct intel_dp *intel_dp);
 static void
 intel_dp_init_panel_power_sequencer_registers(struct drm_device *dev,
-					      struct intel_dp *intel_dp,
-					      struct edp_power_seq *out);
+					      struct intel_dp *intel_dp);
+
+static void pps_lock(struct intel_dp *intel_dp)
+{
+	struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp);
+	struct intel_encoder *encoder = &intel_dig_port->base;
+	struct drm_device *dev = encoder->base.dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	enum intel_display_power_domain power_domain;
+
+	/*
+	 * See vlv_power_sequencer_reset() why we need
+	 * a power domain reference here.
+	 */
+	power_domain = intel_display_port_power_domain(encoder);
+	intel_display_power_get(dev_priv, power_domain);
+
+	mutex_lock(&dev_priv->pps_mutex);
+}
+
+static void pps_unlock(struct intel_dp *intel_dp)
+{
+	struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp);
+	struct intel_encoder *encoder = &intel_dig_port->base;
+	struct drm_device *dev = encoder->base.dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	enum intel_display_power_domain power_domain;
+
+	mutex_unlock(&dev_priv->pps_mutex);
+
+	power_domain = intel_display_port_power_domain(encoder);
+	intel_display_power_put(dev_priv, power_domain);
+}
+
+static void
+vlv_power_sequencer_kick(struct intel_dp *intel_dp)
+{
+	struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp);
+	struct drm_device *dev = intel_dig_port->base.base.dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	enum pipe pipe = intel_dp->pps_pipe;
+	bool pll_enabled;
+	uint32_t DP;
+
+	if (WARN(I915_READ(intel_dp->output_reg) & DP_PORT_EN,
+		 "skipping pipe %c power seqeuncer kick due to port %c being active\n",
+		 pipe_name(pipe), port_name(intel_dig_port->port)))
+		return;
+
+	DRM_DEBUG_KMS("kicking pipe %c power sequencer for port %c\n",
+		      pipe_name(pipe), port_name(intel_dig_port->port));
+
+	/* Preserve the BIOS-computed detected bit. This is
+	 * supposed to be read-only.
+	 */
+	DP = I915_READ(intel_dp->output_reg) & DP_DETECTED;
+	DP |= DP_VOLTAGE_0_4 | DP_PRE_EMPHASIS_0;
+	DP |= DP_PORT_WIDTH(1);
+	DP |= DP_LINK_TRAIN_PAT_1;
+
+	if (IS_CHERRYVIEW(dev))
+		DP |= DP_PIPE_SELECT_CHV(pipe);
+	else if (pipe == PIPE_B)
+		DP |= DP_PIPEB_SELECT;
+
+	pll_enabled = I915_READ(DPLL(pipe)) & DPLL_VCO_ENABLE;
+
+	/*
+	 * The DPLL for the pipe must be enabled for this to work.
+	 * So enable temporarily it if it's not already enabled.
+	 */
+	if (!pll_enabled)
+		vlv_force_pll_on(dev, pipe, IS_CHERRYVIEW(dev) ?
+				 &chv_dpll[0].dpll : &vlv_dpll[0].dpll);
+
+	/*
+	 * Similar magic as in intel_dp_enable_port().
+	 * We _must_ do this port enable + disable trick
+	 * to make this power seqeuencer lock onto the port.
+	 * Otherwise even VDD force bit won't work.
+	 */
+	I915_WRITE(intel_dp->output_reg, DP);
+	POSTING_READ(intel_dp->output_reg);
+
+	I915_WRITE(intel_dp->output_reg, DP | DP_PORT_EN);
+	POSTING_READ(intel_dp->output_reg);
+
+	I915_WRITE(intel_dp->output_reg, DP & ~DP_PORT_EN);
+	POSTING_READ(intel_dp->output_reg);
+
+	if (!pll_enabled)
+		vlv_force_pll_off(dev, pipe);
+}
 
 static enum pipe
 vlv_power_sequencer_pipe(struct intel_dp *intel_dp)
 {
 	struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp);
-	struct drm_crtc *crtc = intel_dig_port->base.base.crtc;
 	struct drm_device *dev = intel_dig_port->base.base.dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	enum port port = intel_dig_port->port;
+	struct intel_encoder *encoder;
+	unsigned int pipes = (1 << PIPE_A) | (1 << PIPE_B);
 	enum pipe pipe;
 
-	/* modeset should have pipe */
-	if (crtc)
-		return to_intel_crtc(crtc)->pipe;
+	lockdep_assert_held(&dev_priv->pps_mutex);
+
+	/* We should never land here with regular DP ports */
+	WARN_ON(!is_edp(intel_dp));
+
+	if (intel_dp->pps_pipe != INVALID_PIPE)
+		return intel_dp->pps_pipe;
+
+	/*
+	 * We don't have power sequencer currently.
+	 * Pick one that's not used by other ports.
+	 */
+	list_for_each_entry(encoder, &dev->mode_config.encoder_list,
+			    base.head) {
+		struct intel_dp *tmp;
+
+		if (encoder->type != INTEL_OUTPUT_EDP)
+			continue;
+
+		tmp = enc_to_intel_dp(&encoder->base);
+
+		if (tmp->pps_pipe != INVALID_PIPE)
+			pipes &= ~(1 << tmp->pps_pipe);
+	}
+
+	/*
+	 * Didn't find one. This should not happen since there
+	 * are two power sequencers and up to two eDP ports.
+	 */
+	if (WARN_ON(pipes == 0))
+		pipe = PIPE_A;
+	else
+		pipe = ffs(pipes) - 1;
+
+	vlv_steal_power_sequencer(dev, pipe);
+	intel_dp->pps_pipe = pipe;
+
+	DRM_DEBUG_KMS("picked pipe %c power sequencer for port %c\n",
+		      pipe_name(intel_dp->pps_pipe),
+		      port_name(intel_dig_port->port));
+
+	/* init power sequencer on this pipe and port */
+	intel_dp_init_panel_power_sequencer(dev, intel_dp);
+	intel_dp_init_panel_power_sequencer_registers(dev, intel_dp);
+
+	/*
+	 * Even vdd force doesn't work until we've made
+	 * the power sequencer lock in on the port.
+	 */
+	vlv_power_sequencer_kick(intel_dp);
+
+	return intel_dp->pps_pipe;
+}
+
+typedef bool (*vlv_pipe_check)(struct drm_i915_private *dev_priv,
+			       enum pipe pipe);
+
+static bool vlv_pipe_has_pp_on(struct drm_i915_private *dev_priv,
+			       enum pipe pipe)
+{
+	return I915_READ(VLV_PIPE_PP_STATUS(pipe)) & PP_ON;
+}
+
+static bool vlv_pipe_has_vdd_on(struct drm_i915_private *dev_priv,
+				enum pipe pipe)
+{
+	return I915_READ(VLV_PIPE_PP_CONTROL(pipe)) & EDP_FORCE_VDD;
+}
+
+static bool vlv_pipe_any(struct drm_i915_private *dev_priv,
+			 enum pipe pipe)
+{
+	return true;
+}
+
+static enum pipe
+vlv_initial_pps_pipe(struct drm_i915_private *dev_priv,
+		     enum port port,
+		     vlv_pipe_check pipe_check)
+{
+	enum pipe pipe;
 
-	/* init time, try to find a pipe with this port selected */
 	for (pipe = PIPE_A; pipe <= PIPE_B; pipe++) {
 		u32 port_sel = I915_READ(VLV_PIPE_PP_ON_DELAYS(pipe)) &
 			PANEL_PORT_SELECT_MASK;
-		if (port_sel == PANEL_PORT_SELECT_DPB_VLV && port == PORT_B)
-			return pipe;
-		if (port_sel == PANEL_PORT_SELECT_DPC_VLV && port == PORT_C)
-			return pipe;
+
+		if (port_sel != PANEL_PORT_SELECT_VLV(port))
+			continue;
+
+		if (!pipe_check(dev_priv, pipe))
+			continue;
+
+		return pipe;
 	}
 
-	/* shrug */
-	return PIPE_A;
+	return INVALID_PIPE;
+}
+
+static void
+vlv_initial_power_sequencer_setup(struct intel_dp *intel_dp)
+{
+	struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp);
+	struct drm_device *dev = intel_dig_port->base.base.dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	enum port port = intel_dig_port->port;
+
+	lockdep_assert_held(&dev_priv->pps_mutex);
+
+	/* try to find a pipe with this port selected */
+	/* first pick one where the panel is on */
+	intel_dp->pps_pipe = vlv_initial_pps_pipe(dev_priv, port,
+						  vlv_pipe_has_pp_on);
+	/* didn't find one? pick one where vdd is on */
+	if (intel_dp->pps_pipe == INVALID_PIPE)
+		intel_dp->pps_pipe = vlv_initial_pps_pipe(dev_priv, port,
+							  vlv_pipe_has_vdd_on);
+	/* didn't find one? pick one with just the correct port */
+	if (intel_dp->pps_pipe == INVALID_PIPE)
+		intel_dp->pps_pipe = vlv_initial_pps_pipe(dev_priv, port,
+							  vlv_pipe_any);
+
+	/* didn't find one? just let vlv_power_sequencer_pipe() pick one when needed */
+	if (intel_dp->pps_pipe == INVALID_PIPE) {
+		DRM_DEBUG_KMS("no initial power sequencer for port %c\n",
+			      port_name(port));
+		return;
+	}
+
+	DRM_DEBUG_KMS("initial power sequencer for port %c: pipe %c\n",
+		      port_name(port), pipe_name(intel_dp->pps_pipe));
+
+	intel_dp_init_panel_power_sequencer(dev, intel_dp);
+	intel_dp_init_panel_power_sequencer_registers(dev, intel_dp);
+}
+
+void vlv_power_sequencer_reset(struct drm_i915_private *dev_priv)
+{
+	struct drm_device *dev = dev_priv->dev;
+	struct intel_encoder *encoder;
+
+	if (WARN_ON(!IS_VALLEYVIEW(dev)))
+		return;
+
+	/*
+	 * We can't grab pps_mutex here due to deadlock with power_domain
+	 * mutex when power_domain functions are called while holding pps_mutex.
+	 * That also means that in order to use pps_pipe the code needs to
+	 * hold both a power domain reference and pps_mutex, and the power domain
+	 * reference get/put must be done while _not_ holding pps_mutex.
+	 * pps_{lock,unlock}() do these steps in the correct order, so one
+	 * should use them always.
+	 */
+
+	list_for_each_entry(encoder, &dev->mode_config.encoder_list, base.head) {
+		struct intel_dp *intel_dp;
+
+		if (encoder->type != INTEL_OUTPUT_EDP)
+			continue;
+
+		intel_dp = enc_to_intel_dp(&encoder->base);
+		intel_dp->pps_pipe = INVALID_PIPE;
+	}
 }
 
 static u32 _pp_ctrl_reg(struct intel_dp *intel_dp)
@@ -349,12 +586,15 @@ static int edp_notify_handler(struct notifier_block *this, unsigned long code,
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	u32 pp_div;
 	u32 pp_ctrl_reg, pp_div_reg;
-	enum pipe pipe = vlv_power_sequencer_pipe(intel_dp);
 
 	if (!is_edp(intel_dp) || code != SYS_RESTART)
 		return 0;
 
+	pps_lock(intel_dp);
+
 	if (IS_VALLEYVIEW(dev)) {
+		enum pipe pipe = vlv_power_sequencer_pipe(intel_dp);
+
 		pp_ctrl_reg = VLV_PIPE_PP_CONTROL(pipe);
 		pp_div_reg  = VLV_PIPE_PP_DIVISOR(pipe);
 		pp_div = I915_READ(pp_div_reg);
@@ -366,6 +606,8 @@ static int edp_notify_handler(struct notifier_block *this, unsigned long code,
 		msleep(intel_dp->panel_power_cycle_delay);
 	}
 
+	pps_unlock(intel_dp);
+
 	return 0;
 }
 
@@ -374,6 +616,12 @@ static bool edp_have_panel_power(struct intel_dp *intel_dp)
 	struct drm_device *dev = intel_dp_to_dev(intel_dp);
 	struct drm_i915_private *dev_priv = dev->dev_private;
 
+	lockdep_assert_held(&dev_priv->pps_mutex);
+
+	if (IS_VALLEYVIEW(dev) &&
+	    intel_dp->pps_pipe == INVALID_PIPE)
+		return false;
+
 	return (I915_READ(_pp_stat_reg(intel_dp)) & PP_ON) != 0;
 }
 
@@ -381,13 +629,14 @@ static bool edp_have_panel_vdd(struct intel_dp *intel_dp)
 {
 	struct drm_device *dev = intel_dp_to_dev(intel_dp);
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp);
-	struct intel_encoder *intel_encoder = &intel_dig_port->base;
-	enum intel_display_power_domain power_domain;
 
-	power_domain = intel_display_port_power_domain(intel_encoder);
-	return intel_display_power_enabled(dev_priv, power_domain) &&
-	       (I915_READ(_pp_ctrl_reg(intel_dp)) & EDP_FORCE_VDD) != 0;
+	lockdep_assert_held(&dev_priv->pps_mutex);
+
+	if (IS_VALLEYVIEW(dev) &&
+	    intel_dp->pps_pipe == INVALID_PIPE)
+		return false;
+
+	return I915_READ(_pp_ctrl_reg(intel_dp)) & EDP_FORCE_VDD;
 }
 
 static void
@@ -488,6 +737,16 @@ static uint32_t vlv_get_aux_clock_divider(struct intel_dp *intel_dp, int index)
 	return index ? 0 : 100;
 }
 
+static uint32_t skl_get_aux_clock_divider(struct intel_dp *intel_dp, int index)
+{
+	/*
+	 * SKL doesn't need us to program the AUX clock divider (Hardware will
+	 * derive the clock from CDCLK automatically). We still implement the
+	 * get_aux_clock_divider vfunc to plug-in into the existing code.
+	 */
+	return index ? 0 : 1;
+}
+
 static uint32_t i9xx_get_aux_send_ctl(struct intel_dp *intel_dp,
 				      bool has_aux_irq,
 				      int send_bytes,
@@ -518,9 +777,24 @@ static uint32_t i9xx_get_aux_send_ctl(struct intel_dp *intel_dp,
 	       (aux_clock_divider << DP_AUX_CH_CTL_BIT_CLOCK_2X_SHIFT);
 }
 
+static uint32_t skl_get_aux_send_ctl(struct intel_dp *intel_dp,
+				      bool has_aux_irq,
+				      int send_bytes,
+				      uint32_t unused)
+{
+	return DP_AUX_CH_CTL_SEND_BUSY |
+	       DP_AUX_CH_CTL_DONE |
+	       (has_aux_irq ? DP_AUX_CH_CTL_INTERRUPT : 0) |
+	       DP_AUX_CH_CTL_TIME_OUT_ERROR |
+	       DP_AUX_CH_CTL_TIME_OUT_1600us |
+	       DP_AUX_CH_CTL_RECEIVE_ERROR |
+	       (send_bytes << DP_AUX_CH_CTL_MESSAGE_SIZE_SHIFT) |
+	       DP_AUX_CH_CTL_SYNC_PULSE_SKL(32);
+}
+
 static int
 intel_dp_aux_ch(struct intel_dp *intel_dp,
-		uint8_t *send, int send_bytes,
+		const uint8_t *send, int send_bytes,
 		uint8_t *recv, int recv_size)
 {
 	struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp);
@@ -535,7 +809,15 @@ intel_dp_aux_ch(struct intel_dp *intel_dp,
 	bool has_aux_irq = HAS_AUX_IRQ(dev);
 	bool vdd;
 
-	vdd = _edp_panel_vdd_on(intel_dp);
+	pps_lock(intel_dp);
+
+	/*
+	 * We will be called with VDD already enabled for dpcd/edid/oui reads.
+	 * In such cases we want to leave VDD enabled and it's up to upper layers
+	 * to turn it off. But for eg. i2c-dev access we need to turn it on/off
+	 * ourselves.
+	 */
+	vdd = edp_panel_vdd_on(intel_dp);
 
 	/* dp aux is extremely sensitive to irq latency, hence request the
 	 * lowest possible wakeup latency and so prevent the cpu from going into
@@ -579,7 +861,8 @@ intel_dp_aux_ch(struct intel_dp *intel_dp,
 			/* Load the send data into the aux channel data registers */
 			for (i = 0; i < send_bytes; i += 4)
 				I915_WRITE(ch_data + i,
-					   pack_aux(send + i, send_bytes - i));
+					   intel_dp_pack_aux(send + i,
+							     send_bytes - i));
 
 			/* Send the command and wait for it to complete */
 			I915_WRITE(ch_ctl, send_ctl);
@@ -633,8 +916,8 @@ intel_dp_aux_ch(struct intel_dp *intel_dp,
 		recv_bytes = recv_size;
 
 	for (i = 0; i < recv_bytes; i += 4)
-		unpack_aux(I915_READ(ch_data + i),
-			   recv + i, recv_bytes - i);
+		intel_dp_unpack_aux(I915_READ(ch_data + i),
+				    recv + i, recv_bytes - i);
 
 	ret = recv_bytes;
 out:
@@ -644,6 +927,8 @@ out:
 	if (vdd)
 		edp_panel_vdd_off(intel_dp, false);
 
+	pps_unlock(intel_dp);
+
 	return ret;
 }
 
@@ -742,7 +1027,16 @@ intel_dp_aux_init(struct intel_dp *intel_dp, struct intel_connector *connector)
 		BUG();
 	}
 
-	if (!HAS_DDI(dev))
+	/*
+	 * The AUX_CTL register is usually DP_CTL + 0x10.
+	 *
+	 * On Haswell and Broadwell though:
+	 *   - Both port A DDI_BUF_CTL and DDI_AUX_CTL are on the CPU
+	 *   - Port B/C/D AUX channels are on the PCH, DDI_BUF_CTL on the CPU
+	 *
+	 * Skylake moves AUX_CTL back next to DDI_BUF_CTL, on the CPU.
+	 */
+	if (!IS_HASWELL(dev) && !IS_BROADWELL(dev))
 		intel_dp->aux_ch_ctl_reg = intel_dp->output_reg + 0x10;
 
 	intel_dp->aux.name = name;
@@ -780,6 +1074,33 @@ intel_dp_connector_unregister(struct intel_connector *intel_connector)
 }
 
 static void
+skl_edp_set_pll_config(struct intel_crtc_config *pipe_config, int link_bw)
+{
+	u32 ctrl1;
+
+	pipe_config->ddi_pll_sel = SKL_DPLL0;
+	pipe_config->dpll_hw_state.cfgcr1 = 0;
+	pipe_config->dpll_hw_state.cfgcr2 = 0;
+
+	ctrl1 = DPLL_CTRL1_OVERRIDE(SKL_DPLL0);
+	switch (link_bw) {
+	case DP_LINK_BW_1_62:
+		ctrl1 |= DPLL_CRTL1_LINK_RATE(DPLL_CRTL1_LINK_RATE_810,
+					      SKL_DPLL0);
+		break;
+	case DP_LINK_BW_2_7:
+		ctrl1 |= DPLL_CRTL1_LINK_RATE(DPLL_CRTL1_LINK_RATE_1350,
+					      SKL_DPLL0);
+		break;
+	case DP_LINK_BW_5_4:
+		ctrl1 |= DPLL_CRTL1_LINK_RATE(DPLL_CRTL1_LINK_RATE_2700,
+					      SKL_DPLL0);
+		break;
+	}
+	pipe_config->dpll_hw_state.ctrl1 = ctrl1;
+}
+
+static void
 hsw_dp_set_ddi_pll_sel(struct intel_crtc_config *pipe_config, int link_bw)
 {
 	switch (link_bw) {
@@ -828,20 +1149,6 @@ intel_dp_set_clock(struct intel_encoder *encoder,
 	}
 }
 
-static void
-intel_dp_set_m2_n2(struct intel_crtc *crtc, struct intel_link_m_n *m_n)
-{
-	struct drm_device *dev = crtc->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	enum transcoder transcoder = crtc->config.cpu_transcoder;
-
-	I915_WRITE(PIPE_DATA_M2(transcoder),
-		TU_SIZE(m_n->tu) | m_n->gmch_m);
-	I915_WRITE(PIPE_DATA_N2(transcoder), m_n->gmch_n);
-	I915_WRITE(PIPE_LINK_M2(transcoder), m_n->link_m);
-	I915_WRITE(PIPE_LINK_N2(transcoder), m_n->link_n);
-}
-
 bool
 intel_dp_compute_config(struct intel_encoder *encoder,
 			struct intel_crtc_config *pipe_config)
@@ -867,6 +1174,7 @@ intel_dp_compute_config(struct intel_encoder *encoder,
 		pipe_config->has_pch_encoder = true;
 
 	pipe_config->has_dp_encoder = true;
+	pipe_config->has_drrs = false;
 	pipe_config->has_audio = intel_dp->has_audio;
 
 	if (is_edp(intel_dp) && intel_connector->panel.fixed_mode) {
@@ -898,23 +1206,15 @@ intel_dp_compute_config(struct intel_encoder *encoder,
 			bpp = dev_priv->vbt.edp_bpp;
 		}
 
-		if (IS_BROADWELL(dev)) {
-			/* Yes, it's an ugly hack. */
-			min_lane_count = max_lane_count;
-			DRM_DEBUG_KMS("forcing lane count to max (%u) on BDW\n",
-				      min_lane_count);
-		} else if (dev_priv->vbt.edp_lanes) {
-			min_lane_count = min(dev_priv->vbt.edp_lanes,
-					     max_lane_count);
-			DRM_DEBUG_KMS("using min %u lanes per VBT\n",
-				      min_lane_count);
-		}
-
-		if (dev_priv->vbt.edp_rate) {
-			min_clock = min(dev_priv->vbt.edp_rate >> 3, max_clock);
-			DRM_DEBUG_KMS("using min %02x link bw per VBT\n",
-				      bws[min_clock]);
-		}
+		/*
+		 * Use the maximum clock and number of lanes the eDP panel
+		 * advertizes being capable of. The panels are generally
+		 * designed to support only a single clock and lane
+		 * configuration, and typically these values correspond to the
+		 * native resolution of the panel.
+		 */
+		min_lane_count = max_lane_count;
+		min_clock = max_clock;
 	}
 
 	for (; bpp >= 6*3; bpp -= 2*3) {
@@ -970,13 +1270,16 @@ found:
 
 	if (intel_connector->panel.downclock_mode != NULL &&
 		intel_dp->drrs_state.type == SEAMLESS_DRRS_SUPPORT) {
+			pipe_config->has_drrs = true;
 			intel_link_compute_m_n(bpp, lane_count,
 				intel_connector->panel.downclock_mode->clock,
 				pipe_config->port_clock,
 				&pipe_config->dp_m2_n2);
 	}
 
-	if (HAS_DDI(dev))
+	if (IS_SKYLAKE(dev) && is_edp(intel_dp))
+		skl_edp_set_pll_config(pipe_config, intel_dp->link_bw);
+	else if (IS_HASWELL(dev) || IS_BROADWELL(dev))
 		hsw_dp_set_ddi_pll_sel(pipe_config, intel_dp->link_bw);
 	else
 		intel_dp_set_clock(encoder, pipe_config, intel_dp->link_bw);
@@ -1049,12 +1352,8 @@ static void intel_dp_prepare(struct intel_encoder *encoder)
 	intel_dp->DP |= DP_VOLTAGE_0_4 | DP_PRE_EMPHASIS_0;
 	intel_dp->DP |= DP_PORT_WIDTH(intel_dp->lane_count);
 
-	if (crtc->config.has_audio) {
-		DRM_DEBUG_DRIVER("Enabling DP audio on pipe %c\n",
-				 pipe_name(crtc->pipe));
+	if (crtc->config.has_audio)
 		intel_dp->DP |= DP_AUDIO_OUTPUT_ENABLE;
-		intel_write_eld(&encoder->base, adjusted_mode);
-	}
 
 	/* Split out the IBX/CPU vs CPT settings */
 
@@ -1110,6 +1409,8 @@ static void wait_panel_status(struct intel_dp *intel_dp,
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	u32 pp_stat_reg, pp_ctrl_reg;
 
+	lockdep_assert_held(&dev_priv->pps_mutex);
+
 	pp_stat_reg = _pp_stat_reg(intel_dp);
 	pp_ctrl_reg = _pp_ctrl_reg(intel_dp);
 
@@ -1173,13 +1474,20 @@ static  u32 ironlake_get_pp_control(struct intel_dp *intel_dp)
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	u32 control;
 
+	lockdep_assert_held(&dev_priv->pps_mutex);
+
 	control = I915_READ(_pp_ctrl_reg(intel_dp));
 	control &= ~PANEL_UNLOCK_MASK;
 	control |= PANEL_UNLOCK_REGS;
 	return control;
 }
 
-static bool _edp_panel_vdd_on(struct intel_dp *intel_dp)
+/*
+ * Must be paired with edp_panel_vdd_off().
+ * Must hold pps_mutex around the whole on/off sequence.
+ * Can be nested with intel_edp_panel_vdd_{on,off}() calls.
+ */
+static bool edp_panel_vdd_on(struct intel_dp *intel_dp)
 {
 	struct drm_device *dev = intel_dp_to_dev(intel_dp);
 	struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp);
@@ -1190,9 +1498,12 @@ static bool _edp_panel_vdd_on(struct intel_dp *intel_dp)
 	u32 pp_stat_reg, pp_ctrl_reg;
 	bool need_to_disable = !intel_dp->want_panel_vdd;
 
+	lockdep_assert_held(&dev_priv->pps_mutex);
+
 	if (!is_edp(intel_dp))
 		return false;
 
+	cancel_delayed_work(&intel_dp->panel_vdd_work);
 	intel_dp->want_panel_vdd = true;
 
 	if (edp_have_panel_vdd(intel_dp))
@@ -1201,7 +1512,8 @@ static bool _edp_panel_vdd_on(struct intel_dp *intel_dp)
 	power_domain = intel_display_port_power_domain(intel_encoder);
 	intel_display_power_get(dev_priv, power_domain);
 
-	DRM_DEBUG_KMS("Turning eDP VDD on\n");
+	DRM_DEBUG_KMS("Turning eDP port %c VDD on\n",
+		      port_name(intel_dig_port->port));
 
 	if (!edp_have_panel_power(intel_dp))
 		wait_panel_power_cycle(intel_dp);
@@ -1220,69 +1532,86 @@ static bool _edp_panel_vdd_on(struct intel_dp *intel_dp)
 	 * If the panel wasn't on, delay before accessing aux channel
 	 */
 	if (!edp_have_panel_power(intel_dp)) {
-		DRM_DEBUG_KMS("eDP was not running\n");
+		DRM_DEBUG_KMS("eDP port %c panel power wasn't enabled\n",
+			      port_name(intel_dig_port->port));
 		msleep(intel_dp->panel_power_up_delay);
 	}
 
 	return need_to_disable;
 }
 
+/*
+ * Must be paired with intel_edp_panel_vdd_off() or
+ * intel_edp_panel_off().
+ * Nested calls to these functions are not allowed since
+ * we drop the lock. Caller must use some higher level
+ * locking to prevent nested calls from other threads.
+ */
 void intel_edp_panel_vdd_on(struct intel_dp *intel_dp)
 {
-	if (is_edp(intel_dp)) {
-		bool vdd = _edp_panel_vdd_on(intel_dp);
+	bool vdd;
 
-		WARN(!vdd, "eDP VDD already requested on\n");
-	}
+	if (!is_edp(intel_dp))
+		return;
+
+	pps_lock(intel_dp);
+	vdd = edp_panel_vdd_on(intel_dp);
+	pps_unlock(intel_dp);
+
+	WARN(!vdd, "eDP port %c VDD already requested on\n",
+	     port_name(dp_to_dig_port(intel_dp)->port));
 }
 
 static void edp_panel_vdd_off_sync(struct intel_dp *intel_dp)
 {
 	struct drm_device *dev = intel_dp_to_dev(intel_dp);
 	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct intel_digital_port *intel_dig_port =
+		dp_to_dig_port(intel_dp);
+	struct intel_encoder *intel_encoder = &intel_dig_port->base;
+	enum intel_display_power_domain power_domain;
 	u32 pp;
 	u32 pp_stat_reg, pp_ctrl_reg;
 
-	WARN_ON(!drm_modeset_is_locked(&dev->mode_config.connection_mutex));
+	lockdep_assert_held(&dev_priv->pps_mutex);
 
-	if (!intel_dp->want_panel_vdd && edp_have_panel_vdd(intel_dp)) {
-		struct intel_digital_port *intel_dig_port =
-						dp_to_dig_port(intel_dp);
-		struct intel_encoder *intel_encoder = &intel_dig_port->base;
-		enum intel_display_power_domain power_domain;
+	WARN_ON(intel_dp->want_panel_vdd);
 
-		DRM_DEBUG_KMS("Turning eDP VDD off\n");
+	if (!edp_have_panel_vdd(intel_dp))
+		return;
 
-		pp = ironlake_get_pp_control(intel_dp);
-		pp &= ~EDP_FORCE_VDD;
+	DRM_DEBUG_KMS("Turning eDP port %c VDD off\n",
+		      port_name(intel_dig_port->port));
 
-		pp_ctrl_reg = _pp_ctrl_reg(intel_dp);
-		pp_stat_reg = _pp_stat_reg(intel_dp);
+	pp = ironlake_get_pp_control(intel_dp);
+	pp &= ~EDP_FORCE_VDD;
 
-		I915_WRITE(pp_ctrl_reg, pp);
-		POSTING_READ(pp_ctrl_reg);
+	pp_ctrl_reg = _pp_ctrl_reg(intel_dp);
+	pp_stat_reg = _pp_stat_reg(intel_dp);
 
-		/* Make sure sequencer is idle before allowing subsequent activity */
-		DRM_DEBUG_KMS("PP_STATUS: 0x%08x PP_CONTROL: 0x%08x\n",
-		I915_READ(pp_stat_reg), I915_READ(pp_ctrl_reg));
+	I915_WRITE(pp_ctrl_reg, pp);
+	POSTING_READ(pp_ctrl_reg);
 
-		if ((pp & POWER_TARGET_ON) == 0)
-			intel_dp->last_power_cycle = jiffies;
+	/* Make sure sequencer is idle before allowing subsequent activity */
+	DRM_DEBUG_KMS("PP_STATUS: 0x%08x PP_CONTROL: 0x%08x\n",
+	I915_READ(pp_stat_reg), I915_READ(pp_ctrl_reg));
 
-		power_domain = intel_display_port_power_domain(intel_encoder);
-		intel_display_power_put(dev_priv, power_domain);
-	}
+	if ((pp & POWER_TARGET_ON) == 0)
+		intel_dp->last_power_cycle = jiffies;
+
+	power_domain = intel_display_port_power_domain(intel_encoder);
+	intel_display_power_put(dev_priv, power_domain);
 }
 
 static void edp_panel_vdd_work(struct work_struct *__work)
 {
 	struct intel_dp *intel_dp = container_of(to_delayed_work(__work),
 						 struct intel_dp, panel_vdd_work);
-	struct drm_device *dev = intel_dp_to_dev(intel_dp);
 
-	drm_modeset_lock(&dev->mode_config.connection_mutex, NULL);
-	edp_panel_vdd_off_sync(intel_dp);
-	drm_modeset_unlock(&dev->mode_config.connection_mutex);
+	pps_lock(intel_dp);
+	if (!intel_dp->want_panel_vdd)
+		edp_panel_vdd_off_sync(intel_dp);
+	pps_unlock(intel_dp);
 }
 
 static void edp_panel_vdd_schedule_off(struct intel_dp *intel_dp)
@@ -1298,12 +1627,23 @@ static void edp_panel_vdd_schedule_off(struct intel_dp *intel_dp)
 	schedule_delayed_work(&intel_dp->panel_vdd_work, delay);
 }
 
+/*
+ * Must be paired with edp_panel_vdd_on().
+ * Must hold pps_mutex around the whole on/off sequence.
+ * Can be nested with intel_edp_panel_vdd_{on,off}() calls.
+ */
 static void edp_panel_vdd_off(struct intel_dp *intel_dp, bool sync)
 {
+	struct drm_i915_private *dev_priv =
+		intel_dp_to_dev(intel_dp)->dev_private;
+
+	lockdep_assert_held(&dev_priv->pps_mutex);
+
 	if (!is_edp(intel_dp))
 		return;
 
-	WARN(!intel_dp->want_panel_vdd, "eDP VDD not forced on");
+	WARN(!intel_dp->want_panel_vdd, "eDP port %c VDD not forced on",
+	     port_name(dp_to_dig_port(intel_dp)->port));
 
 	intel_dp->want_panel_vdd = false;
 
@@ -1313,22 +1653,25 @@ static void edp_panel_vdd_off(struct intel_dp *intel_dp, bool sync)
 		edp_panel_vdd_schedule_off(intel_dp);
 }
 
-void intel_edp_panel_on(struct intel_dp *intel_dp)
+static void edp_panel_on(struct intel_dp *intel_dp)
 {
 	struct drm_device *dev = intel_dp_to_dev(intel_dp);
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	u32 pp;
 	u32 pp_ctrl_reg;
 
+	lockdep_assert_held(&dev_priv->pps_mutex);
+
 	if (!is_edp(intel_dp))
 		return;
 
-	DRM_DEBUG_KMS("Turn eDP power on\n");
+	DRM_DEBUG_KMS("Turn eDP port %c panel power on\n",
+		      port_name(dp_to_dig_port(intel_dp)->port));
 
-	if (edp_have_panel_power(intel_dp)) {
-		DRM_DEBUG_KMS("eDP power already on\n");
+	if (WARN(edp_have_panel_power(intel_dp),
+		 "eDP port %c panel power already on\n",
+		 port_name(dp_to_dig_port(intel_dp)->port)))
 		return;
-	}
 
 	wait_panel_power_cycle(intel_dp);
 
@@ -1358,7 +1701,18 @@ void intel_edp_panel_on(struct intel_dp *intel_dp)
 	}
 }
 
-void intel_edp_panel_off(struct intel_dp *intel_dp)
+void intel_edp_panel_on(struct intel_dp *intel_dp)
+{
+	if (!is_edp(intel_dp))
+		return;
+
+	pps_lock(intel_dp);
+	edp_panel_on(intel_dp);
+	pps_unlock(intel_dp);
+}
+
+
+static void edp_panel_off(struct intel_dp *intel_dp)
 {
 	struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp);
 	struct intel_encoder *intel_encoder = &intel_dig_port->base;
@@ -1368,12 +1722,16 @@ void intel_edp_panel_off(struct intel_dp *intel_dp)
 	u32 pp;
 	u32 pp_ctrl_reg;
 
+	lockdep_assert_held(&dev_priv->pps_mutex);
+
 	if (!is_edp(intel_dp))
 		return;
 
-	DRM_DEBUG_KMS("Turn eDP power off\n");
+	DRM_DEBUG_KMS("Turn eDP port %c panel power off\n",
+		      port_name(dp_to_dig_port(intel_dp)->port));
 
-	WARN(!intel_dp->want_panel_vdd, "Need VDD to turn off panel\n");
+	WARN(!intel_dp->want_panel_vdd, "Need eDP port %c VDD to turn off panel\n",
+	     port_name(dp_to_dig_port(intel_dp)->port));
 
 	pp = ironlake_get_pp_control(intel_dp);
 	/* We need to switch off panel power _and_ force vdd, for otherwise some
@@ -1396,7 +1754,18 @@ void intel_edp_panel_off(struct intel_dp *intel_dp)
 	intel_display_power_put(dev_priv, power_domain);
 }
 
-void intel_edp_backlight_on(struct intel_dp *intel_dp)
+void intel_edp_panel_off(struct intel_dp *intel_dp)
+{
+	if (!is_edp(intel_dp))
+		return;
+
+	pps_lock(intel_dp);
+	edp_panel_off(intel_dp);
+	pps_unlock(intel_dp);
+}
+
+/* Enable backlight in the panel power control. */
+static void _intel_edp_backlight_on(struct intel_dp *intel_dp)
 {
 	struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp);
 	struct drm_device *dev = intel_dig_port->base.base.dev;
@@ -1404,13 +1773,6 @@ void intel_edp_backlight_on(struct intel_dp *intel_dp)
 	u32 pp;
 	u32 pp_ctrl_reg;
 
-	if (!is_edp(intel_dp))
-		return;
-
-	DRM_DEBUG_KMS("\n");
-
-	intel_panel_enable_backlight(intel_dp->attached_connector);
-
 	/*
 	 * If we enable the backlight right away following a panel power
 	 * on, we may see slight flicker as the panel syncs with the eDP
@@ -1418,6 +1780,9 @@ void intel_edp_backlight_on(struct intel_dp *intel_dp)
 	 * allowing it to appear.
 	 */
 	wait_backlight_on(intel_dp);
+
+	pps_lock(intel_dp);
+
 	pp = ironlake_get_pp_control(intel_dp);
 	pp |= EDP_BLC_ENABLE;
 
@@ -1425,9 +1790,24 @@ void intel_edp_backlight_on(struct intel_dp *intel_dp)
 
 	I915_WRITE(pp_ctrl_reg, pp);
 	POSTING_READ(pp_ctrl_reg);
+
+	pps_unlock(intel_dp);
 }
 
-void intel_edp_backlight_off(struct intel_dp *intel_dp)
+/* Enable backlight PWM and backlight PP control. */
+void intel_edp_backlight_on(struct intel_dp *intel_dp)
+{
+	if (!is_edp(intel_dp))
+		return;
+
+	DRM_DEBUG_KMS("\n");
+
+	intel_panel_enable_backlight(intel_dp->attached_connector);
+	_intel_edp_backlight_on(intel_dp);
+}
+
+/* Disable backlight in the panel power control. */
+static void _intel_edp_backlight_off(struct intel_dp *intel_dp)
 {
 	struct drm_device *dev = intel_dp_to_dev(intel_dp);
 	struct drm_i915_private *dev_priv = dev->dev_private;
@@ -1437,7 +1817,8 @@ void intel_edp_backlight_off(struct intel_dp *intel_dp)
 	if (!is_edp(intel_dp))
 		return;
 
-	DRM_DEBUG_KMS("\n");
+	pps_lock(intel_dp);
+
 	pp = ironlake_get_pp_control(intel_dp);
 	pp &= ~EDP_BLC_ENABLE;
 
@@ -1445,13 +1826,51 @@ void intel_edp_backlight_off(struct intel_dp *intel_dp)
 
 	I915_WRITE(pp_ctrl_reg, pp);
 	POSTING_READ(pp_ctrl_reg);
-	intel_dp->last_backlight_off = jiffies;
 
+	pps_unlock(intel_dp);
+
+	intel_dp->last_backlight_off = jiffies;
 	edp_wait_backlight_off(intel_dp);
+}
+
+/* Disable backlight PP control and backlight PWM. */
+void intel_edp_backlight_off(struct intel_dp *intel_dp)
+{
+	if (!is_edp(intel_dp))
+		return;
 
+	DRM_DEBUG_KMS("\n");
+
+	_intel_edp_backlight_off(intel_dp);
 	intel_panel_disable_backlight(intel_dp->attached_connector);
 }
 
+/*
+ * Hook for controlling the panel power control backlight through the bl_power
+ * sysfs attribute. Take care to handle multiple calls.
+ */
+static void intel_edp_backlight_power(struct intel_connector *connector,
+				      bool enable)
+{
+	struct intel_dp *intel_dp = intel_attached_dp(&connector->base);
+	bool is_enabled;
+
+	pps_lock(intel_dp);
+	is_enabled = ironlake_get_pp_control(intel_dp) & EDP_BLC_ENABLE;
+	pps_unlock(intel_dp);
+
+	if (is_enabled == enable)
+		return;
+
+	DRM_DEBUG_KMS("panel power control backlight %s\n",
+		      enable ? "enable" : "disable");
+
+	if (enable)
+		_intel_edp_backlight_on(intel_dp);
+	else
+		_intel_edp_backlight_off(intel_dp);
+}
+
 static void ironlake_edp_pll_on(struct intel_dp *intel_dp)
 {
 	struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp);
@@ -1515,8 +1934,6 @@ void intel_dp_sink_dpms(struct intel_dp *intel_dp, int mode)
 	if (mode != DRM_MODE_DPMS_ON) {
 		ret = drm_dp_dpcd_writeb(&intel_dp->aux, DP_SET_POWER,
 					 DP_SET_POWER_D3);
-		if (ret != 1)
-			DRM_DEBUG_DRIVER("failed to write sink power state\n");
 	} else {
 		/*
 		 * When turning on, we need to retry for 1ms to give the sink
@@ -1530,6 +1947,10 @@ void intel_dp_sink_dpms(struct intel_dp *intel_dp, int mode)
 			msleep(1);
 		}
 	}
+
+	if (ret != 1)
+		DRM_DEBUG_KMS("failed to %s sink power state\n",
+			      mode == DRM_MODE_DPMS_ON ? "enable" : "disable");
 }
 
 static bool intel_dp_get_hw_state(struct intel_encoder *encoder,
@@ -1543,7 +1964,7 @@ static bool intel_dp_get_hw_state(struct intel_encoder *encoder,
 	u32 tmp;
 
 	power_domain = intel_display_port_power_domain(encoder);
-	if (!intel_display_power_enabled(dev_priv, power_domain))
+	if (!intel_display_power_is_enabled(dev_priv, power_domain))
 		return false;
 
 	tmp = I915_READ(intel_dp->output_reg);
@@ -1576,7 +1997,7 @@ static bool intel_dp_get_hw_state(struct intel_encoder *encoder,
 			return true;
 		}
 
-		for_each_pipe(i) {
+		for_each_pipe(dev_priv, i) {
 			trans_dp = I915_READ(TRANS_DP_CTL(i));
 			if ((trans_dp & TRANS_DP_PORT_SEL_MASK) == trans_sel) {
 				*pipe = i;
@@ -1675,369 +2096,14 @@ static void intel_dp_get_config(struct intel_encoder *encoder,
 	}
 }
 
-static bool is_edp_psr(struct intel_dp *intel_dp)
-{
-	return intel_dp->psr_dpcd[0] & DP_PSR_IS_SUPPORTED;
-}
-
-static bool intel_edp_is_psr_enabled(struct drm_device *dev)
-{
-	struct drm_i915_private *dev_priv = dev->dev_private;
-
-	if (!HAS_PSR(dev))
-		return false;
-
-	return I915_READ(EDP_PSR_CTL(dev)) & EDP_PSR_ENABLE;
-}
-
-static void intel_edp_psr_write_vsc(struct intel_dp *intel_dp,
-				    struct edp_vsc_psr *vsc_psr)
-{
-	struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp);
-	struct drm_device *dev = dig_port->base.base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct intel_crtc *crtc = to_intel_crtc(dig_port->base.base.crtc);
-	u32 ctl_reg = HSW_TVIDEO_DIP_CTL(crtc->config.cpu_transcoder);
-	u32 data_reg = HSW_TVIDEO_DIP_VSC_DATA(crtc->config.cpu_transcoder);
-	uint32_t *data = (uint32_t *) vsc_psr;
-	unsigned int i;
-
-	/* As per BSPec (Pipe Video Data Island Packet), we need to disable
-	   the video DIP being updated before program video DIP data buffer
-	   registers for DIP being updated. */
-	I915_WRITE(ctl_reg, 0);
-	POSTING_READ(ctl_reg);
-
-	for (i = 0; i < VIDEO_DIP_VSC_DATA_SIZE; i += 4) {
-		if (i < sizeof(struct edp_vsc_psr))
-			I915_WRITE(data_reg + i, *data++);
-		else
-			I915_WRITE(data_reg + i, 0);
-	}
-
-	I915_WRITE(ctl_reg, VIDEO_DIP_ENABLE_VSC_HSW);
-	POSTING_READ(ctl_reg);
-}
-
-static void intel_edp_psr_setup(struct intel_dp *intel_dp)
-{
-	struct drm_device *dev = intel_dp_to_dev(intel_dp);
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct edp_vsc_psr psr_vsc;
-
-	/* Prepare VSC packet as per EDP 1.3 spec, Table 3.10 */
-	memset(&psr_vsc, 0, sizeof(psr_vsc));
-	psr_vsc.sdp_header.HB0 = 0;
-	psr_vsc.sdp_header.HB1 = 0x7;
-	psr_vsc.sdp_header.HB2 = 0x2;
-	psr_vsc.sdp_header.HB3 = 0x8;
-	intel_edp_psr_write_vsc(intel_dp, &psr_vsc);
-
-	/* Avoid continuous PSR exit by masking memup and hpd */
-	I915_WRITE(EDP_PSR_DEBUG_CTL(dev), EDP_PSR_DEBUG_MASK_MEMUP |
-		   EDP_PSR_DEBUG_MASK_HPD | EDP_PSR_DEBUG_MASK_LPSP);
-}
-
-static void intel_edp_psr_enable_sink(struct intel_dp *intel_dp)
-{
-	struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp);
-	struct drm_device *dev = dig_port->base.base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	uint32_t aux_clock_divider;
-	int precharge = 0x3;
-	int msg_size = 5;       /* Header(4) + Message(1) */
-	bool only_standby = false;
-
-	aux_clock_divider = intel_dp->get_aux_clock_divider(intel_dp, 0);
-
-	if (IS_BROADWELL(dev) && dig_port->port != PORT_A)
-		only_standby = true;
-
-	/* Enable PSR in sink */
-	if (intel_dp->psr_dpcd[1] & DP_PSR_NO_TRAIN_ON_EXIT || only_standby)
-		drm_dp_dpcd_writeb(&intel_dp->aux, DP_PSR_EN_CFG,
-				   DP_PSR_ENABLE & ~DP_PSR_MAIN_LINK_ACTIVE);
-	else
-		drm_dp_dpcd_writeb(&intel_dp->aux, DP_PSR_EN_CFG,
-				   DP_PSR_ENABLE | DP_PSR_MAIN_LINK_ACTIVE);
-
-	/* Setup AUX registers */
-	I915_WRITE(EDP_PSR_AUX_DATA1(dev), EDP_PSR_DPCD_COMMAND);
-	I915_WRITE(EDP_PSR_AUX_DATA2(dev), EDP_PSR_DPCD_NORMAL_OPERATION);
-	I915_WRITE(EDP_PSR_AUX_CTL(dev),
-		   DP_AUX_CH_CTL_TIME_OUT_400us |
-		   (msg_size << DP_AUX_CH_CTL_MESSAGE_SIZE_SHIFT) |
-		   (precharge << DP_AUX_CH_CTL_PRECHARGE_2US_SHIFT) |
-		   (aux_clock_divider << DP_AUX_CH_CTL_BIT_CLOCK_2X_SHIFT));
-}
-
-static void intel_edp_psr_enable_source(struct intel_dp *intel_dp)
-{
-	struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp);
-	struct drm_device *dev = dig_port->base.base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	uint32_t max_sleep_time = 0x1f;
-	uint32_t idle_frames = 1;
-	uint32_t val = 0x0;
-	const uint32_t link_entry_time = EDP_PSR_MIN_LINK_ENTRY_TIME_8_LINES;
-	bool only_standby = false;
-
-	if (IS_BROADWELL(dev) && dig_port->port != PORT_A)
-		only_standby = true;
-
-	if (intel_dp->psr_dpcd[1] & DP_PSR_NO_TRAIN_ON_EXIT || only_standby) {
-		val |= EDP_PSR_LINK_STANDBY;
-		val |= EDP_PSR_TP2_TP3_TIME_0us;
-		val |= EDP_PSR_TP1_TIME_0us;
-		val |= EDP_PSR_SKIP_AUX_EXIT;
-		val |= IS_BROADWELL(dev) ? BDW_PSR_SINGLE_FRAME : 0;
-	} else
-		val |= EDP_PSR_LINK_DISABLE;
-
-	I915_WRITE(EDP_PSR_CTL(dev), val |
-		   (IS_BROADWELL(dev) ? 0 : link_entry_time) |
-		   max_sleep_time << EDP_PSR_MAX_SLEEP_TIME_SHIFT |
-		   idle_frames << EDP_PSR_IDLE_FRAME_SHIFT |
-		   EDP_PSR_ENABLE);
-}
-
-static bool intel_edp_psr_match_conditions(struct intel_dp *intel_dp)
-{
-	struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp);
-	struct drm_device *dev = dig_port->base.base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct drm_crtc *crtc = dig_port->base.base.crtc;
-	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
-
-	lockdep_assert_held(&dev_priv->psr.lock);
-	WARN_ON(!drm_modeset_is_locked(&dev->mode_config.connection_mutex));
-	WARN_ON(!drm_modeset_is_locked(&crtc->mutex));
-
-	dev_priv->psr.source_ok = false;
-
-	if (IS_HASWELL(dev) && dig_port->port != PORT_A) {
-		DRM_DEBUG_KMS("HSW ties PSR to DDI A (eDP)\n");
-		return false;
-	}
-
-	if (!i915.enable_psr) {
-		DRM_DEBUG_KMS("PSR disable by flag\n");
-		return false;
-	}
-
-	/* Below limitations aren't valid for Broadwell */
-	if (IS_BROADWELL(dev))
-		goto out;
-
-	if (I915_READ(HSW_STEREO_3D_CTL(intel_crtc->config.cpu_transcoder)) &
-	    S3D_ENABLE) {
-		DRM_DEBUG_KMS("PSR condition failed: Stereo 3D is Enabled\n");
-		return false;
-	}
-
-	if (intel_crtc->config.adjusted_mode.flags & DRM_MODE_FLAG_INTERLACE) {
-		DRM_DEBUG_KMS("PSR condition failed: Interlaced is Enabled\n");
-		return false;
-	}
-
- out:
-	dev_priv->psr.source_ok = true;
-	return true;
-}
-
-static void intel_edp_psr_do_enable(struct intel_dp *intel_dp)
-{
-	struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp);
-	struct drm_device *dev = intel_dig_port->base.base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
-
-	WARN_ON(I915_READ(EDP_PSR_CTL(dev)) & EDP_PSR_ENABLE);
-	WARN_ON(dev_priv->psr.active);
-	lockdep_assert_held(&dev_priv->psr.lock);
-
-	/* Enable PSR on the panel */
-	intel_edp_psr_enable_sink(intel_dp);
-
-	/* Enable PSR on the host */
-	intel_edp_psr_enable_source(intel_dp);
-
-	dev_priv->psr.active = true;
-}
-
-void intel_edp_psr_enable(struct intel_dp *intel_dp)
-{
-	struct drm_device *dev = intel_dp_to_dev(intel_dp);
-	struct drm_i915_private *dev_priv = dev->dev_private;
-
-	if (!HAS_PSR(dev)) {
-		DRM_DEBUG_KMS("PSR not supported on this platform\n");
-		return;
-	}
-
-	if (!is_edp_psr(intel_dp)) {
-		DRM_DEBUG_KMS("PSR not supported by this panel\n");
-		return;
-	}
-
-	mutex_lock(&dev_priv->psr.lock);
-	if (dev_priv->psr.enabled) {
-		DRM_DEBUG_KMS("PSR already in use\n");
-		mutex_unlock(&dev_priv->psr.lock);
-		return;
-	}
-
-	dev_priv->psr.busy_frontbuffer_bits = 0;
-
-	/* Setup PSR once */
-	intel_edp_psr_setup(intel_dp);
-
-	if (intel_edp_psr_match_conditions(intel_dp))
-		dev_priv->psr.enabled = intel_dp;
-	mutex_unlock(&dev_priv->psr.lock);
-}
-
-void intel_edp_psr_disable(struct intel_dp *intel_dp)
-{
-	struct drm_device *dev = intel_dp_to_dev(intel_dp);
-	struct drm_i915_private *dev_priv = dev->dev_private;
-
-	mutex_lock(&dev_priv->psr.lock);
-	if (!dev_priv->psr.enabled) {
-		mutex_unlock(&dev_priv->psr.lock);
-		return;
-	}
-
-	if (dev_priv->psr.active) {
-		I915_WRITE(EDP_PSR_CTL(dev),
-			   I915_READ(EDP_PSR_CTL(dev)) & ~EDP_PSR_ENABLE);
-
-		/* Wait till PSR is idle */
-		if (_wait_for((I915_READ(EDP_PSR_STATUS_CTL(dev)) &
-			       EDP_PSR_STATUS_STATE_MASK) == 0, 2000, 10))
-			DRM_ERROR("Timed out waiting for PSR Idle State\n");
-
-		dev_priv->psr.active = false;
-	} else {
-		WARN_ON(I915_READ(EDP_PSR_CTL(dev)) & EDP_PSR_ENABLE);
-	}
-
-	dev_priv->psr.enabled = NULL;
-	mutex_unlock(&dev_priv->psr.lock);
-
-	cancel_delayed_work_sync(&dev_priv->psr.work);
-}
-
-static void intel_edp_psr_work(struct work_struct *work)
-{
-	struct drm_i915_private *dev_priv =
-		container_of(work, typeof(*dev_priv), psr.work.work);
-	struct intel_dp *intel_dp = dev_priv->psr.enabled;
-
-	mutex_lock(&dev_priv->psr.lock);
-	intel_dp = dev_priv->psr.enabled;
-
-	if (!intel_dp)
-		goto unlock;
-
-	/*
-	 * The delayed work can race with an invalidate hence we need to
-	 * recheck. Since psr_flush first clears this and then reschedules we
-	 * won't ever miss a flush when bailing out here.
-	 */
-	if (dev_priv->psr.busy_frontbuffer_bits)
-		goto unlock;
-
-	intel_edp_psr_do_enable(intel_dp);
-unlock:
-	mutex_unlock(&dev_priv->psr.lock);
-}
-
-static void intel_edp_psr_do_exit(struct drm_device *dev)
-{
-	struct drm_i915_private *dev_priv = dev->dev_private;
-
-	if (dev_priv->psr.active) {
-		u32 val = I915_READ(EDP_PSR_CTL(dev));
-
-		WARN_ON(!(val & EDP_PSR_ENABLE));
-
-		I915_WRITE(EDP_PSR_CTL(dev), val & ~EDP_PSR_ENABLE);
-
-		dev_priv->psr.active = false;
-	}
-
-}
-
-void intel_edp_psr_invalidate(struct drm_device *dev,
-			      unsigned frontbuffer_bits)
-{
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct drm_crtc *crtc;
-	enum pipe pipe;
-
-	mutex_lock(&dev_priv->psr.lock);
-	if (!dev_priv->psr.enabled) {
-		mutex_unlock(&dev_priv->psr.lock);
-		return;
-	}
-
-	crtc = dp_to_dig_port(dev_priv->psr.enabled)->base.base.crtc;
-	pipe = to_intel_crtc(crtc)->pipe;
-
-	intel_edp_psr_do_exit(dev);
-
-	frontbuffer_bits &= INTEL_FRONTBUFFER_ALL_MASK(pipe);
-
-	dev_priv->psr.busy_frontbuffer_bits |= frontbuffer_bits;
-	mutex_unlock(&dev_priv->psr.lock);
-}
-
-void intel_edp_psr_flush(struct drm_device *dev,
-			 unsigned frontbuffer_bits)
-{
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct drm_crtc *crtc;
-	enum pipe pipe;
-
-	mutex_lock(&dev_priv->psr.lock);
-	if (!dev_priv->psr.enabled) {
-		mutex_unlock(&dev_priv->psr.lock);
-		return;
-	}
-
-	crtc = dp_to_dig_port(dev_priv->psr.enabled)->base.base.crtc;
-	pipe = to_intel_crtc(crtc)->pipe;
-	dev_priv->psr.busy_frontbuffer_bits &= ~frontbuffer_bits;
-
-	/*
-	 * On Haswell sprite plane updates don't result in a psr invalidating
-	 * signal in the hardware. Which means we need to manually fake this in
-	 * software for all flushes, not just when we've seen a preceding
-	 * invalidation through frontbuffer rendering.
-	 */
-	if (IS_HASWELL(dev) &&
-	    (frontbuffer_bits & INTEL_FRONTBUFFER_SPRITE(pipe)))
-		intel_edp_psr_do_exit(dev);
-
-	if (!dev_priv->psr.active && !dev_priv->psr.busy_frontbuffer_bits)
-		schedule_delayed_work(&dev_priv->psr.work,
-				      msecs_to_jiffies(100));
-	mutex_unlock(&dev_priv->psr.lock);
-}
-
-void intel_edp_psr_init(struct drm_device *dev)
-{
-	struct drm_i915_private *dev_priv = dev->dev_private;
-
-	INIT_DELAYED_WORK(&dev_priv->psr.work, intel_edp_psr_work);
-	mutex_init(&dev_priv->psr.lock);
-}
-
 static void intel_disable_dp(struct intel_encoder *encoder)
 {
 	struct intel_dp *intel_dp = enc_to_intel_dp(&encoder->base);
-	enum port port = dp_to_dig_port(intel_dp)->port;
 	struct drm_device *dev = encoder->base.dev;
+	struct intel_crtc *crtc = to_intel_crtc(encoder->base.crtc);
+
+	if (crtc->config.has_audio)
+		intel_audio_codec_disable(encoder);
 
 	/* Make sure the panel is off before trying to change the mode. But also
 	 * ensure that we have vdd while we switch off the panel. */
@@ -2046,21 +2112,19 @@ static void intel_disable_dp(struct intel_encoder *encoder)
 	intel_dp_sink_dpms(intel_dp, DRM_MODE_DPMS_OFF);
 	intel_edp_panel_off(intel_dp);
 
-	/* cpu edp my only be disable _after_ the cpu pipe/plane is disabled. */
-	if (!(port == PORT_A || IS_VALLEYVIEW(dev)))
+	/* disable the port before the pipe on g4x */
+	if (INTEL_INFO(dev)->gen < 5)
 		intel_dp_link_down(intel_dp);
 }
 
-static void g4x_post_disable_dp(struct intel_encoder *encoder)
+static void ilk_post_disable_dp(struct intel_encoder *encoder)
 {
 	struct intel_dp *intel_dp = enc_to_intel_dp(&encoder->base);
 	enum port port = dp_to_dig_port(intel_dp)->port;
 
-	if (port != PORT_A)
-		return;
-
 	intel_dp_link_down(intel_dp);
-	ironlake_edp_pll_off(intel_dp);
+	if (port == PORT_A)
+		ironlake_edp_pll_off(intel_dp);
 }
 
 static void vlv_post_disable_dp(struct intel_encoder *encoder)
@@ -2106,23 +2170,150 @@ static void chv_post_disable_dp(struct intel_encoder *encoder)
 	mutex_unlock(&dev_priv->dpio_lock);
 }
 
+static void
+_intel_dp_set_link_train(struct intel_dp *intel_dp,
+			 uint32_t *DP,
+			 uint8_t dp_train_pat)
+{
+	struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp);
+	struct drm_device *dev = intel_dig_port->base.base.dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	enum port port = intel_dig_port->port;
+
+	if (HAS_DDI(dev)) {
+		uint32_t temp = I915_READ(DP_TP_CTL(port));
+
+		if (dp_train_pat & DP_LINK_SCRAMBLING_DISABLE)
+			temp |= DP_TP_CTL_SCRAMBLE_DISABLE;
+		else
+			temp &= ~DP_TP_CTL_SCRAMBLE_DISABLE;
+
+		temp &= ~DP_TP_CTL_LINK_TRAIN_MASK;
+		switch (dp_train_pat & DP_TRAINING_PATTERN_MASK) {
+		case DP_TRAINING_PATTERN_DISABLE:
+			temp |= DP_TP_CTL_LINK_TRAIN_NORMAL;
+
+			break;
+		case DP_TRAINING_PATTERN_1:
+			temp |= DP_TP_CTL_LINK_TRAIN_PAT1;
+			break;
+		case DP_TRAINING_PATTERN_2:
+			temp |= DP_TP_CTL_LINK_TRAIN_PAT2;
+			break;
+		case DP_TRAINING_PATTERN_3:
+			temp |= DP_TP_CTL_LINK_TRAIN_PAT3;
+			break;
+		}
+		I915_WRITE(DP_TP_CTL(port), temp);
+
+	} else if (HAS_PCH_CPT(dev) && (IS_GEN7(dev) || port != PORT_A)) {
+		*DP &= ~DP_LINK_TRAIN_MASK_CPT;
+
+		switch (dp_train_pat & DP_TRAINING_PATTERN_MASK) {
+		case DP_TRAINING_PATTERN_DISABLE:
+			*DP |= DP_LINK_TRAIN_OFF_CPT;
+			break;
+		case DP_TRAINING_PATTERN_1:
+			*DP |= DP_LINK_TRAIN_PAT_1_CPT;
+			break;
+		case DP_TRAINING_PATTERN_2:
+			*DP |= DP_LINK_TRAIN_PAT_2_CPT;
+			break;
+		case DP_TRAINING_PATTERN_3:
+			DRM_ERROR("DP training pattern 3 not supported\n");
+			*DP |= DP_LINK_TRAIN_PAT_2_CPT;
+			break;
+		}
+
+	} else {
+		if (IS_CHERRYVIEW(dev))
+			*DP &= ~DP_LINK_TRAIN_MASK_CHV;
+		else
+			*DP &= ~DP_LINK_TRAIN_MASK;
+
+		switch (dp_train_pat & DP_TRAINING_PATTERN_MASK) {
+		case DP_TRAINING_PATTERN_DISABLE:
+			*DP |= DP_LINK_TRAIN_OFF;
+			break;
+		case DP_TRAINING_PATTERN_1:
+			*DP |= DP_LINK_TRAIN_PAT_1;
+			break;
+		case DP_TRAINING_PATTERN_2:
+			*DP |= DP_LINK_TRAIN_PAT_2;
+			break;
+		case DP_TRAINING_PATTERN_3:
+			if (IS_CHERRYVIEW(dev)) {
+				*DP |= DP_LINK_TRAIN_PAT_3_CHV;
+			} else {
+				DRM_ERROR("DP training pattern 3 not supported\n");
+				*DP |= DP_LINK_TRAIN_PAT_2;
+			}
+			break;
+		}
+	}
+}
+
+static void intel_dp_enable_port(struct intel_dp *intel_dp)
+{
+	struct drm_device *dev = intel_dp_to_dev(intel_dp);
+	struct drm_i915_private *dev_priv = dev->dev_private;
+
+	/* enable with pattern 1 (as per spec) */
+	_intel_dp_set_link_train(intel_dp, &intel_dp->DP,
+				 DP_TRAINING_PATTERN_1);
+
+	I915_WRITE(intel_dp->output_reg, intel_dp->DP);
+	POSTING_READ(intel_dp->output_reg);
+
+	/*
+	 * Magic for VLV/CHV. We _must_ first set up the register
+	 * without actually enabling the port, and then do another
+	 * write to enable the port. Otherwise link training will
+	 * fail when the power sequencer is freshly used for this port.
+	 */
+	intel_dp->DP |= DP_PORT_EN;
+
+	I915_WRITE(intel_dp->output_reg, intel_dp->DP);
+	POSTING_READ(intel_dp->output_reg);
+}
+
 static void intel_enable_dp(struct intel_encoder *encoder)
 {
 	struct intel_dp *intel_dp = enc_to_intel_dp(&encoder->base);
 	struct drm_device *dev = encoder->base.dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct intel_crtc *crtc = to_intel_crtc(encoder->base.crtc);
 	uint32_t dp_reg = I915_READ(intel_dp->output_reg);
 
 	if (WARN_ON(dp_reg & DP_PORT_EN))
 		return;
 
-	intel_edp_panel_vdd_on(intel_dp);
+	pps_lock(intel_dp);
+
+	if (IS_VALLEYVIEW(dev))
+		vlv_init_panel_power_sequencer(intel_dp);
+
+	intel_dp_enable_port(intel_dp);
+
+	edp_panel_vdd_on(intel_dp);
+	edp_panel_on(intel_dp);
+	edp_panel_vdd_off(intel_dp, true);
+
+	pps_unlock(intel_dp);
+
+	if (IS_VALLEYVIEW(dev))
+		vlv_wait_port_ready(dev_priv, dp_to_dig_port(intel_dp));
+
 	intel_dp_sink_dpms(intel_dp, DRM_MODE_DPMS_ON);
 	intel_dp_start_link_train(intel_dp);
-	intel_edp_panel_on(intel_dp);
-	edp_panel_vdd_off(intel_dp, true);
 	intel_dp_complete_link_train(intel_dp);
 	intel_dp_stop_link_train(intel_dp);
+
+	if (crtc->config.has_audio) {
+		DRM_DEBUG_DRIVER("Enabling DP audio on pipe %c\n",
+				 pipe_name(crtc->pipe));
+		intel_audio_codec_enable(encoder);
+	}
 }
 
 static void g4x_enable_dp(struct intel_encoder *encoder)
@@ -2154,6 +2345,110 @@ static void g4x_pre_enable_dp(struct intel_encoder *encoder)
 	}
 }
 
+static void vlv_detach_power_sequencer(struct intel_dp *intel_dp)
+{
+	struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp);
+	struct drm_i915_private *dev_priv = intel_dig_port->base.base.dev->dev_private;
+	enum pipe pipe = intel_dp->pps_pipe;
+	int pp_on_reg = VLV_PIPE_PP_ON_DELAYS(pipe);
+
+	edp_panel_vdd_off_sync(intel_dp);
+
+	/*
+	 * VLV seems to get confused when multiple power seqeuencers
+	 * have the same port selected (even if only one has power/vdd
+	 * enabled). The failure manifests as vlv_wait_port_ready() failing
+	 * CHV on the other hand doesn't seem to mind having the same port
+	 * selected in multiple power seqeuencers, but let's clear the
+	 * port select always when logically disconnecting a power sequencer
+	 * from a port.
+	 */
+	DRM_DEBUG_KMS("detaching pipe %c power sequencer from port %c\n",
+		      pipe_name(pipe), port_name(intel_dig_port->port));
+	I915_WRITE(pp_on_reg, 0);
+	POSTING_READ(pp_on_reg);
+
+	intel_dp->pps_pipe = INVALID_PIPE;
+}
+
+static void vlv_steal_power_sequencer(struct drm_device *dev,
+				      enum pipe pipe)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct intel_encoder *encoder;
+
+	lockdep_assert_held(&dev_priv->pps_mutex);
+
+	if (WARN_ON(pipe != PIPE_A && pipe != PIPE_B))
+		return;
+
+	list_for_each_entry(encoder, &dev->mode_config.encoder_list,
+			    base.head) {
+		struct intel_dp *intel_dp;
+		enum port port;
+
+		if (encoder->type != INTEL_OUTPUT_EDP)
+			continue;
+
+		intel_dp = enc_to_intel_dp(&encoder->base);
+		port = dp_to_dig_port(intel_dp)->port;
+
+		if (intel_dp->pps_pipe != pipe)
+			continue;
+
+		DRM_DEBUG_KMS("stealing pipe %c power sequencer from port %c\n",
+			      pipe_name(pipe), port_name(port));
+
+		WARN(encoder->connectors_active,
+		     "stealing pipe %c power sequencer from active eDP port %c\n",
+		     pipe_name(pipe), port_name(port));
+
+		/* make sure vdd is off before we steal it */
+		vlv_detach_power_sequencer(intel_dp);
+	}
+}
+
+static void vlv_init_panel_power_sequencer(struct intel_dp *intel_dp)
+{
+	struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp);
+	struct intel_encoder *encoder = &intel_dig_port->base;
+	struct drm_device *dev = encoder->base.dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct intel_crtc *crtc = to_intel_crtc(encoder->base.crtc);
+
+	lockdep_assert_held(&dev_priv->pps_mutex);
+
+	if (!is_edp(intel_dp))
+		return;
+
+	if (intel_dp->pps_pipe == crtc->pipe)
+		return;
+
+	/*
+	 * If another power sequencer was being used on this
+	 * port previously make sure to turn off vdd there while
+	 * we still have control of it.
+	 */
+	if (intel_dp->pps_pipe != INVALID_PIPE)
+		vlv_detach_power_sequencer(intel_dp);
+
+	/*
+	 * We may be stealing the power
+	 * sequencer from another port.
+	 */
+	vlv_steal_power_sequencer(dev, crtc->pipe);
+
+	/* now it's all ours */
+	intel_dp->pps_pipe = crtc->pipe;
+
+	DRM_DEBUG_KMS("initializing pipe %c power sequencer for port %c\n",
+		      pipe_name(intel_dp->pps_pipe), port_name(intel_dig_port->port));
+
+	/* init power sequencer on this pipe and port */
+	intel_dp_init_panel_power_sequencer(dev, intel_dp);
+	intel_dp_init_panel_power_sequencer_registers(dev, intel_dp);
+}
+
 static void vlv_pre_enable_dp(struct intel_encoder *encoder)
 {
 	struct intel_dp *intel_dp = enc_to_intel_dp(&encoder->base);
@@ -2163,7 +2458,6 @@ static void vlv_pre_enable_dp(struct intel_encoder *encoder)
 	struct intel_crtc *intel_crtc = to_intel_crtc(encoder->base.crtc);
 	enum dpio_channel port = vlv_dport_to_channel(dport);
 	int pipe = intel_crtc->pipe;
-	struct edp_power_seq power_seq;
 	u32 val;
 
 	mutex_lock(&dev_priv->dpio_lock);
@@ -2181,16 +2475,7 @@ static void vlv_pre_enable_dp(struct intel_encoder *encoder)
 
 	mutex_unlock(&dev_priv->dpio_lock);
 
-	if (is_edp(intel_dp)) {
-		/* init power sequencer on this pipe and port */
-		intel_dp_init_panel_power_sequencer(dev, intel_dp, &power_seq);
-		intel_dp_init_panel_power_sequencer_registers(dev, intel_dp,
-							      &power_seq);
-	}
-
 	intel_enable_dp(encoder);
-
-	vlv_wait_port_ready(dev_priv, dport);
 }
 
 static void vlv_dp_pre_pll_enable(struct intel_encoder *encoder)
@@ -2229,7 +2514,6 @@ static void chv_pre_enable_dp(struct intel_encoder *encoder)
 	struct intel_digital_port *dport = dp_to_dig_port(intel_dp);
 	struct drm_device *dev = encoder->base.dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct edp_power_seq power_seq;
 	struct intel_crtc *intel_crtc =
 		to_intel_crtc(encoder->base.crtc);
 	enum dpio_channel ch = vlv_dport_to_channel(dport);
@@ -2239,6 +2523,15 @@ static void chv_pre_enable_dp(struct intel_encoder *encoder)
 
 	mutex_lock(&dev_priv->dpio_lock);
 
+	/* allow hardware to manage TX FIFO reset source */
+	val = vlv_dpio_read(dev_priv, pipe, VLV_PCS01_DW11(ch));
+	val &= ~DPIO_LANEDESKEW_STRAP_OVRD;
+	vlv_dpio_write(dev_priv, pipe, VLV_PCS01_DW11(ch), val);
+
+	val = vlv_dpio_read(dev_priv, pipe, VLV_PCS23_DW11(ch));
+	val &= ~DPIO_LANEDESKEW_STRAP_OVRD;
+	vlv_dpio_write(dev_priv, pipe, VLV_PCS23_DW11(ch), val);
+
 	/* Deassert soft data lane reset*/
 	val = vlv_dpio_read(dev_priv, pipe, VLV_PCS01_DW1(ch));
 	val |= CHV_PCS_REQ_SOFTRESET_EN;
@@ -2274,16 +2567,7 @@ static void chv_pre_enable_dp(struct intel_encoder *encoder)
 
 	mutex_unlock(&dev_priv->dpio_lock);
 
-	if (is_edp(intel_dp)) {
-		/* init power sequencer on this pipe and port */
-		intel_dp_init_panel_power_sequencer(dev, intel_dp, &power_seq);
-		intel_dp_init_panel_power_sequencer_registers(dev, intel_dp,
-							      &power_seq);
-	}
-
 	intel_enable_dp(encoder);
-
-	vlv_wait_port_ready(dev_priv, dport);
 }
 
 static void chv_dp_pre_pll_enable(struct intel_encoder *encoder)
@@ -2297,6 +2581,8 @@ static void chv_dp_pre_pll_enable(struct intel_encoder *encoder)
 	enum pipe pipe = intel_crtc->pipe;
 	u32 val;
 
+	intel_dp_prepare(encoder);
+
 	mutex_lock(&dev_priv->dpio_lock);
 
 	/* program left/right clock distribution */
@@ -2364,6 +2650,13 @@ intel_dp_dpcd_read_wake(struct drm_dp_aux *aux, unsigned int offset,
 	ssize_t ret;
 	int i;
 
+	/*
+	 * Sometime we just get the same incorrect byte repeated
+	 * over the entire buffer. Doing just one throw away read
+	 * initially seems to "solve" it.
+	 */
+	drm_dp_dpcd_read(aux, DP_DPCD_REV, buffer, 1);
+
 	for (i = 0; i < 3; i++) {
 		ret = drm_dp_dpcd_read(aux, offset, buffer, size);
 		if (ret == size)
@@ -2394,14 +2687,16 @@ intel_dp_voltage_max(struct intel_dp *intel_dp)
 	struct drm_device *dev = intel_dp_to_dev(intel_dp);
 	enum port port = dp_to_dig_port(intel_dp)->port;
 
-	if (IS_VALLEYVIEW(dev))
-		return DP_TRAIN_VOLTAGE_SWING_1200;
+	if (INTEL_INFO(dev)->gen >= 9)
+		return DP_TRAIN_VOLTAGE_SWING_LEVEL_2;
+	else if (IS_VALLEYVIEW(dev))
+		return DP_TRAIN_VOLTAGE_SWING_LEVEL_3;
 	else if (IS_GEN7(dev) && port == PORT_A)
-		return DP_TRAIN_VOLTAGE_SWING_800;
+		return DP_TRAIN_VOLTAGE_SWING_LEVEL_2;
 	else if (HAS_PCH_CPT(dev) && port != PORT_A)
-		return DP_TRAIN_VOLTAGE_SWING_1200;
+		return DP_TRAIN_VOLTAGE_SWING_LEVEL_3;
 	else
-		return DP_TRAIN_VOLTAGE_SWING_800;
+		return DP_TRAIN_VOLTAGE_SWING_LEVEL_2;
 }
 
 static uint8_t
@@ -2410,51 +2705,62 @@ intel_dp_pre_emphasis_max(struct intel_dp *intel_dp, uint8_t voltage_swing)
 	struct drm_device *dev = intel_dp_to_dev(intel_dp);
 	enum port port = dp_to_dig_port(intel_dp)->port;
 
-	if (IS_HASWELL(dev) || IS_BROADWELL(dev)) {
+	if (INTEL_INFO(dev)->gen >= 9) {
 		switch (voltage_swing & DP_TRAIN_VOLTAGE_SWING_MASK) {
-		case DP_TRAIN_VOLTAGE_SWING_400:
-			return DP_TRAIN_PRE_EMPHASIS_9_5;
-		case DP_TRAIN_VOLTAGE_SWING_600:
-			return DP_TRAIN_PRE_EMPHASIS_6;
-		case DP_TRAIN_VOLTAGE_SWING_800:
-			return DP_TRAIN_PRE_EMPHASIS_3_5;
-		case DP_TRAIN_VOLTAGE_SWING_1200:
+		case DP_TRAIN_VOLTAGE_SWING_LEVEL_0:
+			return DP_TRAIN_PRE_EMPH_LEVEL_3;
+		case DP_TRAIN_VOLTAGE_SWING_LEVEL_1:
+			return DP_TRAIN_PRE_EMPH_LEVEL_2;
+		case DP_TRAIN_VOLTAGE_SWING_LEVEL_2:
+			return DP_TRAIN_PRE_EMPH_LEVEL_1;
 		default:
-			return DP_TRAIN_PRE_EMPHASIS_0;
+			return DP_TRAIN_PRE_EMPH_LEVEL_0;
+		}
+	} else if (IS_HASWELL(dev) || IS_BROADWELL(dev)) {
+		switch (voltage_swing & DP_TRAIN_VOLTAGE_SWING_MASK) {
+		case DP_TRAIN_VOLTAGE_SWING_LEVEL_0:
+			return DP_TRAIN_PRE_EMPH_LEVEL_3;
+		case DP_TRAIN_VOLTAGE_SWING_LEVEL_1:
+			return DP_TRAIN_PRE_EMPH_LEVEL_2;
+		case DP_TRAIN_VOLTAGE_SWING_LEVEL_2:
+			return DP_TRAIN_PRE_EMPH_LEVEL_1;
+		case DP_TRAIN_VOLTAGE_SWING_LEVEL_3:
+		default:
+			return DP_TRAIN_PRE_EMPH_LEVEL_0;
 		}
 	} else if (IS_VALLEYVIEW(dev)) {
 		switch (voltage_swing & DP_TRAIN_VOLTAGE_SWING_MASK) {
-		case DP_TRAIN_VOLTAGE_SWING_400:
-			return DP_TRAIN_PRE_EMPHASIS_9_5;
-		case DP_TRAIN_VOLTAGE_SWING_600:
-			return DP_TRAIN_PRE_EMPHASIS_6;
-		case DP_TRAIN_VOLTAGE_SWING_800:
-			return DP_TRAIN_PRE_EMPHASIS_3_5;
-		case DP_TRAIN_VOLTAGE_SWING_1200:
+		case DP_TRAIN_VOLTAGE_SWING_LEVEL_0:
+			return DP_TRAIN_PRE_EMPH_LEVEL_3;
+		case DP_TRAIN_VOLTAGE_SWING_LEVEL_1:
+			return DP_TRAIN_PRE_EMPH_LEVEL_2;
+		case DP_TRAIN_VOLTAGE_SWING_LEVEL_2:
+			return DP_TRAIN_PRE_EMPH_LEVEL_1;
+		case DP_TRAIN_VOLTAGE_SWING_LEVEL_3:
 		default:
-			return DP_TRAIN_PRE_EMPHASIS_0;
+			return DP_TRAIN_PRE_EMPH_LEVEL_0;
 		}
 	} else if (IS_GEN7(dev) && port == PORT_A) {
 		switch (voltage_swing & DP_TRAIN_VOLTAGE_SWING_MASK) {
-		case DP_TRAIN_VOLTAGE_SWING_400:
-			return DP_TRAIN_PRE_EMPHASIS_6;
-		case DP_TRAIN_VOLTAGE_SWING_600:
-		case DP_TRAIN_VOLTAGE_SWING_800:
-			return DP_TRAIN_PRE_EMPHASIS_3_5;
+		case DP_TRAIN_VOLTAGE_SWING_LEVEL_0:
+			return DP_TRAIN_PRE_EMPH_LEVEL_2;
+		case DP_TRAIN_VOLTAGE_SWING_LEVEL_1:
+		case DP_TRAIN_VOLTAGE_SWING_LEVEL_2:
+			return DP_TRAIN_PRE_EMPH_LEVEL_1;
 		default:
-			return DP_TRAIN_PRE_EMPHASIS_0;
+			return DP_TRAIN_PRE_EMPH_LEVEL_0;
 		}
 	} else {
 		switch (voltage_swing & DP_TRAIN_VOLTAGE_SWING_MASK) {
-		case DP_TRAIN_VOLTAGE_SWING_400:
-			return DP_TRAIN_PRE_EMPHASIS_6;
-		case DP_TRAIN_VOLTAGE_SWING_600:
-			return DP_TRAIN_PRE_EMPHASIS_6;
-		case DP_TRAIN_VOLTAGE_SWING_800:
-			return DP_TRAIN_PRE_EMPHASIS_3_5;
-		case DP_TRAIN_VOLTAGE_SWING_1200:
+		case DP_TRAIN_VOLTAGE_SWING_LEVEL_0:
+			return DP_TRAIN_PRE_EMPH_LEVEL_2;
+		case DP_TRAIN_VOLTAGE_SWING_LEVEL_1:
+			return DP_TRAIN_PRE_EMPH_LEVEL_2;
+		case DP_TRAIN_VOLTAGE_SWING_LEVEL_2:
+			return DP_TRAIN_PRE_EMPH_LEVEL_1;
+		case DP_TRAIN_VOLTAGE_SWING_LEVEL_3:
 		default:
-			return DP_TRAIN_PRE_EMPHASIS_0;
+			return DP_TRAIN_PRE_EMPH_LEVEL_0;
 		}
 	}
 }
@@ -2473,22 +2779,22 @@ static uint32_t intel_vlv_signal_levels(struct intel_dp *intel_dp)
 	int pipe = intel_crtc->pipe;
 
 	switch (train_set & DP_TRAIN_PRE_EMPHASIS_MASK) {
-	case DP_TRAIN_PRE_EMPHASIS_0:
+	case DP_TRAIN_PRE_EMPH_LEVEL_0:
 		preemph_reg_value = 0x0004000;
 		switch (train_set & DP_TRAIN_VOLTAGE_SWING_MASK) {
-		case DP_TRAIN_VOLTAGE_SWING_400:
+		case DP_TRAIN_VOLTAGE_SWING_LEVEL_0:
 			demph_reg_value = 0x2B405555;
 			uniqtranscale_reg_value = 0x552AB83A;
 			break;
-		case DP_TRAIN_VOLTAGE_SWING_600:
+		case DP_TRAIN_VOLTAGE_SWING_LEVEL_1:
 			demph_reg_value = 0x2B404040;
 			uniqtranscale_reg_value = 0x5548B83A;
 			break;
-		case DP_TRAIN_VOLTAGE_SWING_800:
+		case DP_TRAIN_VOLTAGE_SWING_LEVEL_2:
 			demph_reg_value = 0x2B245555;
 			uniqtranscale_reg_value = 0x5560B83A;
 			break;
-		case DP_TRAIN_VOLTAGE_SWING_1200:
+		case DP_TRAIN_VOLTAGE_SWING_LEVEL_3:
 			demph_reg_value = 0x2B405555;
 			uniqtranscale_reg_value = 0x5598DA3A;
 			break;
@@ -2496,18 +2802,18 @@ static uint32_t intel_vlv_signal_levels(struct intel_dp *intel_dp)
 			return 0;
 		}
 		break;
-	case DP_TRAIN_PRE_EMPHASIS_3_5:
+	case DP_TRAIN_PRE_EMPH_LEVEL_1:
 		preemph_reg_value = 0x0002000;
 		switch (train_set & DP_TRAIN_VOLTAGE_SWING_MASK) {
-		case DP_TRAIN_VOLTAGE_SWING_400:
+		case DP_TRAIN_VOLTAGE_SWING_LEVEL_0:
 			demph_reg_value = 0x2B404040;
 			uniqtranscale_reg_value = 0x5552B83A;
 			break;
-		case DP_TRAIN_VOLTAGE_SWING_600:
+		case DP_TRAIN_VOLTAGE_SWING_LEVEL_1:
 			demph_reg_value = 0x2B404848;
 			uniqtranscale_reg_value = 0x5580B83A;
 			break;
-		case DP_TRAIN_VOLTAGE_SWING_800:
+		case DP_TRAIN_VOLTAGE_SWING_LEVEL_2:
 			demph_reg_value = 0x2B404040;
 			uniqtranscale_reg_value = 0x55ADDA3A;
 			break;
@@ -2515,14 +2821,14 @@ static uint32_t intel_vlv_signal_levels(struct intel_dp *intel_dp)
 			return 0;
 		}
 		break;
-	case DP_TRAIN_PRE_EMPHASIS_6:
+	case DP_TRAIN_PRE_EMPH_LEVEL_2:
 		preemph_reg_value = 0x0000000;
 		switch (train_set & DP_TRAIN_VOLTAGE_SWING_MASK) {
-		case DP_TRAIN_VOLTAGE_SWING_400:
+		case DP_TRAIN_VOLTAGE_SWING_LEVEL_0:
 			demph_reg_value = 0x2B305555;
 			uniqtranscale_reg_value = 0x5570B83A;
 			break;
-		case DP_TRAIN_VOLTAGE_SWING_600:
+		case DP_TRAIN_VOLTAGE_SWING_LEVEL_1:
 			demph_reg_value = 0x2B2B4040;
 			uniqtranscale_reg_value = 0x55ADDA3A;
 			break;
@@ -2530,10 +2836,10 @@ static uint32_t intel_vlv_signal_levels(struct intel_dp *intel_dp)
 			return 0;
 		}
 		break;
-	case DP_TRAIN_PRE_EMPHASIS_9_5:
+	case DP_TRAIN_PRE_EMPH_LEVEL_3:
 		preemph_reg_value = 0x0006000;
 		switch (train_set & DP_TRAIN_VOLTAGE_SWING_MASK) {
-		case DP_TRAIN_VOLTAGE_SWING_400:
+		case DP_TRAIN_VOLTAGE_SWING_LEVEL_0:
 			demph_reg_value = 0x1B405555;
 			uniqtranscale_reg_value = 0x55ADDA3A;
 			break;
@@ -2572,21 +2878,21 @@ static uint32_t intel_chv_signal_levels(struct intel_dp *intel_dp)
 	int i;
 
 	switch (train_set & DP_TRAIN_PRE_EMPHASIS_MASK) {
-	case DP_TRAIN_PRE_EMPHASIS_0:
+	case DP_TRAIN_PRE_EMPH_LEVEL_0:
 		switch (train_set & DP_TRAIN_VOLTAGE_SWING_MASK) {
-		case DP_TRAIN_VOLTAGE_SWING_400:
+		case DP_TRAIN_VOLTAGE_SWING_LEVEL_0:
 			deemph_reg_value = 128;
 			margin_reg_value = 52;
 			break;
-		case DP_TRAIN_VOLTAGE_SWING_600:
+		case DP_TRAIN_VOLTAGE_SWING_LEVEL_1:
 			deemph_reg_value = 128;
 			margin_reg_value = 77;
 			break;
-		case DP_TRAIN_VOLTAGE_SWING_800:
+		case DP_TRAIN_VOLTAGE_SWING_LEVEL_2:
 			deemph_reg_value = 128;
 			margin_reg_value = 102;
 			break;
-		case DP_TRAIN_VOLTAGE_SWING_1200:
+		case DP_TRAIN_VOLTAGE_SWING_LEVEL_3:
 			deemph_reg_value = 128;
 			margin_reg_value = 154;
 			/* FIXME extra to set for 1200 */
@@ -2595,17 +2901,17 @@ static uint32_t intel_chv_signal_levels(struct intel_dp *intel_dp)
 			return 0;
 		}
 		break;
-	case DP_TRAIN_PRE_EMPHASIS_3_5:
+	case DP_TRAIN_PRE_EMPH_LEVEL_1:
 		switch (train_set & DP_TRAIN_VOLTAGE_SWING_MASK) {
-		case DP_TRAIN_VOLTAGE_SWING_400:
+		case DP_TRAIN_VOLTAGE_SWING_LEVEL_0:
 			deemph_reg_value = 85;
 			margin_reg_value = 78;
 			break;
-		case DP_TRAIN_VOLTAGE_SWING_600:
+		case DP_TRAIN_VOLTAGE_SWING_LEVEL_1:
 			deemph_reg_value = 85;
 			margin_reg_value = 116;
 			break;
-		case DP_TRAIN_VOLTAGE_SWING_800:
+		case DP_TRAIN_VOLTAGE_SWING_LEVEL_2:
 			deemph_reg_value = 85;
 			margin_reg_value = 154;
 			break;
@@ -2613,13 +2919,13 @@ static uint32_t intel_chv_signal_levels(struct intel_dp *intel_dp)
 			return 0;
 		}
 		break;
-	case DP_TRAIN_PRE_EMPHASIS_6:
+	case DP_TRAIN_PRE_EMPH_LEVEL_2:
 		switch (train_set & DP_TRAIN_VOLTAGE_SWING_MASK) {
-		case DP_TRAIN_VOLTAGE_SWING_400:
+		case DP_TRAIN_VOLTAGE_SWING_LEVEL_0:
 			deemph_reg_value = 64;
 			margin_reg_value = 104;
 			break;
-		case DP_TRAIN_VOLTAGE_SWING_600:
+		case DP_TRAIN_VOLTAGE_SWING_LEVEL_1:
 			deemph_reg_value = 64;
 			margin_reg_value = 154;
 			break;
@@ -2627,9 +2933,9 @@ static uint32_t intel_chv_signal_levels(struct intel_dp *intel_dp)
 			return 0;
 		}
 		break;
-	case DP_TRAIN_PRE_EMPHASIS_9_5:
+	case DP_TRAIN_PRE_EMPH_LEVEL_3:
 		switch (train_set & DP_TRAIN_VOLTAGE_SWING_MASK) {
-		case DP_TRAIN_VOLTAGE_SWING_400:
+		case DP_TRAIN_VOLTAGE_SWING_LEVEL_0:
 			deemph_reg_value = 43;
 			margin_reg_value = 154;
 			break;
@@ -2646,12 +2952,26 @@ static uint32_t intel_chv_signal_levels(struct intel_dp *intel_dp)
 	/* Clear calc init */
 	val = vlv_dpio_read(dev_priv, pipe, VLV_PCS01_DW10(ch));
 	val &= ~(DPIO_PCS_SWING_CALC_TX0_TX2 | DPIO_PCS_SWING_CALC_TX1_TX3);
+	val &= ~(DPIO_PCS_TX1DEEMP_MASK | DPIO_PCS_TX2DEEMP_MASK);
+	val |= DPIO_PCS_TX1DEEMP_9P5 | DPIO_PCS_TX2DEEMP_9P5;
 	vlv_dpio_write(dev_priv, pipe, VLV_PCS01_DW10(ch), val);
 
 	val = vlv_dpio_read(dev_priv, pipe, VLV_PCS23_DW10(ch));
 	val &= ~(DPIO_PCS_SWING_CALC_TX0_TX2 | DPIO_PCS_SWING_CALC_TX1_TX3);
+	val &= ~(DPIO_PCS_TX1DEEMP_MASK | DPIO_PCS_TX2DEEMP_MASK);
+	val |= DPIO_PCS_TX1DEEMP_9P5 | DPIO_PCS_TX2DEEMP_9P5;
 	vlv_dpio_write(dev_priv, pipe, VLV_PCS23_DW10(ch), val);
 
+	val = vlv_dpio_read(dev_priv, pipe, VLV_PCS01_DW9(ch));
+	val &= ~(DPIO_PCS_TX1MARGIN_MASK | DPIO_PCS_TX2MARGIN_MASK);
+	val |= DPIO_PCS_TX1MARGIN_000 | DPIO_PCS_TX2MARGIN_000;
+	vlv_dpio_write(dev_priv, pipe, VLV_PCS01_DW9(ch), val);
+
+	val = vlv_dpio_read(dev_priv, pipe, VLV_PCS23_DW9(ch));
+	val &= ~(DPIO_PCS_TX1MARGIN_MASK | DPIO_PCS_TX2MARGIN_MASK);
+	val |= DPIO_PCS_TX1MARGIN_000 | DPIO_PCS_TX2MARGIN_000;
+	vlv_dpio_write(dev_priv, pipe, VLV_PCS23_DW9(ch), val);
+
 	/* Program swing deemph */
 	for (i = 0; i < 4; i++) {
 		val = vlv_dpio_read(dev_priv, pipe, CHV_TX_DW4(ch, i));
@@ -2663,8 +2983,8 @@ static uint32_t intel_chv_signal_levels(struct intel_dp *intel_dp)
 	/* Program swing margin */
 	for (i = 0; i < 4; i++) {
 		val = vlv_dpio_read(dev_priv, pipe, CHV_TX_DW2(ch, i));
-		val &= ~DPIO_SWING_MARGIN_MASK;
-		val |= margin_reg_value << DPIO_SWING_MARGIN_SHIFT;
+		val &= ~DPIO_SWING_MARGIN000_MASK;
+		val |= margin_reg_value << DPIO_SWING_MARGIN000_SHIFT;
 		vlv_dpio_write(dev_priv, pipe, CHV_TX_DW2(ch, i), val);
 	}
 
@@ -2676,9 +2996,9 @@ static uint32_t intel_chv_signal_levels(struct intel_dp *intel_dp)
 	}
 
 	if (((train_set & DP_TRAIN_PRE_EMPHASIS_MASK)
-			== DP_TRAIN_PRE_EMPHASIS_0) &&
+			== DP_TRAIN_PRE_EMPH_LEVEL_0) &&
 		((train_set & DP_TRAIN_VOLTAGE_SWING_MASK)
-			== DP_TRAIN_VOLTAGE_SWING_1200)) {
+			== DP_TRAIN_VOLTAGE_SWING_LEVEL_3)) {
 
 		/*
 		 * The document said it needs to set bit 27 for ch0 and bit 26
@@ -2757,32 +3077,32 @@ intel_gen4_signal_levels(uint8_t train_set)
 	uint32_t	signal_levels = 0;
 
 	switch (train_set & DP_TRAIN_VOLTAGE_SWING_MASK) {
-	case DP_TRAIN_VOLTAGE_SWING_400:
+	case DP_TRAIN_VOLTAGE_SWING_LEVEL_0:
 	default:
 		signal_levels |= DP_VOLTAGE_0_4;
 		break;
-	case DP_TRAIN_VOLTAGE_SWING_600:
+	case DP_TRAIN_VOLTAGE_SWING_LEVEL_1:
 		signal_levels |= DP_VOLTAGE_0_6;
 		break;
-	case DP_TRAIN_VOLTAGE_SWING_800:
+	case DP_TRAIN_VOLTAGE_SWING_LEVEL_2:
 		signal_levels |= DP_VOLTAGE_0_8;
 		break;
-	case DP_TRAIN_VOLTAGE_SWING_1200:
+	case DP_TRAIN_VOLTAGE_SWING_LEVEL_3:
 		signal_levels |= DP_VOLTAGE_1_2;
 		break;
 	}
 	switch (train_set & DP_TRAIN_PRE_EMPHASIS_MASK) {
-	case DP_TRAIN_PRE_EMPHASIS_0:
+	case DP_TRAIN_PRE_EMPH_LEVEL_0:
 	default:
 		signal_levels |= DP_PRE_EMPHASIS_0;
 		break;
-	case DP_TRAIN_PRE_EMPHASIS_3_5:
+	case DP_TRAIN_PRE_EMPH_LEVEL_1:
 		signal_levels |= DP_PRE_EMPHASIS_3_5;
 		break;
-	case DP_TRAIN_PRE_EMPHASIS_6:
+	case DP_TRAIN_PRE_EMPH_LEVEL_2:
 		signal_levels |= DP_PRE_EMPHASIS_6;
 		break;
-	case DP_TRAIN_PRE_EMPHASIS_9_5:
+	case DP_TRAIN_PRE_EMPH_LEVEL_3:
 		signal_levels |= DP_PRE_EMPHASIS_9_5;
 		break;
 	}
@@ -2796,19 +3116,19 @@ intel_gen6_edp_signal_levels(uint8_t train_set)
 	int signal_levels = train_set & (DP_TRAIN_VOLTAGE_SWING_MASK |
 					 DP_TRAIN_PRE_EMPHASIS_MASK);
 	switch (signal_levels) {
-	case DP_TRAIN_VOLTAGE_SWING_400 | DP_TRAIN_PRE_EMPHASIS_0:
-	case DP_TRAIN_VOLTAGE_SWING_600 | DP_TRAIN_PRE_EMPHASIS_0:
+	case DP_TRAIN_VOLTAGE_SWING_LEVEL_0 | DP_TRAIN_PRE_EMPH_LEVEL_0:
+	case DP_TRAIN_VOLTAGE_SWING_LEVEL_1 | DP_TRAIN_PRE_EMPH_LEVEL_0:
 		return EDP_LINK_TRAIN_400_600MV_0DB_SNB_B;
-	case DP_TRAIN_VOLTAGE_SWING_400 | DP_TRAIN_PRE_EMPHASIS_3_5:
+	case DP_TRAIN_VOLTAGE_SWING_LEVEL_0 | DP_TRAIN_PRE_EMPH_LEVEL_1:
 		return EDP_LINK_TRAIN_400MV_3_5DB_SNB_B;
-	case DP_TRAIN_VOLTAGE_SWING_400 | DP_TRAIN_PRE_EMPHASIS_6:
-	case DP_TRAIN_VOLTAGE_SWING_600 | DP_TRAIN_PRE_EMPHASIS_6:
+	case DP_TRAIN_VOLTAGE_SWING_LEVEL_0 | DP_TRAIN_PRE_EMPH_LEVEL_2:
+	case DP_TRAIN_VOLTAGE_SWING_LEVEL_1 | DP_TRAIN_PRE_EMPH_LEVEL_2:
 		return EDP_LINK_TRAIN_400_600MV_6DB_SNB_B;
-	case DP_TRAIN_VOLTAGE_SWING_600 | DP_TRAIN_PRE_EMPHASIS_3_5:
-	case DP_TRAIN_VOLTAGE_SWING_800 | DP_TRAIN_PRE_EMPHASIS_3_5:
+	case DP_TRAIN_VOLTAGE_SWING_LEVEL_1 | DP_TRAIN_PRE_EMPH_LEVEL_1:
+	case DP_TRAIN_VOLTAGE_SWING_LEVEL_2 | DP_TRAIN_PRE_EMPH_LEVEL_1:
 		return EDP_LINK_TRAIN_600_800MV_3_5DB_SNB_B;
-	case DP_TRAIN_VOLTAGE_SWING_800 | DP_TRAIN_PRE_EMPHASIS_0:
-	case DP_TRAIN_VOLTAGE_SWING_1200 | DP_TRAIN_PRE_EMPHASIS_0:
+	case DP_TRAIN_VOLTAGE_SWING_LEVEL_2 | DP_TRAIN_PRE_EMPH_LEVEL_0:
+	case DP_TRAIN_VOLTAGE_SWING_LEVEL_3 | DP_TRAIN_PRE_EMPH_LEVEL_0:
 		return EDP_LINK_TRAIN_800_1200MV_0DB_SNB_B;
 	default:
 		DRM_DEBUG_KMS("Unsupported voltage swing/pre-emphasis level:"
@@ -2824,21 +3144,21 @@ intel_gen7_edp_signal_levels(uint8_t train_set)
 	int signal_levels = train_set & (DP_TRAIN_VOLTAGE_SWING_MASK |
 					 DP_TRAIN_PRE_EMPHASIS_MASK);
 	switch (signal_levels) {
-	case DP_TRAIN_VOLTAGE_SWING_400 | DP_TRAIN_PRE_EMPHASIS_0:
+	case DP_TRAIN_VOLTAGE_SWING_LEVEL_0 | DP_TRAIN_PRE_EMPH_LEVEL_0:
 		return EDP_LINK_TRAIN_400MV_0DB_IVB;
-	case DP_TRAIN_VOLTAGE_SWING_400 | DP_TRAIN_PRE_EMPHASIS_3_5:
+	case DP_TRAIN_VOLTAGE_SWING_LEVEL_0 | DP_TRAIN_PRE_EMPH_LEVEL_1:
 		return EDP_LINK_TRAIN_400MV_3_5DB_IVB;
-	case DP_TRAIN_VOLTAGE_SWING_400 | DP_TRAIN_PRE_EMPHASIS_6:
+	case DP_TRAIN_VOLTAGE_SWING_LEVEL_0 | DP_TRAIN_PRE_EMPH_LEVEL_2:
 		return EDP_LINK_TRAIN_400MV_6DB_IVB;
 
-	case DP_TRAIN_VOLTAGE_SWING_600 | DP_TRAIN_PRE_EMPHASIS_0:
+	case DP_TRAIN_VOLTAGE_SWING_LEVEL_1 | DP_TRAIN_PRE_EMPH_LEVEL_0:
 		return EDP_LINK_TRAIN_600MV_0DB_IVB;
-	case DP_TRAIN_VOLTAGE_SWING_600 | DP_TRAIN_PRE_EMPHASIS_3_5:
+	case DP_TRAIN_VOLTAGE_SWING_LEVEL_1 | DP_TRAIN_PRE_EMPH_LEVEL_1:
 		return EDP_LINK_TRAIN_600MV_3_5DB_IVB;
 
-	case DP_TRAIN_VOLTAGE_SWING_800 | DP_TRAIN_PRE_EMPHASIS_0:
+	case DP_TRAIN_VOLTAGE_SWING_LEVEL_2 | DP_TRAIN_PRE_EMPH_LEVEL_0:
 		return EDP_LINK_TRAIN_800MV_0DB_IVB;
-	case DP_TRAIN_VOLTAGE_SWING_800 | DP_TRAIN_PRE_EMPHASIS_3_5:
+	case DP_TRAIN_VOLTAGE_SWING_LEVEL_2 | DP_TRAIN_PRE_EMPH_LEVEL_1:
 		return EDP_LINK_TRAIN_800MV_3_5DB_IVB;
 
 	default:
@@ -2855,30 +3175,30 @@ intel_hsw_signal_levels(uint8_t train_set)
 	int signal_levels = train_set & (DP_TRAIN_VOLTAGE_SWING_MASK |
 					 DP_TRAIN_PRE_EMPHASIS_MASK);
 	switch (signal_levels) {
-	case DP_TRAIN_VOLTAGE_SWING_400 | DP_TRAIN_PRE_EMPHASIS_0:
-		return DDI_BUF_EMP_400MV_0DB_HSW;
-	case DP_TRAIN_VOLTAGE_SWING_400 | DP_TRAIN_PRE_EMPHASIS_3_5:
-		return DDI_BUF_EMP_400MV_3_5DB_HSW;
-	case DP_TRAIN_VOLTAGE_SWING_400 | DP_TRAIN_PRE_EMPHASIS_6:
-		return DDI_BUF_EMP_400MV_6DB_HSW;
-	case DP_TRAIN_VOLTAGE_SWING_400 | DP_TRAIN_PRE_EMPHASIS_9_5:
-		return DDI_BUF_EMP_400MV_9_5DB_HSW;
-
-	case DP_TRAIN_VOLTAGE_SWING_600 | DP_TRAIN_PRE_EMPHASIS_0:
-		return DDI_BUF_EMP_600MV_0DB_HSW;
-	case DP_TRAIN_VOLTAGE_SWING_600 | DP_TRAIN_PRE_EMPHASIS_3_5:
-		return DDI_BUF_EMP_600MV_3_5DB_HSW;
-	case DP_TRAIN_VOLTAGE_SWING_600 | DP_TRAIN_PRE_EMPHASIS_6:
-		return DDI_BUF_EMP_600MV_6DB_HSW;
-
-	case DP_TRAIN_VOLTAGE_SWING_800 | DP_TRAIN_PRE_EMPHASIS_0:
-		return DDI_BUF_EMP_800MV_0DB_HSW;
-	case DP_TRAIN_VOLTAGE_SWING_800 | DP_TRAIN_PRE_EMPHASIS_3_5:
-		return DDI_BUF_EMP_800MV_3_5DB_HSW;
+	case DP_TRAIN_VOLTAGE_SWING_LEVEL_0 | DP_TRAIN_PRE_EMPH_LEVEL_0:
+		return DDI_BUF_TRANS_SELECT(0);
+	case DP_TRAIN_VOLTAGE_SWING_LEVEL_0 | DP_TRAIN_PRE_EMPH_LEVEL_1:
+		return DDI_BUF_TRANS_SELECT(1);
+	case DP_TRAIN_VOLTAGE_SWING_LEVEL_0 | DP_TRAIN_PRE_EMPH_LEVEL_2:
+		return DDI_BUF_TRANS_SELECT(2);
+	case DP_TRAIN_VOLTAGE_SWING_LEVEL_0 | DP_TRAIN_PRE_EMPH_LEVEL_3:
+		return DDI_BUF_TRANS_SELECT(3);
+
+	case DP_TRAIN_VOLTAGE_SWING_LEVEL_1 | DP_TRAIN_PRE_EMPH_LEVEL_0:
+		return DDI_BUF_TRANS_SELECT(4);
+	case DP_TRAIN_VOLTAGE_SWING_LEVEL_1 | DP_TRAIN_PRE_EMPH_LEVEL_1:
+		return DDI_BUF_TRANS_SELECT(5);
+	case DP_TRAIN_VOLTAGE_SWING_LEVEL_1 | DP_TRAIN_PRE_EMPH_LEVEL_2:
+		return DDI_BUF_TRANS_SELECT(6);
+
+	case DP_TRAIN_VOLTAGE_SWING_LEVEL_2 | DP_TRAIN_PRE_EMPH_LEVEL_0:
+		return DDI_BUF_TRANS_SELECT(7);
+	case DP_TRAIN_VOLTAGE_SWING_LEVEL_2 | DP_TRAIN_PRE_EMPH_LEVEL_1:
+		return DDI_BUF_TRANS_SELECT(8);
 	default:
 		DRM_DEBUG_KMS("Unsupported voltage swing/pre-emphasis level:"
 			      "0x%x\n", signal_levels);
-		return DDI_BUF_EMP_400MV_0DB_HSW;
+		return DDI_BUF_TRANS_SELECT(0);
 	}
 }
 
@@ -2892,7 +3212,7 @@ intel_dp_set_signal_levels(struct intel_dp *intel_dp, uint32_t *DP)
 	uint32_t signal_levels, mask;
 	uint8_t train_set = intel_dp->train_set[0];
 
-	if (IS_HASWELL(dev) || IS_BROADWELL(dev)) {
+	if (IS_HASWELL(dev) || IS_BROADWELL(dev) || INTEL_INFO(dev)->gen >= 9) {
 		signal_levels = intel_hsw_signal_levels(train_set);
 		mask = DDI_BUF_EMP_MASK;
 	} else if (IS_CHERRYVIEW(dev)) {
@@ -2925,74 +3245,10 @@ intel_dp_set_link_train(struct intel_dp *intel_dp,
 	struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp);
 	struct drm_device *dev = intel_dig_port->base.base.dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	enum port port = intel_dig_port->port;
 	uint8_t buf[sizeof(intel_dp->train_set) + 1];
 	int ret, len;
 
-	if (HAS_DDI(dev)) {
-		uint32_t temp = I915_READ(DP_TP_CTL(port));
-
-		if (dp_train_pat & DP_LINK_SCRAMBLING_DISABLE)
-			temp |= DP_TP_CTL_SCRAMBLE_DISABLE;
-		else
-			temp &= ~DP_TP_CTL_SCRAMBLE_DISABLE;
-
-		temp &= ~DP_TP_CTL_LINK_TRAIN_MASK;
-		switch (dp_train_pat & DP_TRAINING_PATTERN_MASK) {
-		case DP_TRAINING_PATTERN_DISABLE:
-			temp |= DP_TP_CTL_LINK_TRAIN_NORMAL;
-
-			break;
-		case DP_TRAINING_PATTERN_1:
-			temp |= DP_TP_CTL_LINK_TRAIN_PAT1;
-			break;
-		case DP_TRAINING_PATTERN_2:
-			temp |= DP_TP_CTL_LINK_TRAIN_PAT2;
-			break;
-		case DP_TRAINING_PATTERN_3:
-			temp |= DP_TP_CTL_LINK_TRAIN_PAT3;
-			break;
-		}
-		I915_WRITE(DP_TP_CTL(port), temp);
-
-	} else if (HAS_PCH_CPT(dev) && (IS_GEN7(dev) || port != PORT_A)) {
-		*DP &= ~DP_LINK_TRAIN_MASK_CPT;
-
-		switch (dp_train_pat & DP_TRAINING_PATTERN_MASK) {
-		case DP_TRAINING_PATTERN_DISABLE:
-			*DP |= DP_LINK_TRAIN_OFF_CPT;
-			break;
-		case DP_TRAINING_PATTERN_1:
-			*DP |= DP_LINK_TRAIN_PAT_1_CPT;
-			break;
-		case DP_TRAINING_PATTERN_2:
-			*DP |= DP_LINK_TRAIN_PAT_2_CPT;
-			break;
-		case DP_TRAINING_PATTERN_3:
-			DRM_ERROR("DP training pattern 3 not supported\n");
-			*DP |= DP_LINK_TRAIN_PAT_2_CPT;
-			break;
-		}
-
-	} else {
-		*DP &= ~DP_LINK_TRAIN_MASK;
-
-		switch (dp_train_pat & DP_TRAINING_PATTERN_MASK) {
-		case DP_TRAINING_PATTERN_DISABLE:
-			*DP |= DP_LINK_TRAIN_OFF;
-			break;
-		case DP_TRAINING_PATTERN_1:
-			*DP |= DP_LINK_TRAIN_PAT_1;
-			break;
-		case DP_TRAINING_PATTERN_2:
-			*DP |= DP_LINK_TRAIN_PAT_2;
-			break;
-		case DP_TRAINING_PATTERN_3:
-			DRM_ERROR("DP training pattern 3 not supported\n");
-			*DP |= DP_LINK_TRAIN_PAT_2;
-			break;
-		}
-	}
+	_intel_dp_set_link_train(intel_dp, DP, dp_train_pat);
 
 	I915_WRITE(intel_dp->output_reg, *DP);
 	POSTING_READ(intel_dp->output_reg);
@@ -3220,7 +3476,6 @@ intel_dp_complete_link_train(struct intel_dp *intel_dp)
 
 		/* Try 5 times, then try clock recovery if that fails */
 		if (tries > 5) {
-			intel_dp_link_down(intel_dp);
 			intel_dp_start_link_train(intel_dp);
 			intel_dp_set_link_train(intel_dp, &DP,
 						training_pattern |
@@ -3276,7 +3531,10 @@ intel_dp_link_down(struct intel_dp *intel_dp)
 		DP &= ~DP_LINK_TRAIN_MASK_CPT;
 		I915_WRITE(intel_dp->output_reg, DP | DP_LINK_TRAIN_PAT_IDLE_CPT);
 	} else {
-		DP &= ~DP_LINK_TRAIN_MASK;
+		if (IS_CHERRYVIEW(dev))
+			DP &= ~DP_LINK_TRAIN_MASK_CHV;
+		else
+			DP &= ~DP_LINK_TRAIN_MASK;
 		I915_WRITE(intel_dp->output_reg, DP | DP_LINK_TRAIN_PAT_IDLE);
 	}
 	POSTING_READ(intel_dp->output_reg);
@@ -3322,15 +3580,11 @@ intel_dp_get_dpcd(struct intel_dp *intel_dp)
 	struct drm_device *dev = dig_port->base.base.dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 
-	char dpcd_hex_dump[sizeof(intel_dp->dpcd) * 3];
-
 	if (intel_dp_dpcd_read_wake(&intel_dp->aux, 0x000, intel_dp->dpcd,
 				    sizeof(intel_dp->dpcd)) < 0)
 		return false; /* aux transfer failed */
 
-	hex_dump_to_buffer(intel_dp->dpcd, sizeof(intel_dp->dpcd),
-			   32, 1, dpcd_hex_dump, sizeof(dpcd_hex_dump), false);
-	DRM_DEBUG_KMS("DPCD: %s\n", dpcd_hex_dump);
+	DRM_DEBUG_KMS("DPCD: %*ph\n", (int) sizeof(intel_dp->dpcd), intel_dp->dpcd);
 
 	if (intel_dp->dpcd[DP_DPCD_REV] == 0)
 		return false; /* DPCD not present */
@@ -3347,11 +3601,12 @@ intel_dp_get_dpcd(struct intel_dp *intel_dp)
 		}
 	}
 
-	/* Training Pattern 3 support */
+	/* Training Pattern 3 support, both source and sink */
 	if (intel_dp->dpcd[DP_DPCD_REV] >= 0x12 &&
-	    intel_dp->dpcd[DP_MAX_LANE_COUNT] & DP_TPS3_SUPPORTED) {
+	    intel_dp->dpcd[DP_MAX_LANE_COUNT] & DP_TPS3_SUPPORTED &&
+	    (IS_HASWELL(dev_priv) || INTEL_INFO(dev_priv)->gen >= 8)) {
 		intel_dp->use_tps3 = true;
-		DRM_DEBUG_KMS("Displayport TPS3 supported");
+		DRM_DEBUG_KMS("Displayport TPS3 supported\n");
 	} else
 		intel_dp->use_tps3 = false;
 
@@ -3378,8 +3633,6 @@ intel_dp_probe_oui(struct intel_dp *intel_dp)
 	if (!(intel_dp->dpcd[DP_DOWN_STREAM_PORT_COUNT] & DP_OUI_SUPPORT))
 		return;
 
-	intel_edp_panel_vdd_on(intel_dp);
-
 	if (intel_dp_dpcd_read_wake(&intel_dp->aux, DP_SINK_OUI, buf, 3) == 3)
 		DRM_DEBUG_KMS("Sink OUI: %02hx%02hx%02hx\n",
 			      buf[0], buf[1], buf[2]);
@@ -3387,8 +3640,6 @@ intel_dp_probe_oui(struct intel_dp *intel_dp)
 	if (intel_dp_dpcd_read_wake(&intel_dp->aux, DP_BRANCH_OUI, buf, 3) == 3)
 		DRM_DEBUG_KMS("Branch OUI: %02hx%02hx%02hx\n",
 			      buf[0], buf[1], buf[2]);
-
-	edp_panel_vdd_off(intel_dp, false);
 }
 
 static bool
@@ -3402,7 +3653,6 @@ intel_dp_probe_mst(struct intel_dp *intel_dp)
 	if (intel_dp->dpcd[DP_DPCD_REV] < 0x12)
 		return false;
 
-	_edp_panel_vdd_on(intel_dp);
 	if (intel_dp_dpcd_read_wake(&intel_dp->aux, DP_MSTM_CAP, buf, 1)) {
 		if (buf[0] & DP_MST_CAP) {
 			DRM_DEBUG_KMS("Sink is MST capable\n");
@@ -3412,7 +3662,6 @@ intel_dp_probe_mst(struct intel_dp *intel_dp)
 			intel_dp->is_mst = false;
 		}
 	}
-	edp_panel_vdd_off(intel_dp, false);
 
 	drm_dp_mst_topology_mgr_set_mst(&intel_dp->mst_mgr, intel_dp->is_mst);
 	return intel_dp->is_mst;
@@ -3424,26 +3673,48 @@ int intel_dp_sink_crc(struct intel_dp *intel_dp, u8 *crc)
 	struct drm_device *dev = intel_dig_port->base.base.dev;
 	struct intel_crtc *intel_crtc =
 		to_intel_crtc(intel_dig_port->base.base.crtc);
-	u8 buf[1];
+	u8 buf;
+	int test_crc_count;
+	int attempts = 6;
 
-	if (drm_dp_dpcd_readb(&intel_dp->aux, DP_TEST_SINK_MISC, buf) < 0)
-		return -EAGAIN;
+	if (drm_dp_dpcd_readb(&intel_dp->aux, DP_TEST_SINK_MISC, &buf) < 0)
+		return -EIO;
 
-	if (!(buf[0] & DP_TEST_CRC_SUPPORTED))
+	if (!(buf & DP_TEST_CRC_SUPPORTED))
 		return -ENOTTY;
 
+	if (drm_dp_dpcd_readb(&intel_dp->aux, DP_TEST_SINK, &buf) < 0)
+		return -EIO;
+
 	if (drm_dp_dpcd_writeb(&intel_dp->aux, DP_TEST_SINK,
-			       DP_TEST_SINK_START) < 0)
-		return -EAGAIN;
+				buf | DP_TEST_SINK_START) < 0)
+		return -EIO;
+
+	if (drm_dp_dpcd_readb(&intel_dp->aux, DP_TEST_SINK_MISC, &buf) < 0)
+		return -EIO;
+	test_crc_count = buf & DP_TEST_COUNT_MASK;
+
+	do {
+		if (drm_dp_dpcd_readb(&intel_dp->aux,
+				      DP_TEST_SINK_MISC, &buf) < 0)
+			return -EIO;
+		intel_wait_for_vblank(dev, intel_crtc->pipe);
+	} while (--attempts && (buf & DP_TEST_COUNT_MASK) == test_crc_count);
 
-	/* Wait 2 vblanks to be sure we will have the correct CRC value */
-	intel_wait_for_vblank(dev, intel_crtc->pipe);
-	intel_wait_for_vblank(dev, intel_crtc->pipe);
+	if (attempts == 0) {
+		DRM_DEBUG_KMS("Panel is unable to calculate CRC after 6 vblanks\n");
+		return -ETIMEDOUT;
+	}
 
 	if (drm_dp_dpcd_read(&intel_dp->aux, DP_TEST_CRC_R_CR, crc, 6) < 0)
-		return -EAGAIN;
+		return -EIO;
+
+	if (drm_dp_dpcd_readb(&intel_dp->aux, DP_TEST_SINK, &buf) < 0)
+		return -EIO;
+	if (drm_dp_dpcd_writeb(&intel_dp->aux, DP_TEST_SINK,
+			       buf & ~DP_TEST_SINK_START) < 0)
+		return -EIO;
 
-	drm_dp_dpcd_writeb(&intel_dp->aux, DP_TEST_SINK, 0);
 	return 0;
 }
 
@@ -3644,20 +3915,24 @@ intel_dp_detect_dpcd(struct intel_dp *intel_dp)
 }
 
 static enum drm_connector_status
+edp_detect(struct intel_dp *intel_dp)
+{
+	struct drm_device *dev = intel_dp_to_dev(intel_dp);
+	enum drm_connector_status status;
+
+	status = intel_panel_detect(dev);
+	if (status == connector_status_unknown)
+		status = connector_status_connected;
+
+	return status;
+}
+
+static enum drm_connector_status
 ironlake_dp_detect(struct intel_dp *intel_dp)
 {
 	struct drm_device *dev = intel_dp_to_dev(intel_dp);
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp);
-	enum drm_connector_status status;
-
-	/* Can't disconnect eDP, but you can close the lid... */
-	if (is_edp(intel_dp)) {
-		status = intel_panel_detect(dev);
-		if (status == connector_status_unknown)
-			status = connector_status_connected;
-		return status;
-	}
 
 	if (!ibx_digital_port_connected(dev_priv, intel_dig_port))
 		return connector_status_disconnected;
@@ -3733,9 +4008,9 @@ g4x_dp_detect(struct intel_dp *intel_dp)
 }
 
 static struct edid *
-intel_dp_get_edid(struct drm_connector *connector, struct i2c_adapter *adapter)
+intel_dp_get_edid(struct intel_dp *intel_dp)
 {
-	struct intel_connector *intel_connector = to_intel_connector(connector);
+	struct intel_connector *intel_connector = intel_dp->attached_connector;
 
 	/* use cached edid if we have one */
 	if (intel_connector->edid) {
@@ -3744,27 +4019,55 @@ intel_dp_get_edid(struct drm_connector *connector, struct i2c_adapter *adapter)
 			return NULL;
 
 		return drm_edid_duplicate(intel_connector->edid);
-	}
+	} else
+		return drm_get_edid(&intel_connector->base,
+				    &intel_dp->aux.ddc);
+}
 
-	return drm_get_edid(connector, adapter);
+static void
+intel_dp_set_edid(struct intel_dp *intel_dp)
+{
+	struct intel_connector *intel_connector = intel_dp->attached_connector;
+	struct edid *edid;
+
+	edid = intel_dp_get_edid(intel_dp);
+	intel_connector->detect_edid = edid;
+
+	if (intel_dp->force_audio != HDMI_AUDIO_AUTO)
+		intel_dp->has_audio = intel_dp->force_audio == HDMI_AUDIO_ON;
+	else
+		intel_dp->has_audio = drm_detect_monitor_audio(edid);
 }
 
-static int
-intel_dp_get_edid_modes(struct drm_connector *connector, struct i2c_adapter *adapter)
+static void
+intel_dp_unset_edid(struct intel_dp *intel_dp)
 {
-	struct intel_connector *intel_connector = to_intel_connector(connector);
+	struct intel_connector *intel_connector = intel_dp->attached_connector;
 
-	/* use cached edid if we have one */
-	if (intel_connector->edid) {
-		/* invalid edid */
-		if (IS_ERR(intel_connector->edid))
-			return 0;
+	kfree(intel_connector->detect_edid);
+	intel_connector->detect_edid = NULL;
 
-		return intel_connector_update_modes(connector,
-						    intel_connector->edid);
-	}
+	intel_dp->has_audio = false;
+}
+
+static enum intel_display_power_domain
+intel_dp_power_get(struct intel_dp *dp)
+{
+	struct intel_encoder *encoder = &dp_to_dig_port(dp)->base;
+	enum intel_display_power_domain power_domain;
 
-	return intel_ddc_get_modes(connector, adapter);
+	power_domain = intel_display_port_power_domain(encoder);
+	intel_display_power_get(to_i915(encoder->base.dev), power_domain);
+
+	return power_domain;
+}
+
+static void
+intel_dp_power_put(struct intel_dp *dp,
+		   enum intel_display_power_domain power_domain)
+{
+	struct intel_encoder *encoder = &dp_to_dig_port(dp)->base;
+	intel_display_power_put(to_i915(encoder->base.dev), power_domain);
 }
 
 static enum drm_connector_status
@@ -3774,33 +4077,30 @@ intel_dp_detect(struct drm_connector *connector, bool force)
 	struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp);
 	struct intel_encoder *intel_encoder = &intel_dig_port->base;
 	struct drm_device *dev = connector->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
 	enum drm_connector_status status;
 	enum intel_display_power_domain power_domain;
-	struct edid *edid = NULL;
 	bool ret;
 
-	power_domain = intel_display_port_power_domain(intel_encoder);
-	intel_display_power_get(dev_priv, power_domain);
-
 	DRM_DEBUG_KMS("[CONNECTOR:%d:%s]\n",
 		      connector->base.id, connector->name);
+	intel_dp_unset_edid(intel_dp);
 
 	if (intel_dp->is_mst) {
 		/* MST devices are disconnected from a monitor POV */
 		if (intel_encoder->type != INTEL_OUTPUT_EDP)
 			intel_encoder->type = INTEL_OUTPUT_DISPLAYPORT;
-		status = connector_status_disconnected;
-		goto out;
+		return connector_status_disconnected;
 	}
 
-	intel_dp->has_audio = false;
+	power_domain = intel_dp_power_get(intel_dp);
 
-	if (HAS_PCH_SPLIT(dev))
+	/* Can't disconnect eDP, but you can close the lid... */
+	if (is_edp(intel_dp))
+		status = edp_detect(intel_dp);
+	else if (HAS_PCH_SPLIT(dev))
 		status = ironlake_dp_detect(intel_dp);
 	else
 		status = g4x_dp_detect(intel_dp);
-
 	if (status != connector_status_connected)
 		goto out;
 
@@ -3816,82 +4116,78 @@ intel_dp_detect(struct drm_connector *connector, bool force)
 		goto out;
 	}
 
-	if (intel_dp->force_audio != HDMI_AUDIO_AUTO) {
-		intel_dp->has_audio = (intel_dp->force_audio == HDMI_AUDIO_ON);
-	} else {
-		edid = intel_dp_get_edid(connector, &intel_dp->aux.ddc);
-		if (edid) {
-			intel_dp->has_audio = drm_detect_monitor_audio(edid);
-			kfree(edid);
-		}
-	}
+	intel_dp_set_edid(intel_dp);
 
 	if (intel_encoder->type != INTEL_OUTPUT_EDP)
 		intel_encoder->type = INTEL_OUTPUT_DISPLAYPORT;
 	status = connector_status_connected;
 
 out:
-	intel_display_power_put(dev_priv, power_domain);
+	intel_dp_power_put(intel_dp, power_domain);
 	return status;
 }
 
-static int intel_dp_get_modes(struct drm_connector *connector)
+static void
+intel_dp_force(struct drm_connector *connector)
 {
 	struct intel_dp *intel_dp = intel_attached_dp(connector);
-	struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp);
-	struct intel_encoder *intel_encoder = &intel_dig_port->base;
-	struct intel_connector *intel_connector = to_intel_connector(connector);
-	struct drm_device *dev = connector->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct intel_encoder *intel_encoder = &dp_to_dig_port(intel_dp)->base;
 	enum intel_display_power_domain power_domain;
-	int ret;
 
-	/* We should parse the EDID data and find out if it has an audio sink
-	 */
+	DRM_DEBUG_KMS("[CONNECTOR:%d:%s]\n",
+		      connector->base.id, connector->name);
+	intel_dp_unset_edid(intel_dp);
 
-	power_domain = intel_display_port_power_domain(intel_encoder);
-	intel_display_power_get(dev_priv, power_domain);
+	if (connector->status != connector_status_connected)
+		return;
 
-	ret = intel_dp_get_edid_modes(connector, &intel_dp->aux.ddc);
-	intel_display_power_put(dev_priv, power_domain);
-	if (ret)
-		return ret;
+	power_domain = intel_dp_power_get(intel_dp);
+
+	intel_dp_set_edid(intel_dp);
+
+	intel_dp_power_put(intel_dp, power_domain);
+
+	if (intel_encoder->type != INTEL_OUTPUT_EDP)
+		intel_encoder->type = INTEL_OUTPUT_DISPLAYPORT;
+}
+
+static int intel_dp_get_modes(struct drm_connector *connector)
+{
+	struct intel_connector *intel_connector = to_intel_connector(connector);
+	struct edid *edid;
+
+	edid = intel_connector->detect_edid;
+	if (edid) {
+		int ret = intel_connector_update_modes(connector, edid);
+		if (ret)
+			return ret;
+	}
 
 	/* if eDP has no EDID, fall back to fixed mode */
-	if (is_edp(intel_dp) && intel_connector->panel.fixed_mode) {
+	if (is_edp(intel_attached_dp(connector)) &&
+	    intel_connector->panel.fixed_mode) {
 		struct drm_display_mode *mode;
-		mode = drm_mode_duplicate(dev,
+
+		mode = drm_mode_duplicate(connector->dev,
 					  intel_connector->panel.fixed_mode);
 		if (mode) {
 			drm_mode_probed_add(connector, mode);
 			return 1;
 		}
 	}
+
 	return 0;
 }
 
 static bool
 intel_dp_detect_audio(struct drm_connector *connector)
 {
-	struct intel_dp *intel_dp = intel_attached_dp(connector);
-	struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp);
-	struct intel_encoder *intel_encoder = &intel_dig_port->base;
-	struct drm_device *dev = connector->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	enum intel_display_power_domain power_domain;
-	struct edid *edid;
 	bool has_audio = false;
+	struct edid *edid;
 
-	power_domain = intel_display_port_power_domain(intel_encoder);
-	intel_display_power_get(dev_priv, power_domain);
-
-	edid = intel_dp_get_edid(connector, &intel_dp->aux.ddc);
-	if (edid) {
+	edid = to_intel_connector(connector)->detect_edid;
+	if (edid)
 		has_audio = drm_detect_monitor_audio(edid);
-		kfree(edid);
-	}
-
-	intel_display_power_put(dev_priv, power_domain);
 
 	return has_audio;
 }
@@ -3989,6 +4285,8 @@ intel_dp_connector_destroy(struct drm_connector *connector)
 {
 	struct intel_connector *intel_connector = to_intel_connector(connector);
 
+	kfree(intel_connector->detect_edid);
+
 	if (!IS_ERR_OR_NULL(intel_connector->edid))
 		kfree(intel_connector->edid);
 
@@ -4005,16 +4303,20 @@ void intel_dp_encoder_destroy(struct drm_encoder *encoder)
 {
 	struct intel_digital_port *intel_dig_port = enc_to_dig_port(encoder);
 	struct intel_dp *intel_dp = &intel_dig_port->dp;
-	struct drm_device *dev = intel_dp_to_dev(intel_dp);
 
 	drm_dp_aux_unregister(&intel_dp->aux);
 	intel_dp_mst_encoder_cleanup(intel_dig_port);
 	drm_encoder_cleanup(encoder);
 	if (is_edp(intel_dp)) {
 		cancel_delayed_work_sync(&intel_dp->panel_vdd_work);
-		drm_modeset_lock(&dev->mode_config.connection_mutex, NULL);
+		/*
+		 * vdd might still be enabled do to the delayed vdd off.
+		 * Make sure vdd is actually turned off here.
+		 */
+		pps_lock(intel_dp);
 		edp_panel_vdd_off_sync(intel_dp);
-		drm_modeset_unlock(&dev->mode_config.connection_mutex);
+		pps_unlock(intel_dp);
+
 		if (intel_dp->edp_notifier.notifier_call) {
 			unregister_reboot_notifier(&intel_dp->edp_notifier);
 			intel_dp->edp_notifier.notifier_call = NULL;
@@ -4030,17 +4332,68 @@ static void intel_dp_encoder_suspend(struct intel_encoder *intel_encoder)
 	if (!is_edp(intel_dp))
 		return;
 
+	/*
+	 * vdd might still be enabled do to the delayed vdd off.
+	 * Make sure vdd is actually turned off here.
+	 */
+	cancel_delayed_work_sync(&intel_dp->panel_vdd_work);
+	pps_lock(intel_dp);
 	edp_panel_vdd_off_sync(intel_dp);
+	pps_unlock(intel_dp);
+}
+
+static void intel_edp_panel_vdd_sanitize(struct intel_dp *intel_dp)
+{
+	struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp);
+	struct drm_device *dev = intel_dig_port->base.base.dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	enum intel_display_power_domain power_domain;
+
+	lockdep_assert_held(&dev_priv->pps_mutex);
+
+	if (!edp_have_panel_vdd(intel_dp))
+		return;
+
+	/*
+	 * The VDD bit needs a power domain reference, so if the bit is
+	 * already enabled when we boot or resume, grab this reference and
+	 * schedule a vdd off, so we don't hold on to the reference
+	 * indefinitely.
+	 */
+	DRM_DEBUG_KMS("VDD left on by BIOS, adjusting state tracking\n");
+	power_domain = intel_display_port_power_domain(&intel_dig_port->base);
+	intel_display_power_get(dev_priv, power_domain);
+
+	edp_panel_vdd_schedule_off(intel_dp);
 }
 
 static void intel_dp_encoder_reset(struct drm_encoder *encoder)
 {
-	intel_edp_panel_vdd_sanitize(to_intel_encoder(encoder));
+	struct intel_dp *intel_dp;
+
+	if (to_intel_encoder(encoder)->type != INTEL_OUTPUT_EDP)
+		return;
+
+	intel_dp = enc_to_intel_dp(encoder);
+
+	pps_lock(intel_dp);
+
+	/*
+	 * Read out the current power sequencer assignment,
+	 * in case the BIOS did something with it.
+	 */
+	if (IS_VALLEYVIEW(encoder->dev))
+		vlv_initial_power_sequencer_setup(intel_dp);
+
+	intel_edp_panel_vdd_sanitize(intel_dp);
+
+	pps_unlock(intel_dp);
 }
 
 static const struct drm_connector_funcs intel_dp_connector_funcs = {
 	.dpms = intel_connector_dpms,
 	.detect = intel_dp_detect,
+	.force = intel_dp_force,
 	.fill_modes = drm_helper_probe_single_connector_modes,
 	.set_property = intel_dp_set_property,
 	.destroy = intel_dp_connector_destroy,
@@ -4076,7 +4429,20 @@ intel_dp_hpd_pulse(struct intel_digital_port *intel_dig_port, bool long_hpd)
 	if (intel_dig_port->base.type != INTEL_OUTPUT_EDP)
 		intel_dig_port->base.type = INTEL_OUTPUT_DISPLAYPORT;
 
-	DRM_DEBUG_KMS("got hpd irq on port %d - %s\n", intel_dig_port->port,
+	if (long_hpd && intel_dig_port->base.type == INTEL_OUTPUT_EDP) {
+		/*
+		 * vdd off can generate a long pulse on eDP which
+		 * would require vdd on to handle it, and thus we
+		 * would end up in an endless cycle of
+		 * "vdd off -> long hpd -> vdd on -> detect -> vdd off -> ..."
+		 */
+		DRM_DEBUG_KMS("ignoring long hpd on eDP port %c\n",
+			      port_name(intel_dig_port->port));
+		return false;
+	}
+
+	DRM_DEBUG_KMS("got hpd irq on port %c - %s\n",
+		      port_name(intel_dig_port->port),
 		      long_hpd ? "long" : "short");
 
 	power_domain = intel_display_port_power_domain(intel_encoder);
@@ -4208,14 +4574,20 @@ static void intel_dp_init_panel_power_timestamps(struct intel_dp *intel_dp)
 
 static void
 intel_dp_init_panel_power_sequencer(struct drm_device *dev,
-				    struct intel_dp *intel_dp,
-				    struct edp_power_seq *out)
+				    struct intel_dp *intel_dp)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct edp_power_seq cur, vbt, spec, final;
+	struct edp_power_seq cur, vbt, spec,
+		*final = &intel_dp->pps_delays;
 	u32 pp_on, pp_off, pp_div, pp;
 	int pp_ctrl_reg, pp_on_reg, pp_off_reg, pp_div_reg;
 
+	lockdep_assert_held(&dev_priv->pps_mutex);
+
+	/* already initialized? */
+	if (final->t11_t12 != 0)
+		return;
+
 	if (HAS_PCH_SPLIT(dev)) {
 		pp_ctrl_reg = PCH_PP_CONTROL;
 		pp_on_reg = PCH_PP_ON_DELAYS;
@@ -4277,7 +4649,7 @@ intel_dp_init_panel_power_sequencer(struct drm_device *dev,
 
 	/* Use the max of the register settings and vbt. If both are
 	 * unset, fall back to the spec limits. */
-#define assign_final(field)	final.field = (max(cur.field, vbt.field) == 0 ? \
+#define assign_final(field)	final->field = (max(cur.field, vbt.field) == 0 ? \
 				       spec.field : \
 				       max(cur.field, vbt.field))
 	assign_final(t1_t3);
@@ -4287,7 +4659,7 @@ intel_dp_init_panel_power_sequencer(struct drm_device *dev,
 	assign_final(t11_t12);
 #undef assign_final
 
-#define get_delay(field)	(DIV_ROUND_UP(final.field, 10))
+#define get_delay(field)	(DIV_ROUND_UP(final->field, 10))
 	intel_dp->panel_power_up_delay = get_delay(t1_t3);
 	intel_dp->backlight_on_delay = get_delay(t8);
 	intel_dp->backlight_off_delay = get_delay(t9);
@@ -4301,20 +4673,20 @@ intel_dp_init_panel_power_sequencer(struct drm_device *dev,
 
 	DRM_DEBUG_KMS("backlight on delay %d, off delay %d\n",
 		      intel_dp->backlight_on_delay, intel_dp->backlight_off_delay);
-
-	if (out)
-		*out = final;
 }
 
 static void
 intel_dp_init_panel_power_sequencer_registers(struct drm_device *dev,
-					      struct intel_dp *intel_dp,
-					      struct edp_power_seq *seq)
+					      struct intel_dp *intel_dp)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	u32 pp_on, pp_off, pp_div, port_sel = 0;
 	int div = HAS_PCH_SPLIT(dev) ? intel_pch_rawclk(dev) : intel_hrawclk(dev);
 	int pp_on_reg, pp_off_reg, pp_div_reg;
+	enum port port = dp_to_dig_port(intel_dp)->port;
+	const struct edp_power_seq *seq = &intel_dp->pps_delays;
+
+	lockdep_assert_held(&dev_priv->pps_mutex);
 
 	if (HAS_PCH_SPLIT(dev)) {
 		pp_on_reg = PCH_PP_ON_DELAYS;
@@ -4349,12 +4721,9 @@ intel_dp_init_panel_power_sequencer_registers(struct drm_device *dev,
 	/* Haswell doesn't have any port selection bits for the panel
 	 * power sequencer any more. */
 	if (IS_VALLEYVIEW(dev)) {
-		if (dp_to_dig_port(intel_dp)->port == PORT_B)
-			port_sel = PANEL_PORT_SELECT_DPB_VLV;
-		else
-			port_sel = PANEL_PORT_SELECT_DPC_VLV;
+		port_sel = PANEL_PORT_SELECT_VLV(port);
 	} else if (HAS_PCH_IBX(dev) || HAS_PCH_CPT(dev)) {
-		if (dp_to_dig_port(intel_dp)->port == PORT_A)
+		if (port == PORT_A)
 			port_sel = PANEL_PORT_SELECT_DPA;
 		else
 			port_sel = PANEL_PORT_SELECT_DPD;
@@ -4398,7 +4767,7 @@ void intel_dp_set_drrs_state(struct drm_device *dev, int refresh_rate)
 	 * hard to tell without seeing the user of this function of this code.
 	 * Check locking and ordering once that lands.
 	 */
-	if (INTEL_INFO(dev)->gen < 8 && intel_edp_is_psr_enabled(dev)) {
+	if (INTEL_INFO(dev)->gen < 8 && intel_psr_is_enabled(dev)) {
 		DRM_DEBUG_KMS("DRRS is disabled as PSR is enabled\n");
 		return;
 	}
@@ -4438,7 +4807,7 @@ void intel_dp_set_drrs_state(struct drm_device *dev, int refresh_rate)
 		val = I915_READ(reg);
 		if (index > DRRS_HIGH_RR) {
 			val |= PIPECONF_EDP_RR_MODE_SWITCH;
-			intel_dp_set_m2_n2(intel_crtc, &config->dp_m2_n2);
+			intel_dp_set_m_n(intel_crtc);
 		} else {
 			val &= ~PIPECONF_EDP_RR_MODE_SWITCH;
 		}
@@ -4478,7 +4847,7 @@ intel_dp_drrs_init(struct intel_digital_port *intel_dig_port,
 	}
 
 	if (dev_priv->vbt.drrs_type != SEAMLESS_DRRS_SUPPORT) {
-		DRM_INFO("VBT doesn't support DRRS\n");
+		DRM_DEBUG_KMS("VBT doesn't support DRRS\n");
 		return NULL;
 	}
 
@@ -4486,7 +4855,7 @@ intel_dp_drrs_init(struct intel_digital_port *intel_dig_port,
 					(dev, fixed_mode, connector);
 
 	if (!downclock_mode) {
-		DRM_INFO("DRRS not supported\n");
+		DRM_DEBUG_KMS("DRRS not supported\n");
 		return NULL;
 	}
 
@@ -4497,39 +4866,12 @@ intel_dp_drrs_init(struct intel_digital_port *intel_dig_port,
 	intel_dp->drrs_state.type = dev_priv->vbt.drrs_type;
 
 	intel_dp->drrs_state.refresh_rate_type = DRRS_HIGH_RR;
-	DRM_INFO("seamless DRRS supported for eDP panel.\n");
+	DRM_DEBUG_KMS("seamless DRRS supported for eDP panel.\n");
 	return downclock_mode;
 }
 
-void intel_edp_panel_vdd_sanitize(struct intel_encoder *intel_encoder)
-{
-	struct drm_device *dev = intel_encoder->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct intel_dp *intel_dp;
-	enum intel_display_power_domain power_domain;
-
-	if (intel_encoder->type != INTEL_OUTPUT_EDP)
-		return;
-
-	intel_dp = enc_to_intel_dp(&intel_encoder->base);
-	if (!edp_have_panel_vdd(intel_dp))
-		return;
-	/*
-	 * The VDD bit needs a power domain reference, so if the bit is
-	 * already enabled when we boot or resume, grab this reference and
-	 * schedule a vdd off, so we don't hold on to the reference
-	 * indefinitely.
-	 */
-	DRM_DEBUG_KMS("VDD left on by BIOS, adjusting state tracking\n");
-	power_domain = intel_display_port_power_domain(intel_encoder);
-	intel_display_power_get(dev_priv, power_domain);
-
-	edp_panel_vdd_schedule_off(intel_dp);
-}
-
 static bool intel_edp_init_connector(struct intel_dp *intel_dp,
-				     struct intel_connector *intel_connector,
-				     struct edp_power_seq *power_seq)
+				     struct intel_connector *intel_connector)
 {
 	struct drm_connector *connector = &intel_connector->base;
 	struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp);
@@ -4541,18 +4883,19 @@ static bool intel_edp_init_connector(struct intel_dp *intel_dp,
 	bool has_dpcd;
 	struct drm_display_mode *scan;
 	struct edid *edid;
+	enum pipe pipe = INVALID_PIPE;
 
 	intel_dp->drrs_state.type = DRRS_NOT_SUPPORTED;
 
 	if (!is_edp(intel_dp))
 		return true;
 
-	intel_edp_panel_vdd_sanitize(intel_encoder);
+	pps_lock(intel_dp);
+	intel_edp_panel_vdd_sanitize(intel_dp);
+	pps_unlock(intel_dp);
 
 	/* Cache DPCD and EDID for edp. */
-	intel_edp_panel_vdd_on(intel_dp);
 	has_dpcd = intel_dp_get_dpcd(intel_dp);
-	edp_panel_vdd_off(intel_dp, false);
 
 	if (has_dpcd) {
 		if (intel_dp->dpcd[DP_DPCD_REV] >= 0x11)
@@ -4566,7 +4909,9 @@ static bool intel_edp_init_connector(struct intel_dp *intel_dp,
 	}
 
 	/* We now know it's not a ghost, init power sequence regs. */
-	intel_dp_init_panel_power_sequencer_registers(dev, intel_dp, power_seq);
+	pps_lock(intel_dp);
+	intel_dp_init_panel_power_sequencer_registers(dev, intel_dp);
+	pps_unlock(intel_dp);
 
 	mutex_lock(&dev->mode_config.mutex);
 	edid = drm_get_edid(connector, &intel_dp->aux.ddc);
@@ -4607,10 +4952,30 @@ static bool intel_edp_init_connector(struct intel_dp *intel_dp,
 	if (IS_VALLEYVIEW(dev)) {
 		intel_dp->edp_notifier.notifier_call = edp_notify_handler;
 		register_reboot_notifier(&intel_dp->edp_notifier);
+
+		/*
+		 * Figure out the current pipe for the initial backlight setup.
+		 * If the current pipe isn't valid, try the PPS pipe, and if that
+		 * fails just assume pipe A.
+		 */
+		if (IS_CHERRYVIEW(dev))
+			pipe = DP_PORT_TO_PIPE_CHV(intel_dp->DP);
+		else
+			pipe = PORT_TO_PIPE(intel_dp->DP);
+
+		if (pipe != PIPE_A && pipe != PIPE_B)
+			pipe = intel_dp->pps_pipe;
+
+		if (pipe != PIPE_A && pipe != PIPE_B)
+			pipe = PIPE_A;
+
+		DRM_DEBUG_KMS("using pipe %c for initial backlight setup\n",
+			      pipe_name(pipe));
 	}
 
 	intel_panel_init(&intel_connector->panel, fixed_mode, downclock_mode);
-	intel_panel_setup_backlight(connector);
+	intel_connector->panel.backlight_power = intel_edp_backlight_power;
+	intel_panel_setup_backlight(connector, pipe);
 
 	return true;
 }
@@ -4625,11 +4990,14 @@ intel_dp_init_connector(struct intel_digital_port *intel_dig_port,
 	struct drm_device *dev = intel_encoder->base.dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	enum port port = intel_dig_port->port;
-	struct edp_power_seq power_seq = { 0 };
 	int type;
 
+	intel_dp->pps_pipe = INVALID_PIPE;
+
 	/* intel_dp vfuncs */
-	if (IS_VALLEYVIEW(dev))
+	if (INTEL_INFO(dev)->gen >= 9)
+		intel_dp->get_aux_clock_divider = skl_get_aux_clock_divider;
+	else if (IS_VALLEYVIEW(dev))
 		intel_dp->get_aux_clock_divider = vlv_get_aux_clock_divider;
 	else if (IS_HASWELL(dev) || IS_BROADWELL(dev))
 		intel_dp->get_aux_clock_divider = hsw_get_aux_clock_divider;
@@ -4638,7 +5006,10 @@ intel_dp_init_connector(struct intel_digital_port *intel_dig_port,
 	else
 		intel_dp->get_aux_clock_divider = i9xx_get_aux_clock_divider;
 
-	intel_dp->get_aux_send_ctl = i9xx_get_aux_send_ctl;
+	if (INTEL_INFO(dev)->gen >= 9)
+		intel_dp->get_aux_send_ctl = skl_get_aux_send_ctl;
+	else
+		intel_dp->get_aux_send_ctl = i9xx_get_aux_send_ctl;
 
 	/* Preserve the current hw state. */
 	intel_dp->DP = I915_READ(intel_dp->output_reg);
@@ -4657,6 +5028,11 @@ intel_dp_init_connector(struct intel_digital_port *intel_dig_port,
 	if (type == DRM_MODE_CONNECTOR_eDP)
 		intel_encoder->type = INTEL_OUTPUT_EDP;
 
+	/* eDP only on port B and/or C on vlv/chv */
+	if (WARN_ON(IS_VALLEYVIEW(dev) && is_edp(intel_dp) &&
+		    port != PORT_B && port != PORT_C))
+		return false;
+
 	DRM_DEBUG_KMS("Adding %s connector on port %c\n",
 			type == DRM_MODE_CONNECTOR_eDP ? "eDP" : "DP",
 			port_name(port));
@@ -4698,8 +5074,13 @@ intel_dp_init_connector(struct intel_digital_port *intel_dig_port,
 	}
 
 	if (is_edp(intel_dp)) {
+		pps_lock(intel_dp);
 		intel_dp_init_panel_power_timestamps(intel_dp);
-		intel_dp_init_panel_power_sequencer(dev, intel_dp, &power_seq);
+		if (IS_VALLEYVIEW(dev))
+			vlv_initial_power_sequencer_setup(intel_dp);
+		else
+			intel_dp_init_panel_power_sequencer(dev, intel_dp);
+		pps_unlock(intel_dp);
 	}
 
 	intel_dp_aux_init(intel_dp, intel_connector);
@@ -4707,17 +5088,22 @@ intel_dp_init_connector(struct intel_digital_port *intel_dig_port,
 	/* init MST on ports that can support it */
 	if (IS_HASWELL(dev) || IS_BROADWELL(dev)) {
 		if (port == PORT_B || port == PORT_C || port == PORT_D) {
-			intel_dp_mst_encoder_init(intel_dig_port, intel_connector->base.base.id);
+			intel_dp_mst_encoder_init(intel_dig_port,
+						  intel_connector->base.base.id);
 		}
 	}
 
-	if (!intel_edp_init_connector(intel_dp, intel_connector, &power_seq)) {
+	if (!intel_edp_init_connector(intel_dp, intel_connector)) {
 		drm_dp_aux_unregister(&intel_dp->aux);
 		if (is_edp(intel_dp)) {
 			cancel_delayed_work_sync(&intel_dp->panel_vdd_work);
-			drm_modeset_lock(&dev->mode_config.connection_mutex, NULL);
+			/*
+			 * vdd might still be enabled do to the delayed vdd off.
+			 * Make sure vdd is actually turned off here.
+			 */
+			pps_lock(intel_dp);
 			edp_panel_vdd_off_sync(intel_dp);
-			drm_modeset_unlock(&dev->mode_config.connection_mutex);
+			pps_unlock(intel_dp);
 		}
 		drm_connector_unregister(connector);
 		drm_connector_cleanup(connector);
@@ -4781,7 +5167,8 @@ intel_dp_init(struct drm_device *dev, int output_reg, enum port port)
 	} else {
 		intel_encoder->pre_enable = g4x_pre_enable_dp;
 		intel_encoder->enable = g4x_enable_dp;
-		intel_encoder->post_disable = g4x_post_disable_dp;
+		if (INTEL_INFO(dev)->gen >= 5)
+			intel_encoder->post_disable = ilk_post_disable_dp;
 	}
 
 	intel_dig_port->port = port;
diff --git a/drivers/gpu/drm/i915/intel_dp_mst.c b/drivers/gpu/drm/i915/intel_dp_mst.c
index d9a7a7865f66..7f8c6a66680a 100644
--- a/drivers/gpu/drm/i915/intel_dp_mst.c
+++ b/drivers/gpu/drm/i915/intel_dp_mst.c
@@ -278,20 +278,12 @@ static int intel_dp_mst_get_ddc_modes(struct drm_connector *connector)
 }
 
 static enum drm_connector_status
-intel_mst_port_dp_detect(struct drm_connector *connector)
+intel_dp_mst_detect(struct drm_connector *connector, bool force)
 {
 	struct intel_connector *intel_connector = to_intel_connector(connector);
 	struct intel_dp *intel_dp = intel_connector->mst_port;
 
-	return drm_dp_mst_detect_port(&intel_dp->mst_mgr, intel_connector->port);
-}
-
-static enum drm_connector_status
-intel_dp_mst_detect(struct drm_connector *connector, bool force)
-{
-	enum drm_connector_status status;
-	status = intel_mst_port_dp_detect(connector);
-	return status;
+	return drm_dp_mst_detect_port(connector, &intel_dp->mst_mgr, intel_connector->port);
 }
 
 static int
@@ -393,7 +385,7 @@ static void intel_connector_remove_from_fbdev(struct intel_connector *connector)
 #endif
 }
 
-static struct drm_connector *intel_dp_add_mst_connector(struct drm_dp_mst_topology_mgr *mgr, struct drm_dp_mst_port *port, char *pathprop)
+static struct drm_connector *intel_dp_add_mst_connector(struct drm_dp_mst_topology_mgr *mgr, struct drm_dp_mst_port *port, const char *pathprop)
 {
 	struct intel_dp *intel_dp = container_of(mgr, struct intel_dp, mst_mgr);
 	struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp);
@@ -422,6 +414,8 @@ static struct drm_connector *intel_dp_add_mst_connector(struct drm_dp_mst_topolo
 	intel_dp_add_properties(intel_dp, connector);
 
 	drm_object_attach_property(&connector->base, dev->mode_config.path_property, 0);
+	drm_object_attach_property(&connector->base, dev->mode_config.tile_property, 0);
+
 	drm_mode_connector_set_path_property(connector, pathprop);
 	drm_reinit_primary_mode_group(dev);
 	mutex_lock(&dev->mode_config.mutex);
diff --git a/drivers/gpu/drm/i915/intel_drv.h b/drivers/gpu/drm/i915/intel_drv.h
index b8c8bbd8e5f9..25fdbb16d4e0 100644
--- a/drivers/gpu/drm/i915/intel_drv.h
+++ b/drivers/gpu/drm/i915/intel_drv.h
@@ -25,6 +25,7 @@
 #ifndef __INTEL_DRV_H__
 #define __INTEL_DRV_H__
 
+#include <linux/async.h>
 #include <linux/i2c.h>
 #include <linux/hdmi.h>
 #include <drm/i915_drm.h>
@@ -33,6 +34,10 @@
 #include <drm/drm_crtc_helper.h>
 #include <drm/drm_fb_helper.h>
 #include <drm/drm_dp_mst_helper.h>
+#include <drm/drm_rect.h>
+
+#define DIV_ROUND_CLOSEST_ULL(ll, d)	\
+({ unsigned long long _tmp = (ll)+(d)/2; do_div(_tmp, d); _tmp; })
 
 /**
  * _wait_for - magic (register) wait macro
@@ -89,18 +94,20 @@
 
 /* these are outputs from the chip - integrated only
    external chips are via DVO or SDVO output */
-#define INTEL_OUTPUT_UNUSED 0
-#define INTEL_OUTPUT_ANALOG 1
-#define INTEL_OUTPUT_DVO 2
-#define INTEL_OUTPUT_SDVO 3
-#define INTEL_OUTPUT_LVDS 4
-#define INTEL_OUTPUT_TVOUT 5
-#define INTEL_OUTPUT_HDMI 6
-#define INTEL_OUTPUT_DISPLAYPORT 7
-#define INTEL_OUTPUT_EDP 8
-#define INTEL_OUTPUT_DSI 9
-#define INTEL_OUTPUT_UNKNOWN 10
-#define INTEL_OUTPUT_DP_MST 11
+enum intel_output_type {
+	INTEL_OUTPUT_UNUSED = 0,
+	INTEL_OUTPUT_ANALOG = 1,
+	INTEL_OUTPUT_DVO = 2,
+	INTEL_OUTPUT_SDVO = 3,
+	INTEL_OUTPUT_LVDS = 4,
+	INTEL_OUTPUT_TVOUT = 5,
+	INTEL_OUTPUT_HDMI = 6,
+	INTEL_OUTPUT_DISPLAYPORT = 7,
+	INTEL_OUTPUT_EDP = 8,
+	INTEL_OUTPUT_DSI = 9,
+	INTEL_OUTPUT_UNKNOWN = 10,
+	INTEL_OUTPUT_DP_MST = 11,
+};
 
 #define INTEL_DVO_CHIP_NONE 0
 #define INTEL_DVO_CHIP_LVDS 1
@@ -131,7 +138,7 @@ struct intel_encoder {
 	 */
 	struct intel_crtc *new_crtc;
 
-	int type;
+	enum intel_output_type type;
 	unsigned int cloneable;
 	bool connectors_active;
 	void (*hot_plug)(struct intel_encoder *);
@@ -179,6 +186,8 @@ struct intel_panel {
 		bool active_low_pwm;
 		struct backlight_device *device;
 	} backlight;
+
+	void (*backlight_power)(struct intel_connector *, bool enable);
 };
 
 struct intel_connector {
@@ -211,6 +220,7 @@ struct intel_connector {
 
 	/* Cached EDID for eDP and LVDS. May hold ERR_PTR for invalid EDID. */
 	struct edid *edid;
+	struct edid *detect_edid;
 
 	/* since POLL and HPD connectors may use the same HPD line keep the native
 	   state of connector->polled in case hotplug storm detection changes it */
@@ -233,6 +243,17 @@ typedef struct dpll {
 	int	p;
 } intel_clock_t;
 
+struct intel_plane_state {
+	struct drm_crtc *crtc;
+	struct drm_framebuffer *fb;
+	struct drm_rect src;
+	struct drm_rect dst;
+	struct drm_rect clip;
+	struct drm_rect orig_src;
+	struct drm_rect orig_dst;
+	bool visible;
+};
+
 struct intel_plane_config {
 	bool tiled;
 	int size;
@@ -271,6 +292,9 @@ struct intel_crtc_config {
 	 * between pch encoders and cpu encoders. */
 	bool has_pch_encoder;
 
+	/* Are we sending infoframes on the attached port */
+	bool has_infoframe;
+
 	/* CPU Transcoder for the pipe. Currently this can only differ from the
 	 * pipe on Haswell (where we have a special eDP transcoder). */
 	enum transcoder cpu_transcoder;
@@ -319,7 +343,10 @@ struct intel_crtc_config {
 	/* Selected dpll when shared or DPLL_ID_PRIVATE. */
 	enum intel_dpll_id shared_dpll;
 
-	/* PORT_CLK_SEL for DDI ports. */
+	/*
+	 * - PORT_CLK_SEL for DDI ports on HSW/BDW.
+	 * - enum skl_dpll on SKL
+	 */
 	uint32_t ddi_pll_sel;
 
 	/* Actual register state of the dpll, for shared dpll cross-checking. */
@@ -330,6 +357,7 @@ struct intel_crtc_config {
 
 	/* m2_n2 for eDP downclock */
 	struct intel_link_m_n dp_m2_n2;
+	bool has_drrs;
 
 	/*
 	 * Frequence the dpll for the port should run at. Differs from the
@@ -379,7 +407,14 @@ struct intel_pipe_wm {
 
 struct intel_mmio_flip {
 	u32 seqno;
-	u32 ring_id;
+	struct intel_engine_cs *ring;
+	struct work_struct work;
+};
+
+struct skl_pipe_wm {
+	struct skl_wm_level wm[8];
+	struct skl_wm_level trans_wm;
+	uint32_t linetime;
 };
 
 struct intel_crtc {
@@ -410,6 +445,7 @@ struct intel_crtc {
 	uint32_t cursor_addr;
 	int16_t cursor_width, cursor_height;
 	uint32_t cursor_cntl;
+	uint32_t cursor_size;
 	uint32_t cursor_base;
 
 	struct intel_plane_config plane_config;
@@ -428,10 +464,10 @@ struct intel_crtc {
 	struct {
 		/* watermarks currently being used  */
 		struct intel_pipe_wm active;
+		/* SKL wm values currently in use */
+		struct skl_pipe_wm skl_active;
 	} wm;
 
-	wait_queue_head_t vbl_wait;
-
 	int scanline_offset;
 	struct intel_mmio_flip mmio_flip;
 };
@@ -455,6 +491,7 @@ struct intel_plane {
 	unsigned int crtc_w, crtc_h;
 	uint32_t src_x, src_y;
 	uint32_t src_w, src_h;
+	unsigned int rotation;
 
 	/* Since we need to change the watermarks before/after
 	 * enabling/disabling the planes, we need to store the parameters here
@@ -521,6 +558,7 @@ struct intel_hdmi {
 	void (*set_infoframes)(struct drm_encoder *encoder,
 			       bool enable,
 			       struct drm_display_mode *adjusted_mode);
+	bool (*infoframe_enabled)(struct drm_encoder *encoder);
 };
 
 struct intel_dp_mst_encoder;
@@ -565,6 +603,13 @@ struct intel_dp {
 
 	struct notifier_block edp_notifier;
 
+	/*
+	 * Pipe whose power sequencer is currently locked into
+	 * this port. Only relevant on VLV/CHV.
+	 */
+	enum pipe pps_pipe;
+	struct edp_power_seq pps_delays;
+
 	bool use_tps3;
 	bool can_mst; /* this port supports mst */
 	bool is_mst;
@@ -663,6 +708,10 @@ struct intel_unpin_work {
 #define INTEL_FLIP_COMPLETE	2
 	u32 flip_count;
 	u32 gtt_offset;
+	struct intel_engine_cs *flip_queued_ring;
+	u32 flip_queued_seqno;
+	int flip_queued_vblank;
+	int flip_ready_vblank;
 	bool enable_stall_check;
 };
 
@@ -716,32 +765,47 @@ hdmi_to_dig_port(struct intel_hdmi *intel_hdmi)
 	return container_of(intel_hdmi, struct intel_digital_port, hdmi);
 }
 
+/*
+ * Returns the number of planes for this pipe, ie the number of sprites + 1
+ * (primary plane). This doesn't count the cursor plane then.
+ */
+static inline unsigned int intel_num_planes(struct intel_crtc *crtc)
+{
+	return INTEL_INFO(crtc->base.dev)->num_sprites[crtc->pipe] + 1;
+}
 
-/* i915_irq.c */
-bool intel_set_cpu_fifo_underrun_reporting(struct drm_device *dev,
+/* intel_fifo_underrun.c */
+bool intel_set_cpu_fifo_underrun_reporting(struct drm_i915_private *dev_priv,
 					   enum pipe pipe, bool enable);
-bool intel_set_pch_fifo_underrun_reporting(struct drm_device *dev,
+bool intel_set_pch_fifo_underrun_reporting(struct drm_i915_private *dev_priv,
 					   enum transcoder pch_transcoder,
 					   bool enable);
+void intel_cpu_fifo_underrun_irq_handler(struct drm_i915_private *dev_priv,
+					 enum pipe pipe);
+void intel_pch_fifo_underrun_irq_handler(struct drm_i915_private *dev_priv,
+					 enum transcoder pch_transcoder);
+void i9xx_check_fifo_underruns(struct drm_i915_private *dev_priv);
+
+/* i915_irq.c */
 void gen5_enable_gt_irq(struct drm_i915_private *dev_priv, uint32_t mask);
 void gen5_disable_gt_irq(struct drm_i915_private *dev_priv, uint32_t mask);
 void gen6_enable_pm_irq(struct drm_i915_private *dev_priv, uint32_t mask);
 void gen6_disable_pm_irq(struct drm_i915_private *dev_priv, uint32_t mask);
-void gen8_enable_pm_irq(struct drm_i915_private *dev_priv, uint32_t mask);
-void gen8_disable_pm_irq(struct drm_i915_private *dev_priv, uint32_t mask);
-void intel_runtime_pm_disable_interrupts(struct drm_device *dev);
-void intel_runtime_pm_restore_interrupts(struct drm_device *dev);
+void gen6_reset_rps_interrupts(struct drm_device *dev);
+void gen6_enable_rps_interrupts(struct drm_device *dev);
+void gen6_disable_rps_interrupts(struct drm_device *dev);
+void intel_runtime_pm_disable_interrupts(struct drm_i915_private *dev_priv);
+void intel_runtime_pm_enable_interrupts(struct drm_i915_private *dev_priv);
 static inline bool intel_irqs_enabled(struct drm_i915_private *dev_priv)
 {
 	/*
 	 * We only use drm_irq_uninstall() at unload and VT switch, so
 	 * this is the only thing we need to check.
 	 */
-	return !dev_priv->pm._irqs_disabled;
+	return dev_priv->pm.irqs_enabled;
 }
 
 int intel_get_crtc_scanline(struct intel_crtc *crtc);
-void i9xx_check_fifo_underruns(struct drm_device *dev);
 void gen8_irq_power_well_post_enable(struct drm_i915_private *dev_priv);
 
 /* intel_crt.c */
@@ -774,11 +838,7 @@ void intel_ddi_clock_get(struct intel_encoder *encoder,
 			 struct intel_crtc_config *pipe_config);
 void intel_ddi_set_vc_payload_alloc(struct drm_crtc *crtc, bool state);
 
-/* intel_display.c */
-const char *intel_output_name(int output);
-bool intel_has_pending_fb_unpin(struct drm_device *dev);
-int intel_pch_rawclk(struct drm_device *dev);
-void intel_mark_busy(struct drm_device *dev);
+/* intel_frontbuffer.c */
 void intel_fb_obj_invalidate(struct drm_i915_gem_object *obj,
 			     struct intel_engine_cs *ring);
 void intel_frontbuffer_flip_prepare(struct drm_device *dev,
@@ -788,7 +848,7 @@ void intel_frontbuffer_flip_complete(struct drm_device *dev,
 void intel_frontbuffer_flush(struct drm_device *dev,
 			     unsigned frontbuffer_bits);
 /**
- * intel_frontbuffer_flip - prepare frontbuffer flip
+ * intel_frontbuffer_flip - synchronous frontbuffer flip
  * @dev: DRM device
  * @frontbuffer_bits: frontbuffer plane tracking bits
  *
@@ -806,6 +866,18 @@ void intel_frontbuffer_flip(struct drm_device *dev,
 }
 
 void intel_fb_obj_flush(struct drm_i915_gem_object *obj, bool retire);
+
+
+/* intel_audio.c */
+void intel_init_audio(struct drm_device *dev);
+void intel_audio_codec_enable(struct intel_encoder *encoder);
+void intel_audio_codec_disable(struct intel_encoder *encoder);
+
+/* intel_display.c */
+const char *intel_output_name(int output);
+bool intel_has_pending_fb_unpin(struct drm_device *dev);
+int intel_pch_rawclk(struct drm_device *dev);
+void intel_mark_busy(struct drm_device *dev);
 void intel_mark_idle(struct drm_device *dev);
 void intel_crtc_restore_mode(struct drm_crtc *crtc);
 void intel_crtc_control(struct drm_crtc *crtc, bool enable);
@@ -826,8 +898,12 @@ int intel_get_pipe_from_crtc_id(struct drm_device *dev, void *data,
 				struct drm_file *file_priv);
 enum transcoder intel_pipe_to_cpu_transcoder(struct drm_i915_private *dev_priv,
 					     enum pipe pipe);
-void intel_wait_for_vblank(struct drm_device *dev, int pipe);
-void intel_wait_for_pipe_off(struct drm_device *dev, int pipe);
+bool intel_pipe_has_type(struct intel_crtc *crtc, enum intel_output_type type);
+static inline void
+intel_wait_for_vblank(struct drm_device *dev, int pipe)
+{
+	drm_wait_one_vblank(dev, pipe);
+}
 int ironlake_get_lanes_required(int target_clock, int link_bw, int bpp);
 void vlv_wait_port_ready(struct drm_i915_private *dev_priv,
 			 struct intel_digital_port *dport);
@@ -837,8 +913,8 @@ bool intel_get_load_detect_pipe(struct drm_connector *connector,
 				struct drm_modeset_acquire_ctx *ctx);
 void intel_release_load_detect_pipe(struct drm_connector *connector,
 				    struct intel_load_detect_pipe *old);
-int intel_pin_and_fence_fb_obj(struct drm_device *dev,
-			       struct drm_i915_gem_object *obj,
+int intel_pin_and_fence_fb_obj(struct drm_plane *plane,
+			       struct drm_framebuffer *fb,
 			       struct intel_engine_cs *pipelined);
 void intel_unpin_fb_obj(struct drm_i915_gem_object *obj);
 struct drm_framebuffer *
@@ -848,6 +924,7 @@ __intel_framebuffer_create(struct drm_device *dev,
 void intel_prepare_page_flip(struct drm_device *dev, int plane);
 void intel_finish_page_flip(struct drm_device *dev, int pipe);
 void intel_finish_page_flip_plane(struct drm_device *dev, int plane);
+void intel_check_page_flip(struct drm_device *dev, int pipe);
 
 /* shared dpll functions */
 struct intel_shared_dpll *intel_crtc_to_shared_dpll(struct intel_crtc *crtc);
@@ -859,7 +936,13 @@ void assert_shared_dpll(struct drm_i915_private *dev_priv,
 struct intel_shared_dpll *intel_get_shared_dpll(struct intel_crtc *crtc);
 void intel_put_shared_dpll(struct intel_crtc *crtc);
 
+void vlv_force_pll_on(struct drm_device *dev, enum pipe pipe,
+		      const struct dpll *dpll);
+void vlv_force_pll_off(struct drm_device *dev, enum pipe pipe);
+
 /* modesetting asserts */
+void assert_panel_unlocked(struct drm_i915_private *dev_priv,
+			   enum pipe pipe);
 void assert_pll(struct drm_i915_private *dev_priv,
 		enum pipe pipe, bool state);
 #define assert_pll_enabled(d, p) assert_pll(d, p, true)
@@ -871,17 +954,17 @@ void assert_fdi_rx_pll(struct drm_i915_private *dev_priv,
 void assert_pipe(struct drm_i915_private *dev_priv, enum pipe pipe, bool state);
 #define assert_pipe_enabled(d, p) assert_pipe(d, p, true)
 #define assert_pipe_disabled(d, p) assert_pipe(d, p, false)
-void intel_write_eld(struct drm_encoder *encoder,
-		     struct drm_display_mode *mode);
 unsigned long intel_gen4_compute_page_offset(int *x, int *y,
 					     unsigned int tiling_mode,
 					     unsigned int bpp,
 					     unsigned int pitch);
-void intel_display_handle_reset(struct drm_device *dev);
+void intel_prepare_reset(struct drm_device *dev);
+void intel_finish_reset(struct drm_device *dev);
 void hsw_enable_pc8(struct drm_i915_private *dev_priv);
 void hsw_disable_pc8(struct drm_i915_private *dev_priv);
 void intel_dp_get_m_n(struct intel_crtc *crtc,
 		      struct intel_crtc_config *pipe_config);
+void intel_dp_set_m_n(struct intel_crtc *crtc);
 int intel_dotclock_calculate(int link_freq, const struct intel_link_m_n *m_n);
 void
 ironlake_check_encoder_dotclock(const struct intel_crtc_config *pipe_config,
@@ -889,14 +972,13 @@ ironlake_check_encoder_dotclock(const struct intel_crtc_config *pipe_config,
 bool intel_crtc_active(struct drm_crtc *crtc);
 void hsw_enable_ips(struct intel_crtc *crtc);
 void hsw_disable_ips(struct intel_crtc *crtc);
-void intel_display_set_init_power(struct drm_i915_private *dev, bool enable);
 enum intel_display_power_domain
 intel_display_port_power_domain(struct intel_encoder *intel_encoder);
 void intel_mode_from_pipe_config(struct drm_display_mode *mode,
 				 struct intel_crtc_config *pipe_config);
 int intel_format_to_fourcc(int format);
 void intel_crtc_wait_for_pending_flips(struct drm_crtc *crtc);
-
+void intel_modeset_preclose(struct drm_device *dev, struct drm_file *file);
 
 /* intel_dp.c */
 void intel_dp_init(struct drm_device *dev, int output_reg, enum port port);
@@ -917,24 +999,18 @@ bool intel_dp_hpd_pulse(struct intel_digital_port *intel_dig_port,
 void intel_edp_backlight_on(struct intel_dp *intel_dp);
 void intel_edp_backlight_off(struct intel_dp *intel_dp);
 void intel_edp_panel_vdd_on(struct intel_dp *intel_dp);
-void intel_edp_panel_vdd_sanitize(struct intel_encoder *intel_encoder);
 void intel_edp_panel_on(struct intel_dp *intel_dp);
 void intel_edp_panel_off(struct intel_dp *intel_dp);
-void intel_edp_psr_enable(struct intel_dp *intel_dp);
-void intel_edp_psr_disable(struct intel_dp *intel_dp);
 void intel_dp_set_drrs_state(struct drm_device *dev, int refresh_rate);
-void intel_edp_psr_invalidate(struct drm_device *dev,
-			      unsigned frontbuffer_bits);
-void intel_edp_psr_flush(struct drm_device *dev,
-			 unsigned frontbuffer_bits);
-void intel_edp_psr_init(struct drm_device *dev);
-
-int intel_dp_handle_hpd_irq(struct intel_digital_port *digport, bool long_hpd);
 void intel_dp_add_properties(struct intel_dp *intel_dp, struct drm_connector *connector);
 void intel_dp_mst_suspend(struct drm_device *dev);
 void intel_dp_mst_resume(struct drm_device *dev);
 int intel_dp_max_link_bw(struct intel_dp *intel_dp);
 void intel_dp_hot_plug(struct intel_encoder *intel_encoder);
+void vlv_power_sequencer_reset(struct drm_i915_private *dev_priv);
+uint32_t intel_dp_pack_aux(const uint8_t *src, int src_bytes);
+void intel_dp_unpack_aux(uint32_t src, uint8_t *dst, int dst_bytes);
+
 /* intel_dp_mst.c */
 int intel_dp_mst_encoder_init(struct intel_digital_port *intel_dig_port, int conn_id);
 void intel_dp_mst_encoder_cleanup(struct intel_digital_port *intel_dig_port);
@@ -949,9 +1025,9 @@ void intel_dvo_init(struct drm_device *dev);
 /* legacy fbdev emulation in intel_fbdev.c */
 #ifdef CONFIG_DRM_I915_FBDEV
 extern int intel_fbdev_init(struct drm_device *dev);
-extern void intel_fbdev_initial_config(struct drm_device *dev);
+extern void intel_fbdev_initial_config(void *data, async_cookie_t cookie);
 extern void intel_fbdev_fini(struct drm_device *dev);
-extern void intel_fbdev_set_suspend(struct drm_device *dev, int state);
+extern void intel_fbdev_set_suspend(struct drm_device *dev, int state, bool synchronous);
 extern void intel_fbdev_output_poll_changed(struct drm_device *dev);
 extern void intel_fbdev_restore_mode(struct drm_device *dev);
 #else
@@ -960,7 +1036,7 @@ static inline int intel_fbdev_init(struct drm_device *dev)
 	return 0;
 }
 
-static inline void intel_fbdev_initial_config(struct drm_device *dev)
+static inline void intel_fbdev_initial_config(void *data, async_cookie_t cookie)
 {
 }
 
@@ -968,7 +1044,7 @@ static inline void intel_fbdev_fini(struct drm_device *dev)
 {
 }
 
-static inline void intel_fbdev_set_suspend(struct drm_device *dev, int state)
+static inline void intel_fbdev_set_suspend(struct drm_device *dev, int state, bool synchronous)
 {
 }
 
@@ -1024,7 +1100,7 @@ void intel_gmch_panel_fitting(struct intel_crtc *crtc,
 			      int fitting_mode);
 void intel_panel_set_backlight_acpi(struct intel_connector *connector,
 				    u32 level, u32 max);
-int intel_panel_setup_backlight(struct drm_connector *connector);
+int intel_panel_setup_backlight(struct drm_connector *connector, enum pipe pipe);
 void intel_panel_enable_backlight(struct intel_connector *connector);
 void intel_panel_disable_backlight(struct intel_connector *connector);
 void intel_panel_destroy_backlight(struct drm_connector *connector);
@@ -1034,6 +1110,41 @@ extern struct drm_display_mode *intel_find_panel_downclock(
 				struct drm_device *dev,
 				struct drm_display_mode *fixed_mode,
 				struct drm_connector *connector);
+void intel_backlight_register(struct drm_device *dev);
+void intel_backlight_unregister(struct drm_device *dev);
+
+
+/* intel_psr.c */
+bool intel_psr_is_enabled(struct drm_device *dev);
+void intel_psr_enable(struct intel_dp *intel_dp);
+void intel_psr_disable(struct intel_dp *intel_dp);
+void intel_psr_invalidate(struct drm_device *dev,
+			      unsigned frontbuffer_bits);
+void intel_psr_flush(struct drm_device *dev,
+			 unsigned frontbuffer_bits);
+void intel_psr_init(struct drm_device *dev);
+
+/* intel_runtime_pm.c */
+int intel_power_domains_init(struct drm_i915_private *);
+void intel_power_domains_fini(struct drm_i915_private *);
+void intel_power_domains_init_hw(struct drm_i915_private *dev_priv);
+void intel_runtime_pm_enable(struct drm_i915_private *dev_priv);
+
+bool intel_display_power_is_enabled(struct drm_i915_private *dev_priv,
+				    enum intel_display_power_domain domain);
+bool __intel_display_power_is_enabled(struct drm_i915_private *dev_priv,
+				      enum intel_display_power_domain domain);
+void intel_display_power_get(struct drm_i915_private *dev_priv,
+			     enum intel_display_power_domain domain);
+void intel_display_power_put(struct drm_i915_private *dev_priv,
+			     enum intel_display_power_domain domain);
+void intel_aux_display_runtime_get(struct drm_i915_private *dev_priv);
+void intel_aux_display_runtime_put(struct drm_i915_private *dev_priv);
+void intel_runtime_pm_get(struct drm_i915_private *dev_priv);
+void intel_runtime_pm_get_noresume(struct drm_i915_private *dev_priv);
+void intel_runtime_pm_put(struct drm_i915_private *dev_priv);
+
+void intel_display_set_init_power(struct drm_i915_private *dev, bool enable);
 
 /* intel_pm.c */
 void intel_init_clock_gating(struct drm_device *dev);
@@ -1052,17 +1163,6 @@ bool intel_fbc_enabled(struct drm_device *dev);
 void intel_update_fbc(struct drm_device *dev);
 void intel_gpu_ips_init(struct drm_i915_private *dev_priv);
 void intel_gpu_ips_teardown(void);
-int intel_power_domains_init(struct drm_i915_private *);
-void intel_power_domains_remove(struct drm_i915_private *);
-bool intel_display_power_enabled(struct drm_i915_private *dev_priv,
-				 enum intel_display_power_domain domain);
-bool intel_display_power_enabled_unlocked(struct drm_i915_private *dev_priv,
-					  enum intel_display_power_domain domain);
-void intel_display_power_get(struct drm_i915_private *dev_priv,
-			     enum intel_display_power_domain domain);
-void intel_display_power_put(struct drm_i915_private *dev_priv,
-			     enum intel_display_power_domain domain);
-void intel_power_domains_init_hw(struct drm_i915_private *dev_priv);
 void intel_init_gt_powersave(struct drm_device *dev);
 void intel_cleanup_gt_powersave(struct drm_device *dev);
 void intel_enable_gt_powersave(struct drm_device *dev);
@@ -1073,14 +1173,10 @@ void ironlake_teardown_rc6(struct drm_device *dev);
 void gen6_update_ring_freq(struct drm_device *dev);
 void gen6_rps_idle(struct drm_i915_private *dev_priv);
 void gen6_rps_boost(struct drm_i915_private *dev_priv);
-void intel_aux_display_runtime_get(struct drm_i915_private *dev_priv);
-void intel_aux_display_runtime_put(struct drm_i915_private *dev_priv);
-void intel_runtime_pm_get(struct drm_i915_private *dev_priv);
-void intel_runtime_pm_get_noresume(struct drm_i915_private *dev_priv);
-void intel_runtime_pm_put(struct drm_i915_private *dev_priv);
-void intel_init_runtime_pm(struct drm_i915_private *dev_priv);
-void intel_fini_runtime_pm(struct drm_i915_private *dev_priv);
 void ilk_wm_get_hw_state(struct drm_device *dev);
+void skl_wm_get_hw_state(struct drm_device *dev);
+void skl_ddb_get_hw_state(struct drm_i915_private *dev_priv,
+			  struct skl_ddb_allocation *ddb /* out */);
 
 
 /* intel_sdvo.c */
@@ -1091,13 +1187,18 @@ bool intel_sdvo_init(struct drm_device *dev, uint32_t sdvo_reg, bool is_sdvob);
 int intel_plane_init(struct drm_device *dev, enum pipe pipe, int plane);
 void intel_flush_primary_plane(struct drm_i915_private *dev_priv,
 			       enum plane plane);
-void intel_plane_restore(struct drm_plane *plane);
+int intel_plane_set_property(struct drm_plane *plane,
+			     struct drm_property *prop,
+			     uint64_t val);
+int intel_plane_restore(struct drm_plane *plane);
 void intel_plane_disable(struct drm_plane *plane);
 int intel_sprite_set_colorkey(struct drm_device *dev, void *data,
 			      struct drm_file *file_priv);
 int intel_sprite_get_colorkey(struct drm_device *dev, void *data,
 			      struct drm_file *file_priv);
-
+bool intel_pipe_update_start(struct intel_crtc *crtc,
+			     uint32_t *start_vbl_count);
+void intel_pipe_update_end(struct intel_crtc *crtc, u32 start_vbl_count);
 
 /* intel_tv.c */
 void intel_tv_init(struct drm_device *dev);
diff --git a/drivers/gpu/drm/i915/intel_dsi.c b/drivers/gpu/drm/i915/intel_dsi.c
index 670c29a7b5dd..0b184079de14 100644
--- a/drivers/gpu/drm/i915/intel_dsi.c
+++ b/drivers/gpu/drm/i915/intel_dsi.c
@@ -184,7 +184,7 @@ static void intel_dsi_pre_enable(struct intel_encoder *encoder)
 
 	/* update the hw state for DPLL */
 	intel_crtc->config.dpll_hw_state.dpll = DPLL_INTEGRATED_CLOCK_VLV |
-						DPLL_REFA_CLK_ENABLE_VLV;
+		DPLL_REFA_CLK_ENABLE_VLV;
 
 	tmp = I915_READ(DSPCLK_GATE_D);
 	tmp |= DPOUNIT_CLOCK_GATE_DISABLE;
@@ -259,8 +259,8 @@ static void intel_dsi_disable(struct intel_encoder *encoder)
 	temp = I915_READ(MIPI_CTRL(pipe));
 	temp &= ~ESCAPE_CLOCK_DIVIDER_MASK;
 	I915_WRITE(MIPI_CTRL(pipe), temp |
-			intel_dsi->escape_clk_div <<
-			ESCAPE_CLOCK_DIVIDER_SHIFT);
+		   intel_dsi->escape_clk_div <<
+		   ESCAPE_CLOCK_DIVIDER_SHIFT);
 
 	I915_WRITE(MIPI_EOT_DISABLE(pipe), CLOCKSTOP);
 
@@ -297,7 +297,7 @@ static void intel_dsi_clear_device_ready(struct intel_encoder *encoder)
 	usleep_range(2000, 2500);
 
 	if (wait_for(((I915_READ(MIPI_PORT_CTRL(pipe)) & AFE_LATCHOUT)
-					== 0x00000), 30))
+		      == 0x00000), 30))
 		DRM_ERROR("DSI LP not going Low\n");
 
 	val = I915_READ(MIPI_PORT_CTRL(pipe));
@@ -344,7 +344,7 @@ static bool intel_dsi_get_hw_state(struct intel_encoder *encoder,
 	DRM_DEBUG_KMS("\n");
 
 	power_domain = intel_display_port_power_domain(encoder);
-	if (!intel_display_power_enabled(dev_priv, power_domain))
+	if (!intel_display_power_is_enabled(dev_priv, power_domain))
 		return false;
 
 	/* XXX: this only works for one DSI output */
@@ -423,9 +423,11 @@ static u16 txclkesc(u32 divider, unsigned int us)
 }
 
 /* return pixels in terms of txbyteclkhs */
-static u16 txbyteclkhs(u16 pixels, int bpp, int lane_count)
+static u16 txbyteclkhs(u16 pixels, int bpp, int lane_count,
+		       u16 burst_mode_ratio)
 {
-	return DIV_ROUND_UP(DIV_ROUND_UP(pixels * bpp, 8), lane_count);
+	return DIV_ROUND_UP(DIV_ROUND_UP(pixels * bpp * burst_mode_ratio,
+					 8 * 100), lane_count);
 }
 
 static void set_dsi_timings(struct drm_encoder *encoder,
@@ -451,10 +453,12 @@ static void set_dsi_timings(struct drm_encoder *encoder,
 	vbp = mode->vtotal - mode->vsync_end;
 
 	/* horizontal values are in terms of high speed byte clock */
-	hactive = txbyteclkhs(hactive, bpp, lane_count);
-	hfp = txbyteclkhs(hfp, bpp, lane_count);
-	hsync = txbyteclkhs(hsync, bpp, lane_count);
-	hbp = txbyteclkhs(hbp, bpp, lane_count);
+	hactive = txbyteclkhs(hactive, bpp, lane_count,
+			      intel_dsi->burst_mode_ratio);
+	hfp = txbyteclkhs(hfp, bpp, lane_count, intel_dsi->burst_mode_ratio);
+	hsync = txbyteclkhs(hsync, bpp, lane_count,
+			    intel_dsi->burst_mode_ratio);
+	hbp = txbyteclkhs(hbp, bpp, lane_count, intel_dsi->burst_mode_ratio);
 
 	I915_WRITE(MIPI_HACTIVE_AREA_COUNT(pipe), hactive);
 	I915_WRITE(MIPI_HFP_COUNT(pipe), hfp);
@@ -541,12 +545,14 @@ static void intel_dsi_prepare(struct intel_encoder *intel_encoder)
 	    intel_dsi->video_mode_format == VIDEO_MODE_BURST) {
 		I915_WRITE(MIPI_HS_TX_TIMEOUT(pipe),
 			   txbyteclkhs(adjusted_mode->htotal, bpp,
-				       intel_dsi->lane_count) + 1);
+				       intel_dsi->lane_count,
+				       intel_dsi->burst_mode_ratio) + 1);
 	} else {
 		I915_WRITE(MIPI_HS_TX_TIMEOUT(pipe),
 			   txbyteclkhs(adjusted_mode->vtotal *
 				       adjusted_mode->htotal,
-				       bpp, intel_dsi->lane_count) + 1);
+				       bpp, intel_dsi->lane_count,
+				       intel_dsi->burst_mode_ratio) + 1);
 	}
 	I915_WRITE(MIPI_LP_RX_TIMEOUT(pipe), intel_dsi->lp_rx_timeout);
 	I915_WRITE(MIPI_TURN_AROUND_TIMEOUT(pipe), intel_dsi->turn_arnd_val);
@@ -576,7 +582,7 @@ static void intel_dsi_prepare(struct intel_encoder *intel_encoder)
 	 * XXX: write MIPI_STOP_STATE_STALL?
 	 */
 	I915_WRITE(MIPI_HIGH_LOW_SWITCH_COUNT(pipe),
-						intel_dsi->hs_to_lp_count);
+		   intel_dsi->hs_to_lp_count);
 
 	/* XXX: low power clock equivalence in terms of byte clock. the number
 	 * of byte clocks occupied in one low power clock. based on txbyteclkhs
@@ -601,10 +607,10 @@ static void intel_dsi_prepare(struct intel_encoder *intel_encoder)
 		 * 64 like 1366 x 768. Enable RANDOM resolution support for such
 		 * panels by default */
 		I915_WRITE(MIPI_VIDEO_MODE_FORMAT(pipe),
-				intel_dsi->video_frmt_cfg_bits |
-				intel_dsi->video_mode_format |
-				IP_TG_CONFIG |
-				RANDOM_DPI_DISPLAY_RESOLUTION);
+			   intel_dsi->video_frmt_cfg_bits |
+			   intel_dsi->video_mode_format |
+			   IP_TG_CONFIG |
+			   RANDOM_DPI_DISPLAY_RESOLUTION);
 }
 
 static void intel_dsi_pre_pll_enable(struct intel_encoder *encoder)
diff --git a/drivers/gpu/drm/i915/intel_dsi.h b/drivers/gpu/drm/i915/intel_dsi.h
index fd51867fd0d3..657eb5c1b9d8 100644
--- a/drivers/gpu/drm/i915/intel_dsi.h
+++ b/drivers/gpu/drm/i915/intel_dsi.h
@@ -116,6 +116,8 @@ struct intel_dsi {
 	u16 clk_hs_to_lp_count;
 
 	u16 init_count;
+	u32 pclk;
+	u16 burst_mode_ratio;
 
 	/* all delays in ms */
 	u16 backlight_off_delay;
diff --git a/drivers/gpu/drm/i915/intel_dsi_cmd.c b/drivers/gpu/drm/i915/intel_dsi_cmd.c
index 7f1430ac8543..f4767fd2ebeb 100644
--- a/drivers/gpu/drm/i915/intel_dsi_cmd.c
+++ b/drivers/gpu/drm/i915/intel_dsi_cmd.c
@@ -430,7 +430,7 @@ void wait_for_dsi_fifo_empty(struct intel_dsi *intel_dsi)
 	u32 mask;
 
 	mask = LP_CTRL_FIFO_EMPTY | HS_CTRL_FIFO_EMPTY |
-					LP_DATA_FIFO_EMPTY | HS_DATA_FIFO_EMPTY;
+		LP_DATA_FIFO_EMPTY | HS_DATA_FIFO_EMPTY;
 
 	if (wait_for((I915_READ(MIPI_GEN_FIFO_STAT(pipe)) & mask) == mask, 100))
 		DRM_ERROR("DPI FIFOs are not empty\n");
diff --git a/drivers/gpu/drm/i915/intel_dsi_panel_vbt.c b/drivers/gpu/drm/i915/intel_dsi_panel_vbt.c
index 47c7584a4aa0..f6bdd44069ce 100644
--- a/drivers/gpu/drm/i915/intel_dsi_panel_vbt.c
+++ b/drivers/gpu/drm/i915/intel_dsi_panel_vbt.c
@@ -271,6 +271,8 @@ static bool generic_init(struct intel_dsi_device *dsi)
 	u32 ths_prepare_ns, tclk_trail_ns;
 	u32 tclk_prepare_clkzero, ths_prepare_hszero;
 	u32 lp_to_hs_switch, hs_to_lp_switch;
+	u32 pclk, computed_ddr;
+	u16 burst_mode_ratio;
 
 	DRM_DEBUG_KMS("\n");
 
@@ -284,8 +286,6 @@ static bool generic_init(struct intel_dsi_device *dsi)
 	else if (intel_dsi->pixel_format == VID_MODE_FORMAT_RGB565)
 		bits_per_pixel = 16;
 
-	bitrate = (mode->clock * bits_per_pixel) / intel_dsi->lane_count;
-
 	intel_dsi->operation_mode = mipi_config->is_cmd_mode;
 	intel_dsi->video_mode_format = mipi_config->video_transfer_mode;
 	intel_dsi->escape_clk_div = mipi_config->byte_clk_sel;
@@ -297,6 +297,40 @@ static bool generic_init(struct intel_dsi_device *dsi)
 	intel_dsi->video_frmt_cfg_bits =
 		mipi_config->bta_enabled ? DISABLE_VIDEO_BTA : 0;
 
+	pclk = mode->clock;
+
+	/* Burst Mode Ratio
+	 * Target ddr frequency from VBT / non burst ddr freq
+	 * multiply by 100 to preserve remainder
+	 */
+	if (intel_dsi->video_mode_format == VIDEO_MODE_BURST) {
+		if (mipi_config->target_burst_mode_freq) {
+			computed_ddr =
+				(pclk * bits_per_pixel) / intel_dsi->lane_count;
+
+			if (mipi_config->target_burst_mode_freq <
+								computed_ddr) {
+				DRM_ERROR("Burst mode freq is less than computed\n");
+				return false;
+			}
+
+			burst_mode_ratio = DIV_ROUND_UP(
+				mipi_config->target_burst_mode_freq * 100,
+				computed_ddr);
+
+			pclk = DIV_ROUND_UP(pclk * burst_mode_ratio, 100);
+		} else {
+			DRM_ERROR("Burst mode target is not set\n");
+			return false;
+		}
+	} else
+		burst_mode_ratio = 100;
+
+	intel_dsi->burst_mode_ratio = burst_mode_ratio;
+	intel_dsi->pclk = pclk;
+
+	bitrate = (pclk * bits_per_pixel) / intel_dsi->lane_count;
+
 	switch (intel_dsi->escape_clk_div) {
 	case 0:
 		tlpx_ns = 50;
diff --git a/drivers/gpu/drm/i915/intel_dsi_pll.c b/drivers/gpu/drm/i915/intel_dsi_pll.c
index d8bb1ea2f0da..fa7a6ca34cd6 100644
--- a/drivers/gpu/drm/i915/intel_dsi_pll.c
+++ b/drivers/gpu/drm/i915/intel_dsi_pll.c
@@ -134,8 +134,7 @@ static u32 dsi_rr_formula(const struct drm_display_mode *mode,
 #else
 
 /* Get DSI clock from pixel clock */
-static u32 dsi_clk_from_pclk(const struct drm_display_mode *mode,
-			  int pixel_format, int lane_count)
+static u32 dsi_clk_from_pclk(u32 pclk, int pixel_format, int lane_count)
 {
 	u32 dsi_clk_khz;
 	u32 bpp;
@@ -156,7 +155,7 @@ static u32 dsi_clk_from_pclk(const struct drm_display_mode *mode,
 
 	/* DSI data rate = pixel clock * bits per pixel / lane count
 	   pixel clock is converted from KHz to Hz */
-	dsi_clk_khz = DIV_ROUND_CLOSEST(mode->clock * bpp, lane_count);
+	dsi_clk_khz = DIV_ROUND_CLOSEST(pclk * bpp, lane_count);
 
 	return dsi_clk_khz;
 }
@@ -191,7 +190,7 @@ static int dsi_calc_mnp(u32 dsi_clk, struct dsi_mnp *dsi_mnp)
 	for (m = 62; m <= 92; m++) {
 		for (p = 2; p <= 6; p++) {
 			/* Find the optimal m and p divisors
-			with minimal error +/- the required clock */
+			   with minimal error +/- the required clock */
 			calc_dsi_clk = (m * ref_clk) / p;
 			if (calc_dsi_clk == target_dsi_clk) {
 				calc_m = m;
@@ -228,15 +227,13 @@ static int dsi_calc_mnp(u32 dsi_clk, struct dsi_mnp *dsi_mnp)
 static void vlv_configure_dsi_pll(struct intel_encoder *encoder)
 {
 	struct drm_i915_private *dev_priv = encoder->base.dev->dev_private;
-	struct intel_crtc *intel_crtc = to_intel_crtc(encoder->base.crtc);
-	const struct drm_display_mode *mode = &intel_crtc->config.adjusted_mode;
 	struct intel_dsi *intel_dsi = enc_to_intel_dsi(&encoder->base);
 	int ret;
 	struct dsi_mnp dsi_mnp;
 	u32 dsi_clk;
 
-	dsi_clk = dsi_clk_from_pclk(mode, intel_dsi->pixel_format,
-						intel_dsi->lane_count);
+	dsi_clk = dsi_clk_from_pclk(intel_dsi->pclk, intel_dsi->pixel_format,
+				    intel_dsi->lane_count);
 
 	ret = dsi_calc_mnp(dsi_clk, &dsi_mnp);
 	if (ret) {
@@ -318,8 +315,8 @@ static void assert_bpp_mismatch(int pixel_format, int pipe_bpp)
 	}
 
 	WARN(bpp != pipe_bpp,
-		"bpp match assertion failure (expected %d, current %d)\n",
-		bpp, pipe_bpp);
+	     "bpp match assertion failure (expected %d, current %d)\n",
+	     bpp, pipe_bpp);
 }
 
 u32 vlv_get_dsi_pclk(struct intel_encoder *encoder, int pipe_bpp)
diff --git a/drivers/gpu/drm/i915/intel_dvo.c b/drivers/gpu/drm/i915/intel_dvo.c
index 56b47d2ffaf7..e40e3df33517 100644
--- a/drivers/gpu/drm/i915/intel_dvo.c
+++ b/drivers/gpu/drm/i915/intel_dvo.c
@@ -85,7 +85,7 @@ static const struct intel_dvo_device intel_dvo_devices[] = {
 	{
 	        .type = INTEL_DVO_CHIP_TMDS,
 		.name = "ns2501",
-		.dvo_reg = DVOC,
+		.dvo_reg = DVOB,
 		.slave_addr = NS2501_ADDR,
 		.dev_ops = &ns2501_ops,
        }
@@ -185,12 +185,13 @@ static void intel_enable_dvo(struct intel_encoder *encoder)
 	u32 dvo_reg = intel_dvo->dev.dvo_reg;
 	u32 temp = I915_READ(dvo_reg);
 
-	I915_WRITE(dvo_reg, temp | DVO_ENABLE);
-	I915_READ(dvo_reg);
 	intel_dvo->dev.dev_ops->mode_set(&intel_dvo->dev,
 					 &crtc->config.requested_mode,
 					 &crtc->config.adjusted_mode);
 
+	I915_WRITE(dvo_reg, temp | DVO_ENABLE);
+	I915_READ(dvo_reg);
+
 	intel_dvo->dev.dev_ops->dpms(&intel_dvo->dev, true);
 }
 
@@ -226,10 +227,6 @@ static void intel_dvo_dpms(struct drm_connector *connector, int mode)
 
 		intel_crtc_update_dpms(crtc);
 
-		intel_dvo->dev.dev_ops->mode_set(&intel_dvo->dev,
-						 &config->requested_mode,
-						 &config->adjusted_mode);
-
 		intel_dvo->dev.dev_ops->dpms(&intel_dvo->dev, true);
 	} else {
 		intel_dvo->dev.dev_ops->dpms(&intel_dvo->dev, false);
diff --git a/drivers/gpu/drm/i915/intel_fbdev.c b/drivers/gpu/drm/i915/intel_fbdev.c
index f475414671d8..850cf7d6578c 100644
--- a/drivers/gpu/drm/i915/intel_fbdev.c
+++ b/drivers/gpu/drm/i915/intel_fbdev.c
@@ -24,8 +24,10 @@
  *     David Airlie
  */
 
+#include <linux/async.h>
 #include <linux/module.h>
 #include <linux/kernel.h>
+#include <linux/console.h>
 #include <linux/errno.h>
 #include <linux/string.h>
 #include <linux/mm.h>
@@ -117,25 +119,25 @@ static int intelfb_alloc(struct drm_fb_helper *helper,
 		goto out;
 	}
 
-	/* Flush everything out, we'll be doing GTT only from now on */
-	ret = intel_pin_and_fence_fb_obj(dev, obj, NULL);
-	if (ret) {
-		DRM_ERROR("failed to pin obj: %d\n", ret);
-		goto out_unref;
-	}
-
 	fb = __intel_framebuffer_create(dev, &mode_cmd, obj);
 	if (IS_ERR(fb)) {
 		ret = PTR_ERR(fb);
-		goto out_unpin;
+		goto out_unref;
+	}
+
+	/* Flush everything out, we'll be doing GTT only from now on */
+	ret = intel_pin_and_fence_fb_obj(NULL, fb, NULL);
+	if (ret) {
+		DRM_ERROR("failed to pin obj: %d\n", ret);
+		goto out_fb;
 	}
 
 	ifbdev->fb = to_intel_framebuffer(fb);
 
 	return 0;
 
-out_unpin:
-	i915_gem_object_ggtt_unpin(obj);
+out_fb:
+	drm_framebuffer_remove(fb);
 out_unref:
 	drm_gem_object_unreference(&obj->base);
 out:
@@ -322,6 +324,7 @@ intel_fb_helper_crtc(struct drm_fb_helper *fb_helper, struct drm_crtc *crtc)
 static bool intel_fb_initial_config(struct drm_fb_helper *fb_helper,
 				    struct drm_fb_helper_crtc **crtcs,
 				    struct drm_display_mode **modes,
+				    struct drm_fb_offset *offsets,
 				    bool *enabled, int width, int height)
 {
 	struct drm_device *dev = fb_helper->dev;
@@ -330,24 +333,8 @@ static bool intel_fb_initial_config(struct drm_fb_helper *fb_helper,
 	bool fallback = true;
 	int num_connectors_enabled = 0;
 	int num_connectors_detected = 0;
-
-	/*
-	 * If the user specified any force options, just bail here
-	 * and use that config.
-	 */
-	for (i = 0; i < fb_helper->connector_count; i++) {
-		struct drm_fb_helper_connector *fb_conn;
-		struct drm_connector *connector;
-
-		fb_conn = fb_helper->connector_info[i];
-		connector = fb_conn->connector;
-
-		if (!enabled[i])
-			continue;
-
-		if (connector->force != DRM_FORCE_UNSPECIFIED)
-			return false;
-	}
+	uint64_t conn_configured = 0, mask;
+	int pass = 0;
 
 	save_enabled = kcalloc(dev->mode_config.num_connector, sizeof(bool),
 			       GFP_KERNEL);
@@ -355,7 +342,8 @@ static bool intel_fb_initial_config(struct drm_fb_helper *fb_helper,
 		return false;
 
 	memcpy(save_enabled, enabled, dev->mode_config.num_connector);
-
+	mask = (1 << fb_helper->connector_count) - 1;
+retry:
 	for (i = 0; i < fb_helper->connector_count; i++) {
 		struct drm_fb_helper_connector *fb_conn;
 		struct drm_connector *connector;
@@ -365,20 +353,38 @@ static bool intel_fb_initial_config(struct drm_fb_helper *fb_helper,
 		fb_conn = fb_helper->connector_info[i];
 		connector = fb_conn->connector;
 
+		if (conn_configured & (1 << i))
+			continue;
+
+		if (pass == 0 && !connector->has_tile)
+			continue;
+
 		if (connector->status == connector_status_connected)
 			num_connectors_detected++;
 
 		if (!enabled[i]) {
 			DRM_DEBUG_KMS("connector %s not enabled, skipping\n",
 				      connector->name);
+			conn_configured |= (1 << i);
+			continue;
+		}
+
+		if (connector->force == DRM_FORCE_OFF) {
+			DRM_DEBUG_KMS("connector %s is disabled by user, skipping\n",
+				      connector->name);
+			enabled[i] = false;
 			continue;
 		}
 
 		encoder = connector->encoder;
 		if (!encoder || WARN_ON(!encoder->crtc)) {
+			if (connector->force > DRM_FORCE_OFF)
+				goto bail;
+
 			DRM_DEBUG_KMS("connector %s has no encoder or crtc, skipping\n",
 				      connector->name);
 			enabled[i] = false;
+			conn_configured |= (1 << i);
 			continue;
 		}
 
@@ -394,8 +400,7 @@ static bool intel_fb_initial_config(struct drm_fb_helper *fb_helper,
 		for (j = 0; j < fb_helper->connector_count; j++) {
 			if (crtcs[j] == new_crtc) {
 				DRM_DEBUG_KMS("fallback: cloned configuration\n");
-				fallback = true;
-				goto out;
+				goto bail;
 			}
 		}
 
@@ -407,8 +412,8 @@ static bool intel_fb_initial_config(struct drm_fb_helper *fb_helper,
 
 		/* try for preferred next */
 		if (!modes[i]) {
-			DRM_DEBUG_KMS("looking for preferred mode on connector %s\n",
-				      connector->name);
+			DRM_DEBUG_KMS("looking for preferred mode on connector %s %d\n",
+				      connector->name, connector->has_tile);
 			modes[i] = drm_has_preferred_mode(fb_conn, width,
 							  height);
 		}
@@ -451,6 +456,12 @@ static bool intel_fb_initial_config(struct drm_fb_helper *fb_helper,
 			      modes[i]->flags & DRM_MODE_FLAG_INTERLACE ? "i" :"");
 
 		fallback = false;
+		conn_configured |= (1 << i);
+	}
+
+	if ((conn_configured & mask) != mask) {
+		pass++;
+		goto retry;
 	}
 
 	/*
@@ -466,8 +477,8 @@ static bool intel_fb_initial_config(struct drm_fb_helper *fb_helper,
 		fallback = true;
 	}
 
-out:
 	if (fallback) {
+bail:
 		DRM_DEBUG_KMS("Not using firmware configuration\n");
 		memcpy(enabled, save_enabled, dev->mode_config.num_connector);
 		kfree(save_enabled);
@@ -636,6 +647,15 @@ out:
 	return false;
 }
 
+static void intel_fbdev_suspend_worker(struct work_struct *work)
+{
+	intel_fbdev_set_suspend(container_of(work,
+					     struct drm_i915_private,
+					     fbdev_suspend_work)->dev,
+				FBINFO_STATE_RUNNING,
+				true);
+}
+
 int intel_fbdev_init(struct drm_device *dev)
 {
 	struct intel_fbdev *ifbdev;
@@ -662,14 +682,16 @@ int intel_fbdev_init(struct drm_device *dev)
 	}
 
 	dev_priv->fbdev = ifbdev;
+	INIT_WORK(&dev_priv->fbdev_suspend_work, intel_fbdev_suspend_worker);
+
 	drm_fb_helper_single_add_all_connectors(&ifbdev->helper);
 
 	return 0;
 }
 
-void intel_fbdev_initial_config(struct drm_device *dev)
+void intel_fbdev_initial_config(void *data, async_cookie_t cookie)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = data;
 	struct intel_fbdev *ifbdev = dev_priv->fbdev;
 
 	/* Due to peculiar init order wrt to hpd handling this is separate. */
@@ -682,12 +704,15 @@ void intel_fbdev_fini(struct drm_device *dev)
 	if (!dev_priv->fbdev)
 		return;
 
+	flush_work(&dev_priv->fbdev_suspend_work);
+
+	async_synchronize_full();
 	intel_fbdev_destroy(dev, dev_priv->fbdev);
 	kfree(dev_priv->fbdev);
 	dev_priv->fbdev = NULL;
 }
 
-void intel_fbdev_set_suspend(struct drm_device *dev, int state)
+void intel_fbdev_set_suspend(struct drm_device *dev, int state, bool synchronous)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_fbdev *ifbdev = dev_priv->fbdev;
@@ -698,6 +723,33 @@ void intel_fbdev_set_suspend(struct drm_device *dev, int state)
 
 	info = ifbdev->helper.fbdev;
 
+	if (synchronous) {
+		/* Flush any pending work to turn the console on, and then
+		 * wait to turn it off. It must be synchronous as we are
+		 * about to suspend or unload the driver.
+		 *
+		 * Note that from within the work-handler, we cannot flush
+		 * ourselves, so only flush outstanding work upon suspend!
+		 */
+		if (state != FBINFO_STATE_RUNNING)
+			flush_work(&dev_priv->fbdev_suspend_work);
+		console_lock();
+	} else {
+		/*
+		 * The console lock can be pretty contented on resume due
+		 * to all the printk activity.  Try to keep it out of the hot
+		 * path of resume if possible.
+		 */
+		WARN_ON(state != FBINFO_STATE_RUNNING);
+		if (!console_trylock()) {
+			/* Don't block our own workqueue as this can
+			 * be run in parallel with other i915.ko tasks.
+			 */
+			schedule_work(&dev_priv->fbdev_suspend_work);
+			return;
+		}
+	}
+
 	/* On resume from hibernation: If the object is shmemfs backed, it has
 	 * been restored from swap. If the object is stolen however, it will be
 	 * full of whatever garbage was left in there.
@@ -706,6 +758,7 @@ void intel_fbdev_set_suspend(struct drm_device *dev, int state)
 		memset_io(info->screen_base, 0, info->screen_size);
 
 	fb_set_suspend(info, state);
+	console_unlock();
 }
 
 void intel_fbdev_output_poll_changed(struct drm_device *dev)
diff --git a/drivers/gpu/drm/i915/intel_fifo_underrun.c b/drivers/gpu/drm/i915/intel_fifo_underrun.c
new file mode 100644
index 000000000000..77af512d2d35
--- /dev/null
+++ b/drivers/gpu/drm/i915/intel_fifo_underrun.c
@@ -0,0 +1,381 @@
+/*
+ * Copyright © 2014 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ *
+ * Authors:
+ *    Daniel Vetter <daniel.vetter@ffwll.ch>
+ *
+ */
+
+#include "i915_drv.h"
+#include "intel_drv.h"
+
+/**
+ * DOC: fifo underrun handling
+ *
+ * The i915 driver checks for display fifo underruns using the interrupt signals
+ * provided by the hardware. This is enabled by default and fairly useful to
+ * debug display issues, especially watermark settings.
+ *
+ * If an underrun is detected this is logged into dmesg. To avoid flooding logs
+ * and occupying the cpu underrun interrupts are disabled after the first
+ * occurrence until the next modeset on a given pipe.
+ *
+ * Note that underrun detection on gmch platforms is a bit more ugly since there
+ * is no interrupt (despite that the signalling bit is in the PIPESTAT pipe
+ * interrupt register). Also on some other platforms underrun interrupts are
+ * shared, which means that if we detect an underrun we need to disable underrun
+ * reporting on all pipes.
+ *
+ * The code also supports underrun detection on the PCH transcoder.
+ */
+
+static bool ivb_can_enable_err_int(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct intel_crtc *crtc;
+	enum pipe pipe;
+
+	assert_spin_locked(&dev_priv->irq_lock);
+
+	for_each_pipe(dev_priv, pipe) {
+		crtc = to_intel_crtc(dev_priv->pipe_to_crtc_mapping[pipe]);
+
+		if (crtc->cpu_fifo_underrun_disabled)
+			return false;
+	}
+
+	return true;
+}
+
+static bool cpt_can_enable_serr_int(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	enum pipe pipe;
+	struct intel_crtc *crtc;
+
+	assert_spin_locked(&dev_priv->irq_lock);
+
+	for_each_pipe(dev_priv, pipe) {
+		crtc = to_intel_crtc(dev_priv->pipe_to_crtc_mapping[pipe]);
+
+		if (crtc->pch_fifo_underrun_disabled)
+			return false;
+	}
+
+	return true;
+}
+
+/**
+ * i9xx_check_fifo_underruns - check for fifo underruns
+ * @dev_priv: i915 device instance
+ *
+ * This function checks for fifo underruns on GMCH platforms. This needs to be
+ * done manually on modeset to make sure that we catch all underruns since they
+ * do not generate an interrupt by themselves on these platforms.
+ */
+void i9xx_check_fifo_underruns(struct drm_i915_private *dev_priv)
+{
+	struct intel_crtc *crtc;
+
+	spin_lock_irq(&dev_priv->irq_lock);
+
+	for_each_intel_crtc(dev_priv->dev, crtc) {
+		u32 reg = PIPESTAT(crtc->pipe);
+		u32 pipestat;
+
+		if (crtc->cpu_fifo_underrun_disabled)
+			continue;
+
+		pipestat = I915_READ(reg) & 0xffff0000;
+		if ((pipestat & PIPE_FIFO_UNDERRUN_STATUS) == 0)
+			continue;
+
+		I915_WRITE(reg, pipestat | PIPE_FIFO_UNDERRUN_STATUS);
+		POSTING_READ(reg);
+
+		DRM_ERROR("pipe %c underrun\n", pipe_name(crtc->pipe));
+	}
+
+	spin_unlock_irq(&dev_priv->irq_lock);
+}
+
+static void i9xx_set_fifo_underrun_reporting(struct drm_device *dev,
+					     enum pipe pipe,
+					     bool enable, bool old)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	u32 reg = PIPESTAT(pipe);
+	u32 pipestat = I915_READ(reg) & 0xffff0000;
+
+	assert_spin_locked(&dev_priv->irq_lock);
+
+	if (enable) {
+		I915_WRITE(reg, pipestat | PIPE_FIFO_UNDERRUN_STATUS);
+		POSTING_READ(reg);
+	} else {
+		if (old && pipestat & PIPE_FIFO_UNDERRUN_STATUS)
+			DRM_ERROR("pipe %c underrun\n", pipe_name(pipe));
+	}
+}
+
+static void ironlake_set_fifo_underrun_reporting(struct drm_device *dev,
+						 enum pipe pipe, bool enable)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	uint32_t bit = (pipe == PIPE_A) ? DE_PIPEA_FIFO_UNDERRUN :
+					  DE_PIPEB_FIFO_UNDERRUN;
+
+	if (enable)
+		ironlake_enable_display_irq(dev_priv, bit);
+	else
+		ironlake_disable_display_irq(dev_priv, bit);
+}
+
+static void ivybridge_set_fifo_underrun_reporting(struct drm_device *dev,
+						  enum pipe pipe,
+						  bool enable, bool old)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	if (enable) {
+		I915_WRITE(GEN7_ERR_INT, ERR_INT_FIFO_UNDERRUN(pipe));
+
+		if (!ivb_can_enable_err_int(dev))
+			return;
+
+		ironlake_enable_display_irq(dev_priv, DE_ERR_INT_IVB);
+	} else {
+		ironlake_disable_display_irq(dev_priv, DE_ERR_INT_IVB);
+
+		if (old &&
+		    I915_READ(GEN7_ERR_INT) & ERR_INT_FIFO_UNDERRUN(pipe)) {
+			DRM_ERROR("uncleared fifo underrun on pipe %c\n",
+				  pipe_name(pipe));
+		}
+	}
+}
+
+static void broadwell_set_fifo_underrun_reporting(struct drm_device *dev,
+						  enum pipe pipe, bool enable)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+
+	assert_spin_locked(&dev_priv->irq_lock);
+
+	if (enable)
+		dev_priv->de_irq_mask[pipe] &= ~GEN8_PIPE_FIFO_UNDERRUN;
+	else
+		dev_priv->de_irq_mask[pipe] |= GEN8_PIPE_FIFO_UNDERRUN;
+	I915_WRITE(GEN8_DE_PIPE_IMR(pipe), dev_priv->de_irq_mask[pipe]);
+	POSTING_READ(GEN8_DE_PIPE_IMR(pipe));
+}
+
+static void ibx_set_fifo_underrun_reporting(struct drm_device *dev,
+					    enum transcoder pch_transcoder,
+					    bool enable)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	uint32_t bit = (pch_transcoder == TRANSCODER_A) ?
+		       SDE_TRANSA_FIFO_UNDER : SDE_TRANSB_FIFO_UNDER;
+
+	if (enable)
+		ibx_enable_display_interrupt(dev_priv, bit);
+	else
+		ibx_disable_display_interrupt(dev_priv, bit);
+}
+
+static void cpt_set_fifo_underrun_reporting(struct drm_device *dev,
+					    enum transcoder pch_transcoder,
+					    bool enable, bool old)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+
+	if (enable) {
+		I915_WRITE(SERR_INT,
+			   SERR_INT_TRANS_FIFO_UNDERRUN(pch_transcoder));
+
+		if (!cpt_can_enable_serr_int(dev))
+			return;
+
+		ibx_enable_display_interrupt(dev_priv, SDE_ERROR_CPT);
+	} else {
+		ibx_disable_display_interrupt(dev_priv, SDE_ERROR_CPT);
+
+		if (old && I915_READ(SERR_INT) &
+		    SERR_INT_TRANS_FIFO_UNDERRUN(pch_transcoder)) {
+			DRM_ERROR("uncleared pch fifo underrun on pch transcoder %c\n",
+				  transcoder_name(pch_transcoder));
+		}
+	}
+}
+
+static bool __intel_set_cpu_fifo_underrun_reporting(struct drm_device *dev,
+						    enum pipe pipe, bool enable)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_crtc *crtc = dev_priv->pipe_to_crtc_mapping[pipe];
+	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
+	bool old;
+
+	assert_spin_locked(&dev_priv->irq_lock);
+
+	old = !intel_crtc->cpu_fifo_underrun_disabled;
+	intel_crtc->cpu_fifo_underrun_disabled = !enable;
+
+	if (HAS_GMCH_DISPLAY(dev))
+		i9xx_set_fifo_underrun_reporting(dev, pipe, enable, old);
+	else if (IS_GEN5(dev) || IS_GEN6(dev))
+		ironlake_set_fifo_underrun_reporting(dev, pipe, enable);
+	else if (IS_GEN7(dev))
+		ivybridge_set_fifo_underrun_reporting(dev, pipe, enable, old);
+	else if (IS_GEN8(dev) || IS_GEN9(dev))
+		broadwell_set_fifo_underrun_reporting(dev, pipe, enable);
+
+	return old;
+}
+
+/**
+ * intel_set_cpu_fifo_underrun_reporting - set cpu fifo underrrun reporting state
+ * @dev_priv: i915 device instance
+ * @pipe: (CPU) pipe to set state for
+ * @enable: whether underruns should be reported or not
+ *
+ * This function sets the fifo underrun state for @pipe. It is used in the
+ * modeset code to avoid false positives since on many platforms underruns are
+ * expected when disabling or enabling the pipe.
+ *
+ * Notice that on some platforms disabling underrun reports for one pipe
+ * disables for all due to shared interrupts. Actual reporting is still per-pipe
+ * though.
+ *
+ * Returns the previous state of underrun reporting.
+ */
+bool intel_set_cpu_fifo_underrun_reporting(struct drm_i915_private *dev_priv,
+					   enum pipe pipe, bool enable)
+{
+	unsigned long flags;
+	bool ret;
+
+	spin_lock_irqsave(&dev_priv->irq_lock, flags);
+	ret = __intel_set_cpu_fifo_underrun_reporting(dev_priv->dev, pipe,
+						      enable);
+	spin_unlock_irqrestore(&dev_priv->irq_lock, flags);
+
+	return ret;
+}
+
+static bool
+__cpu_fifo_underrun_reporting_enabled(struct drm_i915_private *dev_priv,
+				      enum pipe pipe)
+{
+	struct drm_crtc *crtc = dev_priv->pipe_to_crtc_mapping[pipe];
+	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
+
+	return !intel_crtc->cpu_fifo_underrun_disabled;
+}
+
+/**
+ * intel_set_pch_fifo_underrun_reporting - set PCH fifo underrun reporting state
+ * @dev_priv: i915 device instance
+ * @pch_transcoder: the PCH transcoder (same as pipe on IVB and older)
+ * @enable: whether underruns should be reported or not
+ *
+ * This function makes us disable or enable PCH fifo underruns for a specific
+ * PCH transcoder. Notice that on some PCHs (e.g. CPT/PPT), disabling FIFO
+ * underrun reporting for one transcoder may also disable all the other PCH
+ * error interruts for the other transcoders, due to the fact that there's just
+ * one interrupt mask/enable bit for all the transcoders.
+ *
+ * Returns the previous state of underrun reporting.
+ */
+bool intel_set_pch_fifo_underrun_reporting(struct drm_i915_private *dev_priv,
+					   enum transcoder pch_transcoder,
+					   bool enable)
+{
+	struct drm_crtc *crtc = dev_priv->pipe_to_crtc_mapping[pch_transcoder];
+	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
+	unsigned long flags;
+	bool old;
+
+	/*
+	 * NOTE: Pre-LPT has a fixed cpu pipe -> pch transcoder mapping, but LPT
+	 * has only one pch transcoder A that all pipes can use. To avoid racy
+	 * pch transcoder -> pipe lookups from interrupt code simply store the
+	 * underrun statistics in crtc A. Since we never expose this anywhere
+	 * nor use it outside of the fifo underrun code here using the "wrong"
+	 * crtc on LPT won't cause issues.
+	 */
+
+	spin_lock_irqsave(&dev_priv->irq_lock, flags);
+
+	old = !intel_crtc->pch_fifo_underrun_disabled;
+	intel_crtc->pch_fifo_underrun_disabled = !enable;
+
+	if (HAS_PCH_IBX(dev_priv->dev))
+		ibx_set_fifo_underrun_reporting(dev_priv->dev, pch_transcoder,
+						enable);
+	else
+		cpt_set_fifo_underrun_reporting(dev_priv->dev, pch_transcoder,
+						enable, old);
+
+	spin_unlock_irqrestore(&dev_priv->irq_lock, flags);
+	return old;
+}
+
+/**
+ * intel_pch_fifo_underrun_irq_handler - handle PCH fifo underrun interrupt
+ * @dev_priv: i915 device instance
+ * @pipe: (CPU) pipe to set state for
+ *
+ * This handles a CPU fifo underrun interrupt, generating an underrun warning
+ * into dmesg if underrun reporting is enabled and then disables the underrun
+ * interrupt to avoid an irq storm.
+ */
+void intel_cpu_fifo_underrun_irq_handler(struct drm_i915_private *dev_priv,
+					 enum pipe pipe)
+{
+	/* GMCH can't disable fifo underruns, filter them. */
+	if (HAS_GMCH_DISPLAY(dev_priv->dev) &&
+	    !__cpu_fifo_underrun_reporting_enabled(dev_priv, pipe))
+		return;
+
+	if (intel_set_cpu_fifo_underrun_reporting(dev_priv, pipe, false))
+		DRM_ERROR("CPU pipe %c FIFO underrun\n",
+			  pipe_name(pipe));
+}
+
+/**
+ * intel_pch_fifo_underrun_irq_handler - handle PCH fifo underrun interrupt
+ * @dev_priv: i915 device instance
+ * @pch_transcoder: the PCH transcoder (same as pipe on IVB and older)
+ *
+ * This handles a PCH fifo underrun interrupt, generating an underrun warning
+ * into dmesg if underrun reporting is enabled and then disables the underrun
+ * interrupt to avoid an irq storm.
+ */
+void intel_pch_fifo_underrun_irq_handler(struct drm_i915_private *dev_priv,
+					 enum transcoder pch_transcoder)
+{
+	if (intel_set_pch_fifo_underrun_reporting(dev_priv, pch_transcoder,
+						  false))
+		DRM_ERROR("PCH transcoder %c FIFO underrun\n",
+			  transcoder_name(pch_transcoder));
+}
diff --git a/drivers/gpu/drm/i915/intel_frontbuffer.c b/drivers/gpu/drm/i915/intel_frontbuffer.c
new file mode 100644
index 000000000000..79f6d72179c5
--- /dev/null
+++ b/drivers/gpu/drm/i915/intel_frontbuffer.c
@@ -0,0 +1,279 @@
+/*
+ * Copyright © 2014 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ *
+ * Authors:
+ *	Daniel Vetter <daniel.vetter@ffwll.ch>
+ */
+
+/**
+ * DOC: frontbuffer tracking
+ *
+ * Many features require us to track changes to the currently active
+ * frontbuffer, especially rendering targeted at the frontbuffer.
+ *
+ * To be able to do so GEM tracks frontbuffers using a bitmask for all possible
+ * frontbuffer slots through i915_gem_track_fb(). The function in this file are
+ * then called when the contents of the frontbuffer are invalidated, when
+ * frontbuffer rendering has stopped again to flush out all the changes and when
+ * the frontbuffer is exchanged with a flip. Subsystems interested in
+ * frontbuffer changes (e.g. PSR, FBC, DRRS) should directly put their callbacks
+ * into the relevant places and filter for the frontbuffer slots that they are
+ * interested int.
+ *
+ * On a high level there are two types of powersaving features. The first one
+ * work like a special cache (FBC and PSR) and are interested when they should
+ * stop caching and when to restart caching. This is done by placing callbacks
+ * into the invalidate and the flush functions: At invalidate the caching must
+ * be stopped and at flush time it can be restarted. And maybe they need to know
+ * when the frontbuffer changes (e.g. when the hw doesn't initiate an invalidate
+ * and flush on its own) which can be achieved with placing callbacks into the
+ * flip functions.
+ *
+ * The other type of display power saving feature only cares about busyness
+ * (e.g. DRRS). In that case all three (invalidate, flush and flip) indicate
+ * busyness. There is no direct way to detect idleness. Instead an idle timer
+ * work delayed work should be started from the flush and flip functions and
+ * cancelled as soon as busyness is detected.
+ *
+ * Note that there's also an older frontbuffer activity tracking scheme which
+ * just tracks general activity. This is done by the various mark_busy and
+ * mark_idle functions. For display power management features using these
+ * functions is deprecated and should be avoided.
+ */
+
+#include <drm/drmP.h>
+
+#include "intel_drv.h"
+#include "i915_drv.h"
+
+static void intel_increase_pllclock(struct drm_device *dev,
+				    enum pipe pipe)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	int dpll_reg = DPLL(pipe);
+	int dpll;
+
+	if (!HAS_GMCH_DISPLAY(dev))
+		return;
+
+	if (!dev_priv->lvds_downclock_avail)
+		return;
+
+	dpll = I915_READ(dpll_reg);
+	if (!HAS_PIPE_CXSR(dev) && (dpll & DISPLAY_RATE_SELECT_FPA1)) {
+		DRM_DEBUG_DRIVER("upclocking LVDS\n");
+
+		assert_panel_unlocked(dev_priv, pipe);
+
+		dpll &= ~DISPLAY_RATE_SELECT_FPA1;
+		I915_WRITE(dpll_reg, dpll);
+		intel_wait_for_vblank(dev, pipe);
+
+		dpll = I915_READ(dpll_reg);
+		if (dpll & DISPLAY_RATE_SELECT_FPA1)
+			DRM_DEBUG_DRIVER("failed to upclock LVDS!\n");
+	}
+}
+
+/**
+ * intel_mark_fb_busy - mark given planes as busy
+ * @dev: DRM device
+ * @frontbuffer_bits: bits for the affected planes
+ * @ring: optional ring for asynchronous commands
+ *
+ * This function gets called every time the screen contents change. It can be
+ * used to keep e.g. the update rate at the nominal refresh rate with DRRS.
+ */
+static void intel_mark_fb_busy(struct drm_device *dev,
+			       unsigned frontbuffer_bits,
+			       struct intel_engine_cs *ring)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	enum pipe pipe;
+
+	if (!i915.powersave)
+		return;
+
+	for_each_pipe(dev_priv, pipe) {
+		if (!(frontbuffer_bits & INTEL_FRONTBUFFER_ALL_MASK(pipe)))
+			continue;
+
+		intel_increase_pllclock(dev, pipe);
+		if (ring && intel_fbc_enabled(dev))
+			ring->fbc_dirty = true;
+	}
+}
+
+/**
+ * intel_fb_obj_invalidate - invalidate frontbuffer object
+ * @obj: GEM object to invalidate
+ * @ring: set for asynchronous rendering
+ *
+ * This function gets called every time rendering on the given object starts and
+ * frontbuffer caching (fbc, low refresh rate for DRRS, panel self refresh) must
+ * be invalidated. If @ring is non-NULL any subsequent invalidation will be delayed
+ * until the rendering completes or a flip on this frontbuffer plane is
+ * scheduled.
+ */
+void intel_fb_obj_invalidate(struct drm_i915_gem_object *obj,
+			     struct intel_engine_cs *ring)
+{
+	struct drm_device *dev = obj->base.dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+
+	WARN_ON(!mutex_is_locked(&dev->struct_mutex));
+
+	if (!obj->frontbuffer_bits)
+		return;
+
+	if (ring) {
+		mutex_lock(&dev_priv->fb_tracking.lock);
+		dev_priv->fb_tracking.busy_bits
+			|= obj->frontbuffer_bits;
+		dev_priv->fb_tracking.flip_bits
+			&= ~obj->frontbuffer_bits;
+		mutex_unlock(&dev_priv->fb_tracking.lock);
+	}
+
+	intel_mark_fb_busy(dev, obj->frontbuffer_bits, ring);
+
+	intel_psr_invalidate(dev, obj->frontbuffer_bits);
+}
+
+/**
+ * intel_frontbuffer_flush - flush frontbuffer
+ * @dev: DRM device
+ * @frontbuffer_bits: frontbuffer plane tracking bits
+ *
+ * This function gets called every time rendering on the given planes has
+ * completed and frontbuffer caching can be started again. Flushes will get
+ * delayed if they're blocked by some outstanding asynchronous rendering.
+ *
+ * Can be called without any locks held.
+ */
+void intel_frontbuffer_flush(struct drm_device *dev,
+			     unsigned frontbuffer_bits)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+
+	/* Delay flushing when rings are still busy.*/
+	mutex_lock(&dev_priv->fb_tracking.lock);
+	frontbuffer_bits &= ~dev_priv->fb_tracking.busy_bits;
+	mutex_unlock(&dev_priv->fb_tracking.lock);
+
+	intel_mark_fb_busy(dev, frontbuffer_bits, NULL);
+
+	intel_psr_flush(dev, frontbuffer_bits);
+
+	/*
+	 * FIXME: Unconditional fbc flushing here is a rather gross hack and
+	 * needs to be reworked into a proper frontbuffer tracking scheme like
+	 * psr employs.
+	 */
+	if (dev_priv->fbc.need_sw_cache_clean) {
+		dev_priv->fbc.need_sw_cache_clean = false;
+		bdw_fbc_sw_flush(dev, FBC_REND_CACHE_CLEAN);
+	}
+}
+
+/**
+ * intel_fb_obj_flush - flush frontbuffer object
+ * @obj: GEM object to flush
+ * @retire: set when retiring asynchronous rendering
+ *
+ * This function gets called every time rendering on the given object has
+ * completed and frontbuffer caching can be started again. If @retire is true
+ * then any delayed flushes will be unblocked.
+ */
+void intel_fb_obj_flush(struct drm_i915_gem_object *obj,
+			bool retire)
+{
+	struct drm_device *dev = obj->base.dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	unsigned frontbuffer_bits;
+
+	WARN_ON(!mutex_is_locked(&dev->struct_mutex));
+
+	if (!obj->frontbuffer_bits)
+		return;
+
+	frontbuffer_bits = obj->frontbuffer_bits;
+
+	if (retire) {
+		mutex_lock(&dev_priv->fb_tracking.lock);
+		/* Filter out new bits since rendering started. */
+		frontbuffer_bits &= dev_priv->fb_tracking.busy_bits;
+
+		dev_priv->fb_tracking.busy_bits &= ~frontbuffer_bits;
+		mutex_unlock(&dev_priv->fb_tracking.lock);
+	}
+
+	intel_frontbuffer_flush(dev, frontbuffer_bits);
+}
+
+/**
+ * intel_frontbuffer_flip_prepare - prepare asynchronous frontbuffer flip
+ * @dev: DRM device
+ * @frontbuffer_bits: frontbuffer plane tracking bits
+ *
+ * This function gets called after scheduling a flip on @obj. The actual
+ * frontbuffer flushing will be delayed until completion is signalled with
+ * intel_frontbuffer_flip_complete. If an invalidate happens in between this
+ * flush will be cancelled.
+ *
+ * Can be called without any locks held.
+ */
+void intel_frontbuffer_flip_prepare(struct drm_device *dev,
+				    unsigned frontbuffer_bits)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+
+	mutex_lock(&dev_priv->fb_tracking.lock);
+	dev_priv->fb_tracking.flip_bits |= frontbuffer_bits;
+	/* Remove stale busy bits due to the old buffer. */
+	dev_priv->fb_tracking.busy_bits &= ~frontbuffer_bits;
+	mutex_unlock(&dev_priv->fb_tracking.lock);
+}
+
+/**
+ * intel_frontbuffer_flip_complete - complete asynchronous frontbuffer flip
+ * @dev: DRM device
+ * @frontbuffer_bits: frontbuffer plane tracking bits
+ *
+ * This function gets called after the flip has been latched and will complete
+ * on the next vblank. It will execute the flush if it hasn't been cancelled yet.
+ *
+ * Can be called without any locks held.
+ */
+void intel_frontbuffer_flip_complete(struct drm_device *dev,
+				     unsigned frontbuffer_bits)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+
+	mutex_lock(&dev_priv->fb_tracking.lock);
+	/* Mask any cancelled flips. */
+	frontbuffer_bits &= dev_priv->fb_tracking.flip_bits;
+	dev_priv->fb_tracking.flip_bits &= ~frontbuffer_bits;
+	mutex_unlock(&dev_priv->fb_tracking.lock);
+
+	intel_frontbuffer_flush(dev, frontbuffer_bits);
+}
diff --git a/drivers/gpu/drm/i915/intel_hdmi.c b/drivers/gpu/drm/i915/intel_hdmi.c
index 5a9de21637b7..3abc2000fce9 100644
--- a/drivers/gpu/drm/i915/intel_hdmi.c
+++ b/drivers/gpu/drm/i915/intel_hdmi.c
@@ -166,6 +166,19 @@ static void g4x_write_infoframe(struct drm_encoder *encoder,
 	POSTING_READ(VIDEO_DIP_CTL);
 }
 
+static bool g4x_infoframe_enabled(struct drm_encoder *encoder)
+{
+	struct drm_device *dev = encoder->dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct intel_digital_port *intel_dig_port = enc_to_dig_port(encoder);
+	u32 val = I915_READ(VIDEO_DIP_CTL);
+
+	if (VIDEO_DIP_PORT(intel_dig_port->port) == (val & VIDEO_DIP_PORT_MASK))
+		return val & VIDEO_DIP_ENABLE;
+
+	return false;
+}
+
 static void ibx_write_infoframe(struct drm_encoder *encoder,
 				enum hdmi_infoframe_type type,
 				const void *frame, ssize_t len)
@@ -204,6 +217,17 @@ static void ibx_write_infoframe(struct drm_encoder *encoder,
 	POSTING_READ(reg);
 }
 
+static bool ibx_infoframe_enabled(struct drm_encoder *encoder)
+{
+	struct drm_device *dev = encoder->dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct intel_crtc *intel_crtc = to_intel_crtc(encoder->crtc);
+	int reg = TVIDEO_DIP_CTL(intel_crtc->pipe);
+	u32 val = I915_READ(reg);
+
+	return val & VIDEO_DIP_ENABLE;
+}
+
 static void cpt_write_infoframe(struct drm_encoder *encoder,
 				enum hdmi_infoframe_type type,
 				const void *frame, ssize_t len)
@@ -245,6 +269,17 @@ static void cpt_write_infoframe(struct drm_encoder *encoder,
 	POSTING_READ(reg);
 }
 
+static bool cpt_infoframe_enabled(struct drm_encoder *encoder)
+{
+	struct drm_device *dev = encoder->dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct intel_crtc *intel_crtc = to_intel_crtc(encoder->crtc);
+	int reg = TVIDEO_DIP_CTL(intel_crtc->pipe);
+	u32 val = I915_READ(reg);
+
+	return val & VIDEO_DIP_ENABLE;
+}
+
 static void vlv_write_infoframe(struct drm_encoder *encoder,
 				enum hdmi_infoframe_type type,
 				const void *frame, ssize_t len)
@@ -283,6 +318,17 @@ static void vlv_write_infoframe(struct drm_encoder *encoder,
 	POSTING_READ(reg);
 }
 
+static bool vlv_infoframe_enabled(struct drm_encoder *encoder)
+{
+	struct drm_device *dev = encoder->dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct intel_crtc *intel_crtc = to_intel_crtc(encoder->crtc);
+	int reg = VLV_TVIDEO_DIP_CTL(intel_crtc->pipe);
+	u32 val = I915_READ(reg);
+
+	return val & VIDEO_DIP_ENABLE;
+}
+
 static void hsw_write_infoframe(struct drm_encoder *encoder,
 				enum hdmi_infoframe_type type,
 				const void *frame, ssize_t len)
@@ -320,6 +366,18 @@ static void hsw_write_infoframe(struct drm_encoder *encoder,
 	POSTING_READ(ctl_reg);
 }
 
+static bool hsw_infoframe_enabled(struct drm_encoder *encoder)
+{
+	struct drm_device *dev = encoder->dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct intel_crtc *intel_crtc = to_intel_crtc(encoder->crtc);
+	u32 ctl_reg = HSW_TVIDEO_DIP_CTL(intel_crtc->config.cpu_transcoder);
+	u32 val = I915_READ(ctl_reg);
+
+	return val & (VIDEO_DIP_ENABLE_AVI_HSW | VIDEO_DIP_ENABLE_SPD_HSW |
+		      VIDEO_DIP_ENABLE_VS_HSW);
+}
+
 /*
  * The data we write to the DIP data buffer registers is 1 byte bigger than the
  * HDMI infoframe size because of an ECC/reserved byte at position 3 (starting
@@ -661,14 +719,6 @@ static void intel_hdmi_prepare(struct intel_encoder *encoder)
 	if (crtc->config.has_hdmi_sink)
 		hdmi_val |= HDMI_MODE_SELECT_HDMI;
 
-	if (crtc->config.has_audio) {
-		WARN_ON(!crtc->config.has_hdmi_sink);
-		DRM_DEBUG_DRIVER("Enabling HDMI audio on pipe %c\n",
-				 pipe_name(crtc->pipe));
-		hdmi_val |= SDVO_AUDIO_ENABLE;
-		intel_write_eld(&encoder->base, adjusted_mode);
-	}
-
 	if (HAS_PCH_CPT(dev))
 		hdmi_val |= SDVO_PIPE_SEL_CPT(crtc->pipe);
 	else if (IS_CHERRYVIEW(dev))
@@ -690,7 +740,7 @@ static bool intel_hdmi_get_hw_state(struct intel_encoder *encoder,
 	u32 tmp;
 
 	power_domain = intel_display_port_power_domain(encoder);
-	if (!intel_display_power_enabled(dev_priv, power_domain))
+	if (!intel_display_power_is_enabled(dev_priv, power_domain))
 		return false;
 
 	tmp = I915_READ(intel_hdmi->hdmi_reg);
@@ -732,6 +782,9 @@ static void intel_hdmi_get_config(struct intel_encoder *encoder,
 	if (tmp & HDMI_MODE_SELECT_HDMI)
 		pipe_config->has_hdmi_sink = true;
 
+	if (intel_hdmi->infoframe_enabled(&encoder->base))
+		pipe_config->has_infoframe = true;
+
 	if (tmp & SDVO_AUDIO_ENABLE)
 		pipe_config->has_audio = true;
 
@@ -791,6 +844,13 @@ static void intel_enable_hdmi(struct intel_encoder *encoder)
 		I915_WRITE(intel_hdmi->hdmi_reg, temp);
 		POSTING_READ(intel_hdmi->hdmi_reg);
 	}
+
+	if (intel_crtc->config.has_audio) {
+		WARN_ON(!intel_crtc->config.has_hdmi_sink);
+		DRM_DEBUG_DRIVER("Enabling HDMI audio on pipe %c\n",
+				 pipe_name(intel_crtc->pipe));
+		intel_audio_codec_enable(encoder);
+	}
 }
 
 static void vlv_enable_hdmi(struct intel_encoder *encoder)
@@ -802,9 +862,13 @@ static void intel_disable_hdmi(struct intel_encoder *encoder)
 	struct drm_device *dev = encoder->base.dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_hdmi *intel_hdmi = enc_to_intel_hdmi(&encoder->base);
+	struct intel_crtc *crtc = to_intel_crtc(encoder->base.crtc);
 	u32 temp;
 	u32 enable_bits = SDVO_ENABLE | SDVO_AUDIO_ENABLE;
 
+	if (crtc->config.has_audio)
+		intel_audio_codec_disable(encoder);
+
 	temp = I915_READ(intel_hdmi->hdmi_reg);
 
 	/* HW workaround for IBX, we need to move the port to transcoder A
@@ -869,10 +933,15 @@ static enum drm_mode_status
 intel_hdmi_mode_valid(struct drm_connector *connector,
 		      struct drm_display_mode *mode)
 {
-	if (mode->clock > hdmi_portclock_limit(intel_attached_hdmi(connector),
-					       true))
+	int clock = mode->clock;
+
+	if (mode->flags & DRM_MODE_FLAG_DBLCLK)
+		clock *= 2;
+
+	if (clock > hdmi_portclock_limit(intel_attached_hdmi(connector),
+					 true))
 		return MODE_CLOCK_HIGH;
-	if (mode->clock < 20000)
+	if (clock < 20000)
 		return MODE_CLOCK_LOW;
 
 	if (mode->flags & DRM_MODE_FLAG_DBLSCAN)
@@ -890,7 +959,7 @@ static bool hdmi_12bpc_possible(struct intel_crtc *crtc)
 	if (HAS_GMCH_DISPLAY(dev))
 		return false;
 
-	list_for_each_entry(encoder, &dev->mode_config.encoder_list, base.head) {
+	for_each_intel_encoder(dev, encoder) {
 		if (encoder->new_crtc != crtc)
 			continue;
 
@@ -917,6 +986,9 @@ bool intel_hdmi_compute_config(struct intel_encoder *encoder,
 
 	pipe_config->has_hdmi_sink = intel_hdmi->has_hdmi_sink;
 
+	if (pipe_config->has_hdmi_sink)
+		pipe_config->has_infoframe = true;
+
 	if (intel_hdmi->color_range_auto) {
 		/* See CEA-861-E - 5.1 Default Encoding Parameters */
 		if (pipe_config->has_hdmi_sink &&
@@ -926,6 +998,10 @@ bool intel_hdmi_compute_config(struct intel_encoder *encoder,
 			intel_hdmi->color_range = 0;
 	}
 
+	if (adjusted_mode->flags & DRM_MODE_FLAG_DBLCLK) {
+		pipe_config->pixel_multiplier = 2;
+	}
+
 	if (intel_hdmi->color_range)
 		pipe_config->limited_color_range = true;
 
@@ -967,104 +1043,117 @@ bool intel_hdmi_compute_config(struct intel_encoder *encoder,
 	return true;
 }
 
-static enum drm_connector_status
-intel_hdmi_detect(struct drm_connector *connector, bool force)
+static void
+intel_hdmi_unset_edid(struct drm_connector *connector)
 {
-	struct drm_device *dev = connector->dev;
 	struct intel_hdmi *intel_hdmi = intel_attached_hdmi(connector);
-	struct intel_digital_port *intel_dig_port =
-		hdmi_to_dig_port(intel_hdmi);
-	struct intel_encoder *intel_encoder = &intel_dig_port->base;
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct edid *edid;
-	enum intel_display_power_domain power_domain;
-	enum drm_connector_status status = connector_status_disconnected;
 
-	DRM_DEBUG_KMS("[CONNECTOR:%d:%s]\n",
-		      connector->base.id, connector->name);
+	intel_hdmi->has_hdmi_sink = false;
+	intel_hdmi->has_audio = false;
+	intel_hdmi->rgb_quant_range_selectable = false;
+
+	kfree(to_intel_connector(connector)->detect_edid);
+	to_intel_connector(connector)->detect_edid = NULL;
+}
+
+static bool
+intel_hdmi_set_edid(struct drm_connector *connector)
+{
+	struct drm_i915_private *dev_priv = to_i915(connector->dev);
+	struct intel_hdmi *intel_hdmi = intel_attached_hdmi(connector);
+	struct intel_encoder *intel_encoder =
+		&hdmi_to_dig_port(intel_hdmi)->base;
+	enum intel_display_power_domain power_domain;
+	struct edid *edid;
+	bool connected = false;
 
 	power_domain = intel_display_port_power_domain(intel_encoder);
 	intel_display_power_get(dev_priv, power_domain);
 
-	intel_hdmi->has_hdmi_sink = false;
-	intel_hdmi->has_audio = false;
-	intel_hdmi->rgb_quant_range_selectable = false;
 	edid = drm_get_edid(connector,
 			    intel_gmbus_get_adapter(dev_priv,
 						    intel_hdmi->ddc_bus));
 
-	if (edid) {
-		if (edid->input & DRM_EDID_INPUT_DIGITAL) {
-			status = connector_status_connected;
-			if (intel_hdmi->force_audio != HDMI_AUDIO_OFF_DVI)
-				intel_hdmi->has_hdmi_sink =
-						drm_detect_hdmi_monitor(edid);
-			intel_hdmi->has_audio = drm_detect_monitor_audio(edid);
-			intel_hdmi->rgb_quant_range_selectable =
-				drm_rgb_quant_range_selectable(edid);
-		}
-		kfree(edid);
-	}
+	intel_display_power_put(dev_priv, power_domain);
+
+	to_intel_connector(connector)->detect_edid = edid;
+	if (edid && edid->input & DRM_EDID_INPUT_DIGITAL) {
+		intel_hdmi->rgb_quant_range_selectable =
+			drm_rgb_quant_range_selectable(edid);
 
-	if (status == connector_status_connected) {
+		intel_hdmi->has_audio = drm_detect_monitor_audio(edid);
 		if (intel_hdmi->force_audio != HDMI_AUDIO_AUTO)
 			intel_hdmi->has_audio =
-				(intel_hdmi->force_audio == HDMI_AUDIO_ON);
-		intel_encoder->type = INTEL_OUTPUT_HDMI;
+				intel_hdmi->force_audio == HDMI_AUDIO_ON;
+
+		if (intel_hdmi->force_audio != HDMI_AUDIO_OFF_DVI)
+			intel_hdmi->has_hdmi_sink =
+				drm_detect_hdmi_monitor(edid);
+
+		connected = true;
 	}
 
-	intel_display_power_put(dev_priv, power_domain);
+	return connected;
+}
+
+static enum drm_connector_status
+intel_hdmi_detect(struct drm_connector *connector, bool force)
+{
+	enum drm_connector_status status;
+
+	DRM_DEBUG_KMS("[CONNECTOR:%d:%s]\n",
+		      connector->base.id, connector->name);
+
+	intel_hdmi_unset_edid(connector);
+
+	if (intel_hdmi_set_edid(connector)) {
+		struct intel_hdmi *intel_hdmi = intel_attached_hdmi(connector);
+
+		hdmi_to_dig_port(intel_hdmi)->base.type = INTEL_OUTPUT_HDMI;
+		status = connector_status_connected;
+	} else
+		status = connector_status_disconnected;
 
 	return status;
 }
 
-static int intel_hdmi_get_modes(struct drm_connector *connector)
+static void
+intel_hdmi_force(struct drm_connector *connector)
 {
-	struct intel_encoder *intel_encoder = intel_attached_encoder(connector);
-	struct intel_hdmi *intel_hdmi = enc_to_intel_hdmi(&intel_encoder->base);
-	struct drm_i915_private *dev_priv = connector->dev->dev_private;
-	enum intel_display_power_domain power_domain;
-	int ret;
+	struct intel_hdmi *intel_hdmi = intel_attached_hdmi(connector);
 
-	/* We should parse the EDID data and find out if it's an HDMI sink so
-	 * we can send audio to it.
-	 */
+	DRM_DEBUG_KMS("[CONNECTOR:%d:%s]\n",
+		      connector->base.id, connector->name);
 
-	power_domain = intel_display_port_power_domain(intel_encoder);
-	intel_display_power_get(dev_priv, power_domain);
+	intel_hdmi_unset_edid(connector);
 
-	ret = intel_ddc_get_modes(connector,
-				   intel_gmbus_get_adapter(dev_priv,
-							   intel_hdmi->ddc_bus));
+	if (connector->status != connector_status_connected)
+		return;
 
-	intel_display_power_put(dev_priv, power_domain);
+	intel_hdmi_set_edid(connector);
+	hdmi_to_dig_port(intel_hdmi)->base.type = INTEL_OUTPUT_HDMI;
+}
+
+static int intel_hdmi_get_modes(struct drm_connector *connector)
+{
+	struct edid *edid;
+
+	edid = to_intel_connector(connector)->detect_edid;
+	if (edid == NULL)
+		return 0;
 
-	return ret;
+	return intel_connector_update_modes(connector, edid);
 }
 
 static bool
 intel_hdmi_detect_audio(struct drm_connector *connector)
 {
-	struct intel_encoder *intel_encoder = intel_attached_encoder(connector);
-	struct intel_hdmi *intel_hdmi = enc_to_intel_hdmi(&intel_encoder->base);
-	struct drm_i915_private *dev_priv = connector->dev->dev_private;
-	enum intel_display_power_domain power_domain;
-	struct edid *edid;
 	bool has_audio = false;
+	struct edid *edid;
 
-	power_domain = intel_display_port_power_domain(intel_encoder);
-	intel_display_power_get(dev_priv, power_domain);
-
-	edid = drm_get_edid(connector,
-			    intel_gmbus_get_adapter(dev_priv,
-						    intel_hdmi->ddc_bus));
-	if (edid) {
-		if (edid->input & DRM_EDID_INPUT_DIGITAL)
-			has_audio = drm_detect_monitor_audio(edid);
-		kfree(edid);
-	}
-
-	intel_display_power_put(dev_priv, power_domain);
+	edid = to_intel_connector(connector)->detect_edid;
+	if (edid && edid->input & DRM_EDID_INPUT_DIGITAL)
+		has_audio = drm_detect_monitor_audio(edid);
 
 	return has_audio;
 }
@@ -1265,6 +1354,8 @@ static void chv_hdmi_pre_pll_enable(struct intel_encoder *encoder)
 	enum pipe pipe = intel_crtc->pipe;
 	u32 val;
 
+	intel_hdmi_prepare(encoder);
+
 	mutex_lock(&dev_priv->dpio_lock);
 
 	/* program left/right clock distribution */
@@ -1370,10 +1461,13 @@ static void chv_hdmi_post_disable(struct intel_encoder *encoder)
 static void chv_hdmi_pre_enable(struct intel_encoder *encoder)
 {
 	struct intel_digital_port *dport = enc_to_dig_port(&encoder->base);
+	struct intel_hdmi *intel_hdmi = &dport->hdmi;
 	struct drm_device *dev = encoder->base.dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_crtc *intel_crtc =
 		to_intel_crtc(encoder->base.crtc);
+	struct drm_display_mode *adjusted_mode =
+		&intel_crtc->config.adjusted_mode;
 	enum dpio_channel ch = vlv_dport_to_channel(dport);
 	int pipe = intel_crtc->pipe;
 	int data, i;
@@ -1381,6 +1475,15 @@ static void chv_hdmi_pre_enable(struct intel_encoder *encoder)
 
 	mutex_lock(&dev_priv->dpio_lock);
 
+	/* allow hardware to manage TX FIFO reset source */
+	val = vlv_dpio_read(dev_priv, pipe, VLV_PCS01_DW11(ch));
+	val &= ~DPIO_LANEDESKEW_STRAP_OVRD;
+	vlv_dpio_write(dev_priv, pipe, VLV_PCS01_DW11(ch), val);
+
+	val = vlv_dpio_read(dev_priv, pipe, VLV_PCS23_DW11(ch));
+	val &= ~DPIO_LANEDESKEW_STRAP_OVRD;
+	vlv_dpio_write(dev_priv, pipe, VLV_PCS23_DW11(ch), val);
+
 	/* Deassert soft data lane reset*/
 	val = vlv_dpio_read(dev_priv, pipe, VLV_PCS01_DW1(ch));
 	val |= CHV_PCS_REQ_SOFTRESET_EN;
@@ -1417,12 +1520,26 @@ static void chv_hdmi_pre_enable(struct intel_encoder *encoder)
 	/* Clear calc init */
 	val = vlv_dpio_read(dev_priv, pipe, VLV_PCS01_DW10(ch));
 	val &= ~(DPIO_PCS_SWING_CALC_TX0_TX2 | DPIO_PCS_SWING_CALC_TX1_TX3);
+	val &= ~(DPIO_PCS_TX1DEEMP_MASK | DPIO_PCS_TX2DEEMP_MASK);
+	val |= DPIO_PCS_TX1DEEMP_9P5 | DPIO_PCS_TX2DEEMP_9P5;
 	vlv_dpio_write(dev_priv, pipe, VLV_PCS01_DW10(ch), val);
 
 	val = vlv_dpio_read(dev_priv, pipe, VLV_PCS23_DW10(ch));
 	val &= ~(DPIO_PCS_SWING_CALC_TX0_TX2 | DPIO_PCS_SWING_CALC_TX1_TX3);
+	val &= ~(DPIO_PCS_TX1DEEMP_MASK | DPIO_PCS_TX2DEEMP_MASK);
+	val |= DPIO_PCS_TX1DEEMP_9P5 | DPIO_PCS_TX2DEEMP_9P5;
 	vlv_dpio_write(dev_priv, pipe, VLV_PCS23_DW10(ch), val);
 
+	val = vlv_dpio_read(dev_priv, pipe, VLV_PCS01_DW9(ch));
+	val &= ~(DPIO_PCS_TX1MARGIN_MASK | DPIO_PCS_TX2MARGIN_MASK);
+	val |= DPIO_PCS_TX1MARGIN_000 | DPIO_PCS_TX2MARGIN_000;
+	vlv_dpio_write(dev_priv, pipe, VLV_PCS01_DW9(ch), val);
+
+	val = vlv_dpio_read(dev_priv, pipe, VLV_PCS23_DW9(ch));
+	val &= ~(DPIO_PCS_TX1MARGIN_MASK | DPIO_PCS_TX2MARGIN_MASK);
+	val |= DPIO_PCS_TX1MARGIN_000 | DPIO_PCS_TX2MARGIN_000;
+	vlv_dpio_write(dev_priv, pipe, VLV_PCS23_DW9(ch), val);
+
 	/* FIXME: Program the support xxx V-dB */
 	/* Use 800mV-0dB */
 	for (i = 0; i < 4; i++) {
@@ -1434,8 +1551,8 @@ static void chv_hdmi_pre_enable(struct intel_encoder *encoder)
 
 	for (i = 0; i < 4; i++) {
 		val = vlv_dpio_read(dev_priv, pipe, CHV_TX_DW2(ch, i));
-		val &= ~DPIO_SWING_MARGIN_MASK;
-		val |= 102 << DPIO_SWING_MARGIN_SHIFT;
+		val &= ~DPIO_SWING_MARGIN000_MASK;
+		val |= 102 << DPIO_SWING_MARGIN000_SHIFT;
 		vlv_dpio_write(dev_priv, pipe, CHV_TX_DW2(ch, i), val);
 	}
 
@@ -1475,6 +1592,10 @@ static void chv_hdmi_pre_enable(struct intel_encoder *encoder)
 
 	mutex_unlock(&dev_priv->dpio_lock);
 
+	intel_hdmi->set_infoframes(&encoder->base,
+				   intel_crtc->config.has_hdmi_sink,
+				   adjusted_mode);
+
 	intel_enable_hdmi(encoder);
 
 	vlv_wait_port_ready(dev_priv, dport);
@@ -1482,6 +1603,7 @@ static void chv_hdmi_pre_enable(struct intel_encoder *encoder)
 
 static void intel_hdmi_destroy(struct drm_connector *connector)
 {
+	kfree(to_intel_connector(connector)->detect_edid);
 	drm_connector_cleanup(connector);
 	kfree(connector);
 }
@@ -1489,6 +1611,7 @@ static void intel_hdmi_destroy(struct drm_connector *connector)
 static const struct drm_connector_funcs intel_hdmi_connector_funcs = {
 	.dpms = intel_connector_dpms,
 	.detect = intel_hdmi_detect,
+	.force = intel_hdmi_force,
 	.fill_modes = drm_helper_probe_single_connector_modes,
 	.set_property = intel_hdmi_set_property,
 	.destroy = intel_hdmi_destroy,
@@ -1567,18 +1690,23 @@ void intel_hdmi_init_connector(struct intel_digital_port *intel_dig_port,
 	if (IS_VALLEYVIEW(dev)) {
 		intel_hdmi->write_infoframe = vlv_write_infoframe;
 		intel_hdmi->set_infoframes = vlv_set_infoframes;
+		intel_hdmi->infoframe_enabled = vlv_infoframe_enabled;
 	} else if (IS_G4X(dev)) {
 		intel_hdmi->write_infoframe = g4x_write_infoframe;
 		intel_hdmi->set_infoframes = g4x_set_infoframes;
+		intel_hdmi->infoframe_enabled = g4x_infoframe_enabled;
 	} else if (HAS_DDI(dev)) {
 		intel_hdmi->write_infoframe = hsw_write_infoframe;
 		intel_hdmi->set_infoframes = hsw_set_infoframes;
+		intel_hdmi->infoframe_enabled = hsw_infoframe_enabled;
 	} else if (HAS_PCH_IBX(dev)) {
 		intel_hdmi->write_infoframe = ibx_write_infoframe;
 		intel_hdmi->set_infoframes = ibx_set_infoframes;
+		intel_hdmi->infoframe_enabled = ibx_infoframe_enabled;
 	} else {
 		intel_hdmi->write_infoframe = cpt_write_infoframe;
 		intel_hdmi->set_infoframes = cpt_set_infoframes;
+		intel_hdmi->infoframe_enabled = cpt_infoframe_enabled;
 	}
 
 	if (HAS_DDI(dev))
diff --git a/drivers/gpu/drm/i915/intel_lrc.c b/drivers/gpu/drm/i915/intel_lrc.c
new file mode 100644
index 000000000000..e588376227ea
--- /dev/null
+++ b/drivers/gpu/drm/i915/intel_lrc.c
@@ -0,0 +1,1942 @@
+/*
+ * Copyright © 2014 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ *
+ * Authors:
+ *    Ben Widawsky <ben@bwidawsk.net>
+ *    Michel Thierry <michel.thierry@intel.com>
+ *    Thomas Daniel <thomas.daniel@intel.com>
+ *    Oscar Mateo <oscar.mateo@intel.com>
+ *
+ */
+
+/**
+ * DOC: Logical Rings, Logical Ring Contexts and Execlists
+ *
+ * Motivation:
+ * GEN8 brings an expansion of the HW contexts: "Logical Ring Contexts".
+ * These expanded contexts enable a number of new abilities, especially
+ * "Execlists" (also implemented in this file).
+ *
+ * One of the main differences with the legacy HW contexts is that logical
+ * ring contexts incorporate many more things to the context's state, like
+ * PDPs or ringbuffer control registers:
+ *
+ * The reason why PDPs are included in the context is straightforward: as
+ * PPGTTs (per-process GTTs) are actually per-context, having the PDPs
+ * contained there mean you don't need to do a ppgtt->switch_mm yourself,
+ * instead, the GPU will do it for you on the context switch.
+ *
+ * But, what about the ringbuffer control registers (head, tail, etc..)?
+ * shouldn't we just need a set of those per engine command streamer? This is
+ * where the name "Logical Rings" starts to make sense: by virtualizing the
+ * rings, the engine cs shifts to a new "ring buffer" with every context
+ * switch. When you want to submit a workload to the GPU you: A) choose your
+ * context, B) find its appropriate virtualized ring, C) write commands to it
+ * and then, finally, D) tell the GPU to switch to that context.
+ *
+ * Instead of the legacy MI_SET_CONTEXT, the way you tell the GPU to switch
+ * to a contexts is via a context execution list, ergo "Execlists".
+ *
+ * LRC implementation:
+ * Regarding the creation of contexts, we have:
+ *
+ * - One global default context.
+ * - One local default context for each opened fd.
+ * - One local extra context for each context create ioctl call.
+ *
+ * Now that ringbuffers belong per-context (and not per-engine, like before)
+ * and that contexts are uniquely tied to a given engine (and not reusable,
+ * like before) we need:
+ *
+ * - One ringbuffer per-engine inside each context.
+ * - One backing object per-engine inside each context.
+ *
+ * The global default context starts its life with these new objects fully
+ * allocated and populated. The local default context for each opened fd is
+ * more complex, because we don't know at creation time which engine is going
+ * to use them. To handle this, we have implemented a deferred creation of LR
+ * contexts:
+ *
+ * The local context starts its life as a hollow or blank holder, that only
+ * gets populated for a given engine once we receive an execbuffer. If later
+ * on we receive another execbuffer ioctl for the same context but a different
+ * engine, we allocate/populate a new ringbuffer and context backing object and
+ * so on.
+ *
+ * Finally, regarding local contexts created using the ioctl call: as they are
+ * only allowed with the render ring, we can allocate & populate them right
+ * away (no need to defer anything, at least for now).
+ *
+ * Execlists implementation:
+ * Execlists are the new method by which, on gen8+ hardware, workloads are
+ * submitted for execution (as opposed to the legacy, ringbuffer-based, method).
+ * This method works as follows:
+ *
+ * When a request is committed, its commands (the BB start and any leading or
+ * trailing commands, like the seqno breadcrumbs) are placed in the ringbuffer
+ * for the appropriate context. The tail pointer in the hardware context is not
+ * updated at this time, but instead, kept by the driver in the ringbuffer
+ * structure. A structure representing this request is added to a request queue
+ * for the appropriate engine: this structure contains a copy of the context's
+ * tail after the request was written to the ring buffer and a pointer to the
+ * context itself.
+ *
+ * If the engine's request queue was empty before the request was added, the
+ * queue is processed immediately. Otherwise the queue will be processed during
+ * a context switch interrupt. In any case, elements on the queue will get sent
+ * (in pairs) to the GPU's ExecLists Submit Port (ELSP, for short) with a
+ * globally unique 20-bits submission ID.
+ *
+ * When execution of a request completes, the GPU updates the context status
+ * buffer with a context complete event and generates a context switch interrupt.
+ * During the interrupt handling, the driver examines the events in the buffer:
+ * for each context complete event, if the announced ID matches that on the head
+ * of the request queue, then that request is retired and removed from the queue.
+ *
+ * After processing, if any requests were retired and the queue is not empty
+ * then a new execution list can be submitted. The two requests at the front of
+ * the queue are next to be submitted but since a context may not occur twice in
+ * an execution list, if subsequent requests have the same ID as the first then
+ * the two requests must be combined. This is done simply by discarding requests
+ * at the head of the queue until either only one requests is left (in which case
+ * we use a NULL second context) or the first two requests have unique IDs.
+ *
+ * By always executing the first two requests in the queue the driver ensures
+ * that the GPU is kept as busy as possible. In the case where a single context
+ * completes but a second context is still executing, the request for this second
+ * context will be at the head of the queue when we remove the first one. This
+ * request will then be resubmitted along with a new request for a different context,
+ * which will cause the hardware to continue executing the second request and queue
+ * the new request (the GPU detects the condition of a context getting preempted
+ * with the same context and optimizes the context switch flow by not doing
+ * preemption, but just sampling the new tail pointer).
+ *
+ */
+
+#include <drm/drmP.h>
+#include <drm/i915_drm.h>
+#include "i915_drv.h"
+
+#define GEN9_LR_CONTEXT_RENDER_SIZE (22 * PAGE_SIZE)
+#define GEN8_LR_CONTEXT_RENDER_SIZE (20 * PAGE_SIZE)
+#define GEN8_LR_CONTEXT_OTHER_SIZE (2 * PAGE_SIZE)
+
+#define RING_EXECLIST_QFULL		(1 << 0x2)
+#define RING_EXECLIST1_VALID		(1 << 0x3)
+#define RING_EXECLIST0_VALID		(1 << 0x4)
+#define RING_EXECLIST_ACTIVE_STATUS	(3 << 0xE)
+#define RING_EXECLIST1_ACTIVE		(1 << 0x11)
+#define RING_EXECLIST0_ACTIVE		(1 << 0x12)
+
+#define GEN8_CTX_STATUS_IDLE_ACTIVE	(1 << 0)
+#define GEN8_CTX_STATUS_PREEMPTED	(1 << 1)
+#define GEN8_CTX_STATUS_ELEMENT_SWITCH	(1 << 2)
+#define GEN8_CTX_STATUS_ACTIVE_IDLE	(1 << 3)
+#define GEN8_CTX_STATUS_COMPLETE	(1 << 4)
+#define GEN8_CTX_STATUS_LITE_RESTORE	(1 << 15)
+
+#define CTX_LRI_HEADER_0		0x01
+#define CTX_CONTEXT_CONTROL		0x02
+#define CTX_RING_HEAD			0x04
+#define CTX_RING_TAIL			0x06
+#define CTX_RING_BUFFER_START		0x08
+#define CTX_RING_BUFFER_CONTROL		0x0a
+#define CTX_BB_HEAD_U			0x0c
+#define CTX_BB_HEAD_L			0x0e
+#define CTX_BB_STATE			0x10
+#define CTX_SECOND_BB_HEAD_U		0x12
+#define CTX_SECOND_BB_HEAD_L		0x14
+#define CTX_SECOND_BB_STATE		0x16
+#define CTX_BB_PER_CTX_PTR		0x18
+#define CTX_RCS_INDIRECT_CTX		0x1a
+#define CTX_RCS_INDIRECT_CTX_OFFSET	0x1c
+#define CTX_LRI_HEADER_1		0x21
+#define CTX_CTX_TIMESTAMP		0x22
+#define CTX_PDP3_UDW			0x24
+#define CTX_PDP3_LDW			0x26
+#define CTX_PDP2_UDW			0x28
+#define CTX_PDP2_LDW			0x2a
+#define CTX_PDP1_UDW			0x2c
+#define CTX_PDP1_LDW			0x2e
+#define CTX_PDP0_UDW			0x30
+#define CTX_PDP0_LDW			0x32
+#define CTX_LRI_HEADER_2		0x41
+#define CTX_R_PWR_CLK_STATE		0x42
+#define CTX_GPGPU_CSR_BASE_ADDRESS	0x44
+
+#define GEN8_CTX_VALID (1<<0)
+#define GEN8_CTX_FORCE_PD_RESTORE (1<<1)
+#define GEN8_CTX_FORCE_RESTORE (1<<2)
+#define GEN8_CTX_L3LLC_COHERENT (1<<5)
+#define GEN8_CTX_PRIVILEGE (1<<8)
+enum {
+	ADVANCED_CONTEXT = 0,
+	LEGACY_CONTEXT,
+	ADVANCED_AD_CONTEXT,
+	LEGACY_64B_CONTEXT
+};
+#define GEN8_CTX_MODE_SHIFT 3
+enum {
+	FAULT_AND_HANG = 0,
+	FAULT_AND_HALT, /* Debug only */
+	FAULT_AND_STREAM,
+	FAULT_AND_CONTINUE /* Unsupported */
+};
+#define GEN8_CTX_ID_SHIFT 32
+
+static int intel_lr_context_pin(struct intel_engine_cs *ring,
+		struct intel_context *ctx);
+
+/**
+ * intel_sanitize_enable_execlists() - sanitize i915.enable_execlists
+ * @dev: DRM device.
+ * @enable_execlists: value of i915.enable_execlists module parameter.
+ *
+ * Only certain platforms support Execlists (the prerequisites being
+ * support for Logical Ring Contexts and Aliasing PPGTT or better),
+ * and only when enabled via module parameter.
+ *
+ * Return: 1 if Execlists is supported and has to be enabled.
+ */
+int intel_sanitize_enable_execlists(struct drm_device *dev, int enable_execlists)
+{
+	WARN_ON(i915.enable_ppgtt == -1);
+
+	if (INTEL_INFO(dev)->gen >= 9)
+		return 1;
+
+	if (enable_execlists == 0)
+		return 0;
+
+	if (HAS_LOGICAL_RING_CONTEXTS(dev) && USES_PPGTT(dev) &&
+	    i915.use_mmio_flip >= 0)
+		return 1;
+
+	return 0;
+}
+
+/**
+ * intel_execlists_ctx_id() - get the Execlists Context ID
+ * @ctx_obj: Logical Ring Context backing object.
+ *
+ * Do not confuse with ctx->id! Unfortunately we have a name overload
+ * here: the old context ID we pass to userspace as a handler so that
+ * they can refer to a context, and the new context ID we pass to the
+ * ELSP so that the GPU can inform us of the context status via
+ * interrupts.
+ *
+ * Return: 20-bits globally unique context ID.
+ */
+u32 intel_execlists_ctx_id(struct drm_i915_gem_object *ctx_obj)
+{
+	u32 lrca = i915_gem_obj_ggtt_offset(ctx_obj);
+
+	/* LRCA is required to be 4K aligned so the more significant 20 bits
+	 * are globally unique */
+	return lrca >> 12;
+}
+
+static uint64_t execlists_ctx_descriptor(struct drm_i915_gem_object *ctx_obj)
+{
+	uint64_t desc;
+	uint64_t lrca = i915_gem_obj_ggtt_offset(ctx_obj);
+
+	WARN_ON(lrca & 0xFFFFFFFF00000FFFULL);
+
+	desc = GEN8_CTX_VALID;
+	desc |= LEGACY_CONTEXT << GEN8_CTX_MODE_SHIFT;
+	desc |= GEN8_CTX_L3LLC_COHERENT;
+	desc |= GEN8_CTX_PRIVILEGE;
+	desc |= lrca;
+	desc |= (u64)intel_execlists_ctx_id(ctx_obj) << GEN8_CTX_ID_SHIFT;
+
+	/* TODO: WaDisableLiteRestore when we start using semaphore
+	 * signalling between Command Streamers */
+	/* desc |= GEN8_CTX_FORCE_RESTORE; */
+
+	return desc;
+}
+
+static void execlists_elsp_write(struct intel_engine_cs *ring,
+				 struct drm_i915_gem_object *ctx_obj0,
+				 struct drm_i915_gem_object *ctx_obj1)
+{
+	struct drm_device *dev = ring->dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	uint64_t temp = 0;
+	uint32_t desc[4];
+	unsigned long flags;
+
+	/* XXX: You must always write both descriptors in the order below. */
+	if (ctx_obj1)
+		temp = execlists_ctx_descriptor(ctx_obj1);
+	else
+		temp = 0;
+	desc[1] = (u32)(temp >> 32);
+	desc[0] = (u32)temp;
+
+	temp = execlists_ctx_descriptor(ctx_obj0);
+	desc[3] = (u32)(temp >> 32);
+	desc[2] = (u32)temp;
+
+	/* Set Force Wakeup bit to prevent GT from entering C6 while ELSP writes
+	 * are in progress.
+	 *
+	 * The other problem is that we can't just call gen6_gt_force_wake_get()
+	 * because that function calls intel_runtime_pm_get(), which might sleep.
+	 * Instead, we do the runtime_pm_get/put when creating/destroying requests.
+	 */
+	spin_lock_irqsave(&dev_priv->uncore.lock, flags);
+	if (IS_CHERRYVIEW(dev) || INTEL_INFO(dev)->gen >= 9) {
+		if (dev_priv->uncore.fw_rendercount++ == 0)
+			dev_priv->uncore.funcs.force_wake_get(dev_priv,
+							      FORCEWAKE_RENDER);
+		if (dev_priv->uncore.fw_mediacount++ == 0)
+			dev_priv->uncore.funcs.force_wake_get(dev_priv,
+							      FORCEWAKE_MEDIA);
+		if (INTEL_INFO(dev)->gen >= 9) {
+			if (dev_priv->uncore.fw_blittercount++ == 0)
+				dev_priv->uncore.funcs.force_wake_get(dev_priv,
+							FORCEWAKE_BLITTER);
+		}
+	} else {
+		if (dev_priv->uncore.forcewake_count++ == 0)
+			dev_priv->uncore.funcs.force_wake_get(dev_priv,
+							      FORCEWAKE_ALL);
+	}
+	spin_unlock_irqrestore(&dev_priv->uncore.lock, flags);
+
+	I915_WRITE(RING_ELSP(ring), desc[1]);
+	I915_WRITE(RING_ELSP(ring), desc[0]);
+	I915_WRITE(RING_ELSP(ring), desc[3]);
+	/* The context is automatically loaded after the following */
+	I915_WRITE(RING_ELSP(ring), desc[2]);
+
+	/* ELSP is a wo register, so use another nearby reg for posting instead */
+	POSTING_READ(RING_EXECLIST_STATUS(ring));
+
+	/* Release Force Wakeup (see the big comment above). */
+	spin_lock_irqsave(&dev_priv->uncore.lock, flags);
+	if (IS_CHERRYVIEW(dev) || INTEL_INFO(dev)->gen >= 9) {
+		if (--dev_priv->uncore.fw_rendercount == 0)
+			dev_priv->uncore.funcs.force_wake_put(dev_priv,
+							      FORCEWAKE_RENDER);
+		if (--dev_priv->uncore.fw_mediacount == 0)
+			dev_priv->uncore.funcs.force_wake_put(dev_priv,
+							      FORCEWAKE_MEDIA);
+		if (INTEL_INFO(dev)->gen >= 9) {
+			if (--dev_priv->uncore.fw_blittercount == 0)
+				dev_priv->uncore.funcs.force_wake_put(dev_priv,
+							FORCEWAKE_BLITTER);
+		}
+	} else {
+		if (--dev_priv->uncore.forcewake_count == 0)
+			dev_priv->uncore.funcs.force_wake_put(dev_priv,
+							      FORCEWAKE_ALL);
+	}
+
+	spin_unlock_irqrestore(&dev_priv->uncore.lock, flags);
+}
+
+static int execlists_update_context(struct drm_i915_gem_object *ctx_obj,
+				    struct drm_i915_gem_object *ring_obj,
+				    u32 tail)
+{
+	struct page *page;
+	uint32_t *reg_state;
+
+	page = i915_gem_object_get_page(ctx_obj, 1);
+	reg_state = kmap_atomic(page);
+
+	reg_state[CTX_RING_TAIL+1] = tail;
+	reg_state[CTX_RING_BUFFER_START+1] = i915_gem_obj_ggtt_offset(ring_obj);
+
+	kunmap_atomic(reg_state);
+
+	return 0;
+}
+
+static void execlists_submit_contexts(struct intel_engine_cs *ring,
+				      struct intel_context *to0, u32 tail0,
+				      struct intel_context *to1, u32 tail1)
+{
+	struct drm_i915_gem_object *ctx_obj0 = to0->engine[ring->id].state;
+	struct intel_ringbuffer *ringbuf0 = to0->engine[ring->id].ringbuf;
+	struct drm_i915_gem_object *ctx_obj1 = NULL;
+	struct intel_ringbuffer *ringbuf1 = NULL;
+
+	BUG_ON(!ctx_obj0);
+	WARN_ON(!i915_gem_obj_is_pinned(ctx_obj0));
+	WARN_ON(!i915_gem_obj_is_pinned(ringbuf0->obj));
+
+	execlists_update_context(ctx_obj0, ringbuf0->obj, tail0);
+
+	if (to1) {
+		ringbuf1 = to1->engine[ring->id].ringbuf;
+		ctx_obj1 = to1->engine[ring->id].state;
+		BUG_ON(!ctx_obj1);
+		WARN_ON(!i915_gem_obj_is_pinned(ctx_obj1));
+		WARN_ON(!i915_gem_obj_is_pinned(ringbuf1->obj));
+
+		execlists_update_context(ctx_obj1, ringbuf1->obj, tail1);
+	}
+
+	execlists_elsp_write(ring, ctx_obj0, ctx_obj1);
+}
+
+static void execlists_context_unqueue(struct intel_engine_cs *ring)
+{
+	struct intel_ctx_submit_request *req0 = NULL, *req1 = NULL;
+	struct intel_ctx_submit_request *cursor = NULL, *tmp = NULL;
+
+	assert_spin_locked(&ring->execlist_lock);
+
+	if (list_empty(&ring->execlist_queue))
+		return;
+
+	/* Try to read in pairs */
+	list_for_each_entry_safe(cursor, tmp, &ring->execlist_queue,
+				 execlist_link) {
+		if (!req0) {
+			req0 = cursor;
+		} else if (req0->ctx == cursor->ctx) {
+			/* Same ctx: ignore first request, as second request
+			 * will update tail past first request's workload */
+			cursor->elsp_submitted = req0->elsp_submitted;
+			list_del(&req0->execlist_link);
+			list_add_tail(&req0->execlist_link,
+				&ring->execlist_retired_req_list);
+			req0 = cursor;
+		} else {
+			req1 = cursor;
+			break;
+		}
+	}
+
+	WARN_ON(req1 && req1->elsp_submitted);
+
+	execlists_submit_contexts(ring, req0->ctx, req0->tail,
+				  req1 ? req1->ctx : NULL,
+				  req1 ? req1->tail : 0);
+
+	req0->elsp_submitted++;
+	if (req1)
+		req1->elsp_submitted++;
+}
+
+static bool execlists_check_remove_request(struct intel_engine_cs *ring,
+					   u32 request_id)
+{
+	struct intel_ctx_submit_request *head_req;
+
+	assert_spin_locked(&ring->execlist_lock);
+
+	head_req = list_first_entry_or_null(&ring->execlist_queue,
+					    struct intel_ctx_submit_request,
+					    execlist_link);
+
+	if (head_req != NULL) {
+		struct drm_i915_gem_object *ctx_obj =
+				head_req->ctx->engine[ring->id].state;
+		if (intel_execlists_ctx_id(ctx_obj) == request_id) {
+			WARN(head_req->elsp_submitted == 0,
+			     "Never submitted head request\n");
+
+			if (--head_req->elsp_submitted <= 0) {
+				list_del(&head_req->execlist_link);
+				list_add_tail(&head_req->execlist_link,
+					&ring->execlist_retired_req_list);
+				return true;
+			}
+		}
+	}
+
+	return false;
+}
+
+/**
+ * intel_execlists_handle_ctx_events() - handle Context Switch interrupts
+ * @ring: Engine Command Streamer to handle.
+ *
+ * Check the unread Context Status Buffers and manage the submission of new
+ * contexts to the ELSP accordingly.
+ */
+void intel_execlists_handle_ctx_events(struct intel_engine_cs *ring)
+{
+	struct drm_i915_private *dev_priv = ring->dev->dev_private;
+	u32 status_pointer;
+	u8 read_pointer;
+	u8 write_pointer;
+	u32 status;
+	u32 status_id;
+	u32 submit_contexts = 0;
+
+	status_pointer = I915_READ(RING_CONTEXT_STATUS_PTR(ring));
+
+	read_pointer = ring->next_context_status_buffer;
+	write_pointer = status_pointer & 0x07;
+	if (read_pointer > write_pointer)
+		write_pointer += 6;
+
+	spin_lock(&ring->execlist_lock);
+
+	while (read_pointer < write_pointer) {
+		read_pointer++;
+		status = I915_READ(RING_CONTEXT_STATUS_BUF(ring) +
+				(read_pointer % 6) * 8);
+		status_id = I915_READ(RING_CONTEXT_STATUS_BUF(ring) +
+				(read_pointer % 6) * 8 + 4);
+
+		if (status & GEN8_CTX_STATUS_PREEMPTED) {
+			if (status & GEN8_CTX_STATUS_LITE_RESTORE) {
+				if (execlists_check_remove_request(ring, status_id))
+					WARN(1, "Lite Restored request removed from queue\n");
+			} else
+				WARN(1, "Preemption without Lite Restore\n");
+		}
+
+		 if ((status & GEN8_CTX_STATUS_ACTIVE_IDLE) ||
+		     (status & GEN8_CTX_STATUS_ELEMENT_SWITCH)) {
+			if (execlists_check_remove_request(ring, status_id))
+				submit_contexts++;
+		}
+	}
+
+	if (submit_contexts != 0)
+		execlists_context_unqueue(ring);
+
+	spin_unlock(&ring->execlist_lock);
+
+	WARN(submit_contexts > 2, "More than two context complete events?\n");
+	ring->next_context_status_buffer = write_pointer % 6;
+
+	I915_WRITE(RING_CONTEXT_STATUS_PTR(ring),
+		   ((u32)ring->next_context_status_buffer & 0x07) << 8);
+}
+
+static int execlists_context_queue(struct intel_engine_cs *ring,
+				   struct intel_context *to,
+				   u32 tail)
+{
+	struct intel_ctx_submit_request *req = NULL, *cursor;
+	struct drm_i915_private *dev_priv = ring->dev->dev_private;
+	unsigned long flags;
+	int num_elements = 0;
+
+	req = kzalloc(sizeof(*req), GFP_KERNEL);
+	if (req == NULL)
+		return -ENOMEM;
+	req->ctx = to;
+	i915_gem_context_reference(req->ctx);
+
+	if (to != ring->default_context)
+		intel_lr_context_pin(ring, to);
+
+	req->ring = ring;
+	req->tail = tail;
+
+	intel_runtime_pm_get(dev_priv);
+
+	spin_lock_irqsave(&ring->execlist_lock, flags);
+
+	list_for_each_entry(cursor, &ring->execlist_queue, execlist_link)
+		if (++num_elements > 2)
+			break;
+
+	if (num_elements > 2) {
+		struct intel_ctx_submit_request *tail_req;
+
+		tail_req = list_last_entry(&ring->execlist_queue,
+					   struct intel_ctx_submit_request,
+					   execlist_link);
+
+		if (to == tail_req->ctx) {
+			WARN(tail_req->elsp_submitted != 0,
+				"More than 2 already-submitted reqs queued\n");
+			list_del(&tail_req->execlist_link);
+			list_add_tail(&tail_req->execlist_link,
+				&ring->execlist_retired_req_list);
+		}
+	}
+
+	list_add_tail(&req->execlist_link, &ring->execlist_queue);
+	if (num_elements == 0)
+		execlists_context_unqueue(ring);
+
+	spin_unlock_irqrestore(&ring->execlist_lock, flags);
+
+	return 0;
+}
+
+static int logical_ring_invalidate_all_caches(struct intel_ringbuffer *ringbuf)
+{
+	struct intel_engine_cs *ring = ringbuf->ring;
+	uint32_t flush_domains;
+	int ret;
+
+	flush_domains = 0;
+	if (ring->gpu_caches_dirty)
+		flush_domains = I915_GEM_GPU_DOMAINS;
+
+	ret = ring->emit_flush(ringbuf, I915_GEM_GPU_DOMAINS, flush_domains);
+	if (ret)
+		return ret;
+
+	ring->gpu_caches_dirty = false;
+	return 0;
+}
+
+static int execlists_move_to_gpu(struct intel_ringbuffer *ringbuf,
+				 struct list_head *vmas)
+{
+	struct intel_engine_cs *ring = ringbuf->ring;
+	struct i915_vma *vma;
+	uint32_t flush_domains = 0;
+	bool flush_chipset = false;
+	int ret;
+
+	list_for_each_entry(vma, vmas, exec_list) {
+		struct drm_i915_gem_object *obj = vma->obj;
+
+		ret = i915_gem_object_sync(obj, ring);
+		if (ret)
+			return ret;
+
+		if (obj->base.write_domain & I915_GEM_DOMAIN_CPU)
+			flush_chipset |= i915_gem_clflush_object(obj, false);
+
+		flush_domains |= obj->base.write_domain;
+	}
+
+	if (flush_domains & I915_GEM_DOMAIN_GTT)
+		wmb();
+
+	/* Unconditionally invalidate gpu caches and ensure that we do flush
+	 * any residual writes from the previous batch.
+	 */
+	return logical_ring_invalidate_all_caches(ringbuf);
+}
+
+/**
+ * execlists_submission() - submit a batchbuffer for execution, Execlists style
+ * @dev: DRM device.
+ * @file: DRM file.
+ * @ring: Engine Command Streamer to submit to.
+ * @ctx: Context to employ for this submission.
+ * @args: execbuffer call arguments.
+ * @vmas: list of vmas.
+ * @batch_obj: the batchbuffer to submit.
+ * @exec_start: batchbuffer start virtual address pointer.
+ * @flags: translated execbuffer call flags.
+ *
+ * This is the evil twin version of i915_gem_ringbuffer_submission. It abstracts
+ * away the submission details of the execbuffer ioctl call.
+ *
+ * Return: non-zero if the submission fails.
+ */
+int intel_execlists_submission(struct drm_device *dev, struct drm_file *file,
+			       struct intel_engine_cs *ring,
+			       struct intel_context *ctx,
+			       struct drm_i915_gem_execbuffer2 *args,
+			       struct list_head *vmas,
+			       struct drm_i915_gem_object *batch_obj,
+			       u64 exec_start, u32 flags)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct intel_ringbuffer *ringbuf = ctx->engine[ring->id].ringbuf;
+	int instp_mode;
+	u32 instp_mask;
+	int ret;
+
+	instp_mode = args->flags & I915_EXEC_CONSTANTS_MASK;
+	instp_mask = I915_EXEC_CONSTANTS_MASK;
+	switch (instp_mode) {
+	case I915_EXEC_CONSTANTS_REL_GENERAL:
+	case I915_EXEC_CONSTANTS_ABSOLUTE:
+	case I915_EXEC_CONSTANTS_REL_SURFACE:
+		if (instp_mode != 0 && ring != &dev_priv->ring[RCS]) {
+			DRM_DEBUG("non-0 rel constants mode on non-RCS\n");
+			return -EINVAL;
+		}
+
+		if (instp_mode != dev_priv->relative_constants_mode) {
+			if (instp_mode == I915_EXEC_CONSTANTS_REL_SURFACE) {
+				DRM_DEBUG("rel surface constants mode invalid on gen5+\n");
+				return -EINVAL;
+			}
+
+			/* The HW changed the meaning on this bit on gen6 */
+			instp_mask &= ~I915_EXEC_CONSTANTS_REL_SURFACE;
+		}
+		break;
+	default:
+		DRM_DEBUG("execbuf with unknown constants: %d\n", instp_mode);
+		return -EINVAL;
+	}
+
+	if (args->num_cliprects != 0) {
+		DRM_DEBUG("clip rectangles are only valid on pre-gen5\n");
+		return -EINVAL;
+	} else {
+		if (args->DR4 == 0xffffffff) {
+			DRM_DEBUG("UXA submitting garbage DR4, fixing up\n");
+			args->DR4 = 0;
+		}
+
+		if (args->DR1 || args->DR4 || args->cliprects_ptr) {
+			DRM_DEBUG("0 cliprects but dirt in cliprects fields\n");
+			return -EINVAL;
+		}
+	}
+
+	if (args->flags & I915_EXEC_GEN7_SOL_RESET) {
+		DRM_DEBUG("sol reset is gen7 only\n");
+		return -EINVAL;
+	}
+
+	ret = execlists_move_to_gpu(ringbuf, vmas);
+	if (ret)
+		return ret;
+
+	if (ring == &dev_priv->ring[RCS] &&
+	    instp_mode != dev_priv->relative_constants_mode) {
+		ret = intel_logical_ring_begin(ringbuf, 4);
+		if (ret)
+			return ret;
+
+		intel_logical_ring_emit(ringbuf, MI_NOOP);
+		intel_logical_ring_emit(ringbuf, MI_LOAD_REGISTER_IMM(1));
+		intel_logical_ring_emit(ringbuf, INSTPM);
+		intel_logical_ring_emit(ringbuf, instp_mask << 16 | instp_mode);
+		intel_logical_ring_advance(ringbuf);
+
+		dev_priv->relative_constants_mode = instp_mode;
+	}
+
+	ret = ring->emit_bb_start(ringbuf, exec_start, flags);
+	if (ret)
+		return ret;
+
+	i915_gem_execbuffer_move_to_active(vmas, ring);
+	i915_gem_execbuffer_retire_commands(dev, file, ring, batch_obj);
+
+	return 0;
+}
+
+void intel_execlists_retire_requests(struct intel_engine_cs *ring)
+{
+	struct intel_ctx_submit_request *req, *tmp;
+	struct drm_i915_private *dev_priv = ring->dev->dev_private;
+	unsigned long flags;
+	struct list_head retired_list;
+
+	WARN_ON(!mutex_is_locked(&ring->dev->struct_mutex));
+	if (list_empty(&ring->execlist_retired_req_list))
+		return;
+
+	INIT_LIST_HEAD(&retired_list);
+	spin_lock_irqsave(&ring->execlist_lock, flags);
+	list_replace_init(&ring->execlist_retired_req_list, &retired_list);
+	spin_unlock_irqrestore(&ring->execlist_lock, flags);
+
+	list_for_each_entry_safe(req, tmp, &retired_list, execlist_link) {
+		struct intel_context *ctx = req->ctx;
+		struct drm_i915_gem_object *ctx_obj =
+				ctx->engine[ring->id].state;
+
+		if (ctx_obj && (ctx != ring->default_context))
+			intel_lr_context_unpin(ring, ctx);
+		intel_runtime_pm_put(dev_priv);
+		i915_gem_context_unreference(req->ctx);
+		list_del(&req->execlist_link);
+		kfree(req);
+	}
+}
+
+void intel_logical_ring_stop(struct intel_engine_cs *ring)
+{
+	struct drm_i915_private *dev_priv = ring->dev->dev_private;
+	int ret;
+
+	if (!intel_ring_initialized(ring))
+		return;
+
+	ret = intel_ring_idle(ring);
+	if (ret && !i915_reset_in_progress(&to_i915(ring->dev)->gpu_error))
+		DRM_ERROR("failed to quiesce %s whilst cleaning up: %d\n",
+			  ring->name, ret);
+
+	/* TODO: Is this correct with Execlists enabled? */
+	I915_WRITE_MODE(ring, _MASKED_BIT_ENABLE(STOP_RING));
+	if (wait_for_atomic((I915_READ_MODE(ring) & MODE_IDLE) != 0, 1000)) {
+		DRM_ERROR("%s :timed out trying to stop ring\n", ring->name);
+		return;
+	}
+	I915_WRITE_MODE(ring, _MASKED_BIT_DISABLE(STOP_RING));
+}
+
+int logical_ring_flush_all_caches(struct intel_ringbuffer *ringbuf)
+{
+	struct intel_engine_cs *ring = ringbuf->ring;
+	int ret;
+
+	if (!ring->gpu_caches_dirty)
+		return 0;
+
+	ret = ring->emit_flush(ringbuf, 0, I915_GEM_GPU_DOMAINS);
+	if (ret)
+		return ret;
+
+	ring->gpu_caches_dirty = false;
+	return 0;
+}
+
+/**
+ * intel_logical_ring_advance_and_submit() - advance the tail and submit the workload
+ * @ringbuf: Logical Ringbuffer to advance.
+ *
+ * The tail is updated in our logical ringbuffer struct, not in the actual context. What
+ * really happens during submission is that the context and current tail will be placed
+ * on a queue waiting for the ELSP to be ready to accept a new context submission. At that
+ * point, the tail *inside* the context is updated and the ELSP written to.
+ */
+void intel_logical_ring_advance_and_submit(struct intel_ringbuffer *ringbuf)
+{
+	struct intel_engine_cs *ring = ringbuf->ring;
+	struct intel_context *ctx = ringbuf->FIXME_lrc_ctx;
+
+	intel_logical_ring_advance(ringbuf);
+
+	if (intel_ring_stopped(ring))
+		return;
+
+	execlists_context_queue(ring, ctx, ringbuf->tail);
+}
+
+static int intel_lr_context_pin(struct intel_engine_cs *ring,
+		struct intel_context *ctx)
+{
+	struct drm_i915_gem_object *ctx_obj = ctx->engine[ring->id].state;
+	struct intel_ringbuffer *ringbuf = ctx->engine[ring->id].ringbuf;
+	int ret = 0;
+
+	WARN_ON(!mutex_is_locked(&ring->dev->struct_mutex));
+	if (ctx->engine[ring->id].unpin_count++ == 0) {
+		ret = i915_gem_obj_ggtt_pin(ctx_obj,
+				GEN8_LR_CONTEXT_ALIGN, 0);
+		if (ret)
+			goto reset_unpin_count;
+
+		ret = intel_pin_and_map_ringbuffer_obj(ring->dev, ringbuf);
+		if (ret)
+			goto unpin_ctx_obj;
+	}
+
+	return ret;
+
+unpin_ctx_obj:
+	i915_gem_object_ggtt_unpin(ctx_obj);
+reset_unpin_count:
+	ctx->engine[ring->id].unpin_count = 0;
+
+	return ret;
+}
+
+void intel_lr_context_unpin(struct intel_engine_cs *ring,
+		struct intel_context *ctx)
+{
+	struct drm_i915_gem_object *ctx_obj = ctx->engine[ring->id].state;
+	struct intel_ringbuffer *ringbuf = ctx->engine[ring->id].ringbuf;
+
+	if (ctx_obj) {
+		WARN_ON(!mutex_is_locked(&ring->dev->struct_mutex));
+		if (--ctx->engine[ring->id].unpin_count == 0) {
+			intel_unpin_ringbuffer_obj(ringbuf);
+			i915_gem_object_ggtt_unpin(ctx_obj);
+		}
+	}
+}
+
+static int logical_ring_alloc_seqno(struct intel_engine_cs *ring,
+				    struct intel_context *ctx)
+{
+	int ret;
+
+	if (ring->outstanding_lazy_seqno)
+		return 0;
+
+	if (ring->preallocated_lazy_request == NULL) {
+		struct drm_i915_gem_request *request;
+
+		request = kmalloc(sizeof(*request), GFP_KERNEL);
+		if (request == NULL)
+			return -ENOMEM;
+
+		if (ctx != ring->default_context) {
+			ret = intel_lr_context_pin(ring, ctx);
+			if (ret) {
+				kfree(request);
+				return ret;
+			}
+		}
+
+		/* Hold a reference to the context this request belongs to
+		 * (we will need it when the time comes to emit/retire the
+		 * request).
+		 */
+		request->ctx = ctx;
+		i915_gem_context_reference(request->ctx);
+
+		ring->preallocated_lazy_request = request;
+	}
+
+	return i915_gem_get_seqno(ring->dev, &ring->outstanding_lazy_seqno);
+}
+
+static int logical_ring_wait_request(struct intel_ringbuffer *ringbuf,
+				     int bytes)
+{
+	struct intel_engine_cs *ring = ringbuf->ring;
+	struct drm_i915_gem_request *request;
+	u32 seqno = 0;
+	int ret;
+
+	if (ringbuf->last_retired_head != -1) {
+		ringbuf->head = ringbuf->last_retired_head;
+		ringbuf->last_retired_head = -1;
+
+		ringbuf->space = intel_ring_space(ringbuf);
+		if (ringbuf->space >= bytes)
+			return 0;
+	}
+
+	list_for_each_entry(request, &ring->request_list, list) {
+		if (__intel_ring_space(request->tail, ringbuf->tail,
+				       ringbuf->size) >= bytes) {
+			seqno = request->seqno;
+			break;
+		}
+	}
+
+	if (seqno == 0)
+		return -ENOSPC;
+
+	ret = i915_wait_seqno(ring, seqno);
+	if (ret)
+		return ret;
+
+	i915_gem_retire_requests_ring(ring);
+	ringbuf->head = ringbuf->last_retired_head;
+	ringbuf->last_retired_head = -1;
+
+	ringbuf->space = intel_ring_space(ringbuf);
+	return 0;
+}
+
+static int logical_ring_wait_for_space(struct intel_ringbuffer *ringbuf,
+				       int bytes)
+{
+	struct intel_engine_cs *ring = ringbuf->ring;
+	struct drm_device *dev = ring->dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	unsigned long end;
+	int ret;
+
+	ret = logical_ring_wait_request(ringbuf, bytes);
+	if (ret != -ENOSPC)
+		return ret;
+
+	/* Force the context submission in case we have been skipping it */
+	intel_logical_ring_advance_and_submit(ringbuf);
+
+	/* With GEM the hangcheck timer should kick us out of the loop,
+	 * leaving it early runs the risk of corrupting GEM state (due
+	 * to running on almost untested codepaths). But on resume
+	 * timers don't work yet, so prevent a complete hang in that
+	 * case by choosing an insanely large timeout. */
+	end = jiffies + 60 * HZ;
+
+	do {
+		ringbuf->head = I915_READ_HEAD(ring);
+		ringbuf->space = intel_ring_space(ringbuf);
+		if (ringbuf->space >= bytes) {
+			ret = 0;
+			break;
+		}
+
+		msleep(1);
+
+		if (dev_priv->mm.interruptible && signal_pending(current)) {
+			ret = -ERESTARTSYS;
+			break;
+		}
+
+		ret = i915_gem_check_wedge(&dev_priv->gpu_error,
+					   dev_priv->mm.interruptible);
+		if (ret)
+			break;
+
+		if (time_after(jiffies, end)) {
+			ret = -EBUSY;
+			break;
+		}
+	} while (1);
+
+	return ret;
+}
+
+static int logical_ring_wrap_buffer(struct intel_ringbuffer *ringbuf)
+{
+	uint32_t __iomem *virt;
+	int rem = ringbuf->size - ringbuf->tail;
+
+	if (ringbuf->space < rem) {
+		int ret = logical_ring_wait_for_space(ringbuf, rem);
+
+		if (ret)
+			return ret;
+	}
+
+	virt = ringbuf->virtual_start + ringbuf->tail;
+	rem /= 4;
+	while (rem--)
+		iowrite32(MI_NOOP, virt++);
+
+	ringbuf->tail = 0;
+	ringbuf->space = intel_ring_space(ringbuf);
+
+	return 0;
+}
+
+static int logical_ring_prepare(struct intel_ringbuffer *ringbuf, int bytes)
+{
+	int ret;
+
+	if (unlikely(ringbuf->tail + bytes > ringbuf->effective_size)) {
+		ret = logical_ring_wrap_buffer(ringbuf);
+		if (unlikely(ret))
+			return ret;
+	}
+
+	if (unlikely(ringbuf->space < bytes)) {
+		ret = logical_ring_wait_for_space(ringbuf, bytes);
+		if (unlikely(ret))
+			return ret;
+	}
+
+	return 0;
+}
+
+/**
+ * intel_logical_ring_begin() - prepare the logical ringbuffer to accept some commands
+ *
+ * @ringbuf: Logical ringbuffer.
+ * @num_dwords: number of DWORDs that we plan to write to the ringbuffer.
+ *
+ * The ringbuffer might not be ready to accept the commands right away (maybe it needs to
+ * be wrapped, or wait a bit for the tail to be updated). This function takes care of that
+ * and also preallocates a request (every workload submission is still mediated through
+ * requests, same as it did with legacy ringbuffer submission).
+ *
+ * Return: non-zero if the ringbuffer is not ready to be written to.
+ */
+int intel_logical_ring_begin(struct intel_ringbuffer *ringbuf, int num_dwords)
+{
+	struct intel_engine_cs *ring = ringbuf->ring;
+	struct drm_device *dev = ring->dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	int ret;
+
+	ret = i915_gem_check_wedge(&dev_priv->gpu_error,
+				   dev_priv->mm.interruptible);
+	if (ret)
+		return ret;
+
+	ret = logical_ring_prepare(ringbuf, num_dwords * sizeof(uint32_t));
+	if (ret)
+		return ret;
+
+	/* Preallocate the olr before touching the ring */
+	ret = logical_ring_alloc_seqno(ring, ringbuf->FIXME_lrc_ctx);
+	if (ret)
+		return ret;
+
+	ringbuf->space -= num_dwords * sizeof(uint32_t);
+	return 0;
+}
+
+static int intel_logical_ring_workarounds_emit(struct intel_engine_cs *ring,
+					       struct intel_context *ctx)
+{
+	int ret, i;
+	struct intel_ringbuffer *ringbuf = ctx->engine[ring->id].ringbuf;
+	struct drm_device *dev = ring->dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct i915_workarounds *w = &dev_priv->workarounds;
+
+	if (WARN_ON(w->count == 0))
+		return 0;
+
+	ring->gpu_caches_dirty = true;
+	ret = logical_ring_flush_all_caches(ringbuf);
+	if (ret)
+		return ret;
+
+	ret = intel_logical_ring_begin(ringbuf, w->count * 2 + 2);
+	if (ret)
+		return ret;
+
+	intel_logical_ring_emit(ringbuf, MI_LOAD_REGISTER_IMM(w->count));
+	for (i = 0; i < w->count; i++) {
+		intel_logical_ring_emit(ringbuf, w->reg[i].addr);
+		intel_logical_ring_emit(ringbuf, w->reg[i].value);
+	}
+	intel_logical_ring_emit(ringbuf, MI_NOOP);
+
+	intel_logical_ring_advance(ringbuf);
+
+	ring->gpu_caches_dirty = true;
+	ret = logical_ring_flush_all_caches(ringbuf);
+	if (ret)
+		return ret;
+
+	return 0;
+}
+
+static int gen8_init_common_ring(struct intel_engine_cs *ring)
+{
+	struct drm_device *dev = ring->dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+
+	I915_WRITE_IMR(ring, ~(ring->irq_enable_mask | ring->irq_keep_mask));
+	I915_WRITE(RING_HWSTAM(ring->mmio_base), 0xffffffff);
+
+	I915_WRITE(RING_MODE_GEN7(ring),
+		   _MASKED_BIT_DISABLE(GFX_REPLAY_MODE) |
+		   _MASKED_BIT_ENABLE(GFX_RUN_LIST_ENABLE));
+	POSTING_READ(RING_MODE_GEN7(ring));
+	DRM_DEBUG_DRIVER("Execlists enabled for %s\n", ring->name);
+
+	memset(&ring->hangcheck, 0, sizeof(ring->hangcheck));
+
+	return 0;
+}
+
+static int gen8_init_render_ring(struct intel_engine_cs *ring)
+{
+	struct drm_device *dev = ring->dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	int ret;
+
+	ret = gen8_init_common_ring(ring);
+	if (ret)
+		return ret;
+
+	/* We need to disable the AsyncFlip performance optimisations in order
+	 * to use MI_WAIT_FOR_EVENT within the CS. It should already be
+	 * programmed to '1' on all products.
+	 *
+	 * WaDisableAsyncFlipPerfMode:snb,ivb,hsw,vlv,bdw,chv
+	 */
+	I915_WRITE(MI_MODE, _MASKED_BIT_ENABLE(ASYNC_FLIP_PERF_DISABLE));
+
+	ret = intel_init_pipe_control(ring);
+	if (ret)
+		return ret;
+
+	I915_WRITE(INSTPM, _MASKED_BIT_ENABLE(INSTPM_FORCE_ORDERING));
+
+	return init_workarounds_ring(ring);
+}
+
+static int gen8_emit_bb_start(struct intel_ringbuffer *ringbuf,
+			      u64 offset, unsigned flags)
+{
+	bool ppgtt = !(flags & I915_DISPATCH_SECURE);
+	int ret;
+
+	ret = intel_logical_ring_begin(ringbuf, 4);
+	if (ret)
+		return ret;
+
+	/* FIXME(BDW): Address space and security selectors. */
+	intel_logical_ring_emit(ringbuf, MI_BATCH_BUFFER_START_GEN8 | (ppgtt<<8));
+	intel_logical_ring_emit(ringbuf, lower_32_bits(offset));
+	intel_logical_ring_emit(ringbuf, upper_32_bits(offset));
+	intel_logical_ring_emit(ringbuf, MI_NOOP);
+	intel_logical_ring_advance(ringbuf);
+
+	return 0;
+}
+
+static bool gen8_logical_ring_get_irq(struct intel_engine_cs *ring)
+{
+	struct drm_device *dev = ring->dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	unsigned long flags;
+
+	if (WARN_ON(!intel_irqs_enabled(dev_priv)))
+		return false;
+
+	spin_lock_irqsave(&dev_priv->irq_lock, flags);
+	if (ring->irq_refcount++ == 0) {
+		I915_WRITE_IMR(ring, ~(ring->irq_enable_mask | ring->irq_keep_mask));
+		POSTING_READ(RING_IMR(ring->mmio_base));
+	}
+	spin_unlock_irqrestore(&dev_priv->irq_lock, flags);
+
+	return true;
+}
+
+static void gen8_logical_ring_put_irq(struct intel_engine_cs *ring)
+{
+	struct drm_device *dev = ring->dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	unsigned long flags;
+
+	spin_lock_irqsave(&dev_priv->irq_lock, flags);
+	if (--ring->irq_refcount == 0) {
+		I915_WRITE_IMR(ring, ~ring->irq_keep_mask);
+		POSTING_READ(RING_IMR(ring->mmio_base));
+	}
+	spin_unlock_irqrestore(&dev_priv->irq_lock, flags);
+}
+
+static int gen8_emit_flush(struct intel_ringbuffer *ringbuf,
+			   u32 invalidate_domains,
+			   u32 unused)
+{
+	struct intel_engine_cs *ring = ringbuf->ring;
+	struct drm_device *dev = ring->dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	uint32_t cmd;
+	int ret;
+
+	ret = intel_logical_ring_begin(ringbuf, 4);
+	if (ret)
+		return ret;
+
+	cmd = MI_FLUSH_DW + 1;
+
+	if (ring == &dev_priv->ring[VCS]) {
+		if (invalidate_domains & I915_GEM_GPU_DOMAINS)
+			cmd |= MI_INVALIDATE_TLB | MI_INVALIDATE_BSD |
+				MI_FLUSH_DW_STORE_INDEX |
+				MI_FLUSH_DW_OP_STOREDW;
+	} else {
+		if (invalidate_domains & I915_GEM_DOMAIN_RENDER)
+			cmd |= MI_INVALIDATE_TLB | MI_FLUSH_DW_STORE_INDEX |
+				MI_FLUSH_DW_OP_STOREDW;
+	}
+
+	intel_logical_ring_emit(ringbuf, cmd);
+	intel_logical_ring_emit(ringbuf,
+				I915_GEM_HWS_SCRATCH_ADDR |
+				MI_FLUSH_DW_USE_GTT);
+	intel_logical_ring_emit(ringbuf, 0); /* upper addr */
+	intel_logical_ring_emit(ringbuf, 0); /* value */
+	intel_logical_ring_advance(ringbuf);
+
+	return 0;
+}
+
+static int gen8_emit_flush_render(struct intel_ringbuffer *ringbuf,
+				  u32 invalidate_domains,
+				  u32 flush_domains)
+{
+	struct intel_engine_cs *ring = ringbuf->ring;
+	u32 scratch_addr = ring->scratch.gtt_offset + 2 * CACHELINE_BYTES;
+	u32 flags = 0;
+	int ret;
+
+	flags |= PIPE_CONTROL_CS_STALL;
+
+	if (flush_domains) {
+		flags |= PIPE_CONTROL_RENDER_TARGET_CACHE_FLUSH;
+		flags |= PIPE_CONTROL_DEPTH_CACHE_FLUSH;
+	}
+
+	if (invalidate_domains) {
+		flags |= PIPE_CONTROL_TLB_INVALIDATE;
+		flags |= PIPE_CONTROL_INSTRUCTION_CACHE_INVALIDATE;
+		flags |= PIPE_CONTROL_TEXTURE_CACHE_INVALIDATE;
+		flags |= PIPE_CONTROL_VF_CACHE_INVALIDATE;
+		flags |= PIPE_CONTROL_CONST_CACHE_INVALIDATE;
+		flags |= PIPE_CONTROL_STATE_CACHE_INVALIDATE;
+		flags |= PIPE_CONTROL_QW_WRITE;
+		flags |= PIPE_CONTROL_GLOBAL_GTT_IVB;
+	}
+
+	ret = intel_logical_ring_begin(ringbuf, 6);
+	if (ret)
+		return ret;
+
+	intel_logical_ring_emit(ringbuf, GFX_OP_PIPE_CONTROL(6));
+	intel_logical_ring_emit(ringbuf, flags);
+	intel_logical_ring_emit(ringbuf, scratch_addr);
+	intel_logical_ring_emit(ringbuf, 0);
+	intel_logical_ring_emit(ringbuf, 0);
+	intel_logical_ring_emit(ringbuf, 0);
+	intel_logical_ring_advance(ringbuf);
+
+	return 0;
+}
+
+static u32 gen8_get_seqno(struct intel_engine_cs *ring, bool lazy_coherency)
+{
+	return intel_read_status_page(ring, I915_GEM_HWS_INDEX);
+}
+
+static void gen8_set_seqno(struct intel_engine_cs *ring, u32 seqno)
+{
+	intel_write_status_page(ring, I915_GEM_HWS_INDEX, seqno);
+}
+
+static int gen8_emit_request(struct intel_ringbuffer *ringbuf)
+{
+	struct intel_engine_cs *ring = ringbuf->ring;
+	u32 cmd;
+	int ret;
+
+	ret = intel_logical_ring_begin(ringbuf, 6);
+	if (ret)
+		return ret;
+
+	cmd = MI_STORE_DWORD_IMM_GEN8;
+	cmd |= MI_GLOBAL_GTT;
+
+	intel_logical_ring_emit(ringbuf, cmd);
+	intel_logical_ring_emit(ringbuf,
+				(ring->status_page.gfx_addr +
+				(I915_GEM_HWS_INDEX << MI_STORE_DWORD_INDEX_SHIFT)));
+	intel_logical_ring_emit(ringbuf, 0);
+	intel_logical_ring_emit(ringbuf, ring->outstanding_lazy_seqno);
+	intel_logical_ring_emit(ringbuf, MI_USER_INTERRUPT);
+	intel_logical_ring_emit(ringbuf, MI_NOOP);
+	intel_logical_ring_advance_and_submit(ringbuf);
+
+	return 0;
+}
+
+/**
+ * intel_logical_ring_cleanup() - deallocate the Engine Command Streamer
+ *
+ * @ring: Engine Command Streamer.
+ *
+ */
+void intel_logical_ring_cleanup(struct intel_engine_cs *ring)
+{
+	struct drm_i915_private *dev_priv;
+
+	if (!intel_ring_initialized(ring))
+		return;
+
+	dev_priv = ring->dev->dev_private;
+
+	intel_logical_ring_stop(ring);
+	WARN_ON((I915_READ_MODE(ring) & MODE_IDLE) == 0);
+	ring->preallocated_lazy_request = NULL;
+	ring->outstanding_lazy_seqno = 0;
+
+	if (ring->cleanup)
+		ring->cleanup(ring);
+
+	i915_cmd_parser_fini_ring(ring);
+
+	if (ring->status_page.obj) {
+		kunmap(sg_page(ring->status_page.obj->pages->sgl));
+		ring->status_page.obj = NULL;
+	}
+}
+
+static int logical_ring_init(struct drm_device *dev, struct intel_engine_cs *ring)
+{
+	int ret;
+
+	/* Intentionally left blank. */
+	ring->buffer = NULL;
+
+	ring->dev = dev;
+	INIT_LIST_HEAD(&ring->active_list);
+	INIT_LIST_HEAD(&ring->request_list);
+	init_waitqueue_head(&ring->irq_queue);
+
+	INIT_LIST_HEAD(&ring->execlist_queue);
+	INIT_LIST_HEAD(&ring->execlist_retired_req_list);
+	spin_lock_init(&ring->execlist_lock);
+	ring->next_context_status_buffer = 0;
+
+	ret = i915_cmd_parser_init_ring(ring);
+	if (ret)
+		return ret;
+
+	if (ring->init) {
+		ret = ring->init(ring);
+		if (ret)
+			return ret;
+	}
+
+	ret = intel_lr_context_deferred_create(ring->default_context, ring);
+
+	return ret;
+}
+
+static int logical_render_ring_init(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct intel_engine_cs *ring = &dev_priv->ring[RCS];
+
+	ring->name = "render ring";
+	ring->id = RCS;
+	ring->mmio_base = RENDER_RING_BASE;
+	ring->irq_enable_mask =
+		GT_RENDER_USER_INTERRUPT << GEN8_RCS_IRQ_SHIFT;
+	ring->irq_keep_mask =
+		GT_CONTEXT_SWITCH_INTERRUPT << GEN8_RCS_IRQ_SHIFT;
+	if (HAS_L3_DPF(dev))
+		ring->irq_keep_mask |= GT_RENDER_L3_PARITY_ERROR_INTERRUPT;
+
+	ring->init = gen8_init_render_ring;
+	ring->init_context = intel_logical_ring_workarounds_emit;
+	ring->cleanup = intel_fini_pipe_control;
+	ring->get_seqno = gen8_get_seqno;
+	ring->set_seqno = gen8_set_seqno;
+	ring->emit_request = gen8_emit_request;
+	ring->emit_flush = gen8_emit_flush_render;
+	ring->irq_get = gen8_logical_ring_get_irq;
+	ring->irq_put = gen8_logical_ring_put_irq;
+	ring->emit_bb_start = gen8_emit_bb_start;
+
+	return logical_ring_init(dev, ring);
+}
+
+static int logical_bsd_ring_init(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct intel_engine_cs *ring = &dev_priv->ring[VCS];
+
+	ring->name = "bsd ring";
+	ring->id = VCS;
+	ring->mmio_base = GEN6_BSD_RING_BASE;
+	ring->irq_enable_mask =
+		GT_RENDER_USER_INTERRUPT << GEN8_VCS1_IRQ_SHIFT;
+	ring->irq_keep_mask =
+		GT_CONTEXT_SWITCH_INTERRUPT << GEN8_VCS1_IRQ_SHIFT;
+
+	ring->init = gen8_init_common_ring;
+	ring->get_seqno = gen8_get_seqno;
+	ring->set_seqno = gen8_set_seqno;
+	ring->emit_request = gen8_emit_request;
+	ring->emit_flush = gen8_emit_flush;
+	ring->irq_get = gen8_logical_ring_get_irq;
+	ring->irq_put = gen8_logical_ring_put_irq;
+	ring->emit_bb_start = gen8_emit_bb_start;
+
+	return logical_ring_init(dev, ring);
+}
+
+static int logical_bsd2_ring_init(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct intel_engine_cs *ring = &dev_priv->ring[VCS2];
+
+	ring->name = "bds2 ring";
+	ring->id = VCS2;
+	ring->mmio_base = GEN8_BSD2_RING_BASE;
+	ring->irq_enable_mask =
+		GT_RENDER_USER_INTERRUPT << GEN8_VCS2_IRQ_SHIFT;
+	ring->irq_keep_mask =
+		GT_CONTEXT_SWITCH_INTERRUPT << GEN8_VCS2_IRQ_SHIFT;
+
+	ring->init = gen8_init_common_ring;
+	ring->get_seqno = gen8_get_seqno;
+	ring->set_seqno = gen8_set_seqno;
+	ring->emit_request = gen8_emit_request;
+	ring->emit_flush = gen8_emit_flush;
+	ring->irq_get = gen8_logical_ring_get_irq;
+	ring->irq_put = gen8_logical_ring_put_irq;
+	ring->emit_bb_start = gen8_emit_bb_start;
+
+	return logical_ring_init(dev, ring);
+}
+
+static int logical_blt_ring_init(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct intel_engine_cs *ring = &dev_priv->ring[BCS];
+
+	ring->name = "blitter ring";
+	ring->id = BCS;
+	ring->mmio_base = BLT_RING_BASE;
+	ring->irq_enable_mask =
+		GT_RENDER_USER_INTERRUPT << GEN8_BCS_IRQ_SHIFT;
+	ring->irq_keep_mask =
+		GT_CONTEXT_SWITCH_INTERRUPT << GEN8_BCS_IRQ_SHIFT;
+
+	ring->init = gen8_init_common_ring;
+	ring->get_seqno = gen8_get_seqno;
+	ring->set_seqno = gen8_set_seqno;
+	ring->emit_request = gen8_emit_request;
+	ring->emit_flush = gen8_emit_flush;
+	ring->irq_get = gen8_logical_ring_get_irq;
+	ring->irq_put = gen8_logical_ring_put_irq;
+	ring->emit_bb_start = gen8_emit_bb_start;
+
+	return logical_ring_init(dev, ring);
+}
+
+static int logical_vebox_ring_init(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct intel_engine_cs *ring = &dev_priv->ring[VECS];
+
+	ring->name = "video enhancement ring";
+	ring->id = VECS;
+	ring->mmio_base = VEBOX_RING_BASE;
+	ring->irq_enable_mask =
+		GT_RENDER_USER_INTERRUPT << GEN8_VECS_IRQ_SHIFT;
+	ring->irq_keep_mask =
+		GT_CONTEXT_SWITCH_INTERRUPT << GEN8_VECS_IRQ_SHIFT;
+
+	ring->init = gen8_init_common_ring;
+	ring->get_seqno = gen8_get_seqno;
+	ring->set_seqno = gen8_set_seqno;
+	ring->emit_request = gen8_emit_request;
+	ring->emit_flush = gen8_emit_flush;
+	ring->irq_get = gen8_logical_ring_get_irq;
+	ring->irq_put = gen8_logical_ring_put_irq;
+	ring->emit_bb_start = gen8_emit_bb_start;
+
+	return logical_ring_init(dev, ring);
+}
+
+/**
+ * intel_logical_rings_init() - allocate, populate and init the Engine Command Streamers
+ * @dev: DRM device.
+ *
+ * This function inits the engines for an Execlists submission style (the equivalent in the
+ * legacy ringbuffer submission world would be i915_gem_init_rings). It does it only for
+ * those engines that are present in the hardware.
+ *
+ * Return: non-zero if the initialization failed.
+ */
+int intel_logical_rings_init(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	int ret;
+
+	ret = logical_render_ring_init(dev);
+	if (ret)
+		return ret;
+
+	if (HAS_BSD(dev)) {
+		ret = logical_bsd_ring_init(dev);
+		if (ret)
+			goto cleanup_render_ring;
+	}
+
+	if (HAS_BLT(dev)) {
+		ret = logical_blt_ring_init(dev);
+		if (ret)
+			goto cleanup_bsd_ring;
+	}
+
+	if (HAS_VEBOX(dev)) {
+		ret = logical_vebox_ring_init(dev);
+		if (ret)
+			goto cleanup_blt_ring;
+	}
+
+	if (HAS_BSD2(dev)) {
+		ret = logical_bsd2_ring_init(dev);
+		if (ret)
+			goto cleanup_vebox_ring;
+	}
+
+	ret = i915_gem_set_seqno(dev, ((u32)~0 - 0x1000));
+	if (ret)
+		goto cleanup_bsd2_ring;
+
+	return 0;
+
+cleanup_bsd2_ring:
+	intel_logical_ring_cleanup(&dev_priv->ring[VCS2]);
+cleanup_vebox_ring:
+	intel_logical_ring_cleanup(&dev_priv->ring[VECS]);
+cleanup_blt_ring:
+	intel_logical_ring_cleanup(&dev_priv->ring[BCS]);
+cleanup_bsd_ring:
+	intel_logical_ring_cleanup(&dev_priv->ring[VCS]);
+cleanup_render_ring:
+	intel_logical_ring_cleanup(&dev_priv->ring[RCS]);
+
+	return ret;
+}
+
+int intel_lr_context_render_state_init(struct intel_engine_cs *ring,
+				       struct intel_context *ctx)
+{
+	struct intel_ringbuffer *ringbuf = ctx->engine[ring->id].ringbuf;
+	struct render_state so;
+	struct drm_i915_file_private *file_priv = ctx->file_priv;
+	struct drm_file *file = file_priv ? file_priv->file : NULL;
+	int ret;
+
+	ret = i915_gem_render_state_prepare(ring, &so);
+	if (ret)
+		return ret;
+
+	if (so.rodata == NULL)
+		return 0;
+
+	ret = ring->emit_bb_start(ringbuf,
+			so.ggtt_offset,
+			I915_DISPATCH_SECURE);
+	if (ret)
+		goto out;
+
+	i915_vma_move_to_active(i915_gem_obj_to_ggtt(so.obj), ring);
+
+	ret = __i915_add_request(ring, file, so.obj, NULL);
+	/* intel_logical_ring_add_request moves object to inactive if it
+	 * fails */
+out:
+	i915_gem_render_state_fini(&so);
+	return ret;
+}
+
+static int
+populate_lr_context(struct intel_context *ctx, struct drm_i915_gem_object *ctx_obj,
+		    struct intel_engine_cs *ring, struct intel_ringbuffer *ringbuf)
+{
+	struct drm_device *dev = ring->dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct i915_hw_ppgtt *ppgtt = ctx->ppgtt;
+	struct page *page;
+	uint32_t *reg_state;
+	int ret;
+
+	if (!ppgtt)
+		ppgtt = dev_priv->mm.aliasing_ppgtt;
+
+	ret = i915_gem_object_set_to_cpu_domain(ctx_obj, true);
+	if (ret) {
+		DRM_DEBUG_DRIVER("Could not set to CPU domain\n");
+		return ret;
+	}
+
+	ret = i915_gem_object_get_pages(ctx_obj);
+	if (ret) {
+		DRM_DEBUG_DRIVER("Could not get object pages\n");
+		return ret;
+	}
+
+	i915_gem_object_pin_pages(ctx_obj);
+
+	/* The second page of the context object contains some fields which must
+	 * be set up prior to the first execution. */
+	page = i915_gem_object_get_page(ctx_obj, 1);
+	reg_state = kmap_atomic(page);
+
+	/* A context is actually a big batch buffer with several MI_LOAD_REGISTER_IMM
+	 * commands followed by (reg, value) pairs. The values we are setting here are
+	 * only for the first context restore: on a subsequent save, the GPU will
+	 * recreate this batchbuffer with new values (including all the missing
+	 * MI_LOAD_REGISTER_IMM commands that we are not initializing here). */
+	if (ring->id == RCS)
+		reg_state[CTX_LRI_HEADER_0] = MI_LOAD_REGISTER_IMM(14);
+	else
+		reg_state[CTX_LRI_HEADER_0] = MI_LOAD_REGISTER_IMM(11);
+	reg_state[CTX_LRI_HEADER_0] |= MI_LRI_FORCE_POSTED;
+	reg_state[CTX_CONTEXT_CONTROL] = RING_CONTEXT_CONTROL(ring);
+	reg_state[CTX_CONTEXT_CONTROL+1] =
+			_MASKED_BIT_ENABLE((1<<3) | MI_RESTORE_INHIBIT);
+	reg_state[CTX_RING_HEAD] = RING_HEAD(ring->mmio_base);
+	reg_state[CTX_RING_HEAD+1] = 0;
+	reg_state[CTX_RING_TAIL] = RING_TAIL(ring->mmio_base);
+	reg_state[CTX_RING_TAIL+1] = 0;
+	reg_state[CTX_RING_BUFFER_START] = RING_START(ring->mmio_base);
+	/* Ring buffer start address is not known until the buffer is pinned.
+	 * It is written to the context image in execlists_update_context()
+	 */
+	reg_state[CTX_RING_BUFFER_CONTROL] = RING_CTL(ring->mmio_base);
+	reg_state[CTX_RING_BUFFER_CONTROL+1] =
+			((ringbuf->size - PAGE_SIZE) & RING_NR_PAGES) | RING_VALID;
+	reg_state[CTX_BB_HEAD_U] = ring->mmio_base + 0x168;
+	reg_state[CTX_BB_HEAD_U+1] = 0;
+	reg_state[CTX_BB_HEAD_L] = ring->mmio_base + 0x140;
+	reg_state[CTX_BB_HEAD_L+1] = 0;
+	reg_state[CTX_BB_STATE] = ring->mmio_base + 0x110;
+	reg_state[CTX_BB_STATE+1] = (1<<5);
+	reg_state[CTX_SECOND_BB_HEAD_U] = ring->mmio_base + 0x11c;
+	reg_state[CTX_SECOND_BB_HEAD_U+1] = 0;
+	reg_state[CTX_SECOND_BB_HEAD_L] = ring->mmio_base + 0x114;
+	reg_state[CTX_SECOND_BB_HEAD_L+1] = 0;
+	reg_state[CTX_SECOND_BB_STATE] = ring->mmio_base + 0x118;
+	reg_state[CTX_SECOND_BB_STATE+1] = 0;
+	if (ring->id == RCS) {
+		/* TODO: according to BSpec, the register state context
+		 * for CHV does not have these. OTOH, these registers do
+		 * exist in CHV. I'm waiting for a clarification */
+		reg_state[CTX_BB_PER_CTX_PTR] = ring->mmio_base + 0x1c0;
+		reg_state[CTX_BB_PER_CTX_PTR+1] = 0;
+		reg_state[CTX_RCS_INDIRECT_CTX] = ring->mmio_base + 0x1c4;
+		reg_state[CTX_RCS_INDIRECT_CTX+1] = 0;
+		reg_state[CTX_RCS_INDIRECT_CTX_OFFSET] = ring->mmio_base + 0x1c8;
+		reg_state[CTX_RCS_INDIRECT_CTX_OFFSET+1] = 0;
+	}
+	reg_state[CTX_LRI_HEADER_1] = MI_LOAD_REGISTER_IMM(9);
+	reg_state[CTX_LRI_HEADER_1] |= MI_LRI_FORCE_POSTED;
+	reg_state[CTX_CTX_TIMESTAMP] = ring->mmio_base + 0x3a8;
+	reg_state[CTX_CTX_TIMESTAMP+1] = 0;
+	reg_state[CTX_PDP3_UDW] = GEN8_RING_PDP_UDW(ring, 3);
+	reg_state[CTX_PDP3_LDW] = GEN8_RING_PDP_LDW(ring, 3);
+	reg_state[CTX_PDP2_UDW] = GEN8_RING_PDP_UDW(ring, 2);
+	reg_state[CTX_PDP2_LDW] = GEN8_RING_PDP_LDW(ring, 2);
+	reg_state[CTX_PDP1_UDW] = GEN8_RING_PDP_UDW(ring, 1);
+	reg_state[CTX_PDP1_LDW] = GEN8_RING_PDP_LDW(ring, 1);
+	reg_state[CTX_PDP0_UDW] = GEN8_RING_PDP_UDW(ring, 0);
+	reg_state[CTX_PDP0_LDW] = GEN8_RING_PDP_LDW(ring, 0);
+	reg_state[CTX_PDP3_UDW+1] = upper_32_bits(ppgtt->pd_dma_addr[3]);
+	reg_state[CTX_PDP3_LDW+1] = lower_32_bits(ppgtt->pd_dma_addr[3]);
+	reg_state[CTX_PDP2_UDW+1] = upper_32_bits(ppgtt->pd_dma_addr[2]);
+	reg_state[CTX_PDP2_LDW+1] = lower_32_bits(ppgtt->pd_dma_addr[2]);
+	reg_state[CTX_PDP1_UDW+1] = upper_32_bits(ppgtt->pd_dma_addr[1]);
+	reg_state[CTX_PDP1_LDW+1] = lower_32_bits(ppgtt->pd_dma_addr[1]);
+	reg_state[CTX_PDP0_UDW+1] = upper_32_bits(ppgtt->pd_dma_addr[0]);
+	reg_state[CTX_PDP0_LDW+1] = lower_32_bits(ppgtt->pd_dma_addr[0]);
+	if (ring->id == RCS) {
+		reg_state[CTX_LRI_HEADER_2] = MI_LOAD_REGISTER_IMM(1);
+		reg_state[CTX_R_PWR_CLK_STATE] = 0x20c8;
+		reg_state[CTX_R_PWR_CLK_STATE+1] = 0;
+	}
+
+	kunmap_atomic(reg_state);
+
+	ctx_obj->dirty = 1;
+	set_page_dirty(page);
+	i915_gem_object_unpin_pages(ctx_obj);
+
+	return 0;
+}
+
+/**
+ * intel_lr_context_free() - free the LRC specific bits of a context
+ * @ctx: the LR context to free.
+ *
+ * The real context freeing is done in i915_gem_context_free: this only
+ * takes care of the bits that are LRC related: the per-engine backing
+ * objects and the logical ringbuffer.
+ */
+void intel_lr_context_free(struct intel_context *ctx)
+{
+	int i;
+
+	for (i = 0; i < I915_NUM_RINGS; i++) {
+		struct drm_i915_gem_object *ctx_obj = ctx->engine[i].state;
+
+		if (ctx_obj) {
+			struct intel_ringbuffer *ringbuf =
+					ctx->engine[i].ringbuf;
+			struct intel_engine_cs *ring = ringbuf->ring;
+
+			if (ctx == ring->default_context) {
+				intel_unpin_ringbuffer_obj(ringbuf);
+				i915_gem_object_ggtt_unpin(ctx_obj);
+			}
+			intel_destroy_ringbuffer_obj(ringbuf);
+			kfree(ringbuf);
+			drm_gem_object_unreference(&ctx_obj->base);
+		}
+	}
+}
+
+static uint32_t get_lr_context_size(struct intel_engine_cs *ring)
+{
+	int ret = 0;
+
+	WARN_ON(INTEL_INFO(ring->dev)->gen < 8);
+
+	switch (ring->id) {
+	case RCS:
+		if (INTEL_INFO(ring->dev)->gen >= 9)
+			ret = GEN9_LR_CONTEXT_RENDER_SIZE;
+		else
+			ret = GEN8_LR_CONTEXT_RENDER_SIZE;
+		break;
+	case VCS:
+	case BCS:
+	case VECS:
+	case VCS2:
+		ret = GEN8_LR_CONTEXT_OTHER_SIZE;
+		break;
+	}
+
+	return ret;
+}
+
+static void lrc_setup_hardware_status_page(struct intel_engine_cs *ring,
+		struct drm_i915_gem_object *default_ctx_obj)
+{
+	struct drm_i915_private *dev_priv = ring->dev->dev_private;
+
+	/* The status page is offset 0 from the default context object
+	 * in LRC mode. */
+	ring->status_page.gfx_addr = i915_gem_obj_ggtt_offset(default_ctx_obj);
+	ring->status_page.page_addr =
+			kmap(sg_page(default_ctx_obj->pages->sgl));
+	ring->status_page.obj = default_ctx_obj;
+
+	I915_WRITE(RING_HWS_PGA(ring->mmio_base),
+			(u32)ring->status_page.gfx_addr);
+	POSTING_READ(RING_HWS_PGA(ring->mmio_base));
+}
+
+/**
+ * intel_lr_context_deferred_create() - create the LRC specific bits of a context
+ * @ctx: LR context to create.
+ * @ring: engine to be used with the context.
+ *
+ * This function can be called more than once, with different engines, if we plan
+ * to use the context with them. The context backing objects and the ringbuffers
+ * (specially the ringbuffer backing objects) suck a lot of memory up, and that's why
+ * the creation is a deferred call: it's better to make sure first that we need to use
+ * a given ring with the context.
+ *
+ * Return: non-zero on error.
+ */
+int intel_lr_context_deferred_create(struct intel_context *ctx,
+				     struct intel_engine_cs *ring)
+{
+	const bool is_global_default_ctx = (ctx == ring->default_context);
+	struct drm_device *dev = ring->dev;
+	struct drm_i915_gem_object *ctx_obj;
+	uint32_t context_size;
+	struct intel_ringbuffer *ringbuf;
+	int ret;
+
+	WARN_ON(ctx->legacy_hw_ctx.rcs_state != NULL);
+	if (ctx->engine[ring->id].state)
+		return 0;
+
+	context_size = round_up(get_lr_context_size(ring), 4096);
+
+	ctx_obj = i915_gem_alloc_context_obj(dev, context_size);
+	if (IS_ERR(ctx_obj)) {
+		ret = PTR_ERR(ctx_obj);
+		DRM_DEBUG_DRIVER("Alloc LRC backing obj failed: %d\n", ret);
+		return ret;
+	}
+
+	if (is_global_default_ctx) {
+		ret = i915_gem_obj_ggtt_pin(ctx_obj, GEN8_LR_CONTEXT_ALIGN, 0);
+		if (ret) {
+			DRM_DEBUG_DRIVER("Pin LRC backing obj failed: %d\n",
+					ret);
+			drm_gem_object_unreference(&ctx_obj->base);
+			return ret;
+		}
+	}
+
+	ringbuf = kzalloc(sizeof(*ringbuf), GFP_KERNEL);
+	if (!ringbuf) {
+		DRM_DEBUG_DRIVER("Failed to allocate ringbuffer %s\n",
+				ring->name);
+		ret = -ENOMEM;
+		goto error_unpin_ctx;
+	}
+
+	ringbuf->ring = ring;
+	ringbuf->FIXME_lrc_ctx = ctx;
+
+	ringbuf->size = 32 * PAGE_SIZE;
+	ringbuf->effective_size = ringbuf->size;
+	ringbuf->head = 0;
+	ringbuf->tail = 0;
+	ringbuf->space = ringbuf->size;
+	ringbuf->last_retired_head = -1;
+
+	if (ringbuf->obj == NULL) {
+		ret = intel_alloc_ringbuffer_obj(dev, ringbuf);
+		if (ret) {
+			DRM_DEBUG_DRIVER(
+				"Failed to allocate ringbuffer obj %s: %d\n",
+				ring->name, ret);
+			goto error_free_rbuf;
+		}
+
+		if (is_global_default_ctx) {
+			ret = intel_pin_and_map_ringbuffer_obj(dev, ringbuf);
+			if (ret) {
+				DRM_ERROR(
+					"Failed to pin and map ringbuffer %s: %d\n",
+					ring->name, ret);
+				goto error_destroy_rbuf;
+			}
+		}
+
+	}
+
+	ret = populate_lr_context(ctx, ctx_obj, ring, ringbuf);
+	if (ret) {
+		DRM_DEBUG_DRIVER("Failed to populate LRC: %d\n", ret);
+		goto error;
+	}
+
+	ctx->engine[ring->id].ringbuf = ringbuf;
+	ctx->engine[ring->id].state = ctx_obj;
+
+	if (ctx == ring->default_context)
+		lrc_setup_hardware_status_page(ring, ctx_obj);
+
+	if (ring->id == RCS && !ctx->rcs_initialized) {
+		if (ring->init_context) {
+			ret = ring->init_context(ring, ctx);
+			if (ret)
+				DRM_ERROR("ring init context: %d\n", ret);
+		}
+
+		ret = intel_lr_context_render_state_init(ring, ctx);
+		if (ret) {
+			DRM_ERROR("Init render state failed: %d\n", ret);
+			ctx->engine[ring->id].ringbuf = NULL;
+			ctx->engine[ring->id].state = NULL;
+			goto error;
+		}
+		ctx->rcs_initialized = true;
+	}
+
+	return 0;
+
+error:
+	if (is_global_default_ctx)
+		intel_unpin_ringbuffer_obj(ringbuf);
+error_destroy_rbuf:
+	intel_destroy_ringbuffer_obj(ringbuf);
+error_free_rbuf:
+	kfree(ringbuf);
+error_unpin_ctx:
+	if (is_global_default_ctx)
+		i915_gem_object_ggtt_unpin(ctx_obj);
+	drm_gem_object_unreference(&ctx_obj->base);
+	return ret;
+}
diff --git a/drivers/gpu/drm/i915/intel_lrc.h b/drivers/gpu/drm/i915/intel_lrc.h
new file mode 100644
index 000000000000..14b216b9be7f
--- /dev/null
+++ b/drivers/gpu/drm/i915/intel_lrc.h
@@ -0,0 +1,118 @@
+/*
+ * Copyright © 2014 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ */
+
+#ifndef _INTEL_LRC_H_
+#define _INTEL_LRC_H_
+
+#define GEN8_LR_CONTEXT_ALIGN 4096
+
+/* Execlists regs */
+#define RING_ELSP(ring)			((ring)->mmio_base+0x230)
+#define RING_EXECLIST_STATUS(ring)	((ring)->mmio_base+0x234)
+#define RING_CONTEXT_CONTROL(ring)	((ring)->mmio_base+0x244)
+#define RING_CONTEXT_STATUS_BUF(ring)	((ring)->mmio_base+0x370)
+#define RING_CONTEXT_STATUS_PTR(ring)	((ring)->mmio_base+0x3a0)
+
+/* Logical Rings */
+void intel_logical_ring_stop(struct intel_engine_cs *ring);
+void intel_logical_ring_cleanup(struct intel_engine_cs *ring);
+int intel_logical_rings_init(struct drm_device *dev);
+
+int logical_ring_flush_all_caches(struct intel_ringbuffer *ringbuf);
+void intel_logical_ring_advance_and_submit(struct intel_ringbuffer *ringbuf);
+/**
+ * intel_logical_ring_advance() - advance the ringbuffer tail
+ * @ringbuf: Ringbuffer to advance.
+ *
+ * The tail is only updated in our logical ringbuffer struct.
+ */
+static inline void intel_logical_ring_advance(struct intel_ringbuffer *ringbuf)
+{
+	ringbuf->tail &= ringbuf->size - 1;
+}
+/**
+ * intel_logical_ring_emit() - write a DWORD to the ringbuffer.
+ * @ringbuf: Ringbuffer to write to.
+ * @data: DWORD to write.
+ */
+static inline void intel_logical_ring_emit(struct intel_ringbuffer *ringbuf,
+					   u32 data)
+{
+	iowrite32(data, ringbuf->virtual_start + ringbuf->tail);
+	ringbuf->tail += 4;
+}
+int intel_logical_ring_begin(struct intel_ringbuffer *ringbuf, int num_dwords);
+
+/* Logical Ring Contexts */
+int intel_lr_context_render_state_init(struct intel_engine_cs *ring,
+				       struct intel_context *ctx);
+void intel_lr_context_free(struct intel_context *ctx);
+int intel_lr_context_deferred_create(struct intel_context *ctx,
+				     struct intel_engine_cs *ring);
+void intel_lr_context_unpin(struct intel_engine_cs *ring,
+		struct intel_context *ctx);
+
+/* Execlists */
+int intel_sanitize_enable_execlists(struct drm_device *dev, int enable_execlists);
+int intel_execlists_submission(struct drm_device *dev, struct drm_file *file,
+			       struct intel_engine_cs *ring,
+			       struct intel_context *ctx,
+			       struct drm_i915_gem_execbuffer2 *args,
+			       struct list_head *vmas,
+			       struct drm_i915_gem_object *batch_obj,
+			       u64 exec_start, u32 flags);
+u32 intel_execlists_ctx_id(struct drm_i915_gem_object *ctx_obj);
+
+/**
+ * struct intel_ctx_submit_request - queued context submission request
+ * @ctx: Context to submit to the ELSP.
+ * @ring: Engine to submit it to.
+ * @tail: how far in the context's ringbuffer this request goes to.
+ * @execlist_link: link in the submission queue.
+ * @work: workqueue for processing this request in a bottom half.
+ * @elsp_submitted: no. of times this request has been sent to the ELSP.
+ *
+ * The ELSP only accepts two elements at a time, so we queue context/tail
+ * pairs on a given queue (ring->execlist_queue) until the hardware is
+ * available. The queue serves a double purpose: we also use it to keep track
+ * of the up to 2 contexts currently in the hardware (usually one in execution
+ * and the other queued up by the GPU): We only remove elements from the head
+ * of the queue when the hardware informs us that an element has been
+ * completed.
+ *
+ * All accesses to the queue are mediated by a spinlock (ring->execlist_lock).
+ */
+struct intel_ctx_submit_request {
+	struct intel_context *ctx;
+	struct intel_engine_cs *ring;
+	u32 tail;
+
+	struct list_head execlist_link;
+
+	int elsp_submitted;
+};
+
+void intel_execlists_handle_ctx_events(struct intel_engine_cs *ring);
+void intel_execlists_retire_requests(struct intel_engine_cs *ring);
+
+#endif /* _INTEL_LRC_H_ */
diff --git a/drivers/gpu/drm/i915/intel_lvds.c b/drivers/gpu/drm/i915/intel_lvds.c
index fdf40267249c..14654d628ca4 100644
--- a/drivers/gpu/drm/i915/intel_lvds.c
+++ b/drivers/gpu/drm/i915/intel_lvds.c
@@ -76,7 +76,7 @@ static bool intel_lvds_get_hw_state(struct intel_encoder *encoder,
 	u32 tmp;
 
 	power_domain = intel_display_port_power_domain(encoder);
-	if (!intel_display_power_enabled(dev_priv, power_domain))
+	if (!intel_display_power_is_enabled(dev_priv, power_domain))
 		return false;
 
 	tmp = I915_READ(lvds_encoder->reg);
@@ -823,8 +823,7 @@ bool intel_is_dual_link_lvds(struct drm_device *dev)
 	struct intel_encoder *encoder;
 	struct intel_lvds_encoder *lvds_encoder;
 
-	list_for_each_entry(encoder, &dev->mode_config.encoder_list,
-			    base.head) {
+	for_each_intel_encoder(dev, encoder) {
 		if (encoder->type == INTEL_OUTPUT_LVDS) {
 			lvds_encoder = to_lvds_encoder(&encoder->base);
 
@@ -900,6 +899,17 @@ void intel_lvds_init(struct drm_device *dev)
 	int pipe;
 	u8 pin;
 
+	/*
+	 * Unlock registers and just leave them unlocked. Do this before
+	 * checking quirk lists to avoid bogus WARNINGs.
+	 */
+	if (HAS_PCH_SPLIT(dev)) {
+		I915_WRITE(PCH_PP_CONTROL,
+			   I915_READ(PCH_PP_CONTROL) | PANEL_UNLOCK_REGS);
+	} else {
+		I915_WRITE(PP_CONTROL,
+			   I915_READ(PP_CONTROL) | PANEL_UNLOCK_REGS);
+	}
 	if (!intel_lvds_supported(dev))
 		return;
 
@@ -1098,17 +1108,6 @@ out:
 	lvds_encoder->a3_power = I915_READ(lvds_encoder->reg) &
 				 LVDS_A3_POWER_MASK;
 
-	/*
-	 * Unlock registers and just
-	 * leave them unlocked
-	 */
-	if (HAS_PCH_SPLIT(dev)) {
-		I915_WRITE(PCH_PP_CONTROL,
-			   I915_READ(PCH_PP_CONTROL) | PANEL_UNLOCK_REGS);
-	} else {
-		I915_WRITE(PP_CONTROL,
-			   I915_READ(PP_CONTROL) | PANEL_UNLOCK_REGS);
-	}
 	lvds_connector->lid_notifier.notifier_call = intel_lid_notify;
 	if (acpi_lid_notifier_register(&lvds_connector->lid_notifier)) {
 		DRM_DEBUG_KMS("lid notifier registration failed\n");
@@ -1117,7 +1116,7 @@ out:
 	drm_connector_register(connector);
 
 	intel_panel_init(&intel_connector->panel, fixed_mode, downclock_mode);
-	intel_panel_setup_backlight(connector);
+	intel_panel_setup_backlight(connector, INVALID_PIPE);
 
 	return;
 
diff --git a/drivers/gpu/drm/i915/intel_panel.c b/drivers/gpu/drm/i915/intel_panel.c
index 8e374449c6b5..4d63839bd9b4 100644
--- a/drivers/gpu/drm/i915/intel_panel.c
+++ b/drivers/gpu/drm/i915/intel_panel.c
@@ -419,9 +419,8 @@ static uint32_t scale(uint32_t source_val,
 	source_val = clamp(source_val, source_min, source_max);
 
 	/* avoid overflows */
-	target_val = (uint64_t)(source_val - source_min) *
-		(target_max - target_min);
-	do_div(target_val, source_max - source_min);
+	target_val = DIV_ROUND_CLOSEST_ULL((uint64_t)(source_val - source_min) *
+			(target_max - target_min), source_max - source_min);
 	target_val += target_min;
 
 	return target_val;
@@ -522,6 +521,9 @@ static u32 _vlv_get_backlight(struct drm_device *dev, enum pipe pipe)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 
+	if (WARN_ON(pipe != PIPE_A && pipe != PIPE_B))
+		return 0;
+
 	return I915_READ(VLV_BLC_PWM_CTL(pipe)) & BACKLIGHT_DUTY_CYCLE_MASK;
 }
 
@@ -537,15 +539,17 @@ static u32 intel_panel_get_backlight(struct intel_connector *connector)
 {
 	struct drm_device *dev = connector->base.dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	u32 val;
-	unsigned long flags;
+	struct intel_panel *panel = &connector->panel;
+	u32 val = 0;
 
-	spin_lock_irqsave(&dev_priv->backlight_lock, flags);
+	mutex_lock(&dev_priv->backlight_lock);
 
-	val = dev_priv->display.get_backlight(connector);
-	val = intel_panel_compute_brightness(connector, val);
+	if (panel->backlight.enabled) {
+		val = dev_priv->display.get_backlight(connector);
+		val = intel_panel_compute_brightness(connector, val);
+	}
 
-	spin_unlock_irqrestore(&dev_priv->backlight_lock, flags);
+	mutex_unlock(&dev_priv->backlight_lock);
 
 	DRM_DEBUG_DRIVER("get backlight PWM = %d\n", val);
 	return val;
@@ -604,6 +608,9 @@ static void vlv_set_backlight(struct intel_connector *connector, u32 level)
 	enum pipe pipe = intel_get_pipe_from_connector(connector);
 	u32 tmp;
 
+	if (WARN_ON(pipe != PIPE_A && pipe != PIPE_B))
+		return;
+
 	tmp = I915_READ(VLV_BLC_PWM_CTL(pipe)) & ~BACKLIGHT_DUTY_CYCLE_MASK;
 	I915_WRITE(VLV_BLC_PWM_CTL(pipe), tmp | level);
 }
@@ -627,14 +634,12 @@ static void intel_panel_set_backlight(struct intel_connector *connector,
 	struct drm_device *dev = connector->base.dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_panel *panel = &connector->panel;
-	enum pipe pipe = intel_get_pipe_from_connector(connector);
 	u32 hw_level;
-	unsigned long flags;
 
-	if (!panel->backlight.present || pipe == INVALID_PIPE)
+	if (!panel->backlight.present)
 		return;
 
-	spin_lock_irqsave(&dev_priv->backlight_lock, flags);
+	mutex_lock(&dev_priv->backlight_lock);
 
 	WARN_ON(panel->backlight.max == 0);
 
@@ -644,7 +649,7 @@ static void intel_panel_set_backlight(struct intel_connector *connector,
 	if (panel->backlight.enabled)
 		intel_panel_actually_set_backlight(connector, hw_level);
 
-	spin_unlock_irqrestore(&dev_priv->backlight_lock, flags);
+	mutex_unlock(&dev_priv->backlight_lock);
 }
 
 /* set backlight brightness to level in range [0..max], assuming hw min is
@@ -658,12 +663,17 @@ void intel_panel_set_backlight_acpi(struct intel_connector *connector,
 	struct intel_panel *panel = &connector->panel;
 	enum pipe pipe = intel_get_pipe_from_connector(connector);
 	u32 hw_level;
-	unsigned long flags;
 
+	/*
+	 * INVALID_PIPE may occur during driver init because
+	 * connection_mutex isn't held across the entire backlight
+	 * setup + modeset readout, and the BIOS can issue the
+	 * requests at any time.
+	 */
 	if (!panel->backlight.present || pipe == INVALID_PIPE)
 		return;
 
-	spin_lock_irqsave(&dev_priv->backlight_lock, flags);
+	mutex_lock(&dev_priv->backlight_lock);
 
 	WARN_ON(panel->backlight.max == 0);
 
@@ -679,7 +689,7 @@ void intel_panel_set_backlight_acpi(struct intel_connector *connector,
 	if (panel->backlight.enabled)
 		intel_panel_actually_set_backlight(connector, hw_level);
 
-	spin_unlock_irqrestore(&dev_priv->backlight_lock, flags);
+	mutex_unlock(&dev_priv->backlight_lock);
 }
 
 static void pch_disable_backlight(struct intel_connector *connector)
@@ -721,6 +731,9 @@ static void vlv_disable_backlight(struct intel_connector *connector)
 	enum pipe pipe = intel_get_pipe_from_connector(connector);
 	u32 tmp;
 
+	if (WARN_ON(pipe != PIPE_A && pipe != PIPE_B))
+		return;
+
 	intel_panel_actually_set_backlight(connector, 0);
 
 	tmp = I915_READ(VLV_BLC_PWM_CTL2(pipe));
@@ -732,10 +745,8 @@ void intel_panel_disable_backlight(struct intel_connector *connector)
 	struct drm_device *dev = connector->base.dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_panel *panel = &connector->panel;
-	enum pipe pipe = intel_get_pipe_from_connector(connector);
-	unsigned long flags;
 
-	if (!panel->backlight.present || pipe == INVALID_PIPE)
+	if (!panel->backlight.present)
 		return;
 
 	/*
@@ -749,12 +760,14 @@ void intel_panel_disable_backlight(struct intel_connector *connector)
 		return;
 	}
 
-	spin_lock_irqsave(&dev_priv->backlight_lock, flags);
+	mutex_lock(&dev_priv->backlight_lock);
 
+	if (panel->backlight.device)
+		panel->backlight.device->props.power = FB_BLANK_POWERDOWN;
 	panel->backlight.enabled = false;
 	dev_priv->display.disable_backlight(connector);
 
-	spin_unlock_irqrestore(&dev_priv->backlight_lock, flags);
+	mutex_unlock(&dev_priv->backlight_lock);
 }
 
 static void bdw_enable_backlight(struct intel_connector *connector)
@@ -778,8 +791,9 @@ static void bdw_enable_backlight(struct intel_connector *connector)
 	if (panel->backlight.active_low_pwm)
 		pch_ctl1 |= BLM_PCH_POLARITY;
 
-	/* BDW always uses the pch pwm controls. */
-	pch_ctl1 |= BLM_PCH_OVERRIDE_ENABLE;
+	/* After LPT, override is the default. */
+	if (HAS_PCH_LPT(dev_priv))
+		pch_ctl1 |= BLM_PCH_OVERRIDE_ENABLE;
 
 	I915_WRITE(BLC_PWM_PCH_CTL1, pch_ctl1);
 	POSTING_READ(BLC_PWM_PCH_CTL1);
@@ -908,6 +922,9 @@ static void vlv_enable_backlight(struct intel_connector *connector)
 	enum pipe pipe = intel_get_pipe_from_connector(connector);
 	u32 ctl, ctl2;
 
+	if (WARN_ON(pipe != PIPE_A && pipe != PIPE_B))
+		return;
+
 	ctl2 = I915_READ(VLV_BLC_PWM_CTL2(pipe));
 	if (ctl2 & BLM_PWM_ENABLE) {
 		DRM_DEBUG_KMS("backlight already enabled\n");
@@ -935,14 +952,13 @@ void intel_panel_enable_backlight(struct intel_connector *connector)
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_panel *panel = &connector->panel;
 	enum pipe pipe = intel_get_pipe_from_connector(connector);
-	unsigned long flags;
 
-	if (!panel->backlight.present || pipe == INVALID_PIPE)
+	if (!panel->backlight.present)
 		return;
 
 	DRM_DEBUG_KMS("pipe %c\n", pipe_name(pipe));
 
-	spin_lock_irqsave(&dev_priv->backlight_lock, flags);
+	mutex_lock(&dev_priv->backlight_lock);
 
 	WARN_ON(panel->backlight.max == 0);
 
@@ -957,14 +973,17 @@ void intel_panel_enable_backlight(struct intel_connector *connector)
 
 	dev_priv->display.enable_backlight(connector);
 	panel->backlight.enabled = true;
+	if (panel->backlight.device)
+		panel->backlight.device->props.power = FB_BLANK_UNBLANK;
 
-	spin_unlock_irqrestore(&dev_priv->backlight_lock, flags);
+	mutex_unlock(&dev_priv->backlight_lock);
 }
 
 #if IS_ENABLED(CONFIG_BACKLIGHT_CLASS_DEVICE)
 static int intel_backlight_device_update_status(struct backlight_device *bd)
 {
 	struct intel_connector *connector = bl_get_data(bd);
+	struct intel_panel *panel = &connector->panel;
 	struct drm_device *dev = connector->base.dev;
 
 	drm_modeset_lock(&dev->mode_config.connection_mutex, NULL);
@@ -972,6 +991,23 @@ static int intel_backlight_device_update_status(struct backlight_device *bd)
 		      bd->props.brightness, bd->props.max_brightness);
 	intel_panel_set_backlight(connector, bd->props.brightness,
 				  bd->props.max_brightness);
+
+	/*
+	 * Allow flipping bl_power as a sub-state of enabled. Sadly the
+	 * backlight class device does not make it easy to to differentiate
+	 * between callbacks for brightness and bl_power, so our backlight_power
+	 * callback needs to take this into account.
+	 */
+	if (panel->backlight.enabled) {
+		if (panel->backlight_power) {
+			bool enable = bd->props.power == FB_BLANK_UNBLANK &&
+				bd->props.brightness != 0;
+			panel->backlight_power(connector, enable);
+		}
+	} else {
+		bd->props.power = FB_BLANK_POWERDOWN;
+	}
+
 	drm_modeset_unlock(&dev->mode_config.connection_mutex);
 	return 0;
 }
@@ -1009,6 +1045,9 @@ static int intel_backlight_device_register(struct intel_connector *connector)
 	if (WARN_ON(panel->backlight.device))
 		return -ENODEV;
 
+	if (!panel->backlight.present)
+		return 0;
+
 	WARN_ON(panel->backlight.max == 0);
 
 	memset(&props, 0, sizeof(props));
@@ -1023,6 +1062,11 @@ static int intel_backlight_device_register(struct intel_connector *connector)
 					    panel->backlight.level,
 					    props.max_brightness);
 
+	if (panel->backlight.enabled)
+		props.power = FB_BLANK_UNBLANK;
+	else
+		props.power = FB_BLANK_POWERDOWN;
+
 	/*
 	 * Note: using the same name independent of the connector prevents
 	 * registration of multiple backlight devices in the driver.
@@ -1039,6 +1083,10 @@ static int intel_backlight_device_register(struct intel_connector *connector)
 		panel->backlight.device = NULL;
 		return -ENODEV;
 	}
+
+	DRM_DEBUG_KMS("Connector %s backlight sysfs interface registered\n",
+		      connector->base.name);
+
 	return 0;
 }
 
@@ -1072,15 +1120,28 @@ static u32 get_backlight_min_vbt(struct intel_connector *connector)
 	struct drm_device *dev = connector->base.dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_panel *panel = &connector->panel;
+	int min;
 
 	WARN_ON(panel->backlight.max == 0);
 
+	/*
+	 * XXX: If the vbt value is 255, it makes min equal to max, which leads
+	 * to problems. There are such machines out there. Either our
+	 * interpretation is wrong or the vbt has bogus data. Or both. Safeguard
+	 * against this by letting the minimum be at most (arbitrarily chosen)
+	 * 25% of the max.
+	 */
+	min = clamp_t(int, dev_priv->vbt.backlight.min_brightness, 0, 64);
+	if (min != dev_priv->vbt.backlight.min_brightness) {
+		DRM_DEBUG_KMS("clamping VBT min backlight %d/255 to %d/255\n",
+			      dev_priv->vbt.backlight.min_brightness, min);
+	}
+
 	/* vbt value is a coefficient in range [0..255] */
-	return scale(dev_priv->vbt.backlight.min_brightness, 0, 255,
-		     0, panel->backlight.max);
+	return scale(min, 0, 255, 0, panel->backlight.max);
 }
 
-static int bdw_setup_backlight(struct intel_connector *connector)
+static int bdw_setup_backlight(struct intel_connector *connector, enum pipe unused)
 {
 	struct drm_device *dev = connector->base.dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
@@ -1106,7 +1167,7 @@ static int bdw_setup_backlight(struct intel_connector *connector)
 	return 0;
 }
 
-static int pch_setup_backlight(struct intel_connector *connector)
+static int pch_setup_backlight(struct intel_connector *connector, enum pipe unused)
 {
 	struct drm_device *dev = connector->base.dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
@@ -1133,7 +1194,7 @@ static int pch_setup_backlight(struct intel_connector *connector)
 	return 0;
 }
 
-static int i9xx_setup_backlight(struct intel_connector *connector)
+static int i9xx_setup_backlight(struct intel_connector *connector, enum pipe unused)
 {
 	struct drm_device *dev = connector->base.dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
@@ -1165,7 +1226,7 @@ static int i9xx_setup_backlight(struct intel_connector *connector)
 	return 0;
 }
 
-static int i965_setup_backlight(struct intel_connector *connector)
+static int i965_setup_backlight(struct intel_connector *connector, enum pipe unused)
 {
 	struct drm_device *dev = connector->base.dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
@@ -1195,37 +1256,40 @@ static int i965_setup_backlight(struct intel_connector *connector)
 	return 0;
 }
 
-static int vlv_setup_backlight(struct intel_connector *connector)
+static int vlv_setup_backlight(struct intel_connector *connector, enum pipe pipe)
 {
 	struct drm_device *dev = connector->base.dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_panel *panel = &connector->panel;
-	enum pipe pipe;
+	enum pipe p;
 	u32 ctl, ctl2, val;
 
-	for_each_pipe(pipe) {
-		u32 cur_val = I915_READ(VLV_BLC_PWM_CTL(pipe));
+	for_each_pipe(dev_priv, p) {
+		u32 cur_val = I915_READ(VLV_BLC_PWM_CTL(p));
 
 		/* Skip if the modulation freq is already set */
 		if (cur_val & ~BACKLIGHT_DUTY_CYCLE_MASK)
 			continue;
 
 		cur_val &= BACKLIGHT_DUTY_CYCLE_MASK;
-		I915_WRITE(VLV_BLC_PWM_CTL(pipe), (0xf42 << 16) |
+		I915_WRITE(VLV_BLC_PWM_CTL(p), (0xf42 << 16) |
 			   cur_val);
 	}
 
-	ctl2 = I915_READ(VLV_BLC_PWM_CTL2(PIPE_A));
+	if (WARN_ON(pipe != PIPE_A && pipe != PIPE_B))
+		return -ENODEV;
+
+	ctl2 = I915_READ(VLV_BLC_PWM_CTL2(pipe));
 	panel->backlight.active_low_pwm = ctl2 & BLM_POLARITY_I965;
 
-	ctl = I915_READ(VLV_BLC_PWM_CTL(PIPE_A));
+	ctl = I915_READ(VLV_BLC_PWM_CTL(pipe));
 	panel->backlight.max = ctl >> 16;
 	if (!panel->backlight.max)
 		return -ENODEV;
 
 	panel->backlight.min = get_backlight_min_vbt(connector);
 
-	val = _vlv_get_backlight(dev, PIPE_A);
+	val = _vlv_get_backlight(dev, pipe);
 	panel->backlight.level = intel_panel_compute_brightness(connector, val);
 
 	panel->backlight.enabled = (ctl2 & BLM_PWM_ENABLE) &&
@@ -1234,13 +1298,12 @@ static int vlv_setup_backlight(struct intel_connector *connector)
 	return 0;
 }
 
-int intel_panel_setup_backlight(struct drm_connector *connector)
+int intel_panel_setup_backlight(struct drm_connector *connector, enum pipe pipe)
 {
 	struct drm_device *dev = connector->dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_connector *intel_connector = to_intel_connector(connector);
 	struct intel_panel *panel = &intel_connector->panel;
-	unsigned long flags;
 	int ret;
 
 	if (!dev_priv->vbt.backlight.present) {
@@ -1253,9 +1316,9 @@ int intel_panel_setup_backlight(struct drm_connector *connector)
 	}
 
 	/* set level and max in panel struct */
-	spin_lock_irqsave(&dev_priv->backlight_lock, flags);
-	ret = dev_priv->display.setup_backlight(intel_connector);
-	spin_unlock_irqrestore(&dev_priv->backlight_lock, flags);
+	mutex_lock(&dev_priv->backlight_lock);
+	ret = dev_priv->display.setup_backlight(intel_connector, pipe);
+	mutex_unlock(&dev_priv->backlight_lock);
 
 	if (ret) {
 		DRM_DEBUG_KMS("failed to setup backlight for connector %s\n",
@@ -1263,15 +1326,12 @@ int intel_panel_setup_backlight(struct drm_connector *connector)
 		return ret;
 	}
 
-	intel_backlight_device_register(intel_connector);
-
 	panel->backlight.present = true;
 
-	DRM_DEBUG_KMS("backlight initialized, %s, brightness %u/%u, "
-		      "sysfs interface %sregistered\n",
+	DRM_DEBUG_KMS("Connector %s backlight initialized, %s, brightness %u/%u\n",
+		      connector->name,
 		      panel->backlight.enabled ? "enabled" : "disabled",
-		      panel->backlight.level, panel->backlight.max,
-		      panel->backlight.device ? "" : "not ");
+		      panel->backlight.level, panel->backlight.max);
 
 	return 0;
 }
@@ -1282,7 +1342,6 @@ void intel_panel_destroy_backlight(struct drm_connector *connector)
 	struct intel_panel *panel = &intel_connector->panel;
 
 	panel->backlight.present = false;
-	intel_backlight_device_unregister(intel_connector);
 }
 
 /* Set up chip specific backlight functions */
@@ -1290,7 +1349,7 @@ void intel_panel_init_backlight_funcs(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 
-	if (IS_BROADWELL(dev)) {
+	if (IS_BROADWELL(dev) || (INTEL_INFO(dev)->gen >= 9)) {
 		dev_priv->display.setup_backlight = bdw_setup_backlight;
 		dev_priv->display.enable_backlight = bdw_enable_backlight;
 		dev_priv->display.disable_backlight = pch_disable_backlight;
@@ -1345,3 +1404,19 @@ void intel_panel_fini(struct intel_panel *panel)
 		drm_mode_destroy(intel_connector->base.dev,
 				panel->downclock_mode);
 }
+
+void intel_backlight_register(struct drm_device *dev)
+{
+	struct intel_connector *connector;
+
+	list_for_each_entry(connector, &dev->mode_config.connector_list, base.head)
+		intel_backlight_device_register(connector);
+}
+
+void intel_backlight_unregister(struct drm_device *dev)
+{
+	struct intel_connector *connector;
+
+	list_for_each_entry(connector, &dev->mode_config.connector_list, base.head)
+		intel_backlight_device_unregister(connector);
+}
diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c
index 40c12295c0bd..964b28e3c630 100644
--- a/drivers/gpu/drm/i915/intel_pm.c
+++ b/drivers/gpu/drm/i915/intel_pm.c
@@ -30,9 +30,6 @@
 #include "intel_drv.h"
 #include "../../../platform/x86/intel_ips.h"
 #include <linux/module.h>
-#include <linux/vgaarb.h>
-#include <drm/i915_powerwell.h>
-#include <linux/pm_runtime.h>
 
 /**
  * RC6 is a special power stage which allows the GPU to enter an very
@@ -66,11 +63,37 @@
  * i915.i915_enable_fbc parameter
  */
 
+static void gen9_init_clock_gating(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+
+	/*
+	 * WaDisableSDEUnitClockGating:skl
+	 * This seems to be a pre-production w/a.
+	 */
+	I915_WRITE(GEN8_UCGCTL6, I915_READ(GEN8_UCGCTL6) |
+		   GEN8_SDEUNIT_CLOCK_GATE_DISABLE);
+
+	/*
+	 * WaDisableDgMirrorFixInHalfSliceChicken5:skl
+	 * This is a pre-production w/a.
+	 */
+	I915_WRITE(GEN9_HALF_SLICE_CHICKEN5,
+		   I915_READ(GEN9_HALF_SLICE_CHICKEN5) &
+		   ~GEN9_DG_MIRROR_FIX_ENABLE);
+
+	/* Wa4x4STCOptimizationDisable:skl */
+	I915_WRITE(CACHE_MODE_1,
+		   _MASKED_BIT_ENABLE(GEN8_4x4_STC_OPTIMIZATION_DISABLE));
+}
+
 static void i8xx_disable_fbc(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	u32 fbc_ctl;
 
+	dev_priv->fbc.enabled = false;
+
 	/* Disable compression */
 	fbc_ctl = I915_READ(FBC_CONTROL);
 	if ((fbc_ctl & FBC_CTL_EN) == 0)
@@ -99,6 +122,8 @@ static void i8xx_enable_fbc(struct drm_crtc *crtc)
 	int i;
 	u32 fbc_ctl;
 
+	dev_priv->fbc.enabled = true;
+
 	cfb_pitch = dev_priv->fbc.size / FBC_LL_SIZE;
 	if (fb->pitches[0] < cfb_pitch)
 		cfb_pitch = fb->pitches[0];
@@ -153,6 +178,8 @@ static void g4x_enable_fbc(struct drm_crtc *crtc)
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 	u32 dpfc_ctl;
 
+	dev_priv->fbc.enabled = true;
+
 	dpfc_ctl = DPFC_CTL_PLANE(intel_crtc->plane) | DPFC_SR_EN;
 	if (drm_format_plane_cpp(fb->pixel_format, 0) == 2)
 		dpfc_ctl |= DPFC_CTL_LIMIT_2X;
@@ -173,6 +200,8 @@ static void g4x_disable_fbc(struct drm_device *dev)
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	u32 dpfc_ctl;
 
+	dev_priv->fbc.enabled = false;
+
 	/* Disable compression */
 	dpfc_ctl = I915_READ(DPFC_CONTROL);
 	if (dpfc_ctl & DPFC_CTL_EN) {
@@ -224,6 +253,8 @@ static void ironlake_enable_fbc(struct drm_crtc *crtc)
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 	u32 dpfc_ctl;
 
+	dev_priv->fbc.enabled = true;
+
 	dpfc_ctl = DPFC_CTL_PLANE(intel_crtc->plane);
 	if (drm_format_plane_cpp(fb->pixel_format, 0) == 2)
 		dev_priv->fbc.threshold++;
@@ -264,6 +295,8 @@ static void ironlake_disable_fbc(struct drm_device *dev)
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	u32 dpfc_ctl;
 
+	dev_priv->fbc.enabled = false;
+
 	/* Disable compression */
 	dpfc_ctl = I915_READ(ILK_DPFC_CONTROL);
 	if (dpfc_ctl & DPFC_CTL_EN) {
@@ -290,6 +323,8 @@ static void gen7_enable_fbc(struct drm_crtc *crtc)
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 	u32 dpfc_ctl;
 
+	dev_priv->fbc.enabled = true;
+
 	dpfc_ctl = IVB_DPFC_CTL_PLANE(intel_crtc->plane);
 	if (drm_format_plane_cpp(fb->pixel_format, 0) == 2)
 		dev_priv->fbc.threshold++;
@@ -309,6 +344,9 @@ static void gen7_enable_fbc(struct drm_crtc *crtc)
 
 	dpfc_ctl |= IVB_DPFC_CTL_FENCE_EN;
 
+	if (dev_priv->fbc.false_color)
+		dpfc_ctl |= FBC_CTL_FALSE_COLOR;
+
 	I915_WRITE(ILK_DPFC_CONTROL, dpfc_ctl | DPFC_CTL_EN);
 
 	if (IS_IVYBRIDGE(dev)) {
@@ -336,10 +374,20 @@ bool intel_fbc_enabled(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 
-	if (!dev_priv->display.fbc_enabled)
-		return false;
+	return dev_priv->fbc.enabled;
+}
+
+void bdw_fbc_sw_flush(struct drm_device *dev, u32 value)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+
+	if (!IS_GEN8(dev))
+		return;
+
+	if (!intel_fbc_enabled(dev))
+		return;
 
-	return dev_priv->display.fbc_enabled(dev);
+	I915_WRITE(MSG_FBC_REND_STATE, value);
 }
 
 static void intel_fbc_work_fn(struct work_struct *__work)
@@ -578,6 +626,12 @@ void intel_update_fbc(struct drm_device *dev)
 			DRM_DEBUG_KMS("framebuffer not tiled or fenced, disabling compression\n");
 		goto out_disable;
 	}
+	if (INTEL_INFO(dev)->gen <= 4 && !IS_G4X(dev) &&
+	    to_intel_plane(crtc->primary)->rotation != BIT(DRM_ROTATE_0)) {
+		if (set_no_fbc_reason(dev_priv, FBC_UNSUPPORTED_MODE))
+			DRM_DEBUG_KMS("Rotation unsupported, disabling\n");
+		goto out_disable;
+	}
 
 	/* If the kernel debugger is active, always disable compression */
 	if (in_dbg_master())
@@ -853,7 +907,7 @@ void intel_set_memory_cxsr(struct drm_i915_private *dev_priv, bool enable)
  * A value of 5us seems to be a good balance; safe for very low end
  * platforms but not overly aggressive on lower latency configs.
  */
-static const int latency_ns = 5000;
+static const int pessimal_latency_ns = 5000;
 
 static int i9xx_get_fifo_size(struct drm_device *dev, int plane)
 {
@@ -982,13 +1036,20 @@ static const struct intel_watermark_params i915_wm_info = {
 	.guard_size = 2,
 	.cacheline_size = I915_FIFO_LINE_SIZE,
 };
-static const struct intel_watermark_params i830_wm_info = {
+static const struct intel_watermark_params i830_a_wm_info = {
 	.fifo_size = I855GM_FIFO_SIZE,
 	.max_wm = I915_MAX_WM,
 	.default_wm = 1,
 	.guard_size = 2,
 	.cacheline_size = I830_FIFO_LINE_SIZE,
 };
+static const struct intel_watermark_params i830_bc_wm_info = {
+	.fifo_size = I855GM_FIFO_SIZE,
+	.max_wm = I915_MAX_WM/2,
+	.default_wm = 1,
+	.guard_size = 2,
+	.cacheline_size = I830_FIFO_LINE_SIZE,
+};
 static const struct intel_watermark_params i845_wm_info = {
 	.fifo_size = I830_FIFO_SIZE,
 	.max_wm = I915_MAX_WM,
@@ -1044,6 +1105,17 @@ static unsigned long intel_calculate_wm(unsigned long clock_in_khz,
 		wm_size = wm->max_wm;
 	if (wm_size <= 0)
 		wm_size = wm->default_wm;
+
+	/*
+	 * Bspec seems to indicate that the value shouldn't be lower than
+	 * 'burst size + 1'. Certainly 830 is quite unhappy with low values.
+	 * Lets go for 8 which is the burst size since certain platforms
+	 * already use a hardcoded 8 (which is what the spec says should be
+	 * done).
+	 */
+	if (wm_size <= 8)
+		wm_size = 8;
+
 	return wm_size;
 }
 
@@ -1268,33 +1340,32 @@ static bool g4x_compute_srwm(struct drm_device *dev,
 			      display, cursor);
 }
 
-static bool vlv_compute_drain_latency(struct drm_device *dev,
-				     int plane,
-				     int *plane_prec_mult,
-				     int *plane_dl,
-				     int *cursor_prec_mult,
-				     int *cursor_dl)
+static bool vlv_compute_drain_latency(struct drm_crtc *crtc,
+				      int pixel_size,
+				      int *prec_mult,
+				      int *drain_latency)
 {
-	struct drm_crtc *crtc;
-	int clock, pixel_size;
+	struct drm_device *dev = crtc->dev;
 	int entries;
+	int clock = to_intel_crtc(crtc)->config.adjusted_mode.crtc_clock;
 
-	crtc = intel_get_crtc_for_plane(dev, plane);
-	if (!intel_crtc_active(crtc))
+	if (WARN(clock == 0, "Pixel clock is zero!\n"))
 		return false;
 
-	clock = to_intel_crtc(crtc)->config.adjusted_mode.crtc_clock;
-	pixel_size = crtc->primary->fb->bits_per_pixel / 8;	/* BPP */
+	if (WARN(pixel_size == 0, "Pixel size is zero!\n"))
+		return false;
 
-	entries = (clock / 1000) * pixel_size;
-	*plane_prec_mult = (entries > 128) ?
-		DRAIN_LATENCY_PRECISION_64 : DRAIN_LATENCY_PRECISION_32;
-	*plane_dl = (64 * (*plane_prec_mult) * 4) / entries;
+	entries = DIV_ROUND_UP(clock, 1000) * pixel_size;
+	if (IS_CHERRYVIEW(dev))
+		*prec_mult = (entries > 128) ? DRAIN_LATENCY_PRECISION_32 :
+					       DRAIN_LATENCY_PRECISION_16;
+	else
+		*prec_mult = (entries > 128) ? DRAIN_LATENCY_PRECISION_64 :
+					       DRAIN_LATENCY_PRECISION_32;
+	*drain_latency = (64 * (*prec_mult) * 4) / entries;
 
-	entries = (clock / 1000) * 4;	/* BPP is always 4 for cursor */
-	*cursor_prec_mult = (entries > 128) ?
-		DRAIN_LATENCY_PRECISION_64 : DRAIN_LATENCY_PRECISION_32;
-	*cursor_dl = (64 * (*cursor_prec_mult) * 4) / entries;
+	if (*drain_latency > DRAIN_LATENCY_MASK)
+		*drain_latency = DRAIN_LATENCY_MASK;
 
 	return true;
 }
@@ -1307,39 +1378,51 @@ static bool vlv_compute_drain_latency(struct drm_device *dev,
  * latency value.
  */
 
-static void vlv_update_drain_latency(struct drm_device *dev)
+static void vlv_update_drain_latency(struct drm_crtc *crtc)
 {
+	struct drm_device *dev = crtc->dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	int planea_prec, planea_dl, planeb_prec, planeb_dl;
-	int cursora_prec, cursora_dl, cursorb_prec, cursorb_dl;
-	int plane_prec_mult, cursor_prec_mult; /* Precision multiplier is
-							either 16 or 32 */
+	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
+	int pixel_size;
+	int drain_latency;
+	enum pipe pipe = intel_crtc->pipe;
+	int plane_prec, prec_mult, plane_dl;
+	const int high_precision = IS_CHERRYVIEW(dev) ?
+		DRAIN_LATENCY_PRECISION_32 : DRAIN_LATENCY_PRECISION_64;
 
-	/* For plane A, Cursor A */
-	if (vlv_compute_drain_latency(dev, 0, &plane_prec_mult, &planea_dl,
-				      &cursor_prec_mult, &cursora_dl)) {
-		cursora_prec = (cursor_prec_mult == DRAIN_LATENCY_PRECISION_32) ?
-			DDL_CURSORA_PRECISION_32 : DDL_CURSORA_PRECISION_64;
-		planea_prec = (plane_prec_mult == DRAIN_LATENCY_PRECISION_32) ?
-			DDL_PLANEA_PRECISION_32 : DDL_PLANEA_PRECISION_64;
+	plane_dl = I915_READ(VLV_DDL(pipe)) & ~(DDL_PLANE_PRECISION_HIGH |
+		   DRAIN_LATENCY_MASK | DDL_CURSOR_PRECISION_HIGH |
+		   (DRAIN_LATENCY_MASK << DDL_CURSOR_SHIFT));
 
-		I915_WRITE(VLV_DDL1, cursora_prec |
-				(cursora_dl << DDL_CURSORA_SHIFT) |
-				planea_prec | planea_dl);
+	if (!intel_crtc_active(crtc)) {
+		I915_WRITE(VLV_DDL(pipe), plane_dl);
+		return;
 	}
 
-	/* For plane B, Cursor B */
-	if (vlv_compute_drain_latency(dev, 1, &plane_prec_mult, &planeb_dl,
-				      &cursor_prec_mult, &cursorb_dl)) {
-		cursorb_prec = (cursor_prec_mult == DRAIN_LATENCY_PRECISION_32) ?
-			DDL_CURSORB_PRECISION_32 : DDL_CURSORB_PRECISION_64;
-		planeb_prec = (plane_prec_mult == DRAIN_LATENCY_PRECISION_32) ?
-			DDL_PLANEB_PRECISION_32 : DDL_PLANEB_PRECISION_64;
+	/* Primary plane Drain Latency */
+	pixel_size = crtc->primary->fb->bits_per_pixel / 8;	/* BPP */
+	if (vlv_compute_drain_latency(crtc, pixel_size, &prec_mult, &drain_latency)) {
+		plane_prec = (prec_mult == high_precision) ?
+					   DDL_PLANE_PRECISION_HIGH :
+					   DDL_PLANE_PRECISION_LOW;
+		plane_dl |= plane_prec | drain_latency;
+	}
 
-		I915_WRITE(VLV_DDL2, cursorb_prec |
-				(cursorb_dl << DDL_CURSORB_SHIFT) |
-				planeb_prec | planeb_dl);
+	/* Cursor Drain Latency
+	 * BPP is always 4 for cursor
+	 */
+	pixel_size = 4;
+
+	/* Program cursor DL only if it is enabled */
+	if (intel_crtc->cursor_base &&
+	    vlv_compute_drain_latency(crtc, pixel_size, &prec_mult, &drain_latency)) {
+		plane_prec = (prec_mult == high_precision) ?
+					   DDL_CURSOR_PRECISION_HIGH :
+					   DDL_CURSOR_PRECISION_LOW;
+		plane_dl |= plane_prec | (drain_latency << DDL_CURSOR_SHIFT);
 	}
+
+	I915_WRITE(VLV_DDL(pipe), plane_dl);
 }
 
 #define single_plane_enabled(mask) is_power_of_2(mask)
@@ -1355,20 +1438,92 @@ static void valleyview_update_wm(struct drm_crtc *crtc)
 	unsigned int enabled = 0;
 	bool cxsr_enabled;
 
-	vlv_update_drain_latency(dev);
+	vlv_update_drain_latency(crtc);
+
+	if (g4x_compute_wm0(dev, PIPE_A,
+			    &valleyview_wm_info, pessimal_latency_ns,
+			    &valleyview_cursor_wm_info, pessimal_latency_ns,
+			    &planea_wm, &cursora_wm))
+		enabled |= 1 << PIPE_A;
+
+	if (g4x_compute_wm0(dev, PIPE_B,
+			    &valleyview_wm_info, pessimal_latency_ns,
+			    &valleyview_cursor_wm_info, pessimal_latency_ns,
+			    &planeb_wm, &cursorb_wm))
+		enabled |= 1 << PIPE_B;
+
+	if (single_plane_enabled(enabled) &&
+	    g4x_compute_srwm(dev, ffs(enabled) - 1,
+			     sr_latency_ns,
+			     &valleyview_wm_info,
+			     &valleyview_cursor_wm_info,
+			     &plane_sr, &ignore_cursor_sr) &&
+	    g4x_compute_srwm(dev, ffs(enabled) - 1,
+			     2*sr_latency_ns,
+			     &valleyview_wm_info,
+			     &valleyview_cursor_wm_info,
+			     &ignore_plane_sr, &cursor_sr)) {
+		cxsr_enabled = true;
+	} else {
+		cxsr_enabled = false;
+		intel_set_memory_cxsr(dev_priv, false);
+		plane_sr = cursor_sr = 0;
+	}
+
+	DRM_DEBUG_KMS("Setting FIFO watermarks - A: plane=%d, cursor=%d, "
+		      "B: plane=%d, cursor=%d, SR: plane=%d, cursor=%d\n",
+		      planea_wm, cursora_wm,
+		      planeb_wm, cursorb_wm,
+		      plane_sr, cursor_sr);
+
+	I915_WRITE(DSPFW1,
+		   (plane_sr << DSPFW_SR_SHIFT) |
+		   (cursorb_wm << DSPFW_CURSORB_SHIFT) |
+		   (planeb_wm << DSPFW_PLANEB_SHIFT) |
+		   (planea_wm << DSPFW_PLANEA_SHIFT));
+	I915_WRITE(DSPFW2,
+		   (I915_READ(DSPFW2) & ~DSPFW_CURSORA_MASK) |
+		   (cursora_wm << DSPFW_CURSORA_SHIFT));
+	I915_WRITE(DSPFW3,
+		   (I915_READ(DSPFW3) & ~DSPFW_CURSOR_SR_MASK) |
+		   (cursor_sr << DSPFW_CURSOR_SR_SHIFT));
+
+	if (cxsr_enabled)
+		intel_set_memory_cxsr(dev_priv, true);
+}
+
+static void cherryview_update_wm(struct drm_crtc *crtc)
+{
+	struct drm_device *dev = crtc->dev;
+	static const int sr_latency_ns = 12000;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	int planea_wm, planeb_wm, planec_wm;
+	int cursora_wm, cursorb_wm, cursorc_wm;
+	int plane_sr, cursor_sr;
+	int ignore_plane_sr, ignore_cursor_sr;
+	unsigned int enabled = 0;
+	bool cxsr_enabled;
+
+	vlv_update_drain_latency(crtc);
 
 	if (g4x_compute_wm0(dev, PIPE_A,
-			    &valleyview_wm_info, latency_ns,
-			    &valleyview_cursor_wm_info, latency_ns,
+			    &valleyview_wm_info, pessimal_latency_ns,
+			    &valleyview_cursor_wm_info, pessimal_latency_ns,
 			    &planea_wm, &cursora_wm))
 		enabled |= 1 << PIPE_A;
 
 	if (g4x_compute_wm0(dev, PIPE_B,
-			    &valleyview_wm_info, latency_ns,
-			    &valleyview_cursor_wm_info, latency_ns,
+			    &valleyview_wm_info, pessimal_latency_ns,
+			    &valleyview_cursor_wm_info, pessimal_latency_ns,
 			    &planeb_wm, &cursorb_wm))
 		enabled |= 1 << PIPE_B;
 
+	if (g4x_compute_wm0(dev, PIPE_C,
+			    &valleyview_wm_info, pessimal_latency_ns,
+			    &valleyview_cursor_wm_info, pessimal_latency_ns,
+			    &planec_wm, &cursorc_wm))
+		enabled |= 1 << PIPE_C;
+
 	if (single_plane_enabled(enabled) &&
 	    g4x_compute_srwm(dev, ffs(enabled) - 1,
 			     sr_latency_ns,
@@ -1387,27 +1542,68 @@ static void valleyview_update_wm(struct drm_crtc *crtc)
 		plane_sr = cursor_sr = 0;
 	}
 
-	DRM_DEBUG_KMS("Setting FIFO watermarks - A: plane=%d, cursor=%d, B: plane=%d, cursor=%d, SR: plane=%d, cursor=%d\n",
+	DRM_DEBUG_KMS("Setting FIFO watermarks - A: plane=%d, cursor=%d, "
+		      "B: plane=%d, cursor=%d, C: plane=%d, cursor=%d, "
+		      "SR: plane=%d, cursor=%d\n",
 		      planea_wm, cursora_wm,
 		      planeb_wm, cursorb_wm,
+		      planec_wm, cursorc_wm,
 		      plane_sr, cursor_sr);
 
 	I915_WRITE(DSPFW1,
 		   (plane_sr << DSPFW_SR_SHIFT) |
 		   (cursorb_wm << DSPFW_CURSORB_SHIFT) |
 		   (planeb_wm << DSPFW_PLANEB_SHIFT) |
-		   planea_wm);
+		   (planea_wm << DSPFW_PLANEA_SHIFT));
 	I915_WRITE(DSPFW2,
 		   (I915_READ(DSPFW2) & ~DSPFW_CURSORA_MASK) |
 		   (cursora_wm << DSPFW_CURSORA_SHIFT));
 	I915_WRITE(DSPFW3,
 		   (I915_READ(DSPFW3) & ~DSPFW_CURSOR_SR_MASK) |
 		   (cursor_sr << DSPFW_CURSOR_SR_SHIFT));
+	I915_WRITE(DSPFW9_CHV,
+		   (I915_READ(DSPFW9_CHV) & ~(DSPFW_PLANEC_MASK |
+					      DSPFW_CURSORC_MASK)) |
+		   (planec_wm << DSPFW_PLANEC_SHIFT) |
+		   (cursorc_wm << DSPFW_CURSORC_SHIFT));
 
 	if (cxsr_enabled)
 		intel_set_memory_cxsr(dev_priv, true);
 }
 
+static void valleyview_update_sprite_wm(struct drm_plane *plane,
+					struct drm_crtc *crtc,
+					uint32_t sprite_width,
+					uint32_t sprite_height,
+					int pixel_size,
+					bool enabled, bool scaled)
+{
+	struct drm_device *dev = crtc->dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	int pipe = to_intel_plane(plane)->pipe;
+	int sprite = to_intel_plane(plane)->plane;
+	int drain_latency;
+	int plane_prec;
+	int sprite_dl;
+	int prec_mult;
+	const int high_precision = IS_CHERRYVIEW(dev) ?
+		DRAIN_LATENCY_PRECISION_32 : DRAIN_LATENCY_PRECISION_64;
+
+	sprite_dl = I915_READ(VLV_DDL(pipe)) & ~(DDL_SPRITE_PRECISION_HIGH(sprite) |
+		    (DRAIN_LATENCY_MASK << DDL_SPRITE_SHIFT(sprite)));
+
+	if (enabled && vlv_compute_drain_latency(crtc, pixel_size, &prec_mult,
+						 &drain_latency)) {
+		plane_prec = (prec_mult == high_precision) ?
+					   DDL_SPRITE_PRECISION_HIGH(sprite) :
+					   DDL_SPRITE_PRECISION_LOW(sprite);
+		sprite_dl |= plane_prec |
+			     (drain_latency << DDL_SPRITE_SHIFT(sprite));
+	}
+
+	I915_WRITE(VLV_DDL(pipe), sprite_dl);
+}
+
 static void g4x_update_wm(struct drm_crtc *crtc)
 {
 	struct drm_device *dev = crtc->dev;
@@ -1419,14 +1615,14 @@ static void g4x_update_wm(struct drm_crtc *crtc)
 	bool cxsr_enabled;
 
 	if (g4x_compute_wm0(dev, PIPE_A,
-			    &g4x_wm_info, latency_ns,
-			    &g4x_cursor_wm_info, latency_ns,
+			    &g4x_wm_info, pessimal_latency_ns,
+			    &g4x_cursor_wm_info, pessimal_latency_ns,
 			    &planea_wm, &cursora_wm))
 		enabled |= 1 << PIPE_A;
 
 	if (g4x_compute_wm0(dev, PIPE_B,
-			    &g4x_wm_info, latency_ns,
-			    &g4x_cursor_wm_info, latency_ns,
+			    &g4x_wm_info, pessimal_latency_ns,
+			    &g4x_cursor_wm_info, pessimal_latency_ns,
 			    &planeb_wm, &cursorb_wm))
 		enabled |= 1 << PIPE_B;
 
@@ -1443,7 +1639,8 @@ static void g4x_update_wm(struct drm_crtc *crtc)
 		plane_sr = cursor_sr = 0;
 	}
 
-	DRM_DEBUG_KMS("Setting FIFO watermarks - A: plane=%d, cursor=%d, B: plane=%d, cursor=%d, SR: plane=%d, cursor=%d\n",
+	DRM_DEBUG_KMS("Setting FIFO watermarks - A: plane=%d, cursor=%d, "
+		      "B: plane=%d, cursor=%d, SR: plane=%d, cursor=%d\n",
 		      planea_wm, cursora_wm,
 		      planeb_wm, cursorb_wm,
 		      plane_sr, cursor_sr);
@@ -1452,7 +1649,7 @@ static void g4x_update_wm(struct drm_crtc *crtc)
 		   (plane_sr << DSPFW_SR_SHIFT) |
 		   (cursorb_wm << DSPFW_CURSORB_SHIFT) |
 		   (planeb_wm << DSPFW_PLANEB_SHIFT) |
-		   planea_wm);
+		   (planea_wm << DSPFW_PLANEA_SHIFT));
 	I915_WRITE(DSPFW2,
 		   (I915_READ(DSPFW2) & ~DSPFW_CURSORA_MASK) |
 		   (cursora_wm << DSPFW_CURSORA_SHIFT));
@@ -1526,8 +1723,11 @@ static void i965_update_wm(struct drm_crtc *unused_crtc)
 
 	/* 965 has limitations... */
 	I915_WRITE(DSPFW1, (srwm << DSPFW_SR_SHIFT) |
-		   (8 << 16) | (8 << 8) | (8 << 0));
-	I915_WRITE(DSPFW2, (8 << 8) | (8 << 0));
+		   (8 << DSPFW_CURSORB_SHIFT) |
+		   (8 << DSPFW_PLANEB_SHIFT) |
+		   (8 << DSPFW_PLANEA_SHIFT));
+	I915_WRITE(DSPFW2, (8 << DSPFW_CURSORA_SHIFT) |
+		   (8 << DSPFW_PLANEC_SHIFT_OLD));
 	/* update cursor SR watermark */
 	I915_WRITE(DSPFW3, (cursor_sr << DSPFW_CURSOR_SR_SHIFT));
 
@@ -1552,7 +1752,7 @@ static void i9xx_update_wm(struct drm_crtc *unused_crtc)
 	else if (!IS_GEN2(dev))
 		wm_info = &i915_wm_info;
 	else
-		wm_info = &i830_wm_info;
+		wm_info = &i830_a_wm_info;
 
 	fifo_size = dev_priv->display.get_fifo_size(dev, 0);
 	crtc = intel_get_crtc_for_plane(dev, 0);
@@ -1565,10 +1765,16 @@ static void i9xx_update_wm(struct drm_crtc *unused_crtc)
 		adjusted_mode = &to_intel_crtc(crtc)->config.adjusted_mode;
 		planea_wm = intel_calculate_wm(adjusted_mode->crtc_clock,
 					       wm_info, fifo_size, cpp,
-					       latency_ns);
+					       pessimal_latency_ns);
 		enabled = crtc;
-	} else
+	} else {
 		planea_wm = fifo_size - wm_info->guard_size;
+		if (planea_wm > (long)wm_info->max_wm)
+			planea_wm = wm_info->max_wm;
+	}
+
+	if (IS_GEN2(dev))
+		wm_info = &i830_bc_wm_info;
 
 	fifo_size = dev_priv->display.get_fifo_size(dev, 1);
 	crtc = intel_get_crtc_for_plane(dev, 1);
@@ -1581,13 +1787,16 @@ static void i9xx_update_wm(struct drm_crtc *unused_crtc)
 		adjusted_mode = &to_intel_crtc(crtc)->config.adjusted_mode;
 		planeb_wm = intel_calculate_wm(adjusted_mode->crtc_clock,
 					       wm_info, fifo_size, cpp,
-					       latency_ns);
+					       pessimal_latency_ns);
 		if (enabled == NULL)
 			enabled = crtc;
 		else
 			enabled = NULL;
-	} else
+	} else {
 		planeb_wm = fifo_size - wm_info->guard_size;
+		if (planeb_wm > (long)wm_info->max_wm)
+			planeb_wm = wm_info->max_wm;
+	}
 
 	DRM_DEBUG_KMS("FIFO watermarks - A: %d, B: %d\n", planea_wm, planeb_wm);
 
@@ -1674,7 +1883,7 @@ static void i845_update_wm(struct drm_crtc *unused_crtc)
 	planea_wm = intel_calculate_wm(adjusted_mode->crtc_clock,
 				       &i845_wm_info,
 				       dev_priv->display.get_fifo_size(dev, 0),
-				       4, latency_ns);
+				       4, pessimal_latency_ns);
 	fwater_lo = I915_READ(FW_BLC) & ~0xfff;
 	fwater_lo |= (3<<8) | planea_wm;
 
@@ -1751,6 +1960,14 @@ static uint32_t ilk_wm_fbc(uint32_t pri_val, uint32_t horiz_pixels,
 	return DIV_ROUND_UP(pri_val * 64, horiz_pixels * bytes_per_pixel) + 2;
 }
 
+struct skl_pipe_wm_parameters {
+	bool active;
+	uint32_t pipe_htotal;
+	uint32_t pixel_rate; /* in KHz */
+	struct intel_plane_wm_parameters plane[I915_MAX_PLANES];
+	struct intel_plane_wm_parameters cursor;
+};
+
 struct ilk_pipe_wm_parameters {
 	bool active;
 	uint32_t pipe_htotal;
@@ -2062,11 +2279,82 @@ hsw_compute_linetime_wm(struct drm_device *dev, struct drm_crtc *crtc)
 	       PIPE_WM_LINETIME_TIME(linetime);
 }
 
-static void intel_read_wm_latency(struct drm_device *dev, uint16_t wm[5])
+static void intel_read_wm_latency(struct drm_device *dev, uint16_t wm[8])
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 
-	if (IS_HASWELL(dev) || IS_BROADWELL(dev)) {
+	if (IS_GEN9(dev)) {
+		uint32_t val;
+		int ret, i;
+		int level, max_level = ilk_wm_max_level(dev);
+
+		/* read the first set of memory latencies[0:3] */
+		val = 0; /* data0 to be programmed to 0 for first set */
+		mutex_lock(&dev_priv->rps.hw_lock);
+		ret = sandybridge_pcode_read(dev_priv,
+					     GEN9_PCODE_READ_MEM_LATENCY,
+					     &val);
+		mutex_unlock(&dev_priv->rps.hw_lock);
+
+		if (ret) {
+			DRM_ERROR("SKL Mailbox read error = %d\n", ret);
+			return;
+		}
+
+		wm[0] = val & GEN9_MEM_LATENCY_LEVEL_MASK;
+		wm[1] = (val >> GEN9_MEM_LATENCY_LEVEL_1_5_SHIFT) &
+				GEN9_MEM_LATENCY_LEVEL_MASK;
+		wm[2] = (val >> GEN9_MEM_LATENCY_LEVEL_2_6_SHIFT) &
+				GEN9_MEM_LATENCY_LEVEL_MASK;
+		wm[3] = (val >> GEN9_MEM_LATENCY_LEVEL_3_7_SHIFT) &
+				GEN9_MEM_LATENCY_LEVEL_MASK;
+
+		/* read the second set of memory latencies[4:7] */
+		val = 1; /* data0 to be programmed to 1 for second set */
+		mutex_lock(&dev_priv->rps.hw_lock);
+		ret = sandybridge_pcode_read(dev_priv,
+					     GEN9_PCODE_READ_MEM_LATENCY,
+					     &val);
+		mutex_unlock(&dev_priv->rps.hw_lock);
+		if (ret) {
+			DRM_ERROR("SKL Mailbox read error = %d\n", ret);
+			return;
+		}
+
+		wm[4] = val & GEN9_MEM_LATENCY_LEVEL_MASK;
+		wm[5] = (val >> GEN9_MEM_LATENCY_LEVEL_1_5_SHIFT) &
+				GEN9_MEM_LATENCY_LEVEL_MASK;
+		wm[6] = (val >> GEN9_MEM_LATENCY_LEVEL_2_6_SHIFT) &
+				GEN9_MEM_LATENCY_LEVEL_MASK;
+		wm[7] = (val >> GEN9_MEM_LATENCY_LEVEL_3_7_SHIFT) &
+				GEN9_MEM_LATENCY_LEVEL_MASK;
+
+		/*
+		 * punit doesn't take into account the read latency so we need
+		 * to add 2us to the various latency levels we retrieve from
+		 * the punit.
+		 *   - W0 is a bit special in that it's the only level that
+		 *   can't be disabled if we want to have display working, so
+		 *   we always add 2us there.
+		 *   - For levels >=1, punit returns 0us latency when they are
+		 *   disabled, so we respect that and don't add 2us then
+		 *
+		 * Additionally, if a level n (n > 1) has a 0us latency, all
+		 * levels m (m >= n) need to be disabled. We make sure to
+		 * sanitize the values out of the punit to satisfy this
+		 * requirement.
+		 */
+		wm[0] += 2;
+		for (level = 1; level <= max_level; level++)
+			if (wm[level] != 0)
+				wm[level] += 2;
+			else {
+				for (i = level + 1; i <= max_level; i++)
+					wm[i] = 0;
+
+				break;
+			}
+	} else if (IS_HASWELL(dev) || IS_BROADWELL(dev)) {
 		uint64_t sskpd = I915_READ64(MCH_SSKPD);
 
 		wm[0] = (sskpd >> 56) & 0xFF;
@@ -2114,7 +2402,9 @@ static void intel_fixup_cur_wm_latency(struct drm_device *dev, uint16_t wm[5])
 int ilk_wm_max_level(const struct drm_device *dev)
 {
 	/* how many WM levels are we expecting */
-	if (IS_HASWELL(dev) || IS_BROADWELL(dev))
+	if (IS_GEN9(dev))
+		return 7;
+	else if (IS_HASWELL(dev) || IS_BROADWELL(dev))
 		return 4;
 	else if (INTEL_INFO(dev)->gen >= 6)
 		return 3;
@@ -2124,7 +2414,7 @@ int ilk_wm_max_level(const struct drm_device *dev)
 
 static void intel_print_wm_latency(struct drm_device *dev,
 				   const char *name,
-				   const uint16_t wm[5])
+				   const uint16_t wm[8])
 {
 	int level, max_level = ilk_wm_max_level(dev);
 
@@ -2137,8 +2427,13 @@ static void intel_print_wm_latency(struct drm_device *dev,
 			continue;
 		}
 
-		/* WM1+ latency values in 0.5us units */
-		if (level > 0)
+		/*
+		 * - latencies are in us on gen9.
+		 * - before then, WM1+ latency values are in 0.5us units
+		 */
+		if (IS_GEN9(dev))
+			latency *= 10;
+		else if (level > 0)
 			latency *= 5;
 
 		DRM_DEBUG_KMS("%s WM%d latency %u (%u.%u usec)\n",
@@ -2206,6 +2501,14 @@ static void ilk_setup_wm_latency(struct drm_device *dev)
 		snb_wm_latency_quirk(dev);
 }
 
+static void skl_setup_wm_latency(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+
+	intel_read_wm_latency(dev, dev_priv->wm.skl_latency);
+	intel_print_wm_latency(dev, "Gen9 Plane", dev_priv->wm.skl_latency);
+}
+
 static void ilk_compute_wm_parameters(struct drm_crtc *crtc,
 				      struct ilk_pipe_wm_parameters *p)
 {
@@ -2527,7 +2830,7 @@ static struct intel_pipe_wm *ilk_find_best_result(struct drm_device *dev,
 #define WM_DIRTY_FBC (1 << 24)
 #define WM_DIRTY_DDB (1 << 25)
 
-static unsigned int ilk_compute_wm_dirty(struct drm_device *dev,
+static unsigned int ilk_compute_wm_dirty(struct drm_i915_private *dev_priv,
 					 const struct ilk_wm_values *old,
 					 const struct ilk_wm_values *new)
 {
@@ -2535,7 +2838,7 @@ static unsigned int ilk_compute_wm_dirty(struct drm_device *dev,
 	enum pipe pipe;
 	int wm_lp;
 
-	for_each_pipe(pipe) {
+	for_each_pipe(dev_priv, pipe) {
 		if (old->wm_linetime[pipe] != new->wm_linetime[pipe]) {
 			dirty |= WM_DIRTY_LINETIME(pipe);
 			/* Must disable LP1+ watermarks too */
@@ -2621,7 +2924,7 @@ static void ilk_write_wm_values(struct drm_i915_private *dev_priv,
 	unsigned int dirty;
 	uint32_t val;
 
-	dirty = ilk_compute_wm_dirty(dev, previous, results);
+	dirty = ilk_compute_wm_dirty(dev_priv, previous, results);
 	if (!dirty)
 		return;
 
@@ -2696,6 +2999,769 @@ static bool ilk_disable_lp_wm(struct drm_device *dev)
 	return _ilk_disable_lp_wm(dev_priv, WM_DIRTY_LP_ALL);
 }
 
+/*
+ * On gen9, we need to allocate Display Data Buffer (DDB) portions to the
+ * different active planes.
+ */
+
+#define SKL_DDB_SIZE		896	/* in blocks */
+
+static void
+skl_ddb_get_pipe_allocation_limits(struct drm_device *dev,
+				   struct drm_crtc *for_crtc,
+				   const struct intel_wm_config *config,
+				   const struct skl_pipe_wm_parameters *params,
+				   struct skl_ddb_entry *alloc /* out */)
+{
+	struct drm_crtc *crtc;
+	unsigned int pipe_size, ddb_size;
+	int nth_active_pipe;
+
+	if (!params->active) {
+		alloc->start = 0;
+		alloc->end = 0;
+		return;
+	}
+
+	ddb_size = SKL_DDB_SIZE;
+
+	ddb_size -= 4; /* 4 blocks for bypass path allocation */
+
+	nth_active_pipe = 0;
+	for_each_crtc(dev, crtc) {
+		if (!intel_crtc_active(crtc))
+			continue;
+
+		if (crtc == for_crtc)
+			break;
+
+		nth_active_pipe++;
+	}
+
+	pipe_size = ddb_size / config->num_pipes_active;
+	alloc->start = nth_active_pipe * ddb_size / config->num_pipes_active;
+	alloc->end = alloc->start + pipe_size;
+}
+
+static unsigned int skl_cursor_allocation(const struct intel_wm_config *config)
+{
+	if (config->num_pipes_active == 1)
+		return 32;
+
+	return 8;
+}
+
+static void skl_ddb_entry_init_from_hw(struct skl_ddb_entry *entry, u32 reg)
+{
+	entry->start = reg & 0x3ff;
+	entry->end = (reg >> 16) & 0x3ff;
+	if (entry->end)
+		entry->end += 1;
+}
+
+void skl_ddb_get_hw_state(struct drm_i915_private *dev_priv,
+			  struct skl_ddb_allocation *ddb /* out */)
+{
+	struct drm_device *dev = dev_priv->dev;
+	enum pipe pipe;
+	int plane;
+	u32 val;
+
+	for_each_pipe(dev_priv, pipe) {
+		for_each_plane(pipe, plane) {
+			val = I915_READ(PLANE_BUF_CFG(pipe, plane));
+			skl_ddb_entry_init_from_hw(&ddb->plane[pipe][plane],
+						   val);
+		}
+
+		val = I915_READ(CUR_BUF_CFG(pipe));
+		skl_ddb_entry_init_from_hw(&ddb->cursor[pipe], val);
+	}
+}
+
+static unsigned int
+skl_plane_relative_data_rate(const struct intel_plane_wm_parameters *p)
+{
+	return p->horiz_pixels * p->vert_pixels * p->bytes_per_pixel;
+}
+
+/*
+ * We don't overflow 32 bits. Worst case is 3 planes enabled, each fetching
+ * a 8192x4096@32bpp framebuffer:
+ *   3 * 4096 * 8192  * 4 < 2^32
+ */
+static unsigned int
+skl_get_total_relative_data_rate(struct intel_crtc *intel_crtc,
+				 const struct skl_pipe_wm_parameters *params)
+{
+	unsigned int total_data_rate = 0;
+	int plane;
+
+	for (plane = 0; plane < intel_num_planes(intel_crtc); plane++) {
+		const struct intel_plane_wm_parameters *p;
+
+		p = &params->plane[plane];
+		if (!p->enabled)
+			continue;
+
+		total_data_rate += skl_plane_relative_data_rate(p);
+	}
+
+	return total_data_rate;
+}
+
+static void
+skl_allocate_pipe_ddb(struct drm_crtc *crtc,
+		      const struct intel_wm_config *config,
+		      const struct skl_pipe_wm_parameters *params,
+		      struct skl_ddb_allocation *ddb /* out */)
+{
+	struct drm_device *dev = crtc->dev;
+	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
+	enum pipe pipe = intel_crtc->pipe;
+	struct skl_ddb_entry *alloc = &ddb->pipe[pipe];
+	uint16_t alloc_size, start, cursor_blocks;
+	unsigned int total_data_rate;
+	int plane;
+
+	skl_ddb_get_pipe_allocation_limits(dev, crtc, config, params, alloc);
+	alloc_size = skl_ddb_entry_size(alloc);
+	if (alloc_size == 0) {
+		memset(ddb->plane[pipe], 0, sizeof(ddb->plane[pipe]));
+		memset(&ddb->cursor[pipe], 0, sizeof(ddb->cursor[pipe]));
+		return;
+	}
+
+	cursor_blocks = skl_cursor_allocation(config);
+	ddb->cursor[pipe].start = alloc->end - cursor_blocks;
+	ddb->cursor[pipe].end = alloc->end;
+
+	alloc_size -= cursor_blocks;
+	alloc->end -= cursor_blocks;
+
+	/*
+	 * Each active plane get a portion of the remaining space, in
+	 * proportion to the amount of data they need to fetch from memory.
+	 *
+	 * FIXME: we may not allocate every single block here.
+	 */
+	total_data_rate = skl_get_total_relative_data_rate(intel_crtc, params);
+
+	start = alloc->start;
+	for (plane = 0; plane < intel_num_planes(intel_crtc); plane++) {
+		const struct intel_plane_wm_parameters *p;
+		unsigned int data_rate;
+		uint16_t plane_blocks;
+
+		p = &params->plane[plane];
+		if (!p->enabled)
+			continue;
+
+		data_rate = skl_plane_relative_data_rate(p);
+
+		/*
+		 * promote the expression to 64 bits to avoid overflowing, the
+		 * result is < available as data_rate / total_data_rate < 1
+		 */
+		plane_blocks = div_u64((uint64_t)alloc_size * data_rate,
+				       total_data_rate);
+
+		ddb->plane[pipe][plane].start = start;
+		ddb->plane[pipe][plane].end = start + plane_blocks;
+
+		start += plane_blocks;
+	}
+
+}
+
+static uint32_t skl_pipe_pixel_rate(const struct intel_crtc_config *config)
+{
+	/* TODO: Take into account the scalers once we support them */
+	return config->adjusted_mode.crtc_clock;
+}
+
+/*
+ * The max latency should be 257 (max the punit can code is 255 and we add 2us
+ * for the read latency) and bytes_per_pixel should always be <= 8, so that
+ * should allow pixel_rate up to ~2 GHz which seems sufficient since max
+ * 2xcdclk is 1350 MHz and the pixel rate should never exceed that.
+*/
+static uint32_t skl_wm_method1(uint32_t pixel_rate, uint8_t bytes_per_pixel,
+			       uint32_t latency)
+{
+	uint32_t wm_intermediate_val, ret;
+
+	if (latency == 0)
+		return UINT_MAX;
+
+	wm_intermediate_val = latency * pixel_rate * bytes_per_pixel;
+	ret = DIV_ROUND_UP(wm_intermediate_val, 1000);
+
+	return ret;
+}
+
+static uint32_t skl_wm_method2(uint32_t pixel_rate, uint32_t pipe_htotal,
+			       uint32_t horiz_pixels, uint8_t bytes_per_pixel,
+			       uint32_t latency)
+{
+	uint32_t ret, plane_bytes_per_line, wm_intermediate_val;
+
+	if (latency == 0)
+		return UINT_MAX;
+
+	plane_bytes_per_line = horiz_pixels * bytes_per_pixel;
+	wm_intermediate_val = latency * pixel_rate;
+	ret = DIV_ROUND_UP(wm_intermediate_val, pipe_htotal * 1000) *
+				plane_bytes_per_line;
+
+	return ret;
+}
+
+static bool skl_ddb_allocation_changed(const struct skl_ddb_allocation *new_ddb,
+				       const struct intel_crtc *intel_crtc)
+{
+	struct drm_device *dev = intel_crtc->base.dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	const struct skl_ddb_allocation *cur_ddb = &dev_priv->wm.skl_hw.ddb;
+	enum pipe pipe = intel_crtc->pipe;
+
+	if (memcmp(new_ddb->plane[pipe], cur_ddb->plane[pipe],
+		   sizeof(new_ddb->plane[pipe])))
+		return true;
+
+	if (memcmp(&new_ddb->cursor[pipe], &cur_ddb->cursor[pipe],
+		    sizeof(new_ddb->cursor[pipe])))
+		return true;
+
+	return false;
+}
+
+static void skl_compute_wm_global_parameters(struct drm_device *dev,
+					     struct intel_wm_config *config)
+{
+	struct drm_crtc *crtc;
+	struct drm_plane *plane;
+
+	list_for_each_entry(crtc, &dev->mode_config.crtc_list, head)
+		config->num_pipes_active += intel_crtc_active(crtc);
+
+	/* FIXME: I don't think we need those two global parameters on SKL */
+	list_for_each_entry(plane, &dev->mode_config.plane_list, head) {
+		struct intel_plane *intel_plane = to_intel_plane(plane);
+
+		config->sprites_enabled |= intel_plane->wm.enabled;
+		config->sprites_scaled |= intel_plane->wm.scaled;
+	}
+}
+
+static void skl_compute_wm_pipe_parameters(struct drm_crtc *crtc,
+					   struct skl_pipe_wm_parameters *p)
+{
+	struct drm_device *dev = crtc->dev;
+	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
+	enum pipe pipe = intel_crtc->pipe;
+	struct drm_plane *plane;
+	int i = 1; /* Index for sprite planes start */
+
+	p->active = intel_crtc_active(crtc);
+	if (p->active) {
+		p->pipe_htotal = intel_crtc->config.adjusted_mode.crtc_htotal;
+		p->pixel_rate = skl_pipe_pixel_rate(&intel_crtc->config);
+
+		/*
+		 * For now, assume primary and cursor planes are always enabled.
+		 */
+		p->plane[0].enabled = true;
+		p->plane[0].bytes_per_pixel =
+			crtc->primary->fb->bits_per_pixel / 8;
+		p->plane[0].horiz_pixels = intel_crtc->config.pipe_src_w;
+		p->plane[0].vert_pixels = intel_crtc->config.pipe_src_h;
+
+		p->cursor.enabled = true;
+		p->cursor.bytes_per_pixel = 4;
+		p->cursor.horiz_pixels = intel_crtc->cursor_width ?
+					 intel_crtc->cursor_width : 64;
+	}
+
+	list_for_each_entry(plane, &dev->mode_config.plane_list, head) {
+		struct intel_plane *intel_plane = to_intel_plane(plane);
+
+		if (intel_plane->pipe == pipe)
+			p->plane[i++] = intel_plane->wm;
+	}
+}
+
+static bool skl_compute_plane_wm(struct skl_pipe_wm_parameters *p,
+				 struct intel_plane_wm_parameters *p_params,
+				 uint16_t ddb_allocation,
+				 uint32_t mem_value,
+				 uint16_t *out_blocks, /* out */
+				 uint8_t *out_lines /* out */)
+{
+	uint32_t method1, method2, plane_bytes_per_line, res_blocks, res_lines;
+	uint32_t result_bytes;
+
+	if (mem_value == 0 || !p->active || !p_params->enabled)
+		return false;
+
+	method1 = skl_wm_method1(p->pixel_rate,
+				 p_params->bytes_per_pixel,
+				 mem_value);
+	method2 = skl_wm_method2(p->pixel_rate,
+				 p->pipe_htotal,
+				 p_params->horiz_pixels,
+				 p_params->bytes_per_pixel,
+				 mem_value);
+
+	plane_bytes_per_line = p_params->horiz_pixels *
+					p_params->bytes_per_pixel;
+
+	/* For now xtile and linear */
+	if (((ddb_allocation * 512) / plane_bytes_per_line) >= 1)
+		result_bytes = min(method1, method2);
+	else
+		result_bytes = method1;
+
+	res_blocks = DIV_ROUND_UP(result_bytes, 512) + 1;
+	res_lines = DIV_ROUND_UP(result_bytes, plane_bytes_per_line);
+
+	if (res_blocks > ddb_allocation || res_lines > 31)
+		return false;
+
+	*out_blocks = res_blocks;
+	*out_lines = res_lines;
+
+	return true;
+}
+
+static void skl_compute_wm_level(const struct drm_i915_private *dev_priv,
+				 struct skl_ddb_allocation *ddb,
+				 struct skl_pipe_wm_parameters *p,
+				 enum pipe pipe,
+				 int level,
+				 int num_planes,
+				 struct skl_wm_level *result)
+{
+	uint16_t latency = dev_priv->wm.skl_latency[level];
+	uint16_t ddb_blocks;
+	int i;
+
+	for (i = 0; i < num_planes; i++) {
+		ddb_blocks = skl_ddb_entry_size(&ddb->plane[pipe][i]);
+
+		result->plane_en[i] = skl_compute_plane_wm(p, &p->plane[i],
+						ddb_blocks,
+						latency,
+						&result->plane_res_b[i],
+						&result->plane_res_l[i]);
+	}
+
+	ddb_blocks = skl_ddb_entry_size(&ddb->cursor[pipe]);
+	result->cursor_en = skl_compute_plane_wm(p, &p->cursor, ddb_blocks,
+						 latency, &result->cursor_res_b,
+						 &result->cursor_res_l);
+}
+
+static uint32_t
+skl_compute_linetime_wm(struct drm_crtc *crtc, struct skl_pipe_wm_parameters *p)
+{
+	if (!intel_crtc_active(crtc))
+		return 0;
+
+	return DIV_ROUND_UP(8 * p->pipe_htotal * 1000, p->pixel_rate);
+
+}
+
+static void skl_compute_transition_wm(struct drm_crtc *crtc,
+				      struct skl_pipe_wm_parameters *params,
+				      struct skl_wm_level *trans_wm /* out */)
+{
+	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
+	int i;
+
+	if (!params->active)
+		return;
+
+	/* Until we know more, just disable transition WMs */
+	for (i = 0; i < intel_num_planes(intel_crtc); i++)
+		trans_wm->plane_en[i] = false;
+	trans_wm->cursor_en = false;
+}
+
+static void skl_compute_pipe_wm(struct drm_crtc *crtc,
+				struct skl_ddb_allocation *ddb,
+				struct skl_pipe_wm_parameters *params,
+				struct skl_pipe_wm *pipe_wm)
+{
+	struct drm_device *dev = crtc->dev;
+	const struct drm_i915_private *dev_priv = dev->dev_private;
+	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
+	int level, max_level = ilk_wm_max_level(dev);
+
+	for (level = 0; level <= max_level; level++) {
+		skl_compute_wm_level(dev_priv, ddb, params, intel_crtc->pipe,
+				     level, intel_num_planes(intel_crtc),
+				     &pipe_wm->wm[level]);
+	}
+	pipe_wm->linetime = skl_compute_linetime_wm(crtc, params);
+
+	skl_compute_transition_wm(crtc, params, &pipe_wm->trans_wm);
+}
+
+static void skl_compute_wm_results(struct drm_device *dev,
+				   struct skl_pipe_wm_parameters *p,
+				   struct skl_pipe_wm *p_wm,
+				   struct skl_wm_values *r,
+				   struct intel_crtc *intel_crtc)
+{
+	int level, max_level = ilk_wm_max_level(dev);
+	enum pipe pipe = intel_crtc->pipe;
+	uint32_t temp;
+	int i;
+
+	for (level = 0; level <= max_level; level++) {
+		for (i = 0; i < intel_num_planes(intel_crtc); i++) {
+			temp = 0;
+
+			temp |= p_wm->wm[level].plane_res_l[i] <<
+					PLANE_WM_LINES_SHIFT;
+			temp |= p_wm->wm[level].plane_res_b[i];
+			if (p_wm->wm[level].plane_en[i])
+				temp |= PLANE_WM_EN;
+
+			r->plane[pipe][i][level] = temp;
+		}
+
+		temp = 0;
+
+		temp |= p_wm->wm[level].cursor_res_l << PLANE_WM_LINES_SHIFT;
+		temp |= p_wm->wm[level].cursor_res_b;
+
+		if (p_wm->wm[level].cursor_en)
+			temp |= PLANE_WM_EN;
+
+		r->cursor[pipe][level] = temp;
+
+	}
+
+	/* transition WMs */
+	for (i = 0; i < intel_num_planes(intel_crtc); i++) {
+		temp = 0;
+		temp |= p_wm->trans_wm.plane_res_l[i] << PLANE_WM_LINES_SHIFT;
+		temp |= p_wm->trans_wm.plane_res_b[i];
+		if (p_wm->trans_wm.plane_en[i])
+			temp |= PLANE_WM_EN;
+
+		r->plane_trans[pipe][i] = temp;
+	}
+
+	temp = 0;
+	temp |= p_wm->trans_wm.cursor_res_l << PLANE_WM_LINES_SHIFT;
+	temp |= p_wm->trans_wm.cursor_res_b;
+	if (p_wm->trans_wm.cursor_en)
+		temp |= PLANE_WM_EN;
+
+	r->cursor_trans[pipe] = temp;
+
+	r->wm_linetime[pipe] = p_wm->linetime;
+}
+
+static void skl_ddb_entry_write(struct drm_i915_private *dev_priv, uint32_t reg,
+				const struct skl_ddb_entry *entry)
+{
+	if (entry->end)
+		I915_WRITE(reg, (entry->end - 1) << 16 | entry->start);
+	else
+		I915_WRITE(reg, 0);
+}
+
+static void skl_write_wm_values(struct drm_i915_private *dev_priv,
+				const struct skl_wm_values *new)
+{
+	struct drm_device *dev = dev_priv->dev;
+	struct intel_crtc *crtc;
+
+	list_for_each_entry(crtc, &dev->mode_config.crtc_list, base.head) {
+		int i, level, max_level = ilk_wm_max_level(dev);
+		enum pipe pipe = crtc->pipe;
+
+		if (!new->dirty[pipe])
+			continue;
+
+		I915_WRITE(PIPE_WM_LINETIME(pipe), new->wm_linetime[pipe]);
+
+		for (level = 0; level <= max_level; level++) {
+			for (i = 0; i < intel_num_planes(crtc); i++)
+				I915_WRITE(PLANE_WM(pipe, i, level),
+					   new->plane[pipe][i][level]);
+			I915_WRITE(CUR_WM(pipe, level),
+				   new->cursor[pipe][level]);
+		}
+		for (i = 0; i < intel_num_planes(crtc); i++)
+			I915_WRITE(PLANE_WM_TRANS(pipe, i),
+				   new->plane_trans[pipe][i]);
+		I915_WRITE(CUR_WM_TRANS(pipe), new->cursor_trans[pipe]);
+
+		for (i = 0; i < intel_num_planes(crtc); i++)
+			skl_ddb_entry_write(dev_priv,
+					    PLANE_BUF_CFG(pipe, i),
+					    &new->ddb.plane[pipe][i]);
+
+		skl_ddb_entry_write(dev_priv, CUR_BUF_CFG(pipe),
+				    &new->ddb.cursor[pipe]);
+	}
+}
+
+/*
+ * When setting up a new DDB allocation arrangement, we need to correctly
+ * sequence the times at which the new allocations for the pipes are taken into
+ * account or we'll have pipes fetching from space previously allocated to
+ * another pipe.
+ *
+ * Roughly the sequence looks like:
+ *  1. re-allocate the pipe(s) with the allocation being reduced and not
+ *     overlapping with a previous light-up pipe (another way to put it is:
+ *     pipes with their new allocation strickly included into their old ones).
+ *  2. re-allocate the other pipes that get their allocation reduced
+ *  3. allocate the pipes having their allocation increased
+ *
+ * Steps 1. and 2. are here to take care of the following case:
+ * - Initially DDB looks like this:
+ *     |   B    |   C    |
+ * - enable pipe A.
+ * - pipe B has a reduced DDB allocation that overlaps with the old pipe C
+ *   allocation
+ *     |  A  |  B  |  C  |
+ *
+ * We need to sequence the re-allocation: C, B, A (and not B, C, A).
+ */
+
+static void
+skl_wm_flush_pipe(struct drm_i915_private *dev_priv, enum pipe pipe, int pass)
+{
+	struct drm_device *dev = dev_priv->dev;
+	int plane;
+
+	DRM_DEBUG_KMS("flush pipe %c (pass %d)\n", pipe_name(pipe), pass);
+
+	for_each_plane(pipe, plane) {
+		I915_WRITE(PLANE_SURF(pipe, plane),
+			   I915_READ(PLANE_SURF(pipe, plane)));
+	}
+	I915_WRITE(CURBASE(pipe), I915_READ(CURBASE(pipe)));
+}
+
+static bool
+skl_ddb_allocation_included(const struct skl_ddb_allocation *old,
+			    const struct skl_ddb_allocation *new,
+			    enum pipe pipe)
+{
+	uint16_t old_size, new_size;
+
+	old_size = skl_ddb_entry_size(&old->pipe[pipe]);
+	new_size = skl_ddb_entry_size(&new->pipe[pipe]);
+
+	return old_size != new_size &&
+	       new->pipe[pipe].start >= old->pipe[pipe].start &&
+	       new->pipe[pipe].end <= old->pipe[pipe].end;
+}
+
+static void skl_flush_wm_values(struct drm_i915_private *dev_priv,
+				struct skl_wm_values *new_values)
+{
+	struct drm_device *dev = dev_priv->dev;
+	struct skl_ddb_allocation *cur_ddb, *new_ddb;
+	bool reallocated[I915_MAX_PIPES] = {false, false, false};
+	struct intel_crtc *crtc;
+	enum pipe pipe;
+
+	new_ddb = &new_values->ddb;
+	cur_ddb = &dev_priv->wm.skl_hw.ddb;
+
+	/*
+	 * First pass: flush the pipes with the new allocation contained into
+	 * the old space.
+	 *
+	 * We'll wait for the vblank on those pipes to ensure we can safely
+	 * re-allocate the freed space without this pipe fetching from it.
+	 */
+	for_each_intel_crtc(dev, crtc) {
+		if (!crtc->active)
+			continue;
+
+		pipe = crtc->pipe;
+
+		if (!skl_ddb_allocation_included(cur_ddb, new_ddb, pipe))
+			continue;
+
+		skl_wm_flush_pipe(dev_priv, pipe, 1);
+		intel_wait_for_vblank(dev, pipe);
+
+		reallocated[pipe] = true;
+	}
+
+
+	/*
+	 * Second pass: flush the pipes that are having their allocation
+	 * reduced, but overlapping with a previous allocation.
+	 *
+	 * Here as well we need to wait for the vblank to make sure the freed
+	 * space is not used anymore.
+	 */
+	for_each_intel_crtc(dev, crtc) {
+		if (!crtc->active)
+			continue;
+
+		pipe = crtc->pipe;
+
+		if (reallocated[pipe])
+			continue;
+
+		if (skl_ddb_entry_size(&new_ddb->pipe[pipe]) <
+		    skl_ddb_entry_size(&cur_ddb->pipe[pipe])) {
+			skl_wm_flush_pipe(dev_priv, pipe, 2);
+			intel_wait_for_vblank(dev, pipe);
+		}
+
+		reallocated[pipe] = true;
+	}
+
+	/*
+	 * Third pass: flush the pipes that got more space allocated.
+	 *
+	 * We don't need to actively wait for the update here, next vblank
+	 * will just get more DDB space with the correct WM values.
+	 */
+	for_each_intel_crtc(dev, crtc) {
+		if (!crtc->active)
+			continue;
+
+		pipe = crtc->pipe;
+
+		/*
+		 * At this point, only the pipes more space than before are
+		 * left to re-allocate.
+		 */
+		if (reallocated[pipe])
+			continue;
+
+		skl_wm_flush_pipe(dev_priv, pipe, 3);
+	}
+}
+
+static bool skl_update_pipe_wm(struct drm_crtc *crtc,
+			       struct skl_pipe_wm_parameters *params,
+			       struct intel_wm_config *config,
+			       struct skl_ddb_allocation *ddb, /* out */
+			       struct skl_pipe_wm *pipe_wm /* out */)
+{
+	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
+
+	skl_compute_wm_pipe_parameters(crtc, params);
+	skl_allocate_pipe_ddb(crtc, config, params, ddb);
+	skl_compute_pipe_wm(crtc, ddb, params, pipe_wm);
+
+	if (!memcmp(&intel_crtc->wm.skl_active, pipe_wm, sizeof(*pipe_wm)))
+		return false;
+
+	intel_crtc->wm.skl_active = *pipe_wm;
+	return true;
+}
+
+static void skl_update_other_pipe_wm(struct drm_device *dev,
+				     struct drm_crtc *crtc,
+				     struct intel_wm_config *config,
+				     struct skl_wm_values *r)
+{
+	struct intel_crtc *intel_crtc;
+	struct intel_crtc *this_crtc = to_intel_crtc(crtc);
+
+	/*
+	 * If the WM update hasn't changed the allocation for this_crtc (the
+	 * crtc we are currently computing the new WM values for), other
+	 * enabled crtcs will keep the same allocation and we don't need to
+	 * recompute anything for them.
+	 */
+	if (!skl_ddb_allocation_changed(&r->ddb, this_crtc))
+		return;
+
+	/*
+	 * Otherwise, because of this_crtc being freshly enabled/disabled, the
+	 * other active pipes need new DDB allocation and WM values.
+	 */
+	list_for_each_entry(intel_crtc, &dev->mode_config.crtc_list,
+				base.head) {
+		struct skl_pipe_wm_parameters params = {};
+		struct skl_pipe_wm pipe_wm = {};
+		bool wm_changed;
+
+		if (this_crtc->pipe == intel_crtc->pipe)
+			continue;
+
+		if (!intel_crtc->active)
+			continue;
+
+		wm_changed = skl_update_pipe_wm(&intel_crtc->base,
+						&params, config,
+						&r->ddb, &pipe_wm);
+
+		/*
+		 * If we end up re-computing the other pipe WM values, it's
+		 * because it was really needed, so we expect the WM values to
+		 * be different.
+		 */
+		WARN_ON(!wm_changed);
+
+		skl_compute_wm_results(dev, &params, &pipe_wm, r, intel_crtc);
+		r->dirty[intel_crtc->pipe] = true;
+	}
+}
+
+static void skl_update_wm(struct drm_crtc *crtc)
+{
+	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
+	struct drm_device *dev = crtc->dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct skl_pipe_wm_parameters params = {};
+	struct skl_wm_values *results = &dev_priv->wm.skl_results;
+	struct skl_pipe_wm pipe_wm = {};
+	struct intel_wm_config config = {};
+
+	memset(results, 0, sizeof(*results));
+
+	skl_compute_wm_global_parameters(dev, &config);
+
+	if (!skl_update_pipe_wm(crtc, &params, &config,
+				&results->ddb, &pipe_wm))
+		return;
+
+	skl_compute_wm_results(dev, &params, &pipe_wm, results, intel_crtc);
+	results->dirty[intel_crtc->pipe] = true;
+
+	skl_update_other_pipe_wm(dev, crtc, &config, results);
+	skl_write_wm_values(dev_priv, results);
+	skl_flush_wm_values(dev_priv, results);
+
+	/* store the new configuration */
+	dev_priv->wm.skl_hw = *results;
+}
+
+static void
+skl_update_sprite_wm(struct drm_plane *plane, struct drm_crtc *crtc,
+		     uint32_t sprite_width, uint32_t sprite_height,
+		     int pixel_size, bool enabled, bool scaled)
+{
+	struct intel_plane *intel_plane = to_intel_plane(plane);
+
+	intel_plane->wm.enabled = enabled;
+	intel_plane->wm.scaled = scaled;
+	intel_plane->wm.horiz_pixels = sprite_width;
+	intel_plane->wm.vert_pixels = sprite_height;
+	intel_plane->wm.bytes_per_pixel = pixel_size;
+
+	skl_update_wm(crtc);
+}
+
 static void ilk_update_wm(struct drm_crtc *crtc)
 {
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
@@ -2770,6 +3836,113 @@ ilk_update_sprite_wm(struct drm_plane *plane,
 	ilk_update_wm(crtc);
 }
 
+static void skl_pipe_wm_active_state(uint32_t val,
+				     struct skl_pipe_wm *active,
+				     bool is_transwm,
+				     bool is_cursor,
+				     int i,
+				     int level)
+{
+	bool is_enabled = (val & PLANE_WM_EN) != 0;
+
+	if (!is_transwm) {
+		if (!is_cursor) {
+			active->wm[level].plane_en[i] = is_enabled;
+			active->wm[level].plane_res_b[i] =
+					val & PLANE_WM_BLOCKS_MASK;
+			active->wm[level].plane_res_l[i] =
+					(val >> PLANE_WM_LINES_SHIFT) &
+						PLANE_WM_LINES_MASK;
+		} else {
+			active->wm[level].cursor_en = is_enabled;
+			active->wm[level].cursor_res_b =
+					val & PLANE_WM_BLOCKS_MASK;
+			active->wm[level].cursor_res_l =
+					(val >> PLANE_WM_LINES_SHIFT) &
+						PLANE_WM_LINES_MASK;
+		}
+	} else {
+		if (!is_cursor) {
+			active->trans_wm.plane_en[i] = is_enabled;
+			active->trans_wm.plane_res_b[i] =
+					val & PLANE_WM_BLOCKS_MASK;
+			active->trans_wm.plane_res_l[i] =
+					(val >> PLANE_WM_LINES_SHIFT) &
+						PLANE_WM_LINES_MASK;
+		} else {
+			active->trans_wm.cursor_en = is_enabled;
+			active->trans_wm.cursor_res_b =
+					val & PLANE_WM_BLOCKS_MASK;
+			active->trans_wm.cursor_res_l =
+					(val >> PLANE_WM_LINES_SHIFT) &
+						PLANE_WM_LINES_MASK;
+		}
+	}
+}
+
+static void skl_pipe_wm_get_hw_state(struct drm_crtc *crtc)
+{
+	struct drm_device *dev = crtc->dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct skl_wm_values *hw = &dev_priv->wm.skl_hw;
+	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
+	struct skl_pipe_wm *active = &intel_crtc->wm.skl_active;
+	enum pipe pipe = intel_crtc->pipe;
+	int level, i, max_level;
+	uint32_t temp;
+
+	max_level = ilk_wm_max_level(dev);
+
+	hw->wm_linetime[pipe] = I915_READ(PIPE_WM_LINETIME(pipe));
+
+	for (level = 0; level <= max_level; level++) {
+		for (i = 0; i < intel_num_planes(intel_crtc); i++)
+			hw->plane[pipe][i][level] =
+					I915_READ(PLANE_WM(pipe, i, level));
+		hw->cursor[pipe][level] = I915_READ(CUR_WM(pipe, level));
+	}
+
+	for (i = 0; i < intel_num_planes(intel_crtc); i++)
+		hw->plane_trans[pipe][i] = I915_READ(PLANE_WM_TRANS(pipe, i));
+	hw->cursor_trans[pipe] = I915_READ(CUR_WM_TRANS(pipe));
+
+	if (!intel_crtc_active(crtc))
+		return;
+
+	hw->dirty[pipe] = true;
+
+	active->linetime = hw->wm_linetime[pipe];
+
+	for (level = 0; level <= max_level; level++) {
+		for (i = 0; i < intel_num_planes(intel_crtc); i++) {
+			temp = hw->plane[pipe][i][level];
+			skl_pipe_wm_active_state(temp, active, false,
+						false, i, level);
+		}
+		temp = hw->cursor[pipe][level];
+		skl_pipe_wm_active_state(temp, active, false, true, i, level);
+	}
+
+	for (i = 0; i < intel_num_planes(intel_crtc); i++) {
+		temp = hw->plane_trans[pipe][i];
+		skl_pipe_wm_active_state(temp, active, true, false, i, 0);
+	}
+
+	temp = hw->cursor_trans[pipe];
+	skl_pipe_wm_active_state(temp, active, true, true, i, 0);
+}
+
+void skl_wm_get_hw_state(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct skl_ddb_allocation *ddb = &dev_priv->wm.skl_hw.ddb;
+	struct drm_crtc *crtc;
+
+	skl_ddb_get_hw_state(dev_priv, ddb);
+	list_for_each_entry(crtc, &dev->mode_config.crtc_list, head)
+		skl_pipe_wm_get_hw_state(crtc);
+}
+
 static void ilk_pipe_wm_get_hw_state(struct drm_crtc *crtc)
 {
 	struct drm_device *dev = crtc->dev;
@@ -3278,7 +4451,7 @@ static void vlv_set_rps_idle(struct drm_i915_private *dev_priv)
 					dev_priv->rps.min_freq_softlimit);
 
 	if (wait_for(((vlv_punit_read(dev_priv, PUNIT_REG_GPU_FREQ_STS))
-				& GENFREQSTATUS) == 0, 5))
+				& GENFREQSTATUS) == 0, 100))
 		DRM_ERROR("timed out waiting for Punit\n");
 
 	vlv_force_gfx_clock(dev_priv, false);
@@ -3327,10 +4500,9 @@ void valleyview_set_rps(struct drm_device *dev, u8 val)
 	WARN_ON(val > dev_priv->rps.max_freq_softlimit);
 	WARN_ON(val < dev_priv->rps.min_freq_softlimit);
 
-	DRM_DEBUG_DRIVER("GPU freq request from %d MHz (%u) to %d MHz (%u)\n",
-			 vlv_gpu_freq(dev_priv, dev_priv->rps.cur_freq),
-			 dev_priv->rps.cur_freq,
-			 vlv_gpu_freq(dev_priv, val), val);
+	if (WARN_ONCE(IS_CHERRYVIEW(dev) && (val & 1),
+		      "Odd GPU freq value\n"))
+		val &= ~1;
 
 	if (val != dev_priv->rps.cur_freq)
 		vlv_punit_write(dev_priv, PUNIT_REG_GPU_FREQ_REQ, val);
@@ -3341,43 +4513,11 @@ void valleyview_set_rps(struct drm_device *dev, u8 val)
 	trace_intel_gpu_freq_change(vlv_gpu_freq(dev_priv, val));
 }
 
-static void gen8_disable_rps_interrupts(struct drm_device *dev)
+static void gen9_disable_rps(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 
-	I915_WRITE(GEN6_PMINTRMSK, ~GEN8_PMINTR_REDIRECT_TO_NON_DISP);
-	I915_WRITE(GEN8_GT_IER(2), I915_READ(GEN8_GT_IER(2)) &
-				   ~dev_priv->pm_rps_events);
-	/* Complete PM interrupt masking here doesn't race with the rps work
-	 * item again unmasking PM interrupts because that is using a different
-	 * register (GEN8_GT_IMR(2)) to mask PM interrupts. The only risk is in
-	 * leaving stale bits in GEN8_GT_IIR(2) and GEN8_GT_IMR(2) which
-	 * gen8_enable_rps will clean up. */
-
-	spin_lock_irq(&dev_priv->irq_lock);
-	dev_priv->rps.pm_iir = 0;
-	spin_unlock_irq(&dev_priv->irq_lock);
-
-	I915_WRITE(GEN8_GT_IIR(2), dev_priv->pm_rps_events);
-}
-
-static void gen6_disable_rps_interrupts(struct drm_device *dev)
-{
-	struct drm_i915_private *dev_priv = dev->dev_private;
-
-	I915_WRITE(GEN6_PMINTRMSK, 0xffffffff);
-	I915_WRITE(GEN6_PMIER, I915_READ(GEN6_PMIER) &
-				~dev_priv->pm_rps_events);
-	/* Complete PM interrupt masking here doesn't race with the rps work
-	 * item again unmasking PM interrupts because that is using a different
-	 * register (PMIMR) to mask PM interrupts. The only risk is in leaving
-	 * stale bits in PMIIR and PMIMR which gen6_enable_rps will clean up. */
-
-	spin_lock_irq(&dev_priv->irq_lock);
-	dev_priv->rps.pm_iir = 0;
-	spin_unlock_irq(&dev_priv->irq_lock);
-
-	I915_WRITE(GEN6_PMIIR, dev_priv->pm_rps_events);
+	I915_WRITE(GEN6_RC_CONTROL, 0);
 }
 
 static void gen6_disable_rps(struct drm_device *dev)
@@ -3386,11 +4526,6 @@ static void gen6_disable_rps(struct drm_device *dev)
 
 	I915_WRITE(GEN6_RC_CONTROL, 0);
 	I915_WRITE(GEN6_RPNSWREQ, 1 << 31);
-
-	if (IS_BROADWELL(dev))
-		gen8_disable_rps_interrupts(dev);
-	else
-		gen6_disable_rps_interrupts(dev);
 }
 
 static void cherryview_disable_rps(struct drm_device *dev)
@@ -3398,17 +4533,19 @@ static void cherryview_disable_rps(struct drm_device *dev)
 	struct drm_i915_private *dev_priv = dev->dev_private;
 
 	I915_WRITE(GEN6_RC_CONTROL, 0);
-
-	gen8_disable_rps_interrupts(dev);
 }
 
 static void valleyview_disable_rps(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 
+	/* we're doing forcewake before Disabling RC6,
+	 * This what the BIOS expects when going into suspend */
+	gen6_gt_force_wake_get(dev_priv, FORCEWAKE_ALL);
+
 	I915_WRITE(GEN6_RC_CONTROL, 0);
 
-	gen6_disable_rps_interrupts(dev);
+	gen6_gt_force_wake_put(dev_priv, FORCEWAKE_ALL);
 }
 
 static void intel_print_rc6_info(struct drm_device *dev, u32 mode)
@@ -3419,10 +4556,15 @@ static void intel_print_rc6_info(struct drm_device *dev, u32 mode)
 		else
 			mode = 0;
 	}
-	DRM_DEBUG_KMS("Enabling RC6 states: RC6 %s, RC6p %s, RC6pp %s\n",
-		      (mode & GEN6_RC_CTL_RC6_ENABLE) ? "on" : "off",
-		      (mode & GEN6_RC_CTL_RC6p_ENABLE) ? "on" : "off",
-		      (mode & GEN6_RC_CTL_RC6pp_ENABLE) ? "on" : "off");
+	if (HAS_RC6p(dev))
+		DRM_DEBUG_KMS("Enabling RC6 states: RC6 %s RC6p %s RC6pp %s\n",
+			      (mode & GEN6_RC_CTL_RC6_ENABLE) ? "on" : "off",
+			      (mode & GEN6_RC_CTL_RC6p_ENABLE) ? "on" : "off",
+			      (mode & GEN6_RC_CTL_RC6pp_ENABLE) ? "on" : "off");
+
+	else
+		DRM_DEBUG_KMS("Enabling RC6 states: RC6 %s\n",
+			      (mode & GEN6_RC_CTL_RC6_ENABLE) ? "on" : "off");
 }
 
 static int sanitize_rc6_option(const struct drm_device *dev, int enable_rc6)
@@ -3439,7 +4581,7 @@ static int sanitize_rc6_option(const struct drm_device *dev, int enable_rc6)
 	if (enable_rc6 >= 0) {
 		int mask;
 
-		if (INTEL_INFO(dev)->gen == 6 || IS_IVYBRIDGE(dev))
+		if (HAS_RC6p(dev))
 			mask = INTEL_RC6_ENABLE | INTEL_RC6p_ENABLE |
 			       INTEL_RC6pp_ENABLE;
 		else
@@ -3467,54 +4609,92 @@ int intel_enable_rc6(const struct drm_device *dev)
 	return i915.enable_rc6;
 }
 
-static void gen8_enable_rps_interrupts(struct drm_device *dev)
+static void gen6_init_rps_frequencies(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
+	uint32_t rp_state_cap;
+	u32 ddcc_status = 0;
+	int ret;
 
-	spin_lock_irq(&dev_priv->irq_lock);
-	WARN_ON(dev_priv->rps.pm_iir);
-	gen8_enable_pm_irq(dev_priv, dev_priv->pm_rps_events);
-	I915_WRITE(GEN8_GT_IIR(2), dev_priv->pm_rps_events);
-	spin_unlock_irq(&dev_priv->irq_lock);
-}
-
-static void gen6_enable_rps_interrupts(struct drm_device *dev)
-{
-	struct drm_i915_private *dev_priv = dev->dev_private;
-
-	spin_lock_irq(&dev_priv->irq_lock);
-	WARN_ON(dev_priv->rps.pm_iir);
-	gen6_enable_pm_irq(dev_priv, dev_priv->pm_rps_events);
-	I915_WRITE(GEN6_PMIIR, dev_priv->pm_rps_events);
-	spin_unlock_irq(&dev_priv->irq_lock);
-}
-
-static void parse_rp_state_cap(struct drm_i915_private *dev_priv, u32 rp_state_cap)
-{
+	rp_state_cap = I915_READ(GEN6_RP_STATE_CAP);
 	/* All of these values are in units of 50MHz */
 	dev_priv->rps.cur_freq		= 0;
-	/* static values from HW: RP0 < RPe < RP1 < RPn (min_freq) */
-	dev_priv->rps.rp1_freq		= (rp_state_cap >>  8) & 0xff;
+	/* static values from HW: RP0 > RP1 > RPn (min_freq) */
 	dev_priv->rps.rp0_freq		= (rp_state_cap >>  0) & 0xff;
+	dev_priv->rps.rp1_freq		= (rp_state_cap >>  8) & 0xff;
 	dev_priv->rps.min_freq		= (rp_state_cap >> 16) & 0xff;
-	/* XXX: only BYT has a special efficient freq */
-	dev_priv->rps.efficient_freq	= dev_priv->rps.rp1_freq;
 	/* hw_max = RP0 until we check for overclocking */
 	dev_priv->rps.max_freq		= dev_priv->rps.rp0_freq;
 
+	dev_priv->rps.efficient_freq = dev_priv->rps.rp1_freq;
+	if (IS_HASWELL(dev) || IS_BROADWELL(dev)) {
+		ret = sandybridge_pcode_read(dev_priv,
+					HSW_PCODE_DYNAMIC_DUTY_CYCLE_CONTROL,
+					&ddcc_status);
+		if (0 == ret)
+			dev_priv->rps.efficient_freq =
+				(ddcc_status >> 8) & 0xff;
+	}
+
 	/* Preserve min/max settings in case of re-init */
 	if (dev_priv->rps.max_freq_softlimit == 0)
 		dev_priv->rps.max_freq_softlimit = dev_priv->rps.max_freq;
 
-	if (dev_priv->rps.min_freq_softlimit == 0)
-		dev_priv->rps.min_freq_softlimit = dev_priv->rps.min_freq;
+	if (dev_priv->rps.min_freq_softlimit == 0) {
+		if (IS_HASWELL(dev) || IS_BROADWELL(dev))
+			dev_priv->rps.min_freq_softlimit =
+				/* max(RPe, 450 MHz) */
+				max(dev_priv->rps.efficient_freq, (u8) 9);
+		else
+			dev_priv->rps.min_freq_softlimit =
+				dev_priv->rps.min_freq;
+	}
+}
+
+static void gen9_enable_rps(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct intel_engine_cs *ring;
+	uint32_t rc6_mask = 0;
+	int unused;
+
+	/* 1a: Software RC state - RC0 */
+	I915_WRITE(GEN6_RC_STATE, 0);
+
+	/* 1b: Get forcewake during program sequence. Although the driver
+	 * hasn't enabled a state yet where we need forcewake, BIOS may have.*/
+	gen6_gt_force_wake_get(dev_priv, FORCEWAKE_ALL);
+
+	/* 2a: Disable RC states. */
+	I915_WRITE(GEN6_RC_CONTROL, 0);
+
+	/* 2b: Program RC6 thresholds.*/
+	I915_WRITE(GEN6_RC6_WAKE_RATE_LIMIT, 54 << 16);
+	I915_WRITE(GEN6_RC_EVALUATION_INTERVAL, 125000); /* 12500 * 1280ns */
+	I915_WRITE(GEN6_RC_IDLE_HYSTERSIS, 25); /* 25 * 1280ns */
+	for_each_ring(ring, dev_priv, unused)
+		I915_WRITE(RING_MAX_IDLE(ring->mmio_base), 10);
+	I915_WRITE(GEN6_RC_SLEEP, 0);
+	I915_WRITE(GEN6_RC6_THRESHOLD, 37500); /* 37.5/125ms per EI */
+
+	/* 3a: Enable RC6 */
+	if (intel_enable_rc6(dev) & INTEL_RC6_ENABLE)
+		rc6_mask = GEN6_RC_CTL_RC6_ENABLE;
+	DRM_INFO("RC6 %s\n", (rc6_mask & GEN6_RC_CTL_RC6_ENABLE) ?
+			"on" : "off");
+	I915_WRITE(GEN6_RC_CONTROL, GEN6_RC_CTL_HW_ENABLE |
+				   GEN6_RC_CTL_EI_MODE(1) |
+				   rc6_mask);
+
+	gen6_gt_force_wake_put(dev_priv, FORCEWAKE_ALL);
+
 }
 
 static void gen8_enable_rps(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_engine_cs *ring;
-	uint32_t rc6_mask = 0, rp_state_cap;
+	uint32_t rc6_mask = 0;
 	int unused;
 
 	/* 1a: Software RC state - RC0 */
@@ -3527,8 +4707,8 @@ static void gen8_enable_rps(struct drm_device *dev)
 	/* 2a: Disable RC states. */
 	I915_WRITE(GEN6_RC_CONTROL, 0);
 
-	rp_state_cap = I915_READ(GEN6_RP_STATE_CAP);
-	parse_rp_state_cap(dev_priv, rp_state_cap);
+	/* Initialize rps frequencies */
+	gen6_init_rps_frequencies(dev);
 
 	/* 2b: Program RC6 thresholds.*/
 	I915_WRITE(GEN6_RC6_WAKE_RATE_LIMIT, 40 << 16);
@@ -3586,9 +4766,8 @@ static void gen8_enable_rps(struct drm_device *dev)
 
 	/* 6: Ring frequency + overclocking (our driver does this later */
 
-	gen6_set_rps(dev, (I915_READ(GEN6_GT_PERF_STATUS) & 0xff00) >> 8);
-
-	gen8_enable_rps_interrupts(dev);
+	dev_priv->rps.power = HIGH_POWER; /* force a reset */
+	gen6_set_rps(dev_priv->dev, dev_priv->rps.min_freq_softlimit);
 
 	gen6_gt_force_wake_put(dev_priv, FORCEWAKE_ALL);
 }
@@ -3597,8 +4776,6 @@ static void gen6_enable_rps(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_engine_cs *ring;
-	u32 rp_state_cap;
-	u32 gt_perf_status;
 	u32 rc6vids, pcu_mbox = 0, rc6_mask = 0;
 	u32 gtfifodbg;
 	int rc6_mode;
@@ -3622,10 +4799,8 @@ static void gen6_enable_rps(struct drm_device *dev)
 
 	gen6_gt_force_wake_get(dev_priv, FORCEWAKE_ALL);
 
-	rp_state_cap = I915_READ(GEN6_RP_STATE_CAP);
-	gt_perf_status = I915_READ(GEN6_GT_PERF_STATUS);
-
-	parse_rp_state_cap(dev_priv, rp_state_cap);
+	/* Initialize rps frequencies */
+	gen6_init_rps_frequencies(dev);
 
 	/* disable the counters and set deterministic thresholds */
 	I915_WRITE(GEN6_RC_CONTROL, 0);
@@ -3688,8 +4863,6 @@ static void gen6_enable_rps(struct drm_device *dev)
 	dev_priv->rps.power = HIGH_POWER; /* force a reset */
 	gen6_set_rps(dev_priv->dev, dev_priv->rps.min_freq_softlimit);
 
-	gen6_enable_rps_interrupts(dev);
-
 	rc6vids = 0;
 	ret = sandybridge_pcode_read(dev_priv, GEN6_PCODE_READ_RC6VIDS, &rc6vids);
 	if (IS_GEN6(dev) && ret) {
@@ -3742,9 +4915,9 @@ static void __gen6_update_ring_freq(struct drm_device *dev)
 	 * to use for memory access.  We do this by specifying the IA frequency
 	 * the PCU should use as a reference to determine the ring frequency.
 	 */
-	for (gpu_freq = dev_priv->rps.max_freq_softlimit; gpu_freq >= dev_priv->rps.min_freq_softlimit;
+	for (gpu_freq = dev_priv->rps.max_freq; gpu_freq >= dev_priv->rps.min_freq;
 	     gpu_freq--) {
-		int diff = dev_priv->rps.max_freq_softlimit - gpu_freq;
+		int diff = dev_priv->rps.max_freq - gpu_freq;
 		unsigned int ia_freq = 0, ring_freq = 0;
 
 		if (INTEL_INFO(dev)->gen >= 8) {
@@ -3899,12 +5072,15 @@ static void cherryview_setup_pctx(struct drm_device *dev)
 
 	pcbr = I915_READ(VLV_PCBR);
 	if ((pcbr >> VLV_PCBR_ADDR_SHIFT) == 0) {
+		DRM_DEBUG_DRIVER("BIOS didn't set up PCBR, fixing up\n");
 		paddr = (dev_priv->mm.stolen_base +
 			 (gtt->stolen_size - pctx_size));
 
 		pctx_paddr = (paddr & (~4095));
 		I915_WRITE(VLV_PCBR, pctx_paddr);
 	}
+
+	DRM_DEBUG_DRIVER("PCBR: 0x%08x\n", I915_READ(VLV_PCBR));
 }
 
 static void valleyview_setup_pctx(struct drm_device *dev)
@@ -3930,6 +5106,8 @@ static void valleyview_setup_pctx(struct drm_device *dev)
 		goto out;
 	}
 
+	DRM_DEBUG_DRIVER("BIOS didn't set up PCBR, fixing up\n");
+
 	/*
 	 * From the Gunit register HAS:
 	 * The Gfx driver is expected to program this register and ensure
@@ -3948,6 +5126,7 @@ static void valleyview_setup_pctx(struct drm_device *dev)
 	I915_WRITE(VLV_PCBR, pctx_paddr);
 
 out:
+	DRM_DEBUG_DRIVER("PCBR: 0x%08x\n", I915_READ(VLV_PCBR));
 	dev_priv->vlv_pctx = pctx;
 }
 
@@ -3965,11 +5144,27 @@ static void valleyview_cleanup_pctx(struct drm_device *dev)
 static void valleyview_init_gt_powersave(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
+	u32 val;
 
 	valleyview_setup_pctx(dev);
 
 	mutex_lock(&dev_priv->rps.hw_lock);
 
+	val = vlv_punit_read(dev_priv, PUNIT_REG_GPU_FREQ_STS);
+	switch ((val >> 6) & 3) {
+	case 0:
+	case 1:
+		dev_priv->mem_freq = 800;
+		break;
+	case 2:
+		dev_priv->mem_freq = 1066;
+		break;
+	case 3:
+		dev_priv->mem_freq = 1333;
+		break;
+	}
+	DRM_DEBUG_DRIVER("DDR speed: %d MHz\n", dev_priv->mem_freq);
+
 	dev_priv->rps.max_freq = valleyview_rps_max_freq(dev_priv);
 	dev_priv->rps.rp0_freq = dev_priv->rps.max_freq;
 	DRM_DEBUG_DRIVER("max GPU freq: %d MHz (%u)\n",
@@ -4004,11 +5199,41 @@ static void valleyview_init_gt_powersave(struct drm_device *dev)
 static void cherryview_init_gt_powersave(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
+	u32 val;
 
 	cherryview_setup_pctx(dev);
 
 	mutex_lock(&dev_priv->rps.hw_lock);
 
+	mutex_lock(&dev_priv->dpio_lock);
+	val = vlv_cck_read(dev_priv, CCK_FUSE_REG);
+	mutex_unlock(&dev_priv->dpio_lock);
+
+	switch ((val >> 2) & 0x7) {
+	case 0:
+	case 1:
+		dev_priv->rps.cz_freq = 200;
+		dev_priv->mem_freq = 1600;
+		break;
+	case 2:
+		dev_priv->rps.cz_freq = 267;
+		dev_priv->mem_freq = 1600;
+		break;
+	case 3:
+		dev_priv->rps.cz_freq = 333;
+		dev_priv->mem_freq = 2000;
+		break;
+	case 4:
+		dev_priv->rps.cz_freq = 320;
+		dev_priv->mem_freq = 1600;
+		break;
+	case 5:
+		dev_priv->rps.cz_freq = 400;
+		dev_priv->mem_freq = 1600;
+		break;
+	}
+	DRM_DEBUG_DRIVER("DDR speed: %d MHz\n", dev_priv->mem_freq);
+
 	dev_priv->rps.max_freq = cherryview_rps_max_freq(dev_priv);
 	dev_priv->rps.rp0_freq = dev_priv->rps.max_freq;
 	DRM_DEBUG_DRIVER("max GPU freq: %d MHz (%u)\n",
@@ -4030,6 +5255,12 @@ static void cherryview_init_gt_powersave(struct drm_device *dev)
 			 vlv_gpu_freq(dev_priv, dev_priv->rps.min_freq),
 			 dev_priv->rps.min_freq);
 
+	WARN_ONCE((dev_priv->rps.max_freq |
+		   dev_priv->rps.efficient_freq |
+		   dev_priv->rps.rp1_freq |
+		   dev_priv->rps.min_freq) & 1,
+		  "Odd GPU freq values\n");
+
 	/* Preserve min/max settings in case of re-init */
 	if (dev_priv->rps.max_freq_softlimit == 0)
 		dev_priv->rps.max_freq_softlimit = dev_priv->rps.max_freq;
@@ -4087,8 +5318,6 @@ static void cherryview_enable_rps(struct drm_device *dev)
 	/* For now we assume BIOS is allocating and populating the PCBR  */
 	pcbr = I915_READ(VLV_PCBR);
 
-	DRM_DEBUG_DRIVER("PCBR offset : 0x%x\n", pcbr);
-
 	/* 3: Enable RC6 */
 	if ((intel_enable_rc6(dev) & INTEL_RC6_ENABLE) &&
 						(pcbr >> VLV_PCBR_ADDR_SHIFT))
@@ -4118,7 +5347,10 @@ static void cherryview_enable_rps(struct drm_device *dev)
 
 	val = vlv_punit_read(dev_priv, PUNIT_REG_GPU_FREQ_STS);
 
-	DRM_DEBUG_DRIVER("GPLL enabled? %s\n", val & 0x10 ? "yes" : "no");
+	/* RPS code assumes GPLL is used */
+	WARN_ONCE((val & GPLLENABLE) == 0, "GPLL not enabled\n");
+
+	DRM_DEBUG_DRIVER("GPLL enabled? %s\n", val & GPLLENABLE ? "yes" : "no");
 	DRM_DEBUG_DRIVER("GPU status: 0x%08x\n", val);
 
 	dev_priv->rps.cur_freq = (val >> 8) & 0xff;
@@ -4132,8 +5364,6 @@ static void cherryview_enable_rps(struct drm_device *dev)
 
 	valleyview_set_rps(dev_priv->dev, dev_priv->rps.efficient_freq);
 
-	gen8_enable_rps_interrupts(dev);
-
 	gen6_gt_force_wake_put(dev_priv, FORCEWAKE_ALL);
 }
 
@@ -4198,7 +5428,10 @@ static void valleyview_enable_rps(struct drm_device *dev)
 
 	val = vlv_punit_read(dev_priv, PUNIT_REG_GPU_FREQ_STS);
 
-	DRM_DEBUG_DRIVER("GPLL enabled? %s\n", val & 0x10 ? "yes" : "no");
+	/* RPS code assumes GPLL is used */
+	WARN_ONCE((val & GPLLENABLE) == 0, "GPLL not enabled\n");
+
+	DRM_DEBUG_DRIVER("GPLL enabled? %s\n", val & GPLLENABLE ? "yes" : "no");
 	DRM_DEBUG_DRIVER("GPU status: 0x%08x\n", val);
 
 	dev_priv->rps.cur_freq = (val >> 8) & 0xff;
@@ -4212,8 +5445,6 @@ static void valleyview_enable_rps(struct drm_device *dev)
 
 	valleyview_set_rps(dev_priv->dev, dev_priv->rps.efficient_freq);
 
-	gen6_enable_rps_interrupts(dev);
-
 	gen6_gt_force_wake_put(dev_priv, FORCEWAKE_ALL);
 }
 
@@ -4960,6 +6191,20 @@ void intel_cleanup_gt_powersave(struct drm_device *dev)
 		valleyview_cleanup_gt_powersave(dev);
 }
 
+static void gen6_suspend_rps(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+
+	flush_delayed_work(&dev_priv->rps.delayed_resume_work);
+
+	/*
+	 * TODO: disable RPS interrupts on GEN9+ too once RPS support
+	 * is added for it.
+	 */
+	if (INTEL_INFO(dev)->gen < 9)
+		gen6_disable_rps_interrupts(dev);
+}
+
 /**
  * intel_suspend_gt_powersave - suspend PM work and helper threads
  * @dev: drm device
@@ -4972,12 +6217,10 @@ void intel_suspend_gt_powersave(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 
-	/* Interrupts should be disabled already to avoid re-arming. */
-	WARN_ON(intel_irqs_enabled(dev_priv));
-
-	flush_delayed_work(&dev_priv->rps.delayed_resume_work);
+	if (INTEL_INFO(dev)->gen < 6)
+		return;
 
-	cancel_work_sync(&dev_priv->rps.work);
+	gen6_suspend_rps(dev);
 
 	/* Force GPU to min freq during suspend */
 	gen6_rps_idle(dev_priv);
@@ -4987,9 +6230,6 @@ void intel_disable_gt_powersave(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 
-	/* Interrupts should be disabled already to avoid re-arming. */
-	WARN_ON(intel_irqs_enabled(dev_priv));
-
 	if (IS_IRONLAKE_M(dev)) {
 		ironlake_disable_drps(dev);
 		ironlake_disable_rc6(dev);
@@ -4997,12 +6237,15 @@ void intel_disable_gt_powersave(struct drm_device *dev)
 		intel_suspend_gt_powersave(dev);
 
 		mutex_lock(&dev_priv->rps.hw_lock);
-		if (IS_CHERRYVIEW(dev))
+		if (INTEL_INFO(dev)->gen >= 9)
+			gen9_disable_rps(dev);
+		else if (IS_CHERRYVIEW(dev))
 			cherryview_disable_rps(dev);
 		else if (IS_VALLEYVIEW(dev))
 			valleyview_disable_rps(dev);
 		else
 			gen6_disable_rps(dev);
+
 		dev_priv->rps.enabled = false;
 		mutex_unlock(&dev_priv->rps.hw_lock);
 	}
@@ -5017,10 +6260,19 @@ static void intel_gen6_powersave_work(struct work_struct *work)
 
 	mutex_lock(&dev_priv->rps.hw_lock);
 
+	/*
+	 * TODO: reset/enable RPS interrupts on GEN9+ too, once RPS support is
+	 * added for it.
+	 */
+	if (INTEL_INFO(dev)->gen < 9)
+		gen6_reset_rps_interrupts(dev);
+
 	if (IS_CHERRYVIEW(dev)) {
 		cherryview_enable_rps(dev);
 	} else if (IS_VALLEYVIEW(dev)) {
 		valleyview_enable_rps(dev);
+	} else if (INTEL_INFO(dev)->gen >= 9) {
+		gen9_enable_rps(dev);
 	} else if (IS_BROADWELL(dev)) {
 		gen8_enable_rps(dev);
 		__gen6_update_ring_freq(dev);
@@ -5029,6 +6281,10 @@ static void intel_gen6_powersave_work(struct work_struct *work)
 		__gen6_update_ring_freq(dev);
 	}
 	dev_priv->rps.enabled = true;
+
+	if (INTEL_INFO(dev)->gen < 9)
+		gen6_enable_rps_interrupts(dev);
+
 	mutex_unlock(&dev_priv->rps.hw_lock);
 
 	intel_runtime_pm_put(dev_priv);
@@ -5067,8 +6323,11 @@ void intel_reset_gt_powersave(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 
+	if (INTEL_INFO(dev)->gen < 6)
+		return;
+
+	gen6_suspend_rps(dev);
 	dev_priv->rps.enabled = false;
-	intel_enable_gt_powersave(dev);
 }
 
 static void ibx_init_clock_gating(struct drm_device *dev)
@@ -5088,7 +6347,7 @@ static void g4x_disable_trickle_feed(struct drm_device *dev)
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	int pipe;
 
-	for_each_pipe(pipe) {
+	for_each_pipe(dev_priv, pipe) {
 		I915_WRITE(DSPCNTR(pipe),
 			   I915_READ(DSPCNTR(pipe)) |
 			   DISPPLANE_TRICKLE_FEED_DISABLE);
@@ -5203,7 +6462,7 @@ static void cpt_init_clock_gating(struct drm_device *dev)
 	/* The below fixes the weird display corruption, a few pixels shifted
 	 * downward, on (only) LVDS of some HP laptops with IVY.
 	 */
-	for_each_pipe(pipe) {
+	for_each_pipe(dev_priv, pipe) {
 		val = I915_READ(TRANS_CHICKEN2(pipe));
 		val |= TRANS_CHICKEN2_TIMING_OVERRIDE;
 		val &= ~TRANS_CHICKEN2_FDI_POLARITY_REVERSED;
@@ -5215,7 +6474,7 @@ static void cpt_init_clock_gating(struct drm_device *dev)
 		I915_WRITE(TRANS_CHICKEN2(pipe), val);
 	}
 	/* WADP0ClockGatingDisable */
-	for_each_pipe(pipe) {
+	for_each_pipe(dev_priv, pipe) {
 		I915_WRITE(TRANS_CHICKEN1(pipe),
 			   TRANS_CHICKEN1_DP0UNIT_GC_DISABLE);
 	}
@@ -5247,11 +6506,6 @@ static void gen6_init_clock_gating(struct drm_device *dev)
 	I915_WRITE(_3D_CHICKEN,
 		   _MASKED_BIT_ENABLE(_3D_CHICKEN_HIZ_PLANE_DISABLE_MSAA_4X_SNB));
 
-	/* WaSetupGtModeTdRowDispatch:snb */
-	if (IS_SNB_GT1(dev))
-		I915_WRITE(GEN6_GT_MODE,
-			   _MASKED_BIT_ENABLE(GEN6_TD_FOUR_ROW_DISPATCH_DISABLE));
-
 	/* WaDisable_RenderCache_OperationalFlush:snb */
 	I915_WRITE(CACHE_MODE_0, _MASKED_BIT_DISABLE(RC_OP_FLUSH_ENABLE));
 
@@ -5264,7 +6518,7 @@ static void gen6_init_clock_gating(struct drm_device *dev)
 	 * to keep in mind (see 3DSTATE_PS and 3DSTATE_WM).
 	 */
 	I915_WRITE(GEN6_GT_MODE,
-		   GEN6_WIZ_HASHING_MASK | GEN6_WIZ_HASHING_16x4);
+		   _MASKED_FIELD(GEN6_WIZ_HASHING_MASK, GEN6_WIZ_HASHING_16x4));
 
 	ilk_init_lp_watermarks(dev);
 
@@ -5383,7 +6637,7 @@ static void lpt_suspend_hw(struct drm_device *dev)
 	}
 }
 
-static void gen8_init_clock_gating(struct drm_device *dev)
+static void broadwell_init_clock_gating(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	enum pipe pipe;
@@ -5392,41 +6646,6 @@ static void gen8_init_clock_gating(struct drm_device *dev)
 	I915_WRITE(WM2_LP_ILK, 0);
 	I915_WRITE(WM1_LP_ILK, 0);
 
-	/* FIXME(BDW): Check all the w/a, some might only apply to
-	 * pre-production hw. */
-
-	/* WaDisablePartialInstShootdown:bdw */
-	I915_WRITE(GEN8_ROW_CHICKEN,
-		   _MASKED_BIT_ENABLE(PARTIAL_INSTRUCTION_SHOOTDOWN_DISABLE));
-
-	/* WaDisableThreadStallDopClockGating:bdw */
-	/* FIXME: Unclear whether we really need this on production bdw. */
-	I915_WRITE(GEN8_ROW_CHICKEN,
-		   _MASKED_BIT_ENABLE(STALL_DOP_GATING_DISABLE));
-
-	/*
-	 * This GEN8_CENTROID_PIXEL_OPT_DIS W/A is only needed for
-	 * pre-production hardware
-	 */
-	I915_WRITE(HALF_SLICE_CHICKEN3,
-		   _MASKED_BIT_ENABLE(GEN8_CENTROID_PIXEL_OPT_DIS));
-	I915_WRITE(HALF_SLICE_CHICKEN3,
-		   _MASKED_BIT_ENABLE(GEN8_SAMPLER_POWER_BYPASS_DIS));
-	I915_WRITE(GAMTARBMODE, _MASKED_BIT_ENABLE(ARB_MODE_BWGTLB_DISABLE));
-
-	I915_WRITE(_3D_CHICKEN3,
-		   _MASKED_BIT_ENABLE(_3D_CHICKEN_SDE_LIMIT_FIFO_POLY_DEPTH(2)));
-
-	I915_WRITE(COMMON_SLICE_CHICKEN2,
-		   _MASKED_BIT_ENABLE(GEN8_CSC2_SBE_VUE_CACHE_CONSERVATIVE));
-
-	I915_WRITE(GEN7_HALF_SLICE_CHICKEN1,
-		   _MASKED_BIT_ENABLE(GEN7_SINGLE_SUBSCAN_DISPATCH_ENABLE));
-
-	/* WaDisableDopClockGating:bdw May not be needed for production */
-	I915_WRITE(GEN7_ROW_CHICKEN2,
-		   _MASKED_BIT_ENABLE(DOP_CLOCK_GATING_DISABLE));
-
 	/* WaSwitchSolVfFArbitrationPriority:bdw */
 	I915_WRITE(GAM_ECOCHK, I915_READ(GAM_ECOCHK) | HSW_ECOCHK_ARB_PRIO_SOL);
 
@@ -5435,37 +6654,18 @@ static void gen8_init_clock_gating(struct drm_device *dev)
 		   I915_READ(CHICKEN_PAR1_1) | DPA_MASK_VBLANK_SRD);
 
 	/* WaPsrDPRSUnmaskVBlankInSRD:bdw */
-	for_each_pipe(pipe) {
+	for_each_pipe(dev_priv, pipe) {
 		I915_WRITE(CHICKEN_PIPESL_1(pipe),
 			   I915_READ(CHICKEN_PIPESL_1(pipe)) |
 			   BDW_DPRS_MASK_VBLANK_SRD);
 	}
 
-	/* Use Force Non-Coherent whenever executing a 3D context. This is a
-	 * workaround for for a possible hang in the unlikely event a TLB
-	 * invalidation occurs during a PSD flush.
-	 */
-	I915_WRITE(HDC_CHICKEN0,
-		   I915_READ(HDC_CHICKEN0) |
-		   _MASKED_BIT_ENABLE(HDC_FORCE_NON_COHERENT));
-
 	/* WaVSRefCountFullforceMissDisable:bdw */
 	/* WaDSRefCountFullforceMissDisable:bdw */
 	I915_WRITE(GEN7_FF_THREAD_MODE,
 		   I915_READ(GEN7_FF_THREAD_MODE) &
 		   ~(GEN8_FF_DS_REF_CNT_FFME | GEN7_FF_VS_REF_CNT_FFME));
 
-	/*
-	 * BSpec recommends 8x4 when MSAA is used,
-	 * however in practice 16x4 seems fastest.
-	 *
-	 * Note that PS/WM thread counts depend on the WIZ hashing
-	 * disable bit, which we don't touch here, but it's good
-	 * to keep in mind (see 3DSTATE_PS and 3DSTATE_WM).
-	 */
-	I915_WRITE(GEN7_GT_MODE,
-		   GEN6_WIZ_HASHING_MASK | GEN6_WIZ_HASHING_16x4);
-
 	I915_WRITE(GEN6_RC_SLEEP_PSMI_CONTROL,
 		   _MASKED_BIT_ENABLE(GEN8_RC_SEMA_IDLE_MSG_DISABLE));
 
@@ -5473,9 +6673,7 @@ static void gen8_init_clock_gating(struct drm_device *dev)
 	I915_WRITE(GEN8_UCGCTL6, I915_READ(GEN8_UCGCTL6) |
 		   GEN8_SDEUNIT_CLOCK_GATE_DISABLE);
 
-	/* Wa4x4STCOptimizationDisable:bdw */
-	I915_WRITE(CACHE_MODE_1,
-		   _MASKED_BIT_ENABLE(GEN8_4x4_STC_OPTIMIZATION_DISABLE));
+	lpt_init_clock_gating(dev);
 }
 
 static void haswell_init_clock_gating(struct drm_device *dev)
@@ -5518,7 +6716,7 @@ static void haswell_init_clock_gating(struct drm_device *dev)
 	 * to keep in mind (see 3DSTATE_PS and 3DSTATE_WM).
 	 */
 	I915_WRITE(GEN7_GT_MODE,
-		   GEN6_WIZ_HASHING_MASK | GEN6_WIZ_HASHING_16x4);
+		   _MASKED_FIELD(GEN6_WIZ_HASHING_MASK, GEN6_WIZ_HASHING_16x4));
 
 	/* WaSwitchSolVfFArbitrationPriority:hsw */
 	I915_WRITE(GAM_ECOCHK, I915_READ(GAM_ECOCHK) | HSW_ECOCHK_ARB_PRIO_SOL);
@@ -5615,7 +6813,7 @@ static void ivybridge_init_clock_gating(struct drm_device *dev)
 	 * to keep in mind (see 3DSTATE_PS and 3DSTATE_WM).
 	 */
 	I915_WRITE(GEN7_GT_MODE,
-		   GEN6_WIZ_HASHING_MASK | GEN6_WIZ_HASHING_16x4);
+		   _MASKED_FIELD(GEN6_WIZ_HASHING_MASK, GEN6_WIZ_HASHING_16x4));
 
 	snpcr = I915_READ(GEN6_MBCUNIT_SNPCR);
 	snpcr &= ~GEN6_MBC_SNPCR_MASK;
@@ -5631,24 +6829,6 @@ static void ivybridge_init_clock_gating(struct drm_device *dev)
 static void valleyview_init_clock_gating(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	u32 val;
-
-	mutex_lock(&dev_priv->rps.hw_lock);
-	val = vlv_punit_read(dev_priv, PUNIT_REG_GPU_FREQ_STS);
-	mutex_unlock(&dev_priv->rps.hw_lock);
-	switch ((val >> 6) & 3) {
-	case 0:
-	case 1:
-		dev_priv->mem_freq = 800;
-		break;
-	case 2:
-		dev_priv->mem_freq = 1066;
-		break;
-	case 3:
-		dev_priv->mem_freq = 1333;
-		break;
-	}
-	DRM_DEBUG_DRIVER("DDR speed: %d MHz", dev_priv->mem_freq);
 
 	I915_WRITE(DSPCLK_GATE_D, VRHUNIT_CLOCK_GATE_DISABLE);
 
@@ -5724,48 +6904,11 @@ static void valleyview_init_clock_gating(struct drm_device *dev)
 static void cherryview_init_clock_gating(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	u32 val;
-
-	mutex_lock(&dev_priv->rps.hw_lock);
-	val = vlv_punit_read(dev_priv, CCK_FUSE_REG);
-	mutex_unlock(&dev_priv->rps.hw_lock);
-	switch ((val >> 2) & 0x7) {
-	case 0:
-	case 1:
-			dev_priv->rps.cz_freq = CHV_CZ_CLOCK_FREQ_MODE_200;
-			dev_priv->mem_freq = 1600;
-			break;
-	case 2:
-			dev_priv->rps.cz_freq = CHV_CZ_CLOCK_FREQ_MODE_267;
-			dev_priv->mem_freq = 1600;
-			break;
-	case 3:
-			dev_priv->rps.cz_freq = CHV_CZ_CLOCK_FREQ_MODE_333;
-			dev_priv->mem_freq = 2000;
-			break;
-	case 4:
-			dev_priv->rps.cz_freq = CHV_CZ_CLOCK_FREQ_MODE_320;
-			dev_priv->mem_freq = 1600;
-			break;
-	case 5:
-			dev_priv->rps.cz_freq = CHV_CZ_CLOCK_FREQ_MODE_400;
-			dev_priv->mem_freq = 1600;
-			break;
-	}
-	DRM_DEBUG_DRIVER("DDR speed: %d MHz", dev_priv->mem_freq);
 
 	I915_WRITE(DSPCLK_GATE_D, VRHUNIT_CLOCK_GATE_DISABLE);
 
 	I915_WRITE(MI_ARB_VLV, MI_ARB_DISPLAY_TRICKLE_FEED_DISABLE);
 
-	/* WaDisablePartialInstShootdown:chv */
-	I915_WRITE(GEN8_ROW_CHICKEN,
-		   _MASKED_BIT_ENABLE(PARTIAL_INSTRUCTION_SHOOTDOWN_DISABLE));
-
-	/* WaDisableThreadStallDopClockGating:chv */
-	I915_WRITE(GEN8_ROW_CHICKEN,
-		   _MASKED_BIT_ENABLE(STALL_DOP_GATING_DISABLE));
-
 	/* WaVSRefCountFullforceMissDisable:chv */
 	/* WaDSRefCountFullforceMissDisable:chv */
 	I915_WRITE(GEN7_FF_THREAD_MODE,
@@ -5783,24 +6926,6 @@ static void cherryview_init_clock_gating(struct drm_device *dev)
 	/* WaDisableSDEUnitClockGating:chv */
 	I915_WRITE(GEN8_UCGCTL6, I915_READ(GEN8_UCGCTL6) |
 		   GEN8_SDEUNIT_CLOCK_GATE_DISABLE);
-
-	/* WaDisableSamplerPowerBypass:chv (pre-production hw) */
-	I915_WRITE(HALF_SLICE_CHICKEN3,
-		   _MASKED_BIT_ENABLE(GEN8_SAMPLER_POWER_BYPASS_DIS));
-
-	/* WaDisableGunitClockGating:chv (pre-production hw) */
-	I915_WRITE(VLV_GUNIT_CLOCK_GATE, I915_READ(VLV_GUNIT_CLOCK_GATE) |
-		   GINT_DIS);
-
-	/* WaDisableFfDopClockGating:chv (pre-production hw) */
-	I915_WRITE(GEN6_RC_SLEEP_PSMI_CONTROL,
-		   _MASKED_BIT_ENABLE(GEN8_FF_DOP_CLOCK_GATE_DISABLE));
-
-	/* WaDisableDopClockGating:chv (pre-production hw) */
-	I915_WRITE(GEN7_ROW_CHICKEN2,
-		   _MASKED_BIT_ENABLE(DOP_CLOCK_GATING_DISABLE));
-	I915_WRITE(GEN6_UCGCTL1, I915_READ(GEN6_UCGCTL1) |
-		   GEN6_EU_TCUNIT_CLOCK_GATE_DISABLE);
 }
 
 static void g4x_init_clock_gating(struct drm_device *dev)
@@ -5883,6 +7008,9 @@ static void gen3_init_clock_gating(struct drm_device *dev)
 
 	/* On GEN3 we really need to make sure the ARB C3 LP bit is set */
 	I915_WRITE(MI_ARB_STATE, _MASKED_BIT_ENABLE(MI_ARB_C3_LP_WRITE_ENABLE));
+
+	I915_WRITE(MI_ARB_STATE,
+		   _MASKED_BIT_ENABLE(MI_ARB_DISPLAY_TRICKLE_FEED_DISABLE));
 }
 
 static void i85x_init_clock_gating(struct drm_device *dev)
@@ -5894,6 +7022,9 @@ static void i85x_init_clock_gating(struct drm_device *dev)
 	/* interrupts should cause a wake up from C3 */
 	I915_WRITE(MI_STATE, _MASKED_BIT_ENABLE(MI_AGPBUSY_INT_EN) |
 		   _MASKED_BIT_DISABLE(MI_AGPBUSY_830_MODE));
+
+	I915_WRITE(MEM_MODE,
+		   _MASKED_BIT_ENABLE(MEM_DISPLAY_TRICKLE_FEED_DISABLE));
 }
 
 static void i830_init_clock_gating(struct drm_device *dev)
@@ -5901,6 +7032,10 @@ static void i830_init_clock_gating(struct drm_device *dev)
 	struct drm_i915_private *dev_priv = dev->dev_private;
 
 	I915_WRITE(DSPCLK_GATE_D, OVRUNIT_CLOCK_GATE_DISABLE);
+
+	I915_WRITE(MEM_MODE,
+		   _MASKED_BIT_ENABLE(MEM_DISPLAY_A_TRICKLE_FEED_DISABLE) |
+		   _MASKED_BIT_ENABLE(MEM_DISPLAY_B_TRICKLE_FEED_DISABLE));
 }
 
 void intel_init_clock_gating(struct drm_device *dev)
@@ -5916,863 +7051,35 @@ void intel_suspend_hw(struct drm_device *dev)
 		lpt_suspend_hw(dev);
 }
 
-#define for_each_power_well(i, power_well, domain_mask, power_domains)	\
-	for (i = 0;							\
-	     i < (power_domains)->power_well_count &&			\
-		 ((power_well) = &(power_domains)->power_wells[i]);	\
-	     i++)							\
-		if ((power_well)->domains & (domain_mask))
-
-#define for_each_power_well_rev(i, power_well, domain_mask, power_domains) \
-	for (i = (power_domains)->power_well_count - 1;			 \
-	     i >= 0 && ((power_well) = &(power_domains)->power_wells[i]);\
-	     i--)							 \
-		if ((power_well)->domains & (domain_mask))
-
-/**
- * We should only use the power well if we explicitly asked the hardware to
- * enable it, so check if it's enabled and also check if we've requested it to
- * be enabled.
- */
-static bool hsw_power_well_enabled(struct drm_i915_private *dev_priv,
-				   struct i915_power_well *power_well)
-{
-	return I915_READ(HSW_PWR_WELL_DRIVER) ==
-		     (HSW_PWR_WELL_ENABLE_REQUEST | HSW_PWR_WELL_STATE_ENABLED);
-}
-
-bool intel_display_power_enabled_unlocked(struct drm_i915_private *dev_priv,
-					  enum intel_display_power_domain domain)
-{
-	struct i915_power_domains *power_domains;
-	struct i915_power_well *power_well;
-	bool is_enabled;
-	int i;
-
-	if (dev_priv->pm.suspended)
-		return false;
-
-	power_domains = &dev_priv->power_domains;
-
-	is_enabled = true;
-
-	for_each_power_well_rev(i, power_well, BIT(domain), power_domains) {
-		if (power_well->always_on)
-			continue;
-
-		if (!power_well->hw_enabled) {
-			is_enabled = false;
-			break;
-		}
-	}
-
-	return is_enabled;
-}
-
-bool intel_display_power_enabled(struct drm_i915_private *dev_priv,
-				 enum intel_display_power_domain domain)
-{
-	struct i915_power_domains *power_domains;
-	bool ret;
-
-	power_domains = &dev_priv->power_domains;
-
-	mutex_lock(&power_domains->lock);
-	ret = intel_display_power_enabled_unlocked(dev_priv, domain);
-	mutex_unlock(&power_domains->lock);
-
-	return ret;
-}
-
-/*
- * Starting with Haswell, we have a "Power Down Well" that can be turned off
- * when not needed anymore. We have 4 registers that can request the power well
- * to be enabled, and it will only be disabled if none of the registers is
- * requesting it to be enabled.
- */
-static void hsw_power_well_post_enable(struct drm_i915_private *dev_priv)
-{
-	struct drm_device *dev = dev_priv->dev;
-
-	/*
-	 * After we re-enable the power well, if we touch VGA register 0x3d5
-	 * we'll get unclaimed register interrupts. This stops after we write
-	 * anything to the VGA MSR register. The vgacon module uses this
-	 * register all the time, so if we unbind our driver and, as a
-	 * consequence, bind vgacon, we'll get stuck in an infinite loop at
-	 * console_unlock(). So make here we touch the VGA MSR register, making
-	 * sure vgacon can keep working normally without triggering interrupts
-	 * and error messages.
-	 */
-	vga_get_uninterruptible(dev->pdev, VGA_RSRC_LEGACY_IO);
-	outb(inb(VGA_MSR_READ), VGA_MSR_WRITE);
-	vga_put(dev->pdev, VGA_RSRC_LEGACY_IO);
-
-	if (IS_BROADWELL(dev))
-		gen8_irq_power_well_post_enable(dev_priv);
-}
-
-static void hsw_set_power_well(struct drm_i915_private *dev_priv,
-			       struct i915_power_well *power_well, bool enable)
-{
-	bool is_enabled, enable_requested;
-	uint32_t tmp;
-
-	tmp = I915_READ(HSW_PWR_WELL_DRIVER);
-	is_enabled = tmp & HSW_PWR_WELL_STATE_ENABLED;
-	enable_requested = tmp & HSW_PWR_WELL_ENABLE_REQUEST;
-
-	if (enable) {
-		if (!enable_requested)
-			I915_WRITE(HSW_PWR_WELL_DRIVER,
-				   HSW_PWR_WELL_ENABLE_REQUEST);
-
-		if (!is_enabled) {
-			DRM_DEBUG_KMS("Enabling power well\n");
-			if (wait_for((I915_READ(HSW_PWR_WELL_DRIVER) &
-				      HSW_PWR_WELL_STATE_ENABLED), 20))
-				DRM_ERROR("Timeout enabling power well\n");
-		}
-
-		hsw_power_well_post_enable(dev_priv);
-	} else {
-		if (enable_requested) {
-			I915_WRITE(HSW_PWR_WELL_DRIVER, 0);
-			POSTING_READ(HSW_PWR_WELL_DRIVER);
-			DRM_DEBUG_KMS("Requesting to disable the power well\n");
-		}
-	}
-}
-
-static void hsw_power_well_sync_hw(struct drm_i915_private *dev_priv,
-				   struct i915_power_well *power_well)
-{
-	hsw_set_power_well(dev_priv, power_well, power_well->count > 0);
-
-	/*
-	 * We're taking over the BIOS, so clear any requests made by it since
-	 * the driver is in charge now.
-	 */
-	if (I915_READ(HSW_PWR_WELL_BIOS) & HSW_PWR_WELL_ENABLE_REQUEST)
-		I915_WRITE(HSW_PWR_WELL_BIOS, 0);
-}
-
-static void hsw_power_well_enable(struct drm_i915_private *dev_priv,
-				  struct i915_power_well *power_well)
-{
-	hsw_set_power_well(dev_priv, power_well, true);
-}
-
-static void hsw_power_well_disable(struct drm_i915_private *dev_priv,
-				   struct i915_power_well *power_well)
-{
-	hsw_set_power_well(dev_priv, power_well, false);
-}
-
-static void i9xx_always_on_power_well_noop(struct drm_i915_private *dev_priv,
-					   struct i915_power_well *power_well)
-{
-}
-
-static bool i9xx_always_on_power_well_enabled(struct drm_i915_private *dev_priv,
-					     struct i915_power_well *power_well)
-{
-	return true;
-}
-
-static void vlv_set_power_well(struct drm_i915_private *dev_priv,
-			       struct i915_power_well *power_well, bool enable)
-{
-	enum punit_power_well power_well_id = power_well->data;
-	u32 mask;
-	u32 state;
-	u32 ctrl;
-
-	mask = PUNIT_PWRGT_MASK(power_well_id);
-	state = enable ? PUNIT_PWRGT_PWR_ON(power_well_id) :
-			 PUNIT_PWRGT_PWR_GATE(power_well_id);
-
-	mutex_lock(&dev_priv->rps.hw_lock);
-
-#define COND \
-	((vlv_punit_read(dev_priv, PUNIT_REG_PWRGT_STATUS) & mask) == state)
-
-	if (COND)
-		goto out;
-
-	ctrl = vlv_punit_read(dev_priv, PUNIT_REG_PWRGT_CTRL);
-	ctrl &= ~mask;
-	ctrl |= state;
-	vlv_punit_write(dev_priv, PUNIT_REG_PWRGT_CTRL, ctrl);
-
-	if (wait_for(COND, 100))
-		DRM_ERROR("timout setting power well state %08x (%08x)\n",
-			  state,
-			  vlv_punit_read(dev_priv, PUNIT_REG_PWRGT_CTRL));
-
-#undef COND
-
-out:
-	mutex_unlock(&dev_priv->rps.hw_lock);
-}
-
-static void vlv_power_well_sync_hw(struct drm_i915_private *dev_priv,
-				   struct i915_power_well *power_well)
-{
-	vlv_set_power_well(dev_priv, power_well, power_well->count > 0);
-}
-
-static void vlv_power_well_enable(struct drm_i915_private *dev_priv,
-				  struct i915_power_well *power_well)
-{
-	vlv_set_power_well(dev_priv, power_well, true);
-}
-
-static void vlv_power_well_disable(struct drm_i915_private *dev_priv,
-				   struct i915_power_well *power_well)
-{
-	vlv_set_power_well(dev_priv, power_well, false);
-}
-
-static bool vlv_power_well_enabled(struct drm_i915_private *dev_priv,
-				   struct i915_power_well *power_well)
-{
-	int power_well_id = power_well->data;
-	bool enabled = false;
-	u32 mask;
-	u32 state;
-	u32 ctrl;
-
-	mask = PUNIT_PWRGT_MASK(power_well_id);
-	ctrl = PUNIT_PWRGT_PWR_ON(power_well_id);
-
-	mutex_lock(&dev_priv->rps.hw_lock);
-
-	state = vlv_punit_read(dev_priv, PUNIT_REG_PWRGT_STATUS) & mask;
-	/*
-	 * We only ever set the power-on and power-gate states, anything
-	 * else is unexpected.
-	 */
-	WARN_ON(state != PUNIT_PWRGT_PWR_ON(power_well_id) &&
-		state != PUNIT_PWRGT_PWR_GATE(power_well_id));
-	if (state == ctrl)
-		enabled = true;
-
-	/*
-	 * A transient state at this point would mean some unexpected party
-	 * is poking at the power controls too.
-	 */
-	ctrl = vlv_punit_read(dev_priv, PUNIT_REG_PWRGT_CTRL) & mask;
-	WARN_ON(ctrl != state);
-
-	mutex_unlock(&dev_priv->rps.hw_lock);
-
-	return enabled;
-}
-
-static void vlv_display_power_well_enable(struct drm_i915_private *dev_priv,
-					  struct i915_power_well *power_well)
-{
-	WARN_ON_ONCE(power_well->data != PUNIT_POWER_WELL_DISP2D);
-
-	vlv_set_power_well(dev_priv, power_well, true);
-
-	spin_lock_irq(&dev_priv->irq_lock);
-	valleyview_enable_display_irqs(dev_priv);
-	spin_unlock_irq(&dev_priv->irq_lock);
-
-	/*
-	 * During driver initialization/resume we can avoid restoring the
-	 * part of the HW/SW state that will be inited anyway explicitly.
-	 */
-	if (dev_priv->power_domains.initializing)
-		return;
-
-	intel_hpd_init(dev_priv->dev);
-
-	i915_redisable_vga_power_on(dev_priv->dev);
-}
-
-static void vlv_display_power_well_disable(struct drm_i915_private *dev_priv,
-					   struct i915_power_well *power_well)
-{
-	WARN_ON_ONCE(power_well->data != PUNIT_POWER_WELL_DISP2D);
-
-	spin_lock_irq(&dev_priv->irq_lock);
-	valleyview_disable_display_irqs(dev_priv);
-	spin_unlock_irq(&dev_priv->irq_lock);
-
-	vlv_set_power_well(dev_priv, power_well, false);
-}
-
-static void vlv_dpio_cmn_power_well_enable(struct drm_i915_private *dev_priv,
-					   struct i915_power_well *power_well)
-{
-	WARN_ON_ONCE(power_well->data != PUNIT_POWER_WELL_DPIO_CMN_BC);
-
-	/*
-	 * Enable the CRI clock source so we can get at the
-	 * display and the reference clock for VGA
-	 * hotplug / manual detection.
-	 */
-	I915_WRITE(DPLL(PIPE_B), I915_READ(DPLL(PIPE_B)) |
-		   DPLL_REFA_CLK_ENABLE_VLV | DPLL_INTEGRATED_CRI_CLK_VLV);
-	udelay(1); /* >10ns for cmnreset, >0ns for sidereset */
-
-	vlv_set_power_well(dev_priv, power_well, true);
-
-	/*
-	 * From VLV2A0_DP_eDP_DPIO_driver_vbios_notes_10.docx -
-	 *  6.	De-assert cmn_reset/side_reset. Same as VLV X0.
-	 *   a.	GUnit 0x2110 bit[0] set to 1 (def 0)
-	 *   b.	The other bits such as sfr settings / modesel may all
-	 *	be set to 0.
-	 *
-	 * This should only be done on init and resume from S3 with
-	 * both PLLs disabled, or we risk losing DPIO and PLL
-	 * synchronization.
-	 */
-	I915_WRITE(DPIO_CTL, I915_READ(DPIO_CTL) | DPIO_CMNRST);
-}
-
-static void vlv_dpio_cmn_power_well_disable(struct drm_i915_private *dev_priv,
-					    struct i915_power_well *power_well)
-{
-	struct drm_device *dev = dev_priv->dev;
-	enum pipe pipe;
-
-	WARN_ON_ONCE(power_well->data != PUNIT_POWER_WELL_DPIO_CMN_BC);
-
-	for_each_pipe(pipe)
-		assert_pll_disabled(dev_priv, pipe);
-
-	/* Assert common reset */
-	I915_WRITE(DPIO_CTL, I915_READ(DPIO_CTL) & ~DPIO_CMNRST);
-
-	vlv_set_power_well(dev_priv, power_well, false);
-}
-
-static void check_power_well_state(struct drm_i915_private *dev_priv,
-				   struct i915_power_well *power_well)
+static void intel_init_fbc(struct drm_i915_private *dev_priv)
 {
-	bool enabled = power_well->ops->is_enabled(dev_priv, power_well);
-
-	if (power_well->always_on || !i915.disable_power_well) {
-		if (!enabled)
-			goto mismatch;
-
+	if (!HAS_FBC(dev_priv)) {
+		dev_priv->fbc.enabled = false;
 		return;
 	}
 
-	if (enabled != (power_well->count > 0))
-		goto mismatch;
-
-	return;
-
-mismatch:
-	WARN(1, "state mismatch for '%s' (always_on %d hw state %d use-count %d disable_power_well %d\n",
-		  power_well->name, power_well->always_on, enabled,
-		  power_well->count, i915.disable_power_well);
-}
-
-void intel_display_power_get(struct drm_i915_private *dev_priv,
-			     enum intel_display_power_domain domain)
-{
-	struct i915_power_domains *power_domains;
-	struct i915_power_well *power_well;
-	int i;
-
-	intel_runtime_pm_get(dev_priv);
-
-	power_domains = &dev_priv->power_domains;
-
-	mutex_lock(&power_domains->lock);
-
-	for_each_power_well(i, power_well, BIT(domain), power_domains) {
-		if (!power_well->count++) {
-			DRM_DEBUG_KMS("enabling %s\n", power_well->name);
-			power_well->ops->enable(dev_priv, power_well);
-			power_well->hw_enabled = true;
-		}
-
-		check_power_well_state(dev_priv, power_well);
-	}
-
-	power_domains->domain_use_count[domain]++;
-
-	mutex_unlock(&power_domains->lock);
-}
-
-void intel_display_power_put(struct drm_i915_private *dev_priv,
-			     enum intel_display_power_domain domain)
-{
-	struct i915_power_domains *power_domains;
-	struct i915_power_well *power_well;
-	int i;
-
-	power_domains = &dev_priv->power_domains;
-
-	mutex_lock(&power_domains->lock);
-
-	WARN_ON(!power_domains->domain_use_count[domain]);
-	power_domains->domain_use_count[domain]--;
-
-	for_each_power_well_rev(i, power_well, BIT(domain), power_domains) {
-		WARN_ON(!power_well->count);
-
-		if (!--power_well->count && i915.disable_power_well) {
-			DRM_DEBUG_KMS("disabling %s\n", power_well->name);
-			power_well->hw_enabled = false;
-			power_well->ops->disable(dev_priv, power_well);
-		}
-
-		check_power_well_state(dev_priv, power_well);
-	}
-
-	mutex_unlock(&power_domains->lock);
-
-	intel_runtime_pm_put(dev_priv);
-}
-
-static struct i915_power_domains *hsw_pwr;
-
-/* Display audio driver power well request */
-int i915_request_power_well(void)
-{
-	struct drm_i915_private *dev_priv;
-
-	if (!hsw_pwr)
-		return -ENODEV;
-
-	dev_priv = container_of(hsw_pwr, struct drm_i915_private,
-				power_domains);
-	intel_display_power_get(dev_priv, POWER_DOMAIN_AUDIO);
-	return 0;
-}
-EXPORT_SYMBOL_GPL(i915_request_power_well);
-
-/* Display audio driver power well release */
-int i915_release_power_well(void)
-{
-	struct drm_i915_private *dev_priv;
-
-	if (!hsw_pwr)
-		return -ENODEV;
-
-	dev_priv = container_of(hsw_pwr, struct drm_i915_private,
-				power_domains);
-	intel_display_power_put(dev_priv, POWER_DOMAIN_AUDIO);
-	return 0;
-}
-EXPORT_SYMBOL_GPL(i915_release_power_well);
-
-/*
- * Private interface for the audio driver to get CDCLK in kHz.
- *
- * Caller must request power well using i915_request_power_well() prior to
- * making the call.
- */
-int i915_get_cdclk_freq(void)
-{
-	struct drm_i915_private *dev_priv;
-
-	if (!hsw_pwr)
-		return -ENODEV;
-
-	dev_priv = container_of(hsw_pwr, struct drm_i915_private,
-				power_domains);
-
-	return intel_ddi_get_cdclk_freq(dev_priv);
-}
-EXPORT_SYMBOL_GPL(i915_get_cdclk_freq);
-
-
-#define POWER_DOMAIN_MASK (BIT(POWER_DOMAIN_NUM) - 1)
-
-#define HSW_ALWAYS_ON_POWER_DOMAINS (			\
-	BIT(POWER_DOMAIN_PIPE_A) |			\
-	BIT(POWER_DOMAIN_TRANSCODER_EDP) |		\
-	BIT(POWER_DOMAIN_PORT_DDI_A_2_LANES) |		\
-	BIT(POWER_DOMAIN_PORT_DDI_A_4_LANES) |		\
-	BIT(POWER_DOMAIN_PORT_DDI_B_2_LANES) |		\
-	BIT(POWER_DOMAIN_PORT_DDI_B_4_LANES) |		\
-	BIT(POWER_DOMAIN_PORT_DDI_C_2_LANES) |		\
-	BIT(POWER_DOMAIN_PORT_DDI_C_4_LANES) |		\
-	BIT(POWER_DOMAIN_PORT_DDI_D_2_LANES) |		\
-	BIT(POWER_DOMAIN_PORT_DDI_D_4_LANES) |		\
-	BIT(POWER_DOMAIN_PORT_CRT) |			\
-	BIT(POWER_DOMAIN_PLLS) |			\
-	BIT(POWER_DOMAIN_INIT))
-#define HSW_DISPLAY_POWER_DOMAINS (				\
-	(POWER_DOMAIN_MASK & ~HSW_ALWAYS_ON_POWER_DOMAINS) |	\
-	BIT(POWER_DOMAIN_INIT))
-
-#define BDW_ALWAYS_ON_POWER_DOMAINS (			\
-	HSW_ALWAYS_ON_POWER_DOMAINS |			\
-	BIT(POWER_DOMAIN_PIPE_A_PANEL_FITTER))
-#define BDW_DISPLAY_POWER_DOMAINS (				\
-	(POWER_DOMAIN_MASK & ~BDW_ALWAYS_ON_POWER_DOMAINS) |	\
-	BIT(POWER_DOMAIN_INIT))
-
-#define VLV_ALWAYS_ON_POWER_DOMAINS	BIT(POWER_DOMAIN_INIT)
-#define VLV_DISPLAY_POWER_DOMAINS	POWER_DOMAIN_MASK
-
-#define VLV_DPIO_CMN_BC_POWER_DOMAINS (		\
-	BIT(POWER_DOMAIN_PORT_DDI_B_2_LANES) |	\
-	BIT(POWER_DOMAIN_PORT_DDI_B_4_LANES) |	\
-	BIT(POWER_DOMAIN_PORT_DDI_C_2_LANES) |	\
-	BIT(POWER_DOMAIN_PORT_DDI_C_4_LANES) |	\
-	BIT(POWER_DOMAIN_PORT_CRT) |		\
-	BIT(POWER_DOMAIN_INIT))
-
-#define VLV_DPIO_TX_B_LANES_01_POWER_DOMAINS (	\
-	BIT(POWER_DOMAIN_PORT_DDI_B_2_LANES) |	\
-	BIT(POWER_DOMAIN_PORT_DDI_B_4_LANES) |	\
-	BIT(POWER_DOMAIN_INIT))
-
-#define VLV_DPIO_TX_B_LANES_23_POWER_DOMAINS (	\
-	BIT(POWER_DOMAIN_PORT_DDI_B_4_LANES) |	\
-	BIT(POWER_DOMAIN_INIT))
-
-#define VLV_DPIO_TX_C_LANES_01_POWER_DOMAINS (	\
-	BIT(POWER_DOMAIN_PORT_DDI_C_2_LANES) |	\
-	BIT(POWER_DOMAIN_PORT_DDI_C_4_LANES) |	\
-	BIT(POWER_DOMAIN_INIT))
-
-#define VLV_DPIO_TX_C_LANES_23_POWER_DOMAINS (	\
-	BIT(POWER_DOMAIN_PORT_DDI_C_4_LANES) |	\
-	BIT(POWER_DOMAIN_INIT))
-
-static const struct i915_power_well_ops i9xx_always_on_power_well_ops = {
-	.sync_hw = i9xx_always_on_power_well_noop,
-	.enable = i9xx_always_on_power_well_noop,
-	.disable = i9xx_always_on_power_well_noop,
-	.is_enabled = i9xx_always_on_power_well_enabled,
-};
-
-static struct i915_power_well i9xx_always_on_power_well[] = {
-	{
-		.name = "always-on",
-		.always_on = 1,
-		.domains = POWER_DOMAIN_MASK,
-		.ops = &i9xx_always_on_power_well_ops,
-	},
-};
-
-static const struct i915_power_well_ops hsw_power_well_ops = {
-	.sync_hw = hsw_power_well_sync_hw,
-	.enable = hsw_power_well_enable,
-	.disable = hsw_power_well_disable,
-	.is_enabled = hsw_power_well_enabled,
-};
-
-static struct i915_power_well hsw_power_wells[] = {
-	{
-		.name = "always-on",
-		.always_on = 1,
-		.domains = HSW_ALWAYS_ON_POWER_DOMAINS,
-		.ops = &i9xx_always_on_power_well_ops,
-	},
-	{
-		.name = "display",
-		.domains = HSW_DISPLAY_POWER_DOMAINS,
-		.ops = &hsw_power_well_ops,
-	},
-};
-
-static struct i915_power_well bdw_power_wells[] = {
-	{
-		.name = "always-on",
-		.always_on = 1,
-		.domains = BDW_ALWAYS_ON_POWER_DOMAINS,
-		.ops = &i9xx_always_on_power_well_ops,
-	},
-	{
-		.name = "display",
-		.domains = BDW_DISPLAY_POWER_DOMAINS,
-		.ops = &hsw_power_well_ops,
-	},
-};
-
-static const struct i915_power_well_ops vlv_display_power_well_ops = {
-	.sync_hw = vlv_power_well_sync_hw,
-	.enable = vlv_display_power_well_enable,
-	.disable = vlv_display_power_well_disable,
-	.is_enabled = vlv_power_well_enabled,
-};
-
-static const struct i915_power_well_ops vlv_dpio_cmn_power_well_ops = {
-	.sync_hw = vlv_power_well_sync_hw,
-	.enable = vlv_dpio_cmn_power_well_enable,
-	.disable = vlv_dpio_cmn_power_well_disable,
-	.is_enabled = vlv_power_well_enabled,
-};
-
-static const struct i915_power_well_ops vlv_dpio_power_well_ops = {
-	.sync_hw = vlv_power_well_sync_hw,
-	.enable = vlv_power_well_enable,
-	.disable = vlv_power_well_disable,
-	.is_enabled = vlv_power_well_enabled,
-};
-
-static struct i915_power_well vlv_power_wells[] = {
-	{
-		.name = "always-on",
-		.always_on = 1,
-		.domains = VLV_ALWAYS_ON_POWER_DOMAINS,
-		.ops = &i9xx_always_on_power_well_ops,
-	},
-	{
-		.name = "display",
-		.domains = VLV_DISPLAY_POWER_DOMAINS,
-		.data = PUNIT_POWER_WELL_DISP2D,
-		.ops = &vlv_display_power_well_ops,
-	},
-	{
-		.name = "dpio-tx-b-01",
-		.domains = VLV_DPIO_TX_B_LANES_01_POWER_DOMAINS |
-			   VLV_DPIO_TX_B_LANES_23_POWER_DOMAINS |
-			   VLV_DPIO_TX_C_LANES_01_POWER_DOMAINS |
-			   VLV_DPIO_TX_C_LANES_23_POWER_DOMAINS,
-		.ops = &vlv_dpio_power_well_ops,
-		.data = PUNIT_POWER_WELL_DPIO_TX_B_LANES_01,
-	},
-	{
-		.name = "dpio-tx-b-23",
-		.domains = VLV_DPIO_TX_B_LANES_01_POWER_DOMAINS |
-			   VLV_DPIO_TX_B_LANES_23_POWER_DOMAINS |
-			   VLV_DPIO_TX_C_LANES_01_POWER_DOMAINS |
-			   VLV_DPIO_TX_C_LANES_23_POWER_DOMAINS,
-		.ops = &vlv_dpio_power_well_ops,
-		.data = PUNIT_POWER_WELL_DPIO_TX_B_LANES_23,
-	},
-	{
-		.name = "dpio-tx-c-01",
-		.domains = VLV_DPIO_TX_B_LANES_01_POWER_DOMAINS |
-			   VLV_DPIO_TX_B_LANES_23_POWER_DOMAINS |
-			   VLV_DPIO_TX_C_LANES_01_POWER_DOMAINS |
-			   VLV_DPIO_TX_C_LANES_23_POWER_DOMAINS,
-		.ops = &vlv_dpio_power_well_ops,
-		.data = PUNIT_POWER_WELL_DPIO_TX_C_LANES_01,
-	},
-	{
-		.name = "dpio-tx-c-23",
-		.domains = VLV_DPIO_TX_B_LANES_01_POWER_DOMAINS |
-			   VLV_DPIO_TX_B_LANES_23_POWER_DOMAINS |
-			   VLV_DPIO_TX_C_LANES_01_POWER_DOMAINS |
-			   VLV_DPIO_TX_C_LANES_23_POWER_DOMAINS,
-		.ops = &vlv_dpio_power_well_ops,
-		.data = PUNIT_POWER_WELL_DPIO_TX_C_LANES_23,
-	},
-	{
-		.name = "dpio-common",
-		.domains = VLV_DPIO_CMN_BC_POWER_DOMAINS,
-		.data = PUNIT_POWER_WELL_DPIO_CMN_BC,
-		.ops = &vlv_dpio_cmn_power_well_ops,
-	},
-};
-
-static struct i915_power_well *lookup_power_well(struct drm_i915_private *dev_priv,
-						 enum punit_power_well power_well_id)
-{
-	struct i915_power_domains *power_domains = &dev_priv->power_domains;
-	struct i915_power_well *power_well;
-	int i;
-
-	for_each_power_well(i, power_well, POWER_DOMAIN_MASK, power_domains) {
-		if (power_well->data == power_well_id)
-			return power_well;
-	}
-
-	return NULL;
-}
-
-#define set_power_wells(power_domains, __power_wells) ({		\
-	(power_domains)->power_wells = (__power_wells);			\
-	(power_domains)->power_well_count = ARRAY_SIZE(__power_wells);	\
-})
-
-int intel_power_domains_init(struct drm_i915_private *dev_priv)
-{
-	struct i915_power_domains *power_domains = &dev_priv->power_domains;
-
-	mutex_init(&power_domains->lock);
-
-	/*
-	 * The enabling order will be from lower to higher indexed wells,
-	 * the disabling order is reversed.
-	 */
-	if (IS_HASWELL(dev_priv->dev)) {
-		set_power_wells(power_domains, hsw_power_wells);
-		hsw_pwr = power_domains;
-	} else if (IS_BROADWELL(dev_priv->dev)) {
-		set_power_wells(power_domains, bdw_power_wells);
-		hsw_pwr = power_domains;
-	} else if (IS_VALLEYVIEW(dev_priv->dev)) {
-		set_power_wells(power_domains, vlv_power_wells);
+	if (INTEL_INFO(dev_priv)->gen >= 7) {
+		dev_priv->display.fbc_enabled = ironlake_fbc_enabled;
+		dev_priv->display.enable_fbc = gen7_enable_fbc;
+		dev_priv->display.disable_fbc = ironlake_disable_fbc;
+	} else if (INTEL_INFO(dev_priv)->gen >= 5) {
+		dev_priv->display.fbc_enabled = ironlake_fbc_enabled;
+		dev_priv->display.enable_fbc = ironlake_enable_fbc;
+		dev_priv->display.disable_fbc = ironlake_disable_fbc;
+	} else if (IS_GM45(dev_priv)) {
+		dev_priv->display.fbc_enabled = g4x_fbc_enabled;
+		dev_priv->display.enable_fbc = g4x_enable_fbc;
+		dev_priv->display.disable_fbc = g4x_disable_fbc;
 	} else {
-		set_power_wells(power_domains, i9xx_always_on_power_well);
-	}
-
-	return 0;
-}
-
-void intel_power_domains_remove(struct drm_i915_private *dev_priv)
-{
-	hsw_pwr = NULL;
-}
-
-static void intel_power_domains_resume(struct drm_i915_private *dev_priv)
-{
-	struct i915_power_domains *power_domains = &dev_priv->power_domains;
-	struct i915_power_well *power_well;
-	int i;
-
-	mutex_lock(&power_domains->lock);
-	for_each_power_well(i, power_well, POWER_DOMAIN_MASK, power_domains) {
-		power_well->ops->sync_hw(dev_priv, power_well);
-		power_well->hw_enabled = power_well->ops->is_enabled(dev_priv,
-								     power_well);
-	}
-	mutex_unlock(&power_domains->lock);
-}
-
-static void vlv_cmnlane_wa(struct drm_i915_private *dev_priv)
-{
-	struct i915_power_well *cmn =
-		lookup_power_well(dev_priv, PUNIT_POWER_WELL_DPIO_CMN_BC);
-	struct i915_power_well *disp2d =
-		lookup_power_well(dev_priv, PUNIT_POWER_WELL_DISP2D);
-
-	/* nothing to do if common lane is already off */
-	if (!cmn->ops->is_enabled(dev_priv, cmn))
-		return;
-
-	/* If the display might be already active skip this */
-	if (disp2d->ops->is_enabled(dev_priv, disp2d) &&
-	    I915_READ(DPIO_CTL) & DPIO_CMNRST)
-		return;
-
-	DRM_DEBUG_KMS("toggling display PHY side reset\n");
-
-	/* cmnlane needs DPLL registers */
-	disp2d->ops->enable(dev_priv, disp2d);
-
-	/*
-	 * From VLV2A0_DP_eDP_HDMI_DPIO_driver_vbios_notes_11.docx:
-	 * Need to assert and de-assert PHY SB reset by gating the
-	 * common lane power, then un-gating it.
-	 * Simply ungating isn't enough to reset the PHY enough to get
-	 * ports and lanes running.
-	 */
-	cmn->ops->disable(dev_priv, cmn);
-}
-
-void intel_power_domains_init_hw(struct drm_i915_private *dev_priv)
-{
-	struct drm_device *dev = dev_priv->dev;
-	struct i915_power_domains *power_domains = &dev_priv->power_domains;
-
-	power_domains->initializing = true;
+		dev_priv->display.fbc_enabled = i8xx_fbc_enabled;
+		dev_priv->display.enable_fbc = i8xx_enable_fbc;
+		dev_priv->display.disable_fbc = i8xx_disable_fbc;
 
-	if (IS_VALLEYVIEW(dev) && !IS_CHERRYVIEW(dev)) {
-		mutex_lock(&power_domains->lock);
-		vlv_cmnlane_wa(dev_priv);
-		mutex_unlock(&power_domains->lock);
+		/* This value was pulled out of someone's hat */
+		I915_WRITE(FBC_CONTROL, 500 << FBC_CTL_INTERVAL_SHIFT);
 	}
 
-	/* For now, we need the power well to be always enabled. */
-	intel_display_set_init_power(dev_priv, true);
-	intel_power_domains_resume(dev_priv);
-	power_domains->initializing = false;
-}
-
-void intel_aux_display_runtime_get(struct drm_i915_private *dev_priv)
-{
-	intel_runtime_pm_get(dev_priv);
-}
-
-void intel_aux_display_runtime_put(struct drm_i915_private *dev_priv)
-{
-	intel_runtime_pm_put(dev_priv);
-}
-
-void intel_runtime_pm_get(struct drm_i915_private *dev_priv)
-{
-	struct drm_device *dev = dev_priv->dev;
-	struct device *device = &dev->pdev->dev;
-
-	if (!HAS_RUNTIME_PM(dev))
-		return;
-
-	pm_runtime_get_sync(device);
-	WARN(dev_priv->pm.suspended, "Device still suspended.\n");
-}
-
-void intel_runtime_pm_get_noresume(struct drm_i915_private *dev_priv)
-{
-	struct drm_device *dev = dev_priv->dev;
-	struct device *device = &dev->pdev->dev;
-
-	if (!HAS_RUNTIME_PM(dev))
-		return;
-
-	WARN(dev_priv->pm.suspended, "Getting nosync-ref while suspended.\n");
-	pm_runtime_get_noresume(device);
-}
-
-void intel_runtime_pm_put(struct drm_i915_private *dev_priv)
-{
-	struct drm_device *dev = dev_priv->dev;
-	struct device *device = &dev->pdev->dev;
-
-	if (!HAS_RUNTIME_PM(dev))
-		return;
-
-	pm_runtime_mark_last_busy(device);
-	pm_runtime_put_autosuspend(device);
-}
-
-void intel_init_runtime_pm(struct drm_i915_private *dev_priv)
-{
-	struct drm_device *dev = dev_priv->dev;
-	struct device *device = &dev->pdev->dev;
-
-	if (!HAS_RUNTIME_PM(dev))
-		return;
-
-	pm_runtime_set_active(device);
-
-	/*
-	 * RPM depends on RC6 to save restore the GT HW context, so make RC6 a
-	 * requirement.
-	 */
-	if (!intel_enable_rc6(dev)) {
-		DRM_INFO("RC6 disabled, disabling runtime PM support\n");
-		return;
-	}
-
-	pm_runtime_set_autosuspend_delay(device, 10000); /* 10s */
-	pm_runtime_mark_last_busy(device);
-	pm_runtime_use_autosuspend(device);
-
-	pm_runtime_put_autosuspend(device);
-}
-
-void intel_fini_runtime_pm(struct drm_i915_private *dev_priv)
-{
-	struct drm_device *dev = dev_priv->dev;
-	struct device *device = &dev->pdev->dev;
-
-	if (!HAS_RUNTIME_PM(dev))
-		return;
-
-	if (!intel_enable_rc6(dev))
-		return;
-
-	/* Make sure we're not suspended first. */
-	pm_runtime_get_sync(device);
-	pm_runtime_disable(device);
+	dev_priv->fbc.enabled = dev_priv->display.fbc_enabled(dev_priv->dev);
 }
 
 /* Set up chip specific power management-related functions */
@@ -6780,28 +7087,7 @@ void intel_init_pm(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 
-	if (HAS_FBC(dev)) {
-		if (INTEL_INFO(dev)->gen >= 7) {
-			dev_priv->display.fbc_enabled = ironlake_fbc_enabled;
-			dev_priv->display.enable_fbc = gen7_enable_fbc;
-			dev_priv->display.disable_fbc = ironlake_disable_fbc;
-		} else if (INTEL_INFO(dev)->gen >= 5) {
-			dev_priv->display.fbc_enabled = ironlake_fbc_enabled;
-			dev_priv->display.enable_fbc = ironlake_enable_fbc;
-			dev_priv->display.disable_fbc = ironlake_disable_fbc;
-		} else if (IS_GM45(dev)) {
-			dev_priv->display.fbc_enabled = g4x_fbc_enabled;
-			dev_priv->display.enable_fbc = g4x_enable_fbc;
-			dev_priv->display.disable_fbc = g4x_disable_fbc;
-		} else {
-			dev_priv->display.fbc_enabled = i8xx_fbc_enabled;
-			dev_priv->display.enable_fbc = i8xx_enable_fbc;
-			dev_priv->display.disable_fbc = i8xx_disable_fbc;
-
-			/* This value was pulled out of someone's hat */
-			I915_WRITE(FBC_CONTROL, 500 << FBC_CTL_INTERVAL_SHIFT);
-		}
-	}
+	intel_init_fbc(dev_priv);
 
 	/* For cxsr */
 	if (IS_PINEVIEW(dev))
@@ -6810,7 +7096,13 @@ void intel_init_pm(struct drm_device *dev)
 		i915_ironlake_get_mem_freq(dev);
 
 	/* For FIFO watermark updates */
-	if (HAS_PCH_SPLIT(dev)) {
+	if (INTEL_INFO(dev)->gen >= 9) {
+		skl_setup_wm_latency(dev);
+
+		dev_priv->display.init_clock_gating = gen9_init_clock_gating;
+		dev_priv->display.update_wm = skl_update_wm;
+		dev_priv->display.update_sprite_wm = skl_update_sprite_wm;
+	} else if (HAS_PCH_SPLIT(dev)) {
 		ilk_setup_wm_latency(dev);
 
 		if ((IS_GEN5(dev) && dev_priv->wm.pri_latency[1] &&
@@ -6833,13 +7125,15 @@ void intel_init_pm(struct drm_device *dev)
 		else if (IS_HASWELL(dev))
 			dev_priv->display.init_clock_gating = haswell_init_clock_gating;
 		else if (INTEL_INFO(dev)->gen == 8)
-			dev_priv->display.init_clock_gating = gen8_init_clock_gating;
+			dev_priv->display.init_clock_gating = broadwell_init_clock_gating;
 	} else if (IS_CHERRYVIEW(dev)) {
-		dev_priv->display.update_wm = valleyview_update_wm;
+		dev_priv->display.update_wm = cherryview_update_wm;
+		dev_priv->display.update_sprite_wm = valleyview_update_sprite_wm;
 		dev_priv->display.init_clock_gating =
 			cherryview_init_clock_gating;
 	} else if (IS_VALLEYVIEW(dev)) {
 		dev_priv->display.update_wm = valleyview_update_wm;
+		dev_priv->display.update_sprite_wm = valleyview_update_sprite_wm;
 		dev_priv->display.init_clock_gating =
 			valleyview_init_clock_gating;
 	} else if (IS_PINEVIEW(dev)) {
@@ -6889,7 +7183,7 @@ void intel_init_pm(struct drm_device *dev)
 	}
 }
 
-int sandybridge_pcode_read(struct drm_i915_private *dev_priv, u8 mbox, u32 *val)
+int sandybridge_pcode_read(struct drm_i915_private *dev_priv, u32 mbox, u32 *val)
 {
 	WARN_ON(!mutex_is_locked(&dev_priv->rps.hw_lock));
 
@@ -6899,6 +7193,7 @@ int sandybridge_pcode_read(struct drm_i915_private *dev_priv, u8 mbox, u32 *val)
 	}
 
 	I915_WRITE(GEN6_PCODE_DATA, *val);
+	I915_WRITE(GEN6_PCODE_DATA1, 0);
 	I915_WRITE(GEN6_PCODE_MAILBOX, GEN6_PCODE_READY | mbox);
 
 	if (wait_for((I915_READ(GEN6_PCODE_MAILBOX) & GEN6_PCODE_READY) == 0,
@@ -6913,7 +7208,7 @@ int sandybridge_pcode_read(struct drm_i915_private *dev_priv, u8 mbox, u32 *val)
 	return 0;
 }
 
-int sandybridge_pcode_write(struct drm_i915_private *dev_priv, u8 mbox, u32 val)
+int sandybridge_pcode_write(struct drm_i915_private *dev_priv, u32 mbox, u32 val)
 {
 	WARN_ON(!mutex_is_locked(&dev_priv->rps.hw_lock));
 
@@ -6936,98 +7231,66 @@ int sandybridge_pcode_write(struct drm_i915_private *dev_priv, u8 mbox, u32 val)
 	return 0;
 }
 
-static int byt_gpu_freq(struct drm_i915_private *dev_priv, int val)
+static int vlv_gpu_freq_div(unsigned int czclk_freq)
 {
-	int div;
-
-	/* 4 x czclk */
-	switch (dev_priv->mem_freq) {
-	case 800:
-		div = 10;
-		break;
-	case 1066:
-		div = 12;
-		break;
-	case 1333:
-		div = 16;
-		break;
+	switch (czclk_freq) {
+	case 200:
+		return 10;
+	case 267:
+		return 12;
+	case 320:
+	case 333:
+		return 16;
+	case 400:
+		return 20;
 	default:
 		return -1;
 	}
+}
 
-	return DIV_ROUND_CLOSEST(dev_priv->mem_freq * (val + 6 - 0xbd), 4 * div);
+static int byt_gpu_freq(struct drm_i915_private *dev_priv, int val)
+{
+	int div, czclk_freq = DIV_ROUND_CLOSEST(dev_priv->mem_freq, 4);
+
+	div = vlv_gpu_freq_div(czclk_freq);
+	if (div < 0)
+		return div;
+
+	return DIV_ROUND_CLOSEST(czclk_freq * (val + 6 - 0xbd), div);
 }
 
 static int byt_freq_opcode(struct drm_i915_private *dev_priv, int val)
 {
-	int mul;
+	int mul, czclk_freq = DIV_ROUND_CLOSEST(dev_priv->mem_freq, 4);
 
-	/* 4 x czclk */
-	switch (dev_priv->mem_freq) {
-	case 800:
-		mul = 10;
-		break;
-	case 1066:
-		mul = 12;
-		break;
-	case 1333:
-		mul = 16;
-		break;
-	default:
-		return -1;
-	}
+	mul = vlv_gpu_freq_div(czclk_freq);
+	if (mul < 0)
+		return mul;
 
-	return DIV_ROUND_CLOSEST(4 * mul * val, dev_priv->mem_freq) + 0xbd - 6;
+	return DIV_ROUND_CLOSEST(mul * val, czclk_freq) + 0xbd - 6;
 }
 
 static int chv_gpu_freq(struct drm_i915_private *dev_priv, int val)
 {
-	int div, freq;
+	int div, czclk_freq = dev_priv->rps.cz_freq;
 
-	switch (dev_priv->rps.cz_freq) {
-	case 200:
-		div = 5;
-		break;
-	case 267:
-		div = 6;
-		break;
-	case 320:
-	case 333:
-	case 400:
-		div = 8;
-		break;
-	default:
-		return -1;
-	}
+	div = vlv_gpu_freq_div(czclk_freq) / 2;
+	if (div < 0)
+		return div;
 
-	freq = (DIV_ROUND_CLOSEST((dev_priv->rps.cz_freq * val), 2 * div) / 2);
-
-	return freq;
+	return DIV_ROUND_CLOSEST(czclk_freq * val, 2 * div) / 2;
 }
 
 static int chv_freq_opcode(struct drm_i915_private *dev_priv, int val)
 {
-	int mul, opcode;
-
-	switch (dev_priv->rps.cz_freq) {
-	case 200:
-		mul = 5;
-		break;
-	case 267:
-		mul = 6;
-		break;
-	case 320:
-	case 333:
-	case 400:
-		mul = 8;
-		break;
-	default:
-		return -1;
-	}
+	int mul, czclk_freq = dev_priv->rps.cz_freq;
 
-	opcode = (DIV_ROUND_CLOSEST((val * 2 * mul), dev_priv->rps.cz_freq) * 2);
+	mul = vlv_gpu_freq_div(czclk_freq) / 2;
+	if (mul < 0)
+		return mul;
 
-	return opcode;
+	/* CHV needs even values */
+	return DIV_ROUND_CLOSEST(val * 2 * mul, czclk_freq) * 2;
 }
 
 int vlv_gpu_freq(struct drm_i915_private *dev_priv, int val)
@@ -7064,5 +7327,4 @@ void intel_pm_setup(struct drm_device *dev)
 			  intel_gen6_powersave_work);
 
 	dev_priv->pm.suspended = false;
-	dev_priv->pm._irqs_disabled = false;
 }
diff --git a/drivers/gpu/drm/i915/intel_psr.c b/drivers/gpu/drm/i915/intel_psr.c
new file mode 100644
index 000000000000..716b8a961eea
--- /dev/null
+++ b/drivers/gpu/drm/i915/intel_psr.c
@@ -0,0 +1,481 @@
+/*
+ * Copyright © 2014 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ */
+
+/**
+ * DOC: Panel Self Refresh (PSR/SRD)
+ *
+ * Since Haswell Display controller supports Panel Self-Refresh on display
+ * panels witch have a remote frame buffer (RFB) implemented according to PSR
+ * spec in eDP1.3. PSR feature allows the display to go to lower standby states
+ * when system is idle but display is on as it eliminates display refresh
+ * request to DDR memory completely as long as the frame buffer for that
+ * display is unchanged.
+ *
+ * Panel Self Refresh must be supported by both Hardware (source) and
+ * Panel (sink).
+ *
+ * PSR saves power by caching the framebuffer in the panel RFB, which allows us
+ * to power down the link and memory controller. For DSI panels the same idea
+ * is called "manual mode".
+ *
+ * The implementation uses the hardware-based PSR support which automatically
+ * enters/exits self-refresh mode. The hardware takes care of sending the
+ * required DP aux message and could even retrain the link (that part isn't
+ * enabled yet though). The hardware also keeps track of any frontbuffer
+ * changes to know when to exit self-refresh mode again. Unfortunately that
+ * part doesn't work too well, hence why the i915 PSR support uses the
+ * software frontbuffer tracking to make sure it doesn't miss a screen
+ * update. For this integration intel_psr_invalidate() and intel_psr_flush()
+ * get called by the frontbuffer tracking code. Note that because of locking
+ * issues the self-refresh re-enable code is done from a work queue, which
+ * must be correctly synchronized/cancelled when shutting down the pipe."
+ */
+
+#include <drm/drmP.h>
+
+#include "intel_drv.h"
+#include "i915_drv.h"
+
+static bool is_edp_psr(struct intel_dp *intel_dp)
+{
+	return intel_dp->psr_dpcd[0] & DP_PSR_IS_SUPPORTED;
+}
+
+bool intel_psr_is_enabled(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+
+	if (!HAS_PSR(dev))
+		return false;
+
+	return I915_READ(EDP_PSR_CTL(dev)) & EDP_PSR_ENABLE;
+}
+
+static void intel_psr_write_vsc(struct intel_dp *intel_dp,
+				    struct edp_vsc_psr *vsc_psr)
+{
+	struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp);
+	struct drm_device *dev = dig_port->base.base.dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct intel_crtc *crtc = to_intel_crtc(dig_port->base.base.crtc);
+	u32 ctl_reg = HSW_TVIDEO_DIP_CTL(crtc->config.cpu_transcoder);
+	u32 data_reg = HSW_TVIDEO_DIP_VSC_DATA(crtc->config.cpu_transcoder);
+	uint32_t *data = (uint32_t *) vsc_psr;
+	unsigned int i;
+
+	/* As per BSPec (Pipe Video Data Island Packet), we need to disable
+	   the video DIP being updated before program video DIP data buffer
+	   registers for DIP being updated. */
+	I915_WRITE(ctl_reg, 0);
+	POSTING_READ(ctl_reg);
+
+	for (i = 0; i < VIDEO_DIP_VSC_DATA_SIZE; i += 4) {
+		if (i < sizeof(struct edp_vsc_psr))
+			I915_WRITE(data_reg + i, *data++);
+		else
+			I915_WRITE(data_reg + i, 0);
+	}
+
+	I915_WRITE(ctl_reg, VIDEO_DIP_ENABLE_VSC_HSW);
+	POSTING_READ(ctl_reg);
+}
+
+static void intel_psr_setup_vsc(struct intel_dp *intel_dp)
+{
+	struct edp_vsc_psr psr_vsc;
+
+	/* Prepare VSC packet as per EDP 1.3 spec, Table 3.10 */
+	memset(&psr_vsc, 0, sizeof(psr_vsc));
+	psr_vsc.sdp_header.HB0 = 0;
+	psr_vsc.sdp_header.HB1 = 0x7;
+	psr_vsc.sdp_header.HB2 = 0x2;
+	psr_vsc.sdp_header.HB3 = 0x8;
+	intel_psr_write_vsc(intel_dp, &psr_vsc);
+}
+
+static void intel_psr_enable_sink(struct intel_dp *intel_dp)
+{
+	struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp);
+	struct drm_device *dev = dig_port->base.base.dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	uint32_t aux_clock_divider;
+	int precharge = 0x3;
+	bool only_standby = false;
+	static const uint8_t aux_msg[] = {
+		[0] = DP_AUX_NATIVE_WRITE << 4,
+		[1] = DP_SET_POWER >> 8,
+		[2] = DP_SET_POWER & 0xff,
+		[3] = 1 - 1,
+		[4] = DP_SET_POWER_D0,
+	};
+	int i;
+
+	BUILD_BUG_ON(sizeof(aux_msg) > 20);
+
+	aux_clock_divider = intel_dp->get_aux_clock_divider(intel_dp, 0);
+
+	if (IS_BROADWELL(dev) && dig_port->port != PORT_A)
+		only_standby = true;
+
+	/* Enable PSR in sink */
+	if (intel_dp->psr_dpcd[1] & DP_PSR_NO_TRAIN_ON_EXIT || only_standby)
+		drm_dp_dpcd_writeb(&intel_dp->aux, DP_PSR_EN_CFG,
+				   DP_PSR_ENABLE & ~DP_PSR_MAIN_LINK_ACTIVE);
+	else
+		drm_dp_dpcd_writeb(&intel_dp->aux, DP_PSR_EN_CFG,
+				   DP_PSR_ENABLE | DP_PSR_MAIN_LINK_ACTIVE);
+
+	/* Setup AUX registers */
+	for (i = 0; i < sizeof(aux_msg); i += 4)
+		I915_WRITE(EDP_PSR_AUX_DATA1(dev) + i,
+			   intel_dp_pack_aux(&aux_msg[i], sizeof(aux_msg) - i));
+
+	I915_WRITE(EDP_PSR_AUX_CTL(dev),
+		   DP_AUX_CH_CTL_TIME_OUT_400us |
+		   (sizeof(aux_msg) << DP_AUX_CH_CTL_MESSAGE_SIZE_SHIFT) |
+		   (precharge << DP_AUX_CH_CTL_PRECHARGE_2US_SHIFT) |
+		   (aux_clock_divider << DP_AUX_CH_CTL_BIT_CLOCK_2X_SHIFT));
+}
+
+static void intel_psr_enable_source(struct intel_dp *intel_dp)
+{
+	struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp);
+	struct drm_device *dev = dig_port->base.base.dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	uint32_t max_sleep_time = 0x1f;
+	uint32_t idle_frames = 1;
+	uint32_t val = 0x0;
+	const uint32_t link_entry_time = EDP_PSR_MIN_LINK_ENTRY_TIME_8_LINES;
+	bool only_standby = false;
+
+	if (IS_BROADWELL(dev) && dig_port->port != PORT_A)
+		only_standby = true;
+
+	if (intel_dp->psr_dpcd[1] & DP_PSR_NO_TRAIN_ON_EXIT || only_standby) {
+		val |= EDP_PSR_LINK_STANDBY;
+		val |= EDP_PSR_TP2_TP3_TIME_0us;
+		val |= EDP_PSR_TP1_TIME_0us;
+		val |= EDP_PSR_SKIP_AUX_EXIT;
+		val |= IS_BROADWELL(dev) ? BDW_PSR_SINGLE_FRAME : 0;
+	} else
+		val |= EDP_PSR_LINK_DISABLE;
+
+	I915_WRITE(EDP_PSR_CTL(dev), val |
+		   (IS_BROADWELL(dev) ? 0 : link_entry_time) |
+		   max_sleep_time << EDP_PSR_MAX_SLEEP_TIME_SHIFT |
+		   idle_frames << EDP_PSR_IDLE_FRAME_SHIFT |
+		   EDP_PSR_ENABLE);
+}
+
+static bool intel_psr_match_conditions(struct intel_dp *intel_dp)
+{
+	struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp);
+	struct drm_device *dev = dig_port->base.base.dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_crtc *crtc = dig_port->base.base.crtc;
+	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
+
+	lockdep_assert_held(&dev_priv->psr.lock);
+	WARN_ON(!drm_modeset_is_locked(&dev->mode_config.connection_mutex));
+	WARN_ON(!drm_modeset_is_locked(&crtc->mutex));
+
+	dev_priv->psr.source_ok = false;
+
+	if (IS_HASWELL(dev) && dig_port->port != PORT_A) {
+		DRM_DEBUG_KMS("HSW ties PSR to DDI A (eDP)\n");
+		return false;
+	}
+
+	if (!i915.enable_psr) {
+		DRM_DEBUG_KMS("PSR disable by flag\n");
+		return false;
+	}
+
+	/* Below limitations aren't valid for Broadwell */
+	if (IS_BROADWELL(dev))
+		goto out;
+
+	if (I915_READ(HSW_STEREO_3D_CTL(intel_crtc->config.cpu_transcoder)) &
+	    S3D_ENABLE) {
+		DRM_DEBUG_KMS("PSR condition failed: Stereo 3D is Enabled\n");
+		return false;
+	}
+
+	if (intel_crtc->config.adjusted_mode.flags & DRM_MODE_FLAG_INTERLACE) {
+		DRM_DEBUG_KMS("PSR condition failed: Interlaced is Enabled\n");
+		return false;
+	}
+
+ out:
+	dev_priv->psr.source_ok = true;
+	return true;
+}
+
+static void intel_psr_do_enable(struct intel_dp *intel_dp)
+{
+	struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp);
+	struct drm_device *dev = intel_dig_port->base.base.dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+
+	WARN_ON(I915_READ(EDP_PSR_CTL(dev)) & EDP_PSR_ENABLE);
+	WARN_ON(dev_priv->psr.active);
+	lockdep_assert_held(&dev_priv->psr.lock);
+
+	/* Enable/Re-enable PSR on the host */
+	intel_psr_enable_source(intel_dp);
+
+	dev_priv->psr.active = true;
+}
+
+/**
+ * intel_psr_enable - Enable PSR
+ * @intel_dp: Intel DP
+ *
+ * This function can only be called after the pipe is fully trained and enabled.
+ */
+void intel_psr_enable(struct intel_dp *intel_dp)
+{
+	struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp);
+	struct drm_device *dev = intel_dig_port->base.base.dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+
+	if (!HAS_PSR(dev)) {
+		DRM_DEBUG_KMS("PSR not supported on this platform\n");
+		return;
+	}
+
+	if (!is_edp_psr(intel_dp)) {
+		DRM_DEBUG_KMS("PSR not supported by this panel\n");
+		return;
+	}
+
+	mutex_lock(&dev_priv->psr.lock);
+	if (dev_priv->psr.enabled) {
+		DRM_DEBUG_KMS("PSR already in use\n");
+		goto unlock;
+	}
+
+	if (!intel_psr_match_conditions(intel_dp))
+		goto unlock;
+
+	dev_priv->psr.busy_frontbuffer_bits = 0;
+
+	intel_psr_setup_vsc(intel_dp);
+
+	/* Avoid continuous PSR exit by masking memup and hpd */
+	I915_WRITE(EDP_PSR_DEBUG_CTL(dev), EDP_PSR_DEBUG_MASK_MEMUP |
+		   EDP_PSR_DEBUG_MASK_HPD | EDP_PSR_DEBUG_MASK_LPSP);
+
+	/* Enable PSR on the panel */
+	intel_psr_enable_sink(intel_dp);
+
+	dev_priv->psr.enabled = intel_dp;
+unlock:
+	mutex_unlock(&dev_priv->psr.lock);
+}
+
+/**
+ * intel_psr_disable - Disable PSR
+ * @intel_dp: Intel DP
+ *
+ * This function needs to be called before disabling pipe.
+ */
+void intel_psr_disable(struct intel_dp *intel_dp)
+{
+	struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp);
+	struct drm_device *dev = intel_dig_port->base.base.dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+
+	mutex_lock(&dev_priv->psr.lock);
+	if (!dev_priv->psr.enabled) {
+		mutex_unlock(&dev_priv->psr.lock);
+		return;
+	}
+
+	if (dev_priv->psr.active) {
+		I915_WRITE(EDP_PSR_CTL(dev),
+			   I915_READ(EDP_PSR_CTL(dev)) & ~EDP_PSR_ENABLE);
+
+		/* Wait till PSR is idle */
+		if (_wait_for((I915_READ(EDP_PSR_STATUS_CTL(dev)) &
+			       EDP_PSR_STATUS_STATE_MASK) == 0, 2000, 10))
+			DRM_ERROR("Timed out waiting for PSR Idle State\n");
+
+		dev_priv->psr.active = false;
+	} else {
+		WARN_ON(I915_READ(EDP_PSR_CTL(dev)) & EDP_PSR_ENABLE);
+	}
+
+	dev_priv->psr.enabled = NULL;
+	mutex_unlock(&dev_priv->psr.lock);
+
+	cancel_delayed_work_sync(&dev_priv->psr.work);
+}
+
+static void intel_psr_work(struct work_struct *work)
+{
+	struct drm_i915_private *dev_priv =
+		container_of(work, typeof(*dev_priv), psr.work.work);
+	struct intel_dp *intel_dp = dev_priv->psr.enabled;
+
+	/* We have to make sure PSR is ready for re-enable
+	 * otherwise it keeps disabled until next full enable/disable cycle.
+	 * PSR might take some time to get fully disabled
+	 * and be ready for re-enable.
+	 */
+	if (wait_for((I915_READ(EDP_PSR_STATUS_CTL(dev_priv->dev)) &
+		      EDP_PSR_STATUS_STATE_MASK) == 0, 50)) {
+		DRM_ERROR("Timed out waiting for PSR Idle for re-enable\n");
+		return;
+	}
+
+	mutex_lock(&dev_priv->psr.lock);
+	intel_dp = dev_priv->psr.enabled;
+
+	if (!intel_dp)
+		goto unlock;
+
+	/*
+	 * The delayed work can race with an invalidate hence we need to
+	 * recheck. Since psr_flush first clears this and then reschedules we
+	 * won't ever miss a flush when bailing out here.
+	 */
+	if (dev_priv->psr.busy_frontbuffer_bits)
+		goto unlock;
+
+	intel_psr_do_enable(intel_dp);
+unlock:
+	mutex_unlock(&dev_priv->psr.lock);
+}
+
+static void intel_psr_exit(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+
+	if (dev_priv->psr.active) {
+		u32 val = I915_READ(EDP_PSR_CTL(dev));
+
+		WARN_ON(!(val & EDP_PSR_ENABLE));
+
+		I915_WRITE(EDP_PSR_CTL(dev), val & ~EDP_PSR_ENABLE);
+
+		dev_priv->psr.active = false;
+	}
+
+}
+
+/**
+ * intel_psr_invalidate - Invalidade PSR
+ * @dev: DRM device
+ * @frontbuffer_bits: frontbuffer plane tracking bits
+ *
+ * Since the hardware frontbuffer tracking has gaps we need to integrate
+ * with the software frontbuffer tracking. This function gets called every
+ * time frontbuffer rendering starts and a buffer gets dirtied. PSR must be
+ * disabled if the frontbuffer mask contains a buffer relevant to PSR.
+ *
+ * Dirty frontbuffers relevant to PSR are tracked in busy_frontbuffer_bits."
+ */
+void intel_psr_invalidate(struct drm_device *dev,
+			      unsigned frontbuffer_bits)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_crtc *crtc;
+	enum pipe pipe;
+
+	mutex_lock(&dev_priv->psr.lock);
+	if (!dev_priv->psr.enabled) {
+		mutex_unlock(&dev_priv->psr.lock);
+		return;
+	}
+
+	crtc = dp_to_dig_port(dev_priv->psr.enabled)->base.base.crtc;
+	pipe = to_intel_crtc(crtc)->pipe;
+
+	intel_psr_exit(dev);
+
+	frontbuffer_bits &= INTEL_FRONTBUFFER_ALL_MASK(pipe);
+
+	dev_priv->psr.busy_frontbuffer_bits |= frontbuffer_bits;
+	mutex_unlock(&dev_priv->psr.lock);
+}
+
+/**
+ * intel_psr_flush - Flush PSR
+ * @dev: DRM device
+ * @frontbuffer_bits: frontbuffer plane tracking bits
+ *
+ * Since the hardware frontbuffer tracking has gaps we need to integrate
+ * with the software frontbuffer tracking. This function gets called every
+ * time frontbuffer rendering has completed and flushed out to memory. PSR
+ * can be enabled again if no other frontbuffer relevant to PSR is dirty.
+ *
+ * Dirty frontbuffers relevant to PSR are tracked in busy_frontbuffer_bits.
+ */
+void intel_psr_flush(struct drm_device *dev,
+			 unsigned frontbuffer_bits)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_crtc *crtc;
+	enum pipe pipe;
+
+	mutex_lock(&dev_priv->psr.lock);
+	if (!dev_priv->psr.enabled) {
+		mutex_unlock(&dev_priv->psr.lock);
+		return;
+	}
+
+	crtc = dp_to_dig_port(dev_priv->psr.enabled)->base.base.crtc;
+	pipe = to_intel_crtc(crtc)->pipe;
+	dev_priv->psr.busy_frontbuffer_bits &= ~frontbuffer_bits;
+
+	/*
+	 * On Haswell sprite plane updates don't result in a psr invalidating
+	 * signal in the hardware. Which means we need to manually fake this in
+	 * software for all flushes, not just when we've seen a preceding
+	 * invalidation through frontbuffer rendering.
+	 */
+	if (IS_HASWELL(dev) &&
+	    (frontbuffer_bits & INTEL_FRONTBUFFER_SPRITE(pipe)))
+		intel_psr_exit(dev);
+
+	if (!dev_priv->psr.active && !dev_priv->psr.busy_frontbuffer_bits)
+		schedule_delayed_work(&dev_priv->psr.work,
+				      msecs_to_jiffies(100));
+	mutex_unlock(&dev_priv->psr.lock);
+}
+
+/**
+ * intel_psr_init - Init basic PSR work and mutex.
+ * @dev: DRM device
+ *
+ * This function is  called only once at driver load to initialize basic
+ * PSR stuff.
+ */
+void intel_psr_init(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+
+	INIT_DELAYED_WORK(&dev_priv->psr.work, intel_psr_work);
+	mutex_init(&dev_priv->psr.lock);
+}
diff --git a/drivers/gpu/drm/i915/intel_renderstate.h b/drivers/gpu/drm/i915/intel_renderstate.h
index fd4f66231d30..5bd69852752c 100644
--- a/drivers/gpu/drm/i915/intel_renderstate.h
+++ b/drivers/gpu/drm/i915/intel_renderstate.h
@@ -24,17 +24,12 @@
 #ifndef _INTEL_RENDERSTATE_H
 #define _INTEL_RENDERSTATE_H
 
-#include <linux/types.h>
-
-struct intel_renderstate_rodata {
-	const u32 *reloc;
-	const u32 *batch;
-	const u32 batch_items;
-};
+#include "i915_drv.h"
 
 extern const struct intel_renderstate_rodata gen6_null_state;
 extern const struct intel_renderstate_rodata gen7_null_state;
 extern const struct intel_renderstate_rodata gen8_null_state;
+extern const struct intel_renderstate_rodata gen9_null_state;
 
 #define RO_RENDERSTATE(_g)						\
 	const struct intel_renderstate_rodata gen ## _g ## _null_state = { \
diff --git a/drivers/gpu/drm/i915/intel_renderstate_gen8.c b/drivers/gpu/drm/i915/intel_renderstate_gen8.c
index 75ef1b5de45c..78011d73fa9f 100644
--- a/drivers/gpu/drm/i915/intel_renderstate_gen8.c
+++ b/drivers/gpu/drm/i915/intel_renderstate_gen8.c
@@ -1,16 +1,134 @@
 #include "intel_renderstate.h"
 
 static const u32 gen8_null_state_relocs[] = {
-	0x00000048,
-	0x00000050,
-	0x00000060,
-	0x000003ec,
+	0x00000798,
+	0x000007a4,
+	0x000007ac,
+	0x000007bc,
 	-1,
 };
 
 static const u32 gen8_null_state_batch[] = {
+	0x7a000004,
+	0x01000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
 	0x69040000,
-	0x61020001,
+	0x78140000,
+	0x04000000,
+	0x7820000a,
+	0x00000000,
+	0x00000000,
+	0x80000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x78130002,
+	0x00000000,
+	0x00000000,
+	0x02001808,
+	0x781f0002,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x78510009,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x78100007,
+	0x00000000,
+	0x00000000,
+	0x00010000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x781b0007,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000800,
+	0x00000000,
+	0x78110008,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x781e0003,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x781d0007,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x78120002,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x78500003,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x781c0002,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x780c0000,
+	0x00000000,
+	0x78520003,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x78300000,
+	0x08010040,
+	0x78310000,
+	0x1e000000,
+	0x78320000,
+	0x1e000000,
+	0x78330000,
+	0x1e000000,
+	0x79190002,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x791a0002,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x791b0002,
+	0x00000000,
 	0x00000000,
 	0x00000000,
 	0x79120000,
@@ -23,48 +141,435 @@ static const u32 gen8_null_state_batch[] = {
 	0x00000000,
 	0x79160000,
 	0x00000000,
-	0x6101000e,
-	0x00000001,
+	0x78150009,
 	0x00000000,
-	0x00000001,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x78190009,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x781a0009,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x78160009,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x78170009,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x78490001,
+	0x00000000,
+	0x00000000,
+	0x784a0000,
+	0x00000000,
+	0x784b0000,
+	0x00000004,
+	0x79170101,
+	0x00000000,
+	0x00000080,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x79180006,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x79180006,
+	0x20000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x79180006,
+	0x40000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x79180006,
+	0x60000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x6101000e,
 	0x00000001,	 /* reloc */
 	0x00000000,
+	0x00000000,
 	0x00000001,	 /* reloc */
 	0x00000000,
+	0x00000001,	 /* reloc */
 	0x00000000,
+	0x00000001,
 	0x00000000,
 	0x00000001,	 /* reloc */
 	0x00000000,
-	0xfffff001,
 	0x00001001,
-	0xfffff001,
 	0x00001001,
-	0x78230000,
-	0x000006e0,
-	0x78210000,
-	0x00000700,
-	0x78300000,
-	0x08010040,
-	0x78330000,
-	0x08000000,
-	0x78310000,
-	0x08000000,
-	0x78320000,
-	0x08000000,
-	0x78240000,
-	0x00000641,
-	0x780e0000,
-	0x00000601,
+	0x00000001,
+	0x00001001,
+	0x61020001,
+	0x00000000,
+	0x00000000,
+	0x79000002,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x78050006,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x79040002,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x79040002,
+	0x40000000,
+	0x00000000,
+	0x00000000,
+	0x79040002,
+	0x80000000,
+	0x00000000,
+	0x00000000,
+	0x79040002,
+	0xc0000000,
+	0x00000000,
+	0x00000000,
+	0x79080001,
+	0x00000000,
+	0x00000000,
+	0x790a0001,
+	0x00000000,
+	0x00000000,
+	0x78060003,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x78070003,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x78040001,
+	0x00000000,
+	0x00000000,
+	0x79110000,
+	0x00000000,
 	0x780d0000,
 	0x00000000,
-	0x78180000,
-	0x00000001,
-	0x78520003,
+	0x79060000,
 	0x00000000,
+	0x7907001f,
 	0x00000000,
 	0x00000000,
 	0x00000000,
-	0x78190009,
 	0x00000000,
 	0x00000000,
 	0x00000000,
@@ -75,7 +580,6 @@ static const u32 gen8_null_state_batch[] = {
 	0x00000000,
 	0x00000000,
 	0x00000000,
-	0x781b0007,
 	0x00000000,
 	0x00000000,
 	0x00000000,
@@ -84,26 +588,22 @@ static const u32 gen8_null_state_batch[] = {
 	0x00000000,
 	0x00000000,
 	0x00000000,
-	0x78270000,
 	0x00000000,
-	0x782c0000,
 	0x00000000,
-	0x781c0002,
 	0x00000000,
 	0x00000000,
 	0x00000000,
-	0x78160009,
 	0x00000000,
 	0x00000000,
 	0x00000000,
 	0x00000000,
 	0x00000000,
 	0x00000000,
+	0x7902000f,
 	0x00000000,
 	0x00000000,
 	0x00000000,
 	0x00000000,
-	0x78110008,
 	0x00000000,
 	0x00000000,
 	0x00000000,
@@ -113,12 +613,10 @@ static const u32 gen8_null_state_batch[] = {
 	0x00000000,
 	0x00000000,
 	0x00000000,
-	0x78290000,
 	0x00000000,
-	0x782e0000,
 	0x00000000,
-	0x781a0009,
 	0x00000000,
+	0x790c000f,
 	0x00000000,
 	0x00000000,
 	0x00000000,
@@ -128,7 +626,6 @@ static const u32 gen8_null_state_batch[] = {
 	0x00000000,
 	0x00000000,
 	0x00000000,
-	0x781d0007,
 	0x00000000,
 	0x00000000,
 	0x00000000,
@@ -136,153 +633,153 @@ static const u32 gen8_null_state_batch[] = {
 	0x00000000,
 	0x00000000,
 	0x00000000,
+	0x780a0003,
 	0x00000000,
-	0x78280000,
 	0x00000000,
-	0x782d0000,
 	0x00000000,
-	0x78260000,
 	0x00000000,
-	0x782b0000,
+	0x78080083,
+	0x00004000,
 	0x00000000,
-	0x78150009,
 	0x00000000,
 	0x00000000,
+	0x04004000,
 	0x00000000,
 	0x00000000,
 	0x00000000,
+	0x08004000,
 	0x00000000,
 	0x00000000,
 	0x00000000,
+	0x0c004000,
 	0x00000000,
 	0x00000000,
-	0x78100007,
 	0x00000000,
+	0x10004000,
 	0x00000000,
 	0x00000000,
 	0x00000000,
+	0x14004000,
 	0x00000000,
 	0x00000000,
 	0x00000000,
+	0x18004000,
 	0x00000000,
-	0x781e0003,
 	0x00000000,
 	0x00000000,
+	0x1c004000,
 	0x00000000,
 	0x00000000,
-	0x78120002,
 	0x00000000,
+	0x20004000,
 	0x00000000,
 	0x00000000,
-	0x781f0002,
-	0x30400820,
 	0x00000000,
+	0x24004000,
 	0x00000000,
-	0x78510009,
 	0x00000000,
 	0x00000000,
+	0x28004000,
 	0x00000000,
 	0x00000000,
 	0x00000000,
+	0x2c004000,
 	0x00000000,
 	0x00000000,
 	0x00000000,
+	0x30004000,
 	0x00000000,
 	0x00000000,
-	0x78500003,
-	0x00210000,
 	0x00000000,
+	0x34004000,
 	0x00000000,
 	0x00000000,
-	0x78130002,
 	0x00000000,
+	0x38004000,
 	0x00000000,
 	0x00000000,
-	0x782a0000,
-	0x00000480,
-	0x782f0000,
-	0x00000540,
-	0x78140000,
-	0x00000800,
-	0x78170009,
 	0x00000000,
+	0x3c004000,
 	0x00000000,
 	0x00000000,
 	0x00000000,
+	0x40004000,
 	0x00000000,
 	0x00000000,
 	0x00000000,
+	0x44004000,
 	0x00000000,
 	0x00000000,
 	0x00000000,
-	0x7820000a,
-	0x00000580,
+	0x48004000,
 	0x00000000,
-	0x08080000,
 	0x00000000,
 	0x00000000,
-	0x1f000002,
-	0x00060000,
+	0x4c004000,
 	0x00000000,
 	0x00000000,
 	0x00000000,
+	0x50004000,
 	0x00000000,
-	0x784d0000,
-	0x40000000,
-	0x784f0000,
-	0x80000100,
-	0x780f0000,
-	0x00000740,
-	0x78050006,
 	0x00000000,
 	0x00000000,
+	0x54004000,
 	0x00000000,
 	0x00000000,
 	0x00000000,
+	0x58004000,
 	0x00000000,
 	0x00000000,
-	0x78070003,
 	0x00000000,
+	0x5c004000,
 	0x00000000,
 	0x00000000,
 	0x00000000,
-	0x78060003,
+	0x60004000,
 	0x00000000,
 	0x00000000,
 	0x00000000,
+	0x64004000,
 	0x00000000,
-	0x78040001,
 	0x00000000,
-	0x00000001,
-	0x79000002,
-	0xffffffff,
+	0x00000000,
+	0x68004000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x6c004000,
+	0x00000000,
 	0x00000000,
 	0x00000000,
-	0x78080003,
-	0x00006000,
-	0x000005e0,	 /* reloc */
+	0x70004000,
 	0x00000000,
 	0x00000000,
-	0x78090005,
+	0x00000000,
+	0x74004000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x78004000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x7c004000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x80004000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x78090043,
 	0x02000000,
 	0x22220000,
-	0x02f60000,
-	0x11230000,
-	0x02850004,
-	0x11230000,
-	0x784b0000,
-	0x0000000f,
-	0x78490001,
 	0x00000000,
 	0x00000000,
-	0x7b000005,
 	0x00000000,
-	0x00000003,
 	0x00000000,
-	0x00000001,
 	0x00000000,
 	0x00000000,
-	0x05000000,	 /* cmds end */
 	0x00000000,
 	0x00000000,
 	0x00000000,
@@ -297,8 +794,6 @@ static const u32 gen8_null_state_batch[] = {
 	0x00000000,
 	0x00000000,
 	0x00000000,
-	0x000004c0,	 /* state start */
-	0x00000500,
 	0x00000000,
 	0x00000000,
 	0x00000000,
@@ -345,46 +840,65 @@ static const u32 gen8_null_state_batch[] = {
 	0x00000000,
 	0x00000000,
 	0x00000000,
+	0x680b0001,
+	0x78260000,
+	0x00000000,
+	0x78270000,
+	0x00000000,
+	0x78280000,
+	0x00000000,
+	0x78290000,
+	0x00000000,
+	0x782a0000,
+	0x00000000,
+	0x780e0000,
+	0x00000dc1,
+	0x78240000,
+	0x00000e01,
+	0x784f0000,
+	0x80000100,
+	0x784d0000,
+	0x40000000,
+	0x782b0000,
+	0x00000000,
+	0x782c0000,
+	0x00000000,
+	0x782d0000,
 	0x00000000,
+	0x782e0000,
 	0x00000000,
+	0x782f0000,
 	0x00000000,
-	0x00000092,
+	0x780f0000,
 	0x00000000,
+	0x78230000,
+	0x00000e60,
+	0x78210000,
+	0x00000e80,
+	0x7b000005,
+	0x00000004,
+	0x00000001,
 	0x00000000,
+	0x00000001,
 	0x00000000,
 	0x00000000,
+	0x05000000,	 /* cmds end */
 	0x00000000,
 	0x00000000,
 	0x00000000,
 	0x00000000,
 	0x00000000,
 	0x00000000,
+	0x00000000,	 /* state start */
+	0x00000000,
+	0x3f800000,
+	0x3f800000,
+	0x3f800000,
+	0x3f800000,
+	0x00000000,
+	0x00000000,
 	0x00000000,
 	0x00000000,
-	0x0060005a,
-	0x21403ae8,
-	0x3a0000c0,
-	0x008d0040,
-	0x0060005a,
-	0x21603ae8,
-	0x3a0000c0,
-	0x008d0080,
-	0x0060005a,
-	0x21803ae8,
-	0x3a0000d0,
-	0x008d0040,
-	0x0060005a,
-	0x21a03ae8,
-	0x3a0000d0,
-	0x008d0080,
-	0x02800031,
-	0x2e0022e8,
-	0x0e000140,
-	0x08840001,
-	0x05800031,
-	0x200022e0,
-	0x0e000e00,
-	0x90031000,
 	0x00000000,
 	0x00000000,
 	0x00000000,
@@ -410,38 +924,6 @@ static const u32 gen8_null_state_batch[] = {
 	0x00000000,
 	0x00000000,
 	0x00000000,
-	0x06200000,
-	0x00000002,
-	0x06200000,
-	0x00000002,
-	0x06200000,
-	0x00000002,
-	0x06200000,
-	0x00000002,
-	0x06200000,
-	0x00000002,
-	0x06200000,
-	0x00000002,
-	0x06200000,
-	0x00000002,
-	0x06200000,
-	0x00000002,
-	0x06200000,
-	0x00000002,
-	0x06200000,
-	0x00000002,
-	0x06200000,
-	0x00000002,
-	0x06200000,
-	0x00000002,
-	0x06200000,
-	0x00000002,
-	0x06200000,
-	0x00000002,
-	0x06200000,
-	0x00000002,
-	0x06200000,
-	0x00000002,
 	0x00000000,
 	0x00000000,
 	0x00000000,
@@ -449,8 +931,6 @@ static const u32 gen8_null_state_batch[] = {
 	0x00000000,
 	0x00000000,
 	0x00000000,
-	0xf99a130c,
-	0x799a130c,
 	0x00000000,
 	0x00000000,
 	0x00000000,
@@ -466,9 +946,7 @@ static const u32 gen8_null_state_batch[] = {
 	0x00000000,
 	0x00000000,
 	0x00000000,
-	0x3f800000,
 	0x00000000,
-	0x3f800000,
 	0x00000000,
 	0x00000000,
 	0x00000000,
diff --git a/drivers/gpu/drm/i915/intel_renderstate_gen9.c b/drivers/gpu/drm/i915/intel_renderstate_gen9.c
new file mode 100644
index 000000000000..875075373807
--- /dev/null
+++ b/drivers/gpu/drm/i915/intel_renderstate_gen9.c
@@ -0,0 +1,974 @@
+#include "intel_renderstate.h"
+
+static const u32 gen9_null_state_relocs[] = {
+	0x000007a8,
+	0x000007b4,
+	0x000007bc,
+	0x000007cc,
+	-1,
+};
+
+static const u32 gen9_null_state_batch[] = {
+	0x7a000004,
+	0x01000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x69040300,
+	0x78140000,
+	0x04000000,
+	0x7820000a,
+	0x00000000,
+	0x00000000,
+	0x80000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x78130002,
+	0x00000000,
+	0x00000000,
+	0x02001808,
+	0x781f0004,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x78510009,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x78100007,
+	0x00000000,
+	0x00000000,
+	0x00010000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x781b0007,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000800,
+	0x00000000,
+	0x78110008,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x781e0003,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x781d0009,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x78120002,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x78500003,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x781c0002,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x780c0000,
+	0x00000000,
+	0x78520003,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x78300000,
+	0x08010040,
+	0x78310000,
+	0x1e000000,
+	0x78320000,
+	0x1e000000,
+	0x78330000,
+	0x1e000000,
+	0x79190002,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x791a0002,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x791b0002,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x79120000,
+	0x00000000,
+	0x79130000,
+	0x00000000,
+	0x79140000,
+	0x00000000,
+	0x79150000,
+	0x00000000,
+	0x79160000,
+	0x00000000,
+	0x78150009,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x78190009,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x781a0009,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x78160009,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x78170009,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x78490001,
+	0x00000000,
+	0x00000000,
+	0x784a0000,
+	0x00000000,
+	0x784b0000,
+	0x00000004,
+	0x79170101,
+	0x00000000,
+	0x00000080,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x79180006,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x79180006,
+	0x20000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x79180006,
+	0x40000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x79180006,
+	0x60000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x61010011,
+	0x00000001,	 /* reloc */
+	0x00000000,
+	0x00000000,
+	0x00000001,	 /* reloc */
+	0x00000000,
+	0x00000001,	 /* reloc */
+	0x00000000,
+	0x00000001,
+	0x00000000,
+	0x00000001,	 /* reloc */
+	0x00000000,
+	0x00001001,
+	0x00001001,
+	0x00000001,
+	0x00001001,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x61020001,
+	0x00000000,
+	0x00000000,
+	0x79000002,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x78050006,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x79040002,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x79040002,
+	0x40000000,
+	0x00000000,
+	0x00000000,
+	0x79040002,
+	0x80000000,
+	0x00000000,
+	0x00000000,
+	0x79040002,
+	0xc0000000,
+	0x00000000,
+	0x00000000,
+	0x79080001,
+	0x00000000,
+	0x00000000,
+	0x790a0001,
+	0x00000000,
+	0x00000000,
+	0x78060003,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x78070003,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x78040001,
+	0x00000000,
+	0x00000000,
+	0x79110000,
+	0x00000000,
+	0x780d0000,
+	0x00000000,
+	0x79060000,
+	0x00000000,
+	0x7907001f,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x7902000f,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x790c000f,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x780a0003,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x78080083,
+	0x00004000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x04004000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x08004000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x0c004000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x10004000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x14004000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x18004000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x1c004000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x20004000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x24004000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x28004000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x2c004000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x30004000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x34004000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x38004000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x3c004000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x40004000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x44004000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x48004000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x4c004000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x50004000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x54004000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x58004000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x5c004000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x60004000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x64004000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x68004000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x6c004000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x70004000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x74004000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x78004000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x7c004000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x80004000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x78090043,
+	0x02000000,
+	0x22220000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x78550003,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x680b0001,
+	0x780e0000,
+	0x00000e01,
+	0x78240000,
+	0x00000e41,
+	0x784f0000,
+	0x80000100,
+	0x784d0000,
+	0x40000000,
+	0x782b0000,
+	0x00000000,
+	0x782c0000,
+	0x00000000,
+	0x782d0000,
+	0x00000000,
+	0x782e0000,
+	0x00000000,
+	0x782f0000,
+	0x00000000,
+	0x780f0000,
+	0x00000000,
+	0x78230000,
+	0x00000ea0,
+	0x78210000,
+	0x00000ec0,
+	0x78260000,
+	0x00000000,
+	0x78270000,
+	0x00000000,
+	0x78280000,
+	0x00000000,
+	0x78290000,
+	0x00000000,
+	0x782a0000,
+	0x00000000,
+	0x7b000005,
+	0x00000004,
+	0x00000001,
+	0x00000000,
+	0x00000001,
+	0x00000000,
+	0x00000000,
+	0x05000000,	 /* cmds end */
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,	 /* state start */
+	0x00000000,
+	0x3f800000,
+	0x3f800000,
+	0x3f800000,
+	0x3f800000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,	 /* state end */
+};
+
+RO_RENDERSTATE(9);
diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.c b/drivers/gpu/drm/i915/intel_ringbuffer.c
index 47a126a0493f..c7bc93d28d84 100644
--- a/drivers/gpu/drm/i915/intel_ringbuffer.c
+++ b/drivers/gpu/drm/i915/intel_ringbuffer.c
@@ -33,14 +33,24 @@
 #include "i915_trace.h"
 #include "intel_drv.h"
 
-/* Early gen2 devices have a cacheline of just 32 bytes, using 64 is overkill,
- * but keeps the logic simple. Indeed, the whole purpose of this macro is just
- * to give some inclination as to some of the magic values used in the various
- * workarounds!
- */
-#define CACHELINE_BYTES 64
+bool
+intel_ring_initialized(struct intel_engine_cs *ring)
+{
+	struct drm_device *dev = ring->dev;
+
+	if (!dev)
+		return false;
 
-static inline int __ring_space(int head, int tail, int size)
+	if (i915.enable_execlists) {
+		struct intel_context *dctx = ring->default_context;
+		struct intel_ringbuffer *ringbuf = dctx->engine[ring->id].ringbuf;
+
+		return ringbuf->obj;
+	} else
+		return ring->buffer && ring->buffer->obj;
+}
+
+int __intel_ring_space(int head, int tail, int size)
 {
 	int space = head - (tail + I915_RING_FREE_SPACE);
 	if (space < 0)
@@ -48,12 +58,13 @@ static inline int __ring_space(int head, int tail, int size)
 	return space;
 }
 
-static inline int ring_space(struct intel_ringbuffer *ringbuf)
+int intel_ring_space(struct intel_ringbuffer *ringbuf)
 {
-	return __ring_space(ringbuf->head & HEAD_ADDR, ringbuf->tail, ringbuf->size);
+	return __intel_ring_space(ringbuf->head & HEAD_ADDR,
+				  ringbuf->tail, ringbuf->size);
 }
 
-static bool intel_ring_stopped(struct intel_engine_cs *ring)
+bool intel_ring_stopped(struct intel_engine_cs *ring)
 {
 	struct drm_i915_private *dev_priv = ring->dev->dev_private;
 	return dev_priv->gpu_error.stop_rings & intel_ring_flag(ring);
@@ -351,12 +362,15 @@ gen7_render_ring_flush(struct intel_engine_cs *ring,
 		flags |= PIPE_CONTROL_VF_CACHE_INVALIDATE;
 		flags |= PIPE_CONTROL_CONST_CACHE_INVALIDATE;
 		flags |= PIPE_CONTROL_STATE_CACHE_INVALIDATE;
+		flags |= PIPE_CONTROL_MEDIA_STATE_CLEAR;
 		/*
 		 * TLB invalidate requires a post-sync write.
 		 */
 		flags |= PIPE_CONTROL_QW_WRITE;
 		flags |= PIPE_CONTROL_GLOBAL_GTT_IVB;
 
+		flags |= PIPE_CONTROL_STALL_AT_SCOREBOARD;
+
 		/* Workaround: we must issue a pipe_control with CS-stall bit
 		 * set before a pipe_control command that has the state cache
 		 * invalidate bit set. */
@@ -433,7 +447,14 @@ gen8_render_ring_flush(struct intel_engine_cs *ring,
 			return ret;
 	}
 
-	return gen8_emit_pipe_control(ring, flags, scratch_addr);
+	ret = gen8_emit_pipe_control(ring, flags, scratch_addr);
+	if (ret)
+		return ret;
+
+	if (!invalidate_domains && flush_domains)
+		return gen7_ring_fbc_flush(ring, FBC_REND_NUKE);
+
+	return 0;
 }
 
 static void ring_write_tail(struct intel_engine_cs *ring,
@@ -476,9 +497,14 @@ static bool stop_ring(struct intel_engine_cs *ring)
 
 	if (!IS_GEN2(ring->dev)) {
 		I915_WRITE_MODE(ring, _MASKED_BIT_ENABLE(STOP_RING));
-		if (wait_for_atomic((I915_READ_MODE(ring) & MODE_IDLE) != 0, 1000)) {
-			DRM_ERROR("%s :timed out trying to stop ring\n", ring->name);
-			return false;
+		if (wait_for((I915_READ_MODE(ring) & MODE_IDLE) != 0, 1000)) {
+			DRM_ERROR("%s : timed out trying to stop ring\n", ring->name);
+			/* Sometimes we observe that the idle flag is not
+			 * set even though the ring is empty. So double
+			 * check before giving up.
+			 */
+			if (I915_READ_HEAD(ring) != I915_READ_TAIL(ring))
+				return false;
 		}
 	}
 
@@ -540,6 +566,14 @@ static int init_ring_common(struct intel_engine_cs *ring)
 	 * also enforces ordering), otherwise the hw might lose the new ring
 	 * register values. */
 	I915_WRITE_START(ring, i915_gem_obj_ggtt_offset(obj));
+
+	/* WaClearRingBufHeadRegAtInit:ctg,elk */
+	if (I915_READ_HEAD(ring))
+		DRM_DEBUG("%s initialization failed [head=%08x], fudging\n",
+			  ring->name, I915_READ_HEAD(ring));
+	I915_WRITE_HEAD(ring, 0);
+	(void)I915_READ_HEAD(ring);
+
 	I915_WRITE_CTL(ring,
 			((ringbuf->size - PAGE_SIZE) & RING_NR_PAGES)
 			| RING_VALID);
@@ -558,14 +592,10 @@ static int init_ring_common(struct intel_engine_cs *ring)
 		goto out;
 	}
 
-	if (!drm_core_check_feature(ring->dev, DRIVER_MODESET))
-		i915_kernel_lost_context(ring->dev);
-	else {
-		ringbuf->head = I915_READ_HEAD(ring);
-		ringbuf->tail = I915_READ_TAIL(ring) & TAIL_ADDR;
-		ringbuf->space = ring_space(ringbuf);
-		ringbuf->last_retired_head = -1;
-	}
+	ringbuf->head = I915_READ_HEAD(ring);
+	ringbuf->tail = I915_READ_TAIL(ring) & TAIL_ADDR;
+	ringbuf->space = intel_ring_space(ringbuf);
+	ringbuf->last_retired_head = -1;
 
 	memset(&ring->hangcheck, 0, sizeof(ring->hangcheck));
 
@@ -575,8 +605,25 @@ out:
 	return ret;
 }
 
-static int
-init_pipe_control(struct intel_engine_cs *ring)
+void
+intel_fini_pipe_control(struct intel_engine_cs *ring)
+{
+	struct drm_device *dev = ring->dev;
+
+	if (ring->scratch.obj == NULL)
+		return;
+
+	if (INTEL_INFO(dev)->gen >= 5) {
+		kunmap(sg_page(ring->scratch.obj->pages->sgl));
+		i915_gem_object_ggtt_unpin(ring->scratch.obj);
+	}
+
+	drm_gem_object_unreference(&ring->scratch.obj->base);
+	ring->scratch.obj = NULL;
+}
+
+int
+intel_init_pipe_control(struct intel_engine_cs *ring)
 {
 	int ret;
 
@@ -617,6 +664,170 @@ err:
 	return ret;
 }
 
+static int intel_ring_workarounds_emit(struct intel_engine_cs *ring,
+				       struct intel_context *ctx)
+{
+	int ret, i;
+	struct drm_device *dev = ring->dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct i915_workarounds *w = &dev_priv->workarounds;
+
+	if (WARN_ON(w->count == 0))
+		return 0;
+
+	ring->gpu_caches_dirty = true;
+	ret = intel_ring_flush_all_caches(ring);
+	if (ret)
+		return ret;
+
+	ret = intel_ring_begin(ring, (w->count * 2 + 2));
+	if (ret)
+		return ret;
+
+	intel_ring_emit(ring, MI_LOAD_REGISTER_IMM(w->count));
+	for (i = 0; i < w->count; i++) {
+		intel_ring_emit(ring, w->reg[i].addr);
+		intel_ring_emit(ring, w->reg[i].value);
+	}
+	intel_ring_emit(ring, MI_NOOP);
+
+	intel_ring_advance(ring);
+
+	ring->gpu_caches_dirty = true;
+	ret = intel_ring_flush_all_caches(ring);
+	if (ret)
+		return ret;
+
+	DRM_DEBUG_DRIVER("Number of Workarounds emitted: %d\n", w->count);
+
+	return 0;
+}
+
+static int wa_add(struct drm_i915_private *dev_priv,
+		  const u32 addr, const u32 mask, const u32 val)
+{
+	const u32 idx = dev_priv->workarounds.count;
+
+	if (WARN_ON(idx >= I915_MAX_WA_REGS))
+		return -ENOSPC;
+
+	dev_priv->workarounds.reg[idx].addr = addr;
+	dev_priv->workarounds.reg[idx].value = val;
+	dev_priv->workarounds.reg[idx].mask = mask;
+
+	dev_priv->workarounds.count++;
+
+	return 0;
+}
+
+#define WA_REG(addr, mask, val) { \
+		const int r = wa_add(dev_priv, (addr), (mask), (val)); \
+		if (r) \
+			return r; \
+	}
+
+#define WA_SET_BIT_MASKED(addr, mask) \
+	WA_REG(addr, (mask), _MASKED_BIT_ENABLE(mask))
+
+#define WA_CLR_BIT_MASKED(addr, mask) \
+	WA_REG(addr, (mask), _MASKED_BIT_DISABLE(mask))
+
+#define WA_SET_FIELD_MASKED(addr, mask, value) \
+	WA_REG(addr, mask, _MASKED_FIELD(mask, value))
+
+#define WA_SET_BIT(addr, mask) WA_REG(addr, mask, I915_READ(addr) | (mask))
+#define WA_CLR_BIT(addr, mask) WA_REG(addr, mask, I915_READ(addr) & ~(mask))
+
+#define WA_WRITE(addr, val) WA_REG(addr, 0xffffffff, val)
+
+static int bdw_init_workarounds(struct intel_engine_cs *ring)
+{
+	struct drm_device *dev = ring->dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+
+	/* WaDisablePartialInstShootdown:bdw */
+	/* WaDisableThreadStallDopClockGating:bdw (pre-production) */
+	WA_SET_BIT_MASKED(GEN8_ROW_CHICKEN,
+			  PARTIAL_INSTRUCTION_SHOOTDOWN_DISABLE |
+			  STALL_DOP_GATING_DISABLE);
+
+	/* WaDisableDopClockGating:bdw */
+	WA_SET_BIT_MASKED(GEN7_ROW_CHICKEN2,
+			  DOP_CLOCK_GATING_DISABLE);
+
+	WA_SET_BIT_MASKED(HALF_SLICE_CHICKEN3,
+			  GEN8_SAMPLER_POWER_BYPASS_DIS);
+
+	/* Use Force Non-Coherent whenever executing a 3D context. This is a
+	 * workaround for for a possible hang in the unlikely event a TLB
+	 * invalidation occurs during a PSD flush.
+	 */
+	/* WaDisableFenceDestinationToSLM:bdw (GT3 pre-production) */
+	WA_SET_BIT_MASKED(HDC_CHICKEN0,
+			  HDC_FORCE_NON_COHERENT |
+			  (IS_BDW_GT3(dev) ? HDC_FENCE_DEST_SLM_DISABLE : 0));
+
+	/* Wa4x4STCOptimizationDisable:bdw */
+	WA_SET_BIT_MASKED(CACHE_MODE_1,
+			  GEN8_4x4_STC_OPTIMIZATION_DISABLE);
+
+	/*
+	 * BSpec recommends 8x4 when MSAA is used,
+	 * however in practice 16x4 seems fastest.
+	 *
+	 * Note that PS/WM thread counts depend on the WIZ hashing
+	 * disable bit, which we don't touch here, but it's good
+	 * to keep in mind (see 3DSTATE_PS and 3DSTATE_WM).
+	 */
+	WA_SET_FIELD_MASKED(GEN7_GT_MODE,
+			    GEN6_WIZ_HASHING_MASK,
+			    GEN6_WIZ_HASHING_16x4);
+
+	return 0;
+}
+
+static int chv_init_workarounds(struct intel_engine_cs *ring)
+{
+	struct drm_device *dev = ring->dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+
+	/* WaDisablePartialInstShootdown:chv */
+	/* WaDisableThreadStallDopClockGating:chv */
+	WA_SET_BIT_MASKED(GEN8_ROW_CHICKEN,
+			  PARTIAL_INSTRUCTION_SHOOTDOWN_DISABLE |
+			  STALL_DOP_GATING_DISABLE);
+
+	/* Use Force Non-Coherent whenever executing a 3D context. This is a
+	 * workaround for a possible hang in the unlikely event a TLB
+	 * invalidation occurs during a PSD flush.
+	 */
+	/* WaForceEnableNonCoherent:chv */
+	/* WaHdcDisableFetchWhenMasked:chv */
+	WA_SET_BIT_MASKED(HDC_CHICKEN0,
+			  HDC_FORCE_NON_COHERENT |
+			  HDC_DONOT_FETCH_MEM_WHEN_MASKED);
+
+	return 0;
+}
+
+int init_workarounds_ring(struct intel_engine_cs *ring)
+{
+	struct drm_device *dev = ring->dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+
+	WARN_ON(ring->id != RCS);
+
+	dev_priv->workarounds.count = 0;
+
+	if (IS_BROADWELL(dev))
+		return bdw_init_workarounds(ring);
+
+	if (IS_CHERRYVIEW(dev))
+		return chv_init_workarounds(ring);
+
+	return 0;
+}
+
 static int init_render_ring(struct intel_engine_cs *ring)
 {
 	struct drm_device *dev = ring->dev;
@@ -635,7 +846,7 @@ static int init_render_ring(struct intel_engine_cs *ring)
 	 *
 	 * WaDisableAsyncFlipPerfMode:snb,ivb,hsw,vlv,bdw,chv
 	 */
-	if (INTEL_INFO(dev)->gen >= 6)
+	if (INTEL_INFO(dev)->gen >= 6 && INTEL_INFO(dev)->gen < 9)
 		I915_WRITE(MI_MODE, _MASKED_BIT_ENABLE(ASYNC_FLIP_PERF_DISABLE));
 
 	/* Required for the hardware to program scanline values for waiting */
@@ -651,7 +862,7 @@ static int init_render_ring(struct intel_engine_cs *ring)
 			   _MASKED_BIT_ENABLE(GFX_REPLAY_MODE));
 
 	if (INTEL_INFO(dev)->gen >= 5) {
-		ret = init_pipe_control(ring);
+		ret = intel_init_pipe_control(ring);
 		if (ret)
 			return ret;
 	}
@@ -672,7 +883,7 @@ static int init_render_ring(struct intel_engine_cs *ring)
 	if (HAS_L3_DPF(dev))
 		I915_WRITE_IMR(ring, ~GT_PARITY_ERROR(dev));
 
-	return ret;
+	return init_workarounds_ring(ring);
 }
 
 static void render_ring_cleanup(struct intel_engine_cs *ring)
@@ -686,16 +897,7 @@ static void render_ring_cleanup(struct intel_engine_cs *ring)
 		dev_priv->semaphore_obj = NULL;
 	}
 
-	if (ring->scratch.obj == NULL)
-		return;
-
-	if (INTEL_INFO(dev)->gen >= 5) {
-		kunmap(sg_page(ring->scratch.obj->pages->sgl));
-		i915_gem_object_ggtt_unpin(ring->scratch.obj);
-	}
-
-	drm_gem_object_unreference(&ring->scratch.obj->base);
-	ring->scratch.obj = NULL;
+	intel_fini_pipe_control(ring);
 }
 
 static int gen8_rcs_signal(struct intel_engine_cs *signaller,
@@ -1018,7 +1220,7 @@ gen5_ring_get_irq(struct intel_engine_cs *ring)
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	unsigned long flags;
 
-	if (!dev->irq_enabled)
+	if (WARN_ON(!intel_irqs_enabled(dev_priv)))
 		return false;
 
 	spin_lock_irqsave(&dev_priv->irq_lock, flags);
@@ -1049,7 +1251,7 @@ i9xx_ring_get_irq(struct intel_engine_cs *ring)
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	unsigned long flags;
 
-	if (!dev->irq_enabled)
+	if (!intel_irqs_enabled(dev_priv))
 		return false;
 
 	spin_lock_irqsave(&dev_priv->irq_lock, flags);
@@ -1086,7 +1288,7 @@ i8xx_ring_get_irq(struct intel_engine_cs *ring)
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	unsigned long flags;
 
-	if (!dev->irq_enabled)
+	if (!intel_irqs_enabled(dev_priv))
 		return false;
 
 	spin_lock_irqsave(&dev_priv->irq_lock, flags);
@@ -1220,8 +1422,8 @@ gen6_ring_get_irq(struct intel_engine_cs *ring)
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	unsigned long flags;
 
-	if (!dev->irq_enabled)
-	       return false;
+	if (WARN_ON(!intel_irqs_enabled(dev_priv)))
+		return false;
 
 	spin_lock_irqsave(&dev_priv->irq_lock, flags);
 	if (ring->irq_refcount++ == 0) {
@@ -1263,7 +1465,7 @@ hsw_vebox_get_irq(struct intel_engine_cs *ring)
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	unsigned long flags;
 
-	if (!dev->irq_enabled)
+	if (WARN_ON(!intel_irqs_enabled(dev_priv)))
 		return false;
 
 	spin_lock_irqsave(&dev_priv->irq_lock, flags);
@@ -1283,9 +1485,6 @@ hsw_vebox_put_irq(struct intel_engine_cs *ring)
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	unsigned long flags;
 
-	if (!dev->irq_enabled)
-		return;
-
 	spin_lock_irqsave(&dev_priv->irq_lock, flags);
 	if (--ring->irq_refcount == 0) {
 		I915_WRITE_IMR(ring, ~0);
@@ -1301,7 +1500,7 @@ gen8_ring_get_irq(struct intel_engine_cs *ring)
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	unsigned long flags;
 
-	if (!dev->irq_enabled)
+	if (WARN_ON(!intel_irqs_enabled(dev_priv)))
 		return false;
 
 	spin_lock_irqsave(&dev_priv->irq_lock, flags);
@@ -1526,26 +1725,50 @@ static int init_phys_status_page(struct intel_engine_cs *ring)
 	return 0;
 }
 
-static void intel_destroy_ringbuffer_obj(struct intel_ringbuffer *ringbuf)
+void intel_unpin_ringbuffer_obj(struct intel_ringbuffer *ringbuf)
 {
-	if (!ringbuf->obj)
-		return;
-
 	iounmap(ringbuf->virtual_start);
+	ringbuf->virtual_start = NULL;
 	i915_gem_object_ggtt_unpin(ringbuf->obj);
-	drm_gem_object_unreference(&ringbuf->obj->base);
-	ringbuf->obj = NULL;
 }
 
-static int intel_alloc_ringbuffer_obj(struct drm_device *dev,
-				      struct intel_ringbuffer *ringbuf)
+int intel_pin_and_map_ringbuffer_obj(struct drm_device *dev,
+				     struct intel_ringbuffer *ringbuf)
 {
 	struct drm_i915_private *dev_priv = to_i915(dev);
-	struct drm_i915_gem_object *obj;
+	struct drm_i915_gem_object *obj = ringbuf->obj;
 	int ret;
 
-	if (ringbuf->obj)
-		return 0;
+	ret = i915_gem_obj_ggtt_pin(obj, PAGE_SIZE, PIN_MAPPABLE);
+	if (ret)
+		return ret;
+
+	ret = i915_gem_object_set_to_gtt_domain(obj, true);
+	if (ret) {
+		i915_gem_object_ggtt_unpin(obj);
+		return ret;
+	}
+
+	ringbuf->virtual_start = ioremap_wc(dev_priv->gtt.mappable_base +
+			i915_gem_obj_ggtt_offset(obj), ringbuf->size);
+	if (ringbuf->virtual_start == NULL) {
+		i915_gem_object_ggtt_unpin(obj);
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+void intel_destroy_ringbuffer_obj(struct intel_ringbuffer *ringbuf)
+{
+	drm_gem_object_unreference(&ringbuf->obj->base);
+	ringbuf->obj = NULL;
+}
+
+int intel_alloc_ringbuffer_obj(struct drm_device *dev,
+			       struct intel_ringbuffer *ringbuf)
+{
+	struct drm_i915_gem_object *obj;
 
 	obj = NULL;
 	if (!HAS_LLC(dev))
@@ -1558,30 +1781,9 @@ static int intel_alloc_ringbuffer_obj(struct drm_device *dev,
 	/* mark ring buffers as read-only from GPU side by default */
 	obj->gt_ro = 1;
 
-	ret = i915_gem_obj_ggtt_pin(obj, PAGE_SIZE, PIN_MAPPABLE);
-	if (ret)
-		goto err_unref;
-
-	ret = i915_gem_object_set_to_gtt_domain(obj, true);
-	if (ret)
-		goto err_unpin;
-
-	ringbuf->virtual_start =
-		ioremap_wc(dev_priv->gtt.mappable_base + i915_gem_obj_ggtt_offset(obj),
-				ringbuf->size);
-	if (ringbuf->virtual_start == NULL) {
-		ret = -EINVAL;
-		goto err_unpin;
-	}
-
 	ringbuf->obj = obj;
-	return 0;
 
-err_unpin:
-	i915_gem_object_ggtt_unpin(obj);
-err_unref:
-	drm_gem_object_unreference(&obj->base);
-	return ret;
+	return 0;
 }
 
 static int intel_init_ring_buffer(struct drm_device *dev,
@@ -1600,7 +1802,9 @@ static int intel_init_ring_buffer(struct drm_device *dev,
 	ring->dev = dev;
 	INIT_LIST_HEAD(&ring->active_list);
 	INIT_LIST_HEAD(&ring->request_list);
+	INIT_LIST_HEAD(&ring->execlist_queue);
 	ringbuf->size = 32 * PAGE_SIZE;
+	ringbuf->ring = ring;
 	memset(ring->semaphore.sync_seqno, 0, sizeof(ring->semaphore.sync_seqno));
 
 	init_waitqueue_head(&ring->irq_queue);
@@ -1616,10 +1820,21 @@ static int intel_init_ring_buffer(struct drm_device *dev,
 			goto error;
 	}
 
-	ret = intel_alloc_ringbuffer_obj(dev, ringbuf);
-	if (ret) {
-		DRM_ERROR("Failed to allocate ringbuffer %s: %d\n", ring->name, ret);
-		goto error;
+	if (ringbuf->obj == NULL) {
+		ret = intel_alloc_ringbuffer_obj(dev, ringbuf);
+		if (ret) {
+			DRM_ERROR("Failed to allocate ringbuffer %s: %d\n",
+					ring->name, ret);
+			goto error;
+		}
+
+		ret = intel_pin_and_map_ringbuffer_obj(dev, ringbuf);
+		if (ret) {
+			DRM_ERROR("Failed to pin and map ringbuffer %s: %d\n",
+					ring->name, ret);
+			intel_destroy_ringbuffer_obj(ringbuf);
+			goto error;
+		}
 	}
 
 	/* Workaround an erratum on the i830 which causes a hang if
@@ -1648,15 +1863,19 @@ error:
 
 void intel_cleanup_ring_buffer(struct intel_engine_cs *ring)
 {
-	struct drm_i915_private *dev_priv = to_i915(ring->dev);
-	struct intel_ringbuffer *ringbuf = ring->buffer;
+	struct drm_i915_private *dev_priv;
+	struct intel_ringbuffer *ringbuf;
 
 	if (!intel_ring_initialized(ring))
 		return;
 
+	dev_priv = to_i915(ring->dev);
+	ringbuf = ring->buffer;
+
 	intel_stop_ring_buffer(ring);
 	WARN_ON(!IS_GEN2(ring->dev) && (I915_READ_MODE(ring) & MODE_IDLE) == 0);
 
+	intel_unpin_ringbuffer_obj(ringbuf);
 	intel_destroy_ringbuffer_obj(ringbuf);
 	ring->preallocated_lazy_request = NULL;
 	ring->outstanding_lazy_seqno = 0;
@@ -1683,13 +1902,14 @@ static int intel_ring_wait_request(struct intel_engine_cs *ring, int n)
 		ringbuf->head = ringbuf->last_retired_head;
 		ringbuf->last_retired_head = -1;
 
-		ringbuf->space = ring_space(ringbuf);
+		ringbuf->space = intel_ring_space(ringbuf);
 		if (ringbuf->space >= n)
 			return 0;
 	}
 
 	list_for_each_entry(request, &ring->request_list, list) {
-		if (__ring_space(request->tail, ringbuf->tail, ringbuf->size) >= n) {
+		if (__intel_ring_space(request->tail, ringbuf->tail,
+				       ringbuf->size) >= n) {
 			seqno = request->seqno;
 			break;
 		}
@@ -1706,7 +1926,7 @@ static int intel_ring_wait_request(struct intel_engine_cs *ring, int n)
 	ringbuf->head = ringbuf->last_retired_head;
 	ringbuf->last_retired_head = -1;
 
-	ringbuf->space = ring_space(ringbuf);
+	ringbuf->space = intel_ring_space(ringbuf);
 	return 0;
 }
 
@@ -1735,19 +1955,12 @@ static int ring_wait_for_space(struct intel_engine_cs *ring, int n)
 	trace_i915_ring_wait_begin(ring);
 	do {
 		ringbuf->head = I915_READ_HEAD(ring);
-		ringbuf->space = ring_space(ringbuf);
+		ringbuf->space = intel_ring_space(ringbuf);
 		if (ringbuf->space >= n) {
 			ret = 0;
 			break;
 		}
 
-		if (!drm_core_check_feature(dev, DRIVER_MODESET) &&
-		    dev->primary->master) {
-			struct drm_i915_master_private *master_priv = dev->primary->master->driver_priv;
-			if (master_priv->sarea_priv)
-				master_priv->sarea_priv->perf_boxes |= I915_BOX_WAIT;
-		}
-
 		msleep(1);
 
 		if (dev_priv->mm.interruptible && signal_pending(current)) {
@@ -1787,7 +2000,7 @@ static int intel_wrap_ring_buffer(struct intel_engine_cs *ring)
 		iowrite32(MI_NOOP, virt++);
 
 	ringbuf->tail = 0;
-	ringbuf->space = ring_space(ringbuf);
+	ringbuf->space = intel_ring_space(ringbuf);
 
 	return 0;
 }
@@ -1992,9 +2205,7 @@ gen8_ring_dispatch_execbuffer(struct intel_engine_cs *ring,
 			      u64 offset, u32 len,
 			      unsigned flags)
 {
-	struct drm_i915_private *dev_priv = ring->dev->dev_private;
-	bool ppgtt = dev_priv->mm.aliasing_ppgtt != NULL &&
-		!(flags & I915_DISPATCH_SECURE);
+	bool ppgtt = USES_PPGTT(ring->dev) && !(flags & I915_DISPATCH_SECURE);
 	int ret;
 
 	ret = intel_ring_begin(ring, 4);
@@ -2023,8 +2234,9 @@ hsw_ring_dispatch_execbuffer(struct intel_engine_cs *ring,
 		return ret;
 
 	intel_ring_emit(ring,
-			MI_BATCH_BUFFER_START | MI_BATCH_PPGTT_HSW |
-			(flags & I915_DISPATCH_SECURE ? 0 : MI_BATCH_NON_SECURE_HSW));
+			MI_BATCH_BUFFER_START |
+			(flags & I915_DISPATCH_SECURE ?
+			 0 : MI_BATCH_PPGTT_HSW | MI_BATCH_NON_SECURE_HSW));
 	/* bit0-7 is the length on GEN6+ */
 	intel_ring_emit(ring, offset);
 	intel_ring_advance(ring);
@@ -2059,6 +2271,7 @@ static int gen6_ring_flush(struct intel_engine_cs *ring,
 			   u32 invalidate, u32 flush)
 {
 	struct drm_device *dev = ring->dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
 	uint32_t cmd;
 	int ret;
 
@@ -2089,8 +2302,12 @@ static int gen6_ring_flush(struct intel_engine_cs *ring,
 	}
 	intel_ring_advance(ring);
 
-	if (IS_GEN7(dev) && !invalidate && flush)
-		return gen7_ring_fbc_flush(ring, FBC_REND_CACHE_CLEAN);
+	if (!invalidate && flush) {
+		if (IS_GEN7(dev))
+			return gen7_ring_fbc_flush(ring, FBC_REND_CACHE_CLEAN);
+		else if (IS_BROADWELL(dev))
+			dev_priv->fbc.need_sw_cache_clean = true;
+	}
 
 	return 0;
 }
@@ -2123,6 +2340,8 @@ int intel_init_render_ring_buffer(struct drm_device *dev)
 					dev_priv->semaphore_obj = obj;
 			}
 		}
+
+		ring->init_context = intel_ring_workarounds_emit;
 		ring->add_request = gen6_add_request;
 		ring->flush = gen8_render_ring_flush;
 		ring->irq_get = gen8_ring_get_irq;
@@ -2232,91 +2451,6 @@ int intel_init_render_ring_buffer(struct drm_device *dev)
 	return intel_init_ring_buffer(dev, ring);
 }
 
-int intel_render_ring_init_dri(struct drm_device *dev, u64 start, u32 size)
-{
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct intel_engine_cs *ring = &dev_priv->ring[RCS];
-	struct intel_ringbuffer *ringbuf = ring->buffer;
-	int ret;
-
-	if (ringbuf == NULL) {
-		ringbuf = kzalloc(sizeof(*ringbuf), GFP_KERNEL);
-		if (!ringbuf)
-			return -ENOMEM;
-		ring->buffer = ringbuf;
-	}
-
-	ring->name = "render ring";
-	ring->id = RCS;
-	ring->mmio_base = RENDER_RING_BASE;
-
-	if (INTEL_INFO(dev)->gen >= 6) {
-		/* non-kms not supported on gen6+ */
-		ret = -ENODEV;
-		goto err_ringbuf;
-	}
-
-	/* Note: gem is not supported on gen5/ilk without kms (the corresponding
-	 * gem_init ioctl returns with -ENODEV). Hence we do not need to set up
-	 * the special gen5 functions. */
-	ring->add_request = i9xx_add_request;
-	if (INTEL_INFO(dev)->gen < 4)
-		ring->flush = gen2_render_ring_flush;
-	else
-		ring->flush = gen4_render_ring_flush;
-	ring->get_seqno = ring_get_seqno;
-	ring->set_seqno = ring_set_seqno;
-	if (IS_GEN2(dev)) {
-		ring->irq_get = i8xx_ring_get_irq;
-		ring->irq_put = i8xx_ring_put_irq;
-	} else {
-		ring->irq_get = i9xx_ring_get_irq;
-		ring->irq_put = i9xx_ring_put_irq;
-	}
-	ring->irq_enable_mask = I915_USER_INTERRUPT;
-	ring->write_tail = ring_write_tail;
-	if (INTEL_INFO(dev)->gen >= 4)
-		ring->dispatch_execbuffer = i965_dispatch_execbuffer;
-	else if (IS_I830(dev) || IS_845G(dev))
-		ring->dispatch_execbuffer = i830_dispatch_execbuffer;
-	else
-		ring->dispatch_execbuffer = i915_dispatch_execbuffer;
-	ring->init = init_render_ring;
-	ring->cleanup = render_ring_cleanup;
-
-	ring->dev = dev;
-	INIT_LIST_HEAD(&ring->active_list);
-	INIT_LIST_HEAD(&ring->request_list);
-
-	ringbuf->size = size;
-	ringbuf->effective_size = ringbuf->size;
-	if (IS_I830(ring->dev) || IS_845G(ring->dev))
-		ringbuf->effective_size -= 2 * CACHELINE_BYTES;
-
-	ringbuf->virtual_start = ioremap_wc(start, size);
-	if (ringbuf->virtual_start == NULL) {
-		DRM_ERROR("can not ioremap virtual address for"
-			  " ring buffer\n");
-		ret = -ENOMEM;
-		goto err_ringbuf;
-	}
-
-	if (!I915_NEED_GFX_HWS(dev)) {
-		ret = init_phys_status_page(ring);
-		if (ret)
-			goto err_vstart;
-	}
-
-	return 0;
-
-err_vstart:
-	iounmap(ringbuf->virtual_start);
-err_ringbuf:
-	kfree(ringbuf);
-	ring->buffer = NULL;
-	return ret;
-}
-
 int intel_init_bsd_ring_buffer(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.h b/drivers/gpu/drm/i915/intel_ringbuffer.h
index 70525d0c2c74..fe426cff598b 100644
--- a/drivers/gpu/drm/i915/intel_ringbuffer.h
+++ b/drivers/gpu/drm/i915/intel_ringbuffer.h
@@ -5,6 +5,13 @@
 
 #define I915_CMD_HASH_ORDER 9
 
+/* Early gen2 devices have a cacheline of just 32 bytes, using 64 is overkill,
+ * but keeps the logic simple. Indeed, the whole purpose of this macro is just
+ * to give some inclination as to some of the magic values used in the various
+ * workarounds!
+ */
+#define CACHELINE_BYTES 64
+
 /*
  * Gen2 BSpec "1. Programming Environment" / 1.4.4.6 "Ring Buffer Use"
  * Gen3 BSpec "vol1c Memory Interface Functions" / 2.3.4.5 "Ring Buffer Use"
@@ -90,6 +97,15 @@ struct intel_ringbuffer {
 	struct drm_i915_gem_object *obj;
 	void __iomem *virtual_start;
 
+	struct intel_engine_cs *ring;
+
+	/*
+	 * FIXME: This backpointer is an artifact of the history of how the
+	 * execlist patches came into being. It will get removed once the basic
+	 * code has landed.
+	 */
+	struct intel_context *FIXME_lrc_ctx;
+
 	u32 head;
 	u32 tail;
 	int space;
@@ -132,6 +148,9 @@ struct  intel_engine_cs {
 
 	int		(*init)(struct intel_engine_cs *ring);
 
+	int		(*init_context)(struct intel_engine_cs *ring,
+					struct intel_context *ctx);
+
 	void		(*write_tail)(struct intel_engine_cs *ring,
 				      u32 value);
 	int __must_check (*flush)(struct intel_engine_cs *ring,
@@ -214,6 +233,19 @@ struct  intel_engine_cs {
 				  unsigned int num_dwords);
 	} semaphore;
 
+	/* Execlists */
+	spinlock_t execlist_lock;
+	struct list_head execlist_queue;
+	struct list_head execlist_retired_req_list;
+	u8 next_context_status_buffer;
+	u32             irq_keep_mask; /* bitmask for interrupts that should not be masked */
+	int		(*emit_request)(struct intel_ringbuffer *ringbuf);
+	int		(*emit_flush)(struct intel_ringbuffer *ringbuf,
+				      u32 invalidate_domains,
+				      u32 flush_domains);
+	int		(*emit_bb_start)(struct intel_ringbuffer *ringbuf,
+					 u64 offset, unsigned flags);
+
 	/**
 	 * List of objects currently involved in rendering from the
 	 * ringbuffer.
@@ -287,11 +319,7 @@ struct  intel_engine_cs {
 	u32 (*get_cmd_length_mask)(u32 cmd_header);
 };
 
-static inline bool
-intel_ring_initialized(struct intel_engine_cs *ring)
-{
-	return ring->buffer && ring->buffer->obj;
-}
+bool intel_ring_initialized(struct intel_engine_cs *ring);
 
 static inline unsigned
 intel_ring_flag(struct intel_engine_cs *ring)
@@ -355,6 +383,13 @@ intel_write_status_page(struct intel_engine_cs *ring,
 #define I915_GEM_HWS_SCRATCH_INDEX	0x30
 #define I915_GEM_HWS_SCRATCH_ADDR (I915_GEM_HWS_SCRATCH_INDEX << MI_STORE_DWORD_INDEX_SHIFT)
 
+void intel_unpin_ringbuffer_obj(struct intel_ringbuffer *ringbuf);
+int intel_pin_and_map_ringbuffer_obj(struct drm_device *dev,
+				     struct intel_ringbuffer *ringbuf);
+void intel_destroy_ringbuffer_obj(struct intel_ringbuffer *ringbuf);
+int intel_alloc_ringbuffer_obj(struct drm_device *dev,
+			       struct intel_ringbuffer *ringbuf);
+
 void intel_stop_ring_buffer(struct intel_engine_cs *ring);
 void intel_cleanup_ring_buffer(struct intel_engine_cs *ring);
 
@@ -372,6 +407,9 @@ static inline void intel_ring_advance(struct intel_engine_cs *ring)
 	struct intel_ringbuffer *ringbuf = ring->buffer;
 	ringbuf->tail &= ringbuf->size - 1;
 }
+int __intel_ring_space(int head, int tail, int size);
+int intel_ring_space(struct intel_ringbuffer *ringbuf);
+bool intel_ring_stopped(struct intel_engine_cs *ring);
 void __intel_ring_advance(struct intel_engine_cs *ring);
 
 int __must_check intel_ring_idle(struct intel_engine_cs *ring);
@@ -379,6 +417,9 @@ void intel_ring_init_seqno(struct intel_engine_cs *ring, u32 seqno);
 int intel_ring_flush_all_caches(struct intel_engine_cs *ring);
 int intel_ring_invalidate_all_caches(struct intel_engine_cs *ring);
 
+void intel_fini_pipe_control(struct intel_engine_cs *ring);
+int intel_init_pipe_control(struct intel_engine_cs *ring);
+
 int intel_init_render_ring_buffer(struct drm_device *dev);
 int intel_init_bsd_ring_buffer(struct drm_device *dev);
 int intel_init_bsd2_ring_buffer(struct drm_device *dev);
@@ -388,6 +429,8 @@ int intel_init_vebox_ring_buffer(struct drm_device *dev);
 u64 intel_ring_get_active_head(struct intel_engine_cs *ring);
 void intel_ring_setup_status_page(struct intel_engine_cs *ring);
 
+int init_workarounds_ring(struct intel_engine_cs *ring);
+
 static inline u32 intel_ring_get_tail(struct intel_ringbuffer *ringbuf)
 {
 	return ringbuf->tail;
@@ -405,7 +448,4 @@ static inline void i915_trace_irq_get(struct intel_engine_cs *ring, u32 seqno)
 		ring->trace_irq_seqno = seqno;
 }
 
-/* DRI warts */
-int intel_render_ring_init_dri(struct drm_device *dev, u64 start, u32 size);
-
 #endif /* _INTEL_RINGBUFFER_H_ */
diff --git a/drivers/gpu/drm/i915/intel_runtime_pm.c b/drivers/gpu/drm/i915/intel_runtime_pm.c
new file mode 100644
index 000000000000..ac6da7102fbb
--- /dev/null
+++ b/drivers/gpu/drm/i915/intel_runtime_pm.c
@@ -0,0 +1,1379 @@
+/*
+ * Copyright © 2012-2014 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ *
+ * Authors:
+ *    Eugeni Dodonov <eugeni.dodonov@intel.com>
+ *    Daniel Vetter <daniel.vetter@ffwll.ch>
+ *
+ */
+
+#include <linux/pm_runtime.h>
+#include <linux/vgaarb.h>
+
+#include "i915_drv.h"
+#include "intel_drv.h"
+#include <drm/i915_powerwell.h>
+
+/**
+ * DOC: runtime pm
+ *
+ * The i915 driver supports dynamic enabling and disabling of entire hardware
+ * blocks at runtime. This is especially important on the display side where
+ * software is supposed to control many power gates manually on recent hardware,
+ * since on the GT side a lot of the power management is done by the hardware.
+ * But even there some manual control at the device level is required.
+ *
+ * Since i915 supports a diverse set of platforms with a unified codebase and
+ * hardware engineers just love to shuffle functionality around between power
+ * domains there's a sizeable amount of indirection required. This file provides
+ * generic functions to the driver for grabbing and releasing references for
+ * abstract power domains. It then maps those to the actual power wells
+ * present for a given platform.
+ */
+
+static struct i915_power_domains *hsw_pwr;
+
+#define for_each_power_well(i, power_well, domain_mask, power_domains)	\
+	for (i = 0;							\
+	     i < (power_domains)->power_well_count &&			\
+		 ((power_well) = &(power_domains)->power_wells[i]);	\
+	     i++)							\
+		if ((power_well)->domains & (domain_mask))
+
+#define for_each_power_well_rev(i, power_well, domain_mask, power_domains) \
+	for (i = (power_domains)->power_well_count - 1;			 \
+	     i >= 0 && ((power_well) = &(power_domains)->power_wells[i]);\
+	     i--)							 \
+		if ((power_well)->domains & (domain_mask))
+
+/*
+ * We should only use the power well if we explicitly asked the hardware to
+ * enable it, so check if it's enabled and also check if we've requested it to
+ * be enabled.
+ */
+static bool hsw_power_well_enabled(struct drm_i915_private *dev_priv,
+				   struct i915_power_well *power_well)
+{
+	return I915_READ(HSW_PWR_WELL_DRIVER) ==
+		     (HSW_PWR_WELL_ENABLE_REQUEST | HSW_PWR_WELL_STATE_ENABLED);
+}
+
+/**
+ * __intel_display_power_is_enabled - unlocked check for a power domain
+ * @dev_priv: i915 device instance
+ * @domain: power domain to check
+ *
+ * This is the unlocked version of intel_display_power_is_enabled() and should
+ * only be used from error capture and recovery code where deadlocks are
+ * possible.
+ *
+ * Returns:
+ * True when the power domain is enabled, false otherwise.
+ */
+bool __intel_display_power_is_enabled(struct drm_i915_private *dev_priv,
+				      enum intel_display_power_domain domain)
+{
+	struct i915_power_domains *power_domains;
+	struct i915_power_well *power_well;
+	bool is_enabled;
+	int i;
+
+	if (dev_priv->pm.suspended)
+		return false;
+
+	power_domains = &dev_priv->power_domains;
+
+	is_enabled = true;
+
+	for_each_power_well_rev(i, power_well, BIT(domain), power_domains) {
+		if (power_well->always_on)
+			continue;
+
+		if (!power_well->hw_enabled) {
+			is_enabled = false;
+			break;
+		}
+	}
+
+	return is_enabled;
+}
+
+/**
+ * intel_display_power_is_enabled - unlocked check for a power domain
+ * @dev_priv: i915 device instance
+ * @domain: power domain to check
+ *
+ * This function can be used to check the hw power domain state. It is mostly
+ * used in hardware state readout functions. Everywhere else code should rely
+ * upon explicit power domain reference counting to ensure that the hardware
+ * block is powered up before accessing it.
+ *
+ * Callers must hold the relevant modesetting locks to ensure that concurrent
+ * threads can't disable the power well while the caller tries to read a few
+ * registers.
+ *
+ * Returns:
+ * True when the power domain is enabled, false otherwise.
+ */
+bool intel_display_power_is_enabled(struct drm_i915_private *dev_priv,
+				    enum intel_display_power_domain domain)
+{
+	struct i915_power_domains *power_domains;
+	bool ret;
+
+	power_domains = &dev_priv->power_domains;
+
+	mutex_lock(&power_domains->lock);
+	ret = __intel_display_power_is_enabled(dev_priv, domain);
+	mutex_unlock(&power_domains->lock);
+
+	return ret;
+}
+
+/**
+ * intel_display_set_init_power - set the initial power domain state
+ * @dev_priv: i915 device instance
+ * @enable: whether to enable or disable the initial power domain state
+ *
+ * For simplicity our driver load/unload and system suspend/resume code assumes
+ * that all power domains are always enabled. This functions controls the state
+ * of this little hack. While the initial power domain state is enabled runtime
+ * pm is effectively disabled.
+ */
+void intel_display_set_init_power(struct drm_i915_private *dev_priv,
+				  bool enable)
+{
+	if (dev_priv->power_domains.init_power_on == enable)
+		return;
+
+	if (enable)
+		intel_display_power_get(dev_priv, POWER_DOMAIN_INIT);
+	else
+		intel_display_power_put(dev_priv, POWER_DOMAIN_INIT);
+
+	dev_priv->power_domains.init_power_on = enable;
+}
+
+/*
+ * Starting with Haswell, we have a "Power Down Well" that can be turned off
+ * when not needed anymore. We have 4 registers that can request the power well
+ * to be enabled, and it will only be disabled if none of the registers is
+ * requesting it to be enabled.
+ */
+static void hsw_power_well_post_enable(struct drm_i915_private *dev_priv)
+{
+	struct drm_device *dev = dev_priv->dev;
+
+	/*
+	 * After we re-enable the power well, if we touch VGA register 0x3d5
+	 * we'll get unclaimed register interrupts. This stops after we write
+	 * anything to the VGA MSR register. The vgacon module uses this
+	 * register all the time, so if we unbind our driver and, as a
+	 * consequence, bind vgacon, we'll get stuck in an infinite loop at
+	 * console_unlock(). So make here we touch the VGA MSR register, making
+	 * sure vgacon can keep working normally without triggering interrupts
+	 * and error messages.
+	 */
+	vga_get_uninterruptible(dev->pdev, VGA_RSRC_LEGACY_IO);
+	outb(inb(VGA_MSR_READ), VGA_MSR_WRITE);
+	vga_put(dev->pdev, VGA_RSRC_LEGACY_IO);
+
+	if (IS_BROADWELL(dev) || (INTEL_INFO(dev)->gen >= 9))
+		gen8_irq_power_well_post_enable(dev_priv);
+}
+
+static void hsw_set_power_well(struct drm_i915_private *dev_priv,
+			       struct i915_power_well *power_well, bool enable)
+{
+	bool is_enabled, enable_requested;
+	uint32_t tmp;
+
+	tmp = I915_READ(HSW_PWR_WELL_DRIVER);
+	is_enabled = tmp & HSW_PWR_WELL_STATE_ENABLED;
+	enable_requested = tmp & HSW_PWR_WELL_ENABLE_REQUEST;
+
+	if (enable) {
+		if (!enable_requested)
+			I915_WRITE(HSW_PWR_WELL_DRIVER,
+				   HSW_PWR_WELL_ENABLE_REQUEST);
+
+		if (!is_enabled) {
+			DRM_DEBUG_KMS("Enabling power well\n");
+			if (wait_for((I915_READ(HSW_PWR_WELL_DRIVER) &
+				      HSW_PWR_WELL_STATE_ENABLED), 20))
+				DRM_ERROR("Timeout enabling power well\n");
+			hsw_power_well_post_enable(dev_priv);
+		}
+
+	} else {
+		if (enable_requested) {
+			I915_WRITE(HSW_PWR_WELL_DRIVER, 0);
+			POSTING_READ(HSW_PWR_WELL_DRIVER);
+			DRM_DEBUG_KMS("Requesting to disable the power well\n");
+		}
+	}
+}
+
+static void hsw_power_well_sync_hw(struct drm_i915_private *dev_priv,
+				   struct i915_power_well *power_well)
+{
+	hsw_set_power_well(dev_priv, power_well, power_well->count > 0);
+
+	/*
+	 * We're taking over the BIOS, so clear any requests made by it since
+	 * the driver is in charge now.
+	 */
+	if (I915_READ(HSW_PWR_WELL_BIOS) & HSW_PWR_WELL_ENABLE_REQUEST)
+		I915_WRITE(HSW_PWR_WELL_BIOS, 0);
+}
+
+static void hsw_power_well_enable(struct drm_i915_private *dev_priv,
+				  struct i915_power_well *power_well)
+{
+	hsw_set_power_well(dev_priv, power_well, true);
+}
+
+static void hsw_power_well_disable(struct drm_i915_private *dev_priv,
+				   struct i915_power_well *power_well)
+{
+	hsw_set_power_well(dev_priv, power_well, false);
+}
+
+static void i9xx_always_on_power_well_noop(struct drm_i915_private *dev_priv,
+					   struct i915_power_well *power_well)
+{
+}
+
+static bool i9xx_always_on_power_well_enabled(struct drm_i915_private *dev_priv,
+					     struct i915_power_well *power_well)
+{
+	return true;
+}
+
+static void vlv_set_power_well(struct drm_i915_private *dev_priv,
+			       struct i915_power_well *power_well, bool enable)
+{
+	enum punit_power_well power_well_id = power_well->data;
+	u32 mask;
+	u32 state;
+	u32 ctrl;
+
+	mask = PUNIT_PWRGT_MASK(power_well_id);
+	state = enable ? PUNIT_PWRGT_PWR_ON(power_well_id) :
+			 PUNIT_PWRGT_PWR_GATE(power_well_id);
+
+	mutex_lock(&dev_priv->rps.hw_lock);
+
+#define COND \
+	((vlv_punit_read(dev_priv, PUNIT_REG_PWRGT_STATUS) & mask) == state)
+
+	if (COND)
+		goto out;
+
+	ctrl = vlv_punit_read(dev_priv, PUNIT_REG_PWRGT_CTRL);
+	ctrl &= ~mask;
+	ctrl |= state;
+	vlv_punit_write(dev_priv, PUNIT_REG_PWRGT_CTRL, ctrl);
+
+	if (wait_for(COND, 100))
+		DRM_ERROR("timout setting power well state %08x (%08x)\n",
+			  state,
+			  vlv_punit_read(dev_priv, PUNIT_REG_PWRGT_CTRL));
+
+#undef COND
+
+out:
+	mutex_unlock(&dev_priv->rps.hw_lock);
+}
+
+static void vlv_power_well_sync_hw(struct drm_i915_private *dev_priv,
+				   struct i915_power_well *power_well)
+{
+	vlv_set_power_well(dev_priv, power_well, power_well->count > 0);
+}
+
+static void vlv_power_well_enable(struct drm_i915_private *dev_priv,
+				  struct i915_power_well *power_well)
+{
+	vlv_set_power_well(dev_priv, power_well, true);
+}
+
+static void vlv_power_well_disable(struct drm_i915_private *dev_priv,
+				   struct i915_power_well *power_well)
+{
+	vlv_set_power_well(dev_priv, power_well, false);
+}
+
+static bool vlv_power_well_enabled(struct drm_i915_private *dev_priv,
+				   struct i915_power_well *power_well)
+{
+	int power_well_id = power_well->data;
+	bool enabled = false;
+	u32 mask;
+	u32 state;
+	u32 ctrl;
+
+	mask = PUNIT_PWRGT_MASK(power_well_id);
+	ctrl = PUNIT_PWRGT_PWR_ON(power_well_id);
+
+	mutex_lock(&dev_priv->rps.hw_lock);
+
+	state = vlv_punit_read(dev_priv, PUNIT_REG_PWRGT_STATUS) & mask;
+	/*
+	 * We only ever set the power-on and power-gate states, anything
+	 * else is unexpected.
+	 */
+	WARN_ON(state != PUNIT_PWRGT_PWR_ON(power_well_id) &&
+		state != PUNIT_PWRGT_PWR_GATE(power_well_id));
+	if (state == ctrl)
+		enabled = true;
+
+	/*
+	 * A transient state at this point would mean some unexpected party
+	 * is poking at the power controls too.
+	 */
+	ctrl = vlv_punit_read(dev_priv, PUNIT_REG_PWRGT_CTRL) & mask;
+	WARN_ON(ctrl != state);
+
+	mutex_unlock(&dev_priv->rps.hw_lock);
+
+	return enabled;
+}
+
+static void vlv_display_power_well_enable(struct drm_i915_private *dev_priv,
+					  struct i915_power_well *power_well)
+{
+	WARN_ON_ONCE(power_well->data != PUNIT_POWER_WELL_DISP2D);
+
+	vlv_set_power_well(dev_priv, power_well, true);
+
+	spin_lock_irq(&dev_priv->irq_lock);
+	valleyview_enable_display_irqs(dev_priv);
+	spin_unlock_irq(&dev_priv->irq_lock);
+
+	/*
+	 * During driver initialization/resume we can avoid restoring the
+	 * part of the HW/SW state that will be inited anyway explicitly.
+	 */
+	if (dev_priv->power_domains.initializing)
+		return;
+
+	intel_hpd_init(dev_priv);
+
+	i915_redisable_vga_power_on(dev_priv->dev);
+}
+
+static void vlv_display_power_well_disable(struct drm_i915_private *dev_priv,
+					   struct i915_power_well *power_well)
+{
+	WARN_ON_ONCE(power_well->data != PUNIT_POWER_WELL_DISP2D);
+
+	spin_lock_irq(&dev_priv->irq_lock);
+	valleyview_disable_display_irqs(dev_priv);
+	spin_unlock_irq(&dev_priv->irq_lock);
+
+	vlv_set_power_well(dev_priv, power_well, false);
+
+	vlv_power_sequencer_reset(dev_priv);
+}
+
+static void vlv_dpio_cmn_power_well_enable(struct drm_i915_private *dev_priv,
+					   struct i915_power_well *power_well)
+{
+	WARN_ON_ONCE(power_well->data != PUNIT_POWER_WELL_DPIO_CMN_BC);
+
+	/*
+	 * Enable the CRI clock source so we can get at the
+	 * display and the reference clock for VGA
+	 * hotplug / manual detection.
+	 */
+	I915_WRITE(DPLL(PIPE_B), I915_READ(DPLL(PIPE_B)) |
+		   DPLL_REFA_CLK_ENABLE_VLV | DPLL_INTEGRATED_CRI_CLK_VLV);
+	udelay(1); /* >10ns for cmnreset, >0ns for sidereset */
+
+	vlv_set_power_well(dev_priv, power_well, true);
+
+	/*
+	 * From VLV2A0_DP_eDP_DPIO_driver_vbios_notes_10.docx -
+	 *  6.	De-assert cmn_reset/side_reset. Same as VLV X0.
+	 *   a.	GUnit 0x2110 bit[0] set to 1 (def 0)
+	 *   b.	The other bits such as sfr settings / modesel may all
+	 *	be set to 0.
+	 *
+	 * This should only be done on init and resume from S3 with
+	 * both PLLs disabled, or we risk losing DPIO and PLL
+	 * synchronization.
+	 */
+	I915_WRITE(DPIO_CTL, I915_READ(DPIO_CTL) | DPIO_CMNRST);
+}
+
+static void vlv_dpio_cmn_power_well_disable(struct drm_i915_private *dev_priv,
+					    struct i915_power_well *power_well)
+{
+	enum pipe pipe;
+
+	WARN_ON_ONCE(power_well->data != PUNIT_POWER_WELL_DPIO_CMN_BC);
+
+	for_each_pipe(dev_priv, pipe)
+		assert_pll_disabled(dev_priv, pipe);
+
+	/* Assert common reset */
+	I915_WRITE(DPIO_CTL, I915_READ(DPIO_CTL) & ~DPIO_CMNRST);
+
+	vlv_set_power_well(dev_priv, power_well, false);
+}
+
+static void chv_dpio_cmn_power_well_enable(struct drm_i915_private *dev_priv,
+					   struct i915_power_well *power_well)
+{
+	enum dpio_phy phy;
+
+	WARN_ON_ONCE(power_well->data != PUNIT_POWER_WELL_DPIO_CMN_BC &&
+		     power_well->data != PUNIT_POWER_WELL_DPIO_CMN_D);
+
+	/*
+	 * Enable the CRI clock source so we can get at the
+	 * display and the reference clock for VGA
+	 * hotplug / manual detection.
+	 */
+	if (power_well->data == PUNIT_POWER_WELL_DPIO_CMN_BC) {
+		phy = DPIO_PHY0;
+		I915_WRITE(DPLL(PIPE_B), I915_READ(DPLL(PIPE_B)) |
+			   DPLL_REFA_CLK_ENABLE_VLV);
+		I915_WRITE(DPLL(PIPE_B), I915_READ(DPLL(PIPE_B)) |
+			   DPLL_REFA_CLK_ENABLE_VLV | DPLL_INTEGRATED_CRI_CLK_VLV);
+	} else {
+		phy = DPIO_PHY1;
+		I915_WRITE(DPLL(PIPE_C), I915_READ(DPLL(PIPE_C)) |
+			   DPLL_REFA_CLK_ENABLE_VLV | DPLL_INTEGRATED_CRI_CLK_VLV);
+	}
+	udelay(1); /* >10ns for cmnreset, >0ns for sidereset */
+	vlv_set_power_well(dev_priv, power_well, true);
+
+	/* Poll for phypwrgood signal */
+	if (wait_for(I915_READ(DISPLAY_PHY_STATUS) & PHY_POWERGOOD(phy), 1))
+		DRM_ERROR("Display PHY %d is not power up\n", phy);
+
+	I915_WRITE(DISPLAY_PHY_CONTROL, I915_READ(DISPLAY_PHY_CONTROL) |
+		   PHY_COM_LANE_RESET_DEASSERT(phy));
+}
+
+static void chv_dpio_cmn_power_well_disable(struct drm_i915_private *dev_priv,
+					    struct i915_power_well *power_well)
+{
+	enum dpio_phy phy;
+
+	WARN_ON_ONCE(power_well->data != PUNIT_POWER_WELL_DPIO_CMN_BC &&
+		     power_well->data != PUNIT_POWER_WELL_DPIO_CMN_D);
+
+	if (power_well->data == PUNIT_POWER_WELL_DPIO_CMN_BC) {
+		phy = DPIO_PHY0;
+		assert_pll_disabled(dev_priv, PIPE_A);
+		assert_pll_disabled(dev_priv, PIPE_B);
+	} else {
+		phy = DPIO_PHY1;
+		assert_pll_disabled(dev_priv, PIPE_C);
+	}
+
+	I915_WRITE(DISPLAY_PHY_CONTROL, I915_READ(DISPLAY_PHY_CONTROL) &
+		   ~PHY_COM_LANE_RESET_DEASSERT(phy));
+
+	vlv_set_power_well(dev_priv, power_well, false);
+}
+
+static bool chv_pipe_power_well_enabled(struct drm_i915_private *dev_priv,
+					struct i915_power_well *power_well)
+{
+	enum pipe pipe = power_well->data;
+	bool enabled;
+	u32 state, ctrl;
+
+	mutex_lock(&dev_priv->rps.hw_lock);
+
+	state = vlv_punit_read(dev_priv, PUNIT_REG_DSPFREQ) & DP_SSS_MASK(pipe);
+	/*
+	 * We only ever set the power-on and power-gate states, anything
+	 * else is unexpected.
+	 */
+	WARN_ON(state != DP_SSS_PWR_ON(pipe) && state != DP_SSS_PWR_GATE(pipe));
+	enabled = state == DP_SSS_PWR_ON(pipe);
+
+	/*
+	 * A transient state at this point would mean some unexpected party
+	 * is poking at the power controls too.
+	 */
+	ctrl = vlv_punit_read(dev_priv, PUNIT_REG_DSPFREQ) & DP_SSC_MASK(pipe);
+	WARN_ON(ctrl << 16 != state);
+
+	mutex_unlock(&dev_priv->rps.hw_lock);
+
+	return enabled;
+}
+
+static void chv_set_pipe_power_well(struct drm_i915_private *dev_priv,
+				    struct i915_power_well *power_well,
+				    bool enable)
+{
+	enum pipe pipe = power_well->data;
+	u32 state;
+	u32 ctrl;
+
+	state = enable ? DP_SSS_PWR_ON(pipe) : DP_SSS_PWR_GATE(pipe);
+
+	mutex_lock(&dev_priv->rps.hw_lock);
+
+#define COND \
+	((vlv_punit_read(dev_priv, PUNIT_REG_DSPFREQ) & DP_SSS_MASK(pipe)) == state)
+
+	if (COND)
+		goto out;
+
+	ctrl = vlv_punit_read(dev_priv, PUNIT_REG_DSPFREQ);
+	ctrl &= ~DP_SSC_MASK(pipe);
+	ctrl |= enable ? DP_SSC_PWR_ON(pipe) : DP_SSC_PWR_GATE(pipe);
+	vlv_punit_write(dev_priv, PUNIT_REG_DSPFREQ, ctrl);
+
+	if (wait_for(COND, 100))
+		DRM_ERROR("timout setting power well state %08x (%08x)\n",
+			  state,
+			  vlv_punit_read(dev_priv, PUNIT_REG_DSPFREQ));
+
+#undef COND
+
+out:
+	mutex_unlock(&dev_priv->rps.hw_lock);
+}
+
+static void chv_pipe_power_well_sync_hw(struct drm_i915_private *dev_priv,
+					struct i915_power_well *power_well)
+{
+	chv_set_pipe_power_well(dev_priv, power_well, power_well->count > 0);
+}
+
+static void chv_pipe_power_well_enable(struct drm_i915_private *dev_priv,
+				       struct i915_power_well *power_well)
+{
+	WARN_ON_ONCE(power_well->data != PIPE_A &&
+		     power_well->data != PIPE_B &&
+		     power_well->data != PIPE_C);
+
+	chv_set_pipe_power_well(dev_priv, power_well, true);
+
+	if (power_well->data == PIPE_A) {
+		spin_lock_irq(&dev_priv->irq_lock);
+		valleyview_enable_display_irqs(dev_priv);
+		spin_unlock_irq(&dev_priv->irq_lock);
+
+		/*
+		 * During driver initialization/resume we can avoid restoring the
+		 * part of the HW/SW state that will be inited anyway explicitly.
+		 */
+		if (dev_priv->power_domains.initializing)
+			return;
+
+		intel_hpd_init(dev_priv);
+
+		i915_redisable_vga_power_on(dev_priv->dev);
+	}
+}
+
+static void chv_pipe_power_well_disable(struct drm_i915_private *dev_priv,
+					struct i915_power_well *power_well)
+{
+	WARN_ON_ONCE(power_well->data != PIPE_A &&
+		     power_well->data != PIPE_B &&
+		     power_well->data != PIPE_C);
+
+	if (power_well->data == PIPE_A) {
+		spin_lock_irq(&dev_priv->irq_lock);
+		valleyview_disable_display_irqs(dev_priv);
+		spin_unlock_irq(&dev_priv->irq_lock);
+	}
+
+	chv_set_pipe_power_well(dev_priv, power_well, false);
+
+	if (power_well->data == PIPE_A)
+		vlv_power_sequencer_reset(dev_priv);
+}
+
+/**
+ * intel_display_power_get - grab a power domain reference
+ * @dev_priv: i915 device instance
+ * @domain: power domain to reference
+ *
+ * This function grabs a power domain reference for @domain and ensures that the
+ * power domain and all its parents are powered up. Therefore users should only
+ * grab a reference to the innermost power domain they need.
+ *
+ * Any power domain reference obtained by this function must have a symmetric
+ * call to intel_display_power_put() to release the reference again.
+ */
+void intel_display_power_get(struct drm_i915_private *dev_priv,
+			     enum intel_display_power_domain domain)
+{
+	struct i915_power_domains *power_domains;
+	struct i915_power_well *power_well;
+	int i;
+
+	intel_runtime_pm_get(dev_priv);
+
+	power_domains = &dev_priv->power_domains;
+
+	mutex_lock(&power_domains->lock);
+
+	for_each_power_well(i, power_well, BIT(domain), power_domains) {
+		if (!power_well->count++) {
+			DRM_DEBUG_KMS("enabling %s\n", power_well->name);
+			power_well->ops->enable(dev_priv, power_well);
+			power_well->hw_enabled = true;
+		}
+	}
+
+	power_domains->domain_use_count[domain]++;
+
+	mutex_unlock(&power_domains->lock);
+}
+
+/**
+ * intel_display_power_put - release a power domain reference
+ * @dev_priv: i915 device instance
+ * @domain: power domain to reference
+ *
+ * This function drops the power domain reference obtained by
+ * intel_display_power_get() and might power down the corresponding hardware
+ * block right away if this is the last reference.
+ */
+void intel_display_power_put(struct drm_i915_private *dev_priv,
+			     enum intel_display_power_domain domain)
+{
+	struct i915_power_domains *power_domains;
+	struct i915_power_well *power_well;
+	int i;
+
+	power_domains = &dev_priv->power_domains;
+
+	mutex_lock(&power_domains->lock);
+
+	WARN_ON(!power_domains->domain_use_count[domain]);
+	power_domains->domain_use_count[domain]--;
+
+	for_each_power_well_rev(i, power_well, BIT(domain), power_domains) {
+		WARN_ON(!power_well->count);
+
+		if (!--power_well->count && i915.disable_power_well) {
+			DRM_DEBUG_KMS("disabling %s\n", power_well->name);
+			power_well->hw_enabled = false;
+			power_well->ops->disable(dev_priv, power_well);
+		}
+	}
+
+	mutex_unlock(&power_domains->lock);
+
+	intel_runtime_pm_put(dev_priv);
+}
+
+#define POWER_DOMAIN_MASK (BIT(POWER_DOMAIN_NUM) - 1)
+
+#define HSW_ALWAYS_ON_POWER_DOMAINS (			\
+	BIT(POWER_DOMAIN_PIPE_A) |			\
+	BIT(POWER_DOMAIN_TRANSCODER_EDP) |		\
+	BIT(POWER_DOMAIN_PORT_DDI_A_2_LANES) |		\
+	BIT(POWER_DOMAIN_PORT_DDI_A_4_LANES) |		\
+	BIT(POWER_DOMAIN_PORT_DDI_B_2_LANES) |		\
+	BIT(POWER_DOMAIN_PORT_DDI_B_4_LANES) |		\
+	BIT(POWER_DOMAIN_PORT_DDI_C_2_LANES) |		\
+	BIT(POWER_DOMAIN_PORT_DDI_C_4_LANES) |		\
+	BIT(POWER_DOMAIN_PORT_DDI_D_2_LANES) |		\
+	BIT(POWER_DOMAIN_PORT_DDI_D_4_LANES) |		\
+	BIT(POWER_DOMAIN_PORT_CRT) |			\
+	BIT(POWER_DOMAIN_PLLS) |			\
+	BIT(POWER_DOMAIN_INIT))
+#define HSW_DISPLAY_POWER_DOMAINS (				\
+	(POWER_DOMAIN_MASK & ~HSW_ALWAYS_ON_POWER_DOMAINS) |	\
+	BIT(POWER_DOMAIN_INIT))
+
+#define BDW_ALWAYS_ON_POWER_DOMAINS (			\
+	HSW_ALWAYS_ON_POWER_DOMAINS |			\
+	BIT(POWER_DOMAIN_PIPE_A_PANEL_FITTER))
+#define BDW_DISPLAY_POWER_DOMAINS (				\
+	(POWER_DOMAIN_MASK & ~BDW_ALWAYS_ON_POWER_DOMAINS) |	\
+	BIT(POWER_DOMAIN_INIT))
+
+#define VLV_ALWAYS_ON_POWER_DOMAINS	BIT(POWER_DOMAIN_INIT)
+#define VLV_DISPLAY_POWER_DOMAINS	POWER_DOMAIN_MASK
+
+#define VLV_DPIO_CMN_BC_POWER_DOMAINS (		\
+	BIT(POWER_DOMAIN_PORT_DDI_B_2_LANES) |	\
+	BIT(POWER_DOMAIN_PORT_DDI_B_4_LANES) |	\
+	BIT(POWER_DOMAIN_PORT_DDI_C_2_LANES) |	\
+	BIT(POWER_DOMAIN_PORT_DDI_C_4_LANES) |	\
+	BIT(POWER_DOMAIN_PORT_CRT) |		\
+	BIT(POWER_DOMAIN_INIT))
+
+#define VLV_DPIO_TX_B_LANES_01_POWER_DOMAINS (	\
+	BIT(POWER_DOMAIN_PORT_DDI_B_2_LANES) |	\
+	BIT(POWER_DOMAIN_PORT_DDI_B_4_LANES) |	\
+	BIT(POWER_DOMAIN_INIT))
+
+#define VLV_DPIO_TX_B_LANES_23_POWER_DOMAINS (	\
+	BIT(POWER_DOMAIN_PORT_DDI_B_4_LANES) |	\
+	BIT(POWER_DOMAIN_INIT))
+
+#define VLV_DPIO_TX_C_LANES_01_POWER_DOMAINS (	\
+	BIT(POWER_DOMAIN_PORT_DDI_C_2_LANES) |	\
+	BIT(POWER_DOMAIN_PORT_DDI_C_4_LANES) |	\
+	BIT(POWER_DOMAIN_INIT))
+
+#define VLV_DPIO_TX_C_LANES_23_POWER_DOMAINS (	\
+	BIT(POWER_DOMAIN_PORT_DDI_C_4_LANES) |	\
+	BIT(POWER_DOMAIN_INIT))
+
+#define CHV_PIPE_A_POWER_DOMAINS (	\
+	BIT(POWER_DOMAIN_PIPE_A) |	\
+	BIT(POWER_DOMAIN_INIT))
+
+#define CHV_PIPE_B_POWER_DOMAINS (	\
+	BIT(POWER_DOMAIN_PIPE_B) |	\
+	BIT(POWER_DOMAIN_INIT))
+
+#define CHV_PIPE_C_POWER_DOMAINS (	\
+	BIT(POWER_DOMAIN_PIPE_C) |	\
+	BIT(POWER_DOMAIN_INIT))
+
+#define CHV_DPIO_CMN_BC_POWER_DOMAINS (		\
+	BIT(POWER_DOMAIN_PORT_DDI_B_2_LANES) |	\
+	BIT(POWER_DOMAIN_PORT_DDI_B_4_LANES) |	\
+	BIT(POWER_DOMAIN_PORT_DDI_C_2_LANES) |	\
+	BIT(POWER_DOMAIN_PORT_DDI_C_4_LANES) |	\
+	BIT(POWER_DOMAIN_INIT))
+
+#define CHV_DPIO_CMN_D_POWER_DOMAINS (		\
+	BIT(POWER_DOMAIN_PORT_DDI_D_2_LANES) |	\
+	BIT(POWER_DOMAIN_PORT_DDI_D_4_LANES) |	\
+	BIT(POWER_DOMAIN_INIT))
+
+#define CHV_DPIO_TX_D_LANES_01_POWER_DOMAINS (	\
+	BIT(POWER_DOMAIN_PORT_DDI_D_2_LANES) |	\
+	BIT(POWER_DOMAIN_PORT_DDI_D_4_LANES) |	\
+	BIT(POWER_DOMAIN_INIT))
+
+#define CHV_DPIO_TX_D_LANES_23_POWER_DOMAINS (	\
+	BIT(POWER_DOMAIN_PORT_DDI_D_4_LANES) |	\
+	BIT(POWER_DOMAIN_INIT))
+
+static const struct i915_power_well_ops i9xx_always_on_power_well_ops = {
+	.sync_hw = i9xx_always_on_power_well_noop,
+	.enable = i9xx_always_on_power_well_noop,
+	.disable = i9xx_always_on_power_well_noop,
+	.is_enabled = i9xx_always_on_power_well_enabled,
+};
+
+static const struct i915_power_well_ops chv_pipe_power_well_ops = {
+	.sync_hw = chv_pipe_power_well_sync_hw,
+	.enable = chv_pipe_power_well_enable,
+	.disable = chv_pipe_power_well_disable,
+	.is_enabled = chv_pipe_power_well_enabled,
+};
+
+static const struct i915_power_well_ops chv_dpio_cmn_power_well_ops = {
+	.sync_hw = vlv_power_well_sync_hw,
+	.enable = chv_dpio_cmn_power_well_enable,
+	.disable = chv_dpio_cmn_power_well_disable,
+	.is_enabled = vlv_power_well_enabled,
+};
+
+static struct i915_power_well i9xx_always_on_power_well[] = {
+	{
+		.name = "always-on",
+		.always_on = 1,
+		.domains = POWER_DOMAIN_MASK,
+		.ops = &i9xx_always_on_power_well_ops,
+	},
+};
+
+static const struct i915_power_well_ops hsw_power_well_ops = {
+	.sync_hw = hsw_power_well_sync_hw,
+	.enable = hsw_power_well_enable,
+	.disable = hsw_power_well_disable,
+	.is_enabled = hsw_power_well_enabled,
+};
+
+static struct i915_power_well hsw_power_wells[] = {
+	{
+		.name = "always-on",
+		.always_on = 1,
+		.domains = HSW_ALWAYS_ON_POWER_DOMAINS,
+		.ops = &i9xx_always_on_power_well_ops,
+	},
+	{
+		.name = "display",
+		.domains = HSW_DISPLAY_POWER_DOMAINS,
+		.ops = &hsw_power_well_ops,
+	},
+};
+
+static struct i915_power_well bdw_power_wells[] = {
+	{
+		.name = "always-on",
+		.always_on = 1,
+		.domains = BDW_ALWAYS_ON_POWER_DOMAINS,
+		.ops = &i9xx_always_on_power_well_ops,
+	},
+	{
+		.name = "display",
+		.domains = BDW_DISPLAY_POWER_DOMAINS,
+		.ops = &hsw_power_well_ops,
+	},
+};
+
+static const struct i915_power_well_ops vlv_display_power_well_ops = {
+	.sync_hw = vlv_power_well_sync_hw,
+	.enable = vlv_display_power_well_enable,
+	.disable = vlv_display_power_well_disable,
+	.is_enabled = vlv_power_well_enabled,
+};
+
+static const struct i915_power_well_ops vlv_dpio_cmn_power_well_ops = {
+	.sync_hw = vlv_power_well_sync_hw,
+	.enable = vlv_dpio_cmn_power_well_enable,
+	.disable = vlv_dpio_cmn_power_well_disable,
+	.is_enabled = vlv_power_well_enabled,
+};
+
+static const struct i915_power_well_ops vlv_dpio_power_well_ops = {
+	.sync_hw = vlv_power_well_sync_hw,
+	.enable = vlv_power_well_enable,
+	.disable = vlv_power_well_disable,
+	.is_enabled = vlv_power_well_enabled,
+};
+
+static struct i915_power_well vlv_power_wells[] = {
+	{
+		.name = "always-on",
+		.always_on = 1,
+		.domains = VLV_ALWAYS_ON_POWER_DOMAINS,
+		.ops = &i9xx_always_on_power_well_ops,
+	},
+	{
+		.name = "display",
+		.domains = VLV_DISPLAY_POWER_DOMAINS,
+		.data = PUNIT_POWER_WELL_DISP2D,
+		.ops = &vlv_display_power_well_ops,
+	},
+	{
+		.name = "dpio-tx-b-01",
+		.domains = VLV_DPIO_TX_B_LANES_01_POWER_DOMAINS |
+			   VLV_DPIO_TX_B_LANES_23_POWER_DOMAINS |
+			   VLV_DPIO_TX_C_LANES_01_POWER_DOMAINS |
+			   VLV_DPIO_TX_C_LANES_23_POWER_DOMAINS,
+		.ops = &vlv_dpio_power_well_ops,
+		.data = PUNIT_POWER_WELL_DPIO_TX_B_LANES_01,
+	},
+	{
+		.name = "dpio-tx-b-23",
+		.domains = VLV_DPIO_TX_B_LANES_01_POWER_DOMAINS |
+			   VLV_DPIO_TX_B_LANES_23_POWER_DOMAINS |
+			   VLV_DPIO_TX_C_LANES_01_POWER_DOMAINS |
+			   VLV_DPIO_TX_C_LANES_23_POWER_DOMAINS,
+		.ops = &vlv_dpio_power_well_ops,
+		.data = PUNIT_POWER_WELL_DPIO_TX_B_LANES_23,
+	},
+	{
+		.name = "dpio-tx-c-01",
+		.domains = VLV_DPIO_TX_B_LANES_01_POWER_DOMAINS |
+			   VLV_DPIO_TX_B_LANES_23_POWER_DOMAINS |
+			   VLV_DPIO_TX_C_LANES_01_POWER_DOMAINS |
+			   VLV_DPIO_TX_C_LANES_23_POWER_DOMAINS,
+		.ops = &vlv_dpio_power_well_ops,
+		.data = PUNIT_POWER_WELL_DPIO_TX_C_LANES_01,
+	},
+	{
+		.name = "dpio-tx-c-23",
+		.domains = VLV_DPIO_TX_B_LANES_01_POWER_DOMAINS |
+			   VLV_DPIO_TX_B_LANES_23_POWER_DOMAINS |
+			   VLV_DPIO_TX_C_LANES_01_POWER_DOMAINS |
+			   VLV_DPIO_TX_C_LANES_23_POWER_DOMAINS,
+		.ops = &vlv_dpio_power_well_ops,
+		.data = PUNIT_POWER_WELL_DPIO_TX_C_LANES_23,
+	},
+	{
+		.name = "dpio-common",
+		.domains = VLV_DPIO_CMN_BC_POWER_DOMAINS,
+		.data = PUNIT_POWER_WELL_DPIO_CMN_BC,
+		.ops = &vlv_dpio_cmn_power_well_ops,
+	},
+};
+
+static struct i915_power_well chv_power_wells[] = {
+	{
+		.name = "always-on",
+		.always_on = 1,
+		.domains = VLV_ALWAYS_ON_POWER_DOMAINS,
+		.ops = &i9xx_always_on_power_well_ops,
+	},
+#if 0
+	{
+		.name = "display",
+		.domains = VLV_DISPLAY_POWER_DOMAINS,
+		.data = PUNIT_POWER_WELL_DISP2D,
+		.ops = &vlv_display_power_well_ops,
+	},
+#endif
+	{
+		.name = "pipe-a",
+		/*
+		 * FIXME: pipe A power well seems to be the new disp2d well.
+		 * At least all registers seem to be housed there. Figure
+		 * out if this a a temporary situation in pre-production
+		 * hardware or a permanent state of affairs.
+		 */
+		.domains = CHV_PIPE_A_POWER_DOMAINS | VLV_DISPLAY_POWER_DOMAINS,
+		.data = PIPE_A,
+		.ops = &chv_pipe_power_well_ops,
+	},
+#if 0
+	{
+		.name = "pipe-b",
+		.domains = CHV_PIPE_B_POWER_DOMAINS,
+		.data = PIPE_B,
+		.ops = &chv_pipe_power_well_ops,
+	},
+	{
+		.name = "pipe-c",
+		.domains = CHV_PIPE_C_POWER_DOMAINS,
+		.data = PIPE_C,
+		.ops = &chv_pipe_power_well_ops,
+	},
+#endif
+	{
+		.name = "dpio-common-bc",
+		/*
+		 * XXX: cmnreset for one PHY seems to disturb the other.
+		 * As a workaround keep both powered on at the same
+		 * time for now.
+		 */
+		.domains = CHV_DPIO_CMN_BC_POWER_DOMAINS | CHV_DPIO_CMN_D_POWER_DOMAINS,
+		.data = PUNIT_POWER_WELL_DPIO_CMN_BC,
+		.ops = &chv_dpio_cmn_power_well_ops,
+	},
+	{
+		.name = "dpio-common-d",
+		/*
+		 * XXX: cmnreset for one PHY seems to disturb the other.
+		 * As a workaround keep both powered on at the same
+		 * time for now.
+		 */
+		.domains = CHV_DPIO_CMN_BC_POWER_DOMAINS | CHV_DPIO_CMN_D_POWER_DOMAINS,
+		.data = PUNIT_POWER_WELL_DPIO_CMN_D,
+		.ops = &chv_dpio_cmn_power_well_ops,
+	},
+#if 0
+	{
+		.name = "dpio-tx-b-01",
+		.domains = VLV_DPIO_TX_B_LANES_01_POWER_DOMAINS |
+			   VLV_DPIO_TX_B_LANES_23_POWER_DOMAINS,
+		.ops = &vlv_dpio_power_well_ops,
+		.data = PUNIT_POWER_WELL_DPIO_TX_B_LANES_01,
+	},
+	{
+		.name = "dpio-tx-b-23",
+		.domains = VLV_DPIO_TX_B_LANES_01_POWER_DOMAINS |
+			   VLV_DPIO_TX_B_LANES_23_POWER_DOMAINS,
+		.ops = &vlv_dpio_power_well_ops,
+		.data = PUNIT_POWER_WELL_DPIO_TX_B_LANES_23,
+	},
+	{
+		.name = "dpio-tx-c-01",
+		.domains = VLV_DPIO_TX_C_LANES_01_POWER_DOMAINS |
+			   VLV_DPIO_TX_C_LANES_23_POWER_DOMAINS,
+		.ops = &vlv_dpio_power_well_ops,
+		.data = PUNIT_POWER_WELL_DPIO_TX_C_LANES_01,
+	},
+	{
+		.name = "dpio-tx-c-23",
+		.domains = VLV_DPIO_TX_C_LANES_01_POWER_DOMAINS |
+			   VLV_DPIO_TX_C_LANES_23_POWER_DOMAINS,
+		.ops = &vlv_dpio_power_well_ops,
+		.data = PUNIT_POWER_WELL_DPIO_TX_C_LANES_23,
+	},
+	{
+		.name = "dpio-tx-d-01",
+		.domains = CHV_DPIO_TX_D_LANES_01_POWER_DOMAINS |
+			   CHV_DPIO_TX_D_LANES_23_POWER_DOMAINS,
+		.ops = &vlv_dpio_power_well_ops,
+		.data = PUNIT_POWER_WELL_DPIO_TX_D_LANES_01,
+	},
+	{
+		.name = "dpio-tx-d-23",
+		.domains = CHV_DPIO_TX_D_LANES_01_POWER_DOMAINS |
+			   CHV_DPIO_TX_D_LANES_23_POWER_DOMAINS,
+		.ops = &vlv_dpio_power_well_ops,
+		.data = PUNIT_POWER_WELL_DPIO_TX_D_LANES_23,
+	},
+#endif
+};
+
+static struct i915_power_well *lookup_power_well(struct drm_i915_private *dev_priv,
+						 enum punit_power_well power_well_id)
+{
+	struct i915_power_domains *power_domains = &dev_priv->power_domains;
+	struct i915_power_well *power_well;
+	int i;
+
+	for_each_power_well(i, power_well, POWER_DOMAIN_MASK, power_domains) {
+		if (power_well->data == power_well_id)
+			return power_well;
+	}
+
+	return NULL;
+}
+
+#define set_power_wells(power_domains, __power_wells) ({		\
+	(power_domains)->power_wells = (__power_wells);			\
+	(power_domains)->power_well_count = ARRAY_SIZE(__power_wells);	\
+})
+
+/**
+ * intel_power_domains_init - initializes the power domain structures
+ * @dev_priv: i915 device instance
+ *
+ * Initializes the power domain structures for @dev_priv depending upon the
+ * supported platform.
+ */
+int intel_power_domains_init(struct drm_i915_private *dev_priv)
+{
+	struct i915_power_domains *power_domains = &dev_priv->power_domains;
+
+	mutex_init(&power_domains->lock);
+
+	/*
+	 * The enabling order will be from lower to higher indexed wells,
+	 * the disabling order is reversed.
+	 */
+	if (IS_HASWELL(dev_priv->dev)) {
+		set_power_wells(power_domains, hsw_power_wells);
+		hsw_pwr = power_domains;
+	} else if (IS_BROADWELL(dev_priv->dev)) {
+		set_power_wells(power_domains, bdw_power_wells);
+		hsw_pwr = power_domains;
+	} else if (IS_CHERRYVIEW(dev_priv->dev)) {
+		set_power_wells(power_domains, chv_power_wells);
+	} else if (IS_VALLEYVIEW(dev_priv->dev)) {
+		set_power_wells(power_domains, vlv_power_wells);
+	} else {
+		set_power_wells(power_domains, i9xx_always_on_power_well);
+	}
+
+	return 0;
+}
+
+static void intel_runtime_pm_disable(struct drm_i915_private *dev_priv)
+{
+	struct drm_device *dev = dev_priv->dev;
+	struct device *device = &dev->pdev->dev;
+
+	if (!HAS_RUNTIME_PM(dev))
+		return;
+
+	if (!intel_enable_rc6(dev))
+		return;
+
+	/* Make sure we're not suspended first. */
+	pm_runtime_get_sync(device);
+	pm_runtime_disable(device);
+}
+
+/**
+ * intel_power_domains_fini - finalizes the power domain structures
+ * @dev_priv: i915 device instance
+ *
+ * Finalizes the power domain structures for @dev_priv depending upon the
+ * supported platform. This function also disables runtime pm and ensures that
+ * the device stays powered up so that the driver can be reloaded.
+ */
+void intel_power_domains_fini(struct drm_i915_private *dev_priv)
+{
+	intel_runtime_pm_disable(dev_priv);
+
+	/* The i915.ko module is still not prepared to be loaded when
+	 * the power well is not enabled, so just enable it in case
+	 * we're going to unload/reload. */
+	intel_display_set_init_power(dev_priv, true);
+
+	hsw_pwr = NULL;
+}
+
+static void intel_power_domains_resume(struct drm_i915_private *dev_priv)
+{
+	struct i915_power_domains *power_domains = &dev_priv->power_domains;
+	struct i915_power_well *power_well;
+	int i;
+
+	mutex_lock(&power_domains->lock);
+	for_each_power_well(i, power_well, POWER_DOMAIN_MASK, power_domains) {
+		power_well->ops->sync_hw(dev_priv, power_well);
+		power_well->hw_enabled = power_well->ops->is_enabled(dev_priv,
+								     power_well);
+	}
+	mutex_unlock(&power_domains->lock);
+}
+
+static void vlv_cmnlane_wa(struct drm_i915_private *dev_priv)
+{
+	struct i915_power_well *cmn =
+		lookup_power_well(dev_priv, PUNIT_POWER_WELL_DPIO_CMN_BC);
+	struct i915_power_well *disp2d =
+		lookup_power_well(dev_priv, PUNIT_POWER_WELL_DISP2D);
+
+	/* If the display might be already active skip this */
+	if (cmn->ops->is_enabled(dev_priv, cmn) &&
+	    disp2d->ops->is_enabled(dev_priv, disp2d) &&
+	    I915_READ(DPIO_CTL) & DPIO_CMNRST)
+		return;
+
+	DRM_DEBUG_KMS("toggling display PHY side reset\n");
+
+	/* cmnlane needs DPLL registers */
+	disp2d->ops->enable(dev_priv, disp2d);
+
+	/*
+	 * From VLV2A0_DP_eDP_HDMI_DPIO_driver_vbios_notes_11.docx:
+	 * Need to assert and de-assert PHY SB reset by gating the
+	 * common lane power, then un-gating it.
+	 * Simply ungating isn't enough to reset the PHY enough to get
+	 * ports and lanes running.
+	 */
+	cmn->ops->disable(dev_priv, cmn);
+}
+
+/**
+ * intel_power_domains_init_hw - initialize hardware power domain state
+ * @dev_priv: i915 device instance
+ *
+ * This function initializes the hardware power domain state and enables all
+ * power domains using intel_display_set_init_power().
+ */
+void intel_power_domains_init_hw(struct drm_i915_private *dev_priv)
+{
+	struct drm_device *dev = dev_priv->dev;
+	struct i915_power_domains *power_domains = &dev_priv->power_domains;
+
+	power_domains->initializing = true;
+
+	if (IS_VALLEYVIEW(dev) && !IS_CHERRYVIEW(dev)) {
+		mutex_lock(&power_domains->lock);
+		vlv_cmnlane_wa(dev_priv);
+		mutex_unlock(&power_domains->lock);
+	}
+
+	/* For now, we need the power well to be always enabled. */
+	intel_display_set_init_power(dev_priv, true);
+	intel_power_domains_resume(dev_priv);
+	power_domains->initializing = false;
+}
+
+/**
+ * intel_aux_display_runtime_get - grab an auxilliary power domain reference
+ * @dev_priv: i915 device instance
+ *
+ * This function grabs a power domain reference for the auxiliary power domain
+ * (for access to the GMBUS and DP AUX blocks) and ensures that it and all its
+ * parents are powered up. Therefore users should only grab a reference to the
+ * innermost power domain they need.
+ *
+ * Any power domain reference obtained by this function must have a symmetric
+ * call to intel_aux_display_runtime_put() to release the reference again.
+ */
+void intel_aux_display_runtime_get(struct drm_i915_private *dev_priv)
+{
+	intel_runtime_pm_get(dev_priv);
+}
+
+/**
+ * intel_aux_display_runtime_put - release an auxilliary power domain reference
+ * @dev_priv: i915 device instance
+ *
+ * This function drops the auxilliary power domain reference obtained by
+ * intel_aux_display_runtime_get() and might power down the corresponding
+ * hardware block right away if this is the last reference.
+ */
+void intel_aux_display_runtime_put(struct drm_i915_private *dev_priv)
+{
+	intel_runtime_pm_put(dev_priv);
+}
+
+/**
+ * intel_runtime_pm_get - grab a runtime pm reference
+ * @dev_priv: i915 device instance
+ *
+ * This function grabs a device-level runtime pm reference (mostly used for GEM
+ * code to ensure the GTT or GT is on) and ensures that it is powered up.
+ *
+ * Any runtime pm reference obtained by this function must have a symmetric
+ * call to intel_runtime_pm_put() to release the reference again.
+ */
+void intel_runtime_pm_get(struct drm_i915_private *dev_priv)
+{
+	struct drm_device *dev = dev_priv->dev;
+	struct device *device = &dev->pdev->dev;
+
+	if (!HAS_RUNTIME_PM(dev))
+		return;
+
+	pm_runtime_get_sync(device);
+	WARN(dev_priv->pm.suspended, "Device still suspended.\n");
+}
+
+/**
+ * intel_runtime_pm_get_noresume - grab a runtime pm reference
+ * @dev_priv: i915 device instance
+ *
+ * This function grabs a device-level runtime pm reference (mostly used for GEM
+ * code to ensure the GTT or GT is on).
+ *
+ * It will _not_ power up the device but instead only check that it's powered
+ * on.  Therefore it is only valid to call this functions from contexts where
+ * the device is known to be powered up and where trying to power it up would
+ * result in hilarity and deadlocks. That pretty much means only the system
+ * suspend/resume code where this is used to grab runtime pm references for
+ * delayed setup down in work items.
+ *
+ * Any runtime pm reference obtained by this function must have a symmetric
+ * call to intel_runtime_pm_put() to release the reference again.
+ */
+void intel_runtime_pm_get_noresume(struct drm_i915_private *dev_priv)
+{
+	struct drm_device *dev = dev_priv->dev;
+	struct device *device = &dev->pdev->dev;
+
+	if (!HAS_RUNTIME_PM(dev))
+		return;
+
+	WARN(dev_priv->pm.suspended, "Getting nosync-ref while suspended.\n");
+	pm_runtime_get_noresume(device);
+}
+
+/**
+ * intel_runtime_pm_put - release a runtime pm reference
+ * @dev_priv: i915 device instance
+ *
+ * This function drops the device-level runtime pm reference obtained by
+ * intel_runtime_pm_get() and might power down the corresponding
+ * hardware block right away if this is the last reference.
+ */
+void intel_runtime_pm_put(struct drm_i915_private *dev_priv)
+{
+	struct drm_device *dev = dev_priv->dev;
+	struct device *device = &dev->pdev->dev;
+
+	if (!HAS_RUNTIME_PM(dev))
+		return;
+
+	pm_runtime_mark_last_busy(device);
+	pm_runtime_put_autosuspend(device);
+}
+
+/**
+ * intel_runtime_pm_enable - enable runtime pm
+ * @dev_priv: i915 device instance
+ *
+ * This function enables runtime pm at the end of the driver load sequence.
+ *
+ * Note that this function does currently not enable runtime pm for the
+ * subordinate display power domains. That is only done on the first modeset
+ * using intel_display_set_init_power().
+ */
+void intel_runtime_pm_enable(struct drm_i915_private *dev_priv)
+{
+	struct drm_device *dev = dev_priv->dev;
+	struct device *device = &dev->pdev->dev;
+
+	if (!HAS_RUNTIME_PM(dev))
+		return;
+
+	pm_runtime_set_active(device);
+
+	/*
+	 * RPM depends on RC6 to save restore the GT HW context, so make RC6 a
+	 * requirement.
+	 */
+	if (!intel_enable_rc6(dev)) {
+		DRM_INFO("RC6 disabled, disabling runtime PM support\n");
+		return;
+	}
+
+	pm_runtime_set_autosuspend_delay(device, 10000); /* 10s */
+	pm_runtime_mark_last_busy(device);
+	pm_runtime_use_autosuspend(device);
+
+	pm_runtime_put_autosuspend(device);
+}
+
+/* Display audio driver power well request */
+int i915_request_power_well(void)
+{
+	struct drm_i915_private *dev_priv;
+
+	if (!hsw_pwr)
+		return -ENODEV;
+
+	dev_priv = container_of(hsw_pwr, struct drm_i915_private,
+				power_domains);
+	intel_display_power_get(dev_priv, POWER_DOMAIN_AUDIO);
+	return 0;
+}
+EXPORT_SYMBOL_GPL(i915_request_power_well);
+
+/* Display audio driver power well release */
+int i915_release_power_well(void)
+{
+	struct drm_i915_private *dev_priv;
+
+	if (!hsw_pwr)
+		return -ENODEV;
+
+	dev_priv = container_of(hsw_pwr, struct drm_i915_private,
+				power_domains);
+	intel_display_power_put(dev_priv, POWER_DOMAIN_AUDIO);
+	return 0;
+}
+EXPORT_SYMBOL_GPL(i915_release_power_well);
+
+/*
+ * Private interface for the audio driver to get CDCLK in kHz.
+ *
+ * Caller must request power well using i915_request_power_well() prior to
+ * making the call.
+ */
+int i915_get_cdclk_freq(void)
+{
+	struct drm_i915_private *dev_priv;
+
+	if (!hsw_pwr)
+		return -ENODEV;
+
+	dev_priv = container_of(hsw_pwr, struct drm_i915_private,
+				power_domains);
+
+	return intel_ddi_get_cdclk_freq(dev_priv);
+}
+EXPORT_SYMBOL_GPL(i915_get_cdclk_freq);
diff --git a/drivers/gpu/drm/i915/intel_sdvo.c b/drivers/gpu/drm/i915/intel_sdvo.c
index 9350edd6728d..6d7a277458b5 100644
--- a/drivers/gpu/drm/i915/intel_sdvo.c
+++ b/drivers/gpu/drm/i915/intel_sdvo.c
@@ -1991,57 +1991,10 @@ static int intel_sdvo_get_modes(struct drm_connector *connector)
 	return !list_empty(&connector->probed_modes);
 }
 
-static void
-intel_sdvo_destroy_enhance_property(struct drm_connector *connector)
-{
-	struct intel_sdvo_connector *intel_sdvo_connector = to_intel_sdvo_connector(connector);
-	struct drm_device *dev = connector->dev;
-
-	if (intel_sdvo_connector->left)
-		drm_property_destroy(dev, intel_sdvo_connector->left);
-	if (intel_sdvo_connector->right)
-		drm_property_destroy(dev, intel_sdvo_connector->right);
-	if (intel_sdvo_connector->top)
-		drm_property_destroy(dev, intel_sdvo_connector->top);
-	if (intel_sdvo_connector->bottom)
-		drm_property_destroy(dev, intel_sdvo_connector->bottom);
-	if (intel_sdvo_connector->hpos)
-		drm_property_destroy(dev, intel_sdvo_connector->hpos);
-	if (intel_sdvo_connector->vpos)
-		drm_property_destroy(dev, intel_sdvo_connector->vpos);
-	if (intel_sdvo_connector->saturation)
-		drm_property_destroy(dev, intel_sdvo_connector->saturation);
-	if (intel_sdvo_connector->contrast)
-		drm_property_destroy(dev, intel_sdvo_connector->contrast);
-	if (intel_sdvo_connector->hue)
-		drm_property_destroy(dev, intel_sdvo_connector->hue);
-	if (intel_sdvo_connector->sharpness)
-		drm_property_destroy(dev, intel_sdvo_connector->sharpness);
-	if (intel_sdvo_connector->flicker_filter)
-		drm_property_destroy(dev, intel_sdvo_connector->flicker_filter);
-	if (intel_sdvo_connector->flicker_filter_2d)
-		drm_property_destroy(dev, intel_sdvo_connector->flicker_filter_2d);
-	if (intel_sdvo_connector->flicker_filter_adaptive)
-		drm_property_destroy(dev, intel_sdvo_connector->flicker_filter_adaptive);
-	if (intel_sdvo_connector->tv_luma_filter)
-		drm_property_destroy(dev, intel_sdvo_connector->tv_luma_filter);
-	if (intel_sdvo_connector->tv_chroma_filter)
-		drm_property_destroy(dev, intel_sdvo_connector->tv_chroma_filter);
-	if (intel_sdvo_connector->dot_crawl)
-		drm_property_destroy(dev, intel_sdvo_connector->dot_crawl);
-	if (intel_sdvo_connector->brightness)
-		drm_property_destroy(dev, intel_sdvo_connector->brightness);
-}
-
 static void intel_sdvo_destroy(struct drm_connector *connector)
 {
 	struct intel_sdvo_connector *intel_sdvo_connector = to_intel_sdvo_connector(connector);
 
-	if (intel_sdvo_connector->tv_format)
-		drm_property_destroy(connector->dev,
-				     intel_sdvo_connector->tv_format);
-
-	intel_sdvo_destroy_enhance_property(connector);
 	drm_connector_cleanup(connector);
 	kfree(intel_sdvo_connector);
 }
diff --git a/drivers/gpu/drm/i915/intel_sprite.c b/drivers/gpu/drm/i915/intel_sprite.c
index 168c6652cda1..7d9c340f7693 100644
--- a/drivers/gpu/drm/i915/intel_sprite.c
+++ b/drivers/gpu/drm/i915/intel_sprite.c
@@ -37,6 +37,20 @@
 #include <drm/i915_drm.h>
 #include "i915_drv.h"
 
+static bool
+format_is_yuv(uint32_t format)
+{
+	switch (format) {
+	case DRM_FORMAT_YUYV:
+	case DRM_FORMAT_UYVY:
+	case DRM_FORMAT_VYUY:
+	case DRM_FORMAT_YVYU:
+		return true;
+	default:
+		return false;
+	}
+}
+
 static int usecs_to_scanlines(const struct drm_display_mode *mode, int usecs)
 {
 	/* paranoia */
@@ -46,17 +60,32 @@ static int usecs_to_scanlines(const struct drm_display_mode *mode, int usecs)
 	return DIV_ROUND_UP(usecs * mode->crtc_clock, 1000 * mode->crtc_htotal);
 }
 
-static bool intel_pipe_update_start(struct intel_crtc *crtc, uint32_t *start_vbl_count)
+/**
+ * intel_pipe_update_start() - start update of a set of display registers
+ * @crtc: the crtc of which the registers are going to be updated
+ * @start_vbl_count: vblank counter return pointer used for error checking
+ *
+ * Mark the start of an update to pipe registers that should be updated
+ * atomically regarding vblank. If the next vblank will happens within
+ * the next 100 us, this function waits until the vblank passes.
+ *
+ * After a successful call to this function, interrupts will be disabled
+ * until a subsequent call to intel_pipe_update_end(). That is done to
+ * avoid random delays. The value written to @start_vbl_count should be
+ * supplied to intel_pipe_update_end() for error checking.
+ *
+ * Return: true if the call was successful
+ */
+bool intel_pipe_update_start(struct intel_crtc *crtc, uint32_t *start_vbl_count)
 {
 	struct drm_device *dev = crtc->base.dev;
 	const struct drm_display_mode *mode = &crtc->config.adjusted_mode;
 	enum pipe pipe = crtc->pipe;
 	long timeout = msecs_to_jiffies_timeout(1);
 	int scanline, min, max, vblank_start;
+	wait_queue_head_t *wq = drm_crtc_vblank_waitqueue(&crtc->base);
 	DEFINE_WAIT(wait);
 
-	WARN_ON(!drm_modeset_is_locked(&crtc->base.mutex));
-
 	vblank_start = mode->crtc_vblank_start;
 	if (mode->flags & DRM_MODE_FLAG_INTERLACE)
 		vblank_start = DIV_ROUND_UP(vblank_start, 2);
@@ -81,7 +110,7 @@ static bool intel_pipe_update_start(struct intel_crtc *crtc, uint32_t *start_vbl
 		 * other CPUs can see the task state update by the time we
 		 * read the scanline.
 		 */
-		prepare_to_wait(&crtc->vbl_wait, &wait, TASK_UNINTERRUPTIBLE);
+		prepare_to_wait(wq, &wait, TASK_UNINTERRUPTIBLE);
 
 		scanline = intel_get_crtc_scanline(crtc);
 		if (scanline < min || scanline > max)
@@ -100,7 +129,7 @@ static bool intel_pipe_update_start(struct intel_crtc *crtc, uint32_t *start_vbl
 		local_irq_disable();
 	}
 
-	finish_wait(&crtc->vbl_wait, &wait);
+	finish_wait(wq, &wait);
 
 	drm_vblank_put(dev, pipe);
 
@@ -111,7 +140,16 @@ static bool intel_pipe_update_start(struct intel_crtc *crtc, uint32_t *start_vbl
 	return true;
 }
 
-static void intel_pipe_update_end(struct intel_crtc *crtc, u32 start_vbl_count)
+/**
+ * intel_pipe_update_end() - end update of a set of display registers
+ * @crtc: the crtc of which the registers were updated
+ * @start_vbl_count: start vblank counter (used for error checking)
+ *
+ * Mark the end of an update started with intel_pipe_update_start(). This
+ * re-enables interrupts and verifies the update was actually completed
+ * before a vblank using the value of @start_vbl_count.
+ */
+void intel_pipe_update_end(struct intel_crtc *crtc, u32 start_vbl_count)
 {
 	struct drm_device *dev = crtc->base.dev;
 	enum pipe pipe = crtc->pipe;
@@ -138,6 +176,226 @@ static void intel_update_primary_plane(struct intel_crtc *crtc)
 }
 
 static void
+skl_update_plane(struct drm_plane *drm_plane, struct drm_crtc *crtc,
+		 struct drm_framebuffer *fb,
+		 struct drm_i915_gem_object *obj, int crtc_x, int crtc_y,
+		 unsigned int crtc_w, unsigned int crtc_h,
+		 uint32_t x, uint32_t y,
+		 uint32_t src_w, uint32_t src_h)
+{
+	struct drm_device *dev = drm_plane->dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct intel_plane *intel_plane = to_intel_plane(drm_plane);
+	const int pipe = intel_plane->pipe;
+	const int plane = intel_plane->plane + 1;
+	u32 plane_ctl, stride;
+	int pixel_size = drm_format_plane_cpp(fb->pixel_format, 0);
+
+	plane_ctl = I915_READ(PLANE_CTL(pipe, plane));
+
+	/* Mask out pixel format bits in case we change it */
+	plane_ctl &= ~PLANE_CTL_FORMAT_MASK;
+	plane_ctl &= ~PLANE_CTL_ORDER_RGBX;
+	plane_ctl &= ~PLANE_CTL_YUV422_ORDER_MASK;
+	plane_ctl &= ~PLANE_CTL_TILED_MASK;
+	plane_ctl &= ~PLANE_CTL_ALPHA_MASK;
+	plane_ctl &= ~PLANE_CTL_ROTATE_MASK;
+
+	/* Trickle feed has to be enabled */
+	plane_ctl &= ~PLANE_CTL_TRICKLE_FEED_DISABLE;
+
+	switch (fb->pixel_format) {
+	case DRM_FORMAT_RGB565:
+		plane_ctl |= PLANE_CTL_FORMAT_RGB_565;
+		break;
+	case DRM_FORMAT_XBGR8888:
+		plane_ctl |= PLANE_CTL_FORMAT_XRGB_8888 | PLANE_CTL_ORDER_RGBX;
+		break;
+	case DRM_FORMAT_XRGB8888:
+		plane_ctl |= PLANE_CTL_FORMAT_XRGB_8888;
+		break;
+	/*
+	 * XXX: For ARBG/ABGR formats we default to expecting scanout buffers
+	 * to be already pre-multiplied. We need to add a knob (or a different
+	 * DRM_FORMAT) for user-space to configure that.
+	 */
+	case DRM_FORMAT_ABGR8888:
+		plane_ctl |= PLANE_CTL_FORMAT_XRGB_8888 |
+			     PLANE_CTL_ORDER_RGBX |
+			     PLANE_CTL_ALPHA_SW_PREMULTIPLY;
+		break;
+	case DRM_FORMAT_ARGB8888:
+		plane_ctl |= PLANE_CTL_FORMAT_XRGB_8888 |
+			     PLANE_CTL_ALPHA_SW_PREMULTIPLY;
+		break;
+	case DRM_FORMAT_YUYV:
+		plane_ctl |= PLANE_CTL_FORMAT_YUV422 | PLANE_CTL_YUV422_YUYV;
+		break;
+	case DRM_FORMAT_YVYU:
+		plane_ctl |= PLANE_CTL_FORMAT_YUV422 | PLANE_CTL_YUV422_YVYU;
+		break;
+	case DRM_FORMAT_UYVY:
+		plane_ctl |= PLANE_CTL_FORMAT_YUV422 | PLANE_CTL_YUV422_UYVY;
+		break;
+	case DRM_FORMAT_VYUY:
+		plane_ctl |= PLANE_CTL_FORMAT_YUV422 | PLANE_CTL_YUV422_VYUY;
+		break;
+	default:
+		BUG();
+	}
+
+	switch (obj->tiling_mode) {
+	case I915_TILING_NONE:
+		stride = fb->pitches[0] >> 6;
+		break;
+	case I915_TILING_X:
+		plane_ctl |= PLANE_CTL_TILED_X;
+		stride = fb->pitches[0] >> 9;
+		break;
+	default:
+		BUG();
+	}
+	if (intel_plane->rotation == BIT(DRM_ROTATE_180))
+		plane_ctl |= PLANE_CTL_ROTATE_180;
+
+	plane_ctl |= PLANE_CTL_ENABLE;
+	plane_ctl |= PLANE_CTL_PIPE_CSC_ENABLE;
+
+	intel_update_sprite_watermarks(drm_plane, crtc, src_w, src_h,
+				       pixel_size, true,
+				       src_w != crtc_w || src_h != crtc_h);
+
+	/* Sizes are 0 based */
+	src_w--;
+	src_h--;
+	crtc_w--;
+	crtc_h--;
+
+	I915_WRITE(PLANE_OFFSET(pipe, plane), (y << 16) | x);
+	I915_WRITE(PLANE_STRIDE(pipe, plane), stride);
+	I915_WRITE(PLANE_POS(pipe, plane), (crtc_y << 16) | crtc_x);
+	I915_WRITE(PLANE_SIZE(pipe, plane), (crtc_h << 16) | crtc_w);
+	I915_WRITE(PLANE_CTL(pipe, plane), plane_ctl);
+	I915_WRITE(PLANE_SURF(pipe, plane), i915_gem_obj_ggtt_offset(obj));
+	POSTING_READ(PLANE_SURF(pipe, plane));
+}
+
+static void
+skl_disable_plane(struct drm_plane *drm_plane, struct drm_crtc *crtc)
+{
+	struct drm_device *dev = drm_plane->dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct intel_plane *intel_plane = to_intel_plane(drm_plane);
+	const int pipe = intel_plane->pipe;
+	const int plane = intel_plane->plane + 1;
+
+	I915_WRITE(PLANE_CTL(pipe, plane),
+		   I915_READ(PLANE_CTL(pipe, plane)) & ~PLANE_CTL_ENABLE);
+
+	/* Activate double buffered register update */
+	I915_WRITE(PLANE_CTL(pipe, plane), 0);
+	POSTING_READ(PLANE_CTL(pipe, plane));
+
+	intel_update_sprite_watermarks(drm_plane, crtc, 0, 0, 0, false, false);
+}
+
+static int
+skl_update_colorkey(struct drm_plane *drm_plane,
+		    struct drm_intel_sprite_colorkey *key)
+{
+	struct drm_device *dev = drm_plane->dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct intel_plane *intel_plane = to_intel_plane(drm_plane);
+	const int pipe = intel_plane->pipe;
+	const int plane = intel_plane->plane;
+	u32 plane_ctl;
+
+	I915_WRITE(PLANE_KEYVAL(pipe, plane), key->min_value);
+	I915_WRITE(PLANE_KEYMAX(pipe, plane), key->max_value);
+	I915_WRITE(PLANE_KEYMSK(pipe, plane), key->channel_mask);
+
+	plane_ctl = I915_READ(PLANE_CTL(pipe, plane));
+	plane_ctl &= ~PLANE_CTL_KEY_ENABLE_MASK;
+	if (key->flags & I915_SET_COLORKEY_DESTINATION)
+		plane_ctl |= PLANE_CTL_KEY_ENABLE_DESTINATION;
+	else if (key->flags & I915_SET_COLORKEY_SOURCE)
+		plane_ctl |= PLANE_CTL_KEY_ENABLE_SOURCE;
+	I915_WRITE(PLANE_CTL(pipe, plane), plane_ctl);
+
+	POSTING_READ(PLANE_CTL(pipe, plane));
+
+	return 0;
+}
+
+static void
+skl_get_colorkey(struct drm_plane *drm_plane,
+		 struct drm_intel_sprite_colorkey *key)
+{
+	struct drm_device *dev = drm_plane->dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct intel_plane *intel_plane = to_intel_plane(drm_plane);
+	const int pipe = intel_plane->pipe;
+	const int plane = intel_plane->plane;
+	u32 plane_ctl;
+
+	key->min_value = I915_READ(PLANE_KEYVAL(pipe, plane));
+	key->max_value = I915_READ(PLANE_KEYMAX(pipe, plane));
+	key->channel_mask = I915_READ(PLANE_KEYMSK(pipe, plane));
+
+	plane_ctl = I915_READ(PLANE_CTL(pipe, plane));
+
+	switch (plane_ctl & PLANE_CTL_KEY_ENABLE_MASK) {
+	case PLANE_CTL_KEY_ENABLE_DESTINATION:
+		key->flags = I915_SET_COLORKEY_DESTINATION;
+		break;
+	case PLANE_CTL_KEY_ENABLE_SOURCE:
+		key->flags = I915_SET_COLORKEY_SOURCE;
+		break;
+	default:
+		key->flags = I915_SET_COLORKEY_NONE;
+	}
+}
+
+static void
+chv_update_csc(struct intel_plane *intel_plane, uint32_t format)
+{
+	struct drm_i915_private *dev_priv = intel_plane->base.dev->dev_private;
+	int plane = intel_plane->plane;
+
+	/* Seems RGB data bypasses the CSC always */
+	if (!format_is_yuv(format))
+		return;
+
+	/*
+	 * BT.601 limited range YCbCr -> full range RGB
+	 *
+	 * |r|   | 6537 4769     0|   |cr  |
+	 * |g| = |-3330 4769 -1605| x |y-64|
+	 * |b|   |    0 4769  8263|   |cb  |
+	 *
+	 * Cb and Cr apparently come in as signed already, so no
+	 * need for any offset. For Y we need to remove the offset.
+	 */
+	I915_WRITE(SPCSCYGOFF(plane), SPCSC_OOFF(0) | SPCSC_IOFF(-64));
+	I915_WRITE(SPCSCCBOFF(plane), SPCSC_OOFF(0) | SPCSC_IOFF(0));
+	I915_WRITE(SPCSCCROFF(plane), SPCSC_OOFF(0) | SPCSC_IOFF(0));
+
+	I915_WRITE(SPCSCC01(plane), SPCSC_C1(4769) | SPCSC_C0(6537));
+	I915_WRITE(SPCSCC23(plane), SPCSC_C1(-3330) | SPCSC_C0(0));
+	I915_WRITE(SPCSCC45(plane), SPCSC_C1(-1605) | SPCSC_C0(4769));
+	I915_WRITE(SPCSCC67(plane), SPCSC_C1(4769) | SPCSC_C0(0));
+	I915_WRITE(SPCSCC8(plane), SPCSC_C0(8263));
+
+	I915_WRITE(SPCSCYGICLAMP(plane), SPCSC_IMAX(940) | SPCSC_IMIN(64));
+	I915_WRITE(SPCSCCBICLAMP(plane), SPCSC_IMAX(448) | SPCSC_IMIN(-448));
+	I915_WRITE(SPCSCCRICLAMP(plane), SPCSC_IMAX(448) | SPCSC_IMIN(-448));
+
+	I915_WRITE(SPCSCYGOCLAMP(plane), SPCSC_OMAX(1023) | SPCSC_OMIN(0));
+	I915_WRITE(SPCSCCBOCLAMP(plane), SPCSC_OMAX(1023) | SPCSC_OMIN(0));
+	I915_WRITE(SPCSCCROCLAMP(plane), SPCSC_OMAX(1023) | SPCSC_OMIN(0));
+}
+
+static void
 vlv_update_plane(struct drm_plane *dplane, struct drm_crtc *crtc,
 		 struct drm_framebuffer *fb,
 		 struct drm_i915_gem_object *obj, int crtc_x, int crtc_y,
@@ -163,6 +421,7 @@ vlv_update_plane(struct drm_plane *dplane, struct drm_crtc *crtc,
 	sprctl &= ~SP_PIXFORMAT_MASK;
 	sprctl &= ~SP_YUV_BYTE_ORDER_MASK;
 	sprctl &= ~SP_TILED;
+	sprctl &= ~SP_ROTATE_180;
 
 	switch (fb->pixel_format) {
 	case DRM_FORMAT_YUYV:
@@ -235,10 +494,21 @@ vlv_update_plane(struct drm_plane *dplane, struct drm_crtc *crtc,
 							fb->pitches[0]);
 	linear_offset -= sprsurf_offset;
 
+	if (intel_plane->rotation == BIT(DRM_ROTATE_180)) {
+		sprctl |= SP_ROTATE_180;
+
+		x += src_w;
+		y += src_h;
+		linear_offset += src_h * fb->pitches[0] + src_w * pixel_size;
+	}
+
 	atomic_update = intel_pipe_update_start(intel_crtc, &start_vbl_count);
 
 	intel_update_primary_plane(intel_crtc);
 
+	if (IS_CHERRYVIEW(dev) && pipe == PIPE_B)
+		chv_update_csc(intel_plane, fb->pixel_format);
+
 	I915_WRITE(SPSTRIDE(pipe, plane), fb->pitches[0]);
 	I915_WRITE(SPPOS(pipe, plane), (crtc_y << 16) | crtc_x);
 
@@ -247,6 +517,8 @@ vlv_update_plane(struct drm_plane *dplane, struct drm_crtc *crtc,
 	else
 		I915_WRITE(SPLINOFF(pipe, plane), linear_offset);
 
+	I915_WRITE(SPCONSTALPHA(pipe, plane), 0);
+
 	I915_WRITE(SPSIZE(pipe, plane), (crtc_h << 16) | crtc_w);
 	I915_WRITE(SPCNTR(pipe, plane), sprctl);
 	I915_WRITE(SPSURF(pipe, plane), i915_gem_obj_ggtt_offset(obj) +
@@ -364,6 +636,7 @@ ivb_update_plane(struct drm_plane *plane, struct drm_crtc *crtc,
 	sprctl &= ~SPRITE_RGB_ORDER_RGBX;
 	sprctl &= ~SPRITE_YUV_BYTE_ORDER_MASK;
 	sprctl &= ~SPRITE_TILED;
+	sprctl &= ~SPRITE_ROTATE_180;
 
 	switch (fb->pixel_format) {
 	case DRM_FORMAT_XBGR8888:
@@ -426,6 +699,18 @@ ivb_update_plane(struct drm_plane *plane, struct drm_crtc *crtc,
 					       pixel_size, fb->pitches[0]);
 	linear_offset -= sprsurf_offset;
 
+	if (intel_plane->rotation == BIT(DRM_ROTATE_180)) {
+		sprctl |= SPRITE_ROTATE_180;
+
+		/* HSW and BDW does this automagically in hardware */
+		if (!IS_HASWELL(dev) && !IS_BROADWELL(dev)) {
+			x += src_w;
+			y += src_h;
+			linear_offset += src_h * fb->pitches[0] +
+				src_w * pixel_size;
+		}
+	}
+
 	atomic_update = intel_pipe_update_start(intel_crtc, &start_vbl_count);
 
 	intel_update_primary_plane(intel_crtc);
@@ -571,6 +856,7 @@ ilk_update_plane(struct drm_plane *plane, struct drm_crtc *crtc,
 	dvscntr &= ~DVS_RGB_ORDER_XBGR;
 	dvscntr &= ~DVS_YUV_BYTE_ORDER_MASK;
 	dvscntr &= ~DVS_TILED;
+	dvscntr &= ~DVS_ROTATE_180;
 
 	switch (fb->pixel_format) {
 	case DRM_FORMAT_XBGR8888:
@@ -628,6 +914,14 @@ ilk_update_plane(struct drm_plane *plane, struct drm_crtc *crtc,
 					       pixel_size, fb->pitches[0]);
 	linear_offset -= dvssurf_offset;
 
+	if (intel_plane->rotation == BIT(DRM_ROTATE_180)) {
+		dvscntr |= DVS_ROTATE_180;
+
+		x += src_w;
+		y += src_h;
+		linear_offset += src_h * fb->pitches[0] + src_w * pixel_size;
+	}
+
 	atomic_update = intel_pipe_update_start(intel_crtc, &start_vbl_count);
 
 	intel_update_primary_plane(intel_crtc);
@@ -789,20 +1083,6 @@ ilk_get_colorkey(struct drm_plane *plane, struct drm_intel_sprite_colorkey *key)
 		key->flags = I915_SET_COLORKEY_NONE;
 }
 
-static bool
-format_is_yuv(uint32_t format)
-{
-	switch (format) {
-	case DRM_FORMAT_YUYV:
-	case DRM_FORMAT_UYVY:
-	case DRM_FORMAT_VYUY:
-	case DRM_FORMAT_YVYU:
-		return true;
-	default:
-		return false;
-	}
-}
-
 static bool colorkey_enabled(struct intel_plane *intel_plane)
 {
 	struct drm_intel_sprite_colorkey key;
@@ -813,57 +1093,23 @@ static bool colorkey_enabled(struct intel_plane *intel_plane)
 }
 
 static int
-intel_update_plane(struct drm_plane *plane, struct drm_crtc *crtc,
-		   struct drm_framebuffer *fb, int crtc_x, int crtc_y,
-		   unsigned int crtc_w, unsigned int crtc_h,
-		   uint32_t src_x, uint32_t src_y,
-		   uint32_t src_w, uint32_t src_h)
+intel_check_sprite_plane(struct drm_plane *plane,
+			 struct intel_plane_state *state)
 {
-	struct drm_device *dev = plane->dev;
-	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
+	struct intel_crtc *intel_crtc = to_intel_crtc(state->crtc);
 	struct intel_plane *intel_plane = to_intel_plane(plane);
-	enum pipe pipe = intel_crtc->pipe;
-	struct intel_framebuffer *intel_fb = to_intel_framebuffer(fb);
-	struct drm_i915_gem_object *obj = intel_fb->obj;
-	struct drm_i915_gem_object *old_obj = intel_plane->obj;
-	int ret;
-	bool primary_enabled;
-	bool visible;
+	struct drm_framebuffer *fb = state->fb;
+	struct drm_i915_gem_object *obj = intel_fb_obj(fb);
+	int crtc_x, crtc_y;
+	unsigned int crtc_w, crtc_h;
+	uint32_t src_x, src_y, src_w, src_h;
+	struct drm_rect *src = &state->src;
+	struct drm_rect *dst = &state->dst;
+	struct drm_rect *orig_src = &state->orig_src;
+	const struct drm_rect *clip = &state->clip;
 	int hscale, vscale;
 	int max_scale, min_scale;
 	int pixel_size = drm_format_plane_cpp(fb->pixel_format, 0);
-	struct drm_rect src = {
-		/* sample coordinates in 16.16 fixed point */
-		.x1 = src_x,
-		.x2 = src_x + src_w,
-		.y1 = src_y,
-		.y2 = src_y + src_h,
-	};
-	struct drm_rect dst = {
-		/* integer pixels */
-		.x1 = crtc_x,
-		.x2 = crtc_x + crtc_w,
-		.y1 = crtc_y,
-		.y2 = crtc_y + crtc_h,
-	};
-	const struct drm_rect clip = {
-		.x2 = intel_crtc->active ? intel_crtc->config.pipe_src_w : 0,
-		.y2 = intel_crtc->active ? intel_crtc->config.pipe_src_h : 0,
-	};
-	const struct {
-		int crtc_x, crtc_y;
-		unsigned int crtc_w, crtc_h;
-		uint32_t src_x, src_y, src_w, src_h;
-	} orig = {
-		.crtc_x = crtc_x,
-		.crtc_y = crtc_y,
-		.crtc_w = crtc_w,
-		.crtc_h = crtc_h,
-		.src_x = src_x,
-		.src_y = src_y,
-		.src_w = src_w,
-		.src_h = src_h,
-	};
 
 	/* Don't modify another pipe's plane */
 	if (intel_plane->pipe != intel_crtc->pipe) {
@@ -895,49 +1141,55 @@ intel_update_plane(struct drm_plane *plane, struct drm_crtc *crtc,
 	max_scale = intel_plane->max_downscale << 16;
 	min_scale = intel_plane->can_scale ? 1 : (1 << 16);
 
-	hscale = drm_rect_calc_hscale_relaxed(&src, &dst, min_scale, max_scale);
+	drm_rect_rotate(src, fb->width << 16, fb->height << 16,
+			intel_plane->rotation);
+
+	hscale = drm_rect_calc_hscale_relaxed(src, dst, min_scale, max_scale);
 	BUG_ON(hscale < 0);
 
-	vscale = drm_rect_calc_vscale_relaxed(&src, &dst, min_scale, max_scale);
+	vscale = drm_rect_calc_vscale_relaxed(src, dst, min_scale, max_scale);
 	BUG_ON(vscale < 0);
 
-	visible = drm_rect_clip_scaled(&src, &dst, &clip, hscale, vscale);
+	state->visible =  drm_rect_clip_scaled(src, dst, clip, hscale, vscale);
 
-	crtc_x = dst.x1;
-	crtc_y = dst.y1;
-	crtc_w = drm_rect_width(&dst);
-	crtc_h = drm_rect_height(&dst);
+	crtc_x = dst->x1;
+	crtc_y = dst->y1;
+	crtc_w = drm_rect_width(dst);
+	crtc_h = drm_rect_height(dst);
 
-	if (visible) {
+	if (state->visible) {
 		/* check again in case clipping clamped the results */
-		hscale = drm_rect_calc_hscale(&src, &dst, min_scale, max_scale);
+		hscale = drm_rect_calc_hscale(src, dst, min_scale, max_scale);
 		if (hscale < 0) {
 			DRM_DEBUG_KMS("Horizontal scaling factor out of limits\n");
-			drm_rect_debug_print(&src, true);
-			drm_rect_debug_print(&dst, false);
+			drm_rect_debug_print(src, true);
+			drm_rect_debug_print(dst, false);
 
 			return hscale;
 		}
 
-		vscale = drm_rect_calc_vscale(&src, &dst, min_scale, max_scale);
+		vscale = drm_rect_calc_vscale(src, dst, min_scale, max_scale);
 		if (vscale < 0) {
 			DRM_DEBUG_KMS("Vertical scaling factor out of limits\n");
-			drm_rect_debug_print(&src, true);
-			drm_rect_debug_print(&dst, false);
+			drm_rect_debug_print(src, true);
+			drm_rect_debug_print(dst, false);
 
 			return vscale;
 		}
 
 		/* Make the source viewport size an exact multiple of the scaling factors. */
-		drm_rect_adjust_size(&src,
-				     drm_rect_width(&dst) * hscale - drm_rect_width(&src),
-				     drm_rect_height(&dst) * vscale - drm_rect_height(&src));
+		drm_rect_adjust_size(src,
+				     drm_rect_width(dst) * hscale - drm_rect_width(src),
+				     drm_rect_height(dst) * vscale - drm_rect_height(src));
+
+		drm_rect_rotate_inv(src, fb->width << 16, fb->height << 16,
+				    intel_plane->rotation);
 
 		/* sanity check to make sure the src viewport wasn't enlarged */
-		WARN_ON(src.x1 < (int) src_x ||
-			src.y1 < (int) src_y ||
-			src.x2 > (int) (src_x + src_w) ||
-			src.y2 > (int) (src_y + src_h));
+		WARN_ON(src->x1 < (int) orig_src->x1 ||
+			src->y1 < (int) orig_src->y1 ||
+			src->x2 > (int) orig_src->x2 ||
+			src->y2 > (int) orig_src->y2);
 
 		/*
 		 * Hardware doesn't handle subpixel coordinates.
@@ -945,10 +1197,10 @@ intel_update_plane(struct drm_plane *plane, struct drm_crtc *crtc,
 		 * increase the source viewport size, because that could
 		 * push the downscaling factor out of bounds.
 		 */
-		src_x = src.x1 >> 16;
-		src_w = drm_rect_width(&src) >> 16;
-		src_y = src.y1 >> 16;
-		src_h = drm_rect_height(&src) >> 16;
+		src_x = src->x1 >> 16;
+		src_w = drm_rect_width(src) >> 16;
+		src_y = src->y1 >> 16;
+		src_h = drm_rect_height(src) >> 16;
 
 		if (format_is_yuv(fb->pixel_format)) {
 			src_x &= ~1;
@@ -962,12 +1214,12 @@ intel_update_plane(struct drm_plane *plane, struct drm_crtc *crtc,
 				crtc_w &= ~1;
 
 			if (crtc_w == 0)
-				visible = false;
+				state->visible = false;
 		}
 	}
 
 	/* Check size restrictions when scaling */
-	if (visible && (src_w != crtc_w || src_h != crtc_h)) {
+	if (state->visible && (src_w != crtc_w || src_h != crtc_h)) {
 		unsigned int width_bytes;
 
 		WARN_ON(!intel_plane->can_scale);
@@ -975,12 +1227,13 @@ intel_update_plane(struct drm_plane *plane, struct drm_crtc *crtc,
 		/* FIXME interlacing min height is 6 */
 
 		if (crtc_w < 3 || crtc_h < 3)
-			visible = false;
+			state->visible = false;
 
 		if (src_w < 3 || src_h < 3)
-			visible = false;
+			state->visible = false;
 
-		width_bytes = ((src_x * pixel_size) & 63) + src_w * pixel_size;
+		width_bytes = ((src_x * pixel_size) & 63) +
+					src_w * pixel_size;
 
 		if (src_w > 2048 || src_h > 2048 ||
 		    width_bytes > 4096 || fb->pitches[0] > 4096) {
@@ -989,42 +1242,90 @@ intel_update_plane(struct drm_plane *plane, struct drm_crtc *crtc,
 		}
 	}
 
-	dst.x1 = crtc_x;
-	dst.x2 = crtc_x + crtc_w;
-	dst.y1 = crtc_y;
-	dst.y2 = crtc_y + crtc_h;
+	if (state->visible) {
+		src->x1 = src_x;
+		src->x2 = src_x + src_w;
+		src->y1 = src_y;
+		src->y2 = src_y + src_h;
+	}
 
-	/*
-	 * If the sprite is completely covering the primary plane,
-	 * we can disable the primary and save power.
-	 */
-	primary_enabled = !drm_rect_equals(&dst, &clip) || colorkey_enabled(intel_plane);
-	WARN_ON(!primary_enabled && !visible && intel_crtc->active);
+	dst->x1 = crtc_x;
+	dst->x2 = crtc_x + crtc_w;
+	dst->y1 = crtc_y;
+	dst->y2 = crtc_y + crtc_h;
 
-	mutex_lock(&dev->struct_mutex);
+	return 0;
+}
 
-	/* Note that this will apply the VT-d workaround for scanouts,
-	 * which is more restrictive than required for sprites. (The
-	 * primary plane requires 256KiB alignment with 64 PTE padding,
-	 * the sprite planes only require 128KiB alignment and 32 PTE padding.
-	 */
-	ret = intel_pin_and_fence_fb_obj(dev, obj, NULL);
+static int
+intel_prepare_sprite_plane(struct drm_plane *plane,
+			   struct intel_plane_state *state)
+{
+	struct drm_device *dev = plane->dev;
+	struct drm_crtc *crtc = state->crtc;
+	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
+	struct intel_plane *intel_plane = to_intel_plane(plane);
+	enum pipe pipe = intel_crtc->pipe;
+	struct drm_framebuffer *fb = state->fb;
+	struct drm_i915_gem_object *obj = intel_fb_obj(fb);
+	struct drm_i915_gem_object *old_obj = intel_plane->obj;
+	int ret;
 
-	i915_gem_track_fb(old_obj, obj,
-			  INTEL_FRONTBUFFER_SPRITE(pipe));
-	mutex_unlock(&dev->struct_mutex);
+	if (old_obj != obj) {
+		mutex_lock(&dev->struct_mutex);
 
-	if (ret)
-		return ret;
+		/* Note that this will apply the VT-d workaround for scanouts,
+		 * which is more restrictive than required for sprites. (The
+		 * primary plane requires 256KiB alignment with 64 PTE padding,
+		 * the sprite planes only require 128KiB alignment and 32 PTE
+		 * padding.
+		 */
+		ret = intel_pin_and_fence_fb_obj(plane, fb, NULL);
+		if (ret == 0)
+			i915_gem_track_fb(old_obj, obj,
+					  INTEL_FRONTBUFFER_SPRITE(pipe));
+		mutex_unlock(&dev->struct_mutex);
+		if (ret)
+			return ret;
+	}
+
+	return 0;
+}
+
+static void
+intel_commit_sprite_plane(struct drm_plane *plane,
+			  struct intel_plane_state *state)
+{
+	struct drm_device *dev = plane->dev;
+	struct drm_crtc *crtc = state->crtc;
+	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
+	struct intel_plane *intel_plane = to_intel_plane(plane);
+	enum pipe pipe = intel_crtc->pipe;
+	struct drm_framebuffer *fb = state->fb;
+	struct drm_i915_gem_object *obj = intel_fb_obj(fb);
+	struct drm_i915_gem_object *old_obj = intel_plane->obj;
+	int crtc_x, crtc_y;
+	unsigned int crtc_w, crtc_h;
+	uint32_t src_x, src_y, src_w, src_h;
+	struct drm_rect *dst = &state->dst;
+	const struct drm_rect *clip = &state->clip;
+	bool primary_enabled;
 
-	intel_plane->crtc_x = orig.crtc_x;
-	intel_plane->crtc_y = orig.crtc_y;
-	intel_plane->crtc_w = orig.crtc_w;
-	intel_plane->crtc_h = orig.crtc_h;
-	intel_plane->src_x = orig.src_x;
-	intel_plane->src_y = orig.src_y;
-	intel_plane->src_w = orig.src_w;
-	intel_plane->src_h = orig.src_h;
+	/*
+	 * If the sprite is completely covering the primary plane,
+	 * we can disable the primary and save power.
+	 */
+	primary_enabled = !drm_rect_equals(dst, clip) || colorkey_enabled(intel_plane);
+	WARN_ON(!primary_enabled && !state->visible && intel_crtc->active);
+
+	intel_plane->crtc_x = state->orig_dst.x1;
+	intel_plane->crtc_y = state->orig_dst.y1;
+	intel_plane->crtc_w = drm_rect_width(&state->orig_dst);
+	intel_plane->crtc_h = drm_rect_height(&state->orig_dst);
+	intel_plane->src_x = state->orig_src.x1;
+	intel_plane->src_y = state->orig_src.y1;
+	intel_plane->src_w = drm_rect_width(&state->orig_src);
+	intel_plane->src_h = drm_rect_height(&state->orig_src);
 	intel_plane->obj = obj;
 
 	if (intel_crtc->active) {
@@ -1038,12 +1339,22 @@ intel_update_plane(struct drm_plane *plane, struct drm_crtc *crtc,
 		if (primary_was_enabled && !primary_enabled)
 			intel_pre_disable_primary(crtc);
 
-		if (visible)
+		if (state->visible) {
+			crtc_x = state->dst.x1;
+			crtc_y = state->dst.y1;
+			crtc_w = drm_rect_width(&state->dst);
+			crtc_h = drm_rect_height(&state->dst);
+			src_x = state->src.x1;
+			src_y = state->src.y1;
+			src_w = drm_rect_width(&state->src);
+			src_h = drm_rect_height(&state->src);
 			intel_plane->update_plane(plane, crtc, fb, obj,
 						  crtc_x, crtc_y, crtc_w, crtc_h,
 						  src_x, src_y, src_w, src_h);
-		else
+		} else {
 			intel_plane->disable_plane(plane, crtc);
+		}
+
 
 		intel_frontbuffer_flip(dev, INTEL_FRONTBUFFER_SPRITE(pipe));
 
@@ -1052,21 +1363,65 @@ intel_update_plane(struct drm_plane *plane, struct drm_crtc *crtc,
 	}
 
 	/* Unpin old obj after new one is active to avoid ugliness */
-	if (old_obj) {
+	if (old_obj && old_obj != obj) {
+
 		/*
 		 * It's fairly common to simply update the position of
 		 * an existing object.  In that case, we don't need to
 		 * wait for vblank to avoid ugliness, we only need to
 		 * do the pin & ref bookkeeping.
 		 */
-		if (old_obj != obj && intel_crtc->active)
+		if (intel_crtc->active)
 			intel_wait_for_vblank(dev, intel_crtc->pipe);
 
 		mutex_lock(&dev->struct_mutex);
 		intel_unpin_fb_obj(old_obj);
 		mutex_unlock(&dev->struct_mutex);
 	}
+}
 
+static int
+intel_update_plane(struct drm_plane *plane, struct drm_crtc *crtc,
+		   struct drm_framebuffer *fb, int crtc_x, int crtc_y,
+		   unsigned int crtc_w, unsigned int crtc_h,
+		   uint32_t src_x, uint32_t src_y,
+		   uint32_t src_w, uint32_t src_h)
+{
+	struct intel_plane_state state;
+	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
+	int ret;
+
+	state.crtc = crtc;
+	state.fb = fb;
+
+	/* sample coordinates in 16.16 fixed point */
+	state.src.x1 = src_x;
+	state.src.x2 = src_x + src_w;
+	state.src.y1 = src_y;
+	state.src.y2 = src_y + src_h;
+
+	/* integer pixels */
+	state.dst.x1 = crtc_x;
+	state.dst.x2 = crtc_x + crtc_w;
+	state.dst.y1 = crtc_y;
+	state.dst.y2 = crtc_y + crtc_h;
+
+	state.clip.x1 = 0;
+	state.clip.y1 = 0;
+	state.clip.x2 = intel_crtc->active ? intel_crtc->config.pipe_src_w : 0;
+	state.clip.y2 = intel_crtc->active ? intel_crtc->config.pipe_src_h : 0;
+	state.orig_src = state.src;
+	state.orig_dst = state.dst;
+
+	ret = intel_check_sprite_plane(plane, &state);
+	if (ret)
+		return ret;
+
+	ret = intel_prepare_sprite_plane(plane, &state);
+	if (ret)
+		return ret;
+
+	intel_commit_sprite_plane(plane, &state);
 	return 0;
 }
 
@@ -1180,18 +1535,45 @@ out_unlock:
 	return ret;
 }
 
-void intel_plane_restore(struct drm_plane *plane)
+int intel_plane_set_property(struct drm_plane *plane,
+			     struct drm_property *prop,
+			     uint64_t val)
+{
+	struct drm_device *dev = plane->dev;
+	struct intel_plane *intel_plane = to_intel_plane(plane);
+	uint64_t old_val;
+	int ret = -ENOENT;
+
+	if (prop == dev->mode_config.rotation_property) {
+		/* exactly one rotation angle please */
+		if (hweight32(val & 0xf) != 1)
+			return -EINVAL;
+
+		if (intel_plane->rotation == val)
+			return 0;
+
+		old_val = intel_plane->rotation;
+		intel_plane->rotation = val;
+		ret = intel_plane_restore(plane);
+		if (ret)
+			intel_plane->rotation = old_val;
+	}
+
+	return ret;
+}
+
+int intel_plane_restore(struct drm_plane *plane)
 {
 	struct intel_plane *intel_plane = to_intel_plane(plane);
 
 	if (!plane->crtc || !plane->fb)
-		return;
+		return 0;
 
-	intel_update_plane(plane, plane->crtc, plane->fb,
-			   intel_plane->crtc_x, intel_plane->crtc_y,
-			   intel_plane->crtc_w, intel_plane->crtc_h,
-			   intel_plane->src_x, intel_plane->src_y,
-			   intel_plane->src_w, intel_plane->src_h);
+	return plane->funcs->update_plane(plane, plane->crtc, plane->fb,
+				  intel_plane->crtc_x, intel_plane->crtc_y,
+				  intel_plane->crtc_w, intel_plane->crtc_h,
+				  intel_plane->src_x, intel_plane->src_y,
+				  intel_plane->src_w, intel_plane->src_h);
 }
 
 void intel_plane_disable(struct drm_plane *plane)
@@ -1206,6 +1588,7 @@ static const struct drm_plane_funcs intel_plane_funcs = {
 	.update_plane = intel_update_plane,
 	.disable_plane = intel_disable_plane,
 	.destroy = intel_destroy_plane,
+	.set_property = intel_plane_set_property,
 };
 
 static uint32_t ilk_plane_formats[] = {
@@ -1239,6 +1622,18 @@ static uint32_t vlv_plane_formats[] = {
 	DRM_FORMAT_VYUY,
 };
 
+static uint32_t skl_plane_formats[] = {
+	DRM_FORMAT_RGB565,
+	DRM_FORMAT_ABGR8888,
+	DRM_FORMAT_ARGB8888,
+	DRM_FORMAT_XBGR8888,
+	DRM_FORMAT_XRGB8888,
+	DRM_FORMAT_YUYV,
+	DRM_FORMAT_YVYU,
+	DRM_FORMAT_UYVY,
+	DRM_FORMAT_VYUY,
+};
+
 int
 intel_plane_init(struct drm_device *dev, enum pipe pipe, int plane)
 {
@@ -1302,7 +1697,21 @@ intel_plane_init(struct drm_device *dev, enum pipe pipe, int plane)
 			num_plane_formats = ARRAY_SIZE(snb_plane_formats);
 		}
 		break;
-
+	case 9:
+		/*
+		 * FIXME: Skylake planes can be scaled (with some restrictions),
+		 * but this is for another time.
+		 */
+		intel_plane->can_scale = false;
+		intel_plane->max_downscale = 1;
+		intel_plane->update_plane = skl_update_plane;
+		intel_plane->disable_plane = skl_disable_plane;
+		intel_plane->update_colorkey = skl_update_colorkey;
+		intel_plane->get_colorkey = skl_get_colorkey;
+
+		plane_formats = skl_plane_formats;
+		num_plane_formats = ARRAY_SIZE(skl_plane_formats);
+		break;
 	default:
 		kfree(intel_plane);
 		return -ENODEV;
@@ -1310,13 +1719,28 @@ intel_plane_init(struct drm_device *dev, enum pipe pipe, int plane)
 
 	intel_plane->pipe = pipe;
 	intel_plane->plane = plane;
+	intel_plane->rotation = BIT(DRM_ROTATE_0);
 	possible_crtcs = (1 << pipe);
-	ret = drm_plane_init(dev, &intel_plane->base, possible_crtcs,
-			     &intel_plane_funcs,
-			     plane_formats, num_plane_formats,
-			     false);
-	if (ret)
+	ret = drm_universal_plane_init(dev, &intel_plane->base, possible_crtcs,
+				       &intel_plane_funcs,
+				       plane_formats, num_plane_formats,
+				       DRM_PLANE_TYPE_OVERLAY);
+	if (ret) {
 		kfree(intel_plane);
+		goto out;
+	}
+
+	if (!dev->mode_config.rotation_property)
+		dev->mode_config.rotation_property =
+			drm_mode_create_rotation_property(dev,
+							  BIT(DRM_ROTATE_0) |
+							  BIT(DRM_ROTATE_180));
+
+	if (dev->mode_config.rotation_property)
+		drm_object_attach_property(&intel_plane->base.base,
+					   dev->mode_config.rotation_property,
+					   intel_plane->rotation);
 
+ out:
 	return ret;
 }
diff --git a/drivers/gpu/drm/i915/intel_tv.c b/drivers/gpu/drm/i915/intel_tv.c
index c14341ca3ef9..6f5f59b880f5 100644
--- a/drivers/gpu/drm/i915/intel_tv.c
+++ b/drivers/gpu/drm/i915/intel_tv.c
@@ -1182,18 +1182,17 @@ intel_tv_detect_type(struct intel_tv *intel_tv,
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 	struct drm_device *dev = encoder->dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	unsigned long irqflags;
 	u32 tv_ctl, save_tv_ctl;
 	u32 tv_dac, save_tv_dac;
 	int type;
 
 	/* Disable TV interrupts around load detect or we'll recurse */
 	if (connector->polled & DRM_CONNECTOR_POLL_HPD) {
-		spin_lock_irqsave(&dev_priv->irq_lock, irqflags);
+		spin_lock_irq(&dev_priv->irq_lock);
 		i915_disable_pipestat(dev_priv, 0,
 				      PIPE_HOTPLUG_INTERRUPT_STATUS |
 				      PIPE_HOTPLUG_TV_INTERRUPT_STATUS);
-		spin_unlock_irqrestore(&dev_priv->irq_lock, irqflags);
+		spin_unlock_irq(&dev_priv->irq_lock);
 	}
 
 	save_tv_dac = tv_dac = I915_READ(TV_DAC);
@@ -1266,11 +1265,11 @@ intel_tv_detect_type(struct intel_tv *intel_tv,
 
 	/* Restore interrupt config */
 	if (connector->polled & DRM_CONNECTOR_POLL_HPD) {
-		spin_lock_irqsave(&dev_priv->irq_lock, irqflags);
+		spin_lock_irq(&dev_priv->irq_lock);
 		i915_enable_pipestat(dev_priv, 0,
 				     PIPE_HOTPLUG_INTERRUPT_STATUS |
 				     PIPE_HOTPLUG_TV_INTERRUPT_STATUS);
-		spin_unlock_irqrestore(&dev_priv->irq_lock, irqflags);
+		spin_unlock_irq(&dev_priv->irq_lock);
 	}
 
 	return type;
diff --git a/drivers/gpu/drm/i915/intel_uncore.c b/drivers/gpu/drm/i915/intel_uncore.c
index e81bc3bdc533..46de8d75b4bf 100644
--- a/drivers/gpu/drm/i915/intel_uncore.c
+++ b/drivers/gpu/drm/i915/intel_uncore.c
@@ -43,23 +43,17 @@
 static void
 assert_device_not_suspended(struct drm_i915_private *dev_priv)
 {
-	WARN(HAS_RUNTIME_PM(dev_priv->dev) && dev_priv->pm.suspended,
-	     "Device suspended\n");
+	WARN_ONCE(HAS_RUNTIME_PM(dev_priv->dev) && dev_priv->pm.suspended,
+		  "Device suspended\n");
 }
 
 static void __gen6_gt_wait_for_thread_c0(struct drm_i915_private *dev_priv)
 {
-	u32 gt_thread_status_mask;
-
-	if (IS_HASWELL(dev_priv->dev))
-		gt_thread_status_mask = GEN6_GT_THREAD_STATUS_CORE_MASK_HSW;
-	else
-		gt_thread_status_mask = GEN6_GT_THREAD_STATUS_CORE_MASK;
-
 	/* w/a for a sporadic read returning 0 by waiting for the GT
 	 * thread to wake up.
 	 */
-	if (wait_for_atomic_us((__raw_i915_read32(dev_priv, GEN6_GT_THREAD_STATUS_REG) & gt_thread_status_mask) == 0, 500))
+	if (wait_for_atomic_us((__raw_i915_read32(dev_priv, GEN6_GT_THREAD_STATUS_REG) &
+				GEN6_GT_THREAD_STATUS_CORE_MASK) == 0, 500))
 		DRM_ERROR("GT thread status wait timed out\n");
 }
 
@@ -101,7 +95,7 @@ static void __gen7_gt_force_wake_mt_get(struct drm_i915_private *dev_priv,
 {
 	u32 forcewake_ack;
 
-	if (IS_HASWELL(dev_priv->dev) || IS_GEN8(dev_priv->dev))
+	if (IS_HASWELL(dev_priv->dev) || IS_BROADWELL(dev_priv->dev))
 		forcewake_ack = FORCEWAKE_ACK_HSW;
 	else
 		forcewake_ack = FORCEWAKE_MT_ACK;
@@ -120,8 +114,7 @@ static void __gen7_gt_force_wake_mt_get(struct drm_i915_private *dev_priv,
 		DRM_ERROR("Timed out waiting for forcewake to ack request.\n");
 
 	/* WaRsForcewakeWaitTC0:ivb,hsw */
-	if (INTEL_INFO(dev_priv->dev)->gen < 8)
-		__gen6_gt_wait_for_thread_c0(dev_priv);
+	__gen6_gt_wait_for_thread_c0(dev_priv);
 }
 
 static void gen6_gt_check_fifodbg(struct drm_i915_private *dev_priv)
@@ -229,10 +222,6 @@ static void __vlv_force_wake_get(struct drm_i915_private *dev_priv,
 					FORCEWAKE_ACK_TIMEOUT_MS))
 			DRM_ERROR("Timed out: waiting for media to ack.\n");
 	}
-
-	/* WaRsForcewakeWaitTC0:vlv */
-	if (!IS_CHERRYVIEW(dev_priv->dev))
-		__gen6_gt_wait_for_thread_c0(dev_priv);
 }
 
 static void __vlv_force_wake_put(struct drm_i915_private *dev_priv,
@@ -299,6 +288,154 @@ static void vlv_force_wake_put(struct drm_i915_private *dev_priv, int fw_engine)
 	spin_unlock_irqrestore(&dev_priv->uncore.lock, irqflags);
 }
 
+static void __gen9_gt_force_wake_mt_reset(struct drm_i915_private *dev_priv)
+{
+	__raw_i915_write32(dev_priv, FORCEWAKE_RENDER_GEN9,
+			_MASKED_BIT_DISABLE(0xffff));
+
+	__raw_i915_write32(dev_priv, FORCEWAKE_MEDIA_GEN9,
+			_MASKED_BIT_DISABLE(0xffff));
+
+	__raw_i915_write32(dev_priv, FORCEWAKE_BLITTER_GEN9,
+			_MASKED_BIT_DISABLE(0xffff));
+}
+
+static void
+__gen9_force_wake_get(struct drm_i915_private *dev_priv, int fw_engine)
+{
+	/* Check for Render Engine */
+	if (FORCEWAKE_RENDER & fw_engine) {
+		if (wait_for_atomic((__raw_i915_read32(dev_priv,
+						FORCEWAKE_ACK_RENDER_GEN9) &
+						FORCEWAKE_KERNEL) == 0,
+					FORCEWAKE_ACK_TIMEOUT_MS))
+			DRM_ERROR("Timed out: Render forcewake old ack to clear.\n");
+
+		__raw_i915_write32(dev_priv, FORCEWAKE_RENDER_GEN9,
+				   _MASKED_BIT_ENABLE(FORCEWAKE_KERNEL));
+
+		if (wait_for_atomic((__raw_i915_read32(dev_priv,
+						FORCEWAKE_ACK_RENDER_GEN9) &
+						FORCEWAKE_KERNEL),
+					FORCEWAKE_ACK_TIMEOUT_MS))
+			DRM_ERROR("Timed out: waiting for Render to ack.\n");
+	}
+
+	/* Check for Media Engine */
+	if (FORCEWAKE_MEDIA & fw_engine) {
+		if (wait_for_atomic((__raw_i915_read32(dev_priv,
+						FORCEWAKE_ACK_MEDIA_GEN9) &
+						FORCEWAKE_KERNEL) == 0,
+					FORCEWAKE_ACK_TIMEOUT_MS))
+			DRM_ERROR("Timed out: Media forcewake old ack to clear.\n");
+
+		__raw_i915_write32(dev_priv, FORCEWAKE_MEDIA_GEN9,
+				   _MASKED_BIT_ENABLE(FORCEWAKE_KERNEL));
+
+		if (wait_for_atomic((__raw_i915_read32(dev_priv,
+						FORCEWAKE_ACK_MEDIA_GEN9) &
+						FORCEWAKE_KERNEL),
+					FORCEWAKE_ACK_TIMEOUT_MS))
+			DRM_ERROR("Timed out: waiting for Media to ack.\n");
+	}
+
+	/* Check for Blitter Engine */
+	if (FORCEWAKE_BLITTER & fw_engine) {
+		if (wait_for_atomic((__raw_i915_read32(dev_priv,
+						FORCEWAKE_ACK_BLITTER_GEN9) &
+						FORCEWAKE_KERNEL) == 0,
+					FORCEWAKE_ACK_TIMEOUT_MS))
+			DRM_ERROR("Timed out: Blitter forcewake old ack to clear.\n");
+
+		__raw_i915_write32(dev_priv, FORCEWAKE_BLITTER_GEN9,
+				   _MASKED_BIT_ENABLE(FORCEWAKE_KERNEL));
+
+		if (wait_for_atomic((__raw_i915_read32(dev_priv,
+						FORCEWAKE_ACK_BLITTER_GEN9) &
+						FORCEWAKE_KERNEL),
+					FORCEWAKE_ACK_TIMEOUT_MS))
+			DRM_ERROR("Timed out: waiting for Blitter to ack.\n");
+	}
+}
+
+static void
+__gen9_force_wake_put(struct drm_i915_private *dev_priv, int fw_engine)
+{
+	/* Check for Render Engine */
+	if (FORCEWAKE_RENDER & fw_engine)
+		__raw_i915_write32(dev_priv, FORCEWAKE_RENDER_GEN9,
+				_MASKED_BIT_DISABLE(FORCEWAKE_KERNEL));
+
+	/* Check for Media Engine */
+	if (FORCEWAKE_MEDIA & fw_engine)
+		__raw_i915_write32(dev_priv, FORCEWAKE_MEDIA_GEN9,
+				_MASKED_BIT_DISABLE(FORCEWAKE_KERNEL));
+
+	/* Check for Blitter Engine */
+	if (FORCEWAKE_BLITTER & fw_engine)
+		__raw_i915_write32(dev_priv, FORCEWAKE_BLITTER_GEN9,
+				_MASKED_BIT_DISABLE(FORCEWAKE_KERNEL));
+}
+
+static void
+gen9_force_wake_get(struct drm_i915_private *dev_priv, int fw_engine)
+{
+	unsigned long irqflags;
+
+	spin_lock_irqsave(&dev_priv->uncore.lock, irqflags);
+
+	if (FORCEWAKE_RENDER & fw_engine) {
+		if (dev_priv->uncore.fw_rendercount++ == 0)
+			dev_priv->uncore.funcs.force_wake_get(dev_priv,
+							FORCEWAKE_RENDER);
+	}
+
+	if (FORCEWAKE_MEDIA & fw_engine) {
+		if (dev_priv->uncore.fw_mediacount++ == 0)
+			dev_priv->uncore.funcs.force_wake_get(dev_priv,
+							FORCEWAKE_MEDIA);
+	}
+
+	if (FORCEWAKE_BLITTER & fw_engine) {
+		if (dev_priv->uncore.fw_blittercount++ == 0)
+			dev_priv->uncore.funcs.force_wake_get(dev_priv,
+							FORCEWAKE_BLITTER);
+	}
+
+	spin_unlock_irqrestore(&dev_priv->uncore.lock, irqflags);
+}
+
+static void
+gen9_force_wake_put(struct drm_i915_private *dev_priv, int fw_engine)
+{
+	unsigned long irqflags;
+
+	spin_lock_irqsave(&dev_priv->uncore.lock, irqflags);
+
+	if (FORCEWAKE_RENDER & fw_engine) {
+		WARN_ON(dev_priv->uncore.fw_rendercount == 0);
+		if (--dev_priv->uncore.fw_rendercount == 0)
+			dev_priv->uncore.funcs.force_wake_put(dev_priv,
+							FORCEWAKE_RENDER);
+	}
+
+	if (FORCEWAKE_MEDIA & fw_engine) {
+		WARN_ON(dev_priv->uncore.fw_mediacount == 0);
+		if (--dev_priv->uncore.fw_mediacount == 0)
+			dev_priv->uncore.funcs.force_wake_put(dev_priv,
+							FORCEWAKE_MEDIA);
+	}
+
+	if (FORCEWAKE_BLITTER & fw_engine) {
+		WARN_ON(dev_priv->uncore.fw_blittercount == 0);
+		if (--dev_priv->uncore.fw_blittercount == 0)
+			dev_priv->uncore.funcs.force_wake_put(dev_priv,
+							FORCEWAKE_BLITTER);
+	}
+
+	spin_unlock_irqrestore(&dev_priv->uncore.lock, irqflags);
+}
+
 static void gen6_force_wake_timer(unsigned long arg)
 {
 	struct drm_i915_private *dev_priv = (void *)arg;
@@ -334,9 +471,12 @@ void intel_uncore_forcewake_reset(struct drm_device *dev, bool restore)
 	else if (IS_GEN6(dev) || IS_GEN7(dev))
 		__gen6_gt_force_wake_reset(dev_priv);
 
-	if (IS_IVYBRIDGE(dev) || IS_HASWELL(dev) || IS_GEN8(dev))
+	if (IS_IVYBRIDGE(dev) || IS_HASWELL(dev) || IS_BROADWELL(dev))
 		__gen7_gt_force_wake_mt_reset(dev_priv);
 
+	if (IS_GEN9(dev))
+		__gen9_gt_force_wake_mt_reset(dev_priv);
+
 	if (restore) { /* If reset with a user forcewake, try to restore */
 		unsigned fw = 0;
 
@@ -346,6 +486,15 @@ void intel_uncore_forcewake_reset(struct drm_device *dev, bool restore)
 
 			if (dev_priv->uncore.fw_mediacount)
 				fw |= FORCEWAKE_MEDIA;
+		} else if (IS_GEN9(dev)) {
+			if (dev_priv->uncore.fw_rendercount)
+				fw |= FORCEWAKE_RENDER;
+
+			if (dev_priv->uncore.fw_mediacount)
+				fw |= FORCEWAKE_MEDIA;
+
+			if (dev_priv->uncore.fw_blittercount)
+				fw |= FORCEWAKE_BLITTER;
 		} else {
 			if (dev_priv->uncore.forcewake_count)
 				fw = FORCEWAKE_ALL;
@@ -363,7 +512,8 @@ void intel_uncore_forcewake_reset(struct drm_device *dev, bool restore)
 	spin_unlock_irqrestore(&dev_priv->uncore.lock, irqflags);
 }
 
-void intel_uncore_early_sanitize(struct drm_device *dev, bool restore_forcewake)
+static void __intel_uncore_early_sanitize(struct drm_device *dev,
+					  bool restore_forcewake)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 
@@ -389,6 +539,12 @@ void intel_uncore_early_sanitize(struct drm_device *dev, bool restore_forcewake)
 	intel_uncore_forcewake_reset(dev, restore_forcewake);
 }
 
+void intel_uncore_early_sanitize(struct drm_device *dev, bool restore_forcewake)
+{
+	__intel_uncore_early_sanitize(dev, restore_forcewake);
+	i915_check_and_clear_faults(dev);
+}
+
 void intel_uncore_sanitize(struct drm_device *dev)
 {
 	/* BIOS often leaves RC6 enabled, but disable it for hw init */
@@ -410,6 +566,10 @@ void gen6_gt_force_wake_get(struct drm_i915_private *dev_priv, int fw_engine)
 
 	intel_runtime_pm_get(dev_priv);
 
+	/* Redirect to Gen9 specific routine */
+	if (IS_GEN9(dev_priv->dev))
+		return gen9_force_wake_get(dev_priv, fw_engine);
+
 	/* Redirect to VLV specific routine */
 	if (IS_VALLEYVIEW(dev_priv->dev))
 		return vlv_force_wake_get(dev_priv, fw_engine);
@@ -431,6 +591,12 @@ void gen6_gt_force_wake_put(struct drm_i915_private *dev_priv, int fw_engine)
 	if (!dev_priv->uncore.funcs.force_wake_put)
 		return;
 
+	/* Redirect to Gen9 specific routine */
+	if (IS_GEN9(dev_priv->dev)) {
+		gen9_force_wake_put(dev_priv, fw_engine);
+		goto out;
+	}
+
 	/* Redirect to VLV specific routine */
 	if (IS_VALLEYVIEW(dev_priv->dev)) {
 		vlv_force_wake_put(dev_priv, fw_engine);
@@ -504,6 +670,38 @@ void assert_force_wake_inactive(struct drm_i915_private *dev_priv)
 	 REG_RANGE((reg), 0x14000, 0x14400) || \
 	 REG_RANGE((reg), 0x22000, 0x24000))
 
+#define FORCEWAKE_GEN9_UNCORE_RANGE_OFFSET(reg) \
+	REG_RANGE((reg), 0xB00,  0x2000)
+
+#define FORCEWAKE_GEN9_RENDER_RANGE_OFFSET(reg) \
+	(REG_RANGE((reg), 0x2000, 0x2700) || \
+	 REG_RANGE((reg), 0x3000, 0x4000) || \
+	 REG_RANGE((reg), 0x5200, 0x8000) || \
+	 REG_RANGE((reg), 0x8140, 0x8160) || \
+	 REG_RANGE((reg), 0x8300, 0x8500) || \
+	 REG_RANGE((reg), 0x8C00, 0x8D00) || \
+	 REG_RANGE((reg), 0xB000, 0xB480) || \
+	 REG_RANGE((reg), 0xE000, 0xE900) || \
+	 REG_RANGE((reg), 0x24400, 0x24800))
+
+#define FORCEWAKE_GEN9_MEDIA_RANGE_OFFSET(reg) \
+	(REG_RANGE((reg), 0x8130, 0x8140) || \
+	 REG_RANGE((reg), 0x8800, 0x8A00) || \
+	 REG_RANGE((reg), 0xD000, 0xD800) || \
+	 REG_RANGE((reg), 0x12000, 0x14000) || \
+	 REG_RANGE((reg), 0x1A000, 0x1EA00) || \
+	 REG_RANGE((reg), 0x30000, 0x40000))
+
+#define FORCEWAKE_GEN9_COMMON_RANGE_OFFSET(reg) \
+	REG_RANGE((reg), 0x9400, 0x9800)
+
+#define FORCEWAKE_GEN9_BLITTER_RANGE_OFFSET(reg) \
+	((reg) < 0x40000 &&\
+	 !FORCEWAKE_GEN9_UNCORE_RANGE_OFFSET(reg) && \
+	 !FORCEWAKE_GEN9_RENDER_RANGE_OFFSET(reg) && \
+	 !FORCEWAKE_GEN9_MEDIA_RANGE_OFFSET(reg) && \
+	 !FORCEWAKE_GEN9_COMMON_RANGE_OFFSET(reg))
+
 static void
 ilk_dummy_write(struct drm_i915_private *dev_priv)
 {
@@ -634,6 +832,45 @@ chv_read##x(struct drm_i915_private *dev_priv, off_t reg, bool trace) { \
 	REG_READ_FOOTER; \
 }
 
+#define SKL_NEEDS_FORCE_WAKE(dev_priv, reg)	\
+	 ((reg) < 0x40000 && !FORCEWAKE_GEN9_UNCORE_RANGE_OFFSET(reg))
+
+#define __gen9_read(x) \
+static u##x \
+gen9_read##x(struct drm_i915_private *dev_priv, off_t reg, bool trace) { \
+	REG_READ_HEADER(x); \
+	if (!SKL_NEEDS_FORCE_WAKE((dev_priv), (reg))) { \
+		val = __raw_i915_read##x(dev_priv, reg); \
+	} else { \
+		unsigned fwengine = 0; \
+		if (FORCEWAKE_GEN9_RENDER_RANGE_OFFSET(reg)) { \
+			if (dev_priv->uncore.fw_rendercount == 0) \
+				fwengine = FORCEWAKE_RENDER; \
+		} else if (FORCEWAKE_GEN9_MEDIA_RANGE_OFFSET(reg)) { \
+			if (dev_priv->uncore.fw_mediacount == 0) \
+				fwengine = FORCEWAKE_MEDIA; \
+		} else if (FORCEWAKE_GEN9_COMMON_RANGE_OFFSET(reg)) { \
+			if (dev_priv->uncore.fw_rendercount == 0) \
+				fwengine |= FORCEWAKE_RENDER; \
+			if (dev_priv->uncore.fw_mediacount == 0) \
+				fwengine |= FORCEWAKE_MEDIA; \
+		} else { \
+			if (dev_priv->uncore.fw_blittercount == 0) \
+				fwengine = FORCEWAKE_BLITTER; \
+		} \
+		if (fwengine) \
+			dev_priv->uncore.funcs.force_wake_get(dev_priv, fwengine); \
+		val = __raw_i915_read##x(dev_priv, reg); \
+		if (fwengine) \
+			dev_priv->uncore.funcs.force_wake_put(dev_priv, fwengine); \
+	} \
+	REG_READ_FOOTER; \
+}
+
+__gen9_read(8)
+__gen9_read(16)
+__gen9_read(32)
+__gen9_read(64)
 __chv_read(8)
 __chv_read(16)
 __chv_read(32)
@@ -655,6 +892,7 @@ __gen4_read(16)
 __gen4_read(32)
 __gen4_read(64)
 
+#undef __gen9_read
 #undef __chv_read
 #undef __vlv_read
 #undef __gen6_read
@@ -792,6 +1030,69 @@ chv_write##x(struct drm_i915_private *dev_priv, off_t reg, u##x val, bool trace)
 	REG_WRITE_FOOTER; \
 }
 
+static const u32 gen9_shadowed_regs[] = {
+	RING_TAIL(RENDER_RING_BASE),
+	RING_TAIL(GEN6_BSD_RING_BASE),
+	RING_TAIL(VEBOX_RING_BASE),
+	RING_TAIL(BLT_RING_BASE),
+	FORCEWAKE_BLITTER_GEN9,
+	FORCEWAKE_RENDER_GEN9,
+	FORCEWAKE_MEDIA_GEN9,
+	GEN6_RPNSWREQ,
+	GEN6_RC_VIDEO_FREQ,
+	/* TODO: Other registers are not yet used */
+};
+
+static bool is_gen9_shadowed(struct drm_i915_private *dev_priv, u32 reg)
+{
+	int i;
+	for (i = 0; i < ARRAY_SIZE(gen9_shadowed_regs); i++)
+		if (reg == gen9_shadowed_regs[i])
+			return true;
+
+	return false;
+}
+
+#define __gen9_write(x) \
+static void \
+gen9_write##x(struct drm_i915_private *dev_priv, off_t reg, u##x val, \
+		bool trace) { \
+	REG_WRITE_HEADER; \
+	if (!SKL_NEEDS_FORCE_WAKE((dev_priv), (reg)) || \
+			is_gen9_shadowed(dev_priv, reg)) { \
+		__raw_i915_write##x(dev_priv, reg, val); \
+	} else { \
+		unsigned fwengine = 0; \
+		if (FORCEWAKE_GEN9_RENDER_RANGE_OFFSET(reg)) { \
+			if (dev_priv->uncore.fw_rendercount == 0) \
+				fwengine = FORCEWAKE_RENDER; \
+		} else if (FORCEWAKE_GEN9_MEDIA_RANGE_OFFSET(reg)) { \
+			if (dev_priv->uncore.fw_mediacount == 0) \
+				fwengine = FORCEWAKE_MEDIA; \
+		} else if (FORCEWAKE_GEN9_COMMON_RANGE_OFFSET(reg)) { \
+			if (dev_priv->uncore.fw_rendercount == 0) \
+				fwengine |= FORCEWAKE_RENDER; \
+			if (dev_priv->uncore.fw_mediacount == 0) \
+				fwengine |= FORCEWAKE_MEDIA; \
+		} else { \
+			if (dev_priv->uncore.fw_blittercount == 0) \
+				fwengine = FORCEWAKE_BLITTER; \
+		} \
+		if (fwengine) \
+			dev_priv->uncore.funcs.force_wake_get(dev_priv, \
+					fwengine); \
+		__raw_i915_write##x(dev_priv, reg, val); \
+		if (fwengine) \
+			dev_priv->uncore.funcs.force_wake_put(dev_priv, \
+					fwengine); \
+	} \
+	REG_WRITE_FOOTER; \
+}
+
+__gen9_write(8)
+__gen9_write(16)
+__gen9_write(32)
+__gen9_write(64)
 __chv_write(8)
 __chv_write(16)
 __chv_write(32)
@@ -817,6 +1118,7 @@ __gen4_write(16)
 __gen4_write(32)
 __gen4_write(64)
 
+#undef __gen9_write
 #undef __chv_write
 #undef __gen8_write
 #undef __hsw_write
@@ -826,6 +1128,22 @@ __gen4_write(64)
 #undef REG_WRITE_FOOTER
 #undef REG_WRITE_HEADER
 
+#define ASSIGN_WRITE_MMIO_VFUNCS(x) \
+do { \
+	dev_priv->uncore.funcs.mmio_writeb = x##_write8; \
+	dev_priv->uncore.funcs.mmio_writew = x##_write16; \
+	dev_priv->uncore.funcs.mmio_writel = x##_write32; \
+	dev_priv->uncore.funcs.mmio_writeq = x##_write64; \
+} while (0)
+
+#define ASSIGN_READ_MMIO_VFUNCS(x) \
+do { \
+	dev_priv->uncore.funcs.mmio_readb = x##_read8; \
+	dev_priv->uncore.funcs.mmio_readw = x##_read16; \
+	dev_priv->uncore.funcs.mmio_readl = x##_read32; \
+	dev_priv->uncore.funcs.mmio_readq = x##_read64; \
+} while (0)
+
 void intel_uncore_init(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
@@ -833,12 +1151,15 @@ void intel_uncore_init(struct drm_device *dev)
 	setup_timer(&dev_priv->uncore.force_wake_timer,
 		    gen6_force_wake_timer, (unsigned long)dev_priv);
 
-	intel_uncore_early_sanitize(dev, false);
+	__intel_uncore_early_sanitize(dev, false);
 
-	if (IS_VALLEYVIEW(dev)) {
+	if (IS_GEN9(dev)) {
+		dev_priv->uncore.funcs.force_wake_get = __gen9_force_wake_get;
+		dev_priv->uncore.funcs.force_wake_put = __gen9_force_wake_put;
+	} else if (IS_VALLEYVIEW(dev)) {
 		dev_priv->uncore.funcs.force_wake_get = __vlv_force_wake_get;
 		dev_priv->uncore.funcs.force_wake_put = __vlv_force_wake_put;
-	} else if (IS_HASWELL(dev) || IS_GEN8(dev)) {
+	} else if (IS_HASWELL(dev) || IS_BROADWELL(dev)) {
 		dev_priv->uncore.funcs.force_wake_get = __gen7_gt_force_wake_mt_get;
 		dev_priv->uncore.funcs.force_wake_put = __gen7_gt_force_wake_mt_put;
 	} else if (IS_IVYBRIDGE(dev)) {
@@ -881,77 +1202,52 @@ void intel_uncore_init(struct drm_device *dev)
 
 	switch (INTEL_INFO(dev)->gen) {
 	default:
+		WARN_ON(1);
+		return;
+	case 9:
+		ASSIGN_WRITE_MMIO_VFUNCS(gen9);
+		ASSIGN_READ_MMIO_VFUNCS(gen9);
+		break;
+	case 8:
 		if (IS_CHERRYVIEW(dev)) {
-			dev_priv->uncore.funcs.mmio_writeb  = chv_write8;
-			dev_priv->uncore.funcs.mmio_writew  = chv_write16;
-			dev_priv->uncore.funcs.mmio_writel  = chv_write32;
-			dev_priv->uncore.funcs.mmio_writeq  = chv_write64;
-			dev_priv->uncore.funcs.mmio_readb  = chv_read8;
-			dev_priv->uncore.funcs.mmio_readw  = chv_read16;
-			dev_priv->uncore.funcs.mmio_readl  = chv_read32;
-			dev_priv->uncore.funcs.mmio_readq  = chv_read64;
+			ASSIGN_WRITE_MMIO_VFUNCS(chv);
+			ASSIGN_READ_MMIO_VFUNCS(chv);
 
 		} else {
-			dev_priv->uncore.funcs.mmio_writeb  = gen8_write8;
-			dev_priv->uncore.funcs.mmio_writew  = gen8_write16;
-			dev_priv->uncore.funcs.mmio_writel  = gen8_write32;
-			dev_priv->uncore.funcs.mmio_writeq  = gen8_write64;
-			dev_priv->uncore.funcs.mmio_readb  = gen6_read8;
-			dev_priv->uncore.funcs.mmio_readw  = gen6_read16;
-			dev_priv->uncore.funcs.mmio_readl  = gen6_read32;
-			dev_priv->uncore.funcs.mmio_readq  = gen6_read64;
+			ASSIGN_WRITE_MMIO_VFUNCS(gen8);
+			ASSIGN_READ_MMIO_VFUNCS(gen6);
 		}
 		break;
 	case 7:
 	case 6:
 		if (IS_HASWELL(dev)) {
-			dev_priv->uncore.funcs.mmio_writeb  = hsw_write8;
-			dev_priv->uncore.funcs.mmio_writew  = hsw_write16;
-			dev_priv->uncore.funcs.mmio_writel  = hsw_write32;
-			dev_priv->uncore.funcs.mmio_writeq  = hsw_write64;
+			ASSIGN_WRITE_MMIO_VFUNCS(hsw);
 		} else {
-			dev_priv->uncore.funcs.mmio_writeb  = gen6_write8;
-			dev_priv->uncore.funcs.mmio_writew  = gen6_write16;
-			dev_priv->uncore.funcs.mmio_writel  = gen6_write32;
-			dev_priv->uncore.funcs.mmio_writeq  = gen6_write64;
+			ASSIGN_WRITE_MMIO_VFUNCS(gen6);
 		}
 
 		if (IS_VALLEYVIEW(dev)) {
-			dev_priv->uncore.funcs.mmio_readb  = vlv_read8;
-			dev_priv->uncore.funcs.mmio_readw  = vlv_read16;
-			dev_priv->uncore.funcs.mmio_readl  = vlv_read32;
-			dev_priv->uncore.funcs.mmio_readq  = vlv_read64;
+			ASSIGN_READ_MMIO_VFUNCS(vlv);
 		} else {
-			dev_priv->uncore.funcs.mmio_readb  = gen6_read8;
-			dev_priv->uncore.funcs.mmio_readw  = gen6_read16;
-			dev_priv->uncore.funcs.mmio_readl  = gen6_read32;
-			dev_priv->uncore.funcs.mmio_readq  = gen6_read64;
+			ASSIGN_READ_MMIO_VFUNCS(gen6);
 		}
 		break;
 	case 5:
-		dev_priv->uncore.funcs.mmio_writeb  = gen5_write8;
-		dev_priv->uncore.funcs.mmio_writew  = gen5_write16;
-		dev_priv->uncore.funcs.mmio_writel  = gen5_write32;
-		dev_priv->uncore.funcs.mmio_writeq  = gen5_write64;
-		dev_priv->uncore.funcs.mmio_readb  = gen5_read8;
-		dev_priv->uncore.funcs.mmio_readw  = gen5_read16;
-		dev_priv->uncore.funcs.mmio_readl  = gen5_read32;
-		dev_priv->uncore.funcs.mmio_readq  = gen5_read64;
+		ASSIGN_WRITE_MMIO_VFUNCS(gen5);
+		ASSIGN_READ_MMIO_VFUNCS(gen5);
 		break;
 	case 4:
 	case 3:
 	case 2:
-		dev_priv->uncore.funcs.mmio_writeb  = gen4_write8;
-		dev_priv->uncore.funcs.mmio_writew  = gen4_write16;
-		dev_priv->uncore.funcs.mmio_writel  = gen4_write32;
-		dev_priv->uncore.funcs.mmio_writeq  = gen4_write64;
-		dev_priv->uncore.funcs.mmio_readb  = gen4_read8;
-		dev_priv->uncore.funcs.mmio_readw  = gen4_read16;
-		dev_priv->uncore.funcs.mmio_readl  = gen4_read32;
-		dev_priv->uncore.funcs.mmio_readq  = gen4_read64;
+		ASSIGN_WRITE_MMIO_VFUNCS(gen4);
+		ASSIGN_READ_MMIO_VFUNCS(gen4);
 		break;
 	}
+
+	i915_check_and_clear_faults(dev);
 }
+#undef ASSIGN_WRITE_MMIO_VFUNCS
+#undef ASSIGN_READ_MMIO_VFUNCS
 
 void intel_uncore_fini(struct drm_device *dev)
 {
@@ -968,7 +1264,7 @@ static const struct register_whitelist {
 	/* supported gens, 0x10 for 4, 0x30 for 4 and 5, etc. */
 	uint32_t gen_bitmask;
 } whitelist[] = {
-	{ RING_TIMESTAMP(RENDER_RING_BASE), 8, GEN_RANGE(4, 8) },
+	{ RING_TIMESTAMP(RENDER_RING_BASE), 8, GEN_RANGE(4, 9) },
 };
 
 int i915_reg_read_ioctl(struct drm_device *dev,
@@ -1053,41 +1349,34 @@ int i915_get_reset_stats_ioctl(struct drm_device *dev,
 	return 0;
 }
 
-static int i965_reset_complete(struct drm_device *dev)
+static int i915_reset_complete(struct drm_device *dev)
 {
 	u8 gdrst;
-	pci_read_config_byte(dev->pdev, I965_GDRST, &gdrst);
-	return (gdrst & GRDOM_RESET_ENABLE) == 0;
+	pci_read_config_byte(dev->pdev, I915_GDRST, &gdrst);
+	return (gdrst & GRDOM_RESET_STATUS) == 0;
 }
 
-static int i965_do_reset(struct drm_device *dev)
+static int i915_do_reset(struct drm_device *dev)
 {
-	int ret;
-
-	/* FIXME: i965g/gm need a display save/restore for gpu reset. */
-	return -ENODEV;
+	/* assert reset for at least 20 usec */
+	pci_write_config_byte(dev->pdev, I915_GDRST, GRDOM_RESET_ENABLE);
+	udelay(20);
+	pci_write_config_byte(dev->pdev, I915_GDRST, 0);
 
-	/*
-	 * Set the domains we want to reset (GRDOM/bits 2 and 3) as
-	 * well as the reset bit (GR/bit 0).  Setting the GR bit
-	 * triggers the reset; when done, the hardware will clear it.
-	 */
-	pci_write_config_byte(dev->pdev, I965_GDRST,
-			      GRDOM_RENDER | GRDOM_RESET_ENABLE);
-	ret =  wait_for(i965_reset_complete(dev), 500);
-	if (ret)
-		return ret;
-
-	pci_write_config_byte(dev->pdev, I965_GDRST,
-			      GRDOM_MEDIA | GRDOM_RESET_ENABLE);
-
-	ret =  wait_for(i965_reset_complete(dev), 500);
-	if (ret)
-		return ret;
+	return wait_for(i915_reset_complete(dev), 500);
+}
 
-	pci_write_config_byte(dev->pdev, I965_GDRST, 0);
+static int g4x_reset_complete(struct drm_device *dev)
+{
+	u8 gdrst;
+	pci_read_config_byte(dev->pdev, I915_GDRST, &gdrst);
+	return (gdrst & GRDOM_RESET_ENABLE) == 0;
+}
 
-	return 0;
+static int g33_do_reset(struct drm_device *dev)
+{
+	pci_write_config_byte(dev->pdev, I915_GDRST, GRDOM_RESET_ENABLE);
+	return wait_for(g4x_reset_complete(dev), 500);
 }
 
 static int g4x_do_reset(struct drm_device *dev)
@@ -1095,9 +1384,9 @@ static int g4x_do_reset(struct drm_device *dev)
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	int ret;
 
-	pci_write_config_byte(dev->pdev, I965_GDRST,
+	pci_write_config_byte(dev->pdev, I915_GDRST,
 			      GRDOM_RENDER | GRDOM_RESET_ENABLE);
-	ret =  wait_for(i965_reset_complete(dev), 500);
+	ret =  wait_for(g4x_reset_complete(dev), 500);
 	if (ret)
 		return ret;
 
@@ -1105,9 +1394,9 @@ static int g4x_do_reset(struct drm_device *dev)
 	I915_WRITE(VDECCLK_GATE_D, I915_READ(VDECCLK_GATE_D) | VCP_UNIT_CLOCK_GATE_DISABLE);
 	POSTING_READ(VDECCLK_GATE_D);
 
-	pci_write_config_byte(dev->pdev, I965_GDRST,
+	pci_write_config_byte(dev->pdev, I915_GDRST,
 			      GRDOM_MEDIA | GRDOM_RESET_ENABLE);
-	ret =  wait_for(i965_reset_complete(dev), 500);
+	ret =  wait_for(g4x_reset_complete(dev), 500);
 	if (ret)
 		return ret;
 
@@ -1115,7 +1404,7 @@ static int g4x_do_reset(struct drm_device *dev)
 	I915_WRITE(VDECCLK_GATE_D, I915_READ(VDECCLK_GATE_D) & ~VCP_UNIT_CLOCK_GATE_DISABLE);
 	POSTING_READ(VDECCLK_GATE_D);
 
-	pci_write_config_byte(dev->pdev, I965_GDRST, 0);
+	pci_write_config_byte(dev->pdev, I915_GDRST, 0);
 
 	return 0;
 }
@@ -1173,8 +1462,10 @@ int intel_gpu_reset(struct drm_device *dev)
 		return ironlake_do_reset(dev);
 	else if (IS_G4X(dev))
 		return g4x_do_reset(dev);
-	else if (IS_GEN4(dev))
-		return i965_do_reset(dev);
+	else if (IS_G33(dev))
+		return g33_do_reset(dev);
+	else if (INTEL_INFO(dev)->gen >= 3)
+		return i915_do_reset(dev);
 	else
 		return -ENODEV;
 }
diff --git a/drivers/gpu/drm/imx/Kconfig b/drivers/gpu/drm/imx/Kconfig
new file mode 100644
index 000000000000..ab31848e92cf
--- /dev/null
+++ b/drivers/gpu/drm/imx/Kconfig
@@ -0,0 +1,54 @@
+config DRM_IMX
+	tristate "DRM Support for Freescale i.MX"
+	select DRM_KMS_HELPER
+	select DRM_KMS_FB_HELPER
+	select VIDEOMODE_HELPERS
+	select DRM_GEM_CMA_HELPER
+	select DRM_KMS_CMA_HELPER
+	depends on DRM && (ARCH_MXC || ARCH_MULTIPLATFORM)
+	depends on IMX_IPUV3_CORE
+	help
+	  enable i.MX graphics support
+
+config DRM_IMX_FB_HELPER
+	tristate "provide legacy framebuffer /dev/fb0"
+	select DRM_KMS_CMA_HELPER
+	depends on DRM_IMX
+	help
+	  The DRM framework can provide a legacy /dev/fb0 framebuffer
+	  for your device. This is necessary to get a framebuffer console
+	  and also for applications using the legacy framebuffer API
+
+config DRM_IMX_PARALLEL_DISPLAY
+	tristate "Support for parallel displays"
+	select DRM_PANEL
+	depends on DRM_IMX
+	select VIDEOMODE_HELPERS
+
+config DRM_IMX_TVE
+	tristate "Support for TV and VGA displays"
+	depends on DRM_IMX
+	select REGMAP_MMIO
+	help
+	  Choose this to enable the internal Television Encoder (TVe)
+	  found on i.MX53 processors.
+
+config DRM_IMX_LDB
+	tristate "Support for LVDS displays"
+	depends on DRM_IMX && MFD_SYSCON
+	help
+	  Choose this to enable the internal LVDS Display Bridge (LDB)
+	  found on i.MX53 and i.MX6 processors.
+
+config DRM_IMX_IPUV3
+	tristate
+	depends on DRM_IMX
+	depends on IMX_IPUV3_CORE
+	default y if DRM_IMX=y
+	default m if DRM_IMX=m
+
+config DRM_IMX_HDMI
+	tristate "Freescale i.MX DRM HDMI"
+	depends on DRM_IMX
+	help
+	  Choose this if you want to use HDMI on i.MX6.
diff --git a/drivers/gpu/drm/imx/Makefile b/drivers/gpu/drm/imx/Makefile
new file mode 100644
index 000000000000..582c438d8cbd
--- /dev/null
+++ b/drivers/gpu/drm/imx/Makefile
@@ -0,0 +1,12 @@
+
+imxdrm-objs := imx-drm-core.o
+
+obj-$(CONFIG_DRM_IMX) += imxdrm.o
+
+obj-$(CONFIG_DRM_IMX_PARALLEL_DISPLAY) += parallel-display.o
+obj-$(CONFIG_DRM_IMX_TVE) += imx-tve.o
+obj-$(CONFIG_DRM_IMX_LDB) += imx-ldb.o
+
+imx-ipuv3-crtc-objs  := ipuv3-crtc.o ipuv3-plane.o
+obj-$(CONFIG_DRM_IMX_IPUV3)	+= imx-ipuv3-crtc.o
+obj-$(CONFIG_DRM_IMX_HDMI) += imx-hdmi.o
diff --git a/drivers/gpu/drm/imx/imx-drm-core.c b/drivers/gpu/drm/imx/imx-drm-core.c
new file mode 100644
index 000000000000..b250130debc8
--- /dev/null
+++ b/drivers/gpu/drm/imx/imx-drm-core.c
@@ -0,0 +1,703 @@
+/*
+ * Freescale i.MX drm driver
+ *
+ * Copyright (C) 2011 Sascha Hauer, Pengutronix
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License
+ * as published by the Free Software Foundation; either version 2
+ * of the License, or (at your option) any later version.
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ */
+#include <linux/component.h>
+#include <linux/device.h>
+#include <linux/fb.h>
+#include <linux/module.h>
+#include <linux/of_graph.h>
+#include <linux/platform_device.h>
+#include <drm/drmP.h>
+#include <drm/drm_fb_helper.h>
+#include <drm/drm_crtc_helper.h>
+#include <drm/drm_gem_cma_helper.h>
+#include <drm/drm_fb_cma_helper.h>
+#include <drm/drm_plane_helper.h>
+
+#include "imx-drm.h"
+
+#define MAX_CRTC	4
+
+struct imx_drm_component {
+	struct device_node *of_node;
+	struct list_head list;
+};
+
+struct imx_drm_device {
+	struct drm_device			*drm;
+	struct imx_drm_crtc			*crtc[MAX_CRTC];
+	int					pipes;
+	struct drm_fbdev_cma			*fbhelper;
+};
+
+struct imx_drm_crtc {
+	struct drm_crtc				*crtc;
+	int					pipe;
+	struct imx_drm_crtc_helper_funcs	imx_drm_helper_funcs;
+	struct device_node			*port;
+};
+
+static int legacyfb_depth = 16;
+module_param(legacyfb_depth, int, 0444);
+
+int imx_drm_crtc_id(struct imx_drm_crtc *crtc)
+{
+	return crtc->pipe;
+}
+EXPORT_SYMBOL_GPL(imx_drm_crtc_id);
+
+static void imx_drm_driver_lastclose(struct drm_device *drm)
+{
+#if IS_ENABLED(CONFIG_DRM_IMX_FB_HELPER)
+	struct imx_drm_device *imxdrm = drm->dev_private;
+
+	if (imxdrm->fbhelper)
+		drm_fbdev_cma_restore_mode(imxdrm->fbhelper);
+#endif
+}
+
+static int imx_drm_driver_unload(struct drm_device *drm)
+{
+#if IS_ENABLED(CONFIG_DRM_IMX_FB_HELPER)
+	struct imx_drm_device *imxdrm = drm->dev_private;
+#endif
+
+	drm_kms_helper_poll_fini(drm);
+
+#if IS_ENABLED(CONFIG_DRM_IMX_FB_HELPER)
+	if (imxdrm->fbhelper)
+		drm_fbdev_cma_fini(imxdrm->fbhelper);
+#endif
+
+	component_unbind_all(drm->dev, drm);
+
+	drm_vblank_cleanup(drm);
+	drm_mode_config_cleanup(drm);
+
+	platform_set_drvdata(drm->platformdev, NULL);
+
+	return 0;
+}
+
+static struct imx_drm_crtc *imx_drm_find_crtc(struct drm_crtc *crtc)
+{
+	struct imx_drm_device *imxdrm = crtc->dev->dev_private;
+	unsigned i;
+
+	for (i = 0; i < MAX_CRTC; i++)
+		if (imxdrm->crtc[i] && imxdrm->crtc[i]->crtc == crtc)
+			return imxdrm->crtc[i];
+
+	return NULL;
+}
+
+int imx_drm_panel_format_pins(struct drm_encoder *encoder,
+		u32 interface_pix_fmt, int hsync_pin, int vsync_pin)
+{
+	struct imx_drm_crtc_helper_funcs *helper;
+	struct imx_drm_crtc *imx_crtc;
+
+	imx_crtc = imx_drm_find_crtc(encoder->crtc);
+	if (!imx_crtc)
+		return -EINVAL;
+
+	helper = &imx_crtc->imx_drm_helper_funcs;
+	if (helper->set_interface_pix_fmt)
+		return helper->set_interface_pix_fmt(encoder->crtc,
+				encoder->encoder_type, interface_pix_fmt,
+				hsync_pin, vsync_pin);
+	return 0;
+}
+EXPORT_SYMBOL_GPL(imx_drm_panel_format_pins);
+
+int imx_drm_panel_format(struct drm_encoder *encoder, u32 interface_pix_fmt)
+{
+	return imx_drm_panel_format_pins(encoder, interface_pix_fmt, 2, 3);
+}
+EXPORT_SYMBOL_GPL(imx_drm_panel_format);
+
+int imx_drm_crtc_vblank_get(struct imx_drm_crtc *imx_drm_crtc)
+{
+	return drm_vblank_get(imx_drm_crtc->crtc->dev, imx_drm_crtc->pipe);
+}
+EXPORT_SYMBOL_GPL(imx_drm_crtc_vblank_get);
+
+void imx_drm_crtc_vblank_put(struct imx_drm_crtc *imx_drm_crtc)
+{
+	drm_vblank_put(imx_drm_crtc->crtc->dev, imx_drm_crtc->pipe);
+}
+EXPORT_SYMBOL_GPL(imx_drm_crtc_vblank_put);
+
+void imx_drm_handle_vblank(struct imx_drm_crtc *imx_drm_crtc)
+{
+	drm_handle_vblank(imx_drm_crtc->crtc->dev, imx_drm_crtc->pipe);
+}
+EXPORT_SYMBOL_GPL(imx_drm_handle_vblank);
+
+static int imx_drm_enable_vblank(struct drm_device *drm, int crtc)
+{
+	struct imx_drm_device *imxdrm = drm->dev_private;
+	struct imx_drm_crtc *imx_drm_crtc = imxdrm->crtc[crtc];
+	int ret;
+
+	if (!imx_drm_crtc)
+		return -EINVAL;
+
+	if (!imx_drm_crtc->imx_drm_helper_funcs.enable_vblank)
+		return -ENOSYS;
+
+	ret = imx_drm_crtc->imx_drm_helper_funcs.enable_vblank(
+			imx_drm_crtc->crtc);
+
+	return ret;
+}
+
+static void imx_drm_disable_vblank(struct drm_device *drm, int crtc)
+{
+	struct imx_drm_device *imxdrm = drm->dev_private;
+	struct imx_drm_crtc *imx_drm_crtc = imxdrm->crtc[crtc];
+
+	if (!imx_drm_crtc)
+		return;
+
+	if (!imx_drm_crtc->imx_drm_helper_funcs.disable_vblank)
+		return;
+
+	imx_drm_crtc->imx_drm_helper_funcs.disable_vblank(imx_drm_crtc->crtc);
+}
+
+static void imx_drm_driver_preclose(struct drm_device *drm,
+		struct drm_file *file)
+{
+	int i;
+
+	if (!file->is_master)
+		return;
+
+	for (i = 0; i < MAX_CRTC; i++)
+		imx_drm_disable_vblank(drm, i);
+}
+
+static const struct file_operations imx_drm_driver_fops = {
+	.owner = THIS_MODULE,
+	.open = drm_open,
+	.release = drm_release,
+	.unlocked_ioctl = drm_ioctl,
+	.mmap = drm_gem_cma_mmap,
+	.poll = drm_poll,
+	.read = drm_read,
+	.llseek = noop_llseek,
+};
+
+void imx_drm_connector_destroy(struct drm_connector *connector)
+{
+	drm_connector_unregister(connector);
+	drm_connector_cleanup(connector);
+}
+EXPORT_SYMBOL_GPL(imx_drm_connector_destroy);
+
+void imx_drm_encoder_destroy(struct drm_encoder *encoder)
+{
+	drm_encoder_cleanup(encoder);
+}
+EXPORT_SYMBOL_GPL(imx_drm_encoder_destroy);
+
+static void imx_drm_output_poll_changed(struct drm_device *drm)
+{
+#if IS_ENABLED(CONFIG_DRM_IMX_FB_HELPER)
+	struct imx_drm_device *imxdrm = drm->dev_private;
+
+	drm_fbdev_cma_hotplug_event(imxdrm->fbhelper);
+#endif
+}
+
+static struct drm_mode_config_funcs imx_drm_mode_config_funcs = {
+	.fb_create = drm_fb_cma_create,
+	.output_poll_changed = imx_drm_output_poll_changed,
+};
+
+/*
+ * Main DRM initialisation. This binds, initialises and registers
+ * with DRM the subcomponents of the driver.
+ */
+static int imx_drm_driver_load(struct drm_device *drm, unsigned long flags)
+{
+	struct imx_drm_device *imxdrm;
+	struct drm_connector *connector;
+	int ret;
+
+	imxdrm = devm_kzalloc(drm->dev, sizeof(*imxdrm), GFP_KERNEL);
+	if (!imxdrm)
+		return -ENOMEM;
+
+	imxdrm->drm = drm;
+
+	drm->dev_private = imxdrm;
+
+	/*
+	 * enable drm irq mode.
+	 * - with irq_enabled = true, we can use the vblank feature.
+	 *
+	 * P.S. note that we wouldn't use drm irq handler but
+	 *      just specific driver own one instead because
+	 *      drm framework supports only one irq handler and
+	 *      drivers can well take care of their interrupts
+	 */
+	drm->irq_enabled = true;
+
+	/*
+	 * set max width and height as default value(4096x4096).
+	 * this value would be used to check framebuffer size limitation
+	 * at drm_mode_addfb().
+	 */
+	drm->mode_config.min_width = 64;
+	drm->mode_config.min_height = 64;
+	drm->mode_config.max_width = 4096;
+	drm->mode_config.max_height = 4096;
+	drm->mode_config.funcs = &imx_drm_mode_config_funcs;
+
+	drm_mode_config_init(drm);
+
+	ret = drm_vblank_init(drm, MAX_CRTC);
+	if (ret)
+		goto err_kms;
+
+	/*
+	 * with vblank_disable_allowed = true, vblank interrupt will be
+	 * disabled by drm timer once a current process gives up ownership
+	 * of vblank event. (after drm_vblank_put function is called)
+	 */
+	drm->vblank_disable_allowed = true;
+
+	platform_set_drvdata(drm->platformdev, drm);
+
+	/* Now try and bind all our sub-components */
+	ret = component_bind_all(drm->dev, drm);
+	if (ret)
+		goto err_vblank;
+
+	/*
+	 * All components are now added, we can publish the connector sysfs
+	 * entries to userspace.  This will generate hotplug events and so
+	 * userspace will expect to be able to access DRM at this point.
+	 */
+	list_for_each_entry(connector, &drm->mode_config.connector_list, head) {
+		ret = drm_connector_register(connector);
+		if (ret) {
+			dev_err(drm->dev,
+				"[CONNECTOR:%d:%s] drm_connector_register failed: %d\n",
+				connector->base.id,
+				connector->name, ret);
+			goto err_unbind;
+		}
+	}
+
+	/*
+	 * All components are now initialised, so setup the fb helper.
+	 * The fb helper takes copies of key hardware information, so the
+	 * crtcs/connectors/encoders must not change after this point.
+	 */
+#if IS_ENABLED(CONFIG_DRM_IMX_FB_HELPER)
+	if (legacyfb_depth != 16 && legacyfb_depth != 32) {
+		dev_warn(drm->dev, "Invalid legacyfb_depth.  Defaulting to 16bpp\n");
+		legacyfb_depth = 16;
+	}
+	imxdrm->fbhelper = drm_fbdev_cma_init(drm, legacyfb_depth,
+				drm->mode_config.num_crtc, MAX_CRTC);
+	if (IS_ERR(imxdrm->fbhelper)) {
+		ret = PTR_ERR(imxdrm->fbhelper);
+		imxdrm->fbhelper = NULL;
+		goto err_unbind;
+	}
+#endif
+
+	drm_kms_helper_poll_init(drm);
+
+	return 0;
+
+err_unbind:
+	component_unbind_all(drm->dev, drm);
+err_vblank:
+	drm_vblank_cleanup(drm);
+err_kms:
+	drm_mode_config_cleanup(drm);
+
+	return ret;
+}
+
+/*
+ * imx_drm_add_crtc - add a new crtc
+ */
+int imx_drm_add_crtc(struct drm_device *drm, struct drm_crtc *crtc,
+		struct imx_drm_crtc **new_crtc,
+		const struct imx_drm_crtc_helper_funcs *imx_drm_helper_funcs,
+		struct device_node *port)
+{
+	struct imx_drm_device *imxdrm = drm->dev_private;
+	struct imx_drm_crtc *imx_drm_crtc;
+	int ret;
+
+	/*
+	 * The vblank arrays are dimensioned by MAX_CRTC - we can't
+	 * pass IDs greater than this to those functions.
+	 */
+	if (imxdrm->pipes >= MAX_CRTC)
+		return -EINVAL;
+
+	if (imxdrm->drm->open_count)
+		return -EBUSY;
+
+	imx_drm_crtc = kzalloc(sizeof(*imx_drm_crtc), GFP_KERNEL);
+	if (!imx_drm_crtc)
+		return -ENOMEM;
+
+	imx_drm_crtc->imx_drm_helper_funcs = *imx_drm_helper_funcs;
+	imx_drm_crtc->pipe = imxdrm->pipes++;
+	imx_drm_crtc->port = port;
+	imx_drm_crtc->crtc = crtc;
+
+	imxdrm->crtc[imx_drm_crtc->pipe] = imx_drm_crtc;
+
+	*new_crtc = imx_drm_crtc;
+
+	ret = drm_mode_crtc_set_gamma_size(imx_drm_crtc->crtc, 256);
+	if (ret)
+		goto err_register;
+
+	drm_crtc_helper_add(crtc,
+			imx_drm_crtc->imx_drm_helper_funcs.crtc_helper_funcs);
+
+	drm_crtc_init(drm, crtc,
+			imx_drm_crtc->imx_drm_helper_funcs.crtc_funcs);
+
+	return 0;
+
+err_register:
+	imxdrm->crtc[imx_drm_crtc->pipe] = NULL;
+	kfree(imx_drm_crtc);
+	return ret;
+}
+EXPORT_SYMBOL_GPL(imx_drm_add_crtc);
+
+/*
+ * imx_drm_remove_crtc - remove a crtc
+ */
+int imx_drm_remove_crtc(struct imx_drm_crtc *imx_drm_crtc)
+{
+	struct imx_drm_device *imxdrm = imx_drm_crtc->crtc->dev->dev_private;
+
+	drm_crtc_cleanup(imx_drm_crtc->crtc);
+
+	imxdrm->crtc[imx_drm_crtc->pipe] = NULL;
+
+	kfree(imx_drm_crtc);
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(imx_drm_remove_crtc);
+
+/*
+ * Find the DRM CRTC possible mask for the connected endpoint.
+ *
+ * The encoder possible masks are defined by their position in the
+ * mode_config crtc_list.  This means that CRTCs must not be added
+ * or removed once the DRM device has been fully initialised.
+ */
+static uint32_t imx_drm_find_crtc_mask(struct imx_drm_device *imxdrm,
+	struct device_node *endpoint)
+{
+	struct device_node *port;
+	unsigned i;
+
+	port = of_graph_get_remote_port(endpoint);
+	if (!port)
+		return 0;
+	of_node_put(port);
+
+	for (i = 0; i < MAX_CRTC; i++) {
+		struct imx_drm_crtc *imx_drm_crtc = imxdrm->crtc[i];
+
+		if (imx_drm_crtc && imx_drm_crtc->port == port)
+			return drm_crtc_mask(imx_drm_crtc->crtc);
+	}
+
+	return 0;
+}
+
+static struct device_node *imx_drm_of_get_next_endpoint(
+		const struct device_node *parent, struct device_node *prev)
+{
+	struct device_node *node = of_graph_get_next_endpoint(parent, prev);
+
+	of_node_put(prev);
+	return node;
+}
+
+int imx_drm_encoder_parse_of(struct drm_device *drm,
+	struct drm_encoder *encoder, struct device_node *np)
+{
+	struct imx_drm_device *imxdrm = drm->dev_private;
+	struct device_node *ep = NULL;
+	uint32_t crtc_mask = 0;
+	int i;
+
+	for (i = 0; ; i++) {
+		u32 mask;
+
+		ep = imx_drm_of_get_next_endpoint(np, ep);
+		if (!ep)
+			break;
+
+		mask = imx_drm_find_crtc_mask(imxdrm, ep);
+
+		/*
+		 * If we failed to find the CRTC(s) which this encoder is
+		 * supposed to be connected to, it's because the CRTC has
+		 * not been registered yet.  Defer probing, and hope that
+		 * the required CRTC is added later.
+		 */
+		if (mask == 0)
+			return -EPROBE_DEFER;
+
+		crtc_mask |= mask;
+	}
+
+	of_node_put(ep);
+	if (i == 0)
+		return -ENOENT;
+
+	encoder->possible_crtcs = crtc_mask;
+
+	/* FIXME: this is the mask of outputs which can clone this output. */
+	encoder->possible_clones = ~0;
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(imx_drm_encoder_parse_of);
+
+/*
+ * @node: device tree node containing encoder input ports
+ * @encoder: drm_encoder
+ */
+int imx_drm_encoder_get_mux_id(struct device_node *node,
+			       struct drm_encoder *encoder)
+{
+	struct imx_drm_crtc *imx_crtc = imx_drm_find_crtc(encoder->crtc);
+	struct device_node *ep = NULL;
+	struct of_endpoint endpoint;
+	struct device_node *port;
+	int ret;
+
+	if (!node || !imx_crtc)
+		return -EINVAL;
+
+	do {
+		ep = imx_drm_of_get_next_endpoint(node, ep);
+		if (!ep)
+			break;
+
+		port = of_graph_get_remote_port(ep);
+		of_node_put(port);
+		if (port == imx_crtc->port) {
+			ret = of_graph_parse_endpoint(ep, &endpoint);
+			return ret ? ret : endpoint.port;
+		}
+	} while (ep);
+
+	return -EINVAL;
+}
+EXPORT_SYMBOL_GPL(imx_drm_encoder_get_mux_id);
+
+static const struct drm_ioctl_desc imx_drm_ioctls[] = {
+	/* none so far */
+};
+
+static struct drm_driver imx_drm_driver = {
+	.driver_features	= DRIVER_MODESET | DRIVER_GEM | DRIVER_PRIME,
+	.load			= imx_drm_driver_load,
+	.unload			= imx_drm_driver_unload,
+	.lastclose		= imx_drm_driver_lastclose,
+	.preclose		= imx_drm_driver_preclose,
+	.set_busid		= drm_platform_set_busid,
+	.gem_free_object	= drm_gem_cma_free_object,
+	.gem_vm_ops		= &drm_gem_cma_vm_ops,
+	.dumb_create		= drm_gem_cma_dumb_create,
+	.dumb_map_offset	= drm_gem_cma_dumb_map_offset,
+	.dumb_destroy		= drm_gem_dumb_destroy,
+
+	.prime_handle_to_fd	= drm_gem_prime_handle_to_fd,
+	.prime_fd_to_handle	= drm_gem_prime_fd_to_handle,
+	.gem_prime_import	= drm_gem_prime_import,
+	.gem_prime_export	= drm_gem_prime_export,
+	.gem_prime_get_sg_table	= drm_gem_cma_prime_get_sg_table,
+	.gem_prime_import_sg_table = drm_gem_cma_prime_import_sg_table,
+	.gem_prime_vmap		= drm_gem_cma_prime_vmap,
+	.gem_prime_vunmap	= drm_gem_cma_prime_vunmap,
+	.gem_prime_mmap		= drm_gem_cma_prime_mmap,
+	.get_vblank_counter	= drm_vblank_count,
+	.enable_vblank		= imx_drm_enable_vblank,
+	.disable_vblank		= imx_drm_disable_vblank,
+	.ioctls			= imx_drm_ioctls,
+	.num_ioctls		= ARRAY_SIZE(imx_drm_ioctls),
+	.fops			= &imx_drm_driver_fops,
+	.name			= "imx-drm",
+	.desc			= "i.MX DRM graphics",
+	.date			= "20120507",
+	.major			= 1,
+	.minor			= 0,
+	.patchlevel		= 0,
+};
+
+static int compare_of(struct device *dev, void *data)
+{
+	struct device_node *np = data;
+
+	/* Special case for LDB, one device for two channels */
+	if (of_node_cmp(np->name, "lvds-channel") == 0) {
+		np = of_get_parent(np);
+		of_node_put(np);
+	}
+
+	return dev->of_node == np;
+}
+
+static int imx_drm_bind(struct device *dev)
+{
+	return drm_platform_init(&imx_drm_driver, to_platform_device(dev));
+}
+
+static void imx_drm_unbind(struct device *dev)
+{
+	drm_put_dev(dev_get_drvdata(dev));
+}
+
+static const struct component_master_ops imx_drm_ops = {
+	.bind = imx_drm_bind,
+	.unbind = imx_drm_unbind,
+};
+
+static int imx_drm_platform_probe(struct platform_device *pdev)
+{
+	struct device_node *ep, *port, *remote;
+	struct component_match *match = NULL;
+	int ret;
+	int i;
+
+	/*
+	 * Bind the IPU display interface ports first, so that
+	 * imx_drm_encoder_parse_of called from encoder .bind callbacks
+	 * works as expected.
+	 */
+	for (i = 0; ; i++) {
+		port = of_parse_phandle(pdev->dev.of_node, "ports", i);
+		if (!port)
+			break;
+
+		component_match_add(&pdev->dev, &match, compare_of, port);
+	}
+
+	if (i == 0) {
+		dev_err(&pdev->dev, "missing 'ports' property\n");
+		return -ENODEV;
+	}
+
+	/* Then bind all encoders */
+	for (i = 0; ; i++) {
+		port = of_parse_phandle(pdev->dev.of_node, "ports", i);
+		if (!port)
+			break;
+
+		for_each_child_of_node(port, ep) {
+			remote = of_graph_get_remote_port_parent(ep);
+			if (!remote || !of_device_is_available(remote)) {
+				of_node_put(remote);
+				continue;
+			} else if (!of_device_is_available(remote->parent)) {
+				dev_warn(&pdev->dev, "parent device of %s is not available\n",
+					 remote->full_name);
+				of_node_put(remote);
+				continue;
+			}
+
+			component_match_add(&pdev->dev, &match, compare_of,
+					    remote);
+			of_node_put(remote);
+		}
+		of_node_put(port);
+	}
+
+	ret = dma_set_coherent_mask(&pdev->dev, DMA_BIT_MASK(32));
+	if (ret)
+		return ret;
+
+	return component_master_add_with_match(&pdev->dev, &imx_drm_ops, match);
+}
+
+static int imx_drm_platform_remove(struct platform_device *pdev)
+{
+	component_master_del(&pdev->dev, &imx_drm_ops);
+	return 0;
+}
+
+#ifdef CONFIG_PM_SLEEP
+static int imx_drm_suspend(struct device *dev)
+{
+	struct drm_device *drm_dev = dev_get_drvdata(dev);
+
+	/* The drm_dev is NULL before .load hook is called */
+	if (drm_dev == NULL)
+		return 0;
+
+	drm_kms_helper_poll_disable(drm_dev);
+
+	return 0;
+}
+
+static int imx_drm_resume(struct device *dev)
+{
+	struct drm_device *drm_dev = dev_get_drvdata(dev);
+
+	if (drm_dev == NULL)
+		return 0;
+
+	drm_helper_resume_force_mode(drm_dev);
+	drm_kms_helper_poll_enable(drm_dev);
+
+	return 0;
+}
+#endif
+
+static SIMPLE_DEV_PM_OPS(imx_drm_pm_ops, imx_drm_suspend, imx_drm_resume);
+
+static const struct of_device_id imx_drm_dt_ids[] = {
+	{ .compatible = "fsl,imx-display-subsystem", },
+	{ /* sentinel */ },
+};
+MODULE_DEVICE_TABLE(of, imx_drm_dt_ids);
+
+static struct platform_driver imx_drm_pdrv = {
+	.probe		= imx_drm_platform_probe,
+	.remove		= imx_drm_platform_remove,
+	.driver		= {
+		.name	= "imx-drm",
+		.pm	= &imx_drm_pm_ops,
+		.of_match_table = imx_drm_dt_ids,
+	},
+};
+module_platform_driver(imx_drm_pdrv);
+
+MODULE_AUTHOR("Sascha Hauer <s.hauer@pengutronix.de>");
+MODULE_DESCRIPTION("i.MX drm driver core");
+MODULE_LICENSE("GPL");
diff --git a/drivers/gpu/drm/imx/imx-drm.h b/drivers/gpu/drm/imx/imx-drm.h
new file mode 100644
index 000000000000..7453ae00c412
--- /dev/null
+++ b/drivers/gpu/drm/imx/imx-drm.h
@@ -0,0 +1,56 @@
+#ifndef _IMX_DRM_H_
+#define _IMX_DRM_H_
+
+struct device_node;
+struct drm_crtc;
+struct drm_connector;
+struct drm_device;
+struct drm_display_mode;
+struct drm_encoder;
+struct drm_fbdev_cma;
+struct drm_framebuffer;
+struct imx_drm_crtc;
+struct platform_device;
+
+int imx_drm_crtc_id(struct imx_drm_crtc *crtc);
+
+struct imx_drm_crtc_helper_funcs {
+	int (*enable_vblank)(struct drm_crtc *crtc);
+	void (*disable_vblank)(struct drm_crtc *crtc);
+	int (*set_interface_pix_fmt)(struct drm_crtc *crtc, u32 encoder_type,
+			u32 pix_fmt, int hsync_pin, int vsync_pin);
+	const struct drm_crtc_helper_funcs *crtc_helper_funcs;
+	const struct drm_crtc_funcs *crtc_funcs;
+};
+
+int imx_drm_add_crtc(struct drm_device *drm, struct drm_crtc *crtc,
+		struct imx_drm_crtc **new_crtc,
+		const struct imx_drm_crtc_helper_funcs *imx_helper_funcs,
+		struct device_node *port);
+int imx_drm_remove_crtc(struct imx_drm_crtc *);
+int imx_drm_init_drm(struct platform_device *pdev,
+		int preferred_bpp);
+int imx_drm_exit_drm(void);
+
+int imx_drm_crtc_vblank_get(struct imx_drm_crtc *imx_drm_crtc);
+void imx_drm_crtc_vblank_put(struct imx_drm_crtc *imx_drm_crtc);
+void imx_drm_handle_vblank(struct imx_drm_crtc *imx_drm_crtc);
+
+void imx_drm_mode_config_init(struct drm_device *drm);
+
+struct drm_gem_cma_object *imx_drm_fb_get_obj(struct drm_framebuffer *fb);
+
+int imx_drm_panel_format_pins(struct drm_encoder *encoder,
+		u32 interface_pix_fmt, int hsync_pin, int vsync_pin);
+int imx_drm_panel_format(struct drm_encoder *encoder,
+		u32 interface_pix_fmt);
+
+int imx_drm_encoder_get_mux_id(struct device_node *node,
+		struct drm_encoder *encoder);
+int imx_drm_encoder_parse_of(struct drm_device *drm,
+	struct drm_encoder *encoder, struct device_node *np);
+
+void imx_drm_connector_destroy(struct drm_connector *connector);
+void imx_drm_encoder_destroy(struct drm_encoder *encoder);
+
+#endif /* _IMX_DRM_H_ */
diff --git a/drivers/gpu/drm/imx/imx-hdmi.c b/drivers/gpu/drm/imx/imx-hdmi.c
new file mode 100644
index 000000000000..ddc53e039530
--- /dev/null
+++ b/drivers/gpu/drm/imx/imx-hdmi.c
@@ -0,0 +1,1766 @@
+/*
+ * Copyright (C) 2011-2013 Freescale Semiconductor, Inc.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * SH-Mobile High-Definition Multimedia Interface (HDMI) driver
+ * for SLISHDMI13T and SLIPHDMIT IP cores
+ *
+ * Copyright (C) 2010, Guennadi Liakhovetski <g.liakhovetski@gmx.de>
+ */
+
+#include <linux/component.h>
+#include <linux/irq.h>
+#include <linux/delay.h>
+#include <linux/err.h>
+#include <linux/clk.h>
+#include <linux/hdmi.h>
+#include <linux/regmap.h>
+#include <linux/mfd/syscon.h>
+#include <linux/mfd/syscon/imx6q-iomuxc-gpr.h>
+#include <linux/of_device.h>
+
+#include <drm/drmP.h>
+#include <drm/drm_crtc_helper.h>
+#include <drm/drm_edid.h>
+#include <drm/drm_encoder_slave.h>
+#include <video/imx-ipu-v3.h>
+
+#include "imx-hdmi.h"
+#include "imx-drm.h"
+
+#define HDMI_EDID_LEN		512
+
+#define RGB			0
+#define YCBCR444		1
+#define YCBCR422_16BITS		2
+#define YCBCR422_8BITS		3
+#define XVYCC444		4
+
+enum hdmi_datamap {
+	RGB444_8B = 0x01,
+	RGB444_10B = 0x03,
+	RGB444_12B = 0x05,
+	RGB444_16B = 0x07,
+	YCbCr444_8B = 0x09,
+	YCbCr444_10B = 0x0B,
+	YCbCr444_12B = 0x0D,
+	YCbCr444_16B = 0x0F,
+	YCbCr422_8B = 0x16,
+	YCbCr422_10B = 0x14,
+	YCbCr422_12B = 0x12,
+};
+
+enum imx_hdmi_devtype {
+	IMX6Q_HDMI,
+	IMX6DL_HDMI,
+};
+
+static const u16 csc_coeff_default[3][4] = {
+	{ 0x2000, 0x0000, 0x0000, 0x0000 },
+	{ 0x0000, 0x2000, 0x0000, 0x0000 },
+	{ 0x0000, 0x0000, 0x2000, 0x0000 }
+};
+
+static const u16 csc_coeff_rgb_out_eitu601[3][4] = {
+	{ 0x2000, 0x6926, 0x74fd, 0x010e },
+	{ 0x2000, 0x2cdd, 0x0000, 0x7e9a },
+	{ 0x2000, 0x0000, 0x38b4, 0x7e3b }
+};
+
+static const u16 csc_coeff_rgb_out_eitu709[3][4] = {
+	{ 0x2000, 0x7106, 0x7a02, 0x00a7 },
+	{ 0x2000, 0x3264, 0x0000, 0x7e6d },
+	{ 0x2000, 0x0000, 0x3b61, 0x7e25 }
+};
+
+static const u16 csc_coeff_rgb_in_eitu601[3][4] = {
+	{ 0x2591, 0x1322, 0x074b, 0x0000 },
+	{ 0x6535, 0x2000, 0x7acc, 0x0200 },
+	{ 0x6acd, 0x7534, 0x2000, 0x0200 }
+};
+
+static const u16 csc_coeff_rgb_in_eitu709[3][4] = {
+	{ 0x2dc5, 0x0d9b, 0x049e, 0x0000 },
+	{ 0x62f0, 0x2000, 0x7d11, 0x0200 },
+	{ 0x6756, 0x78ab, 0x2000, 0x0200 }
+};
+
+struct hdmi_vmode {
+	bool mdvi;
+	bool mhsyncpolarity;
+	bool mvsyncpolarity;
+	bool minterlaced;
+	bool mdataenablepolarity;
+
+	unsigned int mpixelclock;
+	unsigned int mpixelrepetitioninput;
+	unsigned int mpixelrepetitionoutput;
+};
+
+struct hdmi_data_info {
+	unsigned int enc_in_format;
+	unsigned int enc_out_format;
+	unsigned int enc_color_depth;
+	unsigned int colorimetry;
+	unsigned int pix_repet_factor;
+	unsigned int hdcp_enable;
+	struct hdmi_vmode video_mode;
+};
+
+struct imx_hdmi {
+	struct drm_connector connector;
+	struct drm_encoder encoder;
+
+	enum imx_hdmi_devtype dev_type;
+	struct device *dev;
+	struct clk *isfr_clk;
+	struct clk *iahb_clk;
+
+	struct hdmi_data_info hdmi_data;
+	int vic;
+
+	u8 edid[HDMI_EDID_LEN];
+	bool cable_plugin;
+
+	bool phy_enabled;
+	struct drm_display_mode previous_mode;
+
+	struct regmap *regmap;
+	struct i2c_adapter *ddc;
+	void __iomem *regs;
+
+	unsigned int sample_rate;
+	int ratio;
+};
+
+static void imx_hdmi_set_ipu_di_mux(struct imx_hdmi *hdmi, int ipu_di)
+{
+	regmap_update_bits(hdmi->regmap, IOMUXC_GPR3,
+			   IMX6Q_GPR3_HDMI_MUX_CTL_MASK,
+			   ipu_di << IMX6Q_GPR3_HDMI_MUX_CTL_SHIFT);
+}
+
+static inline void hdmi_writeb(struct imx_hdmi *hdmi, u8 val, int offset)
+{
+	writeb(val, hdmi->regs + offset);
+}
+
+static inline u8 hdmi_readb(struct imx_hdmi *hdmi, int offset)
+{
+	return readb(hdmi->regs + offset);
+}
+
+static void hdmi_modb(struct imx_hdmi *hdmi, u8 data, u8 mask, unsigned reg)
+{
+	u8 val = hdmi_readb(hdmi, reg) & ~mask;
+
+	val |= data & mask;
+	hdmi_writeb(hdmi, val, reg);
+}
+
+static void hdmi_mask_writeb(struct imx_hdmi *hdmi, u8 data, unsigned int reg,
+		      u8 shift, u8 mask)
+{
+	hdmi_modb(hdmi, data << shift, mask, reg);
+}
+
+static void hdmi_set_clock_regenerator_n(struct imx_hdmi *hdmi,
+					 unsigned int value)
+{
+	hdmi_writeb(hdmi, value & 0xff, HDMI_AUD_N1);
+	hdmi_writeb(hdmi, (value >> 8) & 0xff, HDMI_AUD_N2);
+	hdmi_writeb(hdmi, (value >> 16) & 0x0f, HDMI_AUD_N3);
+
+	/* nshift factor = 0 */
+	hdmi_modb(hdmi, 0, HDMI_AUD_CTS3_N_SHIFT_MASK, HDMI_AUD_CTS3);
+}
+
+static void hdmi_regenerate_cts(struct imx_hdmi *hdmi, unsigned int cts)
+{
+	/* Must be set/cleared first */
+	hdmi_modb(hdmi, 0, HDMI_AUD_CTS3_CTS_MANUAL, HDMI_AUD_CTS3);
+
+	hdmi_writeb(hdmi, cts & 0xff, HDMI_AUD_CTS1);
+	hdmi_writeb(hdmi, (cts >> 8) & 0xff, HDMI_AUD_CTS2);
+	hdmi_writeb(hdmi, ((cts >> 16) & HDMI_AUD_CTS3_AUDCTS19_16_MASK) |
+		    HDMI_AUD_CTS3_CTS_MANUAL, HDMI_AUD_CTS3);
+}
+
+static unsigned int hdmi_compute_n(unsigned int freq, unsigned long pixel_clk,
+				   unsigned int ratio)
+{
+	unsigned int n = (128 * freq) / 1000;
+
+	switch (freq) {
+	case 32000:
+		if (pixel_clk == 25170000)
+			n = (ratio == 150) ? 9152 : 4576;
+		else if (pixel_clk == 27020000)
+			n = (ratio == 150) ? 8192 : 4096;
+		else if (pixel_clk == 74170000 || pixel_clk == 148350000)
+			n = 11648;
+		else
+			n = 4096;
+		break;
+
+	case 44100:
+		if (pixel_clk == 25170000)
+			n = 7007;
+		else if (pixel_clk == 74170000)
+			n = 17836;
+		else if (pixel_clk == 148350000)
+			n = (ratio == 150) ? 17836 : 8918;
+		else
+			n = 6272;
+		break;
+
+	case 48000:
+		if (pixel_clk == 25170000)
+			n = (ratio == 150) ? 9152 : 6864;
+		else if (pixel_clk == 27020000)
+			n = (ratio == 150) ? 8192 : 6144;
+		else if (pixel_clk == 74170000)
+			n = 11648;
+		else if (pixel_clk == 148350000)
+			n = (ratio == 150) ? 11648 : 5824;
+		else
+			n = 6144;
+		break;
+
+	case 88200:
+		n = hdmi_compute_n(44100, pixel_clk, ratio) * 2;
+		break;
+
+	case 96000:
+		n = hdmi_compute_n(48000, pixel_clk, ratio) * 2;
+		break;
+
+	case 176400:
+		n = hdmi_compute_n(44100, pixel_clk, ratio) * 4;
+		break;
+
+	case 192000:
+		n = hdmi_compute_n(48000, pixel_clk, ratio) * 4;
+		break;
+
+	default:
+		break;
+	}
+
+	return n;
+}
+
+static unsigned int hdmi_compute_cts(unsigned int freq, unsigned long pixel_clk,
+				     unsigned int ratio)
+{
+	unsigned int cts = 0;
+
+	pr_debug("%s: freq: %d pixel_clk: %ld ratio: %d\n", __func__, freq,
+		 pixel_clk, ratio);
+
+	switch (freq) {
+	case 32000:
+		if (pixel_clk == 297000000) {
+			cts = 222750;
+			break;
+		}
+	case 48000:
+	case 96000:
+	case 192000:
+		switch (pixel_clk) {
+		case 25200000:
+		case 27000000:
+		case 54000000:
+		case 74250000:
+		case 148500000:
+			cts = pixel_clk / 1000;
+			break;
+		case 297000000:
+			cts = 247500;
+			break;
+		/*
+		 * All other TMDS clocks are not supported by
+		 * DWC_hdmi_tx. The TMDS clocks divided or
+		 * multiplied by 1,001 coefficients are not
+		 * supported.
+		 */
+		default:
+			break;
+		}
+		break;
+	case 44100:
+	case 88200:
+	case 176400:
+		switch (pixel_clk) {
+		case 25200000:
+			cts = 28000;
+			break;
+		case 27000000:
+			cts = 30000;
+			break;
+		case 54000000:
+			cts = 60000;
+			break;
+		case 74250000:
+			cts = 82500;
+			break;
+		case 148500000:
+			cts = 165000;
+			break;
+		case 297000000:
+			cts = 247500;
+			break;
+		default:
+			break;
+		}
+		break;
+	default:
+		break;
+	}
+	if (ratio == 100)
+		return cts;
+	return (cts * ratio) / 100;
+}
+
+static void hdmi_set_clk_regenerator(struct imx_hdmi *hdmi,
+	unsigned long pixel_clk)
+{
+	unsigned int clk_n, clk_cts;
+
+	clk_n = hdmi_compute_n(hdmi->sample_rate, pixel_clk,
+			       hdmi->ratio);
+	clk_cts = hdmi_compute_cts(hdmi->sample_rate, pixel_clk,
+				   hdmi->ratio);
+
+	if (!clk_cts) {
+		dev_dbg(hdmi->dev, "%s: pixel clock not supported: %lu\n",
+			 __func__, pixel_clk);
+		return;
+	}
+
+	dev_dbg(hdmi->dev, "%s: samplerate=%d  ratio=%d  pixelclk=%lu  N=%d cts=%d\n",
+		__func__, hdmi->sample_rate, hdmi->ratio,
+		pixel_clk, clk_n, clk_cts);
+
+	hdmi_set_clock_regenerator_n(hdmi, clk_n);
+	hdmi_regenerate_cts(hdmi, clk_cts);
+}
+
+static void hdmi_init_clk_regenerator(struct imx_hdmi *hdmi)
+{
+	hdmi_set_clk_regenerator(hdmi, 74250000);
+}
+
+static void hdmi_clk_regenerator_update_pixel_clock(struct imx_hdmi *hdmi)
+{
+	hdmi_set_clk_regenerator(hdmi, hdmi->hdmi_data.video_mode.mpixelclock);
+}
+
+/*
+ * this submodule is responsible for the video data synchronization.
+ * for example, for RGB 4:4:4 input, the data map is defined as
+ *			pin{47~40} <==> R[7:0]
+ *			pin{31~24} <==> G[7:0]
+ *			pin{15~8}  <==> B[7:0]
+ */
+static void hdmi_video_sample(struct imx_hdmi *hdmi)
+{
+	int color_format = 0;
+	u8 val;
+
+	if (hdmi->hdmi_data.enc_in_format == RGB) {
+		if (hdmi->hdmi_data.enc_color_depth == 8)
+			color_format = 0x01;
+		else if (hdmi->hdmi_data.enc_color_depth == 10)
+			color_format = 0x03;
+		else if (hdmi->hdmi_data.enc_color_depth == 12)
+			color_format = 0x05;
+		else if (hdmi->hdmi_data.enc_color_depth == 16)
+			color_format = 0x07;
+		else
+			return;
+	} else if (hdmi->hdmi_data.enc_in_format == YCBCR444) {
+		if (hdmi->hdmi_data.enc_color_depth == 8)
+			color_format = 0x09;
+		else if (hdmi->hdmi_data.enc_color_depth == 10)
+			color_format = 0x0B;
+		else if (hdmi->hdmi_data.enc_color_depth == 12)
+			color_format = 0x0D;
+		else if (hdmi->hdmi_data.enc_color_depth == 16)
+			color_format = 0x0F;
+		else
+			return;
+	} else if (hdmi->hdmi_data.enc_in_format == YCBCR422_8BITS) {
+		if (hdmi->hdmi_data.enc_color_depth == 8)
+			color_format = 0x16;
+		else if (hdmi->hdmi_data.enc_color_depth == 10)
+			color_format = 0x14;
+		else if (hdmi->hdmi_data.enc_color_depth == 12)
+			color_format = 0x12;
+		else
+			return;
+	}
+
+	val = HDMI_TX_INVID0_INTERNAL_DE_GENERATOR_DISABLE |
+		((color_format << HDMI_TX_INVID0_VIDEO_MAPPING_OFFSET) &
+		HDMI_TX_INVID0_VIDEO_MAPPING_MASK);
+	hdmi_writeb(hdmi, val, HDMI_TX_INVID0);
+
+	/* Enable TX stuffing: When DE is inactive, fix the output data to 0 */
+	val = HDMI_TX_INSTUFFING_BDBDATA_STUFFING_ENABLE |
+		HDMI_TX_INSTUFFING_RCRDATA_STUFFING_ENABLE |
+		HDMI_TX_INSTUFFING_GYDATA_STUFFING_ENABLE;
+	hdmi_writeb(hdmi, val, HDMI_TX_INSTUFFING);
+	hdmi_writeb(hdmi, 0x0, HDMI_TX_GYDATA0);
+	hdmi_writeb(hdmi, 0x0, HDMI_TX_GYDATA1);
+	hdmi_writeb(hdmi, 0x0, HDMI_TX_RCRDATA0);
+	hdmi_writeb(hdmi, 0x0, HDMI_TX_RCRDATA1);
+	hdmi_writeb(hdmi, 0x0, HDMI_TX_BCBDATA0);
+	hdmi_writeb(hdmi, 0x0, HDMI_TX_BCBDATA1);
+}
+
+static int is_color_space_conversion(struct imx_hdmi *hdmi)
+{
+	return hdmi->hdmi_data.enc_in_format != hdmi->hdmi_data.enc_out_format;
+}
+
+static int is_color_space_decimation(struct imx_hdmi *hdmi)
+{
+	if (hdmi->hdmi_data.enc_out_format != YCBCR422_8BITS)
+		return 0;
+	if (hdmi->hdmi_data.enc_in_format == RGB ||
+	    hdmi->hdmi_data.enc_in_format == YCBCR444)
+		return 1;
+	return 0;
+}
+
+static int is_color_space_interpolation(struct imx_hdmi *hdmi)
+{
+	if (hdmi->hdmi_data.enc_in_format != YCBCR422_8BITS)
+		return 0;
+	if (hdmi->hdmi_data.enc_out_format == RGB ||
+	    hdmi->hdmi_data.enc_out_format == YCBCR444)
+		return 1;
+	return 0;
+}
+
+static void imx_hdmi_update_csc_coeffs(struct imx_hdmi *hdmi)
+{
+	const u16 (*csc_coeff)[3][4] = &csc_coeff_default;
+	unsigned i;
+	u32 csc_scale = 1;
+
+	if (is_color_space_conversion(hdmi)) {
+		if (hdmi->hdmi_data.enc_out_format == RGB) {
+			if (hdmi->hdmi_data.colorimetry ==
+					HDMI_COLORIMETRY_ITU_601)
+				csc_coeff = &csc_coeff_rgb_out_eitu601;
+			else
+				csc_coeff = &csc_coeff_rgb_out_eitu709;
+		} else if (hdmi->hdmi_data.enc_in_format == RGB) {
+			if (hdmi->hdmi_data.colorimetry ==
+					HDMI_COLORIMETRY_ITU_601)
+				csc_coeff = &csc_coeff_rgb_in_eitu601;
+			else
+				csc_coeff = &csc_coeff_rgb_in_eitu709;
+			csc_scale = 0;
+		}
+	}
+
+	/* The CSC registers are sequential, alternating MSB then LSB */
+	for (i = 0; i < ARRAY_SIZE(csc_coeff_default[0]); i++) {
+		u16 coeff_a = (*csc_coeff)[0][i];
+		u16 coeff_b = (*csc_coeff)[1][i];
+		u16 coeff_c = (*csc_coeff)[2][i];
+
+		hdmi_writeb(hdmi, coeff_a & 0xff,
+			HDMI_CSC_COEF_A1_LSB + i * 2);
+		hdmi_writeb(hdmi, coeff_a >> 8, HDMI_CSC_COEF_A1_MSB + i * 2);
+		hdmi_writeb(hdmi, coeff_b & 0xff, HDMI_CSC_COEF_B1_LSB + i * 2);
+		hdmi_writeb(hdmi, coeff_b >> 8, HDMI_CSC_COEF_B1_MSB + i * 2);
+		hdmi_writeb(hdmi, coeff_c & 0xff,
+			HDMI_CSC_COEF_C1_LSB + i * 2);
+		hdmi_writeb(hdmi, coeff_c >> 8, HDMI_CSC_COEF_C1_MSB + i * 2);
+	}
+
+	hdmi_modb(hdmi, csc_scale, HDMI_CSC_SCALE_CSCSCALE_MASK,
+		  HDMI_CSC_SCALE);
+}
+
+static void hdmi_video_csc(struct imx_hdmi *hdmi)
+{
+	int color_depth = 0;
+	int interpolation = HDMI_CSC_CFG_INTMODE_DISABLE;
+	int decimation = 0;
+
+	/* YCC422 interpolation to 444 mode */
+	if (is_color_space_interpolation(hdmi))
+		interpolation = HDMI_CSC_CFG_INTMODE_CHROMA_INT_FORMULA1;
+	else if (is_color_space_decimation(hdmi))
+		decimation = HDMI_CSC_CFG_DECMODE_CHROMA_INT_FORMULA3;
+
+	if (hdmi->hdmi_data.enc_color_depth == 8)
+		color_depth = HDMI_CSC_SCALE_CSC_COLORDE_PTH_24BPP;
+	else if (hdmi->hdmi_data.enc_color_depth == 10)
+		color_depth = HDMI_CSC_SCALE_CSC_COLORDE_PTH_30BPP;
+	else if (hdmi->hdmi_data.enc_color_depth == 12)
+		color_depth = HDMI_CSC_SCALE_CSC_COLORDE_PTH_36BPP;
+	else if (hdmi->hdmi_data.enc_color_depth == 16)
+		color_depth = HDMI_CSC_SCALE_CSC_COLORDE_PTH_48BPP;
+	else
+		return;
+
+	/* Configure the CSC registers */
+	hdmi_writeb(hdmi, interpolation | decimation, HDMI_CSC_CFG);
+	hdmi_modb(hdmi, color_depth, HDMI_CSC_SCALE_CSC_COLORDE_PTH_MASK,
+		  HDMI_CSC_SCALE);
+
+	imx_hdmi_update_csc_coeffs(hdmi);
+}
+
+/*
+ * HDMI video packetizer is used to packetize the data.
+ * for example, if input is YCC422 mode or repeater is used,
+ * data should be repacked this module can be bypassed.
+ */
+static void hdmi_video_packetize(struct imx_hdmi *hdmi)
+{
+	unsigned int color_depth = 0;
+	unsigned int remap_size = HDMI_VP_REMAP_YCC422_16bit;
+	unsigned int output_select = HDMI_VP_CONF_OUTPUT_SELECTOR_PP;
+	struct hdmi_data_info *hdmi_data = &hdmi->hdmi_data;
+	u8 val, vp_conf;
+
+	if (hdmi_data->enc_out_format == RGB
+		|| hdmi_data->enc_out_format == YCBCR444) {
+		if (!hdmi_data->enc_color_depth)
+			output_select = HDMI_VP_CONF_OUTPUT_SELECTOR_BYPASS;
+		else if (hdmi_data->enc_color_depth == 8) {
+			color_depth = 4;
+			output_select = HDMI_VP_CONF_OUTPUT_SELECTOR_BYPASS;
+		} else if (hdmi_data->enc_color_depth == 10)
+			color_depth = 5;
+		else if (hdmi_data->enc_color_depth == 12)
+			color_depth = 6;
+		else if (hdmi_data->enc_color_depth == 16)
+			color_depth = 7;
+		else
+			return;
+	} else if (hdmi_data->enc_out_format == YCBCR422_8BITS) {
+		if (!hdmi_data->enc_color_depth ||
+		    hdmi_data->enc_color_depth == 8)
+			remap_size = HDMI_VP_REMAP_YCC422_16bit;
+		else if (hdmi_data->enc_color_depth == 10)
+			remap_size = HDMI_VP_REMAP_YCC422_20bit;
+		else if (hdmi_data->enc_color_depth == 12)
+			remap_size = HDMI_VP_REMAP_YCC422_24bit;
+		else
+			return;
+		output_select = HDMI_VP_CONF_OUTPUT_SELECTOR_YCC422;
+	} else
+		return;
+
+	/* set the packetizer registers */
+	val = ((color_depth << HDMI_VP_PR_CD_COLOR_DEPTH_OFFSET) &
+		HDMI_VP_PR_CD_COLOR_DEPTH_MASK) |
+		((hdmi_data->pix_repet_factor <<
+		HDMI_VP_PR_CD_DESIRED_PR_FACTOR_OFFSET) &
+		HDMI_VP_PR_CD_DESIRED_PR_FACTOR_MASK);
+	hdmi_writeb(hdmi, val, HDMI_VP_PR_CD);
+
+	hdmi_modb(hdmi, HDMI_VP_STUFF_PR_STUFFING_STUFFING_MODE,
+		  HDMI_VP_STUFF_PR_STUFFING_MASK, HDMI_VP_STUFF);
+
+	/* Data from pixel repeater block */
+	if (hdmi_data->pix_repet_factor > 1) {
+		vp_conf = HDMI_VP_CONF_PR_EN_ENABLE |
+			  HDMI_VP_CONF_BYPASS_SELECT_PIX_REPEATER;
+	} else { /* data from packetizer block */
+		vp_conf = HDMI_VP_CONF_PR_EN_DISABLE |
+			  HDMI_VP_CONF_BYPASS_SELECT_VID_PACKETIZER;
+	}
+
+	hdmi_modb(hdmi, vp_conf,
+		  HDMI_VP_CONF_PR_EN_MASK |
+		  HDMI_VP_CONF_BYPASS_SELECT_MASK, HDMI_VP_CONF);
+
+	hdmi_modb(hdmi, 1 << HDMI_VP_STUFF_IDEFAULT_PHASE_OFFSET,
+		  HDMI_VP_STUFF_IDEFAULT_PHASE_MASK, HDMI_VP_STUFF);
+
+	hdmi_writeb(hdmi, remap_size, HDMI_VP_REMAP);
+
+	if (output_select == HDMI_VP_CONF_OUTPUT_SELECTOR_PP) {
+		vp_conf = HDMI_VP_CONF_BYPASS_EN_DISABLE |
+			  HDMI_VP_CONF_PP_EN_ENABLE |
+			  HDMI_VP_CONF_YCC422_EN_DISABLE;
+	} else if (output_select == HDMI_VP_CONF_OUTPUT_SELECTOR_YCC422) {
+		vp_conf = HDMI_VP_CONF_BYPASS_EN_DISABLE |
+			  HDMI_VP_CONF_PP_EN_DISABLE |
+			  HDMI_VP_CONF_YCC422_EN_ENABLE;
+	} else if (output_select == HDMI_VP_CONF_OUTPUT_SELECTOR_BYPASS) {
+		vp_conf = HDMI_VP_CONF_BYPASS_EN_ENABLE |
+			  HDMI_VP_CONF_PP_EN_DISABLE |
+			  HDMI_VP_CONF_YCC422_EN_DISABLE;
+	} else {
+		return;
+	}
+
+	hdmi_modb(hdmi, vp_conf,
+		  HDMI_VP_CONF_BYPASS_EN_MASK | HDMI_VP_CONF_PP_EN_ENMASK |
+		  HDMI_VP_CONF_YCC422_EN_MASK, HDMI_VP_CONF);
+
+	hdmi_modb(hdmi, HDMI_VP_STUFF_PP_STUFFING_STUFFING_MODE |
+			HDMI_VP_STUFF_YCC422_STUFFING_STUFFING_MODE,
+		  HDMI_VP_STUFF_PP_STUFFING_MASK |
+		  HDMI_VP_STUFF_YCC422_STUFFING_MASK, HDMI_VP_STUFF);
+
+	hdmi_modb(hdmi, output_select, HDMI_VP_CONF_OUTPUT_SELECTOR_MASK,
+		  HDMI_VP_CONF);
+}
+
+static inline void hdmi_phy_test_clear(struct imx_hdmi *hdmi,
+						unsigned char bit)
+{
+	hdmi_modb(hdmi, bit << HDMI_PHY_TST0_TSTCLR_OFFSET,
+		  HDMI_PHY_TST0_TSTCLR_MASK, HDMI_PHY_TST0);
+}
+
+static inline void hdmi_phy_test_enable(struct imx_hdmi *hdmi,
+						unsigned char bit)
+{
+	hdmi_modb(hdmi, bit << HDMI_PHY_TST0_TSTEN_OFFSET,
+		  HDMI_PHY_TST0_TSTEN_MASK, HDMI_PHY_TST0);
+}
+
+static inline void hdmi_phy_test_clock(struct imx_hdmi *hdmi,
+						unsigned char bit)
+{
+	hdmi_modb(hdmi, bit << HDMI_PHY_TST0_TSTCLK_OFFSET,
+		  HDMI_PHY_TST0_TSTCLK_MASK, HDMI_PHY_TST0);
+}
+
+static inline void hdmi_phy_test_din(struct imx_hdmi *hdmi,
+						unsigned char bit)
+{
+	hdmi_writeb(hdmi, bit, HDMI_PHY_TST1);
+}
+
+static inline void hdmi_phy_test_dout(struct imx_hdmi *hdmi,
+						unsigned char bit)
+{
+	hdmi_writeb(hdmi, bit, HDMI_PHY_TST2);
+}
+
+static bool hdmi_phy_wait_i2c_done(struct imx_hdmi *hdmi, int msec)
+{
+	while ((hdmi_readb(hdmi, HDMI_IH_I2CMPHY_STAT0) & 0x3) == 0) {
+		if (msec-- == 0)
+			return false;
+		udelay(1000);
+	}
+	return true;
+}
+
+static void __hdmi_phy_i2c_write(struct imx_hdmi *hdmi, unsigned short data,
+			      unsigned char addr)
+{
+	hdmi_writeb(hdmi, 0xFF, HDMI_IH_I2CMPHY_STAT0);
+	hdmi_writeb(hdmi, addr, HDMI_PHY_I2CM_ADDRESS_ADDR);
+	hdmi_writeb(hdmi, (unsigned char)(data >> 8),
+		HDMI_PHY_I2CM_DATAO_1_ADDR);
+	hdmi_writeb(hdmi, (unsigned char)(data >> 0),
+		HDMI_PHY_I2CM_DATAO_0_ADDR);
+	hdmi_writeb(hdmi, HDMI_PHY_I2CM_OPERATION_ADDR_WRITE,
+		HDMI_PHY_I2CM_OPERATION_ADDR);
+	hdmi_phy_wait_i2c_done(hdmi, 1000);
+}
+
+static int hdmi_phy_i2c_write(struct imx_hdmi *hdmi, unsigned short data,
+				     unsigned char addr)
+{
+	__hdmi_phy_i2c_write(hdmi, data, addr);
+	return 0;
+}
+
+static void imx_hdmi_phy_enable_power(struct imx_hdmi *hdmi, u8 enable)
+{
+	hdmi_mask_writeb(hdmi, enable, HDMI_PHY_CONF0,
+			 HDMI_PHY_CONF0_PDZ_OFFSET,
+			 HDMI_PHY_CONF0_PDZ_MASK);
+}
+
+static void imx_hdmi_phy_enable_tmds(struct imx_hdmi *hdmi, u8 enable)
+{
+	hdmi_mask_writeb(hdmi, enable, HDMI_PHY_CONF0,
+			 HDMI_PHY_CONF0_ENTMDS_OFFSET,
+			 HDMI_PHY_CONF0_ENTMDS_MASK);
+}
+
+static void imx_hdmi_phy_gen2_pddq(struct imx_hdmi *hdmi, u8 enable)
+{
+	hdmi_mask_writeb(hdmi, enable, HDMI_PHY_CONF0,
+			 HDMI_PHY_CONF0_GEN2_PDDQ_OFFSET,
+			 HDMI_PHY_CONF0_GEN2_PDDQ_MASK);
+}
+
+static void imx_hdmi_phy_gen2_txpwron(struct imx_hdmi *hdmi, u8 enable)
+{
+	hdmi_mask_writeb(hdmi, enable, HDMI_PHY_CONF0,
+			 HDMI_PHY_CONF0_GEN2_TXPWRON_OFFSET,
+			 HDMI_PHY_CONF0_GEN2_TXPWRON_MASK);
+}
+
+static void imx_hdmi_phy_sel_data_en_pol(struct imx_hdmi *hdmi, u8 enable)
+{
+	hdmi_mask_writeb(hdmi, enable, HDMI_PHY_CONF0,
+			 HDMI_PHY_CONF0_SELDATAENPOL_OFFSET,
+			 HDMI_PHY_CONF0_SELDATAENPOL_MASK);
+}
+
+static void imx_hdmi_phy_sel_interface_control(struct imx_hdmi *hdmi, u8 enable)
+{
+	hdmi_mask_writeb(hdmi, enable, HDMI_PHY_CONF0,
+			 HDMI_PHY_CONF0_SELDIPIF_OFFSET,
+			 HDMI_PHY_CONF0_SELDIPIF_MASK);
+}
+
+enum {
+	RES_8,
+	RES_10,
+	RES_12,
+	RES_MAX,
+};
+
+struct mpll_config {
+	unsigned long mpixelclock;
+	struct {
+		u16 cpce;
+		u16 gmp;
+	} res[RES_MAX];
+};
+
+static const struct mpll_config mpll_config[] = {
+	{
+		45250000, {
+			{ 0x01e0, 0x0000 },
+			{ 0x21e1, 0x0000 },
+			{ 0x41e2, 0x0000 }
+		},
+	}, {
+		92500000, {
+			{ 0x0140, 0x0005 },
+			{ 0x2141, 0x0005 },
+			{ 0x4142, 0x0005 },
+		},
+	}, {
+		148500000, {
+			{ 0x00a0, 0x000a },
+			{ 0x20a1, 0x000a },
+			{ 0x40a2, 0x000a },
+		},
+	}, {
+		~0UL, {
+			{ 0x00a0, 0x000a },
+			{ 0x2001, 0x000f },
+			{ 0x4002, 0x000f },
+		},
+	}
+};
+
+struct curr_ctrl {
+	unsigned long mpixelclock;
+	u16 curr[RES_MAX];
+};
+
+static const struct curr_ctrl curr_ctrl[] = {
+	/*	pixelclk     bpp8    bpp10   bpp12 */
+	{
+		 54000000, { 0x091c, 0x091c, 0x06dc },
+	}, {
+		 58400000, { 0x091c, 0x06dc, 0x06dc },
+	}, {
+		 72000000, { 0x06dc, 0x06dc, 0x091c },
+	}, {
+		 74250000, { 0x06dc, 0x0b5c, 0x091c },
+	}, {
+		118800000, { 0x091c, 0x091c, 0x06dc },
+	}, {
+		216000000, { 0x06dc, 0x0b5c, 0x091c },
+	}
+};
+
+static int hdmi_phy_configure(struct imx_hdmi *hdmi, unsigned char prep,
+			      unsigned char res, int cscon)
+{
+	unsigned res_idx, i;
+	u8 val, msec;
+
+	if (prep)
+		return -EINVAL;
+
+	switch (res) {
+	case 0:	/* color resolution 0 is 8 bit colour depth */
+	case 8:
+		res_idx = RES_8;
+		break;
+	case 10:
+		res_idx = RES_10;
+		break;
+	case 12:
+		res_idx = RES_12;
+		break;
+	default:
+		return -EINVAL;
+	}
+
+	/* Enable csc path */
+	if (cscon)
+		val = HDMI_MC_FLOWCTRL_FEED_THROUGH_OFF_CSC_IN_PATH;
+	else
+		val = HDMI_MC_FLOWCTRL_FEED_THROUGH_OFF_CSC_BYPASS;
+
+	hdmi_writeb(hdmi, val, HDMI_MC_FLOWCTRL);
+
+	/* gen2 tx power off */
+	imx_hdmi_phy_gen2_txpwron(hdmi, 0);
+
+	/* gen2 pddq */
+	imx_hdmi_phy_gen2_pddq(hdmi, 1);
+
+	/* PHY reset */
+	hdmi_writeb(hdmi, HDMI_MC_PHYRSTZ_DEASSERT, HDMI_MC_PHYRSTZ);
+	hdmi_writeb(hdmi, HDMI_MC_PHYRSTZ_ASSERT, HDMI_MC_PHYRSTZ);
+
+	hdmi_writeb(hdmi, HDMI_MC_HEACPHY_RST_ASSERT, HDMI_MC_HEACPHY_RST);
+
+	hdmi_phy_test_clear(hdmi, 1);
+	hdmi_writeb(hdmi, HDMI_PHY_I2CM_SLAVE_ADDR_PHY_GEN2,
+			HDMI_PHY_I2CM_SLAVE_ADDR);
+	hdmi_phy_test_clear(hdmi, 0);
+
+	/* PLL/MPLL Cfg - always match on final entry */
+	for (i = 0; i < ARRAY_SIZE(mpll_config) - 1; i++)
+		if (hdmi->hdmi_data.video_mode.mpixelclock <=
+		    mpll_config[i].mpixelclock)
+			break;
+
+	hdmi_phy_i2c_write(hdmi, mpll_config[i].res[res_idx].cpce, 0x06);
+	hdmi_phy_i2c_write(hdmi, mpll_config[i].res[res_idx].gmp, 0x15);
+
+	for (i = 0; i < ARRAY_SIZE(curr_ctrl); i++)
+		if (hdmi->hdmi_data.video_mode.mpixelclock <=
+		    curr_ctrl[i].mpixelclock)
+			break;
+
+	if (i >= ARRAY_SIZE(curr_ctrl)) {
+		dev_err(hdmi->dev,
+				"Pixel clock %d - unsupported by HDMI\n",
+				hdmi->hdmi_data.video_mode.mpixelclock);
+		return -EINVAL;
+	}
+
+	/* CURRCTRL */
+	hdmi_phy_i2c_write(hdmi, curr_ctrl[i].curr[res_idx], 0x10);
+
+	hdmi_phy_i2c_write(hdmi, 0x0000, 0x13);  /* PLLPHBYCTRL */
+	hdmi_phy_i2c_write(hdmi, 0x0006, 0x17);
+	/* RESISTANCE TERM 133Ohm Cfg */
+	hdmi_phy_i2c_write(hdmi, 0x0005, 0x19);  /* TXTERM */
+	/* PREEMP Cgf 0.00 */
+	hdmi_phy_i2c_write(hdmi, 0x800d, 0x09);  /* CKSYMTXCTRL */
+	/* TX/CK LVL 10 */
+	hdmi_phy_i2c_write(hdmi, 0x01ad, 0x0E);  /* VLEVCTRL */
+	/* REMOVE CLK TERM */
+	hdmi_phy_i2c_write(hdmi, 0x8000, 0x05);  /* CKCALCTRL */
+
+	imx_hdmi_phy_enable_power(hdmi, 1);
+
+	/* toggle TMDS enable */
+	imx_hdmi_phy_enable_tmds(hdmi, 0);
+	imx_hdmi_phy_enable_tmds(hdmi, 1);
+
+	/* gen2 tx power on */
+	imx_hdmi_phy_gen2_txpwron(hdmi, 1);
+	imx_hdmi_phy_gen2_pddq(hdmi, 0);
+
+	/*Wait for PHY PLL lock */
+	msec = 5;
+	do {
+		val = hdmi_readb(hdmi, HDMI_PHY_STAT0) & HDMI_PHY_TX_PHY_LOCK;
+		if (!val)
+			break;
+
+		if (msec == 0) {
+			dev_err(hdmi->dev, "PHY PLL not locked\n");
+			return -ETIMEDOUT;
+		}
+
+		udelay(1000);
+		msec--;
+	} while (1);
+
+	return 0;
+}
+
+static int imx_hdmi_phy_init(struct imx_hdmi *hdmi)
+{
+	int i, ret;
+	bool cscon = false;
+
+	/*check csc whether needed activated in HDMI mode */
+	cscon = (is_color_space_conversion(hdmi) &&
+			!hdmi->hdmi_data.video_mode.mdvi);
+
+	/* HDMI Phy spec says to do the phy initialization sequence twice */
+	for (i = 0; i < 2; i++) {
+		imx_hdmi_phy_sel_data_en_pol(hdmi, 1);
+		imx_hdmi_phy_sel_interface_control(hdmi, 0);
+		imx_hdmi_phy_enable_tmds(hdmi, 0);
+		imx_hdmi_phy_enable_power(hdmi, 0);
+
+		/* Enable CSC */
+		ret = hdmi_phy_configure(hdmi, 0, 8, cscon);
+		if (ret)
+			return ret;
+	}
+
+	hdmi->phy_enabled = true;
+	return 0;
+}
+
+static void hdmi_tx_hdcp_config(struct imx_hdmi *hdmi)
+{
+	u8 de;
+
+	if (hdmi->hdmi_data.video_mode.mdataenablepolarity)
+		de = HDMI_A_VIDPOLCFG_DATAENPOL_ACTIVE_HIGH;
+	else
+		de = HDMI_A_VIDPOLCFG_DATAENPOL_ACTIVE_LOW;
+
+	/* disable rx detect */
+	hdmi_modb(hdmi, HDMI_A_HDCPCFG0_RXDETECT_DISABLE,
+		  HDMI_A_HDCPCFG0_RXDETECT_MASK, HDMI_A_HDCPCFG0);
+
+	hdmi_modb(hdmi, de, HDMI_A_VIDPOLCFG_DATAENPOL_MASK, HDMI_A_VIDPOLCFG);
+
+	hdmi_modb(hdmi, HDMI_A_HDCPCFG1_ENCRYPTIONDISABLE_DISABLE,
+		  HDMI_A_HDCPCFG1_ENCRYPTIONDISABLE_MASK, HDMI_A_HDCPCFG1);
+}
+
+static void hdmi_config_AVI(struct imx_hdmi *hdmi)
+{
+	u8 val, pix_fmt, under_scan;
+	u8 act_ratio, coded_ratio, colorimetry, ext_colorimetry;
+	bool aspect_16_9;
+
+	aspect_16_9 = false; /* FIXME */
+
+	/* AVI Data Byte 1 */
+	if (hdmi->hdmi_data.enc_out_format == YCBCR444)
+		pix_fmt = HDMI_FC_AVICONF0_PIX_FMT_YCBCR444;
+	else if (hdmi->hdmi_data.enc_out_format == YCBCR422_8BITS)
+		pix_fmt = HDMI_FC_AVICONF0_PIX_FMT_YCBCR422;
+	else
+		pix_fmt = HDMI_FC_AVICONF0_PIX_FMT_RGB;
+
+		under_scan =  HDMI_FC_AVICONF0_SCAN_INFO_NODATA;
+
+	/*
+	 * Active format identification data is present in the AVI InfoFrame.
+	 * Under scan info, no bar data
+	 */
+	val = pix_fmt | under_scan |
+		HDMI_FC_AVICONF0_ACTIVE_FMT_INFO_PRESENT |
+		HDMI_FC_AVICONF0_BAR_DATA_NO_DATA;
+
+	hdmi_writeb(hdmi, val, HDMI_FC_AVICONF0);
+
+	/* AVI Data Byte 2 -Set the Aspect Ratio */
+	if (aspect_16_9) {
+		act_ratio = HDMI_FC_AVICONF1_ACTIVE_ASPECT_RATIO_16_9;
+		coded_ratio = HDMI_FC_AVICONF1_CODED_ASPECT_RATIO_16_9;
+	} else {
+		act_ratio = HDMI_FC_AVICONF1_ACTIVE_ASPECT_RATIO_4_3;
+		coded_ratio = HDMI_FC_AVICONF1_CODED_ASPECT_RATIO_4_3;
+	}
+
+	/* Set up colorimetry */
+	if (hdmi->hdmi_data.enc_out_format == XVYCC444) {
+		colorimetry = HDMI_FC_AVICONF1_COLORIMETRY_EXTENDED_INFO;
+		if (hdmi->hdmi_data.colorimetry == HDMI_COLORIMETRY_ITU_601)
+			ext_colorimetry =
+				HDMI_FC_AVICONF2_EXT_COLORIMETRY_XVYCC601;
+		else /*hdmi->hdmi_data.colorimetry == HDMI_COLORIMETRY_ITU_709*/
+			ext_colorimetry =
+				HDMI_FC_AVICONF2_EXT_COLORIMETRY_XVYCC709;
+	} else if (hdmi->hdmi_data.enc_out_format != RGB) {
+		if (hdmi->hdmi_data.colorimetry == HDMI_COLORIMETRY_ITU_601)
+			colorimetry = HDMI_FC_AVICONF1_COLORIMETRY_SMPTE;
+		else /*hdmi->hdmi_data.colorimetry == HDMI_COLORIMETRY_ITU_709*/
+			colorimetry = HDMI_FC_AVICONF1_COLORIMETRY_ITUR;
+		ext_colorimetry = HDMI_FC_AVICONF2_EXT_COLORIMETRY_XVYCC601;
+	} else { /* Carries no data */
+		colorimetry = HDMI_FC_AVICONF1_COLORIMETRY_NO_DATA;
+		ext_colorimetry = HDMI_FC_AVICONF2_EXT_COLORIMETRY_XVYCC601;
+	}
+
+	val = colorimetry | coded_ratio | act_ratio;
+	hdmi_writeb(hdmi, val, HDMI_FC_AVICONF1);
+
+	/* AVI Data Byte 3 */
+	val = HDMI_FC_AVICONF2_IT_CONTENT_NO_DATA | ext_colorimetry |
+		HDMI_FC_AVICONF2_RGB_QUANT_DEFAULT |
+		HDMI_FC_AVICONF2_SCALING_NONE;
+	hdmi_writeb(hdmi, val, HDMI_FC_AVICONF2);
+
+	/* AVI Data Byte 4 */
+	hdmi_writeb(hdmi, hdmi->vic, HDMI_FC_AVIVID);
+
+	/* AVI Data Byte 5- set up input and output pixel repetition */
+	val = (((hdmi->hdmi_data.video_mode.mpixelrepetitioninput + 1) <<
+		HDMI_FC_PRCONF_INCOMING_PR_FACTOR_OFFSET) &
+		HDMI_FC_PRCONF_INCOMING_PR_FACTOR_MASK) |
+		((hdmi->hdmi_data.video_mode.mpixelrepetitionoutput <<
+		HDMI_FC_PRCONF_OUTPUT_PR_FACTOR_OFFSET) &
+		HDMI_FC_PRCONF_OUTPUT_PR_FACTOR_MASK);
+	hdmi_writeb(hdmi, val, HDMI_FC_PRCONF);
+
+	/* IT Content and quantization range = don't care */
+	val = HDMI_FC_AVICONF3_IT_CONTENT_TYPE_GRAPHICS |
+		HDMI_FC_AVICONF3_QUANT_RANGE_LIMITED;
+	hdmi_writeb(hdmi, val, HDMI_FC_AVICONF3);
+
+	/* AVI Data Bytes 6-13 */
+	hdmi_writeb(hdmi, 0, HDMI_FC_AVIETB0);
+	hdmi_writeb(hdmi, 0, HDMI_FC_AVIETB1);
+	hdmi_writeb(hdmi, 0, HDMI_FC_AVISBB0);
+	hdmi_writeb(hdmi, 0, HDMI_FC_AVISBB1);
+	hdmi_writeb(hdmi, 0, HDMI_FC_AVIELB0);
+	hdmi_writeb(hdmi, 0, HDMI_FC_AVIELB1);
+	hdmi_writeb(hdmi, 0, HDMI_FC_AVISRB0);
+	hdmi_writeb(hdmi, 0, HDMI_FC_AVISRB1);
+}
+
+static void hdmi_av_composer(struct imx_hdmi *hdmi,
+			     const struct drm_display_mode *mode)
+{
+	u8 inv_val;
+	struct hdmi_vmode *vmode = &hdmi->hdmi_data.video_mode;
+	int hblank, vblank, h_de_hs, v_de_vs, hsync_len, vsync_len;
+
+	vmode->mhsyncpolarity = !!(mode->flags & DRM_MODE_FLAG_PHSYNC);
+	vmode->mvsyncpolarity = !!(mode->flags & DRM_MODE_FLAG_PVSYNC);
+	vmode->minterlaced = !!(mode->flags & DRM_MODE_FLAG_INTERLACE);
+	vmode->mpixelclock = mode->clock * 1000;
+
+	dev_dbg(hdmi->dev, "final pixclk = %d\n", vmode->mpixelclock);
+
+	/* Set up HDMI_FC_INVIDCONF */
+	inv_val = (hdmi->hdmi_data.hdcp_enable ?
+		HDMI_FC_INVIDCONF_HDCP_KEEPOUT_ACTIVE :
+		HDMI_FC_INVIDCONF_HDCP_KEEPOUT_INACTIVE);
+
+	inv_val |= (vmode->mvsyncpolarity ?
+		HDMI_FC_INVIDCONF_VSYNC_IN_POLARITY_ACTIVE_HIGH :
+		HDMI_FC_INVIDCONF_VSYNC_IN_POLARITY_ACTIVE_LOW);
+
+	inv_val |= (vmode->mhsyncpolarity ?
+		HDMI_FC_INVIDCONF_HSYNC_IN_POLARITY_ACTIVE_HIGH :
+		HDMI_FC_INVIDCONF_HSYNC_IN_POLARITY_ACTIVE_LOW);
+
+	inv_val |= (vmode->mdataenablepolarity ?
+		HDMI_FC_INVIDCONF_DE_IN_POLARITY_ACTIVE_HIGH :
+		HDMI_FC_INVIDCONF_DE_IN_POLARITY_ACTIVE_LOW);
+
+	if (hdmi->vic == 39)
+		inv_val |= HDMI_FC_INVIDCONF_R_V_BLANK_IN_OSC_ACTIVE_HIGH;
+	else
+		inv_val |= (vmode->minterlaced ?
+			HDMI_FC_INVIDCONF_R_V_BLANK_IN_OSC_ACTIVE_HIGH :
+			HDMI_FC_INVIDCONF_R_V_BLANK_IN_OSC_ACTIVE_LOW);
+
+	inv_val |= (vmode->minterlaced ?
+		HDMI_FC_INVIDCONF_IN_I_P_INTERLACED :
+		HDMI_FC_INVIDCONF_IN_I_P_PROGRESSIVE);
+
+	inv_val |= (vmode->mdvi ?
+		HDMI_FC_INVIDCONF_DVI_MODEZ_DVI_MODE :
+		HDMI_FC_INVIDCONF_DVI_MODEZ_HDMI_MODE);
+
+	hdmi_writeb(hdmi, inv_val, HDMI_FC_INVIDCONF);
+
+	/* Set up horizontal active pixel width */
+	hdmi_writeb(hdmi, mode->hdisplay >> 8, HDMI_FC_INHACTV1);
+	hdmi_writeb(hdmi, mode->hdisplay, HDMI_FC_INHACTV0);
+
+	/* Set up vertical active lines */
+	hdmi_writeb(hdmi, mode->vdisplay >> 8, HDMI_FC_INVACTV1);
+	hdmi_writeb(hdmi, mode->vdisplay, HDMI_FC_INVACTV0);
+
+	/* Set up horizontal blanking pixel region width */
+	hblank = mode->htotal - mode->hdisplay;
+	hdmi_writeb(hdmi, hblank >> 8, HDMI_FC_INHBLANK1);
+	hdmi_writeb(hdmi, hblank, HDMI_FC_INHBLANK0);
+
+	/* Set up vertical blanking pixel region width */
+	vblank = mode->vtotal - mode->vdisplay;
+	hdmi_writeb(hdmi, vblank, HDMI_FC_INVBLANK);
+
+	/* Set up HSYNC active edge delay width (in pixel clks) */
+	h_de_hs = mode->hsync_start - mode->hdisplay;
+	hdmi_writeb(hdmi, h_de_hs >> 8, HDMI_FC_HSYNCINDELAY1);
+	hdmi_writeb(hdmi, h_de_hs, HDMI_FC_HSYNCINDELAY0);
+
+	/* Set up VSYNC active edge delay (in lines) */
+	v_de_vs = mode->vsync_start - mode->vdisplay;
+	hdmi_writeb(hdmi, v_de_vs, HDMI_FC_VSYNCINDELAY);
+
+	/* Set up HSYNC active pulse width (in pixel clks) */
+	hsync_len = mode->hsync_end - mode->hsync_start;
+	hdmi_writeb(hdmi, hsync_len >> 8, HDMI_FC_HSYNCINWIDTH1);
+	hdmi_writeb(hdmi, hsync_len, HDMI_FC_HSYNCINWIDTH0);
+
+	/* Set up VSYNC active edge delay (in lines) */
+	vsync_len = mode->vsync_end - mode->vsync_start;
+	hdmi_writeb(hdmi, vsync_len, HDMI_FC_VSYNCINWIDTH);
+}
+
+static void imx_hdmi_phy_disable(struct imx_hdmi *hdmi)
+{
+	if (!hdmi->phy_enabled)
+		return;
+
+	imx_hdmi_phy_enable_tmds(hdmi, 0);
+	imx_hdmi_phy_enable_power(hdmi, 0);
+
+	hdmi->phy_enabled = false;
+}
+
+/* HDMI Initialization Step B.4 */
+static void imx_hdmi_enable_video_path(struct imx_hdmi *hdmi)
+{
+	u8 clkdis;
+
+	/* control period minimum duration */
+	hdmi_writeb(hdmi, 12, HDMI_FC_CTRLDUR);
+	hdmi_writeb(hdmi, 32, HDMI_FC_EXCTRLDUR);
+	hdmi_writeb(hdmi, 1, HDMI_FC_EXCTRLSPAC);
+
+	/* Set to fill TMDS data channels */
+	hdmi_writeb(hdmi, 0x0B, HDMI_FC_CH0PREAM);
+	hdmi_writeb(hdmi, 0x16, HDMI_FC_CH1PREAM);
+	hdmi_writeb(hdmi, 0x21, HDMI_FC_CH2PREAM);
+
+	/* Enable pixel clock and tmds data path */
+	clkdis = 0x7F;
+	clkdis &= ~HDMI_MC_CLKDIS_PIXELCLK_DISABLE;
+	hdmi_writeb(hdmi, clkdis, HDMI_MC_CLKDIS);
+
+	clkdis &= ~HDMI_MC_CLKDIS_TMDSCLK_DISABLE;
+	hdmi_writeb(hdmi, clkdis, HDMI_MC_CLKDIS);
+
+	/* Enable csc path */
+	if (is_color_space_conversion(hdmi)) {
+		clkdis &= ~HDMI_MC_CLKDIS_CSCCLK_DISABLE;
+		hdmi_writeb(hdmi, clkdis, HDMI_MC_CLKDIS);
+	}
+}
+
+static void hdmi_enable_audio_clk(struct imx_hdmi *hdmi)
+{
+	hdmi_modb(hdmi, 0, HDMI_MC_CLKDIS_AUDCLK_DISABLE, HDMI_MC_CLKDIS);
+}
+
+/* Workaround to clear the overflow condition */
+static void imx_hdmi_clear_overflow(struct imx_hdmi *hdmi)
+{
+	int count;
+	u8 val;
+
+	/* TMDS software reset */
+	hdmi_writeb(hdmi, (u8)~HDMI_MC_SWRSTZ_TMDSSWRST_REQ, HDMI_MC_SWRSTZ);
+
+	val = hdmi_readb(hdmi, HDMI_FC_INVIDCONF);
+	if (hdmi->dev_type == IMX6DL_HDMI) {
+		hdmi_writeb(hdmi, val, HDMI_FC_INVIDCONF);
+		return;
+	}
+
+	for (count = 0; count < 4; count++)
+		hdmi_writeb(hdmi, val, HDMI_FC_INVIDCONF);
+}
+
+static void hdmi_enable_overflow_interrupts(struct imx_hdmi *hdmi)
+{
+	hdmi_writeb(hdmi, 0, HDMI_FC_MASK2);
+	hdmi_writeb(hdmi, 0, HDMI_IH_MUTE_FC_STAT2);
+}
+
+static void hdmi_disable_overflow_interrupts(struct imx_hdmi *hdmi)
+{
+	hdmi_writeb(hdmi, HDMI_IH_MUTE_FC_STAT2_OVERFLOW_MASK,
+		    HDMI_IH_MUTE_FC_STAT2);
+}
+
+static int imx_hdmi_setup(struct imx_hdmi *hdmi, struct drm_display_mode *mode)
+{
+	int ret;
+
+	hdmi_disable_overflow_interrupts(hdmi);
+
+	hdmi->vic = drm_match_cea_mode(mode);
+
+	if (!hdmi->vic) {
+		dev_dbg(hdmi->dev, "Non-CEA mode used in HDMI\n");
+		hdmi->hdmi_data.video_mode.mdvi = true;
+	} else {
+		dev_dbg(hdmi->dev, "CEA mode used vic=%d\n", hdmi->vic);
+		hdmi->hdmi_data.video_mode.mdvi = false;
+	}
+
+	if ((hdmi->vic == 6) || (hdmi->vic == 7) ||
+		(hdmi->vic == 21) || (hdmi->vic == 22) ||
+		(hdmi->vic == 2) || (hdmi->vic == 3) ||
+		(hdmi->vic == 17) || (hdmi->vic == 18))
+		hdmi->hdmi_data.colorimetry = HDMI_COLORIMETRY_ITU_601;
+	else
+		hdmi->hdmi_data.colorimetry = HDMI_COLORIMETRY_ITU_709;
+
+	if ((hdmi->vic == 10) || (hdmi->vic == 11) ||
+		(hdmi->vic == 12) || (hdmi->vic == 13) ||
+		(hdmi->vic == 14) || (hdmi->vic == 15) ||
+		(hdmi->vic == 25) || (hdmi->vic == 26) ||
+		(hdmi->vic == 27) || (hdmi->vic == 28) ||
+		(hdmi->vic == 29) || (hdmi->vic == 30) ||
+		(hdmi->vic == 35) || (hdmi->vic == 36) ||
+		(hdmi->vic == 37) || (hdmi->vic == 38))
+		hdmi->hdmi_data.video_mode.mpixelrepetitionoutput = 1;
+	else
+		hdmi->hdmi_data.video_mode.mpixelrepetitionoutput = 0;
+
+	hdmi->hdmi_data.video_mode.mpixelrepetitioninput = 0;
+
+	/* TODO: Get input format from IPU (via FB driver interface) */
+	hdmi->hdmi_data.enc_in_format = RGB;
+
+	hdmi->hdmi_data.enc_out_format = RGB;
+
+	hdmi->hdmi_data.enc_color_depth = 8;
+	hdmi->hdmi_data.pix_repet_factor = 0;
+	hdmi->hdmi_data.hdcp_enable = 0;
+	hdmi->hdmi_data.video_mode.mdataenablepolarity = true;
+
+	/* HDMI Initialization Step B.1 */
+	hdmi_av_composer(hdmi, mode);
+
+	/* HDMI Initializateion Step B.2 */
+	ret = imx_hdmi_phy_init(hdmi);
+	if (ret)
+		return ret;
+
+	/* HDMI Initialization Step B.3 */
+	imx_hdmi_enable_video_path(hdmi);
+
+	/* not for DVI mode */
+	if (hdmi->hdmi_data.video_mode.mdvi)
+		dev_dbg(hdmi->dev, "%s DVI mode\n", __func__);
+	else {
+		dev_dbg(hdmi->dev, "%s CEA mode\n", __func__);
+
+		/* HDMI Initialization Step E - Configure audio */
+		hdmi_clk_regenerator_update_pixel_clock(hdmi);
+		hdmi_enable_audio_clk(hdmi);
+
+		/* HDMI Initialization Step F - Configure AVI InfoFrame */
+		hdmi_config_AVI(hdmi);
+	}
+
+	hdmi_video_packetize(hdmi);
+	hdmi_video_csc(hdmi);
+	hdmi_video_sample(hdmi);
+	hdmi_tx_hdcp_config(hdmi);
+
+	imx_hdmi_clear_overflow(hdmi);
+	if (hdmi->cable_plugin && !hdmi->hdmi_data.video_mode.mdvi)
+		hdmi_enable_overflow_interrupts(hdmi);
+
+	return 0;
+}
+
+/* Wait until we are registered to enable interrupts */
+static int imx_hdmi_fb_registered(struct imx_hdmi *hdmi)
+{
+	hdmi_writeb(hdmi, HDMI_PHY_I2CM_INT_ADDR_DONE_POL,
+		    HDMI_PHY_I2CM_INT_ADDR);
+
+	hdmi_writeb(hdmi, HDMI_PHY_I2CM_CTLINT_ADDR_NAC_POL |
+		    HDMI_PHY_I2CM_CTLINT_ADDR_ARBITRATION_POL,
+		    HDMI_PHY_I2CM_CTLINT_ADDR);
+
+	/* enable cable hot plug irq */
+	hdmi_writeb(hdmi, (u8)~HDMI_PHY_HPD, HDMI_PHY_MASK0);
+
+	/* Clear Hotplug interrupts */
+	hdmi_writeb(hdmi, HDMI_IH_PHY_STAT0_HPD, HDMI_IH_PHY_STAT0);
+
+	return 0;
+}
+
+static void initialize_hdmi_ih_mutes(struct imx_hdmi *hdmi)
+{
+	u8 ih_mute;
+
+	/*
+	 * Boot up defaults are:
+	 * HDMI_IH_MUTE   = 0x03 (disabled)
+	 * HDMI_IH_MUTE_* = 0x00 (enabled)
+	 *
+	 * Disable top level interrupt bits in HDMI block
+	 */
+	ih_mute = hdmi_readb(hdmi, HDMI_IH_MUTE) |
+		  HDMI_IH_MUTE_MUTE_WAKEUP_INTERRUPT |
+		  HDMI_IH_MUTE_MUTE_ALL_INTERRUPT;
+
+	hdmi_writeb(hdmi, ih_mute, HDMI_IH_MUTE);
+
+	/* by default mask all interrupts */
+	hdmi_writeb(hdmi, 0xff, HDMI_VP_MASK);
+	hdmi_writeb(hdmi, 0xff, HDMI_FC_MASK0);
+	hdmi_writeb(hdmi, 0xff, HDMI_FC_MASK1);
+	hdmi_writeb(hdmi, 0xff, HDMI_FC_MASK2);
+	hdmi_writeb(hdmi, 0xff, HDMI_PHY_MASK0);
+	hdmi_writeb(hdmi, 0xff, HDMI_PHY_I2CM_INT_ADDR);
+	hdmi_writeb(hdmi, 0xff, HDMI_PHY_I2CM_CTLINT_ADDR);
+	hdmi_writeb(hdmi, 0xff, HDMI_AUD_INT);
+	hdmi_writeb(hdmi, 0xff, HDMI_AUD_SPDIFINT);
+	hdmi_writeb(hdmi, 0xff, HDMI_AUD_HBR_MASK);
+	hdmi_writeb(hdmi, 0xff, HDMI_GP_MASK);
+	hdmi_writeb(hdmi, 0xff, HDMI_A_APIINTMSK);
+	hdmi_writeb(hdmi, 0xff, HDMI_CEC_MASK);
+	hdmi_writeb(hdmi, 0xff, HDMI_I2CM_INT);
+	hdmi_writeb(hdmi, 0xff, HDMI_I2CM_CTLINT);
+
+	/* Disable interrupts in the IH_MUTE_* registers */
+	hdmi_writeb(hdmi, 0xff, HDMI_IH_MUTE_FC_STAT0);
+	hdmi_writeb(hdmi, 0xff, HDMI_IH_MUTE_FC_STAT1);
+	hdmi_writeb(hdmi, 0xff, HDMI_IH_MUTE_FC_STAT2);
+	hdmi_writeb(hdmi, 0xff, HDMI_IH_MUTE_AS_STAT0);
+	hdmi_writeb(hdmi, 0xff, HDMI_IH_MUTE_PHY_STAT0);
+	hdmi_writeb(hdmi, 0xff, HDMI_IH_MUTE_I2CM_STAT0);
+	hdmi_writeb(hdmi, 0xff, HDMI_IH_MUTE_CEC_STAT0);
+	hdmi_writeb(hdmi, 0xff, HDMI_IH_MUTE_VP_STAT0);
+	hdmi_writeb(hdmi, 0xff, HDMI_IH_MUTE_I2CMPHY_STAT0);
+	hdmi_writeb(hdmi, 0xff, HDMI_IH_MUTE_AHBDMAAUD_STAT0);
+
+	/* Enable top level interrupt bits in HDMI block */
+	ih_mute &= ~(HDMI_IH_MUTE_MUTE_WAKEUP_INTERRUPT |
+		    HDMI_IH_MUTE_MUTE_ALL_INTERRUPT);
+	hdmi_writeb(hdmi, ih_mute, HDMI_IH_MUTE);
+}
+
+static void imx_hdmi_poweron(struct imx_hdmi *hdmi)
+{
+	imx_hdmi_setup(hdmi, &hdmi->previous_mode);
+}
+
+static void imx_hdmi_poweroff(struct imx_hdmi *hdmi)
+{
+	imx_hdmi_phy_disable(hdmi);
+}
+
+static enum drm_connector_status imx_hdmi_connector_detect(struct drm_connector
+							*connector, bool force)
+{
+	struct imx_hdmi *hdmi = container_of(connector, struct imx_hdmi,
+					     connector);
+
+	return hdmi_readb(hdmi, HDMI_PHY_STAT0) & HDMI_PHY_HPD ?
+		connector_status_connected : connector_status_disconnected;
+}
+
+static int imx_hdmi_connector_get_modes(struct drm_connector *connector)
+{
+	struct imx_hdmi *hdmi = container_of(connector, struct imx_hdmi,
+					     connector);
+	struct edid *edid;
+	int ret;
+
+	if (!hdmi->ddc)
+		return 0;
+
+	edid = drm_get_edid(connector, hdmi->ddc);
+	if (edid) {
+		dev_dbg(hdmi->dev, "got edid: width[%d] x height[%d]\n",
+			edid->width_cm, edid->height_cm);
+
+		drm_mode_connector_update_edid_property(connector, edid);
+		ret = drm_add_edid_modes(connector, edid);
+		kfree(edid);
+	} else {
+		dev_dbg(hdmi->dev, "failed to get edid\n");
+	}
+
+	return 0;
+}
+
+static struct drm_encoder *imx_hdmi_connector_best_encoder(struct drm_connector
+							   *connector)
+{
+	struct imx_hdmi *hdmi = container_of(connector, struct imx_hdmi,
+					     connector);
+
+	return &hdmi->encoder;
+}
+
+static void imx_hdmi_encoder_mode_set(struct drm_encoder *encoder,
+			struct drm_display_mode *mode,
+			struct drm_display_mode *adjusted_mode)
+{
+	struct imx_hdmi *hdmi = container_of(encoder, struct imx_hdmi, encoder);
+
+	imx_hdmi_setup(hdmi, mode);
+
+	/* Store the display mode for plugin/DKMS poweron events */
+	memcpy(&hdmi->previous_mode, mode, sizeof(hdmi->previous_mode));
+}
+
+static bool imx_hdmi_encoder_mode_fixup(struct drm_encoder *encoder,
+			const struct drm_display_mode *mode,
+			struct drm_display_mode *adjusted_mode)
+{
+	return true;
+}
+
+static void imx_hdmi_encoder_disable(struct drm_encoder *encoder)
+{
+}
+
+static void imx_hdmi_encoder_dpms(struct drm_encoder *encoder, int mode)
+{
+	struct imx_hdmi *hdmi = container_of(encoder, struct imx_hdmi, encoder);
+
+	if (mode)
+		imx_hdmi_poweroff(hdmi);
+	else
+		imx_hdmi_poweron(hdmi);
+}
+
+static void imx_hdmi_encoder_prepare(struct drm_encoder *encoder)
+{
+	struct imx_hdmi *hdmi = container_of(encoder, struct imx_hdmi, encoder);
+
+	imx_hdmi_poweroff(hdmi);
+	imx_drm_panel_format(encoder, V4L2_PIX_FMT_RGB24);
+}
+
+static void imx_hdmi_encoder_commit(struct drm_encoder *encoder)
+{
+	struct imx_hdmi *hdmi = container_of(encoder, struct imx_hdmi, encoder);
+	int mux = imx_drm_encoder_get_mux_id(hdmi->dev->of_node, encoder);
+
+	imx_hdmi_set_ipu_di_mux(hdmi, mux);
+
+	imx_hdmi_poweron(hdmi);
+}
+
+static struct drm_encoder_funcs imx_hdmi_encoder_funcs = {
+	.destroy = imx_drm_encoder_destroy,
+};
+
+static struct drm_encoder_helper_funcs imx_hdmi_encoder_helper_funcs = {
+	.dpms = imx_hdmi_encoder_dpms,
+	.prepare = imx_hdmi_encoder_prepare,
+	.commit = imx_hdmi_encoder_commit,
+	.mode_set = imx_hdmi_encoder_mode_set,
+	.mode_fixup = imx_hdmi_encoder_mode_fixup,
+	.disable = imx_hdmi_encoder_disable,
+};
+
+static struct drm_connector_funcs imx_hdmi_connector_funcs = {
+	.dpms = drm_helper_connector_dpms,
+	.fill_modes = drm_helper_probe_single_connector_modes,
+	.detect = imx_hdmi_connector_detect,
+	.destroy = imx_drm_connector_destroy,
+};
+
+static struct drm_connector_helper_funcs imx_hdmi_connector_helper_funcs = {
+	.get_modes = imx_hdmi_connector_get_modes,
+	.best_encoder = imx_hdmi_connector_best_encoder,
+};
+
+static irqreturn_t imx_hdmi_hardirq(int irq, void *dev_id)
+{
+	struct imx_hdmi *hdmi = dev_id;
+	u8 intr_stat;
+
+	intr_stat = hdmi_readb(hdmi, HDMI_IH_PHY_STAT0);
+	if (intr_stat)
+		hdmi_writeb(hdmi, ~0, HDMI_IH_MUTE_PHY_STAT0);
+
+	return intr_stat ? IRQ_WAKE_THREAD : IRQ_NONE;
+}
+
+static irqreturn_t imx_hdmi_irq(int irq, void *dev_id)
+{
+	struct imx_hdmi *hdmi = dev_id;
+	u8 intr_stat;
+	u8 phy_int_pol;
+
+	intr_stat = hdmi_readb(hdmi, HDMI_IH_PHY_STAT0);
+
+	phy_int_pol = hdmi_readb(hdmi, HDMI_PHY_POL0);
+
+	if (intr_stat & HDMI_IH_PHY_STAT0_HPD) {
+		if (phy_int_pol & HDMI_PHY_HPD) {
+			dev_dbg(hdmi->dev, "EVENT=plugin\n");
+
+			hdmi_modb(hdmi, 0, HDMI_PHY_HPD, HDMI_PHY_POL0);
+
+			imx_hdmi_poweron(hdmi);
+		} else {
+			dev_dbg(hdmi->dev, "EVENT=plugout\n");
+
+			hdmi_modb(hdmi, HDMI_PHY_HPD, HDMI_PHY_HPD,
+				HDMI_PHY_POL0);
+
+			imx_hdmi_poweroff(hdmi);
+		}
+		drm_helper_hpd_irq_event(hdmi->connector.dev);
+	}
+
+	hdmi_writeb(hdmi, intr_stat, HDMI_IH_PHY_STAT0);
+	hdmi_writeb(hdmi, ~HDMI_IH_PHY_STAT0_HPD, HDMI_IH_MUTE_PHY_STAT0);
+
+	return IRQ_HANDLED;
+}
+
+static int imx_hdmi_register(struct drm_device *drm, struct imx_hdmi *hdmi)
+{
+	int ret;
+
+	ret = imx_drm_encoder_parse_of(drm, &hdmi->encoder,
+				       hdmi->dev->of_node);
+	if (ret)
+		return ret;
+
+	hdmi->connector.polled = DRM_CONNECTOR_POLL_HPD;
+
+	drm_encoder_helper_add(&hdmi->encoder, &imx_hdmi_encoder_helper_funcs);
+	drm_encoder_init(drm, &hdmi->encoder, &imx_hdmi_encoder_funcs,
+			 DRM_MODE_ENCODER_TMDS);
+
+	drm_connector_helper_add(&hdmi->connector,
+			&imx_hdmi_connector_helper_funcs);
+	drm_connector_init(drm, &hdmi->connector, &imx_hdmi_connector_funcs,
+			   DRM_MODE_CONNECTOR_HDMIA);
+
+	hdmi->connector.encoder = &hdmi->encoder;
+
+	drm_mode_connector_attach_encoder(&hdmi->connector, &hdmi->encoder);
+
+	return 0;
+}
+
+static struct platform_device_id imx_hdmi_devtype[] = {
+	{
+		.name = "imx6q-hdmi",
+		.driver_data = IMX6Q_HDMI,
+	}, {
+		.name = "imx6dl-hdmi",
+		.driver_data = IMX6DL_HDMI,
+	}, { /* sentinel */ }
+};
+MODULE_DEVICE_TABLE(platform, imx_hdmi_devtype);
+
+static const struct of_device_id imx_hdmi_dt_ids[] = {
+{ .compatible = "fsl,imx6q-hdmi", .data = &imx_hdmi_devtype[IMX6Q_HDMI], },
+{ .compatible = "fsl,imx6dl-hdmi", .data = &imx_hdmi_devtype[IMX6DL_HDMI], },
+{ /* sentinel */ }
+};
+MODULE_DEVICE_TABLE(of, imx_hdmi_dt_ids);
+
+static int imx_hdmi_bind(struct device *dev, struct device *master, void *data)
+{
+	struct platform_device *pdev = to_platform_device(dev);
+	const struct of_device_id *of_id =
+				of_match_device(imx_hdmi_dt_ids, dev);
+	struct drm_device *drm = data;
+	struct device_node *np = dev->of_node;
+	struct device_node *ddc_node;
+	struct imx_hdmi *hdmi;
+	struct resource *iores;
+	int ret, irq;
+
+	hdmi = devm_kzalloc(dev, sizeof(*hdmi), GFP_KERNEL);
+	if (!hdmi)
+		return -ENOMEM;
+
+	hdmi->dev = dev;
+	hdmi->sample_rate = 48000;
+	hdmi->ratio = 100;
+
+	if (of_id) {
+		const struct platform_device_id *device_id = of_id->data;
+
+		hdmi->dev_type = device_id->driver_data;
+	}
+
+	ddc_node = of_parse_phandle(np, "ddc-i2c-bus", 0);
+	if (ddc_node) {
+		hdmi->ddc = of_find_i2c_adapter_by_node(ddc_node);
+		if (!hdmi->ddc)
+			dev_dbg(hdmi->dev, "failed to read ddc node\n");
+
+		of_node_put(ddc_node);
+	} else {
+		dev_dbg(hdmi->dev, "no ddc property found\n");
+	}
+
+	irq = platform_get_irq(pdev, 0);
+	if (irq < 0)
+		return irq;
+
+	ret = devm_request_threaded_irq(dev, irq, imx_hdmi_hardirq,
+					imx_hdmi_irq, IRQF_SHARED,
+					dev_name(dev), hdmi);
+	if (ret)
+		return ret;
+
+	iores = platform_get_resource(pdev, IORESOURCE_MEM, 0);
+	hdmi->regs = devm_ioremap_resource(dev, iores);
+	if (IS_ERR(hdmi->regs))
+		return PTR_ERR(hdmi->regs);
+
+	hdmi->regmap = syscon_regmap_lookup_by_phandle(np, "gpr");
+	if (IS_ERR(hdmi->regmap))
+		return PTR_ERR(hdmi->regmap);
+
+	hdmi->isfr_clk = devm_clk_get(hdmi->dev, "isfr");
+	if (IS_ERR(hdmi->isfr_clk)) {
+		ret = PTR_ERR(hdmi->isfr_clk);
+		dev_err(hdmi->dev,
+			"Unable to get HDMI isfr clk: %d\n", ret);
+		return ret;
+	}
+
+	ret = clk_prepare_enable(hdmi->isfr_clk);
+	if (ret) {
+		dev_err(hdmi->dev,
+			"Cannot enable HDMI isfr clock: %d\n", ret);
+		return ret;
+	}
+
+	hdmi->iahb_clk = devm_clk_get(hdmi->dev, "iahb");
+	if (IS_ERR(hdmi->iahb_clk)) {
+		ret = PTR_ERR(hdmi->iahb_clk);
+		dev_err(hdmi->dev,
+			"Unable to get HDMI iahb clk: %d\n", ret);
+		goto err_isfr;
+	}
+
+	ret = clk_prepare_enable(hdmi->iahb_clk);
+	if (ret) {
+		dev_err(hdmi->dev,
+			"Cannot enable HDMI iahb clock: %d\n", ret);
+		goto err_isfr;
+	}
+
+	/* Product and revision IDs */
+	dev_info(dev,
+		"Detected HDMI controller 0x%x:0x%x:0x%x:0x%x\n",
+		hdmi_readb(hdmi, HDMI_DESIGN_ID),
+		hdmi_readb(hdmi, HDMI_REVISION_ID),
+		hdmi_readb(hdmi, HDMI_PRODUCT_ID0),
+		hdmi_readb(hdmi, HDMI_PRODUCT_ID1));
+
+	initialize_hdmi_ih_mutes(hdmi);
+
+	/*
+	 * To prevent overflows in HDMI_IH_FC_STAT2, set the clk regenerator
+	 * N and cts values before enabling phy
+	 */
+	hdmi_init_clk_regenerator(hdmi);
+
+	/*
+	 * Configure registers related to HDMI interrupt
+	 * generation before registering IRQ.
+	 */
+	hdmi_writeb(hdmi, HDMI_PHY_HPD, HDMI_PHY_POL0);
+
+	/* Clear Hotplug interrupts */
+	hdmi_writeb(hdmi, HDMI_IH_PHY_STAT0_HPD, HDMI_IH_PHY_STAT0);
+
+	ret = imx_hdmi_fb_registered(hdmi);
+	if (ret)
+		goto err_iahb;
+
+	ret = imx_hdmi_register(drm, hdmi);
+	if (ret)
+		goto err_iahb;
+
+	/* Unmute interrupts */
+	hdmi_writeb(hdmi, ~HDMI_IH_PHY_STAT0_HPD, HDMI_IH_MUTE_PHY_STAT0);
+
+	dev_set_drvdata(dev, hdmi);
+
+	return 0;
+
+err_iahb:
+	clk_disable_unprepare(hdmi->iahb_clk);
+err_isfr:
+	clk_disable_unprepare(hdmi->isfr_clk);
+
+	return ret;
+}
+
+static void imx_hdmi_unbind(struct device *dev, struct device *master,
+	void *data)
+{
+	struct imx_hdmi *hdmi = dev_get_drvdata(dev);
+
+	/* Disable all interrupts */
+	hdmi_writeb(hdmi, ~0, HDMI_IH_MUTE_PHY_STAT0);
+
+	hdmi->connector.funcs->destroy(&hdmi->connector);
+	hdmi->encoder.funcs->destroy(&hdmi->encoder);
+
+	clk_disable_unprepare(hdmi->iahb_clk);
+	clk_disable_unprepare(hdmi->isfr_clk);
+	i2c_put_adapter(hdmi->ddc);
+}
+
+static const struct component_ops hdmi_ops = {
+	.bind	= imx_hdmi_bind,
+	.unbind	= imx_hdmi_unbind,
+};
+
+static int imx_hdmi_platform_probe(struct platform_device *pdev)
+{
+	return component_add(&pdev->dev, &hdmi_ops);
+}
+
+static int imx_hdmi_platform_remove(struct platform_device *pdev)
+{
+	component_del(&pdev->dev, &hdmi_ops);
+	return 0;
+}
+
+static struct platform_driver imx_hdmi_driver = {
+	.probe  = imx_hdmi_platform_probe,
+	.remove = imx_hdmi_platform_remove,
+	.driver = {
+		.name = "imx-hdmi",
+		.of_match_table = imx_hdmi_dt_ids,
+	},
+};
+
+module_platform_driver(imx_hdmi_driver);
+
+MODULE_AUTHOR("Sascha Hauer <s.hauer@pengutronix.de>");
+MODULE_DESCRIPTION("i.MX6 HDMI transmitter driver");
+MODULE_LICENSE("GPL");
+MODULE_ALIAS("platform:imx-hdmi");
diff --git a/drivers/gpu/drm/imx/imx-hdmi.h b/drivers/gpu/drm/imx/imx-hdmi.h
new file mode 100644
index 000000000000..39b677689db6
--- /dev/null
+++ b/drivers/gpu/drm/imx/imx-hdmi.h
@@ -0,0 +1,1032 @@
+/*
+ * Copyright (C) 2011 Freescale Semiconductor, Inc.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ */
+
+#ifndef __IMX_HDMI_H__
+#define __IMX_HDMI_H__
+
+/* Identification Registers */
+#define HDMI_DESIGN_ID                          0x0000
+#define HDMI_REVISION_ID                        0x0001
+#define HDMI_PRODUCT_ID0                        0x0002
+#define HDMI_PRODUCT_ID1                        0x0003
+#define HDMI_CONFIG0_ID                         0x0004
+#define HDMI_CONFIG1_ID                         0x0005
+#define HDMI_CONFIG2_ID                         0x0006
+#define HDMI_CONFIG3_ID                         0x0007
+
+/* Interrupt Registers */
+#define HDMI_IH_FC_STAT0                        0x0100
+#define HDMI_IH_FC_STAT1                        0x0101
+#define HDMI_IH_FC_STAT2                        0x0102
+#define HDMI_IH_AS_STAT0                        0x0103
+#define HDMI_IH_PHY_STAT0                       0x0104
+#define HDMI_IH_I2CM_STAT0                      0x0105
+#define HDMI_IH_CEC_STAT0                       0x0106
+#define HDMI_IH_VP_STAT0                        0x0107
+#define HDMI_IH_I2CMPHY_STAT0                   0x0108
+#define HDMI_IH_AHBDMAAUD_STAT0                 0x0109
+
+#define HDMI_IH_MUTE_FC_STAT0                   0x0180
+#define HDMI_IH_MUTE_FC_STAT1                   0x0181
+#define HDMI_IH_MUTE_FC_STAT2                   0x0182
+#define HDMI_IH_MUTE_AS_STAT0                   0x0183
+#define HDMI_IH_MUTE_PHY_STAT0                  0x0184
+#define HDMI_IH_MUTE_I2CM_STAT0                 0x0185
+#define HDMI_IH_MUTE_CEC_STAT0                  0x0186
+#define HDMI_IH_MUTE_VP_STAT0                   0x0187
+#define HDMI_IH_MUTE_I2CMPHY_STAT0              0x0188
+#define HDMI_IH_MUTE_AHBDMAAUD_STAT0            0x0189
+#define HDMI_IH_MUTE                            0x01FF
+
+/* Video Sample Registers */
+#define HDMI_TX_INVID0                          0x0200
+#define HDMI_TX_INSTUFFING                      0x0201
+#define HDMI_TX_GYDATA0                         0x0202
+#define HDMI_TX_GYDATA1                         0x0203
+#define HDMI_TX_RCRDATA0                        0x0204
+#define HDMI_TX_RCRDATA1                        0x0205
+#define HDMI_TX_BCBDATA0                        0x0206
+#define HDMI_TX_BCBDATA1                        0x0207
+
+/* Video Packetizer Registers */
+#define HDMI_VP_STATUS                          0x0800
+#define HDMI_VP_PR_CD                           0x0801
+#define HDMI_VP_STUFF                           0x0802
+#define HDMI_VP_REMAP                           0x0803
+#define HDMI_VP_CONF                            0x0804
+#define HDMI_VP_STAT                            0x0805
+#define HDMI_VP_INT                             0x0806
+#define HDMI_VP_MASK                            0x0807
+#define HDMI_VP_POL                             0x0808
+
+/* Frame Composer Registers */
+#define HDMI_FC_INVIDCONF                       0x1000
+#define HDMI_FC_INHACTV0                        0x1001
+#define HDMI_FC_INHACTV1                        0x1002
+#define HDMI_FC_INHBLANK0                       0x1003
+#define HDMI_FC_INHBLANK1                       0x1004
+#define HDMI_FC_INVACTV0                        0x1005
+#define HDMI_FC_INVACTV1                        0x1006
+#define HDMI_FC_INVBLANK                        0x1007
+#define HDMI_FC_HSYNCINDELAY0                   0x1008
+#define HDMI_FC_HSYNCINDELAY1                   0x1009
+#define HDMI_FC_HSYNCINWIDTH0                   0x100A
+#define HDMI_FC_HSYNCINWIDTH1                   0x100B
+#define HDMI_FC_VSYNCINDELAY                    0x100C
+#define HDMI_FC_VSYNCINWIDTH                    0x100D
+#define HDMI_FC_INFREQ0                         0x100E
+#define HDMI_FC_INFREQ1                         0x100F
+#define HDMI_FC_INFREQ2                         0x1010
+#define HDMI_FC_CTRLDUR                         0x1011
+#define HDMI_FC_EXCTRLDUR                       0x1012
+#define HDMI_FC_EXCTRLSPAC                      0x1013
+#define HDMI_FC_CH0PREAM                        0x1014
+#define HDMI_FC_CH1PREAM                        0x1015
+#define HDMI_FC_CH2PREAM                        0x1016
+#define HDMI_FC_AVICONF3                        0x1017
+#define HDMI_FC_GCP                             0x1018
+#define HDMI_FC_AVICONF0                        0x1019
+#define HDMI_FC_AVICONF1                        0x101A
+#define HDMI_FC_AVICONF2                        0x101B
+#define HDMI_FC_AVIVID                          0x101C
+#define HDMI_FC_AVIETB0                         0x101D
+#define HDMI_FC_AVIETB1                         0x101E
+#define HDMI_FC_AVISBB0                         0x101F
+#define HDMI_FC_AVISBB1                         0x1020
+#define HDMI_FC_AVIELB0                         0x1021
+#define HDMI_FC_AVIELB1                         0x1022
+#define HDMI_FC_AVISRB0                         0x1023
+#define HDMI_FC_AVISRB1                         0x1024
+#define HDMI_FC_AUDICONF0                       0x1025
+#define HDMI_FC_AUDICONF1                       0x1026
+#define HDMI_FC_AUDICONF2                       0x1027
+#define HDMI_FC_AUDICONF3                       0x1028
+#define HDMI_FC_VSDIEEEID0                      0x1029
+#define HDMI_FC_VSDSIZE                         0x102A
+#define HDMI_FC_VSDIEEEID1                      0x1030
+#define HDMI_FC_VSDIEEEID2                      0x1031
+#define HDMI_FC_VSDPAYLOAD0                     0x1032
+#define HDMI_FC_VSDPAYLOAD1                     0x1033
+#define HDMI_FC_VSDPAYLOAD2                     0x1034
+#define HDMI_FC_VSDPAYLOAD3                     0x1035
+#define HDMI_FC_VSDPAYLOAD4                     0x1036
+#define HDMI_FC_VSDPAYLOAD5                     0x1037
+#define HDMI_FC_VSDPAYLOAD6                     0x1038
+#define HDMI_FC_VSDPAYLOAD7                     0x1039
+#define HDMI_FC_VSDPAYLOAD8                     0x103A
+#define HDMI_FC_VSDPAYLOAD9                     0x103B
+#define HDMI_FC_VSDPAYLOAD10                    0x103C
+#define HDMI_FC_VSDPAYLOAD11                    0x103D
+#define HDMI_FC_VSDPAYLOAD12                    0x103E
+#define HDMI_FC_VSDPAYLOAD13                    0x103F
+#define HDMI_FC_VSDPAYLOAD14                    0x1040
+#define HDMI_FC_VSDPAYLOAD15                    0x1041
+#define HDMI_FC_VSDPAYLOAD16                    0x1042
+#define HDMI_FC_VSDPAYLOAD17                    0x1043
+#define HDMI_FC_VSDPAYLOAD18                    0x1044
+#define HDMI_FC_VSDPAYLOAD19                    0x1045
+#define HDMI_FC_VSDPAYLOAD20                    0x1046
+#define HDMI_FC_VSDPAYLOAD21                    0x1047
+#define HDMI_FC_VSDPAYLOAD22                    0x1048
+#define HDMI_FC_VSDPAYLOAD23                    0x1049
+#define HDMI_FC_SPDVENDORNAME0                  0x104A
+#define HDMI_FC_SPDVENDORNAME1                  0x104B
+#define HDMI_FC_SPDVENDORNAME2                  0x104C
+#define HDMI_FC_SPDVENDORNAME3                  0x104D
+#define HDMI_FC_SPDVENDORNAME4                  0x104E
+#define HDMI_FC_SPDVENDORNAME5                  0x104F
+#define HDMI_FC_SPDVENDORNAME6                  0x1050
+#define HDMI_FC_SPDVENDORNAME7                  0x1051
+#define HDMI_FC_SDPPRODUCTNAME0                 0x1052
+#define HDMI_FC_SDPPRODUCTNAME1                 0x1053
+#define HDMI_FC_SDPPRODUCTNAME2                 0x1054
+#define HDMI_FC_SDPPRODUCTNAME3                 0x1055
+#define HDMI_FC_SDPPRODUCTNAME4                 0x1056
+#define HDMI_FC_SDPPRODUCTNAME5                 0x1057
+#define HDMI_FC_SDPPRODUCTNAME6                 0x1058
+#define HDMI_FC_SDPPRODUCTNAME7                 0x1059
+#define HDMI_FC_SDPPRODUCTNAME8                 0x105A
+#define HDMI_FC_SDPPRODUCTNAME9                 0x105B
+#define HDMI_FC_SDPPRODUCTNAME10                0x105C
+#define HDMI_FC_SDPPRODUCTNAME11                0x105D
+#define HDMI_FC_SDPPRODUCTNAME12                0x105E
+#define HDMI_FC_SDPPRODUCTNAME13                0x105F
+#define HDMI_FC_SDPPRODUCTNAME14                0x1060
+#define HDMI_FC_SPDPRODUCTNAME15                0x1061
+#define HDMI_FC_SPDDEVICEINF                    0x1062
+#define HDMI_FC_AUDSCONF                        0x1063
+#define HDMI_FC_AUDSSTAT                        0x1064
+#define HDMI_FC_DATACH0FILL                     0x1070
+#define HDMI_FC_DATACH1FILL                     0x1071
+#define HDMI_FC_DATACH2FILL                     0x1072
+#define HDMI_FC_CTRLQHIGH                       0x1073
+#define HDMI_FC_CTRLQLOW                        0x1074
+#define HDMI_FC_ACP0                            0x1075
+#define HDMI_FC_ACP28                           0x1076
+#define HDMI_FC_ACP27                           0x1077
+#define HDMI_FC_ACP26                           0x1078
+#define HDMI_FC_ACP25                           0x1079
+#define HDMI_FC_ACP24                           0x107A
+#define HDMI_FC_ACP23                           0x107B
+#define HDMI_FC_ACP22                           0x107C
+#define HDMI_FC_ACP21                           0x107D
+#define HDMI_FC_ACP20                           0x107E
+#define HDMI_FC_ACP19                           0x107F
+#define HDMI_FC_ACP18                           0x1080
+#define HDMI_FC_ACP17                           0x1081
+#define HDMI_FC_ACP16                           0x1082
+#define HDMI_FC_ACP15                           0x1083
+#define HDMI_FC_ACP14                           0x1084
+#define HDMI_FC_ACP13                           0x1085
+#define HDMI_FC_ACP12                           0x1086
+#define HDMI_FC_ACP11                           0x1087
+#define HDMI_FC_ACP10                           0x1088
+#define HDMI_FC_ACP9                            0x1089
+#define HDMI_FC_ACP8                            0x108A
+#define HDMI_FC_ACP7                            0x108B
+#define HDMI_FC_ACP6                            0x108C
+#define HDMI_FC_ACP5                            0x108D
+#define HDMI_FC_ACP4                            0x108E
+#define HDMI_FC_ACP3                            0x108F
+#define HDMI_FC_ACP2                            0x1090
+#define HDMI_FC_ACP1                            0x1091
+#define HDMI_FC_ISCR1_0                         0x1092
+#define HDMI_FC_ISCR1_16                        0x1093
+#define HDMI_FC_ISCR1_15                        0x1094
+#define HDMI_FC_ISCR1_14                        0x1095
+#define HDMI_FC_ISCR1_13                        0x1096
+#define HDMI_FC_ISCR1_12                        0x1097
+#define HDMI_FC_ISCR1_11                        0x1098
+#define HDMI_FC_ISCR1_10                        0x1099
+#define HDMI_FC_ISCR1_9                         0x109A
+#define HDMI_FC_ISCR1_8                         0x109B
+#define HDMI_FC_ISCR1_7                         0x109C
+#define HDMI_FC_ISCR1_6                         0x109D
+#define HDMI_FC_ISCR1_5                         0x109E
+#define HDMI_FC_ISCR1_4                         0x109F
+#define HDMI_FC_ISCR1_3                         0x10A0
+#define HDMI_FC_ISCR1_2                         0x10A1
+#define HDMI_FC_ISCR1_1                         0x10A2
+#define HDMI_FC_ISCR2_15                        0x10A3
+#define HDMI_FC_ISCR2_14                        0x10A4
+#define HDMI_FC_ISCR2_13                        0x10A5
+#define HDMI_FC_ISCR2_12                        0x10A6
+#define HDMI_FC_ISCR2_11                        0x10A7
+#define HDMI_FC_ISCR2_10                        0x10A8
+#define HDMI_FC_ISCR2_9                         0x10A9
+#define HDMI_FC_ISCR2_8                         0x10AA
+#define HDMI_FC_ISCR2_7                         0x10AB
+#define HDMI_FC_ISCR2_6                         0x10AC
+#define HDMI_FC_ISCR2_5                         0x10AD
+#define HDMI_FC_ISCR2_4                         0x10AE
+#define HDMI_FC_ISCR2_3                         0x10AF
+#define HDMI_FC_ISCR2_2                         0x10B0
+#define HDMI_FC_ISCR2_1                         0x10B1
+#define HDMI_FC_ISCR2_0                         0x10B2
+#define HDMI_FC_DATAUTO0                        0x10B3
+#define HDMI_FC_DATAUTO1                        0x10B4
+#define HDMI_FC_DATAUTO2                        0x10B5
+#define HDMI_FC_DATMAN                          0x10B6
+#define HDMI_FC_DATAUTO3                        0x10B7
+#define HDMI_FC_RDRB0                           0x10B8
+#define HDMI_FC_RDRB1                           0x10B9
+#define HDMI_FC_RDRB2                           0x10BA
+#define HDMI_FC_RDRB3                           0x10BB
+#define HDMI_FC_RDRB4                           0x10BC
+#define HDMI_FC_RDRB5                           0x10BD
+#define HDMI_FC_RDRB6                           0x10BE
+#define HDMI_FC_RDRB7                           0x10BF
+#define HDMI_FC_STAT0                           0x10D0
+#define HDMI_FC_INT0                            0x10D1
+#define HDMI_FC_MASK0                           0x10D2
+#define HDMI_FC_POL0                            0x10D3
+#define HDMI_FC_STAT1                           0x10D4
+#define HDMI_FC_INT1                            0x10D5
+#define HDMI_FC_MASK1                           0x10D6
+#define HDMI_FC_POL1                            0x10D7
+#define HDMI_FC_STAT2                           0x10D8
+#define HDMI_FC_INT2                            0x10D9
+#define HDMI_FC_MASK2                           0x10DA
+#define HDMI_FC_POL2                            0x10DB
+#define HDMI_FC_PRCONF                          0x10E0
+
+#define HDMI_FC_GMD_STAT                        0x1100
+#define HDMI_FC_GMD_EN                          0x1101
+#define HDMI_FC_GMD_UP                          0x1102
+#define HDMI_FC_GMD_CONF                        0x1103
+#define HDMI_FC_GMD_HB                          0x1104
+#define HDMI_FC_GMD_PB0                         0x1105
+#define HDMI_FC_GMD_PB1                         0x1106
+#define HDMI_FC_GMD_PB2                         0x1107
+#define HDMI_FC_GMD_PB3                         0x1108
+#define HDMI_FC_GMD_PB4                         0x1109
+#define HDMI_FC_GMD_PB5                         0x110A
+#define HDMI_FC_GMD_PB6                         0x110B
+#define HDMI_FC_GMD_PB7                         0x110C
+#define HDMI_FC_GMD_PB8                         0x110D
+#define HDMI_FC_GMD_PB9                         0x110E
+#define HDMI_FC_GMD_PB10                        0x110F
+#define HDMI_FC_GMD_PB11                        0x1110
+#define HDMI_FC_GMD_PB12                        0x1111
+#define HDMI_FC_GMD_PB13                        0x1112
+#define HDMI_FC_GMD_PB14                        0x1113
+#define HDMI_FC_GMD_PB15                        0x1114
+#define HDMI_FC_GMD_PB16                        0x1115
+#define HDMI_FC_GMD_PB17                        0x1116
+#define HDMI_FC_GMD_PB18                        0x1117
+#define HDMI_FC_GMD_PB19                        0x1118
+#define HDMI_FC_GMD_PB20                        0x1119
+#define HDMI_FC_GMD_PB21                        0x111A
+#define HDMI_FC_GMD_PB22                        0x111B
+#define HDMI_FC_GMD_PB23                        0x111C
+#define HDMI_FC_GMD_PB24                        0x111D
+#define HDMI_FC_GMD_PB25                        0x111E
+#define HDMI_FC_GMD_PB26                        0x111F
+#define HDMI_FC_GMD_PB27                        0x1120
+
+#define HDMI_FC_DBGFORCE                        0x1200
+#define HDMI_FC_DBGAUD0CH0                      0x1201
+#define HDMI_FC_DBGAUD1CH0                      0x1202
+#define HDMI_FC_DBGAUD2CH0                      0x1203
+#define HDMI_FC_DBGAUD0CH1                      0x1204
+#define HDMI_FC_DBGAUD1CH1                      0x1205
+#define HDMI_FC_DBGAUD2CH1                      0x1206
+#define HDMI_FC_DBGAUD0CH2                      0x1207
+#define HDMI_FC_DBGAUD1CH2                      0x1208
+#define HDMI_FC_DBGAUD2CH2                      0x1209
+#define HDMI_FC_DBGAUD0CH3                      0x120A
+#define HDMI_FC_DBGAUD1CH3                      0x120B
+#define HDMI_FC_DBGAUD2CH3                      0x120C
+#define HDMI_FC_DBGAUD0CH4                      0x120D
+#define HDMI_FC_DBGAUD1CH4                      0x120E
+#define HDMI_FC_DBGAUD2CH4                      0x120F
+#define HDMI_FC_DBGAUD0CH5                      0x1210
+#define HDMI_FC_DBGAUD1CH5                      0x1211
+#define HDMI_FC_DBGAUD2CH5                      0x1212
+#define HDMI_FC_DBGAUD0CH6                      0x1213
+#define HDMI_FC_DBGAUD1CH6                      0x1214
+#define HDMI_FC_DBGAUD2CH6                      0x1215
+#define HDMI_FC_DBGAUD0CH7                      0x1216
+#define HDMI_FC_DBGAUD1CH7                      0x1217
+#define HDMI_FC_DBGAUD2CH7                      0x1218
+#define HDMI_FC_DBGTMDS0                        0x1219
+#define HDMI_FC_DBGTMDS1                        0x121A
+#define HDMI_FC_DBGTMDS2                        0x121B
+
+/* HDMI Source PHY Registers */
+#define HDMI_PHY_CONF0                          0x3000
+#define HDMI_PHY_TST0                           0x3001
+#define HDMI_PHY_TST1                           0x3002
+#define HDMI_PHY_TST2                           0x3003
+#define HDMI_PHY_STAT0                          0x3004
+#define HDMI_PHY_INT0                           0x3005
+#define HDMI_PHY_MASK0                          0x3006
+#define HDMI_PHY_POL0                           0x3007
+
+/* HDMI Master PHY Registers */
+#define HDMI_PHY_I2CM_SLAVE_ADDR                0x3020
+#define HDMI_PHY_I2CM_ADDRESS_ADDR              0x3021
+#define HDMI_PHY_I2CM_DATAO_1_ADDR              0x3022
+#define HDMI_PHY_I2CM_DATAO_0_ADDR              0x3023
+#define HDMI_PHY_I2CM_DATAI_1_ADDR              0x3024
+#define HDMI_PHY_I2CM_DATAI_0_ADDR              0x3025
+#define HDMI_PHY_I2CM_OPERATION_ADDR            0x3026
+#define HDMI_PHY_I2CM_INT_ADDR                  0x3027
+#define HDMI_PHY_I2CM_CTLINT_ADDR               0x3028
+#define HDMI_PHY_I2CM_DIV_ADDR                  0x3029
+#define HDMI_PHY_I2CM_SOFTRSTZ_ADDR             0x302a
+#define HDMI_PHY_I2CM_SS_SCL_HCNT_1_ADDR        0x302b
+#define HDMI_PHY_I2CM_SS_SCL_HCNT_0_ADDR        0x302c
+#define HDMI_PHY_I2CM_SS_SCL_LCNT_1_ADDR        0x302d
+#define HDMI_PHY_I2CM_SS_SCL_LCNT_0_ADDR        0x302e
+#define HDMI_PHY_I2CM_FS_SCL_HCNT_1_ADDR        0x302f
+#define HDMI_PHY_I2CM_FS_SCL_HCNT_0_ADDR        0x3030
+#define HDMI_PHY_I2CM_FS_SCL_LCNT_1_ADDR        0x3031
+#define HDMI_PHY_I2CM_FS_SCL_LCNT_0_ADDR        0x3032
+
+/* Audio Sampler Registers */
+#define HDMI_AUD_CONF0                          0x3100
+#define HDMI_AUD_CONF1                          0x3101
+#define HDMI_AUD_INT                            0x3102
+#define HDMI_AUD_CONF2                          0x3103
+#define HDMI_AUD_N1                             0x3200
+#define HDMI_AUD_N2                             0x3201
+#define HDMI_AUD_N3                             0x3202
+#define HDMI_AUD_CTS1                           0x3203
+#define HDMI_AUD_CTS2                           0x3204
+#define HDMI_AUD_CTS3                           0x3205
+#define HDMI_AUD_INPUTCLKFS                     0x3206
+#define HDMI_AUD_SPDIFINT			0x3302
+#define HDMI_AUD_CONF0_HBR                      0x3400
+#define HDMI_AUD_HBR_STATUS                     0x3401
+#define HDMI_AUD_HBR_INT                        0x3402
+#define HDMI_AUD_HBR_POL                        0x3403
+#define HDMI_AUD_HBR_MASK                       0x3404
+
+/*
+ * Generic Parallel Audio Interface Registers
+ * Not used as GPAUD interface is not enabled in hw
+ */
+#define HDMI_GP_CONF0                           0x3500
+#define HDMI_GP_CONF1                           0x3501
+#define HDMI_GP_CONF2                           0x3502
+#define HDMI_GP_STAT                            0x3503
+#define HDMI_GP_INT                             0x3504
+#define HDMI_GP_MASK                            0x3505
+#define HDMI_GP_POL                             0x3506
+
+/* Audio DMA Registers */
+#define HDMI_AHB_DMA_CONF0                      0x3600
+#define HDMI_AHB_DMA_START                      0x3601
+#define HDMI_AHB_DMA_STOP                       0x3602
+#define HDMI_AHB_DMA_THRSLD                     0x3603
+#define HDMI_AHB_DMA_STRADDR0                   0x3604
+#define HDMI_AHB_DMA_STRADDR1                   0x3605
+#define HDMI_AHB_DMA_STRADDR2                   0x3606
+#define HDMI_AHB_DMA_STRADDR3                   0x3607
+#define HDMI_AHB_DMA_STPADDR0                   0x3608
+#define HDMI_AHB_DMA_STPADDR1                   0x3609
+#define HDMI_AHB_DMA_STPADDR2                   0x360a
+#define HDMI_AHB_DMA_STPADDR3                   0x360b
+#define HDMI_AHB_DMA_BSTADDR0                   0x360c
+#define HDMI_AHB_DMA_BSTADDR1                   0x360d
+#define HDMI_AHB_DMA_BSTADDR2                   0x360e
+#define HDMI_AHB_DMA_BSTADDR3                   0x360f
+#define HDMI_AHB_DMA_MBLENGTH0                  0x3610
+#define HDMI_AHB_DMA_MBLENGTH1                  0x3611
+#define HDMI_AHB_DMA_STAT                       0x3612
+#define HDMI_AHB_DMA_INT                        0x3613
+#define HDMI_AHB_DMA_MASK                       0x3614
+#define HDMI_AHB_DMA_POL                        0x3615
+#define HDMI_AHB_DMA_CONF1                      0x3616
+#define HDMI_AHB_DMA_BUFFSTAT                   0x3617
+#define HDMI_AHB_DMA_BUFFINT                    0x3618
+#define HDMI_AHB_DMA_BUFFMASK                   0x3619
+#define HDMI_AHB_DMA_BUFFPOL                    0x361a
+
+/* Main Controller Registers */
+#define HDMI_MC_SFRDIV                          0x4000
+#define HDMI_MC_CLKDIS                          0x4001
+#define HDMI_MC_SWRSTZ                          0x4002
+#define HDMI_MC_OPCTRL                          0x4003
+#define HDMI_MC_FLOWCTRL                        0x4004
+#define HDMI_MC_PHYRSTZ                         0x4005
+#define HDMI_MC_LOCKONCLOCK                     0x4006
+#define HDMI_MC_HEACPHY_RST                     0x4007
+
+/* Color Space  Converter Registers */
+#define HDMI_CSC_CFG                            0x4100
+#define HDMI_CSC_SCALE                          0x4101
+#define HDMI_CSC_COEF_A1_MSB                    0x4102
+#define HDMI_CSC_COEF_A1_LSB                    0x4103
+#define HDMI_CSC_COEF_A2_MSB                    0x4104
+#define HDMI_CSC_COEF_A2_LSB                    0x4105
+#define HDMI_CSC_COEF_A3_MSB                    0x4106
+#define HDMI_CSC_COEF_A3_LSB                    0x4107
+#define HDMI_CSC_COEF_A4_MSB                    0x4108
+#define HDMI_CSC_COEF_A4_LSB                    0x4109
+#define HDMI_CSC_COEF_B1_MSB                    0x410A
+#define HDMI_CSC_COEF_B1_LSB                    0x410B
+#define HDMI_CSC_COEF_B2_MSB                    0x410C
+#define HDMI_CSC_COEF_B2_LSB                    0x410D
+#define HDMI_CSC_COEF_B3_MSB                    0x410E
+#define HDMI_CSC_COEF_B3_LSB                    0x410F
+#define HDMI_CSC_COEF_B4_MSB                    0x4110
+#define HDMI_CSC_COEF_B4_LSB                    0x4111
+#define HDMI_CSC_COEF_C1_MSB                    0x4112
+#define HDMI_CSC_COEF_C1_LSB                    0x4113
+#define HDMI_CSC_COEF_C2_MSB                    0x4114
+#define HDMI_CSC_COEF_C2_LSB                    0x4115
+#define HDMI_CSC_COEF_C3_MSB                    0x4116
+#define HDMI_CSC_COEF_C3_LSB                    0x4117
+#define HDMI_CSC_COEF_C4_MSB                    0x4118
+#define HDMI_CSC_COEF_C4_LSB                    0x4119
+
+/* HDCP Encryption Engine Registers */
+#define HDMI_A_HDCPCFG0                         0x5000
+#define HDMI_A_HDCPCFG1                         0x5001
+#define HDMI_A_HDCPOBS0                         0x5002
+#define HDMI_A_HDCPOBS1                         0x5003
+#define HDMI_A_HDCPOBS2                         0x5004
+#define HDMI_A_HDCPOBS3                         0x5005
+#define HDMI_A_APIINTCLR                        0x5006
+#define HDMI_A_APIINTSTAT                       0x5007
+#define HDMI_A_APIINTMSK                        0x5008
+#define HDMI_A_VIDPOLCFG                        0x5009
+#define HDMI_A_OESSWCFG                         0x500A
+#define HDMI_A_TIMER1SETUP0                     0x500B
+#define HDMI_A_TIMER1SETUP1                     0x500C
+#define HDMI_A_TIMER2SETUP0                     0x500D
+#define HDMI_A_TIMER2SETUP1                     0x500E
+#define HDMI_A_100MSCFG                         0x500F
+#define HDMI_A_2SCFG0                           0x5010
+#define HDMI_A_2SCFG1                           0x5011
+#define HDMI_A_5SCFG0                           0x5012
+#define HDMI_A_5SCFG1                           0x5013
+#define HDMI_A_SRMVERLSB                        0x5014
+#define HDMI_A_SRMVERMSB                        0x5015
+#define HDMI_A_SRMCTRL                          0x5016
+#define HDMI_A_SFRSETUP                         0x5017
+#define HDMI_A_I2CHSETUP                        0x5018
+#define HDMI_A_INTSETUP                         0x5019
+#define HDMI_A_PRESETUP                         0x501A
+#define HDMI_A_SRM_BASE                         0x5020
+
+/* CEC Engine Registers */
+#define HDMI_CEC_CTRL                           0x7D00
+#define HDMI_CEC_STAT                           0x7D01
+#define HDMI_CEC_MASK                           0x7D02
+#define HDMI_CEC_POLARITY                       0x7D03
+#define HDMI_CEC_INT                            0x7D04
+#define HDMI_CEC_ADDR_L                         0x7D05
+#define HDMI_CEC_ADDR_H                         0x7D06
+#define HDMI_CEC_TX_CNT                         0x7D07
+#define HDMI_CEC_RX_CNT                         0x7D08
+#define HDMI_CEC_TX_DATA0                       0x7D10
+#define HDMI_CEC_TX_DATA1                       0x7D11
+#define HDMI_CEC_TX_DATA2                       0x7D12
+#define HDMI_CEC_TX_DATA3                       0x7D13
+#define HDMI_CEC_TX_DATA4                       0x7D14
+#define HDMI_CEC_TX_DATA5                       0x7D15
+#define HDMI_CEC_TX_DATA6                       0x7D16
+#define HDMI_CEC_TX_DATA7                       0x7D17
+#define HDMI_CEC_TX_DATA8                       0x7D18
+#define HDMI_CEC_TX_DATA9                       0x7D19
+#define HDMI_CEC_TX_DATA10                      0x7D1a
+#define HDMI_CEC_TX_DATA11                      0x7D1b
+#define HDMI_CEC_TX_DATA12                      0x7D1c
+#define HDMI_CEC_TX_DATA13                      0x7D1d
+#define HDMI_CEC_TX_DATA14                      0x7D1e
+#define HDMI_CEC_TX_DATA15                      0x7D1f
+#define HDMI_CEC_RX_DATA0                       0x7D20
+#define HDMI_CEC_RX_DATA1                       0x7D21
+#define HDMI_CEC_RX_DATA2                       0x7D22
+#define HDMI_CEC_RX_DATA3                       0x7D23
+#define HDMI_CEC_RX_DATA4                       0x7D24
+#define HDMI_CEC_RX_DATA5                       0x7D25
+#define HDMI_CEC_RX_DATA6                       0x7D26
+#define HDMI_CEC_RX_DATA7                       0x7D27
+#define HDMI_CEC_RX_DATA8                       0x7D28
+#define HDMI_CEC_RX_DATA9                       0x7D29
+#define HDMI_CEC_RX_DATA10                      0x7D2a
+#define HDMI_CEC_RX_DATA11                      0x7D2b
+#define HDMI_CEC_RX_DATA12                      0x7D2c
+#define HDMI_CEC_RX_DATA13                      0x7D2d
+#define HDMI_CEC_RX_DATA14                      0x7D2e
+#define HDMI_CEC_RX_DATA15                      0x7D2f
+#define HDMI_CEC_LOCK                           0x7D30
+#define HDMI_CEC_WKUPCTRL                       0x7D31
+
+/* I2C Master Registers (E-DDC) */
+#define HDMI_I2CM_SLAVE                         0x7E00
+#define HDMI_I2CMESS                            0x7E01
+#define HDMI_I2CM_DATAO                         0x7E02
+#define HDMI_I2CM_DATAI                         0x7E03
+#define HDMI_I2CM_OPERATION                     0x7E04
+#define HDMI_I2CM_INT                           0x7E05
+#define HDMI_I2CM_CTLINT                        0x7E06
+#define HDMI_I2CM_DIV                           0x7E07
+#define HDMI_I2CM_SEGADDR                       0x7E08
+#define HDMI_I2CM_SOFTRSTZ                      0x7E09
+#define HDMI_I2CM_SEGPTR                        0x7E0A
+#define HDMI_I2CM_SS_SCL_HCNT_1_ADDR            0x7E0B
+#define HDMI_I2CM_SS_SCL_HCNT_0_ADDR            0x7E0C
+#define HDMI_I2CM_SS_SCL_LCNT_1_ADDR            0x7E0D
+#define HDMI_I2CM_SS_SCL_LCNT_0_ADDR            0x7E0E
+#define HDMI_I2CM_FS_SCL_HCNT_1_ADDR            0x7E0F
+#define HDMI_I2CM_FS_SCL_HCNT_0_ADDR            0x7E10
+#define HDMI_I2CM_FS_SCL_LCNT_1_ADDR            0x7E11
+#define HDMI_I2CM_FS_SCL_LCNT_0_ADDR            0x7E12
+
+enum {
+/* IH_FC_INT2 field values */
+	HDMI_IH_FC_INT2_OVERFLOW_MASK = 0x03,
+	HDMI_IH_FC_INT2_LOW_PRIORITY_OVERFLOW = 0x02,
+	HDMI_IH_FC_INT2_HIGH_PRIORITY_OVERFLOW = 0x01,
+
+/* IH_FC_STAT2 field values */
+	HDMI_IH_FC_STAT2_OVERFLOW_MASK = 0x03,
+	HDMI_IH_FC_STAT2_LOW_PRIORITY_OVERFLOW = 0x02,
+	HDMI_IH_FC_STAT2_HIGH_PRIORITY_OVERFLOW = 0x01,
+
+/* IH_PHY_STAT0 field values */
+	HDMI_IH_PHY_STAT0_RX_SENSE3 = 0x20,
+	HDMI_IH_PHY_STAT0_RX_SENSE2 = 0x10,
+	HDMI_IH_PHY_STAT0_RX_SENSE1 = 0x8,
+	HDMI_IH_PHY_STAT0_RX_SENSE0 = 0x4,
+	HDMI_IH_PHY_STAT0_TX_PHY_LOCK = 0x2,
+	HDMI_IH_PHY_STAT0_HPD = 0x1,
+
+/* IH_MUTE_I2CMPHY_STAT0 field values */
+	HDMI_IH_MUTE_I2CMPHY_STAT0_I2CMPHYDONE = 0x2,
+	HDMI_IH_MUTE_I2CMPHY_STAT0_I2CMPHYERROR = 0x1,
+
+/* IH_AHBDMAAUD_STAT0 field values */
+	HDMI_IH_AHBDMAAUD_STAT0_ERROR = 0x20,
+	HDMI_IH_AHBDMAAUD_STAT0_LOST = 0x10,
+	HDMI_IH_AHBDMAAUD_STAT0_RETRY = 0x08,
+	HDMI_IH_AHBDMAAUD_STAT0_DONE = 0x04,
+	HDMI_IH_AHBDMAAUD_STAT0_BUFFFULL = 0x02,
+	HDMI_IH_AHBDMAAUD_STAT0_BUFFEMPTY = 0x01,
+
+/* IH_MUTE_FC_STAT2 field values */
+	HDMI_IH_MUTE_FC_STAT2_OVERFLOW_MASK = 0x03,
+	HDMI_IH_MUTE_FC_STAT2_LOW_PRIORITY_OVERFLOW = 0x02,
+	HDMI_IH_MUTE_FC_STAT2_HIGH_PRIORITY_OVERFLOW = 0x01,
+
+/* IH_MUTE_AHBDMAAUD_STAT0 field values */
+	HDMI_IH_MUTE_AHBDMAAUD_STAT0_ERROR = 0x20,
+	HDMI_IH_MUTE_AHBDMAAUD_STAT0_LOST = 0x10,
+	HDMI_IH_MUTE_AHBDMAAUD_STAT0_RETRY = 0x08,
+	HDMI_IH_MUTE_AHBDMAAUD_STAT0_DONE = 0x04,
+	HDMI_IH_MUTE_AHBDMAAUD_STAT0_BUFFFULL = 0x02,
+	HDMI_IH_MUTE_AHBDMAAUD_STAT0_BUFFEMPTY = 0x01,
+
+/* IH_MUTE field values */
+	HDMI_IH_MUTE_MUTE_WAKEUP_INTERRUPT = 0x2,
+	HDMI_IH_MUTE_MUTE_ALL_INTERRUPT = 0x1,
+
+/* TX_INVID0 field values */
+	HDMI_TX_INVID0_INTERNAL_DE_GENERATOR_MASK = 0x80,
+	HDMI_TX_INVID0_INTERNAL_DE_GENERATOR_ENABLE = 0x80,
+	HDMI_TX_INVID0_INTERNAL_DE_GENERATOR_DISABLE = 0x00,
+	HDMI_TX_INVID0_VIDEO_MAPPING_MASK = 0x1F,
+	HDMI_TX_INVID0_VIDEO_MAPPING_OFFSET = 0,
+
+/* TX_INSTUFFING field values */
+	HDMI_TX_INSTUFFING_BDBDATA_STUFFING_MASK = 0x4,
+	HDMI_TX_INSTUFFING_BDBDATA_STUFFING_ENABLE = 0x4,
+	HDMI_TX_INSTUFFING_BDBDATA_STUFFING_DISABLE = 0x0,
+	HDMI_TX_INSTUFFING_RCRDATA_STUFFING_MASK = 0x2,
+	HDMI_TX_INSTUFFING_RCRDATA_STUFFING_ENABLE = 0x2,
+	HDMI_TX_INSTUFFING_RCRDATA_STUFFING_DISABLE = 0x0,
+	HDMI_TX_INSTUFFING_GYDATA_STUFFING_MASK = 0x1,
+	HDMI_TX_INSTUFFING_GYDATA_STUFFING_ENABLE = 0x1,
+	HDMI_TX_INSTUFFING_GYDATA_STUFFING_DISABLE = 0x0,
+
+/* VP_PR_CD field values */
+	HDMI_VP_PR_CD_COLOR_DEPTH_MASK = 0xF0,
+	HDMI_VP_PR_CD_COLOR_DEPTH_OFFSET = 4,
+	HDMI_VP_PR_CD_DESIRED_PR_FACTOR_MASK = 0x0F,
+	HDMI_VP_PR_CD_DESIRED_PR_FACTOR_OFFSET = 0,
+
+/* VP_STUFF field values */
+	HDMI_VP_STUFF_IDEFAULT_PHASE_MASK = 0x20,
+	HDMI_VP_STUFF_IDEFAULT_PHASE_OFFSET = 5,
+	HDMI_VP_STUFF_IFIX_PP_TO_LAST_MASK = 0x10,
+	HDMI_VP_STUFF_IFIX_PP_TO_LAST_OFFSET = 4,
+	HDMI_VP_STUFF_ICX_GOTO_P0_ST_MASK = 0x8,
+	HDMI_VP_STUFF_ICX_GOTO_P0_ST_OFFSET = 3,
+	HDMI_VP_STUFF_YCC422_STUFFING_MASK = 0x4,
+	HDMI_VP_STUFF_YCC422_STUFFING_STUFFING_MODE = 0x4,
+	HDMI_VP_STUFF_YCC422_STUFFING_DIRECT_MODE = 0x0,
+	HDMI_VP_STUFF_PP_STUFFING_MASK = 0x2,
+	HDMI_VP_STUFF_PP_STUFFING_STUFFING_MODE = 0x2,
+	HDMI_VP_STUFF_PP_STUFFING_DIRECT_MODE = 0x0,
+	HDMI_VP_STUFF_PR_STUFFING_MASK = 0x1,
+	HDMI_VP_STUFF_PR_STUFFING_STUFFING_MODE = 0x1,
+	HDMI_VP_STUFF_PR_STUFFING_DIRECT_MODE = 0x0,
+
+/* VP_CONF field values */
+	HDMI_VP_CONF_BYPASS_EN_MASK = 0x40,
+	HDMI_VP_CONF_BYPASS_EN_ENABLE = 0x40,
+	HDMI_VP_CONF_BYPASS_EN_DISABLE = 0x00,
+	HDMI_VP_CONF_PP_EN_ENMASK = 0x20,
+	HDMI_VP_CONF_PP_EN_ENABLE = 0x20,
+	HDMI_VP_CONF_PP_EN_DISABLE = 0x00,
+	HDMI_VP_CONF_PR_EN_MASK = 0x10,
+	HDMI_VP_CONF_PR_EN_ENABLE = 0x10,
+	HDMI_VP_CONF_PR_EN_DISABLE = 0x00,
+	HDMI_VP_CONF_YCC422_EN_MASK = 0x8,
+	HDMI_VP_CONF_YCC422_EN_ENABLE = 0x8,
+	HDMI_VP_CONF_YCC422_EN_DISABLE = 0x0,
+	HDMI_VP_CONF_BYPASS_SELECT_MASK = 0x4,
+	HDMI_VP_CONF_BYPASS_SELECT_VID_PACKETIZER = 0x4,
+	HDMI_VP_CONF_BYPASS_SELECT_PIX_REPEATER = 0x0,
+	HDMI_VP_CONF_OUTPUT_SELECTOR_MASK = 0x3,
+	HDMI_VP_CONF_OUTPUT_SELECTOR_BYPASS = 0x3,
+	HDMI_VP_CONF_OUTPUT_SELECTOR_YCC422 = 0x1,
+	HDMI_VP_CONF_OUTPUT_SELECTOR_PP = 0x0,
+
+/* VP_REMAP field values */
+	HDMI_VP_REMAP_MASK = 0x3,
+	HDMI_VP_REMAP_YCC422_24bit = 0x2,
+	HDMI_VP_REMAP_YCC422_20bit = 0x1,
+	HDMI_VP_REMAP_YCC422_16bit = 0x0,
+
+/* FC_INVIDCONF field values */
+	HDMI_FC_INVIDCONF_HDCP_KEEPOUT_MASK = 0x80,
+	HDMI_FC_INVIDCONF_HDCP_KEEPOUT_ACTIVE = 0x80,
+	HDMI_FC_INVIDCONF_HDCP_KEEPOUT_INACTIVE = 0x00,
+	HDMI_FC_INVIDCONF_VSYNC_IN_POLARITY_MASK = 0x40,
+	HDMI_FC_INVIDCONF_VSYNC_IN_POLARITY_ACTIVE_HIGH = 0x40,
+	HDMI_FC_INVIDCONF_VSYNC_IN_POLARITY_ACTIVE_LOW = 0x00,
+	HDMI_FC_INVIDCONF_HSYNC_IN_POLARITY_MASK = 0x20,
+	HDMI_FC_INVIDCONF_HSYNC_IN_POLARITY_ACTIVE_HIGH = 0x20,
+	HDMI_FC_INVIDCONF_HSYNC_IN_POLARITY_ACTIVE_LOW = 0x00,
+	HDMI_FC_INVIDCONF_DE_IN_POLARITY_MASK = 0x10,
+	HDMI_FC_INVIDCONF_DE_IN_POLARITY_ACTIVE_HIGH = 0x10,
+	HDMI_FC_INVIDCONF_DE_IN_POLARITY_ACTIVE_LOW = 0x00,
+	HDMI_FC_INVIDCONF_DVI_MODEZ_MASK = 0x8,
+	HDMI_FC_INVIDCONF_DVI_MODEZ_HDMI_MODE = 0x8,
+	HDMI_FC_INVIDCONF_DVI_MODEZ_DVI_MODE = 0x0,
+	HDMI_FC_INVIDCONF_R_V_BLANK_IN_OSC_MASK = 0x2,
+	HDMI_FC_INVIDCONF_R_V_BLANK_IN_OSC_ACTIVE_HIGH = 0x2,
+	HDMI_FC_INVIDCONF_R_V_BLANK_IN_OSC_ACTIVE_LOW = 0x0,
+	HDMI_FC_INVIDCONF_IN_I_P_MASK = 0x1,
+	HDMI_FC_INVIDCONF_IN_I_P_INTERLACED = 0x1,
+	HDMI_FC_INVIDCONF_IN_I_P_PROGRESSIVE = 0x0,
+
+/* FC_AUDICONF0 field values */
+	HDMI_FC_AUDICONF0_CC_OFFSET = 4,
+	HDMI_FC_AUDICONF0_CC_MASK = 0x70,
+	HDMI_FC_AUDICONF0_CT_OFFSET = 0,
+	HDMI_FC_AUDICONF0_CT_MASK = 0xF,
+
+/* FC_AUDICONF1 field values */
+	HDMI_FC_AUDICONF1_SS_OFFSET = 3,
+	HDMI_FC_AUDICONF1_SS_MASK = 0x18,
+	HDMI_FC_AUDICONF1_SF_OFFSET = 0,
+	HDMI_FC_AUDICONF1_SF_MASK = 0x7,
+
+/* FC_AUDICONF3 field values */
+	HDMI_FC_AUDICONF3_LFEPBL_OFFSET = 5,
+	HDMI_FC_AUDICONF3_LFEPBL_MASK = 0x60,
+	HDMI_FC_AUDICONF3_DM_INH_OFFSET = 4,
+	HDMI_FC_AUDICONF3_DM_INH_MASK = 0x10,
+	HDMI_FC_AUDICONF3_LSV_OFFSET = 0,
+	HDMI_FC_AUDICONF3_LSV_MASK = 0xF,
+
+/* FC_AUDSCHNLS0 field values */
+	HDMI_FC_AUDSCHNLS0_CGMSA_OFFSET = 4,
+	HDMI_FC_AUDSCHNLS0_CGMSA_MASK = 0x30,
+	HDMI_FC_AUDSCHNLS0_COPYRIGHT_OFFSET = 0,
+	HDMI_FC_AUDSCHNLS0_COPYRIGHT_MASK = 0x01,
+
+/* FC_AUDSCHNLS3-6 field values */
+	HDMI_FC_AUDSCHNLS3_OIEC_CH0_OFFSET = 0,
+	HDMI_FC_AUDSCHNLS3_OIEC_CH0_MASK = 0x0f,
+	HDMI_FC_AUDSCHNLS3_OIEC_CH1_OFFSET = 4,
+	HDMI_FC_AUDSCHNLS3_OIEC_CH1_MASK = 0xf0,
+	HDMI_FC_AUDSCHNLS4_OIEC_CH2_OFFSET = 0,
+	HDMI_FC_AUDSCHNLS4_OIEC_CH2_MASK = 0x0f,
+	HDMI_FC_AUDSCHNLS4_OIEC_CH3_OFFSET = 4,
+	HDMI_FC_AUDSCHNLS4_OIEC_CH3_MASK = 0xf0,
+
+	HDMI_FC_AUDSCHNLS5_OIEC_CH0_OFFSET = 0,
+	HDMI_FC_AUDSCHNLS5_OIEC_CH0_MASK = 0x0f,
+	HDMI_FC_AUDSCHNLS5_OIEC_CH1_OFFSET = 4,
+	HDMI_FC_AUDSCHNLS5_OIEC_CH1_MASK = 0xf0,
+	HDMI_FC_AUDSCHNLS6_OIEC_CH2_OFFSET = 0,
+	HDMI_FC_AUDSCHNLS6_OIEC_CH2_MASK = 0x0f,
+	HDMI_FC_AUDSCHNLS6_OIEC_CH3_OFFSET = 4,
+	HDMI_FC_AUDSCHNLS6_OIEC_CH3_MASK = 0xf0,
+
+/* HDMI_FC_AUDSCHNLS7 field values */
+	HDMI_FC_AUDSCHNLS7_ACCURACY_OFFSET = 4,
+	HDMI_FC_AUDSCHNLS7_ACCURACY_MASK = 0x30,
+
+/* HDMI_FC_AUDSCHNLS8 field values */
+	HDMI_FC_AUDSCHNLS8_ORIGSAMPFREQ_MASK = 0xf0,
+	HDMI_FC_AUDSCHNLS8_ORIGSAMPFREQ_OFFSET = 4,
+	HDMI_FC_AUDSCHNLS8_WORDLEGNTH_MASK = 0x0f,
+	HDMI_FC_AUDSCHNLS8_WORDLEGNTH_OFFSET = 0,
+
+/* FC_AUDSCONF field values */
+	HDMI_FC_AUDSCONF_AUD_PACKET_SAMPFIT_MASK = 0xF0,
+	HDMI_FC_AUDSCONF_AUD_PACKET_SAMPFIT_OFFSET = 4,
+	HDMI_FC_AUDSCONF_AUD_PACKET_LAYOUT_MASK = 0x1,
+	HDMI_FC_AUDSCONF_AUD_PACKET_LAYOUT_OFFSET = 0,
+	HDMI_FC_AUDSCONF_AUD_PACKET_LAYOUT_LAYOUT1 = 0x1,
+	HDMI_FC_AUDSCONF_AUD_PACKET_LAYOUT_LAYOUT0 = 0x0,
+
+/* FC_STAT2 field values */
+	HDMI_FC_STAT2_OVERFLOW_MASK = 0x03,
+	HDMI_FC_STAT2_LOW_PRIORITY_OVERFLOW = 0x02,
+	HDMI_FC_STAT2_HIGH_PRIORITY_OVERFLOW = 0x01,
+
+/* FC_INT2 field values */
+	HDMI_FC_INT2_OVERFLOW_MASK = 0x03,
+	HDMI_FC_INT2_LOW_PRIORITY_OVERFLOW = 0x02,
+	HDMI_FC_INT2_HIGH_PRIORITY_OVERFLOW = 0x01,
+
+/* FC_MASK2 field values */
+	HDMI_FC_MASK2_OVERFLOW_MASK = 0x03,
+	HDMI_FC_MASK2_LOW_PRIORITY_OVERFLOW = 0x02,
+	HDMI_FC_MASK2_HIGH_PRIORITY_OVERFLOW = 0x01,
+
+/* FC_PRCONF field values */
+	HDMI_FC_PRCONF_INCOMING_PR_FACTOR_MASK = 0xF0,
+	HDMI_FC_PRCONF_INCOMING_PR_FACTOR_OFFSET = 4,
+	HDMI_FC_PRCONF_OUTPUT_PR_FACTOR_MASK = 0x0F,
+	HDMI_FC_PRCONF_OUTPUT_PR_FACTOR_OFFSET = 0,
+
+/* FC_AVICONF0-FC_AVICONF3 field values */
+	HDMI_FC_AVICONF0_PIX_FMT_MASK = 0x03,
+	HDMI_FC_AVICONF0_PIX_FMT_RGB = 0x00,
+	HDMI_FC_AVICONF0_PIX_FMT_YCBCR422 = 0x01,
+	HDMI_FC_AVICONF0_PIX_FMT_YCBCR444 = 0x02,
+	HDMI_FC_AVICONF0_ACTIVE_FMT_MASK = 0x40,
+	HDMI_FC_AVICONF0_ACTIVE_FMT_INFO_PRESENT = 0x40,
+	HDMI_FC_AVICONF0_ACTIVE_FMT_NO_INFO = 0x00,
+	HDMI_FC_AVICONF0_BAR_DATA_MASK = 0x0C,
+	HDMI_FC_AVICONF0_BAR_DATA_NO_DATA = 0x00,
+	HDMI_FC_AVICONF0_BAR_DATA_VERT_BAR = 0x04,
+	HDMI_FC_AVICONF0_BAR_DATA_HORIZ_BAR = 0x08,
+	HDMI_FC_AVICONF0_BAR_DATA_VERT_HORIZ_BAR = 0x0C,
+	HDMI_FC_AVICONF0_SCAN_INFO_MASK = 0x30,
+	HDMI_FC_AVICONF0_SCAN_INFO_OVERSCAN = 0x10,
+	HDMI_FC_AVICONF0_SCAN_INFO_UNDERSCAN = 0x20,
+	HDMI_FC_AVICONF0_SCAN_INFO_NODATA = 0x00,
+
+	HDMI_FC_AVICONF1_ACTIVE_ASPECT_RATIO_MASK = 0x0F,
+	HDMI_FC_AVICONF1_ACTIVE_ASPECT_RATIO_USE_CODED = 0x08,
+	HDMI_FC_AVICONF1_ACTIVE_ASPECT_RATIO_4_3 = 0x09,
+	HDMI_FC_AVICONF1_ACTIVE_ASPECT_RATIO_16_9 = 0x0A,
+	HDMI_FC_AVICONF1_ACTIVE_ASPECT_RATIO_14_9 = 0x0B,
+	HDMI_FC_AVICONF1_CODED_ASPECT_RATIO_MASK = 0x30,
+	HDMI_FC_AVICONF1_CODED_ASPECT_RATIO_NO_DATA = 0x00,
+	HDMI_FC_AVICONF1_CODED_ASPECT_RATIO_4_3 = 0x10,
+	HDMI_FC_AVICONF1_CODED_ASPECT_RATIO_16_9 = 0x20,
+	HDMI_FC_AVICONF1_COLORIMETRY_MASK = 0xC0,
+	HDMI_FC_AVICONF1_COLORIMETRY_NO_DATA = 0x00,
+	HDMI_FC_AVICONF1_COLORIMETRY_SMPTE = 0x40,
+	HDMI_FC_AVICONF1_COLORIMETRY_ITUR = 0x80,
+	HDMI_FC_AVICONF1_COLORIMETRY_EXTENDED_INFO = 0xC0,
+
+	HDMI_FC_AVICONF2_SCALING_MASK = 0x03,
+	HDMI_FC_AVICONF2_SCALING_NONE = 0x00,
+	HDMI_FC_AVICONF2_SCALING_HORIZ = 0x01,
+	HDMI_FC_AVICONF2_SCALING_VERT = 0x02,
+	HDMI_FC_AVICONF2_SCALING_HORIZ_VERT = 0x03,
+	HDMI_FC_AVICONF2_RGB_QUANT_MASK = 0x0C,
+	HDMI_FC_AVICONF2_RGB_QUANT_DEFAULT = 0x00,
+	HDMI_FC_AVICONF2_RGB_QUANT_LIMITED_RANGE = 0x04,
+	HDMI_FC_AVICONF2_RGB_QUANT_FULL_RANGE = 0x08,
+	HDMI_FC_AVICONF2_EXT_COLORIMETRY_MASK = 0x70,
+	HDMI_FC_AVICONF2_EXT_COLORIMETRY_XVYCC601 = 0x00,
+	HDMI_FC_AVICONF2_EXT_COLORIMETRY_XVYCC709 = 0x10,
+	HDMI_FC_AVICONF2_EXT_COLORIMETRY_SYCC601 = 0x20,
+	HDMI_FC_AVICONF2_EXT_COLORIMETRY_ADOBE_YCC601 = 0x30,
+	HDMI_FC_AVICONF2_EXT_COLORIMETRY_ADOBE_RGB = 0x40,
+	HDMI_FC_AVICONF2_IT_CONTENT_MASK = 0x80,
+	HDMI_FC_AVICONF2_IT_CONTENT_NO_DATA = 0x00,
+	HDMI_FC_AVICONF2_IT_CONTENT_VALID = 0x80,
+
+	HDMI_FC_AVICONF3_IT_CONTENT_TYPE_MASK = 0x03,
+	HDMI_FC_AVICONF3_IT_CONTENT_TYPE_GRAPHICS = 0x00,
+	HDMI_FC_AVICONF3_IT_CONTENT_TYPE_PHOTO = 0x01,
+	HDMI_FC_AVICONF3_IT_CONTENT_TYPE_CINEMA = 0x02,
+	HDMI_FC_AVICONF3_IT_CONTENT_TYPE_GAME = 0x03,
+	HDMI_FC_AVICONF3_QUANT_RANGE_MASK = 0x0C,
+	HDMI_FC_AVICONF3_QUANT_RANGE_LIMITED = 0x00,
+	HDMI_FC_AVICONF3_QUANT_RANGE_FULL = 0x04,
+
+/* FC_DBGFORCE field values */
+	HDMI_FC_DBGFORCE_FORCEAUDIO = 0x10,
+	HDMI_FC_DBGFORCE_FORCEVIDEO = 0x1,
+
+/* PHY_CONF0 field values */
+	HDMI_PHY_CONF0_PDZ_MASK = 0x80,
+	HDMI_PHY_CONF0_PDZ_OFFSET = 7,
+	HDMI_PHY_CONF0_ENTMDS_MASK = 0x40,
+	HDMI_PHY_CONF0_ENTMDS_OFFSET = 6,
+	HDMI_PHY_CONF0_SPARECTRL = 0x20,
+	HDMI_PHY_CONF0_GEN2_PDDQ_MASK = 0x10,
+	HDMI_PHY_CONF0_GEN2_PDDQ_OFFSET = 4,
+	HDMI_PHY_CONF0_GEN2_TXPWRON_MASK = 0x8,
+	HDMI_PHY_CONF0_GEN2_TXPWRON_OFFSET = 3,
+	HDMI_PHY_CONF0_GEN2_ENHPDRXSENSE_MASK = 0x4,
+	HDMI_PHY_CONF0_GEN2_ENHPDRXSENSE_OFFSET = 2,
+	HDMI_PHY_CONF0_SELDATAENPOL_MASK = 0x2,
+	HDMI_PHY_CONF0_SELDATAENPOL_OFFSET = 1,
+	HDMI_PHY_CONF0_SELDIPIF_MASK = 0x1,
+	HDMI_PHY_CONF0_SELDIPIF_OFFSET = 0,
+
+/* PHY_TST0 field values */
+	HDMI_PHY_TST0_TSTCLR_MASK = 0x20,
+	HDMI_PHY_TST0_TSTCLR_OFFSET = 5,
+	HDMI_PHY_TST0_TSTEN_MASK = 0x10,
+	HDMI_PHY_TST0_TSTEN_OFFSET = 4,
+	HDMI_PHY_TST0_TSTCLK_MASK = 0x1,
+	HDMI_PHY_TST0_TSTCLK_OFFSET = 0,
+
+/* PHY_STAT0 field values */
+	HDMI_PHY_RX_SENSE3 = 0x80,
+	HDMI_PHY_RX_SENSE2 = 0x40,
+	HDMI_PHY_RX_SENSE1 = 0x20,
+	HDMI_PHY_RX_SENSE0 = 0x10,
+	HDMI_PHY_HPD = 0x02,
+	HDMI_PHY_TX_PHY_LOCK = 0x01,
+
+/* PHY_I2CM_SLAVE_ADDR field values */
+	HDMI_PHY_I2CM_SLAVE_ADDR_PHY_GEN2 = 0x69,
+	HDMI_PHY_I2CM_SLAVE_ADDR_HEAC_PHY = 0x49,
+
+/* PHY_I2CM_OPERATION_ADDR field values */
+	HDMI_PHY_I2CM_OPERATION_ADDR_WRITE = 0x10,
+	HDMI_PHY_I2CM_OPERATION_ADDR_READ = 0x1,
+
+/* HDMI_PHY_I2CM_INT_ADDR */
+	HDMI_PHY_I2CM_INT_ADDR_DONE_POL = 0x08,
+	HDMI_PHY_I2CM_INT_ADDR_DONE_MASK = 0x04,
+
+/* HDMI_PHY_I2CM_CTLINT_ADDR */
+	HDMI_PHY_I2CM_CTLINT_ADDR_NAC_POL = 0x80,
+	HDMI_PHY_I2CM_CTLINT_ADDR_NAC_MASK = 0x40,
+	HDMI_PHY_I2CM_CTLINT_ADDR_ARBITRATION_POL = 0x08,
+	HDMI_PHY_I2CM_CTLINT_ADDR_ARBITRATION_MASK = 0x04,
+
+/* AUD_CTS3 field values */
+	HDMI_AUD_CTS3_N_SHIFT_OFFSET = 5,
+	HDMI_AUD_CTS3_N_SHIFT_MASK = 0xe0,
+	HDMI_AUD_CTS3_N_SHIFT_1 = 0,
+	HDMI_AUD_CTS3_N_SHIFT_16 = 0x20,
+	HDMI_AUD_CTS3_N_SHIFT_32 = 0x40,
+	HDMI_AUD_CTS3_N_SHIFT_64 = 0x60,
+	HDMI_AUD_CTS3_N_SHIFT_128 = 0x80,
+	HDMI_AUD_CTS3_N_SHIFT_256 = 0xa0,
+	/* note that the CTS3 MANUAL bit has been removed
+	   from our part. Can't set it, will read as 0. */
+	HDMI_AUD_CTS3_CTS_MANUAL = 0x10,
+	HDMI_AUD_CTS3_AUDCTS19_16_MASK = 0x0f,
+
+/* AHB_DMA_CONF0 field values */
+	HDMI_AHB_DMA_CONF0_SW_FIFO_RST_OFFSET = 7,
+	HDMI_AHB_DMA_CONF0_SW_FIFO_RST_MASK = 0x80,
+	HDMI_AHB_DMA_CONF0_HBR = 0x10,
+	HDMI_AHB_DMA_CONF0_EN_HLOCK_OFFSET = 3,
+	HDMI_AHB_DMA_CONF0_EN_HLOCK_MASK = 0x08,
+	HDMI_AHB_DMA_CONF0_INCR_TYPE_OFFSET = 1,
+	HDMI_AHB_DMA_CONF0_INCR_TYPE_MASK = 0x06,
+	HDMI_AHB_DMA_CONF0_INCR4 = 0x0,
+	HDMI_AHB_DMA_CONF0_INCR8 = 0x2,
+	HDMI_AHB_DMA_CONF0_INCR16 = 0x4,
+	HDMI_AHB_DMA_CONF0_BURST_MODE = 0x1,
+
+/* HDMI_AHB_DMA_START field values */
+	HDMI_AHB_DMA_START_START_OFFSET = 0,
+	HDMI_AHB_DMA_START_START_MASK = 0x01,
+
+/* HDMI_AHB_DMA_STOP field values */
+	HDMI_AHB_DMA_STOP_STOP_OFFSET = 0,
+	HDMI_AHB_DMA_STOP_STOP_MASK = 0x01,
+
+/* AHB_DMA_STAT, AHB_DMA_INT, AHB_DMA_MASK, AHB_DMA_POL field values */
+	HDMI_AHB_DMA_DONE = 0x80,
+	HDMI_AHB_DMA_RETRY_SPLIT = 0x40,
+	HDMI_AHB_DMA_LOSTOWNERSHIP = 0x20,
+	HDMI_AHB_DMA_ERROR = 0x10,
+	HDMI_AHB_DMA_FIFO_THREMPTY = 0x04,
+	HDMI_AHB_DMA_FIFO_FULL = 0x02,
+	HDMI_AHB_DMA_FIFO_EMPTY = 0x01,
+
+/* AHB_DMA_BUFFSTAT, AHB_DMA_BUFFINT,AHB_DMA_BUFFMASK,AHB_DMA_BUFFPOL values */
+	HDMI_AHB_DMA_BUFFSTAT_FULL = 0x02,
+	HDMI_AHB_DMA_BUFFSTAT_EMPTY = 0x01,
+
+/* MC_CLKDIS field values */
+	HDMI_MC_CLKDIS_HDCPCLK_DISABLE = 0x40,
+	HDMI_MC_CLKDIS_CECCLK_DISABLE = 0x20,
+	HDMI_MC_CLKDIS_CSCCLK_DISABLE = 0x10,
+	HDMI_MC_CLKDIS_AUDCLK_DISABLE = 0x8,
+	HDMI_MC_CLKDIS_PREPCLK_DISABLE = 0x4,
+	HDMI_MC_CLKDIS_TMDSCLK_DISABLE = 0x2,
+	HDMI_MC_CLKDIS_PIXELCLK_DISABLE = 0x1,
+
+/* MC_SWRSTZ field values */
+	HDMI_MC_SWRSTZ_TMDSSWRST_REQ = 0x02,
+
+/* MC_FLOWCTRL field values */
+	HDMI_MC_FLOWCTRL_FEED_THROUGH_OFF_MASK = 0x1,
+	HDMI_MC_FLOWCTRL_FEED_THROUGH_OFF_CSC_IN_PATH = 0x1,
+	HDMI_MC_FLOWCTRL_FEED_THROUGH_OFF_CSC_BYPASS = 0x0,
+
+/* MC_PHYRSTZ field values */
+	HDMI_MC_PHYRSTZ_ASSERT = 0x0,
+	HDMI_MC_PHYRSTZ_DEASSERT = 0x1,
+
+/* MC_HEACPHY_RST field values */
+	HDMI_MC_HEACPHY_RST_ASSERT = 0x1,
+	HDMI_MC_HEACPHY_RST_DEASSERT = 0x0,
+
+/* CSC_CFG field values */
+	HDMI_CSC_CFG_INTMODE_MASK = 0x30,
+	HDMI_CSC_CFG_INTMODE_OFFSET = 4,
+	HDMI_CSC_CFG_INTMODE_DISABLE = 0x00,
+	HDMI_CSC_CFG_INTMODE_CHROMA_INT_FORMULA1 = 0x10,
+	HDMI_CSC_CFG_INTMODE_CHROMA_INT_FORMULA2 = 0x20,
+	HDMI_CSC_CFG_DECMODE_MASK = 0x3,
+	HDMI_CSC_CFG_DECMODE_OFFSET = 0,
+	HDMI_CSC_CFG_DECMODE_DISABLE = 0x0,
+	HDMI_CSC_CFG_DECMODE_CHROMA_INT_FORMULA1 = 0x1,
+	HDMI_CSC_CFG_DECMODE_CHROMA_INT_FORMULA2 = 0x2,
+	HDMI_CSC_CFG_DECMODE_CHROMA_INT_FORMULA3 = 0x3,
+
+/* CSC_SCALE field values */
+	HDMI_CSC_SCALE_CSC_COLORDE_PTH_MASK = 0xF0,
+	HDMI_CSC_SCALE_CSC_COLORDE_PTH_24BPP = 0x00,
+	HDMI_CSC_SCALE_CSC_COLORDE_PTH_30BPP = 0x50,
+	HDMI_CSC_SCALE_CSC_COLORDE_PTH_36BPP = 0x60,
+	HDMI_CSC_SCALE_CSC_COLORDE_PTH_48BPP = 0x70,
+	HDMI_CSC_SCALE_CSCSCALE_MASK = 0x03,
+
+/* A_HDCPCFG0 field values */
+	HDMI_A_HDCPCFG0_ELVENA_MASK = 0x80,
+	HDMI_A_HDCPCFG0_ELVENA_ENABLE = 0x80,
+	HDMI_A_HDCPCFG0_ELVENA_DISABLE = 0x00,
+	HDMI_A_HDCPCFG0_I2CFASTMODE_MASK = 0x40,
+	HDMI_A_HDCPCFG0_I2CFASTMODE_ENABLE = 0x40,
+	HDMI_A_HDCPCFG0_I2CFASTMODE_DISABLE = 0x00,
+	HDMI_A_HDCPCFG0_BYPENCRYPTION_MASK = 0x20,
+	HDMI_A_HDCPCFG0_BYPENCRYPTION_ENABLE = 0x20,
+	HDMI_A_HDCPCFG0_BYPENCRYPTION_DISABLE = 0x00,
+	HDMI_A_HDCPCFG0_SYNCRICHECK_MASK = 0x10,
+	HDMI_A_HDCPCFG0_SYNCRICHECK_ENABLE = 0x10,
+	HDMI_A_HDCPCFG0_SYNCRICHECK_DISABLE = 0x00,
+	HDMI_A_HDCPCFG0_AVMUTE_MASK = 0x8,
+	HDMI_A_HDCPCFG0_AVMUTE_ENABLE = 0x8,
+	HDMI_A_HDCPCFG0_AVMUTE_DISABLE = 0x0,
+	HDMI_A_HDCPCFG0_RXDETECT_MASK = 0x4,
+	HDMI_A_HDCPCFG0_RXDETECT_ENABLE = 0x4,
+	HDMI_A_HDCPCFG0_RXDETECT_DISABLE = 0x0,
+	HDMI_A_HDCPCFG0_EN11FEATURE_MASK = 0x2,
+	HDMI_A_HDCPCFG0_EN11FEATURE_ENABLE = 0x2,
+	HDMI_A_HDCPCFG0_EN11FEATURE_DISABLE = 0x0,
+	HDMI_A_HDCPCFG0_HDMIDVI_MASK = 0x1,
+	HDMI_A_HDCPCFG0_HDMIDVI_HDMI = 0x1,
+	HDMI_A_HDCPCFG0_HDMIDVI_DVI = 0x0,
+
+/* A_HDCPCFG1 field values */
+	HDMI_A_HDCPCFG1_DISSHA1CHECK_MASK = 0x8,
+	HDMI_A_HDCPCFG1_DISSHA1CHECK_DISABLE = 0x8,
+	HDMI_A_HDCPCFG1_DISSHA1CHECK_ENABLE = 0x0,
+	HDMI_A_HDCPCFG1_PH2UPSHFTENC_MASK = 0x4,
+	HDMI_A_HDCPCFG1_PH2UPSHFTENC_ENABLE = 0x4,
+	HDMI_A_HDCPCFG1_PH2UPSHFTENC_DISABLE = 0x0,
+	HDMI_A_HDCPCFG1_ENCRYPTIONDISABLE_MASK = 0x2,
+	HDMI_A_HDCPCFG1_ENCRYPTIONDISABLE_DISABLE = 0x2,
+	HDMI_A_HDCPCFG1_ENCRYPTIONDISABLE_ENABLE = 0x0,
+	HDMI_A_HDCPCFG1_SWRESET_MASK = 0x1,
+	HDMI_A_HDCPCFG1_SWRESET_ASSERT = 0x0,
+
+/* A_VIDPOLCFG field values */
+	HDMI_A_VIDPOLCFG_UNENCRYPTCONF_MASK = 0x60,
+	HDMI_A_VIDPOLCFG_UNENCRYPTCONF_OFFSET = 5,
+	HDMI_A_VIDPOLCFG_DATAENPOL_MASK = 0x10,
+	HDMI_A_VIDPOLCFG_DATAENPOL_ACTIVE_HIGH = 0x10,
+	HDMI_A_VIDPOLCFG_DATAENPOL_ACTIVE_LOW = 0x0,
+	HDMI_A_VIDPOLCFG_VSYNCPOL_MASK = 0x8,
+	HDMI_A_VIDPOLCFG_VSYNCPOL_ACTIVE_HIGH = 0x8,
+	HDMI_A_VIDPOLCFG_VSYNCPOL_ACTIVE_LOW = 0x0,
+	HDMI_A_VIDPOLCFG_HSYNCPOL_MASK = 0x2,
+	HDMI_A_VIDPOLCFG_HSYNCPOL_ACTIVE_HIGH = 0x2,
+	HDMI_A_VIDPOLCFG_HSYNCPOL_ACTIVE_LOW = 0x0,
+};
+#endif /* __IMX_HDMI_H__ */
diff --git a/drivers/gpu/drm/imx/imx-ldb.c b/drivers/gpu/drm/imx/imx-ldb.c
new file mode 100644
index 000000000000..c60460043e24
--- /dev/null
+++ b/drivers/gpu/drm/imx/imx-ldb.c
@@ -0,0 +1,610 @@
+/*
+ * i.MX drm driver - LVDS display bridge
+ *
+ * Copyright (C) 2012 Sascha Hauer, Pengutronix
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License
+ * as published by the Free Software Foundation; either version 2
+ * of the License, or (at your option) any later version.
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ */
+
+#include <linux/module.h>
+#include <linux/clk.h>
+#include <linux/component.h>
+#include <drm/drmP.h>
+#include <drm/drm_fb_helper.h>
+#include <drm/drm_crtc_helper.h>
+#include <linux/mfd/syscon.h>
+#include <linux/mfd/syscon/imx6q-iomuxc-gpr.h>
+#include <linux/of_address.h>
+#include <linux/of_device.h>
+#include <video/of_videomode.h>
+#include <linux/regmap.h>
+#include <linux/videodev2.h>
+
+#include "imx-drm.h"
+
+#define DRIVER_NAME "imx-ldb"
+
+#define LDB_CH0_MODE_EN_TO_DI0		(1 << 0)
+#define LDB_CH0_MODE_EN_TO_DI1		(3 << 0)
+#define LDB_CH0_MODE_EN_MASK		(3 << 0)
+#define LDB_CH1_MODE_EN_TO_DI0		(1 << 2)
+#define LDB_CH1_MODE_EN_TO_DI1		(3 << 2)
+#define LDB_CH1_MODE_EN_MASK		(3 << 2)
+#define LDB_SPLIT_MODE_EN		(1 << 4)
+#define LDB_DATA_WIDTH_CH0_24		(1 << 5)
+#define LDB_BIT_MAP_CH0_JEIDA		(1 << 6)
+#define LDB_DATA_WIDTH_CH1_24		(1 << 7)
+#define LDB_BIT_MAP_CH1_JEIDA		(1 << 8)
+#define LDB_DI0_VS_POL_ACT_LOW		(1 << 9)
+#define LDB_DI1_VS_POL_ACT_LOW		(1 << 10)
+#define LDB_BGREF_RMODE_INT		(1 << 15)
+
+#define con_to_imx_ldb_ch(x) container_of(x, struct imx_ldb_channel, connector)
+#define enc_to_imx_ldb_ch(x) container_of(x, struct imx_ldb_channel, encoder)
+
+struct imx_ldb;
+
+struct imx_ldb_channel {
+	struct imx_ldb *ldb;
+	struct drm_connector connector;
+	struct drm_encoder encoder;
+	struct device_node *child;
+	int chno;
+	void *edid;
+	int edid_len;
+	struct drm_display_mode mode;
+	int mode_valid;
+};
+
+struct bus_mux {
+	int reg;
+	int shift;
+	int mask;
+};
+
+struct imx_ldb {
+	struct regmap *regmap;
+	struct device *dev;
+	struct imx_ldb_channel channel[2];
+	struct clk *clk[2]; /* our own clock */
+	struct clk *clk_sel[4]; /* parent of display clock */
+	struct clk *clk_pll[2]; /* upstream clock we can adjust */
+	u32 ldb_ctrl;
+	const struct bus_mux *lvds_mux;
+};
+
+static enum drm_connector_status imx_ldb_connector_detect(
+		struct drm_connector *connector, bool force)
+{
+	return connector_status_connected;
+}
+
+static int imx_ldb_connector_get_modes(struct drm_connector *connector)
+{
+	struct imx_ldb_channel *imx_ldb_ch = con_to_imx_ldb_ch(connector);
+	int num_modes = 0;
+
+	if (imx_ldb_ch->edid) {
+		drm_mode_connector_update_edid_property(connector,
+							imx_ldb_ch->edid);
+		num_modes = drm_add_edid_modes(connector, imx_ldb_ch->edid);
+	}
+
+	if (imx_ldb_ch->mode_valid) {
+		struct drm_display_mode *mode;
+
+		mode = drm_mode_create(connector->dev);
+		if (!mode)
+			return -EINVAL;
+		drm_mode_copy(mode, &imx_ldb_ch->mode);
+		mode->type |= DRM_MODE_TYPE_DRIVER | DRM_MODE_TYPE_PREFERRED;
+		drm_mode_probed_add(connector, mode);
+		num_modes++;
+	}
+
+	return num_modes;
+}
+
+static struct drm_encoder *imx_ldb_connector_best_encoder(
+		struct drm_connector *connector)
+{
+	struct imx_ldb_channel *imx_ldb_ch = con_to_imx_ldb_ch(connector);
+
+	return &imx_ldb_ch->encoder;
+}
+
+static void imx_ldb_encoder_dpms(struct drm_encoder *encoder, int mode)
+{
+}
+
+static bool imx_ldb_encoder_mode_fixup(struct drm_encoder *encoder,
+			   const struct drm_display_mode *mode,
+			   struct drm_display_mode *adjusted_mode)
+{
+	return true;
+}
+
+static void imx_ldb_set_clock(struct imx_ldb *ldb, int mux, int chno,
+		unsigned long serial_clk, unsigned long di_clk)
+{
+	int ret;
+
+	dev_dbg(ldb->dev, "%s: now: %ld want: %ld\n", __func__,
+			clk_get_rate(ldb->clk_pll[chno]), serial_clk);
+	clk_set_rate(ldb->clk_pll[chno], serial_clk);
+
+	dev_dbg(ldb->dev, "%s after: %ld\n", __func__,
+			clk_get_rate(ldb->clk_pll[chno]));
+
+	dev_dbg(ldb->dev, "%s: now: %ld want: %ld\n", __func__,
+			clk_get_rate(ldb->clk[chno]),
+			(long int)di_clk);
+	clk_set_rate(ldb->clk[chno], di_clk);
+
+	dev_dbg(ldb->dev, "%s after: %ld\n", __func__,
+			clk_get_rate(ldb->clk[chno]));
+
+	/* set display clock mux to LDB input clock */
+	ret = clk_set_parent(ldb->clk_sel[mux], ldb->clk[chno]);
+	if (ret)
+		dev_err(ldb->dev,
+			"unable to set di%d parent clock to ldb_di%d\n", mux,
+			chno);
+}
+
+static void imx_ldb_encoder_prepare(struct drm_encoder *encoder)
+{
+	struct imx_ldb_channel *imx_ldb_ch = enc_to_imx_ldb_ch(encoder);
+	struct imx_ldb *ldb = imx_ldb_ch->ldb;
+	struct drm_display_mode *mode = &encoder->crtc->mode;
+	u32 pixel_fmt;
+	unsigned long serial_clk;
+	unsigned long di_clk = mode->clock * 1000;
+	int mux = imx_drm_encoder_get_mux_id(imx_ldb_ch->child, encoder);
+
+	if (ldb->ldb_ctrl & LDB_SPLIT_MODE_EN) {
+		/* dual channel LVDS mode */
+		serial_clk = 3500UL * mode->clock;
+		imx_ldb_set_clock(ldb, mux, 0, serial_clk, di_clk);
+		imx_ldb_set_clock(ldb, mux, 1, serial_clk, di_clk);
+	} else {
+		serial_clk = 7000UL * mode->clock;
+		imx_ldb_set_clock(ldb, mux, imx_ldb_ch->chno, serial_clk,
+				di_clk);
+	}
+
+	switch (imx_ldb_ch->chno) {
+	case 0:
+		pixel_fmt = (ldb->ldb_ctrl & LDB_DATA_WIDTH_CH0_24) ?
+			V4L2_PIX_FMT_RGB24 : V4L2_PIX_FMT_BGR666;
+		break;
+	case 1:
+		pixel_fmt = (ldb->ldb_ctrl & LDB_DATA_WIDTH_CH1_24) ?
+			V4L2_PIX_FMT_RGB24 : V4L2_PIX_FMT_BGR666;
+		break;
+	default:
+		dev_err(ldb->dev, "unable to config di%d panel format\n",
+			imx_ldb_ch->chno);
+		pixel_fmt = V4L2_PIX_FMT_RGB24;
+	}
+
+	imx_drm_panel_format(encoder, pixel_fmt);
+}
+
+static void imx_ldb_encoder_commit(struct drm_encoder *encoder)
+{
+	struct imx_ldb_channel *imx_ldb_ch = enc_to_imx_ldb_ch(encoder);
+	struct imx_ldb *ldb = imx_ldb_ch->ldb;
+	int dual = ldb->ldb_ctrl & LDB_SPLIT_MODE_EN;
+	int mux = imx_drm_encoder_get_mux_id(imx_ldb_ch->child, encoder);
+
+	if (dual) {
+		clk_prepare_enable(ldb->clk[0]);
+		clk_prepare_enable(ldb->clk[1]);
+	}
+
+	if (imx_ldb_ch == &ldb->channel[0] || dual) {
+		ldb->ldb_ctrl &= ~LDB_CH0_MODE_EN_MASK;
+		if (mux == 0 || ldb->lvds_mux)
+			ldb->ldb_ctrl |= LDB_CH0_MODE_EN_TO_DI0;
+		else if (mux == 1)
+			ldb->ldb_ctrl |= LDB_CH0_MODE_EN_TO_DI1;
+	}
+	if (imx_ldb_ch == &ldb->channel[1] || dual) {
+		ldb->ldb_ctrl &= ~LDB_CH1_MODE_EN_MASK;
+		if (mux == 1 || ldb->lvds_mux)
+			ldb->ldb_ctrl |= LDB_CH1_MODE_EN_TO_DI1;
+		else if (mux == 0)
+			ldb->ldb_ctrl |= LDB_CH1_MODE_EN_TO_DI0;
+	}
+
+	if (ldb->lvds_mux) {
+		const struct bus_mux *lvds_mux = NULL;
+
+		if (imx_ldb_ch == &ldb->channel[0])
+			lvds_mux = &ldb->lvds_mux[0];
+		else if (imx_ldb_ch == &ldb->channel[1])
+			lvds_mux = &ldb->lvds_mux[1];
+
+		regmap_update_bits(ldb->regmap, lvds_mux->reg, lvds_mux->mask,
+				   mux << lvds_mux->shift);
+	}
+
+	regmap_write(ldb->regmap, IOMUXC_GPR2, ldb->ldb_ctrl);
+}
+
+static void imx_ldb_encoder_mode_set(struct drm_encoder *encoder,
+			 struct drm_display_mode *mode,
+			 struct drm_display_mode *adjusted_mode)
+{
+	struct imx_ldb_channel *imx_ldb_ch = enc_to_imx_ldb_ch(encoder);
+	struct imx_ldb *ldb = imx_ldb_ch->ldb;
+	int dual = ldb->ldb_ctrl & LDB_SPLIT_MODE_EN;
+
+	if (mode->clock > 170000) {
+		dev_warn(ldb->dev,
+			 "%s: mode exceeds 170 MHz pixel clock\n", __func__);
+	}
+	if (mode->clock > 85000 && !dual) {
+		dev_warn(ldb->dev,
+			 "%s: mode exceeds 85 MHz pixel clock\n", __func__);
+	}
+
+	/* FIXME - assumes straight connections DI0 --> CH0, DI1 --> CH1 */
+	if (imx_ldb_ch == &ldb->channel[0]) {
+		if (mode->flags & DRM_MODE_FLAG_NVSYNC)
+			ldb->ldb_ctrl |= LDB_DI0_VS_POL_ACT_LOW;
+		else if (mode->flags & DRM_MODE_FLAG_PVSYNC)
+			ldb->ldb_ctrl &= ~LDB_DI0_VS_POL_ACT_LOW;
+	}
+	if (imx_ldb_ch == &ldb->channel[1]) {
+		if (mode->flags & DRM_MODE_FLAG_NVSYNC)
+			ldb->ldb_ctrl |= LDB_DI1_VS_POL_ACT_LOW;
+		else if (mode->flags & DRM_MODE_FLAG_PVSYNC)
+			ldb->ldb_ctrl &= ~LDB_DI1_VS_POL_ACT_LOW;
+	}
+}
+
+static void imx_ldb_encoder_disable(struct drm_encoder *encoder)
+{
+	struct imx_ldb_channel *imx_ldb_ch = enc_to_imx_ldb_ch(encoder);
+	struct imx_ldb *ldb = imx_ldb_ch->ldb;
+
+	/*
+	 * imx_ldb_encoder_disable is called by
+	 * drm_helper_disable_unused_functions without
+	 * the encoder being enabled before.
+	 */
+	if (imx_ldb_ch == &ldb->channel[0] &&
+	    (ldb->ldb_ctrl & LDB_CH0_MODE_EN_MASK) == 0)
+		return;
+	else if (imx_ldb_ch == &ldb->channel[1] &&
+		 (ldb->ldb_ctrl & LDB_CH1_MODE_EN_MASK) == 0)
+		return;
+
+	if (imx_ldb_ch == &ldb->channel[0])
+		ldb->ldb_ctrl &= ~LDB_CH0_MODE_EN_MASK;
+	else if (imx_ldb_ch == &ldb->channel[1])
+		ldb->ldb_ctrl &= ~LDB_CH1_MODE_EN_MASK;
+
+	regmap_write(ldb->regmap, IOMUXC_GPR2, ldb->ldb_ctrl);
+
+	if (ldb->ldb_ctrl & LDB_SPLIT_MODE_EN) {
+		clk_disable_unprepare(ldb->clk[0]);
+		clk_disable_unprepare(ldb->clk[1]);
+	}
+}
+
+static struct drm_connector_funcs imx_ldb_connector_funcs = {
+	.dpms = drm_helper_connector_dpms,
+	.fill_modes = drm_helper_probe_single_connector_modes,
+	.detect = imx_ldb_connector_detect,
+	.destroy = imx_drm_connector_destroy,
+};
+
+static struct drm_connector_helper_funcs imx_ldb_connector_helper_funcs = {
+	.get_modes = imx_ldb_connector_get_modes,
+	.best_encoder = imx_ldb_connector_best_encoder,
+};
+
+static struct drm_encoder_funcs imx_ldb_encoder_funcs = {
+	.destroy = imx_drm_encoder_destroy,
+};
+
+static struct drm_encoder_helper_funcs imx_ldb_encoder_helper_funcs = {
+	.dpms = imx_ldb_encoder_dpms,
+	.mode_fixup = imx_ldb_encoder_mode_fixup,
+	.prepare = imx_ldb_encoder_prepare,
+	.commit = imx_ldb_encoder_commit,
+	.mode_set = imx_ldb_encoder_mode_set,
+	.disable = imx_ldb_encoder_disable,
+};
+
+static int imx_ldb_get_clk(struct imx_ldb *ldb, int chno)
+{
+	char clkname[16];
+
+	snprintf(clkname, sizeof(clkname), "di%d", chno);
+	ldb->clk[chno] = devm_clk_get(ldb->dev, clkname);
+	if (IS_ERR(ldb->clk[chno]))
+		return PTR_ERR(ldb->clk[chno]);
+
+	snprintf(clkname, sizeof(clkname), "di%d_pll", chno);
+	ldb->clk_pll[chno] = devm_clk_get(ldb->dev, clkname);
+
+	return PTR_ERR_OR_ZERO(ldb->clk_pll[chno]);
+}
+
+static int imx_ldb_register(struct drm_device *drm,
+	struct imx_ldb_channel *imx_ldb_ch)
+{
+	struct imx_ldb *ldb = imx_ldb_ch->ldb;
+	int ret;
+
+	ret = imx_drm_encoder_parse_of(drm, &imx_ldb_ch->encoder,
+				       imx_ldb_ch->child);
+	if (ret)
+		return ret;
+
+	ret = imx_ldb_get_clk(ldb, imx_ldb_ch->chno);
+	if (ret)
+		return ret;
+
+	if (ldb->ldb_ctrl & LDB_SPLIT_MODE_EN) {
+		ret = imx_ldb_get_clk(ldb, 1);
+		if (ret)
+			return ret;
+	}
+
+	drm_encoder_helper_add(&imx_ldb_ch->encoder,
+			&imx_ldb_encoder_helper_funcs);
+	drm_encoder_init(drm, &imx_ldb_ch->encoder, &imx_ldb_encoder_funcs,
+			 DRM_MODE_ENCODER_LVDS);
+
+	drm_connector_helper_add(&imx_ldb_ch->connector,
+			&imx_ldb_connector_helper_funcs);
+	drm_connector_init(drm, &imx_ldb_ch->connector,
+			   &imx_ldb_connector_funcs, DRM_MODE_CONNECTOR_LVDS);
+
+	drm_mode_connector_attach_encoder(&imx_ldb_ch->connector,
+			&imx_ldb_ch->encoder);
+
+	return 0;
+}
+
+enum {
+	LVDS_BIT_MAP_SPWG,
+	LVDS_BIT_MAP_JEIDA
+};
+
+static const char * const imx_ldb_bit_mappings[] = {
+	[LVDS_BIT_MAP_SPWG]  = "spwg",
+	[LVDS_BIT_MAP_JEIDA] = "jeida",
+};
+
+static const int of_get_data_mapping(struct device_node *np)
+{
+	const char *bm;
+	int ret, i;
+
+	ret = of_property_read_string(np, "fsl,data-mapping", &bm);
+	if (ret < 0)
+		return ret;
+
+	for (i = 0; i < ARRAY_SIZE(imx_ldb_bit_mappings); i++)
+		if (!strcasecmp(bm, imx_ldb_bit_mappings[i]))
+			return i;
+
+	return -EINVAL;
+}
+
+static struct bus_mux imx6q_lvds_mux[2] = {
+	{
+		.reg = IOMUXC_GPR3,
+		.shift = 6,
+		.mask = IMX6Q_GPR3_LVDS0_MUX_CTL_MASK,
+	}, {
+		.reg = IOMUXC_GPR3,
+		.shift = 8,
+		.mask = IMX6Q_GPR3_LVDS1_MUX_CTL_MASK,
+	}
+};
+
+/*
+ * For a device declaring compatible = "fsl,imx6q-ldb", "fsl,imx53-ldb",
+ * of_match_device will walk through this list and take the first entry
+ * matching any of its compatible values. Therefore, the more generic
+ * entries (in this case fsl,imx53-ldb) need to be ordered last.
+ */
+static const struct of_device_id imx_ldb_dt_ids[] = {
+	{ .compatible = "fsl,imx6q-ldb", .data = imx6q_lvds_mux, },
+	{ .compatible = "fsl,imx53-ldb", .data = NULL, },
+	{ }
+};
+MODULE_DEVICE_TABLE(of, imx_ldb_dt_ids);
+
+static int imx_ldb_bind(struct device *dev, struct device *master, void *data)
+{
+	struct drm_device *drm = data;
+	struct device_node *np = dev->of_node;
+	const struct of_device_id *of_id =
+			of_match_device(imx_ldb_dt_ids, dev);
+	struct device_node *child;
+	const u8 *edidp;
+	struct imx_ldb *imx_ldb;
+	int datawidth;
+	int mapping;
+	int dual;
+	int ret;
+	int i;
+
+	imx_ldb = devm_kzalloc(dev, sizeof(*imx_ldb), GFP_KERNEL);
+	if (!imx_ldb)
+		return -ENOMEM;
+
+	imx_ldb->regmap = syscon_regmap_lookup_by_phandle(np, "gpr");
+	if (IS_ERR(imx_ldb->regmap)) {
+		dev_err(dev, "failed to get parent regmap\n");
+		return PTR_ERR(imx_ldb->regmap);
+	}
+
+	imx_ldb->dev = dev;
+
+	if (of_id)
+		imx_ldb->lvds_mux = of_id->data;
+
+	dual = of_property_read_bool(np, "fsl,dual-channel");
+	if (dual)
+		imx_ldb->ldb_ctrl |= LDB_SPLIT_MODE_EN;
+
+	/*
+	 * There are three different possible clock mux configurations:
+	 * i.MX53:  ipu1_di0_sel, ipu1_di1_sel
+	 * i.MX6q:  ipu1_di0_sel, ipu1_di1_sel, ipu2_di0_sel, ipu2_di1_sel
+	 * i.MX6dl: ipu1_di0_sel, ipu1_di1_sel, lcdif_sel
+	 * Map them all to di0_sel...di3_sel.
+	 */
+	for (i = 0; i < 4; i++) {
+		char clkname[16];
+
+		sprintf(clkname, "di%d_sel", i);
+		imx_ldb->clk_sel[i] = devm_clk_get(imx_ldb->dev, clkname);
+		if (IS_ERR(imx_ldb->clk_sel[i])) {
+			ret = PTR_ERR(imx_ldb->clk_sel[i]);
+			imx_ldb->clk_sel[i] = NULL;
+			break;
+		}
+	}
+	if (i == 0)
+		return ret;
+
+	for_each_child_of_node(np, child) {
+		struct imx_ldb_channel *channel;
+
+		ret = of_property_read_u32(child, "reg", &i);
+		if (ret || i < 0 || i > 1)
+			return -EINVAL;
+
+		if (dual && i > 0) {
+			dev_warn(dev, "dual-channel mode, ignoring second output\n");
+			continue;
+		}
+
+		if (!of_device_is_available(child))
+			continue;
+
+		channel = &imx_ldb->channel[i];
+		channel->ldb = imx_ldb;
+		channel->chno = i;
+		channel->child = child;
+
+		edidp = of_get_property(child, "edid", &channel->edid_len);
+		if (edidp) {
+			channel->edid = kmemdup(edidp, channel->edid_len,
+						GFP_KERNEL);
+		} else {
+			ret = of_get_drm_display_mode(child, &channel->mode, 0);
+			if (!ret)
+				channel->mode_valid = 1;
+		}
+
+		ret = of_property_read_u32(child, "fsl,data-width", &datawidth);
+		if (ret)
+			datawidth = 0;
+		else if (datawidth != 18 && datawidth != 24)
+			return -EINVAL;
+
+		mapping = of_get_data_mapping(child);
+		switch (mapping) {
+		case LVDS_BIT_MAP_SPWG:
+			if (datawidth == 24) {
+				if (i == 0 || dual)
+					imx_ldb->ldb_ctrl |=
+						LDB_DATA_WIDTH_CH0_24;
+				if (i == 1 || dual)
+					imx_ldb->ldb_ctrl |=
+						LDB_DATA_WIDTH_CH1_24;
+			}
+			break;
+		case LVDS_BIT_MAP_JEIDA:
+			if (datawidth == 18) {
+				dev_err(dev, "JEIDA standard only supported in 24 bit\n");
+				return -EINVAL;
+			}
+			if (i == 0 || dual)
+				imx_ldb->ldb_ctrl |= LDB_DATA_WIDTH_CH0_24 |
+					LDB_BIT_MAP_CH0_JEIDA;
+			if (i == 1 || dual)
+				imx_ldb->ldb_ctrl |= LDB_DATA_WIDTH_CH1_24 |
+					LDB_BIT_MAP_CH1_JEIDA;
+			break;
+		default:
+			dev_err(dev, "data mapping not specified or invalid\n");
+			return -EINVAL;
+		}
+
+		ret = imx_ldb_register(drm, channel);
+		if (ret)
+			return ret;
+	}
+
+	dev_set_drvdata(dev, imx_ldb);
+
+	return 0;
+}
+
+static void imx_ldb_unbind(struct device *dev, struct device *master,
+	void *data)
+{
+	struct imx_ldb *imx_ldb = dev_get_drvdata(dev);
+	int i;
+
+	for (i = 0; i < 2; i++) {
+		struct imx_ldb_channel *channel = &imx_ldb->channel[i];
+
+		if (!channel->connector.funcs)
+			continue;
+
+		channel->connector.funcs->destroy(&channel->connector);
+		channel->encoder.funcs->destroy(&channel->encoder);
+	}
+}
+
+static const struct component_ops imx_ldb_ops = {
+	.bind	= imx_ldb_bind,
+	.unbind	= imx_ldb_unbind,
+};
+
+static int imx_ldb_probe(struct platform_device *pdev)
+{
+	return component_add(&pdev->dev, &imx_ldb_ops);
+}
+
+static int imx_ldb_remove(struct platform_device *pdev)
+{
+	component_del(&pdev->dev, &imx_ldb_ops);
+	return 0;
+}
+
+static struct platform_driver imx_ldb_driver = {
+	.probe		= imx_ldb_probe,
+	.remove		= imx_ldb_remove,
+	.driver		= {
+		.of_match_table = imx_ldb_dt_ids,
+		.name	= DRIVER_NAME,
+	},
+};
+
+module_platform_driver(imx_ldb_driver);
+
+MODULE_DESCRIPTION("i.MX LVDS driver");
+MODULE_AUTHOR("Sascha Hauer, Pengutronix");
+MODULE_LICENSE("GPL");
+MODULE_ALIAS("platform:" DRIVER_NAME);
diff --git a/drivers/gpu/drm/imx/imx-tve.c b/drivers/gpu/drm/imx/imx-tve.c
new file mode 100644
index 000000000000..a729f4f7074c
--- /dev/null
+++ b/drivers/gpu/drm/imx/imx-tve.c
@@ -0,0 +1,731 @@
+/*
+ * i.MX drm driver - Television Encoder (TVEv2)
+ *
+ * Copyright (C) 2013 Philipp Zabel, Pengutronix
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License
+ * as published by the Free Software Foundation; either version 2
+ * of the License, or (at your option) any later version.
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ */
+
+#include <linux/clk.h>
+#include <linux/clk-provider.h>
+#include <linux/component.h>
+#include <linux/module.h>
+#include <linux/i2c.h>
+#include <linux/regmap.h>
+#include <linux/regulator/consumer.h>
+#include <linux/spinlock.h>
+#include <linux/videodev2.h>
+#include <drm/drmP.h>
+#include <drm/drm_fb_helper.h>
+#include <drm/drm_crtc_helper.h>
+#include <video/imx-ipu-v3.h>
+
+#include "imx-drm.h"
+
+#define TVE_COM_CONF_REG	0x00
+#define TVE_TVDAC0_CONT_REG	0x28
+#define TVE_TVDAC1_CONT_REG	0x2c
+#define TVE_TVDAC2_CONT_REG	0x30
+#define TVE_CD_CONT_REG		0x34
+#define TVE_INT_CONT_REG	0x64
+#define TVE_STAT_REG		0x68
+#define TVE_TST_MODE_REG	0x6c
+#define TVE_MV_CONT_REG		0xdc
+
+/* TVE_COM_CONF_REG */
+#define TVE_SYNC_CH_2_EN	BIT(22)
+#define TVE_SYNC_CH_1_EN	BIT(21)
+#define TVE_SYNC_CH_0_EN	BIT(20)
+#define TVE_TV_OUT_MODE_MASK	(0x7 << 12)
+#define TVE_TV_OUT_DISABLE	(0x0 << 12)
+#define TVE_TV_OUT_CVBS_0	(0x1 << 12)
+#define TVE_TV_OUT_CVBS_2	(0x2 << 12)
+#define TVE_TV_OUT_CVBS_0_2	(0x3 << 12)
+#define TVE_TV_OUT_SVIDEO_0_1	(0x4 << 12)
+#define TVE_TV_OUT_SVIDEO_0_1_CVBS2_2	(0x5 << 12)
+#define TVE_TV_OUT_YPBPR	(0x6 << 12)
+#define TVE_TV_OUT_RGB		(0x7 << 12)
+#define TVE_TV_STAND_MASK	(0xf << 8)
+#define TVE_TV_STAND_HD_1080P30	(0xc << 8)
+#define TVE_P2I_CONV_EN		BIT(7)
+#define TVE_INP_VIDEO_FORM	BIT(6)
+#define TVE_INP_YCBCR_422	(0x0 << 6)
+#define TVE_INP_YCBCR_444	(0x1 << 6)
+#define TVE_DATA_SOURCE_MASK	(0x3 << 4)
+#define TVE_DATA_SOURCE_BUS1	(0x0 << 4)
+#define TVE_DATA_SOURCE_BUS2	(0x1 << 4)
+#define TVE_DATA_SOURCE_EXT	(0x2 << 4)
+#define TVE_DATA_SOURCE_TESTGEN	(0x3 << 4)
+#define TVE_IPU_CLK_EN_OFS	3
+#define TVE_IPU_CLK_EN		BIT(3)
+#define TVE_DAC_SAMP_RATE_OFS	1
+#define TVE_DAC_SAMP_RATE_WIDTH	2
+#define TVE_DAC_SAMP_RATE_MASK	(0x3 << 1)
+#define TVE_DAC_FULL_RATE	(0x0 << 1)
+#define TVE_DAC_DIV2_RATE	(0x1 << 1)
+#define TVE_DAC_DIV4_RATE	(0x2 << 1)
+#define TVE_EN			BIT(0)
+
+/* TVE_TVDACx_CONT_REG */
+#define TVE_TVDAC_GAIN_MASK	(0x3f << 0)
+
+/* TVE_CD_CONT_REG */
+#define TVE_CD_CH_2_SM_EN	BIT(22)
+#define TVE_CD_CH_1_SM_EN	BIT(21)
+#define TVE_CD_CH_0_SM_EN	BIT(20)
+#define TVE_CD_CH_2_LM_EN	BIT(18)
+#define TVE_CD_CH_1_LM_EN	BIT(17)
+#define TVE_CD_CH_0_LM_EN	BIT(16)
+#define TVE_CD_CH_2_REF_LVL	BIT(10)
+#define TVE_CD_CH_1_REF_LVL	BIT(9)
+#define TVE_CD_CH_0_REF_LVL	BIT(8)
+#define TVE_CD_EN		BIT(0)
+
+/* TVE_INT_CONT_REG */
+#define TVE_FRAME_END_IEN	BIT(13)
+#define TVE_CD_MON_END_IEN	BIT(2)
+#define TVE_CD_SM_IEN		BIT(1)
+#define TVE_CD_LM_IEN		BIT(0)
+
+/* TVE_TST_MODE_REG */
+#define TVE_TVDAC_TEST_MODE_MASK	(0x7 << 0)
+
+#define con_to_tve(x) container_of(x, struct imx_tve, connector)
+#define enc_to_tve(x) container_of(x, struct imx_tve, encoder)
+
+enum {
+	TVE_MODE_TVOUT,
+	TVE_MODE_VGA,
+};
+
+struct imx_tve {
+	struct drm_connector connector;
+	struct drm_encoder encoder;
+	struct device *dev;
+	spinlock_t lock;	/* register lock */
+	bool enabled;
+	int mode;
+
+	struct regmap *regmap;
+	struct regulator *dac_reg;
+	struct i2c_adapter *ddc;
+	struct clk *clk;
+	struct clk *di_sel_clk;
+	struct clk_hw clk_hw_di;
+	struct clk *di_clk;
+	int vsync_pin;
+	int hsync_pin;
+};
+
+static void tve_lock(void *__tve)
+__acquires(&tve->lock)
+{
+	struct imx_tve *tve = __tve;
+
+	spin_lock(&tve->lock);
+}
+
+static void tve_unlock(void *__tve)
+__releases(&tve->lock)
+{
+	struct imx_tve *tve = __tve;
+
+	spin_unlock(&tve->lock);
+}
+
+static void tve_enable(struct imx_tve *tve)
+{
+	int ret;
+
+	if (!tve->enabled) {
+		tve->enabled = true;
+		clk_prepare_enable(tve->clk);
+		ret = regmap_update_bits(tve->regmap, TVE_COM_CONF_REG,
+					 TVE_IPU_CLK_EN | TVE_EN,
+					 TVE_IPU_CLK_EN | TVE_EN);
+	}
+
+	/* clear interrupt status register */
+	regmap_write(tve->regmap, TVE_STAT_REG, 0xffffffff);
+
+	/* cable detection irq disabled in VGA mode, enabled in TVOUT mode */
+	if (tve->mode == TVE_MODE_VGA)
+		regmap_write(tve->regmap, TVE_INT_CONT_REG, 0);
+	else
+		regmap_write(tve->regmap, TVE_INT_CONT_REG,
+			     TVE_CD_SM_IEN |
+			     TVE_CD_LM_IEN |
+			     TVE_CD_MON_END_IEN);
+}
+
+static void tve_disable(struct imx_tve *tve)
+{
+	int ret;
+
+	if (tve->enabled) {
+		tve->enabled = false;
+		ret = regmap_update_bits(tve->regmap, TVE_COM_CONF_REG,
+					 TVE_IPU_CLK_EN | TVE_EN, 0);
+		clk_disable_unprepare(tve->clk);
+	}
+}
+
+static int tve_setup_tvout(struct imx_tve *tve)
+{
+	return -ENOTSUPP;
+}
+
+static int tve_setup_vga(struct imx_tve *tve)
+{
+	unsigned int mask;
+	unsigned int val;
+	int ret;
+
+	/* set gain to (1 + 10/128) to provide 0.7V peak-to-peak amplitude */
+	ret = regmap_update_bits(tve->regmap, TVE_TVDAC0_CONT_REG,
+				 TVE_TVDAC_GAIN_MASK, 0x0a);
+	ret = regmap_update_bits(tve->regmap, TVE_TVDAC1_CONT_REG,
+				 TVE_TVDAC_GAIN_MASK, 0x0a);
+	ret = regmap_update_bits(tve->regmap, TVE_TVDAC2_CONT_REG,
+				 TVE_TVDAC_GAIN_MASK, 0x0a);
+
+	/* set configuration register */
+	mask = TVE_DATA_SOURCE_MASK | TVE_INP_VIDEO_FORM;
+	val  = TVE_DATA_SOURCE_BUS2 | TVE_INP_YCBCR_444;
+	mask |= TVE_TV_STAND_MASK       | TVE_P2I_CONV_EN;
+	val  |= TVE_TV_STAND_HD_1080P30 | 0;
+	mask |= TVE_TV_OUT_MODE_MASK | TVE_SYNC_CH_0_EN;
+	val  |= TVE_TV_OUT_RGB       | TVE_SYNC_CH_0_EN;
+	ret = regmap_update_bits(tve->regmap, TVE_COM_CONF_REG, mask, val);
+	if (ret < 0) {
+		dev_err(tve->dev, "failed to set configuration: %d\n", ret);
+		return ret;
+	}
+
+	/* set test mode (as documented) */
+	ret = regmap_update_bits(tve->regmap, TVE_TST_MODE_REG,
+				 TVE_TVDAC_TEST_MODE_MASK, 1);
+
+	return 0;
+}
+
+static enum drm_connector_status imx_tve_connector_detect(
+				struct drm_connector *connector, bool force)
+{
+	return connector_status_connected;
+}
+
+static int imx_tve_connector_get_modes(struct drm_connector *connector)
+{
+	struct imx_tve *tve = con_to_tve(connector);
+	struct edid *edid;
+	int ret = 0;
+
+	if (!tve->ddc)
+		return 0;
+
+	edid = drm_get_edid(connector, tve->ddc);
+	if (edid) {
+		drm_mode_connector_update_edid_property(connector, edid);
+		ret = drm_add_edid_modes(connector, edid);
+		kfree(edid);
+	}
+
+	return ret;
+}
+
+static int imx_tve_connector_mode_valid(struct drm_connector *connector,
+					struct drm_display_mode *mode)
+{
+	struct imx_tve *tve = con_to_tve(connector);
+	unsigned long rate;
+
+	/* pixel clock with 2x oversampling */
+	rate = clk_round_rate(tve->clk, 2000UL * mode->clock) / 2000;
+	if (rate == mode->clock)
+		return MODE_OK;
+
+	/* pixel clock without oversampling */
+	rate = clk_round_rate(tve->clk, 1000UL * mode->clock) / 1000;
+	if (rate == mode->clock)
+		return MODE_OK;
+
+	dev_warn(tve->dev, "ignoring mode %dx%d\n",
+		 mode->hdisplay, mode->vdisplay);
+
+	return MODE_BAD;
+}
+
+static struct drm_encoder *imx_tve_connector_best_encoder(
+		struct drm_connector *connector)
+{
+	struct imx_tve *tve = con_to_tve(connector);
+
+	return &tve->encoder;
+}
+
+static void imx_tve_encoder_dpms(struct drm_encoder *encoder, int mode)
+{
+	struct imx_tve *tve = enc_to_tve(encoder);
+	int ret;
+
+	ret = regmap_update_bits(tve->regmap, TVE_COM_CONF_REG,
+				 TVE_TV_OUT_MODE_MASK, TVE_TV_OUT_DISABLE);
+	if (ret < 0)
+		dev_err(tve->dev, "failed to disable TVOUT: %d\n", ret);
+}
+
+static bool imx_tve_encoder_mode_fixup(struct drm_encoder *encoder,
+				       const struct drm_display_mode *mode,
+				       struct drm_display_mode *adjusted_mode)
+{
+	return true;
+}
+
+static void imx_tve_encoder_prepare(struct drm_encoder *encoder)
+{
+	struct imx_tve *tve = enc_to_tve(encoder);
+
+	tve_disable(tve);
+
+	switch (tve->mode) {
+	case TVE_MODE_VGA:
+		imx_drm_panel_format_pins(encoder, IPU_PIX_FMT_GBR24,
+				tve->hsync_pin, tve->vsync_pin);
+		break;
+	case TVE_MODE_TVOUT:
+		imx_drm_panel_format(encoder, V4L2_PIX_FMT_YUV444);
+		break;
+	}
+}
+
+static void imx_tve_encoder_mode_set(struct drm_encoder *encoder,
+				     struct drm_display_mode *mode,
+				     struct drm_display_mode *adjusted_mode)
+{
+	struct imx_tve *tve = enc_to_tve(encoder);
+	unsigned long rounded_rate;
+	unsigned long rate;
+	int div = 1;
+	int ret;
+
+	/*
+	 * FIXME
+	 * we should try 4k * mode->clock first,
+	 * and enable 4x oversampling for lower resolutions
+	 */
+	rate = 2000UL * mode->clock;
+	clk_set_rate(tve->clk, rate);
+	rounded_rate = clk_get_rate(tve->clk);
+	if (rounded_rate >= rate)
+		div = 2;
+	clk_set_rate(tve->di_clk, rounded_rate / div);
+
+	ret = clk_set_parent(tve->di_sel_clk, tve->di_clk);
+	if (ret < 0) {
+		dev_err(tve->dev, "failed to set di_sel parent to tve_di: %d\n",
+			ret);
+	}
+
+	if (tve->mode == TVE_MODE_VGA)
+		tve_setup_vga(tve);
+	else
+		tve_setup_tvout(tve);
+}
+
+static void imx_tve_encoder_commit(struct drm_encoder *encoder)
+{
+	struct imx_tve *tve = enc_to_tve(encoder);
+
+	tve_enable(tve);
+}
+
+static void imx_tve_encoder_disable(struct drm_encoder *encoder)
+{
+	struct imx_tve *tve = enc_to_tve(encoder);
+
+	tve_disable(tve);
+}
+
+static struct drm_connector_funcs imx_tve_connector_funcs = {
+	.dpms = drm_helper_connector_dpms,
+	.fill_modes = drm_helper_probe_single_connector_modes,
+	.detect = imx_tve_connector_detect,
+	.destroy = imx_drm_connector_destroy,
+};
+
+static struct drm_connector_helper_funcs imx_tve_connector_helper_funcs = {
+	.get_modes = imx_tve_connector_get_modes,
+	.best_encoder = imx_tve_connector_best_encoder,
+	.mode_valid = imx_tve_connector_mode_valid,
+};
+
+static struct drm_encoder_funcs imx_tve_encoder_funcs = {
+	.destroy = imx_drm_encoder_destroy,
+};
+
+static struct drm_encoder_helper_funcs imx_tve_encoder_helper_funcs = {
+	.dpms = imx_tve_encoder_dpms,
+	.mode_fixup = imx_tve_encoder_mode_fixup,
+	.prepare = imx_tve_encoder_prepare,
+	.mode_set = imx_tve_encoder_mode_set,
+	.commit = imx_tve_encoder_commit,
+	.disable = imx_tve_encoder_disable,
+};
+
+static irqreturn_t imx_tve_irq_handler(int irq, void *data)
+{
+	struct imx_tve *tve = data;
+	unsigned int val;
+
+	regmap_read(tve->regmap, TVE_STAT_REG, &val);
+
+	/* clear interrupt status register */
+	regmap_write(tve->regmap, TVE_STAT_REG, 0xffffffff);
+
+	return IRQ_HANDLED;
+}
+
+static unsigned long clk_tve_di_recalc_rate(struct clk_hw *hw,
+					    unsigned long parent_rate)
+{
+	struct imx_tve *tve = container_of(hw, struct imx_tve, clk_hw_di);
+	unsigned int val;
+	int ret;
+
+	ret = regmap_read(tve->regmap, TVE_COM_CONF_REG, &val);
+	if (ret < 0)
+		return 0;
+
+	switch (val & TVE_DAC_SAMP_RATE_MASK) {
+	case TVE_DAC_DIV4_RATE:
+		return parent_rate / 4;
+	case TVE_DAC_DIV2_RATE:
+		return parent_rate / 2;
+	case TVE_DAC_FULL_RATE:
+	default:
+		return parent_rate;
+	}
+
+	return 0;
+}
+
+static long clk_tve_di_round_rate(struct clk_hw *hw, unsigned long rate,
+				  unsigned long *prate)
+{
+	unsigned long div;
+
+	div = *prate / rate;
+	if (div >= 4)
+		return *prate / 4;
+	else if (div >= 2)
+		return *prate / 2;
+	return *prate;
+}
+
+static int clk_tve_di_set_rate(struct clk_hw *hw, unsigned long rate,
+			       unsigned long parent_rate)
+{
+	struct imx_tve *tve = container_of(hw, struct imx_tve, clk_hw_di);
+	unsigned long div;
+	u32 val;
+	int ret;
+
+	div = parent_rate / rate;
+	if (div >= 4)
+		val = TVE_DAC_DIV4_RATE;
+	else if (div >= 2)
+		val = TVE_DAC_DIV2_RATE;
+	else
+		val = TVE_DAC_FULL_RATE;
+
+	ret = regmap_update_bits(tve->regmap, TVE_COM_CONF_REG,
+				 TVE_DAC_SAMP_RATE_MASK, val);
+
+	if (ret < 0) {
+		dev_err(tve->dev, "failed to set divider: %d\n", ret);
+		return ret;
+	}
+
+	return 0;
+}
+
+static struct clk_ops clk_tve_di_ops = {
+	.round_rate = clk_tve_di_round_rate,
+	.set_rate = clk_tve_di_set_rate,
+	.recalc_rate = clk_tve_di_recalc_rate,
+};
+
+static int tve_clk_init(struct imx_tve *tve, void __iomem *base)
+{
+	const char *tve_di_parent[1];
+	struct clk_init_data init = {
+		.name = "tve_di",
+		.ops = &clk_tve_di_ops,
+		.num_parents = 1,
+		.flags = 0,
+	};
+
+	tve_di_parent[0] = __clk_get_name(tve->clk);
+	init.parent_names = (const char **)&tve_di_parent;
+
+	tve->clk_hw_di.init = &init;
+	tve->di_clk = clk_register(tve->dev, &tve->clk_hw_di);
+	if (IS_ERR(tve->di_clk)) {
+		dev_err(tve->dev, "failed to register TVE output clock: %ld\n",
+			PTR_ERR(tve->di_clk));
+		return PTR_ERR(tve->di_clk);
+	}
+
+	return 0;
+}
+
+static int imx_tve_register(struct drm_device *drm, struct imx_tve *tve)
+{
+	int encoder_type;
+	int ret;
+
+	encoder_type = tve->mode == TVE_MODE_VGA ?
+				DRM_MODE_ENCODER_DAC : DRM_MODE_ENCODER_TVDAC;
+
+	ret = imx_drm_encoder_parse_of(drm, &tve->encoder,
+				       tve->dev->of_node);
+	if (ret)
+		return ret;
+
+	drm_encoder_helper_add(&tve->encoder, &imx_tve_encoder_helper_funcs);
+	drm_encoder_init(drm, &tve->encoder, &imx_tve_encoder_funcs,
+			 encoder_type);
+
+	drm_connector_helper_add(&tve->connector,
+			&imx_tve_connector_helper_funcs);
+	drm_connector_init(drm, &tve->connector, &imx_tve_connector_funcs,
+			   DRM_MODE_CONNECTOR_VGA);
+
+	drm_mode_connector_attach_encoder(&tve->connector, &tve->encoder);
+
+	return 0;
+}
+
+static bool imx_tve_readable_reg(struct device *dev, unsigned int reg)
+{
+	return (reg % 4 == 0) && (reg <= 0xdc);
+}
+
+static struct regmap_config tve_regmap_config = {
+	.reg_bits = 32,
+	.val_bits = 32,
+	.reg_stride = 4,
+
+	.readable_reg = imx_tve_readable_reg,
+
+	.lock = tve_lock,
+	.unlock = tve_unlock,
+
+	.max_register = 0xdc,
+};
+
+static const char * const imx_tve_modes[] = {
+	[TVE_MODE_TVOUT]  = "tvout",
+	[TVE_MODE_VGA] = "vga",
+};
+
+static const int of_get_tve_mode(struct device_node *np)
+{
+	const char *bm;
+	int ret, i;
+
+	ret = of_property_read_string(np, "fsl,tve-mode", &bm);
+	if (ret < 0)
+		return ret;
+
+	for (i = 0; i < ARRAY_SIZE(imx_tve_modes); i++)
+		if (!strcasecmp(bm, imx_tve_modes[i]))
+			return i;
+
+	return -EINVAL;
+}
+
+static int imx_tve_bind(struct device *dev, struct device *master, void *data)
+{
+	struct platform_device *pdev = to_platform_device(dev);
+	struct drm_device *drm = data;
+	struct device_node *np = dev->of_node;
+	struct device_node *ddc_node;
+	struct imx_tve *tve;
+	struct resource *res;
+	void __iomem *base;
+	unsigned int val;
+	int irq;
+	int ret;
+
+	tve = devm_kzalloc(dev, sizeof(*tve), GFP_KERNEL);
+	if (!tve)
+		return -ENOMEM;
+
+	tve->dev = dev;
+	spin_lock_init(&tve->lock);
+
+	ddc_node = of_parse_phandle(np, "ddc-i2c-bus", 0);
+	if (ddc_node) {
+		tve->ddc = of_find_i2c_adapter_by_node(ddc_node);
+		of_node_put(ddc_node);
+	}
+
+	tve->mode = of_get_tve_mode(np);
+	if (tve->mode != TVE_MODE_VGA) {
+		dev_err(dev, "only VGA mode supported, currently\n");
+		return -EINVAL;
+	}
+
+	if (tve->mode == TVE_MODE_VGA) {
+		ret = of_property_read_u32(np, "fsl,hsync-pin",
+					   &tve->hsync_pin);
+
+		if (ret < 0) {
+			dev_err(dev, "failed to get vsync pin\n");
+			return ret;
+		}
+
+		ret |= of_property_read_u32(np, "fsl,vsync-pin",
+					    &tve->vsync_pin);
+
+		if (ret < 0) {
+			dev_err(dev, "failed to get vsync pin\n");
+			return ret;
+		}
+	}
+
+	res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
+	base = devm_ioremap_resource(dev, res);
+	if (IS_ERR(base))
+		return PTR_ERR(base);
+
+	tve_regmap_config.lock_arg = tve;
+	tve->regmap = devm_regmap_init_mmio_clk(dev, "tve", base,
+						&tve_regmap_config);
+	if (IS_ERR(tve->regmap)) {
+		dev_err(dev, "failed to init regmap: %ld\n",
+			PTR_ERR(tve->regmap));
+		return PTR_ERR(tve->regmap);
+	}
+
+	irq = platform_get_irq(pdev, 0);
+	if (irq < 0) {
+		dev_err(dev, "failed to get irq\n");
+		return irq;
+	}
+
+	ret = devm_request_threaded_irq(dev, irq, NULL,
+					imx_tve_irq_handler, IRQF_ONESHOT,
+					"imx-tve", tve);
+	if (ret < 0) {
+		dev_err(dev, "failed to request irq: %d\n", ret);
+		return ret;
+	}
+
+	tve->dac_reg = devm_regulator_get(dev, "dac");
+	if (!IS_ERR(tve->dac_reg)) {
+		regulator_set_voltage(tve->dac_reg, 2750000, 2750000);
+		ret = regulator_enable(tve->dac_reg);
+		if (ret)
+			return ret;
+	}
+
+	tve->clk = devm_clk_get(dev, "tve");
+	if (IS_ERR(tve->clk)) {
+		dev_err(dev, "failed to get high speed tve clock: %ld\n",
+			PTR_ERR(tve->clk));
+		return PTR_ERR(tve->clk);
+	}
+
+	/* this is the IPU DI clock input selector, can be parented to tve_di */
+	tve->di_sel_clk = devm_clk_get(dev, "di_sel");
+	if (IS_ERR(tve->di_sel_clk)) {
+		dev_err(dev, "failed to get ipu di mux clock: %ld\n",
+			PTR_ERR(tve->di_sel_clk));
+		return PTR_ERR(tve->di_sel_clk);
+	}
+
+	ret = tve_clk_init(tve, base);
+	if (ret < 0)
+		return ret;
+
+	ret = regmap_read(tve->regmap, TVE_COM_CONF_REG, &val);
+	if (ret < 0) {
+		dev_err(dev, "failed to read configuration register: %d\n",
+			ret);
+		return ret;
+	}
+	if (val != 0x00100000) {
+		dev_err(dev, "configuration register default value indicates this is not a TVEv2\n");
+		return -ENODEV;
+	}
+
+	/* disable cable detection for VGA mode */
+	ret = regmap_write(tve->regmap, TVE_CD_CONT_REG, 0);
+
+	ret = imx_tve_register(drm, tve);
+	if (ret)
+		return ret;
+
+	dev_set_drvdata(dev, tve);
+
+	return 0;
+}
+
+static void imx_tve_unbind(struct device *dev, struct device *master,
+	void *data)
+{
+	struct imx_tve *tve = dev_get_drvdata(dev);
+
+	tve->connector.funcs->destroy(&tve->connector);
+	tve->encoder.funcs->destroy(&tve->encoder);
+
+	if (!IS_ERR(tve->dac_reg))
+		regulator_disable(tve->dac_reg);
+}
+
+static const struct component_ops imx_tve_ops = {
+	.bind	= imx_tve_bind,
+	.unbind	= imx_tve_unbind,
+};
+
+static int imx_tve_probe(struct platform_device *pdev)
+{
+	return component_add(&pdev->dev, &imx_tve_ops);
+}
+
+static int imx_tve_remove(struct platform_device *pdev)
+{
+	component_del(&pdev->dev, &imx_tve_ops);
+	return 0;
+}
+
+static const struct of_device_id imx_tve_dt_ids[] = {
+	{ .compatible = "fsl,imx53-tve", },
+	{ /* sentinel */ }
+};
+
+static struct platform_driver imx_tve_driver = {
+	.probe		= imx_tve_probe,
+	.remove		= imx_tve_remove,
+	.driver		= {
+		.of_match_table = imx_tve_dt_ids,
+		.name	= "imx-tve",
+	},
+};
+
+module_platform_driver(imx_tve_driver);
+
+MODULE_DESCRIPTION("i.MX Television Encoder driver");
+MODULE_AUTHOR("Philipp Zabel, Pengutronix");
+MODULE_LICENSE("GPL");
+MODULE_ALIAS("platform:imx-tve");
diff --git a/drivers/gpu/drm/imx/ipuv3-crtc.c b/drivers/gpu/drm/imx/ipuv3-crtc.c
new file mode 100644
index 000000000000..ebee59cb96d8
--- /dev/null
+++ b/drivers/gpu/drm/imx/ipuv3-crtc.c
@@ -0,0 +1,513 @@
+/*
+ * i.MX IPUv3 Graphics driver
+ *
+ * Copyright (C) 2011 Sascha Hauer, Pengutronix
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License
+ * as published by the Free Software Foundation; either version 2
+ * of the License, or (at your option) any later version.
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ */
+#include <linux/component.h>
+#include <linux/module.h>
+#include <linux/export.h>
+#include <linux/device.h>
+#include <linux/platform_device.h>
+#include <drm/drmP.h>
+#include <drm/drm_crtc_helper.h>
+#include <linux/fb.h>
+#include <linux/clk.h>
+#include <linux/errno.h>
+#include <drm/drm_gem_cma_helper.h>
+#include <drm/drm_fb_cma_helper.h>
+
+#include <video/imx-ipu-v3.h>
+#include "imx-drm.h"
+#include "ipuv3-plane.h"
+
+#define DRIVER_DESC		"i.MX IPUv3 Graphics"
+
+struct ipu_crtc {
+	struct device		*dev;
+	struct drm_crtc		base;
+	struct imx_drm_crtc	*imx_crtc;
+
+	/* plane[0] is the full plane, plane[1] is the partial plane */
+	struct ipu_plane	*plane[2];
+
+	struct ipu_dc		*dc;
+	struct ipu_di		*di;
+	int			enabled;
+	struct drm_pending_vblank_event *page_flip_event;
+	struct drm_framebuffer	*newfb;
+	int			irq;
+	u32			interface_pix_fmt;
+	unsigned long		di_clkflags;
+	int			di_hsync_pin;
+	int			di_vsync_pin;
+};
+
+#define to_ipu_crtc(x) container_of(x, struct ipu_crtc, base)
+
+static void ipu_fb_enable(struct ipu_crtc *ipu_crtc)
+{
+	struct ipu_soc *ipu = dev_get_drvdata(ipu_crtc->dev->parent);
+
+	if (ipu_crtc->enabled)
+		return;
+
+	ipu_dc_enable(ipu);
+	ipu_plane_enable(ipu_crtc->plane[0]);
+	/* Start DC channel and DI after IDMAC */
+	ipu_dc_enable_channel(ipu_crtc->dc);
+	ipu_di_enable(ipu_crtc->di);
+
+	ipu_crtc->enabled = 1;
+}
+
+static void ipu_fb_disable(struct ipu_crtc *ipu_crtc)
+{
+	struct ipu_soc *ipu = dev_get_drvdata(ipu_crtc->dev->parent);
+
+	if (!ipu_crtc->enabled)
+		return;
+
+	/* Stop DC channel and DI before IDMAC */
+	ipu_dc_disable_channel(ipu_crtc->dc);
+	ipu_di_disable(ipu_crtc->di);
+	ipu_plane_disable(ipu_crtc->plane[0]);
+	ipu_dc_disable(ipu);
+
+	ipu_crtc->enabled = 0;
+}
+
+static void ipu_crtc_dpms(struct drm_crtc *crtc, int mode)
+{
+	struct ipu_crtc *ipu_crtc = to_ipu_crtc(crtc);
+
+	dev_dbg(ipu_crtc->dev, "%s mode: %d\n", __func__, mode);
+
+	switch (mode) {
+	case DRM_MODE_DPMS_ON:
+		ipu_fb_enable(ipu_crtc);
+		break;
+	case DRM_MODE_DPMS_STANDBY:
+	case DRM_MODE_DPMS_SUSPEND:
+	case DRM_MODE_DPMS_OFF:
+		ipu_fb_disable(ipu_crtc);
+		break;
+	}
+}
+
+static int ipu_page_flip(struct drm_crtc *crtc,
+		struct drm_framebuffer *fb,
+		struct drm_pending_vblank_event *event,
+		uint32_t page_flip_flags)
+{
+	struct ipu_crtc *ipu_crtc = to_ipu_crtc(crtc);
+	int ret;
+
+	if (ipu_crtc->newfb)
+		return -EBUSY;
+
+	ret = imx_drm_crtc_vblank_get(ipu_crtc->imx_crtc);
+	if (ret) {
+		dev_dbg(ipu_crtc->dev, "failed to acquire vblank counter\n");
+		list_del(&event->base.link);
+
+		return ret;
+	}
+
+	ipu_crtc->newfb = fb;
+	ipu_crtc->page_flip_event = event;
+	crtc->primary->fb = fb;
+
+	return 0;
+}
+
+static const struct drm_crtc_funcs ipu_crtc_funcs = {
+	.set_config = drm_crtc_helper_set_config,
+	.destroy = drm_crtc_cleanup,
+	.page_flip = ipu_page_flip,
+};
+
+static int ipu_crtc_mode_set(struct drm_crtc *crtc,
+			       struct drm_display_mode *orig_mode,
+			       struct drm_display_mode *mode,
+			       int x, int y,
+			       struct drm_framebuffer *old_fb)
+{
+	struct ipu_crtc *ipu_crtc = to_ipu_crtc(crtc);
+	int ret;
+	struct ipu_di_signal_cfg sig_cfg = {};
+	u32 out_pixel_fmt;
+
+	dev_dbg(ipu_crtc->dev, "%s: mode->hdisplay: %d\n", __func__,
+			mode->hdisplay);
+	dev_dbg(ipu_crtc->dev, "%s: mode->vdisplay: %d\n", __func__,
+			mode->vdisplay);
+
+	out_pixel_fmt = ipu_crtc->interface_pix_fmt;
+
+	if (mode->flags & DRM_MODE_FLAG_INTERLACE)
+		sig_cfg.interlaced = 1;
+	if (mode->flags & DRM_MODE_FLAG_PHSYNC)
+		sig_cfg.Hsync_pol = 1;
+	if (mode->flags & DRM_MODE_FLAG_PVSYNC)
+		sig_cfg.Vsync_pol = 1;
+
+	sig_cfg.enable_pol = 1;
+	sig_cfg.clk_pol = 0;
+	sig_cfg.width = mode->hdisplay;
+	sig_cfg.height = mode->vdisplay;
+	sig_cfg.pixel_fmt = out_pixel_fmt;
+	sig_cfg.h_start_width = mode->htotal - mode->hsync_end;
+	sig_cfg.h_sync_width = mode->hsync_end - mode->hsync_start;
+	sig_cfg.h_end_width = mode->hsync_start - mode->hdisplay;
+
+	sig_cfg.v_start_width = mode->vtotal - mode->vsync_end;
+	sig_cfg.v_sync_width = mode->vsync_end - mode->vsync_start;
+	sig_cfg.v_end_width = mode->vsync_start - mode->vdisplay;
+	sig_cfg.pixelclock = mode->clock * 1000;
+	sig_cfg.clkflags = ipu_crtc->di_clkflags;
+
+	sig_cfg.v_to_h_sync = 0;
+
+	sig_cfg.hsync_pin = ipu_crtc->di_hsync_pin;
+	sig_cfg.vsync_pin = ipu_crtc->di_vsync_pin;
+
+	ret = ipu_dc_init_sync(ipu_crtc->dc, ipu_crtc->di, sig_cfg.interlaced,
+			out_pixel_fmt, mode->hdisplay);
+	if (ret) {
+		dev_err(ipu_crtc->dev,
+				"initializing display controller failed with %d\n",
+				ret);
+		return ret;
+	}
+
+	ret = ipu_di_init_sync_panel(ipu_crtc->di, &sig_cfg);
+	if (ret) {
+		dev_err(ipu_crtc->dev,
+				"initializing panel failed with %d\n", ret);
+		return ret;
+	}
+
+	return ipu_plane_mode_set(ipu_crtc->plane[0], crtc, mode,
+				  crtc->primary->fb,
+				  0, 0, mode->hdisplay, mode->vdisplay,
+				  x, y, mode->hdisplay, mode->vdisplay);
+}
+
+static void ipu_crtc_handle_pageflip(struct ipu_crtc *ipu_crtc)
+{
+	unsigned long flags;
+	struct drm_device *drm = ipu_crtc->base.dev;
+
+	spin_lock_irqsave(&drm->event_lock, flags);
+	if (ipu_crtc->page_flip_event)
+		drm_send_vblank_event(drm, -1, ipu_crtc->page_flip_event);
+	ipu_crtc->page_flip_event = NULL;
+	imx_drm_crtc_vblank_put(ipu_crtc->imx_crtc);
+	spin_unlock_irqrestore(&drm->event_lock, flags);
+}
+
+static irqreturn_t ipu_irq_handler(int irq, void *dev_id)
+{
+	struct ipu_crtc *ipu_crtc = dev_id;
+
+	imx_drm_handle_vblank(ipu_crtc->imx_crtc);
+
+	if (ipu_crtc->newfb) {
+		struct ipu_plane *plane = ipu_crtc->plane[0];
+
+		ipu_crtc->newfb = NULL;
+		ipu_plane_set_base(plane, ipu_crtc->base.primary->fb,
+				   plane->x, plane->y);
+		ipu_crtc_handle_pageflip(ipu_crtc);
+	}
+
+	return IRQ_HANDLED;
+}
+
+static bool ipu_crtc_mode_fixup(struct drm_crtc *crtc,
+				  const struct drm_display_mode *mode,
+				  struct drm_display_mode *adjusted_mode)
+{
+	return true;
+}
+
+static void ipu_crtc_prepare(struct drm_crtc *crtc)
+{
+	struct ipu_crtc *ipu_crtc = to_ipu_crtc(crtc);
+
+	ipu_fb_disable(ipu_crtc);
+}
+
+static void ipu_crtc_commit(struct drm_crtc *crtc)
+{
+	struct ipu_crtc *ipu_crtc = to_ipu_crtc(crtc);
+
+	ipu_fb_enable(ipu_crtc);
+}
+
+static struct drm_crtc_helper_funcs ipu_helper_funcs = {
+	.dpms = ipu_crtc_dpms,
+	.mode_fixup = ipu_crtc_mode_fixup,
+	.mode_set = ipu_crtc_mode_set,
+	.prepare = ipu_crtc_prepare,
+	.commit = ipu_crtc_commit,
+};
+
+static int ipu_enable_vblank(struct drm_crtc *crtc)
+{
+	return 0;
+}
+
+static void ipu_disable_vblank(struct drm_crtc *crtc)
+{
+	struct ipu_crtc *ipu_crtc = to_ipu_crtc(crtc);
+
+	ipu_crtc->page_flip_event = NULL;
+	ipu_crtc->newfb = NULL;
+}
+
+static int ipu_set_interface_pix_fmt(struct drm_crtc *crtc, u32 encoder_type,
+		u32 pixfmt, int hsync_pin, int vsync_pin)
+{
+	struct ipu_crtc *ipu_crtc = to_ipu_crtc(crtc);
+
+	ipu_crtc->interface_pix_fmt = pixfmt;
+	ipu_crtc->di_hsync_pin = hsync_pin;
+	ipu_crtc->di_vsync_pin = vsync_pin;
+
+	switch (encoder_type) {
+	case DRM_MODE_ENCODER_DAC:
+	case DRM_MODE_ENCODER_TVDAC:
+	case DRM_MODE_ENCODER_LVDS:
+		ipu_crtc->di_clkflags = IPU_DI_CLKMODE_SYNC |
+			IPU_DI_CLKMODE_EXT;
+		break;
+	case DRM_MODE_ENCODER_TMDS:
+	case DRM_MODE_ENCODER_NONE:
+		ipu_crtc->di_clkflags = 0;
+		break;
+	}
+
+	return 0;
+}
+
+static const struct imx_drm_crtc_helper_funcs ipu_crtc_helper_funcs = {
+	.enable_vblank = ipu_enable_vblank,
+	.disable_vblank = ipu_disable_vblank,
+	.set_interface_pix_fmt = ipu_set_interface_pix_fmt,
+	.crtc_funcs = &ipu_crtc_funcs,
+	.crtc_helper_funcs = &ipu_helper_funcs,
+};
+
+static void ipu_put_resources(struct ipu_crtc *ipu_crtc)
+{
+	if (!IS_ERR_OR_NULL(ipu_crtc->dc))
+		ipu_dc_put(ipu_crtc->dc);
+	if (!IS_ERR_OR_NULL(ipu_crtc->di))
+		ipu_di_put(ipu_crtc->di);
+}
+
+static int ipu_get_resources(struct ipu_crtc *ipu_crtc,
+		struct ipu_client_platformdata *pdata)
+{
+	struct ipu_soc *ipu = dev_get_drvdata(ipu_crtc->dev->parent);
+	int ret;
+
+	ipu_crtc->dc = ipu_dc_get(ipu, pdata->dc);
+	if (IS_ERR(ipu_crtc->dc)) {
+		ret = PTR_ERR(ipu_crtc->dc);
+		goto err_out;
+	}
+
+	ipu_crtc->di = ipu_di_get(ipu, pdata->di);
+	if (IS_ERR(ipu_crtc->di)) {
+		ret = PTR_ERR(ipu_crtc->di);
+		goto err_out;
+	}
+
+	return 0;
+err_out:
+	ipu_put_resources(ipu_crtc);
+
+	return ret;
+}
+
+static int ipu_crtc_init(struct ipu_crtc *ipu_crtc,
+	struct ipu_client_platformdata *pdata, struct drm_device *drm)
+{
+	struct ipu_soc *ipu = dev_get_drvdata(ipu_crtc->dev->parent);
+	int dp = -EINVAL;
+	int ret;
+	int id;
+
+	ret = ipu_get_resources(ipu_crtc, pdata);
+	if (ret) {
+		dev_err(ipu_crtc->dev, "getting resources failed with %d.\n",
+				ret);
+		return ret;
+	}
+
+	ret = imx_drm_add_crtc(drm, &ipu_crtc->base, &ipu_crtc->imx_crtc,
+			&ipu_crtc_helper_funcs, ipu_crtc->dev->of_node);
+	if (ret) {
+		dev_err(ipu_crtc->dev, "adding crtc failed with %d.\n", ret);
+		goto err_put_resources;
+	}
+
+	if (pdata->dp >= 0)
+		dp = IPU_DP_FLOW_SYNC_BG;
+	id = imx_drm_crtc_id(ipu_crtc->imx_crtc);
+	ipu_crtc->plane[0] = ipu_plane_init(ipu_crtc->base.dev, ipu,
+					    pdata->dma[0], dp, BIT(id), true);
+	ret = ipu_plane_get_resources(ipu_crtc->plane[0]);
+	if (ret) {
+		dev_err(ipu_crtc->dev, "getting plane 0 resources failed with %d.\n",
+			ret);
+		goto err_remove_crtc;
+	}
+
+	/* If this crtc is using the DP, add an overlay plane */
+	if (pdata->dp >= 0 && pdata->dma[1] > 0) {
+		ipu_crtc->plane[1] = ipu_plane_init(ipu_crtc->base.dev, ipu,
+						    pdata->dma[1],
+						    IPU_DP_FLOW_SYNC_FG,
+						    BIT(id), false);
+		if (IS_ERR(ipu_crtc->plane[1]))
+			ipu_crtc->plane[1] = NULL;
+	}
+
+	ipu_crtc->irq = ipu_plane_irq(ipu_crtc->plane[0]);
+	ret = devm_request_irq(ipu_crtc->dev, ipu_crtc->irq, ipu_irq_handler, 0,
+			"imx_drm", ipu_crtc);
+	if (ret < 0) {
+		dev_err(ipu_crtc->dev, "irq request failed with %d.\n", ret);
+		goto err_put_plane_res;
+	}
+
+	return 0;
+
+err_put_plane_res:
+	ipu_plane_put_resources(ipu_crtc->plane[0]);
+err_remove_crtc:
+	imx_drm_remove_crtc(ipu_crtc->imx_crtc);
+err_put_resources:
+	ipu_put_resources(ipu_crtc);
+
+	return ret;
+}
+
+static struct device_node *ipu_drm_get_port_by_id(struct device_node *parent,
+						  int port_id)
+{
+	struct device_node *port;
+	int id, ret;
+
+	port = of_get_child_by_name(parent, "port");
+	while (port) {
+		ret = of_property_read_u32(port, "reg", &id);
+		if (!ret && id == port_id)
+			return port;
+
+		do {
+			port = of_get_next_child(parent, port);
+			if (!port)
+				return NULL;
+		} while (of_node_cmp(port->name, "port"));
+	}
+
+	return NULL;
+}
+
+static int ipu_drm_bind(struct device *dev, struct device *master, void *data)
+{
+	struct ipu_client_platformdata *pdata = dev->platform_data;
+	struct drm_device *drm = data;
+	struct ipu_crtc *ipu_crtc;
+	int ret;
+
+	ipu_crtc = devm_kzalloc(dev, sizeof(*ipu_crtc), GFP_KERNEL);
+	if (!ipu_crtc)
+		return -ENOMEM;
+
+	ipu_crtc->dev = dev;
+
+	ret = ipu_crtc_init(ipu_crtc, pdata, drm);
+	if (ret)
+		return ret;
+
+	dev_set_drvdata(dev, ipu_crtc);
+
+	return 0;
+}
+
+static void ipu_drm_unbind(struct device *dev, struct device *master,
+	void *data)
+{
+	struct ipu_crtc *ipu_crtc = dev_get_drvdata(dev);
+
+	imx_drm_remove_crtc(ipu_crtc->imx_crtc);
+
+	ipu_plane_put_resources(ipu_crtc->plane[0]);
+	ipu_put_resources(ipu_crtc);
+}
+
+static const struct component_ops ipu_crtc_ops = {
+	.bind = ipu_drm_bind,
+	.unbind = ipu_drm_unbind,
+};
+
+static int ipu_drm_probe(struct platform_device *pdev)
+{
+	struct device *dev = &pdev->dev;
+	struct ipu_client_platformdata *pdata = dev->platform_data;
+	int ret;
+
+	if (!dev->platform_data)
+		return -EINVAL;
+
+	if (!dev->of_node) {
+		/* Associate crtc device with the corresponding DI port node */
+		dev->of_node = ipu_drm_get_port_by_id(dev->parent->of_node,
+						      pdata->di + 2);
+		if (!dev->of_node) {
+			dev_err(dev, "missing port@%d node in %s\n",
+				pdata->di + 2, dev->parent->of_node->full_name);
+			return -ENODEV;
+		}
+	}
+
+	ret = dma_set_coherent_mask(dev, DMA_BIT_MASK(32));
+	if (ret)
+		return ret;
+
+	return component_add(dev, &ipu_crtc_ops);
+}
+
+static int ipu_drm_remove(struct platform_device *pdev)
+{
+	component_del(&pdev->dev, &ipu_crtc_ops);
+	return 0;
+}
+
+static struct platform_driver ipu_drm_driver = {
+	.driver = {
+		.name = "imx-ipuv3-crtc",
+	},
+	.probe = ipu_drm_probe,
+	.remove = ipu_drm_remove,
+};
+module_platform_driver(ipu_drm_driver);
+
+MODULE_AUTHOR("Sascha Hauer <s.hauer@pengutronix.de>");
+MODULE_DESCRIPTION(DRIVER_DESC);
+MODULE_LICENSE("GPL");
+MODULE_ALIAS("platform:imx-ipuv3-crtc");
diff --git a/drivers/gpu/drm/imx/ipuv3-plane.c b/drivers/gpu/drm/imx/ipuv3-plane.c
new file mode 100644
index 000000000000..6987e16fe99b
--- /dev/null
+++ b/drivers/gpu/drm/imx/ipuv3-plane.c
@@ -0,0 +1,394 @@
+/*
+ * i.MX IPUv3 DP Overlay Planes
+ *
+ * Copyright (C) 2013 Philipp Zabel, Pengutronix
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License
+ * as published by the Free Software Foundation; either version 2
+ * of the License, or (at your option) any later version.
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ */
+
+#include <drm/drmP.h>
+#include <drm/drm_fb_cma_helper.h>
+#include <drm/drm_gem_cma_helper.h>
+
+#include "video/imx-ipu-v3.h"
+#include "ipuv3-plane.h"
+
+#define to_ipu_plane(x)	container_of(x, struct ipu_plane, base)
+
+static const uint32_t ipu_plane_formats[] = {
+	DRM_FORMAT_XRGB1555,
+	DRM_FORMAT_XBGR1555,
+	DRM_FORMAT_ARGB8888,
+	DRM_FORMAT_XRGB8888,
+	DRM_FORMAT_ABGR8888,
+	DRM_FORMAT_XBGR8888,
+	DRM_FORMAT_YUYV,
+	DRM_FORMAT_YVYU,
+	DRM_FORMAT_YUV420,
+	DRM_FORMAT_YVU420,
+};
+
+int ipu_plane_irq(struct ipu_plane *ipu_plane)
+{
+	return ipu_idmac_channel_irq(ipu_plane->ipu, ipu_plane->ipu_ch,
+				     IPU_IRQ_EOF);
+}
+
+static int calc_vref(struct drm_display_mode *mode)
+{
+	unsigned long htotal, vtotal;
+
+	htotal = mode->htotal;
+	vtotal = mode->vtotal;
+
+	if (!htotal || !vtotal)
+		return 60;
+
+	return DIV_ROUND_UP(mode->clock * 1000, vtotal * htotal);
+}
+
+static inline int calc_bandwidth(int width, int height, unsigned int vref)
+{
+	return width * height * vref;
+}
+
+int ipu_plane_set_base(struct ipu_plane *ipu_plane, struct drm_framebuffer *fb,
+		       int x, int y)
+{
+	struct drm_gem_cma_object *cma_obj;
+	unsigned long eba;
+	int active;
+
+	cma_obj = drm_fb_cma_get_gem_obj(fb, 0);
+	if (!cma_obj) {
+		DRM_DEBUG_KMS("entry is null.\n");
+		return -EFAULT;
+	}
+
+	dev_dbg(ipu_plane->base.dev->dev, "phys = %pad, x = %d, y = %d",
+		&cma_obj->paddr, x, y);
+
+	eba = cma_obj->paddr + fb->offsets[0] +
+	      fb->pitches[0] * y + (fb->bits_per_pixel >> 3) * x;
+
+	if (ipu_plane->enabled) {
+		active = ipu_idmac_get_current_buffer(ipu_plane->ipu_ch);
+		ipu_cpmem_set_buffer(ipu_plane->ipu_ch, !active, eba);
+		ipu_idmac_select_buffer(ipu_plane->ipu_ch, !active);
+	} else {
+		ipu_cpmem_set_buffer(ipu_plane->ipu_ch, 0, eba);
+		ipu_cpmem_set_buffer(ipu_plane->ipu_ch, 1, eba);
+	}
+
+	/* cache offsets for subsequent pageflips */
+	ipu_plane->x = x;
+	ipu_plane->y = y;
+
+	return 0;
+}
+
+int ipu_plane_mode_set(struct ipu_plane *ipu_plane, struct drm_crtc *crtc,
+		       struct drm_display_mode *mode,
+		       struct drm_framebuffer *fb, int crtc_x, int crtc_y,
+		       unsigned int crtc_w, unsigned int crtc_h,
+		       uint32_t src_x, uint32_t src_y,
+		       uint32_t src_w, uint32_t src_h)
+{
+	struct device *dev = ipu_plane->base.dev->dev;
+	int ret;
+
+	/* no scaling */
+	if (src_w != crtc_w || src_h != crtc_h)
+		return -EINVAL;
+
+	/* clip to crtc bounds */
+	if (crtc_x < 0) {
+		if (-crtc_x > crtc_w)
+			return -EINVAL;
+		src_x += -crtc_x;
+		src_w -= -crtc_x;
+		crtc_w -= -crtc_x;
+		crtc_x = 0;
+	}
+	if (crtc_y < 0) {
+		if (-crtc_y > crtc_h)
+			return -EINVAL;
+		src_y += -crtc_y;
+		src_h -= -crtc_y;
+		crtc_h -= -crtc_y;
+		crtc_y = 0;
+	}
+	if (crtc_x + crtc_w > mode->hdisplay) {
+		if (crtc_x > mode->hdisplay)
+			return -EINVAL;
+		crtc_w = mode->hdisplay - crtc_x;
+		src_w = crtc_w;
+	}
+	if (crtc_y + crtc_h > mode->vdisplay) {
+		if (crtc_y > mode->vdisplay)
+			return -EINVAL;
+		crtc_h = mode->vdisplay - crtc_y;
+		src_h = crtc_h;
+	}
+	/* full plane minimum width is 13 pixels */
+	if (crtc_w < 13 && (ipu_plane->dp_flow != IPU_DP_FLOW_SYNC_FG))
+		return -EINVAL;
+	if (crtc_h < 2)
+		return -EINVAL;
+
+	/*
+	 * since we cannot touch active IDMAC channels, we do not support
+	 * resizing the enabled plane or changing its format
+	 */
+	if (ipu_plane->enabled) {
+		if (src_w != ipu_plane->w || src_h != ipu_plane->h ||
+		    fb->pixel_format != ipu_plane->base.fb->pixel_format)
+			return -EINVAL;
+
+		return ipu_plane_set_base(ipu_plane, fb, src_x, src_y);
+	}
+
+	switch (ipu_plane->dp_flow) {
+	case IPU_DP_FLOW_SYNC_BG:
+		ret = ipu_dp_setup_channel(ipu_plane->dp,
+				IPUV3_COLORSPACE_RGB,
+				IPUV3_COLORSPACE_RGB);
+		if (ret) {
+			dev_err(dev,
+				"initializing display processor failed with %d\n",
+				ret);
+			return ret;
+		}
+		ipu_dp_set_global_alpha(ipu_plane->dp, true, 0, true);
+		break;
+	case IPU_DP_FLOW_SYNC_FG:
+		ipu_dp_setup_channel(ipu_plane->dp,
+				ipu_drm_fourcc_to_colorspace(fb->pixel_format),
+				IPUV3_COLORSPACE_UNKNOWN);
+		ipu_dp_set_window_pos(ipu_plane->dp, crtc_x, crtc_y);
+		/* Enable local alpha on partial plane */
+		switch (fb->pixel_format) {
+		case DRM_FORMAT_ARGB8888:
+		case DRM_FORMAT_ABGR8888:
+			ipu_dp_set_global_alpha(ipu_plane->dp, false, 0, false);
+			break;
+		default:
+			break;
+		}
+	}
+
+	ret = ipu_dmfc_init_channel(ipu_plane->dmfc, crtc_w);
+	if (ret) {
+		dev_err(dev, "initializing dmfc channel failed with %d\n", ret);
+		return ret;
+	}
+
+	ret = ipu_dmfc_alloc_bandwidth(ipu_plane->dmfc,
+			calc_bandwidth(crtc_w, crtc_h,
+				       calc_vref(mode)), 64);
+	if (ret) {
+		dev_err(dev, "allocating dmfc bandwidth failed with %d\n", ret);
+		return ret;
+	}
+
+	ipu_cpmem_zero(ipu_plane->ipu_ch);
+	ipu_cpmem_set_resolution(ipu_plane->ipu_ch, src_w, src_h);
+	ret = ipu_cpmem_set_fmt(ipu_plane->ipu_ch, fb->pixel_format);
+	if (ret < 0) {
+		dev_err(dev, "unsupported pixel format 0x%08x\n",
+			fb->pixel_format);
+		return ret;
+	}
+	ipu_cpmem_set_high_priority(ipu_plane->ipu_ch);
+	ipu_idmac_set_double_buffer(ipu_plane->ipu_ch, 1);
+	ipu_cpmem_set_stride(ipu_plane->ipu_ch, fb->pitches[0]);
+
+	ret = ipu_plane_set_base(ipu_plane, fb, src_x, src_y);
+	if (ret < 0)
+		return ret;
+
+	ipu_plane->w = src_w;
+	ipu_plane->h = src_h;
+
+	return 0;
+}
+
+void ipu_plane_put_resources(struct ipu_plane *ipu_plane)
+{
+	if (!IS_ERR_OR_NULL(ipu_plane->dp))
+		ipu_dp_put(ipu_plane->dp);
+	if (!IS_ERR_OR_NULL(ipu_plane->dmfc))
+		ipu_dmfc_put(ipu_plane->dmfc);
+	if (!IS_ERR_OR_NULL(ipu_plane->ipu_ch))
+		ipu_idmac_put(ipu_plane->ipu_ch);
+}
+
+int ipu_plane_get_resources(struct ipu_plane *ipu_plane)
+{
+	int ret;
+
+	ipu_plane->ipu_ch = ipu_idmac_get(ipu_plane->ipu, ipu_plane->dma);
+	if (IS_ERR(ipu_plane->ipu_ch)) {
+		ret = PTR_ERR(ipu_plane->ipu_ch);
+		DRM_ERROR("failed to get idmac channel: %d\n", ret);
+		return ret;
+	}
+
+	ipu_plane->dmfc = ipu_dmfc_get(ipu_plane->ipu, ipu_plane->dma);
+	if (IS_ERR(ipu_plane->dmfc)) {
+		ret = PTR_ERR(ipu_plane->dmfc);
+		DRM_ERROR("failed to get dmfc: ret %d\n", ret);
+		goto err_out;
+	}
+
+	if (ipu_plane->dp_flow >= 0) {
+		ipu_plane->dp = ipu_dp_get(ipu_plane->ipu, ipu_plane->dp_flow);
+		if (IS_ERR(ipu_plane->dp)) {
+			ret = PTR_ERR(ipu_plane->dp);
+			DRM_ERROR("failed to get dp flow: %d\n", ret);
+			goto err_out;
+		}
+	}
+
+	return 0;
+err_out:
+	ipu_plane_put_resources(ipu_plane);
+
+	return ret;
+}
+
+void ipu_plane_enable(struct ipu_plane *ipu_plane)
+{
+	if (ipu_plane->dp)
+		ipu_dp_enable(ipu_plane->ipu);
+	ipu_dmfc_enable_channel(ipu_plane->dmfc);
+	ipu_idmac_enable_channel(ipu_plane->ipu_ch);
+	if (ipu_plane->dp)
+		ipu_dp_enable_channel(ipu_plane->dp);
+
+	ipu_plane->enabled = true;
+}
+
+void ipu_plane_disable(struct ipu_plane *ipu_plane)
+{
+	ipu_plane->enabled = false;
+
+	ipu_idmac_wait_busy(ipu_plane->ipu_ch, 50);
+
+	if (ipu_plane->dp)
+		ipu_dp_disable_channel(ipu_plane->dp);
+	ipu_idmac_disable_channel(ipu_plane->ipu_ch);
+	ipu_dmfc_disable_channel(ipu_plane->dmfc);
+	if (ipu_plane->dp)
+		ipu_dp_disable(ipu_plane->ipu);
+}
+
+/*
+ * drm_plane API
+ */
+
+static int ipu_update_plane(struct drm_plane *plane, struct drm_crtc *crtc,
+			    struct drm_framebuffer *fb, int crtc_x, int crtc_y,
+			    unsigned int crtc_w, unsigned int crtc_h,
+			    uint32_t src_x, uint32_t src_y,
+			    uint32_t src_w, uint32_t src_h)
+{
+	struct ipu_plane *ipu_plane = to_ipu_plane(plane);
+	int ret = 0;
+
+	DRM_DEBUG_KMS("plane - %p\n", plane);
+
+	if (!ipu_plane->enabled)
+		ret = ipu_plane_get_resources(ipu_plane);
+	if (ret < 0)
+		return ret;
+
+	ret = ipu_plane_mode_set(ipu_plane, crtc, &crtc->hwmode, fb,
+			crtc_x, crtc_y, crtc_w, crtc_h,
+			src_x >> 16, src_y >> 16, src_w >> 16, src_h >> 16);
+	if (ret < 0) {
+		ipu_plane_put_resources(ipu_plane);
+		return ret;
+	}
+
+	if (crtc != plane->crtc)
+		dev_info(plane->dev->dev, "crtc change: %p -> %p\n",
+				plane->crtc, crtc);
+	plane->crtc = crtc;
+
+	if (!ipu_plane->enabled)
+		ipu_plane_enable(ipu_plane);
+
+	return 0;
+}
+
+static int ipu_disable_plane(struct drm_plane *plane)
+{
+	struct ipu_plane *ipu_plane = to_ipu_plane(plane);
+
+	DRM_DEBUG_KMS("[%d] %s\n", __LINE__, __func__);
+
+	if (ipu_plane->enabled)
+		ipu_plane_disable(ipu_plane);
+
+	ipu_plane_put_resources(ipu_plane);
+
+	return 0;
+}
+
+static void ipu_plane_destroy(struct drm_plane *plane)
+{
+	struct ipu_plane *ipu_plane = to_ipu_plane(plane);
+
+	DRM_DEBUG_KMS("[%d] %s\n", __LINE__, __func__);
+
+	ipu_disable_plane(plane);
+	drm_plane_cleanup(plane);
+	kfree(ipu_plane);
+}
+
+static struct drm_plane_funcs ipu_plane_funcs = {
+	.update_plane	= ipu_update_plane,
+	.disable_plane	= ipu_disable_plane,
+	.destroy	= ipu_plane_destroy,
+};
+
+struct ipu_plane *ipu_plane_init(struct drm_device *dev, struct ipu_soc *ipu,
+				 int dma, int dp, unsigned int possible_crtcs,
+				 bool priv)
+{
+	struct ipu_plane *ipu_plane;
+	int ret;
+
+	DRM_DEBUG_KMS("channel %d, dp flow %d, possible_crtcs=0x%x\n",
+		      dma, dp, possible_crtcs);
+
+	ipu_plane = kzalloc(sizeof(*ipu_plane), GFP_KERNEL);
+	if (!ipu_plane) {
+		DRM_ERROR("failed to allocate plane\n");
+		return ERR_PTR(-ENOMEM);
+	}
+
+	ipu_plane->ipu = ipu;
+	ipu_plane->dma = dma;
+	ipu_plane->dp_flow = dp;
+
+	ret = drm_plane_init(dev, &ipu_plane->base, possible_crtcs,
+			     &ipu_plane_funcs, ipu_plane_formats,
+			     ARRAY_SIZE(ipu_plane_formats),
+			     priv);
+	if (ret) {
+		DRM_ERROR("failed to initialize plane\n");
+		kfree(ipu_plane);
+		return ERR_PTR(ret);
+	}
+
+	return ipu_plane;
+}
diff --git a/drivers/gpu/drm/imx/ipuv3-plane.h b/drivers/gpu/drm/imx/ipuv3-plane.h
new file mode 100644
index 000000000000..af125fb40ef5
--- /dev/null
+++ b/drivers/gpu/drm/imx/ipuv3-plane.h
@@ -0,0 +1,57 @@
+#ifndef __IPUV3_PLANE_H__
+#define __IPUV3_PLANE_H__
+
+#include <drm/drm_crtc.h> /* drm_plane */
+
+struct drm_plane;
+struct drm_device;
+struct ipu_soc;
+struct drm_crtc;
+struct drm_framebuffer;
+
+struct ipuv3_channel;
+struct dmfc_channel;
+struct ipu_dp;
+
+struct ipu_plane {
+	struct drm_plane	base;
+
+	struct ipu_soc		*ipu;
+	struct ipuv3_channel	*ipu_ch;
+	struct dmfc_channel	*dmfc;
+	struct ipu_dp		*dp;
+
+	int			dma;
+	int			dp_flow;
+
+	int			x;
+	int			y;
+	int			w;
+	int			h;
+
+	bool			enabled;
+};
+
+struct ipu_plane *ipu_plane_init(struct drm_device *dev, struct ipu_soc *ipu,
+				 int dma, int dp, unsigned int possible_crtcs,
+				 bool priv);
+
+/* Init IDMAC, DMFC, DP */
+int ipu_plane_mode_set(struct ipu_plane *plane, struct drm_crtc *crtc,
+		       struct drm_display_mode *mode,
+		       struct drm_framebuffer *fb, int crtc_x, int crtc_y,
+		       unsigned int crtc_w, unsigned int crtc_h,
+		       uint32_t src_x, uint32_t src_y, uint32_t src_w,
+		       uint32_t src_h);
+
+void ipu_plane_enable(struct ipu_plane *plane);
+void ipu_plane_disable(struct ipu_plane *plane);
+int ipu_plane_set_base(struct ipu_plane *plane, struct drm_framebuffer *fb,
+		       int x, int y);
+
+int ipu_plane_get_resources(struct ipu_plane *plane);
+void ipu_plane_put_resources(struct ipu_plane *plane);
+
+int ipu_plane_irq(struct ipu_plane *plane);
+
+#endif
diff --git a/drivers/gpu/drm/imx/parallel-display.c b/drivers/gpu/drm/imx/parallel-display.c
new file mode 100644
index 000000000000..796c3c1c170a
--- /dev/null
+++ b/drivers/gpu/drm/imx/parallel-display.c
@@ -0,0 +1,298 @@
+/*
+ * i.MX drm driver - parallel display implementation
+ *
+ * Copyright (C) 2012 Sascha Hauer, Pengutronix
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License
+ * as published by the Free Software Foundation; either version 2
+ * of the License, or (at your option) any later version.
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ */
+
+#include <linux/component.h>
+#include <linux/module.h>
+#include <drm/drmP.h>
+#include <drm/drm_fb_helper.h>
+#include <drm/drm_crtc_helper.h>
+#include <drm/drm_panel.h>
+#include <linux/videodev2.h>
+#include <video/of_display_timing.h>
+
+#include "imx-drm.h"
+
+#define con_to_imxpd(x) container_of(x, struct imx_parallel_display, connector)
+#define enc_to_imxpd(x) container_of(x, struct imx_parallel_display, encoder)
+
+struct imx_parallel_display {
+	struct drm_connector connector;
+	struct drm_encoder encoder;
+	struct device *dev;
+	void *edid;
+	int edid_len;
+	u32 interface_pix_fmt;
+	int mode_valid;
+	struct drm_display_mode mode;
+	struct drm_panel *panel;
+};
+
+static enum drm_connector_status imx_pd_connector_detect(
+		struct drm_connector *connector, bool force)
+{
+	return connector_status_connected;
+}
+
+static int imx_pd_connector_get_modes(struct drm_connector *connector)
+{
+	struct imx_parallel_display *imxpd = con_to_imxpd(connector);
+	struct device_node *np = imxpd->dev->of_node;
+	int num_modes = 0;
+
+	if (imxpd->panel && imxpd->panel->funcs &&
+	    imxpd->panel->funcs->get_modes) {
+		num_modes = imxpd->panel->funcs->get_modes(imxpd->panel);
+		if (num_modes > 0)
+			return num_modes;
+	}
+
+	if (imxpd->edid) {
+		drm_mode_connector_update_edid_property(connector, imxpd->edid);
+		num_modes = drm_add_edid_modes(connector, imxpd->edid);
+	}
+
+	if (imxpd->mode_valid) {
+		struct drm_display_mode *mode = drm_mode_create(connector->dev);
+
+		if (!mode)
+			return -EINVAL;
+		drm_mode_copy(mode, &imxpd->mode);
+		mode->type |= DRM_MODE_TYPE_DRIVER | DRM_MODE_TYPE_PREFERRED,
+		drm_mode_probed_add(connector, mode);
+		num_modes++;
+	}
+
+	if (np) {
+		struct drm_display_mode *mode = drm_mode_create(connector->dev);
+
+		if (!mode)
+			return -EINVAL;
+		of_get_drm_display_mode(np, &imxpd->mode, OF_USE_NATIVE_MODE);
+		drm_mode_copy(mode, &imxpd->mode);
+		mode->type |= DRM_MODE_TYPE_DRIVER | DRM_MODE_TYPE_PREFERRED,
+		drm_mode_probed_add(connector, mode);
+		num_modes++;
+	}
+
+	return num_modes;
+}
+
+static struct drm_encoder *imx_pd_connector_best_encoder(
+		struct drm_connector *connector)
+{
+	struct imx_parallel_display *imxpd = con_to_imxpd(connector);
+
+	return &imxpd->encoder;
+}
+
+static void imx_pd_encoder_dpms(struct drm_encoder *encoder, int mode)
+{
+	struct imx_parallel_display *imxpd = enc_to_imxpd(encoder);
+
+	if (mode != DRM_MODE_DPMS_ON)
+		drm_panel_disable(imxpd->panel);
+	else
+		drm_panel_enable(imxpd->panel);
+}
+
+static bool imx_pd_encoder_mode_fixup(struct drm_encoder *encoder,
+			   const struct drm_display_mode *mode,
+			   struct drm_display_mode *adjusted_mode)
+{
+	return true;
+}
+
+static void imx_pd_encoder_prepare(struct drm_encoder *encoder)
+{
+	struct imx_parallel_display *imxpd = enc_to_imxpd(encoder);
+
+	imx_drm_panel_format(encoder, imxpd->interface_pix_fmt);
+}
+
+static void imx_pd_encoder_commit(struct drm_encoder *encoder)
+{
+	struct imx_parallel_display *imxpd = enc_to_imxpd(encoder);
+
+	drm_panel_prepare(imxpd->panel);
+	drm_panel_enable(imxpd->panel);
+}
+
+static void imx_pd_encoder_mode_set(struct drm_encoder *encoder,
+			 struct drm_display_mode *mode,
+			 struct drm_display_mode *adjusted_mode)
+{
+}
+
+static void imx_pd_encoder_disable(struct drm_encoder *encoder)
+{
+	struct imx_parallel_display *imxpd = enc_to_imxpd(encoder);
+
+	drm_panel_disable(imxpd->panel);
+	drm_panel_unprepare(imxpd->panel);
+}
+
+static struct drm_connector_funcs imx_pd_connector_funcs = {
+	.dpms = drm_helper_connector_dpms,
+	.fill_modes = drm_helper_probe_single_connector_modes,
+	.detect = imx_pd_connector_detect,
+	.destroy = imx_drm_connector_destroy,
+};
+
+static struct drm_connector_helper_funcs imx_pd_connector_helper_funcs = {
+	.get_modes = imx_pd_connector_get_modes,
+	.best_encoder = imx_pd_connector_best_encoder,
+};
+
+static struct drm_encoder_funcs imx_pd_encoder_funcs = {
+	.destroy = imx_drm_encoder_destroy,
+};
+
+static struct drm_encoder_helper_funcs imx_pd_encoder_helper_funcs = {
+	.dpms = imx_pd_encoder_dpms,
+	.mode_fixup = imx_pd_encoder_mode_fixup,
+	.prepare = imx_pd_encoder_prepare,
+	.commit = imx_pd_encoder_commit,
+	.mode_set = imx_pd_encoder_mode_set,
+	.disable = imx_pd_encoder_disable,
+};
+
+static int imx_pd_register(struct drm_device *drm,
+	struct imx_parallel_display *imxpd)
+{
+	int ret;
+
+	ret = imx_drm_encoder_parse_of(drm, &imxpd->encoder,
+				       imxpd->dev->of_node);
+	if (ret)
+		return ret;
+
+	/* set the connector's dpms to OFF so that
+	 * drm_helper_connector_dpms() won't return
+	 * immediately since the current state is ON
+	 * at this point.
+	 */
+	imxpd->connector.dpms = DRM_MODE_DPMS_OFF;
+
+	drm_encoder_helper_add(&imxpd->encoder, &imx_pd_encoder_helper_funcs);
+	drm_encoder_init(drm, &imxpd->encoder, &imx_pd_encoder_funcs,
+			 DRM_MODE_ENCODER_NONE);
+
+	drm_connector_helper_add(&imxpd->connector,
+			&imx_pd_connector_helper_funcs);
+	drm_connector_init(drm, &imxpd->connector, &imx_pd_connector_funcs,
+			   DRM_MODE_CONNECTOR_VGA);
+
+	if (imxpd->panel)
+		drm_panel_attach(imxpd->panel, &imxpd->connector);
+
+	drm_mode_connector_attach_encoder(&imxpd->connector, &imxpd->encoder);
+
+	imxpd->connector.encoder = &imxpd->encoder;
+
+	return 0;
+}
+
+static int imx_pd_bind(struct device *dev, struct device *master, void *data)
+{
+	struct drm_device *drm = data;
+	struct device_node *np = dev->of_node;
+	struct device_node *panel_node;
+	const u8 *edidp;
+	struct imx_parallel_display *imxpd;
+	int ret;
+	const char *fmt;
+
+	imxpd = devm_kzalloc(dev, sizeof(*imxpd), GFP_KERNEL);
+	if (!imxpd)
+		return -ENOMEM;
+
+	edidp = of_get_property(np, "edid", &imxpd->edid_len);
+	if (edidp)
+		imxpd->edid = kmemdup(edidp, imxpd->edid_len, GFP_KERNEL);
+
+	ret = of_property_read_string(np, "interface-pix-fmt", &fmt);
+	if (!ret) {
+		if (!strcmp(fmt, "rgb24"))
+			imxpd->interface_pix_fmt = V4L2_PIX_FMT_RGB24;
+		else if (!strcmp(fmt, "rgb565"))
+			imxpd->interface_pix_fmt = V4L2_PIX_FMT_RGB565;
+		else if (!strcmp(fmt, "bgr666"))
+			imxpd->interface_pix_fmt = V4L2_PIX_FMT_BGR666;
+		else if (!strcmp(fmt, "lvds666"))
+			imxpd->interface_pix_fmt =
+					v4l2_fourcc('L', 'V', 'D', '6');
+	}
+
+	panel_node = of_parse_phandle(np, "fsl,panel", 0);
+	if (panel_node)
+		imxpd->panel = of_drm_find_panel(panel_node);
+
+	imxpd->dev = dev;
+
+	ret = imx_pd_register(drm, imxpd);
+	if (ret)
+		return ret;
+
+	dev_set_drvdata(dev, imxpd);
+
+	return 0;
+}
+
+static void imx_pd_unbind(struct device *dev, struct device *master,
+	void *data)
+{
+	struct imx_parallel_display *imxpd = dev_get_drvdata(dev);
+
+	imxpd->encoder.funcs->destroy(&imxpd->encoder);
+	imxpd->connector.funcs->destroy(&imxpd->connector);
+}
+
+static const struct component_ops imx_pd_ops = {
+	.bind	= imx_pd_bind,
+	.unbind	= imx_pd_unbind,
+};
+
+static int imx_pd_probe(struct platform_device *pdev)
+{
+	return component_add(&pdev->dev, &imx_pd_ops);
+}
+
+static int imx_pd_remove(struct platform_device *pdev)
+{
+	component_del(&pdev->dev, &imx_pd_ops);
+	return 0;
+}
+
+static const struct of_device_id imx_pd_dt_ids[] = {
+	{ .compatible = "fsl,imx-parallel-display", },
+	{ /* sentinel */ }
+};
+MODULE_DEVICE_TABLE(of, imx_pd_dt_ids);
+
+static struct platform_driver imx_pd_driver = {
+	.probe		= imx_pd_probe,
+	.remove		= imx_pd_remove,
+	.driver		= {
+		.of_match_table = imx_pd_dt_ids,
+		.name	= "imx-parallel-display",
+	},
+};
+
+module_platform_driver(imx_pd_driver);
+
+MODULE_DESCRIPTION("i.MX parallel display driver");
+MODULE_AUTHOR("Sascha Hauer, Pengutronix");
+MODULE_LICENSE("GPL");
+MODULE_ALIAS("platform:imx-parallel-display");
diff --git a/drivers/gpu/drm/mga/mga_dma.c b/drivers/gpu/drm/mga/mga_dma.c
index c3bf059ba720..8cfa9cb74c86 100644
--- a/drivers/gpu/drm/mga/mga_dma.c
+++ b/drivers/gpu/drm/mga/mga_dma.c
@@ -502,31 +502,31 @@ static int mga_do_agp_dma_bootstrap(struct drm_device *dev,
 		return err;
 	}
 
-	/* Make drm_addbufs happy by not trying to create a mapping for less
-	 * than a page.
+	/* Make drm_legacy_addbufs happy by not trying to create a mapping for
+	 * less than a page.
 	 */
 	if (warp_size < PAGE_SIZE)
 		warp_size = PAGE_SIZE;
 
 	offset = 0;
-	err = drm_addmap(dev, offset, warp_size,
-			 _DRM_AGP, _DRM_READ_ONLY, &dev_priv->warp);
+	err = drm_legacy_addmap(dev, offset, warp_size,
+				_DRM_AGP, _DRM_READ_ONLY, &dev_priv->warp);
 	if (err) {
 		DRM_ERROR("Unable to map WARP microcode: %d\n", err);
 		return err;
 	}
 
 	offset += warp_size;
-	err = drm_addmap(dev, offset, dma_bs->primary_size,
-			 _DRM_AGP, _DRM_READ_ONLY, &dev_priv->primary);
+	err = drm_legacy_addmap(dev, offset, dma_bs->primary_size,
+				_DRM_AGP, _DRM_READ_ONLY, &dev_priv->primary);
 	if (err) {
 		DRM_ERROR("Unable to map primary DMA region: %d\n", err);
 		return err;
 	}
 
 	offset += dma_bs->primary_size;
-	err = drm_addmap(dev, offset, secondary_size,
-			 _DRM_AGP, 0, &dev->agp_buffer_map);
+	err = drm_legacy_addmap(dev, offset, secondary_size,
+				_DRM_AGP, 0, &dev->agp_buffer_map);
 	if (err) {
 		DRM_ERROR("Unable to map secondary DMA region: %d\n", err);
 		return err;
@@ -538,7 +538,7 @@ static int mga_do_agp_dma_bootstrap(struct drm_device *dev,
 	req.flags = _DRM_AGP_BUFFER;
 	req.agp_start = offset;
 
-	err = drm_addbufs_agp(dev, &req);
+	err = drm_legacy_addbufs_agp(dev, &req);
 	if (err) {
 		DRM_ERROR("Unable to add secondary DMA buffers: %d\n", err);
 		return err;
@@ -559,16 +559,16 @@ static int mga_do_agp_dma_bootstrap(struct drm_device *dev,
 	}
 
 	offset += secondary_size;
-	err = drm_addmap(dev, offset, agp_size - offset,
-			 _DRM_AGP, 0, &dev_priv->agp_textures);
+	err = drm_legacy_addmap(dev, offset, agp_size - offset,
+				_DRM_AGP, 0, &dev_priv->agp_textures);
 	if (err) {
 		DRM_ERROR("Unable to map AGP texture region %d\n", err);
 		return err;
 	}
 
-	drm_core_ioremap(dev_priv->warp, dev);
-	drm_core_ioremap(dev_priv->primary, dev);
-	drm_core_ioremap(dev->agp_buffer_map, dev);
+	drm_legacy_ioremap(dev_priv->warp, dev);
+	drm_legacy_ioremap(dev_priv->primary, dev);
+	drm_legacy_ioremap(dev->agp_buffer_map, dev);
 
 	if (!dev_priv->warp->handle ||
 	    !dev_priv->primary->handle || !dev->agp_buffer_map->handle) {
@@ -602,7 +602,7 @@ static int mga_do_agp_dma_bootstrap(struct drm_device *dev,
  *
  * \todo
  * Determine whether the maximum address passed to drm_pci_alloc is correct.
- * The same goes for drm_addbufs_pci.
+ * The same goes for drm_legacy_addbufs_pci.
  *
  * \sa mga_do_dma_bootstrap, mga_do_agp_dma_bootstrap
  */
@@ -622,15 +622,15 @@ static int mga_do_pci_dma_bootstrap(struct drm_device *dev,
 		return -EFAULT;
 	}
 
-	/* Make drm_addbufs happy by not trying to create a mapping for less
-	 * than a page.
+	/* Make drm_legacy_addbufs happy by not trying to create a mapping for
+	 * less than a page.
 	 */
 	if (warp_size < PAGE_SIZE)
 		warp_size = PAGE_SIZE;
 
 	/* The proper alignment is 0x100 for this mapping */
-	err = drm_addmap(dev, 0, warp_size, _DRM_CONSISTENT,
-			 _DRM_READ_ONLY, &dev_priv->warp);
+	err = drm_legacy_addmap(dev, 0, warp_size, _DRM_CONSISTENT,
+				_DRM_READ_ONLY, &dev_priv->warp);
 	if (err != 0) {
 		DRM_ERROR("Unable to create mapping for WARP microcode: %d\n",
 			  err);
@@ -645,8 +645,8 @@ static int mga_do_pci_dma_bootstrap(struct drm_device *dev,
 	for (primary_size = dma_bs->primary_size; primary_size != 0;
 	     primary_size >>= 1) {
 		/* The proper alignment for this mapping is 0x04 */
-		err = drm_addmap(dev, 0, primary_size, _DRM_CONSISTENT,
-				 _DRM_READ_ONLY, &dev_priv->primary);
+		err = drm_legacy_addmap(dev, 0, primary_size, _DRM_CONSISTENT,
+					_DRM_READ_ONLY, &dev_priv->primary);
 		if (!err)
 			break;
 	}
@@ -669,7 +669,7 @@ static int mga_do_pci_dma_bootstrap(struct drm_device *dev,
 		req.count = bin_count;
 		req.size = dma_bs->secondary_bin_size;
 
-		err = drm_addbufs_pci(dev, &req);
+		err = drm_legacy_addbufs_pci(dev, &req);
 		if (!err)
 			break;
 	}
@@ -708,15 +708,16 @@ static int mga_do_dma_bootstrap(struct drm_device *dev,
 	/* The first steps are the same for both PCI and AGP based DMA.  Map
 	 * the cards MMIO registers and map a status page.
 	 */
-	err = drm_addmap(dev, dev_priv->mmio_base, dev_priv->mmio_size,
-			 _DRM_REGISTERS, _DRM_READ_ONLY, &dev_priv->mmio);
+	err = drm_legacy_addmap(dev, dev_priv->mmio_base, dev_priv->mmio_size,
+				_DRM_REGISTERS, _DRM_READ_ONLY,
+				&dev_priv->mmio);
 	if (err) {
 		DRM_ERROR("Unable to map MMIO region: %d\n", err);
 		return err;
 	}
 
-	err = drm_addmap(dev, 0, SAREA_MAX, _DRM_SHM,
-			 _DRM_READ_ONLY | _DRM_LOCKED | _DRM_KERNEL,
+	err = drm_legacy_addmap(dev, 0, SAREA_MAX, _DRM_SHM,
+				_DRM_READ_ONLY | _DRM_LOCKED | _DRM_KERNEL,
 			 &dev_priv->status);
 	if (err) {
 		DRM_ERROR("Unable to map status region: %d\n", err);
@@ -809,7 +810,7 @@ static int mga_do_init_dma(struct drm_device *dev, drm_mga_init_t *init)
 	dev_priv->texture_offset = init->texture_offset[0];
 	dev_priv->texture_size = init->texture_size[0];
 
-	dev_priv->sarea = drm_getsarea(dev);
+	dev_priv->sarea = drm_legacy_getsarea(dev);
 	if (!dev_priv->sarea) {
 		DRM_ERROR("failed to find sarea!\n");
 		return -EINVAL;
@@ -820,37 +821,37 @@ static int mga_do_init_dma(struct drm_device *dev, drm_mga_init_t *init)
 		dev_priv->dma_access = MGA_PAGPXFER;
 		dev_priv->wagp_enable = MGA_WAGP_ENABLE;
 
-		dev_priv->status = drm_core_findmap(dev, init->status_offset);
+		dev_priv->status = drm_legacy_findmap(dev, init->status_offset);
 		if (!dev_priv->status) {
 			DRM_ERROR("failed to find status page!\n");
 			return -EINVAL;
 		}
-		dev_priv->mmio = drm_core_findmap(dev, init->mmio_offset);
+		dev_priv->mmio = drm_legacy_findmap(dev, init->mmio_offset);
 		if (!dev_priv->mmio) {
 			DRM_ERROR("failed to find mmio region!\n");
 			return -EINVAL;
 		}
-		dev_priv->warp = drm_core_findmap(dev, init->warp_offset);
+		dev_priv->warp = drm_legacy_findmap(dev, init->warp_offset);
 		if (!dev_priv->warp) {
 			DRM_ERROR("failed to find warp microcode region!\n");
 			return -EINVAL;
 		}
-		dev_priv->primary = drm_core_findmap(dev, init->primary_offset);
+		dev_priv->primary = drm_legacy_findmap(dev, init->primary_offset);
 		if (!dev_priv->primary) {
 			DRM_ERROR("failed to find primary dma region!\n");
 			return -EINVAL;
 		}
 		dev->agp_buffer_token = init->buffers_offset;
 		dev->agp_buffer_map =
-		    drm_core_findmap(dev, init->buffers_offset);
+		    drm_legacy_findmap(dev, init->buffers_offset);
 		if (!dev->agp_buffer_map) {
 			DRM_ERROR("failed to find dma buffer region!\n");
 			return -EINVAL;
 		}
 
-		drm_core_ioremap(dev_priv->warp, dev);
-		drm_core_ioremap(dev_priv->primary, dev);
-		drm_core_ioremap(dev->agp_buffer_map, dev);
+		drm_legacy_ioremap(dev_priv->warp, dev);
+		drm_legacy_ioremap(dev_priv->primary, dev);
+		drm_legacy_ioremap(dev->agp_buffer_map, dev);
 	}
 
 	dev_priv->sarea_priv =
@@ -936,14 +937,14 @@ static int mga_do_cleanup_dma(struct drm_device *dev, int full_cleanup)
 
 		if ((dev_priv->warp != NULL)
 		    && (dev_priv->warp->type != _DRM_CONSISTENT))
-			drm_core_ioremapfree(dev_priv->warp, dev);
+			drm_legacy_ioremapfree(dev_priv->warp, dev);
 
 		if ((dev_priv->primary != NULL)
 		    && (dev_priv->primary->type != _DRM_CONSISTENT))
-			drm_core_ioremapfree(dev_priv->primary, dev);
+			drm_legacy_ioremapfree(dev_priv->primary, dev);
 
 		if (dev->agp_buffer_map != NULL)
-			drm_core_ioremapfree(dev->agp_buffer_map, dev);
+			drm_legacy_ioremapfree(dev->agp_buffer_map, dev);
 
 		if (dev_priv->used_new_dma_init) {
 #if __OS_HAS_AGP
diff --git a/drivers/gpu/drm/mga/mga_drv.c b/drivers/gpu/drm/mga/mga_drv.c
index 6b1a87c8aac5..5e2f131a6a72 100644
--- a/drivers/gpu/drm/mga/mga_drv.c
+++ b/drivers/gpu/drm/mga/mga_drv.c
@@ -48,7 +48,7 @@ static const struct file_operations mga_driver_fops = {
 	.open = drm_open,
 	.release = drm_release,
 	.unlocked_ioctl = drm_ioctl,
-	.mmap = drm_mmap,
+	.mmap = drm_legacy_mmap,
 	.poll = drm_poll,
 #ifdef CONFIG_COMPAT
 	.compat_ioctl = mga_compat_ioctl,
@@ -64,6 +64,7 @@ static struct drm_driver driver = {
 	.load = mga_driver_load,
 	.unload = mga_driver_unload,
 	.lastclose = mga_driver_lastclose,
+	.set_busid = drm_pci_set_busid,
 	.dma_quiescent = mga_driver_dma_quiescent,
 	.device_is_agp = mga_driver_device_is_agp,
 	.get_vblank_counter = mga_get_vblank_counter,
diff --git a/drivers/gpu/drm/mga/mga_drv.h b/drivers/gpu/drm/mga/mga_drv.h
index fe453213600a..b4a2014917e5 100644
--- a/drivers/gpu/drm/mga/mga_drv.h
+++ b/drivers/gpu/drm/mga/mga_drv.h
@@ -31,6 +31,8 @@
 #ifndef __MGA_DRV_H__
 #define __MGA_DRV_H__
 
+#include <drm/drm_legacy.h>
+
 /* General customization:
  */
 
diff --git a/drivers/gpu/drm/mgag200/mgag200_drv.c b/drivers/gpu/drm/mgag200/mgag200_drv.c
index 2d75d6df0789..97745991544d 100644
--- a/drivers/gpu/drm/mgag200/mgag200_drv.c
+++ b/drivers/gpu/drm/mgag200/mgag200_drv.c
@@ -91,6 +91,7 @@ static struct drm_driver driver = {
 	.driver_features = DRIVER_GEM | DRIVER_MODESET,
 	.load = mgag200_driver_load,
 	.unload = mgag200_driver_unload,
+	.set_busid = drm_pci_set_busid,
 	.fops = &mgag200_driver_fops,
 	.name = DRIVER_NAME,
 	.desc = DRIVER_DESC,
diff --git a/drivers/gpu/drm/mgag200/mgag200_drv.h b/drivers/gpu/drm/mgag200/mgag200_drv.h
index 80de23d9b9c9..e9eea1d4e7c3 100644
--- a/drivers/gpu/drm/mgag200/mgag200_drv.h
+++ b/drivers/gpu/drm/mgag200/mgag200_drv.h
@@ -22,6 +22,8 @@
 #include <drm/ttm/ttm_memory.h>
 #include <drm/ttm/ttm_module.h>
 
+#include <drm/drm_gem.h>
+
 #include <linux/i2c.h>
 #include <linux/i2c-algo-bit.h>
 
@@ -190,8 +192,6 @@ struct mga_device {
 	resource_size_t			rmmio_size;
 	void __iomem			*rmmio;
 
-	drm_local_map_t			*framebuffer;
-
 	struct mga_mc			mc;
 	struct mga_mode_info		mode_info;
 
@@ -224,7 +224,7 @@ struct mgag200_bo {
 	struct ttm_placement placement;
 	struct ttm_bo_kmap_obj kmap;
 	struct drm_gem_object gem;
-	u32 placements[3];
+	struct ttm_place placements[3];
 	int pin_count;
 };
 #define gem_to_mga_bo(gobj) container_of((gobj), struct mgag200_bo, gem)
diff --git a/drivers/gpu/drm/mgag200/mgag200_fb.c b/drivers/gpu/drm/mgag200/mgag200_fb.c
index 5451dc58eff1..4415af3666ab 100644
--- a/drivers/gpu/drm/mgag200/mgag200_fb.c
+++ b/drivers/gpu/drm/mgag200/mgag200_fb.c
@@ -158,7 +158,8 @@ static int mgag200fb_create_object(struct mga_fbdev *afbdev,
 static int mgag200fb_create(struct drm_fb_helper *helper,
 			   struct drm_fb_helper_surface_size *sizes)
 {
-	struct mga_fbdev *mfbdev = (struct mga_fbdev *)helper;
+	struct mga_fbdev *mfbdev =
+		container_of(helper, struct mga_fbdev, helper);
 	struct drm_device *dev = mfbdev->helper.dev;
 	struct drm_mode_fb_cmd2 mode_cmd;
 	struct mga_device *mdev = dev->dev_private;
diff --git a/drivers/gpu/drm/mgag200/mgag200_mode.c b/drivers/gpu/drm/mgag200/mgag200_mode.c
index 45f04dea0ac2..9872ba9abf1a 100644
--- a/drivers/gpu/drm/mgag200/mgag200_mode.c
+++ b/drivers/gpu/drm/mgag200/mgag200_mode.c
@@ -15,6 +15,7 @@
 
 #include <drm/drmP.h>
 #include <drm/drm_crtc_helper.h>
+#include <drm/drm_plane_helper.h>
 
 #include "mgag200_drv.h"
 
@@ -1483,11 +1484,7 @@ static int mga_vga_mode_valid(struct drm_connector *connector,
 {
 	struct drm_device *dev = connector->dev;
 	struct mga_device *mdev = (struct mga_device*)dev->dev_private;
-	struct mga_fbdev *mfbdev = mdev->mfbdev;
-	struct drm_fb_helper *fb_helper = &mfbdev->helper;
-	struct drm_fb_helper_connector *fb_helper_conn = NULL;
 	int bpp = 32;
-	int i = 0;
 
 	if (IS_G200_SE(mdev)) {
 		if (mdev->unique_rev_id == 0x01) {
@@ -1537,21 +1534,14 @@ static int mga_vga_mode_valid(struct drm_connector *connector,
 	}
 
 	/* Validate the mode input by the user */
-	for (i = 0; i < fb_helper->connector_count; i++) {
-		if (fb_helper->connector_info[i]->connector == connector) {
-			/* Found the helper for this connector */
-			fb_helper_conn = fb_helper->connector_info[i];
-			if (fb_helper_conn->cmdline_mode.specified) {
-				if (fb_helper_conn->cmdline_mode.bpp_specified) {
-					bpp = fb_helper_conn->cmdline_mode.bpp;
-				}
-			}
-		}
+	if (connector->cmdline_mode.specified) {
+		if (connector->cmdline_mode.bpp_specified)
+			bpp = connector->cmdline_mode.bpp;
 	}
 
 	if ((mode->hdisplay * mode->vdisplay * (bpp/8)) > mdev->mc.vram_size) {
-		if (fb_helper_conn)
-			fb_helper_conn->cmdline_mode.specified = false;
+		if (connector->cmdline_mode.specified)
+			connector->cmdline_mode.specified = false;
 		return MODE_BAD;
 	}
 
diff --git a/drivers/gpu/drm/mgag200/mgag200_ttm.c b/drivers/gpu/drm/mgag200/mgag200_ttm.c
index 5a00e90696de..d16964ea0ed4 100644
--- a/drivers/gpu/drm/mgag200/mgag200_ttm.c
+++ b/drivers/gpu/drm/mgag200/mgag200_ttm.c
@@ -293,18 +293,22 @@ void mgag200_mm_fini(struct mga_device *mdev)
 void mgag200_ttm_placement(struct mgag200_bo *bo, int domain)
 {
 	u32 c = 0;
-	bo->placement.fpfn = 0;
-	bo->placement.lpfn = 0;
+	unsigned i;
+
 	bo->placement.placement = bo->placements;
 	bo->placement.busy_placement = bo->placements;
 	if (domain & TTM_PL_FLAG_VRAM)
-		bo->placements[c++] = TTM_PL_FLAG_WC | TTM_PL_FLAG_UNCACHED | TTM_PL_FLAG_VRAM;
+		bo->placements[c++].flags = TTM_PL_FLAG_WC | TTM_PL_FLAG_UNCACHED | TTM_PL_FLAG_VRAM;
 	if (domain & TTM_PL_FLAG_SYSTEM)
-		bo->placements[c++] = TTM_PL_MASK_CACHING | TTM_PL_FLAG_SYSTEM;
+		bo->placements[c++].flags = TTM_PL_MASK_CACHING | TTM_PL_FLAG_SYSTEM;
 	if (!c)
-		bo->placements[c++] = TTM_PL_MASK_CACHING | TTM_PL_FLAG_SYSTEM;
+		bo->placements[c++].flags = TTM_PL_MASK_CACHING | TTM_PL_FLAG_SYSTEM;
 	bo->placement.num_placement = c;
 	bo->placement.num_busy_placement = c;
+	for (i = 0; i < c; ++i) {
+		bo->placements[i].fpfn = 0;
+		bo->placements[i].lpfn = 0;
+	}
 }
 
 int mgag200_bo_create(struct drm_device *dev, int size, int align,
@@ -335,7 +339,7 @@ int mgag200_bo_create(struct drm_device *dev, int size, int align,
 	ret = ttm_bo_init(&mdev->ttm.bdev, &mgabo->bo, size,
 			  ttm_bo_type_device, &mgabo->placement,
 			  align >> PAGE_SHIFT, false, NULL, acc_size,
-			  NULL, mgag200_bo_ttm_destroy);
+			  NULL, NULL, mgag200_bo_ttm_destroy);
 	if (ret)
 		return ret;
 
@@ -361,7 +365,7 @@ int mgag200_bo_pin(struct mgag200_bo *bo, u32 pl_flag, u64 *gpu_addr)
 
 	mgag200_ttm_placement(bo, pl_flag);
 	for (i = 0; i < bo->placement.num_placement; i++)
-		bo->placements[i] |= TTM_PL_FLAG_NO_EVICT;
+		bo->placements[i].flags |= TTM_PL_FLAG_NO_EVICT;
 	ret = ttm_bo_validate(&bo->bo, &bo->placement, false, false);
 	if (ret)
 		return ret;
@@ -384,7 +388,7 @@ int mgag200_bo_unpin(struct mgag200_bo *bo)
 		return 0;
 
 	for (i = 0; i < bo->placement.num_placement ; i++)
-		bo->placements[i] &= ~TTM_PL_FLAG_NO_EVICT;
+		bo->placements[i].flags &= ~TTM_PL_FLAG_NO_EVICT;
 	ret = ttm_bo_validate(&bo->bo, &bo->placement, false, false);
 	if (ret)
 		return ret;
@@ -408,7 +412,7 @@ int mgag200_bo_push_sysram(struct mgag200_bo *bo)
 
 	mgag200_ttm_placement(bo, TTM_PL_FLAG_SYSTEM);
 	for (i = 0; i < bo->placement.num_placement ; i++)
-		bo->placements[i] |= TTM_PL_FLAG_NO_EVICT;
+		bo->placements[i].flags |= TTM_PL_FLAG_NO_EVICT;
 
 	ret = ttm_bo_validate(&bo->bo, &bo->placement, false, false);
 	if (ret) {
@@ -424,7 +428,7 @@ int mgag200_mmap(struct file *filp, struct vm_area_struct *vma)
 	struct mga_device *mdev;
 
 	if (unlikely(vma->vm_pgoff < DRM_FILE_PAGE_OFFSET))
-		return drm_mmap(filp, vma);
+		return -EINVAL;
 
 	file_priv = filp->private_data;
 	mdev = file_priv->minor->dev->dev_private;
diff --git a/drivers/gpu/drm/msm/Kconfig b/drivers/gpu/drm/msm/Kconfig
index c99c50de3226..5b2a1ff95d3d 100644
--- a/drivers/gpu/drm/msm/Kconfig
+++ b/drivers/gpu/drm/msm/Kconfig
@@ -3,7 +3,9 @@ config DRM_MSM
 	tristate "MSM DRM"
 	depends on DRM
 	depends on ARCH_QCOM || (ARM && COMPILE_TEST)
+	select REGULATOR
 	select DRM_KMS_HELPER
+	select DRM_PANEL
 	select SHMEM
 	select TMPFS
 	default y
diff --git a/drivers/gpu/drm/msm/Makefile b/drivers/gpu/drm/msm/Makefile
index 93ca49c8df44..143d988f8add 100644
--- a/drivers/gpu/drm/msm/Makefile
+++ b/drivers/gpu/drm/msm/Makefile
@@ -4,8 +4,10 @@ ifeq (, $(findstring -W,$(EXTRA_CFLAGS)))
 endif
 
 msm-y := \
+	adreno/adreno_device.o \
 	adreno/adreno_gpu.o \
 	adreno/a3xx_gpu.o \
+	adreno/a4xx_gpu.o \
 	hdmi/hdmi.o \
 	hdmi/hdmi_audio.o \
 	hdmi/hdmi_bridge.o \
@@ -18,15 +20,20 @@ msm-y := \
 	mdp/mdp_kms.o \
 	mdp/mdp4/mdp4_crtc.o \
 	mdp/mdp4/mdp4_dtv_encoder.o \
+	mdp/mdp4/mdp4_lcdc_encoder.o \
+	mdp/mdp4/mdp4_lvds_connector.o \
 	mdp/mdp4/mdp4_irq.o \
 	mdp/mdp4/mdp4_kms.o \
 	mdp/mdp4/mdp4_plane.o \
+	mdp/mdp5/mdp5_cfg.o \
+	mdp/mdp5/mdp5_ctl.o \
 	mdp/mdp5/mdp5_crtc.o \
 	mdp/mdp5/mdp5_encoder.o \
 	mdp/mdp5/mdp5_irq.o \
 	mdp/mdp5/mdp5_kms.o \
 	mdp/mdp5/mdp5_plane.o \
 	mdp/mdp5/mdp5_smp.o \
+	msm_atomic.o \
 	msm_drv.o \
 	msm_fb.o \
 	msm_gem.o \
@@ -39,5 +46,6 @@ msm-y := \
 	msm_ringbuffer.o
 
 msm-$(CONFIG_DRM_MSM_FBDEV) += msm_fbdev.o
+msm-$(CONFIG_COMMON_CLK) += mdp/mdp4/mdp4_lvds_pll.o
 
 obj-$(CONFIG_DRM_MSM)	+= msm.o
diff --git a/drivers/gpu/drm/msm/adreno/a2xx.xml.h b/drivers/gpu/drm/msm/adreno/a2xx.xml.h
index a8a144b38eaa..22882cc0a573 100644
--- a/drivers/gpu/drm/msm/adreno/a2xx.xml.h
+++ b/drivers/gpu/drm/msm/adreno/a2xx.xml.h
@@ -11,10 +11,10 @@ The rules-ng-ng source files this header was generated from are:
 - /home/robclark/src/freedreno/envytools/rnndb/adreno.xml               (    364 bytes, from 2013-11-30 14:47:15)
 - /home/robclark/src/freedreno/envytools/rnndb/freedreno_copyright.xml  (   1453 bytes, from 2013-03-31 16:51:27)
 - /home/robclark/src/freedreno/envytools/rnndb/adreno/a2xx.xml          (  32901 bytes, from 2014-06-02 15:21:30)
-- /home/robclark/src/freedreno/envytools/rnndb/adreno/adreno_common.xml (   9859 bytes, from 2014-06-02 15:21:30)
-- /home/robclark/src/freedreno/envytools/rnndb/adreno/adreno_pm4.xml    (  14477 bytes, from 2014-05-16 11:51:57)
-- /home/robclark/src/freedreno/envytools/rnndb/adreno/a3xx.xml          (  58020 bytes, from 2014-06-25 12:57:16)
-- /home/robclark/src/freedreno/envytools/rnndb/adreno/a4xx.xml          (  26602 bytes, from 2014-06-25 12:57:16)
+- /home/robclark/src/freedreno/envytools/rnndb/adreno/adreno_common.xml (  10551 bytes, from 2014-11-13 22:44:30)
+- /home/robclark/src/freedreno/envytools/rnndb/adreno/adreno_pm4.xml    (  15053 bytes, from 2014-11-09 15:45:47)
+- /home/robclark/src/freedreno/envytools/rnndb/adreno/a3xx.xml          (  63169 bytes, from 2014-11-13 22:44:18)
+- /home/robclark/src/freedreno/envytools/rnndb/adreno/a4xx.xml          (  49097 bytes, from 2014-11-14 15:38:00)
 
 Copyright (C) 2013-2014 by the following authors:
 - Rob Clark <robdclark@gmail.com> (robclark)
@@ -926,11 +926,11 @@ static inline uint32_t A2XX_VGT_DRAW_INITIATOR_INDEX_SIZE(enum pc_di_index_size
 #define A2XX_VGT_DRAW_INITIATOR_NOT_EOP				0x00001000
 #define A2XX_VGT_DRAW_INITIATOR_SMALL_INDEX			0x00002000
 #define A2XX_VGT_DRAW_INITIATOR_PRE_DRAW_INITIATOR_ENABLE	0x00004000
-#define A2XX_VGT_DRAW_INITIATOR_NUM_INDICES__MASK		0xffff0000
-#define A2XX_VGT_DRAW_INITIATOR_NUM_INDICES__SHIFT		16
-static inline uint32_t A2XX_VGT_DRAW_INITIATOR_NUM_INDICES(uint32_t val)
+#define A2XX_VGT_DRAW_INITIATOR_NUM_INSTANCES__MASK		0xff000000
+#define A2XX_VGT_DRAW_INITIATOR_NUM_INSTANCES__SHIFT		24
+static inline uint32_t A2XX_VGT_DRAW_INITIATOR_NUM_INSTANCES(uint32_t val)
 {
-	return ((val) << A2XX_VGT_DRAW_INITIATOR_NUM_INDICES__SHIFT) & A2XX_VGT_DRAW_INITIATOR_NUM_INDICES__MASK;
+	return ((val) << A2XX_VGT_DRAW_INITIATOR_NUM_INSTANCES__SHIFT) & A2XX_VGT_DRAW_INITIATOR_NUM_INSTANCES__MASK;
 }
 
 #define REG_A2XX_VGT_IMMED_DATA					0x000021fd
@@ -1243,13 +1243,13 @@ static inline uint32_t A2XX_CLEAR_COLOR_ALPHA(uint32_t val)
 #define A2XX_PA_SU_POINT_SIZE_HEIGHT__SHIFT			0
 static inline uint32_t A2XX_PA_SU_POINT_SIZE_HEIGHT(float val)
 {
-	return ((((uint32_t)(val * 8.0))) << A2XX_PA_SU_POINT_SIZE_HEIGHT__SHIFT) & A2XX_PA_SU_POINT_SIZE_HEIGHT__MASK;
+	return ((((uint32_t)(val * 16.0))) << A2XX_PA_SU_POINT_SIZE_HEIGHT__SHIFT) & A2XX_PA_SU_POINT_SIZE_HEIGHT__MASK;
 }
 #define A2XX_PA_SU_POINT_SIZE_WIDTH__MASK			0xffff0000
 #define A2XX_PA_SU_POINT_SIZE_WIDTH__SHIFT			16
 static inline uint32_t A2XX_PA_SU_POINT_SIZE_WIDTH(float val)
 {
-	return ((((uint32_t)(val * 8.0))) << A2XX_PA_SU_POINT_SIZE_WIDTH__SHIFT) & A2XX_PA_SU_POINT_SIZE_WIDTH__MASK;
+	return ((((uint32_t)(val * 16.0))) << A2XX_PA_SU_POINT_SIZE_WIDTH__SHIFT) & A2XX_PA_SU_POINT_SIZE_WIDTH__MASK;
 }
 
 #define REG_A2XX_PA_SU_POINT_MINMAX				0x00002281
@@ -1257,13 +1257,13 @@ static inline uint32_t A2XX_PA_SU_POINT_SIZE_WIDTH(float val)
 #define A2XX_PA_SU_POINT_MINMAX_MIN__SHIFT			0
 static inline uint32_t A2XX_PA_SU_POINT_MINMAX_MIN(float val)
 {
-	return ((((uint32_t)(val * 8.0))) << A2XX_PA_SU_POINT_MINMAX_MIN__SHIFT) & A2XX_PA_SU_POINT_MINMAX_MIN__MASK;
+	return ((((uint32_t)(val * 16.0))) << A2XX_PA_SU_POINT_MINMAX_MIN__SHIFT) & A2XX_PA_SU_POINT_MINMAX_MIN__MASK;
 }
 #define A2XX_PA_SU_POINT_MINMAX_MAX__MASK			0xffff0000
 #define A2XX_PA_SU_POINT_MINMAX_MAX__SHIFT			16
 static inline uint32_t A2XX_PA_SU_POINT_MINMAX_MAX(float val)
 {
-	return ((((uint32_t)(val * 8.0))) << A2XX_PA_SU_POINT_MINMAX_MAX__SHIFT) & A2XX_PA_SU_POINT_MINMAX_MAX__MASK;
+	return ((((uint32_t)(val * 16.0))) << A2XX_PA_SU_POINT_MINMAX_MAX__SHIFT) & A2XX_PA_SU_POINT_MINMAX_MAX__MASK;
 }
 
 #define REG_A2XX_PA_SU_LINE_CNTL				0x00002282
@@ -1271,7 +1271,7 @@ static inline uint32_t A2XX_PA_SU_POINT_MINMAX_MAX(float val)
 #define A2XX_PA_SU_LINE_CNTL_WIDTH__SHIFT			0
 static inline uint32_t A2XX_PA_SU_LINE_CNTL_WIDTH(float val)
 {
-	return ((((uint32_t)(val * 8.0))) << A2XX_PA_SU_LINE_CNTL_WIDTH__SHIFT) & A2XX_PA_SU_LINE_CNTL_WIDTH__MASK;
+	return ((((uint32_t)(val * 16.0))) << A2XX_PA_SU_LINE_CNTL_WIDTH__SHIFT) & A2XX_PA_SU_LINE_CNTL_WIDTH__MASK;
 }
 
 #define REG_A2XX_PA_SC_LINE_STIPPLE				0x00002283
diff --git a/drivers/gpu/drm/msm/adreno/a3xx.xml.h b/drivers/gpu/drm/msm/adreno/a3xx.xml.h
index 303e8a9e91a5..109e9a263daf 100644
--- a/drivers/gpu/drm/msm/adreno/a3xx.xml.h
+++ b/drivers/gpu/drm/msm/adreno/a3xx.xml.h
@@ -11,10 +11,10 @@ The rules-ng-ng source files this header was generated from are:
 - /home/robclark/src/freedreno/envytools/rnndb/adreno.xml               (    364 bytes, from 2013-11-30 14:47:15)
 - /home/robclark/src/freedreno/envytools/rnndb/freedreno_copyright.xml  (   1453 bytes, from 2013-03-31 16:51:27)
 - /home/robclark/src/freedreno/envytools/rnndb/adreno/a2xx.xml          (  32901 bytes, from 2014-06-02 15:21:30)
-- /home/robclark/src/freedreno/envytools/rnndb/adreno/adreno_common.xml (   9859 bytes, from 2014-06-02 15:21:30)
-- /home/robclark/src/freedreno/envytools/rnndb/adreno/adreno_pm4.xml    (  14477 bytes, from 2014-05-16 11:51:57)
-- /home/robclark/src/freedreno/envytools/rnndb/adreno/a3xx.xml          (  58020 bytes, from 2014-06-25 12:57:16)
-- /home/robclark/src/freedreno/envytools/rnndb/adreno/a4xx.xml          (  26602 bytes, from 2014-06-25 12:57:16)
+- /home/robclark/src/freedreno/envytools/rnndb/adreno/adreno_common.xml (  10551 bytes, from 2014-11-13 22:44:30)
+- /home/robclark/src/freedreno/envytools/rnndb/adreno/adreno_pm4.xml    (  15053 bytes, from 2014-11-09 15:45:47)
+- /home/robclark/src/freedreno/envytools/rnndb/adreno/a3xx.xml          (  63169 bytes, from 2014-11-13 22:44:18)
+- /home/robclark/src/freedreno/envytools/rnndb/adreno/a4xx.xml          (  49097 bytes, from 2014-11-14 15:38:00)
 
 Copyright (C) 2013-2014 by the following authors:
 - Rob Clark <robdclark@gmail.com> (robclark)
@@ -86,6 +86,14 @@ enum a3xx_vtx_fmt {
 	VFMT_NORM_USHORT_16_16 = 29,
 	VFMT_NORM_USHORT_16_16_16 = 30,
 	VFMT_NORM_USHORT_16_16_16_16 = 31,
+	VFMT_UINT_32 = 32,
+	VFMT_UINT_32_32 = 33,
+	VFMT_UINT_32_32_32 = 34,
+	VFMT_UINT_32_32_32_32 = 35,
+	VFMT_INT_32 = 36,
+	VFMT_INT_32_32 = 37,
+	VFMT_INT_32_32_32 = 38,
+	VFMT_INT_32_32_32_32 = 39,
 	VFMT_UBYTE_8 = 40,
 	VFMT_UBYTE_8_8 = 41,
 	VFMT_UBYTE_8_8_8 = 42,
@@ -112,7 +120,9 @@ enum a3xx_tex_fmt {
 	TFMT_NORM_USHORT_565 = 4,
 	TFMT_NORM_USHORT_5551 = 6,
 	TFMT_NORM_USHORT_4444 = 7,
+	TFMT_NORM_USHORT_Z16 = 9,
 	TFMT_NORM_UINT_X8Z24 = 10,
+	TFMT_FLOAT_Z32 = 11,
 	TFMT_NORM_UINT_NV12_UV_TILED = 17,
 	TFMT_NORM_UINT_NV12_Y_TILED = 19,
 	TFMT_NORM_UINT_NV12_UV = 21,
@@ -121,18 +131,38 @@ enum a3xx_tex_fmt {
 	TFMT_NORM_UINT_I420_U = 26,
 	TFMT_NORM_UINT_I420_V = 27,
 	TFMT_NORM_UINT_2_10_10_10 = 41,
+	TFMT_FLOAT_9_9_9_E5 = 42,
+	TFMT_FLOAT_10_11_11 = 43,
 	TFMT_NORM_UINT_A8 = 44,
 	TFMT_NORM_UINT_L8_A8 = 47,
 	TFMT_NORM_UINT_8 = 48,
 	TFMT_NORM_UINT_8_8 = 49,
 	TFMT_NORM_UINT_8_8_8 = 50,
 	TFMT_NORM_UINT_8_8_8_8 = 51,
+	TFMT_NORM_SINT_8_8 = 53,
+	TFMT_NORM_SINT_8_8_8_8 = 55,
+	TFMT_UINT_8_8 = 57,
+	TFMT_UINT_8_8_8_8 = 59,
+	TFMT_SINT_8_8 = 61,
+	TFMT_SINT_8_8_8_8 = 63,
 	TFMT_FLOAT_16 = 64,
 	TFMT_FLOAT_16_16 = 65,
 	TFMT_FLOAT_16_16_16_16 = 67,
+	TFMT_UINT_16 = 68,
+	TFMT_UINT_16_16 = 69,
+	TFMT_UINT_16_16_16_16 = 71,
+	TFMT_SINT_16 = 72,
+	TFMT_SINT_16_16 = 73,
+	TFMT_SINT_16_16_16_16 = 75,
 	TFMT_FLOAT_32 = 84,
 	TFMT_FLOAT_32_32 = 85,
 	TFMT_FLOAT_32_32_32_32 = 87,
+	TFMT_UINT_32 = 88,
+	TFMT_UINT_32_32 = 89,
+	TFMT_UINT_32_32_32_32 = 91,
+	TFMT_SINT_32 = 92,
+	TFMT_SINT_32_32 = 93,
+	TFMT_SINT_32_32_32_32 = 95,
 };
 
 enum a3xx_tex_fetchsize {
@@ -145,19 +175,34 @@ enum a3xx_tex_fetchsize {
 };
 
 enum a3xx_color_fmt {
+	RB_R5G6B5_UNORM = 0,
+	RB_R5G5B5A1_UNORM = 1,
+	RB_R4G4B4A4_UNORM = 3,
 	RB_R8G8B8_UNORM = 4,
 	RB_R8G8B8A8_UNORM = 8,
-	RB_Z16_UNORM = 12,
+	RB_R8G8B8A8_UINT = 10,
+	RB_R8G8B8A8_SINT = 11,
+	RB_R8G8_UNORM = 12,
+	RB_R8_UINT = 14,
+	RB_R8_SINT = 15,
+	RB_R10G10B10A2_UNORM = 16,
 	RB_A8_UNORM = 20,
+	RB_R8_UNORM = 21,
 	RB_R16G16B16A16_FLOAT = 27,
+	RB_R11G11B10_FLOAT = 28,
+	RB_R16_SINT = 40,
+	RB_R16G16_SINT = 41,
+	RB_R16G16B16A16_SINT = 43,
+	RB_R16_UINT = 44,
+	RB_R16G16_UINT = 45,
+	RB_R16G16B16A16_UINT = 47,
 	RB_R32G32B32A32_FLOAT = 51,
-};
-
-enum a3xx_color_swap {
-	WZYX = 0,
-	WXYZ = 1,
-	ZYXW = 2,
-	XYZW = 3,
+	RB_R32_SINT = 52,
+	RB_R32G32_SINT = 53,
+	RB_R32G32B32A32_SINT = 55,
+	RB_R32_UINT = 56,
+	RB_R32G32_UINT = 57,
+	RB_R32G32B32A32_UINT = 59,
 };
 
 enum a3xx_sp_perfcounter_select {
@@ -194,6 +239,11 @@ enum a3xx_rb_blend_opcode {
 	BLEND_MAX_DST_SRC = 4,
 };
 
+enum a3xx_intp_mode {
+	SMOOTH = 0,
+	FLAT = 1,
+};
+
 enum a3xx_tex_filter {
 	A3XX_TEX_NEAREST = 0,
 	A3XX_TEX_LINEAR = 1,
@@ -536,6 +586,10 @@ enum a3xx_tex_type {
 
 #define REG_A3XX_CP_MEQ_DATA					0x000001db
 
+#define REG_A3XX_CP_WFI_PEND_CTR				0x000001f5
+
+#define REG_A3XX_RBBM_PM_OVERRIDE2				0x0000039d
+
 #define REG_A3XX_CP_PERFCOUNTER_SELECT				0x00000445
 
 #define REG_A3XX_CP_HW_FAULT					0x0000045c
@@ -550,6 +604,12 @@ static inline uint32_t REG_A3XX_CP_PROTECT_REG(uint32_t i0) { return 0x00000460
 
 #define REG_A3XX_CP_AHB_FAULT					0x0000054d
 
+#define REG_A3XX_SQ_GPR_MANAGEMENT				0x00000d00
+
+#define REG_A3XX_SQ_INST_STORE_MANAGMENT			0x00000d02
+
+#define REG_A3XX_TP0_CHICKEN					0x00000e1e
+
 #define REG_A3XX_SP_GLOBAL_MEM_SIZE				0x00000e22
 
 #define REG_A3XX_SP_GLOBAL_MEM_ADDR				0x00000e23
@@ -632,13 +692,13 @@ static inline uint32_t A3XX_GRAS_CL_VPORT_ZSCALE(float val)
 #define A3XX_GRAS_SU_POINT_MINMAX_MIN__SHIFT			0
 static inline uint32_t A3XX_GRAS_SU_POINT_MINMAX_MIN(float val)
 {
-	return ((((uint32_t)(val * 8.0))) << A3XX_GRAS_SU_POINT_MINMAX_MIN__SHIFT) & A3XX_GRAS_SU_POINT_MINMAX_MIN__MASK;
+	return ((((uint32_t)(val * 16.0))) << A3XX_GRAS_SU_POINT_MINMAX_MIN__SHIFT) & A3XX_GRAS_SU_POINT_MINMAX_MIN__MASK;
 }
 #define A3XX_GRAS_SU_POINT_MINMAX_MAX__MASK			0xffff0000
 #define A3XX_GRAS_SU_POINT_MINMAX_MAX__SHIFT			16
 static inline uint32_t A3XX_GRAS_SU_POINT_MINMAX_MAX(float val)
 {
-	return ((((uint32_t)(val * 8.0))) << A3XX_GRAS_SU_POINT_MINMAX_MAX__SHIFT) & A3XX_GRAS_SU_POINT_MINMAX_MAX__MASK;
+	return ((((uint32_t)(val * 16.0))) << A3XX_GRAS_SU_POINT_MINMAX_MAX__SHIFT) & A3XX_GRAS_SU_POINT_MINMAX_MAX__MASK;
 }
 
 #define REG_A3XX_GRAS_SU_POINT_SIZE				0x00002069
@@ -646,7 +706,7 @@ static inline uint32_t A3XX_GRAS_SU_POINT_MINMAX_MAX(float val)
 #define A3XX_GRAS_SU_POINT_SIZE__SHIFT				0
 static inline uint32_t A3XX_GRAS_SU_POINT_SIZE(float val)
 {
-	return ((((uint32_t)(val * 8.0))) << A3XX_GRAS_SU_POINT_SIZE__SHIFT) & A3XX_GRAS_SU_POINT_SIZE__MASK;
+	return ((((int32_t)(val * 16.0))) << A3XX_GRAS_SU_POINT_SIZE__SHIFT) & A3XX_GRAS_SU_POINT_SIZE__MASK;
 }
 
 #define REG_A3XX_GRAS_SU_POLY_OFFSET_SCALE			0x0000206c
@@ -654,7 +714,7 @@ static inline uint32_t A3XX_GRAS_SU_POINT_SIZE(float val)
 #define A3XX_GRAS_SU_POLY_OFFSET_SCALE_VAL__SHIFT		0
 static inline uint32_t A3XX_GRAS_SU_POLY_OFFSET_SCALE_VAL(float val)
 {
-	return ((((uint32_t)(val * 40.0))) << A3XX_GRAS_SU_POLY_OFFSET_SCALE_VAL__SHIFT) & A3XX_GRAS_SU_POLY_OFFSET_SCALE_VAL__MASK;
+	return ((((int32_t)(val * 16384.0))) << A3XX_GRAS_SU_POLY_OFFSET_SCALE_VAL__SHIFT) & A3XX_GRAS_SU_POLY_OFFSET_SCALE_VAL__MASK;
 }
 
 #define REG_A3XX_GRAS_SU_POLY_OFFSET_OFFSET			0x0000206d
@@ -662,7 +722,7 @@ static inline uint32_t A3XX_GRAS_SU_POLY_OFFSET_SCALE_VAL(float val)
 #define A3XX_GRAS_SU_POLY_OFFSET_OFFSET__SHIFT			0
 static inline uint32_t A3XX_GRAS_SU_POLY_OFFSET_OFFSET(float val)
 {
-	return ((((uint32_t)(val * 44.0))) << A3XX_GRAS_SU_POLY_OFFSET_OFFSET__SHIFT) & A3XX_GRAS_SU_POLY_OFFSET_OFFSET__MASK;
+	return ((((int32_t)(val * 16384.0))) << A3XX_GRAS_SU_POLY_OFFSET_OFFSET__SHIFT) & A3XX_GRAS_SU_POLY_OFFSET_OFFSET__MASK;
 }
 
 #define REG_A3XX_GRAS_SU_MODE_CONTROL				0x00002070
@@ -673,7 +733,7 @@ static inline uint32_t A3XX_GRAS_SU_POLY_OFFSET_OFFSET(float val)
 #define A3XX_GRAS_SU_MODE_CONTROL_LINEHALFWIDTH__SHIFT		3
 static inline uint32_t A3XX_GRAS_SU_MODE_CONTROL_LINEHALFWIDTH(float val)
 {
-	return ((((uint32_t)(val * 4.0))) << A3XX_GRAS_SU_MODE_CONTROL_LINEHALFWIDTH__SHIFT) & A3XX_GRAS_SU_MODE_CONTROL_LINEHALFWIDTH__MASK;
+	return ((((int32_t)(val * 4.0))) << A3XX_GRAS_SU_MODE_CONTROL_LINEHALFWIDTH__SHIFT) & A3XX_GRAS_SU_MODE_CONTROL_LINEHALFWIDTH__MASK;
 }
 #define A3XX_GRAS_SU_MODE_CONTROL_POLY_OFFSET			0x00000800
 
@@ -863,6 +923,7 @@ static inline uint32_t A3XX_RB_MRT_BUF_INFO_COLOR_SWAP(enum a3xx_color_swap val)
 {
 	return ((val) << A3XX_RB_MRT_BUF_INFO_COLOR_SWAP__SHIFT) & A3XX_RB_MRT_BUF_INFO_COLOR_SWAP__MASK;
 }
+#define A3XX_RB_MRT_BUF_INFO_COLOR_SRGB				0x00004000
 #define A3XX_RB_MRT_BUF_INFO_COLOR_BUF_PITCH__MASK		0xfffe0000
 #define A3XX_RB_MRT_BUF_INFO_COLOR_BUF_PITCH__SHIFT		17
 static inline uint32_t A3XX_RB_MRT_BUF_INFO_COLOR_BUF_PITCH(uint32_t val)
@@ -1001,6 +1062,7 @@ static inline uint32_t A3XX_RB_COPY_CONTROL_FASTCLEAR(uint32_t val)
 {
 	return ((val) << A3XX_RB_COPY_CONTROL_FASTCLEAR__SHIFT) & A3XX_RB_COPY_CONTROL_FASTCLEAR__MASK;
 }
+#define A3XX_RB_COPY_CONTROL_UNK12				0x00001000
 #define A3XX_RB_COPY_CONTROL_GMEM_BASE__MASK			0xffffc000
 #define A3XX_RB_COPY_CONTROL_GMEM_BASE__SHIFT			14
 static inline uint32_t A3XX_RB_COPY_CONTROL_GMEM_BASE(uint32_t val)
@@ -1079,7 +1141,7 @@ static inline uint32_t A3XX_RB_DEPTH_CONTROL_ZFUNC(enum adreno_compare_func val)
 #define REG_A3XX_RB_DEPTH_CLEAR					0x00002101
 
 #define REG_A3XX_RB_DEPTH_INFO					0x00002102
-#define A3XX_RB_DEPTH_INFO_DEPTH_FORMAT__MASK			0x00000001
+#define A3XX_RB_DEPTH_INFO_DEPTH_FORMAT__MASK			0x00000003
 #define A3XX_RB_DEPTH_INFO_DEPTH_FORMAT__SHIFT			0
 static inline uint32_t A3XX_RB_DEPTH_INFO_DEPTH_FORMAT(enum adreno_rb_depth_format val)
 {
@@ -1265,6 +1327,7 @@ static inline uint32_t A3XX_PC_PRIM_VTX_CNTL_POLYMODE_BACK_PTYPE(enum adreno_pa_
 {
 	return ((val) << A3XX_PC_PRIM_VTX_CNTL_POLYMODE_BACK_PTYPE__SHIFT) & A3XX_PC_PRIM_VTX_CNTL_POLYMODE_BACK_PTYPE__MASK;
 }
+#define A3XX_PC_PRIM_VTX_CNTL_PRIMITIVE_RESTART			0x00100000
 #define A3XX_PC_PRIM_VTX_CNTL_PROVOKING_VTX_LAST		0x02000000
 #define A3XX_PC_PRIM_VTX_CNTL_PSIZE				0x04000000
 
@@ -1281,7 +1344,12 @@ static inline uint32_t A3XX_HLSQ_CONTROL_0_REG_FSTHREADSIZE(enum a3xx_threadsize
 #define A3XX_HLSQ_CONTROL_0_REG_SPSHADERRESTART			0x00000200
 #define A3XX_HLSQ_CONTROL_0_REG_RESERVED2			0x00000400
 #define A3XX_HLSQ_CONTROL_0_REG_CHUNKDISABLE			0x04000000
-#define A3XX_HLSQ_CONTROL_0_REG_CONSTSWITCHMODE			0x08000000
+#define A3XX_HLSQ_CONTROL_0_REG_CONSTMODE__MASK			0x08000000
+#define A3XX_HLSQ_CONTROL_0_REG_CONSTMODE__SHIFT		27
+static inline uint32_t A3XX_HLSQ_CONTROL_0_REG_CONSTMODE(uint32_t val)
+{
+	return ((val) << A3XX_HLSQ_CONTROL_0_REG_CONSTMODE__SHIFT) & A3XX_HLSQ_CONTROL_0_REG_CONSTMODE__MASK;
+}
 #define A3XX_HLSQ_CONTROL_0_REG_LAZYUPDATEDISABLE		0x10000000
 #define A3XX_HLSQ_CONTROL_0_REG_SPCONSTFULLUPDATE		0x20000000
 #define A3XX_HLSQ_CONTROL_0_REG_TPFULLUPDATE			0x40000000
@@ -1484,6 +1552,8 @@ static inline uint32_t A3XX_VFD_CONTROL_1_REGID4INST(uint32_t val)
 
 #define REG_A3XX_VFD_INDEX_OFFSET				0x00002245
 
+#define REG_A3XX_VFD_INDEX_OFFSET				0x00002245
+
 static inline uint32_t REG_A3XX_VFD_FETCH(uint32_t i0) { return 0x00002246 + 0x2*i0; }
 
 static inline uint32_t REG_A3XX_VFD_FETCH_INSTR_0(uint32_t i0) { return 0x00002246 + 0x2*i0; }
@@ -1537,6 +1607,7 @@ static inline uint32_t A3XX_VFD_DECODE_INSTR_REGID(uint32_t val)
 {
 	return ((val) << A3XX_VFD_DECODE_INSTR_REGID__SHIFT) & A3XX_VFD_DECODE_INSTR_REGID__MASK;
 }
+#define A3XX_VFD_DECODE_INSTR_INT				0x00100000
 #define A3XX_VFD_DECODE_INSTR_SWAP__MASK			0x00c00000
 #define A3XX_VFD_DECODE_INSTR_SWAP__SHIFT			22
 static inline uint32_t A3XX_VFD_DECODE_INSTR_SWAP(enum a3xx_color_swap val)
@@ -1604,6 +1675,102 @@ static inline uint32_t A3XX_VPC_PACK_NUMNONPOSVSVAR(uint32_t val)
 static inline uint32_t REG_A3XX_VPC_VARYING_INTERP(uint32_t i0) { return 0x00002282 + 0x1*i0; }
 
 static inline uint32_t REG_A3XX_VPC_VARYING_INTERP_MODE(uint32_t i0) { return 0x00002282 + 0x1*i0; }
+#define A3XX_VPC_VARYING_INTERP_MODE_C0__MASK			0x00000003
+#define A3XX_VPC_VARYING_INTERP_MODE_C0__SHIFT			0
+static inline uint32_t A3XX_VPC_VARYING_INTERP_MODE_C0(enum a3xx_intp_mode val)
+{
+	return ((val) << A3XX_VPC_VARYING_INTERP_MODE_C0__SHIFT) & A3XX_VPC_VARYING_INTERP_MODE_C0__MASK;
+}
+#define A3XX_VPC_VARYING_INTERP_MODE_C1__MASK			0x0000000c
+#define A3XX_VPC_VARYING_INTERP_MODE_C1__SHIFT			2
+static inline uint32_t A3XX_VPC_VARYING_INTERP_MODE_C1(enum a3xx_intp_mode val)
+{
+	return ((val) << A3XX_VPC_VARYING_INTERP_MODE_C1__SHIFT) & A3XX_VPC_VARYING_INTERP_MODE_C1__MASK;
+}
+#define A3XX_VPC_VARYING_INTERP_MODE_C2__MASK			0x00000030
+#define A3XX_VPC_VARYING_INTERP_MODE_C2__SHIFT			4
+static inline uint32_t A3XX_VPC_VARYING_INTERP_MODE_C2(enum a3xx_intp_mode val)
+{
+	return ((val) << A3XX_VPC_VARYING_INTERP_MODE_C2__SHIFT) & A3XX_VPC_VARYING_INTERP_MODE_C2__MASK;
+}
+#define A3XX_VPC_VARYING_INTERP_MODE_C3__MASK			0x000000c0
+#define A3XX_VPC_VARYING_INTERP_MODE_C3__SHIFT			6
+static inline uint32_t A3XX_VPC_VARYING_INTERP_MODE_C3(enum a3xx_intp_mode val)
+{
+	return ((val) << A3XX_VPC_VARYING_INTERP_MODE_C3__SHIFT) & A3XX_VPC_VARYING_INTERP_MODE_C3__MASK;
+}
+#define A3XX_VPC_VARYING_INTERP_MODE_C4__MASK			0x00000300
+#define A3XX_VPC_VARYING_INTERP_MODE_C4__SHIFT			8
+static inline uint32_t A3XX_VPC_VARYING_INTERP_MODE_C4(enum a3xx_intp_mode val)
+{
+	return ((val) << A3XX_VPC_VARYING_INTERP_MODE_C4__SHIFT) & A3XX_VPC_VARYING_INTERP_MODE_C4__MASK;
+}
+#define A3XX_VPC_VARYING_INTERP_MODE_C5__MASK			0x00000c00
+#define A3XX_VPC_VARYING_INTERP_MODE_C5__SHIFT			10
+static inline uint32_t A3XX_VPC_VARYING_INTERP_MODE_C5(enum a3xx_intp_mode val)
+{
+	return ((val) << A3XX_VPC_VARYING_INTERP_MODE_C5__SHIFT) & A3XX_VPC_VARYING_INTERP_MODE_C5__MASK;
+}
+#define A3XX_VPC_VARYING_INTERP_MODE_C6__MASK			0x00003000
+#define A3XX_VPC_VARYING_INTERP_MODE_C6__SHIFT			12
+static inline uint32_t A3XX_VPC_VARYING_INTERP_MODE_C6(enum a3xx_intp_mode val)
+{
+	return ((val) << A3XX_VPC_VARYING_INTERP_MODE_C6__SHIFT) & A3XX_VPC_VARYING_INTERP_MODE_C6__MASK;
+}
+#define A3XX_VPC_VARYING_INTERP_MODE_C7__MASK			0x0000c000
+#define A3XX_VPC_VARYING_INTERP_MODE_C7__SHIFT			14
+static inline uint32_t A3XX_VPC_VARYING_INTERP_MODE_C7(enum a3xx_intp_mode val)
+{
+	return ((val) << A3XX_VPC_VARYING_INTERP_MODE_C7__SHIFT) & A3XX_VPC_VARYING_INTERP_MODE_C7__MASK;
+}
+#define A3XX_VPC_VARYING_INTERP_MODE_C8__MASK			0x00030000
+#define A3XX_VPC_VARYING_INTERP_MODE_C8__SHIFT			16
+static inline uint32_t A3XX_VPC_VARYING_INTERP_MODE_C8(enum a3xx_intp_mode val)
+{
+	return ((val) << A3XX_VPC_VARYING_INTERP_MODE_C8__SHIFT) & A3XX_VPC_VARYING_INTERP_MODE_C8__MASK;
+}
+#define A3XX_VPC_VARYING_INTERP_MODE_C9__MASK			0x000c0000
+#define A3XX_VPC_VARYING_INTERP_MODE_C9__SHIFT			18
+static inline uint32_t A3XX_VPC_VARYING_INTERP_MODE_C9(enum a3xx_intp_mode val)
+{
+	return ((val) << A3XX_VPC_VARYING_INTERP_MODE_C9__SHIFT) & A3XX_VPC_VARYING_INTERP_MODE_C9__MASK;
+}
+#define A3XX_VPC_VARYING_INTERP_MODE_CA__MASK			0x00300000
+#define A3XX_VPC_VARYING_INTERP_MODE_CA__SHIFT			20
+static inline uint32_t A3XX_VPC_VARYING_INTERP_MODE_CA(enum a3xx_intp_mode val)
+{
+	return ((val) << A3XX_VPC_VARYING_INTERP_MODE_CA__SHIFT) & A3XX_VPC_VARYING_INTERP_MODE_CA__MASK;
+}
+#define A3XX_VPC_VARYING_INTERP_MODE_CB__MASK			0x00c00000
+#define A3XX_VPC_VARYING_INTERP_MODE_CB__SHIFT			22
+static inline uint32_t A3XX_VPC_VARYING_INTERP_MODE_CB(enum a3xx_intp_mode val)
+{
+	return ((val) << A3XX_VPC_VARYING_INTERP_MODE_CB__SHIFT) & A3XX_VPC_VARYING_INTERP_MODE_CB__MASK;
+}
+#define A3XX_VPC_VARYING_INTERP_MODE_CC__MASK			0x03000000
+#define A3XX_VPC_VARYING_INTERP_MODE_CC__SHIFT			24
+static inline uint32_t A3XX_VPC_VARYING_INTERP_MODE_CC(enum a3xx_intp_mode val)
+{
+	return ((val) << A3XX_VPC_VARYING_INTERP_MODE_CC__SHIFT) & A3XX_VPC_VARYING_INTERP_MODE_CC__MASK;
+}
+#define A3XX_VPC_VARYING_INTERP_MODE_CD__MASK			0x0c000000
+#define A3XX_VPC_VARYING_INTERP_MODE_CD__SHIFT			26
+static inline uint32_t A3XX_VPC_VARYING_INTERP_MODE_CD(enum a3xx_intp_mode val)
+{
+	return ((val) << A3XX_VPC_VARYING_INTERP_MODE_CD__SHIFT) & A3XX_VPC_VARYING_INTERP_MODE_CD__MASK;
+}
+#define A3XX_VPC_VARYING_INTERP_MODE_CE__MASK			0x30000000
+#define A3XX_VPC_VARYING_INTERP_MODE_CE__SHIFT			28
+static inline uint32_t A3XX_VPC_VARYING_INTERP_MODE_CE(enum a3xx_intp_mode val)
+{
+	return ((val) << A3XX_VPC_VARYING_INTERP_MODE_CE__SHIFT) & A3XX_VPC_VARYING_INTERP_MODE_CE__MASK;
+}
+#define A3XX_VPC_VARYING_INTERP_MODE_CF__MASK			0xc0000000
+#define A3XX_VPC_VARYING_INTERP_MODE_CF__SHIFT			30
+static inline uint32_t A3XX_VPC_VARYING_INTERP_MODE_CF(enum a3xx_intp_mode val)
+{
+	return ((val) << A3XX_VPC_VARYING_INTERP_MODE_CF__SHIFT) & A3XX_VPC_VARYING_INTERP_MODE_CF__MASK;
+}
 
 static inline uint32_t REG_A3XX_VPC_VARYING_PS_REPL(uint32_t i0) { return 0x00002286 + 0x1*i0; }
 
@@ -1696,7 +1863,7 @@ static inline uint32_t A3XX_SP_VS_CTRL_REG1_CONSTFOOTPRINT(uint32_t val)
 {
 	return ((val) << A3XX_SP_VS_CTRL_REG1_CONSTFOOTPRINT__SHIFT) & A3XX_SP_VS_CTRL_REG1_CONSTFOOTPRINT__MASK;
 }
-#define A3XX_SP_VS_CTRL_REG1_INITIALOUTSTANDING__MASK		0x3f000000
+#define A3XX_SP_VS_CTRL_REG1_INITIALOUTSTANDING__MASK		0x7f000000
 #define A3XX_SP_VS_CTRL_REG1_INITIALOUTSTANDING__SHIFT		24
 static inline uint32_t A3XX_SP_VS_CTRL_REG1_INITIALOUTSTANDING(uint32_t val)
 {
@@ -1928,6 +2095,8 @@ static inline uint32_t A3XX_SP_FS_MRT_REG_REGID(uint32_t val)
 	return ((val) << A3XX_SP_FS_MRT_REG_REGID__SHIFT) & A3XX_SP_FS_MRT_REG_REGID__MASK;
 }
 #define A3XX_SP_FS_MRT_REG_HALF_PRECISION			0x00000100
+#define A3XX_SP_FS_MRT_REG_SINT					0x00000400
+#define A3XX_SP_FS_MRT_REG_UINT					0x00000800
 
 static inline uint32_t REG_A3XX_SP_FS_IMAGE_OUTPUT(uint32_t i0) { return 0x000022f4 + 0x1*i0; }
 
@@ -1947,6 +2116,8 @@ static inline uint32_t A3XX_SP_FS_LENGTH_REG_SHADERLENGTH(uint32_t val)
 	return ((val) << A3XX_SP_FS_LENGTH_REG_SHADERLENGTH__SHIFT) & A3XX_SP_FS_LENGTH_REG_SHADERLENGTH__MASK;
 }
 
+#define REG_A3XX_PA_SC_AA_CONFIG				0x00002301
+
 #define REG_A3XX_TPL1_TP_VS_TEX_OFFSET				0x00002340
 #define A3XX_TPL1_TP_VS_TEX_OFFSET_SAMPLEROFFSET__MASK		0x000000ff
 #define A3XX_TPL1_TP_VS_TEX_OFFSET_SAMPLEROFFSET__SHIFT		0
@@ -2297,11 +2468,11 @@ static inline uint32_t A3XX_VGT_DRAW_INITIATOR_INDEX_SIZE(enum pc_di_index_size
 #define A3XX_VGT_DRAW_INITIATOR_NOT_EOP				0x00001000
 #define A3XX_VGT_DRAW_INITIATOR_SMALL_INDEX			0x00002000
 #define A3XX_VGT_DRAW_INITIATOR_PRE_DRAW_INITIATOR_ENABLE	0x00004000
-#define A3XX_VGT_DRAW_INITIATOR_NUM_INDICES__MASK		0xffff0000
-#define A3XX_VGT_DRAW_INITIATOR_NUM_INDICES__SHIFT		16
-static inline uint32_t A3XX_VGT_DRAW_INITIATOR_NUM_INDICES(uint32_t val)
+#define A3XX_VGT_DRAW_INITIATOR_NUM_INSTANCES__MASK		0xff000000
+#define A3XX_VGT_DRAW_INITIATOR_NUM_INSTANCES__SHIFT		24
+static inline uint32_t A3XX_VGT_DRAW_INITIATOR_NUM_INSTANCES(uint32_t val)
 {
-	return ((val) << A3XX_VGT_DRAW_INITIATOR_NUM_INDICES__SHIFT) & A3XX_VGT_DRAW_INITIATOR_NUM_INDICES__MASK;
+	return ((val) << A3XX_VGT_DRAW_INITIATOR_NUM_INSTANCES__SHIFT) & A3XX_VGT_DRAW_INITIATOR_NUM_INSTANCES__MASK;
 }
 
 #define REG_A3XX_VGT_IMMED_DATA					0x000021fd
@@ -2347,17 +2518,23 @@ static inline uint32_t A3XX_TEX_SAMP_0_COMPARE_FUNC(enum adreno_compare_func val
 #define A3XX_TEX_SAMP_0_UNNORM_COORDS				0x80000000
 
 #define REG_A3XX_TEX_SAMP_1					0x00000001
+#define A3XX_TEX_SAMP_1_LOD_BIAS__MASK				0x000007ff
+#define A3XX_TEX_SAMP_1_LOD_BIAS__SHIFT				0
+static inline uint32_t A3XX_TEX_SAMP_1_LOD_BIAS(float val)
+{
+	return ((((int32_t)(val * 64.0))) << A3XX_TEX_SAMP_1_LOD_BIAS__SHIFT) & A3XX_TEX_SAMP_1_LOD_BIAS__MASK;
+}
 #define A3XX_TEX_SAMP_1_MAX_LOD__MASK				0x003ff000
 #define A3XX_TEX_SAMP_1_MAX_LOD__SHIFT				12
 static inline uint32_t A3XX_TEX_SAMP_1_MAX_LOD(float val)
 {
-	return ((((uint32_t)(val * 12.0))) << A3XX_TEX_SAMP_1_MAX_LOD__SHIFT) & A3XX_TEX_SAMP_1_MAX_LOD__MASK;
+	return ((((uint32_t)(val * 64.0))) << A3XX_TEX_SAMP_1_MAX_LOD__SHIFT) & A3XX_TEX_SAMP_1_MAX_LOD__MASK;
 }
 #define A3XX_TEX_SAMP_1_MIN_LOD__MASK				0xffc00000
 #define A3XX_TEX_SAMP_1_MIN_LOD__SHIFT				22
 static inline uint32_t A3XX_TEX_SAMP_1_MIN_LOD(float val)
 {
-	return ((((uint32_t)(val * 12.0))) << A3XX_TEX_SAMP_1_MIN_LOD__SHIFT) & A3XX_TEX_SAMP_1_MIN_LOD__MASK;
+	return ((((uint32_t)(val * 64.0))) << A3XX_TEX_SAMP_1_MIN_LOD__SHIFT) & A3XX_TEX_SAMP_1_MIN_LOD__MASK;
 }
 
 #define REG_A3XX_TEX_CONST_0					0x00000000
@@ -2448,6 +2625,24 @@ static inline uint32_t A3XX_TEX_CONST_2_SWAP(enum a3xx_color_swap val)
 }
 
 #define REG_A3XX_TEX_CONST_3					0x00000003
+#define A3XX_TEX_CONST_3_LAYERSZ1__MASK				0x0000000f
+#define A3XX_TEX_CONST_3_LAYERSZ1__SHIFT			0
+static inline uint32_t A3XX_TEX_CONST_3_LAYERSZ1(uint32_t val)
+{
+	return ((val >> 12) << A3XX_TEX_CONST_3_LAYERSZ1__SHIFT) & A3XX_TEX_CONST_3_LAYERSZ1__MASK;
+}
+#define A3XX_TEX_CONST_3_DEPTH__MASK				0x0ffe0000
+#define A3XX_TEX_CONST_3_DEPTH__SHIFT				17
+static inline uint32_t A3XX_TEX_CONST_3_DEPTH(uint32_t val)
+{
+	return ((val) << A3XX_TEX_CONST_3_DEPTH__SHIFT) & A3XX_TEX_CONST_3_DEPTH__MASK;
+}
+#define A3XX_TEX_CONST_3_LAYERSZ2__MASK				0xf0000000
+#define A3XX_TEX_CONST_3_LAYERSZ2__SHIFT			28
+static inline uint32_t A3XX_TEX_CONST_3_LAYERSZ2(uint32_t val)
+{
+	return ((val >> 12) << A3XX_TEX_CONST_3_LAYERSZ2__SHIFT) & A3XX_TEX_CONST_3_LAYERSZ2__MASK;
+}
 
 
 #endif /* A3XX_XML */
diff --git a/drivers/gpu/drm/msm/adreno/a3xx_gpu.c b/drivers/gpu/drm/msm/adreno/a3xx_gpu.c
index 2773600c9488..b66c53bdc039 100644
--- a/drivers/gpu/drm/msm/adreno/a3xx_gpu.c
+++ b/drivers/gpu/drm/msm/adreno/a3xx_gpu.c
@@ -2,6 +2,8 @@
  * Copyright (C) 2013 Red Hat
  * Author: Rob Clark <robdclark@gmail.com>
  *
+ * Copyright (c) 2014 The Linux Foundation. All rights reserved.
+ *
  * This program is free software; you can redistribute it and/or modify it
  * under the terms of the GNU General Public License version 2 as published by
  * the Free Software Foundation.
@@ -35,10 +37,8 @@
 	 A3XX_INT0_CP_AHB_ERROR_HALT |     \
 	 A3XX_INT0_UCHE_OOB_ACCESS)
 
+extern bool hang_debug;
 
-static bool hang_debug = false;
-MODULE_PARM_DESC(hang_debug, "Dump registers when hang is detected (can be slow!)");
-module_param_named(hang_debug, hang_debug, bool, 0600);
 static void a3xx_dump(struct msm_gpu *gpu);
 
 static void a3xx_me_init(struct msm_gpu *gpu)
@@ -387,59 +387,115 @@ static const unsigned int a3xx_registers[] = {
 	0x2750, 0x2756, 0x2760, 0x2760, 0x300c, 0x300e, 0x301c, 0x301d,
 	0x302a, 0x302a, 0x302c, 0x302d, 0x3030, 0x3031, 0x3034, 0x3036,
 	0x303c, 0x303c, 0x305e, 0x305f,
+	~0   /* sentinel */
 };
 
 #ifdef CONFIG_DEBUG_FS
 static void a3xx_show(struct msm_gpu *gpu, struct seq_file *m)
 {
-	int i;
-
-	adreno_show(gpu, m);
-
 	gpu->funcs->pm_resume(gpu);
-
 	seq_printf(m, "status:   %08x\n",
 			gpu_read(gpu, REG_A3XX_RBBM_STATUS));
-
-	/* dump these out in a form that can be parsed by demsm: */
-	seq_printf(m, "IO:region %s 00000000 00020000\n", gpu->name);
-	for (i = 0; i < ARRAY_SIZE(a3xx_registers); i += 2) {
-		uint32_t start = a3xx_registers[i];
-		uint32_t end   = a3xx_registers[i+1];
-		uint32_t addr;
-
-		for (addr = start; addr <= end; addr++) {
-			uint32_t val = gpu_read(gpu, addr);
-			seq_printf(m, "IO:R %08x %08x\n", addr<<2, val);
-		}
-	}
-
 	gpu->funcs->pm_suspend(gpu);
+	adreno_show(gpu, m);
 }
 #endif
 
 /* would be nice to not have to duplicate the _show() stuff with printk(): */
 static void a3xx_dump(struct msm_gpu *gpu)
 {
-	int i;
-
-	adreno_dump(gpu);
 	printk("status:   %08x\n",
 			gpu_read(gpu, REG_A3XX_RBBM_STATUS));
-
-	/* dump these out in a form that can be parsed by demsm: */
-	printk("IO:region %s 00000000 00020000\n", gpu->name);
-	for (i = 0; i < ARRAY_SIZE(a3xx_registers); i += 2) {
-		uint32_t start = a3xx_registers[i];
-		uint32_t end   = a3xx_registers[i+1];
-		uint32_t addr;
-
-		for (addr = start; addr <= end; addr++) {
-			uint32_t val = gpu_read(gpu, addr);
-			printk("IO:R %08x %08x\n", addr<<2, val);
-		}
-	}
+	adreno_dump(gpu);
 }
+/* Register offset defines for A3XX */
+static const unsigned int a3xx_register_offsets[REG_ADRENO_REGISTER_MAX] = {
+	REG_ADRENO_DEFINE(REG_ADRENO_CP_DEBUG, REG_AXXX_CP_DEBUG),
+	REG_ADRENO_DEFINE(REG_ADRENO_CP_ME_RAM_WADDR, REG_AXXX_CP_ME_RAM_WADDR),
+	REG_ADRENO_DEFINE(REG_ADRENO_CP_ME_RAM_DATA, REG_AXXX_CP_ME_RAM_DATA),
+	REG_ADRENO_DEFINE(REG_ADRENO_CP_PFP_UCODE_DATA,
+			REG_A3XX_CP_PFP_UCODE_DATA),
+	REG_ADRENO_DEFINE(REG_ADRENO_CP_PFP_UCODE_ADDR,
+			REG_A3XX_CP_PFP_UCODE_ADDR),
+	REG_ADRENO_DEFINE(REG_ADRENO_CP_WFI_PEND_CTR, REG_A3XX_CP_WFI_PEND_CTR),
+	REG_ADRENO_DEFINE(REG_ADRENO_CP_RB_BASE, REG_AXXX_CP_RB_BASE),
+	REG_ADRENO_DEFINE(REG_ADRENO_CP_RB_RPTR_ADDR, REG_AXXX_CP_RB_RPTR_ADDR),
+	REG_ADRENO_DEFINE(REG_ADRENO_CP_RB_RPTR, REG_AXXX_CP_RB_RPTR),
+	REG_ADRENO_DEFINE(REG_ADRENO_CP_RB_WPTR, REG_AXXX_CP_RB_WPTR),
+	REG_ADRENO_DEFINE(REG_ADRENO_CP_PROTECT_CTRL, REG_A3XX_CP_PROTECT_CTRL),
+	REG_ADRENO_DEFINE(REG_ADRENO_CP_ME_CNTL, REG_AXXX_CP_ME_CNTL),
+	REG_ADRENO_DEFINE(REG_ADRENO_CP_RB_CNTL, REG_AXXX_CP_RB_CNTL),
+	REG_ADRENO_DEFINE(REG_ADRENO_CP_IB1_BASE, REG_AXXX_CP_IB1_BASE),
+	REG_ADRENO_DEFINE(REG_ADRENO_CP_IB1_BUFSZ, REG_AXXX_CP_IB1_BUFSZ),
+	REG_ADRENO_DEFINE(REG_ADRENO_CP_IB2_BASE, REG_AXXX_CP_IB2_BASE),
+	REG_ADRENO_DEFINE(REG_ADRENO_CP_IB2_BUFSZ, REG_AXXX_CP_IB2_BUFSZ),
+	REG_ADRENO_DEFINE(REG_ADRENO_CP_TIMESTAMP, REG_AXXX_CP_SCRATCH_REG0),
+	REG_ADRENO_DEFINE(REG_ADRENO_CP_ME_RAM_RADDR, REG_AXXX_CP_ME_RAM_RADDR),
+	REG_ADRENO_DEFINE(REG_ADRENO_SCRATCH_ADDR, REG_AXXX_SCRATCH_ADDR),
+	REG_ADRENO_DEFINE(REG_ADRENO_SCRATCH_UMSK, REG_AXXX_SCRATCH_UMSK),
+	REG_ADRENO_DEFINE(REG_ADRENO_CP_ROQ_ADDR, REG_A3XX_CP_ROQ_ADDR),
+	REG_ADRENO_DEFINE(REG_ADRENO_CP_ROQ_DATA, REG_A3XX_CP_ROQ_DATA),
+	REG_ADRENO_DEFINE(REG_ADRENO_CP_MERCIU_ADDR, REG_A3XX_CP_MERCIU_ADDR),
+	REG_ADRENO_DEFINE(REG_ADRENO_CP_MERCIU_DATA, REG_A3XX_CP_MERCIU_DATA),
+	REG_ADRENO_DEFINE(REG_ADRENO_CP_MERCIU_DATA2, REG_A3XX_CP_MERCIU_DATA2),
+	REG_ADRENO_DEFINE(REG_ADRENO_CP_MEQ_ADDR, REG_A3XX_CP_MEQ_ADDR),
+	REG_ADRENO_DEFINE(REG_ADRENO_CP_MEQ_DATA, REG_A3XX_CP_MEQ_DATA),
+	REG_ADRENO_DEFINE(REG_ADRENO_CP_HW_FAULT, REG_A3XX_CP_HW_FAULT),
+	REG_ADRENO_DEFINE(REG_ADRENO_CP_PROTECT_STATUS,
+			REG_A3XX_CP_PROTECT_STATUS),
+	REG_ADRENO_DEFINE(REG_ADRENO_RBBM_STATUS, REG_A3XX_RBBM_STATUS),
+	REG_ADRENO_DEFINE(REG_ADRENO_RBBM_PERFCTR_CTL,
+			REG_A3XX_RBBM_PERFCTR_CTL),
+	REG_ADRENO_DEFINE(REG_ADRENO_RBBM_PERFCTR_LOAD_CMD0,
+			REG_A3XX_RBBM_PERFCTR_LOAD_CMD0),
+	REG_ADRENO_DEFINE(REG_ADRENO_RBBM_PERFCTR_LOAD_CMD1,
+			REG_A3XX_RBBM_PERFCTR_LOAD_CMD1),
+	REG_ADRENO_DEFINE(REG_ADRENO_RBBM_PERFCTR_PWR_1_LO,
+			REG_A3XX_RBBM_PERFCTR_PWR_1_LO),
+	REG_ADRENO_DEFINE(REG_ADRENO_RBBM_INT_0_MASK, REG_A3XX_RBBM_INT_0_MASK),
+	REG_ADRENO_DEFINE(REG_ADRENO_RBBM_INT_0_STATUS,
+			REG_A3XX_RBBM_INT_0_STATUS),
+	REG_ADRENO_DEFINE(REG_ADRENO_RBBM_AHB_ERROR_STATUS,
+			REG_A3XX_RBBM_AHB_ERROR_STATUS),
+	REG_ADRENO_DEFINE(REG_ADRENO_RBBM_AHB_CMD, REG_A3XX_RBBM_AHB_CMD),
+	REG_ADRENO_DEFINE(REG_ADRENO_RBBM_INT_CLEAR_CMD,
+			REG_A3XX_RBBM_INT_CLEAR_CMD),
+	REG_ADRENO_DEFINE(REG_ADRENO_RBBM_CLOCK_CTL, REG_A3XX_RBBM_CLOCK_CTL),
+	REG_ADRENO_DEFINE(REG_ADRENO_VPC_DEBUG_RAM_SEL,
+			REG_A3XX_VPC_VPC_DEBUG_RAM_SEL),
+	REG_ADRENO_DEFINE(REG_ADRENO_VPC_DEBUG_RAM_READ,
+			REG_A3XX_VPC_VPC_DEBUG_RAM_READ),
+	REG_ADRENO_DEFINE(REG_ADRENO_VSC_SIZE_ADDRESS,
+			REG_A3XX_VSC_SIZE_ADDRESS),
+	REG_ADRENO_DEFINE(REG_ADRENO_VFD_CONTROL_0, REG_A3XX_VFD_CONTROL_0),
+	REG_ADRENO_DEFINE(REG_ADRENO_VFD_INDEX_MAX, REG_A3XX_VFD_INDEX_MAX),
+	REG_ADRENO_DEFINE(REG_ADRENO_SP_VS_PVT_MEM_ADDR_REG,
+			REG_A3XX_SP_VS_PVT_MEM_ADDR_REG),
+	REG_ADRENO_DEFINE(REG_ADRENO_SP_FS_PVT_MEM_ADDR_REG,
+			REG_A3XX_SP_FS_PVT_MEM_ADDR_REG),
+	REG_ADRENO_DEFINE(REG_ADRENO_SP_VS_OBJ_START_REG,
+			REG_A3XX_SP_VS_OBJ_START_REG),
+	REG_ADRENO_DEFINE(REG_ADRENO_SP_FS_OBJ_START_REG,
+			REG_A3XX_SP_FS_OBJ_START_REG),
+	REG_ADRENO_DEFINE(REG_ADRENO_PA_SC_AA_CONFIG, REG_A3XX_PA_SC_AA_CONFIG),
+	REG_ADRENO_DEFINE(REG_ADRENO_RBBM_PM_OVERRIDE2,
+			REG_A3XX_RBBM_PM_OVERRIDE2),
+	REG_ADRENO_DEFINE(REG_ADRENO_SCRATCH_REG2, REG_AXXX_CP_SCRATCH_REG2),
+	REG_ADRENO_DEFINE(REG_ADRENO_SQ_GPR_MANAGEMENT,
+			REG_A3XX_SQ_GPR_MANAGEMENT),
+	REG_ADRENO_DEFINE(REG_ADRENO_SQ_INST_STORE_MANAGMENT,
+			REG_A3XX_SQ_INST_STORE_MANAGMENT),
+	REG_ADRENO_DEFINE(REG_ADRENO_TP0_CHICKEN, REG_A3XX_TP0_CHICKEN),
+	REG_ADRENO_DEFINE(REG_ADRENO_RBBM_RBBM_CTL, REG_A3XX_RBBM_RBBM_CTL),
+	REG_ADRENO_DEFINE(REG_ADRENO_RBBM_SW_RESET_CMD,
+			REG_A3XX_RBBM_SW_RESET_CMD),
+	REG_ADRENO_DEFINE(REG_ADRENO_UCHE_INVALIDATE0,
+			REG_A3XX_UCHE_CACHE_INVALIDATE0_REG),
+	REG_ADRENO_DEFINE(REG_ADRENO_RBBM_PERFCTR_LOAD_VALUE_LO,
+			REG_A3XX_RBBM_PERFCTR_LOAD_VALUE_LO),
+	REG_ADRENO_DEFINE(REG_ADRENO_RBBM_PERFCTR_LOAD_VALUE_HI,
+			REG_A3XX_RBBM_PERFCTR_LOAD_VALUE_HI),
+};
 
 static const struct adreno_gpu_funcs funcs = {
 	.base = {
@@ -474,7 +530,6 @@ struct msm_gpu *a3xx_gpu_init(struct drm_device *dev)
 	struct msm_gpu *gpu;
 	struct msm_drm_private *priv = dev->dev_private;
 	struct platform_device *pdev = priv->gpu_pdev;
-	struct adreno_platform_config *config;
 	int ret;
 
 	if (!pdev) {
@@ -483,8 +538,6 @@ struct msm_gpu *a3xx_gpu_init(struct drm_device *dev)
 		goto fail;
 	}
 
-	config = pdev->dev.platform_data;
-
 	a3xx_gpu = kzalloc(sizeof(*a3xx_gpu), GFP_KERNEL);
 	if (!a3xx_gpu) {
 		ret = -ENOMEM;
@@ -496,20 +549,13 @@ struct msm_gpu *a3xx_gpu_init(struct drm_device *dev)
 
 	a3xx_gpu->pdev = pdev;
 
-	gpu->fast_rate = config->fast_rate;
-	gpu->slow_rate = config->slow_rate;
-	gpu->bus_freq  = config->bus_freq;
-#ifdef CONFIG_MSM_BUS_SCALING
-	gpu->bus_scale_table = config->bus_scale_table;
-#endif
-
-	DBG("fast_rate=%u, slow_rate=%u, bus_freq=%u",
-			gpu->fast_rate, gpu->slow_rate, gpu->bus_freq);
-
 	gpu->perfcntrs = perfcntrs;
 	gpu->num_perfcntrs = ARRAY_SIZE(perfcntrs);
 
-	ret = adreno_gpu_init(dev, pdev, adreno_gpu, &funcs, config->rev);
+	adreno_gpu->registers = a3xx_registers;
+	adreno_gpu->reg_offsets = a3xx_register_offsets;
+
+	ret = adreno_gpu_init(dev, pdev, adreno_gpu, &funcs);
 	if (ret)
 		goto fail;
 
@@ -549,158 +595,3 @@ fail:
 
 	return ERR_PTR(ret);
 }
-
-/*
- * The a3xx device:
- */
-
-#if defined(CONFIG_MSM_BUS_SCALING) && !defined(CONFIG_OF)
-#  include <mach/kgsl.h>
-#endif
-
-static void set_gpu_pdev(struct drm_device *dev,
-		struct platform_device *pdev)
-{
-	struct msm_drm_private *priv = dev->dev_private;
-	priv->gpu_pdev = pdev;
-}
-
-static int a3xx_bind(struct device *dev, struct device *master, void *data)
-{
-	static struct adreno_platform_config config = {};
-#ifdef CONFIG_OF
-	struct device_node *child, *node = dev->of_node;
-	u32 val;
-	int ret;
-
-	ret = of_property_read_u32(node, "qcom,chipid", &val);
-	if (ret) {
-		dev_err(dev, "could not find chipid: %d\n", ret);
-		return ret;
-	}
-
-	config.rev = ADRENO_REV((val >> 24) & 0xff,
-			(val >> 16) & 0xff, (val >> 8) & 0xff, val & 0xff);
-
-	/* find clock rates: */
-	config.fast_rate = 0;
-	config.slow_rate = ~0;
-	for_each_child_of_node(node, child) {
-		if (of_device_is_compatible(child, "qcom,gpu-pwrlevels")) {
-			struct device_node *pwrlvl;
-			for_each_child_of_node(child, pwrlvl) {
-				ret = of_property_read_u32(pwrlvl, "qcom,gpu-freq", &val);
-				if (ret) {
-					dev_err(dev, "could not find gpu-freq: %d\n", ret);
-					return ret;
-				}
-				config.fast_rate = max(config.fast_rate, val);
-				config.slow_rate = min(config.slow_rate, val);
-			}
-		}
-	}
-
-	if (!config.fast_rate) {
-		dev_err(dev, "could not find clk rates\n");
-		return -ENXIO;
-	}
-
-#else
-	struct kgsl_device_platform_data *pdata = dev->platform_data;
-	uint32_t version = socinfo_get_version();
-	if (cpu_is_apq8064ab()) {
-		config.fast_rate = 450000000;
-		config.slow_rate = 27000000;
-		config.bus_freq  = 4;
-		config.rev = ADRENO_REV(3, 2, 1, 0);
-	} else if (cpu_is_apq8064()) {
-		config.fast_rate = 400000000;
-		config.slow_rate = 27000000;
-		config.bus_freq  = 4;
-
-		if (SOCINFO_VERSION_MAJOR(version) == 2)
-			config.rev = ADRENO_REV(3, 2, 0, 2);
-		else if ((SOCINFO_VERSION_MAJOR(version) == 1) &&
-				(SOCINFO_VERSION_MINOR(version) == 1))
-			config.rev = ADRENO_REV(3, 2, 0, 1);
-		else
-			config.rev = ADRENO_REV(3, 2, 0, 0);
-
-	} else if (cpu_is_msm8960ab()) {
-		config.fast_rate = 400000000;
-		config.slow_rate = 320000000;
-		config.bus_freq  = 4;
-
-		if (SOCINFO_VERSION_MINOR(version) == 0)
-			config.rev = ADRENO_REV(3, 2, 1, 0);
-		else
-			config.rev = ADRENO_REV(3, 2, 1, 1);
-
-	} else if (cpu_is_msm8930()) {
-		config.fast_rate = 400000000;
-		config.slow_rate = 27000000;
-		config.bus_freq  = 3;
-
-		if ((SOCINFO_VERSION_MAJOR(version) == 1) &&
-			(SOCINFO_VERSION_MINOR(version) == 2))
-			config.rev = ADRENO_REV(3, 0, 5, 2);
-		else
-			config.rev = ADRENO_REV(3, 0, 5, 0);
-
-	}
-#  ifdef CONFIG_MSM_BUS_SCALING
-	config.bus_scale_table = pdata->bus_scale_table;
-#  endif
-#endif
-	dev->platform_data = &config;
-	set_gpu_pdev(dev_get_drvdata(master), to_platform_device(dev));
-	return 0;
-}
-
-static void a3xx_unbind(struct device *dev, struct device *master,
-		void *data)
-{
-	set_gpu_pdev(dev_get_drvdata(master), NULL);
-}
-
-static const struct component_ops a3xx_ops = {
-		.bind   = a3xx_bind,
-		.unbind = a3xx_unbind,
-};
-
-static int a3xx_probe(struct platform_device *pdev)
-{
-	return component_add(&pdev->dev, &a3xx_ops);
-}
-
-static int a3xx_remove(struct platform_device *pdev)
-{
-	component_del(&pdev->dev, &a3xx_ops);
-	return 0;
-}
-
-static const struct of_device_id dt_match[] = {
-	{ .compatible = "qcom,adreno-3xx" },
-	/* for backwards compat w/ downstream kgsl DT files: */
-	{ .compatible = "qcom,kgsl-3d0" },
-	{}
-};
-
-static struct platform_driver a3xx_driver = {
-	.probe = a3xx_probe,
-	.remove = a3xx_remove,
-	.driver = {
-		.name = "kgsl-3d0",
-		.of_match_table = dt_match,
-	},
-};
-
-void __init a3xx_register(void)
-{
-	platform_driver_register(&a3xx_driver);
-}
-
-void __exit a3xx_unregister(void)
-{
-	platform_driver_unregister(&a3xx_driver);
-}
diff --git a/drivers/gpu/drm/msm/adreno/a4xx.xml.h b/drivers/gpu/drm/msm/adreno/a4xx.xml.h
new file mode 100644
index 000000000000..5a24c416d2dd
--- /dev/null
+++ b/drivers/gpu/drm/msm/adreno/a4xx.xml.h
@@ -0,0 +1,2144 @@
+#ifndef A4XX_XML
+#define A4XX_XML
+
+/* Autogenerated file, DO NOT EDIT manually!
+
+This file was generated by the rules-ng-ng headergen tool in this git repository:
+http://github.com/freedreno/envytools/
+git clone https://github.com/freedreno/envytools.git
+
+The rules-ng-ng source files this header was generated from are:
+- /home/robclark/src/freedreno/envytools/rnndb/adreno.xml               (    364 bytes, from 2013-11-30 14:47:15)
+- /home/robclark/src/freedreno/envytools/rnndb/freedreno_copyright.xml  (   1453 bytes, from 2013-03-31 16:51:27)
+- /home/robclark/src/freedreno/envytools/rnndb/adreno/a2xx.xml          (  32901 bytes, from 2014-06-02 15:21:30)
+- /home/robclark/src/freedreno/envytools/rnndb/adreno/adreno_common.xml (  10551 bytes, from 2014-11-13 22:44:30)
+- /home/robclark/src/freedreno/envytools/rnndb/adreno/adreno_pm4.xml    (  15053 bytes, from 2014-11-09 15:45:47)
+- /home/robclark/src/freedreno/envytools/rnndb/adreno/a3xx.xml          (  63169 bytes, from 2014-11-13 22:44:18)
+- /home/robclark/src/freedreno/envytools/rnndb/adreno/a4xx.xml          (  49097 bytes, from 2014-11-14 15:38:00)
+
+Copyright (C) 2013-2014 by the following authors:
+- Rob Clark <robdclark@gmail.com> (robclark)
+
+Permission is hereby granted, free of charge, to any person obtaining
+a copy of this software and associated documentation files (the
+"Software"), to deal in the Software without restriction, including
+without limitation the rights to use, copy, modify, merge, publish,
+distribute, sublicense, and/or sell copies of the Software, and to
+permit persons to whom the Software is furnished to do so, subject to
+the following conditions:
+
+The above copyright notice and this permission notice (including the
+next paragraph) shall be included in all copies or substantial
+portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
+IN NO EVENT SHALL THE COPYRIGHT OWNER(S) AND/OR ITS SUPPLIERS BE
+LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION
+OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION
+WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+*/
+
+
+enum a4xx_color_fmt {
+	RB4_A8_UNORM = 1,
+	RB4_R5G6R5_UNORM = 14,
+	RB4_Z16_UNORM = 15,
+	RB4_R8G8B8_UNORM = 25,
+	RB4_R8G8B8A8_UNORM = 26,
+};
+
+enum a4xx_tile_mode {
+	TILE4_LINEAR = 0,
+	TILE4_3 = 3,
+};
+
+enum a4xx_rb_blend_opcode {
+	BLEND_DST_PLUS_SRC = 0,
+	BLEND_SRC_MINUS_DST = 1,
+	BLEND_DST_MINUS_SRC = 2,
+	BLEND_MIN_DST_SRC = 3,
+	BLEND_MAX_DST_SRC = 4,
+};
+
+enum a4xx_vtx_fmt {
+	VFMT4_FLOAT_32 = 1,
+	VFMT4_FLOAT_32_32 = 2,
+	VFMT4_FLOAT_32_32_32 = 3,
+	VFMT4_FLOAT_32_32_32_32 = 4,
+	VFMT4_FLOAT_16 = 5,
+	VFMT4_FLOAT_16_16 = 6,
+	VFMT4_FLOAT_16_16_16 = 7,
+	VFMT4_FLOAT_16_16_16_16 = 8,
+	VFMT4_FIXED_32 = 9,
+	VFMT4_FIXED_32_32 = 10,
+	VFMT4_FIXED_32_32_32 = 11,
+	VFMT4_FIXED_32_32_32_32 = 12,
+	VFMT4_SHORT_16 = 16,
+	VFMT4_SHORT_16_16 = 17,
+	VFMT4_SHORT_16_16_16 = 18,
+	VFMT4_SHORT_16_16_16_16 = 19,
+	VFMT4_USHORT_16 = 20,
+	VFMT4_USHORT_16_16 = 21,
+	VFMT4_USHORT_16_16_16 = 22,
+	VFMT4_USHORT_16_16_16_16 = 23,
+	VFMT4_NORM_SHORT_16 = 24,
+	VFMT4_NORM_SHORT_16_16 = 25,
+	VFMT4_NORM_SHORT_16_16_16 = 26,
+	VFMT4_NORM_SHORT_16_16_16_16 = 27,
+	VFMT4_NORM_USHORT_16 = 28,
+	VFMT4_NORM_USHORT_16_16 = 29,
+	VFMT4_NORM_USHORT_16_16_16 = 30,
+	VFMT4_NORM_USHORT_16_16_16_16 = 31,
+	VFMT4_UBYTE_8 = 40,
+	VFMT4_UBYTE_8_8 = 41,
+	VFMT4_UBYTE_8_8_8 = 42,
+	VFMT4_UBYTE_8_8_8_8 = 43,
+	VFMT4_NORM_UBYTE_8 = 44,
+	VFMT4_NORM_UBYTE_8_8 = 45,
+	VFMT4_NORM_UBYTE_8_8_8 = 46,
+	VFMT4_NORM_UBYTE_8_8_8_8 = 47,
+	VFMT4_BYTE_8 = 48,
+	VFMT4_BYTE_8_8 = 49,
+	VFMT4_BYTE_8_8_8 = 50,
+	VFMT4_BYTE_8_8_8_8 = 51,
+	VFMT4_NORM_BYTE_8 = 52,
+	VFMT4_NORM_BYTE_8_8 = 53,
+	VFMT4_NORM_BYTE_8_8_8 = 54,
+	VFMT4_NORM_BYTE_8_8_8_8 = 55,
+	VFMT4_UINT_10_10_10_2 = 60,
+	VFMT4_NORM_UINT_10_10_10_2 = 61,
+	VFMT4_INT_10_10_10_2 = 62,
+	VFMT4_NORM_INT_10_10_10_2 = 63,
+};
+
+enum a4xx_tex_fmt {
+	TFMT4_NORM_USHORT_565 = 11,
+	TFMT4_NORM_USHORT_5551 = 10,
+	TFMT4_NORM_USHORT_4444 = 8,
+	TFMT4_NORM_UINT_X8Z24 = 71,
+	TFMT4_NORM_UINT_2_10_10_10 = 33,
+	TFMT4_NORM_UINT_A8 = 3,
+	TFMT4_NORM_UINT_L8_A8 = 13,
+	TFMT4_NORM_UINT_8 = 4,
+	TFMT4_NORM_UINT_8_8_8_8 = 28,
+	TFMT4_FLOAT_16 = 20,
+	TFMT4_FLOAT_16_16 = 40,
+	TFMT4_FLOAT_16_16_16_16 = 53,
+	TFMT4_FLOAT_32 = 43,
+	TFMT4_FLOAT_32_32 = 56,
+	TFMT4_FLOAT_32_32_32_32 = 63,
+};
+
+enum a4xx_depth_format {
+	DEPTH4_NONE = 0,
+	DEPTH4_16 = 1,
+	DEPTH4_24_8 = 2,
+};
+
+enum a4xx_tex_filter {
+	A4XX_TEX_NEAREST = 0,
+	A4XX_TEX_LINEAR = 1,
+};
+
+enum a4xx_tex_clamp {
+	A4XX_TEX_REPEAT = 0,
+	A4XX_TEX_CLAMP_TO_EDGE = 1,
+	A4XX_TEX_MIRROR_REPEAT = 2,
+	A4XX_TEX_CLAMP_NONE = 3,
+};
+
+enum a4xx_tex_swiz {
+	A4XX_TEX_X = 0,
+	A4XX_TEX_Y = 1,
+	A4XX_TEX_Z = 2,
+	A4XX_TEX_W = 3,
+	A4XX_TEX_ZERO = 4,
+	A4XX_TEX_ONE = 5,
+};
+
+enum a4xx_tex_type {
+	A4XX_TEX_1D = 0,
+	A4XX_TEX_2D = 1,
+	A4XX_TEX_CUBE = 2,
+	A4XX_TEX_3D = 3,
+};
+
+#define A4XX_CGC_HLSQ_EARLY_CYC__MASK				0x00700000
+#define A4XX_CGC_HLSQ_EARLY_CYC__SHIFT				20
+static inline uint32_t A4XX_CGC_HLSQ_EARLY_CYC(uint32_t val)
+{
+	return ((val) << A4XX_CGC_HLSQ_EARLY_CYC__SHIFT) & A4XX_CGC_HLSQ_EARLY_CYC__MASK;
+}
+#define A4XX_INT0_RBBM_GPU_IDLE					0x00000001
+#define A4XX_INT0_RBBM_AHB_ERROR				0x00000002
+#define A4XX_INT0_RBBM_REG_TIMEOUT				0x00000004
+#define A4XX_INT0_RBBM_ME_MS_TIMEOUT				0x00000008
+#define A4XX_INT0_RBBM_PFP_MS_TIMEOUT				0x00000010
+#define A4XX_INT0_RBBM_ATB_BUS_OVERFLOW				0x00000020
+#define A4XX_INT0_VFD_ERROR					0x00000040
+#define A4XX_INT0_CP_SW_INT					0x00000080
+#define A4XX_INT0_CP_T0_PACKET_IN_IB				0x00000100
+#define A4XX_INT0_CP_OPCODE_ERROR				0x00000200
+#define A4XX_INT0_CP_RESERVED_BIT_ERROR				0x00000400
+#define A4XX_INT0_CP_HW_FAULT					0x00000800
+#define A4XX_INT0_CP_DMA					0x00001000
+#define A4XX_INT0_CP_IB2_INT					0x00002000
+#define A4XX_INT0_CP_IB1_INT					0x00004000
+#define A4XX_INT0_CP_RB_INT					0x00008000
+#define A4XX_INT0_CP_REG_PROTECT_FAULT				0x00010000
+#define A4XX_INT0_CP_RB_DONE_TS					0x00020000
+#define A4XX_INT0_CP_VS_DONE_TS					0x00040000
+#define A4XX_INT0_CP_PS_DONE_TS					0x00080000
+#define A4XX_INT0_CACHE_FLUSH_TS				0x00100000
+#define A4XX_INT0_CP_AHB_ERROR_HALT				0x00200000
+#define A4XX_INT0_MISC_HANG_DETECT				0x01000000
+#define A4XX_INT0_UCHE_OOB_ACCESS				0x02000000
+#define REG_A4XX_RB_GMEM_BASE_ADDR				0x00000cc0
+
+#define REG_A4XX_RB_PERFCTR_RB_SEL_0				0x00000cc7
+
+#define REG_A4XX_RB_PERFCTR_RB_SEL_1				0x00000cc8
+
+#define REG_A4XX_RB_PERFCTR_RB_SEL_2				0x00000cc9
+
+#define REG_A4XX_RB_PERFCTR_RB_SEL_3				0x00000cca
+
+#define REG_A4XX_RB_PERFCTR_RB_SEL_4				0x00000ccb
+
+#define REG_A4XX_RB_PERFCTR_RB_SEL_5				0x00000ccc
+
+#define REG_A4XX_RB_PERFCTR_RB_SEL_6				0x00000ccd
+
+#define REG_A4XX_RB_PERFCTR_RB_SEL_7				0x00000cce
+
+#define REG_A4XX_RB_PERFCTR_CCU_SEL_3				0x00000cd2
+
+#define REG_A4XX_RB_FRAME_BUFFER_DIMENSION			0x00000ce0
+#define A4XX_RB_FRAME_BUFFER_DIMENSION_WIDTH__MASK		0x00003fff
+#define A4XX_RB_FRAME_BUFFER_DIMENSION_WIDTH__SHIFT		0
+static inline uint32_t A4XX_RB_FRAME_BUFFER_DIMENSION_WIDTH(uint32_t val)
+{
+	return ((val) << A4XX_RB_FRAME_BUFFER_DIMENSION_WIDTH__SHIFT) & A4XX_RB_FRAME_BUFFER_DIMENSION_WIDTH__MASK;
+}
+#define A4XX_RB_FRAME_BUFFER_DIMENSION_HEIGHT__MASK		0x3fff0000
+#define A4XX_RB_FRAME_BUFFER_DIMENSION_HEIGHT__SHIFT		16
+static inline uint32_t A4XX_RB_FRAME_BUFFER_DIMENSION_HEIGHT(uint32_t val)
+{
+	return ((val) << A4XX_RB_FRAME_BUFFER_DIMENSION_HEIGHT__SHIFT) & A4XX_RB_FRAME_BUFFER_DIMENSION_HEIGHT__MASK;
+}
+
+#define REG_A4XX_RB_CLEAR_COLOR_DW0				0x000020cc
+
+#define REG_A4XX_RB_CLEAR_COLOR_DW1				0x000020cd
+
+#define REG_A4XX_RB_CLEAR_COLOR_DW2				0x000020ce
+
+#define REG_A4XX_RB_CLEAR_COLOR_DW3				0x000020cf
+
+#define REG_A4XX_RB_MODE_CONTROL				0x000020a0
+#define A4XX_RB_MODE_CONTROL_WIDTH__MASK			0x0000003f
+#define A4XX_RB_MODE_CONTROL_WIDTH__SHIFT			0
+static inline uint32_t A4XX_RB_MODE_CONTROL_WIDTH(uint32_t val)
+{
+	return ((val >> 5) << A4XX_RB_MODE_CONTROL_WIDTH__SHIFT) & A4XX_RB_MODE_CONTROL_WIDTH__MASK;
+}
+#define A4XX_RB_MODE_CONTROL_HEIGHT__MASK			0x00003f00
+#define A4XX_RB_MODE_CONTROL_HEIGHT__SHIFT			8
+static inline uint32_t A4XX_RB_MODE_CONTROL_HEIGHT(uint32_t val)
+{
+	return ((val >> 5) << A4XX_RB_MODE_CONTROL_HEIGHT__SHIFT) & A4XX_RB_MODE_CONTROL_HEIGHT__MASK;
+}
+
+#define REG_A4XX_RB_RENDER_CONTROL				0x000020a1
+#define A4XX_RB_RENDER_CONTROL_BINNING_PASS			0x00000001
+#define A4XX_RB_RENDER_CONTROL_DISABLE_COLOR_PIPE		0x00000020
+
+#define REG_A4XX_RB_MSAA_CONTROL				0x000020a2
+#define A4XX_RB_MSAA_CONTROL_DISABLE				0x00001000
+#define A4XX_RB_MSAA_CONTROL_SAMPLES__MASK			0x0000e000
+#define A4XX_RB_MSAA_CONTROL_SAMPLES__SHIFT			13
+static inline uint32_t A4XX_RB_MSAA_CONTROL_SAMPLES(uint32_t val)
+{
+	return ((val) << A4XX_RB_MSAA_CONTROL_SAMPLES__SHIFT) & A4XX_RB_MSAA_CONTROL_SAMPLES__MASK;
+}
+
+#define REG_A4XX_RB_MSAA_CONTROL2				0x000020a3
+#define A4XX_RB_MSAA_CONTROL2_MSAA_SAMPLES__MASK		0x00000380
+#define A4XX_RB_MSAA_CONTROL2_MSAA_SAMPLES__SHIFT		7
+static inline uint32_t A4XX_RB_MSAA_CONTROL2_MSAA_SAMPLES(uint32_t val)
+{
+	return ((val) << A4XX_RB_MSAA_CONTROL2_MSAA_SAMPLES__SHIFT) & A4XX_RB_MSAA_CONTROL2_MSAA_SAMPLES__MASK;
+}
+#define A4XX_RB_MSAA_CONTROL2_VARYING				0x00001000
+
+static inline uint32_t REG_A4XX_RB_MRT(uint32_t i0) { return 0x000020a4 + 0x5*i0; }
+
+static inline uint32_t REG_A4XX_RB_MRT_CONTROL(uint32_t i0) { return 0x000020a4 + 0x5*i0; }
+#define A4XX_RB_MRT_CONTROL_READ_DEST_ENABLE			0x00000008
+#define A4XX_RB_MRT_CONTROL_BLEND				0x00000010
+#define A4XX_RB_MRT_CONTROL_BLEND2				0x00000020
+#define A4XX_RB_MRT_CONTROL_FASTCLEAR				0x00000400
+#define A4XX_RB_MRT_CONTROL_B11					0x00000800
+#define A4XX_RB_MRT_CONTROL_COMPONENT_ENABLE__MASK		0x0f000000
+#define A4XX_RB_MRT_CONTROL_COMPONENT_ENABLE__SHIFT		24
+static inline uint32_t A4XX_RB_MRT_CONTROL_COMPONENT_ENABLE(uint32_t val)
+{
+	return ((val) << A4XX_RB_MRT_CONTROL_COMPONENT_ENABLE__SHIFT) & A4XX_RB_MRT_CONTROL_COMPONENT_ENABLE__MASK;
+}
+
+static inline uint32_t REG_A4XX_RB_MRT_BUF_INFO(uint32_t i0) { return 0x000020a5 + 0x5*i0; }
+#define A4XX_RB_MRT_BUF_INFO_COLOR_FORMAT__MASK			0x0000003f
+#define A4XX_RB_MRT_BUF_INFO_COLOR_FORMAT__SHIFT		0
+static inline uint32_t A4XX_RB_MRT_BUF_INFO_COLOR_FORMAT(enum a4xx_color_fmt val)
+{
+	return ((val) << A4XX_RB_MRT_BUF_INFO_COLOR_FORMAT__SHIFT) & A4XX_RB_MRT_BUF_INFO_COLOR_FORMAT__MASK;
+}
+#define A4XX_RB_MRT_BUF_INFO_DITHER_MODE__MASK			0x00000600
+#define A4XX_RB_MRT_BUF_INFO_DITHER_MODE__SHIFT			9
+static inline uint32_t A4XX_RB_MRT_BUF_INFO_DITHER_MODE(enum adreno_rb_dither_mode val)
+{
+	return ((val) << A4XX_RB_MRT_BUF_INFO_DITHER_MODE__SHIFT) & A4XX_RB_MRT_BUF_INFO_DITHER_MODE__MASK;
+}
+#define A4XX_RB_MRT_BUF_INFO_COLOR_SWAP__MASK			0x00001800
+#define A4XX_RB_MRT_BUF_INFO_COLOR_SWAP__SHIFT			11
+static inline uint32_t A4XX_RB_MRT_BUF_INFO_COLOR_SWAP(enum a3xx_color_swap val)
+{
+	return ((val) << A4XX_RB_MRT_BUF_INFO_COLOR_SWAP__SHIFT) & A4XX_RB_MRT_BUF_INFO_COLOR_SWAP__MASK;
+}
+#define A4XX_RB_MRT_BUF_INFO_COLOR_BUF_PITCH__MASK		0x007fc000
+#define A4XX_RB_MRT_BUF_INFO_COLOR_BUF_PITCH__SHIFT		14
+static inline uint32_t A4XX_RB_MRT_BUF_INFO_COLOR_BUF_PITCH(uint32_t val)
+{
+	return ((val >> 4) << A4XX_RB_MRT_BUF_INFO_COLOR_BUF_PITCH__SHIFT) & A4XX_RB_MRT_BUF_INFO_COLOR_BUF_PITCH__MASK;
+}
+
+static inline uint32_t REG_A4XX_RB_MRT_BASE(uint32_t i0) { return 0x000020a6 + 0x5*i0; }
+
+static inline uint32_t REG_A4XX_RB_MRT_CONTROL3(uint32_t i0) { return 0x000020a7 + 0x5*i0; }
+#define A4XX_RB_MRT_CONTROL3_STRIDE__MASK			0x0001fff8
+#define A4XX_RB_MRT_CONTROL3_STRIDE__SHIFT			3
+static inline uint32_t A4XX_RB_MRT_CONTROL3_STRIDE(uint32_t val)
+{
+	return ((val) << A4XX_RB_MRT_CONTROL3_STRIDE__SHIFT) & A4XX_RB_MRT_CONTROL3_STRIDE__MASK;
+}
+
+static inline uint32_t REG_A4XX_RB_MRT_BLEND_CONTROL(uint32_t i0) { return 0x000020a8 + 0x5*i0; }
+#define A4XX_RB_MRT_BLEND_CONTROL_RGB_SRC_FACTOR__MASK		0x0000001f
+#define A4XX_RB_MRT_BLEND_CONTROL_RGB_SRC_FACTOR__SHIFT		0
+static inline uint32_t A4XX_RB_MRT_BLEND_CONTROL_RGB_SRC_FACTOR(enum adreno_rb_blend_factor val)
+{
+	return ((val) << A4XX_RB_MRT_BLEND_CONTROL_RGB_SRC_FACTOR__SHIFT) & A4XX_RB_MRT_BLEND_CONTROL_RGB_SRC_FACTOR__MASK;
+}
+#define A4XX_RB_MRT_BLEND_CONTROL_RGB_BLEND_OPCODE__MASK	0x000000e0
+#define A4XX_RB_MRT_BLEND_CONTROL_RGB_BLEND_OPCODE__SHIFT	5
+static inline uint32_t A4XX_RB_MRT_BLEND_CONTROL_RGB_BLEND_OPCODE(enum a4xx_rb_blend_opcode val)
+{
+	return ((val) << A4XX_RB_MRT_BLEND_CONTROL_RGB_BLEND_OPCODE__SHIFT) & A4XX_RB_MRT_BLEND_CONTROL_RGB_BLEND_OPCODE__MASK;
+}
+#define A4XX_RB_MRT_BLEND_CONTROL_RGB_DEST_FACTOR__MASK		0x00001f00
+#define A4XX_RB_MRT_BLEND_CONTROL_RGB_DEST_FACTOR__SHIFT	8
+static inline uint32_t A4XX_RB_MRT_BLEND_CONTROL_RGB_DEST_FACTOR(enum adreno_rb_blend_factor val)
+{
+	return ((val) << A4XX_RB_MRT_BLEND_CONTROL_RGB_DEST_FACTOR__SHIFT) & A4XX_RB_MRT_BLEND_CONTROL_RGB_DEST_FACTOR__MASK;
+}
+#define A4XX_RB_MRT_BLEND_CONTROL_ALPHA_SRC_FACTOR__MASK	0x001f0000
+#define A4XX_RB_MRT_BLEND_CONTROL_ALPHA_SRC_FACTOR__SHIFT	16
+static inline uint32_t A4XX_RB_MRT_BLEND_CONTROL_ALPHA_SRC_FACTOR(enum adreno_rb_blend_factor val)
+{
+	return ((val) << A4XX_RB_MRT_BLEND_CONTROL_ALPHA_SRC_FACTOR__SHIFT) & A4XX_RB_MRT_BLEND_CONTROL_ALPHA_SRC_FACTOR__MASK;
+}
+#define A4XX_RB_MRT_BLEND_CONTROL_ALPHA_BLEND_OPCODE__MASK	0x00e00000
+#define A4XX_RB_MRT_BLEND_CONTROL_ALPHA_BLEND_OPCODE__SHIFT	21
+static inline uint32_t A4XX_RB_MRT_BLEND_CONTROL_ALPHA_BLEND_OPCODE(enum a4xx_rb_blend_opcode val)
+{
+	return ((val) << A4XX_RB_MRT_BLEND_CONTROL_ALPHA_BLEND_OPCODE__SHIFT) & A4XX_RB_MRT_BLEND_CONTROL_ALPHA_BLEND_OPCODE__MASK;
+}
+#define A4XX_RB_MRT_BLEND_CONTROL_ALPHA_DEST_FACTOR__MASK	0x1f000000
+#define A4XX_RB_MRT_BLEND_CONTROL_ALPHA_DEST_FACTOR__SHIFT	24
+static inline uint32_t A4XX_RB_MRT_BLEND_CONTROL_ALPHA_DEST_FACTOR(enum adreno_rb_blend_factor val)
+{
+	return ((val) << A4XX_RB_MRT_BLEND_CONTROL_ALPHA_DEST_FACTOR__SHIFT) & A4XX_RB_MRT_BLEND_CONTROL_ALPHA_DEST_FACTOR__MASK;
+}
+
+#define REG_A4XX_RB_ALPHA_CONTROL				0x000020f8
+#define A4XX_RB_ALPHA_CONTROL_ALPHA_TEST			0x00000100
+#define A4XX_RB_ALPHA_CONTROL_ALPHA_TEST_FUNC__MASK		0x00000e00
+#define A4XX_RB_ALPHA_CONTROL_ALPHA_TEST_FUNC__SHIFT		9
+static inline uint32_t A4XX_RB_ALPHA_CONTROL_ALPHA_TEST_FUNC(enum adreno_compare_func val)
+{
+	return ((val) << A4XX_RB_ALPHA_CONTROL_ALPHA_TEST_FUNC__SHIFT) & A4XX_RB_ALPHA_CONTROL_ALPHA_TEST_FUNC__MASK;
+}
+
+#define REG_A4XX_RB_FS_OUTPUT					0x000020f9
+#define A4XX_RB_FS_OUTPUT_ENABLE_COLOR_PIPE			0x00000001
+#define A4XX_RB_FS_OUTPUT_FAST_CLEAR				0x00000100
+#define A4XX_RB_FS_OUTPUT_SAMPLE_MASK__MASK			0xffff0000
+#define A4XX_RB_FS_OUTPUT_SAMPLE_MASK__SHIFT			16
+static inline uint32_t A4XX_RB_FS_OUTPUT_SAMPLE_MASK(uint32_t val)
+{
+	return ((val) << A4XX_RB_FS_OUTPUT_SAMPLE_MASK__SHIFT) & A4XX_RB_FS_OUTPUT_SAMPLE_MASK__MASK;
+}
+
+#define REG_A4XX_RB_RENDER_CONTROL3				0x000020fb
+#define A4XX_RB_RENDER_CONTROL3_COMPONENT_ENABLE__MASK		0x0000001f
+#define A4XX_RB_RENDER_CONTROL3_COMPONENT_ENABLE__SHIFT		0
+static inline uint32_t A4XX_RB_RENDER_CONTROL3_COMPONENT_ENABLE(uint32_t val)
+{
+	return ((val) << A4XX_RB_RENDER_CONTROL3_COMPONENT_ENABLE__SHIFT) & A4XX_RB_RENDER_CONTROL3_COMPONENT_ENABLE__MASK;
+}
+
+#define REG_A4XX_RB_COPY_CONTROL				0x000020fc
+#define A4XX_RB_COPY_CONTROL_MSAA_RESOLVE__MASK			0x00000003
+#define A4XX_RB_COPY_CONTROL_MSAA_RESOLVE__SHIFT		0
+static inline uint32_t A4XX_RB_COPY_CONTROL_MSAA_RESOLVE(enum a3xx_msaa_samples val)
+{
+	return ((val) << A4XX_RB_COPY_CONTROL_MSAA_RESOLVE__SHIFT) & A4XX_RB_COPY_CONTROL_MSAA_RESOLVE__MASK;
+}
+#define A4XX_RB_COPY_CONTROL_MODE__MASK				0x00000070
+#define A4XX_RB_COPY_CONTROL_MODE__SHIFT			4
+static inline uint32_t A4XX_RB_COPY_CONTROL_MODE(enum adreno_rb_copy_control_mode val)
+{
+	return ((val) << A4XX_RB_COPY_CONTROL_MODE__SHIFT) & A4XX_RB_COPY_CONTROL_MODE__MASK;
+}
+#define A4XX_RB_COPY_CONTROL_FASTCLEAR__MASK			0x00000f00
+#define A4XX_RB_COPY_CONTROL_FASTCLEAR__SHIFT			8
+static inline uint32_t A4XX_RB_COPY_CONTROL_FASTCLEAR(uint32_t val)
+{
+	return ((val) << A4XX_RB_COPY_CONTROL_FASTCLEAR__SHIFT) & A4XX_RB_COPY_CONTROL_FASTCLEAR__MASK;
+}
+#define A4XX_RB_COPY_CONTROL_GMEM_BASE__MASK			0xffffc000
+#define A4XX_RB_COPY_CONTROL_GMEM_BASE__SHIFT			14
+static inline uint32_t A4XX_RB_COPY_CONTROL_GMEM_BASE(uint32_t val)
+{
+	return ((val >> 14) << A4XX_RB_COPY_CONTROL_GMEM_BASE__SHIFT) & A4XX_RB_COPY_CONTROL_GMEM_BASE__MASK;
+}
+
+#define REG_A4XX_RB_COPY_DEST_BASE				0x000020fd
+#define A4XX_RB_COPY_DEST_BASE_BASE__MASK			0xfffffff0
+#define A4XX_RB_COPY_DEST_BASE_BASE__SHIFT			4
+static inline uint32_t A4XX_RB_COPY_DEST_BASE_BASE(uint32_t val)
+{
+	return ((val >> 4) << A4XX_RB_COPY_DEST_BASE_BASE__SHIFT) & A4XX_RB_COPY_DEST_BASE_BASE__MASK;
+}
+
+#define REG_A4XX_RB_COPY_DEST_PITCH				0x000020fe
+#define A4XX_RB_COPY_DEST_PITCH_PITCH__MASK			0xffffffff
+#define A4XX_RB_COPY_DEST_PITCH_PITCH__SHIFT			0
+static inline uint32_t A4XX_RB_COPY_DEST_PITCH_PITCH(uint32_t val)
+{
+	return ((val >> 5) << A4XX_RB_COPY_DEST_PITCH_PITCH__SHIFT) & A4XX_RB_COPY_DEST_PITCH_PITCH__MASK;
+}
+
+#define REG_A4XX_RB_COPY_DEST_INFO				0x000020ff
+#define A4XX_RB_COPY_DEST_INFO_FORMAT__MASK			0x000000fc
+#define A4XX_RB_COPY_DEST_INFO_FORMAT__SHIFT			2
+static inline uint32_t A4XX_RB_COPY_DEST_INFO_FORMAT(enum a4xx_color_fmt val)
+{
+	return ((val) << A4XX_RB_COPY_DEST_INFO_FORMAT__SHIFT) & A4XX_RB_COPY_DEST_INFO_FORMAT__MASK;
+}
+#define A4XX_RB_COPY_DEST_INFO_SWAP__MASK			0x00000300
+#define A4XX_RB_COPY_DEST_INFO_SWAP__SHIFT			8
+static inline uint32_t A4XX_RB_COPY_DEST_INFO_SWAP(enum a3xx_color_swap val)
+{
+	return ((val) << A4XX_RB_COPY_DEST_INFO_SWAP__SHIFT) & A4XX_RB_COPY_DEST_INFO_SWAP__MASK;
+}
+#define A4XX_RB_COPY_DEST_INFO_DITHER_MODE__MASK		0x00000c00
+#define A4XX_RB_COPY_DEST_INFO_DITHER_MODE__SHIFT		10
+static inline uint32_t A4XX_RB_COPY_DEST_INFO_DITHER_MODE(enum adreno_rb_dither_mode val)
+{
+	return ((val) << A4XX_RB_COPY_DEST_INFO_DITHER_MODE__SHIFT) & A4XX_RB_COPY_DEST_INFO_DITHER_MODE__MASK;
+}
+#define A4XX_RB_COPY_DEST_INFO_COMPONENT_ENABLE__MASK		0x0003c000
+#define A4XX_RB_COPY_DEST_INFO_COMPONENT_ENABLE__SHIFT		14
+static inline uint32_t A4XX_RB_COPY_DEST_INFO_COMPONENT_ENABLE(uint32_t val)
+{
+	return ((val) << A4XX_RB_COPY_DEST_INFO_COMPONENT_ENABLE__SHIFT) & A4XX_RB_COPY_DEST_INFO_COMPONENT_ENABLE__MASK;
+}
+#define A4XX_RB_COPY_DEST_INFO_ENDIAN__MASK			0x001c0000
+#define A4XX_RB_COPY_DEST_INFO_ENDIAN__SHIFT			18
+static inline uint32_t A4XX_RB_COPY_DEST_INFO_ENDIAN(enum adreno_rb_surface_endian val)
+{
+	return ((val) << A4XX_RB_COPY_DEST_INFO_ENDIAN__SHIFT) & A4XX_RB_COPY_DEST_INFO_ENDIAN__MASK;
+}
+#define A4XX_RB_COPY_DEST_INFO_TILE__MASK			0x03000000
+#define A4XX_RB_COPY_DEST_INFO_TILE__SHIFT			24
+static inline uint32_t A4XX_RB_COPY_DEST_INFO_TILE(enum a4xx_tile_mode val)
+{
+	return ((val) << A4XX_RB_COPY_DEST_INFO_TILE__SHIFT) & A4XX_RB_COPY_DEST_INFO_TILE__MASK;
+}
+
+#define REG_A4XX_RB_FS_OUTPUT_REG				0x00002100
+#define A4XX_RB_FS_OUTPUT_REG_COLOR_PIPE_ENABLE			0x00000001
+#define A4XX_RB_FS_OUTPUT_REG_FRAG_WRITES_Z			0x00000020
+
+#define REG_A4XX_RB_DEPTH_CONTROL				0x00002101
+#define A4XX_RB_DEPTH_CONTROL_FRAG_WRITES_Z			0x00000001
+#define A4XX_RB_DEPTH_CONTROL_Z_ENABLE				0x00000002
+#define A4XX_RB_DEPTH_CONTROL_Z_WRITE_ENABLE			0x00000004
+#define A4XX_RB_DEPTH_CONTROL_ZFUNC__MASK			0x00000070
+#define A4XX_RB_DEPTH_CONTROL_ZFUNC__SHIFT			4
+static inline uint32_t A4XX_RB_DEPTH_CONTROL_ZFUNC(enum adreno_compare_func val)
+{
+	return ((val) << A4XX_RB_DEPTH_CONTROL_ZFUNC__SHIFT) & A4XX_RB_DEPTH_CONTROL_ZFUNC__MASK;
+}
+#define A4XX_RB_DEPTH_CONTROL_BF_ENABLE				0x00000080
+#define A4XX_RB_DEPTH_CONTROL_EARLY_Z_DISABLE			0x00010000
+#define A4XX_RB_DEPTH_CONTROL_Z_TEST_ENABLE			0x80000000
+
+#define REG_A4XX_RB_DEPTH_CLEAR					0x00002102
+
+#define REG_A4XX_RB_DEPTH_INFO					0x00002103
+#define A4XX_RB_DEPTH_INFO_DEPTH_FORMAT__MASK			0x00000003
+#define A4XX_RB_DEPTH_INFO_DEPTH_FORMAT__SHIFT			0
+static inline uint32_t A4XX_RB_DEPTH_INFO_DEPTH_FORMAT(enum a4xx_depth_format val)
+{
+	return ((val) << A4XX_RB_DEPTH_INFO_DEPTH_FORMAT__SHIFT) & A4XX_RB_DEPTH_INFO_DEPTH_FORMAT__MASK;
+}
+#define A4XX_RB_DEPTH_INFO_DEPTH_BASE__MASK			0xfffff000
+#define A4XX_RB_DEPTH_INFO_DEPTH_BASE__SHIFT			12
+static inline uint32_t A4XX_RB_DEPTH_INFO_DEPTH_BASE(uint32_t val)
+{
+	return ((val >> 12) << A4XX_RB_DEPTH_INFO_DEPTH_BASE__SHIFT) & A4XX_RB_DEPTH_INFO_DEPTH_BASE__MASK;
+}
+
+#define REG_A4XX_RB_DEPTH_PITCH					0x00002104
+#define A4XX_RB_DEPTH_PITCH__MASK				0xffffffff
+#define A4XX_RB_DEPTH_PITCH__SHIFT				0
+static inline uint32_t A4XX_RB_DEPTH_PITCH(uint32_t val)
+{
+	return ((val >> 4) << A4XX_RB_DEPTH_PITCH__SHIFT) & A4XX_RB_DEPTH_PITCH__MASK;
+}
+
+#define REG_A4XX_RB_DEPTH_PITCH2				0x00002105
+#define A4XX_RB_DEPTH_PITCH2__MASK				0xffffffff
+#define A4XX_RB_DEPTH_PITCH2__SHIFT				0
+static inline uint32_t A4XX_RB_DEPTH_PITCH2(uint32_t val)
+{
+	return ((val >> 4) << A4XX_RB_DEPTH_PITCH2__SHIFT) & A4XX_RB_DEPTH_PITCH2__MASK;
+}
+
+#define REG_A4XX_RB_STENCIL_CONTROL				0x00002106
+#define A4XX_RB_STENCIL_CONTROL_STENCIL_ENABLE			0x00000001
+#define A4XX_RB_STENCIL_CONTROL_STENCIL_ENABLE_BF		0x00000002
+#define A4XX_RB_STENCIL_CONTROL_STENCIL_READ			0x00000004
+#define A4XX_RB_STENCIL_CONTROL_FUNC__MASK			0x00000700
+#define A4XX_RB_STENCIL_CONTROL_FUNC__SHIFT			8
+static inline uint32_t A4XX_RB_STENCIL_CONTROL_FUNC(enum adreno_compare_func val)
+{
+	return ((val) << A4XX_RB_STENCIL_CONTROL_FUNC__SHIFT) & A4XX_RB_STENCIL_CONTROL_FUNC__MASK;
+}
+#define A4XX_RB_STENCIL_CONTROL_FAIL__MASK			0x00003800
+#define A4XX_RB_STENCIL_CONTROL_FAIL__SHIFT			11
+static inline uint32_t A4XX_RB_STENCIL_CONTROL_FAIL(enum adreno_stencil_op val)
+{
+	return ((val) << A4XX_RB_STENCIL_CONTROL_FAIL__SHIFT) & A4XX_RB_STENCIL_CONTROL_FAIL__MASK;
+}
+#define A4XX_RB_STENCIL_CONTROL_ZPASS__MASK			0x0001c000
+#define A4XX_RB_STENCIL_CONTROL_ZPASS__SHIFT			14
+static inline uint32_t A4XX_RB_STENCIL_CONTROL_ZPASS(enum adreno_stencil_op val)
+{
+	return ((val) << A4XX_RB_STENCIL_CONTROL_ZPASS__SHIFT) & A4XX_RB_STENCIL_CONTROL_ZPASS__MASK;
+}
+#define A4XX_RB_STENCIL_CONTROL_ZFAIL__MASK			0x000e0000
+#define A4XX_RB_STENCIL_CONTROL_ZFAIL__SHIFT			17
+static inline uint32_t A4XX_RB_STENCIL_CONTROL_ZFAIL(enum adreno_stencil_op val)
+{
+	return ((val) << A4XX_RB_STENCIL_CONTROL_ZFAIL__SHIFT) & A4XX_RB_STENCIL_CONTROL_ZFAIL__MASK;
+}
+#define A4XX_RB_STENCIL_CONTROL_FUNC_BF__MASK			0x00700000
+#define A4XX_RB_STENCIL_CONTROL_FUNC_BF__SHIFT			20
+static inline uint32_t A4XX_RB_STENCIL_CONTROL_FUNC_BF(enum adreno_compare_func val)
+{
+	return ((val) << A4XX_RB_STENCIL_CONTROL_FUNC_BF__SHIFT) & A4XX_RB_STENCIL_CONTROL_FUNC_BF__MASK;
+}
+#define A4XX_RB_STENCIL_CONTROL_FAIL_BF__MASK			0x03800000
+#define A4XX_RB_STENCIL_CONTROL_FAIL_BF__SHIFT			23
+static inline uint32_t A4XX_RB_STENCIL_CONTROL_FAIL_BF(enum adreno_stencil_op val)
+{
+	return ((val) << A4XX_RB_STENCIL_CONTROL_FAIL_BF__SHIFT) & A4XX_RB_STENCIL_CONTROL_FAIL_BF__MASK;
+}
+#define A4XX_RB_STENCIL_CONTROL_ZPASS_BF__MASK			0x1c000000
+#define A4XX_RB_STENCIL_CONTROL_ZPASS_BF__SHIFT			26
+static inline uint32_t A4XX_RB_STENCIL_CONTROL_ZPASS_BF(enum adreno_stencil_op val)
+{
+	return ((val) << A4XX_RB_STENCIL_CONTROL_ZPASS_BF__SHIFT) & A4XX_RB_STENCIL_CONTROL_ZPASS_BF__MASK;
+}
+#define A4XX_RB_STENCIL_CONTROL_ZFAIL_BF__MASK			0xe0000000
+#define A4XX_RB_STENCIL_CONTROL_ZFAIL_BF__SHIFT			29
+static inline uint32_t A4XX_RB_STENCIL_CONTROL_ZFAIL_BF(enum adreno_stencil_op val)
+{
+	return ((val) << A4XX_RB_STENCIL_CONTROL_ZFAIL_BF__SHIFT) & A4XX_RB_STENCIL_CONTROL_ZFAIL_BF__MASK;
+}
+
+#define REG_A4XX_RB_STENCIL_CONTROL2				0x00002107
+#define A4XX_RB_STENCIL_CONTROL2_STENCIL_BUFFER			0x00000001
+
+#define REG_A4XX_RB_STENCILREFMASK				0x0000210b
+#define A4XX_RB_STENCILREFMASK_STENCILREF__MASK			0x000000ff
+#define A4XX_RB_STENCILREFMASK_STENCILREF__SHIFT		0
+static inline uint32_t A4XX_RB_STENCILREFMASK_STENCILREF(uint32_t val)
+{
+	return ((val) << A4XX_RB_STENCILREFMASK_STENCILREF__SHIFT) & A4XX_RB_STENCILREFMASK_STENCILREF__MASK;
+}
+#define A4XX_RB_STENCILREFMASK_STENCILMASK__MASK		0x0000ff00
+#define A4XX_RB_STENCILREFMASK_STENCILMASK__SHIFT		8
+static inline uint32_t A4XX_RB_STENCILREFMASK_STENCILMASK(uint32_t val)
+{
+	return ((val) << A4XX_RB_STENCILREFMASK_STENCILMASK__SHIFT) & A4XX_RB_STENCILREFMASK_STENCILMASK__MASK;
+}
+#define A4XX_RB_STENCILREFMASK_STENCILWRITEMASK__MASK		0x00ff0000
+#define A4XX_RB_STENCILREFMASK_STENCILWRITEMASK__SHIFT		16
+static inline uint32_t A4XX_RB_STENCILREFMASK_STENCILWRITEMASK(uint32_t val)
+{
+	return ((val) << A4XX_RB_STENCILREFMASK_STENCILWRITEMASK__SHIFT) & A4XX_RB_STENCILREFMASK_STENCILWRITEMASK__MASK;
+}
+
+#define REG_A4XX_RB_STENCILREFMASK_BF				0x0000210c
+#define A4XX_RB_STENCILREFMASK_BF_STENCILREF__MASK		0x000000ff
+#define A4XX_RB_STENCILREFMASK_BF_STENCILREF__SHIFT		0
+static inline uint32_t A4XX_RB_STENCILREFMASK_BF_STENCILREF(uint32_t val)
+{
+	return ((val) << A4XX_RB_STENCILREFMASK_BF_STENCILREF__SHIFT) & A4XX_RB_STENCILREFMASK_BF_STENCILREF__MASK;
+}
+#define A4XX_RB_STENCILREFMASK_BF_STENCILMASK__MASK		0x0000ff00
+#define A4XX_RB_STENCILREFMASK_BF_STENCILMASK__SHIFT		8
+static inline uint32_t A4XX_RB_STENCILREFMASK_BF_STENCILMASK(uint32_t val)
+{
+	return ((val) << A4XX_RB_STENCILREFMASK_BF_STENCILMASK__SHIFT) & A4XX_RB_STENCILREFMASK_BF_STENCILMASK__MASK;
+}
+#define A4XX_RB_STENCILREFMASK_BF_STENCILWRITEMASK__MASK	0x00ff0000
+#define A4XX_RB_STENCILREFMASK_BF_STENCILWRITEMASK__SHIFT	16
+static inline uint32_t A4XX_RB_STENCILREFMASK_BF_STENCILWRITEMASK(uint32_t val)
+{
+	return ((val) << A4XX_RB_STENCILREFMASK_BF_STENCILWRITEMASK__SHIFT) & A4XX_RB_STENCILREFMASK_BF_STENCILWRITEMASK__MASK;
+}
+
+#define REG_A4XX_RB_BIN_OFFSET					0x0000210d
+#define A4XX_RB_BIN_OFFSET_WINDOW_OFFSET_DISABLE		0x80000000
+#define A4XX_RB_BIN_OFFSET_X__MASK				0x00007fff
+#define A4XX_RB_BIN_OFFSET_X__SHIFT				0
+static inline uint32_t A4XX_RB_BIN_OFFSET_X(uint32_t val)
+{
+	return ((val) << A4XX_RB_BIN_OFFSET_X__SHIFT) & A4XX_RB_BIN_OFFSET_X__MASK;
+}
+#define A4XX_RB_BIN_OFFSET_Y__MASK				0x7fff0000
+#define A4XX_RB_BIN_OFFSET_Y__SHIFT				16
+static inline uint32_t A4XX_RB_BIN_OFFSET_Y(uint32_t val)
+{
+	return ((val) << A4XX_RB_BIN_OFFSET_Y__SHIFT) & A4XX_RB_BIN_OFFSET_Y__MASK;
+}
+
+#define REG_A4XX_RB_VPORT_Z_CLAMP_MAX_15			0x0000213f
+
+#define REG_A4XX_RBBM_HW_VERSION				0x00000000
+
+#define REG_A4XX_RBBM_HW_CONFIGURATION				0x00000002
+
+static inline uint32_t REG_A4XX_RBBM_CLOCK_CTL_TP(uint32_t i0) { return 0x00000004 + 0x1*i0; }
+
+static inline uint32_t REG_A4XX_RBBM_CLOCK_CTL_TP_REG(uint32_t i0) { return 0x00000004 + 0x1*i0; }
+
+static inline uint32_t REG_A4XX_RBBM_CLOCK_CTL2_TP(uint32_t i0) { return 0x00000008 + 0x1*i0; }
+
+static inline uint32_t REG_A4XX_RBBM_CLOCK_CTL2_TP_REG(uint32_t i0) { return 0x00000008 + 0x1*i0; }
+
+static inline uint32_t REG_A4XX_RBBM_CLOCK_HYST_TP(uint32_t i0) { return 0x0000000c + 0x1*i0; }
+
+static inline uint32_t REG_A4XX_RBBM_CLOCK_HYST_TP_REG(uint32_t i0) { return 0x0000000c + 0x1*i0; }
+
+static inline uint32_t REG_A4XX_RBBM_CLOCK_DELAY_TP(uint32_t i0) { return 0x00000010 + 0x1*i0; }
+
+static inline uint32_t REG_A4XX_RBBM_CLOCK_DELAY_TP_REG(uint32_t i0) { return 0x00000010 + 0x1*i0; }
+
+#define REG_A4XX_RBBM_CLOCK_CTL_UCHE 				0x00000014
+
+#define REG_A4XX_RBBM_CLOCK_CTL2_UCHE				0x00000015
+
+#define REG_A4XX_RBBM_CLOCK_CTL3_UCHE				0x00000016
+
+#define REG_A4XX_RBBM_CLOCK_CTL4_UCHE				0x00000017
+
+#define REG_A4XX_RBBM_CLOCK_HYST_UCHE				0x00000018
+
+#define REG_A4XX_RBBM_CLOCK_DELAY_UCHE				0x00000019
+
+#define REG_A4XX_RBBM_CLOCK_MODE_GPC				0x0000001a
+
+#define REG_A4XX_RBBM_CLOCK_DELAY_GPC				0x0000001b
+
+#define REG_A4XX_RBBM_CLOCK_HYST_GPC				0x0000001c
+
+#define REG_A4XX_RBBM_CLOCK_CTL_TSE_RAS_RBBM			0x0000001d
+
+#define REG_A4XX_RBBM_CLOCK_HYST_TSE_RAS_RBBM			0x0000001e
+
+#define REG_A4XX_RBBM_CLOCK_DELAY_TSE_RAS_RBBM			0x0000001f
+
+#define REG_A4XX_RBBM_CLOCK_CTL					0x00000020
+
+#define REG_A4XX_RBBM_SP_HYST_CNT				0x00000021
+
+#define REG_A4XX_RBBM_SW_RESET_CMD				0x00000022
+
+#define REG_A4XX_RBBM_AHB_CTL0					0x00000023
+
+#define REG_A4XX_RBBM_AHB_CTL1					0x00000024
+
+#define REG_A4XX_RBBM_AHB_CMD					0x00000025
+
+#define REG_A4XX_RBBM_RB_SUB_BLOCK_SEL_CTL			0x00000026
+
+#define REG_A4XX_RBBM_RAM_ACC_63_32				0x00000028
+
+#define REG_A4XX_RBBM_WAIT_IDLE_CLOCKS_CTL			0x0000002b
+
+#define REG_A4XX_RBBM_INTERFACE_HANG_INT_CTL			0x0000002f
+
+#define REG_A4XX_RBBM_INTERFACE_HANG_MASK_CTL4			0x00000034
+
+#define REG_A4XX_RBBM_INT_CLEAR_CMD				0x00000036
+
+#define REG_A4XX_RBBM_INT_0_MASK				0x00000037
+
+#define REG_A4XX_RBBM_RBBM_CTL					0x0000003e
+
+#define REG_A4XX_RBBM_AHB_DEBUG_CTL				0x0000003f
+
+#define REG_A4XX_RBBM_VBIF_DEBUG_CTL				0x00000041
+
+#define REG_A4XX_RBBM_CLOCK_CTL2				0x00000042
+
+#define REG_A4XX_RBBM_BLOCK_SW_RESET_CMD			0x00000045
+
+#define REG_A4XX_RBBM_RESET_CYCLES				0x00000047
+
+#define REG_A4XX_RBBM_EXT_TRACE_BUS_CTL				0x00000049
+
+#define REG_A4XX_RBBM_CFG_DEBBUS_SEL_A				0x0000004a
+
+#define REG_A4XX_RBBM_CFG_DEBBUS_SEL_B				0x0000004b
+
+#define REG_A4XX_RBBM_CFG_DEBBUS_SEL_C				0x0000004c
+
+#define REG_A4XX_RBBM_CFG_DEBBUS_SEL_D				0x0000004d
+
+#define REG_A4XX_RBBM_PERFCTR_CP_0_LO				0x0000009c
+
+static inline uint32_t REG_A4XX_RBBM_CLOCK_CTL_SP(uint32_t i0) { return 0x00000068 + 0x1*i0; }
+
+static inline uint32_t REG_A4XX_RBBM_CLOCK_CTL_SP_REG(uint32_t i0) { return 0x00000068 + 0x1*i0; }
+
+static inline uint32_t REG_A4XX_RBBM_CLOCK_CTL2_SP(uint32_t i0) { return 0x0000006c + 0x1*i0; }
+
+static inline uint32_t REG_A4XX_RBBM_CLOCK_CTL2_SP_REG(uint32_t i0) { return 0x0000006c + 0x1*i0; }
+
+static inline uint32_t REG_A4XX_RBBM_CLOCK_HYST_SP(uint32_t i0) { return 0x00000070 + 0x1*i0; }
+
+static inline uint32_t REG_A4XX_RBBM_CLOCK_HYST_SP_REG(uint32_t i0) { return 0x00000070 + 0x1*i0; }
+
+static inline uint32_t REG_A4XX_RBBM_CLOCK_DELAY_SP(uint32_t i0) { return 0x00000074 + 0x1*i0; }
+
+static inline uint32_t REG_A4XX_RBBM_CLOCK_DELAY_SP_REG(uint32_t i0) { return 0x00000074 + 0x1*i0; }
+
+static inline uint32_t REG_A4XX_RBBM_CLOCK_CTL_RB(uint32_t i0) { return 0x00000078 + 0x1*i0; }
+
+static inline uint32_t REG_A4XX_RBBM_CLOCK_CTL_RB_REG(uint32_t i0) { return 0x00000078 + 0x1*i0; }
+
+static inline uint32_t REG_A4XX_RBBM_CLOCK_CTL2_RB(uint32_t i0) { return 0x0000007c + 0x1*i0; }
+
+static inline uint32_t REG_A4XX_RBBM_CLOCK_CTL2_RB_REG(uint32_t i0) { return 0x0000007c + 0x1*i0; }
+
+static inline uint32_t REG_A4XX_RBBM_CLOCK_CTL_MARB_CCU(uint32_t i0) { return 0x00000082 + 0x1*i0; }
+
+static inline uint32_t REG_A4XX_RBBM_CLOCK_CTL_MARB_CCU_REG(uint32_t i0) { return 0x00000082 + 0x1*i0; }
+
+static inline uint32_t REG_A4XX_RBBM_CLOCK_HYST_RB_MARB_CCU(uint32_t i0) { return 0x00000086 + 0x1*i0; }
+
+static inline uint32_t REG_A4XX_RBBM_CLOCK_HYST_RB_MARB_CCU_REG(uint32_t i0) { return 0x00000086 + 0x1*i0; }
+
+#define REG_A4XX_RBBM_CLOCK_HYST_COM_DCOM			0x00000080
+
+#define REG_A4XX_RBBM_CLOCK_CTL_COM_DCOM			0x00000081
+
+#define REG_A4XX_RBBM_CLOCK_CTL_HLSQ				0x0000008a
+
+#define REG_A4XX_RBBM_CLOCK_HYST_HLSQ				0x0000008b
+
+#define REG_A4XX_RBBM_CLOCK_DELAY_HLSQ				0x0000008c
+
+#define REG_A4XX_RBBM_CLOCK_DELAY_COM_DCOM			0x0000008d
+
+static inline uint32_t REG_A4XX_RBBM_CLOCK_DELAY_RB_MARB_CCU_L1(uint32_t i0) { return 0x0000008e + 0x1*i0; }
+
+static inline uint32_t REG_A4XX_RBBM_CLOCK_DELAY_RB_MARB_CCU_L1_REG(uint32_t i0) { return 0x0000008e + 0x1*i0; }
+
+#define REG_A4XX_RBBM_PERFCTR_PWR_1_LO				0x00000168
+
+#define REG_A4XX_RBBM_PERFCTR_CTL				0x00000170
+
+#define REG_A4XX_RBBM_PERFCTR_LOAD_CMD0				0x00000171
+
+#define REG_A4XX_RBBM_PERFCTR_LOAD_CMD1				0x00000172
+
+#define REG_A4XX_RBBM_PERFCTR_LOAD_CMD2				0x00000173
+
+#define REG_A4XX_RBBM_PERFCTR_LOAD_VALUE_LO			0x00000174
+
+#define REG_A4XX_RBBM_PERFCTR_LOAD_VALUE_HI			0x00000175
+
+#define REG_A4XX_RBBM_GPU_BUSY_MASKED				0x0000017a
+
+#define REG_A4XX_RBBM_INT_0_STATUS				0x0000017d
+
+#define REG_A4XX_RBBM_CLOCK_STATUS				0x00000182
+
+#define REG_A4XX_RBBM_AHB_STATUS				0x00000189
+
+#define REG_A4XX_RBBM_AHB_ME_SPLIT_STATUS			0x0000018c
+
+#define REG_A4XX_RBBM_AHB_PFP_SPLIT_STATUS			0x0000018d
+
+#define REG_A4XX_RBBM_AHB_ERROR_STATUS				0x0000018f
+
+#define REG_A4XX_RBBM_STATUS					0x00000191
+#define A4XX_RBBM_STATUS_HI_BUSY				0x00000001
+#define A4XX_RBBM_STATUS_CP_ME_BUSY				0x00000002
+#define A4XX_RBBM_STATUS_CP_PFP_BUSY				0x00000004
+#define A4XX_RBBM_STATUS_CP_NRT_BUSY				0x00004000
+#define A4XX_RBBM_STATUS_VBIF_BUSY				0x00008000
+#define A4XX_RBBM_STATUS_TSE_BUSY				0x00010000
+#define A4XX_RBBM_STATUS_RAS_BUSY				0x00020000
+#define A4XX_RBBM_STATUS_RB_BUSY				0x00040000
+#define A4XX_RBBM_STATUS_PC_DCALL_BUSY				0x00080000
+#define A4XX_RBBM_STATUS_PC_VSD_BUSY				0x00100000
+#define A4XX_RBBM_STATUS_VFD_BUSY				0x00200000
+#define A4XX_RBBM_STATUS_VPC_BUSY				0x00400000
+#define A4XX_RBBM_STATUS_UCHE_BUSY				0x00800000
+#define A4XX_RBBM_STATUS_SP_BUSY				0x01000000
+#define A4XX_RBBM_STATUS_TPL1_BUSY				0x02000000
+#define A4XX_RBBM_STATUS_MARB_BUSY				0x04000000
+#define A4XX_RBBM_STATUS_VSC_BUSY				0x08000000
+#define A4XX_RBBM_STATUS_ARB_BUSY				0x10000000
+#define A4XX_RBBM_STATUS_HLSQ_BUSY				0x20000000
+#define A4XX_RBBM_STATUS_GPU_BUSY_NOHC				0x40000000
+#define A4XX_RBBM_STATUS_GPU_BUSY				0x80000000
+
+#define REG_A4XX_RBBM_INTERFACE_RRDY_STATUS5			0x0000019f
+
+#define REG_A4XX_CP_SCRATCH_UMASK				0x00000228
+
+#define REG_A4XX_CP_SCRATCH_ADDR				0x00000229
+
+#define REG_A4XX_CP_RB_BASE					0x00000200
+
+#define REG_A4XX_CP_RB_CNTL					0x00000201
+
+#define REG_A4XX_CP_RB_WPTR					0x00000205
+
+#define REG_A4XX_CP_RB_RPTR_ADDR				0x00000203
+
+#define REG_A4XX_CP_RB_RPTR					0x00000204
+
+#define REG_A4XX_CP_IB1_BASE					0x00000206
+
+#define REG_A4XX_CP_IB1_BUFSZ					0x00000207
+
+#define REG_A4XX_CP_IB2_BASE					0x00000208
+
+#define REG_A4XX_CP_IB2_BUFSZ					0x00000209
+
+#define REG_A4XX_CP_ME_RB_DONE_DATA				0x00000217
+
+#define REG_A4XX_CP_QUEUE_THRESH2				0x00000219
+
+#define REG_A4XX_CP_MERCIU_SIZE					0x0000021b
+
+#define REG_A4XX_CP_ROQ_ADDR					0x0000021c
+
+#define REG_A4XX_CP_ROQ_DATA					0x0000021d
+
+#define REG_A4XX_CP_MEQ_ADDR 					0x0000021e
+
+#define REG_A4XX_CP_MEQ_DATA 					0x0000021f
+
+#define REG_A4XX_CP_MERCIU_ADDR					0x00000220
+
+#define REG_A4XX_CP_MERCIU_DATA					0x00000221
+
+#define REG_A4XX_CP_MERCIU_DATA2				0x00000222
+
+#define REG_A4XX_CP_PFP_UCODE_ADDR				0x00000223
+
+#define REG_A4XX_CP_PFP_UCODE_DATA				0x00000224
+
+#define REG_A4XX_CP_ME_RAM_WADDR				0x00000225
+
+#define REG_A4XX_CP_ME_RAM_RADDR				0x00000226
+
+#define REG_A4XX_CP_ME_RAM_DATA					0x00000227
+
+#define REG_A4XX_CP_PREEMPT					0x0000022a
+
+#define REG_A4XX_CP_CNTL					0x0000022c
+
+#define REG_A4XX_CP_ME_CNTL					0x0000022d
+
+#define REG_A4XX_CP_DEBUG					0x0000022e
+
+#define REG_A4XX_CP_DEBUG_ECO_CONTROL				0x00000231
+
+#define REG_A4XX_CP_DRAW_STATE_ADDR				0x00000232
+
+#define REG_A4XX_CP_PROTECT_REG_0				0x00000240
+
+static inline uint32_t REG_A4XX_CP_PROTECT(uint32_t i0) { return 0x00000240 + 0x1*i0; }
+
+static inline uint32_t REG_A4XX_CP_PROTECT_REG(uint32_t i0) { return 0x00000240 + 0x1*i0; }
+
+#define REG_A4XX_CP_PROTECT_CTRL				0x00000250
+
+#define REG_A4XX_CP_ST_BASE					0x000004c0
+
+#define REG_A4XX_CP_STQ_AVAIL					0x000004ce
+
+#define REG_A4XX_CP_MERCIU_STAT					0x000004d0
+
+#define REG_A4XX_CP_WFI_PEND_CTR				0x000004d2
+
+#define REG_A4XX_CP_HW_FAULT					0x000004d8
+
+#define REG_A4XX_CP_PROTECT_STATUS				0x000004da
+
+#define REG_A4XX_CP_EVENTS_IN_FLIGHT				0x000004dd
+
+#define REG_A4XX_CP_PERFCTR_CP_SEL_0				0x00000500
+
+#define REG_A4XX_CP_PERFCOMBINER_SELECT				0x0000050b
+
+static inline uint32_t REG_A4XX_CP_SCRATCH(uint32_t i0) { return 0x00000578 + 0x1*i0; }
+
+static inline uint32_t REG_A4XX_CP_SCRATCH_REG(uint32_t i0) { return 0x00000578 + 0x1*i0; }
+
+#define REG_A4XX_SP_VS_STATUS					0x00000ec0
+
+#define REG_A4XX_SP_PERFCTR_SP_SEL_11				0x00000ecf
+
+#define REG_A4XX_SP_SP_CTRL_REG					0x000022c0
+#define A4XX_SP_SP_CTRL_REG_BINNING_PASS			0x00080000
+
+#define REG_A4XX_SP_INSTR_CACHE_CTRL				0x000022c1
+
+#define REG_A4XX_SP_VS_CTRL_REG0				0x000022c4
+#define A4XX_SP_VS_CTRL_REG0_THREADMODE__MASK			0x00000001
+#define A4XX_SP_VS_CTRL_REG0_THREADMODE__SHIFT			0
+static inline uint32_t A4XX_SP_VS_CTRL_REG0_THREADMODE(enum a3xx_threadmode val)
+{
+	return ((val) << A4XX_SP_VS_CTRL_REG0_THREADMODE__SHIFT) & A4XX_SP_VS_CTRL_REG0_THREADMODE__MASK;
+}
+#define A4XX_SP_VS_CTRL_REG0_VARYING				0x00000002
+#define A4XX_SP_VS_CTRL_REG0_CACHEINVALID			0x00000004
+#define A4XX_SP_VS_CTRL_REG0_HALFREGFOOTPRINT__MASK		0x000003f0
+#define A4XX_SP_VS_CTRL_REG0_HALFREGFOOTPRINT__SHIFT		4
+static inline uint32_t A4XX_SP_VS_CTRL_REG0_HALFREGFOOTPRINT(uint32_t val)
+{
+	return ((val) << A4XX_SP_VS_CTRL_REG0_HALFREGFOOTPRINT__SHIFT) & A4XX_SP_VS_CTRL_REG0_HALFREGFOOTPRINT__MASK;
+}
+#define A4XX_SP_VS_CTRL_REG0_FULLREGFOOTPRINT__MASK		0x0003fc00
+#define A4XX_SP_VS_CTRL_REG0_FULLREGFOOTPRINT__SHIFT		10
+static inline uint32_t A4XX_SP_VS_CTRL_REG0_FULLREGFOOTPRINT(uint32_t val)
+{
+	return ((val) << A4XX_SP_VS_CTRL_REG0_FULLREGFOOTPRINT__SHIFT) & A4XX_SP_VS_CTRL_REG0_FULLREGFOOTPRINT__MASK;
+}
+#define A4XX_SP_VS_CTRL_REG0_INOUTREGOVERLAP__MASK		0x000c0000
+#define A4XX_SP_VS_CTRL_REG0_INOUTREGOVERLAP__SHIFT		18
+static inline uint32_t A4XX_SP_VS_CTRL_REG0_INOUTREGOVERLAP(uint32_t val)
+{
+	return ((val) << A4XX_SP_VS_CTRL_REG0_INOUTREGOVERLAP__SHIFT) & A4XX_SP_VS_CTRL_REG0_INOUTREGOVERLAP__MASK;
+}
+#define A4XX_SP_VS_CTRL_REG0_THREADSIZE__MASK			0x00100000
+#define A4XX_SP_VS_CTRL_REG0_THREADSIZE__SHIFT			20
+static inline uint32_t A4XX_SP_VS_CTRL_REG0_THREADSIZE(enum a3xx_threadsize val)
+{
+	return ((val) << A4XX_SP_VS_CTRL_REG0_THREADSIZE__SHIFT) & A4XX_SP_VS_CTRL_REG0_THREADSIZE__MASK;
+}
+#define A4XX_SP_VS_CTRL_REG0_SUPERTHREADMODE			0x00200000
+#define A4XX_SP_VS_CTRL_REG0_PIXLODENABLE			0x00400000
+
+#define REG_A4XX_SP_VS_CTRL_REG1				0x000022c5
+#define A4XX_SP_VS_CTRL_REG1_CONSTLENGTH__MASK			0x000000ff
+#define A4XX_SP_VS_CTRL_REG1_CONSTLENGTH__SHIFT			0
+static inline uint32_t A4XX_SP_VS_CTRL_REG1_CONSTLENGTH(uint32_t val)
+{
+	return ((val) << A4XX_SP_VS_CTRL_REG1_CONSTLENGTH__SHIFT) & A4XX_SP_VS_CTRL_REG1_CONSTLENGTH__MASK;
+}
+#define A4XX_SP_VS_CTRL_REG1_INITIALOUTSTANDING__MASK		0x7f000000
+#define A4XX_SP_VS_CTRL_REG1_INITIALOUTSTANDING__SHIFT		24
+static inline uint32_t A4XX_SP_VS_CTRL_REG1_INITIALOUTSTANDING(uint32_t val)
+{
+	return ((val) << A4XX_SP_VS_CTRL_REG1_INITIALOUTSTANDING__SHIFT) & A4XX_SP_VS_CTRL_REG1_INITIALOUTSTANDING__MASK;
+}
+
+#define REG_A4XX_SP_VS_PARAM_REG				0x000022c6
+#define A4XX_SP_VS_PARAM_REG_POSREGID__MASK			0x000000ff
+#define A4XX_SP_VS_PARAM_REG_POSREGID__SHIFT			0
+static inline uint32_t A4XX_SP_VS_PARAM_REG_POSREGID(uint32_t val)
+{
+	return ((val) << A4XX_SP_VS_PARAM_REG_POSREGID__SHIFT) & A4XX_SP_VS_PARAM_REG_POSREGID__MASK;
+}
+#define A4XX_SP_VS_PARAM_REG_PSIZEREGID__MASK			0x0000ff00
+#define A4XX_SP_VS_PARAM_REG_PSIZEREGID__SHIFT			8
+static inline uint32_t A4XX_SP_VS_PARAM_REG_PSIZEREGID(uint32_t val)
+{
+	return ((val) << A4XX_SP_VS_PARAM_REG_PSIZEREGID__SHIFT) & A4XX_SP_VS_PARAM_REG_PSIZEREGID__MASK;
+}
+#define A4XX_SP_VS_PARAM_REG_TOTALVSOUTVAR__MASK		0xfff00000
+#define A4XX_SP_VS_PARAM_REG_TOTALVSOUTVAR__SHIFT		20
+static inline uint32_t A4XX_SP_VS_PARAM_REG_TOTALVSOUTVAR(uint32_t val)
+{
+	return ((val) << A4XX_SP_VS_PARAM_REG_TOTALVSOUTVAR__SHIFT) & A4XX_SP_VS_PARAM_REG_TOTALVSOUTVAR__MASK;
+}
+
+static inline uint32_t REG_A4XX_SP_VS_OUT(uint32_t i0) { return 0x000022c7 + 0x1*i0; }
+
+static inline uint32_t REG_A4XX_SP_VS_OUT_REG(uint32_t i0) { return 0x000022c7 + 0x1*i0; }
+#define A4XX_SP_VS_OUT_REG_A_REGID__MASK			0x000001ff
+#define A4XX_SP_VS_OUT_REG_A_REGID__SHIFT			0
+static inline uint32_t A4XX_SP_VS_OUT_REG_A_REGID(uint32_t val)
+{
+	return ((val) << A4XX_SP_VS_OUT_REG_A_REGID__SHIFT) & A4XX_SP_VS_OUT_REG_A_REGID__MASK;
+}
+#define A4XX_SP_VS_OUT_REG_A_COMPMASK__MASK			0x00001e00
+#define A4XX_SP_VS_OUT_REG_A_COMPMASK__SHIFT			9
+static inline uint32_t A4XX_SP_VS_OUT_REG_A_COMPMASK(uint32_t val)
+{
+	return ((val) << A4XX_SP_VS_OUT_REG_A_COMPMASK__SHIFT) & A4XX_SP_VS_OUT_REG_A_COMPMASK__MASK;
+}
+#define A4XX_SP_VS_OUT_REG_B_REGID__MASK			0x01ff0000
+#define A4XX_SP_VS_OUT_REG_B_REGID__SHIFT			16
+static inline uint32_t A4XX_SP_VS_OUT_REG_B_REGID(uint32_t val)
+{
+	return ((val) << A4XX_SP_VS_OUT_REG_B_REGID__SHIFT) & A4XX_SP_VS_OUT_REG_B_REGID__MASK;
+}
+#define A4XX_SP_VS_OUT_REG_B_COMPMASK__MASK			0x1e000000
+#define A4XX_SP_VS_OUT_REG_B_COMPMASK__SHIFT			25
+static inline uint32_t A4XX_SP_VS_OUT_REG_B_COMPMASK(uint32_t val)
+{
+	return ((val) << A4XX_SP_VS_OUT_REG_B_COMPMASK__SHIFT) & A4XX_SP_VS_OUT_REG_B_COMPMASK__MASK;
+}
+
+static inline uint32_t REG_A4XX_SP_VS_VPC_DST(uint32_t i0) { return 0x000022d8 + 0x1*i0; }
+
+static inline uint32_t REG_A4XX_SP_VS_VPC_DST_REG(uint32_t i0) { return 0x000022d8 + 0x1*i0; }
+#define A4XX_SP_VS_VPC_DST_REG_OUTLOC0__MASK			0x000000ff
+#define A4XX_SP_VS_VPC_DST_REG_OUTLOC0__SHIFT			0
+static inline uint32_t A4XX_SP_VS_VPC_DST_REG_OUTLOC0(uint32_t val)
+{
+	return ((val) << A4XX_SP_VS_VPC_DST_REG_OUTLOC0__SHIFT) & A4XX_SP_VS_VPC_DST_REG_OUTLOC0__MASK;
+}
+#define A4XX_SP_VS_VPC_DST_REG_OUTLOC1__MASK			0x0000ff00
+#define A4XX_SP_VS_VPC_DST_REG_OUTLOC1__SHIFT			8
+static inline uint32_t A4XX_SP_VS_VPC_DST_REG_OUTLOC1(uint32_t val)
+{
+	return ((val) << A4XX_SP_VS_VPC_DST_REG_OUTLOC1__SHIFT) & A4XX_SP_VS_VPC_DST_REG_OUTLOC1__MASK;
+}
+#define A4XX_SP_VS_VPC_DST_REG_OUTLOC2__MASK			0x00ff0000
+#define A4XX_SP_VS_VPC_DST_REG_OUTLOC2__SHIFT			16
+static inline uint32_t A4XX_SP_VS_VPC_DST_REG_OUTLOC2(uint32_t val)
+{
+	return ((val) << A4XX_SP_VS_VPC_DST_REG_OUTLOC2__SHIFT) & A4XX_SP_VS_VPC_DST_REG_OUTLOC2__MASK;
+}
+#define A4XX_SP_VS_VPC_DST_REG_OUTLOC3__MASK			0xff000000
+#define A4XX_SP_VS_VPC_DST_REG_OUTLOC3__SHIFT			24
+static inline uint32_t A4XX_SP_VS_VPC_DST_REG_OUTLOC3(uint32_t val)
+{
+	return ((val) << A4XX_SP_VS_VPC_DST_REG_OUTLOC3__SHIFT) & A4XX_SP_VS_VPC_DST_REG_OUTLOC3__MASK;
+}
+
+#define REG_A4XX_SP_VS_OBJ_OFFSET_REG				0x000022e0
+#define A4XX_SP_VS_OBJ_OFFSET_REG_CONSTOBJECTOFFSET__MASK	0x01ff0000
+#define A4XX_SP_VS_OBJ_OFFSET_REG_CONSTOBJECTOFFSET__SHIFT	16
+static inline uint32_t A4XX_SP_VS_OBJ_OFFSET_REG_CONSTOBJECTOFFSET(uint32_t val)
+{
+	return ((val) << A4XX_SP_VS_OBJ_OFFSET_REG_CONSTOBJECTOFFSET__SHIFT) & A4XX_SP_VS_OBJ_OFFSET_REG_CONSTOBJECTOFFSET__MASK;
+}
+#define A4XX_SP_VS_OBJ_OFFSET_REG_SHADEROBJOFFSET__MASK		0xfe000000
+#define A4XX_SP_VS_OBJ_OFFSET_REG_SHADEROBJOFFSET__SHIFT	25
+static inline uint32_t A4XX_SP_VS_OBJ_OFFSET_REG_SHADEROBJOFFSET(uint32_t val)
+{
+	return ((val) << A4XX_SP_VS_OBJ_OFFSET_REG_SHADEROBJOFFSET__SHIFT) & A4XX_SP_VS_OBJ_OFFSET_REG_SHADEROBJOFFSET__MASK;
+}
+
+#define REG_A4XX_SP_VS_OBJ_START				0x000022e1
+
+#define REG_A4XX_SP_VS_PVT_MEM_PARAM				0x000022e2
+
+#define REG_A4XX_SP_VS_PVT_MEM_ADDR				0x000022e3
+
+#define REG_A4XX_SP_VS_LENGTH_REG				0x000022e5
+
+#define REG_A4XX_SP_FS_CTRL_REG0				0x000022e8
+#define A4XX_SP_FS_CTRL_REG0_THREADMODE__MASK			0x00000001
+#define A4XX_SP_FS_CTRL_REG0_THREADMODE__SHIFT			0
+static inline uint32_t A4XX_SP_FS_CTRL_REG0_THREADMODE(enum a3xx_threadmode val)
+{
+	return ((val) << A4XX_SP_FS_CTRL_REG0_THREADMODE__SHIFT) & A4XX_SP_FS_CTRL_REG0_THREADMODE__MASK;
+}
+#define A4XX_SP_FS_CTRL_REG0_VARYING				0x00000002
+#define A4XX_SP_FS_CTRL_REG0_CACHEINVALID			0x00000004
+#define A4XX_SP_FS_CTRL_REG0_HALFREGFOOTPRINT__MASK		0x000003f0
+#define A4XX_SP_FS_CTRL_REG0_HALFREGFOOTPRINT__SHIFT		4
+static inline uint32_t A4XX_SP_FS_CTRL_REG0_HALFREGFOOTPRINT(uint32_t val)
+{
+	return ((val) << A4XX_SP_FS_CTRL_REG0_HALFREGFOOTPRINT__SHIFT) & A4XX_SP_FS_CTRL_REG0_HALFREGFOOTPRINT__MASK;
+}
+#define A4XX_SP_FS_CTRL_REG0_FULLREGFOOTPRINT__MASK		0x0003fc00
+#define A4XX_SP_FS_CTRL_REG0_FULLREGFOOTPRINT__SHIFT		10
+static inline uint32_t A4XX_SP_FS_CTRL_REG0_FULLREGFOOTPRINT(uint32_t val)
+{
+	return ((val) << A4XX_SP_FS_CTRL_REG0_FULLREGFOOTPRINT__SHIFT) & A4XX_SP_FS_CTRL_REG0_FULLREGFOOTPRINT__MASK;
+}
+#define A4XX_SP_FS_CTRL_REG0_INOUTREGOVERLAP__MASK		0x000c0000
+#define A4XX_SP_FS_CTRL_REG0_INOUTREGOVERLAP__SHIFT		18
+static inline uint32_t A4XX_SP_FS_CTRL_REG0_INOUTREGOVERLAP(uint32_t val)
+{
+	return ((val) << A4XX_SP_FS_CTRL_REG0_INOUTREGOVERLAP__SHIFT) & A4XX_SP_FS_CTRL_REG0_INOUTREGOVERLAP__MASK;
+}
+#define A4XX_SP_FS_CTRL_REG0_THREADSIZE__MASK			0x00100000
+#define A4XX_SP_FS_CTRL_REG0_THREADSIZE__SHIFT			20
+static inline uint32_t A4XX_SP_FS_CTRL_REG0_THREADSIZE(enum a3xx_threadsize val)
+{
+	return ((val) << A4XX_SP_FS_CTRL_REG0_THREADSIZE__SHIFT) & A4XX_SP_FS_CTRL_REG0_THREADSIZE__MASK;
+}
+#define A4XX_SP_FS_CTRL_REG0_SUPERTHREADMODE			0x00200000
+#define A4XX_SP_FS_CTRL_REG0_PIXLODENABLE			0x00400000
+
+#define REG_A4XX_SP_FS_CTRL_REG1				0x000022e9
+#define A4XX_SP_FS_CTRL_REG1_CONSTLENGTH__MASK			0x000000ff
+#define A4XX_SP_FS_CTRL_REG1_CONSTLENGTH__SHIFT			0
+static inline uint32_t A4XX_SP_FS_CTRL_REG1_CONSTLENGTH(uint32_t val)
+{
+	return ((val) << A4XX_SP_FS_CTRL_REG1_CONSTLENGTH__SHIFT) & A4XX_SP_FS_CTRL_REG1_CONSTLENGTH__MASK;
+}
+#define A4XX_SP_FS_CTRL_REG1_VARYING				0x00100000
+
+#define REG_A4XX_SP_FS_OBJ_OFFSET_REG				0x000022ea
+#define A4XX_SP_FS_OBJ_OFFSET_REG_CONSTOBJECTOFFSET__MASK	0x01ff0000
+#define A4XX_SP_FS_OBJ_OFFSET_REG_CONSTOBJECTOFFSET__SHIFT	16
+static inline uint32_t A4XX_SP_FS_OBJ_OFFSET_REG_CONSTOBJECTOFFSET(uint32_t val)
+{
+	return ((val) << A4XX_SP_FS_OBJ_OFFSET_REG_CONSTOBJECTOFFSET__SHIFT) & A4XX_SP_FS_OBJ_OFFSET_REG_CONSTOBJECTOFFSET__MASK;
+}
+#define A4XX_SP_FS_OBJ_OFFSET_REG_SHADEROBJOFFSET__MASK		0xfe000000
+#define A4XX_SP_FS_OBJ_OFFSET_REG_SHADEROBJOFFSET__SHIFT	25
+static inline uint32_t A4XX_SP_FS_OBJ_OFFSET_REG_SHADEROBJOFFSET(uint32_t val)
+{
+	return ((val) << A4XX_SP_FS_OBJ_OFFSET_REG_SHADEROBJOFFSET__SHIFT) & A4XX_SP_FS_OBJ_OFFSET_REG_SHADEROBJOFFSET__MASK;
+}
+
+#define REG_A4XX_SP_FS_OBJ_START				0x000022eb
+
+#define REG_A4XX_SP_FS_PVT_MEM_PARAM				0x000022ec
+
+#define REG_A4XX_SP_FS_PVT_MEM_ADDR				0x000022ed
+
+#define REG_A4XX_SP_FS_LENGTH_REG				0x000022ef
+
+#define REG_A4XX_SP_FS_OUTPUT_REG				0x000022f0
+#define A4XX_SP_FS_OUTPUT_REG_DEPTH_ENABLE			0x00000080
+#define A4XX_SP_FS_OUTPUT_REG_DEPTH_REGID__MASK			0x0000ff00
+#define A4XX_SP_FS_OUTPUT_REG_DEPTH_REGID__SHIFT		8
+static inline uint32_t A4XX_SP_FS_OUTPUT_REG_DEPTH_REGID(uint32_t val)
+{
+	return ((val) << A4XX_SP_FS_OUTPUT_REG_DEPTH_REGID__SHIFT) & A4XX_SP_FS_OUTPUT_REG_DEPTH_REGID__MASK;
+}
+
+static inline uint32_t REG_A4XX_SP_FS_MRT(uint32_t i0) { return 0x000022f1 + 0x1*i0; }
+
+static inline uint32_t REG_A4XX_SP_FS_MRT_REG(uint32_t i0) { return 0x000022f1 + 0x1*i0; }
+#define A4XX_SP_FS_MRT_REG_REGID__MASK				0x000000ff
+#define A4XX_SP_FS_MRT_REG_REGID__SHIFT				0
+static inline uint32_t A4XX_SP_FS_MRT_REG_REGID(uint32_t val)
+{
+	return ((val) << A4XX_SP_FS_MRT_REG_REGID__SHIFT) & A4XX_SP_FS_MRT_REG_REGID__MASK;
+}
+#define A4XX_SP_FS_MRT_REG_HALF_PRECISION			0x00000100
+#define A4XX_SP_FS_MRT_REG_MRTFORMAT__MASK			0x0003f000
+#define A4XX_SP_FS_MRT_REG_MRTFORMAT__SHIFT			12
+static inline uint32_t A4XX_SP_FS_MRT_REG_MRTFORMAT(enum a4xx_color_fmt val)
+{
+	return ((val) << A4XX_SP_FS_MRT_REG_MRTFORMAT__SHIFT) & A4XX_SP_FS_MRT_REG_MRTFORMAT__MASK;
+}
+
+#define REG_A4XX_SP_HS_OBJ_OFFSET_REG				0x0000230d
+#define A4XX_SP_HS_OBJ_OFFSET_REG_CONSTOBJECTOFFSET__MASK	0x01ff0000
+#define A4XX_SP_HS_OBJ_OFFSET_REG_CONSTOBJECTOFFSET__SHIFT	16
+static inline uint32_t A4XX_SP_HS_OBJ_OFFSET_REG_CONSTOBJECTOFFSET(uint32_t val)
+{
+	return ((val) << A4XX_SP_HS_OBJ_OFFSET_REG_CONSTOBJECTOFFSET__SHIFT) & A4XX_SP_HS_OBJ_OFFSET_REG_CONSTOBJECTOFFSET__MASK;
+}
+#define A4XX_SP_HS_OBJ_OFFSET_REG_SHADEROBJOFFSET__MASK		0xfe000000
+#define A4XX_SP_HS_OBJ_OFFSET_REG_SHADEROBJOFFSET__SHIFT	25
+static inline uint32_t A4XX_SP_HS_OBJ_OFFSET_REG_SHADEROBJOFFSET(uint32_t val)
+{
+	return ((val) << A4XX_SP_HS_OBJ_OFFSET_REG_SHADEROBJOFFSET__SHIFT) & A4XX_SP_HS_OBJ_OFFSET_REG_SHADEROBJOFFSET__MASK;
+}
+
+#define REG_A4XX_SP_DS_OBJ_OFFSET_REG				0x00002334
+#define A4XX_SP_DS_OBJ_OFFSET_REG_CONSTOBJECTOFFSET__MASK	0x01ff0000
+#define A4XX_SP_DS_OBJ_OFFSET_REG_CONSTOBJECTOFFSET__SHIFT	16
+static inline uint32_t A4XX_SP_DS_OBJ_OFFSET_REG_CONSTOBJECTOFFSET(uint32_t val)
+{
+	return ((val) << A4XX_SP_DS_OBJ_OFFSET_REG_CONSTOBJECTOFFSET__SHIFT) & A4XX_SP_DS_OBJ_OFFSET_REG_CONSTOBJECTOFFSET__MASK;
+}
+#define A4XX_SP_DS_OBJ_OFFSET_REG_SHADEROBJOFFSET__MASK		0xfe000000
+#define A4XX_SP_DS_OBJ_OFFSET_REG_SHADEROBJOFFSET__SHIFT	25
+static inline uint32_t A4XX_SP_DS_OBJ_OFFSET_REG_SHADEROBJOFFSET(uint32_t val)
+{
+	return ((val) << A4XX_SP_DS_OBJ_OFFSET_REG_SHADEROBJOFFSET__SHIFT) & A4XX_SP_DS_OBJ_OFFSET_REG_SHADEROBJOFFSET__MASK;
+}
+
+#define REG_A4XX_SP_GS_OBJ_OFFSET_REG				0x0000235b
+#define A4XX_SP_GS_OBJ_OFFSET_REG_CONSTOBJECTOFFSET__MASK	0x01ff0000
+#define A4XX_SP_GS_OBJ_OFFSET_REG_CONSTOBJECTOFFSET__SHIFT	16
+static inline uint32_t A4XX_SP_GS_OBJ_OFFSET_REG_CONSTOBJECTOFFSET(uint32_t val)
+{
+	return ((val) << A4XX_SP_GS_OBJ_OFFSET_REG_CONSTOBJECTOFFSET__SHIFT) & A4XX_SP_GS_OBJ_OFFSET_REG_CONSTOBJECTOFFSET__MASK;
+}
+#define A4XX_SP_GS_OBJ_OFFSET_REG_SHADEROBJOFFSET__MASK		0xfe000000
+#define A4XX_SP_GS_OBJ_OFFSET_REG_SHADEROBJOFFSET__SHIFT	25
+static inline uint32_t A4XX_SP_GS_OBJ_OFFSET_REG_SHADEROBJOFFSET(uint32_t val)
+{
+	return ((val) << A4XX_SP_GS_OBJ_OFFSET_REG_SHADEROBJOFFSET__SHIFT) & A4XX_SP_GS_OBJ_OFFSET_REG_SHADEROBJOFFSET__MASK;
+}
+
+#define REG_A4XX_SP_GS_LENGTH_REG				0x00002360
+
+#define REG_A4XX_VPC_DEBUG_RAM_SEL				0x00000e60
+
+#define REG_A4XX_VPC_DEBUG_RAM_READ				0x00000e61
+
+#define REG_A4XX_VPC_DEBUG_ECO_CONTROL				0x00000e64
+
+#define REG_A4XX_VPC_PERFCTR_VPC_SEL_3				0x00000e68
+
+#define REG_A4XX_VPC_ATTR					0x00002140
+#define A4XX_VPC_ATTR_TOTALATTR__MASK				0x000001ff
+#define A4XX_VPC_ATTR_TOTALATTR__SHIFT				0
+static inline uint32_t A4XX_VPC_ATTR_TOTALATTR(uint32_t val)
+{
+	return ((val) << A4XX_VPC_ATTR_TOTALATTR__SHIFT) & A4XX_VPC_ATTR_TOTALATTR__MASK;
+}
+#define A4XX_VPC_ATTR_PSIZE					0x00000200
+#define A4XX_VPC_ATTR_THRDASSIGN__MASK				0x00003000
+#define A4XX_VPC_ATTR_THRDASSIGN__SHIFT				12
+static inline uint32_t A4XX_VPC_ATTR_THRDASSIGN(uint32_t val)
+{
+	return ((val) << A4XX_VPC_ATTR_THRDASSIGN__SHIFT) & A4XX_VPC_ATTR_THRDASSIGN__MASK;
+}
+#define A4XX_VPC_ATTR_ENABLE					0x02000000
+
+#define REG_A4XX_VPC_PACK					0x00002141
+#define A4XX_VPC_PACK_NUMBYPASSVAR__MASK			0x000000ff
+#define A4XX_VPC_PACK_NUMBYPASSVAR__SHIFT			0
+static inline uint32_t A4XX_VPC_PACK_NUMBYPASSVAR(uint32_t val)
+{
+	return ((val) << A4XX_VPC_PACK_NUMBYPASSVAR__SHIFT) & A4XX_VPC_PACK_NUMBYPASSVAR__MASK;
+}
+#define A4XX_VPC_PACK_NUMFPNONPOSVAR__MASK			0x0000ff00
+#define A4XX_VPC_PACK_NUMFPNONPOSVAR__SHIFT			8
+static inline uint32_t A4XX_VPC_PACK_NUMFPNONPOSVAR(uint32_t val)
+{
+	return ((val) << A4XX_VPC_PACK_NUMFPNONPOSVAR__SHIFT) & A4XX_VPC_PACK_NUMFPNONPOSVAR__MASK;
+}
+#define A4XX_VPC_PACK_NUMNONPOSVSVAR__MASK			0x00ff0000
+#define A4XX_VPC_PACK_NUMNONPOSVSVAR__SHIFT			16
+static inline uint32_t A4XX_VPC_PACK_NUMNONPOSVSVAR(uint32_t val)
+{
+	return ((val) << A4XX_VPC_PACK_NUMNONPOSVSVAR__SHIFT) & A4XX_VPC_PACK_NUMNONPOSVSVAR__MASK;
+}
+
+static inline uint32_t REG_A4XX_VPC_VARYING_INTERP(uint32_t i0) { return 0x00002142 + 0x1*i0; }
+
+static inline uint32_t REG_A4XX_VPC_VARYING_INTERP_MODE(uint32_t i0) { return 0x00002142 + 0x1*i0; }
+
+static inline uint32_t REG_A4XX_VPC_VARYING_PS_REPL(uint32_t i0) { return 0x0000214a + 0x1*i0; }
+
+static inline uint32_t REG_A4XX_VPC_VARYING_PS_REPL_MODE(uint32_t i0) { return 0x0000214a + 0x1*i0; }
+
+#define REG_A4XX_VPC_SO_FLUSH_WADDR_3				0x0000216e
+
+#define REG_A4XX_VSC_BIN_SIZE					0x00000c00
+#define A4XX_VSC_BIN_SIZE_WIDTH__MASK				0x0000001f
+#define A4XX_VSC_BIN_SIZE_WIDTH__SHIFT				0
+static inline uint32_t A4XX_VSC_BIN_SIZE_WIDTH(uint32_t val)
+{
+	return ((val >> 5) << A4XX_VSC_BIN_SIZE_WIDTH__SHIFT) & A4XX_VSC_BIN_SIZE_WIDTH__MASK;
+}
+#define A4XX_VSC_BIN_SIZE_HEIGHT__MASK				0x000003e0
+#define A4XX_VSC_BIN_SIZE_HEIGHT__SHIFT				5
+static inline uint32_t A4XX_VSC_BIN_SIZE_HEIGHT(uint32_t val)
+{
+	return ((val >> 5) << A4XX_VSC_BIN_SIZE_HEIGHT__SHIFT) & A4XX_VSC_BIN_SIZE_HEIGHT__MASK;
+}
+
+#define REG_A4XX_VSC_SIZE_ADDRESS				0x00000c01
+
+#define REG_A4XX_VSC_SIZE_ADDRESS2				0x00000c02
+
+#define REG_A4XX_VSC_DEBUG_ECO_CONTROL				0x00000c03
+
+static inline uint32_t REG_A4XX_VSC_PIPE_CONFIG(uint32_t i0) { return 0x00000c08 + 0x1*i0; }
+
+static inline uint32_t REG_A4XX_VSC_PIPE_CONFIG_REG(uint32_t i0) { return 0x00000c08 + 0x1*i0; }
+#define A4XX_VSC_PIPE_CONFIG_REG_X__MASK			0x000003ff
+#define A4XX_VSC_PIPE_CONFIG_REG_X__SHIFT			0
+static inline uint32_t A4XX_VSC_PIPE_CONFIG_REG_X(uint32_t val)
+{
+	return ((val) << A4XX_VSC_PIPE_CONFIG_REG_X__SHIFT) & A4XX_VSC_PIPE_CONFIG_REG_X__MASK;
+}
+#define A4XX_VSC_PIPE_CONFIG_REG_Y__MASK			0x000ffc00
+#define A4XX_VSC_PIPE_CONFIG_REG_Y__SHIFT			10
+static inline uint32_t A4XX_VSC_PIPE_CONFIG_REG_Y(uint32_t val)
+{
+	return ((val) << A4XX_VSC_PIPE_CONFIG_REG_Y__SHIFT) & A4XX_VSC_PIPE_CONFIG_REG_Y__MASK;
+}
+#define A4XX_VSC_PIPE_CONFIG_REG_W__MASK			0x00f00000
+#define A4XX_VSC_PIPE_CONFIG_REG_W__SHIFT			20
+static inline uint32_t A4XX_VSC_PIPE_CONFIG_REG_W(uint32_t val)
+{
+	return ((val) << A4XX_VSC_PIPE_CONFIG_REG_W__SHIFT) & A4XX_VSC_PIPE_CONFIG_REG_W__MASK;
+}
+#define A4XX_VSC_PIPE_CONFIG_REG_H__MASK			0x0f000000
+#define A4XX_VSC_PIPE_CONFIG_REG_H__SHIFT			24
+static inline uint32_t A4XX_VSC_PIPE_CONFIG_REG_H(uint32_t val)
+{
+	return ((val) << A4XX_VSC_PIPE_CONFIG_REG_H__SHIFT) & A4XX_VSC_PIPE_CONFIG_REG_H__MASK;
+}
+
+static inline uint32_t REG_A4XX_VSC_PIPE_DATA_ADDRESS(uint32_t i0) { return 0x00000c10 + 0x1*i0; }
+
+static inline uint32_t REG_A4XX_VSC_PIPE_DATA_ADDRESS_REG(uint32_t i0) { return 0x00000c10 + 0x1*i0; }
+
+static inline uint32_t REG_A4XX_VSC_PIPE_DATA_LENGTH(uint32_t i0) { return 0x00000c18 + 0x1*i0; }
+
+static inline uint32_t REG_A4XX_VSC_PIPE_DATA_LENGTH_REG(uint32_t i0) { return 0x00000c18 + 0x1*i0; }
+
+#define REG_A4XX_VSC_PIPE_PARTIAL_POSN_1			0x00000c41
+
+#define REG_A4XX_VSC_PERFCTR_VSC_SEL_0				0x00000c50
+
+#define REG_A4XX_VSC_PERFCTR_VSC_SEL_1				0x00000c51
+
+#define REG_A4XX_VFD_DEBUG_CONTROL				0x00000e40
+
+#define REG_A4XX_VFD_PERFCTR_VFD_SEL_7				0x00000e4a
+
+#define REG_A4XX_VFD_CONTROL_0					0x00002200
+#define A4XX_VFD_CONTROL_0_TOTALATTRTOVS__MASK			0x000000ff
+#define A4XX_VFD_CONTROL_0_TOTALATTRTOVS__SHIFT			0
+static inline uint32_t A4XX_VFD_CONTROL_0_TOTALATTRTOVS(uint32_t val)
+{
+	return ((val) << A4XX_VFD_CONTROL_0_TOTALATTRTOVS__SHIFT) & A4XX_VFD_CONTROL_0_TOTALATTRTOVS__MASK;
+}
+#define A4XX_VFD_CONTROL_0_BYPASSATTROVS__MASK			0x0001fe00
+#define A4XX_VFD_CONTROL_0_BYPASSATTROVS__SHIFT			9
+static inline uint32_t A4XX_VFD_CONTROL_0_BYPASSATTROVS(uint32_t val)
+{
+	return ((val) << A4XX_VFD_CONTROL_0_BYPASSATTROVS__SHIFT) & A4XX_VFD_CONTROL_0_BYPASSATTROVS__MASK;
+}
+#define A4XX_VFD_CONTROL_0_STRMDECINSTRCNT__MASK		0x03f00000
+#define A4XX_VFD_CONTROL_0_STRMDECINSTRCNT__SHIFT		20
+static inline uint32_t A4XX_VFD_CONTROL_0_STRMDECINSTRCNT(uint32_t val)
+{
+	return ((val) << A4XX_VFD_CONTROL_0_STRMDECINSTRCNT__SHIFT) & A4XX_VFD_CONTROL_0_STRMDECINSTRCNT__MASK;
+}
+#define A4XX_VFD_CONTROL_0_STRMFETCHINSTRCNT__MASK		0xfc000000
+#define A4XX_VFD_CONTROL_0_STRMFETCHINSTRCNT__SHIFT		26
+static inline uint32_t A4XX_VFD_CONTROL_0_STRMFETCHINSTRCNT(uint32_t val)
+{
+	return ((val) << A4XX_VFD_CONTROL_0_STRMFETCHINSTRCNT__SHIFT) & A4XX_VFD_CONTROL_0_STRMFETCHINSTRCNT__MASK;
+}
+
+#define REG_A4XX_VFD_CONTROL_1					0x00002201
+#define A4XX_VFD_CONTROL_1_MAXSTORAGE__MASK			0x0000ffff
+#define A4XX_VFD_CONTROL_1_MAXSTORAGE__SHIFT			0
+static inline uint32_t A4XX_VFD_CONTROL_1_MAXSTORAGE(uint32_t val)
+{
+	return ((val) << A4XX_VFD_CONTROL_1_MAXSTORAGE__SHIFT) & A4XX_VFD_CONTROL_1_MAXSTORAGE__MASK;
+}
+#define A4XX_VFD_CONTROL_1_REGID4VTX__MASK			0x00ff0000
+#define A4XX_VFD_CONTROL_1_REGID4VTX__SHIFT			16
+static inline uint32_t A4XX_VFD_CONTROL_1_REGID4VTX(uint32_t val)
+{
+	return ((val) << A4XX_VFD_CONTROL_1_REGID4VTX__SHIFT) & A4XX_VFD_CONTROL_1_REGID4VTX__MASK;
+}
+#define A4XX_VFD_CONTROL_1_REGID4INST__MASK			0xff000000
+#define A4XX_VFD_CONTROL_1_REGID4INST__SHIFT			24
+static inline uint32_t A4XX_VFD_CONTROL_1_REGID4INST(uint32_t val)
+{
+	return ((val) << A4XX_VFD_CONTROL_1_REGID4INST__SHIFT) & A4XX_VFD_CONTROL_1_REGID4INST__MASK;
+}
+
+#define REG_A4XX_VFD_CONTROL_2					0x00002202
+
+#define REG_A4XX_VFD_CONTROL_3					0x00002203
+
+#define REG_A4XX_VFD_CONTROL_4					0x00002204
+
+#define REG_A4XX_VFD_INDEX_OFFSET				0x00002208
+
+static inline uint32_t REG_A4XX_VFD_FETCH(uint32_t i0) { return 0x0000220a + 0x4*i0; }
+
+static inline uint32_t REG_A4XX_VFD_FETCH_INSTR_0(uint32_t i0) { return 0x0000220a + 0x4*i0; }
+#define A4XX_VFD_FETCH_INSTR_0_FETCHSIZE__MASK			0x0000007f
+#define A4XX_VFD_FETCH_INSTR_0_FETCHSIZE__SHIFT			0
+static inline uint32_t A4XX_VFD_FETCH_INSTR_0_FETCHSIZE(uint32_t val)
+{
+	return ((val) << A4XX_VFD_FETCH_INSTR_0_FETCHSIZE__SHIFT) & A4XX_VFD_FETCH_INSTR_0_FETCHSIZE__MASK;
+}
+#define A4XX_VFD_FETCH_INSTR_0_BUFSTRIDE__MASK			0x0001ff80
+#define A4XX_VFD_FETCH_INSTR_0_BUFSTRIDE__SHIFT			7
+static inline uint32_t A4XX_VFD_FETCH_INSTR_0_BUFSTRIDE(uint32_t val)
+{
+	return ((val) << A4XX_VFD_FETCH_INSTR_0_BUFSTRIDE__SHIFT) & A4XX_VFD_FETCH_INSTR_0_BUFSTRIDE__MASK;
+}
+#define A4XX_VFD_FETCH_INSTR_0_SWITCHNEXT			0x00080000
+#define A4XX_VFD_FETCH_INSTR_0_STEPRATE__MASK			0xff000000
+#define A4XX_VFD_FETCH_INSTR_0_STEPRATE__SHIFT			24
+static inline uint32_t A4XX_VFD_FETCH_INSTR_0_STEPRATE(uint32_t val)
+{
+	return ((val) << A4XX_VFD_FETCH_INSTR_0_STEPRATE__SHIFT) & A4XX_VFD_FETCH_INSTR_0_STEPRATE__MASK;
+}
+
+static inline uint32_t REG_A4XX_VFD_FETCH_INSTR_1(uint32_t i0) { return 0x0000220b + 0x4*i0; }
+
+static inline uint32_t REG_A4XX_VFD_FETCH_INSTR_2(uint32_t i0) { return 0x0000220c + 0x4*i0; }
+#define A4XX_VFD_FETCH_INSTR_2_SIZE__MASK			0xfffffff0
+#define A4XX_VFD_FETCH_INSTR_2_SIZE__SHIFT			4
+static inline uint32_t A4XX_VFD_FETCH_INSTR_2_SIZE(uint32_t val)
+{
+	return ((val >> 4) << A4XX_VFD_FETCH_INSTR_2_SIZE__SHIFT) & A4XX_VFD_FETCH_INSTR_2_SIZE__MASK;
+}
+
+static inline uint32_t REG_A4XX_VFD_FETCH_INSTR_3(uint32_t i0) { return 0x0000220d + 0x4*i0; }
+
+static inline uint32_t REG_A4XX_VFD_DECODE(uint32_t i0) { return 0x0000228a + 0x1*i0; }
+
+static inline uint32_t REG_A4XX_VFD_DECODE_INSTR(uint32_t i0) { return 0x0000228a + 0x1*i0; }
+#define A4XX_VFD_DECODE_INSTR_WRITEMASK__MASK			0x0000000f
+#define A4XX_VFD_DECODE_INSTR_WRITEMASK__SHIFT			0
+static inline uint32_t A4XX_VFD_DECODE_INSTR_WRITEMASK(uint32_t val)
+{
+	return ((val) << A4XX_VFD_DECODE_INSTR_WRITEMASK__SHIFT) & A4XX_VFD_DECODE_INSTR_WRITEMASK__MASK;
+}
+#define A4XX_VFD_DECODE_INSTR_CONSTFILL				0x00000010
+#define A4XX_VFD_DECODE_INSTR_FORMAT__MASK			0x00000fc0
+#define A4XX_VFD_DECODE_INSTR_FORMAT__SHIFT			6
+static inline uint32_t A4XX_VFD_DECODE_INSTR_FORMAT(enum a4xx_vtx_fmt val)
+{
+	return ((val) << A4XX_VFD_DECODE_INSTR_FORMAT__SHIFT) & A4XX_VFD_DECODE_INSTR_FORMAT__MASK;
+}
+#define A4XX_VFD_DECODE_INSTR_REGID__MASK			0x000ff000
+#define A4XX_VFD_DECODE_INSTR_REGID__SHIFT			12
+static inline uint32_t A4XX_VFD_DECODE_INSTR_REGID(uint32_t val)
+{
+	return ((val) << A4XX_VFD_DECODE_INSTR_REGID__SHIFT) & A4XX_VFD_DECODE_INSTR_REGID__MASK;
+}
+#define A4XX_VFD_DECODE_INSTR_SWAP__MASK			0x00c00000
+#define A4XX_VFD_DECODE_INSTR_SWAP__SHIFT			22
+static inline uint32_t A4XX_VFD_DECODE_INSTR_SWAP(enum a3xx_color_swap val)
+{
+	return ((val) << A4XX_VFD_DECODE_INSTR_SWAP__SHIFT) & A4XX_VFD_DECODE_INSTR_SWAP__MASK;
+}
+#define A4XX_VFD_DECODE_INSTR_SHIFTCNT__MASK			0x1f000000
+#define A4XX_VFD_DECODE_INSTR_SHIFTCNT__SHIFT			24
+static inline uint32_t A4XX_VFD_DECODE_INSTR_SHIFTCNT(uint32_t val)
+{
+	return ((val) << A4XX_VFD_DECODE_INSTR_SHIFTCNT__SHIFT) & A4XX_VFD_DECODE_INSTR_SHIFTCNT__MASK;
+}
+#define A4XX_VFD_DECODE_INSTR_LASTCOMPVALID			0x20000000
+#define A4XX_VFD_DECODE_INSTR_SWITCHNEXT			0x40000000
+
+#define REG_A4XX_TPL1_DEBUG_ECO_CONTROL				0x00000f00
+
+#define REG_A4XX_TPL1_PERFCTR_TP_SEL_7				0x00000f0b
+
+#define REG_A4XX_TPL1_TP_TEX_OFFSET				0x00002380
+
+#define REG_A4XX_TPL1_TP_CS_TEXMEMOBJ_BASE_ADDR			0x000023a6
+
+#define REG_A4XX_GRAS_TSE_STATUS				0x00000c80
+
+#define REG_A4XX_GRAS_DEBUG_ECO_CONTROL				0x00000c81
+
+#define REG_A4XX_GRAS_PERFCTR_TSE_SEL_0				0x00000c88
+
+#define REG_A4XX_GRAS_PERFCTR_TSE_SEL_3				0x00000c8b
+
+#define REG_A4XX_GRAS_CL_CLIP_CNTL				0x00002000
+
+#define REG_A4XX_GRAS_CLEAR_CNTL				0x00002003
+#define A4XX_GRAS_CLEAR_CNTL_NOT_FASTCLEAR			0x00000001
+
+#define REG_A4XX_GRAS_CL_GB_CLIP_ADJ				0x00002004
+#define A4XX_GRAS_CL_GB_CLIP_ADJ_HORZ__MASK			0x000003ff
+#define A4XX_GRAS_CL_GB_CLIP_ADJ_HORZ__SHIFT			0
+static inline uint32_t A4XX_GRAS_CL_GB_CLIP_ADJ_HORZ(uint32_t val)
+{
+	return ((val) << A4XX_GRAS_CL_GB_CLIP_ADJ_HORZ__SHIFT) & A4XX_GRAS_CL_GB_CLIP_ADJ_HORZ__MASK;
+}
+#define A4XX_GRAS_CL_GB_CLIP_ADJ_VERT__MASK			0x000ffc00
+#define A4XX_GRAS_CL_GB_CLIP_ADJ_VERT__SHIFT			10
+static inline uint32_t A4XX_GRAS_CL_GB_CLIP_ADJ_VERT(uint32_t val)
+{
+	return ((val) << A4XX_GRAS_CL_GB_CLIP_ADJ_VERT__SHIFT) & A4XX_GRAS_CL_GB_CLIP_ADJ_VERT__MASK;
+}
+
+#define REG_A4XX_GRAS_CL_VPORT_XOFFSET_0			0x00002008
+#define A4XX_GRAS_CL_VPORT_XOFFSET_0__MASK			0xffffffff
+#define A4XX_GRAS_CL_VPORT_XOFFSET_0__SHIFT			0
+static inline uint32_t A4XX_GRAS_CL_VPORT_XOFFSET_0(float val)
+{
+	return ((fui(val)) << A4XX_GRAS_CL_VPORT_XOFFSET_0__SHIFT) & A4XX_GRAS_CL_VPORT_XOFFSET_0__MASK;
+}
+
+#define REG_A4XX_GRAS_CL_VPORT_XSCALE_0				0x00002009
+#define A4XX_GRAS_CL_VPORT_XSCALE_0__MASK			0xffffffff
+#define A4XX_GRAS_CL_VPORT_XSCALE_0__SHIFT			0
+static inline uint32_t A4XX_GRAS_CL_VPORT_XSCALE_0(float val)
+{
+	return ((fui(val)) << A4XX_GRAS_CL_VPORT_XSCALE_0__SHIFT) & A4XX_GRAS_CL_VPORT_XSCALE_0__MASK;
+}
+
+#define REG_A4XX_GRAS_CL_VPORT_YOFFSET_0			0x0000200a
+#define A4XX_GRAS_CL_VPORT_YOFFSET_0__MASK			0xffffffff
+#define A4XX_GRAS_CL_VPORT_YOFFSET_0__SHIFT			0
+static inline uint32_t A4XX_GRAS_CL_VPORT_YOFFSET_0(float val)
+{
+	return ((fui(val)) << A4XX_GRAS_CL_VPORT_YOFFSET_0__SHIFT) & A4XX_GRAS_CL_VPORT_YOFFSET_0__MASK;
+}
+
+#define REG_A4XX_GRAS_CL_VPORT_YSCALE_0				0x0000200b
+#define A4XX_GRAS_CL_VPORT_YSCALE_0__MASK			0xffffffff
+#define A4XX_GRAS_CL_VPORT_YSCALE_0__SHIFT			0
+static inline uint32_t A4XX_GRAS_CL_VPORT_YSCALE_0(float val)
+{
+	return ((fui(val)) << A4XX_GRAS_CL_VPORT_YSCALE_0__SHIFT) & A4XX_GRAS_CL_VPORT_YSCALE_0__MASK;
+}
+
+#define REG_A4XX_GRAS_CL_VPORT_ZOFFSET_0			0x0000200c
+#define A4XX_GRAS_CL_VPORT_ZOFFSET_0__MASK			0xffffffff
+#define A4XX_GRAS_CL_VPORT_ZOFFSET_0__SHIFT			0
+static inline uint32_t A4XX_GRAS_CL_VPORT_ZOFFSET_0(float val)
+{
+	return ((fui(val)) << A4XX_GRAS_CL_VPORT_ZOFFSET_0__SHIFT) & A4XX_GRAS_CL_VPORT_ZOFFSET_0__MASK;
+}
+
+#define REG_A4XX_GRAS_CL_VPORT_ZSCALE_0				0x0000200d
+#define A4XX_GRAS_CL_VPORT_ZSCALE_0__MASK			0xffffffff
+#define A4XX_GRAS_CL_VPORT_ZSCALE_0__SHIFT			0
+static inline uint32_t A4XX_GRAS_CL_VPORT_ZSCALE_0(float val)
+{
+	return ((fui(val)) << A4XX_GRAS_CL_VPORT_ZSCALE_0__SHIFT) & A4XX_GRAS_CL_VPORT_ZSCALE_0__MASK;
+}
+
+#define REG_A4XX_GRAS_SU_POINT_MINMAX				0x00002070
+#define A4XX_GRAS_SU_POINT_MINMAX_MIN__MASK			0x0000ffff
+#define A4XX_GRAS_SU_POINT_MINMAX_MIN__SHIFT			0
+static inline uint32_t A4XX_GRAS_SU_POINT_MINMAX_MIN(float val)
+{
+	return ((((uint32_t)(val * 16.0))) << A4XX_GRAS_SU_POINT_MINMAX_MIN__SHIFT) & A4XX_GRAS_SU_POINT_MINMAX_MIN__MASK;
+}
+#define A4XX_GRAS_SU_POINT_MINMAX_MAX__MASK			0xffff0000
+#define A4XX_GRAS_SU_POINT_MINMAX_MAX__SHIFT			16
+static inline uint32_t A4XX_GRAS_SU_POINT_MINMAX_MAX(float val)
+{
+	return ((((uint32_t)(val * 16.0))) << A4XX_GRAS_SU_POINT_MINMAX_MAX__SHIFT) & A4XX_GRAS_SU_POINT_MINMAX_MAX__MASK;
+}
+
+#define REG_A4XX_GRAS_SU_POINT_SIZE				0x00002071
+#define A4XX_GRAS_SU_POINT_SIZE__MASK				0xffffffff
+#define A4XX_GRAS_SU_POINT_SIZE__SHIFT				0
+static inline uint32_t A4XX_GRAS_SU_POINT_SIZE(float val)
+{
+	return ((((int32_t)(val * 16.0))) << A4XX_GRAS_SU_POINT_SIZE__SHIFT) & A4XX_GRAS_SU_POINT_SIZE__MASK;
+}
+
+#define REG_A4XX_GRAS_ALPHA_CONTROL				0x00002073
+#define A4XX_GRAS_ALPHA_CONTROL_ALPHA_TEST_ENABLE		0x00000004
+
+#define REG_A4XX_GRAS_SU_POLY_OFFSET_SCALE			0x00002074
+#define A4XX_GRAS_SU_POLY_OFFSET_SCALE__MASK			0xffffffff
+#define A4XX_GRAS_SU_POLY_OFFSET_SCALE__SHIFT			0
+static inline uint32_t A4XX_GRAS_SU_POLY_OFFSET_SCALE(float val)
+{
+	return ((fui(val)) << A4XX_GRAS_SU_POLY_OFFSET_SCALE__SHIFT) & A4XX_GRAS_SU_POLY_OFFSET_SCALE__MASK;
+}
+
+#define REG_A4XX_GRAS_SU_POLY_OFFSET_OFFSET			0x00002075
+#define A4XX_GRAS_SU_POLY_OFFSET_OFFSET__MASK			0xffffffff
+#define A4XX_GRAS_SU_POLY_OFFSET_OFFSET__SHIFT			0
+static inline uint32_t A4XX_GRAS_SU_POLY_OFFSET_OFFSET(float val)
+{
+	return ((fui(val)) << A4XX_GRAS_SU_POLY_OFFSET_OFFSET__SHIFT) & A4XX_GRAS_SU_POLY_OFFSET_OFFSET__MASK;
+}
+
+#define REG_A4XX_GRAS_SC_EXTENT_WINDOW_TL			0x0000209f
+
+#define REG_A4XX_GRAS_SC_SCREEN_SCISSOR_TL			0x0000207c
+#define A4XX_GRAS_SC_SCREEN_SCISSOR_TL_WINDOW_OFFSET_DISABLE	0x80000000
+#define A4XX_GRAS_SC_SCREEN_SCISSOR_TL_X__MASK			0x00007fff
+#define A4XX_GRAS_SC_SCREEN_SCISSOR_TL_X__SHIFT			0
+static inline uint32_t A4XX_GRAS_SC_SCREEN_SCISSOR_TL_X(uint32_t val)
+{
+	return ((val) << A4XX_GRAS_SC_SCREEN_SCISSOR_TL_X__SHIFT) & A4XX_GRAS_SC_SCREEN_SCISSOR_TL_X__MASK;
+}
+#define A4XX_GRAS_SC_SCREEN_SCISSOR_TL_Y__MASK			0x7fff0000
+#define A4XX_GRAS_SC_SCREEN_SCISSOR_TL_Y__SHIFT			16
+static inline uint32_t A4XX_GRAS_SC_SCREEN_SCISSOR_TL_Y(uint32_t val)
+{
+	return ((val) << A4XX_GRAS_SC_SCREEN_SCISSOR_TL_Y__SHIFT) & A4XX_GRAS_SC_SCREEN_SCISSOR_TL_Y__MASK;
+}
+
+#define REG_A4XX_GRAS_SC_SCREEN_SCISSOR_BR			0x0000207d
+#define A4XX_GRAS_SC_SCREEN_SCISSOR_BR_WINDOW_OFFSET_DISABLE	0x80000000
+#define A4XX_GRAS_SC_SCREEN_SCISSOR_BR_X__MASK			0x00007fff
+#define A4XX_GRAS_SC_SCREEN_SCISSOR_BR_X__SHIFT			0
+static inline uint32_t A4XX_GRAS_SC_SCREEN_SCISSOR_BR_X(uint32_t val)
+{
+	return ((val) << A4XX_GRAS_SC_SCREEN_SCISSOR_BR_X__SHIFT) & A4XX_GRAS_SC_SCREEN_SCISSOR_BR_X__MASK;
+}
+#define A4XX_GRAS_SC_SCREEN_SCISSOR_BR_Y__MASK			0x7fff0000
+#define A4XX_GRAS_SC_SCREEN_SCISSOR_BR_Y__SHIFT			16
+static inline uint32_t A4XX_GRAS_SC_SCREEN_SCISSOR_BR_Y(uint32_t val)
+{
+	return ((val) << A4XX_GRAS_SC_SCREEN_SCISSOR_BR_Y__SHIFT) & A4XX_GRAS_SC_SCREEN_SCISSOR_BR_Y__MASK;
+}
+
+#define REG_A4XX_GRAS_SC_WINDOW_SCISSOR_BR			0x0000209c
+#define A4XX_GRAS_SC_WINDOW_SCISSOR_BR_WINDOW_OFFSET_DISABLE	0x80000000
+#define A4XX_GRAS_SC_WINDOW_SCISSOR_BR_X__MASK			0x00007fff
+#define A4XX_GRAS_SC_WINDOW_SCISSOR_BR_X__SHIFT			0
+static inline uint32_t A4XX_GRAS_SC_WINDOW_SCISSOR_BR_X(uint32_t val)
+{
+	return ((val) << A4XX_GRAS_SC_WINDOW_SCISSOR_BR_X__SHIFT) & A4XX_GRAS_SC_WINDOW_SCISSOR_BR_X__MASK;
+}
+#define A4XX_GRAS_SC_WINDOW_SCISSOR_BR_Y__MASK			0x7fff0000
+#define A4XX_GRAS_SC_WINDOW_SCISSOR_BR_Y__SHIFT			16
+static inline uint32_t A4XX_GRAS_SC_WINDOW_SCISSOR_BR_Y(uint32_t val)
+{
+	return ((val) << A4XX_GRAS_SC_WINDOW_SCISSOR_BR_Y__SHIFT) & A4XX_GRAS_SC_WINDOW_SCISSOR_BR_Y__MASK;
+}
+
+#define REG_A4XX_GRAS_SC_WINDOW_SCISSOR_TL			0x0000209d
+#define A4XX_GRAS_SC_WINDOW_SCISSOR_TL_WINDOW_OFFSET_DISABLE	0x80000000
+#define A4XX_GRAS_SC_WINDOW_SCISSOR_TL_X__MASK			0x00007fff
+#define A4XX_GRAS_SC_WINDOW_SCISSOR_TL_X__SHIFT			0
+static inline uint32_t A4XX_GRAS_SC_WINDOW_SCISSOR_TL_X(uint32_t val)
+{
+	return ((val) << A4XX_GRAS_SC_WINDOW_SCISSOR_TL_X__SHIFT) & A4XX_GRAS_SC_WINDOW_SCISSOR_TL_X__MASK;
+}
+#define A4XX_GRAS_SC_WINDOW_SCISSOR_TL_Y__MASK			0x7fff0000
+#define A4XX_GRAS_SC_WINDOW_SCISSOR_TL_Y__SHIFT			16
+static inline uint32_t A4XX_GRAS_SC_WINDOW_SCISSOR_TL_Y(uint32_t val)
+{
+	return ((val) << A4XX_GRAS_SC_WINDOW_SCISSOR_TL_Y__SHIFT) & A4XX_GRAS_SC_WINDOW_SCISSOR_TL_Y__MASK;
+}
+
+#define REG_A4XX_GRAS_DEPTH_CONTROL				0x00002077
+#define A4XX_GRAS_DEPTH_CONTROL_FORMAT__MASK			0x00000003
+#define A4XX_GRAS_DEPTH_CONTROL_FORMAT__SHIFT			0
+static inline uint32_t A4XX_GRAS_DEPTH_CONTROL_FORMAT(enum a4xx_depth_format val)
+{
+	return ((val) << A4XX_GRAS_DEPTH_CONTROL_FORMAT__SHIFT) & A4XX_GRAS_DEPTH_CONTROL_FORMAT__MASK;
+}
+
+#define REG_A4XX_GRAS_SU_MODE_CONTROL				0x00002078
+#define A4XX_GRAS_SU_MODE_CONTROL_CULL_FRONT			0x00000001
+#define A4XX_GRAS_SU_MODE_CONTROL_CULL_BACK			0x00000002
+#define A4XX_GRAS_SU_MODE_CONTROL_FRONT_CW			0x00000004
+#define A4XX_GRAS_SU_MODE_CONTROL_LINEHALFWIDTH__MASK		0x000007f8
+#define A4XX_GRAS_SU_MODE_CONTROL_LINEHALFWIDTH__SHIFT		3
+static inline uint32_t A4XX_GRAS_SU_MODE_CONTROL_LINEHALFWIDTH(float val)
+{
+	return ((((int32_t)(val * 4.0))) << A4XX_GRAS_SU_MODE_CONTROL_LINEHALFWIDTH__SHIFT) & A4XX_GRAS_SU_MODE_CONTROL_LINEHALFWIDTH__MASK;
+}
+#define A4XX_GRAS_SU_MODE_CONTROL_POLY_OFFSET			0x00000800
+#define A4XX_GRAS_SU_MODE_CONTROL_RENDERING_PASS		0x00100000
+
+#define REG_A4XX_GRAS_SC_CONTROL				0x0000207b
+#define A4XX_GRAS_SC_CONTROL_RENDER_MODE__MASK			0x0000000c
+#define A4XX_GRAS_SC_CONTROL_RENDER_MODE__SHIFT			2
+static inline uint32_t A4XX_GRAS_SC_CONTROL_RENDER_MODE(enum a3xx_render_mode val)
+{
+	return ((val) << A4XX_GRAS_SC_CONTROL_RENDER_MODE__SHIFT) & A4XX_GRAS_SC_CONTROL_RENDER_MODE__MASK;
+}
+#define A4XX_GRAS_SC_CONTROL_MSAA_SAMPLES__MASK			0x00000380
+#define A4XX_GRAS_SC_CONTROL_MSAA_SAMPLES__SHIFT		7
+static inline uint32_t A4XX_GRAS_SC_CONTROL_MSAA_SAMPLES(uint32_t val)
+{
+	return ((val) << A4XX_GRAS_SC_CONTROL_MSAA_SAMPLES__SHIFT) & A4XX_GRAS_SC_CONTROL_MSAA_SAMPLES__MASK;
+}
+#define A4XX_GRAS_SC_CONTROL_MSAA_DISABLE			0x00000800
+#define A4XX_GRAS_SC_CONTROL_RASTER_MODE__MASK			0x0000f000
+#define A4XX_GRAS_SC_CONTROL_RASTER_MODE__SHIFT			12
+static inline uint32_t A4XX_GRAS_SC_CONTROL_RASTER_MODE(uint32_t val)
+{
+	return ((val) << A4XX_GRAS_SC_CONTROL_RASTER_MODE__SHIFT) & A4XX_GRAS_SC_CONTROL_RASTER_MODE__MASK;
+}
+
+#define REG_A4XX_UCHE_CACHE_MODE_CONTROL			0x00000e80
+
+#define REG_A4XX_UCHE_TRAP_BASE_LO				0x00000e83
+
+#define REG_A4XX_UCHE_TRAP_BASE_HI				0x00000e84
+
+#define REG_A4XX_UCHE_CACHE_STATUS				0x00000e88
+
+#define REG_A4XX_UCHE_INVALIDATE0				0x00000e8a
+
+#define REG_A4XX_UCHE_INVALIDATE1				0x00000e8b
+
+#define REG_A4XX_UCHE_CACHE_WAYS_VFD				0x00000e8c
+
+#define REG_A4XX_UCHE_PERFCTR_UCHE_SEL_7			0x00000e95
+
+#define REG_A4XX_HLSQ_TIMEOUT_THRESHOLD				0x00000e00
+
+#define REG_A4XX_HLSQ_DEBUG_ECO_CONTROL				0x00000e04
+
+#define REG_A4XX_HLSQ_PERF_PIPE_MASK				0x00000e0e
+
+#define REG_A4XX_HLSQ_CONTROL_0_REG				0x000023c0
+#define A4XX_HLSQ_CONTROL_0_REG_FSTHREADSIZE__MASK		0x00000010
+#define A4XX_HLSQ_CONTROL_0_REG_FSTHREADSIZE__SHIFT		4
+static inline uint32_t A4XX_HLSQ_CONTROL_0_REG_FSTHREADSIZE(enum a3xx_threadsize val)
+{
+	return ((val) << A4XX_HLSQ_CONTROL_0_REG_FSTHREADSIZE__SHIFT) & A4XX_HLSQ_CONTROL_0_REG_FSTHREADSIZE__MASK;
+}
+#define A4XX_HLSQ_CONTROL_0_REG_FSSUPERTHREADENABLE		0x00000040
+#define A4XX_HLSQ_CONTROL_0_REG_SPSHADERRESTART			0x00000200
+#define A4XX_HLSQ_CONTROL_0_REG_RESERVED2			0x00000400
+#define A4XX_HLSQ_CONTROL_0_REG_CHUNKDISABLE			0x04000000
+#define A4XX_HLSQ_CONTROL_0_REG_CONSTMODE__MASK			0x08000000
+#define A4XX_HLSQ_CONTROL_0_REG_CONSTMODE__SHIFT		27
+static inline uint32_t A4XX_HLSQ_CONTROL_0_REG_CONSTMODE(uint32_t val)
+{
+	return ((val) << A4XX_HLSQ_CONTROL_0_REG_CONSTMODE__SHIFT) & A4XX_HLSQ_CONTROL_0_REG_CONSTMODE__MASK;
+}
+#define A4XX_HLSQ_CONTROL_0_REG_LAZYUPDATEDISABLE		0x10000000
+#define A4XX_HLSQ_CONTROL_0_REG_SPCONSTFULLUPDATE		0x20000000
+#define A4XX_HLSQ_CONTROL_0_REG_TPFULLUPDATE			0x40000000
+#define A4XX_HLSQ_CONTROL_0_REG_SINGLECONTEXT			0x80000000
+
+#define REG_A4XX_HLSQ_CONTROL_1_REG				0x000023c1
+#define A4XX_HLSQ_CONTROL_1_REG_VSTHREADSIZE__MASK		0x00000040
+#define A4XX_HLSQ_CONTROL_1_REG_VSTHREADSIZE__SHIFT		6
+static inline uint32_t A4XX_HLSQ_CONTROL_1_REG_VSTHREADSIZE(enum a3xx_threadsize val)
+{
+	return ((val) << A4XX_HLSQ_CONTROL_1_REG_VSTHREADSIZE__SHIFT) & A4XX_HLSQ_CONTROL_1_REG_VSTHREADSIZE__MASK;
+}
+#define A4XX_HLSQ_CONTROL_1_REG_VSSUPERTHREADENABLE		0x00000100
+#define A4XX_HLSQ_CONTROL_1_REG_RESERVED1			0x00000200
+#define A4XX_HLSQ_CONTROL_1_REG_ZWCOORD				0x02000000
+
+#define REG_A4XX_HLSQ_CONTROL_2_REG				0x000023c2
+#define A4XX_HLSQ_CONTROL_2_REG_PRIMALLOCTHRESHOLD__MASK	0xfc000000
+#define A4XX_HLSQ_CONTROL_2_REG_PRIMALLOCTHRESHOLD__SHIFT	26
+static inline uint32_t A4XX_HLSQ_CONTROL_2_REG_PRIMALLOCTHRESHOLD(uint32_t val)
+{
+	return ((val) << A4XX_HLSQ_CONTROL_2_REG_PRIMALLOCTHRESHOLD__SHIFT) & A4XX_HLSQ_CONTROL_2_REG_PRIMALLOCTHRESHOLD__MASK;
+}
+
+#define REG_A4XX_HLSQ_CONTROL_3_REG				0x000023c3
+#define A4XX_HLSQ_CONTROL_3_REG_REGID__MASK			0x000000ff
+#define A4XX_HLSQ_CONTROL_3_REG_REGID__SHIFT			0
+static inline uint32_t A4XX_HLSQ_CONTROL_3_REG_REGID(uint32_t val)
+{
+	return ((val) << A4XX_HLSQ_CONTROL_3_REG_REGID__SHIFT) & A4XX_HLSQ_CONTROL_3_REG_REGID__MASK;
+}
+
+#define REG_A4XX_HLSQ_VS_CONTROL_REG				0x000023c5
+#define A4XX_HLSQ_VS_CONTROL_REG_CONSTLENGTH__MASK		0x000000ff
+#define A4XX_HLSQ_VS_CONTROL_REG_CONSTLENGTH__SHIFT		0
+static inline uint32_t A4XX_HLSQ_VS_CONTROL_REG_CONSTLENGTH(uint32_t val)
+{
+	return ((val) << A4XX_HLSQ_VS_CONTROL_REG_CONSTLENGTH__SHIFT) & A4XX_HLSQ_VS_CONTROL_REG_CONSTLENGTH__MASK;
+}
+#define A4XX_HLSQ_VS_CONTROL_REG_CONSTOBJECTOFFSET__MASK	0x0000ff00
+#define A4XX_HLSQ_VS_CONTROL_REG_CONSTOBJECTOFFSET__SHIFT	8
+static inline uint32_t A4XX_HLSQ_VS_CONTROL_REG_CONSTOBJECTOFFSET(uint32_t val)
+{
+	return ((val) << A4XX_HLSQ_VS_CONTROL_REG_CONSTOBJECTOFFSET__SHIFT) & A4XX_HLSQ_VS_CONTROL_REG_CONSTOBJECTOFFSET__MASK;
+}
+#define A4XX_HLSQ_VS_CONTROL_REG_SHADEROBJOFFSET__MASK		0x00fe0000
+#define A4XX_HLSQ_VS_CONTROL_REG_SHADEROBJOFFSET__SHIFT		17
+static inline uint32_t A4XX_HLSQ_VS_CONTROL_REG_SHADEROBJOFFSET(uint32_t val)
+{
+	return ((val) << A4XX_HLSQ_VS_CONTROL_REG_SHADEROBJOFFSET__SHIFT) & A4XX_HLSQ_VS_CONTROL_REG_SHADEROBJOFFSET__MASK;
+}
+#define A4XX_HLSQ_VS_CONTROL_REG_INSTRLENGTH__MASK		0xff000000
+#define A4XX_HLSQ_VS_CONTROL_REG_INSTRLENGTH__SHIFT		24
+static inline uint32_t A4XX_HLSQ_VS_CONTROL_REG_INSTRLENGTH(uint32_t val)
+{
+	return ((val) << A4XX_HLSQ_VS_CONTROL_REG_INSTRLENGTH__SHIFT) & A4XX_HLSQ_VS_CONTROL_REG_INSTRLENGTH__MASK;
+}
+
+#define REG_A4XX_HLSQ_FS_CONTROL_REG				0x000023c6
+#define A4XX_HLSQ_FS_CONTROL_REG_CONSTLENGTH__MASK		0x000000ff
+#define A4XX_HLSQ_FS_CONTROL_REG_CONSTLENGTH__SHIFT		0
+static inline uint32_t A4XX_HLSQ_FS_CONTROL_REG_CONSTLENGTH(uint32_t val)
+{
+	return ((val) << A4XX_HLSQ_FS_CONTROL_REG_CONSTLENGTH__SHIFT) & A4XX_HLSQ_FS_CONTROL_REG_CONSTLENGTH__MASK;
+}
+#define A4XX_HLSQ_FS_CONTROL_REG_CONSTOBJECTOFFSET__MASK	0x0000ff00
+#define A4XX_HLSQ_FS_CONTROL_REG_CONSTOBJECTOFFSET__SHIFT	8
+static inline uint32_t A4XX_HLSQ_FS_CONTROL_REG_CONSTOBJECTOFFSET(uint32_t val)
+{
+	return ((val) << A4XX_HLSQ_FS_CONTROL_REG_CONSTOBJECTOFFSET__SHIFT) & A4XX_HLSQ_FS_CONTROL_REG_CONSTOBJECTOFFSET__MASK;
+}
+#define A4XX_HLSQ_FS_CONTROL_REG_SHADEROBJOFFSET__MASK		0x00fe0000
+#define A4XX_HLSQ_FS_CONTROL_REG_SHADEROBJOFFSET__SHIFT		17
+static inline uint32_t A4XX_HLSQ_FS_CONTROL_REG_SHADEROBJOFFSET(uint32_t val)
+{
+	return ((val) << A4XX_HLSQ_FS_CONTROL_REG_SHADEROBJOFFSET__SHIFT) & A4XX_HLSQ_FS_CONTROL_REG_SHADEROBJOFFSET__MASK;
+}
+#define A4XX_HLSQ_FS_CONTROL_REG_INSTRLENGTH__MASK		0xff000000
+#define A4XX_HLSQ_FS_CONTROL_REG_INSTRLENGTH__SHIFT		24
+static inline uint32_t A4XX_HLSQ_FS_CONTROL_REG_INSTRLENGTH(uint32_t val)
+{
+	return ((val) << A4XX_HLSQ_FS_CONTROL_REG_INSTRLENGTH__SHIFT) & A4XX_HLSQ_FS_CONTROL_REG_INSTRLENGTH__MASK;
+}
+
+#define REG_A4XX_HLSQ_HS_CONTROL_REG				0x000023c7
+#define A4XX_HLSQ_HS_CONTROL_REG_CONSTLENGTH__MASK		0x000000ff
+#define A4XX_HLSQ_HS_CONTROL_REG_CONSTLENGTH__SHIFT		0
+static inline uint32_t A4XX_HLSQ_HS_CONTROL_REG_CONSTLENGTH(uint32_t val)
+{
+	return ((val) << A4XX_HLSQ_HS_CONTROL_REG_CONSTLENGTH__SHIFT) & A4XX_HLSQ_HS_CONTROL_REG_CONSTLENGTH__MASK;
+}
+#define A4XX_HLSQ_HS_CONTROL_REG_CONSTOBJECTOFFSET__MASK	0x0000ff00
+#define A4XX_HLSQ_HS_CONTROL_REG_CONSTOBJECTOFFSET__SHIFT	8
+static inline uint32_t A4XX_HLSQ_HS_CONTROL_REG_CONSTOBJECTOFFSET(uint32_t val)
+{
+	return ((val) << A4XX_HLSQ_HS_CONTROL_REG_CONSTOBJECTOFFSET__SHIFT) & A4XX_HLSQ_HS_CONTROL_REG_CONSTOBJECTOFFSET__MASK;
+}
+#define A4XX_HLSQ_HS_CONTROL_REG_SHADEROBJOFFSET__MASK		0x00fe0000
+#define A4XX_HLSQ_HS_CONTROL_REG_SHADEROBJOFFSET__SHIFT		17
+static inline uint32_t A4XX_HLSQ_HS_CONTROL_REG_SHADEROBJOFFSET(uint32_t val)
+{
+	return ((val) << A4XX_HLSQ_HS_CONTROL_REG_SHADEROBJOFFSET__SHIFT) & A4XX_HLSQ_HS_CONTROL_REG_SHADEROBJOFFSET__MASK;
+}
+#define A4XX_HLSQ_HS_CONTROL_REG_INSTRLENGTH__MASK		0xff000000
+#define A4XX_HLSQ_HS_CONTROL_REG_INSTRLENGTH__SHIFT		24
+static inline uint32_t A4XX_HLSQ_HS_CONTROL_REG_INSTRLENGTH(uint32_t val)
+{
+	return ((val) << A4XX_HLSQ_HS_CONTROL_REG_INSTRLENGTH__SHIFT) & A4XX_HLSQ_HS_CONTROL_REG_INSTRLENGTH__MASK;
+}
+
+#define REG_A4XX_HLSQ_DS_CONTROL_REG				0x000023c8
+#define A4XX_HLSQ_DS_CONTROL_REG_CONSTLENGTH__MASK		0x000000ff
+#define A4XX_HLSQ_DS_CONTROL_REG_CONSTLENGTH__SHIFT		0
+static inline uint32_t A4XX_HLSQ_DS_CONTROL_REG_CONSTLENGTH(uint32_t val)
+{
+	return ((val) << A4XX_HLSQ_DS_CONTROL_REG_CONSTLENGTH__SHIFT) & A4XX_HLSQ_DS_CONTROL_REG_CONSTLENGTH__MASK;
+}
+#define A4XX_HLSQ_DS_CONTROL_REG_CONSTOBJECTOFFSET__MASK	0x0000ff00
+#define A4XX_HLSQ_DS_CONTROL_REG_CONSTOBJECTOFFSET__SHIFT	8
+static inline uint32_t A4XX_HLSQ_DS_CONTROL_REG_CONSTOBJECTOFFSET(uint32_t val)
+{
+	return ((val) << A4XX_HLSQ_DS_CONTROL_REG_CONSTOBJECTOFFSET__SHIFT) & A4XX_HLSQ_DS_CONTROL_REG_CONSTOBJECTOFFSET__MASK;
+}
+#define A4XX_HLSQ_DS_CONTROL_REG_SHADEROBJOFFSET__MASK		0x00fe0000
+#define A4XX_HLSQ_DS_CONTROL_REG_SHADEROBJOFFSET__SHIFT		17
+static inline uint32_t A4XX_HLSQ_DS_CONTROL_REG_SHADEROBJOFFSET(uint32_t val)
+{
+	return ((val) << A4XX_HLSQ_DS_CONTROL_REG_SHADEROBJOFFSET__SHIFT) & A4XX_HLSQ_DS_CONTROL_REG_SHADEROBJOFFSET__MASK;
+}
+#define A4XX_HLSQ_DS_CONTROL_REG_INSTRLENGTH__MASK		0xff000000
+#define A4XX_HLSQ_DS_CONTROL_REG_INSTRLENGTH__SHIFT		24
+static inline uint32_t A4XX_HLSQ_DS_CONTROL_REG_INSTRLENGTH(uint32_t val)
+{
+	return ((val) << A4XX_HLSQ_DS_CONTROL_REG_INSTRLENGTH__SHIFT) & A4XX_HLSQ_DS_CONTROL_REG_INSTRLENGTH__MASK;
+}
+
+#define REG_A4XX_HLSQ_GS_CONTROL_REG				0x000023c9
+#define A4XX_HLSQ_GS_CONTROL_REG_CONSTLENGTH__MASK		0x000000ff
+#define A4XX_HLSQ_GS_CONTROL_REG_CONSTLENGTH__SHIFT		0
+static inline uint32_t A4XX_HLSQ_GS_CONTROL_REG_CONSTLENGTH(uint32_t val)
+{
+	return ((val) << A4XX_HLSQ_GS_CONTROL_REG_CONSTLENGTH__SHIFT) & A4XX_HLSQ_GS_CONTROL_REG_CONSTLENGTH__MASK;
+}
+#define A4XX_HLSQ_GS_CONTROL_REG_CONSTOBJECTOFFSET__MASK	0x0000ff00
+#define A4XX_HLSQ_GS_CONTROL_REG_CONSTOBJECTOFFSET__SHIFT	8
+static inline uint32_t A4XX_HLSQ_GS_CONTROL_REG_CONSTOBJECTOFFSET(uint32_t val)
+{
+	return ((val) << A4XX_HLSQ_GS_CONTROL_REG_CONSTOBJECTOFFSET__SHIFT) & A4XX_HLSQ_GS_CONTROL_REG_CONSTOBJECTOFFSET__MASK;
+}
+#define A4XX_HLSQ_GS_CONTROL_REG_SHADEROBJOFFSET__MASK		0x00fe0000
+#define A4XX_HLSQ_GS_CONTROL_REG_SHADEROBJOFFSET__SHIFT		17
+static inline uint32_t A4XX_HLSQ_GS_CONTROL_REG_SHADEROBJOFFSET(uint32_t val)
+{
+	return ((val) << A4XX_HLSQ_GS_CONTROL_REG_SHADEROBJOFFSET__SHIFT) & A4XX_HLSQ_GS_CONTROL_REG_SHADEROBJOFFSET__MASK;
+}
+#define A4XX_HLSQ_GS_CONTROL_REG_INSTRLENGTH__MASK		0xff000000
+#define A4XX_HLSQ_GS_CONTROL_REG_INSTRLENGTH__SHIFT		24
+static inline uint32_t A4XX_HLSQ_GS_CONTROL_REG_INSTRLENGTH(uint32_t val)
+{
+	return ((val) << A4XX_HLSQ_GS_CONTROL_REG_INSTRLENGTH__SHIFT) & A4XX_HLSQ_GS_CONTROL_REG_INSTRLENGTH__MASK;
+}
+
+#define REG_A4XX_HLSQ_UPDATE_CONTROL				0x000023db
+
+#define REG_A4XX_PC_BINNING_COMMAND				0x00000d00
+#define A4XX_PC_BINNING_COMMAND_BINNING_ENABLE			0x00000001
+
+#define REG_A4XX_PC_DRAWCALL_SETUP_OVERRIDE			0x00000d0c
+
+#define REG_A4XX_PC_PERFCTR_PC_SEL_0				0x00000d10
+
+#define REG_A4XX_PC_PERFCTR_PC_SEL_7				0x00000d17
+
+#define REG_A4XX_PC_BIN_BASE					0x000021c0
+
+#define REG_A4XX_PC_PRIM_VTX_CNTL				0x000021c4
+#define A4XX_PC_PRIM_VTX_CNTL_VAROUT				0x00000001
+#define A4XX_PC_PRIM_VTX_CNTL_PROVOKING_VTX_LAST		0x02000000
+#define A4XX_PC_PRIM_VTX_CNTL_PSIZE				0x04000000
+
+#define REG_A4XX_UNKNOWN_21C5					0x000021c5
+
+#define REG_A4XX_PC_RESTART_INDEX				0x000021c6
+
+#define REG_A4XX_PC_GS_PARAM					0x000021e5
+
+#define REG_A4XX_PC_HS_PARAM					0x000021e7
+
+#define REG_A4XX_VBIF_VERSION					0x00003000
+
+#define REG_A4XX_VBIF_CLKON					0x00003001
+#define A4XX_VBIF_CLKON_FORCE_ON_TESTBUS			0x00000001
+
+#define REG_A4XX_VBIF_ABIT_SORT					0x0000301c
+
+#define REG_A4XX_VBIF_ABIT_SORT_CONF				0x0000301d
+
+#define REG_A4XX_VBIF_GATE_OFF_WRREQ_EN				0x0000302a
+
+#define REG_A4XX_VBIF_IN_RD_LIM_CONF0				0x0000302c
+
+#define REG_A4XX_VBIF_IN_RD_LIM_CONF1				0x0000302d
+
+#define REG_A4XX_VBIF_IN_WR_LIM_CONF0				0x00003030
+
+#define REG_A4XX_VBIF_IN_WR_LIM_CONF1				0x00003031
+
+#define REG_A4XX_VBIF_ROUND_ROBIN_QOS_ARB			0x00003049
+
+#define REG_A4XX_UNKNOWN_0CC5					0x00000cc5
+
+#define REG_A4XX_UNKNOWN_0CC6					0x00000cc6
+
+#define REG_A4XX_UNKNOWN_0D01					0x00000d01
+
+#define REG_A4XX_UNKNOWN_0E05					0x00000e05
+
+#define REG_A4XX_UNKNOWN_0E42					0x00000e42
+
+#define REG_A4XX_UNKNOWN_0EC2					0x00000ec2
+
+#define REG_A4XX_UNKNOWN_0EC3					0x00000ec3
+
+#define REG_A4XX_UNKNOWN_0F03					0x00000f03
+
+#define REG_A4XX_UNKNOWN_2001					0x00002001
+
+#define REG_A4XX_UNKNOWN_209B					0x0000209b
+
+#define REG_A4XX_UNKNOWN_20EF					0x000020ef
+
+#define REG_A4XX_UNKNOWN_20F0					0x000020f0
+
+#define REG_A4XX_UNKNOWN_20F1					0x000020f1
+
+#define REG_A4XX_UNKNOWN_20F2					0x000020f2
+
+#define REG_A4XX_UNKNOWN_20F3					0x000020f3
+
+#define REG_A4XX_UNKNOWN_20F4					0x000020f4
+
+#define REG_A4XX_UNKNOWN_20F5					0x000020f5
+
+#define REG_A4XX_UNKNOWN_20F6					0x000020f6
+
+#define REG_A4XX_UNKNOWN_20F7					0x000020f7
+
+#define REG_A4XX_UNKNOWN_2152					0x00002152
+
+#define REG_A4XX_UNKNOWN_2153					0x00002153
+
+#define REG_A4XX_UNKNOWN_2154					0x00002154
+
+#define REG_A4XX_UNKNOWN_2155					0x00002155
+
+#define REG_A4XX_UNKNOWN_2156					0x00002156
+
+#define REG_A4XX_UNKNOWN_2157					0x00002157
+
+#define REG_A4XX_UNKNOWN_21C3					0x000021c3
+
+#define REG_A4XX_UNKNOWN_21E6					0x000021e6
+
+#define REG_A4XX_UNKNOWN_2209					0x00002209
+
+#define REG_A4XX_UNKNOWN_22D7					0x000022d7
+
+#define REG_A4XX_UNKNOWN_2381					0x00002381
+
+#define REG_A4XX_UNKNOWN_23A0					0x000023a0
+
+#define REG_A4XX_TEX_SAMP_0					0x00000000
+#define A4XX_TEX_SAMP_0_XY_MAG__MASK				0x00000006
+#define A4XX_TEX_SAMP_0_XY_MAG__SHIFT				1
+static inline uint32_t A4XX_TEX_SAMP_0_XY_MAG(enum a4xx_tex_filter val)
+{
+	return ((val) << A4XX_TEX_SAMP_0_XY_MAG__SHIFT) & A4XX_TEX_SAMP_0_XY_MAG__MASK;
+}
+#define A4XX_TEX_SAMP_0_XY_MIN__MASK				0x00000018
+#define A4XX_TEX_SAMP_0_XY_MIN__SHIFT				3
+static inline uint32_t A4XX_TEX_SAMP_0_XY_MIN(enum a4xx_tex_filter val)
+{
+	return ((val) << A4XX_TEX_SAMP_0_XY_MIN__SHIFT) & A4XX_TEX_SAMP_0_XY_MIN__MASK;
+}
+#define A4XX_TEX_SAMP_0_WRAP_S__MASK				0x000000e0
+#define A4XX_TEX_SAMP_0_WRAP_S__SHIFT				5
+static inline uint32_t A4XX_TEX_SAMP_0_WRAP_S(enum a4xx_tex_clamp val)
+{
+	return ((val) << A4XX_TEX_SAMP_0_WRAP_S__SHIFT) & A4XX_TEX_SAMP_0_WRAP_S__MASK;
+}
+#define A4XX_TEX_SAMP_0_WRAP_T__MASK				0x00000700
+#define A4XX_TEX_SAMP_0_WRAP_T__SHIFT				8
+static inline uint32_t A4XX_TEX_SAMP_0_WRAP_T(enum a4xx_tex_clamp val)
+{
+	return ((val) << A4XX_TEX_SAMP_0_WRAP_T__SHIFT) & A4XX_TEX_SAMP_0_WRAP_T__MASK;
+}
+#define A4XX_TEX_SAMP_0_WRAP_R__MASK				0x00003800
+#define A4XX_TEX_SAMP_0_WRAP_R__SHIFT				11
+static inline uint32_t A4XX_TEX_SAMP_0_WRAP_R(enum a4xx_tex_clamp val)
+{
+	return ((val) << A4XX_TEX_SAMP_0_WRAP_R__SHIFT) & A4XX_TEX_SAMP_0_WRAP_R__MASK;
+}
+
+#define REG_A4XX_TEX_SAMP_1					0x00000001
+#define A4XX_TEX_SAMP_1_COMPARE_FUNC__MASK			0x0000000e
+#define A4XX_TEX_SAMP_1_COMPARE_FUNC__SHIFT			1
+static inline uint32_t A4XX_TEX_SAMP_1_COMPARE_FUNC(enum adreno_compare_func val)
+{
+	return ((val) << A4XX_TEX_SAMP_1_COMPARE_FUNC__SHIFT) & A4XX_TEX_SAMP_1_COMPARE_FUNC__MASK;
+}
+#define A4XX_TEX_SAMP_1_MAX_LOD__MASK				0x000fff00
+#define A4XX_TEX_SAMP_1_MAX_LOD__SHIFT				8
+static inline uint32_t A4XX_TEX_SAMP_1_MAX_LOD(float val)
+{
+	return ((((uint32_t)(val * 64.0))) << A4XX_TEX_SAMP_1_MAX_LOD__SHIFT) & A4XX_TEX_SAMP_1_MAX_LOD__MASK;
+}
+#define A4XX_TEX_SAMP_1_MIN_LOD__MASK				0xfff00000
+#define A4XX_TEX_SAMP_1_MIN_LOD__SHIFT				20
+static inline uint32_t A4XX_TEX_SAMP_1_MIN_LOD(float val)
+{
+	return ((((uint32_t)(val * 64.0))) << A4XX_TEX_SAMP_1_MIN_LOD__SHIFT) & A4XX_TEX_SAMP_1_MIN_LOD__MASK;
+}
+
+#define REG_A4XX_TEX_CONST_0					0x00000000
+#define A4XX_TEX_CONST_0_TILED					0x00000001
+#define A4XX_TEX_CONST_0_SWIZ_X__MASK				0x00000070
+#define A4XX_TEX_CONST_0_SWIZ_X__SHIFT				4
+static inline uint32_t A4XX_TEX_CONST_0_SWIZ_X(enum a4xx_tex_swiz val)
+{
+	return ((val) << A4XX_TEX_CONST_0_SWIZ_X__SHIFT) & A4XX_TEX_CONST_0_SWIZ_X__MASK;
+}
+#define A4XX_TEX_CONST_0_SWIZ_Y__MASK				0x00000380
+#define A4XX_TEX_CONST_0_SWIZ_Y__SHIFT				7
+static inline uint32_t A4XX_TEX_CONST_0_SWIZ_Y(enum a4xx_tex_swiz val)
+{
+	return ((val) << A4XX_TEX_CONST_0_SWIZ_Y__SHIFT) & A4XX_TEX_CONST_0_SWIZ_Y__MASK;
+}
+#define A4XX_TEX_CONST_0_SWIZ_Z__MASK				0x00001c00
+#define A4XX_TEX_CONST_0_SWIZ_Z__SHIFT				10
+static inline uint32_t A4XX_TEX_CONST_0_SWIZ_Z(enum a4xx_tex_swiz val)
+{
+	return ((val) << A4XX_TEX_CONST_0_SWIZ_Z__SHIFT) & A4XX_TEX_CONST_0_SWIZ_Z__MASK;
+}
+#define A4XX_TEX_CONST_0_SWIZ_W__MASK				0x0000e000
+#define A4XX_TEX_CONST_0_SWIZ_W__SHIFT				13
+static inline uint32_t A4XX_TEX_CONST_0_SWIZ_W(enum a4xx_tex_swiz val)
+{
+	return ((val) << A4XX_TEX_CONST_0_SWIZ_W__SHIFT) & A4XX_TEX_CONST_0_SWIZ_W__MASK;
+}
+#define A4XX_TEX_CONST_0_FMT__MASK				0x1fc00000
+#define A4XX_TEX_CONST_0_FMT__SHIFT				22
+static inline uint32_t A4XX_TEX_CONST_0_FMT(enum a4xx_tex_fmt val)
+{
+	return ((val) << A4XX_TEX_CONST_0_FMT__SHIFT) & A4XX_TEX_CONST_0_FMT__MASK;
+}
+#define A4XX_TEX_CONST_0_TYPE__MASK				0x60000000
+#define A4XX_TEX_CONST_0_TYPE__SHIFT				29
+static inline uint32_t A4XX_TEX_CONST_0_TYPE(enum a4xx_tex_type val)
+{
+	return ((val) << A4XX_TEX_CONST_0_TYPE__SHIFT) & A4XX_TEX_CONST_0_TYPE__MASK;
+}
+
+#define REG_A4XX_TEX_CONST_1					0x00000001
+#define A4XX_TEX_CONST_1_HEIGHT__MASK				0x00007fff
+#define A4XX_TEX_CONST_1_HEIGHT__SHIFT				0
+static inline uint32_t A4XX_TEX_CONST_1_HEIGHT(uint32_t val)
+{
+	return ((val) << A4XX_TEX_CONST_1_HEIGHT__SHIFT) & A4XX_TEX_CONST_1_HEIGHT__MASK;
+}
+#define A4XX_TEX_CONST_1_WIDTH__MASK				0x1fff8000
+#define A4XX_TEX_CONST_1_WIDTH__SHIFT				15
+static inline uint32_t A4XX_TEX_CONST_1_WIDTH(uint32_t val)
+{
+	return ((val) << A4XX_TEX_CONST_1_WIDTH__SHIFT) & A4XX_TEX_CONST_1_WIDTH__MASK;
+}
+
+#define REG_A4XX_TEX_CONST_2					0x00000002
+#define A4XX_TEX_CONST_2_PITCH__MASK				0x3ffffe00
+#define A4XX_TEX_CONST_2_PITCH__SHIFT				9
+static inline uint32_t A4XX_TEX_CONST_2_PITCH(uint32_t val)
+{
+	return ((val) << A4XX_TEX_CONST_2_PITCH__SHIFT) & A4XX_TEX_CONST_2_PITCH__MASK;
+}
+#define A4XX_TEX_CONST_2_SWAP__MASK				0xc0000000
+#define A4XX_TEX_CONST_2_SWAP__SHIFT				30
+static inline uint32_t A4XX_TEX_CONST_2_SWAP(enum a3xx_color_swap val)
+{
+	return ((val) << A4XX_TEX_CONST_2_SWAP__SHIFT) & A4XX_TEX_CONST_2_SWAP__MASK;
+}
+
+#define REG_A4XX_TEX_CONST_3					0x00000003
+#define A4XX_TEX_CONST_3_LAYERSZ__MASK				0x0000000f
+#define A4XX_TEX_CONST_3_LAYERSZ__SHIFT				0
+static inline uint32_t A4XX_TEX_CONST_3_LAYERSZ(uint32_t val)
+{
+	return ((val >> 12) << A4XX_TEX_CONST_3_LAYERSZ__SHIFT) & A4XX_TEX_CONST_3_LAYERSZ__MASK;
+}
+
+#define REG_A4XX_TEX_CONST_4					0x00000004
+#define A4XX_TEX_CONST_4_BASE__MASK				0xffffffff
+#define A4XX_TEX_CONST_4_BASE__SHIFT				0
+static inline uint32_t A4XX_TEX_CONST_4_BASE(uint32_t val)
+{
+	return ((val) << A4XX_TEX_CONST_4_BASE__SHIFT) & A4XX_TEX_CONST_4_BASE__MASK;
+}
+
+#define REG_A4XX_TEX_CONST_5					0x00000005
+
+#define REG_A4XX_TEX_CONST_6					0x00000006
+
+#define REG_A4XX_TEX_CONST_7					0x00000007
+
+
+#endif /* A4XX_XML */
diff --git a/drivers/gpu/drm/msm/adreno/a4xx_gpu.c b/drivers/gpu/drm/msm/adreno/a4xx_gpu.c
new file mode 100644
index 000000000000..91221836c5ad
--- /dev/null
+++ b/drivers/gpu/drm/msm/adreno/a4xx_gpu.c
@@ -0,0 +1,604 @@
+/* Copyright (c) 2014 The Linux Foundation. All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 and
+ * only version 2 as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ */
+#include "a4xx_gpu.h"
+#ifdef CONFIG_MSM_OCMEM
+#  include <soc/qcom/ocmem.h>
+#endif
+
+#define A4XX_INT0_MASK \
+	(A4XX_INT0_RBBM_AHB_ERROR |        \
+	 A4XX_INT0_RBBM_ATB_BUS_OVERFLOW | \
+	 A4XX_INT0_CP_T0_PACKET_IN_IB |    \
+	 A4XX_INT0_CP_OPCODE_ERROR |       \
+	 A4XX_INT0_CP_RESERVED_BIT_ERROR | \
+	 A4XX_INT0_CP_HW_FAULT |           \
+	 A4XX_INT0_CP_IB1_INT |            \
+	 A4XX_INT0_CP_IB2_INT |            \
+	 A4XX_INT0_CP_RB_INT |             \
+	 A4XX_INT0_CP_REG_PROTECT_FAULT |  \
+	 A4XX_INT0_CP_AHB_ERROR_HALT |     \
+	 A4XX_INT0_UCHE_OOB_ACCESS)
+
+extern bool hang_debug;
+static void a4xx_dump(struct msm_gpu *gpu);
+
+/*
+ * a4xx_enable_hwcg() - Program the clock control registers
+ * @device: The adreno device pointer
+ */
+static void a4xx_enable_hwcg(struct msm_gpu *gpu)
+{
+	struct adreno_gpu *adreno_gpu = to_adreno_gpu(gpu);
+	unsigned int i;
+	for (i = 0; i < 4; i++)
+		gpu_write(gpu, REG_A4XX_RBBM_CLOCK_CTL_TP(i), 0x02222202);
+	for (i = 0; i < 4; i++)
+		gpu_write(gpu, REG_A4XX_RBBM_CLOCK_CTL2_TP(i), 0x00002222);
+	for (i = 0; i < 4; i++)
+		gpu_write(gpu, REG_A4XX_RBBM_CLOCK_HYST_TP(i), 0x0E739CE7);
+	for (i = 0; i < 4; i++)
+		gpu_write(gpu, REG_A4XX_RBBM_CLOCK_DELAY_TP(i), 0x00111111);
+	for (i = 0; i < 4; i++)
+		gpu_write(gpu, REG_A4XX_RBBM_CLOCK_CTL_SP(i), 0x22222222);
+	for (i = 0; i < 4; i++)
+		gpu_write(gpu, REG_A4XX_RBBM_CLOCK_CTL2_SP(i), 0x00222222);
+	for (i = 0; i < 4; i++)
+		gpu_write(gpu, REG_A4XX_RBBM_CLOCK_HYST_SP(i), 0x00000104);
+	for (i = 0; i < 4; i++)
+		gpu_write(gpu, REG_A4XX_RBBM_CLOCK_DELAY_SP(i), 0x00000081);
+	gpu_write(gpu, REG_A4XX_RBBM_CLOCK_CTL_UCHE, 0x22222222);
+	gpu_write(gpu, REG_A4XX_RBBM_CLOCK_CTL2_UCHE, 0x02222222);
+	gpu_write(gpu, REG_A4XX_RBBM_CLOCK_CTL3_UCHE, 0x00000000);
+	gpu_write(gpu, REG_A4XX_RBBM_CLOCK_CTL4_UCHE, 0x00000000);
+	gpu_write(gpu, REG_A4XX_RBBM_CLOCK_HYST_UCHE, 0x00004444);
+	gpu_write(gpu, REG_A4XX_RBBM_CLOCK_DELAY_UCHE, 0x00001112);
+	for (i = 0; i < 4; i++)
+		gpu_write(gpu, REG_A4XX_RBBM_CLOCK_CTL_RB(i), 0x22222222);
+
+	/* Disable L1 clocking in A420 due to CCU issues with it */
+	for (i = 0; i < 4; i++) {
+		if (adreno_is_a420(adreno_gpu)) {
+			gpu_write(gpu, REG_A4XX_RBBM_CLOCK_CTL2_RB(i),
+					0x00002020);
+		} else {
+			gpu_write(gpu, REG_A4XX_RBBM_CLOCK_CTL2_RB(i),
+					0x00022020);
+		}
+	}
+
+	for (i = 0; i < 4; i++) {
+		gpu_write(gpu, REG_A4XX_RBBM_CLOCK_CTL_MARB_CCU(i),
+				0x00000922);
+	}
+
+	for (i = 0; i < 4; i++) {
+		gpu_write(gpu, REG_A4XX_RBBM_CLOCK_HYST_RB_MARB_CCU(i),
+				0x00000000);
+	}
+
+	for (i = 0; i < 4; i++) {
+		gpu_write(gpu, REG_A4XX_RBBM_CLOCK_DELAY_RB_MARB_CCU_L1(i),
+				0x00000001);
+	}
+
+	gpu_write(gpu, REG_A4XX_RBBM_CLOCK_MODE_GPC, 0x02222222);
+	gpu_write(gpu, REG_A4XX_RBBM_CLOCK_HYST_GPC, 0x04100104);
+	gpu_write(gpu, REG_A4XX_RBBM_CLOCK_DELAY_GPC, 0x00022222);
+	gpu_write(gpu, REG_A4XX_RBBM_CLOCK_CTL_COM_DCOM, 0x00000022);
+	gpu_write(gpu, REG_A4XX_RBBM_CLOCK_HYST_COM_DCOM, 0x0000010F);
+	gpu_write(gpu, REG_A4XX_RBBM_CLOCK_DELAY_COM_DCOM, 0x00000022);
+	gpu_write(gpu, REG_A4XX_RBBM_CLOCK_CTL_TSE_RAS_RBBM, 0x00222222);
+	gpu_write(gpu, REG_A4XX_RBBM_CLOCK_HYST_TSE_RAS_RBBM, 0x00004104);
+	gpu_write(gpu, REG_A4XX_RBBM_CLOCK_DELAY_TSE_RAS_RBBM, 0x00000222);
+	gpu_write(gpu, REG_A4XX_RBBM_CLOCK_CTL_HLSQ , 0x00000000);
+	gpu_write(gpu, REG_A4XX_RBBM_CLOCK_HYST_HLSQ, 0x00000000);
+	gpu_write(gpu, REG_A4XX_RBBM_CLOCK_DELAY_HLSQ, 0x00020000);
+	gpu_write(gpu, REG_A4XX_RBBM_CLOCK_CTL, 0xAAAAAAAA);
+	gpu_write(gpu, REG_A4XX_RBBM_CLOCK_CTL2, 0);
+}
+
+static void a4xx_me_init(struct msm_gpu *gpu)
+{
+	struct msm_ringbuffer *ring = gpu->rb;
+
+	OUT_PKT3(ring, CP_ME_INIT, 17);
+	OUT_RING(ring, 0x000003f7);
+	OUT_RING(ring, 0x00000000);
+	OUT_RING(ring, 0x00000000);
+	OUT_RING(ring, 0x00000000);
+	OUT_RING(ring, 0x00000080);
+	OUT_RING(ring, 0x00000100);
+	OUT_RING(ring, 0x00000180);
+	OUT_RING(ring, 0x00006600);
+	OUT_RING(ring, 0x00000150);
+	OUT_RING(ring, 0x0000014e);
+	OUT_RING(ring, 0x00000154);
+	OUT_RING(ring, 0x00000001);
+	OUT_RING(ring, 0x00000000);
+	OUT_RING(ring, 0x00000000);
+	OUT_RING(ring, 0x00000000);
+	OUT_RING(ring, 0x00000000);
+	OUT_RING(ring, 0x00000000);
+
+	gpu->funcs->flush(gpu);
+	gpu->funcs->idle(gpu);
+}
+
+static int a4xx_hw_init(struct msm_gpu *gpu)
+{
+	struct adreno_gpu *adreno_gpu = to_adreno_gpu(gpu);
+	struct a4xx_gpu *a4xx_gpu = to_a4xx_gpu(adreno_gpu);
+	uint32_t *ptr, len;
+	int i, ret;
+
+	if (adreno_is_a4xx(adreno_gpu)) {
+		gpu_write(gpu, REG_A4XX_VBIF_ABIT_SORT, 0x0001001F);
+		gpu_write(gpu, REG_A4XX_VBIF_ABIT_SORT_CONF, 0x000000A4);
+		gpu_write(gpu, REG_A4XX_VBIF_GATE_OFF_WRREQ_EN, 0x00000001);
+		gpu_write(gpu, REG_A4XX_VBIF_IN_RD_LIM_CONF0, 0x18181818);
+		gpu_write(gpu, REG_A4XX_VBIF_IN_RD_LIM_CONF1, 0x00000018);
+		gpu_write(gpu, REG_A4XX_VBIF_IN_WR_LIM_CONF0, 0x18181818);
+		gpu_write(gpu, REG_A4XX_VBIF_IN_WR_LIM_CONF1, 0x00000018);
+		gpu_write(gpu, REG_A4XX_VBIF_ROUND_ROBIN_QOS_ARB, 0x00000003);
+	} else {
+		BUG();
+	}
+
+	/* Make all blocks contribute to the GPU BUSY perf counter */
+	gpu_write(gpu, REG_A4XX_RBBM_GPU_BUSY_MASKED, 0xffffffff);
+
+	/* Tune the hystersis counters for SP and CP idle detection */
+	gpu_write(gpu, REG_A4XX_RBBM_SP_HYST_CNT, 0x10);
+	gpu_write(gpu, REG_A4XX_RBBM_WAIT_IDLE_CLOCKS_CTL, 0x10);
+
+	 /* Enable the RBBM error reporting bits */
+	gpu_write(gpu, REG_A4XX_RBBM_AHB_CTL0, 0x00000001);
+
+	/* Enable AHB error reporting*/
+	gpu_write(gpu, REG_A4XX_RBBM_AHB_CTL1, 0xa6ffffff);
+
+	/* Enable power counters*/
+	gpu_write(gpu, REG_A4XX_RBBM_RBBM_CTL, 0x00000030);
+
+	/*
+	 * Turn on hang detection - this spews a lot of useful information
+	 * into the RBBM registers on a hang:
+	 */
+	gpu_write(gpu, REG_A4XX_RBBM_INTERFACE_HANG_INT_CTL,
+			(1 << 30) | 0xFFFF);
+
+	gpu_write(gpu, REG_A4XX_RB_GMEM_BASE_ADDR,
+			(unsigned int)(a4xx_gpu->ocmem_base >> 14));
+
+	/* Turn on performance counters: */
+	gpu_write(gpu, REG_A4XX_RBBM_PERFCTR_CTL, 0x01);
+
+	/* Disable L2 bypass to avoid UCHE out of bounds errors */
+	gpu_write(gpu, REG_A4XX_UCHE_TRAP_BASE_LO, 0xffff0000);
+	gpu_write(gpu, REG_A4XX_UCHE_TRAP_BASE_HI, 0xffff0000);
+
+	gpu_write(gpu, REG_A4XX_CP_DEBUG, (1 << 25) |
+			(adreno_is_a420(adreno_gpu) ? (1 << 29) : 0));
+
+	a4xx_enable_hwcg(gpu);
+
+	/*
+	 * For A420 set RBBM_CLOCK_DELAY_HLSQ.CGC_HLSQ_TP_EARLY_CYC >= 2
+	 * due to timing issue with HLSQ_TP_CLK_EN
+	 */
+	if (adreno_is_a420(adreno_gpu)) {
+		unsigned int val;
+		val = gpu_read(gpu, REG_A4XX_RBBM_CLOCK_DELAY_HLSQ);
+		val &= ~A4XX_CGC_HLSQ_EARLY_CYC__MASK;
+		val |= 2 << A4XX_CGC_HLSQ_EARLY_CYC__SHIFT;
+		gpu_write(gpu, REG_A4XX_RBBM_CLOCK_DELAY_HLSQ, val);
+	}
+
+	ret = adreno_hw_init(gpu);
+	if (ret)
+		return ret;
+
+	/* setup access protection: */
+	gpu_write(gpu, REG_A4XX_CP_PROTECT_CTRL, 0x00000007);
+
+	/* RBBM registers */
+	gpu_write(gpu, REG_A4XX_CP_PROTECT(0), 0x62000010);
+	gpu_write(gpu, REG_A4XX_CP_PROTECT(1), 0x63000020);
+	gpu_write(gpu, REG_A4XX_CP_PROTECT(2), 0x64000040);
+	gpu_write(gpu, REG_A4XX_CP_PROTECT(3), 0x65000080);
+	gpu_write(gpu, REG_A4XX_CP_PROTECT(4), 0x66000100);
+	gpu_write(gpu, REG_A4XX_CP_PROTECT(5), 0x64000200);
+
+	/* CP registers */
+	gpu_write(gpu, REG_A4XX_CP_PROTECT(6), 0x67000800);
+	gpu_write(gpu, REG_A4XX_CP_PROTECT(7), 0x64001600);
+
+
+	/* RB registers */
+	gpu_write(gpu, REG_A4XX_CP_PROTECT(8), 0x60003300);
+
+	/* HLSQ registers */
+	gpu_write(gpu, REG_A4XX_CP_PROTECT(9), 0x60003800);
+
+	/* VPC registers */
+	gpu_write(gpu, REG_A4XX_CP_PROTECT(10), 0x61003980);
+
+	/* SMMU registers */
+	gpu_write(gpu, REG_A4XX_CP_PROTECT(11), 0x6e010000);
+
+	gpu_write(gpu, REG_A4XX_RBBM_INT_0_MASK, A4XX_INT0_MASK);
+
+	ret = adreno_hw_init(gpu);
+	if (ret)
+		return ret;
+
+	/* Load PM4: */
+	ptr = (uint32_t *)(adreno_gpu->pm4->data);
+	len = adreno_gpu->pm4->size / 4;
+	DBG("loading PM4 ucode version: %u", ptr[0]);
+	gpu_write(gpu, REG_A4XX_CP_ME_RAM_WADDR, 0);
+	for (i = 1; i < len; i++)
+		gpu_write(gpu, REG_A4XX_CP_ME_RAM_DATA, ptr[i]);
+
+	/* Load PFP: */
+	ptr = (uint32_t *)(adreno_gpu->pfp->data);
+	len = adreno_gpu->pfp->size / 4;
+	DBG("loading PFP ucode version: %u", ptr[0]);
+
+	gpu_write(gpu, REG_A4XX_CP_PFP_UCODE_ADDR, 0);
+	for (i = 1; i < len; i++)
+		gpu_write(gpu, REG_A4XX_CP_PFP_UCODE_DATA, ptr[i]);
+
+	/* clear ME_HALT to start micro engine */
+	gpu_write(gpu, REG_A4XX_CP_ME_CNTL, 0);
+
+	a4xx_me_init(gpu);
+	return 0;
+}
+
+static void a4xx_recover(struct msm_gpu *gpu)
+{
+	/* dump registers before resetting gpu, if enabled: */
+	if (hang_debug)
+		a4xx_dump(gpu);
+
+	gpu_write(gpu, REG_A4XX_RBBM_SW_RESET_CMD, 1);
+	gpu_read(gpu, REG_A4XX_RBBM_SW_RESET_CMD);
+	gpu_write(gpu, REG_A4XX_RBBM_SW_RESET_CMD, 0);
+	adreno_recover(gpu);
+}
+
+static void a4xx_destroy(struct msm_gpu *gpu)
+{
+	struct adreno_gpu *adreno_gpu = to_adreno_gpu(gpu);
+	struct a4xx_gpu *a4xx_gpu = to_a4xx_gpu(adreno_gpu);
+
+	DBG("%s", gpu->name);
+
+	adreno_gpu_cleanup(adreno_gpu);
+
+#ifdef CONFIG_MSM_OCMEM
+	if (a4xx_gpu->ocmem_base)
+		ocmem_free(OCMEM_GRAPHICS, a4xx_gpu->ocmem_hdl);
+#endif
+
+	kfree(a4xx_gpu);
+}
+
+static void a4xx_idle(struct msm_gpu *gpu)
+{
+	/* wait for ringbuffer to drain: */
+	adreno_idle(gpu);
+
+	/* then wait for GPU to finish: */
+	if (spin_until(!(gpu_read(gpu, REG_A4XX_RBBM_STATUS) &
+					A4XX_RBBM_STATUS_GPU_BUSY)))
+		DRM_ERROR("%s: timeout waiting for GPU to idle!\n", gpu->name);
+
+	/* TODO maybe we need to reset GPU here to recover from hang? */
+}
+
+static irqreturn_t a4xx_irq(struct msm_gpu *gpu)
+{
+	uint32_t status;
+
+	status = gpu_read(gpu, REG_A4XX_RBBM_INT_0_STATUS);
+	DBG("%s: Int status %08x", gpu->name, status);
+
+	gpu_write(gpu, REG_A4XX_RBBM_INT_CLEAR_CMD, status);
+
+	msm_gpu_retire(gpu);
+
+	return IRQ_HANDLED;
+}
+
+static const unsigned int a4xx_registers[] = {
+	/* RBBM */
+	0x0000, 0x0002, 0x0004, 0x0021, 0x0023, 0x0024, 0x0026, 0x0026,
+	0x0028, 0x002B, 0x002E, 0x0034, 0x0037, 0x0044, 0x0047, 0x0066,
+	0x0068, 0x0095, 0x009C, 0x0170, 0x0174, 0x01AF,
+	/* CP */
+	0x0200, 0x0233, 0x0240, 0x0250, 0x04C0, 0x04DD, 0x0500, 0x050B,
+	0x0578, 0x058F,
+	/* VSC */
+	0x0C00, 0x0C03, 0x0C08, 0x0C41, 0x0C50, 0x0C51,
+	/* GRAS */
+	0x0C80, 0x0C81, 0x0C88, 0x0C8F,
+	/* RB */
+	0x0CC0, 0x0CC0, 0x0CC4, 0x0CD2,
+	/* PC */
+	0x0D00, 0x0D0C, 0x0D10, 0x0D17, 0x0D20, 0x0D23,
+	/* VFD */
+	0x0E40, 0x0E4A,
+	/* VPC */
+	0x0E60, 0x0E61, 0x0E63, 0x0E68,
+	/* UCHE */
+	0x0E80, 0x0E84, 0x0E88, 0x0E95,
+	/* VMIDMT */
+	0x1000, 0x1000, 0x1002, 0x1002, 0x1004, 0x1004, 0x1008, 0x100A,
+	0x100C, 0x100D, 0x100F, 0x1010, 0x1012, 0x1016, 0x1024, 0x1024,
+	0x1027, 0x1027, 0x1100, 0x1100, 0x1102, 0x1102, 0x1104, 0x1104,
+	0x1110, 0x1110, 0x1112, 0x1116, 0x1124, 0x1124, 0x1300, 0x1300,
+	0x1380, 0x1380,
+	/* GRAS CTX 0 */
+	0x2000, 0x2004, 0x2008, 0x2067, 0x2070, 0x2078, 0x207B, 0x216E,
+	/* PC CTX 0 */
+	0x21C0, 0x21C6, 0x21D0, 0x21D0, 0x21D9, 0x21D9, 0x21E5, 0x21E7,
+	/* VFD CTX 0 */
+	0x2200, 0x2204, 0x2208, 0x22A9,
+	/* GRAS CTX 1 */
+	0x2400, 0x2404, 0x2408, 0x2467, 0x2470, 0x2478, 0x247B, 0x256E,
+	/* PC CTX 1 */
+	0x25C0, 0x25C6, 0x25D0, 0x25D0, 0x25D9, 0x25D9, 0x25E5, 0x25E7,
+	/* VFD CTX 1 */
+	0x2600, 0x2604, 0x2608, 0x26A9,
+	/* XPU */
+	0x2C00, 0x2C01, 0x2C10, 0x2C10, 0x2C12, 0x2C16, 0x2C1D, 0x2C20,
+	0x2C28, 0x2C28, 0x2C30, 0x2C30, 0x2C32, 0x2C36, 0x2C40, 0x2C40,
+	0x2C50, 0x2C50, 0x2C52, 0x2C56, 0x2C80, 0x2C80, 0x2C94, 0x2C95,
+	/* VBIF */
+	0x3000, 0x3007, 0x300C, 0x3014, 0x3018, 0x301D, 0x3020, 0x3022,
+	0x3024, 0x3026, 0x3028, 0x302A, 0x302C, 0x302D, 0x3030, 0x3031,
+	0x3034, 0x3036, 0x3038, 0x3038, 0x303C, 0x303D, 0x3040, 0x3040,
+	0x3049, 0x3049, 0x3058, 0x3058, 0x305B, 0x3061, 0x3064, 0x3068,
+	0x306C, 0x306D, 0x3080, 0x3088, 0x308B, 0x308C, 0x3090, 0x3094,
+	0x3098, 0x3098, 0x309C, 0x309C, 0x30C0, 0x30C0, 0x30C8, 0x30C8,
+	0x30D0, 0x30D0, 0x30D8, 0x30D8, 0x30E0, 0x30E0, 0x3100, 0x3100,
+	0x3108, 0x3108, 0x3110, 0x3110, 0x3118, 0x3118, 0x3120, 0x3120,
+	0x3124, 0x3125, 0x3129, 0x3129, 0x3131, 0x3131, 0x330C, 0x330C,
+	0x3310, 0x3310, 0x3400, 0x3401, 0x3410, 0x3410, 0x3412, 0x3416,
+	0x341D, 0x3420, 0x3428, 0x3428, 0x3430, 0x3430, 0x3432, 0x3436,
+	0x3440, 0x3440, 0x3450, 0x3450, 0x3452, 0x3456, 0x3480, 0x3480,
+	0x3494, 0x3495, 0x4000, 0x4000, 0x4002, 0x4002, 0x4004, 0x4004,
+	0x4008, 0x400A, 0x400C, 0x400D, 0x400F, 0x4012, 0x4014, 0x4016,
+	0x401D, 0x401D, 0x4020, 0x4027, 0x4060, 0x4062, 0x4200, 0x4200,
+	0x4300, 0x4300, 0x4400, 0x4400, 0x4500, 0x4500, 0x4800, 0x4802,
+	0x480F, 0x480F, 0x4811, 0x4811, 0x4813, 0x4813, 0x4815, 0x4816,
+	0x482B, 0x482B, 0x4857, 0x4857, 0x4883, 0x4883, 0x48AF, 0x48AF,
+	0x48C5, 0x48C5, 0x48E5, 0x48E5, 0x4905, 0x4905, 0x4925, 0x4925,
+	0x4945, 0x4945, 0x4950, 0x4950, 0x495B, 0x495B, 0x4980, 0x498E,
+	0x4B00, 0x4B00, 0x4C00, 0x4C00, 0x4D00, 0x4D00, 0x4E00, 0x4E00,
+	0x4E80, 0x4E80, 0x4F00, 0x4F00, 0x4F08, 0x4F08, 0x4F10, 0x4F10,
+	0x4F18, 0x4F18, 0x4F20, 0x4F20, 0x4F30, 0x4F30, 0x4F60, 0x4F60,
+	0x4F80, 0x4F81, 0x4F88, 0x4F89, 0x4FEE, 0x4FEE, 0x4FF3, 0x4FF3,
+	0x6000, 0x6001, 0x6008, 0x600F, 0x6014, 0x6016, 0x6018, 0x601B,
+	0x61FD, 0x61FD, 0x623C, 0x623C, 0x6380, 0x6380, 0x63A0, 0x63A0,
+	0x63C0, 0x63C1, 0x63C8, 0x63C9, 0x63D0, 0x63D4, 0x63D6, 0x63D6,
+	0x63EE, 0x63EE, 0x6400, 0x6401, 0x6408, 0x640F, 0x6414, 0x6416,
+	0x6418, 0x641B, 0x65FD, 0x65FD, 0x663C, 0x663C, 0x6780, 0x6780,
+	0x67A0, 0x67A0, 0x67C0, 0x67C1, 0x67C8, 0x67C9, 0x67D0, 0x67D4,
+	0x67D6, 0x67D6, 0x67EE, 0x67EE, 0x6800, 0x6801, 0x6808, 0x680F,
+	0x6814, 0x6816, 0x6818, 0x681B, 0x69FD, 0x69FD, 0x6A3C, 0x6A3C,
+	0x6B80, 0x6B80, 0x6BA0, 0x6BA0, 0x6BC0, 0x6BC1, 0x6BC8, 0x6BC9,
+	0x6BD0, 0x6BD4, 0x6BD6, 0x6BD6, 0x6BEE, 0x6BEE,
+	~0 /* sentinel */
+};
+
+#ifdef CONFIG_DEBUG_FS
+static void a4xx_show(struct msm_gpu *gpu, struct seq_file *m)
+{
+	gpu->funcs->pm_resume(gpu);
+
+	seq_printf(m, "status:   %08x\n",
+			gpu_read(gpu, REG_A4XX_RBBM_STATUS));
+	gpu->funcs->pm_suspend(gpu);
+
+	adreno_show(gpu, m);
+
+}
+#endif
+
+/* Register offset defines for A4XX, in order of enum adreno_regs */
+static const unsigned int a4xx_register_offsets[REG_ADRENO_REGISTER_MAX] = {
+	REG_ADRENO_DEFINE(REG_ADRENO_CP_DEBUG, REG_A4XX_CP_DEBUG),
+	REG_ADRENO_DEFINE(REG_ADRENO_CP_ME_RAM_WADDR, REG_A4XX_CP_ME_RAM_WADDR),
+	REG_ADRENO_DEFINE(REG_ADRENO_CP_ME_RAM_DATA, REG_A4XX_CP_ME_RAM_DATA),
+	REG_ADRENO_DEFINE(REG_ADRENO_CP_PFP_UCODE_DATA,
+			REG_A4XX_CP_PFP_UCODE_DATA),
+	REG_ADRENO_DEFINE(REG_ADRENO_CP_PFP_UCODE_ADDR,
+			REG_A4XX_CP_PFP_UCODE_ADDR),
+	REG_ADRENO_DEFINE(REG_ADRENO_CP_WFI_PEND_CTR, REG_A4XX_CP_WFI_PEND_CTR),
+	REG_ADRENO_DEFINE(REG_ADRENO_CP_RB_BASE, REG_A4XX_CP_RB_BASE),
+	REG_ADRENO_DEFINE(REG_ADRENO_CP_RB_RPTR_ADDR, REG_A4XX_CP_RB_RPTR_ADDR),
+	REG_ADRENO_DEFINE(REG_ADRENO_CP_RB_RPTR, REG_A4XX_CP_RB_RPTR),
+	REG_ADRENO_DEFINE(REG_ADRENO_CP_RB_WPTR, REG_A4XX_CP_RB_WPTR),
+	REG_ADRENO_DEFINE(REG_ADRENO_CP_PROTECT_CTRL, REG_A4XX_CP_PROTECT_CTRL),
+	REG_ADRENO_DEFINE(REG_ADRENO_CP_ME_CNTL, REG_A4XX_CP_ME_CNTL),
+	REG_ADRENO_DEFINE(REG_ADRENO_CP_RB_CNTL, REG_A4XX_CP_RB_CNTL),
+	REG_ADRENO_DEFINE(REG_ADRENO_CP_IB1_BASE, REG_A4XX_CP_IB1_BASE),
+	REG_ADRENO_DEFINE(REG_ADRENO_CP_IB1_BUFSZ, REG_A4XX_CP_IB1_BUFSZ),
+	REG_ADRENO_DEFINE(REG_ADRENO_CP_IB2_BASE, REG_A4XX_CP_IB2_BASE),
+	REG_ADRENO_DEFINE(REG_ADRENO_CP_IB2_BUFSZ, REG_A4XX_CP_IB2_BUFSZ),
+	REG_ADRENO_DEFINE(REG_ADRENO_CP_TIMESTAMP, REG_AXXX_CP_SCRATCH_REG0),
+	REG_ADRENO_DEFINE(REG_ADRENO_CP_ME_RAM_RADDR, REG_A4XX_CP_ME_RAM_RADDR),
+	REG_ADRENO_DEFINE(REG_ADRENO_CP_ROQ_ADDR, REG_A4XX_CP_ROQ_ADDR),
+	REG_ADRENO_DEFINE(REG_ADRENO_CP_ROQ_DATA, REG_A4XX_CP_ROQ_DATA),
+	REG_ADRENO_DEFINE(REG_ADRENO_CP_MERCIU_ADDR, REG_A4XX_CP_MERCIU_ADDR),
+	REG_ADRENO_DEFINE(REG_ADRENO_CP_MERCIU_DATA, REG_A4XX_CP_MERCIU_DATA),
+	REG_ADRENO_DEFINE(REG_ADRENO_CP_MERCIU_DATA2, REG_A4XX_CP_MERCIU_DATA2),
+	REG_ADRENO_DEFINE(REG_ADRENO_CP_MEQ_ADDR, REG_A4XX_CP_MEQ_ADDR),
+	REG_ADRENO_DEFINE(REG_ADRENO_CP_MEQ_DATA, REG_A4XX_CP_MEQ_DATA),
+	REG_ADRENO_DEFINE(REG_ADRENO_CP_HW_FAULT, REG_A4XX_CP_HW_FAULT),
+	REG_ADRENO_DEFINE(REG_ADRENO_CP_PROTECT_STATUS,
+			REG_A4XX_CP_PROTECT_STATUS),
+	REG_ADRENO_DEFINE(REG_ADRENO_SCRATCH_ADDR, REG_A4XX_CP_SCRATCH_ADDR),
+	REG_ADRENO_DEFINE(REG_ADRENO_SCRATCH_UMSK, REG_A4XX_CP_SCRATCH_UMASK),
+	REG_ADRENO_DEFINE(REG_ADRENO_RBBM_STATUS, REG_A4XX_RBBM_STATUS),
+	REG_ADRENO_DEFINE(REG_ADRENO_RBBM_PERFCTR_CTL,
+			REG_A4XX_RBBM_PERFCTR_CTL),
+	REG_ADRENO_DEFINE(REG_ADRENO_RBBM_PERFCTR_LOAD_CMD0,
+			REG_A4XX_RBBM_PERFCTR_LOAD_CMD0),
+	REG_ADRENO_DEFINE(REG_ADRENO_RBBM_PERFCTR_LOAD_CMD1,
+			REG_A4XX_RBBM_PERFCTR_LOAD_CMD1),
+	REG_ADRENO_DEFINE(REG_ADRENO_RBBM_PERFCTR_LOAD_CMD2,
+			REG_A4XX_RBBM_PERFCTR_LOAD_CMD2),
+	REG_ADRENO_DEFINE(REG_ADRENO_RBBM_PERFCTR_PWR_1_LO,
+			REG_A4XX_RBBM_PERFCTR_PWR_1_LO),
+	REG_ADRENO_DEFINE(REG_ADRENO_RBBM_INT_0_MASK, REG_A4XX_RBBM_INT_0_MASK),
+	REG_ADRENO_DEFINE(REG_ADRENO_RBBM_INT_0_STATUS,
+			REG_A4XX_RBBM_INT_0_STATUS),
+	REG_ADRENO_DEFINE(REG_ADRENO_RBBM_AHB_ERROR_STATUS,
+			REG_A4XX_RBBM_AHB_ERROR_STATUS),
+	REG_ADRENO_DEFINE(REG_ADRENO_RBBM_AHB_CMD, REG_A4XX_RBBM_AHB_CMD),
+	REG_ADRENO_DEFINE(REG_ADRENO_RBBM_CLOCK_CTL, REG_A4XX_RBBM_CLOCK_CTL),
+	REG_ADRENO_DEFINE(REG_ADRENO_RBBM_AHB_ME_SPLIT_STATUS,
+			REG_A4XX_RBBM_AHB_ME_SPLIT_STATUS),
+	REG_ADRENO_DEFINE(REG_ADRENO_RBBM_AHB_PFP_SPLIT_STATUS,
+			REG_A4XX_RBBM_AHB_PFP_SPLIT_STATUS),
+	REG_ADRENO_DEFINE(REG_ADRENO_VPC_DEBUG_RAM_SEL,
+			REG_A4XX_VPC_DEBUG_RAM_SEL),
+	REG_ADRENO_DEFINE(REG_ADRENO_VPC_DEBUG_RAM_READ,
+			REG_A4XX_VPC_DEBUG_RAM_READ),
+	REG_ADRENO_DEFINE(REG_ADRENO_RBBM_INT_CLEAR_CMD,
+			REG_A4XX_RBBM_INT_CLEAR_CMD),
+	REG_ADRENO_DEFINE(REG_ADRENO_VSC_SIZE_ADDRESS,
+			REG_A4XX_VSC_SIZE_ADDRESS),
+	REG_ADRENO_DEFINE(REG_ADRENO_VFD_CONTROL_0, REG_A4XX_VFD_CONTROL_0),
+	REG_ADRENO_DEFINE(REG_ADRENO_SP_VS_PVT_MEM_ADDR_REG,
+			REG_A4XX_SP_VS_PVT_MEM_ADDR),
+	REG_ADRENO_DEFINE(REG_ADRENO_SP_FS_PVT_MEM_ADDR_REG,
+			REG_A4XX_SP_FS_PVT_MEM_ADDR),
+	REG_ADRENO_DEFINE(REG_ADRENO_SP_VS_OBJ_START_REG,
+			REG_A4XX_SP_VS_OBJ_START),
+	REG_ADRENO_DEFINE(REG_ADRENO_SP_FS_OBJ_START_REG,
+			REG_A4XX_SP_FS_OBJ_START),
+	REG_ADRENO_DEFINE(REG_ADRENO_RBBM_RBBM_CTL, REG_A4XX_RBBM_RBBM_CTL),
+	REG_ADRENO_DEFINE(REG_ADRENO_RBBM_SW_RESET_CMD,
+			REG_A4XX_RBBM_SW_RESET_CMD),
+	REG_ADRENO_DEFINE(REG_ADRENO_UCHE_INVALIDATE0,
+			REG_A4XX_UCHE_INVALIDATE0),
+	REG_ADRENO_DEFINE(REG_ADRENO_RBBM_PERFCTR_LOAD_VALUE_LO,
+			REG_A4XX_RBBM_PERFCTR_LOAD_VALUE_LO),
+	REG_ADRENO_DEFINE(REG_ADRENO_RBBM_PERFCTR_LOAD_VALUE_HI,
+			REG_A4XX_RBBM_PERFCTR_LOAD_VALUE_HI),
+};
+
+static void a4xx_dump(struct msm_gpu *gpu)
+{
+	adreno_dump(gpu);
+	printk("status:   %08x\n",
+			gpu_read(gpu, REG_A4XX_RBBM_STATUS));
+	adreno_dump(gpu);
+}
+
+static const struct adreno_gpu_funcs funcs = {
+	.base = {
+		.get_param = adreno_get_param,
+		.hw_init = a4xx_hw_init,
+		.pm_suspend = msm_gpu_pm_suspend,
+		.pm_resume = msm_gpu_pm_resume,
+		.recover = a4xx_recover,
+		.last_fence = adreno_last_fence,
+		.submit = adreno_submit,
+		.flush = adreno_flush,
+		.idle = a4xx_idle,
+		.irq = a4xx_irq,
+		.destroy = a4xx_destroy,
+#ifdef CONFIG_DEBUG_FS
+		.show = a4xx_show,
+#endif
+	},
+};
+
+struct msm_gpu *a4xx_gpu_init(struct drm_device *dev)
+{
+	struct a4xx_gpu *a4xx_gpu = NULL;
+	struct adreno_gpu *adreno_gpu;
+	struct msm_gpu *gpu;
+	struct msm_drm_private *priv = dev->dev_private;
+	struct platform_device *pdev = priv->gpu_pdev;
+	int ret;
+
+	if (!pdev) {
+		dev_err(dev->dev, "no a4xx device\n");
+		ret = -ENXIO;
+		goto fail;
+	}
+
+	a4xx_gpu = kzalloc(sizeof(*a4xx_gpu), GFP_KERNEL);
+	if (!a4xx_gpu) {
+		ret = -ENOMEM;
+		goto fail;
+	}
+
+	adreno_gpu = &a4xx_gpu->base;
+	gpu = &adreno_gpu->base;
+
+	a4xx_gpu->pdev = pdev;
+
+	gpu->perfcntrs = NULL;
+	gpu->num_perfcntrs = 0;
+
+	adreno_gpu->registers = a4xx_registers;
+	adreno_gpu->reg_offsets = a4xx_register_offsets;
+
+	ret = adreno_gpu_init(dev, pdev, adreno_gpu, &funcs);
+	if (ret)
+		goto fail;
+
+	/* if needed, allocate gmem: */
+	if (adreno_is_a4xx(adreno_gpu)) {
+#ifdef CONFIG_MSM_OCMEM
+		/* TODO this is different/missing upstream: */
+		struct ocmem_buf *ocmem_hdl =
+				ocmem_allocate(OCMEM_GRAPHICS, adreno_gpu->gmem);
+
+		a4xx_gpu->ocmem_hdl = ocmem_hdl;
+		a4xx_gpu->ocmem_base = ocmem_hdl->addr;
+		adreno_gpu->gmem = ocmem_hdl->len;
+		DBG("using %dK of OCMEM at 0x%08x", adreno_gpu->gmem / 1024,
+				a4xx_gpu->ocmem_base);
+#endif
+	}
+
+	if (!gpu->mmu) {
+		/* TODO we think it is possible to configure the GPU to
+		 * restrict access to VRAM carveout.  But the required
+		 * registers are unknown.  For now just bail out and
+		 * limp along with just modesetting.  If it turns out
+		 * to not be possible to restrict access, then we must
+		 * implement a cmdstream validator.
+		 */
+		dev_err(dev->dev, "No memory protection without IOMMU\n");
+		ret = -ENXIO;
+		goto fail;
+	}
+
+	return gpu;
+
+fail:
+	if (a4xx_gpu)
+		a4xx_destroy(&a4xx_gpu->base.base);
+
+	return ERR_PTR(ret);
+}
diff --git a/drivers/gpu/drm/msm/adreno/a4xx_gpu.h b/drivers/gpu/drm/msm/adreno/a4xx_gpu.h
new file mode 100644
index 000000000000..01247204ac92
--- /dev/null
+++ b/drivers/gpu/drm/msm/adreno/a4xx_gpu.h
@@ -0,0 +1,34 @@
+/* Copyright (c) 2014 The Linux Foundation. All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 and
+ * only version 2 as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ */
+#ifndef __A4XX_GPU_H__
+#define __A4XX_GPU_H__
+
+#include "adreno_gpu.h"
+
+/* arrg, somehow fb.h is getting pulled in: */
+#undef ROP_COPY
+#undef ROP_XOR
+
+#include "a4xx.xml.h"
+
+struct a4xx_gpu {
+	struct adreno_gpu base;
+	struct platform_device *pdev;
+
+	/* if OCMEM is used for GMEM: */
+	uint32_t ocmem_base;
+	void *ocmem_hdl;
+};
+#define to_a4xx_gpu(x) container_of(x, struct a4xx_gpu, base)
+
+#endif /* __A4XX_GPU_H__ */
diff --git a/drivers/gpu/drm/msm/adreno/adreno_common.xml.h b/drivers/gpu/drm/msm/adreno/adreno_common.xml.h
index 9de19ac2e86c..a4b33af9338d 100644
--- a/drivers/gpu/drm/msm/adreno/adreno_common.xml.h
+++ b/drivers/gpu/drm/msm/adreno/adreno_common.xml.h
@@ -11,10 +11,10 @@ The rules-ng-ng source files this header was generated from are:
 - /home/robclark/src/freedreno/envytools/rnndb/adreno.xml               (    364 bytes, from 2013-11-30 14:47:15)
 - /home/robclark/src/freedreno/envytools/rnndb/freedreno_copyright.xml  (   1453 bytes, from 2013-03-31 16:51:27)
 - /home/robclark/src/freedreno/envytools/rnndb/adreno/a2xx.xml          (  32901 bytes, from 2014-06-02 15:21:30)
-- /home/robclark/src/freedreno/envytools/rnndb/adreno/adreno_common.xml (   9859 bytes, from 2014-06-02 15:21:30)
-- /home/robclark/src/freedreno/envytools/rnndb/adreno/adreno_pm4.xml    (  14477 bytes, from 2014-05-16 11:51:57)
-- /home/robclark/src/freedreno/envytools/rnndb/adreno/a3xx.xml          (  58020 bytes, from 2014-06-25 12:57:16)
-- /home/robclark/src/freedreno/envytools/rnndb/adreno/a4xx.xml          (  26602 bytes, from 2014-06-25 12:57:16)
+- /home/robclark/src/freedreno/envytools/rnndb/adreno/adreno_common.xml (  10551 bytes, from 2014-11-13 22:44:30)
+- /home/robclark/src/freedreno/envytools/rnndb/adreno/adreno_pm4.xml    (  15053 bytes, from 2014-11-09 15:45:47)
+- /home/robclark/src/freedreno/envytools/rnndb/adreno/a3xx.xml          (  63169 bytes, from 2014-11-13 22:44:18)
+- /home/robclark/src/freedreno/envytools/rnndb/adreno/a4xx.xml          (  49097 bytes, from 2014-11-14 15:38:00)
 
 Copyright (C) 2013-2014 by the following authors:
 - Rob Clark <robdclark@gmail.com> (robclark)
@@ -105,6 +105,7 @@ enum adreno_rb_dither_mode {
 enum adreno_rb_depth_format {
 	DEPTHX_16 = 0,
 	DEPTHX_24_8 = 1,
+	DEPTHX_32 = 2,
 };
 
 enum adreno_rb_copy_control_mode {
@@ -132,6 +133,7 @@ enum a3xx_threadmode {
 };
 
 enum a3xx_instrbuffermode {
+	CACHE = 0,
 	BUFFER = 1,
 };
 
@@ -140,6 +142,13 @@ enum a3xx_threadsize {
 	FOUR_QUADS = 1,
 };
 
+enum a3xx_color_swap {
+	WZYX = 0,
+	WXYZ = 1,
+	ZYXW = 2,
+	XYZW = 3,
+};
+
 #define REG_AXXX_CP_RB_BASE					0x000001c0
 
 #define REG_AXXX_CP_RB_CNTL					0x000001c1
diff --git a/drivers/gpu/drm/msm/adreno/adreno_device.c b/drivers/gpu/drm/msm/adreno/adreno_device.c
new file mode 100644
index 000000000000..be83dee83d08
--- /dev/null
+++ b/drivers/gpu/drm/msm/adreno/adreno_device.c
@@ -0,0 +1,298 @@
+/*
+ * Copyright (C) 2013-2014 Red Hat
+ * Author: Rob Clark <robdclark@gmail.com>
+ *
+ * Copyright (c) 2014 The Linux Foundation. All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published by
+ * the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along with
+ * this program.  If not, see <http://www.gnu.org/licenses/>.
+ */
+
+#include "adreno_gpu.h"
+
+#if defined(CONFIG_MSM_BUS_SCALING) && !defined(CONFIG_OF)
+#  include <mach/kgsl.h>
+#endif
+
+#define ANY_ID 0xff
+
+bool hang_debug = false;
+MODULE_PARM_DESC(hang_debug, "Dump registers when hang is detected (can be slow!)");
+module_param_named(hang_debug, hang_debug, bool, 0600);
+
+struct msm_gpu *a3xx_gpu_init(struct drm_device *dev);
+struct msm_gpu *a4xx_gpu_init(struct drm_device *dev);
+
+static const struct adreno_info gpulist[] = {
+	{
+		.rev   = ADRENO_REV(3, 0, 5, ANY_ID),
+		.revn  = 305,
+		.name  = "A305",
+		.pm4fw = "a300_pm4.fw",
+		.pfpfw = "a300_pfp.fw",
+		.gmem  = SZ_256K,
+		.init  = a3xx_gpu_init,
+	}, {
+		.rev   = ADRENO_REV(3, 2, ANY_ID, ANY_ID),
+		.revn  = 320,
+		.name  = "A320",
+		.pm4fw = "a300_pm4.fw",
+		.pfpfw = "a300_pfp.fw",
+		.gmem  = SZ_512K,
+		.init  = a3xx_gpu_init,
+	}, {
+		.rev   = ADRENO_REV(3, 3, 0, ANY_ID),
+		.revn  = 330,
+		.name  = "A330",
+		.pm4fw = "a330_pm4.fw",
+		.pfpfw = "a330_pfp.fw",
+		.gmem  = SZ_1M,
+		.init  = a3xx_gpu_init,
+	}, {
+		.rev   = ADRENO_REV(4, 2, 0, ANY_ID),
+		.revn  = 420,
+		.name  = "A420",
+		.pm4fw = "a420_pm4.fw",
+		.pfpfw = "a420_pfp.fw",
+		.gmem  = (SZ_1M + SZ_512K),
+		.init  = a4xx_gpu_init,
+	},
+};
+
+MODULE_FIRMWARE("a300_pm4.fw");
+MODULE_FIRMWARE("a300_pfp.fw");
+MODULE_FIRMWARE("a330_pm4.fw");
+MODULE_FIRMWARE("a330_pfp.fw");
+MODULE_FIRMWARE("a420_pm4.fw");
+MODULE_FIRMWARE("a420_pfp.fw");
+
+static inline bool _rev_match(uint8_t entry, uint8_t id)
+{
+	return (entry == ANY_ID) || (entry == id);
+}
+
+const struct adreno_info *adreno_info(struct adreno_rev rev)
+{
+	int i;
+
+	/* identify gpu: */
+	for (i = 0; i < ARRAY_SIZE(gpulist); i++) {
+		const struct adreno_info *info = &gpulist[i];
+		if (_rev_match(info->rev.core, rev.core) &&
+				_rev_match(info->rev.major, rev.major) &&
+				_rev_match(info->rev.minor, rev.minor) &&
+				_rev_match(info->rev.patchid, rev.patchid))
+			return info;
+	}
+
+	return NULL;
+}
+
+struct msm_gpu *adreno_load_gpu(struct drm_device *dev)
+{
+	struct msm_drm_private *priv = dev->dev_private;
+	struct platform_device *pdev = priv->gpu_pdev;
+	struct adreno_platform_config *config;
+	struct adreno_rev rev;
+	const struct adreno_info *info;
+	struct msm_gpu *gpu = NULL;
+
+	if (!pdev) {
+		dev_err(dev->dev, "no adreno device\n");
+		return NULL;
+	}
+
+	config = pdev->dev.platform_data;
+	rev = config->rev;
+	info = adreno_info(config->rev);
+
+	if (!info) {
+		dev_warn(dev->dev, "Unknown GPU revision: %u.%u.%u.%u\n",
+				rev.core, rev.major, rev.minor, rev.patchid);
+		return NULL;
+	}
+
+	DBG("Found GPU: %u.%u.%u.%u",  rev.core, rev.major,
+			rev.minor, rev.patchid);
+
+	gpu = info->init(dev);
+	if (IS_ERR(gpu)) {
+		dev_warn(dev->dev, "failed to load adreno gpu\n");
+		gpu = NULL;
+		/* not fatal */
+	}
+
+	if (gpu) {
+		int ret;
+		mutex_lock(&dev->struct_mutex);
+		gpu->funcs->pm_resume(gpu);
+		mutex_unlock(&dev->struct_mutex);
+		ret = gpu->funcs->hw_init(gpu);
+		if (ret) {
+			dev_err(dev->dev, "gpu hw init failed: %d\n", ret);
+			gpu->funcs->destroy(gpu);
+			gpu = NULL;
+		} else {
+			/* give inactive pm a chance to kick in: */
+			msm_gpu_retire(gpu);
+		}
+	}
+
+	return gpu;
+}
+
+static void set_gpu_pdev(struct drm_device *dev,
+		struct platform_device *pdev)
+{
+	struct msm_drm_private *priv = dev->dev_private;
+	priv->gpu_pdev = pdev;
+}
+
+static int adreno_bind(struct device *dev, struct device *master, void *data)
+{
+	static struct adreno_platform_config config = {};
+#ifdef CONFIG_OF
+	struct device_node *child, *node = dev->of_node;
+	u32 val;
+	int ret;
+
+	ret = of_property_read_u32(node, "qcom,chipid", &val);
+	if (ret) {
+		dev_err(dev, "could not find chipid: %d\n", ret);
+		return ret;
+	}
+
+	config.rev = ADRENO_REV((val >> 24) & 0xff,
+			(val >> 16) & 0xff, (val >> 8) & 0xff, val & 0xff);
+
+	/* find clock rates: */
+	config.fast_rate = 0;
+	config.slow_rate = ~0;
+	for_each_child_of_node(node, child) {
+		if (of_device_is_compatible(child, "qcom,gpu-pwrlevels")) {
+			struct device_node *pwrlvl;
+			for_each_child_of_node(child, pwrlvl) {
+				ret = of_property_read_u32(pwrlvl, "qcom,gpu-freq", &val);
+				if (ret) {
+					dev_err(dev, "could not find gpu-freq: %d\n", ret);
+					return ret;
+				}
+				config.fast_rate = max(config.fast_rate, val);
+				config.slow_rate = min(config.slow_rate, val);
+			}
+		}
+	}
+
+	if (!config.fast_rate) {
+		dev_err(dev, "could not find clk rates\n");
+		return -ENXIO;
+	}
+
+#else
+	struct kgsl_device_platform_data *pdata = dev->platform_data;
+	uint32_t version = socinfo_get_version();
+	if (cpu_is_apq8064ab()) {
+		config.fast_rate = 450000000;
+		config.slow_rate = 27000000;
+		config.bus_freq  = 4;
+		config.rev = ADRENO_REV(3, 2, 1, 0);
+	} else if (cpu_is_apq8064()) {
+		config.fast_rate = 400000000;
+		config.slow_rate = 27000000;
+		config.bus_freq  = 4;
+
+		if (SOCINFO_VERSION_MAJOR(version) == 2)
+			config.rev = ADRENO_REV(3, 2, 0, 2);
+		else if ((SOCINFO_VERSION_MAJOR(version) == 1) &&
+				(SOCINFO_VERSION_MINOR(version) == 1))
+			config.rev = ADRENO_REV(3, 2, 0, 1);
+		else
+			config.rev = ADRENO_REV(3, 2, 0, 0);
+
+	} else if (cpu_is_msm8960ab()) {
+		config.fast_rate = 400000000;
+		config.slow_rate = 320000000;
+		config.bus_freq  = 4;
+
+		if (SOCINFO_VERSION_MINOR(version) == 0)
+			config.rev = ADRENO_REV(3, 2, 1, 0);
+		else
+			config.rev = ADRENO_REV(3, 2, 1, 1);
+
+	} else if (cpu_is_msm8930()) {
+		config.fast_rate = 400000000;
+		config.slow_rate = 27000000;
+		config.bus_freq  = 3;
+
+		if ((SOCINFO_VERSION_MAJOR(version) == 1) &&
+			(SOCINFO_VERSION_MINOR(version) == 2))
+			config.rev = ADRENO_REV(3, 0, 5, 2);
+		else
+			config.rev = ADRENO_REV(3, 0, 5, 0);
+
+	}
+#  ifdef CONFIG_MSM_BUS_SCALING
+	config.bus_scale_table = pdata->bus_scale_table;
+#  endif
+#endif
+	dev->platform_data = &config;
+	set_gpu_pdev(dev_get_drvdata(master), to_platform_device(dev));
+	return 0;
+}
+
+static void adreno_unbind(struct device *dev, struct device *master,
+		void *data)
+{
+	set_gpu_pdev(dev_get_drvdata(master), NULL);
+}
+
+static const struct component_ops a3xx_ops = {
+		.bind   = adreno_bind,
+		.unbind = adreno_unbind,
+};
+
+static int adreno_probe(struct platform_device *pdev)
+{
+	return component_add(&pdev->dev, &a3xx_ops);
+}
+
+static int adreno_remove(struct platform_device *pdev)
+{
+	component_del(&pdev->dev, &a3xx_ops);
+	return 0;
+}
+
+static const struct of_device_id dt_match[] = {
+	{ .compatible = "qcom,adreno-3xx" },
+	/* for backwards compat w/ downstream kgsl DT files: */
+	{ .compatible = "qcom,kgsl-3d0" },
+	{}
+};
+
+static struct platform_driver adreno_driver = {
+	.probe = adreno_probe,
+	.remove = adreno_remove,
+	.driver = {
+		.name = "adreno",
+		.of_match_table = dt_match,
+	},
+};
+
+void __init adreno_register(void)
+{
+	platform_driver_register(&adreno_driver);
+}
+
+void __exit adreno_unregister(void)
+{
+	platform_driver_unregister(&adreno_driver);
+}
diff --git a/drivers/gpu/drm/msm/adreno/adreno_gpu.c b/drivers/gpu/drm/msm/adreno/adreno_gpu.c
index 655ce5b14ad0..94a5bee69fe7 100644
--- a/drivers/gpu/drm/msm/adreno/adreno_gpu.c
+++ b/drivers/gpu/drm/msm/adreno/adreno_gpu.c
@@ -2,6 +2,8 @@
  * Copyright (C) 2013 Red Hat
  * Author: Rob Clark <robdclark@gmail.com>
  *
+ * Copyright (c) 2014 The Linux Foundation. All rights reserved.
+ *
  * This program is free software; you can redistribute it and/or modify it
  * under the terms of the GNU General Public License version 2 as published by
  * the Free Software Foundation.
@@ -19,46 +21,6 @@
 #include "msm_gem.h"
 #include "msm_mmu.h"
 
-struct adreno_info {
-	struct adreno_rev rev;
-	uint32_t revn;
-	const char *name;
-	const char *pm4fw, *pfpfw;
-	uint32_t gmem;
-};
-
-#define ANY_ID 0xff
-
-static const struct adreno_info gpulist[] = {
-	{
-		.rev   = ADRENO_REV(3, 0, 5, ANY_ID),
-		.revn  = 305,
-		.name  = "A305",
-		.pm4fw = "a300_pm4.fw",
-		.pfpfw = "a300_pfp.fw",
-		.gmem  = SZ_256K,
-	}, {
-		.rev   = ADRENO_REV(3, 2, ANY_ID, ANY_ID),
-		.revn  = 320,
-		.name  = "A320",
-		.pm4fw = "a300_pm4.fw",
-		.pfpfw = "a300_pfp.fw",
-		.gmem  = SZ_512K,
-	}, {
-		.rev   = ADRENO_REV(3, 3, 0, ANY_ID),
-		.revn  = 330,
-		.name  = "A330",
-		.pm4fw = "a330_pm4.fw",
-		.pfpfw = "a330_pfp.fw",
-		.gmem  = SZ_1M,
-	},
-};
-
-MODULE_FIRMWARE("a300_pm4.fw");
-MODULE_FIRMWARE("a300_pfp.fw");
-MODULE_FIRMWARE("a330_pm4.fw");
-MODULE_FIRMWARE("a330_pfp.fw");
-
 #define RB_SIZE    SZ_32K
 #define RB_BLKSIZE 16
 
@@ -103,19 +65,21 @@ int adreno_hw_init(struct msm_gpu *gpu)
 	}
 
 	/* Setup REG_CP_RB_CNTL: */
-	gpu_write(gpu, REG_AXXX_CP_RB_CNTL,
+	adreno_gpu_write(adreno_gpu, REG_ADRENO_CP_RB_CNTL,
 			/* size is log2(quad-words): */
 			AXXX_CP_RB_CNTL_BUFSZ(ilog2(gpu->rb->size / 8)) |
 			AXXX_CP_RB_CNTL_BLKSZ(ilog2(RB_BLKSIZE / 8)));
 
 	/* Setup ringbuffer address: */
-	gpu_write(gpu, REG_AXXX_CP_RB_BASE, gpu->rb_iova);
-	gpu_write(gpu, REG_AXXX_CP_RB_RPTR_ADDR, rbmemptr(adreno_gpu, rptr));
+	adreno_gpu_write(adreno_gpu, REG_ADRENO_CP_RB_BASE, gpu->rb_iova);
+	adreno_gpu_write(adreno_gpu, REG_ADRENO_CP_RB_RPTR_ADDR,
+			rbmemptr(adreno_gpu, rptr));
 
 	/* Setup scratch/timestamp: */
-	gpu_write(gpu, REG_AXXX_SCRATCH_ADDR, rbmemptr(adreno_gpu, fence));
+	adreno_gpu_write(adreno_gpu, REG_ADRENO_SCRATCH_ADDR,
+			rbmemptr(adreno_gpu, fence));
 
-	gpu_write(gpu, REG_AXXX_SCRATCH_UMSK, 0x1);
+	adreno_gpu_write(adreno_gpu, REG_ADRENO_SCRATCH_UMSK, 0x1);
 
 	return 0;
 }
@@ -191,7 +155,7 @@ int adreno_submit(struct msm_gpu *gpu, struct msm_gem_submit *submit,
 	OUT_PKT0(ring, REG_AXXX_CP_SCRATCH_REG2, 1);
 	OUT_RING(ring, submit->fence);
 
-	if (adreno_is_a3xx(adreno_gpu)) {
+	if (adreno_is_a3xx(adreno_gpu) || adreno_is_a4xx(adreno_gpu)) {
 		/* Flush HLSQ lazy updates to make sure there is nothing
 		 * pending for indirect loads after the timestamp has
 		 * passed:
@@ -228,12 +192,13 @@ int adreno_submit(struct msm_gpu *gpu, struct msm_gem_submit *submit,
 
 void adreno_flush(struct msm_gpu *gpu)
 {
+	struct adreno_gpu *adreno_gpu = to_adreno_gpu(gpu);
 	uint32_t wptr = get_wptr(gpu->rb);
 
 	/* ensure writes to ringbuffer have hit system memory: */
 	mb();
 
-	gpu_write(gpu, REG_AXXX_CP_RB_WPTR, wptr);
+	adreno_gpu_write(adreno_gpu, REG_ADRENO_CP_RB_WPTR, wptr);
 }
 
 void adreno_idle(struct msm_gpu *gpu)
@@ -252,6 +217,7 @@ void adreno_idle(struct msm_gpu *gpu)
 void adreno_show(struct msm_gpu *gpu, struct seq_file *m)
 {
 	struct adreno_gpu *adreno_gpu = to_adreno_gpu(gpu);
+	int i;
 
 	seq_printf(m, "revision: %d (%d.%d.%d.%d)\n",
 			adreno_gpu->info->revn, adreno_gpu->rev.core,
@@ -263,6 +229,23 @@ void adreno_show(struct msm_gpu *gpu, struct seq_file *m)
 	seq_printf(m, "rptr:     %d\n", adreno_gpu->memptrs->rptr);
 	seq_printf(m, "wptr:     %d\n", adreno_gpu->memptrs->wptr);
 	seq_printf(m, "rb wptr:  %d\n", get_wptr(gpu->rb));
+
+	gpu->funcs->pm_resume(gpu);
+
+	/* dump these out in a form that can be parsed by demsm: */
+	seq_printf(m, "IO:region %s 00000000 00020000\n", gpu->name);
+	for (i = 0; adreno_gpu->registers[i] != ~0; i += 2) {
+		uint32_t start = adreno_gpu->registers[i];
+		uint32_t end   = adreno_gpu->registers[i+1];
+		uint32_t addr;
+
+		for (addr = start; addr <= end; addr++) {
+			uint32_t val = gpu_read(gpu, addr);
+			seq_printf(m, "IO:R %08x %08x\n", addr<<2, val);
+		}
+	}
+
+	gpu->funcs->pm_suspend(gpu);
 }
 #endif
 
@@ -270,6 +253,7 @@ void adreno_show(struct msm_gpu *gpu, struct seq_file *m)
 void adreno_dump(struct msm_gpu *gpu)
 {
 	struct adreno_gpu *adreno_gpu = to_adreno_gpu(gpu);
+	int i;
 
 	printk("revision: %d (%d.%d.%d.%d)\n",
 			adreno_gpu->info->revn, adreno_gpu->rev.core,
@@ -282,6 +266,18 @@ void adreno_dump(struct msm_gpu *gpu)
 	printk("wptr:     %d\n", adreno_gpu->memptrs->wptr);
 	printk("rb wptr:  %d\n", get_wptr(gpu->rb));
 
+	/* dump these out in a form that can be parsed by demsm: */
+	printk("IO:region %s 00000000 00020000\n", gpu->name);
+	for (i = 0; adreno_gpu->registers[i] != ~0; i += 2) {
+		uint32_t start = adreno_gpu->registers[i];
+		uint32_t end   = adreno_gpu->registers[i+1];
+		uint32_t addr;
+
+		for (addr = start; addr <= end; addr++) {
+			uint32_t val = gpu_read(gpu, addr);
+			printk("IO:R %08x %08x\n", addr<<2, val);
+		}
+	}
 }
 
 static uint32_t ring_freewords(struct msm_gpu *gpu)
@@ -304,65 +300,51 @@ static const char *iommu_ports[] = {
 		"gfx3d1_user", "gfx3d1_priv",
 };
 
-static inline bool _rev_match(uint8_t entry, uint8_t id)
-{
-	return (entry == ANY_ID) || (entry == id);
-}
-
 int adreno_gpu_init(struct drm_device *drm, struct platform_device *pdev,
-		struct adreno_gpu *gpu, const struct adreno_gpu_funcs *funcs,
-		struct adreno_rev rev)
+		struct adreno_gpu *adreno_gpu, const struct adreno_gpu_funcs *funcs)
 {
+	struct adreno_platform_config *config = pdev->dev.platform_data;
+	struct msm_gpu *gpu = &adreno_gpu->base;
 	struct msm_mmu *mmu;
-	int i, ret;
-
-	/* identify gpu: */
-	for (i = 0; i < ARRAY_SIZE(gpulist); i++) {
-		const struct adreno_info *info = &gpulist[i];
-		if (_rev_match(info->rev.core, rev.core) &&
-				_rev_match(info->rev.major, rev.major) &&
-				_rev_match(info->rev.minor, rev.minor) &&
-				_rev_match(info->rev.patchid, rev.patchid)) {
-			gpu->info = info;
-			gpu->revn = info->revn;
-			break;
-		}
-	}
+	int ret;
 
-	if (i == ARRAY_SIZE(gpulist)) {
-		dev_err(drm->dev, "Unknown GPU revision: %u.%u.%u.%u\n",
-				rev.core, rev.major, rev.minor, rev.patchid);
-		return -ENXIO;
-	}
+	adreno_gpu->funcs = funcs;
+	adreno_gpu->info = adreno_info(config->rev);
+	adreno_gpu->gmem = adreno_gpu->info->gmem;
+	adreno_gpu->revn = adreno_gpu->info->revn;
+	adreno_gpu->rev = config->rev;
+
+	gpu->fast_rate = config->fast_rate;
+	gpu->slow_rate = config->slow_rate;
+	gpu->bus_freq  = config->bus_freq;
+#ifdef CONFIG_MSM_BUS_SCALING
+	gpu->bus_scale_table = config->bus_scale_table;
+#endif
 
-	DBG("Found GPU: %s (%u.%u.%u.%u)", gpu->info->name,
-			rev.core, rev.major, rev.minor, rev.patchid);
+	DBG("fast_rate=%u, slow_rate=%u, bus_freq=%u",
+			gpu->fast_rate, gpu->slow_rate, gpu->bus_freq);
 
-	gpu->funcs = funcs;
-	gpu->gmem = gpu->info->gmem;
-	gpu->rev = rev;
+	ret = msm_gpu_init(drm, pdev, &adreno_gpu->base, &funcs->base,
+			adreno_gpu->info->name, "kgsl_3d0_reg_memory", "kgsl_3d0_irq",
+			RB_SIZE);
+	if (ret)
+		return ret;
 
-	ret = request_firmware(&gpu->pm4, gpu->info->pm4fw, drm->dev);
+	ret = request_firmware(&adreno_gpu->pm4, adreno_gpu->info->pm4fw, drm->dev);
 	if (ret) {
 		dev_err(drm->dev, "failed to load %s PM4 firmware: %d\n",
-				gpu->info->pm4fw, ret);
+				adreno_gpu->info->pm4fw, ret);
 		return ret;
 	}
 
-	ret = request_firmware(&gpu->pfp, gpu->info->pfpfw, drm->dev);
+	ret = request_firmware(&adreno_gpu->pfp, adreno_gpu->info->pfpfw, drm->dev);
 	if (ret) {
 		dev_err(drm->dev, "failed to load %s PFP firmware: %d\n",
-				gpu->info->pfpfw, ret);
+				adreno_gpu->info->pfpfw, ret);
 		return ret;
 	}
 
-	ret = msm_gpu_init(drm, pdev, &gpu->base, &funcs->base,
-			gpu->info->name, "kgsl_3d0_reg_memory", "kgsl_3d0_irq",
-			RB_SIZE);
-	if (ret)
-		return ret;
-
-	mmu = gpu->base.mmu;
+	mmu = gpu->mmu;
 	if (mmu) {
 		ret = mmu->funcs->attach(mmu, iommu_ports,
 				ARRAY_SIZE(iommu_ports));
@@ -371,24 +353,24 @@ int adreno_gpu_init(struct drm_device *drm, struct platform_device *pdev,
 	}
 
 	mutex_lock(&drm->struct_mutex);
-	gpu->memptrs_bo = msm_gem_new(drm, sizeof(*gpu->memptrs),
+	adreno_gpu->memptrs_bo = msm_gem_new(drm, sizeof(*adreno_gpu->memptrs),
 			MSM_BO_UNCACHED);
 	mutex_unlock(&drm->struct_mutex);
-	if (IS_ERR(gpu->memptrs_bo)) {
-		ret = PTR_ERR(gpu->memptrs_bo);
-		gpu->memptrs_bo = NULL;
+	if (IS_ERR(adreno_gpu->memptrs_bo)) {
+		ret = PTR_ERR(adreno_gpu->memptrs_bo);
+		adreno_gpu->memptrs_bo = NULL;
 		dev_err(drm->dev, "could not allocate memptrs: %d\n", ret);
 		return ret;
 	}
 
-	gpu->memptrs = msm_gem_vaddr(gpu->memptrs_bo);
-	if (!gpu->memptrs) {
+	adreno_gpu->memptrs = msm_gem_vaddr(adreno_gpu->memptrs_bo);
+	if (!adreno_gpu->memptrs) {
 		dev_err(drm->dev, "could not vmap memptrs\n");
 		return -ENOMEM;
 	}
 
-	ret = msm_gem_get_iova(gpu->memptrs_bo, gpu->base.id,
-			&gpu->memptrs_iova);
+	ret = msm_gem_get_iova(adreno_gpu->memptrs_bo, gpu->id,
+			&adreno_gpu->memptrs_iova);
 	if (ret) {
 		dev_err(drm->dev, "could not map memptrs: %d\n", ret);
 		return ret;
@@ -404,9 +386,7 @@ void adreno_gpu_cleanup(struct adreno_gpu *gpu)
 			msm_gem_put_iova(gpu->memptrs_bo, gpu->base.id);
 		drm_gem_object_unreference(gpu->memptrs_bo);
 	}
-	if (gpu->pm4)
-		release_firmware(gpu->pm4);
-	if (gpu->pfp)
-		release_firmware(gpu->pfp);
+	release_firmware(gpu->pm4);
+	release_firmware(gpu->pfp);
 	msm_gpu_cleanup(&gpu->base);
 }
diff --git a/drivers/gpu/drm/msm/adreno/adreno_gpu.h b/drivers/gpu/drm/msm/adreno/adreno_gpu.h
index 63c36ce33020..a0cc30977e67 100644
--- a/drivers/gpu/drm/msm/adreno/adreno_gpu.h
+++ b/drivers/gpu/drm/msm/adreno/adreno_gpu.h
@@ -2,6 +2,8 @@
  * Copyright (C) 2013 Red Hat
  * Author: Rob Clark <robdclark@gmail.com>
  *
+ * Copyright (c) 2014 The Linux Foundation. All rights reserved.
+ *
  * This program is free software; you can redistribute it and/or modify it
  * under the terms of the GNU General Public License version 2 as published by
  * the Free Software Foundation.
@@ -25,6 +27,81 @@
 #include "adreno_common.xml.h"
 #include "adreno_pm4.xml.h"
 
+#define REG_ADRENO_DEFINE(_offset, _reg) [_offset] = (_reg) + 1
+/**
+ * adreno_regs: List of registers that are used in across all
+ * 3D devices. Each device type has different offset value for the same
+ * register, so an array of register offsets are declared for every device
+ * and are indexed by the enumeration values defined in this enum
+ */
+enum adreno_regs {
+	REG_ADRENO_CP_DEBUG,
+	REG_ADRENO_CP_ME_RAM_WADDR,
+	REG_ADRENO_CP_ME_RAM_DATA,
+	REG_ADRENO_CP_PFP_UCODE_DATA,
+	REG_ADRENO_CP_PFP_UCODE_ADDR,
+	REG_ADRENO_CP_WFI_PEND_CTR,
+	REG_ADRENO_CP_RB_BASE,
+	REG_ADRENO_CP_RB_RPTR_ADDR,
+	REG_ADRENO_CP_RB_RPTR,
+	REG_ADRENO_CP_RB_WPTR,
+	REG_ADRENO_CP_PROTECT_CTRL,
+	REG_ADRENO_CP_ME_CNTL,
+	REG_ADRENO_CP_RB_CNTL,
+	REG_ADRENO_CP_IB1_BASE,
+	REG_ADRENO_CP_IB1_BUFSZ,
+	REG_ADRENO_CP_IB2_BASE,
+	REG_ADRENO_CP_IB2_BUFSZ,
+	REG_ADRENO_CP_TIMESTAMP,
+	REG_ADRENO_CP_ME_RAM_RADDR,
+	REG_ADRENO_CP_ROQ_ADDR,
+	REG_ADRENO_CP_ROQ_DATA,
+	REG_ADRENO_CP_MERCIU_ADDR,
+	REG_ADRENO_CP_MERCIU_DATA,
+	REG_ADRENO_CP_MERCIU_DATA2,
+	REG_ADRENO_CP_MEQ_ADDR,
+	REG_ADRENO_CP_MEQ_DATA,
+	REG_ADRENO_CP_HW_FAULT,
+	REG_ADRENO_CP_PROTECT_STATUS,
+	REG_ADRENO_SCRATCH_ADDR,
+	REG_ADRENO_SCRATCH_UMSK,
+	REG_ADRENO_SCRATCH_REG2,
+	REG_ADRENO_RBBM_STATUS,
+	REG_ADRENO_RBBM_PERFCTR_CTL,
+	REG_ADRENO_RBBM_PERFCTR_LOAD_CMD0,
+	REG_ADRENO_RBBM_PERFCTR_LOAD_CMD1,
+	REG_ADRENO_RBBM_PERFCTR_LOAD_CMD2,
+	REG_ADRENO_RBBM_PERFCTR_PWR_1_LO,
+	REG_ADRENO_RBBM_INT_0_MASK,
+	REG_ADRENO_RBBM_INT_0_STATUS,
+	REG_ADRENO_RBBM_AHB_ERROR_STATUS,
+	REG_ADRENO_RBBM_PM_OVERRIDE2,
+	REG_ADRENO_RBBM_AHB_CMD,
+	REG_ADRENO_RBBM_INT_CLEAR_CMD,
+	REG_ADRENO_RBBM_SW_RESET_CMD,
+	REG_ADRENO_RBBM_CLOCK_CTL,
+	REG_ADRENO_RBBM_AHB_ME_SPLIT_STATUS,
+	REG_ADRENO_RBBM_AHB_PFP_SPLIT_STATUS,
+	REG_ADRENO_VPC_DEBUG_RAM_SEL,
+	REG_ADRENO_VPC_DEBUG_RAM_READ,
+	REG_ADRENO_VSC_SIZE_ADDRESS,
+	REG_ADRENO_VFD_CONTROL_0,
+	REG_ADRENO_VFD_INDEX_MAX,
+	REG_ADRENO_SP_VS_PVT_MEM_ADDR_REG,
+	REG_ADRENO_SP_FS_PVT_MEM_ADDR_REG,
+	REG_ADRENO_SP_VS_OBJ_START_REG,
+	REG_ADRENO_SP_FS_OBJ_START_REG,
+	REG_ADRENO_PA_SC_AA_CONFIG,
+	REG_ADRENO_SQ_GPR_MANAGEMENT,
+	REG_ADRENO_SQ_INST_STORE_MANAGMENT,
+	REG_ADRENO_TP0_CHICKEN,
+	REG_ADRENO_RBBM_RBBM_CTL,
+	REG_ADRENO_UCHE_INVALIDATE0,
+	REG_ADRENO_RBBM_PERFCTR_LOAD_VALUE_LO,
+	REG_ADRENO_RBBM_PERFCTR_LOAD_VALUE_HI,
+	REG_ADRENO_REGISTER_MAX,
+};
+
 struct adreno_rev {
 	uint8_t  core;
 	uint8_t  major;
@@ -39,7 +116,16 @@ struct adreno_gpu_funcs {
 	struct msm_gpu_funcs base;
 };
 
-struct adreno_info;
+struct adreno_info {
+	struct adreno_rev rev;
+	uint32_t revn;
+	const char *name;
+	const char *pm4fw, *pfpfw;
+	uint32_t gmem;
+	struct msm_gpu *(*init)(struct drm_device *dev);
+};
+
+const struct adreno_info *adreno_info(struct adreno_rev rev);
 
 struct adreno_rbmemptrs {
 	volatile uint32_t rptr;
@@ -55,6 +141,9 @@ struct adreno_gpu {
 	uint32_t revn;  /* numeric revision name */
 	const struct adreno_gpu_funcs *funcs;
 
+	/* interesting register offsets to dump: */
+	const unsigned int *registers;
+
 	/* firmware: */
 	const struct firmware *pm4, *pfp;
 
@@ -64,6 +153,13 @@ struct adreno_gpu {
 	struct adreno_rbmemptrs *memptrs;
 	struct drm_gem_object *memptrs_bo;
 	uint32_t memptrs_iova;
+
+	/*
+	 * Register offsets are different between some GPUs.
+	 * GPU specific offsets will be exported by GPU specific
+	 * code (a3xx_gpu.c) and stored in this common location.
+	 */
+	const unsigned int *reg_offsets;
 };
 #define to_adreno_gpu(x) container_of(x, struct adreno_gpu, base)
 
@@ -116,6 +212,16 @@ static inline bool adreno_is_a330v2(struct adreno_gpu *gpu)
 	return adreno_is_a330(gpu) && (gpu->rev.patchid > 0);
 }
 
+static inline bool adreno_is_a4xx(struct adreno_gpu *gpu)
+{
+	return (gpu->revn >= 400) && (gpu->revn < 500);
+}
+
+static inline int adreno_is_a420(struct adreno_gpu *gpu)
+{
+	return gpu->revn == 420;
+}
+
 int adreno_get_param(struct msm_gpu *gpu, uint32_t param, uint64_t *value);
 int adreno_hw_init(struct msm_gpu *gpu);
 uint32_t adreno_last_fence(struct msm_gpu *gpu);
@@ -131,8 +237,7 @@ void adreno_dump(struct msm_gpu *gpu);
 void adreno_wait_ring(struct msm_gpu *gpu, uint32_t ndwords);
 
 int adreno_gpu_init(struct drm_device *drm, struct platform_device *pdev,
-		struct adreno_gpu *gpu, const struct adreno_gpu_funcs *funcs,
-		struct adreno_rev rev);
+		struct adreno_gpu *gpu, const struct adreno_gpu_funcs *funcs);
 void adreno_gpu_cleanup(struct adreno_gpu *gpu);
 
 
@@ -160,5 +265,37 @@ OUT_PKT3(struct msm_ringbuffer *ring, uint8_t opcode, uint16_t cnt)
 	OUT_RING(ring, CP_TYPE3_PKT | ((cnt-1) << 16) | ((opcode & 0xFF) << 8));
 }
 
+/*
+ * adreno_checkreg_off() - Checks the validity of a register enum
+ * @gpu:		Pointer to struct adreno_gpu
+ * @offset_name:	The register enum that is checked
+ */
+static inline bool adreno_reg_check(struct adreno_gpu *gpu,
+		enum adreno_regs offset_name)
+{
+	if (offset_name >= REG_ADRENO_REGISTER_MAX ||
+			!gpu->reg_offsets[offset_name]) {
+		BUG();
+	}
+	return true;
+}
+
+static inline u32 adreno_gpu_read(struct adreno_gpu *gpu,
+		enum adreno_regs offset_name)
+{
+	u32 reg = gpu->reg_offsets[offset_name];
+	u32 val = 0;
+	if(adreno_reg_check(gpu,offset_name))
+		val = gpu_read(&gpu->base, reg - 1);
+	return val;
+}
+
+static inline void adreno_gpu_write(struct adreno_gpu *gpu,
+		enum adreno_regs offset_name, u32 data)
+{
+	u32 reg = gpu->reg_offsets[offset_name];
+	if(adreno_reg_check(gpu, offset_name))
+		gpu_write(&gpu->base, reg - 1, data);
+}
 
 #endif /* __ADRENO_GPU_H__ */
diff --git a/drivers/gpu/drm/msm/adreno/adreno_pm4.xml.h b/drivers/gpu/drm/msm/adreno/adreno_pm4.xml.h
index 4eee0ec8f069..6a75cee94d81 100644
--- a/drivers/gpu/drm/msm/adreno/adreno_pm4.xml.h
+++ b/drivers/gpu/drm/msm/adreno/adreno_pm4.xml.h
@@ -11,10 +11,10 @@ The rules-ng-ng source files this header was generated from are:
 - /home/robclark/src/freedreno/envytools/rnndb/adreno.xml               (    364 bytes, from 2013-11-30 14:47:15)
 - /home/robclark/src/freedreno/envytools/rnndb/freedreno_copyright.xml  (   1453 bytes, from 2013-03-31 16:51:27)
 - /home/robclark/src/freedreno/envytools/rnndb/adreno/a2xx.xml          (  32901 bytes, from 2014-06-02 15:21:30)
-- /home/robclark/src/freedreno/envytools/rnndb/adreno/adreno_common.xml (   9859 bytes, from 2014-06-02 15:21:30)
-- /home/robclark/src/freedreno/envytools/rnndb/adreno/adreno_pm4.xml    (  14477 bytes, from 2014-05-16 11:51:57)
-- /home/robclark/src/freedreno/envytools/rnndb/adreno/a3xx.xml          (  58020 bytes, from 2014-06-25 12:57:16)
-- /home/robclark/src/freedreno/envytools/rnndb/adreno/a4xx.xml          (  26602 bytes, from 2014-06-25 12:57:16)
+- /home/robclark/src/freedreno/envytools/rnndb/adreno/adreno_common.xml (  10551 bytes, from 2014-11-13 22:44:30)
+- /home/robclark/src/freedreno/envytools/rnndb/adreno/adreno_pm4.xml    (  15053 bytes, from 2014-11-09 15:45:47)
+- /home/robclark/src/freedreno/envytools/rnndb/adreno/a3xx.xml          (  63169 bytes, from 2014-11-13 22:44:18)
+- /home/robclark/src/freedreno/envytools/rnndb/adreno/a4xx.xml          (  49097 bytes, from 2014-11-14 15:38:00)
 
 Copyright (C) 2013-2014 by the following authors:
 - Rob Clark <robdclark@gmail.com> (robclark)
@@ -157,18 +157,23 @@ enum adreno_pm4_type3_packets {
 	CP_IM_STORE = 44,
 	CP_SET_DRAW_INIT_FLAGS = 75,
 	CP_SET_PROTECTED_MODE = 95,
+	CP_BOOTSTRAP_UCODE = 111,
 	CP_LOAD_STATE = 48,
 	CP_COND_INDIRECT_BUFFER_PFE = 58,
 	CP_COND_INDIRECT_BUFFER_PFD = 50,
 	CP_INDIRECT_BUFFER_PFE = 63,
 	CP_SET_BIN = 76,
 	CP_TEST_TWO_MEMS = 113,
+	CP_REG_WR_NO_CTXT = 120,
+	CP_RECORD_PFP_TIMESTAMP = 17,
 	CP_WAIT_FOR_ME = 19,
 	CP_SET_DRAW_STATE = 67,
 	CP_DRAW_INDX_OFFSET = 56,
 	CP_DRAW_INDIRECT = 40,
 	CP_DRAW_INDX_INDIRECT = 41,
 	CP_DRAW_AUTO = 36,
+	CP_UNKNOWN_1A = 26,
+	CP_WIDE_REG_WRITE = 116,
 	IN_IB_PREFETCH_END = 23,
 	IN_SUBBLK_PREFETCH = 31,
 	IN_INSTR_PREFETCH = 32,
@@ -274,11 +279,11 @@ static inline uint32_t CP_DRAW_INDX_1_INDEX_SIZE(enum pc_di_index_size val)
 #define CP_DRAW_INDX_1_NOT_EOP					0x00001000
 #define CP_DRAW_INDX_1_SMALL_INDEX				0x00002000
 #define CP_DRAW_INDX_1_PRE_DRAW_INITIATOR_ENABLE		0x00004000
-#define CP_DRAW_INDX_1_NUM_INDICES__MASK			0xffff0000
-#define CP_DRAW_INDX_1_NUM_INDICES__SHIFT			16
-static inline uint32_t CP_DRAW_INDX_1_NUM_INDICES(uint32_t val)
+#define CP_DRAW_INDX_1_NUM_INSTANCES__MASK			0xff000000
+#define CP_DRAW_INDX_1_NUM_INSTANCES__SHIFT			24
+static inline uint32_t CP_DRAW_INDX_1_NUM_INSTANCES(uint32_t val)
 {
-	return ((val) << CP_DRAW_INDX_1_NUM_INDICES__SHIFT) & CP_DRAW_INDX_1_NUM_INDICES__MASK;
+	return ((val) << CP_DRAW_INDX_1_NUM_INSTANCES__SHIFT) & CP_DRAW_INDX_1_NUM_INSTANCES__MASK;
 }
 
 #define REG_CP_DRAW_INDX_2					0x00000002
@@ -289,20 +294,20 @@ static inline uint32_t CP_DRAW_INDX_2_NUM_INDICES(uint32_t val)
 	return ((val) << CP_DRAW_INDX_2_NUM_INDICES__SHIFT) & CP_DRAW_INDX_2_NUM_INDICES__MASK;
 }
 
-#define REG_CP_DRAW_INDX_2					0x00000002
-#define CP_DRAW_INDX_2_INDX_BASE__MASK				0xffffffff
-#define CP_DRAW_INDX_2_INDX_BASE__SHIFT				0
-static inline uint32_t CP_DRAW_INDX_2_INDX_BASE(uint32_t val)
+#define REG_CP_DRAW_INDX_3					0x00000003
+#define CP_DRAW_INDX_3_INDX_BASE__MASK				0xffffffff
+#define CP_DRAW_INDX_3_INDX_BASE__SHIFT				0
+static inline uint32_t CP_DRAW_INDX_3_INDX_BASE(uint32_t val)
 {
-	return ((val) << CP_DRAW_INDX_2_INDX_BASE__SHIFT) & CP_DRAW_INDX_2_INDX_BASE__MASK;
+	return ((val) << CP_DRAW_INDX_3_INDX_BASE__SHIFT) & CP_DRAW_INDX_3_INDX_BASE__MASK;
 }
 
-#define REG_CP_DRAW_INDX_2					0x00000002
-#define CP_DRAW_INDX_2_INDX_SIZE__MASK				0xffffffff
-#define CP_DRAW_INDX_2_INDX_SIZE__SHIFT				0
-static inline uint32_t CP_DRAW_INDX_2_INDX_SIZE(uint32_t val)
+#define REG_CP_DRAW_INDX_4					0x00000004
+#define CP_DRAW_INDX_4_INDX_SIZE__MASK				0xffffffff
+#define CP_DRAW_INDX_4_INDX_SIZE__SHIFT				0
+static inline uint32_t CP_DRAW_INDX_4_INDX_SIZE(uint32_t val)
 {
-	return ((val) << CP_DRAW_INDX_2_INDX_SIZE__SHIFT) & CP_DRAW_INDX_2_INDX_SIZE__MASK;
+	return ((val) << CP_DRAW_INDX_4_INDX_SIZE__SHIFT) & CP_DRAW_INDX_4_INDX_SIZE__MASK;
 }
 
 #define REG_CP_DRAW_INDX_2_0					0x00000000
@@ -341,11 +346,11 @@ static inline uint32_t CP_DRAW_INDX_2_1_INDEX_SIZE(enum pc_di_index_size val)
 #define CP_DRAW_INDX_2_1_NOT_EOP				0x00001000
 #define CP_DRAW_INDX_2_1_SMALL_INDEX				0x00002000
 #define CP_DRAW_INDX_2_1_PRE_DRAW_INITIATOR_ENABLE		0x00004000
-#define CP_DRAW_INDX_2_1_NUM_INDICES__MASK			0xffff0000
-#define CP_DRAW_INDX_2_1_NUM_INDICES__SHIFT			16
-static inline uint32_t CP_DRAW_INDX_2_1_NUM_INDICES(uint32_t val)
+#define CP_DRAW_INDX_2_1_NUM_INSTANCES__MASK			0xff000000
+#define CP_DRAW_INDX_2_1_NUM_INSTANCES__SHIFT			24
+static inline uint32_t CP_DRAW_INDX_2_1_NUM_INSTANCES(uint32_t val)
 {
-	return ((val) << CP_DRAW_INDX_2_1_NUM_INDICES__SHIFT) & CP_DRAW_INDX_2_1_NUM_INDICES__MASK;
+	return ((val) << CP_DRAW_INDX_2_1_NUM_INSTANCES__SHIFT) & CP_DRAW_INDX_2_1_NUM_INSTANCES__MASK;
 }
 
 #define REG_CP_DRAW_INDX_2_2					0x00000002
@@ -384,11 +389,11 @@ static inline uint32_t CP_DRAW_INDX_OFFSET_0_INDEX_SIZE(enum pc_di_index_size va
 #define CP_DRAW_INDX_OFFSET_0_NOT_EOP				0x00001000
 #define CP_DRAW_INDX_OFFSET_0_SMALL_INDEX			0x00002000
 #define CP_DRAW_INDX_OFFSET_0_PRE_DRAW_INITIATOR_ENABLE		0x00004000
-#define CP_DRAW_INDX_OFFSET_0_NUM_INDICES__MASK			0xffff0000
-#define CP_DRAW_INDX_OFFSET_0_NUM_INDICES__SHIFT		16
-static inline uint32_t CP_DRAW_INDX_OFFSET_0_NUM_INDICES(uint32_t val)
+#define CP_DRAW_INDX_OFFSET_0_NUM_INSTANCES__MASK		0xffff0000
+#define CP_DRAW_INDX_OFFSET_0_NUM_INSTANCES__SHIFT		16
+static inline uint32_t CP_DRAW_INDX_OFFSET_0_NUM_INSTANCES(uint32_t val)
 {
-	return ((val) << CP_DRAW_INDX_OFFSET_0_NUM_INDICES__SHIFT) & CP_DRAW_INDX_OFFSET_0_NUM_INDICES__MASK;
+	return ((val) << CP_DRAW_INDX_OFFSET_0_NUM_INSTANCES__SHIFT) & CP_DRAW_INDX_OFFSET_0_NUM_INSTANCES__MASK;
 }
 
 #define REG_CP_DRAW_INDX_OFFSET_1				0x00000001
@@ -401,20 +406,22 @@ static inline uint32_t CP_DRAW_INDX_OFFSET_2_NUM_INDICES(uint32_t val)
 	return ((val) << CP_DRAW_INDX_OFFSET_2_NUM_INDICES__SHIFT) & CP_DRAW_INDX_OFFSET_2_NUM_INDICES__MASK;
 }
 
-#define REG_CP_DRAW_INDX_OFFSET_2				0x00000002
-#define CP_DRAW_INDX_OFFSET_2_INDX_BASE__MASK			0xffffffff
-#define CP_DRAW_INDX_OFFSET_2_INDX_BASE__SHIFT			0
-static inline uint32_t CP_DRAW_INDX_OFFSET_2_INDX_BASE(uint32_t val)
+#define REG_CP_DRAW_INDX_OFFSET_3				0x00000003
+
+#define REG_CP_DRAW_INDX_OFFSET_4				0x00000004
+#define CP_DRAW_INDX_OFFSET_4_INDX_BASE__MASK			0xffffffff
+#define CP_DRAW_INDX_OFFSET_4_INDX_BASE__SHIFT			0
+static inline uint32_t CP_DRAW_INDX_OFFSET_4_INDX_BASE(uint32_t val)
 {
-	return ((val) << CP_DRAW_INDX_OFFSET_2_INDX_BASE__SHIFT) & CP_DRAW_INDX_OFFSET_2_INDX_BASE__MASK;
+	return ((val) << CP_DRAW_INDX_OFFSET_4_INDX_BASE__SHIFT) & CP_DRAW_INDX_OFFSET_4_INDX_BASE__MASK;
 }
 
-#define REG_CP_DRAW_INDX_OFFSET_2				0x00000002
-#define CP_DRAW_INDX_OFFSET_2_INDX_SIZE__MASK			0xffffffff
-#define CP_DRAW_INDX_OFFSET_2_INDX_SIZE__SHIFT			0
-static inline uint32_t CP_DRAW_INDX_OFFSET_2_INDX_SIZE(uint32_t val)
+#define REG_CP_DRAW_INDX_OFFSET_5				0x00000005
+#define CP_DRAW_INDX_OFFSET_5_INDX_SIZE__MASK			0xffffffff
+#define CP_DRAW_INDX_OFFSET_5_INDX_SIZE__SHIFT			0
+static inline uint32_t CP_DRAW_INDX_OFFSET_5_INDX_SIZE(uint32_t val)
 {
-	return ((val) << CP_DRAW_INDX_OFFSET_2_INDX_SIZE__SHIFT) & CP_DRAW_INDX_OFFSET_2_INDX_SIZE__MASK;
+	return ((val) << CP_DRAW_INDX_OFFSET_5_INDX_SIZE__SHIFT) & CP_DRAW_INDX_OFFSET_5_INDX_SIZE__MASK;
 }
 
 #define REG_CP_SET_DRAW_STATE_0					0x00000000
diff --git a/drivers/gpu/drm/msm/dsi/dsi.xml.h b/drivers/gpu/drm/msm/dsi/dsi.xml.h
index 0f1f5b9459a5..448438b759b4 100644
--- a/drivers/gpu/drm/msm/dsi/dsi.xml.h
+++ b/drivers/gpu/drm/msm/dsi/dsi.xml.h
@@ -10,14 +10,14 @@ git clone https://github.com/freedreno/envytools.git
 The rules-ng-ng source files this header was generated from are:
 - /home/robclark/src/freedreno/envytools/rnndb/msm.xml                 (    647 bytes, from 2013-11-30 14:45:35)
 - /home/robclark/src/freedreno/envytools/rnndb/freedreno_copyright.xml (   1453 bytes, from 2013-03-31 16:51:27)
-- /home/robclark/src/freedreno/envytools/rnndb/mdp/mdp4.xml            (  17996 bytes, from 2013-12-01 19:10:31)
-- /home/robclark/src/freedreno/envytools/rnndb/mdp/mdp_common.xml      (   1615 bytes, from 2013-11-30 15:00:52)
-- /home/robclark/src/freedreno/envytools/rnndb/mdp/mdp5.xml            (  22517 bytes, from 2014-06-25 12:55:02)
+- /home/robclark/src/freedreno/envytools/rnndb/mdp/mdp4.xml            (  20136 bytes, from 2014-10-31 16:51:39)
+- /home/robclark/src/freedreno/envytools/rnndb/mdp/mdp_common.xml      (   1940 bytes, from 2014-10-31 16:51:39)
+- /home/robclark/src/freedreno/envytools/rnndb/mdp/mdp5.xml            (  23963 bytes, from 2014-10-31 16:51:46)
 - /home/robclark/src/freedreno/envytools/rnndb/dsi/dsi.xml             (  11712 bytes, from 2013-08-17 17:13:43)
 - /home/robclark/src/freedreno/envytools/rnndb/dsi/sfpb.xml            (    344 bytes, from 2013-08-11 19:26:32)
-- /home/robclark/src/freedreno/envytools/rnndb/dsi/mmss_cc.xml         (   1544 bytes, from 2013-08-16 19:17:05)
+- /home/robclark/src/freedreno/envytools/rnndb/dsi/mmss_cc.xml         (   1686 bytes, from 2014-10-31 16:48:57)
 - /home/robclark/src/freedreno/envytools/rnndb/hdmi/qfprom.xml         (    600 bytes, from 2013-07-05 19:21:12)
-- /home/robclark/src/freedreno/envytools/rnndb/hdmi/hdmi.xml           (  23613 bytes, from 2014-06-25 12:53:44)
+- /home/robclark/src/freedreno/envytools/rnndb/hdmi/hdmi.xml           (  23613 bytes, from 2014-07-17 15:33:30)
 
 Copyright (C) 2013 by the following authors:
 - Rob Clark <robdclark@gmail.com> (robclark)
diff --git a/drivers/gpu/drm/msm/dsi/mmss_cc.xml.h b/drivers/gpu/drm/msm/dsi/mmss_cc.xml.h
index d468f86f637c..c102a7f074ac 100644
--- a/drivers/gpu/drm/msm/dsi/mmss_cc.xml.h
+++ b/drivers/gpu/drm/msm/dsi/mmss_cc.xml.h
@@ -10,16 +10,16 @@ git clone https://github.com/freedreno/envytools.git
 The rules-ng-ng source files this header was generated from are:
 - /home/robclark/src/freedreno/envytools/rnndb/msm.xml                 (    647 bytes, from 2013-11-30 14:45:35)
 - /home/robclark/src/freedreno/envytools/rnndb/freedreno_copyright.xml (   1453 bytes, from 2013-03-31 16:51:27)
-- /home/robclark/src/freedreno/envytools/rnndb/mdp/mdp4.xml            (  17996 bytes, from 2013-12-01 19:10:31)
-- /home/robclark/src/freedreno/envytools/rnndb/mdp/mdp_common.xml      (   1615 bytes, from 2013-11-30 15:00:52)
-- /home/robclark/src/freedreno/envytools/rnndb/mdp/mdp5.xml            (  22517 bytes, from 2014-06-25 12:55:02)
+- /home/robclark/src/freedreno/envytools/rnndb/mdp/mdp4.xml            (  20136 bytes, from 2014-10-31 16:51:39)
+- /home/robclark/src/freedreno/envytools/rnndb/mdp/mdp_common.xml      (   1940 bytes, from 2014-10-31 16:51:39)
+- /home/robclark/src/freedreno/envytools/rnndb/mdp/mdp5.xml            (  23963 bytes, from 2014-10-31 16:51:46)
 - /home/robclark/src/freedreno/envytools/rnndb/dsi/dsi.xml             (  11712 bytes, from 2013-08-17 17:13:43)
 - /home/robclark/src/freedreno/envytools/rnndb/dsi/sfpb.xml            (    344 bytes, from 2013-08-11 19:26:32)
-- /home/robclark/src/freedreno/envytools/rnndb/dsi/mmss_cc.xml         (   1544 bytes, from 2013-08-16 19:17:05)
+- /home/robclark/src/freedreno/envytools/rnndb/dsi/mmss_cc.xml         (   1686 bytes, from 2014-10-31 16:48:57)
 - /home/robclark/src/freedreno/envytools/rnndb/hdmi/qfprom.xml         (    600 bytes, from 2013-07-05 19:21:12)
-- /home/robclark/src/freedreno/envytools/rnndb/hdmi/hdmi.xml           (  23613 bytes, from 2014-06-25 12:53:44)
+- /home/robclark/src/freedreno/envytools/rnndb/hdmi/hdmi.xml           (  23613 bytes, from 2014-07-17 15:33:30)
 
-Copyright (C) 2013 by the following authors:
+Copyright (C) 2013-2014 by the following authors:
 - Rob Clark <robdclark@gmail.com> (robclark)
 
 Permission is hereby granted, free of charge, to any person obtaining
@@ -112,5 +112,11 @@ static inline uint32_t MMSS_CC_CLK_NS_VAL(uint32_t val)
 	return ((val) << MMSS_CC_CLK_NS_VAL__SHIFT) & MMSS_CC_CLK_NS_VAL__MASK;
 }
 
+#define REG_MMSS_CC_DSI2_PIXEL_CC				0x00000094
+
+#define REG_MMSS_CC_DSI2_PIXEL_NS				0x000000e4
+
+#define REG_MMSS_CC_DSI2_PIXEL_CC2				0x00000264
+
 
 #endif /* MMSS_CC_XML */
diff --git a/drivers/gpu/drm/msm/dsi/sfpb.xml.h b/drivers/gpu/drm/msm/dsi/sfpb.xml.h
index da8740054cdf..a900134bdf33 100644
--- a/drivers/gpu/drm/msm/dsi/sfpb.xml.h
+++ b/drivers/gpu/drm/msm/dsi/sfpb.xml.h
@@ -10,14 +10,14 @@ git clone https://github.com/freedreno/envytools.git
 The rules-ng-ng source files this header was generated from are:
 - /home/robclark/src/freedreno/envytools/rnndb/msm.xml                 (    647 bytes, from 2013-11-30 14:45:35)
 - /home/robclark/src/freedreno/envytools/rnndb/freedreno_copyright.xml (   1453 bytes, from 2013-03-31 16:51:27)
-- /home/robclark/src/freedreno/envytools/rnndb/mdp/mdp4.xml            (  17996 bytes, from 2013-12-01 19:10:31)
-- /home/robclark/src/freedreno/envytools/rnndb/mdp/mdp_common.xml      (   1615 bytes, from 2013-11-30 15:00:52)
-- /home/robclark/src/freedreno/envytools/rnndb/mdp/mdp5.xml            (  22517 bytes, from 2014-06-25 12:55:02)
+- /home/robclark/src/freedreno/envytools/rnndb/mdp/mdp4.xml            (  20136 bytes, from 2014-10-31 16:51:39)
+- /home/robclark/src/freedreno/envytools/rnndb/mdp/mdp_common.xml      (   1940 bytes, from 2014-10-31 16:51:39)
+- /home/robclark/src/freedreno/envytools/rnndb/mdp/mdp5.xml            (  23963 bytes, from 2014-10-31 16:51:46)
 - /home/robclark/src/freedreno/envytools/rnndb/dsi/dsi.xml             (  11712 bytes, from 2013-08-17 17:13:43)
 - /home/robclark/src/freedreno/envytools/rnndb/dsi/sfpb.xml            (    344 bytes, from 2013-08-11 19:26:32)
-- /home/robclark/src/freedreno/envytools/rnndb/dsi/mmss_cc.xml         (   1544 bytes, from 2013-08-16 19:17:05)
+- /home/robclark/src/freedreno/envytools/rnndb/dsi/mmss_cc.xml         (   1686 bytes, from 2014-10-31 16:48:57)
 - /home/robclark/src/freedreno/envytools/rnndb/hdmi/qfprom.xml         (    600 bytes, from 2013-07-05 19:21:12)
-- /home/robclark/src/freedreno/envytools/rnndb/hdmi/hdmi.xml           (  23613 bytes, from 2014-06-25 12:53:44)
+- /home/robclark/src/freedreno/envytools/rnndb/hdmi/hdmi.xml           (  23613 bytes, from 2014-07-17 15:33:30)
 
 Copyright (C) 2013 by the following authors:
 - Rob Clark <robdclark@gmail.com> (robclark)
diff --git a/drivers/gpu/drm/msm/hdmi/hdmi.c b/drivers/gpu/drm/msm/hdmi/hdmi.c
index c6c9b02e0ada..062c68725376 100644
--- a/drivers/gpu/drm/msm/hdmi/hdmi.c
+++ b/drivers/gpu/drm/msm/hdmi/hdmi.c
@@ -15,6 +15,7 @@
  * this program.  If not, see <http://www.gnu.org/licenses/>.
  */
 
+#include <linux/of_irq.h>
 #include "hdmi.h"
 
 void hdmi_set_mode(struct hdmi *hdmi, bool power_on)
@@ -39,7 +40,7 @@ void hdmi_set_mode(struct hdmi *hdmi, bool power_on)
 			power_on ? "Enable" : "Disable", ctrl);
 }
 
-irqreturn_t hdmi_irq(int irq, void *dev_id)
+static irqreturn_t hdmi_irq(int irq, void *dev_id)
 {
 	struct hdmi *hdmi = dev_id;
 
@@ -54,9 +55,8 @@ irqreturn_t hdmi_irq(int irq, void *dev_id)
 	return IRQ_HANDLED;
 }
 
-void hdmi_destroy(struct kref *kref)
+static void hdmi_destroy(struct hdmi *hdmi)
 {
-	struct hdmi *hdmi = container_of(kref, struct hdmi, refcount);
 	struct hdmi_phy *phy = hdmi->phy;
 
 	if (phy)
@@ -68,37 +68,24 @@ void hdmi_destroy(struct kref *kref)
 	platform_set_drvdata(hdmi->pdev, NULL);
 }
 
-/* initialize connector */
-struct hdmi *hdmi_init(struct drm_device *dev, struct drm_encoder *encoder)
+/* construct hdmi at bind/probe time, grab all the resources.  If
+ * we are to EPROBE_DEFER we want to do it here, rather than later
+ * at modeset_init() time
+ */
+static struct hdmi *hdmi_init(struct platform_device *pdev)
 {
+	struct hdmi_platform_config *config = pdev->dev.platform_data;
 	struct hdmi *hdmi = NULL;
-	struct msm_drm_private *priv = dev->dev_private;
-	struct platform_device *pdev = priv->hdmi_pdev;
-	struct hdmi_platform_config *config;
 	int i, ret;
 
-	if (!pdev) {
-		dev_err(dev->dev, "no hdmi device\n");
-		ret = -ENXIO;
-		goto fail;
-	}
-
-	config = pdev->dev.platform_data;
-
-	hdmi = kzalloc(sizeof(*hdmi), GFP_KERNEL);
+	hdmi = devm_kzalloc(&pdev->dev, sizeof(*hdmi), GFP_KERNEL);
 	if (!hdmi) {
 		ret = -ENOMEM;
 		goto fail;
 	}
 
-	kref_init(&hdmi->refcount);
-
-	hdmi->dev = dev;
 	hdmi->pdev = pdev;
 	hdmi->config = config;
-	hdmi->encoder = encoder;
-
-	hdmi_audio_infoframe_init(&hdmi->audio.infoframe);
 
 	/* not sure about which phy maps to which msm.. probably I miss some */
 	if (config->phy_init)
@@ -108,7 +95,7 @@ struct hdmi *hdmi_init(struct drm_device *dev, struct drm_encoder *encoder)
 
 	if (IS_ERR(hdmi->phy)) {
 		ret = PTR_ERR(hdmi->phy);
-		dev_err(dev->dev, "failed to load phy: %d\n", ret);
+		dev_err(&pdev->dev, "failed to load phy: %d\n", ret);
 		hdmi->phy = NULL;
 		goto fail;
 	}
@@ -123,11 +110,11 @@ struct hdmi *hdmi_init(struct drm_device *dev, struct drm_encoder *encoder)
 	for (i = 0; i < config->hpd_reg_cnt; i++) {
 		struct regulator *reg;
 
-		reg = devm_regulator_get_exclusive(&pdev->dev,
+		reg = devm_regulator_get(&pdev->dev,
 				config->hpd_reg_names[i]);
 		if (IS_ERR(reg)) {
 			ret = PTR_ERR(reg);
-			dev_err(dev->dev, "failed to get hpd regulator: %s (%d)\n",
+			dev_err(&pdev->dev, "failed to get hpd regulator: %s (%d)\n",
 					config->hpd_reg_names[i], ret);
 			goto fail;
 		}
@@ -139,11 +126,11 @@ struct hdmi *hdmi_init(struct drm_device *dev, struct drm_encoder *encoder)
 	for (i = 0; i < config->pwr_reg_cnt; i++) {
 		struct regulator *reg;
 
-		reg = devm_regulator_get_exclusive(&pdev->dev,
+		reg = devm_regulator_get(&pdev->dev,
 				config->pwr_reg_names[i]);
 		if (IS_ERR(reg)) {
 			ret = PTR_ERR(reg);
-			dev_err(dev->dev, "failed to get pwr regulator: %s (%d)\n",
+			dev_err(&pdev->dev, "failed to get pwr regulator: %s (%d)\n",
 					config->pwr_reg_names[i], ret);
 			goto fail;
 		}
@@ -158,7 +145,7 @@ struct hdmi *hdmi_init(struct drm_device *dev, struct drm_encoder *encoder)
 		clk = devm_clk_get(&pdev->dev, config->hpd_clk_names[i]);
 		if (IS_ERR(clk)) {
 			ret = PTR_ERR(clk);
-			dev_err(dev->dev, "failed to get hpd clk: %s (%d)\n",
+			dev_err(&pdev->dev, "failed to get hpd clk: %s (%d)\n",
 					config->hpd_clk_names[i], ret);
 			goto fail;
 		}
@@ -173,7 +160,7 @@ struct hdmi *hdmi_init(struct drm_device *dev, struct drm_encoder *encoder)
 		clk = devm_clk_get(&pdev->dev, config->pwr_clk_names[i]);
 		if (IS_ERR(clk)) {
 			ret = PTR_ERR(clk);
-			dev_err(dev->dev, "failed to get pwr clk: %s (%d)\n",
+			dev_err(&pdev->dev, "failed to get pwr clk: %s (%d)\n",
 					config->pwr_clk_names[i], ret);
 			goto fail;
 		}
@@ -184,11 +171,40 @@ struct hdmi *hdmi_init(struct drm_device *dev, struct drm_encoder *encoder)
 	hdmi->i2c = hdmi_i2c_init(hdmi);
 	if (IS_ERR(hdmi->i2c)) {
 		ret = PTR_ERR(hdmi->i2c);
-		dev_err(dev->dev, "failed to get i2c: %d\n", ret);
+		dev_err(&pdev->dev, "failed to get i2c: %d\n", ret);
 		hdmi->i2c = NULL;
 		goto fail;
 	}
 
+	return hdmi;
+
+fail:
+	if (hdmi)
+		hdmi_destroy(hdmi);
+
+	return ERR_PTR(ret);
+}
+
+/* Second part of initialization, the drm/kms level modeset_init,
+ * constructs/initializes mode objects, etc, is called from master
+ * driver (not hdmi sub-device's probe/bind!)
+ *
+ * Any resource (regulator/clk/etc) which could be missing at boot
+ * should be handled in hdmi_init() so that failure happens from
+ * hdmi sub-device's probe.
+ */
+int hdmi_modeset_init(struct hdmi *hdmi,
+		struct drm_device *dev, struct drm_encoder *encoder)
+{
+	struct msm_drm_private *priv = dev->dev_private;
+	struct platform_device *pdev = hdmi->pdev;
+	int ret;
+
+	hdmi->dev = dev;
+	hdmi->encoder = encoder;
+
+	hdmi_audio_infoframe_init(&hdmi->audio.infoframe);
+
 	hdmi->bridge = hdmi_bridge_init(hdmi);
 	if (IS_ERR(hdmi->bridge)) {
 		ret = PTR_ERR(hdmi->bridge);
@@ -205,22 +221,20 @@ struct hdmi *hdmi_init(struct drm_device *dev, struct drm_encoder *encoder)
 		goto fail;
 	}
 
-	if (!config->shared_irq) {
-		hdmi->irq = platform_get_irq(pdev, 0);
-		if (hdmi->irq < 0) {
-			ret = hdmi->irq;
-			dev_err(dev->dev, "failed to get irq: %d\n", ret);
-			goto fail;
-		}
+	hdmi->irq = irq_of_parse_and_map(pdev->dev.of_node, 0);
+	if (hdmi->irq < 0) {
+		ret = hdmi->irq;
+		dev_err(dev->dev, "failed to get irq: %d\n", ret);
+		goto fail;
+	}
 
-		ret = devm_request_threaded_irq(&pdev->dev, hdmi->irq,
-				NULL, hdmi_irq, IRQF_TRIGGER_HIGH | IRQF_ONESHOT,
-				"hdmi_isr", hdmi);
-		if (ret < 0) {
-			dev_err(dev->dev, "failed to request IRQ%u: %d\n",
-					hdmi->irq, ret);
-			goto fail;
-		}
+	ret = devm_request_irq(&pdev->dev, hdmi->irq,
+			hdmi_irq, IRQF_TRIGGER_HIGH | IRQF_ONESHOT,
+			"hdmi_isr", hdmi);
+	if (ret < 0) {
+		dev_err(dev->dev, "failed to request IRQ%u: %d\n",
+				hdmi->irq, ret);
+		goto fail;
 	}
 
 	encoder->bridge = hdmi->bridge;
@@ -230,19 +244,20 @@ struct hdmi *hdmi_init(struct drm_device *dev, struct drm_encoder *encoder)
 
 	platform_set_drvdata(pdev, hdmi);
 
-	return hdmi;
+	return 0;
 
 fail:
-	if (hdmi) {
-		/* bridge/connector are normally destroyed by drm: */
-		if (hdmi->bridge)
-			hdmi->bridge->funcs->destroy(hdmi->bridge);
-		if (hdmi->connector)
-			hdmi->connector->funcs->destroy(hdmi->connector);
-		hdmi_destroy(&hdmi->refcount);
+	/* bridge/connector are normally destroyed by drm: */
+	if (hdmi->bridge) {
+		hdmi->bridge->funcs->destroy(hdmi->bridge);
+		hdmi->bridge = NULL;
+	}
+	if (hdmi->connector) {
+		hdmi->connector->funcs->destroy(hdmi->connector);
+		hdmi->connector = NULL;
 	}
 
-	return ERR_PTR(ret);
+	return ret;
 }
 
 /*
@@ -251,13 +266,6 @@ fail:
 
 #include <linux/of_gpio.h>
 
-static void set_hdmi_pdev(struct drm_device *dev,
-		struct platform_device *pdev)
-{
-	struct msm_drm_private *priv = dev->dev_private;
-	priv->hdmi_pdev = pdev;
-}
-
 #ifdef CONFIG_OF
 static int get_gpio(struct device *dev, struct device_node *of_node, const char *name)
 {
@@ -278,7 +286,10 @@ static int get_gpio(struct device *dev, struct device_node *of_node, const char
 
 static int hdmi_bind(struct device *dev, struct device *master, void *data)
 {
+	struct drm_device *drm = dev_get_drvdata(master);
+	struct msm_drm_private *priv = drm->dev_private;
 	static struct hdmi_platform_config config = {};
+	struct hdmi *hdmi;
 #ifdef CONFIG_OF
 	struct device_node *of_node = dev->of_node;
 
@@ -298,7 +309,6 @@ static int hdmi_bind(struct device *dev, struct device *master, void *data)
 		config.hpd_clk_cnt   = ARRAY_SIZE(hpd_clk_names);
 		config.pwr_clk_names = pwr_clk_names;
 		config.pwr_clk_cnt   = ARRAY_SIZE(pwr_clk_names);
-		config.shared_irq    = true;
 	} else if (of_device_is_compatible(of_node, "qcom,hdmi-tx-8960")) {
 		static const char *hpd_clk_names[] = {"core_clk", "master_iface_clk", "slave_iface_clk"};
 		static const char *hpd_reg_names[] = {"core-vdda", "hdmi-mux"};
@@ -369,14 +379,22 @@ static int hdmi_bind(struct device *dev, struct device *master, void *data)
 	}
 #endif
 	dev->platform_data = &config;
-	set_hdmi_pdev(dev_get_drvdata(master), to_platform_device(dev));
+	hdmi = hdmi_init(to_platform_device(dev));
+	if (IS_ERR(hdmi))
+		return PTR_ERR(hdmi);
+	priv->hdmi = hdmi;
 	return 0;
 }
 
 static void hdmi_unbind(struct device *dev, struct device *master,
 		void *data)
 {
-	set_hdmi_pdev(dev_get_drvdata(master), NULL);
+	struct drm_device *drm = dev_get_drvdata(master);
+	struct msm_drm_private *priv = drm->dev_private;
+	if (priv->hdmi) {
+		hdmi_destroy(priv->hdmi);
+		priv->hdmi = NULL;
+	}
 }
 
 static const struct component_ops hdmi_ops = {
diff --git a/drivers/gpu/drm/msm/hdmi/hdmi.h b/drivers/gpu/drm/msm/hdmi/hdmi.h
index b981995410b5..43e654f751b7 100644
--- a/drivers/gpu/drm/msm/hdmi/hdmi.h
+++ b/drivers/gpu/drm/msm/hdmi/hdmi.h
@@ -38,8 +38,6 @@ struct hdmi_audio {
 };
 
 struct hdmi {
-	struct kref refcount;
-
 	struct drm_device *dev;
 	struct platform_device *pdev;
 
@@ -97,13 +95,9 @@ struct hdmi_platform_config {
 	/* gpio's: */
 	int ddc_clk_gpio, ddc_data_gpio, hpd_gpio, mux_en_gpio, mux_sel_gpio;
 	int mux_lpm_gpio;
-
-	/* older devices had their own irq, mdp5+ it is shared w/ mdp: */
-	bool shared_irq;
 };
 
 void hdmi_set_mode(struct hdmi *hdmi, bool power_on);
-void hdmi_destroy(struct kref *kref);
 
 static inline void hdmi_write(struct hdmi *hdmi, u32 reg, u32 data)
 {
@@ -115,17 +109,6 @@ static inline u32 hdmi_read(struct hdmi *hdmi, u32 reg)
 	return msm_readl(hdmi->mmio + reg);
 }
 
-static inline struct hdmi * hdmi_reference(struct hdmi *hdmi)
-{
-	kref_get(&hdmi->refcount);
-	return hdmi;
-}
-
-static inline void hdmi_unreference(struct hdmi *hdmi)
-{
-	kref_put(&hdmi->refcount, hdmi_destroy);
-}
-
 /*
  * The phy appears to be different, for example between 8960 and 8x60,
  * so split the phy related functions out and load the correct one at
diff --git a/drivers/gpu/drm/msm/hdmi/hdmi.xml.h b/drivers/gpu/drm/msm/hdmi/hdmi.xml.h
index e89fe053d375..5b0844befbab 100644
--- a/drivers/gpu/drm/msm/hdmi/hdmi.xml.h
+++ b/drivers/gpu/drm/msm/hdmi/hdmi.xml.h
@@ -10,14 +10,14 @@ git clone https://github.com/freedreno/envytools.git
 The rules-ng-ng source files this header was generated from are:
 - /home/robclark/src/freedreno/envytools/rnndb/msm.xml                 (    647 bytes, from 2013-11-30 14:45:35)
 - /home/robclark/src/freedreno/envytools/rnndb/freedreno_copyright.xml (   1453 bytes, from 2013-03-31 16:51:27)
-- /home/robclark/src/freedreno/envytools/rnndb/mdp/mdp4.xml            (  17996 bytes, from 2013-12-01 19:10:31)
-- /home/robclark/src/freedreno/envytools/rnndb/mdp/mdp_common.xml      (   1615 bytes, from 2013-11-30 15:00:52)
-- /home/robclark/src/freedreno/envytools/rnndb/mdp/mdp5.xml            (  22517 bytes, from 2014-06-25 12:55:02)
+- /home/robclark/src/freedreno/envytools/rnndb/mdp/mdp4.xml            (  20136 bytes, from 2014-10-31 16:51:39)
+- /home/robclark/src/freedreno/envytools/rnndb/mdp/mdp_common.xml      (   1940 bytes, from 2014-10-31 16:51:39)
+- /home/robclark/src/freedreno/envytools/rnndb/mdp/mdp5.xml            (  23963 bytes, from 2014-10-31 16:51:46)
 - /home/robclark/src/freedreno/envytools/rnndb/dsi/dsi.xml             (  11712 bytes, from 2013-08-17 17:13:43)
 - /home/robclark/src/freedreno/envytools/rnndb/dsi/sfpb.xml            (    344 bytes, from 2013-08-11 19:26:32)
-- /home/robclark/src/freedreno/envytools/rnndb/dsi/mmss_cc.xml         (   1544 bytes, from 2013-08-16 19:17:05)
+- /home/robclark/src/freedreno/envytools/rnndb/dsi/mmss_cc.xml         (   1686 bytes, from 2014-10-31 16:48:57)
 - /home/robclark/src/freedreno/envytools/rnndb/hdmi/qfprom.xml         (    600 bytes, from 2013-07-05 19:21:12)
-- /home/robclark/src/freedreno/envytools/rnndb/hdmi/hdmi.xml           (  23613 bytes, from 2014-06-25 12:53:44)
+- /home/robclark/src/freedreno/envytools/rnndb/hdmi/hdmi.xml           (  23613 bytes, from 2014-07-17 15:33:30)
 
 Copyright (C) 2013-2014 by the following authors:
 - Rob Clark <robdclark@gmail.com> (robclark)
diff --git a/drivers/gpu/drm/msm/hdmi/hdmi_bridge.c b/drivers/gpu/drm/msm/hdmi/hdmi_bridge.c
index f6cf745c249e..6902ad6da710 100644
--- a/drivers/gpu/drm/msm/hdmi/hdmi_bridge.c
+++ b/drivers/gpu/drm/msm/hdmi/hdmi_bridge.c
@@ -26,7 +26,6 @@ struct hdmi_bridge {
 static void hdmi_bridge_destroy(struct drm_bridge *bridge)
 {
 	struct hdmi_bridge *hdmi_bridge = to_hdmi_bridge(bridge);
-	hdmi_unreference(hdmi_bridge->hdmi);
 	drm_bridge_cleanup(bridge);
 	kfree(hdmi_bridge);
 }
@@ -218,7 +217,7 @@ struct drm_bridge *hdmi_bridge_init(struct hdmi *hdmi)
 		goto fail;
 	}
 
-	hdmi_bridge->hdmi = hdmi_reference(hdmi);
+	hdmi_bridge->hdmi = hdmi;
 
 	bridge = &hdmi_bridge->base;
 
diff --git a/drivers/gpu/drm/msm/hdmi/hdmi_connector.c b/drivers/gpu/drm/msm/hdmi/hdmi_connector.c
index 4aca2a3c667c..b4e70e0e3cfa 100644
--- a/drivers/gpu/drm/msm/hdmi/hdmi_connector.c
+++ b/drivers/gpu/drm/msm/hdmi/hdmi_connector.c
@@ -141,6 +141,15 @@ static int hpd_enable(struct hdmi_connector *hdmi_connector)
 	uint32_t hpd_ctrl;
 	int i, ret;
 
+	for (i = 0; i < config->hpd_reg_cnt; i++) {
+		ret = regulator_enable(hdmi->hpd_regs[i]);
+		if (ret) {
+			dev_err(dev->dev, "failed to enable hpd regulator: %s (%d)\n",
+					config->hpd_reg_names[i], ret);
+			goto fail;
+		}
+	}
+
 	ret = gpio_config(hdmi, true);
 	if (ret) {
 		dev_err(dev->dev, "failed to configure GPIOs: %d\n", ret);
@@ -164,15 +173,6 @@ static int hpd_enable(struct hdmi_connector *hdmi_connector)
 		}
 	}
 
-	for (i = 0; i < config->hpd_reg_cnt; i++) {
-		ret = regulator_enable(hdmi->hpd_regs[i]);
-		if (ret) {
-			dev_err(dev->dev, "failed to enable hpd regulator: %s (%d)\n",
-					config->hpd_reg_names[i], ret);
-			goto fail;
-		}
-	}
-
 	hdmi_set_mode(hdmi, false);
 	phy->funcs->reset(phy);
 	hdmi_set_mode(hdmi, true);
@@ -200,7 +200,7 @@ fail:
 	return ret;
 }
 
-static int hdp_disable(struct hdmi_connector *hdmi_connector)
+static void hdp_disable(struct hdmi_connector *hdmi_connector)
 {
 	struct hdmi *hdmi = hdmi_connector->hdmi;
 	const struct hdmi_platform_config *config = hdmi->config;
@@ -212,28 +212,19 @@ static int hdp_disable(struct hdmi_connector *hdmi_connector)
 
 	hdmi_set_mode(hdmi, false);
 
-	for (i = 0; i < config->hpd_reg_cnt; i++) {
-		ret = regulator_disable(hdmi->hpd_regs[i]);
-		if (ret) {
-			dev_err(dev->dev, "failed to disable hpd regulator: %s (%d)\n",
-					config->hpd_reg_names[i], ret);
-			goto fail;
-		}
-	}
-
 	for (i = 0; i < config->hpd_clk_cnt; i++)
 		clk_disable_unprepare(hdmi->hpd_clks[i]);
 
 	ret = gpio_config(hdmi, false);
-	if (ret) {
-		dev_err(dev->dev, "failed to unconfigure GPIOs: %d\n", ret);
-		goto fail;
-	}
+	if (ret)
+		dev_warn(dev->dev, "failed to unconfigure GPIOs: %d\n", ret);
 
-	return 0;
-
-fail:
-	return ret;
+	for (i = 0; i < config->hpd_reg_cnt; i++) {
+		ret = regulator_disable(hdmi->hpd_regs[i]);
+		if (ret)
+			dev_warn(dev->dev, "failed to disable hpd regulator: %s (%d)\n",
+					config->hpd_reg_names[i], ret);
+	}
 }
 
 static void
@@ -260,11 +251,11 @@ void hdmi_connector_irq(struct drm_connector *connector)
 			(hpd_int_status & HDMI_HPD_INT_STATUS_INT)) {
 		bool detected = !!(hpd_int_status & HDMI_HPD_INT_STATUS_CABLE_DETECTED);
 
-		DBG("status=%04x, ctrl=%04x", hpd_int_status, hpd_int_ctrl);
-
-		/* ack the irq: */
+		/* ack & disable (temporarily) HPD events: */
 		hdmi_write(hdmi, REG_HDMI_HPD_INT_CTRL,
-				hpd_int_ctrl | HDMI_HPD_INT_CTRL_INT_ACK);
+			HDMI_HPD_INT_CTRL_INT_ACK);
+
+		DBG("status=%04x, ctrl=%04x", hpd_int_status, hpd_int_ctrl);
 
 		/* detect disconnect if we are connected or visa versa: */
 		hpd_int_ctrl = HDMI_HPD_INT_CTRL_INT_EN;
@@ -330,8 +321,6 @@ static void hdmi_connector_destroy(struct drm_connector *connector)
 	drm_connector_unregister(connector);
 	drm_connector_cleanup(connector);
 
-	hdmi_unreference(hdmi_connector->hdmi);
-
 	kfree(hdmi_connector);
 }
 
@@ -401,6 +390,9 @@ static const struct drm_connector_funcs hdmi_connector_funcs = {
 	.detect = hdmi_connector_detect,
 	.fill_modes = drm_helper_probe_single_connector_modes,
 	.destroy = hdmi_connector_destroy,
+	.reset = drm_atomic_helper_connector_reset,
+	.atomic_duplicate_state = drm_atomic_helper_connector_duplicate_state,
+	.atomic_destroy_state = drm_atomic_helper_connector_destroy_state,
 };
 
 static const struct drm_connector_helper_funcs hdmi_connector_helper_funcs = {
@@ -422,7 +414,7 @@ struct drm_connector *hdmi_connector_init(struct hdmi *hdmi)
 		goto fail;
 	}
 
-	hdmi_connector->hdmi = hdmi_reference(hdmi);
+	hdmi_connector->hdmi = hdmi;
 	INIT_WORK(&hdmi_connector->hpd_work, hotplug_work);
 
 	connector = &hdmi_connector->base;
diff --git a/drivers/gpu/drm/msm/hdmi/hdmi_phy_8960.c b/drivers/gpu/drm/msm/hdmi/hdmi_phy_8960.c
index f408b69486a8..eeed006eed13 100644
--- a/drivers/gpu/drm/msm/hdmi/hdmi_phy_8960.c
+++ b/drivers/gpu/drm/msm/hdmi/hdmi_phy_8960.c
@@ -510,7 +510,7 @@ struct hdmi_phy *hdmi_phy_8960_init(struct hdmi *hdmi)
 
 #ifdef CONFIG_COMMON_CLK
 	phy_8960->pll_hw.init = &pll_init;
-	phy_8960->pll = devm_clk_register(hdmi->dev->dev, &phy_8960->pll_hw);
+	phy_8960->pll = devm_clk_register(&hdmi->pdev->dev, &phy_8960->pll_hw);
 	if (IS_ERR(phy_8960->pll)) {
 		ret = PTR_ERR(phy_8960->pll);
 		phy_8960->pll = NULL;
diff --git a/drivers/gpu/drm/msm/hdmi/qfprom.xml.h b/drivers/gpu/drm/msm/hdmi/qfprom.xml.h
index bd81db6a7829..29bd796797de 100644
--- a/drivers/gpu/drm/msm/hdmi/qfprom.xml.h
+++ b/drivers/gpu/drm/msm/hdmi/qfprom.xml.h
@@ -10,14 +10,14 @@ git clone https://github.com/freedreno/envytools.git
 The rules-ng-ng source files this header was generated from are:
 - /home/robclark/src/freedreno/envytools/rnndb/msm.xml                 (    647 bytes, from 2013-11-30 14:45:35)
 - /home/robclark/src/freedreno/envytools/rnndb/freedreno_copyright.xml (   1453 bytes, from 2013-03-31 16:51:27)
-- /home/robclark/src/freedreno/envytools/rnndb/mdp/mdp4.xml            (  17996 bytes, from 2013-12-01 19:10:31)
-- /home/robclark/src/freedreno/envytools/rnndb/mdp/mdp_common.xml      (   1615 bytes, from 2013-11-30 15:00:52)
-- /home/robclark/src/freedreno/envytools/rnndb/mdp/mdp5.xml            (  22517 bytes, from 2014-06-25 12:55:02)
+- /home/robclark/src/freedreno/envytools/rnndb/mdp/mdp4.xml            (  20136 bytes, from 2014-10-31 16:51:39)
+- /home/robclark/src/freedreno/envytools/rnndb/mdp/mdp_common.xml      (   1940 bytes, from 2014-10-31 16:51:39)
+- /home/robclark/src/freedreno/envytools/rnndb/mdp/mdp5.xml            (  23963 bytes, from 2014-10-31 16:51:46)
 - /home/robclark/src/freedreno/envytools/rnndb/dsi/dsi.xml             (  11712 bytes, from 2013-08-17 17:13:43)
 - /home/robclark/src/freedreno/envytools/rnndb/dsi/sfpb.xml            (    344 bytes, from 2013-08-11 19:26:32)
-- /home/robclark/src/freedreno/envytools/rnndb/dsi/mmss_cc.xml         (   1544 bytes, from 2013-08-16 19:17:05)
+- /home/robclark/src/freedreno/envytools/rnndb/dsi/mmss_cc.xml         (   1686 bytes, from 2014-10-31 16:48:57)
 - /home/robclark/src/freedreno/envytools/rnndb/hdmi/qfprom.xml         (    600 bytes, from 2013-07-05 19:21:12)
-- /home/robclark/src/freedreno/envytools/rnndb/hdmi/hdmi.xml           (  23613 bytes, from 2014-06-25 12:53:44)
+- /home/robclark/src/freedreno/envytools/rnndb/hdmi/hdmi.xml           (  23613 bytes, from 2014-07-17 15:33:30)
 
 Copyright (C) 2013 by the following authors:
 - Rob Clark <robdclark@gmail.com> (robclark)
diff --git a/drivers/gpu/drm/msm/mdp/mdp4/mdp4.xml.h b/drivers/gpu/drm/msm/mdp/mdp4/mdp4.xml.h
index 122208e8a2ee..a4a7f8c7122a 100644
--- a/drivers/gpu/drm/msm/mdp/mdp4/mdp4.xml.h
+++ b/drivers/gpu/drm/msm/mdp/mdp4/mdp4.xml.h
@@ -10,16 +10,16 @@ git clone https://github.com/freedreno/envytools.git
 The rules-ng-ng source files this header was generated from are:
 - /home/robclark/src/freedreno/envytools/rnndb/msm.xml                 (    647 bytes, from 2013-11-30 14:45:35)
 - /home/robclark/src/freedreno/envytools/rnndb/freedreno_copyright.xml (   1453 bytes, from 2013-03-31 16:51:27)
-- /home/robclark/src/freedreno/envytools/rnndb/mdp/mdp4.xml            (  17996 bytes, from 2013-12-01 19:10:31)
-- /home/robclark/src/freedreno/envytools/rnndb/mdp/mdp_common.xml      (   1615 bytes, from 2013-11-30 15:00:52)
-- /home/robclark/src/freedreno/envytools/rnndb/mdp/mdp5.xml            (  22517 bytes, from 2014-06-25 12:55:02)
+- /home/robclark/src/freedreno/envytools/rnndb/mdp/mdp4.xml            (  20136 bytes, from 2014-10-31 16:51:39)
+- /home/robclark/src/freedreno/envytools/rnndb/mdp/mdp_common.xml      (   1940 bytes, from 2014-10-31 16:51:39)
+- /home/robclark/src/freedreno/envytools/rnndb/mdp/mdp5.xml            (  23963 bytes, from 2014-10-31 16:51:46)
 - /home/robclark/src/freedreno/envytools/rnndb/dsi/dsi.xml             (  11712 bytes, from 2013-08-17 17:13:43)
 - /home/robclark/src/freedreno/envytools/rnndb/dsi/sfpb.xml            (    344 bytes, from 2013-08-11 19:26:32)
-- /home/robclark/src/freedreno/envytools/rnndb/dsi/mmss_cc.xml         (   1544 bytes, from 2013-08-16 19:17:05)
+- /home/robclark/src/freedreno/envytools/rnndb/dsi/mmss_cc.xml         (   1686 bytes, from 2014-10-31 16:48:57)
 - /home/robclark/src/freedreno/envytools/rnndb/hdmi/qfprom.xml         (    600 bytes, from 2013-07-05 19:21:12)
-- /home/robclark/src/freedreno/envytools/rnndb/hdmi/hdmi.xml           (  23613 bytes, from 2014-06-25 12:53:44)
+- /home/robclark/src/freedreno/envytools/rnndb/hdmi/hdmi.xml           (  23613 bytes, from 2014-07-17 15:33:30)
 
-Copyright (C) 2013 by the following authors:
+Copyright (C) 2013-2014 by the following authors:
 - Rob Clark <robdclark@gmail.com> (robclark)
 
 Permission is hereby granted, free of charge, to any person obtaining
@@ -871,6 +871,101 @@ static inline uint32_t MDP4_LCDC_UNDERFLOW_CLR_COLOR(uint32_t val)
 #define MDP4_LCDC_CTRL_POLARITY_VSYNC_LOW			0x00000002
 #define MDP4_LCDC_CTRL_POLARITY_DATA_EN_LOW			0x00000004
 
+#define REG_MDP4_LCDC_LVDS_INTF_CTL				0x000c2000
+#define MDP4_LCDC_LVDS_INTF_CTL_MODE_SEL			0x00000004
+#define MDP4_LCDC_LVDS_INTF_CTL_RGB_OUT				0x00000008
+#define MDP4_LCDC_LVDS_INTF_CTL_CH_SWAP				0x00000010
+#define MDP4_LCDC_LVDS_INTF_CTL_CH1_RES_BIT			0x00000020
+#define MDP4_LCDC_LVDS_INTF_CTL_CH2_RES_BIT			0x00000040
+#define MDP4_LCDC_LVDS_INTF_CTL_ENABLE				0x00000080
+#define MDP4_LCDC_LVDS_INTF_CTL_CH1_DATA_LANE0_EN		0x00000100
+#define MDP4_LCDC_LVDS_INTF_CTL_CH1_DATA_LANE1_EN		0x00000200
+#define MDP4_LCDC_LVDS_INTF_CTL_CH1_DATA_LANE2_EN		0x00000400
+#define MDP4_LCDC_LVDS_INTF_CTL_CH1_DATA_LANE3_EN		0x00000800
+#define MDP4_LCDC_LVDS_INTF_CTL_CH2_DATA_LANE0_EN		0x00001000
+#define MDP4_LCDC_LVDS_INTF_CTL_CH2_DATA_LANE1_EN		0x00002000
+#define MDP4_LCDC_LVDS_INTF_CTL_CH2_DATA_LANE2_EN		0x00004000
+#define MDP4_LCDC_LVDS_INTF_CTL_CH2_DATA_LANE3_EN		0x00008000
+#define MDP4_LCDC_LVDS_INTF_CTL_CH1_CLK_LANE_EN			0x00010000
+#define MDP4_LCDC_LVDS_INTF_CTL_CH2_CLK_LANE_EN			0x00020000
+
+static inline uint32_t REG_MDP4_LCDC_LVDS_MUX_CTL(uint32_t i0) { return 0x000c2014 + 0x8*i0; }
+
+static inline uint32_t REG_MDP4_LCDC_LVDS_MUX_CTL_3_TO_0(uint32_t i0) { return 0x000c2014 + 0x8*i0; }
+#define MDP4_LCDC_LVDS_MUX_CTL_3_TO_0_BIT0__MASK		0x000000ff
+#define MDP4_LCDC_LVDS_MUX_CTL_3_TO_0_BIT0__SHIFT		0
+static inline uint32_t MDP4_LCDC_LVDS_MUX_CTL_3_TO_0_BIT0(uint32_t val)
+{
+	return ((val) << MDP4_LCDC_LVDS_MUX_CTL_3_TO_0_BIT0__SHIFT) & MDP4_LCDC_LVDS_MUX_CTL_3_TO_0_BIT0__MASK;
+}
+#define MDP4_LCDC_LVDS_MUX_CTL_3_TO_0_BIT1__MASK		0x0000ff00
+#define MDP4_LCDC_LVDS_MUX_CTL_3_TO_0_BIT1__SHIFT		8
+static inline uint32_t MDP4_LCDC_LVDS_MUX_CTL_3_TO_0_BIT1(uint32_t val)
+{
+	return ((val) << MDP4_LCDC_LVDS_MUX_CTL_3_TO_0_BIT1__SHIFT) & MDP4_LCDC_LVDS_MUX_CTL_3_TO_0_BIT1__MASK;
+}
+#define MDP4_LCDC_LVDS_MUX_CTL_3_TO_0_BIT2__MASK		0x00ff0000
+#define MDP4_LCDC_LVDS_MUX_CTL_3_TO_0_BIT2__SHIFT		16
+static inline uint32_t MDP4_LCDC_LVDS_MUX_CTL_3_TO_0_BIT2(uint32_t val)
+{
+	return ((val) << MDP4_LCDC_LVDS_MUX_CTL_3_TO_0_BIT2__SHIFT) & MDP4_LCDC_LVDS_MUX_CTL_3_TO_0_BIT2__MASK;
+}
+#define MDP4_LCDC_LVDS_MUX_CTL_3_TO_0_BIT3__MASK		0xff000000
+#define MDP4_LCDC_LVDS_MUX_CTL_3_TO_0_BIT3__SHIFT		24
+static inline uint32_t MDP4_LCDC_LVDS_MUX_CTL_3_TO_0_BIT3(uint32_t val)
+{
+	return ((val) << MDP4_LCDC_LVDS_MUX_CTL_3_TO_0_BIT3__SHIFT) & MDP4_LCDC_LVDS_MUX_CTL_3_TO_0_BIT3__MASK;
+}
+
+static inline uint32_t REG_MDP4_LCDC_LVDS_MUX_CTL_6_TO_4(uint32_t i0) { return 0x000c2018 + 0x8*i0; }
+#define MDP4_LCDC_LVDS_MUX_CTL_6_TO_4_BIT4__MASK		0x000000ff
+#define MDP4_LCDC_LVDS_MUX_CTL_6_TO_4_BIT4__SHIFT		0
+static inline uint32_t MDP4_LCDC_LVDS_MUX_CTL_6_TO_4_BIT4(uint32_t val)
+{
+	return ((val) << MDP4_LCDC_LVDS_MUX_CTL_6_TO_4_BIT4__SHIFT) & MDP4_LCDC_LVDS_MUX_CTL_6_TO_4_BIT4__MASK;
+}
+#define MDP4_LCDC_LVDS_MUX_CTL_6_TO_4_BIT5__MASK		0x0000ff00
+#define MDP4_LCDC_LVDS_MUX_CTL_6_TO_4_BIT5__SHIFT		8
+static inline uint32_t MDP4_LCDC_LVDS_MUX_CTL_6_TO_4_BIT5(uint32_t val)
+{
+	return ((val) << MDP4_LCDC_LVDS_MUX_CTL_6_TO_4_BIT5__SHIFT) & MDP4_LCDC_LVDS_MUX_CTL_6_TO_4_BIT5__MASK;
+}
+#define MDP4_LCDC_LVDS_MUX_CTL_6_TO_4_BIT6__MASK		0x00ff0000
+#define MDP4_LCDC_LVDS_MUX_CTL_6_TO_4_BIT6__SHIFT		16
+static inline uint32_t MDP4_LCDC_LVDS_MUX_CTL_6_TO_4_BIT6(uint32_t val)
+{
+	return ((val) << MDP4_LCDC_LVDS_MUX_CTL_6_TO_4_BIT6__SHIFT) & MDP4_LCDC_LVDS_MUX_CTL_6_TO_4_BIT6__MASK;
+}
+
+#define REG_MDP4_LCDC_LVDS_PHY_RESET				0x000c2034
+
+#define REG_MDP4_LVDS_PHY_PLL_CTRL_0				0x000c3000
+
+#define REG_MDP4_LVDS_PHY_PLL_CTRL_1				0x000c3004
+
+#define REG_MDP4_LVDS_PHY_PLL_CTRL_2				0x000c3008
+
+#define REG_MDP4_LVDS_PHY_PLL_CTRL_3				0x000c300c
+
+#define REG_MDP4_LVDS_PHY_PLL_CTRL_5				0x000c3014
+
+#define REG_MDP4_LVDS_PHY_PLL_CTRL_6				0x000c3018
+
+#define REG_MDP4_LVDS_PHY_PLL_CTRL_7				0x000c301c
+
+#define REG_MDP4_LVDS_PHY_PLL_CTRL_8				0x000c3020
+
+#define REG_MDP4_LVDS_PHY_PLL_CTRL_9				0x000c3024
+
+#define REG_MDP4_LVDS_PHY_PLL_LOCKED				0x000c3080
+
+#define REG_MDP4_LVDS_PHY_CFG2					0x000c3108
+
+#define REG_MDP4_LVDS_PHY_CFG0					0x000c3100
+#define MDP4_LVDS_PHY_CFG0_SERIALIZATION_ENBLE			0x00000010
+#define MDP4_LVDS_PHY_CFG0_CHANNEL0				0x00000040
+#define MDP4_LVDS_PHY_CFG0_CHANNEL1				0x00000080
+
 #define REG_MDP4_DTV						0x000d0000
 
 #define REG_MDP4_DTV_ENABLE					0x000d0000
diff --git a/drivers/gpu/drm/msm/mdp/mdp4/mdp4_crtc.c b/drivers/gpu/drm/msm/mdp/mdp4/mdp4_crtc.c
index c6c80ea28c35..3449213f1e76 100644
--- a/drivers/gpu/drm/msm/mdp/mdp4/mdp4_crtc.c
+++ b/drivers/gpu/drm/msm/mdp/mdp4/mdp4_crtc.c
@@ -25,8 +25,6 @@
 struct mdp4_crtc {
 	struct drm_crtc base;
 	char name[8];
-	struct drm_plane *plane;
-	struct drm_plane *planes[8];
 	int id;
 	int ovlp;
 	enum mdp4_dma dma;
@@ -52,25 +50,11 @@ struct mdp4_crtc {
 
 	/* if there is a pending flip, these will be non-null: */
 	struct drm_pending_vblank_event *event;
-	struct msm_fence_cb pageflip_cb;
 
 #define PENDING_CURSOR 0x1
 #define PENDING_FLIP   0x2
 	atomic_t pending;
 
-	/* the fb that we logically (from PoV of KMS API) hold a ref
-	 * to.  Which we may not yet be scanning out (we may still
-	 * be scanning out previous in case of page_flip while waiting
-	 * for gpu rendering to complete:
-	 */
-	struct drm_framebuffer *fb;
-
-	/* the fb that we currently hold a scanout ref to: */
-	struct drm_framebuffer *scanout_fb;
-
-	/* for unref'ing framebuffers after scanout completes: */
-	struct drm_flip_work unref_fb_work;
-
 	/* for unref'ing cursor bo's after scanout completes: */
 	struct drm_flip_work unref_cursor_work;
 
@@ -97,15 +81,14 @@ static void crtc_flush(struct drm_crtc *crtc)
 {
 	struct mdp4_crtc *mdp4_crtc = to_mdp4_crtc(crtc);
 	struct mdp4_kms *mdp4_kms = get_kms(crtc);
-	uint32_t i, flush = 0;
+	struct drm_plane *plane;
+	uint32_t flush = 0;
 
-	for (i = 0; i < ARRAY_SIZE(mdp4_crtc->planes); i++) {
-		struct drm_plane *plane = mdp4_crtc->planes[i];
-		if (plane) {
-			enum mdp4_pipe pipe_id = mdp4_plane_pipe(plane);
-			flush |= pipe2flush(pipe_id);
-		}
+	drm_atomic_crtc_for_each_plane(plane, crtc) {
+		enum mdp4_pipe pipe_id = mdp4_plane_pipe(plane);
+		flush |= pipe2flush(pipe_id);
 	}
+
 	flush |= ovlp2flush(mdp4_crtc->ovlp);
 
 	DBG("%s: flush=%08x", mdp4_crtc->name, flush);
@@ -113,47 +96,6 @@ static void crtc_flush(struct drm_crtc *crtc)
 	mdp4_write(mdp4_kms, REG_MDP4_OVERLAY_FLUSH, flush);
 }
 
-static void update_fb(struct drm_crtc *crtc, struct drm_framebuffer *new_fb)
-{
-	struct mdp4_crtc *mdp4_crtc = to_mdp4_crtc(crtc);
-	struct drm_framebuffer *old_fb = mdp4_crtc->fb;
-
-	/* grab reference to incoming scanout fb: */
-	drm_framebuffer_reference(new_fb);
-	mdp4_crtc->base.primary->fb = new_fb;
-	mdp4_crtc->fb = new_fb;
-
-	if (old_fb)
-		drm_flip_work_queue(&mdp4_crtc->unref_fb_work, old_fb);
-}
-
-/* unlike update_fb(), take a ref to the new scanout fb *before* updating
- * plane, then call this.  Needed to ensure we don't unref the buffer that
- * is actually still being scanned out.
- *
- * Note that this whole thing goes away with atomic.. since we can defer
- * calling into driver until rendering is done.
- */
-static void update_scanout(struct drm_crtc *crtc, struct drm_framebuffer *fb)
-{
-	struct mdp4_crtc *mdp4_crtc = to_mdp4_crtc(crtc);
-
-	/* flush updates, to make sure hw is updated to new scanout fb,
-	 * so that we can safely queue unref to current fb (ie. next
-	 * vblank we know hw is done w/ previous scanout_fb).
-	 */
-	crtc_flush(crtc);
-
-	if (mdp4_crtc->scanout_fb)
-		drm_flip_work_queue(&mdp4_crtc->unref_fb_work,
-				mdp4_crtc->scanout_fb);
-
-	mdp4_crtc->scanout_fb = fb;
-
-	/* enable vblank to complete flip: */
-	request_pending(crtc, PENDING_FLIP);
-}
-
 /* if file!=NULL, this is preclose potential cancel-flip path */
 static void complete_flip(struct drm_crtc *crtc, struct drm_file *file)
 {
@@ -171,38 +113,13 @@ static void complete_flip(struct drm_crtc *crtc, struct drm_file *file)
 		 */
 		if (!file || (event->base.file_priv == file)) {
 			mdp4_crtc->event = NULL;
+			DBG("%s: send event: %p", mdp4_crtc->name, event);
 			drm_send_vblank_event(dev, mdp4_crtc->id, event);
 		}
 	}
 	spin_unlock_irqrestore(&dev->event_lock, flags);
 }
 
-static void pageflip_cb(struct msm_fence_cb *cb)
-{
-	struct mdp4_crtc *mdp4_crtc =
-		container_of(cb, struct mdp4_crtc, pageflip_cb);
-	struct drm_crtc *crtc = &mdp4_crtc->base;
-	struct drm_framebuffer *fb = crtc->primary->fb;
-
-	if (!fb)
-		return;
-
-	drm_framebuffer_reference(fb);
-	mdp4_plane_set_scanout(mdp4_crtc->plane, fb);
-	update_scanout(crtc, fb);
-}
-
-static void unref_fb_worker(struct drm_flip_work *work, void *val)
-{
-	struct mdp4_crtc *mdp4_crtc =
-		container_of(work, struct mdp4_crtc, unref_fb_work);
-	struct drm_device *dev = mdp4_crtc->base.dev;
-
-	mutex_lock(&dev->mode_config.mutex);
-	drm_framebuffer_unreference(val);
-	mutex_unlock(&dev->mode_config.mutex);
-}
-
 static void unref_cursor_worker(struct drm_flip_work *work, void *val)
 {
 	struct mdp4_crtc *mdp4_crtc =
@@ -218,7 +135,6 @@ static void mdp4_crtc_destroy(struct drm_crtc *crtc)
 	struct mdp4_crtc *mdp4_crtc = to_mdp4_crtc(crtc);
 
 	drm_crtc_cleanup(crtc);
-	drm_flip_work_cleanup(&mdp4_crtc->unref_fb_work);
 	drm_flip_work_cleanup(&mdp4_crtc->unref_cursor_work);
 
 	kfree(mdp4_crtc);
@@ -251,26 +167,53 @@ static bool mdp4_crtc_mode_fixup(struct drm_crtc *crtc,
 	return true;
 }
 
-static void blend_setup(struct drm_crtc *crtc)
+/* statically (for now) map planes to mixer stage (z-order): */
+static const int idxs[] = {
+		[VG1]  = 1,
+		[VG2]  = 2,
+		[RGB1] = 0,
+		[RGB2] = 0,
+		[RGB3] = 0,
+		[VG3]  = 3,
+		[VG4]  = 4,
+
+};
+
+/* setup mixer config, for which we need to consider all crtc's and
+ * the planes attached to them
+ *
+ * TODO may possibly need some extra locking here
+ */
+static void setup_mixer(struct mdp4_kms *mdp4_kms)
 {
-	struct mdp4_crtc *mdp4_crtc = to_mdp4_crtc(crtc);
-	struct mdp4_kms *mdp4_kms = get_kms(crtc);
-	int i, ovlp = mdp4_crtc->ovlp;
+	struct drm_mode_config *config = &mdp4_kms->dev->mode_config;
+	struct drm_crtc *crtc;
 	uint32_t mixer_cfg = 0;
 	static const enum mdp_mixer_stage_id stages[] = {
 			STAGE_BASE, STAGE0, STAGE1, STAGE2, STAGE3,
 	};
-	/* statically (for now) map planes to mixer stage (z-order): */
-	static const int idxs[] = {
-			[VG1]  = 1,
-			[VG2]  = 2,
-			[RGB1] = 0,
-			[RGB2] = 0,
-			[RGB3] = 0,
-			[VG3]  = 3,
-			[VG4]  = 4,
 
-	};
+	list_for_each_entry(crtc, &config->crtc_list, head) {
+		struct mdp4_crtc *mdp4_crtc = to_mdp4_crtc(crtc);
+		struct drm_plane *plane;
+
+		drm_atomic_crtc_for_each_plane(plane, crtc) {
+			enum mdp4_pipe pipe_id = mdp4_plane_pipe(plane);
+			int idx = idxs[pipe_id];
+			mixer_cfg = mixercfg(mixer_cfg, mdp4_crtc->mixer,
+					pipe_id, stages[idx]);
+		}
+	}
+
+	mdp4_write(mdp4_kms, REG_MDP4_LAYERMIXER_IN_CFG, mixer_cfg);
+}
+
+static void blend_setup(struct drm_crtc *crtc)
+{
+	struct mdp4_crtc *mdp4_crtc = to_mdp4_crtc(crtc);
+	struct mdp4_kms *mdp4_kms = get_kms(crtc);
+	struct drm_plane *plane;
+	int i, ovlp = mdp4_crtc->ovlp;
 	bool alpha[4]= { false, false, false, false };
 
 	mdp4_write(mdp4_kms, REG_MDP4_OVLP_TRANSP_LOW0(ovlp), 0);
@@ -278,26 +221,16 @@ static void blend_setup(struct drm_crtc *crtc)
 	mdp4_write(mdp4_kms, REG_MDP4_OVLP_TRANSP_HIGH0(ovlp), 0);
 	mdp4_write(mdp4_kms, REG_MDP4_OVLP_TRANSP_HIGH1(ovlp), 0);
 
-	/* TODO single register for all CRTCs, so this won't work properly
-	 * when multiple CRTCs are active..
-	 */
-	for (i = 0; i < ARRAY_SIZE(mdp4_crtc->planes); i++) {
-		struct drm_plane *plane = mdp4_crtc->planes[i];
-		if (plane) {
-			enum mdp4_pipe pipe_id = mdp4_plane_pipe(plane);
-			int idx = idxs[pipe_id];
-			if (idx > 0) {
-				const struct mdp_format *format =
+	drm_atomic_crtc_for_each_plane(plane, crtc) {
+		enum mdp4_pipe pipe_id = mdp4_plane_pipe(plane);
+		int idx = idxs[pipe_id];
+		if (idx > 0) {
+			const struct mdp_format *format =
 					to_mdp_format(msm_framebuffer_format(plane->fb));
-				alpha[idx-1] = format->alpha_enable;
-			}
-			mixer_cfg |= mixercfg(mdp4_crtc->mixer, pipe_id, stages[idx]);
+			alpha[idx-1] = format->alpha_enable;
 		}
 	}
 
-	/* this shouldn't happen.. and seems to cause underflow: */
-	WARN_ON(!mixer_cfg);
-
 	for (i = 0; i < 4; i++) {
 		uint32_t op;
 
@@ -320,21 +253,21 @@ static void blend_setup(struct drm_crtc *crtc)
 		mdp4_write(mdp4_kms, REG_MDP4_OVLP_STAGE_TRANSP_HIGH1(ovlp, i), 0);
 	}
 
-	mdp4_write(mdp4_kms, REG_MDP4_LAYERMIXER_IN_CFG, mixer_cfg);
+	setup_mixer(mdp4_kms);
 }
 
-static int mdp4_crtc_mode_set(struct drm_crtc *crtc,
-		struct drm_display_mode *mode,
-		struct drm_display_mode *adjusted_mode,
-		int x, int y,
-		struct drm_framebuffer *old_fb)
+static void mdp4_crtc_mode_set_nofb(struct drm_crtc *crtc)
 {
 	struct mdp4_crtc *mdp4_crtc = to_mdp4_crtc(crtc);
 	struct mdp4_kms *mdp4_kms = get_kms(crtc);
 	enum mdp4_dma dma = mdp4_crtc->dma;
-	int ret, ovlp = mdp4_crtc->ovlp;
+	int ovlp = mdp4_crtc->ovlp;
+	struct drm_display_mode *mode;
+
+	if (WARN_ON(!crtc->state))
+		return;
 
-	mode = adjusted_mode;
+	mode = &crtc->state->adjusted_mode;
 
 	DBG("%s: set mode: %d:\"%s\" %d %d %d %d %d %d %d %d %d %d 0x%x 0x%x",
 			mdp4_crtc->name, mode->base.id, mode->name,
@@ -345,28 +278,13 @@ static int mdp4_crtc_mode_set(struct drm_crtc *crtc,
 			mode->vsync_end, mode->vtotal,
 			mode->type, mode->flags);
 
-	/* grab extra ref for update_scanout() */
-	drm_framebuffer_reference(crtc->primary->fb);
-
-	ret = mdp4_plane_mode_set(mdp4_crtc->plane, crtc, crtc->primary->fb,
-			0, 0, mode->hdisplay, mode->vdisplay,
-			x << 16, y << 16,
-			mode->hdisplay << 16, mode->vdisplay << 16);
-	if (ret) {
-		drm_framebuffer_unreference(crtc->primary->fb);
-		dev_err(crtc->dev->dev, "%s: failed to set mode on plane: %d\n",
-				mdp4_crtc->name, ret);
-		return ret;
-	}
-
 	mdp4_write(mdp4_kms, REG_MDP4_DMA_SRC_SIZE(dma),
 			MDP4_DMA_SRC_SIZE_WIDTH(mode->hdisplay) |
 			MDP4_DMA_SRC_SIZE_HEIGHT(mode->vdisplay));
 
 	/* take data from pipe: */
 	mdp4_write(mdp4_kms, REG_MDP4_DMA_SRC_BASE(dma), 0);
-	mdp4_write(mdp4_kms, REG_MDP4_DMA_SRC_STRIDE(dma),
-			crtc->primary->fb->pitches[0]);
+	mdp4_write(mdp4_kms, REG_MDP4_DMA_SRC_STRIDE(dma), 0);
 	mdp4_write(mdp4_kms, REG_MDP4_DMA_DST_SIZE(dma),
 			MDP4_DMA_DST_SIZE_WIDTH(0) |
 			MDP4_DMA_DST_SIZE_HEIGHT(0));
@@ -375,8 +293,7 @@ static int mdp4_crtc_mode_set(struct drm_crtc *crtc,
 	mdp4_write(mdp4_kms, REG_MDP4_OVLP_SIZE(ovlp),
 			MDP4_OVLP_SIZE_WIDTH(mode->hdisplay) |
 			MDP4_OVLP_SIZE_HEIGHT(mode->vdisplay));
-	mdp4_write(mdp4_kms, REG_MDP4_OVLP_STRIDE(ovlp),
-			crtc->primary->fb->pitches[0]);
+	mdp4_write(mdp4_kms, REG_MDP4_OVLP_STRIDE(ovlp), 0);
 
 	mdp4_write(mdp4_kms, REG_MDP4_OVLP_CFG(ovlp), 1);
 
@@ -385,11 +302,6 @@ static int mdp4_crtc_mode_set(struct drm_crtc *crtc,
 		mdp4_write(mdp4_kms, REG_MDP4_DMA_E_QUANT(1), 0x00ff0000);
 		mdp4_write(mdp4_kms, REG_MDP4_DMA_E_QUANT(2), 0x00ff0000);
 	}
-
-	update_fb(crtc, crtc->primary->fb);
-	update_scanout(crtc, crtc->primary->fb);
-
-	return 0;
 }
 
 static void mdp4_crtc_prepare(struct drm_crtc *crtc)
@@ -411,60 +323,42 @@ static void mdp4_crtc_commit(struct drm_crtc *crtc)
 	drm_crtc_vblank_put(crtc);
 }
 
-static int mdp4_crtc_mode_set_base(struct drm_crtc *crtc, int x, int y,
-		struct drm_framebuffer *old_fb)
+static void mdp4_crtc_load_lut(struct drm_crtc *crtc)
 {
-	struct mdp4_crtc *mdp4_crtc = to_mdp4_crtc(crtc);
-	struct drm_plane *plane = mdp4_crtc->plane;
-	struct drm_display_mode *mode = &crtc->mode;
-	int ret;
-
-	/* grab extra ref for update_scanout() */
-	drm_framebuffer_reference(crtc->primary->fb);
-
-	ret = mdp4_plane_mode_set(plane, crtc, crtc->primary->fb,
-			0, 0, mode->hdisplay, mode->vdisplay,
-			x << 16, y << 16,
-			mode->hdisplay << 16, mode->vdisplay << 16);
-	if (ret) {
-		drm_framebuffer_unreference(crtc->primary->fb);
-		return ret;
-	}
-
-	update_fb(crtc, crtc->primary->fb);
-	update_scanout(crtc, crtc->primary->fb);
+}
 
+static int mdp4_crtc_atomic_check(struct drm_crtc *crtc,
+		struct drm_crtc_state *state)
+{
+	struct mdp4_crtc *mdp4_crtc = to_mdp4_crtc(crtc);
+	DBG("%s: check", mdp4_crtc->name);
+	// TODO anything else to check?
 	return 0;
 }
 
-static void mdp4_crtc_load_lut(struct drm_crtc *crtc)
+static void mdp4_crtc_atomic_begin(struct drm_crtc *crtc)
 {
+	struct mdp4_crtc *mdp4_crtc = to_mdp4_crtc(crtc);
+	DBG("%s: begin", mdp4_crtc->name);
 }
 
-static int mdp4_crtc_page_flip(struct drm_crtc *crtc,
-		struct drm_framebuffer *new_fb,
-		struct drm_pending_vblank_event *event,
-		uint32_t page_flip_flags)
+static void mdp4_crtc_atomic_flush(struct drm_crtc *crtc)
 {
 	struct mdp4_crtc *mdp4_crtc = to_mdp4_crtc(crtc);
 	struct drm_device *dev = crtc->dev;
-	struct drm_gem_object *obj;
 	unsigned long flags;
 
-	if (mdp4_crtc->event) {
-		dev_err(dev->dev, "already pending flip!\n");
-		return -EBUSY;
-	}
+	DBG("%s: event: %p", mdp4_crtc->name, crtc->state->event);
 
-	obj = msm_framebuffer_bo(new_fb, 0);
+	WARN_ON(mdp4_crtc->event);
 
 	spin_lock_irqsave(&dev->event_lock, flags);
-	mdp4_crtc->event = event;
+	mdp4_crtc->event = crtc->state->event;
 	spin_unlock_irqrestore(&dev->event_lock, flags);
 
-	update_fb(crtc, new_fb);
-
-	return msm_gem_queue_inactive_cb(obj, &mdp4_crtc->pageflip_cb);
+	blend_setup(crtc);
+	crtc_flush(crtc);
+	request_pending(crtc, PENDING_FLIP);
 }
 
 static int mdp4_crtc_set_property(struct drm_crtc *crtc,
@@ -602,22 +496,29 @@ static int mdp4_crtc_cursor_move(struct drm_crtc *crtc, int x, int y)
 }
 
 static const struct drm_crtc_funcs mdp4_crtc_funcs = {
-	.set_config = drm_crtc_helper_set_config,
+	.set_config = drm_atomic_helper_set_config,
 	.destroy = mdp4_crtc_destroy,
-	.page_flip = mdp4_crtc_page_flip,
+	.page_flip = drm_atomic_helper_page_flip,
 	.set_property = mdp4_crtc_set_property,
 	.cursor_set = mdp4_crtc_cursor_set,
 	.cursor_move = mdp4_crtc_cursor_move,
+	.reset = drm_atomic_helper_crtc_reset,
+	.atomic_duplicate_state = drm_atomic_helper_crtc_duplicate_state,
+	.atomic_destroy_state = drm_atomic_helper_crtc_destroy_state,
 };
 
 static const struct drm_crtc_helper_funcs mdp4_crtc_helper_funcs = {
 	.dpms = mdp4_crtc_dpms,
 	.mode_fixup = mdp4_crtc_mode_fixup,
-	.mode_set = mdp4_crtc_mode_set,
+	.mode_set_nofb = mdp4_crtc_mode_set_nofb,
+	.mode_set = drm_helper_crtc_mode_set,
+	.mode_set_base = drm_helper_crtc_mode_set_base,
 	.prepare = mdp4_crtc_prepare,
 	.commit = mdp4_crtc_commit,
-	.mode_set_base = mdp4_crtc_mode_set_base,
 	.load_lut = mdp4_crtc_load_lut,
+	.atomic_check = mdp4_crtc_atomic_check,
+	.atomic_begin = mdp4_crtc_atomic_begin,
+	.atomic_flush = mdp4_crtc_atomic_flush,
 };
 
 static void mdp4_crtc_vblank_irq(struct mdp_irq *irq, uint32_t irqstatus)
@@ -633,7 +534,6 @@ static void mdp4_crtc_vblank_irq(struct mdp_irq *irq, uint32_t irqstatus)
 
 	if (pending & PENDING_FLIP) {
 		complete_flip(crtc, NULL);
-		drm_flip_work_commit(&mdp4_crtc->unref_fb_work, priv->wq);
 	}
 
 	if (pending & PENDING_CURSOR) {
@@ -658,7 +558,8 @@ uint32_t mdp4_crtc_vblank(struct drm_crtc *crtc)
 
 void mdp4_crtc_cancel_pending_flip(struct drm_crtc *crtc, struct drm_file *file)
 {
-	DBG("cancel: %p", file);
+	struct mdp4_crtc *mdp4_crtc = to_mdp4_crtc(crtc);
+	DBG("%s: cancel: %p", mdp4_crtc->name, file);
 	complete_flip(crtc, file);
 }
 
@@ -672,7 +573,7 @@ void mdp4_crtc_set_config(struct drm_crtc *crtc, uint32_t config)
 }
 
 /* set interface for routing crtc->encoder: */
-void mdp4_crtc_set_intf(struct drm_crtc *crtc, enum mdp4_intf intf)
+void mdp4_crtc_set_intf(struct drm_crtc *crtc, enum mdp4_intf intf, int mixer)
 {
 	struct mdp4_crtc *mdp4_crtc = to_mdp4_crtc(crtc);
 	struct mdp4_kms *mdp4_kms = get_kms(crtc);
@@ -698,15 +599,13 @@ void mdp4_crtc_set_intf(struct drm_crtc *crtc, enum mdp4_intf intf)
 	if (intf == INTF_DSI_VIDEO) {
 		intf_sel &= ~MDP4_DISP_INTF_SEL_DSI_CMD;
 		intf_sel |= MDP4_DISP_INTF_SEL_DSI_VIDEO;
-		mdp4_crtc->mixer = 0;
 	} else if (intf == INTF_DSI_CMD) {
 		intf_sel &= ~MDP4_DISP_INTF_SEL_DSI_VIDEO;
 		intf_sel |= MDP4_DISP_INTF_SEL_DSI_CMD;
-		mdp4_crtc->mixer = 0;
-	} else if (intf == INTF_LCDC_DTV){
-		mdp4_crtc->mixer = 1;
 	}
 
+	mdp4_crtc->mixer = mixer;
+
 	blend_setup(crtc);
 
 	DBG("%s: intf_sel=%08x", mdp4_crtc->name, intf_sel);
@@ -714,35 +613,6 @@ void mdp4_crtc_set_intf(struct drm_crtc *crtc, enum mdp4_intf intf)
 	mdp4_write(mdp4_kms, REG_MDP4_DISP_INTF_SEL, intf_sel);
 }
 
-static void set_attach(struct drm_crtc *crtc, enum mdp4_pipe pipe_id,
-		struct drm_plane *plane)
-{
-	struct mdp4_crtc *mdp4_crtc = to_mdp4_crtc(crtc);
-
-	BUG_ON(pipe_id >= ARRAY_SIZE(mdp4_crtc->planes));
-
-	if (mdp4_crtc->planes[pipe_id] == plane)
-		return;
-
-	mdp4_crtc->planes[pipe_id] = plane;
-	blend_setup(crtc);
-	if (mdp4_crtc->enabled && (plane != mdp4_crtc->plane))
-		crtc_flush(crtc);
-}
-
-void mdp4_crtc_attach(struct drm_crtc *crtc, struct drm_plane *plane)
-{
-	set_attach(crtc, mdp4_plane_pipe(plane), plane);
-}
-
-void mdp4_crtc_detach(struct drm_crtc *crtc, struct drm_plane *plane)
-{
-	/* don't actually detatch our primary plane: */
-	if (to_mdp4_crtc(crtc)->plane == plane)
-		return;
-	set_attach(crtc, mdp4_plane_pipe(plane), NULL);
-}
-
 static const char *dma_names[] = {
 		"DMA_P", "DMA_S", "DMA_E",
 };
@@ -754,17 +624,13 @@ struct drm_crtc *mdp4_crtc_init(struct drm_device *dev,
 {
 	struct drm_crtc *crtc = NULL;
 	struct mdp4_crtc *mdp4_crtc;
-	int ret;
 
 	mdp4_crtc = kzalloc(sizeof(*mdp4_crtc), GFP_KERNEL);
-	if (!mdp4_crtc) {
-		ret = -ENOMEM;
-		goto fail;
-	}
+	if (!mdp4_crtc)
+		return ERR_PTR(-ENOMEM);
 
 	crtc = &mdp4_crtc->base;
 
-	mdp4_crtc->plane = plane;
 	mdp4_crtc->id = id;
 
 	mdp4_crtc->ovlp = ovlp_id;
@@ -781,26 +647,14 @@ struct drm_crtc *mdp4_crtc_init(struct drm_device *dev,
 
 	spin_lock_init(&mdp4_crtc->cursor.lock);
 
-	ret = drm_flip_work_init(&mdp4_crtc->unref_fb_work, 16,
-			"unref fb", unref_fb_worker);
-	if (ret)
-		goto fail;
-
-	ret = drm_flip_work_init(&mdp4_crtc->unref_cursor_work, 64,
+	drm_flip_work_init(&mdp4_crtc->unref_cursor_work,
 			"unref cursor", unref_cursor_worker);
 
-	INIT_FENCE_CB(&mdp4_crtc->pageflip_cb, pageflip_cb);
-
 	drm_crtc_init_with_planes(dev, crtc, plane, NULL, &mdp4_crtc_funcs);
 	drm_crtc_helper_add(crtc, &mdp4_crtc_helper_funcs);
+	plane->crtc = crtc;
 
-	mdp4_plane_install_properties(mdp4_crtc->plane, &crtc->base);
+	mdp4_plane_install_properties(plane, &crtc->base);
 
 	return crtc;
-
-fail:
-	if (crtc)
-		mdp4_crtc_destroy(crtc);
-
-	return ERR_PTR(ret);
 }
diff --git a/drivers/gpu/drm/msm/mdp/mdp4/mdp4_dtv_encoder.c b/drivers/gpu/drm/msm/mdp/mdp4/mdp4_dtv_encoder.c
index 067ed03b35fe..c3878420180b 100644
--- a/drivers/gpu/drm/msm/mdp/mdp4/mdp4_dtv_encoder.c
+++ b/drivers/gpu/drm/msm/mdp/mdp4/mdp4_dtv_encoder.c
@@ -233,7 +233,7 @@ static void mdp4_dtv_encoder_commit(struct drm_encoder *encoder)
 			MDP4_DMA_CONFIG_G_BPC(BPC8) |
 			MDP4_DMA_CONFIG_B_BPC(BPC8) |
 			MDP4_DMA_CONFIG_PACK(0x21));
-	mdp4_crtc_set_intf(encoder->crtc, INTF_LCDC_DTV);
+	mdp4_crtc_set_intf(encoder->crtc, INTF_LCDC_DTV, 1);
 	mdp4_dtv_encoder_dpms(encoder, DRM_MODE_DPMS_ON);
 }
 
diff --git a/drivers/gpu/drm/msm/mdp/mdp4/mdp4_kms.c b/drivers/gpu/drm/msm/mdp/mdp4/mdp4_kms.c
index 733646c0d3f8..a62109e4ae0d 100644
--- a/drivers/gpu/drm/msm/mdp/mdp4/mdp4_kms.c
+++ b/drivers/gpu/drm/msm/mdp/mdp4/mdp4_kms.c
@@ -106,6 +106,7 @@ static int mdp4_hw_init(struct msm_kms *kms)
 
 	if (mdp4_kms->rev >= 2)
 		mdp4_write(mdp4_kms, REG_MDP4_LAYERMIXER_IN_CFG_UPDATE_METHOD, 1);
+	mdp4_write(mdp4_kms, REG_MDP4_LAYERMIXER_IN_CFG, 0);
 
 	/* disable CSC matrix / YUV by default: */
 	mdp4_write(mdp4_kms, REG_MDP4_PIPE_OP_MODE(VG1), 0);
@@ -196,6 +197,28 @@ int mdp4_enable(struct mdp4_kms *mdp4_kms)
 	return 0;
 }
 
+#ifdef CONFIG_OF
+static struct drm_panel *detect_panel(struct drm_device *dev, const char *name)
+{
+	struct device_node *n;
+	struct drm_panel *panel = NULL;
+
+	n = of_parse_phandle(dev->dev->of_node, name, 0);
+	if (n) {
+		panel = of_drm_find_panel(n);
+		if (!panel)
+			panel = ERR_PTR(-EPROBE_DEFER);
+	}
+
+	return panel;
+}
+#else
+static struct drm_panel *detect_panel(struct drm_device *dev, const char *name)
+{
+	// ??? maybe use a module param to specify which panel is attached?
+}
+#endif
+
 static int modeset_init(struct mdp4_kms *mdp4_kms)
 {
 	struct drm_device *dev = mdp4_kms->dev;
@@ -203,14 +226,10 @@ static int modeset_init(struct mdp4_kms *mdp4_kms)
 	struct drm_plane *plane;
 	struct drm_crtc *crtc;
 	struct drm_encoder *encoder;
-	struct hdmi *hdmi;
+	struct drm_connector *connector;
+	struct drm_panel *panel;
 	int ret;
 
-	/*
-	 *  NOTE: this is a bit simplistic until we add support
-	 * for more than just RGB1->DMA_E->DTV->HDMI
-	 */
-
 	/* construct non-private planes: */
 	plane = mdp4_plane_init(dev, VG1, false);
 	if (IS_ERR(plane)) {
@@ -228,7 +247,57 @@ static int modeset_init(struct mdp4_kms *mdp4_kms)
 	}
 	priv->planes[priv->num_planes++] = plane;
 
-	/* the CRTCs get constructed with a private plane: */
+	/*
+	 * Setup the LCDC/LVDS path: RGB2 -> DMA_P -> LCDC -> LVDS:
+	 */
+
+	panel = detect_panel(dev, "qcom,lvds-panel");
+	if (IS_ERR(panel)) {
+		ret = PTR_ERR(panel);
+		dev_err(dev->dev, "failed to detect LVDS panel: %d\n", ret);
+		goto fail;
+	}
+
+	plane = mdp4_plane_init(dev, RGB2, true);
+	if (IS_ERR(plane)) {
+		dev_err(dev->dev, "failed to construct plane for RGB2\n");
+		ret = PTR_ERR(plane);
+		goto fail;
+	}
+
+	crtc  = mdp4_crtc_init(dev, plane, priv->num_crtcs, 0, DMA_P);
+	if (IS_ERR(crtc)) {
+		dev_err(dev->dev, "failed to construct crtc for DMA_P\n");
+		ret = PTR_ERR(crtc);
+		goto fail;
+	}
+
+	encoder = mdp4_lcdc_encoder_init(dev, panel);
+	if (IS_ERR(encoder)) {
+		dev_err(dev->dev, "failed to construct LCDC encoder\n");
+		ret = PTR_ERR(encoder);
+		goto fail;
+	}
+
+	/* LCDC can be hooked to DMA_P: */
+	encoder->possible_crtcs = 1 << priv->num_crtcs;
+
+	priv->crtcs[priv->num_crtcs++] = crtc;
+	priv->encoders[priv->num_encoders++] = encoder;
+
+	connector = mdp4_lvds_connector_init(dev, panel, encoder);
+	if (IS_ERR(connector)) {
+		ret = PTR_ERR(connector);
+		dev_err(dev->dev, "failed to initialize LVDS connector: %d\n", ret);
+		goto fail;
+	}
+
+	priv->connectors[priv->num_connectors++] = connector;
+
+	/*
+	 * Setup DTV/HDMI path: RGB1 -> DMA_E -> DTV -> HDMI:
+	 */
+
 	plane = mdp4_plane_init(dev, RGB1, true);
 	if (IS_ERR(plane)) {
 		dev_err(dev->dev, "failed to construct plane for RGB1\n");
@@ -242,7 +311,6 @@ static int modeset_init(struct mdp4_kms *mdp4_kms)
 		ret = PTR_ERR(crtc);
 		goto fail;
 	}
-	priv->crtcs[priv->num_crtcs++] = crtc;
 
 	encoder = mdp4_dtv_encoder_init(dev);
 	if (IS_ERR(encoder)) {
@@ -250,14 +318,20 @@ static int modeset_init(struct mdp4_kms *mdp4_kms)
 		ret = PTR_ERR(encoder);
 		goto fail;
 	}
-	encoder->possible_crtcs = 0x1;     /* DTV can be hooked to DMA_E */
+
+	/* DTV can be hooked to DMA_E: */
+	encoder->possible_crtcs = 1 << priv->num_crtcs;
+
+	priv->crtcs[priv->num_crtcs++] = crtc;
 	priv->encoders[priv->num_encoders++] = encoder;
 
-	hdmi = hdmi_init(dev, encoder);
-	if (IS_ERR(hdmi)) {
-		ret = PTR_ERR(hdmi);
-		dev_err(dev->dev, "failed to initialize HDMI: %d\n", ret);
-		goto fail;
+	if (priv->hdmi) {
+		/* Construct bridge/connector for HDMI: */
+		ret = hdmi_modeset_init(priv->hdmi, dev, encoder);
+		if (ret) {
+			dev_err(dev->dev, "failed to initialize HDMI: %d\n", ret);
+			goto fail;
+		}
 	}
 
 	return 0;
@@ -308,6 +382,10 @@ struct msm_kms *mdp4_kms_init(struct drm_device *dev)
 	if (IS_ERR(mdp4_kms->dsi_pll_vddio))
 		mdp4_kms->dsi_pll_vddio = NULL;
 
+	/* NOTE: driver for this regulator still missing upstream.. use
+	 * _get_exclusive() and ignore the error if it does not exist
+	 * (and hope that the bootloader left it on for us)
+	 */
 	mdp4_kms->vdd = devm_regulator_get_exclusive(&pdev->dev, "vdd");
 	if (IS_ERR(mdp4_kms->vdd))
 		mdp4_kms->vdd = NULL;
diff --git a/drivers/gpu/drm/msm/mdp/mdp4/mdp4_kms.h b/drivers/gpu/drm/msm/mdp/mdp4/mdp4_kms.h
index 3225da804c61..cbd77bc626d5 100644
--- a/drivers/gpu/drm/msm/mdp/mdp4/mdp4_kms.h
+++ b/drivers/gpu/drm/msm/mdp/mdp4/mdp4_kms.h
@@ -23,6 +23,8 @@
 #include "mdp/mdp_kms.h"
 #include "mdp4.xml.h"
 
+#include "drm_panel.h"
+
 struct mdp4_kms {
 	struct mdp_kms base;
 
@@ -74,7 +76,7 @@ static inline uint32_t pipe2flush(enum mdp4_pipe pipe)
 	case VG1:      return MDP4_OVERLAY_FLUSH_VG1;
 	case VG2:      return MDP4_OVERLAY_FLUSH_VG2;
 	case RGB1:     return MDP4_OVERLAY_FLUSH_RGB1;
-	case RGB2:     return MDP4_OVERLAY_FLUSH_RGB1;
+	case RGB2:     return MDP4_OVERLAY_FLUSH_RGB2;
 	default:       return 0;
 	}
 }
@@ -108,38 +110,50 @@ static inline uint32_t dma2err(enum mdp4_dma dma)
 	}
 }
 
-static inline uint32_t mixercfg(int mixer, enum mdp4_pipe pipe,
-		enum mdp_mixer_stage_id stage)
+static inline uint32_t mixercfg(uint32_t mixer_cfg, int mixer,
+		enum mdp4_pipe pipe, enum mdp_mixer_stage_id stage)
 {
-	uint32_t mixer_cfg = 0;
-
 	switch (pipe) {
 	case VG1:
-		mixer_cfg = MDP4_LAYERMIXER_IN_CFG_PIPE0(stage) |
+		mixer_cfg &= ~(MDP4_LAYERMIXER_IN_CFG_PIPE0__MASK |
+				MDP4_LAYERMIXER_IN_CFG_PIPE0_MIXER1);
+		mixer_cfg |= MDP4_LAYERMIXER_IN_CFG_PIPE0(stage) |
 			COND(mixer == 1, MDP4_LAYERMIXER_IN_CFG_PIPE0_MIXER1);
 		break;
 	case VG2:
-		mixer_cfg = MDP4_LAYERMIXER_IN_CFG_PIPE1(stage) |
+		mixer_cfg &= ~(MDP4_LAYERMIXER_IN_CFG_PIPE1__MASK |
+				MDP4_LAYERMIXER_IN_CFG_PIPE1_MIXER1);
+		mixer_cfg |= MDP4_LAYERMIXER_IN_CFG_PIPE1(stage) |
 			COND(mixer == 1, MDP4_LAYERMIXER_IN_CFG_PIPE1_MIXER1);
 		break;
 	case RGB1:
-		mixer_cfg = MDP4_LAYERMIXER_IN_CFG_PIPE2(stage) |
+		mixer_cfg &= ~(MDP4_LAYERMIXER_IN_CFG_PIPE2__MASK |
+				MDP4_LAYERMIXER_IN_CFG_PIPE2_MIXER1);
+		mixer_cfg |= MDP4_LAYERMIXER_IN_CFG_PIPE2(stage) |
 			COND(mixer == 1, MDP4_LAYERMIXER_IN_CFG_PIPE2_MIXER1);
 		break;
 	case RGB2:
-		mixer_cfg = MDP4_LAYERMIXER_IN_CFG_PIPE3(stage) |
+		mixer_cfg &= ~(MDP4_LAYERMIXER_IN_CFG_PIPE3__MASK |
+				MDP4_LAYERMIXER_IN_CFG_PIPE3_MIXER1);
+		mixer_cfg |= MDP4_LAYERMIXER_IN_CFG_PIPE3(stage) |
 			COND(mixer == 1, MDP4_LAYERMIXER_IN_CFG_PIPE3_MIXER1);
 		break;
 	case RGB3:
-		mixer_cfg = MDP4_LAYERMIXER_IN_CFG_PIPE4(stage) |
+		mixer_cfg &= ~(MDP4_LAYERMIXER_IN_CFG_PIPE4__MASK |
+				MDP4_LAYERMIXER_IN_CFG_PIPE4_MIXER1);
+		mixer_cfg |= MDP4_LAYERMIXER_IN_CFG_PIPE4(stage) |
 			COND(mixer == 1, MDP4_LAYERMIXER_IN_CFG_PIPE4_MIXER1);
 		break;
 	case VG3:
-		mixer_cfg = MDP4_LAYERMIXER_IN_CFG_PIPE5(stage) |
+		mixer_cfg &= ~(MDP4_LAYERMIXER_IN_CFG_PIPE5__MASK |
+				MDP4_LAYERMIXER_IN_CFG_PIPE5_MIXER1);
+		mixer_cfg |= MDP4_LAYERMIXER_IN_CFG_PIPE5(stage) |
 			COND(mixer == 1, MDP4_LAYERMIXER_IN_CFG_PIPE5_MIXER1);
 		break;
 	case VG4:
-		mixer_cfg = MDP4_LAYERMIXER_IN_CFG_PIPE6(stage) |
+		mixer_cfg &= ~(MDP4_LAYERMIXER_IN_CFG_PIPE6__MASK |
+				MDP4_LAYERMIXER_IN_CFG_PIPE6_MIXER1);
+		mixer_cfg |= MDP4_LAYERMIXER_IN_CFG_PIPE6(stage) |
 			COND(mixer == 1, MDP4_LAYERMIXER_IN_CFG_PIPE6_MIXER1);
 		break;
 	default:
@@ -173,14 +187,6 @@ uint32_t mdp4_get_formats(enum mdp4_pipe pipe_id, uint32_t *pixel_formats,
 
 void mdp4_plane_install_properties(struct drm_plane *plane,
 		struct drm_mode_object *obj);
-void mdp4_plane_set_scanout(struct drm_plane *plane,
-		struct drm_framebuffer *fb);
-int mdp4_plane_mode_set(struct drm_plane *plane,
-		struct drm_crtc *crtc, struct drm_framebuffer *fb,
-		int crtc_x, int crtc_y,
-		unsigned int crtc_w, unsigned int crtc_h,
-		uint32_t src_x, uint32_t src_y,
-		uint32_t src_w, uint32_t src_h);
 enum mdp4_pipe mdp4_plane_pipe(struct drm_plane *plane);
 struct drm_plane *mdp4_plane_init(struct drm_device *dev,
 		enum mdp4_pipe pipe_id, bool private_plane);
@@ -188,9 +194,7 @@ struct drm_plane *mdp4_plane_init(struct drm_device *dev,
 uint32_t mdp4_crtc_vblank(struct drm_crtc *crtc);
 void mdp4_crtc_cancel_pending_flip(struct drm_crtc *crtc, struct drm_file *file);
 void mdp4_crtc_set_config(struct drm_crtc *crtc, uint32_t config);
-void mdp4_crtc_set_intf(struct drm_crtc *crtc, enum mdp4_intf intf);
-void mdp4_crtc_attach(struct drm_crtc *crtc, struct drm_plane *plane);
-void mdp4_crtc_detach(struct drm_crtc *crtc, struct drm_plane *plane);
+void mdp4_crtc_set_intf(struct drm_crtc *crtc, enum mdp4_intf intf, int mixer);
 struct drm_crtc *mdp4_crtc_init(struct drm_device *dev,
 		struct drm_plane *plane, int id, int ovlp_id,
 		enum mdp4_dma dma_id);
@@ -198,6 +202,22 @@ struct drm_crtc *mdp4_crtc_init(struct drm_device *dev,
 long mdp4_dtv_round_pixclk(struct drm_encoder *encoder, unsigned long rate);
 struct drm_encoder *mdp4_dtv_encoder_init(struct drm_device *dev);
 
+long mdp4_lcdc_round_pixclk(struct drm_encoder *encoder, unsigned long rate);
+struct drm_encoder *mdp4_lcdc_encoder_init(struct drm_device *dev,
+		struct drm_panel *panel);
+
+struct drm_connector *mdp4_lvds_connector_init(struct drm_device *dev,
+		struct drm_panel *panel, struct drm_encoder *encoder);
+
+#ifdef CONFIG_COMMON_CLK
+struct clk *mpd4_lvds_pll_init(struct drm_device *dev);
+#else
+static inline struct clk *mpd4_lvds_pll_init(struct drm_device *dev)
+{
+	return ERR_PTR(-ENODEV);
+}
+#endif
+
 #ifdef CONFIG_MSM_BUS_SCALING
 static inline int match_dev_name(struct device *dev, void *data)
 {
diff --git a/drivers/gpu/drm/msm/mdp/mdp4/mdp4_lcdc_encoder.c b/drivers/gpu/drm/msm/mdp/mdp4/mdp4_lcdc_encoder.c
new file mode 100644
index 000000000000..41f6436754fc
--- /dev/null
+++ b/drivers/gpu/drm/msm/mdp/mdp4/mdp4_lcdc_encoder.c
@@ -0,0 +1,506 @@
+/*
+ * Copyright (C) 2014 Red Hat
+ * Author: Rob Clark <robdclark@gmail.com>
+ * Author: Vinay Simha <vinaysimha@inforcecomputing.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published by
+ * the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along with
+ * this program.  If not, see <http://www.gnu.org/licenses/>.
+ */
+
+#include "mdp4_kms.h"
+
+#include "drm_crtc.h"
+#include "drm_crtc_helper.h"
+
+struct mdp4_lcdc_encoder {
+	struct drm_encoder base;
+	struct drm_panel *panel;
+	struct clk *lcdc_clk;
+	unsigned long int pixclock;
+	struct regulator *regs[3];
+	bool enabled;
+	uint32_t bsc;
+};
+#define to_mdp4_lcdc_encoder(x) container_of(x, struct mdp4_lcdc_encoder, base)
+
+static struct mdp4_kms *get_kms(struct drm_encoder *encoder)
+{
+	struct msm_drm_private *priv = encoder->dev->dev_private;
+	return to_mdp4_kms(to_mdp_kms(priv->kms));
+}
+
+#ifdef CONFIG_MSM_BUS_SCALING
+#include <mach/board.h>
+static void bs_init(struct mdp4_lcdc_encoder *mdp4_lcdc_encoder)
+{
+	struct drm_device *dev = mdp4_lcdc_encoder->base.dev;
+	struct lcdc_platform_data *lcdc_pdata = mdp4_find_pdata("lvds.0");
+
+	if (!lcdc_pdata) {
+		dev_err(dev->dev, "could not find lvds pdata\n");
+		return;
+	}
+
+	if (lcdc_pdata->bus_scale_table) {
+		mdp4_lcdc_encoder->bsc = msm_bus_scale_register_client(
+				lcdc_pdata->bus_scale_table);
+		DBG("lvds : bus scale client: %08x", mdp4_lcdc_encoder->bsc);
+	}
+}
+
+static void bs_fini(struct mdp4_lcdc_encoder *mdp4_lcdc_encoder)
+{
+	if (mdp4_lcdc_encoder->bsc) {
+		msm_bus_scale_unregister_client(mdp4_lcdc_encoder->bsc);
+		mdp4_lcdc_encoder->bsc = 0;
+	}
+}
+
+static void bs_set(struct mdp4_lcdc_encoder *mdp4_lcdc_encoder, int idx)
+{
+	if (mdp4_lcdc_encoder->bsc) {
+		DBG("set bus scaling: %d", idx);
+		msm_bus_scale_client_update_request(mdp4_lcdc_encoder->bsc, idx);
+	}
+}
+#else
+static void bs_init(struct mdp4_lcdc_encoder *mdp4_lcdc_encoder) {}
+static void bs_fini(struct mdp4_lcdc_encoder *mdp4_lcdc_encoder) {}
+static void bs_set(struct mdp4_lcdc_encoder *mdp4_lcdc_encoder, int idx) {}
+#endif
+
+static void mdp4_lcdc_encoder_destroy(struct drm_encoder *encoder)
+{
+	struct mdp4_lcdc_encoder *mdp4_lcdc_encoder =
+			to_mdp4_lcdc_encoder(encoder);
+	bs_fini(mdp4_lcdc_encoder);
+	drm_encoder_cleanup(encoder);
+	kfree(mdp4_lcdc_encoder);
+}
+
+static const struct drm_encoder_funcs mdp4_lcdc_encoder_funcs = {
+	.destroy = mdp4_lcdc_encoder_destroy,
+};
+
+/* this should probably be a helper: */
+struct drm_connector *get_connector(struct drm_encoder *encoder)
+{
+	struct drm_device *dev = encoder->dev;
+	struct drm_connector *connector;
+
+	list_for_each_entry(connector, &dev->mode_config.connector_list, head)
+		if (connector->encoder == encoder)
+			return connector;
+
+	return NULL;
+}
+
+static void setup_phy(struct drm_encoder *encoder)
+{
+	struct drm_device *dev = encoder->dev;
+	struct drm_connector *connector = get_connector(encoder);
+	struct mdp4_kms *mdp4_kms = get_kms(encoder);
+	uint32_t lvds_intf = 0, lvds_phy_cfg0 = 0;
+	int bpp, nchan, swap;
+
+	if (!connector)
+		return;
+
+	bpp = 3 * connector->display_info.bpc;
+
+	if (!bpp)
+		bpp = 18;
+
+	/* TODO, these should come from panel somehow: */
+	nchan = 1;
+	swap = 0;
+
+	switch (bpp) {
+	case 24:
+		mdp4_write(mdp4_kms, REG_MDP4_LCDC_LVDS_MUX_CTL_3_TO_0(0),
+				MDP4_LCDC_LVDS_MUX_CTL_3_TO_0_BIT0(0x08) |
+				MDP4_LCDC_LVDS_MUX_CTL_3_TO_0_BIT1(0x05) |
+				MDP4_LCDC_LVDS_MUX_CTL_3_TO_0_BIT2(0x04) |
+				MDP4_LCDC_LVDS_MUX_CTL_3_TO_0_BIT3(0x03));
+		mdp4_write(mdp4_kms, REG_MDP4_LCDC_LVDS_MUX_CTL_6_TO_4(0),
+				MDP4_LCDC_LVDS_MUX_CTL_6_TO_4_BIT4(0x02) |
+				MDP4_LCDC_LVDS_MUX_CTL_6_TO_4_BIT5(0x01) |
+				MDP4_LCDC_LVDS_MUX_CTL_6_TO_4_BIT6(0x00));
+		mdp4_write(mdp4_kms, REG_MDP4_LCDC_LVDS_MUX_CTL_3_TO_0(1),
+				MDP4_LCDC_LVDS_MUX_CTL_3_TO_0_BIT0(0x11) |
+				MDP4_LCDC_LVDS_MUX_CTL_3_TO_0_BIT1(0x10) |
+				MDP4_LCDC_LVDS_MUX_CTL_3_TO_0_BIT2(0x0d) |
+				MDP4_LCDC_LVDS_MUX_CTL_3_TO_0_BIT3(0x0c));
+		mdp4_write(mdp4_kms, REG_MDP4_LCDC_LVDS_MUX_CTL_6_TO_4(1),
+				MDP4_LCDC_LVDS_MUX_CTL_6_TO_4_BIT4(0x0b) |
+				MDP4_LCDC_LVDS_MUX_CTL_6_TO_4_BIT5(0x0a) |
+				MDP4_LCDC_LVDS_MUX_CTL_6_TO_4_BIT6(0x09));
+		mdp4_write(mdp4_kms, REG_MDP4_LCDC_LVDS_MUX_CTL_3_TO_0(2),
+				MDP4_LCDC_LVDS_MUX_CTL_3_TO_0_BIT0(0x1a) |
+				MDP4_LCDC_LVDS_MUX_CTL_3_TO_0_BIT1(0x19) |
+				MDP4_LCDC_LVDS_MUX_CTL_3_TO_0_BIT2(0x18) |
+				MDP4_LCDC_LVDS_MUX_CTL_3_TO_0_BIT3(0x15));
+		mdp4_write(mdp4_kms, REG_MDP4_LCDC_LVDS_MUX_CTL_6_TO_4(2),
+				MDP4_LCDC_LVDS_MUX_CTL_6_TO_4_BIT4(0x14) |
+				MDP4_LCDC_LVDS_MUX_CTL_6_TO_4_BIT5(0x13) |
+				MDP4_LCDC_LVDS_MUX_CTL_6_TO_4_BIT6(0x12));
+		mdp4_write(mdp4_kms, REG_MDP4_LCDC_LVDS_MUX_CTL_3_TO_0(3),
+				MDP4_LCDC_LVDS_MUX_CTL_3_TO_0_BIT0(0x1b) |
+				MDP4_LCDC_LVDS_MUX_CTL_3_TO_0_BIT1(0x17) |
+				MDP4_LCDC_LVDS_MUX_CTL_3_TO_0_BIT2(0x16) |
+				MDP4_LCDC_LVDS_MUX_CTL_3_TO_0_BIT3(0x0f));
+		mdp4_write(mdp4_kms, REG_MDP4_LCDC_LVDS_MUX_CTL_6_TO_4(3),
+				MDP4_LCDC_LVDS_MUX_CTL_6_TO_4_BIT4(0x0e) |
+				MDP4_LCDC_LVDS_MUX_CTL_6_TO_4_BIT5(0x07) |
+				MDP4_LCDC_LVDS_MUX_CTL_6_TO_4_BIT6(0x06));
+		if (nchan == 2) {
+			lvds_intf |= MDP4_LCDC_LVDS_INTF_CTL_CH2_DATA_LANE3_EN |
+					MDP4_LCDC_LVDS_INTF_CTL_CH2_DATA_LANE2_EN |
+					MDP4_LCDC_LVDS_INTF_CTL_CH2_DATA_LANE1_EN |
+					MDP4_LCDC_LVDS_INTF_CTL_CH2_DATA_LANE0_EN |
+					MDP4_LCDC_LVDS_INTF_CTL_CH1_DATA_LANE3_EN |
+					MDP4_LCDC_LVDS_INTF_CTL_CH1_DATA_LANE2_EN |
+					MDP4_LCDC_LVDS_INTF_CTL_CH1_DATA_LANE1_EN |
+					MDP4_LCDC_LVDS_INTF_CTL_CH1_DATA_LANE0_EN;
+		} else {
+			lvds_intf |= MDP4_LCDC_LVDS_INTF_CTL_CH1_DATA_LANE3_EN |
+					MDP4_LCDC_LVDS_INTF_CTL_CH1_DATA_LANE2_EN |
+					MDP4_LCDC_LVDS_INTF_CTL_CH1_DATA_LANE1_EN |
+					MDP4_LCDC_LVDS_INTF_CTL_CH1_DATA_LANE0_EN;
+		}
+		break;
+
+	case 18:
+		mdp4_write(mdp4_kms, REG_MDP4_LCDC_LVDS_MUX_CTL_3_TO_0(0),
+				MDP4_LCDC_LVDS_MUX_CTL_3_TO_0_BIT0(0x0a) |
+				MDP4_LCDC_LVDS_MUX_CTL_3_TO_0_BIT1(0x07) |
+				MDP4_LCDC_LVDS_MUX_CTL_3_TO_0_BIT2(0x06) |
+				MDP4_LCDC_LVDS_MUX_CTL_3_TO_0_BIT3(0x05));
+		mdp4_write(mdp4_kms, REG_MDP4_LCDC_LVDS_MUX_CTL_6_TO_4(0),
+				MDP4_LCDC_LVDS_MUX_CTL_6_TO_4_BIT4(0x04) |
+				MDP4_LCDC_LVDS_MUX_CTL_6_TO_4_BIT5(0x03) |
+				MDP4_LCDC_LVDS_MUX_CTL_6_TO_4_BIT6(0x02));
+		mdp4_write(mdp4_kms, REG_MDP4_LCDC_LVDS_MUX_CTL_3_TO_0(1),
+				MDP4_LCDC_LVDS_MUX_CTL_3_TO_0_BIT0(0x13) |
+				MDP4_LCDC_LVDS_MUX_CTL_3_TO_0_BIT1(0x12) |
+				MDP4_LCDC_LVDS_MUX_CTL_3_TO_0_BIT2(0x0f) |
+				MDP4_LCDC_LVDS_MUX_CTL_3_TO_0_BIT3(0x0e));
+		mdp4_write(mdp4_kms, REG_MDP4_LCDC_LVDS_MUX_CTL_6_TO_4(1),
+				MDP4_LCDC_LVDS_MUX_CTL_6_TO_4_BIT4(0x0d) |
+				MDP4_LCDC_LVDS_MUX_CTL_6_TO_4_BIT5(0x0c) |
+				MDP4_LCDC_LVDS_MUX_CTL_6_TO_4_BIT6(0x0b));
+		mdp4_write(mdp4_kms, REG_MDP4_LCDC_LVDS_MUX_CTL_3_TO_0(2),
+				MDP4_LCDC_LVDS_MUX_CTL_3_TO_0_BIT0(0x1a) |
+				MDP4_LCDC_LVDS_MUX_CTL_3_TO_0_BIT1(0x19) |
+				MDP4_LCDC_LVDS_MUX_CTL_3_TO_0_BIT2(0x18) |
+				MDP4_LCDC_LVDS_MUX_CTL_3_TO_0_BIT3(0x17));
+		mdp4_write(mdp4_kms, REG_MDP4_LCDC_LVDS_MUX_CTL_6_TO_4(2),
+				MDP4_LCDC_LVDS_MUX_CTL_6_TO_4_BIT4(0x16) |
+				MDP4_LCDC_LVDS_MUX_CTL_6_TO_4_BIT5(0x15) |
+				MDP4_LCDC_LVDS_MUX_CTL_6_TO_4_BIT6(0x14));
+		if (nchan == 2) {
+			lvds_intf |= MDP4_LCDC_LVDS_INTF_CTL_CH2_DATA_LANE2_EN |
+					MDP4_LCDC_LVDS_INTF_CTL_CH2_DATA_LANE1_EN |
+					MDP4_LCDC_LVDS_INTF_CTL_CH2_DATA_LANE0_EN |
+					MDP4_LCDC_LVDS_INTF_CTL_CH1_DATA_LANE2_EN |
+					MDP4_LCDC_LVDS_INTF_CTL_CH1_DATA_LANE1_EN |
+					MDP4_LCDC_LVDS_INTF_CTL_CH1_DATA_LANE0_EN;
+		} else {
+			lvds_intf |= MDP4_LCDC_LVDS_INTF_CTL_CH1_DATA_LANE2_EN |
+					MDP4_LCDC_LVDS_INTF_CTL_CH1_DATA_LANE1_EN |
+					MDP4_LCDC_LVDS_INTF_CTL_CH1_DATA_LANE0_EN;
+		}
+		lvds_intf |= MDP4_LCDC_LVDS_INTF_CTL_RGB_OUT;
+		break;
+
+	default:
+		dev_err(dev->dev, "unknown bpp: %d\n", bpp);
+		return;
+	}
+
+	switch (nchan) {
+	case 1:
+		lvds_phy_cfg0 = MDP4_LVDS_PHY_CFG0_CHANNEL0;
+		lvds_intf |= MDP4_LCDC_LVDS_INTF_CTL_CH1_CLK_LANE_EN |
+				MDP4_LCDC_LVDS_INTF_CTL_MODE_SEL;
+		break;
+	case 2:
+		lvds_phy_cfg0 = MDP4_LVDS_PHY_CFG0_CHANNEL0 |
+				MDP4_LVDS_PHY_CFG0_CHANNEL1;
+		lvds_intf |= MDP4_LCDC_LVDS_INTF_CTL_CH2_CLK_LANE_EN |
+				MDP4_LCDC_LVDS_INTF_CTL_CH1_CLK_LANE_EN;
+		break;
+	default:
+		dev_err(dev->dev, "unknown # of channels: %d\n", nchan);
+		return;
+	}
+
+	if (swap)
+		lvds_intf |= MDP4_LCDC_LVDS_INTF_CTL_CH_SWAP;
+
+	lvds_intf |= MDP4_LCDC_LVDS_INTF_CTL_ENABLE;
+
+	mdp4_write(mdp4_kms, REG_MDP4_LVDS_PHY_CFG0, lvds_phy_cfg0);
+	mdp4_write(mdp4_kms, REG_MDP4_LCDC_LVDS_INTF_CTL, lvds_intf);
+	mdp4_write(mdp4_kms, REG_MDP4_LVDS_PHY_CFG2, 0x30);
+
+	mb();
+	udelay(1);
+	lvds_phy_cfg0 |= MDP4_LVDS_PHY_CFG0_SERIALIZATION_ENBLE;
+	mdp4_write(mdp4_kms, REG_MDP4_LVDS_PHY_CFG0, lvds_phy_cfg0);
+}
+
+static void mdp4_lcdc_encoder_dpms(struct drm_encoder *encoder, int mode)
+{
+	struct drm_device *dev = encoder->dev;
+	struct mdp4_lcdc_encoder *mdp4_lcdc_encoder =
+			to_mdp4_lcdc_encoder(encoder);
+	struct mdp4_kms *mdp4_kms = get_kms(encoder);
+	struct drm_panel *panel = mdp4_lcdc_encoder->panel;
+	bool enabled = (mode == DRM_MODE_DPMS_ON);
+	int i, ret;
+
+	DBG("mode=%d", mode);
+
+	if (enabled == mdp4_lcdc_encoder->enabled)
+		return;
+
+	if (enabled) {
+		unsigned long pc = mdp4_lcdc_encoder->pixclock;
+		int ret;
+
+		bs_set(mdp4_lcdc_encoder, 1);
+
+		for (i = 0; i < ARRAY_SIZE(mdp4_lcdc_encoder->regs); i++) {
+			ret = regulator_enable(mdp4_lcdc_encoder->regs[i]);
+			if (ret)
+				dev_err(dev->dev, "failed to enable regulator: %d\n", ret);
+		}
+
+		DBG("setting lcdc_clk=%lu", pc);
+		ret = clk_set_rate(mdp4_lcdc_encoder->lcdc_clk, pc);
+		if (ret)
+			dev_err(dev->dev, "failed to configure lcdc_clk: %d\n", ret);
+		ret = clk_prepare_enable(mdp4_lcdc_encoder->lcdc_clk);
+		if (ret)
+			dev_err(dev->dev, "failed to enable lcdc_clk: %d\n", ret);
+
+		if (panel)
+			drm_panel_enable(panel);
+
+		setup_phy(encoder);
+
+		mdp4_write(mdp4_kms, REG_MDP4_LCDC_ENABLE, 1);
+	} else {
+		mdp4_write(mdp4_kms, REG_MDP4_LCDC_ENABLE, 0);
+
+		if (panel)
+			drm_panel_disable(panel);
+
+		/*
+		 * Wait for a vsync so we know the ENABLE=0 latched before
+		 * the (connector) source of the vsync's gets disabled,
+		 * otherwise we end up in a funny state if we re-enable
+		 * before the disable latches, which results that some of
+		 * the settings changes for the new modeset (like new
+		 * scanout buffer) don't latch properly..
+		 */
+		mdp_irq_wait(&mdp4_kms->base, MDP4_IRQ_PRIMARY_VSYNC);
+
+		clk_disable_unprepare(mdp4_lcdc_encoder->lcdc_clk);
+
+		for (i = 0; i < ARRAY_SIZE(mdp4_lcdc_encoder->regs); i++) {
+			ret = regulator_disable(mdp4_lcdc_encoder->regs[i]);
+			if (ret)
+				dev_err(dev->dev, "failed to disable regulator: %d\n", ret);
+		}
+
+		bs_set(mdp4_lcdc_encoder, 0);
+	}
+
+	mdp4_lcdc_encoder->enabled = enabled;
+}
+
+static bool mdp4_lcdc_encoder_mode_fixup(struct drm_encoder *encoder,
+		const struct drm_display_mode *mode,
+		struct drm_display_mode *adjusted_mode)
+{
+	return true;
+}
+
+static void mdp4_lcdc_encoder_mode_set(struct drm_encoder *encoder,
+		struct drm_display_mode *mode,
+		struct drm_display_mode *adjusted_mode)
+{
+	struct mdp4_lcdc_encoder *mdp4_lcdc_encoder =
+			to_mdp4_lcdc_encoder(encoder);
+	struct mdp4_kms *mdp4_kms = get_kms(encoder);
+	uint32_t lcdc_hsync_skew, vsync_period, vsync_len, ctrl_pol;
+	uint32_t display_v_start, display_v_end;
+	uint32_t hsync_start_x, hsync_end_x;
+
+	mode = adjusted_mode;
+
+	DBG("set mode: %d:\"%s\" %d %d %d %d %d %d %d %d %d %d 0x%x 0x%x",
+			mode->base.id, mode->name,
+			mode->vrefresh, mode->clock,
+			mode->hdisplay, mode->hsync_start,
+			mode->hsync_end, mode->htotal,
+			mode->vdisplay, mode->vsync_start,
+			mode->vsync_end, mode->vtotal,
+			mode->type, mode->flags);
+
+	mdp4_lcdc_encoder->pixclock = mode->clock * 1000;
+
+	DBG("pixclock=%lu", mdp4_lcdc_encoder->pixclock);
+
+	ctrl_pol = 0;
+	if (mode->flags & DRM_MODE_FLAG_NHSYNC)
+		ctrl_pol |= MDP4_LCDC_CTRL_POLARITY_HSYNC_LOW;
+	if (mode->flags & DRM_MODE_FLAG_NVSYNC)
+		ctrl_pol |= MDP4_LCDC_CTRL_POLARITY_VSYNC_LOW;
+	/* probably need to get DATA_EN polarity from panel.. */
+
+	lcdc_hsync_skew = 0;  /* get this from panel? */
+
+	hsync_start_x = (mode->htotal - mode->hsync_start);
+	hsync_end_x = mode->htotal - (mode->hsync_start - mode->hdisplay) - 1;
+
+	vsync_period = mode->vtotal * mode->htotal;
+	vsync_len = (mode->vsync_end - mode->vsync_start) * mode->htotal;
+	display_v_start = (mode->vtotal - mode->vsync_start) * mode->htotal + lcdc_hsync_skew;
+	display_v_end = vsync_period - ((mode->vsync_start - mode->vdisplay) * mode->htotal) + lcdc_hsync_skew - 1;
+
+	mdp4_write(mdp4_kms, REG_MDP4_LCDC_HSYNC_CTRL,
+			MDP4_LCDC_HSYNC_CTRL_PULSEW(mode->hsync_end - mode->hsync_start) |
+			MDP4_LCDC_HSYNC_CTRL_PERIOD(mode->htotal));
+	mdp4_write(mdp4_kms, REG_MDP4_LCDC_VSYNC_PERIOD, vsync_period);
+	mdp4_write(mdp4_kms, REG_MDP4_LCDC_VSYNC_LEN, vsync_len);
+	mdp4_write(mdp4_kms, REG_MDP4_LCDC_DISPLAY_HCTRL,
+			MDP4_LCDC_DISPLAY_HCTRL_START(hsync_start_x) |
+			MDP4_LCDC_DISPLAY_HCTRL_END(hsync_end_x));
+	mdp4_write(mdp4_kms, REG_MDP4_LCDC_DISPLAY_VSTART, display_v_start);
+	mdp4_write(mdp4_kms, REG_MDP4_LCDC_DISPLAY_VEND, display_v_end);
+	mdp4_write(mdp4_kms, REG_MDP4_LCDC_BORDER_CLR, 0);
+	mdp4_write(mdp4_kms, REG_MDP4_LCDC_UNDERFLOW_CLR,
+			MDP4_LCDC_UNDERFLOW_CLR_ENABLE_RECOVERY |
+			MDP4_LCDC_UNDERFLOW_CLR_COLOR(0xff));
+	mdp4_write(mdp4_kms, REG_MDP4_LCDC_HSYNC_SKEW, lcdc_hsync_skew);
+	mdp4_write(mdp4_kms, REG_MDP4_LCDC_CTRL_POLARITY, ctrl_pol);
+	mdp4_write(mdp4_kms, REG_MDP4_LCDC_ACTIVE_HCTL,
+			MDP4_LCDC_ACTIVE_HCTL_START(0) |
+			MDP4_LCDC_ACTIVE_HCTL_END(0));
+	mdp4_write(mdp4_kms, REG_MDP4_LCDC_ACTIVE_VSTART, 0);
+	mdp4_write(mdp4_kms, REG_MDP4_LCDC_ACTIVE_VEND, 0);
+}
+
+static void mdp4_lcdc_encoder_prepare(struct drm_encoder *encoder)
+{
+	mdp4_lcdc_encoder_dpms(encoder, DRM_MODE_DPMS_OFF);
+}
+
+static void mdp4_lcdc_encoder_commit(struct drm_encoder *encoder)
+{
+	/* TODO: hard-coded for 18bpp: */
+	mdp4_crtc_set_config(encoder->crtc,
+			MDP4_DMA_CONFIG_R_BPC(BPC6) |
+			MDP4_DMA_CONFIG_G_BPC(BPC6) |
+			MDP4_DMA_CONFIG_B_BPC(BPC6) |
+			MDP4_DMA_CONFIG_PACK_ALIGN_MSB |
+			MDP4_DMA_CONFIG_PACK(0x21) |
+			MDP4_DMA_CONFIG_DEFLKR_EN |
+			MDP4_DMA_CONFIG_DITHER_EN);
+	mdp4_crtc_set_intf(encoder->crtc, INTF_LCDC_DTV, 0);
+	mdp4_lcdc_encoder_dpms(encoder, DRM_MODE_DPMS_ON);
+}
+
+static const struct drm_encoder_helper_funcs mdp4_lcdc_encoder_helper_funcs = {
+	.dpms = mdp4_lcdc_encoder_dpms,
+	.mode_fixup = mdp4_lcdc_encoder_mode_fixup,
+	.mode_set = mdp4_lcdc_encoder_mode_set,
+	.prepare = mdp4_lcdc_encoder_prepare,
+	.commit = mdp4_lcdc_encoder_commit,
+};
+
+long mdp4_lcdc_round_pixclk(struct drm_encoder *encoder, unsigned long rate)
+{
+	struct mdp4_lcdc_encoder *mdp4_lcdc_encoder =
+			to_mdp4_lcdc_encoder(encoder);
+	return clk_round_rate(mdp4_lcdc_encoder->lcdc_clk, rate);
+}
+
+/* initialize encoder */
+struct drm_encoder *mdp4_lcdc_encoder_init(struct drm_device *dev,
+		struct drm_panel *panel)
+{
+	struct drm_encoder *encoder = NULL;
+	struct mdp4_lcdc_encoder *mdp4_lcdc_encoder;
+	struct regulator *reg;
+	int ret;
+
+	mdp4_lcdc_encoder = kzalloc(sizeof(*mdp4_lcdc_encoder), GFP_KERNEL);
+	if (!mdp4_lcdc_encoder) {
+		ret = -ENOMEM;
+		goto fail;
+	}
+
+	mdp4_lcdc_encoder->panel = panel;
+
+	encoder = &mdp4_lcdc_encoder->base;
+
+	drm_encoder_init(dev, encoder, &mdp4_lcdc_encoder_funcs,
+			 DRM_MODE_ENCODER_LVDS);
+	drm_encoder_helper_add(encoder, &mdp4_lcdc_encoder_helper_funcs);
+
+	/* TODO: do we need different pll in other cases? */
+	mdp4_lcdc_encoder->lcdc_clk = mpd4_lvds_pll_init(dev);
+	if (IS_ERR(mdp4_lcdc_encoder->lcdc_clk)) {
+		dev_err(dev->dev, "failed to get lvds_clk\n");
+		ret = PTR_ERR(mdp4_lcdc_encoder->lcdc_clk);
+		goto fail;
+	}
+
+	/* TODO: different regulators in other cases? */
+	reg = devm_regulator_get(dev->dev, "lvds-vccs-3p3v");
+	if (IS_ERR(reg)) {
+		ret = PTR_ERR(reg);
+		dev_err(dev->dev, "failed to get lvds-vccs-3p3v: %d\n", ret);
+		goto fail;
+	}
+	mdp4_lcdc_encoder->regs[0] = reg;
+
+	reg = devm_regulator_get(dev->dev, "lvds-pll-vdda");
+	if (IS_ERR(reg)) {
+		ret = PTR_ERR(reg);
+		dev_err(dev->dev, "failed to get lvds-pll-vdda: %d\n", ret);
+		goto fail;
+	}
+	mdp4_lcdc_encoder->regs[1] = reg;
+
+	reg = devm_regulator_get(dev->dev, "lvds-vdda");
+	if (IS_ERR(reg)) {
+		ret = PTR_ERR(reg);
+		dev_err(dev->dev, "failed to get lvds-vdda: %d\n", ret);
+		goto fail;
+	}
+	mdp4_lcdc_encoder->regs[2] = reg;
+
+	bs_init(mdp4_lcdc_encoder);
+
+	return encoder;
+
+fail:
+	if (encoder)
+		mdp4_lcdc_encoder_destroy(encoder);
+
+	return ERR_PTR(ret);
+}
diff --git a/drivers/gpu/drm/msm/mdp/mdp4/mdp4_lvds_connector.c b/drivers/gpu/drm/msm/mdp/mdp4/mdp4_lvds_connector.c
new file mode 100644
index 000000000000..4ddc28e1275b
--- /dev/null
+++ b/drivers/gpu/drm/msm/mdp/mdp4/mdp4_lvds_connector.c
@@ -0,0 +1,154 @@
+/*
+ * Copyright (C) 2014 Red Hat
+ * Author: Rob Clark <robdclark@gmail.com>
+ * Author: Vinay Simha <vinaysimha@inforcecomputing.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published by
+ * the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along with
+ * this program.  If not, see <http://www.gnu.org/licenses/>.
+ */
+
+#include <linux/gpio.h>
+
+#include "mdp4_kms.h"
+
+struct mdp4_lvds_connector {
+	struct drm_connector base;
+	struct drm_encoder *encoder;
+	struct drm_panel *panel;
+};
+#define to_mdp4_lvds_connector(x) container_of(x, struct mdp4_lvds_connector, base)
+
+static enum drm_connector_status mdp4_lvds_connector_detect(
+		struct drm_connector *connector, bool force)
+{
+	struct mdp4_lvds_connector *mdp4_lvds_connector =
+			to_mdp4_lvds_connector(connector);
+
+	return mdp4_lvds_connector->panel ?
+			connector_status_connected :
+			connector_status_disconnected;
+}
+
+static void mdp4_lvds_connector_destroy(struct drm_connector *connector)
+{
+	struct mdp4_lvds_connector *mdp4_lvds_connector =
+			to_mdp4_lvds_connector(connector);
+	struct drm_panel *panel = mdp4_lvds_connector->panel;
+
+	if (panel)
+		drm_panel_detach(panel);
+
+	drm_connector_unregister(connector);
+	drm_connector_cleanup(connector);
+
+	kfree(mdp4_lvds_connector);
+}
+
+static int mdp4_lvds_connector_get_modes(struct drm_connector *connector)
+{
+	struct mdp4_lvds_connector *mdp4_lvds_connector =
+			to_mdp4_lvds_connector(connector);
+	struct drm_panel *panel = mdp4_lvds_connector->panel;
+	int ret = 0;
+
+	if (panel)
+		ret = panel->funcs->get_modes(panel);
+
+	return ret;
+}
+
+static int mdp4_lvds_connector_mode_valid(struct drm_connector *connector,
+				 struct drm_display_mode *mode)
+{
+	struct mdp4_lvds_connector *mdp4_lvds_connector =
+			to_mdp4_lvds_connector(connector);
+	struct drm_encoder *encoder = mdp4_lvds_connector->encoder;
+	long actual, requested;
+
+	requested = 1000 * mode->clock;
+	actual = mdp4_lcdc_round_pixclk(encoder, requested);
+
+	DBG("requested=%ld, actual=%ld", requested, actual);
+
+	if (actual != requested)
+		return MODE_CLOCK_RANGE;
+
+	return MODE_OK;
+}
+
+static struct drm_encoder *
+mdp4_lvds_connector_best_encoder(struct drm_connector *connector)
+{
+	struct mdp4_lvds_connector *mdp4_lvds_connector =
+			to_mdp4_lvds_connector(connector);
+	return mdp4_lvds_connector->encoder;
+}
+
+static const struct drm_connector_funcs mdp4_lvds_connector_funcs = {
+	.dpms = drm_helper_connector_dpms,
+	.detect = mdp4_lvds_connector_detect,
+	.fill_modes = drm_helper_probe_single_connector_modes,
+	.destroy = mdp4_lvds_connector_destroy,
+	.reset = drm_atomic_helper_connector_reset,
+	.atomic_duplicate_state = drm_atomic_helper_connector_duplicate_state,
+	.atomic_destroy_state = drm_atomic_helper_connector_destroy_state,
+};
+
+static const struct drm_connector_helper_funcs mdp4_lvds_connector_helper_funcs = {
+	.get_modes = mdp4_lvds_connector_get_modes,
+	.mode_valid = mdp4_lvds_connector_mode_valid,
+	.best_encoder = mdp4_lvds_connector_best_encoder,
+};
+
+/* initialize connector */
+struct drm_connector *mdp4_lvds_connector_init(struct drm_device *dev,
+		struct drm_panel *panel, struct drm_encoder *encoder)
+{
+	struct drm_connector *connector = NULL;
+	struct mdp4_lvds_connector *mdp4_lvds_connector;
+	int ret;
+
+	mdp4_lvds_connector = kzalloc(sizeof(*mdp4_lvds_connector), GFP_KERNEL);
+	if (!mdp4_lvds_connector) {
+		ret = -ENOMEM;
+		goto fail;
+	}
+
+	mdp4_lvds_connector->encoder = encoder;
+	mdp4_lvds_connector->panel = panel;
+
+	connector = &mdp4_lvds_connector->base;
+
+	drm_connector_init(dev, connector, &mdp4_lvds_connector_funcs,
+			DRM_MODE_CONNECTOR_LVDS);
+	drm_connector_helper_add(connector, &mdp4_lvds_connector_helper_funcs);
+
+	connector->polled = 0;
+
+	connector->interlace_allowed = 0;
+	connector->doublescan_allowed = 0;
+
+	drm_connector_register(connector);
+
+	drm_mode_connector_attach_encoder(connector, encoder);
+
+	if (panel)
+		drm_panel_attach(panel, connector);
+
+	return connector;
+
+fail:
+	if (connector)
+		mdp4_lvds_connector_destroy(connector);
+
+	return ERR_PTR(ret);
+}
diff --git a/drivers/gpu/drm/msm/mdp/mdp4/mdp4_lvds_pll.c b/drivers/gpu/drm/msm/mdp/mdp4/mdp4_lvds_pll.c
new file mode 100644
index 000000000000..ce4245971673
--- /dev/null
+++ b/drivers/gpu/drm/msm/mdp/mdp4/mdp4_lvds_pll.c
@@ -0,0 +1,172 @@
+/*
+ * Copyright (C) 2014 Red Hat
+ * Author: Rob Clark <robdclark@gmail.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published by
+ * the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along with
+ * this program.  If not, see <http://www.gnu.org/licenses/>.
+ */
+
+#include <linux/clk.h>
+#include <linux/clk-provider.h>
+
+#include "mdp4_kms.h"
+
+struct mdp4_lvds_pll {
+	struct clk_hw pll_hw;
+	struct drm_device *dev;
+	unsigned long pixclk;
+};
+#define to_mdp4_lvds_pll(x) container_of(x, struct mdp4_lvds_pll, pll_hw)
+
+static struct mdp4_kms *get_kms(struct mdp4_lvds_pll *lvds_pll)
+{
+	struct msm_drm_private *priv = lvds_pll->dev->dev_private;
+	return to_mdp4_kms(to_mdp_kms(priv->kms));
+}
+
+struct pll_rate {
+	unsigned long rate;
+	struct {
+		uint32_t val;
+		uint32_t reg;
+	} conf[32];
+};
+
+/* NOTE: keep sorted highest freq to lowest: */
+static const struct pll_rate freqtbl[] = {
+	{ 72000000, {
+		{ 0x8f, REG_MDP4_LVDS_PHY_PLL_CTRL_1 },
+		{ 0x30, REG_MDP4_LVDS_PHY_PLL_CTRL_2 },
+		{ 0xc6, REG_MDP4_LVDS_PHY_PLL_CTRL_3 },
+		{ 0x10, REG_MDP4_LVDS_PHY_PLL_CTRL_5 },
+		{ 0x07, REG_MDP4_LVDS_PHY_PLL_CTRL_6 },
+		{ 0x62, REG_MDP4_LVDS_PHY_PLL_CTRL_7 },
+		{ 0x41, REG_MDP4_LVDS_PHY_PLL_CTRL_8 },
+		{ 0x0d, REG_MDP4_LVDS_PHY_PLL_CTRL_9 },
+		{ 0, 0 } }
+	},
+};
+
+static const struct pll_rate *find_rate(unsigned long rate)
+{
+	int i;
+	for (i = 1; i < ARRAY_SIZE(freqtbl); i++)
+		if (rate > freqtbl[i].rate)
+			return &freqtbl[i-1];
+	return &freqtbl[i-1];
+}
+
+static int mpd4_lvds_pll_enable(struct clk_hw *hw)
+{
+	struct mdp4_lvds_pll *lvds_pll = to_mdp4_lvds_pll(hw);
+	struct mdp4_kms *mdp4_kms = get_kms(lvds_pll);
+	const struct pll_rate *pll_rate = find_rate(lvds_pll->pixclk);
+	int i;
+
+	DBG("pixclk=%lu (%lu)", lvds_pll->pixclk, pll_rate->rate);
+
+	if (WARN_ON(!pll_rate))
+		return -EINVAL;
+
+	mdp4_write(mdp4_kms, REG_MDP4_LCDC_LVDS_PHY_RESET, 0x33);
+
+	for (i = 0; pll_rate->conf[i].reg; i++)
+		mdp4_write(mdp4_kms, pll_rate->conf[i].reg, pll_rate->conf[i].val);
+
+	mdp4_write(mdp4_kms, REG_MDP4_LVDS_PHY_PLL_CTRL_0, 0x01);
+
+	/* Wait until LVDS PLL is locked and ready */
+	while (!mdp4_read(mdp4_kms, REG_MDP4_LVDS_PHY_PLL_LOCKED))
+		cpu_relax();
+
+	return 0;
+}
+
+static void mpd4_lvds_pll_disable(struct clk_hw *hw)
+{
+	struct mdp4_lvds_pll *lvds_pll = to_mdp4_lvds_pll(hw);
+	struct mdp4_kms *mdp4_kms = get_kms(lvds_pll);
+
+	DBG("");
+
+	mdp4_write(mdp4_kms, REG_MDP4_LVDS_PHY_CFG0, 0x0);
+	mdp4_write(mdp4_kms, REG_MDP4_LVDS_PHY_PLL_CTRL_0, 0x0);
+}
+
+static unsigned long mpd4_lvds_pll_recalc_rate(struct clk_hw *hw,
+				unsigned long parent_rate)
+{
+	struct mdp4_lvds_pll *lvds_pll = to_mdp4_lvds_pll(hw);
+	return lvds_pll->pixclk;
+}
+
+static long mpd4_lvds_pll_round_rate(struct clk_hw *hw, unsigned long rate,
+		unsigned long *parent_rate)
+{
+	const struct pll_rate *pll_rate = find_rate(rate);
+	return pll_rate->rate;
+}
+
+static int mpd4_lvds_pll_set_rate(struct clk_hw *hw, unsigned long rate,
+		unsigned long parent_rate)
+{
+	struct mdp4_lvds_pll *lvds_pll = to_mdp4_lvds_pll(hw);
+	lvds_pll->pixclk = rate;
+	return 0;
+}
+
+
+static const struct clk_ops mpd4_lvds_pll_ops = {
+	.enable = mpd4_lvds_pll_enable,
+	.disable = mpd4_lvds_pll_disable,
+	.recalc_rate = mpd4_lvds_pll_recalc_rate,
+	.round_rate = mpd4_lvds_pll_round_rate,
+	.set_rate = mpd4_lvds_pll_set_rate,
+};
+
+static const char *mpd4_lvds_pll_parents[] = {
+	"pxo",
+};
+
+static struct clk_init_data pll_init = {
+	.name = "mpd4_lvds_pll",
+	.ops = &mpd4_lvds_pll_ops,
+	.parent_names = mpd4_lvds_pll_parents,
+	.num_parents = ARRAY_SIZE(mpd4_lvds_pll_parents),
+};
+
+struct clk *mpd4_lvds_pll_init(struct drm_device *dev)
+{
+	struct mdp4_lvds_pll *lvds_pll;
+	struct clk *clk;
+	int ret;
+
+	lvds_pll = devm_kzalloc(dev->dev, sizeof(*lvds_pll), GFP_KERNEL);
+	if (!lvds_pll) {
+		ret = -ENOMEM;
+		goto fail;
+	}
+
+	lvds_pll->dev = dev;
+
+	lvds_pll->pll_hw.init = &pll_init;
+	clk = devm_clk_register(dev->dev, &lvds_pll->pll_hw);
+	if (IS_ERR(clk)) {
+		ret = PTR_ERR(clk);
+		goto fail;
+	}
+
+	return clk;
+
+fail:
+	return ERR_PTR(ret);
+}
diff --git a/drivers/gpu/drm/msm/mdp/mdp4/mdp4_plane.c b/drivers/gpu/drm/msm/mdp/mdp4/mdp4_plane.c
index 66f33dba1ebb..1e5ebe83647d 100644
--- a/drivers/gpu/drm/msm/mdp/mdp4/mdp4_plane.c
+++ b/drivers/gpu/drm/msm/mdp/mdp4/mdp4_plane.c
@@ -31,47 +31,26 @@ struct mdp4_plane {
 };
 #define to_mdp4_plane(x) container_of(x, struct mdp4_plane, base)
 
-static struct mdp4_kms *get_kms(struct drm_plane *plane)
-{
-	struct msm_drm_private *priv = plane->dev->dev_private;
-	return to_mdp4_kms(to_mdp_kms(priv->kms));
-}
-
-static int mdp4_plane_update(struct drm_plane *plane,
+static void mdp4_plane_set_scanout(struct drm_plane *plane,
+		struct drm_framebuffer *fb);
+static int mdp4_plane_mode_set(struct drm_plane *plane,
 		struct drm_crtc *crtc, struct drm_framebuffer *fb,
 		int crtc_x, int crtc_y,
 		unsigned int crtc_w, unsigned int crtc_h,
 		uint32_t src_x, uint32_t src_y,
-		uint32_t src_w, uint32_t src_h)
-{
-	struct mdp4_plane *mdp4_plane = to_mdp4_plane(plane);
-
-	mdp4_plane->enabled = true;
-
-	if (plane->fb)
-		drm_framebuffer_unreference(plane->fb);
-
-	drm_framebuffer_reference(fb);
-
-	return mdp4_plane_mode_set(plane, crtc, fb,
-			crtc_x, crtc_y, crtc_w, crtc_h,
-			src_x, src_y, src_w, src_h);
-}
+		uint32_t src_w, uint32_t src_h);
 
-static int mdp4_plane_disable(struct drm_plane *plane)
+static struct mdp4_kms *get_kms(struct drm_plane *plane)
 {
-	struct mdp4_plane *mdp4_plane = to_mdp4_plane(plane);
-	DBG("%s: disable", mdp4_plane->name);
-	if (plane->crtc)
-		mdp4_crtc_detach(plane->crtc, plane);
-	return 0;
+	struct msm_drm_private *priv = plane->dev->dev_private;
+	return to_mdp4_kms(to_mdp_kms(priv->kms));
 }
 
 static void mdp4_plane_destroy(struct drm_plane *plane)
 {
 	struct mdp4_plane *mdp4_plane = to_mdp4_plane(plane);
 
-	mdp4_plane_disable(plane);
+	drm_plane_helper_disable(plane);
 	drm_plane_cleanup(plane);
 
 	kfree(mdp4_plane);
@@ -92,19 +71,75 @@ int mdp4_plane_set_property(struct drm_plane *plane,
 }
 
 static const struct drm_plane_funcs mdp4_plane_funcs = {
-		.update_plane = mdp4_plane_update,
-		.disable_plane = mdp4_plane_disable,
+		.update_plane = drm_atomic_helper_update_plane,
+		.disable_plane = drm_atomic_helper_disable_plane,
 		.destroy = mdp4_plane_destroy,
 		.set_property = mdp4_plane_set_property,
+		.reset = drm_atomic_helper_plane_reset,
+		.atomic_duplicate_state = drm_atomic_helper_plane_duplicate_state,
+		.atomic_destroy_state = drm_atomic_helper_plane_destroy_state,
 };
 
-void mdp4_plane_set_scanout(struct drm_plane *plane,
+static int mdp4_plane_prepare_fb(struct drm_plane *plane,
+		struct drm_framebuffer *fb)
+{
+	struct mdp4_plane *mdp4_plane = to_mdp4_plane(plane);
+	struct mdp4_kms *mdp4_kms = get_kms(plane);
+
+	DBG("%s: prepare: FB[%u]", mdp4_plane->name, fb->base.id);
+	return msm_framebuffer_prepare(fb, mdp4_kms->id);
+}
+
+static void mdp4_plane_cleanup_fb(struct drm_plane *plane,
+		struct drm_framebuffer *fb)
+{
+	struct mdp4_plane *mdp4_plane = to_mdp4_plane(plane);
+	struct mdp4_kms *mdp4_kms = get_kms(plane);
+
+	DBG("%s: cleanup: FB[%u]", mdp4_plane->name, fb->base.id);
+	msm_framebuffer_cleanup(fb, mdp4_kms->id);
+}
+
+
+static int mdp4_plane_atomic_check(struct drm_plane *plane,
+		struct drm_plane_state *state)
+{
+	return 0;
+}
+
+static void mdp4_plane_atomic_update(struct drm_plane *plane,
+				     struct drm_plane_state *old_state)
+{
+	struct drm_plane_state *state = plane->state;
+	int ret;
+
+	ret = mdp4_plane_mode_set(plane,
+			state->crtc, state->fb,
+			state->crtc_x, state->crtc_y,
+			state->crtc_w, state->crtc_h,
+			state->src_x,  state->src_y,
+			state->src_w, state->src_h);
+	/* atomic_check should have ensured that this doesn't fail */
+	WARN_ON(ret < 0);
+}
+
+static const struct drm_plane_helper_funcs mdp4_plane_helper_funcs = {
+		.prepare_fb = mdp4_plane_prepare_fb,
+		.cleanup_fb = mdp4_plane_cleanup_fb,
+		.atomic_check = mdp4_plane_atomic_check,
+		.atomic_update = mdp4_plane_atomic_update,
+};
+
+static void mdp4_plane_set_scanout(struct drm_plane *plane,
 		struct drm_framebuffer *fb)
 {
 	struct mdp4_plane *mdp4_plane = to_mdp4_plane(plane);
 	struct mdp4_kms *mdp4_kms = get_kms(plane);
 	enum mdp4_pipe pipe = mdp4_plane->pipe;
-	uint32_t iova;
+	uint32_t iova = msm_framebuffer_iova(fb, mdp4_kms->id, 0);
+
+	DBG("%s: set_scanout: %08x (%u)", mdp4_plane->name,
+			iova, fb->pitches[0]);
 
 	mdp4_write(mdp4_kms, REG_MDP4_PIPE_SRC_STRIDE_A(pipe),
 			MDP4_PIPE_SRC_STRIDE_A_P0(fb->pitches[0]) |
@@ -114,7 +149,6 @@ void mdp4_plane_set_scanout(struct drm_plane *plane,
 			MDP4_PIPE_SRC_STRIDE_B_P2(fb->pitches[2]) |
 			MDP4_PIPE_SRC_STRIDE_B_P3(fb->pitches[3]));
 
-	msm_gem_get_iova(msm_framebuffer_bo(fb, 0), mdp4_kms->id, &iova);
 	mdp4_write(mdp4_kms, REG_MDP4_PIPE_SRCP0_BASE(pipe), iova);
 
 	plane->fb = fb;
@@ -122,7 +156,7 @@ void mdp4_plane_set_scanout(struct drm_plane *plane,
 
 #define MDP4_VG_PHASE_STEP_DEFAULT	0x20000000
 
-int mdp4_plane_mode_set(struct drm_plane *plane,
+static int mdp4_plane_mode_set(struct drm_plane *plane,
 		struct drm_crtc *crtc, struct drm_framebuffer *fb,
 		int crtc_x, int crtc_y,
 		unsigned int crtc_w, unsigned int crtc_h,
@@ -137,6 +171,11 @@ int mdp4_plane_mode_set(struct drm_plane *plane,
 	uint32_t phasex_step = MDP4_VG_PHASE_STEP_DEFAULT;
 	uint32_t phasey_step = MDP4_VG_PHASE_STEP_DEFAULT;
 
+	if (!(crtc && fb)) {
+		DBG("%s: disabled!", mdp4_plane->name);
+		return 0;
+	}
+
 	/* src values are in Q16 fixed point, convert to integer: */
 	src_x = src_x >> 16;
 	src_y = src_y >> 16;
@@ -197,9 +236,6 @@ int mdp4_plane_mode_set(struct drm_plane *plane,
 	mdp4_write(mdp4_kms, REG_MDP4_PIPE_PHASEX_STEP(pipe), phasex_step);
 	mdp4_write(mdp4_kms, REG_MDP4_PIPE_PHASEY_STEP(pipe), phasey_step);
 
-	/* TODO detach from old crtc (if we had more than one) */
-	mdp4_crtc_attach(crtc, plane);
-
 	return 0;
 }
 
@@ -239,9 +275,12 @@ struct drm_plane *mdp4_plane_init(struct drm_device *dev,
 			ARRAY_SIZE(mdp4_plane->formats));
 
 	type = private_plane ? DRM_PLANE_TYPE_PRIMARY : DRM_PLANE_TYPE_OVERLAY;
-	drm_universal_plane_init(dev, plane, 0xff, &mdp4_plane_funcs,
-				 mdp4_plane->formats, mdp4_plane->nformats,
-				 type);
+	ret = drm_universal_plane_init(dev, plane, 0xff, &mdp4_plane_funcs,
+				 mdp4_plane->formats, mdp4_plane->nformats, type);
+	if (ret)
+		goto fail;
+
+	drm_plane_helper_add(plane, &mdp4_plane_helper_funcs);
 
 	mdp4_plane_install_properties(plane, &plane->base);
 
diff --git a/drivers/gpu/drm/msm/mdp/mdp5/mdp5.xml.h b/drivers/gpu/drm/msm/mdp/mdp5/mdp5.xml.h
index 67f4f896ba8c..e87ef5512cb0 100644
--- a/drivers/gpu/drm/msm/mdp/mdp5/mdp5.xml.h
+++ b/drivers/gpu/drm/msm/mdp/mdp5/mdp5.xml.h
@@ -10,14 +10,14 @@ git clone https://github.com/freedreno/envytools.git
 The rules-ng-ng source files this header was generated from are:
 - /home/robclark/src/freedreno/envytools/rnndb/msm.xml                 (    647 bytes, from 2013-11-30 14:45:35)
 - /home/robclark/src/freedreno/envytools/rnndb/freedreno_copyright.xml (   1453 bytes, from 2013-03-31 16:51:27)
-- /home/robclark/src/freedreno/envytools/rnndb/mdp/mdp4.xml            (  17996 bytes, from 2013-12-01 19:10:31)
-- /home/robclark/src/freedreno/envytools/rnndb/mdp/mdp_common.xml      (   1615 bytes, from 2013-11-30 15:00:52)
-- /home/robclark/src/freedreno/envytools/rnndb/mdp/mdp5.xml            (  22517 bytes, from 2014-06-25 12:55:02)
+- /home/robclark/src/freedreno/envytools/rnndb/mdp/mdp4.xml            (  20136 bytes, from 2014-10-31 16:51:39)
+- /home/robclark/src/freedreno/envytools/rnndb/mdp/mdp_common.xml      (   1940 bytes, from 2014-10-31 16:51:39)
+- /home/robclark/src/freedreno/envytools/rnndb/mdp/mdp5.xml            (  23963 bytes, from 2014-10-31 16:51:46)
 - /home/robclark/src/freedreno/envytools/rnndb/dsi/dsi.xml             (  11712 bytes, from 2013-08-17 17:13:43)
 - /home/robclark/src/freedreno/envytools/rnndb/dsi/sfpb.xml            (    344 bytes, from 2013-08-11 19:26:32)
-- /home/robclark/src/freedreno/envytools/rnndb/dsi/mmss_cc.xml         (   1544 bytes, from 2013-08-16 19:17:05)
+- /home/robclark/src/freedreno/envytools/rnndb/dsi/mmss_cc.xml         (   1686 bytes, from 2014-10-31 16:48:57)
 - /home/robclark/src/freedreno/envytools/rnndb/hdmi/qfprom.xml         (    600 bytes, from 2013-07-05 19:21:12)
-- /home/robclark/src/freedreno/envytools/rnndb/hdmi/hdmi.xml           (  23613 bytes, from 2014-06-25 12:53:44)
+- /home/robclark/src/freedreno/envytools/rnndb/hdmi/hdmi.xml           (  23613 bytes, from 2014-07-17 15:33:30)
 
 Copyright (C) 2013-2014 by the following authors:
 - Rob Clark <robdclark@gmail.com> (robclark)
diff --git a/drivers/gpu/drm/msm/mdp/mdp5/mdp5_cfg.c b/drivers/gpu/drm/msm/mdp/mdp5/mdp5_cfg.c
new file mode 100644
index 000000000000..b0a44310cf2a
--- /dev/null
+++ b/drivers/gpu/drm/msm/mdp/mdp5/mdp5_cfg.c
@@ -0,0 +1,207 @@
+/*
+ * Copyright (c) 2014 The Linux Foundation. All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 and
+ * only version 2 as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ */
+
+#include "mdp5_kms.h"
+#include "mdp5_cfg.h"
+
+struct mdp5_cfg_handler {
+	int revision;
+	struct mdp5_cfg config;
+};
+
+/* mdp5_cfg must be exposed (used in mdp5.xml.h) */
+const struct mdp5_cfg_hw *mdp5_cfg = NULL;
+
+const struct mdp5_cfg_hw msm8x74_config = {
+	.name = "msm8x74",
+	.smp = {
+		.mmb_count = 22,
+		.mmb_size = 4096,
+	},
+	.ctl = {
+		.count = 5,
+		.base = { 0x00600, 0x00700, 0x00800, 0x00900, 0x00a00 },
+	},
+	.pipe_vig = {
+		.count = 3,
+		.base = { 0x01200, 0x01600, 0x01a00 },
+	},
+	.pipe_rgb = {
+		.count = 3,
+		.base = { 0x01e00, 0x02200, 0x02600 },
+	},
+	.pipe_dma = {
+		.count = 2,
+		.base = { 0x02a00, 0x02e00 },
+	},
+	.lm = {
+		.count = 5,
+		.base = { 0x03200, 0x03600, 0x03a00, 0x03e00, 0x04200 },
+		.nb_stages = 5,
+	},
+	.dspp = {
+		.count = 3,
+		.base = { 0x04600, 0x04a00, 0x04e00 },
+	},
+	.ad = {
+		.count = 2,
+		.base = { 0x13100, 0x13300 }, /* NOTE: no ad in v1.0 */
+	},
+	.intf = {
+		.count = 4,
+		.base = { 0x12500, 0x12700, 0x12900, 0x12b00 },
+	},
+	.max_clk = 200000000,
+};
+
+const struct mdp5_cfg_hw apq8084_config = {
+	.name = "apq8084",
+	.smp = {
+		.mmb_count = 44,
+		.mmb_size = 8192,
+		.reserved_state[0] = GENMASK(7, 0),	/* first 8 MMBs */
+		.reserved[CID_RGB0] = 2,
+		.reserved[CID_RGB1] = 2,
+		.reserved[CID_RGB2] = 2,
+		.reserved[CID_RGB3] = 2,
+	},
+	.ctl = {
+		.count = 5,
+		.base = { 0x00600, 0x00700, 0x00800, 0x00900, 0x00a00 },
+	},
+	.pipe_vig = {
+		.count = 4,
+		.base = { 0x01200, 0x01600, 0x01a00, 0x01e00 },
+	},
+	.pipe_rgb = {
+		.count = 4,
+		.base = { 0x02200, 0x02600, 0x02a00, 0x02e00 },
+	},
+	.pipe_dma = {
+		.count = 2,
+		.base = { 0x03200, 0x03600 },
+	},
+	.lm = {
+		.count = 6,
+		.base = { 0x03a00, 0x03e00, 0x04200, 0x04600, 0x04a00, 0x04e00 },
+		.nb_stages = 5,
+	},
+	.dspp = {
+		.count = 4,
+		.base = { 0x05200, 0x05600, 0x05a00, 0x05e00 },
+
+	},
+	.ad = {
+		.count = 3,
+		.base = { 0x13500, 0x13700, 0x13900 },
+	},
+	.intf = {
+		.count = 5,
+		.base = { 0x12500, 0x12700, 0x12900, 0x12b00, 0x12d00 },
+	},
+	.max_clk = 320000000,
+};
+
+static const struct mdp5_cfg_handler cfg_handlers[] = {
+	{ .revision = 0, .config = { .hw = &msm8x74_config } },
+	{ .revision = 2, .config = { .hw = &msm8x74_config } },
+	{ .revision = 3, .config = { .hw = &apq8084_config } },
+};
+
+
+static struct mdp5_cfg_platform *mdp5_get_config(struct platform_device *dev);
+
+const struct mdp5_cfg_hw *mdp5_cfg_get_hw_config(struct mdp5_cfg_handler *cfg_handler)
+{
+	return cfg_handler->config.hw;
+}
+
+struct mdp5_cfg *mdp5_cfg_get_config(struct mdp5_cfg_handler *cfg_handler)
+{
+	return &cfg_handler->config;
+}
+
+int mdp5_cfg_get_hw_rev(struct mdp5_cfg_handler *cfg_handler)
+{
+	return cfg_handler->revision;
+}
+
+void mdp5_cfg_destroy(struct mdp5_cfg_handler *cfg_handler)
+{
+	kfree(cfg_handler);
+}
+
+struct mdp5_cfg_handler *mdp5_cfg_init(struct mdp5_kms *mdp5_kms,
+		uint32_t major, uint32_t minor)
+{
+	struct drm_device *dev = mdp5_kms->dev;
+	struct platform_device *pdev = dev->platformdev;
+	struct mdp5_cfg_handler *cfg_handler;
+	struct mdp5_cfg_platform *pconfig;
+	int i, ret = 0;
+
+	cfg_handler = kzalloc(sizeof(*cfg_handler), GFP_KERNEL);
+	if (unlikely(!cfg_handler)) {
+		ret = -ENOMEM;
+		goto fail;
+	}
+
+	if (major != 1) {
+		dev_err(dev->dev, "unexpected MDP major version: v%d.%d\n",
+				major, minor);
+		ret = -ENXIO;
+		goto fail;
+	}
+
+	/* only after mdp5_cfg global pointer's init can we access the hw */
+	for (i = 0; i < ARRAY_SIZE(cfg_handlers); i++) {
+		if (cfg_handlers[i].revision != minor)
+			continue;
+		mdp5_cfg = cfg_handlers[i].config.hw;
+
+		break;
+	}
+	if (unlikely(!mdp5_cfg)) {
+		dev_err(dev->dev, "unexpected MDP minor revision: v%d.%d\n",
+				major, minor);
+		ret = -ENXIO;
+		goto fail;
+	}
+
+	cfg_handler->revision = minor;
+	cfg_handler->config.hw = mdp5_cfg;
+
+	pconfig = mdp5_get_config(pdev);
+	memcpy(&cfg_handler->config.platform, pconfig, sizeof(*pconfig));
+
+	DBG("MDP5: %s hw config selected", mdp5_cfg->name);
+
+	return cfg_handler;
+
+fail:
+	if (cfg_handler)
+		mdp5_cfg_destroy(cfg_handler);
+
+	return NULL;
+}
+
+static struct mdp5_cfg_platform *mdp5_get_config(struct platform_device *dev)
+{
+	static struct mdp5_cfg_platform config = {};
+#ifdef CONFIG_OF
+	/* TODO */
+#endif
+	config.iommu = iommu_domain_alloc(&platform_bus_type);
+
+	return &config;
+}
diff --git a/drivers/gpu/drm/msm/mdp/mdp5/mdp5_cfg.h b/drivers/gpu/drm/msm/mdp/mdp5/mdp5_cfg.h
new file mode 100644
index 000000000000..dba4d52cceeb
--- /dev/null
+++ b/drivers/gpu/drm/msm/mdp/mdp5/mdp5_cfg.h
@@ -0,0 +1,91 @@
+/*
+ * Copyright (c) 2014 The Linux Foundation. All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 and
+ * only version 2 as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ */
+
+#ifndef __MDP5_CFG_H__
+#define __MDP5_CFG_H__
+
+#include "msm_drv.h"
+
+/*
+ * mdp5_cfg
+ *
+ * This module configures the dynamic offsets used by mdp5.xml.h
+ * (initialized in mdp5_cfg.c)
+ */
+extern const struct mdp5_cfg_hw *mdp5_cfg;
+
+#define MAX_CTL			8
+#define MAX_BASES		8
+#define MAX_SMP_BLOCKS		44
+#define MAX_CLIENTS		32
+
+typedef DECLARE_BITMAP(mdp5_smp_state_t, MAX_SMP_BLOCKS);
+
+#define MDP5_SUB_BLOCK_DEFINITION \
+	int count; \
+	uint32_t base[MAX_BASES]
+
+struct mdp5_sub_block {
+	MDP5_SUB_BLOCK_DEFINITION;
+};
+
+struct mdp5_lm_block {
+	MDP5_SUB_BLOCK_DEFINITION;
+	uint32_t nb_stages;		/* number of stages per blender */
+};
+
+struct mdp5_smp_block {
+	int mmb_count;			/* number of SMP MMBs */
+	int mmb_size;			/* MMB: size in bytes */
+	mdp5_smp_state_t reserved_state;/* SMP MMBs statically allocated */
+	int reserved[MAX_CLIENTS];	/* # of MMBs allocated per client */
+};
+
+struct mdp5_cfg_hw {
+	char  *name;
+
+	struct mdp5_smp_block smp;
+	struct mdp5_sub_block ctl;
+	struct mdp5_sub_block pipe_vig;
+	struct mdp5_sub_block pipe_rgb;
+	struct mdp5_sub_block pipe_dma;
+	struct mdp5_lm_block  lm;
+	struct mdp5_sub_block dspp;
+	struct mdp5_sub_block ad;
+	struct mdp5_sub_block intf;
+
+	uint32_t max_clk;
+};
+
+/* platform config data (ie. from DT, or pdata) */
+struct mdp5_cfg_platform {
+	struct iommu_domain *iommu;
+};
+
+struct mdp5_cfg {
+	const struct mdp5_cfg_hw *hw;
+	struct mdp5_cfg_platform platform;
+};
+
+struct mdp5_kms;
+struct mdp5_cfg_handler;
+
+const struct mdp5_cfg_hw *mdp5_cfg_get_hw_config(struct mdp5_cfg_handler *cfg_hnd);
+struct mdp5_cfg *mdp5_cfg_get_config(struct mdp5_cfg_handler *cfg_hnd);
+int mdp5_cfg_get_hw_rev(struct mdp5_cfg_handler *cfg_hnd);
+
+struct mdp5_cfg_handler *mdp5_cfg_init(struct mdp5_kms *mdp5_kms,
+		uint32_t major, uint32_t minor);
+void mdp5_cfg_destroy(struct mdp5_cfg_handler *cfg_hnd);
+
+#endif /* __MDP5_CFG_H__ */
diff --git a/drivers/gpu/drm/msm/mdp/mdp5/mdp5_crtc.c b/drivers/gpu/drm/msm/mdp/mdp5/mdp5_crtc.c
index ebe2e60f3ab1..f021f960a8a2 100644
--- a/drivers/gpu/drm/msm/mdp/mdp5/mdp5_crtc.c
+++ b/drivers/gpu/drm/msm/mdp/mdp5/mdp5_crtc.c
@@ -1,4 +1,5 @@
 /*
+ * Copyright (c) 2014 The Linux Foundation. All rights reserved.
  * Copyright (C) 2013 Red Hat
  * Author: Rob Clark <robdclark@gmail.com>
  *
@@ -17,43 +18,35 @@
 
 #include "mdp5_kms.h"
 
+#include <linux/sort.h>
 #include <drm/drm_mode.h>
 #include "drm_crtc.h"
 #include "drm_crtc_helper.h"
 #include "drm_flip_work.h"
 
+#define SSPP_MAX	(SSPP_RGB3 + 1) /* TODO: Add SSPP_MAX in mdp5.xml.h */
+
 struct mdp5_crtc {
 	struct drm_crtc base;
 	char name[8];
-	struct drm_plane *plane;
-	struct drm_plane *planes[8];
 	int id;
 	bool enabled;
 
-	/* which mixer/encoder we route output to: */
-	int mixer;
+	/* layer mixer used for this CRTC (+ its lock): */
+#define GET_LM_ID(crtc_id)	((crtc_id == 3) ? 5 : crtc_id)
+	int lm;
+	spinlock_t lm_lock;	/* protect REG_MDP5_LM_* registers */
+
+	/* CTL used for this CRTC: */
+	struct mdp5_ctl *ctl;
 
 	/* if there is a pending flip, these will be non-null: */
 	struct drm_pending_vblank_event *event;
-	struct msm_fence_cb pageflip_cb;
 
 #define PENDING_CURSOR 0x1
 #define PENDING_FLIP   0x2
 	atomic_t pending;
 
-	/* the fb that we logically (from PoV of KMS API) hold a ref
-	 * to.  Which we may not yet be scanning out (we may still
-	 * be scanning out previous in case of page_flip while waiting
-	 * for gpu rendering to complete:
-	 */
-	struct drm_framebuffer *fb;
-
-	/* the fb that we currently hold a scanout ref to: */
-	struct drm_framebuffer *scanout_fb;
-
-	/* for unref'ing framebuffers after scanout completes: */
-	struct drm_flip_work unref_fb_work;
-
 	struct mdp_irq vblank;
 	struct mdp_irq err;
 };
@@ -73,67 +66,38 @@ static void request_pending(struct drm_crtc *crtc, uint32_t pending)
 	mdp_irq_register(&get_kms(crtc)->base, &mdp5_crtc->vblank);
 }
 
-static void crtc_flush(struct drm_crtc *crtc)
-{
-	struct mdp5_crtc *mdp5_crtc = to_mdp5_crtc(crtc);
-	struct mdp5_kms *mdp5_kms = get_kms(crtc);
-	int id = mdp5_crtc->id;
-	uint32_t i, flush = 0;
-
-	for (i = 0; i < ARRAY_SIZE(mdp5_crtc->planes); i++) {
-		struct drm_plane *plane = mdp5_crtc->planes[i];
-		if (plane) {
-			enum mdp5_pipe pipe = mdp5_plane_pipe(plane);
-			flush |= pipe2flush(pipe);
-		}
-	}
-	flush |= mixer2flush(mdp5_crtc->id);
-	flush |= MDP5_CTL_FLUSH_CTL;
-
-	DBG("%s: flush=%08x", mdp5_crtc->name, flush);
-
-	mdp5_write(mdp5_kms, REG_MDP5_CTL_FLUSH(id), flush);
-}
+#define mdp5_lm_get_flush(lm)	mdp_ctl_flush_mask_lm(lm)
 
-static void update_fb(struct drm_crtc *crtc, struct drm_framebuffer *new_fb)
+static void crtc_flush(struct drm_crtc *crtc, u32 flush_mask)
 {
 	struct mdp5_crtc *mdp5_crtc = to_mdp5_crtc(crtc);
-	struct drm_framebuffer *old_fb = mdp5_crtc->fb;
 
-	/* grab reference to incoming scanout fb: */
-	drm_framebuffer_reference(new_fb);
-	mdp5_crtc->base.primary->fb = new_fb;
-	mdp5_crtc->fb = new_fb;
-
-	if (old_fb)
-		drm_flip_work_queue(&mdp5_crtc->unref_fb_work, old_fb);
+	DBG("%s: flush=%08x", mdp5_crtc->name, flush_mask);
+	mdp5_ctl_commit(mdp5_crtc->ctl, flush_mask);
 }
 
-/* unlike update_fb(), take a ref to the new scanout fb *before* updating
- * plane, then call this.  Needed to ensure we don't unref the buffer that
- * is actually still being scanned out.
- *
- * Note that this whole thing goes away with atomic.. since we can defer
- * calling into driver until rendering is done.
+/*
+ * flush updates, to make sure hw is updated to new scanout fb,
+ * so that we can safely queue unref to current fb (ie. next
+ * vblank we know hw is done w/ previous scanout_fb).
  */
-static void update_scanout(struct drm_crtc *crtc, struct drm_framebuffer *fb)
+static void crtc_flush_all(struct drm_crtc *crtc)
 {
 	struct mdp5_crtc *mdp5_crtc = to_mdp5_crtc(crtc);
+	struct drm_plane *plane;
+	uint32_t flush_mask = 0;
 
-	/* flush updates, to make sure hw is updated to new scanout fb,
-	 * so that we can safely queue unref to current fb (ie. next
-	 * vblank we know hw is done w/ previous scanout_fb).
-	 */
-	crtc_flush(crtc);
-
-	if (mdp5_crtc->scanout_fb)
-		drm_flip_work_queue(&mdp5_crtc->unref_fb_work,
-				mdp5_crtc->scanout_fb);
+	/* we could have already released CTL in the disable path: */
+	if (!mdp5_crtc->ctl)
+		return;
 
-	mdp5_crtc->scanout_fb = fb;
+	drm_atomic_crtc_for_each_plane(plane, crtc) {
+		flush_mask |= mdp5_plane_get_flush(plane);
+	}
+	flush_mask |= mdp5_ctl_get_flush(mdp5_crtc->ctl);
+	flush_mask |= mdp5_lm_get_flush(mdp5_crtc->lm);
 
-	/* enable vblank to complete flip: */
-	request_pending(crtc, PENDING_FLIP);
+	crtc_flush(crtc, flush_mask);
 }
 
 /* if file!=NULL, this is preclose potential cancel-flip path */
@@ -142,7 +106,8 @@ static void complete_flip(struct drm_crtc *crtc, struct drm_file *file)
 	struct mdp5_crtc *mdp5_crtc = to_mdp5_crtc(crtc);
 	struct drm_device *dev = crtc->dev;
 	struct drm_pending_vblank_event *event;
-	unsigned long flags, i;
+	struct drm_plane *plane;
+	unsigned long flags;
 
 	spin_lock_irqsave(&dev->event_lock, flags);
 	event = mdp5_crtc->event;
@@ -153,50 +118,22 @@ static void complete_flip(struct drm_crtc *crtc, struct drm_file *file)
 		 */
 		if (!file || (event->base.file_priv == file)) {
 			mdp5_crtc->event = NULL;
+			DBG("%s: send event: %p", mdp5_crtc->name, event);
 			drm_send_vblank_event(dev, mdp5_crtc->id, event);
 		}
 	}
 	spin_unlock_irqrestore(&dev->event_lock, flags);
 
-	for (i = 0; i < ARRAY_SIZE(mdp5_crtc->planes); i++) {
-		struct drm_plane *plane = mdp5_crtc->planes[i];
-		if (plane)
-			mdp5_plane_complete_flip(plane);
+	drm_atomic_crtc_for_each_plane(plane, crtc) {
+		mdp5_plane_complete_flip(plane);
 	}
 }
 
-static void pageflip_cb(struct msm_fence_cb *cb)
-{
-	struct mdp5_crtc *mdp5_crtc =
-		container_of(cb, struct mdp5_crtc, pageflip_cb);
-	struct drm_crtc *crtc = &mdp5_crtc->base;
-	struct drm_framebuffer *fb = mdp5_crtc->fb;
-
-	if (!fb)
-		return;
-
-	drm_framebuffer_reference(fb);
-	mdp5_plane_set_scanout(mdp5_crtc->plane, fb);
-	update_scanout(crtc, fb);
-}
-
-static void unref_fb_worker(struct drm_flip_work *work, void *val)
-{
-	struct mdp5_crtc *mdp5_crtc =
-		container_of(work, struct mdp5_crtc, unref_fb_work);
-	struct drm_device *dev = mdp5_crtc->base.dev;
-
-	mutex_lock(&dev->mode_config.mutex);
-	drm_framebuffer_unreference(val);
-	mutex_unlock(&dev->mode_config.mutex);
-}
-
 static void mdp5_crtc_destroy(struct drm_crtc *crtc)
 {
 	struct mdp5_crtc *mdp5_crtc = to_mdp5_crtc(crtc);
 
 	drm_crtc_cleanup(crtc);
-	drm_flip_work_cleanup(&mdp5_crtc->unref_fb_work);
 
 	kfree(mdp5_crtc);
 }
@@ -214,6 +151,8 @@ static void mdp5_crtc_dpms(struct drm_crtc *crtc, int mode)
 			mdp5_enable(mdp5_kms);
 			mdp_irq_register(&mdp5_kms->base, &mdp5_crtc->err);
 		} else {
+			/* set STAGE_UNUSED for all layers */
+			mdp5_ctl_blend(mdp5_crtc->ctl, mdp5_crtc->lm, 0x00000000);
 			mdp_irq_unregister(&mdp5_kms->base, &mdp5_crtc->err);
 			mdp5_disable(mdp5_kms);
 		}
@@ -228,54 +167,78 @@ static bool mdp5_crtc_mode_fixup(struct drm_crtc *crtc,
 	return true;
 }
 
+/*
+ * blend_setup() - blend all the planes of a CRTC
+ *
+ * When border is enabled, the border color will ALWAYS be the base layer.
+ * Therefore, the first plane (private RGB pipe) will start at STAGE0.
+ * If disabled, the first plane starts at STAGE_BASE.
+ *
+ * Note:
+ * Border is not enabled here because the private plane is exactly
+ * the CRTC resolution.
+ */
 static void blend_setup(struct drm_crtc *crtc)
 {
 	struct mdp5_crtc *mdp5_crtc = to_mdp5_crtc(crtc);
 	struct mdp5_kms *mdp5_kms = get_kms(crtc);
-	int id = mdp5_crtc->id;
+	struct drm_plane *plane;
+	const struct mdp5_cfg_hw *hw_cfg;
+	uint32_t lm = mdp5_crtc->lm, blend_cfg = 0;
+	unsigned long flags;
+#define blender(stage)	((stage) - STAGE_BASE)
 
-	/*
-	 * Hard-coded setup for now until I figure out how the
-	 * layer-mixer works
-	 */
+	hw_cfg = mdp5_cfg_get_hw_config(mdp5_kms->cfg);
 
-	/* LM[id]: */
-	mdp5_write(mdp5_kms, REG_MDP5_LM_BLEND_COLOR_OUT(id),
-			MDP5_LM_BLEND_COLOR_OUT_STAGE0_FG_ALPHA);
-	mdp5_write(mdp5_kms, REG_MDP5_LM_BLEND_OP_MODE(id, 0),
-			MDP5_LM_BLEND_OP_MODE_FG_ALPHA(FG_CONST) |
-			MDP5_LM_BLEND_OP_MODE_BG_ALPHA(FG_PIXEL) |
-			MDP5_LM_BLEND_OP_MODE_BG_INV_ALPHA);
-	mdp5_write(mdp5_kms, REG_MDP5_LM_BLEND_FG_ALPHA(id, 0), 0xff);
-	mdp5_write(mdp5_kms, REG_MDP5_LM_BLEND_BG_ALPHA(id, 0), 0x00);
-
-	/* NOTE: seems that LM[n] and CTL[m], we do not need n==m.. but
-	 * we want to be setting CTL[m].LAYER[n].  Not sure what the
-	 * point of having CTL[m].LAYER[o] (for o!=n).. maybe that is
-	 * used when chaining up mixers for high resolution displays?
-	 */
+	spin_lock_irqsave(&mdp5_crtc->lm_lock, flags);
 
-	/* CTL[id]: */
-	mdp5_write(mdp5_kms, REG_MDP5_CTL_LAYER_REG(id, 0),
-			MDP5_CTL_LAYER_REG_RGB0(STAGE0) |
-			MDP5_CTL_LAYER_REG_BORDER_COLOR);
-	mdp5_write(mdp5_kms, REG_MDP5_CTL_LAYER_REG(id, 1), 0);
-	mdp5_write(mdp5_kms, REG_MDP5_CTL_LAYER_REG(id, 2), 0);
-	mdp5_write(mdp5_kms, REG_MDP5_CTL_LAYER_REG(id, 3), 0);
-	mdp5_write(mdp5_kms, REG_MDP5_CTL_LAYER_REG(id, 4), 0);
+	/* ctl could be released already when we are shutting down: */
+	if (!mdp5_crtc->ctl)
+		goto out;
+
+	drm_atomic_crtc_for_each_plane(plane, crtc) {
+		enum mdp_mixer_stage_id stage =
+			to_mdp5_plane_state(plane->state)->stage;
+
+		/*
+		 * Note: This cannot happen with current implementation but
+		 * we need to check this condition once z property is added
+		 */
+		BUG_ON(stage > hw_cfg->lm.nb_stages);
+
+		/* LM */
+		mdp5_write(mdp5_kms,
+				REG_MDP5_LM_BLEND_OP_MODE(lm, blender(stage)),
+				MDP5_LM_BLEND_OP_MODE_FG_ALPHA(FG_CONST) |
+				MDP5_LM_BLEND_OP_MODE_BG_ALPHA(BG_CONST));
+		mdp5_write(mdp5_kms, REG_MDP5_LM_BLEND_FG_ALPHA(lm,
+				blender(stage)), 0xff);
+		mdp5_write(mdp5_kms, REG_MDP5_LM_BLEND_BG_ALPHA(lm,
+				blender(stage)), 0x00);
+		/* CTL */
+		blend_cfg |= mdp_ctl_blend_mask(mdp5_plane_pipe(plane), stage);
+		DBG("%s: blending pipe %s on stage=%d", mdp5_crtc->name,
+				pipe2name(mdp5_plane_pipe(plane)), stage);
+	}
+
+	DBG("%s: lm%d: blend config = 0x%08x", mdp5_crtc->name, lm, blend_cfg);
+	mdp5_ctl_blend(mdp5_crtc->ctl, lm, blend_cfg);
+
+out:
+	spin_unlock_irqrestore(&mdp5_crtc->lm_lock, flags);
 }
 
-static int mdp5_crtc_mode_set(struct drm_crtc *crtc,
-		struct drm_display_mode *mode,
-		struct drm_display_mode *adjusted_mode,
-		int x, int y,
-		struct drm_framebuffer *old_fb)
+static void mdp5_crtc_mode_set_nofb(struct drm_crtc *crtc)
 {
 	struct mdp5_crtc *mdp5_crtc = to_mdp5_crtc(crtc);
 	struct mdp5_kms *mdp5_kms = get_kms(crtc);
-	int ret;
+	unsigned long flags;
+	struct drm_display_mode *mode;
 
-	mode = adjusted_mode;
+	if (WARN_ON(!crtc->state))
+		return;
+
+	mode = &crtc->state->adjusted_mode;
 
 	DBG("%s: set mode: %d:\"%s\" %d %d %d %d %d %d %d %d %d %d 0x%x 0x%x",
 			mdp5_crtc->name, mode->base.id, mode->name,
@@ -286,28 +249,11 @@ static int mdp5_crtc_mode_set(struct drm_crtc *crtc,
 			mode->vsync_end, mode->vtotal,
 			mode->type, mode->flags);
 
-	/* grab extra ref for update_scanout() */
-	drm_framebuffer_reference(crtc->primary->fb);
-
-	ret = mdp5_plane_mode_set(mdp5_crtc->plane, crtc, crtc->primary->fb,
-			0, 0, mode->hdisplay, mode->vdisplay,
-			x << 16, y << 16,
-			mode->hdisplay << 16, mode->vdisplay << 16);
-	if (ret) {
-		drm_framebuffer_unreference(crtc->primary->fb);
-		dev_err(crtc->dev->dev, "%s: failed to set mode on plane: %d\n",
-				mdp5_crtc->name, ret);
-		return ret;
-	}
-
-	mdp5_write(mdp5_kms, REG_MDP5_LM_OUT_SIZE(mdp5_crtc->id),
+	spin_lock_irqsave(&mdp5_crtc->lm_lock, flags);
+	mdp5_write(mdp5_kms, REG_MDP5_LM_OUT_SIZE(mdp5_crtc->lm),
 			MDP5_LM_OUT_SIZE_WIDTH(mode->hdisplay) |
 			MDP5_LM_OUT_SIZE_HEIGHT(mode->vdisplay));
-
-	update_fb(crtc, crtc->primary->fb);
-	update_scanout(crtc, crtc->primary->fb);
-
-	return 0;
+	spin_unlock_irqrestore(&mdp5_crtc->lm_lock, flags);
 }
 
 static void mdp5_crtc_prepare(struct drm_crtc *crtc)
@@ -321,66 +267,114 @@ static void mdp5_crtc_prepare(struct drm_crtc *crtc)
 
 static void mdp5_crtc_commit(struct drm_crtc *crtc)
 {
+	struct mdp5_crtc *mdp5_crtc = to_mdp5_crtc(crtc);
+	DBG("%s", mdp5_crtc->name);
 	mdp5_crtc_dpms(crtc, DRM_MODE_DPMS_ON);
-	crtc_flush(crtc);
+	crtc_flush_all(crtc);
 	/* drop the ref to mdp clk's that we got in prepare: */
 	mdp5_disable(get_kms(crtc));
 }
 
-static int mdp5_crtc_mode_set_base(struct drm_crtc *crtc, int x, int y,
-		struct drm_framebuffer *old_fb)
+static void mdp5_crtc_load_lut(struct drm_crtc *crtc)
+{
+}
+
+struct plane_state {
+	struct drm_plane *plane;
+	struct mdp5_plane_state *state;
+};
+
+static int pstate_cmp(const void *a, const void *b)
+{
+	struct plane_state *pa = (struct plane_state *)a;
+	struct plane_state *pb = (struct plane_state *)b;
+	return pa->state->zpos - pb->state->zpos;
+}
+
+static int mdp5_crtc_atomic_check(struct drm_crtc *crtc,
+		struct drm_crtc_state *state)
 {
 	struct mdp5_crtc *mdp5_crtc = to_mdp5_crtc(crtc);
-	struct drm_plane *plane = mdp5_crtc->plane;
-	struct drm_display_mode *mode = &crtc->mode;
-	int ret;
-
-	/* grab extra ref for update_scanout() */
-	drm_framebuffer_reference(crtc->primary->fb);
-
-	ret = mdp5_plane_mode_set(plane, crtc, crtc->primary->fb,
-			0, 0, mode->hdisplay, mode->vdisplay,
-			x << 16, y << 16,
-			mode->hdisplay << 16, mode->vdisplay << 16);
-	if (ret) {
-		drm_framebuffer_unreference(crtc->primary->fb);
-		return ret;
+	struct mdp5_kms *mdp5_kms = get_kms(crtc);
+	struct drm_plane *plane;
+	struct drm_device *dev = crtc->dev;
+	struct plane_state pstates[STAGE3 + 1];
+	int cnt = 0, i;
+
+	DBG("%s: check", mdp5_crtc->name);
+
+	/* request a free CTL, if none is already allocated for this CRTC */
+	if (state->enable && !mdp5_crtc->ctl) {
+		mdp5_crtc->ctl = mdp5_ctlm_request(mdp5_kms->ctlm, crtc);
+		if (WARN_ON(!mdp5_crtc->ctl))
+			return -EINVAL;
 	}
 
-	update_fb(crtc, crtc->primary->fb);
-	update_scanout(crtc, crtc->primary->fb);
+	/* verify that there are not too many planes attached to crtc
+	 * and that we don't have conflicting mixer stages:
+	 */
+	drm_atomic_crtc_state_for_each_plane(plane, state) {
+		struct drm_plane_state *pstate;
+
+		if (cnt >= ARRAY_SIZE(pstates)) {
+			dev_err(dev->dev, "too many planes!\n");
+			return -EINVAL;
+		}
+
+		pstate = state->state->plane_states[drm_plane_index(plane)];
+
+		/* plane might not have changed, in which case take
+		 * current state:
+		 */
+		if (!pstate)
+			pstate = plane->state;
+
+		pstates[cnt].plane = plane;
+		pstates[cnt].state = to_mdp5_plane_state(pstate);
+
+		cnt++;
+	}
+
+	sort(pstates, cnt, sizeof(pstates[0]), pstate_cmp, NULL);
+
+	for (i = 0; i < cnt; i++) {
+		pstates[i].state->stage = STAGE_BASE + i;
+		DBG("%s: assign pipe %s on stage=%d", mdp5_crtc->name,
+				pipe2name(mdp5_plane_pipe(pstates[i].plane)),
+				pstates[i].state->stage);
+	}
 
 	return 0;
 }
 
-static void mdp5_crtc_load_lut(struct drm_crtc *crtc)
+static void mdp5_crtc_atomic_begin(struct drm_crtc *crtc)
 {
+	struct mdp5_crtc *mdp5_crtc = to_mdp5_crtc(crtc);
+	DBG("%s: begin", mdp5_crtc->name);
 }
 
-static int mdp5_crtc_page_flip(struct drm_crtc *crtc,
-		struct drm_framebuffer *new_fb,
-		struct drm_pending_vblank_event *event,
-		uint32_t page_flip_flags)
+static void mdp5_crtc_atomic_flush(struct drm_crtc *crtc)
 {
 	struct mdp5_crtc *mdp5_crtc = to_mdp5_crtc(crtc);
 	struct drm_device *dev = crtc->dev;
-	struct drm_gem_object *obj;
 	unsigned long flags;
 
-	if (mdp5_crtc->event) {
-		dev_err(dev->dev, "already pending flip!\n");
-		return -EBUSY;
-	}
+	DBG("%s: event: %p", mdp5_crtc->name, crtc->state->event);
 
-	obj = msm_framebuffer_bo(new_fb, 0);
+	WARN_ON(mdp5_crtc->event);
 
 	spin_lock_irqsave(&dev->event_lock, flags);
-	mdp5_crtc->event = event;
+	mdp5_crtc->event = crtc->state->event;
 	spin_unlock_irqrestore(&dev->event_lock, flags);
 
-	update_fb(crtc, new_fb);
+	blend_setup(crtc);
+	crtc_flush_all(crtc);
+	request_pending(crtc, PENDING_FLIP);
 
-	return msm_gem_queue_inactive_cb(obj, &mdp5_crtc->pageflip_cb);
+	if (mdp5_crtc->ctl && !crtc->state->enable) {
+		mdp5_ctl_release(mdp5_crtc->ctl);
+		mdp5_crtc->ctl = NULL;
+	}
 }
 
 static int mdp5_crtc_set_property(struct drm_crtc *crtc,
@@ -391,27 +385,33 @@ static int mdp5_crtc_set_property(struct drm_crtc *crtc,
 }
 
 static const struct drm_crtc_funcs mdp5_crtc_funcs = {
-	.set_config = drm_crtc_helper_set_config,
+	.set_config = drm_atomic_helper_set_config,
 	.destroy = mdp5_crtc_destroy,
-	.page_flip = mdp5_crtc_page_flip,
+	.page_flip = drm_atomic_helper_page_flip,
 	.set_property = mdp5_crtc_set_property,
+	.reset = drm_atomic_helper_crtc_reset,
+	.atomic_duplicate_state = drm_atomic_helper_crtc_duplicate_state,
+	.atomic_destroy_state = drm_atomic_helper_crtc_destroy_state,
 };
 
 static const struct drm_crtc_helper_funcs mdp5_crtc_helper_funcs = {
 	.dpms = mdp5_crtc_dpms,
 	.mode_fixup = mdp5_crtc_mode_fixup,
-	.mode_set = mdp5_crtc_mode_set,
+	.mode_set_nofb = mdp5_crtc_mode_set_nofb,
+	.mode_set = drm_helper_crtc_mode_set,
+	.mode_set_base = drm_helper_crtc_mode_set_base,
 	.prepare = mdp5_crtc_prepare,
 	.commit = mdp5_crtc_commit,
-	.mode_set_base = mdp5_crtc_mode_set_base,
 	.load_lut = mdp5_crtc_load_lut,
+	.atomic_check = mdp5_crtc_atomic_check,
+	.atomic_begin = mdp5_crtc_atomic_begin,
+	.atomic_flush = mdp5_crtc_atomic_flush,
 };
 
 static void mdp5_crtc_vblank_irq(struct mdp_irq *irq, uint32_t irqstatus)
 {
 	struct mdp5_crtc *mdp5_crtc = container_of(irq, struct mdp5_crtc, vblank);
 	struct drm_crtc *crtc = &mdp5_crtc->base;
-	struct msm_drm_private *priv = crtc->dev->dev_private;
 	unsigned pending;
 
 	mdp_irq_unregister(&get_kms(crtc)->base, &mdp5_crtc->vblank);
@@ -420,16 +420,14 @@ static void mdp5_crtc_vblank_irq(struct mdp_irq *irq, uint32_t irqstatus)
 
 	if (pending & PENDING_FLIP) {
 		complete_flip(crtc, NULL);
-		drm_flip_work_commit(&mdp5_crtc->unref_fb_work, priv->wq);
 	}
 }
 
 static void mdp5_crtc_err_irq(struct mdp_irq *irq, uint32_t irqstatus)
 {
 	struct mdp5_crtc *mdp5_crtc = container_of(irq, struct mdp5_crtc, err);
-	struct drm_crtc *crtc = &mdp5_crtc->base;
+
 	DBG("%s: error: %08x", mdp5_crtc->name, irqstatus);
-	crtc_flush(crtc);
 }
 
 uint32_t mdp5_crtc_vblank(struct drm_crtc *crtc)
@@ -450,19 +448,16 @@ void mdp5_crtc_set_intf(struct drm_crtc *crtc, int intf,
 {
 	struct mdp5_crtc *mdp5_crtc = to_mdp5_crtc(crtc);
 	struct mdp5_kms *mdp5_kms = get_kms(crtc);
-	static const enum mdp5_intfnum intfnum[] = {
-			INTF0, INTF1, INTF2, INTF3,
-	};
+	uint32_t flush_mask = 0;
 	uint32_t intf_sel;
+	unsigned long flags;
 
 	/* now that we know what irq's we want: */
 	mdp5_crtc->err.irqmask = intf2err(intf);
 	mdp5_crtc->vblank.irqmask = intf2vblank(intf);
+	mdp_irq_update(&mdp5_kms->base);
 
-	/* when called from modeset_init(), skip the rest until later: */
-	if (!mdp5_kms)
-		return;
-
+	spin_lock_irqsave(&mdp5_kms->resource_lock, flags);
 	intf_sel = mdp5_read(mdp5_kms, REG_MDP5_DISP_INTF_SEL);
 
 	switch (intf) {
@@ -487,45 +482,25 @@ void mdp5_crtc_set_intf(struct drm_crtc *crtc, int intf,
 		break;
 	}
 
-	blend_setup(crtc);
+	mdp5_write(mdp5_kms, REG_MDP5_DISP_INTF_SEL, intf_sel);
+	spin_unlock_irqrestore(&mdp5_kms->resource_lock, flags);
 
 	DBG("%s: intf_sel=%08x", mdp5_crtc->name, intf_sel);
+	mdp5_ctl_set_intf(mdp5_crtc->ctl, intf);
+	flush_mask |= mdp5_ctl_get_flush(mdp5_crtc->ctl);
+	flush_mask |= mdp5_lm_get_flush(mdp5_crtc->lm);
 
-	mdp5_write(mdp5_kms, REG_MDP5_DISP_INTF_SEL, intf_sel);
-	mdp5_write(mdp5_kms, REG_MDP5_CTL_OP(mdp5_crtc->id),
-			MDP5_CTL_OP_MODE(MODE_NONE) |
-			MDP5_CTL_OP_INTF_NUM(intfnum[intf]));
-
-	crtc_flush(crtc);
+	crtc_flush(crtc, flush_mask);
 }
 
-static void set_attach(struct drm_crtc *crtc, enum mdp5_pipe pipe_id,
-		struct drm_plane *plane)
+int mdp5_crtc_get_lm(struct drm_crtc *crtc)
 {
 	struct mdp5_crtc *mdp5_crtc = to_mdp5_crtc(crtc);
 
-	BUG_ON(pipe_id >= ARRAY_SIZE(mdp5_crtc->planes));
-
-	if (mdp5_crtc->planes[pipe_id] == plane)
-		return;
-
-	mdp5_crtc->planes[pipe_id] = plane;
-	blend_setup(crtc);
-	if (mdp5_crtc->enabled && (plane != mdp5_crtc->plane))
-		crtc_flush(crtc);
-}
+	if (WARN_ON(!crtc))
+		return -EINVAL;
 
-void mdp5_crtc_attach(struct drm_crtc *crtc, struct drm_plane *plane)
-{
-	set_attach(crtc, mdp5_plane_pipe(plane), plane);
-}
-
-void mdp5_crtc_detach(struct drm_crtc *crtc, struct drm_plane *plane)
-{
-	/* don't actually detatch our primary plane: */
-	if (to_mdp5_crtc(crtc)->plane == plane)
-		return;
-	set_attach(crtc, mdp5_plane_pipe(plane), NULL);
+	return mdp5_crtc->lm;
 }
 
 /* initialize crtc */
@@ -534,18 +509,17 @@ struct drm_crtc *mdp5_crtc_init(struct drm_device *dev,
 {
 	struct drm_crtc *crtc = NULL;
 	struct mdp5_crtc *mdp5_crtc;
-	int ret;
 
 	mdp5_crtc = kzalloc(sizeof(*mdp5_crtc), GFP_KERNEL);
-	if (!mdp5_crtc) {
-		ret = -ENOMEM;
-		goto fail;
-	}
+	if (!mdp5_crtc)
+		return ERR_PTR(-ENOMEM);
 
 	crtc = &mdp5_crtc->base;
 
-	mdp5_crtc->plane = plane;
 	mdp5_crtc->id = id;
+	mdp5_crtc->lm = GET_LM_ID(id);
+
+	spin_lock_init(&mdp5_crtc->lm_lock);
 
 	mdp5_crtc->vblank.irq = mdp5_crtc_vblank_irq;
 	mdp5_crtc->err.irq = mdp5_crtc_err_irq;
@@ -553,23 +527,11 @@ struct drm_crtc *mdp5_crtc_init(struct drm_device *dev,
 	snprintf(mdp5_crtc->name, sizeof(mdp5_crtc->name), "%s:%d",
 			pipe2name(mdp5_plane_pipe(plane)), id);
 
-	ret = drm_flip_work_init(&mdp5_crtc->unref_fb_work, 16,
-			"unref fb", unref_fb_worker);
-	if (ret)
-		goto fail;
-
-	INIT_FENCE_CB(&mdp5_crtc->pageflip_cb, pageflip_cb);
-
 	drm_crtc_init_with_planes(dev, crtc, plane, NULL, &mdp5_crtc_funcs);
 	drm_crtc_helper_add(crtc, &mdp5_crtc_helper_funcs);
+	plane->crtc = crtc;
 
-	mdp5_plane_install_properties(mdp5_crtc->plane, &crtc->base);
+	mdp5_plane_install_properties(plane, &crtc->base);
 
 	return crtc;
-
-fail:
-	if (crtc)
-		mdp5_crtc_destroy(crtc);
-
-	return ERR_PTR(ret);
 }
diff --git a/drivers/gpu/drm/msm/mdp/mdp5/mdp5_ctl.c b/drivers/gpu/drm/msm/mdp/mdp5/mdp5_ctl.c
new file mode 100644
index 000000000000..dea4505ac963
--- /dev/null
+++ b/drivers/gpu/drm/msm/mdp/mdp5/mdp5_ctl.c
@@ -0,0 +1,322 @@
+/*
+ * Copyright (c) 2014 The Linux Foundation. All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 and
+ * only version 2 as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ */
+
+#include "mdp5_kms.h"
+#include "mdp5_ctl.h"
+
+/*
+ * CTL - MDP Control Pool Manager
+ *
+ * Controls are shared between all CRTCs.
+ *
+ * They are intended to be used for data path configuration.
+ * The top level register programming describes the complete data path for
+ * a specific data path ID - REG_MDP5_CTL_*(<id>, ...)
+ *
+ * Hardware capabilities determine the number of concurrent data paths
+ *
+ * In certain use cases (high-resolution dual pipe), one single CTL can be
+ * shared across multiple CRTCs.
+ *
+ * Because the number of CTLs can be less than the number of CRTCs,
+ * CTLs are dynamically allocated from a pool of CTLs, only once a CRTC is
+ * requested by the client (in mdp5_crtc_mode_set()).
+ */
+
+struct mdp5_ctl {
+	struct mdp5_ctl_manager *ctlm;
+
+	u32 id;
+
+	/* whether this CTL has been allocated or not: */
+	bool busy;
+
+	/* memory output connection (@see mdp5_ctl_mode): */
+	u32 mode;
+
+	/* REG_MDP5_CTL_*(<id>) registers access info + lock: */
+	spinlock_t hw_lock;
+	u32 reg_offset;
+
+	/* flush mask used to commit CTL registers */
+	u32 flush_mask;
+
+	bool cursor_on;
+
+	struct drm_crtc *crtc;
+};
+
+struct mdp5_ctl_manager {
+	struct drm_device *dev;
+
+	/* number of CTL / Layer Mixers in this hw config: */
+	u32 nlm;
+	u32 nctl;
+
+	/* pool of CTLs + lock to protect resource allocation (ctls[i].busy) */
+	spinlock_t pool_lock;
+	struct mdp5_ctl ctls[MAX_CTL];
+};
+
+static inline
+struct mdp5_kms *get_kms(struct mdp5_ctl_manager *ctl_mgr)
+{
+	struct msm_drm_private *priv = ctl_mgr->dev->dev_private;
+
+	return to_mdp5_kms(to_mdp_kms(priv->kms));
+}
+
+static inline
+void ctl_write(struct mdp5_ctl *ctl, u32 reg, u32 data)
+{
+	struct mdp5_kms *mdp5_kms = get_kms(ctl->ctlm);
+
+	(void)ctl->reg_offset; /* TODO use this instead of mdp5_write */
+	mdp5_write(mdp5_kms, reg, data);
+}
+
+static inline
+u32 ctl_read(struct mdp5_ctl *ctl, u32 reg)
+{
+	struct mdp5_kms *mdp5_kms = get_kms(ctl->ctlm);
+
+	(void)ctl->reg_offset; /* TODO use this instead of mdp5_write */
+	return mdp5_read(mdp5_kms, reg);
+}
+
+
+int mdp5_ctl_set_intf(struct mdp5_ctl *ctl, enum mdp5_intf intf)
+{
+	unsigned long flags;
+	static const enum mdp5_intfnum intfnum[] = {
+			INTF0, INTF1, INTF2, INTF3,
+	};
+
+	spin_lock_irqsave(&ctl->hw_lock, flags);
+	ctl_write(ctl, REG_MDP5_CTL_OP(ctl->id),
+			MDP5_CTL_OP_MODE(ctl->mode) |
+			MDP5_CTL_OP_INTF_NUM(intfnum[intf]));
+	spin_unlock_irqrestore(&ctl->hw_lock, flags);
+
+	return 0;
+}
+
+int mdp5_ctl_set_cursor(struct mdp5_ctl *ctl, bool enable)
+{
+	struct mdp5_ctl_manager *ctl_mgr = ctl->ctlm;
+	unsigned long flags;
+	u32 blend_cfg;
+	int lm;
+
+	lm = mdp5_crtc_get_lm(ctl->crtc);
+	if (unlikely(WARN_ON(lm < 0))) {
+		dev_err(ctl_mgr->dev->dev, "CTL %d cannot find LM: %d",
+				ctl->id, lm);
+		return -EINVAL;
+	}
+
+	spin_lock_irqsave(&ctl->hw_lock, flags);
+
+	blend_cfg = ctl_read(ctl, REG_MDP5_CTL_LAYER_REG(ctl->id, lm));
+
+	if (enable)
+		blend_cfg |=  MDP5_CTL_LAYER_REG_CURSOR_OUT;
+	else
+		blend_cfg &= ~MDP5_CTL_LAYER_REG_CURSOR_OUT;
+
+	ctl_write(ctl, REG_MDP5_CTL_LAYER_REG(ctl->id, lm), blend_cfg);
+
+	spin_unlock_irqrestore(&ctl->hw_lock, flags);
+
+	ctl->cursor_on = enable;
+
+	return 0;
+}
+
+
+int mdp5_ctl_blend(struct mdp5_ctl *ctl, u32 lm, u32 blend_cfg)
+{
+	unsigned long flags;
+
+	if (ctl->cursor_on)
+		blend_cfg |=  MDP5_CTL_LAYER_REG_CURSOR_OUT;
+	else
+		blend_cfg &= ~MDP5_CTL_LAYER_REG_CURSOR_OUT;
+
+	spin_lock_irqsave(&ctl->hw_lock, flags);
+	ctl_write(ctl, REG_MDP5_CTL_LAYER_REG(ctl->id, lm), blend_cfg);
+	spin_unlock_irqrestore(&ctl->hw_lock, flags);
+
+	return 0;
+}
+
+int mdp5_ctl_commit(struct mdp5_ctl *ctl, u32 flush_mask)
+{
+	struct mdp5_ctl_manager *ctl_mgr = ctl->ctlm;
+	unsigned long flags;
+
+	if (flush_mask & MDP5_CTL_FLUSH_CURSOR_DUMMY) {
+		int lm = mdp5_crtc_get_lm(ctl->crtc);
+
+		if (unlikely(WARN_ON(lm < 0))) {
+			dev_err(ctl_mgr->dev->dev, "CTL %d cannot find LM: %d",
+					ctl->id, lm);
+			return -EINVAL;
+		}
+
+		/* for current targets, cursor bit is the same as LM bit */
+		flush_mask |= mdp_ctl_flush_mask_lm(lm);
+	}
+
+	spin_lock_irqsave(&ctl->hw_lock, flags);
+	ctl_write(ctl, REG_MDP5_CTL_FLUSH(ctl->id), flush_mask);
+	spin_unlock_irqrestore(&ctl->hw_lock, flags);
+
+	return 0;
+}
+
+u32 mdp5_ctl_get_flush(struct mdp5_ctl *ctl)
+{
+	return ctl->flush_mask;
+}
+
+void mdp5_ctl_release(struct mdp5_ctl *ctl)
+{
+	struct mdp5_ctl_manager *ctl_mgr = ctl->ctlm;
+	unsigned long flags;
+
+	if (unlikely(WARN_ON(ctl->id >= MAX_CTL) || !ctl->busy)) {
+		dev_err(ctl_mgr->dev->dev, "CTL %d in bad state (%d)",
+				ctl->id, ctl->busy);
+		return;
+	}
+
+	spin_lock_irqsave(&ctl_mgr->pool_lock, flags);
+	ctl->busy = false;
+	spin_unlock_irqrestore(&ctl_mgr->pool_lock, flags);
+
+	DBG("CTL %d released", ctl->id);
+}
+
+/*
+ * mdp5_ctl_request() - CTL dynamic allocation
+ *
+ * Note: Current implementation considers that we can only have one CRTC per CTL
+ *
+ * @return first free CTL
+ */
+struct mdp5_ctl *mdp5_ctlm_request(struct mdp5_ctl_manager *ctl_mgr,
+		struct drm_crtc *crtc)
+{
+	struct mdp5_ctl *ctl = NULL;
+	unsigned long flags;
+	int c;
+
+	spin_lock_irqsave(&ctl_mgr->pool_lock, flags);
+
+	for (c = 0; c < ctl_mgr->nctl; c++)
+		if (!ctl_mgr->ctls[c].busy)
+			break;
+
+	if (unlikely(c >= ctl_mgr->nctl)) {
+		dev_err(ctl_mgr->dev->dev, "No more CTL available!");
+		goto unlock;
+	}
+
+	ctl = &ctl_mgr->ctls[c];
+
+	ctl->crtc = crtc;
+	ctl->busy = true;
+	DBG("CTL %d allocated", ctl->id);
+
+unlock:
+	spin_unlock_irqrestore(&ctl_mgr->pool_lock, flags);
+	return ctl;
+}
+
+void mdp5_ctlm_hw_reset(struct mdp5_ctl_manager *ctl_mgr)
+{
+	unsigned long flags;
+	int c;
+
+	for (c = 0; c < ctl_mgr->nctl; c++) {
+		struct mdp5_ctl *ctl = &ctl_mgr->ctls[c];
+
+		spin_lock_irqsave(&ctl->hw_lock, flags);
+		ctl_write(ctl, REG_MDP5_CTL_OP(ctl->id), 0);
+		spin_unlock_irqrestore(&ctl->hw_lock, flags);
+	}
+}
+
+void mdp5_ctlm_destroy(struct mdp5_ctl_manager *ctl_mgr)
+{
+	kfree(ctl_mgr);
+}
+
+struct mdp5_ctl_manager *mdp5_ctlm_init(struct drm_device *dev,
+		void __iomem *mmio_base, const struct mdp5_cfg_hw *hw_cfg)
+{
+	struct mdp5_ctl_manager *ctl_mgr;
+	const struct mdp5_sub_block *ctl_cfg = &hw_cfg->ctl;
+	unsigned long flags;
+	int c, ret;
+
+	ctl_mgr = kzalloc(sizeof(*ctl_mgr), GFP_KERNEL);
+	if (!ctl_mgr) {
+		dev_err(dev->dev, "failed to allocate CTL manager\n");
+		ret = -ENOMEM;
+		goto fail;
+	}
+
+	if (unlikely(WARN_ON(ctl_cfg->count > MAX_CTL))) {
+		dev_err(dev->dev, "Increase static pool size to at least %d\n",
+				ctl_cfg->count);
+		ret = -ENOSPC;
+		goto fail;
+	}
+
+	/* initialize the CTL manager: */
+	ctl_mgr->dev = dev;
+	ctl_mgr->nlm = hw_cfg->lm.count;
+	ctl_mgr->nctl = ctl_cfg->count;
+	spin_lock_init(&ctl_mgr->pool_lock);
+
+	/* initialize each CTL of the pool: */
+	spin_lock_irqsave(&ctl_mgr->pool_lock, flags);
+	for (c = 0; c < ctl_mgr->nctl; c++) {
+		struct mdp5_ctl *ctl = &ctl_mgr->ctls[c];
+
+		if (WARN_ON(!ctl_cfg->base[c])) {
+			dev_err(dev->dev, "CTL_%d: base is null!\n", c);
+			ret = -EINVAL;
+			goto fail;
+		}
+		ctl->ctlm = ctl_mgr;
+		ctl->id = c;
+		ctl->mode = MODE_NONE;
+		ctl->reg_offset = ctl_cfg->base[c];
+		ctl->flush_mask = MDP5_CTL_FLUSH_CTL;
+		ctl->busy = false;
+		spin_lock_init(&ctl->hw_lock);
+	}
+	spin_unlock_irqrestore(&ctl_mgr->pool_lock, flags);
+	DBG("Pool of %d CTLs created.", ctl_mgr->nctl);
+
+	return ctl_mgr;
+
+fail:
+	if (ctl_mgr)
+		mdp5_ctlm_destroy(ctl_mgr);
+
+	return ERR_PTR(ret);
+}
diff --git a/drivers/gpu/drm/msm/mdp/mdp5/mdp5_ctl.h b/drivers/gpu/drm/msm/mdp/mdp5/mdp5_ctl.h
new file mode 100644
index 000000000000..1018519b6af2
--- /dev/null
+++ b/drivers/gpu/drm/msm/mdp/mdp5/mdp5_ctl.h
@@ -0,0 +1,122 @@
+/*
+ * Copyright (c) 2014 The Linux Foundation. All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 and
+ * only version 2 as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ */
+
+#ifndef __MDP5_CTL_H__
+#define __MDP5_CTL_H__
+
+#include "msm_drv.h"
+
+/*
+ * CTL Manager prototypes:
+ * mdp5_ctlm_init() returns a ctlm (CTL Manager) handler,
+ * which is then used to call the other mdp5_ctlm_*(ctlm, ...) functions.
+ */
+struct mdp5_ctl_manager;
+struct mdp5_ctl_manager *mdp5_ctlm_init(struct drm_device *dev,
+		void __iomem *mmio_base, const struct mdp5_cfg_hw *hw_cfg);
+void mdp5_ctlm_hw_reset(struct mdp5_ctl_manager *ctlm);
+void mdp5_ctlm_destroy(struct mdp5_ctl_manager *ctlm);
+
+/*
+ * CTL prototypes:
+ * mdp5_ctl_request(ctlm, ...) returns a ctl (CTL resource) handler,
+ * which is then used to call the other mdp5_ctl_*(ctl, ...) functions.
+ */
+struct mdp5_ctl *mdp5_ctlm_request(struct mdp5_ctl_manager *ctlm, struct drm_crtc *crtc);
+
+int mdp5_ctl_set_intf(struct mdp5_ctl *ctl, enum mdp5_intf intf);
+
+int mdp5_ctl_set_cursor(struct mdp5_ctl *ctl, bool enable);
+
+/* @blend_cfg: see LM blender config definition below */
+int mdp5_ctl_blend(struct mdp5_ctl *ctl, u32 lm, u32 blend_cfg);
+
+/* @flush_mask: see CTL flush masks definitions below */
+int mdp5_ctl_commit(struct mdp5_ctl *ctl, u32 flush_mask);
+u32 mdp5_ctl_get_flush(struct mdp5_ctl *ctl);
+
+void mdp5_ctl_release(struct mdp5_ctl *ctl);
+
+/*
+ * blend_cfg (LM blender config):
+ *
+ * The function below allows the caller of mdp5_ctl_blend() to specify how pipes
+ * are being blended according to their stage (z-order), through @blend_cfg arg.
+ */
+static inline u32 mdp_ctl_blend_mask(enum mdp5_pipe pipe,
+		enum mdp_mixer_stage_id stage)
+{
+	switch (pipe) {
+	case SSPP_VIG0: return MDP5_CTL_LAYER_REG_VIG0(stage);
+	case SSPP_VIG1: return MDP5_CTL_LAYER_REG_VIG1(stage);
+	case SSPP_VIG2: return MDP5_CTL_LAYER_REG_VIG2(stage);
+	case SSPP_RGB0: return MDP5_CTL_LAYER_REG_RGB0(stage);
+	case SSPP_RGB1: return MDP5_CTL_LAYER_REG_RGB1(stage);
+	case SSPP_RGB2: return MDP5_CTL_LAYER_REG_RGB2(stage);
+	case SSPP_DMA0: return MDP5_CTL_LAYER_REG_DMA0(stage);
+	case SSPP_DMA1: return MDP5_CTL_LAYER_REG_DMA1(stage);
+	case SSPP_VIG3: return MDP5_CTL_LAYER_REG_VIG3(stage);
+	case SSPP_RGB3: return MDP5_CTL_LAYER_REG_RGB3(stage);
+	default:	return 0;
+	}
+}
+
+/*
+ * flush_mask (CTL flush masks):
+ *
+ * The following functions allow each DRM entity to get and store
+ * their own flush mask.
+ * Once stored, these masks will then be accessed through each DRM's
+ * interface and used by the caller of mdp5_ctl_commit() to specify
+ * which block(s) need to be flushed through @flush_mask parameter.
+ */
+
+#define MDP5_CTL_FLUSH_CURSOR_DUMMY	0x80000000
+
+static inline u32 mdp_ctl_flush_mask_cursor(int cursor_id)
+{
+	/* TODO: use id once multiple cursor support is present */
+	(void)cursor_id;
+
+	return MDP5_CTL_FLUSH_CURSOR_DUMMY;
+}
+
+static inline u32 mdp_ctl_flush_mask_lm(int lm)
+{
+	switch (lm) {
+	case 0:  return MDP5_CTL_FLUSH_LM0;
+	case 1:  return MDP5_CTL_FLUSH_LM1;
+	case 2:  return MDP5_CTL_FLUSH_LM2;
+	case 5:  return MDP5_CTL_FLUSH_LM5;
+	default: return 0;
+	}
+}
+
+static inline u32 mdp_ctl_flush_mask_pipe(enum mdp5_pipe pipe)
+{
+	switch (pipe) {
+	case SSPP_VIG0: return MDP5_CTL_FLUSH_VIG0;
+	case SSPP_VIG1: return MDP5_CTL_FLUSH_VIG1;
+	case SSPP_VIG2: return MDP5_CTL_FLUSH_VIG2;
+	case SSPP_RGB0: return MDP5_CTL_FLUSH_RGB0;
+	case SSPP_RGB1: return MDP5_CTL_FLUSH_RGB1;
+	case SSPP_RGB2: return MDP5_CTL_FLUSH_RGB2;
+	case SSPP_DMA0: return MDP5_CTL_FLUSH_DMA0;
+	case SSPP_DMA1: return MDP5_CTL_FLUSH_DMA1;
+	case SSPP_VIG3: return MDP5_CTL_FLUSH_VIG3;
+	case SSPP_RGB3: return MDP5_CTL_FLUSH_RGB3;
+	default:        return 0;
+	}
+}
+
+#endif /* __MDP5_CTL_H__ */
diff --git a/drivers/gpu/drm/msm/mdp/mdp5/mdp5_encoder.c b/drivers/gpu/drm/msm/mdp/mdp5/mdp5_encoder.c
index edec7bfaa952..0254bfdeb92f 100644
--- a/drivers/gpu/drm/msm/mdp/mdp5/mdp5_encoder.c
+++ b/drivers/gpu/drm/msm/mdp/mdp5/mdp5_encoder.c
@@ -24,6 +24,7 @@ struct mdp5_encoder {
 	struct drm_encoder base;
 	int intf;
 	enum mdp5_intf intf_id;
+	spinlock_t intf_lock;	/* protect REG_MDP5_INTF_* registers */
 	bool enabled;
 	uint32_t bsc;
 };
@@ -115,6 +116,7 @@ static void mdp5_encoder_dpms(struct drm_encoder *encoder, int mode)
 	struct mdp5_kms *mdp5_kms = get_kms(encoder);
 	int intf = mdp5_encoder->intf;
 	bool enabled = (mode == DRM_MODE_DPMS_ON);
+	unsigned long flags;
 
 	DBG("mode=%d", mode);
 
@@ -123,9 +125,24 @@ static void mdp5_encoder_dpms(struct drm_encoder *encoder, int mode)
 
 	if (enabled) {
 		bs_set(mdp5_encoder, 1);
+		spin_lock_irqsave(&mdp5_encoder->intf_lock, flags);
 		mdp5_write(mdp5_kms, REG_MDP5_INTF_TIMING_ENGINE_EN(intf), 1);
+		spin_unlock_irqrestore(&mdp5_encoder->intf_lock, flags);
 	} else {
+		spin_lock_irqsave(&mdp5_encoder->intf_lock, flags);
 		mdp5_write(mdp5_kms, REG_MDP5_INTF_TIMING_ENGINE_EN(intf), 0);
+		spin_unlock_irqrestore(&mdp5_encoder->intf_lock, flags);
+
+		/*
+		 * Wait for a vsync so we know the ENABLE=0 latched before
+		 * the (connector) source of the vsync's gets disabled,
+		 * otherwise we end up in a funny state if we re-enable
+		 * before the disable latches, which results that some of
+		 * the settings changes for the new modeset (like new
+		 * scanout buffer) don't latch properly..
+		 */
+		mdp_irq_wait(&mdp5_kms->base, intf2vblank(intf));
+
 		bs_set(mdp5_encoder, 0);
 	}
 
@@ -150,6 +167,7 @@ static void mdp5_encoder_mode_set(struct drm_encoder *encoder,
 	uint32_t display_v_start, display_v_end;
 	uint32_t hsync_start_x, hsync_end_x;
 	uint32_t format;
+	unsigned long flags;
 
 	mode = adjusted_mode;
 
@@ -180,6 +198,8 @@ static void mdp5_encoder_mode_set(struct drm_encoder *encoder,
 	display_v_start = (mode->vtotal - mode->vsync_start) * mode->htotal + dtv_hsync_skew;
 	display_v_end = vsync_period - ((mode->vsync_start - mode->vdisplay) * mode->htotal) + dtv_hsync_skew - 1;
 
+	spin_lock_irqsave(&mdp5_encoder->intf_lock, flags);
+
 	mdp5_write(mdp5_kms, REG_MDP5_INTF_HSYNC_CTL(intf),
 			MDP5_INTF_HSYNC_CTL_PULSEW(mode->hsync_end - mode->hsync_start) |
 			MDP5_INTF_HSYNC_CTL_PERIOD(mode->htotal));
@@ -201,6 +221,8 @@ static void mdp5_encoder_mode_set(struct drm_encoder *encoder,
 	mdp5_write(mdp5_kms, REG_MDP5_INTF_ACTIVE_VEND_F0(intf), 0);
 	mdp5_write(mdp5_kms, REG_MDP5_INTF_PANEL_FORMAT(intf), format);
 	mdp5_write(mdp5_kms, REG_MDP5_INTF_FRAME_LINE_COUNT_EN(intf), 0x3);  /* frame+line? */
+
+	spin_unlock_irqrestore(&mdp5_encoder->intf_lock, flags);
 }
 
 static void mdp5_encoder_prepare(struct drm_encoder *encoder)
@@ -242,6 +264,8 @@ struct drm_encoder *mdp5_encoder_init(struct drm_device *dev, int intf,
 	mdp5_encoder->intf_id = intf_id;
 	encoder = &mdp5_encoder->base;
 
+	spin_lock_init(&mdp5_encoder->intf_lock);
+
 	drm_encoder_init(dev, encoder, &mdp5_encoder_funcs,
 			 DRM_MODE_ENCODER_TMDS);
 	drm_encoder_helper_add(encoder, &mdp5_encoder_helper_funcs);
diff --git a/drivers/gpu/drm/msm/mdp/mdp5/mdp5_irq.c b/drivers/gpu/drm/msm/mdp/mdp5/mdp5_irq.c
index f2b985bc2adf..70ac81edd40f 100644
--- a/drivers/gpu/drm/msm/mdp/mdp5/mdp5_irq.c
+++ b/drivers/gpu/drm/msm/mdp/mdp5/mdp5_irq.c
@@ -15,6 +15,8 @@
  * this program.  If not, see <http://www.gnu.org/licenses/>.
  */
 
+#include <linux/irqdomain.h>
+#include <linux/irq.h>
 
 #include "msm_drv.h"
 #include "mdp5_kms.h"
@@ -88,11 +90,17 @@ irqreturn_t mdp5_irq(struct msm_kms *kms)
 
 	VERB("intr=%08x", intr);
 
-	if (intr & MDP5_HW_INTR_STATUS_INTR_MDP)
+	if (intr & MDP5_HW_INTR_STATUS_INTR_MDP) {
 		mdp5_irq_mdp(mdp_kms);
+		intr &= ~MDP5_HW_INTR_STATUS_INTR_MDP;
+	}
 
-	if (intr & MDP5_HW_INTR_STATUS_INTR_HDMI)
-		hdmi_irq(0, mdp5_kms->hdmi);
+	while (intr) {
+		irq_hw_number_t hwirq = fls(intr) - 1;
+		generic_handle_irq(irq_find_mapping(
+				mdp5_kms->irqcontroller.domain, hwirq));
+		intr &= ~(1 << hwirq);
+	}
 
 	return IRQ_HANDLED;
 }
@@ -109,3 +117,82 @@ void mdp5_disable_vblank(struct msm_kms *kms, struct drm_crtc *crtc)
 	mdp_update_vblank_mask(to_mdp_kms(kms),
 			mdp5_crtc_vblank(crtc), false);
 }
+
+/*
+ * interrupt-controller implementation, so sub-blocks (hdmi/eDP/dsi/etc)
+ * can register to get their irq's delivered
+ */
+
+#define VALID_IRQS  (MDP5_HW_INTR_STATUS_INTR_DSI0 | \
+		MDP5_HW_INTR_STATUS_INTR_DSI1 | \
+		MDP5_HW_INTR_STATUS_INTR_HDMI | \
+		MDP5_HW_INTR_STATUS_INTR_EDP)
+
+static void mdp5_hw_mask_irq(struct irq_data *irqd)
+{
+	struct mdp5_kms *mdp5_kms = irq_data_get_irq_chip_data(irqd);
+	smp_mb__before_atomic();
+	clear_bit(irqd->hwirq, &mdp5_kms->irqcontroller.enabled_mask);
+	smp_mb__after_atomic();
+}
+
+static void mdp5_hw_unmask_irq(struct irq_data *irqd)
+{
+	struct mdp5_kms *mdp5_kms = irq_data_get_irq_chip_data(irqd);
+	smp_mb__before_atomic();
+	set_bit(irqd->hwirq, &mdp5_kms->irqcontroller.enabled_mask);
+	smp_mb__after_atomic();
+}
+
+static struct irq_chip mdp5_hw_irq_chip = {
+	.name		= "mdp5",
+	.irq_mask	= mdp5_hw_mask_irq,
+	.irq_unmask	= mdp5_hw_unmask_irq,
+};
+
+static int mdp5_hw_irqdomain_map(struct irq_domain *d,
+		unsigned int irq, irq_hw_number_t hwirq)
+{
+	struct mdp5_kms *mdp5_kms = d->host_data;
+
+	if (!(VALID_IRQS & (1 << hwirq)))
+		return -EPERM;
+
+	irq_set_chip_and_handler(irq, &mdp5_hw_irq_chip, handle_level_irq);
+	irq_set_chip_data(irq, mdp5_kms);
+	set_irq_flags(irq, IRQF_VALID);
+
+	return 0;
+}
+
+static struct irq_domain_ops mdp5_hw_irqdomain_ops = {
+	.map = mdp5_hw_irqdomain_map,
+	.xlate = irq_domain_xlate_onecell,
+};
+
+
+int mdp5_irq_domain_init(struct mdp5_kms *mdp5_kms)
+{
+	struct device *dev = mdp5_kms->dev->dev;
+	struct irq_domain *d;
+
+	d = irq_domain_add_linear(dev->of_node, 32,
+			&mdp5_hw_irqdomain_ops, mdp5_kms);
+	if (!d) {
+		dev_err(dev, "mdp5 irq domain add failed\n");
+		return -ENXIO;
+	}
+
+	mdp5_kms->irqcontroller.enabled_mask = 0;
+	mdp5_kms->irqcontroller.domain = d;
+
+	return 0;
+}
+
+void mdp5_irq_domain_fini(struct mdp5_kms *mdp5_kms)
+{
+	if (mdp5_kms->irqcontroller.domain) {
+		irq_domain_remove(mdp5_kms->irqcontroller.domain);
+		mdp5_kms->irqcontroller.domain = NULL;
+	}
+}
diff --git a/drivers/gpu/drm/msm/mdp/mdp5/mdp5_kms.c b/drivers/gpu/drm/msm/mdp/mdp5/mdp5_kms.c
index 31a2c6331a1d..9f01a4f21af2 100644
--- a/drivers/gpu/drm/msm/mdp/mdp5/mdp5_kms.c
+++ b/drivers/gpu/drm/msm/mdp/mdp5/mdp5_kms.c
@@ -1,4 +1,5 @@
 /*
+ * Copyright (c) 2014, The Linux Foundation. All rights reserved.
  * Copyright (C) 2013 Red Hat
  * Author: Rob Clark <robdclark@gmail.com>
  *
@@ -24,145 +25,11 @@ static const char *iommu_ports[] = {
 		"mdp_0",
 };
 
-static struct mdp5_platform_config *mdp5_get_config(struct platform_device *dev);
-
-const struct mdp5_config *mdp5_cfg;
-
-static const struct mdp5_config msm8x74_config = {
-	.name = "msm8x74",
-	.ctl = {
-		.count = 5,
-		.base = { 0x00600, 0x00700, 0x00800, 0x00900, 0x00a00 },
-	},
-	.pipe_vig = {
-		.count = 3,
-		.base = { 0x01200, 0x01600, 0x01a00 },
-	},
-	.pipe_rgb = {
-		.count = 3,
-		.base = { 0x01e00, 0x02200, 0x02600 },
-	},
-	.pipe_dma = {
-		.count = 2,
-		.base = { 0x02a00, 0x02e00 },
-	},
-	.lm = {
-		.count = 5,
-		.base = { 0x03200, 0x03600, 0x03a00, 0x03e00, 0x04200 },
-	},
-	.dspp = {
-		.count = 3,
-		.base = { 0x04600, 0x04a00, 0x04e00 },
-	},
-	.ad = {
-		.count = 2,
-		.base = { 0x13100, 0x13300 }, /* NOTE: no ad in v1.0 */
-	},
-	.intf = {
-		.count = 4,
-		.base = { 0x12500, 0x12700, 0x12900, 0x12b00 },
-	},
-};
-
-static const struct mdp5_config apq8084_config = {
-	.name = "apq8084",
-	.ctl = {
-		.count = 5,
-		.base = { 0x00600, 0x00700, 0x00800, 0x00900, 0x00a00 },
-	},
-	.pipe_vig = {
-		.count = 4,
-		.base = { 0x01200, 0x01600, 0x01a00, 0x01e00 },
-	},
-	.pipe_rgb = {
-		.count = 4,
-		.base = { 0x02200, 0x02600, 0x02a00, 0x02e00 },
-	},
-	.pipe_dma = {
-		.count = 2,
-		.base = { 0x03200, 0x03600 },
-	},
-	.lm = {
-		.count = 6,
-		.base = { 0x03a00, 0x03e00, 0x04200, 0x04600, 0x04a00, 0x04e00 },
-	},
-	.dspp = {
-		.count = 4,
-		.base = { 0x05200, 0x05600, 0x05a00, 0x05e00 },
-
-	},
-	.ad = {
-		.count = 3,
-		.base = { 0x13500, 0x13700, 0x13900 },
-	},
-	.intf = {
-		.count = 5,
-		.base = { 0x12500, 0x12700, 0x12900, 0x12b00, 0x12d00 },
-	},
-};
-
-struct mdp5_config_entry {
-	int revision;
-	const struct mdp5_config *config;
-};
-
-static const struct mdp5_config_entry mdp5_configs[] = {
-	{ .revision = 0, .config = &msm8x74_config },
-	{ .revision = 2, .config = &msm8x74_config },
-	{ .revision = 3, .config = &apq8084_config },
-};
-
-static int mdp5_select_hw_cfg(struct msm_kms *kms)
-{
-	struct mdp5_kms *mdp5_kms = to_mdp5_kms(to_mdp_kms(kms));
-	struct drm_device *dev = mdp5_kms->dev;
-	uint32_t version, major, minor;
-	int i, ret = 0;
-
-	mdp5_enable(mdp5_kms);
-	version = mdp5_read(mdp5_kms, REG_MDP5_MDP_VERSION);
-	mdp5_disable(mdp5_kms);
-
-	major = FIELD(version, MDP5_MDP_VERSION_MAJOR);
-	minor = FIELD(version, MDP5_MDP_VERSION_MINOR);
-
-	DBG("found MDP5 version v%d.%d", major, minor);
-
-	if (major != 1) {
-		dev_err(dev->dev, "unexpected MDP major version: v%d.%d\n",
-				major, minor);
-		ret = -ENXIO;
-		goto out;
-	}
-
-	mdp5_kms->rev = minor;
-
-	/* only after mdp5_cfg global pointer's init can we access the hw */
-	for (i = 0; i < ARRAY_SIZE(mdp5_configs); i++) {
-		if (mdp5_configs[i].revision != minor)
-			continue;
-		mdp5_kms->hw_cfg = mdp5_cfg = mdp5_configs[i].config;
-		break;
-	}
-	if (unlikely(!mdp5_kms->hw_cfg)) {
-		dev_err(dev->dev, "unexpected MDP minor revision: v%d.%d\n",
-				major, minor);
-		ret = -ENXIO;
-		goto out;
-	}
-
-	DBG("MDP5: %s config selected", mdp5_kms->hw_cfg->name);
-
-	return 0;
-out:
-	return ret;
-}
-
 static int mdp5_hw_init(struct msm_kms *kms)
 {
 	struct mdp5_kms *mdp5_kms = to_mdp5_kms(to_mdp_kms(kms));
 	struct drm_device *dev = mdp5_kms->dev;
-	int i;
+	unsigned long flags;
 
 	pm_runtime_get_sync(dev->dev);
 
@@ -190,10 +57,11 @@ static int mdp5_hw_init(struct msm_kms *kms)
 	 * care.
 	 */
 
+	spin_lock_irqsave(&mdp5_kms->resource_lock, flags);
 	mdp5_write(mdp5_kms, REG_MDP5_DISP_INTF_SEL, 0);
+	spin_unlock_irqrestore(&mdp5_kms->resource_lock, flags);
 
-	for (i = 0; i < mdp5_kms->hw_cfg->ctl.count; i++)
-		mdp5_write(mdp5_kms, REG_MDP5_CTL_OP(i), 0);
+	mdp5_ctlm_hw_reset(mdp5_kms->ctlm);
 
 	pm_runtime_put_sync(dev->dev);
 
@@ -221,10 +89,20 @@ static void mdp5_destroy(struct msm_kms *kms)
 	struct mdp5_kms *mdp5_kms = to_mdp5_kms(to_mdp_kms(kms));
 	struct msm_mmu *mmu = mdp5_kms->mmu;
 
+	mdp5_irq_domain_fini(mdp5_kms);
+
 	if (mmu) {
 		mmu->funcs->detach(mmu, iommu_ports, ARRAY_SIZE(iommu_ports));
 		mmu->funcs->destroy(mmu);
 	}
+
+	if (mdp5_kms->ctlm)
+		mdp5_ctlm_destroy(mdp5_kms->ctlm);
+	if (mdp5_kms->smp)
+		mdp5_smp_destroy(mdp5_kms->smp);
+	if (mdp5_kms->cfg)
+		mdp5_cfg_destroy(mdp5_kms->cfg);
+
 	kfree(mdp5_kms);
 }
 
@@ -274,17 +152,31 @@ static int modeset_init(struct mdp5_kms *mdp5_kms)
 	static const enum mdp5_pipe crtcs[] = {
 			SSPP_RGB0, SSPP_RGB1, SSPP_RGB2, SSPP_RGB3,
 	};
+	static const enum mdp5_pipe pub_planes[] = {
+			SSPP_VIG0, SSPP_VIG1, SSPP_VIG2, SSPP_VIG3,
+	};
 	struct drm_device *dev = mdp5_kms->dev;
 	struct msm_drm_private *priv = dev->dev_private;
 	struct drm_encoder *encoder;
+	const struct mdp5_cfg_hw *hw_cfg;
 	int i, ret;
 
-	/* construct CRTCs: */
-	for (i = 0; i < mdp5_kms->hw_cfg->pipe_rgb.count; i++) {
+	hw_cfg = mdp5_cfg_get_hw_config(mdp5_kms->cfg);
+
+	/* register our interrupt-controller for hdmi/eDP/dsi/etc
+	 * to use for irqs routed through mdp:
+	 */
+	ret = mdp5_irq_domain_init(mdp5_kms);
+	if (ret)
+		goto fail;
+
+	/* construct CRTCs and their private planes: */
+	for (i = 0; i < hw_cfg->pipe_rgb.count; i++) {
 		struct drm_plane *plane;
 		struct drm_crtc *crtc;
 
-		plane = mdp5_plane_init(dev, crtcs[i], true);
+		plane = mdp5_plane_init(dev, crtcs[i], true,
+				hw_cfg->pipe_rgb.base[i]);
 		if (IS_ERR(plane)) {
 			ret = PTR_ERR(plane);
 			dev_err(dev->dev, "failed to construct plane for %s (%d)\n",
@@ -302,6 +194,20 @@ static int modeset_init(struct mdp5_kms *mdp5_kms)
 		priv->crtcs[priv->num_crtcs++] = crtc;
 	}
 
+	/* Construct public planes: */
+	for (i = 0; i < hw_cfg->pipe_vig.count; i++) {
+		struct drm_plane *plane;
+
+		plane = mdp5_plane_init(dev, pub_planes[i], false,
+				hw_cfg->pipe_vig.base[i]);
+		if (IS_ERR(plane)) {
+			ret = PTR_ERR(plane);
+			dev_err(dev->dev, "failed to construct %s plane: %d\n",
+					pipe2name(pub_planes[i]), ret);
+			goto fail;
+		}
+	}
+
 	/* Construct encoder for HDMI: */
 	encoder = mdp5_encoder_init(dev, 3, INTF_HDMI);
 	if (IS_ERR(encoder)) {
@@ -310,25 +216,16 @@ static int modeset_init(struct mdp5_kms *mdp5_kms)
 		goto fail;
 	}
 
-	/* NOTE: the vsync and error irq's are actually associated with
-	 * the INTF/encoder.. the easiest way to deal with this (ie. what
-	 * we do now) is assume a fixed relationship between crtc's and
-	 * encoders.  I'm not sure if there is ever a need to more freely
-	 * assign crtcs to encoders, but if there is then we need to take
-	 * care of error and vblank irq's that the crtc has registered,
-	 * and also update user-requested vblank_mask.
-	 */
-	encoder->possible_crtcs = BIT(0);
-	mdp5_crtc_set_intf(priv->crtcs[0], 3, INTF_HDMI);
-
+	encoder->possible_crtcs = (1 << priv->num_crtcs) - 1;;
 	priv->encoders[priv->num_encoders++] = encoder;
 
 	/* Construct bridge/connector for HDMI: */
-	mdp5_kms->hdmi = hdmi_init(dev, encoder);
-	if (IS_ERR(mdp5_kms->hdmi)) {
-		ret = PTR_ERR(mdp5_kms->hdmi);
-		dev_err(dev->dev, "failed to initialize HDMI: %d\n", ret);
-		goto fail;
+	if (priv->hdmi) {
+		ret = hdmi_modeset_init(priv->hdmi, dev, encoder);
+		if (ret) {
+			dev_err(dev->dev, "failed to initialize HDMI: %d\n", ret);
+			goto fail;
+		}
 	}
 
 	return 0;
@@ -337,6 +234,21 @@ fail:
 	return ret;
 }
 
+static void read_hw_revision(struct mdp5_kms *mdp5_kms,
+		uint32_t *major, uint32_t *minor)
+{
+	uint32_t version;
+
+	mdp5_enable(mdp5_kms);
+	version = mdp5_read(mdp5_kms, REG_MDP5_MDP_VERSION);
+	mdp5_disable(mdp5_kms);
+
+	*major = FIELD(version, MDP5_MDP_VERSION_MAJOR);
+	*minor = FIELD(version, MDP5_MDP_VERSION_MINOR);
+
+	DBG("MDP5 version v%d.%d", *major, *minor);
+}
+
 static int get_clk(struct platform_device *pdev, struct clk **clkp,
 		const char *name)
 {
@@ -353,10 +265,11 @@ static int get_clk(struct platform_device *pdev, struct clk **clkp,
 struct msm_kms *mdp5_kms_init(struct drm_device *dev)
 {
 	struct platform_device *pdev = dev->platformdev;
-	struct mdp5_platform_config *config = mdp5_get_config(pdev);
+	struct mdp5_cfg *config;
 	struct mdp5_kms *mdp5_kms;
 	struct msm_kms *kms = NULL;
 	struct msm_mmu *mmu;
+	uint32_t major, minor;
 	int i, ret;
 
 	mdp5_kms = kzalloc(sizeof(*mdp5_kms), GFP_KERNEL);
@@ -366,12 +279,13 @@ struct msm_kms *mdp5_kms_init(struct drm_device *dev)
 		goto fail;
 	}
 
+	spin_lock_init(&mdp5_kms->resource_lock);
+
 	mdp_kms_init(&mdp5_kms->base, &kms_funcs);
 
 	kms = &mdp5_kms->base.base;
 
 	mdp5_kms->dev = dev;
-	mdp5_kms->smp_blk_cnt = config->smp_blk_cnt;
 
 	mdp5_kms->mmio = msm_ioremap(pdev, "mdp_phys", "MDP5");
 	if (IS_ERR(mdp5_kms->mmio)) {
@@ -416,24 +330,52 @@ struct msm_kms *mdp5_kms_init(struct drm_device *dev)
 	if (ret)
 		goto fail;
 
-	ret = clk_set_rate(mdp5_kms->src_clk, config->max_clk);
+	/* we need to set a default rate before enabling.  Set a safe
+	 * rate first, then figure out hw revision, and then set a
+	 * more optimal rate:
+	 */
+	clk_set_rate(mdp5_kms->src_clk, 200000000);
 
-	ret = mdp5_select_hw_cfg(kms);
-	if (ret)
+	read_hw_revision(mdp5_kms, &major, &minor);
+
+	mdp5_kms->cfg = mdp5_cfg_init(mdp5_kms, major, minor);
+	if (IS_ERR(mdp5_kms->cfg)) {
+		ret = PTR_ERR(mdp5_kms->cfg);
+		mdp5_kms->cfg = NULL;
 		goto fail;
+	}
+
+	config = mdp5_cfg_get_config(mdp5_kms->cfg);
+
+	/* TODO: compute core clock rate at runtime */
+	clk_set_rate(mdp5_kms->src_clk, config->hw->max_clk);
+
+	mdp5_kms->smp = mdp5_smp_init(mdp5_kms->dev, &config->hw->smp);
+	if (IS_ERR(mdp5_kms->smp)) {
+		ret = PTR_ERR(mdp5_kms->smp);
+		mdp5_kms->smp = NULL;
+		goto fail;
+	}
+
+	mdp5_kms->ctlm = mdp5_ctlm_init(dev, mdp5_kms->mmio, config->hw);
+	if (IS_ERR(mdp5_kms->ctlm)) {
+		ret = PTR_ERR(mdp5_kms->ctlm);
+		mdp5_kms->ctlm = NULL;
+		goto fail;
+	}
 
 	/* make sure things are off before attaching iommu (bootloader could
 	 * have left things on, in which case we'll start getting faults if
 	 * we don't disable):
 	 */
 	mdp5_enable(mdp5_kms);
-	for (i = 0; i < mdp5_kms->hw_cfg->intf.count; i++)
+	for (i = 0; i < config->hw->intf.count; i++)
 		mdp5_write(mdp5_kms, REG_MDP5_INTF_TIMING_ENGINE_EN(i), 0);
 	mdp5_disable(mdp5_kms);
 	mdelay(16);
 
-	if (config->iommu) {
-		mmu = msm_iommu_new(&pdev->dev, config->iommu);
+	if (config->platform.iommu) {
+		mmu = msm_iommu_new(&pdev->dev, config->platform.iommu);
 		if (IS_ERR(mmu)) {
 			ret = PTR_ERR(mmu);
 			dev_err(dev->dev, "failed to init iommu: %d\n", ret);
@@ -474,18 +416,3 @@ fail:
 		mdp5_destroy(kms);
 	return ERR_PTR(ret);
 }
-
-static struct mdp5_platform_config *mdp5_get_config(struct platform_device *dev)
-{
-	static struct mdp5_platform_config config = {};
-#ifdef CONFIG_OF
-	/* TODO */
-#endif
-	config.iommu = iommu_domain_alloc(&platform_bus_type);
-	/* TODO hard-coded in downstream mdss, but should it be? */
-	config.max_clk = 200000000;
-	/* TODO get from DT: */
-	config.smp_blk_cnt = 22;
-
-	return &config;
-}
diff --git a/drivers/gpu/drm/msm/mdp/mdp5/mdp5_kms.h b/drivers/gpu/drm/msm/mdp/mdp5/mdp5_kms.h
index 5bf340dd0f00..dd69c77c0d64 100644
--- a/drivers/gpu/drm/msm/mdp/mdp5/mdp5_kms.h
+++ b/drivers/gpu/drm/msm/mdp/mdp5/mdp5_kms.h
@@ -21,25 +21,9 @@
 #include "msm_drv.h"
 #include "msm_kms.h"
 #include "mdp/mdp_kms.h"
-/* dynamic offsets used by mdp5.xml.h (initialized in mdp5_kms.c) */
-#define MDP5_MAX_BASES		8
-struct mdp5_sub_block {
-	int	count;
-	uint32_t base[MDP5_MAX_BASES];
-};
-struct mdp5_config {
-	char  *name;
-	struct mdp5_sub_block ctl;
-	struct mdp5_sub_block pipe_vig;
-	struct mdp5_sub_block pipe_rgb;
-	struct mdp5_sub_block pipe_dma;
-	struct mdp5_sub_block lm;
-	struct mdp5_sub_block dspp;
-	struct mdp5_sub_block ad;
-	struct mdp5_sub_block intf;
-};
-extern const struct mdp5_config *mdp5_cfg;
+#include "mdp5_cfg.h"	/* must be included before mdp5.xml.h */
 #include "mdp5.xml.h"
+#include "mdp5_ctl.h"
 #include "mdp5_smp.h"
 
 struct mdp5_kms {
@@ -47,17 +31,14 @@ struct mdp5_kms {
 
 	struct drm_device *dev;
 
-	int rev;
-	const struct mdp5_config *hw_cfg;
+	struct mdp5_cfg_handler *cfg;
 
 	/* mapper-id used to request GEM buffer mapped for scanout: */
 	int id;
 	struct msm_mmu *mmu;
 
-	/* for tracking smp allocation amongst pipes: */
-	mdp5_smp_state_t smp_state;
-	struct mdp5_client_smp_state smp_client_state[CID_MAX];
-	int smp_blk_cnt;
+	struct mdp5_smp *smp;
+	struct mdp5_ctl_manager *ctlm;
 
 	/* io/register spaces: */
 	void __iomem *mmio, *vbif;
@@ -71,18 +52,47 @@ struct mdp5_kms {
 	struct clk *lut_clk;
 	struct clk *vsync_clk;
 
-	struct hdmi *hdmi;
+	/*
+	 * lock to protect access to global resources: ie., following register:
+	 *	- REG_MDP5_DISP_INTF_SEL
+	 */
+	spinlock_t resource_lock;
 
 	struct mdp_irq error_handler;
+
+	struct {
+		volatile unsigned long enabled_mask;
+		struct irq_domain *domain;
+	} irqcontroller;
 };
 #define to_mdp5_kms(x) container_of(x, struct mdp5_kms, base)
 
-/* platform config data (ie. from DT, or pdata) */
-struct mdp5_platform_config {
-	struct iommu_domain *iommu;
-	uint32_t max_clk;
-	int smp_blk_cnt;
+struct mdp5_plane_state {
+	struct drm_plane_state base;
+
+	/* "virtual" zpos.. we calculate actual mixer-stage at runtime
+	 * by sorting the attached planes by zpos and then assigning
+	 * mixer stage lowest to highest.  Private planes get default
+	 * zpos of zero, and public planes a unique value that is
+	 * greater than zero.  This way, things work out if a naive
+	 * userspace assigns planes to a crtc without setting zpos.
+	 */
+	int zpos;
+
+	/* the actual mixer stage, calculated in crtc->atomic_check()
+	 * NOTE: this should move to mdp5_crtc_state, when that exists
+	 */
+	enum mdp_mixer_stage_id stage;
+
+	/* some additional transactional status to help us know in the
+	 * apply path whether we need to update SMP allocation, and
+	 * whether current update is still pending:
+	 */
+	bool mode_changed : 1;
+	bool pending : 1;
 };
+#define to_mdp5_plane_state(x) \
+		container_of(x, struct mdp5_plane_state, base)
 
 static inline void mdp5_write(struct mdp5_kms *mdp5_kms, u32 reg, u32 data)
 {
@@ -107,23 +117,6 @@ static inline const char *pipe2name(enum mdp5_pipe pipe)
 	return names[pipe];
 }
 
-static inline uint32_t pipe2flush(enum mdp5_pipe pipe)
-{
-	switch (pipe) {
-	case SSPP_VIG0: return MDP5_CTL_FLUSH_VIG0;
-	case SSPP_VIG1: return MDP5_CTL_FLUSH_VIG1;
-	case SSPP_VIG2: return MDP5_CTL_FLUSH_VIG2;
-	case SSPP_RGB0: return MDP5_CTL_FLUSH_RGB0;
-	case SSPP_RGB1: return MDP5_CTL_FLUSH_RGB1;
-	case SSPP_RGB2: return MDP5_CTL_FLUSH_RGB2;
-	case SSPP_DMA0: return MDP5_CTL_FLUSH_DMA0;
-	case SSPP_DMA1: return MDP5_CTL_FLUSH_DMA1;
-	case SSPP_VIG3: return MDP5_CTL_FLUSH_VIG3;
-	case SSPP_RGB3: return MDP5_CTL_FLUSH_RGB3;
-	default:        return 0;
-	}
-}
-
 static inline int pipe2nclients(enum mdp5_pipe pipe)
 {
 	switch (pipe) {
@@ -137,34 +130,6 @@ static inline int pipe2nclients(enum mdp5_pipe pipe)
 	}
 }
 
-static inline enum mdp5_client_id pipe2client(enum mdp5_pipe pipe, int plane)
-{
-	WARN_ON(plane >= pipe2nclients(pipe));
-	switch (pipe) {
-	case SSPP_VIG0: return CID_VIG0_Y + plane;
-	case SSPP_VIG1: return CID_VIG1_Y + plane;
-	case SSPP_VIG2: return CID_VIG2_Y + plane;
-	case SSPP_RGB0: return CID_RGB0;
-	case SSPP_RGB1: return CID_RGB1;
-	case SSPP_RGB2: return CID_RGB2;
-	case SSPP_DMA0: return CID_DMA0_Y + plane;
-	case SSPP_DMA1: return CID_DMA1_Y + plane;
-	case SSPP_VIG3: return CID_VIG3_Y + plane;
-	case SSPP_RGB3: return CID_RGB3;
-	default:        return CID_UNUSED;
-	}
-}
-
-static inline uint32_t mixer2flush(int lm)
-{
-	switch (lm) {
-	case 0:  return MDP5_CTL_FLUSH_LM0;
-	case 1:  return MDP5_CTL_FLUSH_LM1;
-	case 2:  return MDP5_CTL_FLUSH_LM2;
-	default: return 0;
-	}
-}
-
 static inline uint32_t intf2err(int intf)
 {
 	switch (intf) {
@@ -197,6 +162,8 @@ void mdp5_irq_uninstall(struct msm_kms *kms);
 irqreturn_t mdp5_irq(struct msm_kms *kms);
 int mdp5_enable_vblank(struct msm_kms *kms, struct drm_crtc *crtc);
 void mdp5_disable_vblank(struct msm_kms *kms, struct drm_crtc *crtc);
+int mdp5_irq_domain_init(struct mdp5_kms *mdp5_kms);
+void mdp5_irq_domain_fini(struct mdp5_kms *mdp5_kms);
 
 static inline
 uint32_t mdp5_get_formats(enum mdp5_pipe pipe, uint32_t *pixel_formats,
@@ -210,26 +177,18 @@ uint32_t mdp5_get_formats(enum mdp5_pipe pipe, uint32_t *pixel_formats,
 
 void mdp5_plane_install_properties(struct drm_plane *plane,
 		struct drm_mode_object *obj);
-void mdp5_plane_set_scanout(struct drm_plane *plane,
-		struct drm_framebuffer *fb);
-int mdp5_plane_mode_set(struct drm_plane *plane,
-		struct drm_crtc *crtc, struct drm_framebuffer *fb,
-		int crtc_x, int crtc_y,
-		unsigned int crtc_w, unsigned int crtc_h,
-		uint32_t src_x, uint32_t src_y,
-		uint32_t src_w, uint32_t src_h);
+uint32_t mdp5_plane_get_flush(struct drm_plane *plane);
 void mdp5_plane_complete_flip(struct drm_plane *plane);
 enum mdp5_pipe mdp5_plane_pipe(struct drm_plane *plane);
 struct drm_plane *mdp5_plane_init(struct drm_device *dev,
-		enum mdp5_pipe pipe, bool private_plane);
+		enum mdp5_pipe pipe, bool private_plane, uint32_t reg_offset);
 
 uint32_t mdp5_crtc_vblank(struct drm_crtc *crtc);
 
+int mdp5_crtc_get_lm(struct drm_crtc *crtc);
 void mdp5_crtc_cancel_pending_flip(struct drm_crtc *crtc, struct drm_file *file);
 void mdp5_crtc_set_intf(struct drm_crtc *crtc, int intf,
 		enum mdp5_intf intf_id);
-void mdp5_crtc_attach(struct drm_crtc *crtc, struct drm_plane *plane);
-void mdp5_crtc_detach(struct drm_crtc *crtc, struct drm_plane *plane);
 struct drm_crtc *mdp5_crtc_init(struct drm_device *dev,
 		struct drm_plane *plane, int id);
 
diff --git a/drivers/gpu/drm/msm/mdp/mdp5/mdp5_plane.c b/drivers/gpu/drm/msm/mdp/mdp5/mdp5_plane.c
index f3daec4412ad..26e5fdea6594 100644
--- a/drivers/gpu/drm/msm/mdp/mdp5/mdp5_plane.c
+++ b/drivers/gpu/drm/msm/mdp/mdp5/mdp5_plane.c
@@ -1,4 +1,5 @@
 /*
+ * Copyright (c) 2014 The Linux Foundation. All rights reserved.
  * Copyright (C) 2013 Red Hat
  * Author: Rob Clark <robdclark@gmail.com>
  *
@@ -17,6 +18,7 @@
 
 #include "mdp5_kms.h"
 
+#define MAX_PLANE	4
 
 struct mdp5_plane {
 	struct drm_plane base;
@@ -24,6 +26,11 @@ struct mdp5_plane {
 
 	enum mdp5_pipe pipe;
 
+	spinlock_t pipe_lock;	/* protect REG_MDP5_PIPE_* registers */
+	uint32_t reg_offset;
+
+	uint32_t flush_mask;	/* used to commit pipe registers */
+
 	uint32_t nformats;
 	uint32_t formats[32];
 
@@ -31,31 +38,24 @@ struct mdp5_plane {
 };
 #define to_mdp5_plane(x) container_of(x, struct mdp5_plane, base)
 
+static int mdp5_plane_mode_set(struct drm_plane *plane,
+		struct drm_crtc *crtc, struct drm_framebuffer *fb,
+		int crtc_x, int crtc_y,
+		unsigned int crtc_w, unsigned int crtc_h,
+		uint32_t src_x, uint32_t src_y,
+		uint32_t src_w, uint32_t src_h);
+static void set_scanout_locked(struct drm_plane *plane,
+		struct drm_framebuffer *fb);
+
 static struct mdp5_kms *get_kms(struct drm_plane *plane)
 {
 	struct msm_drm_private *priv = plane->dev->dev_private;
 	return to_mdp5_kms(to_mdp_kms(priv->kms));
 }
 
-static int mdp5_plane_update(struct drm_plane *plane,
-		struct drm_crtc *crtc, struct drm_framebuffer *fb,
-		int crtc_x, int crtc_y,
-		unsigned int crtc_w, unsigned int crtc_h,
-		uint32_t src_x, uint32_t src_y,
-		uint32_t src_w, uint32_t src_h)
+static bool plane_enabled(struct drm_plane_state *state)
 {
-	struct mdp5_plane *mdp5_plane = to_mdp5_plane(plane);
-
-	mdp5_plane->enabled = true;
-
-	if (plane->fb)
-		drm_framebuffer_unreference(plane->fb);
-
-	drm_framebuffer_reference(fb);
-
-	return mdp5_plane_mode_set(plane, crtc, fb,
-			crtc_x, crtc_y, crtc_w, crtc_h,
-			src_x, src_y, src_w, src_h);
+	return state->fb && state->crtc;
 }
 
 static int mdp5_plane_disable(struct drm_plane *plane)
@@ -63,21 +63,13 @@ static int mdp5_plane_disable(struct drm_plane *plane)
 	struct mdp5_plane *mdp5_plane = to_mdp5_plane(plane);
 	struct mdp5_kms *mdp5_kms = get_kms(plane);
 	enum mdp5_pipe pipe = mdp5_plane->pipe;
-	int i;
 
 	DBG("%s: disable", mdp5_plane->name);
 
-	/* update our SMP request to zero (release all our blks): */
-	for (i = 0; i < pipe2nclients(pipe); i++)
-		mdp5_smp_request(mdp5_kms, pipe2client(pipe, i), 0);
-
-	/* TODO detaching now will cause us not to get the last
-	 * vblank and mdp5_smp_commit().. so other planes will
-	 * still see smp blocks previously allocated to us as
-	 * in-use..
-	 */
-	if (plane->crtc)
-		mdp5_crtc_detach(plane->crtc, plane);
+	if (mdp5_kms) {
+		/* Release the memory we requested earlier from the SMP: */
+		mdp5_smp_release(mdp5_kms->smp, pipe);
+	}
 
 	return 0;
 }
@@ -85,11 +77,8 @@ static int mdp5_plane_disable(struct drm_plane *plane)
 static void mdp5_plane_destroy(struct drm_plane *plane)
 {
 	struct mdp5_plane *mdp5_plane = to_mdp5_plane(plane);
-	struct msm_drm_private *priv = plane->dev->dev_private;
-
-	if (priv->kms)
-		mdp5_plane_disable(plane);
 
+	drm_plane_helper_disable(plane);
 	drm_plane_cleanup(plane);
 
 	kfree(mdp5_plane);
@@ -109,109 +98,186 @@ int mdp5_plane_set_property(struct drm_plane *plane,
 	return -EINVAL;
 }
 
+static void mdp5_plane_reset(struct drm_plane *plane)
+{
+	struct mdp5_plane_state *mdp5_state;
+
+	if (plane->state && plane->state->fb)
+		drm_framebuffer_unreference(plane->state->fb);
+
+	kfree(to_mdp5_plane_state(plane->state));
+	mdp5_state = kzalloc(sizeof(*mdp5_state), GFP_KERNEL);
+
+	if (plane->type == DRM_PLANE_TYPE_PRIMARY) {
+		mdp5_state->zpos = 0;
+	} else {
+		mdp5_state->zpos = 1 + drm_plane_index(plane);
+	}
+
+	plane->state = &mdp5_state->base;
+}
+
+static struct drm_plane_state *
+mdp5_plane_duplicate_state(struct drm_plane *plane)
+{
+	struct mdp5_plane_state *mdp5_state;
+
+	if (WARN_ON(!plane->state))
+		return NULL;
+
+	mdp5_state = kmemdup(to_mdp5_plane_state(plane->state),
+			sizeof(*mdp5_state), GFP_KERNEL);
+
+	if (mdp5_state && mdp5_state->base.fb)
+		drm_framebuffer_reference(mdp5_state->base.fb);
+
+	mdp5_state->mode_changed = false;
+	mdp5_state->pending = false;
+
+	return &mdp5_state->base;
+}
+
+static void mdp5_plane_destroy_state(struct drm_plane *plane,
+		struct drm_plane_state *state)
+{
+	if (state->fb)
+		drm_framebuffer_unreference(state->fb);
+
+	kfree(to_mdp5_plane_state(state));
+}
+
 static const struct drm_plane_funcs mdp5_plane_funcs = {
-		.update_plane = mdp5_plane_update,
-		.disable_plane = mdp5_plane_disable,
+		.update_plane = drm_atomic_helper_update_plane,
+		.disable_plane = drm_atomic_helper_disable_plane,
 		.destroy = mdp5_plane_destroy,
 		.set_property = mdp5_plane_set_property,
+		.reset = mdp5_plane_reset,
+		.atomic_duplicate_state = mdp5_plane_duplicate_state,
+		.atomic_destroy_state = mdp5_plane_destroy_state,
 };
 
-void mdp5_plane_set_scanout(struct drm_plane *plane,
+static int mdp5_plane_prepare_fb(struct drm_plane *plane,
 		struct drm_framebuffer *fb)
 {
 	struct mdp5_plane *mdp5_plane = to_mdp5_plane(plane);
 	struct mdp5_kms *mdp5_kms = get_kms(plane);
-	enum mdp5_pipe pipe = mdp5_plane->pipe;
-	uint32_t nplanes = drm_format_num_planes(fb->pixel_format);
-	uint32_t iova[4];
-	int i;
-
-	for (i = 0; i < nplanes; i++) {
-		struct drm_gem_object *bo = msm_framebuffer_bo(fb, i);
-		msm_gem_get_iova(bo, mdp5_kms->id, &iova[i]);
-	}
-	for (; i < 4; i++)
-		iova[i] = 0;
 
-	mdp5_write(mdp5_kms, REG_MDP5_PIPE_SRC_STRIDE_A(pipe),
-			MDP5_PIPE_SRC_STRIDE_A_P0(fb->pitches[0]) |
-			MDP5_PIPE_SRC_STRIDE_A_P1(fb->pitches[1]));
-
-	mdp5_write(mdp5_kms, REG_MDP5_PIPE_SRC_STRIDE_B(pipe),
-			MDP5_PIPE_SRC_STRIDE_B_P2(fb->pitches[2]) |
-			MDP5_PIPE_SRC_STRIDE_B_P3(fb->pitches[3]));
-
-	mdp5_write(mdp5_kms, REG_MDP5_PIPE_SRC0_ADDR(pipe), iova[0]);
-	mdp5_write(mdp5_kms, REG_MDP5_PIPE_SRC1_ADDR(pipe), iova[1]);
-	mdp5_write(mdp5_kms, REG_MDP5_PIPE_SRC2_ADDR(pipe), iova[2]);
-	mdp5_write(mdp5_kms, REG_MDP5_PIPE_SRC3_ADDR(pipe), iova[3]);
-
-	plane->fb = fb;
+	DBG("%s: prepare: FB[%u]", mdp5_plane->name, fb->base.id);
+	return msm_framebuffer_prepare(fb, mdp5_kms->id);
 }
 
-/* NOTE: looks like if horizontal decimation is used (if we supported that)
- * then the width used to calculate SMP block requirements is the post-
- * decimated width.  Ie. SMP buffering sits downstream of decimation (which
- * presumably happens during the dma from scanout buffer).
- */
-static int request_smp_blocks(struct drm_plane *plane, uint32_t format,
-		uint32_t nplanes, uint32_t width)
+static void mdp5_plane_cleanup_fb(struct drm_plane *plane,
+		struct drm_framebuffer *fb)
 {
-	struct drm_device *dev = plane->dev;
 	struct mdp5_plane *mdp5_plane = to_mdp5_plane(plane);
 	struct mdp5_kms *mdp5_kms = get_kms(plane);
-	enum mdp5_pipe pipe = mdp5_plane->pipe;
-	int i, hsub, nlines, nblks, ret;
 
-	hsub = drm_format_horz_chroma_subsampling(format);
+	DBG("%s: cleanup: FB[%u]", mdp5_plane->name, fb->base.id);
+	msm_framebuffer_cleanup(fb, mdp5_kms->id);
+}
 
-	/* different if BWC (compressed framebuffer?) enabled: */
-	nlines = 2;
+static int mdp5_plane_atomic_check(struct drm_plane *plane,
+		struct drm_plane_state *state)
+{
+	struct mdp5_plane *mdp5_plane = to_mdp5_plane(plane);
+	struct drm_plane_state *old_state = plane->state;
 
-	for (i = 0, nblks = 0; i < nplanes; i++) {
-		int n, fetch_stride, cpp;
+	DBG("%s: check (%d -> %d)", mdp5_plane->name,
+			plane_enabled(old_state), plane_enabled(state));
 
-		cpp = drm_format_plane_cpp(format, i);
-		fetch_stride = width * cpp / (i ? hsub : 1);
+	if (plane_enabled(state) && plane_enabled(old_state)) {
+		/* we cannot change SMP block configuration during scanout: */
+		bool full_modeset = false;
+		if (state->fb->pixel_format != old_state->fb->pixel_format) {
+			DBG("%s: pixel_format change!", mdp5_plane->name);
+			full_modeset = true;
+		}
+		if (state->src_w != old_state->src_w) {
+			DBG("%s: src_w change!", mdp5_plane->name);
+			full_modeset = true;
+		}
+		if (to_mdp5_plane_state(old_state)->pending) {
+			DBG("%s: still pending!", mdp5_plane->name);
+			full_modeset = true;
+		}
+		if (full_modeset) {
+			struct drm_crtc_state *crtc_state =
+					drm_atomic_get_crtc_state(state->state, state->crtc);
+			crtc_state->mode_changed = true;
+			to_mdp5_plane_state(state)->mode_changed = true;
+		}
+	} else {
+		to_mdp5_plane_state(state)->mode_changed = true;
+	}
 
-		n = DIV_ROUND_UP(fetch_stride * nlines, SMP_BLK_SIZE);
+	return 0;
+}
 
-		/* for hw rev v1.00 */
-		if (mdp5_kms->rev == 0)
-			n = roundup_pow_of_two(n);
+static void mdp5_plane_atomic_update(struct drm_plane *plane,
+				     struct drm_plane_state *old_state)
+{
+	struct mdp5_plane *mdp5_plane = to_mdp5_plane(plane);
+	struct drm_plane_state *state = plane->state;
 
-		DBG("%s[%d]: request %d SMP blocks", mdp5_plane->name, i, n);
-		ret = mdp5_smp_request(mdp5_kms, pipe2client(pipe, i), n);
-		if (ret) {
-			dev_err(dev->dev, "Could not allocate %d SMP blocks: %d\n",
-					n, ret);
-			return ret;
-		}
+	DBG("%s: update", mdp5_plane->name);
 
-		nblks += n;
+	if (!plane_enabled(state)) {
+		to_mdp5_plane_state(state)->pending = true;
+		mdp5_plane_disable(plane);
+	} else if (to_mdp5_plane_state(state)->mode_changed) {
+		int ret;
+		to_mdp5_plane_state(state)->pending = true;
+		ret = mdp5_plane_mode_set(plane,
+				state->crtc, state->fb,
+				state->crtc_x, state->crtc_y,
+				state->crtc_w, state->crtc_h,
+				state->src_x,  state->src_y,
+				state->src_w, state->src_h);
+		/* atomic_check should have ensured that this doesn't fail */
+		WARN_ON(ret < 0);
+	} else {
+		unsigned long flags;
+		spin_lock_irqsave(&mdp5_plane->pipe_lock, flags);
+		set_scanout_locked(plane, state->fb);
+		spin_unlock_irqrestore(&mdp5_plane->pipe_lock, flags);
 	}
-
-	/* in success case, return total # of blocks allocated: */
-	return nblks;
 }
 
-static void set_fifo_thresholds(struct drm_plane *plane, int nblks)
+static const struct drm_plane_helper_funcs mdp5_plane_helper_funcs = {
+		.prepare_fb = mdp5_plane_prepare_fb,
+		.cleanup_fb = mdp5_plane_cleanup_fb,
+		.atomic_check = mdp5_plane_atomic_check,
+		.atomic_update = mdp5_plane_atomic_update,
+};
+
+static void set_scanout_locked(struct drm_plane *plane,
+		struct drm_framebuffer *fb)
 {
 	struct mdp5_plane *mdp5_plane = to_mdp5_plane(plane);
 	struct mdp5_kms *mdp5_kms = get_kms(plane);
 	enum mdp5_pipe pipe = mdp5_plane->pipe;
-	uint32_t val;
 
-	/* 1/4 of SMP pool that is being fetched */
-	val = (nblks * SMP_ENTRIES_PER_BLK) / 4;
+	mdp5_write(mdp5_kms, REG_MDP5_PIPE_SRC_STRIDE_A(pipe),
+			MDP5_PIPE_SRC_STRIDE_A_P0(fb->pitches[0]) |
+			MDP5_PIPE_SRC_STRIDE_A_P1(fb->pitches[1]));
 
-	mdp5_write(mdp5_kms, REG_MDP5_PIPE_REQPRIO_FIFO_WM_0(pipe), val * 1);
-	mdp5_write(mdp5_kms, REG_MDP5_PIPE_REQPRIO_FIFO_WM_1(pipe), val * 2);
-	mdp5_write(mdp5_kms, REG_MDP5_PIPE_REQPRIO_FIFO_WM_2(pipe), val * 3);
+	mdp5_write(mdp5_kms, REG_MDP5_PIPE_SRC_STRIDE_B(pipe),
+			MDP5_PIPE_SRC_STRIDE_B_P2(fb->pitches[2]) |
+			MDP5_PIPE_SRC_STRIDE_B_P3(fb->pitches[3]));
+
+	mdp5_write(mdp5_kms, REG_MDP5_PIPE_SRC0_ADDR(pipe),
+			msm_framebuffer_iova(fb, mdp5_kms->id, 0));
+	mdp5_write(mdp5_kms, REG_MDP5_PIPE_SRC1_ADDR(pipe),
+			msm_framebuffer_iova(fb, mdp5_kms->id, 1));
+	mdp5_write(mdp5_kms, REG_MDP5_PIPE_SRC2_ADDR(pipe),
+			msm_framebuffer_iova(fb, mdp5_kms->id, 2));
+	mdp5_write(mdp5_kms, REG_MDP5_PIPE_SRC3_ADDR(pipe),
+			msm_framebuffer_iova(fb, mdp5_kms->id, 4));
 
+	plane->fb = fb;
 }
 
-int mdp5_plane_mode_set(struct drm_plane *plane,
+static int mdp5_plane_mode_set(struct drm_plane *plane,
 		struct drm_crtc *crtc, struct drm_framebuffer *fb,
 		int crtc_x, int crtc_y,
 		unsigned int crtc_w, unsigned int crtc_h,
@@ -225,7 +291,8 @@ int mdp5_plane_mode_set(struct drm_plane *plane,
 	uint32_t nplanes, config = 0;
 	uint32_t phasex_step = 0, phasey_step = 0;
 	uint32_t hdecm = 0, vdecm = 0;
-	int i, nblks;
+	unsigned long flags;
+	int ret;
 
 	nplanes = drm_format_num_planes(fb->pixel_format);
 
@@ -243,12 +310,11 @@ int mdp5_plane_mode_set(struct drm_plane *plane,
 			fb->base.id, src_x, src_y, src_w, src_h,
 			crtc->base.id, crtc_x, crtc_y, crtc_w, crtc_h);
 
-	/*
-	 * Calculate and request required # of smp blocks:
-	 */
-	nblks = request_smp_blocks(plane, fb->pixel_format, nplanes, src_w);
-	if (nblks < 0)
-		return nblks;
+	/* Request some memory from the SMP: */
+	ret = mdp5_smp_request(mdp5_kms->smp,
+			mdp5_plane->pipe, fb->pixel_format, src_w);
+	if (ret)
+		return ret;
 
 	/*
 	 * Currently we update the hw for allocations/requests immediately,
@@ -256,8 +322,7 @@ int mdp5_plane_mode_set(struct drm_plane *plane,
 	 * would move into atomic->check_plane_state(), while updating the
 	 * hw would remain here:
 	 */
-	for (i = 0; i < pipe2nclients(pipe); i++)
-		mdp5_smp_configure(mdp5_kms, pipe2client(pipe, i));
+	mdp5_smp_configure(mdp5_kms->smp, pipe);
 
 	if (src_w != crtc_w) {
 		config |= MDP5_PIPE_SCALE_CONFIG_SCALEX_EN;
@@ -269,6 +334,8 @@ int mdp5_plane_mode_set(struct drm_plane *plane,
 		/* TODO calc phasey_step, vdecm */
 	}
 
+	spin_lock_irqsave(&mdp5_plane->pipe_lock, flags);
+
 	mdp5_write(mdp5_kms, REG_MDP5_PIPE_SRC_IMG_SIZE(pipe),
 			MDP5_PIPE_SRC_IMG_SIZE_WIDTH(src_w) |
 			MDP5_PIPE_SRC_IMG_SIZE_HEIGHT(src_h));
@@ -289,8 +356,6 @@ int mdp5_plane_mode_set(struct drm_plane *plane,
 			MDP5_PIPE_OUT_XY_X(crtc_x) |
 			MDP5_PIPE_OUT_XY_Y(crtc_y));
 
-	mdp5_plane_set_scanout(plane, fb);
-
 	format = to_mdp_format(msm_framebuffer_format(fb));
 
 	mdp5_write(mdp5_kms, REG_MDP5_PIPE_SRC_FORMAT(pipe),
@@ -330,22 +395,24 @@ int mdp5_plane_mode_set(struct drm_plane *plane,
 			MDP5_PIPE_SCALE_CONFIG_SCALEX_MAX_FILTER(SCALE_FILTER_NEAREST) |
 			MDP5_PIPE_SCALE_CONFIG_SCALEY_MAX_FILTER(SCALE_FILTER_NEAREST));
 
-	set_fifo_thresholds(plane, nblks);
+	set_scanout_locked(plane, fb);
 
-	/* TODO detach from old crtc (if we had more than one) */
-	mdp5_crtc_attach(crtc, plane);
+	spin_unlock_irqrestore(&mdp5_plane->pipe_lock, flags);
 
-	return 0;
+	return ret;
 }
 
 void mdp5_plane_complete_flip(struct drm_plane *plane)
 {
 	struct mdp5_kms *mdp5_kms = get_kms(plane);
-	enum mdp5_pipe pipe = to_mdp5_plane(plane)->pipe;
-	int i;
+	struct mdp5_plane *mdp5_plane = to_mdp5_plane(plane);
+	enum mdp5_pipe pipe = mdp5_plane->pipe;
+
+	DBG("%s: complete flip", mdp5_plane->name);
 
-	for (i = 0; i < pipe2nclients(pipe); i++)
-		mdp5_smp_commit(mdp5_kms, pipe2client(pipe, i));
+	mdp5_smp_commit(mdp5_kms->smp, pipe);
+
+	to_mdp5_plane_state(plane->state)->pending = false;
 }
 
 enum mdp5_pipe mdp5_plane_pipe(struct drm_plane *plane)
@@ -354,9 +421,16 @@ enum mdp5_pipe mdp5_plane_pipe(struct drm_plane *plane)
 	return mdp5_plane->pipe;
 }
 
+uint32_t mdp5_plane_get_flush(struct drm_plane *plane)
+{
+	struct mdp5_plane *mdp5_plane = to_mdp5_plane(plane);
+
+	return mdp5_plane->flush_mask;
+}
+
 /* initialize plane */
 struct drm_plane *mdp5_plane_init(struct drm_device *dev,
-		enum mdp5_pipe pipe, bool private_plane)
+		enum mdp5_pipe pipe, bool private_plane, uint32_t reg_offset)
 {
 	struct drm_plane *plane = NULL;
 	struct mdp5_plane *mdp5_plane;
@@ -377,10 +451,18 @@ struct drm_plane *mdp5_plane_init(struct drm_device *dev,
 	mdp5_plane->nformats = mdp5_get_formats(pipe, mdp5_plane->formats,
 			ARRAY_SIZE(mdp5_plane->formats));
 
+	mdp5_plane->flush_mask = mdp_ctl_flush_mask_pipe(pipe);
+	mdp5_plane->reg_offset = reg_offset;
+	spin_lock_init(&mdp5_plane->pipe_lock);
+
 	type = private_plane ? DRM_PLANE_TYPE_PRIMARY : DRM_PLANE_TYPE_OVERLAY;
-	drm_universal_plane_init(dev, plane, 0xff, &mdp5_plane_funcs,
+	ret = drm_universal_plane_init(dev, plane, 0xff, &mdp5_plane_funcs,
 				 mdp5_plane->formats, mdp5_plane->nformats,
 				 type);
+	if (ret)
+		goto fail;
+
+	drm_plane_helper_add(plane, &mdp5_plane_helper_funcs);
 
 	mdp5_plane_install_properties(plane, &plane->base);
 
diff --git a/drivers/gpu/drm/msm/mdp/mdp5/mdp5_smp.c b/drivers/gpu/drm/msm/mdp/mdp5/mdp5_smp.c
index 2d0236b963a6..bf551885e019 100644
--- a/drivers/gpu/drm/msm/mdp/mdp5/mdp5_smp.c
+++ b/drivers/gpu/drm/msm/mdp/mdp5/mdp5_smp.c
@@ -1,4 +1,5 @@
 /*
+ * Copyright (c) 2014, The Linux Foundation. All rights reserved.
  * Copyright (C) 2013 Red Hat
  * Author: Rob Clark <robdclark@gmail.com>
  *
@@ -29,8 +30,11 @@
  * Based on the size of the attached scanout buffer, a certain # of
  * blocks must be allocated to that client out of the shared pool.
  *
- * For each block, it can be either free, or pending/in-use by a
- * client.  The updates happen in three steps:
+ * In some hw, some blocks are statically allocated for certain pipes
+ * and CANNOT be re-allocated (eg: MMB0 and MMB1 both tied to RGB0).
+ *
+ * For each block that can be dynamically allocated, it can be either
+ * free, or pending/in-use by a client. The updates happen in three steps:
  *
  *  1) mdp5_smp_request():
  *     When plane scanout is setup, calculate required number of
@@ -61,21 +65,68 @@
  * inuse and pending state of all clients..
  */
 
-static DEFINE_SPINLOCK(smp_lock);
+struct mdp5_smp {
+	struct drm_device *dev;
+
+	int blk_cnt;
+	int blk_size;
+
+	spinlock_t state_lock;
+	mdp5_smp_state_t state; /* to track smp allocation amongst pipes: */
+
+	struct mdp5_client_smp_state client_state[CID_MAX];
+};
 
+static inline
+struct mdp5_kms *get_kms(struct mdp5_smp *smp)
+{
+	struct msm_drm_private *priv = smp->dev->dev_private;
+
+	return to_mdp5_kms(to_mdp_kms(priv->kms));
+}
+
+static inline enum mdp5_client_id pipe2client(enum mdp5_pipe pipe, int plane)
+{
+	WARN_ON(plane >= pipe2nclients(pipe));
+	switch (pipe) {
+	case SSPP_VIG0: return CID_VIG0_Y + plane;
+	case SSPP_VIG1: return CID_VIG1_Y + plane;
+	case SSPP_VIG2: return CID_VIG2_Y + plane;
+	case SSPP_RGB0: return CID_RGB0;
+	case SSPP_RGB1: return CID_RGB1;
+	case SSPP_RGB2: return CID_RGB2;
+	case SSPP_DMA0: return CID_DMA0_Y + plane;
+	case SSPP_DMA1: return CID_DMA1_Y + plane;
+	case SSPP_VIG3: return CID_VIG3_Y + plane;
+	case SSPP_RGB3: return CID_RGB3;
+	default:        return CID_UNUSED;
+	}
+}
 
 /* step #1: update # of blocks pending for the client: */
-int mdp5_smp_request(struct mdp5_kms *mdp5_kms,
+static int smp_request_block(struct mdp5_smp *smp,
 		enum mdp5_client_id cid, int nblks)
 {
-	struct mdp5_client_smp_state *ps = &mdp5_kms->smp_client_state[cid];
-	int i, ret, avail, cur_nblks, cnt = mdp5_kms->smp_blk_cnt;
+	struct mdp5_kms *mdp5_kms = get_kms(smp);
+	const struct mdp5_cfg_hw *hw_cfg;
+	struct mdp5_client_smp_state *ps = &smp->client_state[cid];
+	int i, ret, avail, cur_nblks, cnt = smp->blk_cnt;
+	int reserved;
 	unsigned long flags;
 
-	spin_lock_irqsave(&smp_lock, flags);
+	hw_cfg = mdp5_cfg_get_hw_config(mdp5_kms->cfg);
+	reserved = hw_cfg->smp.reserved[cid];
+
+	spin_lock_irqsave(&smp->state_lock, flags);
 
-	avail = cnt - bitmap_weight(mdp5_kms->smp_state, cnt);
+	nblks -= reserved;
+	if (reserved)
+		DBG("%d MMBs allocated (%d reserved)", nblks, reserved);
+
+	avail = cnt - bitmap_weight(smp->state, cnt);
 	if (nblks > avail) {
+		dev_err(mdp5_kms->dev->dev, "out of blks (req=%d > avail=%d)\n",
+				nblks, avail);
 		ret = -ENOSPC;
 		goto fail;
 	}
@@ -84,9 +135,9 @@ int mdp5_smp_request(struct mdp5_kms *mdp5_kms,
 	if (nblks > cur_nblks) {
 		/* grow the existing pending reservation: */
 		for (i = cur_nblks; i < nblks; i++) {
-			int blk = find_first_zero_bit(mdp5_kms->smp_state, cnt);
+			int blk = find_first_zero_bit(smp->state, cnt);
 			set_bit(blk, ps->pending);
-			set_bit(blk, mdp5_kms->smp_state);
+			set_bit(blk, smp->state);
 		}
 	} else {
 		/* shrink the existing pending reservation: */
@@ -98,15 +149,88 @@ int mdp5_smp_request(struct mdp5_kms *mdp5_kms,
 	}
 
 fail:
-	spin_unlock_irqrestore(&smp_lock, flags);
+	spin_unlock_irqrestore(&smp->state_lock, flags);
+	return 0;
+}
+
+static void set_fifo_thresholds(struct mdp5_smp *smp,
+		enum mdp5_pipe pipe, int nblks)
+{
+	struct mdp5_kms *mdp5_kms = get_kms(smp);
+	u32 smp_entries_per_blk = smp->blk_size / (128 / BITS_PER_BYTE);
+	u32 val;
+
+	/* 1/4 of SMP pool that is being fetched */
+	val = (nblks * smp_entries_per_blk) / 4;
+
+	mdp5_write(mdp5_kms, REG_MDP5_PIPE_REQPRIO_FIFO_WM_0(pipe), val * 1);
+	mdp5_write(mdp5_kms, REG_MDP5_PIPE_REQPRIO_FIFO_WM_1(pipe), val * 2);
+	mdp5_write(mdp5_kms, REG_MDP5_PIPE_REQPRIO_FIFO_WM_2(pipe), val * 3);
+}
+
+/*
+ * NOTE: looks like if horizontal decimation is used (if we supported that)
+ * then the width used to calculate SMP block requirements is the post-
+ * decimated width.  Ie. SMP buffering sits downstream of decimation (which
+ * presumably happens during the dma from scanout buffer).
+ */
+int mdp5_smp_request(struct mdp5_smp *smp, enum mdp5_pipe pipe, u32 fmt, u32 width)
+{
+	struct mdp5_kms *mdp5_kms = get_kms(smp);
+	struct drm_device *dev = mdp5_kms->dev;
+	int rev = mdp5_cfg_get_hw_rev(mdp5_kms->cfg);
+	int i, hsub, nplanes, nlines, nblks, ret;
+
+	nplanes = drm_format_num_planes(fmt);
+	hsub = drm_format_horz_chroma_subsampling(fmt);
+
+	/* different if BWC (compressed framebuffer?) enabled: */
+	nlines = 2;
+
+	for (i = 0, nblks = 0; i < nplanes; i++) {
+		int n, fetch_stride, cpp;
+
+		cpp = drm_format_plane_cpp(fmt, i);
+		fetch_stride = width * cpp / (i ? hsub : 1);
+
+		n = DIV_ROUND_UP(fetch_stride * nlines, smp->blk_size);
+
+		/* for hw rev v1.00 */
+		if (rev == 0)
+			n = roundup_pow_of_two(n);
+
+		DBG("%s[%d]: request %d SMP blocks", pipe2name(pipe), i, n);
+		ret = smp_request_block(smp, pipe2client(pipe, i), n);
+		if (ret) {
+			dev_err(dev->dev, "Cannot allocate %d SMP blocks: %d\n",
+					n, ret);
+			return ret;
+		}
+
+		nblks += n;
+	}
+
+	set_fifo_thresholds(smp, pipe, nblks);
+
 	return 0;
 }
 
-static void update_smp_state(struct mdp5_kms *mdp5_kms,
+/* Release SMP blocks for all clients of the pipe */
+void mdp5_smp_release(struct mdp5_smp *smp, enum mdp5_pipe pipe)
+{
+	int i, nblks;
+
+	for (i = 0, nblks = 0; i < pipe2nclients(pipe); i++)
+		smp_request_block(smp, pipe2client(pipe, i), 0);
+	set_fifo_thresholds(smp, pipe, 0);
+}
+
+static void update_smp_state(struct mdp5_smp *smp,
 		enum mdp5_client_id cid, mdp5_smp_state_t *assigned)
 {
-	int cnt = mdp5_kms->smp_blk_cnt;
-	uint32_t blk, val;
+	struct mdp5_kms *mdp5_kms = get_kms(smp);
+	int cnt = smp->blk_cnt;
+	u32 blk, val;
 
 	for_each_set_bit(blk, *assigned, cnt) {
 		int idx = blk / 3;
@@ -135,39 +259,80 @@ static void update_smp_state(struct mdp5_kms *mdp5_kms,
 }
 
 /* step #2: configure hw for union(pending, inuse): */
-void mdp5_smp_configure(struct mdp5_kms *mdp5_kms, enum mdp5_client_id cid)
+void mdp5_smp_configure(struct mdp5_smp *smp, enum mdp5_pipe pipe)
 {
-	struct mdp5_client_smp_state *ps = &mdp5_kms->smp_client_state[cid];
-	int cnt = mdp5_kms->smp_blk_cnt;
+	int cnt = smp->blk_cnt;
 	mdp5_smp_state_t assigned;
+	int i;
 
-	bitmap_or(assigned, ps->inuse, ps->pending, cnt);
-	update_smp_state(mdp5_kms, cid, &assigned);
+	for (i = 0; i < pipe2nclients(pipe); i++) {
+		enum mdp5_client_id cid = pipe2client(pipe, i);
+		struct mdp5_client_smp_state *ps = &smp->client_state[cid];
+
+		bitmap_or(assigned, ps->inuse, ps->pending, cnt);
+		update_smp_state(smp, cid, &assigned);
+	}
 }
 
 /* step #3: after vblank, copy pending -> inuse: */
-void mdp5_smp_commit(struct mdp5_kms *mdp5_kms, enum mdp5_client_id cid)
+void mdp5_smp_commit(struct mdp5_smp *smp, enum mdp5_pipe pipe)
 {
-	struct mdp5_client_smp_state *ps = &mdp5_kms->smp_client_state[cid];
-	int cnt = mdp5_kms->smp_blk_cnt;
+	int cnt = smp->blk_cnt;
 	mdp5_smp_state_t released;
+	int i;
+
+	for (i = 0; i < pipe2nclients(pipe); i++) {
+		enum mdp5_client_id cid = pipe2client(pipe, i);
+		struct mdp5_client_smp_state *ps = &smp->client_state[cid];
+
+		/*
+		 * Figure out if there are any blocks we where previously
+		 * using, which can be released and made available to other
+		 * clients:
+		 */
+		if (bitmap_andnot(released, ps->inuse, ps->pending, cnt)) {
+			unsigned long flags;
+
+			spin_lock_irqsave(&smp->state_lock, flags);
+			/* clear released blocks: */
+			bitmap_andnot(smp->state, smp->state, released, cnt);
+			spin_unlock_irqrestore(&smp->state_lock, flags);
 
-	/*
-	 * Figure out if there are any blocks we where previously
-	 * using, which can be released and made available to other
-	 * clients:
-	 */
-	if (bitmap_andnot(released, ps->inuse, ps->pending, cnt)) {
-		unsigned long flags;
-
-		spin_lock_irqsave(&smp_lock, flags);
-		/* clear released blocks: */
-		bitmap_andnot(mdp5_kms->smp_state, mdp5_kms->smp_state,
-				released, cnt);
-		spin_unlock_irqrestore(&smp_lock, flags);
-
-		update_smp_state(mdp5_kms, CID_UNUSED, &released);
+			update_smp_state(smp, CID_UNUSED, &released);
+		}
+
+		bitmap_copy(ps->inuse, ps->pending, cnt);
 	}
+}
+
+void mdp5_smp_destroy(struct mdp5_smp *smp)
+{
+	kfree(smp);
+}
+
+struct mdp5_smp *mdp5_smp_init(struct drm_device *dev, const struct mdp5_smp_block *cfg)
+{
+	struct mdp5_smp *smp = NULL;
+	int ret;
+
+	smp = kzalloc(sizeof(*smp), GFP_KERNEL);
+	if (unlikely(!smp)) {
+		ret = -ENOMEM;
+		goto fail;
+	}
+
+	smp->dev = dev;
+	smp->blk_cnt = cfg->mmb_count;
+	smp->blk_size = cfg->mmb_size;
+
+	/* statically tied MMBs cannot be re-allocated: */
+	bitmap_copy(smp->state, cfg->reserved_state, smp->blk_cnt);
+	spin_lock_init(&smp->state_lock);
+
+	return smp;
+fail:
+	if (smp)
+		mdp5_smp_destroy(smp);
 
-	bitmap_copy(ps->inuse, ps->pending, cnt);
+	return ERR_PTR(ret);
 }
diff --git a/drivers/gpu/drm/msm/mdp/mdp5/mdp5_smp.h b/drivers/gpu/drm/msm/mdp/mdp5/mdp5_smp.h
index 0ab739e1a1dd..e47179f63585 100644
--- a/drivers/gpu/drm/msm/mdp/mdp5/mdp5_smp.h
+++ b/drivers/gpu/drm/msm/mdp/mdp5/mdp5_smp.h
@@ -1,4 +1,5 @@
 /*
+ * Copyright (c) 2014, The Linux Foundation. All rights reserved.
  * Copyright (C) 2013 Red Hat
  * Author: Rob Clark <robdclark@gmail.com>
  *
@@ -20,22 +21,26 @@
 
 #include "msm_drv.h"
 
-#define MAX_SMP_BLOCKS  22
-#define SMP_BLK_SIZE    4096
-#define SMP_ENTRIES_PER_BLK (SMP_BLK_SIZE / 16)
-
-typedef DECLARE_BITMAP(mdp5_smp_state_t, MAX_SMP_BLOCKS);
-
 struct mdp5_client_smp_state {
 	mdp5_smp_state_t inuse;
 	mdp5_smp_state_t pending;
 };
 
 struct mdp5_kms;
+struct mdp5_smp;
+
+/*
+ * SMP module prototypes:
+ * mdp5_smp_init() returns a SMP @handler,
+ * which is then used to call the other mdp5_smp_*(handler, ...) functions.
+ */
 
-int mdp5_smp_request(struct mdp5_kms *mdp5_kms, enum mdp5_client_id cid, int nblks);
-void mdp5_smp_configure(struct mdp5_kms *mdp5_kms, enum mdp5_client_id cid);
-void mdp5_smp_commit(struct mdp5_kms *mdp5_kms, enum mdp5_client_id cid);
+struct mdp5_smp *mdp5_smp_init(struct drm_device *dev, const struct mdp5_smp_block *cfg);
+void  mdp5_smp_destroy(struct mdp5_smp *smp);
 
+int  mdp5_smp_request(struct mdp5_smp *smp, enum mdp5_pipe pipe, u32 fmt, u32 width);
+void mdp5_smp_configure(struct mdp5_smp *smp, enum mdp5_pipe pipe);
+void mdp5_smp_commit(struct mdp5_smp *smp, enum mdp5_pipe pipe);
+void mdp5_smp_release(struct mdp5_smp *smp, enum mdp5_pipe pipe);
 
 #endif /* __MDP5_SMP_H__ */
diff --git a/drivers/gpu/drm/msm/mdp/mdp_kms.c b/drivers/gpu/drm/msm/mdp/mdp_kms.c
index 03455b64a245..2a731722d840 100644
--- a/drivers/gpu/drm/msm/mdp/mdp_kms.c
+++ b/drivers/gpu/drm/msm/mdp/mdp_kms.c
@@ -42,7 +42,10 @@ static void update_irq(struct mdp_kms *mdp_kms)
 	mdp_kms->funcs->set_irqmask(mdp_kms, irqmask);
 }
 
-static void update_irq_unlocked(struct mdp_kms *mdp_kms)
+/* if an mdp_irq's irqmask has changed, such as when mdp5 crtc<->encoder
+ * link changes, this must be called to figure out the new global irqmask
+ */
+void mdp_irq_update(struct mdp_kms *mdp_kms)
 {
 	unsigned long flags;
 	spin_lock_irqsave(&list_lock, flags);
@@ -122,7 +125,7 @@ void mdp_irq_register(struct mdp_kms *mdp_kms, struct mdp_irq *irq)
 	spin_unlock_irqrestore(&list_lock, flags);
 
 	if (needs_update)
-		update_irq_unlocked(mdp_kms);
+		mdp_irq_update(mdp_kms);
 }
 
 void mdp_irq_unregister(struct mdp_kms *mdp_kms, struct mdp_irq *irq)
@@ -141,5 +144,5 @@ void mdp_irq_unregister(struct mdp_kms *mdp_kms, struct mdp_irq *irq)
 	spin_unlock_irqrestore(&list_lock, flags);
 
 	if (needs_update)
-		update_irq_unlocked(mdp_kms);
+		mdp_irq_update(mdp_kms);
 }
diff --git a/drivers/gpu/drm/msm/mdp/mdp_kms.h b/drivers/gpu/drm/msm/mdp/mdp_kms.h
index 99557b5ad4fd..b268ce95d394 100644
--- a/drivers/gpu/drm/msm/mdp/mdp_kms.h
+++ b/drivers/gpu/drm/msm/mdp/mdp_kms.h
@@ -75,7 +75,7 @@ void mdp_update_vblank_mask(struct mdp_kms *mdp_kms, uint32_t mask, bool enable)
 void mdp_irq_wait(struct mdp_kms *mdp_kms, uint32_t irqmask);
 void mdp_irq_register(struct mdp_kms *mdp_kms, struct mdp_irq *irq);
 void mdp_irq_unregister(struct mdp_kms *mdp_kms, struct mdp_irq *irq);
-
+void mdp_irq_update(struct mdp_kms *mdp_kms);
 
 /*
  * pixel format helpers:
diff --git a/drivers/gpu/drm/msm/msm_atomic.c b/drivers/gpu/drm/msm/msm_atomic.c
new file mode 100644
index 000000000000..191968256c58
--- /dev/null
+++ b/drivers/gpu/drm/msm/msm_atomic.c
@@ -0,0 +1,230 @@
+/*
+ * Copyright (C) 2014 Red Hat
+ * Author: Rob Clark <robdclark@gmail.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published by
+ * the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along with
+ * this program.  If not, see <http://www.gnu.org/licenses/>.
+ */
+
+#include "msm_drv.h"
+#include "msm_kms.h"
+#include "msm_gem.h"
+
+struct msm_commit {
+	struct drm_atomic_state *state;
+	uint32_t fence;
+	struct msm_fence_cb fence_cb;
+	uint32_t crtc_mask;
+};
+
+static void fence_cb(struct msm_fence_cb *cb);
+
+/* block until specified crtcs are no longer pending update, and
+ * atomically mark them as pending update
+ */
+static int start_atomic(struct msm_drm_private *priv, uint32_t crtc_mask)
+{
+	int ret;
+
+	spin_lock(&priv->pending_crtcs_event.lock);
+	ret = wait_event_interruptible_locked(priv->pending_crtcs_event,
+			!(priv->pending_crtcs & crtc_mask));
+	if (ret == 0) {
+		DBG("start: %08x", crtc_mask);
+		priv->pending_crtcs |= crtc_mask;
+	}
+	spin_unlock(&priv->pending_crtcs_event.lock);
+
+	return ret;
+}
+
+/* clear specified crtcs (no longer pending update)
+ */
+static void end_atomic(struct msm_drm_private *priv, uint32_t crtc_mask)
+{
+	spin_lock(&priv->pending_crtcs_event.lock);
+	DBG("end: %08x", crtc_mask);
+	priv->pending_crtcs &= ~crtc_mask;
+	wake_up_all_locked(&priv->pending_crtcs_event);
+	spin_unlock(&priv->pending_crtcs_event.lock);
+}
+
+static struct msm_commit *new_commit(struct drm_atomic_state *state)
+{
+	struct msm_commit *c = kzalloc(sizeof(*c), GFP_KERNEL);
+
+	if (!c)
+		return NULL;
+
+	c->state = state;
+	/* TODO we might need a way to indicate to run the cb on a
+	 * different wq so wait_for_vblanks() doesn't block retiring
+	 * bo's..
+	 */
+	INIT_FENCE_CB(&c->fence_cb, fence_cb);
+
+	return c;
+}
+
+/* The (potentially) asynchronous part of the commit.  At this point
+ * nothing can fail short of armageddon.
+ */
+static void complete_commit(struct msm_commit *c)
+{
+	struct drm_atomic_state *state = c->state;
+	struct drm_device *dev = state->dev;
+
+	drm_atomic_helper_commit_pre_planes(dev, state);
+
+	drm_atomic_helper_commit_planes(dev, state);
+
+	drm_atomic_helper_commit_post_planes(dev, state);
+
+	/* NOTE: _wait_for_vblanks() only waits for vblank on
+	 * enabled CRTCs.  So we end up faulting when disabling
+	 * due to (potentially) unref'ing the outgoing fb's
+	 * before the vblank when the disable has latched.
+	 *
+	 * But if it did wait on disabled (or newly disabled)
+	 * CRTCs, that would be racy (ie. we could have missed
+	 * the irq.  We need some way to poll for pipe shut
+	 * down.  Or just live with occasionally hitting the
+	 * timeout in the CRTC disable path (which really should
+	 * not be critical path)
+	 */
+
+	drm_atomic_helper_wait_for_vblanks(dev, state);
+
+	drm_atomic_helper_cleanup_planes(dev, state);
+
+	drm_atomic_state_free(state);
+
+	end_atomic(dev->dev_private, c->crtc_mask);
+
+	kfree(c);
+}
+
+static void fence_cb(struct msm_fence_cb *cb)
+{
+	struct msm_commit *c =
+			container_of(cb, struct msm_commit, fence_cb);
+	complete_commit(c);
+}
+
+static void add_fb(struct msm_commit *c, struct drm_framebuffer *fb)
+{
+	struct drm_gem_object *obj = msm_framebuffer_bo(fb, 0);
+	c->fence = max(c->fence, msm_gem_fence(to_msm_bo(obj), MSM_PREP_READ));
+}
+
+
+/**
+ * drm_atomic_helper_commit - commit validated state object
+ * @dev: DRM device
+ * @state: the driver state object
+ * @async: asynchronous commit
+ *
+ * This function commits a with drm_atomic_helper_check() pre-validated state
+ * object. This can still fail when e.g. the framebuffer reservation fails. For
+ * now this doesn't implement asynchronous commits.
+ *
+ * RETURNS
+ * Zero for success or -errno.
+ */
+int msm_atomic_commit(struct drm_device *dev,
+		struct drm_atomic_state *state, bool async)
+{
+	int nplanes = dev->mode_config.num_total_plane;
+	int ncrtcs = dev->mode_config.num_crtc;
+	struct msm_commit *c;
+	int i, ret;
+
+	ret = drm_atomic_helper_prepare_planes(dev, state);
+	if (ret)
+		return ret;
+
+	c = new_commit(state);
+	if (!c)
+		return -ENOMEM;
+
+	/*
+	 * Figure out what crtcs we have:
+	 */
+	for (i = 0; i < ncrtcs; i++) {
+		struct drm_crtc *crtc = state->crtcs[i];
+		if (!crtc)
+			continue;
+		c->crtc_mask |= (1 << drm_crtc_index(crtc));
+	}
+
+	/*
+	 * Figure out what fence to wait for:
+	 */
+	for (i = 0; i < nplanes; i++) {
+		struct drm_plane *plane = state->planes[i];
+		struct drm_plane_state *new_state = state->plane_states[i];
+
+		if (!plane)
+			continue;
+
+		if ((plane->state->fb != new_state->fb) && new_state->fb)
+			add_fb(c, new_state->fb);
+	}
+
+	/*
+	 * Wait for pending updates on any of the same crtc's and then
+	 * mark our set of crtc's as busy:
+	 */
+	ret = start_atomic(dev->dev_private, c->crtc_mask);
+	if (ret)
+		return ret;
+
+	/*
+	 * This is the point of no return - everything below never fails except
+	 * when the hw goes bonghits. Which means we can commit the new state on
+	 * the software side now.
+	 */
+
+	drm_atomic_helper_swap_state(dev, state);
+
+	/*
+	 * Everything below can be run asynchronously without the need to grab
+	 * any modeset locks at all under one conditions: It must be guaranteed
+	 * that the asynchronous work has either been cancelled (if the driver
+	 * supports it, which at least requires that the framebuffers get
+	 * cleaned up with drm_atomic_helper_cleanup_planes()) or completed
+	 * before the new state gets committed on the software side with
+	 * drm_atomic_helper_swap_state().
+	 *
+	 * This scheme allows new atomic state updates to be prepared and
+	 * checked in parallel to the asynchronous completion of the previous
+	 * update. Which is important since compositors need to figure out the
+	 * composition of the next frame right after having submitted the
+	 * current layout.
+	 */
+
+	if (async) {
+		msm_queue_fence_cb(dev, &c->fence_cb, c->fence);
+		return 0;
+	}
+
+	ret = msm_wait_fence_interruptable(dev, c->fence, NULL);
+	if (ret) {
+		WARN_ON(ret);  // TODO unswap state back?  or??
+		kfree(c);
+		return ret;
+	}
+
+	complete_commit(c);
+
+	return 0;
+}
diff --git a/drivers/gpu/drm/msm/msm_drv.c b/drivers/gpu/drm/msm/msm_drv.c
index fcf95680413d..9a61546a0b05 100644
--- a/drivers/gpu/drm/msm/msm_drv.c
+++ b/drivers/gpu/drm/msm/msm_drv.c
@@ -29,6 +29,8 @@ static void msm_fb_output_poll_changed(struct drm_device *dev)
 static const struct drm_mode_config_funcs mode_config_funcs = {
 	.fb_create = msm_framebuffer_create,
 	.output_poll_changed = msm_fb_output_poll_changed,
+	.atomic_check = drm_atomic_helper_check,
+	.atomic_commit = msm_atomic_commit,
 };
 
 int msm_register_mmu(struct drm_device *dev, struct msm_mmu *mmu)
@@ -191,6 +193,7 @@ static int msm_load(struct drm_device *dev, unsigned long flags)
 
 	priv->wq = alloc_ordered_workqueue("msm", 0);
 	init_waitqueue_head(&priv->fence_event);
+	init_waitqueue_head(&priv->pending_crtcs_event);
 
 	INIT_LIST_HEAD(&priv->inactive_list);
 	INIT_LIST_HEAD(&priv->fence_cbs);
@@ -280,7 +283,7 @@ static int msm_load(struct drm_device *dev, unsigned long flags)
 	dev->mode_config.max_height = 2048;
 	dev->mode_config.funcs = &mode_config_funcs;
 
-	ret = drm_vblank_init(dev, 1);
+	ret = drm_vblank_init(dev, priv->num_crtcs);
 	if (ret < 0) {
 		dev_err(dev->dev, "failed to initialize vblank\n");
 		goto fail;
@@ -294,6 +297,8 @@ static int msm_load(struct drm_device *dev, unsigned long flags)
 		goto fail;
 	}
 
+	drm_mode_config_reset(dev);
+
 #ifdef CONFIG_DRM_MSM_FBDEV
 	priv->fbdev = msm_fbdev_init(dev);
 #endif
@@ -315,39 +320,12 @@ static void load_gpu(struct drm_device *dev)
 {
 	static DEFINE_MUTEX(init_lock);
 	struct msm_drm_private *priv = dev->dev_private;
-	struct msm_gpu *gpu;
 
 	mutex_lock(&init_lock);
 
-	if (priv->gpu)
-		goto out;
-
-	gpu = a3xx_gpu_init(dev);
-	if (IS_ERR(gpu)) {
-		dev_warn(dev->dev, "failed to load a3xx gpu\n");
-		gpu = NULL;
-		/* not fatal */
-	}
-
-	if (gpu) {
-		int ret;
-		mutex_lock(&dev->struct_mutex);
-		gpu->funcs->pm_resume(gpu);
-		mutex_unlock(&dev->struct_mutex);
-		ret = gpu->funcs->hw_init(gpu);
-		if (ret) {
-			dev_err(dev->dev, "gpu hw init failed: %d\n", ret);
-			gpu->funcs->destroy(gpu);
-			gpu = NULL;
-		} else {
-			/* give inactive pm a chance to kick in: */
-			msm_gpu_retire(gpu);
-		}
-	}
-
-	priv->gpu = gpu;
+	if (!priv->gpu)
+		priv->gpu = adreno_load_gpu(dev);
 
-out:
 	mutex_unlock(&init_lock);
 }
 
@@ -646,6 +624,26 @@ int msm_wait_fence_interruptable(struct drm_device *dev, uint32_t fence,
 	return ret;
 }
 
+int msm_queue_fence_cb(struct drm_device *dev,
+		struct msm_fence_cb *cb, uint32_t fence)
+{
+	struct msm_drm_private *priv = dev->dev_private;
+	int ret = 0;
+
+	mutex_lock(&dev->struct_mutex);
+	if (!list_empty(&cb->work.entry)) {
+		ret = -EINVAL;
+	} else if (fence > priv->completed_fence) {
+		cb->fence = fence;
+		list_add_tail(&cb->work.entry, &priv->fence_cbs);
+	} else {
+		queue_work(priv->wq, &cb->work);
+	}
+	mutex_unlock(&dev->struct_mutex);
+
+	return ret;
+}
+
 /* called from workqueue */
 void msm_update_fence(struct drm_device *dev, uint32_t fence)
 {
@@ -836,6 +834,7 @@ static struct drm_driver msm_driver = {
 	.open               = msm_open,
 	.preclose           = msm_preclose,
 	.lastclose          = msm_lastclose,
+	.set_busid          = drm_platform_set_busid,
 	.irq_handler        = msm_irq,
 	.irq_preinstall     = msm_irq_preinstall,
 	.irq_postinstall    = msm_irq_postinstall,
@@ -858,6 +857,7 @@ static struct drm_driver msm_driver = {
 	.gem_prime_import_sg_table = msm_gem_prime_import_sg_table,
 	.gem_prime_vmap     = msm_gem_prime_vmap,
 	.gem_prime_vunmap   = msm_gem_prime_vunmap,
+	.gem_prime_mmap     = msm_gem_prime_mmap,
 #ifdef CONFIG_DEBUG_FS
 	.debugfs_init       = msm_debugfs_init,
 	.debugfs_cleanup    = msm_debugfs_cleanup,
@@ -1013,7 +1013,6 @@ static struct platform_driver msm_platform_driver = {
 	.probe      = msm_pdev_probe,
 	.remove     = msm_pdev_remove,
 	.driver     = {
-		.owner  = THIS_MODULE,
 		.name   = "msm",
 		.of_match_table = dt_match,
 		.pm     = &msm_pm_ops,
@@ -1025,7 +1024,7 @@ static int __init msm_drm_register(void)
 {
 	DBG("init");
 	hdmi_register();
-	a3xx_register();
+	adreno_register();
 	return platform_driver_register(&msm_platform_driver);
 }
 
@@ -1034,7 +1033,7 @@ static void __exit msm_drm_unregister(void)
 	DBG("fini");
 	platform_driver_unregister(&msm_platform_driver);
 	hdmi_unregister();
-	a3xx_unregister();
+	adreno_unregister();
 }
 
 module_init(msm_drm_register);
diff --git a/drivers/gpu/drm/msm/msm_drv.h b/drivers/gpu/drm/msm/msm_drv.h
index 8a2c5fd0893e..b69ef2d5a26c 100644
--- a/drivers/gpu/drm/msm/msm_drv.h
+++ b/drivers/gpu/drm/msm/msm_drv.h
@@ -32,15 +32,6 @@
 #include <linux/types.h>
 #include <asm/sizes.h>
 
-
-#if defined(CONFIG_COMPILE_TEST) && !defined(CONFIG_ARCH_QCOM)
-/* stubs we need for compile-test: */
-static inline struct device *msm_iommu_get_ctx(const char *ctx_name)
-{
-	return NULL;
-}
-#endif
-
 #ifndef CONFIG_OF
 #include <mach/board.h>
 #include <mach/socinfo.h>
@@ -48,9 +39,13 @@ static inline struct device *msm_iommu_get_ctx(const char *ctx_name)
 #endif
 
 #include <drm/drmP.h>
+#include <drm/drm_atomic.h>
+#include <drm/drm_atomic_helper.h>
 #include <drm/drm_crtc_helper.h>
+#include <drm/drm_plane_helper.h>
 #include <drm/drm_fb_helper.h>
 #include <drm/msm_drm.h>
+#include <drm/drm_gem.h>
 
 struct msm_kms;
 struct msm_gpu;
@@ -74,7 +69,12 @@ struct msm_drm_private {
 	struct msm_kms *kms;
 
 	/* subordinate devices, if present: */
-	struct platform_device *hdmi_pdev, *gpu_pdev;
+	struct platform_device *gpu_pdev;
+
+	/* possibly this should be in the kms component, but it is
+	 * shared by both mdp4 and mdp5..
+	 */
+	struct hdmi *hdmi;
 
 	/* when we have more than one 'msm_gpu' these need to be an array: */
 	struct msm_gpu *gpu;
@@ -96,6 +96,10 @@ struct msm_drm_private {
 	/* callbacks deferred until bo is inactive: */
 	struct list_head fence_cbs;
 
+	/* crtcs pending async atomic updates: */
+	uint32_t pending_crtcs;
+	wait_queue_head_t pending_crtcs_event;
+
 	/* registered MMUs: */
 	unsigned int num_mmus;
 	struct msm_mmu *mmus[NUM_DOMAINS];
@@ -144,21 +148,29 @@ void __msm_fence_worker(struct work_struct *work);
 		(_cb)->func = _func;                         \
 	} while (0)
 
+int msm_atomic_commit(struct drm_device *dev,
+		struct drm_atomic_state *state, bool async);
+
 int msm_register_mmu(struct drm_device *dev, struct msm_mmu *mmu);
 
 int msm_wait_fence_interruptable(struct drm_device *dev, uint32_t fence,
 		struct timespec *timeout);
+int msm_queue_fence_cb(struct drm_device *dev,
+		struct msm_fence_cb *cb, uint32_t fence);
 void msm_update_fence(struct drm_device *dev, uint32_t fence);
 
 int msm_ioctl_gem_submit(struct drm_device *dev, void *data,
 		struct drm_file *file);
 
+int msm_gem_mmap_obj(struct drm_gem_object *obj,
+			struct vm_area_struct *vma);
 int msm_gem_mmap(struct file *filp, struct vm_area_struct *vma);
 int msm_gem_fault(struct vm_area_struct *vma, struct vm_fault *vmf);
 uint64_t msm_gem_mmap_offset(struct drm_gem_object *obj);
 int msm_gem_get_iova_locked(struct drm_gem_object *obj, int id,
 		uint32_t *iova);
 int msm_gem_get_iova(struct drm_gem_object *obj, int id, uint32_t *iova);
+uint32_t msm_gem_iova(struct drm_gem_object *obj, int id);
 struct page **msm_gem_get_pages(struct drm_gem_object *obj);
 void msm_gem_put_pages(struct drm_gem_object *obj);
 void msm_gem_put_iova(struct drm_gem_object *obj, int id);
@@ -169,8 +181,9 @@ int msm_gem_dumb_map_offset(struct drm_file *file, struct drm_device *dev,
 struct sg_table *msm_gem_prime_get_sg_table(struct drm_gem_object *obj);
 void *msm_gem_prime_vmap(struct drm_gem_object *obj);
 void msm_gem_prime_vunmap(struct drm_gem_object *obj, void *vaddr);
+int msm_gem_prime_mmap(struct drm_gem_object *obj, struct vm_area_struct *vma);
 struct drm_gem_object *msm_gem_prime_import_sg_table(struct drm_device *dev,
-		size_t size, struct sg_table *sg);
+		struct dma_buf_attachment *attach, struct sg_table *sg);
 int msm_gem_prime_pin(struct drm_gem_object *obj);
 void msm_gem_prime_unpin(struct drm_gem_object *obj);
 void *msm_gem_vaddr_locked(struct drm_gem_object *obj);
@@ -191,6 +204,9 @@ struct drm_gem_object *msm_gem_new(struct drm_device *dev,
 struct drm_gem_object *msm_gem_import(struct drm_device *dev,
 		uint32_t size, struct sg_table *sgt);
 
+int msm_framebuffer_prepare(struct drm_framebuffer *fb, int id);
+void msm_framebuffer_cleanup(struct drm_framebuffer *fb, int id);
+uint32_t msm_framebuffer_iova(struct drm_framebuffer *fb, int id, int plane);
 struct drm_gem_object *msm_framebuffer_bo(struct drm_framebuffer *fb, int plane);
 const struct msm_format *msm_framebuffer_format(struct drm_framebuffer *fb);
 struct drm_framebuffer *msm_framebuffer_init(struct drm_device *dev,
@@ -201,8 +217,8 @@ struct drm_framebuffer *msm_framebuffer_create(struct drm_device *dev,
 struct drm_fb_helper *msm_fbdev_init(struct drm_device *dev);
 
 struct hdmi;
-struct hdmi *hdmi_init(struct drm_device *dev, struct drm_encoder *encoder);
-irqreturn_t hdmi_irq(int irq, void *dev_id);
+int hdmi_modeset_init(struct hdmi *hdmi, struct drm_device *dev,
+		struct drm_encoder *encoder);
 void __init hdmi_register(void);
 void __exit hdmi_unregister(void);
 
diff --git a/drivers/gpu/drm/msm/msm_fb.c b/drivers/gpu/drm/msm/msm_fb.c
index 81bafdf19ab3..84dec161d836 100644
--- a/drivers/gpu/drm/msm/msm_fb.c
+++ b/drivers/gpu/drm/msm/msm_fb.c
@@ -24,7 +24,7 @@
 struct msm_framebuffer {
 	struct drm_framebuffer base;
 	const struct msm_format *format;
-	struct drm_gem_object *planes[2];
+	struct drm_gem_object *planes[3];
 };
 #define to_msm_framebuffer(x) container_of(x, struct msm_framebuffer, base)
 
@@ -87,6 +87,44 @@ void msm_framebuffer_describe(struct drm_framebuffer *fb, struct seq_file *m)
 }
 #endif
 
+/* prepare/pin all the fb's bo's for scanout.  Note that it is not valid
+ * to prepare an fb more multiple different initiator 'id's.  But that
+ * should be fine, since only the scanout (mdpN) side of things needs
+ * this, the gpu doesn't care about fb's.
+ */
+int msm_framebuffer_prepare(struct drm_framebuffer *fb, int id)
+{
+	struct msm_framebuffer *msm_fb = to_msm_framebuffer(fb);
+	int ret, i, n = drm_format_num_planes(fb->pixel_format);
+	uint32_t iova;
+
+	for (i = 0; i < n; i++) {
+		ret = msm_gem_get_iova(msm_fb->planes[i], id, &iova);
+		DBG("FB[%u]: iova[%d]: %08x (%d)", fb->base.id, i, iova, ret);
+		if (ret)
+			return ret;
+	}
+
+	return 0;
+}
+
+void msm_framebuffer_cleanup(struct drm_framebuffer *fb, int id)
+{
+	struct msm_framebuffer *msm_fb = to_msm_framebuffer(fb);
+	int i, n = drm_format_num_planes(fb->pixel_format);
+
+	for (i = 0; i < n; i++)
+		msm_gem_put_iova(msm_fb->planes[i], id);
+}
+
+uint32_t msm_framebuffer_iova(struct drm_framebuffer *fb, int id, int plane)
+{
+	struct msm_framebuffer *msm_fb = to_msm_framebuffer(fb);
+	if (!msm_fb->planes[plane])
+		return 0;
+	return msm_gem_iova(msm_fb->planes[plane], id);
+}
+
 struct drm_gem_object *msm_framebuffer_bo(struct drm_framebuffer *fb, int plane)
 {
 	struct msm_framebuffer *msm_fb = to_msm_framebuffer(fb);
@@ -166,6 +204,11 @@ struct drm_framebuffer *msm_framebuffer_init(struct drm_device *dev,
 
 	msm_fb->format = format;
 
+	if (n > ARRAY_SIZE(msm_fb->planes)) {
+		ret = -EINVAL;
+		goto fail;
+	}
+
 	for (i = 0; i < n; i++) {
 		unsigned int width = mode_cmd->width / (i ? hsub : 1);
 		unsigned int height = mode_cmd->height / (i ? vsub : 1);
diff --git a/drivers/gpu/drm/msm/msm_fbdev.c b/drivers/gpu/drm/msm/msm_fbdev.c
index ab5bfd2d0ebf..1f3af13ccede 100644
--- a/drivers/gpu/drm/msm/msm_fbdev.c
+++ b/drivers/gpu/drm/msm/msm_fbdev.c
@@ -93,9 +93,6 @@ static int msm_fbdev_create(struct drm_fb_helper *helper,
 	uint32_t paddr;
 	int ret, size;
 
-	sizes->surface_bpp = 32;
-	sizes->surface_depth = 24;
-
 	DBG("create fbdev: %dx%d@%d (%dx%d)", sizes->surface_width,
 			sizes->surface_height, sizes->surface_bpp,
 			sizes->fb_width, sizes->fb_height);
@@ -193,8 +190,7 @@ fail_unlock:
 fail:
 
 	if (ret) {
-		if (fbi)
-			framebuffer_release(fbi);
+		framebuffer_release(fbi);
 		if (fb) {
 			drm_framebuffer_unregister_private(fb);
 			drm_framebuffer_remove(fb);
diff --git a/drivers/gpu/drm/msm/msm_gem.c b/drivers/gpu/drm/msm/msm_gem.c
index 4b1b82adabde..49dea4fb55ac 100644
--- a/drivers/gpu/drm/msm/msm_gem.c
+++ b/drivers/gpu/drm/msm/msm_gem.c
@@ -309,6 +309,7 @@ int msm_gem_get_iova_locked(struct drm_gem_object *obj, int id,
 	return ret;
 }
 
+/* get iova, taking a reference.  Should have a matching put */
 int msm_gem_get_iova(struct drm_gem_object *obj, int id, uint32_t *iova)
 {
 	struct msm_gem_object *msm_obj = to_msm_bo(obj);
@@ -328,6 +329,16 @@ int msm_gem_get_iova(struct drm_gem_object *obj, int id, uint32_t *iova)
 	return ret;
 }
 
+/* get iova without taking a reference, used in places where you have
+ * already done a 'msm_gem_get_iova()'.
+ */
+uint32_t msm_gem_iova(struct drm_gem_object *obj, int id)
+{
+	struct msm_gem_object *msm_obj = to_msm_bo(obj);
+	WARN_ON(!msm_obj->domain[id].iova);
+	return msm_obj->domain[id].iova;
+}
+
 void msm_gem_put_iova(struct drm_gem_object *obj, int id)
 {
 	// XXX TODO ..
@@ -397,23 +408,10 @@ void *msm_gem_vaddr(struct drm_gem_object *obj)
 int msm_gem_queue_inactive_cb(struct drm_gem_object *obj,
 		struct msm_fence_cb *cb)
 {
-	struct drm_device *dev = obj->dev;
-	struct msm_drm_private *priv = dev->dev_private;
 	struct msm_gem_object *msm_obj = to_msm_bo(obj);
-	int ret = 0;
-
-	mutex_lock(&dev->struct_mutex);
-	if (!list_empty(&cb->work.entry)) {
-		ret = -EINVAL;
-	} else if (is_active(msm_obj)) {
-		cb->fence = max(msm_obj->read_fence, msm_obj->write_fence);
-		list_add_tail(&cb->work.entry, &priv->fence_cbs);
-	} else {
-		queue_work(priv->wq, &cb->work);
-	}
-	mutex_unlock(&dev->struct_mutex);
-
-	return ret;
+	uint32_t fence = msm_gem_fence(msm_obj,
+			MSM_PREP_READ | MSM_PREP_WRITE);
+	return msm_queue_fence_cb(obj->dev, cb, fence);
 }
 
 void msm_gem_move_to_active(struct drm_gem_object *obj,
@@ -452,12 +450,8 @@ int msm_gem_cpu_prep(struct drm_gem_object *obj, uint32_t op,
 	int ret = 0;
 
 	if (is_active(msm_obj)) {
-		uint32_t fence = 0;
+		uint32_t fence = msm_gem_fence(msm_obj, op);
 
-		if (op & MSM_PREP_READ)
-			fence = msm_obj->write_fence;
-		if (op & MSM_PREP_WRITE)
-			fence = max(fence, msm_obj->read_fence);
 		if (op & MSM_PREP_NOSYNC)
 			timeout = NULL;
 
@@ -525,13 +519,11 @@ void msm_gem_free_object(struct drm_gem_object *obj)
 	for (id = 0; id < ARRAY_SIZE(msm_obj->domain); id++) {
 		struct msm_mmu *mmu = priv->mmus[id];
 		if (mmu && msm_obj->domain[id].iova) {
-			uint32_t offset = (uint32_t)mmap_offset(obj);
+			uint32_t offset = msm_obj->domain[id].iova;
 			mmu->funcs->unmap(mmu, offset, msm_obj->sgt, obj->size);
 		}
 	}
 
-	drm_gem_free_mmap_offset(obj);
-
 	if (obj->import_attach) {
 		if (msm_obj->vaddr)
 			dma_buf_vunmap(obj->import_attach->dmabuf, msm_obj->vaddr);
@@ -543,8 +535,7 @@ void msm_gem_free_object(struct drm_gem_object *obj)
 			drm_free_large(msm_obj->pages);
 
 	} else {
-		if (msm_obj->vaddr)
-			vunmap(msm_obj->vaddr);
+		vunmap(msm_obj->vaddr);
 		put_pages(obj);
 	}
 
diff --git a/drivers/gpu/drm/msm/msm_gem.h b/drivers/gpu/drm/msm/msm_gem.h
index bfb052688f8e..8fbbd0594c46 100644
--- a/drivers/gpu/drm/msm/msm_gem.h
+++ b/drivers/gpu/drm/msm/msm_gem.h
@@ -70,6 +70,19 @@ static inline bool is_active(struct msm_gem_object *msm_obj)
 	return msm_obj->gpu != NULL;
 }
 
+static inline uint32_t msm_gem_fence(struct msm_gem_object *msm_obj,
+		uint32_t op)
+{
+	uint32_t fence = 0;
+
+	if (op & MSM_PREP_READ)
+		fence = msm_obj->write_fence;
+	if (op & MSM_PREP_WRITE)
+		fence = max(fence, msm_obj->read_fence);
+
+	return fence;
+}
+
 #define MAX_CMDS 4
 
 /* Created per submit-ioctl, to track bo's and cmdstream bufs, etc,
diff --git a/drivers/gpu/drm/msm/msm_gem_prime.c b/drivers/gpu/drm/msm/msm_gem_prime.c
index d48f9fc5129b..dd7a7ab603e2 100644
--- a/drivers/gpu/drm/msm/msm_gem_prime.c
+++ b/drivers/gpu/drm/msm/msm_gem_prime.c
@@ -18,6 +18,7 @@
 #include "msm_drv.h"
 #include "msm_gem.h"
 
+#include <linux/dma-buf.h>
 
 struct sg_table *msm_gem_prime_get_sg_table(struct drm_gem_object *obj)
 {
@@ -36,10 +37,23 @@ void msm_gem_prime_vunmap(struct drm_gem_object *obj, void *vaddr)
 	/* TODO msm_gem_vunmap() */
 }
 
+int msm_gem_prime_mmap(struct drm_gem_object *obj, struct vm_area_struct *vma)
+{
+	int ret;
+
+	mutex_lock(&obj->dev->struct_mutex);
+	ret = drm_gem_mmap_obj(obj, obj->size, vma);
+	mutex_unlock(&obj->dev->struct_mutex);
+	if (ret < 0)
+		return ret;
+
+	return msm_gem_mmap_obj(vma->vm_private_data, vma);
+}
+
 struct drm_gem_object *msm_gem_prime_import_sg_table(struct drm_device *dev,
-		size_t size, struct sg_table *sg)
+		struct dma_buf_attachment *attach, struct sg_table *sg)
 {
-	return msm_gem_import(dev, size, sg);
+	return msm_gem_import(dev, attach->dmabuf->size, sg);
 }
 
 int msm_gem_prime_pin(struct drm_gem_object *obj)
diff --git a/drivers/gpu/drm/msm/msm_gpu.h b/drivers/gpu/drm/msm/msm_gpu.h
index 9b579b792840..fd1e4b4a6d40 100644
--- a/drivers/gpu/drm/msm/msm_gpu.h
+++ b/drivers/gpu/drm/msm/msm_gpu.h
@@ -166,8 +166,8 @@ int msm_gpu_init(struct drm_device *drm, struct platform_device *pdev,
 		const char *name, const char *ioname, const char *irqname, int ringsz);
 void msm_gpu_cleanup(struct msm_gpu *gpu);
 
-struct msm_gpu *a3xx_gpu_init(struct drm_device *dev);
-void __init a3xx_register(void);
-void __exit a3xx_unregister(void);
+struct msm_gpu *adreno_load_gpu(struct drm_device *dev);
+void __init adreno_register(void);
+void __exit adreno_unregister(void);
 
 #endif /* __MSM_GPU_H__ */
diff --git a/drivers/gpu/drm/nouveau/Makefile b/drivers/gpu/drm/nouveau/Makefile
index f5d7f7ce4bc6..6461e3565afe 100644
--- a/drivers/gpu/drm/nouveau/Makefile
+++ b/drivers/gpu/drm/nouveau/Makefile
@@ -38,19 +38,33 @@ nouveau-y += core/subdev/bios/dcb.o
 nouveau-y += core/subdev/bios/disp.o
 nouveau-y += core/subdev/bios/dp.o
 nouveau-y += core/subdev/bios/extdev.o
+nouveau-y += core/subdev/bios/fan.o
 nouveau-y += core/subdev/bios/gpio.o
 nouveau-y += core/subdev/bios/i2c.o
+nouveau-y += core/subdev/bios/image.o
 nouveau-y += core/subdev/bios/init.o
 nouveau-y += core/subdev/bios/mxm.o
+nouveau-y += core/subdev/bios/npde.o
+nouveau-y += core/subdev/bios/pcir.o
 nouveau-y += core/subdev/bios/perf.o
 nouveau-y += core/subdev/bios/pll.o
+nouveau-y += core/subdev/bios/pmu.o
 nouveau-y += core/subdev/bios/ramcfg.o
 nouveau-y += core/subdev/bios/rammap.o
+nouveau-y += core/subdev/bios/shadow.o
+nouveau-y += core/subdev/bios/shadowacpi.o
+nouveau-y += core/subdev/bios/shadowof.o
+nouveau-y += core/subdev/bios/shadowpci.o
+nouveau-y += core/subdev/bios/shadowramin.o
+nouveau-y += core/subdev/bios/shadowrom.o
 nouveau-y += core/subdev/bios/timing.o
 nouveau-y += core/subdev/bios/therm.o
 nouveau-y += core/subdev/bios/vmap.o
 nouveau-y += core/subdev/bios/volt.o
 nouveau-y += core/subdev/bios/xpio.o
+nouveau-y += core/subdev/bios/M0203.o
+nouveau-y += core/subdev/bios/M0205.o
+nouveau-y += core/subdev/bios/M0209.o
 nouveau-y += core/subdev/bios/P0260.o
 nouveau-y += core/subdev/bus/hwsq.o
 nouveau-y += core/subdev/bus/nv04.o
@@ -83,6 +97,7 @@ nouveau-y += core/subdev/devinit/nva3.o
 nouveau-y += core/subdev/devinit/nvaf.o
 nouveau-y += core/subdev/devinit/nvc0.o
 nouveau-y += core/subdev/devinit/gm107.o
+nouveau-y += core/subdev/devinit/gm204.o
 nouveau-y += core/subdev/fb/base.o
 nouveau-y += core/subdev/fb/nv04.o
 nouveau-y += core/subdev/fb/nv10.o
@@ -124,12 +139,18 @@ nouveau-y += core/subdev/fb/ramnvc0.o
 nouveau-y += core/subdev/fb/ramnve0.o
 nouveau-y += core/subdev/fb/ramgk20a.o
 nouveau-y += core/subdev/fb/ramgm107.o
+nouveau-y += core/subdev/fb/sddr2.o
 nouveau-y += core/subdev/fb/sddr3.o
+nouveau-y += core/subdev/fb/gddr3.o
 nouveau-y += core/subdev/fb/gddr5.o
+nouveau-y += core/subdev/fuse/base.o
+nouveau-y += core/subdev/fuse/g80.o
+nouveau-y += core/subdev/fuse/gf100.o
+nouveau-y += core/subdev/fuse/gm107.o
 nouveau-y += core/subdev/gpio/base.o
 nouveau-y += core/subdev/gpio/nv10.o
 nouveau-y += core/subdev/gpio/nv50.o
-nouveau-y += core/subdev/gpio/nv92.o
+nouveau-y += core/subdev/gpio/nv94.o
 nouveau-y += core/subdev/gpio/nvd0.o
 nouveau-y += core/subdev/gpio/nve0.o
 nouveau-y += core/subdev/i2c/base.o
@@ -139,6 +160,7 @@ nouveau-y += core/subdev/i2c/bit.o
 nouveau-y += core/subdev/i2c/pad.o
 nouveau-y += core/subdev/i2c/padnv04.o
 nouveau-y += core/subdev/i2c/padnv94.o
+nouveau-y += core/subdev/i2c/padgm204.o
 nouveau-y += core/subdev/i2c/nv04.o
 nouveau-y += core/subdev/i2c/nv4e.o
 nouveau-y += core/subdev/i2c/nv50.o
@@ -146,6 +168,7 @@ nouveau-y += core/subdev/i2c/nv94.o
 nouveau-y += core/subdev/i2c/nvd0.o
 nouveau-y += core/subdev/i2c/gf117.o
 nouveau-y += core/subdev/i2c/nve0.o
+nouveau-y += core/subdev/i2c/gm204.o
 nouveau-y += core/subdev/ibus/nvc0.o
 nouveau-y += core/subdev/ibus/nve0.o
 nouveau-y += core/subdev/ibus/gk20a.o
@@ -190,6 +213,7 @@ nouveau-y += core/subdev/therm/nv50.o
 nouveau-y += core/subdev/therm/nv84.o
 nouveau-y += core/subdev/therm/nva3.o
 nouveau-y += core/subdev/therm/nvd0.o
+nouveau-y += core/subdev/therm/gm107.o
 nouveau-y += core/subdev/timer/base.o
 nouveau-y += core/subdev/timer/nv04.o
 nouveau-y += core/subdev/timer/gk20a.o
@@ -202,6 +226,7 @@ nouveau-y += core/subdev/vm/nvc0.o
 nouveau-y += core/subdev/volt/base.o
 nouveau-y += core/subdev/volt/gpio.o
 nouveau-y += core/subdev/volt/nv40.o
+nouveau-y += core/subdev/volt/gk20a.o
 
 nouveau-y += core/engine/falcon.o
 nouveau-y += core/engine/xtensa.o
@@ -245,6 +270,7 @@ nouveau-y += core/engine/disp/nvd0.o
 nouveau-y += core/engine/disp/nve0.o
 nouveau-y += core/engine/disp/nvf0.o
 nouveau-y += core/engine/disp/gm107.o
+nouveau-y += core/engine/disp/gm204.o
 nouveau-y += core/engine/disp/dacnv50.o
 nouveau-y += core/engine/disp/dport.o
 nouveau-y += core/engine/disp/hdanva3.o
@@ -252,10 +278,12 @@ nouveau-y += core/engine/disp/hdanvd0.o
 nouveau-y += core/engine/disp/hdminv84.o
 nouveau-y += core/engine/disp/hdminva3.o
 nouveau-y += core/engine/disp/hdminvd0.o
+nouveau-y += core/engine/disp/hdminve0.o
 nouveau-y += core/engine/disp/piornv50.o
 nouveau-y += core/engine/disp/sornv50.o
 nouveau-y += core/engine/disp/sornv94.o
 nouveau-y += core/engine/disp/sornvd0.o
+nouveau-y += core/engine/disp/sorgm204.o
 nouveau-y += core/engine/disp/vga.o
 nouveau-y += core/engine/fifo/base.o
 nouveau-y += core/engine/fifo/nv04.o
diff --git a/drivers/gpu/drm/nouveau/core/core/client.c b/drivers/gpu/drm/nouveau/core/core/client.c
index 68bf06768123..e962433294c3 100644
--- a/drivers/gpu/drm/nouveau/core/core/client.c
+++ b/drivers/gpu/drm/nouveau/core/core/client.c
@@ -91,9 +91,10 @@ nvkm_client_notify_del(struct nouveau_client *client, int index)
 }
 
 int
-nvkm_client_notify_new(struct nouveau_client *client,
+nvkm_client_notify_new(struct nouveau_object *object,
 		       struct nvkm_event *event, void *data, u32 size)
 {
+	struct nouveau_client *client = nouveau_client(object);
 	struct nvkm_client_notify *notify;
 	union {
 		struct nvif_notify_req_v0 v0;
@@ -127,8 +128,8 @@ nvkm_client_notify_new(struct nouveau_client *client,
 	}
 
 	if (ret == 0) {
-		ret = nvkm_notify_init(event, nvkm_client_notify, false,
-				       data, size, reply, &notify->n);
+		ret = nvkm_notify_init(object, event, nvkm_client_notify,
+				       false, data, size, reply, &notify->n);
 		if (ret == 0) {
 			client->notify[index] = notify;
 			notify->client = client;
diff --git a/drivers/gpu/drm/nouveau/core/core/event.c b/drivers/gpu/drm/nouveau/core/core/event.c
index 0540a48c5678..760947e380c9 100644
--- a/drivers/gpu/drm/nouveau/core/core/event.c
+++ b/drivers/gpu/drm/nouveau/core/core/event.c
@@ -20,13 +20,13 @@
  * OTHER DEALINGS IN THE SOFTWARE.
  */
 
-#include <core/os.h>
+#include <core/object.h>
 #include <core/event.h>
 
 void
 nvkm_event_put(struct nvkm_event *event, u32 types, int index)
 {
-	BUG_ON(!spin_is_locked(&event->refs_lock));
+	assert_spin_locked(&event->refs_lock);
 	while (types) {
 		int type = __ffs(types); types &= ~(1 << type);
 		if (--event->refs[index * event->types_nr + type] == 0) {
@@ -39,7 +39,7 @@ nvkm_event_put(struct nvkm_event *event, u32 types, int index)
 void
 nvkm_event_get(struct nvkm_event *event, u32 types, int index)
 {
-	BUG_ON(!spin_is_locked(&event->refs_lock));
+	assert_spin_locked(&event->refs_lock);
 	while (types) {
 		int type = __ffs(types); types &= ~(1 << type);
 		if (++event->refs[index * event->types_nr + type] == 1) {
diff --git a/drivers/gpu/drm/nouveau/core/core/gpuobj.c b/drivers/gpu/drm/nouveau/core/core/gpuobj.c
index 560b2214cf1c..daee87702502 100644
--- a/drivers/gpu/drm/nouveau/core/core/gpuobj.c
+++ b/drivers/gpu/drm/nouveau/core/core/gpuobj.c
@@ -115,7 +115,7 @@ nouveau_gpuobj_create_(struct nouveau_object *parent,
 	gpuobj->size = size;
 
 	if (heap) {
-		ret = nouveau_mm_head(heap, 1, size, size,
+		ret = nouveau_mm_head(heap, 0, 1, size, size,
 				      max(align, (u32)1), &gpuobj->node);
 		if (ret)
 			return ret;
diff --git a/drivers/gpu/drm/nouveau/core/core/handle.c b/drivers/gpu/drm/nouveau/core/core/handle.c
index a490b805d7e3..13f816cb08bd 100644
--- a/drivers/gpu/drm/nouveau/core/core/handle.c
+++ b/drivers/gpu/drm/nouveau/core/core/handle.c
@@ -222,116 +222,3 @@ nouveau_handle_put(struct nouveau_handle *handle)
 	if (handle)
 		nouveau_namedb_put(handle);
 }
-
-int
-nouveau_handle_new(struct nouveau_object *client, u32 _parent, u32 _handle,
-		   u16 _oclass, void *data, u32 size,
-		   struct nouveau_object **pobject)
-{
-	struct nouveau_object *parent = NULL;
-	struct nouveau_object *engctx = NULL;
-	struct nouveau_object *object = NULL;
-	struct nouveau_object *engine;
-	struct nouveau_oclass *oclass;
-	struct nouveau_handle *handle;
-	int ret;
-
-	/* lookup parent object and ensure it *is* a parent */
-	parent = nouveau_handle_ref(client, _parent);
-	if (!parent) {
-		nv_error(client, "parent 0x%08x not found\n", _parent);
-		return -ENOENT;
-	}
-
-	if (!nv_iclass(parent, NV_PARENT_CLASS)) {
-		nv_error(parent, "cannot have children\n");
-		ret = -EINVAL;
-		goto fail_class;
-	}
-
-	/* check that parent supports the requested subclass */
-	ret = nouveau_parent_sclass(parent, _oclass, &engine, &oclass);
-	if (ret) {
-		nv_debug(parent, "illegal class 0x%04x\n", _oclass);
-		goto fail_class;
-	}
-
-	/* make sure engine init has been completed *before* any objects
-	 * it controls are created - the constructors may depend on
-	 * state calculated at init (ie. default context construction)
-	 */
-	if (engine) {
-		ret = nouveau_object_inc(engine);
-		if (ret)
-			goto fail_class;
-	}
-
-	/* if engine requires it, create a context object to insert
-	 * between the parent and its children (eg. PGRAPH context)
-	 */
-	if (engine && nv_engine(engine)->cclass) {
-		ret = nouveau_object_ctor(parent, engine,
-					  nv_engine(engine)->cclass,
-					  data, size, &engctx);
-		if (ret)
-			goto fail_engctx;
-	} else {
-		nouveau_object_ref(parent, &engctx);
-	}
-
-	/* finally, create new object and bind it to its handle */
-	ret = nouveau_object_ctor(engctx, engine, oclass, data, size, &object);
-	*pobject = object;
-	if (ret)
-		goto fail_ctor;
-
-	ret = nouveau_object_inc(object);
-	if (ret)
-		goto fail_init;
-
-	ret = nouveau_handle_create(parent, _parent, _handle, object, &handle);
-	if (ret)
-		goto fail_handle;
-
-	ret = nouveau_handle_init(handle);
-	if (ret)
-		nouveau_handle_destroy(handle);
-
-fail_handle:
-	nouveau_object_dec(object, false);
-fail_init:
-	nouveau_object_ref(NULL, &object);
-fail_ctor:
-	nouveau_object_ref(NULL, &engctx);
-fail_engctx:
-	if (engine)
-		nouveau_object_dec(engine, false);
-fail_class:
-	nouveau_object_ref(NULL, &parent);
-	return ret;
-}
-
-int
-nouveau_handle_del(struct nouveau_object *client, u32 _parent, u32 _handle)
-{
-	struct nouveau_object *parent = NULL;
-	struct nouveau_object *namedb = NULL;
-	struct nouveau_handle *handle = NULL;
-
-	parent = nouveau_handle_ref(client, _parent);
-	if (!parent)
-		return -ENOENT;
-
-	namedb = nv_pclass(parent, NV_NAMEDB_CLASS);
-	if (namedb) {
-		handle = nouveau_namedb_get(nv_namedb(namedb), _handle);
-		if (handle) {
-			nouveau_namedb_put(handle);
-			nouveau_handle_fini(handle, false);
-			nouveau_handle_destroy(handle);
-		}
-	}
-
-	nouveau_object_ref(NULL, &parent);
-	return handle ? 0 : -EINVAL;
-}
diff --git a/drivers/gpu/drm/nouveau/core/core/ioctl.c b/drivers/gpu/drm/nouveau/core/core/ioctl.c
index f7e19bfb489c..692aa92dd850 100644
--- a/drivers/gpu/drm/nouveau/core/core/ioctl.c
+++ b/drivers/gpu/drm/nouveau/core/core/ioctl.c
@@ -349,7 +349,6 @@ nvkm_ioctl_unmap(struct nouveau_handle *handle, void *data, u32 size)
 static int
 nvkm_ioctl_ntfy_new(struct nouveau_handle *handle, void *data, u32 size)
 {
-	struct nouveau_client *client = nouveau_client(handle->object);
 	struct nouveau_object *object = handle->object;
 	struct nouveau_ofuncs *ofuncs = object->oclass->ofuncs;
 	union {
@@ -365,7 +364,7 @@ nvkm_ioctl_ntfy_new(struct nouveau_handle *handle, void *data, u32 size)
 		if (ret = -ENODEV, ofuncs->ntfy)
 			ret = ofuncs->ntfy(object, args->v0.event, &event);
 		if (ret == 0) {
-			ret = nvkm_client_notify_new(client, event, data, size);
+			ret = nvkm_client_notify_new(object, event, data, size);
 			if (ret >= 0) {
 				args->v0.index = ret;
 				ret = 0;
diff --git a/drivers/gpu/drm/nouveau/core/core/mm.c b/drivers/gpu/drm/nouveau/core/core/mm.c
index 7a4e0891c5f8..b4f5db66d5b5 100644
--- a/drivers/gpu/drm/nouveau/core/core/mm.c
+++ b/drivers/gpu/drm/nouveau/core/core/mm.c
@@ -28,6 +28,24 @@
 #define node(root, dir) ((root)->nl_entry.dir == &mm->nodes) ? NULL : \
 	list_entry((root)->nl_entry.dir, struct nouveau_mm_node, nl_entry)
 
+static void
+nouveau_mm_dump(struct nouveau_mm *mm, const char *header)
+{
+	struct nouveau_mm_node *node;
+
+	printk(KERN_ERR "nouveau: %s\n", header);
+	printk(KERN_ERR "nouveau: node list:\n");
+	list_for_each_entry(node, &mm->nodes, nl_entry) {
+		printk(KERN_ERR "nouveau: \t%08x %08x %d\n",
+		       node->offset, node->length, node->type);
+	}
+	printk(KERN_ERR "nouveau: free list:\n");
+	list_for_each_entry(node, &mm->free, fl_entry) {
+		printk(KERN_ERR "nouveau: \t%08x %08x %d\n",
+		       node->offset, node->length, node->type);
+	}
+}
+
 void
 nouveau_mm_free(struct nouveau_mm *mm, struct nouveau_mm_node **pthis)
 {
@@ -37,29 +55,29 @@ nouveau_mm_free(struct nouveau_mm *mm, struct nouveau_mm_node **pthis)
 		struct nouveau_mm_node *prev = node(this, prev);
 		struct nouveau_mm_node *next = node(this, next);
 
-		if (prev && prev->type == 0) {
+		if (prev && prev->type == NVKM_MM_TYPE_NONE) {
 			prev->length += this->length;
 			list_del(&this->nl_entry);
 			kfree(this); this = prev;
 		}
 
-		if (next && next->type == 0) {
+		if (next && next->type == NVKM_MM_TYPE_NONE) {
 			next->offset  = this->offset;
 			next->length += this->length;
-			if (this->type == 0)
+			if (this->type == NVKM_MM_TYPE_NONE)
 				list_del(&this->fl_entry);
 			list_del(&this->nl_entry);
 			kfree(this); this = NULL;
 		}
 
-		if (this && this->type != 0) {
+		if (this && this->type != NVKM_MM_TYPE_NONE) {
 			list_for_each_entry(prev, &mm->free, fl_entry) {
 				if (this->offset < prev->offset)
 					break;
 			}
 
 			list_add_tail(&this->fl_entry, &prev->fl_entry);
-			this->type = 0;
+			this->type = NVKM_MM_TYPE_NONE;
 		}
 	}
 
@@ -80,27 +98,32 @@ region_head(struct nouveau_mm *mm, struct nouveau_mm_node *a, u32 size)
 
 	b->offset = a->offset;
 	b->length = size;
+	b->heap   = a->heap;
 	b->type   = a->type;
 	a->offset += size;
 	a->length -= size;
 	list_add_tail(&b->nl_entry, &a->nl_entry);
-	if (b->type == 0)
+	if (b->type == NVKM_MM_TYPE_NONE)
 		list_add_tail(&b->fl_entry, &a->fl_entry);
 	return b;
 }
 
 int
-nouveau_mm_head(struct nouveau_mm *mm, u8 type, u32 size_max, u32 size_min,
-		u32 align, struct nouveau_mm_node **pnode)
+nouveau_mm_head(struct nouveau_mm *mm, u8 heap, u8 type, u32 size_max,
+		u32 size_min, u32 align, struct nouveau_mm_node **pnode)
 {
 	struct nouveau_mm_node *prev, *this, *next;
 	u32 mask = align - 1;
 	u32 splitoff;
 	u32 s, e;
 
-	BUG_ON(!type);
+	BUG_ON(type == NVKM_MM_TYPE_NONE || type == NVKM_MM_TYPE_HOLE);
 
 	list_for_each_entry(this, &mm->free, fl_entry) {
+		if (unlikely(heap != NVKM_MM_HEAP_ANY)) {
+			if (this->heap != heap)
+				continue;
+		}
 		e = this->offset + this->length;
 		s = this->offset;
 
@@ -149,27 +172,32 @@ region_tail(struct nouveau_mm *mm, struct nouveau_mm_node *a, u32 size)
 	a->length -= size;
 	b->offset  = a->offset + a->length;
 	b->length  = size;
+	b->heap    = a->heap;
 	b->type    = a->type;
 
 	list_add(&b->nl_entry, &a->nl_entry);
-	if (b->type == 0)
+	if (b->type == NVKM_MM_TYPE_NONE)
 		list_add(&b->fl_entry, &a->fl_entry);
 	return b;
 }
 
 int
-nouveau_mm_tail(struct nouveau_mm *mm, u8 type, u32 size_max, u32 size_min,
-		u32 align, struct nouveau_mm_node **pnode)
+nouveau_mm_tail(struct nouveau_mm *mm, u8 heap, u8 type, u32 size_max,
+		u32 size_min, u32 align, struct nouveau_mm_node **pnode)
 {
 	struct nouveau_mm_node *prev, *this, *next;
 	u32 mask = align - 1;
 
-	BUG_ON(!type);
+	BUG_ON(type == NVKM_MM_TYPE_NONE || type == NVKM_MM_TYPE_HOLE);
 
 	list_for_each_entry_reverse(this, &mm->free, fl_entry) {
 		u32 e = this->offset + this->length;
 		u32 s = this->offset;
 		u32 c = 0, a;
+		if (unlikely(heap != NVKM_MM_HEAP_ANY)) {
+			if (this->heap != heap)
+				continue;
+		}
 
 		prev = node(this, prev);
 		if (prev && prev->type != type)
@@ -209,9 +237,23 @@ nouveau_mm_tail(struct nouveau_mm *mm, u8 type, u32 size_max, u32 size_min,
 int
 nouveau_mm_init(struct nouveau_mm *mm, u32 offset, u32 length, u32 block)
 {
-	struct nouveau_mm_node *node;
+	struct nouveau_mm_node *node, *prev;
+	u32 next;
 
-	if (block) {
+	if (nouveau_mm_initialised(mm)) {
+		prev = list_last_entry(&mm->nodes, typeof(*node), nl_entry);
+		next = prev->offset + prev->length;
+		if (next != offset) {
+			BUG_ON(next > offset);
+			if (!(node = kzalloc(sizeof(*node), GFP_KERNEL)))
+				return -ENOMEM;
+			node->type   = NVKM_MM_TYPE_HOLE;
+			node->offset = next;
+			node->length = offset - next;
+			list_add_tail(&node->nl_entry, &mm->nodes);
+		}
+		BUG_ON(block != mm->block_size);
+	} else {
 		INIT_LIST_HEAD(&mm->nodes);
 		INIT_LIST_HEAD(&mm->free);
 		mm->block_size = block;
@@ -230,25 +272,32 @@ nouveau_mm_init(struct nouveau_mm *mm, u32 offset, u32 length, u32 block)
 
 	list_add_tail(&node->nl_entry, &mm->nodes);
 	list_add_tail(&node->fl_entry, &mm->free);
-	mm->heap_nodes++;
+	node->heap = ++mm->heap_nodes;
 	return 0;
 }
 
 int
 nouveau_mm_fini(struct nouveau_mm *mm)
 {
-	if (nouveau_mm_initialised(mm)) {
-		struct nouveau_mm_node *node, *heap =
-			list_first_entry(&mm->nodes, typeof(*heap), nl_entry);
-		int nodes = 0;
+	struct nouveau_mm_node *node, *temp;
+	int nodes = 0;
 
-		list_for_each_entry(node, &mm->nodes, nl_entry) {
-			if (WARN_ON(nodes++ == mm->heap_nodes))
+	if (!nouveau_mm_initialised(mm))
+		return 0;
+
+	list_for_each_entry(node, &mm->nodes, nl_entry) {
+		if (node->type != NVKM_MM_TYPE_HOLE) {
+			if (++nodes > mm->heap_nodes) {
+				nouveau_mm_dump(mm, "mm not clean!");
 				return -EBUSY;
+			}
 		}
-
-		kfree(heap);
 	}
 
+	list_for_each_entry_safe(node, temp, &mm->nodes, nl_entry) {
+		list_del(&node->nl_entry);
+		kfree(node);
+	}
+	mm->heap_nodes = 0;
 	return 0;
 }
diff --git a/drivers/gpu/drm/nouveau/core/core/notify.c b/drivers/gpu/drm/nouveau/core/core/notify.c
index 76adb81bdea2..839a32577680 100644
--- a/drivers/gpu/drm/nouveau/core/core/notify.c
+++ b/drivers/gpu/drm/nouveau/core/core/notify.c
@@ -98,7 +98,7 @@ nvkm_notify_send(struct nvkm_notify *notify, void *data, u32 size)
 	struct nvkm_event *event = notify->event;
 	unsigned long flags;
 
-	BUG_ON(!spin_is_locked(&event->list_lock));
+	assert_spin_locked(&event->list_lock);
 	BUG_ON(size != notify->size);
 
 	spin_lock_irqsave(&event->refs_lock, flags);
@@ -134,14 +134,15 @@ nvkm_notify_fini(struct nvkm_notify *notify)
 }
 
 int
-nvkm_notify_init(struct nvkm_event *event, int (*func)(struct nvkm_notify *),
-		 bool work, void *data, u32 size, u32 reply,
+nvkm_notify_init(struct nouveau_object *object, struct nvkm_event *event,
+		 int (*func)(struct nvkm_notify *), bool work,
+		 void *data, u32 size, u32 reply,
 		 struct nvkm_notify *notify)
 {
 	unsigned long flags;
 	int ret = -ENODEV;
 	if ((notify->event = event), event->refs) {
-		ret = event->func->ctor(data, size, notify);
+		ret = event->func->ctor(object, data, size, notify);
 		if (ret == 0 && (ret = -EINVAL, notify->size == reply)) {
 			notify->flags = 0;
 			notify->block = 1;
diff --git a/drivers/gpu/drm/nouveau/core/engine/device/base.c b/drivers/gpu/drm/nouveau/core/engine/device/base.c
index 8928f7981d4a..137e0b0faeae 100644
--- a/drivers/gpu/drm/nouveau/core/engine/device/base.c
+++ b/drivers/gpu/drm/nouveau/core/engine/device/base.c
@@ -29,6 +29,7 @@
 #include <nvif/unpack.h>
 #include <nvif/class.h>
 
+#include <subdev/bios.h>
 #include <subdev/fb.h>
 #include <subdev/instmem.h>
 
@@ -138,7 +139,7 @@ nouveau_devobj_info(struct nouveau_object *object, void *data, u32 size)
 	}
 
 	args->v0.chipset  = device->chipset;
-	args->v0.revision = device->chipset >= 0x10 ? nv_rd32(device, 0) : 0x00;
+	args->v0.revision = device->chiprev;
 	if (pfb)  args->v0.ram_size = args->v0.ram_user = pfb->ram->size;
 	else      args->v0.ram_size = args->v0.ram_user = 0;
 	if (imem) args->v0.ram_user = args->v0.ram_user - imem->reserved;
@@ -222,6 +223,7 @@ static const u64 disable_map[] = {
 	[NVDEV_SUBDEV_VOLT]	= NV_DEVICE_V0_DISABLE_CORE,
 	[NVDEV_SUBDEV_THERM]	= NV_DEVICE_V0_DISABLE_CORE,
 	[NVDEV_SUBDEV_PWR]	= NV_DEVICE_V0_DISABLE_CORE,
+	[NVDEV_SUBDEV_FUSE]	= NV_DEVICE_V0_DISABLE_CORE,
 	[NVDEV_ENGINE_DMAOBJ]	= NV_DEVICE_V0_DISABLE_CORE,
 	[NVDEV_ENGINE_PERFMON]  = NV_DEVICE_V0_DISABLE_CORE,
 	[NVDEV_ENGINE_FIFO]	= NV_DEVICE_V0_DISABLE_FIFO,
@@ -235,6 +237,7 @@ static const u64 disable_map[] = {
 	[NVDEV_ENGINE_PPP]	= NV_DEVICE_V0_DISABLE_PPP,
 	[NVDEV_ENGINE_COPY0]	= NV_DEVICE_V0_DISABLE_COPY0,
 	[NVDEV_ENGINE_COPY1]	= NV_DEVICE_V0_DISABLE_COPY1,
+	[NVDEV_ENGINE_COPY2]	= NV_DEVICE_V0_DISABLE_COPY1,
 	[NVDEV_ENGINE_VIC]	= NV_DEVICE_V0_DISABLE_VIC,
 	[NVDEV_ENGINE_VENC]	= NV_DEVICE_V0_DISABLE_VENC,
 	[NVDEV_ENGINE_DISP]	= NV_DEVICE_V0_DISABLE_DISP,
@@ -352,12 +355,14 @@ nouveau_devobj_ctor(struct nouveau_object *parent,
 		/* determine chipset and derive architecture from it */
 		if ((boot0 & 0x1f000000) > 0) {
 			device->chipset = (boot0 & 0x1ff00000) >> 20;
+			device->chiprev = (boot0 & 0x000000ff);
 			switch (device->chipset & 0x1f0) {
 			case 0x010: {
 				if (0x461 & (1 << (device->chipset & 0xf)))
 					device->card_type = NV_10;
 				else
 					device->card_type = NV_11;
+				device->chiprev = 0x00;
 				break;
 			}
 			case 0x020: device->card_type = NV_20; break;
@@ -373,7 +378,8 @@ nouveau_devobj_ctor(struct nouveau_object *parent,
 			case 0x0e0:
 			case 0x0f0:
 			case 0x100: device->card_type = NV_E0; break;
-			case 0x110: device->card_type = GM100; break;
+			case 0x110:
+			case 0x120: device->card_type = GM100; break;
 			default:
 				break;
 			}
@@ -427,6 +433,10 @@ nouveau_devobj_ctor(struct nouveau_object *parent,
 		}
 
 		nv_debug(device, "crystal freq: %dKHz\n", device->crystal);
+	} else
+	if ( (args->v0.disable & NV_DEVICE_V0_DISABLE_IDENTIFY)) {
+		device->cname = "NULL";
+		device->oclass[NVDEV_SUBDEV_VBIOS] = &nouveau_bios_oclass;
 	}
 
 	if (!(args->v0.disable & NV_DEVICE_V0_DISABLE_MMIO) &&
@@ -505,7 +515,8 @@ nouveau_device_sclass[] = {
 };
 
 static int
-nouveau_device_event_ctor(void *data, u32 size, struct nvkm_notify *notify)
+nouveau_device_event_ctor(struct nouveau_object *object, void *data, u32 size,
+			  struct nvkm_notify *notify)
 {
 	if (!WARN_ON(size != 0)) {
 		notify->size  = 0;
diff --git a/drivers/gpu/drm/nouveau/core/engine/device/gm100.c b/drivers/gpu/drm/nouveau/core/engine/device/gm100.c
index 377ec0b8851e..4e74a3376de8 100644
--- a/drivers/gpu/drm/nouveau/core/engine/device/gm100.c
+++ b/drivers/gpu/drm/nouveau/core/engine/device/gm100.c
@@ -26,6 +26,7 @@
 #include <subdev/bus.h>
 #include <subdev/gpio.h>
 #include <subdev/i2c.h>
+#include <subdev/fuse.h>
 #include <subdev/clock.h>
 #include <subdev/therm.h>
 #include <subdev/mxm.h>
@@ -62,10 +63,9 @@ gm100_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
 		device->oclass[NVDEV_SUBDEV_GPIO   ] =  nve0_gpio_oclass;
 		device->oclass[NVDEV_SUBDEV_I2C    ] =  nvd0_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_FUSE   ] = &gm107_fuse_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nve0_clock_oclass;
-#if 0
-		device->oclass[NVDEV_SUBDEV_THERM  ] = &nvd0_therm_oclass;
-#endif
+		device->oclass[NVDEV_SUBDEV_THERM  ] = &gm107_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] =  gm107_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] =  gk20a_mc_oclass;
@@ -77,8 +77,9 @@ gm100_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_INSTMEM] =  nv50_instmem_oclass;
 		device->oclass[NVDEV_SUBDEV_VM     ] = &nvc0_vmmgr_oclass;
 		device->oclass[NVDEV_SUBDEV_BAR    ] = &nvc0_bar_oclass;
-#if 0
 		device->oclass[NVDEV_SUBDEV_PWR    ] =  nv108_pwr_oclass;
+
+#if 0
 		device->oclass[NVDEV_SUBDEV_VOLT   ] = &nv40_volt_oclass;
 #endif
 		device->oclass[NVDEV_ENGINE_DMAOBJ ] =  nvd0_dmaeng_oclass;
@@ -97,6 +98,49 @@ gm100_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_ENGINE_PPP    ] = &nvc0_ppp_oclass;
 #endif
 		break;
+	case 0x124:
+		device->cname = "GM204";
+		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
+		device->oclass[NVDEV_SUBDEV_GPIO   ] =  nve0_gpio_oclass;
+		device->oclass[NVDEV_SUBDEV_I2C    ] =  gm204_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_FUSE   ] = &gm107_fuse_oclass;
+#if 0
+		/* looks to be some non-trivial changes */
+		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nve0_clock_oclass;
+		/* priv ring says no to 0x10eb14 writes */
+		device->oclass[NVDEV_SUBDEV_THERM  ] = &gm107_therm_oclass;
+#endif
+		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
+		device->oclass[NVDEV_SUBDEV_DEVINIT] =  gm204_devinit_oclass;
+		device->oclass[NVDEV_SUBDEV_MC     ] =  gk20a_mc_oclass;
+		device->oclass[NVDEV_SUBDEV_BUS    ] =  nvc0_bus_oclass;
+		device->oclass[NVDEV_SUBDEV_TIMER  ] = &gk20a_timer_oclass;
+		device->oclass[NVDEV_SUBDEV_FB     ] =  gm107_fb_oclass;
+		device->oclass[NVDEV_SUBDEV_LTC    ] =  gm107_ltc_oclass;
+		device->oclass[NVDEV_SUBDEV_IBUS   ] = &nve0_ibus_oclass;
+		device->oclass[NVDEV_SUBDEV_INSTMEM] =  nv50_instmem_oclass;
+		device->oclass[NVDEV_SUBDEV_VM     ] = &nvc0_vmmgr_oclass;
+		device->oclass[NVDEV_SUBDEV_BAR    ] = &nvc0_bar_oclass;
+		device->oclass[NVDEV_SUBDEV_PWR    ] =  nv108_pwr_oclass;
+#if 0
+		device->oclass[NVDEV_SUBDEV_VOLT   ] = &nv40_volt_oclass;
+#endif
+		device->oclass[NVDEV_ENGINE_DMAOBJ ] =  nvd0_dmaeng_oclass;
+#if 0
+		device->oclass[NVDEV_ENGINE_FIFO   ] =  nv108_fifo_oclass;
+		device->oclass[NVDEV_ENGINE_SW     ] =  nvc0_software_oclass;
+		device->oclass[NVDEV_ENGINE_GR     ] =  gm107_graph_oclass;
+#endif
+		device->oclass[NVDEV_ENGINE_DISP   ] =  gm204_disp_oclass;
+#if 0
+		device->oclass[NVDEV_ENGINE_COPY0  ] = &gm204_copy0_oclass;
+		device->oclass[NVDEV_ENGINE_COPY1  ] = &gm204_copy1_oclass;
+		device->oclass[NVDEV_ENGINE_COPY2  ] = &gm204_copy2_oclass;
+		device->oclass[NVDEV_ENGINE_BSP    ] = &nve0_bsp_oclass;
+		device->oclass[NVDEV_ENGINE_VP     ] = &nve0_vp_oclass;
+		device->oclass[NVDEV_ENGINE_PPP    ] = &nvc0_ppp_oclass;
+#endif
+		break;
 	default:
 		nv_fatal(device, "unknown Maxwell chipset\n");
 		return -EINVAL;
diff --git a/drivers/gpu/drm/nouveau/core/engine/device/nv50.c b/drivers/gpu/drm/nouveau/core/engine/device/nv50.c
index 932f84fae459..96f568d1321b 100644
--- a/drivers/gpu/drm/nouveau/core/engine/device/nv50.c
+++ b/drivers/gpu/drm/nouveau/core/engine/device/nv50.c
@@ -26,6 +26,7 @@
 #include <subdev/bus.h>
 #include <subdev/gpio.h>
 #include <subdev/i2c.h>
+#include <subdev/fuse.h>
 #include <subdev/clock.h>
 #include <subdev/therm.h>
 #include <subdev/mxm.h>
@@ -62,6 +63,7 @@ nv50_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
 		device->oclass[NVDEV_SUBDEV_GPIO   ] =  nv50_gpio_oclass;
 		device->oclass[NVDEV_SUBDEV_I2C    ] =  nv50_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_FUSE   ] =  &g80_fuse_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] =  nv50_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nv50_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
@@ -87,6 +89,7 @@ nv50_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
 		device->oclass[NVDEV_SUBDEV_GPIO   ] =  nv50_gpio_oclass;
 		device->oclass[NVDEV_SUBDEV_I2C    ] =  nv50_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_FUSE   ] =  &g80_fuse_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] =  nv84_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nv84_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
@@ -115,6 +118,7 @@ nv50_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
 		device->oclass[NVDEV_SUBDEV_GPIO   ] =  nv50_gpio_oclass;
 		device->oclass[NVDEV_SUBDEV_I2C    ] =  nv50_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_FUSE   ] =  &g80_fuse_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] =  nv84_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nv84_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
@@ -141,8 +145,9 @@ nv50_identify(struct nouveau_device *device)
 	case 0x92:
 		device->cname = "G92";
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
-		device->oclass[NVDEV_SUBDEV_GPIO   ] =  nv92_gpio_oclass;
+		device->oclass[NVDEV_SUBDEV_GPIO   ] =  nv50_gpio_oclass;
 		device->oclass[NVDEV_SUBDEV_I2C    ] =  nv50_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_FUSE   ] =  &g80_fuse_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] =  nv84_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nv84_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
@@ -169,8 +174,9 @@ nv50_identify(struct nouveau_device *device)
 	case 0x94:
 		device->cname = "G94";
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
-		device->oclass[NVDEV_SUBDEV_GPIO   ] =  nv92_gpio_oclass;
+		device->oclass[NVDEV_SUBDEV_GPIO   ] =  nv94_gpio_oclass;
 		device->oclass[NVDEV_SUBDEV_I2C    ] =  nv94_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_FUSE   ] =  &g80_fuse_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] =  nv84_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nv84_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
@@ -197,8 +203,9 @@ nv50_identify(struct nouveau_device *device)
 	case 0x96:
 		device->cname = "G96";
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
-		device->oclass[NVDEV_SUBDEV_GPIO   ] =  nv92_gpio_oclass;
+		device->oclass[NVDEV_SUBDEV_GPIO   ] =  nv94_gpio_oclass;
 		device->oclass[NVDEV_SUBDEV_I2C    ] =  nv94_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_FUSE   ] =  &g80_fuse_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] =  nv84_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nv84_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
@@ -225,8 +232,9 @@ nv50_identify(struct nouveau_device *device)
 	case 0x98:
 		device->cname = "G98";
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
-		device->oclass[NVDEV_SUBDEV_GPIO   ] =  nv92_gpio_oclass;
+		device->oclass[NVDEV_SUBDEV_GPIO   ] =  nv94_gpio_oclass;
 		device->oclass[NVDEV_SUBDEV_I2C    ] =  nv94_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_FUSE   ] =  &g80_fuse_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] =  nv84_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nv84_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
@@ -253,8 +261,9 @@ nv50_identify(struct nouveau_device *device)
 	case 0xa0:
 		device->cname = "G200";
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
-		device->oclass[NVDEV_SUBDEV_GPIO   ] =  nv92_gpio_oclass;
+		device->oclass[NVDEV_SUBDEV_GPIO   ] =  nv94_gpio_oclass;
 		device->oclass[NVDEV_SUBDEV_I2C    ] =  nv50_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_FUSE   ] =  &g80_fuse_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] =  nv84_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nv84_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
@@ -281,8 +290,9 @@ nv50_identify(struct nouveau_device *device)
 	case 0xaa:
 		device->cname = "MCP77/MCP78";
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
-		device->oclass[NVDEV_SUBDEV_GPIO   ] =  nv92_gpio_oclass;
+		device->oclass[NVDEV_SUBDEV_GPIO   ] =  nv94_gpio_oclass;
 		device->oclass[NVDEV_SUBDEV_I2C    ] =  nv94_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_FUSE   ] =  &g80_fuse_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] =  nvaa_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nv84_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
@@ -309,8 +319,9 @@ nv50_identify(struct nouveau_device *device)
 	case 0xac:
 		device->cname = "MCP79/MCP7A";
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
-		device->oclass[NVDEV_SUBDEV_GPIO   ] =  nv92_gpio_oclass;
+		device->oclass[NVDEV_SUBDEV_GPIO   ] =  nv94_gpio_oclass;
 		device->oclass[NVDEV_SUBDEV_I2C    ] =  nv94_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_FUSE   ] =  &g80_fuse_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] =  nvaa_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nv84_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
@@ -337,8 +348,9 @@ nv50_identify(struct nouveau_device *device)
 	case 0xa3:
 		device->cname = "GT215";
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
-		device->oclass[NVDEV_SUBDEV_GPIO   ] =  nv92_gpio_oclass;
+		device->oclass[NVDEV_SUBDEV_GPIO   ] =  nv94_gpio_oclass;
 		device->oclass[NVDEV_SUBDEV_I2C    ] =  nv94_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_FUSE   ] =  &g80_fuse_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nva3_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nva3_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
@@ -367,8 +379,9 @@ nv50_identify(struct nouveau_device *device)
 	case 0xa5:
 		device->cname = "GT216";
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
-		device->oclass[NVDEV_SUBDEV_GPIO   ] =  nv92_gpio_oclass;
+		device->oclass[NVDEV_SUBDEV_GPIO   ] =  nv94_gpio_oclass;
 		device->oclass[NVDEV_SUBDEV_I2C    ] =  nv94_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_FUSE   ] =  &g80_fuse_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nva3_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nva3_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
@@ -396,8 +409,9 @@ nv50_identify(struct nouveau_device *device)
 	case 0xa8:
 		device->cname = "GT218";
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
-		device->oclass[NVDEV_SUBDEV_GPIO   ] =  nv92_gpio_oclass;
+		device->oclass[NVDEV_SUBDEV_GPIO   ] =  nv94_gpio_oclass;
 		device->oclass[NVDEV_SUBDEV_I2C    ] =  nv94_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_FUSE   ] =  &g80_fuse_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nva3_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nva3_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
@@ -425,8 +439,9 @@ nv50_identify(struct nouveau_device *device)
 	case 0xaf:
 		device->cname = "MCP89";
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
-		device->oclass[NVDEV_SUBDEV_GPIO   ] =  nv92_gpio_oclass;
+		device->oclass[NVDEV_SUBDEV_GPIO   ] =  nv94_gpio_oclass;
 		device->oclass[NVDEV_SUBDEV_I2C    ] =  nv94_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_FUSE   ] =  &g80_fuse_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nva3_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nva3_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
diff --git a/drivers/gpu/drm/nouveau/core/engine/device/nvc0.c b/drivers/gpu/drm/nouveau/core/engine/device/nvc0.c
index b4a2917ce555..72a40f95d048 100644
--- a/drivers/gpu/drm/nouveau/core/engine/device/nvc0.c
+++ b/drivers/gpu/drm/nouveau/core/engine/device/nvc0.c
@@ -26,6 +26,7 @@
 #include <subdev/bus.h>
 #include <subdev/gpio.h>
 #include <subdev/i2c.h>
+#include <subdev/fuse.h>
 #include <subdev/clock.h>
 #include <subdev/therm.h>
 #include <subdev/mxm.h>
@@ -60,8 +61,9 @@ nvc0_identify(struct nouveau_device *device)
 	case 0xc0:
 		device->cname = "GF100";
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
-		device->oclass[NVDEV_SUBDEV_GPIO   ] =  nv92_gpio_oclass;
+		device->oclass[NVDEV_SUBDEV_GPIO   ] =  nv94_gpio_oclass;
 		device->oclass[NVDEV_SUBDEV_I2C    ] =  nv94_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_FUSE   ] = &gf100_fuse_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nvc0_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nva3_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
@@ -92,8 +94,9 @@ nvc0_identify(struct nouveau_device *device)
 	case 0xc4:
 		device->cname = "GF104";
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
-		device->oclass[NVDEV_SUBDEV_GPIO   ] =  nv92_gpio_oclass;
+		device->oclass[NVDEV_SUBDEV_GPIO   ] =  nv94_gpio_oclass;
 		device->oclass[NVDEV_SUBDEV_I2C    ] =  nv94_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_FUSE   ] = &gf100_fuse_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nvc0_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nva3_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
@@ -124,8 +127,9 @@ nvc0_identify(struct nouveau_device *device)
 	case 0xc3:
 		device->cname = "GF106";
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
-		device->oclass[NVDEV_SUBDEV_GPIO   ] =  nv92_gpio_oclass;
+		device->oclass[NVDEV_SUBDEV_GPIO   ] =  nv94_gpio_oclass;
 		device->oclass[NVDEV_SUBDEV_I2C    ] =  nv94_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_FUSE   ] = &gf100_fuse_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nvc0_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nva3_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
@@ -155,8 +159,9 @@ nvc0_identify(struct nouveau_device *device)
 	case 0xce:
 		device->cname = "GF114";
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
-		device->oclass[NVDEV_SUBDEV_GPIO   ] =  nv92_gpio_oclass;
+		device->oclass[NVDEV_SUBDEV_GPIO   ] =  nv94_gpio_oclass;
 		device->oclass[NVDEV_SUBDEV_I2C    ] =  nv94_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_FUSE   ] = &gf100_fuse_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nvc0_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nva3_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
@@ -187,8 +192,9 @@ nvc0_identify(struct nouveau_device *device)
 	case 0xcf:
 		device->cname = "GF116";
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
-		device->oclass[NVDEV_SUBDEV_GPIO   ] =  nv92_gpio_oclass;
+		device->oclass[NVDEV_SUBDEV_GPIO   ] =  nv94_gpio_oclass;
 		device->oclass[NVDEV_SUBDEV_I2C    ] =  nv94_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_FUSE   ] = &gf100_fuse_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nvc0_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nva3_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
@@ -212,15 +218,15 @@ nvc0_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_ENGINE_BSP    ] = &nvc0_bsp_oclass;
 		device->oclass[NVDEV_ENGINE_PPP    ] = &nvc0_ppp_oclass;
 		device->oclass[NVDEV_ENGINE_COPY0  ] = &nvc0_copy0_oclass;
-		device->oclass[NVDEV_ENGINE_COPY1  ] = &nvc0_copy1_oclass;
 		device->oclass[NVDEV_ENGINE_DISP   ] =  nva3_disp_oclass;
 		device->oclass[NVDEV_ENGINE_PERFMON] = &nvc0_perfmon_oclass;
 		break;
 	case 0xc1:
 		device->cname = "GF108";
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
-		device->oclass[NVDEV_SUBDEV_GPIO   ] =  nv92_gpio_oclass;
+		device->oclass[NVDEV_SUBDEV_GPIO   ] =  nv94_gpio_oclass;
 		device->oclass[NVDEV_SUBDEV_I2C    ] =  nv94_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_FUSE   ] = &gf100_fuse_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nvc0_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nva3_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
@@ -250,8 +256,9 @@ nvc0_identify(struct nouveau_device *device)
 	case 0xc8:
 		device->cname = "GF110";
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
-		device->oclass[NVDEV_SUBDEV_GPIO   ] =  nv92_gpio_oclass;
+		device->oclass[NVDEV_SUBDEV_GPIO   ] =  nv94_gpio_oclass;
 		device->oclass[NVDEV_SUBDEV_I2C    ] =  nv94_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_FUSE   ] = &gf100_fuse_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nvc0_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nva3_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
@@ -284,6 +291,7 @@ nvc0_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
 		device->oclass[NVDEV_SUBDEV_GPIO   ] =  nvd0_gpio_oclass;
 		device->oclass[NVDEV_SUBDEV_I2C    ] =  nvd0_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_FUSE   ] = &gf100_fuse_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nvc0_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nvd0_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
@@ -315,6 +323,7 @@ nvc0_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
 		device->oclass[NVDEV_SUBDEV_GPIO   ] =  nvd0_gpio_oclass;
 		device->oclass[NVDEV_SUBDEV_I2C    ] =  gf117_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_FUSE   ] = &gf100_fuse_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nvc0_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nvd0_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
diff --git a/drivers/gpu/drm/nouveau/core/engine/device/nve0.c b/drivers/gpu/drm/nouveau/core/engine/device/nve0.c
index cdf9147f32a1..732922690653 100644
--- a/drivers/gpu/drm/nouveau/core/engine/device/nve0.c
+++ b/drivers/gpu/drm/nouveau/core/engine/device/nve0.c
@@ -26,6 +26,7 @@
 #include <subdev/bus.h>
 #include <subdev/gpio.h>
 #include <subdev/i2c.h>
+#include <subdev/fuse.h>
 #include <subdev/clock.h>
 #include <subdev/therm.h>
 #include <subdev/mxm.h>
@@ -62,6 +63,7 @@ nve0_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
 		device->oclass[NVDEV_SUBDEV_GPIO   ] =  nve0_gpio_oclass;
 		device->oclass[NVDEV_SUBDEV_I2C    ] =  nve0_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_FUSE   ] = &gf100_fuse_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nve0_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nvd0_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
@@ -95,6 +97,7 @@ nve0_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
 		device->oclass[NVDEV_SUBDEV_GPIO   ] =  nve0_gpio_oclass;
 		device->oclass[NVDEV_SUBDEV_I2C    ] =  nve0_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_FUSE   ] = &gf100_fuse_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nve0_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nvd0_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
@@ -128,6 +131,7 @@ nve0_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
 		device->oclass[NVDEV_SUBDEV_GPIO   ] =  nve0_gpio_oclass;
 		device->oclass[NVDEV_SUBDEV_I2C    ] =  nve0_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_FUSE   ] = &gf100_fuse_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nve0_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nvd0_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
@@ -161,6 +165,7 @@ nve0_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &gk20a_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] =  gk20a_mc_oclass;
 		device->oclass[NVDEV_SUBDEV_BUS    ] =  nvc0_bus_oclass;
+		device->oclass[NVDEV_SUBDEV_FUSE   ] = &gf100_fuse_oclass;
 		device->oclass[NVDEV_SUBDEV_TIMER  ] = &gk20a_timer_oclass;
 		device->oclass[NVDEV_SUBDEV_FB     ] =  gk20a_fb_oclass;
 		device->oclass[NVDEV_SUBDEV_LTC    ] =  gk104_ltc_oclass;
@@ -174,12 +179,14 @@ nve0_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_ENGINE_GR     ] =  gk20a_graph_oclass;
 		device->oclass[NVDEV_ENGINE_COPY2  ] = &nve0_copy2_oclass;
 		device->oclass[NVDEV_ENGINE_PERFMON] = &nve0_perfmon_oclass;
+		device->oclass[NVDEV_SUBDEV_VOLT   ] = &gk20a_volt_oclass;
 		break;
 	case 0xf0:
 		device->cname = "GK110";
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
 		device->oclass[NVDEV_SUBDEV_GPIO   ] =  nve0_gpio_oclass;
 		device->oclass[NVDEV_SUBDEV_I2C    ] =  nve0_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_FUSE   ] = &gf100_fuse_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nve0_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nvd0_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
@@ -213,6 +220,7 @@ nve0_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
 		device->oclass[NVDEV_SUBDEV_GPIO   ] =  nve0_gpio_oclass;
 		device->oclass[NVDEV_SUBDEV_I2C    ] =  nvd0_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_FUSE   ] = &gf100_fuse_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nve0_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nvd0_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
@@ -241,11 +249,45 @@ nve0_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_ENGINE_PPP    ] = &nvc0_ppp_oclass;
 		device->oclass[NVDEV_ENGINE_PERFMON] = &nvf0_perfmon_oclass;
 		break;
+	case 0x106:
+		device->cname = "GK208B";
+		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
+		device->oclass[NVDEV_SUBDEV_GPIO   ] =  nve0_gpio_oclass;
+		device->oclass[NVDEV_SUBDEV_I2C    ] =  nve0_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_FUSE   ] = &gf100_fuse_oclass;
+		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nve0_clock_oclass;
+		device->oclass[NVDEV_SUBDEV_THERM  ] = &nvd0_therm_oclass;
+		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
+		device->oclass[NVDEV_SUBDEV_DEVINIT] =  nvc0_devinit_oclass;
+		device->oclass[NVDEV_SUBDEV_MC     ] =  gk20a_mc_oclass;
+		device->oclass[NVDEV_SUBDEV_BUS    ] =  nvc0_bus_oclass;
+		device->oclass[NVDEV_SUBDEV_TIMER  ] = &nv04_timer_oclass;
+		device->oclass[NVDEV_SUBDEV_FB     ] =  nve0_fb_oclass;
+		device->oclass[NVDEV_SUBDEV_LTC    ] =  gk104_ltc_oclass;
+		device->oclass[NVDEV_SUBDEV_IBUS   ] = &nve0_ibus_oclass;
+		device->oclass[NVDEV_SUBDEV_INSTMEM] =  nv50_instmem_oclass;
+		device->oclass[NVDEV_SUBDEV_VM     ] = &nvc0_vmmgr_oclass;
+		device->oclass[NVDEV_SUBDEV_BAR    ] = &nvc0_bar_oclass;
+		device->oclass[NVDEV_SUBDEV_PWR    ] =  nv108_pwr_oclass;
+		device->oclass[NVDEV_SUBDEV_VOLT   ] = &nv40_volt_oclass;
+		device->oclass[NVDEV_ENGINE_DMAOBJ ] =  nvd0_dmaeng_oclass;
+		device->oclass[NVDEV_ENGINE_FIFO   ] =  nv108_fifo_oclass;
+		device->oclass[NVDEV_ENGINE_SW     ] =  nvc0_software_oclass;
+		device->oclass[NVDEV_ENGINE_GR     ] =  nv108_graph_oclass;
+		device->oclass[NVDEV_ENGINE_DISP   ] =  nvf0_disp_oclass;
+		device->oclass[NVDEV_ENGINE_COPY0  ] = &nve0_copy0_oclass;
+		device->oclass[NVDEV_ENGINE_COPY1  ] = &nve0_copy1_oclass;
+		device->oclass[NVDEV_ENGINE_COPY2  ] = &nve0_copy2_oclass;
+		device->oclass[NVDEV_ENGINE_BSP    ] = &nve0_bsp_oclass;
+		device->oclass[NVDEV_ENGINE_VP     ] = &nve0_vp_oclass;
+		device->oclass[NVDEV_ENGINE_PPP    ] = &nvc0_ppp_oclass;
+		break;
 	case 0x108:
 		device->cname = "GK208";
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
 		device->oclass[NVDEV_SUBDEV_GPIO   ] =  nve0_gpio_oclass;
 		device->oclass[NVDEV_SUBDEV_I2C    ] =  nve0_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_FUSE   ] = &gf100_fuse_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nve0_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nvd0_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/base.c b/drivers/gpu/drm/nouveau/core/engine/disp/base.c
index 22d55f6cde50..64b84667f3a5 100644
--- a/drivers/gpu/drm/nouveau/core/engine/disp/base.c
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/base.c
@@ -32,7 +32,8 @@
 #include "conn.h"
 
 int
-nouveau_disp_vblank_ctor(void *data, u32 size, struct nvkm_notify *notify)
+nouveau_disp_vblank_ctor(struct nouveau_object *object, void *data, u32 size,
+			 struct nvkm_notify *notify)
 {
 	struct nouveau_disp *disp =
 		container_of(notify->event, typeof(*disp), vblank);
@@ -61,7 +62,8 @@ nouveau_disp_vblank(struct nouveau_disp *disp, int head)
 }
 
 static int
-nouveau_disp_hpd_ctor(void *data, u32 size, struct nvkm_notify *notify)
+nouveau_disp_hpd_ctor(struct nouveau_object *object, void *data, u32 size,
+		      struct nvkm_notify *notify)
 {
 	struct nouveau_disp *disp =
 		container_of(notify->event, typeof(*disp), hpd);
diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/conn.c b/drivers/gpu/drm/nouveau/core/engine/disp/conn.c
index 3d1070228977..1496b567dd4a 100644
--- a/drivers/gpu/drm/nouveau/core/engine/disp/conn.c
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/conn.c
@@ -126,8 +126,8 @@ nvkm_connector_create_(struct nouveau_object *parent,
 			return 0;
 		}
 
-		ret = nvkm_notify_init(&gpio->event, nvkm_connector_hpd, true,
-				       &(struct nvkm_gpio_ntfy_req) {
+		ret = nvkm_notify_init(NULL, &gpio->event, nvkm_connector_hpd,
+				       true, &(struct nvkm_gpio_ntfy_req) {
 					.mask = NVKM_GPIO_TOGGLED,
 					.line = func.line,
 				       },
diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/dport.c b/drivers/gpu/drm/nouveau/core/engine/disp/dport.c
index 39890221b91c..16db08dfba6e 100644
--- a/drivers/gpu/drm/nouveau/core/engine/disp/dport.c
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/dport.c
@@ -28,7 +28,7 @@
 #include <subdev/bios/init.h>
 #include <subdev/i2c.h>
 
-#include <engine/disp.h>
+#include "nv50.h"
 
 #include <nvif/class.h>
 
@@ -326,7 +326,7 @@ void
 nouveau_dp_train(struct work_struct *w)
 {
 	struct nvkm_output_dp *outp = container_of(w, typeof(*outp), lt.work);
-	struct nouveau_disp *disp = nouveau_disp(outp);
+	struct nv50_disp_priv *priv = (void *)nouveau_disp(outp);
 	const struct dp_rates *cfg = nouveau_dp_rates;
 	struct dp_state _dp = {
 		.outp = outp,
@@ -334,8 +334,11 @@ nouveau_dp_train(struct work_struct *w)
 	u32 datarate = 0;
 	int ret;
 
+	if (!outp->base.info.location && priv->sor.magic)
+		priv->sor.magic(&outp->base);
+
 	/* bring capabilities within encoder limits */
-	if (nv_mclass(disp) < GF110_DISP)
+	if (nv_mclass(priv) < GF110_DISP)
 		outp->dpcd[2] &= ~DPCD_RC02_TPS3_SUPPORTED;
 	if ((outp->dpcd[2] & 0x1f) > outp->base.info.dpconf.link_nr) {
 		outp->dpcd[2] &= ~DPCD_RC02_MAX_LANE_COUNT;
diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/gm107.c b/drivers/gpu/drm/nouveau/core/engine/disp/gm107.c
index d54da8b5f87e..e2ad0543fb31 100644
--- a/drivers/gpu/drm/nouveau/core/engine/disp/gm107.c
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/gm107.c
@@ -35,8 +35,8 @@
 
 static struct nouveau_oclass
 gm107_disp_sclass[] = {
-	{ GM107_DISP_CORE_CHANNEL_DMA, &nvd0_disp_mast_ofuncs.base },
-	{ GK110_DISP_BASE_CHANNEL_DMA, &nvd0_disp_sync_ofuncs.base },
+	{ GM107_DISP_CORE_CHANNEL_DMA, &nvd0_disp_core_ofuncs.base },
+	{ GK110_DISP_BASE_CHANNEL_DMA, &nvd0_disp_base_ofuncs.base },
 	{ GK104_DISP_OVERLAY_CONTROL_DMA, &nvd0_disp_ovly_ofuncs.base },
 	{ GK104_DISP_OVERLAY, &nvd0_disp_oimm_ofuncs.base },
 	{ GK104_DISP_CURSOR, &nvd0_disp_curs_ofuncs.base },
@@ -44,8 +44,8 @@ gm107_disp_sclass[] = {
 };
 
 static struct nouveau_oclass
-gm107_disp_base_oclass[] = {
-	{ GM107_DISP, &nvd0_disp_base_ofuncs },
+gm107_disp_main_oclass[] = {
+	{ GM107_DISP, &nvd0_disp_main_ofuncs },
 	{}
 };
 
@@ -68,7 +68,11 @@ gm107_disp_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	if (ret)
 		return ret;
 
-	nv_engine(priv)->sclass = gm107_disp_base_oclass;
+	ret = nvkm_event_init(&nvd0_disp_chan_uevent, 1, 17, &priv->uevent);
+	if (ret)
+		return ret;
+
+	nv_engine(priv)->sclass = gm107_disp_main_oclass;
 	nv_engine(priv)->cclass = &nv50_disp_cclass;
 	nv_subdev(priv)->intr = nvd0_disp_intr;
 	INIT_WORK(&priv->supervisor, nvd0_disp_intr_supervisor);
@@ -80,7 +84,7 @@ gm107_disp_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	priv->dac.sense = nv50_dac_sense;
 	priv->sor.power = nv50_sor_power;
 	priv->sor.hda_eld = nvd0_hda_eld;
-	priv->sor.hdmi = nvd0_hdmi_ctrl;
+	priv->sor.hdmi = nve0_hdmi_ctrl;
 	return 0;
 }
 
@@ -95,9 +99,9 @@ gm107_disp_oclass = &(struct nv50_disp_impl) {
 	},
 	.base.vblank = &nvd0_disp_vblank_func,
 	.base.outp =  nvd0_disp_outp_sclass,
-	.mthd.core = &nve0_disp_mast_mthd_chan,
-	.mthd.base = &nvd0_disp_sync_mthd_chan,
+	.mthd.core = &nve0_disp_core_mthd_chan,
+	.mthd.base = &nvd0_disp_base_mthd_chan,
 	.mthd.ovly = &nve0_disp_ovly_mthd_chan,
 	.mthd.prev = -0x020000,
-	.head.scanoutpos = nvd0_disp_base_scanoutpos,
+	.head.scanoutpos = nvd0_disp_main_scanoutpos,
 }.base.base;
diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/gm204.c b/drivers/gpu/drm/nouveau/core/engine/disp/gm204.c
new file mode 100644
index 000000000000..672ded79b2a9
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/gm204.c
@@ -0,0 +1,114 @@
+/*
+ * Copyright 2012 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs
+ */
+
+#include <engine/software.h>
+#include <engine/disp.h>
+
+#include <nvif/class.h>
+
+#include "nv50.h"
+
+/*******************************************************************************
+ * Base display object
+ ******************************************************************************/
+
+static struct nouveau_oclass
+gm204_disp_sclass[] = {
+	{ GM204_DISP_CORE_CHANNEL_DMA, &nvd0_disp_core_ofuncs.base },
+	{ GK110_DISP_BASE_CHANNEL_DMA, &nvd0_disp_base_ofuncs.base },
+	{ GK104_DISP_OVERLAY_CONTROL_DMA, &nvd0_disp_ovly_ofuncs.base },
+	{ GK104_DISP_OVERLAY, &nvd0_disp_oimm_ofuncs.base },
+	{ GK104_DISP_CURSOR, &nvd0_disp_curs_ofuncs.base },
+	{}
+};
+
+static struct nouveau_oclass
+gm204_disp_main_oclass[] = {
+	{ GM204_DISP, &nvd0_disp_main_ofuncs },
+	{}
+};
+
+/*******************************************************************************
+ * Display engine implementation
+ ******************************************************************************/
+
+static int
+gm204_disp_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
+	       struct nouveau_oclass *oclass, void *data, u32 size,
+	       struct nouveau_object **pobject)
+{
+	struct nv50_disp_priv *priv;
+	int heads = nv_rd32(parent, 0x022448);
+	int ret;
+
+	ret = nouveau_disp_create(parent, engine, oclass, heads,
+				  "PDISP", "display", &priv);
+	*pobject = nv_object(priv);
+	if (ret)
+		return ret;
+
+	ret = nvkm_event_init(&nvd0_disp_chan_uevent, 1, 17, &priv->uevent);
+	if (ret)
+		return ret;
+
+	nv_engine(priv)->sclass = gm204_disp_main_oclass;
+	nv_engine(priv)->cclass = &nv50_disp_cclass;
+	nv_subdev(priv)->intr = nvd0_disp_intr;
+	INIT_WORK(&priv->supervisor, nvd0_disp_intr_supervisor);
+	priv->sclass = gm204_disp_sclass;
+	priv->head.nr = heads;
+	priv->dac.nr = 3;
+	priv->sor.nr = 4;
+	priv->dac.power = nv50_dac_power;
+	priv->dac.sense = nv50_dac_sense;
+	priv->sor.power = nv50_sor_power;
+	priv->sor.hda_eld = nvd0_hda_eld;
+	priv->sor.hdmi = nvd0_hdmi_ctrl;
+	priv->sor.magic = gm204_sor_magic;
+	return 0;
+}
+
+struct nouveau_oclass *
+gm204_disp_outp_sclass[] = {
+	&gm204_sor_dp_impl.base.base,
+	NULL
+};
+
+struct nouveau_oclass *
+gm204_disp_oclass = &(struct nv50_disp_impl) {
+	.base.base.handle = NV_ENGINE(DISP, 0x07),
+	.base.base.ofuncs = &(struct nouveau_ofuncs) {
+		.ctor = gm204_disp_ctor,
+		.dtor = _nouveau_disp_dtor,
+		.init = _nouveau_disp_init,
+		.fini = _nouveau_disp_fini,
+	},
+	.base.vblank = &nvd0_disp_vblank_func,
+	.base.outp =  gm204_disp_outp_sclass,
+	.mthd.core = &nve0_disp_core_mthd_chan,
+	.mthd.base = &nvd0_disp_base_mthd_chan,
+	.mthd.ovly = &nve0_disp_ovly_mthd_chan,
+	.mthd.prev = -0x020000,
+	.head.scanoutpos = nvd0_disp_main_scanoutpos,
+}.base.base;
diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/hdanva3.c b/drivers/gpu/drm/nouveau/core/engine/disp/hdanva3.c
index 8b4e06abe533..fe9ef5894dd4 100644
--- a/drivers/gpu/drm/nouveau/core/engine/disp/hdanva3.c
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/hdanva3.c
@@ -26,6 +26,8 @@
 #include <nvif/unpack.h>
 #include <nvif/class.h>
 
+#include <subdev/timer.h>
+
 #include "nv50.h"
 
 int
@@ -46,16 +48,21 @@ nva3_hda_eld(NV50_DISP_MTHD_V1)
 		return ret;
 
 	if (size && args->v0.data[0]) {
+		if (outp->info.type == DCB_OUTPUT_DP) {
+			nv_mask(priv, 0x61c1e0 + soff, 0x8000000d, 0x80000001);
+			nv_wait(priv, 0x61c1e0 + soff, 0x80000000, 0x00000000);
+		}
 		for (i = 0; i < size; i++)
 			nv_wr32(priv, 0x61c440 + soff, (i << 8) | args->v0.data[0]);
 		for (; i < 0x60; i++)
 			nv_wr32(priv, 0x61c440 + soff, (i << 8));
 		nv_mask(priv, 0x61c448 + soff, 0x80000003, 0x80000003);
-	} else
-	if (size) {
-		nv_mask(priv, 0x61c448 + soff, 0x80000003, 0x80000001);
 	} else {
-		nv_mask(priv, 0x61c448 + soff, 0x80000003, 0x80000000);
+		if (outp->info.type == DCB_OUTPUT_DP) {
+			nv_mask(priv, 0x61c1e0 + soff, 0x80000001, 0x80000000);
+			nv_wait(priv, 0x61c1e0 + soff, 0x80000000, 0x00000000);
+		}
+		nv_mask(priv, 0x61c448 + soff, 0x80000003, 0x80000000 | !!size);
 	}
 
 	return 0;
diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/hdanvd0.c b/drivers/gpu/drm/nouveau/core/engine/disp/hdanvd0.c
index baf558fc12fb..1d4e8432d857 100644
--- a/drivers/gpu/drm/nouveau/core/engine/disp/hdanvd0.c
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/hdanvd0.c
@@ -26,10 +26,7 @@
 #include <nvif/unpack.h>
 #include <nvif/class.h>
 
-#include <subdev/bios.h>
-#include <subdev/bios/dcb.h>
-#include <subdev/bios/dp.h>
-#include <subdev/bios/init.h>
+#include <subdev/timer.h>
 
 #include "nv50.h"
 
@@ -40,6 +37,7 @@ nvd0_hda_eld(NV50_DISP_MTHD_V1)
 		struct nv50_disp_sor_hda_eld_v0 v0;
 	} *args = data;
 	const u32 soff = outp->or * 0x030;
+	const u32 hoff = head * 0x800;
 	int ret, i;
 
 	nv_ioctl(object, "disp sor hda eld size %d\n", size);
@@ -51,16 +49,22 @@ nvd0_hda_eld(NV50_DISP_MTHD_V1)
 		return ret;
 
 	if (size && args->v0.data[0]) {
+		if (outp->info.type == DCB_OUTPUT_DP) {
+			nv_mask(priv, 0x616618 + hoff, 0x8000000c, 0x80000001);
+			nv_wait(priv, 0x616618 + hoff, 0x80000000, 0x00000000);
+		}
+		nv_mask(priv, 0x616548 + hoff, 0x00000070, 0x00000000);
 		for (i = 0; i < size; i++)
 			nv_wr32(priv, 0x10ec00 + soff, (i << 8) | args->v0.data[i]);
 		for (; i < 0x60; i++)
 			nv_wr32(priv, 0x10ec00 + soff, (i << 8));
 		nv_mask(priv, 0x10ec10 + soff, 0x80000003, 0x80000003);
-	} else
-	if (size) {
-		nv_mask(priv, 0x10ec10 + soff, 0x80000003, 0x80000001);
 	} else {
-		nv_mask(priv, 0x10ec10 + soff, 0x80000003, 0x80000000);
+		if (outp->info.type == DCB_OUTPUT_DP) {
+			nv_mask(priv, 0x616618 + hoff, 0x80000001, 0x80000000);
+			nv_wait(priv, 0x616618 + hoff, 0x80000000, 0x00000000);
+		}
+		nv_mask(priv, 0x10ec10 + soff, 0x80000003, 0x80000000 | !!size);
 	}
 
 	return 0;
diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/hdminvd0.c b/drivers/gpu/drm/nouveau/core/engine/disp/hdminvd0.c
index 3106d295b48d..bac4fc4570f0 100644
--- a/drivers/gpu/drm/nouveau/core/engine/disp/hdminvd0.c
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/hdminvd0.c
@@ -75,8 +75,5 @@ nvd0_hdmi_ctrl(NV50_DISP_MTHD_V1)
 
 	/* HDMI_CTRL */
 	nv_mask(priv, 0x616798 + hoff, 0x401f007f, ctrl);
-
-	/* NFI, audio doesn't work without it though.. */
-	nv_mask(priv, 0x616548 + hoff, 0x00000070, 0x00000000);
 	return 0;
 }
diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/hdminve0.c b/drivers/gpu/drm/nouveau/core/engine/disp/hdminve0.c
new file mode 100644
index 000000000000..528d14ec2f7f
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/hdminve0.c
@@ -0,0 +1,83 @@
+/*
+ * Copyright 2014 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs
+ */
+
+#include <core/client.h>
+#include <nvif/unpack.h>
+#include <nvif/class.h>
+
+#include "nv50.h"
+
+int
+nve0_hdmi_ctrl(NV50_DISP_MTHD_V1)
+{
+	const u32 hoff = (head * 0x800);
+	const u32 hdmi = (head * 0x400);
+	union {
+		struct nv50_disp_sor_hdmi_pwr_v0 v0;
+	} *args = data;
+	u32 ctrl;
+	int ret;
+
+	nv_ioctl(object, "disp sor hdmi ctrl size %d\n", size);
+	if (nvif_unpack(args->v0, 0, 0, false)) {
+		nv_ioctl(object, "disp sor hdmi ctrl vers %d state %d "
+				 "max_ac_packet %d rekey %d\n",
+			 args->v0.version, args->v0.state,
+			 args->v0.max_ac_packet, args->v0.rekey);
+		if (args->v0.max_ac_packet > 0x1f || args->v0.rekey > 0x7f)
+			return -EINVAL;
+		ctrl  = 0x40000000 * !!args->v0.state;
+		ctrl |= args->v0.max_ac_packet << 16;
+		ctrl |= args->v0.rekey;
+	} else
+		return ret;
+
+	if (!(ctrl & 0x40000000)) {
+		nv_mask(priv, 0x616798 + hoff, 0x40000000, 0x00000000);
+		nv_mask(priv, 0x6900c0 + hdmi, 0x00000001, 0x00000000);
+		nv_mask(priv, 0x690000 + hdmi, 0x00000001, 0x00000000);
+		return 0;
+	}
+
+	/* AVI InfoFrame */
+	nv_mask(priv, 0x690000 + hdmi, 0x00000001, 0x00000000);
+	nv_wr32(priv, 0x690008 + hdmi, 0x000d0282);
+	nv_wr32(priv, 0x69000c + hdmi, 0x0000006f);
+	nv_wr32(priv, 0x690010 + hdmi, 0x00000000);
+	nv_wr32(priv, 0x690014 + hdmi, 0x00000000);
+	nv_wr32(priv, 0x690018 + hdmi, 0x00000000);
+	nv_mask(priv, 0x690000 + hdmi, 0x00000001, 0x00000001);
+
+	/* ??? InfoFrame? */
+	nv_mask(priv, 0x6900c0 + hdmi, 0x00000001, 0x00000000);
+	nv_wr32(priv, 0x6900cc + hdmi, 0x00000010);
+	nv_mask(priv, 0x6900c0 + hdmi, 0x00000001, 0x00000001);
+
+	/* ??? */
+	nv_wr32(priv, 0x690080 + hdmi, 0x82000000);
+
+	/* HDMI_CTRL */
+	nv_mask(priv, 0x616798 + hoff, 0x401f007f, ctrl);
+	return 0;
+}
diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/nv50.c b/drivers/gpu/drm/nouveau/core/engine/disp/nv50.c
index f8cbb512132f..44a8290aaea5 100644
--- a/drivers/gpu/drm/nouveau/core/engine/disp/nv50.c
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/nv50.c
@@ -29,6 +29,7 @@
 #include <core/enum.h>
 #include <nvif/unpack.h>
 #include <nvif/class.h>
+#include <nvif/event.h>
 
 #include <subdev/bios.h>
 #include <subdev/bios/dcb.h>
@@ -82,6 +83,73 @@ nv50_disp_chan_destroy(struct nv50_disp_chan *chan)
 	nouveau_namedb_destroy(&chan->base);
 }
 
+static void
+nv50_disp_chan_uevent_fini(struct nvkm_event *event, int type, int index)
+{
+	struct nv50_disp_priv *priv = container_of(event, typeof(*priv), uevent);
+	nv_mask(priv, 0x610028, 0x00000001 << index, 0x00000000 << index);
+	nv_wr32(priv, 0x610020, 0x00000001 << index);
+}
+
+static void
+nv50_disp_chan_uevent_init(struct nvkm_event *event, int types, int index)
+{
+	struct nv50_disp_priv *priv = container_of(event, typeof(*priv), uevent);
+	nv_wr32(priv, 0x610020, 0x00000001 << index);
+	nv_mask(priv, 0x610028, 0x00000001 << index, 0x00000001 << index);
+}
+
+void
+nv50_disp_chan_uevent_send(struct nv50_disp_priv *priv, int chid)
+{
+	struct nvif_notify_uevent_rep {
+	} rep;
+
+	nvkm_event_send(&priv->uevent, 1, chid, &rep, sizeof(rep));
+}
+
+int
+nv50_disp_chan_uevent_ctor(struct nouveau_object *object, void *data, u32 size,
+			   struct nvkm_notify *notify)
+{
+	struct nv50_disp_dmac *dmac = (void *)object;
+	union {
+		struct nvif_notify_uevent_req none;
+	} *args = data;
+	int ret;
+
+	if (nvif_unvers(args->none)) {
+		notify->size  = sizeof(struct nvif_notify_uevent_rep);
+		notify->types = 1;
+		notify->index = dmac->base.chid;
+		return 0;
+	}
+
+	return ret;
+}
+
+const struct nvkm_event_func
+nv50_disp_chan_uevent = {
+	.ctor = nv50_disp_chan_uevent_ctor,
+	.init = nv50_disp_chan_uevent_init,
+	.fini = nv50_disp_chan_uevent_fini,
+};
+
+int
+nv50_disp_chan_ntfy(struct nouveau_object *object, u32 type,
+		    struct nvkm_event **pevent)
+{
+	struct nv50_disp_priv *priv = (void *)object->engine;
+	switch (type) {
+	case NV50_DISP_CORE_CHANNEL_DMA_V0_NTFY_UEVENT:
+		*pevent = &priv->uevent;
+		return 0;
+	default:
+		break;
+	}
+	return -EINVAL;
+}
+
 int
 nv50_disp_chan_map(struct nouveau_object *object, u64 *addr, u32 *size)
 {
@@ -195,7 +263,7 @@ nv50_disp_dmac_init(struct nouveau_object *object)
 		return ret;
 
 	/* enable error reporting */
-	nv_mask(priv, 0x610028, 0x00010001 << chid, 0x00010001 << chid);
+	nv_mask(priv, 0x610028, 0x00010000 << chid, 0x00010000 << chid);
 
 	/* initialise channel for dma command submission */
 	nv_wr32(priv, 0x610204 + (chid * 0x0010), dmac->push);
@@ -232,7 +300,7 @@ nv50_disp_dmac_fini(struct nouveau_object *object, bool suspend)
 			return -EBUSY;
 	}
 
-	/* disable error reporting */
+	/* disable error reporting and completion notifications */
 	nv_mask(priv, 0x610028, 0x00010001 << chid, 0x00000000 << chid);
 
 	return nv50_disp_chan_fini(&dmac->base, suspend);
@@ -308,7 +376,7 @@ nv50_disp_mthd_chan(struct nv50_disp_priv *priv, int debug, int head,
 }
 
 const struct nv50_disp_mthd_list
-nv50_disp_mast_mthd_base = {
+nv50_disp_core_mthd_base = {
 	.mthd = 0x0000,
 	.addr = 0x000000,
 	.data = {
@@ -321,7 +389,7 @@ nv50_disp_mast_mthd_base = {
 };
 
 static const struct nv50_disp_mthd_list
-nv50_disp_mast_mthd_dac = {
+nv50_disp_core_mthd_dac = {
 	.mthd = 0x0080,
 	.addr = 0x000008,
 	.data = {
@@ -333,7 +401,7 @@ nv50_disp_mast_mthd_dac = {
 };
 
 const struct nv50_disp_mthd_list
-nv50_disp_mast_mthd_sor = {
+nv50_disp_core_mthd_sor = {
 	.mthd = 0x0040,
 	.addr = 0x000008,
 	.data = {
@@ -343,7 +411,7 @@ nv50_disp_mast_mthd_sor = {
 };
 
 const struct nv50_disp_mthd_list
-nv50_disp_mast_mthd_pior = {
+nv50_disp_core_mthd_pior = {
 	.mthd = 0x0040,
 	.addr = 0x000008,
 	.data = {
@@ -353,7 +421,7 @@ nv50_disp_mast_mthd_pior = {
 };
 
 static const struct nv50_disp_mthd_list
-nv50_disp_mast_mthd_head = {
+nv50_disp_core_mthd_head = {
 	.mthd = 0x0400,
 	.addr = 0x000540,
 	.data = {
@@ -400,21 +468,21 @@ nv50_disp_mast_mthd_head = {
 };
 
 static const struct nv50_disp_mthd_chan
-nv50_disp_mast_mthd_chan = {
+nv50_disp_core_mthd_chan = {
 	.name = "Core",
 	.addr = 0x000000,
 	.data = {
-		{ "Global", 1, &nv50_disp_mast_mthd_base },
-		{    "DAC", 3, &nv50_disp_mast_mthd_dac  },
-		{    "SOR", 2, &nv50_disp_mast_mthd_sor  },
-		{   "PIOR", 3, &nv50_disp_mast_mthd_pior },
-		{   "HEAD", 2, &nv50_disp_mast_mthd_head },
+		{ "Global", 1, &nv50_disp_core_mthd_base },
+		{    "DAC", 3, &nv50_disp_core_mthd_dac  },
+		{    "SOR", 2, &nv50_disp_core_mthd_sor  },
+		{   "PIOR", 3, &nv50_disp_core_mthd_pior },
+		{   "HEAD", 2, &nv50_disp_core_mthd_head },
 		{}
 	}
 };
 
 int
-nv50_disp_mast_ctor(struct nouveau_object *parent,
+nv50_disp_core_ctor(struct nouveau_object *parent,
 		    struct nouveau_object *engine,
 		    struct nouveau_oclass *oclass, void *data, u32 size,
 		    struct nouveau_object **pobject)
@@ -443,7 +511,7 @@ nv50_disp_mast_ctor(struct nouveau_object *parent,
 }
 
 static int
-nv50_disp_mast_init(struct nouveau_object *object)
+nv50_disp_core_init(struct nouveau_object *object)
 {
 	struct nv50_disp_priv *priv = (void *)object->engine;
 	struct nv50_disp_dmac *mast = (void *)object;
@@ -454,7 +522,7 @@ nv50_disp_mast_init(struct nouveau_object *object)
 		return ret;
 
 	/* enable error reporting */
-	nv_mask(priv, 0x610028, 0x00010001, 0x00010001);
+	nv_mask(priv, 0x610028, 0x00010000, 0x00010000);
 
 	/* attempt to unstick channel from some unknown state */
 	if ((nv_rd32(priv, 0x610200) & 0x009f0000) == 0x00020000)
@@ -480,7 +548,7 @@ nv50_disp_mast_init(struct nouveau_object *object)
 }
 
 static int
-nv50_disp_mast_fini(struct nouveau_object *object, bool suspend)
+nv50_disp_core_fini(struct nouveau_object *object, bool suspend)
 {
 	struct nv50_disp_priv *priv = (void *)object->engine;
 	struct nv50_disp_dmac *mast = (void *)object;
@@ -494,19 +562,20 @@ nv50_disp_mast_fini(struct nouveau_object *object, bool suspend)
 			return -EBUSY;
 	}
 
-	/* disable error reporting */
+	/* disable error reporting and completion notifications */
 	nv_mask(priv, 0x610028, 0x00010001, 0x00000000);
 
 	return nv50_disp_chan_fini(&mast->base, suspend);
 }
 
 struct nv50_disp_chan_impl
-nv50_disp_mast_ofuncs = {
-	.base.ctor = nv50_disp_mast_ctor,
+nv50_disp_core_ofuncs = {
+	.base.ctor = nv50_disp_core_ctor,
 	.base.dtor = nv50_disp_dmac_dtor,
-	.base.init = nv50_disp_mast_init,
-	.base.fini = nv50_disp_mast_fini,
+	.base.init = nv50_disp_core_init,
+	.base.fini = nv50_disp_core_fini,
 	.base.map  = nv50_disp_chan_map,
+	.base.ntfy = nv50_disp_chan_ntfy,
 	.base.rd32 = nv50_disp_chan_rd32,
 	.base.wr32 = nv50_disp_chan_wr32,
 	.chid = 0,
@@ -519,7 +588,7 @@ nv50_disp_mast_ofuncs = {
  ******************************************************************************/
 
 static const struct nv50_disp_mthd_list
-nv50_disp_sync_mthd_base = {
+nv50_disp_base_mthd_base = {
 	.mthd = 0x0000,
 	.addr = 0x000000,
 	.data = {
@@ -544,7 +613,7 @@ nv50_disp_sync_mthd_base = {
 };
 
 const struct nv50_disp_mthd_list
-nv50_disp_sync_mthd_image = {
+nv50_disp_base_mthd_image = {
 	.mthd = 0x0400,
 	.addr = 0x000000,
 	.data = {
@@ -558,18 +627,18 @@ nv50_disp_sync_mthd_image = {
 };
 
 static const struct nv50_disp_mthd_chan
-nv50_disp_sync_mthd_chan = {
+nv50_disp_base_mthd_chan = {
 	.name = "Base",
 	.addr = 0x000540,
 	.data = {
-		{ "Global", 1, &nv50_disp_sync_mthd_base },
-		{  "Image", 2, &nv50_disp_sync_mthd_image },
+		{ "Global", 1, &nv50_disp_base_mthd_base },
+		{  "Image", 2, &nv50_disp_base_mthd_image },
 		{}
 	}
 };
 
 int
-nv50_disp_sync_ctor(struct nouveau_object *parent,
+nv50_disp_base_ctor(struct nouveau_object *parent,
 		    struct nouveau_object *engine,
 		    struct nouveau_oclass *oclass, void *data, u32 size,
 		    struct nouveau_object **pobject)
@@ -602,11 +671,12 @@ nv50_disp_sync_ctor(struct nouveau_object *parent,
 }
 
 struct nv50_disp_chan_impl
-nv50_disp_sync_ofuncs = {
-	.base.ctor = nv50_disp_sync_ctor,
+nv50_disp_base_ofuncs = {
+	.base.ctor = nv50_disp_base_ctor,
 	.base.dtor = nv50_disp_dmac_dtor,
 	.base.init = nv50_disp_dmac_init,
 	.base.fini = nv50_disp_dmac_fini,
+	.base.ntfy = nv50_disp_chan_ntfy,
 	.base.map  = nv50_disp_chan_map,
 	.base.rd32 = nv50_disp_chan_rd32,
 	.base.wr32 = nv50_disp_chan_wr32,
@@ -696,6 +766,7 @@ nv50_disp_ovly_ofuncs = {
 	.base.dtor = nv50_disp_dmac_dtor,
 	.base.init = nv50_disp_dmac_init,
 	.base.fini = nv50_disp_dmac_fini,
+	.base.ntfy = nv50_disp_chan_ntfy,
 	.base.map  = nv50_disp_chan_map,
 	.base.rd32 = nv50_disp_chan_rd32,
 	.base.wr32 = nv50_disp_chan_wr32,
@@ -813,6 +884,7 @@ nv50_disp_oimm_ofuncs = {
 	.base.dtor = nv50_disp_pioc_dtor,
 	.base.init = nv50_disp_pioc_init,
 	.base.fini = nv50_disp_pioc_fini,
+	.base.ntfy = nv50_disp_chan_ntfy,
 	.base.map  = nv50_disp_chan_map,
 	.base.rd32 = nv50_disp_chan_rd32,
 	.base.wr32 = nv50_disp_chan_wr32,
@@ -860,6 +932,7 @@ nv50_disp_curs_ofuncs = {
 	.base.dtor = nv50_disp_pioc_dtor,
 	.base.init = nv50_disp_pioc_init,
 	.base.fini = nv50_disp_pioc_fini,
+	.base.ntfy = nv50_disp_chan_ntfy,
 	.base.map  = nv50_disp_chan_map,
 	.base.rd32 = nv50_disp_chan_rd32,
 	.base.wr32 = nv50_disp_chan_wr32,
@@ -871,7 +944,7 @@ nv50_disp_curs_ofuncs = {
  ******************************************************************************/
 
 int
-nv50_disp_base_scanoutpos(NV50_DISP_MTHD_V0)
+nv50_disp_main_scanoutpos(NV50_DISP_MTHD_V0)
 {
 	const u32 blanke = nv_rd32(priv, 0x610aec + (head * 0x540));
 	const u32 blanks = nv_rd32(priv, 0x610af4 + (head * 0x540));
@@ -903,7 +976,7 @@ nv50_disp_base_scanoutpos(NV50_DISP_MTHD_V0)
 }
 
 int
-nv50_disp_base_mthd(struct nouveau_object *object, u32 mthd,
+nv50_disp_main_mthd(struct nouveau_object *object, u32 mthd,
 		    void *data, u32 size)
 {
 	const struct nv50_disp_impl *impl = (void *)nv_oclass(object->engine);
@@ -1027,7 +1100,7 @@ nv50_disp_base_mthd(struct nouveau_object *object, u32 mthd,
 }
 
 int
-nv50_disp_base_ctor(struct nouveau_object *parent,
+nv50_disp_main_ctor(struct nouveau_object *parent,
 		    struct nouveau_object *engine,
 		    struct nouveau_oclass *oclass, void *data, u32 size,
 		    struct nouveau_object **pobject)
@@ -1047,7 +1120,7 @@ nv50_disp_base_ctor(struct nouveau_object *parent,
 }
 
 void
-nv50_disp_base_dtor(struct nouveau_object *object)
+nv50_disp_main_dtor(struct nouveau_object *object)
 {
 	struct nv50_disp_base *base = (void *)object;
 	nouveau_ramht_ref(NULL, &base->ramht);
@@ -1055,7 +1128,7 @@ nv50_disp_base_dtor(struct nouveau_object *object)
 }
 
 static int
-nv50_disp_base_init(struct nouveau_object *object)
+nv50_disp_main_init(struct nouveau_object *object)
 {
 	struct nv50_disp_priv *priv = (void *)object->engine;
 	struct nv50_disp_base *base = (void *)object;
@@ -1123,7 +1196,7 @@ nv50_disp_base_init(struct nouveau_object *object)
 }
 
 static int
-nv50_disp_base_fini(struct nouveau_object *object, bool suspend)
+nv50_disp_main_fini(struct nouveau_object *object, bool suspend)
 {
 	struct nv50_disp_priv *priv = (void *)object->engine;
 	struct nv50_disp_base *base = (void *)object;
@@ -1136,25 +1209,25 @@ nv50_disp_base_fini(struct nouveau_object *object, bool suspend)
 }
 
 struct nouveau_ofuncs
-nv50_disp_base_ofuncs = {
-	.ctor = nv50_disp_base_ctor,
-	.dtor = nv50_disp_base_dtor,
-	.init = nv50_disp_base_init,
-	.fini = nv50_disp_base_fini,
-	.mthd = nv50_disp_base_mthd,
+nv50_disp_main_ofuncs = {
+	.ctor = nv50_disp_main_ctor,
+	.dtor = nv50_disp_main_dtor,
+	.init = nv50_disp_main_init,
+	.fini = nv50_disp_main_fini,
+	.mthd = nv50_disp_main_mthd,
 	.ntfy = nouveau_disp_ntfy,
 };
 
 static struct nouveau_oclass
-nv50_disp_base_oclass[] = {
-	{ NV50_DISP, &nv50_disp_base_ofuncs },
+nv50_disp_main_oclass[] = {
+	{ NV50_DISP, &nv50_disp_main_ofuncs },
 	{}
 };
 
 static struct nouveau_oclass
 nv50_disp_sclass[] = {
-	{ NV50_DISP_CORE_CHANNEL_DMA, &nv50_disp_mast_ofuncs.base },
-	{ NV50_DISP_BASE_CHANNEL_DMA, &nv50_disp_sync_ofuncs.base },
+	{ NV50_DISP_CORE_CHANNEL_DMA, &nv50_disp_core_ofuncs.base },
+	{ NV50_DISP_BASE_CHANNEL_DMA, &nv50_disp_base_ofuncs.base },
 	{ NV50_DISP_OVERLAY_CHANNEL_DMA, &nv50_disp_ovly_ofuncs.base },
 	{ NV50_DISP_OVERLAY, &nv50_disp_oimm_ofuncs.base },
 	{ NV50_DISP_CURSOR, &nv50_disp_curs_ofuncs.base },
@@ -1559,7 +1632,7 @@ nv50_disp_intr_unk20_1(struct nv50_disp_priv *priv, int head)
 }
 
 static void
-nv50_disp_intr_unk20_2_dp(struct nv50_disp_priv *priv,
+nv50_disp_intr_unk20_2_dp(struct nv50_disp_priv *priv, int head,
 			  struct dcb_output *outp, u32 pclk)
 {
 	const int link = !(outp->sorconf.link & 1);
@@ -1568,24 +1641,36 @@ nv50_disp_intr_unk20_2_dp(struct nv50_disp_priv *priv,
 	const u32 loff = (link * 0x080) + soff;
 	const u32 ctrl = nv_rd32(priv, 0x610794 + (or * 8));
 	const u32 symbol = 100000;
-	u32 dpctrl = nv_rd32(priv, 0x61c10c + loff) & 0x0000f0000;
+	const s32 vactive = nv_rd32(priv, 0x610af8 + (head * 0x540)) & 0xffff;
+	const s32 vblanke = nv_rd32(priv, 0x610ae8 + (head * 0x540)) & 0xffff;
+	const s32 vblanks = nv_rd32(priv, 0x610af0 + (head * 0x540)) & 0xffff;
+	u32 dpctrl = nv_rd32(priv, 0x61c10c + loff);
 	u32 clksor = nv_rd32(priv, 0x614300 + soff);
 	int bestTU = 0, bestVTUi = 0, bestVTUf = 0, bestVTUa = 0;
 	int TU, VTUi, VTUf, VTUa;
 	u64 link_data_rate, link_ratio, unk;
 	u32 best_diff = 64 * symbol;
 	u32 link_nr, link_bw, bits;
-
-	/* calculate packed data rate for each lane */
-	if      (dpctrl > 0x00030000) link_nr = 4;
-	else if (dpctrl > 0x00010000) link_nr = 2;
-	else			      link_nr = 1;
-
-	if (clksor & 0x000c0000)
-		link_bw = 270000;
-	else
-		link_bw = 162000;
-
+	u64 value;
+
+	link_bw = (clksor & 0x000c0000) ? 270000 : 162000;
+	link_nr = hweight32(dpctrl & 0x000f0000);
+
+	/* symbols/hblank - algorithm taken from comments in tegra driver */
+	value = vblanke + vactive - vblanks - 7;
+	value = value * link_bw;
+	do_div(value, pclk);
+	value = value - (3 * !!(dpctrl & 0x00004000)) - (12 / link_nr);
+	nv_mask(priv, 0x61c1e8 + soff, 0x0000ffff, value);
+
+	/* symbols/vblank - algorithm taken from comments in tegra driver */
+	value = vblanks - vblanke - 25;
+	value = value * link_bw;
+	do_div(value, pclk);
+	value = value - ((36 / link_nr) + 3) - 1;
+	nv_mask(priv, 0x61c1ec + soff, 0x00ffffff, value);
+
+	/* watermark / activesym */
 	if      ((ctrl & 0xf0000) == 0x60000) bits = 30;
 	else if ((ctrl & 0xf0000) == 0x50000) bits = 24;
 	else                                  bits = 18;
@@ -1731,7 +1816,7 @@ nv50_disp_intr_unk20_2(struct nv50_disp_priv *priv, int head)
 	} else
 	if (!outp->info.location) {
 		if (outp->info.type == DCB_OUTPUT_DP)
-			nv50_disp_intr_unk20_2_dp(priv, &outp->info, pclk);
+			nv50_disp_intr_unk20_2_dp(priv, head, &outp->info, pclk);
 		oreg = 0x614300 + (ffs(outp->info.or) - 1) * 0x800;
 		oval = (conf & 0x0100) ? 0x00000101 : 0x00000000;
 		hval = 0x00000000;
@@ -1847,6 +1932,12 @@ nv50_disp_intr(struct nouveau_subdev *subdev)
 		intr0 &= ~(0x00010000 << chid);
 	}
 
+	while (intr0 & 0x0000001f) {
+		u32 chid = __ffs(intr0 & 0x0000001f);
+		nv50_disp_chan_uevent_send(priv, chid);
+		intr0 &= ~(0x00000001 << chid);
+	}
+
 	if (intr1 & 0x00000004) {
 		nouveau_disp_vblank(&priv->base, 0);
 		nv_wr32(priv, 0x610024, 0x00000004);
@@ -1881,7 +1972,11 @@ nv50_disp_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	if (ret)
 		return ret;
 
-	nv_engine(priv)->sclass = nv50_disp_base_oclass;
+	ret = nvkm_event_init(&nv50_disp_chan_uevent, 1, 9, &priv->uevent);
+	if (ret)
+		return ret;
+
+	nv_engine(priv)->sclass = nv50_disp_main_oclass;
 	nv_engine(priv)->cclass = &nv50_disp_cclass;
 	nv_subdev(priv)->intr = nv50_disp_intr;
 	INIT_WORK(&priv->supervisor, nv50_disp_intr_supervisor);
@@ -1914,9 +2009,9 @@ nv50_disp_oclass = &(struct nv50_disp_impl) {
 	},
 	.base.vblank = &nv50_disp_vblank_func,
 	.base.outp =  nv50_disp_outp_sclass,
-	.mthd.core = &nv50_disp_mast_mthd_chan,
-	.mthd.base = &nv50_disp_sync_mthd_chan,
+	.mthd.core = &nv50_disp_core_mthd_chan,
+	.mthd.base = &nv50_disp_base_mthd_chan,
 	.mthd.ovly = &nv50_disp_ovly_mthd_chan,
 	.mthd.prev = 0x000004,
-	.head.scanoutpos = nv50_disp_base_scanoutpos,
+	.head.scanoutpos = nv50_disp_main_scanoutpos,
 }.base.base;
diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/nv50.h b/drivers/gpu/drm/nouveau/core/engine/disp/nv50.h
index 8ab14461f70c..7f08078ee925 100644
--- a/drivers/gpu/drm/nouveau/core/engine/disp/nv50.h
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/nv50.h
@@ -26,6 +26,8 @@ struct nv50_disp_priv {
 	struct work_struct supervisor;
 	u32 super;
 
+	struct nvkm_event uevent;
+
 	struct {
 		int nr;
 	} head;
@@ -40,6 +42,7 @@ struct nv50_disp_priv {
 		int (*hda_eld)(NV50_DISP_MTHD_V1);
 		int (*hdmi)(NV50_DISP_MTHD_V1);
 		u32 lvdsconf;
+		void (*magic)(struct nvkm_output *);
 	} sor;
 	struct {
 		int nr;
@@ -61,10 +64,10 @@ struct nv50_disp_impl {
 	} head;
 };
 
-int nv50_disp_base_scanoutpos(NV50_DISP_MTHD_V0);
-int nv50_disp_base_mthd(struct nouveau_object *, u32, void *, u32);
+int nv50_disp_main_scanoutpos(NV50_DISP_MTHD_V0);
+int nv50_disp_main_mthd(struct nouveau_object *, u32, void *, u32);
 
-int nvd0_disp_base_scanoutpos(NV50_DISP_MTHD_V0);
+int nvd0_disp_main_scanoutpos(NV50_DISP_MTHD_V0);
 
 int nv50_dac_power(NV50_DISP_MTHD_V1);
 int nv50_dac_sense(NV50_DISP_MTHD_V1);
@@ -75,6 +78,7 @@ int nvd0_hda_eld(NV50_DISP_MTHD_V1);
 int nv84_hdmi_ctrl(NV50_DISP_MTHD_V1);
 int nva3_hdmi_ctrl(NV50_DISP_MTHD_V1);
 int nvd0_hdmi_ctrl(NV50_DISP_MTHD_V1);
+int nve0_hdmi_ctrl(NV50_DISP_MTHD_V1);
 
 int nv50_sor_power(NV50_DISP_MTHD_V1);
 
@@ -116,9 +120,16 @@ struct nv50_disp_chan {
 	int chid;
 };
 
+int  nv50_disp_chan_ntfy(struct nouveau_object *, u32, struct nvkm_event **);
 int  nv50_disp_chan_map(struct nouveau_object *, u64 *, u32 *);
 u32  nv50_disp_chan_rd32(struct nouveau_object *, u64);
 void nv50_disp_chan_wr32(struct nouveau_object *, u64, u32);
+extern const struct nvkm_event_func nv50_disp_chan_uevent;
+int  nv50_disp_chan_uevent_ctor(struct nouveau_object *, void *, u32,
+				struct nvkm_notify *);
+void nv50_disp_chan_uevent_send(struct nv50_disp_priv *, int);
+
+extern const struct nvkm_event_func nvd0_disp_chan_uevent;
 
 #define nv50_disp_chan_init(a)                                                 \
 	nouveau_namedb_init(&(a)->base)
@@ -159,18 +170,18 @@ struct nv50_disp_mthd_chan {
 	} data[];
 };
 
-extern struct nv50_disp_chan_impl nv50_disp_mast_ofuncs;
-int nv50_disp_mast_ctor(struct nouveau_object *, struct nouveau_object *,
+extern struct nv50_disp_chan_impl nv50_disp_core_ofuncs;
+int nv50_disp_core_ctor(struct nouveau_object *, struct nouveau_object *,
 			struct nouveau_oclass *, void *, u32,
 			struct nouveau_object **);
-extern const struct nv50_disp_mthd_list nv50_disp_mast_mthd_base;
-extern const struct nv50_disp_mthd_list nv50_disp_mast_mthd_sor;
-extern const struct nv50_disp_mthd_list nv50_disp_mast_mthd_pior;
-extern struct nv50_disp_chan_impl nv50_disp_sync_ofuncs;
-int nv50_disp_sync_ctor(struct nouveau_object *, struct nouveau_object *,
+extern const struct nv50_disp_mthd_list nv50_disp_core_mthd_base;
+extern const struct nv50_disp_mthd_list nv50_disp_core_mthd_sor;
+extern const struct nv50_disp_mthd_list nv50_disp_core_mthd_pior;
+extern struct nv50_disp_chan_impl nv50_disp_base_ofuncs;
+int nv50_disp_base_ctor(struct nouveau_object *, struct nouveau_object *,
 			struct nouveau_oclass *, void *, u32,
 			struct nouveau_object **);
-extern const struct nv50_disp_mthd_list nv50_disp_sync_mthd_image;
+extern const struct nv50_disp_mthd_list nv50_disp_base_mthd_image;
 extern struct nv50_disp_chan_impl nv50_disp_ovly_ofuncs;
 int nv50_disp_ovly_ctor(struct nouveau_object *, struct nouveau_object *,
 			struct nouveau_oclass *, void *, u32,
@@ -184,12 +195,12 @@ extern struct nv50_disp_chan_impl nv50_disp_curs_ofuncs;
 int nv50_disp_curs_ctor(struct nouveau_object *, struct nouveau_object *,
 			struct nouveau_oclass *, void *, u32,
 			struct nouveau_object **);
-extern struct nouveau_ofuncs nv50_disp_base_ofuncs;
-int  nv50_disp_base_ctor(struct nouveau_object *, struct nouveau_object *,
+extern struct nouveau_ofuncs nv50_disp_main_ofuncs;
+int  nv50_disp_main_ctor(struct nouveau_object *, struct nouveau_object *,
 			 struct nouveau_oclass *, void *, u32,
 			 struct nouveau_object **);
-void nv50_disp_base_dtor(struct nouveau_object *);
-extern struct nouveau_omthds nv50_disp_base_omthds[];
+void nv50_disp_main_dtor(struct nouveau_object *);
+extern struct nouveau_omthds nv50_disp_main_omthds[];
 extern struct nouveau_oclass nv50_disp_cclass;
 void nv50_disp_mthd_chan(struct nv50_disp_priv *, int debug, int head,
 			 const struct nv50_disp_mthd_chan *);
@@ -197,31 +208,31 @@ void nv50_disp_intr_supervisor(struct work_struct *);
 void nv50_disp_intr(struct nouveau_subdev *);
 extern const struct nvkm_event_func nv50_disp_vblank_func;
 
-extern const struct nv50_disp_mthd_chan nv84_disp_mast_mthd_chan;
-extern const struct nv50_disp_mthd_list nv84_disp_mast_mthd_dac;
-extern const struct nv50_disp_mthd_list nv84_disp_mast_mthd_head;
-extern const struct nv50_disp_mthd_chan nv84_disp_sync_mthd_chan;
+extern const struct nv50_disp_mthd_chan nv84_disp_core_mthd_chan;
+extern const struct nv50_disp_mthd_list nv84_disp_core_mthd_dac;
+extern const struct nv50_disp_mthd_list nv84_disp_core_mthd_head;
+extern const struct nv50_disp_mthd_chan nv84_disp_base_mthd_chan;
 extern const struct nv50_disp_mthd_chan nv84_disp_ovly_mthd_chan;
 
-extern const struct nv50_disp_mthd_chan nv94_disp_mast_mthd_chan;
+extern const struct nv50_disp_mthd_chan nv94_disp_core_mthd_chan;
 
-extern struct nv50_disp_chan_impl nvd0_disp_mast_ofuncs;
-extern const struct nv50_disp_mthd_list nvd0_disp_mast_mthd_base;
-extern const struct nv50_disp_mthd_list nvd0_disp_mast_mthd_dac;
-extern const struct nv50_disp_mthd_list nvd0_disp_mast_mthd_sor;
-extern const struct nv50_disp_mthd_list nvd0_disp_mast_mthd_pior;
-extern struct nv50_disp_chan_impl nvd0_disp_sync_ofuncs;
+extern struct nv50_disp_chan_impl nvd0_disp_core_ofuncs;
+extern const struct nv50_disp_mthd_list nvd0_disp_core_mthd_base;
+extern const struct nv50_disp_mthd_list nvd0_disp_core_mthd_dac;
+extern const struct nv50_disp_mthd_list nvd0_disp_core_mthd_sor;
+extern const struct nv50_disp_mthd_list nvd0_disp_core_mthd_pior;
+extern struct nv50_disp_chan_impl nvd0_disp_base_ofuncs;
 extern struct nv50_disp_chan_impl nvd0_disp_ovly_ofuncs;
-extern const struct nv50_disp_mthd_chan nvd0_disp_sync_mthd_chan;
+extern const struct nv50_disp_mthd_chan nvd0_disp_base_mthd_chan;
 extern struct nv50_disp_chan_impl nvd0_disp_oimm_ofuncs;
 extern struct nv50_disp_chan_impl nvd0_disp_curs_ofuncs;
-extern struct nouveau_ofuncs nvd0_disp_base_ofuncs;
+extern struct nouveau_ofuncs nvd0_disp_main_ofuncs;
 extern struct nouveau_oclass nvd0_disp_cclass;
 void nvd0_disp_intr_supervisor(struct work_struct *);
 void nvd0_disp_intr(struct nouveau_subdev *);
 extern const struct nvkm_event_func nvd0_disp_vblank_func;
 
-extern const struct nv50_disp_mthd_chan nve0_disp_mast_mthd_chan;
+extern const struct nv50_disp_mthd_chan nve0_disp_core_mthd_chan;
 extern const struct nv50_disp_mthd_chan nve0_disp_ovly_mthd_chan;
 
 extern struct nvkm_output_dp_impl nv50_pior_dp_impl;
@@ -232,6 +243,10 @@ int nv94_sor_dp_lnk_pwr(struct nvkm_output_dp *, int);
 extern struct nouveau_oclass *nv94_disp_outp_sclass[];
 
 extern struct nvkm_output_dp_impl nvd0_sor_dp_impl;
+int nvd0_sor_dp_lnk_ctl(struct nvkm_output_dp *, int, int, bool);
 extern struct nouveau_oclass *nvd0_disp_outp_sclass[];
 
+void gm204_sor_magic(struct nvkm_output *outp);
+extern struct nvkm_output_dp_impl gm204_sor_dp_impl;
+
 #endif
diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/nv84.c b/drivers/gpu/drm/nouveau/core/engine/disp/nv84.c
index 788ced1b6182..13eff5e4ee51 100644
--- a/drivers/gpu/drm/nouveau/core/engine/disp/nv84.c
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/nv84.c
@@ -34,7 +34,7 @@
  ******************************************************************************/
 
 const struct nv50_disp_mthd_list
-nv84_disp_mast_mthd_dac = {
+nv84_disp_core_mthd_dac = {
 	.mthd = 0x0080,
 	.addr = 0x000008,
 	.data = {
@@ -46,7 +46,7 @@ nv84_disp_mast_mthd_dac = {
 };
 
 const struct nv50_disp_mthd_list
-nv84_disp_mast_mthd_head = {
+nv84_disp_core_mthd_head = {
 	.mthd = 0x0400,
 	.addr = 0x000540,
 	.data = {
@@ -98,15 +98,15 @@ nv84_disp_mast_mthd_head = {
 };
 
 const struct nv50_disp_mthd_chan
-nv84_disp_mast_mthd_chan = {
+nv84_disp_core_mthd_chan = {
 	.name = "Core",
 	.addr = 0x000000,
 	.data = {
-		{ "Global", 1, &nv50_disp_mast_mthd_base },
-		{    "DAC", 3, &nv84_disp_mast_mthd_dac  },
-		{    "SOR", 2, &nv50_disp_mast_mthd_sor  },
-		{   "PIOR", 3, &nv50_disp_mast_mthd_pior },
-		{   "HEAD", 2, &nv84_disp_mast_mthd_head },
+		{ "Global", 1, &nv50_disp_core_mthd_base },
+		{    "DAC", 3, &nv84_disp_core_mthd_dac  },
+		{    "SOR", 2, &nv50_disp_core_mthd_sor  },
+		{   "PIOR", 3, &nv50_disp_core_mthd_pior },
+		{   "HEAD", 2, &nv84_disp_core_mthd_head },
 		{}
 	}
 };
@@ -116,7 +116,7 @@ nv84_disp_mast_mthd_chan = {
  ******************************************************************************/
 
 static const struct nv50_disp_mthd_list
-nv84_disp_sync_mthd_base = {
+nv84_disp_base_mthd_base = {
 	.mthd = 0x0000,
 	.addr = 0x000000,
 	.data = {
@@ -146,12 +146,12 @@ nv84_disp_sync_mthd_base = {
 };
 
 const struct nv50_disp_mthd_chan
-nv84_disp_sync_mthd_chan = {
+nv84_disp_base_mthd_chan = {
 	.name = "Base",
 	.addr = 0x000540,
 	.data = {
-		{ "Global", 1, &nv84_disp_sync_mthd_base },
-		{  "Image", 2, &nv50_disp_sync_mthd_image },
+		{ "Global", 1, &nv84_disp_base_mthd_base },
+		{  "Image", 2, &nv50_disp_base_mthd_image },
 		{}
 	}
 };
@@ -204,8 +204,8 @@ nv84_disp_ovly_mthd_chan = {
 
 static struct nouveau_oclass
 nv84_disp_sclass[] = {
-	{ G82_DISP_CORE_CHANNEL_DMA, &nv50_disp_mast_ofuncs.base },
-	{ G82_DISP_BASE_CHANNEL_DMA, &nv50_disp_sync_ofuncs.base },
+	{ G82_DISP_CORE_CHANNEL_DMA, &nv50_disp_core_ofuncs.base },
+	{ G82_DISP_BASE_CHANNEL_DMA, &nv50_disp_base_ofuncs.base },
 	{ G82_DISP_OVERLAY_CHANNEL_DMA, &nv50_disp_ovly_ofuncs.base },
 	{ G82_DISP_OVERLAY, &nv50_disp_oimm_ofuncs.base },
 	{ G82_DISP_CURSOR, &nv50_disp_curs_ofuncs.base },
@@ -213,8 +213,8 @@ nv84_disp_sclass[] = {
 };
 
 static struct nouveau_oclass
-nv84_disp_base_oclass[] = {
-	{ G82_DISP, &nv50_disp_base_ofuncs },
+nv84_disp_main_oclass[] = {
+	{ G82_DISP, &nv50_disp_main_ofuncs },
 	{}
 };
 
@@ -236,7 +236,11 @@ nv84_disp_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	if (ret)
 		return ret;
 
-	nv_engine(priv)->sclass = nv84_disp_base_oclass;
+	ret = nvkm_event_init(&nv50_disp_chan_uevent, 1, 9, &priv->uevent);
+	if (ret)
+		return ret;
+
+	nv_engine(priv)->sclass = nv84_disp_main_oclass;
 	nv_engine(priv)->cclass = &nv50_disp_cclass;
 	nv_subdev(priv)->intr = nv50_disp_intr;
 	INIT_WORK(&priv->supervisor, nv50_disp_intr_supervisor);
@@ -264,9 +268,9 @@ nv84_disp_oclass = &(struct nv50_disp_impl) {
 	},
 	.base.vblank = &nv50_disp_vblank_func,
 	.base.outp =  nv50_disp_outp_sclass,
-	.mthd.core = &nv84_disp_mast_mthd_chan,
-	.mthd.base = &nv84_disp_sync_mthd_chan,
+	.mthd.core = &nv84_disp_core_mthd_chan,
+	.mthd.base = &nv84_disp_base_mthd_chan,
 	.mthd.ovly = &nv84_disp_ovly_mthd_chan,
 	.mthd.prev = 0x000004,
-	.head.scanoutpos = nv50_disp_base_scanoutpos,
+	.head.scanoutpos = nv50_disp_main_scanoutpos,
 }.base.base;
diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/nv94.c b/drivers/gpu/drm/nouveau/core/engine/disp/nv94.c
index fa79de906eae..2bb7ac5cd0e6 100644
--- a/drivers/gpu/drm/nouveau/core/engine/disp/nv94.c
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/nv94.c
@@ -34,7 +34,7 @@
  ******************************************************************************/
 
 const struct nv50_disp_mthd_list
-nv94_disp_mast_mthd_sor = {
+nv94_disp_core_mthd_sor = {
 	.mthd = 0x0040,
 	.addr = 0x000008,
 	.data = {
@@ -44,15 +44,15 @@ nv94_disp_mast_mthd_sor = {
 };
 
 const struct nv50_disp_mthd_chan
-nv94_disp_mast_mthd_chan = {
+nv94_disp_core_mthd_chan = {
 	.name = "Core",
 	.addr = 0x000000,
 	.data = {
-		{ "Global", 1, &nv50_disp_mast_mthd_base },
-		{    "DAC", 3, &nv84_disp_mast_mthd_dac  },
-		{    "SOR", 4, &nv94_disp_mast_mthd_sor  },
-		{   "PIOR", 3, &nv50_disp_mast_mthd_pior },
-		{   "HEAD", 2, &nv84_disp_mast_mthd_head },
+		{ "Global", 1, &nv50_disp_core_mthd_base },
+		{    "DAC", 3, &nv84_disp_core_mthd_dac  },
+		{    "SOR", 4, &nv94_disp_core_mthd_sor  },
+		{   "PIOR", 3, &nv50_disp_core_mthd_pior },
+		{   "HEAD", 2, &nv84_disp_core_mthd_head },
 		{}
 	}
 };
@@ -63,8 +63,8 @@ nv94_disp_mast_mthd_chan = {
 
 static struct nouveau_oclass
 nv94_disp_sclass[] = {
-	{ GT206_DISP_CORE_CHANNEL_DMA, &nv50_disp_mast_ofuncs.base },
-	{ GT200_DISP_BASE_CHANNEL_DMA, &nv50_disp_sync_ofuncs.base },
+	{ GT206_DISP_CORE_CHANNEL_DMA, &nv50_disp_core_ofuncs.base },
+	{ GT200_DISP_BASE_CHANNEL_DMA, &nv50_disp_base_ofuncs.base },
 	{ GT200_DISP_OVERLAY_CHANNEL_DMA, &nv50_disp_ovly_ofuncs.base },
 	{ G82_DISP_OVERLAY, &nv50_disp_oimm_ofuncs.base },
 	{ G82_DISP_CURSOR, &nv50_disp_curs_ofuncs.base },
@@ -72,8 +72,8 @@ nv94_disp_sclass[] = {
 };
 
 static struct nouveau_oclass
-nv94_disp_base_oclass[] = {
-	{ GT206_DISP, &nv50_disp_base_ofuncs },
+nv94_disp_main_oclass[] = {
+	{ GT206_DISP, &nv50_disp_main_ofuncs },
 	{}
 };
 
@@ -95,7 +95,11 @@ nv94_disp_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	if (ret)
 		return ret;
 
-	nv_engine(priv)->sclass = nv94_disp_base_oclass;
+	ret = nvkm_event_init(&nv50_disp_chan_uevent, 1, 9, &priv->uevent);
+	if (ret)
+		return ret;
+
+	nv_engine(priv)->sclass = nv94_disp_main_oclass;
 	nv_engine(priv)->cclass = &nv50_disp_cclass;
 	nv_subdev(priv)->intr = nv50_disp_intr;
 	INIT_WORK(&priv->supervisor, nv50_disp_intr_supervisor);
@@ -130,9 +134,9 @@ nv94_disp_oclass = &(struct nv50_disp_impl) {
 	},
 	.base.vblank = &nv50_disp_vblank_func,
 	.base.outp =  nv94_disp_outp_sclass,
-	.mthd.core = &nv94_disp_mast_mthd_chan,
-	.mthd.base = &nv84_disp_sync_mthd_chan,
+	.mthd.core = &nv94_disp_core_mthd_chan,
+	.mthd.base = &nv84_disp_base_mthd_chan,
 	.mthd.ovly = &nv84_disp_ovly_mthd_chan,
 	.mthd.prev = 0x000004,
-	.head.scanoutpos = nv50_disp_base_scanoutpos,
+	.head.scanoutpos = nv50_disp_main_scanoutpos,
 }.base.base;
diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/nva0.c b/drivers/gpu/drm/nouveau/core/engine/disp/nva0.c
index 7af15f5d48dc..b32456c9494f 100644
--- a/drivers/gpu/drm/nouveau/core/engine/disp/nva0.c
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/nva0.c
@@ -80,8 +80,8 @@ nva0_disp_ovly_mthd_chan = {
 
 static struct nouveau_oclass
 nva0_disp_sclass[] = {
-	{ GT200_DISP_CORE_CHANNEL_DMA, &nv50_disp_mast_ofuncs.base },
-	{ GT200_DISP_BASE_CHANNEL_DMA, &nv50_disp_sync_ofuncs.base },
+	{ GT200_DISP_CORE_CHANNEL_DMA, &nv50_disp_core_ofuncs.base },
+	{ GT200_DISP_BASE_CHANNEL_DMA, &nv50_disp_base_ofuncs.base },
 	{ GT200_DISP_OVERLAY_CHANNEL_DMA, &nv50_disp_ovly_ofuncs.base },
 	{ G82_DISP_OVERLAY, &nv50_disp_oimm_ofuncs.base },
 	{ G82_DISP_CURSOR, &nv50_disp_curs_ofuncs.base },
@@ -89,8 +89,8 @@ nva0_disp_sclass[] = {
 };
 
 static struct nouveau_oclass
-nva0_disp_base_oclass[] = {
-	{ GT200_DISP, &nv50_disp_base_ofuncs },
+nva0_disp_main_oclass[] = {
+	{ GT200_DISP, &nv50_disp_main_ofuncs },
 	{}
 };
 
@@ -112,7 +112,11 @@ nva0_disp_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	if (ret)
 		return ret;
 
-	nv_engine(priv)->sclass = nva0_disp_base_oclass;
+	ret = nvkm_event_init(&nv50_disp_chan_uevent, 1, 9, &priv->uevent);
+	if (ret)
+		return ret;
+
+	nv_engine(priv)->sclass = nva0_disp_main_oclass;
 	nv_engine(priv)->cclass = &nv50_disp_cclass;
 	nv_subdev(priv)->intr = nv50_disp_intr;
 	INIT_WORK(&priv->supervisor, nv50_disp_intr_supervisor);
@@ -140,9 +144,9 @@ nva0_disp_oclass = &(struct nv50_disp_impl) {
 	},
 	.base.vblank = &nv50_disp_vblank_func,
 	.base.outp =  nv50_disp_outp_sclass,
-	.mthd.core = &nv84_disp_mast_mthd_chan,
-	.mthd.base = &nv84_disp_sync_mthd_chan,
+	.mthd.core = &nv84_disp_core_mthd_chan,
+	.mthd.base = &nv84_disp_base_mthd_chan,
 	.mthd.ovly = &nva0_disp_ovly_mthd_chan,
 	.mthd.prev = 0x000004,
-	.head.scanoutpos = nv50_disp_base_scanoutpos,
+	.head.scanoutpos = nv50_disp_main_scanoutpos,
 }.base.base;
diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/nva3.c b/drivers/gpu/drm/nouveau/core/engine/disp/nva3.c
index 6bd39448f8da..951d79f9b781 100644
--- a/drivers/gpu/drm/nouveau/core/engine/disp/nva3.c
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/nva3.c
@@ -35,8 +35,8 @@
 
 static struct nouveau_oclass
 nva3_disp_sclass[] = {
-	{ GT214_DISP_CORE_CHANNEL_DMA, &nv50_disp_mast_ofuncs.base },
-	{ GT214_DISP_BASE_CHANNEL_DMA, &nv50_disp_sync_ofuncs.base },
+	{ GT214_DISP_CORE_CHANNEL_DMA, &nv50_disp_core_ofuncs.base },
+	{ GT214_DISP_BASE_CHANNEL_DMA, &nv50_disp_base_ofuncs.base },
 	{ GT214_DISP_OVERLAY_CHANNEL_DMA, &nv50_disp_ovly_ofuncs.base },
 	{ GT214_DISP_OVERLAY, &nv50_disp_oimm_ofuncs.base },
 	{ GT214_DISP_CURSOR, &nv50_disp_curs_ofuncs.base },
@@ -44,8 +44,8 @@ nva3_disp_sclass[] = {
 };
 
 static struct nouveau_oclass
-nva3_disp_base_oclass[] = {
-	{ GT214_DISP, &nv50_disp_base_ofuncs },
+nva3_disp_main_oclass[] = {
+	{ GT214_DISP, &nv50_disp_main_ofuncs },
 	{}
 };
 
@@ -67,7 +67,11 @@ nva3_disp_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	if (ret)
 		return ret;
 
-	nv_engine(priv)->sclass = nva3_disp_base_oclass;
+	ret = nvkm_event_init(&nv50_disp_chan_uevent, 1, 9, &priv->uevent);
+	if (ret)
+		return ret;
+
+	nv_engine(priv)->sclass = nva3_disp_main_oclass;
 	nv_engine(priv)->cclass = &nv50_disp_cclass;
 	nv_subdev(priv)->intr = nv50_disp_intr;
 	INIT_WORK(&priv->supervisor, nv50_disp_intr_supervisor);
@@ -96,9 +100,9 @@ nva3_disp_oclass = &(struct nv50_disp_impl) {
 	},
 	.base.vblank = &nv50_disp_vblank_func,
 	.base.outp =  nv94_disp_outp_sclass,
-	.mthd.core = &nv94_disp_mast_mthd_chan,
-	.mthd.base = &nv84_disp_sync_mthd_chan,
+	.mthd.core = &nv94_disp_core_mthd_chan,
+	.mthd.base = &nv84_disp_base_mthd_chan,
 	.mthd.ovly = &nv84_disp_ovly_mthd_chan,
 	.mthd.prev = 0x000004,
-	.head.scanoutpos = nv50_disp_base_scanoutpos,
+	.head.scanoutpos = nv50_disp_main_scanoutpos,
 }.base.base;
diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/nvd0.c b/drivers/gpu/drm/nouveau/core/engine/disp/nvd0.c
index a4bb3c774ee1..181a2d57e356 100644
--- a/drivers/gpu/drm/nouveau/core/engine/disp/nvd0.c
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/nvd0.c
@@ -43,6 +43,33 @@
 #include "nv50.h"
 
 /*******************************************************************************
+ * EVO channel base class
+ ******************************************************************************/
+
+static void
+nvd0_disp_chan_uevent_fini(struct nvkm_event *event, int type, int index)
+{
+	struct nv50_disp_priv *priv = container_of(event, typeof(*priv), uevent);
+	nv_mask(priv, 0x610090, 0x00000001 << index, 0x00000000 << index);
+	nv_wr32(priv, 0x61008c, 0x00000001 << index);
+}
+
+static void
+nvd0_disp_chan_uevent_init(struct nvkm_event *event, int types, int index)
+{
+	struct nv50_disp_priv *priv = container_of(event, typeof(*priv), uevent);
+	nv_wr32(priv, 0x61008c, 0x00000001 << index);
+	nv_mask(priv, 0x610090, 0x00000001 << index, 0x00000001 << index);
+}
+
+const struct nvkm_event_func
+nvd0_disp_chan_uevent = {
+	.ctor = nv50_disp_chan_uevent_ctor,
+	.init = nvd0_disp_chan_uevent_init,
+	.fini = nvd0_disp_chan_uevent_fini,
+};
+
+/*******************************************************************************
  * EVO DMA channel base class
  ******************************************************************************/
 
@@ -77,7 +104,6 @@ nvd0_disp_dmac_init(struct nouveau_object *object)
 		return ret;
 
 	/* enable error reporting */
-	nv_mask(priv, 0x610090, 0x00000001 << chid, 0x00000001 << chid);
 	nv_mask(priv, 0x6100a0, 0x00000001 << chid, 0x00000001 << chid);
 
 	/* initialise channel for dma command submission */
@@ -115,7 +141,7 @@ nvd0_disp_dmac_fini(struct nouveau_object *object, bool suspend)
 			return -EBUSY;
 	}
 
-	/* disable error reporting */
+	/* disable error reporting and completion notification */
 	nv_mask(priv, 0x610090, 0x00000001 << chid, 0x00000000);
 	nv_mask(priv, 0x6100a0, 0x00000001 << chid, 0x00000000);
 
@@ -127,7 +153,7 @@ nvd0_disp_dmac_fini(struct nouveau_object *object, bool suspend)
  ******************************************************************************/
 
 const struct nv50_disp_mthd_list
-nvd0_disp_mast_mthd_base = {
+nvd0_disp_core_mthd_base = {
 	.mthd = 0x0000,
 	.addr = 0x000000,
 	.data = {
@@ -140,7 +166,7 @@ nvd0_disp_mast_mthd_base = {
 };
 
 const struct nv50_disp_mthd_list
-nvd0_disp_mast_mthd_dac = {
+nvd0_disp_core_mthd_dac = {
 	.mthd = 0x0020,
 	.addr = 0x000020,
 	.data = {
@@ -153,7 +179,7 @@ nvd0_disp_mast_mthd_dac = {
 };
 
 const struct nv50_disp_mthd_list
-nvd0_disp_mast_mthd_sor = {
+nvd0_disp_core_mthd_sor = {
 	.mthd = 0x0020,
 	.addr = 0x000020,
 	.data = {
@@ -166,7 +192,7 @@ nvd0_disp_mast_mthd_sor = {
 };
 
 const struct nv50_disp_mthd_list
-nvd0_disp_mast_mthd_pior = {
+nvd0_disp_core_mthd_pior = {
 	.mthd = 0x0020,
 	.addr = 0x000020,
 	.data = {
@@ -179,7 +205,7 @@ nvd0_disp_mast_mthd_pior = {
 };
 
 static const struct nv50_disp_mthd_list
-nvd0_disp_mast_mthd_head = {
+nvd0_disp_core_mthd_head = {
 	.mthd = 0x0300,
 	.addr = 0x000300,
 	.data = {
@@ -253,21 +279,21 @@ nvd0_disp_mast_mthd_head = {
 };
 
 static const struct nv50_disp_mthd_chan
-nvd0_disp_mast_mthd_chan = {
+nvd0_disp_core_mthd_chan = {
 	.name = "Core",
 	.addr = 0x000000,
 	.data = {
-		{ "Global", 1, &nvd0_disp_mast_mthd_base },
-		{    "DAC", 3, &nvd0_disp_mast_mthd_dac  },
-		{    "SOR", 8, &nvd0_disp_mast_mthd_sor  },
-		{   "PIOR", 4, &nvd0_disp_mast_mthd_pior },
-		{   "HEAD", 4, &nvd0_disp_mast_mthd_head },
+		{ "Global", 1, &nvd0_disp_core_mthd_base },
+		{    "DAC", 3, &nvd0_disp_core_mthd_dac  },
+		{    "SOR", 8, &nvd0_disp_core_mthd_sor  },
+		{   "PIOR", 4, &nvd0_disp_core_mthd_pior },
+		{   "HEAD", 4, &nvd0_disp_core_mthd_head },
 		{}
 	}
 };
 
 static int
-nvd0_disp_mast_init(struct nouveau_object *object)
+nvd0_disp_core_init(struct nouveau_object *object)
 {
 	struct nv50_disp_priv *priv = (void *)object->engine;
 	struct nv50_disp_dmac *mast = (void *)object;
@@ -278,7 +304,6 @@ nvd0_disp_mast_init(struct nouveau_object *object)
 		return ret;
 
 	/* enable error reporting */
-	nv_mask(priv, 0x610090, 0x00000001, 0x00000001);
 	nv_mask(priv, 0x6100a0, 0x00000001, 0x00000001);
 
 	/* initialise channel for dma command submission */
@@ -299,7 +324,7 @@ nvd0_disp_mast_init(struct nouveau_object *object)
 }
 
 static int
-nvd0_disp_mast_fini(struct nouveau_object *object, bool suspend)
+nvd0_disp_core_fini(struct nouveau_object *object, bool suspend)
 {
 	struct nv50_disp_priv *priv = (void *)object->engine;
 	struct nv50_disp_dmac *mast = (void *)object;
@@ -313,7 +338,7 @@ nvd0_disp_mast_fini(struct nouveau_object *object, bool suspend)
 			return -EBUSY;
 	}
 
-	/* disable error reporting */
+	/* disable error reporting and completion notification */
 	nv_mask(priv, 0x610090, 0x00000001, 0x00000000);
 	nv_mask(priv, 0x6100a0, 0x00000001, 0x00000000);
 
@@ -321,11 +346,12 @@ nvd0_disp_mast_fini(struct nouveau_object *object, bool suspend)
 }
 
 struct nv50_disp_chan_impl
-nvd0_disp_mast_ofuncs = {
-	.base.ctor = nv50_disp_mast_ctor,
+nvd0_disp_core_ofuncs = {
+	.base.ctor = nv50_disp_core_ctor,
 	.base.dtor = nv50_disp_dmac_dtor,
-	.base.init = nvd0_disp_mast_init,
-	.base.fini = nvd0_disp_mast_fini,
+	.base.init = nvd0_disp_core_init,
+	.base.fini = nvd0_disp_core_fini,
+	.base.ntfy = nv50_disp_chan_ntfy,
 	.base.map  = nv50_disp_chan_map,
 	.base.rd32 = nv50_disp_chan_rd32,
 	.base.wr32 = nv50_disp_chan_wr32,
@@ -339,7 +365,7 @@ nvd0_disp_mast_ofuncs = {
  ******************************************************************************/
 
 static const struct nv50_disp_mthd_list
-nvd0_disp_sync_mthd_base = {
+nvd0_disp_base_mthd_base = {
 	.mthd = 0x0000,
 	.addr = 0x000000,
 	.data = {
@@ -389,7 +415,7 @@ nvd0_disp_sync_mthd_base = {
 };
 
 static const struct nv50_disp_mthd_list
-nvd0_disp_sync_mthd_image = {
+nvd0_disp_base_mthd_image = {
 	.mthd = 0x0400,
 	.addr = 0x000400,
 	.data = {
@@ -403,22 +429,23 @@ nvd0_disp_sync_mthd_image = {
 };
 
 const struct nv50_disp_mthd_chan
-nvd0_disp_sync_mthd_chan = {
+nvd0_disp_base_mthd_chan = {
 	.name = "Base",
 	.addr = 0x001000,
 	.data = {
-		{ "Global", 1, &nvd0_disp_sync_mthd_base },
-		{  "Image", 2, &nvd0_disp_sync_mthd_image },
+		{ "Global", 1, &nvd0_disp_base_mthd_base },
+		{  "Image", 2, &nvd0_disp_base_mthd_image },
 		{}
 	}
 };
 
 struct nv50_disp_chan_impl
-nvd0_disp_sync_ofuncs = {
-	.base.ctor = nv50_disp_sync_ctor,
+nvd0_disp_base_ofuncs = {
+	.base.ctor = nv50_disp_base_ctor,
 	.base.dtor = nv50_disp_dmac_dtor,
 	.base.init = nvd0_disp_dmac_init,
 	.base.fini = nvd0_disp_dmac_fini,
+	.base.ntfy = nv50_disp_chan_ntfy,
 	.base.map  = nv50_disp_chan_map,
 	.base.rd32 = nv50_disp_chan_rd32,
 	.base.wr32 = nv50_disp_chan_wr32,
@@ -499,6 +526,7 @@ nvd0_disp_ovly_ofuncs = {
 	.base.dtor = nv50_disp_dmac_dtor,
 	.base.init = nvd0_disp_dmac_init,
 	.base.fini = nvd0_disp_dmac_fini,
+	.base.ntfy = nv50_disp_chan_ntfy,
 	.base.map  = nv50_disp_chan_map,
 	.base.rd32 = nv50_disp_chan_rd32,
 	.base.wr32 = nv50_disp_chan_wr32,
@@ -524,7 +552,6 @@ nvd0_disp_pioc_init(struct nouveau_object *object)
 		return ret;
 
 	/* enable error reporting */
-	nv_mask(priv, 0x610090, 0x00000001 << chid, 0x00000001 << chid);
 	nv_mask(priv, 0x6100a0, 0x00000001 << chid, 0x00000001 << chid);
 
 	/* activate channel */
@@ -553,7 +580,7 @@ nvd0_disp_pioc_fini(struct nouveau_object *object, bool suspend)
 			return -EBUSY;
 	}
 
-	/* disable error reporting */
+	/* disable error reporting and completion notification */
 	nv_mask(priv, 0x610090, 0x00000001 << chid, 0x00000000);
 	nv_mask(priv, 0x6100a0, 0x00000001 << chid, 0x00000000);
 
@@ -570,6 +597,7 @@ nvd0_disp_oimm_ofuncs = {
 	.base.dtor = nv50_disp_pioc_dtor,
 	.base.init = nvd0_disp_pioc_init,
 	.base.fini = nvd0_disp_pioc_fini,
+	.base.ntfy = nv50_disp_chan_ntfy,
 	.base.map  = nv50_disp_chan_map,
 	.base.rd32 = nv50_disp_chan_rd32,
 	.base.wr32 = nv50_disp_chan_wr32,
@@ -586,6 +614,7 @@ nvd0_disp_curs_ofuncs = {
 	.base.dtor = nv50_disp_pioc_dtor,
 	.base.init = nvd0_disp_pioc_init,
 	.base.fini = nvd0_disp_pioc_fini,
+	.base.ntfy = nv50_disp_chan_ntfy,
 	.base.map  = nv50_disp_chan_map,
 	.base.rd32 = nv50_disp_chan_rd32,
 	.base.wr32 = nv50_disp_chan_wr32,
@@ -597,7 +626,7 @@ nvd0_disp_curs_ofuncs = {
  ******************************************************************************/
 
 int
-nvd0_disp_base_scanoutpos(NV50_DISP_MTHD_V0)
+nvd0_disp_main_scanoutpos(NV50_DISP_MTHD_V0)
 {
 	const u32 total  = nv_rd32(priv, 0x640414 + (head * 0x300));
 	const u32 blanke = nv_rd32(priv, 0x64041c + (head * 0x300));
@@ -629,7 +658,7 @@ nvd0_disp_base_scanoutpos(NV50_DISP_MTHD_V0)
 }
 
 static int
-nvd0_disp_base_init(struct nouveau_object *object)
+nvd0_disp_main_init(struct nouveau_object *object)
 {
 	struct nv50_disp_priv *priv = (void *)object->engine;
 	struct nv50_disp_base *base = (void *)object;
@@ -698,7 +727,7 @@ nvd0_disp_base_init(struct nouveau_object *object)
 }
 
 static int
-nvd0_disp_base_fini(struct nouveau_object *object, bool suspend)
+nvd0_disp_main_fini(struct nouveau_object *object, bool suspend)
 {
 	struct nv50_disp_priv *priv = (void *)object->engine;
 	struct nv50_disp_base *base = (void *)object;
@@ -710,25 +739,25 @@ nvd0_disp_base_fini(struct nouveau_object *object, bool suspend)
 }
 
 struct nouveau_ofuncs
-nvd0_disp_base_ofuncs = {
-	.ctor = nv50_disp_base_ctor,
-	.dtor = nv50_disp_base_dtor,
-	.init = nvd0_disp_base_init,
-	.fini = nvd0_disp_base_fini,
-	.mthd = nv50_disp_base_mthd,
+nvd0_disp_main_ofuncs = {
+	.ctor = nv50_disp_main_ctor,
+	.dtor = nv50_disp_main_dtor,
+	.init = nvd0_disp_main_init,
+	.fini = nvd0_disp_main_fini,
+	.mthd = nv50_disp_main_mthd,
 	.ntfy = nouveau_disp_ntfy,
 };
 
 static struct nouveau_oclass
-nvd0_disp_base_oclass[] = {
-	{ GF110_DISP, &nvd0_disp_base_ofuncs },
+nvd0_disp_main_oclass[] = {
+	{ GF110_DISP, &nvd0_disp_main_ofuncs },
 	{}
 };
 
 static struct nouveau_oclass
 nvd0_disp_sclass[] = {
-	{ GF110_DISP_CORE_CHANNEL_DMA, &nvd0_disp_mast_ofuncs.base },
-	{ GF110_DISP_BASE_CHANNEL_DMA, &nvd0_disp_sync_ofuncs.base },
+	{ GF110_DISP_CORE_CHANNEL_DMA, &nvd0_disp_core_ofuncs.base },
+	{ GF110_DISP_BASE_CHANNEL_DMA, &nvd0_disp_base_ofuncs.base },
 	{ GF110_DISP_OVERLAY_CONTROL_DMA, &nvd0_disp_ovly_ofuncs.base },
 	{ GF110_DISP_OVERLAY, &nvd0_disp_oimm_ofuncs.base },
 	{ GF110_DISP_CURSOR, &nvd0_disp_curs_ofuncs.base },
@@ -949,6 +978,9 @@ nvd0_disp_intr_unk2_2_tu(struct nv50_disp_priv *priv, int head,
 	const int or = ffs(outp->or) - 1;
 	const u32 ctrl = nv_rd32(priv, 0x660200 + (or   * 0x020));
 	const u32 conf = nv_rd32(priv, 0x660404 + (head * 0x300));
+	const s32 vactive = nv_rd32(priv, 0x660414 + (head * 0x300)) & 0xffff;
+	const s32 vblanke = nv_rd32(priv, 0x66041c + (head * 0x300)) & 0xffff;
+	const s32 vblanks = nv_rd32(priv, 0x660420 + (head * 0x300)) & 0xffff;
 	const u32 pclk = nv_rd32(priv, 0x660450 + (head * 0x300)) / 1000;
 	const u32 link = ((ctrl & 0xf00) == 0x800) ? 0 : 1;
 	const u32 hoff = (head * 0x800);
@@ -956,23 +988,35 @@ nvd0_disp_intr_unk2_2_tu(struct nv50_disp_priv *priv, int head,
 	const u32 loff = (link * 0x080) + soff;
 	const u32 symbol = 100000;
 	const u32 TU = 64;
-	u32 dpctrl = nv_rd32(priv, 0x61c10c + loff) & 0x000f0000;
+	u32 dpctrl = nv_rd32(priv, 0x61c10c + loff);
 	u32 clksor = nv_rd32(priv, 0x612300 + soff);
 	u32 datarate, link_nr, link_bw, bits;
 	u64 ratio, value;
 
+	link_nr  = hweight32(dpctrl & 0x000f0000);
+	link_bw  = (clksor & 0x007c0000) >> 18;
+	link_bw *= 27000;
+
+	/* symbols/hblank - algorithm taken from comments in tegra driver */
+	value = vblanke + vactive - vblanks - 7;
+	value = value * link_bw;
+	do_div(value, pclk);
+	value = value - (3 * !!(dpctrl & 0x00004000)) - (12 / link_nr);
+	nv_mask(priv, 0x616620 + hoff, 0x0000ffff, value);
+
+	/* symbols/vblank - algorithm taken from comments in tegra driver */
+	value = vblanks - vblanke - 25;
+	value = value * link_bw;
+	do_div(value, pclk);
+	value = value - ((36 / link_nr) + 3) - 1;
+	nv_mask(priv, 0x616624 + hoff, 0x00ffffff, value);
+
+	/* watermark */
 	if      ((conf & 0x3c0) == 0x180) bits = 30;
 	else if ((conf & 0x3c0) == 0x140) bits = 24;
 	else                              bits = 18;
 	datarate = (pclk * bits) / 8;
 
-	if      (dpctrl > 0x00030000) link_nr = 4;
-	else if (dpctrl > 0x00010000) link_nr = 2;
-	else			      link_nr = 1;
-
-	link_bw  = (clksor & 0x007c0000) >> 18;
-	link_bw *= 27000;
-
 	ratio  = datarate;
 	ratio *= symbol;
 	do_div(ratio, link_nr * link_bw);
@@ -1013,6 +1057,9 @@ nvd0_disp_intr_unk2_2(struct nv50_disp_priv *priv, int head)
 
 		if (nvkm_output_dp_train(outp, pclk, true))
 			ERR("link not trained before attach\n");
+	} else {
+		if (priv->sor.magic)
+			priv->sor.magic(outp);
 	}
 
 	exec_clkcmp(priv, head, 0, pclk, &conf);
@@ -1021,10 +1068,18 @@ nvd0_disp_intr_unk2_2(struct nv50_disp_priv *priv, int head)
 		addr = 0x612280 + (ffs(outp->info.or) - 1) * 0x800;
 		data = 0x00000000;
 	} else {
-		if (outp->info.type == DCB_OUTPUT_DP)
-			nvd0_disp_intr_unk2_2_tu(priv, head, &outp->info);
 		addr = 0x612300 + (ffs(outp->info.or) - 1) * 0x800;
 		data = (conf & 0x0100) ? 0x00000101 : 0x00000000;
+		switch (outp->info.type) {
+		case DCB_OUTPUT_TMDS:
+			nv_mask(priv, addr, 0x007c0000, 0x00280000);
+			break;
+		case DCB_OUTPUT_DP:
+			nvd0_disp_intr_unk2_2_tu(priv, head, &outp->info);
+			break;
+		default:
+			break;
+		}
 	}
 
 	nv_mask(priv, addr, 0x00000707, data);
@@ -1153,7 +1208,11 @@ nvd0_disp_intr(struct nouveau_subdev *subdev)
 
 	if (intr & 0x00000001) {
 		u32 stat = nv_rd32(priv, 0x61008c);
-		nv_wr32(priv, 0x61008c, stat);
+		while (stat) {
+			int chid = __ffs(stat); stat &= ~(1 << chid);
+			nv50_disp_chan_uevent_send(priv, chid);
+			nv_wr32(priv, 0x61008c, 1 << chid);
+		}
 		intr &= ~0x00000001;
 	}
 
@@ -1209,7 +1268,11 @@ nvd0_disp_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	if (ret)
 		return ret;
 
-	nv_engine(priv)->sclass = nvd0_disp_base_oclass;
+	ret = nvkm_event_init(&nvd0_disp_chan_uevent, 1, 17, &priv->uevent);
+	if (ret)
+		return ret;
+
+	nv_engine(priv)->sclass = nvd0_disp_main_oclass;
 	nv_engine(priv)->cclass = &nv50_disp_cclass;
 	nv_subdev(priv)->intr = nvd0_disp_intr;
 	INIT_WORK(&priv->supervisor, nvd0_disp_intr_supervisor);
@@ -1242,9 +1305,9 @@ nvd0_disp_oclass = &(struct nv50_disp_impl) {
 	},
 	.base.vblank = &nvd0_disp_vblank_func,
 	.base.outp =  nvd0_disp_outp_sclass,
-	.mthd.core = &nvd0_disp_mast_mthd_chan,
-	.mthd.base = &nvd0_disp_sync_mthd_chan,
+	.mthd.core = &nvd0_disp_core_mthd_chan,
+	.mthd.base = &nvd0_disp_base_mthd_chan,
 	.mthd.ovly = &nvd0_disp_ovly_mthd_chan,
 	.mthd.prev = -0x020000,
-	.head.scanoutpos = nvd0_disp_base_scanoutpos,
+	.head.scanoutpos = nvd0_disp_main_scanoutpos,
 }.base.base;
diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/nve0.c b/drivers/gpu/drm/nouveau/core/engine/disp/nve0.c
index 47fef1e398c4..55debec7e68f 100644
--- a/drivers/gpu/drm/nouveau/core/engine/disp/nve0.c
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/nve0.c
@@ -34,7 +34,7 @@
  ******************************************************************************/
 
 static const struct nv50_disp_mthd_list
-nve0_disp_mast_mthd_head = {
+nve0_disp_core_mthd_head = {
 	.mthd = 0x0300,
 	.addr = 0x000300,
 	.data = {
@@ -113,15 +113,15 @@ nve0_disp_mast_mthd_head = {
 };
 
 const struct nv50_disp_mthd_chan
-nve0_disp_mast_mthd_chan = {
+nve0_disp_core_mthd_chan = {
 	.name = "Core",
 	.addr = 0x000000,
 	.data = {
-		{ "Global", 1, &nvd0_disp_mast_mthd_base },
-		{    "DAC", 3, &nvd0_disp_mast_mthd_dac  },
-		{    "SOR", 8, &nvd0_disp_mast_mthd_sor  },
-		{   "PIOR", 4, &nvd0_disp_mast_mthd_pior },
-		{   "HEAD", 4, &nve0_disp_mast_mthd_head },
+		{ "Global", 1, &nvd0_disp_core_mthd_base },
+		{    "DAC", 3, &nvd0_disp_core_mthd_dac  },
+		{    "SOR", 8, &nvd0_disp_core_mthd_sor  },
+		{   "PIOR", 4, &nvd0_disp_core_mthd_pior },
+		{   "HEAD", 4, &nve0_disp_core_mthd_head },
 		{}
 	}
 };
@@ -200,8 +200,8 @@ nve0_disp_ovly_mthd_chan = {
 
 static struct nouveau_oclass
 nve0_disp_sclass[] = {
-	{ GK104_DISP_CORE_CHANNEL_DMA, &nvd0_disp_mast_ofuncs.base },
-	{ GK104_DISP_BASE_CHANNEL_DMA, &nvd0_disp_sync_ofuncs.base },
+	{ GK104_DISP_CORE_CHANNEL_DMA, &nvd0_disp_core_ofuncs.base },
+	{ GK104_DISP_BASE_CHANNEL_DMA, &nvd0_disp_base_ofuncs.base },
 	{ GK104_DISP_OVERLAY_CONTROL_DMA, &nvd0_disp_ovly_ofuncs.base },
 	{ GK104_DISP_OVERLAY, &nvd0_disp_oimm_ofuncs.base },
 	{ GK104_DISP_CURSOR, &nvd0_disp_curs_ofuncs.base },
@@ -209,8 +209,8 @@ nve0_disp_sclass[] = {
 };
 
 static struct nouveau_oclass
-nve0_disp_base_oclass[] = {
-	{ GK104_DISP, &nvd0_disp_base_ofuncs },
+nve0_disp_main_oclass[] = {
+	{ GK104_DISP, &nvd0_disp_main_ofuncs },
 	{}
 };
 
@@ -233,7 +233,11 @@ nve0_disp_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	if (ret)
 		return ret;
 
-	nv_engine(priv)->sclass = nve0_disp_base_oclass;
+	ret = nvkm_event_init(&nvd0_disp_chan_uevent, 1, 17, &priv->uevent);
+	if (ret)
+		return ret;
+
+	nv_engine(priv)->sclass = nve0_disp_main_oclass;
 	nv_engine(priv)->cclass = &nv50_disp_cclass;
 	nv_subdev(priv)->intr = nvd0_disp_intr;
 	INIT_WORK(&priv->supervisor, nvd0_disp_intr_supervisor);
@@ -245,7 +249,7 @@ nve0_disp_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	priv->dac.sense = nv50_dac_sense;
 	priv->sor.power = nv50_sor_power;
 	priv->sor.hda_eld = nvd0_hda_eld;
-	priv->sor.hdmi = nvd0_hdmi_ctrl;
+	priv->sor.hdmi = nve0_hdmi_ctrl;
 	return 0;
 }
 
@@ -260,9 +264,9 @@ nve0_disp_oclass = &(struct nv50_disp_impl) {
 	},
 	.base.vblank = &nvd0_disp_vblank_func,
 	.base.outp =  nvd0_disp_outp_sclass,
-	.mthd.core = &nve0_disp_mast_mthd_chan,
-	.mthd.base = &nvd0_disp_sync_mthd_chan,
+	.mthd.core = &nve0_disp_core_mthd_chan,
+	.mthd.base = &nvd0_disp_base_mthd_chan,
 	.mthd.ovly = &nve0_disp_ovly_mthd_chan,
 	.mthd.prev = -0x020000,
-	.head.scanoutpos = nvd0_disp_base_scanoutpos,
+	.head.scanoutpos = nvd0_disp_main_scanoutpos,
 }.base.base;
diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/nvf0.c b/drivers/gpu/drm/nouveau/core/engine/disp/nvf0.c
index 04bda4ac4ed3..3e7e2d28744c 100644
--- a/drivers/gpu/drm/nouveau/core/engine/disp/nvf0.c
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/nvf0.c
@@ -35,8 +35,8 @@
 
 static struct nouveau_oclass
 nvf0_disp_sclass[] = {
-	{ GK110_DISP_CORE_CHANNEL_DMA, &nvd0_disp_mast_ofuncs.base },
-	{ GK110_DISP_BASE_CHANNEL_DMA, &nvd0_disp_sync_ofuncs.base },
+	{ GK110_DISP_CORE_CHANNEL_DMA, &nvd0_disp_core_ofuncs.base },
+	{ GK110_DISP_BASE_CHANNEL_DMA, &nvd0_disp_base_ofuncs.base },
 	{ GK104_DISP_OVERLAY_CONTROL_DMA, &nvd0_disp_ovly_ofuncs.base },
 	{ GK104_DISP_OVERLAY, &nvd0_disp_oimm_ofuncs.base },
 	{ GK104_DISP_CURSOR, &nvd0_disp_curs_ofuncs.base },
@@ -44,8 +44,8 @@ nvf0_disp_sclass[] = {
 };
 
 static struct nouveau_oclass
-nvf0_disp_base_oclass[] = {
-	{ GK110_DISP, &nvd0_disp_base_ofuncs },
+nvf0_disp_main_oclass[] = {
+	{ GK110_DISP, &nvd0_disp_main_ofuncs },
 	{}
 };
 
@@ -68,7 +68,11 @@ nvf0_disp_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	if (ret)
 		return ret;
 
-	nv_engine(priv)->sclass = nvf0_disp_base_oclass;
+	ret = nvkm_event_init(&nvd0_disp_chan_uevent, 1, 17, &priv->uevent);
+	if (ret)
+		return ret;
+
+	nv_engine(priv)->sclass = nvf0_disp_main_oclass;
 	nv_engine(priv)->cclass = &nv50_disp_cclass;
 	nv_subdev(priv)->intr = nvd0_disp_intr;
 	INIT_WORK(&priv->supervisor, nvd0_disp_intr_supervisor);
@@ -80,7 +84,7 @@ nvf0_disp_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	priv->dac.sense = nv50_dac_sense;
 	priv->sor.power = nv50_sor_power;
 	priv->sor.hda_eld = nvd0_hda_eld;
-	priv->sor.hdmi = nvd0_hdmi_ctrl;
+	priv->sor.hdmi = nve0_hdmi_ctrl;
 	return 0;
 }
 
@@ -95,9 +99,9 @@ nvf0_disp_oclass = &(struct nv50_disp_impl) {
 	},
 	.base.vblank = &nvd0_disp_vblank_func,
 	.base.outp =  nvd0_disp_outp_sclass,
-	.mthd.core = &nve0_disp_mast_mthd_chan,
-	.mthd.base = &nvd0_disp_sync_mthd_chan,
+	.mthd.core = &nve0_disp_core_mthd_chan,
+	.mthd.base = &nvd0_disp_base_mthd_chan,
 	.mthd.ovly = &nve0_disp_ovly_mthd_chan,
 	.mthd.prev = -0x020000,
-	.head.scanoutpos = nvd0_disp_base_scanoutpos,
+	.head.scanoutpos = nvd0_disp_main_scanoutpos,
 }.base.base;
diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/outp.c b/drivers/gpu/drm/nouveau/core/engine/disp/outp.c
index a5ff00a9cedc..bbd9b6fdc90f 100644
--- a/drivers/gpu/drm/nouveau/core/engine/disp/outp.c
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/outp.c
@@ -85,7 +85,10 @@ nvkm_output_create_(struct nouveau_object *parent,
 	    dcbE->sorconf.link : 0, dcbE->connector, dcbE->i2c_index,
 	    dcbE->bus, dcbE->heads);
 
-	outp->port = i2c->find(i2c, outp->info.i2c_index);
+	if (outp->info.type != DCB_OUTPUT_DP)
+		outp->port = i2c->find(i2c, NV_I2C_PORT(outp->info.i2c_index));
+	else
+		outp->port = i2c->find(i2c, NV_I2C_AUX(outp->info.i2c_index));
 	outp->edid = outp->port;
 
 	data = nvbios_connEp(bios, outp->info.connector, &ver, &hdr, &connE);
diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/outpdp.c b/drivers/gpu/drm/nouveau/core/engine/disp/outpdp.c
index 6f6e2a898270..667a9070e006 100644
--- a/drivers/gpu/drm/nouveau/core/engine/disp/outpdp.c
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/outpdp.c
@@ -254,7 +254,7 @@ nvkm_output_dp_create_(struct nouveau_object *parent,
 	atomic_set(&outp->lt.done, 0);
 
 	/* link maintenance */
-	ret = nvkm_notify_init(&i2c->event, nvkm_output_dp_irq, true,
+	ret = nvkm_notify_init(NULL, &i2c->event, nvkm_output_dp_irq, true,
 			       &(struct nvkm_i2c_ntfy_req) {
 				.mask = NVKM_I2C_IRQ,
 				.port = outp->base.edid->index,
@@ -268,7 +268,7 @@ nvkm_output_dp_create_(struct nouveau_object *parent,
 	}
 
 	/* hotplug detect, replaces gpio-based mechanism with aux events */
-	ret = nvkm_notify_init(&i2c->event, nvkm_output_dp_hpd, true,
+	ret = nvkm_notify_init(NULL, &i2c->event, nvkm_output_dp_hpd, true,
 			       &(struct nvkm_i2c_ntfy_req) {
 				.mask = NVKM_I2C_PLUG | NVKM_I2C_UNPLUG,
 				.port = outp->base.edid->index,
diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/priv.h b/drivers/gpu/drm/nouveau/core/engine/disp/priv.h
index dbd43ae9df81..6a0511d54ce6 100644
--- a/drivers/gpu/drm/nouveau/core/engine/disp/priv.h
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/priv.h
@@ -40,7 +40,8 @@ int  _nouveau_disp_fini(struct nouveau_object *, bool);
 extern struct nouveau_oclass *nvkm_output_oclass;
 extern struct nouveau_oclass *nvkm_connector_oclass;
 
-int  nouveau_disp_vblank_ctor(void *data, u32 size, struct nvkm_notify *);
+int  nouveau_disp_vblank_ctor(struct nouveau_object *, void *data, u32 size,
+			      struct nvkm_notify *);
 void nouveau_disp_vblank(struct nouveau_disp *, int head);
 int  nouveau_disp_ntfy(struct nouveau_object *, u32, struct nvkm_event **);
 
diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/sorgm204.c b/drivers/gpu/drm/nouveau/core/engine/disp/sorgm204.c
new file mode 100644
index 000000000000..0b4fad39e9a6
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/sorgm204.c
@@ -0,0 +1,144 @@
+/*
+ * Copyright 2012 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs
+ */
+
+#include <core/os.h>
+
+#include <subdev/bios.h>
+#include <subdev/bios/dcb.h>
+#include <subdev/bios/dp.h>
+#include <subdev/bios/init.h>
+#include <subdev/timer.h>
+
+#include "nv50.h"
+
+static inline u32
+gm204_sor_soff(struct nvkm_output_dp *outp)
+{
+	return (ffs(outp->base.info.or) - 1) * 0x800;
+}
+
+static inline u32
+gm204_sor_loff(struct nvkm_output_dp *outp)
+{
+	return gm204_sor_soff(outp) + !(outp->base.info.sorconf.link & 1) * 0x80;
+}
+
+void
+gm204_sor_magic(struct nvkm_output *outp)
+{
+	struct nv50_disp_priv *priv = (void *)nouveau_disp(outp);
+	const u32 soff = outp->or * 0x100;
+	const u32 data = outp->or + 1;
+	if (outp->info.sorconf.link & 1)
+		nv_mask(priv, 0x612308 + soff, 0x0000001f, 0x00000000 | data);
+	if (outp->info.sorconf.link & 2)
+		nv_mask(priv, 0x612388 + soff, 0x0000001f, 0x00000010 | data);
+}
+
+static inline u32
+gm204_sor_dp_lane_map(struct nv50_disp_priv *priv, u8 lane)
+{
+	return lane * 0x08;
+}
+
+static int
+gm204_sor_dp_pattern(struct nvkm_output_dp *outp, int pattern)
+{
+	struct nv50_disp_priv *priv = (void *)nouveau_disp(outp);
+	const u32 soff = gm204_sor_soff(outp);
+	const u32 data = 0x01010101 * pattern;
+	if (outp->base.info.sorconf.link & 1)
+		nv_mask(priv, 0x61c110 + soff, 0x0f0f0f0f, data);
+	else
+		nv_mask(priv, 0x61c12c + soff, 0x0f0f0f0f, data);
+	return 0;
+}
+
+static int
+gm204_sor_dp_lnk_pwr(struct nvkm_output_dp *outp, int nr)
+{
+	struct nv50_disp_priv *priv = (void *)nouveau_disp(outp);
+	const u32 soff = gm204_sor_soff(outp);
+	const u32 loff = gm204_sor_loff(outp);
+	u32 mask = 0, i;
+
+	for (i = 0; i < nr; i++)
+		mask |= 1 << (gm204_sor_dp_lane_map(priv, i) >> 3);
+
+	nv_mask(priv, 0x61c130 + loff, 0x0000000f, mask);
+	nv_mask(priv, 0x61c034 + soff, 0x80000000, 0x80000000);
+	nv_wait(priv, 0x61c034 + soff, 0x80000000, 0x00000000);
+	return 0;
+}
+
+static int
+gm204_sor_dp_drv_ctl(struct nvkm_output_dp *outp, int ln, int vs, int pe, int pc)
+{
+	struct nv50_disp_priv *priv = (void *)nouveau_disp(outp);
+	struct nouveau_bios *bios = nouveau_bios(priv);
+	const u32 shift = gm204_sor_dp_lane_map(priv, ln);
+	const u32 loff = gm204_sor_loff(outp);
+	u32 addr, data[4];
+	u8  ver, hdr, cnt, len;
+	struct nvbios_dpout info;
+	struct nvbios_dpcfg ocfg;
+
+	addr = nvbios_dpout_match(bios, outp->base.info.hasht,
+					outp->base.info.hashm,
+				 &ver, &hdr, &cnt, &len, &info);
+	if (!addr)
+		return -ENODEV;
+
+	addr = nvbios_dpcfg_match(bios, addr, pc, vs, pe,
+				 &ver, &hdr, &cnt, &len, &ocfg);
+	if (!addr)
+		return -EINVAL;
+
+	data[0] = nv_rd32(priv, 0x61c118 + loff) & ~(0x000000ff << shift);
+	data[1] = nv_rd32(priv, 0x61c120 + loff) & ~(0x000000ff << shift);
+	data[2] = nv_rd32(priv, 0x61c130 + loff);
+	if ((data[2] & 0x0000ff00) < (ocfg.tx_pu << 8) || ln == 0)
+		data[2] = (data[2] & ~0x0000ff00) | (ocfg.tx_pu << 8);
+	nv_wr32(priv, 0x61c118 + loff, data[0] | (ocfg.dc << shift));
+	nv_wr32(priv, 0x61c120 + loff, data[1] | (ocfg.pe << shift));
+	nv_wr32(priv, 0x61c130 + loff, data[2] | (ocfg.tx_pu << 8));
+	data[3] = nv_rd32(priv, 0x61c13c + loff) & ~(0x000000ff << shift);
+	nv_wr32(priv, 0x61c13c + loff, data[3] | (ocfg.pc << shift));
+	return 0;
+}
+
+struct nvkm_output_dp_impl
+gm204_sor_dp_impl = {
+	.base.base.handle = DCB_OUTPUT_DP,
+	.base.base.ofuncs = &(struct nouveau_ofuncs) {
+		.ctor = _nvkm_output_dp_ctor,
+		.dtor = _nvkm_output_dp_dtor,
+		.init = _nvkm_output_dp_init,
+		.fini = _nvkm_output_dp_fini,
+	},
+	.pattern = gm204_sor_dp_pattern,
+	.lnk_pwr = gm204_sor_dp_lnk_pwr,
+	.lnk_ctl = nvd0_sor_dp_lnk_ctl,
+	.drv_ctl = gm204_sor_dp_drv_ctl,
+};
diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/sornvd0.c b/drivers/gpu/drm/nouveau/core/engine/disp/sornvd0.c
index 7b7bbc3e459e..fdab2939070c 100644
--- a/drivers/gpu/drm/nouveau/core/engine/disp/sornvd0.c
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/sornvd0.c
@@ -60,7 +60,7 @@ nvd0_sor_dp_pattern(struct nvkm_output_dp *outp, int pattern)
 	return 0;
 }
 
-static int
+int
 nvd0_sor_dp_lnk_ctl(struct nvkm_output_dp *outp, int nr, int bw, bool ef)
 {
 	struct nv50_disp_priv *priv = (void *)nouveau_disp(outp);
diff --git a/drivers/gpu/drm/nouveau/core/engine/dmaobj/nvd0.c b/drivers/gpu/drm/nouveau/core/engine/dmaobj/nvd0.c
index 3fc4f0b0eaca..19f5f6522962 100644
--- a/drivers/gpu/drm/nouveau/core/engine/dmaobj/nvd0.c
+++ b/drivers/gpu/drm/nouveau/core/engine/dmaobj/nvd0.c
@@ -51,6 +51,7 @@ nvd0_dmaobj_bind(struct nouveau_dmaobj *dmaobj,
 		case GK104_DISP_CORE_CHANNEL_DMA:
 		case GK110_DISP_CORE_CHANNEL_DMA:
 		case GM107_DISP_CORE_CHANNEL_DMA:
+		case GM204_DISP_CORE_CHANNEL_DMA:
 		case GF110_DISP_BASE_CHANNEL_DMA:
 		case GK104_DISP_BASE_CHANNEL_DMA:
 		case GK110_DISP_BASE_CHANNEL_DMA:
diff --git a/drivers/gpu/drm/nouveau/core/engine/fifo/base.c b/drivers/gpu/drm/nouveau/core/engine/fifo/base.c
index 0f999fc45ab9..ac8375cf4eef 100644
--- a/drivers/gpu/drm/nouveau/core/engine/fifo/base.c
+++ b/drivers/gpu/drm/nouveau/core/engine/fifo/base.c
@@ -34,7 +34,8 @@
 #include <engine/fifo.h>
 
 static int
-nouveau_fifo_event_ctor(void *data, u32 size, struct nvkm_notify *notify)
+nouveau_fifo_event_ctor(struct nouveau_object *object, void *data, u32 size,
+			struct nvkm_notify *notify)
 {
 	if (size == 0) {
 		notify->size  = 0;
@@ -170,7 +171,8 @@ _nouveau_fifo_channel_wr32(struct nouveau_object *object, u64 addr, u32 data)
 }
 
 int
-nouveau_fifo_uevent_ctor(void *data, u32 size, struct nvkm_notify *notify)
+nouveau_fifo_uevent_ctor(struct nouveau_object *object, void *data, u32 size,
+			 struct nvkm_notify *notify)
 {
 	union {
 		struct nvif_notify_uevent_req none;
diff --git a/drivers/gpu/drm/nouveau/core/engine/fifo/nv04.c b/drivers/gpu/drm/nouveau/core/engine/fifo/nv04.c
index 5ae6a43893b5..1931057f9962 100644
--- a/drivers/gpu/drm/nouveau/core/engine/fifo/nv04.c
+++ b/drivers/gpu/drm/nouveau/core/engine/fifo/nv04.c
@@ -551,8 +551,8 @@ nv04_fifo_intr(struct nouveau_subdev *subdev)
 			}
 
 			if (status & 0x40000000) {
-				nouveau_fifo_uevent(&priv->base);
 				nv_wr32(priv, 0x002100, 0x40000000);
+				nouveau_fifo_uevent(&priv->base);
 				status &= ~0x40000000;
 			}
 		}
diff --git a/drivers/gpu/drm/nouveau/core/engine/fifo/nvc0.c b/drivers/gpu/drm/nouveau/core/engine/fifo/nvc0.c
index 1fe1f8fbda0c..074d434c3077 100644
--- a/drivers/gpu/drm/nouveau/core/engine/fifo/nvc0.c
+++ b/drivers/gpu/drm/nouveau/core/engine/fifo/nvc0.c
@@ -740,6 +740,8 @@ nvc0_fifo_intr_engine_unit(struct nvc0_fifo_priv *priv, int engn)
 	u32 inte = nv_rd32(priv, 0x002628);
 	u32 unkn;
 
+	nv_wr32(priv, 0x0025a8 + (engn * 0x04), intr);
+
 	for (unkn = 0; unkn < 8; unkn++) {
 		u32 ints = (intr >> (unkn * 0x04)) & inte;
 		if (ints & 0x1) {
@@ -751,8 +753,6 @@ nvc0_fifo_intr_engine_unit(struct nvc0_fifo_priv *priv, int engn)
 			nv_mask(priv, 0x002628, ints, 0);
 		}
 	}
-
-	nv_wr32(priv, 0x0025a8 + (engn * 0x04), intr);
 }
 
 static void
diff --git a/drivers/gpu/drm/nouveau/core/engine/fifo/nve0.c b/drivers/gpu/drm/nouveau/core/engine/fifo/nve0.c
index d2f0fd39c145..6a8db7c80bd1 100644
--- a/drivers/gpu/drm/nouveau/core/engine/fifo/nve0.c
+++ b/drivers/gpu/drm/nouveau/core/engine/fifo/nve0.c
@@ -792,7 +792,7 @@ nve0_fifo_intr_fault(struct nve0_fifo_priv *priv, int unit)
 	nouveau_engctx_put(engctx);
 }
 
-static const struct nouveau_bitfield nve0_fifo_pbdma_intr[] = {
+static const struct nouveau_bitfield nve0_fifo_pbdma_intr_0[] = {
 	{ 0x00000001, "MEMREQ" },
 	{ 0x00000002, "MEMACK_TIMEOUT" },
 	{ 0x00000004, "MEMACK_EXTRA" },
@@ -827,9 +827,10 @@ static const struct nouveau_bitfield nve0_fifo_pbdma_intr[] = {
 };
 
 static void
-nve0_fifo_intr_pbdma(struct nve0_fifo_priv *priv, int unit)
+nve0_fifo_intr_pbdma_0(struct nve0_fifo_priv *priv, int unit)
 {
-	u32 stat = nv_rd32(priv, 0x040108 + (unit * 0x2000));
+	u32 mask = nv_rd32(priv, 0x04010c + (unit * 0x2000));
+	u32 stat = nv_rd32(priv, 0x040108 + (unit * 0x2000)) & mask;
 	u32 addr = nv_rd32(priv, 0x0400c0 + (unit * 0x2000));
 	u32 data = nv_rd32(priv, 0x0400c4 + (unit * 0x2000));
 	u32 chid = nv_rd32(priv, 0x040120 + (unit * 0x2000)) & 0xfff;
@@ -840,11 +841,12 @@ nve0_fifo_intr_pbdma(struct nve0_fifo_priv *priv, int unit)
 	if (stat & 0x00800000) {
 		if (!nve0_fifo_swmthd(priv, chid, mthd, data))
 			show &= ~0x00800000;
+		nv_wr32(priv, 0x0400c0 + (unit * 0x2000), 0x80600008);
 	}
 
 	if (show) {
 		nv_error(priv, "PBDMA%d:", unit);
-		nouveau_bitfield_print(nve0_fifo_pbdma_intr, show);
+		nouveau_bitfield_print(nve0_fifo_pbdma_intr_0, show);
 		pr_cont("\n");
 		nv_error(priv,
 			 "PBDMA%d: ch %d [%s] subc %d mthd 0x%04x data 0x%08x\n",
@@ -853,10 +855,37 @@ nve0_fifo_intr_pbdma(struct nve0_fifo_priv *priv, int unit)
 			 subc, mthd, data);
 	}
 
-	nv_wr32(priv, 0x0400c0 + (unit * 0x2000), 0x80600008);
 	nv_wr32(priv, 0x040108 + (unit * 0x2000), stat);
 }
 
+static const struct nouveau_bitfield nve0_fifo_pbdma_intr_1[] = {
+	{ 0x00000001, "HCE_RE_ILLEGAL_OP" },
+	{ 0x00000002, "HCE_RE_ALIGNB" },
+	{ 0x00000004, "HCE_PRIV" },
+	{ 0x00000008, "HCE_ILLEGAL_MTHD" },
+	{ 0x00000010, "HCE_ILLEGAL_CLASS" },
+	{}
+};
+
+static void
+nve0_fifo_intr_pbdma_1(struct nve0_fifo_priv *priv, int unit)
+{
+	u32 mask = nv_rd32(priv, 0x04014c + (unit * 0x2000));
+	u32 stat = nv_rd32(priv, 0x040148 + (unit * 0x2000)) & mask;
+	u32 chid = nv_rd32(priv, 0x040120 + (unit * 0x2000)) & 0xfff;
+
+	if (stat) {
+		nv_error(priv, "PBDMA%d:", unit);
+		nouveau_bitfield_print(nve0_fifo_pbdma_intr_1, stat);
+		pr_cont("\n");
+		nv_error(priv, "PBDMA%d: ch %d %08x %08x\n", unit, chid,
+			 nv_rd32(priv, 0x040150 + (unit * 0x2000)),
+			 nv_rd32(priv, 0x040154 + (unit * 0x2000)));
+	}
+
+	nv_wr32(priv, 0x040148 + (unit * 0x2000), stat);
+}
+
 static void
 nve0_fifo_intr_runlist(struct nve0_fifo_priv *priv)
 {
@@ -939,7 +968,8 @@ nve0_fifo_intr(struct nouveau_subdev *subdev)
 		u32 mask = nv_rd32(priv, 0x0025a0);
 		while (mask) {
 			u32 unit = __ffs(mask);
-			nve0_fifo_intr_pbdma(priv, unit);
+			nve0_fifo_intr_pbdma_0(priv, unit);
+			nve0_fifo_intr_pbdma_1(priv, unit);
 			nv_wr32(priv, 0x0025a0, (1 << unit));
 			mask &= ~(1 << unit);
 		}
@@ -952,8 +982,8 @@ nve0_fifo_intr(struct nouveau_subdev *subdev)
 	}
 
 	if (stat & 0x80000000) {
-		nve0_fifo_intr_engine(priv);
 		nv_wr32(priv, 0x002100, 0x80000000);
+		nve0_fifo_intr_engine(priv);
 		stat &= ~0x80000000;
 	}
 
@@ -1022,6 +1052,12 @@ nve0_fifo_init(struct nouveau_object *object)
 		nv_wr32(priv, 0x04010c + (i * 0x2000), 0xfffffeff); /* INTREN */
 	}
 
+	/* PBDMA[n].HCE */
+	for (i = 0; i < priv->spoon_nr; i++) {
+		nv_wr32(priv, 0x040148 + (i * 0x2000), 0xffffffff); /* INTR */
+		nv_wr32(priv, 0x04014c + (i * 0x2000), 0xffffffff); /* INTREN */
+	}
+
 	nv_wr32(priv, 0x002254, 0x10000000 | priv->user.bar.offset >> 12);
 
 	nv_wr32(priv, 0x002100, 0xffffffff);
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/ctxnv50.c b/drivers/gpu/drm/nouveau/core/engine/graph/ctxnv50.c
index 552fdbd45ebe..1d0e33fb5f61 100644
--- a/drivers/gpu/drm/nouveau/core/engine/graph/ctxnv50.c
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/ctxnv50.c
@@ -113,6 +113,8 @@
 #define IS_NVA3F(x) (((x) > 0xa0 && (x) < 0xaa) || (x) == 0xaf)
 #define IS_NVAAF(x) ((x) >= 0xaa && (x) <= 0xac)
 
+#include <subdev/fb.h>
+
 /*
  * This code deals with PGRAPH contexts on NV50 family cards. Like NV40, it's
  * the GPU itself that does context-switching, but it needs a special
@@ -569,8 +571,12 @@ nv50_graph_construct_mmio(struct nouveau_grctx *ctx)
 			gr_def(ctx, 0x407d08, 0x00010040);
 		else if (device->chipset < 0xa0)
 			gr_def(ctx, 0x407d08, 0x00390040);
-		else
-			gr_def(ctx, 0x407d08, 0x003d0040);
+		else {
+			if (nouveau_fb(device)->ram->type != NV_MEM_TYPE_GDDR5)
+				gr_def(ctx, 0x407d08, 0x003d0040);
+			else
+				gr_def(ctx, 0x407d08, 0x003c0040);
+		}
 		gr_def(ctx, 0x407d0c, 0x00000022);
 	}
 
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/nvc0.c b/drivers/gpu/drm/nouveau/core/engine/graph/nvc0.c
index 30fd1dc64f93..17251e4b9e86 100644
--- a/drivers/gpu/drm/nouveau/core/engine/graph/nvc0.c
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/nvc0.c
@@ -1557,7 +1557,7 @@ nvc0_graph_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 		    nvc0_graph_ctor_fw(priv, "fuc409d", &priv->fuc409d) ||
 		    nvc0_graph_ctor_fw(priv, "fuc41ac", &priv->fuc41ac) ||
 		    nvc0_graph_ctor_fw(priv, "fuc41ad", &priv->fuc41ad))
-			return -EINVAL;
+			return -ENODEV;
 		priv->firmware = true;
 	}
 
diff --git a/drivers/gpu/drm/nouveau/core/engine/software/nv50.c b/drivers/gpu/drm/nouveau/core/engine/software/nv50.c
index 4d2994d8cc32..a0fec205f9db 100644
--- a/drivers/gpu/drm/nouveau/core/engine/software/nv50.c
+++ b/drivers/gpu/drm/nouveau/core/engine/software/nv50.c
@@ -175,7 +175,8 @@ nv50_software_context_ctor(struct nouveau_object *parent,
 		return ret;
 
 	for (i = 0; pdisp && i < pdisp->vblank.index_nr; i++) {
-		ret = nvkm_notify_init(&pdisp->vblank, pclass->vblank, false,
+		ret = nvkm_notify_init(NULL, &pdisp->vblank, pclass->vblank,
+				       false,
 				       &(struct nvif_notify_head_req_v0) {
 					.head = i,
 				       },
diff --git a/drivers/gpu/drm/nouveau/core/include/core/client.h b/drivers/gpu/drm/nouveau/core/include/core/client.h
index 1794a05205d8..b0ce9f6680b5 100644
--- a/drivers/gpu/drm/nouveau/core/include/core/client.h
+++ b/drivers/gpu/drm/nouveau/core/include/core/client.h
@@ -48,7 +48,7 @@ int  nouveau_client_init(struct nouveau_client *);
 int  nouveau_client_fini(struct nouveau_client *, bool suspend);
 const char *nouveau_client_name(void *obj);
 
-int nvkm_client_notify_new(struct nouveau_client *, struct nvkm_event *,
+int nvkm_client_notify_new(struct nouveau_object *, struct nvkm_event *,
 			   void *data, u32 size);
 int nvkm_client_notify_del(struct nouveau_client *, int index);
 int nvkm_client_notify_get(struct nouveau_client *, int index);
diff --git a/drivers/gpu/drm/nouveau/core/include/core/device.h b/drivers/gpu/drm/nouveau/core/include/core/device.h
index 8743766454a5..2ec2e50d3676 100644
--- a/drivers/gpu/drm/nouveau/core/include/core/device.h
+++ b/drivers/gpu/drm/nouveau/core/include/core/device.h
@@ -16,6 +16,7 @@ enum nv_subdev_type {
 	 * to during POST.
 	 */
 	NVDEV_SUBDEV_DEVINIT,
+	NVDEV_SUBDEV_IBUS,
 	NVDEV_SUBDEV_GPIO,
 	NVDEV_SUBDEV_I2C,
 	NVDEV_SUBDEV_DEVINIT_LAST = NVDEV_SUBDEV_I2C,
@@ -24,13 +25,13 @@ enum nv_subdev_type {
 	 * been created, and are allowed to assume any subdevs in the
 	 * list above them exist and have been initialised.
 	 */
+	NVDEV_SUBDEV_FUSE,
 	NVDEV_SUBDEV_MXM,
 	NVDEV_SUBDEV_MC,
 	NVDEV_SUBDEV_BUS,
 	NVDEV_SUBDEV_TIMER,
 	NVDEV_SUBDEV_FB,
 	NVDEV_SUBDEV_LTC,
-	NVDEV_SUBDEV_IBUS,
 	NVDEV_SUBDEV_INSTMEM,
 	NVDEV_SUBDEV_VM,
 	NVDEV_SUBDEV_BAR,
@@ -91,6 +92,7 @@ struct nouveau_device {
 		GM100    = 0x110,
 	} card_type;
 	u32 chipset;
+	u8  chiprev;
 	u32 crystal;
 
 	struct nouveau_oclass *oclass[NVDEV_SUBDEV_NR];
@@ -157,6 +159,12 @@ nv_device_is_pci(struct nouveau_device *device)
 	return device->pdev != NULL;
 }
 
+static inline bool
+nv_device_is_cpu_coherent(struct nouveau_device *device)
+{
+	return (!IS_ENABLED(CONFIG_ARM) && nv_device_is_pci(device));
+}
+
 static inline struct device *
 nv_device_base(struct nouveau_device *device)
 {
diff --git a/drivers/gpu/drm/nouveau/core/include/core/event.h b/drivers/gpu/drm/nouveau/core/include/core/event.h
index 51e55d03330a..92876528972f 100644
--- a/drivers/gpu/drm/nouveau/core/include/core/event.h
+++ b/drivers/gpu/drm/nouveau/core/include/core/event.h
@@ -4,7 +4,8 @@
 #include <core/notify.h>
 
 struct nvkm_event_func {
-	int  (*ctor)(void *data, u32 size, struct nvkm_notify *);
+	int  (*ctor)(struct nouveau_object *, void *data, u32 size,
+		     struct nvkm_notify *);
 	void (*send)(void *data, u32 size, struct nvkm_notify *);
 	void (*init)(struct nvkm_event *, int type, int index);
 	void (*fini)(struct nvkm_event *, int type, int index);
diff --git a/drivers/gpu/drm/nouveau/core/include/core/handle.h b/drivers/gpu/drm/nouveau/core/include/core/handle.h
index ceb67d770875..d22a59138a9b 100644
--- a/drivers/gpu/drm/nouveau/core/include/core/handle.h
+++ b/drivers/gpu/drm/nouveau/core/include/core/handle.h
@@ -23,11 +23,6 @@ void nouveau_handle_destroy(struct nouveau_handle *);
 int  nouveau_handle_init(struct nouveau_handle *);
 int  nouveau_handle_fini(struct nouveau_handle *, bool suspend);
 
-int  nouveau_handle_new(struct nouveau_object *, u32 parent, u32 handle,
-			u16 oclass, void *data, u32 size,
-			struct nouveau_object **);
-int  nouveau_handle_del(struct nouveau_object *, u32 parent, u32 handle);
-
 struct nouveau_object *
 nouveau_handle_ref(struct nouveau_object *, u32 name);
 
diff --git a/drivers/gpu/drm/nouveau/core/include/core/mm.h b/drivers/gpu/drm/nouveau/core/include/core/mm.h
index 2bf7d0e32261..bfe6931544fe 100644
--- a/drivers/gpu/drm/nouveau/core/include/core/mm.h
+++ b/drivers/gpu/drm/nouveau/core/include/core/mm.h
@@ -6,6 +6,10 @@ struct nouveau_mm_node {
 	struct list_head fl_entry;
 	struct list_head rl_entry;
 
+#define NVKM_MM_HEAP_ANY 0x00
+	u8  heap;
+#define NVKM_MM_TYPE_NONE 0x00
+#define NVKM_MM_TYPE_HOLE 0xff
 	u8  type;
 	u32 offset;
 	u32 length;
@@ -27,10 +31,10 @@ nouveau_mm_initialised(struct nouveau_mm *mm)
 
 int  nouveau_mm_init(struct nouveau_mm *, u32 offset, u32 length, u32 block);
 int  nouveau_mm_fini(struct nouveau_mm *);
-int  nouveau_mm_head(struct nouveau_mm *, u8 type, u32 size_max, u32 size_min,
-		     u32 align, struct nouveau_mm_node **);
-int  nouveau_mm_tail(struct nouveau_mm *, u8 type, u32 size_max, u32 size_min,
-		     u32 align, struct nouveau_mm_node **);
+int  nouveau_mm_head(struct nouveau_mm *, u8 heap, u8 type, u32 size_max,
+		     u32 size_min, u32 align, struct nouveau_mm_node **);
+int  nouveau_mm_tail(struct nouveau_mm *, u8 heap, u8 type, u32 size_max,
+		     u32 size_min, u32 align, struct nouveau_mm_node **);
 void nouveau_mm_free(struct nouveau_mm *, struct nouveau_mm_node **);
 
 #endif
diff --git a/drivers/gpu/drm/nouveau/core/include/core/notify.h b/drivers/gpu/drm/nouveau/core/include/core/notify.h
index 1262d8f020f3..a7c3c5f578cc 100644
--- a/drivers/gpu/drm/nouveau/core/include/core/notify.h
+++ b/drivers/gpu/drm/nouveau/core/include/core/notify.h
@@ -25,8 +25,9 @@ struct nvkm_notify {
 	const void *data;
 };
 
-int  nvkm_notify_init(struct nvkm_event *, int (*func)(struct nvkm_notify *),
-		      bool work, void *data, u32 size, u32 reply,
+int  nvkm_notify_init(struct nouveau_object *, struct nvkm_event *,
+		      int (*func)(struct nvkm_notify *), bool work,
+		      void *data, u32 size, u32 reply,
 		      struct nvkm_notify *);
 void nvkm_notify_fini(struct nvkm_notify *);
 void nvkm_notify_get(struct nvkm_notify *);
diff --git a/drivers/gpu/drm/nouveau/core/include/core/object.h b/drivers/gpu/drm/nouveau/core/include/core/object.h
index d7039482d6fd..2e2afa502c99 100644
--- a/drivers/gpu/drm/nouveau/core/include/core/object.h
+++ b/drivers/gpu/drm/nouveau/core/include/core/object.h
@@ -203,21 +203,4 @@ nv_memcmp(void *obj, u32 addr, const char *str, u32 len)
 	return 0;
 }
 
-#include <core/handle.h>
-
-static inline int
-nouveau_object_new(struct nouveau_object *client, u32 parent, u32 handle,
-		   u16 oclass, void *data, u32 size,
-		   struct nouveau_object **pobject)
-{
-	return nouveau_handle_new(client, parent, handle, oclass,
-				  data, size, pobject);
-}
-
-static inline int
-nouveau_object_del(struct nouveau_object *client, u32 parent, u32 handle)
-{
-	return nouveau_handle_del(client, parent, handle);
-}
-
 #endif
diff --git a/drivers/gpu/drm/nouveau/core/include/engine/disp.h b/drivers/gpu/drm/nouveau/core/include/engine/disp.h
index 7a64f347b385..fc307f1317ff 100644
--- a/drivers/gpu/drm/nouveau/core/include/engine/disp.h
+++ b/drivers/gpu/drm/nouveau/core/include/engine/disp.h
@@ -31,5 +31,6 @@ extern struct nouveau_oclass *nvd0_disp_oclass;
 extern struct nouveau_oclass *nve0_disp_oclass;
 extern struct nouveau_oclass *nvf0_disp_oclass;
 extern struct nouveau_oclass *gm107_disp_oclass;
+extern struct nouveau_oclass *gm204_disp_oclass;
 
 #endif
diff --git a/drivers/gpu/drm/nouveau/core/include/engine/fifo.h b/drivers/gpu/drm/nouveau/core/include/engine/fifo.h
index e5e4d930b2c2..2007453f6fce 100644
--- a/drivers/gpu/drm/nouveau/core/include/engine/fifo.h
+++ b/drivers/gpu/drm/nouveau/core/include/engine/fifo.h
@@ -116,7 +116,8 @@ extern struct nouveau_oclass *nve0_fifo_oclass;
 extern struct nouveau_oclass *gk20a_fifo_oclass;
 extern struct nouveau_oclass *nv108_fifo_oclass;
 
-int  nouveau_fifo_uevent_ctor(void *, u32, struct nvkm_notify *);
+int  nouveau_fifo_uevent_ctor(struct nouveau_object *, void *, u32,
+			      struct nvkm_notify *);
 void nouveau_fifo_uevent(struct nouveau_fifo *);
 
 void nv04_fifo_intr(struct nouveau_subdev *);
diff --git a/drivers/gpu/drm/nouveau/core/include/subdev/bar.h b/drivers/gpu/drm/nouveau/core/include/subdev/bar.h
index be037fac534c..257ddf6d36d4 100644
--- a/drivers/gpu/drm/nouveau/core/include/subdev/bar.h
+++ b/drivers/gpu/drm/nouveau/core/include/subdev/bar.h
@@ -12,7 +12,6 @@ struct nouveau_bar {
 
 	int (*alloc)(struct nouveau_bar *, struct nouveau_object *,
 		     struct nouveau_mem *, struct nouveau_object **);
-	void __iomem *iomem;
 
 	int (*kmap)(struct nouveau_bar *, struct nouveau_mem *,
 		    u32 flags, struct nouveau_vma *);
diff --git a/drivers/gpu/drm/nouveau/core/include/subdev/bios/M0203.h b/drivers/gpu/drm/nouveau/core/include/subdev/bios/M0203.h
new file mode 100644
index 000000000000..1f84d3612dd8
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/include/subdev/bios/M0203.h
@@ -0,0 +1,31 @@
+#ifndef __NVBIOS_M0203_H__
+#define __NVBIOS_M0203_H__
+
+struct nvbios_M0203T {
+#define M0203T_TYPE_RAMCFG 0x00
+	u8  type;
+	u16 pointer;
+};
+
+u32 nvbios_M0203Te(struct nouveau_bios *, u8 *ver, u8 *hdr, u8 *cnt, u8 *len);
+u32 nvbios_M0203Tp(struct nouveau_bios *, u8 *ver, u8 *hdr, u8 *cnt, u8 *len,
+		   struct nvbios_M0203T *);
+
+struct nvbios_M0203E {
+#define M0203E_TYPE_DDR2  0x0
+#define M0203E_TYPE_DDR3  0x1
+#define M0203E_TYPE_GDDR3 0x2
+#define M0203E_TYPE_GDDR5 0x3
+#define M0203E_TYPE_SKIP  0xf
+	u8 type;
+	u8 strap;
+	u8 group;
+};
+
+u32 nvbios_M0203Ee(struct nouveau_bios *, int idx, u8 *ver, u8 *hdr);
+u32 nvbios_M0203Ep(struct nouveau_bios *, int idx, u8 *ver, u8 *hdr,
+		   struct nvbios_M0203E *);
+u32 nvbios_M0203Em(struct nouveau_bios *, u8 ramcfg, u8 *ver, u8 *hdr,
+		   struct nvbios_M0203E *);
+
+#endif
diff --git a/drivers/gpu/drm/nouveau/core/include/subdev/bios/M0205.h b/drivers/gpu/drm/nouveau/core/include/subdev/bios/M0205.h
new file mode 100644
index 000000000000..e171120cec81
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/include/subdev/bios/M0205.h
@@ -0,0 +1,32 @@
+#ifndef __NVBIOS_M0205_H__
+#define __NVBIOS_M0205_H__
+
+struct nvbios_M0205T {
+	u16 freq;
+};
+
+u32 nvbios_M0205Te(struct nouveau_bios *,
+		   u8 *ver, u8 *hdr, u8 *cnt, u8 *len, u8 *snr, u8 *ssz);
+u32 nvbios_M0205Tp(struct nouveau_bios *,
+		   u8 *ver, u8 *hdr, u8 *cnt, u8 *len, u8 *snr, u8 *ssz,
+		   struct nvbios_M0205T *);
+
+struct nvbios_M0205E {
+	u8 type;
+};
+
+u32 nvbios_M0205Ee(struct nouveau_bios *, int idx,
+		   u8 *ver, u8 *hdr, u8 *cnt, u8 *len);
+u32 nvbios_M0205Ep(struct nouveau_bios *, int idx,
+		   u8 *ver, u8 *hdr, u8 *cnt, u8 *len,
+		   struct nvbios_M0205E *);
+
+struct nvbios_M0205S {
+	u8 data;
+};
+
+u32 nvbios_M0205Se(struct nouveau_bios *, int ent, int idx, u8 *ver, u8 *hdr);
+u32 nvbios_M0205Sp(struct nouveau_bios *, int ent, int idx, u8 *ver, u8 *hdr,
+		   struct nvbios_M0205S *);
+
+#endif
diff --git a/drivers/gpu/drm/nouveau/core/include/subdev/bios/M0209.h b/drivers/gpu/drm/nouveau/core/include/subdev/bios/M0209.h
new file mode 100644
index 000000000000..67dc50d837bc
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/include/subdev/bios/M0209.h
@@ -0,0 +1,30 @@
+#ifndef __NVBIOS_M0209_H__
+#define __NVBIOS_M0209_H__
+
+u32 nvbios_M0209Te(struct nouveau_bios *,
+		   u8 *ver, u8 *hdr, u8 *cnt, u8 *len, u8 *snr, u8 *ssz);
+
+struct nvbios_M0209E {
+	u8 v00_40;
+	u8 bits;
+	u8 modulo;
+	u8 v02_40;
+	u8 v02_07;
+	u8 v03;
+};
+
+u32 nvbios_M0209Ee(struct nouveau_bios *, int idx,
+		   u8 *ver, u8 *hdr, u8 *cnt, u8 *len);
+u32 nvbios_M0209Ep(struct nouveau_bios *, int idx,
+		   u8 *ver, u8 *hdr, u8 *cnt, u8 *len,
+		   struct nvbios_M0209E *);
+
+struct nvbios_M0209S {
+	u32 data[0x200];
+};
+
+u32 nvbios_M0209Se(struct nouveau_bios *, int ent, int idx, u8 *ver, u8 *hdr);
+u32 nvbios_M0209Sp(struct nouveau_bios *, int ent, int idx, u8 *ver, u8 *hdr,
+		   struct nvbios_M0209S *);
+
+#endif
diff --git a/drivers/gpu/drm/nouveau/core/include/subdev/bios/fan.h b/drivers/gpu/drm/nouveau/core/include/subdev/bios/fan.h
new file mode 100644
index 000000000000..119d0874e041
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/include/subdev/bios/fan.h
@@ -0,0 +1,8 @@
+#ifndef __NVBIOS_FAN_H__
+#define __NVBIOS_FAN_H__
+
+#include <subdev/bios/therm.h>
+
+u16 nvbios_fan_parse(struct nouveau_bios *bios, struct nvbios_therm_fan *fan);
+
+#endif
diff --git a/drivers/gpu/drm/nouveau/core/include/subdev/bios/i2c.h b/drivers/gpu/drm/nouveau/core/include/subdev/bios/i2c.h
index 10b57a19a7de..c9bb112895af 100644
--- a/drivers/gpu/drm/nouveau/core/include/subdev/bios/i2c.h
+++ b/drivers/gpu/drm/nouveau/core/include/subdev/bios/i2c.h
@@ -4,11 +4,14 @@
 struct nouveau_bios;
 
 enum dcb_i2c_type {
-	DCB_I2C_NV04_BIT = 0,
-	DCB_I2C_NV4E_BIT = 4,
-	DCB_I2C_NVIO_BIT = 5,
-	DCB_I2C_NVIO_AUX = 6,
-	DCB_I2C_UNUSED = 0xff
+	/* matches bios type field prior to ccb 4.1 */
+	DCB_I2C_NV04_BIT = 0x00,
+	DCB_I2C_NV4E_BIT = 0x04,
+	DCB_I2C_NVIO_BIT = 0x05,
+	DCB_I2C_NVIO_AUX = 0x06,
+	/* made up - mostly */
+	DCB_I2C_PMGR     = 0x80,
+	DCB_I2C_UNUSED   = 0xff
 };
 
 struct dcb_i2c_entry {
@@ -16,6 +19,7 @@ struct dcb_i2c_entry {
 	u8 drive;
 	u8 sense;
 	u8 share;
+	u8 auxch;
 };
 
 u16 dcb_i2c_table(struct nouveau_bios *, u8 *ver, u8 *hdr, u8 *cnt, u8 *len);
diff --git a/drivers/gpu/drm/nouveau/core/include/subdev/bios/image.h b/drivers/gpu/drm/nouveau/core/include/subdev/bios/image.h
new file mode 100644
index 000000000000..3348b4580843
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/include/subdev/bios/image.h
@@ -0,0 +1,13 @@
+#ifndef __NVBIOS_IMAGE_H__
+#define __NVBIOS_IMAGE_H__
+
+struct nvbios_image {
+	u32  base;
+	u32  size;
+	u8   type;
+	bool last;
+};
+
+bool nvbios_image(struct nouveau_bios *, int, struct nvbios_image *);
+
+#endif
diff --git a/drivers/gpu/drm/nouveau/core/include/subdev/bios/npde.h b/drivers/gpu/drm/nouveau/core/include/subdev/bios/npde.h
new file mode 100644
index 000000000000..b18413d951e5
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/include/subdev/bios/npde.h
@@ -0,0 +1,12 @@
+#ifndef __NVBIOS_NPDE_H__
+#define __NVBIOS_NPDE_H__
+
+struct nvbios_npdeT {
+	u32 image_size;
+	bool last;
+};
+
+u32 nvbios_npdeTe(struct nouveau_bios *, u32);
+u32 nvbios_npdeTp(struct nouveau_bios *, u32, struct nvbios_npdeT *);
+
+#endif
diff --git a/drivers/gpu/drm/nouveau/core/include/subdev/bios/pcir.h b/drivers/gpu/drm/nouveau/core/include/subdev/bios/pcir.h
new file mode 100644
index 000000000000..3d634a06dca1
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/include/subdev/bios/pcir.h
@@ -0,0 +1,18 @@
+#ifndef __NVBIOS_PCIR_H__
+#define __NVBIOS_PCIR_H__
+
+struct nvbios_pcirT {
+	u16 vendor_id;
+	u16 device_id;
+	u8  class_code[3];
+	u32 image_size;
+	u16 image_rev;
+	u8  image_type;
+	bool last;
+};
+
+u32 nvbios_pcirTe(struct nouveau_bios *, u32, u8 *ver, u16 *hdr);
+u32 nvbios_pcirTp(struct nouveau_bios *, u32, u8 *ver, u16 *hdr,
+		  struct nvbios_pcirT *);
+
+#endif
diff --git a/drivers/gpu/drm/nouveau/core/include/subdev/bios/pmu.h b/drivers/gpu/drm/nouveau/core/include/subdev/bios/pmu.h
new file mode 100644
index 000000000000..9de593deaea8
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/include/subdev/bios/pmu.h
@@ -0,0 +1,37 @@
+#ifndef __NVBIOS_PMU_H__
+#define __NVBIOS_PMU_H__
+
+struct nvbios_pmuT {
+};
+
+u32 nvbios_pmuTe(struct nouveau_bios *, u8 *ver, u8 *hdr, u8 *cnt, u8 *len);
+u32 nvbios_pmuTp(struct nouveau_bios *, u8 *ver, u8 *hdr, u8 *cnt, u8 *len,
+		 struct nvbios_pmuT *);
+
+struct nvbios_pmuE {
+	u8  type;
+	u32 data;
+};
+
+u32 nvbios_pmuEe(struct nouveau_bios *, int idx, u8 *ver, u8 *hdr);
+u32 nvbios_pmuEp(struct nouveau_bios *, int idx, u8 *ver, u8 *hdr,
+		 struct nvbios_pmuE *);
+
+struct nvbios_pmuR {
+	u32 boot_addr_pmu;
+	u32 boot_addr;
+	u32 boot_size;
+	u32 code_addr_pmu;
+	u32 code_addr;
+	u32 code_size;
+	u32 init_addr_pmu;
+
+	u32 data_addr_pmu;
+	u32 data_addr;
+	u32 data_size;
+	u32 args_addr_pmu;
+};
+
+bool nvbios_pmuRm(struct nouveau_bios *, u8 type, struct nvbios_pmuR *);
+
+#endif
diff --git a/drivers/gpu/drm/nouveau/core/include/subdev/bios/ramcfg.h b/drivers/gpu/drm/nouveau/core/include/subdev/bios/ramcfg.h
index c086ac6d677d..4a0e0ceb41ba 100644
--- a/drivers/gpu/drm/nouveau/core/include/subdev/bios/ramcfg.h
+++ b/drivers/gpu/drm/nouveau/core/include/subdev/bios/ramcfg.h
@@ -4,60 +4,139 @@
 struct nouveau_bios;
 
 struct nvbios_ramcfg {
-	unsigned rammap_11_08_01:1;
-	unsigned rammap_11_08_0c:2;
-	unsigned rammap_11_08_10:1;
-	unsigned rammap_11_11_0c:2;
+	unsigned rammap_ver;
+	unsigned rammap_hdr;
+	unsigned rammap_min;
+	unsigned rammap_max;
+	union {
+		struct {
+			unsigned rammap_10_04_02:1;
+			unsigned rammap_10_04_08:1;
+		};
+		struct {
+			unsigned rammap_11_08_01:1;
+			unsigned rammap_11_08_0c:2;
+			unsigned rammap_11_08_10:1;
+			unsigned rammap_11_09_01ff:9;
+			unsigned rammap_11_0a_03fe:9;
+			unsigned rammap_11_0a_0400:1;
+			unsigned rammap_11_0a_0800:1;
+			unsigned rammap_11_0b_01f0:5;
+			unsigned rammap_11_0b_0200:1;
+			unsigned rammap_11_0b_0400:1;
+			unsigned rammap_11_0b_0800:1;
+			unsigned rammap_11_0d:8;
+			unsigned rammap_11_0e:8;
+			unsigned rammap_11_0f:8;
+			unsigned rammap_11_11_0c:2;
+		};
+	};
 
-	unsigned ramcfg_11_01_01:1;
-	unsigned ramcfg_11_01_02:1;
-	unsigned ramcfg_11_01_04:1;
-	unsigned ramcfg_11_01_08:1;
-	unsigned ramcfg_11_01_10:1;
-	unsigned ramcfg_11_01_20:1;
-	unsigned ramcfg_11_01_40:1;
-	unsigned ramcfg_11_01_80:1;
-	unsigned ramcfg_11_02_03:2;
-	unsigned ramcfg_11_02_04:1;
-	unsigned ramcfg_11_02_08:1;
-	unsigned ramcfg_11_02_10:1;
-	unsigned ramcfg_11_02_40:1;
-	unsigned ramcfg_11_02_80:1;
-	unsigned ramcfg_11_03_0f:4;
-	unsigned ramcfg_11_03_30:2;
-	unsigned ramcfg_11_03_c0:2;
-	unsigned ramcfg_11_03_f0:4;
-	unsigned ramcfg_11_04:8;
-	unsigned ramcfg_11_06:8;
-	unsigned ramcfg_11_07_02:1;
-	unsigned ramcfg_11_07_04:1;
-	unsigned ramcfg_11_07_08:1;
-	unsigned ramcfg_11_07_10:1;
-	unsigned ramcfg_11_07_40:1;
-	unsigned ramcfg_11_07_80:1;
-	unsigned ramcfg_11_08_01:1;
-	unsigned ramcfg_11_08_02:1;
-	unsigned ramcfg_11_08_04:1;
-	unsigned ramcfg_11_08_08:1;
-	unsigned ramcfg_11_08_10:1;
-	unsigned ramcfg_11_08_20:1;
-	unsigned ramcfg_11_09:8;
+	unsigned ramcfg_ver;
+	unsigned ramcfg_hdr;
+	unsigned ramcfg_timing;
+	union {
+		struct {
+			unsigned ramcfg_10_02_01:1;
+			unsigned ramcfg_10_02_02:1;
+			unsigned ramcfg_10_02_04:1;
+			unsigned ramcfg_10_02_08:1;
+			unsigned ramcfg_10_02_10:1;
+			unsigned ramcfg_10_02_20:1;
+			unsigned ramcfg_10_DLLoff:1;
+			unsigned ramcfg_10_03_0f:4;
+			unsigned ramcfg_10_04_01:1;
+			unsigned ramcfg_10_05:8;
+			unsigned ramcfg_10_06:8;
+			unsigned ramcfg_10_07:8;
+			unsigned ramcfg_10_08:8;
+			unsigned ramcfg_10_09_0f:4;
+			unsigned ramcfg_10_09_f0:4;
+		};
+		struct {
+			unsigned ramcfg_11_01_01:1;
+			unsigned ramcfg_11_01_02:1;
+			unsigned ramcfg_11_01_04:1;
+			unsigned ramcfg_11_01_08:1;
+			unsigned ramcfg_11_01_10:1;
+			unsigned ramcfg_11_01_20:1;
+			unsigned ramcfg_11_01_40:1;
+			unsigned ramcfg_11_01_80:1;
+			unsigned ramcfg_11_02_03:2;
+			unsigned ramcfg_11_02_04:1;
+			unsigned ramcfg_11_02_08:1;
+			unsigned ramcfg_11_02_10:1;
+			unsigned ramcfg_11_02_40:1;
+			unsigned ramcfg_11_02_80:1;
+			unsigned ramcfg_11_03_0f:4;
+			unsigned ramcfg_11_03_30:2;
+			unsigned ramcfg_11_03_c0:2;
+			unsigned ramcfg_11_03_f0:4;
+			unsigned ramcfg_11_04:8;
+			unsigned ramcfg_11_06:8;
+			unsigned ramcfg_11_07_02:1;
+			unsigned ramcfg_11_07_04:1;
+			unsigned ramcfg_11_07_08:1;
+			unsigned ramcfg_11_07_10:1;
+			unsigned ramcfg_11_07_40:1;
+			unsigned ramcfg_11_07_80:1;
+			unsigned ramcfg_11_08_01:1;
+			unsigned ramcfg_11_08_02:1;
+			unsigned ramcfg_11_08_04:1;
+			unsigned ramcfg_11_08_08:1;
+			unsigned ramcfg_11_08_10:1;
+			unsigned ramcfg_11_08_20:1;
+			unsigned ramcfg_11_09:8;
+		};
+	};
 
+	unsigned timing_ver;
+	unsigned timing_hdr;
 	unsigned timing[11];
-	unsigned timing_20_2e_03:2;
-	unsigned timing_20_2e_30:2;
-	unsigned timing_20_2e_c0:2;
-	unsigned timing_20_2f_03:2;
-	unsigned timing_20_2c_003f:6;
-	unsigned timing_20_2c_1fc0:7;
-	unsigned timing_20_30_f8:5;
-	unsigned timing_20_30_07:3;
-	unsigned timing_20_31_0007:3;
-	unsigned timing_20_31_0078:4;
-	unsigned timing_20_31_0780:4;
-	unsigned timing_20_31_0800:1;
-	unsigned timing_20_31_7000:3;
-	unsigned timing_20_31_8000:1;
+	union {
+		struct {
+			unsigned timing_10_WR:8;
+			unsigned timing_10_WTR:8;
+			unsigned timing_10_CL:8;
+			unsigned timing_10_RC:8;
+			/*empty: 4 */
+			unsigned timing_10_RFC:8;        /* Byte 5 */
+			/*empty: 6 */
+			unsigned timing_10_RAS:8;        /* Byte 7 */
+			/*empty: 8 */
+			unsigned timing_10_RP:8;         /* Byte 9 */
+			unsigned timing_10_RCDRD:8;
+			unsigned timing_10_RCDWR:8;
+			unsigned timing_10_RRD:8;
+			unsigned timing_10_13:8;
+			unsigned timing_10_ODT:3;
+			/* empty: 15 */
+			unsigned timing_10_16:8;
+			/* empty: 17 */
+			unsigned timing_10_18:8;
+			unsigned timing_10_CWL:8;
+			unsigned timing_10_20:8;
+			unsigned timing_10_21:8;
+			/* empty: 22, 23 */
+			unsigned timing_10_24:8;
+		};
+		struct {
+			unsigned timing_20_2e_03:2;
+			unsigned timing_20_2e_30:2;
+			unsigned timing_20_2e_c0:2;
+			unsigned timing_20_2f_03:2;
+			unsigned timing_20_2c_003f:6;
+			unsigned timing_20_2c_1fc0:7;
+			unsigned timing_20_30_f8:5;
+			unsigned timing_20_30_07:3;
+			unsigned timing_20_31_0007:3;
+			unsigned timing_20_31_0078:4;
+			unsigned timing_20_31_0780:4;
+			unsigned timing_20_31_0800:1;
+			unsigned timing_20_31_7000:3;
+			unsigned timing_20_31_8000:1;
+		};
+	};
 };
 
 u8 nvbios_ramcfg_count(struct nouveau_bios *);
diff --git a/drivers/gpu/drm/nouveau/core/include/subdev/bios/rammap.h b/drivers/gpu/drm/nouveau/core/include/subdev/bios/rammap.h
index 5bdf8e4db40a..47e021d3e20d 100644
--- a/drivers/gpu/drm/nouveau/core/include/subdev/bios/rammap.h
+++ b/drivers/gpu/drm/nouveau/core/include/subdev/bios/rammap.h
@@ -8,9 +8,10 @@ u32 nvbios_rammapTe(struct nouveau_bios *, u8 *ver, u8 *hdr,
 
 u32 nvbios_rammapEe(struct nouveau_bios *, int idx,
 		    u8 *ver, u8 *hdr, u8 *cnt, u8 *len);
+u32 nvbios_rammapEp(struct nouveau_bios *, int idx,
+		    u8 *ver, u8 *hdr, u8 *cnt, u8 *len,
+		    struct nvbios_ramcfg *);
 u32 nvbios_rammapEm(struct nouveau_bios *, u16 mhz,
-		    u8 *ver, u8 *hdr, u8 *cnt, u8 *len);
-u32 nvbios_rammapEp(struct nouveau_bios *, u16 mhz,
 		    u8 *ver, u8 *hdr, u8 *cnt, u8 *len,
 		    struct nvbios_ramcfg *);
 
diff --git a/drivers/gpu/drm/nouveau/core/include/subdev/bios/therm.h b/drivers/gpu/drm/nouveau/core/include/subdev/bios/therm.h
index 8dc5051df55d..295d093f3b30 100644
--- a/drivers/gpu/drm/nouveau/core/include/subdev/bios/therm.h
+++ b/drivers/gpu/drm/nouveau/core/include/subdev/bios/therm.h
@@ -23,6 +23,12 @@ struct nvbios_therm_sensor {
 	struct nvbios_therm_threshold thrs_shutdown;
 };
 
+enum nvbios_therm_fan_type {
+	NVBIOS_THERM_FAN_UNK = 0,
+	NVBIOS_THERM_FAN_TOGGLE = 1,
+	NVBIOS_THERM_FAN_PWM = 2,
+};
+
 /* no vbios have more than 6 */
 #define NOUVEAU_TEMP_FAN_TRIP_MAX 10
 struct nouveau_therm_trip_point {
@@ -38,7 +44,9 @@ enum nvbios_therm_fan_mode {
 };
 
 struct nvbios_therm_fan {
-	u16 pwm_freq;
+	enum nvbios_therm_fan_type type;
+
+	u32 pwm_freq;
 
 	u8 min_duty;
 	u8 max_duty;
diff --git a/drivers/gpu/drm/nouveau/core/include/subdev/clock.h b/drivers/gpu/drm/nouveau/core/include/subdev/clock.h
index a5ca00dd2f61..36ed035d4d42 100644
--- a/drivers/gpu/drm/nouveau/core/include/subdev/clock.h
+++ b/drivers/gpu/drm/nouveau/core/include/subdev/clock.h
@@ -29,6 +29,7 @@ enum nv_clk_src {
 	nv_clk_src_mdiv,
 
 	nv_clk_src_core,
+	nv_clk_src_core_intm,
 	nv_clk_src_shader,
 
 	nv_clk_src_mem,
diff --git a/drivers/gpu/drm/nouveau/core/include/subdev/devinit.h b/drivers/gpu/drm/nouveau/core/include/subdev/devinit.h
index e292271a84e4..e007a9d44683 100644
--- a/drivers/gpu/drm/nouveau/core/include/subdev/devinit.h
+++ b/drivers/gpu/drm/nouveau/core/include/subdev/devinit.h
@@ -30,5 +30,6 @@ extern struct nouveau_oclass *nva3_devinit_oclass;
 extern struct nouveau_oclass *nvaf_devinit_oclass;
 extern struct nouveau_oclass *nvc0_devinit_oclass;
 extern struct nouveau_oclass *gm107_devinit_oclass;
+extern struct nouveau_oclass *gm204_devinit_oclass;
 
 #endif
diff --git a/drivers/gpu/drm/nouveau/core/include/subdev/fb.h b/drivers/gpu/drm/nouveau/core/include/subdev/fb.h
index 871e73914b24..8d0032f15205 100644
--- a/drivers/gpu/drm/nouveau/core/include/subdev/fb.h
+++ b/drivers/gpu/drm/nouveau/core/include/subdev/fb.h
@@ -111,6 +111,7 @@ extern struct nouveau_oclass *gm107_fb_oclass;
 #include <subdev/bios/ramcfg.h>
 
 struct nouveau_ram_data {
+	struct list_head head;
 	struct nvbios_ramcfg bios;
 	u32 freq;
 };
@@ -136,6 +137,7 @@ struct nouveau_ram {
 
 	int ranks;
 	int parts;
+	int part_mask;
 
 	int  (*get)(struct nouveau_fb *, u64 size, u32 align,
 		    u32 size_nc, u32 type, struct nouveau_mem **);
@@ -144,11 +146,6 @@ struct nouveau_ram {
 	int  (*calc)(struct nouveau_fb *, u32 freq);
 	int  (*prog)(struct nouveau_fb *);
 	void (*tidy)(struct nouveau_fb *);
-	struct {
-		u8  version;
-		u32 data;
-		u8  size;
-	} rammap, ramcfg, timing;
 	u32 freq;
 	u32 mr[16];
 	u32 mr1_nuts;
diff --git a/drivers/gpu/drm/nouveau/core/include/subdev/fb/regsnv04.h b/drivers/gpu/drm/nouveau/core/include/subdev/fb/regsnv04.h
new file mode 100644
index 000000000000..0f7fc0c52ab2
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/include/subdev/fb/regsnv04.h
@@ -0,0 +1,21 @@
+#ifndef __NOUVEAU_FB_REGS_04_H__
+#define __NOUVEAU_FB_REGS_04_H__
+
+#define NV04_PFB_BOOT_0						0x00100000
+#	define NV04_PFB_BOOT_0_RAM_AMOUNT			0x00000003
+#	define NV04_PFB_BOOT_0_RAM_AMOUNT_32MB			0x00000000
+#	define NV04_PFB_BOOT_0_RAM_AMOUNT_4MB			0x00000001
+#	define NV04_PFB_BOOT_0_RAM_AMOUNT_8MB			0x00000002
+#	define NV04_PFB_BOOT_0_RAM_AMOUNT_16MB			0x00000003
+#	define NV04_PFB_BOOT_0_RAM_WIDTH_128			0x00000004
+#	define NV04_PFB_BOOT_0_RAM_TYPE				0x00000028
+#	define NV04_PFB_BOOT_0_RAM_TYPE_SGRAM_8MBIT		0x00000000
+#	define NV04_PFB_BOOT_0_RAM_TYPE_SGRAM_16MBIT		0x00000008
+#	define NV04_PFB_BOOT_0_RAM_TYPE_SGRAM_16MBIT_4BANK	0x00000010
+#	define NV04_PFB_BOOT_0_RAM_TYPE_SDRAM_16MBIT		0x00000018
+#	define NV04_PFB_BOOT_0_RAM_TYPE_SDRAM_64MBIT		0x00000020
+#	define NV04_PFB_BOOT_0_RAM_TYPE_SDRAM_64MBITX16		0x00000028
+#	define NV04_PFB_BOOT_0_UMA_ENABLE			0x00000100
+#	define NV04_PFB_BOOT_0_UMA_SIZE				0x0000f000
+
+#endif
diff --git a/drivers/gpu/drm/nouveau/core/include/subdev/fuse.h b/drivers/gpu/drm/nouveau/core/include/subdev/fuse.h
new file mode 100644
index 000000000000..2b1ddb2a9a7d
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/include/subdev/fuse.h
@@ -0,0 +1,30 @@
+#ifndef __NOUVEAU_FUSE_H__
+#define __NOUVEAU_FUSE_H__
+
+#include <core/subdev.h>
+#include <core/device.h>
+
+struct nouveau_fuse {
+	struct nouveau_subdev base;
+};
+
+static inline struct nouveau_fuse *
+nouveau_fuse(void *obj)
+{
+	return (void *)nv_device(obj)->subdev[NVDEV_SUBDEV_FUSE];
+}
+
+#define nouveau_fuse_create(p, e, o, d)                                        \
+	nouveau_fuse_create_((p), (e), (o), sizeof(**d), (void **)d)
+
+int  nouveau_fuse_create_(struct nouveau_object *, struct nouveau_object *,
+			  struct nouveau_oclass *, int, void **);
+void _nouveau_fuse_dtor(struct nouveau_object *);
+int  _nouveau_fuse_init(struct nouveau_object *);
+#define _nouveau_fuse_fini _nouveau_subdev_fini
+
+extern struct nouveau_oclass g80_fuse_oclass;
+extern struct nouveau_oclass gf100_fuse_oclass;
+extern struct nouveau_oclass gm107_fuse_oclass;
+
+#endif
diff --git a/drivers/gpu/drm/nouveau/core/include/subdev/gpio.h b/drivers/gpu/drm/nouveau/core/include/subdev/gpio.h
index b73733d21cc7..f855140dbcb7 100644
--- a/drivers/gpu/drm/nouveau/core/include/subdev/gpio.h
+++ b/drivers/gpu/drm/nouveau/core/include/subdev/gpio.h
@@ -40,7 +40,7 @@ nouveau_gpio(void *obj)
 
 extern struct nouveau_oclass *nv10_gpio_oclass;
 extern struct nouveau_oclass *nv50_gpio_oclass;
-extern struct nouveau_oclass *nv92_gpio_oclass;
+extern struct nouveau_oclass *nv94_gpio_oclass;
 extern struct nouveau_oclass *nvd0_gpio_oclass;
 extern struct nouveau_oclass *nve0_gpio_oclass;
 
diff --git a/drivers/gpu/drm/nouveau/core/include/subdev/i2c.h b/drivers/gpu/drm/nouveau/core/include/subdev/i2c.h
index 1b937c2c25ae..d94ccacb40bf 100644
--- a/drivers/gpu/drm/nouveau/core/include/subdev/i2c.h
+++ b/drivers/gpu/drm/nouveau/core/include/subdev/i2c.h
@@ -8,6 +8,8 @@
 #include <subdev/bios/i2c.h>
 
 #define NV_I2C_PORT(n)    (0x00 + (n))
+#define NV_I2C_AUX(n)     (0x10 + (n))
+#define NV_I2C_EXT(n)     (0x20 + (n))
 #define NV_I2C_DEFAULT(n) (0x80 + (n))
 
 #define NV_I2C_TYPE_DCBI2C(n) (0x0000 | (n))
@@ -89,6 +91,7 @@ extern struct nouveau_oclass *nv94_i2c_oclass;
 extern struct nouveau_oclass *nvd0_i2c_oclass;
 extern struct nouveau_oclass *gf117_i2c_oclass;
 extern struct nouveau_oclass *nve0_i2c_oclass;
+extern struct nouveau_oclass *gm204_i2c_oclass;
 
 static inline int
 nv_rdi2cr(struct nouveau_i2c_port *port, u8 addr, u8 reg)
diff --git a/drivers/gpu/drm/nouveau/core/include/subdev/pwr.h b/drivers/gpu/drm/nouveau/core/include/subdev/pwr.h
index f73feec151db..f2427bf5aeed 100644
--- a/drivers/gpu/drm/nouveau/core/include/subdev/pwr.h
+++ b/drivers/gpu/drm/nouveau/core/include/subdev/pwr.h
@@ -47,5 +47,10 @@ void nouveau_memx_wr32(struct nouveau_memx *, u32 addr, u32 data);
 void nouveau_memx_wait(struct nouveau_memx *,
 		       u32 addr, u32 mask, u32 data, u32 nsec);
 void nouveau_memx_nsec(struct nouveau_memx *, u32 nsec);
+void nouveau_memx_wait_vblank(struct nouveau_memx *);
+void nouveau_memx_train(struct nouveau_memx *);
+int  nouveau_memx_train_result(struct nouveau_pwr *, u32 *, int);
+void nouveau_memx_block(struct nouveau_memx *);
+void nouveau_memx_unblock(struct nouveau_memx *);
 
 #endif
diff --git a/drivers/gpu/drm/nouveau/core/include/subdev/therm.h b/drivers/gpu/drm/nouveau/core/include/subdev/therm.h
index d4a68179e586..a437597dcafc 100644
--- a/drivers/gpu/drm/nouveau/core/include/subdev/therm.h
+++ b/drivers/gpu/drm/nouveau/core/include/subdev/therm.h
@@ -78,5 +78,6 @@ extern struct nouveau_oclass nv50_therm_oclass;
 extern struct nouveau_oclass nv84_therm_oclass;
 extern struct nouveau_oclass nva3_therm_oclass;
 extern struct nouveau_oclass nvd0_therm_oclass;
+extern struct nouveau_oclass gm107_therm_oclass;
 
 #endif
diff --git a/drivers/gpu/drm/nouveau/core/include/subdev/volt.h b/drivers/gpu/drm/nouveau/core/include/subdev/volt.h
index 820b62ffd75b..67db5e58880d 100644
--- a/drivers/gpu/drm/nouveau/core/include/subdev/volt.h
+++ b/drivers/gpu/drm/nouveau/core/include/subdev/volt.h
@@ -52,6 +52,7 @@ int  _nouveau_volt_init(struct nouveau_object *);
 #define _nouveau_volt_fini _nouveau_subdev_fini
 
 extern struct nouveau_oclass nv40_volt_oclass;
+extern struct nouveau_oclass gk20a_volt_oclass;
 
 int nouveau_voltgpio_init(struct nouveau_volt *);
 int nouveau_voltgpio_get(struct nouveau_volt *);
diff --git a/drivers/gpu/drm/nouveau/core/os.h b/drivers/gpu/drm/nouveau/core/os.h
index ccfa21d72ddc..bdd05ee7ec72 100644
--- a/drivers/gpu/drm/nouveau/core/os.h
+++ b/drivers/gpu/drm/nouveau/core/os.h
@@ -23,6 +23,7 @@
 #include <linux/pm_runtime.h>
 #include <linux/power_supply.h>
 #include <linux/clk.h>
+#include <linux/regulator/consumer.h>
 
 #include <asm/unaligned.h>
 
diff --git a/drivers/gpu/drm/nouveau/core/subdev/bar/base.c b/drivers/gpu/drm/nouveau/core/subdev/bar/base.c
index 8bcbdf39cfb2..b1adc69efd88 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/bar/base.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/bar/base.c
@@ -38,10 +38,12 @@ struct nouveau_barobj {
 static int
 nouveau_barobj_ctor(struct nouveau_object *parent,
 		    struct nouveau_object *engine,
-		    struct nouveau_oclass *oclass, void *mem, u32 size,
+		    struct nouveau_oclass *oclass, void *data, u32 size,
 		    struct nouveau_object **pobject)
 {
+	struct nouveau_device *device = nv_device(parent);
 	struct nouveau_bar *bar = (void *)engine;
+	struct nouveau_mem *mem = data;
 	struct nouveau_barobj *barobj;
 	int ret;
 
@@ -54,7 +56,13 @@ nouveau_barobj_ctor(struct nouveau_object *parent,
 	if (ret)
 		return ret;
 
-	barobj->iomem = bar->iomem + (u32)barobj->vma.offset;
+	barobj->iomem = ioremap(nv_device_resource_start(device, 3) +
+				(u32)barobj->vma.offset, mem->size << 12);
+	if (!barobj->iomem) {
+		nv_warn(bar, "PRAMIN ioremap failed\n");
+		return -ENOMEM;
+	}
+
 	return 0;
 }
 
@@ -63,8 +71,11 @@ nouveau_barobj_dtor(struct nouveau_object *object)
 {
 	struct nouveau_bar *bar = (void *)object->engine;
 	struct nouveau_barobj *barobj = (void *)object;
-	if (barobj->vma.node)
+	if (barobj->vma.node) {
+		if (barobj->iomem)
+			iounmap(barobj->iomem);
 		bar->unmap(bar, &barobj->vma);
+	}
 	nouveau_object_destroy(&barobj->base);
 }
 
@@ -99,12 +110,11 @@ nouveau_bar_alloc(struct nouveau_bar *bar, struct nouveau_object *parent,
 		  struct nouveau_mem *mem, struct nouveau_object **pobject)
 {
 	struct nouveau_object *engine = nv_object(bar);
-	int ret = -ENOMEM;
-	if (bar->iomem) {
-		ret = nouveau_object_ctor(parent, engine,
-					  &nouveau_barobj_oclass,
-					  mem, 0, pobject);
-	}
+	struct nouveau_object *gpuobj;
+	int ret = nouveau_object_ctor(parent, engine, &nouveau_barobj_oclass,
+				      mem, 0, &gpuobj);
+	if (ret == 0)
+		*pobject = gpuobj;
 	return ret;
 }
 
@@ -113,7 +123,6 @@ nouveau_bar_create_(struct nouveau_object *parent,
 		    struct nouveau_object *engine,
 		    struct nouveau_oclass *oclass, int length, void **pobject)
 {
-	struct nouveau_device *device = nv_device(parent);
 	struct nouveau_bar *bar;
 	int ret;
 
@@ -123,21 +132,12 @@ nouveau_bar_create_(struct nouveau_object *parent,
 	if (ret)
 		return ret;
 
-	if (nv_device_resource_len(device, 3) != 0) {
-		bar->iomem = ioremap(nv_device_resource_start(device, 3),
-				     nv_device_resource_len(device, 3));
-		if (!bar->iomem)
-			nv_warn(bar, "PRAMIN ioremap failed\n");
-	}
-
 	return 0;
 }
 
 void
 nouveau_bar_destroy(struct nouveau_bar *bar)
 {
-	if (bar->iomem)
-		iounmap(bar->iomem);
 	nouveau_subdev_destroy(&bar->base);
 }
 
diff --git a/drivers/gpu/drm/nouveau/core/subdev/bios/M0203.c b/drivers/gpu/drm/nouveau/core/subdev/bios/M0203.c
new file mode 100644
index 000000000000..28906b16d4e5
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/subdev/bios/M0203.c
@@ -0,0 +1,129 @@
+/*
+ * Copyright 2014 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs
+ */
+
+#include <subdev/bios.h>
+#include <subdev/bios/bit.h>
+#include <subdev/bios/M0203.h>
+
+u32
+nvbios_M0203Te(struct nouveau_bios *bios, u8 *ver, u8 *hdr, u8 *cnt, u8 *len)
+{
+	struct bit_entry bit_M;
+	u32 data = 0x00000000;
+
+	if (!bit_entry(bios, 'M', &bit_M)) {
+		if (bit_M.version == 2 && bit_M.length > 0x04)
+			data = nv_ro16(bios, bit_M.offset + 0x03);
+		if (data) {
+			*ver = nv_ro08(bios, data + 0x00);
+			switch (*ver) {
+			case 0x10:
+				*hdr = nv_ro08(bios, data + 0x01);
+				*len = nv_ro08(bios, data + 0x02);
+				*cnt = nv_ro08(bios, data + 0x03);
+				return data;
+			default:
+				break;
+			}
+		}
+	}
+
+	return 0x00000000;
+}
+
+u32
+nvbios_M0203Tp(struct nouveau_bios *bios, u8 *ver, u8 *hdr, u8 *cnt, u8 *len,
+	       struct nvbios_M0203T *info)
+{
+	u32 data = nvbios_M0203Te(bios, ver, hdr, cnt, len);
+	memset(info, 0x00, sizeof(*info));
+	switch (!!data * *ver) {
+	case 0x10:
+		info->type    = nv_ro08(bios, data + 0x04);
+		info->pointer = nv_ro16(bios, data + 0x05);
+		break;
+	default:
+		break;
+	}
+	return data;
+}
+
+u32
+nvbios_M0203Ee(struct nouveau_bios *bios, int idx, u8 *ver, u8 *hdr)
+{
+	u8  cnt, len;
+	u32 data = nvbios_M0203Te(bios, ver, hdr, &cnt, &len);
+	if (data && idx < cnt) {
+		data = data + *hdr + idx * len;
+		*hdr = len;
+		return data;
+	}
+	return 0x00000000;
+}
+
+u32
+nvbios_M0203Ep(struct nouveau_bios *bios, int idx, u8 *ver, u8 *hdr,
+	       struct nvbios_M0203E *info)
+{
+	u32 data = nvbios_M0203Ee(bios, idx, ver, hdr);
+	memset(info, 0x00, sizeof(*info));
+	switch (!!data * *ver) {
+	case 0x10:
+		info->type  = (nv_ro08(bios, data + 0x00) & 0x0f) >> 0;
+		info->strap = (nv_ro08(bios, data + 0x00) & 0xf0) >> 4;
+		info->group = (nv_ro08(bios, data + 0x01) & 0x0f) >> 0;
+		return data;
+	default:
+		break;
+	}
+	return 0x00000000;
+}
+
+u32
+nvbios_M0203Em(struct nouveau_bios *bios, u8 ramcfg, u8 *ver, u8 *hdr,
+	       struct nvbios_M0203E *info)
+{
+	struct nvbios_M0203T M0203T;
+	u8  cnt, len, idx = 0xff;
+	u32 data;
+
+	if (!nvbios_M0203Tp(bios, ver, hdr, &cnt, &len, &M0203T)) {
+		nv_warn(bios, "M0203T not found\n");
+		return 0x00000000;
+	}
+
+	while ((data = nvbios_M0203Ep(bios, ++idx, ver, hdr, info))) {
+		switch (M0203T.type) {
+		case M0203T_TYPE_RAMCFG:
+			if (info->strap != ramcfg)
+				continue;
+			return data;
+		default:
+			nv_warn(bios, "M0203T type %02x\n", M0203T.type);
+			return 0x00000000;
+		}
+	}
+
+	return data;
+}
diff --git a/drivers/gpu/drm/nouveau/core/subdev/bios/M0205.c b/drivers/gpu/drm/nouveau/core/subdev/bios/M0205.c
new file mode 100644
index 000000000000..ac9617c5fc2a
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/subdev/bios/M0205.c
@@ -0,0 +1,136 @@
+/*
+ * Copyright 2013 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs
+ */
+
+#include <subdev/bios.h>
+#include <subdev/bios/bit.h>
+#include <subdev/bios/M0205.h>
+
+u32
+nvbios_M0205Te(struct nouveau_bios *bios,
+	       u8 *ver, u8 *hdr, u8 *cnt, u8 *len, u8 *snr, u8 *ssz)
+{
+	struct bit_entry bit_M;
+	u32 data = 0x00000000;
+
+	if (!bit_entry(bios, 'M', &bit_M)) {
+		if (bit_M.version == 2 && bit_M.length > 0x08)
+			data = nv_ro32(bios, bit_M.offset + 0x05);
+		if (data) {
+			*ver = nv_ro08(bios, data + 0x00);
+			switch (*ver) {
+			case 0x10:
+				*hdr = nv_ro08(bios, data + 0x01);
+				*len = nv_ro08(bios, data + 0x02);
+				*ssz = nv_ro08(bios, data + 0x03);
+				*snr = nv_ro08(bios, data + 0x04);
+				*cnt = nv_ro08(bios, data + 0x05);
+				return data;
+			default:
+				break;
+			}
+		}
+	}
+
+	return 0x00000000;
+}
+
+u32
+nvbios_M0205Tp(struct nouveau_bios *bios,
+	       u8 *ver, u8 *hdr, u8 *cnt, u8 *len, u8 *snr, u8 *ssz,
+	       struct nvbios_M0205T *info)
+{
+	u32 data = nvbios_M0205Te(bios, ver, hdr, cnt, len, snr, ssz);
+	memset(info, 0x00, sizeof(*info));
+	switch (!!data * *ver) {
+	case 0x10:
+		info->freq = nv_ro16(bios, data + 0x06);
+		break;
+	default:
+		break;
+	}
+	return data;
+}
+
+u32
+nvbios_M0205Ee(struct nouveau_bios *bios, int idx,
+	       u8 *ver, u8 *hdr, u8 *cnt, u8 *len)
+{
+	u8  snr, ssz;
+	u32 data = nvbios_M0205Te(bios, ver, hdr, cnt, len, &snr, &ssz);
+	if (data && idx < *cnt) {
+		data = data + *hdr + idx * (*len + (snr * ssz));
+		*hdr = *len;
+		*cnt = snr;
+		*len = ssz;
+		return data;
+	}
+	return 0x00000000;
+}
+
+u32
+nvbios_M0205Ep(struct nouveau_bios *bios, int idx,
+	       u8 *ver, u8 *hdr, u8 *cnt, u8 *len,
+	       struct nvbios_M0205E *info)
+{
+	u32 data = nvbios_M0205Ee(bios, idx, ver, hdr, cnt, len);
+	memset(info, 0x00, sizeof(*info));
+	switch (!!data * *ver) {
+	case 0x10:
+		info->type = nv_ro08(bios, data + 0x00) & 0x0f;
+		return data;
+	default:
+		break;
+	}
+	return 0x00000000;
+}
+
+u32
+nvbios_M0205Se(struct nouveau_bios *bios, int ent, int idx, u8 *ver, u8 *hdr)
+{
+
+	u8  cnt, len;
+	u32 data = nvbios_M0205Ee(bios, ent, ver, hdr, &cnt, &len);
+	if (data && idx < cnt) {
+		data = data + *hdr + idx * len;
+		*hdr = len;
+		return data;
+	}
+	return 0x00000000;
+}
+
+u32
+nvbios_M0205Sp(struct nouveau_bios *bios, int ent, int idx, u8 *ver, u8 *hdr,
+	       struct nvbios_M0205S *info)
+{
+	u32 data = nvbios_M0205Se(bios, ent, idx, ver, hdr);
+	memset(info, 0x00, sizeof(*info));
+	switch (!!data * *ver) {
+	case 0x10:
+		info->data = nv_ro08(bios, data + 0x00);
+		return data;
+	default:
+		break;
+	}
+	return 0x00000000;
+}
diff --git a/drivers/gpu/drm/nouveau/core/subdev/bios/M0209.c b/drivers/gpu/drm/nouveau/core/subdev/bios/M0209.c
new file mode 100644
index 000000000000..b142a510e89f
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/subdev/bios/M0209.c
@@ -0,0 +1,137 @@
+/*
+ * Copyright 2013 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs
+ */
+
+#include <subdev/bios.h>
+#include <subdev/bios/bit.h>
+#include <subdev/bios/M0209.h>
+
+u32
+nvbios_M0209Te(struct nouveau_bios *bios,
+	       u8 *ver, u8 *hdr, u8 *cnt, u8 *len, u8 *snr, u8 *ssz)
+{
+	struct bit_entry bit_M;
+	u32 data = 0x00000000;
+
+	if (!bit_entry(bios, 'M', &bit_M)) {
+		if (bit_M.version == 2 && bit_M.length > 0x0c)
+			data = nv_ro32(bios, bit_M.offset + 0x09);
+		if (data) {
+			*ver = nv_ro08(bios, data + 0x00);
+			switch (*ver) {
+			case 0x10:
+				*hdr = nv_ro08(bios, data + 0x01);
+				*len = nv_ro08(bios, data + 0x02);
+				*ssz = nv_ro08(bios, data + 0x03);
+				*snr = 1;
+				*cnt = nv_ro08(bios, data + 0x04);
+				return data;
+			default:
+				break;
+			}
+		}
+	}
+
+	return 0x00000000;
+}
+
+u32
+nvbios_M0209Ee(struct nouveau_bios *bios, int idx,
+	       u8 *ver, u8 *hdr, u8 *cnt, u8 *len)
+{
+	u8  snr, ssz;
+	u32 data = nvbios_M0209Te(bios, ver, hdr, cnt, len, &snr, &ssz);
+	if (data && idx < *cnt) {
+		data = data + *hdr + idx * (*len + (snr * ssz));
+		*hdr = *len;
+		*cnt = snr;
+		*len = ssz;
+		return data;
+	}
+	return 0x00000000;
+}
+
+u32
+nvbios_M0209Ep(struct nouveau_bios *bios, int idx,
+	       u8 *ver, u8 *hdr, u8 *cnt, u8 *len,
+	       struct nvbios_M0209E *info)
+{
+	u32 data = nvbios_M0209Ee(bios, idx, ver, hdr, cnt, len);
+	memset(info, 0x00, sizeof(*info));
+	switch (!!data * *ver) {
+	case 0x10:
+		info->v00_40 = (nv_ro08(bios, data + 0x00) & 0x40) >> 6;
+		info->bits   =  nv_ro08(bios, data + 0x00) & 0x3f;
+		info->modulo =  nv_ro08(bios, data + 0x01);
+		info->v02_40 = (nv_ro08(bios, data + 0x02) & 0x40) >> 6;
+		info->v02_07 =  nv_ro08(bios, data + 0x02) & 0x07;
+		info->v03    =  nv_ro08(bios, data + 0x03);
+		return data;
+	default:
+		break;
+	}
+	return 0x00000000;
+}
+
+u32
+nvbios_M0209Se(struct nouveau_bios *bios, int ent, int idx, u8 *ver, u8 *hdr)
+{
+
+	u8  cnt, len;
+	u32 data = nvbios_M0209Ee(bios, ent, ver, hdr, &cnt, &len);
+	if (data && idx < cnt) {
+		data = data + *hdr + idx * len;
+		*hdr = len;
+		return data;
+	}
+	return 0x00000000;
+}
+
+u32
+nvbios_M0209Sp(struct nouveau_bios *bios, int ent, int idx, u8 *ver, u8 *hdr,
+	       struct nvbios_M0209S *info)
+{
+	struct nvbios_M0209E M0209E;
+	u8  cnt, len;
+	u32 data = nvbios_M0209Ep(bios, ent, ver, hdr, &cnt, &len, &M0209E);
+	if (data) {
+		u32 i, data = nvbios_M0209Se(bios, ent, idx, ver, hdr);
+		memset(info, 0x00, sizeof(*info));
+		switch (!!data * *ver) {
+		case 0x10:
+			for (i = 0; i < ARRAY_SIZE(info->data); i++) {
+				u32 bits = (i % M0209E.modulo) * M0209E.bits;
+				u32 mask = (1ULL << M0209E.bits) - 1;
+				u16  off = bits / 8;
+				u8   mod = bits % 8;
+				info->data[i] = nv_ro32(bios, data + off);
+				info->data[i] = info->data[i] >> mod;
+				info->data[i] = info->data[i] & mask;
+			}
+			return data;
+		default:
+			break;
+		}
+	}
+	return 0x00000000;
+}
diff --git a/drivers/gpu/drm/nouveau/core/subdev/bios/base.c b/drivers/gpu/drm/nouveau/core/subdev/bios/base.c
index d45704a2c2df..7df3a273553d 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/bios/base.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/bios/base.c
@@ -31,6 +31,8 @@
 #include <subdev/bios/bmp.h>
 #include <subdev/bios/bit.h>
 
+#include "priv.h"
+
 u8
 nvbios_checksum(const u8 *data, int size)
 {
@@ -56,362 +58,21 @@ nvbios_findstr(const u8 *data, int size, const char *str, int len)
 	return 0;
 }
 
-#if defined(__powerpc__)
-static void
-nouveau_bios_shadow_of(struct nouveau_bios *bios)
+int
+nvbios_extend(struct nouveau_bios *bios, u32 length)
 {
-	struct pci_dev *pdev = nv_device(bios)->pdev;
-	struct device_node *dn;
-	const u32 *data;
-	int size;
-
-	dn = pci_device_to_OF_node(pdev);
-	if (!dn) {
-		nv_info(bios, "Unable to get the OF node\n");
-		return;
-	}
-
-	data = of_get_property(dn, "NVDA,BMP", &size);
-	if (data && size) {
-		bios->size = size;
-		bios->data = kmalloc(bios->size, GFP_KERNEL);
-		if (bios->data)
-			memcpy(bios->data, data, size);
-	}
-}
-#endif
-
-static void
-nouveau_bios_shadow_pramin(struct nouveau_bios *bios)
-{
-	struct nouveau_device *device = nv_device(bios);
-	u64 addr = 0;
-	u32 bar0 = 0;
-	int i;
-
-	if (device->card_type >= NV_50) {
-		if (device->card_type >= NV_C0 && device->card_type < GM100) {
-			if (nv_rd32(bios, 0x022500) & 0x00000001)
-				return;
-		} else
-		if (device->card_type >= GM100) {
-			if (nv_rd32(bios, 0x021c04) & 0x00000001)
-				return;
-		}
-
-		addr = nv_rd32(bios, 0x619f04);
-		if (!(addr & 0x00000008)) {
-			nv_debug(bios, "... not enabled\n");
-			return;
+	if (bios->size < length) {
+		u8 *prev = bios->data;
+		if (!(bios->data = kmalloc(length, GFP_KERNEL))) {
+			bios->data = prev;
+			return -ENOMEM;
 		}
-		if ( (addr & 0x00000003) != 1) {
-			nv_debug(bios, "... not in vram\n");
-			return;
-		}
-
-		addr = (addr & 0xffffff00) << 8;
-		if (!addr) {
-			addr  = (u64)nv_rd32(bios, 0x001700) << 16;
-			addr += 0xf0000;
-		}
-
-		bar0 = nv_mask(bios, 0x001700, 0xffffffff, addr >> 16);
-	}
-
-	/* bail if no rom signature */
-	if (nv_rd08(bios, 0x700000) != 0x55 ||
-	    nv_rd08(bios, 0x700001) != 0xaa)
-		goto out;
-
-	bios->size = nv_rd08(bios, 0x700002) * 512;
-	if (!bios->size)
-		goto out;
-
-	bios->data = kmalloc(bios->size, GFP_KERNEL);
-	if (bios->data) {
-		for (i = 0; i < bios->size; i++)
-			nv_wo08(bios, i, nv_rd08(bios, 0x700000 + i));
-	}
-
-out:
-	if (device->card_type >= NV_50)
-		nv_wr32(bios, 0x001700, bar0);
-}
-
-static void
-nouveau_bios_shadow_prom(struct nouveau_bios *bios)
-{
-	struct nouveau_device *device = nv_device(bios);
-	u32 pcireg, access;
-	u16 pcir;
-	int i;
-
-	/* there is no prom on nv4x IGP's */
-	if (device->card_type == NV_40 && device->chipset >= 0x4c)
-		return;
-
-	/* enable access to rom */
-	if (device->card_type >= NV_50)
-		pcireg = 0x088050;
-	else
-		pcireg = 0x001850;
-	access = nv_mask(bios, pcireg, 0x00000001, 0x00000000);
-
-	/* WARNING: PROM accesses should always be 32-bits aligned. Other
-	 * accesses work on most chipset but do not on Kepler chipsets
-	 */
-
-	/* bail if no rom signature, with a workaround for a PROM reading
-	 * issue on some chipsets.  the first read after a period of
-	 * inactivity returns the wrong result, so retry the first header
-	 * byte a few times before giving up as a workaround
-	 */
-	i = 16;
-	do {
-		u32 data = le32_to_cpu(nv_rd32(bios, 0x300000)) & 0xffff;
-		if (data == 0xaa55)
-			break;
-	} while (i--);
-
-	if (!i)
-		goto out;
-
-	/* read entire bios image to system memory */
-	bios->size = (le32_to_cpu(nv_rd32(bios, 0x300000)) >> 16) & 0xff;
-	bios->size = bios->size * 512;
-	if (!bios->size)
-		goto out;
-
-	bios->data = kmalloc(bios->size, GFP_KERNEL);
-	if (!bios->data)
-		goto out;
-
-	for (i = 0; i < bios->size; i += 4)
-		((u32 *)bios->data)[i/4] = nv_rd32(bios, 0x300000 + i);
-
-	/* check the PCI record header */
-	pcir = nv_ro16(bios, 0x0018);
-	if (bios->data[pcir + 0] != 'P' ||
-	    bios->data[pcir + 1] != 'C' ||
-	    bios->data[pcir + 2] != 'I' ||
-	    bios->data[pcir + 3] != 'R') {
-		bios->size = 0;
-		kfree(bios->data);
-	}
-
-out:
-	/* disable access to rom */
-	nv_wr32(bios, pcireg, access);
-}
-
-#if defined(CONFIG_ACPI) && defined(CONFIG_X86)
-int nouveau_acpi_get_bios_chunk(uint8_t *bios, int offset, int len);
-bool nouveau_acpi_rom_supported(struct pci_dev *pdev);
-#else
-static inline bool
-nouveau_acpi_rom_supported(struct pci_dev *pdev) {
-	return false;
-}
-
-static inline int
-nouveau_acpi_get_bios_chunk(uint8_t *bios, int offset, int len) {
-	return -EINVAL;
-}
-#endif
-
-static void
-nouveau_bios_shadow_acpi(struct nouveau_bios *bios)
-{
-	struct pci_dev *pdev = nv_device(bios)->pdev;
-	int ret, cnt, i;
-
-	if (!nouveau_acpi_rom_supported(pdev)) {
-		bios->data = NULL;
-		return;
-	}
-
-	bios->size = 0;
-	bios->data = kmalloc(4096, GFP_KERNEL);
-	if (bios->data) {
-		if (nouveau_acpi_get_bios_chunk(bios->data, 0, 4096) == 4096)
-			bios->size = bios->data[2] * 512;
-		kfree(bios->data);
+		memcpy(bios->data, prev, bios->size);
+		bios->size = length;
+		kfree(prev);
+		return 1;
 	}
-
-	if (!bios->size)
-		return;
-
-	bios->data = kmalloc(bios->size, GFP_KERNEL);
-	if (bios->data) {
-		/* disobey the acpi spec - much faster on at least w530 ... */
-		ret = nouveau_acpi_get_bios_chunk(bios->data, 0, bios->size);
-		if (ret != bios->size ||
-		    nvbios_checksum(bios->data, bios->size)) {
-			/* ... that didn't work, ok, i'll be good now */
-			for (i = 0; i < bios->size; i += cnt) {
-				cnt = min((bios->size - i), (u32)4096);
-				ret = nouveau_acpi_get_bios_chunk(bios->data, i, cnt);
-				if (ret != cnt)
-					break;
-			}
-		}
-	}
-}
-
-static void
-nouveau_bios_shadow_pci(struct nouveau_bios *bios)
-{
-	struct pci_dev *pdev = nv_device(bios)->pdev;
-	size_t size;
-
-	if (!pci_enable_rom(pdev)) {
-		void __iomem *rom = pci_map_rom(pdev, &size);
-		if (rom && size) {
-			bios->data = kmalloc(size, GFP_KERNEL);
-			if (bios->data) {
-				memcpy_fromio(bios->data, rom, size);
-				bios->size = size;
-			}
-		}
-		if (rom)
-			pci_unmap_rom(pdev, rom);
-
-		pci_disable_rom(pdev);
-	}
-}
-
-static void
-nouveau_bios_shadow_platform(struct nouveau_bios *bios)
-{
-	struct pci_dev *pdev = nv_device(bios)->pdev;
-	size_t size;
-
-	void __iomem *rom = pci_platform_rom(pdev, &size);
-	if (rom && size) {
-		bios->data = kmalloc(size, GFP_KERNEL);
-		if (bios->data) {
-			memcpy_fromio(bios->data, rom, size);
-			bios->size = size;
-		}
-	}
-}
-
-static int
-nouveau_bios_score(struct nouveau_bios *bios, const bool writeable)
-{
-	if (bios->size < 3 || !bios->data || bios->data[0] != 0x55 ||
-			bios->data[1] != 0xAA) {
-		nv_info(bios, "... signature not found\n");
-		return 0;
-	}
-
-	if (nvbios_checksum(bios->data,
-			min_t(u32, bios->data[2] * 512, bios->size))) {
-		nv_info(bios, "... checksum invalid\n");
-		/* if a ro image is somewhat bad, it's probably all rubbish */
-		return writeable ? 2 : 1;
-	}
-
-	nv_info(bios, "... appears to be valid\n");
-	return 3;
-}
-
-struct methods {
-	const char desc[16];
-	void (*shadow)(struct nouveau_bios *);
-	const bool rw;
-	int score;
-	u32 size;
-	u8 *data;
-};
-
-static int
-nouveau_bios_shadow(struct nouveau_bios *bios)
-{
-	struct methods shadow_methods[] = {
-#if defined(__powerpc__)
-		{ "OpenFirmware", nouveau_bios_shadow_of, true, 0, 0, NULL },
-#endif
-		{ "PRAMIN", nouveau_bios_shadow_pramin, true, 0, 0, NULL },
-		{ "PROM", nouveau_bios_shadow_prom, false, 0, 0, NULL },
-		{ "ACPI", nouveau_bios_shadow_acpi, true, 0, 0, NULL },
-		{ "PCIROM", nouveau_bios_shadow_pci, true, 0, 0, NULL },
-		{ "PLATFORM", nouveau_bios_shadow_platform, true, 0, 0, NULL },
-		{}
-	};
-	struct methods *mthd, *best;
-	const struct firmware *fw;
-	const char *optarg;
-	int optlen, ret;
-	char *source;
-
-	optarg = nouveau_stropt(nv_device(bios)->cfgopt, "NvBios", &optlen);
-	source = optarg ? kstrndup(optarg, optlen, GFP_KERNEL) : NULL;
-	if (source) {
-		/* try to match one of the built-in methods */
-		mthd = shadow_methods;
-		do {
-			if (strcasecmp(source, mthd->desc))
-				continue;
-			nv_info(bios, "source: %s\n", mthd->desc);
-
-			mthd->shadow(bios);
-			mthd->score = nouveau_bios_score(bios, mthd->rw);
-			if (mthd->score) {
-				kfree(source);
-				return 0;
-			}
-		} while ((++mthd)->shadow);
-
-		/* attempt to load firmware image */
-		ret = request_firmware(&fw, source, &nv_device(bios)->pdev->dev);
-		if (ret == 0) {
-			bios->size = fw->size;
-			bios->data = kmemdup(fw->data, fw->size, GFP_KERNEL);
-			release_firmware(fw);
-
-			nv_info(bios, "image: %s\n", source);
-			if (nouveau_bios_score(bios, 1)) {
-				kfree(source);
-				return 0;
-			}
-
-			kfree(bios->data);
-			bios->data = NULL;
-		}
-
-		nv_error(bios, "source \'%s\' invalid\n", source);
-		kfree(source);
-	}
-
-	mthd = shadow_methods;
-	do {
-		nv_info(bios, "checking %s for image...\n", mthd->desc);
-		mthd->shadow(bios);
-		mthd->score = nouveau_bios_score(bios, mthd->rw);
-		mthd->size = bios->size;
-		mthd->data = bios->data;
-		bios->data = NULL;
-	} while (mthd->score != 3 && (++mthd)->shadow);
-
-	mthd = shadow_methods;
-	best = mthd;
-	do {
-		if (mthd->score > best->score) {
-			kfree(best->data);
-			best = mthd;
-		}
-	} while ((++mthd)->shadow);
-
-	if (best->score) {
-		nv_info(bios, "using image from %s\n", best->desc);
-		bios->size = best->size;
-		bios->data = best->data;
-		return 0;
-	}
-
-	nv_error(bios, "unable to locate usable image\n");
-	return -EINVAL;
+	return 0;
 }
 
 static u8
@@ -472,7 +133,7 @@ nouveau_bios_ctor(struct nouveau_object *parent,
 	if (ret)
 		return ret;
 
-	ret = nouveau_bios_shadow(bios);
+	ret = nvbios_shadow(bios);
 	if (ret)
 		return ret;
 
diff --git a/drivers/gpu/drm/nouveau/core/subdev/bios/dcb.c b/drivers/gpu/drm/nouveau/core/subdev/bios/dcb.c
index 88606bfaf847..96099aff8b41 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/bios/dcb.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/bios/dcb.c
@@ -42,7 +42,7 @@ dcb_table(struct nouveau_bios *bios, u8 *ver, u8 *hdr, u8 *cnt, u8 *len)
 
 	*ver = nv_ro08(bios, dcb);
 
-	if (*ver >= 0x41) {
+	if (*ver >= 0x42) {
 		nv_warn(bios, "DCB version 0x%02x unknown\n", *ver);
 		return 0x0000;
 	} else
@@ -124,6 +124,7 @@ dcb_outp_parse(struct nouveau_bios *bios, u8 idx, u8 *ver, u8 *len,
 	       struct dcb_output *outp)
 {
 	u16 dcb = dcb_outp(bios, idx, ver, len);
+	memset(outp, 0x00, sizeof(*outp));
 	if (dcb) {
 		if (*ver >= 0x20) {
 			u32 conn = nv_ro32(bios, dcb + 0x00);
@@ -156,17 +157,20 @@ dcb_outp_parse(struct nouveau_bios *bios, u8 idx, u8 *ver, u8 *len,
 					break;
 				}
 
-				switch (conf & 0x0f000000) {
-				case 0x0f000000:
-					outp->dpconf.link_nr = 4;
-					break;
-				case 0x03000000:
-					outp->dpconf.link_nr = 2;
-					break;
-				case 0x01000000:
-				default:
-					outp->dpconf.link_nr = 1;
-					break;
+				outp->dpconf.link_nr = (conf & 0x0f000000) >> 24;
+				if (*ver < 0x41) {
+					switch (outp->dpconf.link_nr) {
+					case 0x0f:
+						outp->dpconf.link_nr = 4;
+						break;
+					case 0x03:
+						outp->dpconf.link_nr = 2;
+						break;
+					case 0x01:
+					default:
+						outp->dpconf.link_nr = 1;
+						break;
+					}
 				}
 
 				/* fall-through... */
diff --git a/drivers/gpu/drm/nouveau/core/subdev/bios/disp.c b/drivers/gpu/drm/nouveau/core/subdev/bios/disp.c
index 7f16e52d9bea..51f355599694 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/bios/disp.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/bios/disp.c
@@ -40,6 +40,7 @@ nvbios_disp_table(struct nouveau_bios *bios,
 				switch (*ver) {
 				case 0x20:
 				case 0x21:
+				case 0x22:
 					*hdr = nv_ro08(bios, data + 0x01);
 					*len = nv_ro08(bios, data + 0x02);
 					*cnt = nv_ro08(bios, data + 0x03);
diff --git a/drivers/gpu/drm/nouveau/core/subdev/bios/dp.c b/drivers/gpu/drm/nouveau/core/subdev/bios/dp.c
index f309dd657250..cef53f81f12b 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/bios/dp.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/bios/dp.c
@@ -41,6 +41,7 @@ nvbios_dp_table(struct nouveau_bios *bios, u8 *ver, u8 *hdr, u8 *cnt, u8 *len)
 				case 0x21:
 				case 0x30:
 				case 0x40:
+				case 0x41:
 					*hdr = nv_ro08(bios, data + 0x01);
 					*len = nv_ro08(bios, data + 0x02);
 					*cnt = nv_ro08(bios, data + 0x03);
@@ -70,6 +71,7 @@ nvbios_dpout_entry(struct nouveau_bios *bios, u8 idx,
 			*cnt = nv_ro08(bios, outp + 0x04);
 			break;
 		case 0x40:
+		case 0x41:
 			*hdr = nv_ro08(bios, data + 0x04);
 			*cnt = 0;
 			*len = 0;
@@ -108,6 +110,7 @@ nvbios_dpout_parse(struct nouveau_bios *bios, u8 idx,
 				info->script[4] = nv_ro16(bios, data + 0x10);
 			break;
 		case 0x40:
+		case 0x41:
 			info->flags     = nv_ro08(bios, data + 0x04);
 			info->script[0] = nv_ro16(bios, data + 0x05);
 			info->script[1] = nv_ro16(bios, data + 0x07);
@@ -172,10 +175,11 @@ nvbios_dpcfg_parse(struct nouveau_bios *bios, u16 outp, u8 idx,
 			break;
 		case 0x30:
 		case 0x40:
+		case 0x41:
 			info->pc    = nv_ro08(bios, data + 0x00);
 			info->dc    = nv_ro08(bios, data + 0x01);
 			info->pe    = nv_ro08(bios, data + 0x02);
-			info->tx_pu = nv_ro08(bios, data + 0x03);
+			info->tx_pu = nv_ro08(bios, data + 0x03) & 0x0f;
 			break;
 		default:
 			data = 0x0000;
@@ -194,6 +198,10 @@ nvbios_dpcfg_match(struct nouveau_bios *bios, u16 outp, u8 pc, u8 vs, u8 pe,
 	u16 data;
 
 	if (*ver >= 0x30) {
+		/*XXX: there's a second set of these on at least 4.1, that
+		 *     i've witnessed nvidia using instead of the first
+		 *     on gm204.  figure out what/why
+		 */
 		const u8 vsoff[] = { 0, 4, 7, 9 };
 		idx = (pc * 10) + vsoff[vs] + pe;
 	} else {
diff --git a/drivers/gpu/drm/nouveau/core/subdev/bios/extdev.c b/drivers/gpu/drm/nouveau/core/subdev/bios/extdev.c
index b2a676e53580..49285d4f7ca5 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/bios/extdev.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/bios/extdev.c
@@ -90,7 +90,7 @@ nvbios_extdev_find(struct nouveau_bios *bios, enum nvbios_extdev_type type,
 	u16 entry;
 
 	i = 0;
-	while (!(entry = nvbios_extdev_entry(bios, i++, &ver, &len))) {
+	while ((entry = nvbios_extdev_entry(bios, i++, &ver, &len))) {
 		extdev_parse_entry(bios, entry, func);
 		if (func->type == type)
 			return 0;
diff --git a/drivers/gpu/drm/nouveau/core/subdev/bios/fan.c b/drivers/gpu/drm/nouveau/core/subdev/bios/fan.c
new file mode 100644
index 000000000000..e419892240f5
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/subdev/bios/fan.c
@@ -0,0 +1,93 @@
+/*
+ * Copyright 2014 Martin Peres
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Martin Peres
+ */
+
+#include <subdev/bios.h>
+#include <subdev/bios/bit.h>
+#include <subdev/bios/fan.h>
+
+u16
+nvbios_fan_table(struct nouveau_bios *bios, u8 *ver, u8 *hdr, u8 *cnt, u8 *len)
+{
+	struct bit_entry bit_P;
+	u16 fan = 0x0000;
+
+	if (!bit_entry(bios, 'P', &bit_P)) {
+		if (bit_P.version == 2 && bit_P.length >= 0x5a)
+			fan = nv_ro16(bios, bit_P.offset + 0x58);
+
+		if (fan) {
+			*ver = nv_ro08(bios, fan + 0);
+			switch (*ver) {
+			case 0x10:
+				*hdr = nv_ro08(bios, fan + 1);
+				*len = nv_ro08(bios, fan + 2);
+				*cnt = nv_ro08(bios, fan + 3);
+				return fan;
+			default:
+				break;
+			}
+		}
+	}
+
+	return 0x0000;
+}
+
+u16
+nvbios_fan_entry(struct nouveau_bios *bios, int idx, u8 *ver, u8 *hdr,
+		 u8 *cnt, u8 *len)
+{
+	u16 data = nvbios_fan_table(bios, ver, hdr, cnt, len);
+	if (data && idx < *cnt)
+		return data + *hdr + (idx * (*len));
+	return 0x0000;
+}
+
+u16
+nvbios_fan_parse(struct nouveau_bios *bios, struct nvbios_therm_fan *fan)
+{
+	u8 ver, hdr, cnt, len;
+
+	u16 data = nvbios_fan_entry(bios, 0, &ver, &hdr, &cnt, &len);
+	if (data) {
+		u8 type = nv_ro08(bios, data + 0x00);
+		switch (type) {
+		case 0:
+			fan->type = NVBIOS_THERM_FAN_TOGGLE;
+			break;
+		case 1:
+		case 2:
+			/* TODO: Understand the difference between the two! */
+			fan->type = NVBIOS_THERM_FAN_PWM;
+			break;
+		default:
+			fan->type = NVBIOS_THERM_FAN_UNK;
+		}
+
+		fan->min_duty = nv_ro08(bios, data + 0x02);
+		fan->max_duty = nv_ro08(bios, data + 0x03);
+
+		fan->pwm_freq = nv_ro32(bios, data + 0x0b) & 0xffffff;
+	}
+	return data;
+}
diff --git a/drivers/gpu/drm/nouveau/core/subdev/bios/i2c.c b/drivers/gpu/drm/nouveau/core/subdev/bios/i2c.c
index cfb9288c6d28..282320ba9264 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/bios/i2c.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/bios/i2c.c
@@ -39,6 +39,11 @@ dcb_i2c_table(struct nouveau_bios *bios, u8 *ver, u8 *hdr, u8 *cnt, u8 *len)
 			i2c = nv_ro16(bios, dcb + 4);
 	}
 
+	if (i2c && *ver >= 0x42) {
+		nv_warn(bios, "ccb %02x not supported\n", *ver);
+		return 0x0000;
+	}
+
 	if (i2c && *ver >= 0x30) {
 		*ver = nv_ro08(bios, i2c + 0);
 		*hdr = nv_ro08(bios, i2c + 1);
@@ -70,14 +75,25 @@ dcb_i2c_parse(struct nouveau_bios *bios, u8 idx, struct dcb_i2c_entry *info)
 	u8  ver, len;
 	u16 ent = dcb_i2c_entry(bios, idx, &ver, &len);
 	if (ent) {
-		info->type  = nv_ro08(bios, ent + 3);
-		info->share = DCB_I2C_UNUSED;
-		if (ver < 0x30) {
-			info->type &= 0x07;
+		if (ver >= 0x41) {
+			if (!(nv_ro32(bios, ent) & 0x80000000))
+				info->type = DCB_I2C_UNUSED;
+			else
+				info->type = DCB_I2C_PMGR;
+		} else
+		if (ver >= 0x30) {
+			info->type = nv_ro08(bios, ent + 0x03);
+		} else {
+			info->type = nv_ro08(bios, ent + 0x03) & 0x07;
 			if (info->type == 0x07)
 				info->type = DCB_I2C_UNUSED;
 		}
 
+		info->drive = DCB_I2C_UNUSED;
+		info->sense = DCB_I2C_UNUSED;
+		info->share = DCB_I2C_UNUSED;
+		info->auxch = DCB_I2C_UNUSED;
+
 		switch (info->type) {
 		case DCB_I2C_NV04_BIT:
 			info->drive = nv_ro08(bios, ent + 0);
@@ -87,12 +103,23 @@ dcb_i2c_parse(struct nouveau_bios *bios, u8 idx, struct dcb_i2c_entry *info)
 			info->drive = nv_ro08(bios, ent + 1);
 			return 0;
 		case DCB_I2C_NVIO_BIT:
-		case DCB_I2C_NVIO_AUX:
 			info->drive = nv_ro08(bios, ent + 0) & 0x0f;
-			if (nv_ro08(bios, ent + 1) & 0x01) {
-				info->share  = nv_ro08(bios, ent + 1) >> 1;
-				info->share &= 0x0f;
-			}
+			if (nv_ro08(bios, ent + 1) & 0x01)
+				info->share = nv_ro08(bios, ent + 1) >> 1;
+			return 0;
+		case DCB_I2C_NVIO_AUX:
+			info->auxch = nv_ro08(bios, ent + 0) & 0x0f;
+			if (nv_ro08(bios, ent + 1) & 0x01)
+					info->share = info->auxch;
+			return 0;
+		case DCB_I2C_PMGR:
+			info->drive = (nv_ro16(bios, ent + 0) & 0x01f) >> 0;
+			if (info->drive == 0x1f)
+				info->drive = DCB_I2C_UNUSED;
+			info->auxch = (nv_ro16(bios, ent + 0) & 0x3e0) >> 5;
+			if (info->auxch == 0x1f)
+				info->auxch = DCB_I2C_UNUSED;
+			info->share = info->auxch;
 			return 0;
 		case DCB_I2C_UNUSED:
 			return 0;
diff --git a/drivers/gpu/drm/nouveau/core/subdev/bios/image.c b/drivers/gpu/drm/nouveau/core/subdev/bios/image.c
new file mode 100644
index 000000000000..373f9a564ac9
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/subdev/bios/image.c
@@ -0,0 +1,78 @@
+/*
+ * Copyright 2014 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs <bskeggs@redhat.com>
+ */
+
+#include <subdev/bios.h>
+#include <subdev/bios/image.h>
+#include <subdev/bios/pcir.h>
+#include <subdev/bios/npde.h>
+
+static bool
+nvbios_imagen(struct nouveau_bios *bios, struct nvbios_image *image)
+{
+	struct nvbios_pcirT pcir;
+	struct nvbios_npdeT npde;
+	u8  ver;
+	u16 hdr;
+	u32 data;
+
+	switch ((data = nv_ro16(bios, image->base + 0x00))) {
+	case 0xaa55:
+	case 0xbb77:
+	case 0x4e56: /* NV */
+		break;
+	default:
+		nv_debug(bios, "%08x: ROM signature (%04x) unknown\n",
+			 image->base, data);
+		return false;
+	}
+
+	if (!(data = nvbios_pcirTp(bios, image->base, &ver, &hdr, &pcir)))
+		return false;
+	image->size = pcir.image_size;
+	image->type = pcir.image_type;
+	image->last = pcir.last;
+
+	if (image->type != 0x70) {
+		if (!(data = nvbios_npdeTp(bios, image->base, &npde)))
+			return true;
+		image->size = npde.image_size;
+		image->last = npde.last;
+	} else {
+		image->last = true;
+	}
+
+	return true;
+}
+
+bool
+nvbios_image(struct nouveau_bios *bios, int idx, struct nvbios_image *image)
+{
+	memset(image, 0x00, sizeof(*image));
+	do {
+		image->base += image->size;
+		if (image->last || !nvbios_imagen(bios, image))
+			return false;
+	} while(idx--);
+	return true;
+}
diff --git a/drivers/gpu/drm/nouveau/core/subdev/bios/init.c b/drivers/gpu/drm/nouveau/core/subdev/bios/init.c
index 626380f9e4c0..c6579ef32cd1 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/bios/init.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/bios/init.c
@@ -255,6 +255,8 @@ init_i2c(struct nvbios_init *init, int index)
 		}
 
 		index = init->outp->i2c_index;
+		if (init->outp->type == DCB_OUTPUT_DP)
+			index += NV_I2C_AUX(0);
 	}
 
 	return i2c->find(i2c, index);
@@ -278,7 +280,7 @@ init_wri2cr(struct nvbios_init *init, u8 index, u8 addr, u8 reg, u8 val)
 	return -ENODEV;
 }
 
-static int
+static u8
 init_rdauxr(struct nvbios_init *init, u32 addr)
 {
 	struct nouveau_i2c_port *port = init_i2c(init, -2);
@@ -286,20 +288,24 @@ init_rdauxr(struct nvbios_init *init, u32 addr)
 
 	if (port && init_exec(init)) {
 		int ret = nv_rdaux(port, addr, &data, 1);
-		if (ret)
-			return ret;
-		return data;
+		if (ret == 0)
+			return data;
+		trace("auxch read failed with %d\n", ret);
 	}
 
-	return -ENODEV;
+	return 0x00;
 }
 
 static int
 init_wrauxr(struct nvbios_init *init, u32 addr, u8 data)
 {
 	struct nouveau_i2c_port *port = init_i2c(init, -2);
-	if (port && init_exec(init))
-		return nv_wraux(port, addr, &data, 1);
+	if (port && init_exec(init)) {
+		int ret = nv_wraux(port, addr, &data, 1);
+		if (ret)
+			trace("auxch write failed with %d\n", ret);
+		return ret;
+	}
 	return -ENODEV;
 }
 
@@ -838,6 +844,40 @@ init_io_or(struct nvbios_init *init)
 }
 
 /**
+ * INIT_ANDN_REG - opcode 0x47
+ *
+ */
+static void
+init_andn_reg(struct nvbios_init *init)
+{
+	struct nouveau_bios *bios = init->bios;
+	u32  reg = nv_ro32(bios, init->offset + 1);
+	u32 mask = nv_ro32(bios, init->offset + 5);
+
+	trace("ANDN_REG\tR[0x%06x] &= ~0x%08x\n", reg, mask);
+	init->offset += 9;
+
+	init_mask(init, reg, mask, 0);
+}
+
+/**
+ * INIT_OR_REG - opcode 0x48
+ *
+ */
+static void
+init_or_reg(struct nvbios_init *init)
+{
+	struct nouveau_bios *bios = init->bios;
+	u32  reg = nv_ro32(bios, init->offset + 1);
+	u32 mask = nv_ro32(bios, init->offset + 5);
+
+	trace("OR_REG\tR[0x%06x] |= 0x%08x\n", reg, mask);
+	init->offset += 9;
+
+	init_mask(init, reg, 0, mask);
+}
+
+/**
  * INIT_INDEX_ADDRESS_LATCHED - opcode 0x49
  *
  */
@@ -2068,6 +2108,8 @@ static struct nvbios_init_opcode {
 	[0x3a] = { init_dp_condition },
 	[0x3b] = { init_io_mask_or },
 	[0x3c] = { init_io_or },
+	[0x47] = { init_andn_reg },
+	[0x48] = { init_or_reg },
 	[0x49] = { init_idx_addr_latched },
 	[0x4a] = { init_io_restrict_pll2 },
 	[0x4b] = { init_pll2 },
diff --git a/drivers/gpu/drm/nouveau/core/subdev/bios/npde.c b/drivers/gpu/drm/nouveau/core/subdev/bios/npde.c
new file mode 100644
index 000000000000..d694716a166c
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/subdev/bios/npde.c
@@ -0,0 +1,59 @@
+/*
+ * Copyright 2014 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs <bskeggs@redhat.com>
+ */
+
+#include <subdev/bios.h>
+#include <subdev/bios/npde.h>
+#include <subdev/bios/pcir.h>
+
+u32
+nvbios_npdeTe(struct nouveau_bios *bios, u32 base)
+{
+	struct nvbios_pcirT pcir;
+	u8  ver; u16 hdr;
+	u32 data = nvbios_pcirTp(bios, base, &ver, &hdr, &pcir);
+	if (data = (data + hdr + 0x0f) & ~0x0f, data) {
+		switch (nv_ro32(bios, data + 0x00)) {
+		case 0x4544504e: /* NPDE */
+			break;
+		default:
+			nv_debug(bios, "%08x: NPDE signature (%08x) unknown\n",
+				 data, nv_ro32(bios, data + 0x00));
+			data = 0;
+			break;
+		}
+	}
+	return data;
+}
+
+u32
+nvbios_npdeTp(struct nouveau_bios *bios, u32 base, struct nvbios_npdeT *info)
+{
+	u32 data = nvbios_npdeTe(bios, base);
+	memset(info, 0x00, sizeof(*info));
+	if (data) {
+		info->image_size = nv_ro16(bios, data + 0x08) * 512;
+		info->last = nv_ro08(bios, data + 0x0a) & 0x80;
+	}
+	return data;
+}
diff --git a/drivers/gpu/drm/nouveau/core/subdev/bios/pcir.c b/drivers/gpu/drm/nouveau/core/subdev/bios/pcir.c
new file mode 100644
index 000000000000..91dae26bc50f
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/subdev/bios/pcir.c
@@ -0,0 +1,69 @@
+/*
+ * Copyright 2014 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs <bskeggs@redhat.com>
+ */
+
+#include <subdev/bios.h>
+#include <subdev/bios/pcir.h>
+
+u32
+nvbios_pcirTe(struct nouveau_bios *bios, u32 base, u8 *ver, u16 *hdr)
+{
+	u32 data = nv_ro16(bios, base + 0x18);
+	if (data) {
+		data += base;
+		switch (nv_ro32(bios, data + 0x00)) {
+		case 0x52494350: /* PCIR */
+		case 0x53494752: /* RGIS */
+		case 0x5344504e: /* NPDS */
+			*hdr = nv_ro16(bios, data + 0x0a);
+			*ver = nv_ro08(bios, data + 0x0c);
+			break;
+		default:
+			nv_debug(bios, "%08x: PCIR signature (%08x) unknown\n",
+				 data, nv_ro32(bios, data + 0x00));
+			data = 0;
+			break;
+		}
+	}
+	return data;
+}
+
+u32
+nvbios_pcirTp(struct nouveau_bios *bios, u32 base, u8 *ver, u16 *hdr,
+	      struct nvbios_pcirT *info)
+{
+	u32 data = nvbios_pcirTe(bios, base, ver, hdr);
+	memset(info, 0x00, sizeof(*info));
+	if (data) {
+		info->vendor_id = nv_ro16(bios, data + 0x04);
+		info->device_id = nv_ro16(bios, data + 0x06);
+		info->class_code[0] = nv_ro08(bios, data + 0x0d);
+		info->class_code[1] = nv_ro08(bios, data + 0x0e);
+		info->class_code[2] = nv_ro08(bios, data + 0x0f);
+		info->image_size = nv_ro16(bios, data + 0x10) * 512;
+		info->image_rev = nv_ro16(bios, data + 0x12);
+		info->image_type = nv_ro08(bios, data + 0x14);
+		info->last = nv_ro08(bios, data + 0x15) & 0x80;
+	}
+	return data;
+}
diff --git a/drivers/gpu/drm/nouveau/core/subdev/bios/pmu.c b/drivers/gpu/drm/nouveau/core/subdev/bios/pmu.c
new file mode 100644
index 000000000000..66c56ba07d1b
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/subdev/bios/pmu.c
@@ -0,0 +1,135 @@
+/*
+ * Copyright 2014 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs <bskeggs@redhat.com>
+ */
+
+#include <subdev/bios.h>
+#include <subdev/bios/bit.h>
+#include <subdev/bios/image.h>
+#include <subdev/bios/pmu.h>
+
+static u32
+weirdo_pointer(struct nouveau_bios *bios, u32 data)
+{
+	struct nvbios_image image;
+	int idx = 0;
+	if (nvbios_image(bios, idx++, &image)) {
+		data -= image.size;
+		while (nvbios_image(bios, idx++, &image)) {
+			if (image.type == 0xe0)
+				return image.base + data;
+		}
+	}
+	return 0;
+}
+
+u32
+nvbios_pmuTe(struct nouveau_bios *bios, u8 *ver, u8 *hdr, u8 *cnt, u8 *len)
+{
+	struct bit_entry bit_p;
+	u32 data = 0;
+
+	if (!bit_entry(bios, 'p', &bit_p)) {
+		if (bit_p.version == 2 && bit_p.length >= 4)
+			data = nv_ro32(bios, bit_p.offset + 0x00);
+		if ((data = weirdo_pointer(bios, data))) {
+			*ver = nv_ro08(bios, data + 0x00); /* maybe? */
+			*hdr = nv_ro08(bios, data + 0x01);
+			*len = nv_ro08(bios, data + 0x02);
+			*cnt = nv_ro08(bios, data + 0x03);
+		}
+	}
+
+	return data;
+}
+
+u32
+nvbios_pmuTp(struct nouveau_bios *bios, u8 *ver, u8 *hdr, u8 *cnt, u8 *len,
+	     struct nvbios_pmuT *info)
+{
+	u32 data = nvbios_pmuTe(bios, ver, hdr, cnt, len);
+	memset(info, 0x00, sizeof(*info));
+	switch (!!data * *ver) {
+	default:
+		break;
+	}
+	return data;
+}
+
+u32
+nvbios_pmuEe(struct nouveau_bios *bios, int idx, u8 *ver, u8 *hdr)
+{
+	u8  cnt, len;
+	u32 data = nvbios_pmuTe(bios, ver, hdr, &cnt, &len);
+	if (data && idx < cnt) {
+		data = data + *hdr + (idx * len);
+		*hdr = len;
+		return data;
+	}
+	return 0;
+}
+
+u32
+nvbios_pmuEp(struct nouveau_bios *bios, int idx, u8 *ver, u8 *hdr,
+	     struct nvbios_pmuE *info)
+{
+	u32 data = nvbios_pmuEe(bios, idx, ver, hdr);
+	memset(info, 0x00, sizeof(*info));
+	switch (!!data * *ver) {
+	default:
+		info->type = nv_ro08(bios, data + 0x00);
+		info->data = nv_ro32(bios, data + 0x02);
+		break;
+	}
+	return data;
+}
+
+bool
+nvbios_pmuRm(struct nouveau_bios *bios, u8 type, struct nvbios_pmuR *info)
+{
+	struct nvbios_pmuE pmuE;
+	u8  ver, hdr, idx = 0;
+	u32 data;
+	memset(info, 0x00, sizeof(*info));
+	while ((data = nvbios_pmuEp(bios, idx++, &ver, &hdr, &pmuE))) {
+		if ( pmuE.type == type &&
+		    (data = weirdo_pointer(bios, pmuE.data))) {
+			info->init_addr_pmu = nv_ro32(bios, data + 0x08);
+			info->args_addr_pmu = nv_ro32(bios, data + 0x0c);
+			info->boot_addr     = data + 0x30;
+			info->boot_addr_pmu = nv_ro32(bios, data + 0x10) +
+					      nv_ro32(bios, data + 0x18);
+			info->boot_size     = nv_ro32(bios, data + 0x1c) -
+					      nv_ro32(bios, data + 0x18);
+			info->code_addr     = info->boot_addr + info->boot_size;
+			info->code_addr_pmu = info->boot_addr_pmu +
+					      info->boot_size;
+			info->code_size     = nv_ro32(bios, data + 0x20);
+			info->data_addr     = data + 0x30 +
+					      nv_ro32(bios, data + 0x24);
+			info->data_addr_pmu = nv_ro32(bios, data + 0x28);
+			info->data_size     = nv_ro32(bios, data + 0x2c);
+			return true;
+		}
+	}
+	return false;
+}
diff --git a/drivers/gpu/drm/nouveau/core/subdev/bios/priv.h b/drivers/gpu/drm/nouveau/core/subdev/bios/priv.h
new file mode 100644
index 000000000000..187d225bd1e9
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/subdev/bios/priv.h
@@ -0,0 +1,25 @@
+#ifndef __NVKM_BIOS_PRIV_H__
+#define __NVKM_BIOS_PRIV_H__
+
+#include <subdev/bios.h>
+
+struct nvbios_source {
+	const char *name;
+	void *(*init)(struct nouveau_bios *, const char *);
+	void  (*fini)(void *);
+	u32   (*read)(void *, u32 offset, u32 length, struct nouveau_bios *);
+	bool rw;
+};
+
+int nvbios_extend(struct nouveau_bios *, u32 length);
+int nvbios_shadow(struct nouveau_bios *);
+
+extern const struct nvbios_source nvbios_rom;
+extern const struct nvbios_source nvbios_ramin;
+extern const struct nvbios_source nvbios_acpi_fast;
+extern const struct nvbios_source nvbios_acpi_slow;
+extern const struct nvbios_source nvbios_pcirom;
+extern const struct nvbios_source nvbios_platform;
+extern const struct nvbios_source nvbios_of;
+
+#endif
diff --git a/drivers/gpu/drm/nouveau/core/subdev/bios/ramcfg.c b/drivers/gpu/drm/nouveau/core/subdev/bios/ramcfg.c
index 6c401f70ab99..1623c8dfe797 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/bios/ramcfg.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/bios/ramcfg.c
@@ -25,6 +25,7 @@
 #include <subdev/bios.h>
 #include <subdev/bios/bit.h>
 #include <subdev/bios/ramcfg.h>
+#include <subdev/bios/M0203.h>
 
 static u8
 nvbios_ramcfg_strap(struct nouveau_subdev *subdev)
@@ -54,12 +55,22 @@ nvbios_ramcfg_index(struct nouveau_subdev *subdev)
 	u8 strap = nvbios_ramcfg_strap(subdev);
 	u32 xlat = 0x00000000;
 	struct bit_entry bit_M;
+	struct nvbios_M0203E M0203E;
+	u8 ver, hdr;
 
 	if (!bit_entry(bios, 'M', &bit_M)) {
 		if (bit_M.version == 1 && bit_M.length >= 5)
 			xlat = nv_ro16(bios, bit_M.offset + 3);
-		if (bit_M.version == 2 && bit_M.length >= 3)
+		if (bit_M.version == 2 && bit_M.length >= 3) {
+			/*XXX: is M ever shorter than this?
+			 *     if not - what is xlat used for now?
+			 *     also - sigh..
+			 */
+			if (bit_M.length >= 7 &&
+			    nvbios_M0203Em(bios, strap, &ver, &hdr, &M0203E))
+				return M0203E.group;
 			xlat = nv_ro16(bios, bit_M.offset + 1);
+		}
 	}
 
 	if (xlat)
diff --git a/drivers/gpu/drm/nouveau/core/subdev/bios/rammap.c b/drivers/gpu/drm/nouveau/core/subdev/bios/rammap.c
index 1811b2cb0472..c5685228c322 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/bios/rammap.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/bios/rammap.c
@@ -75,31 +75,39 @@ nvbios_rammapEe(struct nouveau_bios *bios, int idx,
 }
 
 u32
-nvbios_rammapEm(struct nouveau_bios *bios, u16 khz,
-		u8 *ver, u8 *hdr, u8 *cnt, u8 *len)
-{
-	int idx = 0;
-	u32 data;
-	while ((data = nvbios_rammapEe(bios, idx++, ver, hdr, cnt, len))) {
-		if (khz >= nv_ro16(bios, data + 0x00) &&
-		    khz <= nv_ro16(bios, data + 0x02))
-			break;
-	}
-	return data;
-}
-
-u32
-nvbios_rammapEp(struct nouveau_bios *bios, u16 khz,
+nvbios_rammapEp(struct nouveau_bios *bios, int idx,
 		u8 *ver, u8 *hdr, u8 *cnt, u8 *len,
 		struct nvbios_ramcfg *p)
 {
-	u32 data = nvbios_rammapEm(bios, khz, ver, hdr, cnt, len);
+	u32 data = nvbios_rammapEe(bios, idx, ver, hdr, cnt, len), temp;
 	memset(p, 0x00, sizeof(*p));
+	p->rammap_ver = *ver;
+	p->rammap_hdr = *hdr;
 	switch (!!data * *ver) {
+	case 0x10:
+		p->rammap_min      =  nv_ro16(bios, data + 0x00);
+		p->rammap_max      =  nv_ro16(bios, data + 0x02);
+		p->rammap_10_04_02 = (nv_ro08(bios, data + 0x04) & 0x02) >> 1;
+		p->rammap_10_04_08 = (nv_ro08(bios, data + 0x04) & 0x08) >> 3;
+		break;
 	case 0x11:
+		p->rammap_min      =  nv_ro16(bios, data + 0x00);
+		p->rammap_max      =  nv_ro16(bios, data + 0x02);
 		p->rammap_11_08_01 = (nv_ro08(bios, data + 0x08) & 0x01) >> 0;
 		p->rammap_11_08_0c = (nv_ro08(bios, data + 0x08) & 0x0c) >> 2;
 		p->rammap_11_08_10 = (nv_ro08(bios, data + 0x08) & 0x10) >> 4;
+		temp = nv_ro32(bios, data + 0x09);
+		p->rammap_11_09_01ff = (temp & 0x000001ff) >> 0;
+		p->rammap_11_0a_03fe = (temp & 0x0003fe00) >> 9;
+		p->rammap_11_0a_0400 = (temp & 0x00040000) >> 18;
+		p->rammap_11_0a_0800 = (temp & 0x00080000) >> 19;
+		p->rammap_11_0b_01f0 = (temp & 0x01f00000) >> 20;
+		p->rammap_11_0b_0200 = (temp & 0x02000000) >> 25;
+		p->rammap_11_0b_0400 = (temp & 0x04000000) >> 26;
+		p->rammap_11_0b_0800 = (temp & 0x08000000) >> 27;
+		p->rammap_11_0d    =  nv_ro08(bios, data + 0x0d);
+		p->rammap_11_0e    =  nv_ro08(bios, data + 0x0e);
+		p->rammap_11_0f    =  nv_ro08(bios, data + 0x0f);
 		p->rammap_11_11_0c = (nv_ro08(bios, data + 0x11) & 0x0c) >> 2;
 		break;
 	default:
@@ -110,6 +118,20 @@ nvbios_rammapEp(struct nouveau_bios *bios, u16 khz,
 }
 
 u32
+nvbios_rammapEm(struct nouveau_bios *bios, u16 mhz,
+		u8 *ver, u8 *hdr, u8 *cnt, u8 *len,
+		struct nvbios_ramcfg *info)
+{
+	int idx = 0;
+	u32 data;
+	while ((data = nvbios_rammapEp(bios, idx++, ver, hdr, cnt, len, info))) {
+		if (mhz >= info->rammap_min && mhz <= info->rammap_max)
+			break;
+	}
+	return data;
+}
+
+u32
 nvbios_rammapSe(struct nouveau_bios *bios, u32 data,
 		u8 ever, u8 ehdr, u8 ecnt, u8 elen, int idx,
 		u8 *ver, u8 *hdr)
@@ -129,8 +151,29 @@ nvbios_rammapSp(struct nouveau_bios *bios, u32 data,
 		u8 *ver, u8 *hdr, struct nvbios_ramcfg *p)
 {
 	data = nvbios_rammapSe(bios, data, ever, ehdr, ecnt, elen, idx, ver, hdr);
+	p->ramcfg_ver = *ver;
+	p->ramcfg_hdr = *hdr;
 	switch (!!data * *ver) {
+	case 0x10:
+		p->ramcfg_timing   =  nv_ro08(bios, data + 0x01);
+		p->ramcfg_10_02_01 = (nv_ro08(bios, data + 0x02) & 0x01) >> 0;
+		p->ramcfg_10_02_02 = (nv_ro08(bios, data + 0x02) & 0x02) >> 1;
+		p->ramcfg_10_02_04 = (nv_ro08(bios, data + 0x02) & 0x04) >> 2;
+		p->ramcfg_10_02_08 = (nv_ro08(bios, data + 0x02) & 0x08) >> 3;
+		p->ramcfg_10_02_10 = (nv_ro08(bios, data + 0x02) & 0x10) >> 4;
+		p->ramcfg_10_02_20 = (nv_ro08(bios, data + 0x02) & 0x20) >> 5;
+		p->ramcfg_10_DLLoff = (nv_ro08(bios, data + 0x02) & 0x40) >> 6;
+		p->ramcfg_10_03_0f = (nv_ro08(bios, data + 0x03) & 0x0f) >> 0;
+		p->ramcfg_10_04_01 = (nv_ro08(bios, data + 0x04) & 0x01) >> 0;
+		p->ramcfg_10_05    = (nv_ro08(bios, data + 0x05) & 0xff) >> 0;
+		p->ramcfg_10_06    = (nv_ro08(bios, data + 0x06) & 0xff) >> 0;
+		p->ramcfg_10_07    = (nv_ro08(bios, data + 0x07) & 0xff) >> 0;
+		p->ramcfg_10_08    = (nv_ro08(bios, data + 0x08) & 0xff) >> 0;
+		p->ramcfg_10_09_0f = (nv_ro08(bios, data + 0x09) & 0x0f) >> 0;
+		p->ramcfg_10_09_f0 = (nv_ro08(bios, data + 0x09) & 0xf0) >> 4;
+		break;
 	case 0x11:
+		p->ramcfg_timing   =  nv_ro08(bios, data + 0x00);
 		p->ramcfg_11_01_01 = (nv_ro08(bios, data + 0x01) & 0x01) >> 0;
 		p->ramcfg_11_01_02 = (nv_ro08(bios, data + 0x01) & 0x02) >> 1;
 		p->ramcfg_11_01_04 = (nv_ro08(bios, data + 0x01) & 0x04) >> 2;
diff --git a/drivers/gpu/drm/nouveau/core/subdev/bios/shadow.c b/drivers/gpu/drm/nouveau/core/subdev/bios/shadow.c
new file mode 100644
index 000000000000..bb9e0018d936
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/subdev/bios/shadow.c
@@ -0,0 +1,270 @@
+/*
+ * Copyright 2014 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs <bskeggs@redhat.com>
+ */
+
+#include "priv.h"
+#include <core/option.h>
+#include <subdev/bios/image.h>
+
+struct shadow {
+	struct nouveau_oclass base;
+	u32 skip;
+	const struct nvbios_source *func;
+	void *data;
+	u32 size;
+	int score;
+};
+
+static bool
+shadow_fetch(struct nouveau_bios *bios, u32 upto)
+{
+	struct shadow *mthd = (void *)nv_object(bios)->oclass;
+	const u32 limit = (upto + 3) & ~3;
+	const u32 start = bios->size;
+	void *data = mthd->data;
+	if (nvbios_extend(bios, limit) > 0) {
+		u32 read = mthd->func->read(data, start, limit - start, bios);
+		bios->size = start + read;
+	}
+	return bios->size >= limit;
+}
+
+static u8
+shadow_rd08(struct nouveau_object *object, u64 addr)
+{
+	struct nouveau_bios *bios = (void *)object;
+	if (shadow_fetch(bios, addr + 1))
+		return bios->data[addr];
+	return 0x00;
+}
+
+static u16
+shadow_rd16(struct nouveau_object *object, u64 addr)
+{
+	struct nouveau_bios *bios = (void *)object;
+	if (shadow_fetch(bios, addr + 2))
+		return get_unaligned_le16(&bios->data[addr]);
+	return 0x0000;
+}
+
+static u32
+shadow_rd32(struct nouveau_object *object, u64 addr)
+{
+	struct nouveau_bios *bios = (void *)object;
+	if (shadow_fetch(bios, addr + 4))
+		return get_unaligned_le32(&bios->data[addr]);
+	return 0x00000000;
+}
+
+static struct nouveau_oclass
+shadow_class = {
+	.handle = NV_SUBDEV(VBIOS, 0x00),
+	.ofuncs = &(struct nouveau_ofuncs) {
+		.rd08 = shadow_rd08,
+		.rd16 = shadow_rd16,
+		.rd32 = shadow_rd32,
+	},
+};
+
+static int
+shadow_image(struct nouveau_bios *bios, int idx, struct shadow *mthd)
+{
+	struct nvbios_image image;
+	int score = 1;
+
+	if (!nvbios_image(bios, idx, &image)) {
+		nv_debug(bios, "image %d invalid\n", idx);
+		return 0;
+	}
+	nv_debug(bios, "%08x: type %02x, %d bytes\n",
+		 image.base, image.type, image.size);
+
+	if (!shadow_fetch(bios, image.size)) {
+		nv_debug(bios, "%08x: fetch failed\n", image.base);
+		return 0;
+	}
+
+	switch (image.type) {
+	case 0x00:
+		if (nvbios_checksum(&bios->data[image.base], image.size)) {
+			nv_debug(bios, "%08x: checksum failed\n", image.base);
+			if (mthd->func->rw)
+				score += 1;
+			score += 1;
+		} else {
+			score += 3;
+		}
+		break;
+	default:
+		score += 3;
+		break;
+	}
+
+	if (!image.last)
+		score += shadow_image(bios, idx + 1, mthd);
+	return score;
+}
+
+static int
+shadow_score(struct nouveau_bios *bios, struct shadow *mthd)
+{
+	struct nouveau_oclass *oclass = nv_object(bios)->oclass;
+	int score;
+	nv_object(bios)->oclass = &mthd->base;
+	score = shadow_image(bios, 0, mthd);
+	nv_object(bios)->oclass = oclass;
+	return score;
+
+}
+
+static int
+shadow_method(struct nouveau_bios *bios, struct shadow *mthd, const char *name)
+{
+	const struct nvbios_source *func = mthd->func;
+	if (func->name) {
+		nv_debug(bios, "trying %s...\n", name ? name : func->name);
+		if (func->init) {
+			mthd->data = func->init(bios, name);
+			if (IS_ERR(mthd->data)) {
+				mthd->data = NULL;
+				return 0;
+			}
+		}
+		mthd->score = shadow_score(bios, mthd);
+		if (func->fini)
+			func->fini(mthd->data);
+		nv_debug(bios, "scored %d\n", mthd->score);
+		mthd->data = bios->data;
+		mthd->size = bios->size;
+		bios->data  = NULL;
+		bios->size  = 0;
+	}
+	return mthd->score;
+}
+
+static u32
+shadow_fw_read(void *data, u32 offset, u32 length, struct nouveau_bios *bios)
+{
+	const struct firmware *fw = data;
+	if (offset + length <= fw->size) {
+		memcpy(bios->data + offset, fw->data + offset, length);
+		return length;
+	}
+	return 0;
+}
+
+static void *
+shadow_fw_init(struct nouveau_bios *bios, const char *name)
+{
+	struct device *dev = &nv_device(bios)->pdev->dev;
+	const struct firmware *fw;
+	int ret = request_firmware(&fw, name, dev);
+	if (ret)
+		return ERR_PTR(-ENOENT);
+	return (void *)fw;
+}
+
+static const struct nvbios_source
+shadow_fw = {
+	.name = "firmware",
+	.init = shadow_fw_init,
+	.fini = (void(*)(void *))release_firmware,
+	.read = shadow_fw_read,
+	.rw = false,
+};
+
+int
+nvbios_shadow(struct nouveau_bios *bios)
+{
+	struct shadow mthds[] = {
+		{ shadow_class, 0, &nvbios_of },
+		{ shadow_class, 0, &nvbios_ramin },
+		{ shadow_class, 0, &nvbios_rom },
+		{ shadow_class, 0, &nvbios_acpi_fast },
+		{ shadow_class, 4, &nvbios_acpi_slow },
+		{ shadow_class, 1, &nvbios_pcirom },
+		{ shadow_class, 1, &nvbios_platform },
+		{ shadow_class }
+	}, *mthd = mthds, *best = NULL;
+	const char *optarg;
+	char *source;
+	int optlen;
+
+	/* handle user-specified bios source */
+	optarg = nouveau_stropt(nv_device(bios)->cfgopt, "NvBios", &optlen);
+	source = optarg ? kstrndup(optarg, optlen, GFP_KERNEL) : NULL;
+	if (source) {
+		/* try to match one of the built-in methods */
+		for (mthd = mthds; mthd->func; mthd++) {
+			if (mthd->func->name &&
+			    !strcasecmp(source, mthd->func->name)) {
+				best = mthd;
+				if (shadow_method(bios, mthd, NULL))
+					break;
+			}
+		}
+
+		/* otherwise, attempt to load as firmware */
+		if (!best && (best = mthd)) {
+			mthd->func = &shadow_fw;
+			shadow_method(bios, mthd, source);
+			mthd->func = NULL;
+		}
+
+		if (!best->score) {
+			nv_error(bios, "%s invalid\n", source);
+			kfree(source);
+			source = NULL;
+		}
+	}
+
+	/* scan all potential bios sources, looking for best image */
+	if (!best || !best->score) {
+		for (mthd = mthds, best = mthd; mthd->func; mthd++) {
+			if (!mthd->skip || best->score < mthd->skip) {
+				if (shadow_method(bios, mthd, NULL)) {
+					if (mthd->score > best->score)
+						best = mthd;
+				}
+			}
+		}
+	}
+
+	/* cleanup the ones we didn't use */
+	for (mthd = mthds; mthd->func; mthd++) {
+		if (mthd != best)
+			kfree(mthd->data);
+	}
+
+	if (!best->score) {
+		nv_fatal(bios, "unable to locate usable image\n");
+		return -EINVAL;
+	}
+
+	nv_info(bios, "using image from %s\n", best->func ?
+		best->func->name : source);
+	bios->data = best->data;
+	bios->size = best->size;
+	kfree(source);
+	return 0;
+}
diff --git a/drivers/gpu/drm/nouveau/core/subdev/bios/shadowacpi.c b/drivers/gpu/drm/nouveau/core/subdev/bios/shadowacpi.c
new file mode 100644
index 000000000000..bc130c12ec06
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/subdev/bios/shadowacpi.c
@@ -0,0 +1,111 @@
+/*
+ * Copyright 2012 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#include "priv.h"
+
+#if defined(CONFIG_ACPI) && defined(CONFIG_X86)
+int nouveau_acpi_get_bios_chunk(uint8_t *bios, int offset, int len);
+bool nouveau_acpi_rom_supported(struct pci_dev *pdev);
+#else
+static inline bool
+nouveau_acpi_rom_supported(struct pci_dev *pdev)
+{
+	return false;
+}
+
+static inline int
+nouveau_acpi_get_bios_chunk(uint8_t *bios, int offset, int len)
+{
+	return -EINVAL;
+}
+#endif
+
+/* This version of the shadow function disobeys the ACPI spec and tries
+ * to fetch in units of more than 4KiB at a time.  This is a LOT faster
+ * on some systems, such as Lenovo W530.
+ */
+static u32
+acpi_read_fast(void *data, u32 offset, u32 length, struct nouveau_bios *bios)
+{
+	u32 limit = (offset + length + 0xfff) & ~0xfff;
+	u32 start = offset & ~0x00000fff;
+	u32 fetch = limit - start;
+
+	if (nvbios_extend(bios, limit) > 0) {
+		int ret = nouveau_acpi_get_bios_chunk(bios->data, start, fetch);
+		if (ret == fetch)
+			return fetch;
+	}
+
+	return 0;
+}
+
+/* Other systems, such as the one in fdo#55948, will report a success
+ * but only return 4KiB of data.  The common bios fetching logic will
+ * detect an invalid image, and fall back to this version of the read
+ * function.
+ */
+static u32
+acpi_read_slow(void *data, u32 offset, u32 length, struct nouveau_bios *bios)
+{
+	u32 limit = (offset + length + 0xfff) & ~0xfff;
+	u32 start = offset & ~0xfff;
+	u32 fetch = 0;
+
+	if (nvbios_extend(bios, limit) > 0) {
+		while (start + fetch < limit) {
+			int ret = nouveau_acpi_get_bios_chunk(bios->data,
+							      start + fetch,
+							      0x1000);
+			if (ret != 0x1000)
+				break;
+			fetch += 0x1000;
+		}
+	}
+
+	return fetch;
+}
+
+static void *
+acpi_init(struct nouveau_bios *bios, const char *name)
+{
+	if (!nouveau_acpi_rom_supported(nv_device(bios)->pdev))
+		return ERR_PTR(-ENODEV);
+	return NULL;
+}
+
+const struct nvbios_source
+nvbios_acpi_fast = {
+	.name = "ACPI",
+	.init = acpi_init,
+	.read = acpi_read_fast,
+	.rw = false,
+};
+
+const struct nvbios_source
+nvbios_acpi_slow = {
+	.name = "ACPI",
+	.init = acpi_init,
+	.read = acpi_read_slow,
+	.rw = false,
+};
diff --git a/drivers/gpu/drm/nouveau/core/subdev/bios/shadowof.c b/drivers/gpu/drm/nouveau/core/subdev/bios/shadowof.c
new file mode 100644
index 000000000000..3abe487a6025
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/subdev/bios/shadowof.c
@@ -0,0 +1,71 @@
+/*
+ * Copyright 2012 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#include "priv.h"
+
+#if defined(__powerpc__)
+struct priv {
+	const void __iomem *data;
+	int size;
+};
+
+static u32
+of_read(void *data, u32 offset, u32 length, struct nouveau_bios *bios)
+{
+	struct priv *priv = data;
+	if (offset + length <= priv->size) {
+		memcpy_fromio(bios->data + offset, priv->data + offset, length);
+		return length;
+	}
+	return 0;
+}
+
+static void *
+of_init(struct nouveau_bios *bios, const char *name)
+{
+	struct pci_dev *pdev = nv_device(bios)->pdev;
+	struct device_node *dn;
+	struct priv *priv;
+	if (!(dn = pci_device_to_OF_node(pdev)))
+		return ERR_PTR(-ENODEV);
+	if (!(priv = kzalloc(sizeof(*priv), GFP_KERNEL)))
+		return ERR_PTR(-ENOMEM);
+	if ((priv->data = of_get_property(dn, "NVDA,BMP", &priv->size)))
+		return priv;
+	kfree(priv);
+	return ERR_PTR(-EINVAL);
+}
+
+const struct nvbios_source
+nvbios_of = {
+	.name = "OpenFirmware",
+	.init = of_init,
+	.fini = (void(*)(void *))kfree,
+	.read = of_read,
+	.rw = false,
+};
+#else
+const struct nvbios_source
+nvbios_of = {
+};
+#endif
diff --git a/drivers/gpu/drm/nouveau/core/subdev/bios/shadowpci.c b/drivers/gpu/drm/nouveau/core/subdev/bios/shadowpci.c
new file mode 100644
index 000000000000..1d0389c0abef
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/subdev/bios/shadowpci.c
@@ -0,0 +1,108 @@
+/*
+ * Copyright 2012 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#include "priv.h"
+
+struct priv {
+	struct pci_dev *pdev;
+	void __iomem *rom;
+	size_t size;
+};
+
+static u32
+pcirom_read(void *data, u32 offset, u32 length, struct nouveau_bios *bios)
+{
+	struct priv *priv = data;
+	if (offset + length <= priv->size) {
+		memcpy_fromio(bios->data + offset, priv->rom + offset, length);
+		return length;
+	}
+	return 0;
+}
+
+static void
+pcirom_fini(void *data)
+{
+	struct priv *priv = data;
+	pci_unmap_rom(priv->pdev, priv->rom);
+	pci_disable_rom(priv->pdev);
+	kfree(priv);
+}
+
+static void *
+pcirom_init(struct nouveau_bios *bios, const char *name)
+{
+	struct pci_dev *pdev = nv_device(bios)->pdev;
+	struct priv *priv = NULL;
+	int ret;
+
+	if (!(ret = pci_enable_rom(pdev))) {
+		if (ret = -ENOMEM,
+		    (priv = kmalloc(sizeof(*priv), GFP_KERNEL))) {
+			if (ret = -EFAULT,
+			    (priv->rom = pci_map_rom(pdev, &priv->size))) {
+				priv->pdev = pdev;
+				return priv;
+			}
+			kfree(priv);
+		}
+		pci_disable_rom(pdev);
+	}
+
+	return ERR_PTR(ret);
+}
+
+const struct nvbios_source
+nvbios_pcirom = {
+	.name = "PCIROM",
+	.init = pcirom_init,
+	.fini = pcirom_fini,
+	.read = pcirom_read,
+	.rw = true,
+};
+
+static void *
+platform_init(struct nouveau_bios *bios, const char *name)
+{
+	struct pci_dev *pdev = nv_device(bios)->pdev;
+	struct priv *priv;
+	int ret = -ENOMEM;
+
+	if ((priv = kmalloc(sizeof(*priv), GFP_KERNEL))) {
+		if (ret = -ENODEV,
+		    (priv->rom = pci_platform_rom(pdev, &priv->size)))
+			return priv;
+		kfree(priv);
+	}
+
+	return ERR_PTR(ret);
+}
+
+const struct nvbios_source
+nvbios_platform = {
+	.name = "PLATFORM",
+	.init = platform_init,
+	.fini = (void(*)(void *))kfree,
+	.read = pcirom_read,
+	.rw = true,
+};
diff --git a/drivers/gpu/drm/nouveau/core/subdev/bios/shadowramin.c b/drivers/gpu/drm/nouveau/core/subdev/bios/shadowramin.c
new file mode 100644
index 000000000000..a7a890fad1e5
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/subdev/bios/shadowramin.c
@@ -0,0 +1,114 @@
+/*
+ * Copyright 2012 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#include "priv.h"
+
+struct priv {
+	struct nouveau_bios *bios;
+	u32 bar0;
+};
+
+static u32
+pramin_read(void *data, u32 offset, u32 length, struct nouveau_bios *bios)
+{
+	u32 i;
+	if (offset + length <= 0x00100000) {
+		for (i = offset; i < offset + length; i += 4)
+			*(u32 *)&bios->data[i] = nv_rd32(bios, 0x700000 + i);
+		return length;
+	}
+	return 0;
+}
+
+static void
+pramin_fini(void *data)
+{
+	struct priv *priv = data;
+	if (priv) {
+		nv_wr32(priv->bios, 0x001700, priv->bar0);
+		kfree(priv);
+	}
+}
+
+static void *
+pramin_init(struct nouveau_bios *bios, const char *name)
+{
+	struct priv *priv = NULL;
+	u64 addr = 0;
+
+	/* PRAMIN always potentially available prior to nv50 */
+	if (nv_device(bios)->card_type < NV_50)
+		return NULL;
+
+	/* we can't get the bios image pointer without PDISP */
+	if (nv_device(bios)->card_type >= GM100)
+		addr = nv_rd32(bios, 0x021c04);
+	else
+	if (nv_device(bios)->card_type >= NV_C0)
+		addr = nv_rd32(bios, 0x022500);
+	if (addr & 0x00000001) {
+		nv_debug(bios, "... display disabled\n");
+		return ERR_PTR(-ENODEV);
+	}
+
+	/* check that the window is enabled and in vram, particularly
+	 * important as we don't want to be touching vram on an
+	 * uninitialised board
+	 */
+	addr = nv_rd32(bios, 0x619f04);
+	if (!(addr & 0x00000008)) {
+		nv_debug(bios, "... not enabled\n");
+		return ERR_PTR(-ENODEV);
+	}
+	if ( (addr & 0x00000003) != 1) {
+		nv_debug(bios, "... not in vram\n");
+		return ERR_PTR(-ENODEV);
+	}
+
+	/* some alternate method inherited from xf86-video-nv... */
+	addr = (addr & 0xffffff00) << 8;
+	if (!addr) {
+		addr  = (u64)nv_rd32(bios, 0x001700) << 16;
+		addr += 0xf0000;
+	}
+
+	/* modify bar0 PRAMIN window to cover the bios image */
+	if (!(priv = kmalloc(sizeof(*priv), GFP_KERNEL))) {
+		nv_error(bios, "... out of memory\n");
+		return ERR_PTR(-ENOMEM);
+	}
+
+	priv->bios = bios;
+	priv->bar0 = nv_rd32(bios, 0x001700);
+	nv_wr32(bios, 0x001700, addr >> 16);
+	return priv;
+}
+
+const struct nvbios_source
+nvbios_ramin = {
+	.name = "PRAMIN",
+	.init = pramin_init,
+	.fini = pramin_fini,
+	.read = pramin_read,
+	.rw = true,
+};
diff --git a/drivers/gpu/drm/nouveau/core/subdev/bios/shadowrom.c b/drivers/gpu/drm/nouveau/core/subdev/bios/shadowrom.c
new file mode 100644
index 000000000000..b7992bc3ffa5
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/subdev/bios/shadowrom.c
@@ -0,0 +1,69 @@
+/*
+ * Copyright 2012 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#include "priv.h"
+
+static u32
+prom_read(void *data, u32 offset, u32 length, struct nouveau_bios *bios)
+{
+	u32 i;
+	if (offset + length <= 0x00100000) {
+		for (i = offset; i < offset + length; i += 4)
+			*(u32 *)&bios->data[i] = nv_rd32(bios, 0x300000 + i);
+		return length;
+	}
+	return 0;
+}
+
+static void
+prom_fini(void *data)
+{
+	struct nouveau_bios *bios = data;
+	if (nv_device(bios)->card_type < NV_50)
+		nv_mask(bios, 0x001850, 0x00000001, 0x00000001);
+	else
+		nv_mask(bios, 0x088050, 0x00000001, 0x00000001);
+}
+
+static void *
+prom_init(struct nouveau_bios *bios, const char *name)
+{
+	if (nv_device(bios)->card_type < NV_50) {
+		if (nv_device(bios)->card_type == NV_40 &&
+		    nv_device(bios)->chipset >= 0x4c)
+			return ERR_PTR(-ENODEV);
+		nv_mask(bios, 0x001850, 0x00000001, 0x00000000);
+	} else {
+		nv_mask(bios, 0x088050, 0x00000001, 0x00000000);
+	}
+	return bios;
+}
+
+const struct nvbios_source
+nvbios_rom = {
+	.name = "PROM",
+	.init = prom_init,
+	.fini = prom_fini,
+	.read = prom_read,
+	.rw = false,
+};
diff --git a/drivers/gpu/drm/nouveau/core/subdev/bios/timing.c b/drivers/gpu/drm/nouveau/core/subdev/bios/timing.c
index 350d44ab2ba2..8521eca1ed9c 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/bios/timing.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/bios/timing.c
@@ -89,7 +89,49 @@ nvbios_timingEp(struct nouveau_bios *bios, int idx,
 		struct nvbios_ramcfg *p)
 {
 	u16 data = nvbios_timingEe(bios, idx, ver, hdr, cnt, len), temp;
+	p->timing_ver = *ver;
+	p->timing_hdr = *hdr;
 	switch (!!data * *ver) {
+	case 0x10:
+		p->timing_10_WR    = nv_ro08(bios, data + 0x00);
+		p->timing_10_WTR   = nv_ro08(bios, data + 0x01);
+		p->timing_10_CL    = nv_ro08(bios, data + 0x02);
+		p->timing_10_RC    = nv_ro08(bios, data + 0x03);
+		p->timing_10_RFC   = nv_ro08(bios, data + 0x05);
+		p->timing_10_RAS   = nv_ro08(bios, data + 0x07);
+		p->timing_10_RP    = nv_ro08(bios, data + 0x09);
+		p->timing_10_RCDRD = nv_ro08(bios, data + 0x0a);
+		p->timing_10_RCDWR = nv_ro08(bios, data + 0x0b);
+		p->timing_10_RRD   = nv_ro08(bios, data + 0x0c);
+		p->timing_10_13    = nv_ro08(bios, data + 0x0d);
+		p->timing_10_ODT   = nv_ro08(bios, data + 0x0e) & 0x07;
+
+		p->timing_10_24  = 0xff;
+		p->timing_10_21  = 0;
+		p->timing_10_20  = 0;
+		p->timing_10_CWL = 0;
+		p->timing_10_18  = 0;
+		p->timing_10_16  = 0;
+
+		switch (min_t(u8, *hdr, 25)) {
+		case 25:
+			p->timing_10_24  = nv_ro08(bios, data + 0x18);
+		case 24:
+		case 23:
+		case 22:
+			p->timing_10_21  = nv_ro08(bios, data + 0x15);
+		case 21:
+			p->timing_10_20  = nv_ro08(bios, data + 0x14);
+		case 20:
+			p->timing_10_CWL = nv_ro08(bios, data + 0x13);
+		case 19:
+			p->timing_10_18  = nv_ro08(bios, data + 0x12);
+		case 18:
+		case 17:
+			p->timing_10_16  = nv_ro08(bios, data + 0x10);
+		}
+
+		break;
 	case 0x20:
 		p->timing[0] = nv_ro32(bios, data + 0x00);
 		p->timing[1] = nv_ro32(bios, data + 0x04);
diff --git a/drivers/gpu/drm/nouveau/core/subdev/clock/base.c b/drivers/gpu/drm/nouveau/core/subdev/clock/base.c
index a276a711294a..e51b72d47129 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/clock/base.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/clock/base.c
@@ -573,7 +573,7 @@ nouveau_clock_create_(struct nouveau_object *parent,
 
 	clk->allow_reclock = allow_reclock;
 
-	ret = nvkm_notify_init(&device->event, nouveau_clock_pwrsrc, true,
+	ret = nvkm_notify_init(NULL, &device->event, nouveau_clock_pwrsrc, true,
 			       NULL, 0, 0, &clk->pwrsrc_ntfy);
 	if (ret)
 		return ret;
diff --git a/drivers/gpu/drm/nouveau/core/subdev/clock/gk20a.c b/drivers/gpu/drm/nouveau/core/subdev/clock/gk20a.c
index 425a8d5e9129..fb4fad374bdd 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/clock/gk20a.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/clock/gk20a.c
@@ -109,7 +109,7 @@ struct gk20a_clk_pllg_params {
 };
 
 static const struct gk20a_clk_pllg_params gk20a_pllg_params = {
-	.min_vco = 1000, .max_vco = 1700,
+	.min_vco = 1000, .max_vco = 2064,
 	.min_u = 12, .max_u = 38,
 	.min_m = 1, .max_m = 255,
 	.min_n = 8, .max_n = 255,
@@ -470,76 +470,91 @@ gk20a_pstates[] = {
 	{
 		.base = {
 			.domain[nv_clk_src_gpc] = 72000,
+			.voltage = 0,
 		},
 	},
 	{
 		.base = {
 			.domain[nv_clk_src_gpc] = 108000,
+			.voltage = 1,
 		},
 	},
 	{
 		.base = {
 			.domain[nv_clk_src_gpc] = 180000,
+			.voltage = 2,
 		},
 	},
 	{
 		.base = {
 			.domain[nv_clk_src_gpc] = 252000,
+			.voltage = 3,
 		},
 	},
 	{
 		.base = {
 			.domain[nv_clk_src_gpc] = 324000,
+			.voltage = 4,
 		},
 	},
 	{
 		.base = {
 			.domain[nv_clk_src_gpc] = 396000,
+			.voltage = 5,
 		},
 	},
 	{
 		.base = {
 			.domain[nv_clk_src_gpc] = 468000,
+			.voltage = 6,
 		},
 	},
 	{
 		.base = {
 			.domain[nv_clk_src_gpc] = 540000,
+			.voltage = 7,
 		},
 	},
 	{
 		.base = {
 			.domain[nv_clk_src_gpc] = 612000,
+			.voltage = 8,
 		},
 	},
 	{
 		.base = {
 			.domain[nv_clk_src_gpc] = 648000,
+			.voltage = 9,
 		},
 	},
 	{
 		.base = {
 			.domain[nv_clk_src_gpc] = 684000,
+			.voltage = 10,
 		},
 	},
 	{
 		.base = {
 			.domain[nv_clk_src_gpc] = 708000,
+			.voltage = 11,
 		},
 	},
 	{
 		.base = {
 			.domain[nv_clk_src_gpc] = 756000,
+			.voltage = 12,
 		},
 	},
 	{
 		.base = {
 			.domain[nv_clk_src_gpc] = 804000,
+			.voltage = 13,
 		},
 	},
 	{
 		.base = {
 			.domain[nv_clk_src_gpc] = 852000,
+			.voltage = 14,
 		},
 	},
 };
diff --git a/drivers/gpu/drm/nouveau/core/subdev/clock/nva3.c b/drivers/gpu/drm/nouveau/core/subdev/clock/nva3.c
index 087012b18956..07ad01247675 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/clock/nva3.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/clock/nva3.c
@@ -20,8 +20,10 @@
  * OTHER DEALINGS IN THE SOFTWARE.
  *
  * Authors: Ben Skeggs
+ *          Roy Spliet
  */
 
+#include <engine/fifo.h>
 #include <subdev/bios.h>
 #include <subdev/bios/pll.h>
 #include <subdev/timer.h>
@@ -42,9 +44,17 @@ static u32
 read_vco(struct nva3_clock_priv *priv, int clk)
 {
 	u32 sctl = nv_rd32(priv, 0x4120 + (clk * 4));
-	if ((sctl & 0x00000030) != 0x00000030)
+
+	switch (sctl & 0x00000030) {
+	case 0x00000000:
+		return nv_device(priv)->crystal;
+	case 0x00000020:
 		return read_pll(priv, 0x41, 0x00e820);
-	return read_pll(priv, 0x42, 0x00e8a0);
+	case 0x00000030:
+		return read_pll(priv, 0x42, 0x00e8a0);
+	default:
+		return 0;
+	}
 }
 
 static u32
@@ -66,14 +76,25 @@ read_clk(struct nva3_clock_priv *priv, int clk, bool ignore_en)
 	if (!ignore_en && !(sctl & 0x00000100))
 		return 0;
 
+	/* out_alt */
+	if (sctl & 0x00000400)
+		return 108000;
+
+	/* vco_out */
 	switch (sctl & 0x00003000) {
 	case 0x00000000:
-		return nv_device(priv)->crystal;
+		if (!(sctl & 0x00000200))
+			return nv_device(priv)->crystal;
+		return 0;
 	case 0x00002000:
 		if (sctl & 0x00000040)
 			return 108000;
 		return 100000;
 	case 0x00003000:
+		/* vco_enable */
+		if (!(sctl & 0x00000001))
+			return 0;
+
 		sclk = read_vco(priv, clk);
 		sdiv = ((sctl & 0x003f0000) >> 16) + 2;
 		return (sclk * 2) / sdiv;
@@ -95,7 +116,9 @@ read_pll(struct nva3_clock_priv *priv, int clk, u32 pll)
 			N = (coef & 0x0000ff00) >> 8;
 			P = (coef & 0x003f0000) >> 16;
 
-			/* no post-divider on these.. */
+			/* no post-divider on these..
+			 * XXX: it looks more like two post-"dividers" that
+			 * cross each other out in the default RPLL config */
 			if ((pll & 0x00ff00) == 0x00e800)
 				P = 1;
 
@@ -114,13 +137,13 @@ static int
 nva3_clock_read(struct nouveau_clock *clk, enum nv_clk_src src)
 {
 	struct nva3_clock_priv *priv = (void *)clk;
+	u32 hsrc;
 
 	switch (src) {
 	case nv_clk_src_crystal:
 		return nv_device(priv)->crystal;
-	case nv_clk_src_href:
-		return 100000;
 	case nv_clk_src_core:
+	case nv_clk_src_core_intm:
 		return read_pll(priv, 0x00, 0x4200);
 	case nv_clk_src_shader:
 		return read_pll(priv, 0x01, 0x4220);
@@ -132,24 +155,33 @@ nva3_clock_read(struct nouveau_clock *clk, enum nv_clk_src src)
 		return read_clk(priv, 0x21, false);
 	case nv_clk_src_daemon:
 		return read_clk(priv, 0x25, false);
+	case nv_clk_src_host:
+		hsrc = (nv_rd32(priv, 0xc040) & 0x30000000) >> 28;
+		switch (hsrc) {
+		case 0:
+			return read_clk(priv, 0x1d, false);
+		case 2:
+		case 3:
+			return 277000;
+		default:
+			nv_error(clk, "unknown HOST clock source %d\n", hsrc);
+			return -EINVAL;
+		}
 	default:
 		nv_error(clk, "invalid clock source %d\n", src);
 		return -EINVAL;
 	}
+
+	return 0;
 }
 
 int
-nva3_clock_info(struct nouveau_clock *clock, int clk, u32 pll, u32 khz,
+nva3_clk_info(struct nouveau_clock *clock, int clk, u32 khz,
 		struct nva3_clock_info *info)
 {
-	struct nouveau_bios *bios = nouveau_bios(clock);
 	struct nva3_clock_priv *priv = (void *)clock;
-	struct nvbios_pll limits;
-	u32 oclk, sclk, sdiv;
-	int P, N, M, diff;
-	int ret;
+	u32 oclk, sclk, sdiv, diff;
 
-	info->pll = 0;
 	info->clk = 0;
 
 	switch (khz) {
@@ -164,43 +196,69 @@ nva3_clock_info(struct nouveau_clock *clock, int clk, u32 pll, u32 khz,
 		return khz;
 	default:
 		sclk = read_vco(priv, clk);
-		sdiv = min((sclk * 2) / (khz - 2999), (u32)65);
-		/* if the clock has a PLL attached, and we can get a within
-		 * [-2, 3) MHz of a divider, we'll disable the PLL and use
-		 * the divider instead.
-		 *
-		 * divider can go as low as 2, limited here because NVIDIA
+		sdiv = min((sclk * 2) / khz, (u32)65);
+		oclk = (sclk * 2) / sdiv;
+		diff = ((khz + 3000) - oclk);
+
+		/* When imprecise, play it safe and aim for a clock lower than
+		 * desired rather than higher */
+		if (diff < 0) {
+			sdiv++;
+			oclk = (sclk * 2) / sdiv;
+		}
+
+		/* divider can go as low as 2, limited here because NVIDIA
 		 * and the VBIOS on my NVA8 seem to prefer using the PLL
 		 * for 810MHz - is there a good reason?
-		 */
+		 * XXX: PLLs with refclk 810MHz?  */
 		if (sdiv > 4) {
-			oclk = (sclk * 2) / sdiv;
-			diff = khz - oclk;
-			if (!pll || (diff >= -2000 && diff < 3000)) {
-				info->clk = (((sdiv - 2) << 16) | 0x00003100);
-				return oclk;
-			}
+			info->clk = (((sdiv - 2) << 16) | 0x00003100);
+			return oclk;
 		}
 
-		if (!pll)
-			return -ERANGE;
 		break;
 	}
 
+	return -ERANGE;
+}
+
+int
+nva3_pll_info(struct nouveau_clock *clock, int clk, u32 pll, u32 khz,
+		struct nva3_clock_info *info)
+{
+	struct nouveau_bios *bios = nouveau_bios(clock);
+	struct nva3_clock_priv *priv = (void *)clock;
+	struct nvbios_pll limits;
+	int P, N, M, diff;
+	int ret;
+
+	info->pll = 0;
+
+	/* If we can get a within [-2, 3) MHz of a divider, we'll disable the
+	 * PLL and use the divider instead. */
+	ret = nva3_clk_info(clock, clk, khz, info);
+	diff = khz - ret;
+	if (!pll || (diff >= -2000 && diff < 3000)) {
+		goto out;
+	}
+
+	/* Try with PLL */
 	ret = nvbios_pll_parse(bios, pll, &limits);
 	if (ret)
 		return ret;
 
-	limits.refclk = read_clk(priv, clk - 0x10, true);
-	if (!limits.refclk)
+	ret = nva3_clk_info(clock, clk - 0x10, limits.refclk, info);
+	if (ret != limits.refclk)
 		return -EINVAL;
 
 	ret = nva3_pll_calc(nv_subdev(priv), &limits, khz, &N, NULL, &M, &P);
 	if (ret >= 0) {
-		info->clk = nv_rd32(priv, 0x4120 + (clk * 4));
 		info->pll = (P << 16) | (N << 8) | M;
 	}
 
+out:
+	info->fb_delay = max(((khz + 7566) / 15133), (u32) 18);
+
 	return ret ? ret : -ERANGE;
 }
 
@@ -208,13 +266,76 @@ static int
 calc_clk(struct nva3_clock_priv *priv, struct nouveau_cstate *cstate,
 	 int clk, u32 pll, int idx)
 {
-	int ret = nva3_clock_info(&priv->base, clk, pll, cstate->domain[idx],
+	int ret = nva3_pll_info(&priv->base, clk, pll, cstate->domain[idx],
 				  &priv->eng[idx]);
 	if (ret >= 0)
 		return 0;
 	return ret;
 }
 
+static int
+calc_host(struct nva3_clock_priv *priv, struct nouveau_cstate *cstate)
+{
+	int ret = 0;
+	u32 kHz = cstate->domain[nv_clk_src_host];
+	struct nva3_clock_info *info = &priv->eng[nv_clk_src_host];
+
+	if (kHz == 277000) {
+		info->clk = 0;
+		info->host_out = NVA3_HOST_277;
+		return 0;
+	}
+
+	info->host_out = NVA3_HOST_CLK;
+
+	ret = nva3_clk_info(&priv->base, 0x1d, kHz, info);
+	if (ret >= 0)
+		return 0;
+	return ret;
+}
+
+int
+nva3_clock_pre(struct nouveau_clock *clk, unsigned long *flags)
+{
+	struct nouveau_fifo *pfifo = nouveau_fifo(clk);
+
+	/* halt and idle execution engines */
+	nv_mask(clk, 0x020060, 0x00070000, 0x00000000);
+	nv_mask(clk, 0x002504, 0x00000001, 0x00000001);
+	/* Wait until the interrupt handler is finished */
+	if (!nv_wait(clk, 0x000100, 0xffffffff, 0x00000000))
+		return -EBUSY;
+
+	if (pfifo)
+		pfifo->pause(pfifo, flags);
+
+	if (!nv_wait(clk, 0x002504, 0x00000010, 0x00000010))
+		return -EIO;
+	if (!nv_wait(clk, 0x00251c, 0x0000003f, 0x0000003f))
+		return -EIO;
+
+	return 0;
+}
+
+void
+nva3_clock_post(struct nouveau_clock *clk, unsigned long *flags)
+{
+	struct nouveau_fifo *pfifo = nouveau_fifo(clk);
+
+	if (pfifo && flags)
+		pfifo->start(pfifo, flags);
+
+	nv_mask(clk, 0x002504, 0x00000001, 0x00000000);
+	nv_mask(clk, 0x020060, 0x00070000, 0x00040000);
+}
+
+static void
+disable_clk_src(struct nva3_clock_priv *priv, u32 src)
+{
+	nv_mask(priv, src, 0x00000100, 0x00000000);
+	nv_mask(priv, src, 0x00000001, 0x00000000);
+}
+
 static void
 prog_pll(struct nva3_clock_priv *priv, int clk, u32 pll, int idx)
 {
@@ -223,24 +344,35 @@ prog_pll(struct nva3_clock_priv *priv, int clk, u32 pll, int idx)
 	const u32 src1 = 0x004160 + (clk * 4);
 	const u32 ctrl = pll + 0;
 	const u32 coef = pll + 4;
+	u32 bypass;
 
 	if (info->pll) {
-		nv_mask(priv, src0, 0x00000101, 0x00000101);
+		/* Always start from a non-PLL clock */
+		bypass = nv_rd32(priv, ctrl)  & 0x00000008;
+		if (!bypass) {
+			nv_mask(priv, src1, 0x00000101, 0x00000101);
+			nv_mask(priv, ctrl, 0x00000008, 0x00000008);
+			udelay(20);
+		}
+
+		nv_mask(priv, src0, 0x003f3141, 0x00000101 | info->clk);
 		nv_wr32(priv, coef, info->pll);
 		nv_mask(priv, ctrl, 0x00000015, 0x00000015);
 		nv_mask(priv, ctrl, 0x00000010, 0x00000000);
-		nv_wait(priv, ctrl, 0x00020000, 0x00020000);
+		if (!nv_wait(priv, ctrl, 0x00020000, 0x00020000)) {
+			nv_mask(priv, ctrl, 0x00000010, 0x00000010);
+			nv_mask(priv, src0, 0x00000101, 0x00000000);
+			return;
+		}
 		nv_mask(priv, ctrl, 0x00000010, 0x00000010);
 		nv_mask(priv, ctrl, 0x00000008, 0x00000000);
-		nv_mask(priv, src1, 0x00000100, 0x00000000);
-		nv_mask(priv, src1, 0x00000001, 0x00000000);
+		disable_clk_src(priv, src1);
 	} else {
 		nv_mask(priv, src1, 0x003f3141, 0x00000101 | info->clk);
 		nv_mask(priv, ctrl, 0x00000018, 0x00000018);
 		udelay(20);
 		nv_mask(priv, ctrl, 0x00000001, 0x00000000);
-		nv_mask(priv, src0, 0x00000100, 0x00000000);
-		nv_mask(priv, src0, 0x00000001, 0x00000000);
+		disable_clk_src(priv, src0);
 	}
 }
 
@@ -251,18 +383,72 @@ prog_clk(struct nva3_clock_priv *priv, int clk, int idx)
 	nv_mask(priv, 0x004120 + (clk * 4), 0x003f3141, 0x00000101 | info->clk);
 }
 
+static void
+prog_host(struct nva3_clock_priv *priv)
+{
+	struct nva3_clock_info *info = &priv->eng[nv_clk_src_host];
+	u32 hsrc = (nv_rd32(priv, 0xc040));
+
+	switch (info->host_out) {
+	case NVA3_HOST_277:
+		if ((hsrc & 0x30000000) == 0) {
+			nv_wr32(priv, 0xc040, hsrc | 0x20000000);
+			disable_clk_src(priv, 0x4194);
+		}
+		break;
+	case NVA3_HOST_CLK:
+		prog_clk(priv, 0x1d, nv_clk_src_host);
+		if ((hsrc & 0x30000000) >= 0x20000000) {
+			nv_wr32(priv, 0xc040, hsrc & ~0x30000000);
+		}
+		break;
+	default:
+		break;
+	}
+
+	/* This seems to be a clock gating factor on idle, always set to 64 */
+	nv_wr32(priv, 0xc044, 0x3e);
+}
+
+static void
+prog_core(struct nva3_clock_priv *priv, int idx)
+{
+	struct nva3_clock_info *info = &priv->eng[idx];
+	u32 fb_delay = nv_rd32(priv, 0x10002c);
+
+	if (fb_delay < info->fb_delay)
+		nv_wr32(priv, 0x10002c, info->fb_delay);
+
+	prog_pll(priv, 0x00, 0x004200, idx);
+
+	if (fb_delay > info->fb_delay)
+		nv_wr32(priv, 0x10002c, info->fb_delay);
+}
+
 static int
 nva3_clock_calc(struct nouveau_clock *clk, struct nouveau_cstate *cstate)
 {
 	struct nva3_clock_priv *priv = (void *)clk;
+	struct nva3_clock_info *core = &priv->eng[nv_clk_src_core];
 	int ret;
 
 	if ((ret = calc_clk(priv, cstate, 0x10, 0x4200, nv_clk_src_core)) ||
 	    (ret = calc_clk(priv, cstate, 0x11, 0x4220, nv_clk_src_shader)) ||
 	    (ret = calc_clk(priv, cstate, 0x20, 0x0000, nv_clk_src_disp)) ||
-	    (ret = calc_clk(priv, cstate, 0x21, 0x0000, nv_clk_src_vdec)))
+	    (ret = calc_clk(priv, cstate, 0x21, 0x0000, nv_clk_src_vdec)) ||
+	    (ret = calc_host(priv, cstate)))
 		return ret;
 
+	/* XXX: Should be reading the highest bit in the VBIOS clock to decide
+	 * whether to use a PLL or not... but using a PLL defeats the purpose */
+	if (core->pll) {
+		ret = nva3_clk_info(clk, 0x10,
+				cstate->domain[nv_clk_src_core_intm],
+				&priv->eng[nv_clk_src_core_intm]);
+		if (ret < 0)
+			return ret;
+	}
+
 	return 0;
 }
 
@@ -270,11 +456,31 @@ static int
 nva3_clock_prog(struct nouveau_clock *clk)
 {
 	struct nva3_clock_priv *priv = (void *)clk;
-	prog_pll(priv, 0x00, 0x004200, nv_clk_src_core);
+	struct nva3_clock_info *core = &priv->eng[nv_clk_src_core];
+	int ret = 0;
+	unsigned long flags;
+	unsigned long *f = &flags;
+
+	ret = nva3_clock_pre(clk, f);
+	if (ret)
+		goto out;
+
+	if (core->pll)
+		prog_core(priv, nv_clk_src_core_intm);
+
+	prog_core(priv,  nv_clk_src_core);
 	prog_pll(priv, 0x01, 0x004220, nv_clk_src_shader);
 	prog_clk(priv, 0x20, nv_clk_src_disp);
 	prog_clk(priv, 0x21, nv_clk_src_vdec);
-	return 0;
+	prog_host(priv);
+
+out:
+	if (ret == -EBUSY)
+		f = NULL;
+
+	nva3_clock_post(clk, f);
+
+	return ret;
 }
 
 static void
@@ -284,13 +490,14 @@ nva3_clock_tidy(struct nouveau_clock *clk)
 
 static struct nouveau_clocks
 nva3_domain[] = {
-	{ nv_clk_src_crystal, 0xff },
-	{ nv_clk_src_href   , 0xff },
-	{ nv_clk_src_core   , 0x00, 0, "core", 1000 },
-	{ nv_clk_src_shader , 0x01, 0, "shader", 1000 },
-	{ nv_clk_src_mem    , 0x02, 0, "memory", 1000 },
-	{ nv_clk_src_vdec   , 0x03 },
-	{ nv_clk_src_disp   , 0x04 },
+	{ nv_clk_src_crystal  , 0xff },
+	{ nv_clk_src_core     , 0x00, 0, "core", 1000 },
+	{ nv_clk_src_shader   , 0x01, 0, "shader", 1000 },
+	{ nv_clk_src_mem      , 0x02, 0, "memory", 1000 },
+	{ nv_clk_src_vdec     , 0x03 },
+	{ nv_clk_src_disp     , 0x04 },
+	{ nv_clk_src_host     , 0x05 },
+	{ nv_clk_src_core_intm, 0x06 },
 	{ nv_clk_src_max }
 };
 
@@ -303,7 +510,7 @@ nva3_clock_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	int ret;
 
 	ret = nouveau_clock_create(parent, engine, oclass, nva3_domain, NULL, 0,
-				   false, &priv);
+				   true, &priv);
 	*pobject = nv_object(priv);
 	if (ret)
 		return ret;
diff --git a/drivers/gpu/drm/nouveau/core/subdev/clock/nva3.h b/drivers/gpu/drm/nouveau/core/subdev/clock/nva3.h
index 6229a509b42e..a45a1038b12f 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/clock/nva3.h
+++ b/drivers/gpu/drm/nouveau/core/subdev/clock/nva3.h
@@ -6,9 +6,15 @@
 struct nva3_clock_info {
 	u32 clk;
 	u32 pll;
+	enum {
+		NVA3_HOST_277,
+		NVA3_HOST_CLK,
+	} host_out;
+	u32 fb_delay;
 };
 
-int nva3_clock_info(struct nouveau_clock *, int, u32, u32,
+int nva3_pll_info(struct nouveau_clock *, int, u32, u32,
 		    struct nva3_clock_info *);
-
+int nva3_clock_pre(struct nouveau_clock *clk, unsigned long *flags);
+void nva3_clock_post(struct nouveau_clock *clk, unsigned long *flags);
 #endif
diff --git a/drivers/gpu/drm/nouveau/core/subdev/clock/nvaa.c b/drivers/gpu/drm/nouveau/core/subdev/clock/nvaa.c
index 74e19731b1b7..54aeab8005a0 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/clock/nvaa.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/clock/nvaa.c
@@ -28,6 +28,7 @@
 #include <subdev/timer.h>
 #include <subdev/clock.h>
 
+#include "nva3.h"
 #include "pll.h"
 
 struct nvaa_clock_priv {
@@ -299,25 +300,14 @@ static int
 nvaa_clock_prog(struct nouveau_clock *clk)
 {
 	struct nvaa_clock_priv *priv = (void *)clk;
-	struct nouveau_fifo *pfifo = nouveau_fifo(clk);
+	u32 pllmask = 0, mast;
 	unsigned long flags;
-	u32 pllmask = 0, mast, ptherm_gate;
-	int ret = -EBUSY;
-
-	/* halt and idle execution engines */
-	ptherm_gate = nv_mask(clk, 0x020060, 0x00070000, 0x00000000);
-	nv_mask(clk, 0x002504, 0x00000001, 0x00000001);
-	/* Wait until the interrupt handler is finished */
-	if (!nv_wait(clk, 0x000100, 0xffffffff, 0x00000000))
-		goto resume;
-
-	if (pfifo)
-		pfifo->pause(pfifo, &flags);
+	unsigned long *f = &flags;
+	int ret = 0;
 
-	if (!nv_wait(clk, 0x002504, 0x00000010, 0x00000010))
-		goto resume;
-	if (!nv_wait(clk, 0x00251c, 0x0000003f, 0x0000003f))
-		goto resume;
+	ret = nva3_clock_pre(clk, f);
+	if (ret)
+		goto out;
 
 	/* First switch to safe clocks: href */
 	mast = nv_mask(clk, 0xc054, 0x03400e70, 0x03400640);
@@ -375,15 +365,8 @@ nvaa_clock_prog(struct nouveau_clock *clk)
 	}
 
 	nv_wr32(clk, 0xc054, mast);
-	ret = 0;
 
 resume:
-	if (pfifo)
-		pfifo->start(pfifo, &flags);
-
-	nv_mask(clk, 0x002504, 0x00000001, 0x00000000);
-	nv_wr32(clk, 0x020060, ptherm_gate);
-
 	/* Disable some PLLs and dividers when unused */
 	if (priv->csrc != nv_clk_src_core) {
 		nv_wr32(clk, 0x4040, 0x00000000);
@@ -395,6 +378,12 @@ resume:
 		nv_mask(clk, 0x4020, 0x80000000, 0x00000000);
 	}
 
+out:
+	if (ret == -EBUSY)
+		f = NULL;
+
+	nva3_clock_post(clk, f);
+
 	return ret;
 }
 
diff --git a/drivers/gpu/drm/nouveau/core/subdev/devinit/base.c b/drivers/gpu/drm/nouveau/core/subdev/devinit/base.c
index 239acfe876c3..0e45cee82463 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/devinit/base.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/devinit/base.c
@@ -24,8 +24,6 @@
 
 #include <core/option.h>
 
-#include <subdev/bios.h>
-#include <subdev/bios/init.h>
 #include <subdev/vga.h>
 
 #include "priv.h"
@@ -56,7 +54,7 @@ _nouveau_devinit_init(struct nouveau_object *object)
 	if (ret)
 		return ret;
 
-	ret = nvbios_init(&devinit->base, devinit->post);
+	ret = impl->post(&devinit->base, devinit->post);
 	if (ret)
 		return ret;
 
diff --git a/drivers/gpu/drm/nouveau/core/subdev/devinit/fbmem.h b/drivers/gpu/drm/nouveau/core/subdev/devinit/fbmem.h
index 4fe49cf4c99a..6103484fea72 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/devinit/fbmem.h
+++ b/drivers/gpu/drm/nouveau/core/subdev/devinit/fbmem.h
@@ -26,22 +26,8 @@
 
 #include <core/device.h>
 
-#define NV04_PFB_BOOT_0						0x00100000
-#	define NV04_PFB_BOOT_0_RAM_AMOUNT			0x00000003
-#	define NV04_PFB_BOOT_0_RAM_AMOUNT_32MB			0x00000000
-#	define NV04_PFB_BOOT_0_RAM_AMOUNT_4MB			0x00000001
-#	define NV04_PFB_BOOT_0_RAM_AMOUNT_8MB			0x00000002
-#	define NV04_PFB_BOOT_0_RAM_AMOUNT_16MB			0x00000003
-#	define NV04_PFB_BOOT_0_RAM_WIDTH_128			0x00000004
-#	define NV04_PFB_BOOT_0_RAM_TYPE				0x00000028
-#	define NV04_PFB_BOOT_0_RAM_TYPE_SGRAM_8MBIT		0x00000000
-#	define NV04_PFB_BOOT_0_RAM_TYPE_SGRAM_16MBIT		0x00000008
-#	define NV04_PFB_BOOT_0_RAM_TYPE_SGRAM_16MBIT_4BANK	0x00000010
-#	define NV04_PFB_BOOT_0_RAM_TYPE_SDRAM_16MBIT		0x00000018
-#	define NV04_PFB_BOOT_0_RAM_TYPE_SDRAM_64MBIT		0x00000020
-#	define NV04_PFB_BOOT_0_RAM_TYPE_SDRAM_64MBITX16		0x00000028
-#	define NV04_PFB_BOOT_0_UMA_ENABLE			0x00000100
-#	define NV04_PFB_BOOT_0_UMA_SIZE				0x0000f000
+#include <subdev/fb/regsnv04.h>
+
 #define NV04_PFB_DEBUG_0					0x00100080
 #	define NV04_PFB_DEBUG_0_PAGE_MODE			0x00000001
 #	define NV04_PFB_DEBUG_0_REFRESH_OFF			0x00000010
diff --git a/drivers/gpu/drm/nouveau/core/subdev/devinit/gm107.c b/drivers/gpu/drm/nouveau/core/subdev/devinit/gm107.c
index c69bc7f54e37..4ba43d6a1ec8 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/devinit/gm107.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/devinit/gm107.c
@@ -24,7 +24,7 @@
 
 #include "nv50.h"
 
-static u64
+u64
 gm107_devinit_disable(struct nouveau_devinit *devinit)
 {
 	struct nv50_devinit_priv *priv = (void *)devinit;
@@ -53,4 +53,5 @@ gm107_devinit_oclass = &(struct nouveau_devinit_impl) {
 	},
 	.pll_set = nvc0_devinit_pll_set,
 	.disable = gm107_devinit_disable,
+	.post = nvbios_init,
 }.base;
diff --git a/drivers/gpu/drm/nouveau/core/subdev/devinit/gm204.c b/drivers/gpu/drm/nouveau/core/subdev/devinit/gm204.c
new file mode 100644
index 000000000000..e44a86662a2a
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/subdev/devinit/gm204.c
@@ -0,0 +1,173 @@
+/*
+ * Copyright 2013 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs
+ */
+
+#include <subdev/bios.h>
+#include <subdev/bios/bit.h>
+#include <subdev/bios/pmu.h>
+
+#include "nv50.h"
+
+static void
+pmu_code(struct nv50_devinit_priv *priv, u32 pmu, u32 img, u32 len, bool sec)
+{
+	struct nouveau_bios *bios = nouveau_bios(priv);
+	int i;
+
+	nv_wr32(priv, 0x10a180, 0x01000000 | (sec ? 0x10000000 : 0) | pmu);
+	for (i = 0; i < len; i += 4) {
+		if ((i & 0xff) == 0)
+			nv_wr32(priv, 0x10a188, (pmu + i) >> 8);
+		nv_wr32(priv, 0x10a184, nv_ro32(bios, img + i));
+	}
+
+	while (i & 0xff) {
+		nv_wr32(priv, 0x10a184, 0x00000000);
+		i += 4;
+	}
+}
+
+static void
+pmu_data(struct nv50_devinit_priv *priv, u32 pmu, u32 img, u32 len)
+{
+	struct nouveau_bios *bios = nouveau_bios(priv);
+	int i;
+
+	nv_wr32(priv, 0x10a1c0, 0x01000000 | pmu);
+	for (i = 0; i < len; i += 4)
+		nv_wr32(priv, 0x10a1c4, nv_ro32(bios, img + i));
+}
+
+static u32
+pmu_args(struct nv50_devinit_priv *priv, u32 argp, u32 argi)
+{
+	nv_wr32(priv, 0x10a1c0, argp);
+	nv_wr32(priv, 0x10a1c0, nv_rd32(priv, 0x10a1c4) + argi);
+	return nv_rd32(priv, 0x10a1c4);
+}
+
+static void
+pmu_exec(struct nv50_devinit_priv *priv, u32 init_addr)
+{
+	nv_wr32(priv, 0x10a104, init_addr);
+	nv_wr32(priv, 0x10a10c, 0x00000000);
+	nv_wr32(priv, 0x10a100, 0x00000002);
+}
+
+static int
+pmu_load(struct nv50_devinit_priv *priv, u8 type, bool post,
+	 u32 *init_addr_pmu, u32 *args_addr_pmu)
+{
+	struct nouveau_bios *bios = nouveau_bios(priv);
+	struct nvbios_pmuR pmu;
+
+	if (!nvbios_pmuRm(bios, type, &pmu)) {
+		nv_error(priv, "VBIOS PMU fuc %02x not found\n", type);
+		return -EINVAL;
+	}
+
+	if (!post)
+		return 0;
+
+	pmu_code(priv, pmu.boot_addr_pmu, pmu.boot_addr, pmu.boot_size, false);
+	pmu_code(priv, pmu.code_addr_pmu, pmu.code_addr, pmu.code_size, true);
+	pmu_data(priv, pmu.data_addr_pmu, pmu.data_addr, pmu.data_size);
+
+	if (init_addr_pmu) {
+		*init_addr_pmu = pmu.init_addr_pmu;
+		*args_addr_pmu = pmu.args_addr_pmu;
+		return 0;
+	}
+
+	return pmu_exec(priv, pmu.init_addr_pmu), 0;
+}
+
+static int
+gm204_devinit_post(struct nouveau_subdev *subdev, bool post)
+{
+	struct nv50_devinit_priv *priv = (void *)nouveau_devinit(subdev);
+	struct nouveau_bios *bios = nouveau_bios(priv);
+	struct bit_entry bit_I;
+	u32 init, args;
+	int ret;
+
+	if (bit_entry(bios, 'I', &bit_I) || bit_I.version != 1 ||
+					    bit_I.length < 0x1c) {
+		nv_error(priv, "VBIOS PMU init data not found\n");
+		return -EINVAL;
+	}
+
+	/* reset PMU and load init table parser ucode */
+	if (post) {
+		nv_mask(priv, 0x000200, 0x00002000, 0x00000000);
+		nv_mask(priv, 0x000200, 0x00002000, 0x00002000);
+		nv_rd32(priv, 0x000200);
+		while (nv_rd32(priv, 0x10a10c) & 0x00000006) {
+		}
+	}
+
+	ret = pmu_load(priv, 0x04, post, &init, &args);
+	if (ret)
+		return ret;
+
+	/* upload first chunk of init data */
+	if (post) {
+		u32 pmu = pmu_args(priv, args + 0x08, 0x08);
+		u32 img = nv_ro16(bios, bit_I.offset + 0x14);
+		u32 len = nv_ro16(bios, bit_I.offset + 0x16);
+		pmu_data(priv, pmu, img, len);
+	}
+
+	/* upload second chunk of init data */
+	if (post) {
+		u32 pmu = pmu_args(priv, args + 0x08, 0x10);
+		u32 img = nv_ro16(bios, bit_I.offset + 0x18);
+		u32 len = nv_ro16(bios, bit_I.offset + 0x1a);
+		pmu_data(priv, pmu, img, len);
+	}
+
+	/* execute init tables */
+	if (post) {
+		nv_wr32(priv, 0x10a040, 0x00005000);
+		pmu_exec(priv, init);
+		while (!(nv_rd32(priv, 0x10a040) & 0x00002000)) {
+		}
+	}
+
+	/* load and execute some other ucode image (bios therm?) */
+	return pmu_load(priv, 0x01, post, NULL, NULL);
+}
+
+struct nouveau_oclass *
+gm204_devinit_oclass = &(struct nouveau_devinit_impl) {
+	.base.handle = NV_SUBDEV(DEVINIT, 0x07),
+	.base.ofuncs = &(struct nouveau_ofuncs) {
+		.ctor = nv50_devinit_ctor,
+		.dtor = _nouveau_devinit_dtor,
+		.init = nv50_devinit_init,
+		.fini = _nouveau_devinit_fini,
+	},
+	.pll_set = nvc0_devinit_pll_set,
+	.disable = gm107_devinit_disable,
+	.post = gm204_devinit_post,
+}.base;
diff --git a/drivers/gpu/drm/nouveau/core/subdev/devinit/nv04.c b/drivers/gpu/drm/nouveau/core/subdev/devinit/nv04.c
index 052ad690b468..65651c50f6ea 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/devinit/nv04.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/devinit/nv04.c
@@ -464,4 +464,5 @@ nv04_devinit_oclass = &(struct nouveau_devinit_impl) {
 	},
 	.meminit = nv04_devinit_meminit,
 	.pll_set = nv04_devinit_pll_set,
+	.post = nvbios_init,
 }.base;
diff --git a/drivers/gpu/drm/nouveau/core/subdev/devinit/nv05.c b/drivers/gpu/drm/nouveau/core/subdev/devinit/nv05.c
index 4a19c10e5178..a2007a3efc4d 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/devinit/nv05.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/devinit/nv05.c
@@ -136,4 +136,5 @@ nv05_devinit_oclass = &(struct nouveau_devinit_impl) {
 	},
 	.meminit = nv05_devinit_meminit,
 	.pll_set = nv04_devinit_pll_set,
+	.post = nvbios_init,
 }.base;
diff --git a/drivers/gpu/drm/nouveau/core/subdev/devinit/nv10.c b/drivers/gpu/drm/nouveau/core/subdev/devinit/nv10.c
index 3b8d657da279..178b46f79b50 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/devinit/nv10.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/devinit/nv10.c
@@ -107,4 +107,5 @@ nv10_devinit_oclass = &(struct nouveau_devinit_impl) {
 	},
 	.meminit = nv10_devinit_meminit,
 	.pll_set = nv04_devinit_pll_set,
+	.post = nvbios_init,
 }.base;
diff --git a/drivers/gpu/drm/nouveau/core/subdev/devinit/nv1a.c b/drivers/gpu/drm/nouveau/core/subdev/devinit/nv1a.c
index 526d0c6faacd..995dd97af3e9 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/devinit/nv1a.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/devinit/nv1a.c
@@ -34,4 +34,5 @@ nv1a_devinit_oclass = &(struct nouveau_devinit_impl) {
 		.fini = nv04_devinit_fini,
 	},
 	.pll_set = nv04_devinit_pll_set,
+	.post = nvbios_init,
 }.base;
diff --git a/drivers/gpu/drm/nouveau/core/subdev/devinit/nv20.c b/drivers/gpu/drm/nouveau/core/subdev/devinit/nv20.c
index 04bc9732644c..915089fb46f7 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/devinit/nv20.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/devinit/nv20.c
@@ -71,4 +71,5 @@ nv20_devinit_oclass = &(struct nouveau_devinit_impl) {
 	},
 	.meminit = nv20_devinit_meminit,
 	.pll_set = nv04_devinit_pll_set,
+	.post = nvbios_init,
 }.base;
diff --git a/drivers/gpu/drm/nouveau/core/subdev/devinit/nv50.c b/drivers/gpu/drm/nouveau/core/subdev/devinit/nv50.c
index b46c62a1d5d8..968334d1dca4 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/devinit/nv50.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/devinit/nv50.c
@@ -26,6 +26,7 @@
 #include <subdev/bios/dcb.h>
 #include <subdev/bios/disp.h>
 #include <subdev/bios/init.h>
+#include <subdev/ibus.h>
 #include <subdev/vga.h>
 
 #include "nv50.h"
@@ -91,6 +92,7 @@ int
 nv50_devinit_init(struct nouveau_object *object)
 {
 	struct nouveau_bios *bios = nouveau_bios(object);
+	struct nouveau_ibus *ibus = nouveau_ibus(object);
 	struct nv50_devinit_priv *priv = (void *)object;
 	struct nvbios_outp info;
 	struct dcb_output outp;
@@ -105,6 +107,13 @@ nv50_devinit_init(struct nouveau_object *object)
 		}
 	}
 
+	/* some boards appear to require certain priv register timeouts
+	 * to be bumped before runing devinit scripts.  not a clue why
+	 * the vbios engineers didn't make the scripts just work...
+	 */
+	if (priv->base.post && ibus)
+		nv_ofuncs(ibus)->init(nv_object(ibus));
+
 	ret = nouveau_devinit_init(&priv->base);
 	if (ret)
 		return ret;
@@ -160,4 +169,5 @@ nv50_devinit_oclass = &(struct nouveau_devinit_impl) {
 	},
 	.pll_set = nv50_devinit_pll_set,
 	.disable = nv50_devinit_disable,
+	.post = nvbios_init,
 }.base;
diff --git a/drivers/gpu/drm/nouveau/core/subdev/devinit/nv50.h b/drivers/gpu/drm/nouveau/core/subdev/devinit/nv50.h
index 51d5076333ec..f412bb7f780e 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/devinit/nv50.h
+++ b/drivers/gpu/drm/nouveau/core/subdev/devinit/nv50.h
@@ -18,4 +18,6 @@ int  nva3_devinit_pll_set(struct nouveau_devinit *, u32, u32);
 
 int  nvc0_devinit_pll_set(struct nouveau_devinit *, u32, u32);
 
+u64  gm107_devinit_disable(struct nouveau_devinit *);
+
 #endif
diff --git a/drivers/gpu/drm/nouveau/core/subdev/devinit/nv84.c b/drivers/gpu/drm/nouveau/core/subdev/devinit/nv84.c
index 787422505d87..a7c80ded77cd 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/devinit/nv84.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/devinit/nv84.c
@@ -60,4 +60,5 @@ nv84_devinit_oclass = &(struct nouveau_devinit_impl) {
 	},
 	.pll_set = nv50_devinit_pll_set,
 	.disable = nv84_devinit_disable,
+	.post = nvbios_init,
 }.base;
diff --git a/drivers/gpu/drm/nouveau/core/subdev/devinit/nv98.c b/drivers/gpu/drm/nouveau/core/subdev/devinit/nv98.c
index 2b0e963fc6f0..a773253a17f6 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/devinit/nv98.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/devinit/nv98.c
@@ -59,4 +59,5 @@ nv98_devinit_oclass = &(struct nouveau_devinit_impl) {
 	},
 	.pll_set = nv50_devinit_pll_set,
 	.disable = nv98_devinit_disable,
+	.post = nvbios_init,
 }.base;
diff --git a/drivers/gpu/drm/nouveau/core/subdev/devinit/nva3.c b/drivers/gpu/drm/nouveau/core/subdev/devinit/nva3.c
index 006cf348bda7..b9cd9e53f760 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/devinit/nva3.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/devinit/nva3.c
@@ -142,4 +142,5 @@ nva3_devinit_oclass = &(struct nouveau_devinit_impl) {
 	.pll_set = nva3_devinit_pll_set,
 	.disable = nva3_devinit_disable,
 	.mmio    = nva3_devinit_mmio,
+	.post = nvbios_init,
 }.base;
diff --git a/drivers/gpu/drm/nouveau/core/subdev/devinit/nvaf.c b/drivers/gpu/drm/nouveau/core/subdev/devinit/nvaf.c
index 4fc68d27eff3..3729846a8e5c 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/devinit/nvaf.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/devinit/nvaf.c
@@ -60,4 +60,5 @@ nvaf_devinit_oclass = &(struct nouveau_devinit_impl) {
 	},
 	.pll_set = nva3_devinit_pll_set,
 	.disable = nvaf_devinit_disable,
+	.post = nvbios_init,
 }.base;
diff --git a/drivers/gpu/drm/nouveau/core/subdev/devinit/nvc0.c b/drivers/gpu/drm/nouveau/core/subdev/devinit/nvc0.c
index 30c765747eea..80bd7f5eda3d 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/devinit/nvc0.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/devinit/nvc0.c
@@ -115,4 +115,5 @@ nvc0_devinit_oclass = &(struct nouveau_devinit_impl) {
 	},
 	.pll_set = nvc0_devinit_pll_set,
 	.disable = nvc0_devinit_disable,
+	.post = nvbios_init,
 }.base;
diff --git a/drivers/gpu/drm/nouveau/core/subdev/devinit/priv.h b/drivers/gpu/drm/nouveau/core/subdev/devinit/priv.h
index f0e8683ad840..cbcd51852472 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/devinit/priv.h
+++ b/drivers/gpu/drm/nouveau/core/subdev/devinit/priv.h
@@ -3,6 +3,7 @@
 
 #include <subdev/bios.h>
 #include <subdev/bios/pll.h>
+#include <subdev/bios/init.h>
 #include <subdev/clock/pll.h>
 #include <subdev/devinit.h>
 
@@ -12,6 +13,7 @@ struct nouveau_devinit_impl {
 	int  (*pll_set)(struct nouveau_devinit *, u32 type, u32 freq);
 	u64  (*disable)(struct nouveau_devinit *);
 	u32  (*mmio)(struct nouveau_devinit *, u32);
+	int  (*post)(struct nouveau_subdev *, bool);
 };
 
 #define nouveau_devinit_create(p,e,o,d)                                        \
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fb/base.c b/drivers/gpu/drm/nouveau/core/subdev/fb/base.c
index f009d8a39d9d..c866148c440f 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/fb/base.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/fb/base.c
@@ -23,37 +23,30 @@
  */
 
 #include <subdev/bios.h>
-#include <subdev/bios/bit.h>
+#include <subdev/bios/M0203.h>
 
 #include "priv.h"
 
 int
 nouveau_fb_bios_memtype(struct nouveau_bios *bios)
 {
-	struct bit_entry M;
-	u8 ramcfg;
-
-	ramcfg = (nv_rd32(bios, 0x101000) & 0x0000003c) >> 2;
-	if (!bit_entry(bios, 'M', &M) && M.version == 2 && M.length >= 5) {
-		u16 table   = nv_ro16(bios, M.offset + 3);
-		u8  version = nv_ro08(bios, table + 0);
-		u8  header  = nv_ro08(bios, table + 1);
-		u8  record  = nv_ro08(bios, table + 2);
-		u8  entries = nv_ro08(bios, table + 3);
-		if (table && version == 0x10 && ramcfg < entries) {
-			u16 entry = table + header + (ramcfg * record);
-			switch (nv_ro08(bios, entry) & 0x0f) {
-			case 0: return NV_MEM_TYPE_DDR2;
-			case 1: return NV_MEM_TYPE_DDR3;
-			case 2: return NV_MEM_TYPE_GDDR3;
-			case 3: return NV_MEM_TYPE_GDDR5;
-			default:
-				break;
-			}
-
+	const u8 ramcfg = (nv_rd32(bios, 0x101000) & 0x0000003c) >> 2;
+	struct nvbios_M0203E M0203E;
+	u8 ver, hdr;
+
+	if (nvbios_M0203Em(bios, ramcfg, &ver, &hdr, &M0203E)) {
+		switch (M0203E.type) {
+		case M0203E_TYPE_DDR2 : return NV_MEM_TYPE_DDR2;
+		case M0203E_TYPE_DDR3 : return NV_MEM_TYPE_DDR3;
+		case M0203E_TYPE_GDDR3: return NV_MEM_TYPE_GDDR3;
+		case M0203E_TYPE_GDDR5: return NV_MEM_TYPE_GDDR5;
+		default:
+			nv_warn(bios, "M0203E type %02x\n", M0203E.type);
+			return NV_MEM_TYPE_UNKNOWN;
 		}
 	}
 
+	nv_warn(bios, "M0203E not matched!\n");
 	return NV_MEM_TYPE_UNKNOWN;
 }
 
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fb/gddr3.c b/drivers/gpu/drm/nouveau/core/subdev/fb/gddr3.c
new file mode 100644
index 000000000000..d85a25d027ee
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/subdev/fb/gddr3.c
@@ -0,0 +1,117 @@
+/*
+ * Copyright 2013 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs <bskeggs@redhat.com>
+ * 	    Roy Spliet <rspliet@eclipso.eu>
+ */
+
+#include <subdev/bios.h>
+#include "priv.h"
+
+struct ramxlat {
+	int id;
+	u8 enc;
+};
+
+static inline int
+ramxlat(const struct ramxlat *xlat, int id)
+{
+	while (xlat->id >= 0) {
+		if (xlat->id == id)
+			return xlat->enc;
+		xlat++;
+	}
+	return -EINVAL;
+}
+
+static const struct ramxlat
+ramgddr3_cl_lo[] = {
+	{ 7, 7 }, { 8, 0 }, { 9, 1 }, { 10, 2 }, { 11, 3 },
+	/* the below are mentioned in some, but not all, gddr3 docs */
+	{ 12, 4 }, { 13, 5 }, { 14, 6 },
+	/* XXX: Per Samsung docs, are these used? They overlap with Qimonda */
+	/* { 4, 4 }, { 5, 5 }, { 6, 6 }, { 12, 8 }, { 13, 9 }, { 14, 10 },
+	 * { 15, 11 }, */
+	{ -1 }
+};
+
+static const struct ramxlat
+ramgddr3_cl_hi[] = {
+	{ 10, 2 }, { 11, 3 }, { 12, 4 }, { 13, 5 }, { 14, 6 }, { 15, 7 },
+	{ 16, 0 }, { 17, 1 },
+	{ -1 }
+};
+
+static const struct ramxlat
+ramgddr3_wr_lo[] = {
+	{ 5, 2 }, { 7, 4 }, { 8, 5 }, { 9, 6 }, { 10, 7 },
+	{ 11, 0 },
+	/* the below are mentioned in some, but not all, gddr3 docs */
+	{ 4, 1 }, { 6, 3 }, { 12, 1 }, { 13 , 2 },
+	{ -1 }
+};
+
+int
+nouveau_gddr3_calc(struct nouveau_ram *ram)
+{
+	int CL, WR, CWL, DLL = 0, ODT = 0, hi;
+
+	switch (ram->next->bios.timing_ver) {
+	case 0x10:
+		CWL = ram->next->bios.timing_10_CWL;
+		CL  = ram->next->bios.timing_10_CL;
+		WR  = ram->next->bios.timing_10_WR;
+		DLL = !ram->next->bios.ramcfg_10_DLLoff;
+		ODT = ram->next->bios.timing_10_ODT;
+		break;
+	case 0x20:
+		CWL = (ram->next->bios.timing[1] & 0x00000f80) >> 7;
+		CL  = (ram->next->bios.timing[1] & 0x0000001f) >> 0;
+		WR  = (ram->next->bios.timing[2] & 0x007f0000) >> 16;
+		/* XXX: Get these values from the VBIOS instead */
+		DLL = !(ram->mr[1] & 0x1);
+		ODT =  (ram->mr[1] & 0x004) >> 2 |
+		       (ram->mr[1] & 0x040) >> 5 |
+		       (ram->mr[1] & 0x200) >> 7;
+		break;
+	default:
+		return -ENOSYS;
+	}
+
+	hi = ram->mr[2] & 0x1;
+	CL  = ramxlat(hi ? ramgddr3_cl_hi : ramgddr3_cl_lo, CL);
+	WR  = ramxlat(ramgddr3_wr_lo, WR);
+	if (CL < 0 || CWL < 1 || CWL > 7 || WR < 0)
+		return -EINVAL;
+
+	ram->mr[0] &= ~0xf74;
+	ram->mr[0] |= (CWL & 0x07) << 9;
+	ram->mr[0] |= (CL & 0x07) << 4;
+	ram->mr[0] |= (CL & 0x08) >> 1;
+
+	ram->mr[1] &= ~0x3fc;
+	ram->mr[1] |= (ODT & 0x03) << 2;
+	ram->mr[1] |= (ODT & 0x03) << 8;
+	ram->mr[1] |= (WR  & 0x03) << 4;
+	ram->mr[1] |= (WR  & 0x04) << 5;
+	ram->mr[1] |= !DLL << 6;
+	return 0;
+}
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fb/gddr5.c b/drivers/gpu/drm/nouveau/core/subdev/fb/gddr5.c
index 66fe959b4f74..7fbbe05d5c60 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/fb/gddr5.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/fb/gddr5.c
@@ -40,7 +40,7 @@ nouveau_gddr5_calc(struct nouveau_ram *ram, bool nuts)
 	int WL, CL, WR, at[2], dt, ds;
 	int rq = ram->freq < 1000000; /* XXX */
 
-	switch (ram->ramcfg.version) {
+	switch (ram->next->bios.ramcfg_ver) {
 	case 0x11:
 		pd =  ram->next->bios.ramcfg_11_01_80;
 		lf =  ram->next->bios.ramcfg_11_01_40;
@@ -54,7 +54,7 @@ nouveau_gddr5_calc(struct nouveau_ram *ram, bool nuts)
 		return -ENOSYS;
 	}
 
-	switch (ram->timing.version) {
+	switch (ram->next->bios.timing_ver) {
 	case 0x20:
 		WL = (ram->next->bios.timing[1] & 0x00000f80) >> 7;
 		CL = (ram->next->bios.timing[1] & 0x0000001f);
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fb/gk20a.c b/drivers/gpu/drm/nouveau/core/subdev/fb/gk20a.c
index a16024a74771..fde42e4d1b56 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/fb/gk20a.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/fb/gk20a.c
@@ -27,6 +27,20 @@ struct gk20a_fb_priv {
 };
 
 static int
+gk20a_fb_init(struct nouveau_object *object)
+{
+	struct gk20a_fb_priv *priv = (void *)object;
+	int ret;
+
+	ret = nouveau_fb_init(&priv->base);
+	if (ret)
+		return ret;
+
+	nv_mask(priv, 0x100c80, 0x00000001, 0x00000000); /* 128KiB lpg */
+	return 0;
+}
+
+static int
 gk20a_fb_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	     struct nouveau_oclass *oclass, void *data, u32 size,
 	     struct nouveau_object **pobject)
@@ -48,7 +62,7 @@ gk20a_fb_oclass = &(struct nouveau_fb_impl) {
 	.base.ofuncs = &(struct nouveau_ofuncs) {
 		.ctor = gk20a_fb_ctor,
 		.dtor = _nouveau_fb_dtor,
-		.init = _nouveau_fb_init,
+		.init = gk20a_fb_init,
 		.fini = _nouveau_fb_fini,
 	},
 	.memtype = nvc0_fb_memtype_valid,
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fb/nv20.c b/drivers/gpu/drm/nouveau/core/subdev/fb/nv20.c
index f003c1b1893f..2209ade63339 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/fb/nv20.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/fb/nv20.c
@@ -45,7 +45,7 @@ nv20_fb_tile_comp(struct nouveau_fb *pfb, int i, u32 size, u32 flags,
 {
 	u32 tiles = DIV_ROUND_UP(size, 0x40);
 	u32 tags  = round_up(tiles / pfb->ram->parts, 0x40);
-	if (!nouveau_mm_head(&pfb->tags, 1, tags, tags, 1, &tile->tag)) {
+	if (!nouveau_mm_head(&pfb->tags, 0, 1, tags, tags, 1, &tile->tag)) {
 		if (!(flags & 2)) tile->zcomp = 0x00000000; /* Z16 */
 		else              tile->zcomp = 0x04000000; /* Z24S8 */
 		tile->zcomp |= tile->tag->offset;
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fb/nv25.c b/drivers/gpu/drm/nouveau/core/subdev/fb/nv25.c
index f34f4223210b..e2a66c355c50 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/fb/nv25.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/fb/nv25.c
@@ -32,7 +32,7 @@ nv25_fb_tile_comp(struct nouveau_fb *pfb, int i, u32 size, u32 flags,
 {
 	u32 tiles = DIV_ROUND_UP(size, 0x40);
 	u32 tags  = round_up(tiles / pfb->ram->parts, 0x40);
-	if (!nouveau_mm_head(&pfb->tags, 1, tags, tags, 1, &tile->tag)) {
+	if (!nouveau_mm_head(&pfb->tags, 0, 1, tags, tags, 1, &tile->tag)) {
 		if (!(flags & 2)) tile->zcomp = 0x00100000; /* Z16 */
 		else              tile->zcomp = 0x00200000; /* Z24S8 */
 		tile->zcomp |= tile->tag->offset;
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fb/nv30.c b/drivers/gpu/drm/nouveau/core/subdev/fb/nv30.c
index 69093f7151f0..cbec402ba5b9 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/fb/nv30.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/fb/nv30.c
@@ -51,7 +51,7 @@ nv30_fb_tile_comp(struct nouveau_fb *pfb, int i, u32 size, u32 flags,
 {
 	u32 tiles = DIV_ROUND_UP(size, 0x40);
 	u32 tags  = round_up(tiles / pfb->ram->parts, 0x40);
-	if (!nouveau_mm_head(&pfb->tags, 1, tags, tags, 1, &tile->tag)) {
+	if (!nouveau_mm_head(&pfb->tags, 0, 1, tags, tags, 1, &tile->tag)) {
 		if (flags & 2) tile->zcomp |= 0x01000000; /* Z16 */
 		else           tile->zcomp |= 0x02000000; /* Z24S8 */
 		tile->zcomp |= ((tile->tag->offset           ) >> 6);
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fb/nv35.c b/drivers/gpu/drm/nouveau/core/subdev/fb/nv35.c
index 161b06e8fc3f..b2cf8c69fb2e 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/fb/nv35.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/fb/nv35.c
@@ -32,7 +32,7 @@ nv35_fb_tile_comp(struct nouveau_fb *pfb, int i, u32 size, u32 flags,
 {
 	u32 tiles = DIV_ROUND_UP(size, 0x40);
 	u32 tags  = round_up(tiles / pfb->ram->parts, 0x40);
-	if (!nouveau_mm_head(&pfb->tags, 1, tags, tags, 1, &tile->tag)) {
+	if (!nouveau_mm_head(&pfb->tags, 0, 1, tags, tags, 1, &tile->tag)) {
 		if (flags & 2) tile->zcomp |= 0x04000000; /* Z16 */
 		else           tile->zcomp |= 0x08000000; /* Z24S8 */
 		tile->zcomp |= ((tile->tag->offset           ) >> 6);
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fb/nv36.c b/drivers/gpu/drm/nouveau/core/subdev/fb/nv36.c
index 2dd3d0aab6bb..b4cdae2a3b2f 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/fb/nv36.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/fb/nv36.c
@@ -32,7 +32,7 @@ nv36_fb_tile_comp(struct nouveau_fb *pfb, int i, u32 size, u32 flags,
 {
 	u32 tiles = DIV_ROUND_UP(size, 0x40);
 	u32 tags  = round_up(tiles / pfb->ram->parts, 0x40);
-	if (!nouveau_mm_head(&pfb->tags, 1, tags, tags, 1, &tile->tag)) {
+	if (!nouveau_mm_head(&pfb->tags, 0, 1, tags, tags, 1, &tile->tag)) {
 		if (flags & 2) tile->zcomp |= 0x10000000; /* Z16 */
 		else           tile->zcomp |= 0x20000000; /* Z24S8 */
 		tile->zcomp |= ((tile->tag->offset           ) >> 6);
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fb/nv40.c b/drivers/gpu/drm/nouveau/core/subdev/fb/nv40.c
index 95a115ab0c86..52814258c212 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/fb/nv40.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/fb/nv40.c
@@ -33,7 +33,7 @@ nv40_fb_tile_comp(struct nouveau_fb *pfb, int i, u32 size, u32 flags,
 	u32 tiles = DIV_ROUND_UP(size, 0x80);
 	u32 tags  = round_up(tiles / pfb->ram->parts, 0x100);
 	if ( (flags & 2) &&
-	    !nouveau_mm_head(&pfb->tags, 1, tags, tags, 1, &tile->tag)) {
+	    !nouveau_mm_head(&pfb->tags, 0, 1, tags, tags, 1, &tile->tag)) {
 		tile->zcomp  = 0x28000000; /* Z24S8_SPLIT_GRAD */
 		tile->zcomp |= ((tile->tag->offset           ) >> 8);
 		tile->zcomp |= ((tile->tag->offset + tags - 1) >> 8) << 13;
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fb/priv.h b/drivers/gpu/drm/nouveau/core/subdev/fb/priv.h
index 82273f832e42..283863f7aa9b 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/fb/priv.h
+++ b/drivers/gpu/drm/nouveau/core/subdev/fb/priv.h
@@ -35,7 +35,9 @@ extern struct nouveau_oclass nve0_ram_oclass;
 extern struct nouveau_oclass gk20a_ram_oclass;
 extern struct nouveau_oclass gm107_ram_oclass;
 
+int nouveau_sddr2_calc(struct nouveau_ram *ram);
 int nouveau_sddr3_calc(struct nouveau_ram *ram);
+int nouveau_gddr3_calc(struct nouveau_ram *ram);
 int nouveau_gddr5_calc(struct nouveau_ram *ram, bool nuts);
 
 #define nouveau_fb_create(p,e,c,d)                                             \
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fb/ramfuc.h b/drivers/gpu/drm/nouveau/core/subdev/fb/ramfuc.h
index 2af9cfd2c60f..0ac7256443bb 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/fb/ramfuc.h
+++ b/drivers/gpu/drm/nouveau/core/subdev/fb/ramfuc.h
@@ -12,16 +12,32 @@ struct ramfuc {
 struct ramfuc_reg {
 	int sequence;
 	bool force;
-	u32 addr[2];
+	u32 addr;
+	u32 stride; /* in bytes */
+	u32 mask;
 	u32 data;
 };
 
 static inline struct ramfuc_reg
+ramfuc_stride(u32 addr, u32 stride, u32 mask)
+{
+	return (struct ramfuc_reg) {
+		.sequence = 0,
+		.addr = addr,
+		.stride = stride,
+		.mask = mask,
+		.data = 0xdeadbeef,
+	};
+}
+
+static inline struct ramfuc_reg
 ramfuc_reg2(u32 addr1, u32 addr2)
 {
 	return (struct ramfuc_reg) {
 		.sequence = 0,
-		.addr = { addr1, addr2 },
+		.addr = addr1,
+		.stride = addr2 - addr1,
+		.mask = 0x3,
 		.data = 0xdeadbeef,
 	};
 }
@@ -29,7 +45,13 @@ ramfuc_reg2(u32 addr1, u32 addr2)
 static noinline struct ramfuc_reg
 ramfuc_reg(u32 addr)
 {
-	return ramfuc_reg2(addr, addr);
+	return (struct ramfuc_reg) {
+		.sequence = 0,
+		.addr = addr,
+		.stride = 0,
+		.mask = 0x1,
+		.data = 0xdeadbeef,
+	};
 }
 
 static inline int
@@ -62,18 +84,25 @@ static inline u32
 ramfuc_rd32(struct ramfuc *ram, struct ramfuc_reg *reg)
 {
 	if (reg->sequence != ram->sequence)
-		reg->data = nv_rd32(ram->pfb, reg->addr[0]);
+		reg->data = nv_rd32(ram->pfb, reg->addr);
 	return reg->data;
 }
 
 static inline void
 ramfuc_wr32(struct ramfuc *ram, struct ramfuc_reg *reg, u32 data)
 {
+	unsigned int mask, off = 0;
+
 	reg->sequence = ram->sequence;
 	reg->data = data;
-	if (reg->addr[0] != reg->addr[1])
-		nouveau_memx_wr32(ram->memx, reg->addr[1], reg->data);
-	nouveau_memx_wr32(ram->memx, reg->addr[0], reg->data);
+
+	for (mask = reg->mask; mask > 0; mask = (mask & ~1) >> 1) {
+		if (mask & 1) {
+			nouveau_memx_wr32(ram->memx, reg->addr+off, reg->data);
+		}
+
+		off += reg->stride;
+	}
 }
 
 static inline void
@@ -105,14 +134,51 @@ ramfuc_nsec(struct ramfuc *ram, u32 nsec)
 	nouveau_memx_nsec(ram->memx, nsec);
 }
 
-#define ram_init(s,p)       ramfuc_init(&(s)->base, (p))
-#define ram_exec(s,e)       ramfuc_exec(&(s)->base, (e))
-#define ram_have(s,r)       ((s)->r_##r.addr[0] != 0x000000)
-#define ram_rd32(s,r)       ramfuc_rd32(&(s)->base, &(s)->r_##r)
-#define ram_wr32(s,r,d)     ramfuc_wr32(&(s)->base, &(s)->r_##r, (d))
-#define ram_nuke(s,r)       ramfuc_nuke(&(s)->base, &(s)->r_##r)
-#define ram_mask(s,r,m,d)   ramfuc_mask(&(s)->base, &(s)->r_##r, (m), (d))
-#define ram_wait(s,r,m,d,n) ramfuc_wait(&(s)->base, (r), (m), (d), (n))
-#define ram_nsec(s,n)       ramfuc_nsec(&(s)->base, (n))
+static inline void
+ramfuc_wait_vblank(struct ramfuc *ram)
+{
+	nouveau_memx_wait_vblank(ram->memx);
+}
+
+static inline void
+ramfuc_train(struct ramfuc *ram)
+{
+	nouveau_memx_train(ram->memx);
+}
+
+static inline int
+ramfuc_train_result(struct nouveau_fb *pfb, u32 *result, u32 rsize)
+{
+	struct nouveau_pwr *ppwr = nouveau_pwr(pfb);
+
+	return nouveau_memx_train_result(ppwr, result, rsize);
+}
+
+static inline void
+ramfuc_block(struct ramfuc *ram)
+{
+	nouveau_memx_block(ram->memx);
+}
+
+static inline void
+ramfuc_unblock(struct ramfuc *ram)
+{
+	nouveau_memx_unblock(ram->memx);
+}
+
+#define ram_init(s,p)        ramfuc_init(&(s)->base, (p))
+#define ram_exec(s,e)        ramfuc_exec(&(s)->base, (e))
+#define ram_have(s,r)        ((s)->r_##r.addr != 0x000000)
+#define ram_rd32(s,r)        ramfuc_rd32(&(s)->base, &(s)->r_##r)
+#define ram_wr32(s,r,d)      ramfuc_wr32(&(s)->base, &(s)->r_##r, (d))
+#define ram_nuke(s,r)        ramfuc_nuke(&(s)->base, &(s)->r_##r)
+#define ram_mask(s,r,m,d)    ramfuc_mask(&(s)->base, &(s)->r_##r, (m), (d))
+#define ram_wait(s,r,m,d,n)  ramfuc_wait(&(s)->base, (r), (m), (d), (n))
+#define ram_nsec(s,n)        ramfuc_nsec(&(s)->base, (n))
+#define ram_wait_vblank(s)   ramfuc_wait_vblank(&(s)->base)
+#define ram_train(s)         ramfuc_train(&(s)->base)
+#define ram_train_result(s,r,l) ramfuc_train_result((s), (r), (l))
+#define ram_block(s)         ramfuc_block(&(s)->base)
+#define ram_unblock(s)       ramfuc_unblock(&(s)->base)
 
 #endif
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fb/ramnv04.c b/drivers/gpu/drm/nouveau/core/subdev/fb/ramnv04.c
index e781080d3327..1972268d1410 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/fb/ramnv04.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/fb/ramnv04.c
@@ -22,22 +22,7 @@
  * Authors: Ben Skeggs
  */
 
-#define NV04_PFB_BOOT_0						0x00100000
-#	define NV04_PFB_BOOT_0_RAM_AMOUNT			0x00000003
-#	define NV04_PFB_BOOT_0_RAM_AMOUNT_32MB			0x00000000
-#	define NV04_PFB_BOOT_0_RAM_AMOUNT_4MB			0x00000001
-#	define NV04_PFB_BOOT_0_RAM_AMOUNT_8MB			0x00000002
-#	define NV04_PFB_BOOT_0_RAM_AMOUNT_16MB			0x00000003
-#	define NV04_PFB_BOOT_0_RAM_WIDTH_128			0x00000004
-#	define NV04_PFB_BOOT_0_RAM_TYPE				0x00000028
-#	define NV04_PFB_BOOT_0_RAM_TYPE_SGRAM_8MBIT		0x00000000
-#	define NV04_PFB_BOOT_0_RAM_TYPE_SGRAM_16MBIT		0x00000008
-#	define NV04_PFB_BOOT_0_RAM_TYPE_SGRAM_16MBIT_4BANK	0x00000010
-#	define NV04_PFB_BOOT_0_RAM_TYPE_SDRAM_16MBIT		0x00000018
-#	define NV04_PFB_BOOT_0_RAM_TYPE_SDRAM_64MBIT		0x00000020
-#	define NV04_PFB_BOOT_0_RAM_TYPE_SDRAM_64MBITX16		0x00000028
-#	define NV04_PFB_BOOT_0_UMA_ENABLE			0x00000100
-#	define NV04_PFB_BOOT_0_UMA_SIZE				0x0000f000
+#include <subdev/fb/regsnv04.h>
 
 #include "priv.h"
 
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fb/ramnv50.c b/drivers/gpu/drm/nouveau/core/subdev/fb/ramnv50.c
index e5d12c24cc43..64a983c96625 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/fb/ramnv50.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/fb/ramnv50.c
@@ -280,7 +280,7 @@ nv50_ram_get(struct nouveau_fb *pfb, u64 size, u32 align, u32 ncmin,
 		if (align == 16) {
 			int n = (max >> 4) * comp;
 
-			ret = nouveau_mm_head(tags, 1, n, n, 1, &mem->tag);
+			ret = nouveau_mm_head(tags, 0, 1, n, n, 1, &mem->tag);
 			if (ret)
 				mem->tag = NULL;
 		}
@@ -296,9 +296,9 @@ nv50_ram_get(struct nouveau_fb *pfb, u64 size, u32 align, u32 ncmin,
 	type = nv50_fb_memtype[type];
 	do {
 		if (back)
-			ret = nouveau_mm_tail(heap, type, max, min, align, &r);
+			ret = nouveau_mm_tail(heap, 0, type, max, min, align, &r);
 		else
-			ret = nouveau_mm_head(heap, type, max, min, align, &r);
+			ret = nouveau_mm_head(heap, 0, type, max, min, align, &r);
 		if (ret) {
 			mutex_unlock(&pfb->base.mutex);
 			pfb->ram->put(pfb, &mem);
@@ -319,27 +319,22 @@ nv50_ram_get(struct nouveau_fb *pfb, u64 size, u32 align, u32 ncmin,
 static u32
 nv50_fb_vram_rblock(struct nouveau_fb *pfb, struct nouveau_ram *ram)
 {
-	int i, parts, colbits, rowbitsa, rowbitsb, banks;
+	int colbits, rowbitsa, rowbitsb, banks;
 	u64 rowsize, predicted;
-	u32 r0, r4, rt, ru, rblock_size;
+	u32 r0, r4, rt, rblock_size;
 
 	r0 = nv_rd32(pfb, 0x100200);
 	r4 = nv_rd32(pfb, 0x100204);
 	rt = nv_rd32(pfb, 0x100250);
-	ru = nv_rd32(pfb, 0x001540);
-	nv_debug(pfb, "memcfg 0x%08x 0x%08x 0x%08x 0x%08x\n", r0, r4, rt, ru);
-
-	for (i = 0, parts = 0; i < 8; i++) {
-		if (ru & (0x00010000 << i))
-			parts++;
-	}
+	nv_debug(pfb, "memcfg 0x%08x 0x%08x 0x%08x 0x%08x\n", r0, r4, rt,
+			nv_rd32(pfb, 0x001540));
 
 	colbits  =  (r4 & 0x0000f000) >> 12;
 	rowbitsa = ((r4 & 0x000f0000) >> 16) + 8;
 	rowbitsb = ((r4 & 0x00f00000) >> 20) + 8;
 	banks    = 1 << (((r4 & 0x03000000) >> 24) + 2);
 
-	rowsize = parts * banks * (1 << colbits) * 8;
+	rowsize = ram->parts * banks * (1 << colbits) * 8;
 	predicted = rowsize << rowbitsa;
 	if (r0 & 0x00000004)
 		predicted += rowsize << rowbitsb;
@@ -376,6 +371,9 @@ nv50_ram_create_(struct nouveau_object *parent, struct nouveau_object *engine,
 	ram->size = nv_rd32(pfb, 0x10020c);
 	ram->size = (ram->size & 0xffffff00) | ((ram->size & 0x000000ff) << 32);
 
+	ram->part_mask = (nv_rd32(pfb, 0x001540) & 0x00ff0000) >> 16;
+	ram->parts = hweight8(ram->part_mask);
+
 	switch (nv_rd32(pfb, 0x100714) & 0x00000007) {
 	case 0: ram->type = NV_MEM_TYPE_DDR1; break;
 	case 1:
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fb/ramnva3.c b/drivers/gpu/drm/nouveau/core/subdev/fb/ramnva3.c
index 8076fb195dd5..3b38a538845d 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/fb/ramnva3.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/fb/ramnva3.c
@@ -20,79 +20,512 @@
  * OTHER DEALINGS IN THE SOFTWARE.
  *
  * Authors: Ben Skeggs
+ * 	    Roy Spliet <rspliet@eclipso.eu>
  */
 
 #include <subdev/bios.h>
 #include <subdev/bios/bit.h>
 #include <subdev/bios/pll.h>
 #include <subdev/bios/rammap.h>
+#include <subdev/bios/M0205.h>
 #include <subdev/bios/timing.h>
 
 #include <subdev/clock/nva3.h>
 #include <subdev/clock/pll.h>
 
+#include <subdev/gpio.h>
+
+#include <subdev/timer.h>
+
+#include <engine/fifo.h>
+
 #include <core/option.h>
 
 #include "ramfuc.h"
 
 #include "nv50.h"
 
+/* XXX: Remove when memx gains GPIO support */
+extern int nv50_gpio_location(int line, u32 *reg, u32 *shift);
+
 struct nva3_ramfuc {
 	struct ramfuc base;
+	struct ramfuc_reg r_0x001610;
+	struct ramfuc_reg r_0x001700;
+	struct ramfuc_reg r_0x002504;
 	struct ramfuc_reg r_0x004000;
 	struct ramfuc_reg r_0x004004;
 	struct ramfuc_reg r_0x004018;
 	struct ramfuc_reg r_0x004128;
 	struct ramfuc_reg r_0x004168;
+	struct ramfuc_reg r_0x100080;
 	struct ramfuc_reg r_0x100200;
 	struct ramfuc_reg r_0x100210;
 	struct ramfuc_reg r_0x100220[9];
+	struct ramfuc_reg r_0x100264;
 	struct ramfuc_reg r_0x1002d0;
 	struct ramfuc_reg r_0x1002d4;
 	struct ramfuc_reg r_0x1002dc;
 	struct ramfuc_reg r_0x10053c;
 	struct ramfuc_reg r_0x1005a0;
 	struct ramfuc_reg r_0x1005a4;
+	struct ramfuc_reg r_0x100700;
 	struct ramfuc_reg r_0x100714;
 	struct ramfuc_reg r_0x100718;
 	struct ramfuc_reg r_0x10071c;
+	struct ramfuc_reg r_0x100720;
 	struct ramfuc_reg r_0x100760;
 	struct ramfuc_reg r_0x1007a0;
 	struct ramfuc_reg r_0x1007e0;
+	struct ramfuc_reg r_0x100da0;
 	struct ramfuc_reg r_0x10f804;
 	struct ramfuc_reg r_0x1110e0;
 	struct ramfuc_reg r_0x111100;
 	struct ramfuc_reg r_0x111104;
+	struct ramfuc_reg r_0x1111e0;
+	struct ramfuc_reg r_0x111400;
 	struct ramfuc_reg r_0x611200;
 	struct ramfuc_reg r_mr[4];
+	struct ramfuc_reg r_gpioFBVREF;
+};
+
+struct nva3_ltrain {
+	enum {
+		NVA3_TRAIN_UNKNOWN,
+		NVA3_TRAIN_UNSUPPORTED,
+		NVA3_TRAIN_ONCE,
+		NVA3_TRAIN_EXEC,
+		NVA3_TRAIN_DONE
+	} state;
+	u32 r_100720;
+	u32 r_1111e0;
+	u32 r_111400;
+	struct nouveau_mem *mem;
 };
 
 struct nva3_ram {
 	struct nouveau_ram base;
 	struct nva3_ramfuc fuc;
+	struct nva3_ltrain ltrain;
 };
 
+void
+nva3_link_train_calc(u32 *vals, struct nva3_ltrain *train)
+{
+	int i, lo, hi;
+	u8 median[8], bins[4] = {0, 0, 0, 0}, bin = 0, qty = 0;
+
+	for (i = 0; i < 8; i++) {
+		for (lo = 0; lo < 0x40; lo++) {
+			if (!(vals[lo] & 0x80000000))
+				continue;
+			if (vals[lo] & (0x101 << i))
+				break;
+		}
+
+		if (lo == 0x40)
+			return;
+
+		for (hi = lo + 1; hi < 0x40; hi++) {
+			if (!(vals[lo] & 0x80000000))
+				continue;
+			if (!(vals[hi] & (0x101 << i))) {
+				hi--;
+				break;
+			}
+		}
+
+		median[i] = ((hi - lo) >> 1) + lo;
+		bins[(median[i] & 0xf0) >> 4]++;
+		median[i] += 0x30;
+	}
+
+	/* Find the best value for 0x1111e0 */
+	for (i = 0; i < 4; i++) {
+		if (bins[i] > qty) {
+			bin = i + 3;
+			qty = bins[i];
+		}
+	}
+
+	train->r_100720 = 0;
+	for (i = 0; i < 8; i++) {
+		median[i] = max(median[i], (u8) (bin << 4));
+		median[i] = min(median[i], (u8) ((bin << 4) | 0xf));
+
+		train->r_100720 |= ((median[i] & 0x0f) << (i << 2));
+	}
+
+	train->r_1111e0 = 0x02000000 | (bin * 0x101);
+	train->r_111400 = 0x0;
+}
+
+/*
+ * Link training for (at least) DDR3
+ */
+int
+nva3_link_train(struct nouveau_fb *pfb)
+{
+	struct nouveau_bios *bios = nouveau_bios(pfb);
+	struct nva3_ram *ram = (void *)pfb->ram;
+	struct nouveau_clock *clk = nouveau_clock(pfb);
+	struct nva3_ltrain *train = &ram->ltrain;
+	struct nouveau_device *device = nv_device(pfb);
+	struct nva3_ramfuc *fuc = &ram->fuc;
+	u32 *result, r1700;
+	int ret, i;
+	struct nvbios_M0205T M0205T = { 0 };
+	u8 ver, hdr, cnt, len, snr, ssz;
+	unsigned int clk_current;
+	unsigned long flags;
+	unsigned long *f = &flags;
+
+	if (nouveau_boolopt(device->cfgopt, "NvMemExec", true) != true)
+		return -ENOSYS;
+
+	/* XXX: Multiple partitions? */
+	result = kmalloc(64 * sizeof(u32), GFP_KERNEL);
+	if (!result)
+		return -ENOMEM;
+
+	train->state = NVA3_TRAIN_EXEC;
+
+	/* Clock speeds for training and back */
+	nvbios_M0205Tp(bios, &ver, &hdr, &cnt, &len, &snr, &ssz, &M0205T);
+	if (M0205T.freq == 0)
+		return -ENOENT;
+
+	clk_current = clk->read(clk, nv_clk_src_mem);
+
+	ret = nva3_clock_pre(clk, f);
+	if (ret)
+		goto out;
+
+	/* First: clock up/down */
+	ret = ram->base.calc(pfb, (u32) M0205T.freq * 1000);
+	if (ret)
+		goto out;
+
+	/* Do this *after* calc, eliminates write in script */
+	nv_wr32(pfb, 0x111400, 0x00000000);
+	/* XXX: Magic writes that improve train reliability? */
+	nv_mask(pfb, 0x100674, 0x0000ffff, 0x00000000);
+	nv_mask(pfb, 0x1005e4, 0x0000ffff, 0x00000000);
+	nv_mask(pfb, 0x100b0c, 0x000000ff, 0x00000000);
+	nv_wr32(pfb, 0x100c04, 0x00000400);
+
+	/* Now the training script */
+	r1700 = ram_rd32(fuc, 0x001700);
+
+	ram_mask(fuc, 0x100200, 0x00000800, 0x00000000);
+	ram_wr32(fuc, 0x611200, 0x3300);
+	ram_wait_vblank(fuc);
+	ram_wait(fuc, 0x611200, 0x00000003, 0x00000000, 500000);
+	ram_mask(fuc, 0x001610, 0x00000083, 0x00000003);
+	ram_mask(fuc, 0x100080, 0x00000020, 0x00000000);
+	ram_mask(fuc, 0x10f804, 0x80000000, 0x00000000);
+	ram_wr32(fuc, 0x001700, 0x00000000);
+
+	ram_train(fuc);
+
+	/* Reset */
+	ram_mask(fuc, 0x10f804, 0x80000000, 0x80000000);
+	ram_wr32(fuc, 0x10053c, 0x0);
+	ram_wr32(fuc, 0x100720, train->r_100720);
+	ram_wr32(fuc, 0x1111e0, train->r_1111e0);
+	ram_wr32(fuc, 0x111400, train->r_111400);
+	ram_nuke(fuc, 0x100080);
+	ram_mask(fuc, 0x100080, 0x00000020, 0x00000020);
+	ram_nsec(fuc, 1000);
+
+	ram_wr32(fuc, 0x001700, r1700);
+	ram_mask(fuc, 0x001610, 0x00000083, 0x00000080);
+	ram_wr32(fuc, 0x611200, 0x3330);
+	ram_mask(fuc, 0x100200, 0x00000800, 0x00000800);
+
+	ram_exec(fuc, true);
+
+	ram->base.calc(pfb, clk_current);
+	ram_exec(fuc, true);
+
+	/* Post-processing, avoids flicker */
+	nv_mask(pfb, 0x616308, 0x10, 0x10);
+	nv_mask(pfb, 0x616b08, 0x10, 0x10);
+
+	nva3_clock_post(clk, f);
+
+	ram_train_result(pfb, result, 64);
+	for (i = 0; i < 64; i++)
+		nv_debug(pfb, "Train: %08x", result[i]);
+	nva3_link_train_calc(result, train);
+
+	nv_debug(pfb, "Train: %08x %08x %08x", train->r_100720,
+			train->r_1111e0, train->r_111400);
+
+	kfree(result);
+
+	train->state = NVA3_TRAIN_DONE;
+
+	return ret;
+
+out:
+	if(ret == -EBUSY)
+		f = NULL;
+
+	train->state = NVA3_TRAIN_UNSUPPORTED;
+
+	nva3_clock_post(clk, f);
+	return ret;
+}
+
+int
+nva3_link_train_init(struct nouveau_fb *pfb)
+{
+	static const u32 pattern[16] = {
+		0xaaaaaaaa, 0xcccccccc, 0xdddddddd, 0xeeeeeeee,
+		0x00000000, 0x11111111, 0x44444444, 0xdddddddd,
+		0x33333333, 0x55555555, 0x77777777, 0x66666666,
+		0x99999999, 0x88888888, 0xeeeeeeee, 0xbbbbbbbb,
+	};
+	struct nouveau_bios *bios = nouveau_bios(pfb);
+	struct nva3_ram *ram = (void *)pfb->ram;
+	struct nva3_ltrain *train = &ram->ltrain;
+	struct nouveau_mem *mem;
+	struct nvbios_M0205E M0205E;
+	u8 ver, hdr, cnt, len;
+	u32 r001700;
+	int ret, i = 0;
+
+	train->state = NVA3_TRAIN_UNSUPPORTED;
+
+	/* We support type "5"
+	 * XXX: training pattern table appears to be unused for this routine */
+	if (!nvbios_M0205Ep(bios, i, &ver, &hdr, &cnt, &len, &M0205E))
+		return -ENOENT;
+
+	if (M0205E.type != 5)
+		return 0;
+
+	train->state = NVA3_TRAIN_ONCE;
+
+	ret = pfb->ram->get(pfb, 0x8000, 0x10000, 0, 0x800, &ram->ltrain.mem);
+	if (ret)
+		return ret;
+
+	mem = ram->ltrain.mem;
+
+	nv_wr32(pfb, 0x100538, 0x10000000 | (mem->offset >> 16));
+	nv_wr32(pfb, 0x1005a8, 0x0000ffff);
+	nv_mask(pfb, 0x10f800, 0x00000001, 0x00000001);
+
+	for (i = 0; i < 0x30; i++) {
+		nv_wr32(pfb, 0x10f8c0, (i << 8) | i);
+		nv_wr32(pfb, 0x10f900, pattern[i % 16]);
+	}
+
+	for (i = 0; i < 0x30; i++) {
+		nv_wr32(pfb, 0x10f8e0, (i << 8) | i);
+		nv_wr32(pfb, 0x10f920, pattern[i % 16]);
+	}
+
+	/* And upload the pattern */
+	r001700 = nv_rd32(pfb, 0x1700);
+	nv_wr32(pfb, 0x1700, mem->offset >> 16);
+	for (i = 0; i < 16; i++)
+		nv_wr32(pfb, 0x700000 + (i << 2), pattern[i]);
+	for (i = 0; i < 16; i++)
+		nv_wr32(pfb, 0x700100 + (i << 2), pattern[i]);
+	nv_wr32(pfb, 0x1700, r001700);
+
+	train->r_100720 = nv_rd32(pfb, 0x100720);
+	train->r_1111e0 = nv_rd32(pfb, 0x1111e0);
+	train->r_111400 = nv_rd32(pfb, 0x111400);
+
+	return 0;
+}
+
+void
+nva3_link_train_fini(struct nouveau_fb *pfb)
+{
+	struct nva3_ram *ram = (void *)pfb->ram;
+
+	if (ram->ltrain.mem)
+		pfb->ram->put(pfb, &ram->ltrain.mem);
+}
+
+/*
+ * RAM reclocking
+ */
+#define T(t) cfg->timing_10_##t
+static int
+nva3_ram_timing_calc(struct nouveau_fb *pfb, u32 *timing)
+{
+	struct nva3_ram *ram = (void *)pfb->ram;
+	struct nvbios_ramcfg *cfg = &ram->base.target.bios;
+	int tUNK_base, tUNK_40_0, prevCL;
+	u32 cur2, cur3, cur7, cur8;
+
+	cur2 = nv_rd32(pfb, 0x100228);
+	cur3 = nv_rd32(pfb, 0x10022c);
+	cur7 = nv_rd32(pfb, 0x10023c);
+	cur8 = nv_rd32(pfb, 0x100240);
+
+
+	switch ((!T(CWL)) * ram->base.type) {
+	case NV_MEM_TYPE_DDR2:
+		T(CWL) = T(CL) - 1;
+		break;
+	case NV_MEM_TYPE_GDDR3:
+		T(CWL) = ((cur2 & 0xff000000) >> 24) + 1;
+		break;
+	}
+
+	prevCL = (cur3 & 0x000000ff) + 1;
+	tUNK_base = ((cur7 & 0x00ff0000) >> 16) - prevCL;
+
+	timing[0] = (T(RP) << 24 | T(RAS) << 16 | T(RFC) << 8 | T(RC));
+	timing[1] = (T(WR) + 1 + T(CWL)) << 24 |
+		    max_t(u8,T(18), 1) << 16 |
+		    (T(WTR) + 1 + T(CWL)) << 8 |
+		    (5 + T(CL) - T(CWL));
+	timing[2] = (T(CWL) - 1) << 24 |
+		    (T(RRD) << 16) |
+		    (T(RCDWR) << 8) |
+		    T(RCDRD);
+	timing[3] = (cur3 & 0x00ff0000) |
+		    (0x30 + T(CL)) << 24 |
+		    (0xb + T(CL)) << 8 |
+		    (T(CL) - 1);
+	timing[4] = T(20) << 24 |
+		    T(21) << 16 |
+		    T(13) << 8 |
+		    T(13);
+	timing[5] = T(RFC) << 24 |
+		    max_t(u8,T(RCDRD), T(RCDWR)) << 16 |
+		    max_t(u8, (T(CWL) + 6), (T(CL) + 2)) << 8 |
+		    T(RP);
+	timing[6] = (0x5a + T(CL)) << 16 |
+		    max_t(u8, 1, (6 - T(CL) + T(CWL))) << 8 |
+		    (0x50 + T(CL) - T(CWL));
+	timing[7] = (cur7 & 0xff000000) |
+		    ((tUNK_base + T(CL)) << 16) |
+		    0x202;
+	timing[8] = cur8 & 0xffffff00;
+
+	switch (ram->base.type) {
+	case NV_MEM_TYPE_DDR2:
+	case NV_MEM_TYPE_GDDR3:
+		tUNK_40_0 = prevCL - (cur8 & 0xff);
+		if (tUNK_40_0 > 0)
+			timing[8] |= T(CL);
+		break;
+	default:
+		break;
+	}
+
+	nv_debug(pfb, "Entry: 220: %08x %08x %08x %08x\n",
+			timing[0], timing[1], timing[2], timing[3]);
+	nv_debug(pfb, "  230: %08x %08x %08x %08x\n",
+			timing[4], timing[5], timing[6], timing[7]);
+	nv_debug(pfb, "  240: %08x\n", timing[8]);
+	return 0;
+}
+#undef T
+
+static void
+nouveau_sddr2_dll_reset(struct nva3_ramfuc *fuc)
+{
+	ram_mask(fuc, mr[0], 0x100, 0x100);
+	ram_nsec(fuc, 1000);
+	ram_mask(fuc, mr[0], 0x100, 0x000);
+	ram_nsec(fuc, 1000);
+}
+
+static void
+nouveau_sddr3_dll_disable(struct nva3_ramfuc *fuc, u32 *mr)
+{
+	u32 mr1_old = ram_rd32(fuc, mr[1]);
+
+	if (!(mr1_old & 0x1)) {
+		ram_wr32(fuc, 0x1002d4, 0x00000001);
+		ram_wr32(fuc, mr[1], mr[1]);
+		ram_nsec(fuc, 1000);
+	}
+}
+
+static void
+nouveau_gddr3_dll_disable(struct nva3_ramfuc *fuc, u32 *mr)
+{
+	u32 mr1_old = ram_rd32(fuc, mr[1]);
+
+	if (!(mr1_old & 0x40)) {
+		ram_wr32(fuc, mr[1], mr[1]);
+		ram_nsec(fuc, 1000);
+	}
+}
+
+static void
+nva3_ram_lock_pll(struct nva3_ramfuc *fuc, struct nva3_clock_info *mclk)
+{
+	ram_wr32(fuc, 0x004004, mclk->pll);
+	ram_mask(fuc, 0x004000, 0x00000001, 0x00000001);
+	ram_mask(fuc, 0x004000, 0x00000010, 0x00000000);
+	ram_wait(fuc, 0x004000, 0x00020000, 0x00020000, 64000);
+	ram_mask(fuc, 0x004000, 0x00000010, 0x00000010);
+}
+
+static void
+nva3_ram_fbvref(struct nva3_ramfuc *fuc, u32 val)
+{
+	struct nouveau_gpio *gpio = nouveau_gpio(fuc->base.pfb);
+	struct dcb_gpio_func func;
+	u32 reg, sh, gpio_val;
+	int ret;
+
+	if (gpio->get(gpio, 0, 0x2e, DCB_GPIO_UNUSED) != val) {
+		ret = gpio->find(gpio, 0, 0x2e, DCB_GPIO_UNUSED, &func);
+		if (ret)
+			return;
+
+		nv50_gpio_location(func.line, &reg, &sh);
+		gpio_val = ram_rd32(fuc, gpioFBVREF);
+		if (gpio_val & (8 << sh))
+			val = !val;
+
+		ram_mask(fuc, gpioFBVREF, (0x3 << sh), ((val | 0x2) << sh));
+		ram_nsec(fuc, 20000);
+	}
+}
+
 static int
 nva3_ram_calc(struct nouveau_fb *pfb, u32 freq)
 {
 	struct nouveau_bios *bios = nouveau_bios(pfb);
 	struct nva3_ram *ram = (void *)pfb->ram;
 	struct nva3_ramfuc *fuc = &ram->fuc;
+	struct nva3_ltrain *train = &ram->ltrain;
 	struct nva3_clock_info mclk;
-	u8  ver, cnt, len, strap;
+	struct nouveau_ram_data *next;
+	u8  ver, hdr, cnt, len, strap;
 	u32 data;
-	struct {
-		u32 data;
-		u8  size;
-	} rammap, ramcfg, timing;
-	u32 r004018, r100760, ctrl;
+	u32 r004018, r100760, r100da0, r111100, ctrl;
 	u32 unk714, unk718, unk71c;
-	int ret;
+	int ret, i;
+	u32 timing[9];
+	bool pll2pll;
+
+	next = &ram->base.target;
+	next->freq = freq;
+	ram->base.next = next;
+
+	if (ram->ltrain.state == NVA3_TRAIN_ONCE)
+		nva3_link_train(pfb);
 
 	/* lookup memory config data relevant to the target frequency */
-	rammap.data = nvbios_rammapEm(bios, freq / 1000, &ver, &rammap.size,
-				     &cnt, &ramcfg.size);
-	if (!rammap.data || ver != 0x10 || rammap.size < 0x0e) {
+	i = 0;
+	data = nvbios_rammapEm(bios, freq / 1000, &ver, &hdr, &cnt, &len,
+				      &next->bios);
+	if (!data || ver != 0x10 || hdr < 0x05) {
 		nv_error(pfb, "invalid/missing rammap entry\n");
 		return -EINVAL;
 	}
@@ -104,78 +537,123 @@ nva3_ram_calc(struct nouveau_fb *pfb, u32 freq)
 		return -EINVAL;
 	}
 
-	ramcfg.data = rammap.data + rammap.size + (strap * ramcfg.size);
-	if (!ramcfg.data || ver != 0x10 || ramcfg.size < 0x0e) {
+	data = nvbios_rammapSp(bios, data, ver, hdr, cnt, len, strap,
+			       &ver, &hdr, &next->bios);
+	if (!data || ver != 0x10 || hdr < 0x09) {
 		nv_error(pfb, "invalid/missing ramcfg entry\n");
 		return -EINVAL;
 	}
 
 	/* lookup memory timings, if bios says they're present */
-	strap = nv_ro08(bios, ramcfg.data + 0x01);
-	if (strap != 0xff) {
-		timing.data = nvbios_timingEe(bios, strap, &ver, &timing.size,
-					     &cnt, &len);
-		if (!timing.data || ver != 0x10 || timing.size < 0x19) {
+	if (next->bios.ramcfg_timing != 0xff) {
+		data = nvbios_timingEp(bios, next->bios.ramcfg_timing,
+				       &ver, &hdr, &cnt, &len,
+				       &next->bios);
+		if (!data || ver != 0x10 || hdr < 0x17) {
 			nv_error(pfb, "invalid/missing timing entry\n");
 			return -EINVAL;
 		}
-	} else {
-		timing.data = 0;
 	}
 
-	ret = nva3_clock_info(nouveau_clock(pfb), 0x12, 0x4000, freq, &mclk);
+	ret = nva3_pll_info(nouveau_clock(pfb), 0x12, 0x4000, freq, &mclk);
 	if (ret < 0) {
 		nv_error(pfb, "failed mclk calculation\n");
 		return ret;
 	}
 
+	nva3_ram_timing_calc(pfb, timing);
+
 	ret = ram_init(fuc, pfb);
 	if (ret)
 		return ret;
 
+	/* Determine ram-specific MR values */
+	ram->base.mr[0] = ram_rd32(fuc, mr[0]);
+	ram->base.mr[1] = ram_rd32(fuc, mr[1]);
+	ram->base.mr[2] = ram_rd32(fuc, mr[2]);
+
+	switch (ram->base.type) {
+	case NV_MEM_TYPE_DDR2:
+		ret = nouveau_sddr2_calc(&ram->base);
+		break;
+	case NV_MEM_TYPE_DDR3:
+		ret = nouveau_sddr3_calc(&ram->base);
+		break;
+	case NV_MEM_TYPE_GDDR3:
+		ret = nouveau_gddr3_calc(&ram->base);
+		break;
+	default:
+		ret = -ENOSYS;
+		break;
+	}
+
+	if (ret)
+		return ret;
+
 	/* XXX: where the fuck does 750MHz come from? */
 	if (freq <= 750000) {
 		r004018 = 0x10000000;
 		r100760 = 0x22222222;
+		r100da0 = 0x00000010;
 	} else {
 		r004018 = 0x00000000;
 		r100760 = 0x00000000;
+		r100da0 = 0x00000000;
 	}
 
+	if (!next->bios.ramcfg_10_DLLoff)
+		r004018 |= 0x00004000;
+
+	/* pll2pll requires to switch to a safe clock first */
 	ctrl = ram_rd32(fuc, 0x004000);
-	if (ctrl & 0x00000008) {
-		if (mclk.pll) {
-			ram_mask(fuc, 0x004128, 0x00000101, 0x00000101);
-			ram_wr32(fuc, 0x004004, mclk.pll);
-			ram_wr32(fuc, 0x004000, (ctrl |= 0x00000001));
-			ram_wr32(fuc, 0x004000, (ctrl &= 0xffffffef));
-			ram_wait(fuc, 0x004000, 0x00020000, 0x00020000, 64000);
-			ram_wr32(fuc, 0x004000, (ctrl |= 0x00000010));
-			ram_wr32(fuc, 0x004018, 0x00005000 | r004018);
-			ram_wr32(fuc, 0x004000, (ctrl |= 0x00000004));
-		}
-	} else {
-		u32 ssel = 0x00000101;
-		if (mclk.clk)
-			ssel |= mclk.clk;
-		else
-			ssel |= 0x00080000; /* 324MHz, shouldn't matter... */
-		ram_mask(fuc, 0x004168, 0x003f3141, ctrl);
-	}
+	pll2pll = (!(ctrl & 0x00000008)) && mclk.pll;
 
-	if ( (nv_ro08(bios, ramcfg.data + 0x02) & 0x10)) {
+	/* Pre, NVIDIA does this outside the script */
+	if (next->bios.ramcfg_10_02_10) {
 		ram_mask(fuc, 0x111104, 0x00000600, 0x00000000);
 	} else {
 		ram_mask(fuc, 0x111100, 0x40000000, 0x40000000);
 		ram_mask(fuc, 0x111104, 0x00000180, 0x00000000);
 	}
+	/* Always disable this bit during reclock */
+	ram_mask(fuc, 0x100200, 0x00000800, 0x00000000);
 
-	if (!(nv_ro08(bios, rammap.data + 0x04) & 0x02))
-		ram_mask(fuc, 0x100200, 0x00000800, 0x00000000);
-	ram_wr32(fuc, 0x611200, 0x00003300);
-	if (!(nv_ro08(bios, ramcfg.data + 0x02) & 0x10))
-		ram_wr32(fuc, 0x111100, 0x4c020000); /*XXX*/
+	/* If switching from non-pll to pll, lock before disabling FB */
+	if (mclk.pll && !pll2pll) {
+		ram_mask(fuc, 0x004128, 0x003f3141, mclk.clk | 0x00000101);
+		nva3_ram_lock_pll(fuc, &mclk);
+	}
+
+	/* Start with disabling some CRTCs and PFIFO? */
+	ram_wait_vblank(fuc);
+	ram_wr32(fuc, 0x611200, 0x3300);
+	ram_mask(fuc, 0x002504, 0x1, 0x1);
+	ram_nsec(fuc, 10000);
+	ram_wait(fuc, 0x002504, 0x10, 0x10, 20000); /* XXX: or longer? */
+	ram_block(fuc);
+	ram_nsec(fuc, 2000);
 
+	if (!next->bios.ramcfg_10_02_10) {
+		if (ram->base.type == NV_MEM_TYPE_GDDR3)
+			ram_mask(fuc, 0x111100, 0x04020000, 0x00020000);
+		else
+			ram_mask(fuc, 0x111100, 0x04020000, 0x04020000);
+	}
+
+	/* If we're disabling the DLL, do it now */
+	switch (next->bios.ramcfg_10_DLLoff * ram->base.type) {
+	case NV_MEM_TYPE_DDR3:
+		nouveau_sddr3_dll_disable(fuc, ram->base.mr);
+		break;
+	case NV_MEM_TYPE_GDDR3:
+		nouveau_gddr3_dll_disable(fuc, ram->base.mr);
+		break;
+	}
+
+	if (fuc->r_gpioFBVREF.addr && next->bios.timing_10_ODT)
+		nva3_ram_fbvref(fuc, 0);
+
+	/* Brace RAM for impact */
 	ram_wr32(fuc, 0x1002d4, 0x00000001);
 	ram_wr32(fuc, 0x1002d0, 0x00000001);
 	ram_wr32(fuc, 0x1002d0, 0x00000001);
@@ -183,38 +661,57 @@ nva3_ram_calc(struct nouveau_fb *pfb, u32 freq)
 	ram_wr32(fuc, 0x1002dc, 0x00000001);
 	ram_nsec(fuc, 2000);
 
-	ctrl = ram_rd32(fuc, 0x004000);
-	if (!(ctrl & 0x00000008) && mclk.pll) {
-		ram_wr32(fuc, 0x004000, (ctrl |=  0x00000008));
+	if (nv_device(pfb)->chipset == 0xa3 && freq <= 500000)
+		ram_mask(fuc, 0x100700, 0x00000006, 0x00000006);
+
+	/* Fiddle with clocks */
+	/* There's 4 scenario's
+	 * pll->pll: first switch to a 324MHz clock, set up new PLL, switch
+	 * clk->pll: Set up new PLL, switch
+	 * pll->clk: Set up clock, switch
+	 * clk->clk: Overwrite ctrl and other bits, switch */
+
+	/* Switch to regular clock - 324MHz */
+	if (pll2pll) {
+		ram_mask(fuc, 0x004000, 0x00000004, 0x00000004);
+		ram_mask(fuc, 0x004168, 0x003f3141, 0x00083101);
+		ram_mask(fuc, 0x004000, 0x00000008, 0x00000008);
 		ram_mask(fuc, 0x1110e0, 0x00088000, 0x00088000);
 		ram_wr32(fuc, 0x004018, 0x00001000);
-		ram_wr32(fuc, 0x004000, (ctrl &= ~0x00000001));
-		ram_wr32(fuc, 0x004004, mclk.pll);
-		ram_wr32(fuc, 0x004000, (ctrl |=  0x00000001));
-		udelay(64);
-		ram_wr32(fuc, 0x004018, 0x00005000 | r004018);
-		udelay(20);
-	} else
-	if (!mclk.pll) {
-		ram_mask(fuc, 0x004168, 0x003f3040, mclk.clk);
-		ram_wr32(fuc, 0x004000, (ctrl |= 0x00000008));
+		nva3_ram_lock_pll(fuc, &mclk);
+	}
+
+	if (mclk.pll) {
+		ram_mask(fuc, 0x004000, 0x00000105, 0x00000105);
+		ram_wr32(fuc, 0x004018, 0x00001000 | r004018);
+		ram_wr32(fuc, 0x100da0, r100da0);
+	} else {
+		ram_mask(fuc, 0x004168, 0x003f3141, mclk.clk | 0x00000101);
+		ram_mask(fuc, 0x004000, 0x00000108, 0x00000008);
 		ram_mask(fuc, 0x1110e0, 0x00088000, 0x00088000);
-		ram_wr32(fuc, 0x004018, 0x0000d000 | r004018);
-	}
-
-	if ( (nv_ro08(bios, rammap.data + 0x04) & 0x08)) {
-		u32 unk5a0 = (nv_ro16(bios, ramcfg.data + 0x05) << 8) |
-			      nv_ro08(bios, ramcfg.data + 0x05);
-		u32 unk5a4 = (nv_ro16(bios, ramcfg.data + 0x07));
-		u32 unk804 = (nv_ro08(bios, ramcfg.data + 0x09) & 0xf0) << 16 |
-			     (nv_ro08(bios, ramcfg.data + 0x03) & 0x0f) << 16 |
-			     (nv_ro08(bios, ramcfg.data + 0x09) & 0x0f) |
-			     0x80000000;
-		ram_wr32(fuc, 0x1005a0, unk5a0);
-		ram_wr32(fuc, 0x1005a4, unk5a4);
-		ram_wr32(fuc, 0x10f804, unk804);
+		ram_wr32(fuc, 0x004018, 0x00009000 | r004018);
+		ram_wr32(fuc, 0x100da0, r100da0);
+	}
+	ram_nsec(fuc, 20000);
+
+	if (next->bios.rammap_10_04_08) {
+		ram_wr32(fuc, 0x1005a0, next->bios.ramcfg_10_06 << 16 |
+					next->bios.ramcfg_10_05 << 8 |
+					next->bios.ramcfg_10_05);
+		ram_wr32(fuc, 0x1005a4, next->bios.ramcfg_10_08 << 8 |
+					next->bios.ramcfg_10_07);
+		ram_wr32(fuc, 0x10f804, next->bios.ramcfg_10_09_f0 << 20 |
+					next->bios.ramcfg_10_03_0f << 16 |
+					next->bios.ramcfg_10_09_0f |
+					0x80000000);
 		ram_mask(fuc, 0x10053c, 0x00001000, 0x00000000);
 	} else {
+		if (train->state == NVA3_TRAIN_DONE) {
+			ram_wr32(fuc, 0x100080, 0x1020);
+			ram_mask(fuc, 0x111400, 0xffffffff, train->r_111400);
+			ram_mask(fuc, 0x1111e0, 0xffffffff, train->r_1111e0);
+			ram_mask(fuc, 0x100720, 0xffffffff, train->r_100720);
+		}
 		ram_mask(fuc, 0x10053c, 0x00001000, 0x00001000);
 		ram_mask(fuc, 0x10f804, 0x80000000, 0x00000000);
 		ram_mask(fuc, 0x100760, 0x22222222, r100760);
@@ -222,69 +719,134 @@ nva3_ram_calc(struct nouveau_fb *pfb, u32 freq)
 		ram_mask(fuc, 0x1007e0, 0x22222222, r100760);
 	}
 
+	if (nv_device(pfb)->chipset == 0xa3 && freq > 500000) {
+		ram_mask(fuc, 0x100700, 0x00000006, 0x00000000);
+	}
+
+	/* Final switch */
 	if (mclk.pll) {
 		ram_mask(fuc, 0x1110e0, 0x00088000, 0x00011000);
-		ram_wr32(fuc, 0x004000, (ctrl &= ~0x00000008));
+		ram_mask(fuc, 0x004000, 0x00000008, 0x00000000);
 	}
 
-	/*XXX: LEAVE */
 	ram_wr32(fuc, 0x1002dc, 0x00000000);
 	ram_wr32(fuc, 0x1002d4, 0x00000001);
 	ram_wr32(fuc, 0x100210, 0x80000000);
-	ram_nsec(fuc, 1000);
-	ram_nsec(fuc, 1000);
+	ram_nsec(fuc, 2000);
 
-	ram_mask(fuc, mr[2], 0x00000000, 0x00000000);
-	ram_nsec(fuc, 1000);
-	ram_nuke(fuc, mr[0]);
-	ram_mask(fuc, mr[0], 0x00000000, 0x00000000);
-	ram_nsec(fuc, 1000);
+	/* Set RAM MR parameters and timings */
+	for (i = 2; i >= 0; i--) {
+		if (ram_rd32(fuc, mr[i]) != ram->base.mr[i]) {
+			ram_wr32(fuc, mr[i], ram->base.mr[i]);
+			ram_nsec(fuc, 1000);
+		}
+	}
 
-	ram_mask(fuc, 0x100220[3], 0x00000000, 0x00000000);
-	ram_mask(fuc, 0x100220[1], 0x00000000, 0x00000000);
-	ram_mask(fuc, 0x100220[6], 0x00000000, 0x00000000);
-	ram_mask(fuc, 0x100220[7], 0x00000000, 0x00000000);
-	ram_mask(fuc, 0x100220[2], 0x00000000, 0x00000000);
-	ram_mask(fuc, 0x100220[4], 0x00000000, 0x00000000);
-	ram_mask(fuc, 0x100220[5], 0x00000000, 0x00000000);
-	ram_mask(fuc, 0x100220[0], 0x00000000, 0x00000000);
-	ram_mask(fuc, 0x100220[8], 0x00000000, 0x00000000);
-
-	data = (nv_ro08(bios, ramcfg.data + 0x02) & 0x08) ? 0x00000000 : 0x00001000;
-	ram_mask(fuc, 0x100200, 0x00001000, data);
-
-	unk714 = ram_rd32(fuc, 0x100714) & ~0xf0000010;
-	unk718 = ram_rd32(fuc, 0x100718) & ~0x00000100;
-	unk71c = ram_rd32(fuc, 0x10071c) & ~0x00000100;
-	if ( (nv_ro08(bios, ramcfg.data + 0x02) & 0x20))
-		unk714 |= 0xf0000000;
-	if (!(nv_ro08(bios, ramcfg.data + 0x02) & 0x04))
-		unk714 |= 0x00000010;
-	ram_wr32(fuc, 0x100714, unk714);
+	ram_wr32(fuc, 0x100220[3], timing[3]);
+	ram_wr32(fuc, 0x100220[1], timing[1]);
+	ram_wr32(fuc, 0x100220[6], timing[6]);
+	ram_wr32(fuc, 0x100220[7], timing[7]);
+	ram_wr32(fuc, 0x100220[2], timing[2]);
+	ram_wr32(fuc, 0x100220[4], timing[4]);
+	ram_wr32(fuc, 0x100220[5], timing[5]);
+	ram_wr32(fuc, 0x100220[0], timing[0]);
+	ram_wr32(fuc, 0x100220[8], timing[8]);
 
-	if (nv_ro08(bios, ramcfg.data + 0x02) & 0x01)
-		unk71c |= 0x00000100;
-	ram_wr32(fuc, 0x10071c, unk71c);
+	/* Misc */
+	ram_mask(fuc, 0x100200, 0x00001000, !next->bios.ramcfg_10_02_08 << 12);
+
+	/* XXX: A lot of "chipset"/"ram type" specific stuff...? */
+	unk714  = ram_rd32(fuc, 0x100714) & ~0xf0000130;
+	unk718  = ram_rd32(fuc, 0x100718) & ~0x00000100;
+	unk71c  = ram_rd32(fuc, 0x10071c) & ~0x00000100;
+	r111100 = ram_rd32(fuc, 0x111100) & ~0x3a800000;
+
+	if (next->bios.ramcfg_10_02_04) {
+		switch (ram->base.type) {
+		case NV_MEM_TYPE_DDR3:
+			if (nv_device(pfb)->chipset != 0xa8)
+				r111100 |= 0x00000004;
+			/* no break */
+		case NV_MEM_TYPE_DDR2:
+			r111100 |= 0x08000000;
+			break;
+		default:
+			break;
+		}
+	} else {
+		switch (ram->base.type) {
+		case NV_MEM_TYPE_DDR2:
+			r111100 |= 0x1a800000;
+			unk714  |= 0x00000010;
+			break;
+		case NV_MEM_TYPE_DDR3:
+			if (nv_device(pfb)->chipset == 0xa8) {
+				r111100 |=  0x08000000;
+			} else {
+				r111100 &= ~0x00000004;
+				r111100 |=  0x12800000;
+			}
+			unk714  |= 0x00000010;
+			break;
+		case NV_MEM_TYPE_GDDR3:
+			r111100 |= 0x30000000;
+			unk714  |= 0x00000020;
+			break;
+		default:
+			break;
+		}
+	}
+
+	unk714 |= (next->bios.ramcfg_10_04_01) << 8;
 
-	if (nv_ro08(bios, ramcfg.data + 0x02) & 0x02)
+	if (next->bios.ramcfg_10_02_20)
+		unk714 |= 0xf0000000;
+	if (next->bios.ramcfg_10_02_02)
 		unk718 |= 0x00000100;
-	ram_wr32(fuc, 0x100718, unk718);
+	if (next->bios.ramcfg_10_02_01)
+		unk71c |= 0x00000100;
+	if (next->bios.timing_10_24 != 0xff) {
+		unk718 &= ~0xf0000000;
+		unk718 |= next->bios.timing_10_24 << 28;
+	}
+	if (next->bios.ramcfg_10_02_10)
+		r111100 &= ~0x04020000;
 
-	if (nv_ro08(bios, ramcfg.data + 0x02) & 0x10)
-		ram_wr32(fuc, 0x111100, 0x48000000); /*XXX*/
+	ram_mask(fuc, 0x100714, 0xffffffff, unk714);
+	ram_mask(fuc, 0x10071c, 0xffffffff, unk71c);
+	ram_mask(fuc, 0x100718, 0xffffffff, unk718);
+	ram_mask(fuc, 0x111100, 0xffffffff, r111100);
 
-	ram_mask(fuc, mr[0], 0x100, 0x100);
-	ram_nsec(fuc, 1000);
-	ram_mask(fuc, mr[0], 0x100, 0x000);
-	ram_nsec(fuc, 1000);
+	if (fuc->r_gpioFBVREF.addr && !next->bios.timing_10_ODT)
+		nva3_ram_fbvref(fuc, 1);
 
-	ram_nsec(fuc, 2000);
-	ram_nsec(fuc, 12000);
+	/* Reset DLL */
+	if (!next->bios.ramcfg_10_DLLoff)
+		nouveau_sddr2_dll_reset(fuc);
 
-	ram_wr32(fuc, 0x611200, 0x00003330);
-	if ( (nv_ro08(bios, rammap.data + 0x04) & 0x02))
+	if (ram->base.type == NV_MEM_TYPE_GDDR3) {
+		ram_nsec(fuc, 31000);
+	} else {
+		ram_nsec(fuc, 14000);
+	}
+
+	if (ram->base.type == NV_MEM_TYPE_DDR3) {
+		ram_wr32(fuc, 0x100264, 0x1);
+		ram_nsec(fuc, 2000);
+	}
+
+	ram_nuke(fuc, 0x100700);
+	ram_mask(fuc, 0x100700, 0x01000000, 0x01000000);
+	ram_mask(fuc, 0x100700, 0x01000000, 0x00000000);
+
+	/* Re-enable FB */
+	ram_unblock(fuc);
+	ram_wr32(fuc, 0x611200, 0x3330);
+
+	/* Post fiddlings */
+	if (next->bios.rammap_10_04_02)
 		ram_mask(fuc, 0x100200, 0x00000800, 0x00000800);
-	if ( (nv_ro08(bios, ramcfg.data + 0x02) & 0x10)) {
+	if (next->bios.ramcfg_10_02_10) {
 		ram_mask(fuc, 0x111104, 0x00000180, 0x00000180);
 		ram_mask(fuc, 0x111100, 0x40000000, 0x00000000);
 	} else {
@@ -309,7 +871,22 @@ nva3_ram_prog(struct nouveau_fb *pfb)
 	struct nouveau_device *device = nv_device(pfb);
 	struct nva3_ram *ram = (void *)pfb->ram;
 	struct nva3_ramfuc *fuc = &ram->fuc;
-	ram_exec(fuc, nouveau_boolopt(device->cfgopt, "NvMemExec", true));
+	bool exec = nouveau_boolopt(device->cfgopt, "NvMemExec", true);
+
+	if (exec) {
+		nv_mask(pfb, 0x001534, 0x2, 0x2);
+
+		ram_exec(fuc, true);
+
+		/* Post-processing, avoids flicker */
+		nv_mask(pfb, 0x002504, 0x1, 0x0);
+		nv_mask(pfb, 0x001534, 0x2, 0x0);
+
+		nv_mask(pfb, 0x616308, 0x10, 0x10);
+		nv_mask(pfb, 0x616b08, 0x10, 0x10);
+	} else {
+		ram_exec(fuc, false);
+	}
 	return 0;
 }
 
@@ -326,38 +903,24 @@ nva3_ram_init(struct nouveau_object *object)
 {
 	struct nouveau_fb *pfb = (void *)object->parent;
 	struct nva3_ram   *ram = (void *)object;
-	int ret, i;
+	int ret;
 
 	ret = nouveau_ram_init(&ram->base);
 	if (ret)
 		return ret;
 
-	/* prepare for ddr link training, and load training patterns */
-	switch (ram->base.type) {
-	case NV_MEM_TYPE_DDR3: {
-		if (nv_device(pfb)->chipset == 0xa8) {
-			static const u32 pattern[16] = {
-				0xaaaaaaaa, 0xcccccccc, 0xdddddddd, 0xeeeeeeee,
-				0x00000000, 0x11111111, 0x44444444, 0xdddddddd,
-				0x33333333, 0x55555555, 0x77777777, 0x66666666,
-				0x99999999, 0x88888888, 0xeeeeeeee, 0xbbbbbbbb,
-			};
-
-			nv_wr32(pfb, 0x100538, 0x10001ff6); /*XXX*/
-			nv_wr32(pfb, 0x1005a8, 0x0000ffff);
-			nv_mask(pfb, 0x10f800, 0x00000001, 0x00000001);
-			for (i = 0; i < 0x30; i++) {
-				nv_wr32(pfb, 0x10f8c0, (i << 8) | i);
-				nv_wr32(pfb, 0x10f8e0, (i << 8) | i);
-				nv_wr32(pfb, 0x10f900, pattern[i % 16]);
-				nv_wr32(pfb, 0x10f920, pattern[i % 16]);
-			}
-		}
-	}
-		break;
-	default:
-		break;
-	}
+	nva3_link_train_init(pfb);
+
+	return 0;
+}
+
+static int
+nva3_ram_fini(struct nouveau_object *object, bool suspend)
+{
+	struct nouveau_fb *pfb = (void *)object->parent;
+
+	if (!suspend)
+		nva3_link_train_fini(pfb);
 
 	return 0;
 }
@@ -367,8 +930,12 @@ nva3_ram_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	      struct nouveau_oclass *oclass, void *data, u32 datasize,
 	      struct nouveau_object **pobject)
 {
+	struct nouveau_fb *pfb = nouveau_fb(parent);
+	struct nouveau_gpio *gpio = nouveau_gpio(pfb);
+	struct dcb_gpio_func func;
 	struct nva3_ram *ram;
 	int ret, i;
+	u32 reg, shift;
 
 	ret = nv50_ram_create(parent, engine, oclass, &ram);
 	*pobject = nv_object(ram);
@@ -376,7 +943,9 @@ nva3_ram_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 		return ret;
 
 	switch (ram->base.type) {
+	case NV_MEM_TYPE_DDR2:
 	case NV_MEM_TYPE_DDR3:
+	case NV_MEM_TYPE_GDDR3:
 		ram->base.calc = nva3_ram_calc;
 		ram->base.prog = nva3_ram_prog;
 		ram->base.tidy = nva3_ram_tidy;
@@ -386,31 +955,41 @@ nva3_ram_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 		return 0;
 	}
 
+	ram->fuc.r_0x001610 = ramfuc_reg(0x001610);
+	ram->fuc.r_0x001700 = ramfuc_reg(0x001700);
+	ram->fuc.r_0x002504 = ramfuc_reg(0x002504);
 	ram->fuc.r_0x004000 = ramfuc_reg(0x004000);
 	ram->fuc.r_0x004004 = ramfuc_reg(0x004004);
 	ram->fuc.r_0x004018 = ramfuc_reg(0x004018);
 	ram->fuc.r_0x004128 = ramfuc_reg(0x004128);
 	ram->fuc.r_0x004168 = ramfuc_reg(0x004168);
+	ram->fuc.r_0x100080 = ramfuc_reg(0x100080);
 	ram->fuc.r_0x100200 = ramfuc_reg(0x100200);
 	ram->fuc.r_0x100210 = ramfuc_reg(0x100210);
 	for (i = 0; i < 9; i++)
 		ram->fuc.r_0x100220[i] = ramfuc_reg(0x100220 + (i * 4));
+	ram->fuc.r_0x100264 = ramfuc_reg(0x100264);
 	ram->fuc.r_0x1002d0 = ramfuc_reg(0x1002d0);
 	ram->fuc.r_0x1002d4 = ramfuc_reg(0x1002d4);
 	ram->fuc.r_0x1002dc = ramfuc_reg(0x1002dc);
 	ram->fuc.r_0x10053c = ramfuc_reg(0x10053c);
 	ram->fuc.r_0x1005a0 = ramfuc_reg(0x1005a0);
 	ram->fuc.r_0x1005a4 = ramfuc_reg(0x1005a4);
+	ram->fuc.r_0x100700 = ramfuc_reg(0x100700);
 	ram->fuc.r_0x100714 = ramfuc_reg(0x100714);
 	ram->fuc.r_0x100718 = ramfuc_reg(0x100718);
 	ram->fuc.r_0x10071c = ramfuc_reg(0x10071c);
-	ram->fuc.r_0x100760 = ramfuc_reg(0x100760);
-	ram->fuc.r_0x1007a0 = ramfuc_reg(0x1007a0);
-	ram->fuc.r_0x1007e0 = ramfuc_reg(0x1007e0);
+	ram->fuc.r_0x100720 = ramfuc_reg(0x100720);
+	ram->fuc.r_0x100760 = ramfuc_stride(0x100760, 4, ram->base.part_mask);
+	ram->fuc.r_0x1007a0 = ramfuc_stride(0x1007a0, 4, ram->base.part_mask);
+	ram->fuc.r_0x1007e0 = ramfuc_stride(0x1007e0, 4, ram->base.part_mask);
+	ram->fuc.r_0x100da0 = ramfuc_stride(0x100da0, 4, ram->base.part_mask);
 	ram->fuc.r_0x10f804 = ramfuc_reg(0x10f804);
-	ram->fuc.r_0x1110e0 = ramfuc_reg(0x1110e0);
+	ram->fuc.r_0x1110e0 = ramfuc_stride(0x1110e0, 4, ram->base.part_mask);
 	ram->fuc.r_0x111100 = ramfuc_reg(0x111100);
 	ram->fuc.r_0x111104 = ramfuc_reg(0x111104);
+	ram->fuc.r_0x1111e0 = ramfuc_reg(0x1111e0);
+	ram->fuc.r_0x111400 = ramfuc_reg(0x111400);
 	ram->fuc.r_0x611200 = ramfuc_reg(0x611200);
 
 	if (ram->base.ranks > 1) {
@@ -425,6 +1004,12 @@ nva3_ram_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 		ram->fuc.r_mr[3] = ramfuc_reg(0x1002e4);
 	}
 
+	ret = gpio->find(gpio, 0, 0x2e, DCB_GPIO_UNUSED, &func);
+	if (ret == 0) {
+		nv50_gpio_location(func.line, &reg, &shift);
+		ram->fuc.r_gpioFBVREF = ramfuc_reg(reg);
+	}
+
 	return 0;
 }
 
@@ -434,6 +1019,6 @@ nva3_ram_oclass = {
 		.ctor = nva3_ram_ctor,
 		.dtor = _nouveau_ram_dtor,
 		.init = nva3_ram_init,
-		.fini = _nouveau_ram_fini,
+		.fini = nva3_ram_fini,
 	},
 };
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fb/ramnvaa.c b/drivers/gpu/drm/nouveau/core/subdev/fb/ramnvaa.c
index 00f2ca7e44a5..033a8e999497 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/fb/ramnvaa.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/fb/ramnvaa.c
@@ -24,34 +24,71 @@
 
 #include "nv50.h"
 
+struct nvaa_ram_priv {
+	struct nouveau_ram base;
+	u64 poller_base;
+};
+
 static int
 nvaa_ram_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	      struct nouveau_oclass *oclass, void *data, u32 datasize,
 	      struct nouveau_object **pobject)
 {
-	const u32 rsvd_head = ( 256 * 1024) >> 12; /* vga memory */
-	const u32 rsvd_tail = (1024 * 1024) >> 12; /* vbios etc */
+	u32 rsvd_head = ( 256 * 1024); /* vga memory */
+	u32 rsvd_tail = (1024 * 1024); /* vbios etc */
 	struct nouveau_fb *pfb = nouveau_fb(parent);
-	struct nouveau_ram *ram;
+	struct nvaa_ram_priv *priv;
 	int ret;
 
-	ret = nouveau_ram_create(parent, engine, oclass, &ram);
-	*pobject = nv_object(ram);
+	ret = nouveau_ram_create(parent, engine, oclass, &priv);
+	*pobject = nv_object(priv);
 	if (ret)
 		return ret;
 
-	ram->size = nv_rd32(pfb, 0x10020c);
-	ram->size = (ram->size & 0xffffff00) | ((ram->size & 0x000000ff) << 32);
+	priv->base.type   = NV_MEM_TYPE_STOLEN;
+	priv->base.stolen = (u64)nv_rd32(pfb, 0x100e10) << 12;
+	priv->base.size   = (u64)nv_rd32(pfb, 0x100e14) << 12;
 
-	ret = nouveau_mm_init(&pfb->vram, rsvd_head, (ram->size >> 12) -
-			      (rsvd_head + rsvd_tail), 1);
+	rsvd_tail += 0x1000;
+	priv->poller_base = priv->base.size - rsvd_tail;
+
+	ret = nouveau_mm_init(&pfb->vram, rsvd_head >> 12,
+			      (priv->base.size  - (rsvd_head + rsvd_tail)) >> 12,
+			      1);
 	if (ret)
 		return ret;
 
-	ram->type   = NV_MEM_TYPE_STOLEN;
-	ram->stolen = (u64)nv_rd32(pfb, 0x100e10) << 12;
-	ram->get = nv50_ram_get;
-	ram->put = nv50_ram_put;
+	priv->base.get = nv50_ram_get;
+	priv->base.put = nv50_ram_put;
+	return 0;
+}
+
+static int
+nvaa_ram_init(struct nouveau_object *object)
+{
+	struct nouveau_fb *pfb = nouveau_fb(object);
+	struct nvaa_ram_priv *priv = (void *)object;
+	int ret;
+	u64 dniso, hostnb, flush;
+
+	ret = nouveau_ram_init(&priv->base);
+	if (ret)
+		return ret;
+
+	dniso  = ((priv->base.size - (priv->poller_base + 0x00)) >> 5) - 1;
+	hostnb = ((priv->base.size - (priv->poller_base + 0x20)) >> 5) - 1;
+	flush  = ((priv->base.size - (priv->poller_base + 0x40)) >> 5) - 1;
+
+	/* Enable NISO poller for various clients and set their associated
+	 * read address, only for MCP77/78 and MCP79/7A. (fd#25701)
+	 */
+	nv_wr32(pfb, 0x100c18, dniso);
+	nv_mask(pfb, 0x100c14, 0x00000000, 0x00000001);
+	nv_wr32(pfb, 0x100c1c, hostnb);
+	nv_mask(pfb, 0x100c14, 0x00000000, 0x00000002);
+	nv_wr32(pfb, 0x100c24, flush);
+	nv_mask(pfb, 0x100c14, 0x00000000, 0x00010000);
+
 	return 0;
 }
 
@@ -60,7 +97,7 @@ nvaa_ram_oclass = {
 	.ofuncs = &(struct nouveau_ofuncs) {
 		.ctor = nvaa_ram_ctor,
 		.dtor = _nouveau_ram_dtor,
-		.init = _nouveau_ram_init,
+		.init = nvaa_ram_init,
 		.fini = _nouveau_ram_fini,
 	},
 };
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fb/ramnvc0.c b/drivers/gpu/drm/nouveau/core/subdev/fb/ramnvc0.c
index 2b284b192763..735cb9580abe 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/fb/ramnvc0.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/fb/ramnvc0.c
@@ -133,6 +133,7 @@ nvc0_ram_calc(struct nouveau_fb *pfb, u32 freq)
 	struct nouveau_bios *bios = nouveau_bios(pfb);
 	struct nvc0_ram *ram = (void *)pfb->ram;
 	struct nvc0_ramfuc *fuc = &ram->fuc;
+	struct nvbios_ramcfg cfg;
 	u8  ver, cnt, len, strap;
 	struct {
 		u32 data;
@@ -145,7 +146,7 @@ nvc0_ram_calc(struct nouveau_fb *pfb, u32 freq)
 
 	/* lookup memory config data relevant to the target frequency */
 	rammap.data = nvbios_rammapEm(bios, freq / 1000, &ver, &rammap.size,
-				     &cnt, &ramcfg.size);
+				     &cnt, &ramcfg.size, &cfg);
 	if (!rammap.data || ver != 0x10 || rammap.size < 0x0e) {
 		nv_error(pfb, "invalid/missing rammap entry\n");
 		return -EINVAL;
@@ -483,9 +484,9 @@ nvc0_ram_get(struct nouveau_fb *pfb, u64 size, u32 align, u32 ncmin,
 
 	do {
 		if (back)
-			ret = nouveau_mm_tail(mm, 1, size, ncmin, align, &r);
+			ret = nouveau_mm_tail(mm, 0, 1, size, ncmin, align, &r);
 		else
-			ret = nouveau_mm_head(mm, 1, size, ncmin, align, &r);
+			ret = nouveau_mm_head(mm, 0, 1, size, ncmin, align, &r);
 		if (ret) {
 			mutex_unlock(&pfb->base.mutex);
 			pfb->ram->put(pfb, &mem);
@@ -562,7 +563,7 @@ nvc0_ram_create_(struct nouveau_object *parent, struct nouveau_object *engine,
 		offset = (0x0200000000ULL >> 12) + (bsize << 8);
 		length = (ram->size >> 12) - ((bsize * parts) << 8) - rsvd_tail;
 
-		ret = nouveau_mm_init(&pfb->vram, offset, length, 0);
+		ret = nouveau_mm_init(&pfb->vram, offset, length, 1);
 		if (ret)
 			nouveau_mm_fini(&pfb->vram);
 	}
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fb/ramnve0.c b/drivers/gpu/drm/nouveau/core/subdev/fb/ramnve0.c
index c5b46e302319..6bae474abb44 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/fb/ramnve0.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/fb/ramnve0.c
@@ -29,6 +29,8 @@
 #include <subdev/bios/init.h>
 #include <subdev/bios/rammap.h>
 #include <subdev/bios/timing.h>
+#include <subdev/bios/M0205.h>
+#include <subdev/bios/M0209.h>
 
 #include <subdev/clock.h>
 #include <subdev/clock/pll.h>
@@ -41,14 +43,6 @@
 
 #include "ramfuc.h"
 
-/* binary driver only executes this path if the condition (a) is true
- * for any configuration (combination of rammap+ramcfg+timing) that
- * can be reached on a given card.  for now, we will execute the branch
- * unconditionally in the hope that a "false everywhere" in the bios
- * tables doesn't actually mean "don't touch this".
- */
-#define NOTE00(a) 1
-
 struct nve0_ramfuc {
 	struct ramfuc base;
 
@@ -134,10 +128,12 @@ struct nve0_ram {
 	struct nouveau_ram base;
 	struct nve0_ramfuc fuc;
 
+	struct list_head cfg;
 	u32 parts;
 	u32 pmask;
 	u32 pnuts;
 
+	struct nvbios_ramcfg diff;
 	int from;
 	int mode;
 	int N1, fN1, M1, P1;
@@ -241,7 +237,7 @@ nve0_ram_nuts(struct nve0_ram *ram, struct ramfuc_reg *reg,
 {
 	struct nve0_fb_priv *priv = (void *)nouveau_fb(ram);
 	struct ramfuc *fuc = &ram->fuc.base;
-	u32 addr = 0x110000 + (reg->addr[0] & 0xfff);
+	u32 addr = 0x110000 + (reg->addr & 0xfff);
 	u32 mask = _mask | _copy;
 	u32 data = (_data & _mask) | (reg->data & _copy);
 	u32 i;
@@ -268,6 +264,7 @@ nve0_ram_calc_gddr5(struct nouveau_fb *pfb, u32 freq)
 	u32 mask, data;
 
 	ram_mask(fuc, 0x10f808, 0x40000000, 0x40000000);
+	ram_block(fuc);
 	ram_wr32(fuc, 0x62c000, 0x0f0f0000);
 
 	/* MR1: turn termination on early, for some reason.. */
@@ -478,7 +475,7 @@ nve0_ram_calc_gddr5(struct nouveau_fb *pfb, u32 freq)
 	ram_mask(fuc, 0x10f2e8, 0xffffffff, next->bios.timing[9]);
 
 	data = mask = 0x00000000;
-	if (NOTE00(ramcfg_08_20)) {
+	if (ram->diff.ramcfg_11_08_20) {
 		if (next->bios.ramcfg_11_08_20)
 			data |= 0x01000000;
 		mask |= 0x01000000;
@@ -486,11 +483,11 @@ nve0_ram_calc_gddr5(struct nouveau_fb *pfb, u32 freq)
 	ram_mask(fuc, 0x10f200, mask, data);
 
 	data = mask = 0x00000000;
-	if (NOTE00(ramcfg_02_03 != 0)) {
+	if (ram->diff.ramcfg_11_02_03) {
 		data |= next->bios.ramcfg_11_02_03 << 8;
 		mask |= 0x00000300;
 	}
-	if (NOTE00(ramcfg_01_10)) {
+	if (ram->diff.ramcfg_11_01_10) {
 		if (next->bios.ramcfg_11_01_10)
 			data |= 0x70000000;
 		mask |= 0x70000000;
@@ -498,11 +495,11 @@ nve0_ram_calc_gddr5(struct nouveau_fb *pfb, u32 freq)
 	ram_mask(fuc, 0x10f604, mask, data);
 
 	data = mask = 0x00000000;
-	if (NOTE00(timing_30_07 != 0)) {
+	if (ram->diff.timing_20_30_07) {
 		data |= next->bios.timing_20_30_07 << 28;
 		mask |= 0x70000000;
 	}
-	if (NOTE00(ramcfg_01_01)) {
+	if (ram->diff.ramcfg_11_01_01) {
 		if (next->bios.ramcfg_11_01_01)
 			data |= 0x00000100;
 		mask |= 0x00000100;
@@ -510,11 +507,11 @@ nve0_ram_calc_gddr5(struct nouveau_fb *pfb, u32 freq)
 	ram_mask(fuc, 0x10f614, mask, data);
 
 	data = mask = 0x00000000;
-	if (NOTE00(timing_30_07 != 0)) {
+	if (ram->diff.timing_20_30_07) {
 		data |= next->bios.timing_20_30_07 << 28;
 		mask |= 0x70000000;
 	}
-	if (NOTE00(ramcfg_01_02)) {
+	if (ram->diff.ramcfg_11_01_02) {
 		if (next->bios.ramcfg_11_01_02)
 			data |= 0x00000100;
 		mask |= 0x00000100;
@@ -548,11 +545,11 @@ nve0_ram_calc_gddr5(struct nouveau_fb *pfb, u32 freq)
 	ram_wr32(fuc, 0x10f870, 0x11111111 * next->bios.ramcfg_11_03_0f);
 
 	data = mask = 0x00000000;
-	if (NOTE00(ramcfg_02_03 != 0)) {
+	if (ram->diff.ramcfg_11_02_03) {
 		data |= next->bios.ramcfg_11_02_03;
 		mask |= 0x00000003;
 	}
-	if (NOTE00(ramcfg_01_10)) {
+	if (ram->diff.ramcfg_11_01_10) {
 		if (next->bios.ramcfg_11_01_10)
 			data |= 0x00000004;
 		mask |= 0x00000004;
@@ -666,6 +663,7 @@ nve0_ram_calc_gddr5(struct nouveau_fb *pfb, u32 freq)
 	if (next->bios.ramcfg_11_07_02)
 		nve0_ram_train(fuc, 0x80020000, 0x01000000);
 
+	ram_unblock(fuc);
 	ram_wr32(fuc, 0x62c000, 0x0f0f0f00);
 
 	if (next->bios.rammap_11_08_01)
@@ -695,6 +693,7 @@ nve0_ram_calc_sddr3(struct nouveau_fb *pfb, u32 freq)
 	u32 mask, data;
 
 	ram_mask(fuc, 0x10f808, 0x40000000, 0x40000000);
+	ram_block(fuc);
 	ram_wr32(fuc, 0x62c000, 0x0f0f0000);
 
 	if (vc == 1 && ram_have(fuc, gpio2E)) {
@@ -917,6 +916,7 @@ nve0_ram_calc_sddr3(struct nouveau_fb *pfb, u32 freq)
 	ram_mask(fuc, 0x10f200, 0x80000000, 0x00000000);
 	ram_nsec(fuc, 1000);
 
+	ram_unblock(fuc);
 	ram_wr32(fuc, 0x62c000, 0x0f0f0f00);
 
 	if (next->bios.rammap_11_08_01)
@@ -932,58 +932,24 @@ nve0_ram_calc_sddr3(struct nouveau_fb *pfb, u32 freq)
  ******************************************************************************/
 
 static int
-nve0_ram_calc_data(struct nouveau_fb *pfb, u32 freq,
+nve0_ram_calc_data(struct nouveau_fb *pfb, u32 khz,
 		   struct nouveau_ram_data *data)
 {
-	struct nouveau_bios *bios = nouveau_bios(pfb);
 	struct nve0_ram *ram = (void *)pfb->ram;
-	u8 strap, cnt, len;
-
-	/* lookup memory config data relevant to the target frequency */
-	ram->base.rammap.data = nvbios_rammapEp(bios, freq / 1000,
-					       &ram->base.rammap.version,
-					       &ram->base.rammap.size,
-					       &cnt, &len, &data->bios);
-	if (!ram->base.rammap.data || ram->base.rammap.version != 0x11 ||
-	     ram->base.rammap.size < 0x09) {
-		nv_error(pfb, "invalid/missing rammap entry\n");
-		return -EINVAL;
-	}
-
-	/* locate specific data set for the attached memory */
-	strap = nvbios_ramcfg_index(nv_subdev(pfb));
-	ram->base.ramcfg.data = nvbios_rammapSp(bios, ram->base.rammap.data,
-						ram->base.rammap.version,
-						ram->base.rammap.size,
-						cnt, len, strap,
-						&ram->base.ramcfg.version,
-						&ram->base.ramcfg.size,
-						&data->bios);
-	if (!ram->base.ramcfg.data || ram->base.ramcfg.version != 0x11 ||
-	     ram->base.ramcfg.size < 0x08) {
-		nv_error(pfb, "invalid/missing ramcfg entry\n");
-		return -EINVAL;
-	}
-
-	/* lookup memory timings, if bios says they're present */
-	strap = nv_ro08(bios, ram->base.ramcfg.data + 0x00);
-	if (strap != 0xff) {
-		ram->base.timing.data =
-			nvbios_timingEp(bios, strap, &ram->base.timing.version,
-				       &ram->base.timing.size, &cnt, &len,
-				       &data->bios);
-		if (!ram->base.timing.data ||
-		     ram->base.timing.version != 0x20 ||
-		     ram->base.timing.size < 0x33) {
-			nv_error(pfb, "invalid/missing timing entry\n");
-			return -EINVAL;
+	struct nouveau_ram_data *cfg;
+	u32 mhz = khz / 1000;
+
+	list_for_each_entry(cfg, &ram->cfg, head) {
+		if (mhz >= cfg->bios.rammap_min &&
+		    mhz <= cfg->bios.rammap_max) {
+			*data = *cfg;
+			data->freq = khz;
+			return 0;
 		}
-	} else {
-		ram->base.timing.data = 0;
 	}
 
-	data->freq = freq;
-	return 0;
+	nv_error(ram, "ramcfg data for %dMHz not found\n", mhz);
+	return -EINVAL;
 }
 
 static int
@@ -1106,13 +1072,99 @@ nve0_ram_calc(struct nouveau_fb *pfb, u32 freq)
 	return nve0_ram_calc_xits(pfb, ram->base.next);
 }
 
+static void
+nve0_ram_prog_0(struct nouveau_fb *pfb, u32 freq)
+{
+	struct nve0_ram *ram = (void *)pfb->ram;
+	struct nouveau_ram_data *cfg;
+	u32 mhz = freq / 1000;
+	u32 mask, data;
+
+	list_for_each_entry(cfg, &ram->cfg, head) {
+		if (mhz >= cfg->bios.rammap_min &&
+		    mhz <= cfg->bios.rammap_max)
+			break;
+	}
+
+	if (&cfg->head == &ram->cfg)
+		return;
+
+	if (mask = 0, data = 0, ram->diff.rammap_11_0a_03fe) {
+		data |= cfg->bios.rammap_11_0a_03fe << 12;
+		mask |= 0x001ff000;
+	}
+	if (ram->diff.rammap_11_09_01ff) {
+		data |= cfg->bios.rammap_11_09_01ff;
+		mask |= 0x000001ff;
+	}
+	nv_mask(pfb, 0x10f468, mask, data);
+
+	if (mask = 0, data = 0, ram->diff.rammap_11_0a_0400) {
+		data |= cfg->bios.rammap_11_0a_0400;
+		mask |= 0x00000001;
+	}
+	nv_mask(pfb, 0x10f420, mask, data);
+
+	if (mask = 0, data = 0, ram->diff.rammap_11_0a_0800) {
+		data |= cfg->bios.rammap_11_0a_0800;
+		mask |= 0x00000001;
+	}
+	nv_mask(pfb, 0x10f430, mask, data);
+
+	if (mask = 0, data = 0, ram->diff.rammap_11_0b_01f0) {
+		data |= cfg->bios.rammap_11_0b_01f0;
+		mask |= 0x0000001f;
+	}
+	nv_mask(pfb, 0x10f400, mask, data);
+
+	if (mask = 0, data = 0, ram->diff.rammap_11_0b_0200) {
+		data |= cfg->bios.rammap_11_0b_0200 << 9;
+		mask |= 0x00000200;
+	}
+	nv_mask(pfb, 0x10f410, mask, data);
+
+	if (mask = 0, data = 0, ram->diff.rammap_11_0d) {
+		data |= cfg->bios.rammap_11_0d << 16;
+		mask |= 0x00ff0000;
+	}
+	if (ram->diff.rammap_11_0f) {
+		data |= cfg->bios.rammap_11_0f << 8;
+		mask |= 0x0000ff00;
+	}
+	nv_mask(pfb, 0x10f440, mask, data);
+
+	if (mask = 0, data = 0, ram->diff.rammap_11_0e) {
+		data |= cfg->bios.rammap_11_0e << 8;
+		mask |= 0x0000ff00;
+	}
+	if (ram->diff.rammap_11_0b_0800) {
+		data |= cfg->bios.rammap_11_0b_0800 << 7;
+		mask |= 0x00000080;
+	}
+	if (ram->diff.rammap_11_0b_0400) {
+		data |= cfg->bios.rammap_11_0b_0400 << 5;
+		mask |= 0x00000020;
+	}
+	nv_mask(pfb, 0x10f444, mask, data);
+}
+
 static int
 nve0_ram_prog(struct nouveau_fb *pfb)
 {
 	struct nouveau_device *device = nv_device(pfb);
 	struct nve0_ram *ram = (void *)pfb->ram;
 	struct nve0_ramfuc *fuc = &ram->fuc;
-	ram_exec(fuc, nouveau_boolopt(device->cfgopt, "NvMemExec", true));
+	struct nouveau_ram_data *next = ram->base.next;
+
+	if (!nouveau_boolopt(device->cfgopt, "NvMemExec", true)) {
+		ram_exec(fuc, false);
+		return (ram->base.next == &ram->base.xition);
+	}
+
+	nve0_ram_prog_0(pfb, 1000);
+	ram_exec(fuc, true);
+	nve0_ram_prog_0(pfb, next->freq);
+
 	return (ram->base.next == &ram->base.xition);
 }
 
@@ -1125,24 +1177,147 @@ nve0_ram_tidy(struct nouveau_fb *pfb)
 	ram_exec(fuc, false);
 }
 
+struct nve0_ram_train {
+	u16 mask;
+	struct nvbios_M0209S remap;
+	struct nvbios_M0209S type00;
+	struct nvbios_M0209S type01;
+	struct nvbios_M0209S type04;
+	struct nvbios_M0209S type06;
+	struct nvbios_M0209S type07;
+	struct nvbios_M0209S type08;
+	struct nvbios_M0209S type09;
+};
+
+static int
+nve0_ram_train_type(struct nouveau_fb *pfb, int i, u8 ramcfg,
+		    struct nve0_ram_train *train)
+{
+	struct nouveau_bios *bios = nouveau_bios(pfb);
+	struct nvbios_M0205E M0205E;
+	struct nvbios_M0205S M0205S;
+	struct nvbios_M0209E M0209E;
+	struct nvbios_M0209S *remap = &train->remap;
+	struct nvbios_M0209S *value;
+	u8  ver, hdr, cnt, len;
+	u32 data;
+
+	/* determine type of data for this index */
+	if (!(data = nvbios_M0205Ep(bios, i, &ver, &hdr, &cnt, &len, &M0205E)))
+		return -ENOENT;
+
+	switch (M0205E.type) {
+	case 0x00: value = &train->type00; break;
+	case 0x01: value = &train->type01; break;
+	case 0x04: value = &train->type04; break;
+	case 0x06: value = &train->type06; break;
+	case 0x07: value = &train->type07; break;
+	case 0x08: value = &train->type08; break;
+	case 0x09: value = &train->type09; break;
+	default:
+		return 0;
+	}
+
+	/* training data index determined by ramcfg strap */
+	if (!(data = nvbios_M0205Sp(bios, i, ramcfg, &ver, &hdr, &M0205S)))
+		return -EINVAL;
+	i = M0205S.data;
+
+	/* training data format information */
+	if (!(data = nvbios_M0209Ep(bios, i, &ver, &hdr, &cnt, &len, &M0209E)))
+		return -EINVAL;
+
+	/* ... and the raw data */
+	if (!(data = nvbios_M0209Sp(bios, i, 0, &ver, &hdr, value)))
+		return -EINVAL;
+
+	if (M0209E.v02_07 == 2) {
+		/* of course! why wouldn't we have a pointer to another entry
+		 * in the same table, and use the first one as an array of
+		 * remap indices...
+		 */
+		if (!(data = nvbios_M0209Sp(bios, M0209E.v03, 0, &ver, &hdr,
+					    remap)))
+			return -EINVAL;
+
+		for (i = 0; i < ARRAY_SIZE(value->data); i++)
+			value->data[i] = remap->data[value->data[i]];
+	} else
+	if (M0209E.v02_07 != 1)
+		return -EINVAL;
+
+	train->mask |= 1 << M0205E.type;
+	return 0;
+}
+
+static int
+nve0_ram_train_init_0(struct nouveau_fb *pfb, struct nve0_ram_train *train)
+{
+	int i, j;
+
+	if ((train->mask & 0x03d3) != 0x03d3) {
+		nv_warn(pfb, "missing link training data\n");
+		return -EINVAL;
+	}
+
+	for (i = 0; i < 0x30; i++) {
+		for (j = 0; j < 8; j += 4) {
+			nv_wr32(pfb, 0x10f968 + j, 0x00000000 | (i << 8));
+			nv_wr32(pfb, 0x10f920 + j, 0x00000000 |
+						   train->type08.data[i] << 4 |
+						   train->type06.data[i]);
+			nv_wr32(pfb, 0x10f918 + j, train->type00.data[i]);
+			nv_wr32(pfb, 0x10f920 + j, 0x00000100 |
+						   train->type09.data[i] << 4 |
+						   train->type07.data[i]);
+			nv_wr32(pfb, 0x10f918 + j, train->type01.data[i]);
+		}
+	}
+
+	for (j = 0; j < 8; j += 4) {
+		for (i = 0; i < 0x100; i++) {
+			nv_wr32(pfb, 0x10f968 + j, i);
+			nv_wr32(pfb, 0x10f900 + j, train->type04.data[i]);
+		}
+	}
+
+	return 0;
+}
+
+static int
+nve0_ram_train_init(struct nouveau_fb *pfb)
+{
+	u8 ramcfg = nvbios_ramcfg_index(nv_subdev(pfb));
+	struct nve0_ram_train *train;
+	int ret = -ENOMEM, i;
+
+	if ((train = kzalloc(sizeof(*train), GFP_KERNEL))) {
+		for (i = 0; i < 0x100; i++) {
+			ret = nve0_ram_train_type(pfb, i, ramcfg, train);
+			if (ret && ret != -ENOENT)
+				break;
+		}
+	}
+
+	switch (pfb->ram->type) {
+	case NV_MEM_TYPE_GDDR5:
+		ret = nve0_ram_train_init_0(pfb, train);
+		break;
+	default:
+		ret = 0;
+		break;
+	}
+
+	kfree(train);
+	return ret;
+}
+
 int
 nve0_ram_init(struct nouveau_object *object)
 {
 	struct nouveau_fb *pfb = (void *)object->parent;
 	struct nve0_ram *ram   = (void *)object;
 	struct nouveau_bios *bios = nouveau_bios(pfb);
-	static const u8  train0[] = {
-		0x00, 0xff, 0xff, 0x00, 0xff, 0x00,
-		0x00, 0xff, 0xff, 0x00, 0xff, 0x00,
-	};
-	static const u32 train1[] = {
-		0x00000000, 0xffffffff,
-		0x55555555, 0xaaaaaaaa,
-		0x33333333, 0xcccccccc,
-		0xf0f0f0f0, 0x0f0f0f0f,
-		0x00ff00ff, 0xff00ff00,
-		0x0000ffff, 0xffff0000,
-	};
 	u8  ver, hdr, cnt, len, snr, ssz;
 	u32 data, save;
 	int ret, i;
@@ -1168,51 +1343,107 @@ nve0_ram_init(struct nouveau_object *object)
 
 	cnt  = nv_ro08(bios, data + 0x14); /* guess at count */
 	data = nv_ro32(bios, data + 0x10); /* guess u32... */
-	save = nv_rd32(pfb, 0x10f65c);
-	for (i = 0; i < cnt; i++) {
-		nv_mask(pfb, 0x10f65c, 0x000000f0, i << 4);
-		nvbios_exec(&(struct nvbios_init) {
-				.subdev = nv_subdev(pfb),
-				.bios = bios,
-				.offset = nv_ro32(bios, data), /* guess u32 */
-				.execute = 1,
-			    });
-		data += 4;
-	}
-	nv_wr32(pfb, 0x10f65c, save);
+	save = nv_rd32(pfb, 0x10f65c) & 0x000000f0;
+	for (i = 0; i < cnt; i++, data += 4) {
+		if (i != save >> 4) {
+			nv_mask(pfb, 0x10f65c, 0x000000f0, i << 4);
+			nvbios_exec(&(struct nvbios_init) {
+					.subdev = nv_subdev(pfb),
+					.bios = bios,
+					.offset = nv_ro32(bios, data),
+					.execute = 1,
+				    });
+		}
+	}
+	nv_mask(pfb, 0x10f65c, 0x000000f0, save);
 	nv_mask(pfb, 0x10f584, 0x11000000, 0x00000000);
+	nv_wr32(pfb, 0x10ecc0, 0xffffffff);
+	nv_mask(pfb, 0x10f160, 0x00000010, 0x00000010);
 
-	switch (ram->base.type) {
-	case NV_MEM_TYPE_GDDR5:
-		for (i = 0; i < 0x30; i++) {
-			nv_wr32(pfb, 0x10f968, 0x00000000 | (i << 8));
-			nv_wr32(pfb, 0x10f920, 0x00000000 | train0[i % 12]);
-			nv_wr32(pfb, 0x10f918,              train1[i % 12]);
-			nv_wr32(pfb, 0x10f920, 0x00000100 | train0[i % 12]);
-			nv_wr32(pfb, 0x10f918,              train1[i % 12]);
-
-			nv_wr32(pfb, 0x10f96c, 0x00000000 | (i << 8));
-			nv_wr32(pfb, 0x10f924, 0x00000000 | train0[i % 12]);
-			nv_wr32(pfb, 0x10f91c,              train1[i % 12]);
-			nv_wr32(pfb, 0x10f924, 0x00000100 | train0[i % 12]);
-			nv_wr32(pfb, 0x10f91c,              train1[i % 12]);
-		}
+	return nve0_ram_train_init(pfb);
+}
 
-		for (i = 0; i < 0x100; i++) {
-			nv_wr32(pfb, 0x10f968, i);
-			nv_wr32(pfb, 0x10f900, train1[2 + (i & 1)]);
-		}
+static int
+nve0_ram_ctor_data(struct nve0_ram *ram, u8 ramcfg, int i)
+{
+	struct nouveau_fb *pfb = (void *)nv_object(ram)->parent;
+	struct nouveau_bios *bios = nouveau_bios(pfb);
+	struct nouveau_ram_data *cfg;
+	struct nvbios_ramcfg *d = &ram->diff;
+	struct nvbios_ramcfg *p, *n;
+	u8  ver, hdr, cnt, len;
+	u32 data;
+	int ret;
 
-		for (i = 0; i < 0x100; i++) {
-			nv_wr32(pfb, 0x10f96c, i);
-			nv_wr32(pfb, 0x10f900, train1[2 + (i & 1)]);
-		}
-		break;
-	default:
-		break;
+	if (!(cfg = kmalloc(sizeof(*cfg), GFP_KERNEL)))
+		return -ENOMEM;
+	p = &list_last_entry(&ram->cfg, typeof(*cfg), head)->bios;
+	n = &cfg->bios;
+
+	/* memory config data for a range of target frequencies */
+	data = nvbios_rammapEp(bios, i, &ver, &hdr, &cnt, &len, &cfg->bios);
+	if (ret = -ENOENT, !data)
+		goto done;
+	if (ret = -ENOSYS, ver != 0x11 || hdr < 0x12)
+		goto done;
+
+	/* ... and a portion specific to the attached memory */
+	data = nvbios_rammapSp(bios, data, ver, hdr, cnt, len, ramcfg,
+			       &ver, &hdr, &cfg->bios);
+	if (ret = -EINVAL, !data)
+		goto done;
+	if (ret = -ENOSYS, ver != 0x11 || hdr < 0x0a)
+		goto done;
+
+	/* lookup memory timings, if bios says they're present */
+	if (cfg->bios.ramcfg_timing != 0xff) {
+		data = nvbios_timingEp(bios, cfg->bios.ramcfg_timing,
+				       &ver, &hdr, &cnt, &len,
+				       &cfg->bios);
+		if (ret = -EINVAL, !data)
+			goto done;
+		if (ret = -ENOSYS, ver != 0x20 || hdr < 0x33)
+			goto done;
 	}
 
-	return 0;
+	list_add_tail(&cfg->head, &ram->cfg);
+	if (ret = 0, i == 0)
+		goto done;
+
+	d->rammap_11_0a_03fe |= p->rammap_11_0a_03fe != n->rammap_11_0a_03fe;
+	d->rammap_11_09_01ff |= p->rammap_11_09_01ff != n->rammap_11_09_01ff;
+	d->rammap_11_0a_0400 |= p->rammap_11_0a_0400 != n->rammap_11_0a_0400;
+	d->rammap_11_0a_0800 |= p->rammap_11_0a_0800 != n->rammap_11_0a_0800;
+	d->rammap_11_0b_01f0 |= p->rammap_11_0b_01f0 != n->rammap_11_0b_01f0;
+	d->rammap_11_0b_0200 |= p->rammap_11_0b_0200 != n->rammap_11_0b_0200;
+	d->rammap_11_0d |= p->rammap_11_0d != n->rammap_11_0d;
+	d->rammap_11_0f |= p->rammap_11_0f != n->rammap_11_0f;
+	d->rammap_11_0e |= p->rammap_11_0e != n->rammap_11_0e;
+	d->rammap_11_0b_0800 |= p->rammap_11_0b_0800 != n->rammap_11_0b_0800;
+	d->rammap_11_0b_0400 |= p->rammap_11_0b_0400 != n->rammap_11_0b_0400;
+	d->ramcfg_11_01_01 |= p->ramcfg_11_01_01 != n->ramcfg_11_01_01;
+	d->ramcfg_11_01_02 |= p->ramcfg_11_01_02 != n->ramcfg_11_01_02;
+	d->ramcfg_11_01_10 |= p->ramcfg_11_01_10 != n->ramcfg_11_01_10;
+	d->ramcfg_11_02_03 |= p->ramcfg_11_02_03 != n->ramcfg_11_02_03;
+	d->ramcfg_11_08_20 |= p->ramcfg_11_08_20 != n->ramcfg_11_08_20;
+	d->timing_20_30_07 |= p->timing_20_30_07 != n->timing_20_30_07;
+done:
+	if (ret)
+		kfree(cfg);
+	return ret;
+}
+
+static void
+nve0_ram_dtor(struct nouveau_object *object)
+{
+	struct nve0_ram *ram = (void *)object;
+	struct nouveau_ram_data *cfg, *tmp;
+
+	list_for_each_entry_safe(cfg, tmp, &ram->cfg, head) {
+		kfree(cfg);
+	}
+
+	nouveau_ram_destroy(&ram->base);
 }
 
 static int
@@ -1226,6 +1457,7 @@ nve0_ram_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	struct dcb_gpio_func func;
 	struct nve0_ram *ram;
 	int ret, i;
+	u8  ramcfg = nvbios_ramcfg_index(nv_subdev(pfb));
 	u32 tmp;
 
 	ret = nvc0_ram_create(parent, engine, oclass, 0x022554, &ram);
@@ -1233,6 +1465,8 @@ nve0_ram_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	if (ret)
 		return ret;
 
+	INIT_LIST_HEAD(&ram->cfg);
+
 	switch (ram->base.type) {
 	case NV_MEM_TYPE_DDR3:
 	case NV_MEM_TYPE_GDDR5:
@@ -1264,7 +1498,26 @@ nve0_ram_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 		}
 	}
 
-	// parse bios data for both pll's
+	/* parse bios data for all rammap table entries up-front, and
+	 * build information on whether certain fields differ between
+	 * any of the entries.
+	 *
+	 * the binary driver appears to completely ignore some fields
+	 * when all entries contain the same value.  at first, it was
+	 * hoped that these were mere optimisations and the bios init
+	 * tables had configured as per the values here, but there is
+	 * evidence now to suggest that this isn't the case and we do
+	 * need to treat this condition as a "don't touch" indicator.
+	 */
+	for (i = 0; !ret; i++) {
+		ret = nve0_ram_ctor_data(ram, ramcfg, i);
+		if (ret && ret != -ENOENT) {
+			nv_error(pfb, "failed to parse ramcfg data\n");
+			return ret;
+		}
+	}
+
+	/* parse bios data for both pll's */
 	ret = nvbios_pll_parse(bios, 0x0c, &ram->fuc.refpll);
 	if (ret) {
 		nv_error(pfb, "mclk refpll data not found\n");
@@ -1277,6 +1530,7 @@ nve0_ram_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 		return ret;
 	}
 
+	/* lookup memory voltage gpios */
 	ret = gpio->find(gpio, 0, 0x18, DCB_GPIO_UNUSED, &func);
 	if (ret == 0) {
 		ram->fuc.r_gpioMV = ramfuc_reg(0x00d610 + (func.line * 0x04));
@@ -1385,7 +1639,7 @@ nve0_ram_oclass = {
 	.handle = 0,
 	.ofuncs = &(struct nouveau_ofuncs) {
 		.ctor = nve0_ram_ctor,
-		.dtor = _nouveau_ram_dtor,
+		.dtor = nve0_ram_dtor,
 		.init = nve0_ram_init,
 		.fini = _nouveau_ram_fini,
 	}
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fb/sddr2.c b/drivers/gpu/drm/nouveau/core/subdev/fb/sddr2.c
new file mode 100644
index 000000000000..252575f3aa29
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/subdev/fb/sddr2.c
@@ -0,0 +1,94 @@
+/*
+ * Copyright 2014 Roy Spliet
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Roy Spliet <rspliet@eclipso.eu>
+ *          Ben Skeggs
+ */
+
+#include "priv.h"
+
+struct ramxlat {
+	int id;
+	u8 enc;
+};
+
+static inline int
+ramxlat(const struct ramxlat *xlat, int id)
+{
+	while (xlat->id >= 0) {
+		if (xlat->id == id)
+			return xlat->enc;
+		xlat++;
+	}
+	return -EINVAL;
+}
+
+static const struct ramxlat
+ramddr2_cl[] = {
+	{ 2, 2 }, { 3, 3 }, { 4, 4 }, { 5, 5 }, { 6, 6 },
+	/* The following are available in some, but not all DDR2 docs */
+	{ 7, 7 },
+	{ -1 }
+};
+
+static const struct ramxlat
+ramddr2_wr[] = {
+	{ 2, 1 }, { 3, 2 }, { 4, 3 }, { 5, 4 }, { 6, 5 },
+	/* The following are available in some, but not all DDR2 docs */
+	{ 7, 6 },
+	{ -1 }
+};
+
+int
+nouveau_sddr2_calc(struct nouveau_ram *ram)
+{
+	int CL, WR, DLL = 0, ODT = 0;
+
+	switch (ram->next->bios.timing_ver) {
+	case 0x10:
+		CL  = ram->next->bios.timing_10_CL;
+		WR  = ram->next->bios.timing_10_WR;
+		DLL = !ram->next->bios.ramcfg_10_DLLoff;
+		ODT = ram->next->bios.timing_10_ODT & 3;
+		break;
+	case 0x20:
+		CL  = (ram->next->bios.timing[1] & 0x0000001f);
+		WR  = (ram->next->bios.timing[2] & 0x007f0000) >> 16;
+		break;
+	default:
+		return -ENOSYS;
+	}
+
+	CL  = ramxlat(ramddr2_cl, CL);
+	WR  = ramxlat(ramddr2_wr, WR);
+	if (CL < 0 || WR < 0)
+		return -EINVAL;
+
+	ram->mr[0] &= ~0xf70;
+	ram->mr[0] |= (WR & 0x07) << 9;
+	ram->mr[0] |= (CL & 0x07) << 4;
+
+	ram->mr[1] &= ~0x045;
+	ram->mr[1] |= (ODT & 0x1) << 2;
+	ram->mr[1] |= (ODT & 0x2) << 5;
+	ram->mr[1] |= !DLL;
+	return 0;
+}
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fb/sddr3.c b/drivers/gpu/drm/nouveau/core/subdev/fb/sddr3.c
index ebd4cd9c35d9..a2dca4869e52 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/fb/sddr3.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/fb/sddr3.c
@@ -20,9 +20,9 @@
  * OTHER DEALINGS IN THE SOFTWARE.
  *
  * Authors: Ben Skeggs <bskeggs@redhat.com>
+ * 	    Roy Spliet <rspliet@eclipso.eu>
  */
 
-#include <subdev/bios.h>
 #include "priv.h"
 
 struct ramxlat {
@@ -69,31 +69,52 @@ ramddr3_cwl[] = {
 int
 nouveau_sddr3_calc(struct nouveau_ram *ram)
 {
-	struct nouveau_bios *bios = nouveau_bios(ram);
-	int WL, CL, WR;
+	int CWL, CL, WR, DLL = 0, ODT = 0;
 
-	switch (!!ram->timing.data * ram->timing.version) {
+	switch (ram->next->bios.timing_ver) {
+	case 0x10:
+		if (ram->next->bios.timing_hdr < 0x17) {
+			/* XXX: NV50: Get CWL from the timing register */
+			return -ENOSYS;
+		}
+		CWL = ram->next->bios.timing_10_CWL;
+		CL  = ram->next->bios.timing_10_CL;
+		WR  = ram->next->bios.timing_10_WR;
+		DLL = !ram->next->bios.ramcfg_10_DLLoff;
+		ODT = ram->next->bios.timing_10_ODT;
+		break;
 	case 0x20:
-		WL = (nv_ro16(bios, ram->timing.data + 0x04) & 0x0f80) >> 7;
-		CL =  nv_ro08(bios, ram->timing.data + 0x04) & 0x1f;
-		WR =  nv_ro08(bios, ram->timing.data + 0x0a) & 0x7f;
+		CWL = (ram->next->bios.timing[1] & 0x00000f80) >> 7;
+		CL  = (ram->next->bios.timing[1] & 0x0000001f) >> 0;
+		WR  = (ram->next->bios.timing[2] & 0x007f0000) >> 16;
+		/* XXX: Get these values from the VBIOS instead */
+		DLL = !(ram->mr[1] & 0x1);
+		ODT =   (ram->mr[1] & 0x004) >> 2 |
+			(ram->mr[1] & 0x040) >> 5 |
+		        (ram->mr[1] & 0x200) >> 7;
 		break;
 	default:
 		return -ENOSYS;
 	}
 
-	WL = ramxlat(ramddr3_cwl, WL);
-	CL = ramxlat(ramddr3_cl, CL);
-	WR = ramxlat(ramddr3_wr, WR);
-	if (WL < 0 || CL < 0 || WR < 0)
+	CWL = ramxlat(ramddr3_cwl, CWL);
+	CL  = ramxlat(ramddr3_cl, CL);
+	WR  = ramxlat(ramddr3_wr, WR);
+	if (CL < 0 || CWL < 0 || WR < 0)
 		return -EINVAL;
 
-	ram->mr[0] &= ~0xe74;
+	ram->mr[0] &= ~0xf74;
 	ram->mr[0] |= (WR & 0x07) << 9;
 	ram->mr[0] |= (CL & 0x0e) << 3;
 	ram->mr[0] |= (CL & 0x01) << 2;
 
+	ram->mr[1] &= ~0x245;
+	ram->mr[1] |= (ODT & 0x1) << 2;
+	ram->mr[1] |= (ODT & 0x2) << 5;
+	ram->mr[1] |= (ODT & 0x4) << 7;
+	ram->mr[1] |= !DLL;
+
 	ram->mr[2] &= ~0x038;
-	ram->mr[2] |= (WL & 0x07) << 3;
+	ram->mr[2] |= (CWL & 0x07) << 3;
 	return 0;
 }
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fuse/base.c b/drivers/gpu/drm/nouveau/core/subdev/fuse/base.c
new file mode 100644
index 000000000000..9e8e92127715
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/subdev/fuse/base.c
@@ -0,0 +1,54 @@
+/*
+ * Copyright 2014 Martin Peres
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Martin Peres
+ */
+
+#include <subdev/fuse.h>
+
+int
+_nouveau_fuse_init(struct nouveau_object *object)
+{
+	struct nouveau_fuse *fuse = (void *)object;
+	return nouveau_subdev_init(&fuse->base);
+}
+
+void
+_nouveau_fuse_dtor(struct nouveau_object *object)
+{
+	struct nouveau_fuse *fuse = (void *)object;
+	nouveau_subdev_destroy(&fuse->base);
+}
+
+int
+nouveau_fuse_create_(struct nouveau_object *parent,
+		     struct nouveau_object *engine,
+		     struct nouveau_oclass *oclass, int length, void **pobject)
+{
+	struct nouveau_fuse *fuse;
+	int ret;
+
+	ret = nouveau_subdev_create_(parent, engine, oclass, 0, "FUSE",
+				     "fuse", length, pobject);
+	fuse = *pobject;
+
+	return ret;
+}
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fuse/g80.c b/drivers/gpu/drm/nouveau/core/subdev/fuse/g80.c
new file mode 100644
index 000000000000..a374ade485be
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/subdev/fuse/g80.c
@@ -0,0 +1,81 @@
+/*
+ * Copyright 2014 Martin Peres
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Martin Peres
+ */
+
+#include "priv.h"
+
+struct g80_fuse_priv {
+	struct nouveau_fuse base;
+
+	spinlock_t fuse_enable_lock;
+};
+
+static u32
+g80_fuse_rd32(struct nouveau_object *object, u64 addr)
+{
+	struct g80_fuse_priv *priv = (void *)object;
+	unsigned long flags;
+	u32 fuse_enable, val;
+
+	spin_lock_irqsave(&priv->fuse_enable_lock, flags);
+
+	/* racy if another part of nouveau start writing to this reg */
+	fuse_enable = nv_mask(priv, 0x1084, 0x800, 0x800);
+	val = nv_rd32(priv, 0x21000 + addr);
+	nv_wr32(priv, 0x1084, fuse_enable);
+
+	spin_unlock_irqrestore(&priv->fuse_enable_lock, flags);
+
+	return val;
+}
+
+
+static int
+g80_fuse_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
+	       struct nouveau_oclass *oclass, void *data, u32 size,
+	       struct nouveau_object **pobject)
+{
+	struct g80_fuse_priv *priv;
+	int ret;
+
+	ret = nouveau_fuse_create(parent, engine, oclass, &priv);
+	*pobject = nv_object(priv);
+	if (ret)
+		return ret;
+
+	spin_lock_init(&priv->fuse_enable_lock);
+
+	return 0;
+}
+
+struct nouveau_oclass
+g80_fuse_oclass = {
+	.handle = NV_SUBDEV(FUSE, 0x50),
+	.ofuncs = &(struct nouveau_ofuncs) {
+		.ctor = g80_fuse_ctor,
+		.dtor = _nouveau_fuse_dtor,
+		.init = _nouveau_fuse_init,
+		.fini = _nouveau_fuse_fini,
+		.rd32 = g80_fuse_rd32,
+	},
+};
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fuse/gf100.c b/drivers/gpu/drm/nouveau/core/subdev/fuse/gf100.c
new file mode 100644
index 000000000000..5ed03f54b3d4
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/subdev/fuse/gf100.c
@@ -0,0 +1,83 @@
+/*
+ * Copyright 2014 Martin Peres
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Martin Peres
+ */
+
+#include "priv.h"
+
+struct gf100_fuse_priv {
+	struct nouveau_fuse base;
+
+	spinlock_t fuse_enable_lock;
+};
+
+static u32
+gf100_fuse_rd32(struct nouveau_object *object, u64 addr)
+{
+	struct gf100_fuse_priv *priv = (void *)object;
+	unsigned long flags;
+	u32 fuse_enable, unk, val;
+
+	spin_lock_irqsave(&priv->fuse_enable_lock, flags);
+
+	/* racy if another part of nouveau start writing to these regs */
+	fuse_enable = nv_mask(priv, 0x22400, 0x800, 0x800);
+	unk = nv_mask(priv, 0x21000, 0x1, 0x1);
+	val = nv_rd32(priv, 0x21100 + addr);
+	nv_wr32(priv, 0x21000, unk);
+	nv_wr32(priv, 0x22400, fuse_enable);
+
+	spin_unlock_irqrestore(&priv->fuse_enable_lock, flags);
+
+	return val;
+}
+
+
+static int
+gf100_fuse_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
+	       struct nouveau_oclass *oclass, void *data, u32 size,
+	       struct nouveau_object **pobject)
+{
+	struct gf100_fuse_priv *priv;
+	int ret;
+
+	ret = nouveau_fuse_create(parent, engine, oclass, &priv);
+	*pobject = nv_object(priv);
+	if (ret)
+		return ret;
+
+	spin_lock_init(&priv->fuse_enable_lock);
+
+	return 0;
+}
+
+struct nouveau_oclass
+gf100_fuse_oclass = {
+	.handle = NV_SUBDEV(FUSE, 0xC0),
+	.ofuncs = &(struct nouveau_ofuncs) {
+		.ctor = gf100_fuse_ctor,
+		.dtor = _nouveau_fuse_dtor,
+		.init = _nouveau_fuse_init,
+		.fini = _nouveau_fuse_fini,
+		.rd32 = gf100_fuse_rd32,
+	},
+};
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fuse/gm107.c b/drivers/gpu/drm/nouveau/core/subdev/fuse/gm107.c
new file mode 100644
index 000000000000..4f1a636c6538
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/subdev/fuse/gm107.c
@@ -0,0 +1,66 @@
+/*
+ * Copyright 2014 Martin Peres
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Martin Peres
+ */
+
+#include "priv.h"
+
+struct gm107_fuse_priv {
+	struct nouveau_fuse base;
+};
+
+static u32
+gm107_fuse_rd32(struct nouveau_object *object, u64 addr)
+{
+	struct gf100_fuse_priv *priv = (void *)object;
+
+	return nv_rd32(priv, 0x21100 + addr);
+}
+
+
+static int
+gm107_fuse_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
+	       struct nouveau_oclass *oclass, void *data, u32 size,
+	       struct nouveau_object **pobject)
+{
+	struct gm107_fuse_priv *priv;
+	int ret;
+
+	ret = nouveau_fuse_create(parent, engine, oclass, &priv);
+	*pobject = nv_object(priv);
+	if (ret)
+		return ret;
+
+	return 0;
+}
+
+struct nouveau_oclass
+gm107_fuse_oclass = {
+	.handle = NV_SUBDEV(FUSE, 0x117),
+	.ofuncs = &(struct nouveau_ofuncs) {
+		.ctor = gm107_fuse_ctor,
+		.dtor = _nouveau_fuse_dtor,
+		.init = _nouveau_fuse_init,
+		.fini = _nouveau_fuse_fini,
+		.rd32 = gm107_fuse_rd32,
+	},
+};
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fuse/priv.h b/drivers/gpu/drm/nouveau/core/subdev/fuse/priv.h
new file mode 100644
index 000000000000..d2085411a5cb
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/subdev/fuse/priv.h
@@ -0,0 +1,9 @@
+#ifndef __NVKM_FUSE_PRIV_H__
+#define __NVKM_FUSE_PRIV_H__
+
+#include <subdev/fuse.h>
+
+int _nouveau_fuse_init(struct nouveau_object *object);
+void _nouveau_fuse_dtor(struct nouveau_object *object);
+
+#endif
diff --git a/drivers/gpu/drm/nouveau/core/subdev/gpio/base.c b/drivers/gpu/drm/nouveau/core/subdev/gpio/base.c
index b1e3ed7c8beb..7ad99b763f4c 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/gpio/base.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/gpio/base.c
@@ -122,7 +122,8 @@ nouveau_gpio_intr_init(struct nvkm_event *event, int type, int index)
 }
 
 static int
-nouveau_gpio_intr_ctor(void *data, u32 size, struct nvkm_notify *notify)
+nouveau_gpio_intr_ctor(struct nouveau_object *object, void *data, u32 size,
+		       struct nvkm_notify *notify)
 {
 	struct nvkm_gpio_ntfy_req *req = data;
 	if (!WARN_ON(size != sizeof(*req))) {
diff --git a/drivers/gpu/drm/nouveau/core/subdev/gpio/nv50.c b/drivers/gpu/drm/nouveau/core/subdev/gpio/nv50.c
index 1864fa98e6b1..2e30d5a62d6e 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/gpio/nv50.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/gpio/nv50.c
@@ -54,7 +54,7 @@ nv50_gpio_reset(struct nouveau_gpio *gpio, u8 match)
 	}
 }
 
-static int
+int
 nv50_gpio_location(int line, u32 *reg, u32 *shift)
 {
 	const u32 nv50_gpio_reg[4] = { 0xe104, 0xe108, 0xe280, 0xe284 };
diff --git a/drivers/gpu/drm/nouveau/core/subdev/gpio/nv92.c b/drivers/gpu/drm/nouveau/core/subdev/gpio/nv94.c
index 252083d376f5..cae404ccadac 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/gpio/nv92.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/gpio/nv94.c
@@ -25,7 +25,7 @@
 #include "priv.h"
 
 void
-nv92_gpio_intr_stat(struct nouveau_gpio *gpio, u32 *hi, u32 *lo)
+nv94_gpio_intr_stat(struct nouveau_gpio *gpio, u32 *hi, u32 *lo)
 {
 	u32 intr0 = nv_rd32(gpio, 0x00e054);
 	u32 intr1 = nv_rd32(gpio, 0x00e074);
@@ -38,7 +38,7 @@ nv92_gpio_intr_stat(struct nouveau_gpio *gpio, u32 *hi, u32 *lo)
 }
 
 void
-nv92_gpio_intr_mask(struct nouveau_gpio *gpio, u32 type, u32 mask, u32 data)
+nv94_gpio_intr_mask(struct nouveau_gpio *gpio, u32 type, u32 mask, u32 data)
 {
 	u32 inte0 = nv_rd32(gpio, 0x00e050);
 	u32 inte1 = nv_rd32(gpio, 0x00e070);
@@ -57,8 +57,8 @@ nv92_gpio_intr_mask(struct nouveau_gpio *gpio, u32 type, u32 mask, u32 data)
 }
 
 struct nouveau_oclass *
-nv92_gpio_oclass = &(struct nouveau_gpio_impl) {
-	.base.handle = NV_SUBDEV(GPIO, 0x92),
+nv94_gpio_oclass = &(struct nouveau_gpio_impl) {
+	.base.handle = NV_SUBDEV(GPIO, 0x94),
 	.base.ofuncs = &(struct nouveau_ofuncs) {
 		.ctor = _nouveau_gpio_ctor,
 		.dtor = _nouveau_gpio_dtor,
@@ -66,8 +66,8 @@ nv92_gpio_oclass = &(struct nouveau_gpio_impl) {
 		.fini = _nouveau_gpio_fini,
 	},
 	.lines = 32,
-	.intr_stat = nv92_gpio_intr_stat,
-	.intr_mask = nv92_gpio_intr_mask,
+	.intr_stat = nv94_gpio_intr_stat,
+	.intr_mask = nv94_gpio_intr_mask,
 	.drive = nv50_gpio_drive,
 	.sense = nv50_gpio_sense,
 	.reset = nv50_gpio_reset,
diff --git a/drivers/gpu/drm/nouveau/core/subdev/gpio/nvd0.c b/drivers/gpu/drm/nouveau/core/subdev/gpio/nvd0.c
index a4682b0956ad..480d6d2af770 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/gpio/nvd0.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/gpio/nvd0.c
@@ -77,8 +77,8 @@ nvd0_gpio_oclass = &(struct nouveau_gpio_impl) {
 		.fini = _nouveau_gpio_fini,
 	},
 	.lines = 32,
-	.intr_stat = nv92_gpio_intr_stat,
-	.intr_mask = nv92_gpio_intr_mask,
+	.intr_stat = nv94_gpio_intr_stat,
+	.intr_mask = nv94_gpio_intr_mask,
 	.drive = nvd0_gpio_drive,
 	.sense = nvd0_gpio_sense,
 	.reset = nvd0_gpio_reset,
diff --git a/drivers/gpu/drm/nouveau/core/subdev/gpio/priv.h b/drivers/gpu/drm/nouveau/core/subdev/gpio/priv.h
index e1724dfc86ae..bff98b86e2b5 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/gpio/priv.h
+++ b/drivers/gpu/drm/nouveau/core/subdev/gpio/priv.h
@@ -56,8 +56,8 @@ void nv50_gpio_reset(struct nouveau_gpio *, u8);
 int  nv50_gpio_drive(struct nouveau_gpio *, int, int, int);
 int  nv50_gpio_sense(struct nouveau_gpio *, int);
 
-void nv92_gpio_intr_stat(struct nouveau_gpio *, u32 *, u32 *);
-void nv92_gpio_intr_mask(struct nouveau_gpio *, u32, u32, u32);
+void nv94_gpio_intr_stat(struct nouveau_gpio *, u32 *, u32 *);
+void nv94_gpio_intr_mask(struct nouveau_gpio *, u32, u32, u32);
 
 void nvd0_gpio_reset(struct nouveau_gpio *, u8);
 int  nvd0_gpio_drive(struct nouveau_gpio *, int, int, int);
diff --git a/drivers/gpu/drm/nouveau/core/subdev/i2c/base.c b/drivers/gpu/drm/nouveau/core/subdev/i2c/base.c
index a652cafde3d6..0dc605db7ec8 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/i2c/base.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/i2c/base.c
@@ -23,6 +23,7 @@
  */
 
 #include <core/option.h>
+#include <core/object.h>
 #include <core/event.h>
 
 #include <subdev/bios.h>
@@ -346,7 +347,8 @@ nouveau_i2c_intr_init(struct nvkm_event *event, int type, int index)
 }
 
 static int
-nouveau_i2c_intr_ctor(void *data, u32 size, struct nvkm_notify *notify)
+nouveau_i2c_intr_ctor(struct nouveau_object *object, void *data, u32 size,
+		      struct nvkm_notify *notify)
 {
 	struct nvkm_i2c_ntfy_req *req = data;
 	if (!WARN_ON(size != sizeof(*req))) {
@@ -471,18 +473,56 @@ nouveau_i2c_extdev_sclass[] = {
 	nouveau_anx9805_sclass,
 };
 
+static void
+nouveau_i2c_create_port(struct nouveau_i2c *i2c, int index, u8 type,
+			struct dcb_i2c_entry *info)
+{
+	const struct nouveau_i2c_impl *impl = (void *)nv_oclass(i2c);
+	struct nouveau_oclass *oclass;
+	struct nouveau_object *parent;
+	struct nouveau_object *object;
+	int ret, pad;
+
+	if (info->share != DCB_I2C_UNUSED) {
+		pad    = info->share;
+		oclass = impl->pad_s;
+	} else {
+		if (type != DCB_I2C_NVIO_AUX)
+			pad = 0x100 + info->drive;
+		else
+			pad = 0x100 + info->auxch;
+		oclass = impl->pad_x;
+	}
+
+	ret = nouveau_object_ctor(NULL, nv_object(i2c), oclass, NULL, pad,
+				 &parent);
+	if (ret < 0)
+		return;
+
+	oclass = impl->sclass;
+	do {
+		ret = -EINVAL;
+		if (oclass->handle == type) {
+			ret = nouveau_object_ctor(parent, nv_object(i2c),
+						  oclass, info, index,
+						 &object);
+		}
+	} while (ret && (++oclass)->handle);
+
+	nouveau_object_ref(NULL, &parent);
+}
+
 int
 nouveau_i2c_create_(struct nouveau_object *parent,
 		    struct nouveau_object *engine,
 		    struct nouveau_oclass *oclass,
 		    int length, void **pobject)
 {
-	const struct nouveau_i2c_impl *impl = (void *)oclass;
 	struct nouveau_bios *bios = nouveau_bios(parent);
 	struct nouveau_i2c *i2c;
 	struct nouveau_object *object;
 	struct dcb_i2c_entry info;
-	int ret, i, j, index = -1, pad;
+	int ret, i, j, index = -1;
 	struct dcb_output outp;
 	u8  ver, hdr;
 	u32 data;
@@ -505,43 +545,40 @@ nouveau_i2c_create_(struct nouveau_object *parent,
 	INIT_LIST_HEAD(&i2c->ports);
 
 	while (!dcb_i2c_parse(bios, ++index, &info)) {
-		if (info.type == DCB_I2C_UNUSED)
+		switch (info.type) {
+		case DCB_I2C_NV04_BIT:
+		case DCB_I2C_NV4E_BIT:
+		case DCB_I2C_NVIO_BIT:
+			nouveau_i2c_create_port(i2c, NV_I2C_PORT(index),
+						info.type, &info);
+			break;
+		case DCB_I2C_NVIO_AUX:
+			nouveau_i2c_create_port(i2c, NV_I2C_AUX(index),
+						info.type, &info);
+			break;
+		case DCB_I2C_PMGR:
+			if (info.drive != DCB_I2C_UNUSED) {
+				nouveau_i2c_create_port(i2c, NV_I2C_PORT(index),
+							DCB_I2C_NVIO_BIT,
+							&info);
+			}
+			if (info.auxch != DCB_I2C_UNUSED) {
+				nouveau_i2c_create_port(i2c, NV_I2C_AUX(index),
+							DCB_I2C_NVIO_AUX,
+							&info);
+			}
+			break;
+		case DCB_I2C_UNUSED:
+		default:
 			continue;
-
-		if (info.share != DCB_I2C_UNUSED) {
-			if (info.type == DCB_I2C_NVIO_AUX)
-				pad = info.drive;
-			else
-				pad = info.share;
-			oclass = impl->pad_s;
-		} else {
-			pad = 0x100 + info.drive;
-			oclass = impl->pad_x;
 		}
-
-		ret = nouveau_object_ctor(NULL, *pobject, oclass,
-					  NULL, pad, &parent);
-		if (ret < 0)
-			continue;
-
-		oclass = impl->sclass;
-		do {
-			ret = -EINVAL;
-			if (oclass->handle == info.type) {
-				ret = nouveau_object_ctor(parent, *pobject,
-							  oclass, &info,
-							  index, &object);
-			}
-		} while (ret && (++oclass)->handle);
-
-		nouveau_object_ref(NULL, &parent);
 	}
 
 	/* in addition to the busses specified in the i2c table, there
 	 * may be ddc/aux channels hiding behind external tmds/dp/etc
 	 * transmitters.
 	 */
-	index = ((index + 0x0f) / 0x10) * 0x10;
+	index = NV_I2C_EXT(0);
 	i = -1;
 	while ((data = dcb_outp_parse(bios, ++i, &ver, &hdr, &outp))) {
 		if (!outp.location || !outp.extdev)
diff --git a/drivers/gpu/drm/nouveau/core/subdev/i2c/gm204.c b/drivers/gpu/drm/nouveau/core/subdev/i2c/gm204.c
new file mode 100644
index 000000000000..06a2b87ccbf1
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/subdev/i2c/gm204.c
@@ -0,0 +1,221 @@
+/*
+ * Copyright 2012 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs
+ */
+
+#include "nv50.h"
+
+#define AUX_DBG(fmt, args...) nv_debug(aux, "AUXCH(%d): " fmt, ch, ##args)
+#define AUX_ERR(fmt, args...) nv_error(aux, "AUXCH(%d): " fmt, ch, ##args)
+
+static void
+auxch_fini(struct nouveau_i2c *aux, int ch)
+{
+	nv_mask(aux, 0x00d954 + (ch * 0x50), 0x00310000, 0x00000000);
+}
+
+static int
+auxch_init(struct nouveau_i2c *aux, int ch)
+{
+	const u32 unksel = 1; /* nfi which to use, or if it matters.. */
+	const u32 ureq = unksel ? 0x00100000 : 0x00200000;
+	const u32 urep = unksel ? 0x01000000 : 0x02000000;
+	u32 ctrl, timeout;
+
+	/* wait up to 1ms for any previous transaction to be done... */
+	timeout = 1000;
+	do {
+		ctrl = nv_rd32(aux, 0x00d954 + (ch * 0x50));
+		udelay(1);
+		if (!timeout--) {
+			AUX_ERR("begin idle timeout 0x%08x\n", ctrl);
+			return -EBUSY;
+		}
+	} while (ctrl & 0x03010000);
+
+	/* set some magic, and wait up to 1ms for it to appear */
+	nv_mask(aux, 0x00d954 + (ch * 0x50), 0x00300000, ureq);
+	timeout = 1000;
+	do {
+		ctrl = nv_rd32(aux, 0x00d954 + (ch * 0x50));
+		udelay(1);
+		if (!timeout--) {
+			AUX_ERR("magic wait 0x%08x\n", ctrl);
+			auxch_fini(aux, ch);
+			return -EBUSY;
+		}
+	} while ((ctrl & 0x03000000) != urep);
+
+	return 0;
+}
+
+int
+gm204_aux(struct nouveau_i2c_port *base, bool retry,
+	 u8 type, u32 addr, u8 *data, u8 size)
+{
+	struct nouveau_i2c *aux = nouveau_i2c(base);
+	struct nv50_i2c_port *port = (void *)base;
+	u32 ctrl, stat, timeout, retries;
+	u32 xbuf[4] = {};
+	int ch = port->addr;
+	int ret, i;
+
+	AUX_DBG("%d: 0x%08x %d\n", type, addr, size);
+
+	ret = auxch_init(aux, ch);
+	if (ret)
+		goto out;
+
+	stat = nv_rd32(aux, 0x00d958 + (ch * 0x50));
+	if (!(stat & 0x10000000)) {
+		AUX_DBG("sink not detected\n");
+		ret = -ENXIO;
+		goto out;
+	}
+
+	if (!(type & 1)) {
+		memcpy(xbuf, data, size);
+		for (i = 0; i < 16; i += 4) {
+			AUX_DBG("wr 0x%08x\n", xbuf[i / 4]);
+			nv_wr32(aux, 0x00d930 + (ch * 0x50) + i, xbuf[i / 4]);
+		}
+	}
+
+	ctrl  = nv_rd32(aux, 0x00d954 + (ch * 0x50));
+	ctrl &= ~0x0001f0ff;
+	ctrl |= type << 12;
+	ctrl |= size - 1;
+	nv_wr32(aux, 0x00d950 + (ch * 0x50), addr);
+
+	/* (maybe) retry transaction a number of times on failure... */
+	for (retries = 0; !ret && retries < 32; retries++) {
+		/* reset, and delay a while if this is a retry */
+		nv_wr32(aux, 0x00d954 + (ch * 0x50), 0x80000000 | ctrl);
+		nv_wr32(aux, 0x00d954 + (ch * 0x50), 0x00000000 | ctrl);
+		if (retries)
+			udelay(400);
+
+		/* transaction request, wait up to 1ms for it to complete */
+		nv_wr32(aux, 0x00d954 + (ch * 0x50), 0x00010000 | ctrl);
+
+		timeout = 1000;
+		do {
+			ctrl = nv_rd32(aux, 0x00d954 + (ch * 0x50));
+			udelay(1);
+			if (!timeout--) {
+				AUX_ERR("tx req timeout 0x%08x\n", ctrl);
+				ret = -EIO;
+				goto out;
+			}
+		} while (ctrl & 0x00010000);
+		ret = 1;
+
+		/* read status, and check if transaction completed ok */
+		stat = nv_mask(aux, 0x00d958 + (ch * 0x50), 0, 0);
+		if ((stat & 0x000f0000) == 0x00080000 ||
+		    (stat & 0x000f0000) == 0x00020000)
+			ret = retry ? 0 : 1;
+		if ((stat & 0x00000100))
+			ret = -ETIMEDOUT;
+		if ((stat & 0x00000e00))
+			ret = -EIO;
+
+		AUX_DBG("%02d 0x%08x 0x%08x\n", retries, ctrl, stat);
+	}
+
+	if (type & 1) {
+		for (i = 0; i < 16; i += 4) {
+			xbuf[i / 4] = nv_rd32(aux, 0x00d940 + (ch * 0x50) + i);
+			AUX_DBG("rd 0x%08x\n", xbuf[i / 4]);
+		}
+		memcpy(data, xbuf, size);
+	}
+
+out:
+	auxch_fini(aux, ch);
+	return ret < 0 ? ret : (stat & 0x000f0000) >> 16;
+}
+
+static const struct nouveau_i2c_func
+gm204_aux_func = {
+	.aux       = gm204_aux,
+};
+
+int
+gm204_aux_port_ctor(struct nouveau_object *parent,
+		    struct nouveau_object *engine,
+		    struct nouveau_oclass *oclass, void *data, u32 index,
+		    struct nouveau_object **pobject)
+{
+	struct dcb_i2c_entry *info = data;
+	struct nv50_i2c_port *port;
+	int ret;
+
+	ret = nouveau_i2c_port_create(parent, engine, oclass, index,
+				      &nouveau_i2c_aux_algo, &gm204_aux_func,
+				      &port);
+	*pobject = nv_object(port);
+	if (ret)
+		return ret;
+
+	port->base.aux = info->auxch;
+	port->addr = info->auxch;
+	return 0;
+}
+
+struct nouveau_oclass
+gm204_i2c_sclass[] = {
+	{ .handle = NV_I2C_TYPE_DCBI2C(DCB_I2C_NVIO_BIT),
+	  .ofuncs = &(struct nouveau_ofuncs) {
+		  .ctor = nvd0_i2c_port_ctor,
+		  .dtor = _nouveau_i2c_port_dtor,
+		  .init = nv50_i2c_port_init,
+		  .fini = _nouveau_i2c_port_fini,
+	  },
+	},
+	{ .handle = NV_I2C_TYPE_DCBI2C(DCB_I2C_NVIO_AUX),
+	  .ofuncs = &(struct nouveau_ofuncs) {
+		  .ctor = gm204_aux_port_ctor,
+		  .dtor = _nouveau_i2c_port_dtor,
+		  .init = _nouveau_i2c_port_init,
+		  .fini = _nouveau_i2c_port_fini,
+	  },
+	},
+	{}
+};
+
+struct nouveau_oclass *
+gm204_i2c_oclass = &(struct nouveau_i2c_impl) {
+	.base.handle = NV_SUBDEV(I2C, 0x24),
+	.base.ofuncs = &(struct nouveau_ofuncs) {
+		.ctor = _nouveau_i2c_ctor,
+		.dtor = _nouveau_i2c_dtor,
+		.init = _nouveau_i2c_init,
+		.fini = _nouveau_i2c_fini,
+	},
+	.sclass = gm204_i2c_sclass,
+	.pad_x = &nv04_i2c_pad_oclass,
+	.pad_s = &gm204_i2c_pad_oclass,
+	.aux = 8,
+	.aux_stat = nve0_aux_stat,
+	.aux_mask = nve0_aux_mask,
+}.base;
diff --git a/drivers/gpu/drm/nouveau/core/subdev/i2c/nv50.h b/drivers/gpu/drm/nouveau/core/subdev/i2c/nv50.h
index 5d2a77421c74..9ef965692fb1 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/i2c/nv50.h
+++ b/drivers/gpu/drm/nouveau/core/subdev/i2c/nv50.h
@@ -10,8 +10,6 @@ struct nv50_i2c_priv {
 struct nv50_i2c_port {
 	struct nouveau_i2c_port base;
 	u32 addr;
-	u32 ctrl;
-	u32 data;
 	u32 state;
 };
 
@@ -29,4 +27,8 @@ int  nv94_aux_port_ctor(struct nouveau_object *, struct nouveau_object *,
 void nv94_i2c_acquire(struct nouveau_i2c_port *);
 void nv94_i2c_release(struct nouveau_i2c_port *);
 
+int  nvd0_i2c_port_ctor(struct nouveau_object *, struct nouveau_object *,
+			struct nouveau_oclass *, void *, u32,
+			struct nouveau_object **);
+
 #endif
diff --git a/drivers/gpu/drm/nouveau/core/subdev/i2c/nv94.c b/drivers/gpu/drm/nouveau/core/subdev/i2c/nv94.c
index f59c3a255462..e383ee81f4d2 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/i2c/nv94.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/i2c/nv94.c
@@ -214,10 +214,6 @@ nv94_i2c_port_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 
 	port->state = 7;
 	port->addr = nv50_i2c_addr[info->drive];
-	if (info->share != DCB_I2C_UNUSED) {
-		port->ctrl = 0x00e500 + (info->share * 0x50);
-		port->data = 0x0000e001;
-	}
 	return 0;
 }
 
@@ -242,13 +238,8 @@ nv94_aux_port_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	if (ret)
 		return ret;
 
-	port->base.aux = info->drive;
-	port->addr = info->drive;
-	if (info->share != DCB_I2C_UNUSED) {
-		port->ctrl = 0x00e500 + (info->drive * 0x50);
-		port->data = 0x00002002;
-	}
-
+	port->base.aux = info->auxch;
+	port->addr = info->auxch;
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/nouveau/core/subdev/i2c/nvd0.c b/drivers/gpu/drm/nouveau/core/subdev/i2c/nvd0.c
index 364ddb1c5f03..fd99380502ec 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/i2c/nvd0.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/i2c/nvd0.c
@@ -48,7 +48,7 @@ nvd0_i2c_func = {
 	.sense_sda = nvd0_i2c_sense_sda,
 };
 
-static int
+int
 nvd0_i2c_port_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 		   struct nouveau_oclass *oclass, void *data, u32 index,
 		   struct nouveau_object **pobject)
@@ -66,10 +66,6 @@ nvd0_i2c_port_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 
 	port->state = 0x00000007;
 	port->addr = 0x00d014 + (info->drive * 0x20);
-	if (info->share != DCB_I2C_UNUSED) {
-		port->ctrl = 0x00e500 + (info->share * 0x50);
-		port->data = 0x0000e001;
-	}
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/nouveau/core/subdev/i2c/nve0.c b/drivers/gpu/drm/nouveau/core/subdev/i2c/nve0.c
index cae77e1ad8dc..25fe5c2d110e 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/i2c/nve0.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/i2c/nve0.c
@@ -24,7 +24,7 @@
 
 #include "nv50.h"
 
-static void
+void
 nve0_aux_stat(struct nouveau_i2c *i2c, u32 *hi, u32 *lo, u32 *rq, u32 *tx)
 {
 	u32 intr = nv_rd32(i2c, 0x00dc60);
@@ -38,7 +38,7 @@ nve0_aux_stat(struct nouveau_i2c *i2c, u32 *hi, u32 *lo, u32 *rq, u32 *tx)
 	nv_wr32(i2c, 0x00dc60, intr);
 }
 
-static void
+void
 nve0_aux_mask(struct nouveau_i2c *i2c, u32 type, u32 mask, u32 data)
 {
 	u32 temp = nv_rd32(i2c, 0x00dc68), i;
diff --git a/drivers/gpu/drm/nouveau/core/subdev/i2c/padgm204.c b/drivers/gpu/drm/nouveau/core/subdev/i2c/padgm204.c
new file mode 100644
index 000000000000..f0e6fbbaa8cd
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/subdev/i2c/padgm204.c
@@ -0,0 +1,86 @@
+/*
+ * Copyright 2014 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs
+ */
+
+#include "pad.h"
+
+struct gm204_i2c_pad {
+	struct nvkm_i2c_pad base;
+	int addr;
+};
+
+static int
+gm204_i2c_pad_fini(struct nouveau_object *object, bool suspend)
+{
+	struct nouveau_i2c *i2c = (void *)object->engine;
+	struct gm204_i2c_pad *pad = (void *)object;
+	nv_mask(i2c, 0x00d97c + pad->addr, 0x00000001, 0x00000001);
+	return nvkm_i2c_pad_fini(&pad->base, suspend);
+}
+
+static int
+gm204_i2c_pad_init(struct nouveau_object *object)
+{
+	struct nouveau_i2c *i2c = (void *)object->engine;
+	struct gm204_i2c_pad *pad = (void *)object;
+
+	switch (nv_oclass(pad->base.next)->handle) {
+	case NV_I2C_TYPE_DCBI2C(DCB_I2C_NVIO_AUX):
+		nv_mask(i2c, 0x00d970 + pad->addr, 0x0000c003, 0x00000002);
+		break;
+	case NV_I2C_TYPE_DCBI2C(DCB_I2C_NVIO_BIT):
+	default:
+		nv_mask(i2c, 0x00d970 + pad->addr, 0x0000c003, 0x0000c001);
+		break;
+	}
+
+	nv_mask(i2c, 0x00d97c + pad->addr, 0x00000001, 0x00000000);
+	return nvkm_i2c_pad_init(&pad->base);
+}
+
+static int
+gm204_i2c_pad_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
+		  struct nouveau_oclass *oclass, void *data, u32 index,
+		  struct nouveau_object **pobject)
+{
+	struct gm204_i2c_pad *pad;
+	int ret;
+
+	ret = nvkm_i2c_pad_create(parent, engine, oclass, index, &pad);
+	*pobject = nv_object(pad);
+	if (ret)
+		return ret;
+
+	pad->addr = index * 0x50;;
+	return 0;
+}
+
+struct nouveau_oclass
+gm204_i2c_pad_oclass = {
+	.ofuncs = &(struct nouveau_ofuncs) {
+		.ctor = gm204_i2c_pad_ctor,
+		.dtor = _nvkm_i2c_pad_dtor,
+		.init = gm204_i2c_pad_init,
+		.fini = gm204_i2c_pad_fini,
+	},
+};
diff --git a/drivers/gpu/drm/nouveau/core/subdev/i2c/priv.h b/drivers/gpu/drm/nouveau/core/subdev/i2c/priv.h
index 780090b6425a..4fe7ae3fde4e 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/i2c/priv.h
+++ b/drivers/gpu/drm/nouveau/core/subdev/i2c/priv.h
@@ -5,6 +5,7 @@
 
 extern struct nouveau_oclass nv04_i2c_pad_oclass;
 extern struct nouveau_oclass nv94_i2c_pad_oclass;
+extern struct nouveau_oclass gm204_i2c_pad_oclass;
 
 #define nouveau_i2c_port_create(p,e,o,i,a,f,d)                                 \
 	nouveau_i2c_port_create_((p), (e), (o), (i), (a), (f),                 \
@@ -82,4 +83,7 @@ struct nouveau_i2c_impl {
 void nv94_aux_stat(struct nouveau_i2c *, u32 *, u32 *, u32 *, u32 *);
 void nv94_aux_mask(struct nouveau_i2c *, u32, u32, u32);
 
+void nve0_aux_stat(struct nouveau_i2c *, u32 *, u32 *, u32 *, u32 *);
+void nve0_aux_mask(struct nouveau_i2c *, u32, u32, u32);
+
 #endif
diff --git a/drivers/gpu/drm/nouveau/core/subdev/instmem/nv04.c b/drivers/gpu/drm/nouveau/core/subdev/instmem/nv04.c
index 7b64befee48f..e8b1401c59c0 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/instmem/nv04.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/instmem/nv04.c
@@ -69,7 +69,7 @@ nv04_instobj_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	if (ret)
 		return ret;
 
-	ret = nouveau_mm_head(&priv->heap, 1, args->size, args->size,
+	ret = nouveau_mm_head(&priv->heap, 0, 1, args->size, args->size,
 			      args->align, &node->mem);
 	if (ret)
 		return ret;
diff --git a/drivers/gpu/drm/nouveau/core/subdev/ltc/base.c b/drivers/gpu/drm/nouveau/core/subdev/ltc/base.c
index 32ed442c5913..7fa331516f84 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/ltc/base.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/ltc/base.c
@@ -31,7 +31,7 @@ nvkm_ltc_tags_alloc(struct nouveau_ltc *ltc, u32 n,
 	struct nvkm_ltc_priv *priv = (void *)ltc;
 	int ret;
 
-	ret = nouveau_mm_head(&priv->tags, 1, n, n, 1, pnode);
+	ret = nouveau_mm_head(&priv->tags, 0, 1, n, n, 1, pnode);
 	if (ret)
 		*pnode = NULL;
 
diff --git a/drivers/gpu/drm/nouveau/core/subdev/ltc/gf100.c b/drivers/gpu/drm/nouveau/core/subdev/ltc/gf100.c
index d5d65285efe5..2db0977284f8 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/ltc/gf100.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/ltc/gf100.c
@@ -62,16 +62,38 @@ gf100_ltc_zbc_clear_depth(struct nvkm_ltc_priv *priv, int i, const u32 depth)
 	nv_wr32(priv, 0x17ea58, depth);
 }
 
+static const struct nouveau_bitfield
+gf100_ltc_lts_intr_name[] = {
+	{ 0x00000001, "IDLE_ERROR_IQ" },
+	{ 0x00000002, "IDLE_ERROR_CBC" },
+	{ 0x00000004, "IDLE_ERROR_TSTG" },
+	{ 0x00000008, "IDLE_ERROR_DSTG" },
+	{ 0x00000010, "EVICTED_CB" },
+	{ 0x00000020, "ILLEGAL_COMPSTAT" },
+	{ 0x00000040, "BLOCKLINEAR_CB" },
+	{ 0x00000100, "ECC_SEC_ERROR" },
+	{ 0x00000200, "ECC_DED_ERROR" },
+	{ 0x00000400, "DEBUG" },
+	{ 0x00000800, "ATOMIC_TO_Z" },
+	{ 0x00001000, "ILLEGAL_ATOMIC" },
+	{ 0x00002000, "BLKACTIVITY_ERR" },
+	{}
+};
+
 static void
-gf100_ltc_lts_isr(struct nvkm_ltc_priv *priv, int ltc, int lts)
+gf100_ltc_lts_intr(struct nvkm_ltc_priv *priv, int ltc, int lts)
 {
 	u32 base = 0x141000 + (ltc * 0x2000) + (lts * 0x400);
-	u32 stat = nv_rd32(priv, base + 0x020);
+	u32 intr = nv_rd32(priv, base + 0x020);
+	u32 stat = intr & 0x0000ffff;
 
 	if (stat) {
-		nv_info(priv, "LTC%d_LTS%d: 0x%08x\n", ltc, lts, stat);
-		nv_wr32(priv, base + 0x020, stat);
+		nv_info(priv, "LTC%d_LTS%d:", ltc, lts);
+		nouveau_bitfield_print(gf100_ltc_lts_intr_name, stat);
+		pr_cont("\n");
 	}
+
+	nv_wr32(priv, base + 0x020, intr);
 }
 
 void
@@ -84,14 +106,9 @@ gf100_ltc_intr(struct nouveau_subdev *subdev)
 	while (mask) {
 		u32 lts, ltc = __ffs(mask);
 		for (lts = 0; lts < priv->lts_nr; lts++)
-			gf100_ltc_lts_isr(priv, ltc, lts);
+			gf100_ltc_lts_intr(priv, ltc, lts);
 		mask &= ~(1 << ltc);
 	}
-
-	/* we do something horribly wrong and upset PMFB a lot, so mask off
-	 * interrupts from it after the first one until it's fixed
-	 */
-	nv_mask(priv, 0x000640, 0x02000000, 0x00000000);
 }
 
 static int
@@ -153,7 +170,7 @@ gf100_ltc_init_tag_ram(struct nouveau_fb *pfb, struct nvkm_ltc_priv *priv)
 	tag_size += tag_align;
 	tag_size  = (tag_size + 0xfff) >> 12; /* round up */
 
-	ret = nouveau_mm_tail(&pfb->vram, 1, tag_size, tag_size, 1,
+	ret = nouveau_mm_tail(&pfb->vram, 1, 1, tag_size, tag_size, 1,
 	                      &priv->tag_ram);
 	if (ret) {
 		priv->num_tags = 0;
diff --git a/drivers/gpu/drm/nouveau/core/subdev/ltc/gm107.c b/drivers/gpu/drm/nouveau/core/subdev/ltc/gm107.c
index a4de64289762..89fc4238f50c 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/ltc/gm107.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/ltc/gm107.c
@@ -87,11 +87,6 @@ gm107_ltc_intr(struct nouveau_subdev *subdev)
 			gm107_ltc_lts_isr(priv, ltc, lts);
 		mask &= ~(1 << ltc);
 	}
-
-	/* we do something horribly wrong and upset PMFB a lot, so mask off
-	 * interrupts from it after the first one until it's fixed
-	 */
-	nv_mask(priv, 0x000640, 0x02000000, 0x00000000);
 }
 
 static int
diff --git a/drivers/gpu/drm/nouveau/core/subdev/ltc/priv.h b/drivers/gpu/drm/nouveau/core/subdev/ltc/priv.h
index 594924f39126..41f179d93da6 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/ltc/priv.h
+++ b/drivers/gpu/drm/nouveau/core/subdev/ltc/priv.h
@@ -4,6 +4,8 @@
 #include <subdev/ltc.h>
 #include <subdev/fb.h>
 
+#include <core/enum.h>
+
 struct nvkm_ltc_priv {
 	struct nouveau_ltc base;
 	u32 ltc_nr;
diff --git a/drivers/gpu/drm/nouveau/core/subdev/mc/nv4c.c b/drivers/gpu/drm/nouveau/core/subdev/mc/nv4c.c
index a75c35ccf25c..165401c4045c 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/mc/nv4c.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/mc/nv4c.c
@@ -24,13 +24,6 @@
 
 #include "nv04.h"
 
-static void
-nv4c_mc_msi_rearm(struct nouveau_mc *pmc)
-{
-	struct nv04_mc_priv *priv = (void *)pmc;
-	nv_wr08(priv, 0x088050, 0xff);
-}
-
 struct nouveau_oclass *
 nv4c_mc_oclass = &(struct nouveau_mc_oclass) {
 	.base.handle = NV_SUBDEV(MC, 0x4c),
@@ -41,5 +34,4 @@ nv4c_mc_oclass = &(struct nouveau_mc_oclass) {
 		.fini = _nouveau_mc_fini,
 	},
 	.intr = nv04_mc_intr,
-	.msi_rearm = nv4c_mc_msi_rearm,
 }.base;
diff --git a/drivers/gpu/drm/nouveau/core/subdev/pwr/base.c b/drivers/gpu/drm/nouveau/core/subdev/pwr/base.c
index 69f1f34f6931..0ab55f27ec45 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/pwr/base.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/pwr/base.c
@@ -203,6 +203,8 @@ _nouveau_pwr_init(struct nouveau_object *object)
 	nv_wait(ppwr, 0x10a04c, 0xffffffff, 0x00000000);
 	nv_mask(ppwr, 0x000200, 0x00002000, 0x00000000);
 	nv_mask(ppwr, 0x000200, 0x00002000, 0x00002000);
+	nv_rd32(ppwr, 0x000200);
+	nv_wait(ppwr, 0x10a10c, 0x00000006, 0x00000000);
 
 	/* upload data segment */
 	nv_wr32(ppwr, 0x10a1c0, 0x01000000);
diff --git a/drivers/gpu/drm/nouveau/core/subdev/pwr/fuc/arith.fuc b/drivers/gpu/drm/nouveau/core/subdev/pwr/fuc/arith.fuc
new file mode 100644
index 000000000000..214a6d9e088d
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/subdev/pwr/fuc/arith.fuc
@@ -0,0 +1,94 @@
+/*
+ * Copyright 2014 Martin Peres <martin.peres@free.fr>
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the folloing conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Martin Peres
+ */
+
+/******************************************************************************
+ * arith data segment
+ *****************************************************************************/
+#ifdef INCLUDE_PROC
+#endif
+
+#ifdef INCLUDE_DATA
+#endif
+
+/******************************************************************************
+ * arith code segment
+ *****************************************************************************/
+#ifdef INCLUDE_CODE
+
+// does a 32x32 -> 64 multiplication
+//
+// A * B = A_lo * B_lo
+//        + ( A_hi * B_lo ) << 16
+//        + ( A_lo * B_hi ) << 16
+//        + ( A_hi * B_hi ) << 32
+//
+// $r15 - current
+// $r14 - A
+// $r13 - B
+// $r12 - mul_lo (return)
+// $r11 - mul_hi (return)
+// $r0  - zero
+mulu32_32_64:
+	push $r1 // A_hi
+	push $r2 // B_hi
+	push $r3 // tmp0
+	push $r4 // tmp1
+
+	shr b32 $r1 $r14 16
+	shr b32 $r2 $r13 16
+
+	clear b32 $r12
+	clear b32 $r11
+
+	// A_lo * B_lo
+	mulu $r12 $r14 $r13
+
+	// ( A_hi * B_lo ) << 16
+	mulu $r3 $r1 $r13 // tmp0 = A_hi * B_lo
+	mov b32 $r4 $r3
+	and $r3 0xffff // tmp0 = tmp0_lo
+	shl b32 $r3 16
+	shr b32 $r4 16 // tmp1 = tmp0_hi
+	add b32 $r12 $r3
+	adc b32 $r11 $r4
+
+	// ( A_lo * B_hi ) << 16
+	mulu $r3 $r14 $r2 // tmp0 = A_lo * B_hi
+	mov b32 $r4 $r3
+	and $r3 0xffff // tmp0 = tmp0_lo
+	shl b32 $r3 16
+	shr b32 $r4 16 // tmp1 = tmp0_hi
+	add b32 $r12 $r3
+	adc b32 $r11 $r4
+
+	// ( A_hi * B_hi ) << 32
+	mulu $r3 $r1 $r2 // tmp0 = A_hi * B_hi
+	add b32 $r11 $r3
+
+	pop $r4
+	pop $r3
+	pop $r2
+	pop $r1
+	ret
+#endif
diff --git a/drivers/gpu/drm/nouveau/core/subdev/pwr/fuc/kernel.fuc b/drivers/gpu/drm/nouveau/core/subdev/pwr/fuc/kernel.fuc
index 8f29badd785f..5cf5be63cbef 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/pwr/fuc/kernel.fuc
+++ b/drivers/gpu/drm/nouveau/core/subdev/pwr/fuc/kernel.fuc
@@ -98,12 +98,16 @@ wr32:
 // $r14 - ns
 // $r0  - zero
 nsec:
+	push $r9
+	push $r8
 	nv_iord($r8, NV_PPWR_TIMER_LOW)
 	nsec_loop:
 		nv_iord($r9, NV_PPWR_TIMER_LOW)
 		sub b32 $r9 $r8
 		cmp b32 $r9 $r14
 		bra l #nsec_loop
+	pop $r8
+	pop $r9
 	ret
 
 // busy-wait for a period of time
@@ -115,6 +119,8 @@ nsec:
 // $r11 - timeout (ns)
 // $r0  - zero
 wait:
+	push $r9
+	push $r8
 	nv_iord($r8, NV_PPWR_TIMER_LOW)
 	wait_loop:
 		nv_rd32($r10, $r14)
@@ -126,6 +132,8 @@ wait:
 		cmp b32 $r9 $r11
 		bra l #wait_loop
 	wait_done:
+	pop $r8
+	pop $r9
 	ret
 
 // $r15 - current (kern)
@@ -242,12 +250,89 @@ intr:
 	bclr $flags $p0
 	iret
 
-// request the current process be sent a message after a timeout expires
+// calculate the number of ticks in the specified nanoseconds delay
+//
+// $r15 - current
+// $r14 - ns
+// $r14 - ticks (return)
+// $r0  - zero
+ticks_from_ns:
+	push $r12
+	push $r11
+
+	/* try not losing precision (multiply then divide) */
+	imm32($r13, HW_TICKS_PER_US)
+	call #mulu32_32_64
+
+	/* use an immeditate, it's ok because HW_TICKS_PER_US < 16 bits */
+	div $r12 $r12 1000
+
+	/* check if there wasn't any overflow */
+	cmpu b32 $r11 0
+	bra e #ticks_from_ns_quit
+
+	/* let's divide then multiply, too bad for the precision! */
+	div $r14 $r14 1000
+	imm32($r13, HW_TICKS_PER_US)
+	call #mulu32_32_64
+
+	/* this cannot overflow as long as HW_TICKS_PER_US < 1000 */
+
+ticks_from_ns_quit:
+	mov b32 $r14 $r12
+	pop $r11
+	pop $r12
+	ret
+
+// calculate the number of ticks in the specified microsecond delay
+//
+// $r15 - current
+// $r14 - us
+// $r14 - ticks (return)
+// $r0  - zero
+ticks_from_us:
+	push $r12
+	push $r11
+
+	/* simply multiply $us by HW_TICKS_PER_US */
+	imm32($r13, HW_TICKS_PER_US)
+	call #mulu32_32_64
+	mov b32 $r14 $r12
+
+	/* check if there wasn't any overflow */
+	cmpu b32 $r11 0
+	bra e #ticks_from_us_quit
+
+	/* Overflow! */
+	clear b32 $r14
+
+ticks_from_us_quit:
+	pop $r11
+	pop $r12
+	ret
+
+// calculate the number of ticks in the specified microsecond delay
 //
 // $r15 - current
 // $r14 - ticks
+// $r14 - us (return)
+// $r0  - zero
+ticks_to_us:
+	/* simply divide $ticks by HW_TICKS_PER_US */
+	imm32($r13, HW_TICKS_PER_US)
+	div $r14 $r14 $r13
+
+	ret
+
+// request the current process be sent a message after a timeout expires
+//
+// $r15 - current
+// $r14 - ticks (make sure it is < 2^31 to avoid any possible overflow)
 // $r0  - zero
 timer:
+	push $r9
+	push $r8
+
 	// interrupts off to prevent racing with timer isr
 	bclr $flags ie0
 
@@ -255,13 +340,22 @@ timer:
 	ld b32 $r8 D[$r15 + #proc_time]
 	cmp b32 $r8 0
 	bra g #timer_done
-	st b32 D[$r15 + #proc_time] $r14
 
-	// halt watchdog timer temporarily and check for a pending
-	// interrupt.  if there's one already pending, we can just
-	// bail since the timer isr will queue the next soonest
-	// right after it's done
+	// halt watchdog timer temporarily
+	clear b32 $r8
 	nv_iowr(NV_PPWR_WATCHDOG_ENABLE, $r8)
+
+	// find out how much time elapsed since the last update
+	// of the watchdog and add this time to the wanted ticks
+	nv_iord($r8, NV_PPWR_WATCHDOG_TIME)
+	ld b32 $r9 D[$r0 + #time_prev]
+	sub b32 $r9 $r8
+	add b32 $r14 $r9
+	st b32 D[$r15 + #proc_time] $r14
+
+	// check for a pending interrupt.  if there's one already
+	// pending, we can just bail since the timer isr will
+	// queue the next soonest right after it's done
 	nv_iord($r8, NV_PPWR_INTR)
 	and $r8 NV_PPWR_INTR_WATCHDOG
 	bra nz #timer_enable
@@ -272,10 +366,10 @@ timer:
 	cmp b32 $r14 $r0
 	bra e #timer_reset
 	cmp b32 $r14 $r8
-	bra l #timer_done
-	timer_reset:
-	nv_iowr(NV_PPWR_WATCHDOG_TIME, $r14)
-	st b32 D[$r0 + #time_prev] $r14
+	bra g #timer_enable
+		timer_reset:
+		nv_iowr(NV_PPWR_WATCHDOG_TIME, $r14)
+		st b32 D[$r0 + #time_prev] $r14
 
 	// re-enable the watchdog timer
 	timer_enable:
@@ -285,6 +379,9 @@ timer:
 	// interrupts back on
 	timer_done:
 	bset $flags ie0
+
+	pop $r8
+	pop $r9
 	ret
 
 // send message to another process
@@ -371,6 +468,9 @@ send:
 // $r14 - process
 // $r0  - zero
 recv:
+	push $r9
+	push $r8
+
 	ld b32 $r8 D[$r14 + #proc_qget]
 	ld b32 $r9 D[$r14 + #proc_qput]
 	bclr $flags $p1
@@ -403,6 +503,8 @@ recv:
 		bset $flags $p1
 		pop $r15
 	recv_done:
+	pop $r8
+	pop $r9
 	ret
 
 init:
diff --git a/drivers/gpu/drm/nouveau/core/subdev/pwr/fuc/macros.fuc b/drivers/gpu/drm/nouveau/core/subdev/pwr/fuc/macros.fuc
index 5668e045bac1..96fc984dafdc 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/pwr/fuc/macros.fuc
+++ b/drivers/gpu/drm/nouveau/core/subdev/pwr/fuc/macros.fuc
@@ -250,3 +250,23 @@
 */	st b32 D[$r0] reg /*
 */	clear b32 $r0
 #endif
+
+#define st(size, addr, reg) /*
+*/	movw $r0 addr /*
+*/	st size D[$r0] reg /*
+*/	clear b32 $r0
+
+#define ld(size, reg, addr) /*
+*/	movw $r0 addr /*
+*/	ld size reg D[$r0] /*
+*/	clear b32 $r0
+
+// does a 64+64 -> 64 unsigned addition (C = A + B)
+#define addu64(reg_a_c_hi, reg_a_c_lo, b_hi, b_lo) /*
+*/    add b32 reg_a_c_lo b_lo /*
+*/    adc b32 reg_a_c_hi b_hi
+
+// does a 64+64 -> 64 substraction (C = A - B)
+#define subu64(reg_a_c_hi, reg_a_c_lo, b_hi, b_lo) /*
+*/    sub b32 reg_a_c_lo b_lo /*
+*/    sbb b32 reg_a_c_hi b_hi
diff --git a/drivers/gpu/drm/nouveau/core/subdev/pwr/fuc/memx.fuc b/drivers/gpu/drm/nouveau/core/subdev/pwr/fuc/memx.fuc
index d43741eccb11..ec03f9a4290b 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/pwr/fuc/memx.fuc
+++ b/drivers/gpu/drm/nouveau/core/subdev/pwr/fuc/memx.fuc
@@ -43,20 +43,31 @@ process(PROC_MEMX, #memx_init, #memx_recv)
 */	.b32 func
 
 memx_func_head:
-handler(ENTER , 0x0001, 0x0000, #memx_func_enter)
+handler(ENTER , 0x0000, 0x0000, #memx_func_enter)
 memx_func_next:
 handler(LEAVE , 0x0000, 0x0000, #memx_func_leave)
 handler(WR32  , 0x0000, 0x0002, #memx_func_wr32)
 handler(WAIT  , 0x0004, 0x0000, #memx_func_wait)
 handler(DELAY , 0x0001, 0x0000, #memx_func_delay)
+handler(VBLANK, 0x0001, 0x0000, #memx_func_wait_vblank)
+handler(TRAIN , 0x0000, 0x0000, #memx_func_train)
 memx_func_tail:
 
 .equ #memx_func_size #memx_func_next - #memx_func_head
 .equ #memx_func_num (#memx_func_tail - #memx_func_head) / #memx_func_size
 
+memx_ts_start:
+.b32 0
+memx_ts_end:
+.b32 0
+
 memx_data_head:
 .skip 0x0800
 memx_data_tail:
+
+memx_train_head:
+.skip 0x0100
+memx_train_tail:
 #endif
 
 /******************************************************************************
@@ -67,19 +78,44 @@ memx_data_tail:
 //
 // $r15 - current (memx)
 // $r4  - packet length
-//	+00: bitmask of heads to wait for vblank on
 // $r3  - opcode desciption
 // $r0  - zero
 memx_func_enter:
+#if NVKM_PPWR_CHIPSET == GT215
+	movw $r8 0x1610
+	nv_rd32($r7, $r8)
+	imm32($r6, 0xfffffffc)
+	and $r7 $r6
+	movw $r6 0x2
+	or $r7 $r6
+	nv_wr32($r8, $r7)
+#else
+	movw $r6 0x001620
+	imm32($r7, ~0x00000aa2);
+	nv_rd32($r8, $r6)
+	and $r8 $r7
+	nv_wr32($r6, $r8)
+
+	imm32($r7, ~0x00000001)
+	nv_rd32($r8, $r6)
+	and $r8 $r7
+	nv_wr32($r6, $r8)
+
+	movw $r6 0x0026f0
+	nv_rd32($r8, $r6)
+	and $r8 $r7
+	nv_wr32($r6, $r8)
+#endif
+
 	mov $r6 NV_PPWR_OUTPUT_SET_FB_PAUSE
 	nv_iowr(NV_PPWR_OUTPUT_SET, $r6)
 	memx_func_enter_wait:
 		nv_iord($r6, NV_PPWR_OUTPUT)
 		and $r6 NV_PPWR_OUTPUT_FB_PAUSE
 		bra z #memx_func_enter_wait
-	//XXX: TODO
-	ld b32 $r6 D[$r1 + 0x00]
-	add b32 $r1 0x04
+
+	nv_iord($r6, NV_PPWR_TIMER_LOW)
+	st b32 D[$r0 + #memx_ts_start] $r6
 	ret
 
 // description
@@ -89,14 +125,93 @@ memx_func_enter:
 // $r3  - opcode desciption
 // $r0  - zero
 memx_func_leave:
+	nv_iord($r6, NV_PPWR_TIMER_LOW)
+	st b32 D[$r0 + #memx_ts_end] $r6
+
 	mov $r6 NV_PPWR_OUTPUT_CLR_FB_PAUSE
 	nv_iowr(NV_PPWR_OUTPUT_CLR, $r6)
 	memx_func_leave_wait:
 		nv_iord($r6, NV_PPWR_OUTPUT)
 		and $r6 NV_PPWR_OUTPUT_FB_PAUSE
 		bra nz #memx_func_leave_wait
+
+#if NVKM_PPWR_CHIPSET == GT215
+	movw $r8 0x1610
+	nv_rd32($r7, $r8)
+	imm32($r6, 0xffffffcc)
+	and $r7 $r6
+	nv_wr32($r8, $r7)
+#else
+	movw $r6 0x0026f0
+	imm32($r7, 0x00000001)
+	nv_rd32($r8, $r6)
+	or $r8 $r7
+	nv_wr32($r6, $r8)
+
+	movw $r6 0x001620
+	nv_rd32($r8, $r6)
+	or $r8 $r7
+	nv_wr32($r6, $r8)
+
+	imm32($r7, 0x00000aa2);
+	nv_rd32($r8, $r6)
+	or $r8 $r7
+	nv_wr32($r6, $r8)
+#endif
 	ret
 
+#if NVKM_PPWR_CHIPSET < GF119
+// description
+//
+// $r15 - current (memx)
+// $r4  - packet length
+//	+00: head to wait for vblank on
+// $r3  - opcode desciption
+// $r0  - zero
+memx_func_wait_vblank:
+	ld b32 $r6 D[$r1 + 0x00]
+	cmp b32 $r6 0x0
+	bra z #memx_func_wait_vblank_head0
+	cmp b32 $r6 0x1
+	bra z #memx_func_wait_vblank_head1
+	bra #memx_func_wait_vblank_fini
+
+	memx_func_wait_vblank_head1:
+	movw $r7 0x20
+	bra #memx_func_wait_vblank_0
+
+	memx_func_wait_vblank_head0:
+	movw $r7 0x8
+
+	memx_func_wait_vblank_0:
+		nv_iord($r6, NV_PPWR_INPUT)
+		and $r6 $r7
+		bra nz #memx_func_wait_vblank_0
+
+	memx_func_wait_vblank_1:
+		nv_iord($r6, NV_PPWR_INPUT)
+		and $r6 $r7
+		bra z #memx_func_wait_vblank_1
+
+	memx_func_wait_vblank_fini:
+	add b32 $r1 0x4
+	ret
+
+#else
+
+// XXX: currently no-op
+//
+// $r15 - current (memx)
+// $r4  - packet length
+//	+00: head to wait for vblank on
+// $r3  - opcode desciption
+// $r0  - zero
+memx_func_wait_vblank:
+	add b32 $r1 0x4
+	ret
+
+#endif
+
 // description
 //
 // $r15 - current (memx)
@@ -150,6 +265,101 @@ memx_func_delay:
 // description
 //
 // $r15 - current (memx)
+// $r4  - packet length
+// $r3  - opcode desciption
+// $r0  - zero
+memx_func_train:
+#if NVKM_PPWR_CHIPSET == GT215
+// $r5 - outer loop counter
+// $r6 - inner loop counter
+// $r7 - entry counter (#memx_train_head + $r7)
+	movw $r5 0x3
+	movw $r7 0x0
+
+// Read random memory to wake up... things
+	imm32($r9, 0x700000)
+	nv_rd32($r8,$r9)
+	movw $r14 0x2710
+	call(nsec)
+
+	memx_func_train_loop_outer:
+		mulu $r8 $r5 0x101
+		sethi $r8 0x02000000
+		imm32($r9, 0x1111e0)
+		nv_wr32($r9, $r8)
+		push $r5
+
+		movw $r6 0x0
+		memx_func_train_loop_inner:
+			movw $r8 0x1111
+			mulu $r9 $r6 $r8
+			shl b32 $r8 $r9 0x10
+			or $r8 $r9
+			imm32($r9, 0x100720)
+			nv_wr32($r9, $r8)
+
+			imm32($r9, 0x100080)
+			nv_rd32($r8, $r9)
+			or $r8 $r8 0x20
+			nv_wr32($r9, $r8)
+
+			imm32($r9, 0x10053c)
+			imm32($r8, 0x80003002)
+			nv_wr32($r9, $r8)
+
+			imm32($r14, 0x100560)
+			imm32($r13, 0x80000000)
+			add b32 $r12 $r13 0
+			imm32($r11, 0x001e8480)
+			call(wait)
+
+			// $r5 - inner inner loop counter
+			// $r9 - result
+			movw $r5 0
+			imm32($r9, 0x8300ffff)
+			memx_func_train_loop_4x:
+				imm32($r10, 0x100080)
+				nv_rd32($r8, $r10)
+				imm32($r11, 0xffffffdf)
+				and $r8 $r11
+				nv_wr32($r10, $r8)
+
+				imm32($r10, 0x10053c)
+				imm32($r8, 0x80003002)
+				nv_wr32($r10, $r8)
+
+				imm32($r14, 0x100560)
+				imm32($r13, 0x80000000)
+				mov b32 $r12 $r13
+				imm32($r11, 0x00002710)
+				call(wait)
+
+				nv_rd32($r13, $r14)
+				and $r9 $r9 $r13
+
+				add b32 $r5 1
+				cmp b16 $r5 0x4
+				bra l #memx_func_train_loop_4x
+
+			add b32 $r10 $r7 #memx_train_head
+			st b32 D[$r10 + 0] $r9
+			add b32 $r6 1
+			add b32 $r7 4
+
+			cmp b16 $r6 0x10
+			bra l #memx_func_train_loop_inner
+
+		pop $r5
+		add b32 $r5 1
+		cmp b16 $r5 7
+		bra l #memx_func_train_loop_outer
+
+#endif
+	ret
+
+// description
+//
+// $r15 - current (memx)
 // $r14 - sender process name
 // $r13 - message (exec)
 // $r12 - head of script
@@ -160,14 +370,17 @@ memx_exec:
 	push $r13
 	mov b32 $r1 $r12
 	mov b32 $r2 $r11
+
 	memx_exec_next:
-		// fetch the packet header, and locate opcode info
+		// fetch the packet header
 		ld b32 $r3 D[$r1]
 		add b32 $r1 4
-		shr b32 $r4 $r3 16
-		mulu $r3 #memx_func_size
+		extr $r4 $r3 16:31
+		extr $r3 $r3 0:15
 
 		// execute the opcode handler
+		sub b32 $r3 1
+		mulu $r3 #memx_func_size
 		ld b32 $r5 D[$r3 + #memx_func_head + #memx_func]
 		call $r5
 
@@ -176,6 +389,10 @@ memx_exec:
 		bra l #memx_exec_next
 
 	// send completion reply
+	ld b32 $r11 D[$r0 + #memx_ts_start]
+	ld b32 $r12 D[$r0 + #memx_ts_end]
+	sub b32 $r12 $r11
+	nv_iord($r11, NV_PPWR_INPUT)
 	pop $r13
 	pop $r14
 	call(send)
@@ -190,8 +407,19 @@ memx_exec:
 // $r11 - data1
 // $r0  - zero
 memx_info:
+	cmp b16 $r12 0x1
+	bra e #memx_info_train
+
+	memx_info_data:
 	mov $r12 #memx_data_head
 	mov $r11 #memx_data_tail - #memx_data_head
+	bra #memx_info_send
+
+	memx_info_train:
+	mov $r12 #memx_train_head
+	mov $r11 #memx_train_tail - #memx_train_head
+
+	memx_info_send:
 	call(send)
 	ret
 
diff --git a/drivers/gpu/drm/nouveau/core/subdev/pwr/fuc/nv108.fuc b/drivers/gpu/drm/nouveau/core/subdev/pwr/fuc/nv108.fuc
index 17a8a383d91a..b439519ec866 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/pwr/fuc/nv108.fuc
+++ b/drivers/gpu/drm/nouveau/core/subdev/pwr/fuc/nv108.fuc
@@ -23,6 +23,7 @@
  */
 
 #define NVKM_PPWR_CHIPSET GK208
+#define HW_TICKS_PER_US 324
 
 #define NVKM_FALCON_PC24
 #define NVKM_FALCON_UNSHIFTED_IO
@@ -34,6 +35,7 @@
 .section #nv108_pwr_data
 #define INCLUDE_PROC
 #include "kernel.fuc"
+#include "arith.fuc"
 #include "host.fuc"
 #include "memx.fuc"
 #include "perf.fuc"
@@ -44,6 +46,7 @@
 
 #define INCLUDE_DATA
 #include "kernel.fuc"
+#include "arith.fuc"
 #include "host.fuc"
 #include "memx.fuc"
 #include "perf.fuc"
@@ -56,6 +59,7 @@
 .section #nv108_pwr_code
 #define INCLUDE_CODE
 #include "kernel.fuc"
+#include "arith.fuc"
 #include "host.fuc"
 #include "memx.fuc"
 #include "perf.fuc"
diff --git a/drivers/gpu/drm/nouveau/core/subdev/pwr/fuc/nv108.fuc.h b/drivers/gpu/drm/nouveau/core/subdev/pwr/fuc/nv108.fuc.h
index 986495d533dd..713e11e2953d 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/pwr/fuc/nv108.fuc.h
+++ b/drivers/gpu/drm/nouveau/core/subdev/pwr/fuc/nv108.fuc.h
@@ -24,8 +24,8 @@ uint32_t nv108_pwr_data[] = {
 	0x00000000,
 /* 0x0058: proc_list_head */
 	0x54534f48,
-	0x00000379,
-	0x0000032a,
+	0x00000453,
+	0x00000404,
 	0x00000000,
 	0x00000000,
 	0x00000000,
@@ -46,8 +46,8 @@ uint32_t nv108_pwr_data[] = {
 	0x00000000,
 	0x00000000,
 	0x584d454d,
-	0x00000464,
-	0x00000456,
+	0x0000062d,
+	0x0000061f,
 	0x00000000,
 	0x00000000,
 	0x00000000,
@@ -68,8 +68,8 @@ uint32_t nv108_pwr_data[] = {
 	0x00000000,
 	0x00000000,
 	0x46524550,
-	0x00000468,
-	0x00000466,
+	0x00000631,
+	0x0000062f,
 	0x00000000,
 	0x00000000,
 	0x00000000,
@@ -90,8 +90,8 @@ uint32_t nv108_pwr_data[] = {
 	0x00000000,
 	0x00000000,
 	0x5f433249,
-	0x0000086c,
-	0x00000713,
+	0x00000a35,
+	0x000008dc,
 	0x00000000,
 	0x00000000,
 	0x00000000,
@@ -112,8 +112,8 @@ uint32_t nv108_pwr_data[] = {
 	0x00000000,
 	0x00000000,
 	0x54534554,
-	0x0000088d,
-	0x0000086e,
+	0x00000a56,
+	0x00000a37,
 	0x00000000,
 	0x00000000,
 	0x00000000,
@@ -134,8 +134,8 @@ uint32_t nv108_pwr_data[] = {
 	0x00000000,
 	0x00000000,
 	0x454c4449,
-	0x00000898,
-	0x00000896,
+	0x00000a61,
+	0x00000a5f,
 	0x00000000,
 	0x00000000,
 	0x00000000,
@@ -227,24 +227,98 @@ uint32_t nv108_pwr_data[] = {
 	0x00000000,
 	0x00000000,
 /* 0x0370: memx_func_head */
-	0x00010000,
-	0x00000000,
-	0x000003a9,
-/* 0x037c: memx_func_next */
 	0x00000001,
 	0x00000000,
-	0x000003c7,
+	0x00000483,
+/* 0x037c: memx_func_next */
 	0x00000002,
+	0x00000000,
+	0x00000500,
+	0x00000003,
 	0x00000002,
-	0x000003df,
-	0x00040003,
+	0x00000580,
+	0x00040004,
+	0x00000000,
+	0x0000059d,
+	0x00010005,
+	0x00000000,
+	0x000005b7,
+	0x00010006,
+	0x00000000,
+	0x0000057b,
+	0x00000007,
+	0x00000000,
+	0x000005c3,
+/* 0x03c4: memx_func_tail */
+/* 0x03c4: memx_ts_start */
+	0x00000000,
+/* 0x03c8: memx_ts_end */
+	0x00000000,
+/* 0x03cc: memx_data_head */
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
 	0x00000000,
-	0x000003fc,
-	0x00010004,
 	0x00000000,
-	0x00000416,
-/* 0x03ac: memx_func_tail */
-/* 0x03ac: memx_data_head */
 	0x00000000,
 	0x00000000,
 	0x00000000,
@@ -693,6 +767,8 @@ uint32_t nv108_pwr_data[] = {
 	0x00000000,
 	0x00000000,
 	0x00000000,
+/* 0x0bcc: memx_data_tail */
+/* 0x0bcc: memx_train_head */
 	0x00000000,
 	0x00000000,
 	0x00000000,
@@ -757,8 +833,8 @@ uint32_t nv108_pwr_data[] = {
 	0x00000000,
 	0x00000000,
 	0x00000000,
-/* 0x0bac: memx_data_tail */
-/* 0x0bac: i2c_scl_map */
+/* 0x0ccc: memx_train_tail */
+/* 0x0ccc: i2c_scl_map */
 	0x00000400,
 	0x00000800,
 	0x00001000,
@@ -769,7 +845,7 @@ uint32_t nv108_pwr_data[] = {
 	0x00020000,
 	0x00040000,
 	0x00080000,
-/* 0x0bd4: i2c_sda_map */
+/* 0x0cf4: i2c_sda_map */
 	0x00100000,
 	0x00200000,
 	0x00400000,
@@ -781,10 +857,66 @@ uint32_t nv108_pwr_data[] = {
 	0x10000000,
 	0x20000000,
 	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
 };
 
 uint32_t nv108_pwr_code[] = {
-	0x02910ef5,
+	0x031c0ef5,
 /* 0x0004: rd32 */
 	0xf607a040,
 	0x04bd000e,
@@ -812,15 +944,18 @@ uint32_t nv108_pwr_code[] = {
 	0x7000d4f1,
 	0xf8f61bf4,
 /* 0x005d: nsec */
-	0xcf2c0800,
-/* 0x0062: nsec_loop */
+	0xf990f900,
+	0xcf2c0880,
+/* 0x0066: nsec_loop */
 	0x2c090088,
 	0xbb0099cf,
 	0x9ea60298,
-	0xf8f61ef4,
-/* 0x0071: wait */
-	0xcf2c0800,
-/* 0x0076: wait_loop */
+	0xfcf61ef4,
+	0xf890fc80,
+/* 0x0079: wait */
+	0xf990f900,
+	0xcf2c0880,
+/* 0x0082: wait_loop */
 	0xeeb20088,
 	0x0000047e,
 	0xadfddab2,
@@ -828,28 +963,29 @@ uint32_t nv108_pwr_code[] = {
 	0x2c09100b,
 	0xbb0099cf,
 	0x9ba60298,
-/* 0x0093: wait_done */
-	0xf8e61ef4,
-/* 0x0095: intr_watchdog */
+/* 0x009f: wait_done */
+	0xfce61ef4,
+	0xf890fc80,
+/* 0x00a5: intr_watchdog */
 	0x03e99800,
 	0xf40096b0,
 	0x0a98280b,
 	0x029abb9a,
 	0x0d0e1cf4,
-	0x01de7e01,
+	0x02617e01,
 	0xf494bd00,
-/* 0x00b2: intr_watchdog_next_time */
+/* 0x00c2: intr_watchdog_next_time */
 	0x0a98140e,
 	0x00a6b09b,
 	0xa6080bf4,
 	0x061cf49a,
-/* 0x00c0: intr_watchdog_next_time_set */
-/* 0x00c3: intr_watchdog_next_proc */
+/* 0x00d0: intr_watchdog_next_time_set */
+/* 0x00d3: intr_watchdog_next_proc */
 	0xb59b09b5,
 	0xe0b603e9,
 	0x68e6b158,
 	0xc81bf402,
-/* 0x00d2: intr */
+/* 0x00e2: intr */
 	0x00f900f8,
 	0x80f904bd,
 	0xa0f990f9,
@@ -865,13 +1001,13 @@ uint32_t nv108_pwr_code[] = {
 	0xc40088cf,
 	0x0bf40289,
 	0x9b00b51f,
-	0x957e580e,
+	0xa57e580e,
 	0x09980000,
 	0x0096b09b,
 	0x000d0bf4,
 	0x0009f634,
 	0x09b504bd,
-/* 0x0125: intr_skip_watchdog */
+/* 0x0135: intr_skip_watchdog */
 	0x0089e49a,
 	0x360bf408,
 	0xcf068849,
@@ -881,20 +1017,20 @@ uint32_t nv108_pwr_code[] = {
 	0xc0f900cc,
 	0xf14f484e,
 	0x0d5453e3,
-	0x023f7e00,
+	0x02c27e00,
 	0x40c0fc00,
 	0x0cf604c0,
-/* 0x0157: intr_subintr_skip_fifo */
+/* 0x0167: intr_subintr_skip_fifo */
 	0x4004bd00,
 	0x09f60688,
-/* 0x015f: intr_skip_subintr */
+/* 0x016f: intr_skip_subintr */
 	0xc404bd00,
 	0x0bf42089,
 	0xbfa4f107,
-/* 0x0169: intr_skip_pause */
+/* 0x0179: intr_skip_pause */
 	0x4089c4ff,
 	0xf1070bf4,
-/* 0x0173: intr_skip_user0 */
+/* 0x0183: intr_skip_user0 */
 	0x00ffbfa4,
 	0x0008f604,
 	0x80fc04bd,
@@ -904,549 +1040,684 @@ uint32_t nv108_pwr_code[] = {
 	0xfca0fcb0,
 	0xfc80fc90,
 	0x0032f400,
-/* 0x0196: timer */
-	0x32f401f8,
-	0x03f89810,
-	0xf40086b0,
-	0xfeb53a1c,
-	0xf6380003,
+/* 0x01a6: ticks_from_ns */
+	0xc0f901f8,
+	0xd7f1b0f9,
+	0xd3f00144,
+	0x7721f500,
+	0xe8ccec03,
+	0x00b4b003,
+	0xec120bf4,
+	0xf103e8ee,
+	0xf00144d7,
+	0x21f500d3,
+/* 0x01ce: ticks_from_ns_quit */
+	0xceb20377,
+	0xc0fcb0fc,
+/* 0x01d6: ticks_from_us */
+	0xc0f900f8,
+	0xd7f1b0f9,
+	0xd3f00144,
+	0x7721f500,
+	0xb0ceb203,
+	0x0bf400b4,
+/* 0x01ef: ticks_from_us_quit */
+	0xfce4bd05,
+	0xf8c0fcb0,
+/* 0x01f5: ticks_to_us */
+	0x44d7f100,
+	0x00d3f001,
+	0xf8ecedff,
+/* 0x0201: timer */
+	0xf990f900,
+	0x1032f480,
+	0xb003f898,
+	0x1cf40086,
+	0x0084bd4a,
+	0x0008f638,
+	0x340804bd,
+	0x980088cf,
+	0x98bb9a09,
+	0x00e9bb02,
+	0x0803feb5,
+	0x0088cf08,
+	0xf40284f0,
+	0x34081c1b,
+	0xa60088cf,
+	0x080bf4e0,
+	0x1cf4e8a6,
+/* 0x0245: timer_reset */
+	0xf634000d,
+	0x04bd000e,
+/* 0x024f: timer_enable */
+	0x089a0eb5,
+	0xf6380001,
 	0x04bd0008,
-	0x88cf0808,
-	0x0284f000,
-	0x081c1bf4,
-	0x0088cf34,
-	0x0bf4e0a6,
-	0xf4e8a608,
-/* 0x01c6: timer_reset */
-	0x3400161e,
-	0xbd000ef6,
-	0x9a0eb504,
-/* 0x01d0: timer_enable */
-	0x38000108,
-	0xbd0008f6,
-/* 0x01d9: timer_done */
-	0x1031f404,
-/* 0x01de: send_proc */
-	0x80f900f8,
-	0xe89890f9,
-	0x04e99805,
-	0xa60486f0,
-	0x2a0bf489,
-	0x940398c4,
-	0x80b60488,
-	0x008ebb18,
-	0xb500fa98,
-	0x8db5008a,
-	0x028cb501,
-	0xb6038bb5,
-	0x94f00190,
-	0x04e9b507,
-/* 0x0217: send_done */
-	0xfc0231f4,
-	0xf880fc90,
-/* 0x021d: find */
-	0x0880f900,
-	0x0131f458,
-/* 0x0224: find_loop */
-	0xa6008a98,
-	0x100bf4ae,
-	0xb15880b6,
-	0xf4026886,
-	0x32f4f11b,
-/* 0x0239: find_done */
-	0xfc8eb201,
-/* 0x023f: send */
-	0x7e00f880,
-	0xf400021d,
-	0x00f89b01,
-/* 0x0248: recv */
-	0x9805e898,
-	0x32f404e9,
-	0xf489a601,
-	0x89c43c0b,
-	0x0180b603,
-	0xb50784f0,
-	0xea9805e8,
-	0xfef0f902,
-	0xf0f9018f,
-	0x9994efb2,
-	0x00e9bb04,
-	0x9818e0b6,
-	0xec9803eb,
-	0x01ed9802,
-	0xf900ee98,
-	0xfef0fca5,
-	0x31f400f8,
-/* 0x028f: recv_done */
-	0xf8f0fc01,
-/* 0x0291: init */
-	0x01084100,
-	0xe70011cf,
-	0xb6010911,
-	0x14fe0814,
-	0x00e04100,
-	0x000013f0,
-	0x0001f61c,
-	0xff0104bd,
-	0x01f61400,
-	0x0104bd00,
-	0x0015f102,
-	0xf6100008,
-	0x04bd0001,
-	0xf000d241,
-	0x10fe0013,
-	0x1031f400,
-	0x38000101,
+/* 0x0258: timer_done */
+	0xfc1031f4,
+	0xf890fc80,
+/* 0x0261: send_proc */
+	0xf980f900,
+	0x05e89890,
+	0xf004e998,
+	0x89a60486,
+	0xc42a0bf4,
+	0x88940398,
+	0x1880b604,
+	0x98008ebb,
+	0x8ab500fa,
+	0x018db500,
+	0xb5028cb5,
+	0x90b6038b,
+	0x0794f001,
+	0xf404e9b5,
+/* 0x029a: send_done */
+	0x90fc0231,
+	0x00f880fc,
+/* 0x02a0: find */
+	0x580880f9,
+/* 0x02a7: find_loop */
+	0x980131f4,
+	0xaea6008a,
+	0xb6100bf4,
+	0x86b15880,
+	0x1bf40268,
+	0x0132f4f1,
+/* 0x02bc: find_done */
+	0x80fc8eb2,
+/* 0x02c2: send */
+	0xa07e00f8,
+	0x01f40002,
+/* 0x02cb: recv */
+	0xf900f89b,
+	0x9880f990,
+	0xe99805e8,
+	0x0132f404,
+	0x0bf489a6,
+	0x0389c43c,
+	0xf00180b6,
+	0xe8b50784,
+	0x02ea9805,
+	0x8ffef0f9,
+	0xb2f0f901,
+	0x049994ef,
+	0xb600e9bb,
+	0xeb9818e0,
+	0x02ec9803,
+	0x9801ed98,
+	0xa5f900ee,
+	0xf8fef0fc,
+	0x0131f400,
+/* 0x0316: recv_done */
+	0x80fcf0fc,
+	0x00f890fc,
+/* 0x031c: init */
+	0xcf010841,
+	0x11e70011,
+	0x14b60109,
+	0x0014fe08,
+	0xf000e041,
+	0x1c000013,
 	0xbd0001f6,
-/* 0x02db: init_proc */
-	0x98580f04,
-	0x16b001f1,
-	0xfa0bf400,
-	0xf0b615f9,
-	0xf20ef458,
-/* 0x02ec: host_send */
-	0xcf04b041,
-	0xa0420011,
-	0x0022cf04,
-	0x0bf412a6,
-	0x071ec42e,
-	0xb704ee94,
-	0x980270e0,
-	0xec9803eb,
-	0x01ed9802,
-	0x7e00ee98,
-	0xb600023f,
-	0x1ec40110,
-	0x04b0400f,
-	0xbd000ef6,
-	0xc70ef404,
-/* 0x0328: host_send_done */
-/* 0x032a: host_recv */
-	0x494100f8,
-	0x5413f14e,
-	0xf4e1a652,
-/* 0x0336: host_recv_wait */
-	0xcc41b90b,
-	0x0011cf04,
-	0xcf04c842,
-	0x16f00022,
-	0xf412a608,
-	0x23c4ef0b,
-	0x0434b607,
-	0x02f030b7,
-	0xb5033bb5,
-	0x3db5023c,
-	0x003eb501,
-	0xf00120b6,
-	0xc8400f24,
-	0x0002f604,
-	0x400204bd,
-	0x02f60000,
-	0xf804bd00,
-/* 0x0379: host_init */
-	0x00804100,
-	0xf11014b6,
-	0x40027015,
-	0x01f604d0,
+	0x00ff0104,
+	0x0001f614,
+	0x020104bd,
+	0x080015f1,
+	0x01f61000,
 	0x4104bd00,
+	0x13f000e2,
+	0x0010fe00,
+	0x011031f4,
+	0xf6380001,
+	0x04bd0001,
+/* 0x0366: init_proc */
+	0xf198580f,
+	0x0016b001,
+	0xf9fa0bf4,
+	0x58f0b615,
+/* 0x0377: mulu32_32_64 */
+	0xf9f20ef4,
+	0xf920f910,
+	0x9540f930,
+	0xd29510e1,
+	0xbdc4bd10,
+	0xc0edffb4,
+	0xb2301dff,
+	0xff34f134,
+	0x1034b6ff,
+	0xbb1045b6,
+	0xb4bb00c3,
+	0x30e2ff01,
+	0x34f134b2,
+	0x34b6ffff,
+	0x1045b610,
+	0xbb00c3bb,
+	0x12ff01b4,
+	0x00b3bb30,
+	0x30fc40fc,
+	0x10fc20fc,
+/* 0x03c6: host_send */
+	0xb04100f8,
+	0x0011cf04,
+	0xcf04a042,
+	0x12a60022,
+	0xc42e0bf4,
+	0xee94071e,
+	0x70e0b704,
+	0x03eb9802,
+	0x9802ec98,
+	0xee9801ed,
+	0x02c27e00,
+	0x0110b600,
+	0x400f1ec4,
+	0x0ef604b0,
+	0xf404bd00,
+/* 0x0402: host_send_done */
+	0x00f8c70e,
+/* 0x0404: host_recv */
+	0xf14e4941,
+	0xa6525413,
+	0xb90bf4e1,
+/* 0x0410: host_recv_wait */
+	0xcf04cc41,
+	0xc8420011,
+	0x0022cf04,
+	0xa60816f0,
+	0xef0bf412,
+	0xb60723c4,
+	0x30b70434,
+	0x3bb502f0,
+	0x023cb503,
+	0xb5013db5,
+	0x20b6003e,
+	0x0f24f001,
+	0xf604c840,
+	0x04bd0002,
+	0x00004002,
+	0xbd0002f6,
+/* 0x0453: host_init */
+	0x4100f804,
 	0x14b60080,
-	0xf015f110,
-	0x04dc4002,
+	0x7015f110,
+	0x04d04002,
 	0xbd0001f6,
-	0x40010104,
-	0x01f604c4,
-	0xf804bd00,
-/* 0x03a9: memx_func_enter */
-	0x40040600,
-	0x06f607e0,
-/* 0x03b3: memx_func_enter_wait */
-	0x4604bd00,
-	0x66cf07c0,
-	0x0464f000,
-	0x98f70bf4,
-	0x10b60016,
-/* 0x03c7: memx_func_leave */
-	0x0600f804,
-	0x07e44004,
-	0xbd0006f6,
-/* 0x03d1: memx_func_leave_wait */
-	0x07c04604,
-	0xf00066cf,
-	0x1bf40464,
-/* 0x03df: memx_func_wr32 */
-	0x9800f8f7,
-	0x15980016,
-	0x0810b601,
-	0x50f960f9,
+	0x00804104,
+	0xf11014b6,
+	0x4002f015,
+	0x01f604dc,
+	0x0104bd00,
+	0x04c44001,
+	0xbd0001f6,
+/* 0x0483: memx_func_enter */
+	0xf100f804,
+	0xf1162067,
+	0xf1f55d77,
+	0xb2ffff73,
+	0x00047e6e,
+	0xfdd8b200,
+	0x60f90487,
+	0xd0fc80f9,
+	0x2e7ee0fc,
+	0x77f10000,
+	0x73f1fffe,
+	0x6eb2ffff,
+	0x0000047e,
+	0x87fdd8b2,
+	0xf960f904,
+	0xfcd0fc80,
+	0x002e7ee0,
+	0xf067f100,
+	0x7e6eb226,
+	0xb2000004,
+	0x0487fdd8,
+	0x80f960f9,
 	0xe0fcd0fc,
 	0x00002e7e,
-	0xf40242b6,
-	0x00f8e81b,
-/* 0x03fc: memx_func_wait */
-	0x88cf2c08,
-	0x001e9800,
-	0x98011d98,
-	0x1b98021c,
-	0x1010b603,
-	0x0000717e,
-/* 0x0416: memx_func_delay */
-	0x1e9800f8,
-	0x0410b600,
-	0x00005d7e,
-/* 0x0422: memx_exec */
-	0xe0f900f8,
-	0xc1b2d0f9,
-/* 0x042a: memx_exec_next */
-	0x1398b2b2,
-	0x0410b600,
-	0xf0103495,
-	0x35980c30,
-	0xa655f9de,
-	0xed1ef412,
+	0xe0400406,
+	0x0006f607,
+/* 0x04ea: memx_func_enter_wait */
+	0xc04604bd,
+	0x0066cf07,
+	0xf40464f0,
+	0x2c06f70b,
+	0xb50066cf,
+	0x00f8f106,
+/* 0x0500: memx_func_leave */
+	0x66cf2c06,
+	0xf206b500,
+	0xe4400406,
+	0x0006f607,
+/* 0x0512: memx_func_leave_wait */
+	0xc04604bd,
+	0x0066cf07,
+	0xf40464f0,
+	0x67f1f71b,
+	0x77f126f0,
+	0x73f00001,
+	0x7e6eb200,
+	0xb2000004,
+	0x0587fdd8,
+	0x80f960f9,
 	0xe0fcd0fc,
-	0x00023f7e,
-/* 0x044a: memx_info */
-	0xac4c00f8,
+	0x00002e7e,
+	0x162067f1,
+	0x047e6eb2,
+	0xd8b20000,
+	0xf90587fd,
+	0xfc80f960,
+	0x7ee0fcd0,
+	0xf100002e,
+	0xf00aa277,
+	0x6eb20073,
+	0x0000047e,
+	0x87fdd8b2,
+	0xf960f905,
+	0xfcd0fc80,
+	0x002e7ee0,
+/* 0x057b: memx_func_wait_vblank */
+	0xb600f800,
+	0x00f80410,
+/* 0x0580: memx_func_wr32 */
+	0x98001698,
+	0x10b60115,
+	0xf960f908,
+	0xfcd0fc50,
+	0x002e7ee0,
+	0x0242b600,
+	0xf8e81bf4,
+/* 0x059d: memx_func_wait */
+	0xcf2c0800,
+	0x1e980088,
+	0x011d9800,
+	0x98021c98,
+	0x10b6031b,
+	0x00797e10,
+/* 0x05b7: memx_func_delay */
+	0x9800f800,
+	0x10b6001e,
+	0x005d7e04,
+/* 0x05c3: memx_func_train */
+	0xf800f800,
+/* 0x05c5: memx_exec */
+	0xf9e0f900,
+	0xb2c1b2d0,
+/* 0x05cd: memx_exec_next */
+	0x001398b2,
+	0xe70410b6,
+	0xe701f034,
+	0xb601e033,
+	0x30f00132,
+	0xde35980c,
+	0x12a655f9,
+	0x98e51ef4,
+	0x0c98f10b,
+	0x02cbbbf2,
+	0xcf07c44b,
+	0xd0fc00bb,
+	0xc27ee0fc,
+	0x00f80002,
+/* 0x0604: memx_info */
+	0xf401c670,
+/* 0x060a: memx_info_data */
+	0xcc4c0c0b,
 	0x08004b03,
-	0x00023f7e,
-/* 0x0456: memx_recv */
-	0xd6b000f8,
-	0xc90bf401,
-	0xf400d6b0,
-	0x00f8eb0b,
-/* 0x0464: memx_init */
-/* 0x0466: perf_recv */
-	0x00f800f8,
-/* 0x0468: perf_init */
-/* 0x046a: i2c_drive_scl */
-	0x36b000f8,
-	0x0d0bf400,
-	0xf607e040,
-	0x04bd0001,
-/* 0x047a: i2c_drive_scl_lo */
-	0xe44000f8,
-	0x0001f607,
-	0x00f804bd,
-/* 0x0484: i2c_drive_sda */
-	0xf40036b0,
-	0xe0400d0b,
-	0x0002f607,
-	0x00f804bd,
-/* 0x0494: i2c_drive_sda_lo */
-	0xf607e440,
-	0x04bd0002,
-/* 0x049e: i2c_sense_scl */
-	0x32f400f8,
-	0x07c44301,
-	0xfd0033cf,
-	0x0bf40431,
-	0x0131f406,
-/* 0x04b0: i2c_sense_scl_done */
-/* 0x04b2: i2c_sense_sda */
-	0x32f400f8,
-	0x07c44301,
-	0xfd0033cf,
-	0x0bf40432,
-	0x0131f406,
-/* 0x04c4: i2c_sense_sda_done */
-/* 0x04c6: i2c_raise_scl */
-	0x40f900f8,
-	0x03089844,
-	0x046a7e01,
-/* 0x04d1: i2c_raise_scl_wait */
-	0x03e84e00,
-	0x00005d7e,
-	0x00049e7e,
-	0xb60901f4,
-	0x1bf40142,
-/* 0x04e5: i2c_raise_scl_done */
-	0xf840fcef,
-/* 0x04e9: i2c_start */
-	0x049e7e00,
-	0x0d11f400,
-	0x0004b27e,
-	0xf40611f4,
-/* 0x04fa: i2c_start_rep */
-	0x00032e0e,
-	0x00046a7e,
-	0x847e0103,
-	0x76bb0004,
-	0x0465b600,
-	0x659450f9,
-	0x0256bb04,
-	0x75fd50bd,
-	0x7e50fc04,
-	0xb60004c6,
-	0x11f40464,
-/* 0x0525: i2c_start_send */
-	0x7e00031d,
-	0x4e000484,
-	0x5d7e1388,
-	0x00030000,
-	0x00046a7e,
-	0x7e13884e,
-/* 0x053f: i2c_start_out */
-	0xf800005d,
-/* 0x0541: i2c_stop */
-	0x7e000300,
-	0x0300046a,
-	0x04847e00,
-	0x03e84e00,
-	0x00005d7e,
-	0x6a7e0103,
-	0x884e0004,
-	0x005d7e13,
+/* 0x0613: memx_info_train */
+	0x4c090ef4,
+	0x004b0bcc,
+/* 0x0619: memx_info_send */
+	0x02c27e01,
+/* 0x061f: memx_recv */
+	0xb000f800,
+	0x0bf401d6,
+	0x00d6b0a3,
+	0xf8dc0bf4,
+/* 0x062d: memx_init */
+/* 0x062f: perf_recv */
+	0xf800f800,
+/* 0x0631: perf_init */
+/* 0x0633: i2c_drive_scl */
+	0xb000f800,
+	0x0bf40036,
+	0x07e0400d,
+	0xbd0001f6,
+/* 0x0643: i2c_drive_scl_lo */
+	0x4000f804,
+	0x01f607e4,
+	0xf804bd00,
+/* 0x064d: i2c_drive_sda */
+	0x0036b000,
+	0x400d0bf4,
+	0x02f607e0,
+	0xf804bd00,
+/* 0x065d: i2c_drive_sda_lo */
+	0x07e44000,
+	0xbd0002f6,
+/* 0x0667: i2c_sense_scl */
+	0xf400f804,
+	0xc4430132,
+	0x0033cf07,
+	0xf40431fd,
+	0x31f4060b,
+/* 0x0679: i2c_sense_scl_done */
+/* 0x067b: i2c_sense_sda */
+	0xf400f801,
+	0xc4430132,
+	0x0033cf07,
+	0xf40432fd,
+	0x31f4060b,
+/* 0x068d: i2c_sense_sda_done */
+/* 0x068f: i2c_raise_scl */
+	0xf900f801,
+	0x08984440,
+	0x337e0103,
+/* 0x069a: i2c_raise_scl_wait */
+	0xe84e0006,
+	0x005d7e03,
+	0x06677e00,
+	0x0901f400,
+	0xf40142b6,
+/* 0x06ae: i2c_raise_scl_done */
+	0x40fcef1b,
+/* 0x06b2: i2c_start */
+	0x677e00f8,
+	0x11f40006,
+	0x067b7e0d,
+	0x0611f400,
+/* 0x06c3: i2c_start_rep */
+	0x032e0ef4,
+	0x06337e00,
 	0x7e010300,
-	0x4e000484,
-	0x5d7e1388,
-	0x00f80000,
-/* 0x0570: i2c_bitw */
-	0x0004847e,
-	0x7e03e84e,
-	0xbb00005d,
+	0xbb00064d,
 	0x65b60076,
 	0x9450f904,
 	0x56bb0465,
 	0xfd50bd02,
 	0x50fc0475,
-	0x0004c67e,
+	0x00068f7e,
 	0xf40464b6,
-	0x884e1711,
-	0x005d7e13,
-	0x7e000300,
-	0x4e00046a,
-	0x5d7e1388,
-/* 0x05ae: i2c_bitw_out */
-	0x00f80000,
-/* 0x05b0: i2c_bitr */
-	0x847e0103,
-	0xe84e0004,
+/* 0x06ee: i2c_start_send */
+	0x00031d11,
+	0x00064d7e,
+	0x7e13884e,
+	0x0300005d,
+	0x06337e00,
+	0x13884e00,
+	0x00005d7e,
+/* 0x0708: i2c_start_out */
+/* 0x070a: i2c_stop */
+	0x000300f8,
+	0x0006337e,
+	0x4d7e0003,
+	0xe84e0006,
 	0x005d7e03,
-	0x0076bb00,
-	0xf90465b6,
-	0x04659450,
-	0xbd0256bb,
-	0x0475fd50,
-	0xc67e50fc,
-	0x64b60004,
-	0x1a11f404,
-	0x0004b27e,
-	0x6a7e0003,
-	0x884e0004,
-	0x005d7e13,
-	0x013cf000,
-/* 0x05f3: i2c_bitr_done */
-	0xf80131f4,
-/* 0x05f5: i2c_get_byte */
-	0x04000500,
-/* 0x05f9: i2c_get_byte_next */
-	0x0154b608,
+	0x7e010300,
+	0x4e000633,
+	0x5d7e1388,
+	0x01030000,
+	0x00064d7e,
+	0x7e13884e,
+	0xf800005d,
+/* 0x0739: i2c_bitw */
+	0x064d7e00,
+	0x03e84e00,
+	0x00005d7e,
 	0xb60076bb,
 	0x50f90465,
 	0xbb046594,
 	0x50bd0256,
 	0xfc0475fd,
-	0x05b07e50,
+	0x068f7e50,
 	0x0464b600,
-	0xfd2a11f4,
-	0x42b60553,
-	0xd81bf401,
-	0x76bb0103,
+	0x4e1711f4,
+	0x5d7e1388,
+	0x00030000,
+	0x0006337e,
+	0x7e13884e,
+/* 0x0777: i2c_bitw_out */
+	0xf800005d,
+/* 0x0779: i2c_bitr */
+	0x7e010300,
+	0x4e00064d,
+	0x5d7e03e8,
+	0x76bb0000,
 	0x0465b600,
 	0x659450f9,
 	0x0256bb04,
 	0x75fd50bd,
 	0x7e50fc04,
-	0xb6000570,
-/* 0x0642: i2c_get_byte_done */
-	0x00f80464,
-/* 0x0644: i2c_put_byte */
-/* 0x0646: i2c_put_byte_next */
-	0x42b60804,
-	0x3854ff01,
-	0xb60076bb,
-	0x50f90465,
-	0xbb046594,
-	0x50bd0256,
-	0xfc0475fd,
-	0x05707e50,
-	0x0464b600,
-	0xb03411f4,
-	0x1bf40046,
-	0x0076bbd8,
+	0xb600068f,
+	0x11f40464,
+	0x067b7e1a,
+	0x7e000300,
+	0x4e000633,
+	0x5d7e1388,
+	0x3cf00000,
+	0x0131f401,
+/* 0x07bc: i2c_bitr_done */
+/* 0x07be: i2c_get_byte */
+	0x000500f8,
+/* 0x07c2: i2c_get_byte_next */
+	0x54b60804,
+	0x0076bb01,
+	0xf90465b6,
+	0x04659450,
+	0xbd0256bb,
+	0x0475fd50,
+	0x797e50fc,
+	0x64b60007,
+	0x2a11f404,
+	0xb60553fd,
+	0x1bf40142,
+	0xbb0103d8,
+	0x65b60076,
+	0x9450f904,
+	0x56bb0465,
+	0xfd50bd02,
+	0x50fc0475,
+	0x0007397e,
+/* 0x080b: i2c_get_byte_done */
+	0xf80464b6,
+/* 0x080d: i2c_put_byte */
+/* 0x080f: i2c_put_byte_next */
+	0xb6080400,
+	0x54ff0142,
+	0x0076bb38,
 	0xf90465b6,
 	0x04659450,
 	0xbd0256bb,
 	0x0475fd50,
-	0xb07e50fc,
-	0x64b60005,
-	0x0f11f404,
-	0xb00076bb,
-	0x1bf40136,
-	0x0132f406,
-/* 0x069c: i2c_put_byte_done */
-/* 0x069e: i2c_addr */
-	0x76bb00f8,
+	0x397e50fc,
+	0x64b60007,
+	0x3411f404,
+	0xf40046b0,
+	0x76bbd81b,
 	0x0465b600,
 	0x659450f9,
 	0x0256bb04,
 	0x75fd50bd,
 	0x7e50fc04,
-	0xb60004e9,
+	0xb6000779,
 	0x11f40464,
-	0x2ec3e729,
-	0x0134b601,
-	0xbb0553fd,
+	0x0076bb0f,
+	0xf40136b0,
+	0x32f4061b,
+/* 0x0865: i2c_put_byte_done */
+/* 0x0867: i2c_addr */
+	0xbb00f801,
 	0x65b60076,
 	0x9450f904,
 	0x56bb0465,
 	0xfd50bd02,
 	0x50fc0475,
-	0x0006447e,
-/* 0x06e3: i2c_addr_done */
-	0xf80464b6,
-/* 0x06e5: i2c_acquire_addr */
-	0xf8cec700,
-	0xb705e4b6,
-	0xf8d014e0,
-/* 0x06f1: i2c_acquire */
-	0x06e57e00,
-	0x00047e00,
-	0x03d9f000,
-	0x00002e7e,
-/* 0x0702: i2c_release */
-	0xe57e00f8,
-	0x047e0006,
-	0xdaf00000,
+	0x0006b27e,
+	0xf40464b6,
+	0xc3e72911,
+	0x34b6012e,
+	0x0553fd01,
+	0xb60076bb,
+	0x50f90465,
+	0xbb046594,
+	0x50bd0256,
+	0xfc0475fd,
+	0x080d7e50,
+	0x0464b600,
+/* 0x08ac: i2c_addr_done */
+/* 0x08ae: i2c_acquire_addr */
+	0xcec700f8,
+	0x05e4b6f8,
+	0xd014e0b7,
+/* 0x08ba: i2c_acquire */
+	0xae7e00f8,
+	0x047e0008,
+	0xd9f00000,
 	0x002e7e03,
-/* 0x0713: i2c_recv */
-	0xf400f800,
-	0xc1c70132,
-	0x0214b6f8,
-	0xf52816b0,
-	0xb801371f,
-	0x000bd413,
-	0xb8003298,
-	0x000bac13,
-	0xf4003198,
-	0xd0f90231,
-	0xd0f9e0f9,
-	0x000067f1,
-	0x100063f1,
-	0xbb016792,
+/* 0x08cb: i2c_release */
+	0x7e00f800,
+	0x7e0008ae,
+	0xf0000004,
+	0x2e7e03da,
+	0x00f80000,
+/* 0x08dc: i2c_recv */
+	0xc70132f4,
+	0x14b6f8c1,
+	0x2816b002,
+	0x01371ff5,
+	0x0cf413b8,
+	0x00329800,
+	0x0ccc13b8,
+	0x00319800,
+	0xf90231f4,
+	0xf9e0f9d0,
+	0x0067f1d0,
+	0x0063f100,
+	0x01679210,
+	0xb60076bb,
+	0x50f90465,
+	0xbb046594,
+	0x50bd0256,
+	0xfc0475fd,
+	0x08ba7e50,
+	0x0464b600,
+	0xd6b0d0fc,
+	0xb01bf500,
+	0xbb000500,
 	0x65b60076,
 	0x9450f904,
 	0x56bb0465,
 	0xfd50bd02,
 	0x50fc0475,
-	0x0006f17e,
-	0xfc0464b6,
-	0x00d6b0d0,
-	0x00b01bf5,
-	0x76bb0005,
+	0x0008677e,
+	0xf50464b6,
+	0xc700cc11,
+	0x76bbe0c5,
 	0x0465b600,
 	0x659450f9,
 	0x0256bb04,
 	0x75fd50bd,
 	0x7e50fc04,
-	0xb600069e,
+	0xb600080d,
 	0x11f50464,
-	0xc5c700cc,
-	0x0076bbe0,
-	0xf90465b6,
-	0x04659450,
-	0xbd0256bb,
-	0x0475fd50,
-	0x447e50fc,
-	0x64b60006,
-	0xa911f504,
-	0xbb010500,
-	0x65b60076,
-	0x9450f904,
-	0x56bb0465,
-	0xfd50bd02,
-	0x50fc0475,
-	0x00069e7e,
-	0xf50464b6,
-	0xbb008711,
-	0x65b60076,
-	0x9450f904,
-	0x56bb0465,
-	0xfd50bd02,
-	0x50fc0475,
-	0x0005f57e,
-	0xf40464b6,
-	0x5bcb6711,
-	0x0076bbe0,
-	0xf90465b6,
-	0x04659450,
-	0xbd0256bb,
-	0x0475fd50,
-	0x417e50fc,
-	0x64b60005,
-	0xbd5bb204,
-	0x410ef474,
-/* 0x0818: i2c_recv_not_rd08 */
-	0xf401d6b0,
-	0x00053b1b,
-	0x00069e7e,
-	0xc73211f4,
-	0x447ee0c5,
-	0x11f40006,
-	0x7e000528,
-	0xf400069e,
-	0xb5c71f11,
-	0x06447ee0,
-	0x1511f400,
-	0x0005417e,
-	0xc5c774bd,
-	0x091bf408,
-	0xf40232f4,
-/* 0x0856: i2c_recv_not_wr08 */
-/* 0x0856: i2c_recv_done */
-	0xcec7030e,
-	0x07027ef8,
-	0xfce0fc00,
-	0x0912f4d0,
-	0x3f7e7cb2,
-/* 0x086a: i2c_recv_exit */
-	0x00f80002,
-/* 0x086c: i2c_init */
-/* 0x086e: test_recv */
-	0x584100f8,
-	0x0011cf04,
-	0x400110b6,
-	0x01f60458,
-	0xf104bd00,
-	0xf1d900e7,
-	0x7e134fe3,
-	0xf8000196,
-/* 0x088d: test_init */
-	0x08004e00,
-	0x0001967e,
-/* 0x0896: idle_recv */
-	0x00f800f8,
-/* 0x0898: idle */
-	0x410031f4,
-	0x11cf0454,
+	0x010500a9,
+	0xb60076bb,
+	0x50f90465,
+	0xbb046594,
+	0x50bd0256,
+	0xfc0475fd,
+	0x08677e50,
+	0x0464b600,
+	0x008711f5,
+	0xb60076bb,
+	0x50f90465,
+	0xbb046594,
+	0x50bd0256,
+	0xfc0475fd,
+	0x07be7e50,
+	0x0464b600,
+	0xcb6711f4,
+	0x76bbe05b,
+	0x0465b600,
+	0x659450f9,
+	0x0256bb04,
+	0x75fd50bd,
+	0x7e50fc04,
+	0xb600070a,
+	0x5bb20464,
+	0x0ef474bd,
+/* 0x09e1: i2c_recv_not_rd08 */
+	0x01d6b041,
+	0x053b1bf4,
+	0x08677e00,
+	0x3211f400,
+	0x7ee0c5c7,
+	0xf400080d,
+	0x00052811,
+	0x0008677e,
+	0xc71f11f4,
+	0x0d7ee0b5,
+	0x11f40008,
+	0x070a7e15,
+	0xc774bd00,
+	0x1bf408c5,
+	0x0232f409,
+/* 0x0a1f: i2c_recv_not_wr08 */
+/* 0x0a1f: i2c_recv_done */
+	0xc7030ef4,
+	0xcb7ef8ce,
+	0xe0fc0008,
+	0x12f4d0fc,
+	0x7e7cb209,
+/* 0x0a33: i2c_recv_exit */
+	0xf80002c2,
+/* 0x0a35: i2c_init */
+/* 0x0a37: test_recv */
+	0x4100f800,
+	0x11cf0458,
 	0x0110b600,
-	0xf6045440,
+	0xf6045840,
 	0x04bd0001,
-/* 0x08ac: idle_loop */
-	0x32f45801,
-/* 0x08b1: idle_proc */
-/* 0x08b1: idle_proc_exec */
-	0xb210f902,
-	0x02487e1e,
-	0xf410fc00,
-	0x31f40911,
-	0xf00ef402,
-/* 0x08c4: idle_proc_next */
-	0xa65810b6,
-	0xe81bf41f,
-	0xf4e002f4,
-	0x0ef40028,
-	0x000000c6,
+	0xd900e7f1,
+	0x134fe3f1,
+	0x0002017e,
+/* 0x0a56: test_init */
+	0x004e00f8,
+	0x02017e08,
+/* 0x0a5f: idle_recv */
+	0xf800f800,
+/* 0x0a61: idle */
+	0x0031f400,
+	0xcf045441,
+	0x10b60011,
+	0x04544001,
+	0xbd0001f6,
+/* 0x0a75: idle_loop */
+	0xf4580104,
+/* 0x0a7a: idle_proc */
+/* 0x0a7a: idle_proc_exec */
+	0x10f90232,
+	0xcb7e1eb2,
+	0x10fc0002,
+	0xf40911f4,
+	0x0ef40231,
+/* 0x0a8d: idle_proc_next */
+	0x5810b6f0,
+	0x1bf41fa6,
+	0xe002f4e8,
+	0xf40028f4,
+	0x0000c60e,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
 	0x00000000,
 	0x00000000,
 	0x00000000,
diff --git a/drivers/gpu/drm/nouveau/core/subdev/pwr/fuc/nva3.fuc b/drivers/gpu/drm/nouveau/core/subdev/pwr/fuc/nva3.fuc
index 6744fcc06151..daa06c1c655e 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/pwr/fuc/nva3.fuc
+++ b/drivers/gpu/drm/nouveau/core/subdev/pwr/fuc/nva3.fuc
@@ -23,6 +23,7 @@
  */
 
 #define NVKM_PPWR_CHIPSET GT215
+#define HW_TICKS_PER_US 203 // should be 202.5
 
 //#define NVKM_FALCON_PC24
 //#define NVKM_FALCON_UNSHIFTED_IO
@@ -34,6 +35,7 @@
 .section #nva3_pwr_data
 #define INCLUDE_PROC
 #include "kernel.fuc"
+#include "arith.fuc"
 #include "host.fuc"
 #include "memx.fuc"
 #include "perf.fuc"
@@ -44,6 +46,7 @@
 
 #define INCLUDE_DATA
 #include "kernel.fuc"
+#include "arith.fuc"
 #include "host.fuc"
 #include "memx.fuc"
 #include "perf.fuc"
@@ -56,6 +59,7 @@
 .section #nva3_pwr_code
 #define INCLUDE_CODE
 #include "kernel.fuc"
+#include "arith.fuc"
 #include "host.fuc"
 #include "memx.fuc"
 #include "perf.fuc"
diff --git a/drivers/gpu/drm/nouveau/core/subdev/pwr/fuc/nva3.fuc.h b/drivers/gpu/drm/nouveau/core/subdev/pwr/fuc/nva3.fuc.h
index e087ce3041be..d1f9b6cb66d7 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/pwr/fuc/nva3.fuc.h
+++ b/drivers/gpu/drm/nouveau/core/subdev/pwr/fuc/nva3.fuc.h
@@ -24,8 +24,8 @@ uint32_t nva3_pwr_data[] = {
 	0x00000000,
 /* 0x0058: proc_list_head */
 	0x54534f48,
-	0x00000430,
-	0x000003cd,
+	0x00000512,
+	0x000004af,
 	0x00000000,
 	0x00000000,
 	0x00000000,
@@ -46,8 +46,8 @@ uint32_t nva3_pwr_data[] = {
 	0x00000000,
 	0x00000000,
 	0x584d454d,
-	0x00000542,
-	0x00000534,
+	0x00000842,
+	0x00000834,
 	0x00000000,
 	0x00000000,
 	0x00000000,
@@ -68,8 +68,8 @@ uint32_t nva3_pwr_data[] = {
 	0x00000000,
 	0x00000000,
 	0x46524550,
-	0x00000546,
-	0x00000544,
+	0x00000846,
+	0x00000844,
 	0x00000000,
 	0x00000000,
 	0x00000000,
@@ -90,8 +90,8 @@ uint32_t nva3_pwr_data[] = {
 	0x00000000,
 	0x00000000,
 	0x5f433249,
-	0x00000976,
-	0x00000819,
+	0x00000c76,
+	0x00000b19,
 	0x00000000,
 	0x00000000,
 	0x00000000,
@@ -112,8 +112,8 @@ uint32_t nva3_pwr_data[] = {
 	0x00000000,
 	0x00000000,
 	0x54534554,
-	0x0000099f,
-	0x00000978,
+	0x00000c9f,
+	0x00000c78,
 	0x00000000,
 	0x00000000,
 	0x00000000,
@@ -134,8 +134,8 @@ uint32_t nva3_pwr_data[] = {
 	0x00000000,
 	0x00000000,
 	0x454c4449,
-	0x000009ab,
-	0x000009a9,
+	0x00000cab,
+	0x00000ca9,
 	0x00000000,
 	0x00000000,
 	0x00000000,
@@ -227,24 +227,88 @@ uint32_t nva3_pwr_data[] = {
 	0x00000000,
 	0x00000000,
 /* 0x0370: memx_func_head */
-	0x00010000,
-	0x00000000,
-	0x0000046f,
-/* 0x037c: memx_func_next */
 	0x00000001,
 	0x00000000,
-	0x00000496,
+	0x00000551,
+/* 0x037c: memx_func_next */
 	0x00000002,
+	0x00000000,
+	0x000005a8,
+	0x00000003,
 	0x00000002,
-	0x000004b7,
-	0x00040003,
+	0x0000063a,
+	0x00040004,
+	0x00000000,
+	0x00000656,
+	0x00010005,
+	0x00000000,
+	0x00000673,
+	0x00010006,
+	0x00000000,
+	0x000005f8,
+	0x00000007,
+	0x00000000,
+	0x0000067e,
+/* 0x03c4: memx_func_tail */
+/* 0x03c4: memx_ts_start */
+	0x00000000,
+/* 0x03c8: memx_ts_end */
+	0x00000000,
+/* 0x03cc: memx_data_head */
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
 	0x00000000,
-	0x000004d3,
-	0x00010004,
 	0x00000000,
-	0x000004f0,
-/* 0x03ac: memx_func_tail */
-/* 0x03ac: memx_data_head */
 	0x00000000,
 	0x00000000,
 	0x00000000,
@@ -703,6 +767,8 @@ uint32_t nva3_pwr_data[] = {
 	0x00000000,
 	0x00000000,
 	0x00000000,
+/* 0x0bcc: memx_data_tail */
+/* 0x0bcc: memx_train_head */
 	0x00000000,
 	0x00000000,
 	0x00000000,
@@ -757,8 +823,18 @@ uint32_t nva3_pwr_data[] = {
 	0x00000000,
 	0x00000000,
 	0x00000000,
-/* 0x0bac: memx_data_tail */
-/* 0x0bac: i2c_scl_map */
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+/* 0x0ccc: memx_train_tail */
+/* 0x0ccc: i2c_scl_map */
 	0x00001000,
 	0x00004000,
 	0x00010000,
@@ -769,7 +845,7 @@ uint32_t nva3_pwr_data[] = {
 	0x01000000,
 	0x04000000,
 	0x10000000,
-/* 0x0bd4: i2c_sda_map */
+/* 0x0cf4: i2c_sda_map */
 	0x00002000,
 	0x00008000,
 	0x00020000,
@@ -780,7 +856,7 @@ uint32_t nva3_pwr_data[] = {
 	0x02000000,
 	0x08000000,
 	0x20000000,
-/* 0x0bfc: i2c_ctrl */
+/* 0x0d1c: i2c_ctrl */
 	0x0000e138,
 	0x0000e150,
 	0x0000e168,
@@ -838,18 +914,10 @@ uint32_t nva3_pwr_data[] = {
 	0x00000000,
 	0x00000000,
 	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
 };
 
 uint32_t nva3_pwr_code[] = {
-	0x030d0ef5,
+	0x039e0ef5,
 /* 0x0004: rd32 */
 	0x07a007f1,
 	0xd00604b6,
@@ -885,19 +953,22 @@ uint32_t nva3_pwr_code[] = {
 	0xd4f100dd,
 	0x1bf47000,
 /* 0x007f: nsec */
-	0xf000f8f2,
+	0xf900f8f2,
+	0xf080f990,
 	0x84b62c87,
 	0x0088cf06,
-/* 0x0088: nsec_loop */
+/* 0x008c: nsec_loop */
 	0xb62c97f0,
 	0x99cf0694,
 	0x0298bb00,
 	0xf4069eb8,
-	0x00f8f11e,
-/* 0x009c: wait */
+	0x80fcf11e,
+	0x00f890fc,
+/* 0x00a4: wait */
+	0x80f990f9,
 	0xb62c87f0,
 	0x88cf0684,
-/* 0x00a5: wait_loop */
+/* 0x00b1: wait_loop */
 	0x02eeb900,
 	0xb90421f4,
 	0xadfd02da,
@@ -907,28 +978,29 @@ uint32_t nva3_pwr_code[] = {
 	0x0099cf06,
 	0xb80298bb,
 	0x1ef4069b,
-/* 0x00c9: wait_done */
-/* 0x00cb: intr_watchdog */
-	0x9800f8df,
+/* 0x00d5: wait_done */
+	0xfc80fcdf,
+/* 0x00db: intr_watchdog */
+	0x9800f890,
 	0x96b003e9,
 	0x2a0bf400,
 	0xbb9a0a98,
 	0x1cf4029a,
 	0x01d7f00f,
-	0x025421f5,
+	0x02dd21f5,
 	0x0ef494bd,
-/* 0x00e9: intr_watchdog_next_time */
+/* 0x00f9: intr_watchdog_next_time */
 	0x9b0a9815,
 	0xf400a6b0,
 	0x9ab8090b,
 	0x061cf406,
-/* 0x00f8: intr_watchdog_next_time_set */
-/* 0x00fb: intr_watchdog_next_proc */
+/* 0x0108: intr_watchdog_next_time_set */
+/* 0x010b: intr_watchdog_next_proc */
 	0x809b0980,
 	0xe0b603e9,
 	0x68e6b158,
 	0xc61bf402,
-/* 0x010a: intr */
+/* 0x011a: intr */
 	0x00f900f8,
 	0x80f904bd,
 	0xa0f990f9,
@@ -948,13 +1020,13 @@ uint32_t nva3_pwr_code[] = {
 	0xf40289c4,
 	0x0080230b,
 	0x58e7f09b,
-	0x98cb21f4,
+	0x98db21f4,
 	0x96b09b09,
 	0x110bf400,
 	0xb63407f0,
 	0x09d00604,
 	0x8004bd00,
-/* 0x016e: intr_skip_watchdog */
+/* 0x017e: intr_skip_watchdog */
 	0x89e49a09,
 	0x0bf40800,
 	0x8897f148,
@@ -967,22 +1039,22 @@ uint32_t nva3_pwr_code[] = {
 	0x48e7f1c0,
 	0x53e3f14f,
 	0x00d7f054,
-	0x02b921f5,
+	0x034221f5,
 	0x07f1c0fc,
 	0x04b604c0,
 	0x000cd006,
-/* 0x01ae: intr_subintr_skip_fifo */
+/* 0x01be: intr_subintr_skip_fifo */
 	0x07f104bd,
 	0x04b60688,
 	0x0009d006,
-/* 0x01ba: intr_skip_subintr */
+/* 0x01ca: intr_skip_subintr */
 	0x89c404bd,
 	0x070bf420,
 	0xffbfa4f1,
-/* 0x01c4: intr_skip_pause */
+/* 0x01d4: intr_skip_pause */
 	0xf44089c4,
 	0xa4f1070b,
-/* 0x01ce: intr_skip_user0 */
+/* 0x01de: intr_skip_user0 */
 	0x07f0ffbf,
 	0x0604b604,
 	0xbd0008d0,
@@ -993,344 +1065,551 @@ uint32_t nva3_pwr_code[] = {
 	0x90fca0fc,
 	0x00fc80fc,
 	0xf80032f4,
-/* 0x01f5: timer */
-	0x1032f401,
-	0xb003f898,
-	0x1cf40086,
-	0x03fe8051,
+/* 0x0205: ticks_from_ns */
+	0xf9c0f901,
+	0xcbd7f1b0,
+	0x00d3f000,
+	0x041321f5,
+	0x03e8ccec,
+	0xf400b4b0,
+	0xeeec120b,
+	0xd7f103e8,
+	0xd3f000cb,
+	0x1321f500,
+/* 0x022d: ticks_from_ns_quit */
+	0x02ceb904,
+	0xc0fcb0fc,
+/* 0x0236: ticks_from_us */
+	0xc0f900f8,
+	0xd7f1b0f9,
+	0xd3f000cb,
+	0x1321f500,
+	0x02ceb904,
+	0xf400b4b0,
+	0xe4bd050b,
+/* 0x0250: ticks_from_us_quit */
+	0xc0fcb0fc,
+/* 0x0256: ticks_to_us */
+	0xd7f100f8,
+	0xd3f000cb,
+	0xecedff00,
+/* 0x0262: timer */
+	0x90f900f8,
+	0x32f480f9,
+	0x03f89810,
+	0xf40086b0,
+	0x84bd651c,
 	0xb63807f0,
 	0x08d00604,
 	0xf004bd00,
-	0x84b60887,
+	0x84b63487,
 	0x0088cf06,
-	0xf40284f0,
-	0x87f0261b,
-	0x0684b634,
-	0xb80088cf,
-	0x0bf406e0,
-	0x06e8b809,
-/* 0x0233: timer_reset */
-	0xf01f1ef4,
-	0x04b63407,
-	0x000ed006,
-	0x0e8004bd,
-/* 0x0241: timer_enable */
-	0x0187f09a,
+	0xbb9a0998,
+	0xe9bb0298,
+	0x03fe8000,
+	0xb60887f0,
+	0x88cf0684,
+	0x0284f000,
+	0xf0261bf4,
+	0x84b63487,
+	0x0088cf06,
+	0xf406e0b8,
+	0xe8b8090b,
+	0x111cf406,
+/* 0x02b8: timer_reset */
+	0xb63407f0,
+	0x0ed00604,
+	0x8004bd00,
+/* 0x02c6: timer_enable */
+	0x87f09a0e,
+	0x3807f001,
+	0xd00604b6,
+	0x04bd0008,
+/* 0x02d4: timer_done */
+	0xfc1031f4,
+	0xf890fc80,
+/* 0x02dd: send_proc */
+	0xf980f900,
+	0x05e89890,
+	0xf004e998,
+	0x89b80486,
+	0x2a0bf406,
+	0x940398c4,
+	0x80b60488,
+	0x008ebb18,
+	0x8000fa98,
+	0x8d80008a,
+	0x028c8001,
+	0xb6038b80,
+	0x94f00190,
+	0x04e98007,
+/* 0x0317: send_done */
+	0xfc0231f4,
+	0xf880fc90,
+/* 0x031d: find */
+	0xf080f900,
+	0x31f45887,
+/* 0x0325: find_loop */
+	0x008a9801,
+	0xf406aeb8,
+	0x80b6100b,
+	0x6886b158,
+	0xf01bf402,
+/* 0x033b: find_done */
+	0xb90132f4,
+	0x80fc028e,
+/* 0x0342: send */
+	0x21f500f8,
+	0x01f4031d,
+/* 0x034b: recv */
+	0xf900f897,
+	0x9880f990,
+	0xe99805e8,
+	0x0132f404,
+	0xf40689b8,
+	0x89c43d0b,
+	0x0180b603,
+	0x800784f0,
+	0xea9805e8,
+	0xfef0f902,
+	0xf0f9018f,
+	0x9402efb9,
+	0xe9bb0499,
+	0x18e0b600,
+	0x9803eb98,
+	0xed9802ec,
+	0x00ee9801,
+	0xf0fca5f9,
+	0xf400f8fe,
+	0xf0fc0131,
+/* 0x0398: recv_done */
+	0x90fc80fc,
+/* 0x039e: init */
+	0x17f100f8,
+	0x14b60108,
+	0x0011cf06,
+	0x010911e7,
+	0xfe0814b6,
+	0x17f10014,
+	0x13f000e0,
+	0x1c07f000,
+	0xd00604b6,
+	0x04bd0001,
+	0xf0ff17f0,
+	0x04b61407,
+	0x0001d006,
+	0x17f004bd,
+	0x0015f102,
+	0x1007f008,
+	0xd00604b6,
+	0x04bd0001,
+	0x011a17f1,
+	0xfe0013f0,
+	0x31f40010,
+	0x0117f010,
 	0xb63807f0,
-	0x08d00604,
-/* 0x024f: timer_done */
-	0xf404bd00,
-	0x00f81031,
-/* 0x0254: send_proc */
-	0x90f980f9,
-	0x9805e898,
-	0x86f004e9,
-	0x0689b804,
-	0xc42a0bf4,
-	0x88940398,
-	0x1880b604,
-	0x98008ebb,
-	0x8a8000fa,
-	0x018d8000,
-	0x80028c80,
-	0x90b6038b,
-	0x0794f001,
-	0xf404e980,
-/* 0x028e: send_done */
-	0x90fc0231,
-	0x00f880fc,
-/* 0x0294: find */
-	0x87f080f9,
-	0x0131f458,
-/* 0x029c: find_loop */
-	0xb8008a98,
-	0x0bf406ae,
-	0x5880b610,
-	0x026886b1,
-	0xf4f01bf4,
-/* 0x02b2: find_done */
-	0x8eb90132,
-	0xf880fc02,
-/* 0x02b9: send */
-	0x9421f500,
-	0x9701f402,
-/* 0x02c2: recv */
-	0xe89800f8,
-	0x04e99805,
-	0xb80132f4,
-	0x0bf40689,
-	0x0389c43d,
-	0xf00180b6,
-	0xe8800784,
-	0x02ea9805,
-	0x8ffef0f9,
-	0xb9f0f901,
-	0x999402ef,
-	0x00e9bb04,
-	0x9818e0b6,
-	0xec9803eb,
-	0x01ed9802,
-	0xf900ee98,
-	0xfef0fca5,
-	0x31f400f8,
-/* 0x030b: recv_done */
-	0xf8f0fc01,
-/* 0x030d: init */
-	0x0817f100,
-	0x0614b601,
-	0xe70011cf,
-	0xb6010911,
-	0x14fe0814,
-	0xe017f100,
-	0x0013f000,
-	0xb61c07f0,
 	0x01d00604,
 	0xf004bd00,
-	0x07f0ff17,
-	0x0604b614,
-	0xbd0001d0,
-	0x0217f004,
-	0x080015f1,
-	0xb61007f0,
-	0x01d00604,
-	0xf104bd00,
-	0xf0010a17,
-	0x10fe0013,
-	0x1031f400,
-	0xf00117f0,
-	0x04b63807,
-	0x0001d006,
-	0xf7f004bd,
-/* 0x0371: init_proc */
-	0x01f19858,
-	0xf40016b0,
-	0x15f9fa0b,
-	0xf458f0b6,
-/* 0x0382: host_send */
-	0x17f1f20e,
-	0x14b604b0,
-	0x0011cf06,
-	0x04a027f1,
-	0xcf0624b6,
-	0x12b80022,
-	0x320bf406,
-	0x94071ec4,
-	0xe0b704ee,
-	0xeb980270,
-	0x02ec9803,
-	0x9801ed98,
-	0x21f500ee,
-	0x10b602b9,
-	0x0f1ec401,
-	0x04b007f1,
-	0xd00604b6,
-	0x04bd000e,
-/* 0x03cb: host_send_done */
-	0xf8ba0ef4,
-/* 0x03cd: host_recv */
-	0x4917f100,
-	0x5413f14e,
-	0x06e1b852,
-/* 0x03db: host_recv_wait */
-	0xf1aa0bf4,
-	0xb604cc17,
-	0x11cf0614,
-	0xc827f100,
-	0x0624b604,
-	0xf00022cf,
-	0x12b80816,
-	0xe60bf406,
-	0xb60723c4,
-	0x30b70434,
-	0x3b8002f0,
-	0x023c8003,
-	0x80013d80,
-	0x20b6003e,
-	0x0f24f001,
-	0x04c807f1,
+/* 0x0402: init_proc */
+	0xf19858f7,
+	0x0016b001,
+	0xf9fa0bf4,
+	0x58f0b615,
+/* 0x0413: mulu32_32_64 */
+	0xf9f20ef4,
+	0xf920f910,
+	0x9540f930,
+	0xd29510e1,
+	0xbdc4bd10,
+	0xc0edffb4,
+	0xb9301dff,
+	0x34f10234,
+	0x34b6ffff,
+	0x1045b610,
+	0xbb00c3bb,
+	0xe2ff01b4,
+	0x0234b930,
+	0xffff34f1,
+	0xb61034b6,
+	0xc3bb1045,
+	0x01b4bb00,
+	0xbb3012ff,
+	0x40fc00b3,
+	0x20fc30fc,
+	0x00f810fc,
+/* 0x0464: host_send */
+	0x04b017f1,
+	0xcf0614b6,
+	0x27f10011,
+	0x24b604a0,
+	0x0022cf06,
+	0xf40612b8,
+	0x1ec4320b,
+	0x04ee9407,
+	0x0270e0b7,
+	0x9803eb98,
+	0xed9802ec,
+	0x00ee9801,
+	0x034221f5,
+	0xc40110b6,
+	0x07f10f1e,
+	0x04b604b0,
+	0x000ed006,
+	0x0ef404bd,
+/* 0x04ad: host_send_done */
+/* 0x04af: host_recv */
+	0xf100f8ba,
+	0xf14e4917,
+	0xb8525413,
+	0x0bf406e1,
+/* 0x04bd: host_recv_wait */
+	0xcc17f1aa,
+	0x0614b604,
+	0xf10011cf,
+	0xb604c827,
+	0x22cf0624,
+	0x0816f000,
+	0xf40612b8,
+	0x23c4e60b,
+	0x0434b607,
+	0x02f030b7,
+	0x80033b80,
+	0x3d80023c,
+	0x003e8001,
+	0xf00120b6,
+	0x07f10f24,
+	0x04b604c8,
+	0x0002d006,
+	0x27f004bd,
+	0x0007f040,
 	0xd00604b6,
 	0x04bd0002,
-	0xf04027f0,
-	0x04b60007,
-	0x0002d006,
-	0x00f804bd,
-/* 0x0430: host_init */
-	0x008017f1,
-	0xf11014b6,
-	0xf1027015,
-	0xb604d007,
-	0x01d00604,
-	0xf104bd00,
-	0xb6008017,
-	0x15f11014,
-	0x07f102f0,
-	0x04b604dc,
-	0x0001d006,
-	0x17f004bd,
-	0xc407f101,
+/* 0x0512: host_init */
+	0x17f100f8,
+	0x14b60080,
+	0x7015f110,
+	0xd007f102,
 	0x0604b604,
 	0xbd0001d0,
-/* 0x046f: memx_func_enter */
-	0xf000f804,
+	0x8017f104,
+	0x1014b600,
+	0x02f015f1,
+	0x04dc07f1,
+	0xd00604b6,
+	0x04bd0001,
+	0xf10117f0,
+	0xb604c407,
+	0x01d00604,
+	0xf804bd00,
+/* 0x0551: memx_func_enter */
+	0x1087f100,
+	0x028eb916,
+	0xb90421f4,
+	0x67f102d7,
+	0x63f1fffc,
+	0x76fdffff,
+	0x0267f104,
+	0x0576fd00,
+	0x70f980f9,
+	0xe0fcd0fc,
+	0xf03f21f4,
 	0x07f10467,
 	0x04b607e0,
 	0x0006d006,
-/* 0x047e: memx_func_enter_wait */
+/* 0x058a: memx_func_enter_wait */
 	0x67f104bd,
 	0x64b607c0,
 	0x0066cf06,
 	0xf40464f0,
-	0x1698f30b,
-	0x0410b600,
-/* 0x0496: memx_func_leave */
-	0x67f000f8,
-	0xe407f104,
-	0x0604b607,
-	0xbd0006d0,
-/* 0x04a5: memx_func_leave_wait */
-	0xc067f104,
+	0x67f0f30b,
+	0x0664b62c,
+	0x800066cf,
+	0x00f8f106,
+/* 0x05a8: memx_func_leave */
+	0xb62c67f0,
+	0x66cf0664,
+	0xf2068000,
+	0xf10467f0,
+	0xb607e407,
+	0x06d00604,
+/* 0x05c3: memx_func_leave_wait */
+	0xf104bd00,
+	0xb607c067,
+	0x66cf0664,
+	0x0464f000,
+	0xf1f31bf4,
+	0xb9161087,
+	0x21f4028e,
+	0x02d7b904,
+	0xffcc67f1,
+	0xffff63f1,
+	0xf90476fd,
+	0xfc70f980,
+	0xf4e0fcd0,
+	0x00f83f21,
+/* 0x05f8: memx_func_wait_vblank */
+	0xb0001698,
+	0x0bf40066,
+	0x0166b013,
+	0xf4060bf4,
+/* 0x060a: memx_func_wait_vblank_head1 */
+	0x77f12e0e,
+	0x0ef40020,
+/* 0x0611: memx_func_wait_vblank_head0 */
+	0x0877f107,
+/* 0x0615: memx_func_wait_vblank_0 */
+	0xc467f100,
 	0x0664b607,
-	0xf00066cf,
-	0x1bf40464,
-/* 0x04b7: memx_func_wr32 */
-	0x9800f8f3,
-	0x15980016,
-	0x0810b601,
-	0x50f960f9,
+	0xfd0066cf,
+	0x1bf40467,
+/* 0x0625: memx_func_wait_vblank_1 */
+	0xc467f1f3,
+	0x0664b607,
+	0xfd0066cf,
+	0x0bf40467,
+/* 0x0635: memx_func_wait_vblank_fini */
+	0x0410b6f3,
+/* 0x063a: memx_func_wr32 */
+	0x169800f8,
+	0x01159800,
+	0xf90810b6,
+	0xfc50f960,
+	0xf4e0fcd0,
+	0x42b63f21,
+	0xe91bf402,
+/* 0x0656: memx_func_wait */
+	0x87f000f8,
+	0x0684b62c,
+	0x980088cf,
+	0x1d98001e,
+	0x021c9801,
+	0xb6031b98,
+	0x21f41010,
+/* 0x0673: memx_func_delay */
+	0x9800f8a4,
+	0x10b6001e,
+	0x7f21f404,
+/* 0x067e: memx_func_train */
+	0x57f100f8,
+	0x77f10003,
+	0x97f10000,
+	0x93f00000,
+	0x029eb970,
+	0xb90421f4,
+	0xe7f102d8,
+	0x21f42710,
+/* 0x069d: memx_func_train_loop_outer */
+	0x0158e07f,
+	0x0083f101,
+	0xe097f102,
+	0x1193f011,
+	0x80f990f9,
 	0xe0fcd0fc,
-	0xb63f21f4,
-	0x1bf40242,
-/* 0x04d3: memx_func_wait */
-	0xf000f8e9,
-	0x84b62c87,
-	0x0088cf06,
-	0x98001e98,
-	0x1c98011d,
-	0x031b9802,
-	0xf41010b6,
-	0x00f89c21,
-/* 0x04f0: memx_func_delay */
-	0xb6001e98,
-	0x21f40410,
-/* 0x04fb: memx_exec */
-	0xf900f87f,
+	0xf93f21f4,
+	0x0067f150,
+/* 0x06bd: memx_func_train_loop_inner */
+	0x1187f100,
+	0x9068ff11,
+	0xfd109894,
+	0x97f10589,
+	0x93f00720,
+	0xf990f910,
+	0xfcd0fc80,
+	0x3f21f4e0,
+	0x008097f1,
+	0xb91093f0,
+	0x21f4029e,
+	0x02d8b904,
+	0xf92088c5,
+	0xfc80f990,
+	0xf4e0fcd0,
+	0x97f13f21,
+	0x93f0053c,
+	0x0287f110,
+	0x0083f130,
+	0xf990f980,
+	0xfcd0fc80,
+	0x3f21f4e0,
+	0x0560e7f1,
+	0xf110e3f0,
+	0xf10000d7,
+	0x908000d3,
+	0xb7f100dc,
+	0xb3f08480,
+	0xa421f41e,
+	0x000057f1,
+	0xffff97f1,
+	0x830093f1,
+/* 0x073c: memx_func_train_loop_4x */
+	0x0080a7f1,
+	0xb910a3f0,
+	0x21f402ae,
+	0x02d8b904,
+	0xffdfb7f1,
+	0xffffb3f1,
+	0xf9048bfd,
+	0xfc80f9a0,
+	0xf4e0fcd0,
+	0xa7f13f21,
+	0xa3f0053c,
+	0x0287f110,
+	0x0083f130,
+	0xf9a0f980,
+	0xfcd0fc80,
+	0x3f21f4e0,
+	0x0560e7f1,
+	0xf110e3f0,
+	0xf10000d7,
+	0xb98000d3,
+	0xb7f102dc,
+	0xb3f02710,
+	0xa421f400,
+	0xf402eeb9,
+	0xddb90421,
+	0x949dff02,
+	0x700150b6,
+	0x1ef40456,
+	0xcc7aa092,
+	0x00a9800b,
+	0xb60160b6,
+	0x66700470,
+	0x001ef510,
+	0xb650fcff,
+	0x56700150,
+	0xd41ef507,
+/* 0x07cf: memx_exec */
+	0xf900f8fe,
 	0xb9d0f9e0,
 	0xb2b902c1,
-/* 0x0505: memx_exec_next */
+/* 0x07d9: memx_exec_next */
 	0x00139802,
-	0x950410b6,
-	0x30f01034,
+	0xe70410b6,
+	0xe701f034,
+	0xb601e033,
+	0x30f00132,
 	0xde35980c,
 	0x12b855f9,
-	0xec1ef406,
-	0xe0fcd0fc,
-	0x02b921f5,
-/* 0x0526: memx_info */
-	0xc7f100f8,
-	0xb7f103ac,
-	0x21f50800,
-	0x00f802b9,
-/* 0x0534: memx_recv */
+	0xe41ef406,
+	0x98f10b98,
+	0xcbbbf20c,
+	0xc4b7f102,
+	0x06b4b607,
+	0xfc00bbcf,
+	0xf5e0fcd0,
+	0xf8034221,
+/* 0x0815: memx_info */
+	0x01c67000,
+/* 0x081b: memx_info_data */
+	0xf10e0bf4,
+	0xf103ccc7,
+	0xf40800b7,
+/* 0x0826: memx_info_train */
+	0xc7f10b0e,
+	0xb7f10bcc,
+/* 0x082e: memx_info_send */
+	0x21f50100,
+	0x00f80342,
+/* 0x0834: memx_recv */
 	0xf401d6b0,
-	0xd6b0c40b,
-	0xe90bf400,
-/* 0x0542: memx_init */
+	0xd6b0980b,
+	0xd80bf400,
+/* 0x0842: memx_init */
 	0x00f800f8,
-/* 0x0544: perf_recv */
-/* 0x0546: perf_init */
+/* 0x0844: perf_recv */
+/* 0x0846: perf_init */
 	0x00f800f8,
-/* 0x0548: i2c_drive_scl */
+/* 0x0848: i2c_drive_scl */
 	0xf40036b0,
 	0x07f1110b,
 	0x04b607e0,
 	0x0001d006,
 	0x00f804bd,
-/* 0x055c: i2c_drive_scl_lo */
+/* 0x085c: i2c_drive_scl_lo */
 	0x07e407f1,
 	0xd00604b6,
 	0x04bd0001,
-/* 0x056a: i2c_drive_sda */
+/* 0x086a: i2c_drive_sda */
 	0x36b000f8,
 	0x110bf400,
 	0x07e007f1,
 	0xd00604b6,
 	0x04bd0002,
-/* 0x057e: i2c_drive_sda_lo */
+/* 0x087e: i2c_drive_sda_lo */
 	0x07f100f8,
 	0x04b607e4,
 	0x0002d006,
 	0x00f804bd,
-/* 0x058c: i2c_sense_scl */
+/* 0x088c: i2c_sense_scl */
 	0xf10132f4,
 	0xb607c437,
 	0x33cf0634,
 	0x0431fd00,
 	0xf4060bf4,
-/* 0x05a2: i2c_sense_scl_done */
+/* 0x08a2: i2c_sense_scl_done */
 	0x00f80131,
-/* 0x05a4: i2c_sense_sda */
+/* 0x08a4: i2c_sense_sda */
 	0xf10132f4,
 	0xb607c437,
 	0x33cf0634,
 	0x0432fd00,
 	0xf4060bf4,
-/* 0x05ba: i2c_sense_sda_done */
+/* 0x08ba: i2c_sense_sda_done */
 	0x00f80131,
-/* 0x05bc: i2c_raise_scl */
+/* 0x08bc: i2c_raise_scl */
 	0x47f140f9,
 	0x37f00898,
 	0x4821f501,
-/* 0x05c9: i2c_raise_scl_wait */
-	0xe8e7f105,
+/* 0x08c9: i2c_raise_scl_wait */
+	0xe8e7f108,
 	0x7f21f403,
-	0x058c21f5,
+	0x088c21f5,
 	0xb60901f4,
 	0x1bf40142,
-/* 0x05dd: i2c_raise_scl_done */
+/* 0x08dd: i2c_raise_scl_done */
 	0xf840fcef,
-/* 0x05e1: i2c_start */
+/* 0x08e1: i2c_start */
 	0x8c21f500,
-	0x0d11f405,
-	0x05a421f5,
+	0x0d11f408,
+	0x08a421f5,
 	0xf40611f4,
-/* 0x05f2: i2c_start_rep */
+/* 0x08f2: i2c_start_rep */
 	0x37f0300e,
 	0x4821f500,
-	0x0137f005,
-	0x056a21f5,
+	0x0137f008,
+	0x086a21f5,
 	0xb60076bb,
 	0x50f90465,
 	0xbb046594,
 	0x50bd0256,
 	0xfc0475fd,
 	0xbc21f550,
-	0x0464b605,
-/* 0x061f: i2c_start_send */
+	0x0464b608,
+/* 0x091f: i2c_start_send */
 	0xf01f11f4,
 	0x21f50037,
-	0xe7f1056a,
+	0xe7f1086a,
 	0x21f41388,
 	0x0037f07f,
-	0x054821f5,
+	0x084821f5,
 	0x1388e7f1,
-/* 0x063b: i2c_start_out */
+/* 0x093b: i2c_start_out */
 	0xf87f21f4,
-/* 0x063d: i2c_stop */
+/* 0x093d: i2c_stop */
 	0x0037f000,
-	0x054821f5,
+	0x084821f5,
 	0xf50037f0,
-	0xf1056a21,
+	0xf1086a21,
 	0xf403e8e7,
 	0x37f07f21,
 	0x4821f501,
-	0x88e7f105,
+	0x88e7f108,
 	0x7f21f413,
 	0xf50137f0,
-	0xf1056a21,
+	0xf1086a21,
 	0xf41388e7,
 	0x00f87f21,
-/* 0x0670: i2c_bitw */
-	0x056a21f5,
+/* 0x0970: i2c_bitw */
+	0x086a21f5,
 	0x03e8e7f1,
 	0xbb7f21f4,
 	0x65b60076,
@@ -1338,18 +1617,18 @@ uint32_t nva3_pwr_code[] = {
 	0x56bb0465,
 	0xfd50bd02,
 	0x50fc0475,
-	0x05bc21f5,
+	0x08bc21f5,
 	0xf40464b6,
 	0xe7f11811,
 	0x21f41388,
 	0x0037f07f,
-	0x054821f5,
+	0x084821f5,
 	0x1388e7f1,
-/* 0x06af: i2c_bitw_out */
+/* 0x09af: i2c_bitw_out */
 	0xf87f21f4,
-/* 0x06b1: i2c_bitr */
+/* 0x09b1: i2c_bitr */
 	0x0137f000,
-	0x056a21f5,
+	0x086a21f5,
 	0x03e8e7f1,
 	0xbb7f21f4,
 	0x65b60076,
@@ -1357,19 +1636,19 @@ uint32_t nva3_pwr_code[] = {
 	0x56bb0465,
 	0xfd50bd02,
 	0x50fc0475,
-	0x05bc21f5,
+	0x08bc21f5,
 	0xf40464b6,
 	0x21f51b11,
-	0x37f005a4,
+	0x37f008a4,
 	0x4821f500,
-	0x88e7f105,
+	0x88e7f108,
 	0x7f21f413,
 	0xf4013cf0,
-/* 0x06f6: i2c_bitr_done */
+/* 0x09f6: i2c_bitr_done */
 	0x00f80131,
-/* 0x06f8: i2c_get_byte */
+/* 0x09f8: i2c_get_byte */
 	0xf00057f0,
-/* 0x06fe: i2c_get_byte_next */
+/* 0x09fe: i2c_get_byte_next */
 	0x54b60847,
 	0x0076bb01,
 	0xf90465b6,
@@ -1377,7 +1656,7 @@ uint32_t nva3_pwr_code[] = {
 	0xbd0256bb,
 	0x0475fd50,
 	0x21f550fc,
-	0x64b606b1,
+	0x64b609b1,
 	0x2b11f404,
 	0xb60553fd,
 	0x1bf40142,
@@ -1388,11 +1667,11 @@ uint32_t nva3_pwr_code[] = {
 	0x50bd0256,
 	0xfc0475fd,
 	0x7021f550,
-	0x0464b606,
-/* 0x0748: i2c_get_byte_done */
-/* 0x074a: i2c_put_byte */
+	0x0464b609,
+/* 0x0a48: i2c_get_byte_done */
+/* 0x0a4a: i2c_put_byte */
 	0x47f000f8,
-/* 0x074d: i2c_put_byte_next */
+/* 0x0a4d: i2c_put_byte_next */
 	0x0142b608,
 	0xbb3854ff,
 	0x65b60076,
@@ -1400,7 +1679,7 @@ uint32_t nva3_pwr_code[] = {
 	0x56bb0465,
 	0xfd50bd02,
 	0x50fc0475,
-	0x067021f5,
+	0x097021f5,
 	0xf40464b6,
 	0x46b03411,
 	0xd81bf400,
@@ -1410,20 +1689,20 @@ uint32_t nva3_pwr_code[] = {
 	0x50bd0256,
 	0xfc0475fd,
 	0xb121f550,
-	0x0464b606,
+	0x0464b609,
 	0xbb0f11f4,
 	0x36b00076,
 	0x061bf401,
-/* 0x07a3: i2c_put_byte_done */
+/* 0x0aa3: i2c_put_byte_done */
 	0xf80132f4,
-/* 0x07a5: i2c_addr */
+/* 0x0aa5: i2c_addr */
 	0x0076bb00,
 	0xf90465b6,
 	0x04659450,
 	0xbd0256bb,
 	0x0475fd50,
 	0x21f550fc,
-	0x64b605e1,
+	0x64b608e1,
 	0x2911f404,
 	0x012ec3e7,
 	0xfd0134b6,
@@ -1433,31 +1712,31 @@ uint32_t nva3_pwr_code[] = {
 	0x0256bb04,
 	0x75fd50bd,
 	0xf550fc04,
-	0xb6074a21,
-/* 0x07ea: i2c_addr_done */
+	0xb60a4a21,
+/* 0x0aea: i2c_addr_done */
 	0x00f80464,
-/* 0x07ec: i2c_acquire_addr */
+/* 0x0aec: i2c_acquire_addr */
 	0xb6f8cec7,
 	0xe0b702e4,
-	0xee980bfc,
-/* 0x07fb: i2c_acquire */
+	0xee980d1c,
+/* 0x0afb: i2c_acquire */
 	0xf500f800,
-	0xf407ec21,
+	0xf40aec21,
 	0xd9f00421,
 	0x3f21f403,
-/* 0x080a: i2c_release */
+/* 0x0b0a: i2c_release */
 	0x21f500f8,
-	0x21f407ec,
+	0x21f40aec,
 	0x03daf004,
 	0xf83f21f4,
-/* 0x0819: i2c_recv */
+/* 0x0b19: i2c_recv */
 	0x0132f400,
 	0xb6f8c1c7,
 	0x16b00214,
 	0x3a1ff528,
-	0xd413a001,
-	0x0032980b,
-	0x0bac13a0,
+	0xf413a001,
+	0x0032980c,
+	0x0ccc13a0,
 	0xf4003198,
 	0xd0f90231,
 	0xd0f9e0f9,
@@ -1469,7 +1748,7 @@ uint32_t nva3_pwr_code[] = {
 	0x56bb0465,
 	0xfd50bd02,
 	0x50fc0475,
-	0x07fb21f5,
+	0x0afb21f5,
 	0xfc0464b6,
 	0x00d6b0d0,
 	0x00b31bf5,
@@ -1479,7 +1758,7 @@ uint32_t nva3_pwr_code[] = {
 	0x56bb0465,
 	0xfd50bd02,
 	0x50fc0475,
-	0x07a521f5,
+	0x0aa521f5,
 	0xf50464b6,
 	0xc700d011,
 	0x76bbe0c5,
@@ -1488,7 +1767,7 @@ uint32_t nva3_pwr_code[] = {
 	0x0256bb04,
 	0x75fd50bd,
 	0xf550fc04,
-	0xb6074a21,
+	0xb60a4a21,
 	0x11f50464,
 	0x57f000ad,
 	0x0076bb01,
@@ -1497,7 +1776,7 @@ uint32_t nva3_pwr_code[] = {
 	0xbd0256bb,
 	0x0475fd50,
 	0x21f550fc,
-	0x64b607a5,
+	0x64b60aa5,
 	0x8a11f504,
 	0x0076bb00,
 	0xf90465b6,
@@ -1505,7 +1784,7 @@ uint32_t nva3_pwr_code[] = {
 	0xbd0256bb,
 	0x0475fd50,
 	0x21f550fc,
-	0x64b606f8,
+	0x64b609f8,
 	0x6a11f404,
 	0xbbe05bcb,
 	0x65b60076,
@@ -1513,38 +1792,38 @@ uint32_t nva3_pwr_code[] = {
 	0x56bb0465,
 	0xfd50bd02,
 	0x50fc0475,
-	0x063d21f5,
+	0x093d21f5,
 	0xb90464b6,
 	0x74bd025b,
-/* 0x091f: i2c_recv_not_rd08 */
+/* 0x0c1f: i2c_recv_not_rd08 */
 	0xb0430ef4,
 	0x1bf401d6,
 	0x0057f03d,
-	0x07a521f5,
+	0x0aa521f5,
 	0xc73311f4,
 	0x21f5e0c5,
-	0x11f4074a,
+	0x11f40a4a,
 	0x0057f029,
-	0x07a521f5,
+	0x0aa521f5,
 	0xc71f11f4,
 	0x21f5e0b5,
-	0x11f4074a,
+	0x11f40a4a,
 	0x3d21f515,
-	0xc774bd06,
+	0xc774bd09,
 	0x1bf408c5,
 	0x0232f409,
-/* 0x095f: i2c_recv_not_wr08 */
-/* 0x095f: i2c_recv_done */
+/* 0x0c5f: i2c_recv_not_wr08 */
+/* 0x0c5f: i2c_recv_done */
 	0xc7030ef4,
 	0x21f5f8ce,
-	0xe0fc080a,
+	0xe0fc0b0a,
 	0x12f4d0fc,
 	0x027cb90a,
-	0x02b921f5,
-/* 0x0974: i2c_recv_exit */
-/* 0x0976: i2c_init */
+	0x034221f5,
+/* 0x0c74: i2c_recv_exit */
+/* 0x0c76: i2c_init */
 	0x00f800f8,
-/* 0x0978: test_recv */
+/* 0x0c78: test_recv */
 	0x05d817f1,
 	0xcf0614b6,
 	0x10b60011,
@@ -1553,13 +1832,13 @@ uint32_t nva3_pwr_code[] = {
 	0xbd0001d0,
 	0x00e7f104,
 	0x4fe3f1d9,
-	0xf521f513,
-/* 0x099f: test_init */
-	0xf100f801,
+	0x6221f513,
+/* 0x0c9f: test_init */
+	0xf100f802,
 	0xf50800e7,
-	0xf801f521,
-/* 0x09a9: idle_recv */
-/* 0x09ab: idle */
+	0xf8026221,
+/* 0x0ca9: idle_recv */
+/* 0x0cab: idle */
 	0xf400f800,
 	0x17f10031,
 	0x14b605d4,
@@ -1567,17 +1846,17 @@ uint32_t nva3_pwr_code[] = {
 	0xf10110b6,
 	0xb605d407,
 	0x01d00604,
-/* 0x09c7: idle_loop */
+/* 0x0cc7: idle_loop */
 	0xf004bd00,
 	0x32f45817,
-/* 0x09cd: idle_proc */
-/* 0x09cd: idle_proc_exec */
+/* 0x0ccd: idle_proc */
+/* 0x0ccd: idle_proc_exec */
 	0xb910f902,
 	0x21f5021e,
-	0x10fc02c2,
+	0x10fc034b,
 	0xf40911f4,
 	0x0ef40231,
-/* 0x09e1: idle_proc_next */
+/* 0x0ce1: idle_proc_next */
 	0x5810b6ef,
 	0xf4061fb8,
 	0x02f4e61b,
diff --git a/drivers/gpu/drm/nouveau/core/subdev/pwr/fuc/nvc0.fuc b/drivers/gpu/drm/nouveau/core/subdev/pwr/fuc/nvc0.fuc
index 48f79434a449..21bf8cc7618f 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/pwr/fuc/nvc0.fuc
+++ b/drivers/gpu/drm/nouveau/core/subdev/pwr/fuc/nvc0.fuc
@@ -23,6 +23,7 @@
  */
 
 #define NVKM_PPWR_CHIPSET GF100
+#define HW_TICKS_PER_US 203 // should be 202.5
 
 //#define NVKM_FALCON_PC24
 //#define NVKM_FALCON_UNSHIFTED_IO
@@ -34,6 +35,7 @@
 .section #nvc0_pwr_data
 #define INCLUDE_PROC
 #include "kernel.fuc"
+#include "arith.fuc"
 #include "host.fuc"
 #include "memx.fuc"
 #include "perf.fuc"
@@ -44,6 +46,7 @@
 
 #define INCLUDE_DATA
 #include "kernel.fuc"
+#include "arith.fuc"
 #include "host.fuc"
 #include "memx.fuc"
 #include "perf.fuc"
@@ -56,6 +59,7 @@
 .section #nvc0_pwr_code
 #define INCLUDE_CODE
 #include "kernel.fuc"
+#include "arith.fuc"
 #include "host.fuc"
 #include "memx.fuc"
 #include "perf.fuc"
diff --git a/drivers/gpu/drm/nouveau/core/subdev/pwr/fuc/nvc0.fuc.h b/drivers/gpu/drm/nouveau/core/subdev/pwr/fuc/nvc0.fuc.h
index 0773ff0e3dc3..90221d973f84 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/pwr/fuc/nvc0.fuc.h
+++ b/drivers/gpu/drm/nouveau/core/subdev/pwr/fuc/nvc0.fuc.h
@@ -24,8 +24,8 @@ uint32_t nvc0_pwr_data[] = {
 	0x00000000,
 /* 0x0058: proc_list_head */
 	0x54534f48,
-	0x00000430,
-	0x000003cd,
+	0x00000512,
+	0x000004af,
 	0x00000000,
 	0x00000000,
 	0x00000000,
@@ -46,8 +46,8 @@ uint32_t nvc0_pwr_data[] = {
 	0x00000000,
 	0x00000000,
 	0x584d454d,
-	0x00000542,
-	0x00000534,
+	0x0000075e,
+	0x00000750,
 	0x00000000,
 	0x00000000,
 	0x00000000,
@@ -68,8 +68,8 @@ uint32_t nvc0_pwr_data[] = {
 	0x00000000,
 	0x00000000,
 	0x46524550,
-	0x00000546,
-	0x00000544,
+	0x00000762,
+	0x00000760,
 	0x00000000,
 	0x00000000,
 	0x00000000,
@@ -90,8 +90,8 @@ uint32_t nvc0_pwr_data[] = {
 	0x00000000,
 	0x00000000,
 	0x5f433249,
-	0x00000976,
-	0x00000819,
+	0x00000b92,
+	0x00000a35,
 	0x00000000,
 	0x00000000,
 	0x00000000,
@@ -112,8 +112,8 @@ uint32_t nvc0_pwr_data[] = {
 	0x00000000,
 	0x00000000,
 	0x54534554,
-	0x0000099f,
-	0x00000978,
+	0x00000bbb,
+	0x00000b94,
 	0x00000000,
 	0x00000000,
 	0x00000000,
@@ -134,8 +134,8 @@ uint32_t nvc0_pwr_data[] = {
 	0x00000000,
 	0x00000000,
 	0x454c4449,
-	0x000009ab,
-	0x000009a9,
+	0x00000bc7,
+	0x00000bc5,
 	0x00000000,
 	0x00000000,
 	0x00000000,
@@ -227,24 +227,74 @@ uint32_t nvc0_pwr_data[] = {
 	0x00000000,
 	0x00000000,
 /* 0x0370: memx_func_head */
-	0x00010000,
-	0x00000000,
-	0x0000046f,
-/* 0x037c: memx_func_next */
 	0x00000001,
 	0x00000000,
-	0x00000496,
+	0x00000551,
+/* 0x037c: memx_func_next */
 	0x00000002,
+	0x00000000,
+	0x000005db,
+	0x00000003,
 	0x00000002,
-	0x000004b7,
-	0x00040003,
+	0x000006a5,
+	0x00040004,
+	0x00000000,
+	0x000006c1,
+	0x00010005,
+	0x00000000,
+	0x000006de,
+	0x00010006,
+	0x00000000,
+	0x00000663,
+	0x00000007,
+	0x00000000,
+	0x000006e9,
+/* 0x03c4: memx_func_tail */
+/* 0x03c4: memx_ts_start */
+	0x00000000,
+/* 0x03c8: memx_ts_end */
+	0x00000000,
+/* 0x03cc: memx_data_head */
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
 	0x00000000,
-	0x000004d3,
-	0x00010004,
 	0x00000000,
-	0x000004f0,
-/* 0x03ac: memx_func_tail */
-/* 0x03ac: memx_data_head */
 	0x00000000,
 	0x00000000,
 	0x00000000,
@@ -717,6 +767,8 @@ uint32_t nvc0_pwr_data[] = {
 	0x00000000,
 	0x00000000,
 	0x00000000,
+/* 0x0bcc: memx_data_tail */
+/* 0x0bcc: memx_train_head */
 	0x00000000,
 	0x00000000,
 	0x00000000,
@@ -757,8 +809,32 @@ uint32_t nvc0_pwr_data[] = {
 	0x00000000,
 	0x00000000,
 	0x00000000,
-/* 0x0bac: memx_data_tail */
-/* 0x0bac: i2c_scl_map */
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+/* 0x0ccc: memx_train_tail */
+/* 0x0ccc: i2c_scl_map */
 	0x00001000,
 	0x00004000,
 	0x00010000,
@@ -769,7 +845,7 @@ uint32_t nvc0_pwr_data[] = {
 	0x01000000,
 	0x04000000,
 	0x10000000,
-/* 0x0bd4: i2c_sda_map */
+/* 0x0cf4: i2c_sda_map */
 	0x00002000,
 	0x00008000,
 	0x00020000,
@@ -780,7 +856,7 @@ uint32_t nvc0_pwr_data[] = {
 	0x02000000,
 	0x08000000,
 	0x20000000,
-/* 0x0bfc: i2c_ctrl */
+/* 0x0d1c: i2c_ctrl */
 	0x0000e138,
 	0x0000e150,
 	0x0000e168,
@@ -838,18 +914,10 @@ uint32_t nvc0_pwr_data[] = {
 	0x00000000,
 	0x00000000,
 	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
 };
 
 uint32_t nvc0_pwr_code[] = {
-	0x030d0ef5,
+	0x039e0ef5,
 /* 0x0004: rd32 */
 	0x07a007f1,
 	0xd00604b6,
@@ -885,19 +953,22 @@ uint32_t nvc0_pwr_code[] = {
 	0xd4f100dd,
 	0x1bf47000,
 /* 0x007f: nsec */
-	0xf000f8f2,
+	0xf900f8f2,
+	0xf080f990,
 	0x84b62c87,
 	0x0088cf06,
-/* 0x0088: nsec_loop */
+/* 0x008c: nsec_loop */
 	0xb62c97f0,
 	0x99cf0694,
 	0x0298bb00,
 	0xf4069eb8,
-	0x00f8f11e,
-/* 0x009c: wait */
+	0x80fcf11e,
+	0x00f890fc,
+/* 0x00a4: wait */
+	0x80f990f9,
 	0xb62c87f0,
 	0x88cf0684,
-/* 0x00a5: wait_loop */
+/* 0x00b1: wait_loop */
 	0x02eeb900,
 	0xb90421f4,
 	0xadfd02da,
@@ -907,28 +978,29 @@ uint32_t nvc0_pwr_code[] = {
 	0x0099cf06,
 	0xb80298bb,
 	0x1ef4069b,
-/* 0x00c9: wait_done */
-/* 0x00cb: intr_watchdog */
-	0x9800f8df,
+/* 0x00d5: wait_done */
+	0xfc80fcdf,
+/* 0x00db: intr_watchdog */
+	0x9800f890,
 	0x96b003e9,
 	0x2a0bf400,
 	0xbb9a0a98,
 	0x1cf4029a,
 	0x01d7f00f,
-	0x025421f5,
+	0x02dd21f5,
 	0x0ef494bd,
-/* 0x00e9: intr_watchdog_next_time */
+/* 0x00f9: intr_watchdog_next_time */
 	0x9b0a9815,
 	0xf400a6b0,
 	0x9ab8090b,
 	0x061cf406,
-/* 0x00f8: intr_watchdog_next_time_set */
-/* 0x00fb: intr_watchdog_next_proc */
+/* 0x0108: intr_watchdog_next_time_set */
+/* 0x010b: intr_watchdog_next_proc */
 	0x809b0980,
 	0xe0b603e9,
 	0x68e6b158,
 	0xc61bf402,
-/* 0x010a: intr */
+/* 0x011a: intr */
 	0x00f900f8,
 	0x80f904bd,
 	0xa0f990f9,
@@ -948,13 +1020,13 @@ uint32_t nvc0_pwr_code[] = {
 	0xf40289c4,
 	0x0080230b,
 	0x58e7f09b,
-	0x98cb21f4,
+	0x98db21f4,
 	0x96b09b09,
 	0x110bf400,
 	0xb63407f0,
 	0x09d00604,
 	0x8004bd00,
-/* 0x016e: intr_skip_watchdog */
+/* 0x017e: intr_skip_watchdog */
 	0x89e49a09,
 	0x0bf40800,
 	0x8897f148,
@@ -967,22 +1039,22 @@ uint32_t nvc0_pwr_code[] = {
 	0x48e7f1c0,
 	0x53e3f14f,
 	0x00d7f054,
-	0x02b921f5,
+	0x034221f5,
 	0x07f1c0fc,
 	0x04b604c0,
 	0x000cd006,
-/* 0x01ae: intr_subintr_skip_fifo */
+/* 0x01be: intr_subintr_skip_fifo */
 	0x07f104bd,
 	0x04b60688,
 	0x0009d006,
-/* 0x01ba: intr_skip_subintr */
+/* 0x01ca: intr_skip_subintr */
 	0x89c404bd,
 	0x070bf420,
 	0xffbfa4f1,
-/* 0x01c4: intr_skip_pause */
+/* 0x01d4: intr_skip_pause */
 	0xf44089c4,
 	0xa4f1070b,
-/* 0x01ce: intr_skip_user0 */
+/* 0x01de: intr_skip_user0 */
 	0x07f0ffbf,
 	0x0604b604,
 	0xbd0008d0,
@@ -993,344 +1065,491 @@ uint32_t nvc0_pwr_code[] = {
 	0x90fca0fc,
 	0x00fc80fc,
 	0xf80032f4,
-/* 0x01f5: timer */
-	0x1032f401,
-	0xb003f898,
-	0x1cf40086,
-	0x03fe8051,
+/* 0x0205: ticks_from_ns */
+	0xf9c0f901,
+	0xcbd7f1b0,
+	0x00d3f000,
+	0x041321f5,
+	0x03e8ccec,
+	0xf400b4b0,
+	0xeeec120b,
+	0xd7f103e8,
+	0xd3f000cb,
+	0x1321f500,
+/* 0x022d: ticks_from_ns_quit */
+	0x02ceb904,
+	0xc0fcb0fc,
+/* 0x0236: ticks_from_us */
+	0xc0f900f8,
+	0xd7f1b0f9,
+	0xd3f000cb,
+	0x1321f500,
+	0x02ceb904,
+	0xf400b4b0,
+	0xe4bd050b,
+/* 0x0250: ticks_from_us_quit */
+	0xc0fcb0fc,
+/* 0x0256: ticks_to_us */
+	0xd7f100f8,
+	0xd3f000cb,
+	0xecedff00,
+/* 0x0262: timer */
+	0x90f900f8,
+	0x32f480f9,
+	0x03f89810,
+	0xf40086b0,
+	0x84bd651c,
 	0xb63807f0,
 	0x08d00604,
 	0xf004bd00,
-	0x84b60887,
+	0x84b63487,
 	0x0088cf06,
-	0xf40284f0,
-	0x87f0261b,
-	0x0684b634,
-	0xb80088cf,
-	0x0bf406e0,
-	0x06e8b809,
-/* 0x0233: timer_reset */
-	0xf01f1ef4,
-	0x04b63407,
-	0x000ed006,
-	0x0e8004bd,
-/* 0x0241: timer_enable */
-	0x0187f09a,
+	0xbb9a0998,
+	0xe9bb0298,
+	0x03fe8000,
+	0xb60887f0,
+	0x88cf0684,
+	0x0284f000,
+	0xf0261bf4,
+	0x84b63487,
+	0x0088cf06,
+	0xf406e0b8,
+	0xe8b8090b,
+	0x111cf406,
+/* 0x02b8: timer_reset */
+	0xb63407f0,
+	0x0ed00604,
+	0x8004bd00,
+/* 0x02c6: timer_enable */
+	0x87f09a0e,
+	0x3807f001,
+	0xd00604b6,
+	0x04bd0008,
+/* 0x02d4: timer_done */
+	0xfc1031f4,
+	0xf890fc80,
+/* 0x02dd: send_proc */
+	0xf980f900,
+	0x05e89890,
+	0xf004e998,
+	0x89b80486,
+	0x2a0bf406,
+	0x940398c4,
+	0x80b60488,
+	0x008ebb18,
+	0x8000fa98,
+	0x8d80008a,
+	0x028c8001,
+	0xb6038b80,
+	0x94f00190,
+	0x04e98007,
+/* 0x0317: send_done */
+	0xfc0231f4,
+	0xf880fc90,
+/* 0x031d: find */
+	0xf080f900,
+	0x31f45887,
+/* 0x0325: find_loop */
+	0x008a9801,
+	0xf406aeb8,
+	0x80b6100b,
+	0x6886b158,
+	0xf01bf402,
+/* 0x033b: find_done */
+	0xb90132f4,
+	0x80fc028e,
+/* 0x0342: send */
+	0x21f500f8,
+	0x01f4031d,
+/* 0x034b: recv */
+	0xf900f897,
+	0x9880f990,
+	0xe99805e8,
+	0x0132f404,
+	0xf40689b8,
+	0x89c43d0b,
+	0x0180b603,
+	0x800784f0,
+	0xea9805e8,
+	0xfef0f902,
+	0xf0f9018f,
+	0x9402efb9,
+	0xe9bb0499,
+	0x18e0b600,
+	0x9803eb98,
+	0xed9802ec,
+	0x00ee9801,
+	0xf0fca5f9,
+	0xf400f8fe,
+	0xf0fc0131,
+/* 0x0398: recv_done */
+	0x90fc80fc,
+/* 0x039e: init */
+	0x17f100f8,
+	0x14b60108,
+	0x0011cf06,
+	0x010911e7,
+	0xfe0814b6,
+	0x17f10014,
+	0x13f000e0,
+	0x1c07f000,
+	0xd00604b6,
+	0x04bd0001,
+	0xf0ff17f0,
+	0x04b61407,
+	0x0001d006,
+	0x17f004bd,
+	0x0015f102,
+	0x1007f008,
+	0xd00604b6,
+	0x04bd0001,
+	0x011a17f1,
+	0xfe0013f0,
+	0x31f40010,
+	0x0117f010,
 	0xb63807f0,
-	0x08d00604,
-/* 0x024f: timer_done */
-	0xf404bd00,
-	0x00f81031,
-/* 0x0254: send_proc */
-	0x90f980f9,
-	0x9805e898,
-	0x86f004e9,
-	0x0689b804,
-	0xc42a0bf4,
-	0x88940398,
-	0x1880b604,
-	0x98008ebb,
-	0x8a8000fa,
-	0x018d8000,
-	0x80028c80,
-	0x90b6038b,
-	0x0794f001,
-	0xf404e980,
-/* 0x028e: send_done */
-	0x90fc0231,
-	0x00f880fc,
-/* 0x0294: find */
-	0x87f080f9,
-	0x0131f458,
-/* 0x029c: find_loop */
-	0xb8008a98,
-	0x0bf406ae,
-	0x5880b610,
-	0x026886b1,
-	0xf4f01bf4,
-/* 0x02b2: find_done */
-	0x8eb90132,
-	0xf880fc02,
-/* 0x02b9: send */
-	0x9421f500,
-	0x9701f402,
-/* 0x02c2: recv */
-	0xe89800f8,
-	0x04e99805,
-	0xb80132f4,
-	0x0bf40689,
-	0x0389c43d,
-	0xf00180b6,
-	0xe8800784,
-	0x02ea9805,
-	0x8ffef0f9,
-	0xb9f0f901,
-	0x999402ef,
-	0x00e9bb04,
-	0x9818e0b6,
-	0xec9803eb,
-	0x01ed9802,
-	0xf900ee98,
-	0xfef0fca5,
-	0x31f400f8,
-/* 0x030b: recv_done */
-	0xf8f0fc01,
-/* 0x030d: init */
-	0x0817f100,
-	0x0614b601,
-	0xe70011cf,
-	0xb6010911,
-	0x14fe0814,
-	0xe017f100,
-	0x0013f000,
-	0xb61c07f0,
 	0x01d00604,
 	0xf004bd00,
-	0x07f0ff17,
-	0x0604b614,
-	0xbd0001d0,
-	0x0217f004,
-	0x080015f1,
-	0xb61007f0,
-	0x01d00604,
-	0xf104bd00,
-	0xf0010a17,
-	0x10fe0013,
-	0x1031f400,
-	0xf00117f0,
-	0x04b63807,
-	0x0001d006,
-	0xf7f004bd,
-/* 0x0371: init_proc */
-	0x01f19858,
-	0xf40016b0,
-	0x15f9fa0b,
-	0xf458f0b6,
-/* 0x0382: host_send */
-	0x17f1f20e,
-	0x14b604b0,
-	0x0011cf06,
-	0x04a027f1,
-	0xcf0624b6,
-	0x12b80022,
-	0x320bf406,
-	0x94071ec4,
-	0xe0b704ee,
-	0xeb980270,
-	0x02ec9803,
-	0x9801ed98,
-	0x21f500ee,
-	0x10b602b9,
-	0x0f1ec401,
-	0x04b007f1,
-	0xd00604b6,
-	0x04bd000e,
-/* 0x03cb: host_send_done */
-	0xf8ba0ef4,
-/* 0x03cd: host_recv */
-	0x4917f100,
-	0x5413f14e,
-	0x06e1b852,
-/* 0x03db: host_recv_wait */
-	0xf1aa0bf4,
-	0xb604cc17,
-	0x11cf0614,
-	0xc827f100,
-	0x0624b604,
-	0xf00022cf,
-	0x12b80816,
-	0xe60bf406,
-	0xb60723c4,
-	0x30b70434,
-	0x3b8002f0,
-	0x023c8003,
-	0x80013d80,
-	0x20b6003e,
-	0x0f24f001,
-	0x04c807f1,
+/* 0x0402: init_proc */
+	0xf19858f7,
+	0x0016b001,
+	0xf9fa0bf4,
+	0x58f0b615,
+/* 0x0413: mulu32_32_64 */
+	0xf9f20ef4,
+	0xf920f910,
+	0x9540f930,
+	0xd29510e1,
+	0xbdc4bd10,
+	0xc0edffb4,
+	0xb9301dff,
+	0x34f10234,
+	0x34b6ffff,
+	0x1045b610,
+	0xbb00c3bb,
+	0xe2ff01b4,
+	0x0234b930,
+	0xffff34f1,
+	0xb61034b6,
+	0xc3bb1045,
+	0x01b4bb00,
+	0xbb3012ff,
+	0x40fc00b3,
+	0x20fc30fc,
+	0x00f810fc,
+/* 0x0464: host_send */
+	0x04b017f1,
+	0xcf0614b6,
+	0x27f10011,
+	0x24b604a0,
+	0x0022cf06,
+	0xf40612b8,
+	0x1ec4320b,
+	0x04ee9407,
+	0x0270e0b7,
+	0x9803eb98,
+	0xed9802ec,
+	0x00ee9801,
+	0x034221f5,
+	0xc40110b6,
+	0x07f10f1e,
+	0x04b604b0,
+	0x000ed006,
+	0x0ef404bd,
+/* 0x04ad: host_send_done */
+/* 0x04af: host_recv */
+	0xf100f8ba,
+	0xf14e4917,
+	0xb8525413,
+	0x0bf406e1,
+/* 0x04bd: host_recv_wait */
+	0xcc17f1aa,
+	0x0614b604,
+	0xf10011cf,
+	0xb604c827,
+	0x22cf0624,
+	0x0816f000,
+	0xf40612b8,
+	0x23c4e60b,
+	0x0434b607,
+	0x02f030b7,
+	0x80033b80,
+	0x3d80023c,
+	0x003e8001,
+	0xf00120b6,
+	0x07f10f24,
+	0x04b604c8,
+	0x0002d006,
+	0x27f004bd,
+	0x0007f040,
 	0xd00604b6,
 	0x04bd0002,
-	0xf04027f0,
-	0x04b60007,
-	0x0002d006,
-	0x00f804bd,
-/* 0x0430: host_init */
-	0x008017f1,
-	0xf11014b6,
-	0xf1027015,
-	0xb604d007,
-	0x01d00604,
-	0xf104bd00,
-	0xb6008017,
-	0x15f11014,
-	0x07f102f0,
-	0x04b604dc,
-	0x0001d006,
-	0x17f004bd,
-	0xc407f101,
+/* 0x0512: host_init */
+	0x17f100f8,
+	0x14b60080,
+	0x7015f110,
+	0xd007f102,
 	0x0604b604,
 	0xbd0001d0,
-/* 0x046f: memx_func_enter */
-	0xf000f804,
+	0x8017f104,
+	0x1014b600,
+	0x02f015f1,
+	0x04dc07f1,
+	0xd00604b6,
+	0x04bd0001,
+	0xf10117f0,
+	0xb604c407,
+	0x01d00604,
+	0xf804bd00,
+/* 0x0551: memx_func_enter */
+	0x2067f100,
+	0x5d77f116,
+	0xff73f1f5,
+	0x026eb9ff,
+	0xb90421f4,
+	0x87fd02d8,
+	0xf960f904,
+	0xfcd0fc80,
+	0x3f21f4e0,
+	0xfffe77f1,
+	0xffff73f1,
+	0xf4026eb9,
+	0xd8b90421,
+	0x0487fd02,
+	0x80f960f9,
+	0xe0fcd0fc,
+	0xf13f21f4,
+	0xb926f067,
+	0x21f4026e,
+	0x02d8b904,
+	0xf90487fd,
+	0xfc80f960,
+	0xf4e0fcd0,
+	0x67f03f21,
+	0xe007f104,
+	0x0604b607,
+	0xbd0006d0,
+/* 0x05bd: memx_func_enter_wait */
+	0xc067f104,
+	0x0664b607,
+	0xf00066cf,
+	0x0bf40464,
+	0x2c67f0f3,
+	0xcf0664b6,
+	0x06800066,
+/* 0x05db: memx_func_leave */
+	0xf000f8f1,
+	0x64b62c67,
+	0x0066cf06,
+	0xf0f20680,
 	0x07f10467,
-	0x04b607e0,
+	0x04b607e4,
 	0x0006d006,
-/* 0x047e: memx_func_enter_wait */
+/* 0x05f6: memx_func_leave_wait */
 	0x67f104bd,
 	0x64b607c0,
 	0x0066cf06,
 	0xf40464f0,
-	0x1698f30b,
-	0x0410b600,
-/* 0x0496: memx_func_leave */
-	0x67f000f8,
-	0xe407f104,
-	0x0604b607,
-	0xbd0006d0,
-/* 0x04a5: memx_func_leave_wait */
-	0xc067f104,
-	0x0664b607,
-	0xf00066cf,
-	0x1bf40464,
-/* 0x04b7: memx_func_wr32 */
-	0x9800f8f3,
-	0x15980016,
-	0x0810b601,
-	0x50f960f9,
+	0x67f1f31b,
+	0x77f126f0,
+	0x73f00001,
+	0x026eb900,
+	0xb90421f4,
+	0x87fd02d8,
+	0xf960f905,
+	0xfcd0fc80,
+	0x3f21f4e0,
+	0x162067f1,
+	0xf4026eb9,
+	0xd8b90421,
+	0x0587fd02,
+	0x80f960f9,
 	0xe0fcd0fc,
-	0xb63f21f4,
-	0x1bf40242,
-/* 0x04d3: memx_func_wait */
-	0xf000f8e9,
-	0x84b62c87,
-	0x0088cf06,
-	0x98001e98,
-	0x1c98011d,
-	0x031b9802,
-	0xf41010b6,
-	0x00f89c21,
-/* 0x04f0: memx_func_delay */
-	0xb6001e98,
-	0x21f40410,
-/* 0x04fb: memx_exec */
-	0xf900f87f,
+	0xf13f21f4,
+	0xf00aa277,
+	0x6eb90073,
+	0x0421f402,
+	0xfd02d8b9,
+	0x60f90587,
+	0xd0fc80f9,
+	0x21f4e0fc,
+/* 0x0663: memx_func_wait_vblank */
+	0x9800f83f,
+	0x66b00016,
+	0x130bf400,
+	0xf40166b0,
+	0x0ef4060b,
+/* 0x0675: memx_func_wait_vblank_head1 */
+	0x2077f12e,
+	0x070ef400,
+/* 0x067c: memx_func_wait_vblank_head0 */
+	0x000877f1,
+/* 0x0680: memx_func_wait_vblank_0 */
+	0x07c467f1,
+	0xcf0664b6,
+	0x67fd0066,
+	0xf31bf404,
+/* 0x0690: memx_func_wait_vblank_1 */
+	0x07c467f1,
+	0xcf0664b6,
+	0x67fd0066,
+	0xf30bf404,
+/* 0x06a0: memx_func_wait_vblank_fini */
+	0xf80410b6,
+/* 0x06a5: memx_func_wr32 */
+	0x00169800,
+	0xb6011598,
+	0x60f90810,
+	0xd0fc50f9,
+	0x21f4e0fc,
+	0x0242b63f,
+	0xf8e91bf4,
+/* 0x06c1: memx_func_wait */
+	0x2c87f000,
+	0xcf0684b6,
+	0x1e980088,
+	0x011d9800,
+	0x98021c98,
+	0x10b6031b,
+	0xa421f410,
+/* 0x06de: memx_func_delay */
+	0x1e9800f8,
+	0x0410b600,
+	0xf87f21f4,
+/* 0x06e9: memx_func_train */
+/* 0x06eb: memx_exec */
+	0xf900f800,
 	0xb9d0f9e0,
 	0xb2b902c1,
-/* 0x0505: memx_exec_next */
+/* 0x06f5: memx_exec_next */
 	0x00139802,
-	0x950410b6,
-	0x30f01034,
+	0xe70410b6,
+	0xe701f034,
+	0xb601e033,
+	0x30f00132,
 	0xde35980c,
 	0x12b855f9,
-	0xec1ef406,
-	0xe0fcd0fc,
-	0x02b921f5,
-/* 0x0526: memx_info */
-	0xc7f100f8,
-	0xb7f103ac,
-	0x21f50800,
-	0x00f802b9,
-/* 0x0534: memx_recv */
+	0xe41ef406,
+	0x98f10b98,
+	0xcbbbf20c,
+	0xc4b7f102,
+	0x06b4b607,
+	0xfc00bbcf,
+	0xf5e0fcd0,
+	0xf8034221,
+/* 0x0731: memx_info */
+	0x01c67000,
+/* 0x0737: memx_info_data */
+	0xf10e0bf4,
+	0xf103ccc7,
+	0xf40800b7,
+/* 0x0742: memx_info_train */
+	0xc7f10b0e,
+	0xb7f10bcc,
+/* 0x074a: memx_info_send */
+	0x21f50100,
+	0x00f80342,
+/* 0x0750: memx_recv */
 	0xf401d6b0,
-	0xd6b0c40b,
-	0xe90bf400,
-/* 0x0542: memx_init */
+	0xd6b0980b,
+	0xd80bf400,
+/* 0x075e: memx_init */
 	0x00f800f8,
-/* 0x0544: perf_recv */
-/* 0x0546: perf_init */
+/* 0x0760: perf_recv */
+/* 0x0762: perf_init */
 	0x00f800f8,
-/* 0x0548: i2c_drive_scl */
+/* 0x0764: i2c_drive_scl */
 	0xf40036b0,
 	0x07f1110b,
 	0x04b607e0,
 	0x0001d006,
 	0x00f804bd,
-/* 0x055c: i2c_drive_scl_lo */
+/* 0x0778: i2c_drive_scl_lo */
 	0x07e407f1,
 	0xd00604b6,
 	0x04bd0001,
-/* 0x056a: i2c_drive_sda */
+/* 0x0786: i2c_drive_sda */
 	0x36b000f8,
 	0x110bf400,
 	0x07e007f1,
 	0xd00604b6,
 	0x04bd0002,
-/* 0x057e: i2c_drive_sda_lo */
+/* 0x079a: i2c_drive_sda_lo */
 	0x07f100f8,
 	0x04b607e4,
 	0x0002d006,
 	0x00f804bd,
-/* 0x058c: i2c_sense_scl */
+/* 0x07a8: i2c_sense_scl */
 	0xf10132f4,
 	0xb607c437,
 	0x33cf0634,
 	0x0431fd00,
 	0xf4060bf4,
-/* 0x05a2: i2c_sense_scl_done */
+/* 0x07be: i2c_sense_scl_done */
 	0x00f80131,
-/* 0x05a4: i2c_sense_sda */
+/* 0x07c0: i2c_sense_sda */
 	0xf10132f4,
 	0xb607c437,
 	0x33cf0634,
 	0x0432fd00,
 	0xf4060bf4,
-/* 0x05ba: i2c_sense_sda_done */
+/* 0x07d6: i2c_sense_sda_done */
 	0x00f80131,
-/* 0x05bc: i2c_raise_scl */
+/* 0x07d8: i2c_raise_scl */
 	0x47f140f9,
 	0x37f00898,
-	0x4821f501,
-/* 0x05c9: i2c_raise_scl_wait */
-	0xe8e7f105,
+	0x6421f501,
+/* 0x07e5: i2c_raise_scl_wait */
+	0xe8e7f107,
 	0x7f21f403,
-	0x058c21f5,
+	0x07a821f5,
 	0xb60901f4,
 	0x1bf40142,
-/* 0x05dd: i2c_raise_scl_done */
+/* 0x07f9: i2c_raise_scl_done */
 	0xf840fcef,
-/* 0x05e1: i2c_start */
-	0x8c21f500,
-	0x0d11f405,
-	0x05a421f5,
+/* 0x07fd: i2c_start */
+	0xa821f500,
+	0x0d11f407,
+	0x07c021f5,
 	0xf40611f4,
-/* 0x05f2: i2c_start_rep */
+/* 0x080e: i2c_start_rep */
 	0x37f0300e,
-	0x4821f500,
-	0x0137f005,
-	0x056a21f5,
+	0x6421f500,
+	0x0137f007,
+	0x078621f5,
 	0xb60076bb,
 	0x50f90465,
 	0xbb046594,
 	0x50bd0256,
 	0xfc0475fd,
-	0xbc21f550,
-	0x0464b605,
-/* 0x061f: i2c_start_send */
+	0xd821f550,
+	0x0464b607,
+/* 0x083b: i2c_start_send */
 	0xf01f11f4,
 	0x21f50037,
-	0xe7f1056a,
+	0xe7f10786,
 	0x21f41388,
 	0x0037f07f,
-	0x054821f5,
+	0x076421f5,
 	0x1388e7f1,
-/* 0x063b: i2c_start_out */
+/* 0x0857: i2c_start_out */
 	0xf87f21f4,
-/* 0x063d: i2c_stop */
+/* 0x0859: i2c_stop */
 	0x0037f000,
-	0x054821f5,
+	0x076421f5,
 	0xf50037f0,
-	0xf1056a21,
+	0xf1078621,
 	0xf403e8e7,
 	0x37f07f21,
-	0x4821f501,
-	0x88e7f105,
+	0x6421f501,
+	0x88e7f107,
 	0x7f21f413,
 	0xf50137f0,
-	0xf1056a21,
+	0xf1078621,
 	0xf41388e7,
 	0x00f87f21,
-/* 0x0670: i2c_bitw */
-	0x056a21f5,
+/* 0x088c: i2c_bitw */
+	0x078621f5,
 	0x03e8e7f1,
 	0xbb7f21f4,
 	0x65b60076,
@@ -1338,18 +1557,18 @@ uint32_t nvc0_pwr_code[] = {
 	0x56bb0465,
 	0xfd50bd02,
 	0x50fc0475,
-	0x05bc21f5,
+	0x07d821f5,
 	0xf40464b6,
 	0xe7f11811,
 	0x21f41388,
 	0x0037f07f,
-	0x054821f5,
+	0x076421f5,
 	0x1388e7f1,
-/* 0x06af: i2c_bitw_out */
+/* 0x08cb: i2c_bitw_out */
 	0xf87f21f4,
-/* 0x06b1: i2c_bitr */
+/* 0x08cd: i2c_bitr */
 	0x0137f000,
-	0x056a21f5,
+	0x078621f5,
 	0x03e8e7f1,
 	0xbb7f21f4,
 	0x65b60076,
@@ -1357,19 +1576,19 @@ uint32_t nvc0_pwr_code[] = {
 	0x56bb0465,
 	0xfd50bd02,
 	0x50fc0475,
-	0x05bc21f5,
+	0x07d821f5,
 	0xf40464b6,
 	0x21f51b11,
-	0x37f005a4,
-	0x4821f500,
-	0x88e7f105,
+	0x37f007c0,
+	0x6421f500,
+	0x88e7f107,
 	0x7f21f413,
 	0xf4013cf0,
-/* 0x06f6: i2c_bitr_done */
+/* 0x0912: i2c_bitr_done */
 	0x00f80131,
-/* 0x06f8: i2c_get_byte */
+/* 0x0914: i2c_get_byte */
 	0xf00057f0,
-/* 0x06fe: i2c_get_byte_next */
+/* 0x091a: i2c_get_byte_next */
 	0x54b60847,
 	0x0076bb01,
 	0xf90465b6,
@@ -1377,7 +1596,7 @@ uint32_t nvc0_pwr_code[] = {
 	0xbd0256bb,
 	0x0475fd50,
 	0x21f550fc,
-	0x64b606b1,
+	0x64b608cd,
 	0x2b11f404,
 	0xb60553fd,
 	0x1bf40142,
@@ -1387,12 +1606,12 @@ uint32_t nvc0_pwr_code[] = {
 	0xbb046594,
 	0x50bd0256,
 	0xfc0475fd,
-	0x7021f550,
-	0x0464b606,
-/* 0x0748: i2c_get_byte_done */
-/* 0x074a: i2c_put_byte */
+	0x8c21f550,
+	0x0464b608,
+/* 0x0964: i2c_get_byte_done */
+/* 0x0966: i2c_put_byte */
 	0x47f000f8,
-/* 0x074d: i2c_put_byte_next */
+/* 0x0969: i2c_put_byte_next */
 	0x0142b608,
 	0xbb3854ff,
 	0x65b60076,
@@ -1400,7 +1619,7 @@ uint32_t nvc0_pwr_code[] = {
 	0x56bb0465,
 	0xfd50bd02,
 	0x50fc0475,
-	0x067021f5,
+	0x088c21f5,
 	0xf40464b6,
 	0x46b03411,
 	0xd81bf400,
@@ -1409,21 +1628,21 @@ uint32_t nvc0_pwr_code[] = {
 	0xbb046594,
 	0x50bd0256,
 	0xfc0475fd,
-	0xb121f550,
-	0x0464b606,
+	0xcd21f550,
+	0x0464b608,
 	0xbb0f11f4,
 	0x36b00076,
 	0x061bf401,
-/* 0x07a3: i2c_put_byte_done */
+/* 0x09bf: i2c_put_byte_done */
 	0xf80132f4,
-/* 0x07a5: i2c_addr */
+/* 0x09c1: i2c_addr */
 	0x0076bb00,
 	0xf90465b6,
 	0x04659450,
 	0xbd0256bb,
 	0x0475fd50,
 	0x21f550fc,
-	0x64b605e1,
+	0x64b607fd,
 	0x2911f404,
 	0x012ec3e7,
 	0xfd0134b6,
@@ -1433,31 +1652,31 @@ uint32_t nvc0_pwr_code[] = {
 	0x0256bb04,
 	0x75fd50bd,
 	0xf550fc04,
-	0xb6074a21,
-/* 0x07ea: i2c_addr_done */
+	0xb6096621,
+/* 0x0a06: i2c_addr_done */
 	0x00f80464,
-/* 0x07ec: i2c_acquire_addr */
+/* 0x0a08: i2c_acquire_addr */
 	0xb6f8cec7,
 	0xe0b702e4,
-	0xee980bfc,
-/* 0x07fb: i2c_acquire */
+	0xee980d1c,
+/* 0x0a17: i2c_acquire */
 	0xf500f800,
-	0xf407ec21,
+	0xf40a0821,
 	0xd9f00421,
 	0x3f21f403,
-/* 0x080a: i2c_release */
+/* 0x0a26: i2c_release */
 	0x21f500f8,
-	0x21f407ec,
+	0x21f40a08,
 	0x03daf004,
 	0xf83f21f4,
-/* 0x0819: i2c_recv */
+/* 0x0a35: i2c_recv */
 	0x0132f400,
 	0xb6f8c1c7,
 	0x16b00214,
 	0x3a1ff528,
-	0xd413a001,
-	0x0032980b,
-	0x0bac13a0,
+	0xf413a001,
+	0x0032980c,
+	0x0ccc13a0,
 	0xf4003198,
 	0xd0f90231,
 	0xd0f9e0f9,
@@ -1469,7 +1688,7 @@ uint32_t nvc0_pwr_code[] = {
 	0x56bb0465,
 	0xfd50bd02,
 	0x50fc0475,
-	0x07fb21f5,
+	0x0a1721f5,
 	0xfc0464b6,
 	0x00d6b0d0,
 	0x00b31bf5,
@@ -1479,7 +1698,7 @@ uint32_t nvc0_pwr_code[] = {
 	0x56bb0465,
 	0xfd50bd02,
 	0x50fc0475,
-	0x07a521f5,
+	0x09c121f5,
 	0xf50464b6,
 	0xc700d011,
 	0x76bbe0c5,
@@ -1488,7 +1707,7 @@ uint32_t nvc0_pwr_code[] = {
 	0x0256bb04,
 	0x75fd50bd,
 	0xf550fc04,
-	0xb6074a21,
+	0xb6096621,
 	0x11f50464,
 	0x57f000ad,
 	0x0076bb01,
@@ -1497,7 +1716,7 @@ uint32_t nvc0_pwr_code[] = {
 	0xbd0256bb,
 	0x0475fd50,
 	0x21f550fc,
-	0x64b607a5,
+	0x64b609c1,
 	0x8a11f504,
 	0x0076bb00,
 	0xf90465b6,
@@ -1505,7 +1724,7 @@ uint32_t nvc0_pwr_code[] = {
 	0xbd0256bb,
 	0x0475fd50,
 	0x21f550fc,
-	0x64b606f8,
+	0x64b60914,
 	0x6a11f404,
 	0xbbe05bcb,
 	0x65b60076,
@@ -1513,38 +1732,38 @@ uint32_t nvc0_pwr_code[] = {
 	0x56bb0465,
 	0xfd50bd02,
 	0x50fc0475,
-	0x063d21f5,
+	0x085921f5,
 	0xb90464b6,
 	0x74bd025b,
-/* 0x091f: i2c_recv_not_rd08 */
+/* 0x0b3b: i2c_recv_not_rd08 */
 	0xb0430ef4,
 	0x1bf401d6,
 	0x0057f03d,
-	0x07a521f5,
+	0x09c121f5,
 	0xc73311f4,
 	0x21f5e0c5,
-	0x11f4074a,
+	0x11f40966,
 	0x0057f029,
-	0x07a521f5,
+	0x09c121f5,
 	0xc71f11f4,
 	0x21f5e0b5,
-	0x11f4074a,
-	0x3d21f515,
-	0xc774bd06,
+	0x11f40966,
+	0x5921f515,
+	0xc774bd08,
 	0x1bf408c5,
 	0x0232f409,
-/* 0x095f: i2c_recv_not_wr08 */
-/* 0x095f: i2c_recv_done */
+/* 0x0b7b: i2c_recv_not_wr08 */
+/* 0x0b7b: i2c_recv_done */
 	0xc7030ef4,
 	0x21f5f8ce,
-	0xe0fc080a,
+	0xe0fc0a26,
 	0x12f4d0fc,
 	0x027cb90a,
-	0x02b921f5,
-/* 0x0974: i2c_recv_exit */
-/* 0x0976: i2c_init */
+	0x034221f5,
+/* 0x0b90: i2c_recv_exit */
+/* 0x0b92: i2c_init */
 	0x00f800f8,
-/* 0x0978: test_recv */
+/* 0x0b94: test_recv */
 	0x05d817f1,
 	0xcf0614b6,
 	0x10b60011,
@@ -1553,13 +1772,13 @@ uint32_t nvc0_pwr_code[] = {
 	0xbd0001d0,
 	0x00e7f104,
 	0x4fe3f1d9,
-	0xf521f513,
-/* 0x099f: test_init */
-	0xf100f801,
+	0x6221f513,
+/* 0x0bbb: test_init */
+	0xf100f802,
 	0xf50800e7,
-	0xf801f521,
-/* 0x09a9: idle_recv */
-/* 0x09ab: idle */
+	0xf8026221,
+/* 0x0bc5: idle_recv */
+/* 0x0bc7: idle */
 	0xf400f800,
 	0x17f10031,
 	0x14b605d4,
@@ -1567,17 +1786,17 @@ uint32_t nvc0_pwr_code[] = {
 	0xf10110b6,
 	0xb605d407,
 	0x01d00604,
-/* 0x09c7: idle_loop */
+/* 0x0be3: idle_loop */
 	0xf004bd00,
 	0x32f45817,
-/* 0x09cd: idle_proc */
-/* 0x09cd: idle_proc_exec */
+/* 0x0be9: idle_proc */
+/* 0x0be9: idle_proc_exec */
 	0xb910f902,
 	0x21f5021e,
-	0x10fc02c2,
+	0x10fc034b,
 	0xf40911f4,
 	0x0ef40231,
-/* 0x09e1: idle_proc_next */
+/* 0x0bfd: idle_proc_next */
 	0x5810b6ef,
 	0xf4061fb8,
 	0x02f4e61b,
@@ -1586,4 +1805,61 @@ uint32_t nvc0_pwr_code[] = {
 	0x00000000,
 	0x00000000,
 	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
 };
diff --git a/drivers/gpu/drm/nouveau/core/subdev/pwr/fuc/nvd0.fuc b/drivers/gpu/drm/nouveau/core/subdev/pwr/fuc/nvd0.fuc
index 8a89dfe41ce1..b85443261569 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/pwr/fuc/nvd0.fuc
+++ b/drivers/gpu/drm/nouveau/core/subdev/pwr/fuc/nvd0.fuc
@@ -23,6 +23,7 @@
  */
 
 #define NVKM_PPWR_CHIPSET GF119
+#define HW_TICKS_PER_US 324
 
 //#define NVKM_FALCON_PC24
 #define NVKM_FALCON_UNSHIFTED_IO
@@ -34,6 +35,7 @@
 .section #nvd0_pwr_data
 #define INCLUDE_PROC
 #include "kernel.fuc"
+#include "arith.fuc"
 #include "host.fuc"
 #include "memx.fuc"
 #include "perf.fuc"
@@ -44,6 +46,7 @@
 
 #define INCLUDE_DATA
 #include "kernel.fuc"
+#include "arith.fuc"
 #include "host.fuc"
 #include "memx.fuc"
 #include "perf.fuc"
@@ -56,6 +59,7 @@
 .section #nvd0_pwr_code
 #define INCLUDE_CODE
 #include "kernel.fuc"
+#include "arith.fuc"
 #include "host.fuc"
 #include "memx.fuc"
 #include "perf.fuc"
diff --git a/drivers/gpu/drm/nouveau/core/subdev/pwr/fuc/nvd0.fuc.h b/drivers/gpu/drm/nouveau/core/subdev/pwr/fuc/nvd0.fuc.h
index 8d369b3faaba..7e16aab44d85 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/pwr/fuc/nvd0.fuc.h
+++ b/drivers/gpu/drm/nouveau/core/subdev/pwr/fuc/nvd0.fuc.h
@@ -24,8 +24,8 @@ uint32_t nvd0_pwr_data[] = {
 	0x00000000,
 /* 0x0058: proc_list_head */
 	0x54534f48,
-	0x000003be,
-	0x00000367,
+	0x0000049d,
+	0x00000446,
 	0x00000000,
 	0x00000000,
 	0x00000000,
@@ -46,8 +46,8 @@ uint32_t nvd0_pwr_data[] = {
 	0x00000000,
 	0x00000000,
 	0x584d454d,
-	0x000004b8,
-	0x000004aa,
+	0x0000068b,
+	0x0000067d,
 	0x00000000,
 	0x00000000,
 	0x00000000,
@@ -68,8 +68,8 @@ uint32_t nvd0_pwr_data[] = {
 	0x00000000,
 	0x00000000,
 	0x46524550,
-	0x000004bc,
-	0x000004ba,
+	0x0000068f,
+	0x0000068d,
 	0x00000000,
 	0x00000000,
 	0x00000000,
@@ -90,8 +90,8 @@ uint32_t nvd0_pwr_data[] = {
 	0x00000000,
 	0x00000000,
 	0x5f433249,
-	0x000008d7,
-	0x0000077a,
+	0x00000aaa,
+	0x0000094d,
 	0x00000000,
 	0x00000000,
 	0x00000000,
@@ -112,8 +112,8 @@ uint32_t nvd0_pwr_data[] = {
 	0x00000000,
 	0x00000000,
 	0x54534554,
-	0x000008fa,
-	0x000008d9,
+	0x00000acd,
+	0x00000aac,
 	0x00000000,
 	0x00000000,
 	0x00000000,
@@ -134,8 +134,8 @@ uint32_t nvd0_pwr_data[] = {
 	0x00000000,
 	0x00000000,
 	0x454c4449,
-	0x00000906,
-	0x00000904,
+	0x00000ad9,
+	0x00000ad7,
 	0x00000000,
 	0x00000000,
 	0x00000000,
@@ -227,24 +227,63 @@ uint32_t nvd0_pwr_data[] = {
 	0x00000000,
 	0x00000000,
 /* 0x0370: memx_func_head */
-	0x00010000,
-	0x00000000,
-	0x000003f4,
-/* 0x037c: memx_func_next */
 	0x00000001,
 	0x00000000,
-	0x00000415,
+	0x000004d3,
+/* 0x037c: memx_func_next */
 	0x00000002,
+	0x00000000,
+	0x00000554,
+	0x00000003,
 	0x00000002,
-	0x00000430,
-	0x00040003,
+	0x000005d8,
+	0x00040004,
+	0x00000000,
+	0x000005f4,
+	0x00010005,
+	0x00000000,
+	0x0000060e,
+	0x00010006,
+	0x00000000,
+	0x000005d3,
+	0x00000007,
+	0x00000000,
+	0x00000619,
+/* 0x03c4: memx_func_tail */
+/* 0x03c4: memx_ts_start */
+	0x00000000,
+/* 0x03c8: memx_ts_end */
+	0x00000000,
+/* 0x03cc: memx_data_head */
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
 	0x00000000,
-	0x0000044c,
-	0x00010004,
 	0x00000000,
-	0x00000466,
-/* 0x03ac: memx_func_tail */
-/* 0x03ac: memx_data_head */
 	0x00000000,
 	0x00000000,
 	0x00000000,
@@ -728,6 +767,8 @@ uint32_t nvd0_pwr_data[] = {
 	0x00000000,
 	0x00000000,
 	0x00000000,
+/* 0x0bcc: memx_data_tail */
+/* 0x0bcc: memx_train_head */
 	0x00000000,
 	0x00000000,
 	0x00000000,
@@ -757,8 +798,43 @@ uint32_t nvd0_pwr_data[] = {
 	0x00000000,
 	0x00000000,
 	0x00000000,
-/* 0x0bac: memx_data_tail */
-/* 0x0bac: i2c_scl_map */
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+/* 0x0ccc: memx_train_tail */
+/* 0x0ccc: i2c_scl_map */
 	0x00000400,
 	0x00000800,
 	0x00001000,
@@ -769,7 +845,7 @@ uint32_t nvd0_pwr_data[] = {
 	0x00020000,
 	0x00040000,
 	0x00080000,
-/* 0x0bd4: i2c_sda_map */
+/* 0x0cf4: i2c_sda_map */
 	0x00100000,
 	0x00200000,
 	0x00400000,
@@ -781,10 +857,66 @@ uint32_t nvd0_pwr_data[] = {
 	0x10000000,
 	0x20000000,
 	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
 };
 
 uint32_t nvd0_pwr_code[] = {
-	0x02bf0ef5,
+	0x034d0ef5,
 /* 0x0004: rd32 */
 	0x07a007f1,
 	0xbd000ed0,
@@ -814,17 +946,20 @@ uint32_t nvd0_pwr_code[] = {
 	0xd4f100dd,
 	0x1bf47000,
 /* 0x0067: nsec */
-	0xf000f8f5,
+	0xf900f8f5,
+	0xf080f990,
 	0x88cf2c87,
-/* 0x006d: nsec_loop */
+/* 0x0071: nsec_loop */
 	0x2c97f000,
 	0xbb0099cf,
 	0x9eb80298,
 	0xf41ef406,
-/* 0x007e: wait */
-	0x87f000f8,
+	0x90fc80fc,
+/* 0x0086: wait */
+	0x90f900f8,
+	0x87f080f9,
 	0x0088cf2c,
-/* 0x0084: wait_loop */
+/* 0x0090: wait_loop */
 	0xf402eeb9,
 	0xdab90421,
 	0x04adfd02,
@@ -833,28 +968,29 @@ uint32_t nvd0_pwr_code[] = {
 	0x0099cf2c,
 	0xb80298bb,
 	0x1ef4069b,
-/* 0x00a5: wait_done */
-/* 0x00a7: intr_watchdog */
-	0x9800f8e2,
+/* 0x00b1: wait_done */
+	0xfc80fce2,
+/* 0x00b7: intr_watchdog */
+	0x9800f890,
 	0x96b003e9,
 	0x2a0bf400,
 	0xbb9a0a98,
 	0x1cf4029a,
 	0x01d7f00f,
-	0x020621f5,
+	0x028c21f5,
 	0x0ef494bd,
-/* 0x00c5: intr_watchdog_next_time */
+/* 0x00d5: intr_watchdog_next_time */
 	0x9b0a9815,
 	0xf400a6b0,
 	0x9ab8090b,
 	0x061cf406,
-/* 0x00d4: intr_watchdog_next_time_set */
-/* 0x00d7: intr_watchdog_next_proc */
+/* 0x00e4: intr_watchdog_next_time_set */
+/* 0x00e7: intr_watchdog_next_proc */
 	0x809b0980,
 	0xe0b603e9,
 	0x68e6b158,
 	0xc61bf402,
-/* 0x00e6: intr */
+/* 0x00f6: intr */
 	0x00f900f8,
 	0x80f904bd,
 	0xa0f990f9,
@@ -872,12 +1008,12 @@ uint32_t nvd0_pwr_code[] = {
 	0x0bf40289,
 	0x9b008020,
 	0xf458e7f0,
-	0x0998a721,
+	0x0998b721,
 	0x0096b09b,
 	0xf00e0bf4,
 	0x09d03407,
 	0x8004bd00,
-/* 0x013e: intr_skip_watchdog */
+/* 0x014e: intr_skip_watchdog */
 	0x89e49a09,
 	0x0bf40800,
 	0x8897f13c,
@@ -889,20 +1025,20 @@ uint32_t nvd0_pwr_code[] = {
 	0xf14f48e7,
 	0xf05453e3,
 	0x21f500d7,
-	0xc0fc026b,
+	0xc0fc02f1,
 	0x04c007f1,
 	0xbd000cd0,
-/* 0x0175: intr_subintr_skip_fifo */
+/* 0x0185: intr_subintr_skip_fifo */
 	0x8807f104,
 	0x0009d006,
-/* 0x017e: intr_skip_subintr */
+/* 0x018e: intr_skip_subintr */
 	0x89c404bd,
 	0x070bf420,
 	0xffbfa4f1,
-/* 0x0188: intr_skip_pause */
+/* 0x0198: intr_skip_pause */
 	0xf44089c4,
 	0xa4f1070b,
-/* 0x0192: intr_skip_user0 */
+/* 0x01a2: intr_skip_user0 */
 	0x07f0ffbf,
 	0x0008d004,
 	0x80fc04bd,
@@ -912,189 +1048,298 @@ uint32_t nvd0_pwr_code[] = {
 	0xfca0fcb0,
 	0xfc80fc90,
 	0x0032f400,
-/* 0x01b6: timer */
-	0x32f401f8,
-	0x03f89810,
-	0xf40086b0,
-	0xfe80421c,
-	0x3807f003,
+/* 0x01c6: ticks_from_ns */
+	0xc0f901f8,
+	0xd7f1b0f9,
+	0xd3f00144,
+	0xb321f500,
+	0xe8ccec03,
+	0x00b4b003,
+	0xec120bf4,
+	0xf103e8ee,
+	0xf00144d7,
+	0x21f500d3,
+/* 0x01ee: ticks_from_ns_quit */
+	0xceb903b3,
+	0xfcb0fc02,
+/* 0x01f7: ticks_from_us */
+	0xf900f8c0,
+	0xf1b0f9c0,
+	0xf00144d7,
+	0x21f500d3,
+	0xceb903b3,
+	0x00b4b002,
+	0xbd050bf4,
+/* 0x0211: ticks_from_us_quit */
+	0xfcb0fce4,
+/* 0x0217: ticks_to_us */
+	0xf100f8c0,
+	0xf00144d7,
+	0xedff00d3,
+/* 0x0223: timer */
+	0xf900f8ec,
+	0xf480f990,
+	0xf8981032,
+	0x0086b003,
+	0xbd531cf4,
+	0x3807f084,
 	0xbd0008d0,
-	0x0887f004,
-	0xf00088cf,
-	0x1bf40284,
-	0x3487f020,
-	0xb80088cf,
-	0x0bf406e0,
-	0x06e8b809,
-/* 0x01eb: timer_reset */
-	0xf0191ef4,
-	0x0ed03407,
-	0x8004bd00,
-/* 0x01f6: timer_enable */
-	0x87f09a0e,
-	0x3807f001,
-	0xbd0008d0,
-/* 0x0201: timer_done */
-	0x1031f404,
-/* 0x0206: send_proc */
-	0x80f900f8,
-	0xe89890f9,
+	0x3487f004,
+	0x980088cf,
+	0x98bb9a09,
+	0x00e9bb02,
+	0xf003fe80,
+	0x88cf0887,
+	0x0284f000,
+	0xf0201bf4,
+	0x88cf3487,
+	0x06e0b800,
+	0xb8090bf4,
+	0x1cf406e8,
+/* 0x026d: timer_reset */
+	0x3407f00e,
+	0xbd000ed0,
+	0x9a0e8004,
+/* 0x0278: timer_enable */
+	0xf00187f0,
+	0x08d03807,
+/* 0x0283: timer_done */
+	0xf404bd00,
+	0x80fc1031,
+	0x00f890fc,
+/* 0x028c: send_proc */
+	0x90f980f9,
+	0x9805e898,
+	0x86f004e9,
+	0x0689b804,
+	0xc42a0bf4,
+	0x88940398,
+	0x1880b604,
+	0x98008ebb,
+	0x8a8000fa,
+	0x018d8000,
+	0x80028c80,
+	0x90b6038b,
+	0x0794f001,
+	0xf404e980,
+/* 0x02c6: send_done */
+	0x90fc0231,
+	0x00f880fc,
+/* 0x02cc: find */
+	0x87f080f9,
+	0x0131f458,
+/* 0x02d4: find_loop */
+	0xb8008a98,
+	0x0bf406ae,
+	0x5880b610,
+	0x026886b1,
+	0xf4f01bf4,
+/* 0x02ea: find_done */
+	0x8eb90132,
+	0xf880fc02,
+/* 0x02f1: send */
+	0xcc21f500,
+	0x9701f402,
+/* 0x02fa: recv */
+	0x90f900f8,
+	0xe89880f9,
 	0x04e99805,
-	0xb80486f0,
+	0xb80132f4,
 	0x0bf40689,
-	0x0398c42a,
-	0xb6048894,
-	0x8ebb1880,
-	0x00fa9800,
-	0x80008a80,
-	0x8c80018d,
-	0x038b8002,
-	0xf00190b6,
-	0xe9800794,
-	0x0231f404,
-/* 0x0240: send_done */
-	0x80fc90fc,
-/* 0x0246: find */
-	0x80f900f8,
-	0xf45887f0,
-/* 0x024e: find_loop */
-	0x8a980131,
-	0x06aeb800,
-	0xb6100bf4,
-	0x86b15880,
-	0x1bf40268,
-	0x0132f4f0,
-/* 0x0264: find_done */
-	0xfc028eb9,
-/* 0x026b: send */
-	0xf500f880,
-	0xf4024621,
-	0x00f89701,
-/* 0x0274: recv */
-	0x9805e898,
-	0x32f404e9,
-	0x0689b801,
-	0xc43d0bf4,
-	0x80b60389,
-	0x0784f001,
-	0x9805e880,
-	0xf0f902ea,
-	0xf9018ffe,
-	0x02efb9f0,
-	0xbb049994,
-	0xe0b600e9,
-	0x03eb9818,
-	0x9802ec98,
-	0xee9801ed,
-	0xfca5f900,
-	0x00f8fef0,
-	0xfc0131f4,
-/* 0x02bd: recv_done */
-/* 0x02bf: init */
-	0xf100f8f0,
-	0xcf010817,
-	0x11e70011,
-	0x14b60109,
-	0x0014fe08,
-	0x00e017f1,
-	0xf00013f0,
-	0x01d01c07,
-	0xf004bd00,
-	0x07f0ff17,
-	0x0001d014,
-	0x17f004bd,
-	0x0015f102,
-	0x1007f008,
-	0xbd0001d0,
-	0xe617f104,
-	0x0013f000,
-	0xf40010fe,
-	0x17f01031,
-	0x3807f001,
-	0xbd0001d0,
-	0x58f7f004,
-/* 0x0314: init_proc */
-	0xb001f198,
-	0x0bf40016,
-	0xb615f9fa,
-	0x0ef458f0,
-/* 0x0325: host_send */
-	0xb017f1f2,
-	0x0011cf04,
-	0x04a027f1,
-	0xb80022cf,
-	0x0bf40612,
-	0x071ec42f,
-	0xb704ee94,
-	0x980270e0,
+	0x0389c43d,
+	0xf00180b6,
+	0xe8800784,
+	0x02ea9805,
+	0x8ffef0f9,
+	0xb9f0f901,
+	0x999402ef,
+	0x00e9bb04,
+	0x9818e0b6,
 	0xec9803eb,
 	0x01ed9802,
-	0xf500ee98,
-	0xb6026b21,
-	0x1ec40110,
-	0xb007f10f,
-	0x000ed004,
-	0x0ef404bd,
-/* 0x0365: host_send_done */
-/* 0x0367: host_recv */
-	0xf100f8c3,
-	0xf14e4917,
-	0xb8525413,
-	0x0bf406e1,
-/* 0x0375: host_recv_wait */
-	0xcc17f1b3,
-	0x0011cf04,
-	0x04c827f1,
-	0xf00022cf,
-	0x12b80816,
-	0xec0bf406,
-	0xb60723c4,
-	0x30b70434,
-	0x3b8002f0,
-	0x023c8003,
-	0x80013d80,
-	0x20b6003e,
-	0x0f24f001,
-	0x04c807f1,
-	0xbd0002d0,
-	0x4027f004,
-	0xd00007f0,
-	0x04bd0002,
-/* 0x03be: host_init */
+	0xf900ee98,
+	0xfef0fca5,
+	0x31f400f8,
+/* 0x0347: recv_done */
+	0xfcf0fc01,
+	0xf890fc80,
+/* 0x034d: init */
+	0x0817f100,
+	0x0011cf01,
+	0x010911e7,
+	0xfe0814b6,
+	0x17f10014,
+	0x13f000e0,
+	0x1c07f000,
+	0xbd0001d0,
+	0xff17f004,
+	0xd01407f0,
+	0x04bd0001,
+	0xf10217f0,
+	0xf0080015,
+	0x01d01007,
+	0xf104bd00,
+	0xf000f617,
+	0x10fe0013,
+	0x1031f400,
+	0xf00117f0,
+	0x01d03807,
+	0xf004bd00,
+/* 0x03a2: init_proc */
+	0xf19858f7,
+	0x0016b001,
+	0xf9fa0bf4,
+	0x58f0b615,
+/* 0x03b3: mulu32_32_64 */
+	0xf9f20ef4,
+	0xf920f910,
+	0x9540f930,
+	0xd29510e1,
+	0xbdc4bd10,
+	0xc0edffb4,
+	0xb9301dff,
+	0x34f10234,
+	0x34b6ffff,
+	0x1045b610,
+	0xbb00c3bb,
+	0xe2ff01b4,
+	0x0234b930,
+	0xffff34f1,
+	0xb61034b6,
+	0xc3bb1045,
+	0x01b4bb00,
+	0xbb3012ff,
+	0x40fc00b3,
+	0x20fc30fc,
+	0x00f810fc,
+/* 0x0404: host_send */
+	0x04b017f1,
+	0xf10011cf,
+	0xcf04a027,
+	0x12b80022,
+	0x2f0bf406,
+	0x94071ec4,
+	0xe0b704ee,
+	0xeb980270,
+	0x02ec9803,
+	0x9801ed98,
+	0x21f500ee,
+	0x10b602f1,
+	0x0f1ec401,
+	0x04b007f1,
+	0xbd000ed0,
+	0xc30ef404,
+/* 0x0444: host_send_done */
+/* 0x0446: host_recv */
 	0x17f100f8,
-	0x14b60080,
-	0x7015f110,
-	0xd007f102,
-	0x0001d004,
-	0x17f104bd,
-	0x14b60080,
-	0xf015f110,
-	0xdc07f102,
-	0x0001d004,
-	0x17f004bd,
-	0xc407f101,
-	0x0001d004,
-	0x00f804bd,
-/* 0x03f4: memx_func_enter */
+	0x13f14e49,
+	0xe1b85254,
+	0xb30bf406,
+/* 0x0454: host_recv_wait */
+	0x04cc17f1,
+	0xf10011cf,
+	0xcf04c827,
+	0x16f00022,
+	0x0612b808,
+	0xc4ec0bf4,
+	0x34b60723,
+	0xf030b704,
+	0x033b8002,
+	0x80023c80,
+	0x3e80013d,
+	0x0120b600,
+	0xf10f24f0,
+	0xd004c807,
+	0x04bd0002,
+	0xf04027f0,
+	0x02d00007,
+	0xf804bd00,
+/* 0x049d: host_init */
+	0x8017f100,
+	0x1014b600,
+	0x027015f1,
+	0x04d007f1,
+	0xbd0001d0,
+	0x8017f104,
+	0x1014b600,
+	0x02f015f1,
+	0x04dc07f1,
+	0xbd0001d0,
+	0x0117f004,
+	0x04c407f1,
+	0xbd0001d0,
+/* 0x04d3: memx_func_enter */
+	0xf100f804,
+	0xf1162067,
+	0xf1f55d77,
+	0xb9ffff73,
+	0x21f4026e,
+	0x02d8b904,
+	0xf90487fd,
+	0xfc80f960,
+	0xf4e0fcd0,
+	0x77f13321,
+	0x73f1fffe,
+	0x6eb9ffff,
+	0x0421f402,
+	0xfd02d8b9,
+	0x60f90487,
+	0xd0fc80f9,
+	0x21f4e0fc,
+	0xf067f133,
+	0x026eb926,
+	0xb90421f4,
+	0x87fd02d8,
+	0xf960f904,
+	0xfcd0fc80,
+	0x3321f4e0,
 	0xf10467f0,
 	0xd007e007,
 	0x04bd0006,
-/* 0x0400: memx_func_enter_wait */
+/* 0x053c: memx_func_enter_wait */
 	0x07c067f1,
 	0xf00066cf,
 	0x0bf40464,
-	0x001698f6,
-	0xf80410b6,
-/* 0x0415: memx_func_leave */
-	0x0467f000,
+	0x2c67f0f6,
+	0x800066cf,
+	0x00f8f106,
+/* 0x0554: memx_func_leave */
+	0xcf2c67f0,
+	0x06800066,
+	0x0467f0f2,
 	0x07e407f1,
 	0xbd0006d0,
-/* 0x0421: memx_func_leave_wait */
+/* 0x0569: memx_func_leave_wait */
 	0xc067f104,
 	0x0066cf07,
 	0xf40464f0,
-	0x00f8f61b,
-/* 0x0430: memx_func_wr32 */
+	0x67f1f61b,
+	0x77f126f0,
+	0x73f00001,
+	0x026eb900,
+	0xb90421f4,
+	0x87fd02d8,
+	0xf960f905,
+	0xfcd0fc80,
+	0x3321f4e0,
+	0x162067f1,
+	0xf4026eb9,
+	0xd8b90421,
+	0x0587fd02,
+	0x80f960f9,
+	0xe0fcd0fc,
+	0xf13321f4,
+	0xf00aa277,
+	0x6eb90073,
+	0x0421f402,
+	0xfd02d8b9,
+	0x60f90587,
+	0xd0fc80f9,
+	0x21f4e0fc,
+/* 0x05d3: memx_func_wait_vblank */
+	0xb600f833,
+	0x00f80410,
+/* 0x05d8: memx_func_wr32 */
 	0x98001698,
 	0x10b60115,
 	0xf960f908,
@@ -1102,379 +1347,405 @@ uint32_t nvd0_pwr_code[] = {
 	0x3321f4e0,
 	0xf40242b6,
 	0x00f8e91b,
-/* 0x044c: memx_func_wait */
+/* 0x05f4: memx_func_wait */
 	0xcf2c87f0,
 	0x1e980088,
 	0x011d9800,
 	0x98021c98,
 	0x10b6031b,
-	0x7e21f410,
-/* 0x0466: memx_func_delay */
+	0x8621f410,
+/* 0x060e: memx_func_delay */
 	0x1e9800f8,
 	0x0410b600,
 	0xf86721f4,
-/* 0x0471: memx_exec */
-	0xf9e0f900,
-	0x02c1b9d0,
-/* 0x047b: memx_exec_next */
-	0x9802b2b9,
-	0x10b60013,
-	0x10349504,
-	0x980c30f0,
-	0x55f9de35,
-	0xf40612b8,
-	0xd0fcec1e,
-	0x21f5e0fc,
-	0x00f8026b,
-/* 0x049c: memx_info */
-	0x03acc7f1,
+/* 0x0619: memx_func_train */
+/* 0x061b: memx_exec */
+	0xf900f800,
+	0xb9d0f9e0,
+	0xb2b902c1,
+/* 0x0625: memx_exec_next */
+	0x00139802,
+	0xe70410b6,
+	0xe701f034,
+	0xb601e033,
+	0x30f00132,
+	0xde35980c,
+	0x12b855f9,
+	0xe41ef406,
+	0x98f10b98,
+	0xcbbbf20c,
+	0xc4b7f102,
+	0x00bbcf07,
+	0xe0fcd0fc,
+	0x02f121f5,
+/* 0x065e: memx_info */
+	0xc67000f8,
+	0x0e0bf401,
+/* 0x0664: memx_info_data */
+	0x03ccc7f1,
 	0x0800b7f1,
-	0x026b21f5,
-/* 0x04aa: memx_recv */
-	0xd6b000f8,
-	0xc40bf401,
-	0xf400d6b0,
-	0x00f8e90b,
-/* 0x04b8: memx_init */
-/* 0x04ba: perf_recv */
-	0x00f800f8,
-/* 0x04bc: perf_init */
-/* 0x04be: i2c_drive_scl */
-	0x36b000f8,
-	0x0e0bf400,
-	0x07e007f1,
-	0xbd0001d0,
-/* 0x04cf: i2c_drive_scl_lo */
-	0xf100f804,
-	0xd007e407,
+/* 0x066f: memx_info_train */
+	0xf10b0ef4,
+	0xf10bccc7,
+/* 0x0677: memx_info_send */
+	0xf50100b7,
+	0xf802f121,
+/* 0x067d: memx_recv */
+	0x01d6b000,
+	0xb09b0bf4,
+	0x0bf400d6,
+/* 0x068b: memx_init */
+	0xf800f8d8,
+/* 0x068d: perf_recv */
+/* 0x068f: perf_init */
+	0xf800f800,
+/* 0x0691: i2c_drive_scl */
+	0x0036b000,
+	0xf10e0bf4,
+	0xd007e007,
 	0x04bd0001,
-/* 0x04da: i2c_drive_sda */
-	0x36b000f8,
-	0x0e0bf400,
-	0x07e007f1,
-	0xbd0002d0,
-/* 0x04eb: i2c_drive_sda_lo */
-	0xf100f804,
-	0xd007e407,
+/* 0x06a2: i2c_drive_scl_lo */
+	0x07f100f8,
+	0x01d007e4,
+	0xf804bd00,
+/* 0x06ad: i2c_drive_sda */
+	0x0036b000,
+	0xf10e0bf4,
+	0xd007e007,
 	0x04bd0002,
-/* 0x04f6: i2c_sense_scl */
+/* 0x06be: i2c_drive_sda_lo */
+	0x07f100f8,
+	0x02d007e4,
+	0xf804bd00,
+/* 0x06c9: i2c_sense_scl */
+	0x0132f400,
+	0x07c437f1,
+	0xfd0033cf,
+	0x0bf40431,
+	0x0131f406,
+/* 0x06dc: i2c_sense_scl_done */
+/* 0x06de: i2c_sense_sda */
 	0x32f400f8,
 	0xc437f101,
 	0x0033cf07,
-	0xf40431fd,
+	0xf40432fd,
 	0x31f4060b,
-/* 0x0509: i2c_sense_scl_done */
-/* 0x050b: i2c_sense_sda */
-	0xf400f801,
-	0x37f10132,
-	0x33cf07c4,
-	0x0432fd00,
-	0xf4060bf4,
-/* 0x051e: i2c_sense_sda_done */
-	0x00f80131,
-/* 0x0520: i2c_raise_scl */
-	0x47f140f9,
-	0x37f00898,
-	0xbe21f501,
-/* 0x052d: i2c_raise_scl_wait */
-	0xe8e7f104,
-	0x6721f403,
-	0x04f621f5,
-	0xb60901f4,
-	0x1bf40142,
-/* 0x0541: i2c_raise_scl_done */
-	0xf840fcef,
-/* 0x0545: i2c_start */
-	0xf621f500,
-	0x0d11f404,
-	0x050b21f5,
-	0xf40611f4,
-/* 0x0556: i2c_start_rep */
-	0x37f0300e,
-	0xbe21f500,
-	0x0137f004,
-	0x04da21f5,
-	0xb60076bb,
-	0x50f90465,
-	0xbb046594,
-	0x50bd0256,
-	0xfc0475fd,
-	0x2021f550,
-	0x0464b605,
-/* 0x0583: i2c_start_send */
-	0xf01f11f4,
-	0x21f50037,
-	0xe7f104da,
-	0x21f41388,
-	0x0037f067,
-	0x04be21f5,
-	0x1388e7f1,
-/* 0x059f: i2c_start_out */
-	0xf86721f4,
-/* 0x05a1: i2c_stop */
-	0x0037f000,
-	0x04be21f5,
-	0xf50037f0,
-	0xf104da21,
-	0xf403e8e7,
-	0x37f06721,
-	0xbe21f501,
-	0x88e7f104,
-	0x6721f413,
-	0xf50137f0,
-	0xf104da21,
-	0xf41388e7,
-	0x00f86721,
-/* 0x05d4: i2c_bitw */
-	0x04da21f5,
+/* 0x06f1: i2c_sense_sda_done */
+/* 0x06f3: i2c_raise_scl */
+	0xf900f801,
+	0x9847f140,
+	0x0137f008,
+	0x069121f5,
+/* 0x0700: i2c_raise_scl_wait */
 	0x03e8e7f1,
-	0xbb6721f4,
+	0xf56721f4,
+	0xf406c921,
+	0x42b60901,
+	0xef1bf401,
+/* 0x0714: i2c_raise_scl_done */
+	0x00f840fc,
+/* 0x0718: i2c_start */
+	0x06c921f5,
+	0xf50d11f4,
+	0xf406de21,
+	0x0ef40611,
+/* 0x0729: i2c_start_rep */
+	0x0037f030,
+	0x069121f5,
+	0xf50137f0,
+	0xbb06ad21,
 	0x65b60076,
 	0x9450f904,
 	0x56bb0465,
 	0xfd50bd02,
 	0x50fc0475,
-	0x052021f5,
+	0x06f321f5,
 	0xf40464b6,
-	0xe7f11811,
-	0x21f41388,
-	0x0037f067,
-	0x04be21f5,
+/* 0x0756: i2c_start_send */
+	0x37f01f11,
+	0xad21f500,
+	0x88e7f106,
+	0x6721f413,
+	0xf50037f0,
+	0xf1069121,
+	0xf41388e7,
+/* 0x0772: i2c_start_out */
+	0x00f86721,
+/* 0x0774: i2c_stop */
+	0xf50037f0,
+	0xf0069121,
+	0x21f50037,
+	0xe7f106ad,
+	0x21f403e8,
+	0x0137f067,
+	0x069121f5,
 	0x1388e7f1,
-/* 0x0613: i2c_bitw_out */
-	0xf86721f4,
-/* 0x0615: i2c_bitr */
-	0x0137f000,
-	0x04da21f5,
-	0x03e8e7f1,
-	0xbb6721f4,
-	0x65b60076,
-	0x9450f904,
-	0x56bb0465,
-	0xfd50bd02,
-	0x50fc0475,
-	0x052021f5,
-	0xf40464b6,
-	0x21f51b11,
-	0x37f0050b,
-	0xbe21f500,
-	0x88e7f104,
+	0xf06721f4,
+	0x21f50137,
+	0xe7f106ad,
+	0x21f41388,
+/* 0x07a7: i2c_bitw */
+	0xf500f867,
+	0xf106ad21,
+	0xf403e8e7,
+	0x76bb6721,
+	0x0465b600,
+	0x659450f9,
+	0x0256bb04,
+	0x75fd50bd,
+	0xf550fc04,
+	0xb606f321,
+	0x11f40464,
+	0x88e7f118,
 	0x6721f413,
-	0xf4013cf0,
-/* 0x065a: i2c_bitr_done */
-	0x00f80131,
-/* 0x065c: i2c_get_byte */
-	0xf00057f0,
-/* 0x0662: i2c_get_byte_next */
-	0x54b60847,
-	0x0076bb01,
-	0xf90465b6,
-	0x04659450,
-	0xbd0256bb,
-	0x0475fd50,
-	0x21f550fc,
-	0x64b60615,
-	0x2b11f404,
-	0xb60553fd,
-	0x1bf40142,
-	0x0137f0d8,
+	0xf50037f0,
+	0xf1069121,
+	0xf41388e7,
+/* 0x07e6: i2c_bitw_out */
+	0x00f86721,
+/* 0x07e8: i2c_bitr */
+	0xf50137f0,
+	0xf106ad21,
+	0xf403e8e7,
+	0x76bb6721,
+	0x0465b600,
+	0x659450f9,
+	0x0256bb04,
+	0x75fd50bd,
+	0xf550fc04,
+	0xb606f321,
+	0x11f40464,
+	0xde21f51b,
+	0x0037f006,
+	0x069121f5,
+	0x1388e7f1,
+	0xf06721f4,
+	0x31f4013c,
+/* 0x082d: i2c_bitr_done */
+/* 0x082f: i2c_get_byte */
+	0xf000f801,
+	0x47f00057,
+/* 0x0835: i2c_get_byte_next */
+	0x0154b608,
 	0xb60076bb,
 	0x50f90465,
 	0xbb046594,
 	0x50bd0256,
 	0xfc0475fd,
-	0xd421f550,
-	0x0464b605,
-/* 0x06ac: i2c_get_byte_done */
-/* 0x06ae: i2c_put_byte */
-	0x47f000f8,
-/* 0x06b1: i2c_put_byte_next */
-	0x0142b608,
-	0xbb3854ff,
+	0xe821f550,
+	0x0464b607,
+	0xfd2b11f4,
+	0x42b60553,
+	0xd81bf401,
+	0xbb0137f0,
+	0x65b60076,
+	0x9450f904,
+	0x56bb0465,
+	0xfd50bd02,
+	0x50fc0475,
+	0x07a721f5,
+/* 0x087f: i2c_get_byte_done */
+	0xf80464b6,
+/* 0x0881: i2c_put_byte */
+	0x0847f000,
+/* 0x0884: i2c_put_byte_next */
+	0xff0142b6,
+	0x76bb3854,
+	0x0465b600,
+	0x659450f9,
+	0x0256bb04,
+	0x75fd50bd,
+	0xf550fc04,
+	0xb607a721,
+	0x11f40464,
+	0x0046b034,
+	0xbbd81bf4,
 	0x65b60076,
 	0x9450f904,
 	0x56bb0465,
 	0xfd50bd02,
 	0x50fc0475,
-	0x05d421f5,
+	0x07e821f5,
 	0xf40464b6,
-	0x46b03411,
-	0xd81bf400,
+	0x76bb0f11,
+	0x0136b000,
+	0xf4061bf4,
+/* 0x08da: i2c_put_byte_done */
+	0x00f80132,
+/* 0x08dc: i2c_addr */
 	0xb60076bb,
 	0x50f90465,
 	0xbb046594,
 	0x50bd0256,
 	0xfc0475fd,
-	0x1521f550,
-	0x0464b606,
-	0xbb0f11f4,
-	0x36b00076,
-	0x061bf401,
-/* 0x0707: i2c_put_byte_done */
-	0xf80132f4,
-/* 0x0709: i2c_addr */
-	0x0076bb00,
+	0x1821f550,
+	0x0464b607,
+	0xe72911f4,
+	0xb6012ec3,
+	0x53fd0134,
+	0x0076bb05,
 	0xf90465b6,
 	0x04659450,
 	0xbd0256bb,
 	0x0475fd50,
 	0x21f550fc,
-	0x64b60545,
-	0x2911f404,
-	0x012ec3e7,
-	0xfd0134b6,
-	0x76bb0553,
-	0x0465b600,
-	0x659450f9,
-	0x0256bb04,
-	0x75fd50bd,
-	0xf550fc04,
-	0xb606ae21,
-/* 0x074e: i2c_addr_done */
-	0x00f80464,
-/* 0x0750: i2c_acquire_addr */
-	0xb6f8cec7,
-	0xe0b705e4,
-	0x00f8d014,
-/* 0x075c: i2c_acquire */
-	0x075021f5,
-	0xf00421f4,
-	0x21f403d9,
-/* 0x076b: i2c_release */
-	0xf500f833,
-	0xf4075021,
-	0xdaf00421,
+	0x64b60881,
+/* 0x0921: i2c_addr_done */
+/* 0x0923: i2c_acquire_addr */
+	0xc700f804,
+	0xe4b6f8ce,
+	0x14e0b705,
+/* 0x092f: i2c_acquire */
+	0xf500f8d0,
+	0xf4092321,
+	0xd9f00421,
 	0x3321f403,
-/* 0x077a: i2c_recv */
-	0x32f400f8,
-	0xf8c1c701,
-	0xb00214b6,
-	0x1ff52816,
-	0x13a0013a,
-	0x32980bd4,
-	0xac13a000,
-	0x0031980b,
-	0xf90231f4,
-	0xf9e0f9d0,
-	0x0067f1d0,
-	0x0063f100,
-	0x01679210,
-	0xb60076bb,
-	0x50f90465,
-	0xbb046594,
-	0x50bd0256,
-	0xfc0475fd,
-	0x5c21f550,
-	0x0464b607,
-	0xd6b0d0fc,
-	0xb31bf500,
-	0x0057f000,
-	0xb60076bb,
-	0x50f90465,
-	0xbb046594,
-	0x50bd0256,
-	0xfc0475fd,
-	0x0921f550,
-	0x0464b607,
-	0x00d011f5,
-	0xbbe0c5c7,
+/* 0x093e: i2c_release */
+	0x21f500f8,
+	0x21f40923,
+	0x03daf004,
+	0xf83321f4,
+/* 0x094d: i2c_recv */
+	0x0132f400,
+	0xb6f8c1c7,
+	0x16b00214,
+	0x3a1ff528,
+	0xf413a001,
+	0x0032980c,
+	0x0ccc13a0,
+	0xf4003198,
+	0xd0f90231,
+	0xd0f9e0f9,
+	0x000067f1,
+	0x100063f1,
+	0xbb016792,
+	0x65b60076,
+	0x9450f904,
+	0x56bb0465,
+	0xfd50bd02,
+	0x50fc0475,
+	0x092f21f5,
+	0xfc0464b6,
+	0x00d6b0d0,
+	0x00b31bf5,
+	0xbb0057f0,
 	0x65b60076,
 	0x9450f904,
 	0x56bb0465,
 	0xfd50bd02,
 	0x50fc0475,
-	0x06ae21f5,
+	0x08dc21f5,
 	0xf50464b6,
-	0xf000ad11,
-	0x76bb0157,
+	0xc700d011,
+	0x76bbe0c5,
 	0x0465b600,
 	0x659450f9,
 	0x0256bb04,
 	0x75fd50bd,
 	0xf550fc04,
-	0xb6070921,
+	0xb6088121,
 	0x11f50464,
-	0x76bb008a,
-	0x0465b600,
-	0x659450f9,
-	0x0256bb04,
-	0x75fd50bd,
-	0xf550fc04,
-	0xb6065c21,
-	0x11f40464,
-	0xe05bcb6a,
-	0xb60076bb,
-	0x50f90465,
-	0xbb046594,
-	0x50bd0256,
-	0xfc0475fd,
-	0xa121f550,
-	0x0464b605,
-	0xbd025bb9,
-	0x430ef474,
-/* 0x0880: i2c_recv_not_rd08 */
-	0xf401d6b0,
-	0x57f03d1b,
-	0x0921f500,
-	0x3311f407,
-	0xf5e0c5c7,
-	0xf406ae21,
-	0x57f02911,
-	0x0921f500,
-	0x1f11f407,
-	0xf5e0b5c7,
-	0xf406ae21,
-	0x21f51511,
-	0x74bd05a1,
-	0xf408c5c7,
-	0x32f4091b,
-	0x030ef402,
-/* 0x08c0: i2c_recv_not_wr08 */
-/* 0x08c0: i2c_recv_done */
-	0xf5f8cec7,
-	0xfc076b21,
-	0xf4d0fce0,
-	0x7cb90a12,
-	0x6b21f502,
-/* 0x08d5: i2c_recv_exit */
-/* 0x08d7: i2c_init */
+	0x57f000ad,
+	0x0076bb01,
+	0xf90465b6,
+	0x04659450,
+	0xbd0256bb,
+	0x0475fd50,
+	0x21f550fc,
+	0x64b608dc,
+	0x8a11f504,
+	0x0076bb00,
+	0xf90465b6,
+	0x04659450,
+	0xbd0256bb,
+	0x0475fd50,
+	0x21f550fc,
+	0x64b6082f,
+	0x6a11f404,
+	0xbbe05bcb,
+	0x65b60076,
+	0x9450f904,
+	0x56bb0465,
+	0xfd50bd02,
+	0x50fc0475,
+	0x077421f5,
+	0xb90464b6,
+	0x74bd025b,
+/* 0x0a53: i2c_recv_not_rd08 */
+	0xb0430ef4,
+	0x1bf401d6,
+	0x0057f03d,
+	0x08dc21f5,
+	0xc73311f4,
+	0x21f5e0c5,
+	0x11f40881,
+	0x0057f029,
+	0x08dc21f5,
+	0xc71f11f4,
+	0x21f5e0b5,
+	0x11f40881,
+	0x7421f515,
+	0xc774bd07,
+	0x1bf408c5,
+	0x0232f409,
+/* 0x0a93: i2c_recv_not_wr08 */
+/* 0x0a93: i2c_recv_done */
+	0xc7030ef4,
+	0x21f5f8ce,
+	0xe0fc093e,
+	0x12f4d0fc,
+	0x027cb90a,
+	0x02f121f5,
+/* 0x0aa8: i2c_recv_exit */
+/* 0x0aaa: i2c_init */
+	0x00f800f8,
+/* 0x0aac: test_recv */
+	0x05d817f1,
+	0xb60011cf,
+	0x07f10110,
+	0x01d005d8,
+	0xf104bd00,
+	0xf1d900e7,
+	0xf5134fe3,
+	0xf8022321,
+/* 0x0acd: test_init */
+	0x00e7f100,
+	0x2321f508,
+/* 0x0ad7: idle_recv */
 	0xf800f802,
-/* 0x08d9: test_recv */
-	0xd817f100,
-	0x0011cf05,
-	0xf10110b6,
-	0xd005d807,
-	0x04bd0001,
-	0xd900e7f1,
-	0x134fe3f1,
-	0x01b621f5,
-/* 0x08fa: test_init */
-	0xe7f100f8,
-	0x21f50800,
-	0x00f801b6,
-/* 0x0904: idle_recv */
-/* 0x0906: idle */
-	0x31f400f8,
-	0xd417f100,
-	0x0011cf05,
-	0xf10110b6,
-	0xd005d407,
-	0x04bd0001,
-/* 0x091c: idle_loop */
-	0xf45817f0,
-/* 0x0922: idle_proc */
-/* 0x0922: idle_proc_exec */
-	0x10f90232,
-	0xf5021eb9,
-	0xfc027421,
-	0x0911f410,
-	0xf40231f4,
-/* 0x0936: idle_proc_next */
-	0x10b6ef0e,
-	0x061fb858,
-	0xf4e61bf4,
-	0x28f4dd02,
-	0xc10ef400,
+/* 0x0ad9: idle */
+	0x0031f400,
+	0x05d417f1,
+	0xb60011cf,
+	0x07f10110,
+	0x01d005d4,
+/* 0x0aef: idle_loop */
+	0xf004bd00,
+	0x32f45817,
+/* 0x0af5: idle_proc */
+/* 0x0af5: idle_proc_exec */
+	0xb910f902,
+	0x21f5021e,
+	0x10fc02fa,
+	0xf40911f4,
+	0x0ef40231,
+/* 0x0b09: idle_proc_next */
+	0x5810b6ef,
+	0xf4061fb8,
+	0x02f4e61b,
+	0x0028f4dd,
+	0x00c10ef4,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
 	0x00000000,
 	0x00000000,
 	0x00000000,
diff --git a/drivers/gpu/drm/nouveau/core/subdev/pwr/fuc/os.h b/drivers/gpu/drm/nouveau/core/subdev/pwr/fuc/os.h
index 574acfa44c8c..c8b06cb77e72 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/pwr/fuc/os.h
+++ b/drivers/gpu/drm/nouveau/core/subdev/pwr/fuc/os.h
@@ -18,12 +18,18 @@
 #define MEMX_MSG_INFO 0
 #define MEMX_MSG_EXEC 1
 
+/* MEMX: info types */
+#define MEMX_INFO_DATA  0
+#define MEMX_INFO_TRAIN 1
+
 /* MEMX: script opcode definitions */
-#define MEMX_ENTER  0
-#define MEMX_LEAVE  1
-#define MEMX_WR32   2
-#define MEMX_WAIT   3
-#define MEMX_DELAY  4
+#define MEMX_ENTER  1
+#define MEMX_LEAVE  2
+#define MEMX_WR32   3
+#define MEMX_WAIT   4
+#define MEMX_DELAY  5
+#define MEMX_VBLANK 6
+#define MEMX_TRAIN  7
 
 /* I2C_: message identifiers */
 #define I2C__MSG_RD08 0
diff --git a/drivers/gpu/drm/nouveau/core/subdev/pwr/memx.c b/drivers/gpu/drm/nouveau/core/subdev/pwr/memx.c
index def6a9ac68cf..7a9299d7159f 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/pwr/memx.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/pwr/memx.c
@@ -20,10 +20,11 @@ memx_out(struct nouveau_memx *memx)
 	struct nouveau_pwr *ppwr = memx->ppwr;
 	int i;
 
-	if (memx->c.size) {
+	if (memx->c.mthd) {
 		nv_wr32(ppwr, 0x10a1c4, (memx->c.size << 16) | memx->c.mthd);
 		for (i = 0; i < memx->c.size; i++)
 			nv_wr32(ppwr, 0x10a1c4, memx->c.data[i]);
+		memx->c.mthd = 0;
 		memx->c.size = 0;
 	}
 }
@@ -32,7 +33,7 @@ static void
 memx_cmd(struct nouveau_memx *memx, u32 mthd, u32 size, u32 data[])
 {
 	if ((memx->c.size + size >= ARRAY_SIZE(memx->c.data)) ||
-	    (memx->c.size && memx->c.mthd != mthd))
+	    (memx->c.mthd && memx->c.mthd != mthd))
 		memx_out(memx);
 	memcpy(&memx->c.data[memx->c.size], data, size * sizeof(data[0]));
 	memx->c.size += size;
@@ -46,7 +47,8 @@ nouveau_memx_init(struct nouveau_pwr *ppwr, struct nouveau_memx **pmemx)
 	u32 reply[2];
 	int ret;
 
-	ret = ppwr->message(ppwr, reply, PROC_MEMX, MEMX_MSG_INFO, 0, 0);
+	ret = ppwr->message(ppwr, reply, PROC_MEMX, MEMX_MSG_INFO,
+					MEMX_INFO_DATA, 0);
 	if (ret)
 		return ret;
 
@@ -62,8 +64,7 @@ nouveau_memx_init(struct nouveau_pwr *ppwr, struct nouveau_memx **pmemx)
 		nv_wr32(ppwr, 0x10a580, 0x00000003);
 	} while (nv_rd32(ppwr, 0x10a580) != 0x00000003);
 	nv_wr32(ppwr, 0x10a1c0, 0x01000000 | memx->base);
-	nv_wr32(ppwr, 0x10a1c4, 0x00010000 | MEMX_ENTER);
-	nv_wr32(ppwr, 0x10a1c4, 0x00000000);
+
 	return 0;
 }
 
@@ -78,7 +79,6 @@ nouveau_memx_fini(struct nouveau_memx **pmemx, bool exec)
 	memx_out(memx);
 
 	/* release data segment access */
-	nv_wr32(ppwr, 0x10a1c4, 0x00000000 | MEMX_LEAVE);
 	finish = nv_rd32(ppwr, 0x10a1c0) & 0x00ffffff;
 	nv_wr32(ppwr, 0x10a580, 0x00000000);
 
@@ -88,6 +88,8 @@ nouveau_memx_fini(struct nouveau_memx **pmemx, bool exec)
 				 memx->base, finish);
 	}
 
+	nv_debug(memx->ppwr, "Exec took %uns, PPWR_IN %08x\n",
+		 reply[0], reply[1]);
 	kfree(memx);
 	return 0;
 }
@@ -105,7 +107,7 @@ nouveau_memx_wait(struct nouveau_memx *memx,
 {
 	nv_debug(memx->ppwr, "R[%06x] & 0x%08x == 0x%08x, %d us\n",
 				addr, mask, data, nsec);
-	memx_cmd(memx, MEMX_WAIT, 4, (u32[]){ addr, ~mask, data, nsec });
+	memx_cmd(memx, MEMX_WAIT, 4, (u32[]){ addr, mask, data, nsec });
 	memx_out(memx); /* fuc can't handle multiple */
 }
 
@@ -117,4 +119,83 @@ nouveau_memx_nsec(struct nouveau_memx *memx, u32 nsec)
 	memx_out(memx); /* fuc can't handle multiple */
 }
 
+void
+nouveau_memx_wait_vblank(struct nouveau_memx *memx)
+{
+	struct nouveau_pwr *ppwr = memx->ppwr;
+	u32 heads, x, y, px = 0;
+	int i, head_sync;
+
+	if (nv_device(ppwr)->chipset < 0xd0) {
+		heads = nv_rd32(ppwr, 0x610050);
+		for (i = 0; i < 2; i++) {
+			/* Heuristic: sync to head with biggest resolution */
+			if (heads & (2 << (i << 3))) {
+				x = nv_rd32(ppwr, 0x610b40 + (0x540 * i));
+				y = (x & 0xffff0000) >> 16;
+				x &= 0x0000ffff;
+				if ((x * y) > px) {
+					px = (x * y);
+					head_sync = i;
+				}
+			}
+		}
+	}
+
+	if (px == 0) {
+		nv_debug(memx->ppwr, "WAIT VBLANK !NO ACTIVE HEAD\n");
+		return;
+	}
+
+	nv_debug(memx->ppwr, "WAIT VBLANK HEAD%d\n", head_sync);
+	memx_cmd(memx, MEMX_VBLANK, 1, (u32[]){ head_sync });
+	memx_out(memx); /* fuc can't handle multiple */
+}
+
+void
+nouveau_memx_train(struct nouveau_memx *memx)
+{
+	nv_debug(memx->ppwr, "   MEM TRAIN\n");
+	memx_cmd(memx, MEMX_TRAIN, 0, NULL);
+}
+
+int
+nouveau_memx_train_result(struct nouveau_pwr *ppwr, u32 *res, int rsize)
+{
+	u32 reply[2], base, size, i;
+	int ret;
+
+	ret = ppwr->message(ppwr, reply, PROC_MEMX, MEMX_MSG_INFO,
+					MEMX_INFO_TRAIN, 0);
+	if (ret)
+		return ret;
+
+	base = reply[0];
+	size = reply[1] >> 2;
+	if (size > rsize)
+		return -ENOMEM;
+
+	/* read the packet */
+	nv_wr32(ppwr, 0x10a1c0, 0x02000000 | base);
+
+	for (i = 0; i < size; i++)
+		res[i] = nv_rd32(ppwr, 0x10a1c4);
+
+	return 0;
+}
+
+void
+nouveau_memx_block(struct nouveau_memx *memx)
+{
+	nv_debug(memx->ppwr, "   HOST BLOCKED\n");
+	memx_cmd(memx, MEMX_ENTER, 0, NULL);
+}
+
+void
+nouveau_memx_unblock(struct nouveau_memx *memx)
+{
+	nv_debug(memx->ppwr, "   HOST UNBLOCKED\n");
+	memx_cmd(memx, MEMX_LEAVE, 0, NULL);
+}
+
 #endif
diff --git a/drivers/gpu/drm/nouveau/core/subdev/therm/fan.c b/drivers/gpu/drm/nouveau/core/subdev/therm/fan.c
index 016990a8252c..3656d605168f 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/therm/fan.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/therm/fan.c
@@ -31,6 +31,8 @@
 #include <subdev/gpio.h>
 #include <subdev/timer.h>
 
+#include <subdev/bios/fan.h>
+
 static int
 nouveau_fan_update(struct nouveau_fan *fan, bool immediate, int target)
 {
@@ -275,8 +277,11 @@ nouveau_therm_fan_ctor(struct nouveau_therm *therm)
 	/* other random init... */
 	nouveau_therm_fan_set_defaults(therm);
 	nvbios_perf_fan_parse(bios, &priv->fan->perf);
-	if (nvbios_therm_fan_parse(bios, &priv->fan->bios))
-		nv_error(therm, "parsing the thermal table failed\n");
+	if (!nvbios_fan_parse(bios, &priv->fan->bios)) {
+		nv_debug(therm, "parsing the fan table failed\n");
+		if (nvbios_therm_fan_parse(bios, &priv->fan->bios))
+			nv_error(therm, "parsing both fan tables failed\n");
+	}
 	nouveau_therm_fan_safety_checks(therm);
 	return 0;
 }
diff --git a/drivers/gpu/drm/nouveau/core/subdev/therm/fanpwm.c b/drivers/gpu/drm/nouveau/core/subdev/therm/fanpwm.c
index 9a5c07340263..c629d7f2a6a4 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/therm/fanpwm.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/therm/fanpwm.c
@@ -25,6 +25,8 @@
 
 #include <core/option.h>
 #include <subdev/gpio.h>
+#include <subdev/bios.h>
+#include <subdev/bios/fan.h>
 
 #include "priv.h"
 
@@ -86,11 +88,15 @@ nouveau_fanpwm_create(struct nouveau_therm *therm, struct dcb_gpio_func *func)
 {
 	struct nouveau_device *device = nv_device(therm);
 	struct nouveau_therm_priv *tpriv = (void *)therm;
+	struct nouveau_bios *bios = nouveau_bios(therm);
 	struct nouveau_fanpwm_priv *priv;
+	struct nvbios_therm_fan fan;
 	u32 divs, duty;
 
+	nvbios_fan_parse(bios, &fan);
+
 	if (!nouveau_boolopt(device->cfgopt, "NvFanPWM", func->param) ||
-	    !therm->pwm_ctrl ||
+	    !therm->pwm_ctrl || fan.type == NVBIOS_THERM_FAN_TOGGLE ||
 	     therm->pwm_get(therm, func->line, &divs, &duty) == -ENODEV)
 		return -ENODEV;
 
diff --git a/drivers/gpu/drm/nouveau/core/subdev/therm/gm107.c b/drivers/gpu/drm/nouveau/core/subdev/therm/gm107.c
new file mode 100644
index 000000000000..668cf3322285
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/subdev/therm/gm107.c
@@ -0,0 +1,93 @@
+/*
+ * Copyright 2014 Martin Peres
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Martin Peres
+ */
+
+#include "priv.h"
+
+struct gm107_therm_priv {
+	struct nouveau_therm_priv base;
+};
+
+static int
+gm107_fan_pwm_ctrl(struct nouveau_therm *therm, int line, bool enable)
+{
+	/* nothing to do, it seems hardwired */
+	return 0;
+}
+
+static int
+gm107_fan_pwm_get(struct nouveau_therm *therm, int line, u32 *divs, u32 *duty)
+{
+	*divs = nv_rd32(therm, 0x10eb20) & 0x1fff;
+	*duty = nv_rd32(therm, 0x10eb24) & 0x1fff;
+	return 0;
+}
+
+static int
+gm107_fan_pwm_set(struct nouveau_therm *therm, int line, u32 divs, u32 duty)
+{
+	nv_mask(therm, 0x10eb10, 0x1fff, divs); /* keep the high bits */
+	nv_wr32(therm, 0x10eb14, duty | 0x80000000);
+	return 0;
+}
+
+static int
+gm107_fan_pwm_clock(struct nouveau_therm *therm, int line)
+{
+	return nv_device(therm)->crystal * 1000;
+}
+
+static int
+gm107_therm_ctor(struct nouveau_object *parent,
+		struct nouveau_object *engine,
+		struct nouveau_oclass *oclass, void *data, u32 size,
+		struct nouveau_object **pobject)
+{
+	struct gm107_therm_priv *priv;
+	int ret;
+
+	ret = nouveau_therm_create(parent, engine, oclass, &priv);
+	*pobject = nv_object(priv);
+	if (ret)
+		return ret;
+
+	priv->base.base.pwm_ctrl = gm107_fan_pwm_ctrl;
+	priv->base.base.pwm_get = gm107_fan_pwm_get;
+	priv->base.base.pwm_set = gm107_fan_pwm_set;
+	priv->base.base.pwm_clock = gm107_fan_pwm_clock;
+	priv->base.base.temp_get = nv84_temp_get;
+	priv->base.base.fan_sense = nva3_therm_fan_sense;
+	priv->base.sensor.program_alarms = nouveau_therm_program_alarms_polling;
+	return nouveau_therm_preinit(&priv->base.base);
+}
+
+struct nouveau_oclass
+gm107_therm_oclass = {
+	.handle = NV_SUBDEV(THERM, 0x117),
+	.ofuncs = &(struct nouveau_ofuncs) {
+		.ctor = gm107_therm_ctor,
+		.dtor = _nouveau_therm_dtor,
+		.init = nvd0_therm_init,
+		.fini = nv84_therm_fini,
+	},
+};
diff --git a/drivers/gpu/drm/nouveau/core/subdev/therm/nv84.c b/drivers/gpu/drm/nouveau/core/subdev/therm/nv84.c
index 1d15c52fad0c..14e2e09bfc24 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/therm/nv84.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/therm/nv84.c
@@ -24,6 +24,7 @@
  */
 
 #include "priv.h"
+#include <subdev/fuse.h>
 
 struct nv84_therm_priv {
 	struct nouveau_therm_priv base;
@@ -32,7 +33,25 @@ struct nv84_therm_priv {
 int
 nv84_temp_get(struct nouveau_therm *therm)
 {
-	return nv_rd32(therm, 0x20400);
+	struct nouveau_fuse *fuse = nouveau_fuse(therm);
+
+	if (nv_ro32(fuse, 0x1a8) == 1)
+		return nv_rd32(therm, 0x20400);
+	else
+		return -ENODEV;
+}
+
+void
+nv84_sensor_setup(struct nouveau_therm *therm)
+{
+	struct nouveau_fuse *fuse = nouveau_fuse(therm);
+
+	/* enable temperature reading for cards with insane defaults */
+	if (nv_ro32(fuse, 0x1a8) == 1) {
+		nv_mask(therm, 0x20008, 0x80008000, 0x80000000);
+		nv_mask(therm, 0x2000c, 0x80000003, 0x00000000);
+		mdelay(20); /* wait for the temperature to stabilize */
+	}
 }
 
 static void
@@ -171,6 +190,21 @@ nv84_therm_intr(struct nouveau_subdev *subdev)
 }
 
 static int
+nv84_therm_init(struct nouveau_object *object)
+{
+	struct nv84_therm_priv *priv = (void *)object;
+	int ret;
+
+	ret = nouveau_therm_init(&priv->base.base);
+	if (ret)
+		return ret;
+
+	nv84_sensor_setup(&priv->base.base);
+
+	return 0;
+}
+
+static int
 nv84_therm_ctor(struct nouveau_object *parent,
 		struct nouveau_object *engine,
 		struct nouveau_oclass *oclass, void *data, u32 size,
@@ -228,7 +262,7 @@ nv84_therm_oclass = {
 	.ofuncs = &(struct nouveau_ofuncs) {
 		.ctor = nv84_therm_ctor,
 		.dtor = _nouveau_therm_dtor,
-		.init = _nouveau_therm_init,
+		.init = nv84_therm_init,
 		.fini = nv84_therm_fini,
 	},
 };
diff --git a/drivers/gpu/drm/nouveau/core/subdev/therm/nva3.c b/drivers/gpu/drm/nouveau/core/subdev/therm/nva3.c
index 0478b2e3fb1d..7893357a7e9f 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/therm/nva3.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/therm/nva3.c
@@ -51,6 +51,8 @@ nva3_therm_init(struct nouveau_object *object)
 	if (ret)
 		return ret;
 
+	nv84_sensor_setup(&priv->base.base);
+
 	/* enable fan tach, count revolutions per-second */
 	nv_mask(priv, 0x00e720, 0x00000003, 0x00000002);
 	if (tach->func != DCB_GPIO_UNUSED) {
diff --git a/drivers/gpu/drm/nouveau/core/subdev/therm/nvd0.c b/drivers/gpu/drm/nouveau/core/subdev/therm/nvd0.c
index bbf117be572f..b70f7cc649b8 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/therm/nvd0.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/therm/nvd0.c
@@ -114,7 +114,7 @@ nvd0_fan_pwm_clock(struct nouveau_therm *therm, int line)
 		return nv_device(therm)->crystal * 1000 / 10;
 }
 
-static int
+int
 nvd0_therm_init(struct nouveau_object *object)
 {
 	struct nvd0_therm_priv *priv = (void *)object;
@@ -150,6 +150,8 @@ nvd0_therm_ctor(struct nouveau_object *parent,
 	if (ret)
 		return ret;
 
+	nv84_sensor_setup(&priv->base.base);
+
 	priv->base.base.pwm_ctrl = nvd0_fan_pwm_ctrl;
 	priv->base.base.pwm_get = nvd0_fan_pwm_get;
 	priv->base.base.pwm_set = nvd0_fan_pwm_set;
diff --git a/drivers/gpu/drm/nouveau/core/subdev/therm/priv.h b/drivers/gpu/drm/nouveau/core/subdev/therm/priv.h
index 916fca5c7816..7dba8c281a0b 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/therm/priv.h
+++ b/drivers/gpu/drm/nouveau/core/subdev/therm/priv.h
@@ -145,10 +145,13 @@ int nv50_fan_pwm_get(struct nouveau_therm *, int, u32 *, u32 *);
 int nv50_fan_pwm_set(struct nouveau_therm *, int, u32, u32);
 int nv50_fan_pwm_clock(struct nouveau_therm *, int);
 int nv84_temp_get(struct nouveau_therm *therm);
+void nv84_sensor_setup(struct nouveau_therm *therm);
 int nv84_therm_fini(struct nouveau_object *object, bool suspend);
 
 int nva3_therm_fan_sense(struct nouveau_therm *);
 
+int nvd0_therm_init(struct nouveau_object *object);
+
 int nouveau_fanpwm_create(struct nouveau_therm *, struct dcb_gpio_func *);
 int nouveau_fantog_create(struct nouveau_therm *, struct dcb_gpio_func *);
 int nouveau_fannil_create(struct nouveau_therm *);
diff --git a/drivers/gpu/drm/nouveau/core/subdev/vm/base.c b/drivers/gpu/drm/nouveau/core/subdev/vm/base.c
index 7dd680ff2f6f..f75a683bd47a 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/vm/base.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/vm/base.c
@@ -296,7 +296,7 @@ nouveau_vm_get(struct nouveau_vm *vm, u64 size, u32 page_shift,
 	int ret;
 
 	mutex_lock(&nv_subdev(vmm)->mutex);
-	ret = nouveau_mm_head(&vm->mm, page_shift, msize, msize, align,
+	ret = nouveau_mm_head(&vm->mm, 0, page_shift, msize, msize, align,
 			     &vma->node);
 	if (unlikely(ret != 0)) {
 		mutex_unlock(&nv_subdev(vmm)->mutex);
diff --git a/drivers/gpu/drm/nouveau/core/subdev/volt/base.c b/drivers/gpu/drm/nouveau/core/subdev/volt/base.c
index 32794a999106..26ccd8df193f 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/volt/base.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/volt/base.c
@@ -101,6 +101,41 @@ nouveau_volt_set_id(struct nouveau_volt *volt, u8 id, int condition)
 	return ret;
 }
 
+static void nouveau_volt_parse_bios(struct nouveau_bios *bios,
+		struct nouveau_volt *volt)
+{
+	struct nvbios_volt_entry ivid;
+	struct nvbios_volt info;
+	u8  ver, hdr, cnt, len;
+	u16 data;
+	int i;
+
+	data = nvbios_volt_parse(bios, &ver, &hdr, &cnt, &len, &info);
+	if (data && info.vidmask && info.base && info.step) {
+		for (i = 0; i < info.vidmask + 1; i++) {
+			if (info.base >= info.min &&
+				info.base <= info.max) {
+				volt->vid[volt->vid_nr].uv = info.base;
+				volt->vid[volt->vid_nr].vid = i;
+				volt->vid_nr++;
+			}
+			info.base += info.step;
+		}
+		volt->vid_mask = info.vidmask;
+	} else if (data && info.vidmask) {
+		for (i = 0; i < cnt; i++) {
+			data = nvbios_volt_entry_parse(bios, i, &ver, &hdr,
+							  &ivid);
+			if (data) {
+				volt->vid[volt->vid_nr].uv = ivid.voltage;
+				volt->vid[volt->vid_nr].vid = ivid.vid;
+				volt->vid_nr++;
+			}
+		}
+		volt->vid_mask = info.vidmask;
+	}
+}
+
 int
 _nouveau_volt_init(struct nouveau_object *object)
 {
@@ -136,10 +171,6 @@ nouveau_volt_create_(struct nouveau_object *parent,
 {
 	struct nouveau_bios *bios = nouveau_bios(parent);
 	struct nouveau_volt *volt;
-	struct nvbios_volt_entry ivid;
-	struct nvbios_volt info;
-	u8  ver, hdr, cnt, len;
-	u16 data;
 	int ret, i;
 
 	ret = nouveau_subdev_create_(parent, engine, oclass, 0, "VOLT",
@@ -152,31 +183,9 @@ nouveau_volt_create_(struct nouveau_object *parent,
 	volt->set = nouveau_volt_set;
 	volt->set_id = nouveau_volt_set_id;
 
-	data = nvbios_volt_parse(bios, &ver, &hdr, &cnt, &len, &info);
-	if (data && info.vidmask && info.base && info.step) {
-		for (i = 0; i < info.vidmask + 1; i++) {
-			if (info.base >= info.min &&
-			    info.base <= info.max) {
-				volt->vid[volt->vid_nr].uv = info.base;
-				volt->vid[volt->vid_nr].vid = i;
-				volt->vid_nr++;
-			}
-			info.base += info.step;
-		}
-		volt->vid_mask = info.vidmask;
-	} else
-	if (data && info.vidmask) {
-		for (i = 0; i < cnt; i++) {
-			data = nvbios_volt_entry_parse(bios, i, &ver, &hdr,
-						      &ivid);
-			if (data) {
-				volt->vid[volt->vid_nr].uv = ivid.voltage;
-				volt->vid[volt->vid_nr].vid = ivid.vid;
-				volt->vid_nr++;
-			}
-		}
-		volt->vid_mask = info.vidmask;
-	}
+	/* Assuming the non-bios device should build the voltage table later */
+	if (bios)
+		nouveau_volt_parse_bios(bios, volt);
 
 	if (volt->vid_nr) {
 		for (i = 0; i < volt->vid_nr; i++) {
diff --git a/drivers/gpu/drm/nouveau/core/subdev/volt/gk20a.c b/drivers/gpu/drm/nouveau/core/subdev/volt/gk20a.c
new file mode 100644
index 000000000000..717368ef31ac
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/subdev/volt/gk20a.c
@@ -0,0 +1,199 @@
+/*
+ * Copyright (c) 2014, NVIDIA CORPORATION. All rights reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ */
+
+#ifdef __KERNEL__
+#include <nouveau_platform.h>
+#endif
+#include <subdev/volt.h>
+
+struct cvb_coef {
+	int c0;
+	int c1;
+	int c2;
+	int c3;
+	int c4;
+	int c5;
+};
+
+struct gk20a_volt_priv {
+	struct nouveau_volt base;
+	struct regulator *vdd;
+};
+
+const struct cvb_coef gk20a_cvb_coef[] = {
+	/* MHz,        c0,     c1,   c2,    c3,     c4,   c5 */
+	/*  72 */ { 1209886, -36468,  515,   417, -13123,  203},
+	/* 108 */ { 1130804, -27659,  296,   298, -10834,  221},
+	/* 180 */ { 1162871, -27110,  247,   238, -10681,  268},
+	/* 252 */ { 1220458, -28654,  247,   179, -10376,  298},
+	/* 324 */ { 1280953, -30204,  247,   119,  -9766,  304},
+	/* 396 */ { 1344547, -31777,  247,   119,  -8545,  292},
+	/* 468 */ { 1420168, -34227,  269,    60,  -7172,  256},
+	/* 540 */ { 1490757, -35955,  274,    60,  -5188,  197},
+	/* 612 */ { 1599112, -42583,  398,     0,  -1831,  119},
+	/* 648 */ { 1366986, -16459, -274,     0,  -3204,   72},
+	/* 684 */ { 1391884, -17078, -274,   -60,  -1526,   30},
+	/* 708 */ { 1415522, -17497, -274,   -60,   -458,    0},
+	/* 756 */ { 1464061, -18331, -274,  -119,   1831,  -72},
+	/* 804 */ { 1524225, -20064, -254,  -119,   4272, -155},
+	/* 852 */ { 1608418, -21643, -269,     0,    763,  -48},
+};
+
+/**
+ * cvb_mv = ((c2 * speedo / s_scale + c1) * speedo / s_scale + c0)
+ */
+static inline int
+gk20a_volt_get_cvb_voltage(int speedo, int s_scale,
+		const struct cvb_coef *coef)
+{
+	int mv;
+
+	mv = DIV_ROUND_CLOSEST(coef->c2 * speedo, s_scale);
+	mv = DIV_ROUND_CLOSEST((mv + coef->c1) * speedo, s_scale) + coef->c0;
+	return mv;
+}
+
+/**
+ * cvb_t_mv =
+ * ((c2 * speedo / s_scale + c1) * speedo / s_scale + c0) +
+ * ((c3 * speedo / s_scale + c4 + c5 * T / t_scale) * T / t_scale)
+ */
+static inline int
+gk20a_volt_get_cvb_t_voltage(int speedo, int temp, int s_scale, int t_scale,
+		const struct cvb_coef *coef)
+{
+	int cvb_mv, mv;
+
+	cvb_mv = gk20a_volt_get_cvb_voltage(speedo, s_scale, coef);
+
+	mv = DIV_ROUND_CLOSEST(coef->c3 * speedo, s_scale) + coef->c4 +
+		DIV_ROUND_CLOSEST(coef->c5 * temp, t_scale);
+	mv = DIV_ROUND_CLOSEST(mv * temp, t_scale) + cvb_mv;
+	return mv;
+}
+
+static int
+gk20a_volt_calc_voltage(const struct cvb_coef *coef, int speedo)
+{
+	int mv;
+
+	mv = gk20a_volt_get_cvb_t_voltage(speedo, -10, 100, 10, coef);
+	mv = DIV_ROUND_UP(mv, 1000);
+
+	return mv * 1000;
+}
+
+static int
+gk20a_volt_vid_get(struct nouveau_volt *volt)
+{
+	struct gk20a_volt_priv *priv = (void *)volt;
+	int i, uv;
+
+	uv = regulator_get_voltage(priv->vdd);
+
+	for (i = 0; i < volt->vid_nr; i++)
+		if (volt->vid[i].uv >= uv)
+			return i;
+
+	return -EINVAL;
+}
+
+static int
+gk20a_volt_vid_set(struct nouveau_volt *volt, u8 vid)
+{
+	struct gk20a_volt_priv *priv = (void *)volt;
+
+	nv_debug(volt, "set voltage as %duv\n", volt->vid[vid].uv);
+	return regulator_set_voltage(priv->vdd, volt->vid[vid].uv, 1200000);
+}
+
+static int
+gk20a_volt_set_id(struct nouveau_volt *volt, u8 id, int condition)
+{
+	struct gk20a_volt_priv *priv = (void *)volt;
+	int prev_uv = regulator_get_voltage(priv->vdd);
+	int target_uv = volt->vid[id].uv;
+	int ret;
+
+	nv_debug(volt, "prev=%d, target=%d, condition=%d\n",
+			prev_uv, target_uv, condition);
+	if (!condition ||
+		(condition < 0 && target_uv < prev_uv) ||
+		(condition > 0 && target_uv > prev_uv)) {
+		ret = gk20a_volt_vid_set(volt, volt->vid[id].vid);
+	} else {
+		ret = 0;
+	}
+
+	return ret;
+}
+
+static int
+gk20a_volt_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
+	       struct nouveau_oclass *oclass, void *data, u32 size,
+	       struct nouveau_object **pobject)
+{
+	struct gk20a_volt_priv *priv;
+	struct nouveau_volt *volt;
+	struct nouveau_platform_device *plat;
+	int i, ret, uv;
+
+	ret = nouveau_volt_create(parent, engine, oclass, &priv);
+	*pobject = nv_object(priv);
+	if (ret)
+		return ret;
+
+	volt = &priv->base;
+
+	plat = nv_device_to_platform(nv_device(parent));
+
+	uv = regulator_get_voltage(plat->gpu->vdd);
+	nv_info(priv, "The default voltage is %duV\n", uv);
+
+	priv->vdd = plat->gpu->vdd;
+	priv->base.vid_get = gk20a_volt_vid_get;
+	priv->base.vid_set = gk20a_volt_vid_set;
+	priv->base.set_id = gk20a_volt_set_id;
+
+	volt->vid_nr = ARRAY_SIZE(gk20a_cvb_coef);
+	nv_debug(priv, "%s - vid_nr = %d\n", __func__, volt->vid_nr);
+	for (i = 0; i < volt->vid_nr; i++) {
+		volt->vid[i].vid = i;
+		volt->vid[i].uv = gk20a_volt_calc_voltage(&gk20a_cvb_coef[i],
+					plat->gpu_speedo);
+		nv_debug(priv, "%2d: vid=%d, uv=%d\n", i, volt->vid[i].vid,
+					volt->vid[i].uv);
+	}
+
+	return 0;
+}
+
+struct nouveau_oclass
+gk20a_volt_oclass = {
+	.handle = NV_SUBDEV(VOLT, 0xea),
+	.ofuncs = &(struct nouveau_ofuncs) {
+		.ctor = gk20a_volt_ctor,
+		.dtor = _nouveau_volt_dtor,
+		.init = _nouveau_volt_init,
+		.fini = _nouveau_volt_fini,
+	},
+};
diff --git a/drivers/gpu/drm/nouveau/dispnv04/crtc.c b/drivers/gpu/drm/nouveau/dispnv04/crtc.c
index b90aa5c1f90a..38402ade6835 100644
--- a/drivers/gpu/drm/nouveau/dispnv04/crtc.c
+++ b/drivers/gpu/drm/nouveau/dispnv04/crtc.c
@@ -26,6 +26,7 @@
 
 #include <drm/drmP.h>
 #include <drm/drm_crtc_helper.h>
+#include <drm/drm_plane_helper.h>
 
 #include "nouveau_drm.h"
 #include "nouveau_reg.h"
@@ -613,7 +614,7 @@ nv_crtc_swap_fbs(struct drm_crtc *crtc, struct drm_framebuffer *old_fb)
 	struct nouveau_crtc *nv_crtc = nouveau_crtc(crtc);
 	int ret;
 
-	ret = nouveau_bo_pin(nvfb->nvbo, TTM_PL_FLAG_VRAM);
+	ret = nouveau_bo_pin(nvfb->nvbo, TTM_PL_FLAG_VRAM, false);
 	if (ret == 0) {
 		if (disp->image[nv_crtc->index])
 			nouveau_bo_unpin(disp->image[nv_crtc->index]);
@@ -1127,9 +1128,9 @@ nv04_crtc_create(struct drm_device *dev, int crtc_num)
 	drm_mode_crtc_set_gamma_size(&nv_crtc->base, 256);
 
 	ret = nouveau_bo_new(dev, 64*64*4, 0x100, TTM_PL_FLAG_VRAM,
-			     0, 0x0000, NULL, &nv_crtc->cursor.nvbo);
+			     0, 0x0000, NULL, NULL, &nv_crtc->cursor.nvbo);
 	if (!ret) {
-		ret = nouveau_bo_pin(nv_crtc->cursor.nvbo, TTM_PL_FLAG_VRAM);
+		ret = nouveau_bo_pin(nv_crtc->cursor.nvbo, TTM_PL_FLAG_VRAM, false);
 		if (!ret) {
 			ret = nouveau_bo_map(nv_crtc->cursor.nvbo);
 			if (ret)
diff --git a/drivers/gpu/drm/nouveau/dispnv04/overlay.c b/drivers/gpu/drm/nouveau/dispnv04/overlay.c
index b36afcbbc83f..9f2498571d09 100644
--- a/drivers/gpu/drm/nouveau/dispnv04/overlay.c
+++ b/drivers/gpu/drm/nouveau/dispnv04/overlay.c
@@ -97,7 +97,8 @@ nv10_update_plane(struct drm_plane *plane, struct drm_crtc *crtc,
 		  uint32_t src_w, uint32_t src_h)
 {
 	struct nvif_device *dev = &nouveau_drm(plane->dev)->device;
-	struct nouveau_plane *nv_plane = (struct nouveau_plane *)plane;
+	struct nouveau_plane *nv_plane =
+		container_of(plane, struct nouveau_plane, base);
 	struct nouveau_framebuffer *nv_fb = nouveau_framebuffer(fb);
 	struct nouveau_crtc *nv_crtc = nouveau_crtc(crtc);
 	struct nouveau_bo *cur = nv_plane->cur;
@@ -125,7 +126,7 @@ nv10_update_plane(struct drm_plane *plane, struct drm_crtc *crtc,
 			return -ERANGE;
 	}
 
-	ret = nouveau_bo_pin(nv_fb->nvbo, TTM_PL_FLAG_VRAM);
+	ret = nouveau_bo_pin(nv_fb->nvbo, TTM_PL_FLAG_VRAM, false);
 	if (ret)
 		return ret;
 
@@ -173,7 +174,8 @@ static int
 nv10_disable_plane(struct drm_plane *plane)
 {
 	struct nvif_device *dev = &nouveau_drm(plane->dev)->device;
-	struct nouveau_plane *nv_plane = (struct nouveau_plane *)plane;
+	struct nouveau_plane *nv_plane =
+		container_of(plane, struct nouveau_plane, base);
 
 	nvif_wr32(dev, NV_PVIDEO_STOP, 1);
 	if (nv_plane->cur) {
@@ -224,7 +226,8 @@ nv_set_property(struct drm_plane *plane,
 		struct drm_property *property,
 		uint64_t value)
 {
-	struct nouveau_plane *nv_plane = (struct nouveau_plane *)plane;
+	struct nouveau_plane *nv_plane =
+		container_of(plane, struct nouveau_plane, base);
 
 	if (property == nv_plane->props.colorkey)
 		nv_plane->colorkey = value;
@@ -344,7 +347,8 @@ nv04_update_plane(struct drm_plane *plane, struct drm_crtc *crtc,
 		  uint32_t src_w, uint32_t src_h)
 {
 	struct nvif_device *dev = &nouveau_drm(plane->dev)->device;
-	struct nouveau_plane *nv_plane = (struct nouveau_plane *)plane;
+	struct nouveau_plane *nv_plane =
+		container_of(plane, struct nouveau_plane, base);
 	struct nouveau_framebuffer *nv_fb = nouveau_framebuffer(fb);
 	struct nouveau_bo *cur = nv_plane->cur;
 	uint32_t overlay = 1;
@@ -369,7 +373,7 @@ nv04_update_plane(struct drm_plane *plane, struct drm_crtc *crtc,
 	if (crtc_w < src_w || crtc_h < src_h)
 		return -ERANGE;
 
-	ret = nouveau_bo_pin(nv_fb->nvbo, TTM_PL_FLAG_VRAM);
+	ret = nouveau_bo_pin(nv_fb->nvbo, TTM_PL_FLAG_VRAM, false);
 	if (ret)
 		return ret;
 
@@ -423,7 +427,8 @@ static int
 nv04_disable_plane(struct drm_plane *plane)
 {
 	struct nvif_device *dev = &nouveau_drm(plane->dev)->device;
-	struct nouveau_plane *nv_plane = (struct nouveau_plane *)plane;
+	struct nouveau_plane *nv_plane =
+		container_of(plane, struct nouveau_plane, base);
 
 	nvif_mask(dev, NV_PVIDEO_OVERLAY, 1, 0);
 	nvif_wr32(dev, NV_PVIDEO_OE_STATE, 0);
diff --git a/drivers/gpu/drm/nouveau/nouveau_abi16.c b/drivers/gpu/drm/nouveau/nouveau_abi16.c
index 615714c1727d..d39a15000068 100644
--- a/drivers/gpu/drm/nouveau/nouveau_abi16.c
+++ b/drivers/gpu/drm/nouveau/nouveau_abi16.c
@@ -308,7 +308,7 @@ nouveau_abi16_ioctl_channel_alloc(ABI16_IOCTL_ARGS)
 	ret = nouveau_gem_new(dev, PAGE_SIZE, 0, NOUVEAU_GEM_DOMAIN_GART,
 			      0, 0, &chan->ntfy);
 	if (ret == 0)
-		ret = nouveau_bo_pin(chan->ntfy, TTM_PL_FLAG_TT);
+		ret = nouveau_bo_pin(chan->ntfy, TTM_PL_FLAG_TT, false);
 	if (ret)
 		goto done;
 
@@ -448,7 +448,7 @@ nouveau_abi16_ioctl_notifierobj_alloc(ABI16_IOCTL_ARGS)
 	list_add(&ntfy->head, &chan->notifiers);
 	ntfy->handle = info->handle;
 
-	ret = nouveau_mm_head(&chan->heap, 1, info->size, info->size, 1,
+	ret = nouveau_mm_head(&chan->heap, 0, 1, info->size, info->size, 1,
 			      &ntfy->node);
 	if (ret)
 		goto done;
diff --git a/drivers/gpu/drm/nouveau/nouveau_bios.c b/drivers/gpu/drm/nouveau/nouveau_bios.c
index dae2c96deef8..7df6acc8bb34 100644
--- a/drivers/gpu/drm/nouveau/nouveau_bios.c
+++ b/drivers/gpu/drm/nouveau/nouveau_bios.c
@@ -1258,7 +1258,7 @@ olddcb_table(struct drm_device *dev)
 		return NULL;
 	}
 
-	if (dcb[0] >= 0x41) {
+	if (dcb[0] >= 0x42) {
 		NV_WARN(drm, "DCB version 0x%02x unknown\n", dcb[0]);
 		return NULL;
 	} else
@@ -1481,18 +1481,22 @@ parse_dcb20_entry(struct drm_device *dev, struct dcb_table *dcb,
 			entry->dpconf.link_bw = 540000;
 			break;
 		}
-		switch ((conf & 0x0f000000) >> 24) {
-		case 0xf:
-			entry->dpconf.link_nr = 4;
-			break;
-		case 0x3:
-			entry->dpconf.link_nr = 2;
-			break;
-		default:
-			entry->dpconf.link_nr = 1;
-			break;
+		entry->dpconf.link_nr = (conf & 0x0f000000) >> 24;
+		if (dcb->version < 0x41) {
+			switch (entry->dpconf.link_nr) {
+			case 0xf:
+				entry->dpconf.link_nr = 4;
+				break;
+			case 0x3:
+				entry->dpconf.link_nr = 2;
+				break;
+			default:
+				entry->dpconf.link_nr = 1;
+				break;
+			}
 		}
 		link = entry->dpconf.sor.link;
+		entry->i2c_index += NV_I2C_AUX(0);
 		break;
 	case DCB_OUTPUT_TMDS:
 		if (dcb->version >= 0x40) {
diff --git a/drivers/gpu/drm/nouveau/nouveau_bo.c b/drivers/gpu/drm/nouveau/nouveau_bo.c
index 01da508625f2..bba2960d3dfb 100644
--- a/drivers/gpu/drm/nouveau/nouveau_bo.c
+++ b/drivers/gpu/drm/nouveau/nouveau_bo.c
@@ -88,13 +88,13 @@ nv10_bo_get_tile_region(struct drm_device *dev, int i)
 
 static void
 nv10_bo_put_tile_region(struct drm_device *dev, struct nouveau_drm_tile *tile,
-			struct nouveau_fence *fence)
+			struct fence *fence)
 {
 	struct nouveau_drm *drm = nouveau_drm(dev);
 
 	if (tile) {
 		spin_lock(&drm->tile.lock);
-		tile->fence = nouveau_fence_ref(fence);
+		tile->fence = (struct nouveau_fence *)fence_get(fence);
 		tile->used = false;
 		spin_unlock(&drm->tile.lock);
 	}
@@ -181,7 +181,7 @@ nouveau_bo_fixup_align(struct nouveau_bo *nvbo, u32 flags,
 int
 nouveau_bo_new(struct drm_device *dev, int size, int align,
 	       uint32_t flags, uint32_t tile_mode, uint32_t tile_flags,
-	       struct sg_table *sg,
+	       struct sg_table *sg, struct reservation_object *robj,
 	       struct nouveau_bo **pnvbo)
 {
 	struct nouveau_drm *drm = nouveau_drm(dev);
@@ -214,6 +214,9 @@ nouveau_bo_new(struct drm_device *dev, int size, int align,
 	nvbo->tile_flags = tile_flags;
 	nvbo->bo.bdev = &drm->ttm.bdev;
 
+	if (!nv_device_is_cpu_coherent(nvkm_device(&drm->device)))
+		nvbo->force_coherent = flags & TTM_PL_FLAG_UNCACHED;
+
 	nvbo->page_shift = 12;
 	if (drm->client.vm) {
 		if (!(flags & TTM_PL_FLAG_TT) && size > 256 * 1024)
@@ -230,7 +233,7 @@ nouveau_bo_new(struct drm_device *dev, int size, int align,
 	ret = ttm_bo_init(&drm->ttm.bdev, &nvbo->bo, size,
 			  type, &nvbo->placement,
 			  align >> PAGE_SHIFT, false, NULL, acc_size, sg,
-			  nouveau_bo_del_ttm);
+			  robj, nouveau_bo_del_ttm);
 	if (ret) {
 		/* ttm will call nouveau_bo_del_ttm if it fails.. */
 		return ret;
@@ -241,16 +244,16 @@ nouveau_bo_new(struct drm_device *dev, int size, int align,
 }
 
 static void
-set_placement_list(uint32_t *pl, unsigned *n, uint32_t type, uint32_t flags)
+set_placement_list(struct ttm_place *pl, unsigned *n, uint32_t type, uint32_t flags)
 {
 	*n = 0;
 
 	if (type & TTM_PL_FLAG_VRAM)
-		pl[(*n)++] = TTM_PL_FLAG_VRAM | flags;
+		pl[(*n)++].flags = TTM_PL_FLAG_VRAM | flags;
 	if (type & TTM_PL_FLAG_TT)
-		pl[(*n)++] = TTM_PL_FLAG_TT | flags;
+		pl[(*n)++].flags = TTM_PL_FLAG_TT | flags;
 	if (type & TTM_PL_FLAG_SYSTEM)
-		pl[(*n)++] = TTM_PL_FLAG_SYSTEM | flags;
+		pl[(*n)++].flags = TTM_PL_FLAG_SYSTEM | flags;
 }
 
 static void
@@ -258,6 +261,7 @@ set_placement_range(struct nouveau_bo *nvbo, uint32_t type)
 {
 	struct nouveau_drm *drm = nouveau_bdev(nvbo->bo.bdev);
 	u32 vram_pages = drm->device.info.ram_size >> PAGE_SHIFT;
+	unsigned i, fpfn, lpfn;
 
 	if (drm->device.info.family == NV_DEVICE_INFO_V0_CELSIUS &&
 	    nvbo->tile_mode && (type & TTM_PL_FLAG_VRAM) &&
@@ -269,11 +273,19 @@ set_placement_range(struct nouveau_bo *nvbo, uint32_t type)
 		 * at the same time.
 		 */
 		if (nvbo->tile_flags & NOUVEAU_GEM_TILE_ZETA) {
-			nvbo->placement.fpfn = vram_pages / 2;
-			nvbo->placement.lpfn = ~0;
+			fpfn = vram_pages / 2;
+			lpfn = ~0;
 		} else {
-			nvbo->placement.fpfn = 0;
-			nvbo->placement.lpfn = vram_pages / 2;
+			fpfn = 0;
+			lpfn = vram_pages / 2;
+		}
+		for (i = 0; i < nvbo->placement.num_placement; ++i) {
+			nvbo->placements[i].fpfn = fpfn;
+			nvbo->placements[i].lpfn = lpfn;
+		}
+		for (i = 0; i < nvbo->placement.num_busy_placement; ++i) {
+			nvbo->busy_placements[i].fpfn = fpfn;
+			nvbo->busy_placements[i].lpfn = lpfn;
 		}
 	}
 }
@@ -282,8 +294,9 @@ void
 nouveau_bo_placement_set(struct nouveau_bo *nvbo, uint32_t type, uint32_t busy)
 {
 	struct ttm_placement *pl = &nvbo->placement;
-	uint32_t flags = TTM_PL_MASK_CACHING |
-		(nvbo->pin_refcnt ? TTM_PL_FLAG_NO_EVICT : 0);
+	uint32_t flags = (nvbo->force_coherent ? TTM_PL_FLAG_UNCACHED :
+						 TTM_PL_MASK_CACHING) |
+			 (nvbo->pin_refcnt ? TTM_PL_FLAG_NO_EVICT : 0);
 
 	pl->placement = nvbo->placements;
 	set_placement_list(nvbo->placements, &pl->num_placement,
@@ -297,42 +310,75 @@ nouveau_bo_placement_set(struct nouveau_bo *nvbo, uint32_t type, uint32_t busy)
 }
 
 int
-nouveau_bo_pin(struct nouveau_bo *nvbo, uint32_t memtype)
+nouveau_bo_pin(struct nouveau_bo *nvbo, uint32_t memtype, bool contig)
 {
 	struct nouveau_drm *drm = nouveau_bdev(nvbo->bo.bdev);
 	struct ttm_buffer_object *bo = &nvbo->bo;
+	bool force = false, evict = false;
 	int ret;
 
 	ret = ttm_bo_reserve(bo, false, false, false, NULL);
 	if (ret)
-		goto out;
+		return ret;
 
-	if (nvbo->pin_refcnt && !(memtype & (1 << bo->mem.mem_type))) {
-		NV_ERROR(drm, "bo %p pinned elsewhere: 0x%08x vs 0x%08x\n", bo,
-			 1 << bo->mem.mem_type, memtype);
-		ret = -EINVAL;
-		goto out;
+	if (drm->device.info.family >= NV_DEVICE_INFO_V0_TESLA &&
+	    memtype == TTM_PL_FLAG_VRAM && contig) {
+		if (nvbo->tile_flags & NOUVEAU_GEM_TILE_NONCONTIG) {
+			if (bo->mem.mem_type == TTM_PL_VRAM) {
+				struct nouveau_mem *mem = bo->mem.mm_node;
+				if (!list_is_singular(&mem->regions))
+					evict = true;
+			}
+			nvbo->tile_flags &= ~NOUVEAU_GEM_TILE_NONCONTIG;
+			force = true;
+		}
 	}
 
-	if (nvbo->pin_refcnt++)
+	if (nvbo->pin_refcnt) {
+		if (!(memtype & (1 << bo->mem.mem_type)) || evict) {
+			NV_ERROR(drm, "bo %p pinned elsewhere: "
+				      "0x%08x vs 0x%08x\n", bo,
+				 1 << bo->mem.mem_type, memtype);
+			ret = -EBUSY;
+		}
+		nvbo->pin_refcnt++;
 		goto out;
+	}
+
+	if (evict) {
+		nouveau_bo_placement_set(nvbo, TTM_PL_FLAG_TT, 0);
+		ret = nouveau_bo_validate(nvbo, false, false);
+		if (ret)
+			goto out;
+	}
 
+	nvbo->pin_refcnt++;
 	nouveau_bo_placement_set(nvbo, memtype, 0);
 
+	/* drop pin_refcnt temporarily, so we don't trip the assertion
+	 * in nouveau_bo_move() that makes sure we're not trying to
+	 * move a pinned buffer
+	 */
+	nvbo->pin_refcnt--;
 	ret = nouveau_bo_validate(nvbo, false, false);
-	if (ret == 0) {
-		switch (bo->mem.mem_type) {
-		case TTM_PL_VRAM:
-			drm->gem.vram_available -= bo->mem.size;
-			break;
-		case TTM_PL_TT:
-			drm->gem.gart_available -= bo->mem.size;
-			break;
-		default:
-			break;
-		}
+	if (ret)
+		goto out;
+	nvbo->pin_refcnt++;
+
+	switch (bo->mem.mem_type) {
+	case TTM_PL_VRAM:
+		drm->gem.vram_available -= bo->mem.size;
+		break;
+	case TTM_PL_TT:
+		drm->gem.gart_available -= bo->mem.size;
+		break;
+	default:
+		break;
 	}
+
 out:
+	if (force && ret)
+		nvbo->tile_flags |= NOUVEAU_GEM_TILE_NONCONTIG;
 	ttm_bo_unreserve(bo);
 	return ret;
 }
@@ -383,7 +429,14 @@ nouveau_bo_map(struct nouveau_bo *nvbo)
 	if (ret)
 		return ret;
 
-	ret = ttm_bo_kmap(&nvbo->bo, 0, nvbo->bo.mem.num_pages, &nvbo->kmap);
+	/*
+	 * TTM buffers allocated using the DMA API already have a mapping, let's
+	 * use it instead.
+	 */
+	if (!nvbo->force_coherent)
+		ret = ttm_bo_kmap(&nvbo->bo, 0, nvbo->bo.mem.num_pages,
+				  &nvbo->kmap);
+
 	ttm_bo_unreserve(&nvbo->bo);
 	return ret;
 }
@@ -391,10 +444,57 @@ nouveau_bo_map(struct nouveau_bo *nvbo)
 void
 nouveau_bo_unmap(struct nouveau_bo *nvbo)
 {
-	if (nvbo)
+	if (!nvbo)
+		return;
+
+	/*
+	 * TTM buffers allocated using the DMA API already had a coherent
+	 * mapping which we used, no need to unmap.
+	 */
+	if (!nvbo->force_coherent)
 		ttm_bo_kunmap(&nvbo->kmap);
 }
 
+void
+nouveau_bo_sync_for_device(struct nouveau_bo *nvbo)
+{
+	struct nouveau_drm *drm = nouveau_bdev(nvbo->bo.bdev);
+	struct nouveau_device *device = nvkm_device(&drm->device);
+	struct ttm_dma_tt *ttm_dma = (struct ttm_dma_tt *)nvbo->bo.ttm;
+	int i;
+
+	if (!ttm_dma)
+		return;
+
+	/* Don't waste time looping if the object is coherent */
+	if (nvbo->force_coherent)
+		return;
+
+	for (i = 0; i < ttm_dma->ttm.num_pages; i++)
+		dma_sync_single_for_device(nv_device_base(device),
+			ttm_dma->dma_address[i], PAGE_SIZE, DMA_TO_DEVICE);
+}
+
+void
+nouveau_bo_sync_for_cpu(struct nouveau_bo *nvbo)
+{
+	struct nouveau_drm *drm = nouveau_bdev(nvbo->bo.bdev);
+	struct nouveau_device *device = nvkm_device(&drm->device);
+	struct ttm_dma_tt *ttm_dma = (struct ttm_dma_tt *)nvbo->bo.ttm;
+	int i;
+
+	if (!ttm_dma)
+		return;
+
+	/* Don't waste time looping if the object is coherent */
+	if (nvbo->force_coherent)
+		return;
+
+	for (i = 0; i < ttm_dma->ttm.num_pages; i++)
+		dma_sync_single_for_cpu(nv_device_base(device),
+			ttm_dma->dma_address[i], PAGE_SIZE, DMA_FROM_DEVICE);
+}
+
 int
 nouveau_bo_validate(struct nouveau_bo *nvbo, bool interruptible,
 		    bool no_wait_gpu)
@@ -406,15 +506,41 @@ nouveau_bo_validate(struct nouveau_bo *nvbo, bool interruptible,
 	if (ret)
 		return ret;
 
+	nouveau_bo_sync_for_device(nvbo);
+
 	return 0;
 }
 
+static inline void *
+_nouveau_bo_mem_index(struct nouveau_bo *nvbo, unsigned index, void *mem, u8 sz)
+{
+	struct ttm_dma_tt *dma_tt;
+	u8 *m = mem;
+
+	index *= sz;
+
+	if (m) {
+		/* kmap'd address, return the corresponding offset */
+		m += index;
+	} else {
+		/* DMA-API mapping, lookup the right address */
+		dma_tt = (struct ttm_dma_tt *)nvbo->bo.ttm;
+		m = dma_tt->cpu_address[index / PAGE_SIZE];
+		m += index % PAGE_SIZE;
+	}
+
+	return m;
+}
+#define nouveau_bo_mem_index(o, i, m) _nouveau_bo_mem_index(o, i, m, sizeof(*m))
+
 u16
 nouveau_bo_rd16(struct nouveau_bo *nvbo, unsigned index)
 {
 	bool is_iomem;
 	u16 *mem = ttm_kmap_obj_virtual(&nvbo->kmap, &is_iomem);
-	mem = &mem[index];
+
+	mem = nouveau_bo_mem_index(nvbo, index, mem);
+
 	if (is_iomem)
 		return ioread16_native((void __force __iomem *)mem);
 	else
@@ -426,7 +552,9 @@ nouveau_bo_wr16(struct nouveau_bo *nvbo, unsigned index, u16 val)
 {
 	bool is_iomem;
 	u16 *mem = ttm_kmap_obj_virtual(&nvbo->kmap, &is_iomem);
-	mem = &mem[index];
+
+	mem = nouveau_bo_mem_index(nvbo, index, mem);
+
 	if (is_iomem)
 		iowrite16_native(val, (void __force __iomem *)mem);
 	else
@@ -438,7 +566,9 @@ nouveau_bo_rd32(struct nouveau_bo *nvbo, unsigned index)
 {
 	bool is_iomem;
 	u32 *mem = ttm_kmap_obj_virtual(&nvbo->kmap, &is_iomem);
-	mem = &mem[index];
+
+	mem = nouveau_bo_mem_index(nvbo, index, mem);
+
 	if (is_iomem)
 		return ioread32_native((void __force __iomem *)mem);
 	else
@@ -450,7 +580,9 @@ nouveau_bo_wr32(struct nouveau_bo *nvbo, unsigned index, u32 val)
 {
 	bool is_iomem;
 	u32 *mem = ttm_kmap_obj_virtual(&nvbo->kmap, &is_iomem);
-	mem = &mem[index];
+
+	mem = nouveau_bo_mem_index(nvbo, index, mem);
+
 	if (is_iomem)
 		iowrite32_native(val, (void __force __iomem *)mem);
 	else
@@ -961,13 +1093,14 @@ nouveau_bo_move_m2mf(struct ttm_buffer_object *bo, int evict, bool intr,
 	}
 
 	mutex_lock_nested(&cli->mutex, SINGLE_DEPTH_NESTING);
-	ret = nouveau_fence_sync(bo->sync_obj, chan);
+	ret = nouveau_fence_sync(nouveau_bo(bo), chan, true, intr);
 	if (ret == 0) {
 		ret = drm->ttm.move(chan, bo, &bo->mem, new_mem);
 		if (ret == 0) {
 			ret = nouveau_fence_new(chan, false, &fence);
 			if (ret == 0) {
-				ret = ttm_bo_move_accel_cleanup(bo, fence,
+				ret = ttm_bo_move_accel_cleanup(bo,
+								&fence->base,
 								evict,
 								no_wait_gpu,
 								new_mem);
@@ -1041,12 +1174,15 @@ static int
 nouveau_bo_move_flipd(struct ttm_buffer_object *bo, bool evict, bool intr,
 		      bool no_wait_gpu, struct ttm_mem_reg *new_mem)
 {
-	u32 placement_memtype = TTM_PL_FLAG_TT | TTM_PL_MASK_CACHING;
+	struct ttm_place placement_memtype = {
+		.fpfn = 0,
+		.lpfn = 0,
+		.flags = TTM_PL_FLAG_TT | TTM_PL_MASK_CACHING
+	};
 	struct ttm_placement placement;
 	struct ttm_mem_reg tmp_mem;
 	int ret;
 
-	placement.fpfn = placement.lpfn = 0;
 	placement.num_placement = placement.num_busy_placement = 1;
 	placement.placement = placement.busy_placement = &placement_memtype;
 
@@ -1074,12 +1210,15 @@ static int
 nouveau_bo_move_flips(struct ttm_buffer_object *bo, bool evict, bool intr,
 		      bool no_wait_gpu, struct ttm_mem_reg *new_mem)
 {
-	u32 placement_memtype = TTM_PL_FLAG_TT | TTM_PL_MASK_CACHING;
+	struct ttm_place placement_memtype = {
+		.fpfn = 0,
+		.lpfn = 0,
+		.flags = TTM_PL_FLAG_TT | TTM_PL_MASK_CACHING
+	};
 	struct ttm_placement placement;
 	struct ttm_mem_reg tmp_mem;
 	int ret;
 
-	placement.fpfn = placement.lpfn = 0;
 	placement.num_placement = placement.num_busy_placement = 1;
 	placement.placement = placement.busy_placement = &placement_memtype;
 
@@ -1152,8 +1291,9 @@ nouveau_bo_vm_cleanup(struct ttm_buffer_object *bo,
 {
 	struct nouveau_drm *drm = nouveau_bdev(bo->bdev);
 	struct drm_device *dev = drm->dev;
+	struct fence *fence = reservation_object_get_excl(bo->resv);
 
-	nv10_bo_put_tile_region(dev, *old_tile, bo->sync_obj);
+	nv10_bo_put_tile_region(dev, *old_tile, fence);
 	*old_tile = new_tile;
 }
 
@@ -1167,6 +1307,9 @@ nouveau_bo_move(struct ttm_buffer_object *bo, bool evict, bool intr,
 	struct nouveau_drm_tile *new_tile = NULL;
 	int ret = 0;
 
+	if (nvbo->pin_refcnt)
+		NV_WARN(drm, "Moving pinned object %p!\n", nvbo);
+
 	if (drm->device.info.family < NV_DEVICE_INFO_V0_TESLA) {
 		ret = nouveau_bo_vm_bind(bo, new_mem, &new_tile);
 		if (ret)
@@ -1197,9 +1340,7 @@ nouveau_bo_move(struct ttm_buffer_object *bo, bool evict, bool intr,
 	}
 
 	/* Fallback to software copy. */
-	spin_lock(&bo->bdev->fence_lock);
 	ret = ttm_bo_wait(bo, true, intr, no_wait_gpu);
-	spin_unlock(&bo->bdev->fence_lock);
 	if (ret == 0)
 		ret = ttm_bo_move_memcpy(bo, evict, no_wait_gpu, new_mem);
 
@@ -1294,7 +1435,7 @@ nouveau_ttm_fault_reserve_notify(struct ttm_buffer_object *bo)
 	struct nouveau_bo *nvbo = nouveau_bo(bo);
 	struct nvif_device *device = &drm->device;
 	u32 mappable = nv_device_resource_len(nvkm_device(device), 1) >> PAGE_SHIFT;
-	int ret;
+	int i, ret;
 
 	/* as long as the bo isn't in vram, and isn't tiled, we've got
 	 * nothing to do here.
@@ -1319,9 +1460,16 @@ nouveau_ttm_fault_reserve_notify(struct ttm_buffer_object *bo)
 	    bo->mem.start + bo->mem.num_pages < mappable)
 		return 0;
 
+	for (i = 0; i < nvbo->placement.num_placement; ++i) {
+		nvbo->placements[i].fpfn = 0;
+		nvbo->placements[i].lpfn = mappable;
+	}
+
+	for (i = 0; i < nvbo->placement.num_busy_placement; ++i) {
+		nvbo->busy_placements[i].fpfn = 0;
+		nvbo->busy_placements[i].lpfn = mappable;
+	}
 
-	nvbo->placement.fpfn = 0;
-	nvbo->placement.lpfn = mappable;
 	nouveau_bo_placement_set(nvbo, TTM_PL_FLAG_VRAM, 0);
 	return nouveau_bo_validate(nvbo, false, false);
 }
@@ -1354,6 +1502,14 @@ nouveau_ttm_tt_populate(struct ttm_tt *ttm)
 	dev = drm->dev;
 	pdev = nv_device_base(device);
 
+	/*
+	 * Objects matching this condition have been marked as force_coherent,
+	 * so use the DMA API for them.
+	 */
+	if (!nv_device_is_cpu_coherent(device) &&
+	    ttm->caching_state == tt_uncached)
+		return ttm_dma_populate(ttm_dma, dev->dev);
+
 #if __OS_HAS_AGP
 	if (drm->agp.stat == ENABLED) {
 		return ttm_agp_tt_populate(ttm);
@@ -1411,6 +1567,16 @@ nouveau_ttm_tt_unpopulate(struct ttm_tt *ttm)
 	dev = drm->dev;
 	pdev = nv_device_base(device);
 
+	/*
+	 * Objects matching this condition have been marked as force_coherent,
+	 * so use the DMA API for them.
+	 */
+	if (!nv_device_is_cpu_coherent(device) &&
+	    ttm->caching_state == tt_uncached) {
+		ttm_dma_unpopulate(ttm_dma, dev->dev);
+		return;
+	}
+
 #if __OS_HAS_AGP
 	if (drm->agp.stat == ENABLED) {
 		ttm_agp_tt_unpopulate(ttm);
@@ -1436,47 +1602,14 @@ nouveau_ttm_tt_unpopulate(struct ttm_tt *ttm)
 }
 
 void
-nouveau_bo_fence(struct nouveau_bo *nvbo, struct nouveau_fence *fence)
+nouveau_bo_fence(struct nouveau_bo *nvbo, struct nouveau_fence *fence, bool exclusive)
 {
-	struct nouveau_fence *new_fence = nouveau_fence_ref(fence);
-	struct nouveau_fence *old_fence = NULL;
-
-	spin_lock(&nvbo->bo.bdev->fence_lock);
-	old_fence = nvbo->bo.sync_obj;
-	nvbo->bo.sync_obj = new_fence;
-	spin_unlock(&nvbo->bo.bdev->fence_lock);
-
-	nouveau_fence_unref(&old_fence);
-}
-
-static void
-nouveau_bo_fence_unref(void **sync_obj)
-{
-	nouveau_fence_unref((struct nouveau_fence **)sync_obj);
-}
+	struct reservation_object *resv = nvbo->bo.resv;
 
-static void *
-nouveau_bo_fence_ref(void *sync_obj)
-{
-	return nouveau_fence_ref(sync_obj);
-}
-
-static bool
-nouveau_bo_fence_signalled(void *sync_obj)
-{
-	return nouveau_fence_done(sync_obj);
-}
-
-static int
-nouveau_bo_fence_wait(void *sync_obj, bool lazy, bool intr)
-{
-	return nouveau_fence_wait(sync_obj, lazy, intr);
-}
-
-static int
-nouveau_bo_fence_flush(void *sync_obj)
-{
-	return 0;
+	if (exclusive)
+		reservation_object_add_excl_fence(resv, &fence->base);
+	else if (fence)
+		reservation_object_add_shared_fence(resv, &fence->base);
 }
 
 struct ttm_bo_driver nouveau_bo_driver = {
@@ -1489,11 +1622,6 @@ struct ttm_bo_driver nouveau_bo_driver = {
 	.move_notify = nouveau_bo_move_ntfy,
 	.move = nouveau_bo_move,
 	.verify_access = nouveau_bo_verify_access,
-	.sync_obj_signaled = nouveau_bo_fence_signalled,
-	.sync_obj_wait = nouveau_bo_fence_wait,
-	.sync_obj_flush = nouveau_bo_fence_flush,
-	.sync_obj_unref = nouveau_bo_fence_unref,
-	.sync_obj_ref = nouveau_bo_fence_ref,
 	.fault_reserve_notify = &nouveau_ttm_fault_reserve_notify,
 	.io_mem_reserve = &nouveau_ttm_io_mem_reserve,
 	.io_mem_free = &nouveau_ttm_io_mem_free,
diff --git a/drivers/gpu/drm/nouveau/nouveau_bo.h b/drivers/gpu/drm/nouveau/nouveau_bo.h
index ff17c1f432fc..072222efeeb7 100644
--- a/drivers/gpu/drm/nouveau/nouveau_bo.h
+++ b/drivers/gpu/drm/nouveau/nouveau_bo.h
@@ -1,6 +1,8 @@
 #ifndef __NOUVEAU_BO_H__
 #define __NOUVEAU_BO_H__
 
+#include <drm/drm_gem.h>
+
 struct nouveau_channel;
 struct nouveau_fence;
 struct nouveau_vma;
@@ -9,8 +11,9 @@ struct nouveau_bo {
 	struct ttm_buffer_object bo;
 	struct ttm_placement placement;
 	u32 valid_domains;
-	u32 placements[3];
-	u32 busy_placements[3];
+	struct ttm_place placements[3];
+	struct ttm_place busy_placements[3];
+	bool force_coherent;
 	struct ttm_bo_kmap_obj kmap;
 	struct list_head head;
 
@@ -68,8 +71,9 @@ extern struct ttm_bo_driver nouveau_bo_driver;
 void nouveau_bo_move_init(struct nouveau_drm *);
 int  nouveau_bo_new(struct drm_device *, int size, int align, u32 flags,
 		    u32 tile_mode, u32 tile_flags, struct sg_table *sg,
+		    struct reservation_object *robj,
 		    struct nouveau_bo **);
-int  nouveau_bo_pin(struct nouveau_bo *, u32 flags);
+int  nouveau_bo_pin(struct nouveau_bo *, u32 flags, bool contig);
 int  nouveau_bo_unpin(struct nouveau_bo *);
 int  nouveau_bo_map(struct nouveau_bo *);
 void nouveau_bo_unmap(struct nouveau_bo *);
@@ -78,9 +82,11 @@ u16  nouveau_bo_rd16(struct nouveau_bo *, unsigned index);
 void nouveau_bo_wr16(struct nouveau_bo *, unsigned index, u16 val);
 u32  nouveau_bo_rd32(struct nouveau_bo *, unsigned index);
 void nouveau_bo_wr32(struct nouveau_bo *, unsigned index, u32 val);
-void nouveau_bo_fence(struct nouveau_bo *, struct nouveau_fence *);
+void nouveau_bo_fence(struct nouveau_bo *, struct nouveau_fence *, bool exclusive);
 int  nouveau_bo_validate(struct nouveau_bo *, bool interruptible,
 			 bool no_wait_gpu);
+void nouveau_bo_sync_for_device(struct nouveau_bo *nvbo);
+void nouveau_bo_sync_for_cpu(struct nouveau_bo *nvbo);
 
 struct nouveau_vma *
 nouveau_bo_vma_find(struct nouveau_bo *, struct nouveau_vm *);
diff --git a/drivers/gpu/drm/nouveau/nouveau_chan.c b/drivers/gpu/drm/nouveau/nouveau_chan.c
index 3440fc999f2f..aff9099aae6c 100644
--- a/drivers/gpu/drm/nouveau/nouveau_chan.c
+++ b/drivers/gpu/drm/nouveau/nouveau_chan.c
@@ -36,7 +36,7 @@
 #include "nouveau_abi16.h"
 
 MODULE_PARM_DESC(vram_pushbuf, "Create DMA push buffers in VRAM");
-static int nouveau_vram_pushbuf;
+int nouveau_vram_pushbuf;
 module_param_named(vram_pushbuf, nouveau_vram_pushbuf, int, 0400);
 
 int
@@ -102,14 +102,14 @@ nouveau_channel_prep(struct nouveau_drm *drm, struct nvif_device *device,
 	chan->drm = drm;
 
 	/* allocate memory for dma push buffer */
-	target = TTM_PL_FLAG_TT;
+	target = TTM_PL_FLAG_TT | TTM_PL_FLAG_UNCACHED;
 	if (nouveau_vram_pushbuf)
 		target = TTM_PL_FLAG_VRAM;
 
-	ret = nouveau_bo_new(drm->dev, size, 0, target, 0, 0, NULL,
+	ret = nouveau_bo_new(drm->dev, size, 0, target, 0, 0, NULL, NULL,
 			    &chan->push.buffer);
 	if (ret == 0) {
-		ret = nouveau_bo_pin(chan->push.buffer, target);
+		ret = nouveau_bo_pin(chan->push.buffer, target, false);
 		if (ret == 0)
 			ret = nouveau_bo_map(chan->push.buffer);
 	}
@@ -285,7 +285,6 @@ nouveau_channel_init(struct nouveau_channel *chan, u32 vram, u32 gart)
 	struct nouveau_software_chan *swch;
 	struct nv_dma_v0 args = {};
 	int ret, i;
-	bool save;
 
 	nvif_object_map(chan->object);
 
@@ -387,11 +386,7 @@ nouveau_channel_init(struct nouveau_channel *chan, u32 vram, u32 gart)
 	}
 
 	/* initialise synchronisation */
-	save = cli->base.super;
-	cli->base.super = true; /* hack until fencenv50 fixed */
-	ret = nouveau_fence(chan->drm)->context_new(chan);
-	cli->base.super = save;
-	return ret;
+	return nouveau_fence(chan->drm)->context_new(chan);
 }
 
 int
@@ -400,15 +395,20 @@ nouveau_channel_new(struct nouveau_drm *drm, struct nvif_device *device,
 		    struct nouveau_channel **pchan)
 {
 	struct nouveau_cli *cli = (void *)nvif_client(&device->base);
+	bool super;
 	int ret;
 
+	/* hack until fencenv50 is fixed, and agp access relaxed */
+	super = cli->base.super;
+	cli->base.super = true;
+
 	ret = nouveau_channel_ind(drm, device, handle, arg0, pchan);
 	if (ret) {
 		NV_PRINTK(debug, cli, "ib channel create, %d\n", ret);
 		ret = nouveau_channel_dma(drm, device, handle, pchan);
 		if (ret) {
 			NV_PRINTK(debug, cli, "dma channel create, %d\n", ret);
-			return ret;
+			goto done;
 		}
 	}
 
@@ -416,8 +416,9 @@ nouveau_channel_new(struct nouveau_drm *drm, struct nvif_device *device,
 	if (ret) {
 		NV_PRINTK(error, cli, "channel failed to initialise, %d\n", ret);
 		nouveau_channel_del(pchan);
-		return ret;
 	}
 
-	return 0;
+done:
+	cli->base.super = super;
+	return ret;
 }
diff --git a/drivers/gpu/drm/nouveau/nouveau_chan.h b/drivers/gpu/drm/nouveau/nouveau_chan.h
index 20163709d608..8309c24ee698 100644
--- a/drivers/gpu/drm/nouveau/nouveau_chan.h
+++ b/drivers/gpu/drm/nouveau/nouveau_chan.h
@@ -47,4 +47,6 @@ int  nouveau_channel_new(struct nouveau_drm *, struct nvif_device *,
 void nouveau_channel_del(struct nouveau_channel **);
 int  nouveau_channel_idle(struct nouveau_channel *);
 
+extern int nouveau_vram_pushbuf;
+
 #endif
diff --git a/drivers/gpu/drm/nouveau/nouveau_connector.c b/drivers/gpu/drm/nouveau/nouveau_connector.c
index 1ec44c83e919..c8ac9482cf2e 100644
--- a/drivers/gpu/drm/nouveau/nouveau_connector.c
+++ b/drivers/gpu/drm/nouveau/nouveau_connector.c
@@ -45,15 +45,15 @@
 #include <nvif/event.h>
 
 MODULE_PARM_DESC(tv_disable, "Disable TV-out detection");
-static int nouveau_tv_disable = 0;
+int nouveau_tv_disable = 0;
 module_param_named(tv_disable, nouveau_tv_disable, int, 0400);
 
 MODULE_PARM_DESC(ignorelid, "Ignore ACPI lid status");
-static int nouveau_ignorelid = 0;
+int nouveau_ignorelid = 0;
 module_param_named(ignorelid, nouveau_ignorelid, int, 0400);
 
 MODULE_PARM_DESC(duallink, "Allow dual-link TMDS (default: enabled)");
-static int nouveau_duallink = 1;
+int nouveau_duallink = 1;
 module_param_named(duallink, nouveau_duallink, int, 0400);
 
 struct nouveau_encoder *
diff --git a/drivers/gpu/drm/nouveau/nouveau_connector.h b/drivers/gpu/drm/nouveau/nouveau_connector.h
index 68029d041dd2..629a380c7085 100644
--- a/drivers/gpu/drm/nouveau/nouveau_connector.h
+++ b/drivers/gpu/drm/nouveau/nouveau_connector.h
@@ -105,4 +105,8 @@ nouveau_crtc_connector_get(struct nouveau_crtc *nv_crtc)
 struct drm_connector *
 nouveau_connector_create(struct drm_device *, int index);
 
+extern int nouveau_tv_disable;
+extern int nouveau_ignorelid;
+extern int nouveau_duallink;
+
 #endif /* __NOUVEAU_CONNECTOR_H__ */
diff --git a/drivers/gpu/drm/nouveau/nouveau_display.c b/drivers/gpu/drm/nouveau/nouveau_display.c
index 4a21b2b06ce2..f8042433752b 100644
--- a/drivers/gpu/drm/nouveau/nouveau_display.c
+++ b/drivers/gpu/drm/nouveau/nouveau_display.c
@@ -126,7 +126,7 @@ nouveau_display_scanoutpos_head(struct drm_crtc *crtc, int *vpos, int *hpos,
 	if (etime) *etime = ns_to_ktime(args.scan.time[1]);
 
 	if (*vpos < 0)
-		ret |= DRM_SCANOUTPOS_INVBL;
+		ret |= DRM_SCANOUTPOS_IN_VBLANK;
 	return ret;
 }
 
@@ -479,6 +479,7 @@ nouveau_display_create(struct drm_device *dev)
 
 	if (nouveau_modeset != 2 && drm->vbios.dcb.entries) {
 		static const u16 oclass[] = {
+			GM204_DISP,
 			GM107_DISP,
 			GK110_DISP,
 			GK104_DISP,
@@ -568,9 +569,10 @@ nouveau_display_suspend(struct drm_device *dev, bool runtime)
 
 	list_for_each_entry(crtc, &dev->mode_config.crtc_list, head) {
 		struct nouveau_crtc *nv_crtc = nouveau_crtc(crtc);
-
-		nouveau_bo_unmap(nv_crtc->cursor.nvbo);
-		nouveau_bo_unpin(nv_crtc->cursor.nvbo);
+		if (nv_crtc->cursor.nvbo) {
+			nouveau_bo_unmap(nv_crtc->cursor.nvbo);
+			nouveau_bo_unpin(nv_crtc->cursor.nvbo);
+		}
 	}
 
 	return 0;
@@ -591,15 +593,17 @@ nouveau_display_resume(struct drm_device *dev, bool runtime)
 		if (!nouveau_fb || !nouveau_fb->nvbo)
 			continue;
 
-		ret = nouveau_bo_pin(nouveau_fb->nvbo, TTM_PL_FLAG_VRAM);
+		ret = nouveau_bo_pin(nouveau_fb->nvbo, TTM_PL_FLAG_VRAM, true);
 		if (ret)
 			NV_ERROR(drm, "Could not pin framebuffer\n");
 	}
 
 	list_for_each_entry(crtc, &dev->mode_config.crtc_list, head) {
 		struct nouveau_crtc *nv_crtc = nouveau_crtc(crtc);
+		if (!nv_crtc->cursor.nvbo)
+			continue;
 
-		ret = nouveau_bo_pin(nv_crtc->cursor.nvbo, TTM_PL_FLAG_VRAM);
+		ret = nouveau_bo_pin(nv_crtc->cursor.nvbo, TTM_PL_FLAG_VRAM, true);
 		if (!ret)
 			ret = nouveau_bo_map(nv_crtc->cursor.nvbo);
 		if (ret)
@@ -630,9 +634,10 @@ nouveau_display_resume(struct drm_device *dev, bool runtime)
 
 	list_for_each_entry(crtc, &dev->mode_config.crtc_list, head) {
 		struct nouveau_crtc *nv_crtc = nouveau_crtc(crtc);
-		u32 offset = nv_crtc->cursor.nvbo->bo.offset;
 
-		nv_crtc->cursor.set_offset(nv_crtc, offset);
+		if (!nv_crtc->cursor.nvbo)
+			continue;
+		nv_crtc->cursor.set_offset(nv_crtc, nv_crtc->cursor.nvbo->bo.offset);
 		nv_crtc->cursor.set_pos(nv_crtc, nv_crtc->cursor_saved_x,
 						 nv_crtc->cursor_saved_y);
 	}
@@ -657,7 +662,7 @@ nouveau_page_flip_emit(struct nouveau_channel *chan,
 	spin_unlock_irqrestore(&dev->event_lock, flags);
 
 	/* Synchronize with the old framebuffer */
-	ret = nouveau_fence_sync(old_bo->bo.sync_obj, chan);
+	ret = nouveau_fence_sync(old_bo, chan, false, false);
 	if (ret)
 		goto fail;
 
@@ -710,25 +715,30 @@ nouveau_crtc_page_flip(struct drm_crtc *crtc, struct drm_framebuffer *fb,
 		return -ENOMEM;
 
 	if (new_bo != old_bo) {
-		ret = nouveau_bo_pin(new_bo, TTM_PL_FLAG_VRAM);
+		ret = nouveau_bo_pin(new_bo, TTM_PL_FLAG_VRAM, true);
 		if (ret)
 			goto fail_free;
 	}
 
 	mutex_lock(&cli->mutex);
-
-	/* synchronise rendering channel with the kernel's channel */
-	spin_lock(&new_bo->bo.bdev->fence_lock);
-	fence = nouveau_fence_ref(new_bo->bo.sync_obj);
-	spin_unlock(&new_bo->bo.bdev->fence_lock);
-	ret = nouveau_fence_sync(fence, chan);
-	nouveau_fence_unref(&fence);
+	ret = ttm_bo_reserve(&new_bo->bo, true, false, false, NULL);
 	if (ret)
 		goto fail_unpin;
 
-	ret = ttm_bo_reserve(&old_bo->bo, true, false, false, NULL);
-	if (ret)
+	/* synchronise rendering channel with the kernel's channel */
+	ret = nouveau_fence_sync(new_bo, chan, false, true);
+	if (ret) {
+		ttm_bo_unreserve(&new_bo->bo);
 		goto fail_unpin;
+	}
+
+	if (new_bo != old_bo) {
+		ttm_bo_unreserve(&new_bo->bo);
+
+		ret = ttm_bo_reserve(&old_bo->bo, true, false, false, NULL);
+		if (ret)
+			goto fail_unpin;
+	}
 
 	/* Initialize a page flip struct */
 	*s = (struct nouveau_page_flip_state)
@@ -774,7 +784,7 @@ nouveau_crtc_page_flip(struct drm_crtc *crtc, struct drm_framebuffer *fb,
 	/* Update the crtc struct and cleanup */
 	crtc->primary->fb = fb;
 
-	nouveau_bo_fence(old_bo, fence);
+	nouveau_bo_fence(old_bo, fence, false);
 	ttm_bo_unreserve(&old_bo->bo);
 	if (old_bo != new_bo)
 		nouveau_bo_unpin(old_bo);
diff --git a/drivers/gpu/drm/nouveau/nouveau_drm.c b/drivers/gpu/drm/nouveau/nouveau_drm.c
index 3ed32dd90303..65910e3aed0c 100644
--- a/drivers/gpu/drm/nouveau/nouveau_drm.c
+++ b/drivers/gpu/drm/nouveau/nouveau_drm.c
@@ -51,6 +51,7 @@
 #include "nouveau_fence.h"
 #include "nouveau_debugfs.h"
 #include "nouveau_usif.h"
+#include "nouveau_connector.h"
 
 MODULE_PARM_DESC(config, "option string to pass to driver core");
 static char *nouveau_config;
@@ -73,7 +74,9 @@ MODULE_PARM_DESC(runpm, "disable (0), force enable (1), optimus only default (-1
 int nouveau_runtime_pm = -1;
 module_param_named(runpm, nouveau_runtime_pm, int, 0400);
 
-static struct drm_driver driver;
+static struct drm_driver driver_stub;
+static struct drm_driver driver_pci;
+static struct drm_driver driver_platform;
 
 static u64
 nouveau_pci_name(struct pci_dev *pdev)
@@ -322,7 +325,7 @@ static int nouveau_drm_probe(struct pci_dev *pdev,
 
 	pci_set_master(pdev);
 
-	ret = drm_get_pci_dev(pdev, pent, &driver);
+	ret = drm_get_pci_dev(pdev, pent, &driver_pci);
 	if (ret) {
 		nouveau_object_ref(NULL, (struct nouveau_object **)&device);
 		return ret;
@@ -610,27 +613,6 @@ fail_display:
 	return ret;
 }
 
-int nouveau_pmops_suspend(struct device *dev)
-{
-	struct pci_dev *pdev = to_pci_dev(dev);
-	struct drm_device *drm_dev = pci_get_drvdata(pdev);
-	int ret;
-
-	if (drm_dev->switch_power_state == DRM_SWITCH_POWER_OFF ||
-	    drm_dev->switch_power_state == DRM_SWITCH_POWER_DYNAMIC_OFF)
-		return 0;
-
-	ret = nouveau_do_suspend(drm_dev, false);
-	if (ret)
-		return ret;
-
-	pci_save_state(pdev);
-	pci_disable_device(pdev);
-	pci_ignore_hotplug(pdev);
-	pci_set_power_state(pdev, PCI_D3hot);
-	return 0;
-}
-
 static int
 nouveau_do_resume(struct drm_device *dev, bool runtime)
 {
@@ -665,7 +647,29 @@ nouveau_do_resume(struct drm_device *dev, bool runtime)
 	return 0;
 }
 
-int nouveau_pmops_resume(struct device *dev)
+int
+nouveau_pmops_suspend(struct device *dev)
+{
+	struct pci_dev *pdev = to_pci_dev(dev);
+	struct drm_device *drm_dev = pci_get_drvdata(pdev);
+	int ret;
+
+	if (drm_dev->switch_power_state == DRM_SWITCH_POWER_OFF ||
+	    drm_dev->switch_power_state == DRM_SWITCH_POWER_DYNAMIC_OFF)
+		return 0;
+
+	ret = nouveau_do_suspend(drm_dev, false);
+	if (ret)
+		return ret;
+
+	pci_save_state(pdev);
+	pci_disable_device(pdev);
+	pci_set_power_state(pdev, PCI_D3hot);
+	return 0;
+}
+
+int
+nouveau_pmops_resume(struct device *dev)
 {
 	struct pci_dev *pdev = to_pci_dev(dev);
 	struct drm_device *drm_dev = pci_get_drvdata(pdev);
@@ -685,20 +689,122 @@ int nouveau_pmops_resume(struct device *dev)
 	return nouveau_do_resume(drm_dev, false);
 }
 
-static int nouveau_pmops_freeze(struct device *dev)
+static int
+nouveau_pmops_freeze(struct device *dev)
 {
 	struct pci_dev *pdev = to_pci_dev(dev);
 	struct drm_device *drm_dev = pci_get_drvdata(pdev);
 	return nouveau_do_suspend(drm_dev, false);
 }
 
-static int nouveau_pmops_thaw(struct device *dev)
+static int
+nouveau_pmops_thaw(struct device *dev)
 {
 	struct pci_dev *pdev = to_pci_dev(dev);
 	struct drm_device *drm_dev = pci_get_drvdata(pdev);
 	return nouveau_do_resume(drm_dev, false);
 }
 
+static int
+nouveau_pmops_runtime_suspend(struct device *dev)
+{
+	struct pci_dev *pdev = to_pci_dev(dev);
+	struct drm_device *drm_dev = pci_get_drvdata(pdev);
+	int ret;
+
+	if (nouveau_runtime_pm == 0) {
+		pm_runtime_forbid(dev);
+		return -EBUSY;
+	}
+
+	/* are we optimus enabled? */
+	if (nouveau_runtime_pm == -1 && !nouveau_is_optimus() && !nouveau_is_v1_dsm()) {
+		DRM_DEBUG_DRIVER("failing to power off - not optimus\n");
+		pm_runtime_forbid(dev);
+		return -EBUSY;
+	}
+
+	nv_debug_level(SILENT);
+	drm_kms_helper_poll_disable(drm_dev);
+	vga_switcheroo_set_dynamic_switch(pdev, VGA_SWITCHEROO_OFF);
+	nouveau_switcheroo_optimus_dsm();
+	ret = nouveau_do_suspend(drm_dev, true);
+	pci_save_state(pdev);
+	pci_disable_device(pdev);
+	pci_ignore_hotplug(pdev);
+	pci_set_power_state(pdev, PCI_D3cold);
+	drm_dev->switch_power_state = DRM_SWITCH_POWER_DYNAMIC_OFF;
+	return ret;
+}
+
+static int
+nouveau_pmops_runtime_resume(struct device *dev)
+{
+	struct pci_dev *pdev = to_pci_dev(dev);
+	struct drm_device *drm_dev = pci_get_drvdata(pdev);
+	struct nvif_device *device = &nouveau_drm(drm_dev)->device;
+	int ret;
+
+	if (nouveau_runtime_pm == 0)
+		return -EINVAL;
+
+	pci_set_power_state(pdev, PCI_D0);
+	pci_restore_state(pdev);
+	ret = pci_enable_device(pdev);
+	if (ret)
+		return ret;
+	pci_set_master(pdev);
+
+	ret = nouveau_do_resume(drm_dev, true);
+	drm_kms_helper_poll_enable(drm_dev);
+	/* do magic */
+	nvif_mask(device, 0x88488, (1 << 25), (1 << 25));
+	vga_switcheroo_set_dynamic_switch(pdev, VGA_SWITCHEROO_ON);
+	drm_dev->switch_power_state = DRM_SWITCH_POWER_ON;
+	nv_debug_level(NORMAL);
+	return ret;
+}
+
+static int
+nouveau_pmops_runtime_idle(struct device *dev)
+{
+	struct pci_dev *pdev = to_pci_dev(dev);
+	struct drm_device *drm_dev = pci_get_drvdata(pdev);
+	struct nouveau_drm *drm = nouveau_drm(drm_dev);
+	struct drm_crtc *crtc;
+
+	if (nouveau_runtime_pm == 0) {
+		pm_runtime_forbid(dev);
+		return -EBUSY;
+	}
+
+	/* are we optimus enabled? */
+	if (nouveau_runtime_pm == -1 && !nouveau_is_optimus() && !nouveau_is_v1_dsm()) {
+		DRM_DEBUG_DRIVER("failing to power off - not optimus\n");
+		pm_runtime_forbid(dev);
+		return -EBUSY;
+	}
+
+	/* if we have a hdmi audio device - make sure it has a driver loaded */
+	if (drm->hdmi_device) {
+		if (!drm->hdmi_device->driver) {
+			DRM_DEBUG_DRIVER("failing to power off - no HDMI audio driver loaded\n");
+			pm_runtime_mark_last_busy(dev);
+			return -EBUSY;
+		}
+	}
+
+	list_for_each_entry(crtc, &drm->dev->mode_config.crtc_list, head) {
+		if (crtc->enabled) {
+			DRM_DEBUG_DRIVER("failing to power off - crtc active\n");
+			return -EBUSY;
+		}
+	}
+	pm_runtime_mark_last_busy(dev);
+	pm_runtime_autosuspend(dev);
+	/* we don't want the main rpm_idle to call suspend - we want to autosuspend */
+	return 1;
+}
 
 static int
 nouveau_drm_open(struct drm_device *dev, struct drm_file *fpriv)
@@ -831,7 +937,7 @@ nouveau_driver_fops = {
 };
 
 static struct drm_driver
-driver = {
+driver_stub = {
 	.driver_features =
 		DRIVER_USE_AGP |
 		DRIVER_GEM | DRIVER_MODESET | DRIVER_PRIME | DRIVER_RENDER,
@@ -905,101 +1011,21 @@ nouveau_drm_pci_table[] = {
 	{}
 };
 
-static int nouveau_pmops_runtime_suspend(struct device *dev)
+static void nouveau_display_options(void)
 {
-	struct pci_dev *pdev = to_pci_dev(dev);
-	struct drm_device *drm_dev = pci_get_drvdata(pdev);
-	int ret;
-
-	if (nouveau_runtime_pm == 0) {
-		pm_runtime_forbid(dev);
-		return -EBUSY;
-	}
-
-	/* are we optimus enabled? */
-	if (nouveau_runtime_pm == -1 && !nouveau_is_optimus() && !nouveau_is_v1_dsm()) {
-		DRM_DEBUG_DRIVER("failing to power off - not optimus\n");
-		pm_runtime_forbid(dev);
-		return -EBUSY;
-	}
-
-	nv_debug_level(SILENT);
-	drm_kms_helper_poll_disable(drm_dev);
-	vga_switcheroo_set_dynamic_switch(pdev, VGA_SWITCHEROO_OFF);
-	nouveau_switcheroo_optimus_dsm();
-	ret = nouveau_do_suspend(drm_dev, true);
-	pci_save_state(pdev);
-	pci_disable_device(pdev);
-	pci_set_power_state(pdev, PCI_D3cold);
-	drm_dev->switch_power_state = DRM_SWITCH_POWER_DYNAMIC_OFF;
-	return ret;
-}
-
-static int nouveau_pmops_runtime_resume(struct device *dev)
-{
-	struct pci_dev *pdev = to_pci_dev(dev);
-	struct drm_device *drm_dev = pci_get_drvdata(pdev);
-	struct nvif_device *device = &nouveau_drm(drm_dev)->device;
-	int ret;
-
-	if (nouveau_runtime_pm == 0)
-		return -EINVAL;
-
-	pci_set_power_state(pdev, PCI_D0);
-	pci_restore_state(pdev);
-	ret = pci_enable_device(pdev);
-	if (ret)
-		return ret;
-	pci_set_master(pdev);
-
-	ret = nouveau_do_resume(drm_dev, true);
-	drm_kms_helper_poll_enable(drm_dev);
-	/* do magic */
-	nvif_mask(device, 0x88488, (1 << 25), (1 << 25));
-	vga_switcheroo_set_dynamic_switch(pdev, VGA_SWITCHEROO_ON);
-	drm_dev->switch_power_state = DRM_SWITCH_POWER_ON;
-	nv_debug_level(NORMAL);
-	return ret;
-}
-
-static int nouveau_pmops_runtime_idle(struct device *dev)
-{
-	struct pci_dev *pdev = to_pci_dev(dev);
-	struct drm_device *drm_dev = pci_get_drvdata(pdev);
-	struct nouveau_drm *drm = nouveau_drm(drm_dev);
-	struct drm_crtc *crtc;
-
-	if (nouveau_runtime_pm == 0) {
-		pm_runtime_forbid(dev);
-		return -EBUSY;
-	}
-
-	/* are we optimus enabled? */
-	if (nouveau_runtime_pm == -1 && !nouveau_is_optimus() && !nouveau_is_v1_dsm()) {
-		DRM_DEBUG_DRIVER("failing to power off - not optimus\n");
-		pm_runtime_forbid(dev);
-		return -EBUSY;
-	}
-
-	/* if we have a hdmi audio device - make sure it has a driver loaded */
-	if (drm->hdmi_device) {
-		if (!drm->hdmi_device->driver) {
-			DRM_DEBUG_DRIVER("failing to power off - no HDMI audio driver loaded\n");
-			pm_runtime_mark_last_busy(dev);
-			return -EBUSY;
-		}
-	}
-
-	list_for_each_entry(crtc, &drm->dev->mode_config.crtc_list, head) {
-		if (crtc->enabled) {
-			DRM_DEBUG_DRIVER("failing to power off - crtc active\n");
-			return -EBUSY;
-		}
-	}
-	pm_runtime_mark_last_busy(dev);
-	pm_runtime_autosuspend(dev);
-	/* we don't want the main rpm_idle to call suspend - we want to autosuspend */
-	return 1;
+	DRM_DEBUG_DRIVER("Loading Nouveau with parameters:\n");
+
+	DRM_DEBUG_DRIVER("... tv_disable   : %d\n", nouveau_tv_disable);
+	DRM_DEBUG_DRIVER("... ignorelid    : %d\n", nouveau_ignorelid);
+	DRM_DEBUG_DRIVER("... duallink     : %d\n", nouveau_duallink);
+	DRM_DEBUG_DRIVER("... nofbaccel    : %d\n", nouveau_nofbaccel);
+	DRM_DEBUG_DRIVER("... config       : %s\n", nouveau_config);
+	DRM_DEBUG_DRIVER("... debug        : %s\n", nouveau_debug);
+	DRM_DEBUG_DRIVER("... noaccel      : %d\n", nouveau_noaccel);
+	DRM_DEBUG_DRIVER("... modeset      : %d\n", nouveau_modeset);
+	DRM_DEBUG_DRIVER("... runpm        : %d\n", nouveau_runtime_pm);
+	DRM_DEBUG_DRIVER("... vram_pushbuf : %d\n", nouveau_vram_pushbuf);
+	DRM_DEBUG_DRIVER("... pstate       : %d\n", nouveau_pstate);
 }
 
 static const struct dev_pm_ops nouveau_pm_ops = {
@@ -1037,7 +1063,7 @@ nouveau_platform_device_create_(struct platform_device *pdev, int size,
 	if (err)
 		return ERR_PTR(err);
 
-	drm = drm_dev_alloc(&driver, &pdev->dev);
+	drm = drm_dev_alloc(&driver_platform, &pdev->dev);
 	if (!drm) {
 		err = -ENOMEM;
 		goto err_free;
@@ -1062,6 +1088,13 @@ EXPORT_SYMBOL(nouveau_platform_device_create_);
 static int __init
 nouveau_drm_init(void)
 {
+	driver_pci = driver_stub;
+	driver_pci.set_busid = drm_pci_set_busid;
+	driver_platform = driver_stub;
+	driver_platform.set_busid = drm_platform_set_busid;
+
+	nouveau_display_options();
+
 	if (nouveau_modeset == -1) {
 #ifdef CONFIG_VGA_CONSOLE
 		if (vgacon_text_force())
@@ -1073,7 +1106,7 @@ nouveau_drm_init(void)
 		return 0;
 
 	nouveau_register_dsm_handler();
-	return drm_pci_init(&driver, &nouveau_drm_pci_driver);
+	return drm_pci_init(&driver_pci, &nouveau_drm_pci_driver);
 }
 
 static void __exit
@@ -1082,7 +1115,7 @@ nouveau_drm_exit(void)
 	if (!nouveau_modeset)
 		return;
 
-	drm_pci_exit(&driver, &nouveau_drm_pci_driver);
+	drm_pci_exit(&driver_pci, &nouveau_drm_pci_driver);
 	nouveau_unregister_dsm_handler();
 }
 
diff --git a/drivers/gpu/drm/nouveau/nouveau_drm.h b/drivers/gpu/drm/nouveau/nouveau_drm.h
index b02b02452c85..8ae36f265fb8 100644
--- a/drivers/gpu/drm/nouveau/nouveau_drm.h
+++ b/drivers/gpu/drm/nouveau/nouveau_drm.h
@@ -10,7 +10,7 @@
 
 #define DRIVER_MAJOR		1
 #define DRIVER_MINOR		2
-#define DRIVER_PATCHLEVEL	0
+#define DRIVER_PATCHLEVEL	1
 
 /*
  * 1.1.1:
@@ -26,6 +26,8 @@
  * 1.2.0:
  * 	- object api exposed to userspace
  * 	- fermi,kepler,maxwell zbc
+ * 1.2.1:
+ *      - allow concurrent access to bo's mapped read/write.
  */
 
 #include <nvif/client.h>
diff --git a/drivers/gpu/drm/nouveau/nouveau_fbcon.c b/drivers/gpu/drm/nouveau/nouveau_fbcon.c
index 49fe6075cc7c..3ed12a8cfc91 100644
--- a/drivers/gpu/drm/nouveau/nouveau_fbcon.c
+++ b/drivers/gpu/drm/nouveau/nouveau_fbcon.c
@@ -52,7 +52,7 @@
 #include "nouveau_crtc.h"
 
 MODULE_PARM_DESC(nofbaccel, "Disable fbcon acceleration");
-static int nouveau_nofbaccel = 0;
+int nouveau_nofbaccel = 0;
 module_param_named(nofbaccel, nouveau_nofbaccel, int, 0400);
 
 static void
@@ -308,7 +308,8 @@ static int
 nouveau_fbcon_create(struct drm_fb_helper *helper,
 		     struct drm_fb_helper_surface_size *sizes)
 {
-	struct nouveau_fbdev *fbcon = (struct nouveau_fbdev *)helper;
+	struct nouveau_fbdev *fbcon =
+		container_of(helper, struct nouveau_fbdev, helper);
 	struct drm_device *dev = fbcon->dev;
 	struct nouveau_drm *drm = nouveau_drm(dev);
 	struct nvif_device *device = &drm->device;
@@ -340,7 +341,7 @@ nouveau_fbcon_create(struct drm_fb_helper *helper,
 		goto out;
 	}
 
-	ret = nouveau_bo_pin(nvbo, TTM_PL_FLAG_VRAM);
+	ret = nouveau_bo_pin(nvbo, TTM_PL_FLAG_VRAM, false);
 	if (ret) {
 		NV_ERROR(drm, "failed to pin fb: %d\n", ret);
 		goto out_unref;
@@ -497,6 +498,23 @@ nouveau_fbcon_set_suspend_work(struct work_struct *work)
 	console_unlock();
 }
 
+void
+nouveau_fbcon_set_suspend(struct drm_device *dev, int state)
+{
+	struct nouveau_drm *drm = nouveau_drm(dev);
+	if (drm->fbcon) {
+		if (state == FBINFO_STATE_RUNNING) {
+			schedule_work(&drm->fbcon->work);
+			return;
+		}
+		flush_work(&drm->fbcon->work);
+		console_lock();
+		fb_set_suspend(drm->fbcon->helper.fbdev, state);
+		nouveau_fbcon_accel_save_disable(dev);
+		console_unlock();
+	}
+}
+
 int
 nouveau_fbcon_init(struct drm_device *dev)
 {
@@ -556,20 +574,3 @@ nouveau_fbcon_fini(struct drm_device *dev)
 	kfree(drm->fbcon);
 	drm->fbcon = NULL;
 }
-
-void
-nouveau_fbcon_set_suspend(struct drm_device *dev, int state)
-{
-	struct nouveau_drm *drm = nouveau_drm(dev);
-	if (drm->fbcon) {
-		if (state == FBINFO_STATE_RUNNING) {
-			schedule_work(&drm->fbcon->work);
-			return;
-		}
-		flush_work(&drm->fbcon->work);
-		console_lock();
-		fb_set_suspend(drm->fbcon->helper.fbdev, state);
-		nouveau_fbcon_accel_save_disable(dev);
-		console_unlock();
-	}
-}
diff --git a/drivers/gpu/drm/nouveau/nouveau_fbcon.h b/drivers/gpu/drm/nouveau/nouveau_fbcon.h
index 0b465c7d3907..6208e70e4a1c 100644
--- a/drivers/gpu/drm/nouveau/nouveau_fbcon.h
+++ b/drivers/gpu/drm/nouveau/nouveau_fbcon.h
@@ -73,5 +73,8 @@ void nouveau_fbcon_accel_save_disable(struct drm_device *dev);
 void nouveau_fbcon_accel_restore(struct drm_device *dev);
 
 void nouveau_fbcon_output_poll_changed(struct drm_device *dev);
+
+extern int nouveau_nofbaccel;
+
 #endif /* __NV50_FBCON_H__ */
 
diff --git a/drivers/gpu/drm/nouveau/nouveau_fence.c b/drivers/gpu/drm/nouveau/nouveau_fence.c
index 0a93114158cd..f32a434724e3 100644
--- a/drivers/gpu/drm/nouveau/nouveau_fence.c
+++ b/drivers/gpu/drm/nouveau/nouveau_fence.c
@@ -28,6 +28,7 @@
 
 #include <linux/ktime.h>
 #include <linux/hrtimer.h>
+#include <trace/events/fence.h>
 
 #include <nvif/notify.h>
 #include <nvif/event.h>
@@ -36,123 +37,253 @@
 #include "nouveau_dma.h"
 #include "nouveau_fence.h"
 
-struct fence_work {
-	struct work_struct base;
-	struct list_head head;
-	void (*func)(void *);
-	void *data;
-};
+static const struct fence_ops nouveau_fence_ops_uevent;
+static const struct fence_ops nouveau_fence_ops_legacy;
 
-static void
+static inline struct nouveau_fence *
+from_fence(struct fence *fence)
+{
+	return container_of(fence, struct nouveau_fence, base);
+}
+
+static inline struct nouveau_fence_chan *
+nouveau_fctx(struct nouveau_fence *fence)
+{
+	return container_of(fence->base.lock, struct nouveau_fence_chan, lock);
+}
+
+static int
 nouveau_fence_signal(struct nouveau_fence *fence)
 {
-	struct fence_work *work, *temp;
+	int drop = 0;
+
+	fence_signal_locked(&fence->base);
+	list_del(&fence->head);
+	rcu_assign_pointer(fence->channel, NULL);
+
+	if (test_bit(FENCE_FLAG_USER_BITS, &fence->base.flags)) {
+		struct nouveau_fence_chan *fctx = nouveau_fctx(fence);
 
-	list_for_each_entry_safe(work, temp, &fence->work, head) {
-		schedule_work(&work->base);
-		list_del(&work->head);
+		if (!--fctx->notify_ref)
+			drop = 1;
 	}
 
-	fence->channel = NULL;
-	list_del(&fence->head);
+	fence_put(&fence->base);
+	return drop;
+}
+
+static struct nouveau_fence *
+nouveau_local_fence(struct fence *fence, struct nouveau_drm *drm) {
+	struct nouveau_fence_priv *priv = (void*)drm->fence;
+
+	if (fence->ops != &nouveau_fence_ops_legacy &&
+	    fence->ops != &nouveau_fence_ops_uevent)
+		return NULL;
+
+	if (fence->context < priv->context_base ||
+	    fence->context >= priv->context_base + priv->contexts)
+		return NULL;
+
+	return from_fence(fence);
 }
 
 void
 nouveau_fence_context_del(struct nouveau_fence_chan *fctx)
 {
-	struct nouveau_fence *fence, *fnext;
-	spin_lock(&fctx->lock);
-	list_for_each_entry_safe(fence, fnext, &fctx->pending, head) {
-		nouveau_fence_signal(fence);
+	struct nouveau_fence *fence;
+
+	spin_lock_irq(&fctx->lock);
+	while (!list_empty(&fctx->pending)) {
+		fence = list_entry(fctx->pending.next, typeof(*fence), head);
+
+		if (nouveau_fence_signal(fence))
+			nvif_notify_put(&fctx->notify);
+	}
+	spin_unlock_irq(&fctx->lock);
+
+	nvif_notify_fini(&fctx->notify);
+	fctx->dead = 1;
+
+	/*
+	 * Ensure that all accesses to fence->channel complete before freeing
+	 * the channel.
+	 */
+	synchronize_rcu();
+}
+
+static void
+nouveau_fence_context_put(struct kref *fence_ref)
+{
+	kfree(container_of(fence_ref, struct nouveau_fence_chan, fence_ref));
+}
+
+void
+nouveau_fence_context_free(struct nouveau_fence_chan *fctx)
+{
+	kref_put(&fctx->fence_ref, nouveau_fence_context_put);
+}
+
+static int
+nouveau_fence_update(struct nouveau_channel *chan, struct nouveau_fence_chan *fctx)
+{
+	struct nouveau_fence *fence;
+	int drop = 0;
+	u32 seq = fctx->read(chan);
+
+	while (!list_empty(&fctx->pending)) {
+		fence = list_entry(fctx->pending.next, typeof(*fence), head);
+
+		if ((int)(seq - fence->base.seqno) < 0)
+			break;
+
+		drop |= nouveau_fence_signal(fence);
+	}
+
+	return drop;
+}
+
+static int
+nouveau_fence_wait_uevent_handler(struct nvif_notify *notify)
+{
+	struct nouveau_fence_chan *fctx =
+		container_of(notify, typeof(*fctx), notify);
+	unsigned long flags;
+	int ret = NVIF_NOTIFY_KEEP;
+
+	spin_lock_irqsave(&fctx->lock, flags);
+	if (!list_empty(&fctx->pending)) {
+		struct nouveau_fence *fence;
+		struct nouveau_channel *chan;
+
+		fence = list_entry(fctx->pending.next, typeof(*fence), head);
+		chan = rcu_dereference_protected(fence->channel, lockdep_is_held(&fctx->lock));
+		if (nouveau_fence_update(fence->channel, fctx))
+			ret = NVIF_NOTIFY_DROP;
 	}
-	spin_unlock(&fctx->lock);
+	spin_unlock_irqrestore(&fctx->lock, flags);
+
+	return ret;
 }
 
 void
-nouveau_fence_context_new(struct nouveau_fence_chan *fctx)
+nouveau_fence_context_new(struct nouveau_channel *chan, struct nouveau_fence_chan *fctx)
 {
+	struct nouveau_fence_priv *priv = (void*)chan->drm->fence;
+	struct nouveau_cli *cli = (void *)nvif_client(chan->object);
+	int ret;
+
 	INIT_LIST_HEAD(&fctx->flip);
 	INIT_LIST_HEAD(&fctx->pending);
 	spin_lock_init(&fctx->lock);
+	fctx->context = priv->context_base + chan->chid;
+
+	if (chan == chan->drm->cechan)
+		strcpy(fctx->name, "copy engine channel");
+	else if (chan == chan->drm->channel)
+		strcpy(fctx->name, "generic kernel channel");
+	else
+		strcpy(fctx->name, nvkm_client(&cli->base)->name);
+
+	kref_init(&fctx->fence_ref);
+	if (!priv->uevent)
+		return;
+
+	ret = nvif_notify_init(chan->object, NULL,
+			 nouveau_fence_wait_uevent_handler, false,
+			 G82_CHANNEL_DMA_V0_NTFY_UEVENT,
+			 &(struct nvif_notify_uevent_req) { },
+			 sizeof(struct nvif_notify_uevent_req),
+			 sizeof(struct nvif_notify_uevent_rep),
+			 &fctx->notify);
+
+	WARN_ON(ret);
 }
 
+struct nouveau_fence_work {
+	struct work_struct work;
+	struct fence_cb cb;
+	void (*func)(void *);
+	void *data;
+};
+
 static void
 nouveau_fence_work_handler(struct work_struct *kwork)
 {
-	struct fence_work *work = container_of(kwork, typeof(*work), base);
+	struct nouveau_fence_work *work = container_of(kwork, typeof(*work), work);
 	work->func(work->data);
 	kfree(work);
 }
 
+static void nouveau_fence_work_cb(struct fence *fence, struct fence_cb *cb)
+{
+	struct nouveau_fence_work *work = container_of(cb, typeof(*work), cb);
+
+	schedule_work(&work->work);
+}
+
 void
-nouveau_fence_work(struct nouveau_fence *fence,
+nouveau_fence_work(struct fence *fence,
 		   void (*func)(void *), void *data)
 {
-	struct nouveau_channel *chan = fence->channel;
-	struct nouveau_fence_chan *fctx;
-	struct fence_work *work = NULL;
+	struct nouveau_fence_work *work;
 
-	if (nouveau_fence_done(fence)) {
-		func(data);
-		return;
-	}
+	if (fence_is_signaled(fence))
+		goto err;
 
-	fctx = chan->fence;
 	work = kmalloc(sizeof(*work), GFP_KERNEL);
 	if (!work) {
-		WARN_ON(nouveau_fence_wait(fence, false, false));
-		func(data);
-		return;
+		/*
+		 * this might not be a nouveau fence any more,
+		 * so force a lazy wait here
+		 */
+		WARN_ON(nouveau_fence_wait((struct nouveau_fence *)fence,
+					   true, false));
+		goto err;
 	}
 
-	spin_lock(&fctx->lock);
-	if (!fence->channel) {
-		spin_unlock(&fctx->lock);
-		kfree(work);
-		func(data);
-		return;
-	}
-
-	INIT_WORK(&work->base, nouveau_fence_work_handler);
+	INIT_WORK(&work->work, nouveau_fence_work_handler);
 	work->func = func;
 	work->data = data;
-	list_add(&work->head, &fence->work);
-	spin_unlock(&fctx->lock);
-}
-
-static void
-nouveau_fence_update(struct nouveau_channel *chan)
-{
-	struct nouveau_fence_chan *fctx = chan->fence;
-	struct nouveau_fence *fence, *fnext;
 
-	spin_lock(&fctx->lock);
-	list_for_each_entry_safe(fence, fnext, &fctx->pending, head) {
-		if (fctx->read(chan) < fence->sequence)
-			break;
+	if (fence_add_callback(fence, &work->cb, nouveau_fence_work_cb) < 0)
+		goto err_free;
+	return;
 
-		nouveau_fence_signal(fence);
-		nouveau_fence_unref(&fence);
-	}
-	spin_unlock(&fctx->lock);
+err_free:
+	kfree(work);
+err:
+	func(data);
 }
 
 int
 nouveau_fence_emit(struct nouveau_fence *fence, struct nouveau_channel *chan)
 {
 	struct nouveau_fence_chan *fctx = chan->fence;
+	struct nouveau_fence_priv *priv = (void*)chan->drm->fence;
 	int ret;
 
 	fence->channel  = chan;
 	fence->timeout  = jiffies + (15 * HZ);
-	fence->sequence = ++fctx->sequence;
 
+	if (priv->uevent)
+		fence_init(&fence->base, &nouveau_fence_ops_uevent,
+			   &fctx->lock, fctx->context, ++fctx->sequence);
+	else
+		fence_init(&fence->base, &nouveau_fence_ops_legacy,
+			   &fctx->lock, fctx->context, ++fctx->sequence);
+	kref_get(&fctx->fence_ref);
+
+	trace_fence_emit(&fence->base);
 	ret = fctx->emit(fence);
 	if (!ret) {
-		kref_get(&fence->kref);
-		spin_lock(&fctx->lock);
+		fence_get(&fence->base);
+		spin_lock_irq(&fctx->lock);
+
+		if (nouveau_fence_update(chan, fctx))
+			nvif_notify_put(&fctx->notify);
+
 		list_add_tail(&fence->head, &fctx->pending);
-		spin_unlock(&fctx->lock);
+		spin_unlock_irq(&fctx->lock);
 	}
 
 	return ret;
@@ -161,114 +292,73 @@ nouveau_fence_emit(struct nouveau_fence *fence, struct nouveau_channel *chan)
 bool
 nouveau_fence_done(struct nouveau_fence *fence)
 {
-	if (fence->channel)
-		nouveau_fence_update(fence->channel);
-	return !fence->channel;
+	if (fence->base.ops == &nouveau_fence_ops_legacy ||
+	    fence->base.ops == &nouveau_fence_ops_uevent) {
+		struct nouveau_fence_chan *fctx = nouveau_fctx(fence);
+		struct nouveau_channel *chan;
+		unsigned long flags;
+
+		if (test_bit(FENCE_FLAG_SIGNALED_BIT, &fence->base.flags))
+			return true;
+
+		spin_lock_irqsave(&fctx->lock, flags);
+		chan = rcu_dereference_protected(fence->channel, lockdep_is_held(&fctx->lock));
+		if (chan && nouveau_fence_update(chan, fctx))
+			nvif_notify_put(&fctx->notify);
+		spin_unlock_irqrestore(&fctx->lock, flags);
+	}
+	return fence_is_signaled(&fence->base);
 }
 
-struct nouveau_fence_wait {
-	struct nouveau_fence_priv *priv;
-	struct nvif_notify notify;
-};
-
-static int
-nouveau_fence_wait_uevent_handler(struct nvif_notify *notify)
+static long
+nouveau_fence_wait_legacy(struct fence *f, bool intr, long wait)
 {
-	struct nouveau_fence_wait *wait =
-		container_of(notify, typeof(*wait), notify);
-	wake_up_all(&wait->priv->waiting);
-	return NVIF_NOTIFY_KEEP;
-}
-
-static int
-nouveau_fence_wait_uevent(struct nouveau_fence *fence, bool intr)
+	struct nouveau_fence *fence = from_fence(f);
+	unsigned long sleep_time = NSEC_PER_MSEC / 1000;
+	unsigned long t = jiffies, timeout = t + wait;
 
-{
-	struct nouveau_channel *chan = fence->channel;
-	struct nouveau_fence_priv *priv = chan->drm->fence;
-	struct nouveau_fence_wait wait = { .priv = priv };
-	int ret = 0;
+	while (!nouveau_fence_done(fence)) {
+		ktime_t kt;
 
-	ret = nvif_notify_init(chan->object, NULL,
-			       nouveau_fence_wait_uevent_handler, false,
-			       G82_CHANNEL_DMA_V0_NTFY_UEVENT,
-			       &(struct nvif_notify_uevent_req) {
-			       },
-			       sizeof(struct nvif_notify_uevent_req),
-			       sizeof(struct nvif_notify_uevent_rep),
-			       &wait.notify);
-	if (ret)
-		return ret;
+		t = jiffies;
 
-	nvif_notify_get(&wait.notify);
-
-	if (fence->timeout) {
-		unsigned long timeout = fence->timeout - jiffies;
-
-		if (time_before(jiffies, fence->timeout)) {
-			if (intr) {
-				ret = wait_event_interruptible_timeout(
-						priv->waiting,
-						nouveau_fence_done(fence),
-						timeout);
-			} else {
-				ret = wait_event_timeout(priv->waiting,
-						nouveau_fence_done(fence),
-						timeout);
-			}
+		if (wait != MAX_SCHEDULE_TIMEOUT && time_after_eq(t, timeout)) {
+			__set_current_state(TASK_RUNNING);
+			return 0;
 		}
 
-		if (ret >= 0) {
-			fence->timeout = jiffies + ret;
-			if (time_after_eq(jiffies, fence->timeout))
-				ret = -EBUSY;
-		}
-	} else {
-		if (intr) {
-			ret = wait_event_interruptible(priv->waiting,
-					nouveau_fence_done(fence));
-		} else {
-			wait_event(priv->waiting, nouveau_fence_done(fence));
-		}
+		__set_current_state(intr ? TASK_INTERRUPTIBLE :
+					   TASK_UNINTERRUPTIBLE);
+
+		kt = ktime_set(0, sleep_time);
+		schedule_hrtimeout(&kt, HRTIMER_MODE_REL);
+		sleep_time *= 2;
+		if (sleep_time > NSEC_PER_MSEC)
+			sleep_time = NSEC_PER_MSEC;
+
+		if (intr && signal_pending(current))
+			return -ERESTARTSYS;
 	}
 
-	nvif_notify_fini(&wait.notify);
-	if (unlikely(ret < 0))
-		return ret;
+	__set_current_state(TASK_RUNNING);
 
-	return 0;
+	return timeout - t;
 }
 
-int
-nouveau_fence_wait(struct nouveau_fence *fence, bool lazy, bool intr)
+static int
+nouveau_fence_wait_busy(struct nouveau_fence *fence, bool intr)
 {
-	struct nouveau_channel *chan = fence->channel;
-	struct nouveau_fence_priv *priv = chan ? chan->drm->fence : NULL;
-	unsigned long sleep_time = NSEC_PER_MSEC / 1000;
-	ktime_t t;
 	int ret = 0;
 
-	while (priv && priv->uevent && lazy && !nouveau_fence_done(fence)) {
-		ret = nouveau_fence_wait_uevent(fence, intr);
-		if (ret < 0)
-			return ret;
-	}
-
 	while (!nouveau_fence_done(fence)) {
-		if (fence->timeout && time_after_eq(jiffies, fence->timeout)) {
+		if (time_after_eq(jiffies, fence->timeout)) {
 			ret = -EBUSY;
 			break;
 		}
 
-		__set_current_state(intr ? TASK_INTERRUPTIBLE :
-					   TASK_UNINTERRUPTIBLE);
-		if (lazy) {
-			t = ktime_set(0, sleep_time);
-			schedule_hrtimeout(&t, HRTIMER_MODE_REL);
-			sleep_time *= 2;
-			if (sleep_time > NSEC_PER_MSEC)
-				sleep_time = NSEC_PER_MSEC;
-		}
+		__set_current_state(intr ?
+				    TASK_INTERRUPTIBLE :
+				    TASK_UNINTERRUPTIBLE);
 
 		if (intr && signal_pending(current)) {
 			ret = -ERESTARTSYS;
@@ -281,47 +371,95 @@ nouveau_fence_wait(struct nouveau_fence *fence, bool lazy, bool intr)
 }
 
 int
-nouveau_fence_sync(struct nouveau_fence *fence, struct nouveau_channel *chan)
+nouveau_fence_wait(struct nouveau_fence *fence, bool lazy, bool intr)
+{
+	long ret;
+
+	if (!lazy)
+		return nouveau_fence_wait_busy(fence, intr);
+
+	ret = fence_wait_timeout(&fence->base, intr, 15 * HZ);
+	if (ret < 0)
+		return ret;
+	else if (!ret)
+		return -EBUSY;
+	else
+		return 0;
+}
+
+int
+nouveau_fence_sync(struct nouveau_bo *nvbo, struct nouveau_channel *chan, bool exclusive, bool intr)
 {
 	struct nouveau_fence_chan *fctx = chan->fence;
-	struct nouveau_channel *prev;
-	int ret = 0;
+	struct fence *fence;
+	struct reservation_object *resv = nvbo->bo.resv;
+	struct reservation_object_list *fobj;
+	struct nouveau_fence *f;
+	int ret = 0, i;
 
-	prev = fence ? fence->channel : NULL;
-	if (prev) {
-		if (unlikely(prev != chan && !nouveau_fence_done(fence))) {
-			ret = fctx->sync(fence, prev, chan);
-			if (unlikely(ret))
-				ret = nouveau_fence_wait(fence, true, false);
+	if (!exclusive) {
+		ret = reservation_object_reserve_shared(resv);
+
+		if (ret)
+			return ret;
+	}
+
+	fobj = reservation_object_get_list(resv);
+	fence = reservation_object_get_excl(resv);
+
+	if (fence && (!exclusive || !fobj || !fobj->shared_count)) {
+		struct nouveau_channel *prev = NULL;
+		bool must_wait = true;
+
+		f = nouveau_local_fence(fence, chan->drm);
+		if (f) {
+			rcu_read_lock();
+			prev = rcu_dereference(f->channel);
+			if (prev && (prev == chan || fctx->sync(f, prev, chan) == 0))
+				must_wait = false;
+			rcu_read_unlock();
 		}
+
+		if (must_wait)
+			ret = fence_wait(fence, intr);
+
+		return ret;
 	}
 
-	return ret;
-}
+	if (!exclusive || !fobj)
+		return ret;
 
-static void
-nouveau_fence_del(struct kref *kref)
-{
-	struct nouveau_fence *fence = container_of(kref, typeof(*fence), kref);
-	kfree(fence);
+	for (i = 0; i < fobj->shared_count && !ret; ++i) {
+		struct nouveau_channel *prev = NULL;
+		bool must_wait = true;
+
+		fence = rcu_dereference_protected(fobj->shared[i],
+						reservation_object_held(resv));
+
+		f = nouveau_local_fence(fence, chan->drm);
+		if (f) {
+			rcu_read_lock();
+			prev = rcu_dereference(f->channel);
+			if (prev && (prev == chan || fctx->sync(f, prev, chan) == 0))
+				must_wait = false;
+			rcu_read_unlock();
+		}
+
+		if (must_wait)
+			ret = fence_wait(fence, intr);
+	}
+
+	return ret;
 }
 
 void
 nouveau_fence_unref(struct nouveau_fence **pfence)
 {
 	if (*pfence)
-		kref_put(&(*pfence)->kref, nouveau_fence_del);
+		fence_put(&(*pfence)->base);
 	*pfence = NULL;
 }
 
-struct nouveau_fence *
-nouveau_fence_ref(struct nouveau_fence *fence)
-{
-	if (fence)
-		kref_get(&fence->kref);
-	return fence;
-}
-
 int
 nouveau_fence_new(struct nouveau_channel *chan, bool sysmem,
 		  struct nouveau_fence **pfence)
@@ -336,9 +474,7 @@ nouveau_fence_new(struct nouveau_channel *chan, bool sysmem,
 	if (!fence)
 		return -ENOMEM;
 
-	INIT_LIST_HEAD(&fence->work);
 	fence->sysmem = sysmem;
-	kref_init(&fence->kref);
 
 	ret = nouveau_fence_emit(fence, chan);
 	if (ret)
@@ -347,3 +483,108 @@ nouveau_fence_new(struct nouveau_channel *chan, bool sysmem,
 	*pfence = fence;
 	return ret;
 }
+
+static const char *nouveau_fence_get_get_driver_name(struct fence *fence)
+{
+	return "nouveau";
+}
+
+static const char *nouveau_fence_get_timeline_name(struct fence *f)
+{
+	struct nouveau_fence *fence = from_fence(f);
+	struct nouveau_fence_chan *fctx = nouveau_fctx(fence);
+
+	return !fctx->dead ? fctx->name : "dead channel";
+}
+
+/*
+ * In an ideal world, read would not assume the channel context is still alive.
+ * This function may be called from another device, running into free memory as a
+ * result. The drm node should still be there, so we can derive the index from
+ * the fence context.
+ */
+static bool nouveau_fence_is_signaled(struct fence *f)
+{
+	struct nouveau_fence *fence = from_fence(f);
+	struct nouveau_fence_chan *fctx = nouveau_fctx(fence);
+	struct nouveau_channel *chan;
+	bool ret = false;
+
+	rcu_read_lock();
+	chan = rcu_dereference(fence->channel);
+	if (chan)
+		ret = (int)(fctx->read(chan) - fence->base.seqno) >= 0;
+	rcu_read_unlock();
+
+	return ret;
+}
+
+static bool nouveau_fence_no_signaling(struct fence *f)
+{
+	struct nouveau_fence *fence = from_fence(f);
+
+	/*
+	 * caller should have a reference on the fence,
+	 * else fence could get freed here
+	 */
+	WARN_ON(atomic_read(&fence->base.refcount.refcount) <= 1);
+
+	/*
+	 * This needs uevents to work correctly, but fence_add_callback relies on
+	 * being able to enable signaling. It will still get signaled eventually,
+	 * just not right away.
+	 */
+	if (nouveau_fence_is_signaled(f)) {
+		list_del(&fence->head);
+
+		fence_put(&fence->base);
+		return false;
+	}
+
+	return true;
+}
+
+static void nouveau_fence_release(struct fence *f)
+{
+	struct nouveau_fence *fence = from_fence(f);
+	struct nouveau_fence_chan *fctx = nouveau_fctx(fence);
+
+	kref_put(&fctx->fence_ref, nouveau_fence_context_put);
+	fence_free(&fence->base);
+}
+
+static const struct fence_ops nouveau_fence_ops_legacy = {
+	.get_driver_name = nouveau_fence_get_get_driver_name,
+	.get_timeline_name = nouveau_fence_get_timeline_name,
+	.enable_signaling = nouveau_fence_no_signaling,
+	.signaled = nouveau_fence_is_signaled,
+	.wait = nouveau_fence_wait_legacy,
+	.release = nouveau_fence_release
+};
+
+static bool nouveau_fence_enable_signaling(struct fence *f)
+{
+	struct nouveau_fence *fence = from_fence(f);
+	struct nouveau_fence_chan *fctx = nouveau_fctx(fence);
+	bool ret;
+
+	if (!fctx->notify_ref++)
+		nvif_notify_get(&fctx->notify);
+
+	ret = nouveau_fence_no_signaling(f);
+	if (ret)
+		set_bit(FENCE_FLAG_USER_BITS, &fence->base.flags);
+	else if (!--fctx->notify_ref)
+		nvif_notify_put(&fctx->notify);
+
+	return ret;
+}
+
+static const struct fence_ops nouveau_fence_ops_uevent = {
+	.get_driver_name = nouveau_fence_get_get_driver_name,
+	.get_timeline_name = nouveau_fence_get_timeline_name,
+	.enable_signaling = nouveau_fence_enable_signaling,
+	.signaled = nouveau_fence_is_signaled,
+	.wait = fence_default_wait,
+	.release = NULL
+};
diff --git a/drivers/gpu/drm/nouveau/nouveau_fence.h b/drivers/gpu/drm/nouveau/nouveau_fence.h
index c57bb61da58c..96e461c6f68f 100644
--- a/drivers/gpu/drm/nouveau/nouveau_fence.h
+++ b/drivers/gpu/drm/nouveau/nouveau_fence.h
@@ -1,33 +1,37 @@
 #ifndef __NOUVEAU_FENCE_H__
 #define __NOUVEAU_FENCE_H__
 
+#include <linux/fence.h>
+#include <nvif/notify.h>
+
 struct nouveau_drm;
+struct nouveau_bo;
 
 struct nouveau_fence {
+	struct fence base;
+
 	struct list_head head;
-	struct list_head work;
-	struct kref kref;
 
 	bool sysmem;
 
-	struct nouveau_channel *channel;
+	struct nouveau_channel __rcu *channel;
 	unsigned long timeout;
-	u32 sequence;
 };
 
 int  nouveau_fence_new(struct nouveau_channel *, bool sysmem,
 		       struct nouveau_fence **);
-struct nouveau_fence *
-nouveau_fence_ref(struct nouveau_fence *);
 void nouveau_fence_unref(struct nouveau_fence **);
 
 int  nouveau_fence_emit(struct nouveau_fence *, struct nouveau_channel *);
 bool nouveau_fence_done(struct nouveau_fence *);
-void nouveau_fence_work(struct nouveau_fence *, void (*)(void *), void *);
+void nouveau_fence_work(struct fence *, void (*)(void *), void *);
 int  nouveau_fence_wait(struct nouveau_fence *, bool lazy, bool intr);
-int  nouveau_fence_sync(struct nouveau_fence *, struct nouveau_channel *);
+int  nouveau_fence_sync(struct nouveau_bo *, struct nouveau_channel *, bool exclusive, bool intr);
 
 struct nouveau_fence_chan {
+	spinlock_t lock;
+	struct kref fence_ref;
+
 	struct list_head pending;
 	struct list_head flip;
 
@@ -38,8 +42,12 @@ struct nouveau_fence_chan {
 	int  (*emit32)(struct nouveau_channel *, u64, u32);
 	int  (*sync32)(struct nouveau_channel *, u64, u32);
 
-	spinlock_t lock;
 	u32 sequence;
+	u32 context;
+	char name[32];
+
+	struct nvif_notify notify;
+	int notify_ref, dead;
 };
 
 struct nouveau_fence_priv {
@@ -49,14 +57,15 @@ struct nouveau_fence_priv {
 	int  (*context_new)(struct nouveau_channel *);
 	void (*context_del)(struct nouveau_channel *);
 
-	wait_queue_head_t waiting;
+	u32 contexts, context_base;
 	bool uevent;
 };
 
 #define nouveau_fence(drm) ((struct nouveau_fence_priv *)(drm)->fence)
 
-void nouveau_fence_context_new(struct nouveau_fence_chan *);
+void nouveau_fence_context_new(struct nouveau_channel *, struct nouveau_fence_chan *);
 void nouveau_fence_context_del(struct nouveau_fence_chan *);
+void nouveau_fence_context_free(struct nouveau_fence_chan *);
 
 int nv04_fence_create(struct nouveau_drm *);
 int nv04_fence_mthd(struct nouveau_channel *, u32, u32, u32);
diff --git a/drivers/gpu/drm/nouveau/nouveau_gem.c b/drivers/gpu/drm/nouveau/nouveau_gem.c
index 292a677bfed4..bf0f9e21d714 100644
--- a/drivers/gpu/drm/nouveau/nouveau_gem.c
+++ b/drivers/gpu/drm/nouveau/nouveau_gem.c
@@ -36,7 +36,14 @@ void
 nouveau_gem_object_del(struct drm_gem_object *gem)
 {
 	struct nouveau_bo *nvbo = nouveau_gem_object(gem);
+	struct nouveau_drm *drm = nouveau_bdev(nvbo->bo.bdev);
 	struct ttm_buffer_object *bo = &nvbo->bo;
+	struct device *dev = drm->dev->dev;
+	int ret;
+
+	ret = pm_runtime_get_sync(dev);
+	if (WARN_ON(ret < 0 && ret != -EACCES))
+		return;
 
 	if (gem->import_attach)
 		drm_prime_gem_destroy(gem, nvbo->bo.sg);
@@ -46,6 +53,9 @@ nouveau_gem_object_del(struct drm_gem_object *gem)
 	/* reset filp so nouveau_bo_del_ttm() can test for it */
 	gem->filp = NULL;
 	ttm_bo_unref(&bo);
+
+	pm_runtime_mark_last_busy(dev);
+	pm_runtime_put_autosuspend(dev);
 }
 
 int
@@ -53,7 +63,9 @@ nouveau_gem_object_open(struct drm_gem_object *gem, struct drm_file *file_priv)
 {
 	struct nouveau_cli *cli = nouveau_cli(file_priv);
 	struct nouveau_bo *nvbo = nouveau_gem_object(gem);
+	struct nouveau_drm *drm = nouveau_bdev(nvbo->bo.bdev);
 	struct nouveau_vma *vma;
+	struct device *dev = drm->dev->dev;
 	int ret;
 
 	if (!cli->vm)
@@ -71,11 +83,16 @@ nouveau_gem_object_open(struct drm_gem_object *gem, struct drm_file *file_priv)
 			goto out;
 		}
 
+		ret = pm_runtime_get_sync(dev);
+		if (ret < 0 && ret != -EACCES)
+			goto out;
+
 		ret = nouveau_bo_vma_add(nvbo, cli->vm, vma);
-		if (ret) {
+		if (ret)
 			kfree(vma);
-			goto out;
-		}
+
+		pm_runtime_mark_last_busy(dev);
+		pm_runtime_put_autosuspend(dev);
 	} else {
 		vma->refcount++;
 	}
@@ -98,17 +115,23 @@ static void
 nouveau_gem_object_unmap(struct nouveau_bo *nvbo, struct nouveau_vma *vma)
 {
 	const bool mapped = nvbo->bo.mem.mem_type != TTM_PL_SYSTEM;
-	struct nouveau_fence *fence = NULL;
+	struct reservation_object *resv = nvbo->bo.resv;
+	struct reservation_object_list *fobj;
+	struct fence *fence = NULL;
+
+	fobj = reservation_object_get_list(resv);
 
 	list_del(&vma->head);
 
-	if (mapped) {
-		spin_lock(&nvbo->bo.bdev->fence_lock);
-		fence = nouveau_fence_ref(nvbo->bo.sync_obj);
-		spin_unlock(&nvbo->bo.bdev->fence_lock);
-	}
+	if (fobj && fobj->shared_count > 1)
+		ttm_bo_wait(&nvbo->bo, true, false, false);
+	else if (fobj && fobj->shared_count == 1)
+		fence = rcu_dereference_protected(fobj->shared[0],
+						reservation_object_held(resv));
+	else
+		fence = reservation_object_get_excl(nvbo->bo.resv);
 
-	if (fence) {
+	if (fence && mapped) {
 		nouveau_fence_work(fence, nouveau_gem_object_delete, vma);
 	} else {
 		if (mapped)
@@ -116,7 +139,6 @@ nouveau_gem_object_unmap(struct nouveau_bo *nvbo, struct nouveau_vma *vma)
 		nouveau_vm_put(vma);
 		kfree(vma);
 	}
-	nouveau_fence_unref(&fence);
 }
 
 void
@@ -124,6 +146,8 @@ nouveau_gem_object_close(struct drm_gem_object *gem, struct drm_file *file_priv)
 {
 	struct nouveau_cli *cli = nouveau_cli(file_priv);
 	struct nouveau_bo *nvbo = nouveau_gem_object(gem);
+	struct nouveau_drm *drm = nouveau_bdev(nvbo->bo.bdev);
+	struct device *dev = drm->dev->dev;
 	struct nouveau_vma *vma;
 	int ret;
 
@@ -136,8 +160,14 @@ nouveau_gem_object_close(struct drm_gem_object *gem, struct drm_file *file_priv)
 
 	vma = nouveau_bo_vma_find(nvbo, cli->vm);
 	if (vma) {
-		if (--vma->refcount == 0)
-			nouveau_gem_object_unmap(nvbo, vma);
+		if (--vma->refcount == 0) {
+			ret = pm_runtime_get_sync(dev);
+			if (!WARN_ON(ret < 0 && ret != -EACCES)) {
+				nouveau_gem_object_unmap(nvbo, vma);
+				pm_runtime_mark_last_busy(dev);
+				pm_runtime_put_autosuspend(dev);
+			}
+		}
 	}
 	ttm_bo_unreserve(&nvbo->bo);
 }
@@ -160,7 +190,7 @@ nouveau_gem_new(struct drm_device *dev, int size, int align, uint32_t domain,
 		flags |= TTM_PL_FLAG_SYSTEM;
 
 	ret = nouveau_bo_new(dev, size, align, flags, tile_mode,
-			     tile_flags, NULL, pnvbo);
+			     tile_flags, NULL, NULL, pnvbo);
 	if (ret)
 		return ret;
 	nvbo = *pnvbo;
@@ -288,24 +318,23 @@ nouveau_gem_set_domain(struct drm_gem_object *gem, uint32_t read_domains,
 }
 
 struct validate_op {
-	struct list_head vram_list;
-	struct list_head gart_list;
-	struct list_head both_list;
+	struct list_head list;
 	struct ww_acquire_ctx ticket;
 };
 
 static void
-validate_fini_list(struct list_head *list, struct nouveau_fence *fence,
-		   struct ww_acquire_ctx *ticket)
+validate_fini_no_ticket(struct validate_op *op, struct nouveau_fence *fence,
+			struct drm_nouveau_gem_pushbuf_bo *pbbo)
 {
-	struct list_head *entry, *tmp;
 	struct nouveau_bo *nvbo;
+	struct drm_nouveau_gem_pushbuf_bo *b;
 
-	list_for_each_safe(entry, tmp, list) {
-		nvbo = list_entry(entry, struct nouveau_bo, entry);
+	while (!list_empty(&op->list)) {
+		nvbo = list_entry(op->list.next, struct nouveau_bo, entry);
+		b = &pbbo[nvbo->pbbo_index];
 
 		if (likely(fence))
-			nouveau_bo_fence(nvbo, fence);
+			nouveau_bo_fence(nvbo, fence, !!b->write_domains);
 
 		if (unlikely(nvbo->validate_mapped)) {
 			ttm_bo_kunmap(&nvbo->kmap);
@@ -314,23 +343,16 @@ validate_fini_list(struct list_head *list, struct nouveau_fence *fence,
 
 		list_del(&nvbo->entry);
 		nvbo->reserved_by = NULL;
-		ttm_bo_unreserve_ticket(&nvbo->bo, ticket);
+		ttm_bo_unreserve_ticket(&nvbo->bo, &op->ticket);
 		drm_gem_object_unreference_unlocked(&nvbo->gem);
 	}
 }
 
 static void
-validate_fini_no_ticket(struct validate_op *op, struct nouveau_fence *fence)
+validate_fini(struct validate_op *op, struct nouveau_fence *fence,
+	      struct drm_nouveau_gem_pushbuf_bo *pbbo)
 {
-	validate_fini_list(&op->vram_list, fence, &op->ticket);
-	validate_fini_list(&op->gart_list, fence, &op->ticket);
-	validate_fini_list(&op->both_list, fence, &op->ticket);
-}
-
-static void
-validate_fini(struct validate_op *op, struct nouveau_fence *fence)
-{
-	validate_fini_no_ticket(op, fence);
+	validate_fini_no_ticket(op, fence, pbbo);
 	ww_acquire_fini(&op->ticket);
 }
 
@@ -344,6 +366,9 @@ validate_init(struct nouveau_channel *chan, struct drm_file *file_priv,
 	int trycnt = 0;
 	int ret, i;
 	struct nouveau_bo *res_bo = NULL;
+	LIST_HEAD(gart_list);
+	LIST_HEAD(vram_list);
+	LIST_HEAD(both_list);
 
 	ww_acquire_init(&op->ticket, &reservation_ww_class);
 retry:
@@ -360,9 +385,8 @@ retry:
 		gem = drm_gem_object_lookup(dev, file_priv, b->handle);
 		if (!gem) {
 			NV_PRINTK(error, cli, "Unknown handle 0x%08x\n", b->handle);
-			ww_acquire_done(&op->ticket);
-			validate_fini(op, NULL);
-			return -ENOENT;
+			ret = -ENOENT;
+			break;
 		}
 		nvbo = nouveau_gem_object(gem);
 		if (nvbo == res_bo) {
@@ -375,14 +399,16 @@ retry:
 			NV_PRINTK(error, cli, "multiple instances of buffer %d on "
 				      "validation list\n", b->handle);
 			drm_gem_object_unreference_unlocked(gem);
-			ww_acquire_done(&op->ticket);
-			validate_fini(op, NULL);
-			return -EINVAL;
+			ret = -EINVAL;
+			break;
 		}
 
 		ret = ttm_bo_reserve(&nvbo->bo, true, false, true, &op->ticket);
 		if (ret) {
-			validate_fini_no_ticket(op, NULL);
+			list_splice_tail_init(&vram_list, &op->list);
+			list_splice_tail_init(&gart_list, &op->list);
+			list_splice_tail_init(&both_list, &op->list);
+			validate_fini_no_ticket(op, NULL, NULL);
 			if (unlikely(ret == -EDEADLK)) {
 				ret = ttm_bo_reserve_slowpath(&nvbo->bo, true,
 							      &op->ticket);
@@ -390,12 +416,9 @@ retry:
 					res_bo = nvbo;
 			}
 			if (unlikely(ret)) {
-				ww_acquire_done(&op->ticket);
-				ww_acquire_fini(&op->ticket);
-				drm_gem_object_unreference_unlocked(gem);
 				if (ret != -ERESTARTSYS)
 					NV_PRINTK(error, cli, "fail reserve\n");
-				return ret;
+				break;
 			}
 		}
 
@@ -404,45 +427,32 @@ retry:
 		nvbo->pbbo_index = i;
 		if ((b->valid_domains & NOUVEAU_GEM_DOMAIN_VRAM) &&
 		    (b->valid_domains & NOUVEAU_GEM_DOMAIN_GART))
-			list_add_tail(&nvbo->entry, &op->both_list);
+			list_add_tail(&nvbo->entry, &both_list);
 		else
 		if (b->valid_domains & NOUVEAU_GEM_DOMAIN_VRAM)
-			list_add_tail(&nvbo->entry, &op->vram_list);
+			list_add_tail(&nvbo->entry, &vram_list);
 		else
 		if (b->valid_domains & NOUVEAU_GEM_DOMAIN_GART)
-			list_add_tail(&nvbo->entry, &op->gart_list);
+			list_add_tail(&nvbo->entry, &gart_list);
 		else {
 			NV_PRINTK(error, cli, "invalid valid domains: 0x%08x\n",
 				 b->valid_domains);
-			list_add_tail(&nvbo->entry, &op->both_list);
-			ww_acquire_done(&op->ticket);
-			validate_fini(op, NULL);
-			return -EINVAL;
+			list_add_tail(&nvbo->entry, &both_list);
+			ret = -EINVAL;
+			break;
 		}
 		if (nvbo == res_bo)
 			goto retry;
 	}
 
 	ww_acquire_done(&op->ticket);
-	return 0;
-}
-
-static int
-validate_sync(struct nouveau_channel *chan, struct nouveau_bo *nvbo)
-{
-	struct nouveau_fence *fence = NULL;
-	int ret = 0;
-
-	spin_lock(&nvbo->bo.bdev->fence_lock);
-	fence = nouveau_fence_ref(nvbo->bo.sync_obj);
-	spin_unlock(&nvbo->bo.bdev->fence_lock);
-
-	if (fence) {
-		ret = nouveau_fence_sync(fence, chan);
-		nouveau_fence_unref(&fence);
-	}
-
+	list_splice_tail(&vram_list, &op->list);
+	list_splice_tail(&gart_list, &op->list);
+	list_splice_tail(&both_list, &op->list);
+	if (ret)
+		validate_fini(op, NULL, NULL);
 	return ret;
+
 }
 
 static int
@@ -474,9 +484,10 @@ validate_list(struct nouveau_channel *chan, struct nouveau_cli *cli,
 			return ret;
 		}
 
-		ret = validate_sync(chan, nvbo);
+		ret = nouveau_fence_sync(nvbo, chan, !!b->write_domains, true);
 		if (unlikely(ret)) {
-			NV_PRINTK(error, cli, "fail post-validate sync\n");
+			if (ret != -ERESTARTSYS)
+				NV_PRINTK(error, cli, "fail post-validate sync\n");
 			return ret;
 		}
 
@@ -513,11 +524,9 @@ nouveau_gem_pushbuf_validate(struct nouveau_channel *chan,
 			     struct validate_op *op, int *apply_relocs)
 {
 	struct nouveau_cli *cli = nouveau_cli(file_priv);
-	int ret, relocs = 0;
+	int ret;
 
-	INIT_LIST_HEAD(&op->vram_list);
-	INIT_LIST_HEAD(&op->gart_list);
-	INIT_LIST_HEAD(&op->both_list);
+	INIT_LIST_HEAD(&op->list);
 
 	if (nr_buffers == 0)
 		return 0;
@@ -529,34 +538,14 @@ nouveau_gem_pushbuf_validate(struct nouveau_channel *chan,
 		return ret;
 	}
 
-	ret = validate_list(chan, cli, &op->vram_list, pbbo, user_buffers);
+	ret = validate_list(chan, cli, &op->list, pbbo, user_buffers);
 	if (unlikely(ret < 0)) {
 		if (ret != -ERESTARTSYS)
-			NV_PRINTK(error, cli, "validate vram_list\n");
-		validate_fini(op, NULL);
+			NV_PRINTK(error, cli, "validating bo list\n");
+		validate_fini(op, NULL, NULL);
 		return ret;
 	}
-	relocs += ret;
-
-	ret = validate_list(chan, cli, &op->gart_list, pbbo, user_buffers);
-	if (unlikely(ret < 0)) {
-		if (ret != -ERESTARTSYS)
-			NV_PRINTK(error, cli, "validate gart_list\n");
-		validate_fini(op, NULL);
-		return ret;
-	}
-	relocs += ret;
-
-	ret = validate_list(chan, cli, &op->both_list, pbbo, user_buffers);
-	if (unlikely(ret < 0)) {
-		if (ret != -ERESTARTSYS)
-			NV_PRINTK(error, cli, "validate both_list\n");
-		validate_fini(op, NULL);
-		return ret;
-	}
-	relocs += ret;
-
-	*apply_relocs = relocs;
+	*apply_relocs = ret;
 	return 0;
 }
 
@@ -659,9 +648,7 @@ nouveau_gem_pushbuf_reloc_apply(struct nouveau_cli *cli,
 				data |= r->vor;
 		}
 
-		spin_lock(&nvbo->bo.bdev->fence_lock);
-		ret = ttm_bo_wait(&nvbo->bo, false, false, false);
-		spin_unlock(&nvbo->bo.bdev->fence_lock);
+		ret = ttm_bo_wait(&nvbo->bo, true, false, false);
 		if (ret) {
 			NV_PRINTK(error, cli, "reloc wait_idle failed: %d\n", ret);
 			break;
@@ -839,7 +826,7 @@ nouveau_gem_ioctl_pushbuf(struct drm_device *dev, void *data,
 	}
 
 out:
-	validate_fini(&op, fence);
+	validate_fini(&op, fence, bo);
 	nouveau_fence_unref(&fence);
 
 out_prevalid:
@@ -884,17 +871,30 @@ nouveau_gem_ioctl_cpu_prep(struct drm_device *dev, void *data,
 	struct drm_gem_object *gem;
 	struct nouveau_bo *nvbo;
 	bool no_wait = !!(req->flags & NOUVEAU_GEM_CPU_PREP_NOWAIT);
-	int ret = -EINVAL;
+	bool write = !!(req->flags & NOUVEAU_GEM_CPU_PREP_WRITE);
+	int ret;
 
 	gem = drm_gem_object_lookup(dev, file_priv, req->handle);
 	if (!gem)
 		return -ENOENT;
 	nvbo = nouveau_gem_object(gem);
 
-	spin_lock(&nvbo->bo.bdev->fence_lock);
-	ret = ttm_bo_wait(&nvbo->bo, true, true, no_wait);
-	spin_unlock(&nvbo->bo.bdev->fence_lock);
+	if (no_wait)
+		ret = reservation_object_test_signaled_rcu(nvbo->bo.resv, write) ? 0 : -EBUSY;
+	else {
+		long lret;
+
+		lret = reservation_object_wait_timeout_rcu(nvbo->bo.resv, write, true, 30 * HZ);
+		if (!lret)
+			ret = -EBUSY;
+		else if (lret > 0)
+			ret = 0;
+		else
+			ret = lret;
+	}
+	nouveau_bo_sync_for_cpu(nvbo);
 	drm_gem_object_unreference_unlocked(gem);
+
 	return ret;
 }
 
@@ -902,6 +902,17 @@ int
 nouveau_gem_ioctl_cpu_fini(struct drm_device *dev, void *data,
 			   struct drm_file *file_priv)
 {
+	struct drm_nouveau_gem_cpu_fini *req = data;
+	struct drm_gem_object *gem;
+	struct nouveau_bo *nvbo;
+
+	gem = drm_gem_object_lookup(dev, file_priv, req->handle);
+	if (!gem)
+		return -ENOENT;
+	nvbo = nouveau_gem_object(gem);
+
+	nouveau_bo_sync_for_device(nvbo);
+	drm_gem_object_unreference_unlocked(gem);
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/nouveau/nouveau_gem.h b/drivers/gpu/drm/nouveau/nouveau_gem.h
index ddab762d81fe..e4049faca780 100644
--- a/drivers/gpu/drm/nouveau/nouveau_gem.h
+++ b/drivers/gpu/drm/nouveau/nouveau_gem.h
@@ -39,7 +39,7 @@ struct reservation_object *nouveau_gem_prime_res_obj(struct drm_gem_object *);
 extern void nouveau_gem_prime_unpin(struct drm_gem_object *);
 extern struct sg_table *nouveau_gem_prime_get_sg_table(struct drm_gem_object *);
 extern struct drm_gem_object *nouveau_gem_prime_import_sg_table(
-	struct drm_device *, size_t size, struct sg_table *);
+	struct drm_device *, struct dma_buf_attachment *, struct sg_table *);
 extern void *nouveau_gem_prime_vmap(struct drm_gem_object *);
 extern void nouveau_gem_prime_vunmap(struct drm_gem_object *, void *);
 
diff --git a/drivers/gpu/drm/nouveau/nouveau_nvif.c b/drivers/gpu/drm/nouveau/nouveau_nvif.c
index 47ca88623753..6544b84f0303 100644
--- a/drivers/gpu/drm/nouveau/nouveau_nvif.c
+++ b/drivers/gpu/drm/nouveau/nouveau_nvif.c
@@ -40,12 +40,12 @@
 #include "nouveau_usif.h"
 
 static void
-nvkm_client_unmap(void *priv, void *ptr, u32 size)
+nvkm_client_unmap(void *priv, void __iomem *ptr, u32 size)
 {
 	iounmap(ptr);
 }
 
-static void *
+static void __iomem *
 nvkm_client_map(void *priv, u64 handle, u32 size)
 {
 	return ioremap(handle, size);
diff --git a/drivers/gpu/drm/nouveau/nouveau_platform.c b/drivers/gpu/drm/nouveau/nouveau_platform.c
index 246a824c16ca..b307bbedd4c4 100644
--- a/drivers/gpu/drm/nouveau/nouveau_platform.c
+++ b/drivers/gpu/drm/nouveau/nouveau_platform.c
@@ -27,6 +27,7 @@
 #include <linux/of.h>
 #include <linux/reset.h>
 #include <linux/regulator/consumer.h>
+#include <soc/tegra/fuse.h>
 #include <soc/tegra/pmc.h>
 
 #include "nouveau_drm.h"
@@ -128,6 +129,7 @@ static int nouveau_platform_probe(struct platform_device *pdev)
 	}
 
 	device->gpu = gpu;
+	device->gpu_speedo = tegra_sku_info.gpu_speedo_value;
 
 	err = drm_dev_register(drm, 0);
 	if (err < 0)
diff --git a/drivers/gpu/drm/nouveau/nouveau_platform.h b/drivers/gpu/drm/nouveau/nouveau_platform.h
index 91f66504900e..58c28b5653d5 100644
--- a/drivers/gpu/drm/nouveau/nouveau_platform.h
+++ b/drivers/gpu/drm/nouveau/nouveau_platform.h
@@ -41,6 +41,8 @@ struct nouveau_platform_device {
 	struct nouveau_device device;
 
 	struct nouveau_platform_gpu *gpu;
+
+	int gpu_speedo;
 };
 
 #define nv_device_to_platform(d)                                               \
diff --git a/drivers/gpu/drm/nouveau/nouveau_prime.c b/drivers/gpu/drm/nouveau/nouveau_prime.c
index 1f51008e4d26..dd32ad6db53d 100644
--- a/drivers/gpu/drm/nouveau/nouveau_prime.c
+++ b/drivers/gpu/drm/nouveau/nouveau_prime.c
@@ -23,6 +23,7 @@
  */
 
 #include <drm/drmP.h>
+#include <linux/dma-buf.h>
 
 #include "nouveau_drm.h"
 #include "nouveau_gem.h"
@@ -56,17 +57,20 @@ void nouveau_gem_prime_vunmap(struct drm_gem_object *obj, void *vaddr)
 }
 
 struct drm_gem_object *nouveau_gem_prime_import_sg_table(struct drm_device *dev,
-							 size_t size,
+							 struct dma_buf_attachment *attach,
 							 struct sg_table *sg)
 {
 	struct nouveau_bo *nvbo;
+	struct reservation_object *robj = attach->dmabuf->resv;
 	u32 flags = 0;
 	int ret;
 
 	flags = TTM_PL_FLAG_TT;
 
-	ret = nouveau_bo_new(dev, size, 0, flags, 0, 0,
-			     sg, &nvbo);
+	ww_mutex_lock(&robj->lock, NULL);
+	ret = nouveau_bo_new(dev, attach->dmabuf->size, 0, flags, 0, 0,
+			     sg, robj, &nvbo);
+	ww_mutex_unlock(&robj->lock);
 	if (ret)
 		return ERR_PTR(ret);
 
@@ -89,7 +93,7 @@ int nouveau_gem_prime_pin(struct drm_gem_object *obj)
 	int ret;
 
 	/* pin buffer into GTT */
-	ret = nouveau_bo_pin(nvbo, TTM_PL_FLAG_TT);
+	ret = nouveau_bo_pin(nvbo, TTM_PL_FLAG_TT, false);
 	if (ret)
 		return -EINVAL;
 
diff --git a/drivers/gpu/drm/nouveau/nouveau_sysfs.c b/drivers/gpu/drm/nouveau/nouveau_sysfs.c
index 3c6962d15b26..8fbbf3093d86 100644
--- a/drivers/gpu/drm/nouveau/nouveau_sysfs.c
+++ b/drivers/gpu/drm/nouveau/nouveau_sysfs.c
@@ -29,7 +29,7 @@
 #include "nouveau_sysfs.h"
 
 MODULE_PARM_DESC(pstate, "enable sysfs pstate file, which will be moved in the future");
-static int nouveau_pstate;
+int nouveau_pstate;
 module_param_named(pstate, nouveau_pstate, int, 0400);
 
 static inline struct drm_device *
diff --git a/drivers/gpu/drm/nouveau/nouveau_sysfs.h b/drivers/gpu/drm/nouveau/nouveau_sysfs.h
index f973378160f8..4e5ea9241b28 100644
--- a/drivers/gpu/drm/nouveau/nouveau_sysfs.h
+++ b/drivers/gpu/drm/nouveau/nouveau_sysfs.h
@@ -16,4 +16,6 @@ nouveau_sysfs(struct drm_device *dev)
 int  nouveau_sysfs_init(struct drm_device *);
 void nouveau_sysfs_fini(struct drm_device *);
 
+extern int nouveau_pstate;
+
 #endif
diff --git a/drivers/gpu/drm/nouveau/nouveau_ttm.c b/drivers/gpu/drm/nouveau/nouveau_ttm.c
index 53874b76b031..3d1cfcb96b6b 100644
--- a/drivers/gpu/drm/nouveau/nouveau_ttm.c
+++ b/drivers/gpu/drm/nouveau/nouveau_ttm.c
@@ -28,6 +28,7 @@
 #include "nouveau_ttm.h"
 #include "nouveau_gem.h"
 
+#include "drm_legacy.h"
 static int
 nouveau_vram_manager_init(struct ttm_mem_type_manager *man, unsigned long psize)
 {
@@ -71,8 +72,7 @@ nouveau_vram_manager_del(struct ttm_mem_type_manager *man,
 static int
 nouveau_vram_manager_new(struct ttm_mem_type_manager *man,
 			 struct ttm_buffer_object *bo,
-			 struct ttm_placement *placement,
-			 uint32_t flags,
+			 const struct ttm_place *place,
 			 struct ttm_mem_reg *mem)
 {
 	struct nouveau_drm *drm = nouveau_bdev(man->bdev);
@@ -158,8 +158,7 @@ nouveau_gart_manager_del(struct ttm_mem_type_manager *man,
 static int
 nouveau_gart_manager_new(struct ttm_mem_type_manager *man,
 			 struct ttm_buffer_object *bo,
-			 struct ttm_placement *placement,
-			 uint32_t flags,
+			 const struct ttm_place *place,
 			 struct ttm_mem_reg *mem)
 {
 	struct nouveau_drm *drm = nouveau_bdev(bo->bdev);
@@ -239,8 +238,7 @@ nv04_gart_manager_del(struct ttm_mem_type_manager *man, struct ttm_mem_reg *mem)
 static int
 nv04_gart_manager_new(struct ttm_mem_type_manager *man,
 		      struct ttm_buffer_object *bo,
-		      struct ttm_placement *placement,
-		      uint32_t flags,
+		      const struct ttm_place *place,
 		      struct ttm_mem_reg *mem)
 {
 	struct nouveau_mem *node;
@@ -284,7 +282,7 @@ nouveau_ttm_mmap(struct file *filp, struct vm_area_struct *vma)
 	struct nouveau_drm *drm = nouveau_drm(file_priv->minor->dev);
 
 	if (unlikely(vma->vm_pgoff < DRM_FILE_PAGE_OFFSET))
-		return drm_mmap(filp, vma);
+		return drm_legacy_mmap(filp, vma);
 
 	return ttm_bo_mmap(filp, vma, &drm->ttm.bdev);
 }
diff --git a/drivers/gpu/drm/nouveau/nv04_fence.c b/drivers/gpu/drm/nouveau/nv04_fence.c
index 239c2c5a9615..f9859deb108a 100644
--- a/drivers/gpu/drm/nouveau/nv04_fence.c
+++ b/drivers/gpu/drm/nouveau/nv04_fence.c
@@ -41,7 +41,7 @@ nv04_fence_emit(struct nouveau_fence *fence)
 	int ret = RING_SPACE(chan, 2);
 	if (ret == 0) {
 		BEGIN_NV04(chan, NvSubSw, 0x0150, 1);
-		OUT_RING  (chan, fence->sequence);
+		OUT_RING  (chan, fence->base.seqno);
 		FIRE_RING (chan);
 	}
 	return ret;
@@ -67,7 +67,7 @@ nv04_fence_context_del(struct nouveau_channel *chan)
 	struct nv04_fence_chan *fctx = chan->fence;
 	nouveau_fence_context_del(&fctx->base);
 	chan->fence = NULL;
-	kfree(fctx);
+	nouveau_fence_context_free(&fctx->base);
 }
 
 static int
@@ -75,7 +75,7 @@ nv04_fence_context_new(struct nouveau_channel *chan)
 {
 	struct nv04_fence_chan *fctx = kzalloc(sizeof(*fctx), GFP_KERNEL);
 	if (fctx) {
-		nouveau_fence_context_new(&fctx->base);
+		nouveau_fence_context_new(chan, &fctx->base);
 		fctx->base.emit = nv04_fence_emit;
 		fctx->base.sync = nv04_fence_sync;
 		fctx->base.read = nv04_fence_read;
@@ -105,5 +105,7 @@ nv04_fence_create(struct nouveau_drm *drm)
 	priv->base.dtor = nv04_fence_destroy;
 	priv->base.context_new = nv04_fence_context_new;
 	priv->base.context_del = nv04_fence_context_del;
+	priv->base.contexts = 15;
+	priv->base.context_base = fence_context_alloc(priv->base.contexts);
 	return 0;
 }
diff --git a/drivers/gpu/drm/nouveau/nv10_fence.c b/drivers/gpu/drm/nouveau/nv10_fence.c
index 4faaf0acf5d7..5e1ea1cdce75 100644
--- a/drivers/gpu/drm/nouveau/nv10_fence.c
+++ b/drivers/gpu/drm/nouveau/nv10_fence.c
@@ -33,7 +33,7 @@ nv10_fence_emit(struct nouveau_fence *fence)
 	int ret = RING_SPACE(chan, 2);
 	if (ret == 0) {
 		BEGIN_NV04(chan, 0, NV10_SUBCHAN_REF_CNT, 1);
-		OUT_RING  (chan, fence->sequence);
+		OUT_RING  (chan, fence->base.seqno);
 		FIRE_RING (chan);
 	}
 	return ret;
@@ -63,7 +63,7 @@ nv10_fence_context_del(struct nouveau_channel *chan)
 		nvif_object_fini(&fctx->head[i]);
 	nvif_object_fini(&fctx->sema);
 	chan->fence = NULL;
-	kfree(fctx);
+	nouveau_fence_context_free(&fctx->base);
 }
 
 int
@@ -75,7 +75,7 @@ nv10_fence_context_new(struct nouveau_channel *chan)
 	if (!fctx)
 		return -ENOMEM;
 
-	nouveau_fence_context_new(&fctx->base);
+	nouveau_fence_context_new(chan, &fctx->base);
 	fctx->base.emit = nv10_fence_emit;
 	fctx->base.read = nv10_fence_read;
 	fctx->base.sync = nv10_fence_sync;
@@ -106,6 +106,8 @@ nv10_fence_create(struct nouveau_drm *drm)
 	priv->base.dtor = nv10_fence_destroy;
 	priv->base.context_new = nv10_fence_context_new;
 	priv->base.context_del = nv10_fence_context_del;
+	priv->base.contexts = 31;
+	priv->base.context_base = fence_context_alloc(priv->base.contexts);
 	spin_lock_init(&priv->lock);
 	return 0;
 }
diff --git a/drivers/gpu/drm/nouveau/nv17_fence.c b/drivers/gpu/drm/nouveau/nv17_fence.c
index ca907479f92f..57860cfa1de5 100644
--- a/drivers/gpu/drm/nouveau/nv17_fence.c
+++ b/drivers/gpu/drm/nouveau/nv17_fence.c
@@ -84,7 +84,7 @@ nv17_fence_context_new(struct nouveau_channel *chan)
 	if (!fctx)
 		return -ENOMEM;
 
-	nouveau_fence_context_new(&fctx->base);
+	nouveau_fence_context_new(chan, &fctx->base);
 	fctx->base.emit = nv10_fence_emit;
 	fctx->base.read = nv10_fence_read;
 	fctx->base.sync = nv17_fence_sync;
@@ -124,12 +124,14 @@ nv17_fence_create(struct nouveau_drm *drm)
 	priv->base.resume = nv17_fence_resume;
 	priv->base.context_new = nv17_fence_context_new;
 	priv->base.context_del = nv10_fence_context_del;
+	priv->base.contexts = 31;
+	priv->base.context_base = fence_context_alloc(priv->base.contexts);
 	spin_lock_init(&priv->lock);
 
 	ret = nouveau_bo_new(drm->dev, 4096, 0x1000, TTM_PL_FLAG_VRAM,
-			     0, 0x0000, NULL, &priv->bo);
+			     0, 0x0000, NULL, NULL, &priv->bo);
 	if (!ret) {
-		ret = nouveau_bo_pin(priv->bo, TTM_PL_FLAG_VRAM);
+		ret = nouveau_bo_pin(priv->bo, TTM_PL_FLAG_VRAM, false);
 		if (!ret) {
 			ret = nouveau_bo_map(priv->bo);
 			if (ret)
diff --git a/drivers/gpu/drm/nouveau/nv50_display.c b/drivers/gpu/drm/nouveau/nv50_display.c
index 03949eaa629f..490b90866baf 100644
--- a/drivers/gpu/drm/nouveau/nv50_display.c
+++ b/drivers/gpu/drm/nouveau/nv50_display.c
@@ -26,6 +26,7 @@
 
 #include <drm/drmP.h>
 #include <drm/drm_crtc_helper.h>
+#include <drm/drm_plane_helper.h>
 #include <drm/drm_dp_helper.h>
 
 #include <nvif/class.h>
@@ -65,15 +66,29 @@ static int
 nv50_chan_create(struct nvif_object *disp, const u32 *oclass, u8 head,
 		 void *data, u32 size, struct nv50_chan *chan)
 {
+	const u32 handle = (oclass[0] << 16) | head;
+	u32 sclass[8];
+	int ret, i;
+
+	ret = nvif_object_sclass(disp, sclass, ARRAY_SIZE(sclass));
+	WARN_ON(ret > ARRAY_SIZE(sclass));
+	if (ret < 0)
+		return ret;
+
 	while (oclass[0]) {
-		int ret = nvif_object_init(disp, NULL, (oclass[0] << 16) | head,
-					   oclass[0], data, size,
-					  &chan->user);
-		if (oclass++, ret == 0) {
-			nvif_object_map(&chan->user);
-			return ret;
+		for (i = 0; i < ARRAY_SIZE(sclass); i++) {
+			if (sclass[i] == oclass[0]) {
+				ret = nvif_object_init(disp, NULL, handle,
+						       oclass[0], data, size,
+						       &chan->user);
+				if (ret == 0)
+					nvif_object_map(&chan->user);
+				return ret;
+			}
 		}
+		oclass++;
 	}
+
 	return -ENOSYS;
 }
 
@@ -110,6 +125,7 @@ nv50_pioc_create(struct nvif_object *disp, const u32 *oclass, u8 head,
 
 struct nv50_curs {
 	struct nv50_pioc base;
+	struct nouveau_bo *image;
 };
 
 static int
@@ -265,6 +281,7 @@ nv50_core_create(struct nvif_object *disp, u64 syncbuf, struct nv50_mast *core)
 		.pushbuf = 0xb0007d00,
 	};
 	static const u32 oclass[] = {
+		GM204_DISP_CORE_CHANNEL_DMA,
 		GM107_DISP_CORE_CHANNEL_DMA,
 		GK110_DISP_CORE_CHANNEL_DMA,
 		GK104_DISP_CORE_CHANNEL_DMA,
@@ -424,8 +441,21 @@ evo_kick(u32 *push, void *evoc)
 	mutex_unlock(&dmac->lock);
 }
 
+#if 1
 #define evo_mthd(p,m,s) *((p)++) = (((s) << 18) | (m))
 #define evo_data(p,d)   *((p)++) = (d)
+#else
+#define evo_mthd(p,m,s) do {                                                   \
+	const u32 _m = (m), _s = (s);                                          \
+	printk(KERN_ERR "%04x %d %s\n", _m, _s, __func__);                     \
+	*((p)++) = ((_s << 18) | _m);                                          \
+} while(0)
+#define evo_data(p,d) do {                                                     \
+	const u32 _d = (d);                                                    \
+	printk(KERN_ERR "\t%08x\n", _d);                                       \
+	*((p)++) = _d;                                                         \
+} while(0)
+#endif
 
 static bool
 evo_sync_wait(void *data)
@@ -791,6 +821,22 @@ nv50_crtc_set_scale(struct nouveau_crtc *nv_crtc, bool update)
 }
 
 static int
+nv50_crtc_set_raster_vblank_dmi(struct nouveau_crtc *nv_crtc, u32 usec)
+{
+	struct nv50_mast *mast = nv50_mast(nv_crtc->base.dev);
+	u32 *push;
+
+	push = evo_wait(mast, 8);
+	if (!push)
+		return -ENOMEM;
+
+	evo_mthd(push, 0x0828 + (nv_crtc->index * 0x400), 1);
+	evo_data(push, usec);
+	evo_kick(push, mast);
+	return 0;
+}
+
+static int
 nv50_crtc_set_color_vibrance(struct nouveau_crtc *nv_crtc, bool update)
 {
 	struct nv50_mast *mast = nv50_mast(nv_crtc->base.dev);
@@ -871,23 +917,24 @@ static void
 nv50_crtc_cursor_show(struct nouveau_crtc *nv_crtc)
 {
 	struct nv50_mast *mast = nv50_mast(nv_crtc->base.dev);
+	struct nv50_curs *curs = nv50_curs(&nv_crtc->base);
 	u32 *push = evo_wait(mast, 16);
 	if (push) {
 		if (nv50_vers(mast) < G82_DISP_CORE_CHANNEL_DMA) {
 			evo_mthd(push, 0x0880 + (nv_crtc->index * 0x400), 2);
 			evo_data(push, 0x85000000);
-			evo_data(push, nv_crtc->cursor.nvbo->bo.offset >> 8);
+			evo_data(push, curs->image->bo.offset >> 8);
 		} else
 		if (nv50_vers(mast) < GF110_DISP_CORE_CHANNEL_DMA) {
 			evo_mthd(push, 0x0880 + (nv_crtc->index * 0x400), 2);
 			evo_data(push, 0x85000000);
-			evo_data(push, nv_crtc->cursor.nvbo->bo.offset >> 8);
+			evo_data(push, curs->image->bo.offset >> 8);
 			evo_mthd(push, 0x089c + (nv_crtc->index * 0x400), 1);
 			evo_data(push, mast->base.vram.handle);
 		} else {
 			evo_mthd(push, 0x0480 + (nv_crtc->index * 0x300), 2);
 			evo_data(push, 0x85000000);
-			evo_data(push, nv_crtc->cursor.nvbo->bo.offset >> 8);
+			evo_data(push, curs->image->bo.offset >> 8);
 			evo_mthd(push, 0x048c + (nv_crtc->index * 0x300), 1);
 			evo_data(push, mast->base.vram.handle);
 		}
@@ -924,8 +971,9 @@ static void
 nv50_crtc_cursor_show_hide(struct nouveau_crtc *nv_crtc, bool show, bool update)
 {
 	struct nv50_mast *mast = nv50_mast(nv_crtc->base.dev);
+	struct nv50_curs *curs = nv50_curs(&nv_crtc->base);
 
-	if (show)
+	if (show && curs->image)
 		nv50_crtc_cursor_show(nv_crtc);
 	else
 		nv50_crtc_cursor_hide(nv_crtc);
@@ -1025,7 +1073,7 @@ nv50_crtc_commit(struct drm_crtc *crtc)
 		evo_kick(push, mast);
 	}
 
-	nv50_crtc_cursor_show_hide(nv_crtc, nv_crtc->cursor.visible, true);
+	nv50_crtc_cursor_show_hide(nv_crtc, true, true);
 	nv50_display_flip_next(crtc, crtc->primary->fb, NULL, 1);
 }
 
@@ -1044,7 +1092,7 @@ nv50_crtc_swap_fbs(struct drm_crtc *crtc, struct drm_framebuffer *old_fb)
 	struct nv50_head *head = nv50_head(crtc);
 	int ret;
 
-	ret = nouveau_bo_pin(nvfb->nvbo, TTM_PL_FLAG_VRAM);
+	ret = nouveau_bo_pin(nvfb->nvbo, TTM_PL_FLAG_VRAM, true);
 	if (ret == 0) {
 		if (head->image)
 			nouveau_bo_unpin(head->image);
@@ -1066,7 +1114,7 @@ nv50_crtc_mode_set(struct drm_crtc *crtc, struct drm_display_mode *umode,
 	u32 vscan = (mode->flags & DRM_MODE_FLAG_DBLSCAN) ? 2 : 1;
 	u32 hactive, hsynce, hbackp, hfrontp, hblanke, hblanks;
 	u32 vactive, vsynce, vbackp, vfrontp, vblanke, vblanks;
-	u32 vblan2e = 0, vblan2s = 1;
+	u32 vblan2e = 0, vblan2s = 1, vblankus = 0;
 	u32 *push;
 	int ret;
 
@@ -1083,6 +1131,11 @@ nv50_crtc_mode_set(struct drm_crtc *crtc, struct drm_display_mode *umode,
 	vblanke = vsynce + vbackp;
 	vfrontp = (mode->vsync_start - mode->vdisplay) * vscan / ilace;
 	vblanks = vactive - vfrontp - 1;
+	/* XXX: Safe underestimate, even "0" works */
+	vblankus = (vactive - mode->vdisplay - 2) * hactive;
+	vblankus *= 1000;
+	vblankus /= mode->clock;
+
 	if (mode->flags & DRM_MODE_FLAG_INTERLACE) {
 		vblan2e = vactive + vsynce + vbackp;
 		vblan2s = vblan2e + (mode->vdisplay * vscan / ilace);
@@ -1136,6 +1189,11 @@ nv50_crtc_mode_set(struct drm_crtc *crtc, struct drm_display_mode *umode,
 	nv_connector = nouveau_crtc_connector_get(nv_crtc);
 	nv50_crtc_set_dither(nv_crtc, false);
 	nv50_crtc_set_scale(nv_crtc, false);
+
+	/* G94 only accepts this after setting scale */
+	if (nv50_vers(mast) < GF110_DISP_CORE_CHANNEL_DMA)
+		nv50_crtc_set_raster_vblank_dmi(nv_crtc, vblankus);
+
 	nv50_crtc_set_color_vibrance(nv_crtc, false);
 	nv50_crtc_set_image(nv_crtc, crtc->primary->fb, x, y, false);
 	return 0;
@@ -1215,13 +1273,13 @@ nv50_crtc_cursor_set(struct drm_crtc *crtc, struct drm_file *file_priv,
 		     uint32_t handle, uint32_t width, uint32_t height)
 {
 	struct nouveau_crtc *nv_crtc = nouveau_crtc(crtc);
+	struct nv50_curs *curs = nv50_curs(crtc);
 	struct drm_device *dev = crtc->dev;
-	struct drm_gem_object *gem;
-	struct nouveau_bo *nvbo;
-	bool visible = (handle != 0);
-	int i, ret = 0;
+	struct drm_gem_object *gem = NULL;
+	struct nouveau_bo *nvbo = NULL;
+	int ret = 0;
 
-	if (visible) {
+	if (handle) {
 		if (width != 64 || height != 64)
 			return -EINVAL;
 
@@ -1230,23 +1288,17 @@ nv50_crtc_cursor_set(struct drm_crtc *crtc, struct drm_file *file_priv,
 			return -ENOENT;
 		nvbo = nouveau_gem_object(gem);
 
-		ret = nouveau_bo_map(nvbo);
-		if (ret == 0) {
-			for (i = 0; i < 64 * 64; i++) {
-				u32 v = nouveau_bo_rd32(nvbo, i);
-				nouveau_bo_wr32(nv_crtc->cursor.nvbo, i, v);
-			}
-			nouveau_bo_unmap(nvbo);
-		}
-
-		drm_gem_object_unreference_unlocked(gem);
+		ret = nouveau_bo_pin(nvbo, TTM_PL_FLAG_VRAM, true);
 	}
 
-	if (visible != nv_crtc->cursor.visible) {
-		nv50_crtc_cursor_show_hide(nv_crtc, visible, true);
-		nv_crtc->cursor.visible = visible;
+	if (ret == 0) {
+		if (curs->image)
+			nouveau_bo_unpin(curs->image);
+		nouveau_bo_ref(nvbo, &curs->image);
 	}
+	drm_gem_object_unreference_unlocked(gem);
 
+	nv50_crtc_cursor_show_hide(nv_crtc, true, true);
 	return ret;
 }
 
@@ -1301,10 +1353,10 @@ nv50_crtc_destroy(struct drm_crtc *crtc)
 		nouveau_bo_unpin(head->image);
 	nouveau_bo_ref(NULL, &head->image);
 
-	nouveau_bo_unmap(nv_crtc->cursor.nvbo);
-	if (nv_crtc->cursor.nvbo)
-		nouveau_bo_unpin(nv_crtc->cursor.nvbo);
-	nouveau_bo_ref(NULL, &nv_crtc->cursor.nvbo);
+	/*XXX: ditto */
+	if (head->curs.image)
+		nouveau_bo_unpin(head->curs.image);
+	nouveau_bo_ref(NULL, &head->curs.image);
 
 	nouveau_bo_unmap(nv_crtc->lut.nvbo);
 	if (nv_crtc->lut.nvbo)
@@ -1336,16 +1388,6 @@ static const struct drm_crtc_funcs nv50_crtc_func = {
 	.page_flip = nouveau_crtc_page_flip,
 };
 
-static void
-nv50_cursor_set_pos(struct nouveau_crtc *nv_crtc, int x, int y)
-{
-}
-
-static void
-nv50_cursor_set_offset(struct nouveau_crtc *nv_crtc, uint32_t offset)
-{
-}
-
 static int
 nv50_crtc_create(struct drm_device *dev, int index)
 {
@@ -1364,8 +1406,6 @@ nv50_crtc_create(struct drm_device *dev, int index)
 	head->base.set_color_vibrance = nv50_crtc_set_color_vibrance;
 	head->base.color_vibrance = 50;
 	head->base.vibrant_hue = 0;
-	head->base.cursor.set_offset = nv50_cursor_set_offset;
-	head->base.cursor.set_pos = nv50_cursor_set_pos;
 	for (i = 0; i < 256; i++) {
 		head->base.lut.r[i] = i << 8;
 		head->base.lut.g[i] = i << 8;
@@ -1378,9 +1418,9 @@ nv50_crtc_create(struct drm_device *dev, int index)
 	drm_mode_crtc_set_gamma_size(crtc, 256);
 
 	ret = nouveau_bo_new(dev, 8192, 0x100, TTM_PL_FLAG_VRAM,
-			     0, 0x0000, NULL, &head->base.lut.nvbo);
+			     0, 0x0000, NULL, NULL, &head->base.lut.nvbo);
 	if (!ret) {
-		ret = nouveau_bo_pin(head->base.lut.nvbo, TTM_PL_FLAG_VRAM);
+		ret = nouveau_bo_pin(head->base.lut.nvbo, TTM_PL_FLAG_VRAM, true);
 		if (!ret) {
 			ret = nouveau_bo_map(head->base.lut.nvbo);
 			if (ret)
@@ -1400,22 +1440,6 @@ nv50_crtc_create(struct drm_device *dev, int index)
 	if (ret)
 		goto out;
 
-	ret = nouveau_bo_new(dev, 64 * 64 * 4, 0x100, TTM_PL_FLAG_VRAM,
-			     0, 0x0000, NULL, &head->base.cursor.nvbo);
-	if (!ret) {
-		ret = nouveau_bo_pin(head->base.cursor.nvbo, TTM_PL_FLAG_VRAM);
-		if (!ret) {
-			ret = nouveau_bo_map(head->base.cursor.nvbo);
-			if (ret)
-				nouveau_bo_unpin(head->base.lut.nvbo);
-		}
-		if (ret)
-			nouveau_bo_ref(NULL, &head->base.cursor.nvbo);
-	}
-
-	if (ret)
-		goto out;
-
 	/* allocate page flip / sync resources */
 	ret = nv50_base_create(disp->disp, index, disp->sync->bo.offset,
 			      &head->sync);
@@ -1651,17 +1675,21 @@ static void
 nv50_audio_mode_set(struct drm_encoder *encoder, struct drm_display_mode *mode)
 {
 	struct nouveau_encoder *nv_encoder = nouveau_encoder(encoder);
+	struct nouveau_crtc *nv_crtc = nouveau_crtc(encoder->crtc);
 	struct nouveau_connector *nv_connector;
 	struct nv50_disp *disp = nv50_disp(encoder->dev);
-	struct {
-		struct nv50_disp_mthd_v1 base;
-		struct nv50_disp_sor_hda_eld_v0 eld;
+	struct __packed {
+		struct {
+			struct nv50_disp_mthd_v1 mthd;
+			struct nv50_disp_sor_hda_eld_v0 eld;
+		} base;
 		u8 data[sizeof(nv_connector->base.eld)];
 	} args = {
-		.base.version = 1,
-		.base.method  = NV50_DISP_MTHD_V1_SOR_HDA_ELD,
-		.base.hasht   = nv_encoder->dcb->hasht,
-		.base.hashm   = nv_encoder->dcb->hashm,
+		.base.mthd.version = 1,
+		.base.mthd.method  = NV50_DISP_MTHD_V1_SOR_HDA_ELD,
+		.base.mthd.hasht   = nv_encoder->dcb->hasht,
+		.base.mthd.hashm   = (0xf0ff & nv_encoder->dcb->hashm) |
+				     (0x0100 << nv_crtc->index),
 	};
 
 	nv_connector = nouveau_encoder_connector_get(nv_encoder);
@@ -1671,11 +1699,12 @@ nv50_audio_mode_set(struct drm_encoder *encoder, struct drm_display_mode *mode)
 	drm_edid_to_eld(&nv_connector->base, nv_connector->edid);
 	memcpy(args.data, nv_connector->base.eld, sizeof(args.data));
 
-	nvif_mthd(disp->disp, 0, &args, sizeof(args));
+	nvif_mthd(disp->disp, 0, &args,
+		  sizeof(args.base) + drm_eld_size(args.data));
 }
 
 static void
-nv50_audio_disconnect(struct drm_encoder *encoder)
+nv50_audio_disconnect(struct drm_encoder *encoder, struct nouveau_crtc *nv_crtc)
 {
 	struct nouveau_encoder *nv_encoder = nouveau_encoder(encoder);
 	struct nv50_disp *disp = nv50_disp(encoder->dev);
@@ -1686,7 +1715,8 @@ nv50_audio_disconnect(struct drm_encoder *encoder)
 		.base.version = 1,
 		.base.method  = NV50_DISP_MTHD_V1_SOR_HDA_ELD,
 		.base.hasht   = nv_encoder->dcb->hasht,
-		.base.hashm   = nv_encoder->dcb->hashm,
+		.base.hashm   = (0xf0ff & nv_encoder->dcb->hashm) |
+				(0x0100 << nv_crtc->index),
 	};
 
 	nvif_mthd(disp->disp, 0, &args, sizeof(args));
@@ -1745,8 +1775,6 @@ nv50_hdmi_disconnect(struct drm_encoder *encoder, struct nouveau_crtc *nv_crtc)
 			       (0x0100 << nv_crtc->index),
 	};
 
-	nv50_audio_disconnect(encoder);
-
 	nvif_mthd(disp->disp, 0, &args, sizeof(args));
 }
 
@@ -1855,6 +1883,7 @@ nv50_sor_disconnect(struct drm_encoder *encoder)
 	if (nv_crtc) {
 		nv50_crtc_prepare(&nv_crtc->base);
 		nv50_sor_ctrl(nv_encoder, 1 << nv_crtc->index, 0);
+		nv50_audio_disconnect(encoder, nv_crtc);
 		nv50_hdmi_disconnect(&nv_encoder->base.base, nv_crtc);
 	}
 }
@@ -1954,6 +1983,7 @@ nv50_sor_mode_set(struct drm_encoder *encoder, struct drm_display_mode *umode,
 			proto = 0x8;
 		else
 			proto = 0x9;
+		nv50_audio_mode_set(encoder, mode);
 		break;
 	default:
 		BUG_ON(1);
@@ -2342,11 +2372,6 @@ nv50_fb_ctor(struct drm_framebuffer *fb)
 	u8 kind = nouveau_bo_tile_layout(nvbo) >> 8;
 	u8 tile = nvbo->tile_mode;
 
-	if (nvbo->tile_flags & NOUVEAU_GEM_TILE_NONCONTIG) {
-		NV_ERROR(drm, "framebuffer requires contiguous bo\n");
-		return -EINVAL;
-	}
-
 	if (drm->device.info.chipset >= 0xc0)
 		tile >>= 4; /* yep.. */
 
@@ -2458,9 +2483,9 @@ nv50_display_create(struct drm_device *dev)
 
 	/* small shared memory area we use for notifiers and semaphores */
 	ret = nouveau_bo_new(dev, 4096, 0x1000, TTM_PL_FLAG_VRAM,
-			     0, 0x0000, NULL, &disp->sync);
+			     0, 0x0000, NULL, NULL, &disp->sync);
 	if (!ret) {
-		ret = nouveau_bo_pin(disp->sync, TTM_PL_FLAG_VRAM);
+		ret = nouveau_bo_pin(disp->sync, TTM_PL_FLAG_VRAM, true);
 		if (!ret) {
 			ret = nouveau_bo_map(disp->sync);
 			if (ret)
diff --git a/drivers/gpu/drm/nouveau/nv50_fence.c b/drivers/gpu/drm/nouveau/nv50_fence.c
index 195cf51a7c31..a82d9ea7c6fd 100644
--- a/drivers/gpu/drm/nouveau/nv50_fence.c
+++ b/drivers/gpu/drm/nouveau/nv50_fence.c
@@ -46,7 +46,7 @@ nv50_fence_context_new(struct nouveau_channel *chan)
 	if (!fctx)
 		return -ENOMEM;
 
-	nouveau_fence_context_new(&fctx->base);
+	nouveau_fence_context_new(chan, &fctx->base);
 	fctx->base.emit = nv10_fence_emit;
 	fctx->base.read = nv10_fence_read;
 	fctx->base.sync = nv17_fence_sync;
@@ -95,12 +95,14 @@ nv50_fence_create(struct nouveau_drm *drm)
 	priv->base.resume = nv17_fence_resume;
 	priv->base.context_new = nv50_fence_context_new;
 	priv->base.context_del = nv10_fence_context_del;
+	priv->base.contexts = 127;
+	priv->base.context_base = fence_context_alloc(priv->base.contexts);
 	spin_lock_init(&priv->lock);
 
 	ret = nouveau_bo_new(drm->dev, 4096, 0x1000, TTM_PL_FLAG_VRAM,
-			     0, 0x0000, NULL, &priv->bo);
+			     0, 0x0000, NULL, NULL, &priv->bo);
 	if (!ret) {
-		ret = nouveau_bo_pin(priv->bo, TTM_PL_FLAG_VRAM);
+		ret = nouveau_bo_pin(priv->bo, TTM_PL_FLAG_VRAM, false);
 		if (!ret) {
 			ret = nouveau_bo_map(priv->bo);
 			if (ret)
diff --git a/drivers/gpu/drm/nouveau/nv84_fence.c b/drivers/gpu/drm/nouveau/nv84_fence.c
index 933a779c93ab..cb5b88938d45 100644
--- a/drivers/gpu/drm/nouveau/nv84_fence.c
+++ b/drivers/gpu/drm/nouveau/nv84_fence.c
@@ -82,7 +82,7 @@ nv84_fence_emit(struct nouveau_fence *fence)
 	else
 		addr += fctx->vma.offset;
 
-	return fctx->base.emit32(chan, addr, fence->sequence);
+	return fctx->base.emit32(chan, addr, fence->base.seqno);
 }
 
 static int
@@ -97,7 +97,7 @@ nv84_fence_sync(struct nouveau_fence *fence,
 	else
 		addr += fctx->vma.offset;
 
-	return fctx->base.sync32(chan, addr, fence->sequence);
+	return fctx->base.sync32(chan, addr, fence->base.seqno);
 }
 
 static u32
@@ -120,11 +120,12 @@ nv84_fence_context_del(struct nouveau_channel *chan)
 		nouveau_bo_vma_del(bo, &fctx->dispc_vma[i]);
 	}
 
+	nouveau_bo_wr32(priv->bo, chan->chid * 16 / 4, fctx->base.sequence);
 	nouveau_bo_vma_del(priv->bo, &fctx->vma_gart);
 	nouveau_bo_vma_del(priv->bo, &fctx->vma);
 	nouveau_fence_context_del(&fctx->base);
 	chan->fence = NULL;
-	kfree(fctx);
+	nouveau_fence_context_free(&fctx->base);
 }
 
 int
@@ -139,12 +140,13 @@ nv84_fence_context_new(struct nouveau_channel *chan)
 	if (!fctx)
 		return -ENOMEM;
 
-	nouveau_fence_context_new(&fctx->base);
+	nouveau_fence_context_new(chan, &fctx->base);
 	fctx->base.emit = nv84_fence_emit;
 	fctx->base.sync = nv84_fence_sync;
 	fctx->base.read = nv84_fence_read;
 	fctx->base.emit32 = nv84_fence_emit32;
 	fctx->base.sync32 = nv84_fence_sync32;
+	fctx->base.sequence = nv84_fence_read(chan);
 
 	ret = nouveau_bo_vma_add(priv->bo, cli->vm, &fctx->vma);
 	if (ret == 0) {
@@ -158,8 +160,6 @@ nv84_fence_context_new(struct nouveau_channel *chan)
 		ret = nouveau_bo_vma_add(bo, cli->vm, &fctx->dispc_vma[i]);
 	}
 
-	nouveau_bo_wr32(priv->bo, chan->chid * 16/4, 0x00000000);
-
 	if (ret)
 		nv84_fence_context_del(chan);
 	return ret;
@@ -168,13 +168,12 @@ nv84_fence_context_new(struct nouveau_channel *chan)
 static bool
 nv84_fence_suspend(struct nouveau_drm *drm)
 {
-	struct nouveau_fifo *pfifo = nvkm_fifo(&drm->device);
 	struct nv84_fence_priv *priv = drm->fence;
 	int i;
 
-	priv->suspend = vmalloc((pfifo->max + 1) * sizeof(u32));
+	priv->suspend = vmalloc(priv->base.contexts * sizeof(u32));
 	if (priv->suspend) {
-		for (i = 0; i <= pfifo->max; i++)
+		for (i = 0; i < priv->base.contexts; i++)
 			priv->suspend[i] = nouveau_bo_rd32(priv->bo, i*4);
 	}
 
@@ -184,12 +183,11 @@ nv84_fence_suspend(struct nouveau_drm *drm)
 static void
 nv84_fence_resume(struct nouveau_drm *drm)
 {
-	struct nouveau_fifo *pfifo = nvkm_fifo(&drm->device);
 	struct nv84_fence_priv *priv = drm->fence;
 	int i;
 
 	if (priv->suspend) {
-		for (i = 0; i <= pfifo->max; i++)
+		for (i = 0; i < priv->base.contexts; i++)
 			nouveau_bo_wr32(priv->bo, i*4, priv->suspend[i]);
 		vfree(priv->suspend);
 		priv->suspend = NULL;
@@ -229,13 +227,14 @@ nv84_fence_create(struct nouveau_drm *drm)
 	priv->base.context_new = nv84_fence_context_new;
 	priv->base.context_del = nv84_fence_context_del;
 
-	init_waitqueue_head(&priv->base.waiting);
+	priv->base.contexts = pfifo->max + 1;
+	priv->base.context_base = fence_context_alloc(priv->base.contexts);
 	priv->base.uevent = true;
 
-	ret = nouveau_bo_new(drm->dev, 16 * (pfifo->max + 1), 0,
-			     TTM_PL_FLAG_VRAM, 0, 0, NULL, &priv->bo);
+	ret = nouveau_bo_new(drm->dev, 16 * priv->base.contexts, 0,
+			     TTM_PL_FLAG_VRAM, 0, 0, NULL, NULL, &priv->bo);
 	if (ret == 0) {
-		ret = nouveau_bo_pin(priv->bo, TTM_PL_FLAG_VRAM);
+		ret = nouveau_bo_pin(priv->bo, TTM_PL_FLAG_VRAM, false);
 		if (ret == 0) {
 			ret = nouveau_bo_map(priv->bo);
 			if (ret)
@@ -246,11 +245,11 @@ nv84_fence_create(struct nouveau_drm *drm)
 	}
 
 	if (ret == 0)
-		ret = nouveau_bo_new(drm->dev, 16 * (pfifo->max + 1), 0,
-				     TTM_PL_FLAG_TT, 0, 0, NULL,
-				     &priv->bo_gart);
+		ret = nouveau_bo_new(drm->dev, 16 * priv->base.contexts, 0,
+				     TTM_PL_FLAG_TT | TTM_PL_FLAG_UNCACHED, 0,
+				     0, NULL, NULL, &priv->bo_gart);
 	if (ret == 0) {
-		ret = nouveau_bo_pin(priv->bo_gart, TTM_PL_FLAG_TT);
+		ret = nouveau_bo_pin(priv->bo_gart, TTM_PL_FLAG_TT, false);
 		if (ret == 0) {
 			ret = nouveau_bo_map(priv->bo_gart);
 			if (ret)
diff --git a/drivers/gpu/drm/nouveau/nvif/class.h b/drivers/gpu/drm/nouveau/nvif/class.h
index 573491f84792..4e308eacb27a 100644
--- a/drivers/gpu/drm/nouveau/nvif/class.h
+++ b/drivers/gpu/drm/nouveau/nvif/class.h
@@ -35,6 +35,7 @@
 #define GK104_DISP                                                   0x00009170
 #define GK110_DISP                                                   0x00009270
 #define GM107_DISP                                                   0x00009470
+#define GM204_DISP                                                   0x00009570
 
 #define NV50_DISP_CURSOR                                             0x0000507a
 #define G82_DISP_CURSOR                                              0x0000827a
@@ -65,6 +66,7 @@
 #define GK104_DISP_CORE_CHANNEL_DMA                                  0x0000917d
 #define GK110_DISP_CORE_CHANNEL_DMA                                  0x0000927d
 #define GM107_DISP_CORE_CHANNEL_DMA                                  0x0000947d
+#define GM204_DISP_CORE_CHANNEL_DMA                                  0x0000957d
 
 #define NV50_DISP_OVERLAY_CHANNEL_DMA                                0x0000507e
 #define G82_DISP_OVERLAY_CHANNEL_DMA                                 0x0000827e
@@ -131,6 +133,7 @@ struct nv_device_v0 {
 #define NV_DEVICE_V0_DISABLE_COPY1                        0x0000010000000000ULL
 #define NV_DEVICE_V0_DISABLE_VIC                          0x0000020000000000ULL
 #define NV_DEVICE_V0_DISABLE_VENC                         0x0000040000000000ULL
+#define NV_DEVICE_V0_DISABLE_COPY2                        0x0000080000000000ULL
 	__u64 disable;	/* disable particular subsystems */
 	__u64 debug0;	/* as above, but *internal* ids, and *NOT* ABI */
 };
@@ -479,6 +482,8 @@ struct nv50_disp_core_channel_dma_v0 {
 	__u32 pushbuf;
 };
 
+#define NV50_DISP_CORE_CHANNEL_DMA_V0_NTFY_UEVENT                          0x00
+
 /* cursor immediate */
 struct nv50_disp_cursor_v0 {
 	__u8  version;
@@ -486,6 +491,8 @@ struct nv50_disp_cursor_v0 {
 	__u8  pad02[6];
 };
 
+#define NV50_DISP_CURSOR_V0_NTFY_UEVENT                                    0x00
+
 /* base */
 struct nv50_disp_base_channel_dma_v0 {
 	__u8  version;
@@ -494,6 +501,8 @@ struct nv50_disp_base_channel_dma_v0 {
 	__u32 pushbuf;
 };
 
+#define NV50_DISP_BASE_CHANNEL_DMA_V0_NTFY_UEVENT                          0x00
+
 /* overlay */
 struct nv50_disp_overlay_channel_dma_v0 {
 	__u8  version;
@@ -502,6 +511,8 @@ struct nv50_disp_overlay_channel_dma_v0 {
 	__u32 pushbuf;
 };
 
+#define NV50_DISP_OVERLAY_CHANNEL_DMA_V0_NTFY_UEVENT                       0x00
+
 /* overlay immediate */
 struct nv50_disp_overlay_v0 {
 	__u8  version;
@@ -509,6 +520,7 @@ struct nv50_disp_overlay_v0 {
 	__u8  pad02[6];
 };
 
+#define NV50_DISP_OVERLAY_V0_NTFY_UEVENT                                   0x00
 
 /*******************************************************************************
  * fermi
diff --git a/drivers/gpu/drm/nouveau/nvif/client.c b/drivers/gpu/drm/nouveau/nvif/client.c
index 3c4df1fc26dc..3f7ac5bc8e03 100644
--- a/drivers/gpu/drm/nouveau/nvif/client.c
+++ b/drivers/gpu/drm/nouveau/nvif/client.c
@@ -62,6 +62,7 @@ nvif_drivers[] = {
 #else
 	&nvif_driver_drm,
 	&nvif_driver_lib,
+	&nvif_driver_null,
 #endif
 	NULL
 };
diff --git a/drivers/gpu/drm/nouveau/nvif/driver.h b/drivers/gpu/drm/nouveau/nvif/driver.h
index b72a8f0c2758..8bd39e69229c 100644
--- a/drivers/gpu/drm/nouveau/nvif/driver.h
+++ b/drivers/gpu/drm/nouveau/nvif/driver.h
@@ -9,13 +9,14 @@ struct nvif_driver {
 	int (*suspend)(void *priv);
 	int (*resume)(void *priv);
 	int (*ioctl)(void *priv, bool super, void *data, u32 size, void **hack);
-	void *(*map)(void *priv, u64 handle, u32 size);
-	void (*unmap)(void *priv, void *ptr, u32 size);
+	void __iomem *(*map)(void *priv, u64 handle, u32 size);
+	void (*unmap)(void *priv, void __iomem *ptr, u32 size);
 	bool keep;
 };
 
 extern const struct nvif_driver nvif_driver_nvkm;
 extern const struct nvif_driver nvif_driver_drm;
 extern const struct nvif_driver nvif_driver_lib;
+extern const struct nvif_driver nvif_driver_null;
 
 #endif
diff --git a/drivers/gpu/drm/nouveau/nvif/object.h b/drivers/gpu/drm/nouveau/nvif/object.h
index fac3a3bbec44..fe519179b76c 100644
--- a/drivers/gpu/drm/nouveau/nvif/object.h
+++ b/drivers/gpu/drm/nouveau/nvif/object.h
@@ -14,7 +14,7 @@ struct nvif_object {
 	void *priv; /*XXX: hack */
 	void (*dtor)(struct nvif_object *);
 	struct {
-		void *ptr;
+		void __iomem *ptr;
 		u32 size;
 	} map;
 };
@@ -42,7 +42,7 @@ void nvif_object_unmap(struct nvif_object *);
 	struct nvif_object *_object = nvif_object(a);                          \
 	u32 _data;                                                             \
 	if (likely(_object->map.ptr))                                          \
-		_data = ioread##b##_native((u8 *)_object->map.ptr + (c));      \
+		_data = ioread##b##_native((u8 __iomem *)_object->map.ptr + (c));      \
 	else                                                                   \
 		_data = nvif_object_rd(_object, (b) / 8, (c));                 \
 	_data;                                                                 \
@@ -50,7 +50,7 @@ void nvif_object_unmap(struct nvif_object *);
 #define nvif_wr(a,b,c,d) ({                                                    \
 	struct nvif_object *_object = nvif_object(a);                          \
 	if (likely(_object->map.ptr))                                          \
-		iowrite##b##_native((d), (u8 *)_object->map.ptr + (c));        \
+		iowrite##b##_native((d), (u8 __iomem *)_object->map.ptr + (c));        \
 	else                                                                   \
 		nvif_object_wr(_object, (b) / 8, (c), (d));                    \
 })
diff --git a/drivers/gpu/drm/omapdrm/omap_crtc.c b/drivers/gpu/drm/omapdrm/omap_crtc.c
index 2d28dc337cfb..b0566a1ca28f 100644
--- a/drivers/gpu/drm/omapdrm/omap_crtc.c
+++ b/drivers/gpu/drm/omapdrm/omap_crtc.c
@@ -20,6 +20,7 @@
 #include "omap_drv.h"
 
 #include <drm/drm_mode.h>
+#include <drm/drm_plane_helper.h>
 #include "drm_crtc.h"
 #include "drm_crtc_helper.h"
 
diff --git a/drivers/gpu/drm/omapdrm/omap_drv.c b/drivers/gpu/drm/omapdrm/omap_drv.c
index 002b9721e85a..8241ed9b353c 100644
--- a/drivers/gpu/drm/omapdrm/omap_drv.c
+++ b/drivers/gpu/drm/omapdrm/omap_drv.c
@@ -629,6 +629,7 @@ static struct drm_driver omap_drm_driver = {
 		.lastclose = dev_lastclose,
 		.preclose = dev_preclose,
 		.postclose = dev_postclose,
+		.set_busid = drm_platform_set_busid,
 		.get_vblank_counter = drm_vblank_count,
 		.enable_vblank = omap_irq_enable_vblank,
 		.disable_vblank = omap_irq_disable_vblank,
@@ -717,7 +718,6 @@ static const struct dev_pm_ops omapdrm_pm_ops = {
 static struct platform_driver pdev = {
 		.driver = {
 			.name = DRIVER_NAME,
-			.owner = THIS_MODULE,
 #ifdef CONFIG_PM
 			.pm = &omapdrm_pm_ops,
 #endif
diff --git a/drivers/gpu/drm/omapdrm/omap_drv.h b/drivers/gpu/drm/omapdrm/omap_drv.h
index 84d73a61b34b..60e47b33c801 100644
--- a/drivers/gpu/drm/omapdrm/omap_drv.h
+++ b/drivers/gpu/drm/omapdrm/omap_drv.h
@@ -26,6 +26,7 @@
 #include <drm/drmP.h>
 #include <drm/drm_crtc_helper.h>
 #include <drm/omap_drm.h>
+#include <drm/drm_gem.h>
 #include <linux/platform_data/omap_drm.h>
 
 
diff --git a/drivers/gpu/drm/omapdrm/omap_gem.c b/drivers/gpu/drm/omapdrm/omap_gem.c
index e4849413ee80..aeb91ed653c9 100644
--- a/drivers/gpu/drm/omapdrm/omap_gem.c
+++ b/drivers/gpu/drm/omapdrm/omap_gem.c
@@ -612,8 +612,7 @@ int omap_gem_dumb_create(struct drm_file *file, struct drm_device *dev,
 {
 	union omap_gem_size gsize;
 
-	/* in case someone tries to feed us a completely bogus stride: */
-	args->pitch = align_pitch(args->pitch, args->width, args->bpp);
+	args->pitch = align_pitch(0, args->width, args->bpp);
 	args->size = PAGE_ALIGN(args->pitch * args->height);
 
 	gsize = (union omap_gem_size){
diff --git a/drivers/gpu/drm/omapdrm/omap_plane.c b/drivers/gpu/drm/omapdrm/omap_plane.c
index 891a4dc608af..ee8e2b3a117e 100644
--- a/drivers/gpu/drm/omapdrm/omap_plane.c
+++ b/drivers/gpu/drm/omapdrm/omap_plane.c
@@ -388,20 +388,15 @@ struct drm_plane *omap_plane_init(struct drm_device *dev,
 	struct drm_plane *plane = NULL;
 	struct omap_plane *omap_plane;
 	struct omap_overlay_info *info;
-	int ret;
 
 	DBG("%s: priv=%d", plane_names[id], private_plane);
 
 	omap_plane = kzalloc(sizeof(*omap_plane), GFP_KERNEL);
 	if (!omap_plane)
-		goto fail;
+		return NULL;
 
-	ret = drm_flip_work_init(&omap_plane->unpin_work, 16,
+	drm_flip_work_init(&omap_plane->unpin_work,
 			"unpin", unpin_worker);
-	if (ret) {
-		dev_err(dev->dev, "could not allocate unpin FIFO\n");
-		goto fail;
-	}
 
 	omap_plane->nformats = omap_framebuffer_get_formats(
 			omap_plane->formats, ARRAY_SIZE(omap_plane->formats),
@@ -443,10 +438,4 @@ struct drm_plane *omap_plane_init(struct drm_device *dev,
 		omap_plane->info.zorder = id;
 
 	return plane;
-
-fail:
-	if (plane)
-		omap_plane_destroy(plane);
-
-	return NULL;
 }
diff --git a/drivers/gpu/drm/panel/Kconfig b/drivers/gpu/drm/panel/Kconfig
index bee9f72b3a93..024e98ef8e4d 100644
--- a/drivers/gpu/drm/panel/Kconfig
+++ b/drivers/gpu/drm/panel/Kconfig
@@ -27,4 +27,17 @@ config DRM_PANEL_S6E8AA0
 	select DRM_MIPI_DSI
 	select VIDEOMODE_HELPERS
 
+config DRM_PANEL_SHARP_LQ101R1SX01
+	tristate "Sharp LQ101R1SX01 panel"
+	depends on OF
+	depends on DRM_MIPI_DSI
+	help
+	  Say Y here if you want to enable support for Sharp LQ101R1SX01
+	  TFT-LCD modules. The panel has a 2560x1600 resolution and uses
+	  24 bit RGB per pixel. It provides a dual MIPI DSI interface to
+	  the host and has a built-in LED backlight.
+
+	  To compile this driver as a module, choose M here: the module
+	  will be called panel-sharp-lq101r1sx01.
+
 endmenu
diff --git a/drivers/gpu/drm/panel/Makefile b/drivers/gpu/drm/panel/Makefile
index 8b929212fad7..4b2a0430804b 100644
--- a/drivers/gpu/drm/panel/Makefile
+++ b/drivers/gpu/drm/panel/Makefile
@@ -1,3 +1,4 @@
 obj-$(CONFIG_DRM_PANEL_SIMPLE) += panel-simple.o
 obj-$(CONFIG_DRM_PANEL_LD9040) += panel-ld9040.o
 obj-$(CONFIG_DRM_PANEL_S6E8AA0) += panel-s6e8aa0.o
+obj-$(CONFIG_DRM_PANEL_SHARP_LQ101R1SX01) += panel-sharp-lq101r1sx01.o
diff --git a/drivers/gpu/drm/panel/panel-ld9040.c b/drivers/gpu/drm/panel/panel-ld9040.c
index 42ac67b21e9f..08cf2c588c3d 100644
--- a/drivers/gpu/drm/panel/panel-ld9040.c
+++ b/drivers/gpu/drm/panel/panel-ld9040.c
@@ -145,7 +145,7 @@ static void ld9040_dcs_write(struct ld9040 *ctx, const u8 *data, size_t len)
 	if (ctx->error < 0 || len == 0)
 		return;
 
-	dev_dbg(ctx->dev, "writing dcs seq: %*ph\n", len, data);
+	dev_dbg(ctx->dev, "writing dcs seq: %*ph\n", (int)len, data);
 	ret = ld9040_spi_write_word(ctx, *data);
 
 	while (!ret && --len) {
@@ -154,8 +154,8 @@ static void ld9040_dcs_write(struct ld9040 *ctx, const u8 *data, size_t len)
 	}
 
 	if (ret) {
-		dev_err(ctx->dev, "error %d writing dcs seq: %*ph\n", ret, len,
-			data);
+		dev_err(ctx->dev, "error %d writing dcs seq: %*ph\n", ret,
+			(int)len, data);
 		ctx->error = ret;
 	}
 
@@ -336,17 +336,12 @@ static int ld9040_probe(struct spi_device *spi)
 	if (ret < 0)
 		return ret;
 
-	ctx->reset_gpio = devm_gpiod_get(dev, "reset");
+	ctx->reset_gpio = devm_gpiod_get(dev, "reset", GPIOD_OUT_HIGH);
 	if (IS_ERR(ctx->reset_gpio)) {
 		dev_err(dev, "cannot get reset-gpios %ld\n",
 			PTR_ERR(ctx->reset_gpio));
 		return PTR_ERR(ctx->reset_gpio);
 	}
-	ret = gpiod_direction_output(ctx->reset_gpio, 1);
-	if (ret < 0) {
-		dev_err(dev, "cannot configure reset-gpios %d\n", ret);
-		return ret;
-	}
 
 	spi->bits_per_word = 9;
 	ret = spi_setup(spi);
diff --git a/drivers/gpu/drm/panel/panel-s6e8aa0.c b/drivers/gpu/drm/panel/panel-s6e8aa0.c
index b5217fe37f02..144b2733e3d7 100644
--- a/drivers/gpu/drm/panel/panel-s6e8aa0.c
+++ b/drivers/gpu/drm/panel/panel-s6e8aa0.c
@@ -141,10 +141,10 @@ static void s6e8aa0_dcs_write(struct s6e8aa0 *ctx, const void *data, size_t len)
 	if (ctx->error < 0)
 		return;
 
-	ret = mipi_dsi_dcs_write(dsi, data, len);
+	ret = mipi_dsi_dcs_write_buffer(dsi, data, len);
 	if (ret < 0) {
-		dev_err(ctx->dev, "error %zd writing dcs seq: %*ph\n", ret, len,
-			data);
+		dev_err(ctx->dev, "error %zd writing dcs seq: %*ph\n", ret,
+			(int)len, data);
 		ctx->error = ret;
 	}
 }
@@ -800,27 +800,15 @@ static void s6e8aa0_panel_init(struct s6e8aa0 *ctx)
 }
 
 static void s6e8aa0_set_maximum_return_packet_size(struct s6e8aa0 *ctx,
-						   int size)
+						   u16 size)
 {
 	struct mipi_dsi_device *dsi = to_mipi_dsi_device(ctx->dev);
-	const struct mipi_dsi_host_ops *ops = dsi->host->ops;
-	u8 buf[] = {size, 0};
-	struct mipi_dsi_msg msg = {
-		.channel = dsi->channel,
-		.type = MIPI_DSI_SET_MAXIMUM_RETURN_PACKET_SIZE,
-		.tx_len = sizeof(buf),
-		.tx_buf = buf
-	};
 	int ret;
 
 	if (ctx->error < 0)
 		return;
 
-	if (!ops || !ops->transfer)
-		ret = -EIO;
-	else
-		ret = ops->transfer(dsi->host, &msg);
-
+	ret = mipi_dsi_set_maximum_return_packet_size(dsi, size);
 	if (ret < 0) {
 		dev_err(ctx->dev,
 			"error %d setting maximum return packet size to %d\n",
@@ -1019,17 +1007,12 @@ static int s6e8aa0_probe(struct mipi_dsi_device *dsi)
 		return ret;
 	}
 
-	ctx->reset_gpio = devm_gpiod_get(dev, "reset");
+	ctx->reset_gpio = devm_gpiod_get(dev, "reset", GPIOD_OUT_HIGH);
 	if (IS_ERR(ctx->reset_gpio)) {
 		dev_err(dev, "cannot get reset-gpios %ld\n",
 			PTR_ERR(ctx->reset_gpio));
 		return PTR_ERR(ctx->reset_gpio);
 	}
-	ret = gpiod_direction_output(ctx->reset_gpio, 1);
-	if (ret < 0) {
-		dev_err(dev, "cannot configure reset-gpios %d\n", ret);
-		return ret;
-	}
 
 	ctx->brightness = GAMMA_LEVEL_NUM - 1;
 
@@ -1069,7 +1052,6 @@ static struct mipi_dsi_driver s6e8aa0_driver = {
 	.remove = s6e8aa0_remove,
 	.driver = {
 		.name = "panel_s6e8aa0",
-		.owner = THIS_MODULE,
 		.of_match_table = s6e8aa0_of_match,
 	},
 };
diff --git a/drivers/gpu/drm/panel/panel-sharp-lq101r1sx01.c b/drivers/gpu/drm/panel/panel-sharp-lq101r1sx01.c
new file mode 100644
index 000000000000..9d81759d82fc
--- /dev/null
+++ b/drivers/gpu/drm/panel/panel-sharp-lq101r1sx01.c
@@ -0,0 +1,464 @@
+/*
+ * Copyright (C) 2014 NVIDIA Corporation
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+
+#include <linux/backlight.h>
+#include <linux/gpio/consumer.h>
+#include <linux/module.h>
+#include <linux/of.h>
+#include <linux/regulator/consumer.h>
+
+#include <drm/drmP.h>
+#include <drm/drm_crtc.h>
+#include <drm/drm_mipi_dsi.h>
+#include <drm/drm_panel.h>
+
+#include <video/mipi_display.h>
+
+#include <linux/host1x.h>
+
+struct sharp_panel {
+	struct drm_panel base;
+	/* the datasheet refers to them as DSI-LINK1 and DSI-LINK2 */
+	struct mipi_dsi_device *link1;
+	struct mipi_dsi_device *link2;
+
+	struct backlight_device *backlight;
+	struct regulator *supply;
+
+	bool prepared;
+	bool enabled;
+
+	const struct drm_display_mode *mode;
+};
+
+static inline struct sharp_panel *to_sharp_panel(struct drm_panel *panel)
+{
+	return container_of(panel, struct sharp_panel, base);
+}
+
+static int sharp_panel_write(struct sharp_panel *sharp, u16 offset, u8 value)
+{
+	u8 payload[3] = { offset >> 8, offset & 0xff, value };
+	struct mipi_dsi_device *dsi = sharp->link1;
+	ssize_t err;
+
+	err = mipi_dsi_generic_write(dsi, payload, sizeof(payload));
+	if (err < 0) {
+		dev_err(&dsi->dev, "failed to write %02x to %04x: %zd\n",
+			value, offset, err);
+		return err;
+	}
+
+	err = mipi_dsi_dcs_nop(dsi);
+	if (err < 0) {
+		dev_err(&dsi->dev, "failed to send DCS nop: %zd\n", err);
+		return err;
+	}
+
+	usleep_range(10, 20);
+
+	return 0;
+}
+
+static __maybe_unused int sharp_panel_read(struct sharp_panel *sharp,
+					   u16 offset, u8 *value)
+{
+	ssize_t err;
+
+	cpu_to_be16s(&offset);
+
+	err = mipi_dsi_generic_read(sharp->link1, &offset, sizeof(offset),
+				    value, sizeof(*value));
+	if (err < 0)
+		dev_err(&sharp->link1->dev, "failed to read from %04x: %zd\n",
+			offset, err);
+
+	return err;
+}
+
+static int sharp_panel_disable(struct drm_panel *panel)
+{
+	struct sharp_panel *sharp = to_sharp_panel(panel);
+
+	if (!sharp->enabled)
+		return 0;
+
+	if (sharp->backlight) {
+		sharp->backlight->props.power = FB_BLANK_POWERDOWN;
+		backlight_update_status(sharp->backlight);
+	}
+
+	sharp->enabled = false;
+
+	return 0;
+}
+
+static int sharp_panel_unprepare(struct drm_panel *panel)
+{
+	struct sharp_panel *sharp = to_sharp_panel(panel);
+	int err;
+
+	if (!sharp->prepared)
+		return 0;
+
+	err = mipi_dsi_dcs_set_display_off(sharp->link1);
+	if (err < 0)
+		dev_err(panel->dev, "failed to set display off: %d\n", err);
+
+	err = mipi_dsi_dcs_enter_sleep_mode(sharp->link1);
+	if (err < 0)
+		dev_err(panel->dev, "failed to enter sleep mode: %d\n", err);
+
+	msleep(120);
+
+	regulator_disable(sharp->supply);
+
+	sharp->prepared = false;
+
+	return 0;
+}
+
+static int sharp_setup_symmetrical_split(struct mipi_dsi_device *left,
+					 struct mipi_dsi_device *right,
+					 const struct drm_display_mode *mode)
+{
+	int err;
+
+	err = mipi_dsi_dcs_set_column_address(left, 0, mode->hdisplay / 2 - 1);
+	if (err < 0) {
+		dev_err(&left->dev, "failed to set column address: %d\n", err);
+		return err;
+	}
+
+	err = mipi_dsi_dcs_set_page_address(left, 0, mode->vdisplay - 1);
+	if (err < 0) {
+		dev_err(&left->dev, "failed to set page address: %d\n", err);
+		return err;
+	}
+
+	err = mipi_dsi_dcs_set_column_address(right, mode->hdisplay / 2,
+					      mode->hdisplay - 1);
+	if (err < 0) {
+		dev_err(&right->dev, "failed to set column address: %d\n", err);
+		return err;
+	}
+
+	err = mipi_dsi_dcs_set_page_address(right, 0, mode->vdisplay - 1);
+	if (err < 0) {
+		dev_err(&right->dev, "failed to set page address: %d\n", err);
+		return err;
+	}
+
+	return 0;
+}
+
+static int sharp_panel_prepare(struct drm_panel *panel)
+{
+	struct sharp_panel *sharp = to_sharp_panel(panel);
+	u8 format = MIPI_DCS_PIXEL_FMT_24BIT;
+	int err;
+
+	if (sharp->prepared)
+		return 0;
+
+	err = regulator_enable(sharp->supply);
+	if (err < 0)
+		return err;
+
+	usleep_range(10000, 20000);
+
+	err = mipi_dsi_dcs_soft_reset(sharp->link1);
+	if (err < 0) {
+		dev_err(panel->dev, "soft reset failed: %d\n", err);
+		goto poweroff;
+	}
+
+	msleep(120);
+
+	err = mipi_dsi_dcs_exit_sleep_mode(sharp->link1);
+	if (err < 0) {
+		dev_err(panel->dev, "failed to exit sleep mode: %d\n", err);
+		goto poweroff;
+	}
+
+	/*
+	 * The MIPI DCS specification mandates this delay only between the
+	 * exit_sleep_mode and enter_sleep_mode commands, so it isn't strictly
+	 * necessary here.
+	 */
+	/*
+	msleep(120);
+	*/
+
+	/* set left-right mode */
+	err = sharp_panel_write(sharp, 0x1000, 0x2a);
+	if (err < 0) {
+		dev_err(panel->dev, "failed to set left-right mode: %d\n", err);
+		goto poweroff;
+	}
+
+	/* enable command mode */
+	err = sharp_panel_write(sharp, 0x1001, 0x01);
+	if (err < 0) {
+		dev_err(panel->dev, "failed to enable command mode: %d\n", err);
+		goto poweroff;
+	}
+
+	err = mipi_dsi_dcs_set_pixel_format(sharp->link1, format);
+	if (err < 0) {
+		dev_err(panel->dev, "failed to set pixel format: %d\n", err);
+		goto poweroff;
+	}
+
+	/*
+	 * TODO: The device supports both left-right and even-odd split
+	 * configurations, but this driver currently supports only the left-
+	 * right split. To support a different mode a mechanism needs to be
+	 * put in place to communicate the configuration back to the DSI host
+	 * controller.
+	 */
+	err = sharp_setup_symmetrical_split(sharp->link1, sharp->link2,
+					    sharp->mode);
+	if (err < 0) {
+		dev_err(panel->dev, "failed to set up symmetrical split: %d\n",
+			err);
+		goto poweroff;
+	}
+
+	err = mipi_dsi_dcs_set_display_on(sharp->link1);
+	if (err < 0) {
+		dev_err(panel->dev, "failed to set display on: %d\n", err);
+		goto poweroff;
+	}
+
+	sharp->prepared = true;
+
+	return 0;
+
+poweroff:
+	regulator_disable(sharp->supply);
+	return err;
+}
+
+static int sharp_panel_enable(struct drm_panel *panel)
+{
+	struct sharp_panel *sharp = to_sharp_panel(panel);
+
+	if (sharp->enabled)
+		return 0;
+
+	if (sharp->backlight) {
+		sharp->backlight->props.power = FB_BLANK_UNBLANK;
+		backlight_update_status(sharp->backlight);
+	}
+
+	sharp->enabled = true;
+
+	return 0;
+}
+
+static const struct drm_display_mode default_mode = {
+	.clock = 278000,
+	.hdisplay = 2560,
+	.hsync_start = 2560 + 128,
+	.hsync_end = 2560 + 128 + 64,
+	.htotal = 2560 + 128 + 64 + 64,
+	.vdisplay = 1600,
+	.vsync_start = 1600 + 4,
+	.vsync_end = 1600 + 4 + 8,
+	.vtotal = 1600 + 4 + 8 + 32,
+	.vrefresh = 60,
+};
+
+static int sharp_panel_get_modes(struct drm_panel *panel)
+{
+	struct drm_display_mode *mode;
+
+	mode = drm_mode_duplicate(panel->drm, &default_mode);
+	if (!mode) {
+		dev_err(panel->drm->dev, "failed to add mode %ux%ux@%u\n",
+			default_mode.hdisplay, default_mode.vdisplay,
+			default_mode.vrefresh);
+		return -ENOMEM;
+	}
+
+	drm_mode_set_name(mode);
+
+	drm_mode_probed_add(panel->connector, mode);
+
+	panel->connector->display_info.width_mm = 217;
+	panel->connector->display_info.height_mm = 136;
+
+	return 1;
+}
+
+static const struct drm_panel_funcs sharp_panel_funcs = {
+	.disable = sharp_panel_disable,
+	.unprepare = sharp_panel_unprepare,
+	.prepare = sharp_panel_prepare,
+	.enable = sharp_panel_enable,
+	.get_modes = sharp_panel_get_modes,
+};
+
+static const struct of_device_id sharp_of_match[] = {
+	{ .compatible = "sharp,lq101r1sx01", },
+	{ }
+};
+MODULE_DEVICE_TABLE(of, sharp_of_match);
+
+static int sharp_panel_add(struct sharp_panel *sharp)
+{
+	struct device_node *np;
+	int err;
+
+	sharp->mode = &default_mode;
+
+	sharp->supply = devm_regulator_get(&sharp->link1->dev, "power");
+	if (IS_ERR(sharp->supply))
+		return PTR_ERR(sharp->supply);
+
+	np = of_parse_phandle(sharp->link1->dev.of_node, "backlight", 0);
+	if (np) {
+		sharp->backlight = of_find_backlight_by_node(np);
+		of_node_put(np);
+
+		if (!sharp->backlight)
+			return -EPROBE_DEFER;
+	}
+
+	drm_panel_init(&sharp->base);
+	sharp->base.funcs = &sharp_panel_funcs;
+	sharp->base.dev = &sharp->link1->dev;
+
+	err = drm_panel_add(&sharp->base);
+	if (err < 0)
+		goto put_backlight;
+
+	return 0;
+
+put_backlight:
+	if (sharp->backlight)
+		put_device(&sharp->backlight->dev);
+
+	return err;
+}
+
+static void sharp_panel_del(struct sharp_panel *sharp)
+{
+	if (sharp->base.dev)
+		drm_panel_remove(&sharp->base);
+
+	if (sharp->backlight)
+		put_device(&sharp->backlight->dev);
+
+	if (sharp->link2)
+		put_device(&sharp->link2->dev);
+}
+
+static int sharp_panel_probe(struct mipi_dsi_device *dsi)
+{
+	struct mipi_dsi_device *secondary = NULL;
+	struct sharp_panel *sharp;
+	struct device_node *np;
+	int err;
+
+	dsi->lanes = 4;
+	dsi->format = MIPI_DSI_FMT_RGB888;
+	dsi->mode_flags = MIPI_DSI_MODE_LPM;
+
+	/* Find DSI-LINK1 */
+	np = of_parse_phandle(dsi->dev.of_node, "link2", 0);
+	if (np) {
+		secondary = of_find_mipi_dsi_device_by_node(np);
+		of_node_put(np);
+
+		if (!secondary)
+			return -EPROBE_DEFER;
+	}
+
+	/* register a panel for only the DSI-LINK1 interface */
+	if (secondary) {
+		sharp = devm_kzalloc(&dsi->dev, sizeof(*sharp), GFP_KERNEL);
+		if (!sharp) {
+			put_device(&secondary->dev);
+			return -ENOMEM;
+		}
+
+		mipi_dsi_set_drvdata(dsi, sharp);
+
+		sharp->link2 = secondary;
+		sharp->link1 = dsi;
+
+		err = sharp_panel_add(sharp);
+		if (err < 0) {
+			put_device(&secondary->dev);
+			return err;
+		}
+	}
+
+	err = mipi_dsi_attach(dsi);
+	if (err < 0) {
+		if (secondary)
+			sharp_panel_del(sharp);
+
+		return err;
+	}
+
+	return 0;
+}
+
+static int sharp_panel_remove(struct mipi_dsi_device *dsi)
+{
+	struct sharp_panel *sharp = mipi_dsi_get_drvdata(dsi);
+	int err;
+
+	/* only detach from host for the DSI-LINK2 interface */
+	if (!sharp) {
+		mipi_dsi_detach(dsi);
+		return 0;
+	}
+
+	err = sharp_panel_disable(&sharp->base);
+	if (err < 0)
+		dev_err(&dsi->dev, "failed to disable panel: %d\n", err);
+
+	err = mipi_dsi_detach(dsi);
+	if (err < 0)
+		dev_err(&dsi->dev, "failed to detach from DSI host: %d\n", err);
+
+	drm_panel_detach(&sharp->base);
+	sharp_panel_del(sharp);
+
+	return 0;
+}
+
+static void sharp_panel_shutdown(struct mipi_dsi_device *dsi)
+{
+	struct sharp_panel *sharp = mipi_dsi_get_drvdata(dsi);
+
+	/* nothing to do for DSI-LINK2 */
+	if (!sharp)
+		return;
+
+	sharp_panel_disable(&sharp->base);
+}
+
+static struct mipi_dsi_driver sharp_panel_driver = {
+	.driver = {
+		.name = "panel-sharp-lq101r1sx01",
+		.of_match_table = sharp_of_match,
+	},
+	.probe = sharp_panel_probe,
+	.remove = sharp_panel_remove,
+	.shutdown = sharp_panel_shutdown,
+};
+module_mipi_dsi_driver(sharp_panel_driver);
+
+MODULE_AUTHOR("Thierry Reding <treding@nvidia.com>");
+MODULE_DESCRIPTION("Sharp LQ101R1SX01 panel driver");
+MODULE_LICENSE("GPL v2");
diff --git a/drivers/gpu/drm/panel/panel-simple.c b/drivers/gpu/drm/panel/panel-simple.c
index 4ce1db0a68ff..e95385bf8356 100644
--- a/drivers/gpu/drm/panel/panel-simple.c
+++ b/drivers/gpu/drm/panel/panel-simple.c
@@ -247,21 +247,14 @@ static int panel_simple_probe(struct device *dev, const struct panel_desc *desc)
 	if (IS_ERR(panel->supply))
 		return PTR_ERR(panel->supply);
 
-	panel->enable_gpio = devm_gpiod_get_optional(dev, "enable");
+	panel->enable_gpio = devm_gpiod_get_optional(dev, "enable",
+						     GPIOD_OUT_LOW);
 	if (IS_ERR(panel->enable_gpio)) {
 		err = PTR_ERR(panel->enable_gpio);
 		dev_err(dev, "failed to request GPIO: %d\n", err);
 		return err;
 	}
 
-	if (panel->enable_gpio) {
-		err = gpiod_direction_output(panel->enable_gpio, 0);
-		if (err < 0) {
-			dev_err(dev, "failed to setup GPIO: %d\n", err);
-			return err;
-		}
-	}
-
 	backlight = of_parse_phandle(dev->of_node, "backlight", 0);
 	if (backlight) {
 		panel->backlight = of_find_backlight_by_node(backlight);
@@ -352,6 +345,53 @@ static const struct panel_desc auo_b101aw03 = {
 	},
 };
 
+static const struct drm_display_mode auo_b101xtn01_mode = {
+	.clock = 72000,
+	.hdisplay = 1366,
+	.hsync_start = 1366 + 20,
+	.hsync_end = 1366 + 20 + 70,
+	.htotal = 1366 + 20 + 70,
+	.vdisplay = 768,
+	.vsync_start = 768 + 14,
+	.vsync_end = 768 + 14 + 42,
+	.vtotal = 768 + 14 + 42,
+	.vrefresh = 60,
+	.flags = DRM_MODE_FLAG_NVSYNC | DRM_MODE_FLAG_NHSYNC,
+};
+
+static const struct panel_desc auo_b101xtn01 = {
+	.modes = &auo_b101xtn01_mode,
+	.num_modes = 1,
+	.bpc = 6,
+	.size = {
+		.width = 223,
+		.height = 125,
+	},
+};
+
+static const struct drm_display_mode auo_b116xw03_mode = {
+	.clock = 70589,
+	.hdisplay = 1366,
+	.hsync_start = 1366 + 40,
+	.hsync_end = 1366 + 40 + 40,
+	.htotal = 1366 + 40 + 40 + 32,
+	.vdisplay = 768,
+	.vsync_start = 768 + 10,
+	.vsync_end = 768 + 10 + 12,
+	.vtotal = 768 + 10 + 12 + 6,
+	.vrefresh = 60,
+};
+
+static const struct panel_desc auo_b116xw03 = {
+	.modes = &auo_b116xw03_mode,
+	.num_modes = 1,
+	.bpc = 6,
+	.size = {
+		.width = 256,
+		.height = 144,
+	},
+};
+
 static const struct drm_display_mode auo_b133xtn01_mode = {
 	.clock = 69500,
 	.hdisplay = 1366,
@@ -391,6 +431,7 @@ static const struct drm_display_mode auo_b133htn01_mode = {
 static const struct panel_desc auo_b133htn01 = {
 	.modes = &auo_b133htn01_mode,
 	.num_modes = 1,
+	.bpc = 6,
 	.size = {
 		.width = 293,
 		.height = 165,
@@ -512,22 +553,92 @@ static const struct drm_display_mode foxlink_fl500wvr00_a0t_mode = {
 static const struct panel_desc foxlink_fl500wvr00_a0t = {
 	.modes = &foxlink_fl500wvr00_a0t_mode,
 	.num_modes = 1,
+	.bpc = 8,
 	.size = {
 		.width = 108,
 		.height = 65,
 	},
 };
 
-static const struct drm_display_mode innolux_n116bge_mode = {
+static const struct drm_display_mode hannstar_hsd070pww1_mode = {
+	.clock = 71100,
+	.hdisplay = 1280,
+	.hsync_start = 1280 + 1,
+	.hsync_end = 1280 + 1 + 158,
+	.htotal = 1280 + 1 + 158 + 1,
+	.vdisplay = 800,
+	.vsync_start = 800 + 1,
+	.vsync_end = 800 + 1 + 21,
+	.vtotal = 800 + 1 + 21 + 1,
+	.vrefresh = 60,
+};
+
+static const struct panel_desc hannstar_hsd070pww1 = {
+	.modes = &hannstar_hsd070pww1_mode,
+	.num_modes = 1,
+	.bpc = 6,
+	.size = {
+		.width = 151,
+		.height = 94,
+	},
+};
+
+static const struct drm_display_mode hitachi_tx23d38vm0caa_mode = {
+	.clock = 33333,
+	.hdisplay = 800,
+	.hsync_start = 800 + 85,
+	.hsync_end = 800 + 85 + 86,
+	.htotal = 800 + 85 + 86 + 85,
+	.vdisplay = 480,
+	.vsync_start = 480 + 16,
+	.vsync_end = 480 + 16 + 13,
+	.vtotal = 480 + 16 + 13 + 16,
+	.vrefresh = 60,
+};
+
+static const struct panel_desc hitachi_tx23d38vm0caa = {
+	.modes = &hitachi_tx23d38vm0caa_mode,
+	.num_modes = 1,
+	.bpc = 6,
+	.size = {
+		.width = 195,
+		.height = 117,
+	},
+};
+
+static const struct drm_display_mode innolux_g121i1_l01_mode = {
 	.clock = 71000,
+	.hdisplay = 1280,
+	.hsync_start = 1280 + 64,
+	.hsync_end = 1280 + 64 + 32,
+	.htotal = 1280 + 64 + 32 + 64,
+	.vdisplay = 800,
+	.vsync_start = 800 + 9,
+	.vsync_end = 800 + 9 + 6,
+	.vtotal = 800 + 9 + 6 + 9,
+	.vrefresh = 60,
+};
+
+static const struct panel_desc innolux_g121i1_l01 = {
+	.modes = &innolux_g121i1_l01_mode,
+	.num_modes = 1,
+	.bpc = 6,
+	.size = {
+		.width = 261,
+		.height = 163,
+	},
+};
+
+static const struct drm_display_mode innolux_n116bge_mode = {
+	.clock = 76420,
 	.hdisplay = 1366,
-	.hsync_start = 1366 + 64,
-	.hsync_end = 1366 + 64 + 6,
-	.htotal = 1366 + 64 + 6 + 64,
+	.hsync_start = 1366 + 136,
+	.hsync_end = 1366 + 136 + 30,
+	.htotal = 1366 + 136 + 30 + 60,
 	.vdisplay = 768,
 	.vsync_start = 768 + 8,
-	.vsync_end = 768 + 8 + 4,
-	.vtotal = 768 + 8 + 4 + 8,
+	.vsync_end = 768 + 8 + 12,
+	.vtotal = 768 + 8 + 12 + 12,
 	.vrefresh = 60,
 	.flags = DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC,
 };
@@ -616,6 +727,12 @@ static const struct of_device_id platform_of_match[] = {
 		.compatible = "auo,b101aw03",
 		.data = &auo_b101aw03,
 	}, {
+		.compatible = "auo,b101xtn01",
+		.data = &auo_b101xtn01,
+	}, {
+		.compatible = "auo,b116xw03",
+		.data = &auo_b116xw03,
+	}, {
 		.compatible = "auo,b133htn01",
 		.data = &auo_b133htn01,
 	}, {
@@ -640,6 +757,15 @@ static const struct of_device_id platform_of_match[] = {
 		.compatible = "foxlink,fl500wvr00-a0t",
 		.data = &foxlink_fl500wvr00_a0t,
 	}, {
+		.compatible = "hannstar,hsd070pww1",
+		.data = &hannstar_hsd070pww1,
+	}, {
+		.compatible = "hit,tx23d38vm0caa",
+		.data = &hitachi_tx23d38vm0caa
+	}, {
+		.compatible ="innolux,g121i1-l01",
+		.data = &innolux_g121i1_l01
+	}, {
 		.compatible = "innolux,n116bge",
 		.data = &innolux_n116bge,
 	}, {
@@ -681,7 +807,6 @@ static void panel_simple_platform_shutdown(struct platform_device *pdev)
 static struct platform_driver panel_simple_platform_driver = {
 	.driver = {
 		.name = "panel-simple",
-		.owner = THIS_MODULE,
 		.of_match_table = platform_of_match,
 	},
 	.probe = panel_simple_platform_probe,
@@ -714,6 +839,7 @@ static const struct panel_desc_dsi lg_ld070wx3_sl01 = {
 	.desc = {
 		.modes = &lg_ld070wx3_sl01_mode,
 		.num_modes = 1,
+		.bpc = 8,
 		.size = {
 			.width = 94,
 			.height = 151,
@@ -741,6 +867,7 @@ static const struct panel_desc_dsi lg_lh500wx1_sd03 = {
 	.desc = {
 		.modes = &lg_lh500wx1_sd03_mode,
 		.num_modes = 1,
+		.bpc = 8,
 		.size = {
 			.width = 62,
 			.height = 110,
@@ -768,6 +895,7 @@ static const struct panel_desc_dsi panasonic_vvx10f004b00 = {
 	.desc = {
 		.modes = &panasonic_vvx10f004b00_mode,
 		.num_modes = 1,
+		.bpc = 8,
 		.size = {
 			.width = 217,
 			.height = 136,
@@ -837,7 +965,6 @@ static void panel_simple_dsi_shutdown(struct mipi_dsi_device *dsi)
 static struct mipi_dsi_driver panel_simple_dsi_driver = {
 	.driver = {
 		.name = "panel-simple-dsi",
-		.owner = THIS_MODULE,
 		.of_match_table = dsi_of_match,
 	},
 	.probe = panel_simple_dsi_probe,
diff --git a/drivers/gpu/drm/qxl/Makefile b/drivers/gpu/drm/qxl/Makefile
index ea046ba691d2..bacc4aff1201 100644
--- a/drivers/gpu/drm/qxl/Makefile
+++ b/drivers/gpu/drm/qxl/Makefile
@@ -4,6 +4,6 @@
 
 ccflags-y := -Iinclude/drm
 
-qxl-y := qxl_drv.o qxl_kms.o qxl_display.o qxl_ttm.o qxl_fb.o qxl_object.o qxl_gem.o qxl_cmd.o qxl_image.o qxl_draw.o qxl_debugfs.o qxl_irq.o qxl_dumb.o qxl_ioctl.o qxl_fence.o qxl_release.o
+qxl-y := qxl_drv.o qxl_kms.o qxl_display.o qxl_ttm.o qxl_fb.o qxl_object.o qxl_gem.o qxl_cmd.o qxl_image.o qxl_draw.o qxl_debugfs.o qxl_irq.o qxl_dumb.o qxl_ioctl.o qxl_release.o qxl_prime.o
 
 obj-$(CONFIG_DRM_QXL)+= qxl.o
diff --git a/drivers/gpu/drm/qxl/qxl_cmd.c b/drivers/gpu/drm/qxl/qxl_cmd.c
index eb89653a7a17..97823644d347 100644
--- a/drivers/gpu/drm/qxl/qxl_cmd.c
+++ b/drivers/gpu/drm/qxl/qxl_cmd.c
@@ -620,17 +620,10 @@ static int qxl_reap_surf(struct qxl_device *qdev, struct qxl_bo *surf, bool stal
 	if (ret == -EBUSY)
 		return -EBUSY;
 
-	if (surf->fence.num_active_releases > 0 && stall == false) {
-		qxl_bo_unreserve(surf);
-		return -EBUSY;
-	}
-
 	if (stall)
 		mutex_unlock(&qdev->surf_evict_mutex);
 
-	spin_lock(&surf->tbo.bdev->fence_lock);
 	ret = ttm_bo_wait(&surf->tbo, true, true, !stall);
-	spin_unlock(&surf->tbo.bdev->fence_lock);
 
 	if (stall)
 		mutex_lock(&qdev->surf_evict_mutex);
diff --git a/drivers/gpu/drm/qxl/qxl_debugfs.c b/drivers/gpu/drm/qxl/qxl_debugfs.c
index c3c2bbdc6674..6911b8c44492 100644
--- a/drivers/gpu/drm/qxl/qxl_debugfs.c
+++ b/drivers/gpu/drm/qxl/qxl_debugfs.c
@@ -58,9 +58,17 @@ qxl_debugfs_buffers_info(struct seq_file *m, void *data)
 	struct qxl_bo *bo;
 
 	list_for_each_entry(bo, &qdev->gem.objects, list) {
-		seq_printf(m, "size %ld, pc %d, sync obj %p, num releases %d\n",
-			   (unsigned long)bo->gem_base.size, bo->pin_count,
-			   bo->tbo.sync_obj, bo->fence.num_active_releases);
+		struct reservation_object_list *fobj;
+		int rel;
+
+		rcu_read_lock();
+		fobj = rcu_dereference(bo->tbo.resv->fence);
+		rel = fobj ? fobj->shared_count : 0;
+		rcu_read_unlock();
+
+		seq_printf(m, "size %ld, pc %d, num releases %d\n",
+			   (unsigned long)bo->gem_base.size,
+			   bo->pin_count, rel);
 	}
 	return 0;
 }
diff --git a/drivers/gpu/drm/qxl/qxl_display.c b/drivers/gpu/drm/qxl/qxl_display.c
index b8ced08b6291..4a0a8b29b0a1 100644
--- a/drivers/gpu/drm/qxl/qxl_display.c
+++ b/drivers/gpu/drm/qxl/qxl_display.c
@@ -29,6 +29,7 @@
 #include "qxl_drv.h"
 #include "qxl_object.h"
 #include "drm_crtc_helper.h"
+#include <drm/drm_plane_helper.h>
 
 static bool qxl_head_enabled(struct qxl_head *head)
 {
@@ -100,14 +101,37 @@ static int qxl_display_copy_rom_client_monitors_config(struct qxl_device *qdev)
 	return 0;
 }
 
+static void qxl_update_offset_props(struct qxl_device *qdev)
+{
+	struct drm_device *dev = qdev->ddev;
+	struct drm_connector *connector;
+	struct qxl_output *output;
+	struct qxl_head *head;
+
+	list_for_each_entry(connector, &dev->mode_config.connector_list, head) {
+		output = drm_connector_to_qxl_output(connector);
+
+		head = &qdev->client_monitors_config->heads[output->index];
+
+		drm_object_property_set_value(&connector->base,
+			dev->mode_config.suggested_x_property, head->x);
+		drm_object_property_set_value(&connector->base,
+			dev->mode_config.suggested_y_property, head->y);
+	}
+}
+
 void qxl_display_read_client_monitors_config(struct qxl_device *qdev)
 {
 
+	struct drm_device *dev = qdev->ddev;
 	while (qxl_display_copy_rom_client_monitors_config(qdev)) {
 		qxl_io_log(qdev, "failed crc check for client_monitors_config,"
 				 " retrying\n");
 	}
 
+	drm_modeset_lock_all(dev);
+	qxl_update_offset_props(qdev);
+	drm_modeset_unlock_all(dev);
 	if (!drm_helper_hpd_irq_event(qdev->ddev)) {
 		/* notify that the monitor configuration changed, to
 		   adjust at the arbitrary resolution */
@@ -187,6 +211,54 @@ static void qxl_crtc_destroy(struct drm_crtc *crtc)
 	kfree(qxl_crtc);
 }
 
+static int qxl_crtc_page_flip(struct drm_crtc *crtc,
+                              struct drm_framebuffer *fb,
+                              struct drm_pending_vblank_event *event,
+                              uint32_t page_flip_flags)
+{
+	struct drm_device *dev = crtc->dev;
+	struct qxl_device *qdev = dev->dev_private;
+	struct qxl_crtc *qcrtc = to_qxl_crtc(crtc);
+	struct qxl_framebuffer *qfb_src = to_qxl_framebuffer(fb);
+	struct qxl_framebuffer *qfb_old = to_qxl_framebuffer(crtc->primary->fb);
+	struct qxl_bo *bo_old = gem_to_qxl_bo(qfb_old->obj);
+	struct qxl_bo *bo = gem_to_qxl_bo(qfb_src->obj);
+	unsigned long flags;
+	struct drm_clip_rect norect = {
+	    .x1 = 0,
+	    .y1 = 0,
+	    .x2 = fb->width,
+	    .y2 = fb->height
+	};
+	int inc = 1;
+	int one_clip_rect = 1;
+	int ret = 0;
+
+	crtc->primary->fb = fb;
+	bo_old->is_primary = false;
+	bo->is_primary = true;
+
+	ret = qxl_bo_reserve(bo, false);
+	if (ret)
+		return ret;
+
+	qxl_draw_dirty_fb(qdev, qfb_src, bo, 0, 0,
+			  &norect, one_clip_rect, inc);
+
+	drm_vblank_get(dev, qcrtc->index);
+
+	if (event) {
+		spin_lock_irqsave(&dev->event_lock, flags);
+		drm_send_vblank_event(dev, qcrtc->index, event);
+		spin_unlock_irqrestore(&dev->event_lock, flags);
+	}
+	drm_vblank_put(dev, qcrtc->index);
+
+	qxl_bo_unreserve(bo);
+
+	return 0;
+}
+
 static int
 qxl_hide_cursor(struct qxl_device *qdev)
 {
@@ -374,6 +446,7 @@ static const struct drm_crtc_funcs qxl_crtc_funcs = {
 	.cursor_move = qxl_crtc_cursor_move,
 	.set_config = drm_crtc_helper_set_config,
 	.destroy = qxl_crtc_destroy,
+	.page_flip = qxl_crtc_page_flip,
 };
 
 static void qxl_user_framebuffer_destroy(struct drm_framebuffer *fb)
@@ -519,11 +592,9 @@ static int qxl_crtc_mode_set(struct drm_crtc *crtc,
 {
 	struct drm_device *dev = crtc->dev;
 	struct qxl_device *qdev = dev->dev_private;
-	struct qxl_mode *m = (void *)mode->private;
 	struct qxl_framebuffer *qfb;
 	struct qxl_bo *bo, *old_bo = NULL;
 	struct qxl_crtc *qcrtc = to_qxl_crtc(crtc);
-	uint32_t width, height, base_offset;
 	bool recreate_primary = false;
 	int ret;
 	int surf_id;
@@ -538,12 +609,6 @@ static int qxl_crtc_mode_set(struct drm_crtc *crtc,
 	}
 	qfb = to_qxl_framebuffer(crtc->primary->fb);
 	bo = gem_to_qxl_bo(qfb->obj);
-	if (!m)
-		/* and do we care? */
-		DRM_DEBUG("%dx%d: not a native mode\n", x, y);
-	else
-		DRM_DEBUG("%dx%d: qxl id %d\n",
-			  mode->hdisplay, mode->vdisplay, m->id);
 	DRM_DEBUG("+%d+%d (%d,%d) => (%d,%d)\n",
 		  x, y,
 		  mode->hdisplay, mode->vdisplay,
@@ -553,9 +618,10 @@ static int qxl_crtc_mode_set(struct drm_crtc *crtc,
 	if (qcrtc->index == 0)
 		recreate_primary = true;
 
-	width = mode->hdisplay;
-	height = mode->vdisplay;
-	base_offset = 0;
+	if (bo->surf.stride * bo->surf.height > qdev->vram_size) {
+		DRM_ERROR("Mode doesn't fit in vram size (vgamem)");
+		return -EINVAL;
+        }
 
 	ret = qxl_bo_reserve(bo, false);
 	if (ret != 0)
@@ -569,10 +635,10 @@ static int qxl_crtc_mode_set(struct drm_crtc *crtc,
 	if (recreate_primary) {
 		qxl_io_destroy_primary(qdev);
 		qxl_io_log(qdev,
-			   "recreate primary: %dx%d (was %dx%d,%d,%d)\n",
-			   width, height, bo->surf.width,
-			   bo->surf.height, bo->surf.stride, bo->surf.format);
-		qxl_io_create_primary(qdev, base_offset, bo);
+			   "recreate primary: %dx%d,%d,%d\n",
+			   bo->surf.width, bo->surf.height,
+			   bo->surf.stride, bo->surf.format);
+		qxl_io_create_primary(qdev, 0, bo);
 		bo->is_primary = true;
 	}
 
@@ -902,6 +968,10 @@ static int qdev_output_init(struct drm_device *dev, int num_output)
 
 	drm_object_attach_property(&connector->base,
 				   qdev->hotplug_mode_update_property, 0);
+	drm_object_attach_property(&connector->base,
+				   dev->mode_config.suggested_x_property, 0);
+	drm_object_attach_property(&connector->base,
+				   dev->mode_config.suggested_y_property, 0);
 	drm_connector_register(connector);
 	return 0;
 }
@@ -1015,6 +1085,7 @@ int qxl_modeset_init(struct qxl_device *qdev)
 
 	qdev->ddev->mode_config.fb_base = qdev->vram_base;
 
+	drm_mode_create_suggested_offset_properties(qdev->ddev);
 	qxl_mode_create_hotplug_mode_update_property(qdev);
 
 	for (i = 0 ; i < qxl_num_crtc; ++i) {
diff --git a/drivers/gpu/drm/qxl/qxl_drv.c b/drivers/gpu/drm/qxl/qxl_drv.c
index a3fd92029a14..1d9b80c91a15 100644
--- a/drivers/gpu/drm/qxl/qxl_drv.c
+++ b/drivers/gpu/drm/qxl/qxl_drv.c
@@ -84,6 +84,7 @@ static const struct file_operations qxl_fops = {
 	.release = drm_release,
 	.unlocked_ioctl = drm_ioctl,
 	.poll = drm_poll,
+	.read = drm_read,
 	.mmap = qxl_mmap,
 };
 
@@ -195,6 +196,20 @@ static int qxl_pm_restore(struct device *dev)
 	return qxl_drm_resume(drm_dev, false);
 }
 
+static u32 qxl_noop_get_vblank_counter(struct drm_device *dev, int crtc)
+{
+	return dev->vblank[crtc].count.counter;
+}
+
+static int qxl_noop_enable_vblank(struct drm_device *dev, int crtc)
+{
+	return 0;
+}
+
+static void qxl_noop_disable_vblank(struct drm_device *dev, int crtc)
+{
+}
+
 static const struct dev_pm_ops qxl_pm_ops = {
 	.suspend = qxl_pm_suspend,
 	.resume = qxl_pm_resume,
@@ -212,10 +227,15 @@ static struct pci_driver qxl_pci_driver = {
 };
 
 static struct drm_driver qxl_driver = {
-	.driver_features = DRIVER_GEM | DRIVER_MODESET |
+	.driver_features = DRIVER_GEM | DRIVER_MODESET | DRIVER_PRIME |
 			   DRIVER_HAVE_IRQ | DRIVER_IRQ_SHARED,
 	.load = qxl_driver_load,
 	.unload = qxl_driver_unload,
+	.get_vblank_counter = qxl_noop_get_vblank_counter,
+	.enable_vblank = qxl_noop_enable_vblank,
+	.disable_vblank = qxl_noop_disable_vblank,
+
+	.set_busid = drm_pci_set_busid,
 
 	.dumb_create = qxl_mode_dumb_create,
 	.dumb_map_offset = qxl_mode_dumb_mmap,
@@ -224,6 +244,17 @@ static struct drm_driver qxl_driver = {
 	.debugfs_init = qxl_debugfs_init,
 	.debugfs_cleanup = qxl_debugfs_takedown,
 #endif
+	.prime_handle_to_fd = drm_gem_prime_handle_to_fd,
+	.prime_fd_to_handle = drm_gem_prime_fd_to_handle,
+	.gem_prime_export = drm_gem_prime_export,
+	.gem_prime_import = drm_gem_prime_import,
+	.gem_prime_pin = qxl_gem_prime_pin,
+	.gem_prime_unpin = qxl_gem_prime_unpin,
+	.gem_prime_get_sg_table = qxl_gem_prime_get_sg_table,
+	.gem_prime_import_sg_table = qxl_gem_prime_import_sg_table,
+	.gem_prime_vmap = qxl_gem_prime_vmap,
+	.gem_prime_vunmap = qxl_gem_prime_vunmap,
+	.gem_prime_mmap = qxl_gem_prime_mmap,
 	.gem_free_object = qxl_gem_object_free,
 	.gem_open_object = qxl_gem_object_open,
 	.gem_close_object = qxl_gem_object_close,
diff --git a/drivers/gpu/drm/qxl/qxl_drv.h b/drivers/gpu/drm/qxl/qxl_drv.h
index 36ed40ba773f..7c6cafe21f5f 100644
--- a/drivers/gpu/drm/qxl/qxl_drv.h
+++ b/drivers/gpu/drm/qxl/qxl_drv.h
@@ -31,6 +31,7 @@
  * Definitions taken from spice-protocol, plus kernel driver specific bits.
  */
 
+#include <linux/fence.h>
 #include <linux/workqueue.h>
 #include <linux/firmware.h>
 #include <linux/platform_device.h>
@@ -42,6 +43,8 @@
 #include <ttm/ttm_placement.h>
 #include <ttm/ttm_module.h>
 
+#include <drm/drm_gem.h>
+
 /* just for ttm_validate_buffer */
 #include <ttm/ttm_execbuf_util.h>
 
@@ -95,31 +98,24 @@ enum {
 	QXL_INTERRUPT_IO_CMD |\
 	QXL_INTERRUPT_CLIENT_MONITORS_CONFIG)
 
-struct qxl_fence {
-	struct qxl_device *qdev;
-	uint32_t num_active_releases;
-	uint32_t *release_ids;
-	struct radix_tree_root tree;
-};
-
 struct qxl_bo {
 	/* Protected by gem.mutex */
 	struct list_head		list;
 	/* Protected by tbo.reserved */
-	u32				placements[3];
+	struct ttm_place		placements[3];
 	struct ttm_placement		placement;
 	struct ttm_buffer_object	tbo;
 	struct ttm_bo_kmap_obj		kmap;
 	unsigned			pin_count;
 	void				*kptr;
 	int                             type;
+
 	/* Constant after initialization */
 	struct drm_gem_object		gem_base;
 	bool is_primary; /* is this now a primary surface */
 	bool hw_surf_alloc;
 	struct qxl_surface surf;
 	uint32_t surface_id;
-	struct qxl_fence fence; /* per bo fence  - list of releases */
 	struct qxl_release *surf_create;
 };
 #define gem_to_qxl_bo(gobj) container_of((gobj), struct qxl_bo, gem_base)
@@ -191,6 +187,8 @@ enum {
  * spice-protocol/qxl_dev.h */
 #define QXL_MAX_RES 96
 struct qxl_release {
+	struct fence base;
+
 	int id;
 	int type;
 	uint32_t release_offset;
@@ -284,7 +282,9 @@ struct qxl_device {
 	uint8_t		slot_gen_bits;
 	uint64_t	va_slot_mask;
 
+	spinlock_t	release_lock;
 	struct idr	release_idr;
+	uint32_t	release_seqno;
 	spinlock_t release_idr_lock;
 	struct mutex	async_io_mutex;
 	unsigned int last_sent_io_cmd;
@@ -532,6 +532,18 @@ int qxl_garbage_collect(struct qxl_device *qdev);
 int qxl_debugfs_init(struct drm_minor *minor);
 void qxl_debugfs_takedown(struct drm_minor *minor);
 
+/* qxl_prime.c */
+int qxl_gem_prime_pin(struct drm_gem_object *obj);
+void qxl_gem_prime_unpin(struct drm_gem_object *obj);
+struct sg_table *qxl_gem_prime_get_sg_table(struct drm_gem_object *obj);
+struct drm_gem_object *qxl_gem_prime_import_sg_table(
+	struct drm_device *dev, struct dma_buf_attachment *attach,
+	struct sg_table *sgt);
+void *qxl_gem_prime_vmap(struct drm_gem_object *obj);
+void qxl_gem_prime_vunmap(struct drm_gem_object *obj, void *vaddr);
+int qxl_gem_prime_mmap(struct drm_gem_object *obj,
+				struct vm_area_struct *vma);
+
 /* qxl_irq.c */
 int qxl_irq_init(struct qxl_device *qdev);
 irqreturn_t qxl_irq_handler(int irq, void *arg);
@@ -561,10 +573,4 @@ qxl_surface_lookup(struct drm_device *dev, int surface_id);
 void qxl_surface_evict(struct qxl_device *qdev, struct qxl_bo *surf, bool freeing);
 int qxl_update_surface(struct qxl_device *qdev, struct qxl_bo *surf);
 
-/* qxl_fence.c */
-void qxl_fence_add_release_locked(struct qxl_fence *qfence, uint32_t rel_id);
-int qxl_fence_remove_release(struct qxl_fence *qfence, uint32_t rel_id);
-int qxl_fence_init(struct qxl_device *qdev, struct qxl_fence *qfence);
-void qxl_fence_fini(struct qxl_fence *qfence);
-
 #endif
diff --git a/drivers/gpu/drm/qxl/qxl_fb.c b/drivers/gpu/drm/qxl/qxl_fb.c
index df567888bb1e..3d7c1d00a424 100644
--- a/drivers/gpu/drm/qxl/qxl_fb.c
+++ b/drivers/gpu/drm/qxl/qxl_fb.c
@@ -625,7 +625,8 @@ static int qxl_fb_find_or_create_single(
 		struct drm_fb_helper *helper,
 		struct drm_fb_helper_surface_size *sizes)
 {
-	struct qxl_fbdev *qfbdev = (struct qxl_fbdev *)helper;
+	struct qxl_fbdev *qfbdev =
+		container_of(helper, struct qxl_fbdev, helper);
 	int new_fb = 0;
 	int ret;
 
diff --git a/drivers/gpu/drm/qxl/qxl_fence.c b/drivers/gpu/drm/qxl/qxl_fence.c
deleted file mode 100644
index ae59e91cfb9a..000000000000
--- a/drivers/gpu/drm/qxl/qxl_fence.c
+++ /dev/null
@@ -1,91 +0,0 @@
-/*
- * Copyright 2013 Red Hat Inc.
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * the rights to use, copy, modify, merge, publish, distribute, sublicense,
- * and/or sell copies of the Software, and to permit persons to whom the
- * Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in
- * all copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
- * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
- * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
- * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
- * OTHER DEALINGS IN THE SOFTWARE.
- *
- * Authors: Dave Airlie
- *          Alon Levy
- */
-
-
-#include "qxl_drv.h"
-
-/* QXL fencing-
-
-   When we submit operations to the GPU we pass a release reference to the GPU
-   with them, the release reference is then added to the release ring when
-   the GPU is finished with that particular operation and has removed it from
-   its tree.
-
-   So we have can have multiple outstanding non linear fences per object.
-
-   From a TTM POV we only care if the object has any outstanding releases on
-   it.
-
-   we wait until all outstanding releases are processeed.
-
-   sync object is just a list of release ids that represent that fence on
-   that buffer.
-
-   we just add new releases onto the sync object attached to the object.
-
-   This currently uses a radix tree to store the list of release ids.
-
-   For some reason every so often qxl hw fails to release, things go wrong.
-*/
-/* must be called with the fence lock held */
-void qxl_fence_add_release_locked(struct qxl_fence *qfence, uint32_t rel_id)
-{
-	radix_tree_insert(&qfence->tree, rel_id, qfence);
-	qfence->num_active_releases++;
-}
-
-int qxl_fence_remove_release(struct qxl_fence *qfence, uint32_t rel_id)
-{
-	void *ret;
-	int retval = 0;
-	struct qxl_bo *bo = container_of(qfence, struct qxl_bo, fence);
-
-	spin_lock(&bo->tbo.bdev->fence_lock);
-
-	ret = radix_tree_delete(&qfence->tree, rel_id);
-	if (ret == qfence)
-		qfence->num_active_releases--;
-	else {
-		DRM_DEBUG("didn't find fence in radix tree for %d\n", rel_id);
-		retval = -ENOENT;
-	}
-	spin_unlock(&bo->tbo.bdev->fence_lock);
-	return retval;
-}
-
-
-int qxl_fence_init(struct qxl_device *qdev, struct qxl_fence *qfence)
-{
-	qfence->qdev = qdev;
-	qfence->num_active_releases = 0;
-	INIT_RADIX_TREE(&qfence->tree, GFP_ATOMIC);
-	return 0;
-}
-
-void qxl_fence_fini(struct qxl_fence *qfence)
-{
-	kfree(qfence->release_ids);
-	qfence->num_active_releases = 0;
-}
diff --git a/drivers/gpu/drm/qxl/qxl_kms.c b/drivers/gpu/drm/qxl/qxl_kms.c
index fd88eb4a3f79..b2977a181935 100644
--- a/drivers/gpu/drm/qxl/qxl_kms.c
+++ b/drivers/gpu/drm/qxl/qxl_kms.c
@@ -223,6 +223,7 @@ static int qxl_device_init(struct qxl_device *qdev,
 
 	idr_init(&qdev->release_idr);
 	spin_lock_init(&qdev->release_idr_lock);
+	spin_lock_init(&qdev->release_lock);
 
 	idr_init(&qdev->surf_id_idr);
 	spin_lock_init(&qdev->surf_id_idr_lock);
@@ -297,6 +298,9 @@ int qxl_driver_unload(struct drm_device *dev)
 
 	if (qdev == NULL)
 		return 0;
+
+	drm_vblank_cleanup(dev);
+
 	qxl_modeset_fini(qdev);
 	qxl_device_fini(qdev);
 
@@ -324,15 +328,20 @@ int qxl_driver_load(struct drm_device *dev, unsigned long flags)
 	if (r)
 		goto out;
 
+	r = drm_vblank_init(dev, 1);
+	if (r)
+		goto unload;
+
 	r = qxl_modeset_init(qdev);
-	if (r) {
-		qxl_driver_unload(dev);
-		goto out;
-	}
+	if (r)
+		goto unload;
 
 	drm_kms_helper_poll_init(qdev->ddev);
 
 	return 0;
+unload:
+	qxl_driver_unload(dev);
+
 out:
 	kfree(qdev);
 	return r;
diff --git a/drivers/gpu/drm/qxl/qxl_object.c b/drivers/gpu/drm/qxl/qxl_object.c
index b95f144f0b49..cdeaf08fdc74 100644
--- a/drivers/gpu/drm/qxl/qxl_object.c
+++ b/drivers/gpu/drm/qxl/qxl_object.c
@@ -36,7 +36,6 @@ static void qxl_ttm_bo_destroy(struct ttm_buffer_object *tbo)
 	qdev = (struct qxl_device *)bo->gem_base.dev->dev_private;
 
 	qxl_surface_evict(qdev, bo, false);
-	qxl_fence_fini(&bo->fence);
 	mutex_lock(&qdev->gem.mutex);
 	list_del_init(&bo->list);
 	mutex_unlock(&qdev->gem.mutex);
@@ -55,21 +54,24 @@ void qxl_ttm_placement_from_domain(struct qxl_bo *qbo, u32 domain, bool pinned)
 {
 	u32 c = 0;
 	u32 pflag = pinned ? TTM_PL_FLAG_NO_EVICT : 0;
+	unsigned i;
 
-	qbo->placement.fpfn = 0;
-	qbo->placement.lpfn = 0;
 	qbo->placement.placement = qbo->placements;
 	qbo->placement.busy_placement = qbo->placements;
 	if (domain == QXL_GEM_DOMAIN_VRAM)
-		qbo->placements[c++] = TTM_PL_FLAG_CACHED | TTM_PL_FLAG_VRAM | pflag;
+		qbo->placements[c++].flags = TTM_PL_FLAG_CACHED | TTM_PL_FLAG_VRAM | pflag;
 	if (domain == QXL_GEM_DOMAIN_SURFACE)
-		qbo->placements[c++] = TTM_PL_FLAG_CACHED | TTM_PL_FLAG_PRIV0 | pflag;
+		qbo->placements[c++].flags = TTM_PL_FLAG_CACHED | TTM_PL_FLAG_PRIV0 | pflag;
 	if (domain == QXL_GEM_DOMAIN_CPU)
-		qbo->placements[c++] = TTM_PL_MASK_CACHING | TTM_PL_FLAG_SYSTEM | pflag;
+		qbo->placements[c++].flags = TTM_PL_MASK_CACHING | TTM_PL_FLAG_SYSTEM | pflag;
 	if (!c)
-		qbo->placements[c++] = TTM_PL_MASK_CACHING | TTM_PL_FLAG_SYSTEM;
+		qbo->placements[c++].flags = TTM_PL_MASK_CACHING | TTM_PL_FLAG_SYSTEM;
 	qbo->placement.num_placement = c;
 	qbo->placement.num_busy_placement = c;
+	for (i = 0; i < c; ++i) {
+		qbo->placements[i].fpfn = 0;
+		qbo->placements[i].lpfn = 0;
+	}
 }
 
 
@@ -99,7 +101,6 @@ int qxl_bo_create(struct qxl_device *qdev,
 	bo->type = domain;
 	bo->pin_count = pinned ? 1 : 0;
 	bo->surface_id = 0;
-	qxl_fence_init(qdev, &bo->fence);
 	INIT_LIST_HEAD(&bo->list);
 
 	if (surf)
@@ -109,7 +110,7 @@ int qxl_bo_create(struct qxl_device *qdev,
 
 	r = ttm_bo_init(&qdev->mman.bdev, &bo->tbo, size, type,
 			&bo->placement, 0, !kernel, NULL, size,
-			NULL, &qxl_ttm_bo_destroy);
+			NULL, NULL, &qxl_ttm_bo_destroy);
 	if (unlikely(r != 0)) {
 		if (r != -ERESTARTSYS)
 			dev_err(qdev->dev,
@@ -259,7 +260,7 @@ int qxl_bo_unpin(struct qxl_bo *bo)
 	if (bo->pin_count)
 		return 0;
 	for (i = 0; i < bo->placement.num_placement; i++)
-		bo->placements[i] &= ~TTM_PL_FLAG_NO_EVICT;
+		bo->placements[i].flags &= ~TTM_PL_FLAG_NO_EVICT;
 	r = ttm_bo_validate(&bo->tbo, &bo->placement, false, false);
 	if (unlikely(r != 0))
 		dev_err(qdev->dev, "%p validate failed for unpin\n", bo);
diff --git a/drivers/gpu/drm/qxl/qxl_object.h b/drivers/gpu/drm/qxl/qxl_object.h
index 83a423293afd..37af1bc0dd00 100644
--- a/drivers/gpu/drm/qxl/qxl_object.h
+++ b/drivers/gpu/drm/qxl/qxl_object.h
@@ -76,12 +76,10 @@ static inline int qxl_bo_wait(struct qxl_bo *bo, u32 *mem_type,
 		}
 		return r;
 	}
-	spin_lock(&bo->tbo.bdev->fence_lock);
 	if (mem_type)
 		*mem_type = bo->tbo.mem.mem_type;
-	if (bo->tbo.sync_obj)
-		r = ttm_bo_wait(&bo->tbo, true, true, no_wait);
-	spin_unlock(&bo->tbo.bdev->fence_lock);
+
+	r = ttm_bo_wait(&bo->tbo, true, true, no_wait);
 	ttm_bo_unreserve(&bo->tbo);
 	return r;
 }
diff --git a/drivers/gpu/drm/qxl/qxl_prime.c b/drivers/gpu/drm/qxl/qxl_prime.c
new file mode 100644
index 000000000000..3d031b50a8fd
--- /dev/null
+++ b/drivers/gpu/drm/qxl/qxl_prime.c
@@ -0,0 +1,72 @@
+/*
+ * Copyright 2014 Canonical
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Andreas Pokorny
+ */
+
+#include "qxl_drv.h"
+
+/* Empty Implementations as there should not be any other driver for a virtual
+ * device that might share buffers with qxl */
+
+int qxl_gem_prime_pin(struct drm_gem_object *obj)
+{
+	WARN_ONCE(1, "not implemented");
+	return -ENOSYS;
+}
+
+void qxl_gem_prime_unpin(struct drm_gem_object *obj)
+{
+	WARN_ONCE(1, "not implemented");
+}
+
+
+struct sg_table *qxl_gem_prime_get_sg_table(struct drm_gem_object *obj)
+{
+	WARN_ONCE(1, "not implemented");
+	return ERR_PTR(-ENOSYS);
+}
+
+struct drm_gem_object *qxl_gem_prime_import_sg_table(
+	struct drm_device *dev, struct dma_buf_attachment *attach,
+	struct sg_table *table)
+{
+	WARN_ONCE(1, "not implemented");
+	return ERR_PTR(-ENOSYS);
+}
+
+void *qxl_gem_prime_vmap(struct drm_gem_object *obj)
+{
+	WARN_ONCE(1, "not implemented");
+	return ERR_PTR(-ENOSYS);
+}
+
+void qxl_gem_prime_vunmap(struct drm_gem_object *obj, void *vaddr)
+{
+	WARN_ONCE(1, "not implemented");
+}
+
+int qxl_gem_prime_mmap(struct drm_gem_object *obj,
+		       struct vm_area_struct *area)
+{
+	WARN_ONCE(1, "not implemented");
+	return ENOSYS;
+}
diff --git a/drivers/gpu/drm/qxl/qxl_release.c b/drivers/gpu/drm/qxl/qxl_release.c
index 14e776f1d14e..d9b25684ac98 100644
--- a/drivers/gpu/drm/qxl/qxl_release.c
+++ b/drivers/gpu/drm/qxl/qxl_release.c
@@ -21,6 +21,7 @@
  */
 #include "qxl_drv.h"
 #include "qxl_object.h"
+#include <trace/events/fence.h>
 
 /*
  * drawable cmd cache - allocate a bunch of VRAM pages, suballocate
@@ -39,6 +40,88 @@
 static const int release_size_per_bo[] = { RELEASE_SIZE, SURFACE_RELEASE_SIZE, RELEASE_SIZE };
 static const int releases_per_bo[] = { RELEASES_PER_BO, SURFACE_RELEASES_PER_BO, RELEASES_PER_BO };
 
+static const char *qxl_get_driver_name(struct fence *fence)
+{
+	return "qxl";
+}
+
+static const char *qxl_get_timeline_name(struct fence *fence)
+{
+	return "release";
+}
+
+static bool qxl_nop_signaling(struct fence *fence)
+{
+	/* fences are always automatically signaled, so just pretend we did this.. */
+	return true;
+}
+
+static long qxl_fence_wait(struct fence *fence, bool intr, signed long timeout)
+{
+	struct qxl_device *qdev;
+	struct qxl_release *release;
+	int count = 0, sc = 0;
+	bool have_drawable_releases;
+	unsigned long cur, end = jiffies + timeout;
+
+	qdev = container_of(fence->lock, struct qxl_device, release_lock);
+	release = container_of(fence, struct qxl_release, base);
+	have_drawable_releases = release->type == QXL_RELEASE_DRAWABLE;
+
+retry:
+	sc++;
+
+	if (fence_is_signaled(fence))
+		goto signaled;
+
+	qxl_io_notify_oom(qdev);
+
+	for (count = 0; count < 11; count++) {
+		if (!qxl_queue_garbage_collect(qdev, true))
+			break;
+
+		if (fence_is_signaled(fence))
+			goto signaled;
+	}
+
+	if (fence_is_signaled(fence))
+		goto signaled;
+
+	if (have_drawable_releases || sc < 4) {
+		if (sc > 2)
+			/* back off */
+			usleep_range(500, 1000);
+
+		if (time_after(jiffies, end))
+			return 0;
+
+		if (have_drawable_releases && sc > 300) {
+			FENCE_WARN(fence, "failed to wait on release %d "
+					  "after spincount %d\n",
+					  fence->context & ~0xf0000000, sc);
+			goto signaled;
+		}
+		goto retry;
+	}
+	/*
+	 * yeah, original sync_obj_wait gave up after 3 spins when
+	 * have_drawable_releases is not set.
+	 */
+
+signaled:
+	cur = jiffies;
+	if (time_after(cur, end))
+		return 0;
+	return end - cur;
+}
+
+static const struct fence_ops qxl_fence_ops = {
+	.get_driver_name = qxl_get_driver_name,
+	.get_timeline_name = qxl_get_timeline_name,
+	.enable_signaling = qxl_nop_signaling,
+	.wait = qxl_fence_wait,
+};
+
 static uint64_t
 qxl_release_alloc(struct qxl_device *qdev, int type,
 		  struct qxl_release **ret)
@@ -46,13 +129,13 @@ qxl_release_alloc(struct qxl_device *qdev, int type,
 	struct qxl_release *release;
 	int handle;
 	size_t size = sizeof(*release);
-	int idr_ret;
 
 	release = kmalloc(size, GFP_KERNEL);
 	if (!release) {
 		DRM_ERROR("Out of memory\n");
 		return 0;
 	}
+	release->base.ops = NULL;
 	release->type = type;
 	release->release_offset = 0;
 	release->surface_release_id = 0;
@@ -60,44 +143,61 @@ qxl_release_alloc(struct qxl_device *qdev, int type,
 
 	idr_preload(GFP_KERNEL);
 	spin_lock(&qdev->release_idr_lock);
-	idr_ret = idr_alloc(&qdev->release_idr, release, 1, 0, GFP_NOWAIT);
+	handle = idr_alloc(&qdev->release_idr, release, 1, 0, GFP_NOWAIT);
+	release->base.seqno = ++qdev->release_seqno;
 	spin_unlock(&qdev->release_idr_lock);
 	idr_preload_end();
-	handle = idr_ret;
-	if (idr_ret < 0)
-		goto release_fail;
+	if (handle < 0) {
+		kfree(release);
+		*ret = NULL;
+		return handle;
+	}
 	*ret = release;
 	QXL_INFO(qdev, "allocated release %lld\n", handle);
 	release->id = handle;
-release_fail:
-
 	return handle;
 }
 
+static void
+qxl_release_free_list(struct qxl_release *release)
+{
+	while (!list_empty(&release->bos)) {
+		struct qxl_bo_list *entry;
+		struct qxl_bo *bo;
+
+		entry = container_of(release->bos.next,
+				     struct qxl_bo_list, tv.head);
+		bo = to_qxl_bo(entry->tv.bo);
+		qxl_bo_unref(&bo);
+		list_del(&entry->tv.head);
+		kfree(entry);
+	}
+}
+
 void
 qxl_release_free(struct qxl_device *qdev,
 		 struct qxl_release *release)
 {
-	struct qxl_bo_list *entry, *tmp;
 	QXL_INFO(qdev, "release %d, type %d\n", release->id,
 		 release->type);
 
 	if (release->surface_release_id)
 		qxl_surface_id_dealloc(qdev, release->surface_release_id);
 
-	list_for_each_entry_safe(entry, tmp, &release->bos, tv.head) {
-		struct qxl_bo *bo = to_qxl_bo(entry->tv.bo);
-		QXL_INFO(qdev, "release %llx\n",
-			drm_vma_node_offset_addr(&entry->tv.bo->vma_node)
-						- DRM_FILE_OFFSET);
-		qxl_fence_remove_release(&bo->fence, release->id);
-		qxl_bo_unref(&bo);
-		kfree(entry);
-	}
 	spin_lock(&qdev->release_idr_lock);
 	idr_remove(&qdev->release_idr, release->id);
 	spin_unlock(&qdev->release_idr_lock);
-	kfree(release);
+
+	if (release->base.ops) {
+		WARN_ON(list_empty(&release->bos));
+		qxl_release_free_list(release);
+
+		fence_signal(&release->base);
+		fence_put(&release->base);
+	} else {
+		qxl_release_free_list(release);
+		kfree(release);
+	}
 }
 
 static int qxl_release_bo_alloc(struct qxl_device *qdev,
@@ -126,6 +226,7 @@ int qxl_release_list_add(struct qxl_release *release, struct qxl_bo *bo)
 
 	qxl_bo_ref(bo);
 	entry->tv.bo = &bo->tbo;
+	entry->tv.shared = false;
 	list_add_tail(&entry->tv.head, &release->bos);
 	return 0;
 }
@@ -142,6 +243,10 @@ static int qxl_release_validate_bo(struct qxl_bo *bo)
 			return ret;
 	}
 
+	ret = reservation_object_reserve_shared(bo->tbo.resv);
+	if (ret)
+		return ret;
+
 	/* allocate a surface for reserved + validated buffers */
 	ret = qxl_bo_check_id(bo->gem_base.dev->dev_private, bo);
 	if (ret)
@@ -159,7 +264,8 @@ int qxl_release_reserve_list(struct qxl_release *release, bool no_intr)
 	if (list_is_singular(&release->bos))
 		return 0;
 
-	ret = ttm_eu_reserve_buffers(&release->ticket, &release->bos);
+	ret = ttm_eu_reserve_buffers(&release->ticket, &release->bos,
+				     !no_intr, NULL);
 	if (ret)
 		return ret;
 
@@ -199,6 +305,8 @@ int qxl_alloc_surface_release_reserved(struct qxl_device *qdev,
 
 		/* stash the release after the create command */
 		idr_ret = qxl_release_alloc(qdev, QXL_RELEASE_SURFACE_CMD, release);
+		if (idr_ret < 0)
+			return idr_ret;
 		bo = qxl_bo_ref(to_qxl_bo(entry->tv.bo));
 
 		(*release)->release_offset = create_rel->release_offset + 64;
@@ -239,6 +347,11 @@ int qxl_alloc_release_reserved(struct qxl_device *qdev, unsigned long size,
 	}
 
 	idr_ret = qxl_release_alloc(qdev, type, release);
+	if (idr_ret < 0) {
+		if (rbo)
+			*rbo = NULL;
+		return idr_ret;
+	}
 
 	mutex_lock(&qdev->release_mutex);
 	if (qdev->current_release_bo_offset[cur_idx] + 1 >= releases_per_bo[cur_idx]) {
@@ -319,40 +432,44 @@ void qxl_release_unmap(struct qxl_device *qdev,
 
 void qxl_release_fence_buffer_objects(struct qxl_release *release)
 {
-	struct ttm_validate_buffer *entry;
 	struct ttm_buffer_object *bo;
 	struct ttm_bo_global *glob;
 	struct ttm_bo_device *bdev;
 	struct ttm_bo_driver *driver;
 	struct qxl_bo *qbo;
+	struct ttm_validate_buffer *entry;
+	struct qxl_device *qdev;
 
 	/* if only one object on the release its the release itself
 	   since these objects are pinned no need to reserve */
-	if (list_is_singular(&release->bos))
+	if (list_is_singular(&release->bos) || list_empty(&release->bos))
 		return;
 
 	bo = list_first_entry(&release->bos, struct ttm_validate_buffer, head)->bo;
 	bdev = bo->bdev;
+	qdev = container_of(bdev, struct qxl_device, mman.bdev);
+
+	/*
+	 * Since we never really allocated a context and we don't want to conflict,
+	 * set the highest bits. This will break if we really allow exporting of dma-bufs.
+	 */
+	fence_init(&release->base, &qxl_fence_ops, &qdev->release_lock,
+		   release->id | 0xf0000000, release->base.seqno);
+	trace_fence_emit(&release->base);
+
 	driver = bdev->driver;
 	glob = bo->glob;
 
 	spin_lock(&glob->lru_lock);
-	spin_lock(&bdev->fence_lock);
 
 	list_for_each_entry(entry, &release->bos, head) {
 		bo = entry->bo;
 		qbo = to_qxl_bo(bo);
 
-		if (!entry->bo->sync_obj)
-			entry->bo->sync_obj = &qbo->fence;
-
-		qxl_fence_add_release_locked(&qbo->fence, release->id);
-
+		reservation_object_add_shared_fence(bo->resv, &release->base);
 		ttm_bo_add_to_lru(bo);
 		__ttm_bo_unreserve(bo);
-		entry->reserved = false;
 	}
-	spin_unlock(&bdev->fence_lock);
 	spin_unlock(&glob->lru_lock);
 	ww_acquire_fini(&release->ticket);
 }
diff --git a/drivers/gpu/drm/qxl/qxl_ttm.c b/drivers/gpu/drm/qxl/qxl_ttm.c
index 71a1baeac14e..0cbc4c987164 100644
--- a/drivers/gpu/drm/qxl/qxl_ttm.c
+++ b/drivers/gpu/drm/qxl/qxl_ttm.c
@@ -127,7 +127,7 @@ int qxl_mmap(struct file *filp, struct vm_area_struct *vma)
 	if (unlikely(vma->vm_pgoff < DRM_FILE_PAGE_OFFSET)) {
 		pr_info("%s: vma->vm_pgoff (%ld) < DRM_FILE_PAGE_OFFSET\n",
 			__func__, vma->vm_pgoff);
-		return drm_mmap(filp, vma);
+		return -EINVAL;
 	}
 
 	file_priv = filp->private_data;
@@ -188,11 +188,13 @@ static void qxl_evict_flags(struct ttm_buffer_object *bo,
 				struct ttm_placement *placement)
 {
 	struct qxl_bo *qbo;
-	static u32 placements = TTM_PL_MASK_CACHING | TTM_PL_FLAG_SYSTEM;
+	static struct ttm_place placements = {
+		.fpfn = 0,
+		.lpfn = 0,
+		.flags = TTM_PL_MASK_CACHING | TTM_PL_FLAG_SYSTEM
+	};
 
 	if (!qxl_ttm_bo_is_qxl_bo(bo)) {
-		placement->fpfn = 0;
-		placement->lpfn = 0;
 		placement->placement = &placements;
 		placement->busy_placement = &placements;
 		placement->num_placement = 1;
@@ -355,92 +357,6 @@ static int qxl_bo_move(struct ttm_buffer_object *bo,
 	return ttm_bo_move_memcpy(bo, evict, no_wait_gpu, new_mem);
 }
 
-
-static int qxl_sync_obj_wait(void *sync_obj,
-			     bool lazy, bool interruptible)
-{
-	struct qxl_fence *qfence = (struct qxl_fence *)sync_obj;
-	int count = 0, sc = 0;
-	struct qxl_bo *bo = container_of(qfence, struct qxl_bo, fence);
-
-	if (qfence->num_active_releases == 0)
-		return 0;
-
-retry:
-	if (sc == 0) {
-		if (bo->type == QXL_GEM_DOMAIN_SURFACE)
-			qxl_update_surface(qfence->qdev, bo);
-	} else if (sc >= 1) {
-		qxl_io_notify_oom(qfence->qdev);
-	}
-
-	sc++;
-
-	for (count = 0; count < 10; count++) {
-		bool ret;
-		ret = qxl_queue_garbage_collect(qfence->qdev, true);
-		if (ret == false)
-			break;
-
-		if (qfence->num_active_releases == 0)
-			return 0;
-	}
-
-	if (qfence->num_active_releases) {
-		bool have_drawable_releases = false;
-		void **slot;
-		struct radix_tree_iter iter;
-		int release_id;
-
-		radix_tree_for_each_slot(slot, &qfence->tree, &iter, 0) {
-			struct qxl_release *release;
-
-			release_id = iter.index;
-			release = qxl_release_from_id_locked(qfence->qdev, release_id);
-			if (release == NULL)
-				continue;
-
-			if (release->type == QXL_RELEASE_DRAWABLE)
-				have_drawable_releases = true;
-		}
-
-		qxl_queue_garbage_collect(qfence->qdev, true);
-
-		if (have_drawable_releases || sc < 4) {
-			if (sc > 2)
-				/* back off */
-				usleep_range(500, 1000);
-			if (have_drawable_releases && sc > 300) {
-				WARN(1, "sync obj %d still has outstanding releases %d %d %d %ld %d\n", sc, bo->surface_id, bo->is_primary, bo->pin_count, (unsigned long)bo->gem_base.size, qfence->num_active_releases);
-				return -EBUSY;
-			}
-			goto retry;
-		}
-	}
-	return 0;
-}
-
-static int qxl_sync_obj_flush(void *sync_obj)
-{
-	return 0;
-}
-
-static void qxl_sync_obj_unref(void **sync_obj)
-{
-	*sync_obj = NULL;
-}
-
-static void *qxl_sync_obj_ref(void *sync_obj)
-{
-	return sync_obj;
-}
-
-static bool qxl_sync_obj_signaled(void *sync_obj)
-{
-	struct qxl_fence *qfence = (struct qxl_fence *)sync_obj;
-	return (qfence->num_active_releases == 0);
-}
-
 static void qxl_bo_move_notify(struct ttm_buffer_object *bo,
 			       struct ttm_mem_reg *new_mem)
 {
@@ -467,16 +383,9 @@ static struct ttm_bo_driver qxl_bo_driver = {
 	.verify_access = &qxl_verify_access,
 	.io_mem_reserve = &qxl_ttm_io_mem_reserve,
 	.io_mem_free = &qxl_ttm_io_mem_free,
-	.sync_obj_signaled = &qxl_sync_obj_signaled,
-	.sync_obj_wait = &qxl_sync_obj_wait,
-	.sync_obj_flush = &qxl_sync_obj_flush,
-	.sync_obj_unref = &qxl_sync_obj_unref,
-	.sync_obj_ref = &qxl_sync_obj_ref,
 	.move_notify = &qxl_bo_move_notify,
 };
 
-
-
 int qxl_ttm_init(struct qxl_device *qdev)
 {
 	int r;
diff --git a/drivers/gpu/drm/r128/r128_cce.c b/drivers/gpu/drm/r128/r128_cce.c
index 59459fe4e8c5..2c45ac9c1dc3 100644
--- a/drivers/gpu/drm/r128/r128_cce.c
+++ b/drivers/gpu/drm/r128/r128_cce.c
@@ -452,7 +452,7 @@ static int r128_do_init_cce(struct drm_device *dev, drm_r128_init_t *init)
 	dev_priv->span_pitch_offset_c = (((dev_priv->depth_pitch / 8) << 21) |
 					 (dev_priv->span_offset >> 5));
 
-	dev_priv->sarea = drm_getsarea(dev);
+	dev_priv->sarea = drm_legacy_getsarea(dev);
 	if (!dev_priv->sarea) {
 		DRM_ERROR("could not find sarea!\n");
 		dev->dev_private = (void *)dev_priv;
@@ -460,21 +460,21 @@ static int r128_do_init_cce(struct drm_device *dev, drm_r128_init_t *init)
 		return -EINVAL;
 	}
 
-	dev_priv->mmio = drm_core_findmap(dev, init->mmio_offset);
+	dev_priv->mmio = drm_legacy_findmap(dev, init->mmio_offset);
 	if (!dev_priv->mmio) {
 		DRM_ERROR("could not find mmio region!\n");
 		dev->dev_private = (void *)dev_priv;
 		r128_do_cleanup_cce(dev);
 		return -EINVAL;
 	}
-	dev_priv->cce_ring = drm_core_findmap(dev, init->ring_offset);
+	dev_priv->cce_ring = drm_legacy_findmap(dev, init->ring_offset);
 	if (!dev_priv->cce_ring) {
 		DRM_ERROR("could not find cce ring region!\n");
 		dev->dev_private = (void *)dev_priv;
 		r128_do_cleanup_cce(dev);
 		return -EINVAL;
 	}
-	dev_priv->ring_rptr = drm_core_findmap(dev, init->ring_rptr_offset);
+	dev_priv->ring_rptr = drm_legacy_findmap(dev, init->ring_rptr_offset);
 	if (!dev_priv->ring_rptr) {
 		DRM_ERROR("could not find ring read pointer!\n");
 		dev->dev_private = (void *)dev_priv;
@@ -482,7 +482,7 @@ static int r128_do_init_cce(struct drm_device *dev, drm_r128_init_t *init)
 		return -EINVAL;
 	}
 	dev->agp_buffer_token = init->buffers_offset;
-	dev->agp_buffer_map = drm_core_findmap(dev, init->buffers_offset);
+	dev->agp_buffer_map = drm_legacy_findmap(dev, init->buffers_offset);
 	if (!dev->agp_buffer_map) {
 		DRM_ERROR("could not find dma buffer region!\n");
 		dev->dev_private = (void *)dev_priv;
@@ -492,7 +492,7 @@ static int r128_do_init_cce(struct drm_device *dev, drm_r128_init_t *init)
 
 	if (!dev_priv->is_pci) {
 		dev_priv->agp_textures =
-		    drm_core_findmap(dev, init->agp_textures_offset);
+		    drm_legacy_findmap(dev, init->agp_textures_offset);
 		if (!dev_priv->agp_textures) {
 			DRM_ERROR("could not find agp texture region!\n");
 			dev->dev_private = (void *)dev_priv;
@@ -507,9 +507,9 @@ static int r128_do_init_cce(struct drm_device *dev, drm_r128_init_t *init)
 
 #if __OS_HAS_AGP
 	if (!dev_priv->is_pci) {
-		drm_core_ioremap_wc(dev_priv->cce_ring, dev);
-		drm_core_ioremap_wc(dev_priv->ring_rptr, dev);
-		drm_core_ioremap_wc(dev->agp_buffer_map, dev);
+		drm_legacy_ioremap_wc(dev_priv->cce_ring, dev);
+		drm_legacy_ioremap_wc(dev_priv->ring_rptr, dev);
+		drm_legacy_ioremap_wc(dev->agp_buffer_map, dev);
 		if (!dev_priv->cce_ring->handle ||
 		    !dev_priv->ring_rptr->handle ||
 		    !dev->agp_buffer_map->handle) {
@@ -603,11 +603,11 @@ int r128_do_cleanup_cce(struct drm_device *dev)
 #if __OS_HAS_AGP
 		if (!dev_priv->is_pci) {
 			if (dev_priv->cce_ring != NULL)
-				drm_core_ioremapfree(dev_priv->cce_ring, dev);
+				drm_legacy_ioremapfree(dev_priv->cce_ring, dev);
 			if (dev_priv->ring_rptr != NULL)
-				drm_core_ioremapfree(dev_priv->ring_rptr, dev);
+				drm_legacy_ioremapfree(dev_priv->ring_rptr, dev);
 			if (dev->agp_buffer_map != NULL) {
-				drm_core_ioremapfree(dev->agp_buffer_map, dev);
+				drm_legacy_ioremapfree(dev->agp_buffer_map, dev);
 				dev->agp_buffer_map = NULL;
 			}
 		} else
diff --git a/drivers/gpu/drm/r128/r128_drv.c b/drivers/gpu/drm/r128/r128_drv.c
index 5bd307cd8da1..c57b4de63caf 100644
--- a/drivers/gpu/drm/r128/r128_drv.c
+++ b/drivers/gpu/drm/r128/r128_drv.c
@@ -46,7 +46,7 @@ static const struct file_operations r128_driver_fops = {
 	.open = drm_open,
 	.release = drm_release,
 	.unlocked_ioctl = drm_ioctl,
-	.mmap = drm_mmap,
+	.mmap = drm_legacy_mmap,
 	.poll = drm_poll,
 #ifdef CONFIG_COMPAT
 	.compat_ioctl = r128_compat_ioctl,
@@ -62,6 +62,7 @@ static struct drm_driver driver = {
 	.load = r128_driver_load,
 	.preclose = r128_driver_preclose,
 	.lastclose = r128_driver_lastclose,
+	.set_busid = drm_pci_set_busid,
 	.get_vblank_counter = r128_get_vblank_counter,
 	.enable_vblank = r128_enable_vblank,
 	.disable_vblank = r128_disable_vblank,
diff --git a/drivers/gpu/drm/r128/r128_drv.h b/drivers/gpu/drm/r128/r128_drv.h
index 5bf3f5ff805d..723e5d6f10a4 100644
--- a/drivers/gpu/drm/r128/r128_drv.h
+++ b/drivers/gpu/drm/r128/r128_drv.h
@@ -35,6 +35,9 @@
 #ifndef __R128_DRV_H__
 #define __R128_DRV_H__
 
+#include <drm/ati_pcigart.h>
+#include <drm/drm_legacy.h>
+
 /* General customization:
  */
 #define DRIVER_AUTHOR		"Gareth Hughes, VA Linux Systems Inc."
diff --git a/drivers/gpu/drm/r128/r128_state.c b/drivers/gpu/drm/r128/r128_state.c
index 575e986f82a7..8fd2d9f58f77 100644
--- a/drivers/gpu/drm/r128/r128_state.c
+++ b/drivers/gpu/drm/r128/r128_state.c
@@ -905,7 +905,7 @@ static int r128_cce_dispatch_write_span(struct drm_device *dev,
 	if (IS_ERR(buffer))
 		return PTR_ERR(buffer);
 
-	mask_size = depth->n * sizeof(u8);
+	mask_size = depth->n;
 	if (depth->mask) {
 		mask = memdup_user(depth->mask, mask_size);
 		if (IS_ERR(mask)) {
@@ -1010,7 +1010,7 @@ static int r128_cce_dispatch_write_pixels(struct drm_device *dev,
 	}
 
 	if (depth->mask) {
-		mask_size = depth->n * sizeof(u8);
+		mask_size = depth->n;
 		mask = memdup_user(depth->mask, mask_size);
 		if (IS_ERR(mask)) {
 			kfree(x);
diff --git a/drivers/gpu/drm/radeon/Makefile b/drivers/gpu/drm/radeon/Makefile
index f77b7135ee4c..12bc21219a0e 100644
--- a/drivers/gpu/drm/radeon/Makefile
+++ b/drivers/gpu/drm/radeon/Makefile
@@ -60,7 +60,7 @@ radeon-y := radeon_drv.o
 
 # add UMS driver
 radeon-$(CONFIG_DRM_RADEON_UMS)+= radeon_cp.o radeon_state.o radeon_mem.o \
-	radeon_irq.o r300_cmdbuf.o r600_cp.o r600_blit.o
+	radeon_irq.o r300_cmdbuf.o r600_cp.o r600_blit.o drm_buffer.o
 
 # add KMS driver
 radeon-y += radeon_device.o radeon_asic.o radeon_kms.o \
@@ -72,7 +72,7 @@ radeon-y += radeon_device.o radeon_asic.o radeon_kms.o \
 	radeon_cs.o radeon_bios.o radeon_benchmark.o r100.o r300.o r420.o \
 	rs400.o rs600.o rs690.o rv515.o r520.o r600.o rv770.o radeon_test.o \
 	r200.o radeon_legacy_tv.o r600_cs.o r600_blit_shaders.o \
-	radeon_pm.o atombios_dp.o r600_audio.o r600_hdmi.o dce3_1_afmt.o \
+	radeon_pm.o atombios_dp.o r600_hdmi.o dce3_1_afmt.o \
 	evergreen.o evergreen_cs.o evergreen_blit_shaders.o \
 	evergreen_hdmi.o radeon_trace_points.o ni.o cayman_blit_shaders.o \
 	atombios_encoders.o radeon_semaphore.o radeon_sa.o atombios_i2c.o si.o \
@@ -80,7 +80,8 @@ radeon-y += radeon_device.o radeon_asic.o radeon_kms.o \
 	r600_dpm.o rs780_dpm.o rv6xx_dpm.o rv770_dpm.o rv730_dpm.o rv740_dpm.o \
 	rv770_smc.o cypress_dpm.o btc_dpm.o sumo_dpm.o sumo_smc.o trinity_dpm.o \
 	trinity_smc.o ni_dpm.o si_smc.o si_dpm.o kv_smc.o kv_dpm.o ci_smc.o \
-	ci_dpm.o dce6_afmt.o radeon_vm.o radeon_ucode.o radeon_ib.o
+	ci_dpm.o dce6_afmt.o radeon_vm.o radeon_ucode.o radeon_ib.o radeon_mn.o \
+	radeon_sync.o
 
 # add async DMA block
 radeon-y += \
@@ -104,6 +105,7 @@ radeon-y += \
 	radeon_vce.o \
 	vce_v1_0.o \
 	vce_v2_0.o \
+	radeon_kfd.o
 
 radeon-$(CONFIG_COMPAT) += radeon_ioc32.o
 radeon-$(CONFIG_VGA_SWITCHEROO) += radeon_atpx_handler.o
diff --git a/drivers/gpu/drm/radeon/atom.c b/drivers/gpu/drm/radeon/atom.c
index 15da7ef344a4..ec1593a6a561 100644
--- a/drivers/gpu/drm/radeon/atom.c
+++ b/drivers/gpu/drm/radeon/atom.c
@@ -1217,7 +1217,7 @@ free:
 	return ret;
 }
 
-int atom_execute_table(struct atom_context *ctx, int index, uint32_t * params)
+int atom_execute_table_scratch_unlocked(struct atom_context *ctx, int index, uint32_t * params)
 {
 	int r;
 
@@ -1238,6 +1238,15 @@ int atom_execute_table(struct atom_context *ctx, int index, uint32_t * params)
 	return r;
 }
 
+int atom_execute_table(struct atom_context *ctx, int index, uint32_t * params)
+{
+	int r;
+	mutex_lock(&ctx->scratch_mutex);
+	r = atom_execute_table_scratch_unlocked(ctx, index, params);
+	mutex_unlock(&ctx->scratch_mutex);
+	return r;
+}
+
 static int atom_iio_len[] = { 1, 2, 3, 3, 3, 3, 4, 4, 4, 3 };
 
 static void atom_index_iio(struct atom_context *ctx, int base)
diff --git a/drivers/gpu/drm/radeon/atom.h b/drivers/gpu/drm/radeon/atom.h
index feba6b8d36b3..6d014ddb6b78 100644
--- a/drivers/gpu/drm/radeon/atom.h
+++ b/drivers/gpu/drm/radeon/atom.h
@@ -125,6 +125,7 @@ struct card_info {
 struct atom_context {
 	struct card_info *card;
 	struct mutex mutex;
+	struct mutex scratch_mutex;
 	void *bios;
 	uint32_t cmd_table, data_table;
 	uint16_t *iio;
@@ -145,6 +146,7 @@ extern int atom_debug;
 
 struct atom_context *atom_parse(struct card_info *, void *);
 int atom_execute_table(struct atom_context *, int, uint32_t *);
+int atom_execute_table_scratch_unlocked(struct atom_context *, int, uint32_t *);
 int atom_asic_init(struct atom_context *);
 void atom_destroy(struct atom_context *);
 bool atom_parse_data_header(struct atom_context *ctx, int index, uint16_t *size,
diff --git a/drivers/gpu/drm/radeon/atombios_crtc.c b/drivers/gpu/drm/radeon/atombios_crtc.c
index 30d242b25078..ed644a4f6f57 100644
--- a/drivers/gpu/drm/radeon/atombios_crtc.c
+++ b/drivers/gpu/drm/radeon/atombios_crtc.c
@@ -1851,10 +1851,9 @@ static int radeon_atom_pick_pll(struct drm_crtc *crtc)
 				return pll;
 		}
 		/* otherwise, pick one of the plls */
-		if ((rdev->family == CHIP_KAVERI) ||
-		    (rdev->family == CHIP_KABINI) ||
+		if ((rdev->family == CHIP_KABINI) ||
 		    (rdev->family == CHIP_MULLINS)) {
-			/* KB/KV/ML has PPLL1 and PPLL2 */
+			/* KB/ML has PPLL1 and PPLL2 */
 			pll_in_use = radeon_get_pll_use_mask(crtc);
 			if (!(pll_in_use & (1 << ATOM_PPLL2)))
 				return ATOM_PPLL2;
@@ -1863,7 +1862,7 @@ static int radeon_atom_pick_pll(struct drm_crtc *crtc)
 			DRM_ERROR("unable to allocate a PPLL\n");
 			return ATOM_PPLL_INVALID;
 		} else {
-			/* CI has PPLL0, PPLL1, and PPLL2 */
+			/* CI/KV has PPLL0, PPLL1, and PPLL2 */
 			pll_in_use = radeon_get_pll_use_mask(crtc);
 			if (!(pll_in_use & (1 << ATOM_PPLL2)))
 				return ATOM_PPLL2;
@@ -2039,6 +2038,7 @@ int atombios_crtc_mode_set(struct drm_crtc *crtc,
 	atombios_crtc_set_base(crtc, x, y, old_fb);
 	atombios_overscan_setup(crtc, mode, adjusted_mode);
 	atombios_scaler_setup(crtc);
+	radeon_cursor_reset(crtc);
 	/* update the hw version fpr dpm */
 	radeon_crtc->hw_mode = *adjusted_mode;
 
@@ -2154,6 +2154,7 @@ static void atombios_crtc_disable(struct drm_crtc *crtc)
 	case ATOM_PPLL0:
 		/* disable the ppll */
 		if ((rdev->family == CHIP_ARUBA) ||
+		    (rdev->family == CHIP_KAVERI) ||
 		    (rdev->family == CHIP_BONAIRE) ||
 		    (rdev->family == CHIP_HAWAII))
 			atombios_crtc_program_pll(crtc, radeon_crtc->crtc_id, radeon_crtc->pll_id,
diff --git a/drivers/gpu/drm/radeon/atombios_dp.c b/drivers/gpu/drm/radeon/atombios_dp.c
index ac14b67621d3..db42a670f995 100644
--- a/drivers/gpu/drm/radeon/atombios_dp.c
+++ b/drivers/gpu/drm/radeon/atombios_dp.c
@@ -100,6 +100,7 @@ static int radeon_process_aux_ch(struct radeon_i2c_chan *chan,
 	memset(&args, 0, sizeof(args));
 
 	mutex_lock(&chan->mutex);
+	mutex_lock(&rdev->mode_info.atom_context->scratch_mutex);
 
 	base = (unsigned char *)(rdev->mode_info.atom_context->scratch + 1);
 
@@ -113,7 +114,7 @@ static int radeon_process_aux_ch(struct radeon_i2c_chan *chan,
 	if (ASIC_IS_DCE4(rdev))
 		args.v2.ucHPD_ID = chan->rec.hpd;
 
-	atom_execute_table(rdev->mode_info.atom_context, index, (uint32_t *)&args);
+	atom_execute_table_scratch_unlocked(rdev->mode_info.atom_context, index, (uint32_t *)&args);
 
 	*ack = args.v1.ucReplyStatus;
 
@@ -147,6 +148,7 @@ static int radeon_process_aux_ch(struct radeon_i2c_chan *chan,
 
 	r = recv_bytes;
 done:
+	mutex_unlock(&rdev->mode_info.atom_context->scratch_mutex);
 	mutex_unlock(&chan->mutex);
 
 	return r;
@@ -232,8 +234,8 @@ void radeon_dp_aux_init(struct radeon_connector *radeon_connector)
 
 /***** general DP utility functions *****/
 
-#define DP_VOLTAGE_MAX         DP_TRAIN_VOLTAGE_SWING_1200
-#define DP_PRE_EMPHASIS_MAX    DP_TRAIN_PRE_EMPHASIS_9_5
+#define DP_VOLTAGE_MAX         DP_TRAIN_VOLTAGE_SWING_LEVEL_3
+#define DP_PRE_EMPHASIS_MAX    DP_TRAIN_PRE_EMPH_LEVEL_3
 
 static void dp_get_adjust_train(u8 link_status[DP_LINK_STATUS_SIZE],
 				int lane_count,
@@ -490,6 +492,10 @@ int radeon_dp_mode_valid_helper(struct drm_connector *connector,
 	struct radeon_connector_atom_dig *dig_connector;
 	int dp_clock;
 
+	if ((mode->clock > 340000) &&
+	    (!radeon_connector_is_dp12_capable(connector)))
+		return MODE_CLOCK_HIGH;
+
 	if (!radeon_connector->con_priv)
 		return MODE_CLOCK_HIGH;
 	dig_connector = radeon_connector->con_priv;
diff --git a/drivers/gpu/drm/radeon/atombios_encoders.c b/drivers/gpu/drm/radeon/atombios_encoders.c
index a7f2ddf09a9d..b8cd7975f797 100644
--- a/drivers/gpu/drm/radeon/atombios_encoders.c
+++ b/drivers/gpu/drm/radeon/atombios_encoders.c
@@ -291,29 +291,6 @@ static void radeon_atom_backlight_exit(struct radeon_encoder *encoder)
 bool radeon_atom_get_tv_timings(struct radeon_device *rdev, int index,
 				struct drm_display_mode *mode);
 
-
-static inline bool radeon_encoder_is_digital(struct drm_encoder *encoder)
-{
-	struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder);
-	switch (radeon_encoder->encoder_id) {
-	case ENCODER_OBJECT_ID_INTERNAL_LVDS:
-	case ENCODER_OBJECT_ID_INTERNAL_TMDS1:
-	case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_TMDS1:
-	case ENCODER_OBJECT_ID_INTERNAL_LVTM1:
-	case ENCODER_OBJECT_ID_INTERNAL_DVO1:
-	case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_DVO1:
-	case ENCODER_OBJECT_ID_INTERNAL_DDI:
-	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY:
-	case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_LVTMA:
-	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY1:
-	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY2:
-	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY3:
-		return true;
-	default:
-		return false;
-	}
-}
-
 static bool radeon_atom_mode_fixup(struct drm_encoder *encoder,
 				   const struct drm_display_mode *mode,
 				   struct drm_display_mode *adjusted_mode)
diff --git a/drivers/gpu/drm/radeon/atombios_i2c.c b/drivers/gpu/drm/radeon/atombios_i2c.c
index 9c570fb15b8c..4157780585a0 100644
--- a/drivers/gpu/drm/radeon/atombios_i2c.c
+++ b/drivers/gpu/drm/radeon/atombios_i2c.c
@@ -48,6 +48,7 @@ static int radeon_process_i2c_ch(struct radeon_i2c_chan *chan,
 	memset(&args, 0, sizeof(args));
 
 	mutex_lock(&chan->mutex);
+	mutex_lock(&rdev->mode_info.atom_context->scratch_mutex);
 
 	base = (unsigned char *)rdev->mode_info.atom_context->scratch;
 
@@ -82,7 +83,7 @@ static int radeon_process_i2c_ch(struct radeon_i2c_chan *chan,
 	args.ucSlaveAddr = slave_addr << 1;
 	args.ucLineNumber = chan->rec.i2c_id;
 
-	atom_execute_table(rdev->mode_info.atom_context, index, (uint32_t *)&args);
+	atom_execute_table_scratch_unlocked(rdev->mode_info.atom_context, index, (uint32_t *)&args);
 
 	/* error */
 	if (args.ucStatus != HW_ASSISTED_I2C_STATUS_SUCCESS) {
@@ -95,6 +96,7 @@ static int radeon_process_i2c_ch(struct radeon_i2c_chan *chan,
 		radeon_atom_copy_swap(buf, base, num, false);
 
 done:
+	mutex_unlock(&rdev->mode_info.atom_context->scratch_mutex);
 	mutex_unlock(&chan->mutex);
 
 	return r;
diff --git a/drivers/gpu/drm/radeon/btc_dpm.c b/drivers/gpu/drm/radeon/btc_dpm.c
index f81d7ca134db..0b2929de9f41 100644
--- a/drivers/gpu/drm/radeon/btc_dpm.c
+++ b/drivers/gpu/drm/radeon/btc_dpm.c
@@ -24,6 +24,7 @@
 
 #include "drmP.h"
 #include "radeon.h"
+#include "radeon_asic.h"
 #include "btcd.h"
 #include "r600_dpm.h"
 #include "cypress_dpm.h"
@@ -2099,7 +2100,6 @@ static void btc_apply_state_adjust_rules(struct radeon_device *rdev,
 	bool disable_mclk_switching;
 	u32 mclk, sclk;
 	u16 vddc, vddci;
-	u32 max_sclk_vddc, max_mclk_vddci, max_mclk_vddc;
 
 	if ((rdev->pm.dpm.new_active_crtc_count > 1) ||
 	    btc_dpm_vblank_too_short(rdev))
@@ -2141,39 +2141,6 @@ static void btc_apply_state_adjust_rules(struct radeon_device *rdev,
 			ps->low.vddci = max_limits->vddci;
 	}
 
-	/* limit clocks to max supported clocks based on voltage dependency tables */
-	btc_get_max_clock_from_voltage_dependency_table(&rdev->pm.dpm.dyn_state.vddc_dependency_on_sclk,
-							&max_sclk_vddc);
-	btc_get_max_clock_from_voltage_dependency_table(&rdev->pm.dpm.dyn_state.vddci_dependency_on_mclk,
-							&max_mclk_vddci);
-	btc_get_max_clock_from_voltage_dependency_table(&rdev->pm.dpm.dyn_state.vddc_dependency_on_mclk,
-							&max_mclk_vddc);
-
-	if (max_sclk_vddc) {
-		if (ps->low.sclk > max_sclk_vddc)
-			ps->low.sclk = max_sclk_vddc;
-		if (ps->medium.sclk > max_sclk_vddc)
-			ps->medium.sclk = max_sclk_vddc;
-		if (ps->high.sclk > max_sclk_vddc)
-			ps->high.sclk = max_sclk_vddc;
-	}
-	if (max_mclk_vddci) {
-		if (ps->low.mclk > max_mclk_vddci)
-			ps->low.mclk = max_mclk_vddci;
-		if (ps->medium.mclk > max_mclk_vddci)
-			ps->medium.mclk = max_mclk_vddci;
-		if (ps->high.mclk > max_mclk_vddci)
-			ps->high.mclk = max_mclk_vddci;
-	}
-	if (max_mclk_vddc) {
-		if (ps->low.mclk > max_mclk_vddc)
-			ps->low.mclk = max_mclk_vddc;
-		if (ps->medium.mclk > max_mclk_vddc)
-			ps->medium.mclk = max_mclk_vddc;
-		if (ps->high.mclk > max_mclk_vddc)
-			ps->high.mclk = max_mclk_vddc;
-	}
-
 	/* XXX validate the min clocks required for display */
 
 	if (disable_mclk_switching) {
diff --git a/drivers/gpu/drm/radeon/ci_dpm.c b/drivers/gpu/drm/radeon/ci_dpm.c
index d416bb2ff48d..f373a81ba3d5 100644
--- a/drivers/gpu/drm/radeon/ci_dpm.c
+++ b/drivers/gpu/drm/radeon/ci_dpm.c
@@ -24,6 +24,7 @@
 #include <linux/firmware.h>
 #include "drmP.h"
 #include "radeon.h"
+#include "radeon_asic.h"
 #include "radeon_ucode.h"
 #include "cikd.h"
 #include "r600_dpm.h"
@@ -45,15 +46,15 @@
 static const struct ci_pt_defaults defaults_hawaii_xt =
 {
 	1, 0xF, 0xFD, 0x19, 5, 0x14, 0, 0xB0000,
-	{ 0x84,  0x0,   0x0,   0x7F,  0x0,   0x0,   0x5A,  0x60,  0x51,  0x8E,  0x79,  0x6B,  0x5F,  0x90,  0x79  },
-	{ 0x1EA, 0x1EA, 0x1EA, 0x224, 0x224, 0x224, 0x24F, 0x24F, 0x24F, 0x28E, 0x28E, 0x28E, 0x2BC, 0x2BC, 0x2BC }
+	{ 0x2E,  0x00,  0x00,  0x88,  0x00,  0x00,  0x72,  0x60,  0x51,  0xA7,  0x79,  0x6B,  0x90,  0xBD,  0x79  },
+	{ 0x217, 0x217, 0x217, 0x242, 0x242, 0x242, 0x269, 0x269, 0x269, 0x2A1, 0x2A1, 0x2A1, 0x2C9, 0x2C9, 0x2C9 }
 };
 
 static const struct ci_pt_defaults defaults_hawaii_pro =
 {
 	1, 0xF, 0xFD, 0x19, 5, 0x14, 0, 0x65062,
-	{ 0x93,  0x0,   0x0,   0x97,  0x0,   0x0,   0x6B,  0x60,  0x51,  0x95,  0x79,  0x6B,  0x5F,  0x90,  0x79  },
-	{ 0x1EA, 0x1EA, 0x1EA, 0x224, 0x224, 0x224, 0x24F, 0x24F, 0x24F, 0x28E, 0x28E, 0x28E, 0x2BC, 0x2BC, 0x2BC }
+	{ 0x2E,  0x00,  0x00,  0x88,  0x00,  0x00,  0x72,  0x60,  0x51,  0xA7,  0x79,  0x6B,  0x90,  0xBD,  0x79  },
+	{ 0x217, 0x217, 0x217, 0x242, 0x242, 0x242, 0x269, 0x269, 0x269, 0x2A1, 0x2A1, 0x2A1, 0x2C9, 0x2C9, 0x2C9 }
 };
 
 static const struct ci_pt_defaults defaults_bonaire_xt =
@@ -162,8 +163,6 @@ static const struct ci_pt_config_reg didt_config_ci[] =
 };
 
 extern u8 rv770_get_memory_module_index(struct radeon_device *rdev);
-extern void btc_get_max_clock_from_voltage_dependency_table(struct radeon_clock_voltage_dependency_table *table,
-							    u32 *max_clock);
 extern int ni_copy_and_switch_arb_sets(struct radeon_device *rdev,
 				       u32 arb_freq_src, u32 arb_freq_dest);
 extern u8 si_get_ddr3_mclk_frequency_ratio(u32 memory_clock);
@@ -185,6 +184,9 @@ static int ci_set_overdrive_target_tdp(struct radeon_device *rdev,
 				       u32 target_tdp);
 static int ci_update_uvd_dpm(struct radeon_device *rdev, bool gate);
 
+static PPSMC_Result ci_send_msg_to_smc_with_parameter(struct radeon_device *rdev,
+						      PPSMC_Msg msg, u32 parameter);
+
 static struct ci_power_info *ci_get_pi(struct radeon_device *rdev)
 {
         struct ci_power_info *pi = rdev->pm.dpm.priv;
@@ -250,7 +252,10 @@ static void ci_initialize_powertune_defaults(struct radeon_device *rdev)
 
 	if (pi->caps_power_containment) {
 		pi->caps_cac = true;
-		pi->enable_bapm_feature = true;
+		if (rdev->family == CHIP_HAWAII)
+			pi->enable_bapm_feature = false;
+		else
+			pi->enable_bapm_feature = true;
 		pi->enable_tdc_limit_feature = true;
 		pi->enable_pkg_pwr_tracking_feature = true;
 	}
@@ -353,6 +358,21 @@ static int ci_populate_dw8(struct radeon_device *rdev)
 	return 0;
 }
 
+static int ci_populate_fuzzy_fan(struct radeon_device *rdev)
+{
+	struct ci_power_info *pi = ci_get_pi(rdev);
+
+	if ((rdev->pm.dpm.fan.fan_output_sensitivity & (1 << 15)) ||
+	    (rdev->pm.dpm.fan.fan_output_sensitivity == 0))
+		rdev->pm.dpm.fan.fan_output_sensitivity =
+			rdev->pm.dpm.fan.default_fan_output_sensitivity;
+
+	pi->smc_powertune_table.FuzzyFan_PwmSetDelta =
+		cpu_to_be16(rdev->pm.dpm.fan.fan_output_sensitivity);
+
+	return 0;
+}
+
 static int ci_min_max_v_gnbl_pm_lid_from_bapm_vddc(struct radeon_device *rdev)
 {
 	struct ci_power_info *pi = ci_get_pi(rdev);
@@ -478,6 +498,9 @@ static int ci_populate_pm_base(struct radeon_device *rdev)
 		ret = ci_populate_dw8(rdev);
 		if (ret)
 			return ret;
+		ret = ci_populate_fuzzy_fan(rdev);
+		if (ret)
+			return ret;
 		ret = ci_min_max_v_gnbl_pm_lid_from_bapm_vddc(rdev);
 		if (ret)
 			return ret;
@@ -691,6 +714,25 @@ static int ci_enable_smc_cac(struct radeon_device *rdev, bool enable)
 	return ret;
 }
 
+static int ci_enable_thermal_based_sclk_dpm(struct radeon_device *rdev,
+					    bool enable)
+{
+	struct ci_power_info *pi = ci_get_pi(rdev);
+	PPSMC_Result smc_result = PPSMC_Result_OK;
+
+	if (pi->thermal_sclk_dpm_enabled) {
+		if (enable)
+			smc_result = ci_send_msg_to_smc(rdev, PPSMC_MSG_ENABLE_THERMAL_DPM);
+		else
+			smc_result = ci_send_msg_to_smc(rdev, PPSMC_MSG_DISABLE_THERMAL_DPM);
+	}
+
+	if (smc_result == PPSMC_Result_OK)
+		return 0;
+	else
+		return -EINVAL;
+}
+
 static int ci_power_control_set_level(struct radeon_device *rdev)
 {
 	struct ci_power_info *pi = ci_get_pi(rdev);
@@ -701,13 +743,11 @@ static int ci_power_control_set_level(struct radeon_device *rdev)
 	int ret = 0;
 	bool adjust_polarity = false; /* ??? */
 
-	if (pi->caps_power_containment &&
-	    (pi->power_containment_features & POWERCONTAINMENT_FEATURE_BAPM)) {
+	if (pi->caps_power_containment) {
 		adjust_percent = adjust_polarity ?
 			rdev->pm.dpm.tdp_adjustment : (-1 * rdev->pm.dpm.tdp_adjustment);
 		target_tdp = ((100 + adjust_percent) *
 			      (s32)cac_tdp_table->configurable_tdp) / 100;
-		target_tdp *= 256;
 
 		ret = ci_set_overdrive_target_tdp(rdev, (u32)target_tdp);
 	}
@@ -748,7 +788,6 @@ static void ci_apply_state_adjust_rules(struct radeon_device *rdev,
 	struct radeon_clock_and_voltage_limits *max_limits;
 	bool disable_mclk_switching;
 	u32 sclk, mclk;
-	u32 max_sclk_vddc, max_mclk_vddci, max_mclk_vddc;
 	int i;
 
 	if (rps->vce_active) {
@@ -784,29 +823,6 @@ static void ci_apply_state_adjust_rules(struct radeon_device *rdev,
 		}
 	}
 
-	/* limit clocks to max supported clocks based on voltage dependency tables */
-	btc_get_max_clock_from_voltage_dependency_table(&rdev->pm.dpm.dyn_state.vddc_dependency_on_sclk,
-							&max_sclk_vddc);
-	btc_get_max_clock_from_voltage_dependency_table(&rdev->pm.dpm.dyn_state.vddci_dependency_on_mclk,
-							&max_mclk_vddci);
-	btc_get_max_clock_from_voltage_dependency_table(&rdev->pm.dpm.dyn_state.vddc_dependency_on_mclk,
-							&max_mclk_vddc);
-
-	for (i = 0; i < ps->performance_level_count; i++) {
-		if (max_sclk_vddc) {
-			if (ps->performance_levels[i].sclk > max_sclk_vddc)
-				ps->performance_levels[i].sclk = max_sclk_vddc;
-		}
-		if (max_mclk_vddci) {
-			if (ps->performance_levels[i].mclk > max_mclk_vddci)
-				ps->performance_levels[i].mclk = max_mclk_vddci;
-		}
-		if (max_mclk_vddc) {
-			if (ps->performance_levels[i].mclk > max_mclk_vddc)
-				ps->performance_levels[i].mclk = max_mclk_vddc;
-		}
-	}
-
 	/* XXX validate the min clocks required for display */
 
 	if (disable_mclk_switching) {
@@ -839,7 +855,7 @@ static void ci_apply_state_adjust_rules(struct radeon_device *rdev,
 	}
 }
 
-static int ci_set_thermal_temperature_range(struct radeon_device *rdev,
+static int ci_thermal_set_temperature_range(struct radeon_device *rdev,
 					    int min_temp, int max_temp)
 {
 	int low_temp = 0 * 1000;
@@ -875,6 +891,350 @@ static int ci_set_thermal_temperature_range(struct radeon_device *rdev,
 	return 0;
 }
 
+static int ci_thermal_enable_alert(struct radeon_device *rdev,
+				   bool enable)
+{
+	u32 thermal_int = RREG32_SMC(CG_THERMAL_INT);
+	PPSMC_Result result;
+
+	if (enable) {
+		thermal_int &= ~(THERM_INT_MASK_HIGH | THERM_INT_MASK_LOW);
+		WREG32_SMC(CG_THERMAL_INT, thermal_int);
+		rdev->irq.dpm_thermal = false;
+		result = ci_send_msg_to_smc(rdev, PPSMC_MSG_Thermal_Cntl_Enable);
+		if (result != PPSMC_Result_OK) {
+			DRM_DEBUG_KMS("Could not enable thermal interrupts.\n");
+			return -EINVAL;
+		}
+	} else {
+		thermal_int |= THERM_INT_MASK_HIGH | THERM_INT_MASK_LOW;
+		WREG32_SMC(CG_THERMAL_INT, thermal_int);
+		rdev->irq.dpm_thermal = true;
+		result = ci_send_msg_to_smc(rdev, PPSMC_MSG_Thermal_Cntl_Disable);
+		if (result != PPSMC_Result_OK) {
+			DRM_DEBUG_KMS("Could not disable thermal interrupts.\n");
+			return -EINVAL;
+		}
+	}
+
+	return 0;
+}
+
+static void ci_fan_ctrl_set_static_mode(struct radeon_device *rdev, u32 mode)
+{
+	struct ci_power_info *pi = ci_get_pi(rdev);
+	u32 tmp;
+
+	if (pi->fan_ctrl_is_in_default_mode) {
+		tmp = (RREG32_SMC(CG_FDO_CTRL2) & FDO_PWM_MODE_MASK) >> FDO_PWM_MODE_SHIFT;
+		pi->fan_ctrl_default_mode = tmp;
+		tmp = (RREG32_SMC(CG_FDO_CTRL2) & TMIN_MASK) >> TMIN_SHIFT;
+		pi->t_min = tmp;
+		pi->fan_ctrl_is_in_default_mode = false;
+	}
+
+	tmp = RREG32_SMC(CG_FDO_CTRL2) & ~TMIN_MASK;
+	tmp |= TMIN(0);
+	WREG32_SMC(CG_FDO_CTRL2, tmp);
+
+	tmp = RREG32_SMC(CG_FDO_CTRL2) & ~FDO_PWM_MODE_MASK;
+	tmp |= FDO_PWM_MODE(mode);
+	WREG32_SMC(CG_FDO_CTRL2, tmp);
+}
+
+static int ci_thermal_setup_fan_table(struct radeon_device *rdev)
+{
+	struct ci_power_info *pi = ci_get_pi(rdev);
+	SMU7_Discrete_FanTable fan_table = { FDO_MODE_HARDWARE };
+	u32 duty100;
+	u32 t_diff1, t_diff2, pwm_diff1, pwm_diff2;
+	u16 fdo_min, slope1, slope2;
+	u32 reference_clock, tmp;
+	int ret;
+	u64 tmp64;
+
+	if (!pi->fan_table_start) {
+		rdev->pm.dpm.fan.ucode_fan_control = false;
+		return 0;
+	}
+
+	duty100 = (RREG32_SMC(CG_FDO_CTRL1) & FMAX_DUTY100_MASK) >> FMAX_DUTY100_SHIFT;
+
+	if (duty100 == 0) {
+		rdev->pm.dpm.fan.ucode_fan_control = false;
+		return 0;
+	}
+
+	tmp64 = (u64)rdev->pm.dpm.fan.pwm_min * duty100;
+	do_div(tmp64, 10000);
+	fdo_min = (u16)tmp64;
+
+	t_diff1 = rdev->pm.dpm.fan.t_med - rdev->pm.dpm.fan.t_min;
+	t_diff2 = rdev->pm.dpm.fan.t_high - rdev->pm.dpm.fan.t_med;
+
+	pwm_diff1 = rdev->pm.dpm.fan.pwm_med - rdev->pm.dpm.fan.pwm_min;
+	pwm_diff2 = rdev->pm.dpm.fan.pwm_high - rdev->pm.dpm.fan.pwm_med;
+
+	slope1 = (u16)((50 + ((16 * duty100 * pwm_diff1) / t_diff1)) / 100);
+	slope2 = (u16)((50 + ((16 * duty100 * pwm_diff2) / t_diff2)) / 100);
+
+	fan_table.TempMin = cpu_to_be16((50 + rdev->pm.dpm.fan.t_min) / 100);
+	fan_table.TempMed = cpu_to_be16((50 + rdev->pm.dpm.fan.t_med) / 100);
+	fan_table.TempMax = cpu_to_be16((50 + rdev->pm.dpm.fan.t_max) / 100);
+
+	fan_table.Slope1 = cpu_to_be16(slope1);
+	fan_table.Slope2 = cpu_to_be16(slope2);
+
+	fan_table.FdoMin = cpu_to_be16(fdo_min);
+
+	fan_table.HystDown = cpu_to_be16(rdev->pm.dpm.fan.t_hyst);
+
+	fan_table.HystUp = cpu_to_be16(1);
+
+	fan_table.HystSlope = cpu_to_be16(1);
+
+	fan_table.TempRespLim = cpu_to_be16(5);
+
+	reference_clock = radeon_get_xclk(rdev);
+
+	fan_table.RefreshPeriod = cpu_to_be32((rdev->pm.dpm.fan.cycle_delay *
+					       reference_clock) / 1600);
+
+	fan_table.FdoMax = cpu_to_be16((u16)duty100);
+
+	tmp = (RREG32_SMC(CG_MULT_THERMAL_CTRL) & TEMP_SEL_MASK) >> TEMP_SEL_SHIFT;
+	fan_table.TempSrc = (uint8_t)tmp;
+
+	ret = ci_copy_bytes_to_smc(rdev,
+				   pi->fan_table_start,
+				   (u8 *)(&fan_table),
+				   sizeof(fan_table),
+				   pi->sram_end);
+
+	if (ret) {
+		DRM_ERROR("Failed to load fan table to the SMC.");
+		rdev->pm.dpm.fan.ucode_fan_control = false;
+	}
+
+	return 0;
+}
+
+static int ci_fan_ctrl_start_smc_fan_control(struct radeon_device *rdev)
+{
+	struct ci_power_info *pi = ci_get_pi(rdev);
+	PPSMC_Result ret;
+
+	if (pi->caps_od_fuzzy_fan_control_support) {
+		ret = ci_send_msg_to_smc_with_parameter(rdev,
+							PPSMC_StartFanControl,
+							FAN_CONTROL_FUZZY);
+		if (ret != PPSMC_Result_OK)
+			return -EINVAL;
+		ret = ci_send_msg_to_smc_with_parameter(rdev,
+							PPSMC_MSG_SetFanPwmMax,
+							rdev->pm.dpm.fan.default_max_fan_pwm);
+		if (ret != PPSMC_Result_OK)
+			return -EINVAL;
+	} else {
+		ret = ci_send_msg_to_smc_with_parameter(rdev,
+							PPSMC_StartFanControl,
+							FAN_CONTROL_TABLE);
+		if (ret != PPSMC_Result_OK)
+			return -EINVAL;
+	}
+
+	return 0;
+}
+
+#if 0
+static int ci_fan_ctrl_stop_smc_fan_control(struct radeon_device *rdev)
+{
+	PPSMC_Result ret;
+
+	ret = ci_send_msg_to_smc(rdev, PPSMC_StopFanControl);
+	if (ret == PPSMC_Result_OK)
+		return 0;
+	else
+		return -EINVAL;
+}
+
+static int ci_fan_ctrl_get_fan_speed_percent(struct radeon_device *rdev,
+					     u32 *speed)
+{
+	u32 duty, duty100;
+	u64 tmp64;
+
+	if (rdev->pm.no_fan)
+		return -ENOENT;
+
+	duty100 = (RREG32_SMC(CG_FDO_CTRL1) & FMAX_DUTY100_MASK) >> FMAX_DUTY100_SHIFT;
+	duty = (RREG32_SMC(CG_THERMAL_STATUS) & FDO_PWM_DUTY_MASK) >> FDO_PWM_DUTY_SHIFT;
+
+	if (duty100 == 0)
+		return -EINVAL;
+
+	tmp64 = (u64)duty * 100;
+	do_div(tmp64, duty100);
+	*speed = (u32)tmp64;
+
+	if (*speed > 100)
+		*speed = 100;
+
+	return 0;
+}
+
+static int ci_fan_ctrl_set_fan_speed_percent(struct radeon_device *rdev,
+					     u32 speed)
+{
+	u32 tmp;
+	u32 duty, duty100;
+	u64 tmp64;
+
+	if (rdev->pm.no_fan)
+		return -ENOENT;
+
+	if (speed > 100)
+		return -EINVAL;
+
+	if (rdev->pm.dpm.fan.ucode_fan_control)
+		ci_fan_ctrl_stop_smc_fan_control(rdev);
+
+	duty100 = (RREG32_SMC(CG_FDO_CTRL1) & FMAX_DUTY100_MASK) >> FMAX_DUTY100_SHIFT;
+
+	if (duty100 == 0)
+		return -EINVAL;
+
+	tmp64 = (u64)speed * duty100;
+	do_div(tmp64, 100);
+	duty = (u32)tmp64;
+
+	tmp = RREG32_SMC(CG_FDO_CTRL0) & ~FDO_STATIC_DUTY_MASK;
+	tmp |= FDO_STATIC_DUTY(duty);
+	WREG32_SMC(CG_FDO_CTRL0, tmp);
+
+	ci_fan_ctrl_set_static_mode(rdev, FDO_PWM_MODE_STATIC);
+
+	return 0;
+}
+
+static int ci_fan_ctrl_get_fan_speed_rpm(struct radeon_device *rdev,
+					 u32 *speed)
+{
+	u32 tach_period;
+	u32 xclk = radeon_get_xclk(rdev);
+
+	if (rdev->pm.no_fan)
+		return -ENOENT;
+
+	if (rdev->pm.fan_pulses_per_revolution == 0)
+		return -ENOENT;
+
+	tach_period = (RREG32_SMC(CG_TACH_STATUS) & TACH_PERIOD_MASK) >> TACH_PERIOD_SHIFT;
+	if (tach_period == 0)
+		return -ENOENT;
+
+	*speed = 60 * xclk * 10000 / tach_period;
+
+	return 0;
+}
+
+static int ci_fan_ctrl_set_fan_speed_rpm(struct radeon_device *rdev,
+					 u32 speed)
+{
+	u32 tach_period, tmp;
+	u32 xclk = radeon_get_xclk(rdev);
+
+	if (rdev->pm.no_fan)
+		return -ENOENT;
+
+	if (rdev->pm.fan_pulses_per_revolution == 0)
+		return -ENOENT;
+
+	if ((speed < rdev->pm.fan_min_rpm) ||
+	    (speed > rdev->pm.fan_max_rpm))
+		return -EINVAL;
+
+	if (rdev->pm.dpm.fan.ucode_fan_control)
+		ci_fan_ctrl_stop_smc_fan_control(rdev);
+
+	tach_period = 60 * xclk * 10000 / (8 * speed);
+	tmp = RREG32_SMC(CG_TACH_CTRL) & ~TARGET_PERIOD_MASK;
+	tmp |= TARGET_PERIOD(tach_period);
+	WREG32_SMC(CG_TACH_CTRL, tmp);
+
+	ci_fan_ctrl_set_static_mode(rdev, FDO_PWM_MODE_STATIC_RPM);
+
+	return 0;
+}
+#endif
+
+static void ci_fan_ctrl_set_default_mode(struct radeon_device *rdev)
+{
+	struct ci_power_info *pi = ci_get_pi(rdev);
+	u32 tmp;
+
+	if (!pi->fan_ctrl_is_in_default_mode) {
+		tmp = RREG32_SMC(CG_FDO_CTRL2) & ~FDO_PWM_MODE_MASK;
+		tmp |= FDO_PWM_MODE(pi->fan_ctrl_default_mode);
+		WREG32_SMC(CG_FDO_CTRL2, tmp);
+
+		tmp = RREG32_SMC(CG_FDO_CTRL2) & ~TMIN_MASK;
+		tmp |= TMIN(pi->t_min);
+		WREG32_SMC(CG_FDO_CTRL2, tmp);
+		pi->fan_ctrl_is_in_default_mode = true;
+	}
+}
+
+static void ci_thermal_start_smc_fan_control(struct radeon_device *rdev)
+{
+	if (rdev->pm.dpm.fan.ucode_fan_control) {
+		ci_fan_ctrl_start_smc_fan_control(rdev);
+		ci_fan_ctrl_set_static_mode(rdev, FDO_PWM_MODE_STATIC);
+	}
+}
+
+static void ci_thermal_initialize(struct radeon_device *rdev)
+{
+	u32 tmp;
+
+	if (rdev->pm.fan_pulses_per_revolution) {
+		tmp = RREG32_SMC(CG_TACH_CTRL) & ~EDGE_PER_REV_MASK;
+		tmp |= EDGE_PER_REV(rdev->pm.fan_pulses_per_revolution -1);
+		WREG32_SMC(CG_TACH_CTRL, tmp);
+	}
+
+	tmp = RREG32_SMC(CG_FDO_CTRL2) & ~TACH_PWM_RESP_RATE_MASK;
+	tmp |= TACH_PWM_RESP_RATE(0x28);
+	WREG32_SMC(CG_FDO_CTRL2, tmp);
+}
+
+static int ci_thermal_start_thermal_controller(struct radeon_device *rdev)
+{
+	int ret;
+
+	ci_thermal_initialize(rdev);
+	ret = ci_thermal_set_temperature_range(rdev, R600_TEMP_RANGE_MIN, R600_TEMP_RANGE_MAX);
+	if (ret)
+		return ret;
+	ret = ci_thermal_enable_alert(rdev, true);
+	if (ret)
+		return ret;
+	if (rdev->pm.dpm.fan.ucode_fan_control) {
+		ret = ci_thermal_setup_fan_table(rdev);
+		if (ret)
+			return ret;
+		ci_thermal_start_smc_fan_control(rdev);
+	}
+
+	return 0;
+}
+
+static void ci_thermal_stop_thermal_controller(struct radeon_device *rdev)
+{
+	if (!rdev->pm.no_fan)
+		ci_fan_ctrl_set_default_mode(rdev);
+}
+
 #if 0
 static int ci_read_smc_soft_register(struct radeon_device *rdev,
 				     u16 reg_offset, u32 *value)
@@ -1278,7 +1638,7 @@ static int ci_dpm_force_state_sclk(struct radeon_device *rdev, u32 n)
 
 	if (!pi->sclk_dpm_key_disabled) {
 		PPSMC_Result smc_result =
-			ci_send_msg_to_smc_with_parameter(rdev, PPSMC_MSG_DPM_ForceState, n);
+			ci_send_msg_to_smc_with_parameter(rdev, PPSMC_MSG_SCLKDPM_SetEnabledMask, 1 << n);
 		if (smc_result != PPSMC_Result_OK)
 			return -EINVAL;
 	}
@@ -1292,7 +1652,7 @@ static int ci_dpm_force_state_mclk(struct radeon_device *rdev, u32 n)
 
 	if (!pi->mclk_dpm_key_disabled) {
 		PPSMC_Result smc_result =
-			ci_send_msg_to_smc_with_parameter(rdev, PPSMC_MSG_MCLKDPM_ForceState, n);
+			ci_send_msg_to_smc_with_parameter(rdev, PPSMC_MSG_MCLKDPM_SetEnabledMask, 1 << n);
 		if (smc_result != PPSMC_Result_OK)
 			return -EINVAL;
 	}
@@ -2067,6 +2427,33 @@ static int ci_force_switch_to_arb_f0(struct radeon_device *rdev)
 	return ni_copy_and_switch_arb_sets(rdev, tmp, MC_CG_ARB_FREQ_F0);
 }
 
+static void ci_register_patching_mc_arb(struct radeon_device *rdev,
+					const u32 engine_clock,
+					const u32 memory_clock,
+					u32 *dram_timimg2)
+{
+	bool patch;
+	u32 tmp, tmp2;
+
+	tmp = RREG32(MC_SEQ_MISC0);
+	patch = ((tmp & 0x0000f00) == 0x300) ? true : false;
+
+	if (patch &&
+	    ((rdev->pdev->device == 0x67B0) ||
+	     (rdev->pdev->device == 0x67B1))) {
+		if ((memory_clock > 100000) && (memory_clock <= 125000)) {
+			tmp2 = (((0x31 * engine_clock) / 125000) - 1) & 0xff;
+			*dram_timimg2 &= ~0x00ff0000;
+			*dram_timimg2 |= tmp2 << 16;
+		} else if ((memory_clock > 125000) && (memory_clock <= 137500)) {
+			tmp2 = (((0x36 * engine_clock) / 137500) - 1) & 0xff;
+			*dram_timimg2 &= ~0x00ff0000;
+			*dram_timimg2 |= tmp2 << 16;
+		}
+	}
+}
+
+
 static int ci_populate_memory_timing_parameters(struct radeon_device *rdev,
 						u32 sclk,
 						u32 mclk,
@@ -2082,6 +2469,8 @@ static int ci_populate_memory_timing_parameters(struct radeon_device *rdev,
 	dram_timing2 = RREG32(MC_ARB_DRAM_TIMING2);
 	burst_time = RREG32(MC_ARB_BURST_TIME) & STATE0_MASK;
 
+	ci_register_patching_mc_arb(rdev, sclk, mclk, &dram_timing2);
+
 	arb_regs->McArbDramTiming  = cpu_to_be32(dram_timing);
 	arb_regs->McArbDramTiming2 = cpu_to_be32(dram_timing2);
 	arb_regs->McArbBurstTime = (u8)burst_time;
@@ -2376,10 +2765,10 @@ static int ci_calculate_mclk_params(struct radeon_device *rdev,
 		u32 tmp;
 		u32 reference_clock = rdev->clock.mpll.reference_freq;
 
-		if (pi->mem_gddr5)
-			freq_nom = memory_clock * 4;
+		if (mpll_param.qdr == 1)
+			freq_nom = memory_clock * 4 * (1 << mpll_param.post_div);
 		else
-			freq_nom = memory_clock * 2;
+			freq_nom = memory_clock * 2 * (1 << mpll_param.post_div);
 
 		tmp = (freq_nom / reference_clock);
 		tmp = tmp * tmp;
@@ -2459,7 +2848,6 @@ static int ci_populate_single_memory_level(struct radeon_device *rdev,
 						      &memory_level->MinVddcPhases);
 
 	memory_level->EnabledForThrottle = 1;
-	memory_level->EnabledForActivity = 1;
 	memory_level->UpH = 0;
 	memory_level->DownH = 100;
 	memory_level->VoltageDownH = 0;
@@ -2792,7 +3180,6 @@ static int ci_populate_single_graphic_level(struct radeon_device *rdev,
 
 	graphic_level->CcPwrDynRm = 0;
 	graphic_level->CcPwrDynRm1 = 0;
-	graphic_level->EnabledForActivity = 1;
 	graphic_level->EnabledForThrottle = 1;
 	graphic_level->UpH = 0;
 	graphic_level->DownH = 0;
@@ -2841,10 +3228,13 @@ static int ci_populate_all_graphic_levels(struct radeon_device *rdev)
 						       &pi->smc_state_table.GraphicsLevel[i]);
 		if (ret)
 			return ret;
+		if (i > 1)
+			pi->smc_state_table.GraphicsLevel[i].DeepSleepDivId = 0;
 		if (i == (dpm_table->sclk_table.count - 1))
 			pi->smc_state_table.GraphicsLevel[i].DisplayWatermark =
 				PPSMC_DISPLAY_WATERMARK_HIGH;
 	}
+	pi->smc_state_table.GraphicsLevel[0].EnabledForActivity = 1;
 
 	pi->smc_state_table.GraphicsDpmLevelCount = (u8)dpm_table->sclk_table.count;
 	pi->dpm_level_enable_mask.sclk_dpm_enable_mask =
@@ -2888,6 +3278,16 @@ static int ci_populate_all_memory_levels(struct radeon_device *rdev)
 			return ret;
 	}
 
+	pi->smc_state_table.MemoryLevel[0].EnabledForActivity = 1;
+
+	if ((dpm_table->mclk_table.count >= 2) &&
+	    ((rdev->pdev->device == 0x67B0) || (rdev->pdev->device == 0x67B1))) {
+		pi->smc_state_table.MemoryLevel[1].MinVddc =
+			pi->smc_state_table.MemoryLevel[0].MinVddc;
+		pi->smc_state_table.MemoryLevel[1].MinVddcPhases =
+			pi->smc_state_table.MemoryLevel[0].MinVddcPhases;
+	}
+
 	pi->smc_state_table.MemoryLevel[0].ActivityLevel = cpu_to_be16(0x1F);
 
 	pi->smc_state_table.MemoryDpmLevelCount = (u8)dpm_table->mclk_table.count;
@@ -2944,9 +3344,14 @@ static int ci_setup_default_pcie_tables(struct radeon_device *rdev)
 				  &pi->dpm_table.pcie_speed_table,
 				  SMU7_MAX_LEVELS_LINK);
 
-	ci_setup_pcie_table_entry(&pi->dpm_table.pcie_speed_table, 0,
-				  pi->pcie_gen_powersaving.min,
-				  pi->pcie_lane_powersaving.min);
+	if (rdev->family == CHIP_BONAIRE)
+		ci_setup_pcie_table_entry(&pi->dpm_table.pcie_speed_table, 0,
+					  pi->pcie_gen_powersaving.min,
+					  pi->pcie_lane_powersaving.max);
+	else
+		ci_setup_pcie_table_entry(&pi->dpm_table.pcie_speed_table, 0,
+					  pi->pcie_gen_powersaving.min,
+					  pi->pcie_lane_powersaving.min);
 	ci_setup_pcie_table_entry(&pi->dpm_table.pcie_speed_table, 1,
 				  pi->pcie_gen_performance.min,
 				  pi->pcie_lane_performance.min);
@@ -3013,19 +3418,21 @@ static int ci_setup_default_dpm_tables(struct radeon_device *rdev)
 		     allowed_sclk_vddc_table->entries[i].clk)) {
 			pi->dpm_table.sclk_table.dpm_levels[pi->dpm_table.sclk_table.count].value =
 				allowed_sclk_vddc_table->entries[i].clk;
-			pi->dpm_table.sclk_table.dpm_levels[pi->dpm_table.sclk_table.count].enabled = true;
+			pi->dpm_table.sclk_table.dpm_levels[pi->dpm_table.sclk_table.count].enabled =
+				(i == 0) ? true : false;
 			pi->dpm_table.sclk_table.count++;
 		}
 	}
 
 	pi->dpm_table.mclk_table.count = 0;
 	for (i = 0; i < allowed_mclk_table->count; i++) {
-		if ((i==0) ||
+		if ((i == 0) ||
 		    (pi->dpm_table.mclk_table.dpm_levels[pi->dpm_table.mclk_table.count-1].value !=
 		     allowed_mclk_table->entries[i].clk)) {
 			pi->dpm_table.mclk_table.dpm_levels[pi->dpm_table.mclk_table.count].value =
 				allowed_mclk_table->entries[i].clk;
-			pi->dpm_table.mclk_table.dpm_levels[pi->dpm_table.mclk_table.count].enabled = true;
+			pi->dpm_table.mclk_table.dpm_levels[pi->dpm_table.mclk_table.count].enabled =
+				(i == 0) ? true : false;
 			pi->dpm_table.mclk_table.count++;
 		}
 	}
@@ -3191,7 +3598,7 @@ static int ci_init_smc_table(struct radeon_device *rdev)
 	table->VddcVddciDelta = 4000;
 	table->PhaseResponseTime = 0;
 	table->MemoryThermThrottleEnable = 1;
-	table->PCIeBootLinkLevel = 0;
+	table->PCIeBootLinkLevel = pi->dpm_table.pcie_speed_table.count - 1;
 	table->PCIeGenInterval = 1;
 	if (pi->voltage_control == CISLANDS_VOLTAGE_CONTROL_BY_SVID2)
 		table->SVI2Enable  = 1;
@@ -3345,6 +3752,8 @@ static int ci_upload_dpm_level_enable_mask(struct radeon_device *rdev)
 	struct ci_power_info *pi = ci_get_pi(rdev);
 	PPSMC_Result result;
 
+	ci_apply_disp_minimum_voltage_request(rdev);
+
 	if (!pi->sclk_dpm_key_disabled) {
 		if (pi->dpm_level_enable_mask.sclk_dpm_enable_mask) {
 			result = ci_send_msg_to_smc_with_parameter(rdev,
@@ -3364,7 +3773,7 @@ static int ci_upload_dpm_level_enable_mask(struct radeon_device *rdev)
 				return -EINVAL;
 		}
 	}
-
+#if 0
 	if (!pi->pcie_dpm_key_disabled) {
 		if (pi->dpm_level_enable_mask.pcie_dpm_enable_mask) {
 			result = ci_send_msg_to_smc_with_parameter(rdev,
@@ -3374,9 +3783,7 @@ static int ci_upload_dpm_level_enable_mask(struct radeon_device *rdev)
 				return -EINVAL;
 		}
 	}
-
-	ci_apply_disp_minimum_voltage_request(rdev);
-
+#endif
 	return 0;
 }
 
@@ -3402,7 +3809,7 @@ static void ci_find_dpm_states_clocks_in_dpm_table(struct radeon_device *rdev,
 		pi->need_update_smu7_dpm_table |= DPMTABLE_OD_UPDATE_SCLK;
 	} else {
 		/* XXX check display min clock requirements */
-		if (0 != CISLAND_MINIMUM_ENGINE_CLOCK)
+		if (CISLAND_MINIMUM_ENGINE_CLOCK != CISLAND_MINIMUM_ENGINE_CLOCK)
 			pi->need_update_smu7_dpm_table |= DPMTABLE_UPDATE_SCLK;
 	}
 
@@ -3732,62 +4139,61 @@ int ci_dpm_force_performance_level(struct radeon_device *rdev,
 				   enum radeon_dpm_forced_level level)
 {
 	struct ci_power_info *pi = ci_get_pi(rdev);
-	PPSMC_Result smc_result;
 	u32 tmp, levels, i;
 	int ret;
 
 	if (level == RADEON_DPM_FORCED_LEVEL_HIGH) {
-		if ((!pi->sclk_dpm_key_disabled) &&
-		    pi->dpm_level_enable_mask.sclk_dpm_enable_mask) {
+		if ((!pi->pcie_dpm_key_disabled) &&
+		    pi->dpm_level_enable_mask.pcie_dpm_enable_mask) {
 			levels = 0;
-			tmp = pi->dpm_level_enable_mask.sclk_dpm_enable_mask;
+			tmp = pi->dpm_level_enable_mask.pcie_dpm_enable_mask;
 			while (tmp >>= 1)
 				levels++;
 			if (levels) {
-				ret = ci_dpm_force_state_sclk(rdev, levels);
+				ret = ci_dpm_force_state_pcie(rdev, level);
 				if (ret)
 					return ret;
 				for (i = 0; i < rdev->usec_timeout; i++) {
-					tmp = (RREG32_SMC(TARGET_AND_CURRENT_PROFILE_INDEX) &
-					       CURR_SCLK_INDEX_MASK) >> CURR_SCLK_INDEX_SHIFT;
+					tmp = (RREG32_SMC(TARGET_AND_CURRENT_PROFILE_INDEX_1) &
+					       CURR_PCIE_INDEX_MASK) >> CURR_PCIE_INDEX_SHIFT;
 					if (tmp == levels)
 						break;
 					udelay(1);
 				}
 			}
 		}
-		if ((!pi->mclk_dpm_key_disabled) &&
-		    pi->dpm_level_enable_mask.mclk_dpm_enable_mask) {
+		if ((!pi->sclk_dpm_key_disabled) &&
+		    pi->dpm_level_enable_mask.sclk_dpm_enable_mask) {
 			levels = 0;
-			tmp = pi->dpm_level_enable_mask.mclk_dpm_enable_mask;
+			tmp = pi->dpm_level_enable_mask.sclk_dpm_enable_mask;
 			while (tmp >>= 1)
 				levels++;
 			if (levels) {
-				ret = ci_dpm_force_state_mclk(rdev, levels);
+				ret = ci_dpm_force_state_sclk(rdev, levels);
 				if (ret)
 					return ret;
 				for (i = 0; i < rdev->usec_timeout; i++) {
 					tmp = (RREG32_SMC(TARGET_AND_CURRENT_PROFILE_INDEX) &
-					       CURR_MCLK_INDEX_MASK) >> CURR_MCLK_INDEX_SHIFT;
+					       CURR_SCLK_INDEX_MASK) >> CURR_SCLK_INDEX_SHIFT;
 					if (tmp == levels)
 						break;
 					udelay(1);
 				}
 			}
 		}
-		if ((!pi->pcie_dpm_key_disabled) &&
-		    pi->dpm_level_enable_mask.pcie_dpm_enable_mask) {
+		if ((!pi->mclk_dpm_key_disabled) &&
+		    pi->dpm_level_enable_mask.mclk_dpm_enable_mask) {
 			levels = 0;
-			tmp = pi->dpm_level_enable_mask.pcie_dpm_enable_mask;
+			tmp = pi->dpm_level_enable_mask.mclk_dpm_enable_mask;
 			while (tmp >>= 1)
 				levels++;
 			if (levels) {
-				ret = ci_dpm_force_state_pcie(rdev, level);
+				ret = ci_dpm_force_state_mclk(rdev, levels);
 				if (ret)
 					return ret;
 				for (i = 0; i < rdev->usec_timeout; i++) {
-					tmp = (RREG32_SMC(TARGET_AND_CURRENT_PROFILE_INDEX_1) &
-					       CURR_PCIE_INDEX_MASK) >> CURR_PCIE_INDEX_SHIFT;
+					tmp = (RREG32_SMC(TARGET_AND_CURRENT_PROFILE_INDEX) &
+					       CURR_MCLK_INDEX_MASK) >> CURR_MCLK_INDEX_SHIFT;
 					if (tmp == levels)
 						break;
 					udelay(1);
@@ -3841,21 +4247,17 @@ int ci_dpm_force_performance_level(struct radeon_device *rdev,
 			}
 		}
 	} else if (level == RADEON_DPM_FORCED_LEVEL_AUTO) {
-		if (!pi->sclk_dpm_key_disabled) {
-			smc_result = ci_send_msg_to_smc(rdev, PPSMC_MSG_NoForcedLevel);
-			if (smc_result != PPSMC_Result_OK)
-				return -EINVAL;
-		}
-		if (!pi->mclk_dpm_key_disabled) {
-			smc_result = ci_send_msg_to_smc(rdev, PPSMC_MSG_MCLKDPM_NoForcedLevel);
-			if (smc_result != PPSMC_Result_OK)
-				return -EINVAL;
-		}
 		if (!pi->pcie_dpm_key_disabled) {
-			smc_result = ci_send_msg_to_smc(rdev, PPSMC_MSG_PCIeDPM_UnForceLevel);
+			PPSMC_Result smc_result;
+
+			smc_result = ci_send_msg_to_smc(rdev,
+							PPSMC_MSG_PCIeDPM_UnForceLevel);
 			if (smc_result != PPSMC_Result_OK)
 				return -EINVAL;
 		}
+		ret = ci_upload_dpm_level_enable_mask(rdev);
+		if (ret)
+			return ret;
 	}
 
 	rdev->pm.dpm.forced_level = level;
@@ -4061,6 +4463,96 @@ static int ci_copy_vbios_mc_reg_table(const struct atom_mc_reg_table *table,
 	return 0;
 }
 
+static int ci_register_patching_mc_seq(struct radeon_device *rdev,
+				       struct ci_mc_reg_table *table)
+{
+	u8 i, k;
+	u32 tmp;
+	bool patch;
+
+	tmp = RREG32(MC_SEQ_MISC0);
+	patch = ((tmp & 0x0000f00) == 0x300) ? true : false;
+
+	if (patch &&
+	    ((rdev->pdev->device == 0x67B0) ||
+	     (rdev->pdev->device == 0x67B1))) {
+		for (i = 0; i < table->last; i++) {
+			if (table->last >= SMU7_DISCRETE_MC_REGISTER_ARRAY_SIZE)
+				return -EINVAL;
+			switch(table->mc_reg_address[i].s1 >> 2) {
+			case MC_SEQ_MISC1:
+				for (k = 0; k < table->num_entries; k++) {
+					if ((table->mc_reg_table_entry[k].mclk_max == 125000) ||
+					    (table->mc_reg_table_entry[k].mclk_max == 137500))
+						table->mc_reg_table_entry[k].mc_data[i] =
+							(table->mc_reg_table_entry[k].mc_data[i] & 0xFFFFFFF8) |
+							0x00000007;
+				}
+				break;
+			case MC_SEQ_WR_CTL_D0:
+				for (k = 0; k < table->num_entries; k++) {
+					if ((table->mc_reg_table_entry[k].mclk_max == 125000) ||
+					    (table->mc_reg_table_entry[k].mclk_max == 137500))
+						table->mc_reg_table_entry[k].mc_data[i] =
+							(table->mc_reg_table_entry[k].mc_data[i] & 0xFFFF0F00) |
+							0x0000D0DD;
+				}
+				break;
+			case MC_SEQ_WR_CTL_D1:
+				for (k = 0; k < table->num_entries; k++) {
+					if ((table->mc_reg_table_entry[k].mclk_max == 125000) ||
+					    (table->mc_reg_table_entry[k].mclk_max == 137500))
+						table->mc_reg_table_entry[k].mc_data[i] =
+							(table->mc_reg_table_entry[k].mc_data[i] & 0xFFFF0F00) |
+							0x0000D0DD;
+				}
+				break;
+			case MC_SEQ_WR_CTL_2:
+				for (k = 0; k < table->num_entries; k++) {
+					if ((table->mc_reg_table_entry[k].mclk_max == 125000) ||
+					    (table->mc_reg_table_entry[k].mclk_max == 137500))
+						table->mc_reg_table_entry[k].mc_data[i] = 0;
+				}
+				break;
+			case MC_SEQ_CAS_TIMING:
+				for (k = 0; k < table->num_entries; k++) {
+					if (table->mc_reg_table_entry[k].mclk_max == 125000)
+						table->mc_reg_table_entry[k].mc_data[i] =
+							(table->mc_reg_table_entry[k].mc_data[i] & 0xFFE0FE0F) |
+							0x000C0140;
+					else if (table->mc_reg_table_entry[k].mclk_max == 137500)
+						table->mc_reg_table_entry[k].mc_data[i] =
+							(table->mc_reg_table_entry[k].mc_data[i] & 0xFFE0FE0F) |
+							0x000C0150;
+				}
+				break;
+			case MC_SEQ_MISC_TIMING:
+				for (k = 0; k < table->num_entries; k++) {
+					if (table->mc_reg_table_entry[k].mclk_max == 125000)
+						table->mc_reg_table_entry[k].mc_data[i] =
+							(table->mc_reg_table_entry[k].mc_data[i] & 0xFFFFFFE0) |
+							0x00000030;
+					else if (table->mc_reg_table_entry[k].mclk_max == 137500)
+						table->mc_reg_table_entry[k].mc_data[i] =
+							(table->mc_reg_table_entry[k].mc_data[i] & 0xFFFFFFE0) |
+							0x00000035;
+				}
+				break;
+			default:
+				break;
+			}
+		}
+
+		WREG32(MC_SEQ_IO_DEBUG_INDEX, 3);
+		tmp = RREG32(MC_SEQ_IO_DEBUG_DATA);
+		tmp = (tmp & 0xFFF8FFFF) | (1 << 16);
+		WREG32(MC_SEQ_IO_DEBUG_INDEX, 3);
+		WREG32(MC_SEQ_IO_DEBUG_DATA, tmp);
+	}
+
+	return 0;
+}
+
 static int ci_initialize_mc_reg_table(struct radeon_device *rdev)
 {
 	struct ci_power_info *pi = ci_get_pi(rdev);
@@ -4104,6 +4596,10 @@ static int ci_initialize_mc_reg_table(struct radeon_device *rdev)
 
 	ci_set_s0_mc_reg_index(ci_table);
 
+	ret = ci_register_patching_mc_seq(rdev, ci_table);
+	if (ret)
+		goto init_mc_done;
+
 	ret = ci_set_mc_special_registers(rdev, ci_table);
 	if (ret)
 		goto init_mc_done;
@@ -4700,36 +5196,51 @@ int ci_dpm_enable(struct radeon_device *rdev)
 		return ret;
 	}
 
+	ret = ci_power_control_set_level(rdev);
+	if (ret) {
+		DRM_ERROR("ci_power_control_set_level failed\n");
+		return ret;
+	}
+
 	ci_enable_auto_throttle_source(rdev, RADEON_DPM_AUTO_THROTTLE_SRC_THERMAL, true);
 
+	ret = ci_enable_thermal_based_sclk_dpm(rdev, true);
+	if (ret) {
+		DRM_ERROR("ci_enable_thermal_based_sclk_dpm failed\n");
+		return ret;
+	}
+
+	ci_thermal_start_thermal_controller(rdev);
+
 	ci_update_current_ps(rdev, boot_ps);
 
 	return 0;
 }
 
-int ci_dpm_late_enable(struct radeon_device *rdev)
+static int ci_set_temperature_range(struct radeon_device *rdev)
 {
 	int ret;
 
-	if (rdev->irq.installed &&
-	    r600_is_internal_thermal_sensor(rdev->pm.int_thermal_type)) {
-#if 0
-		PPSMC_Result result;
-#endif
-		ret = ci_set_thermal_temperature_range(rdev, R600_TEMP_RANGE_MIN, R600_TEMP_RANGE_MAX);
-		if (ret) {
-			DRM_ERROR("ci_set_thermal_temperature_range failed\n");
-			return ret;
-		}
-		rdev->irq.dpm_thermal = true;
-		radeon_irq_set(rdev);
-#if 0
-		result = ci_send_msg_to_smc(rdev, PPSMC_MSG_EnableThermalInterrupt);
+	ret = ci_thermal_enable_alert(rdev, false);
+	if (ret)
+		return ret;
+	ret = ci_thermal_set_temperature_range(rdev, R600_TEMP_RANGE_MIN, R600_TEMP_RANGE_MAX);
+	if (ret)
+		return ret;
+	ret = ci_thermal_enable_alert(rdev, true);
+	if (ret)
+		return ret;
 
-		if (result != PPSMC_Result_OK)
-			DRM_DEBUG_KMS("Could not enable thermal interrupts.\n");
-#endif
-	}
+	return ret;
+}
+
+int ci_dpm_late_enable(struct radeon_device *rdev)
+{
+	int ret;
+
+	ret = ci_set_temperature_range(rdev);
+	if (ret)
+		return ret;
 
 	ci_dpm_powergate_uvd(rdev, true);
 
@@ -4746,6 +5257,8 @@ void ci_dpm_disable(struct radeon_device *rdev)
 	if (!ci_is_smc_running(rdev))
 		return;
 
+	ci_thermal_stop_thermal_controller(rdev);
+
 	if (pi->thermal_protection)
 		ci_enable_thermal_protection(rdev, false);
 	ci_enable_power_containment(rdev, false);
@@ -4754,12 +5267,13 @@ void ci_dpm_disable(struct radeon_device *rdev)
 	ci_enable_spread_spectrum(rdev, false);
 	ci_enable_auto_throttle_source(rdev, RADEON_DPM_AUTO_THROTTLE_SRC_THERMAL, false);
 	ci_stop_dpm(rdev);
-	ci_enable_ds_master_switch(rdev, true);
+	ci_enable_ds_master_switch(rdev, false);
 	ci_enable_ulv(rdev, false);
 	ci_clear_vc(rdev);
 	ci_reset_to_default(rdev);
 	ci_dpm_stop_smc(rdev);
 	ci_force_switch_to_arb_f0(rdev);
+	ci_enable_thermal_based_sclk_dpm(rdev, false);
 
 	ci_update_current_ps(rdev, boot_ps);
 }
@@ -4829,11 +5343,6 @@ int ci_dpm_set_power_state(struct radeon_device *rdev)
 	return 0;
 }
 
-int ci_dpm_power_control_set_level(struct radeon_device *rdev)
-{
-	return ci_power_control_set_level(rdev);
-}
-
 void ci_dpm_reset_asic(struct radeon_device *rdev)
 {
 	ci_set_boot_state(rdev);
@@ -5093,6 +5602,8 @@ void ci_dpm_fini(struct radeon_device *rdev)
 int ci_dpm_init(struct radeon_device *rdev)
 {
 	int index = GetIndexIntoMasterTable(DATA, ASIC_InternalSS_Info);
+	SMU7_Discrete_DpmTable  *dpm_table;
+	struct radeon_gpio_rec gpio;
 	u16 data_offset, size;
 	u8 frev, crev;
 	struct ci_power_info *pi;
@@ -5162,6 +5673,7 @@ int ci_dpm_init(struct radeon_device *rdev)
 	pi->sclk_dpm_key_disabled = 0;
 	pi->mclk_dpm_key_disabled = 0;
 	pi->pcie_dpm_key_disabled = 0;
+	pi->thermal_sclk_dpm_enabled = 0;
 
 	/* mclk dpm is unstable on some R7 260X cards with the old mc ucode */
 	if ((rdev->pdev->device == 0x6658) &&
@@ -5226,6 +5738,55 @@ int ci_dpm_init(struct radeon_device *rdev)
 
 	pi->uvd_enabled = false;
 
+	dpm_table = &pi->smc_state_table;
+
+	gpio = radeon_atombios_lookup_gpio(rdev, VDDC_VRHOT_GPIO_PINID);
+	if (gpio.valid) {
+		dpm_table->VRHotGpio = gpio.shift;
+		rdev->pm.dpm.platform_caps |= ATOM_PP_PLATFORM_CAP_REGULATOR_HOT;
+	} else {
+		dpm_table->VRHotGpio = CISLANDS_UNUSED_GPIO_PIN;
+		rdev->pm.dpm.platform_caps &= ~ATOM_PP_PLATFORM_CAP_REGULATOR_HOT;
+	}
+
+	gpio = radeon_atombios_lookup_gpio(rdev, PP_AC_DC_SWITCH_GPIO_PINID);
+	if (gpio.valid) {
+		dpm_table->AcDcGpio = gpio.shift;
+		rdev->pm.dpm.platform_caps |= ATOM_PP_PLATFORM_CAP_HARDWAREDC;
+	} else {
+		dpm_table->AcDcGpio = CISLANDS_UNUSED_GPIO_PIN;
+		rdev->pm.dpm.platform_caps &= ~ATOM_PP_PLATFORM_CAP_HARDWAREDC;
+	}
+
+	gpio = radeon_atombios_lookup_gpio(rdev, VDDC_PCC_GPIO_PINID);
+	if (gpio.valid) {
+		u32 tmp = RREG32_SMC(CNB_PWRMGT_CNTL);
+
+		switch (gpio.shift) {
+		case 0:
+			tmp &= ~GNB_SLOW_MODE_MASK;
+			tmp |= GNB_SLOW_MODE(1);
+			break;
+		case 1:
+			tmp &= ~GNB_SLOW_MODE_MASK;
+			tmp |= GNB_SLOW_MODE(2);
+			break;
+		case 2:
+			tmp |= GNB_SLOW;
+			break;
+		case 3:
+			tmp |= FORCE_NB_PS1;
+			break;
+		case 4:
+			tmp |= DPM_ENABLED;
+			break;
+		default:
+			DRM_ERROR("Invalid PCC GPIO: %u!\n", gpio.shift);
+			break;
+		}
+		WREG32_SMC(CNB_PWRMGT_CNTL, tmp);
+	}
+
 	pi->voltage_control = CISLANDS_VOLTAGE_CONTROL_NONE;
 	pi->vddci_control = CISLANDS_VOLTAGE_CONTROL_NONE;
 	pi->mvdd_control = CISLANDS_VOLTAGE_CONTROL_NONE;
@@ -5287,15 +5848,21 @@ int ci_dpm_init(struct radeon_device *rdev)
 		rdev->pm.dpm.dyn_state.max_clock_voltage_on_dc =
 			rdev->pm.dpm.dyn_state.max_clock_voltage_on_ac;
 
+	pi->fan_ctrl_is_in_default_mode = true;
+
 	return 0;
 }
 
 void ci_dpm_debugfs_print_current_performance_level(struct radeon_device *rdev,
 						    struct seq_file *m)
 {
+	struct ci_power_info *pi = ci_get_pi(rdev);
+	struct radeon_ps *rps = &pi->current_rps;
 	u32 sclk = ci_get_average_sclk_freq(rdev);
 	u32 mclk = ci_get_average_mclk_freq(rdev);
 
+	seq_printf(m, "uvd    %sabled\n", pi->uvd_enabled ? "en" : "dis");
+	seq_printf(m, "vce    %sabled\n", rps->vce_active ? "en" : "dis");
 	seq_printf(m, "power level avg    sclk: %u mclk: %u\n",
 		   sclk, mclk);
 }
diff --git a/drivers/gpu/drm/radeon/ci_dpm.h b/drivers/gpu/drm/radeon/ci_dpm.h
index 93bbed977ffb..84e3d3bcf9f3 100644
--- a/drivers/gpu/drm/radeon/ci_dpm.h
+++ b/drivers/gpu/drm/radeon/ci_dpm.h
@@ -33,6 +33,8 @@
 
 #define CISLANDS_MAX_HARDWARE_POWERLEVELS 2
 
+#define CISLANDS_UNUSED_GPIO_PIN 0x7F
+
 struct ci_pl {
 	u32 mclk;
 	u32 sclk;
@@ -237,6 +239,7 @@ struct ci_power_info {
 	u32 sclk_dpm_key_disabled;
 	u32 mclk_dpm_key_disabled;
 	u32 pcie_dpm_key_disabled;
+	u32 thermal_sclk_dpm_enabled;
 	struct ci_pcie_perf_range pcie_gen_performance;
 	struct ci_pcie_perf_range pcie_lane_performance;
 	struct ci_pcie_perf_range pcie_gen_powersaving;
@@ -264,6 +267,7 @@ struct ci_power_info {
 	bool caps_automatic_dc_transition;
 	bool caps_sclk_throttle_low_notification;
 	bool caps_dynamic_ac_timing;
+	bool caps_od_fuzzy_fan_control_support;
 	/* flags */
 	bool thermal_protection;
 	bool pcie_performance_request;
@@ -285,6 +289,10 @@ struct ci_power_info {
 	struct ci_ps current_ps;
 	struct radeon_ps requested_rps;
 	struct ci_ps requested_ps;
+	/* fan control */
+	bool fan_ctrl_is_in_default_mode;
+	u32 t_min;
+	u32 fan_ctrl_default_mode;
 };
 
 #define CISLANDS_VOLTAGE_CONTROL_NONE                   0x0
diff --git a/drivers/gpu/drm/radeon/ci_smc.c b/drivers/gpu/drm/radeon/ci_smc.c
index b630edc2fd0c..e78bcad7a43e 100644
--- a/drivers/gpu/drm/radeon/ci_smc.c
+++ b/drivers/gpu/drm/radeon/ci_smc.c
@@ -129,7 +129,7 @@ void ci_reset_smc(struct radeon_device *rdev)
 
 int ci_program_jump_on_start(struct radeon_device *rdev)
 {
-	static u8 data[] = { 0xE0, 0x00, 0x80, 0x40 };
+	static const u8 data[] = { 0xE0, 0x00, 0x80, 0x40 };
 
 	return ci_copy_bytes_to_smc(rdev, 0x0, data, 4, sizeof(data)+1);
 }
diff --git a/drivers/gpu/drm/radeon/cik.c b/drivers/gpu/drm/radeon/cik.c
index 3d546c606b43..6dcde3798b45 100644
--- a/drivers/gpu/drm/radeon/cik.c
+++ b/drivers/gpu/drm/radeon/cik.c
@@ -32,6 +32,7 @@
 #include "cik_blit_shaders.h"
 #include "radeon_ucode.h"
 #include "clearstate_ci.h"
+#include "radeon_kfd.h"
 
 MODULE_FIRMWARE("radeon/BONAIRE_pfp.bin");
 MODULE_FIRMWARE("radeon/BONAIRE_me.bin");
@@ -1563,6 +1564,8 @@ static const u32 godavari_golden_registers[] =
 
 static void cik_init_golden_registers(struct radeon_device *rdev)
 {
+	/* Some of the registers might be dependent on GRBM_GFX_INDEX */
+	mutex_lock(&rdev->grbm_idx_mutex);
 	switch (rdev->family) {
 	case CHIP_BONAIRE:
 		radeon_program_register_sequence(rdev,
@@ -1637,6 +1640,7 @@ static void cik_init_golden_registers(struct radeon_device *rdev)
 	default:
 		break;
 	}
+	mutex_unlock(&rdev->grbm_idx_mutex);
 }
 
 /**
@@ -1806,7 +1810,7 @@ int ci_mc_load_microcode(struct radeon_device *rdev)
 {
 	const __be32 *fw_data = NULL;
 	const __le32 *new_fw_data = NULL;
-	u32 running, blackout = 0;
+	u32 running, blackout = 0, tmp;
 	u32 *io_mc_regs = NULL;
 	const __le32 *new_io_mc_regs = NULL;
 	int i, regs_size, ucode_size;
@@ -1866,6 +1870,15 @@ int ci_mc_load_microcode(struct radeon_device *rdev)
 				WREG32(MC_SEQ_IO_DEBUG_DATA, io_mc_regs[(i << 1) + 1]);
 			}
 		}
+
+		tmp = RREG32(MC_SEQ_MISC0);
+		if ((rdev->pdev->device == 0x6649) && ((tmp & 0xff00) == 0x5600)) {
+			WREG32(MC_SEQ_IO_DEBUG_INDEX, 5);
+			WREG32(MC_SEQ_IO_DEBUG_DATA, 0x00000023);
+			WREG32(MC_SEQ_IO_DEBUG_INDEX, 9);
+			WREG32(MC_SEQ_IO_DEBUG_DATA, 0x000001f0);
+		}
+
 		/* load the MC ucode */
 		for (i = 0; i < ucode_size; i++) {
 			if (rdev->new_fw)
@@ -3419,6 +3432,7 @@ static void cik_setup_rb(struct radeon_device *rdev,
 	u32 disabled_rbs = 0;
 	u32 enabled_rbs = 0;
 
+	mutex_lock(&rdev->grbm_idx_mutex);
 	for (i = 0; i < se_num; i++) {
 		for (j = 0; j < sh_per_se; j++) {
 			cik_select_se_sh(rdev, i, j);
@@ -3430,6 +3444,7 @@ static void cik_setup_rb(struct radeon_device *rdev,
 		}
 	}
 	cik_select_se_sh(rdev, 0xffffffff, 0xffffffff);
+	mutex_unlock(&rdev->grbm_idx_mutex);
 
 	mask = 1;
 	for (i = 0; i < max_rb_num_per_se * se_num; i++) {
@@ -3440,6 +3455,7 @@ static void cik_setup_rb(struct radeon_device *rdev,
 
 	rdev->config.cik.backend_enable_mask = enabled_rbs;
 
+	mutex_lock(&rdev->grbm_idx_mutex);
 	for (i = 0; i < se_num; i++) {
 		cik_select_se_sh(rdev, i, 0xffffffff);
 		data = 0;
@@ -3467,6 +3483,7 @@ static void cik_setup_rb(struct radeon_device *rdev,
 		WREG32(PA_SC_RASTER_CONFIG, data);
 	}
 	cik_select_se_sh(rdev, 0xffffffff, 0xffffffff);
+	mutex_unlock(&rdev->grbm_idx_mutex);
 }
 
 /**
@@ -3684,6 +3701,12 @@ static void cik_gpu_init(struct radeon_device *rdev)
 	/* set HW defaults for 3D engine */
 	WREG32(CP_MEQ_THRESHOLDS, MEQ1_START(0x30) | MEQ2_START(0x60));
 
+	mutex_lock(&rdev->grbm_idx_mutex);
+	/*
+	 * making sure that the following register writes will be broadcasted
+	 * to all the shaders
+	 */
+	cik_select_se_sh(rdev, 0xffffffff, 0xffffffff);
 	WREG32(SX_DEBUG_1, 0x20);
 
 	WREG32(TA_CNTL_AUX, 0x00010000);
@@ -3739,6 +3762,7 @@ static void cik_gpu_init(struct radeon_device *rdev)
 
 	WREG32(PA_CL_ENHANCE, CLIP_VTX_REORDER_ENA | NUM_CLIP_SEQ(3));
 	WREG32(PA_SC_ENHANCE, ENABLE_PA_SC_OUT_OF_ORDER);
+	mutex_unlock(&rdev->grbm_idx_mutex);
 
 	udelay(50);
 }
@@ -3959,41 +3983,38 @@ bool cik_semaphore_ring_emit(struct radeon_device *rdev,
  * @src_offset: src GPU address
  * @dst_offset: dst GPU address
  * @num_gpu_pages: number of GPU pages to xfer
- * @fence: radeon fence object
+ * @resv: reservation object to sync to
  *
  * Copy GPU paging using the CP DMA engine (CIK+).
  * Used by the radeon ttm implementation to move pages if
  * registered as the asic copy callback.
  */
-int cik_copy_cpdma(struct radeon_device *rdev,
-		   uint64_t src_offset, uint64_t dst_offset,
-		   unsigned num_gpu_pages,
-		   struct radeon_fence **fence)
+struct radeon_fence *cik_copy_cpdma(struct radeon_device *rdev,
+				    uint64_t src_offset, uint64_t dst_offset,
+				    unsigned num_gpu_pages,
+				    struct reservation_object *resv)
 {
-	struct radeon_semaphore *sem = NULL;
+	struct radeon_fence *fence;
+	struct radeon_sync sync;
 	int ring_index = rdev->asic->copy.blit_ring_index;
 	struct radeon_ring *ring = &rdev->ring[ring_index];
 	u32 size_in_bytes, cur_size_in_bytes, control;
 	int i, num_loops;
 	int r = 0;
 
-	r = radeon_semaphore_create(rdev, &sem);
-	if (r) {
-		DRM_ERROR("radeon: moving bo (%d).\n", r);
-		return r;
-	}
+	radeon_sync_create(&sync);
 
 	size_in_bytes = (num_gpu_pages << RADEON_GPU_PAGE_SHIFT);
 	num_loops = DIV_ROUND_UP(size_in_bytes, 0x1fffff);
 	r = radeon_ring_lock(rdev, ring, num_loops * 7 + 18);
 	if (r) {
 		DRM_ERROR("radeon: moving bo (%d).\n", r);
-		radeon_semaphore_free(rdev, &sem, NULL);
-		return r;
+		radeon_sync_free(rdev, &sync, NULL);
+		return ERR_PTR(r);
 	}
 
-	radeon_semaphore_sync_to(sem, *fence);
-	radeon_semaphore_sync_rings(rdev, sem, ring->idx);
+	radeon_sync_resv(rdev, &sync, resv, false);
+	radeon_sync_rings(rdev, &sync, ring->idx);
 
 	for (i = 0; i < num_loops; i++) {
 		cur_size_in_bytes = size_in_bytes;
@@ -4014,17 +4035,17 @@ int cik_copy_cpdma(struct radeon_device *rdev,
 		dst_offset += cur_size_in_bytes;
 	}
 
-	r = radeon_fence_emit(rdev, fence, ring->idx);
+	r = radeon_fence_emit(rdev, &fence, ring->idx);
 	if (r) {
 		radeon_ring_unlock_undo(rdev, ring);
-		radeon_semaphore_free(rdev, &sem, NULL);
-		return r;
+		radeon_sync_free(rdev, &sync, NULL);
+		return ERR_PTR(r);
 	}
 
 	radeon_ring_unlock_commit(rdev, ring, false);
-	radeon_semaphore_free(rdev, &sem, *fence);
+	radeon_sync_free(rdev, &sync, fence);
 
-	return r;
+	return fence;
 }
 
 /*
@@ -4045,6 +4066,7 @@ int cik_copy_cpdma(struct radeon_device *rdev,
 void cik_ring_ib_execute(struct radeon_device *rdev, struct radeon_ib *ib)
 {
 	struct radeon_ring *ring = &rdev->ring[ib->ring];
+	unsigned vm_id = ib->vm ? ib->vm->ids[ib->ring].id : 0;
 	u32 header, control = INDIRECT_BUFFER_VALID;
 
 	if (ib->is_const_ib) {
@@ -4073,8 +4095,7 @@ void cik_ring_ib_execute(struct radeon_device *rdev, struct radeon_ib *ib)
 		header = PACKET3(PACKET3_INDIRECT_BUFFER, 2);
 	}
 
-	control |= ib->length_dw |
-		(ib->vm ? (ib->vm->id << 24) : 0);
+	control |= ib->length_dw | (vm_id << 24);
 
 	radeon_ring_write(ring, header);
 	radeon_ring_write(ring,
@@ -4234,7 +4255,7 @@ static int cik_cp_gfx_load_microcode(struct radeon_device *rdev)
 		WREG32(CP_PFP_UCODE_ADDR, 0);
 		for (i = 0; i < fw_size; i++)
 			WREG32(CP_PFP_UCODE_DATA, le32_to_cpup(fw_data++));
-		WREG32(CP_PFP_UCODE_ADDR, 0);
+		WREG32(CP_PFP_UCODE_ADDR, le32_to_cpu(pfp_hdr->header.ucode_version));
 
 		/* CE */
 		fw_data = (const __le32 *)
@@ -4243,7 +4264,7 @@ static int cik_cp_gfx_load_microcode(struct radeon_device *rdev)
 		WREG32(CP_CE_UCODE_ADDR, 0);
 		for (i = 0; i < fw_size; i++)
 			WREG32(CP_CE_UCODE_DATA, le32_to_cpup(fw_data++));
-		WREG32(CP_CE_UCODE_ADDR, 0);
+		WREG32(CP_CE_UCODE_ADDR, le32_to_cpu(ce_hdr->header.ucode_version));
 
 		/* ME */
 		fw_data = (const __be32 *)
@@ -4252,7 +4273,8 @@ static int cik_cp_gfx_load_microcode(struct radeon_device *rdev)
 		WREG32(CP_ME_RAM_WADDR, 0);
 		for (i = 0; i < fw_size; i++)
 			WREG32(CP_ME_RAM_DATA, le32_to_cpup(fw_data++));
-		WREG32(CP_ME_RAM_WADDR, 0);
+		WREG32(CP_ME_RAM_WADDR, le32_to_cpu(me_hdr->header.ucode_version));
+		WREG32(CP_ME_RAM_RADDR, le32_to_cpu(me_hdr->header.ucode_version));
 	} else {
 		const __be32 *fw_data;
 
@@ -4278,10 +4300,6 @@ static int cik_cp_gfx_load_microcode(struct radeon_device *rdev)
 		WREG32(CP_ME_RAM_WADDR, 0);
 	}
 
-	WREG32(CP_PFP_UCODE_ADDR, 0);
-	WREG32(CP_CE_UCODE_ADDR, 0);
-	WREG32(CP_ME_RAM_WADDR, 0);
-	WREG32(CP_ME_RAM_RADDR, 0);
 	return 0;
 }
 
@@ -4315,8 +4333,8 @@ static int cik_cp_gfx_start(struct radeon_device *rdev)
 	/* init the CE partitions.  CE only used for gfx on CIK */
 	radeon_ring_write(ring, PACKET3(PACKET3_SET_BASE, 2));
 	radeon_ring_write(ring, PACKET3_BASE_INDEX(CE_PARTITION_BASE));
-	radeon_ring_write(ring, 0xc000);
-	radeon_ring_write(ring, 0xc000);
+	radeon_ring_write(ring, 0x8000);
+	radeon_ring_write(ring, 0x8000);
 
 	/* setup clear context state */
 	radeon_ring_write(ring, PACKET3(PACKET3_PREAMBLE_CNTL, 0));
@@ -4563,7 +4581,7 @@ static int cik_cp_compute_load_microcode(struct radeon_device *rdev)
 		WREG32(CP_MEC_ME1_UCODE_ADDR, 0);
 		for (i = 0; i < fw_size; i++)
 			WREG32(CP_MEC_ME1_UCODE_DATA, le32_to_cpup(fw_data++));
-		WREG32(CP_MEC_ME1_UCODE_ADDR, 0);
+		WREG32(CP_MEC_ME1_UCODE_ADDR, le32_to_cpu(mec_hdr->header.ucode_version));
 
 		/* MEC2 */
 		if (rdev->family == CHIP_KAVERI) {
@@ -4577,7 +4595,7 @@ static int cik_cp_compute_load_microcode(struct radeon_device *rdev)
 			WREG32(CP_MEC_ME2_UCODE_ADDR, 0);
 			for (i = 0; i < fw_size; i++)
 				WREG32(CP_MEC_ME2_UCODE_DATA, le32_to_cpup(fw_data++));
-			WREG32(CP_MEC_ME2_UCODE_ADDR, 0);
+			WREG32(CP_MEC_ME2_UCODE_ADDR, le32_to_cpu(mec2_hdr->header.ucode_version));
 		}
 	} else {
 		const __be32 *fw_data;
@@ -4677,19 +4695,18 @@ static int cik_mec_init(struct radeon_device *rdev)
 	/*
 	 * KV:    2 MEC, 4 Pipes/MEC, 8 Queues/Pipe - 64 Queues total
 	 * CI/KB: 1 MEC, 4 Pipes/MEC, 8 Queues/Pipe - 32 Queues total
+	 * Nonetheless, we assign only 1 pipe because all other pipes will
+	 * be handled by KFD
 	 */
-	if (rdev->family == CHIP_KAVERI)
-		rdev->mec.num_mec = 2;
-	else
-		rdev->mec.num_mec = 1;
-	rdev->mec.num_pipe = 4;
+	rdev->mec.num_mec = 1;
+	rdev->mec.num_pipe = 1;
 	rdev->mec.num_queue = rdev->mec.num_mec * rdev->mec.num_pipe * 8;
 
 	if (rdev->mec.hpd_eop_obj == NULL) {
 		r = radeon_bo_create(rdev,
 				     rdev->mec.num_mec *rdev->mec.num_pipe * MEC_HPD_SIZE * 2,
 				     PAGE_SIZE, true,
-				     RADEON_GEM_DOMAIN_GTT, 0, NULL,
+				     RADEON_GEM_DOMAIN_GTT, 0, NULL, NULL,
 				     &rdev->mec.hpd_eop_obj);
 		if (r) {
 			dev_warn(rdev->dev, "(%d) create HDP EOP bo failed\n", r);
@@ -4824,28 +4841,24 @@ static int cik_cp_compute_resume(struct radeon_device *rdev)
 
 	/* init the pipes */
 	mutex_lock(&rdev->srbm_mutex);
-	for (i = 0; i < (rdev->mec.num_pipe * rdev->mec.num_mec); i++) {
-		int me = (i < 4) ? 1 : 2;
-		int pipe = (i < 4) ? i : (i - 4);
 
-		eop_gpu_addr = rdev->mec.hpd_eop_gpu_addr + (i * MEC_HPD_SIZE * 2);
+	eop_gpu_addr = rdev->mec.hpd_eop_gpu_addr;
 
-		cik_srbm_select(rdev, me, pipe, 0, 0);
+	cik_srbm_select(rdev, 0, 0, 0, 0);
 
-		/* write the EOP addr */
-		WREG32(CP_HPD_EOP_BASE_ADDR, eop_gpu_addr >> 8);
-		WREG32(CP_HPD_EOP_BASE_ADDR_HI, upper_32_bits(eop_gpu_addr) >> 8);
+	/* write the EOP addr */
+	WREG32(CP_HPD_EOP_BASE_ADDR, eop_gpu_addr >> 8);
+	WREG32(CP_HPD_EOP_BASE_ADDR_HI, upper_32_bits(eop_gpu_addr) >> 8);
 
-		/* set the VMID assigned */
-		WREG32(CP_HPD_EOP_VMID, 0);
+	/* set the VMID assigned */
+	WREG32(CP_HPD_EOP_VMID, 0);
+
+	/* set the EOP size, register value is 2^(EOP_SIZE+1) dwords */
+	tmp = RREG32(CP_HPD_EOP_CONTROL);
+	tmp &= ~EOP_SIZE_MASK;
+	tmp |= order_base_2(MEC_HPD_SIZE / 8);
+	WREG32(CP_HPD_EOP_CONTROL, tmp);
 
-		/* set the EOP size, register value is 2^(EOP_SIZE+1) dwords */
-		tmp = RREG32(CP_HPD_EOP_CONTROL);
-		tmp &= ~EOP_SIZE_MASK;
-		tmp |= order_base_2(MEC_HPD_SIZE / 8);
-		WREG32(CP_HPD_EOP_CONTROL, tmp);
-	}
-	cik_srbm_select(rdev, 0, 0, 0, 0);
 	mutex_unlock(&rdev->srbm_mutex);
 
 	/* init the queues.  Just two for now. */
@@ -4860,7 +4873,7 @@ static int cik_cp_compute_resume(struct radeon_device *rdev)
 					     sizeof(struct bonaire_mqd),
 					     PAGE_SIZE, true,
 					     RADEON_GEM_DOMAIN_GTT, 0, NULL,
-					     &rdev->ring[idx].mqd_obj);
+					     NULL, &rdev->ring[idx].mqd_obj);
 			if (r) {
 				dev_warn(rdev->dev, "(%d) create MQD bo failed\n", r);
 				return r;
@@ -5899,8 +5912,13 @@ int cik_ib_parse(struct radeon_device *rdev, struct radeon_ib *ib)
  */
 int cik_vm_init(struct radeon_device *rdev)
 {
-	/* number of VMs */
-	rdev->vm_manager.nvm = 16;
+	/*
+	 * number of VMs
+	 * VMID 0 is reserved for System
+	 * radeon graphics/compute will use VMIDs 1-7
+	 * amdkfd will use VMIDs 8-15
+	 */
+	rdev->vm_manager.nvm = RADEON_NUM_OF_VMIDS;
 	/* base offset of vram pages */
 	if (rdev->flags & RADEON_IS_IGP) {
 		u64 tmp = RREG32(MC_VM_FB_OFFSET);
@@ -5960,26 +5978,23 @@ static void cik_vm_decode_fault(struct radeon_device *rdev,
  * Update the page table base and flush the VM TLB
  * using the CP (CIK).
  */
-void cik_vm_flush(struct radeon_device *rdev, int ridx, struct radeon_vm *vm)
+void cik_vm_flush(struct radeon_device *rdev, struct radeon_ring *ring,
+		  unsigned vm_id, uint64_t pd_addr)
 {
-	struct radeon_ring *ring = &rdev->ring[ridx];
-	int usepfp = (ridx == RADEON_RING_TYPE_GFX_INDEX);
-
-	if (vm == NULL)
-		return;
+	int usepfp = (ring->idx == RADEON_RING_TYPE_GFX_INDEX);
 
 	radeon_ring_write(ring, PACKET3(PACKET3_WRITE_DATA, 3));
 	radeon_ring_write(ring, (WRITE_DATA_ENGINE_SEL(usepfp) |
 				 WRITE_DATA_DST_SEL(0)));
-	if (vm->id < 8) {
+	if (vm_id < 8) {
 		radeon_ring_write(ring,
-				  (VM_CONTEXT0_PAGE_TABLE_BASE_ADDR + (vm->id << 2)) >> 2);
+				  (VM_CONTEXT0_PAGE_TABLE_BASE_ADDR + (vm_id << 2)) >> 2);
 	} else {
 		radeon_ring_write(ring,
-				  (VM_CONTEXT8_PAGE_TABLE_BASE_ADDR + ((vm->id - 8) << 2)) >> 2);
+				  (VM_CONTEXT8_PAGE_TABLE_BASE_ADDR + ((vm_id - 8) << 2)) >> 2);
 	}
 	radeon_ring_write(ring, 0);
-	radeon_ring_write(ring, vm->pd_gpu_addr >> 12);
+	radeon_ring_write(ring, pd_addr >> 12);
 
 	/* update SH_MEM_* regs */
 	radeon_ring_write(ring, PACKET3(PACKET3_WRITE_DATA, 3));
@@ -5987,7 +6002,7 @@ void cik_vm_flush(struct radeon_device *rdev, int ridx, struct radeon_vm *vm)
 				 WRITE_DATA_DST_SEL(0)));
 	radeon_ring_write(ring, SRBM_GFX_CNTL >> 2);
 	radeon_ring_write(ring, 0);
-	radeon_ring_write(ring, VMID(vm->id));
+	radeon_ring_write(ring, VMID(vm_id));
 
 	radeon_ring_write(ring, PACKET3(PACKET3_WRITE_DATA, 6));
 	radeon_ring_write(ring, (WRITE_DATA_ENGINE_SEL(usepfp) |
@@ -6008,7 +6023,7 @@ void cik_vm_flush(struct radeon_device *rdev, int ridx, struct radeon_vm *vm)
 	radeon_ring_write(ring, VMID(0));
 
 	/* HDP flush */
-	cik_hdp_flush_cp_ring_emit(rdev, ridx);
+	cik_hdp_flush_cp_ring_emit(rdev, ring->idx);
 
 	/* bits 0-15 are the VM contexts0-15 */
 	radeon_ring_write(ring, PACKET3(PACKET3_WRITE_DATA, 3));
@@ -6016,7 +6031,7 @@ void cik_vm_flush(struct radeon_device *rdev, int ridx, struct radeon_vm *vm)
 				 WRITE_DATA_DST_SEL(0)));
 	radeon_ring_write(ring, VM_INVALIDATE_REQUEST >> 2);
 	radeon_ring_write(ring, 0);
-	radeon_ring_write(ring, 1 << vm->id);
+	radeon_ring_write(ring, 1 << vm_id);
 
 	/* compute doesn't have PFP */
 	if (usepfp) {
@@ -6061,6 +6076,7 @@ static void cik_wait_for_rlc_serdes(struct radeon_device *rdev)
 	u32 i, j, k;
 	u32 mask;
 
+	mutex_lock(&rdev->grbm_idx_mutex);
 	for (i = 0; i < rdev->config.cik.max_shader_engines; i++) {
 		for (j = 0; j < rdev->config.cik.max_sh_per_se; j++) {
 			cik_select_se_sh(rdev, i, j);
@@ -6072,6 +6088,7 @@ static void cik_wait_for_rlc_serdes(struct radeon_device *rdev)
 		}
 	}
 	cik_select_se_sh(rdev, 0xffffffff, 0xffffffff);
+	mutex_unlock(&rdev->grbm_idx_mutex);
 
 	mask = SE_MASTER_BUSY_MASK | GC_MASTER_BUSY | TC0_MASTER_BUSY | TC1_MASTER_BUSY;
 	for (k = 0; k < rdev->usec_timeout; k++) {
@@ -6206,10 +6223,12 @@ static int cik_rlc_resume(struct radeon_device *rdev)
 	WREG32(RLC_LB_CNTR_INIT, 0);
 	WREG32(RLC_LB_CNTR_MAX, 0x00008000);
 
+	mutex_lock(&rdev->grbm_idx_mutex);
 	cik_select_se_sh(rdev, 0xffffffff, 0xffffffff);
 	WREG32(RLC_LB_INIT_CU_MASK, 0xffffffff);
 	WREG32(RLC_LB_PARAMS, 0x00600408);
 	WREG32(RLC_LB_CNTL, 0x80000004);
+	mutex_unlock(&rdev->grbm_idx_mutex);
 
 	WREG32(RLC_MC_CNTL, 0);
 	WREG32(RLC_UCODE_CNTL, 0);
@@ -6226,7 +6245,7 @@ static int cik_rlc_resume(struct radeon_device *rdev)
 		WREG32(RLC_GPM_UCODE_ADDR, 0);
 		for (i = 0; i < size; i++)
 			WREG32(RLC_GPM_UCODE_DATA, le32_to_cpup(fw_data++));
-		WREG32(RLC_GPM_UCODE_ADDR, 0);
+		WREG32(RLC_GPM_UCODE_ADDR, le32_to_cpu(hdr->header.ucode_version));
 	} else {
 		const __be32 *fw_data;
 
@@ -6276,11 +6295,13 @@ static void cik_enable_cgcg(struct radeon_device *rdev, bool enable)
 
 		tmp = cik_halt_rlc(rdev);
 
+		mutex_lock(&rdev->grbm_idx_mutex);
 		cik_select_se_sh(rdev, 0xffffffff, 0xffffffff);
 		WREG32(RLC_SERDES_WR_CU_MASTER_MASK, 0xffffffff);
 		WREG32(RLC_SERDES_WR_NONCU_MASTER_MASK, 0xffffffff);
 		tmp2 = BPM_ADDR_MASK | CGCG_OVERRIDE_0 | CGLS_ENABLE;
 		WREG32(RLC_SERDES_WR_CTRL, tmp2);
+		mutex_unlock(&rdev->grbm_idx_mutex);
 
 		cik_update_rlc(rdev, tmp);
 
@@ -6316,17 +6337,20 @@ static void cik_enable_mgcg(struct radeon_device *rdev, bool enable)
 		}
 
 		orig = data = RREG32(RLC_CGTT_MGCG_OVERRIDE);
+		data |= 0x00000001;
 		data &= 0xfffffffd;
 		if (orig != data)
 			WREG32(RLC_CGTT_MGCG_OVERRIDE, data);
 
 		tmp = cik_halt_rlc(rdev);
 
+		mutex_lock(&rdev->grbm_idx_mutex);
 		cik_select_se_sh(rdev, 0xffffffff, 0xffffffff);
 		WREG32(RLC_SERDES_WR_CU_MASTER_MASK, 0xffffffff);
 		WREG32(RLC_SERDES_WR_NONCU_MASTER_MASK, 0xffffffff);
 		data = BPM_ADDR_MASK | MGCG_OVERRIDE_0;
 		WREG32(RLC_SERDES_WR_CTRL, data);
+		mutex_unlock(&rdev->grbm_idx_mutex);
 
 		cik_update_rlc(rdev, tmp);
 
@@ -6347,7 +6371,7 @@ static void cik_enable_mgcg(struct radeon_device *rdev, bool enable)
 		}
 	} else {
 		orig = data = RREG32(RLC_CGTT_MGCG_OVERRIDE);
-		data |= 0x00000002;
+		data |= 0x00000003;
 		if (orig != data)
 			WREG32(RLC_CGTT_MGCG_OVERRIDE, data);
 
@@ -6370,11 +6394,13 @@ static void cik_enable_mgcg(struct radeon_device *rdev, bool enable)
 
 		tmp = cik_halt_rlc(rdev);
 
+		mutex_lock(&rdev->grbm_idx_mutex);
 		cik_select_se_sh(rdev, 0xffffffff, 0xffffffff);
 		WREG32(RLC_SERDES_WR_CU_MASTER_MASK, 0xffffffff);
 		WREG32(RLC_SERDES_WR_NONCU_MASTER_MASK, 0xffffffff);
 		data = BPM_ADDR_MASK | MGCG_OVERRIDE_1;
 		WREG32(RLC_SERDES_WR_CTRL, data);
+		mutex_unlock(&rdev->grbm_idx_mutex);
 
 		cik_update_rlc(rdev, tmp);
 	}
@@ -6803,10 +6829,12 @@ static u32 cik_get_cu_active_bitmap(struct radeon_device *rdev, u32 se, u32 sh)
 	u32 mask = 0, tmp, tmp1;
 	int i;
 
+	mutex_lock(&rdev->grbm_idx_mutex);
 	cik_select_se_sh(rdev, se, sh);
 	tmp = RREG32(CC_GC_SHADER_ARRAY_CONFIG);
 	tmp1 = RREG32(GC_USER_SHADER_ARRAY_CONFIG);
 	cik_select_se_sh(rdev, 0xffffffff, 0xffffffff);
+	mutex_unlock(&rdev->grbm_idx_mutex);
 
 	tmp &= 0xffff0000;
 
@@ -7290,8 +7318,7 @@ static int cik_irq_init(struct radeon_device *rdev)
 int cik_irq_set(struct radeon_device *rdev)
 {
 	u32 cp_int_cntl;
-	u32 cp_m1p0, cp_m1p1, cp_m1p2, cp_m1p3;
-	u32 cp_m2p0, cp_m2p1, cp_m2p2, cp_m2p3;
+	u32 cp_m1p0;
 	u32 crtc1 = 0, crtc2 = 0, crtc3 = 0, crtc4 = 0, crtc5 = 0, crtc6 = 0;
 	u32 hpd1, hpd2, hpd3, hpd4, hpd5, hpd6;
 	u32 grbm_int_cntl = 0;
@@ -7325,13 +7352,6 @@ int cik_irq_set(struct radeon_device *rdev)
 	dma_cntl1 = RREG32(SDMA0_CNTL + SDMA1_REGISTER_OFFSET) & ~TRAP_ENABLE;
 
 	cp_m1p0 = RREG32(CP_ME1_PIPE0_INT_CNTL) & ~TIME_STAMP_INT_ENABLE;
-	cp_m1p1 = RREG32(CP_ME1_PIPE1_INT_CNTL) & ~TIME_STAMP_INT_ENABLE;
-	cp_m1p2 = RREG32(CP_ME1_PIPE2_INT_CNTL) & ~TIME_STAMP_INT_ENABLE;
-	cp_m1p3 = RREG32(CP_ME1_PIPE3_INT_CNTL) & ~TIME_STAMP_INT_ENABLE;
-	cp_m2p0 = RREG32(CP_ME2_PIPE0_INT_CNTL) & ~TIME_STAMP_INT_ENABLE;
-	cp_m2p1 = RREG32(CP_ME2_PIPE1_INT_CNTL) & ~TIME_STAMP_INT_ENABLE;
-	cp_m2p2 = RREG32(CP_ME2_PIPE2_INT_CNTL) & ~TIME_STAMP_INT_ENABLE;
-	cp_m2p3 = RREG32(CP_ME2_PIPE3_INT_CNTL) & ~TIME_STAMP_INT_ENABLE;
 
 	if (rdev->flags & RADEON_IS_IGP)
 		thermal_int = RREG32_SMC(CG_THERMAL_INT_CTRL) &
@@ -7353,33 +7373,6 @@ int cik_irq_set(struct radeon_device *rdev)
 			case 0:
 				cp_m1p0 |= TIME_STAMP_INT_ENABLE;
 				break;
-			case 1:
-				cp_m1p1 |= TIME_STAMP_INT_ENABLE;
-				break;
-			case 2:
-				cp_m1p2 |= TIME_STAMP_INT_ENABLE;
-				break;
-			case 3:
-				cp_m1p2 |= TIME_STAMP_INT_ENABLE;
-				break;
-			default:
-				DRM_DEBUG("si_irq_set: sw int cp1 invalid pipe %d\n", ring->pipe);
-				break;
-			}
-		} else if (ring->me == 2) {
-			switch (ring->pipe) {
-			case 0:
-				cp_m2p0 |= TIME_STAMP_INT_ENABLE;
-				break;
-			case 1:
-				cp_m2p1 |= TIME_STAMP_INT_ENABLE;
-				break;
-			case 2:
-				cp_m2p2 |= TIME_STAMP_INT_ENABLE;
-				break;
-			case 3:
-				cp_m2p2 |= TIME_STAMP_INT_ENABLE;
-				break;
 			default:
 				DRM_DEBUG("si_irq_set: sw int cp1 invalid pipe %d\n", ring->pipe);
 				break;
@@ -7396,33 +7389,6 @@ int cik_irq_set(struct radeon_device *rdev)
 			case 0:
 				cp_m1p0 |= TIME_STAMP_INT_ENABLE;
 				break;
-			case 1:
-				cp_m1p1 |= TIME_STAMP_INT_ENABLE;
-				break;
-			case 2:
-				cp_m1p2 |= TIME_STAMP_INT_ENABLE;
-				break;
-			case 3:
-				cp_m1p2 |= TIME_STAMP_INT_ENABLE;
-				break;
-			default:
-				DRM_DEBUG("si_irq_set: sw int cp2 invalid pipe %d\n", ring->pipe);
-				break;
-			}
-		} else if (ring->me == 2) {
-			switch (ring->pipe) {
-			case 0:
-				cp_m2p0 |= TIME_STAMP_INT_ENABLE;
-				break;
-			case 1:
-				cp_m2p1 |= TIME_STAMP_INT_ENABLE;
-				break;
-			case 2:
-				cp_m2p2 |= TIME_STAMP_INT_ENABLE;
-				break;
-			case 3:
-				cp_m2p2 |= TIME_STAMP_INT_ENABLE;
-				break;
 			default:
 				DRM_DEBUG("si_irq_set: sw int cp2 invalid pipe %d\n", ring->pipe);
 				break;
@@ -7511,13 +7477,6 @@ int cik_irq_set(struct radeon_device *rdev)
 	WREG32(SDMA0_CNTL + SDMA1_REGISTER_OFFSET, dma_cntl1);
 
 	WREG32(CP_ME1_PIPE0_INT_CNTL, cp_m1p0);
-	WREG32(CP_ME1_PIPE1_INT_CNTL, cp_m1p1);
-	WREG32(CP_ME1_PIPE2_INT_CNTL, cp_m1p2);
-	WREG32(CP_ME1_PIPE3_INT_CNTL, cp_m1p3);
-	WREG32(CP_ME2_PIPE0_INT_CNTL, cp_m2p0);
-	WREG32(CP_ME2_PIPE1_INT_CNTL, cp_m2p1);
-	WREG32(CP_ME2_PIPE2_INT_CNTL, cp_m2p2);
-	WREG32(CP_ME2_PIPE3_INT_CNTL, cp_m2p3);
 
 	WREG32(GRBM_INT_CNTL, grbm_int_cntl);
 
@@ -7834,6 +7793,10 @@ restart_ih:
 	while (rptr != wptr) {
 		/* wptr/rptr are in bytes! */
 		ring_index = rptr / 4;
+
+		radeon_kfd_interrupt(rdev,
+				(const void *) &rdev->ih.ring[ring_index]);
+
 		src_id =  le32_to_cpu(rdev->ih.ring[ring_index]) & 0xff;
 		src_data = le32_to_cpu(rdev->ih.ring[ring_index + 1]) & 0xfffffff;
 		ring_id = le32_to_cpu(rdev->ih.ring[ring_index + 2]) & 0xff;
@@ -8255,8 +8218,10 @@ restart_ih:
 	}
 	if (queue_hotplug)
 		schedule_work(&rdev->hotplug_work);
-	if (queue_reset)
-		schedule_work(&rdev->reset_work);
+	if (queue_reset) {
+		rdev->needs_reset = true;
+		wake_up_all(&rdev->fence_queue);
+	}
 	if (queue_thermal)
 		schedule_work(&rdev->pm.dpm.thermal.work);
 	rdev->ih.rptr = rptr;
@@ -8521,6 +8486,10 @@ static int cik_startup(struct radeon_device *rdev)
 	if (r)
 		return r;
 
+	r = radeon_kfd_resume(rdev);
+	if (r)
+		return r;
+
 	return 0;
 }
 
@@ -8569,6 +8538,7 @@ int cik_resume(struct radeon_device *rdev)
  */
 int cik_suspend(struct radeon_device *rdev)
 {
+	radeon_kfd_suspend(rdev);
 	radeon_pm_suspend(rdev);
 	dce6_audio_fini(rdev);
 	radeon_vm_manager_fini(rdev);
@@ -9447,6 +9417,9 @@ void dce8_bandwidth_update(struct radeon_device *rdev)
 	u32 num_heads = 0, lb_size;
 	int i;
 
+	if (!rdev->mode_info.mode_config_initialized)
+		return;
+
 	radeon_update_display_priority(rdev);
 
 	for (i = 0; i < rdev->num_crtc; i++) {
diff --git a/drivers/gpu/drm/radeon/cik_reg.h b/drivers/gpu/drm/radeon/cik_reg.h
index ca1bb6133580..79c45e8a536b 100644
--- a/drivers/gpu/drm/radeon/cik_reg.h
+++ b/drivers/gpu/drm/radeon/cik_reg.h
@@ -147,4 +147,140 @@
 
 #define CIK_LB_DESKTOP_HEIGHT                     0x6b0c
 
+#define CP_HQD_IQ_RPTR					0xC970u
+#define AQL_ENABLE					(1U << 0)
+
+#define IDLE					(1 << 2)
+
+struct cik_mqd {
+	uint32_t header;
+	uint32_t compute_dispatch_initiator;
+	uint32_t compute_dim_x;
+	uint32_t compute_dim_y;
+	uint32_t compute_dim_z;
+	uint32_t compute_start_x;
+	uint32_t compute_start_y;
+	uint32_t compute_start_z;
+	uint32_t compute_num_thread_x;
+	uint32_t compute_num_thread_y;
+	uint32_t compute_num_thread_z;
+	uint32_t compute_pipelinestat_enable;
+	uint32_t compute_perfcount_enable;
+	uint32_t compute_pgm_lo;
+	uint32_t compute_pgm_hi;
+	uint32_t compute_tba_lo;
+	uint32_t compute_tba_hi;
+	uint32_t compute_tma_lo;
+	uint32_t compute_tma_hi;
+	uint32_t compute_pgm_rsrc1;
+	uint32_t compute_pgm_rsrc2;
+	uint32_t compute_vmid;
+	uint32_t compute_resource_limits;
+	uint32_t compute_static_thread_mgmt_se0;
+	uint32_t compute_static_thread_mgmt_se1;
+	uint32_t compute_tmpring_size;
+	uint32_t compute_static_thread_mgmt_se2;
+	uint32_t compute_static_thread_mgmt_se3;
+	uint32_t compute_restart_x;
+	uint32_t compute_restart_y;
+	uint32_t compute_restart_z;
+	uint32_t compute_thread_trace_enable;
+	uint32_t compute_misc_reserved;
+	uint32_t compute_user_data_0;
+	uint32_t compute_user_data_1;
+	uint32_t compute_user_data_2;
+	uint32_t compute_user_data_3;
+	uint32_t compute_user_data_4;
+	uint32_t compute_user_data_5;
+	uint32_t compute_user_data_6;
+	uint32_t compute_user_data_7;
+	uint32_t compute_user_data_8;
+	uint32_t compute_user_data_9;
+	uint32_t compute_user_data_10;
+	uint32_t compute_user_data_11;
+	uint32_t compute_user_data_12;
+	uint32_t compute_user_data_13;
+	uint32_t compute_user_data_14;
+	uint32_t compute_user_data_15;
+	uint32_t cp_compute_csinvoc_count_lo;
+	uint32_t cp_compute_csinvoc_count_hi;
+	uint32_t cp_mqd_base_addr_lo;
+	uint32_t cp_mqd_base_addr_hi;
+	uint32_t cp_hqd_active;
+	uint32_t cp_hqd_vmid;
+	uint32_t cp_hqd_persistent_state;
+	uint32_t cp_hqd_pipe_priority;
+	uint32_t cp_hqd_queue_priority;
+	uint32_t cp_hqd_quantum;
+	uint32_t cp_hqd_pq_base_lo;
+	uint32_t cp_hqd_pq_base_hi;
+	uint32_t cp_hqd_pq_rptr;
+	uint32_t cp_hqd_pq_rptr_report_addr_lo;
+	uint32_t cp_hqd_pq_rptr_report_addr_hi;
+	uint32_t cp_hqd_pq_wptr_poll_addr_lo;
+	uint32_t cp_hqd_pq_wptr_poll_addr_hi;
+	uint32_t cp_hqd_pq_doorbell_control;
+	uint32_t cp_hqd_pq_wptr;
+	uint32_t cp_hqd_pq_control;
+	uint32_t cp_hqd_ib_base_addr_lo;
+	uint32_t cp_hqd_ib_base_addr_hi;
+	uint32_t cp_hqd_ib_rptr;
+	uint32_t cp_hqd_ib_control;
+	uint32_t cp_hqd_iq_timer;
+	uint32_t cp_hqd_iq_rptr;
+	uint32_t cp_hqd_dequeue_request;
+	uint32_t cp_hqd_dma_offload;
+	uint32_t cp_hqd_sema_cmd;
+	uint32_t cp_hqd_msg_type;
+	uint32_t cp_hqd_atomic0_preop_lo;
+	uint32_t cp_hqd_atomic0_preop_hi;
+	uint32_t cp_hqd_atomic1_preop_lo;
+	uint32_t cp_hqd_atomic1_preop_hi;
+	uint32_t cp_hqd_hq_status0;
+	uint32_t cp_hqd_hq_control0;
+	uint32_t cp_mqd_control;
+	uint32_t cp_mqd_query_time_lo;
+	uint32_t cp_mqd_query_time_hi;
+	uint32_t cp_mqd_connect_start_time_lo;
+	uint32_t cp_mqd_connect_start_time_hi;
+	uint32_t cp_mqd_connect_end_time_lo;
+	uint32_t cp_mqd_connect_end_time_hi;
+	uint32_t cp_mqd_connect_end_wf_count;
+	uint32_t cp_mqd_connect_end_pq_rptr;
+	uint32_t cp_mqd_connect_end_pq_wptr;
+	uint32_t cp_mqd_connect_end_ib_rptr;
+	uint32_t reserved_96;
+	uint32_t reserved_97;
+	uint32_t reserved_98;
+	uint32_t reserved_99;
+	uint32_t iqtimer_pkt_header;
+	uint32_t iqtimer_pkt_dw0;
+	uint32_t iqtimer_pkt_dw1;
+	uint32_t iqtimer_pkt_dw2;
+	uint32_t iqtimer_pkt_dw3;
+	uint32_t iqtimer_pkt_dw4;
+	uint32_t iqtimer_pkt_dw5;
+	uint32_t iqtimer_pkt_dw6;
+	uint32_t reserved_108;
+	uint32_t reserved_109;
+	uint32_t reserved_110;
+	uint32_t reserved_111;
+	uint32_t queue_doorbell_id0;
+	uint32_t queue_doorbell_id1;
+	uint32_t queue_doorbell_id2;
+	uint32_t queue_doorbell_id3;
+	uint32_t queue_doorbell_id4;
+	uint32_t queue_doorbell_id5;
+	uint32_t queue_doorbell_id6;
+	uint32_t queue_doorbell_id7;
+	uint32_t queue_doorbell_id8;
+	uint32_t queue_doorbell_id9;
+	uint32_t queue_doorbell_id10;
+	uint32_t queue_doorbell_id11;
+	uint32_t queue_doorbell_id12;
+	uint32_t queue_doorbell_id13;
+	uint32_t queue_doorbell_id14;
+	uint32_t queue_doorbell_id15;
+};
+
 #endif
diff --git a/drivers/gpu/drm/radeon/cik_sdma.c b/drivers/gpu/drm/radeon/cik_sdma.c
index c4ffa54b1e3d..dde5c7e29eb2 100644
--- a/drivers/gpu/drm/radeon/cik_sdma.c
+++ b/drivers/gpu/drm/radeon/cik_sdma.c
@@ -134,7 +134,7 @@ void cik_sdma_ring_ib_execute(struct radeon_device *rdev,
 			      struct radeon_ib *ib)
 {
 	struct radeon_ring *ring = &rdev->ring[ib->ring];
-	u32 extra_bits = (ib->vm ? ib->vm->id : 0) & 0xf;
+	u32 extra_bits = (ib->vm ? ib->vm->ids[ib->ring].id : 0) & 0xf;
 
 	if (rdev->wb.enabled) {
 		u32 next_rptr = ring->wptr + 5;
@@ -530,41 +530,38 @@ void cik_sdma_fini(struct radeon_device *rdev)
  * @src_offset: src GPU address
  * @dst_offset: dst GPU address
  * @num_gpu_pages: number of GPU pages to xfer
- * @fence: radeon fence object
+ * @resv: reservation object to sync to
  *
  * Copy GPU paging using the DMA engine (CIK).
  * Used by the radeon ttm implementation to move pages if
  * registered as the asic copy callback.
  */
-int cik_copy_dma(struct radeon_device *rdev,
-		 uint64_t src_offset, uint64_t dst_offset,
-		 unsigned num_gpu_pages,
-		 struct radeon_fence **fence)
+struct radeon_fence *cik_copy_dma(struct radeon_device *rdev,
+				  uint64_t src_offset, uint64_t dst_offset,
+				  unsigned num_gpu_pages,
+				  struct reservation_object *resv)
 {
-	struct radeon_semaphore *sem = NULL;
+	struct radeon_fence *fence;
+	struct radeon_sync sync;
 	int ring_index = rdev->asic->copy.dma_ring_index;
 	struct radeon_ring *ring = &rdev->ring[ring_index];
 	u32 size_in_bytes, cur_size_in_bytes;
 	int i, num_loops;
 	int r = 0;
 
-	r = radeon_semaphore_create(rdev, &sem);
-	if (r) {
-		DRM_ERROR("radeon: moving bo (%d).\n", r);
-		return r;
-	}
+	radeon_sync_create(&sync);
 
 	size_in_bytes = (num_gpu_pages << RADEON_GPU_PAGE_SHIFT);
 	num_loops = DIV_ROUND_UP(size_in_bytes, 0x1fffff);
 	r = radeon_ring_lock(rdev, ring, num_loops * 7 + 14);
 	if (r) {
 		DRM_ERROR("radeon: moving bo (%d).\n", r);
-		radeon_semaphore_free(rdev, &sem, NULL);
-		return r;
+		radeon_sync_free(rdev, &sync, NULL);
+		return ERR_PTR(r);
 	}
 
-	radeon_semaphore_sync_to(sem, *fence);
-	radeon_semaphore_sync_rings(rdev, sem, ring->idx);
+	radeon_sync_resv(rdev, &sync, resv, false);
+	radeon_sync_rings(rdev, &sync, ring->idx);
 
 	for (i = 0; i < num_loops; i++) {
 		cur_size_in_bytes = size_in_bytes;
@@ -582,17 +579,17 @@ int cik_copy_dma(struct radeon_device *rdev,
 		dst_offset += cur_size_in_bytes;
 	}
 
-	r = radeon_fence_emit(rdev, fence, ring->idx);
+	r = radeon_fence_emit(rdev, &fence, ring->idx);
 	if (r) {
 		radeon_ring_unlock_undo(rdev, ring);
-		radeon_semaphore_free(rdev, &sem, NULL);
-		return r;
+		radeon_sync_free(rdev, &sync, NULL);
+		return ERR_PTR(r);
 	}
 
 	radeon_ring_unlock_commit(rdev, ring, false);
-	radeon_semaphore_free(rdev, &sem, *fence);
+	radeon_sync_free(rdev, &sync, fence);
 
-	return r;
+	return fence;
 }
 
 /**
@@ -610,16 +607,19 @@ int cik_sdma_ring_test(struct radeon_device *rdev,
 {
 	unsigned i;
 	int r;
-	void __iomem *ptr = (void *)rdev->vram_scratch.ptr;
+	unsigned index;
 	u32 tmp;
+	u64 gpu_addr;
 
-	if (!ptr) {
-		DRM_ERROR("invalid vram scratch pointer\n");
-		return -EINVAL;
-	}
+	if (ring->idx == R600_RING_TYPE_DMA_INDEX)
+		index = R600_WB_DMA_RING_TEST_OFFSET;
+	else
+		index = CAYMAN_WB_DMA1_RING_TEST_OFFSET;
+
+	gpu_addr = rdev->wb.gpu_addr + index;
 
 	tmp = 0xCAFEDEAD;
-	writel(tmp, ptr);
+	rdev->wb.wb[index/4] = cpu_to_le32(tmp);
 
 	r = radeon_ring_lock(rdev, ring, 5);
 	if (r) {
@@ -627,14 +627,14 @@ int cik_sdma_ring_test(struct radeon_device *rdev,
 		return r;
 	}
 	radeon_ring_write(ring, SDMA_PACKET(SDMA_OPCODE_WRITE, SDMA_WRITE_SUB_OPCODE_LINEAR, 0));
-	radeon_ring_write(ring, rdev->vram_scratch.gpu_addr & 0xfffffffc);
-	radeon_ring_write(ring, upper_32_bits(rdev->vram_scratch.gpu_addr));
+	radeon_ring_write(ring, lower_32_bits(gpu_addr));
+	radeon_ring_write(ring, upper_32_bits(gpu_addr));
 	radeon_ring_write(ring, 1); /* number of DWs to follow */
 	radeon_ring_write(ring, 0xDEADBEEF);
 	radeon_ring_unlock_commit(rdev, ring, false);
 
 	for (i = 0; i < rdev->usec_timeout; i++) {
-		tmp = readl(ptr);
+		tmp = le32_to_cpu(rdev->wb.wb[index/4]);
 		if (tmp == 0xDEADBEEF)
 			break;
 		DRM_UDELAY(1);
@@ -663,17 +663,20 @@ int cik_sdma_ib_test(struct radeon_device *rdev, struct radeon_ring *ring)
 {
 	struct radeon_ib ib;
 	unsigned i;
+	unsigned index;
 	int r;
-	void __iomem *ptr = (void *)rdev->vram_scratch.ptr;
 	u32 tmp = 0;
+	u64 gpu_addr;
 
-	if (!ptr) {
-		DRM_ERROR("invalid vram scratch pointer\n");
-		return -EINVAL;
-	}
+	if (ring->idx == R600_RING_TYPE_DMA_INDEX)
+		index = R600_WB_DMA_RING_TEST_OFFSET;
+	else
+		index = CAYMAN_WB_DMA1_RING_TEST_OFFSET;
+
+	gpu_addr = rdev->wb.gpu_addr + index;
 
 	tmp = 0xCAFEDEAD;
-	writel(tmp, ptr);
+	rdev->wb.wb[index/4] = cpu_to_le32(tmp);
 
 	r = radeon_ib_get(rdev, ring->idx, &ib, NULL, 256);
 	if (r) {
@@ -682,8 +685,8 @@ int cik_sdma_ib_test(struct radeon_device *rdev, struct radeon_ring *ring)
 	}
 
 	ib.ptr[0] = SDMA_PACKET(SDMA_OPCODE_WRITE, SDMA_WRITE_SUB_OPCODE_LINEAR, 0);
-	ib.ptr[1] = rdev->vram_scratch.gpu_addr & 0xfffffffc;
-	ib.ptr[2] = upper_32_bits(rdev->vram_scratch.gpu_addr);
+	ib.ptr[1] = lower_32_bits(gpu_addr);
+	ib.ptr[2] = upper_32_bits(gpu_addr);
 	ib.ptr[3] = 1;
 	ib.ptr[4] = 0xDEADBEEF;
 	ib.length_dw = 5;
@@ -700,7 +703,7 @@ int cik_sdma_ib_test(struct radeon_device *rdev, struct radeon_ring *ring)
 		return r;
 	}
 	for (i = 0; i < rdev->usec_timeout; i++) {
-		tmp = readl(ptr);
+		tmp = le32_to_cpu(rdev->wb.wb[index/4]);
 		if (tmp == 0xDEADBEEF)
 			break;
 		DRM_UDELAY(1);
@@ -897,25 +900,21 @@ void cik_sdma_vm_pad_ib(struct radeon_ib *ib)
  * Update the page table base and flush the VM TLB
  * using sDMA (CIK).
  */
-void cik_dma_vm_flush(struct radeon_device *rdev, int ridx, struct radeon_vm *vm)
+void cik_dma_vm_flush(struct radeon_device *rdev, struct radeon_ring *ring,
+		      unsigned vm_id, uint64_t pd_addr)
 {
-	struct radeon_ring *ring = &rdev->ring[ridx];
-
-	if (vm == NULL)
-		return;
-
 	radeon_ring_write(ring, SDMA_PACKET(SDMA_OPCODE_SRBM_WRITE, 0, 0xf000));
-	if (vm->id < 8) {
-		radeon_ring_write(ring, (VM_CONTEXT0_PAGE_TABLE_BASE_ADDR + (vm->id << 2)) >> 2);
+	if (vm_id < 8) {
+		radeon_ring_write(ring, (VM_CONTEXT0_PAGE_TABLE_BASE_ADDR + (vm_id << 2)) >> 2);
 	} else {
-		radeon_ring_write(ring, (VM_CONTEXT8_PAGE_TABLE_BASE_ADDR + ((vm->id - 8) << 2)) >> 2);
+		radeon_ring_write(ring, (VM_CONTEXT8_PAGE_TABLE_BASE_ADDR + ((vm_id - 8) << 2)) >> 2);
 	}
-	radeon_ring_write(ring, vm->pd_gpu_addr >> 12);
+	radeon_ring_write(ring, pd_addr >> 12);
 
 	/* update SH_MEM_* regs */
 	radeon_ring_write(ring, SDMA_PACKET(SDMA_OPCODE_SRBM_WRITE, 0, 0xf000));
 	radeon_ring_write(ring, SRBM_GFX_CNTL >> 2);
-	radeon_ring_write(ring, VMID(vm->id));
+	radeon_ring_write(ring, VMID(vm_id));
 
 	radeon_ring_write(ring, SDMA_PACKET(SDMA_OPCODE_SRBM_WRITE, 0, 0xf000));
 	radeon_ring_write(ring, SH_MEM_BASES >> 2);
@@ -938,11 +937,11 @@ void cik_dma_vm_flush(struct radeon_device *rdev, int ridx, struct radeon_vm *vm
 	radeon_ring_write(ring, VMID(0));
 
 	/* flush HDP */
-	cik_sdma_hdp_flush_ring_emit(rdev, ridx);
+	cik_sdma_hdp_flush_ring_emit(rdev, ring->idx);
 
 	/* flush TLB */
 	radeon_ring_write(ring, SDMA_PACKET(SDMA_OPCODE_SRBM_WRITE, 0, 0xf000));
 	radeon_ring_write(ring, VM_INVALIDATE_REQUEST >> 2);
-	radeon_ring_write(ring, 1 << vm->id);
+	radeon_ring_write(ring, 1 << vm_id);
 }
 
diff --git a/drivers/gpu/drm/radeon/cikd.h b/drivers/gpu/drm/radeon/cikd.h
index 0c6e1b55d968..03003f8a6de6 100644
--- a/drivers/gpu/drm/radeon/cikd.h
+++ b/drivers/gpu/drm/radeon/cikd.h
@@ -30,6 +30,8 @@
 #define CIK_RB_BITMAP_WIDTH_PER_SH     2
 #define HAWAII_RB_BITMAP_WIDTH_PER_SH  4
 
+#define RADEON_NUM_OF_VMIDS	8
+
 /* DIDT IND registers */
 #define DIDT_SQ_CTRL0                                     0x0
 #       define DIDT_CTRL_EN                               (1 << 0)
@@ -184,7 +186,10 @@
 #define		DIG_THERM_DPM(x)			((x) << 14)
 #define		DIG_THERM_DPM_MASK			0x003FC000
 #define		DIG_THERM_DPM_SHIFT			14
-
+#define	CG_THERMAL_STATUS				0xC0300008
+#define		FDO_PWM_DUTY(x)				((x) << 9)
+#define		FDO_PWM_DUTY_MASK			(0xff << 9)
+#define		FDO_PWM_DUTY_SHIFT			9
 #define	CG_THERMAL_INT					0xC030000C
 #define		CI_DIG_THERM_INTH(x)			((x) << 8)
 #define		CI_DIG_THERM_INTH_MASK			0x0000FF00
@@ -194,7 +199,10 @@
 #define		CI_DIG_THERM_INTL_SHIFT			16
 #define 	THERM_INT_MASK_HIGH			(1 << 24)
 #define 	THERM_INT_MASK_LOW			(1 << 25)
-
+#define	CG_MULT_THERMAL_CTRL				0xC0300010
+#define		TEMP_SEL(x)				((x) << 20)
+#define		TEMP_SEL_MASK				(0xff << 20)
+#define		TEMP_SEL_SHIFT				20
 #define	CG_MULT_THERMAL_STATUS				0xC0300014
 #define		ASIC_MAX_TEMP(x)			((x) << 0)
 #define		ASIC_MAX_TEMP_MASK			0x000001ff
@@ -203,6 +211,36 @@
 #define		CTF_TEMP_MASK				0x0003fe00
 #define		CTF_TEMP_SHIFT				9
 
+#define	CG_FDO_CTRL0					0xC0300064
+#define		FDO_STATIC_DUTY(x)			((x) << 0)
+#define		FDO_STATIC_DUTY_MASK			0x000000FF
+#define		FDO_STATIC_DUTY_SHIFT			0
+#define	CG_FDO_CTRL1					0xC0300068
+#define		FMAX_DUTY100(x)				((x) << 0)
+#define		FMAX_DUTY100_MASK			0x000000FF
+#define		FMAX_DUTY100_SHIFT			0
+#define	CG_FDO_CTRL2					0xC030006C
+#define		TMIN(x)					((x) << 0)
+#define		TMIN_MASK				0x000000FF
+#define		TMIN_SHIFT				0
+#define		FDO_PWM_MODE(x)				((x) << 11)
+#define		FDO_PWM_MODE_MASK			(7 << 11)
+#define		FDO_PWM_MODE_SHIFT			11
+#define		TACH_PWM_RESP_RATE(x)			((x) << 25)
+#define		TACH_PWM_RESP_RATE_MASK			(0x7f << 25)
+#define		TACH_PWM_RESP_RATE_SHIFT		25
+#define CG_TACH_CTRL                                    0xC0300070
+#       define EDGE_PER_REV(x)                          ((x) << 0)
+#       define EDGE_PER_REV_MASK                        (0x7 << 0)
+#       define EDGE_PER_REV_SHIFT                       0
+#       define TARGET_PERIOD(x)                         ((x) << 3)
+#       define TARGET_PERIOD_MASK                       0xfffffff8
+#       define TARGET_PERIOD_SHIFT                      3
+#define CG_TACH_STATUS                                  0xC0300074
+#       define TACH_PERIOD(x)                           ((x) << 0)
+#       define TACH_PERIOD_MASK                         0xffffffff
+#       define TACH_PERIOD_SHIFT                        0
+
 #define CG_ECLK_CNTL                                    0xC05000AC
 #       define ECLK_DIVIDER_MASK                        0x7f
 #       define ECLK_DIR_CNTL_EN                         (1 << 8)
@@ -1137,6 +1175,9 @@
 #define			SH_MEM_ALIGNMENT_MODE_UNALIGNED			3
 #define		DEFAULT_MTYPE(x)				((x) << 4)
 #define		APE1_MTYPE(x)					((x) << 7)
+/* valid for both DEFAULT_MTYPE and APE1_MTYPE */
+#define	MTYPE_CACHED					0
+#define	MTYPE_NONCACHED					3
 
 #define	SX_DEBUG_1					0x9060
 
@@ -1447,6 +1488,16 @@
 #define CP_HQD_ACTIVE                                     0xC91C
 #define CP_HQD_VMID                                       0xC920
 
+#define CP_HQD_PERSISTENT_STATE				0xC924u
+#define	DEFAULT_CP_HQD_PERSISTENT_STATE			(0x33U << 8)
+
+#define CP_HQD_PIPE_PRIORITY				0xC928u
+#define CP_HQD_QUEUE_PRIORITY				0xC92Cu
+#define CP_HQD_QUANTUM					0xC930u
+#define	QUANTUM_EN					1U
+#define	QUANTUM_SCALE_1MS				(1U << 4)
+#define	QUANTUM_DURATION(x)				((x) << 8)
+
 #define CP_HQD_PQ_BASE                                    0xC934
 #define CP_HQD_PQ_BASE_HI                                 0xC938
 #define CP_HQD_PQ_RPTR                                    0xC93C
@@ -1474,12 +1525,32 @@
 #define		PRIV_STATE      			(1 << 30)
 #define		KMD_QUEUE      				(1 << 31)
 
-#define CP_HQD_DEQUEUE_REQUEST                          0xC974
+#define CP_HQD_IB_BASE_ADDR				0xC95Cu
+#define CP_HQD_IB_BASE_ADDR_HI			0xC960u
+#define CP_HQD_IB_RPTR					0xC964u
+#define CP_HQD_IB_CONTROL				0xC968u
+#define	IB_ATC_EN					(1U << 23)
+#define	DEFAULT_MIN_IB_AVAIL_SIZE			(3U << 20)
+
+#define CP_HQD_DEQUEUE_REQUEST			0xC974
+#define	DEQUEUE_REQUEST_DRAIN				1
+#define DEQUEUE_REQUEST_RESET				2
 
 #define CP_MQD_CONTROL                                  0xC99C
 #define		MQD_VMID(x)				((x) << 0)
 #define		MQD_VMID_MASK      			(0xf << 0)
 
+#define CP_HQD_SEMA_CMD					0xC97Cu
+#define CP_HQD_MSG_TYPE					0xC980u
+#define CP_HQD_ATOMIC0_PREOP_LO			0xC984u
+#define CP_HQD_ATOMIC0_PREOP_HI			0xC988u
+#define CP_HQD_ATOMIC1_PREOP_LO			0xC98Cu
+#define CP_HQD_ATOMIC1_PREOP_HI			0xC990u
+#define CP_HQD_HQ_SCHEDULER0			0xC994u
+#define CP_HQD_HQ_SCHEDULER1			0xC998u
+
+#define SH_STATIC_MEM_CONFIG			0x9604u
+
 #define DB_RENDER_CONTROL                               0x28000
 
 #define PA_SC_RASTER_CONFIG                             0x28350
@@ -2069,4 +2140,22 @@
 #define VCE_CMD_IB_AUTO		0x00000005
 #define VCE_CMD_SEMAPHORE	0x00000006
 
+#define ATC_VMID0_PASID_MAPPING					0x339Cu
+#define	ATC_VMID_PASID_MAPPING_UPDATE_STATUS	0x3398u
+#define	ATC_VMID_PASID_MAPPING_VALID				(1U << 31)
+
+#define ATC_VM_APERTURE0_CNTL					0x3310u
+#define	ATS_ACCESS_MODE_NEVER						0
+#define	ATS_ACCESS_MODE_ALWAYS						1
+
+#define ATC_VM_APERTURE0_CNTL2					0x3318u
+#define ATC_VM_APERTURE0_HIGH_ADDR				0x3308u
+#define ATC_VM_APERTURE0_LOW_ADDR				0x3300u
+#define ATC_VM_APERTURE1_CNTL					0x3314u
+#define ATC_VM_APERTURE1_CNTL2					0x331Cu
+#define ATC_VM_APERTURE1_HIGH_ADDR				0x330Cu
+#define ATC_VM_APERTURE1_LOW_ADDR				0x3304u
+
+#define IH_VMID_0_LUT						0x3D40u
+
 #endif
diff --git a/drivers/gpu/drm/radeon/cypress_dpm.c b/drivers/gpu/drm/radeon/cypress_dpm.c
index 47d31e915758..9aad0327e4d1 100644
--- a/drivers/gpu/drm/radeon/cypress_dpm.c
+++ b/drivers/gpu/drm/radeon/cypress_dpm.c
@@ -24,6 +24,7 @@
 
 #include "drmP.h"
 #include "radeon.h"
+#include "radeon_asic.h"
 #include "evergreend.h"
 #include "r600_dpm.h"
 #include "cypress_dpm.h"
diff --git a/drivers/gpu/drm/radeon/dce3_1_afmt.c b/drivers/gpu/drm/radeon/dce3_1_afmt.c
index 51800e340a57..bafdf92a5732 100644
--- a/drivers/gpu/drm/radeon/dce3_1_afmt.c
+++ b/drivers/gpu/drm/radeon/dce3_1_afmt.c
@@ -32,7 +32,7 @@ static void dce3_2_afmt_write_speaker_allocation(struct drm_encoder *encoder)
 	struct drm_connector *connector;
 	struct radeon_connector *radeon_connector = NULL;
 	u32 tmp;
-	u8 *sadb;
+	u8 *sadb = NULL;
 	int sad_count;
 
 	list_for_each_entry(connector, &encoder->dev->mode_config.connector_list, head) {
@@ -49,8 +49,8 @@ static void dce3_2_afmt_write_speaker_allocation(struct drm_encoder *encoder)
 
 	sad_count = drm_edid_to_speaker_allocation(radeon_connector->edid, &sadb);
 	if (sad_count < 0) {
-		DRM_ERROR("Couldn't read Speaker Allocation Data Block: %d\n", sad_count);
-		return;
+		DRM_DEBUG("Couldn't read Speaker Allocation Data Block: %d\n", sad_count);
+		sad_count = 0;
 	}
 
 	/* program the speaker allocation */
@@ -103,7 +103,7 @@ static void dce3_2_afmt_write_sad_regs(struct drm_encoder *encoder)
 	}
 
 	sad_count = drm_edid_to_sad(radeon_connector->edid, &sads);
-	if (sad_count < 0) {
+	if (sad_count <= 0) {
 		DRM_ERROR("Couldn't read SADs: %d\n", sad_count);
 		return;
 	}
@@ -165,7 +165,7 @@ void dce3_1_hdmi_setmode(struct drm_encoder *encoder, struct drm_display_mode *m
 
 	/* disable audio prior to setting up hw */
 	dig->afmt->pin = r600_audio_get_pin(rdev);
-	r600_audio_enable(rdev, dig->afmt->pin, false);
+	r600_audio_enable(rdev, dig->afmt->pin, 0);
 
 	r600_audio_set_dto(encoder, mode->clock);
 
@@ -240,5 +240,5 @@ void dce3_1_hdmi_setmode(struct drm_encoder *encoder, struct drm_display_mode *m
 	r600_hdmi_audio_workaround(encoder);
 
 	/* enable audio after to setting up hw */
-	r600_audio_enable(rdev, dig->afmt->pin, true);
+	r600_audio_enable(rdev, dig->afmt->pin, 0xf);
 }
diff --git a/drivers/gpu/drm/radeon/dce6_afmt.c b/drivers/gpu/drm/radeon/dce6_afmt.c
index ab29f953a767..f312edf4d50e 100644
--- a/drivers/gpu/drm/radeon/dce6_afmt.c
+++ b/drivers/gpu/drm/radeon/dce6_afmt.c
@@ -155,7 +155,7 @@ void dce6_afmt_write_speaker_allocation(struct drm_encoder *encoder)
 	struct drm_connector *connector;
 	struct radeon_connector *radeon_connector = NULL;
 	u32 offset, tmp;
-	u8 *sadb;
+	u8 *sadb = NULL;
 	int sad_count;
 
 	if (!dig || !dig->afmt || !dig->afmt->pin)
@@ -176,9 +176,9 @@ void dce6_afmt_write_speaker_allocation(struct drm_encoder *encoder)
 	}
 
 	sad_count = drm_edid_to_speaker_allocation(radeon_connector_edid(connector), &sadb);
-	if (sad_count <= 0) {
-		DRM_ERROR("Couldn't read Speaker Allocation Data Block: %d\n", sad_count);
-		return;
+	if (sad_count < 0) {
+		DRM_DEBUG("Couldn't read Speaker Allocation Data Block: %d\n", sad_count);
+		sad_count = 0;
 	}
 
 	/* program the speaker allocation */
@@ -284,13 +284,13 @@ static int dce6_audio_chipset_supported(struct radeon_device *rdev)
 
 void dce6_audio_enable(struct radeon_device *rdev,
 		       struct r600_audio_pin *pin,
-		       bool enable)
+		       u8 enable_mask)
 {
 	if (!pin)
 		return;
 
-	WREG32_ENDPOINT(pin->offset, AZ_F0_CODEC_PIN_CONTROL_HOTPLUG_CONTROL,
-			enable ? AUDIO_ENABLED : 0);
+	WREG32_ENDPOINT(pin->offset, AZ_F0_CODEC_PIN_CONTROL_HOT_PLUG_CONTROL,
+			enable_mask ? AUDIO_ENABLED : 0);
 }
 
 static const u32 pin_offsets[7] =
diff --git a/drivers/gpu/drm/drm_buffer.c b/drivers/gpu/drm/radeon/drm_buffer.c
index 86a4a4a60afc..f4e0f3a3d7b1 100644
--- a/drivers/gpu/drm/drm_buffer.c
+++ b/drivers/gpu/drm/radeon/drm_buffer.c
@@ -33,7 +33,7 @@
  */
 
 #include <linux/export.h>
-#include <drm/drm_buffer.h>
+#include "drm_buffer.h"
 
 /**
  * Allocate the drm buffer object.
@@ -86,7 +86,6 @@ error_out:
 	kfree(*buf);
 	return -ENOMEM;
 }
-EXPORT_SYMBOL(drm_buffer_alloc);
 
 /**
  * Copy the user data to the begin of the buffer and reset the processing
@@ -123,7 +122,6 @@ int drm_buffer_copy_from_user(struct drm_buffer *buf,
 	buf->iterator = 0;
 	return 0;
 }
-EXPORT_SYMBOL(drm_buffer_copy_from_user);
 
 /**
  * Free the drm buffer object
@@ -141,7 +139,6 @@ void drm_buffer_free(struct drm_buffer *buf)
 		kfree(buf);
 	}
 }
-EXPORT_SYMBOL(drm_buffer_free);
 
 /**
  * Read an object from buffer that may be split to multiple parts. If object
@@ -178,4 +175,3 @@ void *drm_buffer_read_object(struct drm_buffer *buf,
 	drm_buffer_advance(buf, objsize);
 	return obj;
 }
-EXPORT_SYMBOL(drm_buffer_read_object);
diff --git a/drivers/gpu/drm/radeon/drm_buffer.h b/drivers/gpu/drm/radeon/drm_buffer.h
new file mode 100644
index 000000000000..c80d3a340b94
--- /dev/null
+++ b/drivers/gpu/drm/radeon/drm_buffer.h
@@ -0,0 +1,148 @@
+/**************************************************************************
+ *
+ * Copyright 2010 Pauli Nieminen.
+ * All Rights Reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the
+ * "Software"), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sub license, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the
+ * next paragraph) shall be included in all copies or substantial portions
+ * of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDERS, AUTHORS AND/OR ITS SUPPLIERS BE LIABLE FOR ANY CLAIM,
+ * DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR
+ * OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE
+ * USE OR OTHER DEALINGS IN THE SOFTWARE.
+ *
+ *
+ **************************************************************************/
+/*
+ * Multipart buffer for coping data which is larger than the page size.
+ *
+ * Authors:
+ * Pauli Nieminen <suokkos-at-gmail-dot-com>
+ */
+
+#ifndef _DRM_BUFFER_H_
+#define _DRM_BUFFER_H_
+
+#include <drm/drmP.h>
+
+struct drm_buffer {
+	int iterator;
+	int size;
+	char *data[];
+};
+
+
+/**
+ * Return the index of page that buffer is currently pointing at.
+ */
+static inline int drm_buffer_page(struct drm_buffer *buf)
+{
+	return buf->iterator / PAGE_SIZE;
+}
+/**
+ * Return the index of the current byte in the page
+ */
+static inline int drm_buffer_index(struct drm_buffer *buf)
+{
+	return buf->iterator & (PAGE_SIZE - 1);
+}
+/**
+ * Return number of bytes that is left to process
+ */
+static inline int drm_buffer_unprocessed(struct drm_buffer *buf)
+{
+	return buf->size - buf->iterator;
+}
+
+/**
+ * Advance the buffer iterator number of bytes that is given.
+ */
+static inline void drm_buffer_advance(struct drm_buffer *buf, int bytes)
+{
+	buf->iterator += bytes;
+}
+
+/**
+ * Allocate the drm buffer object.
+ *
+ *   buf: A pointer to a pointer where the object is stored.
+ *   size: The number of bytes to allocate.
+ */
+extern int drm_buffer_alloc(struct drm_buffer **buf, int size);
+
+/**
+ * Copy the user data to the begin of the buffer and reset the processing
+ * iterator.
+ *
+ *   user_data: A pointer the data that is copied to the buffer.
+ *   size: The Number of bytes to copy.
+ */
+extern int drm_buffer_copy_from_user(struct drm_buffer *buf,
+		void __user *user_data, int size);
+
+/**
+ * Free the drm buffer object
+ */
+extern void drm_buffer_free(struct drm_buffer *buf);
+
+/**
+ * Read an object from buffer that may be split to multiple parts. If object
+ * is not split function just returns the pointer to object in buffer. But in
+ * case of split object data is copied to given stack object that is suplied
+ * by caller.
+ *
+ * The processing location of the buffer is also advanced to the next byte
+ * after the object.
+ *
+ *   objsize: The size of the objet in bytes.
+ *   stack_obj: A pointer to a memory location where object can be copied.
+ */
+extern void *drm_buffer_read_object(struct drm_buffer *buf,
+		int objsize, void *stack_obj);
+
+/**
+ * Returns the pointer to the dword which is offset number of elements from the
+ * current processing location.
+ *
+ * Caller must make sure that dword is not split in the buffer. This
+ * requirement is easily met if all the sizes of objects in buffer are
+ * multiples of dword and PAGE_SIZE is multiple dword.
+ *
+ * Call to this function doesn't change the processing location.
+ *
+ *   offset: The index of the dword relative to the internat iterator.
+ */
+static inline void *drm_buffer_pointer_to_dword(struct drm_buffer *buffer,
+		int offset)
+{
+	int iter = buffer->iterator + offset * 4;
+	return &buffer->data[iter / PAGE_SIZE][iter & (PAGE_SIZE - 1)];
+}
+/**
+ * Returns the pointer to the dword which is offset number of elements from
+ * the current processing location.
+ *
+ * Call to this function doesn't change the processing location.
+ *
+ *   offset: The index of the byte relative to the internat iterator.
+ */
+static inline void *drm_buffer_pointer_to_byte(struct drm_buffer *buffer,
+		int offset)
+{
+	int iter = buffer->iterator + offset;
+	return &buffer->data[iter / PAGE_SIZE][iter & (PAGE_SIZE - 1)];
+}
+
+#endif
diff --git a/drivers/gpu/drm/radeon/evergreen.c b/drivers/gpu/drm/radeon/evergreen.c
index e50807c29f69..85995b4e3338 100644
--- a/drivers/gpu/drm/radeon/evergreen.c
+++ b/drivers/gpu/drm/radeon/evergreen.c
@@ -22,7 +22,6 @@
  * Authors: Alex Deucher
  */
 #include <linux/firmware.h>
-#include <linux/platform_device.h>
 #include <linux/slab.h>
 #include <drm/drmP.h>
 #include "radeon.h"
@@ -2346,6 +2345,9 @@ void evergreen_bandwidth_update(struct radeon_device *rdev)
 	u32 num_heads = 0, lb_size;
 	int i;
 
+	if (!rdev->mode_info.mode_config_initialized)
+		return;
+
 	radeon_update_display_priority(rdev);
 
 	for (i = 0; i < rdev->num_crtc; i++) {
@@ -2553,6 +2555,7 @@ void evergreen_mc_stop(struct radeon_device *rdev, struct evergreen_mc_save *sav
 					WREG32(EVERGREEN_CRTC_UPDATE_LOCK + crtc_offsets[i], 1);
 					tmp |= EVERGREEN_CRTC_BLANK_DATA_EN;
 					WREG32(EVERGREEN_CRTC_BLANK_CONTROL + crtc_offsets[i], tmp);
+					WREG32(EVERGREEN_CRTC_UPDATE_LOCK + crtc_offsets[i], 0);
 				}
 			} else {
 				tmp = RREG32(EVERGREEN_CRTC_CONTROL + crtc_offsets[i]);
@@ -3006,7 +3009,7 @@ static void evergreen_gpu_init(struct radeon_device *rdev)
 	u32 vgt_cache_invalidation;
 	u32 hdp_host_path_cntl, tmp;
 	u32 disabled_rb_mask;
-	int i, j, num_shader_engines, ps_thread_count;
+	int i, j, ps_thread_count;
 
 	switch (rdev->family) {
 	case CHIP_CYPRESS:
@@ -3304,8 +3307,6 @@ static void evergreen_gpu_init(struct radeon_device *rdev)
 	rdev->config.evergreen.tile_config |=
 		((gb_addr_config & 0x30000000) >> 28) << 12;
 
-	num_shader_engines = (gb_addr_config & NUM_SHADER_ENGINES(3) >> 12) + 1;
-
 	if ((rdev->family >= CHIP_CEDAR) && (rdev->family <= CHIP_HEMLOCK)) {
 		u32 efuse_straps_4;
 		u32 efuse_straps_3;
@@ -4023,7 +4024,7 @@ int sumo_rlc_init(struct radeon_device *rdev)
 		if (rdev->rlc.save_restore_obj == NULL) {
 			r = radeon_bo_create(rdev, dws * 4, PAGE_SIZE, true,
 					     RADEON_GEM_DOMAIN_VRAM, 0, NULL,
-					     &rdev->rlc.save_restore_obj);
+					     NULL, &rdev->rlc.save_restore_obj);
 			if (r) {
 				dev_warn(rdev->dev, "(%d) create RLC sr bo failed\n", r);
 				return r;
@@ -4102,7 +4103,7 @@ int sumo_rlc_init(struct radeon_device *rdev)
 		if (rdev->rlc.clear_state_obj == NULL) {
 			r = radeon_bo_create(rdev, dws * 4, PAGE_SIZE, true,
 					     RADEON_GEM_DOMAIN_VRAM, 0, NULL,
-					     &rdev->rlc.clear_state_obj);
+					     NULL, &rdev->rlc.clear_state_obj);
 			if (r) {
 				dev_warn(rdev->dev, "(%d) create RLC c bo failed\n", r);
 				sumo_rlc_fini(rdev);
@@ -4179,7 +4180,7 @@ int sumo_rlc_init(struct radeon_device *rdev)
 			r = radeon_bo_create(rdev, rdev->rlc.cp_table_size,
 					     PAGE_SIZE, true,
 					     RADEON_GEM_DOMAIN_VRAM, 0, NULL,
-					     &rdev->rlc.cp_table_obj);
+					     NULL, &rdev->rlc.cp_table_obj);
 			if (r) {
 				dev_warn(rdev->dev, "(%d) create RLC cp table bo failed\n", r);
 				sumo_rlc_fini(rdev);
diff --git a/drivers/gpu/drm/radeon/evergreen_cs.c b/drivers/gpu/drm/radeon/evergreen_cs.c
index 5c8b358f9fba..924b1b7ab455 100644
--- a/drivers/gpu/drm/radeon/evergreen_cs.c
+++ b/drivers/gpu/drm/radeon/evergreen_cs.c
@@ -35,7 +35,7 @@
 #define MIN(a,b)                   (((a)<(b))?(a):(b))
 
 int r600_dma_cs_next_reloc(struct radeon_cs_parser *p,
-			   struct radeon_cs_reloc **cs_reloc);
+			   struct radeon_bo_list **cs_reloc);
 struct evergreen_cs_track {
 	u32			group_size;
 	u32			nbanks;
@@ -1094,7 +1094,7 @@ static int evergreen_cs_parse_packet0(struct radeon_cs_parser *p,
 static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
 {
 	struct evergreen_cs_track *track = (struct evergreen_cs_track *)p->track;
-	struct radeon_cs_reloc *reloc;
+	struct radeon_bo_list *reloc;
 	u32 last_reg;
 	u32 m, i, tmp, *ib;
 	int r;
@@ -1792,7 +1792,7 @@ static bool evergreen_is_safe_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
 static int evergreen_packet3_check(struct radeon_cs_parser *p,
 				   struct radeon_cs_packet *pkt)
 {
-	struct radeon_cs_reloc *reloc;
+	struct radeon_bo_list *reloc;
 	struct evergreen_cs_track *track;
 	volatile u32 *ib;
 	unsigned idx;
@@ -2661,7 +2661,7 @@ int evergreen_cs_parse(struct radeon_cs_parser *p)
 			p->track = NULL;
 			return r;
 		}
-	} while (p->idx < p->chunks[p->chunk_ib_idx].length_dw);
+	} while (p->idx < p->chunk_ib->length_dw);
 #if 0
 	for (r = 0; r < p->ib.length_dw; r++) {
 		printk(KERN_INFO "%05d  0x%08X\n", r, p->ib.ptr[r]);
@@ -2684,8 +2684,8 @@ int evergreen_cs_parse(struct radeon_cs_parser *p)
  **/
 int evergreen_dma_cs_parse(struct radeon_cs_parser *p)
 {
-	struct radeon_cs_chunk *ib_chunk = &p->chunks[p->chunk_ib_idx];
-	struct radeon_cs_reloc *src_reloc, *dst_reloc, *dst2_reloc;
+	struct radeon_cs_chunk *ib_chunk = p->chunk_ib;
+	struct radeon_bo_list *src_reloc, *dst_reloc, *dst2_reloc;
 	u32 header, cmd, count, sub_cmd;
 	volatile u32 *ib = p->ib.ptr;
 	u32 idx;
@@ -3100,7 +3100,7 @@ int evergreen_dma_cs_parse(struct radeon_cs_parser *p)
 			DRM_ERROR("Unknown packet type %d at %d !\n", cmd, idx);
 			return -EINVAL;
 		}
-	} while (p->idx < p->chunks[p->chunk_ib_idx].length_dw);
+	} while (p->idx < p->chunk_ib->length_dw);
 #if 0
 	for (r = 0; r < p->ib->length_dw; r++) {
 		printk(KERN_INFO "%05d  0x%08X\n", r, p->ib.ptr[r]);
diff --git a/drivers/gpu/drm/radeon/evergreen_dma.c b/drivers/gpu/drm/radeon/evergreen_dma.c
index afaba388c36d..96535aa8659c 100644
--- a/drivers/gpu/drm/radeon/evergreen_dma.c
+++ b/drivers/gpu/drm/radeon/evergreen_dma.c
@@ -104,35 +104,33 @@ void evergreen_dma_ring_ib_execute(struct radeon_device *rdev,
  * Used by the radeon ttm implementation to move pages if
  * registered as the asic copy callback.
  */
-int evergreen_copy_dma(struct radeon_device *rdev,
-		       uint64_t src_offset, uint64_t dst_offset,
-		       unsigned num_gpu_pages,
-		       struct radeon_fence **fence)
+struct radeon_fence *evergreen_copy_dma(struct radeon_device *rdev,
+					uint64_t src_offset,
+					uint64_t dst_offset,
+					unsigned num_gpu_pages,
+					struct reservation_object *resv)
 {
-	struct radeon_semaphore *sem = NULL;
+	struct radeon_fence *fence;
+	struct radeon_sync sync;
 	int ring_index = rdev->asic->copy.dma_ring_index;
 	struct radeon_ring *ring = &rdev->ring[ring_index];
 	u32 size_in_dw, cur_size_in_dw;
 	int i, num_loops;
 	int r = 0;
 
-	r = radeon_semaphore_create(rdev, &sem);
-	if (r) {
-		DRM_ERROR("radeon: moving bo (%d).\n", r);
-		return r;
-	}
+	radeon_sync_create(&sync);
 
 	size_in_dw = (num_gpu_pages << RADEON_GPU_PAGE_SHIFT) / 4;
 	num_loops = DIV_ROUND_UP(size_in_dw, 0xfffff);
 	r = radeon_ring_lock(rdev, ring, num_loops * 5 + 11);
 	if (r) {
 		DRM_ERROR("radeon: moving bo (%d).\n", r);
-		radeon_semaphore_free(rdev, &sem, NULL);
-		return r;
+		radeon_sync_free(rdev, &sync, NULL);
+		return ERR_PTR(r);
 	}
 
-	radeon_semaphore_sync_to(sem, *fence);
-	radeon_semaphore_sync_rings(rdev, sem, ring->idx);
+	radeon_sync_resv(rdev, &sync, resv, false);
+	radeon_sync_rings(rdev, &sync, ring->idx);
 
 	for (i = 0; i < num_loops; i++) {
 		cur_size_in_dw = size_in_dw;
@@ -148,17 +146,17 @@ int evergreen_copy_dma(struct radeon_device *rdev,
 		dst_offset += cur_size_in_dw * 4;
 	}
 
-	r = radeon_fence_emit(rdev, fence, ring->idx);
+	r = radeon_fence_emit(rdev, &fence, ring->idx);
 	if (r) {
 		radeon_ring_unlock_undo(rdev, ring);
-		radeon_semaphore_free(rdev, &sem, NULL);
-		return r;
+		radeon_sync_free(rdev, &sync, NULL);
+		return ERR_PTR(r);
 	}
 
 	radeon_ring_unlock_commit(rdev, ring, false);
-	radeon_semaphore_free(rdev, &sem, *fence);
+	radeon_sync_free(rdev, &sync, fence);
 
-	return r;
+	return fence;
 }
 
 /**
diff --git a/drivers/gpu/drm/radeon/evergreen_hdmi.c b/drivers/gpu/drm/radeon/evergreen_hdmi.c
index 278c7a139d74..53abd9b17a50 100644
--- a/drivers/gpu/drm/radeon/evergreen_hdmi.c
+++ b/drivers/gpu/drm/radeon/evergreen_hdmi.c
@@ -38,6 +38,37 @@ extern void dce6_afmt_select_pin(struct drm_encoder *encoder);
 extern void dce6_afmt_write_latency_fields(struct drm_encoder *encoder,
 					   struct drm_display_mode *mode);
 
+/* enable the audio stream */
+static void dce4_audio_enable(struct radeon_device *rdev,
+			      struct r600_audio_pin *pin,
+			      u8 enable_mask)
+{
+	u32 tmp = RREG32(AZ_HOT_PLUG_CONTROL);
+
+	if (!pin)
+		return;
+
+	if (enable_mask) {
+		tmp |= AUDIO_ENABLED;
+		if (enable_mask & 1)
+			tmp |= PIN0_AUDIO_ENABLED;
+		if (enable_mask & 2)
+			tmp |= PIN1_AUDIO_ENABLED;
+		if (enable_mask & 4)
+			tmp |= PIN2_AUDIO_ENABLED;
+		if (enable_mask & 8)
+			tmp |= PIN3_AUDIO_ENABLED;
+	} else {
+		tmp &= ~(AUDIO_ENABLED |
+			 PIN0_AUDIO_ENABLED |
+			 PIN1_AUDIO_ENABLED |
+			 PIN2_AUDIO_ENABLED |
+			 PIN3_AUDIO_ENABLED);
+	}
+
+	WREG32(AZ_HOT_PLUG_CONTROL, tmp);
+}
+
 /*
  * update the N and CTS parameters for a given pixel clock rate
  */
@@ -102,7 +133,7 @@ static void dce4_afmt_write_speaker_allocation(struct drm_encoder *encoder)
 	struct drm_connector *connector;
 	struct radeon_connector *radeon_connector = NULL;
 	u32 tmp;
-	u8 *sadb;
+	u8 *sadb = NULL;
 	int sad_count;
 
 	list_for_each_entry(connector, &encoder->dev->mode_config.connector_list, head) {
@@ -118,9 +149,9 @@ static void dce4_afmt_write_speaker_allocation(struct drm_encoder *encoder)
 	}
 
 	sad_count = drm_edid_to_speaker_allocation(radeon_connector_edid(connector), &sadb);
-	if (sad_count <= 0) {
-		DRM_ERROR("Couldn't read Speaker Allocation Data Block: %d\n", sad_count);
-		return;
+	if (sad_count < 0) {
+		DRM_DEBUG("Couldn't read Speaker Allocation Data Block: %d\n", sad_count);
+		sad_count = 0;
 	}
 
 	/* program the speaker allocation */
@@ -318,10 +349,10 @@ void evergreen_hdmi_setmode(struct drm_encoder *encoder, struct drm_display_mode
 	/* disable audio prior to setting up hw */
 	if (ASIC_IS_DCE6(rdev)) {
 		dig->afmt->pin = dce6_audio_get_pin(rdev);
-		dce6_audio_enable(rdev, dig->afmt->pin, false);
+		dce6_audio_enable(rdev, dig->afmt->pin, 0);
 	} else {
 		dig->afmt->pin = r600_audio_get_pin(rdev);
-		r600_audio_enable(rdev, dig->afmt->pin, false);
+		dce4_audio_enable(rdev, dig->afmt->pin, 0);
 	}
 
 	evergreen_audio_set_dto(encoder, mode->clock);
@@ -463,13 +494,15 @@ void evergreen_hdmi_setmode(struct drm_encoder *encoder, struct drm_display_mode
 
 	/* enable audio after to setting up hw */
 	if (ASIC_IS_DCE6(rdev))
-		dce6_audio_enable(rdev, dig->afmt->pin, true);
+		dce6_audio_enable(rdev, dig->afmt->pin, 1);
 	else
-		r600_audio_enable(rdev, dig->afmt->pin, true);
+		dce4_audio_enable(rdev, dig->afmt->pin, 0xf);
 }
 
 void evergreen_hdmi_enable(struct drm_encoder *encoder, bool enable)
 {
+	struct drm_device *dev = encoder->dev;
+	struct radeon_device *rdev = dev->dev_private;
 	struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder);
 	struct radeon_encoder_atom_dig *dig = radeon_encoder->enc_priv;
 
@@ -482,6 +515,14 @@ void evergreen_hdmi_enable(struct drm_encoder *encoder, bool enable)
 	if (!enable && !dig->afmt->enabled)
 		return;
 
+	if (!enable && dig->afmt->pin) {
+		if (ASIC_IS_DCE6(rdev))
+			dce6_audio_enable(rdev, dig->afmt->pin, 0);
+		else
+			dce4_audio_enable(rdev, dig->afmt->pin, 0);
+		dig->afmt->pin = NULL;
+	}
+
 	dig->afmt->enabled = enable;
 
 	DRM_DEBUG("%sabling HDMI interface @ 0x%04X for encoder 0x%x\n",
diff --git a/drivers/gpu/drm/radeon/kv_dpm.c b/drivers/gpu/drm/radeon/kv_dpm.c
index 67cb472d188c..e3e9c10cfba9 100644
--- a/drivers/gpu/drm/radeon/kv_dpm.c
+++ b/drivers/gpu/drm/radeon/kv_dpm.c
@@ -2725,7 +2725,11 @@ int kv_dpm_init(struct radeon_device *rdev)
 
         pi->sram_end = SMC_RAM_END;
 
-	pi->enable_nb_dpm = true;
+	/* Enabling nb dpm on an asrock system prevents dpm from working */
+	if (rdev->pdev->subsystem_vendor == 0x1849)
+		pi->enable_nb_dpm = false;
+	else
+		pi->enable_nb_dpm = true;
 
 	pi->caps_power_containment = true;
 	pi->caps_cac = true;
@@ -2740,10 +2744,17 @@ int kv_dpm_init(struct radeon_device *rdev)
 	pi->caps_sclk_ds = true;
 	pi->enable_auto_thermal_throttling = true;
 	pi->disable_nb_ps3_in_battery = false;
-	if (radeon_bapm == 0)
+	if (radeon_bapm == -1) {
+		/* only enable bapm on KB, ML by default */
+		if (rdev->family == CHIP_KABINI || rdev->family == CHIP_MULLINS)
+			pi->bapm_enable = true;
+		else
+			pi->bapm_enable = false;
+	} else if (radeon_bapm == 0) {
 		pi->bapm_enable = false;
-	else
+	} else {
 		pi->bapm_enable = true;
+	}
 	pi->voltage_drop_t = 0;
 	pi->caps_sclk_throttle_low_notification = false;
 	pi->caps_fps = false; /* true? */
@@ -2787,6 +2798,8 @@ void kv_dpm_debugfs_print_current_performance_level(struct radeon_device *rdev,
 		tmp = (RREG32_SMC(SMU_VOLTAGE_STATUS) & SMU_VOLTAGE_CURRENT_LEVEL_MASK) >>
 			SMU_VOLTAGE_CURRENT_LEVEL_SHIFT;
 		vddc = kv_convert_8bit_index_to_voltage(rdev, (u16)tmp);
+		seq_printf(m, "uvd    %sabled\n", pi->uvd_power_gated ? "dis" : "en");
+		seq_printf(m, "vce    %sabled\n", pi->vce_power_gated ? "dis" : "en");
 		seq_printf(m, "power level %d    sclk: %u vddc: %u\n",
 			   current_index, sclk, vddc);
 	}
diff --git a/drivers/gpu/drm/radeon/mkregtable.c b/drivers/gpu/drm/radeon/mkregtable.c
index 4a85bb644e24..b928c17bdeed 100644
--- a/drivers/gpu/drm/radeon/mkregtable.c
+++ b/drivers/gpu/drm/radeon/mkregtable.c
@@ -347,7 +347,7 @@ static inline void list_splice_tail_init(struct list_head *list,
  * list_entry - get the struct for this entry
  * @ptr:	the &struct list_head pointer.
  * @type:	the type of the struct this is embedded in.
- * @member:	the name of the list_struct within the struct.
+ * @member:	the name of the list_head within the struct.
  */
 #define list_entry(ptr, type, member) \
 	container_of(ptr, type, member)
@@ -356,7 +356,7 @@ static inline void list_splice_tail_init(struct list_head *list,
  * list_first_entry - get the first element from a list
  * @ptr:	the list head to take the element from.
  * @type:	the type of the struct this is embedded in.
- * @member:	the name of the list_struct within the struct.
+ * @member:	the name of the list_head within the struct.
  *
  * Note, that list is expected to be not empty.
  */
@@ -406,7 +406,7 @@ static inline void list_splice_tail_init(struct list_head *list,
  * list_for_each_entry	-	iterate over list of given type
  * @pos:	the type * to use as a loop cursor.
  * @head:	the head for your list.
- * @member:	the name of the list_struct within the struct.
+ * @member:	the name of the list_head within the struct.
  */
 #define list_for_each_entry(pos, head, member)				\
 	for (pos = list_entry((head)->next, typeof(*pos), member);	\
@@ -417,7 +417,7 @@ static inline void list_splice_tail_init(struct list_head *list,
  * list_for_each_entry_reverse - iterate backwards over list of given type.
  * @pos:	the type * to use as a loop cursor.
  * @head:	the head for your list.
- * @member:	the name of the list_struct within the struct.
+ * @member:	the name of the list_head within the struct.
  */
 #define list_for_each_entry_reverse(pos, head, member)			\
 	for (pos = list_entry((head)->prev, typeof(*pos), member);	\
@@ -428,7 +428,7 @@ static inline void list_splice_tail_init(struct list_head *list,
  * list_prepare_entry - prepare a pos entry for use in list_for_each_entry_continue()
  * @pos:	the type * to use as a start point
  * @head:	the head of the list
- * @member:	the name of the list_struct within the struct.
+ * @member:	the name of the list_head within the struct.
  *
  * Prepares a pos entry for use as a start point in list_for_each_entry_continue().
  */
@@ -439,7 +439,7 @@ static inline void list_splice_tail_init(struct list_head *list,
  * list_for_each_entry_continue - continue iteration over list of given type
  * @pos:	the type * to use as a loop cursor.
  * @head:	the head for your list.
- * @member:	the name of the list_struct within the struct.
+ * @member:	the name of the list_head within the struct.
  *
  * Continue to iterate over list of given type, continuing after
  * the current position.
@@ -453,7 +453,7 @@ static inline void list_splice_tail_init(struct list_head *list,
  * list_for_each_entry_continue_reverse - iterate backwards from the given point
  * @pos:	the type * to use as a loop cursor.
  * @head:	the head for your list.
- * @member:	the name of the list_struct within the struct.
+ * @member:	the name of the list_head within the struct.
  *
  * Start to iterate over list of given type backwards, continuing after
  * the current position.
@@ -467,7 +467,7 @@ static inline void list_splice_tail_init(struct list_head *list,
  * list_for_each_entry_from - iterate over list of given type from the current point
  * @pos:	the type * to use as a loop cursor.
  * @head:	the head for your list.
- * @member:	the name of the list_struct within the struct.
+ * @member:	the name of the list_head within the struct.
  *
  * Iterate over list of given type, continuing from current position.
  */
@@ -480,7 +480,7 @@ static inline void list_splice_tail_init(struct list_head *list,
  * @pos:	the type * to use as a loop cursor.
  * @n:		another type * to use as temporary storage
  * @head:	the head for your list.
- * @member:	the name of the list_struct within the struct.
+ * @member:	the name of the list_head within the struct.
  */
 #define list_for_each_entry_safe(pos, n, head, member)			\
 	for (pos = list_entry((head)->next, typeof(*pos), member),	\
@@ -493,7 +493,7 @@ static inline void list_splice_tail_init(struct list_head *list,
  * @pos:	the type * to use as a loop cursor.
  * @n:		another type * to use as temporary storage
  * @head:	the head for your list.
- * @member:	the name of the list_struct within the struct.
+ * @member:	the name of the list_head within the struct.
  *
  * Iterate over list of given type, continuing after current point,
  * safe against removal of list entry.
@@ -509,7 +509,7 @@ static inline void list_splice_tail_init(struct list_head *list,
  * @pos:	the type * to use as a loop cursor.
  * @n:		another type * to use as temporary storage
  * @head:	the head for your list.
- * @member:	the name of the list_struct within the struct.
+ * @member:	the name of the list_head within the struct.
  *
  * Iterate over list of given type from current point, safe against
  * removal of list entry.
@@ -524,7 +524,7 @@ static inline void list_splice_tail_init(struct list_head *list,
  * @pos:	the type * to use as a loop cursor.
  * @n:		another type * to use as temporary storage
  * @head:	the head for your list.
- * @member:	the name of the list_struct within the struct.
+ * @member:	the name of the list_head within the struct.
  *
  * Iterate backwards over list of given type, safe against removal
  * of list entry.
diff --git a/drivers/gpu/drm/radeon/ni.c b/drivers/gpu/drm/radeon/ni.c
index 3faee58946dd..360de9f1f491 100644
--- a/drivers/gpu/drm/radeon/ni.c
+++ b/drivers/gpu/drm/radeon/ni.c
@@ -1373,6 +1373,7 @@ void cayman_fence_ring_emit(struct radeon_device *rdev,
 void cayman_ring_ib_execute(struct radeon_device *rdev, struct radeon_ib *ib)
 {
 	struct radeon_ring *ring = &rdev->ring[ib->ring];
+	unsigned vm_id = ib->vm ? ib->vm->ids[ib->ring].id : 0;
 	u32 cp_coher_cntl = PACKET3_FULL_CACHE_ENA | PACKET3_TC_ACTION_ENA |
 		PACKET3_SH_ACTION_ENA;
 
@@ -1395,15 +1396,14 @@ void cayman_ring_ib_execute(struct radeon_device *rdev, struct radeon_ib *ib)
 #endif
 			  (ib->gpu_addr & 0xFFFFFFFC));
 	radeon_ring_write(ring, upper_32_bits(ib->gpu_addr) & 0xFF);
-	radeon_ring_write(ring, ib->length_dw | 
-			  (ib->vm ? (ib->vm->id << 24) : 0));
+	radeon_ring_write(ring, ib->length_dw | (vm_id << 24));
 
 	/* flush read cache over gart for this vmid */
 	radeon_ring_write(ring, PACKET3(PACKET3_SURFACE_SYNC, 3));
 	radeon_ring_write(ring, PACKET3_ENGINE_ME | cp_coher_cntl);
 	radeon_ring_write(ring, 0xFFFFFFFF);
 	radeon_ring_write(ring, 0);
-	radeon_ring_write(ring, ((ib->vm ? ib->vm->id : 0) << 24) | 10); /* poll interval */
+	radeon_ring_write(ring, (vm_id << 24) | 10); /* poll interval */
 }
 
 static void cayman_cp_enable(struct radeon_device *rdev, bool enable)
@@ -2502,15 +2502,11 @@ void cayman_vm_decode_fault(struct radeon_device *rdev,
  * Update the page table base and flush the VM TLB
  * using the CP (cayman-si).
  */
-void cayman_vm_flush(struct radeon_device *rdev, int ridx, struct radeon_vm *vm)
+void cayman_vm_flush(struct radeon_device *rdev, struct radeon_ring *ring,
+		     unsigned vm_id, uint64_t pd_addr)
 {
-	struct radeon_ring *ring = &rdev->ring[ridx];
-
-	if (vm == NULL)
-		return;
-
-	radeon_ring_write(ring, PACKET0(VM_CONTEXT0_PAGE_TABLE_BASE_ADDR + (vm->id << 2), 0));
-	radeon_ring_write(ring, vm->pd_gpu_addr >> 12);
+	radeon_ring_write(ring, PACKET0(VM_CONTEXT0_PAGE_TABLE_BASE_ADDR + (vm_id << 2), 0));
+	radeon_ring_write(ring, pd_addr >> 12);
 
 	/* flush hdp cache */
 	radeon_ring_write(ring, PACKET0(HDP_MEM_COHERENCY_FLUSH_CNTL, 0));
@@ -2518,7 +2514,7 @@ void cayman_vm_flush(struct radeon_device *rdev, int ridx, struct radeon_vm *vm)
 
 	/* bits 0-7 are the VM contexts0-7 */
 	radeon_ring_write(ring, PACKET0(VM_INVALIDATE_REQUEST, 0));
-	radeon_ring_write(ring, 1 << vm->id);
+	radeon_ring_write(ring, 1 << vm_id);
 
 	/* sync PFP to ME, otherwise we might get invalid PFP reads */
 	radeon_ring_write(ring, PACKET3(PACKET3_PFP_SYNC_ME, 0));
diff --git a/drivers/gpu/drm/radeon/ni_dma.c b/drivers/gpu/drm/radeon/ni_dma.c
index f26f0a9fb522..50f88611ff60 100644
--- a/drivers/gpu/drm/radeon/ni_dma.c
+++ b/drivers/gpu/drm/radeon/ni_dma.c
@@ -123,6 +123,7 @@ void cayman_dma_ring_ib_execute(struct radeon_device *rdev,
 				struct radeon_ib *ib)
 {
 	struct radeon_ring *ring = &rdev->ring[ib->ring];
+	unsigned vm_id = ib->vm ? ib->vm->ids[ib->ring].id : 0;
 
 	if (rdev->wb.enabled) {
 		u32 next_rptr = ring->wptr + 4;
@@ -140,7 +141,7 @@ void cayman_dma_ring_ib_execute(struct radeon_device *rdev,
 	 */
 	while ((ring->wptr & 7) != 5)
 		radeon_ring_write(ring, DMA_PACKET(DMA_PACKET_NOP, 0, 0, 0));
-	radeon_ring_write(ring, DMA_IB_PACKET(DMA_PACKET_INDIRECT_BUFFER, ib->vm ? ib->vm->id : 0, 0));
+	radeon_ring_write(ring, DMA_IB_PACKET(DMA_PACKET_INDIRECT_BUFFER, vm_id, 0));
 	radeon_ring_write(ring, (ib->gpu_addr & 0xFFFFFFE0));
 	radeon_ring_write(ring, (ib->length_dw << 12) | (upper_32_bits(ib->gpu_addr) & 0xFF));
 
@@ -446,16 +447,12 @@ void cayman_dma_vm_pad_ib(struct radeon_ib *ib)
 		ib->ptr[ib->length_dw++] = DMA_PACKET(DMA_PACKET_NOP, 0, 0, 0);
 }
 
-void cayman_dma_vm_flush(struct radeon_device *rdev, int ridx, struct radeon_vm *vm)
+void cayman_dma_vm_flush(struct radeon_device *rdev, struct radeon_ring *ring,
+			 unsigned vm_id, uint64_t pd_addr)
 {
-	struct radeon_ring *ring = &rdev->ring[ridx];
-
-	if (vm == NULL)
-		return;
-
 	radeon_ring_write(ring, DMA_PACKET(DMA_PACKET_SRBM_WRITE, 0, 0, 0));
-	radeon_ring_write(ring, (0xf << 16) | ((VM_CONTEXT0_PAGE_TABLE_BASE_ADDR + (vm->id << 2)) >> 2));
-	radeon_ring_write(ring, vm->pd_gpu_addr >> 12);
+	radeon_ring_write(ring, (0xf << 16) | ((VM_CONTEXT0_PAGE_TABLE_BASE_ADDR + (vm_id << 2)) >> 2));
+	radeon_ring_write(ring, pd_addr >> 12);
 
 	/* flush hdp cache */
 	radeon_ring_write(ring, DMA_PACKET(DMA_PACKET_SRBM_WRITE, 0, 0, 0));
@@ -465,6 +462,6 @@ void cayman_dma_vm_flush(struct radeon_device *rdev, int ridx, struct radeon_vm
 	/* bits 0-7 are the VM contexts0-7 */
 	radeon_ring_write(ring, DMA_PACKET(DMA_PACKET_SRBM_WRITE, 0, 0, 0));
 	radeon_ring_write(ring, (0xf << 16) | (VM_INVALIDATE_REQUEST >> 2));
-	radeon_ring_write(ring, 1 << vm->id);
+	radeon_ring_write(ring, 1 << vm_id);
 }
 
diff --git a/drivers/gpu/drm/radeon/ni_dpm.c b/drivers/gpu/drm/radeon/ni_dpm.c
index 01fc4888e6fe..6d2f16cf2c1c 100644
--- a/drivers/gpu/drm/radeon/ni_dpm.c
+++ b/drivers/gpu/drm/radeon/ni_dpm.c
@@ -23,6 +23,7 @@
 
 #include "drmP.h"
 #include "radeon.h"
+#include "radeon_asic.h"
 #include "nid.h"
 #include "r600_dpm.h"
 #include "ni_dpm.h"
@@ -789,7 +790,6 @@ static void ni_apply_state_adjust_rules(struct radeon_device *rdev,
 	bool disable_mclk_switching;
 	u32 mclk;
 	u16 vddci;
-	u32 max_sclk_vddc, max_mclk_vddci, max_mclk_vddc;
 	int i;
 
 	if ((rdev->pm.dpm.new_active_crtc_count > 1) ||
@@ -816,29 +816,6 @@ static void ni_apply_state_adjust_rules(struct radeon_device *rdev,
 		}
 	}
 
-	/* limit clocks to max supported clocks based on voltage dependency tables */
-	btc_get_max_clock_from_voltage_dependency_table(&rdev->pm.dpm.dyn_state.vddc_dependency_on_sclk,
-							&max_sclk_vddc);
-	btc_get_max_clock_from_voltage_dependency_table(&rdev->pm.dpm.dyn_state.vddci_dependency_on_mclk,
-							&max_mclk_vddci);
-	btc_get_max_clock_from_voltage_dependency_table(&rdev->pm.dpm.dyn_state.vddc_dependency_on_mclk,
-							&max_mclk_vddc);
-
-	for (i = 0; i < ps->performance_level_count; i++) {
-		if (max_sclk_vddc) {
-			if (ps->performance_levels[i].sclk > max_sclk_vddc)
-				ps->performance_levels[i].sclk = max_sclk_vddc;
-		}
-		if (max_mclk_vddci) {
-			if (ps->performance_levels[i].mclk > max_mclk_vddci)
-				ps->performance_levels[i].mclk = max_mclk_vddci;
-		}
-		if (max_mclk_vddc) {
-			if (ps->performance_levels[i].mclk > max_mclk_vddc)
-				ps->performance_levels[i].mclk = max_mclk_vddc;
-		}
-	}
-
 	/* XXX validate the min clocks required for display */
 
 	/* adjust low state */
diff --git a/drivers/gpu/drm/radeon/ppsmc.h b/drivers/gpu/drm/radeon/ppsmc.h
index 5670b8291285..7e5724a12f8b 100644
--- a/drivers/gpu/drm/radeon/ppsmc.h
+++ b/drivers/gpu/drm/radeon/ppsmc.h
@@ -56,6 +56,14 @@
 #define PPSMC_STATEFLAG_DEEPSLEEP_THROTTLE 0x20
 #define PPSMC_STATEFLAG_DEEPSLEEP_BYPASS   0x40
 
+#define FDO_MODE_HARDWARE 0
+#define FDO_MODE_PIECE_WISE_LINEAR 1
+
+enum FAN_CONTROL {
+	FAN_CONTROL_FUZZY,
+	FAN_CONTROL_TABLE
+};
+
 #define PPSMC_Result_OK             ((uint8_t)0x01)
 #define PPSMC_Result_Failed         ((uint8_t)0xFF)
 
@@ -79,6 +87,8 @@ typedef uint8_t PPSMC_Result;
 #define PPSMC_MSG_DisableCac                ((uint8_t)0x54)
 #define PPSMC_TDPClampingActive             ((uint8_t)0x59)
 #define PPSMC_TDPClampingInactive           ((uint8_t)0x5A)
+#define PPSMC_StartFanControl               ((uint8_t)0x5B)
+#define PPSMC_StopFanControl                ((uint8_t)0x5C)
 #define PPSMC_MSG_NoDisplay                 ((uint8_t)0x5D)
 #define PPSMC_MSG_HasDisplay                ((uint8_t)0x5E)
 #define PPSMC_MSG_UVDPowerOFF               ((uint8_t)0x60)
@@ -106,6 +116,7 @@ typedef uint8_t PPSMC_Result;
 #define PPSMC_MSG_SAMUDPM_SetEnabledMask      ((uint16_t) 0x130)
 #define PPSMC_MSG_MCLKDPM_ForceState          ((uint16_t) 0x131)
 #define PPSMC_MSG_MCLKDPM_NoForcedLevel       ((uint16_t) 0x132)
+#define PPSMC_MSG_Thermal_Cntl_Disable        ((uint16_t) 0x133)
 #define PPSMC_MSG_Voltage_Cntl_Disable        ((uint16_t) 0x135)
 #define PPSMC_MSG_PCIeDPM_Enable              ((uint16_t) 0x136)
 #define PPSMC_MSG_PCIeDPM_Disable             ((uint16_t) 0x13d)
@@ -149,6 +160,10 @@ typedef uint8_t PPSMC_Result;
 #define PPSMC_MSG_MASTER_DeepSleep_ON         ((uint16_t) 0x18F)
 #define PPSMC_MSG_MASTER_DeepSleep_OFF        ((uint16_t) 0x190)
 #define PPSMC_MSG_Remove_DC_Clamp             ((uint16_t) 0x191)
+#define PPSMC_MSG_SetFanPwmMax                ((uint16_t) 0x19A)
+
+#define PPSMC_MSG_ENABLE_THERMAL_DPM          ((uint16_t) 0x19C)
+#define PPSMC_MSG_DISABLE_THERMAL_DPM         ((uint16_t) 0x19D)
 
 #define PPSMC_MSG_API_GetSclkFrequency        ((uint16_t) 0x200)
 #define PPSMC_MSG_API_GetMclkFrequency        ((uint16_t) 0x201)
@@ -157,10 +172,11 @@ typedef uint8_t PPSMC_Result;
 #define PPSMC_MSG_DPM_Config                ((uint32_t) 0x102)
 #define PPSMC_MSG_DPM_ForceState            ((uint32_t) 0x104)
 #define PPSMC_MSG_PG_SIMD_Config            ((uint32_t) 0x108)
-#define PPSMC_MSG_DPM_N_LevelsDisabled      ((uint32_t) 0x112)
+#define PPSMC_MSG_Thermal_Cntl_Enable       ((uint32_t) 0x10a)
 #define PPSMC_MSG_Voltage_Cntl_Enable       ((uint32_t) 0x109)
 #define PPSMC_MSG_VCEPowerOFF               ((uint32_t) 0x10e)
 #define PPSMC_MSG_VCEPowerON                ((uint32_t) 0x10f)
+#define PPSMC_MSG_DPM_N_LevelsDisabled      ((uint32_t) 0x112)
 #define PPSMC_MSG_DCE_RemoveVoltageAdjustment   ((uint32_t) 0x11d)
 #define PPSMC_MSG_DCE_AllowVoltageAdjustment    ((uint32_t) 0x11e)
 #define PPSMC_MSG_EnableBAPM                ((uint32_t) 0x120)
diff --git a/drivers/gpu/drm/radeon/pptable.h b/drivers/gpu/drm/radeon/pptable.h
index 2d532996c697..4c2eec49dadc 100644
--- a/drivers/gpu/drm/radeon/pptable.h
+++ b/drivers/gpu/drm/radeon/pptable.h
@@ -96,6 +96,14 @@ typedef struct _ATOM_PPLIB_FANTABLE2
     USHORT  usTMax;                          // The max temperature
 } ATOM_PPLIB_FANTABLE2;
 
+typedef struct _ATOM_PPLIB_FANTABLE3
+{
+	ATOM_PPLIB_FANTABLE2 basicTable2;
+	UCHAR ucFanControlMode;
+	USHORT usFanPWMMax;
+	USHORT usFanOutputSensitivity;
+} ATOM_PPLIB_FANTABLE3;
+
 typedef struct _ATOM_PPLIB_EXTENDEDHEADER
 {
     USHORT  usSize;
diff --git a/drivers/gpu/drm/radeon/r100.c b/drivers/gpu/drm/radeon/r100.c
index b0098e792e62..74f06d540591 100644
--- a/drivers/gpu/drm/radeon/r100.c
+++ b/drivers/gpu/drm/radeon/r100.c
@@ -869,13 +869,14 @@ bool r100_semaphore_ring_emit(struct radeon_device *rdev,
 	return false;
 }
 
-int r100_copy_blit(struct radeon_device *rdev,
-		   uint64_t src_offset,
-		   uint64_t dst_offset,
-		   unsigned num_gpu_pages,
-		   struct radeon_fence **fence)
+struct radeon_fence *r100_copy_blit(struct radeon_device *rdev,
+				    uint64_t src_offset,
+				    uint64_t dst_offset,
+				    unsigned num_gpu_pages,
+				    struct reservation_object *resv)
 {
 	struct radeon_ring *ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX];
+	struct radeon_fence *fence;
 	uint32_t cur_pages;
 	uint32_t stride_bytes = RADEON_GPU_PAGE_SIZE;
 	uint32_t pitch;
@@ -896,7 +897,7 @@ int r100_copy_blit(struct radeon_device *rdev,
 	r = radeon_ring_lock(rdev, ring, ndw);
 	if (r) {
 		DRM_ERROR("radeon: moving bo (%d) asking for %u dw.\n", r, ndw);
-		return -EINVAL;
+		return ERR_PTR(-EINVAL);
 	}
 	while (num_gpu_pages > 0) {
 		cur_pages = num_gpu_pages;
@@ -936,11 +937,13 @@ int r100_copy_blit(struct radeon_device *rdev,
 			  RADEON_WAIT_2D_IDLECLEAN |
 			  RADEON_WAIT_HOST_IDLECLEAN |
 			  RADEON_WAIT_DMA_GUI_IDLE);
-	if (fence) {
-		r = radeon_fence_emit(rdev, fence, RADEON_RING_TYPE_GFX_INDEX);
+	r = radeon_fence_emit(rdev, &fence, RADEON_RING_TYPE_GFX_INDEX);
+	if (r) {
+		radeon_ring_unlock_undo(rdev, ring);
+		return ERR_PTR(r);
 	}
 	radeon_ring_unlock_commit(rdev, ring, false);
-	return r;
+	return fence;
 }
 
 static int r100_cp_wait_for_idle(struct radeon_device *rdev)
@@ -1251,7 +1254,7 @@ int r100_reloc_pitch_offset(struct radeon_cs_parser *p,
 	int r;
 	u32 tile_flags = 0;
 	u32 tmp;
-	struct radeon_cs_reloc *reloc;
+	struct radeon_bo_list *reloc;
 	u32 value;
 
 	r = radeon_cs_packet_next_reloc(p, &reloc, 0);
@@ -1290,7 +1293,7 @@ int r100_packet3_load_vbpntr(struct radeon_cs_parser *p,
 			     int idx)
 {
 	unsigned c, i;
-	struct radeon_cs_reloc *reloc;
+	struct radeon_bo_list *reloc;
 	struct r100_cs_track *track;
 	int r = 0;
 	volatile uint32_t *ib;
@@ -1539,7 +1542,7 @@ static int r100_packet0_check(struct radeon_cs_parser *p,
 			      struct radeon_cs_packet *pkt,
 			      unsigned idx, unsigned reg)
 {
-	struct radeon_cs_reloc *reloc;
+	struct radeon_bo_list *reloc;
 	struct r100_cs_track *track;
 	volatile uint32_t *ib;
 	uint32_t tmp;
@@ -1898,7 +1901,7 @@ int r100_cs_track_check_pkt3_indx_buffer(struct radeon_cs_parser *p,
 static int r100_packet3_check(struct radeon_cs_parser *p,
 			      struct radeon_cs_packet *pkt)
 {
-	struct radeon_cs_reloc *reloc;
+	struct radeon_bo_list *reloc;
 	struct r100_cs_track *track;
 	unsigned idx;
 	volatile uint32_t *ib;
@@ -2058,7 +2061,7 @@ int r100_cs_parse(struct radeon_cs_parser *p)
 		}
 		if (r)
 			return r;
-	} while (p->idx < p->chunks[p->chunk_ib_idx].length_dw);
+	} while (p->idx < p->chunk_ib->length_dw);
 	return 0;
 }
 
@@ -3204,6 +3207,9 @@ void r100_bandwidth_update(struct radeon_device *rdev)
 	uint32_t pixel_bytes1 = 0;
 	uint32_t pixel_bytes2 = 0;
 
+	if (!rdev->mode_info.mode_config_initialized)
+		return;
+
 	radeon_update_display_priority(rdev);
 
 	if (rdev->mode_info.crtcs[0]->base.enabled) {
diff --git a/drivers/gpu/drm/radeon/r200.c b/drivers/gpu/drm/radeon/r200.c
index 67780374a652..c70e6d5bcd19 100644
--- a/drivers/gpu/drm/radeon/r200.c
+++ b/drivers/gpu/drm/radeon/r200.c
@@ -80,13 +80,14 @@ static int r200_get_vtx_size_0(uint32_t vtx_fmt_0)
 	return vtx_size;
 }
 
-int r200_copy_dma(struct radeon_device *rdev,
-		  uint64_t src_offset,
-		  uint64_t dst_offset,
-		  unsigned num_gpu_pages,
-		  struct radeon_fence **fence)
+struct radeon_fence *r200_copy_dma(struct radeon_device *rdev,
+				   uint64_t src_offset,
+				   uint64_t dst_offset,
+				   unsigned num_gpu_pages,
+				   struct reservation_object *resv)
 {
 	struct radeon_ring *ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX];
+	struct radeon_fence *fence;
 	uint32_t size;
 	uint32_t cur_size;
 	int i, num_loops;
@@ -98,7 +99,7 @@ int r200_copy_dma(struct radeon_device *rdev,
 	r = radeon_ring_lock(rdev, ring, num_loops * 4 + 64);
 	if (r) {
 		DRM_ERROR("radeon: moving bo (%d).\n", r);
-		return r;
+		return ERR_PTR(r);
 	}
 	/* Must wait for 2D idle & clean before DMA or hangs might happen */
 	radeon_ring_write(ring, PACKET0(RADEON_WAIT_UNTIL, 0));
@@ -118,11 +119,13 @@ int r200_copy_dma(struct radeon_device *rdev,
 	}
 	radeon_ring_write(ring, PACKET0(RADEON_WAIT_UNTIL, 0));
 	radeon_ring_write(ring, RADEON_WAIT_DMA_GUI_IDLE);
-	if (fence) {
-		r = radeon_fence_emit(rdev, fence, RADEON_RING_TYPE_GFX_INDEX);
+	r = radeon_fence_emit(rdev, &fence, RADEON_RING_TYPE_GFX_INDEX);
+	if (r) {
+		radeon_ring_unlock_undo(rdev, ring);
+		return ERR_PTR(r);
 	}
 	radeon_ring_unlock_commit(rdev, ring, false);
-	return r;
+	return fence;
 }
 
 
@@ -143,7 +146,7 @@ int r200_packet0_check(struct radeon_cs_parser *p,
 		       struct radeon_cs_packet *pkt,
 		       unsigned idx, unsigned reg)
 {
-	struct radeon_cs_reloc *reloc;
+	struct radeon_bo_list *reloc;
 	struct r100_cs_track *track;
 	volatile uint32_t *ib;
 	uint32_t tmp;
diff --git a/drivers/gpu/drm/radeon/r300.c b/drivers/gpu/drm/radeon/r300.c
index 1bc4704034ce..064ad5569cca 100644
--- a/drivers/gpu/drm/radeon/r300.c
+++ b/drivers/gpu/drm/radeon/r300.c
@@ -598,7 +598,7 @@ static int r300_packet0_check(struct radeon_cs_parser *p,
 		struct radeon_cs_packet *pkt,
 		unsigned idx, unsigned reg)
 {
-	struct radeon_cs_reloc *reloc;
+	struct radeon_bo_list *reloc;
 	struct r100_cs_track *track;
 	volatile uint32_t *ib;
 	uint32_t tmp, tile_flags = 0;
@@ -1142,7 +1142,7 @@ fail:
 static int r300_packet3_check(struct radeon_cs_parser *p,
 			      struct radeon_cs_packet *pkt)
 {
-	struct radeon_cs_reloc *reloc;
+	struct radeon_bo_list *reloc;
 	struct r100_cs_track *track;
 	volatile uint32_t *ib;
 	unsigned idx;
@@ -1283,7 +1283,7 @@ int r300_cs_parse(struct radeon_cs_parser *p)
 		if (r) {
 			return r;
 		}
-	} while (p->idx < p->chunks[p->chunk_ib_idx].length_dw);
+	} while (p->idx < p->chunk_ib->length_dw);
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/radeon/r300_cmdbuf.c b/drivers/gpu/drm/radeon/r300_cmdbuf.c
index 84b1d5367a11..9418e388b045 100644
--- a/drivers/gpu/drm/radeon/r300_cmdbuf.c
+++ b/drivers/gpu/drm/radeon/r300_cmdbuf.c
@@ -34,10 +34,10 @@
  */
 
 #include <drm/drmP.h>
-#include <drm/drm_buffer.h>
 #include <drm/radeon_drm.h>
 #include "radeon_drv.h"
 #include "r300_reg.h"
+#include "drm_buffer.h"
 
 #include <asm/unaligned.h>
 
diff --git a/drivers/gpu/drm/radeon/r600.c b/drivers/gpu/drm/radeon/r600.c
index ea5c9af722ef..ef5d6066fa5b 100644
--- a/drivers/gpu/drm/radeon/r600.c
+++ b/drivers/gpu/drm/radeon/r600.c
@@ -122,6 +122,94 @@ u32 r600_get_xclk(struct radeon_device *rdev)
 
 int r600_set_uvd_clocks(struct radeon_device *rdev, u32 vclk, u32 dclk)
 {
+	unsigned fb_div = 0, ref_div, vclk_div = 0, dclk_div = 0;
+	int r;
+
+	/* bypass vclk and dclk with bclk */
+	WREG32_P(CG_UPLL_FUNC_CNTL_2,
+		 VCLK_SRC_SEL(1) | DCLK_SRC_SEL(1),
+		 ~(VCLK_SRC_SEL_MASK | DCLK_SRC_SEL_MASK));
+
+	/* assert BYPASS_EN, deassert UPLL_RESET, UPLL_SLEEP and UPLL_CTLREQ */
+	WREG32_P(CG_UPLL_FUNC_CNTL, UPLL_BYPASS_EN_MASK, ~(
+		 UPLL_RESET_MASK | UPLL_SLEEP_MASK | UPLL_CTLREQ_MASK));
+
+	if (rdev->family >= CHIP_RS780)
+		WREG32_P(GFX_MACRO_BYPASS_CNTL, UPLL_BYPASS_CNTL,
+			 ~UPLL_BYPASS_CNTL);
+
+	if (!vclk || !dclk) {
+		/* keep the Bypass mode, put PLL to sleep */
+		WREG32_P(CG_UPLL_FUNC_CNTL, UPLL_SLEEP_MASK, ~UPLL_SLEEP_MASK);
+		return 0;
+	}
+
+	if (rdev->clock.spll.reference_freq == 10000)
+		ref_div = 34;
+	else
+		ref_div = 4;
+
+	r = radeon_uvd_calc_upll_dividers(rdev, vclk, dclk, 50000, 160000,
+					  ref_div + 1, 0xFFF, 2, 30, ~0,
+					  &fb_div, &vclk_div, &dclk_div);
+	if (r)
+		return r;
+
+	if (rdev->family >= CHIP_RV670 && rdev->family < CHIP_RS780)
+		fb_div >>= 1;
+	else
+		fb_div |= 1;
+
+	r = radeon_uvd_send_upll_ctlreq(rdev, CG_UPLL_FUNC_CNTL);
+        if (r)
+                return r;
+
+	/* assert PLL_RESET */
+	WREG32_P(CG_UPLL_FUNC_CNTL, UPLL_RESET_MASK, ~UPLL_RESET_MASK);
+
+	/* For RS780 we have to choose ref clk */
+	if (rdev->family >= CHIP_RS780)
+		WREG32_P(CG_UPLL_FUNC_CNTL, UPLL_REFCLK_SRC_SEL_MASK,
+			 ~UPLL_REFCLK_SRC_SEL_MASK);
+
+	/* set the required fb, ref and post divder values */
+	WREG32_P(CG_UPLL_FUNC_CNTL,
+		 UPLL_FB_DIV(fb_div) |
+		 UPLL_REF_DIV(ref_div),
+		 ~(UPLL_FB_DIV_MASK | UPLL_REF_DIV_MASK));
+	WREG32_P(CG_UPLL_FUNC_CNTL_2,
+		 UPLL_SW_HILEN(vclk_div >> 1) |
+		 UPLL_SW_LOLEN((vclk_div >> 1) + (vclk_div & 1)) |
+		 UPLL_SW_HILEN2(dclk_div >> 1) |
+		 UPLL_SW_LOLEN2((dclk_div >> 1) + (dclk_div & 1)) |
+		 UPLL_DIVEN_MASK | UPLL_DIVEN2_MASK,
+		 ~UPLL_SW_MASK);
+
+	/* give the PLL some time to settle */
+	mdelay(15);
+
+	/* deassert PLL_RESET */
+	WREG32_P(CG_UPLL_FUNC_CNTL, 0, ~UPLL_RESET_MASK);
+
+	mdelay(15);
+
+	/* deassert BYPASS EN */
+	WREG32_P(CG_UPLL_FUNC_CNTL, 0, ~UPLL_BYPASS_EN_MASK);
+
+	if (rdev->family >= CHIP_RS780)
+		WREG32_P(GFX_MACRO_BYPASS_CNTL, 0, ~UPLL_BYPASS_CNTL);
+
+	r = radeon_uvd_send_upll_ctlreq(rdev, CG_UPLL_FUNC_CNTL);
+	if (r)
+		return r;
+
+	/* switch VCLK and DCLK selection */
+	WREG32_P(CG_UPLL_FUNC_CNTL_2,
+		 VCLK_SRC_SEL(2) | DCLK_SRC_SEL(2),
+		 ~(VCLK_SRC_SEL_MASK | DCLK_SRC_SEL_MASK));
+
+	mdelay(100);
+
 	return 0;
 }
 
@@ -992,6 +1080,8 @@ static int r600_pcie_gart_enable(struct radeon_device *rdev)
 	WREG32(MC_VM_L1_TLB_MCB_WR_GFX_CNTL, tmp);
 	WREG32(MC_VM_L1_TLB_MCB_RD_PDMA_CNTL, tmp);
 	WREG32(MC_VM_L1_TLB_MCB_WR_PDMA_CNTL, tmp);
+	WREG32(MC_VM_L1_TLB_MCB_RD_UVD_CNTL, tmp);
+	WREG32(MC_VM_L1_TLB_MCB_WR_UVD_CNTL, tmp);
 	WREG32(MC_VM_L1_TLB_MCB_RD_SEM_CNTL, tmp | ENABLE_SEMAPHORE_MODE);
 	WREG32(MC_VM_L1_TLB_MCB_WR_SEM_CNTL, tmp | ENABLE_SEMAPHORE_MODE);
 	WREG32(VM_CONTEXT0_PAGE_TABLE_START_ADDR, rdev->mc.gtt_start >> 12);
@@ -1042,6 +1132,8 @@ static void r600_pcie_gart_disable(struct radeon_device *rdev)
 	WREG32(MC_VM_L1_TLB_MCB_WR_SYS_CNTL, tmp);
 	WREG32(MC_VM_L1_TLB_MCB_RD_HDP_CNTL, tmp);
 	WREG32(MC_VM_L1_TLB_MCB_WR_HDP_CNTL, tmp);
+	WREG32(MC_VM_L1_TLB_MCB_RD_UVD_CNTL, tmp);
+	WREG32(MC_VM_L1_TLB_MCB_WR_UVD_CNTL, tmp);
 	radeon_gart_table_vram_unpin(rdev);
 }
 
@@ -1338,7 +1430,7 @@ int r600_vram_scratch_init(struct radeon_device *rdev)
 	if (rdev->vram_scratch.robj == NULL) {
 		r = radeon_bo_create(rdev, RADEON_GPU_PAGE_SIZE,
 				     PAGE_SIZE, true, RADEON_GEM_DOMAIN_VRAM,
-				     0, NULL, &rdev->vram_scratch.robj);
+				     0, NULL, NULL, &rdev->vram_scratch.robj);
 		if (r) {
 			return r;
 		}
@@ -2792,35 +2884,32 @@ bool r600_semaphore_ring_emit(struct radeon_device *rdev,
  * Used by the radeon ttm implementation to move pages if
  * registered as the asic copy callback.
  */
-int r600_copy_cpdma(struct radeon_device *rdev,
-		    uint64_t src_offset, uint64_t dst_offset,
-		    unsigned num_gpu_pages,
-		    struct radeon_fence **fence)
+struct radeon_fence *r600_copy_cpdma(struct radeon_device *rdev,
+				     uint64_t src_offset, uint64_t dst_offset,
+				     unsigned num_gpu_pages,
+				     struct reservation_object *resv)
 {
-	struct radeon_semaphore *sem = NULL;
+	struct radeon_fence *fence;
+	struct radeon_sync sync;
 	int ring_index = rdev->asic->copy.blit_ring_index;
 	struct radeon_ring *ring = &rdev->ring[ring_index];
 	u32 size_in_bytes, cur_size_in_bytes, tmp;
 	int i, num_loops;
 	int r = 0;
 
-	r = radeon_semaphore_create(rdev, &sem);
-	if (r) {
-		DRM_ERROR("radeon: moving bo (%d).\n", r);
-		return r;
-	}
+	radeon_sync_create(&sync);
 
 	size_in_bytes = (num_gpu_pages << RADEON_GPU_PAGE_SHIFT);
 	num_loops = DIV_ROUND_UP(size_in_bytes, 0x1fffff);
 	r = radeon_ring_lock(rdev, ring, num_loops * 6 + 24);
 	if (r) {
 		DRM_ERROR("radeon: moving bo (%d).\n", r);
-		radeon_semaphore_free(rdev, &sem, NULL);
-		return r;
+		radeon_sync_free(rdev, &sync, NULL);
+		return ERR_PTR(r);
 	}
 
-	radeon_semaphore_sync_to(sem, *fence);
-	radeon_semaphore_sync_rings(rdev, sem, ring->idx);
+	radeon_sync_resv(rdev, &sync, resv, false);
+	radeon_sync_rings(rdev, &sync, ring->idx);
 
 	radeon_ring_write(ring, PACKET3(PACKET3_SET_CONFIG_REG, 1));
 	radeon_ring_write(ring, (WAIT_UNTIL - PACKET3_SET_CONFIG_REG_OFFSET) >> 2);
@@ -2846,17 +2935,17 @@ int r600_copy_cpdma(struct radeon_device *rdev,
 	radeon_ring_write(ring, (WAIT_UNTIL - PACKET3_SET_CONFIG_REG_OFFSET) >> 2);
 	radeon_ring_write(ring, WAIT_CP_DMA_IDLE_bit);
 
-	r = radeon_fence_emit(rdev, fence, ring->idx);
+	r = radeon_fence_emit(rdev, &fence, ring->idx);
 	if (r) {
 		radeon_ring_unlock_undo(rdev, ring);
-		radeon_semaphore_free(rdev, &sem, NULL);
-		return r;
+		radeon_sync_free(rdev, &sync, NULL);
+		return ERR_PTR(r);
 	}
 
 	radeon_ring_unlock_commit(rdev, ring, false);
-	radeon_semaphore_free(rdev, &sem, *fence);
+	radeon_sync_free(rdev, &sync, fence);
 
-	return r;
+	return fence;
 }
 
 int r600_set_surface_reg(struct radeon_device *rdev, int reg,
@@ -2907,6 +2996,18 @@ static int r600_startup(struct radeon_device *rdev)
 		return r;
 	}
 
+	if (rdev->has_uvd) {
+		r = uvd_v1_0_resume(rdev);
+		if (!r) {
+			r = radeon_fence_driver_start_ring(rdev, R600_RING_TYPE_UVD_INDEX);
+			if (r) {
+				dev_err(rdev->dev, "failed initializing UVD fences (%d).\n", r);
+			}
+		}
+		if (r)
+			rdev->ring[R600_RING_TYPE_UVD_INDEX].ring_size = 0;
+	}
+
 	/* Enable IRQ */
 	if (!rdev->irq.installed) {
 		r = radeon_irq_kms_init(rdev);
@@ -2935,6 +3036,18 @@ static int r600_startup(struct radeon_device *rdev)
 	if (r)
 		return r;
 
+	if (rdev->has_uvd) {
+		ring = &rdev->ring[R600_RING_TYPE_UVD_INDEX];
+		if (ring->ring_size) {
+			r = radeon_ring_init(rdev, ring, ring->ring_size, 0,
+					     RADEON_CP_PACKET2);
+			if (!r)
+				r = uvd_v1_0_init(rdev);
+			if (r)
+				DRM_ERROR("radeon: failed initializing UVD (%d).\n", r);
+		}
+	}
+
 	r = radeon_ib_pool_init(rdev);
 	if (r) {
 		dev_err(rdev->dev, "IB initialization failed (%d).\n", r);
@@ -2994,6 +3107,10 @@ int r600_suspend(struct radeon_device *rdev)
 	radeon_pm_suspend(rdev);
 	r600_audio_fini(rdev);
 	r600_cp_stop(rdev);
+	if (rdev->has_uvd) {
+		uvd_v1_0_fini(rdev);
+		radeon_uvd_suspend(rdev);
+	}
 	r600_irq_suspend(rdev);
 	radeon_wb_disable(rdev);
 	r600_pcie_gart_disable(rdev);
@@ -3073,6 +3190,14 @@ int r600_init(struct radeon_device *rdev)
 	rdev->ring[RADEON_RING_TYPE_GFX_INDEX].ring_obj = NULL;
 	r600_ring_init(rdev, &rdev->ring[RADEON_RING_TYPE_GFX_INDEX], 1024 * 1024);
 
+	if (rdev->has_uvd) {
+		r = radeon_uvd_init(rdev);
+		if (!r) {
+			rdev->ring[R600_RING_TYPE_UVD_INDEX].ring_obj = NULL;
+			r600_ring_init(rdev, &rdev->ring[R600_RING_TYPE_UVD_INDEX], 4096);
+		}
+	}
+
 	rdev->ih.ring_obj = NULL;
 	r600_ih_ring_init(rdev, 64 * 1024);
 
@@ -3102,6 +3227,10 @@ void r600_fini(struct radeon_device *rdev)
 	r600_audio_fini(rdev);
 	r600_cp_fini(rdev);
 	r600_irq_fini(rdev);
+	if (rdev->has_uvd) {
+		uvd_v1_0_fini(rdev);
+		radeon_uvd_fini(rdev);
+	}
 	radeon_wb_fini(rdev);
 	radeon_ib_pool_fini(rdev);
 	radeon_irq_kms_fini(rdev);
@@ -3235,7 +3364,7 @@ int r600_ih_ring_alloc(struct radeon_device *rdev)
 		r = radeon_bo_create(rdev, rdev->ih.ring_size,
 				     PAGE_SIZE, true,
 				     RADEON_GEM_DOMAIN_GTT, 0,
-				     NULL, &rdev->ih.ring_obj);
+				     NULL, NULL, &rdev->ih.ring_obj);
 		if (r) {
 			DRM_ERROR("radeon: failed to create ih ring buffer (%d).\n", r);
 			return r;
diff --git a/drivers/gpu/drm/radeon/r600_audio.c b/drivers/gpu/drm/radeon/r600_audio.c
deleted file mode 100644
index bffac10c4296..000000000000
--- a/drivers/gpu/drm/radeon/r600_audio.c
+++ /dev/null
@@ -1,207 +0,0 @@
-/*
- * Copyright 2008 Advanced Micro Devices, Inc.
- * Copyright 2008 Red Hat Inc.
- * Copyright 2009 Christian König.
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * the rights to use, copy, modify, merge, publish, distribute, sublicense,
- * and/or sell copies of the Software, and to permit persons to whom the
- * Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in
- * all copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
- * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
- * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
- * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
- * OTHER DEALINGS IN THE SOFTWARE.
- *
- * Authors: Christian König
- */
-#include <drm/drmP.h>
-#include "radeon.h"
-#include "radeon_reg.h"
-#include "radeon_asic.h"
-#include "atom.h"
-
-/*
- * check if enc_priv stores radeon_encoder_atom_dig
- */
-static bool radeon_dig_encoder(struct drm_encoder *encoder)
-{
-	struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder);
-	switch (radeon_encoder->encoder_id) {
-	case ENCODER_OBJECT_ID_INTERNAL_LVDS:
-	case ENCODER_OBJECT_ID_INTERNAL_TMDS1:
-	case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_TMDS1:
-	case ENCODER_OBJECT_ID_INTERNAL_LVTM1:
-	case ENCODER_OBJECT_ID_INTERNAL_DVO1:
-	case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_DVO1:
-	case ENCODER_OBJECT_ID_INTERNAL_DDI:
-	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY:
-	case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_LVTMA:
-	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY1:
-	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY2:
-		return true;
-	}
-	return false;
-}
-
-/*
- * check if the chipset is supported
- */
-static int r600_audio_chipset_supported(struct radeon_device *rdev)
-{
-	return ASIC_IS_DCE2(rdev) && !ASIC_IS_NODCE(rdev);
-}
-
-struct r600_audio_pin r600_audio_status(struct radeon_device *rdev)
-{
-	struct r600_audio_pin status;
-	uint32_t value;
-
-	value = RREG32(R600_AUDIO_RATE_BPS_CHANNEL);
-
-	/* number of channels */
-	status.channels = (value & 0x7) + 1;
-
-	/* bits per sample */
-	switch ((value & 0xF0) >> 4) {
-	case 0x0:
-		status.bits_per_sample = 8;
-		break;
-	case 0x1:
-		status.bits_per_sample = 16;
-		break;
-	case 0x2:
-		status.bits_per_sample = 20;
-		break;
-	case 0x3:
-		status.bits_per_sample = 24;
-		break;
-	case 0x4:
-		status.bits_per_sample = 32;
-		break;
-	default:
-		dev_err(rdev->dev, "Unknown bits per sample 0x%x, using 16\n",
-			(int)value);
-		status.bits_per_sample = 16;
-	}
-
-	/* current sampling rate in HZ */
-	if (value & 0x4000)
-		status.rate = 44100;
-	else
-		status.rate = 48000;
-	status.rate *= ((value >> 11) & 0x7) + 1;
-	status.rate /= ((value >> 8) & 0x7) + 1;
-
-	value = RREG32(R600_AUDIO_STATUS_BITS);
-
-	/* iec 60958 status bits */
-	status.status_bits = value & 0xff;
-
-	/* iec 60958 category code */
-	status.category_code = (value >> 8) & 0xff;
-
-	return status;
-}
-
-/*
- * update all hdmi interfaces with current audio parameters
- */
-void r600_audio_update_hdmi(struct work_struct *work)
-{
-	struct radeon_device *rdev = container_of(work, struct radeon_device,
-						  audio_work);
-	struct drm_device *dev = rdev->ddev;
-	struct r600_audio_pin audio_status = r600_audio_status(rdev);
-	struct drm_encoder *encoder;
-	bool changed = false;
-
-	if (rdev->audio.pin[0].channels != audio_status.channels ||
-	    rdev->audio.pin[0].rate != audio_status.rate ||
-	    rdev->audio.pin[0].bits_per_sample != audio_status.bits_per_sample ||
-	    rdev->audio.pin[0].status_bits != audio_status.status_bits ||
-	    rdev->audio.pin[0].category_code != audio_status.category_code) {
-		rdev->audio.pin[0] = audio_status;
-		changed = true;
-	}
-
-	list_for_each_entry(encoder, &dev->mode_config.encoder_list, head) {
-		if (!radeon_dig_encoder(encoder))
-			continue;
-		if (changed || r600_hdmi_buffer_status_changed(encoder))
-			r600_hdmi_update_audio_settings(encoder);
-	}
-}
-
-/* enable the audio stream */
-void r600_audio_enable(struct radeon_device *rdev,
-		       struct r600_audio_pin *pin,
-		       bool enable)
-{
-	u32 value = 0;
-
-	if (!pin)
-		return;
-
-	if (ASIC_IS_DCE4(rdev)) {
-		if (enable) {
-			value |= 0x81000000; /* Required to enable audio */
-			value |= 0x0e1000f0; /* fglrx sets that too */
-		}
-		WREG32(EVERGREEN_AUDIO_ENABLE, value);
-	} else {
-		WREG32_P(R600_AUDIO_ENABLE,
-			 enable ? 0x81000000 : 0x0, ~0x81000000);
-	}
-}
-
-/*
- * initialize the audio vars
- */
-int r600_audio_init(struct radeon_device *rdev)
-{
-	if (!radeon_audio || !r600_audio_chipset_supported(rdev))
-		return 0;
-
-	rdev->audio.enabled = true;
-
-	rdev->audio.num_pins = 1;
-	rdev->audio.pin[0].channels = -1;
-	rdev->audio.pin[0].rate = -1;
-	rdev->audio.pin[0].bits_per_sample = -1;
-	rdev->audio.pin[0].status_bits = 0;
-	rdev->audio.pin[0].category_code = 0;
-	rdev->audio.pin[0].id = 0;
-	/* disable audio.  it will be set up later */
-	r600_audio_enable(rdev, &rdev->audio.pin[0], false);
-
-	return 0;
-}
-
-/*
- * release the audio timer
- * TODO: How to do this correctly on SMP systems?
- */
-void r600_audio_fini(struct radeon_device *rdev)
-{
-	if (!rdev->audio.enabled)
-		return;
-
-	r600_audio_enable(rdev, &rdev->audio.pin[0], false);
-
-	rdev->audio.enabled = false;
-}
-
-struct r600_audio_pin *r600_audio_get_pin(struct radeon_device *rdev)
-{
-	/* only one pin on 6xx-NI */
-	return &rdev->audio.pin[0];
-}
diff --git a/drivers/gpu/drm/radeon/r600_cp.c b/drivers/gpu/drm/radeon/r600_cp.c
index 8c9b7e26533c..09e3f39925fa 100644
--- a/drivers/gpu/drm/radeon/r600_cp.c
+++ b/drivers/gpu/drm/radeon/r600_cp.c
@@ -1949,15 +1949,15 @@ int r600_do_cleanup_cp(struct drm_device *dev)
 #if __OS_HAS_AGP
 	if (dev_priv->flags & RADEON_IS_AGP) {
 		if (dev_priv->cp_ring != NULL) {
-			drm_core_ioremapfree(dev_priv->cp_ring, dev);
+			drm_legacy_ioremapfree(dev_priv->cp_ring, dev);
 			dev_priv->cp_ring = NULL;
 		}
 		if (dev_priv->ring_rptr != NULL) {
-			drm_core_ioremapfree(dev_priv->ring_rptr, dev);
+			drm_legacy_ioremapfree(dev_priv->ring_rptr, dev);
 			dev_priv->ring_rptr = NULL;
 		}
 		if (dev->agp_buffer_map != NULL) {
-			drm_core_ioremapfree(dev->agp_buffer_map, dev);
+			drm_legacy_ioremapfree(dev->agp_buffer_map, dev);
 			dev->agp_buffer_map = NULL;
 		}
 	} else
@@ -1968,7 +1968,7 @@ int r600_do_cleanup_cp(struct drm_device *dev)
 			r600_page_table_cleanup(dev, &dev_priv->gart_info);
 
 		if (dev_priv->gart_info.gart_table_location == DRM_ATI_GART_FB) {
-			drm_core_ioremapfree(&dev_priv->gart_info.mapping, dev);
+			drm_legacy_ioremapfree(&dev_priv->gart_info.mapping, dev);
 			dev_priv->gart_info.addr = NULL;
 		}
 	}
@@ -2052,27 +2052,27 @@ int r600_do_init_cp(struct drm_device *dev, drm_radeon_init_t *init,
 	dev_priv->buffers_offset = init->buffers_offset;
 	dev_priv->gart_textures_offset = init->gart_textures_offset;
 
-	master_priv->sarea = drm_getsarea(dev);
+	master_priv->sarea = drm_legacy_getsarea(dev);
 	if (!master_priv->sarea) {
 		DRM_ERROR("could not find sarea!\n");
 		r600_do_cleanup_cp(dev);
 		return -EINVAL;
 	}
 
-	dev_priv->cp_ring = drm_core_findmap(dev, init->ring_offset);
+	dev_priv->cp_ring = drm_legacy_findmap(dev, init->ring_offset);
 	if (!dev_priv->cp_ring) {
 		DRM_ERROR("could not find cp ring region!\n");
 		r600_do_cleanup_cp(dev);
 		return -EINVAL;
 	}
-	dev_priv->ring_rptr = drm_core_findmap(dev, init->ring_rptr_offset);
+	dev_priv->ring_rptr = drm_legacy_findmap(dev, init->ring_rptr_offset);
 	if (!dev_priv->ring_rptr) {
 		DRM_ERROR("could not find ring read pointer!\n");
 		r600_do_cleanup_cp(dev);
 		return -EINVAL;
 	}
 	dev->agp_buffer_token = init->buffers_offset;
-	dev->agp_buffer_map = drm_core_findmap(dev, init->buffers_offset);
+	dev->agp_buffer_map = drm_legacy_findmap(dev, init->buffers_offset);
 	if (!dev->agp_buffer_map) {
 		DRM_ERROR("could not find dma buffer region!\n");
 		r600_do_cleanup_cp(dev);
@@ -2081,7 +2081,7 @@ int r600_do_init_cp(struct drm_device *dev, drm_radeon_init_t *init,
 
 	if (init->gart_textures_offset) {
 		dev_priv->gart_textures =
-		    drm_core_findmap(dev, init->gart_textures_offset);
+		    drm_legacy_findmap(dev, init->gart_textures_offset);
 		if (!dev_priv->gart_textures) {
 			DRM_ERROR("could not find GART texture region!\n");
 			r600_do_cleanup_cp(dev);
@@ -2092,9 +2092,9 @@ int r600_do_init_cp(struct drm_device *dev, drm_radeon_init_t *init,
 #if __OS_HAS_AGP
 	/* XXX */
 	if (dev_priv->flags & RADEON_IS_AGP) {
-		drm_core_ioremap_wc(dev_priv->cp_ring, dev);
-		drm_core_ioremap_wc(dev_priv->ring_rptr, dev);
-		drm_core_ioremap_wc(dev->agp_buffer_map, dev);
+		drm_legacy_ioremap_wc(dev_priv->cp_ring, dev);
+		drm_legacy_ioremap_wc(dev_priv->ring_rptr, dev);
+		drm_legacy_ioremap_wc(dev->agp_buffer_map, dev);
 		if (!dev_priv->cp_ring->handle ||
 		    !dev_priv->ring_rptr->handle ||
 		    !dev->agp_buffer_map->handle) {
@@ -2235,7 +2235,7 @@ int r600_do_init_cp(struct drm_device *dev, drm_radeon_init_t *init,
 		dev_priv->gart_info.mapping.size =
 			dev_priv->gart_info.table_size;
 
-		drm_core_ioremap_wc(&dev_priv->gart_info.mapping, dev);
+		drm_legacy_ioremap_wc(&dev_priv->gart_info.mapping, dev);
 		if (!dev_priv->gart_info.mapping.handle) {
 			DRM_ERROR("ioremap failed.\n");
 			r600_do_cleanup_cp(dev);
diff --git a/drivers/gpu/drm/radeon/r600_cs.c b/drivers/gpu/drm/radeon/r600_cs.c
index c47537a1ddba..acc1f99c84d9 100644
--- a/drivers/gpu/drm/radeon/r600_cs.c
+++ b/drivers/gpu/drm/radeon/r600_cs.c
@@ -969,7 +969,7 @@ static int r600_cs_parse_packet0(struct radeon_cs_parser *p,
 static int r600_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
 {
 	struct r600_cs_track *track = (struct r600_cs_track *)p->track;
-	struct radeon_cs_reloc *reloc;
+	struct radeon_bo_list *reloc;
 	u32 m, i, tmp, *ib;
 	int r;
 
@@ -1626,7 +1626,7 @@ static bool r600_is_safe_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
 static int r600_packet3_check(struct radeon_cs_parser *p,
 				struct radeon_cs_packet *pkt)
 {
-	struct radeon_cs_reloc *reloc;
+	struct radeon_bo_list *reloc;
 	struct r600_cs_track *track;
 	volatile u32 *ib;
 	unsigned idx;
@@ -2316,7 +2316,7 @@ int r600_cs_parse(struct radeon_cs_parser *p)
 			p->track = NULL;
 			return r;
 		}
-	} while (p->idx < p->chunks[p->chunk_ib_idx].length_dw);
+	} while (p->idx < p->chunk_ib->length_dw);
 #if 0
 	for (r = 0; r < p->ib.length_dw; r++) {
 		printk(KERN_INFO "%05d  0x%08X\n", r, p->ib.ptr[r]);
@@ -2351,10 +2351,10 @@ static void r600_cs_parser_fini(struct radeon_cs_parser *parser, int error)
 
 static int r600_cs_parser_relocs_legacy(struct radeon_cs_parser *p)
 {
-	if (p->chunk_relocs_idx == -1) {
+	if (p->chunk_relocs == NULL) {
 		return 0;
 	}
-	p->relocs = kzalloc(sizeof(struct radeon_cs_reloc), GFP_KERNEL);
+	p->relocs = kzalloc(sizeof(struct radeon_bo_list), GFP_KERNEL);
 	if (p->relocs == NULL) {
 		return -ENOMEM;
 	}
@@ -2398,7 +2398,7 @@ int r600_cs_legacy(struct drm_device *dev, void *data, struct drm_file *filp,
 	/* Copy the packet into the IB, the parser will read from the
 	 * input memory (cached) and write to the IB (which can be
 	 * uncached). */
-	ib_chunk = &parser.chunks[parser.chunk_ib_idx];
+	ib_chunk = parser.chunk_ib;
 	parser.ib.length_dw = ib_chunk->length_dw;
 	*l = parser.ib.length_dw;
 	if (copy_from_user(ib, ib_chunk->user_ptr, ib_chunk->length_dw * 4)) {
@@ -2435,24 +2435,24 @@ void r600_cs_legacy_init(void)
  * GPU offset using the provided start.
  **/
 int r600_dma_cs_next_reloc(struct radeon_cs_parser *p,
-			   struct radeon_cs_reloc **cs_reloc)
+			   struct radeon_bo_list **cs_reloc)
 {
 	struct radeon_cs_chunk *relocs_chunk;
 	unsigned idx;
 
 	*cs_reloc = NULL;
-	if (p->chunk_relocs_idx == -1) {
+	if (p->chunk_relocs == NULL) {
 		DRM_ERROR("No relocation chunk !\n");
 		return -EINVAL;
 	}
-	relocs_chunk = &p->chunks[p->chunk_relocs_idx];
+	relocs_chunk = p->chunk_relocs;
 	idx = p->dma_reloc_idx;
 	if (idx >= p->nrelocs) {
 		DRM_ERROR("Relocs at %d after relocations chunk end %d !\n",
 			  idx, p->nrelocs);
 		return -EINVAL;
 	}
-	*cs_reloc = p->relocs_ptr[idx];
+	*cs_reloc = &p->relocs[idx];
 	p->dma_reloc_idx++;
 	return 0;
 }
@@ -2472,8 +2472,8 @@ int r600_dma_cs_next_reloc(struct radeon_cs_parser *p,
  **/
 int r600_dma_cs_parse(struct radeon_cs_parser *p)
 {
-	struct radeon_cs_chunk *ib_chunk = &p->chunks[p->chunk_ib_idx];
-	struct radeon_cs_reloc *src_reloc, *dst_reloc;
+	struct radeon_cs_chunk *ib_chunk = p->chunk_ib;
+	struct radeon_bo_list *src_reloc, *dst_reloc;
 	u32 header, cmd, count, tiled;
 	volatile u32 *ib = p->ib.ptr;
 	u32 idx, idx_value;
@@ -2619,7 +2619,7 @@ int r600_dma_cs_parse(struct radeon_cs_parser *p)
 			DRM_ERROR("Unknown packet type %d at %d !\n", cmd, idx);
 			return -EINVAL;
 		}
-	} while (p->idx < p->chunks[p->chunk_ib_idx].length_dw);
+	} while (p->idx < p->chunk_ib->length_dw);
 #if 0
 	for (r = 0; r < p->ib->length_dw; r++) {
 		printk(KERN_INFO "%05d  0x%08X\n", r, p->ib.ptr[r]);
diff --git a/drivers/gpu/drm/radeon/r600_dma.c b/drivers/gpu/drm/radeon/r600_dma.c
index a908daa006d2..d2dd29ab24fa 100644
--- a/drivers/gpu/drm/radeon/r600_dma.c
+++ b/drivers/gpu/drm/radeon/r600_dma.c
@@ -232,16 +232,19 @@ int r600_dma_ring_test(struct radeon_device *rdev,
 {
 	unsigned i;
 	int r;
-	void __iomem *ptr = (void *)rdev->vram_scratch.ptr;
+	unsigned index;
 	u32 tmp;
+	u64 gpu_addr;
 
-	if (!ptr) {
-		DRM_ERROR("invalid vram scratch pointer\n");
-		return -EINVAL;
-	}
+	if (ring->idx == R600_RING_TYPE_DMA_INDEX)
+		index = R600_WB_DMA_RING_TEST_OFFSET;
+	else
+		index = CAYMAN_WB_DMA1_RING_TEST_OFFSET;
+
+	gpu_addr = rdev->wb.gpu_addr + index;
 
 	tmp = 0xCAFEDEAD;
-	writel(tmp, ptr);
+	rdev->wb.wb[index/4] = cpu_to_le32(tmp);
 
 	r = radeon_ring_lock(rdev, ring, 4);
 	if (r) {
@@ -249,13 +252,13 @@ int r600_dma_ring_test(struct radeon_device *rdev,
 		return r;
 	}
 	radeon_ring_write(ring, DMA_PACKET(DMA_PACKET_WRITE, 0, 0, 1));
-	radeon_ring_write(ring, rdev->vram_scratch.gpu_addr & 0xfffffffc);
-	radeon_ring_write(ring, upper_32_bits(rdev->vram_scratch.gpu_addr) & 0xff);
+	radeon_ring_write(ring, lower_32_bits(gpu_addr));
+	radeon_ring_write(ring, upper_32_bits(gpu_addr) & 0xff);
 	radeon_ring_write(ring, 0xDEADBEEF);
 	radeon_ring_unlock_commit(rdev, ring, false);
 
 	for (i = 0; i < rdev->usec_timeout; i++) {
-		tmp = readl(ptr);
+		tmp = le32_to_cpu(rdev->wb.wb[index/4]);
 		if (tmp == 0xDEADBEEF)
 			break;
 		DRM_UDELAY(1);
@@ -335,17 +338,17 @@ int r600_dma_ib_test(struct radeon_device *rdev, struct radeon_ring *ring)
 {
 	struct radeon_ib ib;
 	unsigned i;
+	unsigned index;
 	int r;
-	void __iomem *ptr = (void *)rdev->vram_scratch.ptr;
 	u32 tmp = 0;
+	u64 gpu_addr;
 
-	if (!ptr) {
-		DRM_ERROR("invalid vram scratch pointer\n");
-		return -EINVAL;
-	}
+	if (ring->idx == R600_RING_TYPE_DMA_INDEX)
+		index = R600_WB_DMA_RING_TEST_OFFSET;
+	else
+		index = CAYMAN_WB_DMA1_RING_TEST_OFFSET;
 
-	tmp = 0xCAFEDEAD;
-	writel(tmp, ptr);
+	gpu_addr = rdev->wb.gpu_addr + index;
 
 	r = radeon_ib_get(rdev, ring->idx, &ib, NULL, 256);
 	if (r) {
@@ -354,8 +357,8 @@ int r600_dma_ib_test(struct radeon_device *rdev, struct radeon_ring *ring)
 	}
 
 	ib.ptr[0] = DMA_PACKET(DMA_PACKET_WRITE, 0, 0, 1);
-	ib.ptr[1] = rdev->vram_scratch.gpu_addr & 0xfffffffc;
-	ib.ptr[2] = upper_32_bits(rdev->vram_scratch.gpu_addr) & 0xff;
+	ib.ptr[1] = lower_32_bits(gpu_addr);
+	ib.ptr[2] = upper_32_bits(gpu_addr) & 0xff;
 	ib.ptr[3] = 0xDEADBEEF;
 	ib.length_dw = 4;
 
@@ -371,7 +374,7 @@ int r600_dma_ib_test(struct radeon_device *rdev, struct radeon_ring *ring)
 		return r;
 	}
 	for (i = 0; i < rdev->usec_timeout; i++) {
-		tmp = readl(ptr);
+		tmp = le32_to_cpu(rdev->wb.wb[index/4]);
 		if (tmp == 0xDEADBEEF)
 			break;
 		DRM_UDELAY(1);
@@ -427,41 +430,38 @@ void r600_dma_ring_ib_execute(struct radeon_device *rdev, struct radeon_ib *ib)
  * @src_offset: src GPU address
  * @dst_offset: dst GPU address
  * @num_gpu_pages: number of GPU pages to xfer
- * @fence: radeon fence object
+ * @resv: reservation object to sync to
  *
  * Copy GPU paging using the DMA engine (r6xx).
  * Used by the radeon ttm implementation to move pages if
  * registered as the asic copy callback.
  */
-int r600_copy_dma(struct radeon_device *rdev,
-		  uint64_t src_offset, uint64_t dst_offset,
-		  unsigned num_gpu_pages,
-		  struct radeon_fence **fence)
+struct radeon_fence *r600_copy_dma(struct radeon_device *rdev,
+				   uint64_t src_offset, uint64_t dst_offset,
+				   unsigned num_gpu_pages,
+				   struct reservation_object *resv)
 {
-	struct radeon_semaphore *sem = NULL;
+	struct radeon_fence *fence;
+	struct radeon_sync sync;
 	int ring_index = rdev->asic->copy.dma_ring_index;
 	struct radeon_ring *ring = &rdev->ring[ring_index];
 	u32 size_in_dw, cur_size_in_dw;
 	int i, num_loops;
 	int r = 0;
 
-	r = radeon_semaphore_create(rdev, &sem);
-	if (r) {
-		DRM_ERROR("radeon: moving bo (%d).\n", r);
-		return r;
-	}
+	radeon_sync_create(&sync);
 
 	size_in_dw = (num_gpu_pages << RADEON_GPU_PAGE_SHIFT) / 4;
 	num_loops = DIV_ROUND_UP(size_in_dw, 0xFFFE);
 	r = radeon_ring_lock(rdev, ring, num_loops * 4 + 8);
 	if (r) {
 		DRM_ERROR("radeon: moving bo (%d).\n", r);
-		radeon_semaphore_free(rdev, &sem, NULL);
-		return r;
+		radeon_sync_free(rdev, &sync, NULL);
+		return ERR_PTR(r);
 	}
 
-	radeon_semaphore_sync_to(sem, *fence);
-	radeon_semaphore_sync_rings(rdev, sem, ring->idx);
+	radeon_sync_resv(rdev, &sync, resv, false);
+	radeon_sync_rings(rdev, &sync, ring->idx);
 
 	for (i = 0; i < num_loops; i++) {
 		cur_size_in_dw = size_in_dw;
@@ -477,15 +477,15 @@ int r600_copy_dma(struct radeon_device *rdev,
 		dst_offset += cur_size_in_dw * 4;
 	}
 
-	r = radeon_fence_emit(rdev, fence, ring->idx);
+	r = radeon_fence_emit(rdev, &fence, ring->idx);
 	if (r) {
 		radeon_ring_unlock_undo(rdev, ring);
-		radeon_semaphore_free(rdev, &sem, NULL);
-		return r;
+		radeon_sync_free(rdev, &sync, NULL);
+		return ERR_PTR(r);
 	}
 
 	radeon_ring_unlock_commit(rdev, ring, false);
-	radeon_semaphore_free(rdev, &sem, *fence);
+	radeon_sync_free(rdev, &sync, fence);
 
-	return r;
+	return fence;
 }
diff --git a/drivers/gpu/drm/radeon/r600_dpm.c b/drivers/gpu/drm/radeon/r600_dpm.c
index 9c61b74ef441..843b65f46ece 100644
--- a/drivers/gpu/drm/radeon/r600_dpm.c
+++ b/drivers/gpu/drm/radeon/r600_dpm.c
@@ -24,6 +24,7 @@
 
 #include "drmP.h"
 #include "radeon.h"
+#include "radeon_asic.h"
 #include "r600d.h"
 #include "r600_dpm.h"
 #include "atom.h"
@@ -810,6 +811,7 @@ union power_info {
 union fan_info {
 	struct _ATOM_PPLIB_FANTABLE fan;
 	struct _ATOM_PPLIB_FANTABLE2 fan2;
+	struct _ATOM_PPLIB_FANTABLE3 fan3;
 };
 
 static int r600_parse_clk_voltage_dep_table(struct radeon_clock_voltage_dependency_table *radeon_table,
@@ -899,6 +901,14 @@ int r600_parse_extended_power_table(struct radeon_device *rdev)
 			else
 				rdev->pm.dpm.fan.t_max = 10900;
 			rdev->pm.dpm.fan.cycle_delay = 100000;
+			if (fan_info->fan.ucFanTableFormat >= 3) {
+				rdev->pm.dpm.fan.control_mode = fan_info->fan3.ucFanControlMode;
+				rdev->pm.dpm.fan.default_max_fan_pwm =
+					le16_to_cpu(fan_info->fan3.usFanPWMMax);
+				rdev->pm.dpm.fan.default_fan_output_sensitivity = 4836;
+				rdev->pm.dpm.fan.fan_output_sensitivity =
+					le16_to_cpu(fan_info->fan3.usFanOutputSensitivity);
+			}
 			rdev->pm.dpm.fan.ucode_fan_control = true;
 		}
 	}
@@ -1255,7 +1265,7 @@ int r600_parse_extended_power_table(struct radeon_device *rdev)
 					(mode_info->atom_context->bios + data_offset +
 					 le16_to_cpu(ext_hdr->usPowerTuneTableOffset));
 				rdev->pm.dpm.dyn_state.cac_tdp_table->maximum_power_delivery_limit =
-					ppt->usMaximumPowerDeliveryLimit;
+					le16_to_cpu(ppt->usMaximumPowerDeliveryLimit);
 				pt = &ppt->power_tune_table;
 			} else {
 				ATOM_PPLIB_POWERTUNE_Table *ppt = (ATOM_PPLIB_POWERTUNE_Table *)
diff --git a/drivers/gpu/drm/radeon/r600_dpm.h b/drivers/gpu/drm/radeon/r600_dpm.h
index 46b9d2a03018..bd499d749bc9 100644
--- a/drivers/gpu/drm/radeon/r600_dpm.h
+++ b/drivers/gpu/drm/radeon/r600_dpm.h
@@ -96,6 +96,9 @@
 #define R600_TEMP_RANGE_MIN (90 * 1000)
 #define R600_TEMP_RANGE_MAX (120 * 1000)
 
+#define FDO_PWM_MODE_STATIC  1
+#define FDO_PWM_MODE_STATIC_RPM 5
+
 enum r600_power_level {
 	R600_POWER_LEVEL_LOW = 0,
 	R600_POWER_LEVEL_MEDIUM = 1,
diff --git a/drivers/gpu/drm/radeon/r600_hdmi.c b/drivers/gpu/drm/radeon/r600_hdmi.c
index 26ef8ced6f89..b90dc0eb08e6 100644
--- a/drivers/gpu/drm/radeon/r600_hdmi.c
+++ b/drivers/gpu/drm/radeon/r600_hdmi.c
@@ -72,6 +72,169 @@ static const struct radeon_hdmi_acr r600_hdmi_predefined_acr[] = {
 
 
 /*
+ * check if the chipset is supported
+ */
+static int r600_audio_chipset_supported(struct radeon_device *rdev)
+{
+	return ASIC_IS_DCE2(rdev) && !ASIC_IS_NODCE(rdev);
+}
+
+static struct r600_audio_pin r600_audio_status(struct radeon_device *rdev)
+{
+	struct r600_audio_pin status;
+	uint32_t value;
+
+	value = RREG32(R600_AUDIO_RATE_BPS_CHANNEL);
+
+	/* number of channels */
+	status.channels = (value & 0x7) + 1;
+
+	/* bits per sample */
+	switch ((value & 0xF0) >> 4) {
+	case 0x0:
+		status.bits_per_sample = 8;
+		break;
+	case 0x1:
+		status.bits_per_sample = 16;
+		break;
+	case 0x2:
+		status.bits_per_sample = 20;
+		break;
+	case 0x3:
+		status.bits_per_sample = 24;
+		break;
+	case 0x4:
+		status.bits_per_sample = 32;
+		break;
+	default:
+		dev_err(rdev->dev, "Unknown bits per sample 0x%x, using 16\n",
+			(int)value);
+		status.bits_per_sample = 16;
+	}
+
+	/* current sampling rate in HZ */
+	if (value & 0x4000)
+		status.rate = 44100;
+	else
+		status.rate = 48000;
+	status.rate *= ((value >> 11) & 0x7) + 1;
+	status.rate /= ((value >> 8) & 0x7) + 1;
+
+	value = RREG32(R600_AUDIO_STATUS_BITS);
+
+	/* iec 60958 status bits */
+	status.status_bits = value & 0xff;
+
+	/* iec 60958 category code */
+	status.category_code = (value >> 8) & 0xff;
+
+	return status;
+}
+
+/*
+ * update all hdmi interfaces with current audio parameters
+ */
+void r600_audio_update_hdmi(struct work_struct *work)
+{
+	struct radeon_device *rdev = container_of(work, struct radeon_device,
+						  audio_work);
+	struct drm_device *dev = rdev->ddev;
+	struct r600_audio_pin audio_status = r600_audio_status(rdev);
+	struct drm_encoder *encoder;
+	bool changed = false;
+
+	if (rdev->audio.pin[0].channels != audio_status.channels ||
+	    rdev->audio.pin[0].rate != audio_status.rate ||
+	    rdev->audio.pin[0].bits_per_sample != audio_status.bits_per_sample ||
+	    rdev->audio.pin[0].status_bits != audio_status.status_bits ||
+	    rdev->audio.pin[0].category_code != audio_status.category_code) {
+		rdev->audio.pin[0] = audio_status;
+		changed = true;
+	}
+
+	list_for_each_entry(encoder, &dev->mode_config.encoder_list, head) {
+		if (!radeon_encoder_is_digital(encoder))
+			continue;
+		if (changed || r600_hdmi_buffer_status_changed(encoder))
+			r600_hdmi_update_audio_settings(encoder);
+	}
+}
+
+/* enable the audio stream */
+void r600_audio_enable(struct radeon_device *rdev,
+		       struct r600_audio_pin *pin,
+		       u8 enable_mask)
+{
+	u32 tmp = RREG32(AZ_HOT_PLUG_CONTROL);
+
+	if (!pin)
+		return;
+
+	if (enable_mask) {
+		tmp |= AUDIO_ENABLED;
+		if (enable_mask & 1)
+			tmp |= PIN0_AUDIO_ENABLED;
+		if (enable_mask & 2)
+			tmp |= PIN1_AUDIO_ENABLED;
+		if (enable_mask & 4)
+			tmp |= PIN2_AUDIO_ENABLED;
+		if (enable_mask & 8)
+			tmp |= PIN3_AUDIO_ENABLED;
+	} else {
+		tmp &= ~(AUDIO_ENABLED |
+			 PIN0_AUDIO_ENABLED |
+			 PIN1_AUDIO_ENABLED |
+			 PIN2_AUDIO_ENABLED |
+			 PIN3_AUDIO_ENABLED);
+	}
+
+	WREG32(AZ_HOT_PLUG_CONTROL, tmp);
+}
+
+/*
+ * initialize the audio vars
+ */
+int r600_audio_init(struct radeon_device *rdev)
+{
+	if (!radeon_audio || !r600_audio_chipset_supported(rdev))
+		return 0;
+
+	rdev->audio.enabled = true;
+
+	rdev->audio.num_pins = 1;
+	rdev->audio.pin[0].channels = -1;
+	rdev->audio.pin[0].rate = -1;
+	rdev->audio.pin[0].bits_per_sample = -1;
+	rdev->audio.pin[0].status_bits = 0;
+	rdev->audio.pin[0].category_code = 0;
+	rdev->audio.pin[0].id = 0;
+	/* disable audio.  it will be set up later */
+	r600_audio_enable(rdev, &rdev->audio.pin[0], 0);
+
+	return 0;
+}
+
+/*
+ * release the audio timer
+ * TODO: How to do this correctly on SMP systems?
+ */
+void r600_audio_fini(struct radeon_device *rdev)
+{
+	if (!rdev->audio.enabled)
+		return;
+
+	r600_audio_enable(rdev, &rdev->audio.pin[0], 0);
+
+	rdev->audio.enabled = false;
+}
+
+struct r600_audio_pin *r600_audio_get_pin(struct radeon_device *rdev)
+{
+	/* only one pin on 6xx-NI */
+	return &rdev->audio.pin[0];
+}
+
+/*
  * calculate CTS and N values if they are not found in the table
  */
 static void r600_hdmi_calc_cts(uint32_t clock, int *CTS, int *N, int freq)
@@ -357,7 +520,7 @@ void r600_hdmi_setmode(struct drm_encoder *encoder, struct drm_display_mode *mod
 
 	/* disable audio prior to setting up hw */
 	dig->afmt->pin = r600_audio_get_pin(rdev);
-	r600_audio_enable(rdev, dig->afmt->pin, false);
+	r600_audio_enable(rdev, dig->afmt->pin, 0xf);
 
 	r600_audio_set_dto(encoder, mode->clock);
 
@@ -443,7 +606,7 @@ void r600_hdmi_setmode(struct drm_encoder *encoder, struct drm_display_mode *mod
 	WREG32(HDMI0_RAMP_CONTROL3 + offset, 0x00000001);
 
 	/* enable audio after to setting up hw */
-	r600_audio_enable(rdev, dig->afmt->pin, true);
+	r600_audio_enable(rdev, dig->afmt->pin, 0xf);
 }
 
 /**
@@ -528,6 +691,11 @@ void r600_hdmi_enable(struct drm_encoder *encoder, bool enable)
 	if (!enable && !dig->afmt->enabled)
 		return;
 
+	if (!enable && dig->afmt->pin) {
+		r600_audio_enable(rdev, dig->afmt->pin, 0);
+		dig->afmt->pin = NULL;
+	}
+
 	/* Older chipsets require setting HDMI and routing manually */
 	if (!ASIC_IS_DCE3(rdev)) {
 		if (enable)
diff --git a/drivers/gpu/drm/radeon/r600d.h b/drivers/gpu/drm/radeon/r600d.h
index 31e1052ad3e3..1e8495cca41e 100644
--- a/drivers/gpu/drm/radeon/r600d.h
+++ b/drivers/gpu/drm/radeon/r600d.h
@@ -323,11 +323,12 @@
 #define	HDP_TILING_CONFIG				0x2F3C
 #define HDP_DEBUG1                                      0x2F34
 
+#define MC_CONFIG					0x2000
 #define MC_VM_AGP_TOP					0x2184
 #define MC_VM_AGP_BOT					0x2188
 #define	MC_VM_AGP_BASE					0x218C
 #define MC_VM_FB_LOCATION				0x2180
-#define MC_VM_L1_TLB_MCD_RD_A_CNTL			0x219C
+#define MC_VM_L1_TLB_MCB_RD_UVD_CNTL			0x2124
 #define 	ENABLE_L1_TLB					(1 << 0)
 #define		ENABLE_L1_FRAGMENT_PROCESSING			(1 << 1)
 #define		ENABLE_L1_STRICT_ORDERING			(1 << 2)
@@ -347,12 +348,14 @@
 #define		EFFECTIVE_L1_QUEUE_SIZE(x)			(((x) & 7) << 15)
 #define		EFFECTIVE_L1_QUEUE_SIZE_MASK			0x00038000
 #define		EFFECTIVE_L1_QUEUE_SIZE_SHIFT			15
+#define MC_VM_L1_TLB_MCD_RD_A_CNTL			0x219C
 #define MC_VM_L1_TLB_MCD_RD_B_CNTL			0x21A0
 #define MC_VM_L1_TLB_MCB_RD_GFX_CNTL			0x21FC
 #define MC_VM_L1_TLB_MCB_RD_HDP_CNTL			0x2204
 #define MC_VM_L1_TLB_MCB_RD_PDMA_CNTL			0x2208
 #define MC_VM_L1_TLB_MCB_RD_SEM_CNTL			0x220C
 #define	MC_VM_L1_TLB_MCB_RD_SYS_CNTL			0x2200
+#define MC_VM_L1_TLB_MCB_WR_UVD_CNTL			0x212c
 #define MC_VM_L1_TLB_MCD_WR_A_CNTL			0x21A4
 #define MC_VM_L1_TLB_MCD_WR_B_CNTL			0x21A8
 #define MC_VM_L1_TLB_MCB_WR_GFX_CNTL			0x2210
@@ -366,6 +369,8 @@
 #define MC_VM_SYSTEM_APERTURE_HIGH_ADDR			0x2194
 #define MC_VM_SYSTEM_APERTURE_DEFAULT_ADDR		0x2198
 
+#define RS_DQ_RD_RET_CONF				0x2348
+
 #define	PA_CL_ENHANCE					0x8A14
 #define		CLIP_VTX_REORDER_ENA				(1 << 0)
 #define		NUM_CLIP_SEQ(x)					((x) << 1)
@@ -922,6 +927,23 @@
 #       define TARGET_LINK_SPEED_MASK                     (0xf << 0)
 #       define SELECTABLE_DEEMPHASIS                      (1 << 6)
 
+/* Audio */
+#define AZ_HOT_PLUG_CONTROL               0x7300
+#       define AZ_FORCE_CODEC_WAKE        (1 << 0)
+#       define JACK_DETECTION_ENABLE      (1 << 4)
+#       define UNSOLICITED_RESPONSE_ENABLE (1 << 8)
+#       define CODEC_HOT_PLUG_ENABLE      (1 << 12)
+#       define AUDIO_ENABLED              (1 << 31)
+/* DCE3 adds */
+#       define PIN0_JACK_DETECTION_ENABLE (1 << 4)
+#       define PIN1_JACK_DETECTION_ENABLE (1 << 5)
+#       define PIN2_JACK_DETECTION_ENABLE (1 << 6)
+#       define PIN3_JACK_DETECTION_ENABLE (1 << 7)
+#       define PIN0_AUDIO_ENABLED         (1 << 24)
+#       define PIN1_AUDIO_ENABLED         (1 << 25)
+#       define PIN2_AUDIO_ENABLED         (1 << 26)
+#       define PIN3_AUDIO_ENABLED         (1 << 27)
+
 /* Audio clocks DCE 2.0/3.0 */
 #define AUDIO_DTO                         0x7340
 #       define AUDIO_DTO_PHASE(x)         (((x) & 0xffff) << 0)
@@ -1476,6 +1498,7 @@
 #define UVD_CGC_GATE					0xf4a8
 #define UVD_LMI_CTRL2					0xf4f4
 #define UVD_MASTINT_EN					0xf500
+#define UVD_FW_START					0xf51C
 #define UVD_LMI_ADDR_EXT				0xf594
 #define UVD_LMI_CTRL					0xf598
 #define UVD_LMI_SWAP_CNTL				0xf5b4
@@ -1488,6 +1511,13 @@
 #define UVD_MPC_SET_MUX					0xf5f4
 #define UVD_MPC_SET_ALU					0xf5f8
 
+#define UVD_VCPU_CACHE_OFFSET0				0xf608
+#define UVD_VCPU_CACHE_SIZE0				0xf60c
+#define UVD_VCPU_CACHE_OFFSET1				0xf610
+#define UVD_VCPU_CACHE_SIZE1				0xf614
+#define UVD_VCPU_CACHE_OFFSET2				0xf618
+#define UVD_VCPU_CACHE_SIZE2				0xf61c
+
 #define UVD_VCPU_CNTL					0xf660
 #define UVD_SOFT_RESET					0xf680
 #define		RBC_SOFT_RESET					(1<<0)
@@ -1517,9 +1547,35 @@
 
 #define UVD_CONTEXT_ID					0xf6f4
 
+/* rs780 only */
+#define	GFX_MACRO_BYPASS_CNTL				0x30c0
+#define		SPLL_BYPASS_CNTL			(1 << 0)
+#define		UPLL_BYPASS_CNTL			(1 << 1)
+
+#define CG_UPLL_FUNC_CNTL				0x7e0
+#	define UPLL_RESET_MASK				0x00000001
+#	define UPLL_SLEEP_MASK				0x00000002
+#	define UPLL_BYPASS_EN_MASK			0x00000004
 #	define UPLL_CTLREQ_MASK				0x00000008
+#	define UPLL_FB_DIV(x)				((x) << 4)
+#	define UPLL_FB_DIV_MASK				0x0000FFF0
+#	define UPLL_REF_DIV(x)				((x) << 16)
+#	define UPLL_REF_DIV_MASK			0x003F0000
+#	define UPLL_REFCLK_SRC_SEL_MASK			0x20000000
 #	define UPLL_CTLACK_MASK				0x40000000
 #	define UPLL_CTLACK2_MASK			0x80000000
+#define CG_UPLL_FUNC_CNTL_2				0x7e4
+#	define UPLL_SW_HILEN(x)				((x) << 0)
+#	define UPLL_SW_LOLEN(x)				((x) << 4)
+#	define UPLL_SW_HILEN2(x)			((x) << 8)
+#	define UPLL_SW_LOLEN2(x)			((x) << 12)
+#	define UPLL_DIVEN_MASK				0x00010000
+#	define UPLL_DIVEN2_MASK				0x00020000
+#	define UPLL_SW_MASK				0x0003FFFF
+#	define VCLK_SRC_SEL(x)				((x) << 20)
+#	define VCLK_SRC_SEL_MASK			0x01F00000
+#	define DCLK_SRC_SEL(x)				((x) << 25)
+#	define DCLK_SRC_SEL_MASK			0x3E000000
 
 /*
  * PM4
diff --git a/drivers/gpu/drm/radeon/radeon.h b/drivers/gpu/drm/radeon/radeon.h
index 3247bfd14410..54529b837afa 100644
--- a/drivers/gpu/drm/radeon/radeon.h
+++ b/drivers/gpu/drm/radeon/radeon.h
@@ -65,6 +65,8 @@
 #include <linux/list.h>
 #include <linux/kref.h>
 #include <linux/interval_tree.h>
+#include <linux/hashtable.h>
+#include <linux/fence.h>
 
 #include <ttm/ttm_bo_api.h>
 #include <ttm/ttm_bo_driver.h>
@@ -72,6 +74,8 @@
 #include <ttm/ttm_module.h>
 #include <ttm/ttm_execbuf_util.h>
 
+#include <drm/drm_gem.h>
+
 #include "radeon_family.h"
 #include "radeon_mode.h"
 #include "radeon_reg.h"
@@ -120,9 +124,6 @@ extern int radeon_backlight;
 #define RADEONFB_CONN_LIMIT			4
 #define RADEON_BIOS_NUM_SCRATCH			8
 
-/* fence seq are set to this number when signaled */
-#define RADEON_FENCE_SIGNALED_SEQ		0LL
-
 /* internal ring indices */
 /* r1xx+ has gfx CP ring */
 #define RADEON_RING_TYPE_GFX_INDEX		0
@@ -149,9 +150,6 @@ extern int radeon_backlight;
 /* number of hw syncs before falling back on blocking */
 #define RADEON_NUM_SYNCS			4
 
-/* number of hw syncs before falling back on blocking */
-#define RADEON_NUM_SYNCS			4
-
 /* hardcode those limit for now */
 #define RADEON_VA_IB_OFFSET			(1 << 20)
 #define RADEON_VA_RESERVED_SIZE			(8 << 20)
@@ -350,28 +348,33 @@ extern void evergreen_tiling_fields(unsigned tiling_flags, unsigned *bankw,
  * Fences.
  */
 struct radeon_fence_driver {
+	struct radeon_device		*rdev;
 	uint32_t			scratch_reg;
 	uint64_t			gpu_addr;
 	volatile uint32_t		*cpu_addr;
 	/* sync_seq is protected by ring emission lock */
 	uint64_t			sync_seq[RADEON_NUM_RINGS];
 	atomic64_t			last_seq;
-	bool				initialized;
+	bool				initialized, delayed_irq;
+	struct delayed_work		lockup_work;
 };
 
 struct radeon_fence {
-	struct radeon_device		*rdev;
-	struct kref			kref;
-	/* protected by radeon_fence.lock */
-	uint64_t			seq;
+	struct fence		base;
+
+	struct radeon_device	*rdev;
+	uint64_t		seq;
 	/* RB, DMA, etc. */
-	unsigned			ring;
+	unsigned		ring;
+	bool			is_vm_update;
+
+	wait_queue_t		fence_wake;
 };
 
 int radeon_fence_driver_start_ring(struct radeon_device *rdev, int ring);
 int radeon_fence_driver_init(struct radeon_device *rdev);
 void radeon_fence_driver_fini(struct radeon_device *rdev);
-void radeon_fence_driver_force_completion(struct radeon_device *rdev);
+void radeon_fence_driver_force_completion(struct radeon_device *rdev, int ring);
 int radeon_fence_emit(struct radeon_device *rdev, struct radeon_fence **fence, int ring);
 void radeon_fence_process(struct radeon_device *rdev, int ring);
 bool radeon_fence_signaled(struct radeon_fence *fence);
@@ -447,12 +450,22 @@ struct radeon_mman {
 #endif
 };
 
+struct radeon_bo_list {
+	struct radeon_bo		*robj;
+	struct ttm_validate_buffer	tv;
+	uint64_t			gpu_offset;
+	unsigned			prefered_domains;
+	unsigned			allowed_domains;
+	uint32_t			tiling_flags;
+};
+
 /* bo virtual address in a specific vm */
 struct radeon_bo_va {
 	/* protected by bo being reserved */
 	struct list_head		bo_list;
 	uint32_t			flags;
 	uint64_t			addr;
+	struct radeon_fence		*last_pt_update;
 	unsigned			ref_count;
 
 	/* protected by vm mutex */
@@ -469,7 +482,7 @@ struct radeon_bo {
 	struct list_head		list;
 	/* Protected by tbo.reserved */
 	u32				initial_domain;
-	u32				placements[3];
+	struct ttm_place		placements[4];
 	struct ttm_placement		placement;
 	struct ttm_buffer_object	tbo;
 	struct ttm_bo_kmap_obj		kmap;
@@ -489,6 +502,9 @@ struct radeon_bo {
 
 	struct ttm_bo_kmap_obj		dma_buf_vmap;
 	pid_t				pid;
+
+	struct radeon_mn		*mn;
+	struct interval_tree_node	mn_it;
 };
 #define gem_to_radeon_bo(gobj) container_of((gobj), struct radeon_bo, gem_base)
 
@@ -568,10 +584,9 @@ int radeon_mode_dumb_mmap(struct drm_file *filp,
  * Semaphores.
  */
 struct radeon_semaphore {
-	struct radeon_sa_bo		*sa_bo;
-	signed				waiters;
-	uint64_t			gpu_addr;
-	struct radeon_fence		*sync_to[RADEON_NUM_RINGS];
+	struct radeon_sa_bo	*sa_bo;
+	signed			waiters;
+	uint64_t		gpu_addr;
 };
 
 int radeon_semaphore_create(struct radeon_device *rdev,
@@ -580,16 +595,33 @@ bool radeon_semaphore_emit_signal(struct radeon_device *rdev, int ring,
 				  struct radeon_semaphore *semaphore);
 bool radeon_semaphore_emit_wait(struct radeon_device *rdev, int ring,
 				struct radeon_semaphore *semaphore);
-void radeon_semaphore_sync_to(struct radeon_semaphore *semaphore,
-			      struct radeon_fence *fence);
-int radeon_semaphore_sync_rings(struct radeon_device *rdev,
-				struct radeon_semaphore *semaphore,
-				int waiting_ring);
 void radeon_semaphore_free(struct radeon_device *rdev,
 			   struct radeon_semaphore **semaphore,
 			   struct radeon_fence *fence);
 
 /*
+ * Synchronization
+ */
+struct radeon_sync {
+	struct radeon_semaphore *semaphores[RADEON_NUM_SYNCS];
+	struct radeon_fence	*sync_to[RADEON_NUM_RINGS];
+	struct radeon_fence	*last_vm_update;
+};
+
+void radeon_sync_create(struct radeon_sync *sync);
+void radeon_sync_fence(struct radeon_sync *sync,
+		       struct radeon_fence *fence);
+int radeon_sync_resv(struct radeon_device *rdev,
+		     struct radeon_sync *sync,
+		     struct reservation_object *resv,
+		     bool shared);
+int radeon_sync_rings(struct radeon_device *rdev,
+		      struct radeon_sync *sync,
+		      int waiting_ring);
+void radeon_sync_free(struct radeon_device *rdev, struct radeon_sync *sync,
+		      struct radeon_fence *fence);
+
+/*
  * GART structures, functions & helpers
  */
 struct radeon_mc;
@@ -689,6 +721,10 @@ struct radeon_doorbell {
 
 int radeon_doorbell_get(struct radeon_device *rdev, u32 *page);
 void radeon_doorbell_free(struct radeon_device *rdev, u32 doorbell);
+void radeon_doorbell_get_kfd_info(struct radeon_device *rdev,
+				  phys_addr_t *aperture_base,
+				  size_t *aperture_size,
+				  size_t *start_offset);
 
 /*
  * IRQS.
@@ -702,7 +738,7 @@ struct radeon_flip_work {
 	uint64_t			base;
 	struct drm_pending_vblank_event *event;
 	struct radeon_bo		*old_rbo;
-	struct radeon_fence		*fence;
+	struct fence			*fence;
 };
 
 struct r500_irq_stat_regs {
@@ -780,6 +816,7 @@ struct radeon_irq {
 int radeon_irq_kms_init(struct radeon_device *rdev);
 void radeon_irq_kms_fini(struct radeon_device *rdev);
 void radeon_irq_kms_sw_irq_get(struct radeon_device *rdev, int ring);
+bool radeon_irq_kms_sw_irq_get_delayed(struct radeon_device *rdev, int ring);
 void radeon_irq_kms_sw_irq_put(struct radeon_device *rdev, int ring);
 void radeon_irq_kms_pflip_irq_get(struct radeon_device *rdev, int crtc);
 void radeon_irq_kms_pflip_irq_put(struct radeon_device *rdev, int crtc);
@@ -801,7 +838,7 @@ struct radeon_ib {
 	struct radeon_fence		*fence;
 	struct radeon_vm		*vm;
 	bool				is_const_ib;
-	struct radeon_semaphore		*semaphore;
+	struct radeon_sync		sync;
 };
 
 struct radeon_ring {
@@ -878,33 +915,40 @@ struct radeon_vm_pt {
 	uint64_t			addr;
 };
 
+struct radeon_vm_id {
+	unsigned		id;
+	uint64_t		pd_gpu_addr;
+	/* last flushed PD/PT update */
+	struct radeon_fence	*flushed_updates;
+	/* last use of vmid */
+	struct radeon_fence	*last_id_use;
+};
+
 struct radeon_vm {
-	struct rb_root			va;
-	unsigned			id;
+	struct mutex		mutex;
+
+	struct rb_root		va;
+
+	/* protecting invalidated and freed */
+	spinlock_t		status_lock;
 
 	/* BOs moved, but not yet updated in the PT */
-	struct list_head		invalidated;
+	struct list_head	invalidated;
 
 	/* BOs freed, but not yet updated in the PT */
-	struct list_head		freed;
+	struct list_head	freed;
 
 	/* contains the page directory */
-	struct radeon_bo		*page_directory;
-	uint64_t			pd_gpu_addr;
-	unsigned			max_pde_used;
+	struct radeon_bo	*page_directory;
+	unsigned		max_pde_used;
 
 	/* array of page tables, one for each page directory entry */
-	struct radeon_vm_pt		*page_tables;
+	struct radeon_vm_pt	*page_tables;
 
-	struct radeon_bo_va		*ib_bo_va;
+	struct radeon_bo_va	*ib_bo_va;
 
-	struct mutex			mutex;
-	/* last fence for cs using this vm */
-	struct radeon_fence		*fence;
-	/* last flush or NULL if we still need to flush */
-	struct radeon_fence		*last_flush;
-	/* last use of vmid */
-	struct radeon_fence		*last_id_use;
+	/* for id and flush management per ring */
+	struct radeon_vm_id	ids[RADEON_NUM_RINGS];
 };
 
 struct radeon_vm_manager {
@@ -1012,19 +1056,7 @@ void cayman_dma_fini(struct radeon_device *rdev);
 /*
  * CS.
  */
-struct radeon_cs_reloc {
-	struct drm_gem_object		*gobj;
-	struct radeon_bo		*robj;
-	struct ttm_validate_buffer	tv;
-	uint64_t			gpu_offset;
-	unsigned			prefered_domains;
-	unsigned			allowed_domains;
-	uint32_t			tiling_flags;
-	uint32_t			handle;
-};
-
 struct radeon_cs_chunk {
-	uint32_t		chunk_id;
 	uint32_t		length_dw;
 	uint32_t		*kdata;
 	void __user		*user_ptr;
@@ -1042,16 +1074,15 @@ struct radeon_cs_parser {
 	unsigned		idx;
 	/* relocations */
 	unsigned		nrelocs;
-	struct radeon_cs_reloc	*relocs;
-	struct radeon_cs_reloc	**relocs_ptr;
-	struct radeon_cs_reloc	*vm_bos;
+	struct radeon_bo_list	*relocs;
+	struct radeon_bo_list	*vm_bos;
 	struct list_head	validated;
 	unsigned		dma_reloc_idx;
 	/* indices of various chunks */
-	int			chunk_ib_idx;
-	int			chunk_relocs_idx;
-	int			chunk_flags_idx;
-	int			chunk_const_ib_idx;
+	struct radeon_cs_chunk  *chunk_ib;
+	struct radeon_cs_chunk  *chunk_relocs;
+	struct radeon_cs_chunk  *chunk_flags;
+	struct radeon_cs_chunk  *chunk_const_ib;
 	struct radeon_ib	ib;
 	struct radeon_ib	const_ib;
 	void			*track;
@@ -1065,7 +1096,7 @@ struct radeon_cs_parser {
 
 static inline u32 radeon_get_ib_value(struct radeon_cs_parser *p, int idx)
 {
-	struct radeon_cs_chunk *ibc = &p->chunks[p->chunk_ib_idx];
+	struct radeon_cs_chunk *ibc = p->chunk_ib;
 
 	if (ibc->kdata)
 		return ibc->kdata[idx];
@@ -1120,6 +1151,8 @@ struct radeon_wb {
 #define R600_WB_EVENT_OFFSET     3072
 #define CIK_WB_CP1_WPTR_OFFSET     3328
 #define CIK_WB_CP2_WPTR_OFFSET     3584
+#define R600_WB_DMA_RING_TEST_OFFSET 3588
+#define CAYMAN_WB_DMA1_RING_TEST_OFFSET 3592
 
 /**
  * struct radeon_pm - power management datas
@@ -1475,6 +1508,10 @@ struct radeon_dpm_fan {
 	u8 t_hyst;
 	u32 cycle_delay;
 	u16 t_max;
+	u8 control_mode;
+	u16 default_max_fan_pwm;
+	u16 default_fan_output_sensitivity;
+	u16 fan_output_sensitivity;
 	bool ucode_fan_control;
 };
 
@@ -1608,6 +1645,11 @@ struct radeon_pm {
 	/* internal thermal controller on rv6xx+ */
 	enum radeon_int_thermal_type int_thermal_type;
 	struct device	        *int_hwmon_dev;
+	/* fan control parameters */
+	bool                    no_fan;
+	u8                      fan_pulses_per_revolution;
+	u8                      fan_min_rpm;
+	u8                      fan_max_rpm;
 	/* dpm */
 	bool                    dpm_enabled;
 	struct radeon_dpm       dpm;
@@ -1642,7 +1684,8 @@ int radeon_uvd_get_create_msg(struct radeon_device *rdev, int ring,
 			      uint32_t handle, struct radeon_fence **fence);
 int radeon_uvd_get_destroy_msg(struct radeon_device *rdev, int ring,
 			       uint32_t handle, struct radeon_fence **fence);
-void radeon_uvd_force_into_uvd_segment(struct radeon_bo *rbo);
+void radeon_uvd_force_into_uvd_segment(struct radeon_bo *rbo,
+				       uint32_t allowed_domains);
 void radeon_uvd_free_handles(struct radeon_device *rdev,
 			     struct drm_file *filp);
 int radeon_uvd_cs_parse(struct radeon_cs_parser *parser);
@@ -1731,6 +1774,11 @@ void radeon_test_ring_sync(struct radeon_device *rdev,
 			   struct radeon_ring *cpB);
 void radeon_test_syncing(struct radeon_device *rdev);
 
+/*
+ * MMU Notifier
+ */
+int radeon_mn_register(struct radeon_bo *bo, unsigned long addr);
+void radeon_mn_unregister(struct radeon_bo *bo);
 
 /*
  * Debugfs
@@ -1764,7 +1812,8 @@ struct radeon_asic_ring {
 	void (*hdp_flush)(struct radeon_device *rdev, struct radeon_ring *ring);
 	bool (*emit_semaphore)(struct radeon_device *rdev, struct radeon_ring *cp,
 			       struct radeon_semaphore *semaphore, bool emit_wait);
-	void (*vm_flush)(struct radeon_device *rdev, int ridx, struct radeon_vm *vm);
+	void (*vm_flush)(struct radeon_device *rdev, struct radeon_ring *ring,
+			 unsigned vm_id, uint64_t pd_addr);
 
 	/* testing functions */
 	int (*ring_test)(struct radeon_device *rdev, struct radeon_ring *cp);
@@ -1845,24 +1894,24 @@ struct radeon_asic {
 	} display;
 	/* copy functions for bo handling */
 	struct {
-		int (*blit)(struct radeon_device *rdev,
-			    uint64_t src_offset,
-			    uint64_t dst_offset,
-			    unsigned num_gpu_pages,
-			    struct radeon_fence **fence);
+		struct radeon_fence *(*blit)(struct radeon_device *rdev,
+					     uint64_t src_offset,
+					     uint64_t dst_offset,
+					     unsigned num_gpu_pages,
+					     struct reservation_object *resv);
 		u32 blit_ring_index;
-		int (*dma)(struct radeon_device *rdev,
-			   uint64_t src_offset,
-			   uint64_t dst_offset,
-			   unsigned num_gpu_pages,
-			   struct radeon_fence **fence);
+		struct radeon_fence *(*dma)(struct radeon_device *rdev,
+					    uint64_t src_offset,
+					    uint64_t dst_offset,
+					    unsigned num_gpu_pages,
+					    struct reservation_object *resv);
 		u32 dma_ring_index;
 		/* method used for bo copy */
-		int (*copy)(struct radeon_device *rdev,
-			    uint64_t src_offset,
-			    uint64_t dst_offset,
-			    unsigned num_gpu_pages,
-			    struct radeon_fence **fence);
+		struct radeon_fence *(*copy)(struct radeon_device *rdev,
+					     uint64_t src_offset,
+					     uint64_t dst_offset,
+					     unsigned num_gpu_pages,
+					     struct reservation_object *resv);
 		/* ring used for bo copies */
 		u32 copy_ring_index;
 	} copy;
@@ -2144,6 +2193,8 @@ int radeon_gem_info_ioctl(struct drm_device *dev, void *data,
 			  struct drm_file *filp);
 int radeon_gem_create_ioctl(struct drm_device *dev, void *data,
 			    struct drm_file *filp);
+int radeon_gem_userptr_ioctl(struct drm_device *dev, void *data,
+			     struct drm_file *filp);
 int radeon_gem_pin_ioctl(struct drm_device *dev, void *data,
 			 struct drm_file *file_priv);
 int radeon_gem_unpin_ioctl(struct drm_device *dev, void *data,
@@ -2300,6 +2351,7 @@ struct radeon_device {
 	struct radeon_mman		mman;
 	struct radeon_fence_driver	fence_drv[RADEON_NUM_RINGS];
 	wait_queue_head_t		fence_queue;
+	unsigned			fence_context;
 	struct mutex			ring_lock;
 	struct radeon_ring		ring[RADEON_NUM_RINGS];
 	bool				ib_pool_ready;
@@ -2318,7 +2370,7 @@ struct radeon_device {
 	bool				need_dma32;
 	bool				accel_working;
 	bool				fastfb_working; /* IGP feature*/
-	bool				needs_reset;
+	bool				needs_reset, in_reset;
 	struct radeon_surface_reg surface_regs[RADEON_GEM_MAX_SURFACES];
 	const struct firmware *me_fw;	/* all family ME firmware */
 	const struct firmware *pfp_fw;	/* r6/700 PFP firmware */
@@ -2339,7 +2391,6 @@ struct radeon_device {
 	struct radeon_mec mec;
 	struct work_struct hotplug_work;
 	struct work_struct audio_work;
-	struct work_struct reset_work;
 	int num_crtc; /* number of crtcs */
 	struct mutex dc_hw_i2c_mutex; /* display controller hw i2c mutex */
 	bool has_uvd;
@@ -2365,6 +2416,8 @@ struct radeon_device {
 	struct radeon_atcs		atcs;
 	/* srbm instance registers */
 	struct mutex			srbm_mutex;
+	/* GRBM index mutex. Protects concurrents access to GRBM index */
+	struct mutex			grbm_idx_mutex;
 	/* clock, powergating flags */
 	u32 cg_flags;
 	u32 pg_flags;
@@ -2376,6 +2429,13 @@ struct radeon_device {
 	/* tracking pinned memory */
 	u64 vram_pin_size;
 	u64 gart_pin_size;
+
+	/* amdkfd interface */
+	struct kfd_dev		*kfd;
+	struct radeon_sa_manager	kfd_bo;
+
+	struct mutex	mn_lock;
+	DECLARE_HASHTABLE(mn_hash, 7);
 };
 
 bool radeon_is_px(struct drm_device *dev);
@@ -2431,7 +2491,17 @@ void cik_mm_wdoorbell(struct radeon_device *rdev, u32 index, u32 v);
 /*
  * Cast helper
  */
-#define to_radeon_fence(p) ((struct radeon_fence *)(p))
+extern const struct fence_ops radeon_fence_ops;
+
+static inline struct radeon_fence *to_radeon_fence(struct fence *f)
+{
+	struct radeon_fence *__f = container_of(f, struct radeon_fence, base);
+
+	if (__f->base.ops == &radeon_fence_ops)
+		return __f;
+
+	return NULL;
+}
 
 /*
  * Registers read & write functions.
@@ -2751,18 +2821,25 @@ void radeon_atombios_fini(struct radeon_device *rdev);
 /*
  * RING helpers.
  */
-#if DRM_DEBUG_CODE == 0
+
+/**
+ * radeon_ring_write - write a value to the ring
+ *
+ * @ring: radeon_ring structure holding ring information
+ * @v: dword (dw) value to write
+ *
+ * Write a value to the requested ring buffer (all asics).
+ */
 static inline void radeon_ring_write(struct radeon_ring *ring, uint32_t v)
 {
+	if (ring->count_dw <= 0)
+		DRM_ERROR("radeon: writing more dwords to the ring than expected!\n");
+
 	ring->ring[ring->wptr++] = v;
 	ring->wptr &= ring->ptr_mask;
 	ring->count_dw--;
 	ring->ring_free_dw--;
 }
-#else
-/* With debugging this is just too big to inline */
-void radeon_ring_write(struct radeon_ring *ring, uint32_t v);
-#endif
 
 /*
  * ASICs macro.
@@ -2788,7 +2865,7 @@ void radeon_ring_write(struct radeon_ring *ring, uint32_t v);
 #define radeon_ring_ib_execute(rdev, r, ib) (rdev)->asic->ring[(r)]->ib_execute((rdev), (ib))
 #define radeon_ring_ib_parse(rdev, r, ib) (rdev)->asic->ring[(r)]->ib_parse((rdev), (ib))
 #define radeon_ring_is_lockup(rdev, r, cp) (rdev)->asic->ring[(r)]->is_lockup((rdev), (cp))
-#define radeon_ring_vm_flush(rdev, r, vm) (rdev)->asic->ring[(r)]->vm_flush((rdev), (r), (vm))
+#define radeon_ring_vm_flush(rdev, r, vm_id, pd_addr) (rdev)->asic->ring[(r)->idx]->vm_flush((rdev), (r), (vm_id), (pd_addr))
 #define radeon_ring_get_rptr(rdev, r) (rdev)->asic->ring[(r)->idx]->get_rptr((rdev), (r))
 #define radeon_ring_get_wptr(rdev, r) (rdev)->asic->ring[(r)->idx]->get_wptr((rdev), (r))
 #define radeon_ring_set_wptr(rdev, r) (rdev)->asic->ring[(r)->idx]->set_wptr((rdev), (r))
@@ -2801,9 +2878,9 @@ void radeon_ring_write(struct radeon_ring *ring, uint32_t v);
 #define radeon_hdmi_setmode(rdev, e, m) (rdev)->asic->display.hdmi_setmode((e), (m))
 #define radeon_fence_ring_emit(rdev, r, fence) (rdev)->asic->ring[(r)]->emit_fence((rdev), (fence))
 #define radeon_semaphore_ring_emit(rdev, r, cp, semaphore, emit_wait) (rdev)->asic->ring[(r)]->emit_semaphore((rdev), (cp), (semaphore), (emit_wait))
-#define radeon_copy_blit(rdev, s, d, np, f) (rdev)->asic->copy.blit((rdev), (s), (d), (np), (f))
-#define radeon_copy_dma(rdev, s, d, np, f) (rdev)->asic->copy.dma((rdev), (s), (d), (np), (f))
-#define radeon_copy(rdev, s, d, np, f) (rdev)->asic->copy.copy((rdev), (s), (d), (np), (f))
+#define radeon_copy_blit(rdev, s, d, np, resv) (rdev)->asic->copy.blit((rdev), (s), (d), (np), (resv))
+#define radeon_copy_dma(rdev, s, d, np, resv) (rdev)->asic->copy.dma((rdev), (s), (d), (np), (resv))
+#define radeon_copy(rdev, s, d, np, resv) (rdev)->asic->copy.copy((rdev), (s), (d), (np), (resv))
 #define radeon_copy_blit_ring_index(rdev) (rdev)->asic->copy.blit_ring_index
 #define radeon_copy_dma_ring_index(rdev) (rdev)->asic->copy.dma_ring_index
 #define radeon_copy_ring_index(rdev) (rdev)->asic->copy.copy_ring_index
@@ -2877,6 +2954,10 @@ extern void radeon_legacy_set_clock_gating(struct radeon_device *rdev, int enabl
 extern void radeon_atom_set_clock_gating(struct radeon_device *rdev, int enable);
 extern void radeon_ttm_placement_from_domain(struct radeon_bo *rbo, u32 domain);
 extern bool radeon_ttm_bo_is_radeon_bo(struct ttm_buffer_object *bo);
+extern int radeon_ttm_tt_set_userptr(struct ttm_tt *ttm, uint64_t addr,
+				     uint32_t flags);
+extern bool radeon_ttm_tt_has_userptr(struct ttm_tt *ttm);
+extern bool radeon_ttm_tt_is_readonly(struct ttm_tt *ttm);
 extern void radeon_vram_location(struct radeon_device *rdev, struct radeon_mc *mc, u64 base);
 extern void radeon_gtt_location(struct radeon_device *rdev, struct radeon_mc *mc);
 extern int radeon_resume_kms(struct drm_device *dev, bool resume, bool fbcon);
@@ -2893,14 +2974,14 @@ int radeon_vm_manager_init(struct radeon_device *rdev);
 void radeon_vm_manager_fini(struct radeon_device *rdev);
 int radeon_vm_init(struct radeon_device *rdev, struct radeon_vm *vm);
 void radeon_vm_fini(struct radeon_device *rdev, struct radeon_vm *vm);
-struct radeon_cs_reloc *radeon_vm_get_bos(struct radeon_device *rdev,
+struct radeon_bo_list *radeon_vm_get_bos(struct radeon_device *rdev,
 					  struct radeon_vm *vm,
                                           struct list_head *head);
 struct radeon_fence *radeon_vm_grab_id(struct radeon_device *rdev,
 				       struct radeon_vm *vm, int ring);
 void radeon_vm_flush(struct radeon_device *rdev,
                      struct radeon_vm *vm,
-                     int ring);
+		     int ring, struct radeon_fence *fence);
 void radeon_vm_fence(struct radeon_device *rdev,
 		     struct radeon_vm *vm,
 		     struct radeon_fence *fence);
@@ -2934,10 +3015,10 @@ struct r600_audio_pin *r600_audio_get_pin(struct radeon_device *rdev);
 struct r600_audio_pin *dce6_audio_get_pin(struct radeon_device *rdev);
 void r600_audio_enable(struct radeon_device *rdev,
 		       struct r600_audio_pin *pin,
-		       bool enable);
+		       u8 enable_mask);
 void dce6_audio_enable(struct radeon_device *rdev,
 		       struct r600_audio_pin *pin,
-		       bool enable);
+		       u8 enable_mask);
 
 /*
  * R600 vram scratch functions
@@ -3007,7 +3088,7 @@ bool radeon_cs_packet_next_is_pkt3_nop(struct radeon_cs_parser *p);
 void radeon_cs_dump_packet(struct radeon_cs_parser *p,
 			   struct radeon_cs_packet *pkt);
 int radeon_cs_packet_next_reloc(struct radeon_cs_parser *p,
-				struct radeon_cs_reloc **cs_reloc,
+				struct radeon_bo_list **cs_reloc,
 				int nomm);
 int r600_cs_common_vline_parse(struct radeon_cs_parser *p,
 			       uint32_t *vline_start_end,
diff --git a/drivers/gpu/drm/radeon/radeon_asic.c b/drivers/gpu/drm/radeon/radeon_asic.c
index 2dd5847f9b98..850de57069be 100644
--- a/drivers/gpu/drm/radeon/radeon_asic.c
+++ b/drivers/gpu/drm/radeon/radeon_asic.c
@@ -963,6 +963,19 @@ static struct radeon_asic r600_asic = {
 	},
 };
 
+static struct radeon_asic_ring rv6xx_uvd_ring = {
+	.ib_execute = &uvd_v1_0_ib_execute,
+	.emit_fence = &uvd_v1_0_fence_emit,
+	.emit_semaphore = &uvd_v1_0_semaphore_emit,
+	.cs_parse = &radeon_uvd_cs_parse,
+	.ring_test = &uvd_v1_0_ring_test,
+	.ib_test = &uvd_v1_0_ib_test,
+	.is_lockup = &radeon_ring_test_lockup,
+	.get_rptr = &uvd_v1_0_get_rptr,
+	.get_wptr = &uvd_v1_0_get_wptr,
+	.set_wptr = &uvd_v1_0_set_wptr,
+};
+
 static struct radeon_asic rv6xx_asic = {
 	.init = &r600_init,
 	.fini = &r600_fini,
@@ -982,6 +995,7 @@ static struct radeon_asic rv6xx_asic = {
 	.ring = {
 		[RADEON_RING_TYPE_GFX_INDEX] = &r600_gfx_ring,
 		[R600_RING_TYPE_DMA_INDEX] = &r600_dma_ring,
+		[R600_RING_TYPE_UVD_INDEX] = &rv6xx_uvd_ring,
 	},
 	.irq = {
 		.set = &r600_irq_set,
@@ -1072,6 +1086,7 @@ static struct radeon_asic rs780_asic = {
 	.ring = {
 		[RADEON_RING_TYPE_GFX_INDEX] = &r600_gfx_ring,
 		[R600_RING_TYPE_DMA_INDEX] = &r600_dma_ring,
+		[R600_RING_TYPE_UVD_INDEX] = &rv6xx_uvd_ring,
 	},
 	.irq = {
 		.set = &r600_irq_set,
@@ -2296,7 +2311,15 @@ int radeon_asic_init(struct radeon_device *rdev)
 	case CHIP_RS780:
 	case CHIP_RS880:
 		rdev->asic = &rs780_asic;
-		rdev->has_uvd = true;
+		/* 760G/780V/880V don't have UVD */
+		if ((rdev->pdev->device == 0x9616)||
+		    (rdev->pdev->device == 0x9611)||
+		    (rdev->pdev->device == 0x9613)||
+		    (rdev->pdev->device == 0x9711)||
+		    (rdev->pdev->device == 0x9713))
+			rdev->has_uvd = false;
+		else
+			rdev->has_uvd = true;
 		break;
 	case CHIP_RV770:
 	case CHIP_RV730:
diff --git a/drivers/gpu/drm/radeon/radeon_asic.h b/drivers/gpu/drm/radeon/radeon_asic.h
index 7756bc1e1cd3..2a45d548d5ec 100644
--- a/drivers/gpu/drm/radeon/radeon_asic.h
+++ b/drivers/gpu/drm/radeon/radeon_asic.h
@@ -81,11 +81,11 @@ bool r100_semaphore_ring_emit(struct radeon_device *rdev,
 int r100_cs_parse(struct radeon_cs_parser *p);
 void r100_pll_wreg(struct radeon_device *rdev, uint32_t reg, uint32_t v);
 uint32_t r100_pll_rreg(struct radeon_device *rdev, uint32_t reg);
-int r100_copy_blit(struct radeon_device *rdev,
-		   uint64_t src_offset,
-		   uint64_t dst_offset,
-		   unsigned num_gpu_pages,
-		   struct radeon_fence **fence);
+struct radeon_fence *r100_copy_blit(struct radeon_device *rdev,
+				    uint64_t src_offset,
+				    uint64_t dst_offset,
+				    unsigned num_gpu_pages,
+				    struct reservation_object *resv);
 int r100_set_surface_reg(struct radeon_device *rdev, int reg,
 			 uint32_t tiling_flags, uint32_t pitch,
 			 uint32_t offset, uint32_t obj_size);
@@ -152,11 +152,11 @@ void r100_gfx_set_wptr(struct radeon_device *rdev,
 /*
  * r200,rv250,rs300,rv280
  */
-extern int r200_copy_dma(struct radeon_device *rdev,
-			 uint64_t src_offset,
-			 uint64_t dst_offset,
-			 unsigned num_gpu_pages,
-			 struct radeon_fence **fence);
+struct radeon_fence *r200_copy_dma(struct radeon_device *rdev,
+				   uint64_t src_offset,
+				   uint64_t dst_offset,
+				   unsigned num_gpu_pages,
+				   struct reservation_object *resv);
 void r200_set_safe_registers(struct radeon_device *rdev);
 
 /*
@@ -340,12 +340,14 @@ int r600_dma_ib_test(struct radeon_device *rdev, struct radeon_ring *ring);
 void r600_ring_ib_execute(struct radeon_device *rdev, struct radeon_ib *ib);
 int r600_ring_test(struct radeon_device *rdev, struct radeon_ring *cp);
 int r600_dma_ring_test(struct radeon_device *rdev, struct radeon_ring *cp);
-int r600_copy_cpdma(struct radeon_device *rdev,
-		    uint64_t src_offset, uint64_t dst_offset,
-		    unsigned num_gpu_pages, struct radeon_fence **fence);
-int r600_copy_dma(struct radeon_device *rdev,
-		  uint64_t src_offset, uint64_t dst_offset,
-		  unsigned num_gpu_pages, struct radeon_fence **fence);
+struct radeon_fence *r600_copy_cpdma(struct radeon_device *rdev,
+				     uint64_t src_offset, uint64_t dst_offset,
+				     unsigned num_gpu_pages,
+				     struct reservation_object *resv);
+struct radeon_fence *r600_copy_dma(struct radeon_device *rdev,
+				   uint64_t src_offset, uint64_t dst_offset,
+				   unsigned num_gpu_pages,
+				   struct reservation_object *resv);
 void r600_hpd_init(struct radeon_device *rdev);
 void r600_hpd_fini(struct radeon_device *rdev);
 bool r600_hpd_sense(struct radeon_device *rdev, enum radeon_hpd_id hpd);
@@ -389,7 +391,6 @@ void r600_disable_interrupts(struct radeon_device *rdev);
 void r600_rlc_stop(struct radeon_device *rdev);
 /* r600 audio */
 int r600_audio_init(struct radeon_device *rdev);
-struct r600_audio_pin r600_audio_status(struct radeon_device *rdev);
 void r600_audio_fini(struct radeon_device *rdev);
 void r600_audio_set_dto(struct drm_encoder *encoder, u32 clock);
 void r600_hdmi_update_avi_infoframe(struct drm_encoder *encoder, void *buffer,
@@ -461,10 +462,10 @@ bool rv770_page_flip_pending(struct radeon_device *rdev, int crtc);
 void r700_vram_gtt_location(struct radeon_device *rdev, struct radeon_mc *mc);
 void r700_cp_stop(struct radeon_device *rdev);
 void r700_cp_fini(struct radeon_device *rdev);
-int rv770_copy_dma(struct radeon_device *rdev,
-		  uint64_t src_offset, uint64_t dst_offset,
-		  unsigned num_gpu_pages,
-		   struct radeon_fence **fence);
+struct radeon_fence *rv770_copy_dma(struct radeon_device *rdev,
+				    uint64_t src_offset, uint64_t dst_offset,
+				    unsigned num_gpu_pages,
+				    struct reservation_object *resv);
 u32 rv770_get_xclk(struct radeon_device *rdev);
 int rv770_set_uvd_clocks(struct radeon_device *rdev, u32 vclk, u32 dclk);
 int rv770_get_temp(struct radeon_device *rdev);
@@ -535,10 +536,10 @@ void evergreen_dma_fence_ring_emit(struct radeon_device *rdev,
 				   struct radeon_fence *fence);
 void evergreen_dma_ring_ib_execute(struct radeon_device *rdev,
 				   struct radeon_ib *ib);
-int evergreen_copy_dma(struct radeon_device *rdev,
-		       uint64_t src_offset, uint64_t dst_offset,
-		       unsigned num_gpu_pages,
-		       struct radeon_fence **fence);
+struct radeon_fence *evergreen_copy_dma(struct radeon_device *rdev,
+					uint64_t src_offset, uint64_t dst_offset,
+					unsigned num_gpu_pages,
+					struct reservation_object *resv);
 void evergreen_hdmi_enable(struct drm_encoder *encoder, bool enable);
 void evergreen_hdmi_setmode(struct drm_encoder *encoder, struct drm_display_mode *mode);
 int evergreen_get_temp(struct radeon_device *rdev);
@@ -598,7 +599,8 @@ int cayman_asic_reset(struct radeon_device *rdev);
 void cayman_ring_ib_execute(struct radeon_device *rdev, struct radeon_ib *ib);
 int cayman_vm_init(struct radeon_device *rdev);
 void cayman_vm_fini(struct radeon_device *rdev);
-void cayman_vm_flush(struct radeon_device *rdev, int ridx, struct radeon_vm *vm);
+void cayman_vm_flush(struct radeon_device *rdev, struct radeon_ring *ring,
+		     unsigned vm_id, uint64_t pd_addr);
 uint32_t cayman_vm_page_flags(struct radeon_device *rdev, uint32_t flags);
 int evergreen_ib_parse(struct radeon_device *rdev, struct radeon_ib *ib);
 int evergreen_dma_ib_parse(struct radeon_device *rdev, struct radeon_ib *ib);
@@ -623,7 +625,8 @@ void cayman_dma_vm_set_pages(struct radeon_device *rdev,
 			     uint32_t incr, uint32_t flags);
 void cayman_dma_vm_pad_ib(struct radeon_ib *ib);
 
-void cayman_dma_vm_flush(struct radeon_device *rdev, int ridx, struct radeon_vm *vm);
+void cayman_dma_vm_flush(struct radeon_device *rdev, struct radeon_ring *ring,
+			 unsigned vm_id, uint64_t pd_addr);
 
 u32 cayman_gfx_get_rptr(struct radeon_device *rdev,
 			struct radeon_ring *ring);
@@ -698,12 +701,13 @@ int si_irq_set(struct radeon_device *rdev);
 int si_irq_process(struct radeon_device *rdev);
 int si_vm_init(struct radeon_device *rdev);
 void si_vm_fini(struct radeon_device *rdev);
-void si_vm_flush(struct radeon_device *rdev, int ridx, struct radeon_vm *vm);
+void si_vm_flush(struct radeon_device *rdev, struct radeon_ring *ring,
+		 unsigned vm_id, uint64_t pd_addr);
 int si_ib_parse(struct radeon_device *rdev, struct radeon_ib *ib);
-int si_copy_dma(struct radeon_device *rdev,
-		uint64_t src_offset, uint64_t dst_offset,
-		unsigned num_gpu_pages,
-		struct radeon_fence **fence);
+struct radeon_fence *si_copy_dma(struct radeon_device *rdev,
+				 uint64_t src_offset, uint64_t dst_offset,
+				 unsigned num_gpu_pages,
+				 struct reservation_object *resv);
 
 void si_dma_vm_copy_pages(struct radeon_device *rdev,
 			  struct radeon_ib *ib,
@@ -720,7 +724,8 @@ void si_dma_vm_set_pages(struct radeon_device *rdev,
 			 uint64_t addr, unsigned count,
 			 uint32_t incr, uint32_t flags);
 
-void si_dma_vm_flush(struct radeon_device *rdev, int ridx, struct radeon_vm *vm);
+void si_dma_vm_flush(struct radeon_device *rdev, struct radeon_ring *ring,
+		     unsigned vm_id, uint64_t pd_addr);
 u32 si_get_xclk(struct radeon_device *rdev);
 uint64_t si_get_gpu_clock_counter(struct radeon_device *rdev);
 int si_set_uvd_clocks(struct radeon_device *rdev, u32 vclk, u32 dclk);
@@ -759,14 +764,14 @@ bool cik_sdma_semaphore_ring_emit(struct radeon_device *rdev,
 				  struct radeon_semaphore *semaphore,
 				  bool emit_wait);
 void cik_sdma_ring_ib_execute(struct radeon_device *rdev, struct radeon_ib *ib);
-int cik_copy_dma(struct radeon_device *rdev,
-		 uint64_t src_offset, uint64_t dst_offset,
-		 unsigned num_gpu_pages,
-		 struct radeon_fence **fence);
-int cik_copy_cpdma(struct radeon_device *rdev,
-		   uint64_t src_offset, uint64_t dst_offset,
-		   unsigned num_gpu_pages,
-		   struct radeon_fence **fence);
+struct radeon_fence *cik_copy_dma(struct radeon_device *rdev,
+				  uint64_t src_offset, uint64_t dst_offset,
+				  unsigned num_gpu_pages,
+				  struct reservation_object *resv);
+struct radeon_fence *cik_copy_cpdma(struct radeon_device *rdev,
+				    uint64_t src_offset, uint64_t dst_offset,
+				    unsigned num_gpu_pages,
+				    struct reservation_object *resv);
 int cik_sdma_ring_test(struct radeon_device *rdev, struct radeon_ring *ring);
 int cik_sdma_ib_test(struct radeon_device *rdev, struct radeon_ring *ring);
 bool cik_sdma_is_lockup(struct radeon_device *rdev, struct radeon_ring *ring);
@@ -792,7 +797,8 @@ int cik_irq_set(struct radeon_device *rdev);
 int cik_irq_process(struct radeon_device *rdev);
 int cik_vm_init(struct radeon_device *rdev);
 void cik_vm_fini(struct radeon_device *rdev);
-void cik_vm_flush(struct radeon_device *rdev, int ridx, struct radeon_vm *vm);
+void cik_vm_flush(struct radeon_device *rdev, struct radeon_ring *ring,
+		  unsigned vm_id, uint64_t pd_addr);
 
 void cik_sdma_vm_copy_pages(struct radeon_device *rdev,
 			    struct radeon_ib *ib,
@@ -810,7 +816,8 @@ void cik_sdma_vm_set_pages(struct radeon_device *rdev,
 			   uint32_t incr, uint32_t flags);
 void cik_sdma_vm_pad_ib(struct radeon_ib *ib);
 
-void cik_dma_vm_flush(struct radeon_device *rdev, int ridx, struct radeon_vm *vm);
+void cik_dma_vm_flush(struct radeon_device *rdev, struct radeon_ring *ring,
+		      unsigned vm_id, uint64_t pd_addr);
 int cik_ib_parse(struct radeon_device *rdev, struct radeon_ib *ib);
 u32 cik_gfx_get_rptr(struct radeon_device *rdev,
 		     struct radeon_ring *ring);
@@ -882,6 +889,7 @@ uint32_t uvd_v1_0_get_wptr(struct radeon_device *rdev,
                            struct radeon_ring *ring);
 void uvd_v1_0_set_wptr(struct radeon_device *rdev,
                        struct radeon_ring *ring);
+int uvd_v1_0_resume(struct radeon_device *rdev);
 
 int uvd_v1_0_init(struct radeon_device *rdev);
 void uvd_v1_0_fini(struct radeon_device *rdev);
@@ -889,6 +897,8 @@ int uvd_v1_0_start(struct radeon_device *rdev);
 void uvd_v1_0_stop(struct radeon_device *rdev);
 
 int uvd_v1_0_ring_test(struct radeon_device *rdev, struct radeon_ring *ring);
+void uvd_v1_0_fence_emit(struct radeon_device *rdev,
+			 struct radeon_fence *fence);
 int uvd_v1_0_ib_test(struct radeon_device *rdev, struct radeon_ring *ring);
 bool uvd_v1_0_semaphore_emit(struct radeon_device *rdev,
 			     struct radeon_ring *ring,
diff --git a/drivers/gpu/drm/radeon/radeon_atombios.c b/drivers/gpu/drm/radeon/radeon_atombios.c
index e74c7e387dde..dbc94f300297 100644
--- a/drivers/gpu/drm/radeon/radeon_atombios.c
+++ b/drivers/gpu/drm/radeon/radeon_atombios.c
@@ -196,8 +196,8 @@ void radeon_atombios_i2c_init(struct radeon_device *rdev)
 	}
 }
 
-static struct radeon_gpio_rec radeon_lookup_gpio(struct radeon_device *rdev,
-						 u8 id)
+struct radeon_gpio_rec radeon_atombios_lookup_gpio(struct radeon_device *rdev,
+						   u8 id)
 {
 	struct atom_context *ctx = rdev->mode_info.atom_context;
 	struct radeon_gpio_rec gpio;
@@ -221,6 +221,7 @@ static struct radeon_gpio_rec radeon_lookup_gpio(struct radeon_device *rdev,
 			if (id == pin->ucGPIO_ID) {
 				gpio.id = pin->ucGPIO_ID;
 				gpio.reg = le16_to_cpu(pin->usGpioPin_AIndex) * 4;
+				gpio.shift = pin->ucGpioPinBitShift;
 				gpio.mask = (1 << pin->ucGpioPinBitShift);
 				gpio.valid = true;
 				break;
@@ -458,7 +459,7 @@ static bool radeon_atom_apply_quirks(struct drm_device *dev,
 	return true;
 }
 
-const int supported_devices_connector_convert[] = {
+static const int supported_devices_connector_convert[] = {
 	DRM_MODE_CONNECTOR_Unknown,
 	DRM_MODE_CONNECTOR_VGA,
 	DRM_MODE_CONNECTOR_DVII,
@@ -477,7 +478,7 @@ const int supported_devices_connector_convert[] = {
 	DRM_MODE_CONNECTOR_DisplayPort
 };
 
-const uint16_t supported_devices_connector_object_id_convert[] = {
+static const uint16_t supported_devices_connector_object_id_convert[] = {
 	CONNECTOR_OBJECT_ID_NONE,
 	CONNECTOR_OBJECT_ID_VGA,
 	CONNECTOR_OBJECT_ID_DUAL_LINK_DVI_I, /* not all boards support DL */
@@ -494,7 +495,7 @@ const uint16_t supported_devices_connector_object_id_convert[] = {
 	CONNECTOR_OBJECT_ID_SVIDEO
 };
 
-const int object_connector_convert[] = {
+static const int object_connector_convert[] = {
 	DRM_MODE_CONNECTOR_Unknown,
 	DRM_MODE_CONNECTOR_DVII,
 	DRM_MODE_CONNECTOR_DVII,
@@ -801,7 +802,7 @@ bool radeon_get_atom_connector_info_from_object_table(struct drm_device *dev)
 								hpd_record =
 									(ATOM_HPD_INT_RECORD *)
 									record;
-								gpio = radeon_lookup_gpio(rdev,
+								gpio = radeon_atombios_lookup_gpio(rdev,
 											  hpd_record->ucHPDIntGPIOID);
 								hpd = radeon_atom_get_hpd_info_from_gpio(rdev, &gpio);
 								hpd.plugged_state = hpd_record->ucPlugged_PinState;
@@ -2128,7 +2129,7 @@ static int radeon_atombios_parse_power_table_1_3(struct radeon_device *rdev)
 				rdev->pm.power_state[state_index].clock_info[0].voltage.type =
 					VOLTAGE_GPIO;
 				rdev->pm.power_state[state_index].clock_info[0].voltage.gpio =
-					radeon_lookup_gpio(rdev,
+					radeon_atombios_lookup_gpio(rdev,
 							   power_info->info.asPowerPlayInfo[i].ucVoltageDropIndex);
 				if (misc & ATOM_PM_MISCINFO_VOLTAGE_DROP_ACTIVE_HIGH)
 					rdev->pm.power_state[state_index].clock_info[0].voltage.active_high =
@@ -2164,7 +2165,7 @@ static int radeon_atombios_parse_power_table_1_3(struct radeon_device *rdev)
 				rdev->pm.power_state[state_index].clock_info[0].voltage.type =
 					VOLTAGE_GPIO;
 				rdev->pm.power_state[state_index].clock_info[0].voltage.gpio =
-					radeon_lookup_gpio(rdev,
+					radeon_atombios_lookup_gpio(rdev,
 							   power_info->info_2.asPowerPlayInfo[i].ucVoltageDropIndex);
 				if (misc & ATOM_PM_MISCINFO_VOLTAGE_DROP_ACTIVE_HIGH)
 					rdev->pm.power_state[state_index].clock_info[0].voltage.active_high =
@@ -2200,7 +2201,7 @@ static int radeon_atombios_parse_power_table_1_3(struct radeon_device *rdev)
 				rdev->pm.power_state[state_index].clock_info[0].voltage.type =
 					VOLTAGE_GPIO;
 				rdev->pm.power_state[state_index].clock_info[0].voltage.gpio =
-					radeon_lookup_gpio(rdev,
+					radeon_atombios_lookup_gpio(rdev,
 							   power_info->info_3.asPowerPlayInfo[i].ucVoltageDropIndex);
 				if (misc & ATOM_PM_MISCINFO_VOLTAGE_DROP_ACTIVE_HIGH)
 					rdev->pm.power_state[state_index].clock_info[0].voltage.active_high =
@@ -2248,6 +2249,14 @@ static void radeon_atombios_add_pplib_thermal_controller(struct radeon_device *r
 
 	/* add the i2c bus for thermal/fan chip */
 	if (controller->ucType > 0) {
+		if (controller->ucFanParameters & ATOM_PP_FANPARAMETERS_NOFAN)
+			rdev->pm.no_fan = true;
+		rdev->pm.fan_pulses_per_revolution =
+			controller->ucFanParameters & ATOM_PP_FANPARAMETERS_TACHOMETER_PULSES_PER_REVOLUTION_MASK;
+		if (rdev->pm.fan_pulses_per_revolution) {
+			rdev->pm.fan_min_rpm = controller->ucFanMinRPM;
+			rdev->pm.fan_max_rpm = controller->ucFanMaxRPM;
+		}
 		if (controller->ucType == ATOM_PP_THERMALCONTROLLER_RV6xx) {
 			DRM_INFO("Internal thermal controller %s fan control\n",
 				 (controller->ucFanParameters &
diff --git a/drivers/gpu/drm/radeon/radeon_benchmark.c b/drivers/gpu/drm/radeon/radeon_benchmark.c
index 69f5695bdab9..9e7f23dd14bd 100644
--- a/drivers/gpu/drm/radeon/radeon_benchmark.c
+++ b/drivers/gpu/drm/radeon/radeon_benchmark.c
@@ -45,33 +45,29 @@ static int radeon_benchmark_do_move(struct radeon_device *rdev, unsigned size,
 	for (i = 0; i < n; i++) {
 		switch (flag) {
 		case RADEON_BENCHMARK_COPY_DMA:
-			r = radeon_copy_dma(rdev, saddr, daddr,
-					    size / RADEON_GPU_PAGE_SIZE,
-					    &fence);
+			fence = radeon_copy_dma(rdev, saddr, daddr,
+						size / RADEON_GPU_PAGE_SIZE,
+						NULL);
 			break;
 		case RADEON_BENCHMARK_COPY_BLIT:
-			r = radeon_copy_blit(rdev, saddr, daddr,
-					     size / RADEON_GPU_PAGE_SIZE,
-					     &fence);
+			fence = radeon_copy_blit(rdev, saddr, daddr,
+						 size / RADEON_GPU_PAGE_SIZE,
+						 NULL);
 			break;
 		default:
 			DRM_ERROR("Unknown copy method\n");
-			r = -EINVAL;
+			return -EINVAL;
 		}
-		if (r)
-			goto exit_do_move;
+		if (IS_ERR(fence))
+			return PTR_ERR(fence);
+
 		r = radeon_fence_wait(fence, false);
-		if (r)
-			goto exit_do_move;
 		radeon_fence_unref(&fence);
+		if (r)
+			return r;
 	}
 	end_jiffies = jiffies;
-	r = jiffies_to_msecs(end_jiffies - start_jiffies);
-
-exit_do_move:
-	if (fence)
-		radeon_fence_unref(&fence);
-	return r;
+	return jiffies_to_msecs(end_jiffies - start_jiffies);
 }
 
 
@@ -97,7 +93,7 @@ static void radeon_benchmark_move(struct radeon_device *rdev, unsigned size,
 	int time;
 
 	n = RADEON_BENCHMARK_ITERATIONS;
-	r = radeon_bo_create(rdev, size, PAGE_SIZE, true, sdomain, 0, NULL, &sobj);
+	r = radeon_bo_create(rdev, size, PAGE_SIZE, true, sdomain, 0, NULL, NULL, &sobj);
 	if (r) {
 		goto out_cleanup;
 	}
@@ -109,7 +105,7 @@ static void radeon_benchmark_move(struct radeon_device *rdev, unsigned size,
 	if (r) {
 		goto out_cleanup;
 	}
-	r = radeon_bo_create(rdev, size, PAGE_SIZE, true, ddomain, 0, NULL, &dobj);
+	r = radeon_bo_create(rdev, size, PAGE_SIZE, true, ddomain, 0, NULL, NULL, &dobj);
 	if (r) {
 		goto out_cleanup;
 	}
diff --git a/drivers/gpu/drm/radeon/radeon_bios.c b/drivers/gpu/drm/radeon/radeon_bios.c
index 6a03624fadaa..63ccb8fa799c 100644
--- a/drivers/gpu/drm/radeon/radeon_bios.c
+++ b/drivers/gpu/drm/radeon/radeon_bios.c
@@ -658,12 +658,10 @@ bool radeon_get_bios(struct radeon_device *rdev)
 		r = igp_read_bios_from_vram(rdev);
 	if (r == false)
 		r = radeon_read_bios(rdev);
-	if (r == false) {
+	if (r == false)
 		r = radeon_read_disabled_bios(rdev);
-	}
-	if (r == false) {
+	if (r == false)
 		r = radeon_read_platform_bios(rdev);
-	}
 	if (r == false || rdev->bios == NULL) {
 		DRM_ERROR("Unable to locate a BIOS ROM\n");
 		rdev->bios = NULL;
diff --git a/drivers/gpu/drm/radeon/radeon_combios.c b/drivers/gpu/drm/radeon/radeon_combios.c
index 6651177110f0..3e5f6b71f3ad 100644
--- a/drivers/gpu/drm/radeon/radeon_combios.c
+++ b/drivers/gpu/drm/radeon/radeon_combios.c
@@ -116,7 +116,7 @@ enum radeon_combios_connector {
 	CONNECTOR_UNSUPPORTED_LEGACY
 };
 
-const int legacy_connector_convert[] = {
+static const int legacy_connector_convert[] = {
 	DRM_MODE_CONNECTOR_Unknown,
 	DRM_MODE_CONNECTOR_DVID,
 	DRM_MODE_CONNECTOR_VGA,
diff --git a/drivers/gpu/drm/radeon/radeon_connectors.c b/drivers/gpu/drm/radeon/radeon_connectors.c
index 300c4b3d4669..26baa9c05f6c 100644
--- a/drivers/gpu/drm/radeon/radeon_connectors.c
+++ b/drivers/gpu/drm/radeon/radeon_connectors.c
@@ -322,6 +322,12 @@ static void radeon_connector_get_edid(struct drm_connector *connector)
 	}
 
 	if (!radeon_connector->edid) {
+		/* don't fetch the edid from the vbios if ddc fails and runpm is
+		 * enabled so we report disconnected.
+		 */
+		if ((rdev->flags & RADEON_IS_PX) && (radeon_runtime_pm != 0))
+			return;
+
 		if (rdev->is_atom_bios) {
 			/* some laptops provide a hardcoded edid in rom for LCDs */
 			if (((connector->connector_type == DRM_MODE_CONNECTOR_LVDS) ||
@@ -826,6 +832,8 @@ static int radeon_lvds_mode_valid(struct drm_connector *connector,
 static enum drm_connector_status
 radeon_lvds_detect(struct drm_connector *connector, bool force)
 {
+	struct drm_device *dev = connector->dev;
+	struct radeon_device *rdev = dev->dev_private;
 	struct radeon_connector *radeon_connector = to_radeon_connector(connector);
 	struct drm_encoder *encoder = radeon_best_single_encoder(connector);
 	enum drm_connector_status ret = connector_status_disconnected;
@@ -842,7 +850,11 @@ radeon_lvds_detect(struct drm_connector *connector, bool force)
 		/* check if panel is valid */
 		if (native_mode->hdisplay >= 320 && native_mode->vdisplay >= 240)
 			ret = connector_status_connected;
-
+		/* don't fetch the edid from the vbios if ddc fails and runpm is
+		 * enabled so we report disconnected.
+		 */
+		if ((rdev->flags & RADEON_IS_PX) && (radeon_runtime_pm != 0))
+			ret = connector_status_disconnected;
 	}
 
 	/* check for edid as well */
@@ -1589,6 +1601,11 @@ radeon_dp_detect(struct drm_connector *connector, bool force)
 			/* check if panel is valid */
 			if (native_mode->hdisplay >= 320 && native_mode->vdisplay >= 240)
 				ret = connector_status_connected;
+			/* don't fetch the edid from the vbios if ddc fails and runpm is
+			 * enabled so we report disconnected.
+			 */
+			if ((rdev->flags & RADEON_IS_PX) && (radeon_runtime_pm != 0))
+				ret = connector_status_disconnected;
 		}
 		/* eDP is always DP */
 		radeon_dig_connector->dp_sink_type = CONNECTOR_OBJECT_ID_DISPLAYPORT;
diff --git a/drivers/gpu/drm/radeon/radeon_cp.c b/drivers/gpu/drm/radeon/radeon_cp.c
index bb0d5c3a8311..ea134a7d51a5 100644
--- a/drivers/gpu/drm/radeon/radeon_cp.c
+++ b/drivers/gpu/drm/radeon/radeon_cp.c
@@ -1298,27 +1298,27 @@ static int radeon_do_init_cp(struct drm_device *dev, drm_radeon_init_t *init,
 	dev_priv->buffers_offset = init->buffers_offset;
 	dev_priv->gart_textures_offset = init->gart_textures_offset;
 
-	master_priv->sarea = drm_getsarea(dev);
+	master_priv->sarea = drm_legacy_getsarea(dev);
 	if (!master_priv->sarea) {
 		DRM_ERROR("could not find sarea!\n");
 		radeon_do_cleanup_cp(dev);
 		return -EINVAL;
 	}
 
-	dev_priv->cp_ring = drm_core_findmap(dev, init->ring_offset);
+	dev_priv->cp_ring = drm_legacy_findmap(dev, init->ring_offset);
 	if (!dev_priv->cp_ring) {
 		DRM_ERROR("could not find cp ring region!\n");
 		radeon_do_cleanup_cp(dev);
 		return -EINVAL;
 	}
-	dev_priv->ring_rptr = drm_core_findmap(dev, init->ring_rptr_offset);
+	dev_priv->ring_rptr = drm_legacy_findmap(dev, init->ring_rptr_offset);
 	if (!dev_priv->ring_rptr) {
 		DRM_ERROR("could not find ring read pointer!\n");
 		radeon_do_cleanup_cp(dev);
 		return -EINVAL;
 	}
 	dev->agp_buffer_token = init->buffers_offset;
-	dev->agp_buffer_map = drm_core_findmap(dev, init->buffers_offset);
+	dev->agp_buffer_map = drm_legacy_findmap(dev, init->buffers_offset);
 	if (!dev->agp_buffer_map) {
 		DRM_ERROR("could not find dma buffer region!\n");
 		radeon_do_cleanup_cp(dev);
@@ -1327,7 +1327,7 @@ static int radeon_do_init_cp(struct drm_device *dev, drm_radeon_init_t *init,
 
 	if (init->gart_textures_offset) {
 		dev_priv->gart_textures =
-		    drm_core_findmap(dev, init->gart_textures_offset);
+		    drm_legacy_findmap(dev, init->gart_textures_offset);
 		if (!dev_priv->gart_textures) {
 			DRM_ERROR("could not find GART texture region!\n");
 			radeon_do_cleanup_cp(dev);
@@ -1337,9 +1337,9 @@ static int radeon_do_init_cp(struct drm_device *dev, drm_radeon_init_t *init,
 
 #if __OS_HAS_AGP
 	if (dev_priv->flags & RADEON_IS_AGP) {
-		drm_core_ioremap_wc(dev_priv->cp_ring, dev);
-		drm_core_ioremap_wc(dev_priv->ring_rptr, dev);
-		drm_core_ioremap_wc(dev->agp_buffer_map, dev);
+		drm_legacy_ioremap_wc(dev_priv->cp_ring, dev);
+		drm_legacy_ioremap_wc(dev_priv->ring_rptr, dev);
+		drm_legacy_ioremap_wc(dev->agp_buffer_map, dev);
 		if (!dev_priv->cp_ring->handle ||
 		    !dev_priv->ring_rptr->handle ||
 		    !dev->agp_buffer_map->handle) {
@@ -1475,7 +1475,7 @@ static int radeon_do_init_cp(struct drm_device *dev, drm_radeon_init_t *init,
 			dev_priv->gart_info.mapping.size =
 			    dev_priv->gart_info.table_size;
 
-			drm_core_ioremap_wc(&dev_priv->gart_info.mapping, dev);
+			drm_legacy_ioremap_wc(&dev_priv->gart_info.mapping, dev);
 			dev_priv->gart_info.addr =
 			    dev_priv->gart_info.mapping.handle;
 
@@ -1569,15 +1569,15 @@ static int radeon_do_cleanup_cp(struct drm_device * dev)
 #if __OS_HAS_AGP
 	if (dev_priv->flags & RADEON_IS_AGP) {
 		if (dev_priv->cp_ring != NULL) {
-			drm_core_ioremapfree(dev_priv->cp_ring, dev);
+			drm_legacy_ioremapfree(dev_priv->cp_ring, dev);
 			dev_priv->cp_ring = NULL;
 		}
 		if (dev_priv->ring_rptr != NULL) {
-			drm_core_ioremapfree(dev_priv->ring_rptr, dev);
+			drm_legacy_ioremapfree(dev_priv->ring_rptr, dev);
 			dev_priv->ring_rptr = NULL;
 		}
 		if (dev->agp_buffer_map != NULL) {
-			drm_core_ioremapfree(dev->agp_buffer_map, dev);
+			drm_legacy_ioremapfree(dev->agp_buffer_map, dev);
 			dev->agp_buffer_map = NULL;
 		}
 	} else
@@ -1597,7 +1597,7 @@ static int radeon_do_cleanup_cp(struct drm_device * dev)
 
 		if (dev_priv->gart_info.gart_table_location == DRM_ATI_GART_FB)
 		{
-			drm_core_ioremapfree(&dev_priv->gart_info.mapping, dev);
+			drm_legacy_ioremapfree(&dev_priv->gart_info.mapping, dev);
 			dev_priv->gart_info.addr = NULL;
 		}
 	}
@@ -2106,9 +2106,9 @@ int radeon_driver_load(struct drm_device *dev, unsigned long flags)
 	else
 		dev_priv->flags |= RADEON_IS_PCI;
 
-	ret = drm_addmap(dev, pci_resource_start(dev->pdev, 2),
-			 pci_resource_len(dev->pdev, 2), _DRM_REGISTERS,
-			 _DRM_READ_ONLY | _DRM_DRIVER, &dev_priv->mmio);
+	ret = drm_legacy_addmap(dev, pci_resource_start(dev->pdev, 2),
+				pci_resource_len(dev->pdev, 2), _DRM_REGISTERS,
+				_DRM_READ_ONLY | _DRM_DRIVER, &dev_priv->mmio);
 	if (ret != 0)
 		return ret;
 
@@ -2135,8 +2135,8 @@ int radeon_master_create(struct drm_device *dev, struct drm_master *master)
 
 	/* prebuild the SAREA */
 	sareapage = max_t(unsigned long, SAREA_MAX, PAGE_SIZE);
-	ret = drm_addmap(dev, 0, sareapage, _DRM_SHM, _DRM_CONTAINS_LOCK,
-			 &master_priv->sarea);
+	ret = drm_legacy_addmap(dev, 0, sareapage, _DRM_SHM, _DRM_CONTAINS_LOCK,
+				&master_priv->sarea);
 	if (ret) {
 		DRM_ERROR("SAREA setup failed\n");
 		kfree(master_priv);
@@ -2162,7 +2162,7 @@ void radeon_master_destroy(struct drm_device *dev, struct drm_master *master)
 
 	master_priv->sarea_priv = NULL;
 	if (master_priv->sarea)
-		drm_rmmap_locked(dev, master_priv->sarea);
+		drm_legacy_rmmap_locked(dev, master_priv->sarea);
 
 	kfree(master_priv);
 
@@ -2181,9 +2181,9 @@ int radeon_driver_firstopen(struct drm_device *dev)
 	dev_priv->gart_info.table_size = RADEON_PCIGART_TABLE_SIZE;
 
 	dev_priv->fb_aper_offset = pci_resource_start(dev->pdev, 0);
-	ret = drm_addmap(dev, dev_priv->fb_aper_offset,
-			 pci_resource_len(dev->pdev, 0), _DRM_FRAME_BUFFER,
-			 _DRM_WRITE_COMBINING, &map);
+	ret = drm_legacy_addmap(dev, dev_priv->fb_aper_offset,
+				pci_resource_len(dev->pdev, 0),
+				_DRM_FRAME_BUFFER, _DRM_WRITE_COMBINING, &map);
 	if (ret != 0)
 		return ret;
 
@@ -2196,7 +2196,7 @@ int radeon_driver_unload(struct drm_device *dev)
 
 	DRM_DEBUG("\n");
 
-	drm_rmmap(dev, dev_priv->mmio);
+	drm_legacy_rmmap(dev, dev_priv->mmio);
 
 	kfree(dev_priv);
 
diff --git a/drivers/gpu/drm/radeon/radeon_cs.c b/drivers/gpu/drm/radeon/radeon_cs.c
index 83f382e8e40e..c830863bc98a 100644
--- a/drivers/gpu/drm/radeon/radeon_cs.c
+++ b/drivers/gpu/drm/radeon/radeon_cs.c
@@ -77,21 +77,18 @@ static int radeon_cs_parser_relocs(struct radeon_cs_parser *p)
 	struct drm_device *ddev = p->rdev->ddev;
 	struct radeon_cs_chunk *chunk;
 	struct radeon_cs_buckets buckets;
-	unsigned i, j;
-	bool duplicate;
+	unsigned i;
+	bool need_mmap_lock = false;
+	int r;
 
-	if (p->chunk_relocs_idx == -1) {
+	if (p->chunk_relocs == NULL) {
 		return 0;
 	}
-	chunk = &p->chunks[p->chunk_relocs_idx];
+	chunk = p->chunk_relocs;
 	p->dma_reloc_idx = 0;
 	/* FIXME: we assume that each relocs use 4 dwords */
 	p->nrelocs = chunk->length_dw / 4;
-	p->relocs_ptr = kcalloc(p->nrelocs, sizeof(void *), GFP_KERNEL);
-	if (p->relocs_ptr == NULL) {
-		return -ENOMEM;
-	}
-	p->relocs = kcalloc(p->nrelocs, sizeof(struct radeon_cs_reloc), GFP_KERNEL);
+	p->relocs = kcalloc(p->nrelocs, sizeof(struct radeon_bo_list), GFP_KERNEL);
 	if (p->relocs == NULL) {
 		return -ENOMEM;
 	}
@@ -100,31 +97,17 @@ static int radeon_cs_parser_relocs(struct radeon_cs_parser *p)
 
 	for (i = 0; i < p->nrelocs; i++) {
 		struct drm_radeon_cs_reloc *r;
+		struct drm_gem_object *gobj;
 		unsigned priority;
 
-		duplicate = false;
 		r = (struct drm_radeon_cs_reloc *)&chunk->kdata[i*4];
-		for (j = 0; j < i; j++) {
-			if (r->handle == p->relocs[j].handle) {
-				p->relocs_ptr[i] = &p->relocs[j];
-				duplicate = true;
-				break;
-			}
-		}
-		if (duplicate) {
-			p->relocs[i].handle = 0;
-			continue;
-		}
-
-		p->relocs[i].gobj = drm_gem_object_lookup(ddev, p->filp,
-							  r->handle);
-		if (p->relocs[i].gobj == NULL) {
+		gobj = drm_gem_object_lookup(ddev, p->filp, r->handle);
+		if (gobj == NULL) {
 			DRM_ERROR("gem object lookup failed 0x%x\n",
 				  r->handle);
 			return -ENOENT;
 		}
-		p->relocs_ptr[i] = &p->relocs[i];
-		p->relocs[i].robj = gem_to_radeon_bo(p->relocs[i].gobj);
+		p->relocs[i].robj = gem_to_radeon_bo(gobj);
 
 		/* The userspace buffer priorities are from 0 to 15. A higher
 		 * number means the buffer is more important.
@@ -136,10 +119,13 @@ static int radeon_cs_parser_relocs(struct radeon_cs_parser *p)
 			   + !!r->write_domain;
 
 		/* the first reloc of an UVD job is the msg and that must be in
-		   VRAM, also but everything into VRAM on AGP cards to avoid
-		   image corruptions */
+		   VRAM, also but everything into VRAM on AGP cards and older
+		   IGP chips to avoid image corruptions */
 		if (p->ring == R600_RING_TYPE_UVD_INDEX &&
-		    (i == 0 || drm_pci_device_is_agp(p->rdev->ddev))) {
+		    (i == 0 || drm_pci_device_is_agp(p->rdev->ddev) ||
+		     p->rdev->family == CHIP_RS780 ||
+		     p->rdev->family == CHIP_RS880)) {
+
 			/* TODO: is this still needed for NI+ ? */
 			p->relocs[i].prefered_domains =
 				RADEON_GEM_DOMAIN_VRAM;
@@ -165,8 +151,21 @@ static int radeon_cs_parser_relocs(struct radeon_cs_parser *p)
 			p->relocs[i].allowed_domains = domain;
 		}
 
+		if (radeon_ttm_tt_has_userptr(p->relocs[i].robj->tbo.ttm)) {
+			uint32_t domain = p->relocs[i].prefered_domains;
+			if (!(domain & RADEON_GEM_DOMAIN_GTT)) {
+				DRM_ERROR("Only RADEON_GEM_DOMAIN_GTT is "
+					  "allowed for userptr BOs\n");
+				return -EINVAL;
+			}
+			need_mmap_lock = true;
+			domain = RADEON_GEM_DOMAIN_GTT;
+			p->relocs[i].prefered_domains = domain;
+			p->relocs[i].allowed_domains = domain;
+		}
+
 		p->relocs[i].tv.bo = &p->relocs[i].robj->tbo;
-		p->relocs[i].handle = r->handle;
+		p->relocs[i].tv.shared = !r->write_domain;
 
 		radeon_cs_buckets_add(&buckets, &p->relocs[i].tv.head,
 				      priority);
@@ -177,8 +176,15 @@ static int radeon_cs_parser_relocs(struct radeon_cs_parser *p)
 	if (p->cs_flags & RADEON_CS_USE_VM)
 		p->vm_bos = radeon_vm_get_bos(p->rdev, p->ib.vm,
 					      &p->validated);
+	if (need_mmap_lock)
+		down_read(&current->mm->mmap_sem);
+
+	r = radeon_bo_list_validate(p->rdev, &p->ticket, &p->validated, p->ring);
 
-	return radeon_bo_list_validate(p->rdev, &p->ticket, &p->validated, p->ring);
+	if (need_mmap_lock)
+		up_read(&current->mm->mmap_sem);
+
+	return r;
 }
 
 static int radeon_cs_get_ring(struct radeon_cs_parser *p, u32 ring, s32 priority)
@@ -224,17 +230,21 @@ static int radeon_cs_get_ring(struct radeon_cs_parser *p, u32 ring, s32 priority
 	return 0;
 }
 
-static void radeon_cs_sync_rings(struct radeon_cs_parser *p)
+static int radeon_cs_sync_rings(struct radeon_cs_parser *p)
 {
-	int i;
+	struct radeon_bo_list *reloc;
+	int r;
 
-	for (i = 0; i < p->nrelocs; i++) {
-		if (!p->relocs[i].robj)
-			continue;
+	list_for_each_entry(reloc, &p->validated, tv.head) {
+		struct reservation_object *resv;
 
-		radeon_semaphore_sync_to(p->ib.semaphore,
-					 p->relocs[i].robj->tbo.sync_obj);
+		resv = reloc->robj->tbo.resv;
+		r = radeon_sync_resv(p->rdev, &p->ib.sync, resv,
+				     reloc->tv.shared);
+		if (r)
+			return r;
 	}
+	return 0;
 }
 
 /* XXX: note that this is called from the legacy UMS CS ioctl as well */
@@ -253,13 +263,11 @@ int radeon_cs_parser_init(struct radeon_cs_parser *p, void *data)
 	INIT_LIST_HEAD(&p->validated);
 	p->idx = 0;
 	p->ib.sa_bo = NULL;
-	p->ib.semaphore = NULL;
 	p->const_ib.sa_bo = NULL;
-	p->const_ib.semaphore = NULL;
-	p->chunk_ib_idx = -1;
-	p->chunk_relocs_idx = -1;
-	p->chunk_flags_idx = -1;
-	p->chunk_const_ib_idx = -1;
+	p->chunk_ib = NULL;
+	p->chunk_relocs = NULL;
+	p->chunk_flags = NULL;
+	p->chunk_const_ib = NULL;
 	p->chunks_array = kcalloc(cs->num_chunks, sizeof(uint64_t), GFP_KERNEL);
 	if (p->chunks_array == NULL) {
 		return -ENOMEM;
@@ -286,24 +294,23 @@ int radeon_cs_parser_init(struct radeon_cs_parser *p, void *data)
 			return -EFAULT;
 		}
 		p->chunks[i].length_dw = user_chunk.length_dw;
-		p->chunks[i].chunk_id = user_chunk.chunk_id;
-		if (p->chunks[i].chunk_id == RADEON_CHUNK_ID_RELOCS) {
-			p->chunk_relocs_idx = i;
+		if (user_chunk.chunk_id == RADEON_CHUNK_ID_RELOCS) {
+			p->chunk_relocs = &p->chunks[i];
 		}
-		if (p->chunks[i].chunk_id == RADEON_CHUNK_ID_IB) {
-			p->chunk_ib_idx = i;
+		if (user_chunk.chunk_id == RADEON_CHUNK_ID_IB) {
+			p->chunk_ib = &p->chunks[i];
 			/* zero length IB isn't useful */
 			if (p->chunks[i].length_dw == 0)
 				return -EINVAL;
 		}
-		if (p->chunks[i].chunk_id == RADEON_CHUNK_ID_CONST_IB) {
-			p->chunk_const_ib_idx = i;
+		if (user_chunk.chunk_id == RADEON_CHUNK_ID_CONST_IB) {
+			p->chunk_const_ib = &p->chunks[i];
 			/* zero length CONST IB isn't useful */
 			if (p->chunks[i].length_dw == 0)
 				return -EINVAL;
 		}
-		if (p->chunks[i].chunk_id == RADEON_CHUNK_ID_FLAGS) {
-			p->chunk_flags_idx = i;
+		if (user_chunk.chunk_id == RADEON_CHUNK_ID_FLAGS) {
+			p->chunk_flags = &p->chunks[i];
 			/* zero length flags aren't useful */
 			if (p->chunks[i].length_dw == 0)
 				return -EINVAL;
@@ -312,10 +319,10 @@ int radeon_cs_parser_init(struct radeon_cs_parser *p, void *data)
 		size = p->chunks[i].length_dw;
 		cdata = (void __user *)(unsigned long)user_chunk.chunk_data;
 		p->chunks[i].user_ptr = cdata;
-		if (p->chunks[i].chunk_id == RADEON_CHUNK_ID_CONST_IB)
+		if (user_chunk.chunk_id == RADEON_CHUNK_ID_CONST_IB)
 			continue;
 
-		if (p->chunks[i].chunk_id == RADEON_CHUNK_ID_IB) {
+		if (user_chunk.chunk_id == RADEON_CHUNK_ID_IB) {
 			if (!p->rdev || !(p->rdev->flags & RADEON_IS_AGP))
 				continue;
 		}
@@ -328,7 +335,7 @@ int radeon_cs_parser_init(struct radeon_cs_parser *p, void *data)
 		if (copy_from_user(p->chunks[i].kdata, cdata, size)) {
 			return -EFAULT;
 		}
-		if (p->chunks[i].chunk_id == RADEON_CHUNK_ID_FLAGS) {
+		if (user_chunk.chunk_id == RADEON_CHUNK_ID_FLAGS) {
 			p->cs_flags = p->chunks[i].kdata[0];
 			if (p->chunks[i].length_dw > 1)
 				ring = p->chunks[i].kdata[1];
@@ -369,8 +376,8 @@ int radeon_cs_parser_init(struct radeon_cs_parser *p, void *data)
 static int cmp_size_smaller_first(void *priv, struct list_head *a,
 				  struct list_head *b)
 {
-	struct radeon_cs_reloc *la = list_entry(a, struct radeon_cs_reloc, tv.head);
-	struct radeon_cs_reloc *lb = list_entry(b, struct radeon_cs_reloc, tv.head);
+	struct radeon_bo_list *la = list_entry(a, struct radeon_bo_list, tv.head);
+	struct radeon_bo_list *lb = list_entry(b, struct radeon_bo_list, tv.head);
 
 	/* Sort A before B if A is smaller. */
 	return (int)la->robj->tbo.num_pages - (int)lb->robj->tbo.num_pages;
@@ -403,7 +410,7 @@ static void radeon_cs_parser_fini(struct radeon_cs_parser *parser, int error, bo
 
 		ttm_eu_fence_buffer_objects(&parser->ticket,
 					    &parser->validated,
-					    parser->ib.fence);
+					    &parser->ib.fence->base);
 	} else if (backoff) {
 		ttm_eu_backoff_reservation(&parser->ticket,
 					   &parser->validated);
@@ -411,14 +418,16 @@ static void radeon_cs_parser_fini(struct radeon_cs_parser *parser, int error, bo
 
 	if (parser->relocs != NULL) {
 		for (i = 0; i < parser->nrelocs; i++) {
-			if (parser->relocs[i].gobj)
-				drm_gem_object_unreference_unlocked(parser->relocs[i].gobj);
+			struct radeon_bo *bo = parser->relocs[i].robj;
+			if (bo == NULL)
+				continue;
+
+			drm_gem_object_unreference_unlocked(&bo->gem_base);
 		}
 	}
 	kfree(parser->track);
 	kfree(parser->relocs);
-	kfree(parser->relocs_ptr);
-	kfree(parser->vm_bos);
+	drm_free_large(parser->vm_bos);
 	for (i = 0; i < parser->nchunks; i++)
 		drm_free_large(parser->chunks[i].kdata);
 	kfree(parser->chunks);
@@ -432,7 +441,7 @@ static int radeon_cs_ib_chunk(struct radeon_device *rdev,
 {
 	int r;
 
-	if (parser->chunk_ib_idx == -1)
+	if (parser->chunk_ib == NULL)
 		return 0;
 
 	if (parser->cs_flags & RADEON_CS_USE_VM)
@@ -444,13 +453,19 @@ static int radeon_cs_ib_chunk(struct radeon_device *rdev,
 		return r;
 	}
 
+	r = radeon_cs_sync_rings(parser);
+	if (r) {
+		if (r != -ERESTARTSYS)
+			DRM_ERROR("Failed to sync rings: %i\n", r);
+		return r;
+	}
+
 	if (parser->ring == R600_RING_TYPE_UVD_INDEX)
 		radeon_uvd_note_usage(rdev);
 	else if ((parser->ring == TN_RING_TYPE_VCE1_INDEX) ||
 		 (parser->ring == TN_RING_TYPE_VCE2_INDEX))
 		radeon_vce_note_usage(rdev);
 
-	radeon_cs_sync_rings(parser);
 	r = radeon_ib_schedule(rdev, &parser->ib, NULL, true);
 	if (r) {
 		DRM_ERROR("Failed to schedule IB !\n");
@@ -486,10 +501,6 @@ static int radeon_bo_vm_update_pte(struct radeon_cs_parser *p,
 	for (i = 0; i < p->nrelocs; i++) {
 		struct radeon_bo *bo;
 
-		/* ignore duplicates */
-		if (p->relocs_ptr[i] != &p->relocs[i])
-			continue;
-
 		bo = p->relocs[i].robj;
 		bo_va = radeon_vm_bo_find(vm, bo);
 		if (bo_va == NULL) {
@@ -500,6 +511,8 @@ static int radeon_bo_vm_update_pte(struct radeon_cs_parser *p,
 		r = radeon_vm_bo_update(rdev, bo_va, &bo->tbo.mem);
 		if (r)
 			return r;
+
+		radeon_sync_fence(&p->ib.sync, bo_va->last_pt_update);
 	}
 
 	return radeon_vm_clear_invalids(rdev, vm);
@@ -512,7 +525,7 @@ static int radeon_cs_ib_vm_chunk(struct radeon_device *rdev,
 	struct radeon_vm *vm = &fpriv->vm;
 	int r;
 
-	if (parser->chunk_ib_idx == -1)
+	if (parser->chunk_ib == NULL)
 		return 0;
 	if ((parser->cs_flags & RADEON_CS_USE_VM) == 0)
 		return 0;
@@ -537,11 +550,16 @@ static int radeon_cs_ib_vm_chunk(struct radeon_device *rdev,
 	if (r) {
 		goto out;
 	}
-	radeon_cs_sync_rings(parser);
-	radeon_semaphore_sync_to(parser->ib.semaphore, vm->fence);
+
+	r = radeon_cs_sync_rings(parser);
+	if (r) {
+		if (r != -ERESTARTSYS)
+			DRM_ERROR("Failed to sync rings: %i\n", r);
+		goto out;
+	}
 
 	if ((rdev->family >= CHIP_TAHITI) &&
-	    (parser->chunk_const_ib_idx != -1)) {
+	    (parser->chunk_const_ib != NULL)) {
 		r = radeon_ib_schedule(rdev, &parser->ib, &parser->const_ib, true);
 	} else {
 		r = radeon_ib_schedule(rdev, &parser->ib, NULL, true);
@@ -568,7 +586,7 @@ static int radeon_cs_ib_fill(struct radeon_device *rdev, struct radeon_cs_parser
 	struct radeon_vm *vm = NULL;
 	int r;
 
-	if (parser->chunk_ib_idx == -1)
+	if (parser->chunk_ib == NULL)
 		return 0;
 
 	if (parser->cs_flags & RADEON_CS_USE_VM) {
@@ -576,8 +594,8 @@ static int radeon_cs_ib_fill(struct radeon_device *rdev, struct radeon_cs_parser
 		vm = &fpriv->vm;
 
 		if ((rdev->family >= CHIP_TAHITI) &&
-		    (parser->chunk_const_ib_idx != -1)) {
-			ib_chunk = &parser->chunks[parser->chunk_const_ib_idx];
+		    (parser->chunk_const_ib != NULL)) {
+			ib_chunk = parser->chunk_const_ib;
 			if (ib_chunk->length_dw > RADEON_IB_VM_MAX_SIZE) {
 				DRM_ERROR("cs IB CONST too big: %d\n", ib_chunk->length_dw);
 				return -EINVAL;
@@ -596,13 +614,13 @@ static int radeon_cs_ib_fill(struct radeon_device *rdev, struct radeon_cs_parser
 				return -EFAULT;
 		}
 
-		ib_chunk = &parser->chunks[parser->chunk_ib_idx];
+		ib_chunk = parser->chunk_ib;
 		if (ib_chunk->length_dw > RADEON_IB_VM_MAX_SIZE) {
 			DRM_ERROR("cs IB too big: %d\n", ib_chunk->length_dw);
 			return -EINVAL;
 		}
 	}
-	ib_chunk = &parser->chunks[parser->chunk_ib_idx];
+	ib_chunk = parser->chunk_ib;
 
 	r =  radeon_ib_get(rdev, parser->ring, &parser->ib,
 			   vm, ib_chunk->length_dw * 4);
@@ -629,6 +647,13 @@ int radeon_cs_ioctl(struct drm_device *dev, void *data, struct drm_file *filp)
 		up_read(&rdev->exclusive_lock);
 		return -EBUSY;
 	}
+	if (rdev->in_reset) {
+		up_read(&rdev->exclusive_lock);
+		r = radeon_gpu_reset(rdev);
+		if (!r)
+			r = -EAGAIN;
+		return r;
+	}
 	/* initialize parser */
 	memset(&parser, 0, sizeof(struct radeon_cs_parser));
 	parser.filp = filp;
@@ -687,7 +712,7 @@ int radeon_cs_packet_parse(struct radeon_cs_parser *p,
 			   struct radeon_cs_packet *pkt,
 			   unsigned idx)
 {
-	struct radeon_cs_chunk *ib_chunk = &p->chunks[p->chunk_ib_idx];
+	struct radeon_cs_chunk *ib_chunk = p->chunk_ib;
 	struct radeon_device *rdev = p->rdev;
 	uint32_t header;
 
@@ -781,7 +806,7 @@ void radeon_cs_dump_packet(struct radeon_cs_parser *p,
  * GPU offset using the provided start.
  **/
 int radeon_cs_packet_next_reloc(struct radeon_cs_parser *p,
-				struct radeon_cs_reloc **cs_reloc,
+				struct radeon_bo_list **cs_reloc,
 				int nomm)
 {
 	struct radeon_cs_chunk *relocs_chunk;
@@ -789,12 +814,12 @@ int radeon_cs_packet_next_reloc(struct radeon_cs_parser *p,
 	unsigned idx;
 	int r;
 
-	if (p->chunk_relocs_idx == -1) {
+	if (p->chunk_relocs == NULL) {
 		DRM_ERROR("No relocation chunk !\n");
 		return -EINVAL;
 	}
 	*cs_reloc = NULL;
-	relocs_chunk = &p->chunks[p->chunk_relocs_idx];
+	relocs_chunk = p->chunk_relocs;
 	r = radeon_cs_packet_parse(p, &p3reloc, p->idx);
 	if (r)
 		return r;
@@ -820,6 +845,6 @@ int radeon_cs_packet_next_reloc(struct radeon_cs_parser *p,
 			(u64)relocs_chunk->kdata[idx + 3] << 32;
 		(*cs_reloc)->gpu_offset |= relocs_chunk->kdata[idx + 0];
 	} else
-		*cs_reloc = p->relocs_ptr[(idx / 4)];
+		*cs_reloc = &p->relocs[(idx / 4)];
 	return 0;
 }
diff --git a/drivers/gpu/drm/radeon/radeon_cursor.c b/drivers/gpu/drm/radeon/radeon_cursor.c
index 9630e8d95fb4..45e54060ee97 100644
--- a/drivers/gpu/drm/radeon/radeon_cursor.c
+++ b/drivers/gpu/drm/radeon/radeon_cursor.c
@@ -117,106 +117,7 @@ static void radeon_show_cursor(struct drm_crtc *crtc)
 	}
 }
 
-static void radeon_set_cursor(struct drm_crtc *crtc, struct drm_gem_object *obj,
-			      uint64_t gpu_addr)
-{
-	struct radeon_crtc *radeon_crtc = to_radeon_crtc(crtc);
-	struct radeon_device *rdev = crtc->dev->dev_private;
-
-	if (ASIC_IS_DCE4(rdev)) {
-		WREG32(EVERGREEN_CUR_SURFACE_ADDRESS_HIGH + radeon_crtc->crtc_offset,
-		       upper_32_bits(gpu_addr));
-		WREG32(EVERGREEN_CUR_SURFACE_ADDRESS + radeon_crtc->crtc_offset,
-		       gpu_addr & 0xffffffff);
-	} else if (ASIC_IS_AVIVO(rdev)) {
-		if (rdev->family >= CHIP_RV770) {
-			if (radeon_crtc->crtc_id)
-				WREG32(R700_D2CUR_SURFACE_ADDRESS_HIGH, upper_32_bits(gpu_addr));
-			else
-				WREG32(R700_D1CUR_SURFACE_ADDRESS_HIGH, upper_32_bits(gpu_addr));
-		}
-		WREG32(AVIVO_D1CUR_SURFACE_ADDRESS + radeon_crtc->crtc_offset,
-		       gpu_addr & 0xffffffff);
-	} else {
-		radeon_crtc->legacy_cursor_offset = gpu_addr - radeon_crtc->legacy_display_base_addr;
-		/* offset is from DISP(2)_BASE_ADDRESS */
-		WREG32(RADEON_CUR_OFFSET + radeon_crtc->crtc_offset, radeon_crtc->legacy_cursor_offset);
-	}
-}
-
-int radeon_crtc_cursor_set(struct drm_crtc *crtc,
-			   struct drm_file *file_priv,
-			   uint32_t handle,
-			   uint32_t width,
-			   uint32_t height)
-{
-	struct radeon_crtc *radeon_crtc = to_radeon_crtc(crtc);
-	struct radeon_device *rdev = crtc->dev->dev_private;
-	struct drm_gem_object *obj;
-	struct radeon_bo *robj;
-	uint64_t gpu_addr;
-	int ret;
-
-	if (!handle) {
-		/* turn off cursor */
-		radeon_hide_cursor(crtc);
-		obj = NULL;
-		goto unpin;
-	}
-
-	if ((width > radeon_crtc->max_cursor_width) ||
-	    (height > radeon_crtc->max_cursor_height)) {
-		DRM_ERROR("bad cursor width or height %d x %d\n", width, height);
-		return -EINVAL;
-	}
-
-	obj = drm_gem_object_lookup(crtc->dev, file_priv, handle);
-	if (!obj) {
-		DRM_ERROR("Cannot find cursor object %x for crtc %d\n", handle, radeon_crtc->crtc_id);
-		return -ENOENT;
-	}
-
-	robj = gem_to_radeon_bo(obj);
-	ret = radeon_bo_reserve(robj, false);
-	if (unlikely(ret != 0))
-		goto fail;
-	/* Only 27 bit offset for legacy cursor */
-	ret = radeon_bo_pin_restricted(robj, RADEON_GEM_DOMAIN_VRAM,
-				       ASIC_IS_AVIVO(rdev) ? 0 : 1 << 27,
-				       &gpu_addr);
-	radeon_bo_unreserve(robj);
-	if (ret)
-		goto fail;
-
-	radeon_crtc->cursor_width = width;
-	radeon_crtc->cursor_height = height;
-
-	radeon_lock_cursor(crtc, true);
-	radeon_set_cursor(crtc, obj, gpu_addr);
-	radeon_show_cursor(crtc);
-	radeon_lock_cursor(crtc, false);
-
-unpin:
-	if (radeon_crtc->cursor_bo) {
-		robj = gem_to_radeon_bo(radeon_crtc->cursor_bo);
-		ret = radeon_bo_reserve(robj, false);
-		if (likely(ret == 0)) {
-			radeon_bo_unpin(robj);
-			radeon_bo_unreserve(robj);
-		}
-		drm_gem_object_unreference_unlocked(radeon_crtc->cursor_bo);
-	}
-
-	radeon_crtc->cursor_bo = obj;
-	return 0;
-fail:
-	drm_gem_object_unreference_unlocked(obj);
-
-	return ret;
-}
-
-int radeon_crtc_cursor_move(struct drm_crtc *crtc,
-			    int x, int y)
+static int radeon_cursor_move_locked(struct drm_crtc *crtc, int x, int y)
 {
 	struct radeon_crtc *radeon_crtc = to_radeon_crtc(crtc);
 	struct radeon_device *rdev = crtc->dev->dev_private;
@@ -281,7 +182,6 @@ int radeon_crtc_cursor_move(struct drm_crtc *crtc,
 		}
 	}
 
-	radeon_lock_cursor(crtc, true);
 	if (ASIC_IS_DCE4(rdev)) {
 		WREG32(EVERGREEN_CUR_POSITION + radeon_crtc->crtc_offset, (x << 16) | y);
 		WREG32(EVERGREEN_CUR_HOT_SPOT + radeon_crtc->crtc_offset, (xorigin << 16) | yorigin);
@@ -308,7 +208,173 @@ int radeon_crtc_cursor_move(struct drm_crtc *crtc,
 		WREG32(RADEON_CUR_OFFSET + radeon_crtc->crtc_offset, (radeon_crtc->legacy_cursor_offset +
 								      (yorigin * 256)));
 	}
+
+	radeon_crtc->cursor_x = x;
+	radeon_crtc->cursor_y = y;
+
+	return 0;
+}
+
+int radeon_crtc_cursor_move(struct drm_crtc *crtc,
+			    int x, int y)
+{
+	int ret;
+
+	radeon_lock_cursor(crtc, true);
+	ret = radeon_cursor_move_locked(crtc, x, y);
 	radeon_lock_cursor(crtc, false);
 
+	return ret;
+}
+
+static int radeon_set_cursor(struct drm_crtc *crtc, struct drm_gem_object *obj)
+{
+	struct radeon_crtc *radeon_crtc = to_radeon_crtc(crtc);
+	struct radeon_device *rdev = crtc->dev->dev_private;
+	struct radeon_bo *robj = gem_to_radeon_bo(obj);
+	uint64_t gpu_addr;
+	int ret;
+
+	ret = radeon_bo_reserve(robj, false);
+	if (unlikely(ret != 0))
+		goto fail;
+	/* Only 27 bit offset for legacy cursor */
+	ret = radeon_bo_pin_restricted(robj, RADEON_GEM_DOMAIN_VRAM,
+				       ASIC_IS_AVIVO(rdev) ? 0 : 1 << 27,
+				       &gpu_addr);
+	radeon_bo_unreserve(robj);
+	if (ret)
+		goto fail;
+
+	if (ASIC_IS_DCE4(rdev)) {
+		WREG32(EVERGREEN_CUR_SURFACE_ADDRESS_HIGH + radeon_crtc->crtc_offset,
+		       upper_32_bits(gpu_addr));
+		WREG32(EVERGREEN_CUR_SURFACE_ADDRESS + radeon_crtc->crtc_offset,
+		       gpu_addr & 0xffffffff);
+	} else if (ASIC_IS_AVIVO(rdev)) {
+		if (rdev->family >= CHIP_RV770) {
+			if (radeon_crtc->crtc_id)
+				WREG32(R700_D2CUR_SURFACE_ADDRESS_HIGH, upper_32_bits(gpu_addr));
+			else
+				WREG32(R700_D1CUR_SURFACE_ADDRESS_HIGH, upper_32_bits(gpu_addr));
+		}
+		WREG32(AVIVO_D1CUR_SURFACE_ADDRESS + radeon_crtc->crtc_offset,
+		       gpu_addr & 0xffffffff);
+	} else {
+		radeon_crtc->legacy_cursor_offset = gpu_addr - radeon_crtc->legacy_display_base_addr;
+		/* offset is from DISP(2)_BASE_ADDRESS */
+		WREG32(RADEON_CUR_OFFSET + radeon_crtc->crtc_offset, radeon_crtc->legacy_cursor_offset);
+	}
+
 	return 0;
+
+fail:
+	drm_gem_object_unreference_unlocked(obj);
+
+	return ret;
+}
+
+int radeon_crtc_cursor_set2(struct drm_crtc *crtc,
+			    struct drm_file *file_priv,
+			    uint32_t handle,
+			    uint32_t width,
+			    uint32_t height,
+			    int32_t hot_x,
+			    int32_t hot_y)
+{
+	struct radeon_crtc *radeon_crtc = to_radeon_crtc(crtc);
+	struct drm_gem_object *obj;
+	int ret;
+
+	if (!handle) {
+		/* turn off cursor */
+		radeon_hide_cursor(crtc);
+		obj = NULL;
+		goto unpin;
+	}
+
+	if ((width > radeon_crtc->max_cursor_width) ||
+	    (height > radeon_crtc->max_cursor_height)) {
+		DRM_ERROR("bad cursor width or height %d x %d\n", width, height);
+		return -EINVAL;
+	}
+
+	obj = drm_gem_object_lookup(crtc->dev, file_priv, handle);
+	if (!obj) {
+		DRM_ERROR("Cannot find cursor object %x for crtc %d\n", handle, radeon_crtc->crtc_id);
+		return -ENOENT;
+	}
+
+	radeon_crtc->cursor_width = width;
+	radeon_crtc->cursor_height = height;
+
+	radeon_lock_cursor(crtc, true);
+
+	if (hot_x != radeon_crtc->cursor_hot_x ||
+	    hot_y != radeon_crtc->cursor_hot_y) {
+		int x, y;
+
+		x = radeon_crtc->cursor_x + radeon_crtc->cursor_hot_x - hot_x;
+		y = radeon_crtc->cursor_y + radeon_crtc->cursor_hot_y - hot_y;
+
+		radeon_cursor_move_locked(crtc, x, y);
+
+		radeon_crtc->cursor_hot_x = hot_x;
+		radeon_crtc->cursor_hot_y = hot_y;
+	}
+
+	ret = radeon_set_cursor(crtc, obj);
+
+	if (ret)
+		DRM_ERROR("radeon_set_cursor returned %d, not changing cursor\n",
+			  ret);
+	else
+		radeon_show_cursor(crtc);
+
+	radeon_lock_cursor(crtc, false);
+
+unpin:
+	if (radeon_crtc->cursor_bo) {
+		struct radeon_bo *robj = gem_to_radeon_bo(radeon_crtc->cursor_bo);
+		ret = radeon_bo_reserve(robj, false);
+		if (likely(ret == 0)) {
+			radeon_bo_unpin(robj);
+			radeon_bo_unreserve(robj);
+		}
+		if (radeon_crtc->cursor_bo != obj)
+			drm_gem_object_unreference_unlocked(radeon_crtc->cursor_bo);
+	}
+
+	radeon_crtc->cursor_bo = obj;
+	return 0;
+}
+
+/**
+ * radeon_cursor_reset - Re-set the current cursor, if any.
+ *
+ * @crtc: drm crtc
+ *
+ * If the CRTC passed in currently has a cursor assigned, this function
+ * makes sure it's visible.
+ */
+void radeon_cursor_reset(struct drm_crtc *crtc)
+{
+	struct radeon_crtc *radeon_crtc = to_radeon_crtc(crtc);
+	int ret;
+
+	if (radeon_crtc->cursor_bo) {
+		radeon_lock_cursor(crtc, true);
+
+		radeon_cursor_move_locked(crtc, radeon_crtc->cursor_x,
+					  radeon_crtc->cursor_y);
+
+		ret = radeon_set_cursor(crtc, radeon_crtc->cursor_bo);
+		if (ret)
+			DRM_ERROR("radeon_set_cursor returned %d, not showing "
+				  "cursor\n", ret);
+		else
+			radeon_show_cursor(crtc);
+
+		radeon_lock_cursor(crtc, false);
+	}
 }
diff --git a/drivers/gpu/drm/radeon/radeon_device.c b/drivers/gpu/drm/radeon/radeon_device.c
index 12c8329644c4..0ec65168f331 100644
--- a/drivers/gpu/drm/radeon/radeon_device.c
+++ b/drivers/gpu/drm/radeon/radeon_device.c
@@ -377,6 +377,37 @@ void radeon_doorbell_free(struct radeon_device *rdev, u32 doorbell)
 		__clear_bit(doorbell, rdev->doorbell.used);
 }
 
+/**
+ * radeon_doorbell_get_kfd_info - Report doorbell configuration required to
+ *                                setup KFD
+ *
+ * @rdev: radeon_device pointer
+ * @aperture_base: output returning doorbell aperture base physical address
+ * @aperture_size: output returning doorbell aperture size in bytes
+ * @start_offset: output returning # of doorbell bytes reserved for radeon.
+ *
+ * Radeon and the KFD share the doorbell aperture. Radeon sets it up,
+ * takes doorbells required for its own rings and reports the setup to KFD.
+ * Radeon reserved doorbells are at the start of the doorbell aperture.
+ */
+void radeon_doorbell_get_kfd_info(struct radeon_device *rdev,
+				  phys_addr_t *aperture_base,
+				  size_t *aperture_size,
+				  size_t *start_offset)
+{
+	/* The first num_doorbells are used by radeon.
+	 * KFD takes whatever's left in the aperture. */
+	if (rdev->doorbell.size > rdev->doorbell.num_doorbells * sizeof(u32)) {
+		*aperture_base = rdev->doorbell.base;
+		*aperture_size = rdev->doorbell.size;
+		*start_offset = rdev->doorbell.num_doorbells * sizeof(u32);
+	} else {
+		*aperture_base = 0;
+		*aperture_size = 0;
+		*start_offset = 0;
+	}
+}
+
 /*
  * radeon_wb_*()
  * Writeback is the the method by which the the GPU updates special pages
@@ -434,7 +465,7 @@ int radeon_wb_init(struct radeon_device *rdev)
 
 	if (rdev->wb.wb_obj == NULL) {
 		r = radeon_bo_create(rdev, RADEON_GPU_PAGE_SIZE, PAGE_SIZE, true,
-				     RADEON_GEM_DOMAIN_GTT, 0, NULL,
+				     RADEON_GEM_DOMAIN_GTT, 0, NULL, NULL,
 				     &rdev->wb.wb_obj);
 		if (r) {
 			dev_warn(rdev->dev, "(%d) create WB bo failed\n", r);
@@ -952,6 +983,7 @@ int radeon_atombios_init(struct radeon_device *rdev)
 	}
 
 	mutex_init(&rdev->mode_info.atom_context->mutex);
+	mutex_init(&rdev->mode_info.atom_context->scratch_mutex);
 	radeon_atom_initialize_bios_scratch_regs(rdev->ddev);
 	atom_allocate_fb_scratch(rdev->mode_info.atom_context);
 	return 0;
@@ -1130,7 +1162,7 @@ static void radeon_check_arguments(struct radeon_device *rdev)
 	if (radeon_vm_block_size == -1) {
 
 		/* Total bits covered by PD + PTs */
-		unsigned bits = ilog2(radeon_vm_size) + 17;
+		unsigned bits = ilog2(radeon_vm_size) + 18;
 
 		/* Make sure the PD is 4K in size up to 8GB address space.
 		   Above that split equal between PD and PTs */
@@ -1257,6 +1289,7 @@ int radeon_device_init(struct radeon_device *rdev,
 	for (i = 0; i < RADEON_NUM_RINGS; i++) {
 		rdev->ring[i].idx = i;
 	}
+	rdev->fence_context = fence_context_alloc(RADEON_NUM_RINGS);
 
 	DRM_INFO("initializing kernel modesetting (%s 0x%04X:0x%04X 0x%04X:0x%04X).\n",
 		radeon_family_name[rdev->family], pdev->vendor, pdev->device,
@@ -1271,9 +1304,12 @@ int radeon_device_init(struct radeon_device *rdev,
 	mutex_init(&rdev->pm.mutex);
 	mutex_init(&rdev->gpu_clock_mutex);
 	mutex_init(&rdev->srbm_mutex);
+	mutex_init(&rdev->grbm_idx_mutex);
 	init_rwsem(&rdev->pm.mclk_lock);
 	init_rwsem(&rdev->exclusive_lock);
 	init_waitqueue_head(&rdev->irq.vblank_queue);
+	mutex_init(&rdev->mn_lock);
+	hash_init(rdev->mn_hash);
 	r = radeon_gem_init(rdev);
 	if (r)
 		return r;
@@ -1399,10 +1435,6 @@ int radeon_device_init(struct radeon_device *rdev,
 	if (r)
 		goto failed;
 
-	r = radeon_ib_ring_tests(rdev);
-	if (r)
-		DRM_ERROR("ib ring test failed (%d).\n", r);
-
 	r = radeon_gem_debugfs_init(rdev);
 	if (r) {
 		DRM_ERROR("registering gem debugfs failed (%d).\n", r);
@@ -1420,6 +1452,10 @@ int radeon_device_init(struct radeon_device *rdev,
 			goto failed;
 	}
 
+	r = radeon_ib_ring_tests(rdev);
+	if (r)
+		DRM_ERROR("ib ring test failed (%d).\n", r);
+
 	if ((radeon_testing & 1)) {
 		if (rdev->accel_working)
 			radeon_test_moves(rdev);
@@ -1497,7 +1533,6 @@ int radeon_suspend_kms(struct drm_device *dev, bool suspend, bool fbcon)
 	struct drm_crtc *crtc;
 	struct drm_connector *connector;
 	int i, r;
-	bool force_completion = false;
 
 	if (dev == NULL || dev->dev_private == NULL) {
 		return -ENODEV;
@@ -1541,12 +1576,9 @@ int radeon_suspend_kms(struct drm_device *dev, bool suspend, bool fbcon)
 		r = radeon_fence_wait_empty(rdev, i);
 		if (r) {
 			/* delay GPU reset to resume */
-			force_completion = true;
+			radeon_fence_driver_force_completion(rdev, i);
 		}
 	}
-	if (force_completion) {
-		radeon_fence_driver_force_completion(rdev);
-	}
 
 	radeon_save_bios_scratch_regs(rdev);
 
@@ -1686,8 +1718,6 @@ int radeon_gpu_reset(struct radeon_device *rdev)
 		return 0;
 	}
 
-	rdev->needs_reset = false;
-
 	radeon_save_bios_scratch_regs(rdev);
 	/* block TTM */
 	resched = ttm_bo_lock_delayed_workqueue(&rdev->mman.bdev);
@@ -1704,7 +1734,6 @@ int radeon_gpu_reset(struct radeon_device *rdev)
 		}
 	}
 
-retry:
 	r = radeon_asic_reset(rdev);
 	if (!r) {
 		dev_info(rdev->dev, "GPU reset succeeded, trying to resume\n");
@@ -1713,26 +1742,12 @@ retry:
 
 	radeon_restore_bios_scratch_regs(rdev);
 
-	if (!r) {
-		for (i = 0; i < RADEON_NUM_RINGS; ++i) {
+	for (i = 0; i < RADEON_NUM_RINGS; ++i) {
+		if (!r && ring_data[i]) {
 			radeon_ring_restore(rdev, &rdev->ring[i],
 					    ring_sizes[i], ring_data[i]);
-			ring_sizes[i] = 0;
-			ring_data[i] = NULL;
-		}
-
-		r = radeon_ib_ring_tests(rdev);
-		if (r) {
-			dev_err(rdev->dev, "ib ring test failed (%d).\n", r);
-			if (saved) {
-				saved = false;
-				radeon_suspend(rdev);
-				goto retry;
-			}
-		}
-	} else {
-		radeon_fence_driver_force_completion(rdev);
-		for (i = 0; i < RADEON_NUM_RINGS; ++i) {
+		} else {
+			radeon_fence_driver_force_completion(rdev, i);
 			kfree(ring_data[i]);
 		}
 	}
@@ -1764,19 +1779,32 @@ retry:
 	/* reset hpd state */
 	radeon_hpd_init(rdev);
 
+	ttm_bo_unlock_delayed_workqueue(&rdev->mman.bdev, resched);
+
+	rdev->in_reset = true;
+	rdev->needs_reset = false;
+
+	downgrade_write(&rdev->exclusive_lock);
+
 	drm_helper_resume_force_mode(rdev->ddev);
 
 	/* set the power state here in case we are a PX system or headless */
 	if ((rdev->pm.pm_method == PM_METHOD_DPM) && rdev->pm.dpm_enabled)
 		radeon_pm_compute_clocks(rdev);
 
-	ttm_bo_unlock_delayed_workqueue(&rdev->mman.bdev, resched);
-	if (r) {
+	if (!r) {
+		r = radeon_ib_ring_tests(rdev);
+		if (r && saved)
+			r = -EAGAIN;
+	} else {
 		/* bad news, how to tell it to userspace ? */
 		dev_info(rdev->dev, "GPU reset failed\n");
 	}
 
-	up_write(&rdev->exclusive_lock);
+	rdev->needs_reset = r == -EAGAIN;
+	rdev->in_reset = false;
+
+	up_read(&rdev->exclusive_lock);
 	return r;
 }
 
diff --git a/drivers/gpu/drm/radeon/radeon_display.c b/drivers/gpu/drm/radeon/radeon_display.c
index 3fdf87318069..102116902a07 100644
--- a/drivers/gpu/drm/radeon/radeon_display.c
+++ b/drivers/gpu/drm/radeon/radeon_display.c
@@ -32,6 +32,7 @@
 
 #include <linux/pm_runtime.h>
 #include <drm/drm_crtc_helper.h>
+#include <drm/drm_plane_helper.h>
 #include <drm/drm_edid.h>
 
 #include <linux/gcd.h>
@@ -402,12 +403,21 @@ static void radeon_flip_work_func(struct work_struct *__work)
 
         down_read(&rdev->exclusive_lock);
 	if (work->fence) {
-		r = radeon_fence_wait(work->fence, false);
-		if (r == -EDEADLK) {
-			up_read(&rdev->exclusive_lock);
-			r = radeon_gpu_reset(rdev);
-			down_read(&rdev->exclusive_lock);
-		}
+		struct radeon_fence *fence;
+
+		fence = to_radeon_fence(work->fence);
+		if (fence && fence->rdev == rdev) {
+			r = radeon_fence_wait(fence, false);
+			if (r == -EDEADLK) {
+				up_read(&rdev->exclusive_lock);
+				do {
+					r = radeon_gpu_reset(rdev);
+				} while (r == -EAGAIN);
+				down_read(&rdev->exclusive_lock);
+			}
+		} else
+			r = fence_wait(work->fence, false);
+
 		if (r)
 			DRM_ERROR("failed to wait on page flip fence (%d)!\n", r);
 
@@ -416,7 +426,8 @@ static void radeon_flip_work_func(struct work_struct *__work)
 		 * confused about which BO the CRTC is scanning out
 		 */
 
-		radeon_fence_unref(&work->fence);
+		fence_put(work->fence);
+		work->fence = NULL;
 	}
 
 	/* We borrow the event spin lock for protecting flip_status */
@@ -474,11 +485,6 @@ static int radeon_crtc_page_flip(struct drm_crtc *crtc,
 	obj = new_radeon_fb->obj;
 	new_rbo = gem_to_radeon_bo(obj);
 
-	spin_lock(&new_rbo->tbo.bdev->fence_lock);
-	if (new_rbo->tbo.sync_obj)
-		work->fence = radeon_fence_ref(new_rbo->tbo.sync_obj);
-	spin_unlock(&new_rbo->tbo.bdev->fence_lock);
-
 	/* pin the new buffer */
 	DRM_DEBUG_DRIVER("flip-ioctl() cur_rbo = %p, new_rbo = %p\n",
 			 work->old_rbo, new_rbo);
@@ -497,6 +503,7 @@ static int radeon_crtc_page_flip(struct drm_crtc *crtc,
 		DRM_ERROR("failed to pin new rbo buffer before flip\n");
 		goto cleanup;
 	}
+	work->fence = fence_get(reservation_object_get_excl(new_rbo->tbo.resv));
 	radeon_bo_get_tiling_flags(new_rbo, &tiling_flags, NULL);
 	radeon_bo_unreserve(new_rbo);
 
@@ -578,9 +585,8 @@ pflip_cleanup:
 
 cleanup:
 	drm_gem_object_unreference_unlocked(&work->old_rbo->gem_base);
-	radeon_fence_unref(&work->fence);
+	fence_put(work->fence);
 	kfree(work);
-
 	return r;
 }
 
@@ -629,7 +635,7 @@ radeon_crtc_set_config(struct drm_mode_set *set)
 	return ret;
 }
 static const struct drm_crtc_funcs radeon_crtc_funcs = {
-	.cursor_set = radeon_crtc_cursor_set,
+	.cursor_set2 = radeon_crtc_cursor_set2,
 	.cursor_move = radeon_crtc_cursor_move,
 	.gamma_set = radeon_crtc_gamma_set,
 	.set_config = radeon_crtc_set_config,
@@ -1917,7 +1923,7 @@ int radeon_get_crtc_scanoutpos(struct drm_device *dev, int crtc, unsigned int fl
 
 	/* In vblank? */
 	if (in_vbl)
-		ret |= DRM_SCANOUTPOS_INVBL;
+		ret |= DRM_SCANOUTPOS_IN_VBLANK;
 
 	/* Is vpos outside nominal vblank area, but less than
 	 * 1/100 of a frame height away from start of vblank?
diff --git a/drivers/gpu/drm/radeon/radeon_drv.c b/drivers/gpu/drm/radeon/radeon_drv.c
index f9d17b29b343..4f50fb0e3d93 100644
--- a/drivers/gpu/drm/radeon/radeon_drv.c
+++ b/drivers/gpu/drm/radeon/radeon_drv.c
@@ -38,7 +38,11 @@
 #include <linux/module.h>
 #include <linux/pm_runtime.h>
 #include <linux/vga_switcheroo.h>
+#include <drm/drm_gem.h>
+
 #include "drm_crtc_helper.h"
+#include "radeon_kfd.h"
+
 /*
  * KMS wrapper.
  * - 2.0.0 - initial interface
@@ -114,6 +118,9 @@ int radeon_gem_object_open(struct drm_gem_object *obj,
 				struct drm_file *file_priv);
 void radeon_gem_object_close(struct drm_gem_object *obj,
 				struct drm_file *file_priv);
+struct dma_buf *radeon_gem_prime_export(struct drm_device *dev,
+					struct drm_gem_object *gobj,
+					int flags);
 extern int radeon_get_crtc_scanoutpos(struct drm_device *dev, int crtc,
 				      unsigned int flags,
 				      int *vpos, int *hpos, ktime_t *stime,
@@ -130,7 +137,7 @@ int radeon_mode_dumb_create(struct drm_file *file_priv,
 			    struct drm_mode_create_dumb *args);
 struct sg_table *radeon_gem_prime_get_sg_table(struct drm_gem_object *obj);
 struct drm_gem_object *radeon_gem_prime_import_sg_table(struct drm_device *dev,
-							size_t size,
+							struct dma_buf_attachment *,
 							struct sg_table *sg);
 int radeon_gem_prime_pin(struct drm_gem_object *obj);
 void radeon_gem_prime_unpin(struct drm_gem_object *obj);
@@ -309,7 +316,7 @@ static const struct file_operations radeon_driver_old_fops = {
 	.open = drm_open,
 	.release = drm_release,
 	.unlocked_ioctl = drm_ioctl,
-	.mmap = drm_mmap,
+	.mmap = drm_legacy_mmap,
 	.poll = drm_poll,
 	.read = drm_read,
 #ifdef CONFIG_COMPAT
@@ -329,6 +336,7 @@ static struct drm_driver driver_old = {
 	.preclose = radeon_driver_preclose,
 	.postclose = radeon_driver_postclose,
 	.lastclose = radeon_driver_lastclose,
+	.set_busid = drm_pci_set_busid,
 	.unload = radeon_driver_unload,
 	.suspend = radeon_suspend,
 	.resume = radeon_resume,
@@ -553,6 +561,7 @@ static struct drm_driver kms_driver = {
 	.preclose = radeon_driver_preclose_kms,
 	.postclose = radeon_driver_postclose_kms,
 	.lastclose = radeon_driver_lastclose_kms,
+	.set_busid = drm_pci_set_busid,
 	.unload = radeon_driver_unload_kms,
 	.get_vblank_counter = radeon_get_vblank_counter_kms,
 	.enable_vblank = radeon_enable_vblank_kms,
@@ -578,7 +587,7 @@ static struct drm_driver kms_driver = {
 
 	.prime_handle_to_fd = drm_gem_prime_handle_to_fd,
 	.prime_fd_to_handle = drm_gem_prime_fd_to_handle,
-	.gem_prime_export = drm_gem_prime_export,
+	.gem_prime_export = radeon_gem_prime_export,
 	.gem_prime_import = drm_gem_prime_import,
 	.gem_prime_pin = radeon_gem_prime_pin,
 	.gem_prime_unpin = radeon_gem_prime_unpin,
@@ -647,12 +656,15 @@ static int __init radeon_init(void)
 #endif
 	}
 
+	radeon_kfd_init();
+
 	/* let modprobe override vga console setting */
 	return drm_pci_init(driver, pdriver);
 }
 
 static void __exit radeon_exit(void)
 {
+	radeon_kfd_fini();
 	drm_pci_exit(driver, pdriver);
 	radeon_unregister_atpx_handler();
 }
diff --git a/drivers/gpu/drm/radeon/radeon_drv.h b/drivers/gpu/drm/radeon/radeon_drv.h
index dafd812e4571..46bd3938282c 100644
--- a/drivers/gpu/drm/radeon/radeon_drv.h
+++ b/drivers/gpu/drm/radeon/radeon_drv.h
@@ -33,7 +33,9 @@
 
 #include <linux/firmware.h>
 #include <linux/platform_device.h>
+#include <drm/drm_legacy.h>
 
+#include <drm/ati_pcigart.h>
 #include "radeon_family.h"
 
 /* General customization:
diff --git a/drivers/gpu/drm/radeon/radeon_encoders.c b/drivers/gpu/drm/radeon/radeon_encoders.c
index 15edf23b465c..6b670b0bc47b 100644
--- a/drivers/gpu/drm/radeon/radeon_encoders.c
+++ b/drivers/gpu/drm/radeon/radeon_encoders.c
@@ -179,6 +179,9 @@ static void radeon_encoder_add_backlight(struct radeon_encoder *radeon_encoder,
 		    (rdev->pdev->subsystem_vendor == 0x1734) &&
 		    (rdev->pdev->subsystem_device == 0x1107))
 			use_bl = false;
+		/* disable native backlight control on older asics */
+		else if (rdev->family < CHIP_R600)
+			use_bl = false;
 		else
 			use_bl = true;
 	}
@@ -410,3 +413,24 @@ bool radeon_dig_monitor_is_duallink(struct drm_encoder *encoder,
 	}
 }
 
+bool radeon_encoder_is_digital(struct drm_encoder *encoder)
+{
+	struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder);
+	switch (radeon_encoder->encoder_id) {
+	case ENCODER_OBJECT_ID_INTERNAL_LVDS:
+	case ENCODER_OBJECT_ID_INTERNAL_TMDS1:
+	case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_TMDS1:
+	case ENCODER_OBJECT_ID_INTERNAL_LVTM1:
+	case ENCODER_OBJECT_ID_INTERNAL_DVO1:
+	case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_DVO1:
+	case ENCODER_OBJECT_ID_INTERNAL_DDI:
+	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY:
+	case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_LVTMA:
+	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY1:
+	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY2:
+	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY3:
+		return true;
+	default:
+		return false;
+	}
+}
diff --git a/drivers/gpu/drm/radeon/radeon_fb.c b/drivers/gpu/drm/radeon/radeon_fb.c
index 94b0f2aa3d7c..29b9220ec399 100644
--- a/drivers/gpu/drm/radeon/radeon_fb.c
+++ b/drivers/gpu/drm/radeon/radeon_fb.c
@@ -48,10 +48,40 @@ struct radeon_fbdev {
 	struct radeon_device *rdev;
 };
 
+/**
+ * radeon_fb_helper_set_par - Hide cursor on CRTCs used by fbdev.
+ *
+ * @info: fbdev info
+ *
+ * This function hides the cursor on all CRTCs used by fbdev.
+ */
+static int radeon_fb_helper_set_par(struct fb_info *info)
+{
+	int ret;
+
+	ret = drm_fb_helper_set_par(info);
+
+	/* XXX: with universal plane support fbdev will automatically disable
+	 * all non-primary planes (including the cursor)
+	 */
+	if (ret == 0) {
+		struct drm_fb_helper *fb_helper = info->par;
+		int i;
+
+		for (i = 0; i < fb_helper->crtc_count; i++) {
+			struct drm_crtc *crtc = fb_helper->crtc_info[i].mode_set.crtc;
+
+			radeon_crtc_cursor_set2(crtc, NULL, 0, 0, 0, 0, 0);
+		}
+	}
+
+	return ret;
+}
+
 static struct fb_ops radeonfb_ops = {
 	.owner = THIS_MODULE,
 	.fb_check_var = drm_fb_helper_check_var,
-	.fb_set_par = drm_fb_helper_set_par,
+	.fb_set_par = radeon_fb_helper_set_par,
 	.fb_fillrect = cfb_fillrect,
 	.fb_copyarea = cfb_copyarea,
 	.fb_imageblit = cfb_imageblit,
@@ -189,7 +219,8 @@ out_unref:
 static int radeonfb_create(struct drm_fb_helper *helper,
 			   struct drm_fb_helper_surface_size *sizes)
 {
-	struct radeon_fbdev *rfbdev = (struct radeon_fbdev *)helper;
+	struct radeon_fbdev *rfbdev =
+		container_of(helper, struct radeon_fbdev, helper);
 	struct radeon_device *rdev = rfbdev->rdev;
 	struct fb_info *info;
 	struct drm_framebuffer *fb = NULL;
diff --git a/drivers/gpu/drm/radeon/radeon_fence.c b/drivers/gpu/drm/radeon/radeon_fence.c
index 913787085dfa..d13d1b5a859f 100644
--- a/drivers/gpu/drm/radeon/radeon_fence.c
+++ b/drivers/gpu/drm/radeon/radeon_fence.c
@@ -98,6 +98,25 @@ static u32 radeon_fence_read(struct radeon_device *rdev, int ring)
 }
 
 /**
+ * radeon_fence_schedule_check - schedule lockup check
+ *
+ * @rdev: radeon_device pointer
+ * @ring: ring index we should work with
+ *
+ * Queues a delayed work item to check for lockups.
+ */
+static void radeon_fence_schedule_check(struct radeon_device *rdev, int ring)
+{
+	/*
+	 * Do not reset the timer here with mod_delayed_work,
+	 * this can livelock in an interaction with TTM delayed destroy.
+	 */
+	queue_delayed_work(system_power_efficient_wq,
+			   &rdev->fence_drv[ring].lockup_work,
+			   RADEON_FENCE_JIFFIES_TIMEOUT);
+}
+
+/**
  * radeon_fence_emit - emit a fence on the requested ring
  *
  * @rdev: radeon_device pointer
@@ -111,30 +130,71 @@ int radeon_fence_emit(struct radeon_device *rdev,
 		      struct radeon_fence **fence,
 		      int ring)
 {
+	u64 seq = ++rdev->fence_drv[ring].sync_seq[ring];
+
 	/* we are protected by the ring emission mutex */
 	*fence = kmalloc(sizeof(struct radeon_fence), GFP_KERNEL);
 	if ((*fence) == NULL) {
 		return -ENOMEM;
 	}
-	kref_init(&((*fence)->kref));
 	(*fence)->rdev = rdev;
-	(*fence)->seq = ++rdev->fence_drv[ring].sync_seq[ring];
+	(*fence)->seq = seq;
 	(*fence)->ring = ring;
+	(*fence)->is_vm_update = false;
+	fence_init(&(*fence)->base, &radeon_fence_ops,
+		   &rdev->fence_queue.lock, rdev->fence_context + ring, seq);
 	radeon_fence_ring_emit(rdev, ring, *fence);
 	trace_radeon_fence_emit(rdev->ddev, ring, (*fence)->seq);
+	radeon_fence_schedule_check(rdev, ring);
 	return 0;
 }
 
 /**
- * radeon_fence_process - process a fence
+ * radeon_fence_check_signaled - callback from fence_queue
+ *
+ * this function is called with fence_queue lock held, which is also used
+ * for the fence locking itself, so unlocked variants are used for
+ * fence_signal, and remove_wait_queue.
+ */
+static int radeon_fence_check_signaled(wait_queue_t *wait, unsigned mode, int flags, void *key)
+{
+	struct radeon_fence *fence;
+	u64 seq;
+
+	fence = container_of(wait, struct radeon_fence, fence_wake);
+
+	/*
+	 * We cannot use radeon_fence_process here because we're already
+	 * in the waitqueue, in a call from wake_up_all.
+	 */
+	seq = atomic64_read(&fence->rdev->fence_drv[fence->ring].last_seq);
+	if (seq >= fence->seq) {
+		int ret = fence_signal_locked(&fence->base);
+
+		if (!ret)
+			FENCE_TRACE(&fence->base, "signaled from irq context\n");
+		else
+			FENCE_TRACE(&fence->base, "was already signaled\n");
+
+		radeon_irq_kms_sw_irq_put(fence->rdev, fence->ring);
+		__remove_wait_queue(&fence->rdev->fence_queue, &fence->fence_wake);
+		fence_put(&fence->base);
+	} else
+		FENCE_TRACE(&fence->base, "pending\n");
+	return 0;
+}
+
+/**
+ * radeon_fence_activity - check for fence activity
  *
  * @rdev: radeon_device pointer
  * @ring: ring index the fence is associated with
  *
- * Checks the current fence value and wakes the fence queue
- * if the sequence number has increased (all asics).
+ * Checks the current fence value and calculates the last
+ * signalled fence value. Returns true if activity occured
+ * on the ring, and the fence_queue should be waken up.
  */
-void radeon_fence_process(struct radeon_device *rdev, int ring)
+static bool radeon_fence_activity(struct radeon_device *rdev, int ring)
 {
 	uint64_t seq, last_seq, last_emitted;
 	unsigned count_loop = 0;
@@ -190,23 +250,77 @@ void radeon_fence_process(struct radeon_device *rdev, int ring)
 		}
 	} while (atomic64_xchg(&rdev->fence_drv[ring].last_seq, seq) > seq);
 
-	if (wake)
-		wake_up_all(&rdev->fence_queue);
+	if (seq < last_emitted)
+		radeon_fence_schedule_check(rdev, ring);
+
+	return wake;
 }
 
 /**
- * radeon_fence_destroy - destroy a fence
+ * radeon_fence_check_lockup - check for hardware lockup
  *
- * @kref: fence kref
+ * @work: delayed work item
  *
- * Frees the fence object (all asics).
+ * Checks for fence activity and if there is none probe
+ * the hardware if a lockup occured.
  */
-static void radeon_fence_destroy(struct kref *kref)
+static void radeon_fence_check_lockup(struct work_struct *work)
 {
-	struct radeon_fence *fence;
+	struct radeon_fence_driver *fence_drv;
+	struct radeon_device *rdev;
+	int ring;
+
+	fence_drv = container_of(work, struct radeon_fence_driver,
+				 lockup_work.work);
+	rdev = fence_drv->rdev;
+	ring = fence_drv - &rdev->fence_drv[0];
+
+	if (!down_read_trylock(&rdev->exclusive_lock)) {
+		/* just reschedule the check if a reset is going on */
+		radeon_fence_schedule_check(rdev, ring);
+		return;
+	}
+
+	if (fence_drv->delayed_irq && rdev->ddev->irq_enabled) {
+		unsigned long irqflags;
+
+		fence_drv->delayed_irq = false;
+		spin_lock_irqsave(&rdev->irq.lock, irqflags);
+		radeon_irq_set(rdev);
+		spin_unlock_irqrestore(&rdev->irq.lock, irqflags);
+	}
+
+	if (radeon_fence_activity(rdev, ring))
+		wake_up_all(&rdev->fence_queue);
 
-	fence = container_of(kref, struct radeon_fence, kref);
-	kfree(fence);
+	else if (radeon_ring_is_lockup(rdev, ring, &rdev->ring[ring])) {
+
+		/* good news we believe it's a lockup */
+		dev_warn(rdev->dev, "GPU lockup (current fence id "
+			 "0x%016llx last fence id 0x%016llx on ring %d)\n",
+			 (uint64_t)atomic64_read(&fence_drv->last_seq),
+			 fence_drv->sync_seq[ring], ring);
+
+		/* remember that we need an reset */
+		rdev->needs_reset = true;
+		wake_up_all(&rdev->fence_queue);
+	}
+	up_read(&rdev->exclusive_lock);
+}
+
+/**
+ * radeon_fence_process - process a fence
+ *
+ * @rdev: radeon_device pointer
+ * @ring: ring index the fence is associated with
+ *
+ * Checks the current fence value and wakes the fence queue
+ * if the sequence number has increased (all asics).
+ */
+void radeon_fence_process(struct radeon_device *rdev, int ring)
+{
+	if (radeon_fence_activity(rdev, ring))
+		wake_up_all(&rdev->fence_queue);
 }
 
 /**
@@ -237,6 +351,75 @@ static bool radeon_fence_seq_signaled(struct radeon_device *rdev,
 	return false;
 }
 
+static bool radeon_fence_is_signaled(struct fence *f)
+{
+	struct radeon_fence *fence = to_radeon_fence(f);
+	struct radeon_device *rdev = fence->rdev;
+	unsigned ring = fence->ring;
+	u64 seq = fence->seq;
+
+	if (atomic64_read(&rdev->fence_drv[ring].last_seq) >= seq) {
+		return true;
+	}
+
+	if (down_read_trylock(&rdev->exclusive_lock)) {
+		radeon_fence_process(rdev, ring);
+		up_read(&rdev->exclusive_lock);
+
+		if (atomic64_read(&rdev->fence_drv[ring].last_seq) >= seq) {
+			return true;
+		}
+	}
+	return false;
+}
+
+/**
+ * radeon_fence_enable_signaling - enable signalling on fence
+ * @fence: fence
+ *
+ * This function is called with fence_queue lock held, and adds a callback
+ * to fence_queue that checks if this fence is signaled, and if so it
+ * signals the fence and removes itself.
+ */
+static bool radeon_fence_enable_signaling(struct fence *f)
+{
+	struct radeon_fence *fence = to_radeon_fence(f);
+	struct radeon_device *rdev = fence->rdev;
+
+	if (atomic64_read(&rdev->fence_drv[fence->ring].last_seq) >= fence->seq)
+		return false;
+
+	if (down_read_trylock(&rdev->exclusive_lock)) {
+		radeon_irq_kms_sw_irq_get(rdev, fence->ring);
+
+		if (radeon_fence_activity(rdev, fence->ring))
+			wake_up_all_locked(&rdev->fence_queue);
+
+		/* did fence get signaled after we enabled the sw irq? */
+		if (atomic64_read(&rdev->fence_drv[fence->ring].last_seq) >= fence->seq) {
+			radeon_irq_kms_sw_irq_put(rdev, fence->ring);
+			up_read(&rdev->exclusive_lock);
+			return false;
+		}
+
+		up_read(&rdev->exclusive_lock);
+	} else {
+		/* we're probably in a lockup, lets not fiddle too much */
+		if (radeon_irq_kms_sw_irq_get_delayed(rdev, fence->ring))
+			rdev->fence_drv[fence->ring].delayed_irq = true;
+		radeon_fence_schedule_check(rdev, fence->ring);
+	}
+
+	fence->fence_wake.flags = 0;
+	fence->fence_wake.private = NULL;
+	fence->fence_wake.func = radeon_fence_check_signaled;
+	__add_wait_queue(&rdev->fence_queue, &fence->fence_wake);
+	fence_get(f);
+
+	FENCE_TRACE(&fence->base, "armed on ring %i!\n", fence->ring);
+	return true;
+}
+
 /**
  * radeon_fence_signaled - check if a fence has signaled
  *
@@ -247,14 +430,15 @@ static bool radeon_fence_seq_signaled(struct radeon_device *rdev,
  */
 bool radeon_fence_signaled(struct radeon_fence *fence)
 {
-	if (!fence) {
+	if (!fence)
 		return true;
-	}
-	if (fence->seq == RADEON_FENCE_SIGNALED_SEQ) {
-		return true;
-	}
+
 	if (radeon_fence_seq_signaled(fence->rdev, fence->seq, fence->ring)) {
-		fence->seq = RADEON_FENCE_SIGNALED_SEQ;
+		int ret;
+
+		ret = fence_signal(&fence->base);
+		if (!ret)
+			FENCE_TRACE(&fence->base, "signaled from radeon_fence_signaled\n");
 		return true;
 	}
 	return false;
@@ -283,110 +467,70 @@ static bool radeon_fence_any_seq_signaled(struct radeon_device *rdev, u64 *seq)
 }
 
 /**
- * radeon_fence_wait_seq - wait for a specific sequence numbers
+ * radeon_fence_wait_seq_timeout - wait for a specific sequence numbers
  *
  * @rdev: radeon device pointer
  * @target_seq: sequence number(s) we want to wait for
  * @intr: use interruptable sleep
+ * @timeout: maximum time to wait, or MAX_SCHEDULE_TIMEOUT for infinite wait
  *
  * Wait for the requested sequence number(s) to be written by any ring
  * (all asics).  Sequnce number array is indexed by ring id.
  * @intr selects whether to use interruptable (true) or non-interruptable
  * (false) sleep when waiting for the sequence number.  Helper function
  * for radeon_fence_wait_*().
- * Returns 0 if the sequence number has passed, error for all other cases.
+ * Returns remaining time if the sequence number has passed, 0 when
+ * the wait timeout, or an error for all other cases.
  * -EDEADLK is returned when a GPU lockup has been detected.
  */
-static int radeon_fence_wait_seq(struct radeon_device *rdev, u64 *target_seq,
-				 bool intr)
+static long radeon_fence_wait_seq_timeout(struct radeon_device *rdev,
+					  u64 *target_seq, bool intr,
+					  long timeout)
 {
-	uint64_t last_seq[RADEON_NUM_RINGS];
-	bool signaled;
-	int i, r;
-
-	while (!radeon_fence_any_seq_signaled(rdev, target_seq)) {
+	long r;
+	int i;
 
-		/* Save current sequence values, used to check for GPU lockups */
-		for (i = 0; i < RADEON_NUM_RINGS; ++i) {
-			if (!target_seq[i])
-				continue;
+	if (radeon_fence_any_seq_signaled(rdev, target_seq))
+		return timeout;
 
-			last_seq[i] = atomic64_read(&rdev->fence_drv[i].last_seq);
-			trace_radeon_fence_wait_begin(rdev->ddev, i, target_seq[i]);
-			radeon_irq_kms_sw_irq_get(rdev, i);
-		}
+	/* enable IRQs and tracing */
+	for (i = 0; i < RADEON_NUM_RINGS; ++i) {
+		if (!target_seq[i])
+			continue;
 
-		if (intr) {
-			r = wait_event_interruptible_timeout(rdev->fence_queue, (
-				(signaled = radeon_fence_any_seq_signaled(rdev, target_seq))
-				 || rdev->needs_reset), RADEON_FENCE_JIFFIES_TIMEOUT);
-		} else {
-			r = wait_event_timeout(rdev->fence_queue, (
-				(signaled = radeon_fence_any_seq_signaled(rdev, target_seq))
-				 || rdev->needs_reset), RADEON_FENCE_JIFFIES_TIMEOUT);
-		}
+		trace_radeon_fence_wait_begin(rdev->ddev, i, target_seq[i]);
+		radeon_irq_kms_sw_irq_get(rdev, i);
+	}
 
-		for (i = 0; i < RADEON_NUM_RINGS; ++i) {
-			if (!target_seq[i])
-				continue;
+	if (intr) {
+		r = wait_event_interruptible_timeout(rdev->fence_queue, (
+			radeon_fence_any_seq_signaled(rdev, target_seq)
+			 || rdev->needs_reset), timeout);
+	} else {
+		r = wait_event_timeout(rdev->fence_queue, (
+			radeon_fence_any_seq_signaled(rdev, target_seq)
+			 || rdev->needs_reset), timeout);
+	}
 
-			radeon_irq_kms_sw_irq_put(rdev, i);
-			trace_radeon_fence_wait_end(rdev->ddev, i, target_seq[i]);
-		}
+	if (rdev->needs_reset)
+		r = -EDEADLK;
 
-		if (unlikely(r < 0))
-			return r;
+	for (i = 0; i < RADEON_NUM_RINGS; ++i) {
+		if (!target_seq[i])
+			continue;
 
-		if (unlikely(!signaled)) {
-			if (rdev->needs_reset)
-				return -EDEADLK;
-
-			/* we were interrupted for some reason and fence
-			 * isn't signaled yet, resume waiting */
-			if (r)
-				continue;
-
-			for (i = 0; i < RADEON_NUM_RINGS; ++i) {
-				if (!target_seq[i])
-					continue;
-
-				if (last_seq[i] != atomic64_read(&rdev->fence_drv[i].last_seq))
-					break;
-			}
-
-			if (i != RADEON_NUM_RINGS)
-				continue;
-
-			for (i = 0; i < RADEON_NUM_RINGS; ++i) {
-				if (!target_seq[i])
-					continue;
-
-				if (radeon_ring_is_lockup(rdev, i, &rdev->ring[i]))
-					break;
-			}
-
-			if (i < RADEON_NUM_RINGS) {
-				/* good news we believe it's a lockup */
-				dev_warn(rdev->dev, "GPU lockup (waiting for "
-					 "0x%016llx last fence id 0x%016llx on"
-					 " ring %d)\n",
-					 target_seq[i], last_seq[i], i);
-
-				/* remember that we need an reset */
-				rdev->needs_reset = true;
-				wake_up_all(&rdev->fence_queue);
-				return -EDEADLK;
-			}
-		}
+		radeon_irq_kms_sw_irq_put(rdev, i);
+		trace_radeon_fence_wait_end(rdev->ddev, i, target_seq[i]);
 	}
-	return 0;
+
+	return r;
 }
 
 /**
  * radeon_fence_wait - wait for a fence to signal
  *
  * @fence: radeon fence object
- * @intr: use interruptable sleep
+ * @intr: use interruptible sleep
  *
  * Wait for the requested fence to signal (all asics).
  * @intr selects whether to use interruptable (true) or non-interruptable
@@ -396,22 +540,26 @@ static int radeon_fence_wait_seq(struct radeon_device *rdev, u64 *target_seq,
 int radeon_fence_wait(struct radeon_fence *fence, bool intr)
 {
 	uint64_t seq[RADEON_NUM_RINGS] = {};
-	int r;
+	long r;
 
-	if (fence == NULL) {
-		WARN(1, "Querying an invalid fence : %p !\n", fence);
-		return -EINVAL;
-	}
+	/*
+	 * This function should not be called on !radeon fences.
+	 * If this is the case, it would mean this function can
+	 * also be called on radeon fences belonging to another card.
+	 * exclusive_lock is not held in that case.
+	 */
+	if (WARN_ON_ONCE(!to_radeon_fence(&fence->base)))
+		return fence_wait(&fence->base, intr);
 
 	seq[fence->ring] = fence->seq;
-	if (seq[fence->ring] == RADEON_FENCE_SIGNALED_SEQ)
-		return 0;
-
-	r = radeon_fence_wait_seq(fence->rdev, seq, intr);
-	if (r)
+	r = radeon_fence_wait_seq_timeout(fence->rdev, seq, intr, MAX_SCHEDULE_TIMEOUT);
+	if (r < 0) {
 		return r;
+	}
 
-	fence->seq = RADEON_FENCE_SIGNALED_SEQ;
+	r = fence_signal(&fence->base);
+	if (!r)
+		FENCE_TRACE(&fence->base, "signaled from fence_wait\n");
 	return 0;
 }
 
@@ -434,7 +582,7 @@ int radeon_fence_wait_any(struct radeon_device *rdev,
 {
 	uint64_t seq[RADEON_NUM_RINGS];
 	unsigned i, num_rings = 0;
-	int r;
+	long r;
 
 	for (i = 0; i < RADEON_NUM_RINGS; ++i) {
 		seq[i] = 0;
@@ -445,18 +593,14 @@ int radeon_fence_wait_any(struct radeon_device *rdev,
 
 		seq[i] = fences[i]->seq;
 		++num_rings;
-
-		/* test if something was allready signaled */
-		if (seq[i] == RADEON_FENCE_SIGNALED_SEQ)
-			return 0;
 	}
 
 	/* nothing to wait for ? */
 	if (num_rings == 0)
 		return -ENOENT;
 
-	r = radeon_fence_wait_seq(rdev, seq, intr);
-	if (r) {
+	r = radeon_fence_wait_seq_timeout(rdev, seq, intr, MAX_SCHEDULE_TIMEOUT);
+	if (r < 0) {
 		return r;
 	}
 	return 0;
@@ -475,6 +619,7 @@ int radeon_fence_wait_any(struct radeon_device *rdev,
 int radeon_fence_wait_next(struct radeon_device *rdev, int ring)
 {
 	uint64_t seq[RADEON_NUM_RINGS] = {};
+	long r;
 
 	seq[ring] = atomic64_read(&rdev->fence_drv[ring].last_seq) + 1ULL;
 	if (seq[ring] >= rdev->fence_drv[ring].sync_seq[ring]) {
@@ -482,7 +627,10 @@ int radeon_fence_wait_next(struct radeon_device *rdev, int ring)
 		   already the last emited fence */
 		return -ENOENT;
 	}
-	return radeon_fence_wait_seq(rdev, seq, false);
+	r = radeon_fence_wait_seq_timeout(rdev, seq, false, MAX_SCHEDULE_TIMEOUT);
+	if (r < 0)
+		return r;
+	return 0;
 }
 
 /**
@@ -498,18 +646,18 @@ int radeon_fence_wait_next(struct radeon_device *rdev, int ring)
 int radeon_fence_wait_empty(struct radeon_device *rdev, int ring)
 {
 	uint64_t seq[RADEON_NUM_RINGS] = {};
-	int r;
+	long r;
 
 	seq[ring] = rdev->fence_drv[ring].sync_seq[ring];
 	if (!seq[ring])
 		return 0;
 
-	r = radeon_fence_wait_seq(rdev, seq, false);
-	if (r) {
+	r = radeon_fence_wait_seq_timeout(rdev, seq, false, MAX_SCHEDULE_TIMEOUT);
+	if (r < 0) {
 		if (r == -EDEADLK)
 			return -EDEADLK;
 
-		dev_err(rdev->dev, "error waiting for ring[%d] to become idle (%d)\n",
+		dev_err(rdev->dev, "error waiting for ring[%d] to become idle (%ld)\n",
 			ring, r);
 	}
 	return 0;
@@ -525,7 +673,7 @@ int radeon_fence_wait_empty(struct radeon_device *rdev, int ring)
  */
 struct radeon_fence *radeon_fence_ref(struct radeon_fence *fence)
 {
-	kref_get(&fence->kref);
+	fence_get(&fence->base);
 	return fence;
 }
 
@@ -542,7 +690,7 @@ void radeon_fence_unref(struct radeon_fence **fence)
 
 	*fence = NULL;
 	if (tmp) {
-		kref_put(&tmp->kref, radeon_fence_destroy);
+		fence_put(&tmp->base);
 	}
 }
 
@@ -711,6 +859,9 @@ static void radeon_fence_driver_init_ring(struct radeon_device *rdev, int ring)
 		rdev->fence_drv[ring].sync_seq[i] = 0;
 	atomic64_set(&rdev->fence_drv[ring].last_seq, 0);
 	rdev->fence_drv[ring].initialized = false;
+	INIT_DELAYED_WORK(&rdev->fence_drv[ring].lockup_work,
+			  radeon_fence_check_lockup);
+	rdev->fence_drv[ring].rdev = rdev;
 }
 
 /**
@@ -758,8 +909,9 @@ void radeon_fence_driver_fini(struct radeon_device *rdev)
 		r = radeon_fence_wait_empty(rdev, ring);
 		if (r) {
 			/* no need to trigger GPU reset as we are unloading */
-			radeon_fence_driver_force_completion(rdev);
+			radeon_fence_driver_force_completion(rdev, ring);
 		}
+		cancel_delayed_work_sync(&rdev->fence_drv[ring].lockup_work);
 		wake_up_all(&rdev->fence_queue);
 		radeon_scratch_free(rdev, rdev->fence_drv[ring].scratch_reg);
 		rdev->fence_drv[ring].initialized = false;
@@ -771,18 +923,16 @@ void radeon_fence_driver_fini(struct radeon_device *rdev)
  * radeon_fence_driver_force_completion - force all fence waiter to complete
  *
  * @rdev: radeon device pointer
+ * @ring: the ring to complete
  *
  * In case of GPU reset failure make sure no process keep waiting on fence
  * that will never complete.
  */
-void radeon_fence_driver_force_completion(struct radeon_device *rdev)
+void radeon_fence_driver_force_completion(struct radeon_device *rdev, int ring)
 {
-	int ring;
-
-	for (ring = 0; ring < RADEON_NUM_RINGS; ring++) {
-		if (!rdev->fence_drv[ring].initialized)
-			continue;
+	if (rdev->fence_drv[ring].initialized) {
 		radeon_fence_write(rdev, rdev->fence_drv[ring].sync_seq[ring], ring);
+		cancel_delayed_work_sync(&rdev->fence_drv[ring].lockup_work);
 	}
 }
 
@@ -833,6 +983,7 @@ static int radeon_debugfs_gpu_reset(struct seq_file *m, void *data)
 	down_read(&rdev->exclusive_lock);
 	seq_printf(m, "%d\n", rdev->needs_reset);
 	rdev->needs_reset = true;
+	wake_up_all(&rdev->fence_queue);
 	up_read(&rdev->exclusive_lock);
 
 	return 0;
@@ -852,3 +1003,72 @@ int radeon_debugfs_fence_init(struct radeon_device *rdev)
 	return 0;
 #endif
 }
+
+static const char *radeon_fence_get_driver_name(struct fence *fence)
+{
+	return "radeon";
+}
+
+static const char *radeon_fence_get_timeline_name(struct fence *f)
+{
+	struct radeon_fence *fence = to_radeon_fence(f);
+	switch (fence->ring) {
+	case RADEON_RING_TYPE_GFX_INDEX: return "radeon.gfx";
+	case CAYMAN_RING_TYPE_CP1_INDEX: return "radeon.cp1";
+	case CAYMAN_RING_TYPE_CP2_INDEX: return "radeon.cp2";
+	case R600_RING_TYPE_DMA_INDEX: return "radeon.dma";
+	case CAYMAN_RING_TYPE_DMA1_INDEX: return "radeon.dma1";
+	case R600_RING_TYPE_UVD_INDEX: return "radeon.uvd";
+	case TN_RING_TYPE_VCE1_INDEX: return "radeon.vce1";
+	case TN_RING_TYPE_VCE2_INDEX: return "radeon.vce2";
+	default: WARN_ON_ONCE(1); return "radeon.unk";
+	}
+}
+
+static inline bool radeon_test_signaled(struct radeon_fence *fence)
+{
+	return test_bit(FENCE_FLAG_SIGNALED_BIT, &fence->base.flags);
+}
+
+static signed long radeon_fence_default_wait(struct fence *f, bool intr,
+					     signed long t)
+{
+	struct radeon_fence *fence = to_radeon_fence(f);
+	struct radeon_device *rdev = fence->rdev;
+	bool signaled;
+
+	fence_enable_sw_signaling(&fence->base);
+
+	/*
+	 * This function has to return -EDEADLK, but cannot hold
+	 * exclusive_lock during the wait because some callers
+	 * may already hold it. This means checking needs_reset without
+	 * lock, and not fiddling with any gpu internals.
+	 *
+	 * The callback installed with fence_enable_sw_signaling will
+	 * run before our wait_event_*timeout call, so we will see
+	 * both the signaled fence and the changes to needs_reset.
+	 */
+
+	if (intr)
+		t = wait_event_interruptible_timeout(rdev->fence_queue,
+			((signaled = radeon_test_signaled(fence)) ||
+			 rdev->needs_reset), t);
+	else
+		t = wait_event_timeout(rdev->fence_queue,
+			((signaled = radeon_test_signaled(fence)) ||
+			 rdev->needs_reset), t);
+
+	if (t > 0 && !signaled)
+		return -EDEADLK;
+	return t;
+}
+
+const struct fence_ops radeon_fence_ops = {
+	.get_driver_name = radeon_fence_get_driver_name,
+	.get_timeline_name = radeon_fence_get_timeline_name,
+	.enable_signaling = radeon_fence_enable_signaling,
+	.signaled = radeon_fence_is_signaled,
+	.wait = radeon_fence_default_wait,
+	.release = NULL,
+};
diff --git a/drivers/gpu/drm/radeon/radeon_gart.c b/drivers/gpu/drm/radeon/radeon_gart.c
index a053a0779aac..84146d5901aa 100644
--- a/drivers/gpu/drm/radeon/radeon_gart.c
+++ b/drivers/gpu/drm/radeon/radeon_gart.c
@@ -128,7 +128,7 @@ int radeon_gart_table_vram_alloc(struct radeon_device *rdev)
 	if (rdev->gart.robj == NULL) {
 		r = radeon_bo_create(rdev, rdev->gart.table_size,
 				     PAGE_SIZE, true, RADEON_GEM_DOMAIN_VRAM,
-				     0, NULL, &rdev->gart.robj);
+				     0, NULL, NULL, &rdev->gart.robj);
 		if (r) {
 			return r;
 		}
diff --git a/drivers/gpu/drm/radeon/radeon_gem.c b/drivers/gpu/drm/radeon/radeon_gem.c
index bfd7e1b0ff3f..a46f73737994 100644
--- a/drivers/gpu/drm/radeon/radeon_gem.c
+++ b/drivers/gpu/drm/radeon/radeon_gem.c
@@ -67,7 +67,7 @@ int radeon_gem_object_create(struct radeon_device *rdev, unsigned long size,
 
 retry:
 	r = radeon_bo_create(rdev, size, alignment, kernel, initial_domain,
-			     flags, NULL, &robj);
+			     flags, NULL, NULL, &robj);
 	if (r) {
 		if (r != -ERESTARTSYS) {
 			if (initial_domain == RADEON_GEM_DOMAIN_VRAM) {
@@ -94,7 +94,7 @@ static int radeon_gem_set_domain(struct drm_gem_object *gobj,
 {
 	struct radeon_bo *robj;
 	uint32_t domain;
-	int r;
+	long r;
 
 	/* FIXME: reeimplement */
 	robj = gem_to_radeon_bo(gobj);
@@ -110,9 +110,12 @@ static int radeon_gem_set_domain(struct drm_gem_object *gobj,
 	}
 	if (domain == RADEON_GEM_DOMAIN_CPU) {
 		/* Asking for cpu access wait for object idle */
-		r = radeon_bo_wait(robj, NULL, false);
-		if (r) {
-			printk(KERN_ERR "Failed to wait for object !\n");
+		r = reservation_object_wait_timeout_rcu(robj->tbo.resv, true, true, 30 * HZ);
+		if (!r)
+			r = -EBUSY;
+
+		if (r < 0 && r != -EINTR) {
+			printk(KERN_ERR "Failed to wait for object: %li\n", r);
 			return r;
 		}
 	}
@@ -272,6 +275,94 @@ int radeon_gem_create_ioctl(struct drm_device *dev, void *data,
 	return 0;
 }
 
+int radeon_gem_userptr_ioctl(struct drm_device *dev, void *data,
+			     struct drm_file *filp)
+{
+	struct radeon_device *rdev = dev->dev_private;
+	struct drm_radeon_gem_userptr *args = data;
+	struct drm_gem_object *gobj;
+	struct radeon_bo *bo;
+	uint32_t handle;
+	int r;
+
+	if (offset_in_page(args->addr | args->size))
+		return -EINVAL;
+
+	/* reject unknown flag values */
+	if (args->flags & ~(RADEON_GEM_USERPTR_READONLY |
+	    RADEON_GEM_USERPTR_ANONONLY | RADEON_GEM_USERPTR_VALIDATE |
+	    RADEON_GEM_USERPTR_REGISTER))
+		return -EINVAL;
+
+	if (args->flags & RADEON_GEM_USERPTR_READONLY) {
+		/* readonly pages not tested on older hardware */
+		if (rdev->family < CHIP_R600)
+			return -EINVAL;
+
+	} else if (!(args->flags & RADEON_GEM_USERPTR_ANONONLY) ||
+		   !(args->flags & RADEON_GEM_USERPTR_REGISTER)) {
+
+		/* if we want to write to it we must require anonymous
+		   memory and install a MMU notifier */
+		return -EACCES;
+	}
+
+	down_read(&rdev->exclusive_lock);
+
+	/* create a gem object to contain this object in */
+	r = radeon_gem_object_create(rdev, args->size, 0,
+				     RADEON_GEM_DOMAIN_CPU, 0,
+				     false, &gobj);
+	if (r)
+		goto handle_lockup;
+
+	bo = gem_to_radeon_bo(gobj);
+	r = radeon_ttm_tt_set_userptr(bo->tbo.ttm, args->addr, args->flags);
+	if (r)
+		goto release_object;
+
+	if (args->flags & RADEON_GEM_USERPTR_REGISTER) {
+		r = radeon_mn_register(bo, args->addr);
+		if (r)
+			goto release_object;
+	}
+
+	if (args->flags & RADEON_GEM_USERPTR_VALIDATE) {
+		down_read(&current->mm->mmap_sem);
+		r = radeon_bo_reserve(bo, true);
+		if (r) {
+			up_read(&current->mm->mmap_sem);
+			goto release_object;
+		}
+
+		radeon_ttm_placement_from_domain(bo, RADEON_GEM_DOMAIN_GTT);
+		r = ttm_bo_validate(&bo->tbo, &bo->placement, true, false);
+		radeon_bo_unreserve(bo);
+		up_read(&current->mm->mmap_sem);
+		if (r)
+			goto release_object;
+	}
+
+	r = drm_gem_handle_create(filp, gobj, &handle);
+	/* drop reference from allocate - handle holds it now */
+	drm_gem_object_unreference_unlocked(gobj);
+	if (r)
+		goto handle_lockup;
+
+	args->handle = handle;
+	up_read(&rdev->exclusive_lock);
+	return 0;
+
+release_object:
+	drm_gem_object_unreference_unlocked(gobj);
+
+handle_lockup:
+	up_read(&rdev->exclusive_lock);
+	r = radeon_gem_handle_lockup(rdev, r);
+
+	return r;
+}
+
 int radeon_gem_set_domain_ioctl(struct drm_device *dev, void *data,
 				struct drm_file *filp)
 {
@@ -315,6 +406,10 @@ int radeon_mode_dumb_mmap(struct drm_file *filp,
 		return -ENOENT;
 	}
 	robj = gem_to_radeon_bo(gobj);
+	if (radeon_ttm_tt_has_userptr(robj->tbo.ttm)) {
+		drm_gem_object_unreference_unlocked(gobj);
+		return -EPERM;
+	}
 	*offset_p = radeon_bo_mmap_offset(robj);
 	drm_gem_object_unreference_unlocked(gobj);
 	return 0;
@@ -357,15 +452,22 @@ int radeon_gem_wait_idle_ioctl(struct drm_device *dev, void *data,
 	struct drm_radeon_gem_wait_idle *args = data;
 	struct drm_gem_object *gobj;
 	struct radeon_bo *robj;
-	int r;
+	int r = 0;
 	uint32_t cur_placement = 0;
+	long ret;
 
 	gobj = drm_gem_object_lookup(dev, filp, args->handle);
 	if (gobj == NULL) {
 		return -ENOENT;
 	}
 	robj = gem_to_radeon_bo(gobj);
-	r = radeon_bo_wait(robj, &cur_placement, false);
+
+	ret = reservation_object_wait_timeout_rcu(robj->tbo.resv, true, true, 30 * HZ);
+	if (ret == 0)
+		r = -EBUSY;
+	else if (ret < 0)
+		r = ret;
+
 	/* Flush HDP cache via MMIO if necessary */
 	if (rdev->asic->mmio_hdp_flush &&
 	    radeon_mem_type_to_domain(cur_placement) == RADEON_GEM_DOMAIN_VRAM)
@@ -416,6 +518,68 @@ out:
 	return r;
 }
 
+/**
+ * radeon_gem_va_update_vm -update the bo_va in its VM
+ *
+ * @rdev: radeon_device pointer
+ * @bo_va: bo_va to update
+ *
+ * Update the bo_va directly after setting it's address. Errors are not
+ * vital here, so they are not reported back to userspace.
+ */
+static void radeon_gem_va_update_vm(struct radeon_device *rdev,
+				    struct radeon_bo_va *bo_va)
+{
+	struct ttm_validate_buffer tv, *entry;
+	struct radeon_bo_list *vm_bos;
+	struct ww_acquire_ctx ticket;
+	struct list_head list;
+	unsigned domain;
+	int r;
+
+	INIT_LIST_HEAD(&list);
+
+	tv.bo = &bo_va->bo->tbo;
+	tv.shared = true;
+	list_add(&tv.head, &list);
+
+	vm_bos = radeon_vm_get_bos(rdev, bo_va->vm, &list);
+	if (!vm_bos)
+		return;
+
+	r = ttm_eu_reserve_buffers(&ticket, &list, true, NULL);
+	if (r)
+		goto error_free;
+
+	list_for_each_entry(entry, &list, head) {
+		domain = radeon_mem_type_to_domain(entry->bo->mem.mem_type);
+		/* if anything is swapped out don't swap it in here,
+		   just abort and wait for the next CS */
+		if (domain == RADEON_GEM_DOMAIN_CPU)
+			goto error_unreserve;
+	}
+
+	mutex_lock(&bo_va->vm->mutex);
+	r = radeon_vm_clear_freed(rdev, bo_va->vm);
+	if (r)
+		goto error_unlock;
+
+	if (bo_va->it.start)
+		r = radeon_vm_bo_update(rdev, bo_va, &bo_va->bo->tbo.mem);
+
+error_unlock:
+	mutex_unlock(&bo_va->vm->mutex);
+
+error_unreserve:
+	ttm_eu_backoff_reservation(&ticket, &list);
+
+error_free:
+	drm_free_large(vm_bos);
+
+	if (r)
+		DRM_ERROR("Couldn't update BO_VA (%d)\n", r);
+}
+
 int radeon_gem_va_ioctl(struct drm_device *dev, void *data,
 			  struct drm_file *filp)
 {
@@ -499,6 +663,7 @@ int radeon_gem_va_ioctl(struct drm_device *dev, void *data,
 		if (bo_va->it.start) {
 			args->operation = RADEON_VA_RESULT_VA_EXIST;
 			args->offset = bo_va->it.start * RADEON_GPU_PAGE_SIZE;
+			radeon_bo_unreserve(rbo);
 			goto out;
 		}
 		r = radeon_vm_bo_set_addr(rdev, bo_va, args->offset, args->flags);
@@ -509,12 +674,13 @@ int radeon_gem_va_ioctl(struct drm_device *dev, void *data,
 	default:
 		break;
 	}
+	if (!r)
+		radeon_gem_va_update_vm(rdev, bo_va);
 	args->operation = RADEON_VA_RESULT_OK;
 	if (r) {
 		args->operation = RADEON_VA_RESULT_ERROR;
 	}
 out:
-	radeon_bo_unreserve(rbo);
 	drm_gem_object_unreference_unlocked(gobj);
 	return r;
 }
@@ -532,6 +698,11 @@ int radeon_gem_op_ioctl(struct drm_device *dev, void *data,
 		return -ENOENT;
 	}
 	robj = gem_to_radeon_bo(gobj);
+
+	r = -EPERM;
+	if (radeon_ttm_tt_has_userptr(robj->tbo.ttm))
+		goto out;
+
 	r = radeon_bo_reserve(robj, false);
 	if (unlikely(r))
 		goto out;
diff --git a/drivers/gpu/drm/radeon/radeon_ib.c b/drivers/gpu/drm/radeon/radeon_ib.c
index 5bf2c0a05827..c39ce1f05703 100644
--- a/drivers/gpu/drm/radeon/radeon_ib.c
+++ b/drivers/gpu/drm/radeon/radeon_ib.c
@@ -64,10 +64,7 @@ int radeon_ib_get(struct radeon_device *rdev, int ring,
 		return r;
 	}
 
-	r = radeon_semaphore_create(rdev, &ib->semaphore);
-	if (r) {
-		return r;
-	}
+	radeon_sync_create(&ib->sync);
 
 	ib->ring = ring;
 	ib->fence = NULL;
@@ -96,7 +93,7 @@ int radeon_ib_get(struct radeon_device *rdev, int ring,
  */
 void radeon_ib_free(struct radeon_device *rdev, struct radeon_ib *ib)
 {
-	radeon_semaphore_free(rdev, &ib->semaphore, ib->fence);
+	radeon_sync_free(rdev, &ib->sync, ib->fence);
 	radeon_sa_bo_free(rdev, &ib->sa_bo, ib->fence);
 	radeon_fence_unref(&ib->fence);
 }
@@ -145,11 +142,11 @@ int radeon_ib_schedule(struct radeon_device *rdev, struct radeon_ib *ib,
 	if (ib->vm) {
 		struct radeon_fence *vm_id_fence;
 		vm_id_fence = radeon_vm_grab_id(rdev, ib->vm, ib->ring);
-        	radeon_semaphore_sync_to(ib->semaphore, vm_id_fence);
+		radeon_sync_fence(&ib->sync, vm_id_fence);
 	}
 
 	/* sync with other rings */
-	r = radeon_semaphore_sync_rings(rdev, ib->semaphore, ib->ring);
+	r = radeon_sync_rings(rdev, &ib->sync, ib->ring);
 	if (r) {
 		dev_err(rdev->dev, "failed to sync rings (%d)\n", r);
 		radeon_ring_unlock_undo(rdev, ring);
@@ -157,11 +154,12 @@ int radeon_ib_schedule(struct radeon_device *rdev, struct radeon_ib *ib,
 	}
 
 	if (ib->vm)
-		radeon_vm_flush(rdev, ib->vm, ib->ring);
+		radeon_vm_flush(rdev, ib->vm, ib->ring,
+				ib->sync.last_vm_update);
 
 	if (const_ib) {
 		radeon_ring_ib_execute(rdev, const_ib->ring, const_ib);
-		radeon_semaphore_free(rdev, &const_ib->semaphore, NULL);
+		radeon_sync_free(rdev, &const_ib->sync, NULL);
 	}
 	radeon_ring_ib_execute(rdev, ib->ring, ib);
 	r = radeon_fence_emit(rdev, &ib->fence, ib->ring);
@@ -269,6 +267,7 @@ int radeon_ib_ring_tests(struct radeon_device *rdev)
 
 		r = radeon_ib_test(rdev, i, ring);
 		if (r) {
+			radeon_fence_driver_force_completion(rdev, i);
 			ring->ready = false;
 			rdev->needs_reset = false;
 
diff --git a/drivers/gpu/drm/radeon/radeon_irq_kms.c b/drivers/gpu/drm/radeon/radeon_irq_kms.c
index 16807afab362..00fc59762e0d 100644
--- a/drivers/gpu/drm/radeon/radeon_irq_kms.c
+++ b/drivers/gpu/drm/radeon/radeon_irq_kms.c
@@ -88,23 +88,6 @@ static void radeon_hotplug_work_func(struct work_struct *work)
 }
 
 /**
- * radeon_irq_reset_work_func - execute gpu reset
- *
- * @work: work struct
- *
- * Execute scheduled gpu reset (cayman+).
- * This function is called when the irq handler
- * thinks we need a gpu reset.
- */
-static void radeon_irq_reset_work_func(struct work_struct *work)
-{
-	struct radeon_device *rdev = container_of(work, struct radeon_device,
-						  reset_work);
-
-	radeon_gpu_reset(rdev);
-}
-
-/**
  * radeon_driver_irq_preinstall_kms - drm irq preinstall callback
  *
  * @dev: drm dev pointer
@@ -202,6 +185,16 @@ static bool radeon_msi_ok(struct radeon_device *rdev)
 	if (rdev->flags & RADEON_IS_AGP)
 		return false;
 
+	/*
+	 * Older chips have a HW limitation, they can only generate 40 bits
+	 * of address for "64-bit" MSIs which breaks on some platforms, notably
+	 * IBM POWER servers, so we limit them
+	 */
+	if (rdev->family < CHIP_BONAIRE) {
+		dev_info(rdev->dev, "radeon: MSI limited to 32-bit\n");
+		rdev->pdev->no_64bit_msi = 1;
+	}
+
 	/* force MSI on */
 	if (radeon_msi == 1)
 		return true;
@@ -284,7 +277,6 @@ int radeon_irq_kms_init(struct radeon_device *rdev)
 
 	INIT_WORK(&rdev->hotplug_work, radeon_hotplug_work_func);
 	INIT_WORK(&rdev->audio_work, r600_audio_update_hdmi);
-	INIT_WORK(&rdev->reset_work, radeon_irq_reset_work_func);
 
 	rdev->irq.installed = true;
 	r = drm_irq_install(rdev->ddev, rdev->ddev->pdev->irq);
@@ -342,6 +334,21 @@ void radeon_irq_kms_sw_irq_get(struct radeon_device *rdev, int ring)
 }
 
 /**
+ * radeon_irq_kms_sw_irq_get_delayed - enable software interrupt
+ *
+ * @rdev: radeon device pointer
+ * @ring: ring whose interrupt you want to enable
+ *
+ * Enables the software interrupt for a specific ring (all asics).
+ * The software interrupt is generally used to signal a fence on
+ * a particular ring.
+ */
+bool radeon_irq_kms_sw_irq_get_delayed(struct radeon_device *rdev, int ring)
+{
+	return atomic_inc_return(&rdev->irq.ring_int[ring]) == 1;
+}
+
+/**
  * radeon_irq_kms_sw_irq_put - disable software interrupt
  *
  * @rdev: radeon device pointer
diff --git a/drivers/gpu/drm/radeon/radeon_kfd.c b/drivers/gpu/drm/radeon/radeon_kfd.c
new file mode 100644
index 000000000000..8bf87f1203cc
--- /dev/null
+++ b/drivers/gpu/drm/radeon/radeon_kfd.c
@@ -0,0 +1,633 @@
+/*
+ * Copyright 2014 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ */
+
+#include <linux/module.h>
+#include <linux/fdtable.h>
+#include <linux/uaccess.h>
+#include <drm/drmP.h>
+#include "radeon.h"
+#include "cikd.h"
+#include "cik_reg.h"
+#include "radeon_kfd.h"
+#include "radeon_ucode.h"
+#include <linux/firmware.h>
+
+#define CIK_PIPE_PER_MEC	(4)
+
+struct kgd_mem {
+	struct radeon_sa_bo *sa_bo;
+	uint64_t gpu_addr;
+	void *ptr;
+};
+
+static int init_sa_manager(struct kgd_dev *kgd, unsigned int size);
+static void fini_sa_manager(struct kgd_dev *kgd);
+
+static int allocate_mem(struct kgd_dev *kgd, size_t size, size_t alignment,
+		enum kgd_memory_pool pool, struct kgd_mem **mem);
+
+static void free_mem(struct kgd_dev *kgd, struct kgd_mem *mem);
+
+static uint64_t get_vmem_size(struct kgd_dev *kgd);
+static uint64_t get_gpu_clock_counter(struct kgd_dev *kgd);
+
+static uint32_t get_max_engine_clock_in_mhz(struct kgd_dev *kgd);
+static uint16_t get_fw_version(struct kgd_dev *kgd, enum kgd_engine_type type);
+
+/*
+ * Register access functions
+ */
+
+static void kgd_program_sh_mem_settings(struct kgd_dev *kgd, uint32_t vmid,
+		uint32_t sh_mem_config,	uint32_t sh_mem_ape1_base,
+		uint32_t sh_mem_ape1_limit, uint32_t sh_mem_bases);
+
+static int kgd_set_pasid_vmid_mapping(struct kgd_dev *kgd, unsigned int pasid,
+					unsigned int vmid);
+
+static int kgd_init_memory(struct kgd_dev *kgd);
+
+static int kgd_init_pipeline(struct kgd_dev *kgd, uint32_t pipe_id,
+				uint32_t hpd_size, uint64_t hpd_gpu_addr);
+
+static int kgd_hqd_load(struct kgd_dev *kgd, void *mqd, uint32_t pipe_id,
+			uint32_t queue_id, uint32_t __user *wptr);
+
+static bool kgd_hqd_is_occupied(struct kgd_dev *kgd, uint64_t queue_address,
+				uint32_t pipe_id, uint32_t queue_id);
+
+static int kgd_hqd_destroy(struct kgd_dev *kgd, uint32_t reset_type,
+				unsigned int timeout, uint32_t pipe_id,
+				uint32_t queue_id);
+
+static const struct kfd2kgd_calls kfd2kgd = {
+	.init_sa_manager = init_sa_manager,
+	.fini_sa_manager = fini_sa_manager,
+	.allocate_mem = allocate_mem,
+	.free_mem = free_mem,
+	.get_vmem_size = get_vmem_size,
+	.get_gpu_clock_counter = get_gpu_clock_counter,
+	.get_max_engine_clock_in_mhz = get_max_engine_clock_in_mhz,
+	.program_sh_mem_settings = kgd_program_sh_mem_settings,
+	.set_pasid_vmid_mapping = kgd_set_pasid_vmid_mapping,
+	.init_memory = kgd_init_memory,
+	.init_pipeline = kgd_init_pipeline,
+	.hqd_load = kgd_hqd_load,
+	.hqd_is_occupied = kgd_hqd_is_occupied,
+	.hqd_destroy = kgd_hqd_destroy,
+	.get_fw_version = get_fw_version
+};
+
+static const struct kgd2kfd_calls *kgd2kfd;
+
+bool radeon_kfd_init(void)
+{
+#if defined(CONFIG_HSA_AMD_MODULE)
+	bool (*kgd2kfd_init_p)(unsigned, const struct kfd2kgd_calls*,
+				const struct kgd2kfd_calls**);
+
+	kgd2kfd_init_p = symbol_request(kgd2kfd_init);
+
+	if (kgd2kfd_init_p == NULL)
+		return false;
+
+	if (!kgd2kfd_init_p(KFD_INTERFACE_VERSION, &kfd2kgd, &kgd2kfd)) {
+		symbol_put(kgd2kfd_init);
+		kgd2kfd = NULL;
+
+		return false;
+	}
+
+	return true;
+#elif defined(CONFIG_HSA_AMD)
+	if (!kgd2kfd_init(KFD_INTERFACE_VERSION, &kfd2kgd, &kgd2kfd)) {
+		kgd2kfd = NULL;
+
+		return false;
+	}
+
+	return true;
+#else
+	return false;
+#endif
+}
+
+void radeon_kfd_fini(void)
+{
+	if (kgd2kfd) {
+		kgd2kfd->exit();
+		symbol_put(kgd2kfd_init);
+	}
+}
+
+void radeon_kfd_device_probe(struct radeon_device *rdev)
+{
+	if (kgd2kfd)
+		rdev->kfd = kgd2kfd->probe((struct kgd_dev *)rdev, rdev->pdev);
+}
+
+void radeon_kfd_device_init(struct radeon_device *rdev)
+{
+	if (rdev->kfd) {
+		struct kgd2kfd_shared_resources gpu_resources = {
+			.compute_vmid_bitmap = 0xFF00,
+
+			.first_compute_pipe = 1,
+			.compute_pipe_count = 8 - 1,
+		};
+
+		radeon_doorbell_get_kfd_info(rdev,
+				&gpu_resources.doorbell_physical_address,
+				&gpu_resources.doorbell_aperture_size,
+				&gpu_resources.doorbell_start_offset);
+
+		kgd2kfd->device_init(rdev->kfd, &gpu_resources);
+	}
+}
+
+void radeon_kfd_device_fini(struct radeon_device *rdev)
+{
+	if (rdev->kfd) {
+		kgd2kfd->device_exit(rdev->kfd);
+		rdev->kfd = NULL;
+	}
+}
+
+void radeon_kfd_interrupt(struct radeon_device *rdev, const void *ih_ring_entry)
+{
+	if (rdev->kfd)
+		kgd2kfd->interrupt(rdev->kfd, ih_ring_entry);
+}
+
+void radeon_kfd_suspend(struct radeon_device *rdev)
+{
+	if (rdev->kfd)
+		kgd2kfd->suspend(rdev->kfd);
+}
+
+int radeon_kfd_resume(struct radeon_device *rdev)
+{
+	int r = 0;
+
+	if (rdev->kfd)
+		r = kgd2kfd->resume(rdev->kfd);
+
+	return r;
+}
+
+static u32 pool_to_domain(enum kgd_memory_pool p)
+{
+	switch (p) {
+	case KGD_POOL_FRAMEBUFFER: return RADEON_GEM_DOMAIN_VRAM;
+	default: return RADEON_GEM_DOMAIN_GTT;
+	}
+}
+
+static int init_sa_manager(struct kgd_dev *kgd, unsigned int size)
+{
+	struct radeon_device *rdev = (struct radeon_device *)kgd;
+	int r;
+
+	BUG_ON(kgd == NULL);
+
+	r = radeon_sa_bo_manager_init(rdev, &rdev->kfd_bo,
+				      size,
+				      RADEON_GPU_PAGE_SIZE,
+				      RADEON_GEM_DOMAIN_GTT,
+				      RADEON_GEM_GTT_WC);
+
+	if (r)
+		return r;
+
+	r = radeon_sa_bo_manager_start(rdev, &rdev->kfd_bo);
+	if (r)
+		radeon_sa_bo_manager_fini(rdev, &rdev->kfd_bo);
+
+	return r;
+}
+
+static void fini_sa_manager(struct kgd_dev *kgd)
+{
+	struct radeon_device *rdev = (struct radeon_device *)kgd;
+
+	BUG_ON(kgd == NULL);
+
+	radeon_sa_bo_manager_suspend(rdev, &rdev->kfd_bo);
+	radeon_sa_bo_manager_fini(rdev, &rdev->kfd_bo);
+}
+
+static int allocate_mem(struct kgd_dev *kgd, size_t size, size_t alignment,
+		enum kgd_memory_pool pool, struct kgd_mem **mem)
+{
+	struct radeon_device *rdev = (struct radeon_device *)kgd;
+	u32 domain;
+	int r;
+
+	BUG_ON(kgd == NULL);
+
+	domain = pool_to_domain(pool);
+	if (domain != RADEON_GEM_DOMAIN_GTT) {
+		dev_err(rdev->dev,
+			"Only allowed to allocate gart memory for kfd\n");
+		return -EINVAL;
+	}
+
+	*mem = kmalloc(sizeof(struct kgd_mem), GFP_KERNEL);
+	if ((*mem) == NULL)
+		return -ENOMEM;
+
+	r = radeon_sa_bo_new(rdev, &rdev->kfd_bo, &(*mem)->sa_bo, size,
+				alignment);
+	if (r) {
+		dev_err(rdev->dev, "failed to get memory for kfd (%d)\n", r);
+		return r;
+	}
+
+	(*mem)->ptr = radeon_sa_bo_cpu_addr((*mem)->sa_bo);
+	(*mem)->gpu_addr = radeon_sa_bo_gpu_addr((*mem)->sa_bo);
+
+	return 0;
+}
+
+static void free_mem(struct kgd_dev *kgd, struct kgd_mem *mem)
+{
+	struct radeon_device *rdev = (struct radeon_device *)kgd;
+
+	BUG_ON(kgd == NULL);
+
+	radeon_sa_bo_free(rdev, &mem->sa_bo, NULL);
+	kfree(mem);
+}
+
+static uint64_t get_vmem_size(struct kgd_dev *kgd)
+{
+	struct radeon_device *rdev = (struct radeon_device *)kgd;
+
+	BUG_ON(kgd == NULL);
+
+	return rdev->mc.real_vram_size;
+}
+
+static uint64_t get_gpu_clock_counter(struct kgd_dev *kgd)
+{
+	struct radeon_device *rdev = (struct radeon_device *)kgd;
+
+	return rdev->asic->get_gpu_clock_counter(rdev);
+}
+
+static uint32_t get_max_engine_clock_in_mhz(struct kgd_dev *kgd)
+{
+	struct radeon_device *rdev = (struct radeon_device *)kgd;
+
+	/* The sclk is in quantas of 10kHz */
+	return rdev->pm.dpm.dyn_state.max_clock_voltage_on_ac.sclk / 100;
+}
+
+static inline struct radeon_device *get_radeon_device(struct kgd_dev *kgd)
+{
+	return (struct radeon_device *)kgd;
+}
+
+static void write_register(struct kgd_dev *kgd, uint32_t offset, uint32_t value)
+{
+	struct radeon_device *rdev = get_radeon_device(kgd);
+
+	writel(value, (void __iomem *)(rdev->rmmio + offset));
+}
+
+static uint32_t read_register(struct kgd_dev *kgd, uint32_t offset)
+{
+	struct radeon_device *rdev = get_radeon_device(kgd);
+
+	return readl((void __iomem *)(rdev->rmmio + offset));
+}
+
+static void lock_srbm(struct kgd_dev *kgd, uint32_t mec, uint32_t pipe,
+			uint32_t queue, uint32_t vmid)
+{
+	struct radeon_device *rdev = get_radeon_device(kgd);
+	uint32_t value = PIPEID(pipe) | MEID(mec) | VMID(vmid) | QUEUEID(queue);
+
+	mutex_lock(&rdev->srbm_mutex);
+	write_register(kgd, SRBM_GFX_CNTL, value);
+}
+
+static void unlock_srbm(struct kgd_dev *kgd)
+{
+	struct radeon_device *rdev = get_radeon_device(kgd);
+
+	write_register(kgd, SRBM_GFX_CNTL, 0);
+	mutex_unlock(&rdev->srbm_mutex);
+}
+
+static void acquire_queue(struct kgd_dev *kgd, uint32_t pipe_id,
+				uint32_t queue_id)
+{
+	uint32_t mec = (++pipe_id / CIK_PIPE_PER_MEC) + 1;
+	uint32_t pipe = (pipe_id % CIK_PIPE_PER_MEC);
+
+	lock_srbm(kgd, mec, pipe, queue_id, 0);
+}
+
+static void release_queue(struct kgd_dev *kgd)
+{
+	unlock_srbm(kgd);
+}
+
+static void kgd_program_sh_mem_settings(struct kgd_dev *kgd, uint32_t vmid,
+					uint32_t sh_mem_config,
+					uint32_t sh_mem_ape1_base,
+					uint32_t sh_mem_ape1_limit,
+					uint32_t sh_mem_bases)
+{
+	lock_srbm(kgd, 0, 0, 0, vmid);
+
+	write_register(kgd, SH_MEM_CONFIG, sh_mem_config);
+	write_register(kgd, SH_MEM_APE1_BASE, sh_mem_ape1_base);
+	write_register(kgd, SH_MEM_APE1_LIMIT, sh_mem_ape1_limit);
+	write_register(kgd, SH_MEM_BASES, sh_mem_bases);
+
+	unlock_srbm(kgd);
+}
+
+static int kgd_set_pasid_vmid_mapping(struct kgd_dev *kgd, unsigned int pasid,
+					unsigned int vmid)
+{
+	/*
+	 * We have to assume that there is no outstanding mapping.
+	 * The ATC_VMID_PASID_MAPPING_UPDATE_STATUS bit could be 0
+	 * because a mapping is in progress or because a mapping finished and
+	 * the SW cleared it.
+	 * So the protocol is to always wait & clear.
+	 */
+	uint32_t pasid_mapping = (pasid == 0) ? 0 :
+				(uint32_t)pasid | ATC_VMID_PASID_MAPPING_VALID;
+
+	write_register(kgd, ATC_VMID0_PASID_MAPPING + vmid*sizeof(uint32_t),
+			pasid_mapping);
+
+	while (!(read_register(kgd, ATC_VMID_PASID_MAPPING_UPDATE_STATUS) &
+								(1U << vmid)))
+		cpu_relax();
+	write_register(kgd, ATC_VMID_PASID_MAPPING_UPDATE_STATUS, 1U << vmid);
+
+	/* Mapping vmid to pasid also for IH block */
+	write_register(kgd, IH_VMID_0_LUT + vmid * sizeof(uint32_t),
+			pasid_mapping);
+
+	return 0;
+}
+
+static int kgd_init_memory(struct kgd_dev *kgd)
+{
+	/*
+	 * Configure apertures:
+	 * LDS:         0x60000000'00000000 - 0x60000001'00000000 (4GB)
+	 * Scratch:     0x60000001'00000000 - 0x60000002'00000000 (4GB)
+	 * GPUVM:       0x60010000'00000000 - 0x60020000'00000000 (1TB)
+	 */
+	int i;
+	uint32_t sh_mem_bases = PRIVATE_BASE(0x6000) | SHARED_BASE(0x6000);
+
+	for (i = 8; i < 16; i++) {
+		uint32_t sh_mem_config;
+
+		lock_srbm(kgd, 0, 0, 0, i);
+
+		sh_mem_config = ALIGNMENT_MODE(SH_MEM_ALIGNMENT_MODE_UNALIGNED);
+		sh_mem_config |= DEFAULT_MTYPE(MTYPE_NONCACHED);
+
+		write_register(kgd, SH_MEM_CONFIG, sh_mem_config);
+
+		write_register(kgd, SH_MEM_BASES, sh_mem_bases);
+
+		/* Scratch aperture is not supported for now. */
+		write_register(kgd, SH_STATIC_MEM_CONFIG, 0);
+
+		/* APE1 disabled for now. */
+		write_register(kgd, SH_MEM_APE1_BASE, 1);
+		write_register(kgd, SH_MEM_APE1_LIMIT, 0);
+
+		unlock_srbm(kgd);
+	}
+
+	return 0;
+}
+
+static int kgd_init_pipeline(struct kgd_dev *kgd, uint32_t pipe_id,
+				uint32_t hpd_size, uint64_t hpd_gpu_addr)
+{
+	uint32_t mec = (++pipe_id / CIK_PIPE_PER_MEC) + 1;
+	uint32_t pipe = (pipe_id % CIK_PIPE_PER_MEC);
+
+	lock_srbm(kgd, mec, pipe, 0, 0);
+	write_register(kgd, CP_HPD_EOP_BASE_ADDR,
+			lower_32_bits(hpd_gpu_addr >> 8));
+	write_register(kgd, CP_HPD_EOP_BASE_ADDR_HI,
+			upper_32_bits(hpd_gpu_addr >> 8));
+	write_register(kgd, CP_HPD_EOP_VMID, 0);
+	write_register(kgd, CP_HPD_EOP_CONTROL, hpd_size);
+	unlock_srbm(kgd);
+
+	return 0;
+}
+
+static inline struct cik_mqd *get_mqd(void *mqd)
+{
+	return (struct cik_mqd *)mqd;
+}
+
+static int kgd_hqd_load(struct kgd_dev *kgd, void *mqd, uint32_t pipe_id,
+			uint32_t queue_id, uint32_t __user *wptr)
+{
+	uint32_t wptr_shadow, is_wptr_shadow_valid;
+	struct cik_mqd *m;
+
+	m = get_mqd(mqd);
+
+	is_wptr_shadow_valid = !get_user(wptr_shadow, wptr);
+
+	acquire_queue(kgd, pipe_id, queue_id);
+	write_register(kgd, CP_MQD_BASE_ADDR, m->cp_mqd_base_addr_lo);
+	write_register(kgd, CP_MQD_BASE_ADDR_HI, m->cp_mqd_base_addr_hi);
+	write_register(kgd, CP_MQD_CONTROL, m->cp_mqd_control);
+
+	write_register(kgd, CP_HQD_PQ_BASE, m->cp_hqd_pq_base_lo);
+	write_register(kgd, CP_HQD_PQ_BASE_HI, m->cp_hqd_pq_base_hi);
+	write_register(kgd, CP_HQD_PQ_CONTROL, m->cp_hqd_pq_control);
+
+	write_register(kgd, CP_HQD_IB_CONTROL, m->cp_hqd_ib_control);
+	write_register(kgd, CP_HQD_IB_BASE_ADDR, m->cp_hqd_ib_base_addr_lo);
+	write_register(kgd, CP_HQD_IB_BASE_ADDR_HI, m->cp_hqd_ib_base_addr_hi);
+
+	write_register(kgd, CP_HQD_IB_RPTR, m->cp_hqd_ib_rptr);
+
+	write_register(kgd, CP_HQD_PERSISTENT_STATE,
+			m->cp_hqd_persistent_state);
+	write_register(kgd, CP_HQD_SEMA_CMD, m->cp_hqd_sema_cmd);
+	write_register(kgd, CP_HQD_MSG_TYPE, m->cp_hqd_msg_type);
+
+	write_register(kgd, CP_HQD_ATOMIC0_PREOP_LO,
+			m->cp_hqd_atomic0_preop_lo);
+
+	write_register(kgd, CP_HQD_ATOMIC0_PREOP_HI,
+			m->cp_hqd_atomic0_preop_hi);
+
+	write_register(kgd, CP_HQD_ATOMIC1_PREOP_LO,
+			m->cp_hqd_atomic1_preop_lo);
+
+	write_register(kgd, CP_HQD_ATOMIC1_PREOP_HI,
+			m->cp_hqd_atomic1_preop_hi);
+
+	write_register(kgd, CP_HQD_PQ_RPTR_REPORT_ADDR,
+			m->cp_hqd_pq_rptr_report_addr_lo);
+
+	write_register(kgd, CP_HQD_PQ_RPTR_REPORT_ADDR_HI,
+			m->cp_hqd_pq_rptr_report_addr_hi);
+
+	write_register(kgd, CP_HQD_PQ_RPTR, m->cp_hqd_pq_rptr);
+
+	write_register(kgd, CP_HQD_PQ_WPTR_POLL_ADDR,
+			m->cp_hqd_pq_wptr_poll_addr_lo);
+
+	write_register(kgd, CP_HQD_PQ_WPTR_POLL_ADDR_HI,
+			m->cp_hqd_pq_wptr_poll_addr_hi);
+
+	write_register(kgd, CP_HQD_PQ_DOORBELL_CONTROL,
+			m->cp_hqd_pq_doorbell_control);
+
+	write_register(kgd, CP_HQD_VMID, m->cp_hqd_vmid);
+
+	write_register(kgd, CP_HQD_QUANTUM, m->cp_hqd_quantum);
+
+	write_register(kgd, CP_HQD_PIPE_PRIORITY, m->cp_hqd_pipe_priority);
+	write_register(kgd, CP_HQD_QUEUE_PRIORITY, m->cp_hqd_queue_priority);
+
+	write_register(kgd, CP_HQD_IQ_RPTR, m->cp_hqd_iq_rptr);
+
+	if (is_wptr_shadow_valid)
+		write_register(kgd, CP_HQD_PQ_WPTR, wptr_shadow);
+
+	write_register(kgd, CP_HQD_ACTIVE, m->cp_hqd_active);
+	release_queue(kgd);
+
+	return 0;
+}
+
+static bool kgd_hqd_is_occupied(struct kgd_dev *kgd, uint64_t queue_address,
+				uint32_t pipe_id, uint32_t queue_id)
+{
+	uint32_t act;
+	bool retval = false;
+	uint32_t low, high;
+
+	acquire_queue(kgd, pipe_id, queue_id);
+	act = read_register(kgd, CP_HQD_ACTIVE);
+	if (act) {
+		low = lower_32_bits(queue_address >> 8);
+		high = upper_32_bits(queue_address >> 8);
+
+		if (low == read_register(kgd, CP_HQD_PQ_BASE) &&
+				high == read_register(kgd, CP_HQD_PQ_BASE_HI))
+			retval = true;
+	}
+	release_queue(kgd);
+	return retval;
+}
+
+static int kgd_hqd_destroy(struct kgd_dev *kgd, uint32_t reset_type,
+				unsigned int timeout, uint32_t pipe_id,
+				uint32_t queue_id)
+{
+	uint32_t temp;
+
+	acquire_queue(kgd, pipe_id, queue_id);
+	write_register(kgd, CP_HQD_PQ_DOORBELL_CONTROL, 0);
+
+	write_register(kgd, CP_HQD_DEQUEUE_REQUEST, reset_type);
+
+	while (true) {
+		temp = read_register(kgd, CP_HQD_ACTIVE);
+		if (temp & 0x1)
+			break;
+		if (timeout == 0) {
+			pr_err("kfd: cp queue preemption time out (%dms)\n",
+				temp);
+			release_queue(kgd);
+			return -ETIME;
+		}
+		msleep(20);
+		timeout -= 20;
+	}
+
+	release_queue(kgd);
+	return 0;
+}
+
+static uint16_t get_fw_version(struct kgd_dev *kgd, enum kgd_engine_type type)
+{
+	struct radeon_device *rdev = (struct radeon_device *) kgd;
+	const union radeon_firmware_header *hdr;
+
+	BUG_ON(kgd == NULL || rdev->mec_fw == NULL);
+
+	switch (type) {
+	case KGD_ENGINE_PFP:
+		hdr = (const union radeon_firmware_header *) rdev->pfp_fw->data;
+		break;
+
+	case KGD_ENGINE_ME:
+		hdr = (const union radeon_firmware_header *) rdev->me_fw->data;
+		break;
+
+	case KGD_ENGINE_CE:
+		hdr = (const union radeon_firmware_header *) rdev->ce_fw->data;
+		break;
+
+	case KGD_ENGINE_MEC1:
+		hdr = (const union radeon_firmware_header *) rdev->mec_fw->data;
+		break;
+
+	case KGD_ENGINE_MEC2:
+		hdr = (const union radeon_firmware_header *)
+							rdev->mec2_fw->data;
+		break;
+
+	case KGD_ENGINE_RLC:
+		hdr = (const union radeon_firmware_header *) rdev->rlc_fw->data;
+		break;
+
+	case KGD_ENGINE_SDMA:
+		hdr = (const union radeon_firmware_header *)
+							rdev->sdma_fw->data;
+		break;
+
+	default:
+		return 0;
+	}
+
+	if (hdr == NULL)
+		return 0;
+
+	/* Only 12 bit in use*/
+	return hdr->common.ucode_version;
+}
diff --git a/drivers/gpu/drm/radeon/radeon_kfd.h b/drivers/gpu/drm/radeon/radeon_kfd.h
new file mode 100644
index 000000000000..f90e161ca507
--- /dev/null
+++ b/drivers/gpu/drm/radeon/radeon_kfd.h
@@ -0,0 +1,47 @@
+/*
+ * Copyright 2014 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ */
+
+/*
+ * radeon_kfd.h defines the private interface between the
+ * AMD kernel graphics drivers and the AMD KFD.
+ */
+
+#ifndef RADEON_KFD_H_INCLUDED
+#define RADEON_KFD_H_INCLUDED
+
+#include <linux/types.h>
+#include "../amd/include/kgd_kfd_interface.h"
+
+struct radeon_device;
+
+bool radeon_kfd_init(void);
+void radeon_kfd_fini(void);
+
+void radeon_kfd_suspend(struct radeon_device *rdev);
+int radeon_kfd_resume(struct radeon_device *rdev);
+void radeon_kfd_interrupt(struct radeon_device *rdev,
+			const void *ih_ring_entry);
+void radeon_kfd_device_probe(struct radeon_device *rdev);
+void radeon_kfd_device_init(struct radeon_device *rdev);
+void radeon_kfd_device_fini(struct radeon_device *rdev);
+
+#endif /* RADEON_KFD_H_INCLUDED */
diff --git a/drivers/gpu/drm/radeon/radeon_kms.c b/drivers/gpu/drm/radeon/radeon_kms.c
index eb7164d07985..3cf9c1fa6475 100644
--- a/drivers/gpu/drm/radeon/radeon_kms.c
+++ b/drivers/gpu/drm/radeon/radeon_kms.c
@@ -34,6 +34,8 @@
 #include <linux/slab.h>
 #include <linux/pm_runtime.h>
 
+#include "radeon_kfd.h"
+
 #if defined(CONFIG_VGA_SWITCHEROO)
 bool radeon_has_atpx(void);
 #else
@@ -63,6 +65,8 @@ int radeon_driver_unload_kms(struct drm_device *dev)
 
 	pm_runtime_get_sync(dev->dev);
 
+	radeon_kfd_device_fini(rdev);
+
 	radeon_acpi_fini(rdev);
 	
 	radeon_modeset_fini(rdev);
@@ -142,6 +146,9 @@ int radeon_driver_load_kms(struct drm_device *dev, unsigned long flags)
 				"Error during ACPI methods call\n");
 	}
 
+	radeon_kfd_device_probe(rdev);
+	radeon_kfd_device_init(rdev);
+
 	if (radeon_is_px(dev)) {
 		pm_runtime_use_autosuspend(dev->dev);
 		pm_runtime_set_autosuspend_delay(dev->dev, 5000);
@@ -621,8 +628,6 @@ int radeon_driver_open_kms(struct drm_device *dev, struct drm_file *file_priv)
 						  RADEON_VA_IB_OFFSET,
 						  RADEON_VM_PAGE_READABLE |
 						  RADEON_VM_PAGE_SNOOPED);
-
-			radeon_bo_unreserve(rdev->ring_tmp_bo.bo);
 			if (r) {
 				radeon_vm_fini(rdev, vm);
 				kfree(fpriv);
@@ -795,6 +800,8 @@ int radeon_get_vblank_timestamp_kms(struct drm_device *dev, int crtc,
 
 	/* Get associated drm_crtc: */
 	drmcrtc = &rdev->mode_info.crtcs[crtc]->base;
+	if (!drmcrtc)
+		return -EINVAL;
 
 	/* Helper routine in DRM core does all the work: */
 	return drm_calc_vbltimestamp_from_scanoutpos(dev, crtc, max_error,
@@ -885,5 +892,6 @@ const struct drm_ioctl_desc radeon_ioctls_kms[] = {
 	DRM_IOCTL_DEF_DRV(RADEON_GEM_BUSY, radeon_gem_busy_ioctl, DRM_AUTH|DRM_UNLOCKED|DRM_RENDER_ALLOW),
 	DRM_IOCTL_DEF_DRV(RADEON_GEM_VA, radeon_gem_va_ioctl, DRM_AUTH|DRM_UNLOCKED|DRM_RENDER_ALLOW),
 	DRM_IOCTL_DEF_DRV(RADEON_GEM_OP, radeon_gem_op_ioctl, DRM_AUTH|DRM_UNLOCKED|DRM_RENDER_ALLOW),
+	DRM_IOCTL_DEF_DRV(RADEON_GEM_USERPTR, radeon_gem_userptr_ioctl, DRM_AUTH|DRM_UNLOCKED|DRM_RENDER_ALLOW),
 };
 int radeon_max_kms_ioctl = ARRAY_SIZE(radeon_ioctls_kms);
diff --git a/drivers/gpu/drm/radeon/radeon_legacy_crtc.c b/drivers/gpu/drm/radeon/radeon_legacy_crtc.c
index cafb1ccf2ec3..678b4386540d 100644
--- a/drivers/gpu/drm/radeon/radeon_legacy_crtc.c
+++ b/drivers/gpu/drm/radeon/radeon_legacy_crtc.c
@@ -1054,6 +1054,7 @@ static int radeon_crtc_mode_set(struct drm_crtc *crtc,
 			DRM_ERROR("Mode need scaling but only first crtc can do that.\n");
 		}
 	}
+	radeon_cursor_reset(crtc);
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/radeon/radeon_mn.c b/drivers/gpu/drm/radeon/radeon_mn.c
new file mode 100644
index 000000000000..a69bd441dd2d
--- /dev/null
+++ b/drivers/gpu/drm/radeon/radeon_mn.c
@@ -0,0 +1,274 @@
+/*
+ * Copyright 2014 Advanced Micro Devices, Inc.
+ * All Rights Reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the
+ * "Software"), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sub license, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDERS, AUTHORS AND/OR ITS SUPPLIERS BE LIABLE FOR ANY CLAIM,
+ * DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR
+ * OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE
+ * USE OR OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * The above copyright notice and this permission notice (including the
+ * next paragraph) shall be included in all copies or substantial portions
+ * of the Software.
+ *
+ */
+/*
+ * Authors:
+ *    Christian König <christian.koenig@amd.com>
+ */
+
+#include <linux/firmware.h>
+#include <linux/module.h>
+#include <linux/mmu_notifier.h>
+#include <drm/drmP.h>
+#include <drm/drm.h>
+
+#include "radeon.h"
+
+struct radeon_mn {
+	/* constant after initialisation */
+	struct radeon_device	*rdev;
+	struct mm_struct	*mm;
+	struct mmu_notifier	mn;
+
+	/* only used on destruction */
+	struct work_struct	work;
+
+	/* protected by rdev->mn_lock */
+	struct hlist_node	node;
+
+	/* objects protected by lock */
+	struct mutex		lock;
+	struct rb_root		objects;
+};
+
+/**
+ * radeon_mn_destroy - destroy the rmn
+ *
+ * @work: previously sheduled work item
+ *
+ * Lazy destroys the notifier from a work item
+ */
+static void radeon_mn_destroy(struct work_struct *work)
+{
+	struct radeon_mn *rmn = container_of(work, struct radeon_mn, work);
+	struct radeon_device *rdev = rmn->rdev;
+	struct radeon_bo *bo, *next;
+
+	mutex_lock(&rdev->mn_lock);
+	mutex_lock(&rmn->lock);
+	hash_del(&rmn->node);
+	rbtree_postorder_for_each_entry_safe(bo, next, &rmn->objects, mn_it.rb) {
+		interval_tree_remove(&bo->mn_it, &rmn->objects);
+		bo->mn = NULL;
+	}
+	mutex_unlock(&rmn->lock);
+	mutex_unlock(&rdev->mn_lock);
+	mmu_notifier_unregister(&rmn->mn, rmn->mm);
+	kfree(rmn);
+}
+
+/**
+ * radeon_mn_release - callback to notify about mm destruction
+ *
+ * @mn: our notifier
+ * @mn: the mm this callback is about
+ *
+ * Shedule a work item to lazy destroy our notifier.
+ */
+static void radeon_mn_release(struct mmu_notifier *mn,
+			      struct mm_struct *mm)
+{
+	struct radeon_mn *rmn = container_of(mn, struct radeon_mn, mn);
+	INIT_WORK(&rmn->work, radeon_mn_destroy);
+	schedule_work(&rmn->work);
+}
+
+/**
+ * radeon_mn_invalidate_range_start - callback to notify about mm change
+ *
+ * @mn: our notifier
+ * @mn: the mm this callback is about
+ * @start: start of updated range
+ * @end: end of updated range
+ *
+ * We block for all BOs between start and end to be idle and
+ * unmap them by move them into system domain again.
+ */
+static void radeon_mn_invalidate_range_start(struct mmu_notifier *mn,
+					     struct mm_struct *mm,
+					     unsigned long start,
+					     unsigned long end)
+{
+	struct radeon_mn *rmn = container_of(mn, struct radeon_mn, mn);
+	struct interval_tree_node *it;
+
+	/* notification is exclusive, but interval is inclusive */
+	end -= 1;
+
+	mutex_lock(&rmn->lock);
+
+	it = interval_tree_iter_first(&rmn->objects, start, end);
+	while (it) {
+		struct radeon_bo *bo;
+		struct fence *fence;
+		int r;
+
+		bo = container_of(it, struct radeon_bo, mn_it);
+		it = interval_tree_iter_next(it, start, end);
+
+		r = radeon_bo_reserve(bo, true);
+		if (r) {
+			DRM_ERROR("(%d) failed to reserve user bo\n", r);
+			continue;
+		}
+
+		fence = reservation_object_get_excl(bo->tbo.resv);
+		if (fence) {
+			r = radeon_fence_wait((struct radeon_fence *)fence, false);
+			if (r)
+				DRM_ERROR("(%d) failed to wait for user bo\n", r);
+		}
+
+		radeon_ttm_placement_from_domain(bo, RADEON_GEM_DOMAIN_CPU);
+		r = ttm_bo_validate(&bo->tbo, &bo->placement, false, false);
+		if (r)
+			DRM_ERROR("(%d) failed to validate user bo\n", r);
+
+		radeon_bo_unreserve(bo);
+	}
+	
+	mutex_unlock(&rmn->lock);
+}
+
+static const struct mmu_notifier_ops radeon_mn_ops = {
+	.release = radeon_mn_release,
+	.invalidate_range_start = radeon_mn_invalidate_range_start,
+};
+
+/**
+ * radeon_mn_get - create notifier context
+ *
+ * @rdev: radeon device pointer
+ *
+ * Creates a notifier context for current->mm.
+ */
+static struct radeon_mn *radeon_mn_get(struct radeon_device *rdev)
+{
+	struct mm_struct *mm = current->mm;
+	struct radeon_mn *rmn;
+	int r;
+
+	down_write(&mm->mmap_sem);
+	mutex_lock(&rdev->mn_lock);
+
+	hash_for_each_possible(rdev->mn_hash, rmn, node, (unsigned long)mm)
+		if (rmn->mm == mm)
+			goto release_locks;
+
+	rmn = kzalloc(sizeof(*rmn), GFP_KERNEL);
+	if (!rmn) {
+		rmn = ERR_PTR(-ENOMEM);
+		goto release_locks;
+	}
+
+	rmn->rdev = rdev;
+	rmn->mm = mm;
+	rmn->mn.ops = &radeon_mn_ops;
+	mutex_init(&rmn->lock);
+	rmn->objects = RB_ROOT;
+	
+	r = __mmu_notifier_register(&rmn->mn, mm);
+	if (r)
+		goto free_rmn;
+
+	hash_add(rdev->mn_hash, &rmn->node, (unsigned long)mm);
+
+release_locks:
+	mutex_unlock(&rdev->mn_lock);
+	up_write(&mm->mmap_sem);
+
+	return rmn;
+
+free_rmn:
+	mutex_unlock(&rdev->mn_lock);
+	up_write(&mm->mmap_sem);
+	kfree(rmn);
+
+	return ERR_PTR(r);
+}
+
+/**
+ * radeon_mn_register - register a BO for notifier updates
+ *
+ * @bo: radeon buffer object
+ * @addr: userptr addr we should monitor
+ *
+ * Registers an MMU notifier for the given BO at the specified address.
+ * Returns 0 on success, -ERRNO if anything goes wrong.
+ */
+int radeon_mn_register(struct radeon_bo *bo, unsigned long addr)
+{
+	unsigned long end = addr + radeon_bo_size(bo) - 1;
+	struct radeon_device *rdev = bo->rdev;
+	struct radeon_mn *rmn;
+	struct interval_tree_node *it;
+
+	rmn = radeon_mn_get(rdev);
+	if (IS_ERR(rmn))
+		return PTR_ERR(rmn);
+
+	mutex_lock(&rmn->lock);
+
+	it = interval_tree_iter_first(&rmn->objects, addr, end);
+	if (it) {
+		mutex_unlock(&rmn->lock);
+		return -EEXIST;
+	}
+
+	bo->mn = rmn;
+	bo->mn_it.start = addr;
+	bo->mn_it.last = end;
+	interval_tree_insert(&bo->mn_it, &rmn->objects);
+
+	mutex_unlock(&rmn->lock);
+
+	return 0;
+}
+
+/**
+ * radeon_mn_unregister - unregister a BO for notifier updates
+ *
+ * @bo: radeon buffer object
+ *
+ * Remove any registration of MMU notifier updates from the buffer object.
+ */
+void radeon_mn_unregister(struct radeon_bo *bo)
+{
+	struct radeon_device *rdev = bo->rdev;
+	struct radeon_mn *rmn;
+
+	mutex_lock(&rdev->mn_lock);
+	rmn = bo->mn;
+	if (rmn == NULL) {
+		mutex_unlock(&rdev->mn_lock);
+		return;
+	}
+
+	mutex_lock(&rmn->lock);
+	interval_tree_remove(&bo->mn_it, &rmn->objects);
+	bo->mn = NULL;
+	mutex_unlock(&rmn->lock);
+	mutex_unlock(&rdev->mn_lock);
+}
diff --git a/drivers/gpu/drm/radeon/radeon_mode.h b/drivers/gpu/drm/radeon/radeon_mode.h
index e27608c29c11..390db897f322 100644
--- a/drivers/gpu/drm/radeon/radeon_mode.h
+++ b/drivers/gpu/drm/radeon/radeon_mode.h
@@ -321,6 +321,10 @@ struct radeon_crtc {
 	uint32_t crtc_offset;
 	struct drm_gem_object *cursor_bo;
 	uint64_t cursor_addr;
+	int cursor_x;
+	int cursor_y;
+	int cursor_hot_x;
+	int cursor_hot_y;
 	int cursor_width;
 	int cursor_height;
 	int max_cursor_width;
@@ -462,6 +466,7 @@ struct radeon_gpio_rec {
 	u8 id;
 	u32 reg;
 	u32 mask;
+	u32 shift;
 };
 
 struct radeon_hpd {
@@ -748,6 +753,8 @@ extern bool radeon_atombios_get_ppll_ss_info(struct radeon_device *rdev,
 extern bool radeon_atombios_get_asic_ss_info(struct radeon_device *rdev,
 					     struct radeon_atom_ss *ss,
 					     int id, u32 clock);
+extern struct radeon_gpio_rec radeon_atombios_lookup_gpio(struct radeon_device *rdev,
+							  u8 id);
 
 extern void radeon_compute_pll_legacy(struct radeon_pll *pll,
 				      uint64_t freq,
@@ -777,6 +784,7 @@ extern void atombios_digital_setup(struct drm_encoder *encoder, int action);
 extern int atombios_get_encoder_mode(struct drm_encoder *encoder);
 extern bool atombios_set_edp_panel_power(struct drm_connector *connector, int action);
 extern void radeon_encoder_set_active_device(struct drm_encoder *encoder);
+extern bool radeon_encoder_is_digital(struct drm_encoder *encoder);
 
 extern void radeon_crtc_load_lut(struct drm_crtc *crtc);
 extern int atombios_crtc_set_base(struct drm_crtc *crtc, int x, int y,
@@ -801,13 +809,16 @@ extern int radeon_crtc_set_base_atomic(struct drm_crtc *crtc,
 extern int radeon_crtc_do_set_base(struct drm_crtc *crtc,
 				   struct drm_framebuffer *fb,
 				   int x, int y, int atomic);
-extern int radeon_crtc_cursor_set(struct drm_crtc *crtc,
-				  struct drm_file *file_priv,
-				  uint32_t handle,
-				  uint32_t width,
-				  uint32_t height);
+extern int radeon_crtc_cursor_set2(struct drm_crtc *crtc,
+				   struct drm_file *file_priv,
+				   uint32_t handle,
+				   uint32_t width,
+				   uint32_t height,
+				   int32_t hot_x,
+				   int32_t hot_y);
 extern int radeon_crtc_cursor_move(struct drm_crtc *crtc,
 				   int x, int y);
+extern void radeon_cursor_reset(struct drm_crtc *crtc);
 
 extern int radeon_get_crtc_scanoutpos(struct drm_device *dev, int crtc,
 				      unsigned int flags,
diff --git a/drivers/gpu/drm/radeon/radeon_object.c b/drivers/gpu/drm/radeon/radeon_object.c
index 480c87d8edc5..86fc56434b28 100644
--- a/drivers/gpu/drm/radeon/radeon_object.c
+++ b/drivers/gpu/drm/radeon/radeon_object.c
@@ -75,6 +75,7 @@ static void radeon_ttm_bo_destroy(struct ttm_buffer_object *tbo)
 	bo = container_of(tbo, struct radeon_bo, tbo);
 
 	radeon_update_memory_usage(bo, bo->tbo.mem.mem_type, -1);
+	radeon_mn_unregister(bo);
 
 	mutex_lock(&bo->rdev->gem.mutex);
 	list_del_init(&bo->list);
@@ -96,40 +97,83 @@ void radeon_ttm_placement_from_domain(struct radeon_bo *rbo, u32 domain)
 {
 	u32 c = 0, i;
 
-	rbo->placement.fpfn = 0;
-	rbo->placement.lpfn = 0;
 	rbo->placement.placement = rbo->placements;
 	rbo->placement.busy_placement = rbo->placements;
-	if (domain & RADEON_GEM_DOMAIN_VRAM)
-		rbo->placements[c++] = TTM_PL_FLAG_WC | TTM_PL_FLAG_UNCACHED |
-					TTM_PL_FLAG_VRAM;
+	if (domain & RADEON_GEM_DOMAIN_VRAM) {
+		/* Try placing BOs which don't need CPU access outside of the
+		 * CPU accessible part of VRAM
+		 */
+		if ((rbo->flags & RADEON_GEM_NO_CPU_ACCESS) &&
+		    rbo->rdev->mc.visible_vram_size < rbo->rdev->mc.real_vram_size) {
+			rbo->placements[c].fpfn =
+				rbo->rdev->mc.visible_vram_size >> PAGE_SHIFT;
+			rbo->placements[c++].flags = TTM_PL_FLAG_WC |
+						     TTM_PL_FLAG_UNCACHED |
+						     TTM_PL_FLAG_VRAM;
+		}
+
+		rbo->placements[c].fpfn = 0;
+		rbo->placements[c++].flags = TTM_PL_FLAG_WC |
+					     TTM_PL_FLAG_UNCACHED |
+					     TTM_PL_FLAG_VRAM;
+	}
+
 	if (domain & RADEON_GEM_DOMAIN_GTT) {
 		if (rbo->flags & RADEON_GEM_GTT_UC) {
-			rbo->placements[c++] = TTM_PL_FLAG_UNCACHED | TTM_PL_FLAG_TT;
+			rbo->placements[c].fpfn = 0;
+			rbo->placements[c++].flags = TTM_PL_FLAG_UNCACHED |
+				TTM_PL_FLAG_TT;
+
 		} else if ((rbo->flags & RADEON_GEM_GTT_WC) ||
 			   (rbo->rdev->flags & RADEON_IS_AGP)) {
-			rbo->placements[c++] = TTM_PL_FLAG_WC | TTM_PL_FLAG_UNCACHED |
+			rbo->placements[c].fpfn = 0;
+			rbo->placements[c++].flags = TTM_PL_FLAG_WC |
+				TTM_PL_FLAG_UNCACHED |
 				TTM_PL_FLAG_TT;
 		} else {
-			rbo->placements[c++] = TTM_PL_FLAG_CACHED | TTM_PL_FLAG_TT;
+			rbo->placements[c].fpfn = 0;
+			rbo->placements[c++].flags = TTM_PL_FLAG_CACHED |
+						     TTM_PL_FLAG_TT;
 		}
 	}
+
 	if (domain & RADEON_GEM_DOMAIN_CPU) {
 		if (rbo->flags & RADEON_GEM_GTT_UC) {
-			rbo->placements[c++] = TTM_PL_FLAG_UNCACHED | TTM_PL_FLAG_SYSTEM;
+			rbo->placements[c].fpfn = 0;
+			rbo->placements[c++].flags = TTM_PL_FLAG_UNCACHED |
+				TTM_PL_FLAG_SYSTEM;
+
 		} else if ((rbo->flags & RADEON_GEM_GTT_WC) ||
 		    rbo->rdev->flags & RADEON_IS_AGP) {
-			rbo->placements[c++] = TTM_PL_FLAG_WC | TTM_PL_FLAG_UNCACHED |
+			rbo->placements[c].fpfn = 0;
+			rbo->placements[c++].flags = TTM_PL_FLAG_WC |
+				TTM_PL_FLAG_UNCACHED |
 				TTM_PL_FLAG_SYSTEM;
 		} else {
-			rbo->placements[c++] = TTM_PL_FLAG_CACHED | TTM_PL_FLAG_SYSTEM;
+			rbo->placements[c].fpfn = 0;
+			rbo->placements[c++].flags = TTM_PL_FLAG_CACHED |
+						     TTM_PL_FLAG_SYSTEM;
 		}
 	}
-	if (!c)
-		rbo->placements[c++] = TTM_PL_MASK_CACHING | TTM_PL_FLAG_SYSTEM;
+	if (!c) {
+		rbo->placements[c].fpfn = 0;
+		rbo->placements[c++].flags = TTM_PL_MASK_CACHING |
+					     TTM_PL_FLAG_SYSTEM;
+	}
+
 	rbo->placement.num_placement = c;
 	rbo->placement.num_busy_placement = c;
 
+	for (i = 0; i < c; ++i) {
+		if ((rbo->flags & RADEON_GEM_CPU_ACCESS) &&
+		    (rbo->placements[i].flags & TTM_PL_FLAG_VRAM) &&
+		    !rbo->placements[i].fpfn)
+			rbo->placements[i].lpfn =
+				rbo->rdev->mc.visible_vram_size >> PAGE_SHIFT;
+		else
+			rbo->placements[i].lpfn = 0;
+	}
+
 	/*
 	 * Use two-ended allocation depending on the buffer size to
 	 * improve fragmentation quality.
@@ -137,14 +181,16 @@ void radeon_ttm_placement_from_domain(struct radeon_bo *rbo, u32 domain)
 	 */
 	if (rbo->tbo.mem.size > 512 * 1024) {
 		for (i = 0; i < c; i++) {
-			rbo->placements[i] |= TTM_PL_FLAG_TOPDOWN;
+			rbo->placements[i].flags |= TTM_PL_FLAG_TOPDOWN;
 		}
 	}
 }
 
 int radeon_bo_create(struct radeon_device *rdev,
-		     unsigned long size, int byte_align, bool kernel, u32 domain,
-		     u32 flags, struct sg_table *sg, struct radeon_bo **bo_ptr)
+		     unsigned long size, int byte_align, bool kernel,
+		     u32 domain, u32 flags, struct sg_table *sg,
+		     struct reservation_object *resv,
+		     struct radeon_bo **bo_ptr)
 {
 	struct radeon_bo *bo;
 	enum ttm_bo_type type;
@@ -187,12 +233,19 @@ int radeon_bo_create(struct radeon_device *rdev,
 	if (!(rdev->flags & RADEON_IS_PCIE))
 		bo->flags &= ~(RADEON_GEM_GTT_WC | RADEON_GEM_GTT_UC);
 
+#ifdef CONFIG_X86_32
+	/* XXX: Write-combined CPU mappings of GTT seem broken on 32-bit
+	 * See https://bugs.freedesktop.org/show_bug.cgi?id=84627
+	 */
+	bo->flags &= ~RADEON_GEM_GTT_WC;
+#endif
+
 	radeon_ttm_placement_from_domain(bo, domain);
 	/* Kernel allocation are uninterruptible */
 	down_read(&rdev->pm.mclk_lock);
 	r = ttm_bo_init(&rdev->mman.bdev, &bo->tbo, size, type,
 			&bo->placement, page_align, !kernel, NULL,
-			acc_size, sg, &radeon_ttm_bo_destroy);
+			acc_size, sg, resv, &radeon_ttm_bo_destroy);
 	up_read(&rdev->pm.mclk_lock);
 	if (unlikely(r != 0)) {
 		return r;
@@ -264,6 +317,9 @@ int radeon_bo_pin_restricted(struct radeon_bo *bo, u32 domain, u64 max_offset,
 {
 	int r, i;
 
+	if (radeon_ttm_tt_has_userptr(bo->tbo.ttm))
+		return -EPERM;
+
 	if (bo->pin_count) {
 		bo->pin_count++;
 		if (gpu_addr)
@@ -283,21 +339,19 @@ int radeon_bo_pin_restricted(struct radeon_bo *bo, u32 domain, u64 max_offset,
 		return 0;
 	}
 	radeon_ttm_placement_from_domain(bo, domain);
-	if (domain == RADEON_GEM_DOMAIN_VRAM) {
+	for (i = 0; i < bo->placement.num_placement; i++) {
 		/* force to pin into visible video ram */
-		bo->placement.lpfn = bo->rdev->mc.visible_vram_size >> PAGE_SHIFT;
-	}
-	if (max_offset) {
-		u64 lpfn = max_offset >> PAGE_SHIFT;
-
-		if (!bo->placement.lpfn)
-			bo->placement.lpfn = bo->rdev->mc.gtt_size >> PAGE_SHIFT;
+		if ((bo->placements[i].flags & TTM_PL_FLAG_VRAM) &&
+		    !(bo->flags & RADEON_GEM_NO_CPU_ACCESS) &&
+		    (!max_offset || max_offset > bo->rdev->mc.visible_vram_size))
+			bo->placements[i].lpfn =
+				bo->rdev->mc.visible_vram_size >> PAGE_SHIFT;
+		else
+			bo->placements[i].lpfn = max_offset >> PAGE_SHIFT;
 
-		if (lpfn < bo->placement.lpfn)
-			bo->placement.lpfn = lpfn;
+		bo->placements[i].flags |= TTM_PL_FLAG_NO_EVICT;
 	}
-	for (i = 0; i < bo->placement.num_placement; i++)
-		bo->placements[i] |= TTM_PL_FLAG_NO_EVICT;
+
 	r = ttm_bo_validate(&bo->tbo, &bo->placement, false, false);
 	if (likely(r == 0)) {
 		bo->pin_count = 1;
@@ -329,8 +383,10 @@ int radeon_bo_unpin(struct radeon_bo *bo)
 	bo->pin_count--;
 	if (bo->pin_count)
 		return 0;
-	for (i = 0; i < bo->placement.num_placement; i++)
-		bo->placements[i] &= ~TTM_PL_FLAG_NO_EVICT;
+	for (i = 0; i < bo->placement.num_placement; i++) {
+		bo->placements[i].lpfn = 0;
+		bo->placements[i].flags &= ~TTM_PL_FLAG_NO_EVICT;
+	}
 	r = ttm_bo_validate(&bo->tbo, &bo->placement, false, false);
 	if (likely(r == 0)) {
 		if (bo->tbo.mem.mem_type == TTM_PL_VRAM)
@@ -453,21 +509,23 @@ int radeon_bo_list_validate(struct radeon_device *rdev,
 			    struct ww_acquire_ctx *ticket,
 			    struct list_head *head, int ring)
 {
-	struct radeon_cs_reloc *lobj;
-	struct radeon_bo *bo;
+	struct radeon_bo_list *lobj;
+	struct list_head duplicates;
 	int r;
 	u64 bytes_moved = 0, initial_bytes_moved;
 	u64 bytes_moved_threshold = radeon_bo_get_threshold_for_moves(rdev);
 
-	r = ttm_eu_reserve_buffers(ticket, head);
+	INIT_LIST_HEAD(&duplicates);
+	r = ttm_eu_reserve_buffers(ticket, head, true, &duplicates);
 	if (unlikely(r != 0)) {
 		return r;
 	}
 
 	list_for_each_entry(lobj, head, tv.head) {
-		bo = lobj->robj;
+		struct radeon_bo *bo = lobj->robj;
 		if (!bo->pin_count) {
 			u32 domain = lobj->prefered_domains;
+			u32 allowed = lobj->allowed_domains;
 			u32 current_domain =
 				radeon_mem_type_to_domain(bo->tbo.mem.mem_type);
 
@@ -479,7 +537,7 @@ int radeon_bo_list_validate(struct radeon_device *rdev,
 			 * into account. We don't want to disallow buffer moves
 			 * completely.
 			 */
-			if ((lobj->allowed_domains & current_domain) != 0 &&
+			if ((allowed & current_domain) != 0 &&
 			    (domain & current_domain) == 0 && /* will be moved */
 			    bytes_moved > bytes_moved_threshold) {
 				/* don't move it */
@@ -489,7 +547,7 @@ int radeon_bo_list_validate(struct radeon_device *rdev,
 		retry:
 			radeon_ttm_placement_from_domain(bo, domain);
 			if (ring == R600_RING_TYPE_UVD_INDEX)
-				radeon_uvd_force_into_uvd_segment(bo);
+				radeon_uvd_force_into_uvd_segment(bo, allowed);
 
 			initial_bytes_moved = atomic64_read(&rdev->num_bytes_moved);
 			r = ttm_bo_validate(&bo->tbo, &bo->placement, true, false);
@@ -509,6 +567,12 @@ int radeon_bo_list_validate(struct radeon_device *rdev,
 		lobj->gpu_offset = radeon_bo_gpu_offset(bo);
 		lobj->tiling_flags = bo->tiling_flags;
 	}
+
+	list_for_each_entry(lobj, &duplicates, tv.head) {
+		lobj->gpu_offset = radeon_bo_gpu_offset(lobj->robj);
+		lobj->tiling_flags = lobj->robj->tiling_flags;
+	}
+
 	return 0;
 }
 
@@ -713,8 +777,8 @@ int radeon_bo_fault_reserve_notify(struct ttm_buffer_object *bo)
 {
 	struct radeon_device *rdev;
 	struct radeon_bo *rbo;
-	unsigned long offset, size;
-	int r;
+	unsigned long offset, size, lpfn;
+	int i, r;
 
 	if (!radeon_ttm_bo_is_radeon_bo(bo))
 		return 0;
@@ -731,7 +795,13 @@ int radeon_bo_fault_reserve_notify(struct ttm_buffer_object *bo)
 
 	/* hurrah the memory is not visible ! */
 	radeon_ttm_placement_from_domain(rbo, RADEON_GEM_DOMAIN_VRAM);
-	rbo->placement.lpfn = rdev->mc.visible_vram_size >> PAGE_SHIFT;
+	lpfn =	rdev->mc.visible_vram_size >> PAGE_SHIFT;
+	for (i = 0; i < rbo->placement.num_placement; i++) {
+		/* Force into visible VRAM */
+		if ((rbo->placements[i].flags & TTM_PL_FLAG_VRAM) &&
+		    (!rbo->placements[i].lpfn || rbo->placements[i].lpfn > lpfn))
+			rbo->placements[i].lpfn = lpfn;
+	}
 	r = ttm_bo_validate(bo, &rbo->placement, false, false);
 	if (unlikely(r == -ENOMEM)) {
 		radeon_ttm_placement_from_domain(rbo, RADEON_GEM_DOMAIN_GTT);
@@ -755,12 +825,29 @@ int radeon_bo_wait(struct radeon_bo *bo, u32 *mem_type, bool no_wait)
 	r = ttm_bo_reserve(&bo->tbo, true, no_wait, false, NULL);
 	if (unlikely(r != 0))
 		return r;
-	spin_lock(&bo->tbo.bdev->fence_lock);
 	if (mem_type)
 		*mem_type = bo->tbo.mem.mem_type;
-	if (bo->tbo.sync_obj)
-		r = ttm_bo_wait(&bo->tbo, true, true, no_wait);
-	spin_unlock(&bo->tbo.bdev->fence_lock);
+
+	r = ttm_bo_wait(&bo->tbo, true, true, no_wait);
 	ttm_bo_unreserve(&bo->tbo);
 	return r;
 }
+
+/**
+ * radeon_bo_fence - add fence to buffer object
+ *
+ * @bo: buffer object in question
+ * @fence: fence to add
+ * @shared: true if fence should be added shared
+ *
+ */
+void radeon_bo_fence(struct radeon_bo *bo, struct radeon_fence *fence,
+                     bool shared)
+{
+	struct reservation_object *resv = bo->tbo.resv;
+
+	if (shared)
+		reservation_object_add_shared_fence(resv, &fence->base);
+	else
+		reservation_object_add_excl_fence(resv, &fence->base);
+}
diff --git a/drivers/gpu/drm/radeon/radeon_object.h b/drivers/gpu/drm/radeon/radeon_object.h
index 98a47fdf3625..3b0b377f76cb 100644
--- a/drivers/gpu/drm/radeon/radeon_object.h
+++ b/drivers/gpu/drm/radeon/radeon_object.h
@@ -126,6 +126,7 @@ extern int radeon_bo_create(struct radeon_device *rdev,
 			    unsigned long size, int byte_align,
 			    bool kernel, u32 domain, u32 flags,
 			    struct sg_table *sg,
+			    struct reservation_object *resv,
 			    struct radeon_bo **bo_ptr);
 extern int radeon_bo_kmap(struct radeon_bo *bo, void **ptr);
 extern void radeon_bo_kunmap(struct radeon_bo *bo);
@@ -154,6 +155,8 @@ extern void radeon_bo_move_notify(struct ttm_buffer_object *bo,
 				  struct ttm_mem_reg *new_mem);
 extern int radeon_bo_fault_reserve_notify(struct ttm_buffer_object *bo);
 extern int radeon_bo_get_surface_reg(struct radeon_bo *bo);
+extern void radeon_bo_fence(struct radeon_bo *bo, struct radeon_fence *fence,
+			    bool shared);
 
 /*
  * sub allocation
diff --git a/drivers/gpu/drm/radeon/radeon_pm.c b/drivers/gpu/drm/radeon/radeon_pm.c
index 164898b0010c..32522cc940a1 100644
--- a/drivers/gpu/drm/radeon/radeon_pm.c
+++ b/drivers/gpu/drm/radeon/radeon_pm.c
@@ -1556,7 +1556,7 @@ static bool radeon_pm_in_vbl(struct radeon_device *rdev)
 		if (rdev->pm.active_crtcs & (1 << crtc)) {
 			vbl_status = radeon_get_crtc_scanoutpos(rdev->ddev, crtc, 0, &vpos, &hpos, NULL, NULL);
 			if ((vbl_status & DRM_SCANOUTPOS_VALID) &&
-			    !(vbl_status & DRM_SCANOUTPOS_INVBL))
+			    !(vbl_status & DRM_SCANOUTPOS_IN_VBLANK))
 				in_vbl = false;
 		}
 	}
diff --git a/drivers/gpu/drm/radeon/radeon_prime.c b/drivers/gpu/drm/radeon/radeon_prime.c
index 0b16f2cbcf17..f3609c97496b 100644
--- a/drivers/gpu/drm/radeon/radeon_prime.c
+++ b/drivers/gpu/drm/radeon/radeon_prime.c
@@ -27,6 +27,7 @@
 
 #include "radeon.h"
 #include <drm/radeon_drm.h>
+#include <linux/dma-buf.h>
 
 struct sg_table *radeon_gem_prime_get_sg_table(struct drm_gem_object *obj)
 {
@@ -57,15 +58,18 @@ void radeon_gem_prime_vunmap(struct drm_gem_object *obj, void *vaddr)
 }
 
 struct drm_gem_object *radeon_gem_prime_import_sg_table(struct drm_device *dev,
-							size_t size,
+							struct dma_buf_attachment *attach,
 							struct sg_table *sg)
 {
+	struct reservation_object *resv = attach->dmabuf->resv;
 	struct radeon_device *rdev = dev->dev_private;
 	struct radeon_bo *bo;
 	int ret;
 
-	ret = radeon_bo_create(rdev, size, PAGE_SIZE, false,
-			       RADEON_GEM_DOMAIN_GTT, 0, sg, &bo);
+	ww_mutex_lock(&resv->lock, NULL);
+	ret = radeon_bo_create(rdev, attach->dmabuf->size, PAGE_SIZE, false,
+			       RADEON_GEM_DOMAIN_GTT, 0, sg, resv, &bo);
+	ww_mutex_unlock(&resv->lock);
 	if (ret)
 		return ERR_PTR(ret);
 
@@ -111,3 +115,13 @@ struct reservation_object *radeon_gem_prime_res_obj(struct drm_gem_object *obj)
 
 	return bo->tbo.resv;
 }
+
+struct dma_buf *radeon_gem_prime_export(struct drm_device *dev,
+					struct drm_gem_object *gobj,
+					int flags)
+{
+	struct radeon_bo *bo = gem_to_radeon_bo(gobj);
+	if (radeon_ttm_tt_has_userptr(bo->tbo.ttm))
+		return ERR_PTR(-EPERM);
+	return drm_gem_prime_export(dev, gobj, flags);
+}
diff --git a/drivers/gpu/drm/radeon/radeon_ring.c b/drivers/gpu/drm/radeon/radeon_ring.c
index d65607902537..2456f69efd23 100644
--- a/drivers/gpu/drm/radeon/radeon_ring.c
+++ b/drivers/gpu/drm/radeon/radeon_ring.c
@@ -45,27 +45,6 @@
 static int radeon_debugfs_ring_init(struct radeon_device *rdev, struct radeon_ring *ring);
 
 /**
- * radeon_ring_write - write a value to the ring
- *
- * @ring: radeon_ring structure holding ring information
- * @v: dword (dw) value to write
- *
- * Write a value to the requested ring buffer (all asics).
- */
-void radeon_ring_write(struct radeon_ring *ring, uint32_t v)
-{
-#if DRM_DEBUG_CODE
-	if (ring->count_dw <= 0) {
-		DRM_ERROR("radeon: writing more dwords to the ring than expected!\n");
-	}
-#endif
-	ring->ring[ring->wptr++] = v;
-	ring->wptr &= ring->ptr_mask;
-	ring->count_dw--;
-	ring->ring_free_dw--;
-}
-
-/**
  * radeon_ring_supports_scratch_reg - check if the ring supports
  * writing to scratch registers
  *
@@ -335,7 +314,7 @@ unsigned radeon_ring_backup(struct radeon_device *rdev, struct radeon_ring *ring
 	}
 
 	/* and then save the content of the ring */
-	*data = kmalloc_array(size, sizeof(uint32_t), GFP_KERNEL);
+	*data = drm_malloc_ab(size, sizeof(uint32_t));
 	if (!*data) {
 		mutex_unlock(&rdev->ring_lock);
 		return 0;
@@ -377,7 +356,7 @@ int radeon_ring_restore(struct radeon_device *rdev, struct radeon_ring *ring,
 	}
 
 	radeon_ring_unlock_commit(rdev, ring, false);
-	kfree(data);
+	drm_free_large(data);
 	return 0;
 }
 
@@ -404,7 +383,7 @@ int radeon_ring_init(struct radeon_device *rdev, struct radeon_ring *ring, unsig
 	/* Allocate ring buffer */
 	if (ring->ring_obj == NULL) {
 		r = radeon_bo_create(rdev, ring->ring_size, PAGE_SIZE, true,
-				     RADEON_GEM_DOMAIN_GTT, 0,
+				     RADEON_GEM_DOMAIN_GTT, 0, NULL,
 				     NULL, &ring->ring_obj);
 		if (r) {
 			dev_err(rdev->dev, "(%d) ring create failed\n", r);
diff --git a/drivers/gpu/drm/radeon/radeon_sa.c b/drivers/gpu/drm/radeon/radeon_sa.c
index b84f97c8718c..c507896aca45 100644
--- a/drivers/gpu/drm/radeon/radeon_sa.c
+++ b/drivers/gpu/drm/radeon/radeon_sa.c
@@ -65,7 +65,7 @@ int radeon_sa_bo_manager_init(struct radeon_device *rdev,
 	}
 
 	r = radeon_bo_create(rdev, size, align, true,
-			     domain, flags, NULL, &sa_manager->bo);
+			     domain, flags, NULL, NULL, &sa_manager->bo);
 	if (r) {
 		dev_err(rdev->dev, "(%d) failed to allocate bo for manager\n", r);
 		return r;
diff --git a/drivers/gpu/drm/radeon/radeon_semaphore.c b/drivers/gpu/drm/radeon/radeon_semaphore.c
index abd6753a570a..e6ad54cdfa62 100644
--- a/drivers/gpu/drm/radeon/radeon_semaphore.c
+++ b/drivers/gpu/drm/radeon/radeon_semaphore.c
@@ -34,15 +34,14 @@
 int radeon_semaphore_create(struct radeon_device *rdev,
 			    struct radeon_semaphore **semaphore)
 {
-	uint64_t *cpu_addr;
-	int i, r;
+	int r;
 
 	*semaphore = kmalloc(sizeof(struct radeon_semaphore), GFP_KERNEL);
 	if (*semaphore == NULL) {
 		return -ENOMEM;
 	}
-	r = radeon_sa_bo_new(rdev, &rdev->ring_tmp_bo, &(*semaphore)->sa_bo,
-			     8 * RADEON_NUM_SYNCS, 8);
+	r = radeon_sa_bo_new(rdev, &rdev->ring_tmp_bo,
+			     &(*semaphore)->sa_bo, 8, 8);
 	if (r) {
 		kfree(*semaphore);
 		*semaphore = NULL;
@@ -51,12 +50,7 @@ int radeon_semaphore_create(struct radeon_device *rdev,
 	(*semaphore)->waiters = 0;
 	(*semaphore)->gpu_addr = radeon_sa_bo_gpu_addr((*semaphore)->sa_bo);
 
-	cpu_addr = radeon_sa_bo_cpu_addr((*semaphore)->sa_bo);
-	for (i = 0; i < RADEON_NUM_SYNCS; ++i)
-		cpu_addr[i] = 0;
-
-	for (i = 0; i < RADEON_NUM_RINGS; ++i)
-		(*semaphore)->sync_to[i] = NULL;
+	*((uint64_t *)radeon_sa_bo_cpu_addr((*semaphore)->sa_bo)) = 0;
 
 	return 0;
 }
@@ -95,99 +89,6 @@ bool radeon_semaphore_emit_wait(struct radeon_device *rdev, int ridx,
 	return false;
 }
 
-/**
- * radeon_semaphore_sync_to - use the semaphore to sync to a fence
- *
- * @semaphore: semaphore object to add fence to
- * @fence: fence to sync to
- *
- * Sync to the fence using this semaphore object
- */
-void radeon_semaphore_sync_to(struct radeon_semaphore *semaphore,
-			      struct radeon_fence *fence)
-{
-        struct radeon_fence *other;
-
-        if (!fence)
-                return;
-
-        other = semaphore->sync_to[fence->ring];
-        semaphore->sync_to[fence->ring] = radeon_fence_later(fence, other);
-}
-
-/**
- * radeon_semaphore_sync_rings - sync ring to all registered fences
- *
- * @rdev: radeon_device pointer
- * @semaphore: semaphore object to use for sync
- * @ring: ring that needs sync
- *
- * Ensure that all registered fences are signaled before letting
- * the ring continue. The caller must hold the ring lock.
- */
-int radeon_semaphore_sync_rings(struct radeon_device *rdev,
-				struct radeon_semaphore *semaphore,
-				int ring)
-{
-	unsigned count = 0;
-	int i, r;
-
-        for (i = 0; i < RADEON_NUM_RINGS; ++i) {
-		struct radeon_fence *fence = semaphore->sync_to[i];
-
-		/* check if we really need to sync */
-                if (!radeon_fence_need_sync(fence, ring))
-			continue;
-
-		/* prevent GPU deadlocks */
-		if (!rdev->ring[i].ready) {
-			dev_err(rdev->dev, "Syncing to a disabled ring!");
-			return -EINVAL;
-		}
-
-		if (++count > RADEON_NUM_SYNCS) {
-			/* not enough room, wait manually */
-			r = radeon_fence_wait(fence, false);
-			if (r)
-				return r;
-			continue;
-		}
-
-		/* allocate enough space for sync command */
-		r = radeon_ring_alloc(rdev, &rdev->ring[i], 16);
-		if (r) {
-			return r;
-		}
-
-		/* emit the signal semaphore */
-		if (!radeon_semaphore_emit_signal(rdev, i, semaphore)) {
-			/* signaling wasn't successful wait manually */
-			radeon_ring_undo(&rdev->ring[i]);
-			r = radeon_fence_wait(fence, false);
-			if (r)
-				return r;
-			continue;
-		}
-
-		/* we assume caller has already allocated space on waiters ring */
-		if (!radeon_semaphore_emit_wait(rdev, ring, semaphore)) {
-			/* waiting wasn't successful wait manually */
-			radeon_ring_undo(&rdev->ring[i]);
-			r = radeon_fence_wait(fence, false);
-			if (r)
-				return r;
-			continue;
-		}
-
-		radeon_ring_commit(rdev, &rdev->ring[i], false);
-		radeon_fence_note_sync(fence, ring);
-
-		semaphore->gpu_addr += 8;
-	}
-
-	return 0;
-}
-
 void radeon_semaphore_free(struct radeon_device *rdev,
 			   struct radeon_semaphore **semaphore,
 			   struct radeon_fence *fence)
diff --git a/drivers/gpu/drm/radeon/radeon_state.c b/drivers/gpu/drm/radeon/radeon_state.c
index 23bb64fd775f..15aee723db77 100644
--- a/drivers/gpu/drm/radeon/radeon_state.c
+++ b/drivers/gpu/drm/radeon/radeon_state.c
@@ -30,9 +30,9 @@
  */
 
 #include <drm/drmP.h>
-#include <drm/drm_buffer.h>
 #include <drm/radeon_drm.h>
 #include "radeon_drv.h"
+#include "drm_buffer.h"
 
 /* ================================================================
  * Helper functions for client state checking and fixup
@@ -1703,7 +1703,7 @@ static int radeon_cp_dispatch_texture(struct drm_device * dev,
 	u32 format;
 	u32 *buffer;
 	const u8 __user *data;
-	int size, dwords, tex_width, blit_width, spitch;
+	unsigned int size, dwords, tex_width, blit_width, spitch;
 	u32 height;
 	int i;
 	u32 texpitch, microtile;
diff --git a/drivers/gpu/drm/radeon/radeon_sync.c b/drivers/gpu/drm/radeon/radeon_sync.c
new file mode 100644
index 000000000000..02ac8a1de4ff
--- /dev/null
+++ b/drivers/gpu/drm/radeon/radeon_sync.c
@@ -0,0 +1,220 @@
+/*
+ * Copyright 2014 Advanced Micro Devices, Inc.
+ * All Rights Reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the
+ * "Software"), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sub license, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDERS, AUTHORS AND/OR ITS SUPPLIERS BE LIABLE FOR ANY CLAIM,
+ * DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR
+ * OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE
+ * USE OR OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * The above copyright notice and this permission notice (including the
+ * next paragraph) shall be included in all copies or substantial portions
+ * of the Software.
+ *
+ */
+/*
+ * Authors:
+ *    Christian König <christian.koenig@amd.com>
+ */
+
+#include <drm/drmP.h>
+#include "radeon.h"
+#include "radeon_trace.h"
+
+/**
+ * radeon_sync_create - zero init sync object
+ *
+ * @sync: sync object to initialize
+ *
+ * Just clear the sync object for now.
+ */
+void radeon_sync_create(struct radeon_sync *sync)
+{
+	unsigned i;
+
+	for (i = 0; i < RADEON_NUM_SYNCS; ++i)
+		sync->semaphores[i] = NULL;
+
+	for (i = 0; i < RADEON_NUM_RINGS; ++i)
+		sync->sync_to[i] = NULL;
+
+	sync->last_vm_update = NULL;
+}
+
+/**
+ * radeon_sync_fence - use the semaphore to sync to a fence
+ *
+ * @sync: sync object to add fence to
+ * @fence: fence to sync to
+ *
+ * Sync to the fence using the semaphore objects
+ */
+void radeon_sync_fence(struct radeon_sync *sync,
+		       struct radeon_fence *fence)
+{
+	struct radeon_fence *other;
+
+	if (!fence)
+		return;
+
+	other = sync->sync_to[fence->ring];
+	sync->sync_to[fence->ring] = radeon_fence_later(fence, other);
+
+	if (fence->is_vm_update) {
+		other = sync->last_vm_update;
+		sync->last_vm_update = radeon_fence_later(fence, other);
+	}
+}
+
+/**
+ * radeon_sync_resv - use the semaphores to sync to a reservation object
+ *
+ * @sync: sync object to add fences from reservation object to
+ * @resv: reservation object with embedded fence
+ * @shared: true if we should only sync to the exclusive fence
+ *
+ * Sync to the fence using the semaphore objects
+ */
+int radeon_sync_resv(struct radeon_device *rdev,
+		     struct radeon_sync *sync,
+		     struct reservation_object *resv,
+		     bool shared)
+{
+	struct reservation_object_list *flist;
+	struct fence *f;
+	struct radeon_fence *fence;
+	unsigned i;
+	int r = 0;
+
+	/* always sync to the exclusive fence */
+	f = reservation_object_get_excl(resv);
+	fence = f ? to_radeon_fence(f) : NULL;
+	if (fence && fence->rdev == rdev)
+		radeon_sync_fence(sync, fence);
+	else if (f)
+		r = fence_wait(f, true);
+
+	flist = reservation_object_get_list(resv);
+	if (shared || !flist || r)
+		return r;
+
+	for (i = 0; i < flist->shared_count; ++i) {
+		f = rcu_dereference_protected(flist->shared[i],
+					      reservation_object_held(resv));
+		fence = to_radeon_fence(f);
+		if (fence && fence->rdev == rdev)
+			radeon_sync_fence(sync, fence);
+		else
+			r = fence_wait(f, true);
+
+		if (r)
+			break;
+	}
+	return r;
+}
+
+/**
+ * radeon_sync_rings - sync ring to all registered fences
+ *
+ * @rdev: radeon_device pointer
+ * @sync: sync object to use
+ * @ring: ring that needs sync
+ *
+ * Ensure that all registered fences are signaled before letting
+ * the ring continue. The caller must hold the ring lock.
+ */
+int radeon_sync_rings(struct radeon_device *rdev,
+		      struct radeon_sync *sync,
+		      int ring)
+{
+	unsigned count = 0;
+	int i, r;
+
+	for (i = 0; i < RADEON_NUM_RINGS; ++i) {
+		struct radeon_fence *fence = sync->sync_to[i];
+		struct radeon_semaphore *semaphore;
+
+		/* check if we really need to sync */
+		if (!radeon_fence_need_sync(fence, ring))
+			continue;
+
+		/* prevent GPU deadlocks */
+		if (!rdev->ring[i].ready) {
+			dev_err(rdev->dev, "Syncing to a disabled ring!");
+			return -EINVAL;
+		}
+
+		if (count >= RADEON_NUM_SYNCS) {
+			/* not enough room, wait manually */
+			r = radeon_fence_wait(fence, false);
+			if (r)
+				return r;
+			continue;
+		}
+		r = radeon_semaphore_create(rdev, &semaphore);
+		if (r)
+			return r;
+
+		sync->semaphores[count++] = semaphore;
+
+		/* allocate enough space for sync command */
+		r = radeon_ring_alloc(rdev, &rdev->ring[i], 16);
+		if (r)
+			return r;
+
+		/* emit the signal semaphore */
+		if (!radeon_semaphore_emit_signal(rdev, i, semaphore)) {
+			/* signaling wasn't successful wait manually */
+			radeon_ring_undo(&rdev->ring[i]);
+			r = radeon_fence_wait(fence, false);
+			if (r)
+				return r;
+			continue;
+		}
+
+		/* we assume caller has already allocated space on waiters ring */
+		if (!radeon_semaphore_emit_wait(rdev, ring, semaphore)) {
+			/* waiting wasn't successful wait manually */
+			radeon_ring_undo(&rdev->ring[i]);
+			r = radeon_fence_wait(fence, false);
+			if (r)
+				return r;
+			continue;
+		}
+
+		radeon_ring_commit(rdev, &rdev->ring[i], false);
+		radeon_fence_note_sync(fence, ring);
+	}
+
+	return 0;
+}
+
+/**
+ * radeon_sync_free - free the sync object
+ *
+ * @rdev: radeon_device pointer
+ * @sync: sync object to use
+ * @fence: fence to use for the free
+ *
+ * Free the sync object by freeing all semaphores in it.
+ */
+void radeon_sync_free(struct radeon_device *rdev,
+		      struct radeon_sync *sync,
+		      struct radeon_fence *fence)
+{
+	unsigned i;
+
+	for (i = 0; i < RADEON_NUM_SYNCS; ++i)
+		radeon_semaphore_free(rdev, &sync->semaphores[i], fence);
+}
diff --git a/drivers/gpu/drm/radeon/radeon_test.c b/drivers/gpu/drm/radeon/radeon_test.c
index 17bc3dced9f1..07b506b41008 100644
--- a/drivers/gpu/drm/radeon/radeon_test.c
+++ b/drivers/gpu/drm/radeon/radeon_test.c
@@ -67,7 +67,7 @@ static void radeon_do_test_moves(struct radeon_device *rdev, int flag)
 	}
 
 	r = radeon_bo_create(rdev, size, PAGE_SIZE, true, RADEON_GEM_DOMAIN_VRAM,
-			     0, NULL, &vram_obj);
+			     0, NULL, NULL, &vram_obj);
 	if (r) {
 		DRM_ERROR("Failed to create VRAM object\n");
 		goto out_cleanup;
@@ -87,7 +87,8 @@ static void radeon_do_test_moves(struct radeon_device *rdev, int flag)
 		struct radeon_fence *fence = NULL;
 
 		r = radeon_bo_create(rdev, size, PAGE_SIZE, true,
-				     RADEON_GEM_DOMAIN_GTT, 0, NULL, gtt_obj + i);
+				     RADEON_GEM_DOMAIN_GTT, 0, NULL, NULL,
+				     gtt_obj + i);
 		if (r) {
 			DRM_ERROR("Failed to create GTT object %d\n", i);
 			goto out_lclean;
@@ -116,11 +117,16 @@ static void radeon_do_test_moves(struct radeon_device *rdev, int flag)
 		radeon_bo_kunmap(gtt_obj[i]);
 
 		if (ring == R600_RING_TYPE_DMA_INDEX)
-			r = radeon_copy_dma(rdev, gtt_addr, vram_addr, size / RADEON_GPU_PAGE_SIZE, &fence);
+			fence = radeon_copy_dma(rdev, gtt_addr, vram_addr,
+						size / RADEON_GPU_PAGE_SIZE,
+						NULL);
 		else
-			r = radeon_copy_blit(rdev, gtt_addr, vram_addr, size / RADEON_GPU_PAGE_SIZE, &fence);
-		if (r) {
+			fence = radeon_copy_blit(rdev, gtt_addr, vram_addr,
+						 size / RADEON_GPU_PAGE_SIZE,
+						 NULL);
+		if (IS_ERR(fence)) {
 			DRM_ERROR("Failed GTT->VRAM copy %d\n", i);
+			r = PTR_ERR(fence);
 			goto out_lclean_unpin;
 		}
 
@@ -162,11 +168,16 @@ static void radeon_do_test_moves(struct radeon_device *rdev, int flag)
 		radeon_bo_kunmap(vram_obj);
 
 		if (ring == R600_RING_TYPE_DMA_INDEX)
-			r = radeon_copy_dma(rdev, vram_addr, gtt_addr, size / RADEON_GPU_PAGE_SIZE, &fence);
+			fence = radeon_copy_dma(rdev, vram_addr, gtt_addr,
+						size / RADEON_GPU_PAGE_SIZE,
+						NULL);
 		else
-			r = radeon_copy_blit(rdev, vram_addr, gtt_addr, size / RADEON_GPU_PAGE_SIZE, &fence);
-		if (r) {
+			fence = radeon_copy_blit(rdev, vram_addr, gtt_addr,
+						 size / RADEON_GPU_PAGE_SIZE,
+						 NULL);
+		if (IS_ERR(fence)) {
 			DRM_ERROR("Failed VRAM->GTT copy %d\n", i);
+			r = PTR_ERR(fence);
 			goto out_lclean_unpin;
 		}
 
@@ -222,7 +233,7 @@ out_lclean:
 			radeon_bo_unreserve(gtt_obj[i]);
 			radeon_bo_unref(&gtt_obj[i]);
 		}
-		if (fence)
+		if (fence && !IS_ERR(fence))
 			radeon_fence_unref(&fence);
 		break;
 	}
diff --git a/drivers/gpu/drm/radeon/radeon_trace.h b/drivers/gpu/drm/radeon/radeon_trace.h
index 9db74a96ef61..ce075cb08cb2 100644
--- a/drivers/gpu/drm/radeon/radeon_trace.h
+++ b/drivers/gpu/drm/radeon/radeon_trace.h
@@ -38,7 +38,7 @@ TRACE_EVENT(radeon_cs,
 
 	    TP_fast_assign(
 			   __entry->ring = p->ring;
-			   __entry->dw = p->chunks[p->chunk_ib_idx].length_dw;
+			   __entry->dw = p->chunk_ib->length_dw;
 			   __entry->fences = radeon_fence_count_emitted(
 				p->rdev, p->ring);
 			   ),
diff --git a/drivers/gpu/drm/radeon/radeon_ttm.c b/drivers/gpu/drm/radeon/radeon_ttm.c
index 72afe82a95c9..d02aa1d0f588 100644
--- a/drivers/gpu/drm/radeon/radeon_ttm.c
+++ b/drivers/gpu/drm/radeon/radeon_ttm.c
@@ -39,6 +39,8 @@
 #include <linux/seq_file.h>
 #include <linux/slab.h>
 #include <linux/swiotlb.h>
+#include <linux/swap.h>
+#include <linux/pagemap.h>
 #include <linux/debugfs.h>
 #include "radeon_reg.h"
 #include "radeon.h"
@@ -176,12 +178,15 @@ static int radeon_init_mem_type(struct ttm_bo_device *bdev, uint32_t type,
 static void radeon_evict_flags(struct ttm_buffer_object *bo,
 				struct ttm_placement *placement)
 {
+	static struct ttm_place placements = {
+		.fpfn = 0,
+		.lpfn = 0,
+		.flags = TTM_PL_MASK_CACHING | TTM_PL_FLAG_SYSTEM
+	};
+
 	struct radeon_bo *rbo;
-	static u32 placements = TTM_PL_MASK_CACHING | TTM_PL_FLAG_SYSTEM;
 
 	if (!radeon_ttm_bo_is_radeon_bo(bo)) {
-		placement->fpfn = 0;
-		placement->lpfn = 0;
 		placement->placement = &placements;
 		placement->busy_placement = &placements;
 		placement->num_placement = 1;
@@ -191,9 +196,32 @@ static void radeon_evict_flags(struct ttm_buffer_object *bo,
 	rbo = container_of(bo, struct radeon_bo, tbo);
 	switch (bo->mem.mem_type) {
 	case TTM_PL_VRAM:
-		if (rbo->rdev->ring[RADEON_RING_TYPE_GFX_INDEX].ready == false)
+		if (rbo->rdev->ring[radeon_copy_ring_index(rbo->rdev)].ready == false)
 			radeon_ttm_placement_from_domain(rbo, RADEON_GEM_DOMAIN_CPU);
-		else
+		else if (rbo->rdev->mc.visible_vram_size < rbo->rdev->mc.real_vram_size &&
+			 bo->mem.start < (rbo->rdev->mc.visible_vram_size >> PAGE_SHIFT)) {
+			unsigned fpfn = rbo->rdev->mc.visible_vram_size >> PAGE_SHIFT;
+			int i;
+
+			/* Try evicting to the CPU inaccessible part of VRAM
+			 * first, but only set GTT as busy placement, so this
+			 * BO will be evicted to GTT rather than causing other
+			 * BOs to be evicted from VRAM
+			 */
+			radeon_ttm_placement_from_domain(rbo, RADEON_GEM_DOMAIN_VRAM |
+							 RADEON_GEM_DOMAIN_GTT);
+			rbo->placement.num_busy_placement = 0;
+			for (i = 0; i < rbo->placement.num_placement; i++) {
+				if (rbo->placements[i].flags & TTM_PL_FLAG_VRAM) {
+					if (rbo->placements[0].fpfn < fpfn)
+						rbo->placements[0].fpfn = fpfn;
+				} else {
+					rbo->placement.busy_placement =
+						&rbo->placements[i];
+					rbo->placement.num_busy_placement = 1;
+				}
+			}
+		} else
 			radeon_ttm_placement_from_domain(rbo, RADEON_GEM_DOMAIN_GTT);
 		break;
 	case TTM_PL_TT:
@@ -228,6 +256,7 @@ static int radeon_move_blit(struct ttm_buffer_object *bo,
 	struct radeon_device *rdev;
 	uint64_t old_start, new_start;
 	struct radeon_fence *fence;
+	unsigned num_pages;
 	int r, ridx;
 
 	rdev = radeon_get_rdev(bo->bdev);
@@ -264,13 +293,12 @@ static int radeon_move_blit(struct ttm_buffer_object *bo,
 
 	BUILD_BUG_ON((PAGE_SIZE % RADEON_GPU_PAGE_SIZE) != 0);
 
-	/* sync other rings */
-	fence = bo->sync_obj;
-	r = radeon_copy(rdev, old_start, new_start,
-			new_mem->num_pages * (PAGE_SIZE / RADEON_GPU_PAGE_SIZE), /* GPU pages */
-			&fence);
-	/* FIXME: handle copy error */
-	r = ttm_bo_move_accel_cleanup(bo, (void *)fence,
+	num_pages = new_mem->num_pages * (PAGE_SIZE / RADEON_GPU_PAGE_SIZE);
+	fence = radeon_copy(rdev, old_start, new_start, num_pages, bo->resv);
+	if (IS_ERR(fence))
+		return PTR_ERR(fence);
+
+	r = ttm_bo_move_accel_cleanup(bo, &fence->base,
 				      evict, no_wait_gpu, new_mem);
 	radeon_fence_unref(&fence);
 	return r;
@@ -284,20 +312,20 @@ static int radeon_move_vram_ram(struct ttm_buffer_object *bo,
 	struct radeon_device *rdev;
 	struct ttm_mem_reg *old_mem = &bo->mem;
 	struct ttm_mem_reg tmp_mem;
-	u32 placements;
+	struct ttm_place placements;
 	struct ttm_placement placement;
 	int r;
 
 	rdev = radeon_get_rdev(bo->bdev);
 	tmp_mem = *new_mem;
 	tmp_mem.mm_node = NULL;
-	placement.fpfn = 0;
-	placement.lpfn = 0;
 	placement.num_placement = 1;
 	placement.placement = &placements;
 	placement.num_busy_placement = 1;
 	placement.busy_placement = &placements;
-	placements = TTM_PL_MASK_CACHING | TTM_PL_FLAG_TT;
+	placements.fpfn = 0;
+	placements.lpfn = 0;
+	placements.flags = TTM_PL_MASK_CACHING | TTM_PL_FLAG_TT;
 	r = ttm_bo_mem_space(bo, &placement, &tmp_mem,
 			     interruptible, no_wait_gpu);
 	if (unlikely(r)) {
@@ -332,19 +360,19 @@ static int radeon_move_ram_vram(struct ttm_buffer_object *bo,
 	struct ttm_mem_reg *old_mem = &bo->mem;
 	struct ttm_mem_reg tmp_mem;
 	struct ttm_placement placement;
-	u32 placements;
+	struct ttm_place placements;
 	int r;
 
 	rdev = radeon_get_rdev(bo->bdev);
 	tmp_mem = *new_mem;
 	tmp_mem.mm_node = NULL;
-	placement.fpfn = 0;
-	placement.lpfn = 0;
 	placement.num_placement = 1;
 	placement.placement = &placements;
 	placement.num_busy_placement = 1;
 	placement.busy_placement = &placements;
-	placements = TTM_PL_MASK_CACHING | TTM_PL_FLAG_TT;
+	placements.fpfn = 0;
+	placements.lpfn = 0;
+	placements.flags = TTM_PL_MASK_CACHING | TTM_PL_FLAG_TT;
 	r = ttm_bo_mem_space(bo, &placement, &tmp_mem,
 			     interruptible, no_wait_gpu);
 	if (unlikely(r)) {
@@ -483,39 +511,108 @@ static void radeon_ttm_io_mem_free(struct ttm_bo_device *bdev, struct ttm_mem_re
 {
 }
 
-static int radeon_sync_obj_wait(void *sync_obj, bool lazy, bool interruptible)
-{
-	return radeon_fence_wait((struct radeon_fence *)sync_obj, interruptible);
-}
+/*
+ * TTM backend functions.
+ */
+struct radeon_ttm_tt {
+	struct ttm_dma_tt		ttm;
+	struct radeon_device		*rdev;
+	u64				offset;
 
-static int radeon_sync_obj_flush(void *sync_obj)
+	uint64_t			userptr;
+	struct mm_struct		*usermm;
+	uint32_t			userflags;
+};
+
+/* prepare the sg table with the user pages */
+static int radeon_ttm_tt_pin_userptr(struct ttm_tt *ttm)
 {
+	struct radeon_device *rdev = radeon_get_rdev(ttm->bdev);
+	struct radeon_ttm_tt *gtt = (void *)ttm;
+	unsigned pinned = 0, nents;
+	int r;
+
+	int write = !(gtt->userflags & RADEON_GEM_USERPTR_READONLY);
+	enum dma_data_direction direction = write ?
+		DMA_BIDIRECTIONAL : DMA_TO_DEVICE;
+
+	if (current->mm != gtt->usermm)
+		return -EPERM;
+
+	if (gtt->userflags & RADEON_GEM_USERPTR_ANONONLY) {
+		/* check that we only pin down anonymous memory
+		   to prevent problems with writeback */
+		unsigned long end = gtt->userptr + ttm->num_pages * PAGE_SIZE;
+		struct vm_area_struct *vma;
+		vma = find_vma(gtt->usermm, gtt->userptr);
+		if (!vma || vma->vm_file || vma->vm_end < end)
+			return -EPERM;
+	}
+
+	do {
+		unsigned num_pages = ttm->num_pages - pinned;
+		uint64_t userptr = gtt->userptr + pinned * PAGE_SIZE;
+		struct page **pages = ttm->pages + pinned;
+
+		r = get_user_pages(current, current->mm, userptr, num_pages,
+				   write, 0, pages, NULL);
+		if (r < 0)
+			goto release_pages;
+
+		pinned += r;
+
+	} while (pinned < ttm->num_pages);
+
+	r = sg_alloc_table_from_pages(ttm->sg, ttm->pages, ttm->num_pages, 0,
+				      ttm->num_pages << PAGE_SHIFT,
+				      GFP_KERNEL);
+	if (r)
+		goto release_sg;
+
+	r = -ENOMEM;
+	nents = dma_map_sg(rdev->dev, ttm->sg->sgl, ttm->sg->nents, direction);
+	if (nents != ttm->sg->nents)
+		goto release_sg;
+
+	drm_prime_sg_to_page_addr_arrays(ttm->sg, ttm->pages,
+					 gtt->ttm.dma_address, ttm->num_pages);
+
 	return 0;
-}
 
-static void radeon_sync_obj_unref(void **sync_obj)
-{
-	radeon_fence_unref((struct radeon_fence **)sync_obj);
-}
+release_sg:
+	kfree(ttm->sg);
 
-static void *radeon_sync_obj_ref(void *sync_obj)
-{
-	return radeon_fence_ref((struct radeon_fence *)sync_obj);
+release_pages:
+	release_pages(ttm->pages, pinned, 0);
+	return r;
 }
 
-static bool radeon_sync_obj_signaled(void *sync_obj)
+static void radeon_ttm_tt_unpin_userptr(struct ttm_tt *ttm)
 {
-	return radeon_fence_signaled((struct radeon_fence *)sync_obj);
-}
+	struct radeon_device *rdev = radeon_get_rdev(ttm->bdev);
+	struct radeon_ttm_tt *gtt = (void *)ttm;
+	struct scatterlist *sg;
+	int i;
 
-/*
- * TTM backend functions.
- */
-struct radeon_ttm_tt {
-	struct ttm_dma_tt		ttm;
-	struct radeon_device		*rdev;
-	u64				offset;
-};
+	int write = !(gtt->userflags & RADEON_GEM_USERPTR_READONLY);
+	enum dma_data_direction direction = write ?
+		DMA_BIDIRECTIONAL : DMA_TO_DEVICE;
+
+	/* free the sg table and pages again */
+	dma_unmap_sg(rdev->dev, ttm->sg->sgl, ttm->sg->nents, direction);
+
+	for_each_sg(ttm->sg->sgl, sg, ttm->sg->nents, i) {
+		struct page *page = sg_page(sg);
+
+		if (!(gtt->userflags & RADEON_GEM_USERPTR_READONLY))
+			set_page_dirty(page);
+
+		mark_page_accessed(page);
+		page_cache_release(page);
+	}
+
+	sg_free_table(ttm->sg);
+}
 
 static int radeon_ttm_backend_bind(struct ttm_tt *ttm,
 				   struct ttm_mem_reg *bo_mem)
@@ -525,6 +622,11 @@ static int radeon_ttm_backend_bind(struct ttm_tt *ttm,
 		RADEON_GART_PAGE_WRITE;
 	int r;
 
+	if (gtt->userptr) {
+		radeon_ttm_tt_pin_userptr(ttm);
+		flags &= ~RADEON_GART_PAGE_WRITE;
+	}
+
 	gtt->offset = (unsigned long)(bo_mem->start << PAGE_SHIFT);
 	if (!ttm->num_pages) {
 		WARN(1, "nothing to bind %lu pages for mreg %p back %p!\n",
@@ -547,6 +649,10 @@ static int radeon_ttm_backend_unbind(struct ttm_tt *ttm)
 	struct radeon_ttm_tt *gtt = (void *)ttm;
 
 	radeon_gart_unbind(gtt->rdev, gtt->offset, ttm->num_pages);
+
+	if (gtt->userptr)
+		radeon_ttm_tt_unpin_userptr(ttm);
+
 	return 0;
 }
 
@@ -592,10 +698,17 @@ static struct ttm_tt *radeon_ttm_tt_create(struct ttm_bo_device *bdev,
 	return &gtt->ttm.ttm;
 }
 
+static struct radeon_ttm_tt *radeon_ttm_tt_to_gtt(struct ttm_tt *ttm)
+{
+	if (!ttm || ttm->func != &radeon_backend_func)
+		return NULL;
+	return (struct radeon_ttm_tt *)ttm;
+}
+
 static int radeon_ttm_tt_populate(struct ttm_tt *ttm)
 {
+	struct radeon_ttm_tt *gtt = radeon_ttm_tt_to_gtt(ttm);
 	struct radeon_device *rdev;
-	struct radeon_ttm_tt *gtt = (void *)ttm;
 	unsigned i;
 	int r;
 	bool slave = !!(ttm->page_flags & TTM_PAGE_FLAG_SG);
@@ -603,6 +716,16 @@ static int radeon_ttm_tt_populate(struct ttm_tt *ttm)
 	if (ttm->state != tt_unpopulated)
 		return 0;
 
+	if (gtt && gtt->userptr) {
+		ttm->sg = kcalloc(1, sizeof(struct sg_table), GFP_KERNEL);
+		if (!ttm->sg)
+			return -ENOMEM;
+
+		ttm->page_flags |= TTM_PAGE_FLAG_SG;
+		ttm->state = tt_unbound;
+		return 0;
+	}
+
 	if (slave && ttm->sg) {
 		drm_prime_sg_to_page_addr_arrays(ttm->sg, ttm->pages,
 						 gtt->ttm.dma_address, ttm->num_pages);
@@ -648,10 +771,16 @@ static int radeon_ttm_tt_populate(struct ttm_tt *ttm)
 static void radeon_ttm_tt_unpopulate(struct ttm_tt *ttm)
 {
 	struct radeon_device *rdev;
-	struct radeon_ttm_tt *gtt = (void *)ttm;
+	struct radeon_ttm_tt *gtt = radeon_ttm_tt_to_gtt(ttm);
 	unsigned i;
 	bool slave = !!(ttm->page_flags & TTM_PAGE_FLAG_SG);
 
+	if (gtt && gtt->userptr) {
+		kfree(ttm->sg);
+		ttm->page_flags &= ~TTM_PAGE_FLAG_SG;
+		return;
+	}
+
 	if (slave)
 		return;
 
@@ -680,6 +809,40 @@ static void radeon_ttm_tt_unpopulate(struct ttm_tt *ttm)
 	ttm_pool_unpopulate(ttm);
 }
 
+int radeon_ttm_tt_set_userptr(struct ttm_tt *ttm, uint64_t addr,
+			      uint32_t flags)
+{
+	struct radeon_ttm_tt *gtt = radeon_ttm_tt_to_gtt(ttm);
+
+	if (gtt == NULL)
+		return -EINVAL;
+
+	gtt->userptr = addr;
+	gtt->usermm = current->mm;
+	gtt->userflags = flags;
+	return 0;
+}
+
+bool radeon_ttm_tt_has_userptr(struct ttm_tt *ttm)
+{
+	struct radeon_ttm_tt *gtt = radeon_ttm_tt_to_gtt(ttm);
+
+	if (gtt == NULL)
+		return false;
+
+	return !!gtt->userptr;
+}
+
+bool radeon_ttm_tt_is_readonly(struct ttm_tt *ttm)
+{
+	struct radeon_ttm_tt *gtt = radeon_ttm_tt_to_gtt(ttm);
+
+	if (gtt == NULL)
+		return false;
+
+	return !!(gtt->userflags & RADEON_GEM_USERPTR_READONLY);
+}
+
 static struct ttm_bo_driver radeon_bo_driver = {
 	.ttm_tt_create = &radeon_ttm_tt_create,
 	.ttm_tt_populate = &radeon_ttm_tt_populate,
@@ -689,11 +852,6 @@ static struct ttm_bo_driver radeon_bo_driver = {
 	.evict_flags = &radeon_evict_flags,
 	.move = &radeon_bo_move,
 	.verify_access = &radeon_verify_access,
-	.sync_obj_signaled = &radeon_sync_obj_signaled,
-	.sync_obj_wait = &radeon_sync_obj_wait,
-	.sync_obj_flush = &radeon_sync_obj_flush,
-	.sync_obj_unref = &radeon_sync_obj_unref,
-	.sync_obj_ref = &radeon_sync_obj_ref,
 	.move_notify = &radeon_bo_move_notify,
 	.fault_reserve_notify = &radeon_bo_fault_reserve_notify,
 	.io_mem_reserve = &radeon_ttm_io_mem_reserve,
@@ -730,7 +888,7 @@ int radeon_ttm_init(struct radeon_device *rdev)
 	radeon_ttm_set_active_vram_size(rdev, rdev->mc.visible_vram_size);
 
 	r = radeon_bo_create(rdev, 256 * 1024, PAGE_SIZE, true,
-			     RADEON_GEM_DOMAIN_VRAM, 0,
+			     RADEON_GEM_DOMAIN_VRAM, 0, NULL,
 			     NULL, &rdev->stollen_vga_memory);
 	if (r) {
 		return r;
@@ -828,7 +986,7 @@ int radeon_mmap(struct file *filp, struct vm_area_struct *vma)
 	int r;
 
 	if (unlikely(vma->vm_pgoff < DRM_FILE_PAGE_OFFSET)) {
-		return drm_mmap(filp, vma);
+		return -EINVAL;
 	}
 
 	file_priv = filp->private_data;
diff --git a/drivers/gpu/drm/radeon/radeon_uvd.c b/drivers/gpu/drm/radeon/radeon_uvd.c
index 341848a14376..c10b2aec6450 100644
--- a/drivers/gpu/drm/radeon/radeon_uvd.c
+++ b/drivers/gpu/drm/radeon/radeon_uvd.c
@@ -40,12 +40,18 @@
 #define UVD_IDLE_TIMEOUT_MS	1000
 
 /* Firmware Names */
+#define FIRMWARE_R600		"radeon/R600_uvd.bin"
+#define FIRMWARE_RS780		"radeon/RS780_uvd.bin"
+#define FIRMWARE_RV770		"radeon/RV770_uvd.bin"
 #define FIRMWARE_RV710		"radeon/RV710_uvd.bin"
 #define FIRMWARE_CYPRESS	"radeon/CYPRESS_uvd.bin"
 #define FIRMWARE_SUMO		"radeon/SUMO_uvd.bin"
 #define FIRMWARE_TAHITI		"radeon/TAHITI_uvd.bin"
 #define FIRMWARE_BONAIRE	"radeon/BONAIRE_uvd.bin"
 
+MODULE_FIRMWARE(FIRMWARE_R600);
+MODULE_FIRMWARE(FIRMWARE_RS780);
+MODULE_FIRMWARE(FIRMWARE_RV770);
 MODULE_FIRMWARE(FIRMWARE_RV710);
 MODULE_FIRMWARE(FIRMWARE_CYPRESS);
 MODULE_FIRMWARE(FIRMWARE_SUMO);
@@ -63,6 +69,23 @@ int radeon_uvd_init(struct radeon_device *rdev)
 	INIT_DELAYED_WORK(&rdev->uvd.idle_work, radeon_uvd_idle_work_handler);
 
 	switch (rdev->family) {
+	case CHIP_RV610:
+	case CHIP_RV630:
+	case CHIP_RV670:
+	case CHIP_RV620:
+	case CHIP_RV635:
+		fw_name = FIRMWARE_R600;
+		break;
+
+	case CHIP_RS780:
+	case CHIP_RS880:
+		fw_name = FIRMWARE_RS780;
+		break;
+
+	case CHIP_RV770:
+		fw_name = FIRMWARE_RV770;
+		break;
+
 	case CHIP_RV710:
 	case CHIP_RV730:
 	case CHIP_RV740:
@@ -115,9 +138,11 @@ int radeon_uvd_init(struct radeon_device *rdev)
 	}
 
 	bo_size = RADEON_GPU_PAGE_ALIGN(rdev->uvd_fw->size + 8) +
-		  RADEON_UVD_STACK_SIZE + RADEON_UVD_HEAP_SIZE;
+		  RADEON_UVD_STACK_SIZE + RADEON_UVD_HEAP_SIZE +
+		  RADEON_GPU_PAGE_SIZE;
 	r = radeon_bo_create(rdev, bo_size, PAGE_SIZE, true,
-			     RADEON_GEM_DOMAIN_VRAM, 0, NULL, &rdev->uvd.vcpu_bo);
+			     RADEON_GEM_DOMAIN_VRAM, 0, NULL,
+			     NULL, &rdev->uvd.vcpu_bo);
 	if (r) {
 		dev_err(rdev->dev, "(%d) failed to allocate UVD bo\n", r);
 		return r;
@@ -231,10 +256,30 @@ int radeon_uvd_resume(struct radeon_device *rdev)
 	return 0;
 }
 
-void radeon_uvd_force_into_uvd_segment(struct radeon_bo *rbo)
+void radeon_uvd_force_into_uvd_segment(struct radeon_bo *rbo,
+				       uint32_t allowed_domains)
 {
-	rbo->placement.fpfn = 0 >> PAGE_SHIFT;
-	rbo->placement.lpfn = (256 * 1024 * 1024) >> PAGE_SHIFT;
+	int i;
+
+	for (i = 0; i < rbo->placement.num_placement; ++i) {
+		rbo->placements[i].fpfn = 0 >> PAGE_SHIFT;
+		rbo->placements[i].lpfn = (256 * 1024 * 1024) >> PAGE_SHIFT;
+	}
+
+	/* If it must be in VRAM it must be in the first segment as well */
+	if (allowed_domains == RADEON_GEM_DOMAIN_VRAM)
+		return;
+
+	/* abort if we already have more than one placement */
+	if (rbo->placement.num_placement > 1)
+		return;
+
+	/* add another 256MB segment */
+	rbo->placements[1] = rbo->placements[0];
+	rbo->placements[1].fpfn += (256 * 1024 * 1024) >> PAGE_SHIFT;
+	rbo->placements[1].lpfn += (256 * 1024 * 1024) >> PAGE_SHIFT;
+	rbo->placement.num_placement++;
+	rbo->placement.num_busy_placement++;
 }
 
 void radeon_uvd_free_handles(struct radeon_device *rdev, struct drm_file *filp)
@@ -356,6 +401,7 @@ static int radeon_uvd_cs_msg(struct radeon_cs_parser *p, struct radeon_bo *bo,
 {
 	int32_t *msg, msg_type, handle;
 	unsigned img_size = 0;
+	struct fence *f;
 	void *ptr;
 
 	int i, r;
@@ -365,8 +411,9 @@ static int radeon_uvd_cs_msg(struct radeon_cs_parser *p, struct radeon_bo *bo,
 		return -EINVAL;
 	}
 
-	if (bo->tbo.sync_obj) {
-		r = radeon_fence_wait(bo->tbo.sync_obj, false);
+	f = reservation_object_get_excl(bo->tbo.resv);
+	if (f) {
+		r = radeon_fence_wait((struct radeon_fence *)f, false);
 		if (r) {
 			DRM_ERROR("Failed waiting for UVD message (%d)!\n", r);
 			return r;
@@ -441,12 +488,12 @@ static int radeon_uvd_cs_reloc(struct radeon_cs_parser *p,
 			       unsigned buf_sizes[], bool *has_msg_cmd)
 {
 	struct radeon_cs_chunk *relocs_chunk;
-	struct radeon_cs_reloc *reloc;
+	struct radeon_bo_list *reloc;
 	unsigned idx, cmd, offset;
 	uint64_t start, end;
 	int r;
 
-	relocs_chunk = &p->chunks[p->chunk_relocs_idx];
+	relocs_chunk = p->chunk_relocs;
 	offset = radeon_get_ib_value(p, data0);
 	idx = radeon_get_ib_value(p, data1);
 	if (idx >= relocs_chunk->length_dw) {
@@ -455,7 +502,7 @@ static int radeon_uvd_cs_reloc(struct radeon_cs_parser *p,
 		return -EINVAL;
 	}
 
-	reloc = p->relocs_ptr[(idx / 4)];
+	reloc = &p->relocs[(idx / 4)];
 	start = reloc->gpu_offset;
 	end = start + radeon_bo_size(reloc->robj);
 	start += offset;
@@ -563,13 +610,13 @@ int radeon_uvd_cs_parse(struct radeon_cs_parser *p)
 		[0x00000003]	=	2048,
 	};
 
-	if (p->chunks[p->chunk_ib_idx].length_dw % 16) {
+	if (p->chunk_ib->length_dw % 16) {
 		DRM_ERROR("UVD IB length (%d) not 16 dwords aligned!\n",
-			  p->chunks[p->chunk_ib_idx].length_dw);
+			  p->chunk_ib->length_dw);
 		return -EINVAL;
 	}
 
-	if (p->chunk_relocs_idx == -1) {
+	if (p->chunk_relocs == NULL) {
 		DRM_ERROR("No relocation chunk !\n");
 		return -EINVAL;
 	}
@@ -593,7 +640,7 @@ int radeon_uvd_cs_parse(struct radeon_cs_parser *p)
 			DRM_ERROR("Unknown packet type %d !\n", pkt.type);
 			return -EINVAL;
 		}
-	} while (p->idx < p->chunks[p->chunk_ib_idx].length_dw);
+	} while (p->idx < p->chunk_ib->length_dw);
 
 	if (!has_msg_cmd) {
 		DRM_ERROR("UVD-IBs need a msg command!\n");
@@ -604,38 +651,16 @@ int radeon_uvd_cs_parse(struct radeon_cs_parser *p)
 }
 
 static int radeon_uvd_send_msg(struct radeon_device *rdev,
-			       int ring, struct radeon_bo *bo,
+			       int ring, uint64_t addr,
 			       struct radeon_fence **fence)
 {
-	struct ttm_validate_buffer tv;
-	struct ww_acquire_ctx ticket;
-	struct list_head head;
 	struct radeon_ib ib;
-	uint64_t addr;
 	int i, r;
 
-	memset(&tv, 0, sizeof(tv));
-	tv.bo = &bo->tbo;
-
-	INIT_LIST_HEAD(&head);
-	list_add(&tv.head, &head);
-
-	r = ttm_eu_reserve_buffers(&ticket, &head);
-	if (r)
-		return r;
-
-	radeon_ttm_placement_from_domain(bo, RADEON_GEM_DOMAIN_VRAM);
-	radeon_uvd_force_into_uvd_segment(bo);
-
-	r = ttm_bo_validate(&bo->tbo, &bo->placement, true, false);
-	if (r) 
-		goto err;
-
 	r = radeon_ib_get(rdev, ring, &ib, NULL, 64);
 	if (r)
-		goto err;
+		return r;
 
-	addr = radeon_bo_gpu_offset(bo);
 	ib.ptr[0] = PACKET0(UVD_GPCOM_VCPU_DATA0, 0);
 	ib.ptr[1] = addr;
 	ib.ptr[2] = PACKET0(UVD_GPCOM_VCPU_DATA1, 0);
@@ -647,19 +672,11 @@ static int radeon_uvd_send_msg(struct radeon_device *rdev,
 	ib.length_dw = 16;
 
 	r = radeon_ib_schedule(rdev, &ib, NULL, false);
-	if (r)
-		goto err;
-	ttm_eu_fence_buffer_objects(&ticket, &head, ib.fence);
 
 	if (fence)
 		*fence = radeon_fence_ref(ib.fence);
 
 	radeon_ib_free(rdev, &ib);
-	radeon_bo_unref(&bo);
-	return 0;
-
-err:
-	ttm_eu_backoff_reservation(&ticket, &head);
 	return r;
 }
 
@@ -669,27 +686,18 @@ err:
 int radeon_uvd_get_create_msg(struct radeon_device *rdev, int ring,
 			      uint32_t handle, struct radeon_fence **fence)
 {
-	struct radeon_bo *bo;
-	uint32_t *msg;
-	int r, i;
+	/* we use the last page of the vcpu bo for the UVD message */
+	uint64_t offs = radeon_bo_size(rdev->uvd.vcpu_bo) -
+		RADEON_GPU_PAGE_SIZE;
 
-	r = radeon_bo_create(rdev, 1024, PAGE_SIZE, true,
-			     RADEON_GEM_DOMAIN_VRAM, 0, NULL, &bo);
-	if (r)
-		return r;
+	uint32_t *msg = rdev->uvd.cpu_addr + offs;
+	uint64_t addr = rdev->uvd.gpu_addr + offs;
 
-	r = radeon_bo_reserve(bo, false);
-	if (r) {
-		radeon_bo_unref(&bo);
-		return r;
-	}
+	int r, i;
 
-	r = radeon_bo_kmap(bo, (void **)&msg);
-	if (r) {
-		radeon_bo_unreserve(bo);
-		radeon_bo_unref(&bo);
+	r = radeon_bo_reserve(rdev->uvd.vcpu_bo, true);
+	if (r)
 		return r;
-	}
 
 	/* stitch together an UVD create msg */
 	msg[0] = cpu_to_le32(0x00000de4);
@@ -706,36 +714,26 @@ int radeon_uvd_get_create_msg(struct radeon_device *rdev, int ring,
 	for (i = 11; i < 1024; ++i)
 		msg[i] = cpu_to_le32(0x0);
 
-	radeon_bo_kunmap(bo);
-	radeon_bo_unreserve(bo);
-
-	return radeon_uvd_send_msg(rdev, ring, bo, fence);
+	r = radeon_uvd_send_msg(rdev, ring, addr, fence);
+	radeon_bo_unreserve(rdev->uvd.vcpu_bo);
+	return r;
 }
 
 int radeon_uvd_get_destroy_msg(struct radeon_device *rdev, int ring,
 			       uint32_t handle, struct radeon_fence **fence)
 {
-	struct radeon_bo *bo;
-	uint32_t *msg;
-	int r, i;
+	/* we use the last page of the vcpu bo for the UVD message */
+	uint64_t offs = radeon_bo_size(rdev->uvd.vcpu_bo) -
+		RADEON_GPU_PAGE_SIZE;
 
-	r = radeon_bo_create(rdev, 1024, PAGE_SIZE, true,
-			     RADEON_GEM_DOMAIN_VRAM, 0, NULL, &bo);
-	if (r)
-		return r;
+	uint32_t *msg = rdev->uvd.cpu_addr + offs;
+	uint64_t addr = rdev->uvd.gpu_addr + offs;
 
-	r = radeon_bo_reserve(bo, false);
-	if (r) {
-		radeon_bo_unref(&bo);
-		return r;
-	}
+	int r, i;
 
-	r = radeon_bo_kmap(bo, (void **)&msg);
-	if (r) {
-		radeon_bo_unreserve(bo);
-		radeon_bo_unref(&bo);
+	r = radeon_bo_reserve(rdev->uvd.vcpu_bo, true);
+	if (r)
 		return r;
-	}
 
 	/* stitch together an UVD destroy msg */
 	msg[0] = cpu_to_le32(0x00000de4);
@@ -745,10 +743,9 @@ int radeon_uvd_get_destroy_msg(struct radeon_device *rdev, int ring,
 	for (i = 4; i < 1024; ++i)
 		msg[i] = cpu_to_le32(0x0);
 
-	radeon_bo_kunmap(bo);
-	radeon_bo_unreserve(bo);
-
-	return radeon_uvd_send_msg(rdev, ring, bo, fence);
+	r = radeon_uvd_send_msg(rdev, ring, addr, fence);
+	radeon_bo_unreserve(rdev->uvd.vcpu_bo);
+	return r;
 }
 
 /**
diff --git a/drivers/gpu/drm/radeon/radeon_vce.c b/drivers/gpu/drm/radeon/radeon_vce.c
index c7190aadbd89..976fe432f4e2 100644
--- a/drivers/gpu/drm/radeon/radeon_vce.c
+++ b/drivers/gpu/drm/radeon/radeon_vce.c
@@ -126,7 +126,8 @@ int radeon_vce_init(struct radeon_device *rdev)
 	size = RADEON_GPU_PAGE_ALIGN(rdev->vce_fw->size) +
 	       RADEON_VCE_STACK_SIZE + RADEON_VCE_HEAP_SIZE;
 	r = radeon_bo_create(rdev, size, PAGE_SIZE, true,
-			     RADEON_GEM_DOMAIN_VRAM, 0, NULL, &rdev->vce.vcpu_bo);
+			     RADEON_GEM_DOMAIN_VRAM, 0, NULL, NULL,
+			     &rdev->vce.vcpu_bo);
 	if (r) {
 		dev_err(rdev->dev, "(%d) failed to allocate VCE bo\n", r);
 		return r;
@@ -452,11 +453,11 @@ int radeon_vce_cs_reloc(struct radeon_cs_parser *p, int lo, int hi,
 			unsigned size)
 {
 	struct radeon_cs_chunk *relocs_chunk;
-	struct radeon_cs_reloc *reloc;
+	struct radeon_bo_list *reloc;
 	uint64_t start, end, offset;
 	unsigned idx;
 
-	relocs_chunk = &p->chunks[p->chunk_relocs_idx];
+	relocs_chunk = p->chunk_relocs;
 	offset = radeon_get_ib_value(p, lo);
 	idx = radeon_get_ib_value(p, hi);
 
@@ -466,7 +467,7 @@ int radeon_vce_cs_reloc(struct radeon_cs_parser *p, int lo, int hi,
 		return -EINVAL;
 	}
 
-	reloc = p->relocs_ptr[(idx / 4)];
+	reloc = &p->relocs[(idx / 4)];
 	start = reloc->gpu_offset;
 	end = start + radeon_bo_size(reloc->robj);
 	start += offset;
@@ -533,7 +534,7 @@ int radeon_vce_cs_parse(struct radeon_cs_parser *p)
 	uint32_t *size = &tmp;
 	int i, r;
 
-	while (p->idx < p->chunks[p->chunk_ib_idx].length_dw) {
+	while (p->idx < p->chunk_ib->length_dw) {
 		uint32_t len = radeon_get_ib_value(p, p->idx);
 		uint32_t cmd = radeon_get_ib_value(p, p->idx + 1);
 
diff --git a/drivers/gpu/drm/radeon/radeon_vm.c b/drivers/gpu/drm/radeon/radeon_vm.c
index 088ffdc2f577..cde48c42b30a 100644
--- a/drivers/gpu/drm/radeon/radeon_vm.c
+++ b/drivers/gpu/drm/radeon/radeon_vm.c
@@ -125,39 +125,37 @@ void radeon_vm_manager_fini(struct radeon_device *rdev)
  * Add the page directory to the list of BOs to
  * validate for command submission (cayman+).
  */
-struct radeon_cs_reloc *radeon_vm_get_bos(struct radeon_device *rdev,
+struct radeon_bo_list *radeon_vm_get_bos(struct radeon_device *rdev,
 					  struct radeon_vm *vm,
 					  struct list_head *head)
 {
-	struct radeon_cs_reloc *list;
+	struct radeon_bo_list *list;
 	unsigned i, idx;
 
-	list = kmalloc_array(vm->max_pde_used + 2,
-			     sizeof(struct radeon_cs_reloc), GFP_KERNEL);
+	list = drm_malloc_ab(vm->max_pde_used + 2,
+			     sizeof(struct radeon_bo_list));
 	if (!list)
 		return NULL;
 
 	/* add the vm page table to the list */
-	list[0].gobj = NULL;
 	list[0].robj = vm->page_directory;
 	list[0].prefered_domains = RADEON_GEM_DOMAIN_VRAM;
 	list[0].allowed_domains = RADEON_GEM_DOMAIN_VRAM;
 	list[0].tv.bo = &vm->page_directory->tbo;
+	list[0].tv.shared = true;
 	list[0].tiling_flags = 0;
-	list[0].handle = 0;
 	list_add(&list[0].tv.head, head);
 
 	for (i = 0, idx = 1; i <= vm->max_pde_used; i++) {
 		if (!vm->page_tables[i].bo)
 			continue;
 
-		list[idx].gobj = NULL;
 		list[idx].robj = vm->page_tables[i].bo;
 		list[idx].prefered_domains = RADEON_GEM_DOMAIN_VRAM;
 		list[idx].allowed_domains = RADEON_GEM_DOMAIN_VRAM;
 		list[idx].tv.bo = &list[idx].robj->tbo;
+		list[idx].tv.shared = true;
 		list[idx].tiling_flags = 0;
-		list[idx].handle = 0;
 		list_add(&list[idx++].tv.head, head);
 	}
 
@@ -180,15 +178,18 @@ struct radeon_fence *radeon_vm_grab_id(struct radeon_device *rdev,
 				       struct radeon_vm *vm, int ring)
 {
 	struct radeon_fence *best[RADEON_NUM_RINGS] = {};
+	struct radeon_vm_id *vm_id = &vm->ids[ring];
+
 	unsigned choices[2] = {};
 	unsigned i;
 
 	/* check if the id is still valid */
-	if (vm->last_id_use && vm->last_id_use == rdev->vm_manager.active[vm->id])
+	if (vm_id->id && vm_id->last_id_use &&
+	    vm_id->last_id_use == rdev->vm_manager.active[vm_id->id])
 		return NULL;
 
 	/* we definately need to flush */
-	radeon_fence_unref(&vm->last_flush);
+	vm_id->pd_gpu_addr = ~0ll;
 
 	/* skip over VMID 0, since it is the system VM */
 	for (i = 1; i < rdev->vm_manager.nvm; ++i) {
@@ -196,8 +197,8 @@ struct radeon_fence *radeon_vm_grab_id(struct radeon_device *rdev,
 
 		if (fence == NULL) {
 			/* found a free one */
-			vm->id = i;
-			trace_radeon_vm_grab_id(vm->id, ring);
+			vm_id->id = i;
+			trace_radeon_vm_grab_id(i, ring);
 			return NULL;
 		}
 
@@ -209,8 +210,8 @@ struct radeon_fence *radeon_vm_grab_id(struct radeon_device *rdev,
 
 	for (i = 0; i < 2; ++i) {
 		if (choices[i]) {
-			vm->id = choices[i];
-			trace_radeon_vm_grab_id(vm->id, ring);
+			vm_id->id = choices[i];
+			trace_radeon_vm_grab_id(choices[i], ring);
 			return rdev->vm_manager.active[choices[i]];
 		}
 	}
@@ -226,6 +227,7 @@ struct radeon_fence *radeon_vm_grab_id(struct radeon_device *rdev,
  * @rdev: radeon_device pointer
  * @vm: vm we want to flush
  * @ring: ring to use for flush
+ * @updates: last vm update that is waited for
  *
  * Flush the vm (cayman+).
  *
@@ -233,15 +235,21 @@ struct radeon_fence *radeon_vm_grab_id(struct radeon_device *rdev,
  */
 void radeon_vm_flush(struct radeon_device *rdev,
 		     struct radeon_vm *vm,
-		     int ring)
+		     int ring, struct radeon_fence *updates)
 {
 	uint64_t pd_addr = radeon_bo_gpu_offset(vm->page_directory);
+	struct radeon_vm_id *vm_id = &vm->ids[ring];
+
+	if (pd_addr != vm_id->pd_gpu_addr || !vm_id->flushed_updates ||
+	    radeon_fence_is_earlier(vm_id->flushed_updates, updates)) {
+
+		trace_radeon_vm_flush(pd_addr, ring, vm->ids[ring].id);
+		radeon_fence_unref(&vm_id->flushed_updates);
+		vm_id->flushed_updates = radeon_fence_ref(updates);
+		vm_id->pd_gpu_addr = pd_addr;
+		radeon_ring_vm_flush(rdev, &rdev->ring[ring],
+				     vm_id->id, vm_id->pd_gpu_addr);
 
-	/* if we can't remember our last VM flush then flush now! */
-	if (!vm->last_flush || pd_addr != vm->pd_gpu_addr) {
-		trace_radeon_vm_flush(pd_addr, ring, vm->id);
-		vm->pd_gpu_addr = pd_addr;
-		radeon_ring_vm_flush(rdev, ring, vm);
 	}
 }
 
@@ -261,18 +269,13 @@ void radeon_vm_fence(struct radeon_device *rdev,
 		     struct radeon_vm *vm,
 		     struct radeon_fence *fence)
 {
-	radeon_fence_unref(&vm->fence);
-	vm->fence = radeon_fence_ref(fence);
-
-	radeon_fence_unref(&rdev->vm_manager.active[vm->id]);
-	rdev->vm_manager.active[vm->id] = radeon_fence_ref(fence);
+	unsigned vm_id = vm->ids[fence->ring].id;
 
-	radeon_fence_unref(&vm->last_id_use);
-	vm->last_id_use = radeon_fence_ref(fence);
+	radeon_fence_unref(&rdev->vm_manager.active[vm_id]);
+	rdev->vm_manager.active[vm_id] = radeon_fence_ref(fence);
 
-        /* we just flushed the VM, remember that */
-        if (!vm->last_flush)
-                vm->last_flush = radeon_fence_ref(fence);
+	radeon_fence_unref(&vm->ids[fence->ring].last_id_use);
+	vm->ids[fence->ring].last_id_use = radeon_fence_ref(fence);
 }
 
 /**
@@ -385,34 +388,25 @@ static void radeon_vm_set_pages(struct radeon_device *rdev,
 static int radeon_vm_clear_bo(struct radeon_device *rdev,
 			      struct radeon_bo *bo)
 {
-        struct ttm_validate_buffer tv;
-        struct ww_acquire_ctx ticket;
-        struct list_head head;
 	struct radeon_ib ib;
 	unsigned entries;
 	uint64_t addr;
 	int r;
 
-        memset(&tv, 0, sizeof(tv));
-        tv.bo = &bo->tbo;
-
-        INIT_LIST_HEAD(&head);
-        list_add(&tv.head, &head);
-
-        r = ttm_eu_reserve_buffers(&ticket, &head);
-        if (r)
+	r = radeon_bo_reserve(bo, false);
+	if (r)
 		return r;
 
-        r = ttm_bo_validate(&bo->tbo, &bo->placement, true, false);
-        if (r)
-                goto error;
+	r = ttm_bo_validate(&bo->tbo, &bo->placement, true, false);
+	if (r)
+		goto error_unreserve;
 
 	addr = radeon_bo_gpu_offset(bo);
 	entries = radeon_bo_size(bo) / 8;
 
 	r = radeon_ib_get(rdev, R600_RING_TYPE_DMA_INDEX, &ib, NULL, 256);
 	if (r)
-                goto error;
+		goto error_unreserve;
 
 	ib.length_dw = 0;
 
@@ -422,15 +416,16 @@ static int radeon_vm_clear_bo(struct radeon_device *rdev,
 
 	r = radeon_ib_schedule(rdev, &ib, NULL, false);
 	if (r)
-                goto error;
+		goto error_free;
 
-	ttm_eu_fence_buffer_objects(&ticket, &head, ib.fence);
-	radeon_ib_free(rdev, &ib);
+	ib.fence->is_vm_update = true;
+	radeon_bo_fence(bo, ib.fence, false);
 
-	return 0;
+error_free:
+	radeon_ib_free(rdev, &ib);
 
-error:
-	ttm_eu_backoff_reservation(&ticket, &head);
+error_unreserve:
+	radeon_bo_unreserve(bo);
 	return r;
 }
 
@@ -446,7 +441,7 @@ error:
  * Validate and set the offset requested within the vm address space.
  * Returns 0 for success, error for failure.
  *
- * Object has to be reserved!
+ * Object has to be reserved and gets unreserved by this function!
  */
 int radeon_vm_bo_set_addr(struct radeon_device *rdev,
 			  struct radeon_bo_va *bo_va,
@@ -492,7 +487,9 @@ int radeon_vm_bo_set_addr(struct radeon_device *rdev,
 			tmp->vm = vm;
 			tmp->addr = bo_va->addr;
 			tmp->bo = radeon_bo_ref(bo_va->bo);
+			spin_lock(&vm->status_lock);
 			list_add(&tmp->vm_status, &vm->freed);
+			spin_unlock(&vm->status_lock);
 		}
 
 		interval_tree_remove(&bo_va->it, &vm->va);
@@ -545,7 +542,8 @@ int radeon_vm_bo_set_addr(struct radeon_device *rdev,
 
 		r = radeon_bo_create(rdev, RADEON_VM_PTE_COUNT * 8,
 				     RADEON_GPU_PAGE_SIZE, true,
-				     RADEON_GEM_DOMAIN_VRAM, 0, NULL, &pt);
+				     RADEON_GEM_DOMAIN_VRAM, 0,
+				     NULL, NULL, &pt);
 		if (r)
 			return r;
 
@@ -571,7 +569,7 @@ int radeon_vm_bo_set_addr(struct radeon_device *rdev,
 	}
 
 	mutex_unlock(&vm->mutex);
-	return radeon_bo_reserve(bo_va->bo, false);
+	return 0;
 }
 
 /**
@@ -694,17 +692,16 @@ int radeon_vm_update_page_directory(struct radeon_device *rdev,
 
 	if (ib.length_dw != 0) {
 		radeon_asic_vm_pad_ib(rdev, &ib);
-		radeon_semaphore_sync_to(ib.semaphore, pd->tbo.sync_obj);
-		radeon_semaphore_sync_to(ib.semaphore, vm->last_id_use);
+
+		radeon_sync_resv(rdev, &ib.sync, pd->tbo.resv, true);
 		WARN_ON(ib.length_dw > ndw);
 		r = radeon_ib_schedule(rdev, &ib, NULL, false);
 		if (r) {
 			radeon_ib_free(rdev, &ib);
 			return r;
 		}
-		radeon_fence_unref(&vm->fence);
-		vm->fence = radeon_fence_ref(ib.fence);
-		radeon_fence_unref(&vm->last_flush);
+		ib.fence->is_vm_update = true;
+		radeon_bo_fence(pd, ib.fence, false);
 	}
 	radeon_ib_free(rdev, &ib);
 
@@ -803,11 +800,11 @@ static void radeon_vm_frag_ptes(struct radeon_device *rdev,
  *
  * Global and local mutex must be locked!
  */
-static void radeon_vm_update_ptes(struct radeon_device *rdev,
-				  struct radeon_vm *vm,
-				  struct radeon_ib *ib,
-				  uint64_t start, uint64_t end,
-				  uint64_t dst, uint32_t flags)
+static int radeon_vm_update_ptes(struct radeon_device *rdev,
+				 struct radeon_vm *vm,
+				 struct radeon_ib *ib,
+				 uint64_t start, uint64_t end,
+				 uint64_t dst, uint32_t flags)
 {
 	uint64_t mask = RADEON_VM_PTE_COUNT - 1;
 	uint64_t last_pte = ~0, last_dst = ~0;
@@ -820,8 +817,12 @@ static void radeon_vm_update_ptes(struct radeon_device *rdev,
 		struct radeon_bo *pt = vm->page_tables[pt_idx].bo;
 		unsigned nptes;
 		uint64_t pte;
+		int r;
 
-		radeon_semaphore_sync_to(ib->semaphore, pt->tbo.sync_obj);
+		radeon_sync_resv(rdev, &ib->sync, pt->tbo.resv, true);
+		r = reservation_object_reserve_shared(pt->tbo.resv);
+		if (r)
+			return r;
 
 		if ((addr & ~mask) == (end & ~mask))
 			nptes = end - addr;
@@ -855,6 +856,33 @@ static void radeon_vm_update_ptes(struct radeon_device *rdev,
 				    last_pte + 8 * count,
 				    last_dst, flags);
 	}
+
+	return 0;
+}
+
+/**
+ * radeon_vm_fence_pts - fence page tables after an update
+ *
+ * @vm: requested vm
+ * @start: start of GPU address range
+ * @end: end of GPU address range
+ * @fence: fence to use
+ *
+ * Fence the page tables in the range @start - @end (cayman+).
+ *
+ * Global and local mutex must be locked!
+ */
+static void radeon_vm_fence_pts(struct radeon_vm *vm,
+				uint64_t start, uint64_t end,
+				struct radeon_fence *fence)
+{
+	unsigned i;
+
+	start >>= radeon_vm_block_size;
+	end >>= radeon_vm_block_size;
+
+	for (i = start; i <= end; ++i)
+		radeon_bo_fence(vm->page_tables[i].bo, fence, true);
 }
 
 /**
@@ -887,11 +915,16 @@ int radeon_vm_bo_update(struct radeon_device *rdev,
 		return -EINVAL;
 	}
 
+	spin_lock(&vm->status_lock);
 	list_del_init(&bo_va->vm_status);
+	spin_unlock(&vm->status_lock);
 
 	bo_va->flags &= ~RADEON_VM_PAGE_VALID;
 	bo_va->flags &= ~RADEON_VM_PAGE_SYSTEM;
 	bo_va->flags &= ~RADEON_VM_PAGE_SNOOPED;
+	if (bo_va->bo && radeon_ttm_tt_is_readonly(bo_va->bo->tbo.ttm))
+		bo_va->flags &= ~RADEON_VM_PAGE_WRITEABLE;
+
 	if (mem) {
 		addr = mem->start << PAGE_SHIFT;
 		if (mem->mem_type != TTM_PL_SYSTEM) {
@@ -953,23 +986,34 @@ int radeon_vm_bo_update(struct radeon_device *rdev,
 		return r;
 	ib.length_dw = 0;
 
-	radeon_vm_update_ptes(rdev, vm, &ib, bo_va->it.start,
-			      bo_va->it.last + 1, addr,
-			      radeon_vm_page_flags(bo_va->flags));
+	if (!(bo_va->flags & RADEON_VM_PAGE_VALID)) {
+		unsigned i;
+
+		for (i = 0; i < RADEON_NUM_RINGS; ++i)
+			radeon_sync_fence(&ib.sync, vm->ids[i].last_id_use);
+	}
+
+	r = radeon_vm_update_ptes(rdev, vm, &ib, bo_va->it.start,
+				  bo_va->it.last + 1, addr,
+				  radeon_vm_page_flags(bo_va->flags));
+	if (r) {
+		radeon_ib_free(rdev, &ib);
+		return r;
+	}
 
 	radeon_asic_vm_pad_ib(rdev, &ib);
 	WARN_ON(ib.length_dw > ndw);
 
-	radeon_semaphore_sync_to(ib.semaphore, vm->fence);
 	r = radeon_ib_schedule(rdev, &ib, NULL, false);
 	if (r) {
 		radeon_ib_free(rdev, &ib);
 		return r;
 	}
-	radeon_fence_unref(&vm->fence);
-	vm->fence = radeon_fence_ref(ib.fence);
+	ib.fence->is_vm_update = true;
+	radeon_vm_fence_pts(vm, bo_va->it.start, bo_va->it.last + 1, ib.fence);
+	radeon_fence_unref(&bo_va->last_pt_update);
+	bo_va->last_pt_update = radeon_fence_ref(ib.fence);
 	radeon_ib_free(rdev, &ib);
-	radeon_fence_unref(&vm->last_flush);
 
 	return 0;
 }
@@ -988,16 +1032,25 @@ int radeon_vm_bo_update(struct radeon_device *rdev,
 int radeon_vm_clear_freed(struct radeon_device *rdev,
 			  struct radeon_vm *vm)
 {
-	struct radeon_bo_va *bo_va, *tmp;
+	struct radeon_bo_va *bo_va;
 	int r;
 
-	list_for_each_entry_safe(bo_va, tmp, &vm->freed, vm_status) {
+	spin_lock(&vm->status_lock);
+	while (!list_empty(&vm->freed)) {
+		bo_va = list_first_entry(&vm->freed,
+			struct radeon_bo_va, vm_status);
+		spin_unlock(&vm->status_lock);
+
 		r = radeon_vm_bo_update(rdev, bo_va, NULL);
 		radeon_bo_unref(&bo_va->bo);
+		radeon_fence_unref(&bo_va->last_pt_update);
 		kfree(bo_va);
 		if (r)
 			return r;
+
+		spin_lock(&vm->status_lock);
 	}
+	spin_unlock(&vm->status_lock);
 	return 0;
 
 }
@@ -1016,14 +1069,23 @@ int radeon_vm_clear_freed(struct radeon_device *rdev,
 int radeon_vm_clear_invalids(struct radeon_device *rdev,
 			     struct radeon_vm *vm)
 {
-	struct radeon_bo_va *bo_va, *tmp;
+	struct radeon_bo_va *bo_va;
 	int r;
 
-	list_for_each_entry_safe(bo_va, tmp, &vm->invalidated, vm_status) {
+	spin_lock(&vm->status_lock);
+	while (!list_empty(&vm->invalidated)) {
+		bo_va = list_first_entry(&vm->invalidated,
+			struct radeon_bo_va, vm_status);
+		spin_unlock(&vm->status_lock);
+
 		r = radeon_vm_bo_update(rdev, bo_va, NULL);
 		if (r)
 			return r;
+
+		spin_lock(&vm->status_lock);
 	}
+	spin_unlock(&vm->status_lock);
+
 	return 0;
 }
 
@@ -1046,14 +1108,17 @@ void radeon_vm_bo_rmv(struct radeon_device *rdev,
 
 	mutex_lock(&vm->mutex);
 	interval_tree_remove(&bo_va->it, &vm->va);
+	spin_lock(&vm->status_lock);
 	list_del(&bo_va->vm_status);
 
 	if (bo_va->addr) {
 		bo_va->bo = radeon_bo_ref(bo_va->bo);
 		list_add(&bo_va->vm_status, &vm->freed);
 	} else {
+		radeon_fence_unref(&bo_va->last_pt_update);
 		kfree(bo_va);
 	}
+	spin_unlock(&vm->status_lock);
 
 	mutex_unlock(&vm->mutex);
 }
@@ -1074,10 +1139,10 @@ void radeon_vm_bo_invalidate(struct radeon_device *rdev,
 
 	list_for_each_entry(bo_va, &bo->va, bo_list) {
 		if (bo_va->addr) {
-			mutex_lock(&bo_va->vm->mutex);
+			spin_lock(&bo_va->vm->status_lock);
 			list_del(&bo_va->vm_status);
 			list_add(&bo_va->vm_status, &bo_va->vm->invalidated);
-			mutex_unlock(&bo_va->vm->mutex);
+			spin_unlock(&bo_va->vm->status_lock);
 		}
 	}
 }
@@ -1095,15 +1160,17 @@ int radeon_vm_init(struct radeon_device *rdev, struct radeon_vm *vm)
 	const unsigned align = min(RADEON_VM_PTB_ALIGN_SIZE,
 		RADEON_VM_PTE_COUNT * 8);
 	unsigned pd_size, pd_entries, pts_size;
-	int r;
+	int i, r;
 
-	vm->id = 0;
 	vm->ib_bo_va = NULL;
-	vm->fence = NULL;
-	vm->last_flush = NULL;
-	vm->last_id_use = NULL;
+	for (i = 0; i < RADEON_NUM_RINGS; ++i) {
+		vm->ids[i].id = 0;
+		vm->ids[i].flushed_updates = NULL;
+		vm->ids[i].last_id_use = NULL;
+	}
 	mutex_init(&vm->mutex);
 	vm->va = RB_ROOT;
+	spin_lock_init(&vm->status_lock);
 	INIT_LIST_HEAD(&vm->invalidated);
 	INIT_LIST_HEAD(&vm->freed);
 
@@ -1120,7 +1187,7 @@ int radeon_vm_init(struct radeon_device *rdev, struct radeon_vm *vm)
 
 	r = radeon_bo_create(rdev, pd_size, align, true,
 			     RADEON_GEM_DOMAIN_VRAM, 0, NULL,
-			     &vm->page_directory);
+			     NULL, &vm->page_directory);
 	if (r)
 		return r;
 
@@ -1157,11 +1224,13 @@ void radeon_vm_fini(struct radeon_device *rdev, struct radeon_vm *vm)
 		if (!r) {
 			list_del_init(&bo_va->bo_list);
 			radeon_bo_unreserve(bo_va->bo);
+			radeon_fence_unref(&bo_va->last_pt_update);
 			kfree(bo_va);
 		}
 	}
 	list_for_each_entry_safe(bo_va, tmp, &vm->freed, vm_status) {
 		radeon_bo_unref(&bo_va->bo);
+		radeon_fence_unref(&bo_va->last_pt_update);
 		kfree(bo_va);
 	}
 
@@ -1171,9 +1240,10 @@ void radeon_vm_fini(struct radeon_device *rdev, struct radeon_vm *vm)
 
 	radeon_bo_unref(&vm->page_directory);
 
-	radeon_fence_unref(&vm->fence);
-	radeon_fence_unref(&vm->last_flush);
-	radeon_fence_unref(&vm->last_id_use);
+	for (i = 0; i < RADEON_NUM_RINGS; ++i) {
+		radeon_fence_unref(&vm->ids[i].flushed_updates);
+		radeon_fence_unref(&vm->ids[i].last_id_use);
+	}
 
 	mutex_destroy(&vm->mutex);
 }
diff --git a/drivers/gpu/drm/radeon/rs600.c b/drivers/gpu/drm/radeon/rs600.c
index 5f6db4629aaa..9acb1c3c005b 100644
--- a/drivers/gpu/drm/radeon/rs600.c
+++ b/drivers/gpu/drm/radeon/rs600.c
@@ -879,6 +879,9 @@ void rs600_bandwidth_update(struct radeon_device *rdev)
 	u32 d1mode_priority_a_cnt, d2mode_priority_a_cnt;
 	/* FIXME: implement full support */
 
+	if (!rdev->mode_info.mode_config_initialized)
+		return;
+
 	radeon_update_display_priority(rdev);
 
 	if (rdev->mode_info.crtcs[0]->base.enabled)
diff --git a/drivers/gpu/drm/radeon/rs690.c b/drivers/gpu/drm/radeon/rs690.c
index 3462b64369bf..0a2d36e81108 100644
--- a/drivers/gpu/drm/radeon/rs690.c
+++ b/drivers/gpu/drm/radeon/rs690.c
@@ -579,6 +579,9 @@ void rs690_bandwidth_update(struct radeon_device *rdev)
 	u32 d1mode_priority_a_cnt, d1mode_priority_b_cnt;
 	u32 d2mode_priority_a_cnt, d2mode_priority_b_cnt;
 
+	if (!rdev->mode_info.mode_config_initialized)
+		return;
+
 	radeon_update_display_priority(rdev);
 
 	if (rdev->mode_info.crtcs[0]->base.enabled)
diff --git a/drivers/gpu/drm/radeon/rs780_dpm.c b/drivers/gpu/drm/radeon/rs780_dpm.c
index 02f7710de470..9031f4b69824 100644
--- a/drivers/gpu/drm/radeon/rs780_dpm.c
+++ b/drivers/gpu/drm/radeon/rs780_dpm.c
@@ -24,6 +24,7 @@
 
 #include "drmP.h"
 #include "radeon.h"
+#include "radeon_asic.h"
 #include "rs780d.h"
 #include "r600_dpm.h"
 #include "rs780_dpm.h"
diff --git a/drivers/gpu/drm/radeon/rv515.c b/drivers/gpu/drm/radeon/rv515.c
index 8a477bf1fdb3..c55d653aaf5f 100644
--- a/drivers/gpu/drm/radeon/rv515.c
+++ b/drivers/gpu/drm/radeon/rv515.c
@@ -1277,6 +1277,9 @@ void rv515_bandwidth_update(struct radeon_device *rdev)
 	struct drm_display_mode *mode0 = NULL;
 	struct drm_display_mode *mode1 = NULL;
 
+	if (!rdev->mode_info.mode_config_initialized)
+		return;
+
 	radeon_update_display_priority(rdev);
 
 	if (rdev->mode_info.crtcs[0]->base.enabled)
diff --git a/drivers/gpu/drm/radeon/rv6xx_dpm.c b/drivers/gpu/drm/radeon/rv6xx_dpm.c
index e7045b085715..6a5c233361e9 100644
--- a/drivers/gpu/drm/radeon/rv6xx_dpm.c
+++ b/drivers/gpu/drm/radeon/rv6xx_dpm.c
@@ -24,6 +24,7 @@
 
 #include "drmP.h"
 #include "radeon.h"
+#include "radeon_asic.h"
 #include "rv6xxd.h"
 #include "r600_dpm.h"
 #include "rv6xx_dpm.h"
diff --git a/drivers/gpu/drm/radeon/rv770.c b/drivers/gpu/drm/radeon/rv770.c
index d9f5ce715c9b..372016e266d0 100644
--- a/drivers/gpu/drm/radeon/rv770.c
+++ b/drivers/gpu/drm/radeon/rv770.c
@@ -26,7 +26,6 @@
  *          Jerome Glisse
  */
 #include <linux/firmware.h>
-#include <linux/platform_device.h>
 #include <linux/slab.h>
 #include <drm/drmP.h>
 #include "radeon.h"
diff --git a/drivers/gpu/drm/radeon/rv770_dma.c b/drivers/gpu/drm/radeon/rv770_dma.c
index 74426ac2bb5c..acff6e09cc40 100644
--- a/drivers/gpu/drm/radeon/rv770_dma.c
+++ b/drivers/gpu/drm/radeon/rv770_dma.c
@@ -33,41 +33,38 @@
  * @src_offset: src GPU address
  * @dst_offset: dst GPU address
  * @num_gpu_pages: number of GPU pages to xfer
- * @fence: radeon fence object
+ * @resv: reservation object to sync to
  *
  * Copy GPU paging using the DMA engine (r7xx).
  * Used by the radeon ttm implementation to move pages if
  * registered as the asic copy callback.
  */
-int rv770_copy_dma(struct radeon_device *rdev,
-		  uint64_t src_offset, uint64_t dst_offset,
-		  unsigned num_gpu_pages,
-		  struct radeon_fence **fence)
+struct radeon_fence *rv770_copy_dma(struct radeon_device *rdev,
+				    uint64_t src_offset, uint64_t dst_offset,
+				    unsigned num_gpu_pages,
+				    struct reservation_object *resv)
 {
-	struct radeon_semaphore *sem = NULL;
+	struct radeon_fence *fence;
+	struct radeon_sync sync;
 	int ring_index = rdev->asic->copy.dma_ring_index;
 	struct radeon_ring *ring = &rdev->ring[ring_index];
 	u32 size_in_dw, cur_size_in_dw;
 	int i, num_loops;
 	int r = 0;
 
-	r = radeon_semaphore_create(rdev, &sem);
-	if (r) {
-		DRM_ERROR("radeon: moving bo (%d).\n", r);
-		return r;
-	}
+	radeon_sync_create(&sync);
 
 	size_in_dw = (num_gpu_pages << RADEON_GPU_PAGE_SHIFT) / 4;
 	num_loops = DIV_ROUND_UP(size_in_dw, 0xFFFF);
 	r = radeon_ring_lock(rdev, ring, num_loops * 5 + 8);
 	if (r) {
 		DRM_ERROR("radeon: moving bo (%d).\n", r);
-		radeon_semaphore_free(rdev, &sem, NULL);
-		return r;
+		radeon_sync_free(rdev, &sync, NULL);
+		return ERR_PTR(r);
 	}
 
-	radeon_semaphore_sync_to(sem, *fence);
-	radeon_semaphore_sync_rings(rdev, sem, ring->idx);
+	radeon_sync_resv(rdev, &sync, resv, false);
+	radeon_sync_rings(rdev, &sync, ring->idx);
 
 	for (i = 0; i < num_loops; i++) {
 		cur_size_in_dw = size_in_dw;
@@ -83,15 +80,15 @@ int rv770_copy_dma(struct radeon_device *rdev,
 		dst_offset += cur_size_in_dw * 4;
 	}
 
-	r = radeon_fence_emit(rdev, fence, ring->idx);
+	r = radeon_fence_emit(rdev, &fence, ring->idx);
 	if (r) {
 		radeon_ring_unlock_undo(rdev, ring);
-		radeon_semaphore_free(rdev, &sem, NULL);
-		return r;
+		radeon_sync_free(rdev, &sync, NULL);
+		return ERR_PTR(r);
 	}
 
 	radeon_ring_unlock_commit(rdev, ring, false);
-	radeon_semaphore_free(rdev, &sem, *fence);
+	radeon_sync_free(rdev, &sync, fence);
 
-	return r;
+	return fence;
 }
diff --git a/drivers/gpu/drm/radeon/rv770_dpm.c b/drivers/gpu/drm/radeon/rv770_dpm.c
index 3c76e1dcdf04..755a8f96fe46 100644
--- a/drivers/gpu/drm/radeon/rv770_dpm.c
+++ b/drivers/gpu/drm/radeon/rv770_dpm.c
@@ -24,6 +24,7 @@
 
 #include "drmP.h"
 #include "radeon.h"
+#include "radeon_asic.h"
 #include "rv770d.h"
 #include "r600_dpm.h"
 #include "rv770_dpm.h"
diff --git a/drivers/gpu/drm/radeon/si.c b/drivers/gpu/drm/radeon/si.c
index 3a0b973e8a96..60df444bd075 100644
--- a/drivers/gpu/drm/radeon/si.c
+++ b/drivers/gpu/drm/radeon/si.c
@@ -2384,6 +2384,9 @@ void dce6_bandwidth_update(struct radeon_device *rdev)
 	u32 num_heads = 0, lb_size;
 	int i;
 
+	if (!rdev->mode_info.mode_config_initialized)
+		return;
+
 	radeon_update_display_priority(rdev);
 
 	for (i = 0; i < rdev->num_crtc; i++) {
@@ -3362,6 +3365,7 @@ void si_fence_ring_emit(struct radeon_device *rdev,
 void si_ring_ib_execute(struct radeon_device *rdev, struct radeon_ib *ib)
 {
 	struct radeon_ring *ring = &rdev->ring[ib->ring];
+	unsigned vm_id = ib->vm ? ib->vm->ids[ib->ring].id : 0;
 	u32 header;
 
 	if (ib->is_const_ib) {
@@ -3397,14 +3401,13 @@ void si_ring_ib_execute(struct radeon_device *rdev, struct radeon_ib *ib)
 #endif
 			  (ib->gpu_addr & 0xFFFFFFFC));
 	radeon_ring_write(ring, upper_32_bits(ib->gpu_addr) & 0xFFFF);
-	radeon_ring_write(ring, ib->length_dw |
-			  (ib->vm ? (ib->vm->id << 24) : 0));
+	radeon_ring_write(ring, ib->length_dw | (vm_id << 24));
 
 	if (!ib->is_const_ib) {
 		/* flush read cache over gart for this vmid */
 		radeon_ring_write(ring, PACKET3(PACKET3_SET_CONFIG_REG, 1));
 		radeon_ring_write(ring, (CP_COHER_CNTL2 - PACKET3_SET_CONFIG_REG_START) >> 2);
-		radeon_ring_write(ring, ib->vm ? ib->vm->id : 0);
+		radeon_ring_write(ring, vm_id);
 		radeon_ring_write(ring, PACKET3(PACKET3_SURFACE_SYNC, 3));
 		radeon_ring_write(ring, PACKET3_TCL1_ACTION_ENA |
 				  PACKET3_TC_ACTION_ENA |
@@ -4684,7 +4687,7 @@ static int si_vm_packet3_compute_check(struct radeon_device *rdev,
 int si_ib_parse(struct radeon_device *rdev, struct radeon_ib *ib)
 {
 	int ret = 0;
-	u32 idx = 0;
+	u32 idx = 0, i;
 	struct radeon_cs_packet pkt;
 
 	do {
@@ -4695,6 +4698,12 @@ int si_ib_parse(struct radeon_device *rdev, struct radeon_ib *ib)
 		switch (pkt.type) {
 		case RADEON_PACKET_TYPE0:
 			dev_err(rdev->dev, "Packet0 not allowed!\n");
+			for (i = 0; i < ib->length_dw; i++) {
+				if (i == idx)
+					printk("\t0x%08x <---\n", ib->ptr[i]);
+				else
+					printk("\t0x%08x\n", ib->ptr[i]);
+			}
 			ret = -EINVAL;
 			break;
 		case RADEON_PACKET_TYPE2:
@@ -5014,27 +5023,23 @@ static void si_vm_decode_fault(struct radeon_device *rdev,
 	       block, mc_id);
 }
 
-void si_vm_flush(struct radeon_device *rdev, int ridx, struct radeon_vm *vm)
+void si_vm_flush(struct radeon_device *rdev, struct radeon_ring *ring,
+		 unsigned vm_id, uint64_t pd_addr)
 {
-	struct radeon_ring *ring = &rdev->ring[ridx];
-
-	if (vm == NULL)
-		return;
-
 	/* write new base address */
 	radeon_ring_write(ring, PACKET3(PACKET3_WRITE_DATA, 3));
 	radeon_ring_write(ring, (WRITE_DATA_ENGINE_SEL(1) |
 				 WRITE_DATA_DST_SEL(0)));
 
-	if (vm->id < 8) {
+	if (vm_id < 8) {
 		radeon_ring_write(ring,
-				  (VM_CONTEXT0_PAGE_TABLE_BASE_ADDR + (vm->id << 2)) >> 2);
+				  (VM_CONTEXT0_PAGE_TABLE_BASE_ADDR + (vm_id << 2)) >> 2);
 	} else {
 		radeon_ring_write(ring,
-				  (VM_CONTEXT8_PAGE_TABLE_BASE_ADDR + ((vm->id - 8) << 2)) >> 2);
+				  (VM_CONTEXT8_PAGE_TABLE_BASE_ADDR + ((vm_id - 8) << 2)) >> 2);
 	}
 	radeon_ring_write(ring, 0);
-	radeon_ring_write(ring, vm->pd_gpu_addr >> 12);
+	radeon_ring_write(ring, pd_addr >> 12);
 
 	/* flush hdp cache */
 	radeon_ring_write(ring, PACKET3(PACKET3_WRITE_DATA, 3));
@@ -5050,7 +5055,7 @@ void si_vm_flush(struct radeon_device *rdev, int ridx, struct radeon_vm *vm)
 				 WRITE_DATA_DST_SEL(0)));
 	radeon_ring_write(ring, VM_INVALIDATE_REQUEST >> 2);
 	radeon_ring_write(ring, 0);
-	radeon_ring_write(ring, 1 << vm->id);
+	radeon_ring_write(ring, 1 << vm_id);
 
 	/* sync PFP to ME, otherwise we might get invalid PFP reads */
 	radeon_ring_write(ring, PACKET3(PACKET3_PFP_SYNC_ME, 0));
diff --git a/drivers/gpu/drm/radeon/si_dma.c b/drivers/gpu/drm/radeon/si_dma.c
index 7c22baaf94db..f5cc777e1c5f 100644
--- a/drivers/gpu/drm/radeon/si_dma.c
+++ b/drivers/gpu/drm/radeon/si_dma.c
@@ -185,20 +185,17 @@ void si_dma_vm_set_pages(struct radeon_device *rdev,
 	}
 }
 
-void si_dma_vm_flush(struct radeon_device *rdev, int ridx, struct radeon_vm *vm)
-{
-	struct radeon_ring *ring = &rdev->ring[ridx];
-
-	if (vm == NULL)
-		return;
+void si_dma_vm_flush(struct radeon_device *rdev, struct radeon_ring *ring,
+		     unsigned vm_id, uint64_t pd_addr)
 
+{
 	radeon_ring_write(ring, DMA_PACKET(DMA_PACKET_SRBM_WRITE, 0, 0, 0, 0));
-	if (vm->id < 8) {
-		radeon_ring_write(ring, (0xf << 16) | ((VM_CONTEXT0_PAGE_TABLE_BASE_ADDR + (vm->id << 2)) >> 2));
+	if (vm_id < 8) {
+		radeon_ring_write(ring, (0xf << 16) | ((VM_CONTEXT0_PAGE_TABLE_BASE_ADDR + (vm_id << 2)) >> 2));
 	} else {
-		radeon_ring_write(ring, (0xf << 16) | ((VM_CONTEXT8_PAGE_TABLE_BASE_ADDR + ((vm->id - 8) << 2)) >> 2));
+		radeon_ring_write(ring, (0xf << 16) | ((VM_CONTEXT8_PAGE_TABLE_BASE_ADDR + ((vm_id - 8) << 2)) >> 2));
 	}
-	radeon_ring_write(ring, vm->pd_gpu_addr >> 12);
+	radeon_ring_write(ring, pd_addr >> 12);
 
 	/* flush hdp cache */
 	radeon_ring_write(ring, DMA_PACKET(DMA_PACKET_SRBM_WRITE, 0, 0, 0, 0));
@@ -208,7 +205,7 @@ void si_dma_vm_flush(struct radeon_device *rdev, int ridx, struct radeon_vm *vm)
 	/* bits 0-7 are the VM contexts0-7 */
 	radeon_ring_write(ring, DMA_PACKET(DMA_PACKET_SRBM_WRITE, 0, 0, 0, 0));
 	radeon_ring_write(ring, (0xf << 16) | (VM_INVALIDATE_REQUEST >> 2));
-	radeon_ring_write(ring, 1 << vm->id);
+	radeon_ring_write(ring, 1 << vm_id);
 }
 
 /**
@@ -218,41 +215,38 @@ void si_dma_vm_flush(struct radeon_device *rdev, int ridx, struct radeon_vm *vm)
  * @src_offset: src GPU address
  * @dst_offset: dst GPU address
  * @num_gpu_pages: number of GPU pages to xfer
- * @fence: radeon fence object
+ * @resv: reservation object to sync to
  *
  * Copy GPU paging using the DMA engine (SI).
  * Used by the radeon ttm implementation to move pages if
  * registered as the asic copy callback.
  */
-int si_copy_dma(struct radeon_device *rdev,
-		uint64_t src_offset, uint64_t dst_offset,
-		unsigned num_gpu_pages,
-		struct radeon_fence **fence)
+struct radeon_fence *si_copy_dma(struct radeon_device *rdev,
+				 uint64_t src_offset, uint64_t dst_offset,
+				 unsigned num_gpu_pages,
+				 struct reservation_object *resv)
 {
-	struct radeon_semaphore *sem = NULL;
+	struct radeon_fence *fence;
+	struct radeon_sync sync;
 	int ring_index = rdev->asic->copy.dma_ring_index;
 	struct radeon_ring *ring = &rdev->ring[ring_index];
 	u32 size_in_bytes, cur_size_in_bytes;
 	int i, num_loops;
 	int r = 0;
 
-	r = radeon_semaphore_create(rdev, &sem);
-	if (r) {
-		DRM_ERROR("radeon: moving bo (%d).\n", r);
-		return r;
-	}
+	radeon_sync_create(&sync);
 
 	size_in_bytes = (num_gpu_pages << RADEON_GPU_PAGE_SHIFT);
 	num_loops = DIV_ROUND_UP(size_in_bytes, 0xfffff);
 	r = radeon_ring_lock(rdev, ring, num_loops * 5 + 11);
 	if (r) {
 		DRM_ERROR("radeon: moving bo (%d).\n", r);
-		radeon_semaphore_free(rdev, &sem, NULL);
-		return r;
+		radeon_sync_free(rdev, &sync, NULL);
+		return ERR_PTR(r);
 	}
 
-	radeon_semaphore_sync_to(sem, *fence);
-	radeon_semaphore_sync_rings(rdev, sem, ring->idx);
+	radeon_sync_resv(rdev, &sync, resv, false);
+	radeon_sync_rings(rdev, &sync, ring->idx);
 
 	for (i = 0; i < num_loops; i++) {
 		cur_size_in_bytes = size_in_bytes;
@@ -268,16 +262,16 @@ int si_copy_dma(struct radeon_device *rdev,
 		dst_offset += cur_size_in_bytes;
 	}
 
-	r = radeon_fence_emit(rdev, fence, ring->idx);
+	r = radeon_fence_emit(rdev, &fence, ring->idx);
 	if (r) {
 		radeon_ring_unlock_undo(rdev, ring);
-		radeon_semaphore_free(rdev, &sem, NULL);
-		return r;
+		radeon_sync_free(rdev, &sync, NULL);
+		return ERR_PTR(r);
 	}
 
 	radeon_ring_unlock_commit(rdev, ring, false);
-	radeon_semaphore_free(rdev, &sem, *fence);
+	radeon_sync_free(rdev, &sync, fence);
 
-	return r;
+	return fence;
 }
 
diff --git a/drivers/gpu/drm/radeon/si_dpm.c b/drivers/gpu/drm/radeon/si_dpm.c
index 70e61ffeace2..32e354b8b0ab 100644
--- a/drivers/gpu/drm/radeon/si_dpm.c
+++ b/drivers/gpu/drm/radeon/si_dpm.c
@@ -23,6 +23,7 @@
 
 #include "drmP.h"
 #include "radeon.h"
+#include "radeon_asic.h"
 #include "sid.h"
 #include "r600_dpm.h"
 #include "si_dpm.h"
@@ -3397,6 +3398,15 @@ static int si_process_firmware_header(struct radeon_device *rdev)
 
 	ret = si_read_smc_sram_dword(rdev,
 				     SISLANDS_SMC_FIRMWARE_HEADER_LOCATION +
+				     SISLANDS_SMC_FIRMWARE_HEADER_fanTable,
+				     &tmp, si_pi->sram_end);
+	if (ret)
+		return ret;
+
+	si_pi->fan_table_start = tmp;
+
+	ret = si_read_smc_sram_dword(rdev,
+				     SISLANDS_SMC_FIRMWARE_HEADER_LOCATION +
 				     SISLANDS_SMC_FIRMWARE_HEADER_mcArbDramAutoRefreshTable,
 				     &tmp, si_pi->sram_end);
 	if (ret)
@@ -5816,8 +5826,33 @@ void si_dpm_setup_asic(struct radeon_device *rdev)
 	si_enable_acpi_power_management(rdev);
 }
 
-static int si_set_thermal_temperature_range(struct radeon_device *rdev,
-					int min_temp, int max_temp)
+static int si_thermal_enable_alert(struct radeon_device *rdev,
+				   bool enable)
+{
+	u32 thermal_int = RREG32(CG_THERMAL_INT);
+
+	if (enable) {
+		PPSMC_Result result;
+
+		thermal_int &= ~(THERM_INT_MASK_HIGH | THERM_INT_MASK_LOW);
+		WREG32(CG_THERMAL_INT, thermal_int);
+		rdev->irq.dpm_thermal = false;
+		result = si_send_msg_to_smc(rdev, PPSMC_MSG_EnableThermalInterrupt);
+		if (result != PPSMC_Result_OK) {
+			DRM_DEBUG_KMS("Could not enable thermal interrupts.\n");
+			return -EINVAL;
+		}
+	} else {
+		thermal_int |= THERM_INT_MASK_HIGH | THERM_INT_MASK_LOW;
+		WREG32(CG_THERMAL_INT, thermal_int);
+		rdev->irq.dpm_thermal = true;
+	}
+
+	return 0;
+}
+
+static int si_thermal_set_temperature_range(struct radeon_device *rdev,
+					    int min_temp, int max_temp)
 {
 	int low_temp = 0 * 1000;
 	int high_temp = 255 * 1000;
@@ -5841,6 +5876,309 @@ static int si_set_thermal_temperature_range(struct radeon_device *rdev,
 	return 0;
 }
 
+static void si_fan_ctrl_set_static_mode(struct radeon_device *rdev, u32 mode)
+{
+	struct si_power_info *si_pi = si_get_pi(rdev);
+	u32 tmp;
+
+	if (si_pi->fan_ctrl_is_in_default_mode) {
+		tmp = (RREG32(CG_FDO_CTRL2) & FDO_PWM_MODE_MASK) >> FDO_PWM_MODE_SHIFT;
+		si_pi->fan_ctrl_default_mode = tmp;
+		tmp = (RREG32(CG_FDO_CTRL2) & TMIN_MASK) >> TMIN_SHIFT;
+		si_pi->t_min = tmp;
+		si_pi->fan_ctrl_is_in_default_mode = false;
+	}
+
+	tmp = RREG32(CG_FDO_CTRL2) & ~TMIN_MASK;
+	tmp |= TMIN(0);
+	WREG32(CG_FDO_CTRL2, tmp);
+
+	tmp = RREG32(CG_FDO_CTRL2) & ~FDO_PWM_MODE_MASK;
+	tmp |= FDO_PWM_MODE(mode);
+	WREG32(CG_FDO_CTRL2, tmp);
+}
+
+static int si_thermal_setup_fan_table(struct radeon_device *rdev)
+{
+	struct si_power_info *si_pi = si_get_pi(rdev);
+	PP_SIslands_FanTable fan_table = { FDO_MODE_HARDWARE };
+	u32 duty100;
+	u32 t_diff1, t_diff2, pwm_diff1, pwm_diff2;
+	u16 fdo_min, slope1, slope2;
+	u32 reference_clock, tmp;
+	int ret;
+	u64 tmp64;
+
+	if (!si_pi->fan_table_start) {
+		rdev->pm.dpm.fan.ucode_fan_control = false;
+		return 0;
+	}
+
+	duty100 = (RREG32(CG_FDO_CTRL1) & FMAX_DUTY100_MASK) >> FMAX_DUTY100_SHIFT;
+
+	if (duty100 == 0) {
+		rdev->pm.dpm.fan.ucode_fan_control = false;
+		return 0;
+	}
+
+	tmp64 = (u64)rdev->pm.dpm.fan.pwm_min * duty100;
+	do_div(tmp64, 10000);
+	fdo_min = (u16)tmp64;
+
+	t_diff1 = rdev->pm.dpm.fan.t_med - rdev->pm.dpm.fan.t_min;
+	t_diff2 = rdev->pm.dpm.fan.t_high - rdev->pm.dpm.fan.t_med;
+
+	pwm_diff1 = rdev->pm.dpm.fan.pwm_med - rdev->pm.dpm.fan.pwm_min;
+	pwm_diff2 = rdev->pm.dpm.fan.pwm_high - rdev->pm.dpm.fan.pwm_med;
+
+	slope1 = (u16)((50 + ((16 * duty100 * pwm_diff1) / t_diff1)) / 100);
+	slope2 = (u16)((50 + ((16 * duty100 * pwm_diff2) / t_diff2)) / 100);
+
+	fan_table.slope1 = cpu_to_be16(slope1);
+	fan_table.slope2 = cpu_to_be16(slope2);
+
+	fan_table.fdo_min = cpu_to_be16(fdo_min);
+
+	fan_table.hys_down = cpu_to_be16(rdev->pm.dpm.fan.t_hyst);
+
+	fan_table.hys_up = cpu_to_be16(1);
+
+	fan_table.hys_slope = cpu_to_be16(1);
+
+	fan_table.temp_resp_lim = cpu_to_be16(5);
+
+	reference_clock = radeon_get_xclk(rdev);
+
+	fan_table.refresh_period = cpu_to_be32((rdev->pm.dpm.fan.cycle_delay *
+						reference_clock) / 1600);
+
+	fan_table.fdo_max = cpu_to_be16((u16)duty100);
+
+	tmp = (RREG32(CG_MULT_THERMAL_CTRL) & TEMP_SEL_MASK) >> TEMP_SEL_SHIFT;
+	fan_table.temp_src = (uint8_t)tmp;
+
+	ret = si_copy_bytes_to_smc(rdev,
+				   si_pi->fan_table_start,
+				   (u8 *)(&fan_table),
+				   sizeof(fan_table),
+				   si_pi->sram_end);
+
+	if (ret) {
+		DRM_ERROR("Failed to load fan table to the SMC.");
+		rdev->pm.dpm.fan.ucode_fan_control = false;
+	}
+
+	return 0;
+}
+
+static int si_fan_ctrl_start_smc_fan_control(struct radeon_device *rdev)
+{
+	PPSMC_Result ret;
+
+	ret = si_send_msg_to_smc(rdev, PPSMC_StartFanControl);
+	if (ret == PPSMC_Result_OK)
+		return 0;
+	else
+		return -EINVAL;
+}
+
+static int si_fan_ctrl_stop_smc_fan_control(struct radeon_device *rdev)
+{
+	PPSMC_Result ret;
+
+	ret = si_send_msg_to_smc(rdev, PPSMC_StopFanControl);
+	if (ret == PPSMC_Result_OK)
+		return 0;
+	else
+		return -EINVAL;
+}
+
+#if 0
+static int si_fan_ctrl_get_fan_speed_percent(struct radeon_device *rdev,
+					     u32 *speed)
+{
+	u32 duty, duty100;
+	u64 tmp64;
+
+	if (rdev->pm.no_fan)
+		return -ENOENT;
+
+	duty100 = (RREG32(CG_FDO_CTRL1) & FMAX_DUTY100_MASK) >> FMAX_DUTY100_SHIFT;
+	duty = (RREG32(CG_THERMAL_STATUS) & FDO_PWM_DUTY_MASK) >> FDO_PWM_DUTY_SHIFT;
+
+	if (duty100 == 0)
+		return -EINVAL;
+
+	tmp64 = (u64)duty * 100;
+	do_div(tmp64, duty100);
+	*speed = (u32)tmp64;
+
+	if (*speed > 100)
+		*speed = 100;
+
+	return 0;
+}
+
+static int si_fan_ctrl_set_fan_speed_percent(struct radeon_device *rdev,
+					     u32 speed)
+{
+	u32 tmp;
+	u32 duty, duty100;
+	u64 tmp64;
+
+	if (rdev->pm.no_fan)
+		return -ENOENT;
+
+	if (speed > 100)
+		return -EINVAL;
+
+	if (rdev->pm.dpm.fan.ucode_fan_control)
+		si_fan_ctrl_stop_smc_fan_control(rdev);
+
+	duty100 = (RREG32(CG_FDO_CTRL1) & FMAX_DUTY100_MASK) >> FMAX_DUTY100_SHIFT;
+
+	if (duty100 == 0)
+		return -EINVAL;
+
+	tmp64 = (u64)speed * duty100;
+	do_div(tmp64, 100);
+	duty = (u32)tmp64;
+
+	tmp = RREG32(CG_FDO_CTRL0) & ~FDO_STATIC_DUTY_MASK;
+	tmp |= FDO_STATIC_DUTY(duty);
+	WREG32(CG_FDO_CTRL0, tmp);
+
+	si_fan_ctrl_set_static_mode(rdev, FDO_PWM_MODE_STATIC);
+
+	return 0;
+}
+
+static int si_fan_ctrl_get_fan_speed_rpm(struct radeon_device *rdev,
+					 u32 *speed)
+{
+	u32 tach_period;
+	u32 xclk = radeon_get_xclk(rdev);
+
+	if (rdev->pm.no_fan)
+		return -ENOENT;
+
+	if (rdev->pm.fan_pulses_per_revolution == 0)
+		return -ENOENT;
+
+	tach_period = (RREG32(CG_TACH_STATUS) & TACH_PERIOD_MASK) >> TACH_PERIOD_SHIFT;
+	if (tach_period == 0)
+		return -ENOENT;
+
+	*speed = 60 * xclk * 10000 / tach_period;
+
+	return 0;
+}
+
+static int si_fan_ctrl_set_fan_speed_rpm(struct radeon_device *rdev,
+					 u32 speed)
+{
+	u32 tach_period, tmp;
+	u32 xclk = radeon_get_xclk(rdev);
+
+	if (rdev->pm.no_fan)
+		return -ENOENT;
+
+	if (rdev->pm.fan_pulses_per_revolution == 0)
+		return -ENOENT;
+
+	if ((speed < rdev->pm.fan_min_rpm) ||
+	    (speed > rdev->pm.fan_max_rpm))
+		return -EINVAL;
+
+	if (rdev->pm.dpm.fan.ucode_fan_control)
+		si_fan_ctrl_stop_smc_fan_control(rdev);
+
+	tach_period = 60 * xclk * 10000 / (8 * speed);
+	tmp = RREG32(CG_TACH_CTRL) & ~TARGET_PERIOD_MASK;
+	tmp |= TARGET_PERIOD(tach_period);
+	WREG32(CG_TACH_CTRL, tmp);
+
+	si_fan_ctrl_set_static_mode(rdev, FDO_PWM_MODE_STATIC_RPM);
+
+	return 0;
+}
+#endif
+
+static void si_fan_ctrl_set_default_mode(struct radeon_device *rdev)
+{
+	struct si_power_info *si_pi = si_get_pi(rdev);
+	u32 tmp;
+
+	if (!si_pi->fan_ctrl_is_in_default_mode) {
+		tmp = RREG32(CG_FDO_CTRL2) & ~FDO_PWM_MODE_MASK;
+		tmp |= FDO_PWM_MODE(si_pi->fan_ctrl_default_mode);
+		WREG32(CG_FDO_CTRL2, tmp);
+
+		tmp = RREG32(CG_FDO_CTRL2) & ~TMIN_MASK;
+		tmp |= TMIN(si_pi->t_min);
+		WREG32(CG_FDO_CTRL2, tmp);
+		si_pi->fan_ctrl_is_in_default_mode = true;
+	}
+}
+
+static void si_thermal_start_smc_fan_control(struct radeon_device *rdev)
+{
+	if (rdev->pm.dpm.fan.ucode_fan_control) {
+		si_fan_ctrl_start_smc_fan_control(rdev);
+		si_fan_ctrl_set_static_mode(rdev, FDO_PWM_MODE_STATIC);
+	}
+}
+
+static void si_thermal_initialize(struct radeon_device *rdev)
+{
+	u32 tmp;
+
+	if (rdev->pm.fan_pulses_per_revolution) {
+		tmp = RREG32(CG_TACH_CTRL) & ~EDGE_PER_REV_MASK;
+		tmp |= EDGE_PER_REV(rdev->pm.fan_pulses_per_revolution -1);
+		WREG32(CG_TACH_CTRL, tmp);
+	}
+
+	tmp = RREG32(CG_FDO_CTRL2) & ~TACH_PWM_RESP_RATE_MASK;
+	tmp |= TACH_PWM_RESP_RATE(0x28);
+	WREG32(CG_FDO_CTRL2, tmp);
+}
+
+static int si_thermal_start_thermal_controller(struct radeon_device *rdev)
+{
+	int ret;
+
+	si_thermal_initialize(rdev);
+	ret = si_thermal_set_temperature_range(rdev, R600_TEMP_RANGE_MIN, R600_TEMP_RANGE_MAX);
+	if (ret)
+		return ret;
+	ret = si_thermal_enable_alert(rdev, true);
+	if (ret)
+		return ret;
+	if (rdev->pm.dpm.fan.ucode_fan_control) {
+		ret = si_halt_smc(rdev);
+		if (ret)
+			return ret;
+		ret = si_thermal_setup_fan_table(rdev);
+		if (ret)
+			return ret;
+		ret = si_resume_smc(rdev);
+		if (ret)
+			return ret;
+		si_thermal_start_smc_fan_control(rdev);
+	}
+
+	return 0;
+}
+
+static void si_thermal_stop_thermal_controller(struct radeon_device *rdev)
+{
+	if (!rdev->pm.no_fan) {
+		si_fan_ctrl_set_default_mode(rdev);
+		si_fan_ctrl_stop_smc_fan_control(rdev);
+	}
+}
+
 int si_dpm_enable(struct radeon_device *rdev)
 {
 	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
@@ -5953,31 +6291,39 @@ int si_dpm_enable(struct radeon_device *rdev)
 
 	si_enable_auto_throttle_source(rdev, RADEON_DPM_AUTO_THROTTLE_SRC_THERMAL, true);
 
+	si_thermal_start_thermal_controller(rdev);
+
 	ni_update_current_ps(rdev, boot_ps);
 
 	return 0;
 }
 
-int si_dpm_late_enable(struct radeon_device *rdev)
+static int si_set_temperature_range(struct radeon_device *rdev)
 {
 	int ret;
 
-	if (rdev->irq.installed &&
-	    r600_is_internal_thermal_sensor(rdev->pm.int_thermal_type)) {
-		PPSMC_Result result;
+	ret = si_thermal_enable_alert(rdev, false);
+	if (ret)
+		return ret;
+	ret = si_thermal_set_temperature_range(rdev, R600_TEMP_RANGE_MIN, R600_TEMP_RANGE_MAX);
+	if (ret)
+		return ret;
+	ret = si_thermal_enable_alert(rdev, true);
+	if (ret)
+		return ret;
 
-		ret = si_set_thermal_temperature_range(rdev, R600_TEMP_RANGE_MIN, R600_TEMP_RANGE_MAX);
-		if (ret)
-			return ret;
-		rdev->irq.dpm_thermal = true;
-		radeon_irq_set(rdev);
-		result = si_send_msg_to_smc(rdev, PPSMC_MSG_EnableThermalInterrupt);
+	return ret;
+}
 
-		if (result != PPSMC_Result_OK)
-			DRM_DEBUG_KMS("Could not enable thermal interrupts.\n");
-	}
+int si_dpm_late_enable(struct radeon_device *rdev)
+{
+	int ret;
 
-	return 0;
+	ret = si_set_temperature_range(rdev);
+	if (ret)
+		return ret;
+
+	return ret;
 }
 
 void si_dpm_disable(struct radeon_device *rdev)
@@ -5987,6 +6333,7 @@ void si_dpm_disable(struct radeon_device *rdev)
 
 	if (!si_is_smc_running(rdev))
 		return;
+	si_thermal_stop_thermal_controller(rdev);
 	si_disable_ulv(rdev);
 	si_clear_vc(rdev);
 	if (pi->thermal_protection)
@@ -6255,7 +6602,7 @@ static void si_parse_pplib_clock_info(struct radeon_device *rdev,
 	if ((rps->class2 & ATOM_PPLIB_CLASSIFICATION2_ULV) &&
 	    index == 0) {
 		/* XXX disable for A0 tahiti */
-		si_pi->ulv.supported = true;
+		si_pi->ulv.supported = false;
 		si_pi->ulv.pl = *pl;
 		si_pi->ulv.one_pcie_lane_in_ulv = false;
 		si_pi->ulv.volt_change_delay = SISLANDS_ULVVOLTAGECHANGEDELAY_DFLT;
@@ -6525,6 +6872,9 @@ int si_dpm_init(struct radeon_device *rdev)
 		rdev->pm.dpm.dyn_state.max_clock_voltage_on_dc =
 			rdev->pm.dpm.dyn_state.max_clock_voltage_on_ac;
 
+	si_pi->fan_ctrl_is_in_default_mode = true;
+	rdev->pm.dpm.fan.ucode_fan_control = false;
+
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/radeon/si_dpm.h b/drivers/gpu/drm/radeon/si_dpm.h
index 8b5c06a0832d..d16bb1b5f10f 100644
--- a/drivers/gpu/drm/radeon/si_dpm.h
+++ b/drivers/gpu/drm/radeon/si_dpm.h
@@ -182,6 +182,7 @@ struct si_power_info {
 	u32 dte_table_start;
 	u32 spll_table_start;
 	u32 papm_cfg_table_start;
+	u32 fan_table_start;
 	/* CAC stuff */
 	const struct si_cac_config_reg *cac_weights;
 	const struct si_cac_config_reg *lcac_config;
@@ -197,6 +198,10 @@ struct si_power_info {
 	/* SVI2 */
 	u8 svd_gpio_id;
 	u8 svc_gpio_id;
+	/* fan control */
+	bool fan_ctrl_is_in_default_mode;
+	u32 t_min;
+	u32 fan_ctrl_default_mode;
 };
 
 #define SISLANDS_INITIAL_STATE_ARB_INDEX    0
diff --git a/drivers/gpu/drm/radeon/si_smc.c b/drivers/gpu/drm/radeon/si_smc.c
index 73dbc79c959d..e5bb92f16775 100644
--- a/drivers/gpu/drm/radeon/si_smc.c
+++ b/drivers/gpu/drm/radeon/si_smc.c
@@ -135,7 +135,7 @@ void si_reset_smc(struct radeon_device *rdev)
 
 int si_program_jump_on_start(struct radeon_device *rdev)
 {
-	static u8 data[] = { 0x0E, 0x00, 0x40, 0x40 };
+	static const u8 data[] = { 0x0E, 0x00, 0x40, 0x40 };
 
 	return si_copy_bytes_to_smc(rdev, 0x0, data, 4, sizeof(data)+1);
 }
diff --git a/drivers/gpu/drm/radeon/sid.h b/drivers/gpu/drm/radeon/sid.h
index fd414d34d885..4069be89e585 100644
--- a/drivers/gpu/drm/radeon/sid.h
+++ b/drivers/gpu/drm/radeon/sid.h
@@ -180,7 +180,10 @@
 #define		DIG_THERM_DPM(x)			((x) << 14)
 #define		DIG_THERM_DPM_MASK			0x003FC000
 #define		DIG_THERM_DPM_SHIFT			14
-
+#define	CG_THERMAL_STATUS				0x704
+#define		FDO_PWM_DUTY(x)				((x) << 9)
+#define		FDO_PWM_DUTY_MASK			(0xff << 9)
+#define		FDO_PWM_DUTY_SHIFT			9
 #define	CG_THERMAL_INT					0x708
 #define		DIG_THERM_INTH(x)			((x) << 8)
 #define		DIG_THERM_INTH_MASK			0x0000FF00
@@ -191,6 +194,10 @@
 #define 	THERM_INT_MASK_HIGH			(1 << 24)
 #define 	THERM_INT_MASK_LOW			(1 << 25)
 
+#define	CG_MULT_THERMAL_CTRL					0x710
+#define		TEMP_SEL(x)					((x) << 20)
+#define		TEMP_SEL_MASK					(0xff << 20)
+#define		TEMP_SEL_SHIFT					20
 #define	CG_MULT_THERMAL_STATUS					0x714
 #define		ASIC_MAX_TEMP(x)				((x) << 0)
 #define		ASIC_MAX_TEMP_MASK				0x000001ff
@@ -199,6 +206,37 @@
 #define		CTF_TEMP_MASK					0x0003fe00
 #define		CTF_TEMP_SHIFT					9
 
+#define	CG_FDO_CTRL0					0x754
+#define		FDO_STATIC_DUTY(x)			((x) << 0)
+#define		FDO_STATIC_DUTY_MASK			0x000000FF
+#define		FDO_STATIC_DUTY_SHIFT			0
+#define	CG_FDO_CTRL1					0x758
+#define		FMAX_DUTY100(x)				((x) << 0)
+#define		FMAX_DUTY100_MASK			0x000000FF
+#define		FMAX_DUTY100_SHIFT			0
+#define	CG_FDO_CTRL2					0x75C
+#define		TMIN(x)					((x) << 0)
+#define		TMIN_MASK				0x000000FF
+#define		TMIN_SHIFT				0
+#define		FDO_PWM_MODE(x)				((x) << 11)
+#define		FDO_PWM_MODE_MASK			(7 << 11)
+#define		FDO_PWM_MODE_SHIFT			11
+#define		TACH_PWM_RESP_RATE(x)			((x) << 25)
+#define		TACH_PWM_RESP_RATE_MASK			(0x7f << 25)
+#define		TACH_PWM_RESP_RATE_SHIFT		25
+
+#define CG_TACH_CTRL                                    0x770
+#       define EDGE_PER_REV(x)                          ((x) << 0)
+#       define EDGE_PER_REV_MASK                        (0x7 << 0)
+#       define EDGE_PER_REV_SHIFT                       0
+#       define TARGET_PERIOD(x)                         ((x) << 3)
+#       define TARGET_PERIOD_MASK                       0xfffffff8
+#       define TARGET_PERIOD_SHIFT                      3
+#define CG_TACH_STATUS                                  0x774
+#       define TACH_PERIOD(x)                           ((x) << 0)
+#       define TACH_PERIOD_MASK                         0xffffffff
+#       define TACH_PERIOD_SHIFT                        0
+
 #define GENERAL_PWRMGT                                  0x780
 #       define GLOBAL_PWRMGT_EN                         (1 << 0)
 #       define STATIC_PM_EN                             (1 << 1)
@@ -736,7 +774,7 @@
 #       define DESCRIPTION16(x)                          (((x) & 0xff) << 0)
 #       define DESCRIPTION17(x)                          (((x) & 0xff) << 8)
 
-#define AZ_F0_CODEC_PIN_CONTROL_HOTPLUG_CONTROL          0x54
+#define AZ_F0_CODEC_PIN_CONTROL_HOT_PLUG_CONTROL         0x54
 #       define AUDIO_ENABLED                             (1 << 31)
 
 #define AZ_F0_CODEC_PIN_CONTROL_RESPONSE_CONFIGURATION_DEFAULT  0x56
diff --git a/drivers/gpu/drm/radeon/sislands_smc.h b/drivers/gpu/drm/radeon/sislands_smc.h
index 623a0b1e2d9d..3c779838d9ab 100644
--- a/drivers/gpu/drm/radeon/sislands_smc.h
+++ b/drivers/gpu/drm/radeon/sislands_smc.h
@@ -245,6 +245,31 @@ typedef struct SISLANDS_SMC_STATETABLE SISLANDS_SMC_STATETABLE;
 #define SI_SMC_SOFT_REGISTER_svi_rework_gpio_id_svd   0x11c
 #define SI_SMC_SOFT_REGISTER_svi_rework_gpio_id_svc   0x120
 
+struct PP_SIslands_FanTable
+{
+	uint8_t  fdo_mode;
+	uint8_t  padding;
+	int16_t  temp_min;
+	int16_t  temp_med;
+	int16_t  temp_max;
+	int16_t  slope1;
+	int16_t  slope2;
+	int16_t  fdo_min;
+	int16_t  hys_up;
+	int16_t  hys_down;
+	int16_t  hys_slope;
+	int16_t  temp_resp_lim;
+	int16_t  temp_curr;
+	int16_t  slope_curr;
+	int16_t  pwm_curr;
+	uint32_t refresh_period;
+	int16_t  fdo_max;
+	uint8_t  temp_src;
+	int8_t  padding2;
+};
+
+typedef struct PP_SIslands_FanTable PP_SIslands_FanTable;
+
 #define SMC_SISLANDS_LKGE_LUT_NUM_OF_TEMP_ENTRIES 16
 #define SMC_SISLANDS_LKGE_LUT_NUM_OF_VOLT_ENTRIES 32
 
diff --git a/drivers/gpu/drm/radeon/smu7_discrete.h b/drivers/gpu/drm/radeon/smu7_discrete.h
index 82f70c90a9ee..0b0b404ff091 100644
--- a/drivers/gpu/drm/radeon/smu7_discrete.h
+++ b/drivers/gpu/drm/radeon/smu7_discrete.h
@@ -431,6 +431,31 @@ struct SMU7_Discrete_MCRegisters
 
 typedef struct SMU7_Discrete_MCRegisters SMU7_Discrete_MCRegisters;
 
+struct SMU7_Discrete_FanTable
+{
+	uint16_t FdoMode;
+	int16_t  TempMin;
+	int16_t  TempMed;
+	int16_t  TempMax;
+	int16_t  Slope1;
+	int16_t  Slope2;
+	int16_t  FdoMin;
+	int16_t  HystUp;
+	int16_t  HystDown;
+	int16_t  HystSlope;
+	int16_t  TempRespLim;
+	int16_t  TempCurr;
+	int16_t  SlopeCurr;
+	int16_t  PwmCurr;
+	uint32_t RefreshPeriod;
+	int16_t  FdoMax;
+	uint8_t  TempSrc;
+	int8_t   Padding;
+};
+
+typedef struct SMU7_Discrete_FanTable SMU7_Discrete_FanTable;
+
+
 struct SMU7_Discrete_PmFuses {
   // dw0-dw1
   uint8_t BapmVddCVidHiSidd[8];
@@ -462,7 +487,10 @@ struct SMU7_Discrete_PmFuses {
   uint8_t BapmVddCVidHiSidd2[8];
 
   // dw11-dw12
-  uint32_t Reserved6[2];
+  int16_t FuzzyFan_ErrorSetDelta;
+  int16_t FuzzyFan_ErrorRateSetDelta;
+  int16_t FuzzyFan_PwmSetDelta;
+  uint16_t CalcMeasPowerBlend;
 
   // dw13-dw16
   uint8_t GnbLPML[16];
diff --git a/drivers/gpu/drm/radeon/sumo_dpm.c b/drivers/gpu/drm/radeon/sumo_dpm.c
index 3f0e8d7b8dbe..1f8a8833e1be 100644
--- a/drivers/gpu/drm/radeon/sumo_dpm.c
+++ b/drivers/gpu/drm/radeon/sumo_dpm.c
@@ -23,6 +23,7 @@
 
 #include "drmP.h"
 #include "radeon.h"
+#include "radeon_asic.h"
 #include "sumod.h"
 #include "r600_dpm.h"
 #include "cypress_dpm.h"
diff --git a/drivers/gpu/drm/radeon/trinity_dpm.c b/drivers/gpu/drm/radeon/trinity_dpm.c
index 57f780053b3e..b4ec5c4e7969 100644
--- a/drivers/gpu/drm/radeon/trinity_dpm.c
+++ b/drivers/gpu/drm/radeon/trinity_dpm.c
@@ -23,6 +23,7 @@
 
 #include "drmP.h"
 #include "radeon.h"
+#include "radeon_asic.h"
 #include "trinityd.h"
 #include "r600_dpm.h"
 #include "trinity_dpm.h"
diff --git a/drivers/gpu/drm/radeon/uvd_v1_0.c b/drivers/gpu/drm/radeon/uvd_v1_0.c
index cda391347286..e72b3cb59358 100644
--- a/drivers/gpu/drm/radeon/uvd_v1_0.c
+++ b/drivers/gpu/drm/radeon/uvd_v1_0.c
@@ -22,6 +22,7 @@
  * Authors: Christian König <christian.koenig@amd.com>
  */
 
+#include <linux/firmware.h>
 #include <drm/drmP.h>
 #include "radeon.h"
 #include "radeon_asic.h"
@@ -70,6 +71,82 @@ void uvd_v1_0_set_wptr(struct radeon_device *rdev,
 }
 
 /**
+ * uvd_v1_0_fence_emit - emit an fence & trap command
+ *
+ * @rdev: radeon_device pointer
+ * @fence: fence to emit
+ *
+ * Write a fence and a trap command to the ring.
+ */
+void uvd_v1_0_fence_emit(struct radeon_device *rdev,
+			 struct radeon_fence *fence)
+{
+	struct radeon_ring *ring = &rdev->ring[fence->ring];
+	uint64_t addr = rdev->fence_drv[fence->ring].gpu_addr;
+
+	radeon_ring_write(ring, PACKET0(UVD_GPCOM_VCPU_DATA0, 0));
+	radeon_ring_write(ring, addr & 0xffffffff);
+	radeon_ring_write(ring, PACKET0(UVD_GPCOM_VCPU_DATA1, 0));
+	radeon_ring_write(ring, fence->seq);
+	radeon_ring_write(ring, PACKET0(UVD_GPCOM_VCPU_CMD, 0));
+	radeon_ring_write(ring, 0);
+
+	radeon_ring_write(ring, PACKET0(UVD_GPCOM_VCPU_DATA0, 0));
+	radeon_ring_write(ring, 0);
+	radeon_ring_write(ring, PACKET0(UVD_GPCOM_VCPU_DATA1, 0));
+	radeon_ring_write(ring, 0);
+	radeon_ring_write(ring, PACKET0(UVD_GPCOM_VCPU_CMD, 0));
+	radeon_ring_write(ring, 2);
+	return;
+}
+
+/**
+ * uvd_v1_0_resume - memory controller programming
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Let the UVD memory controller know it's offsets
+ */
+int uvd_v1_0_resume(struct radeon_device *rdev)
+{
+	uint64_t addr;
+	uint32_t size;
+	int r;
+
+	r = radeon_uvd_resume(rdev);
+	if (r)
+		return r;
+
+	/* programm the VCPU memory controller bits 0-27 */
+	addr = (rdev->uvd.gpu_addr >> 3) + 16;
+	size = RADEON_GPU_PAGE_ALIGN(rdev->uvd_fw->size) >> 3;
+	WREG32(UVD_VCPU_CACHE_OFFSET0, addr);
+	WREG32(UVD_VCPU_CACHE_SIZE0, size);
+
+	addr += size;
+	size = RADEON_UVD_STACK_SIZE >> 3;
+	WREG32(UVD_VCPU_CACHE_OFFSET1, addr);
+	WREG32(UVD_VCPU_CACHE_SIZE1, size);
+
+	addr += size;
+	size = RADEON_UVD_HEAP_SIZE >> 3;
+	WREG32(UVD_VCPU_CACHE_OFFSET2, addr);
+	WREG32(UVD_VCPU_CACHE_SIZE2, size);
+
+	/* bits 28-31 */
+	addr = (rdev->uvd.gpu_addr >> 28) & 0xF;
+	WREG32(UVD_LMI_ADDR_EXT, (addr << 12) | (addr << 0));
+
+	/* bits 32-39 */
+	addr = (rdev->uvd.gpu_addr >> 32) & 0xFF;
+	WREG32(UVD_LMI_EXT40_ADDR, addr | (0x9 << 16) | (0x1 << 31));
+
+	WREG32(UVD_FW_START, *((uint32_t*)rdev->uvd.cpu_addr));
+
+	return 0;
+}
+
+/**
  * uvd_v1_0_init - start and test UVD block
  *
  * @rdev: radeon_device pointer
@@ -130,8 +207,32 @@ done:
 	/* lower clocks again */
 	radeon_set_uvd_clocks(rdev, 0, 0);
 
-	if (!r)
+	if (!r) {
+		switch (rdev->family) {
+		case CHIP_RV610:
+		case CHIP_RV630:
+		case CHIP_RV620:
+			/* 64byte granularity workaround */
+			WREG32(MC_CONFIG, 0);
+			WREG32(MC_CONFIG, 1 << 4);
+			WREG32(RS_DQ_RD_RET_CONF, 0x3f);
+			WREG32(MC_CONFIG, 0x1f);
+
+			/* fall through */
+		case CHIP_RV670:
+		case CHIP_RV635:
+
+			/* write clean workaround */
+			WREG32_P(UVD_VCPU_CNTL, 0x10, ~0x10);
+			break;
+
+		default:
+			/* TODO: Do we need more? */
+			break;
+		}
+
 		DRM_INFO("UVD initialized successfully.\n");
+	}
 
 	return r;
 }
@@ -218,12 +319,12 @@ int uvd_v1_0_start(struct radeon_device *rdev)
 	/* enable UMC */
 	WREG32_P(UVD_LMI_CTRL2, 0, ~(1 << 8));
 
+	WREG32_P(UVD_RB_ARB_CTRL, 0, ~(1 << 3));
+
 	/* boot up the VCPU */
 	WREG32(UVD_SOFT_RESET, 0);
 	mdelay(10);
 
-	WREG32_P(UVD_RB_ARB_CTRL, 0, ~(1 << 3));
-
 	for (i = 0; i < 10; ++i) {
 		uint32_t status;
 		for (j = 0; j < 100; ++j) {
diff --git a/drivers/gpu/drm/radeon/uvd_v2_2.c b/drivers/gpu/drm/radeon/uvd_v2_2.c
index 8bfdadd56598..89193519f8a1 100644
--- a/drivers/gpu/drm/radeon/uvd_v2_2.c
+++ b/drivers/gpu/drm/radeon/uvd_v2_2.c
@@ -72,6 +72,10 @@ int uvd_v2_2_resume(struct radeon_device *rdev)
 	uint32_t chip_id, size;
 	int r;
 
+	/* RV770 uses V1.0 MC */
+	if (rdev->family == CHIP_RV770)
+		return uvd_v1_0_resume(rdev);
+
 	r = radeon_uvd_resume(rdev);
 	if (r)
 		return r;
diff --git a/drivers/gpu/drm/rcar-du/Kconfig b/drivers/gpu/drm/rcar-du/Kconfig
index 2e3d7b5b0ad7..2324a526de65 100644
--- a/drivers/gpu/drm/rcar-du/Kconfig
+++ b/drivers/gpu/drm/rcar-du/Kconfig
@@ -6,14 +6,22 @@ config DRM_RCAR_DU
 	select DRM_KMS_CMA_HELPER
 	select DRM_GEM_CMA_HELPER
 	select DRM_KMS_FB_HELPER
+	select VIDEOMODE_HELPERS
 	help
 	  Choose this option if you have an R-Car chipset.
 	  If M is selected the module will be called rcar-du-drm.
 
+config DRM_RCAR_HDMI
+	bool "R-Car DU HDMI Encoder Support"
+	depends on DRM_RCAR_DU
+	depends on OF
+	help
+	  Enable support for external HDMI encoders.
+
 config DRM_RCAR_LVDS
 	bool "R-Car DU LVDS Encoder Support"
 	depends on DRM_RCAR_DU
 	depends on ARCH_R8A7790 || ARCH_R8A7791 || COMPILE_TEST
 	help
-	  Enable support the R-Car Display Unit embedded LVDS encoders
-	  (currently only on R8A7790).
+	  Enable support for the R-Car Display Unit embedded LVDS encoders
+	  (currently only on R8A7790 and R8A7791).
diff --git a/drivers/gpu/drm/rcar-du/Makefile b/drivers/gpu/drm/rcar-du/Makefile
index 12b8d4477835..05de1c4097af 100644
--- a/drivers/gpu/drm/rcar-du/Makefile
+++ b/drivers/gpu/drm/rcar-du/Makefile
@@ -7,6 +7,8 @@ rcar-du-drm-y := rcar_du_crtc.o \
 		 rcar_du_plane.o \
 		 rcar_du_vgacon.o
 
+rcar-du-drm-$(CONFIG_DRM_RCAR_HDMI)	+= rcar_du_hdmicon.o \
+					   rcar_du_hdmienc.o
 rcar-du-drm-$(CONFIG_DRM_RCAR_LVDS)	+= rcar_du_lvdsenc.o
 
 obj-$(CONFIG_DRM_RCAR_DU)		+= rcar-du-drm.o
diff --git a/drivers/gpu/drm/rcar-du/rcar_du_crtc.c b/drivers/gpu/drm/rcar-du/rcar_du_crtc.c
index 299267db2898..23cc910951f4 100644
--- a/drivers/gpu/drm/rcar-du/rcar_du_crtc.c
+++ b/drivers/gpu/drm/rcar-du/rcar_du_crtc.c
@@ -1,7 +1,7 @@
 /*
  * rcar_du_crtc.c  --  R-Car Display Unit CRTCs
  *
- * Copyright (C) 2013 Renesas Corporation
+ * Copyright (C) 2013-2014 Renesas Electronics Corporation
  *
  * Contact: Laurent Pinchart (laurent.pinchart@ideasonboard.com)
  *
@@ -19,6 +19,7 @@
 #include <drm/drm_crtc_helper.h>
 #include <drm/drm_fb_cma_helper.h>
 #include <drm/drm_gem_cma_helper.h>
+#include <drm/drm_plane_helper.h>
 
 #include "rcar_du_crtc.h"
 #include "rcar_du_drv.h"
@@ -585,7 +586,7 @@ int rcar_du_crtc_create(struct rcar_du_group *rgrp, unsigned int index)
 
 	if (irq < 0) {
 		dev_err(rcdu->dev, "no IRQ for CRTC %u\n", index);
-		return ret;
+		return irq;
 	}
 
 	ret = devm_request_irq(rcdu->dev, irq, rcar_du_crtc_irq, irqflags,
diff --git a/drivers/gpu/drm/rcar-du/rcar_du_crtc.h b/drivers/gpu/drm/rcar-du/rcar_du_crtc.h
index 43e7575c700c..984e6083699f 100644
--- a/drivers/gpu/drm/rcar-du/rcar_du_crtc.h
+++ b/drivers/gpu/drm/rcar-du/rcar_du_crtc.h
@@ -1,7 +1,7 @@
 /*
  * rcar_du_crtc.h  --  R-Car Display Unit CRTCs
  *
- * Copyright (C) 2013 Renesas Corporation
+ * Copyright (C) 2013-2014 Renesas Electronics Corporation
  *
  * Contact: Laurent Pinchart (laurent.pinchart@ideasonboard.com)
  *
@@ -15,7 +15,6 @@
 #define __RCAR_DU_CRTC_H__
 
 #include <linux/mutex.h>
-#include <linux/platform_data/rcar-du.h>
 
 #include <drm/drmP.h>
 #include <drm/drm_crtc.h>
@@ -41,6 +40,15 @@ struct rcar_du_crtc {
 
 #define to_rcar_crtc(c)	container_of(c, struct rcar_du_crtc, crtc)
 
+enum rcar_du_output {
+	RCAR_DU_OUTPUT_DPAD0,
+	RCAR_DU_OUTPUT_DPAD1,
+	RCAR_DU_OUTPUT_LVDS0,
+	RCAR_DU_OUTPUT_LVDS1,
+	RCAR_DU_OUTPUT_TCON,
+	RCAR_DU_OUTPUT_MAX,
+};
+
 int rcar_du_crtc_create(struct rcar_du_group *rgrp, unsigned int index);
 void rcar_du_crtc_enable_vblank(struct rcar_du_crtc *rcrtc, bool enable);
 void rcar_du_crtc_cancel_page_flip(struct rcar_du_crtc *rcrtc,
diff --git a/drivers/gpu/drm/rcar-du/rcar_du_drv.c b/drivers/gpu/drm/rcar-du/rcar_du_drv.c
index fda64b7b73e8..7bfa09cf18d5 100644
--- a/drivers/gpu/drm/rcar-du/rcar_du_drv.c
+++ b/drivers/gpu/drm/rcar-du/rcar_du_drv.c
@@ -1,7 +1,7 @@
 /*
  * rcar_du_drv.c  --  R-Car Display Unit DRM driver
  *
- * Copyright (C) 2013 Renesas Corporation
+ * Copyright (C) 2013-2014 Renesas Electronics Corporation
  *
  * Contact: Laurent Pinchart (laurent.pinchart@ideasonboard.com)
  *
@@ -15,6 +15,7 @@
 #include <linux/io.h>
 #include <linux/mm.h>
 #include <linux/module.h>
+#include <linux/of_device.h>
 #include <linux/platform_device.h>
 #include <linux/pm.h>
 #include <linux/slab.h>
@@ -30,6 +31,97 @@
 #include "rcar_du_regs.h"
 
 /* -----------------------------------------------------------------------------
+ * Device Information
+ */
+
+static const struct rcar_du_device_info rcar_du_r8a7779_info = {
+	.features = 0,
+	.num_crtcs = 2,
+	.routes = {
+		/* R8A7779 has two RGB outputs and one (currently unsupported)
+		 * TCON output.
+		 */
+		[RCAR_DU_OUTPUT_DPAD0] = {
+			.possible_crtcs = BIT(0),
+			.encoder_type = DRM_MODE_ENCODER_NONE,
+			.port = 0,
+		},
+		[RCAR_DU_OUTPUT_DPAD1] = {
+			.possible_crtcs = BIT(1) | BIT(0),
+			.encoder_type = DRM_MODE_ENCODER_NONE,
+			.port = 1,
+		},
+	},
+	.num_lvds = 0,
+};
+
+static const struct rcar_du_device_info rcar_du_r8a7790_info = {
+	.features = RCAR_DU_FEATURE_CRTC_IRQ_CLOCK | RCAR_DU_FEATURE_DEFR8,
+	.quirks = RCAR_DU_QUIRK_ALIGN_128B | RCAR_DU_QUIRK_LVDS_LANES,
+	.num_crtcs = 3,
+	.routes = {
+		/* R8A7790 has one RGB output, two LVDS outputs and one
+		 * (currently unsupported) TCON output.
+		 */
+		[RCAR_DU_OUTPUT_DPAD0] = {
+			.possible_crtcs = BIT(2) | BIT(1) | BIT(0),
+			.encoder_type = DRM_MODE_ENCODER_NONE,
+			.port = 0,
+		},
+		[RCAR_DU_OUTPUT_LVDS0] = {
+			.possible_crtcs = BIT(0),
+			.encoder_type = DRM_MODE_ENCODER_LVDS,
+			.port = 1,
+		},
+		[RCAR_DU_OUTPUT_LVDS1] = {
+			.possible_crtcs = BIT(2) | BIT(1),
+			.encoder_type = DRM_MODE_ENCODER_LVDS,
+			.port = 2,
+		},
+	},
+	.num_lvds = 2,
+};
+
+static const struct rcar_du_device_info rcar_du_r8a7791_info = {
+	.features = RCAR_DU_FEATURE_CRTC_IRQ_CLOCK | RCAR_DU_FEATURE_DEFR8,
+	.num_crtcs = 2,
+	.routes = {
+		/* R8A7791 has one RGB output, one LVDS output and one
+		 * (currently unsupported) TCON output.
+		 */
+		[RCAR_DU_OUTPUT_DPAD0] = {
+			.possible_crtcs = BIT(1),
+			.encoder_type = DRM_MODE_ENCODER_NONE,
+			.port = 0,
+		},
+		[RCAR_DU_OUTPUT_LVDS0] = {
+			.possible_crtcs = BIT(0),
+			.encoder_type = DRM_MODE_ENCODER_LVDS,
+			.port = 1,
+		},
+	},
+	.num_lvds = 1,
+};
+
+static const struct platform_device_id rcar_du_id_table[] = {
+	{ "rcar-du-r8a7779", (kernel_ulong_t)&rcar_du_r8a7779_info },
+	{ "rcar-du-r8a7790", (kernel_ulong_t)&rcar_du_r8a7790_info },
+	{ "rcar-du-r8a7791", (kernel_ulong_t)&rcar_du_r8a7791_info },
+	{ }
+};
+
+MODULE_DEVICE_TABLE(platform, rcar_du_id_table);
+
+static const struct of_device_id rcar_du_of_table[] = {
+	{ .compatible = "renesas,du-r8a7779", .data = &rcar_du_r8a7779_info },
+	{ .compatible = "renesas,du-r8a7790", .data = &rcar_du_r8a7790_info },
+	{ .compatible = "renesas,du-r8a7791", .data = &rcar_du_r8a7791_info },
+	{ }
+};
+
+MODULE_DEVICE_TABLE(of, rcar_du_of_table);
+
+/* -----------------------------------------------------------------------------
  * DRM operations
  */
 
@@ -53,12 +145,12 @@ static int rcar_du_unload(struct drm_device *dev)
 static int rcar_du_load(struct drm_device *dev, unsigned long flags)
 {
 	struct platform_device *pdev = dev->platformdev;
-	struct rcar_du_platform_data *pdata = pdev->dev.platform_data;
+	struct device_node *np = pdev->dev.of_node;
 	struct rcar_du_device *rcdu;
 	struct resource *mem;
 	int ret;
 
-	if (pdata == NULL) {
+	if (np == NULL) {
 		dev_err(dev->dev, "no platform data\n");
 		return -ENODEV;
 	}
@@ -70,8 +162,8 @@ static int rcar_du_load(struct drm_device *dev, unsigned long flags)
 	}
 
 	rcdu->dev = &pdev->dev;
-	rcdu->pdata = pdata;
-	rcdu->info = (struct rcar_du_device_info *)pdev->id_entry->driver_data;
+	rcdu->info = np ? of_match_device(rcar_du_of_table, rcdu->dev)->data
+		   : (void *)platform_get_device_id(pdev)->driver_data;
 	rcdu->ddev = dev;
 	dev->dev_private = rcdu;
 
@@ -158,6 +250,7 @@ static struct drm_driver rcar_du_driver = {
 	.unload			= rcar_du_unload,
 	.preclose		= rcar_du_preclose,
 	.lastclose		= rcar_du_lastclose,
+	.set_busid		= drm_platform_set_busid,
 	.get_vblank_counter	= drm_vblank_count,
 	.enable_vblank		= rcar_du_enable_vblank,
 	.disable_vblank		= rcar_du_disable_vblank,
@@ -231,84 +324,13 @@ static int rcar_du_remove(struct platform_device *pdev)
 	return 0;
 }
 
-static const struct rcar_du_device_info rcar_du_r8a7779_info = {
-	.features = 0,
-	.num_crtcs = 2,
-	.routes = {
-		/* R8A7779 has two RGB outputs and one (currently unsupported)
-		 * TCON output.
-		 */
-		[RCAR_DU_OUTPUT_DPAD0] = {
-			.possible_crtcs = BIT(0),
-			.encoder_type = DRM_MODE_ENCODER_NONE,
-		},
-		[RCAR_DU_OUTPUT_DPAD1] = {
-			.possible_crtcs = BIT(1) | BIT(0),
-			.encoder_type = DRM_MODE_ENCODER_NONE,
-		},
-	},
-	.num_lvds = 0,
-};
-
-static const struct rcar_du_device_info rcar_du_r8a7790_info = {
-	.features = RCAR_DU_FEATURE_CRTC_IRQ_CLOCK | RCAR_DU_FEATURE_DEFR8,
-	.quirks = RCAR_DU_QUIRK_ALIGN_128B | RCAR_DU_QUIRK_LVDS_LANES,
-	.num_crtcs = 3,
-	.routes = {
-		/* R8A7790 has one RGB output, two LVDS outputs and one
-		 * (currently unsupported) TCON output.
-		 */
-		[RCAR_DU_OUTPUT_DPAD0] = {
-			.possible_crtcs = BIT(2) | BIT(1) | BIT(0),
-			.encoder_type = DRM_MODE_ENCODER_NONE,
-		},
-		[RCAR_DU_OUTPUT_LVDS0] = {
-			.possible_crtcs = BIT(0),
-			.encoder_type = DRM_MODE_ENCODER_LVDS,
-		},
-		[RCAR_DU_OUTPUT_LVDS1] = {
-			.possible_crtcs = BIT(2) | BIT(1),
-			.encoder_type = DRM_MODE_ENCODER_LVDS,
-		},
-	},
-	.num_lvds = 2,
-};
-
-static const struct rcar_du_device_info rcar_du_r8a7791_info = {
-	.features = RCAR_DU_FEATURE_CRTC_IRQ_CLOCK | RCAR_DU_FEATURE_DEFR8,
-	.num_crtcs = 2,
-	.routes = {
-		/* R8A7791 has one RGB output, one LVDS output and one
-		 * (currently unsupported) TCON output.
-		 */
-		[RCAR_DU_OUTPUT_DPAD0] = {
-			.possible_crtcs = BIT(1),
-			.encoder_type = DRM_MODE_ENCODER_NONE,
-		},
-		[RCAR_DU_OUTPUT_LVDS0] = {
-			.possible_crtcs = BIT(0),
-			.encoder_type = DRM_MODE_ENCODER_LVDS,
-		},
-	},
-	.num_lvds = 1,
-};
-
-static const struct platform_device_id rcar_du_id_table[] = {
-	{ "rcar-du-r8a7779", (kernel_ulong_t)&rcar_du_r8a7779_info },
-	{ "rcar-du-r8a7790", (kernel_ulong_t)&rcar_du_r8a7790_info },
-	{ "rcar-du-r8a7791", (kernel_ulong_t)&rcar_du_r8a7791_info },
-	{ }
-};
-
-MODULE_DEVICE_TABLE(platform, rcar_du_id_table);
-
 static struct platform_driver rcar_du_platform_driver = {
 	.probe		= rcar_du_probe,
 	.remove		= rcar_du_remove,
 	.driver		= {
-		.owner	= THIS_MODULE,
 		.name	= "rcar-du",
 		.pm	= &rcar_du_pm_ops,
+		.of_match_table = rcar_du_of_table,
 	},
 	.id_table	= rcar_du_id_table,
 };
diff --git a/drivers/gpu/drm/rcar-du/rcar_du_drv.h b/drivers/gpu/drm/rcar-du/rcar_du_drv.h
index e31b735d3f25..0a724669f02d 100644
--- a/drivers/gpu/drm/rcar-du/rcar_du_drv.h
+++ b/drivers/gpu/drm/rcar-du/rcar_du_drv.h
@@ -1,7 +1,7 @@
 /*
  * rcar_du_drv.h  --  R-Car Display Unit DRM driver
  *
- * Copyright (C) 2013 Renesas Corporation
+ * Copyright (C) 2013-2014 Renesas Electronics Corporation
  *
  * Contact: Laurent Pinchart (laurent.pinchart@ideasonboard.com)
  *
@@ -15,7 +15,6 @@
 #define __RCAR_DU_DRV_H__
 
 #include <linux/kernel.h>
-#include <linux/platform_data/rcar-du.h>
 
 #include "rcar_du_crtc.h"
 #include "rcar_du_group.h"
@@ -37,6 +36,7 @@ struct rcar_du_lvdsenc;
  * struct rcar_du_output_routing - Output routing specification
  * @possible_crtcs: bitmask of possible CRTCs for the output
  * @encoder_type: DRM type of the internal encoder associated with the output
+ * @port: device tree port number corresponding to this output route
  *
  * The DU has 5 possible outputs (DPAD0/1, LVDS0/1, TCON). Output routing data
  * specify the valid SoC outputs, which CRTCs can drive the output, and the type
@@ -45,6 +45,7 @@ struct rcar_du_lvdsenc;
 struct rcar_du_output_routing {
 	unsigned int possible_crtcs;
 	unsigned int encoder_type;
+	unsigned int port;
 };
 
 /*
@@ -65,7 +66,6 @@ struct rcar_du_device_info {
 
 struct rcar_du_device {
 	struct device *dev;
-	const struct rcar_du_platform_data *pdata;
 	const struct rcar_du_device_info *info;
 
 	void __iomem *mmio;
diff --git a/drivers/gpu/drm/rcar-du/rcar_du_encoder.c b/drivers/gpu/drm/rcar-du/rcar_du_encoder.c
index 3daa7a168dc6..34a122a39664 100644
--- a/drivers/gpu/drm/rcar-du/rcar_du_encoder.c
+++ b/drivers/gpu/drm/rcar-du/rcar_du_encoder.c
@@ -1,7 +1,7 @@
 /*
  * rcar_du_encoder.c  --  R-Car Display Unit Encoder
  *
- * Copyright (C) 2013 Renesas Corporation
+ * Copyright (C) 2013-2014 Renesas Electronics Corporation
  *
  * Contact: Laurent Pinchart (laurent.pinchart@ideasonboard.com)
  *
@@ -19,6 +19,8 @@
 
 #include "rcar_du_drv.h"
 #include "rcar_du_encoder.h"
+#include "rcar_du_hdmicon.h"
+#include "rcar_du_hdmienc.h"
 #include "rcar_du_kms.h"
 #include "rcar_du_lvdscon.h"
 #include "rcar_du_lvdsenc.h"
@@ -33,7 +35,7 @@ rcar_du_connector_best_encoder(struct drm_connector *connector)
 {
 	struct rcar_du_connector *rcon = to_rcar_connector(connector);
 
-	return &rcon->encoder->encoder;
+	return rcar_encoder_to_drm_encoder(rcon->encoder);
 }
 
 /* -----------------------------------------------------------------------------
@@ -142,9 +144,11 @@ static const struct drm_encoder_funcs encoder_funcs = {
 int rcar_du_encoder_init(struct rcar_du_device *rcdu,
 			 enum rcar_du_encoder_type type,
 			 enum rcar_du_output output,
-			 const struct rcar_du_encoder_data *data)
+			 struct device_node *enc_node,
+			 struct device_node *con_node)
 {
 	struct rcar_du_encoder *renc;
+	struct drm_encoder *encoder;
 	unsigned int encoder_type;
 	int ret;
 
@@ -153,6 +157,7 @@ int rcar_du_encoder_init(struct rcar_du_device *rcdu,
 		return -ENOMEM;
 
 	renc->output = output;
+	encoder = rcar_encoder_to_drm_encoder(renc);
 
 	switch (output) {
 	case RCAR_DU_OUTPUT_LVDS0:
@@ -174,6 +179,9 @@ int rcar_du_encoder_init(struct rcar_du_device *rcdu,
 	case RCAR_DU_ENCODER_LVDS:
 		encoder_type = DRM_MODE_ENCODER_LVDS;
 		break;
+	case RCAR_DU_ENCODER_HDMI:
+		encoder_type = DRM_MODE_ENCODER_TMDS;
+		break;
 	case RCAR_DU_ENCODER_NONE:
 	default:
 		/* No external encoder, use the internal encoder type. */
@@ -181,21 +189,35 @@ int rcar_du_encoder_init(struct rcar_du_device *rcdu,
 		break;
 	}
 
-	ret = drm_encoder_init(rcdu->ddev, &renc->encoder, &encoder_funcs,
-			       encoder_type);
-	if (ret < 0)
-		return ret;
+	if (type == RCAR_DU_ENCODER_HDMI) {
+		if (renc->lvds) {
+			dev_err(rcdu->dev,
+				"Chaining LVDS and HDMI encoders not supported\n");
+			return -EINVAL;
+		}
 
-	drm_encoder_helper_add(&renc->encoder, &encoder_helper_funcs);
+		ret = rcar_du_hdmienc_init(rcdu, renc, enc_node);
+		if (ret < 0)
+			return ret;
+	} else {
+		ret = drm_encoder_init(rcdu->ddev, encoder, &encoder_funcs,
+				       encoder_type);
+		if (ret < 0)
+			return ret;
+
+		drm_encoder_helper_add(encoder, &encoder_helper_funcs);
+	}
 
 	switch (encoder_type) {
 	case DRM_MODE_ENCODER_LVDS:
-		return rcar_du_lvds_connector_init(rcdu, renc,
-						   &data->connector.lvds.panel);
+		return rcar_du_lvds_connector_init(rcdu, renc, con_node);
 
 	case DRM_MODE_ENCODER_DAC:
 		return rcar_du_vga_connector_init(rcdu, renc);
 
+	case DRM_MODE_ENCODER_TMDS:
+		return rcar_du_hdmi_connector_init(rcdu, renc);
+
 	default:
 		return -EINVAL;
 	}
diff --git a/drivers/gpu/drm/rcar-du/rcar_du_encoder.h b/drivers/gpu/drm/rcar-du/rcar_du_encoder.h
index 0e5a65e45d0e..719b6f2a031c 100644
--- a/drivers/gpu/drm/rcar-du/rcar_du_encoder.h
+++ b/drivers/gpu/drm/rcar-du/rcar_du_encoder.h
@@ -1,7 +1,7 @@
 /*
  * rcar_du_encoder.h  --  R-Car Display Unit Encoder
  *
- * Copyright (C) 2013 Renesas Corporation
+ * Copyright (C) 2013-2014 Renesas Electronics Corporation
  *
  * Contact: Laurent Pinchart (laurent.pinchart@ideasonboard.com)
  *
@@ -14,21 +14,32 @@
 #ifndef __RCAR_DU_ENCODER_H__
 #define __RCAR_DU_ENCODER_H__
 
-#include <linux/platform_data/rcar-du.h>
-
 #include <drm/drm_crtc.h>
+#include <drm/drm_encoder_slave.h>
 
 struct rcar_du_device;
+struct rcar_du_hdmienc;
 struct rcar_du_lvdsenc;
 
+enum rcar_du_encoder_type {
+	RCAR_DU_ENCODER_UNUSED = 0,
+	RCAR_DU_ENCODER_NONE,
+	RCAR_DU_ENCODER_VGA,
+	RCAR_DU_ENCODER_LVDS,
+	RCAR_DU_ENCODER_HDMI,
+};
+
 struct rcar_du_encoder {
-	struct drm_encoder encoder;
+	struct drm_encoder_slave slave;
 	enum rcar_du_output output;
+	struct rcar_du_hdmienc *hdmi;
 	struct rcar_du_lvdsenc *lvds;
 };
 
 #define to_rcar_encoder(e) \
-	container_of(e, struct rcar_du_encoder, encoder)
+	container_of(e, struct rcar_du_encoder, slave.base)
+
+#define rcar_encoder_to_drm_encoder(e)	(&(e)->slave.base)
 
 struct rcar_du_connector {
 	struct drm_connector connector;
@@ -44,6 +55,7 @@ rcar_du_connector_best_encoder(struct drm_connector *connector);
 int rcar_du_encoder_init(struct rcar_du_device *rcdu,
 			 enum rcar_du_encoder_type type,
 			 enum rcar_du_output output,
-			 const struct rcar_du_encoder_data *data);
+			 struct device_node *enc_node,
+			 struct device_node *con_node);
 
 #endif /* __RCAR_DU_ENCODER_H__ */
diff --git a/drivers/gpu/drm/rcar-du/rcar_du_group.c b/drivers/gpu/drm/rcar-du/rcar_du_group.c
index eb53cd97e8c6..4e7614b145db 100644
--- a/drivers/gpu/drm/rcar-du/rcar_du_group.c
+++ b/drivers/gpu/drm/rcar-du/rcar_du_group.c
@@ -1,7 +1,7 @@
 /*
  * rcar_du_group.c  --  R-Car Display Unit Channels Pair
  *
- * Copyright (C) 2013 Renesas Corporation
+ * Copyright (C) 2013-2014 Renesas Electronics Corporation
  *
  * Contact: Laurent Pinchart (laurent.pinchart@ideasonboard.com)
  *
diff --git a/drivers/gpu/drm/rcar-du/rcar_du_group.h b/drivers/gpu/drm/rcar-du/rcar_du_group.h
index 5025930972ec..0c38cdcda4ca 100644
--- a/drivers/gpu/drm/rcar-du/rcar_du_group.h
+++ b/drivers/gpu/drm/rcar-du/rcar_du_group.h
@@ -1,7 +1,7 @@
 /*
  * rcar_du_group.c  --  R-Car Display Unit Planes and CRTCs Group
  *
- * Copyright (C) 2013 Renesas Corporation
+ * Copyright (C) 2013-2014 Renesas Electronics Corporation
  *
  * Contact: Laurent Pinchart (laurent.pinchart@ideasonboard.com)
  *
diff --git a/drivers/gpu/drm/rcar-du/rcar_du_hdmicon.c b/drivers/gpu/drm/rcar-du/rcar_du_hdmicon.c
new file mode 100644
index 000000000000..4d7d4dd46d26
--- /dev/null
+++ b/drivers/gpu/drm/rcar-du/rcar_du_hdmicon.c
@@ -0,0 +1,121 @@
+/*
+ * R-Car Display Unit HDMI Connector
+ *
+ * Copyright (C) 2014 Renesas Electronics Corporation
+ *
+ * Contact: Laurent Pinchart (laurent.pinchart@ideasonboard.com)
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ */
+
+#include <drm/drmP.h>
+#include <drm/drm_crtc.h>
+#include <drm/drm_crtc_helper.h>
+#include <drm/drm_encoder_slave.h>
+
+#include "rcar_du_drv.h"
+#include "rcar_du_encoder.h"
+#include "rcar_du_hdmicon.h"
+#include "rcar_du_kms.h"
+
+#define to_slave_funcs(e)	(to_rcar_encoder(e)->slave.slave_funcs)
+
+static int rcar_du_hdmi_connector_get_modes(struct drm_connector *connector)
+{
+	struct rcar_du_connector *con = to_rcar_connector(connector);
+	struct drm_encoder *encoder = rcar_encoder_to_drm_encoder(con->encoder);
+	struct drm_encoder_slave_funcs *sfuncs = to_slave_funcs(encoder);
+
+	if (sfuncs->get_modes == NULL)
+		return 0;
+
+	return sfuncs->get_modes(encoder, connector);
+}
+
+static int rcar_du_hdmi_connector_mode_valid(struct drm_connector *connector,
+					     struct drm_display_mode *mode)
+{
+	struct rcar_du_connector *con = to_rcar_connector(connector);
+	struct drm_encoder *encoder = rcar_encoder_to_drm_encoder(con->encoder);
+	struct drm_encoder_slave_funcs *sfuncs = to_slave_funcs(encoder);
+
+	if (sfuncs->mode_valid == NULL)
+		return MODE_OK;
+
+	return sfuncs->mode_valid(encoder, mode);
+}
+
+static const struct drm_connector_helper_funcs connector_helper_funcs = {
+	.get_modes = rcar_du_hdmi_connector_get_modes,
+	.mode_valid = rcar_du_hdmi_connector_mode_valid,
+	.best_encoder = rcar_du_connector_best_encoder,
+};
+
+static void rcar_du_hdmi_connector_destroy(struct drm_connector *connector)
+{
+	drm_connector_unregister(connector);
+	drm_connector_cleanup(connector);
+}
+
+static enum drm_connector_status
+rcar_du_hdmi_connector_detect(struct drm_connector *connector, bool force)
+{
+	struct rcar_du_connector *con = to_rcar_connector(connector);
+	struct drm_encoder *encoder = rcar_encoder_to_drm_encoder(con->encoder);
+	struct drm_encoder_slave_funcs *sfuncs = to_slave_funcs(encoder);
+
+	if (sfuncs->detect == NULL)
+		return connector_status_unknown;
+
+	return sfuncs->detect(encoder, connector);
+}
+
+static const struct drm_connector_funcs connector_funcs = {
+	.dpms = drm_helper_connector_dpms,
+	.detect = rcar_du_hdmi_connector_detect,
+	.fill_modes = drm_helper_probe_single_connector_modes,
+	.destroy = rcar_du_hdmi_connector_destroy,
+};
+
+int rcar_du_hdmi_connector_init(struct rcar_du_device *rcdu,
+				struct rcar_du_encoder *renc)
+{
+	struct drm_encoder *encoder = rcar_encoder_to_drm_encoder(renc);
+	struct rcar_du_connector *rcon;
+	struct drm_connector *connector;
+	int ret;
+
+	rcon = devm_kzalloc(rcdu->dev, sizeof(*rcon), GFP_KERNEL);
+	if (rcon == NULL)
+		return -ENOMEM;
+
+	connector = &rcon->connector;
+	connector->display_info.width_mm = 0;
+	connector->display_info.height_mm = 0;
+
+	ret = drm_connector_init(rcdu->ddev, connector, &connector_funcs,
+				 DRM_MODE_CONNECTOR_HDMIA);
+	if (ret < 0)
+		return ret;
+
+	drm_connector_helper_add(connector, &connector_helper_funcs);
+	ret = drm_connector_register(connector);
+	if (ret < 0)
+		return ret;
+
+	drm_helper_connector_dpms(connector, DRM_MODE_DPMS_OFF);
+	drm_object_property_set_value(&connector->base,
+		rcdu->ddev->mode_config.dpms_property, DRM_MODE_DPMS_OFF);
+
+	ret = drm_mode_connector_attach_encoder(connector, encoder);
+	if (ret < 0)
+		return ret;
+
+	connector->encoder = encoder;
+	rcon->encoder = renc;
+
+	return 0;
+}
diff --git a/drivers/gpu/drm/rcar-du/rcar_du_hdmicon.h b/drivers/gpu/drm/rcar-du/rcar_du_hdmicon.h
new file mode 100644
index 000000000000..87daa949227f
--- /dev/null
+++ b/drivers/gpu/drm/rcar-du/rcar_du_hdmicon.h
@@ -0,0 +1,31 @@
+/*
+ * R-Car Display Unit HDMI Connector
+ *
+ * Copyright (C) 2014 Renesas Electronics Corporation
+ *
+ * Contact: Laurent Pinchart (laurent.pinchart@ideasonboard.com)
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ */
+
+#ifndef __RCAR_DU_HDMICON_H__
+#define __RCAR_DU_HDMICON_H__
+
+struct rcar_du_device;
+struct rcar_du_encoder;
+
+#if IS_ENABLED(CONFIG_DRM_RCAR_HDMI)
+int rcar_du_hdmi_connector_init(struct rcar_du_device *rcdu,
+				struct rcar_du_encoder *renc);
+#else
+static inline int rcar_du_hdmi_connector_init(struct rcar_du_device *rcdu,
+					      struct rcar_du_encoder *renc)
+{
+	return -ENOSYS;
+}
+#endif
+
+#endif /* __RCAR_DU_HDMICON_H__ */
diff --git a/drivers/gpu/drm/rcar-du/rcar_du_hdmienc.c b/drivers/gpu/drm/rcar-du/rcar_du_hdmienc.c
new file mode 100644
index 000000000000..359bc999a9c8
--- /dev/null
+++ b/drivers/gpu/drm/rcar-du/rcar_du_hdmienc.c
@@ -0,0 +1,151 @@
+/*
+ * R-Car Display Unit HDMI Encoder
+ *
+ * Copyright (C) 2014 Renesas Electronics Corporation
+ *
+ * Contact: Laurent Pinchart (laurent.pinchart@ideasonboard.com)
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ */
+
+#include <linux/slab.h>
+
+#include <drm/drmP.h>
+#include <drm/drm_crtc.h>
+#include <drm/drm_crtc_helper.h>
+#include <drm/drm_encoder_slave.h>
+
+#include "rcar_du_drv.h"
+#include "rcar_du_encoder.h"
+#include "rcar_du_hdmienc.h"
+
+struct rcar_du_hdmienc {
+	struct rcar_du_encoder *renc;
+	struct device *dev;
+	int dpms;
+};
+
+#define to_rcar_hdmienc(e)	(to_rcar_encoder(e)->hdmi)
+#define to_slave_funcs(e)	(to_rcar_encoder(e)->slave.slave_funcs)
+
+static void rcar_du_hdmienc_dpms(struct drm_encoder *encoder, int mode)
+{
+	struct rcar_du_hdmienc *hdmienc = to_rcar_hdmienc(encoder);
+	struct drm_encoder_slave_funcs *sfuncs = to_slave_funcs(encoder);
+
+	if (hdmienc->dpms == mode)
+		return;
+
+	if (sfuncs->dpms)
+		sfuncs->dpms(encoder, mode);
+
+	hdmienc->dpms = mode;
+}
+
+static bool rcar_du_hdmienc_mode_fixup(struct drm_encoder *encoder,
+				       const struct drm_display_mode *mode,
+				       struct drm_display_mode *adjusted_mode)
+{
+	struct drm_encoder_slave_funcs *sfuncs = to_slave_funcs(encoder);
+
+	if (sfuncs->mode_fixup == NULL)
+		return true;
+
+	return sfuncs->mode_fixup(encoder, mode, adjusted_mode);
+}
+
+static void rcar_du_hdmienc_mode_prepare(struct drm_encoder *encoder)
+{
+	rcar_du_hdmienc_dpms(encoder, DRM_MODE_DPMS_OFF);
+}
+
+static void rcar_du_hdmienc_mode_commit(struct drm_encoder *encoder)
+{
+	rcar_du_hdmienc_dpms(encoder, DRM_MODE_DPMS_ON);
+}
+
+static void rcar_du_hdmienc_mode_set(struct drm_encoder *encoder,
+				     struct drm_display_mode *mode,
+				     struct drm_display_mode *adjusted_mode)
+{
+	struct rcar_du_hdmienc *hdmienc = to_rcar_hdmienc(encoder);
+	struct drm_encoder_slave_funcs *sfuncs = to_slave_funcs(encoder);
+
+	if (sfuncs->mode_set)
+		sfuncs->mode_set(encoder, mode, adjusted_mode);
+
+	rcar_du_crtc_route_output(encoder->crtc, hdmienc->renc->output);
+}
+
+static const struct drm_encoder_helper_funcs encoder_helper_funcs = {
+	.dpms = rcar_du_hdmienc_dpms,
+	.mode_fixup = rcar_du_hdmienc_mode_fixup,
+	.prepare = rcar_du_hdmienc_mode_prepare,
+	.commit = rcar_du_hdmienc_mode_commit,
+	.mode_set = rcar_du_hdmienc_mode_set,
+};
+
+static void rcar_du_hdmienc_cleanup(struct drm_encoder *encoder)
+{
+	struct rcar_du_hdmienc *hdmienc = to_rcar_hdmienc(encoder);
+
+	rcar_du_hdmienc_dpms(encoder, DRM_MODE_DPMS_OFF);
+
+	drm_encoder_cleanup(encoder);
+	put_device(hdmienc->dev);
+}
+
+static const struct drm_encoder_funcs encoder_funcs = {
+	.destroy = rcar_du_hdmienc_cleanup,
+};
+
+int rcar_du_hdmienc_init(struct rcar_du_device *rcdu,
+			 struct rcar_du_encoder *renc, struct device_node *np)
+{
+	struct drm_encoder *encoder = rcar_encoder_to_drm_encoder(renc);
+	struct drm_i2c_encoder_driver *driver;
+	struct i2c_client *i2c_slave;
+	struct rcar_du_hdmienc *hdmienc;
+	int ret;
+
+	hdmienc = devm_kzalloc(rcdu->dev, sizeof(*hdmienc), GFP_KERNEL);
+	if (hdmienc == NULL)
+		return -ENOMEM;
+
+	/* Locate the slave I2C device and driver. */
+	i2c_slave = of_find_i2c_device_by_node(np);
+	if (!i2c_slave || !i2c_get_clientdata(i2c_slave))
+		return -EPROBE_DEFER;
+
+	hdmienc->dev = &i2c_slave->dev;
+
+	if (hdmienc->dev->driver == NULL) {
+		ret = -EPROBE_DEFER;
+		goto error;
+	}
+
+	/* Initialize the slave encoder. */
+	driver = to_drm_i2c_encoder_driver(to_i2c_driver(hdmienc->dev->driver));
+	ret = driver->encoder_init(i2c_slave, rcdu->ddev, &renc->slave);
+	if (ret < 0)
+		goto error;
+
+	ret = drm_encoder_init(rcdu->ddev, encoder, &encoder_funcs,
+			       DRM_MODE_ENCODER_TMDS);
+	if (ret < 0)
+		goto error;
+
+	drm_encoder_helper_add(encoder, &encoder_helper_funcs);
+
+	renc->hdmi = hdmienc;
+	hdmienc->renc = renc;
+
+	return 0;
+
+error:
+	put_device(hdmienc->dev);
+	return ret;
+}
diff --git a/drivers/gpu/drm/rcar-du/rcar_du_hdmienc.h b/drivers/gpu/drm/rcar-du/rcar_du_hdmienc.h
new file mode 100644
index 000000000000..2ff0128ac8e1
--- /dev/null
+++ b/drivers/gpu/drm/rcar-du/rcar_du_hdmienc.h
@@ -0,0 +1,35 @@
+/*
+ * R-Car Display Unit HDMI Encoder
+ *
+ * Copyright (C) 2014 Renesas Electronics Corporation
+ *
+ * Contact: Laurent Pinchart (laurent.pinchart@ideasonboard.com)
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ */
+
+#ifndef __RCAR_DU_HDMIENC_H__
+#define __RCAR_DU_HDMIENC_H__
+
+#include <linux/module.h>
+
+struct device_node;
+struct rcar_du_device;
+struct rcar_du_encoder;
+
+#if IS_ENABLED(CONFIG_DRM_RCAR_HDMI)
+int rcar_du_hdmienc_init(struct rcar_du_device *rcdu,
+			 struct rcar_du_encoder *renc, struct device_node *np);
+#else
+static inline int rcar_du_hdmienc_init(struct rcar_du_device *rcdu,
+				       struct rcar_du_encoder *renc,
+				       struct device_node *np)
+{
+	return -ENOSYS;
+}
+#endif
+
+#endif /* __RCAR_DU_HDMIENC_H__ */
diff --git a/drivers/gpu/drm/rcar-du/rcar_du_kms.c b/drivers/gpu/drm/rcar-du/rcar_du_kms.c
index 76026104d000..0c5ee616b5a3 100644
--- a/drivers/gpu/drm/rcar-du/rcar_du_kms.c
+++ b/drivers/gpu/drm/rcar-du/rcar_du_kms.c
@@ -1,7 +1,7 @@
 /*
  * rcar_du_kms.c  --  R-Car Display Unit Mode Setting
  *
- * Copyright (C) 2013 Renesas Corporation
+ * Copyright (C) 2013-2014 Renesas Electronics Corporation
  *
  * Contact: Laurent Pinchart (laurent.pinchart@ideasonboard.com)
  *
@@ -17,6 +17,8 @@
 #include <drm/drm_fb_cma_helper.h>
 #include <drm/drm_gem_cma_helper.h>
 
+#include <linux/of_graph.h>
+
 #include "rcar_du_crtc.h"
 #include "rcar_du_drv.h"
 #include "rcar_du_encoder.h"
@@ -124,9 +126,9 @@ int rcar_du_dumb_create(struct drm_file *file, struct drm_device *dev,
 	else
 		align = 16 * args->bpp / 8;
 
-	args->pitch = roundup(max(args->pitch, min_pitch), align);
+	args->pitch = roundup(min_pitch, align);
 
-	return drm_gem_cma_dumb_create(file, dev, args);
+	return drm_gem_cma_dumb_create_internal(file, dev, args);
 }
 
 static struct drm_framebuffer *
@@ -188,6 +190,172 @@ static const struct drm_mode_config_funcs rcar_du_mode_config_funcs = {
 	.output_poll_changed = rcar_du_output_poll_changed,
 };
 
+static int rcar_du_encoders_init_one(struct rcar_du_device *rcdu,
+				     enum rcar_du_output output,
+				     struct of_endpoint *ep)
+{
+	static const struct {
+		const char *compatible;
+		enum rcar_du_encoder_type type;
+	} encoders[] = {
+		{ "adi,adv7123", RCAR_DU_ENCODER_VGA },
+		{ "adi,adv7511w", RCAR_DU_ENCODER_HDMI },
+		{ "thine,thc63lvdm83d", RCAR_DU_ENCODER_LVDS },
+	};
+
+	enum rcar_du_encoder_type enc_type = RCAR_DU_ENCODER_NONE;
+	struct device_node *connector = NULL;
+	struct device_node *encoder = NULL;
+	struct device_node *prev = NULL;
+	struct device_node *entity_ep_node;
+	struct device_node *entity;
+	int ret;
+
+	/*
+	 * Locate the connected entity and infer its type from the number of
+	 * endpoints.
+	 */
+	entity = of_graph_get_remote_port_parent(ep->local_node);
+	if (!entity) {
+		dev_dbg(rcdu->dev, "unconnected endpoint %s, skipping\n",
+			ep->local_node->full_name);
+		return 0;
+	}
+
+	entity_ep_node = of_parse_phandle(ep->local_node, "remote-endpoint", 0);
+
+	while (1) {
+		struct device_node *ep_node;
+
+		ep_node = of_graph_get_next_endpoint(entity, prev);
+		of_node_put(prev);
+		prev = ep_node;
+
+		if (!ep_node)
+			break;
+
+		if (ep_node == entity_ep_node)
+			continue;
+
+		/*
+		 * We've found one endpoint other than the input, this must
+		 * be an encoder. Locate the connector.
+		 */
+		encoder = entity;
+		connector = of_graph_get_remote_port_parent(ep_node);
+		of_node_put(ep_node);
+
+		if (!connector) {
+			dev_warn(rcdu->dev,
+				 "no connector for encoder %s, skipping\n",
+				 encoder->full_name);
+			of_node_put(entity_ep_node);
+			of_node_put(encoder);
+			return 0;
+		}
+
+		break;
+	}
+
+	of_node_put(entity_ep_node);
+
+	if (encoder) {
+		/*
+		 * If an encoder has been found, get its type based on its
+		 * compatible string.
+		 */
+		unsigned int i;
+
+		for (i = 0; i < ARRAY_SIZE(encoders); ++i) {
+			if (of_device_is_compatible(encoder,
+						    encoders[i].compatible)) {
+				enc_type = encoders[i].type;
+				break;
+			}
+		}
+
+		if (i == ARRAY_SIZE(encoders)) {
+			dev_warn(rcdu->dev,
+				 "unknown encoder type for %s, skipping\n",
+				 encoder->full_name);
+			of_node_put(encoder);
+			of_node_put(connector);
+			return 0;
+		}
+	} else {
+		/*
+		 * If no encoder has been found the entity must be the
+		 * connector.
+		 */
+		connector = entity;
+	}
+
+	ret = rcar_du_encoder_init(rcdu, enc_type, output, encoder, connector);
+	of_node_put(encoder);
+	of_node_put(connector);
+
+	return ret < 0 ? ret : 1;
+}
+
+static int rcar_du_encoders_init(struct rcar_du_device *rcdu)
+{
+	struct device_node *np = rcdu->dev->of_node;
+	struct device_node *prev = NULL;
+	unsigned int num_encoders = 0;
+
+	/*
+	 * Iterate over the endpoints and create one encoder for each output
+	 * pipeline.
+	 */
+	while (1) {
+		struct device_node *ep_node;
+		enum rcar_du_output output;
+		struct of_endpoint ep;
+		unsigned int i;
+		int ret;
+
+		ep_node = of_graph_get_next_endpoint(np, prev);
+		of_node_put(prev);
+		prev = ep_node;
+
+		if (ep_node == NULL)
+			break;
+
+		ret = of_graph_parse_endpoint(ep_node, &ep);
+		if (ret < 0) {
+			of_node_put(ep_node);
+			return ret;
+		}
+
+		/* Find the output route corresponding to the port number. */
+		for (i = 0; i < RCAR_DU_OUTPUT_MAX; ++i) {
+			if (rcdu->info->routes[i].possible_crtcs &&
+			    rcdu->info->routes[i].port == ep.port) {
+				output = i;
+				break;
+			}
+		}
+
+		if (i == RCAR_DU_OUTPUT_MAX) {
+			dev_warn(rcdu->dev,
+				 "port %u references unexisting output, skipping\n",
+				 ep.port);
+			continue;
+		}
+
+		/* Process the output pipeline. */
+		ret = rcar_du_encoders_init_one(rcdu, output, &ep);
+		if (ret < 0) {
+			of_node_put(ep_node);
+			return ret;
+		}
+
+		num_encoders += ret;
+	}
+
+	return num_encoders;
+}
+
 int rcar_du_modeset_init(struct rcar_du_device *rcdu)
 {
 	static const unsigned int mmio_offsets[] = {
@@ -197,6 +365,7 @@ int rcar_du_modeset_init(struct rcar_du_device *rcdu)
 	struct drm_device *dev = rcdu->ddev;
 	struct drm_encoder *encoder;
 	struct drm_fbdev_cma *fbdev;
+	unsigned int num_encoders;
 	unsigned int num_groups;
 	unsigned int i;
 	int ret;
@@ -240,28 +409,11 @@ int rcar_du_modeset_init(struct rcar_du_device *rcdu)
 	if (ret < 0)
 		return ret;
 
-	for (i = 0; i < rcdu->pdata->num_encoders; ++i) {
-		const struct rcar_du_encoder_data *pdata =
-			&rcdu->pdata->encoders[i];
-		const struct rcar_du_output_routing *route =
-			&rcdu->info->routes[pdata->output];
-
-		if (pdata->type == RCAR_DU_ENCODER_UNUSED)
-			continue;
-
-		if (pdata->output >= RCAR_DU_OUTPUT_MAX ||
-		    route->possible_crtcs == 0) {
-			dev_warn(rcdu->dev,
-				 "encoder %u references unexisting output %u, skipping\n",
-				 i, pdata->output);
-			continue;
-		}
+	ret = rcar_du_encoders_init(rcdu);
+	if (ret < 0)
+		return ret;
 
-		ret = rcar_du_encoder_init(rcdu, pdata->type, pdata->output,
-					   pdata);
-		if (ret < 0)
-			return ret;
-	}
+	num_encoders = ret;
 
 	/* Set the possible CRTCs and possible clones. There's always at least
 	 * one way for all encoders to clone each other, set all bits in the
@@ -273,7 +425,7 @@ int rcar_du_modeset_init(struct rcar_du_device *rcdu)
 			&rcdu->info->routes[renc->output];
 
 		encoder->possible_crtcs = route->possible_crtcs;
-		encoder->possible_clones = (1 << rcdu->pdata->num_encoders) - 1;
+		encoder->possible_clones = (1 << num_encoders) - 1;
 	}
 
 	/* Now that the CRTCs have been initialized register the planes. */
diff --git a/drivers/gpu/drm/rcar-du/rcar_du_kms.h b/drivers/gpu/drm/rcar-du/rcar_du_kms.h
index 5750e6af5655..07951d5fe38b 100644
--- a/drivers/gpu/drm/rcar-du/rcar_du_kms.h
+++ b/drivers/gpu/drm/rcar-du/rcar_du_kms.h
@@ -1,7 +1,7 @@
 /*
  * rcar_du_kms.h  --  R-Car Display Unit Mode Setting
  *
- * Copyright (C) 2013 Renesas Corporation
+ * Copyright (C) 2013-2014 Renesas Electronics Corporation
  *
  * Contact: Laurent Pinchart (laurent.pinchart@ideasonboard.com)
  *
diff --git a/drivers/gpu/drm/rcar-du/rcar_du_lvdscon.c b/drivers/gpu/drm/rcar-du/rcar_du_lvdscon.c
index 21426bd234eb..6d9811c052c4 100644
--- a/drivers/gpu/drm/rcar-du/rcar_du_lvdscon.c
+++ b/drivers/gpu/drm/rcar-du/rcar_du_lvdscon.c
@@ -1,7 +1,7 @@
 /*
  * rcar_du_lvdscon.c  --  R-Car Display Unit LVDS Connector
  *
- * Copyright (C) 2013 Renesas Corporation
+ * Copyright (C) 2013-2014 Renesas Electronics Corporation
  *
  * Contact: Laurent Pinchart (laurent.pinchart@ideasonboard.com)
  *
@@ -15,6 +15,10 @@
 #include <drm/drm_crtc.h>
 #include <drm/drm_crtc_helper.h>
 
+#include <video/display_timing.h>
+#include <video/of_display_timing.h>
+#include <video/videomode.h>
+
 #include "rcar_du_drv.h"
 #include "rcar_du_encoder.h"
 #include "rcar_du_kms.h"
@@ -23,7 +27,11 @@
 struct rcar_du_lvds_connector {
 	struct rcar_du_connector connector;
 
-	const struct rcar_du_panel_data *panel;
+	struct {
+		unsigned int width_mm;		/* Panel width in mm */
+		unsigned int height_mm;		/* Panel height in mm */
+		struct videomode mode;
+	} panel;
 };
 
 #define to_rcar_lvds_connector(c) \
@@ -40,18 +48,9 @@ static int rcar_du_lvds_connector_get_modes(struct drm_connector *connector)
 		return 0;
 
 	mode->type = DRM_MODE_TYPE_PREFERRED | DRM_MODE_TYPE_DRIVER;
-	mode->clock = lvdscon->panel->mode.clock;
-	mode->hdisplay = lvdscon->panel->mode.hdisplay;
-	mode->hsync_start = lvdscon->panel->mode.hsync_start;
-	mode->hsync_end = lvdscon->panel->mode.hsync_end;
-	mode->htotal = lvdscon->panel->mode.htotal;
-	mode->vdisplay = lvdscon->panel->mode.vdisplay;
-	mode->vsync_start = lvdscon->panel->mode.vsync_start;
-	mode->vsync_end = lvdscon->panel->mode.vsync_end;
-	mode->vtotal = lvdscon->panel->mode.vtotal;
-	mode->flags = lvdscon->panel->mode.flags;
-
-	drm_mode_set_name(mode);
+
+	drm_display_mode_from_videomode(&lvdscon->panel.mode, mode);
+
 	drm_mode_probed_add(connector, mode);
 
 	return 1;
@@ -83,21 +82,30 @@ static const struct drm_connector_funcs connector_funcs = {
 
 int rcar_du_lvds_connector_init(struct rcar_du_device *rcdu,
 				struct rcar_du_encoder *renc,
-				const struct rcar_du_panel_data *panel)
+				/* TODO const */ struct device_node *np)
 {
+	struct drm_encoder *encoder = rcar_encoder_to_drm_encoder(renc);
 	struct rcar_du_lvds_connector *lvdscon;
 	struct drm_connector *connector;
+	struct display_timing timing;
 	int ret;
 
 	lvdscon = devm_kzalloc(rcdu->dev, sizeof(*lvdscon), GFP_KERNEL);
 	if (lvdscon == NULL)
 		return -ENOMEM;
 
-	lvdscon->panel = panel;
+	ret = of_get_display_timing(np, "panel-timing", &timing);
+	if (ret < 0)
+		return ret;
+
+	videomode_from_timing(&timing, &lvdscon->panel.mode);
+
+	of_property_read_u32(np, "width-mm", &lvdscon->panel.width_mm);
+	of_property_read_u32(np, "height-mm", &lvdscon->panel.height_mm);
 
 	connector = &lvdscon->connector.connector;
-	connector->display_info.width_mm = panel->width_mm;
-	connector->display_info.height_mm = panel->height_mm;
+	connector->display_info.width_mm = lvdscon->panel.width_mm;
+	connector->display_info.height_mm = lvdscon->panel.height_mm;
 
 	ret = drm_connector_init(rcdu->ddev, connector, &connector_funcs,
 				 DRM_MODE_CONNECTOR_LVDS);
@@ -113,11 +121,11 @@ int rcar_du_lvds_connector_init(struct rcar_du_device *rcdu,
 	drm_object_property_set_value(&connector->base,
 		rcdu->ddev->mode_config.dpms_property, DRM_MODE_DPMS_OFF);
 
-	ret = drm_mode_connector_attach_encoder(connector, &renc->encoder);
+	ret = drm_mode_connector_attach_encoder(connector, encoder);
 	if (ret < 0)
 		return ret;
 
-	connector->encoder = &renc->encoder;
+	connector->encoder = encoder;
 	lvdscon->connector.encoder = renc;
 
 	return 0;
diff --git a/drivers/gpu/drm/rcar-du/rcar_du_lvdscon.h b/drivers/gpu/drm/rcar-du/rcar_du_lvdscon.h
index bff8683699ca..d4881ee0be7e 100644
--- a/drivers/gpu/drm/rcar-du/rcar_du_lvdscon.h
+++ b/drivers/gpu/drm/rcar-du/rcar_du_lvdscon.h
@@ -1,7 +1,7 @@
 /*
  * rcar_du_lvdscon.h  --  R-Car Display Unit LVDS Connector
  *
- * Copyright (C) 2013 Renesas Corporation
+ * Copyright (C) 2013-2014 Renesas Electronics Corporation
  *
  * Contact: Laurent Pinchart (laurent.pinchart@ideasonboard.com)
  *
@@ -16,10 +16,9 @@
 
 struct rcar_du_device;
 struct rcar_du_encoder;
-struct rcar_du_panel_data;
 
 int rcar_du_lvds_connector_init(struct rcar_du_device *rcdu,
 				struct rcar_du_encoder *renc,
-				const struct rcar_du_panel_data *panel);
+				struct device_node *np);
 
 #endif /* __RCAR_DU_LVDSCON_H__ */
diff --git a/drivers/gpu/drm/rcar-du/rcar_du_lvdsenc.c b/drivers/gpu/drm/rcar-du/rcar_du_lvdsenc.c
index df30a075d793..7cfb48ce1791 100644
--- a/drivers/gpu/drm/rcar-du/rcar_du_lvdsenc.c
+++ b/drivers/gpu/drm/rcar-du/rcar_du_lvdsenc.c
@@ -1,7 +1,7 @@
 /*
  * rcar_du_lvdsenc.c  --  R-Car Display Unit LVDS Encoder
  *
- * Copyright (C) 2013 Renesas Corporation
+ * Copyright (C) 2013-2014 Renesas Electronics Corporation
  *
  * Contact: Laurent Pinchart (laurent.pinchart@ideasonboard.com)
  *
diff --git a/drivers/gpu/drm/rcar-du/rcar_du_lvdsenc.h b/drivers/gpu/drm/rcar-du/rcar_du_lvdsenc.h
index 7051c6de19ae..f65aabda0796 100644
--- a/drivers/gpu/drm/rcar-du/rcar_du_lvdsenc.h
+++ b/drivers/gpu/drm/rcar-du/rcar_du_lvdsenc.h
@@ -1,7 +1,7 @@
 /*
  * rcar_du_lvdsenc.h  --  R-Car Display Unit LVDS Encoder
  *
- * Copyright (C) 2013 Renesas Corporation
+ * Copyright (C) 2013-2014 Renesas Electronics Corporation
  *
  * Contact: Laurent Pinchart (laurent.pinchart@ideasonboard.com)
  *
@@ -16,7 +16,6 @@
 
 #include <linux/io.h>
 #include <linux/module.h>
-#include <linux/platform_data/rcar-du.h>
 
 struct rcar_drm_crtc;
 struct rcar_du_lvdsenc;
diff --git a/drivers/gpu/drm/rcar-du/rcar_du_plane.c b/drivers/gpu/drm/rcar-du/rcar_du_plane.c
index 3fb69d9ae61b..72a7cb47bd9f 100644
--- a/drivers/gpu/drm/rcar-du/rcar_du_plane.c
+++ b/drivers/gpu/drm/rcar-du/rcar_du_plane.c
@@ -1,7 +1,7 @@
 /*
  * rcar_du_plane.c  --  R-Car Display Unit Planes
  *
- * Copyright (C) 2013 Renesas Corporation
+ * Copyright (C) 2013-2014 Renesas Electronics Corporation
  *
  * Contact: Laurent Pinchart (laurent.pinchart@ideasonboard.com)
  *
diff --git a/drivers/gpu/drm/rcar-du/rcar_du_plane.h b/drivers/gpu/drm/rcar-du/rcar_du_plane.h
index f94f9ce84998..3021288b1a89 100644
--- a/drivers/gpu/drm/rcar-du/rcar_du_plane.h
+++ b/drivers/gpu/drm/rcar-du/rcar_du_plane.h
@@ -1,7 +1,7 @@
 /*
  * rcar_du_plane.h  --  R-Car Display Unit Planes
  *
- * Copyright (C) 2013 Renesas Corporation
+ * Copyright (C) 2013-2014 Renesas Electronics Corporation
  *
  * Contact: Laurent Pinchart (laurent.pinchart@ideasonboard.com)
  *
diff --git a/drivers/gpu/drm/rcar-du/rcar_du_vgacon.c b/drivers/gpu/drm/rcar-du/rcar_du_vgacon.c
index 8af3944d31b9..752747a5e920 100644
--- a/drivers/gpu/drm/rcar-du/rcar_du_vgacon.c
+++ b/drivers/gpu/drm/rcar-du/rcar_du_vgacon.c
@@ -1,7 +1,7 @@
 /*
  * rcar_du_vgacon.c  --  R-Car Display Unit VGA Connector
  *
- * Copyright (C) 2013 Renesas Corporation
+ * Copyright (C) 2013-2014 Renesas Electronics Corporation
  *
  * Contact: Laurent Pinchart (laurent.pinchart@ideasonboard.com)
  *
@@ -52,6 +52,7 @@ static const struct drm_connector_funcs connector_funcs = {
 int rcar_du_vga_connector_init(struct rcar_du_device *rcdu,
 			       struct rcar_du_encoder *renc)
 {
+	struct drm_encoder *encoder = rcar_encoder_to_drm_encoder(renc);
 	struct rcar_du_connector *rcon;
 	struct drm_connector *connector;
 	int ret;
@@ -78,11 +79,11 @@ int rcar_du_vga_connector_init(struct rcar_du_device *rcdu,
 	drm_object_property_set_value(&connector->base,
 		rcdu->ddev->mode_config.dpms_property, DRM_MODE_DPMS_OFF);
 
-	ret = drm_mode_connector_attach_encoder(connector, &renc->encoder);
+	ret = drm_mode_connector_attach_encoder(connector, encoder);
 	if (ret < 0)
 		return ret;
 
-	connector->encoder = &renc->encoder;
+	connector->encoder = encoder;
 	rcon->encoder = renc;
 
 	return 0;
diff --git a/drivers/gpu/drm/rcar-du/rcar_du_vgacon.h b/drivers/gpu/drm/rcar-du/rcar_du_vgacon.h
index b12b0cf7f117..112f50316e01 100644
--- a/drivers/gpu/drm/rcar-du/rcar_du_vgacon.h
+++ b/drivers/gpu/drm/rcar-du/rcar_du_vgacon.h
@@ -1,7 +1,7 @@
 /*
  * rcar_du_vgacon.h  --  R-Car Display Unit VGA Connector
  *
- * Copyright (C) 2013 Renesas Corporation
+ * Copyright (C) 2013-2014 Renesas Electronics Corporation
  *
  * Contact: Laurent Pinchart (laurent.pinchart@ideasonboard.com)
  *
diff --git a/drivers/gpu/drm/rockchip/Kconfig b/drivers/gpu/drm/rockchip/Kconfig
new file mode 100644
index 000000000000..ca9f085efa92
--- /dev/null
+++ b/drivers/gpu/drm/rockchip/Kconfig
@@ -0,0 +1,17 @@
+config DRM_ROCKCHIP
+	tristate "DRM Support for Rockchip"
+	depends on DRM && ROCKCHIP_IOMMU
+	select DRM_KMS_HELPER
+	select DRM_KMS_FB_HELPER
+	select DRM_PANEL
+	select FB_CFB_FILLRECT
+	select FB_CFB_COPYAREA
+	select FB_CFB_IMAGEBLIT
+	select VT_HW_CONSOLE_BINDING if FRAMEBUFFER_CONSOLE
+	select VIDEOMODE_HELPERS
+	help
+	  Choose this option if you have a Rockchip soc chipset.
+	  This driver provides kernel mode setting and buffer
+	  management to userspace. This driver does not provide
+	  2D or 3D acceleration; acceleration is performed by other
+	  IP found on the SoC.
diff --git a/drivers/gpu/drm/rockchip/Makefile b/drivers/gpu/drm/rockchip/Makefile
new file mode 100644
index 000000000000..2cb0672f57ed
--- /dev/null
+++ b/drivers/gpu/drm/rockchip/Makefile
@@ -0,0 +1,8 @@
+#
+# Makefile for the drm device driver.  This driver provides support for the
+# Direct Rendering Infrastructure (DRI) in XFree86 4.1.0 and higher.
+
+rockchipdrm-y := rockchip_drm_drv.o rockchip_drm_fb.o rockchip_drm_fbdev.o \
+		rockchip_drm_gem.o
+
+obj-$(CONFIG_DRM_ROCKCHIP) += rockchipdrm.o rockchip_drm_vop.o
diff --git a/drivers/gpu/drm/rockchip/rockchip_drm_drv.c b/drivers/gpu/drm/rockchip/rockchip_drm_drv.c
new file mode 100644
index 000000000000..a798c7c71f91
--- /dev/null
+++ b/drivers/gpu/drm/rockchip/rockchip_drm_drv.c
@@ -0,0 +1,551 @@
+/*
+ * Copyright (C) Fuzhou Rockchip Electronics Co.Ltd
+ * Author:Mark Yao <mark.yao@rock-chips.com>
+ *
+ * based on exynos_drm_drv.c
+ *
+ * This software is licensed under the terms of the GNU General Public
+ * License version 2, as published by the Free Software Foundation, and
+ * may be copied, distributed, and modified under those terms.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ */
+
+#include <asm/dma-iommu.h>
+
+#include <drm/drmP.h>
+#include <drm/drm_crtc_helper.h>
+#include <drm/drm_fb_helper.h>
+#include <linux/dma-mapping.h>
+#include <linux/pm_runtime.h>
+#include <linux/of_graph.h>
+#include <linux/component.h>
+
+#include "rockchip_drm_drv.h"
+#include "rockchip_drm_fb.h"
+#include "rockchip_drm_fbdev.h"
+#include "rockchip_drm_gem.h"
+
+#define DRIVER_NAME	"rockchip"
+#define DRIVER_DESC	"RockChip Soc DRM"
+#define DRIVER_DATE	"20140818"
+#define DRIVER_MAJOR	1
+#define DRIVER_MINOR	0
+
+/*
+ * Attach a (component) device to the shared drm dma mapping from master drm
+ * device.  This is used by the VOPs to map GEM buffers to a common DMA
+ * mapping.
+ */
+int rockchip_drm_dma_attach_device(struct drm_device *drm_dev,
+				   struct device *dev)
+{
+	struct dma_iommu_mapping *mapping = drm_dev->dev->archdata.mapping;
+	int ret;
+
+	ret = dma_set_coherent_mask(dev, DMA_BIT_MASK(32));
+	if (ret)
+		return ret;
+
+	dma_set_max_seg_size(dev, DMA_BIT_MASK(32));
+
+	return arm_iommu_attach_device(dev, mapping);
+}
+EXPORT_SYMBOL_GPL(rockchip_drm_dma_attach_device);
+
+void rockchip_drm_dma_detach_device(struct drm_device *drm_dev,
+				    struct device *dev)
+{
+	arm_iommu_detach_device(dev);
+}
+EXPORT_SYMBOL_GPL(rockchip_drm_dma_detach_device);
+
+int rockchip_register_crtc_funcs(struct drm_device *dev,
+				 const struct rockchip_crtc_funcs *crtc_funcs,
+				 int pipe)
+{
+	struct rockchip_drm_private *priv = dev->dev_private;
+
+	if (pipe > ROCKCHIP_MAX_CRTC)
+		return -EINVAL;
+
+	priv->crtc_funcs[pipe] = crtc_funcs;
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(rockchip_register_crtc_funcs);
+
+void rockchip_unregister_crtc_funcs(struct drm_device *dev, int pipe)
+{
+	struct rockchip_drm_private *priv = dev->dev_private;
+
+	if (pipe > ROCKCHIP_MAX_CRTC)
+		return;
+
+	priv->crtc_funcs[pipe] = NULL;
+}
+EXPORT_SYMBOL_GPL(rockchip_unregister_crtc_funcs);
+
+static struct drm_crtc *rockchip_crtc_from_pipe(struct drm_device *drm,
+						int pipe)
+{
+	struct drm_crtc *crtc;
+	int i = 0;
+
+	list_for_each_entry(crtc, &drm->mode_config.crtc_list, head)
+		if (i++ == pipe)
+			return crtc;
+
+	return NULL;
+}
+
+static int rockchip_drm_crtc_enable_vblank(struct drm_device *dev, int pipe)
+{
+	struct rockchip_drm_private *priv = dev->dev_private;
+	struct drm_crtc *crtc = rockchip_crtc_from_pipe(dev, pipe);
+
+	if (crtc && priv->crtc_funcs[pipe] &&
+	    priv->crtc_funcs[pipe]->enable_vblank)
+		return priv->crtc_funcs[pipe]->enable_vblank(crtc);
+
+	return 0;
+}
+
+static void rockchip_drm_crtc_disable_vblank(struct drm_device *dev, int pipe)
+{
+	struct rockchip_drm_private *priv = dev->dev_private;
+	struct drm_crtc *crtc = rockchip_crtc_from_pipe(dev, pipe);
+
+	if (crtc && priv->crtc_funcs[pipe] &&
+	    priv->crtc_funcs[pipe]->enable_vblank)
+		priv->crtc_funcs[pipe]->disable_vblank(crtc);
+}
+
+static int rockchip_drm_load(struct drm_device *drm_dev, unsigned long flags)
+{
+	struct rockchip_drm_private *private;
+	struct dma_iommu_mapping *mapping;
+	struct device *dev = drm_dev->dev;
+	int ret;
+
+	private = devm_kzalloc(drm_dev->dev, sizeof(*private), GFP_KERNEL);
+	if (!private)
+		return -ENOMEM;
+
+	drm_dev->dev_private = private;
+
+	drm_mode_config_init(drm_dev);
+
+	rockchip_drm_mode_config_init(drm_dev);
+
+	dev->dma_parms = devm_kzalloc(dev, sizeof(*dev->dma_parms),
+				      GFP_KERNEL);
+	if (!dev->dma_parms) {
+		ret = -ENOMEM;
+		goto err_config_cleanup;
+	}
+
+	/* TODO(djkurtz): fetch the mapping start/size from somewhere */
+	mapping = arm_iommu_create_mapping(&platform_bus_type, 0x00000000,
+					   SZ_2G);
+	if (IS_ERR(mapping)) {
+		ret = PTR_ERR(mapping);
+		goto err_config_cleanup;
+	}
+
+	ret = dma_set_mask_and_coherent(dev, DMA_BIT_MASK(32));
+	if (ret)
+		goto err_release_mapping;
+
+	dma_set_max_seg_size(dev, DMA_BIT_MASK(32));
+
+	ret = arm_iommu_attach_device(dev, mapping);
+	if (ret)
+		goto err_release_mapping;
+
+	/* Try to bind all sub drivers. */
+	ret = component_bind_all(dev, drm_dev);
+	if (ret)
+		goto err_detach_device;
+
+	/* init kms poll for handling hpd */
+	drm_kms_helper_poll_init(drm_dev);
+
+	/*
+	 * enable drm irq mode.
+	 * - with irq_enabled = true, we can use the vblank feature.
+	 */
+	drm_dev->irq_enabled = true;
+
+	ret = drm_vblank_init(drm_dev, ROCKCHIP_MAX_CRTC);
+	if (ret)
+		goto err_kms_helper_poll_fini;
+
+	/*
+	 * with vblank_disable_allowed = true, vblank interrupt will be disabled
+	 * by drm timer once a current process gives up ownership of
+	 * vblank event.(after drm_vblank_put function is called)
+	 */
+	drm_dev->vblank_disable_allowed = true;
+
+	ret = rockchip_drm_fbdev_init(drm_dev);
+	if (ret)
+		goto err_vblank_cleanup;
+
+	return 0;
+err_vblank_cleanup:
+	drm_vblank_cleanup(drm_dev);
+err_kms_helper_poll_fini:
+	drm_kms_helper_poll_fini(drm_dev);
+	component_unbind_all(dev, drm_dev);
+err_detach_device:
+	arm_iommu_detach_device(dev);
+err_release_mapping:
+	arm_iommu_release_mapping(dev->archdata.mapping);
+err_config_cleanup:
+	drm_mode_config_cleanup(drm_dev);
+	drm_dev->dev_private = NULL;
+	return ret;
+}
+
+static int rockchip_drm_unload(struct drm_device *drm_dev)
+{
+	struct device *dev = drm_dev->dev;
+
+	rockchip_drm_fbdev_fini(drm_dev);
+	drm_vblank_cleanup(drm_dev);
+	drm_kms_helper_poll_fini(drm_dev);
+	component_unbind_all(dev, drm_dev);
+	arm_iommu_detach_device(dev);
+	arm_iommu_release_mapping(dev->archdata.mapping);
+	drm_mode_config_cleanup(drm_dev);
+	drm_dev->dev_private = NULL;
+
+	return 0;
+}
+
+void rockchip_drm_lastclose(struct drm_device *dev)
+{
+	struct rockchip_drm_private *priv = dev->dev_private;
+
+	drm_fb_helper_restore_fbdev_mode_unlocked(&priv->fbdev_helper);
+}
+
+static const struct file_operations rockchip_drm_driver_fops = {
+	.owner = THIS_MODULE,
+	.open = drm_open,
+	.mmap = rockchip_gem_mmap,
+	.poll = drm_poll,
+	.read = drm_read,
+	.unlocked_ioctl = drm_ioctl,
+#ifdef CONFIG_COMPAT
+	.compat_ioctl = drm_compat_ioctl,
+#endif
+	.release = drm_release,
+};
+
+const struct vm_operations_struct rockchip_drm_vm_ops = {
+	.open = drm_gem_vm_open,
+	.close = drm_gem_vm_close,
+};
+
+static struct drm_driver rockchip_drm_driver = {
+	.driver_features	= DRIVER_MODESET | DRIVER_GEM | DRIVER_PRIME,
+	.load			= rockchip_drm_load,
+	.unload			= rockchip_drm_unload,
+	.lastclose		= rockchip_drm_lastclose,
+	.get_vblank_counter	= drm_vblank_count,
+	.enable_vblank		= rockchip_drm_crtc_enable_vblank,
+	.disable_vblank		= rockchip_drm_crtc_disable_vblank,
+	.gem_vm_ops		= &rockchip_drm_vm_ops,
+	.gem_free_object	= rockchip_gem_free_object,
+	.dumb_create		= rockchip_gem_dumb_create,
+	.dumb_map_offset	= rockchip_gem_dumb_map_offset,
+	.dumb_destroy		= drm_gem_dumb_destroy,
+	.prime_handle_to_fd	= drm_gem_prime_handle_to_fd,
+	.prime_fd_to_handle	= drm_gem_prime_fd_to_handle,
+	.gem_prime_import	= drm_gem_prime_import,
+	.gem_prime_export	= drm_gem_prime_export,
+	.gem_prime_get_sg_table	= rockchip_gem_prime_get_sg_table,
+	.gem_prime_vmap		= rockchip_gem_prime_vmap,
+	.gem_prime_vunmap	= rockchip_gem_prime_vunmap,
+	.gem_prime_mmap		= rockchip_gem_mmap_buf,
+	.fops			= &rockchip_drm_driver_fops,
+	.name	= DRIVER_NAME,
+	.desc	= DRIVER_DESC,
+	.date	= DRIVER_DATE,
+	.major	= DRIVER_MAJOR,
+	.minor	= DRIVER_MINOR,
+};
+
+#ifdef CONFIG_PM_SLEEP
+static int rockchip_drm_sys_suspend(struct device *dev)
+{
+	struct drm_device *drm = dev_get_drvdata(dev);
+	struct drm_connector *connector;
+
+	if (!drm)
+		return 0;
+
+	drm_modeset_lock_all(drm);
+	list_for_each_entry(connector, &drm->mode_config.connector_list, head) {
+		int old_dpms = connector->dpms;
+
+		if (connector->funcs->dpms)
+			connector->funcs->dpms(connector, DRM_MODE_DPMS_OFF);
+
+		/* Set the old mode back to the connector for resume */
+		connector->dpms = old_dpms;
+	}
+	drm_modeset_unlock_all(drm);
+
+	return 0;
+}
+
+static int rockchip_drm_sys_resume(struct device *dev)
+{
+	struct drm_device *drm = dev_get_drvdata(dev);
+	struct drm_connector *connector;
+	enum drm_connector_status status;
+	bool changed = false;
+
+	if (!drm)
+		return 0;
+
+	drm_modeset_lock_all(drm);
+	list_for_each_entry(connector, &drm->mode_config.connector_list, head) {
+		int desired_mode = connector->dpms;
+
+		/*
+		 * at suspend time, we save dpms to connector->dpms,
+		 * restore the old_dpms, and at current time, the connector
+		 * dpms status must be DRM_MODE_DPMS_OFF.
+		 */
+		connector->dpms = DRM_MODE_DPMS_OFF;
+
+		/*
+		 * If the connector has been disconnected during suspend,
+		 * disconnect it from the encoder and leave it off. We'll notify
+		 * userspace at the end.
+		 */
+		if (desired_mode == DRM_MODE_DPMS_ON) {
+			status = connector->funcs->detect(connector, true);
+			if (status == connector_status_disconnected) {
+				connector->encoder = NULL;
+				connector->status = status;
+				changed = true;
+				continue;
+			}
+		}
+		if (connector->funcs->dpms)
+			connector->funcs->dpms(connector, desired_mode);
+	}
+	drm_modeset_unlock_all(drm);
+
+	drm_helper_resume_force_mode(drm);
+
+	if (changed)
+		drm_kms_helper_hotplug_event(drm);
+
+	return 0;
+}
+#endif
+
+static const struct dev_pm_ops rockchip_drm_pm_ops = {
+	SET_SYSTEM_SLEEP_PM_OPS(rockchip_drm_sys_suspend,
+				rockchip_drm_sys_resume)
+};
+
+/*
+ * @node: device tree node containing encoder input ports
+ * @encoder: drm_encoder
+ */
+int rockchip_drm_encoder_get_mux_id(struct device_node *node,
+				    struct drm_encoder *encoder)
+{
+	struct device_node *ep = NULL;
+	struct drm_crtc *crtc = encoder->crtc;
+	struct of_endpoint endpoint;
+	struct device_node *port;
+	int ret;
+
+	if (!node || !crtc)
+		return -EINVAL;
+
+	do {
+		ep = of_graph_get_next_endpoint(node, ep);
+		if (!ep)
+			break;
+
+		port = of_graph_get_remote_port(ep);
+		of_node_put(port);
+		if (port == crtc->port) {
+			ret = of_graph_parse_endpoint(ep, &endpoint);
+			return ret ?: endpoint.id;
+		}
+	} while (ep);
+
+	return -EINVAL;
+}
+
+static int compare_of(struct device *dev, void *data)
+{
+	struct device_node *np = data;
+
+	return dev->of_node == np;
+}
+
+static void rockchip_add_endpoints(struct device *dev,
+				   struct component_match **match,
+				   struct device_node *port)
+{
+	struct device_node *ep, *remote;
+
+	for_each_child_of_node(port, ep) {
+		remote = of_graph_get_remote_port_parent(ep);
+		if (!remote || !of_device_is_available(remote)) {
+			of_node_put(remote);
+			continue;
+		} else if (!of_device_is_available(remote->parent)) {
+			dev_warn(dev, "parent device of %s is not available\n",
+				 remote->full_name);
+			of_node_put(remote);
+			continue;
+		}
+
+		component_match_add(dev, match, compare_of, remote);
+		of_node_put(remote);
+	}
+}
+
+static int rockchip_drm_bind(struct device *dev)
+{
+	struct drm_device *drm;
+	int ret;
+
+	drm = drm_dev_alloc(&rockchip_drm_driver, dev);
+	if (!drm)
+		return -ENOMEM;
+
+	ret = drm_dev_set_unique(drm, "%s", dev_name(dev));
+	if (ret)
+		goto err_free;
+
+	ret = drm_dev_register(drm, 0);
+	if (ret)
+		goto err_free;
+
+	dev_set_drvdata(dev, drm);
+
+	return 0;
+
+err_free:
+	drm_dev_unref(drm);
+	return ret;
+}
+
+static void rockchip_drm_unbind(struct device *dev)
+{
+	struct drm_device *drm = dev_get_drvdata(dev);
+
+	drm_dev_unregister(drm);
+	drm_dev_unref(drm);
+	dev_set_drvdata(dev, NULL);
+}
+
+static const struct component_master_ops rockchip_drm_ops = {
+	.bind = rockchip_drm_bind,
+	.unbind = rockchip_drm_unbind,
+};
+
+static int rockchip_drm_platform_probe(struct platform_device *pdev)
+{
+	struct device *dev = &pdev->dev;
+	struct component_match *match = NULL;
+	struct device_node *np = dev->of_node;
+	struct device_node *port;
+	int i;
+
+	if (!np)
+		return -ENODEV;
+	/*
+	 * Bind the crtc ports first, so that
+	 * drm_of_find_possible_crtcs called from encoder .bind callbacks
+	 * works as expected.
+	 */
+	for (i = 0;; i++) {
+		port = of_parse_phandle(np, "ports", i);
+		if (!port)
+			break;
+
+		if (!of_device_is_available(port->parent)) {
+			of_node_put(port);
+			continue;
+		}
+
+		component_match_add(dev, &match, compare_of, port->parent);
+		of_node_put(port);
+	}
+
+	if (i == 0) {
+		dev_err(dev, "missing 'ports' property\n");
+		return -ENODEV;
+	}
+
+	if (!match) {
+		dev_err(dev, "No available vop found for display-subsystem.\n");
+		return -ENODEV;
+	}
+	/*
+	 * For each bound crtc, bind the encoders attached to its
+	 * remote endpoint.
+	 */
+	for (i = 0;; i++) {
+		port = of_parse_phandle(np, "ports", i);
+		if (!port)
+			break;
+
+		if (!of_device_is_available(port->parent)) {
+			of_node_put(port);
+			continue;
+		}
+
+		rockchip_add_endpoints(dev, &match, port);
+		of_node_put(port);
+	}
+
+	return component_master_add_with_match(dev, &rockchip_drm_ops, match);
+}
+
+static int rockchip_drm_platform_remove(struct platform_device *pdev)
+{
+	component_master_del(&pdev->dev, &rockchip_drm_ops);
+
+	return 0;
+}
+
+static const struct of_device_id rockchip_drm_dt_ids[] = {
+	{ .compatible = "rockchip,display-subsystem", },
+	{ /* sentinel */ },
+};
+MODULE_DEVICE_TABLE(of, rockchip_drm_dt_ids);
+
+static struct platform_driver rockchip_drm_platform_driver = {
+	.probe = rockchip_drm_platform_probe,
+	.remove = rockchip_drm_platform_remove,
+	.driver = {
+		.owner = THIS_MODULE,
+		.name = "rockchip-drm",
+		.of_match_table = rockchip_drm_dt_ids,
+		.pm = &rockchip_drm_pm_ops,
+	},
+};
+
+module_platform_driver(rockchip_drm_platform_driver);
+
+MODULE_AUTHOR("Mark Yao <mark.yao@rock-chips.com>");
+MODULE_DESCRIPTION("ROCKCHIP DRM Driver");
+MODULE_LICENSE("GPL v2");
diff --git a/drivers/gpu/drm/rockchip/rockchip_drm_drv.h b/drivers/gpu/drm/rockchip/rockchip_drm_drv.h
new file mode 100644
index 000000000000..dc4e5f03ac79
--- /dev/null
+++ b/drivers/gpu/drm/rockchip/rockchip_drm_drv.h
@@ -0,0 +1,68 @@
+/*
+ * Copyright (C) Fuzhou Rockchip Electronics Co.Ltd
+ * Author:Mark Yao <mark.yao@rock-chips.com>
+ *
+ * based on exynos_drm_drv.h
+ *
+ * This software is licensed under the terms of the GNU General Public
+ * License version 2, as published by the Free Software Foundation, and
+ * may be copied, distributed, and modified under those terms.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ */
+
+#ifndef _ROCKCHIP_DRM_DRV_H
+#define _ROCKCHIP_DRM_DRV_H
+
+#include <drm/drm_fb_helper.h>
+#include <drm/drm_gem.h>
+
+#include <linux/module.h>
+#include <linux/component.h>
+
+#define ROCKCHIP_MAX_FB_BUFFER	3
+#define ROCKCHIP_MAX_CONNECTOR	2
+#define ROCKCHIP_MAX_CRTC	2
+
+struct drm_device;
+struct drm_connector;
+
+/*
+ * Rockchip drm private crtc funcs.
+ * @enable_vblank: enable crtc vblank irq.
+ * @disable_vblank: disable crtc vblank irq.
+ */
+struct rockchip_crtc_funcs {
+	int (*enable_vblank)(struct drm_crtc *crtc);
+	void (*disable_vblank)(struct drm_crtc *crtc);
+};
+
+/*
+ * Rockchip drm private structure.
+ *
+ * @crtc: array of enabled CRTCs, used to map from "pipe" to drm_crtc.
+ * @num_pipe: number of pipes for this device.
+ */
+struct rockchip_drm_private {
+	struct drm_fb_helper fbdev_helper;
+	struct drm_gem_object *fbdev_bo;
+	const struct rockchip_crtc_funcs *crtc_funcs[ROCKCHIP_MAX_CRTC];
+};
+
+int rockchip_register_crtc_funcs(struct drm_device *dev,
+				 const struct rockchip_crtc_funcs *crtc_funcs,
+				 int pipe);
+void rockchip_unregister_crtc_funcs(struct drm_device *dev, int pipe);
+int rockchip_drm_encoder_get_mux_id(struct device_node *node,
+				    struct drm_encoder *encoder);
+int rockchip_drm_crtc_mode_config(struct drm_crtc *crtc, int connector_type,
+				  int out_mode);
+int rockchip_drm_dma_attach_device(struct drm_device *drm_dev,
+				   struct device *dev);
+void rockchip_drm_dma_detach_device(struct drm_device *drm_dev,
+				    struct device *dev);
+
+#endif /* _ROCKCHIP_DRM_DRV_H_ */
diff --git a/drivers/gpu/drm/rockchip/rockchip_drm_fb.c b/drivers/gpu/drm/rockchip/rockchip_drm_fb.c
new file mode 100644
index 000000000000..77d52893d40f
--- /dev/null
+++ b/drivers/gpu/drm/rockchip/rockchip_drm_fb.c
@@ -0,0 +1,201 @@
+/*
+ * Copyright (C) Fuzhou Rockchip Electronics Co.Ltd
+ * Author:Mark Yao <mark.yao@rock-chips.com>
+ *
+ * This software is licensed under the terms of the GNU General Public
+ * License version 2, as published by the Free Software Foundation, and
+ * may be copied, distributed, and modified under those terms.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ */
+
+#include <linux/kernel.h>
+#include <drm/drm.h>
+#include <drm/drmP.h>
+#include <drm/drm_fb_helper.h>
+#include <drm/drm_crtc_helper.h>
+
+#include "rockchip_drm_drv.h"
+#include "rockchip_drm_gem.h"
+
+#define to_rockchip_fb(x) container_of(x, struct rockchip_drm_fb, fb)
+
+struct rockchip_drm_fb {
+	struct drm_framebuffer fb;
+	struct drm_gem_object *obj[ROCKCHIP_MAX_FB_BUFFER];
+};
+
+struct drm_gem_object *rockchip_fb_get_gem_obj(struct drm_framebuffer *fb,
+					       unsigned int plane)
+{
+	struct rockchip_drm_fb *rk_fb = to_rockchip_fb(fb);
+
+	if (plane >= ROCKCHIP_MAX_FB_BUFFER)
+		return NULL;
+
+	return rk_fb->obj[plane];
+}
+EXPORT_SYMBOL_GPL(rockchip_fb_get_gem_obj);
+
+static void rockchip_drm_fb_destroy(struct drm_framebuffer *fb)
+{
+	struct rockchip_drm_fb *rockchip_fb = to_rockchip_fb(fb);
+	struct drm_gem_object *obj;
+	int i;
+
+	for (i = 0; i < ROCKCHIP_MAX_FB_BUFFER; i++) {
+		obj = rockchip_fb->obj[i];
+		if (obj)
+			drm_gem_object_unreference_unlocked(obj);
+	}
+
+	drm_framebuffer_cleanup(fb);
+	kfree(rockchip_fb);
+}
+
+static int rockchip_drm_fb_create_handle(struct drm_framebuffer *fb,
+					 struct drm_file *file_priv,
+					 unsigned int *handle)
+{
+	struct rockchip_drm_fb *rockchip_fb = to_rockchip_fb(fb);
+
+	return drm_gem_handle_create(file_priv,
+				     rockchip_fb->obj[0], handle);
+}
+
+static struct drm_framebuffer_funcs rockchip_drm_fb_funcs = {
+	.destroy	= rockchip_drm_fb_destroy,
+	.create_handle	= rockchip_drm_fb_create_handle,
+};
+
+static struct rockchip_drm_fb *
+rockchip_fb_alloc(struct drm_device *dev, struct drm_mode_fb_cmd2 *mode_cmd,
+		  struct drm_gem_object **obj, unsigned int num_planes)
+{
+	struct rockchip_drm_fb *rockchip_fb;
+	int ret;
+	int i;
+
+	rockchip_fb = kzalloc(sizeof(*rockchip_fb), GFP_KERNEL);
+	if (!rockchip_fb)
+		return ERR_PTR(-ENOMEM);
+
+	drm_helper_mode_fill_fb_struct(&rockchip_fb->fb, mode_cmd);
+
+	for (i = 0; i < num_planes; i++)
+		rockchip_fb->obj[i] = obj[i];
+
+	ret = drm_framebuffer_init(dev, &rockchip_fb->fb,
+				   &rockchip_drm_fb_funcs);
+	if (ret) {
+		dev_err(dev->dev, "Failed to initialize framebuffer: %d\n",
+			ret);
+		kfree(rockchip_fb);
+		return ERR_PTR(ret);
+	}
+
+	return rockchip_fb;
+}
+
+static struct drm_framebuffer *
+rockchip_user_fb_create(struct drm_device *dev, struct drm_file *file_priv,
+			struct drm_mode_fb_cmd2 *mode_cmd)
+{
+	struct rockchip_drm_fb *rockchip_fb;
+	struct drm_gem_object *objs[ROCKCHIP_MAX_FB_BUFFER];
+	struct drm_gem_object *obj;
+	unsigned int hsub;
+	unsigned int vsub;
+	int num_planes;
+	int ret;
+	int i;
+
+	hsub = drm_format_horz_chroma_subsampling(mode_cmd->pixel_format);
+	vsub = drm_format_vert_chroma_subsampling(mode_cmd->pixel_format);
+	num_planes = min(drm_format_num_planes(mode_cmd->pixel_format),
+			 ROCKCHIP_MAX_FB_BUFFER);
+
+	for (i = 0; i < num_planes; i++) {
+		unsigned int width = mode_cmd->width / (i ? hsub : 1);
+		unsigned int height = mode_cmd->height / (i ? vsub : 1);
+		unsigned int min_size;
+
+		obj = drm_gem_object_lookup(dev, file_priv,
+					    mode_cmd->handles[i]);
+		if (!obj) {
+			dev_err(dev->dev, "Failed to lookup GEM object\n");
+			ret = -ENXIO;
+			goto err_gem_object_unreference;
+		}
+
+		min_size = (height - 1) * mode_cmd->pitches[i] +
+			mode_cmd->offsets[i] +
+			width * drm_format_plane_cpp(mode_cmd->pixel_format, i);
+
+		if (obj->size < min_size) {
+			drm_gem_object_unreference_unlocked(obj);
+			ret = -EINVAL;
+			goto err_gem_object_unreference;
+		}
+		objs[i] = obj;
+	}
+
+	rockchip_fb = rockchip_fb_alloc(dev, mode_cmd, objs, i);
+	if (IS_ERR(rockchip_fb)) {
+		ret = PTR_ERR(rockchip_fb);
+		goto err_gem_object_unreference;
+	}
+
+	return &rockchip_fb->fb;
+
+err_gem_object_unreference:
+	for (i--; i >= 0; i--)
+		drm_gem_object_unreference_unlocked(objs[i]);
+	return ERR_PTR(ret);
+}
+
+static void rockchip_drm_output_poll_changed(struct drm_device *dev)
+{
+	struct rockchip_drm_private *private = dev->dev_private;
+	struct drm_fb_helper *fb_helper = &private->fbdev_helper;
+
+	drm_fb_helper_hotplug_event(fb_helper);
+}
+
+static const struct drm_mode_config_funcs rockchip_drm_mode_config_funcs = {
+	.fb_create = rockchip_user_fb_create,
+	.output_poll_changed = rockchip_drm_output_poll_changed,
+};
+
+struct drm_framebuffer *
+rockchip_drm_framebuffer_init(struct drm_device *dev,
+			      struct drm_mode_fb_cmd2 *mode_cmd,
+			      struct drm_gem_object *obj)
+{
+	struct rockchip_drm_fb *rockchip_fb;
+
+	rockchip_fb = rockchip_fb_alloc(dev, mode_cmd, &obj, 1);
+	if (IS_ERR(rockchip_fb))
+		return NULL;
+
+	return &rockchip_fb->fb;
+}
+
+void rockchip_drm_mode_config_init(struct drm_device *dev)
+{
+	dev->mode_config.min_width = 0;
+	dev->mode_config.min_height = 0;
+
+	/*
+	 * set max width and height as default value(4096x4096).
+	 * this value would be used to check framebuffer size limitation
+	 * at drm_mode_addfb().
+	 */
+	dev->mode_config.max_width = 4096;
+	dev->mode_config.max_height = 4096;
+
+	dev->mode_config.funcs = &rockchip_drm_mode_config_funcs;
+}
diff --git a/drivers/gpu/drm/rockchip/rockchip_drm_fb.h b/drivers/gpu/drm/rockchip/rockchip_drm_fb.h
new file mode 100644
index 000000000000..09574d48226f
--- /dev/null
+++ b/drivers/gpu/drm/rockchip/rockchip_drm_fb.h
@@ -0,0 +1,28 @@
+/*
+ * Copyright (C) Fuzhou Rockchip Electronics Co.Ltd
+ * Author:Mark Yao <mark.yao@rock-chips.com>
+ *
+ * This software is licensed under the terms of the GNU General Public
+ * License version 2, as published by the Free Software Foundation, and
+ * may be copied, distributed, and modified under those terms.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ */
+
+#ifndef _ROCKCHIP_DRM_FB_H
+#define _ROCKCHIP_DRM_FB_H
+
+struct drm_framebuffer *
+rockchip_drm_framebuffer_init(struct drm_device *dev,
+			      struct drm_mode_fb_cmd2 *mode_cmd,
+			      struct drm_gem_object *obj);
+void rockchip_drm_framebuffer_fini(struct drm_framebuffer *fb);
+
+void rockchip_drm_mode_config_init(struct drm_device *dev);
+
+struct drm_gem_object *rockchip_fb_get_gem_obj(struct drm_framebuffer *fb,
+					       unsigned int plane);
+#endif /* _ROCKCHIP_DRM_FB_H */
diff --git a/drivers/gpu/drm/rockchip/rockchip_drm_fbdev.c b/drivers/gpu/drm/rockchip/rockchip_drm_fbdev.c
new file mode 100644
index 000000000000..a5d889a8716b
--- /dev/null
+++ b/drivers/gpu/drm/rockchip/rockchip_drm_fbdev.c
@@ -0,0 +1,210 @@
+/*
+ * Copyright (C) Fuzhou Rockchip Electronics Co.Ltd
+ * Author:Mark Yao <mark.yao@rock-chips.com>
+ *
+ * This software is licensed under the terms of the GNU General Public
+ * License version 2, as published by the Free Software Foundation, and
+ * may be copied, distributed, and modified under those terms.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ */
+
+#include <drm/drm.h>
+#include <drm/drmP.h>
+#include <drm/drm_fb_helper.h>
+#include <drm/drm_crtc_helper.h>
+
+#include "rockchip_drm_drv.h"
+#include "rockchip_drm_gem.h"
+#include "rockchip_drm_fb.h"
+
+#define PREFERRED_BPP		32
+#define to_drm_private(x) \
+		container_of(x, struct rockchip_drm_private, fbdev_helper)
+
+static int rockchip_fbdev_mmap(struct fb_info *info,
+			       struct vm_area_struct *vma)
+{
+	struct drm_fb_helper *helper = info->par;
+	struct rockchip_drm_private *private = to_drm_private(helper);
+
+	return rockchip_gem_mmap_buf(private->fbdev_bo, vma);
+}
+
+static struct fb_ops rockchip_drm_fbdev_ops = {
+	.owner		= THIS_MODULE,
+	.fb_mmap	= rockchip_fbdev_mmap,
+	.fb_fillrect	= cfb_fillrect,
+	.fb_copyarea	= cfb_copyarea,
+	.fb_imageblit	= cfb_imageblit,
+	.fb_check_var	= drm_fb_helper_check_var,
+	.fb_set_par	= drm_fb_helper_set_par,
+	.fb_blank	= drm_fb_helper_blank,
+	.fb_pan_display	= drm_fb_helper_pan_display,
+	.fb_setcmap	= drm_fb_helper_setcmap,
+};
+
+static int rockchip_drm_fbdev_create(struct drm_fb_helper *helper,
+				     struct drm_fb_helper_surface_size *sizes)
+{
+	struct rockchip_drm_private *private = to_drm_private(helper);
+	struct drm_mode_fb_cmd2 mode_cmd = { 0 };
+	struct drm_device *dev = helper->dev;
+	struct rockchip_gem_object *rk_obj;
+	struct drm_framebuffer *fb;
+	unsigned int bytes_per_pixel;
+	unsigned long offset;
+	struct fb_info *fbi;
+	size_t size;
+	int ret;
+
+	bytes_per_pixel = DIV_ROUND_UP(sizes->surface_bpp, 8);
+
+	mode_cmd.width = sizes->surface_width;
+	mode_cmd.height = sizes->surface_height;
+	mode_cmd.pitches[0] = sizes->surface_width * bytes_per_pixel;
+	mode_cmd.pixel_format = drm_mode_legacy_fb_format(sizes->surface_bpp,
+		sizes->surface_depth);
+
+	size = mode_cmd.pitches[0] * mode_cmd.height;
+
+	rk_obj = rockchip_gem_create_object(dev, size);
+	if (IS_ERR(rk_obj))
+		return -ENOMEM;
+
+	private->fbdev_bo = &rk_obj->base;
+
+	fbi = framebuffer_alloc(0, dev->dev);
+	if (!fbi) {
+		dev_err(dev->dev, "Failed to allocate framebuffer info.\n");
+		ret = -ENOMEM;
+		goto err_rockchip_gem_free_object;
+	}
+
+	helper->fb = rockchip_drm_framebuffer_init(dev, &mode_cmd,
+						   private->fbdev_bo);
+	if (IS_ERR(helper->fb)) {
+		dev_err(dev->dev, "Failed to allocate DRM framebuffer.\n");
+		ret = PTR_ERR(helper->fb);
+		goto err_framebuffer_release;
+	}
+
+	helper->fbdev = fbi;
+
+	fbi->par = helper;
+	fbi->flags = FBINFO_FLAG_DEFAULT;
+	fbi->fbops = &rockchip_drm_fbdev_ops;
+
+	ret = fb_alloc_cmap(&fbi->cmap, 256, 0);
+	if (ret) {
+		dev_err(dev->dev, "Failed to allocate color map.\n");
+		goto err_drm_framebuffer_unref;
+	}
+
+	fb = helper->fb;
+	drm_fb_helper_fill_fix(fbi, fb->pitches[0], fb->depth);
+	drm_fb_helper_fill_var(fbi, helper, fb->width, fb->height);
+
+	offset = fbi->var.xoffset * bytes_per_pixel;
+	offset += fbi->var.yoffset * fb->pitches[0];
+
+	dev->mode_config.fb_base = 0;
+	fbi->screen_base = rk_obj->kvaddr + offset;
+	fbi->screen_size = rk_obj->base.size;
+	fbi->fix.smem_len = rk_obj->base.size;
+
+	DRM_DEBUG_KMS("FB [%dx%d]-%d kvaddr=%p offset=%ld size=%d\n",
+		      fb->width, fb->height, fb->depth, rk_obj->kvaddr,
+		      offset, size);
+	return 0;
+
+err_drm_framebuffer_unref:
+	drm_framebuffer_unreference(helper->fb);
+err_framebuffer_release:
+	framebuffer_release(fbi);
+err_rockchip_gem_free_object:
+	rockchip_gem_free_object(&rk_obj->base);
+	return ret;
+}
+
+static const struct drm_fb_helper_funcs rockchip_drm_fb_helper_funcs = {
+	.fb_probe = rockchip_drm_fbdev_create,
+};
+
+int rockchip_drm_fbdev_init(struct drm_device *dev)
+{
+	struct rockchip_drm_private *private = dev->dev_private;
+	struct drm_fb_helper *helper;
+	unsigned int num_crtc;
+	int ret;
+
+	if (!dev->mode_config.num_crtc || !dev->mode_config.num_connector)
+		return -EINVAL;
+
+	num_crtc = dev->mode_config.num_crtc;
+
+	helper = &private->fbdev_helper;
+
+	drm_fb_helper_prepare(dev, helper, &rockchip_drm_fb_helper_funcs);
+
+	ret = drm_fb_helper_init(dev, helper, num_crtc, ROCKCHIP_MAX_CONNECTOR);
+	if (ret < 0) {
+		dev_err(dev->dev, "Failed to initialize drm fb helper - %d.\n",
+			ret);
+		return ret;
+	}
+
+	ret = drm_fb_helper_single_add_all_connectors(helper);
+	if (ret < 0) {
+		dev_err(dev->dev, "Failed to add connectors - %d.\n", ret);
+		goto err_drm_fb_helper_fini;
+	}
+
+	/* disable all the possible outputs/crtcs before entering KMS mode */
+	drm_helper_disable_unused_functions(dev);
+
+	ret = drm_fb_helper_initial_config(helper, PREFERRED_BPP);
+	if (ret < 0) {
+		dev_err(dev->dev, "Failed to set initial hw config - %d.\n",
+			ret);
+		goto err_drm_fb_helper_fini;
+	}
+
+	return 0;
+
+err_drm_fb_helper_fini:
+	drm_fb_helper_fini(helper);
+	return ret;
+}
+
+void rockchip_drm_fbdev_fini(struct drm_device *dev)
+{
+	struct rockchip_drm_private *private = dev->dev_private;
+	struct drm_fb_helper *helper;
+
+	helper = &private->fbdev_helper;
+
+	if (helper->fbdev) {
+		struct fb_info *info;
+		int ret;
+
+		info = helper->fbdev;
+		ret = unregister_framebuffer(info);
+		if (ret < 0)
+			DRM_DEBUG_KMS("failed unregister_framebuffer() - %d\n",
+				      ret);
+
+		if (info->cmap.len)
+			fb_dealloc_cmap(&info->cmap);
+
+		framebuffer_release(info);
+	}
+
+	if (helper->fb)
+		drm_framebuffer_unreference(helper->fb);
+
+	drm_fb_helper_fini(helper);
+}
diff --git a/drivers/gpu/drm/rockchip/rockchip_drm_fbdev.h b/drivers/gpu/drm/rockchip/rockchip_drm_fbdev.h
new file mode 100644
index 000000000000..50432e9b5b37
--- /dev/null
+++ b/drivers/gpu/drm/rockchip/rockchip_drm_fbdev.h
@@ -0,0 +1,21 @@
+/*
+ * Copyright (C) Fuzhou Rockchip Electronics Co.Ltd
+ * Author:Mark Yao <mark.yao@rock-chips.com>
+ *
+ * This software is licensed under the terms of the GNU General Public
+ * License version 2, as published by the Free Software Foundation, and
+ * may be copied, distributed, and modified under those terms.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ */
+
+#ifndef _ROCKCHIP_DRM_FBDEV_H
+#define _ROCKCHIP_DRM_FBDEV_H
+
+int rockchip_drm_fbdev_init(struct drm_device *dev);
+void rockchip_drm_fbdev_fini(struct drm_device *dev);
+
+#endif /* _ROCKCHIP_DRM_FBDEV_H */
diff --git a/drivers/gpu/drm/rockchip/rockchip_drm_gem.c b/drivers/gpu/drm/rockchip/rockchip_drm_gem.c
new file mode 100644
index 000000000000..bc98a227dc76
--- /dev/null
+++ b/drivers/gpu/drm/rockchip/rockchip_drm_gem.c
@@ -0,0 +1,294 @@
+/*
+ * Copyright (C) Fuzhou Rockchip Electronics Co.Ltd
+ * Author:Mark Yao <mark.yao@rock-chips.com>
+ *
+ * This software is licensed under the terms of the GNU General Public
+ * License version 2, as published by the Free Software Foundation, and
+ * may be copied, distributed, and modified under those terms.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ */
+
+#include <drm/drm.h>
+#include <drm/drmP.h>
+#include <drm/drm_gem.h>
+#include <drm/drm_vma_manager.h>
+
+#include <linux/dma-attrs.h>
+
+#include "rockchip_drm_drv.h"
+#include "rockchip_drm_gem.h"
+
+static int rockchip_gem_alloc_buf(struct rockchip_gem_object *rk_obj)
+{
+	struct drm_gem_object *obj = &rk_obj->base;
+	struct drm_device *drm = obj->dev;
+
+	init_dma_attrs(&rk_obj->dma_attrs);
+	dma_set_attr(DMA_ATTR_WRITE_COMBINE, &rk_obj->dma_attrs);
+
+	/* TODO(djkurtz): Use DMA_ATTR_NO_KERNEL_MAPPING except for fbdev */
+	rk_obj->kvaddr = dma_alloc_attrs(drm->dev, obj->size,
+					 &rk_obj->dma_addr, GFP_KERNEL,
+					 &rk_obj->dma_attrs);
+	if (IS_ERR(rk_obj->kvaddr)) {
+		int ret = PTR_ERR(rk_obj->kvaddr);
+
+		DRM_ERROR("failed to allocate %#x byte dma buffer, %d",
+			  obj->size, ret);
+		return ret;
+	}
+
+	return 0;
+}
+
+static void rockchip_gem_free_buf(struct rockchip_gem_object *rk_obj)
+{
+	struct drm_gem_object *obj = &rk_obj->base;
+	struct drm_device *drm = obj->dev;
+
+	dma_free_attrs(drm->dev, obj->size, rk_obj->kvaddr, rk_obj->dma_addr,
+		       &rk_obj->dma_attrs);
+}
+
+int rockchip_gem_mmap_buf(struct drm_gem_object *obj,
+			  struct vm_area_struct *vma)
+{
+	struct rockchip_gem_object *rk_obj = to_rockchip_obj(obj);
+	struct drm_device *drm = obj->dev;
+	unsigned long vm_size;
+
+	vma->vm_flags |= VM_IO | VM_DONTEXPAND | VM_DONTDUMP;
+	vm_size = vma->vm_end - vma->vm_start;
+
+	if (vm_size > obj->size)
+		return -EINVAL;
+
+	return dma_mmap_attrs(drm->dev, vma, rk_obj->kvaddr, rk_obj->dma_addr,
+			     obj->size, &rk_obj->dma_attrs);
+}
+
+/* drm driver mmap file operations */
+int rockchip_gem_mmap(struct file *filp, struct vm_area_struct *vma)
+{
+	struct drm_file *priv = filp->private_data;
+	struct drm_device *dev = priv->minor->dev;
+	struct drm_gem_object *obj;
+	struct drm_vma_offset_node *node;
+	int ret;
+
+	if (drm_device_is_unplugged(dev))
+		return -ENODEV;
+
+	mutex_lock(&dev->struct_mutex);
+
+	node = drm_vma_offset_exact_lookup(dev->vma_offset_manager,
+					   vma->vm_pgoff,
+					   vma_pages(vma));
+	if (!node) {
+		mutex_unlock(&dev->struct_mutex);
+		DRM_ERROR("failed to find vma node.\n");
+		return -EINVAL;
+	} else if (!drm_vma_node_is_allowed(node, filp)) {
+		mutex_unlock(&dev->struct_mutex);
+		return -EACCES;
+	}
+
+	obj = container_of(node, struct drm_gem_object, vma_node);
+	ret = rockchip_gem_mmap_buf(obj, vma);
+
+	mutex_unlock(&dev->struct_mutex);
+
+	return ret;
+}
+
+struct rockchip_gem_object *
+	rockchip_gem_create_object(struct drm_device *drm, unsigned int size)
+{
+	struct rockchip_gem_object *rk_obj;
+	struct drm_gem_object *obj;
+	int ret;
+
+	size = round_up(size, PAGE_SIZE);
+
+	rk_obj = kzalloc(sizeof(*rk_obj), GFP_KERNEL);
+	if (!rk_obj)
+		return ERR_PTR(-ENOMEM);
+
+	obj = &rk_obj->base;
+
+	drm_gem_private_object_init(drm, obj, size);
+
+	ret = rockchip_gem_alloc_buf(rk_obj);
+	if (ret)
+		goto err_free_rk_obj;
+
+	return rk_obj;
+
+err_free_rk_obj:
+	kfree(rk_obj);
+	return ERR_PTR(ret);
+}
+
+/*
+ * rockchip_gem_free_object - (struct drm_driver)->gem_free_object callback
+ * function
+ */
+void rockchip_gem_free_object(struct drm_gem_object *obj)
+{
+	struct rockchip_gem_object *rk_obj;
+
+	drm_gem_free_mmap_offset(obj);
+
+	rk_obj = to_rockchip_obj(obj);
+
+	rockchip_gem_free_buf(rk_obj);
+
+	kfree(rk_obj);
+}
+
+/*
+ * rockchip_gem_create_with_handle - allocate an object with the given
+ * size and create a gem handle on it
+ *
+ * returns a struct rockchip_gem_object* on success or ERR_PTR values
+ * on failure.
+ */
+static struct rockchip_gem_object *
+rockchip_gem_create_with_handle(struct drm_file *file_priv,
+				struct drm_device *drm, unsigned int size,
+				unsigned int *handle)
+{
+	struct rockchip_gem_object *rk_obj;
+	struct drm_gem_object *obj;
+	int ret;
+
+	rk_obj = rockchip_gem_create_object(drm, size);
+	if (IS_ERR(rk_obj))
+		return ERR_CAST(rk_obj);
+
+	obj = &rk_obj->base;
+
+	/*
+	 * allocate a id of idr table where the obj is registered
+	 * and handle has the id what user can see.
+	 */
+	ret = drm_gem_handle_create(file_priv, obj, handle);
+	if (ret)
+		goto err_handle_create;
+
+	/* drop reference from allocate - handle holds it now. */
+	drm_gem_object_unreference_unlocked(obj);
+
+	return rk_obj;
+
+err_handle_create:
+	rockchip_gem_free_object(obj);
+
+	return ERR_PTR(ret);
+}
+
+int rockchip_gem_dumb_map_offset(struct drm_file *file_priv,
+				 struct drm_device *dev, uint32_t handle,
+				 uint64_t *offset)
+{
+	struct drm_gem_object *obj;
+	int ret;
+
+	mutex_lock(&dev->struct_mutex);
+
+	obj = drm_gem_object_lookup(dev, file_priv, handle);
+	if (!obj) {
+		DRM_ERROR("failed to lookup gem object.\n");
+		ret = -EINVAL;
+		goto unlock;
+	}
+
+	ret = drm_gem_create_mmap_offset(obj);
+	if (ret)
+		goto out;
+
+	*offset = drm_vma_node_offset_addr(&obj->vma_node);
+	DRM_DEBUG_KMS("offset = 0x%llx\n", *offset);
+
+out:
+	drm_gem_object_unreference(obj);
+unlock:
+	mutex_unlock(&dev->struct_mutex);
+	return ret;
+}
+
+/*
+ * rockchip_gem_dumb_create - (struct drm_driver)->dumb_create callback
+ * function
+ *
+ * This aligns the pitch and size arguments to the minimum required. wrap
+ * this into your own function if you need bigger alignment.
+ */
+int rockchip_gem_dumb_create(struct drm_file *file_priv,
+			     struct drm_device *dev,
+			     struct drm_mode_create_dumb *args)
+{
+	struct rockchip_gem_object *rk_obj;
+	int min_pitch = DIV_ROUND_UP(args->width * args->bpp, 8);
+
+	/*
+	 * align to 64 bytes since Mali requires it.
+	 */
+	min_pitch = ALIGN(min_pitch, 64);
+
+	if (args->pitch < min_pitch)
+		args->pitch = min_pitch;
+
+	if (args->size < args->pitch * args->height)
+		args->size = args->pitch * args->height;
+
+	rk_obj = rockchip_gem_create_with_handle(file_priv, dev, args->size,
+						 &args->handle);
+
+	return PTR_ERR_OR_ZERO(rk_obj);
+}
+
+/*
+ * Allocate a sg_table for this GEM object.
+ * Note: Both the table's contents, and the sg_table itself must be freed by
+ *       the caller.
+ * Returns a pointer to the newly allocated sg_table, or an ERR_PTR() error.
+ */
+struct sg_table *rockchip_gem_prime_get_sg_table(struct drm_gem_object *obj)
+{
+	struct rockchip_gem_object *rk_obj = to_rockchip_obj(obj);
+	struct drm_device *drm = obj->dev;
+	struct sg_table *sgt;
+	int ret;
+
+	sgt = kzalloc(sizeof(*sgt), GFP_KERNEL);
+	if (!sgt)
+		return ERR_PTR(-ENOMEM);
+
+	ret = dma_get_sgtable_attrs(drm->dev, sgt, rk_obj->kvaddr,
+				    rk_obj->dma_addr, obj->size,
+				    &rk_obj->dma_attrs);
+	if (ret) {
+		DRM_ERROR("failed to allocate sgt, %d\n", ret);
+		kfree(sgt);
+		return ERR_PTR(ret);
+	}
+
+	return sgt;
+}
+
+void *rockchip_gem_prime_vmap(struct drm_gem_object *obj)
+{
+	struct rockchip_gem_object *rk_obj = to_rockchip_obj(obj);
+
+	return rk_obj->kvaddr;
+}
+
+void rockchip_gem_prime_vunmap(struct drm_gem_object *obj, void *vaddr)
+{
+	/* Nothing to do */
+}
diff --git a/drivers/gpu/drm/rockchip/rockchip_drm_gem.h b/drivers/gpu/drm/rockchip/rockchip_drm_gem.h
new file mode 100644
index 000000000000..67bcebe90003
--- /dev/null
+++ b/drivers/gpu/drm/rockchip/rockchip_drm_gem.h
@@ -0,0 +1,54 @@
+/*
+ * Copyright (C) Fuzhou Rockchip Electronics Co.Ltd
+ * Author:Mark Yao <mark.yao@rock-chips.com>
+ *
+ * This software is licensed under the terms of the GNU General Public
+ * License version 2, as published by the Free Software Foundation, and
+ * may be copied, distributed, and modified under those terms.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ */
+
+#ifndef _ROCKCHIP_DRM_GEM_H
+#define _ROCKCHIP_DRM_GEM_H
+
+#define to_rockchip_obj(x) container_of(x, struct rockchip_gem_object, base)
+
+struct rockchip_gem_object {
+	struct drm_gem_object base;
+	unsigned int flags;
+
+	void *kvaddr;
+	dma_addr_t dma_addr;
+	struct dma_attrs dma_attrs;
+};
+
+struct sg_table *rockchip_gem_prime_get_sg_table(struct drm_gem_object *obj);
+struct drm_gem_object *
+rockchip_gem_prime_import_sg_table(struct drm_device *dev, size_t size,
+				   struct sg_table *sgt);
+void *rockchip_gem_prime_vmap(struct drm_gem_object *obj);
+void rockchip_gem_prime_vunmap(struct drm_gem_object *obj, void *vaddr);
+
+/* drm driver mmap file operations */
+int rockchip_gem_mmap(struct file *filp, struct vm_area_struct *vma);
+
+/* mmap a gem object to userspace. */
+int rockchip_gem_mmap_buf(struct drm_gem_object *obj,
+			  struct vm_area_struct *vma);
+
+struct rockchip_gem_object *
+	rockchip_gem_create_object(struct drm_device *drm, unsigned int size);
+
+void rockchip_gem_free_object(struct drm_gem_object *obj);
+
+int rockchip_gem_dumb_create(struct drm_file *file_priv,
+			     struct drm_device *dev,
+			     struct drm_mode_create_dumb *args);
+int rockchip_gem_dumb_map_offset(struct drm_file *file_priv,
+				 struct drm_device *dev, uint32_t handle,
+				 uint64_t *offset);
+#endif /* _ROCKCHIP_DRM_GEM_H */
diff --git a/drivers/gpu/drm/rockchip/rockchip_drm_vop.c b/drivers/gpu/drm/rockchip/rockchip_drm_vop.c
new file mode 100644
index 000000000000..e7ca25b3fb38
--- /dev/null
+++ b/drivers/gpu/drm/rockchip/rockchip_drm_vop.c
@@ -0,0 +1,1455 @@
+/*
+ * Copyright (C) Fuzhou Rockchip Electronics Co.Ltd
+ * Author:Mark Yao <mark.yao@rock-chips.com>
+ *
+ * This software is licensed under the terms of the GNU General Public
+ * License version 2, as published by the Free Software Foundation, and
+ * may be copied, distributed, and modified under those terms.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ */
+
+#include <drm/drm.h>
+#include <drm/drmP.h>
+#include <drm/drm_crtc.h>
+#include <drm/drm_crtc_helper.h>
+#include <drm/drm_plane_helper.h>
+
+#include <linux/kernel.h>
+#include <linux/platform_device.h>
+#include <linux/clk.h>
+#include <linux/of.h>
+#include <linux/of_device.h>
+#include <linux/pm_runtime.h>
+#include <linux/component.h>
+
+#include <linux/reset.h>
+#include <linux/delay.h>
+
+#include "rockchip_drm_drv.h"
+#include "rockchip_drm_gem.h"
+#include "rockchip_drm_fb.h"
+#include "rockchip_drm_vop.h"
+
+#define VOP_REG(off, _mask, s) \
+		{.offset = off, \
+		 .mask = _mask, \
+		 .shift = s,}
+
+#define __REG_SET_RELAXED(x, off, mask, shift, v) \
+		vop_mask_write_relaxed(x, off, (mask) << shift, (v) << shift)
+#define __REG_SET_NORMAL(x, off, mask, shift, v) \
+		vop_mask_write(x, off, (mask) << shift, (v) << shift)
+
+#define REG_SET(x, base, reg, v, mode) \
+		__REG_SET_##mode(x, base + reg.offset, reg.mask, reg.shift, v)
+
+#define VOP_WIN_SET(x, win, name, v) \
+		REG_SET(x, win->base, win->phy->name, v, RELAXED)
+#define VOP_CTRL_SET(x, name, v) \
+		REG_SET(x, 0, (x)->data->ctrl->name, v, NORMAL)
+
+#define VOP_WIN_GET(x, win, name) \
+		vop_read_reg(x, win->base, &win->phy->name)
+
+#define VOP_WIN_GET_YRGBADDR(vop, win) \
+		vop_readl(vop, win->base + win->phy->yrgb_mst.offset)
+
+#define to_vop(x) container_of(x, struct vop, crtc)
+#define to_vop_win(x) container_of(x, struct vop_win, base)
+
+struct vop_win_state {
+	struct list_head head;
+	struct drm_framebuffer *fb;
+	dma_addr_t yrgb_mst;
+	struct drm_pending_vblank_event *event;
+};
+
+struct vop_win {
+	struct drm_plane base;
+	const struct vop_win_data *data;
+	struct vop *vop;
+
+	struct list_head pending;
+	struct vop_win_state *active;
+};
+
+struct vop {
+	struct drm_crtc crtc;
+	struct device *dev;
+	struct drm_device *drm_dev;
+	unsigned int dpms;
+
+	int connector_type;
+	int connector_out_mode;
+
+	/* mutex vsync_ work */
+	struct mutex vsync_mutex;
+	bool vsync_work_pending;
+
+	const struct vop_data *data;
+
+	uint32_t *regsbak;
+	void __iomem *regs;
+
+	/* physical map length of vop register */
+	uint32_t len;
+
+	/* one time only one process allowed to config the register */
+	spinlock_t reg_lock;
+	/* lock vop irq reg */
+	spinlock_t irq_lock;
+
+	unsigned int irq;
+
+	/* vop AHP clk */
+	struct clk *hclk;
+	/* vop dclk */
+	struct clk *dclk;
+	/* vop share memory frequency */
+	struct clk *aclk;
+
+	/* vop dclk reset */
+	struct reset_control *dclk_rst;
+
+	int pipe;
+
+	struct vop_win win[];
+};
+
+enum vop_data_format {
+	VOP_FMT_ARGB8888 = 0,
+	VOP_FMT_RGB888,
+	VOP_FMT_RGB565,
+	VOP_FMT_YUV420SP = 4,
+	VOP_FMT_YUV422SP,
+	VOP_FMT_YUV444SP,
+};
+
+struct vop_reg_data {
+	uint32_t offset;
+	uint32_t value;
+};
+
+struct vop_reg {
+	uint32_t offset;
+	uint32_t shift;
+	uint32_t mask;
+};
+
+struct vop_ctrl {
+	struct vop_reg standby;
+	struct vop_reg data_blank;
+	struct vop_reg gate_en;
+	struct vop_reg mmu_en;
+	struct vop_reg rgb_en;
+	struct vop_reg edp_en;
+	struct vop_reg hdmi_en;
+	struct vop_reg mipi_en;
+	struct vop_reg out_mode;
+	struct vop_reg dither_down;
+	struct vop_reg dither_up;
+	struct vop_reg pin_pol;
+
+	struct vop_reg htotal_pw;
+	struct vop_reg hact_st_end;
+	struct vop_reg vtotal_pw;
+	struct vop_reg vact_st_end;
+	struct vop_reg hpost_st_end;
+	struct vop_reg vpost_st_end;
+};
+
+struct vop_win_phy {
+	const uint32_t *data_formats;
+	uint32_t nformats;
+
+	struct vop_reg enable;
+	struct vop_reg format;
+	struct vop_reg act_info;
+	struct vop_reg dsp_info;
+	struct vop_reg dsp_st;
+	struct vop_reg yrgb_mst;
+	struct vop_reg uv_mst;
+	struct vop_reg yrgb_vir;
+	struct vop_reg uv_vir;
+
+	struct vop_reg dst_alpha_ctl;
+	struct vop_reg src_alpha_ctl;
+};
+
+struct vop_win_data {
+	uint32_t base;
+	const struct vop_win_phy *phy;
+	enum drm_plane_type type;
+};
+
+struct vop_data {
+	const struct vop_reg_data *init_table;
+	unsigned int table_size;
+	const struct vop_ctrl *ctrl;
+	const struct vop_win_data *win;
+	unsigned int win_size;
+};
+
+static const uint32_t formats_01[] = {
+	DRM_FORMAT_XRGB8888,
+	DRM_FORMAT_ARGB8888,
+	DRM_FORMAT_RGB888,
+	DRM_FORMAT_RGB565,
+	DRM_FORMAT_NV12,
+	DRM_FORMAT_NV16,
+	DRM_FORMAT_NV24,
+};
+
+static const uint32_t formats_234[] = {
+	DRM_FORMAT_XRGB8888,
+	DRM_FORMAT_ARGB8888,
+	DRM_FORMAT_RGB888,
+	DRM_FORMAT_RGB565,
+};
+
+static const struct vop_win_phy win01_data = {
+	.data_formats = formats_01,
+	.nformats = ARRAY_SIZE(formats_01),
+	.enable = VOP_REG(WIN0_CTRL0, 0x1, 0),
+	.format = VOP_REG(WIN0_CTRL0, 0x7, 1),
+	.act_info = VOP_REG(WIN0_ACT_INFO, 0x1fff1fff, 0),
+	.dsp_info = VOP_REG(WIN0_DSP_INFO, 0x0fff0fff, 0),
+	.dsp_st = VOP_REG(WIN0_DSP_ST, 0x1fff1fff, 0),
+	.yrgb_mst = VOP_REG(WIN0_YRGB_MST, 0xffffffff, 0),
+	.uv_mst = VOP_REG(WIN0_CBR_MST, 0xffffffff, 0),
+	.yrgb_vir = VOP_REG(WIN0_VIR, 0x3fff, 0),
+	.uv_vir = VOP_REG(WIN0_VIR, 0x3fff, 16),
+	.src_alpha_ctl = VOP_REG(WIN0_SRC_ALPHA_CTRL, 0xff, 0),
+	.dst_alpha_ctl = VOP_REG(WIN0_DST_ALPHA_CTRL, 0xff, 0),
+};
+
+static const struct vop_win_phy win23_data = {
+	.data_formats = formats_234,
+	.nformats = ARRAY_SIZE(formats_234),
+	.enable = VOP_REG(WIN2_CTRL0, 0x1, 0),
+	.format = VOP_REG(WIN2_CTRL0, 0x7, 1),
+	.dsp_info = VOP_REG(WIN2_DSP_INFO0, 0x0fff0fff, 0),
+	.dsp_st = VOP_REG(WIN2_DSP_ST0, 0x1fff1fff, 0),
+	.yrgb_mst = VOP_REG(WIN2_MST0, 0xffffffff, 0),
+	.yrgb_vir = VOP_REG(WIN2_VIR0_1, 0x1fff, 0),
+	.src_alpha_ctl = VOP_REG(WIN2_SRC_ALPHA_CTRL, 0xff, 0),
+	.dst_alpha_ctl = VOP_REG(WIN2_DST_ALPHA_CTRL, 0xff, 0),
+};
+
+static const struct vop_win_phy cursor_data = {
+	.data_formats = formats_234,
+	.nformats = ARRAY_SIZE(formats_234),
+	.enable = VOP_REG(HWC_CTRL0, 0x1, 0),
+	.format = VOP_REG(HWC_CTRL0, 0x7, 1),
+	.dsp_st = VOP_REG(HWC_DSP_ST, 0x1fff1fff, 0),
+	.yrgb_mst = VOP_REG(HWC_MST, 0xffffffff, 0),
+};
+
+static const struct vop_ctrl ctrl_data = {
+	.standby = VOP_REG(SYS_CTRL, 0x1, 22),
+	.gate_en = VOP_REG(SYS_CTRL, 0x1, 23),
+	.mmu_en = VOP_REG(SYS_CTRL, 0x1, 20),
+	.rgb_en = VOP_REG(SYS_CTRL, 0x1, 12),
+	.hdmi_en = VOP_REG(SYS_CTRL, 0x1, 13),
+	.edp_en = VOP_REG(SYS_CTRL, 0x1, 14),
+	.mipi_en = VOP_REG(SYS_CTRL, 0x1, 15),
+	.dither_down = VOP_REG(DSP_CTRL1, 0xf, 1),
+	.dither_up = VOP_REG(DSP_CTRL1, 0x1, 6),
+	.data_blank = VOP_REG(DSP_CTRL0, 0x1, 19),
+	.out_mode = VOP_REG(DSP_CTRL0, 0xf, 0),
+	.pin_pol = VOP_REG(DSP_CTRL0, 0xf, 4),
+	.htotal_pw = VOP_REG(DSP_HTOTAL_HS_END, 0x1fff1fff, 0),
+	.hact_st_end = VOP_REG(DSP_HACT_ST_END, 0x1fff1fff, 0),
+	.vtotal_pw = VOP_REG(DSP_VTOTAL_VS_END, 0x1fff1fff, 0),
+	.vact_st_end = VOP_REG(DSP_VACT_ST_END, 0x1fff1fff, 0),
+	.hpost_st_end = VOP_REG(POST_DSP_HACT_INFO, 0x1fff1fff, 0),
+	.vpost_st_end = VOP_REG(POST_DSP_VACT_INFO, 0x1fff1fff, 0),
+};
+
+static const struct vop_reg_data vop_init_reg_table[] = {
+	{SYS_CTRL, 0x00c00000},
+	{DSP_CTRL0, 0x00000000},
+	{WIN0_CTRL0, 0x00000080},
+	{WIN1_CTRL0, 0x00000080},
+};
+
+/*
+ * Note: rk3288 has a dedicated 'cursor' window, however, that window requires
+ * special support to get alpha blending working.  For now, just use overlay
+ * window 1 for the drm cursor.
+ */
+static const struct vop_win_data rk3288_vop_win_data[] = {
+	{ .base = 0x00, .phy = &win01_data, .type = DRM_PLANE_TYPE_PRIMARY },
+	{ .base = 0x40, .phy = &win01_data, .type = DRM_PLANE_TYPE_CURSOR },
+	{ .base = 0x00, .phy = &win23_data, .type = DRM_PLANE_TYPE_OVERLAY },
+	{ .base = 0x50, .phy = &win23_data, .type = DRM_PLANE_TYPE_OVERLAY },
+	{ .base = 0x00, .phy = &cursor_data, .type = DRM_PLANE_TYPE_OVERLAY },
+};
+
+static const struct vop_data rk3288_vop = {
+	.init_table = vop_init_reg_table,
+	.table_size = ARRAY_SIZE(vop_init_reg_table),
+	.ctrl = &ctrl_data,
+	.win = rk3288_vop_win_data,
+	.win_size = ARRAY_SIZE(rk3288_vop_win_data),
+};
+
+static const struct of_device_id vop_driver_dt_match[] = {
+	{ .compatible = "rockchip,rk3288-vop",
+	  .data = &rk3288_vop },
+	{},
+};
+
+static inline void vop_writel(struct vop *vop, uint32_t offset, uint32_t v)
+{
+	writel(v, vop->regs + offset);
+	vop->regsbak[offset >> 2] = v;
+}
+
+static inline uint32_t vop_readl(struct vop *vop, uint32_t offset)
+{
+	return readl(vop->regs + offset);
+}
+
+static inline uint32_t vop_read_reg(struct vop *vop, uint32_t base,
+				    const struct vop_reg *reg)
+{
+	return (vop_readl(vop, base + reg->offset) >> reg->shift) & reg->mask;
+}
+
+static inline void vop_cfg_done(struct vop *vop)
+{
+	writel(0x01, vop->regs + REG_CFG_DONE);
+}
+
+static inline void vop_mask_write(struct vop *vop, uint32_t offset,
+				  uint32_t mask, uint32_t v)
+{
+	if (mask) {
+		uint32_t cached_val = vop->regsbak[offset >> 2];
+
+		cached_val = (cached_val & ~mask) | v;
+		writel(cached_val, vop->regs + offset);
+		vop->regsbak[offset >> 2] = cached_val;
+	}
+}
+
+static inline void vop_mask_write_relaxed(struct vop *vop, uint32_t offset,
+					  uint32_t mask, uint32_t v)
+{
+	if (mask) {
+		uint32_t cached_val = vop->regsbak[offset >> 2];
+
+		cached_val = (cached_val & ~mask) | v;
+		writel_relaxed(cached_val, vop->regs + offset);
+		vop->regsbak[offset >> 2] = cached_val;
+	}
+}
+
+static enum vop_data_format vop_convert_format(uint32_t format)
+{
+	switch (format) {
+	case DRM_FORMAT_XRGB8888:
+	case DRM_FORMAT_ARGB8888:
+		return VOP_FMT_ARGB8888;
+	case DRM_FORMAT_RGB888:
+		return VOP_FMT_RGB888;
+	case DRM_FORMAT_RGB565:
+		return VOP_FMT_RGB565;
+	case DRM_FORMAT_NV12:
+		return VOP_FMT_YUV420SP;
+	case DRM_FORMAT_NV16:
+		return VOP_FMT_YUV422SP;
+	case DRM_FORMAT_NV24:
+		return VOP_FMT_YUV444SP;
+	default:
+		DRM_ERROR("unsupport format[%08x]\n", format);
+		return -EINVAL;
+	}
+}
+
+static bool is_alpha_support(uint32_t format)
+{
+	switch (format) {
+	case DRM_FORMAT_ARGB8888:
+		return true;
+	default:
+		return false;
+	}
+}
+
+static void vop_enable(struct drm_crtc *crtc)
+{
+	struct vop *vop = to_vop(crtc);
+	int ret;
+
+	ret = clk_enable(vop->hclk);
+	if (ret < 0) {
+		dev_err(vop->dev, "failed to enable hclk - %d\n", ret);
+		return;
+	}
+
+	ret = clk_enable(vop->dclk);
+	if (ret < 0) {
+		dev_err(vop->dev, "failed to enable dclk - %d\n", ret);
+		goto err_disable_hclk;
+	}
+
+	ret = clk_enable(vop->aclk);
+	if (ret < 0) {
+		dev_err(vop->dev, "failed to enable aclk - %d\n", ret);
+		goto err_disable_dclk;
+	}
+
+	/*
+	 * Slave iommu shares power, irq and clock with vop.  It was associated
+	 * automatically with this master device via common driver code.
+	 * Now that we have enabled the clock we attach it to the shared drm
+	 * mapping.
+	 */
+	ret = rockchip_drm_dma_attach_device(vop->drm_dev, vop->dev);
+	if (ret) {
+		dev_err(vop->dev, "failed to attach dma mapping, %d\n", ret);
+		goto err_disable_aclk;
+	}
+
+	spin_lock(&vop->reg_lock);
+
+	VOP_CTRL_SET(vop, standby, 0);
+
+	spin_unlock(&vop->reg_lock);
+
+	enable_irq(vop->irq);
+
+	drm_vblank_on(vop->drm_dev, vop->pipe);
+
+	return;
+
+err_disable_aclk:
+	clk_disable(vop->aclk);
+err_disable_dclk:
+	clk_disable(vop->dclk);
+err_disable_hclk:
+	clk_disable(vop->hclk);
+}
+
+static void vop_disable(struct drm_crtc *crtc)
+{
+	struct vop *vop = to_vop(crtc);
+
+	drm_vblank_off(crtc->dev, vop->pipe);
+
+	disable_irq(vop->irq);
+
+	/*
+	 * TODO: Since standby doesn't take effect until the next vblank,
+	 * when we turn off dclk below, the vop is probably still active.
+	 */
+	spin_lock(&vop->reg_lock);
+
+	VOP_CTRL_SET(vop, standby, 1);
+
+	spin_unlock(&vop->reg_lock);
+	/*
+	 * disable dclk to stop frame scan, so we can safely detach iommu,
+	 */
+	clk_disable(vop->dclk);
+
+	rockchip_drm_dma_detach_device(vop->drm_dev, vop->dev);
+
+	clk_disable(vop->aclk);
+	clk_disable(vop->hclk);
+}
+
+/*
+ * Caller must hold vsync_mutex.
+ */
+static struct drm_framebuffer *vop_win_last_pending_fb(struct vop_win *vop_win)
+{
+	struct vop_win_state *last;
+	struct vop_win_state *active = vop_win->active;
+
+	if (list_empty(&vop_win->pending))
+		return active ? active->fb : NULL;
+
+	last = list_last_entry(&vop_win->pending, struct vop_win_state, head);
+	return last ? last->fb : NULL;
+}
+
+/*
+ * Caller must hold vsync_mutex.
+ */
+static int vop_win_queue_fb(struct vop_win *vop_win,
+			    struct drm_framebuffer *fb, dma_addr_t yrgb_mst,
+			    struct drm_pending_vblank_event *event)
+{
+	struct vop_win_state *state;
+
+	state = kzalloc(sizeof(*state), GFP_KERNEL);
+	if (!state)
+		return -ENOMEM;
+
+	state->fb = fb;
+	state->yrgb_mst = yrgb_mst;
+	state->event = event;
+
+	list_add_tail(&state->head, &vop_win->pending);
+
+	return 0;
+}
+
+static int vop_update_plane_event(struct drm_plane *plane,
+				  struct drm_crtc *crtc,
+				  struct drm_framebuffer *fb, int crtc_x,
+				  int crtc_y, unsigned int crtc_w,
+				  unsigned int crtc_h, uint32_t src_x,
+				  uint32_t src_y, uint32_t src_w,
+				  uint32_t src_h,
+				  struct drm_pending_vblank_event *event)
+{
+	struct vop_win *vop_win = to_vop_win(plane);
+	const struct vop_win_data *win = vop_win->data;
+	struct vop *vop = to_vop(crtc);
+	struct drm_gem_object *obj;
+	struct rockchip_gem_object *rk_obj;
+	unsigned long offset;
+	unsigned int actual_w;
+	unsigned int actual_h;
+	unsigned int dsp_stx;
+	unsigned int dsp_sty;
+	unsigned int y_vir_stride;
+	dma_addr_t yrgb_mst;
+	enum vop_data_format format;
+	uint32_t val;
+	bool is_alpha;
+	bool visible;
+	int ret;
+	struct drm_rect dest = {
+		.x1 = crtc_x,
+		.y1 = crtc_y,
+		.x2 = crtc_x + crtc_w,
+		.y2 = crtc_y + crtc_h,
+	};
+	struct drm_rect src = {
+		/* 16.16 fixed point */
+		.x1 = src_x,
+		.y1 = src_y,
+		.x2 = src_x + src_w,
+		.y2 = src_y + src_h,
+	};
+	const struct drm_rect clip = {
+		.x2 = crtc->mode.hdisplay,
+		.y2 = crtc->mode.vdisplay,
+	};
+	bool can_position = plane->type != DRM_PLANE_TYPE_PRIMARY;
+
+	ret = drm_plane_helper_check_update(plane, crtc, fb,
+					    &src, &dest, &clip,
+					    DRM_PLANE_HELPER_NO_SCALING,
+					    DRM_PLANE_HELPER_NO_SCALING,
+					    can_position, false, &visible);
+	if (ret)
+		return ret;
+
+	if (!visible)
+		return 0;
+
+	is_alpha = is_alpha_support(fb->pixel_format);
+	format = vop_convert_format(fb->pixel_format);
+	if (format < 0)
+		return format;
+
+	obj = rockchip_fb_get_gem_obj(fb, 0);
+	if (!obj) {
+		DRM_ERROR("fail to get rockchip gem object from framebuffer\n");
+		return -EINVAL;
+	}
+
+	rk_obj = to_rockchip_obj(obj);
+
+	actual_w = (src.x2 - src.x1) >> 16;
+	actual_h = (src.y2 - src.y1) >> 16;
+	crtc_x = max(0, crtc_x);
+	crtc_y = max(0, crtc_y);
+
+	dsp_stx = crtc_x + crtc->mode.htotal - crtc->mode.hsync_start;
+	dsp_sty = crtc_y + crtc->mode.vtotal - crtc->mode.vsync_start;
+
+	offset = (src.x1 >> 16) * (fb->bits_per_pixel >> 3);
+	offset += (src.y1 >> 16) * fb->pitches[0];
+	yrgb_mst = rk_obj->dma_addr + offset;
+
+	y_vir_stride = fb->pitches[0] / (fb->bits_per_pixel >> 3);
+
+	/*
+	 * If this plane update changes the plane's framebuffer, (or more
+	 * precisely, if this update has a different framebuffer than the last
+	 * update), enqueue it so we can track when it completes.
+	 *
+	 * Only when we discover that this update has completed, can we
+	 * unreference any previous framebuffers.
+	 */
+	mutex_lock(&vop->vsync_mutex);
+	if (fb != vop_win_last_pending_fb(vop_win)) {
+		ret = drm_vblank_get(plane->dev, vop->pipe);
+		if (ret) {
+			DRM_ERROR("failed to get vblank, %d\n", ret);
+			mutex_unlock(&vop->vsync_mutex);
+			return ret;
+		}
+
+		drm_framebuffer_reference(fb);
+
+		ret = vop_win_queue_fb(vop_win, fb, yrgb_mst, event);
+		if (ret) {
+			drm_vblank_put(plane->dev, vop->pipe);
+			mutex_unlock(&vop->vsync_mutex);
+			return ret;
+		}
+
+		vop->vsync_work_pending = true;
+	}
+	mutex_unlock(&vop->vsync_mutex);
+
+	spin_lock(&vop->reg_lock);
+
+	VOP_WIN_SET(vop, win, format, format);
+	VOP_WIN_SET(vop, win, yrgb_vir, y_vir_stride);
+	VOP_WIN_SET(vop, win, yrgb_mst, yrgb_mst);
+	val = (actual_h - 1) << 16;
+	val |= (actual_w - 1) & 0xffff;
+	VOP_WIN_SET(vop, win, act_info, val);
+	VOP_WIN_SET(vop, win, dsp_info, val);
+	val = (dsp_sty - 1) << 16;
+	val |= (dsp_stx - 1) & 0xffff;
+	VOP_WIN_SET(vop, win, dsp_st, val);
+
+	if (is_alpha) {
+		VOP_WIN_SET(vop, win, dst_alpha_ctl,
+			    DST_FACTOR_M0(ALPHA_SRC_INVERSE));
+		val = SRC_ALPHA_EN(1) | SRC_COLOR_M0(ALPHA_SRC_PRE_MUL) |
+			SRC_ALPHA_M0(ALPHA_STRAIGHT) |
+			SRC_BLEND_M0(ALPHA_PER_PIX) |
+			SRC_ALPHA_CAL_M0(ALPHA_NO_SATURATION) |
+			SRC_FACTOR_M0(ALPHA_ONE);
+		VOP_WIN_SET(vop, win, src_alpha_ctl, val);
+	} else {
+		VOP_WIN_SET(vop, win, src_alpha_ctl, SRC_ALPHA_EN(0));
+	}
+
+	VOP_WIN_SET(vop, win, enable, 1);
+
+	vop_cfg_done(vop);
+	spin_unlock(&vop->reg_lock);
+
+	return 0;
+}
+
+static int vop_update_plane(struct drm_plane *plane, struct drm_crtc *crtc,
+			    struct drm_framebuffer *fb, int crtc_x, int crtc_y,
+			    unsigned int crtc_w, unsigned int crtc_h,
+			    uint32_t src_x, uint32_t src_y, uint32_t src_w,
+			    uint32_t src_h)
+{
+	return vop_update_plane_event(plane, crtc, fb, crtc_x, crtc_y, crtc_w,
+				      crtc_h, src_x, src_y, src_w, src_h,
+				      NULL);
+}
+
+static int vop_update_primary_plane(struct drm_crtc *crtc,
+				    struct drm_pending_vblank_event *event)
+{
+	unsigned int crtc_w, crtc_h;
+
+	crtc_w = crtc->primary->fb->width - crtc->x;
+	crtc_h = crtc->primary->fb->height - crtc->y;
+
+	return vop_update_plane_event(crtc->primary, crtc, crtc->primary->fb,
+				      0, 0, crtc_w, crtc_h, crtc->x << 16,
+				      crtc->y << 16, crtc_w << 16,
+				      crtc_h << 16, event);
+}
+
+static int vop_disable_plane(struct drm_plane *plane)
+{
+	struct vop_win *vop_win = to_vop_win(plane);
+	const struct vop_win_data *win = vop_win->data;
+	struct vop *vop;
+	int ret;
+
+	if (!plane->crtc)
+		return 0;
+
+	vop = to_vop(plane->crtc);
+
+	ret = drm_vblank_get(plane->dev, vop->pipe);
+	if (ret) {
+		DRM_ERROR("failed to get vblank, %d\n", ret);
+		return ret;
+	}
+
+	mutex_lock(&vop->vsync_mutex);
+
+	ret = vop_win_queue_fb(vop_win, NULL, 0, NULL);
+	if (ret) {
+		drm_vblank_put(plane->dev, vop->pipe);
+		mutex_unlock(&vop->vsync_mutex);
+		return ret;
+	}
+
+	vop->vsync_work_pending = true;
+	mutex_unlock(&vop->vsync_mutex);
+
+	spin_lock(&vop->reg_lock);
+	VOP_WIN_SET(vop, win, enable, 0);
+	vop_cfg_done(vop);
+	spin_unlock(&vop->reg_lock);
+
+	return 0;
+}
+
+static void vop_plane_destroy(struct drm_plane *plane)
+{
+	vop_disable_plane(plane);
+	drm_plane_cleanup(plane);
+}
+
+static const struct drm_plane_funcs vop_plane_funcs = {
+	.update_plane = vop_update_plane,
+	.disable_plane = vop_disable_plane,
+	.destroy = vop_plane_destroy,
+};
+
+int rockchip_drm_crtc_mode_config(struct drm_crtc *crtc,
+				  int connector_type,
+				  int out_mode)
+{
+	struct vop *vop = to_vop(crtc);
+
+	vop->connector_type = connector_type;
+	vop->connector_out_mode = out_mode;
+
+	return 0;
+}
+
+static int vop_crtc_enable_vblank(struct drm_crtc *crtc)
+{
+	struct vop *vop = to_vop(crtc);
+	unsigned long flags;
+
+	if (vop->dpms != DRM_MODE_DPMS_ON)
+		return -EPERM;
+
+	spin_lock_irqsave(&vop->irq_lock, flags);
+
+	vop_mask_write(vop, INTR_CTRL0, FS_INTR_MASK, FS_INTR_EN(1));
+
+	spin_unlock_irqrestore(&vop->irq_lock, flags);
+
+	return 0;
+}
+
+static void vop_crtc_disable_vblank(struct drm_crtc *crtc)
+{
+	struct vop *vop = to_vop(crtc);
+	unsigned long flags;
+
+	if (vop->dpms != DRM_MODE_DPMS_ON)
+		return;
+	spin_lock_irqsave(&vop->irq_lock, flags);
+	vop_mask_write(vop, INTR_CTRL0, FS_INTR_MASK, FS_INTR_EN(0));
+	spin_unlock_irqrestore(&vop->irq_lock, flags);
+}
+
+static const struct rockchip_crtc_funcs private_crtc_funcs = {
+	.enable_vblank = vop_crtc_enable_vblank,
+	.disable_vblank = vop_crtc_disable_vblank,
+};
+
+static void vop_crtc_dpms(struct drm_crtc *crtc, int mode)
+{
+	struct vop *vop = to_vop(crtc);
+
+	DRM_DEBUG_KMS("crtc[%d] mode[%d]\n", crtc->base.id, mode);
+
+	if (vop->dpms == mode) {
+		DRM_DEBUG_KMS("desired dpms mode is same as previous one.\n");
+		return;
+	}
+
+	switch (mode) {
+	case DRM_MODE_DPMS_ON:
+		vop_enable(crtc);
+		break;
+	case DRM_MODE_DPMS_STANDBY:
+	case DRM_MODE_DPMS_SUSPEND:
+	case DRM_MODE_DPMS_OFF:
+		vop_disable(crtc);
+		break;
+	default:
+		DRM_DEBUG_KMS("unspecified mode %d\n", mode);
+		break;
+	}
+
+	vop->dpms = mode;
+}
+
+static void vop_crtc_prepare(struct drm_crtc *crtc)
+{
+	vop_crtc_dpms(crtc, DRM_MODE_DPMS_ON);
+}
+
+static bool vop_crtc_mode_fixup(struct drm_crtc *crtc,
+				const struct drm_display_mode *mode,
+				struct drm_display_mode *adjusted_mode)
+{
+	if (adjusted_mode->htotal == 0 || adjusted_mode->vtotal == 0)
+		return false;
+
+	return true;
+}
+
+static int vop_crtc_mode_set_base(struct drm_crtc *crtc, int x, int y,
+				  struct drm_framebuffer *old_fb)
+{
+	int ret;
+
+	crtc->x = x;
+	crtc->y = y;
+
+	ret = vop_update_primary_plane(crtc, NULL);
+	if (ret < 0) {
+		DRM_ERROR("fail to update plane\n");
+		return ret;
+	}
+
+	return 0;
+}
+
+static int vop_crtc_mode_set(struct drm_crtc *crtc,
+			     struct drm_display_mode *mode,
+			     struct drm_display_mode *adjusted_mode,
+			     int x, int y, struct drm_framebuffer *fb)
+{
+	struct vop *vop = to_vop(crtc);
+	u16 hsync_len = adjusted_mode->hsync_end - adjusted_mode->hsync_start;
+	u16 hdisplay = adjusted_mode->hdisplay;
+	u16 htotal = adjusted_mode->htotal;
+	u16 hact_st = adjusted_mode->htotal - adjusted_mode->hsync_start;
+	u16 hact_end = hact_st + hdisplay;
+	u16 vdisplay = adjusted_mode->vdisplay;
+	u16 vtotal = adjusted_mode->vtotal;
+	u16 vsync_len = adjusted_mode->vsync_end - adjusted_mode->vsync_start;
+	u16 vact_st = adjusted_mode->vtotal - adjusted_mode->vsync_start;
+	u16 vact_end = vact_st + vdisplay;
+	int ret;
+	uint32_t val;
+
+	/*
+	 * disable dclk to stop frame scan, so that we can safe config mode and
+	 * enable iommu.
+	 */
+	clk_disable(vop->dclk);
+
+	switch (vop->connector_type) {
+	case DRM_MODE_CONNECTOR_LVDS:
+		VOP_CTRL_SET(vop, rgb_en, 1);
+		break;
+	case DRM_MODE_CONNECTOR_eDP:
+		VOP_CTRL_SET(vop, edp_en, 1);
+		break;
+	case DRM_MODE_CONNECTOR_HDMIA:
+		VOP_CTRL_SET(vop, hdmi_en, 1);
+		break;
+	default:
+		DRM_ERROR("unsupport connector_type[%d]\n",
+			  vop->connector_type);
+		return -EINVAL;
+	};
+	VOP_CTRL_SET(vop, out_mode, vop->connector_out_mode);
+
+	val = 0x8;
+	val |= (adjusted_mode->flags & DRM_MODE_FLAG_NHSYNC) ? 1 : 0;
+	val |= (adjusted_mode->flags & DRM_MODE_FLAG_NVSYNC) ? (1 << 1) : 0;
+	VOP_CTRL_SET(vop, pin_pol, val);
+
+	VOP_CTRL_SET(vop, htotal_pw, (htotal << 16) | hsync_len);
+	val = hact_st << 16;
+	val |= hact_end;
+	VOP_CTRL_SET(vop, hact_st_end, val);
+	VOP_CTRL_SET(vop, hpost_st_end, val);
+
+	VOP_CTRL_SET(vop, vtotal_pw, (vtotal << 16) | vsync_len);
+	val = vact_st << 16;
+	val |= vact_end;
+	VOP_CTRL_SET(vop, vact_st_end, val);
+	VOP_CTRL_SET(vop, vpost_st_end, val);
+
+	ret = vop_crtc_mode_set_base(crtc, x, y, fb);
+	if (ret)
+		return ret;
+
+	/*
+	 * reset dclk, take all mode config affect, so the clk would run in
+	 * correct frame.
+	 */
+	reset_control_assert(vop->dclk_rst);
+	usleep_range(10, 20);
+	reset_control_deassert(vop->dclk_rst);
+
+	clk_set_rate(vop->dclk, adjusted_mode->clock * 1000);
+	ret = clk_enable(vop->dclk);
+	if (ret < 0) {
+		dev_err(vop->dev, "failed to enable dclk - %d\n", ret);
+		return ret;
+	}
+
+	return 0;
+}
+
+static void vop_crtc_commit(struct drm_crtc *crtc)
+{
+}
+
+static const struct drm_crtc_helper_funcs vop_crtc_helper_funcs = {
+	.dpms = vop_crtc_dpms,
+	.prepare = vop_crtc_prepare,
+	.mode_fixup = vop_crtc_mode_fixup,
+	.mode_set = vop_crtc_mode_set,
+	.mode_set_base = vop_crtc_mode_set_base,
+	.commit = vop_crtc_commit,
+};
+
+static int vop_crtc_page_flip(struct drm_crtc *crtc,
+			      struct drm_framebuffer *fb,
+			      struct drm_pending_vblank_event *event,
+			      uint32_t page_flip_flags)
+{
+	struct vop *vop = to_vop(crtc);
+	struct drm_framebuffer *old_fb = crtc->primary->fb;
+	int ret;
+
+	/* when the page flip is requested, crtc's dpms should be on */
+	if (vop->dpms > DRM_MODE_DPMS_ON) {
+		DRM_DEBUG("failed page flip request at dpms[%d].\n", vop->dpms);
+		return 0;
+	}
+
+	crtc->primary->fb = fb;
+
+	ret = vop_update_primary_plane(crtc, event);
+	if (ret)
+		crtc->primary->fb = old_fb;
+
+	return ret;
+}
+
+static void vop_win_state_complete(struct vop_win *vop_win,
+				   struct vop_win_state *state)
+{
+	struct vop *vop = vop_win->vop;
+	struct drm_crtc *crtc = &vop->crtc;
+	struct drm_device *drm = crtc->dev;
+	unsigned long flags;
+
+	if (state->event) {
+		spin_lock_irqsave(&drm->event_lock, flags);
+		drm_send_vblank_event(drm, -1, state->event);
+		spin_unlock_irqrestore(&drm->event_lock, flags);
+	}
+
+	list_del(&state->head);
+	drm_vblank_put(crtc->dev, vop->pipe);
+}
+
+static void vop_crtc_destroy(struct drm_crtc *crtc)
+{
+	drm_crtc_cleanup(crtc);
+}
+
+static const struct drm_crtc_funcs vop_crtc_funcs = {
+	.set_config = drm_crtc_helper_set_config,
+	.page_flip = vop_crtc_page_flip,
+	.destroy = vop_crtc_destroy,
+};
+
+static bool vop_win_state_is_active(struct vop_win *vop_win,
+				    struct vop_win_state *state)
+{
+	bool active = false;
+
+	if (state->fb) {
+		dma_addr_t yrgb_mst;
+
+		/* check yrgb_mst to tell if pending_fb is now front */
+		yrgb_mst = VOP_WIN_GET_YRGBADDR(vop_win->vop, vop_win->data);
+
+		active = (yrgb_mst == state->yrgb_mst);
+	} else {
+		bool enabled;
+
+		/* if enable bit is clear, plane is now disabled */
+		enabled = VOP_WIN_GET(vop_win->vop, vop_win->data, enable);
+
+		active = (enabled == 0);
+	}
+
+	return active;
+}
+
+static void vop_win_state_destroy(struct vop_win_state *state)
+{
+	struct drm_framebuffer *fb = state->fb;
+
+	if (fb)
+		drm_framebuffer_unreference(fb);
+
+	kfree(state);
+}
+
+static void vop_win_update_state(struct vop_win *vop_win)
+{
+	struct vop_win_state *state, *n, *new_active = NULL;
+
+	/* Check if any pending states are now active */
+	list_for_each_entry(state, &vop_win->pending, head)
+		if (vop_win_state_is_active(vop_win, state)) {
+			new_active = state;
+			break;
+		}
+
+	if (!new_active)
+		return;
+
+	/*
+	 * Destroy any 'skipped' pending states - states that were queued
+	 * before the newly active state.
+	 */
+	list_for_each_entry_safe(state, n, &vop_win->pending, head) {
+		if (state == new_active)
+			break;
+		vop_win_state_complete(vop_win, state);
+		vop_win_state_destroy(state);
+	}
+
+	vop_win_state_complete(vop_win, new_active);
+
+	if (vop_win->active)
+		vop_win_state_destroy(vop_win->active);
+	vop_win->active = new_active;
+}
+
+static bool vop_win_has_pending_state(struct vop_win *vop_win)
+{
+	return !list_empty(&vop_win->pending);
+}
+
+static irqreturn_t vop_isr_thread(int irq, void *data)
+{
+	struct vop *vop = data;
+	const struct vop_data *vop_data = vop->data;
+	unsigned int i;
+
+	mutex_lock(&vop->vsync_mutex);
+
+	if (!vop->vsync_work_pending)
+		goto done;
+
+	vop->vsync_work_pending = false;
+
+	for (i = 0; i < vop_data->win_size; i++) {
+		struct vop_win *vop_win = &vop->win[i];
+
+		vop_win_update_state(vop_win);
+		if (vop_win_has_pending_state(vop_win))
+			vop->vsync_work_pending = true;
+	}
+
+done:
+	mutex_unlock(&vop->vsync_mutex);
+
+	return IRQ_HANDLED;
+}
+
+static irqreturn_t vop_isr(int irq, void *data)
+{
+	struct vop *vop = data;
+	uint32_t intr0_reg, active_irqs;
+	unsigned long flags;
+
+	/*
+	 * INTR_CTRL0 register has interrupt status, enable and clear bits, we
+	 * must hold irq_lock to avoid a race with enable/disable_vblank().
+	*/
+	spin_lock_irqsave(&vop->irq_lock, flags);
+	intr0_reg = vop_readl(vop, INTR_CTRL0);
+	active_irqs = intr0_reg & INTR_MASK;
+	/* Clear all active interrupt sources */
+	if (active_irqs)
+		vop_writel(vop, INTR_CTRL0,
+			   intr0_reg | (active_irqs << INTR_CLR_SHIFT));
+	spin_unlock_irqrestore(&vop->irq_lock, flags);
+
+	/* This is expected for vop iommu irqs, since the irq is shared */
+	if (!active_irqs)
+		return IRQ_NONE;
+
+	/* Only Frame Start Interrupt is enabled; other irqs are spurious. */
+	if (!(active_irqs & FS_INTR)) {
+		DRM_ERROR("Unknown VOP IRQs: %#02x\n", active_irqs);
+		return IRQ_NONE;
+	}
+
+	drm_handle_vblank(vop->drm_dev, vop->pipe);
+
+	return (vop->vsync_work_pending) ? IRQ_WAKE_THREAD : IRQ_HANDLED;
+}
+
+static int vop_create_crtc(struct vop *vop)
+{
+	const struct vop_data *vop_data = vop->data;
+	struct device *dev = vop->dev;
+	struct drm_device *drm_dev = vop->drm_dev;
+	struct drm_plane *primary = NULL, *cursor = NULL, *plane;
+	struct drm_crtc *crtc = &vop->crtc;
+	struct device_node *port;
+	int ret;
+	int i;
+
+	/*
+	 * Create drm_plane for primary and cursor planes first, since we need
+	 * to pass them to drm_crtc_init_with_planes, which sets the
+	 * "possible_crtcs" to the newly initialized crtc.
+	 */
+	for (i = 0; i < vop_data->win_size; i++) {
+		struct vop_win *vop_win = &vop->win[i];
+		const struct vop_win_data *win_data = vop_win->data;
+
+		if (win_data->type != DRM_PLANE_TYPE_PRIMARY &&
+		    win_data->type != DRM_PLANE_TYPE_CURSOR)
+			continue;
+
+		ret = drm_universal_plane_init(vop->drm_dev, &vop_win->base,
+					       0, &vop_plane_funcs,
+					       win_data->phy->data_formats,
+					       win_data->phy->nformats,
+					       win_data->type);
+		if (ret) {
+			DRM_ERROR("failed to initialize plane\n");
+			goto err_cleanup_planes;
+		}
+
+		plane = &vop_win->base;
+		if (plane->type == DRM_PLANE_TYPE_PRIMARY)
+			primary = plane;
+		else if (plane->type == DRM_PLANE_TYPE_CURSOR)
+			cursor = plane;
+	}
+
+	ret = drm_crtc_init_with_planes(drm_dev, crtc, primary, cursor,
+					&vop_crtc_funcs);
+	if (ret)
+		return ret;
+
+	drm_crtc_helper_add(crtc, &vop_crtc_helper_funcs);
+
+	/*
+	 * Create drm_planes for overlay windows with possible_crtcs restricted
+	 * to the newly created crtc.
+	 */
+	for (i = 0; i < vop_data->win_size; i++) {
+		struct vop_win *vop_win = &vop->win[i];
+		const struct vop_win_data *win_data = vop_win->data;
+		unsigned long possible_crtcs = 1 << drm_crtc_index(crtc);
+
+		if (win_data->type != DRM_PLANE_TYPE_OVERLAY)
+			continue;
+
+		ret = drm_universal_plane_init(vop->drm_dev, &vop_win->base,
+					       possible_crtcs,
+					       &vop_plane_funcs,
+					       win_data->phy->data_formats,
+					       win_data->phy->nformats,
+					       win_data->type);
+		if (ret) {
+			DRM_ERROR("failed to initialize overlay plane\n");
+			goto err_cleanup_crtc;
+		}
+	}
+
+	port = of_get_child_by_name(dev->of_node, "port");
+	if (!port) {
+		DRM_ERROR("no port node found in %s\n",
+			  dev->of_node->full_name);
+		goto err_cleanup_crtc;
+	}
+
+	crtc->port = port;
+	vop->pipe = drm_crtc_index(crtc);
+	rockchip_register_crtc_funcs(drm_dev, &private_crtc_funcs, vop->pipe);
+
+	return 0;
+
+err_cleanup_crtc:
+	drm_crtc_cleanup(crtc);
+err_cleanup_planes:
+	list_for_each_entry(plane, &drm_dev->mode_config.plane_list, head)
+		drm_plane_cleanup(plane);
+	return ret;
+}
+
+static void vop_destroy_crtc(struct vop *vop)
+{
+	struct drm_crtc *crtc = &vop->crtc;
+
+	rockchip_unregister_crtc_funcs(vop->drm_dev, vop->pipe);
+	of_node_put(crtc->port);
+	drm_crtc_cleanup(crtc);
+}
+
+static int vop_initial(struct vop *vop)
+{
+	const struct vop_data *vop_data = vop->data;
+	const struct vop_reg_data *init_table = vop_data->init_table;
+	struct reset_control *ahb_rst;
+	int i, ret;
+
+	vop->hclk = devm_clk_get(vop->dev, "hclk_vop");
+	if (IS_ERR(vop->hclk)) {
+		dev_err(vop->dev, "failed to get hclk source\n");
+		return PTR_ERR(vop->hclk);
+	}
+	vop->aclk = devm_clk_get(vop->dev, "aclk_vop");
+	if (IS_ERR(vop->aclk)) {
+		dev_err(vop->dev, "failed to get aclk source\n");
+		return PTR_ERR(vop->aclk);
+	}
+	vop->dclk = devm_clk_get(vop->dev, "dclk_vop");
+	if (IS_ERR(vop->dclk)) {
+		dev_err(vop->dev, "failed to get dclk source\n");
+		return PTR_ERR(vop->dclk);
+	}
+
+	ret = clk_prepare(vop->hclk);
+	if (ret < 0) {
+		dev_err(vop->dev, "failed to prepare hclk\n");
+		return ret;
+	}
+
+	ret = clk_prepare(vop->dclk);
+	if (ret < 0) {
+		dev_err(vop->dev, "failed to prepare dclk\n");
+		goto err_unprepare_hclk;
+	}
+
+	ret = clk_prepare(vop->aclk);
+	if (ret < 0) {
+		dev_err(vop->dev, "failed to prepare aclk\n");
+		goto err_unprepare_dclk;
+	}
+
+	/*
+	 * enable hclk, so that we can config vop register.
+	 */
+	ret = clk_enable(vop->hclk);
+	if (ret < 0) {
+		dev_err(vop->dev, "failed to prepare aclk\n");
+		goto err_unprepare_aclk;
+	}
+	/*
+	 * do hclk_reset, reset all vop registers.
+	 */
+	ahb_rst = devm_reset_control_get(vop->dev, "ahb");
+	if (IS_ERR(ahb_rst)) {
+		dev_err(vop->dev, "failed to get ahb reset\n");
+		ret = PTR_ERR(ahb_rst);
+		goto err_disable_hclk;
+	}
+	reset_control_assert(ahb_rst);
+	usleep_range(10, 20);
+	reset_control_deassert(ahb_rst);
+
+	memcpy(vop->regsbak, vop->regs, vop->len);
+
+	for (i = 0; i < vop_data->table_size; i++)
+		vop_writel(vop, init_table[i].offset, init_table[i].value);
+
+	for (i = 0; i < vop_data->win_size; i++) {
+		const struct vop_win_data *win = &vop_data->win[i];
+
+		VOP_WIN_SET(vop, win, enable, 0);
+	}
+
+	vop_cfg_done(vop);
+
+	/*
+	 * do dclk_reset, let all config take affect.
+	 */
+	vop->dclk_rst = devm_reset_control_get(vop->dev, "dclk");
+	if (IS_ERR(vop->dclk_rst)) {
+		dev_err(vop->dev, "failed to get dclk reset\n");
+		ret = PTR_ERR(vop->dclk_rst);
+		goto err_unprepare_aclk;
+	}
+	reset_control_assert(vop->dclk_rst);
+	usleep_range(10, 20);
+	reset_control_deassert(vop->dclk_rst);
+
+	clk_disable(vop->hclk);
+
+	vop->dpms = DRM_MODE_DPMS_OFF;
+
+	return 0;
+
+err_disable_hclk:
+	clk_disable(vop->hclk);
+err_unprepare_aclk:
+	clk_unprepare(vop->aclk);
+err_unprepare_dclk:
+	clk_unprepare(vop->dclk);
+err_unprepare_hclk:
+	clk_unprepare(vop->hclk);
+	return ret;
+}
+
+/*
+ * Initialize the vop->win array elements.
+ */
+static void vop_win_init(struct vop *vop)
+{
+	const struct vop_data *vop_data = vop->data;
+	unsigned int i;
+
+	for (i = 0; i < vop_data->win_size; i++) {
+		struct vop_win *vop_win = &vop->win[i];
+		const struct vop_win_data *win_data = &vop_data->win[i];
+
+		vop_win->data = win_data;
+		vop_win->vop = vop;
+		INIT_LIST_HEAD(&vop_win->pending);
+	}
+}
+
+static int vop_bind(struct device *dev, struct device *master, void *data)
+{
+	struct platform_device *pdev = to_platform_device(dev);
+	const struct of_device_id *of_id;
+	const struct vop_data *vop_data;
+	struct drm_device *drm_dev = data;
+	struct vop *vop;
+	struct resource *res;
+	size_t alloc_size;
+	int ret;
+
+	of_id = of_match_device(vop_driver_dt_match, dev);
+	vop_data = of_id->data;
+	if (!vop_data)
+		return -ENODEV;
+
+	/* Allocate vop struct and its vop_win array */
+	alloc_size = sizeof(*vop) + sizeof(*vop->win) * vop_data->win_size;
+	vop = devm_kzalloc(dev, alloc_size, GFP_KERNEL);
+	if (!vop)
+		return -ENOMEM;
+
+	vop->dev = dev;
+	vop->data = vop_data;
+	vop->drm_dev = drm_dev;
+	dev_set_drvdata(dev, vop);
+
+	vop_win_init(vop);
+
+	res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
+	vop->len = resource_size(res);
+	vop->regs = devm_ioremap_resource(dev, res);
+	if (IS_ERR(vop->regs))
+		return PTR_ERR(vop->regs);
+
+	vop->regsbak = devm_kzalloc(dev, vop->len, GFP_KERNEL);
+	if (!vop->regsbak)
+		return -ENOMEM;
+
+	ret = vop_initial(vop);
+	if (ret < 0) {
+		dev_err(&pdev->dev, "cannot initial vop dev - err %d\n", ret);
+		return ret;
+	}
+
+	vop->irq = platform_get_irq(pdev, 0);
+	if (vop->irq < 0) {
+		dev_err(dev, "cannot find irq for vop\n");
+		return vop->irq;
+	}
+
+	spin_lock_init(&vop->reg_lock);
+	spin_lock_init(&vop->irq_lock);
+
+	mutex_init(&vop->vsync_mutex);
+
+	ret = devm_request_threaded_irq(dev, vop->irq, vop_isr, vop_isr_thread,
+					IRQF_SHARED, dev_name(dev), vop);
+	if (ret)
+		return ret;
+
+	/* IRQ is initially disabled; it gets enabled in power_on */
+	disable_irq(vop->irq);
+
+	ret = vop_create_crtc(vop);
+	if (ret)
+		return ret;
+
+	pm_runtime_enable(&pdev->dev);
+	return 0;
+}
+
+static void vop_unbind(struct device *dev, struct device *master, void *data)
+{
+	struct vop *vop = dev_get_drvdata(dev);
+
+	pm_runtime_disable(dev);
+	vop_destroy_crtc(vop);
+}
+
+static const struct component_ops vop_component_ops = {
+	.bind = vop_bind,
+	.unbind = vop_unbind,
+};
+
+static int vop_probe(struct platform_device *pdev)
+{
+	struct device *dev = &pdev->dev;
+
+	if (!dev->of_node) {
+		dev_err(dev, "can't find vop devices\n");
+		return -ENODEV;
+	}
+
+	return component_add(dev, &vop_component_ops);
+}
+
+static int vop_remove(struct platform_device *pdev)
+{
+	component_del(&pdev->dev, &vop_component_ops);
+
+	return 0;
+}
+
+struct platform_driver vop_platform_driver = {
+	.probe = vop_probe,
+	.remove = vop_remove,
+	.driver = {
+		.name = "rockchip-vop",
+		.owner = THIS_MODULE,
+		.of_match_table = of_match_ptr(vop_driver_dt_match),
+	},
+};
+
+module_platform_driver(vop_platform_driver);
+
+MODULE_AUTHOR("Mark Yao <mark.yao@rock-chips.com>");
+MODULE_DESCRIPTION("ROCKCHIP VOP Driver");
+MODULE_LICENSE("GPL v2");
diff --git a/drivers/gpu/drm/rockchip/rockchip_drm_vop.h b/drivers/gpu/drm/rockchip/rockchip_drm_vop.h
new file mode 100644
index 000000000000..63e9b3a084c5
--- /dev/null
+++ b/drivers/gpu/drm/rockchip/rockchip_drm_vop.h
@@ -0,0 +1,201 @@
+/*
+ * Copyright (C) Fuzhou Rockchip Electronics Co.Ltd
+ * Author:Mark Yao <mark.yao@rock-chips.com>
+ *
+ * This software is licensed under the terms of the GNU General Public
+ * License version 2, as published by the Free Software Foundation, and
+ * may be copied, distributed, and modified under those terms.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ */
+
+#ifndef _ROCKCHIP_DRM_VOP_H
+#define _ROCKCHIP_DRM_VOP_H
+
+/* register definition */
+#define REG_CFG_DONE			0x0000
+#define VERSION_INFO			0x0004
+#define SYS_CTRL			0x0008
+#define SYS_CTRL1			0x000c
+#define DSP_CTRL0			0x0010
+#define DSP_CTRL1			0x0014
+#define DSP_BG				0x0018
+#define MCU_CTRL			0x001c
+#define INTR_CTRL0			0x0020
+#define INTR_CTRL1			0x0024
+#define WIN0_CTRL0			0x0030
+#define WIN0_CTRL1			0x0034
+#define WIN0_COLOR_KEY			0x0038
+#define WIN0_VIR			0x003c
+#define WIN0_YRGB_MST			0x0040
+#define WIN0_CBR_MST			0x0044
+#define WIN0_ACT_INFO			0x0048
+#define WIN0_DSP_INFO			0x004c
+#define WIN0_DSP_ST			0x0050
+#define WIN0_SCL_FACTOR_YRGB		0x0054
+#define WIN0_SCL_FACTOR_CBR		0x0058
+#define WIN0_SCL_OFFSET			0x005c
+#define WIN0_SRC_ALPHA_CTRL		0x0060
+#define WIN0_DST_ALPHA_CTRL		0x0064
+#define WIN0_FADING_CTRL		0x0068
+/* win1 register */
+#define WIN1_CTRL0			0x0070
+#define WIN1_CTRL1			0x0074
+#define WIN1_COLOR_KEY			0x0078
+#define WIN1_VIR			0x007c
+#define WIN1_YRGB_MST			0x0080
+#define WIN1_CBR_MST			0x0084
+#define WIN1_ACT_INFO			0x0088
+#define WIN1_DSP_INFO			0x008c
+#define WIN1_DSP_ST			0x0090
+#define WIN1_SCL_FACTOR_YRGB		0x0094
+#define WIN1_SCL_FACTOR_CBR		0x0098
+#define WIN1_SCL_OFFSET			0x009c
+#define WIN1_SRC_ALPHA_CTRL		0x00a0
+#define WIN1_DST_ALPHA_CTRL		0x00a4
+#define WIN1_FADING_CTRL		0x00a8
+/* win2 register */
+#define WIN2_CTRL0			0x00b0
+#define WIN2_CTRL1			0x00b4
+#define WIN2_VIR0_1			0x00b8
+#define WIN2_VIR2_3			0x00bc
+#define WIN2_MST0			0x00c0
+#define WIN2_DSP_INFO0			0x00c4
+#define WIN2_DSP_ST0			0x00c8
+#define WIN2_COLOR_KEY			0x00cc
+#define WIN2_MST1			0x00d0
+#define WIN2_DSP_INFO1			0x00d4
+#define WIN2_DSP_ST1			0x00d8
+#define WIN2_SRC_ALPHA_CTRL		0x00dc
+#define WIN2_MST2			0x00e0
+#define WIN2_DSP_INFO2			0x00e4
+#define WIN2_DSP_ST2			0x00e8
+#define WIN2_DST_ALPHA_CTRL		0x00ec
+#define WIN2_MST3			0x00f0
+#define WIN2_DSP_INFO3			0x00f4
+#define WIN2_DSP_ST3			0x00f8
+#define WIN2_FADING_CTRL		0x00fc
+/* win3 register */
+#define WIN3_CTRL0			0x0100
+#define WIN3_CTRL1			0x0104
+#define WIN3_VIR0_1			0x0108
+#define WIN3_VIR2_3			0x010c
+#define WIN3_MST0			0x0110
+#define WIN3_DSP_INFO0			0x0114
+#define WIN3_DSP_ST0			0x0118
+#define WIN3_COLOR_KEY			0x011c
+#define WIN3_MST1			0x0120
+#define WIN3_DSP_INFO1			0x0124
+#define WIN3_DSP_ST1			0x0128
+#define WIN3_SRC_ALPHA_CTRL		0x012c
+#define WIN3_MST2			0x0130
+#define WIN3_DSP_INFO2			0x0134
+#define WIN3_DSP_ST2			0x0138
+#define WIN3_DST_ALPHA_CTRL		0x013c
+#define WIN3_MST3			0x0140
+#define WIN3_DSP_INFO3			0x0144
+#define WIN3_DSP_ST3			0x0148
+#define WIN3_FADING_CTRL		0x014c
+/* hwc register */
+#define HWC_CTRL0			0x0150
+#define HWC_CTRL1			0x0154
+#define HWC_MST				0x0158
+#define HWC_DSP_ST			0x015c
+#define HWC_SRC_ALPHA_CTRL		0x0160
+#define HWC_DST_ALPHA_CTRL		0x0164
+#define HWC_FADING_CTRL			0x0168
+/* post process register */
+#define POST_DSP_HACT_INFO		0x0170
+#define POST_DSP_VACT_INFO		0x0174
+#define POST_SCL_FACTOR_YRGB		0x0178
+#define POST_SCL_CTRL			0x0180
+#define POST_DSP_VACT_INFO_F1		0x0184
+#define DSP_HTOTAL_HS_END		0x0188
+#define DSP_HACT_ST_END			0x018c
+#define DSP_VTOTAL_VS_END		0x0190
+#define DSP_VACT_ST_END			0x0194
+#define DSP_VS_ST_END_F1		0x0198
+#define DSP_VACT_ST_END_F1		0x019c
+/* register definition end */
+
+/* interrupt define */
+#define DSP_HOLD_VALID_INTR		(1 << 0)
+#define FS_INTR				(1 << 1)
+#define LINE_FLAG_INTR			(1 << 2)
+#define BUS_ERROR_INTR			(1 << 3)
+
+#define INTR_MASK			(DSP_HOLD_VALID_INTR | FS_INTR | \
+					 LINE_FLAG_INTR | BUS_ERROR_INTR)
+
+#define DSP_HOLD_VALID_INTR_EN(x)	((x) << 4)
+#define FS_INTR_EN(x)			((x) << 5)
+#define LINE_FLAG_INTR_EN(x)		((x) << 6)
+#define BUS_ERROR_INTR_EN(x)		((x) << 7)
+#define DSP_HOLD_VALID_INTR_MASK	(1 << 4)
+#define FS_INTR_MASK			(1 << 5)
+#define LINE_FLAG_INTR_MASK		(1 << 6)
+#define BUS_ERROR_INTR_MASK		(1 << 7)
+
+#define INTR_CLR_SHIFT			8
+#define DSP_HOLD_VALID_INTR_CLR		(1 << (INTR_CLR_SHIFT + 0))
+#define FS_INTR_CLR			(1 << (INTR_CLR_SHIFT + 1))
+#define LINE_FLAG_INTR_CLR		(1 << (INTR_CLR_SHIFT + 2))
+#define BUS_ERROR_INTR_CLR		(1 << (INTR_CLR_SHIFT + 3))
+
+#define DSP_LINE_NUM(x)			(((x) & 0x1fff) << 12)
+#define DSP_LINE_NUM_MASK		(0x1fff << 12)
+
+/* src alpha ctrl define */
+#define SRC_FADING_VALUE(x)		(((x) & 0xff) << 24)
+#define SRC_GLOBAL_ALPHA(x)		(((x) & 0xff) << 16)
+#define SRC_FACTOR_M0(x)		(((x) & 0x7) << 6)
+#define SRC_ALPHA_CAL_M0(x)		(((x) & 0x1) << 5)
+#define SRC_BLEND_M0(x)			(((x) & 0x3) << 3)
+#define SRC_ALPHA_M0(x)			(((x) & 0x1) << 2)
+#define SRC_COLOR_M0(x)			(((x) & 0x1) << 1)
+#define SRC_ALPHA_EN(x)			(((x) & 0x1) << 0)
+/* dst alpha ctrl define */
+#define DST_FACTOR_M0(x)		(((x) & 0x7) << 6)
+
+/*
+ * display output interface supported by rockchip lcdc
+ */
+#define ROCKCHIP_OUT_MODE_P888	0
+#define ROCKCHIP_OUT_MODE_P666	1
+#define ROCKCHIP_OUT_MODE_P565	2
+/* for use special outface */
+#define ROCKCHIP_OUT_MODE_AAAA	15
+
+enum alpha_mode {
+	ALPHA_STRAIGHT,
+	ALPHA_INVERSE,
+};
+
+enum global_blend_mode {
+	ALPHA_GLOBAL,
+	ALPHA_PER_PIX,
+	ALPHA_PER_PIX_GLOBAL,
+};
+
+enum alpha_cal_mode {
+	ALPHA_SATURATION,
+	ALPHA_NO_SATURATION,
+};
+
+enum color_mode {
+	ALPHA_SRC_PRE_MUL,
+	ALPHA_SRC_NO_PRE_MUL,
+};
+
+enum factor_mode {
+	ALPHA_ZERO,
+	ALPHA_ONE,
+	ALPHA_SRC,
+	ALPHA_SRC_INVERSE,
+	ALPHA_SRC_GLOBAL,
+};
+
+#endif /* _ROCKCHIP_DRM_VOP_H */
diff --git a/drivers/gpu/drm/savage/savage_bci.c b/drivers/gpu/drm/savage/savage_bci.c
index c97cdc9ab239..d47dff95fe52 100644
--- a/drivers/gpu/drm/savage/savage_bci.c
+++ b/drivers/gpu/drm/savage/savage_bci.c
@@ -556,7 +556,7 @@ int savage_driver_load(struct drm_device *dev, unsigned long chipset)
 /*
  * Initialize mappings. On Savage4 and SavageIX the alignment
  * and size of the aperture is not suitable for automatic MTRR setup
- * in drm_addmap. Therefore we add them manually before the maps are
+ * in drm_legacy_addmap. Therefore we add them manually before the maps are
  * initialized, and tear them down on last close.
  */
 int savage_driver_firstopen(struct drm_device *dev)
@@ -624,19 +624,20 @@ int savage_driver_firstopen(struct drm_device *dev)
 		/* Automatic MTRR setup will do the right thing. */
 	}
 
-	ret = drm_addmap(dev, mmio_base, SAVAGE_MMIO_SIZE, _DRM_REGISTERS,
-			 _DRM_READ_ONLY, &dev_priv->mmio);
+	ret = drm_legacy_addmap(dev, mmio_base, SAVAGE_MMIO_SIZE,
+				_DRM_REGISTERS, _DRM_READ_ONLY,
+				&dev_priv->mmio);
 	if (ret)
 		return ret;
 
-	ret = drm_addmap(dev, fb_base, fb_size, _DRM_FRAME_BUFFER,
-			 _DRM_WRITE_COMBINING, &dev_priv->fb);
+	ret = drm_legacy_addmap(dev, fb_base, fb_size, _DRM_FRAME_BUFFER,
+				_DRM_WRITE_COMBINING, &dev_priv->fb);
 	if (ret)
 		return ret;
 
-	ret = drm_addmap(dev, aperture_base, SAVAGE_APERTURE_SIZE,
-			 _DRM_FRAME_BUFFER, _DRM_WRITE_COMBINING,
-			 &dev_priv->aperture);
+	ret = drm_legacy_addmap(dev, aperture_base, SAVAGE_APERTURE_SIZE,
+				_DRM_FRAME_BUFFER, _DRM_WRITE_COMBINING,
+				&dev_priv->aperture);
 	return ret;
 }
 
@@ -698,14 +699,14 @@ static int savage_do_init_bci(struct drm_device * dev, drm_savage_init_t * init)
 	dev_priv->texture_offset = init->texture_offset;
 	dev_priv->texture_size = init->texture_size;
 
-	dev_priv->sarea = drm_getsarea(dev);
+	dev_priv->sarea = drm_legacy_getsarea(dev);
 	if (!dev_priv->sarea) {
 		DRM_ERROR("could not find sarea!\n");
 		savage_do_cleanup_bci(dev);
 		return -EINVAL;
 	}
 	if (init->status_offset != 0) {
-		dev_priv->status = drm_core_findmap(dev, init->status_offset);
+		dev_priv->status = drm_legacy_findmap(dev, init->status_offset);
 		if (!dev_priv->status) {
 			DRM_ERROR("could not find shadow status region!\n");
 			savage_do_cleanup_bci(dev);
@@ -716,14 +717,14 @@ static int savage_do_init_bci(struct drm_device * dev, drm_savage_init_t * init)
 	}
 	if (dev_priv->dma_type == SAVAGE_DMA_AGP && init->buffers_offset) {
 		dev->agp_buffer_token = init->buffers_offset;
-		dev->agp_buffer_map = drm_core_findmap(dev,
+		dev->agp_buffer_map = drm_legacy_findmap(dev,
 						       init->buffers_offset);
 		if (!dev->agp_buffer_map) {
 			DRM_ERROR("could not find DMA buffer region!\n");
 			savage_do_cleanup_bci(dev);
 			return -EINVAL;
 		}
-		drm_core_ioremap(dev->agp_buffer_map, dev);
+		drm_legacy_ioremap(dev->agp_buffer_map, dev);
 		if (!dev->agp_buffer_map->handle) {
 			DRM_ERROR("failed to ioremap DMA buffer region!\n");
 			savage_do_cleanup_bci(dev);
@@ -732,7 +733,7 @@ static int savage_do_init_bci(struct drm_device * dev, drm_savage_init_t * init)
 	}
 	if (init->agp_textures_offset) {
 		dev_priv->agp_textures =
-		    drm_core_findmap(dev, init->agp_textures_offset);
+		    drm_legacy_findmap(dev, init->agp_textures_offset);
 		if (!dev_priv->agp_textures) {
 			DRM_ERROR("could not find agp texture region!\n");
 			savage_do_cleanup_bci(dev);
@@ -755,7 +756,7 @@ static int savage_do_init_bci(struct drm_device * dev, drm_savage_init_t * init)
 			savage_do_cleanup_bci(dev);
 			return -EINVAL;
 		}
-		dev_priv->cmd_dma = drm_core_findmap(dev, init->cmd_dma_offset);
+		dev_priv->cmd_dma = drm_legacy_findmap(dev, init->cmd_dma_offset);
 		if (!dev_priv->cmd_dma) {
 			DRM_ERROR("could not find command DMA region!\n");
 			savage_do_cleanup_bci(dev);
@@ -768,7 +769,7 @@ static int savage_do_init_bci(struct drm_device * dev, drm_savage_init_t * init)
 				savage_do_cleanup_bci(dev);
 				return -EINVAL;
 			}
-			drm_core_ioremap(dev_priv->cmd_dma, dev);
+			drm_legacy_ioremap(dev_priv->cmd_dma, dev);
 			if (!dev_priv->cmd_dma->handle) {
 				DRM_ERROR("failed to ioremap command "
 					  "DMA region!\n");
@@ -894,11 +895,11 @@ static int savage_do_cleanup_bci(struct drm_device * dev)
 	} else if (dev_priv->cmd_dma && dev_priv->cmd_dma->handle &&
 		   dev_priv->cmd_dma->type == _DRM_AGP &&
 		   dev_priv->dma_type == SAVAGE_DMA_AGP)
-		drm_core_ioremapfree(dev_priv->cmd_dma, dev);
+		drm_legacy_ioremapfree(dev_priv->cmd_dma, dev);
 
 	if (dev_priv->dma_type == SAVAGE_DMA_AGP &&
 	    dev->agp_buffer_map && dev->agp_buffer_map->handle) {
-		drm_core_ioremapfree(dev->agp_buffer_map, dev);
+		drm_legacy_ioremapfree(dev->agp_buffer_map, dev);
 		/* make sure the next instance (which may be running
 		 * in PCI mode) doesn't try to use an old
 		 * agp_buffer_map. */
@@ -1050,7 +1051,7 @@ void savage_reclaim_buffers(struct drm_device *dev, struct drm_file *file_priv)
 		return;
 
 	if (file_priv->master && file_priv->master->lock.hw_lock) {
-		drm_idlelock_take(&file_priv->master->lock);
+		drm_legacy_idlelock_take(&file_priv->master->lock);
 		release_idlelock = 1;
 	}
 
@@ -1069,7 +1070,7 @@ void savage_reclaim_buffers(struct drm_device *dev, struct drm_file *file_priv)
 	}
 
 	if (release_idlelock)
-		drm_idlelock_release(&file_priv->master->lock);
+		drm_legacy_idlelock_release(&file_priv->master->lock);
 }
 
 const struct drm_ioctl_desc savage_ioctls[] = {
diff --git a/drivers/gpu/drm/savage/savage_drv.c b/drivers/gpu/drm/savage/savage_drv.c
index 3c030216e888..21aed1febeb4 100644
--- a/drivers/gpu/drm/savage/savage_drv.c
+++ b/drivers/gpu/drm/savage/savage_drv.c
@@ -40,7 +40,7 @@ static const struct file_operations savage_driver_fops = {
 	.open = drm_open,
 	.release = drm_release,
 	.unlocked_ioctl = drm_ioctl,
-	.mmap = drm_mmap,
+	.mmap = drm_legacy_mmap,
 	.poll = drm_poll,
 #ifdef CONFIG_COMPAT
 	.compat_ioctl = drm_compat_ioctl,
@@ -57,6 +57,7 @@ static struct drm_driver driver = {
 	.preclose = savage_reclaim_buffers,
 	.lastclose = savage_driver_lastclose,
 	.unload = savage_driver_unload,
+	.set_busid = drm_pci_set_busid,
 	.ioctls = savage_ioctls,
 	.dma_ioctl = savage_bci_buffers,
 	.fops = &savage_driver_fops,
diff --git a/drivers/gpu/drm/savage/savage_drv.h b/drivers/gpu/drm/savage/savage_drv.h
index 335f8fcf1041..37b699571ad0 100644
--- a/drivers/gpu/drm/savage/savage_drv.h
+++ b/drivers/gpu/drm/savage/savage_drv.h
@@ -26,6 +26,8 @@
 #ifndef __SAVAGE_DRV_H__
 #define __SAVAGE_DRV_H__
 
+#include <drm/drm_legacy.h>
+
 #define DRIVER_AUTHOR	"Felix Kuehling"
 
 #define DRIVER_NAME	"savage"
diff --git a/drivers/gpu/drm/shmobile/shmob_drm_backlight.c b/drivers/gpu/drm/shmobile/shmob_drm_backlight.c
index 463aee18f774..33dd41afea0e 100644
--- a/drivers/gpu/drm/shmobile/shmob_drm_backlight.c
+++ b/drivers/gpu/drm/shmobile/shmob_drm_backlight.c
@@ -1,7 +1,7 @@
 /*
  * shmob_drm_backlight.c  --  SH Mobile DRM Backlight
  *
- * Copyright (C) 2012 Renesas Corporation
+ * Copyright (C) 2012 Renesas Electronics Corporation
  *
  * Laurent Pinchart (laurent.pinchart@ideasonboard.com)
  *
diff --git a/drivers/gpu/drm/shmobile/shmob_drm_backlight.h b/drivers/gpu/drm/shmobile/shmob_drm_backlight.h
index 9477595d2ff3..bac719ecc301 100644
--- a/drivers/gpu/drm/shmobile/shmob_drm_backlight.h
+++ b/drivers/gpu/drm/shmobile/shmob_drm_backlight.h
@@ -1,7 +1,7 @@
 /*
  * shmob_drm_backlight.h  --  SH Mobile DRM Backlight
  *
- * Copyright (C) 2012 Renesas Corporation
+ * Copyright (C) 2012 Renesas Electronics Corporation
  *
  * Laurent Pinchart (laurent.pinchart@ideasonboard.com)
  *
diff --git a/drivers/gpu/drm/shmobile/shmob_drm_crtc.c b/drivers/gpu/drm/shmobile/shmob_drm_crtc.c
index 47875de89010..859ccb658601 100644
--- a/drivers/gpu/drm/shmobile/shmob_drm_crtc.c
+++ b/drivers/gpu/drm/shmobile/shmob_drm_crtc.c
@@ -1,7 +1,7 @@
 /*
  * shmob_drm_crtc.c  --  SH Mobile DRM CRTCs
  *
- * Copyright (C) 2012 Renesas Corporation
+ * Copyright (C) 2012 Renesas Electronics Corporation
  *
  * Laurent Pinchart (laurent.pinchart@ideasonboard.com)
  *
@@ -19,6 +19,7 @@
 #include <drm/drm_crtc_helper.h>
 #include <drm/drm_fb_cma_helper.h>
 #include <drm/drm_gem_cma_helper.h>
+#include <drm/drm_plane_helper.h>
 
 #include <video/sh_mobile_meram.h>
 
diff --git a/drivers/gpu/drm/shmobile/shmob_drm_crtc.h b/drivers/gpu/drm/shmobile/shmob_drm_crtc.h
index e5bd109c4c38..eddad6dcc88a 100644
--- a/drivers/gpu/drm/shmobile/shmob_drm_crtc.h
+++ b/drivers/gpu/drm/shmobile/shmob_drm_crtc.h
@@ -1,7 +1,7 @@
 /*
  * shmob_drm_crtc.h  --  SH Mobile DRM CRTCs
  *
- * Copyright (C) 2012 Renesas Corporation
+ * Copyright (C) 2012 Renesas Electronics Corporation
  *
  * Laurent Pinchart (laurent.pinchart@ideasonboard.com)
  *
diff --git a/drivers/gpu/drm/shmobile/shmob_drm_drv.c b/drivers/gpu/drm/shmobile/shmob_drm_drv.c
index ff4ba483b602..666321de7b99 100644
--- a/drivers/gpu/drm/shmobile/shmob_drm_drv.c
+++ b/drivers/gpu/drm/shmobile/shmob_drm_drv.c
@@ -1,7 +1,7 @@
 /*
  * shmob_drm_drv.c  --  SH Mobile DRM driver
  *
- * Copyright (C) 2012 Renesas Corporation
+ * Copyright (C) 2012 Renesas Electronics Corporation
  *
  * Laurent Pinchart (laurent.pinchart@ideasonboard.com)
  *
@@ -267,6 +267,7 @@ static struct drm_driver shmob_drm_driver = {
 	.load			= shmob_drm_load,
 	.unload			= shmob_drm_unload,
 	.preclose		= shmob_drm_preclose,
+	.set_busid		= drm_platform_set_busid,
 	.irq_handler		= shmob_drm_irq,
 	.get_vblank_counter	= drm_vblank_count,
 	.enable_vblank		= shmob_drm_enable_vblank,
@@ -347,7 +348,6 @@ static struct platform_driver shmob_drm_platform_driver = {
 	.probe		= shmob_drm_probe,
 	.remove		= shmob_drm_remove,
 	.driver		= {
-		.owner	= THIS_MODULE,
 		.name	= "shmob-drm",
 		.pm	= &shmob_drm_pm_ops,
 	},
diff --git a/drivers/gpu/drm/shmobile/shmob_drm_drv.h b/drivers/gpu/drm/shmobile/shmob_drm_drv.h
index 4d46b811b5a7..02ea315ba69a 100644
--- a/drivers/gpu/drm/shmobile/shmob_drm_drv.h
+++ b/drivers/gpu/drm/shmobile/shmob_drm_drv.h
@@ -1,7 +1,7 @@
 /*
  * shmob_drm.h  --  SH Mobile DRM driver
  *
- * Copyright (C) 2012 Renesas Corporation
+ * Copyright (C) 2012 Renesas Electronics Corporation
  *
  * Laurent Pinchart (laurent.pinchart@ideasonboard.com)
  *
diff --git a/drivers/gpu/drm/shmobile/shmob_drm_kms.c b/drivers/gpu/drm/shmobile/shmob_drm_kms.c
index fc0ef0ca7d04..aaf98ace4a90 100644
--- a/drivers/gpu/drm/shmobile/shmob_drm_kms.c
+++ b/drivers/gpu/drm/shmobile/shmob_drm_kms.c
@@ -1,7 +1,7 @@
 /*
  * shmob_drm_kms.c  --  SH Mobile DRM Mode Setting
  *
- * Copyright (C) 2012 Renesas Corporation
+ * Copyright (C) 2012 Renesas Electronics Corporation
  *
  * Laurent Pinchart (laurent.pinchart@ideasonboard.com)
  *
diff --git a/drivers/gpu/drm/shmobile/shmob_drm_kms.h b/drivers/gpu/drm/shmobile/shmob_drm_kms.h
index 9495c9111308..06d5b7caa026 100644
--- a/drivers/gpu/drm/shmobile/shmob_drm_kms.h
+++ b/drivers/gpu/drm/shmobile/shmob_drm_kms.h
@@ -1,7 +1,7 @@
 /*
  * shmob_drm_kms.h  --  SH Mobile DRM Mode Setting
  *
- * Copyright (C) 2012 Renesas Corporation
+ * Copyright (C) 2012 Renesas Electronics Corporation
  *
  * Laurent Pinchart (laurent.pinchart@ideasonboard.com)
  *
diff --git a/drivers/gpu/drm/shmobile/shmob_drm_plane.c b/drivers/gpu/drm/shmobile/shmob_drm_plane.c
index 060ae03e5f9b..1805bb23b113 100644
--- a/drivers/gpu/drm/shmobile/shmob_drm_plane.c
+++ b/drivers/gpu/drm/shmobile/shmob_drm_plane.c
@@ -1,7 +1,7 @@
 /*
  * shmob_drm_plane.c  --  SH Mobile DRM Planes
  *
- * Copyright (C) 2012 Renesas Corporation
+ * Copyright (C) 2012 Renesas Electronics Corporation
  *
  * Laurent Pinchart (laurent.pinchart@ideasonboard.com)
  *
diff --git a/drivers/gpu/drm/shmobile/shmob_drm_plane.h b/drivers/gpu/drm/shmobile/shmob_drm_plane.h
index 99623d05e3b0..a58cc1fc3240 100644
--- a/drivers/gpu/drm/shmobile/shmob_drm_plane.h
+++ b/drivers/gpu/drm/shmobile/shmob_drm_plane.h
@@ -1,7 +1,7 @@
 /*
  * shmob_drm_plane.h  --  SH Mobile DRM Planes
  *
- * Copyright (C) 2012 Renesas Corporation
+ * Copyright (C) 2012 Renesas Electronics Corporation
  *
  * Laurent Pinchart (laurent.pinchart@ideasonboard.com)
  *
diff --git a/drivers/gpu/drm/shmobile/shmob_drm_regs.h b/drivers/gpu/drm/shmobile/shmob_drm_regs.h
index 7923cdd6368e..ea17d4415b9e 100644
--- a/drivers/gpu/drm/shmobile/shmob_drm_regs.h
+++ b/drivers/gpu/drm/shmobile/shmob_drm_regs.h
@@ -1,7 +1,7 @@
 /*
  * shmob_drm_regs.h  --  SH Mobile DRM registers
  *
- * Copyright (C) 2012 Renesas Corporation
+ * Copyright (C) 2012 Renesas Electronics Corporation
  *
  * Laurent Pinchart (laurent.pinchart@ideasonboard.com)
  *
diff --git a/drivers/gpu/drm/sis/sis_drv.c b/drivers/gpu/drm/sis/sis_drv.c
index 756f787b7143..79bce76cb8f7 100644
--- a/drivers/gpu/drm/sis/sis_drv.c
+++ b/drivers/gpu/drm/sis/sis_drv.c
@@ -70,7 +70,7 @@ static const struct file_operations sis_driver_fops = {
 	.open = drm_open,
 	.release = drm_release,
 	.unlocked_ioctl = drm_ioctl,
-	.mmap = drm_mmap,
+	.mmap = drm_legacy_mmap,
 	.poll = drm_poll,
 #ifdef CONFIG_COMPAT
 	.compat_ioctl = drm_compat_ioctl,
@@ -108,6 +108,7 @@ static struct drm_driver driver = {
 	.open = sis_driver_open,
 	.preclose = sis_reclaim_buffers_locked,
 	.postclose = sis_driver_postclose,
+	.set_busid = drm_pci_set_busid,
 	.dma_quiescent = sis_idle,
 	.lastclose = sis_lastclose,
 	.ioctls = sis_ioctls,
diff --git a/drivers/gpu/drm/sis/sis_drv.h b/drivers/gpu/drm/sis/sis_drv.h
index c31c0253054d..16f972b2a76a 100644
--- a/drivers/gpu/drm/sis/sis_drv.h
+++ b/drivers/gpu/drm/sis/sis_drv.h
@@ -28,6 +28,8 @@
 #ifndef _SIS_DRV_H_
 #define _SIS_DRV_H_
 
+#include <drm/drm_legacy.h>
+
 /* General customization:
  */
 
diff --git a/drivers/gpu/drm/sis/sis_mm.c b/drivers/gpu/drm/sis/sis_mm.c
index 77f288e4a0a6..93ad8a5704d1 100644
--- a/drivers/gpu/drm/sis/sis_mm.c
+++ b/drivers/gpu/drm/sis/sis_mm.c
@@ -319,12 +319,12 @@ void sis_reclaim_buffers_locked(struct drm_device *dev,
 	if (!(file->minor->master && file->master->lock.hw_lock))
 		return;
 
-	drm_idlelock_take(&file->master->lock);
+	drm_legacy_idlelock_take(&file->master->lock);
 
 	mutex_lock(&dev->struct_mutex);
 	if (list_empty(&file_priv->obj_list)) {
 		mutex_unlock(&dev->struct_mutex);
-		drm_idlelock_release(&file->master->lock);
+		drm_legacy_idlelock_release(&file->master->lock);
 
 		return;
 	}
@@ -345,7 +345,7 @@ void sis_reclaim_buffers_locked(struct drm_device *dev,
 	}
 	mutex_unlock(&dev->struct_mutex);
 
-	drm_idlelock_release(&file->master->lock);
+	drm_legacy_idlelock_release(&file->master->lock);
 
 	return;
 }
diff --git a/drivers/gpu/drm/sti/Kconfig b/drivers/gpu/drm/sti/Kconfig
index ae8850f3e63b..d6d6b705b8c1 100644
--- a/drivers/gpu/drm/sti/Kconfig
+++ b/drivers/gpu/drm/sti/Kconfig
@@ -5,6 +5,7 @@ config DRM_STI
 	select DRM_KMS_HELPER
 	select DRM_GEM_CMA_HELPER
 	select DRM_KMS_CMA_HELPER
+	select FW_LOADER_USER_HELPER_FALLBACK
 	help
 	  Choose this option to enable DRM on STM stiH41x chipset
 
diff --git a/drivers/gpu/drm/sti/Makefile b/drivers/gpu/drm/sti/Makefile
index 04ac2ceef27f..6ba9d27c1b90 100644
--- a/drivers/gpu/drm/sti/Makefile
+++ b/drivers/gpu/drm/sti/Makefile
@@ -3,6 +3,7 @@ sticompositor-y := \
 	sti_mixer.o \
 	sti_gdp.o \
 	sti_vid.o \
+	sti_cursor.o \
 	sti_compositor.o \
 	sti_drm_crtc.o \
 	sti_drm_plane.o
@@ -18,4 +19,5 @@ obj-$(CONFIG_DRM_STI) = \
 	sti_hda.o \
 	sti_tvout.o \
 	sticompositor.o \
-	sti_drm_drv.o
\ No newline at end of file
+	sti_hqvdp.o \
+	sti_drm_drv.o
diff --git a/drivers/gpu/drm/sti/sti_compositor.c b/drivers/gpu/drm/sti/sti_compositor.c
index 390d93e9a06c..43215d3020fb 100644
--- a/drivers/gpu/drm/sti/sti_compositor.c
+++ b/drivers/gpu/drm/sti/sti_compositor.c
@@ -24,14 +24,16 @@
  * stiH407 compositor properties
  */
 struct sti_compositor_data stih407_compositor_data = {
-	.nb_subdev = 6,
+	.nb_subdev = 8,
 	.subdev_desc = {
+			{STI_CURSOR_SUBDEV, (int)STI_CURSOR, 0x000},
 			{STI_GPD_SUBDEV, (int)STI_GDP_0, 0x100},
 			{STI_GPD_SUBDEV, (int)STI_GDP_1, 0x200},
 			{STI_GPD_SUBDEV, (int)STI_GDP_2, 0x300},
 			{STI_GPD_SUBDEV, (int)STI_GDP_3, 0x400},
 			{STI_VID_SUBDEV, (int)STI_VID_0, 0x700},
-			{STI_MIXER_MAIN_SUBDEV, STI_MIXER_MAIN, 0xC00}
+			{STI_MIXER_MAIN_SUBDEV, STI_MIXER_MAIN, 0xC00},
+			{STI_MIXER_AUX_SUBDEV, STI_MIXER_AUX, 0xD00},
 	},
 };
 
@@ -67,11 +69,11 @@ static int sti_compositor_init_subdev(struct sti_compositor *compo,
 			break;
 		case STI_GPD_SUBDEV:
 		case STI_VID_SUBDEV:
+		case STI_CURSOR_SUBDEV:
 			compo->layer[layer_id++] =
 			    sti_layer_create(compo->dev, desc[i].id,
 					     compo->regs + desc[i].offset);
 			break;
-			/* case STI_CURSOR_SUBDEV : TODO */
 		default:
 			DRM_ERROR("Unknow subdev compoment type\n");
 			return 1;
@@ -102,33 +104,35 @@ static int sti_compositor_bind(struct device *dev, struct device *master,
 			enum sti_layer_type type = desc & STI_LAYER_TYPE_MASK;
 			enum drm_plane_type plane_type = DRM_PLANE_TYPE_OVERLAY;
 
-			if (compo->mixer[crtc])
+			if (crtc < compo->nb_mixers)
 				plane_type = DRM_PLANE_TYPE_PRIMARY;
 
 			switch (type) {
 			case STI_CUR:
 				cursor = sti_drm_plane_init(drm_dev,
 						compo->layer[i],
-						(1 << crtc) - 1,
-						DRM_PLANE_TYPE_CURSOR);
+						1, DRM_PLANE_TYPE_CURSOR);
 				break;
 			case STI_GDP:
 			case STI_VID:
 				primary = sti_drm_plane_init(drm_dev,
 						compo->layer[i],
-						(1 << crtc) - 1, plane_type);
+						(1 << compo->nb_mixers) - 1,
+						plane_type);
 				plane++;
 				break;
 			case STI_BCK:
+			case STI_VDP:
 				break;
 			}
 
 			/* The first planes are reserved for primary planes*/
-			if (compo->mixer[crtc]) {
+			if (crtc < compo->nb_mixers && primary) {
 				sti_drm_crtc_init(drm_dev, compo->mixer[crtc],
 						primary, cursor);
 				crtc++;
 				cursor = NULL;
+				primary = NULL;
 			}
 		}
 	}
@@ -267,7 +271,6 @@ static int sti_compositor_remove(struct platform_device *pdev)
 static struct platform_driver sti_compositor_driver = {
 	.driver = {
 		.name = "sti-compositor",
-		.owner = THIS_MODULE,
 		.of_match_table = compositor_of_match,
 	},
 	.probe = sti_compositor_probe,
diff --git a/drivers/gpu/drm/sti/sti_compositor.h b/drivers/gpu/drm/sti/sti_compositor.h
index 3ea19db72e0f..019eb44c62cc 100644
--- a/drivers/gpu/drm/sti/sti_compositor.h
+++ b/drivers/gpu/drm/sti/sti_compositor.h
@@ -64,7 +64,6 @@ struct sti_compositor_data {
  * @layer: array of layers
  * @nb_mixers: number of mixers for this compositor
  * @nb_layers: number of layers (GDP,VID,...) for this compositor
- * @enable: true if compositor is enable else false
  * @vtg_vblank_nb: callback for VTG VSYNC notification
  */
 struct sti_compositor {
@@ -83,7 +82,6 @@ struct sti_compositor {
 	struct sti_layer *layer[STI_MAX_LAYER];
 	int nb_mixers;
 	int nb_layers;
-	bool enable;
 	struct notifier_block vtg_vblank_nb;
 };
 
diff --git a/drivers/gpu/drm/sti/sti_cursor.c b/drivers/gpu/drm/sti/sti_cursor.c
new file mode 100644
index 000000000000..010eaee60bf7
--- /dev/null
+++ b/drivers/gpu/drm/sti/sti_cursor.c
@@ -0,0 +1,242 @@
+/*
+ * Copyright (C) STMicroelectronics SA 2014
+ * Authors: Vincent Abriou <vincent.abriou@st.com>
+ *          Fabien Dessenne <fabien.dessenne@st.com>
+ *          for STMicroelectronics.
+ * License terms:  GNU General Public License (GPL), version 2
+ */
+#include <drm/drmP.h>
+
+#include "sti_cursor.h"
+#include "sti_layer.h"
+#include "sti_vtg.h"
+
+/* Registers */
+#define CUR_CTL             0x00
+#define CUR_VPO             0x0C
+#define CUR_PML             0x14
+#define CUR_PMP             0x18
+#define CUR_SIZE            0x1C
+#define CUR_CML             0x20
+#define CUR_AWS             0x28
+#define CUR_AWE             0x2C
+
+#define CUR_CTL_CLUT_UPDATE BIT(1)
+
+#define STI_CURS_MIN_SIZE   1
+#define STI_CURS_MAX_SIZE   128
+
+/*
+ * pixmap dma buffer stucture
+ *
+ * @paddr:  physical address
+ * @size:   buffer size
+ * @base:   virtual address
+ */
+struct dma_pixmap {
+	dma_addr_t paddr;
+	size_t size;
+	void *base;
+};
+
+/**
+ * STI Cursor structure
+ *
+ * @layer:      layer structure
+ * @width:      cursor width
+ * @height:     cursor height
+ * @clut:       color look up table
+ * @clut_paddr: color look up table physical address
+ * @pixmap:     pixmap dma buffer (clut8-format cursor)
+ */
+struct sti_cursor {
+	struct sti_layer layer;
+	unsigned int width;
+	unsigned int height;
+	unsigned short *clut;
+	dma_addr_t clut_paddr;
+	struct dma_pixmap pixmap;
+};
+
+static const uint32_t cursor_supported_formats[] = {
+	DRM_FORMAT_ARGB8888,
+};
+
+#define to_sti_cursor(x) container_of(x, struct sti_cursor, layer)
+
+static const uint32_t *sti_cursor_get_formats(struct sti_layer *layer)
+{
+	return cursor_supported_formats;
+}
+
+static unsigned int sti_cursor_get_nb_formats(struct sti_layer *layer)
+{
+	return ARRAY_SIZE(cursor_supported_formats);
+}
+
+static void sti_cursor_argb8888_to_clut8(struct sti_layer *layer)
+{
+	struct sti_cursor *cursor = to_sti_cursor(layer);
+	u32 *src = layer->vaddr;
+	u8  *dst = cursor->pixmap.base;
+	unsigned int i, j;
+	u32 a, r, g, b;
+
+	for (i = 0; i < cursor->height; i++) {
+		for (j = 0; j < cursor->width; j++) {
+			/* Pick the 2 higher bits of each component */
+			a = (*src >> 30) & 3;
+			r = (*src >> 22) & 3;
+			g = (*src >> 14) & 3;
+			b = (*src >> 6) & 3;
+			*dst = a << 6 | r << 4 | g << 2 | b;
+			src++;
+			dst++;
+		}
+	}
+}
+
+static int sti_cursor_prepare_layer(struct sti_layer *layer, bool first_prepare)
+{
+	struct sti_cursor *cursor = to_sti_cursor(layer);
+	struct drm_display_mode *mode = layer->mode;
+	u32 y, x;
+	u32 val;
+
+	DRM_DEBUG_DRIVER("\n");
+
+	dev_dbg(layer->dev, "%s %s\n", __func__, sti_layer_to_str(layer));
+
+	if (layer->src_w < STI_CURS_MIN_SIZE ||
+	    layer->src_h < STI_CURS_MIN_SIZE ||
+	    layer->src_w > STI_CURS_MAX_SIZE ||
+	    layer->src_h > STI_CURS_MAX_SIZE) {
+		DRM_ERROR("Invalid cursor size (%dx%d)\n",
+				layer->src_w, layer->src_h);
+		return -EINVAL;
+	}
+
+	/* If the cursor size has changed, re-allocated the pixmap */
+	if (!cursor->pixmap.base ||
+	    (cursor->width != layer->src_w) ||
+	    (cursor->height != layer->src_h)) {
+		cursor->width = layer->src_w;
+		cursor->height = layer->src_h;
+
+		if (cursor->pixmap.base)
+			dma_free_writecombine(layer->dev,
+					      cursor->pixmap.size,
+					      cursor->pixmap.base,
+					      cursor->pixmap.paddr);
+
+		cursor->pixmap.size = cursor->width * cursor->height;
+
+		cursor->pixmap.base = dma_alloc_writecombine(layer->dev,
+							cursor->pixmap.size,
+							&cursor->pixmap.paddr,
+							GFP_KERNEL | GFP_DMA);
+		if (!cursor->pixmap.base) {
+			DRM_ERROR("Failed to allocate memory for pixmap\n");
+			return -ENOMEM;
+		}
+	}
+
+	/* Convert ARGB8888 to CLUT8 */
+	sti_cursor_argb8888_to_clut8(layer);
+
+	/* AWS and AWE depend on the mode */
+	y = sti_vtg_get_line_number(*mode, 0);
+	x = sti_vtg_get_pixel_number(*mode, 0);
+	val = y << 16 | x;
+	writel(val, layer->regs + CUR_AWS);
+	y = sti_vtg_get_line_number(*mode, mode->vdisplay - 1);
+	x = sti_vtg_get_pixel_number(*mode, mode->hdisplay - 1);
+	val = y << 16 | x;
+	writel(val, layer->regs + CUR_AWE);
+
+	if (first_prepare) {
+		/* Set and fetch CLUT */
+		writel(cursor->clut_paddr, layer->regs + CUR_CML);
+		writel(CUR_CTL_CLUT_UPDATE, layer->regs + CUR_CTL);
+	}
+
+	return 0;
+}
+
+static int sti_cursor_commit_layer(struct sti_layer *layer)
+{
+	struct sti_cursor *cursor = to_sti_cursor(layer);
+	struct drm_display_mode *mode = layer->mode;
+	u32 ydo, xdo;
+
+	dev_dbg(layer->dev, "%s %s\n", __func__, sti_layer_to_str(layer));
+
+	/* Set memory location, size, and position */
+	writel(cursor->pixmap.paddr, layer->regs + CUR_PML);
+	writel(cursor->width, layer->regs + CUR_PMP);
+	writel(cursor->height << 16 | cursor->width, layer->regs + CUR_SIZE);
+
+	ydo = sti_vtg_get_line_number(*mode, layer->dst_y);
+	xdo = sti_vtg_get_pixel_number(*mode, layer->dst_y);
+	writel((ydo << 16) | xdo, layer->regs + CUR_VPO);
+
+	return 0;
+}
+
+static int sti_cursor_disable_layer(struct sti_layer *layer)
+{
+	return 0;
+}
+
+static void sti_cursor_init(struct sti_layer *layer)
+{
+	struct sti_cursor *cursor = to_sti_cursor(layer);
+	unsigned short *base = cursor->clut;
+	unsigned int a, r, g, b;
+
+	/* Assign CLUT values, ARGB444 format */
+	for (a = 0; a < 4; a++)
+		for (r = 0; r < 4; r++)
+			for (g = 0; g < 4; g++)
+				for (b = 0; b < 4; b++)
+					*base++ = (a * 5) << 12 |
+						  (r * 5) << 8 |
+						  (g * 5) << 4 |
+						  (b * 5);
+}
+
+static const struct sti_layer_funcs cursor_ops = {
+	.get_formats = sti_cursor_get_formats,
+	.get_nb_formats = sti_cursor_get_nb_formats,
+	.init = sti_cursor_init,
+	.prepare = sti_cursor_prepare_layer,
+	.commit = sti_cursor_commit_layer,
+	.disable = sti_cursor_disable_layer,
+};
+
+struct sti_layer *sti_cursor_create(struct device *dev)
+{
+	struct sti_cursor *cursor;
+
+	cursor = devm_kzalloc(dev, sizeof(*cursor), GFP_KERNEL);
+	if (!cursor) {
+		DRM_ERROR("Failed to allocate memory for cursor\n");
+		return NULL;
+	}
+
+	/* Allocate clut buffer */
+	cursor->clut = dma_alloc_writecombine(dev,
+			0x100 * sizeof(unsigned short),
+			&cursor->clut_paddr,
+			GFP_KERNEL | GFP_DMA);
+
+	if (!cursor->clut) {
+		DRM_ERROR("Failed to allocate memory for cursor clut\n");
+		devm_kfree(dev, cursor);
+		return NULL;
+	}
+
+	cursor->layer.ops = &cursor_ops;
+
+	return (struct sti_layer *)cursor;
+}
diff --git a/drivers/gpu/drm/sti/sti_cursor.h b/drivers/gpu/drm/sti/sti_cursor.h
new file mode 100644
index 000000000000..3c9827404f27
--- /dev/null
+++ b/drivers/gpu/drm/sti/sti_cursor.h
@@ -0,0 +1,12 @@
+/*
+ * Copyright (C) STMicroelectronics SA 2013
+ * Authors: Vincent Abriou <vincent.abriou@st.com> for STMicroelectronics.
+ * License terms:  GNU General Public License (GPL), version 2
+ */
+
+#ifndef _STI_CURSOR_H_
+#define _STI_CURSOR_H_
+
+struct sti_layer *sti_cursor_create(struct device *dev);
+
+#endif
diff --git a/drivers/gpu/drm/sti/sti_drm_crtc.c b/drivers/gpu/drm/sti/sti_drm_crtc.c
index d2ae0c0e13be..4c651c200f20 100644
--- a/drivers/gpu/drm/sti/sti_drm_crtc.c
+++ b/drivers/gpu/drm/sti/sti_drm_crtc.c
@@ -10,6 +10,7 @@
 
 #include <drm/drmP.h>
 #include <drm/drm_crtc_helper.h>
+#include <drm/drm_plane_helper.h>
 
 #include "sti_compositor.h"
 #include "sti_drm_drv.h"
@@ -27,7 +28,7 @@ static void sti_drm_crtc_prepare(struct drm_crtc *crtc)
 	struct device *dev = mixer->dev;
 	struct sti_compositor *compo = dev_get_drvdata(dev);
 
-	compo->enable = true;
+	mixer->enabled = true;
 
 	/* Prepare and enable the compo IP clock */
 	if (mixer->id == STI_MIXER_MAIN) {
@@ -37,6 +38,8 @@ static void sti_drm_crtc_prepare(struct drm_crtc *crtc)
 		if (clk_prepare_enable(compo->clk_compo_aux))
 			DRM_INFO("Failed to prepare/enable compo_aux clk\n");
 	}
+
+	sti_mixer_clear_all_layers(mixer);
 }
 
 static void sti_drm_crtc_commit(struct drm_crtc *crtc)
@@ -61,6 +64,8 @@ static void sti_drm_crtc_commit(struct drm_crtc *crtc)
 	/* Enable layer on mixer */
 	if (sti_mixer_set_layer_status(mixer, layer, true))
 		DRM_ERROR("Can not enable layer at mixer\n");
+
+	drm_crtc_vblank_on(crtc);
 }
 
 static bool sti_drm_crtc_mode_fixup(struct drm_crtc *crtc,
@@ -143,7 +148,8 @@ sti_drm_crtc_mode_set(struct drm_crtc *crtc, struct drm_display_mode *mode,
 	w = crtc->primary->fb->width - x;
 	h = crtc->primary->fb->height - y;
 
-	return sti_layer_prepare(layer, crtc->primary->fb, &crtc->mode,
+	return sti_layer_prepare(layer, crtc,
+			crtc->primary->fb, &crtc->mode,
 			mixer->id, 0, 0, w, h, x, y, w, h);
 }
 
@@ -170,7 +176,8 @@ static int sti_drm_crtc_mode_set_base(struct drm_crtc *crtc, int x, int y,
 	w = crtc->primary->fb->width - crtc->x;
 	h = crtc->primary->fb->height - crtc->y;
 
-	ret = sti_layer_prepare(layer, crtc->primary->fb, &crtc->mode,
+	ret = sti_layer_prepare(layer, crtc,
+				crtc->primary->fb, &crtc->mode,
 				mixer->id, 0, 0, w, h,
 				crtc->x, crtc->y, w, h);
 	if (ret) {
@@ -195,7 +202,7 @@ static void sti_drm_crtc_disable(struct drm_crtc *crtc)
 	struct sti_compositor *compo = dev_get_drvdata(dev);
 	struct sti_layer *layer;
 
-	if (!compo->enable)
+	if (!mixer->enabled)
 		return;
 
 	DRM_DEBUG_KMS("CRTC:%d (%s)\n", crtc->base.id, sti_mixer_to_str(mixer));
@@ -221,7 +228,7 @@ static void sti_drm_crtc_disable(struct drm_crtc *crtc)
 	/* Then disable layer itself */
 	sti_layer_disable(layer);
 
-	drm_vblank_off(crtc->dev, mixer->id);
+	drm_crtc_vblank_off(crtc);
 
 	/* Disable pixel clock and compo IP clocks */
 	if (mixer->id == STI_MIXER_MAIN) {
@@ -232,7 +239,7 @@ static void sti_drm_crtc_disable(struct drm_crtc *crtc)
 		clk_disable_unprepare(compo->clk_compo_aux);
 	}
 
-	compo->enable = false;
+	mixer->enabled = false;
 }
 
 static struct drm_crtc_helper_funcs sti_crtc_helper_funcs = {
@@ -363,7 +370,6 @@ void sti_drm_crtc_disable_vblank(struct drm_device *dev, int crtc)
 	struct sti_drm_private *priv = dev->dev_private;
 	struct sti_compositor *compo = priv->compo;
 	struct notifier_block *vtg_vblank_nb = &compo->vtg_vblank_nb;
-	unsigned long flags;
 
 	DRM_DEBUG_DRIVER("\n");
 
@@ -372,13 +378,10 @@ void sti_drm_crtc_disable_vblank(struct drm_device *dev, int crtc)
 		DRM_DEBUG_DRIVER("Warning: cannot unregister VTG notifier\n");
 
 	/* free the resources of the pending requests */
-	spin_lock_irqsave(&dev->event_lock, flags);
 	if (compo->mixer[crtc]->pending_event) {
 		drm_vblank_put(dev, crtc);
 		compo->mixer[crtc]->pending_event = NULL;
 	}
-	spin_unlock_irqrestore(&dev->event_lock, flags);
-
 }
 EXPORT_SYMBOL(sti_drm_crtc_disable_vblank);
 
@@ -398,6 +401,7 @@ bool sti_drm_crtc_is_main(struct drm_crtc *crtc)
 
 	return false;
 }
+EXPORT_SYMBOL(sti_drm_crtc_is_main);
 
 int sti_drm_crtc_init(struct drm_device *drm_dev, struct sti_mixer *mixer,
 		struct drm_plane *primary, struct drm_plane *cursor)
diff --git a/drivers/gpu/drm/sti/sti_drm_drv.c b/drivers/gpu/drm/sti/sti_drm_drv.c
index 223d93c3a05d..5239fa121726 100644
--- a/drivers/gpu/drm/sti/sti_drm_drv.c
+++ b/drivers/gpu/drm/sti/sti_drm_drv.c
@@ -67,8 +67,12 @@ static int sti_drm_load(struct drm_device *dev, unsigned long flags)
 	sti_drm_mode_config_init(dev);
 
 	ret = component_bind_all(dev->dev, dev);
-	if (ret)
+	if (ret) {
+		drm_kms_helper_poll_fini(dev);
+		drm_mode_config_cleanup(dev);
+		kfree(private);
 		return ret;
+	}
 
 	drm_helper_disable_unused_functions(dev);
 
@@ -184,7 +188,6 @@ static struct platform_driver sti_drm_master_driver = {
 	.probe = sti_drm_master_probe,
 	.remove = sti_drm_master_remove,
 	.driver = {
-		.owner = THIS_MODULE,
 		.name = DRIVER_NAME "__master",
 	},
 };
@@ -228,7 +231,6 @@ static struct platform_driver sti_drm_platform_driver = {
 	.probe = sti_drm_platform_probe,
 	.remove = sti_drm_platform_remove,
 	.driver = {
-		.owner = THIS_MODULE,
 		.name = DRIVER_NAME,
 		.of_match_table = sti_drm_dt_ids,
 	},
diff --git a/drivers/gpu/drm/sti/sti_drm_plane.c b/drivers/gpu/drm/sti/sti_drm_plane.c
index f4118d4cac22..bb6a29339e10 100644
--- a/drivers/gpu/drm/sti/sti_drm_plane.c
+++ b/drivers/gpu/drm/sti/sti_drm_plane.c
@@ -45,7 +45,8 @@ sti_drm_update_plane(struct drm_plane *plane, struct drm_crtc *crtc,
 	}
 
 	/* src_x are in 16.16 format. */
-	res = sti_layer_prepare(layer, fb, &crtc->mode, mixer->id,
+	res = sti_layer_prepare(layer, crtc, fb,
+			&crtc->mode, mixer->id,
 			crtc_x, crtc_y, crtc_w, crtc_h,
 			src_x >> 16, src_y >> 16,
 			src_w >> 16, src_h >> 16);
@@ -193,3 +194,4 @@ struct drm_plane *sti_drm_plane_init(struct drm_device *dev,
 
 	return &layer->plane;
 }
+EXPORT_SYMBOL(sti_drm_plane_init);
diff --git a/drivers/gpu/drm/sti/sti_gdp.c b/drivers/gpu/drm/sti/sti_gdp.c
index 4e30b74559f5..32448d1d1e8f 100644
--- a/drivers/gpu/drm/sti/sti_gdp.c
+++ b/drivers/gpu/drm/sti/sti_gdp.c
@@ -73,7 +73,9 @@ struct sti_gdp_node {
 
 struct sti_gdp_node_list {
 	struct sti_gdp_node *top_field;
+	dma_addr_t top_field_paddr;
 	struct sti_gdp_node *btm_field;
+	dma_addr_t btm_field_paddr;
 };
 
 /**
@@ -81,6 +83,8 @@ struct sti_gdp_node_list {
  *
  * @layer:		layer structure
  * @clk_pix:            pixel clock for the current gdp
+ * @clk_main_parent:    gdp parent clock if main path used
+ * @clk_aux_parent:     gdp parent clock if aux path used
  * @vtg_field_nb:       callback for VTG FIELD (top or bottom) notification
  * @is_curr_top:        true if the current node processed is the top field
  * @node_list:		array of node list
@@ -88,6 +92,8 @@ struct sti_gdp_node_list {
 struct sti_gdp {
 	struct sti_layer layer;
 	struct clk *clk_pix;
+	struct clk *clk_main_parent;
+	struct clk *clk_aux_parent;
 	struct notifier_block vtg_field_nb;
 	bool is_curr_top;
 	struct sti_gdp_node_list node_list[GDP_NODE_NB_BANK];
@@ -168,7 +174,6 @@ static int sti_gdp_get_alpharange(int format)
 static struct sti_gdp_node_list *sti_gdp_get_free_nodes(struct sti_layer *layer)
 {
 	int hw_nvn;
-	void *virt_nvn;
 	struct sti_gdp *gdp = to_sti_gdp(layer);
 	unsigned int i;
 
@@ -176,11 +181,9 @@ static struct sti_gdp_node_list *sti_gdp_get_free_nodes(struct sti_layer *layer)
 	if (!hw_nvn)
 		goto end;
 
-	virt_nvn = dma_to_virt(layer->dev, (dma_addr_t) hw_nvn);
-
 	for (i = 0; i < GDP_NODE_NB_BANK; i++)
-		if ((virt_nvn != gdp->node_list[i].btm_field) &&
-		    (virt_nvn != gdp->node_list[i].top_field))
+		if ((hw_nvn != gdp->node_list[i].btm_field_paddr) &&
+		    (hw_nvn != gdp->node_list[i].top_field_paddr))
 			return &gdp->node_list[i];
 
 	/* in hazardious cases restart with the first node */
@@ -204,7 +207,6 @@ static
 struct sti_gdp_node_list *sti_gdp_get_current_nodes(struct sti_layer *layer)
 {
 	int hw_nvn;
-	void *virt_nvn;
 	struct sti_gdp *gdp = to_sti_gdp(layer);
 	unsigned int i;
 
@@ -212,11 +214,9 @@ struct sti_gdp_node_list *sti_gdp_get_current_nodes(struct sti_layer *layer)
 	if (!hw_nvn)
 		goto end;
 
-	virt_nvn = dma_to_virt(layer->dev, (dma_addr_t) hw_nvn);
-
 	for (i = 0; i < GDP_NODE_NB_BANK; i++)
-		if ((virt_nvn == gdp->node_list[i].btm_field) ||
-				(virt_nvn == gdp->node_list[i].top_field))
+		if ((hw_nvn == gdp->node_list[i].btm_field_paddr) ||
+				(hw_nvn == gdp->node_list[i].top_field_paddr))
 			return &gdp->node_list[i];
 
 end:
@@ -292,8 +292,8 @@ static int sti_gdp_prepare_layer(struct sti_layer *layer, bool first_prepare)
 
 	/* Same content and chained together */
 	memcpy(btm_field, top_field, sizeof(*btm_field));
-	top_field->gam_gdp_nvn = virt_to_dma(dev, btm_field);
-	btm_field->gam_gdp_nvn = virt_to_dma(dev, top_field);
+	top_field->gam_gdp_nvn = list->btm_field_paddr;
+	btm_field->gam_gdp_nvn = list->top_field_paddr;
 
 	/* Interlaced mode */
 	if (layer->mode->flags & DRM_MODE_FLAG_INTERLACE)
@@ -311,6 +311,17 @@ static int sti_gdp_prepare_layer(struct sti_layer *layer, bool first_prepare)
 
 		/* Set and enable gdp clock */
 		if (gdp->clk_pix) {
+			struct clk *clkp;
+			/* According to the mixer used, the gdp pixel clock
+			 * should have a different parent clock. */
+			if (layer->mixer_id == STI_MIXER_MAIN)
+				clkp = gdp->clk_main_parent;
+			else
+				clkp = gdp->clk_aux_parent;
+
+			if (clkp)
+				clk_set_parent(gdp->clk_pix, clkp);
+
 			res = clk_set_rate(gdp->clk_pix, rate);
 			if (res < 0) {
 				DRM_ERROR("Cannot set rate (%dHz) for gdp\n",
@@ -349,8 +360,8 @@ static int sti_gdp_commit_layer(struct sti_layer *layer)
 	struct sti_gdp_node *updated_top_node = updated_list->top_field;
 	struct sti_gdp_node *updated_btm_node = updated_list->btm_field;
 	struct sti_gdp *gdp = to_sti_gdp(layer);
-	u32 dma_updated_top = virt_to_dma(layer->dev, updated_top_node);
-	u32 dma_updated_btm = virt_to_dma(layer->dev, updated_btm_node);
+	u32 dma_updated_top = updated_list->top_field_paddr;
+	u32 dma_updated_btm = updated_list->btm_field_paddr;
 	struct sti_gdp_node_list *curr_list = sti_gdp_get_current_nodes(layer);
 
 	dev_dbg(layer->dev, "%s %s top/btm_node:0x%p/0x%p\n", __func__,
@@ -461,16 +472,16 @@ static void sti_gdp_init(struct sti_layer *layer)
 {
 	struct sti_gdp *gdp = to_sti_gdp(layer);
 	struct device_node *np = layer->dev->of_node;
-	dma_addr_t dma;
+	dma_addr_t dma_addr;
 	void *base;
 	unsigned int i, size;
 
 	/* Allocate all the nodes within a single memory page */
 	size = sizeof(struct sti_gdp_node) *
 	    GDP_NODE_PER_FIELD * GDP_NODE_NB_BANK;
-
 	base = dma_alloc_writecombine(layer->dev,
-			size, &dma, GFP_KERNEL | GFP_DMA);
+			size, &dma_addr, GFP_KERNEL | GFP_DMA);
+
 	if (!base) {
 		DRM_ERROR("Failed to allocate memory for GDP node\n");
 		return;
@@ -478,21 +489,26 @@ static void sti_gdp_init(struct sti_layer *layer)
 	memset(base, 0, size);
 
 	for (i = 0; i < GDP_NODE_NB_BANK; i++) {
-		if (virt_to_dma(layer->dev, base) & 0xF) {
+		if (dma_addr & 0xF) {
 			DRM_ERROR("Mem alignment failed\n");
 			return;
 		}
 		gdp->node_list[i].top_field = base;
+		gdp->node_list[i].top_field_paddr = dma_addr;
+
 		DRM_DEBUG_DRIVER("node[%d].top_field=%p\n", i, base);
 		base += sizeof(struct sti_gdp_node);
+		dma_addr += sizeof(struct sti_gdp_node);
 
-		if (virt_to_dma(layer->dev, base) & 0xF) {
+		if (dma_addr & 0xF) {
 			DRM_ERROR("Mem alignment failed\n");
 			return;
 		}
 		gdp->node_list[i].btm_field = base;
+		gdp->node_list[i].btm_field_paddr = dma_addr;
 		DRM_DEBUG_DRIVER("node[%d].btm_field=%p\n", i, base);
 		base += sizeof(struct sti_gdp_node);
+		dma_addr += sizeof(struct sti_gdp_node);
 	}
 
 	if (of_device_is_compatible(np, "st,stih407-compositor")) {
@@ -520,6 +536,14 @@ static void sti_gdp_init(struct sti_layer *layer)
 		gdp->clk_pix = devm_clk_get(layer->dev, clk_name);
 		if (IS_ERR(gdp->clk_pix))
 			DRM_ERROR("Cannot get %s clock\n", clk_name);
+
+		gdp->clk_main_parent = devm_clk_get(layer->dev, "main_parent");
+		if (IS_ERR(gdp->clk_main_parent))
+			DRM_ERROR("Cannot get main_parent clock\n");
+
+		gdp->clk_aux_parent = devm_clk_get(layer->dev, "aux_parent");
+		if (IS_ERR(gdp->clk_aux_parent))
+			DRM_ERROR("Cannot get aux_parent clock\n");
 	}
 }
 
diff --git a/drivers/gpu/drm/sti/sti_hdmi.c b/drivers/gpu/drm/sti/sti_hdmi.c
index b22968c08d1f..d032e024b0b8 100644
--- a/drivers/gpu/drm/sti/sti_hdmi.c
+++ b/drivers/gpu/drm/sti/sti_hdmi.c
@@ -130,8 +130,7 @@ static irqreturn_t hdmi_irq_thread(int irq, void *arg)
 
 	/* Hot plug/unplug IRQ */
 	if (hdmi->irq_status & HDMI_INT_HOT_PLUG) {
-		/* read gpio to get the status */
-		hdmi->hpd = gpio_get_value(hdmi->hpd_gpio);
+		hdmi->hpd = readl(hdmi->regs + HDMI_STA) & HDMI_STA_HOT_PLUG;
 		if (hdmi->drm_dev)
 			drm_helper_hpd_irq_event(hdmi->drm_dev);
 	}
@@ -273,31 +272,32 @@ static int hdmi_avi_infoframe_config(struct sti_hdmi *hdmi)
 	hdmi_write(hdmi, val, HDMI_SW_DI_CFG);
 
 	/* Infoframe header */
-	val = buffer[0x0];
-	val |= buffer[0x1] << 8;
-	val |= buffer[0x2] << 16;
+	val =  buffer[0];
+	val |= buffer[1] << 8;
+	val |= buffer[2] << 16;
 	hdmi_write(hdmi, val, HDMI_SW_DI_N_HEAD_WORD(HDMI_IFRAME_SLOT_AVI));
 
 	/* Infoframe packet bytes */
-	val = frame[0x0];
-	val |= frame[0x1] << 8;
-	val |= frame[0x2] << 16;
-	val |= frame[0x3] << 24;
+	val =  buffer[3];
+	val |= *(frame++) << 8;
+	val |= *(frame++) << 16;
+	val |= *(frame++) << 24;
 	hdmi_write(hdmi, val, HDMI_SW_DI_N_PKT_WORD0(HDMI_IFRAME_SLOT_AVI));
 
-	val = frame[0x4];
-	val |= frame[0x5] << 8;
-	val |= frame[0x6] << 16;
-	val |= frame[0x7] << 24;
+	val =  *(frame++);
+	val |= *(frame++) << 8;
+	val |= *(frame++) << 16;
+	val |= *(frame++) << 24;
 	hdmi_write(hdmi, val, HDMI_SW_DI_N_PKT_WORD1(HDMI_IFRAME_SLOT_AVI));
 
-	val = frame[0x8];
-	val |= frame[0x9] << 8;
-	val |= frame[0xA] << 16;
-	val |= frame[0xB] << 24;
+	val =  *(frame++);
+	val |= *(frame++) << 8;
+	val |= *(frame++) << 16;
+	val |= *(frame++) << 24;
 	hdmi_write(hdmi, val, HDMI_SW_DI_N_PKT_WORD2(HDMI_IFRAME_SLOT_AVI));
 
-	val = frame[0xC];
+	val = *(frame++);
+	val |= *(frame) << 8;
 	hdmi_write(hdmi, val, HDMI_SW_DI_N_PKT_WORD3(HDMI_IFRAME_SLOT_AVI));
 
 	/* Enable transmission slot for AVI infoframe
@@ -480,17 +480,15 @@ static const struct drm_bridge_funcs sti_hdmi_bridge_funcs = {
 
 static int sti_hdmi_connector_get_modes(struct drm_connector *connector)
 {
-	struct i2c_adapter *i2c_adap;
+	struct sti_hdmi_connector *hdmi_connector
+		= to_sti_hdmi_connector(connector);
+	struct sti_hdmi *hdmi = hdmi_connector->hdmi;
 	struct edid *edid;
 	int count;
 
 	DRM_DEBUG_DRIVER("\n");
 
-	i2c_adap = i2c_get_adapter(1);
-	if (!i2c_adap)
-		goto fail;
-
-	edid = drm_get_edid(connector, i2c_adap);
+	edid = drm_get_edid(connector, hdmi->ddc_adapt);
 	if (!edid)
 		goto fail;
 
@@ -603,29 +601,38 @@ static int sti_hdmi_bind(struct device *dev, struct device *master, void *data)
 	struct sti_hdmi_connector *connector;
 	struct drm_connector *drm_connector;
 	struct drm_bridge *bridge;
-	struct i2c_adapter *i2c_adap;
+	struct device_node *ddc;
 	int err;
 
-	i2c_adap = i2c_get_adapter(1);
-	if (!i2c_adap)
-		return -EPROBE_DEFER;
+	ddc = of_parse_phandle(dev->of_node, "ddc", 0);
+	if (ddc) {
+		hdmi->ddc_adapt = of_find_i2c_adapter_by_node(ddc);
+		if (!hdmi->ddc_adapt) {
+			err = -EPROBE_DEFER;
+			of_node_put(ddc);
+			return err;
+		}
+
+		of_node_put(ddc);
+	}
 
 	/* Set the drm device handle */
 	hdmi->drm_dev = drm_dev;
 
 	encoder = sti_hdmi_find_encoder(drm_dev);
 	if (!encoder)
-		return -ENOMEM;
+		goto err_adapt;
 
 	connector = devm_kzalloc(dev, sizeof(*connector), GFP_KERNEL);
 	if (!connector)
-		return -ENOMEM;
+		goto err_adapt;
+
 
 	connector->hdmi = hdmi;
 
 	bridge = devm_kzalloc(dev, sizeof(*bridge), GFP_KERNEL);
 	if (!bridge)
-		return -ENOMEM;
+		goto err_adapt;
 
 	bridge->driver_private = hdmi;
 	drm_bridge_init(drm_dev, bridge, &sti_hdmi_bridge_funcs);
@@ -662,6 +669,8 @@ err_sysfs:
 err_connector:
 	drm_bridge_cleanup(bridge);
 	drm_connector_cleanup(drm_connector);
+err_adapt:
+	put_device(&hdmi->ddc_adapt->dev);
 	return -EINVAL;
 }
 
@@ -757,13 +766,7 @@ static int sti_hdmi_probe(struct platform_device *pdev)
 		return PTR_ERR(hdmi->clk_audio);
 	}
 
-	hdmi->hpd_gpio = of_get_named_gpio(np, "hdmi,hpd-gpio", 0);
-	if (hdmi->hpd_gpio < 0) {
-		DRM_ERROR("Failed to get hdmi hpd-gpio\n");
-		return -EIO;
-	}
-
-	hdmi->hpd = gpio_get_value(hdmi->hpd_gpio);
+	hdmi->hpd = readl(hdmi->regs + HDMI_STA) & HDMI_STA_HOT_PLUG;
 
 	init_waitqueue_head(&hdmi->wait_event);
 
@@ -788,6 +791,11 @@ static int sti_hdmi_probe(struct platform_device *pdev)
 
 static int sti_hdmi_remove(struct platform_device *pdev)
 {
+	struct sti_hdmi *hdmi = dev_get_drvdata(&pdev->dev);
+
+	if (hdmi->ddc_adapt)
+		put_device(&hdmi->ddc_adapt->dev);
+
 	component_del(&pdev->dev, &sti_hdmi_ops);
 	return 0;
 }
diff --git a/drivers/gpu/drm/sti/sti_hdmi.h b/drivers/gpu/drm/sti/sti_hdmi.h
index 61bec6557ceb..3d22390e1f3b 100644
--- a/drivers/gpu/drm/sti/sti_hdmi.h
+++ b/drivers/gpu/drm/sti/sti_hdmi.h
@@ -14,6 +14,9 @@
 #define HDMI_STA           0x0010
 #define HDMI_STA_DLL_LCK   BIT(5)
 
+#define HDMI_STA_HOT_PLUG_SHIFT 4
+#define HDMI_STA_HOT_PLUG	(1 << HDMI_STA_HOT_PLUG_SHIFT)
+
 struct sti_hdmi;
 
 struct hdmi_phy_ops {
@@ -37,7 +40,6 @@ struct hdmi_phy_ops {
  * @irq_status: interrupt status register
  * @phy_ops: phy start/stop operations
  * @enabled: true if hdmi is enabled else false
- * @hpd_gpio: hdmi hot plug detect gpio number
  * @hpd: hot plug detect status
  * @wait_event: wait event
  * @event_received: wait event status
@@ -57,11 +59,11 @@ struct sti_hdmi {
 	u32 irq_status;
 	struct hdmi_phy_ops *phy_ops;
 	bool enabled;
-	int hpd_gpio;
 	bool hpd;
 	wait_queue_head_t wait_event;
 	bool event_received;
 	struct reset_control *reset;
+	struct i2c_adapter *ddc_adapt;
 };
 
 u32 hdmi_read(struct sti_hdmi *hdmi, int offset);
diff --git a/drivers/gpu/drm/sti/sti_hqvdp.c b/drivers/gpu/drm/sti/sti_hqvdp.c
new file mode 100644
index 000000000000..f3db05dab0ab
--- /dev/null
+++ b/drivers/gpu/drm/sti/sti_hqvdp.c
@@ -0,0 +1,1073 @@
+/*
+ * Copyright (C) STMicroelectronics SA 2014
+ * Authors: Fabien Dessenne <fabien.dessenne@st.com> for STMicroelectronics.
+ * License terms:  GNU General Public License (GPL), version 2
+ */
+
+#include <linux/clk.h>
+#include <linux/component.h>
+#include <linux/firmware.h>
+#include <linux/module.h>
+#include <linux/platform_device.h>
+#include <linux/reset.h>
+
+#include <drm/drmP.h>
+
+#include "sti_drm_plane.h"
+#include "sti_hqvdp.h"
+#include "sti_hqvdp_lut.h"
+#include "sti_layer.h"
+#include "sti_vtg.h"
+
+/* Firmware name */
+#define HQVDP_FMW_NAME          "hqvdp-stih407.bin"
+
+/* Regs address */
+#define HQVDP_DMEM              0x00000000               /* 0x00000000 */
+#define HQVDP_PMEM              0x00040000               /* 0x00040000 */
+#define HQVDP_RD_PLUG           0x000E0000               /* 0x000E0000 */
+#define HQVDP_RD_PLUG_CONTROL   (HQVDP_RD_PLUG + 0x1000) /* 0x000E1000 */
+#define HQVDP_RD_PLUG_PAGE_SIZE (HQVDP_RD_PLUG + 0x1004) /* 0x000E1004 */
+#define HQVDP_RD_PLUG_MIN_OPC   (HQVDP_RD_PLUG + 0x1008) /* 0x000E1008 */
+#define HQVDP_RD_PLUG_MAX_OPC   (HQVDP_RD_PLUG + 0x100C) /* 0x000E100C */
+#define HQVDP_RD_PLUG_MAX_CHK   (HQVDP_RD_PLUG + 0x1010) /* 0x000E1010 */
+#define HQVDP_RD_PLUG_MAX_MSG   (HQVDP_RD_PLUG + 0x1014) /* 0x000E1014 */
+#define HQVDP_RD_PLUG_MIN_SPACE (HQVDP_RD_PLUG + 0x1018) /* 0x000E1018 */
+#define HQVDP_WR_PLUG           0x000E2000               /* 0x000E2000 */
+#define HQVDP_WR_PLUG_CONTROL   (HQVDP_WR_PLUG + 0x1000) /* 0x000E3000 */
+#define HQVDP_WR_PLUG_PAGE_SIZE (HQVDP_WR_PLUG + 0x1004) /* 0x000E3004 */
+#define HQVDP_WR_PLUG_MIN_OPC   (HQVDP_WR_PLUG + 0x1008) /* 0x000E3008 */
+#define HQVDP_WR_PLUG_MAX_OPC   (HQVDP_WR_PLUG + 0x100C) /* 0x000E300C */
+#define HQVDP_WR_PLUG_MAX_CHK   (HQVDP_WR_PLUG + 0x1010) /* 0x000E3010 */
+#define HQVDP_WR_PLUG_MAX_MSG   (HQVDP_WR_PLUG + 0x1014) /* 0x000E3014 */
+#define HQVDP_WR_PLUG_MIN_SPACE (HQVDP_WR_PLUG + 0x1018) /* 0x000E3018 */
+#define HQVDP_MBX               0x000E4000               /* 0x000E4000 */
+#define HQVDP_MBX_IRQ_TO_XP70   (HQVDP_MBX + 0x0000)     /* 0x000E4000 */
+#define HQVDP_MBX_INFO_HOST     (HQVDP_MBX + 0x0004)     /* 0x000E4004 */
+#define HQVDP_MBX_IRQ_TO_HOST   (HQVDP_MBX + 0x0008)     /* 0x000E4008 */
+#define HQVDP_MBX_INFO_XP70     (HQVDP_MBX + 0x000C)     /* 0x000E400C */
+#define HQVDP_MBX_SW_RESET_CTRL (HQVDP_MBX + 0x0010)     /* 0x000E4010 */
+#define HQVDP_MBX_STARTUP_CTRL1 (HQVDP_MBX + 0x0014)     /* 0x000E4014 */
+#define HQVDP_MBX_STARTUP_CTRL2 (HQVDP_MBX + 0x0018)     /* 0x000E4018 */
+#define HQVDP_MBX_GP_STATUS     (HQVDP_MBX + 0x001C)     /* 0x000E401C */
+#define HQVDP_MBX_NEXT_CMD      (HQVDP_MBX + 0x0020)     /* 0x000E4020 */
+#define HQVDP_MBX_CURRENT_CMD   (HQVDP_MBX + 0x0024)     /* 0x000E4024 */
+#define HQVDP_MBX_SOFT_VSYNC    (HQVDP_MBX + 0x0028)     /* 0x000E4028 */
+
+/* Plugs config */
+#define PLUG_CONTROL_ENABLE     0x00000001
+#define PLUG_PAGE_SIZE_256      0x00000002
+#define PLUG_MIN_OPC_8          0x00000003
+#define PLUG_MAX_OPC_64         0x00000006
+#define PLUG_MAX_CHK_2X         0x00000001
+#define PLUG_MAX_MSG_1X         0x00000000
+#define PLUG_MIN_SPACE_1        0x00000000
+
+/* SW reset CTRL */
+#define SW_RESET_CTRL_FULL      BIT(0)
+#define SW_RESET_CTRL_CORE      BIT(1)
+
+/* Startup ctrl 1 */
+#define STARTUP_CTRL1_RST_DONE  BIT(0)
+#define STARTUP_CTRL1_AUTH_IDLE BIT(2)
+
+/* Startup ctrl 2 */
+#define STARTUP_CTRL2_FETCH_EN  BIT(1)
+
+/* Info xP70 */
+#define INFO_XP70_FW_READY      BIT(15)
+#define INFO_XP70_FW_PROCESSING BIT(14)
+#define INFO_XP70_FW_INITQUEUES BIT(13)
+
+/* SOFT_VSYNC */
+#define SOFT_VSYNC_HW           0x00000000
+#define SOFT_VSYNC_SW_CMD       0x00000001
+#define SOFT_VSYNC_SW_CTRL_IRQ  0x00000003
+
+/* Reset & boot poll config */
+#define POLL_MAX_ATTEMPT        50
+#define POLL_DELAY_MS           20
+
+#define SCALE_FACTOR            8192
+#define SCALE_MAX_FOR_LEG_LUT_F 4096
+#define SCALE_MAX_FOR_LEG_LUT_E 4915
+#define SCALE_MAX_FOR_LEG_LUT_D 6654
+#define SCALE_MAX_FOR_LEG_LUT_C 8192
+
+enum sti_hvsrc_orient {
+	HVSRC_HORI,
+	HVSRC_VERT
+};
+
+/* Command structures */
+struct sti_hqvdp_top {
+	u32 config;
+	u32 mem_format;
+	u32 current_luma;
+	u32 current_enh_luma;
+	u32 current_right_luma;
+	u32 current_enh_right_luma;
+	u32 current_chroma;
+	u32 current_enh_chroma;
+	u32 current_right_chroma;
+	u32 current_enh_right_chroma;
+	u32 output_luma;
+	u32 output_chroma;
+	u32 luma_src_pitch;
+	u32 luma_enh_src_pitch;
+	u32 luma_right_src_pitch;
+	u32 luma_enh_right_src_pitch;
+	u32 chroma_src_pitch;
+	u32 chroma_enh_src_pitch;
+	u32 chroma_right_src_pitch;
+	u32 chroma_enh_right_src_pitch;
+	u32 luma_processed_pitch;
+	u32 chroma_processed_pitch;
+	u32 input_frame_size;
+	u32 input_viewport_ori;
+	u32 input_viewport_ori_right;
+	u32 input_viewport_size;
+	u32 left_view_border_width;
+	u32 right_view_border_width;
+	u32 left_view_3d_offset_width;
+	u32 right_view_3d_offset_width;
+	u32 side_stripe_color;
+	u32 crc_reset_ctrl;
+};
+
+/* Configs for interlaced : no IT, no pass thru, 3 fields */
+#define TOP_CONFIG_INTER_BTM            0x00000000
+#define TOP_CONFIG_INTER_TOP            0x00000002
+
+/* Config for progressive : no IT, no pass thru, 3 fields */
+#define TOP_CONFIG_PROGRESSIVE          0x00000001
+
+/* Default MemFormat: in=420_raster_dual out=444_raster;opaque Mem2Tv mode */
+#define TOP_MEM_FORMAT_DFLT             0x00018060
+
+/* Min/Max size */
+#define MAX_WIDTH                       0x1FFF
+#define MAX_HEIGHT                      0x0FFF
+#define MIN_WIDTH                       0x0030
+#define MIN_HEIGHT                      0x0010
+
+struct sti_hqvdp_vc1re {
+	u32 ctrl_prv_csdi;
+	u32 ctrl_cur_csdi;
+	u32 ctrl_nxt_csdi;
+	u32 ctrl_cur_fmd;
+	u32 ctrl_nxt_fmd;
+};
+
+struct sti_hqvdp_fmd {
+	u32 config;
+	u32 viewport_ori;
+	u32 viewport_size;
+	u32 next_next_luma;
+	u32 next_next_right_luma;
+	u32 next_next_next_luma;
+	u32 next_next_next_right_luma;
+	u32 threshold_scd;
+	u32 threshold_rfd;
+	u32 threshold_move;
+	u32 threshold_cfd;
+};
+
+struct sti_hqvdp_csdi {
+	u32 config;
+	u32 config2;
+	u32 dcdi_config;
+	u32 prev_luma;
+	u32 prev_enh_luma;
+	u32 prev_right_luma;
+	u32 prev_enh_right_luma;
+	u32 next_luma;
+	u32 next_enh_luma;
+	u32 next_right_luma;
+	u32 next_enh_right_luma;
+	u32 prev_chroma;
+	u32 prev_enh_chroma;
+	u32 prev_right_chroma;
+	u32 prev_enh_right_chroma;
+	u32 next_chroma;
+	u32 next_enh_chroma;
+	u32 next_right_chroma;
+	u32 next_enh_right_chroma;
+	u32 prev_motion;
+	u32 prev_right_motion;
+	u32 cur_motion;
+	u32 cur_right_motion;
+	u32 next_motion;
+	u32 next_right_motion;
+};
+
+/* Config for progressive: by pass */
+#define CSDI_CONFIG_PROG                0x00000000
+/* Config for directional deinterlacing without motion */
+#define CSDI_CONFIG_INTER_DIR           0x00000016
+/* Additional configs for fader, blender, motion,... deinterlace algorithms */
+#define CSDI_CONFIG2_DFLT               0x000001B3
+#define CSDI_DCDI_CONFIG_DFLT           0x00203803
+
+struct sti_hqvdp_hvsrc {
+	u32 hor_panoramic_ctrl;
+	u32 output_picture_size;
+	u32 init_horizontal;
+	u32 init_vertical;
+	u32 param_ctrl;
+	u32 yh_coef[NB_COEF];
+	u32 ch_coef[NB_COEF];
+	u32 yv_coef[NB_COEF];
+	u32 cv_coef[NB_COEF];
+	u32 hori_shift;
+	u32 vert_shift;
+};
+
+/* Default ParamCtrl: all controls enabled */
+#define HVSRC_PARAM_CTRL_DFLT           0xFFFFFFFF
+
+struct sti_hqvdp_iqi {
+	u32 config;
+	u32 demo_wind_size;
+	u32 pk_config;
+	u32 coeff0_coeff1;
+	u32 coeff2_coeff3;
+	u32 coeff4;
+	u32 pk_lut;
+	u32 pk_gain;
+	u32 pk_coring_level;
+	u32 cti_config;
+	u32 le_config;
+	u32 le_lut[64];
+	u32 con_bri;
+	u32 sat_gain;
+	u32 pxf_conf;
+	u32 default_color;
+};
+
+/* Default Config : IQI bypassed */
+#define IQI_CONFIG_DFLT                 0x00000001
+/* Default Contrast & Brightness gain = 256 */
+#define IQI_CON_BRI_DFLT                0x00000100
+/* Default Saturation gain = 256 */
+#define IQI_SAT_GAIN_DFLT               0x00000100
+/* Default PxfConf : P2I bypassed */
+#define IQI_PXF_CONF_DFLT               0x00000001
+
+struct sti_hqvdp_top_status {
+	u32 processing_time;
+	u32 input_y_crc;
+	u32 input_uv_crc;
+};
+
+struct sti_hqvdp_fmd_status {
+	u32 fmd_repeat_move_status;
+	u32 fmd_scene_count_status;
+	u32 cfd_sum;
+	u32 field_sum;
+	u32 next_y_fmd_crc;
+	u32 next_next_y_fmd_crc;
+	u32 next_next_next_y_fmd_crc;
+};
+
+struct sti_hqvdp_csdi_status {
+	u32 prev_y_csdi_crc;
+	u32 cur_y_csdi_crc;
+	u32 next_y_csdi_crc;
+	u32 prev_uv_csdi_crc;
+	u32 cur_uv_csdi_crc;
+	u32 next_uv_csdi_crc;
+	u32 y_csdi_crc;
+	u32 uv_csdi_crc;
+	u32 uv_cup_crc;
+	u32 mot_csdi_crc;
+	u32 mot_cur_csdi_crc;
+	u32 mot_prev_csdi_crc;
+};
+
+struct sti_hqvdp_hvsrc_status {
+	u32 y_hvsrc_crc;
+	u32 u_hvsrc_crc;
+	u32 v_hvsrc_crc;
+};
+
+struct sti_hqvdp_iqi_status {
+	u32 pxf_it_status;
+	u32 y_iqi_crc;
+	u32 u_iqi_crc;
+	u32 v_iqi_crc;
+};
+
+/* Main commands. We use 2 commands one being processed by the firmware, one
+ * ready to be fetched upon next Vsync*/
+#define NB_VDP_CMD	2
+
+struct sti_hqvdp_cmd {
+	struct sti_hqvdp_top top;
+	struct sti_hqvdp_vc1re vc1re;
+	struct sti_hqvdp_fmd fmd;
+	struct sti_hqvdp_csdi csdi;
+	struct sti_hqvdp_hvsrc hvsrc;
+	struct sti_hqvdp_iqi iqi;
+	struct sti_hqvdp_top_status top_status;
+	struct sti_hqvdp_fmd_status fmd_status;
+	struct sti_hqvdp_csdi_status csdi_status;
+	struct sti_hqvdp_hvsrc_status hvsrc_status;
+	struct sti_hqvdp_iqi_status iqi_status;
+};
+
+/*
+ * STI HQVDP structure
+ *
+ * @dev:               driver device
+ * @drm_dev:           the drm device
+ * @regs:              registers
+ * @layer:             layer structure for hqvdp it self
+ * @vid_plane:         VID plug used as link with compositor IP
+ * @clk:               IP clock
+ * @clk_pix_main:      pix main clock
+ * @reset:             reset control
+ * @vtg_nb:            notifier to handle VTG Vsync
+ * @btm_field_pending: is there any bottom field (interlaced frame) to display
+ * @curr_field_count:  number of field updates
+ * @last_field_count:  number of field updates since last fps measure
+ * @hqvdp_cmd:         buffer of commands
+ * @hqvdp_cmd_paddr:   physical address of hqvdp_cmd
+ * @vtg:               vtg for main data path
+ */
+struct sti_hqvdp {
+	struct device *dev;
+	struct drm_device *drm_dev;
+	void __iomem *regs;
+	struct sti_layer layer;
+	struct drm_plane *vid_plane;
+	struct clk *clk;
+	struct clk *clk_pix_main;
+	struct reset_control *reset;
+	struct notifier_block vtg_nb;
+	bool btm_field_pending;
+	unsigned int curr_field_count;
+	unsigned int last_field_count;
+	void *hqvdp_cmd;
+	dma_addr_t hqvdp_cmd_paddr;
+	struct sti_vtg *vtg;
+};
+
+#define to_sti_hqvdp(x) container_of(x, struct sti_hqvdp, layer)
+
+static const uint32_t hqvdp_supported_formats[] = {
+	DRM_FORMAT_NV12,
+};
+
+static const uint32_t *sti_hqvdp_get_formats(struct sti_layer *layer)
+{
+	return hqvdp_supported_formats;
+}
+
+static unsigned int sti_hqvdp_get_nb_formats(struct sti_layer *layer)
+{
+	return ARRAY_SIZE(hqvdp_supported_formats);
+}
+
+/**
+ * sti_hqvdp_get_free_cmd
+ * @hqvdp: hqvdp structure
+ *
+ * Look for a hqvdp_cmd that is not being used (or about to be used) by the FW.
+ *
+ * RETURNS:
+ * the offset of the command to be used.
+ * -1 in error cases
+ */
+static int sti_hqvdp_get_free_cmd(struct sti_hqvdp *hqvdp)
+{
+	int curr_cmd, next_cmd;
+	dma_addr_t cmd = hqvdp->hqvdp_cmd_paddr;
+	int i;
+
+	curr_cmd = readl(hqvdp->regs + HQVDP_MBX_CURRENT_CMD);
+	next_cmd = readl(hqvdp->regs + HQVDP_MBX_NEXT_CMD);
+
+	for (i = 0; i < NB_VDP_CMD; i++) {
+		if ((cmd != curr_cmd) && (cmd != next_cmd))
+			return i * sizeof(struct sti_hqvdp_cmd);
+		cmd += sizeof(struct sti_hqvdp_cmd);
+	}
+
+	return -1;
+}
+
+/**
+ * sti_hqvdp_get_curr_cmd
+ * @hqvdp: hqvdp structure
+ *
+ * Look for the hqvdp_cmd that is being used by the FW.
+ *
+ * RETURNS:
+ *  the offset of the command to be used.
+ * -1 in error cases
+ */
+static int sti_hqvdp_get_curr_cmd(struct sti_hqvdp *hqvdp)
+{
+	int curr_cmd;
+	dma_addr_t cmd = hqvdp->hqvdp_cmd_paddr;
+	unsigned int i;
+
+	curr_cmd = readl(hqvdp->regs + HQVDP_MBX_CURRENT_CMD);
+
+	for (i = 0; i < NB_VDP_CMD; i++) {
+		if (cmd == curr_cmd)
+			return i * sizeof(struct sti_hqvdp_cmd);
+
+		cmd += sizeof(struct sti_hqvdp_cmd);
+	}
+
+	return -1;
+}
+
+/**
+ * sti_hqvdp_update_hvsrc
+ * @orient: horizontal or vertical
+ * @scale:  scaling/zoom factor
+ * @hvsrc:  the structure containing the LUT coef
+ *
+ * Update the Y and C Lut coef, as well as the shift param
+ *
+ * RETURNS:
+ * None.
+ */
+static void sti_hqvdp_update_hvsrc(enum sti_hvsrc_orient orient, int scale,
+		struct sti_hqvdp_hvsrc *hvsrc)
+{
+	const int *coef_c, *coef_y;
+	int shift_c, shift_y;
+
+	/* Get the appropriate coef tables */
+	if (scale < SCALE_MAX_FOR_LEG_LUT_F) {
+		coef_y = coef_lut_f_y_legacy;
+		coef_c = coef_lut_f_c_legacy;
+		shift_y = SHIFT_LUT_F_Y_LEGACY;
+		shift_c = SHIFT_LUT_F_C_LEGACY;
+	} else if (scale < SCALE_MAX_FOR_LEG_LUT_E) {
+		coef_y = coef_lut_e_y_legacy;
+		coef_c = coef_lut_e_c_legacy;
+		shift_y = SHIFT_LUT_E_Y_LEGACY;
+		shift_c = SHIFT_LUT_E_C_LEGACY;
+	} else if (scale < SCALE_MAX_FOR_LEG_LUT_D) {
+		coef_y = coef_lut_d_y_legacy;
+		coef_c = coef_lut_d_c_legacy;
+		shift_y = SHIFT_LUT_D_Y_LEGACY;
+		shift_c = SHIFT_LUT_D_C_LEGACY;
+	} else if (scale < SCALE_MAX_FOR_LEG_LUT_C) {
+		coef_y = coef_lut_c_y_legacy;
+		coef_c = coef_lut_c_c_legacy;
+		shift_y = SHIFT_LUT_C_Y_LEGACY;
+		shift_c = SHIFT_LUT_C_C_LEGACY;
+	} else if (scale == SCALE_MAX_FOR_LEG_LUT_C) {
+		coef_y = coef_c = coef_lut_b;
+		shift_y = shift_c = SHIFT_LUT_B;
+	} else {
+		coef_y = coef_c = coef_lut_a_legacy;
+		shift_y = shift_c = SHIFT_LUT_A_LEGACY;
+	}
+
+	if (orient == HVSRC_HORI) {
+		hvsrc->hori_shift = (shift_c << 16) | shift_y;
+		memcpy(hvsrc->yh_coef, coef_y, sizeof(hvsrc->yh_coef));
+		memcpy(hvsrc->ch_coef, coef_c, sizeof(hvsrc->ch_coef));
+	} else {
+		hvsrc->vert_shift = (shift_c << 16) | shift_y;
+		memcpy(hvsrc->yv_coef, coef_y, sizeof(hvsrc->yv_coef));
+		memcpy(hvsrc->cv_coef, coef_c, sizeof(hvsrc->cv_coef));
+	}
+}
+
+/**
+ * sti_hqvdp_check_hw_scaling
+ * @layer: hqvdp layer
+ *
+ * Check if the HW is able to perform the scaling request
+ * The firmware scaling limitation is "CEIL(1/Zy) <= FLOOR(LFW)" where:
+ *   Zy = OutputHeight / InputHeight
+ *   LFW = (Tx * IPClock) / (MaxNbCycles * Cp)
+ *     Tx : Total video mode horizontal resolution
+ *     IPClock : HQVDP IP clock (Mhz)
+ *     MaxNbCycles: max(InputWidth, OutputWidth)
+ *     Cp: Video mode pixel clock (Mhz)
+ *
+ * RETURNS:
+ * True if the HW can scale.
+ */
+static bool sti_hqvdp_check_hw_scaling(struct sti_layer *layer)
+{
+	struct sti_hqvdp *hqvdp = to_sti_hqvdp(layer);
+	unsigned long lfw;
+	unsigned int inv_zy;
+
+	lfw = layer->mode->htotal * (clk_get_rate(hqvdp->clk) / 1000000);
+	lfw /= max(layer->src_w, layer->dst_w) * layer->mode->clock / 1000;
+
+	inv_zy = DIV_ROUND_UP(layer->src_h, layer->dst_h);
+
+	return (inv_zy <= lfw) ? true : false;
+}
+
+/**
+ * sti_hqvdp_prepare_layer
+ * @layer: hqvdp layer
+ * @first_prepare: true if it is the first time this function is called
+ *
+ * Prepares a command for the firmware
+ *
+ * RETURNS:
+ * 0 on success.
+ */
+static int sti_hqvdp_prepare_layer(struct sti_layer *layer, bool first_prepare)
+{
+	struct sti_hqvdp *hqvdp = to_sti_hqvdp(layer);
+	struct sti_hqvdp_cmd *cmd;
+	int scale_h, scale_v;
+	int cmd_offset;
+
+	dev_dbg(hqvdp->dev, "%s %s\n", __func__, sti_layer_to_str(layer));
+
+	/* prepare and commit VID plane */
+	hqvdp->vid_plane->funcs->update_plane(hqvdp->vid_plane,
+					layer->crtc, layer->fb,
+					layer->dst_x, layer->dst_y,
+					layer->dst_w, layer->dst_h,
+					layer->src_x, layer->src_y,
+					layer->src_w, layer->src_h);
+
+	cmd_offset = sti_hqvdp_get_free_cmd(hqvdp);
+	if (cmd_offset == -1) {
+		DRM_ERROR("No available hqvdp_cmd now\n");
+		return -EBUSY;
+	}
+	cmd = hqvdp->hqvdp_cmd + cmd_offset;
+
+	if (!sti_hqvdp_check_hw_scaling(layer)) {
+		DRM_ERROR("Scaling beyond HW capabilities\n");
+		return -EINVAL;
+	}
+
+	/* Static parameters, defaulting to progressive mode */
+	cmd->top.config = TOP_CONFIG_PROGRESSIVE;
+	cmd->top.mem_format = TOP_MEM_FORMAT_DFLT;
+	cmd->hvsrc.param_ctrl = HVSRC_PARAM_CTRL_DFLT;
+	cmd->csdi.config = CSDI_CONFIG_PROG;
+
+	/* VC1RE, FMD bypassed : keep everything set to 0
+	 * IQI/P2I bypassed */
+	cmd->iqi.config = IQI_CONFIG_DFLT;
+	cmd->iqi.con_bri = IQI_CON_BRI_DFLT;
+	cmd->iqi.sat_gain = IQI_SAT_GAIN_DFLT;
+	cmd->iqi.pxf_conf = IQI_PXF_CONF_DFLT;
+
+	/* Buffer planes address */
+	cmd->top.current_luma = (u32) layer->paddr + layer->offsets[0];
+	cmd->top.current_chroma = (u32) layer->paddr + layer->offsets[1];
+
+	/* Pitches */
+	cmd->top.luma_processed_pitch = cmd->top.luma_src_pitch =
+			layer->pitches[0];
+	cmd->top.chroma_processed_pitch = cmd->top.chroma_src_pitch =
+			layer->pitches[1];
+
+	/* Input / output size
+	 * Align to upper even value */
+	layer->dst_w = ALIGN(layer->dst_w, 2);
+	layer->dst_h = ALIGN(layer->dst_h, 2);
+
+	if ((layer->src_w > MAX_WIDTH) || (layer->src_w < MIN_WIDTH) ||
+	    (layer->src_h > MAX_HEIGHT) || (layer->src_h < MIN_HEIGHT) ||
+	    (layer->dst_w > MAX_WIDTH) || (layer->dst_w < MIN_WIDTH) ||
+	    (layer->dst_h > MAX_HEIGHT) || (layer->dst_h < MIN_HEIGHT)) {
+		DRM_ERROR("Invalid in/out size %dx%d -> %dx%d\n",
+				layer->src_w, layer->src_h,
+				layer->dst_w, layer->dst_h);
+		return -EINVAL;
+	}
+	cmd->top.input_viewport_size = cmd->top.input_frame_size =
+			layer->src_h << 16 | layer->src_w;
+	cmd->hvsrc.output_picture_size = layer->dst_h << 16 | layer->dst_w;
+	cmd->top.input_viewport_ori = layer->src_y << 16 | layer->src_x;
+
+	/* Handle interlaced */
+	if (layer->fb->flags & DRM_MODE_FB_INTERLACED) {
+		/* Top field to display */
+		cmd->top.config = TOP_CONFIG_INTER_TOP;
+
+		/* Update pitches and vert size */
+		cmd->top.input_frame_size = (layer->src_h / 2) << 16 |
+					     layer->src_w;
+		cmd->top.luma_processed_pitch *= 2;
+		cmd->top.luma_src_pitch *= 2;
+		cmd->top.chroma_processed_pitch *= 2;
+		cmd->top.chroma_src_pitch *= 2;
+
+		/* Enable directional deinterlacing processing */
+		cmd->csdi.config = CSDI_CONFIG_INTER_DIR;
+		cmd->csdi.config2 = CSDI_CONFIG2_DFLT;
+		cmd->csdi.dcdi_config = CSDI_DCDI_CONFIG_DFLT;
+	}
+
+	/* Update hvsrc lut coef */
+	scale_h = SCALE_FACTOR * layer->dst_w / layer->src_w;
+	sti_hqvdp_update_hvsrc(HVSRC_HORI, scale_h, &cmd->hvsrc);
+
+	scale_v = SCALE_FACTOR * layer->dst_h / layer->src_h;
+	sti_hqvdp_update_hvsrc(HVSRC_VERT, scale_v, &cmd->hvsrc);
+
+	if (first_prepare) {
+		/* Prevent VTG shutdown */
+		if (clk_prepare_enable(hqvdp->clk_pix_main)) {
+			DRM_ERROR("Failed to prepare/enable pix main clk\n");
+			return -ENXIO;
+		}
+
+		/* Register VTG Vsync callback to handle bottom fields */
+		if ((layer->fb->flags & DRM_MODE_FB_INTERLACED) &&
+				sti_vtg_register_client(hqvdp->vtg,
+					&hqvdp->vtg_nb, layer->mixer_id)) {
+			DRM_ERROR("Cannot register VTG notifier\n");
+			return -ENXIO;
+		}
+	}
+
+	return 0;
+}
+
+static int sti_hqvdp_commit_layer(struct sti_layer *layer)
+{
+	struct sti_hqvdp *hqvdp = to_sti_hqvdp(layer);
+	int cmd_offset;
+
+	dev_dbg(hqvdp->dev, "%s %s\n", __func__, sti_layer_to_str(layer));
+
+	cmd_offset = sti_hqvdp_get_free_cmd(hqvdp);
+	if (cmd_offset == -1) {
+		DRM_ERROR("No available hqvdp_cmd now\n");
+		return -EBUSY;
+	}
+
+	writel(hqvdp->hqvdp_cmd_paddr + cmd_offset,
+			hqvdp->regs + HQVDP_MBX_NEXT_CMD);
+
+	hqvdp->curr_field_count++;
+
+	/* Interlaced : get ready to display the bottom field at next Vsync */
+	if (layer->fb->flags & DRM_MODE_FB_INTERLACED)
+		hqvdp->btm_field_pending = true;
+
+	dev_dbg(hqvdp->dev, "%s Posted command:0x%x\n",
+			__func__, hqvdp->hqvdp_cmd_paddr + cmd_offset);
+
+	return 0;
+}
+
+static int sti_hqvdp_disable_layer(struct sti_layer *layer)
+{
+	struct sti_hqvdp *hqvdp = to_sti_hqvdp(layer);
+	int i;
+
+	DRM_DEBUG_DRIVER("%s\n", sti_layer_to_str(layer));
+
+	/* Unregister VTG Vsync callback */
+	if ((layer->fb->flags & DRM_MODE_FB_INTERLACED) &&
+		sti_vtg_unregister_client(hqvdp->vtg, &hqvdp->vtg_nb))
+		DRM_DEBUG_DRIVER("Warning: cannot unregister VTG notifier\n");
+
+	/* Set next cmd to NULL */
+	writel(0, hqvdp->regs + HQVDP_MBX_NEXT_CMD);
+
+	for (i = 0; i < POLL_MAX_ATTEMPT; i++) {
+		if (readl(hqvdp->regs + HQVDP_MBX_INFO_XP70)
+				& INFO_XP70_FW_READY)
+			break;
+		msleep(POLL_DELAY_MS);
+	}
+
+	/* VTG can stop now */
+	clk_disable_unprepare(hqvdp->clk_pix_main);
+
+	if (i == POLL_MAX_ATTEMPT) {
+		DRM_ERROR("XP70 could not revert to idle\n");
+		return -ENXIO;
+	}
+
+	/* disable VID plane */
+	hqvdp->vid_plane->funcs->disable_plane(hqvdp->vid_plane);
+
+	return 0;
+}
+
+/**
+ * sti_vdp_vtg_cb
+ * @nb: notifier block
+ * @evt: event message
+ * @data: private data
+ *
+ * Handle VTG Vsync event, display pending bottom field
+ *
+ * RETURNS:
+ * 0 on success.
+ */
+int sti_hqvdp_vtg_cb(struct notifier_block *nb, unsigned long evt, void *data)
+{
+	struct sti_hqvdp *hqvdp = container_of(nb, struct sti_hqvdp, vtg_nb);
+	int btm_cmd_offset, top_cmd_offest;
+	struct sti_hqvdp_cmd *btm_cmd, *top_cmd;
+
+	if ((evt != VTG_TOP_FIELD_EVENT) && (evt != VTG_BOTTOM_FIELD_EVENT)) {
+		DRM_DEBUG_DRIVER("Unknown event\n");
+		return 0;
+	}
+
+	if (hqvdp->btm_field_pending) {
+		/* Create the btm field command from the current one */
+		btm_cmd_offset = sti_hqvdp_get_free_cmd(hqvdp);
+		top_cmd_offest = sti_hqvdp_get_curr_cmd(hqvdp);
+		if ((btm_cmd_offset == -1) || (top_cmd_offest == -1)) {
+			DRM_ERROR("Cannot get cmds, skip btm field\n");
+			return -EBUSY;
+		}
+
+		btm_cmd = hqvdp->hqvdp_cmd + btm_cmd_offset;
+		top_cmd = hqvdp->hqvdp_cmd + top_cmd_offest;
+
+		memcpy(btm_cmd, top_cmd, sizeof(*btm_cmd));
+
+		btm_cmd->top.config = TOP_CONFIG_INTER_BTM;
+		btm_cmd->top.current_luma +=
+				btm_cmd->top.luma_src_pitch / 2;
+		btm_cmd->top.current_chroma +=
+				btm_cmd->top.chroma_src_pitch / 2;
+
+		/* Post the command to mailbox */
+		writel(hqvdp->hqvdp_cmd_paddr + btm_cmd_offset,
+				hqvdp->regs + HQVDP_MBX_NEXT_CMD);
+
+		hqvdp->curr_field_count++;
+		hqvdp->btm_field_pending = false;
+
+		dev_dbg(hqvdp->dev, "%s Posted command:0x%x\n",
+				__func__, hqvdp->hqvdp_cmd_paddr);
+	}
+
+	return 0;
+}
+
+static struct drm_plane *sti_hqvdp_find_vid(struct drm_device *dev, int id)
+{
+	struct drm_plane *plane;
+
+	list_for_each_entry(plane, &dev->mode_config.plane_list, head) {
+		struct sti_layer *layer = to_sti_layer(plane);
+
+		if (layer->desc == id)
+			return plane;
+	}
+
+	return NULL;
+}
+
+static void sti_hqvd_init(struct sti_layer *layer)
+{
+	struct sti_hqvdp *hqvdp = to_sti_hqvdp(layer);
+	int size;
+
+	/* find the plane macthing with vid 0 */
+	hqvdp->vid_plane = sti_hqvdp_find_vid(hqvdp->drm_dev, STI_VID_0);
+	if (!hqvdp->vid_plane) {
+		DRM_ERROR("Cannot find Main video layer\n");
+		return;
+	}
+
+	hqvdp->vtg_nb.notifier_call = sti_hqvdp_vtg_cb;
+
+	/* Allocate memory for the VDP commands */
+	size = NB_VDP_CMD * sizeof(struct sti_hqvdp_cmd);
+	hqvdp->hqvdp_cmd = dma_alloc_writecombine(hqvdp->dev, size,
+					 &hqvdp->hqvdp_cmd_paddr,
+					 GFP_KERNEL | GFP_DMA);
+	if (!hqvdp->hqvdp_cmd) {
+		DRM_ERROR("Failed to allocate memory for VDP cmd\n");
+		return;
+	}
+
+	memset(hqvdp->hqvdp_cmd, 0, size);
+}
+
+static const struct sti_layer_funcs hqvdp_ops = {
+	.get_formats = sti_hqvdp_get_formats,
+	.get_nb_formats = sti_hqvdp_get_nb_formats,
+	.init = sti_hqvd_init,
+	.prepare = sti_hqvdp_prepare_layer,
+	.commit = sti_hqvdp_commit_layer,
+	.disable = sti_hqvdp_disable_layer,
+};
+
+struct sti_layer *sti_hqvdp_create(struct device *dev)
+{
+	struct sti_hqvdp *hqvdp = dev_get_drvdata(dev);
+
+	hqvdp->layer.ops = &hqvdp_ops;
+
+	return &hqvdp->layer;
+}
+EXPORT_SYMBOL(sti_hqvdp_create);
+
+static void sti_hqvdp_init_plugs(struct sti_hqvdp *hqvdp)
+{
+	/* Configure Plugs (same for RD & WR) */
+	writel(PLUG_PAGE_SIZE_256, hqvdp->regs + HQVDP_RD_PLUG_PAGE_SIZE);
+	writel(PLUG_MIN_OPC_8, hqvdp->regs + HQVDP_RD_PLUG_MIN_OPC);
+	writel(PLUG_MAX_OPC_64, hqvdp->regs + HQVDP_RD_PLUG_MAX_OPC);
+	writel(PLUG_MAX_CHK_2X, hqvdp->regs + HQVDP_RD_PLUG_MAX_CHK);
+	writel(PLUG_MAX_MSG_1X, hqvdp->regs + HQVDP_RD_PLUG_MAX_MSG);
+	writel(PLUG_MIN_SPACE_1, hqvdp->regs + HQVDP_RD_PLUG_MIN_SPACE);
+	writel(PLUG_CONTROL_ENABLE, hqvdp->regs + HQVDP_RD_PLUG_CONTROL);
+
+	writel(PLUG_PAGE_SIZE_256, hqvdp->regs + HQVDP_WR_PLUG_PAGE_SIZE);
+	writel(PLUG_MIN_OPC_8, hqvdp->regs + HQVDP_WR_PLUG_MIN_OPC);
+	writel(PLUG_MAX_OPC_64, hqvdp->regs + HQVDP_WR_PLUG_MAX_OPC);
+	writel(PLUG_MAX_CHK_2X, hqvdp->regs + HQVDP_WR_PLUG_MAX_CHK);
+	writel(PLUG_MAX_MSG_1X, hqvdp->regs + HQVDP_WR_PLUG_MAX_MSG);
+	writel(PLUG_MIN_SPACE_1, hqvdp->regs + HQVDP_WR_PLUG_MIN_SPACE);
+	writel(PLUG_CONTROL_ENABLE, hqvdp->regs + HQVDP_WR_PLUG_CONTROL);
+}
+
+/**
+ * sti_hqvdp_start_xp70
+ * @firmware: firmware found
+ * @ctxt:     hqvdp structure
+ *
+ * Run the xP70 initialization sequence
+ */
+static void sti_hqvdp_start_xp70(const struct firmware *firmware, void *ctxt)
+{
+	struct sti_hqvdp *hqvdp = ctxt;
+	u32 *fw_rd_plug, *fw_wr_plug, *fw_pmem, *fw_dmem;
+	u8 *data;
+	int i;
+	struct fw_header {
+		int rd_size;
+		int wr_size;
+		int pmem_size;
+		int dmem_size;
+	} *header;
+
+	DRM_DEBUG_DRIVER("\n");
+	/* Check firmware parts */
+	if (!firmware) {
+		DRM_ERROR("Firmware not available\n");
+		return;
+	}
+
+	header = (struct fw_header *) firmware->data;
+	if (firmware->size < sizeof(*header)) {
+		DRM_ERROR("Invalid firmware size (%d)\n", firmware->size);
+		goto out;
+	}
+	if ((sizeof(*header) + header->rd_size + header->wr_size +
+		header->pmem_size + header->dmem_size) != firmware->size) {
+		DRM_ERROR("Invalid fmw structure (%d+%d+%d+%d+%d != %d)\n",
+			   sizeof(*header), header->rd_size, header->wr_size,
+			   header->pmem_size, header->dmem_size,
+			   firmware->size);
+		goto out;
+	}
+
+	data = (u8 *) firmware->data;
+	data += sizeof(*header);
+	fw_rd_plug = (void *) data;
+	data += header->rd_size;
+	fw_wr_plug = (void *) data;
+	data += header->wr_size;
+	fw_pmem = (void *) data;
+	data += header->pmem_size;
+	fw_dmem = (void *) data;
+
+	/* Enable clock */
+	if (clk_prepare_enable(hqvdp->clk))
+		DRM_ERROR("Failed to prepare/enable HQVDP clk\n");
+
+	/* Reset */
+	writel(SW_RESET_CTRL_FULL, hqvdp->regs + HQVDP_MBX_SW_RESET_CTRL);
+
+	for (i = 0; i < POLL_MAX_ATTEMPT; i++) {
+		if (readl(hqvdp->regs + HQVDP_MBX_STARTUP_CTRL1)
+				& STARTUP_CTRL1_RST_DONE)
+			break;
+		msleep(POLL_DELAY_MS);
+	}
+	if (i == POLL_MAX_ATTEMPT) {
+		DRM_ERROR("Could not reset\n");
+		goto out;
+	}
+
+	/* Init Read & Write plugs */
+	for (i = 0; i < header->rd_size / 4; i++)
+		writel(fw_rd_plug[i], hqvdp->regs + HQVDP_RD_PLUG + i * 4);
+	for (i = 0; i < header->wr_size / 4; i++)
+		writel(fw_wr_plug[i], hqvdp->regs + HQVDP_WR_PLUG + i * 4);
+
+	sti_hqvdp_init_plugs(hqvdp);
+
+	/* Authorize Idle Mode */
+	writel(STARTUP_CTRL1_AUTH_IDLE, hqvdp->regs + HQVDP_MBX_STARTUP_CTRL1);
+
+	/* Prevent VTG interruption during the boot */
+	writel(SOFT_VSYNC_SW_CTRL_IRQ, hqvdp->regs + HQVDP_MBX_SOFT_VSYNC);
+	writel(0, hqvdp->regs + HQVDP_MBX_NEXT_CMD);
+
+	/* Download PMEM & DMEM */
+	for (i = 0; i < header->pmem_size / 4; i++)
+		writel(fw_pmem[i], hqvdp->regs + HQVDP_PMEM + i * 4);
+	for (i = 0; i < header->dmem_size / 4; i++)
+		writel(fw_dmem[i], hqvdp->regs + HQVDP_DMEM + i * 4);
+
+	/* Enable fetch */
+	writel(STARTUP_CTRL2_FETCH_EN, hqvdp->regs + HQVDP_MBX_STARTUP_CTRL2);
+
+	/* Wait end of boot */
+	for (i = 0; i < POLL_MAX_ATTEMPT; i++) {
+		if (readl(hqvdp->regs + HQVDP_MBX_INFO_XP70)
+				& INFO_XP70_FW_READY)
+			break;
+		msleep(POLL_DELAY_MS);
+	}
+	if (i == POLL_MAX_ATTEMPT) {
+		DRM_ERROR("Could not boot\n");
+		goto out;
+	}
+
+	/* Launch Vsync */
+	writel(SOFT_VSYNC_HW, hqvdp->regs + HQVDP_MBX_SOFT_VSYNC);
+
+	DRM_INFO("HQVDP XP70 started\n");
+out:
+	release_firmware(firmware);
+}
+
+int sti_hqvdp_bind(struct device *dev, struct device *master, void *data)
+{
+	struct sti_hqvdp *hqvdp = dev_get_drvdata(dev);
+	struct drm_device *drm_dev = data;
+	struct sti_layer *layer;
+	int err;
+
+	DRM_DEBUG_DRIVER("\n");
+
+	hqvdp->drm_dev = drm_dev;
+
+	/* Request for firmware */
+	err = request_firmware_nowait(THIS_MODULE, FW_ACTION_HOTPLUG,
+				HQVDP_FMW_NAME,	hqvdp->dev,
+				GFP_KERNEL, hqvdp, sti_hqvdp_start_xp70);
+	if (err) {
+		DRM_ERROR("Can't get HQVDP firmware\n");
+		return err;
+	}
+
+	layer = sti_layer_create(hqvdp->dev, STI_HQVDP_0, hqvdp->regs);
+	if (!layer) {
+		DRM_ERROR("Can't create HQVDP plane\n");
+		return -ENOMEM;
+	}
+
+	sti_drm_plane_init(drm_dev, layer, 1, DRM_PLANE_TYPE_OVERLAY);
+
+	return 0;
+}
+
+static void sti_hqvdp_unbind(struct device *dev,
+		struct device *master, void *data)
+{
+	/* do nothing */
+}
+
+static const struct component_ops sti_hqvdp_ops = {
+	.bind = sti_hqvdp_bind,
+	.unbind = sti_hqvdp_unbind,
+};
+
+static int sti_hqvdp_probe(struct platform_device *pdev)
+{
+	struct device *dev = &pdev->dev;
+	struct device_node *vtg_np;
+	struct sti_hqvdp *hqvdp;
+	struct resource *res;
+
+	DRM_DEBUG_DRIVER("\n");
+
+	hqvdp = devm_kzalloc(dev, sizeof(*hqvdp), GFP_KERNEL);
+	if (!hqvdp) {
+		DRM_ERROR("Failed to allocate HQVDP context\n");
+		return -ENOMEM;
+	}
+
+	hqvdp->dev = dev;
+
+	/* Get Memory resources */
+	res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
+	if (res == NULL) {
+		DRM_ERROR("Get memory resource failed\n");
+		return -ENXIO;
+	}
+	hqvdp->regs = devm_ioremap(dev, res->start, resource_size(res));
+	if (hqvdp->regs == NULL) {
+		DRM_ERROR("Register mapping failed\n");
+		return -ENXIO;
+	}
+
+	/* Get clock resources */
+	hqvdp->clk = devm_clk_get(dev, "hqvdp");
+	hqvdp->clk_pix_main = devm_clk_get(dev, "pix_main");
+	if (IS_ERR(hqvdp->clk) || IS_ERR(hqvdp->clk)) {
+		DRM_ERROR("Cannot get clocks\n");
+		return -ENXIO;
+	}
+
+	/* Get reset resources */
+	hqvdp->reset = devm_reset_control_get(dev, "hqvdp");
+	if (!IS_ERR(hqvdp->reset))
+		reset_control_deassert(hqvdp->reset);
+
+	vtg_np = of_parse_phandle(pdev->dev.of_node, "st,vtg", 0);
+	if (vtg_np)
+		hqvdp->vtg = of_vtg_find(vtg_np);
+
+	platform_set_drvdata(pdev, hqvdp);
+
+	return component_add(&pdev->dev, &sti_hqvdp_ops);
+}
+
+static int sti_hqvdp_remove(struct platform_device *pdev)
+{
+	component_del(&pdev->dev, &sti_hqvdp_ops);
+	return 0;
+}
+
+static struct of_device_id hqvdp_of_match[] = {
+	{ .compatible = "st,stih407-hqvdp", },
+	{ /* end node */ }
+};
+MODULE_DEVICE_TABLE(of, hqvdp_of_match);
+
+struct platform_driver sti_hqvdp_driver = {
+	.driver = {
+		.name = "sti-hqvdp",
+		.owner = THIS_MODULE,
+		.of_match_table = hqvdp_of_match,
+	},
+	.probe = sti_hqvdp_probe,
+	.remove = sti_hqvdp_remove,
+};
+
+module_platform_driver(sti_hqvdp_driver);
+
+MODULE_AUTHOR("Benjamin Gaignard <benjamin.gaignard@st.com>");
+MODULE_DESCRIPTION("STMicroelectronics SoC DRM driver");
+MODULE_LICENSE("GPL");
diff --git a/drivers/gpu/drm/sti/sti_hqvdp.h b/drivers/gpu/drm/sti/sti_hqvdp.h
new file mode 100644
index 000000000000..cd5ecd0a6dea
--- /dev/null
+++ b/drivers/gpu/drm/sti/sti_hqvdp.h
@@ -0,0 +1,12 @@
+/*
+ * Copyright (C) STMicroelectronics SA 2014
+ * Authors: Fabien Dessenne <fabien.dessenne@st.com> for STMicroelectronics.
+ * License terms:  GNU General Public License (GPL), version 2
+ */
+
+#ifndef _STI_HQVDP_H_
+#define _STI_HQVDP_H_
+
+struct sti_layer *sti_hqvdp_create(struct device *dev);
+
+#endif
diff --git a/drivers/gpu/drm/sti/sti_hqvdp_lut.h b/drivers/gpu/drm/sti/sti_hqvdp_lut.h
new file mode 100644
index 000000000000..619af7f4384e
--- /dev/null
+++ b/drivers/gpu/drm/sti/sti_hqvdp_lut.h
@@ -0,0 +1,373 @@
+/*
+ * Copyright (C) STMicroelectronics SA 2014
+ * Authors: Fabien Dessenne <fabien.dessenne@st.com> for STMicroelectronics.
+ * License terms:  GNU General Public License (GPL), version 2
+ */
+
+#ifndef _STI_HQVDP_LUT_H_
+#define _STI_HQVDP_LUT_H_
+
+#define NB_COEF                 128
+
+#define SHIFT_LUT_A_LEGACY      8
+#define SHIFT_LUT_B             8
+#define SHIFT_LUT_C_Y_LEGACY    8
+#define SHIFT_LUT_C_C_LEGACY    8
+#define SHIFT_LUT_D_Y_LEGACY    8
+#define SHIFT_LUT_D_C_LEGACY    8
+#define SHIFT_LUT_E_Y_LEGACY    8
+#define SHIFT_LUT_E_C_LEGACY    8
+#define SHIFT_LUT_F_Y_LEGACY    8
+#define SHIFT_LUT_F_C_LEGACY    8
+
+static const u32 coef_lut_a_legacy[NB_COEF] = {
+	0x0000ffff, 0x00010000, 0x000100ff, 0x00000000,
+	0x00000000, 0x00050000, 0xfffc00ff, 0x00000000,
+	0x00000000, 0x00090000, 0xfff900fe, 0x00000000,
+	0x00000000, 0x0010ffff, 0xfff600fb, 0x00000000,
+	0x00000000, 0x0017fffe, 0xfff400f7, 0x00000000,
+	0x00000000, 0x001ffffd, 0xfff200f2, 0x00000000,
+	0x00000000, 0x0027fffc, 0xfff100ec, 0x00000000,
+	0x00000000, 0x0030fffb, 0xfff000e5, 0x00000000,
+	0x00000000, 0x003afffa, 0xffee00de, 0x00000000,
+	0x00000000, 0x0044fff9, 0xffed00d6, 0x00000000,
+	0x00000000, 0x004efff8, 0xffed00cd, 0x00000000,
+	0x00000000, 0x0059fff6, 0xffed00c4, 0x00000000,
+	0x00000000, 0x0064fff5, 0xffed00ba, 0x00000000,
+	0x00000000, 0x006ffff3, 0xffee00b0, 0x00000000,
+	0x00000000, 0x007afff2, 0xffee00a6, 0x00000000,
+	0x00000000, 0x0085fff1, 0xffef009b, 0x00000000,
+	0x00000000, 0x0090fff0, 0xfff00090, 0x00000000,
+	0x00000000, 0x009bffef, 0xfff10085, 0x00000000,
+	0x00000000, 0x00a6ffee, 0xfff2007a, 0x00000000,
+	0x00000000, 0x00b0ffee, 0xfff3006f, 0x00000000,
+	0x00000000, 0x00baffed, 0xfff50064, 0x00000000,
+	0x00000000, 0x00c4ffed, 0xfff60059, 0x00000000,
+	0x00000000, 0x00cdffed, 0xfff8004e, 0x00000000,
+	0x00000000, 0x00d6ffed, 0xfff90044, 0x00000000,
+	0x00000000, 0x00deffee, 0xfffa003a, 0x00000000,
+	0x00000000, 0x00e5fff0, 0xfffb0030, 0x00000000,
+	0x00000000, 0x00ecfff1, 0xfffc0027, 0x00000000,
+	0x00000000, 0x00f2fff2, 0xfffd001f, 0x00000000,
+	0x00000000, 0x00f7fff4, 0xfffe0017, 0x00000000,
+	0x00000000, 0x00fbfff6, 0xffff0010, 0x00000000,
+	0x00000000, 0x00fefff9, 0x00000009, 0x00000000,
+	0x00000000, 0x00fffffc, 0x00000005, 0x00000000
+};
+
+static const u32 coef_lut_b[NB_COEF] = {
+	0x00000000, 0x00000000, 0x00000100, 0x00000000,
+	0x00000000, 0x00000000, 0x00000100, 0x00000000,
+	0x00000000, 0x00000000, 0x00000100, 0x00000000,
+	0x00000000, 0x00000000, 0x00000100, 0x00000000,
+	0x00000000, 0x00000000, 0x00000100, 0x00000000,
+	0x00000000, 0x00000000, 0x00000100, 0x00000000,
+	0x00000000, 0x00000000, 0x00000100, 0x00000000,
+	0x00000000, 0x00000000, 0x00000100, 0x00000000,
+	0x00000000, 0x00000000, 0x00000100, 0x00000000,
+	0x00000000, 0x00000000, 0x00000100, 0x00000000,
+	0x00000000, 0x00000000, 0x00000100, 0x00000000,
+	0x00000000, 0x00000000, 0x00000100, 0x00000000,
+	0x00000000, 0x00000000, 0x00000100, 0x00000000,
+	0x00000000, 0x00000000, 0x00000100, 0x00000000,
+	0x00000000, 0x00000000, 0x00000100, 0x00000000,
+	0x00000000, 0x00000000, 0x00000100, 0x00000000,
+	0x00000000, 0x00000000, 0x00000100, 0x00000000,
+	0x00000000, 0x00000000, 0x00000100, 0x00000000,
+	0x00000000, 0x00000000, 0x00000100, 0x00000000,
+	0x00000000, 0x00000000, 0x00000100, 0x00000000,
+	0x00000000, 0x00000000, 0x00000100, 0x00000000,
+	0x00000000, 0x00000000, 0x00000100, 0x00000000,
+	0x00000000, 0x00000000, 0x00000100, 0x00000000,
+	0x00000000, 0x00000000, 0x00000100, 0x00000000,
+	0x00000000, 0x00000000, 0x00000100, 0x00000000,
+	0x00000000, 0x00000000, 0x00000100, 0x00000000,
+	0x00000000, 0x00000000, 0x00000100, 0x00000000,
+	0x00000000, 0x00000000, 0x00000100, 0x00000000,
+	0x00000000, 0x00000000, 0x00000100, 0x00000000,
+	0x00000000, 0x00000000, 0x00000100, 0x00000000,
+	0x00000000, 0x00000000, 0x00000100, 0x00000000,
+	0x00000000, 0x00000000, 0x00000100, 0x00000000
+};
+
+static const u32 coef_lut_c_y_legacy[NB_COEF] = {
+	0x00060004, 0x0038ffe1, 0x003800be, 0x0006ffe1,
+	0x00050005, 0x0042ffe1, 0x003800b3, 0x0007ffe1,
+	0x00040006, 0x0046ffe1, 0x003300b2, 0x0008ffe2,
+	0x00030007, 0x004cffe1, 0x002e00b1, 0x0008ffe2,
+	0x00020006, 0x0051ffe2, 0x002900b0, 0x0009ffe3,
+	0x00010008, 0x0056ffe2, 0x002400ae, 0x0009ffe4,
+	0xffff0008, 0x005cffe3, 0x001f00ad, 0x000affe4,
+	0xfffe0008, 0x0062ffe4, 0x001a00ab, 0x000affe5,
+	0xfffd000a, 0x0066ffe5, 0x001500a8, 0x000bffe6,
+	0xfffc0009, 0x006bffe7, 0x001100a5, 0x000bffe8,
+	0xfffa000a, 0x0070ffe8, 0x000d00a3, 0x000bffe9,
+	0xfff9000b, 0x0076ffea, 0x0008009f, 0x000bffea,
+	0xfff7000b, 0x007affec, 0x0005009b, 0x000cffec,
+	0xfff6000b, 0x007effef, 0x00010098, 0x000cffed,
+	0xfff4000b, 0x0084fff1, 0xfffd0095, 0x000cffee,
+	0xfff3000b, 0x0088fff4, 0xfffa0090, 0x000cfff0,
+	0xfff1000b, 0x008dfff7, 0xfff7008d, 0x000bfff1,
+	0xfff0000c, 0x0090fffa, 0xfff40088, 0x000bfff3,
+	0xffee000c, 0x0095fffd, 0xfff10084, 0x000bfff4,
+	0xffed000c, 0x00980001, 0xffef007e, 0x000bfff6,
+	0xffec000c, 0x009b0005, 0xffec007a, 0x000bfff7,
+	0xffea000b, 0x009f0008, 0xffea0076, 0x000bfff9,
+	0xffe9000b, 0x00a3000d, 0xffe80070, 0x000afffa,
+	0xffe8000b, 0x00a50011, 0xffe7006b, 0x0009fffc,
+	0xffe6000b, 0x00a80015, 0xffe50066, 0x000afffd,
+	0xffe5000a, 0x00ab001a, 0xffe40062, 0x0008fffe,
+	0xffe4000a, 0x00ad001f, 0xffe3005c, 0x0008ffff,
+	0xffe40009, 0x00ae0024, 0xffe20056, 0x00080001,
+	0xffe30009, 0x00b00029, 0xffe20051, 0x00060002,
+	0xffe20008, 0x00b1002e, 0xffe1004c, 0x00070003,
+	0xffe20008, 0x00b20033, 0xffe10046, 0x00060004,
+	0xffe10007, 0x00b30038, 0xffe10042, 0x00050005
+};
+
+static const u32 coef_lut_c_c_legacy[NB_COEF] = {
+	0x0001fff3, 0x003afffb, 0x003a00a1, 0x0001fffb,
+	0x0001fff5, 0x0041fffb, 0x0038009a, 0x0001fffb,
+	0x0001fff5, 0x0046fffb, 0x00340099, 0x0001fffb,
+	0x0001fff7, 0x0049fffb, 0x00300098, 0x0001fffb,
+	0x0001fff9, 0x004cfffb, 0x002d0096, 0x0001fffb,
+	0x0001fffa, 0x004ffffc, 0x00290095, 0x0001fffb,
+	0x0001fff9, 0x0054fffd, 0x00250093, 0x0001fffc,
+	0x0001fffa, 0x0058fffd, 0x00220092, 0x0000fffc,
+	0x0001fffb, 0x005bfffe, 0x001f0090, 0x0000fffc,
+	0x0001fffd, 0x005effff, 0x001c008c, 0x0000fffd,
+	0x0001fffd, 0x00620000, 0x0019008a, 0x0000fffd,
+	0x0001fffe, 0x00660001, 0x00160088, 0xfffffffd,
+	0x0000fffe, 0x006a0003, 0x00130085, 0xfffffffe,
+	0x0000fffe, 0x006e0004, 0x00100083, 0xfffffffe,
+	0x0000fffe, 0x00710006, 0x000e007f, 0xffffffff,
+	0x0000fffe, 0x00750008, 0x000c007c, 0xfffeffff,
+	0xfffffffe, 0x0079000a, 0x000a0079, 0xfffeffff,
+	0xfffffffe, 0x007c000c, 0x00080075, 0xfffe0000,
+	0xffffffff, 0x007f000e, 0x00060071, 0xfffe0000,
+	0xfffeffff, 0x00830010, 0x0004006e, 0xfffe0000,
+	0xfffeffff, 0x00850013, 0x0003006a, 0xfffe0000,
+	0xfffdffff, 0x00880016, 0x00010066, 0xfffe0001,
+	0xfffd0000, 0x008a0019, 0x00000062, 0xfffd0001,
+	0xfffd0000, 0x008c001c, 0xffff005e, 0xfffd0001,
+	0xfffc0000, 0x0090001f, 0xfffe005b, 0xfffb0001,
+	0xfffc0000, 0x00920022, 0xfffd0058, 0xfffa0001,
+	0xfffc0001, 0x00930025, 0xfffd0054, 0xfff90001,
+	0xfffb0001, 0x00950029, 0xfffc004f, 0xfffa0001,
+	0xfffb0001, 0x0096002d, 0xfffb004c, 0xfff90001,
+	0xfffb0001, 0x00980030, 0xfffb0049, 0xfff70001,
+	0xfffb0001, 0x00990034, 0xfffb0046, 0xfff50001,
+	0xfffb0001, 0x009a0038, 0xfffb0041, 0xfff50001
+};
+
+static const u32 coef_lut_d_y_legacy[NB_COEF] = {
+	0xfff80009, 0x0046ffec, 0x004600a3, 0xfff8ffec,
+	0xfff70009, 0x004effed, 0x0044009d, 0xfff9ffeb,
+	0xfff6000a, 0x0052ffee, 0x003f009d, 0xfffaffea,
+	0xfff50009, 0x0057ffef, 0x003b009d, 0xfffbffe9,
+	0xfff50008, 0x005bfff0, 0x0037009c, 0xfffcffe9,
+	0xfff40008, 0x005ffff2, 0x0033009b, 0xfffcffe9,
+	0xfff30007, 0x0064fff3, 0x002f009b, 0xfffdffe8,
+	0xfff20007, 0x0068fff5, 0x002b0099, 0xfffeffe8,
+	0xfff10008, 0x006bfff7, 0x00270097, 0xffffffe8,
+	0xfff00007, 0x006ffff9, 0x00230097, 0xffffffe8,
+	0xffef0006, 0x0073fffb, 0x00200095, 0x0000ffe8,
+	0xffee0005, 0x0077fffe, 0x001c0093, 0x0000ffe9,
+	0xffee0005, 0x007a0000, 0x00180091, 0x0001ffe9,
+	0xffed0005, 0x007d0003, 0x0015008e, 0x0002ffe9,
+	0xffec0005, 0x00800006, 0x0012008b, 0x0002ffea,
+	0xffeb0004, 0x00840008, 0x000e008a, 0x0003ffea,
+	0xffeb0003, 0x0087000b, 0x000b0087, 0x0003ffeb,
+	0xffea0003, 0x008a000e, 0x00080084, 0x0004ffeb,
+	0xffea0002, 0x008b0012, 0x00060080, 0x0005ffec,
+	0xffe90002, 0x008e0015, 0x0003007d, 0x0005ffed,
+	0xffe90001, 0x00910018, 0x0000007a, 0x0005ffee,
+	0xffe90000, 0x0093001c, 0xfffe0077, 0x0005ffee,
+	0xffe80000, 0x00950020, 0xfffb0073, 0x0006ffef,
+	0xffe8ffff, 0x00970023, 0xfff9006f, 0x0007fff0,
+	0xffe8ffff, 0x00970027, 0xfff7006b, 0x0008fff1,
+	0xffe8fffe, 0x0099002b, 0xfff50068, 0x0007fff2,
+	0xffe8fffd, 0x009b002f, 0xfff30064, 0x0007fff3,
+	0xffe9fffc, 0x009b0033, 0xfff2005f, 0x0008fff4,
+	0xffe9fffc, 0x009c0037, 0xfff0005b, 0x0008fff5,
+	0xffe9fffb, 0x009d003b, 0xffef0057, 0x0009fff5,
+	0xffeafffa, 0x009d003f, 0xffee0052, 0x000afff6,
+	0xffebfff9, 0x009d0044, 0xffed004e, 0x0009fff7
+};
+
+static const u32 coef_lut_d_c_legacy[NB_COEF] = {
+	0xfffeffff, 0x003fffff, 0x003f0089, 0xfffeffff,
+	0xfffe0000, 0x00460000, 0x0042007d, 0xfffffffe,
+	0xfffe0000, 0x00490001, 0x003f007d, 0xfffffffd,
+	0xfffd0001, 0x004b0002, 0x003c007d, 0x0000fffc,
+	0xfffd0001, 0x004e0003, 0x0039007c, 0x0000fffc,
+	0xfffc0001, 0x00510005, 0x0036007c, 0x0000fffb,
+	0xfffc0001, 0x00540006, 0x0033007b, 0x0001fffa,
+	0xfffc0003, 0x00550008, 0x00310078, 0x0001fffa,
+	0xfffb0003, 0x00580009, 0x002e0078, 0x0001fffa,
+	0xfffb0002, 0x005b000b, 0x002b0077, 0x0002fff9,
+	0xfffa0003, 0x005e000d, 0x00280075, 0x0002fff9,
+	0xfffa0002, 0x0060000f, 0x00260074, 0x0002fff9,
+	0xfffa0004, 0x00610011, 0x00230072, 0x0002fff9,
+	0xfffa0004, 0x00640013, 0x00200070, 0x0002fff9,
+	0xfff90004, 0x00660015, 0x001e006e, 0x0003fff9,
+	0xfff90004, 0x00680017, 0x001c006c, 0x0003fff9,
+	0xfff90003, 0x006b0019, 0x0019006b, 0x0003fff9,
+	0xfff90003, 0x006c001c, 0x00170068, 0x0004fff9,
+	0xfff90003, 0x006e001e, 0x00150066, 0x0004fff9,
+	0xfff90002, 0x00700020, 0x00130064, 0x0004fffa,
+	0xfff90002, 0x00720023, 0x00110061, 0x0004fffa,
+	0xfff90002, 0x00740026, 0x000f0060, 0x0002fffa,
+	0xfff90002, 0x00750028, 0x000d005e, 0x0003fffa,
+	0xfff90002, 0x0077002b, 0x000b005b, 0x0002fffb,
+	0xfffa0001, 0x0078002e, 0x00090058, 0x0003fffb,
+	0xfffa0001, 0x00780031, 0x00080055, 0x0003fffc,
+	0xfffa0001, 0x007b0033, 0x00060054, 0x0001fffc,
+	0xfffb0000, 0x007c0036, 0x00050051, 0x0001fffc,
+	0xfffc0000, 0x007c0039, 0x0003004e, 0x0001fffd,
+	0xfffc0000, 0x007d003c, 0x0002004b, 0x0001fffd,
+	0xfffdffff, 0x007d003f, 0x00010049, 0x0000fffe,
+	0xfffeffff, 0x007d0042, 0x00000046, 0x0000fffe
+};
+
+static const u32 coef_lut_e_y_legacy[NB_COEF] = {
+	0xfff10001, 0x00490004, 0x00490083, 0xfff10004,
+	0xfff10000, 0x00500006, 0x004b007b, 0xfff10002,
+	0xfff10000, 0x00530007, 0x0048007b, 0xfff10001,
+	0xfff10000, 0x00550009, 0x0046007a, 0xfff10000,
+	0xfff1fffe, 0x0058000b, 0x0043007b, 0xfff2fffe,
+	0xfff1ffff, 0x005a000d, 0x0040007a, 0xfff2fffd,
+	0xfff1fffd, 0x005d000f, 0x003e007a, 0xfff2fffc,
+	0xfff1fffd, 0x005f0011, 0x003b0079, 0xfff3fffb,
+	0xfff1fffc, 0x00610013, 0x00390079, 0xfff3fffa,
+	0xfff1fffb, 0x00640015, 0x00360079, 0xfff3fff9,
+	0xfff1fffa, 0x00660017, 0x00340078, 0xfff4fff8,
+	0xfff1fffb, 0x00680019, 0x00310077, 0xfff4fff7,
+	0xfff2fff9, 0x006a001b, 0x002f0076, 0xfff5fff6,
+	0xfff2fff9, 0x006c001e, 0x002c0075, 0xfff5fff5,
+	0xfff2fff9, 0x006d0020, 0x002a0073, 0xfff6fff5,
+	0xfff3fff7, 0x00700022, 0x00270073, 0xfff6fff4,
+	0xfff3fff7, 0x00710025, 0x00250071, 0xfff7fff3,
+	0xfff4fff6, 0x00730027, 0x00220070, 0xfff7fff3,
+	0xfff5fff6, 0x0073002a, 0x0020006d, 0xfff9fff2,
+	0xfff5fff5, 0x0075002c, 0x001e006c, 0xfff9fff2,
+	0xfff6fff5, 0x0076002f, 0x001b006a, 0xfff9fff2,
+	0xfff7fff4, 0x00770031, 0x00190068, 0xfffbfff1,
+	0xfff8fff4, 0x00780034, 0x00170066, 0xfffafff1,
+	0xfff9fff3, 0x00790036, 0x00150064, 0xfffbfff1,
+	0xfffafff3, 0x00790039, 0x00130061, 0xfffcfff1,
+	0xfffbfff3, 0x0079003b, 0x0011005f, 0xfffdfff1,
+	0xfffcfff2, 0x007a003e, 0x000f005d, 0xfffdfff1,
+	0xfffdfff2, 0x007a0040, 0x000d005a, 0xfffffff1,
+	0xfffefff2, 0x007b0043, 0x000b0058, 0xfffefff1,
+	0x0000fff1, 0x007a0046, 0x00090055, 0x0000fff1,
+	0x0001fff1, 0x007b0048, 0x00070053, 0x0000fff1,
+	0x0002fff1, 0x007b004b, 0x00060050, 0x0000fff1
+};
+
+static const u32 coef_lut_e_c_legacy[NB_COEF] = {
+	0xfffa0001, 0x003f0010, 0x003f006d, 0xfffa0010,
+	0xfffb0002, 0x00440011, 0x00440062, 0xfffa000e,
+	0xfffb0001, 0x00460013, 0x00420062, 0xfffa000d,
+	0xfffb0000, 0x00480014, 0x00410062, 0xfffa000c,
+	0xfffb0001, 0x00490015, 0x003f0061, 0xfffb000b,
+	0xfffb0000, 0x004b0017, 0x003d0061, 0xfffb000a,
+	0xfffb0000, 0x004d0018, 0x003b0062, 0xfffb0008,
+	0xfffcffff, 0x004f001a, 0x00390061, 0xfffb0007,
+	0xfffc0000, 0x004f001c, 0x00380060, 0xfffb0006,
+	0xfffcffff, 0x0052001d, 0x00360060, 0xfffb0005,
+	0xfffdfffe, 0x0053001f, 0x00340060, 0xfffb0004,
+	0xfffdfffe, 0x00540021, 0x0032005e, 0xfffc0004,
+	0xfffeffff, 0x00550022, 0x0030005d, 0xfffc0003,
+	0xfffeffff, 0x00560024, 0x002f005c, 0xfffc0002,
+	0xfffffffd, 0x00580026, 0x002d005c, 0xfffc0001,
+	0xfffffffd, 0x005a0027, 0x002b005c, 0xfffc0000,
+	0x0000fffd, 0x005a0029, 0x0029005a, 0xfffd0000,
+	0x0000fffc, 0x005c002b, 0x0027005a, 0xfffdffff,
+	0x0001fffc, 0x005c002d, 0x00260058, 0xfffdffff,
+	0x0002fffc, 0x005c002f, 0x00240056, 0xfffffffe,
+	0x0003fffc, 0x005d0030, 0x00220055, 0xfffffffe,
+	0x0004fffc, 0x005e0032, 0x00210054, 0xfffefffd,
+	0x0004fffb, 0x00600034, 0x001f0053, 0xfffefffd,
+	0x0005fffb, 0x00600036, 0x001d0052, 0xfffffffc,
+	0x0006fffb, 0x00600038, 0x001c004f, 0x0000fffc,
+	0x0007fffb, 0x00610039, 0x001a004f, 0xfffffffc,
+	0x0008fffb, 0x0062003b, 0x0018004d, 0x0000fffb,
+	0x000afffb, 0x0061003d, 0x0017004b, 0x0000fffb,
+	0x000bfffb, 0x0061003f, 0x00150049, 0x0001fffb,
+	0x000cfffa, 0x00620041, 0x00140048, 0x0000fffb,
+	0x000dfffa, 0x00620042, 0x00130046, 0x0001fffb,
+	0x000efffa, 0x00620044, 0x00110044, 0x0002fffb
+};
+
+static const u32 coef_lut_f_y_legacy[NB_COEF] = {
+	0xfff6fff0, 0x00490012, 0x0049006e, 0xfff60012,
+	0xfff7fff1, 0x004e0013, 0x00490068, 0xfff60010,
+	0xfff7fff2, 0x004f0015, 0x00470067, 0xfff6000f,
+	0xfff7fff5, 0x004f0017, 0x00450065, 0xfff6000e,
+	0xfff8fff5, 0x00500018, 0x00440065, 0xfff6000c,
+	0xfff8fff6, 0x0051001a, 0x00420064, 0xfff6000b,
+	0xfff8fff6, 0x0052001c, 0x00400064, 0xfff6000a,
+	0xfff9fff6, 0x0054001d, 0x003e0064, 0xfff60008,
+	0xfff9fff8, 0x0054001f, 0x003c0063, 0xfff60007,
+	0xfffafff8, 0x00550021, 0x003a0062, 0xfff60006,
+	0xfffbfff7, 0x00560022, 0x00390062, 0xfff60005,
+	0xfffbfff8, 0x00570024, 0x00370061, 0xfff60004,
+	0xfffcfff8, 0x00580026, 0x00350060, 0xfff60003,
+	0xfffdfff8, 0x00590028, 0x0033005f, 0xfff60002,
+	0xfffdfff7, 0x005b002a, 0x0031005f, 0xfff60001,
+	0xfffefff7, 0x005c002c, 0x002f005e, 0xfff60000,
+	0xfffffff6, 0x005e002d, 0x002d005e, 0xfff6ffff,
+	0x0000fff6, 0x005e002f, 0x002c005c, 0xfff7fffe,
+	0x0001fff6, 0x005f0031, 0x002a005b, 0xfff7fffd,
+	0x0002fff6, 0x005f0033, 0x00280059, 0xfff8fffd,
+	0x0003fff6, 0x00600035, 0x00260058, 0xfff8fffc,
+	0x0004fff6, 0x00610037, 0x00240057, 0xfff8fffb,
+	0x0005fff6, 0x00620039, 0x00220056, 0xfff7fffb,
+	0x0006fff6, 0x0062003a, 0x00210055, 0xfff8fffa,
+	0x0007fff6, 0x0063003c, 0x001f0054, 0xfff8fff9,
+	0x0008fff6, 0x0064003e, 0x001d0054, 0xfff6fff9,
+	0x000afff6, 0x00640040, 0x001c0052, 0xfff6fff8,
+	0x000bfff6, 0x00640042, 0x001a0051, 0xfff6fff8,
+	0x000cfff6, 0x00650044, 0x00180050, 0xfff5fff8,
+	0x000efff6, 0x00650045, 0x0017004f, 0xfff5fff7,
+	0x000ffff6, 0x00670047, 0x0015004f, 0xfff2fff7,
+	0x0010fff6, 0x00680049, 0x0013004e, 0xfff1fff7
+};
+
+static const u32 coef_lut_f_c_legacy[NB_COEF] = {
+	0x0000fffb, 0x003a001a, 0x003a005d, 0x0000001a,
+	0x0001fffb, 0x003f001b, 0x00400051, 0x00000019,
+	0x0001fffc, 0x0040001c, 0x003f0051, 0x00000017,
+	0x0002fffb, 0x0042001d, 0x003e0051, 0xffff0016,
+	0x0002fffb, 0x0043001e, 0x003d0051, 0xffff0015,
+	0x0003fffc, 0x00430020, 0x003b0050, 0xffff0014,
+	0x0003fffb, 0x00450021, 0x003a0051, 0xfffe0013,
+	0x0004fffc, 0x00450022, 0x00390050, 0xfffe0012,
+	0x0005fffc, 0x00460023, 0x0038004f, 0xfffe0011,
+	0x0005fffb, 0x00480025, 0x00360050, 0xfffd0010,
+	0x0006fffc, 0x00480026, 0x0035004f, 0xfffd000f,
+	0x0006fffc, 0x00490027, 0x0034004f, 0xfffd000e,
+	0x0007fffd, 0x00490028, 0x0033004e, 0xfffd000d,
+	0x0008fffc, 0x004a002a, 0x0031004d, 0xfffd000d,
+	0x0009fffd, 0x004a002b, 0x0030004d, 0xfffc000c,
+	0x0009fffc, 0x004c002c, 0x002f004d, 0xfffc000b,
+	0x000afffc, 0x004c002e, 0x002e004c, 0xfffc000a,
+	0x000bfffc, 0x004d002f, 0x002c004c, 0xfffc0009,
+	0x000cfffc, 0x004d0030, 0x002b004a, 0xfffd0009,
+	0x000dfffd, 0x004d0031, 0x002a004a, 0xfffc0008,
+	0x000dfffd, 0x004e0033, 0x00280049, 0xfffd0007,
+	0x000efffd, 0x004f0034, 0x00270049, 0xfffc0006,
+	0x000ffffd, 0x004f0035, 0x00260048, 0xfffc0006,
+	0x0010fffd, 0x00500036, 0x00250048, 0xfffb0005,
+	0x0011fffe, 0x004f0038, 0x00230046, 0xfffc0005,
+	0x0012fffe, 0x00500039, 0x00220045, 0xfffc0004,
+	0x0013fffe, 0x0051003a, 0x00210045, 0xfffb0003,
+	0x0014ffff, 0x0050003b, 0x00200043, 0xfffc0003,
+	0x0015ffff, 0x0051003d, 0x001e0043, 0xfffb0002,
+	0x0016ffff, 0x0051003e, 0x001d0042, 0xfffb0002,
+	0x00170000, 0x0051003f, 0x001c0040, 0xfffc0001,
+	0x00190000, 0x00510040, 0x001b003f, 0xfffb0001
+};
+
+#endif
diff --git a/drivers/gpu/drm/sti/sti_layer.c b/drivers/gpu/drm/sti/sti_layer.c
index 06a587c4f1bb..899104f9d4bc 100644
--- a/drivers/gpu/drm/sti/sti_layer.c
+++ b/drivers/gpu/drm/sti/sti_layer.c
@@ -11,7 +11,9 @@
 #include <drm/drm_fb_cma_helper.h>
 
 #include "sti_compositor.h"
+#include "sti_cursor.h"
 #include "sti_gdp.h"
+#include "sti_hqvdp.h"
 #include "sti_layer.h"
 #include "sti_vid.h"
 
@@ -32,10 +34,13 @@ const char *sti_layer_to_str(struct sti_layer *layer)
 		return "VID1";
 	case STI_CURSOR:
 		return "CURSOR";
+	case STI_HQVDP_0:
+		return "HQVDP0";
 	default:
 		return "<UNKNOWN LAYER>";
 	}
 }
+EXPORT_SYMBOL(sti_layer_to_str);
 
 struct sti_layer *sti_layer_create(struct device *dev, int desc,
 				   void __iomem *baseaddr)
@@ -50,6 +55,12 @@ struct sti_layer *sti_layer_create(struct device *dev, int desc,
 	case STI_VID:
 		layer = sti_vid_create(dev);
 		break;
+	case STI_CUR:
+		layer = sti_cursor_create(dev);
+		break;
+	case STI_VDP:
+		layer = sti_hqvdp_create(dev);
+		break;
 	}
 
 	if (!layer) {
@@ -67,8 +78,11 @@ struct sti_layer *sti_layer_create(struct device *dev, int desc,
 
 	return layer;
 }
+EXPORT_SYMBOL(sti_layer_create);
 
-int sti_layer_prepare(struct sti_layer *layer, struct drm_framebuffer *fb,
+int sti_layer_prepare(struct sti_layer *layer,
+		      struct drm_crtc *crtc,
+		      struct drm_framebuffer *fb,
 		      struct drm_display_mode *mode, int mixer_id,
 		      int dest_x, int dest_y, int dest_w, int dest_h,
 		      int src_x, int src_y, int src_w, int src_h)
@@ -88,6 +102,7 @@ int sti_layer_prepare(struct sti_layer *layer, struct drm_framebuffer *fb,
 		return 1;
 	}
 
+	layer->crtc = crtc;
 	layer->fb = fb;
 	layer->mode = mode;
 	layer->mixer_id = mixer_id;
@@ -100,6 +115,7 @@ int sti_layer_prepare(struct sti_layer *layer, struct drm_framebuffer *fb,
 	layer->src_w = src_w;
 	layer->src_h = src_h;
 	layer->format = fb->pixel_format;
+	layer->vaddr = cma_obj->vaddr;
 	layer->paddr = cma_obj->paddr;
 	for (i = 0; i < 4; i++) {
 		layer->pitches[i] = fb->pitches[i];
diff --git a/drivers/gpu/drm/sti/sti_layer.h b/drivers/gpu/drm/sti/sti_layer.h
index 198c3774cc12..ceff497f557e 100644
--- a/drivers/gpu/drm/sti/sti_layer.h
+++ b/drivers/gpu/drm/sti/sti_layer.h
@@ -22,7 +22,8 @@ enum sti_layer_type {
 	STI_GDP = 1 << STI_LAYER_TYPE_SHIFT,
 	STI_VID = 2 << STI_LAYER_TYPE_SHIFT,
 	STI_CUR = 3 << STI_LAYER_TYPE_SHIFT,
-	STI_BCK = 4 << STI_LAYER_TYPE_SHIFT
+	STI_BCK = 4 << STI_LAYER_TYPE_SHIFT,
+	STI_VDP = 5 << STI_LAYER_TYPE_SHIFT
 };
 
 enum sti_layer_id_of_type {
@@ -39,6 +40,7 @@ enum sti_layer_desc {
 	STI_GDP_3       = STI_GDP | STI_ID_3,
 	STI_VID_0       = STI_VID | STI_ID_0,
 	STI_VID_1       = STI_VID | STI_ID_1,
+	STI_HQVDP_0     = STI_VDP | STI_ID_0,
 	STI_CURSOR      = STI_CUR,
 	STI_BACK        = STI_BCK
 };
@@ -67,6 +69,7 @@ struct sti_layer_funcs {
  *
  * @plane:              drm plane it is bound to (if any)
  * @fb:                 drm fb it is bound to
+ * @crtc:               crtc it is bound to
  * @mode:               display mode
  * @desc:               layer type & id
  * @device:		driver device
@@ -82,11 +85,13 @@ struct sti_layer_funcs {
  * @format:             format
  * @pitches:            pitch of 'planes' (eg: Y, U, V)
  * @offsets:            offset of 'planes'
+ * @vaddr:              virtual address of the input buffer
  * @paddr:              physical address of the input buffer
  */
 struct sti_layer {
 	struct drm_plane plane;
 	struct drm_framebuffer *fb;
+	struct drm_crtc *crtc;
 	struct drm_display_mode *mode;
 	enum sti_layer_desc desc;
 	struct device *dev;
@@ -102,12 +107,15 @@ struct sti_layer {
 	uint32_t format;
 	unsigned int pitches[4];
 	unsigned int offsets[4];
+	void *vaddr;
 	dma_addr_t paddr;
 };
 
 struct sti_layer *sti_layer_create(struct device *dev, int desc,
 			void __iomem *baseaddr);
-int sti_layer_prepare(struct sti_layer *layer, struct drm_framebuffer *fb,
+int sti_layer_prepare(struct sti_layer *layer,
+			struct drm_crtc *crtc,
+			struct drm_framebuffer *fb,
 			struct drm_display_mode *mode,
 			int mixer_id,
 			int dest_x, int dest_y,
diff --git a/drivers/gpu/drm/sti/sti_mixer.c b/drivers/gpu/drm/sti/sti_mixer.c
index 79f369db9fb6..13a4b84deab6 100644
--- a/drivers/gpu/drm/sti/sti_mixer.c
+++ b/drivers/gpu/drm/sti/sti_mixer.c
@@ -45,6 +45,7 @@ static const u32 mixerColorSpaceMatIdentity[] = {
 #define GAM_CTL_GDP1_MASK  BIT(4)
 #define GAM_CTL_GDP2_MASK  BIT(5)
 #define GAM_CTL_GDP3_MASK  BIT(6)
+#define GAM_CTL_CURSOR_MASK BIT(9)
 
 const char *sti_mixer_to_str(struct sti_mixer *mixer)
 {
@@ -122,11 +123,15 @@ int sti_mixer_set_layer_depth(struct sti_mixer *mixer, struct sti_layer *layer)
 		layer_id = GAM_DEPTH_GDP3_ID;
 		break;
 	case STI_VID_0:
+	case STI_HQVDP_0:
 		layer_id = GAM_DEPTH_VID0_ID;
 		break;
 	case STI_VID_1:
 		layer_id = GAM_DEPTH_VID1_ID;
 		break;
+	case STI_CURSOR:
+		/* no need to set depth for cursor */
+		return 0;
 	default:
 		DRM_ERROR("Unknown layer %d\n", layer->desc);
 		return 1;
@@ -185,9 +190,12 @@ static u32 sti_mixer_get_layer_mask(struct sti_layer *layer)
 	case STI_GDP_3:
 		return GAM_CTL_GDP3_MASK;
 	case STI_VID_0:
+	case STI_HQVDP_0:
 		return GAM_CTL_VID0_MASK;
 	case STI_VID_1:
 		return GAM_CTL_VID1_MASK;
+	case STI_CURSOR:
+		return GAM_CTL_CURSOR_MASK;
 	default:
 		return 0;
 	}
@@ -215,6 +223,15 @@ int sti_mixer_set_layer_status(struct sti_mixer *mixer,
 	return 0;
 }
 
+void sti_mixer_clear_all_layers(struct sti_mixer *mixer)
+{
+	u32 val;
+
+	DRM_DEBUG_DRIVER("%s clear all layer\n", sti_mixer_to_str(mixer));
+	val = sti_mixer_reg_read(mixer, GAM_MIXER_CTL) & 0xFFFF0000;
+	sti_mixer_reg_write(mixer, GAM_MIXER_CTL, val);
+}
+
 void sti_mixer_set_matrix(struct sti_mixer *mixer)
 {
 	unsigned int i;
diff --git a/drivers/gpu/drm/sti/sti_mixer.h b/drivers/gpu/drm/sti/sti_mixer.h
index 874372102e52..b97282182908 100644
--- a/drivers/gpu/drm/sti/sti_mixer.h
+++ b/drivers/gpu/drm/sti/sti_mixer.h
@@ -23,6 +23,7 @@
  * @id: id of the mixer
  * @drm_crtc: crtc object link to the mixer
  * @pending_event: set if a flip event is pending on crtc
+ * @enabled: to know if the mixer is active or not
  */
 struct sti_mixer {
 	struct device *dev;
@@ -30,6 +31,7 @@ struct sti_mixer {
 	int id;
 	struct drm_crtc	drm_crtc;
 	struct drm_pending_vblank_event *pending_event;
+	bool enabled;
 };
 
 const char *sti_mixer_to_str(struct sti_mixer *mixer);
@@ -39,6 +41,7 @@ struct sti_mixer *sti_mixer_create(struct device *dev, int id,
 
 int sti_mixer_set_layer_status(struct sti_mixer *mixer,
 		struct sti_layer *layer, bool status);
+void sti_mixer_clear_all_layers(struct sti_mixer *mixer);
 int sti_mixer_set_layer_depth(struct sti_mixer *mixer, struct sti_layer *layer);
 int sti_mixer_active_video_area(struct sti_mixer *mixer,
 		struct drm_display_mode *mode);
diff --git a/drivers/gpu/drm/sti/sti_tvout.c b/drivers/gpu/drm/sti/sti_tvout.c
index b8afe490356a..cb924aa2b321 100644
--- a/drivers/gpu/drm/sti/sti_tvout.c
+++ b/drivers/gpu/drm/sti/sti_tvout.c
@@ -16,6 +16,8 @@
 #include <drm/drmP.h>
 #include <drm/drm_crtc_helper.h>
 
+#include "sti_drm_crtc.h"
+
 /* glue registers */
 #define TVO_CSC_MAIN_M0                  0x000
 #define TVO_CSC_MAIN_M1                  0x004
@@ -96,7 +98,7 @@
 
 #define TVO_SYNC_HD_DCS_SHIFT            8
 
-#define ENCODER_MAIN_CRTC_MASK           BIT(0)
+#define ENCODER_CRTC_MASK                (BIT(0) | BIT(1))
 
 /* enum listing the supported output data format */
 enum sti_tvout_video_out_type {
@@ -149,14 +151,15 @@ static void tvout_write(struct sti_tvout *tvout, u32 val, int offset)
  * Set the clipping mode of a VIP
  *
  * @tvout: tvout structure
+ * @reg: register to set
  * @cr_r:
  * @y_g:
  * @cb_b:
  */
-static void tvout_vip_set_color_order(struct sti_tvout *tvout,
+static void tvout_vip_set_color_order(struct sti_tvout *tvout, int reg,
 				      u32 cr_r, u32 y_g, u32 cb_b)
 {
-	u32 val = tvout_read(tvout, TVO_VIP_HDMI);
+	u32 val = tvout_read(tvout, reg);
 
 	val &= ~(TVO_VIP_REORDER_MASK << TVO_VIP_REORDER_R_SHIFT);
 	val &= ~(TVO_VIP_REORDER_MASK << TVO_VIP_REORDER_G_SHIFT);
@@ -165,52 +168,58 @@ static void tvout_vip_set_color_order(struct sti_tvout *tvout,
 	val |= y_g << TVO_VIP_REORDER_G_SHIFT;
 	val |= cb_b << TVO_VIP_REORDER_B_SHIFT;
 
-	tvout_write(tvout, val, TVO_VIP_HDMI);
+	tvout_write(tvout, val, reg);
 }
 
 /**
  * Set the clipping mode of a VIP
  *
  * @tvout: tvout structure
+ * @reg: register to set
  * @range: clipping range
  */
-static void tvout_vip_set_clip_mode(struct sti_tvout *tvout, u32 range)
+static void tvout_vip_set_clip_mode(struct sti_tvout *tvout, int reg, u32 range)
 {
-	u32 val = tvout_read(tvout, TVO_VIP_HDMI);
+	u32 val = tvout_read(tvout, reg);
 
 	val &= ~(TVO_VIP_CLIP_MASK << TVO_VIP_CLIP_SHIFT);
 	val |= range << TVO_VIP_CLIP_SHIFT;
-	tvout_write(tvout, val, TVO_VIP_HDMI);
+	tvout_write(tvout, val, reg);
 }
 
 /**
  * Set the rounded value of a VIP
  *
  * @tvout: tvout structure
+ * @reg: register to set
  * @rnd: rounded val per component
  */
-static void tvout_vip_set_rnd(struct sti_tvout *tvout, u32 rnd)
+static void tvout_vip_set_rnd(struct sti_tvout *tvout, int reg, u32 rnd)
 {
-	u32 val = tvout_read(tvout, TVO_VIP_HDMI);
+	u32 val = tvout_read(tvout, reg);
 
 	val &= ~(TVO_VIP_RND_MASK << TVO_VIP_RND_SHIFT);
 	val |= rnd << TVO_VIP_RND_SHIFT;
-	tvout_write(tvout, val, TVO_VIP_HDMI);
+	tvout_write(tvout, val, reg);
 }
 
 /**
  * Select the VIP input
  *
  * @tvout: tvout structure
+ * @reg: register to set
+ * @main_path: main or auxiliary path
+ * @sel_input_logic_inverted: need to invert the logic
  * @sel_input: selected_input (main/aux + conv)
  */
 static void tvout_vip_set_sel_input(struct sti_tvout *tvout,
+				    int reg,
 				    bool main_path,
 				    bool sel_input_logic_inverted,
 				    enum sti_tvout_video_out_type video_out)
 {
 	u32 sel_input;
-	u32 val = tvout_read(tvout, TVO_VIP_HDMI);
+	u32 val = tvout_read(tvout, reg);
 
 	if (main_path)
 		sel_input = TVO_VIP_SEL_INPUT_MAIN;
@@ -232,22 +241,24 @@ static void tvout_vip_set_sel_input(struct sti_tvout *tvout,
 
 	val &= ~TVO_VIP_SEL_INPUT_MASK;
 	val |= sel_input;
-	tvout_write(tvout, val, TVO_VIP_HDMI);
+	tvout_write(tvout, val, reg);
 }
 
 /**
  * Select the input video signed or unsigned
  *
  * @tvout: tvout structure
+ * @reg: register to set
  * @in_vid_signed: used video input format
  */
-static void tvout_vip_set_in_vid_fmt(struct sti_tvout *tvout, u32 in_vid_fmt)
+static void tvout_vip_set_in_vid_fmt(struct sti_tvout *tvout,
+		int reg, u32 in_vid_fmt)
 {
-	u32 val = tvout_read(tvout, TVO_VIP_HDMI);
+	u32 val = tvout_read(tvout, reg);
 
 	val &= ~TVO_IN_FMT_SIGNED;
 	val |= in_vid_fmt;
-	tvout_write(tvout, val, TVO_MAIN_IN_VID_FORMAT);
+	tvout_write(tvout, val, reg);
 }
 
 /**
@@ -261,6 +272,7 @@ static void tvout_hdmi_start(struct sti_tvout *tvout, bool main_path)
 {
 	struct device_node *node = tvout->dev->of_node;
 	bool sel_input_logic_inverted = false;
+	u32 tvo_in_vid_format;
 
 	dev_dbg(tvout->dev, "%s\n", __func__);
 
@@ -268,33 +280,36 @@ static void tvout_hdmi_start(struct sti_tvout *tvout, bool main_path)
 		DRM_DEBUG_DRIVER("main vip for hdmi\n");
 		/* select the input sync for hdmi = VTG set 1 */
 		tvout_write(tvout, TVO_SYNC_MAIN_VTG_SET_1, TVO_HDMI_SYNC_SEL);
+		tvo_in_vid_format = TVO_MAIN_IN_VID_FORMAT;
 	} else {
 		DRM_DEBUG_DRIVER("aux vip for hdmi\n");
 		/* select the input sync for hdmi = VTG set 1 */
 		tvout_write(tvout, TVO_SYNC_AUX_VTG_SET_1, TVO_HDMI_SYNC_SEL);
+		tvo_in_vid_format = TVO_AUX_IN_VID_FORMAT;
 	}
 
 	/* set color channel order */
-	tvout_vip_set_color_order(tvout,
+	tvout_vip_set_color_order(tvout, TVO_VIP_HDMI,
 				  TVO_VIP_REORDER_CR_R_SEL,
 				  TVO_VIP_REORDER_Y_G_SEL,
 				  TVO_VIP_REORDER_CB_B_SEL);
 
 	/* set clipping mode (Limited range RGB/Y) */
-	tvout_vip_set_clip_mode(tvout, TVO_VIP_CLIP_LIMITED_RANGE_RGB_Y);
+	tvout_vip_set_clip_mode(tvout, TVO_VIP_HDMI,
+			TVO_VIP_CLIP_LIMITED_RANGE_RGB_Y);
 
 	/* set round mode (rounded to 8-bit per component) */
-	tvout_vip_set_rnd(tvout, TVO_VIP_RND_8BIT_ROUNDED);
+	tvout_vip_set_rnd(tvout, TVO_VIP_HDMI, TVO_VIP_RND_8BIT_ROUNDED);
 
 	if (of_device_is_compatible(node, "st,stih407-tvout")) {
 		/* set input video format */
-		tvout_vip_set_in_vid_fmt(tvout->regs + TVO_MAIN_IN_VID_FORMAT,
-					 TVO_IN_FMT_SIGNED);
+		tvout_vip_set_in_vid_fmt(tvout, tvo_in_vid_format,
+					TVO_IN_FMT_SIGNED);
 		sel_input_logic_inverted = true;
 	}
 
 	/* input selection */
-	tvout_vip_set_sel_input(tvout, main_path,
+	tvout_vip_set_sel_input(tvout, TVO_VIP_HDMI, main_path,
 			sel_input_logic_inverted, STI_TVOUT_VIDEO_OUT_RGB);
 }
 
@@ -309,48 +324,47 @@ static void tvout_hda_start(struct sti_tvout *tvout, bool main_path)
 {
 	struct device_node *node = tvout->dev->of_node;
 	bool sel_input_logic_inverted = false;
+	u32 tvo_in_vid_format;
+	int val;
 
 	dev_dbg(tvout->dev, "%s\n", __func__);
 
-	if (!main_path) {
-		DRM_ERROR("HD Analog on aux not implemented\n");
-		return;
+	if (main_path) {
+		val = TVO_SYNC_MAIN_VTG_SET_2 << TVO_SYNC_HD_DCS_SHIFT;
+		val |= TVO_SYNC_MAIN_VTG_SET_3;
+		tvout_write(tvout, val, TVO_HD_SYNC_SEL);
+		tvo_in_vid_format = TVO_MAIN_IN_VID_FORMAT;
+	} else {
+		val = TVO_SYNC_AUX_VTG_SET_2 << TVO_SYNC_HD_DCS_SHIFT;
+		val |= TVO_SYNC_AUX_VTG_SET_3;
+		tvout_write(tvout, val, TVO_HD_SYNC_SEL);
+		tvo_in_vid_format = TVO_AUX_IN_VID_FORMAT;
 	}
 
-	DRM_DEBUG_DRIVER("main vip for HDF\n");
-
 	/* set color channel order */
-	tvout_vip_set_color_order(tvout->regs + TVO_VIP_HDF,
+	tvout_vip_set_color_order(tvout, TVO_VIP_HDF,
 				  TVO_VIP_REORDER_CR_R_SEL,
 				  TVO_VIP_REORDER_Y_G_SEL,
 				  TVO_VIP_REORDER_CB_B_SEL);
 
-	/* set clipping mode (Limited range RGB/Y) */
-	tvout_vip_set_clip_mode(tvout->regs + TVO_VIP_HDF,
-				TVO_VIP_CLIP_LIMITED_RANGE_CB_CR);
+	/* set clipping mode (EAV/SAV clipping) */
+	tvout_vip_set_clip_mode(tvout, TVO_VIP_HDF, TVO_VIP_CLIP_EAV_SAV);
 
 	/* set round mode (rounded to 10-bit per component) */
-	tvout_vip_set_rnd(tvout->regs + TVO_VIP_HDF, TVO_VIP_RND_10BIT_ROUNDED);
+	tvout_vip_set_rnd(tvout, TVO_VIP_HDF, TVO_VIP_RND_10BIT_ROUNDED);
 
 	if (of_device_is_compatible(node, "st,stih407-tvout")) {
 		/* set input video format */
-		tvout_vip_set_in_vid_fmt(tvout, TVO_IN_FMT_SIGNED);
+		tvout_vip_set_in_vid_fmt(tvout,
+			tvo_in_vid_format, TVO_IN_FMT_SIGNED);
 		sel_input_logic_inverted = true;
 	}
 
 	/* Input selection */
-	tvout_vip_set_sel_input(tvout->regs + TVO_VIP_HDF,
-				main_path,
+	tvout_vip_set_sel_input(tvout, TVO_VIP_HDF, main_path,
 				sel_input_logic_inverted,
 				STI_TVOUT_VIDEO_OUT_YUV);
 
-	/* select the input sync for HD analog = VTG set 3
-	 * and HD DCS = VTG set 2 */
-	tvout_write(tvout,
-		(TVO_SYNC_MAIN_VTG_SET_2 << TVO_SYNC_HD_DCS_SHIFT)
-		| TVO_SYNC_MAIN_VTG_SET_3,
-		TVO_HD_SYNC_SEL);
-
 	/* power up HD DAC */
 	tvout_write(tvout, 0, TVO_HD_DAC_CFG_OFF);
 }
@@ -392,7 +406,7 @@ static void sti_hda_encoder_commit(struct drm_encoder *encoder)
 {
 	struct sti_tvout *tvout = to_sti_tvout(encoder);
 
-	tvout_hda_start(tvout, true);
+	tvout_hda_start(tvout, sti_drm_crtc_is_main(encoder->crtc));
 }
 
 static void sti_hda_encoder_disable(struct drm_encoder *encoder)
@@ -429,7 +443,7 @@ static struct drm_encoder *sti_tvout_create_hda_encoder(struct drm_device *dev,
 
 	drm_encoder = (struct drm_encoder *) encoder;
 
-	drm_encoder->possible_crtcs = ENCODER_MAIN_CRTC_MASK;
+	drm_encoder->possible_crtcs = ENCODER_CRTC_MASK;
 	drm_encoder->possible_clones = 1 << 0;
 
 	drm_encoder_init(dev, drm_encoder,
@@ -444,7 +458,7 @@ static void sti_hdmi_encoder_commit(struct drm_encoder *encoder)
 {
 	struct sti_tvout *tvout = to_sti_tvout(encoder);
 
-	tvout_hdmi_start(tvout, true);
+	tvout_hdmi_start(tvout, sti_drm_crtc_is_main(encoder->crtc));
 }
 
 static void sti_hdmi_encoder_disable(struct drm_encoder *encoder)
@@ -478,7 +492,7 @@ static struct drm_encoder *sti_tvout_create_hdmi_encoder(struct drm_device *dev,
 
 	drm_encoder = (struct drm_encoder *) encoder;
 
-	drm_encoder->possible_crtcs = ENCODER_MAIN_CRTC_MASK;
+	drm_encoder->possible_crtcs = ENCODER_CRTC_MASK;
 	drm_encoder->possible_clones = 1 << 1;
 
 	drm_encoder_init(dev, drm_encoder,
diff --git a/drivers/gpu/drm/sti/sti_vtac.c b/drivers/gpu/drm/sti/sti_vtac.c
index 82a51d488434..97bcdac23ae1 100644
--- a/drivers/gpu/drm/sti/sti_vtac.c
+++ b/drivers/gpu/drm/sti/sti_vtac.c
@@ -56,8 +56,16 @@ struct sti_vtac_mode {
 	u32 phyts_per_pixel;
 };
 
-static const struct sti_vtac_mode vtac_mode_main = {0x2, 0x2, VTAC_5_PPP};
-static const struct sti_vtac_mode vtac_mode_aux = {0x1, 0x0, VTAC_17_PPP};
+static const struct sti_vtac_mode vtac_mode_main = {
+	.vid_in_width = 0x2,
+	.phyts_width = 0x2,
+	.phyts_per_pixel = VTAC_5_PPP,
+};
+static const struct sti_vtac_mode vtac_mode_aux = {
+	.vid_in_width = 0x1,
+	.phyts_width = 0x0,
+	.phyts_per_pixel = VTAC_17_PPP,
+};
 
 /**
  * VTAC structure
diff --git a/drivers/gpu/drm/sti/sti_vtg.c b/drivers/gpu/drm/sti/sti_vtg.c
index 740d6e347a62..9564f2568e2c 100644
--- a/drivers/gpu/drm/sti/sti_vtg.c
+++ b/drivers/gpu/drm/sti/sti_vtg.c
@@ -51,10 +51,19 @@
 #define VTG_TOP_V_HD_3      0x010C
 #define VTG_BOT_V_HD_3      0x0110
 
+#define VTG_H_HD_4          0x0120
+#define VTG_TOP_V_VD_4      0x0124
+#define VTG_BOT_V_VD_4      0x0128
+#define VTG_TOP_V_HD_4      0x012c
+#define VTG_BOT_V_HD_4      0x0130
+
 #define VTG_IRQ_BOTTOM      BIT(0)
 #define VTG_IRQ_TOP         BIT(1)
 #define VTG_IRQ_MASK        (VTG_IRQ_TOP | VTG_IRQ_BOTTOM)
 
+/* Delay introduced by the HDMI in nb of pixel */
+#define HDMI_DELAY          (6)
+
 /* delay introduced by the Arbitrary Waveform Generator in nb of pixels */
 #define AWG_DELAY_HD        (-9)
 #define AWG_DELAY_ED        (-8)
@@ -133,10 +142,10 @@ static void vtg_set_mode(struct sti_vtg *vtg,
 	writel(tmp, vtg->regs + VTG_VID_TFS);
 	writel(tmp, vtg->regs + VTG_VID_BFS);
 
-	/* prepare VTG set 1 and 2 for HDMI and VTG set 3 for HD DAC */
-	tmp = (mode->hsync_end - mode->hsync_start) << 16;
+	/* prepare VTG set 1 for HDMI */
+	tmp = (mode->hsync_end - mode->hsync_start + HDMI_DELAY) << 16;
+	tmp |= HDMI_DELAY;
 	writel(tmp, vtg->regs + VTG_H_HD_1);
-	writel(tmp, vtg->regs + VTG_H_HD_2);
 
 	tmp = (mode->vsync_end - mode->vsync_start + 1) << 16;
 	tmp |= 1;
@@ -146,6 +155,11 @@ static void vtg_set_mode(struct sti_vtg *vtg,
 	writel(0, vtg->regs + VTG_BOT_V_HD_1);
 
 	/* prepare VTG set 2 for for HD DCS */
+	tmp = (mode->hsync_end - mode->hsync_start) << 16;
+	writel(tmp, vtg->regs + VTG_H_HD_2);
+
+	tmp = (mode->vsync_end - mode->vsync_start + 1) << 16;
+	tmp |= 1;
 	writel(tmp, vtg->regs + VTG_TOP_V_VD_2);
 	writel(tmp, vtg->regs + VTG_BOT_V_VD_2);
 	writel(0, vtg->regs + VTG_TOP_V_HD_2);
@@ -166,6 +180,17 @@ static void vtg_set_mode(struct sti_vtg *vtg,
 	writel(tmp, vtg->regs + VTG_TOP_V_HD_3);
 	writel(tmp, vtg->regs + VTG_BOT_V_HD_3);
 
+	/* Prepare VTG set 4 for DVO */
+	tmp = (mode->hsync_end - mode->hsync_start) << 16;
+	writel(tmp, vtg->regs + VTG_H_HD_4);
+
+	tmp = (mode->vsync_end - mode->vsync_start + 1) << 16;
+	tmp |= 1;
+	writel(tmp, vtg->regs + VTG_TOP_V_VD_4);
+	writel(tmp, vtg->regs + VTG_BOT_V_VD_4);
+	writel(0, vtg->regs + VTG_TOP_V_HD_4);
+	writel(0, vtg->regs + VTG_BOT_V_HD_4);
+
 	/* mode */
 	writel(type, vtg->regs + VTG_MODE);
 }
diff --git a/drivers/gpu/drm/tdfx/tdfx_drv.c b/drivers/gpu/drm/tdfx/tdfx_drv.c
index 3492ca5c46d3..fab5ebcb0fef 100644
--- a/drivers/gpu/drm/tdfx/tdfx_drv.c
+++ b/drivers/gpu/drm/tdfx/tdfx_drv.c
@@ -36,6 +36,7 @@
 #include "tdfx_drv.h"
 
 #include <drm/drm_pciids.h>
+#include <drm/drm_legacy.h>
 
 static struct pci_device_id pciidlist[] = {
 	tdfx_PCI_IDS
@@ -46,7 +47,7 @@ static const struct file_operations tdfx_driver_fops = {
 	.open = drm_open,
 	.release = drm_release,
 	.unlocked_ioctl = drm_ioctl,
-	.mmap = drm_mmap,
+	.mmap = drm_legacy_mmap,
 	.poll = drm_poll,
 #ifdef CONFIG_COMPAT
 	.compat_ioctl = drm_compat_ioctl,
@@ -55,6 +56,7 @@ static const struct file_operations tdfx_driver_fops = {
 };
 
 static struct drm_driver driver = {
+	.set_busid = drm_pci_set_busid,
 	.fops = &tdfx_driver_fops,
 	.name = DRIVER_NAME,
 	.desc = DRIVER_DESC,
diff --git a/drivers/gpu/drm/tegra/Kconfig b/drivers/gpu/drm/tegra/Kconfig
index 354ddb29231f..74d9d621453d 100644
--- a/drivers/gpu/drm/tegra/Kconfig
+++ b/drivers/gpu/drm/tegra/Kconfig
@@ -1,6 +1,7 @@
 config DRM_TEGRA
 	tristate "NVIDIA Tegra DRM"
 	depends on ARCH_TEGRA || (ARM && COMPILE_TEST)
+	depends on COMMON_CLK
 	depends on DRM
 	depends on RESET_CONTROLLER
 	select DRM_KMS_HELPER
diff --git a/drivers/gpu/drm/tegra/dc.c b/drivers/gpu/drm/tegra/dc.c
index 6553fd238685..978993fa3a36 100644
--- a/drivers/gpu/drm/tegra/dc.c
+++ b/drivers/gpu/drm/tegra/dc.c
@@ -9,17 +9,23 @@
 
 #include <linux/clk.h>
 #include <linux/debugfs.h>
+#include <linux/iommu.h>
 #include <linux/reset.h>
 
+#include <soc/tegra/pmc.h>
+
 #include "dc.h"
 #include "drm.h"
 #include "gem.h"
 
+#include <drm/drm_plane_helper.h>
+
 struct tegra_dc_soc_info {
 	bool supports_interlacing;
 	bool supports_cursor;
 	bool supports_block_linear;
 	unsigned int pitch_align;
+	bool has_powergate;
 };
 
 struct tegra_plane {
@@ -32,6 +38,26 @@ static inline struct tegra_plane *to_tegra_plane(struct drm_plane *plane)
 	return container_of(plane, struct tegra_plane, base);
 }
 
+static void tegra_dc_window_commit(struct tegra_dc *dc, unsigned int index)
+{
+	u32 value = WIN_A_ACT_REQ << index;
+
+	tegra_dc_writel(dc, value << 8, DC_CMD_STATE_CONTROL);
+	tegra_dc_writel(dc, value, DC_CMD_STATE_CONTROL);
+}
+
+static void tegra_dc_cursor_commit(struct tegra_dc *dc)
+{
+	tegra_dc_writel(dc, CURSOR_ACT_REQ << 8, DC_CMD_STATE_CONTROL);
+	tegra_dc_writel(dc, CURSOR_ACT_REQ, DC_CMD_STATE_CONTROL);
+}
+
+static void tegra_dc_commit(struct tegra_dc *dc)
+{
+	tegra_dc_writel(dc, GENERAL_ACT_REQ << 8, DC_CMD_STATE_CONTROL);
+	tegra_dc_writel(dc, GENERAL_ACT_REQ, DC_CMD_STATE_CONTROL);
+}
+
 static unsigned int tegra_dc_format(uint32_t format, uint32_t *swap)
 {
 	/* assume no swapping of fetched data */
@@ -142,7 +168,7 @@ static int tegra_dc_setup_window(struct tegra_dc *dc, unsigned int index,
 				 const struct tegra_dc_window *window)
 {
 	unsigned h_offset, v_offset, h_size, v_size, h_dda, v_dda, bpp;
-	unsigned long value;
+	unsigned long value, flags;
 	bool yuv, planar;
 
 	/*
@@ -155,6 +181,8 @@ static int tegra_dc_setup_window(struct tegra_dc *dc, unsigned int index,
 	else
 		bpp = planar ? 1 : 2;
 
+	spin_lock_irqsave(&dc->lock, flags);
+
 	value = WINDOW_A_SELECT << index;
 	tegra_dc_writel(dc, value, DC_CMD_DISPLAY_WINDOW_HEADER);
 
@@ -247,6 +275,7 @@ static int tegra_dc_setup_window(struct tegra_dc *dc, unsigned int index,
 
 		case TEGRA_BO_TILING_MODE_BLOCK:
 			DRM_ERROR("hardware doesn't support block linear mode\n");
+			spin_unlock_irqrestore(&dc->lock, flags);
 			return -EINVAL;
 		}
 
@@ -303,17 +332,267 @@ static int tegra_dc_setup_window(struct tegra_dc *dc, unsigned int index,
 		break;
 	}
 
-	tegra_dc_writel(dc, WIN_A_UPDATE << index, DC_CMD_STATE_CONTROL);
-	tegra_dc_writel(dc, WIN_A_ACT_REQ << index, DC_CMD_STATE_CONTROL);
+	tegra_dc_window_commit(dc, index);
+
+	spin_unlock_irqrestore(&dc->lock, flags);
+
+	return 0;
+}
+
+static int tegra_window_plane_disable(struct drm_plane *plane)
+{
+	struct tegra_dc *dc = to_tegra_dc(plane->crtc);
+	struct tegra_plane *p = to_tegra_plane(plane);
+	unsigned long flags;
+	u32 value;
+
+	if (!plane->crtc)
+		return 0;
+
+	spin_lock_irqsave(&dc->lock, flags);
+
+	value = WINDOW_A_SELECT << p->index;
+	tegra_dc_writel(dc, value, DC_CMD_DISPLAY_WINDOW_HEADER);
+
+	value = tegra_dc_readl(dc, DC_WIN_WIN_OPTIONS);
+	value &= ~WIN_ENABLE;
+	tegra_dc_writel(dc, value, DC_WIN_WIN_OPTIONS);
+
+	tegra_dc_window_commit(dc, p->index);
+
+	spin_unlock_irqrestore(&dc->lock, flags);
 
 	return 0;
 }
 
-static int tegra_plane_update(struct drm_plane *plane, struct drm_crtc *crtc,
-			      struct drm_framebuffer *fb, int crtc_x,
-			      int crtc_y, unsigned int crtc_w,
-			      unsigned int crtc_h, uint32_t src_x,
-			      uint32_t src_y, uint32_t src_w, uint32_t src_h)
+static void tegra_plane_destroy(struct drm_plane *plane)
+{
+	struct tegra_plane *p = to_tegra_plane(plane);
+
+	drm_plane_cleanup(plane);
+	kfree(p);
+}
+
+static const u32 tegra_primary_plane_formats[] = {
+	DRM_FORMAT_XBGR8888,
+	DRM_FORMAT_XRGB8888,
+	DRM_FORMAT_RGB565,
+};
+
+static int tegra_primary_plane_update(struct drm_plane *plane,
+				      struct drm_crtc *crtc,
+				      struct drm_framebuffer *fb, int crtc_x,
+				      int crtc_y, unsigned int crtc_w,
+				      unsigned int crtc_h, uint32_t src_x,
+				      uint32_t src_y, uint32_t src_w,
+				      uint32_t src_h)
+{
+	struct tegra_bo *bo = tegra_fb_get_plane(fb, 0);
+	struct tegra_plane *p = to_tegra_plane(plane);
+	struct tegra_dc *dc = to_tegra_dc(crtc);
+	struct tegra_dc_window window;
+	int err;
+
+	memset(&window, 0, sizeof(window));
+	window.src.x = src_x >> 16;
+	window.src.y = src_y >> 16;
+	window.src.w = src_w >> 16;
+	window.src.h = src_h >> 16;
+	window.dst.x = crtc_x;
+	window.dst.y = crtc_y;
+	window.dst.w = crtc_w;
+	window.dst.h = crtc_h;
+	window.format = tegra_dc_format(fb->pixel_format, &window.swap);
+	window.bits_per_pixel = fb->bits_per_pixel;
+	window.bottom_up = tegra_fb_is_bottom_up(fb);
+
+	err = tegra_fb_get_tiling(fb, &window.tiling);
+	if (err < 0)
+		return err;
+
+	window.base[0] = bo->paddr + fb->offsets[0];
+	window.stride[0] = fb->pitches[0];
+
+	err = tegra_dc_setup_window(dc, p->index, &window);
+	if (err < 0)
+		return err;
+
+	return 0;
+}
+
+static void tegra_primary_plane_destroy(struct drm_plane *plane)
+{
+	tegra_window_plane_disable(plane);
+	tegra_plane_destroy(plane);
+}
+
+static const struct drm_plane_funcs tegra_primary_plane_funcs = {
+	.update_plane = tegra_primary_plane_update,
+	.disable_plane = tegra_window_plane_disable,
+	.destroy = tegra_primary_plane_destroy,
+};
+
+static struct drm_plane *tegra_dc_primary_plane_create(struct drm_device *drm,
+						       struct tegra_dc *dc)
+{
+	struct tegra_plane *plane;
+	unsigned int num_formats;
+	const u32 *formats;
+	int err;
+
+	plane = kzalloc(sizeof(*plane), GFP_KERNEL);
+	if (!plane)
+		return ERR_PTR(-ENOMEM);
+
+	num_formats = ARRAY_SIZE(tegra_primary_plane_formats);
+	formats = tegra_primary_plane_formats;
+
+	err = drm_universal_plane_init(drm, &plane->base, 1 << dc->pipe,
+				       &tegra_primary_plane_funcs, formats,
+				       num_formats, DRM_PLANE_TYPE_PRIMARY);
+	if (err < 0) {
+		kfree(plane);
+		return ERR_PTR(err);
+	}
+
+	return &plane->base;
+}
+
+static const u32 tegra_cursor_plane_formats[] = {
+	DRM_FORMAT_RGBA8888,
+};
+
+static int tegra_cursor_plane_update(struct drm_plane *plane,
+				     struct drm_crtc *crtc,
+				     struct drm_framebuffer *fb, int crtc_x,
+				     int crtc_y, unsigned int crtc_w,
+				     unsigned int crtc_h, uint32_t src_x,
+				     uint32_t src_y, uint32_t src_w,
+				     uint32_t src_h)
+{
+	struct tegra_bo *bo = tegra_fb_get_plane(fb, 0);
+	struct tegra_dc *dc = to_tegra_dc(crtc);
+	u32 value = CURSOR_CLIP_DISPLAY;
+
+	/* scaling not supported for cursor */
+	if ((src_w >> 16 != crtc_w) || (src_h >> 16 != crtc_h))
+		return -EINVAL;
+
+	/* only square cursors supported */
+	if (src_w != src_h)
+		return -EINVAL;
+
+	switch (crtc_w) {
+	case 32:
+		value |= CURSOR_SIZE_32x32;
+		break;
+
+	case 64:
+		value |= CURSOR_SIZE_64x64;
+		break;
+
+	case 128:
+		value |= CURSOR_SIZE_128x128;
+		break;
+
+	case 256:
+		value |= CURSOR_SIZE_256x256;
+		break;
+
+	default:
+		return -EINVAL;
+	}
+
+	value |= (bo->paddr >> 10) & 0x3fffff;
+	tegra_dc_writel(dc, value, DC_DISP_CURSOR_START_ADDR);
+
+#ifdef CONFIG_ARCH_DMA_ADDR_T_64BIT
+	value = (bo->paddr >> 32) & 0x3;
+	tegra_dc_writel(dc, value, DC_DISP_CURSOR_START_ADDR_HI);
+#endif
+
+	/* enable cursor and set blend mode */
+	value = tegra_dc_readl(dc, DC_DISP_DISP_WIN_OPTIONS);
+	value |= CURSOR_ENABLE;
+	tegra_dc_writel(dc, value, DC_DISP_DISP_WIN_OPTIONS);
+
+	value = tegra_dc_readl(dc, DC_DISP_BLEND_CURSOR_CONTROL);
+	value &= ~CURSOR_DST_BLEND_MASK;
+	value &= ~CURSOR_SRC_BLEND_MASK;
+	value |= CURSOR_MODE_NORMAL;
+	value |= CURSOR_DST_BLEND_NEG_K1_TIMES_SRC;
+	value |= CURSOR_SRC_BLEND_K1_TIMES_SRC;
+	value |= CURSOR_ALPHA;
+	tegra_dc_writel(dc, value, DC_DISP_BLEND_CURSOR_CONTROL);
+
+	/* position the cursor */
+	value = (crtc_y & 0x3fff) << 16 | (crtc_x & 0x3fff);
+	tegra_dc_writel(dc, value, DC_DISP_CURSOR_POSITION);
+
+	/* apply changes */
+	tegra_dc_cursor_commit(dc);
+	tegra_dc_commit(dc);
+
+	return 0;
+}
+
+static int tegra_cursor_plane_disable(struct drm_plane *plane)
+{
+	struct tegra_dc *dc = to_tegra_dc(plane->crtc);
+	u32 value;
+
+	if (!plane->crtc)
+		return 0;
+
+	value = tegra_dc_readl(dc, DC_DISP_DISP_WIN_OPTIONS);
+	value &= ~CURSOR_ENABLE;
+	tegra_dc_writel(dc, value, DC_DISP_DISP_WIN_OPTIONS);
+
+	tegra_dc_cursor_commit(dc);
+	tegra_dc_commit(dc);
+
+	return 0;
+}
+
+static const struct drm_plane_funcs tegra_cursor_plane_funcs = {
+	.update_plane = tegra_cursor_plane_update,
+	.disable_plane = tegra_cursor_plane_disable,
+	.destroy = tegra_plane_destroy,
+};
+
+static struct drm_plane *tegra_dc_cursor_plane_create(struct drm_device *drm,
+						      struct tegra_dc *dc)
+{
+	struct tegra_plane *plane;
+	unsigned int num_formats;
+	const u32 *formats;
+	int err;
+
+	plane = kzalloc(sizeof(*plane), GFP_KERNEL);
+	if (!plane)
+		return ERR_PTR(-ENOMEM);
+
+	num_formats = ARRAY_SIZE(tegra_cursor_plane_formats);
+	formats = tegra_cursor_plane_formats;
+
+	err = drm_universal_plane_init(drm, &plane->base, 1 << dc->pipe,
+				       &tegra_cursor_plane_funcs, formats,
+				       num_formats, DRM_PLANE_TYPE_CURSOR);
+	if (err < 0) {
+		kfree(plane);
+		return ERR_PTR(err);
+	}
+
+	return &plane->base;
+}
+
+static int tegra_overlay_plane_update(struct drm_plane *plane,
+				      struct drm_crtc *crtc,
+				      struct drm_framebuffer *fb, int crtc_x,
+				      int crtc_y, unsigned int crtc_w,
+				      unsigned int crtc_h, uint32_t src_x,
+				      uint32_t src_y, uint32_t src_w,
+				      uint32_t src_h)
 {
 	struct tegra_plane *p = to_tegra_plane(plane);
 	struct tegra_dc *dc = to_tegra_dc(crtc);
@@ -359,44 +638,19 @@ static int tegra_plane_update(struct drm_plane *plane, struct drm_crtc *crtc,
 	return tegra_dc_setup_window(dc, p->index, &window);
 }
 
-static int tegra_plane_disable(struct drm_plane *plane)
-{
-	struct tegra_dc *dc = to_tegra_dc(plane->crtc);
-	struct tegra_plane *p = to_tegra_plane(plane);
-	unsigned long value;
-
-	if (!plane->crtc)
-		return 0;
-
-	value = WINDOW_A_SELECT << p->index;
-	tegra_dc_writel(dc, value, DC_CMD_DISPLAY_WINDOW_HEADER);
-
-	value = tegra_dc_readl(dc, DC_WIN_WIN_OPTIONS);
-	value &= ~WIN_ENABLE;
-	tegra_dc_writel(dc, value, DC_WIN_WIN_OPTIONS);
-
-	tegra_dc_writel(dc, WIN_A_UPDATE << p->index, DC_CMD_STATE_CONTROL);
-	tegra_dc_writel(dc, WIN_A_ACT_REQ << p->index, DC_CMD_STATE_CONTROL);
-
-	return 0;
-}
-
-static void tegra_plane_destroy(struct drm_plane *plane)
+static void tegra_overlay_plane_destroy(struct drm_plane *plane)
 {
-	struct tegra_plane *p = to_tegra_plane(plane);
-
-	tegra_plane_disable(plane);
-	drm_plane_cleanup(plane);
-	kfree(p);
+	tegra_window_plane_disable(plane);
+	tegra_plane_destroy(plane);
 }
 
-static const struct drm_plane_funcs tegra_plane_funcs = {
-	.update_plane = tegra_plane_update,
-	.disable_plane = tegra_plane_disable,
-	.destroy = tegra_plane_destroy,
+static const struct drm_plane_funcs tegra_overlay_plane_funcs = {
+	.update_plane = tegra_overlay_plane_update,
+	.disable_plane = tegra_window_plane_disable,
+	.destroy = tegra_overlay_plane_destroy,
 };
 
-static const uint32_t plane_formats[] = {
+static const uint32_t tegra_overlay_plane_formats[] = {
 	DRM_FORMAT_XBGR8888,
 	DRM_FORMAT_XRGB8888,
 	DRM_FORMAT_RGB565,
@@ -406,27 +660,44 @@ static const uint32_t plane_formats[] = {
 	DRM_FORMAT_YUV422,
 };
 
-static int tegra_dc_add_planes(struct drm_device *drm, struct tegra_dc *dc)
+static struct drm_plane *tegra_dc_overlay_plane_create(struct drm_device *drm,
+						       struct tegra_dc *dc,
+						       unsigned int index)
 {
-	unsigned int i;
-	int err = 0;
+	struct tegra_plane *plane;
+	unsigned int num_formats;
+	const u32 *formats;
+	int err;
 
-	for (i = 0; i < 2; i++) {
-		struct tegra_plane *plane;
+	plane = kzalloc(sizeof(*plane), GFP_KERNEL);
+	if (!plane)
+		return ERR_PTR(-ENOMEM);
 
-		plane = kzalloc(sizeof(*plane), GFP_KERNEL);
-		if (!plane)
-			return -ENOMEM;
+	plane->index = index;
 
-		plane->index = 1 + i;
+	num_formats = ARRAY_SIZE(tegra_overlay_plane_formats);
+	formats = tegra_overlay_plane_formats;
 
-		err = drm_plane_init(drm, &plane->base, 1 << dc->pipe,
-				     &tegra_plane_funcs, plane_formats,
-				     ARRAY_SIZE(plane_formats), false);
-		if (err < 0) {
-			kfree(plane);
-			return err;
-		}
+	err = drm_universal_plane_init(drm, &plane->base, 1 << dc->pipe,
+				       &tegra_overlay_plane_funcs, formats,
+				       num_formats, DRM_PLANE_TYPE_OVERLAY);
+	if (err < 0) {
+		kfree(plane);
+		return ERR_PTR(err);
+	}
+
+	return &plane->base;
+}
+
+static int tegra_dc_add_planes(struct drm_device *drm, struct tegra_dc *dc)
+{
+	struct drm_plane *plane;
+	unsigned int i;
+
+	for (i = 0; i < 2; i++) {
+		plane = tegra_dc_overlay_plane_create(drm, dc, 1 + i);
+		if (IS_ERR(plane))
+			return PTR_ERR(plane);
 	}
 
 	return 0;
@@ -438,14 +709,16 @@ static int tegra_dc_set_base(struct tegra_dc *dc, int x, int y,
 	struct tegra_bo *bo = tegra_fb_get_plane(fb, 0);
 	unsigned int h_offset = 0, v_offset = 0;
 	struct tegra_bo_tiling tiling;
+	unsigned long value, flags;
 	unsigned int format, swap;
-	unsigned long value;
 	int err;
 
 	err = tegra_fb_get_tiling(fb, &tiling);
 	if (err < 0)
 		return err;
 
+	spin_lock_irqsave(&dc->lock, flags);
+
 	tegra_dc_writel(dc, WINDOW_A_SELECT, DC_CMD_DISPLAY_WINDOW_HEADER);
 
 	value = fb->offsets[0] + y * fb->pitches[0] +
@@ -491,6 +764,7 @@ static int tegra_dc_set_base(struct tegra_dc *dc, int x, int y,
 
 		case TEGRA_BO_TILING_MODE_BLOCK:
 			DRM_ERROR("hardware doesn't support block linear mode\n");
+			spin_unlock_irqrestore(&dc->lock, flags);
 			return -EINVAL;
 		}
 
@@ -513,12 +787,12 @@ static int tegra_dc_set_base(struct tegra_dc *dc, int x, int y,
 	tegra_dc_writel(dc, h_offset, DC_WINBUF_ADDR_H_OFFSET);
 	tegra_dc_writel(dc, v_offset, DC_WINBUF_ADDR_V_OFFSET);
 
-	value = GENERAL_UPDATE | WIN_A_UPDATE;
-	tegra_dc_writel(dc, value, DC_CMD_STATE_CONTROL);
-
 	value = GENERAL_ACT_REQ | WIN_A_ACT_REQ;
+	tegra_dc_writel(dc, value << 8, DC_CMD_STATE_CONTROL);
 	tegra_dc_writel(dc, value, DC_CMD_STATE_CONTROL);
 
+	spin_unlock_irqrestore(&dc->lock, flags);
+
 	return 0;
 }
 
@@ -548,109 +822,6 @@ void tegra_dc_disable_vblank(struct tegra_dc *dc)
 	spin_unlock_irqrestore(&dc->lock, flags);
 }
 
-static int tegra_dc_cursor_set2(struct drm_crtc *crtc, struct drm_file *file,
-				uint32_t handle, uint32_t width,
-				uint32_t height, int32_t hot_x, int32_t hot_y)
-{
-	unsigned long value = CURSOR_CLIP_DISPLAY;
-	struct tegra_dc *dc = to_tegra_dc(crtc);
-	struct drm_gem_object *gem;
-	struct tegra_bo *bo = NULL;
-
-	if (!dc->soc->supports_cursor)
-		return -ENXIO;
-
-	if (width != height)
-		return -EINVAL;
-
-	switch (width) {
-	case 32:
-		value |= CURSOR_SIZE_32x32;
-		break;
-
-	case 64:
-		value |= CURSOR_SIZE_64x64;
-		break;
-
-	case 128:
-		value |= CURSOR_SIZE_128x128;
-
-	case 256:
-		value |= CURSOR_SIZE_256x256;
-		break;
-
-	default:
-		return -EINVAL;
-	}
-
-	if (handle) {
-		gem = drm_gem_object_lookup(crtc->dev, file, handle);
-		if (!gem)
-			return -ENOENT;
-
-		bo = to_tegra_bo(gem);
-	}
-
-	if (bo) {
-		unsigned long addr = (bo->paddr & 0xfffffc00) >> 10;
-#ifdef CONFIG_ARCH_DMA_ADDR_T_64BIT
-		unsigned long high = (bo->paddr & 0xfffffffc) >> 32;
-#endif
-
-		tegra_dc_writel(dc, value | addr, DC_DISP_CURSOR_START_ADDR);
-
-#ifdef CONFIG_ARCH_DMA_ADDR_T_64BIT
-		tegra_dc_writel(dc, high, DC_DISP_CURSOR_START_ADDR_HI);
-#endif
-
-		value = tegra_dc_readl(dc, DC_DISP_DISP_WIN_OPTIONS);
-		value |= CURSOR_ENABLE;
-		tegra_dc_writel(dc, value, DC_DISP_DISP_WIN_OPTIONS);
-
-		value = tegra_dc_readl(dc, DC_DISP_BLEND_CURSOR_CONTROL);
-		value &= ~CURSOR_DST_BLEND_MASK;
-		value &= ~CURSOR_SRC_BLEND_MASK;
-		value |= CURSOR_MODE_NORMAL;
-		value |= CURSOR_DST_BLEND_NEG_K1_TIMES_SRC;
-		value |= CURSOR_SRC_BLEND_K1_TIMES_SRC;
-		value |= CURSOR_ALPHA;
-		tegra_dc_writel(dc, value, DC_DISP_BLEND_CURSOR_CONTROL);
-	} else {
-		value = tegra_dc_readl(dc, DC_DISP_DISP_WIN_OPTIONS);
-		value &= ~CURSOR_ENABLE;
-		tegra_dc_writel(dc, value, DC_DISP_DISP_WIN_OPTIONS);
-	}
-
-	tegra_dc_writel(dc, CURSOR_ACT_REQ << 8, DC_CMD_STATE_CONTROL);
-	tegra_dc_writel(dc, CURSOR_ACT_REQ, DC_CMD_STATE_CONTROL);
-
-	tegra_dc_writel(dc, GENERAL_ACT_REQ << 8, DC_CMD_STATE_CONTROL);
-	tegra_dc_writel(dc, GENERAL_ACT_REQ, DC_CMD_STATE_CONTROL);
-
-	return 0;
-}
-
-static int tegra_dc_cursor_move(struct drm_crtc *crtc, int x, int y)
-{
-	struct tegra_dc *dc = to_tegra_dc(crtc);
-	unsigned long value;
-
-	if (!dc->soc->supports_cursor)
-		return -ENXIO;
-
-	value = ((y & 0x3fff) << 16) | (x & 0x3fff);
-	tegra_dc_writel(dc, value, DC_DISP_CURSOR_POSITION);
-
-	tegra_dc_writel(dc, CURSOR_ACT_REQ << 8, DC_CMD_STATE_CONTROL);
-	tegra_dc_writel(dc, CURSOR_ACT_REQ, DC_CMD_STATE_CONTROL);
-
-	/* XXX: only required on generations earlier than Tegra124? */
-	tegra_dc_writel(dc, GENERAL_ACT_REQ << 8, DC_CMD_STATE_CONTROL);
-	tegra_dc_writel(dc, GENERAL_ACT_REQ, DC_CMD_STATE_CONTROL);
-
-	return 0;
-}
-
 static void tegra_dc_finish_page_flip(struct tegra_dc *dc)
 {
 	struct drm_device *drm = dc->base.dev;
@@ -658,23 +829,32 @@ static void tegra_dc_finish_page_flip(struct tegra_dc *dc)
 	unsigned long flags, base;
 	struct tegra_bo *bo;
 
-	if (!dc->event)
+	spin_lock_irqsave(&drm->event_lock, flags);
+
+	if (!dc->event) {
+		spin_unlock_irqrestore(&drm->event_lock, flags);
 		return;
+	}
 
 	bo = tegra_fb_get_plane(crtc->primary->fb, 0);
 
+	spin_lock_irqsave(&dc->lock, flags);
+
 	/* check if new start address has been latched */
+	tegra_dc_writel(dc, WINDOW_A_SELECT, DC_CMD_DISPLAY_WINDOW_HEADER);
 	tegra_dc_writel(dc, READ_MUX, DC_CMD_STATE_ACCESS);
 	base = tegra_dc_readl(dc, DC_WINBUF_START_ADDR);
 	tegra_dc_writel(dc, 0, DC_CMD_STATE_ACCESS);
 
+	spin_unlock_irqrestore(&dc->lock, flags);
+
 	if (base == bo->paddr + crtc->primary->fb->offsets[0]) {
-		spin_lock_irqsave(&drm->event_lock, flags);
-		drm_send_vblank_event(drm, dc->pipe, dc->event);
-		drm_vblank_put(drm, dc->pipe);
+		drm_crtc_send_vblank_event(crtc, dc->event);
+		drm_crtc_vblank_put(crtc);
 		dc->event = NULL;
-		spin_unlock_irqrestore(&drm->event_lock, flags);
 	}
+
+	spin_unlock_irqrestore(&drm->event_lock, flags);
 }
 
 void tegra_dc_cancel_page_flip(struct drm_crtc *crtc, struct drm_file *file)
@@ -687,7 +867,7 @@ void tegra_dc_cancel_page_flip(struct drm_crtc *crtc, struct drm_file *file)
 
 	if (dc->event && dc->event->base.file_priv == file) {
 		dc->event->base.destroy(&dc->event->base);
-		drm_vblank_put(drm, dc->pipe);
+		drm_crtc_vblank_put(crtc);
 		dc->event = NULL;
 	}
 
@@ -697,16 +877,16 @@ void tegra_dc_cancel_page_flip(struct drm_crtc *crtc, struct drm_file *file)
 static int tegra_dc_page_flip(struct drm_crtc *crtc, struct drm_framebuffer *fb,
 			      struct drm_pending_vblank_event *event, uint32_t page_flip_flags)
 {
+	unsigned int pipe = drm_crtc_index(crtc);
 	struct tegra_dc *dc = to_tegra_dc(crtc);
-	struct drm_device *drm = crtc->dev;
 
 	if (dc->event)
 		return -EBUSY;
 
 	if (event) {
-		event->pipe = dc->pipe;
+		event->pipe = pipe;
 		dc->event = event;
-		drm_vblank_get(drm, dc->pipe);
+		drm_crtc_vblank_get(crtc);
 	}
 
 	tegra_dc_set_base(dc, 0, 0, fb);
@@ -727,8 +907,6 @@ static void tegra_dc_destroy(struct drm_crtc *crtc)
 }
 
 static const struct drm_crtc_funcs tegra_crtc_funcs = {
-	.cursor_set2 = tegra_dc_cursor_set2,
-	.cursor_move = tegra_dc_cursor_move,
 	.page_flip = tegra_dc_page_flip,
 	.set_config = drm_crtc_helper_set_config,
 	.destroy = tegra_dc_destroy,
@@ -742,7 +920,7 @@ static void tegra_crtc_disable(struct drm_crtc *crtc)
 
 	drm_for_each_legacy_plane(plane, &drm->mode_config.plane_list) {
 		if (plane->crtc == crtc) {
-			tegra_plane_disable(plane);
+			tegra_window_plane_disable(plane);
 			plane->crtc = NULL;
 
 			if (plane->fb) {
@@ -752,7 +930,8 @@ static void tegra_crtc_disable(struct drm_crtc *crtc)
 		}
 	}
 
-	drm_vblank_off(drm, dc->pipe);
+	drm_crtc_vblank_off(crtc);
+	tegra_dc_commit(dc);
 }
 
 static bool tegra_crtc_mode_fixup(struct drm_crtc *crtc,
@@ -841,8 +1020,6 @@ static int tegra_crtc_mode_set(struct drm_crtc *crtc,
 	u32 value;
 	int err;
 
-	drm_vblank_pre_modeset(crtc->dev, dc->pipe);
-
 	err = tegra_crtc_setup_clk(crtc, mode);
 	if (err) {
 		dev_err(dc->dev, "failed to setup clock for CRTC: %d\n", err);
@@ -896,6 +1073,8 @@ static void tegra_crtc_prepare(struct drm_crtc *crtc)
 	unsigned int syncpt;
 	unsigned long value;
 
+	drm_crtc_vblank_off(crtc);
+
 	/* hardware initialization */
 	reset_control_deassert(dc->rst);
 	usleep_range(10000, 20000);
@@ -935,15 +1114,9 @@ static void tegra_crtc_prepare(struct drm_crtc *crtc)
 static void tegra_crtc_commit(struct drm_crtc *crtc)
 {
 	struct tegra_dc *dc = to_tegra_dc(crtc);
-	unsigned long value;
-
-	value = GENERAL_UPDATE | WIN_A_UPDATE;
-	tegra_dc_writel(dc, value, DC_CMD_STATE_CONTROL);
 
-	value = GENERAL_ACT_REQ | WIN_A_ACT_REQ;
-	tegra_dc_writel(dc, value, DC_CMD_STATE_CONTROL);
-
-	drm_vblank_post_modeset(crtc->dev, dc->pipe);
+	drm_crtc_vblank_on(crtc);
+	tegra_dc_commit(dc);
 }
 
 static void tegra_crtc_load_lut(struct drm_crtc *crtc)
@@ -978,7 +1151,7 @@ static irqreturn_t tegra_dc_irq(int irq, void *data)
 		/*
 		dev_dbg(dc->dev, "%s(): vertical blank\n", __func__);
 		*/
-		drm_handle_vblank(dc->base.dev, dc->pipe);
+		drm_crtc_handle_vblank(&dc->base);
 		tegra_dc_finish_page_flip(dc);
 	}
 
@@ -997,7 +1170,7 @@ static int tegra_dc_show_regs(struct seq_file *s, void *data)
 	struct tegra_dc *dc = node->info_ent->data;
 
 #define DUMP_REG(name)						\
-	seq_printf(s, "%-40s %#05x %08lx\n", #name, name,	\
+	seq_printf(s, "%-40s %#05x %08x\n", #name, name,	\
 		   tegra_dc_readl(dc, name))
 
 	DUMP_REG(DC_CMD_GENERAL_INCR_SYNCPT);
@@ -1285,9 +1458,40 @@ static int tegra_dc_init(struct host1x_client *client)
 	struct drm_device *drm = dev_get_drvdata(client->parent);
 	struct tegra_dc *dc = host1x_client_to_dc(client);
 	struct tegra_drm *tegra = drm->dev_private;
+	struct drm_plane *primary = NULL;
+	struct drm_plane *cursor = NULL;
 	int err;
 
-	drm_crtc_init(drm, &dc->base, &tegra_crtc_funcs);
+	if (tegra->domain) {
+		err = iommu_attach_device(tegra->domain, dc->dev);
+		if (err < 0) {
+			dev_err(dc->dev, "failed to attach to domain: %d\n",
+				err);
+			return err;
+		}
+
+		dc->domain = tegra->domain;
+	}
+
+	primary = tegra_dc_primary_plane_create(drm, dc);
+	if (IS_ERR(primary)) {
+		err = PTR_ERR(primary);
+		goto cleanup;
+	}
+
+	if (dc->soc->supports_cursor) {
+		cursor = tegra_dc_cursor_plane_create(drm, dc);
+		if (IS_ERR(cursor)) {
+			err = PTR_ERR(cursor);
+			goto cleanup;
+		}
+	}
+
+	err = drm_crtc_init_with_planes(drm, &dc->base, primary, cursor,
+					&tegra_crtc_funcs);
+	if (err < 0)
+		goto cleanup;
+
 	drm_mode_crtc_set_gamma_size(&dc->base, 256);
 	drm_crtc_helper_add(&dc->base, &tegra_crtc_helper_funcs);
 
@@ -1301,12 +1505,12 @@ static int tegra_dc_init(struct host1x_client *client)
 	err = tegra_dc_rgb_init(drm, dc);
 	if (err < 0 && err != -ENODEV) {
 		dev_err(dc->dev, "failed to initialize RGB output: %d\n", err);
-		return err;
+		goto cleanup;
 	}
 
 	err = tegra_dc_add_planes(drm, dc);
 	if (err < 0)
-		return err;
+		goto cleanup;
 
 	if (IS_ENABLED(CONFIG_DEBUG_FS)) {
 		err = tegra_dc_debugfs_init(dc, drm->primary);
@@ -1319,10 +1523,24 @@ static int tegra_dc_init(struct host1x_client *client)
 	if (err < 0) {
 		dev_err(dc->dev, "failed to request IRQ#%u: %d\n", dc->irq,
 			err);
-		return err;
+		goto cleanup;
 	}
 
 	return 0;
+
+cleanup:
+	if (cursor)
+		drm_plane_cleanup(cursor);
+
+	if (primary)
+		drm_plane_cleanup(primary);
+
+	if (tegra->domain) {
+		iommu_detach_device(tegra->domain, dc->dev);
+		dc->domain = NULL;
+	}
+
+	return err;
 }
 
 static int tegra_dc_exit(struct host1x_client *client)
@@ -1344,6 +1562,11 @@ static int tegra_dc_exit(struct host1x_client *client)
 		return err;
 	}
 
+	if (dc->domain) {
+		iommu_detach_device(dc->domain, dc->dev);
+		dc->domain = NULL;
+	}
+
 	return 0;
 }
 
@@ -1357,6 +1580,7 @@ static const struct tegra_dc_soc_info tegra20_dc_soc_info = {
 	.supports_cursor = false,
 	.supports_block_linear = false,
 	.pitch_align = 8,
+	.has_powergate = false,
 };
 
 static const struct tegra_dc_soc_info tegra30_dc_soc_info = {
@@ -1364,6 +1588,7 @@ static const struct tegra_dc_soc_info tegra30_dc_soc_info = {
 	.supports_cursor = false,
 	.supports_block_linear = false,
 	.pitch_align = 8,
+	.has_powergate = false,
 };
 
 static const struct tegra_dc_soc_info tegra114_dc_soc_info = {
@@ -1371,6 +1596,7 @@ static const struct tegra_dc_soc_info tegra114_dc_soc_info = {
 	.supports_cursor = false,
 	.supports_block_linear = false,
 	.pitch_align = 64,
+	.has_powergate = true,
 };
 
 static const struct tegra_dc_soc_info tegra124_dc_soc_info = {
@@ -1378,6 +1604,7 @@ static const struct tegra_dc_soc_info tegra124_dc_soc_info = {
 	.supports_cursor = true,
 	.supports_block_linear = true,
 	.pitch_align = 64,
+	.has_powergate = true,
 };
 
 static const struct of_device_id tegra_dc_of_match[] = {
@@ -1385,6 +1612,9 @@ static const struct of_device_id tegra_dc_of_match[] = {
 		.compatible = "nvidia,tegra124-dc",
 		.data = &tegra124_dc_soc_info,
 	}, {
+		.compatible = "nvidia,tegra114-dc",
+		.data = &tegra114_dc_soc_info,
+	}, {
 		.compatible = "nvidia,tegra30-dc",
 		.data = &tegra30_dc_soc_info,
 	}, {
@@ -1467,9 +1697,34 @@ static int tegra_dc_probe(struct platform_device *pdev)
 		return PTR_ERR(dc->rst);
 	}
 
-	err = clk_prepare_enable(dc->clk);
-	if (err < 0)
-		return err;
+	if (dc->soc->has_powergate) {
+		if (dc->pipe == 0)
+			dc->powergate = TEGRA_POWERGATE_DIS;
+		else
+			dc->powergate = TEGRA_POWERGATE_DISB;
+
+		err = tegra_powergate_sequence_power_up(dc->powergate, dc->clk,
+							dc->rst);
+		if (err < 0) {
+			dev_err(&pdev->dev, "failed to power partition: %d\n",
+				err);
+			return err;
+		}
+	} else {
+		err = clk_prepare_enable(dc->clk);
+		if (err < 0) {
+			dev_err(&pdev->dev, "failed to enable clock: %d\n",
+				err);
+			return err;
+		}
+
+		err = reset_control_deassert(dc->rst);
+		if (err < 0) {
+			dev_err(&pdev->dev, "failed to deassert reset: %d\n",
+				err);
+			return err;
+		}
+	}
 
 	regs = platform_get_resource(pdev, IORESOURCE_MEM, 0);
 	dc->regs = devm_ioremap_resource(&pdev->dev, regs);
@@ -1523,6 +1778,10 @@ static int tegra_dc_remove(struct platform_device *pdev)
 	}
 
 	reset_control_assert(dc->rst);
+
+	if (dc->soc->has_powergate)
+		tegra_powergate_power_off(dc->powergate);
+
 	clk_disable_unprepare(dc->clk);
 
 	return 0;
diff --git a/drivers/gpu/drm/tegra/dpaux.c b/drivers/gpu/drm/tegra/dpaux.c
index 708f783ead47..d6b55e3e3716 100644
--- a/drivers/gpu/drm/tegra/dpaux.c
+++ b/drivers/gpu/drm/tegra/dpaux.c
@@ -533,9 +533,9 @@ int tegra_dpaux_train(struct tegra_dpaux *dpaux, struct drm_dp_link *link,
 
 	for (i = 0; i < link->num_lanes; i++)
 		values[i] = DP_TRAIN_MAX_PRE_EMPHASIS_REACHED |
-			    DP_TRAIN_PRE_EMPHASIS_0 |
+			    DP_TRAIN_PRE_EMPH_LEVEL_0 |
 			    DP_TRAIN_MAX_SWING_REACHED |
-			    DP_TRAIN_VOLTAGE_SWING_400;
+			    DP_TRAIN_VOLTAGE_SWING_LEVEL_0;
 
 	err = drm_dp_dpcd_write(&dpaux->aux, DP_TRAINING_LANE0_SET, values,
 				link->num_lanes);
diff --git a/drivers/gpu/drm/tegra/drm.c b/drivers/gpu/drm/tegra/drm.c
index 59736bb810cd..d4f827593dfa 100644
--- a/drivers/gpu/drm/tegra/drm.c
+++ b/drivers/gpu/drm/tegra/drm.c
@@ -8,6 +8,7 @@
  */
 
 #include <linux/host1x.h>
+#include <linux/iommu.h>
 
 #include "drm.h"
 #include "gem.h"
@@ -33,6 +34,17 @@ static int tegra_drm_load(struct drm_device *drm, unsigned long flags)
 	if (!tegra)
 		return -ENOMEM;
 
+	if (iommu_present(&platform_bus_type)) {
+		tegra->domain = iommu_domain_alloc(&platform_bus_type);
+		if (IS_ERR(tegra->domain)) {
+			err = PTR_ERR(tegra->domain);
+			goto free;
+		}
+
+		DRM_DEBUG("IOMMU context initialized\n");
+		drm_mm_init(&tegra->mm, 0, SZ_2G);
+	}
+
 	mutex_init(&tegra->clients_lock);
 	INIT_LIST_HEAD(&tegra->clients);
 	drm->dev_private = tegra;
@@ -42,13 +54,13 @@ static int tegra_drm_load(struct drm_device *drm, unsigned long flags)
 
 	err = tegra_drm_fb_prepare(drm);
 	if (err < 0)
-		return err;
+		goto config;
 
 	drm_kms_helper_poll_init(drm);
 
 	err = host1x_device_init(device);
 	if (err < 0)
-		return err;
+		goto fbdev;
 
 	/*
 	 * We don't use the drm_irq_install() helpers provided by the DRM
@@ -59,18 +71,37 @@ static int tegra_drm_load(struct drm_device *drm, unsigned long flags)
 
 	err = drm_vblank_init(drm, drm->mode_config.num_crtc);
 	if (err < 0)
-		return err;
+		goto device;
 
 	err = tegra_drm_fb_init(drm);
 	if (err < 0)
-		return err;
+		goto vblank;
 
 	return 0;
+
+vblank:
+	drm_vblank_cleanup(drm);
+device:
+	host1x_device_exit(device);
+fbdev:
+	drm_kms_helper_poll_fini(drm);
+	tegra_drm_fb_free(drm);
+config:
+	drm_mode_config_cleanup(drm);
+
+	if (tegra->domain) {
+		iommu_domain_free(tegra->domain);
+		drm_mm_takedown(&tegra->mm);
+	}
+free:
+	kfree(tegra);
+	return err;
 }
 
 static int tegra_drm_unload(struct drm_device *drm)
 {
 	struct host1x_device *device = to_host1x_device(drm->dev);
+	struct tegra_drm *tegra = drm->dev_private;
 	int err;
 
 	drm_kms_helper_poll_fini(drm);
@@ -82,6 +113,13 @@ static int tegra_drm_unload(struct drm_device *drm)
 	if (err < 0)
 		return err;
 
+	if (tegra->domain) {
+		iommu_domain_free(tegra->domain);
+		drm_mm_takedown(&tegra->mm);
+	}
+
+	kfree(tegra);
+
 	return 0;
 }
 
@@ -656,24 +694,28 @@ static const struct file_operations tegra_drm_fops = {
 	.llseek = noop_llseek,
 };
 
-static struct drm_crtc *tegra_crtc_from_pipe(struct drm_device *drm, int pipe)
+static struct drm_crtc *tegra_crtc_from_pipe(struct drm_device *drm,
+					     unsigned int pipe)
 {
 	struct drm_crtc *crtc;
 
 	list_for_each_entry(crtc, &drm->mode_config.crtc_list, head) {
-		struct tegra_dc *dc = to_tegra_dc(crtc);
-
-		if (dc->pipe == pipe)
+		if (pipe == drm_crtc_index(crtc))
 			return crtc;
 	}
 
 	return NULL;
 }
 
-static u32 tegra_drm_get_vblank_counter(struct drm_device *dev, int crtc)
+static u32 tegra_drm_get_vblank_counter(struct drm_device *drm, int pipe)
 {
+	struct drm_crtc *crtc = tegra_crtc_from_pipe(drm, pipe);
+
+	if (!crtc)
+		return 0;
+
 	/* TODO: implement real hardware counter using syncpoints */
-	return drm_vblank_count(dev, crtc);
+	return drm_crtc_vblank_count(crtc);
 }
 
 static int tegra_drm_enable_vblank(struct drm_device *drm, int pipe)
diff --git a/drivers/gpu/drm/tegra/drm.h b/drivers/gpu/drm/tegra/drm.h
index e89c70fa82d5..3a3b2e7b5b3f 100644
--- a/drivers/gpu/drm/tegra/drm.h
+++ b/drivers/gpu/drm/tegra/drm.h
@@ -39,6 +39,9 @@ struct tegra_fbdev {
 struct tegra_drm {
 	struct drm_device *drm;
 
+	struct iommu_domain *domain;
+	struct drm_mm mm;
+
 	struct mutex clients_lock;
 	struct list_head clients;
 
@@ -101,6 +104,7 @@ struct tegra_dc {
 	spinlock_t lock;
 
 	struct drm_crtc base;
+	int powergate;
 	int pipe;
 
 	struct clk *clk;
@@ -120,6 +124,8 @@ struct tegra_dc {
 	struct drm_pending_vblank_event *event;
 
 	const struct tegra_dc_soc_info *soc;
+
+	struct iommu_domain *domain;
 };
 
 static inline struct tegra_dc *
@@ -133,16 +139,15 @@ static inline struct tegra_dc *to_tegra_dc(struct drm_crtc *crtc)
 	return crtc ? container_of(crtc, struct tegra_dc, base) : NULL;
 }
 
-static inline void tegra_dc_writel(struct tegra_dc *dc, unsigned long value,
-				   unsigned long reg)
+static inline void tegra_dc_writel(struct tegra_dc *dc, u32 value,
+				   unsigned long offset)
 {
-	writel(value, dc->regs + (reg << 2));
+	writel(value, dc->regs + (offset << 2));
 }
 
-static inline unsigned long tegra_dc_readl(struct tegra_dc *dc,
-					   unsigned long reg)
+static inline u32 tegra_dc_readl(struct tegra_dc *dc, unsigned long offset)
 {
-	return readl(dc->regs + (reg << 2));
+	return readl(dc->regs + (offset << 2));
 }
 
 struct tegra_dc_window {
@@ -287,6 +292,7 @@ bool tegra_fb_is_bottom_up(struct drm_framebuffer *framebuffer);
 int tegra_fb_get_tiling(struct drm_framebuffer *framebuffer,
 			struct tegra_bo_tiling *tiling);
 int tegra_drm_fb_prepare(struct drm_device *drm);
+void tegra_drm_fb_free(struct drm_device *drm);
 int tegra_drm_fb_init(struct drm_device *drm);
 void tegra_drm_fb_exit(struct drm_device *drm);
 #ifdef CONFIG_DRM_TEGRA_FBDEV
diff --git a/drivers/gpu/drm/tegra/dsi.c b/drivers/gpu/drm/tegra/dsi.c
index f7874458926a..33f67fd601c6 100644
--- a/drivers/gpu/drm/tegra/dsi.c
+++ b/drivers/gpu/drm/tegra/dsi.c
@@ -11,6 +11,7 @@
 #include <linux/host1x.h>
 #include <linux/module.h>
 #include <linux/of.h>
+#include <linux/of_platform.h>
 #include <linux/platform_device.h>
 #include <linux/reset.h>
 
@@ -26,9 +27,6 @@
 #include "dsi.h"
 #include "mipi-phy.h"
 
-#define DSI_VIDEO_FIFO_DEPTH (1920 / 4)
-#define DSI_HOST_FIFO_DEPTH 64
-
 struct tegra_dsi {
 	struct host1x_client client;
 	struct tegra_output output;
@@ -54,6 +52,13 @@ struct tegra_dsi {
 
 	struct regulator *vdd;
 	bool enabled;
+
+	unsigned int video_fifo_depth;
+	unsigned int host_fifo_depth;
+
+	/* for ganged-mode support */
+	struct tegra_dsi *master;
+	struct tegra_dsi *slave;
 };
 
 static inline struct tegra_dsi *
@@ -318,6 +323,21 @@ static const u32 pkt_seq_video_non_burst_sync_events[NUM_PKT_SEQ] = {
 	[11] = PKT_ID0(MIPI_DSI_BLANKING_PACKET) | PKT_LEN0(4),
 };
 
+static const u32 pkt_seq_command_mode[NUM_PKT_SEQ] = {
+	[ 0] = 0,
+	[ 1] = 0,
+	[ 2] = 0,
+	[ 3] = 0,
+	[ 4] = 0,
+	[ 5] = 0,
+	[ 6] = PKT_ID0(MIPI_DSI_DCS_LONG_WRITE) | PKT_LEN0(3) | PKT_LP,
+	[ 7] = 0,
+	[ 8] = 0,
+	[ 9] = 0,
+	[10] = PKT_ID0(MIPI_DSI_DCS_LONG_WRITE) | PKT_LEN0(5) | PKT_LP,
+	[11] = 0,
+};
+
 static int tegra_dsi_set_phy_timing(struct tegra_dsi *dsi)
 {
 	struct mipi_dphy_timing timing;
@@ -329,7 +349,7 @@ static int tegra_dsi_set_phy_timing(struct tegra_dsi *dsi)
 	if (rate < 0)
 		return rate;
 
-	period = DIV_ROUND_CLOSEST(1000000000UL, rate * 2);
+	period = DIV_ROUND_CLOSEST(NSEC_PER_SEC, rate * 2);
 
 	err = mipi_dphy_timing_get_default(&timing, period);
 	if (err < 0)
@@ -369,6 +389,9 @@ static int tegra_dsi_set_phy_timing(struct tegra_dsi *dsi)
 		DSI_TIMING_FIELD(timing.tago, period, 1);
 	tegra_dsi_writel(dsi, value, DSI_BTA_TIMING);
 
+	if (dsi->slave)
+		return tegra_dsi_set_phy_timing(dsi->slave);
+
 	return 0;
 }
 
@@ -426,26 +449,59 @@ static int tegra_dsi_get_format(enum mipi_dsi_pixel_format format,
 	return 0;
 }
 
-static int tegra_output_dsi_enable(struct tegra_output *output)
+static void tegra_dsi_ganged_enable(struct tegra_dsi *dsi, unsigned int start,
+				    unsigned int size)
+{
+	u32 value;
+
+	tegra_dsi_writel(dsi, start, DSI_GANGED_MODE_START);
+	tegra_dsi_writel(dsi, size << 16 | size, DSI_GANGED_MODE_SIZE);
+
+	value = DSI_GANGED_MODE_CONTROL_ENABLE;
+	tegra_dsi_writel(dsi, value, DSI_GANGED_MODE_CONTROL);
+}
+
+static void tegra_dsi_enable(struct tegra_dsi *dsi)
+{
+	u32 value;
+
+	value = tegra_dsi_readl(dsi, DSI_POWER_CONTROL);
+	value |= DSI_POWER_CONTROL_ENABLE;
+	tegra_dsi_writel(dsi, value, DSI_POWER_CONTROL);
+
+	if (dsi->slave)
+		tegra_dsi_enable(dsi->slave);
+}
+
+static unsigned int tegra_dsi_get_lanes(struct tegra_dsi *dsi)
+{
+	if (dsi->master)
+		return dsi->master->lanes + dsi->lanes;
+
+	if (dsi->slave)
+		return dsi->lanes + dsi->slave->lanes;
+
+	return dsi->lanes;
+}
+
+static int tegra_dsi_configure(struct tegra_dsi *dsi, unsigned int pipe,
+			       const struct drm_display_mode *mode)
 {
-	struct tegra_dc *dc = to_tegra_dc(output->encoder.crtc);
-	struct drm_display_mode *mode = &dc->base.mode;
 	unsigned int hact, hsw, hbp, hfp, i, mul, div;
-	struct tegra_dsi *dsi = to_dsi(output);
 	enum tegra_dsi_format format;
-	unsigned long value;
 	const u32 *pkt_seq;
+	u32 value;
 	int err;
 
-	if (dsi->enabled)
-		return 0;
-
 	if (dsi->flags & MIPI_DSI_MODE_VIDEO_SYNC_PULSE) {
 		DRM_DEBUG_KMS("Non-burst video mode with sync pulses\n");
 		pkt_seq = pkt_seq_video_non_burst_sync_pulses;
-	} else {
+	} else if (dsi->flags & MIPI_DSI_MODE_VIDEO) {
 		DRM_DEBUG_KMS("Non-burst video mode with sync events\n");
 		pkt_seq = pkt_seq_video_non_burst_sync_events;
+	} else {
+		DRM_DEBUG_KMS("Command mode\n");
+		pkt_seq = pkt_seq_command_mode;
 	}
 
 	err = tegra_dsi_get_muldiv(dsi->format, &mul, &div);
@@ -456,61 +512,136 @@ static int tegra_output_dsi_enable(struct tegra_output *output)
 	if (err < 0)
 		return err;
 
-	err = clk_enable(dsi->clk);
-	if (err < 0)
-		return err;
-
-	reset_control_deassert(dsi->rst);
-
 	value = DSI_CONTROL_CHANNEL(0) | DSI_CONTROL_FORMAT(format) |
 		DSI_CONTROL_LANES(dsi->lanes - 1) |
-		DSI_CONTROL_SOURCE(dc->pipe);
+		DSI_CONTROL_SOURCE(pipe);
 	tegra_dsi_writel(dsi, value, DSI_CONTROL);
 
-	tegra_dsi_writel(dsi, DSI_VIDEO_FIFO_DEPTH, DSI_MAX_THRESHOLD);
+	tegra_dsi_writel(dsi, dsi->video_fifo_depth, DSI_MAX_THRESHOLD);
 
-	value = DSI_HOST_CONTROL_HS | DSI_HOST_CONTROL_CS |
-		DSI_HOST_CONTROL_ECC;
+	value = DSI_HOST_CONTROL_HS;
 	tegra_dsi_writel(dsi, value, DSI_HOST_CONTROL);
 
 	value = tegra_dsi_readl(dsi, DSI_CONTROL);
+
 	if (dsi->flags & MIPI_DSI_CLOCK_NON_CONTINUOUS)
 		value |= DSI_CONTROL_HS_CLK_CTRL;
+
 	value &= ~DSI_CONTROL_TX_TRIG(3);
-	value &= ~DSI_CONTROL_DCS_ENABLE;
+
+	/* enable DCS commands for command mode */
+	if (dsi->flags & MIPI_DSI_MODE_VIDEO)
+		value &= ~DSI_CONTROL_DCS_ENABLE;
+	else
+		value |= DSI_CONTROL_DCS_ENABLE;
+
 	value |= DSI_CONTROL_VIDEO_ENABLE;
 	value &= ~DSI_CONTROL_HOST_ENABLE;
 	tegra_dsi_writel(dsi, value, DSI_CONTROL);
 
-	err = tegra_dsi_set_phy_timing(dsi);
-	if (err < 0)
-		return err;
-
 	for (i = 0; i < NUM_PKT_SEQ; i++)
 		tegra_dsi_writel(dsi, pkt_seq[i], DSI_PKT_SEQ_0_LO + i);
 
-	/* horizontal active pixels */
-	hact = mode->hdisplay * mul / div;
+	if (dsi->flags & MIPI_DSI_MODE_VIDEO) {
+		/* horizontal active pixels */
+		hact = mode->hdisplay * mul / div;
 
-	/* horizontal sync width */
-	hsw = (mode->hsync_end - mode->hsync_start) * mul / div;
-	hsw -= 10;
+		/* horizontal sync width */
+		hsw = (mode->hsync_end - mode->hsync_start) * mul / div;
+		hsw -= 10;
 
-	/* horizontal back porch */
-	hbp = (mode->htotal - mode->hsync_end) * mul / div;
-	hbp -= 14;
+		/* horizontal back porch */
+		hbp = (mode->htotal - mode->hsync_end) * mul / div;
+		hbp -= 14;
 
-	/* horizontal front porch */
-	hfp = (mode->hsync_start  - mode->hdisplay) * mul / div;
-	hfp -= 8;
+		/* horizontal front porch */
+		hfp = (mode->hsync_start - mode->hdisplay) * mul / div;
+		hfp -= 8;
 
-	tegra_dsi_writel(dsi, hsw << 16 | 0, DSI_PKT_LEN_0_1);
-	tegra_dsi_writel(dsi, hact << 16 | hbp, DSI_PKT_LEN_2_3);
-	tegra_dsi_writel(dsi, hfp, DSI_PKT_LEN_4_5);
-	tegra_dsi_writel(dsi, 0x0f0f << 16, DSI_PKT_LEN_6_7);
+		tegra_dsi_writel(dsi, hsw << 16 | 0, DSI_PKT_LEN_0_1);
+		tegra_dsi_writel(dsi, hact << 16 | hbp, DSI_PKT_LEN_2_3);
+		tegra_dsi_writel(dsi, hfp, DSI_PKT_LEN_4_5);
+		tegra_dsi_writel(dsi, 0x0f0f << 16, DSI_PKT_LEN_6_7);
 
-	/* set SOL delay */
-	tegra_dsi_writel(dsi, 8 * mul / div, DSI_SOL_DELAY);
+		/* set SOL delay (for non-burst mode only) */
+		tegra_dsi_writel(dsi, 8 * mul / div, DSI_SOL_DELAY);
+
+		/* TODO: implement ganged mode */
+	} else {
+		u16 bytes;
+
+		if (dsi->master || dsi->slave) {
+			/*
+			 * For ganged mode, assume symmetric left-right mode.
+			 */
+			bytes = 1 + (mode->hdisplay / 2) * mul / div;
+		} else {
+			/* 1 byte (DCS command) + pixel data */
+			bytes = 1 + mode->hdisplay * mul / div;
+		}
+
+		tegra_dsi_writel(dsi, 0, DSI_PKT_LEN_0_1);
+		tegra_dsi_writel(dsi, bytes << 16, DSI_PKT_LEN_2_3);
+		tegra_dsi_writel(dsi, bytes << 16, DSI_PKT_LEN_4_5);
+		tegra_dsi_writel(dsi, 0, DSI_PKT_LEN_6_7);
+
+		value = MIPI_DCS_WRITE_MEMORY_START << 8 |
+			MIPI_DCS_WRITE_MEMORY_CONTINUE;
+		tegra_dsi_writel(dsi, value, DSI_DCS_CMDS);
+
+		/* set SOL delay */
+		if (dsi->master || dsi->slave) {
+			unsigned int lanes = tegra_dsi_get_lanes(dsi);
+			unsigned long delay, bclk, bclk_ganged;
+
+			/* SOL to valid, valid to FIFO and FIFO write delay */
+			delay = 4 + 4 + 2;
+			delay = DIV_ROUND_UP(delay * mul, div * lanes);
+			/* FIFO read delay */
+			delay = delay + 6;
+
+			bclk = DIV_ROUND_UP(mode->htotal * mul, div * lanes);
+			bclk_ganged = DIV_ROUND_UP(bclk * lanes / 2, lanes);
+			value = bclk - bclk_ganged + delay + 20;
+		} else {
+			/* TODO: revisit for non-ganged mode */
+			value = 8 * mul / div;
+		}
+
+		tegra_dsi_writel(dsi, value, DSI_SOL_DELAY);
+	}
+
+	if (dsi->slave) {
+		err = tegra_dsi_configure(dsi->slave, pipe, mode);
+		if (err < 0)
+			return err;
+
+		/*
+		 * TODO: Support modes other than symmetrical left-right
+		 * split.
+		 */
+		tegra_dsi_ganged_enable(dsi, 0, mode->hdisplay / 2);
+		tegra_dsi_ganged_enable(dsi->slave, mode->hdisplay / 2,
+					mode->hdisplay / 2);
+	}
+
+	return 0;
+}
+
+static int tegra_output_dsi_enable(struct tegra_output *output)
+{
+	struct tegra_dc *dc = to_tegra_dc(output->encoder.crtc);
+	const struct drm_display_mode *mode = &dc->base.mode;
+	struct tegra_dsi *dsi = to_dsi(output);
+	u32 value;
+	int err;
+
+	if (dsi->enabled)
+		return 0;
+
+	err = tegra_dsi_configure(dsi, dc->pipe, mode);
+	if (err < 0)
+		return err;
 
 	/* enable display controller */
 	value = tegra_dc_readl(dc, DC_DISP_DISP_WIN_OPTIONS);
@@ -531,28 +662,79 @@ static int tegra_output_dsi_enable(struct tegra_output *output)
 	tegra_dc_writel(dc, GENERAL_ACT_REQ, DC_CMD_STATE_CONTROL);
 
 	/* enable DSI controller */
-	value = tegra_dsi_readl(dsi, DSI_POWER_CONTROL);
-	value |= DSI_POWER_CONTROL_ENABLE;
-	tegra_dsi_writel(dsi, value, DSI_POWER_CONTROL);
+	tegra_dsi_enable(dsi);
 
 	dsi->enabled = true;
 
 	return 0;
 }
 
+static int tegra_dsi_wait_idle(struct tegra_dsi *dsi, unsigned long timeout)
+{
+	u32 value;
+
+	timeout = jiffies + msecs_to_jiffies(timeout);
+
+	while (time_before(jiffies, timeout)) {
+		value = tegra_dsi_readl(dsi, DSI_STATUS);
+		if (value & DSI_STATUS_IDLE)
+			return 0;
+
+		usleep_range(1000, 2000);
+	}
+
+	return -ETIMEDOUT;
+}
+
+static void tegra_dsi_video_disable(struct tegra_dsi *dsi)
+{
+	u32 value;
+
+	value = tegra_dsi_readl(dsi, DSI_CONTROL);
+	value &= ~DSI_CONTROL_VIDEO_ENABLE;
+	tegra_dsi_writel(dsi, value, DSI_CONTROL);
+
+	if (dsi->slave)
+		tegra_dsi_video_disable(dsi->slave);
+}
+
+static void tegra_dsi_ganged_disable(struct tegra_dsi *dsi)
+{
+	tegra_dsi_writel(dsi, 0, DSI_GANGED_MODE_START);
+	tegra_dsi_writel(dsi, 0, DSI_GANGED_MODE_SIZE);
+	tegra_dsi_writel(dsi, 0, DSI_GANGED_MODE_CONTROL);
+}
+
+static void tegra_dsi_disable(struct tegra_dsi *dsi)
+{
+	u32 value;
+
+	if (dsi->slave) {
+		tegra_dsi_ganged_disable(dsi->slave);
+		tegra_dsi_ganged_disable(dsi);
+	}
+
+	value = tegra_dsi_readl(dsi, DSI_POWER_CONTROL);
+	value &= ~DSI_POWER_CONTROL_ENABLE;
+	tegra_dsi_writel(dsi, value, DSI_POWER_CONTROL);
+
+	if (dsi->slave)
+		tegra_dsi_disable(dsi->slave);
+
+	usleep_range(5000, 10000);
+}
+
 static int tegra_output_dsi_disable(struct tegra_output *output)
 {
 	struct tegra_dc *dc = to_tegra_dc(output->encoder.crtc);
 	struct tegra_dsi *dsi = to_dsi(output);
 	unsigned long value;
+	int err;
 
 	if (!dsi->enabled)
 		return 0;
 
-	/* disable DSI controller */
-	value = tegra_dsi_readl(dsi, DSI_POWER_CONTROL);
-	value &= ~DSI_POWER_CONTROL_ENABLE;
-	tegra_dsi_writel(dsi, value, DSI_POWER_CONTROL);
+	tegra_dsi_video_disable(dsi);
 
 	/*
 	 * The following accesses registers of the display controller, so make
@@ -576,39 +758,68 @@ static int tegra_output_dsi_disable(struct tegra_output *output)
 		tegra_dc_writel(dc, GENERAL_ACT_REQ, DC_CMD_STATE_CONTROL);
 	}
 
-	clk_disable(dsi->clk);
+	err = tegra_dsi_wait_idle(dsi, 100);
+	if (err < 0)
+		dev_dbg(dsi->dev, "failed to idle DSI: %d\n", err);
+
+	tegra_dsi_disable(dsi);
 
 	dsi->enabled = false;
 
 	return 0;
 }
 
+static void tegra_dsi_set_timeout(struct tegra_dsi *dsi, unsigned long bclk,
+				  unsigned int vrefresh)
+{
+	unsigned int timeout;
+	u32 value;
+
+	/* one frame high-speed transmission timeout */
+	timeout = (bclk / vrefresh) / 512;
+	value = DSI_TIMEOUT_LRX(0x2000) | DSI_TIMEOUT_HTX(timeout);
+	tegra_dsi_writel(dsi, value, DSI_TIMEOUT_0);
+
+	/* 2 ms peripheral timeout for panel */
+	timeout = 2 * bclk / 512 * 1000;
+	value = DSI_TIMEOUT_PR(timeout) | DSI_TIMEOUT_TA(0x2000);
+	tegra_dsi_writel(dsi, value, DSI_TIMEOUT_1);
+
+	value = DSI_TALLY_TA(0) | DSI_TALLY_LRX(0) | DSI_TALLY_HTX(0);
+	tegra_dsi_writel(dsi, value, DSI_TO_TALLY);
+
+	if (dsi->slave)
+		tegra_dsi_set_timeout(dsi->slave, bclk, vrefresh);
+}
+
 static int tegra_output_dsi_setup_clock(struct tegra_output *output,
 					struct clk *clk, unsigned long pclk,
 					unsigned int *divp)
 {
 	struct tegra_dc *dc = to_tegra_dc(output->encoder.crtc);
 	struct drm_display_mode *mode = &dc->base.mode;
-	unsigned int timeout, mul, div, vrefresh;
 	struct tegra_dsi *dsi = to_dsi(output);
-	unsigned long bclk, plld, value;
+	unsigned int mul, div, vrefresh, lanes;
+	unsigned long bclk, plld;
 	int err;
 
+	lanes = tegra_dsi_get_lanes(dsi);
+
 	err = tegra_dsi_get_muldiv(dsi->format, &mul, &div);
 	if (err < 0)
 		return err;
 
-	DRM_DEBUG_KMS("mul: %u, div: %u, lanes: %u\n", mul, div, dsi->lanes);
+	DRM_DEBUG_KMS("mul: %u, div: %u, lanes: %u\n", mul, div, lanes);
 	vrefresh = drm_mode_vrefresh(mode);
 	DRM_DEBUG_KMS("vrefresh: %u\n", vrefresh);
 
 	/* compute byte clock */
-	bclk = (pclk * mul) / (div * dsi->lanes);
+	bclk = (pclk * mul) / (div * lanes);
 
 	/*
 	 * Compute bit clock and round up to the next MHz.
 	 */
-	plld = DIV_ROUND_UP(bclk * 8, 1000000) * 1000000;
+	plld = DIV_ROUND_UP(bclk * 8, USEC_PER_SEC) * USEC_PER_SEC;
 
 	/*
 	 * We divide the frequency by two here, but we make up for that by
@@ -640,25 +851,17 @@ static int tegra_output_dsi_setup_clock(struct tegra_output *output,
 	 * not working properly otherwise. Perhaps the PLLs cannot generate
 	 * frequencies sufficiently high.
 	 */
-	*divp = ((8 * mul) / (div * dsi->lanes)) - 2;
+	*divp = ((8 * mul) / (div * lanes)) - 2;
 
 	/*
 	 * XXX: Move the below somewhere else so that we don't need to have
 	 * access to the vrefresh in this function?
 	 */
+	tegra_dsi_set_timeout(dsi, bclk, vrefresh);
 
-	/* one frame high-speed transmission timeout */
-	timeout = (bclk / vrefresh) / 512;
-	value = DSI_TIMEOUT_LRX(0x2000) | DSI_TIMEOUT_HTX(timeout);
-	tegra_dsi_writel(dsi, value, DSI_TIMEOUT_0);
-
-	/* 2 ms peripheral timeout for panel */
-	timeout = 2 * bclk / 512 * 1000;
-	value = DSI_TIMEOUT_PR(timeout) | DSI_TIMEOUT_TA(0x2000);
-	tegra_dsi_writel(dsi, value, DSI_TIMEOUT_1);
-
-	value = DSI_TALLY_TA(0) | DSI_TALLY_LRX(0) | DSI_TALLY_HTX(0);
-	tegra_dsi_writel(dsi, value, DSI_TO_TALLY);
+	err = tegra_dsi_set_phy_timing(dsi);
+	if (err < 0)
+		return err;
 
 	return 0;
 }
@@ -695,7 +898,7 @@ static int tegra_dsi_pad_enable(struct tegra_dsi *dsi)
 
 static int tegra_dsi_pad_calibrate(struct tegra_dsi *dsi)
 {
-	unsigned long value;
+	u32 value;
 
 	tegra_dsi_writel(dsi, 0, DSI_PAD_CONTROL_0);
 	tegra_dsi_writel(dsi, 0, DSI_PAD_CONTROL_1);
@@ -720,14 +923,17 @@ static int tegra_dsi_init(struct host1x_client *client)
 	struct tegra_dsi *dsi = host1x_client_to_dsi(client);
 	int err;
 
-	dsi->output.type = TEGRA_OUTPUT_DSI;
-	dsi->output.dev = client->dev;
-	dsi->output.ops = &dsi_ops;
-
-	err = tegra_output_init(drm, &dsi->output);
-	if (err < 0) {
-		dev_err(client->dev, "output setup failed: %d\n", err);
-		return err;
+	/* Gangsters must not register their own outputs. */
+	if (!dsi->master) {
+		dsi->output.type = TEGRA_OUTPUT_DSI;
+		dsi->output.dev = client->dev;
+		dsi->output.ops = &dsi_ops;
+
+		err = tegra_output_init(drm, &dsi->output);
+		if (err < 0) {
+			dev_err(client->dev, "output setup failed: %d\n", err);
+			return err;
+		}
 	}
 
 	if (IS_ENABLED(CONFIG_DEBUG_FS)) {
@@ -736,12 +942,6 @@ static int tegra_dsi_init(struct host1x_client *client)
 			dev_err(dsi->dev, "debugfs setup failed: %d\n", err);
 	}
 
-	err = tegra_dsi_pad_calibrate(dsi);
-	if (err < 0) {
-		dev_err(dsi->dev, "MIPI calibration failed: %d\n", err);
-		return err;
-	}
-
 	return 0;
 }
 
@@ -756,16 +956,20 @@ static int tegra_dsi_exit(struct host1x_client *client)
 			dev_err(dsi->dev, "debugfs cleanup failed: %d\n", err);
 	}
 
-	err = tegra_output_disable(&dsi->output);
-	if (err < 0) {
-		dev_err(client->dev, "output failed to disable: %d\n", err);
-		return err;
-	}
-
-	err = tegra_output_exit(&dsi->output);
-	if (err < 0) {
-		dev_err(client->dev, "output cleanup failed: %d\n", err);
-		return err;
+	if (!dsi->master) {
+		err = tegra_output_disable(&dsi->output);
+		if (err < 0) {
+			dev_err(client->dev, "output failed to disable: %d\n",
+				err);
+			return err;
+		}
+
+		err = tegra_output_exit(&dsi->output);
+		if (err < 0) {
+			dev_err(client->dev, "output cleanup failed: %d\n",
+				err);
+			return err;
+		}
 	}
 
 	return 0;
@@ -792,20 +996,324 @@ static int tegra_dsi_setup_clocks(struct tegra_dsi *dsi)
 	return 0;
 }
 
+static const char * const error_report[16] = {
+	"SoT Error",
+	"SoT Sync Error",
+	"EoT Sync Error",
+	"Escape Mode Entry Command Error",
+	"Low-Power Transmit Sync Error",
+	"Peripheral Timeout Error",
+	"False Control Error",
+	"Contention Detected",
+	"ECC Error, single-bit",
+	"ECC Error, multi-bit",
+	"Checksum Error",
+	"DSI Data Type Not Recognized",
+	"DSI VC ID Invalid",
+	"Invalid Transmission Length",
+	"Reserved",
+	"DSI Protocol Violation",
+};
+
+static ssize_t tegra_dsi_read_response(struct tegra_dsi *dsi,
+				       const struct mipi_dsi_msg *msg,
+				       size_t count)
+{
+	u8 *rx = msg->rx_buf;
+	unsigned int i, j, k;
+	size_t size = 0;
+	u16 errors;
+	u32 value;
+
+	/* read and parse packet header */
+	value = tegra_dsi_readl(dsi, DSI_RD_DATA);
+
+	switch (value & 0x3f) {
+	case MIPI_DSI_RX_ACKNOWLEDGE_AND_ERROR_REPORT:
+		errors = (value >> 8) & 0xffff;
+		dev_dbg(dsi->dev, "Acknowledge and error report: %04x\n",
+			errors);
+		for (i = 0; i < ARRAY_SIZE(error_report); i++)
+			if (errors & BIT(i))
+				dev_dbg(dsi->dev, "  %2u: %s\n", i,
+					error_report[i]);
+		break;
+
+	case MIPI_DSI_RX_DCS_SHORT_READ_RESPONSE_1BYTE:
+		rx[0] = (value >> 8) & 0xff;
+		size = 1;
+		break;
+
+	case MIPI_DSI_RX_DCS_SHORT_READ_RESPONSE_2BYTE:
+		rx[0] = (value >>  8) & 0xff;
+		rx[1] = (value >> 16) & 0xff;
+		size = 2;
+		break;
+
+	case MIPI_DSI_RX_DCS_LONG_READ_RESPONSE:
+		size = ((value >> 8) & 0xff00) | ((value >> 8) & 0xff);
+		break;
+
+	case MIPI_DSI_RX_GENERIC_LONG_READ_RESPONSE:
+		size = ((value >> 8) & 0xff00) | ((value >> 8) & 0xff);
+		break;
+
+	default:
+		dev_err(dsi->dev, "unhandled response type: %02x\n",
+			value & 0x3f);
+		return -EPROTO;
+	}
+
+	size = min(size, msg->rx_len);
+
+	if (msg->rx_buf && size > 0) {
+		for (i = 0, j = 0; i < count - 1; i++, j += 4) {
+			u8 *rx = msg->rx_buf + j;
+
+			value = tegra_dsi_readl(dsi, DSI_RD_DATA);
+
+			for (k = 0; k < 4 && (j + k) < msg->rx_len; k++)
+				rx[j + k] = (value >> (k << 3)) & 0xff;
+		}
+	}
+
+	return size;
+}
+
+static int tegra_dsi_transmit(struct tegra_dsi *dsi, unsigned long timeout)
+{
+	tegra_dsi_writel(dsi, DSI_TRIGGER_HOST, DSI_TRIGGER);
+
+	timeout = jiffies + msecs_to_jiffies(timeout);
+
+	while (time_before(jiffies, timeout)) {
+		u32 value = tegra_dsi_readl(dsi, DSI_TRIGGER);
+		if ((value & DSI_TRIGGER_HOST) == 0)
+			return 0;
+
+		usleep_range(1000, 2000);
+	}
+
+	DRM_DEBUG_KMS("timeout waiting for transmission to complete\n");
+	return -ETIMEDOUT;
+}
+
+static int tegra_dsi_wait_for_response(struct tegra_dsi *dsi,
+				       unsigned long timeout)
+{
+	timeout = jiffies + msecs_to_jiffies(250);
+
+	while (time_before(jiffies, timeout)) {
+		u32 value = tegra_dsi_readl(dsi, DSI_STATUS);
+		u8 count = value & 0x1f;
+
+		if (count > 0)
+			return count;
+
+		usleep_range(1000, 2000);
+	}
+
+	DRM_DEBUG_KMS("peripheral returned no data\n");
+	return -ETIMEDOUT;
+}
+
+static void tegra_dsi_writesl(struct tegra_dsi *dsi, unsigned long offset,
+			      const void *buffer, size_t size)
+{
+	const u8 *buf = buffer;
+	size_t i, j;
+	u32 value;
+
+	for (j = 0; j < size; j += 4) {
+		value = 0;
+
+		for (i = 0; i < 4 && j + i < size; i++)
+			value |= buf[j + i] << (i << 3);
+
+		tegra_dsi_writel(dsi, value, DSI_WR_DATA);
+	}
+}
+
+static ssize_t tegra_dsi_host_transfer(struct mipi_dsi_host *host,
+				       const struct mipi_dsi_msg *msg)
+{
+	struct tegra_dsi *dsi = host_to_tegra(host);
+	struct mipi_dsi_packet packet;
+	const u8 *header;
+	size_t count;
+	ssize_t err;
+	u32 value;
+
+	err = mipi_dsi_create_packet(&packet, msg);
+	if (err < 0)
+		return err;
+
+	header = packet.header;
+
+	/* maximum FIFO depth is 1920 words */
+	if (packet.size > dsi->video_fifo_depth * 4)
+		return -ENOSPC;
+
+	/* reset underflow/overflow flags */
+	value = tegra_dsi_readl(dsi, DSI_STATUS);
+	if (value & (DSI_STATUS_UNDERFLOW | DSI_STATUS_OVERFLOW)) {
+		value = DSI_HOST_CONTROL_FIFO_RESET;
+		tegra_dsi_writel(dsi, value, DSI_HOST_CONTROL);
+		usleep_range(10, 20);
+	}
+
+	value = tegra_dsi_readl(dsi, DSI_POWER_CONTROL);
+	value |= DSI_POWER_CONTROL_ENABLE;
+	tegra_dsi_writel(dsi, value, DSI_POWER_CONTROL);
+
+	usleep_range(5000, 10000);
+
+	value = DSI_HOST_CONTROL_CRC_RESET | DSI_HOST_CONTROL_TX_TRIG_HOST |
+		DSI_HOST_CONTROL_CS | DSI_HOST_CONTROL_ECC;
+
+	if ((msg->flags & MIPI_DSI_MSG_USE_LPM) == 0)
+		value |= DSI_HOST_CONTROL_HS;
+
+	/*
+	 * The host FIFO has a maximum of 64 words, so larger transmissions
+	 * need to use the video FIFO.
+	 */
+	if (packet.size > dsi->host_fifo_depth * 4)
+		value |= DSI_HOST_CONTROL_FIFO_SEL;
+
+	tegra_dsi_writel(dsi, value, DSI_HOST_CONTROL);
+
+	/*
+	 * For reads and messages with explicitly requested ACK, generate a
+	 * BTA sequence after the transmission of the packet.
+	 */
+	if ((msg->flags & MIPI_DSI_MSG_REQ_ACK) ||
+	    (msg->rx_buf && msg->rx_len > 0)) {
+		value = tegra_dsi_readl(dsi, DSI_HOST_CONTROL);
+		value |= DSI_HOST_CONTROL_PKT_BTA;
+		tegra_dsi_writel(dsi, value, DSI_HOST_CONTROL);
+	}
+
+	value = DSI_CONTROL_LANES(0) | DSI_CONTROL_HOST_ENABLE;
+	tegra_dsi_writel(dsi, value, DSI_CONTROL);
+
+	/* write packet header, ECC is generated by hardware */
+	value = header[2] << 16 | header[1] << 8 | header[0];
+	tegra_dsi_writel(dsi, value, DSI_WR_DATA);
+
+	/* write payload (if any) */
+	if (packet.payload_length > 0)
+		tegra_dsi_writesl(dsi, DSI_WR_DATA, packet.payload,
+				  packet.payload_length);
+
+	err = tegra_dsi_transmit(dsi, 250);
+	if (err < 0)
+		return err;
+
+	if ((msg->flags & MIPI_DSI_MSG_REQ_ACK) ||
+	    (msg->rx_buf && msg->rx_len > 0)) {
+		err = tegra_dsi_wait_for_response(dsi, 250);
+		if (err < 0)
+			return err;
+
+		count = err;
+
+		value = tegra_dsi_readl(dsi, DSI_RD_DATA);
+		switch (value) {
+		case 0x84:
+			/*
+			dev_dbg(dsi->dev, "ACK\n");
+			*/
+			break;
+
+		case 0x87:
+			/*
+			dev_dbg(dsi->dev, "ESCAPE\n");
+			*/
+			break;
+
+		default:
+			dev_err(dsi->dev, "unknown status: %08x\n", value);
+			break;
+		}
+
+		if (count > 1) {
+			err = tegra_dsi_read_response(dsi, msg, count);
+			if (err < 0)
+				dev_err(dsi->dev,
+					"failed to parse response: %zd\n",
+					err);
+			else {
+				/*
+				 * For read commands, return the number of
+				 * bytes returned by the peripheral.
+				 */
+				count = err;
+			}
+		}
+	} else {
+		/*
+		 * For write commands, we have transmitted the 4-byte header
+		 * plus the variable-length payload.
+		 */
+		count = 4 + packet.payload_length;
+	}
+
+	return count;
+}
+
+static int tegra_dsi_ganged_setup(struct tegra_dsi *dsi)
+{
+	struct clk *parent;
+	int err;
+
+	/* make sure both DSI controllers share the same PLL */
+	parent = clk_get_parent(dsi->slave->clk);
+	if (!parent)
+		return -EINVAL;
+
+	err = clk_set_parent(parent, dsi->clk_parent);
+	if (err < 0)
+		return err;
+
+	return 0;
+}
+
 static int tegra_dsi_host_attach(struct mipi_dsi_host *host,
 				 struct mipi_dsi_device *device)
 {
 	struct tegra_dsi *dsi = host_to_tegra(host);
-	struct tegra_output *output = &dsi->output;
 
 	dsi->flags = device->mode_flags;
 	dsi->format = device->format;
 	dsi->lanes = device->lanes;
 
-	output->panel = of_drm_find_panel(device->dev.of_node);
-	if (output->panel) {
-		if (output->connector.dev)
+	if (dsi->slave) {
+		int err;
+
+		dev_dbg(dsi->dev, "attaching dual-channel device %s\n",
+			dev_name(&device->dev));
+
+		err = tegra_dsi_ganged_setup(dsi);
+		if (err < 0) {
+			dev_err(dsi->dev, "failed to set up ganged mode: %d\n",
+				err);
+			return err;
+		}
+	}
+
+	/*
+	 * Slaves don't have a panel associated with them, so they provide
+	 * merely the second channel.
+	 */
+	if (!dsi->master) {
+		struct tegra_output *output = &dsi->output;
+
+		output->panel = of_drm_find_panel(device->dev.of_node);
+		if (output->panel && output->connector.dev) {
+			drm_panel_attach(output->panel, &output->connector);
 			drm_helper_hpd_irq_event(output->connector.dev);
+		}
 	}
 
 	return 0;
@@ -818,10 +1326,10 @@ static int tegra_dsi_host_detach(struct mipi_dsi_host *host,
 	struct tegra_output *output = &dsi->output;
 
 	if (output->panel && &device->dev == output->panel->dev) {
+		output->panel = NULL;
+
 		if (output->connector.dev)
 			drm_helper_hpd_irq_event(output->connector.dev);
-
-		output->panel = NULL;
 	}
 
 	return 0;
@@ -830,8 +1338,29 @@ static int tegra_dsi_host_detach(struct mipi_dsi_host *host,
 static const struct mipi_dsi_host_ops tegra_dsi_host_ops = {
 	.attach = tegra_dsi_host_attach,
 	.detach = tegra_dsi_host_detach,
+	.transfer = tegra_dsi_host_transfer,
 };
 
+static int tegra_dsi_ganged_probe(struct tegra_dsi *dsi)
+{
+	struct device_node *np;
+
+	np = of_parse_phandle(dsi->dev->of_node, "nvidia,ganged-mode", 0);
+	if (np) {
+		struct platform_device *gangster = of_find_device_by_node(np);
+
+		dsi->slave = platform_get_drvdata(gangster);
+		of_node_put(np);
+
+		if (!dsi->slave)
+			return -EPROBE_DEFER;
+
+		dsi->slave->master = dsi;
+	}
+
+	return 0;
+}
+
 static int tegra_dsi_probe(struct platform_device *pdev)
 {
 	struct tegra_dsi *dsi;
@@ -843,11 +1372,19 @@ static int tegra_dsi_probe(struct platform_device *pdev)
 		return -ENOMEM;
 
 	dsi->output.dev = dsi->dev = &pdev->dev;
+	dsi->video_fifo_depth = 1920;
+	dsi->host_fifo_depth = 64;
+
+	err = tegra_dsi_ganged_probe(dsi);
+	if (err < 0)
+		return err;
 
 	err = tegra_output_probe(&dsi->output);
 	if (err < 0)
 		return err;
 
+	dsi->output.connector.polled = DRM_CONNECTOR_POLL_HPD;
+
 	/*
 	 * Assume these values by default. When a DSI peripheral driver
 	 * attaches to the DSI host, the parameters will be taken from
@@ -861,68 +1398,83 @@ static int tegra_dsi_probe(struct platform_device *pdev)
 	if (IS_ERR(dsi->rst))
 		return PTR_ERR(dsi->rst);
 
+	err = reset_control_deassert(dsi->rst);
+	if (err < 0) {
+		dev_err(&pdev->dev, "failed to bring DSI out of reset: %d\n",
+			err);
+		return err;
+	}
+
 	dsi->clk = devm_clk_get(&pdev->dev, NULL);
 	if (IS_ERR(dsi->clk)) {
 		dev_err(&pdev->dev, "cannot get DSI clock\n");
-		return PTR_ERR(dsi->clk);
+		err = PTR_ERR(dsi->clk);
+		goto reset;
 	}
 
 	err = clk_prepare_enable(dsi->clk);
 	if (err < 0) {
 		dev_err(&pdev->dev, "cannot enable DSI clock\n");
-		return err;
+		goto reset;
 	}
 
 	dsi->clk_lp = devm_clk_get(&pdev->dev, "lp");
 	if (IS_ERR(dsi->clk_lp)) {
 		dev_err(&pdev->dev, "cannot get low-power clock\n");
-		return PTR_ERR(dsi->clk_lp);
+		err = PTR_ERR(dsi->clk_lp);
+		goto disable_clk;
 	}
 
 	err = clk_prepare_enable(dsi->clk_lp);
 	if (err < 0) {
 		dev_err(&pdev->dev, "cannot enable low-power clock\n");
-		return err;
+		goto disable_clk;
 	}
 
 	dsi->clk_parent = devm_clk_get(&pdev->dev, "parent");
 	if (IS_ERR(dsi->clk_parent)) {
 		dev_err(&pdev->dev, "cannot get parent clock\n");
-		return PTR_ERR(dsi->clk_parent);
-	}
-
-	err = clk_prepare_enable(dsi->clk_parent);
-	if (err < 0) {
-		dev_err(&pdev->dev, "cannot enable parent clock\n");
-		return err;
+		err = PTR_ERR(dsi->clk_parent);
+		goto disable_clk_lp;
 	}
 
 	dsi->vdd = devm_regulator_get(&pdev->dev, "avdd-dsi-csi");
 	if (IS_ERR(dsi->vdd)) {
 		dev_err(&pdev->dev, "cannot get VDD supply\n");
-		return PTR_ERR(dsi->vdd);
+		err = PTR_ERR(dsi->vdd);
+		goto disable_clk_lp;
 	}
 
 	err = regulator_enable(dsi->vdd);
 	if (err < 0) {
 		dev_err(&pdev->dev, "cannot enable VDD supply\n");
-		return err;
+		goto disable_clk_lp;
 	}
 
 	err = tegra_dsi_setup_clocks(dsi);
 	if (err < 0) {
 		dev_err(&pdev->dev, "cannot setup clocks\n");
-		return err;
+		goto disable_vdd;
 	}
 
 	regs = platform_get_resource(pdev, IORESOURCE_MEM, 0);
 	dsi->regs = devm_ioremap_resource(&pdev->dev, regs);
-	if (IS_ERR(dsi->regs))
-		return PTR_ERR(dsi->regs);
+	if (IS_ERR(dsi->regs)) {
+		err = PTR_ERR(dsi->regs);
+		goto disable_vdd;
+	}
 
 	dsi->mipi = tegra_mipi_request(&pdev->dev);
-	if (IS_ERR(dsi->mipi))
-		return PTR_ERR(dsi->mipi);
+	if (IS_ERR(dsi->mipi)) {
+		err = PTR_ERR(dsi->mipi);
+		goto disable_vdd;
+	}
+
+	err = tegra_dsi_pad_calibrate(dsi);
+	if (err < 0) {
+		dev_err(dsi->dev, "MIPI calibration failed: %d\n", err);
+		goto mipi_free;
+	}
 
 	dsi->host.ops = &tegra_dsi_host_ops;
 	dsi->host.dev = &pdev->dev;
@@ -930,7 +1482,7 @@ static int tegra_dsi_probe(struct platform_device *pdev)
 	err = mipi_dsi_host_register(&dsi->host);
 	if (err < 0) {
 		dev_err(&pdev->dev, "failed to register DSI host: %d\n", err);
-		return err;
+		goto mipi_free;
 	}
 
 	INIT_LIST_HEAD(&dsi->client.list);
@@ -941,12 +1493,26 @@ static int tegra_dsi_probe(struct platform_device *pdev)
 	if (err < 0) {
 		dev_err(&pdev->dev, "failed to register host1x client: %d\n",
 			err);
-		return err;
+		goto unregister;
 	}
 
 	platform_set_drvdata(pdev, dsi);
 
 	return 0;
+
+unregister:
+	mipi_dsi_host_unregister(&dsi->host);
+mipi_free:
+	tegra_mipi_free(dsi->mipi);
+disable_vdd:
+	regulator_disable(dsi->vdd);
+disable_clk_lp:
+	clk_disable_unprepare(dsi->clk_lp);
+disable_clk:
+	clk_disable_unprepare(dsi->clk);
+reset:
+	reset_control_assert(dsi->rst);
+	return err;
 }
 
 static int tegra_dsi_remove(struct platform_device *pdev)
@@ -965,7 +1531,6 @@ static int tegra_dsi_remove(struct platform_device *pdev)
 	tegra_mipi_free(dsi->mipi);
 
 	regulator_disable(dsi->vdd);
-	clk_disable_unprepare(dsi->clk_parent);
 	clk_disable_unprepare(dsi->clk_lp);
 	clk_disable_unprepare(dsi->clk);
 	reset_control_assert(dsi->rst);
diff --git a/drivers/gpu/drm/tegra/dsi.h b/drivers/gpu/drm/tegra/dsi.h
index 5ce610d08d77..bad1006a5150 100644
--- a/drivers/gpu/drm/tegra/dsi.h
+++ b/drivers/gpu/drm/tegra/dsi.h
@@ -21,9 +21,16 @@
 #define DSI_INT_STATUS			0x0d
 #define DSI_INT_MASK			0x0e
 #define DSI_HOST_CONTROL		0x0f
+#define DSI_HOST_CONTROL_FIFO_RESET	(1 << 21)
+#define DSI_HOST_CONTROL_CRC_RESET	(1 << 20)
+#define DSI_HOST_CONTROL_TX_TRIG_SOL	(0 << 12)
+#define DSI_HOST_CONTROL_TX_TRIG_FIFO	(1 << 12)
+#define DSI_HOST_CONTROL_TX_TRIG_HOST	(2 << 12)
 #define DSI_HOST_CONTROL_RAW		(1 << 6)
 #define DSI_HOST_CONTROL_HS		(1 << 5)
-#define DSI_HOST_CONTROL_BTA		(1 << 2)
+#define DSI_HOST_CONTROL_FIFO_SEL	(1 << 4)
+#define DSI_HOST_CONTROL_IMM_BTA	(1 << 3)
+#define DSI_HOST_CONTROL_PKT_BTA	(1 << 2)
 #define DSI_HOST_CONTROL_CS		(1 << 1)
 #define DSI_HOST_CONTROL_ECC		(1 << 0)
 #define DSI_CONTROL			0x10
@@ -39,9 +46,13 @@
 #define DSI_SOL_DELAY			0x11
 #define DSI_MAX_THRESHOLD		0x12
 #define DSI_TRIGGER			0x13
+#define DSI_TRIGGER_HOST		(1 << 1)
+#define DSI_TRIGGER_VIDEO		(1 << 0)
 #define DSI_TX_CRC			0x14
 #define DSI_STATUS			0x15
 #define DSI_STATUS_IDLE			(1 << 10)
+#define DSI_STATUS_UNDERFLOW		(1 <<  9)
+#define DSI_STATUS_OVERFLOW		(1 <<  8)
 #define DSI_INIT_SEQ_CONTROL		0x1a
 #define DSI_INIT_SEQ_DATA_0		0x1b
 #define DSI_INIT_SEQ_DATA_1		0x1c
@@ -104,6 +115,7 @@
 #define DSI_PAD_CONTROL_3		0x51
 #define DSI_PAD_CONTROL_4		0x52
 #define DSI_GANGED_MODE_CONTROL		0x53
+#define DSI_GANGED_MODE_CONTROL_ENABLE	(1 << 0)
 #define DSI_GANGED_MODE_START		0x54
 #define DSI_GANGED_MODE_SIZE		0x55
 #define DSI_RAW_DATA_BYTE_COUNT		0x56
diff --git a/drivers/gpu/drm/tegra/fb.c b/drivers/gpu/drm/tegra/fb.c
index 3513d12d5aa1..e9c715d89261 100644
--- a/drivers/gpu/drm/tegra/fb.c
+++ b/drivers/gpu/drm/tegra/fb.c
@@ -65,8 +65,12 @@ static void tegra_fb_destroy(struct drm_framebuffer *framebuffer)
 	for (i = 0; i < fb->num_planes; i++) {
 		struct tegra_bo *bo = fb->planes[i];
 
-		if (bo)
+		if (bo) {
+			if (bo->pages && bo->vaddr)
+				vunmap(bo->vaddr);
+
 			drm_gem_object_unreference_unlocked(&bo->gem);
+		}
 	}
 
 	drm_framebuffer_cleanup(framebuffer);
@@ -223,14 +227,16 @@ static int tegra_fbdev_probe(struct drm_fb_helper *helper,
 	info = framebuffer_alloc(0, drm->dev);
 	if (!info) {
 		dev_err(drm->dev, "failed to allocate framebuffer info\n");
-		tegra_bo_free_object(&bo->gem);
+		drm_gem_object_unreference_unlocked(&bo->gem);
 		return -ENOMEM;
 	}
 
 	fbdev->fb = tegra_fb_alloc(drm, &cmd, &bo, 1);
 	if (IS_ERR(fbdev->fb)) {
-		dev_err(drm->dev, "failed to allocate DRM framebuffer\n");
 		err = PTR_ERR(fbdev->fb);
+		dev_err(drm->dev, "failed to allocate DRM framebuffer: %d\n",
+			err);
+		drm_gem_object_unreference_unlocked(&bo->gem);
 		goto release;
 	}
 
@@ -254,6 +260,16 @@ static int tegra_fbdev_probe(struct drm_fb_helper *helper,
 	offset = info->var.xoffset * bytes_per_pixel +
 		 info->var.yoffset * fb->pitches[0];
 
+	if (bo->pages) {
+		bo->vaddr = vmap(bo->pages, bo->num_pages, VM_MAP,
+				 pgprot_writecombine(PAGE_KERNEL));
+		if (!bo->vaddr) {
+			dev_err(drm->dev, "failed to vmap() framebuffer\n");
+			err = -ENOMEM;
+			goto destroy;
+		}
+	}
+
 	drm->mode_config.fb_base = (resource_size_t)bo->paddr;
 	info->screen_base = (void __iomem *)bo->vaddr + offset;
 	info->screen_size = size;
@@ -289,6 +305,11 @@ static struct tegra_fbdev *tegra_fbdev_create(struct drm_device *drm)
 	return fbdev;
 }
 
+static void tegra_fbdev_free(struct tegra_fbdev *fbdev)
+{
+	kfree(fbdev);
+}
+
 static int tegra_fbdev_init(struct tegra_fbdev *fbdev,
 			    unsigned int preferred_bpp,
 			    unsigned int num_crtc,
@@ -299,19 +320,21 @@ static int tegra_fbdev_init(struct tegra_fbdev *fbdev,
 
 	err = drm_fb_helper_init(drm, &fbdev->base, num_crtc, max_connectors);
 	if (err < 0) {
-		dev_err(drm->dev, "failed to initialize DRM FB helper\n");
+		dev_err(drm->dev, "failed to initialize DRM FB helper: %d\n",
+			err);
 		return err;
 	}
 
 	err = drm_fb_helper_single_add_all_connectors(&fbdev->base);
 	if (err < 0) {
-		dev_err(drm->dev, "failed to add connectors\n");
+		dev_err(drm->dev, "failed to add connectors: %d\n", err);
 		goto fini;
 	}
 
 	err = drm_fb_helper_initial_config(&fbdev->base, preferred_bpp);
 	if (err < 0) {
-		dev_err(drm->dev, "failed to set initial configuration\n");
+		dev_err(drm->dev, "failed to set initial configuration: %d\n",
+			err);
 		goto fini;
 	}
 
@@ -322,7 +345,7 @@ fini:
 	return err;
 }
 
-static void tegra_fbdev_free(struct tegra_fbdev *fbdev)
+static void tegra_fbdev_exit(struct tegra_fbdev *fbdev)
 {
 	struct fb_info *info = fbdev->base.fbdev;
 
@@ -341,11 +364,11 @@ static void tegra_fbdev_free(struct tegra_fbdev *fbdev)
 
 	if (fbdev->fb) {
 		drm_framebuffer_unregister_private(&fbdev->fb->base);
-		tegra_fb_destroy(&fbdev->fb->base);
+		drm_framebuffer_remove(&fbdev->fb->base);
 	}
 
 	drm_fb_helper_fini(&fbdev->base);
-	kfree(fbdev);
+	tegra_fbdev_free(fbdev);
 }
 
 void tegra_fbdev_restore_mode(struct tegra_fbdev *fbdev)
@@ -393,6 +416,15 @@ int tegra_drm_fb_prepare(struct drm_device *drm)
 	return 0;
 }
 
+void tegra_drm_fb_free(struct drm_device *drm)
+{
+#ifdef CONFIG_DRM_TEGRA_FBDEV
+	struct tegra_drm *tegra = drm->dev_private;
+
+	tegra_fbdev_free(tegra->fbdev);
+#endif
+}
+
 int tegra_drm_fb_init(struct drm_device *drm)
 {
 #ifdef CONFIG_DRM_TEGRA_FBDEV
@@ -413,6 +445,6 @@ void tegra_drm_fb_exit(struct drm_device *drm)
 #ifdef CONFIG_DRM_TEGRA_FBDEV
 	struct tegra_drm *tegra = drm->dev_private;
 
-	tegra_fbdev_free(tegra->fbdev);
+	tegra_fbdev_exit(tegra->fbdev);
 #endif
 }
diff --git a/drivers/gpu/drm/tegra/gem.c b/drivers/gpu/drm/tegra/gem.c
index ce023fa3e8ae..8777b7f75791 100644
--- a/drivers/gpu/drm/tegra/gem.c
+++ b/drivers/gpu/drm/tegra/gem.c
@@ -14,6 +14,7 @@
  */
 
 #include <linux/dma-buf.h>
+#include <linux/iommu.h>
 #include <drm/tegra_drm.h>
 
 #include "drm.h"
@@ -91,13 +92,90 @@ static const struct host1x_bo_ops tegra_bo_ops = {
 	.kunmap = tegra_bo_kunmap,
 };
 
-static void tegra_bo_destroy(struct drm_device *drm, struct tegra_bo *bo)
+/*
+ * A generic iommu_map_sg() function is being reviewed and will hopefully be
+ * merged soon. At that point this function can be dropped in favour of the
+ * one provided by the IOMMU API.
+ */
+static ssize_t __iommu_map_sg(struct iommu_domain *domain, unsigned long iova,
+			      struct scatterlist *sg, unsigned int nents,
+			      int prot)
 {
-	dma_free_writecombine(drm->dev, bo->gem.size, bo->vaddr, bo->paddr);
+	struct scatterlist *s;
+	size_t offset = 0;
+	unsigned int i;
+	int err;
+
+	for_each_sg(sg, s, nents, i) {
+		phys_addr_t phys = page_to_phys(sg_page(s));
+		size_t length = s->offset + s->length;
+
+		err = iommu_map(domain, iova + offset, phys, length, prot);
+		if (err < 0) {
+			iommu_unmap(domain, iova, offset);
+			return err;
+		}
+
+		offset += length;
+	}
+
+	return offset;
 }
 
-struct tegra_bo *tegra_bo_create(struct drm_device *drm, unsigned int size,
-				 unsigned long flags)
+static int tegra_bo_iommu_map(struct tegra_drm *tegra, struct tegra_bo *bo)
+{
+	int prot = IOMMU_READ | IOMMU_WRITE;
+	ssize_t err;
+
+	if (bo->mm)
+		return -EBUSY;
+
+	bo->mm = kzalloc(sizeof(*bo->mm), GFP_KERNEL);
+	if (!bo->mm)
+		return -ENOMEM;
+
+	err = drm_mm_insert_node_generic(&tegra->mm, bo->mm, bo->gem.size,
+					 PAGE_SIZE, 0, 0, 0);
+	if (err < 0) {
+		dev_err(tegra->drm->dev, "out of I/O virtual memory: %zd\n",
+			err);
+		goto free;
+	}
+
+	bo->paddr = bo->mm->start;
+
+	err = __iommu_map_sg(tegra->domain, bo->paddr, bo->sgt->sgl,
+			     bo->sgt->nents, prot);
+	if (err < 0) {
+		dev_err(tegra->drm->dev, "failed to map buffer: %zd\n", err);
+		goto remove;
+	}
+
+	bo->size = err;
+
+	return 0;
+
+remove:
+	drm_mm_remove_node(bo->mm);
+free:
+	kfree(bo->mm);
+	return err;
+}
+
+static int tegra_bo_iommu_unmap(struct tegra_drm *tegra, struct tegra_bo *bo)
+{
+	if (!bo->mm)
+		return 0;
+
+	iommu_unmap(tegra->domain, bo->paddr, bo->size);
+	drm_mm_remove_node(bo->mm);
+	kfree(bo->mm);
+
+	return 0;
+}
+
+static struct tegra_bo *tegra_bo_alloc_object(struct drm_device *drm,
+					      size_t size)
 {
 	struct tegra_bo *bo;
 	int err;
@@ -109,22 +187,124 @@ struct tegra_bo *tegra_bo_create(struct drm_device *drm, unsigned int size,
 	host1x_bo_init(&bo->base, &tegra_bo_ops);
 	size = round_up(size, PAGE_SIZE);
 
-	bo->vaddr = dma_alloc_writecombine(drm->dev, size, &bo->paddr,
-					   GFP_KERNEL | __GFP_NOWARN);
-	if (!bo->vaddr) {
-		dev_err(drm->dev, "failed to allocate buffer with size %u\n",
-			size);
-		err = -ENOMEM;
-		goto err_dma;
-	}
-
 	err = drm_gem_object_init(drm, &bo->gem, size);
-	if (err)
-		goto err_init;
+	if (err < 0)
+		goto free;
 
 	err = drm_gem_create_mmap_offset(&bo->gem);
-	if (err)
-		goto err_mmap;
+	if (err < 0)
+		goto release;
+
+	return bo;
+
+release:
+	drm_gem_object_release(&bo->gem);
+free:
+	kfree(bo);
+	return ERR_PTR(err);
+}
+
+static void tegra_bo_free(struct drm_device *drm, struct tegra_bo *bo)
+{
+	if (bo->pages) {
+		drm_gem_put_pages(&bo->gem, bo->pages, true, true);
+		sg_free_table(bo->sgt);
+		kfree(bo->sgt);
+	} else if (bo->vaddr) {
+		dma_free_writecombine(drm->dev, bo->gem.size, bo->vaddr,
+				      bo->paddr);
+	}
+}
+
+static int tegra_bo_get_pages(struct drm_device *drm, struct tegra_bo *bo)
+{
+	struct scatterlist *s;
+	struct sg_table *sgt;
+	unsigned int i;
+
+	bo->pages = drm_gem_get_pages(&bo->gem);
+	if (IS_ERR(bo->pages))
+		return PTR_ERR(bo->pages);
+
+	bo->num_pages = bo->gem.size >> PAGE_SHIFT;
+
+	sgt = drm_prime_pages_to_sg(bo->pages, bo->num_pages);
+	if (IS_ERR(sgt))
+		goto put_pages;
+
+	/*
+	 * Fake up the SG table so that dma_map_sg() can be used to flush the
+	 * pages associated with it. Note that this relies on the fact that
+	 * the DMA API doesn't hook into IOMMU on Tegra, therefore mapping is
+	 * only cache maintenance.
+	 *
+	 * TODO: Replace this by drm_clflash_sg() once it can be implemented
+	 * without relying on symbols that are not exported.
+	 */
+	for_each_sg(sgt->sgl, s, sgt->nents, i)
+		sg_dma_address(s) = sg_phys(s);
+
+	if (dma_map_sg(drm->dev, sgt->sgl, sgt->nents, DMA_TO_DEVICE) == 0) {
+		sgt = ERR_PTR(-ENOMEM);
+		goto release_sgt;
+	}
+
+	bo->sgt = sgt;
+
+	return 0;
+
+release_sgt:
+	sg_free_table(sgt);
+	kfree(sgt);
+put_pages:
+	drm_gem_put_pages(&bo->gem, bo->pages, false, false);
+	return PTR_ERR(sgt);
+}
+
+static int tegra_bo_alloc(struct drm_device *drm, struct tegra_bo *bo)
+{
+	struct tegra_drm *tegra = drm->dev_private;
+	int err;
+
+	if (tegra->domain) {
+		err = tegra_bo_get_pages(drm, bo);
+		if (err < 0)
+			return err;
+
+		err = tegra_bo_iommu_map(tegra, bo);
+		if (err < 0) {
+			tegra_bo_free(drm, bo);
+			return err;
+		}
+	} else {
+		size_t size = bo->gem.size;
+
+		bo->vaddr = dma_alloc_writecombine(drm->dev, size, &bo->paddr,
+						   GFP_KERNEL | __GFP_NOWARN);
+		if (!bo->vaddr) {
+			dev_err(drm->dev,
+				"failed to allocate buffer of size %zu\n",
+				size);
+			return -ENOMEM;
+		}
+	}
+
+	return 0;
+}
+
+struct tegra_bo *tegra_bo_create(struct drm_device *drm, size_t size,
+				 unsigned long flags)
+{
+	struct tegra_bo *bo;
+	int err;
+
+	bo = tegra_bo_alloc_object(drm, size);
+	if (IS_ERR(bo))
+		return bo;
+
+	err = tegra_bo_alloc(drm, bo);
+	if (err < 0)
+		goto release;
 
 	if (flags & DRM_TEGRA_GEM_CREATE_TILED)
 		bo->tiling.mode = TEGRA_BO_TILING_MODE_TILED;
@@ -134,69 +314,52 @@ struct tegra_bo *tegra_bo_create(struct drm_device *drm, unsigned int size,
 
 	return bo;
 
-err_mmap:
+release:
 	drm_gem_object_release(&bo->gem);
-err_init:
-	tegra_bo_destroy(drm, bo);
-err_dma:
 	kfree(bo);
-
 	return ERR_PTR(err);
 }
 
 struct tegra_bo *tegra_bo_create_with_handle(struct drm_file *file,
 					     struct drm_device *drm,
-					     unsigned int size,
+					     size_t size,
 					     unsigned long flags,
-					     unsigned int *handle)
+					     u32 *handle)
 {
 	struct tegra_bo *bo;
-	int ret;
+	int err;
 
 	bo = tegra_bo_create(drm, size, flags);
 	if (IS_ERR(bo))
 		return bo;
 
-	ret = drm_gem_handle_create(file, &bo->gem, handle);
-	if (ret)
-		goto err;
+	err = drm_gem_handle_create(file, &bo->gem, handle);
+	if (err) {
+		tegra_bo_free_object(&bo->gem);
+		return ERR_PTR(err);
+	}
 
 	drm_gem_object_unreference_unlocked(&bo->gem);
 
 	return bo;
-
-err:
-	tegra_bo_free_object(&bo->gem);
-	return ERR_PTR(ret);
 }
 
 static struct tegra_bo *tegra_bo_import(struct drm_device *drm,
 					struct dma_buf *buf)
 {
+	struct tegra_drm *tegra = drm->dev_private;
 	struct dma_buf_attachment *attach;
 	struct tegra_bo *bo;
-	ssize_t size;
 	int err;
 
-	bo = kzalloc(sizeof(*bo), GFP_KERNEL);
-	if (!bo)
-		return ERR_PTR(-ENOMEM);
-
-	host1x_bo_init(&bo->base, &tegra_bo_ops);
-	size = round_up(buf->size, PAGE_SIZE);
-
-	err = drm_gem_object_init(drm, &bo->gem, size);
-	if (err < 0)
-		goto free;
-
-	err = drm_gem_create_mmap_offset(&bo->gem);
-	if (err < 0)
-		goto release;
+	bo = tegra_bo_alloc_object(drm, buf->size);
+	if (IS_ERR(bo))
+		return bo;
 
 	attach = dma_buf_attach(buf, drm->dev);
 	if (IS_ERR(attach)) {
 		err = PTR_ERR(attach);
-		goto free_mmap;
+		goto free;
 	}
 
 	get_dma_buf(buf);
@@ -212,12 +375,19 @@ static struct tegra_bo *tegra_bo_import(struct drm_device *drm,
 		goto detach;
 	}
 
-	if (bo->sgt->nents > 1) {
-		err = -EINVAL;
-		goto detach;
+	if (tegra->domain) {
+		err = tegra_bo_iommu_map(tegra, bo);
+		if (err < 0)
+			goto detach;
+	} else {
+		if (bo->sgt->nents > 1) {
+			err = -EINVAL;
+			goto detach;
+		}
+
+		bo->paddr = sg_dma_address(bo->sgt->sgl);
 	}
 
-	bo->paddr = sg_dma_address(bo->sgt->sgl);
 	bo->gem.import_attach = attach;
 
 	return bo;
@@ -228,47 +398,41 @@ detach:
 
 	dma_buf_detach(buf, attach);
 	dma_buf_put(buf);
-free_mmap:
-	drm_gem_free_mmap_offset(&bo->gem);
-release:
-	drm_gem_object_release(&bo->gem);
 free:
+	drm_gem_object_release(&bo->gem);
 	kfree(bo);
-
 	return ERR_PTR(err);
 }
 
 void tegra_bo_free_object(struct drm_gem_object *gem)
 {
+	struct tegra_drm *tegra = gem->dev->dev_private;
 	struct tegra_bo *bo = to_tegra_bo(gem);
 
+	if (tegra->domain)
+		tegra_bo_iommu_unmap(tegra, bo);
+
 	if (gem->import_attach) {
 		dma_buf_unmap_attachment(gem->import_attach, bo->sgt,
 					 DMA_TO_DEVICE);
 		drm_prime_gem_destroy(gem, NULL);
 	} else {
-		tegra_bo_destroy(gem->dev, bo);
+		tegra_bo_free(gem->dev, bo);
 	}
 
-	drm_gem_free_mmap_offset(gem);
 	drm_gem_object_release(gem);
-
 	kfree(bo);
 }
 
 int tegra_bo_dumb_create(struct drm_file *file, struct drm_device *drm,
 			 struct drm_mode_create_dumb *args)
 {
-	int min_pitch = DIV_ROUND_UP(args->width * args->bpp, 8);
+	unsigned int min_pitch = DIV_ROUND_UP(args->width * args->bpp, 8);
 	struct tegra_drm *tegra = drm->dev_private;
 	struct tegra_bo *bo;
 
-	min_pitch = round_up(min_pitch, tegra->pitch_align);
-	if (args->pitch < min_pitch)
-		args->pitch = min_pitch;
-
-	if (args->size < args->pitch * args->height)
-		args->size = args->pitch * args->height;
+	args->pitch = round_up(min_pitch, tegra->pitch_align);
+	args->size = args->pitch * args->height;
 
 	bo = tegra_bo_create_with_handle(file, drm, args->size, 0,
 					 &args->handle);
@@ -279,7 +443,7 @@ int tegra_bo_dumb_create(struct drm_file *file, struct drm_device *drm,
 }
 
 int tegra_bo_dumb_map_offset(struct drm_file *file, struct drm_device *drm,
-			     uint32_t handle, uint64_t *offset)
+			     u32 handle, u64 *offset)
 {
 	struct drm_gem_object *gem;
 	struct tegra_bo *bo;
@@ -304,7 +468,38 @@ int tegra_bo_dumb_map_offset(struct drm_file *file, struct drm_device *drm,
 	return 0;
 }
 
+static int tegra_bo_fault(struct vm_area_struct *vma, struct vm_fault *vmf)
+{
+	struct drm_gem_object *gem = vma->vm_private_data;
+	struct tegra_bo *bo = to_tegra_bo(gem);
+	struct page *page;
+	pgoff_t offset;
+	int err;
+
+	if (!bo->pages)
+		return VM_FAULT_SIGBUS;
+
+	offset = ((unsigned long)vmf->virtual_address - vma->vm_start) >> PAGE_SHIFT;
+	page = bo->pages[offset];
+
+	err = vm_insert_page(vma, (unsigned long)vmf->virtual_address, page);
+	switch (err) {
+	case -EAGAIN:
+	case 0:
+	case -ERESTARTSYS:
+	case -EINTR:
+	case -EBUSY:
+		return VM_FAULT_NOPAGE;
+
+	case -ENOMEM:
+		return VM_FAULT_OOM;
+	}
+
+	return VM_FAULT_SIGBUS;
+}
+
 const struct vm_operations_struct tegra_bo_vm_ops = {
+	.fault = tegra_bo_fault,
 	.open = drm_gem_vm_open,
 	.close = drm_gem_vm_close,
 };
@@ -322,12 +517,30 @@ int tegra_drm_mmap(struct file *file, struct vm_area_struct *vma)
 	gem = vma->vm_private_data;
 	bo = to_tegra_bo(gem);
 
-	ret = remap_pfn_range(vma, vma->vm_start, bo->paddr >> PAGE_SHIFT,
-			      vma->vm_end - vma->vm_start, vma->vm_page_prot);
-	if (ret)
-		drm_gem_vm_close(vma);
+	if (!bo->pages) {
+		unsigned long vm_pgoff = vma->vm_pgoff;
+
+		vma->vm_flags &= ~VM_PFNMAP;
+		vma->vm_pgoff = 0;
+
+		ret = dma_mmap_writecombine(gem->dev->dev, vma, bo->vaddr,
+					    bo->paddr, gem->size);
+		if (ret) {
+			drm_gem_vm_close(vma);
+			return ret;
+		}
+
+		vma->vm_pgoff = vm_pgoff;
+	} else {
+		pgprot_t prot = vm_get_page_prot(vma->vm_flags);
+
+		vma->vm_flags |= VM_MIXEDMAP;
+		vma->vm_flags &= ~VM_PFNMAP;
 
-	return ret;
+		vma->vm_page_prot = pgprot_writecombine(prot);
+	}
+
+	return 0;
 }
 
 static struct sg_table *
@@ -342,21 +555,44 @@ tegra_gem_prime_map_dma_buf(struct dma_buf_attachment *attach,
 	if (!sgt)
 		return NULL;
 
-	if (sg_alloc_table(sgt, 1, GFP_KERNEL)) {
-		kfree(sgt);
-		return NULL;
-	}
+	if (bo->pages) {
+		struct scatterlist *sg;
+		unsigned int i;
 
-	sg_dma_address(sgt->sgl) = bo->paddr;
-	sg_dma_len(sgt->sgl) = gem->size;
+		if (sg_alloc_table(sgt, bo->num_pages, GFP_KERNEL))
+			goto free;
+
+		for_each_sg(sgt->sgl, sg, bo->num_pages, i)
+			sg_set_page(sg, bo->pages[i], PAGE_SIZE, 0);
+
+		if (dma_map_sg(attach->dev, sgt->sgl, sgt->nents, dir) == 0)
+			goto free;
+	} else {
+		if (sg_alloc_table(sgt, 1, GFP_KERNEL))
+			goto free;
+
+		sg_dma_address(sgt->sgl) = bo->paddr;
+		sg_dma_len(sgt->sgl) = gem->size;
+	}
 
 	return sgt;
+
+free:
+	sg_free_table(sgt);
+	kfree(sgt);
+	return NULL;
 }
 
 static void tegra_gem_prime_unmap_dma_buf(struct dma_buf_attachment *attach,
 					  struct sg_table *sgt,
 					  enum dma_data_direction dir)
 {
+	struct drm_gem_object *gem = attach->dmabuf->priv;
+	struct tegra_bo *bo = to_tegra_bo(gem);
+
+	if (bo->pages)
+		dma_unmap_sg(attach->dev, sgt->sgl, sgt->nents, dir);
+
 	sg_free_table(sgt);
 	kfree(sgt);
 }
diff --git a/drivers/gpu/drm/tegra/gem.h b/drivers/gpu/drm/tegra/gem.h
index 43a25c853357..6c5f12ac0087 100644
--- a/drivers/gpu/drm/tegra/gem.h
+++ b/drivers/gpu/drm/tegra/gem.h
@@ -15,6 +15,7 @@
 
 #include <drm/drm.h>
 #include <drm/drmP.h>
+#include <drm/drm_gem.h>
 
 #define TEGRA_BO_BOTTOM_UP (1 << 0)
 
@@ -37,6 +38,12 @@ struct tegra_bo {
 	dma_addr_t paddr;
 	void *vaddr;
 
+	struct drm_mm_node *mm;
+	unsigned long num_pages;
+	struct page **pages;
+	/* size of IOMMU mapping */
+	size_t size;
+
 	struct tegra_bo_tiling tiling;
 };
 
@@ -45,18 +52,18 @@ static inline struct tegra_bo *to_tegra_bo(struct drm_gem_object *gem)
 	return container_of(gem, struct tegra_bo, gem);
 }
 
-struct tegra_bo *tegra_bo_create(struct drm_device *drm, unsigned int size,
+struct tegra_bo *tegra_bo_create(struct drm_device *drm, size_t size,
 				 unsigned long flags);
 struct tegra_bo *tegra_bo_create_with_handle(struct drm_file *file,
 					     struct drm_device *drm,
-					     unsigned int size,
+					     size_t size,
 					     unsigned long flags,
-					     unsigned int *handle);
+					     u32 *handle);
 void tegra_bo_free_object(struct drm_gem_object *gem);
 int tegra_bo_dumb_create(struct drm_file *file, struct drm_device *drm,
 			 struct drm_mode_create_dumb *args);
 int tegra_bo_dumb_map_offset(struct drm_file *file, struct drm_device *drm,
-			     uint32_t handle, uint64_t *offset);
+			     u32 handle, u64 *offset);
 
 int tegra_drm_mmap(struct file *file, struct vm_area_struct *vma);
 
diff --git a/drivers/gpu/drm/tegra/output.c b/drivers/gpu/drm/tegra/output.c
index 0c67d7eebc94..6a5c7b81fbc5 100644
--- a/drivers/gpu/drm/tegra/output.c
+++ b/drivers/gpu/drm/tegra/output.c
@@ -157,22 +157,18 @@ static bool tegra_encoder_mode_fixup(struct drm_encoder *encoder,
 
 static void tegra_encoder_prepare(struct drm_encoder *encoder)
 {
+	tegra_encoder_dpms(encoder, DRM_MODE_DPMS_OFF);
 }
 
 static void tegra_encoder_commit(struct drm_encoder *encoder)
 {
+	tegra_encoder_dpms(encoder, DRM_MODE_DPMS_ON);
 }
 
 static void tegra_encoder_mode_set(struct drm_encoder *encoder,
 				   struct drm_display_mode *mode,
 				   struct drm_display_mode *adjusted)
 {
-	struct tegra_output *output = encoder_to_output(encoder);
-	int err;
-
-	err = tegra_output_enable(output);
-	if (err < 0)
-		dev_err(encoder->dev->dev, "tegra_output_enable(): %d\n", err);
 }
 
 static const struct drm_encoder_helper_funcs encoder_helper_funcs = {
@@ -187,7 +183,8 @@ static irqreturn_t hpd_irq(int irq, void *data)
 {
 	struct tegra_output *output = data;
 
-	drm_helper_hpd_irq_event(output->connector.dev);
+	if (output->connector.dev)
+		drm_helper_hpd_irq_event(output->connector.dev);
 
 	return IRQ_HANDLED;
 }
@@ -259,6 +256,13 @@ int tegra_output_probe(struct tegra_output *output)
 		}
 
 		output->connector.polled = DRM_CONNECTOR_POLL_HPD;
+
+		/*
+		 * Disable the interrupt until the connector has been
+		 * initialized to avoid a race in the hotplug interrupt
+		 * handler.
+		 */
+		disable_irq(output->hpd_irq);
 	}
 
 	return 0;
@@ -324,10 +328,27 @@ int tegra_output_init(struct drm_device *drm, struct tegra_output *output)
 
 	output->encoder.possible_crtcs = 0x3;
 
+	/*
+	 * The connector is now registered and ready to receive hotplug events
+	 * so the hotplug interrupt can be enabled.
+	 */
+	if (gpio_is_valid(output->hpd_gpio))
+		enable_irq(output->hpd_irq);
+
 	return 0;
 }
 
 int tegra_output_exit(struct tegra_output *output)
 {
+	/*
+	 * The connector is going away, so the interrupt must be disabled to
+	 * prevent the hotplug interrupt handler from potentially crashing.
+	 */
+	if (gpio_is_valid(output->hpd_gpio))
+		disable_irq(output->hpd_irq);
+
+	if (output->panel)
+		drm_panel_detach(output->panel);
+
 	return 0;
 }
diff --git a/drivers/gpu/drm/tilcdc/tilcdc_crtc.c b/drivers/gpu/drm/tilcdc/tilcdc_crtc.c
index d642d4a02134..c73588483be0 100644
--- a/drivers/gpu/drm/tilcdc/tilcdc_crtc.c
+++ b/drivers/gpu/drm/tilcdc/tilcdc_crtc.c
@@ -16,6 +16,7 @@
  */
 
 #include "drm_flip_work.h"
+#include <drm/drm_plane_helper.h>
 
 #include "tilcdc_drv.h"
 #include "tilcdc_regs.h"
@@ -664,12 +665,8 @@ struct drm_crtc *tilcdc_crtc_create(struct drm_device *dev)
 	tilcdc_crtc->dpms = DRM_MODE_DPMS_OFF;
 	init_waitqueue_head(&tilcdc_crtc->frame_done_wq);
 
-	ret = drm_flip_work_init(&tilcdc_crtc->unref_work, 16,
+	drm_flip_work_init(&tilcdc_crtc->unref_work,
 			"unref", unref_worker);
-	if (ret) {
-		dev_err(dev->dev, "could not allocate unref FIFO\n");
-		goto fail;
-	}
 
 	ret = drm_crtc_init(dev, crtc, &tilcdc_crtc_funcs);
 	if (ret < 0)
diff --git a/drivers/gpu/drm/tilcdc/tilcdc_drv.c b/drivers/gpu/drm/tilcdc/tilcdc_drv.c
index 6be623b4a86f..095fca91525c 100644
--- a/drivers/gpu/drm/tilcdc/tilcdc_drv.c
+++ b/drivers/gpu/drm/tilcdc/tilcdc_drv.c
@@ -58,8 +58,7 @@ static struct drm_framebuffer *tilcdc_fb_create(struct drm_device *dev,
 static void tilcdc_fb_output_poll_changed(struct drm_device *dev)
 {
 	struct tilcdc_drm_private *priv = dev->dev_private;
-	if (priv->fbdev)
-		drm_fbdev_cma_hotplug_event(priv->fbdev);
+	drm_fbdev_cma_hotplug_event(priv->fbdev);
 }
 
 static const struct drm_mode_config_funcs mode_config_funcs = {
@@ -84,6 +83,7 @@ static int modeset_init(struct drm_device *dev)
 	if ((priv->num_encoders == 0) || (priv->num_connectors == 0)) {
 		/* oh nos! */
 		dev_err(dev->dev, "no encoders/connectors found\n");
+		drm_mode_config_cleanup(dev);
 		return -ENXIO;
 	}
 
@@ -172,33 +172,37 @@ static int tilcdc_load(struct drm_device *dev, unsigned long flags)
 	dev->dev_private = priv;
 
 	priv->wq = alloc_ordered_workqueue("tilcdc", 0);
+	if (!priv->wq) {
+		ret = -ENOMEM;
+		goto fail_free_priv;
+	}
 
 	res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
 	if (!res) {
 		dev_err(dev->dev, "failed to get memory resource\n");
 		ret = -EINVAL;
-		goto fail;
+		goto fail_free_wq;
 	}
 
 	priv->mmio = ioremap_nocache(res->start, resource_size(res));
 	if (!priv->mmio) {
 		dev_err(dev->dev, "failed to ioremap\n");
 		ret = -ENOMEM;
-		goto fail;
+		goto fail_free_wq;
 	}
 
 	priv->clk = clk_get(dev->dev, "fck");
 	if (IS_ERR(priv->clk)) {
 		dev_err(dev->dev, "failed to get functional clock\n");
 		ret = -ENODEV;
-		goto fail;
+		goto fail_iounmap;
 	}
 
 	priv->disp_clk = clk_get(dev->dev, "dpll_disp_ck");
 	if (IS_ERR(priv->clk)) {
 		dev_err(dev->dev, "failed to get display clock\n");
 		ret = -ENODEV;
-		goto fail;
+		goto fail_put_clk;
 	}
 
 #ifdef CONFIG_CPU_FREQ
@@ -208,7 +212,7 @@ static int tilcdc_load(struct drm_device *dev, unsigned long flags)
 			CPUFREQ_TRANSITION_NOTIFIER);
 	if (ret) {
 		dev_err(dev->dev, "failed to register cpufreq notifier\n");
-		goto fail;
+		goto fail_put_disp_clk;
 	}
 #endif
 
@@ -253,13 +257,13 @@ static int tilcdc_load(struct drm_device *dev, unsigned long flags)
 	ret = modeset_init(dev);
 	if (ret < 0) {
 		dev_err(dev->dev, "failed to initialize mode setting\n");
-		goto fail;
+		goto fail_cpufreq_unregister;
 	}
 
 	ret = drm_vblank_init(dev, 1);
 	if (ret < 0) {
 		dev_err(dev->dev, "failed to initialize vblank\n");
-		goto fail;
+		goto fail_mode_config_cleanup;
 	}
 
 	pm_runtime_get_sync(dev->dev);
@@ -267,7 +271,7 @@ static int tilcdc_load(struct drm_device *dev, unsigned long flags)
 	pm_runtime_put_sync(dev->dev);
 	if (ret < 0) {
 		dev_err(dev->dev, "failed to install IRQ handler\n");
-		goto fail;
+		goto fail_vblank_cleanup;
 	}
 
 	platform_set_drvdata(pdev, dev);
@@ -283,13 +287,48 @@ static int tilcdc_load(struct drm_device *dev, unsigned long flags)
 	priv->fbdev = drm_fbdev_cma_init(dev, bpp,
 			dev->mode_config.num_crtc,
 			dev->mode_config.num_connector);
+	if (IS_ERR(priv->fbdev)) {
+		ret = PTR_ERR(priv->fbdev);
+		goto fail_irq_uninstall;
+	}
 
 	drm_kms_helper_poll_init(dev);
 
 	return 0;
 
-fail:
-	tilcdc_unload(dev);
+fail_irq_uninstall:
+	pm_runtime_get_sync(dev->dev);
+	drm_irq_uninstall(dev);
+	pm_runtime_put_sync(dev->dev);
+
+fail_vblank_cleanup:
+	drm_vblank_cleanup(dev);
+
+fail_mode_config_cleanup:
+	drm_mode_config_cleanup(dev);
+
+fail_cpufreq_unregister:
+	pm_runtime_disable(dev->dev);
+#ifdef CONFIG_CPU_FREQ
+	cpufreq_unregister_notifier(&priv->freq_transition,
+			CPUFREQ_TRANSITION_NOTIFIER);
+fail_put_disp_clk:
+	clk_put(priv->disp_clk);
+#endif
+
+fail_put_clk:
+	clk_put(priv->clk);
+
+fail_iounmap:
+	iounmap(priv->mmio);
+
+fail_free_wq:
+	flush_workqueue(priv->wq);
+	destroy_workqueue(priv->wq);
+
+fail_free_priv:
+	dev->dev_private = NULL;
+	kfree(priv);
 	return ret;
 }
 
@@ -502,6 +541,7 @@ static struct drm_driver tilcdc_driver = {
 	.unload             = tilcdc_unload,
 	.preclose           = tilcdc_preclose,
 	.lastclose          = tilcdc_lastclose,
+	.set_busid          = drm_platform_set_busid,
 	.irq_handler        = tilcdc_irq,
 	.irq_preinstall     = tilcdc_irq_preinstall,
 	.irq_postinstall    = tilcdc_irq_postinstall,
@@ -604,7 +644,6 @@ static struct platform_driver tilcdc_platform_driver = {
 	.probe      = tilcdc_pdev_probe,
 	.remove     = tilcdc_pdev_remove,
 	.driver     = {
-		.owner  = THIS_MODULE,
 		.name   = "tilcdc",
 		.pm     = &tilcdc_pm_ops,
 		.of_match_table = tilcdc_of_match,
diff --git a/drivers/gpu/drm/tilcdc/tilcdc_panel.c b/drivers/gpu/drm/tilcdc/tilcdc_panel.c
index 4c7aa1d8134f..7a0315855e90 100644
--- a/drivers/gpu/drm/tilcdc/tilcdc_panel.c
+++ b/drivers/gpu/drm/tilcdc/tilcdc_panel.c
@@ -18,6 +18,7 @@
 #include <linux/pinctrl/pinmux.h>
 #include <linux/pinctrl/consumer.h>
 #include <linux/backlight.h>
+#include <linux/gpio/consumer.h>
 #include <video/display_timing.h>
 #include <video/of_display_timing.h>
 #include <video/videomode.h>
@@ -29,6 +30,7 @@ struct panel_module {
 	struct tilcdc_panel_info *info;
 	struct display_timings *timings;
 	struct backlight_device *backlight;
+	struct gpio_desc *enable_gpio;
 };
 #define to_panel_module(x) container_of(x, struct panel_module, base)
 
@@ -55,13 +57,17 @@ static void panel_encoder_dpms(struct drm_encoder *encoder, int mode)
 {
 	struct panel_encoder *panel_encoder = to_panel_encoder(encoder);
 	struct backlight_device *backlight = panel_encoder->mod->backlight;
+	struct gpio_desc *gpio = panel_encoder->mod->enable_gpio;
 
-	if (!backlight)
-		return;
+	if (backlight) {
+		backlight->props.power = mode == DRM_MODE_DPMS_ON ?
+					 FB_BLANK_UNBLANK : FB_BLANK_POWERDOWN;
+		backlight_update_status(backlight);
+	}
 
-	backlight->props.power = mode == DRM_MODE_DPMS_ON
-				     ? FB_BLANK_UNBLANK : FB_BLANK_POWERDOWN;
-	backlight_update_status(backlight);
+	if (gpio)
+		gpiod_set_value_cansleep(gpio,
+					 mode == DRM_MODE_DPMS_ON ? 1 : 0);
 }
 
 static bool panel_encoder_mode_fixup(struct drm_encoder *encoder,
@@ -311,6 +317,7 @@ static struct tilcdc_panel_info *of_get_panel_info(struct device_node *np)
 	info = kzalloc(sizeof(*info), GFP_KERNEL);
 	if (!info) {
 		pr_err("%s: allocation failed\n", __func__);
+		of_node_put(info_np);
 		return NULL;
 	}
 
@@ -331,22 +338,21 @@ static struct tilcdc_panel_info *of_get_panel_info(struct device_node *np)
 	if (ret) {
 		pr_err("%s: error reading panel-info properties\n", __func__);
 		kfree(info);
+		of_node_put(info_np);
 		return NULL;
 	}
+	of_node_put(info_np);
 
 	return info;
 }
 
-static struct of_device_id panel_of_match[];
-
 static int panel_probe(struct platform_device *pdev)
 {
-	struct device_node *node = pdev->dev.of_node;
+	struct device_node *bl_node, *node = pdev->dev.of_node;
 	struct panel_module *panel_mod;
 	struct tilcdc_module *mod;
 	struct pinctrl *pinctrl;
-	int ret = -EINVAL;
-
+	int ret;
 
 	/* bail out early if no DT data: */
 	if (!node) {
@@ -354,10 +360,40 @@ static int panel_probe(struct platform_device *pdev)
 		return -ENXIO;
 	}
 
-	panel_mod = kzalloc(sizeof(*panel_mod), GFP_KERNEL);
+	panel_mod = devm_kzalloc(&pdev->dev, sizeof(*panel_mod), GFP_KERNEL);
 	if (!panel_mod)
 		return -ENOMEM;
 
+	bl_node = of_parse_phandle(node, "backlight", 0);
+	if (bl_node) {
+		panel_mod->backlight = of_find_backlight_by_node(bl_node);
+		of_node_put(bl_node);
+
+		if (!panel_mod->backlight)
+			return -EPROBE_DEFER;
+
+		dev_info(&pdev->dev, "found backlight\n");
+	}
+
+	panel_mod->enable_gpio = devm_gpiod_get(&pdev->dev, "enable");
+	if (IS_ERR(panel_mod->enable_gpio)) {
+		ret = PTR_ERR(panel_mod->enable_gpio);
+		if (ret != -ENOENT) {
+			dev_err(&pdev->dev, "failed to request enable GPIO\n");
+			goto fail_backlight;
+		}
+
+		/* Optional GPIO is not here, continue silently. */
+		panel_mod->enable_gpio = NULL;
+	} else {
+		ret = gpiod_direction_output(panel_mod->enable_gpio, 0);
+		if (ret < 0) {
+			dev_err(&pdev->dev, "failed to setup GPIO\n");
+			goto fail_backlight;
+		}
+		dev_info(&pdev->dev, "found enable GPIO\n");
+	}
+
 	mod = &panel_mod->base;
 	pdev->dev.platform_data = mod;
 
@@ -370,29 +406,30 @@ static int panel_probe(struct platform_device *pdev)
 	panel_mod->timings = of_get_display_timings(node);
 	if (!panel_mod->timings) {
 		dev_err(&pdev->dev, "could not get panel timings\n");
+		ret = -EINVAL;
 		goto fail_free;
 	}
 
 	panel_mod->info = of_get_panel_info(node);
 	if (!panel_mod->info) {
 		dev_err(&pdev->dev, "could not get panel info\n");
+		ret = -EINVAL;
 		goto fail_timings;
 	}
 
 	mod->preferred_bpp = panel_mod->info->bpp;
 
-	panel_mod->backlight = of_find_backlight_by_node(node);
-	if (panel_mod->backlight)
-		dev_info(&pdev->dev, "found backlight\n");
-
 	return 0;
 
 fail_timings:
 	display_timings_release(panel_mod->timings);
 
 fail_free:
-	kfree(panel_mod);
 	tilcdc_module_cleanup(mod);
+
+fail_backlight:
+	if (panel_mod->backlight)
+		put_device(&panel_mod->backlight->dev);
 	return ret;
 }
 
@@ -400,12 +437,15 @@ static int panel_remove(struct platform_device *pdev)
 {
 	struct tilcdc_module *mod = dev_get_platdata(&pdev->dev);
 	struct panel_module *panel_mod = to_panel_module(mod);
+	struct backlight_device *backlight = panel_mod->backlight;
+
+	if (backlight)
+		put_device(&backlight->dev);
 
 	display_timings_release(panel_mod->timings);
 
 	tilcdc_module_cleanup(mod);
 	kfree(panel_mod->info);
-	kfree(panel_mod);
 
 	return 0;
 }
diff --git a/drivers/gpu/drm/ttm/ttm_bo.c b/drivers/gpu/drm/ttm/ttm_bo.c
index 3da89d5dab60..d395b0bef73b 100644
--- a/drivers/gpu/drm/ttm/ttm_bo.c
+++ b/drivers/gpu/drm/ttm/ttm_bo.c
@@ -40,6 +40,7 @@
 #include <linux/file.h>
 #include <linux/module.h>
 #include <linux/atomic.h>
+#include <linux/reservation.h>
 
 #define TTM_ASSERT_LOCKED(param)
 #define TTM_DEBUG(fmt, arg...)
@@ -53,12 +54,13 @@ static struct attribute ttm_bo_count = {
 	.mode = S_IRUGO
 };
 
-static inline int ttm_mem_type_from_flags(uint32_t flags, uint32_t *mem_type)
+static inline int ttm_mem_type_from_place(const struct ttm_place *place,
+					  uint32_t *mem_type)
 {
 	int i;
 
 	for (i = 0; i <= TTM_PL_PRIV5; i++)
-		if (flags & (1 << i)) {
+		if (place->flags & (1 << i)) {
 			*mem_type = i;
 			return 0;
 		}
@@ -89,12 +91,12 @@ static void ttm_bo_mem_space_debug(struct ttm_buffer_object *bo,
 	       bo, bo->mem.num_pages, bo->mem.size >> 10,
 	       bo->mem.size >> 20);
 	for (i = 0; i < placement->num_placement; i++) {
-		ret = ttm_mem_type_from_flags(placement->placement[i],
+		ret = ttm_mem_type_from_place(&placement->placement[i],
 						&mem_type);
 		if (ret)
 			return;
 		pr_err("  placement[%d]=0x%08X (%d)\n",
-		       i, placement->placement[i], mem_type);
+		       i, placement->placement[i].flags, mem_type);
 		ttm_mem_type_debug(bo->bdev, mem_type);
 	}
 }
@@ -141,7 +143,6 @@ static void ttm_bo_release_list(struct kref *list_kref)
 	BUG_ON(atomic_read(&bo->list_kref.refcount));
 	BUG_ON(atomic_read(&bo->kref.refcount));
 	BUG_ON(atomic_read(&bo->cpu_writers));
-	BUG_ON(bo->sync_obj != NULL);
 	BUG_ON(bo->mem.mm_node != NULL);
 	BUG_ON(!list_empty(&bo->lru));
 	BUG_ON(!list_empty(&bo->ddestroy));
@@ -402,36 +403,48 @@ static void ttm_bo_cleanup_memtype_use(struct ttm_buffer_object *bo)
 	ww_mutex_unlock (&bo->resv->lock);
 }
 
+static void ttm_bo_flush_all_fences(struct ttm_buffer_object *bo)
+{
+	struct reservation_object_list *fobj;
+	struct fence *fence;
+	int i;
+
+	fobj = reservation_object_get_list(bo->resv);
+	fence = reservation_object_get_excl(bo->resv);
+	if (fence && !fence->ops->signaled)
+		fence_enable_sw_signaling(fence);
+
+	for (i = 0; fobj && i < fobj->shared_count; ++i) {
+		fence = rcu_dereference_protected(fobj->shared[i],
+					reservation_object_held(bo->resv));
+
+		if (!fence->ops->signaled)
+			fence_enable_sw_signaling(fence);
+	}
+}
+
 static void ttm_bo_cleanup_refs_or_queue(struct ttm_buffer_object *bo)
 {
 	struct ttm_bo_device *bdev = bo->bdev;
 	struct ttm_bo_global *glob = bo->glob;
-	struct ttm_bo_driver *driver = bdev->driver;
-	void *sync_obj = NULL;
 	int put_count;
 	int ret;
 
 	spin_lock(&glob->lru_lock);
 	ret = __ttm_bo_reserve(bo, false, true, false, NULL);
 
-	spin_lock(&bdev->fence_lock);
-	(void) ttm_bo_wait(bo, false, false, true);
-	if (!ret && !bo->sync_obj) {
-		spin_unlock(&bdev->fence_lock);
-		put_count = ttm_bo_del_from_lru(bo);
-
-		spin_unlock(&glob->lru_lock);
-		ttm_bo_cleanup_memtype_use(bo);
+	if (!ret) {
+		if (!ttm_bo_wait(bo, false, false, true)) {
+			put_count = ttm_bo_del_from_lru(bo);
 
-		ttm_bo_list_ref_sub(bo, put_count, true);
+			spin_unlock(&glob->lru_lock);
+			ttm_bo_cleanup_memtype_use(bo);
 
-		return;
-	}
-	if (bo->sync_obj)
-		sync_obj = driver->sync_obj_ref(bo->sync_obj);
-	spin_unlock(&bdev->fence_lock);
+			ttm_bo_list_ref_sub(bo, put_count, true);
 
-	if (!ret) {
+			return;
+		} else
+			ttm_bo_flush_all_fences(bo);
 
 		/*
 		 * Make NO_EVICT bos immediately available to
@@ -450,10 +463,6 @@ static void ttm_bo_cleanup_refs_or_queue(struct ttm_buffer_object *bo)
 	list_add_tail(&bo->ddestroy, &bdev->ddestroy);
 	spin_unlock(&glob->lru_lock);
 
-	if (sync_obj) {
-		driver->sync_obj_flush(sync_obj);
-		driver->sync_obj_unref(&sync_obj);
-	}
 	schedule_delayed_work(&bdev->wq,
 			      ((HZ / 100) < 1) ? 1 : HZ / 100);
 }
@@ -474,44 +483,26 @@ static int ttm_bo_cleanup_refs_and_unlock(struct ttm_buffer_object *bo,
 					  bool interruptible,
 					  bool no_wait_gpu)
 {
-	struct ttm_bo_device *bdev = bo->bdev;
-	struct ttm_bo_driver *driver = bdev->driver;
 	struct ttm_bo_global *glob = bo->glob;
 	int put_count;
 	int ret;
 
-	spin_lock(&bdev->fence_lock);
 	ret = ttm_bo_wait(bo, false, false, true);
 
 	if (ret && !no_wait_gpu) {
-		void *sync_obj;
-
-		/*
-		 * Take a reference to the fence and unreserve,
-		 * at this point the buffer should be dead, so
-		 * no new sync objects can be attached.
-		 */
-		sync_obj = driver->sync_obj_ref(bo->sync_obj);
-		spin_unlock(&bdev->fence_lock);
-
-		__ttm_bo_unreserve(bo);
+		long lret;
+		ww_mutex_unlock(&bo->resv->lock);
 		spin_unlock(&glob->lru_lock);
 
-		ret = driver->sync_obj_wait(sync_obj, false, interruptible);
-		driver->sync_obj_unref(&sync_obj);
-		if (ret)
-			return ret;
+		lret = reservation_object_wait_timeout_rcu(bo->resv,
+							   true,
+							   interruptible,
+							   30 * HZ);
 
-		/*
-		 * remove sync_obj with ttm_bo_wait, the wait should be
-		 * finished, and no new wait object should have been added.
-		 */
-		spin_lock(&bdev->fence_lock);
-		ret = ttm_bo_wait(bo, false, false, true);
-		WARN_ON(ret);
-		spin_unlock(&bdev->fence_lock);
-		if (ret)
-			return ret;
+		if (lret < 0)
+			return lret;
+		else if (lret == 0)
+			return -EBUSY;
 
 		spin_lock(&glob->lru_lock);
 		ret = __ttm_bo_reserve(bo, false, true, false, NULL);
@@ -528,8 +519,14 @@ static int ttm_bo_cleanup_refs_and_unlock(struct ttm_buffer_object *bo,
 			spin_unlock(&glob->lru_lock);
 			return 0;
 		}
-	} else
-		spin_unlock(&bdev->fence_lock);
+
+		/*
+		 * remove sync_obj with ttm_bo_wait, the wait should be
+		 * finished, and no new wait object should have been added.
+		 */
+		ret = ttm_bo_wait(bo, false, false, true);
+		WARN_ON(ret);
+	}
 
 	if (ret || unlikely(list_empty(&bo->ddestroy))) {
 		__ttm_bo_unreserve(bo);
@@ -667,9 +664,7 @@ static int ttm_bo_evict(struct ttm_buffer_object *bo, bool interruptible,
 	struct ttm_placement placement;
 	int ret = 0;
 
-	spin_lock(&bdev->fence_lock);
 	ret = ttm_bo_wait(bo, false, interruptible, no_wait_gpu);
-	spin_unlock(&bdev->fence_lock);
 
 	if (unlikely(ret != 0)) {
 		if (ret != -ERESTARTSYS) {
@@ -685,8 +680,6 @@ static int ttm_bo_evict(struct ttm_buffer_object *bo, bool interruptible,
 	evict_mem.bus.io_reserved_vm = false;
 	evict_mem.bus.io_reserved_count = 0;
 
-	placement.fpfn = 0;
-	placement.lpfn = 0;
 	placement.num_placement = 0;
 	placement.num_busy_placement = 0;
 	bdev->driver->evict_flags(bo, &placement);
@@ -716,6 +709,7 @@ out:
 
 static int ttm_mem_evict_first(struct ttm_bo_device *bdev,
 				uint32_t mem_type,
+				const struct ttm_place *place,
 				bool interruptible,
 				bool no_wait_gpu)
 {
@@ -727,8 +721,21 @@ static int ttm_mem_evict_first(struct ttm_bo_device *bdev,
 	spin_lock(&glob->lru_lock);
 	list_for_each_entry(bo, &man->lru, lru) {
 		ret = __ttm_bo_reserve(bo, false, true, false, NULL);
-		if (!ret)
+		if (!ret) {
+			if (place && (place->fpfn || place->lpfn)) {
+				/* Don't evict this BO if it's outside of the
+				 * requested placement range
+				 */
+				if (place->fpfn >= (bo->mem.start + bo->mem.size) ||
+				    (place->lpfn && place->lpfn <= bo->mem.start)) {
+					__ttm_bo_unreserve(bo);
+					ret = -EBUSY;
+					continue;
+				}
+			}
+
 			break;
+		}
 	}
 
 	if (ret) {
@@ -774,7 +781,7 @@ EXPORT_SYMBOL(ttm_bo_mem_put);
  */
 static int ttm_bo_mem_force_space(struct ttm_buffer_object *bo,
 					uint32_t mem_type,
-					struct ttm_placement *placement,
+					const struct ttm_place *place,
 					struct ttm_mem_reg *mem,
 					bool interruptible,
 					bool no_wait_gpu)
@@ -784,12 +791,12 @@ static int ttm_bo_mem_force_space(struct ttm_buffer_object *bo,
 	int ret;
 
 	do {
-		ret = (*man->func->get_node)(man, bo, placement, 0, mem);
+		ret = (*man->func->get_node)(man, bo, place, mem);
 		if (unlikely(ret != 0))
 			return ret;
 		if (mem->mm_node)
 			break;
-		ret = ttm_mem_evict_first(bdev, mem_type,
+		ret = ttm_mem_evict_first(bdev, mem_type, place,
 					  interruptible, no_wait_gpu);
 		if (unlikely(ret != 0))
 			return ret;
@@ -827,18 +834,18 @@ static uint32_t ttm_bo_select_caching(struct ttm_mem_type_manager *man,
 
 static bool ttm_bo_mt_compatible(struct ttm_mem_type_manager *man,
 				 uint32_t mem_type,
-				 uint32_t proposed_placement,
+				 const struct ttm_place *place,
 				 uint32_t *masked_placement)
 {
 	uint32_t cur_flags = ttm_bo_type_flags(mem_type);
 
-	if ((cur_flags & proposed_placement & TTM_PL_MASK_MEM) == 0)
+	if ((cur_flags & place->flags & TTM_PL_MASK_MEM) == 0)
 		return false;
 
-	if ((proposed_placement & man->available_caching) == 0)
+	if ((place->flags & man->available_caching) == 0)
 		return false;
 
-	cur_flags |= (proposed_placement & man->available_caching);
+	cur_flags |= (place->flags & man->available_caching);
 
 	*masked_placement = cur_flags;
 	return true;
@@ -869,15 +876,14 @@ int ttm_bo_mem_space(struct ttm_buffer_object *bo,
 
 	mem->mm_node = NULL;
 	for (i = 0; i < placement->num_placement; ++i) {
-		ret = ttm_mem_type_from_flags(placement->placement[i],
-						&mem_type);
+		const struct ttm_place *place = &placement->placement[i];
+
+		ret = ttm_mem_type_from_place(place, &mem_type);
 		if (ret)
 			return ret;
 		man = &bdev->man[mem_type];
 
-		type_ok = ttm_bo_mt_compatible(man,
-						mem_type,
-						placement->placement[i],
+		type_ok = ttm_bo_mt_compatible(man, mem_type, place,
 						&cur_flags);
 
 		if (!type_ok)
@@ -889,7 +895,7 @@ int ttm_bo_mem_space(struct ttm_buffer_object *bo,
 		 * Use the access and other non-mapping-related flag bits from
 		 * the memory placement flags to the current flags
 		 */
-		ttm_flag_masked(&cur_flags, placement->placement[i],
+		ttm_flag_masked(&cur_flags, place->flags,
 				~TTM_PL_MASK_MEMTYPE);
 
 		if (mem_type == TTM_PL_SYSTEM)
@@ -897,8 +903,7 @@ int ttm_bo_mem_space(struct ttm_buffer_object *bo,
 
 		if (man->has_type && man->use_type) {
 			type_found = true;
-			ret = (*man->func->get_node)(man, bo, placement,
-						     cur_flags, mem);
+			ret = (*man->func->get_node)(man, bo, place, mem);
 			if (unlikely(ret))
 				return ret;
 		}
@@ -916,17 +921,15 @@ int ttm_bo_mem_space(struct ttm_buffer_object *bo,
 		return -EINVAL;
 
 	for (i = 0; i < placement->num_busy_placement; ++i) {
-		ret = ttm_mem_type_from_flags(placement->busy_placement[i],
-						&mem_type);
+		const struct ttm_place *place = &placement->busy_placement[i];
+
+		ret = ttm_mem_type_from_place(place, &mem_type);
 		if (ret)
 			return ret;
 		man = &bdev->man[mem_type];
 		if (!man->has_type)
 			continue;
-		if (!ttm_bo_mt_compatible(man,
-						mem_type,
-						placement->busy_placement[i],
-						&cur_flags))
+		if (!ttm_bo_mt_compatible(man, mem_type, place, &cur_flags))
 			continue;
 
 		cur_flags = ttm_bo_select_caching(man, bo->mem.placement,
@@ -935,7 +938,7 @@ int ttm_bo_mem_space(struct ttm_buffer_object *bo,
 		 * Use the access and other non-mapping-related flag bits from
 		 * the memory placement flags to the current flags
 		 */
-		ttm_flag_masked(&cur_flags, placement->busy_placement[i],
+		ttm_flag_masked(&cur_flags, place->flags,
 				~TTM_PL_MASK_MEMTYPE);
 
 		if (mem_type == TTM_PL_SYSTEM) {
@@ -945,7 +948,7 @@ int ttm_bo_mem_space(struct ttm_buffer_object *bo,
 			return 0;
 		}
 
-		ret = ttm_bo_mem_force_space(bo, mem_type, placement, mem,
+		ret = ttm_bo_mem_force_space(bo, mem_type, place, mem,
 						interruptible, no_wait_gpu);
 		if (ret == 0 && mem->mm_node) {
 			mem->placement = cur_flags;
@@ -966,7 +969,6 @@ static int ttm_bo_move_buffer(struct ttm_buffer_object *bo,
 {
 	int ret = 0;
 	struct ttm_mem_reg mem;
-	struct ttm_bo_device *bdev = bo->bdev;
 
 	lockdep_assert_held(&bo->resv->lock.base);
 
@@ -975,9 +977,7 @@ static int ttm_bo_move_buffer(struct ttm_buffer_object *bo,
 	 * Have the driver move function wait for idle when necessary,
 	 * instead of doing it here.
 	 */
-	spin_lock(&bdev->fence_lock);
 	ret = ttm_bo_wait(bo, false, interruptible, no_wait_gpu);
-	spin_unlock(&bdev->fence_lock);
 	if (ret)
 		return ret;
 	mem.num_pages = bo->num_pages;
@@ -1006,20 +1006,27 @@ static bool ttm_bo_mem_compat(struct ttm_placement *placement,
 {
 	int i;
 
-	if (mem->mm_node && placement->lpfn != 0 &&
-	    (mem->start < placement->fpfn ||
-	     mem->start + mem->num_pages > placement->lpfn))
-		return false;
-
 	for (i = 0; i < placement->num_placement; i++) {
-		*new_flags = placement->placement[i];
+		const struct ttm_place *heap = &placement->placement[i];
+		if (mem->mm_node &&
+		    (mem->start < heap->fpfn ||
+		     (heap->lpfn != 0 && (mem->start + mem->num_pages) > heap->lpfn)))
+			continue;
+
+		*new_flags = heap->flags;
 		if ((*new_flags & mem->placement & TTM_PL_MASK_CACHING) &&
 		    (*new_flags & mem->placement & TTM_PL_MASK_MEM))
 			return true;
 	}
 
 	for (i = 0; i < placement->num_busy_placement; i++) {
-		*new_flags = placement->busy_placement[i];
+		const struct ttm_place *heap = &placement->busy_placement[i];
+		if (mem->mm_node &&
+		    (mem->start < heap->fpfn ||
+		     (heap->lpfn != 0 && (mem->start + mem->num_pages) > heap->lpfn)))
+			continue;
+
+		*new_flags = heap->flags;
 		if ((*new_flags & mem->placement & TTM_PL_MASK_CACHING) &&
 		    (*new_flags & mem->placement & TTM_PL_MASK_MEM))
 			return true;
@@ -1037,11 +1044,6 @@ int ttm_bo_validate(struct ttm_buffer_object *bo,
 	uint32_t new_flags;
 
 	lockdep_assert_held(&bo->resv->lock.base);
-	/* Check that range is valid */
-	if (placement->lpfn || placement->fpfn)
-		if (placement->fpfn > placement->lpfn ||
-			(placement->lpfn - placement->fpfn) < bo->num_pages)
-			return -EINVAL;
 	/*
 	 * Check whether we need to move buffer.
 	 */
@@ -1070,15 +1072,6 @@ int ttm_bo_validate(struct ttm_buffer_object *bo,
 }
 EXPORT_SYMBOL(ttm_bo_validate);
 
-int ttm_bo_check_placement(struct ttm_buffer_object *bo,
-				struct ttm_placement *placement)
-{
-	BUG_ON((placement->fpfn || placement->lpfn) &&
-	       (bo->mem.num_pages > (placement->lpfn - placement->fpfn)));
-
-	return 0;
-}
-
 int ttm_bo_init(struct ttm_bo_device *bdev,
 		struct ttm_buffer_object *bo,
 		unsigned long size,
@@ -1089,6 +1082,7 @@ int ttm_bo_init(struct ttm_bo_device *bdev,
 		struct file *persistent_swap_storage,
 		size_t acc_size,
 		struct sg_table *sg,
+		struct reservation_object *resv,
 		void (*destroy) (struct ttm_buffer_object *))
 {
 	int ret = 0;
@@ -1142,30 +1136,38 @@ int ttm_bo_init(struct ttm_bo_device *bdev,
 	bo->persistent_swap_storage = persistent_swap_storage;
 	bo->acc_size = acc_size;
 	bo->sg = sg;
-	bo->resv = &bo->ttm_resv;
-	reservation_object_init(bo->resv);
+	if (resv) {
+		bo->resv = resv;
+		lockdep_assert_held(&bo->resv->lock.base);
+	} else {
+		bo->resv = &bo->ttm_resv;
+		reservation_object_init(&bo->ttm_resv);
+	}
 	atomic_inc(&bo->glob->bo_count);
 	drm_vma_node_reset(&bo->vma_node);
 
-	ret = ttm_bo_check_placement(bo, placement);
-
 	/*
 	 * For ttm_bo_type_device buffers, allocate
 	 * address space from the device.
 	 */
-	if (likely(!ret) &&
-	    (bo->type == ttm_bo_type_device ||
-	     bo->type == ttm_bo_type_sg))
+	if (bo->type == ttm_bo_type_device ||
+	    bo->type == ttm_bo_type_sg)
 		ret = drm_vma_offset_add(&bdev->vma_manager, &bo->vma_node,
 					 bo->mem.num_pages);
 
-	locked = ww_mutex_trylock(&bo->resv->lock);
-	WARN_ON(!locked);
+	/* passed reservation objects should already be locked,
+	 * since otherwise lockdep will be angered in radeon.
+	 */
+	if (!resv) {
+		locked = ww_mutex_trylock(&bo->resv->lock);
+		WARN_ON(!locked);
+	}
 
 	if (likely(!ret))
 		ret = ttm_bo_validate(bo, placement, interruptible, false);
 
-	ttm_bo_unreserve(bo);
+	if (!resv)
+		ttm_bo_unreserve(bo);
 
 	if (unlikely(ret))
 		ttm_bo_unref(&bo);
@@ -1223,7 +1225,7 @@ int ttm_bo_create(struct ttm_bo_device *bdev,
 	acc_size = ttm_bo_acc_size(bdev, size, sizeof(struct ttm_buffer_object));
 	ret = ttm_bo_init(bdev, bo, size, type, placement, page_alignment,
 			  interruptible, persistent_swap_storage, acc_size,
-			  NULL, NULL);
+			  NULL, NULL, NULL);
 	if (likely(ret == 0))
 		*p_bo = bo;
 
@@ -1245,7 +1247,7 @@ static int ttm_bo_force_list_clean(struct ttm_bo_device *bdev,
 	spin_lock(&glob->lru_lock);
 	while (!list_empty(&man->lru)) {
 		spin_unlock(&glob->lru_lock);
-		ret = ttm_mem_evict_first(bdev, mem_type, false, false);
+		ret = ttm_mem_evict_first(bdev, mem_type, NULL, false, false);
 		if (ret) {
 			if (allow_errors) {
 				return ret;
@@ -1477,7 +1479,6 @@ int ttm_bo_device_init(struct ttm_bo_device *bdev,
 	bdev->glob = glob;
 	bdev->need_dma32 = need_dma32;
 	bdev->val_seq = 0;
-	spin_lock_init(&bdev->fence_lock);
 	mutex_lock(&glob->device_list_mutex);
 	list_add_tail(&bdev->device_list, &glob->device_list);
 	mutex_unlock(&glob->device_list_mutex);
@@ -1530,65 +1531,56 @@ void ttm_bo_unmap_virtual(struct ttm_buffer_object *bo)
 
 EXPORT_SYMBOL(ttm_bo_unmap_virtual);
 
-
 int ttm_bo_wait(struct ttm_buffer_object *bo,
 		bool lazy, bool interruptible, bool no_wait)
 {
-	struct ttm_bo_driver *driver = bo->bdev->driver;
-	struct ttm_bo_device *bdev = bo->bdev;
-	void *sync_obj;
-	int ret = 0;
-
-	if (likely(bo->sync_obj == NULL))
-		return 0;
+	struct reservation_object_list *fobj;
+	struct reservation_object *resv;
+	struct fence *excl;
+	long timeout = 15 * HZ;
+	int i;
 
-	while (bo->sync_obj) {
+	resv = bo->resv;
+	fobj = reservation_object_get_list(resv);
+	excl = reservation_object_get_excl(resv);
+	if (excl) {
+		if (!fence_is_signaled(excl)) {
+			if (no_wait)
+				return -EBUSY;
 
-		if (driver->sync_obj_signaled(bo->sync_obj)) {
-			void *tmp_obj = bo->sync_obj;
-			bo->sync_obj = NULL;
-			clear_bit(TTM_BO_PRIV_FLAG_MOVING, &bo->priv_flags);
-			spin_unlock(&bdev->fence_lock);
-			driver->sync_obj_unref(&tmp_obj);
-			spin_lock(&bdev->fence_lock);
-			continue;
+			timeout = fence_wait_timeout(excl,
+						     interruptible, timeout);
 		}
+	}
 
-		if (no_wait)
-			return -EBUSY;
+	for (i = 0; fobj && timeout > 0 && i < fobj->shared_count; ++i) {
+		struct fence *fence;
+		fence = rcu_dereference_protected(fobj->shared[i],
+						reservation_object_held(resv));
 
-		sync_obj = driver->sync_obj_ref(bo->sync_obj);
-		spin_unlock(&bdev->fence_lock);
-		ret = driver->sync_obj_wait(sync_obj,
-					    lazy, interruptible);
-		if (unlikely(ret != 0)) {
-			driver->sync_obj_unref(&sync_obj);
-			spin_lock(&bdev->fence_lock);
-			return ret;
-		}
-		spin_lock(&bdev->fence_lock);
-		if (likely(bo->sync_obj == sync_obj)) {
-			void *tmp_obj = bo->sync_obj;
-			bo->sync_obj = NULL;
-			clear_bit(TTM_BO_PRIV_FLAG_MOVING,
-				  &bo->priv_flags);
-			spin_unlock(&bdev->fence_lock);
-			driver->sync_obj_unref(&sync_obj);
-			driver->sync_obj_unref(&tmp_obj);
-			spin_lock(&bdev->fence_lock);
-		} else {
-			spin_unlock(&bdev->fence_lock);
-			driver->sync_obj_unref(&sync_obj);
-			spin_lock(&bdev->fence_lock);
+		if (!fence_is_signaled(fence)) {
+			if (no_wait)
+				return -EBUSY;
+
+			timeout = fence_wait_timeout(fence,
+						     interruptible, timeout);
 		}
 	}
+
+	if (timeout < 0)
+		return timeout;
+
+	if (timeout == 0)
+		return -EBUSY;
+
+	reservation_object_add_excl_fence(resv, NULL);
+	clear_bit(TTM_BO_PRIV_FLAG_MOVING, &bo->priv_flags);
 	return 0;
 }
 EXPORT_SYMBOL(ttm_bo_wait);
 
 int ttm_bo_synccpu_write_grab(struct ttm_buffer_object *bo, bool no_wait)
 {
-	struct ttm_bo_device *bdev = bo->bdev;
 	int ret = 0;
 
 	/*
@@ -1598,9 +1590,7 @@ int ttm_bo_synccpu_write_grab(struct ttm_buffer_object *bo, bool no_wait)
 	ret = ttm_bo_reserve(bo, true, no_wait, false, NULL);
 	if (unlikely(ret != 0))
 		return ret;
-	spin_lock(&bdev->fence_lock);
 	ret = ttm_bo_wait(bo, false, true, no_wait);
-	spin_unlock(&bdev->fence_lock);
 	if (likely(ret == 0))
 		atomic_inc(&bo->cpu_writers);
 	ttm_bo_unreserve(bo);
@@ -1657,9 +1647,7 @@ static int ttm_bo_swapout(struct ttm_mem_shrink *shrink)
 	 * Wait for GPU, then move to system cached.
 	 */
 
-	spin_lock(&bo->bdev->fence_lock);
 	ret = ttm_bo_wait(bo, false, false, false);
-	spin_unlock(&bo->bdev->fence_lock);
 
 	if (unlikely(ret != 0))
 		goto out;
diff --git a/drivers/gpu/drm/ttm/ttm_bo_manager.c b/drivers/gpu/drm/ttm/ttm_bo_manager.c
index 9e103a4875c8..aa0bd054d3e9 100644
--- a/drivers/gpu/drm/ttm/ttm_bo_manager.c
+++ b/drivers/gpu/drm/ttm/ttm_bo_manager.c
@@ -49,18 +49,18 @@ struct ttm_range_manager {
 
 static int ttm_bo_man_get_node(struct ttm_mem_type_manager *man,
 			       struct ttm_buffer_object *bo,
-			       struct ttm_placement *placement,
-			       uint32_t flags,
+			       const struct ttm_place *place,
 			       struct ttm_mem_reg *mem)
 {
 	struct ttm_range_manager *rman = (struct ttm_range_manager *) man->priv;
 	struct drm_mm *mm = &rman->mm;
 	struct drm_mm_node *node = NULL;
+	enum drm_mm_search_flags sflags = DRM_MM_SEARCH_BEST;
 	enum drm_mm_allocator_flags aflags = DRM_MM_CREATE_DEFAULT;
 	unsigned long lpfn;
 	int ret;
 
-	lpfn = placement->lpfn;
+	lpfn = place->lpfn;
 	if (!lpfn)
 		lpfn = man->size;
 
@@ -68,15 +68,16 @@ static int ttm_bo_man_get_node(struct ttm_mem_type_manager *man,
 	if (!node)
 		return -ENOMEM;
 
-	if (flags & TTM_PL_FLAG_TOPDOWN)
+	if (place->flags & TTM_PL_FLAG_TOPDOWN) {
+		sflags = DRM_MM_SEARCH_BELOW;
 		aflags = DRM_MM_CREATE_TOP;
+	}
 
 	spin_lock(&rman->lock);
 	ret = drm_mm_insert_node_in_range_generic(mm, node, mem->num_pages,
 					  mem->page_alignment, 0,
-					  placement->fpfn, lpfn,
-					  DRM_MM_SEARCH_BEST,
-					  aflags);
+					  place->fpfn, lpfn,
+					  sflags, aflags);
 	spin_unlock(&rman->lock);
 
 	if (unlikely(ret)) {
diff --git a/drivers/gpu/drm/ttm/ttm_bo_util.c b/drivers/gpu/drm/ttm/ttm_bo_util.c
index 30e5d90cb7bc..882cccdad272 100644
--- a/drivers/gpu/drm/ttm/ttm_bo_util.c
+++ b/drivers/gpu/drm/ttm/ttm_bo_util.c
@@ -37,6 +37,7 @@
 #include <linux/slab.h>
 #include <linux/vmalloc.h>
 #include <linux/module.h>
+#include <linux/reservation.h>
 
 void ttm_bo_free_old_node(struct ttm_buffer_object *bo)
 {
@@ -444,8 +445,6 @@ static int ttm_buffer_object_transfer(struct ttm_buffer_object *bo,
 				      struct ttm_buffer_object **new_obj)
 {
 	struct ttm_buffer_object *fbo;
-	struct ttm_bo_device *bdev = bo->bdev;
-	struct ttm_bo_driver *driver = bdev->driver;
 	int ret;
 
 	fbo = kmalloc(sizeof(*fbo), GFP_KERNEL);
@@ -466,12 +465,6 @@ static int ttm_buffer_object_transfer(struct ttm_buffer_object *bo,
 	drm_vma_node_reset(&fbo->vma_node);
 	atomic_set(&fbo->cpu_writers, 0);
 
-	spin_lock(&bdev->fence_lock);
-	if (bo->sync_obj)
-		fbo->sync_obj = driver->sync_obj_ref(bo->sync_obj);
-	else
-		fbo->sync_obj = NULL;
-	spin_unlock(&bdev->fence_lock);
 	kref_init(&fbo->list_kref);
 	kref_init(&fbo->kref);
 	fbo->destroy = &ttm_transfered_destroy;
@@ -487,28 +480,24 @@ static int ttm_buffer_object_transfer(struct ttm_buffer_object *bo,
 
 pgprot_t ttm_io_prot(uint32_t caching_flags, pgprot_t tmp)
 {
+	/* Cached mappings need no adjustment */
+	if (caching_flags & TTM_PL_FLAG_CACHED)
+		return tmp;
+
 #if defined(__i386__) || defined(__x86_64__)
 	if (caching_flags & TTM_PL_FLAG_WC)
 		tmp = pgprot_writecombine(tmp);
 	else if (boot_cpu_data.x86 > 3)
 		tmp = pgprot_noncached(tmp);
-
-#elif defined(__powerpc__)
-	if (!(caching_flags & TTM_PL_FLAG_CACHED)) {
-		pgprot_val(tmp) |= _PAGE_NO_CACHE;
-		if (caching_flags & TTM_PL_FLAG_UNCACHED)
-			pgprot_val(tmp) |= _PAGE_GUARDED;
-	}
 #endif
-#if defined(__ia64__) || defined(__arm__)
+#if defined(__ia64__) || defined(__arm__) || defined(__powerpc__)
 	if (caching_flags & TTM_PL_FLAG_WC)
 		tmp = pgprot_writecombine(tmp);
 	else
 		tmp = pgprot_noncached(tmp);
 #endif
 #if defined(__sparc__) || defined(__mips__)
-	if (!(caching_flags & TTM_PL_FLAG_CACHED))
-		tmp = pgprot_noncached(tmp);
+	tmp = pgprot_noncached(tmp);
 #endif
 	return tmp;
 }
@@ -567,9 +556,7 @@ static int ttm_bo_kmap_ttm(struct ttm_buffer_object *bo,
 		 * We need to use vmap to get the desired page protection
 		 * or to make the buffer object look contiguous.
 		 */
-		prot = (mem->placement & TTM_PL_FLAG_CACHED) ?
-			PAGE_KERNEL :
-			ttm_io_prot(mem->placement, PAGE_KERNEL);
+		prot = ttm_io_prot(mem->placement, PAGE_KERNEL);
 		map->bo_kmap_type = ttm_bo_map_vmap;
 		map->virtual = vmap(ttm->pages + start_page, num_pages,
 				    0, prot);
@@ -644,30 +631,20 @@ void ttm_bo_kunmap(struct ttm_bo_kmap_obj *map)
 EXPORT_SYMBOL(ttm_bo_kunmap);
 
 int ttm_bo_move_accel_cleanup(struct ttm_buffer_object *bo,
-			      void *sync_obj,
+			      struct fence *fence,
 			      bool evict,
 			      bool no_wait_gpu,
 			      struct ttm_mem_reg *new_mem)
 {
 	struct ttm_bo_device *bdev = bo->bdev;
-	struct ttm_bo_driver *driver = bdev->driver;
 	struct ttm_mem_type_manager *man = &bdev->man[new_mem->mem_type];
 	struct ttm_mem_reg *old_mem = &bo->mem;
 	int ret;
 	struct ttm_buffer_object *ghost_obj;
-	void *tmp_obj = NULL;
 
-	spin_lock(&bdev->fence_lock);
-	if (bo->sync_obj) {
-		tmp_obj = bo->sync_obj;
-		bo->sync_obj = NULL;
-	}
-	bo->sync_obj = driver->sync_obj_ref(sync_obj);
+	reservation_object_add_excl_fence(bo->resv, fence);
 	if (evict) {
 		ret = ttm_bo_wait(bo, false, false, false);
-		spin_unlock(&bdev->fence_lock);
-		if (tmp_obj)
-			driver->sync_obj_unref(&tmp_obj);
 		if (ret)
 			return ret;
 
@@ -688,14 +665,13 @@ int ttm_bo_move_accel_cleanup(struct ttm_buffer_object *bo,
 		 */
 
 		set_bit(TTM_BO_PRIV_FLAG_MOVING, &bo->priv_flags);
-		spin_unlock(&bdev->fence_lock);
-		if (tmp_obj)
-			driver->sync_obj_unref(&tmp_obj);
 
 		ret = ttm_buffer_object_transfer(bo, &ghost_obj);
 		if (ret)
 			return ret;
 
+		reservation_object_add_excl_fence(ghost_obj->resv, fence);
+
 		/**
 		 * If we're not moving to fixed memory, the TTM object
 		 * needs to stay alive. Otherwhise hang it on the ghost
diff --git a/drivers/gpu/drm/ttm/ttm_bo_vm.c b/drivers/gpu/drm/ttm/ttm_bo_vm.c
index 0ce48e5a9cb4..8fb7213277cc 100644
--- a/drivers/gpu/drm/ttm/ttm_bo_vm.c
+++ b/drivers/gpu/drm/ttm/ttm_bo_vm.c
@@ -45,10 +45,8 @@ static int ttm_bo_vm_fault_idle(struct ttm_buffer_object *bo,
 				struct vm_area_struct *vma,
 				struct vm_fault *vmf)
 {
-	struct ttm_bo_device *bdev = bo->bdev;
 	int ret = 0;
 
-	spin_lock(&bdev->fence_lock);
 	if (likely(!test_bit(TTM_BO_PRIV_FLAG_MOVING, &bo->priv_flags)))
 		goto out_unlock;
 
@@ -82,7 +80,6 @@ static int ttm_bo_vm_fault_idle(struct ttm_buffer_object *bo,
 			VM_FAULT_NOPAGE;
 
 out_unlock:
-	spin_unlock(&bdev->fence_lock);
 	return ret;
 }
 
@@ -200,9 +197,8 @@ static int ttm_bo_vm_fault(struct vm_area_struct *vma, struct vm_fault *vmf)
 						cvma.vm_page_prot);
 	} else {
 		ttm = bo->ttm;
-		if (!(bo->mem.placement & TTM_PL_FLAG_CACHED))
-			cvma.vm_page_prot = ttm_io_prot(bo->mem.placement,
-							cvma.vm_page_prot);
+		cvma.vm_page_prot = ttm_io_prot(bo->mem.placement,
+						cvma.vm_page_prot);
 
 		/* Allocate all page at once, most common usage */
 		if (ttm->bdev->driver->ttm_tt_populate(ttm)) {
diff --git a/drivers/gpu/drm/ttm/ttm_execbuf_util.c b/drivers/gpu/drm/ttm/ttm_execbuf_util.c
index e8dac8758528..3820ae97a030 100644
--- a/drivers/gpu/drm/ttm/ttm_execbuf_util.c
+++ b/drivers/gpu/drm/ttm/ttm_execbuf_util.c
@@ -32,20 +32,12 @@
 #include <linux/sched.h>
 #include <linux/module.h>
 
-static void ttm_eu_backoff_reservation_locked(struct list_head *list)
+static void ttm_eu_backoff_reservation_reverse(struct list_head *list,
+					      struct ttm_validate_buffer *entry)
 {
-	struct ttm_validate_buffer *entry;
-
-	list_for_each_entry(entry, list, head) {
+	list_for_each_entry_continue_reverse(entry, list, head) {
 		struct ttm_buffer_object *bo = entry->bo;
-		if (!entry->reserved)
-			continue;
 
-		entry->reserved = false;
-		if (entry->removed) {
-			ttm_bo_add_to_lru(bo);
-			entry->removed = false;
-		}
 		__ttm_bo_unreserve(bo);
 	}
 }
@@ -56,27 +48,9 @@ static void ttm_eu_del_from_lru_locked(struct list_head *list)
 
 	list_for_each_entry(entry, list, head) {
 		struct ttm_buffer_object *bo = entry->bo;
-		if (!entry->reserved)
-			continue;
-
-		if (!entry->removed) {
-			entry->put_count = ttm_bo_del_from_lru(bo);
-			entry->removed = true;
-		}
-	}
-}
-
-static void ttm_eu_list_ref_sub(struct list_head *list)
-{
-	struct ttm_validate_buffer *entry;
-
-	list_for_each_entry(entry, list, head) {
-		struct ttm_buffer_object *bo = entry->bo;
+		unsigned put_count = ttm_bo_del_from_lru(bo);
 
-		if (entry->put_count) {
-			ttm_bo_list_ref_sub(bo, entry->put_count, true);
-			entry->put_count = 0;
-		}
+		ttm_bo_list_ref_sub(bo, put_count, true);
 	}
 }
 
@@ -91,11 +65,18 @@ void ttm_eu_backoff_reservation(struct ww_acquire_ctx *ticket,
 
 	entry = list_first_entry(list, struct ttm_validate_buffer, head);
 	glob = entry->bo->glob;
+
 	spin_lock(&glob->lru_lock);
-	ttm_eu_backoff_reservation_locked(list);
+	list_for_each_entry(entry, list, head) {
+		struct ttm_buffer_object *bo = entry->bo;
+
+		ttm_bo_add_to_lru(bo);
+		__ttm_bo_unreserve(bo);
+	}
+	spin_unlock(&glob->lru_lock);
+
 	if (ticket)
 		ww_acquire_fini(ticket);
-	spin_unlock(&glob->lru_lock);
 }
 EXPORT_SYMBOL(ttm_eu_backoff_reservation);
 
@@ -112,7 +93,8 @@ EXPORT_SYMBOL(ttm_eu_backoff_reservation);
  */
 
 int ttm_eu_reserve_buffers(struct ww_acquire_ctx *ticket,
-			   struct list_head *list)
+			   struct list_head *list, bool intr,
+			   struct list_head *dups)
 {
 	struct ttm_bo_global *glob;
 	struct ttm_validate_buffer *entry;
@@ -121,60 +103,71 @@ int ttm_eu_reserve_buffers(struct ww_acquire_ctx *ticket,
 	if (list_empty(list))
 		return 0;
 
-	list_for_each_entry(entry, list, head) {
-		entry->reserved = false;
-		entry->put_count = 0;
-		entry->removed = false;
-	}
-
 	entry = list_first_entry(list, struct ttm_validate_buffer, head);
 	glob = entry->bo->glob;
 
 	if (ticket)
 		ww_acquire_init(ticket, &reservation_ww_class);
-retry:
+
 	list_for_each_entry(entry, list, head) {
 		struct ttm_buffer_object *bo = entry->bo;
 
-		/* already slowpath reserved? */
-		if (entry->reserved)
+		ret = __ttm_bo_reserve(bo, intr, (ticket == NULL), true,
+				       ticket);
+		if (!ret && unlikely(atomic_read(&bo->cpu_writers) > 0)) {
+			__ttm_bo_unreserve(bo);
+
+			ret = -EBUSY;
+
+		} else if (ret == -EALREADY && dups) {
+			struct ttm_validate_buffer *safe = entry;
+			entry = list_prev_entry(entry, head);
+			list_del(&safe->head);
+			list_add(&safe->head, dups);
 			continue;
+		}
 
-		ret = __ttm_bo_reserve(bo, true, (ticket == NULL), true,
-				       ticket);
+		if (!ret) {
+			if (!entry->shared)
+				continue;
+
+			ret = reservation_object_reserve_shared(bo->resv);
+			if (!ret)
+				continue;
+		}
 
-		if (ret == -EDEADLK) {
-			/* uh oh, we lost out, drop every reservation and try
-			 * to only reserve this buffer, then start over if
-			 * this succeeds.
-			 */
-			BUG_ON(ticket == NULL);
-			spin_lock(&glob->lru_lock);
-			ttm_eu_backoff_reservation_locked(list);
-			spin_unlock(&glob->lru_lock);
-			ttm_eu_list_ref_sub(list);
+		/* uh oh, we lost out, drop every reservation and try
+		 * to only reserve this buffer, then start over if
+		 * this succeeds.
+		 */
+		ttm_eu_backoff_reservation_reverse(list, entry);
+
+		if (ret == -EDEADLK && intr) {
 			ret = ww_mutex_lock_slow_interruptible(&bo->resv->lock,
 							       ticket);
-			if (unlikely(ret != 0)) {
-				if (ret == -EINTR)
-					ret = -ERESTARTSYS;
-				goto err_fini;
-			}
+		} else if (ret == -EDEADLK) {
+			ww_mutex_lock_slow(&bo->resv->lock, ticket);
+			ret = 0;
+		}
 
-			entry->reserved = true;
-			if (unlikely(atomic_read(&bo->cpu_writers) > 0)) {
-				ret = -EBUSY;
-				goto err;
-			}
-			goto retry;
-		} else if (ret)
-			goto err;
+		if (!ret && entry->shared)
+			ret = reservation_object_reserve_shared(bo->resv);
 
-		entry->reserved = true;
-		if (unlikely(atomic_read(&bo->cpu_writers) > 0)) {
-			ret = -EBUSY;
-			goto err;
+		if (unlikely(ret != 0)) {
+			if (ret == -EINTR)
+				ret = -ERESTARTSYS;
+			if (ticket) {
+				ww_acquire_done(ticket);
+				ww_acquire_fini(ticket);
+			}
+			return ret;
 		}
+
+		/* move this item to the front of the list,
+		 * forces correct iteration of the loop without keeping track
+		 */
+		list_del(&entry->head);
+		list_add(&entry->head, list);
 	}
 
 	if (ticket)
@@ -182,25 +175,12 @@ retry:
 	spin_lock(&glob->lru_lock);
 	ttm_eu_del_from_lru_locked(list);
 	spin_unlock(&glob->lru_lock);
-	ttm_eu_list_ref_sub(list);
 	return 0;
-
-err:
-	spin_lock(&glob->lru_lock);
-	ttm_eu_backoff_reservation_locked(list);
-	spin_unlock(&glob->lru_lock);
-	ttm_eu_list_ref_sub(list);
-err_fini:
-	if (ticket) {
-		ww_acquire_done(ticket);
-		ww_acquire_fini(ticket);
-	}
-	return ret;
 }
 EXPORT_SYMBOL(ttm_eu_reserve_buffers);
 
 void ttm_eu_fence_buffer_objects(struct ww_acquire_ctx *ticket,
-				 struct list_head *list, void *sync_obj)
+				 struct list_head *list, struct fence *fence)
 {
 	struct ttm_validate_buffer *entry;
 	struct ttm_buffer_object *bo;
@@ -217,24 +197,18 @@ void ttm_eu_fence_buffer_objects(struct ww_acquire_ctx *ticket,
 	glob = bo->glob;
 
 	spin_lock(&glob->lru_lock);
-	spin_lock(&bdev->fence_lock);
 
 	list_for_each_entry(entry, list, head) {
 		bo = entry->bo;
-		entry->old_sync_obj = bo->sync_obj;
-		bo->sync_obj = driver->sync_obj_ref(sync_obj);
+		if (entry->shared)
+			reservation_object_add_shared_fence(bo->resv, fence);
+		else
+			reservation_object_add_excl_fence(bo->resv, fence);
 		ttm_bo_add_to_lru(bo);
 		__ttm_bo_unreserve(bo);
-		entry->reserved = false;
 	}
-	spin_unlock(&bdev->fence_lock);
 	spin_unlock(&glob->lru_lock);
 	if (ticket)
 		ww_acquire_fini(ticket);
-
-	list_for_each_entry(entry, list, head) {
-		if (entry->old_sync_obj)
-			driver->sync_obj_unref(&entry->old_sync_obj);
-	}
 }
 EXPORT_SYMBOL(ttm_eu_fence_buffer_objects);
diff --git a/drivers/gpu/drm/ttm/ttm_memory.c b/drivers/gpu/drm/ttm/ttm_memory.c
index dbc2def887cd..a1803fbcc898 100644
--- a/drivers/gpu/drm/ttm/ttm_memory.c
+++ b/drivers/gpu/drm/ttm/ttm_memory.c
@@ -300,7 +300,8 @@ static int ttm_mem_init_highmem_zone(struct ttm_mem_global *glob,
 	zone->glob = glob;
 	glob->zone_highmem = zone;
 	ret = kobject_init_and_add(
-		&zone->kobj, &ttm_mem_zone_kobj_type, &glob->kobj, zone->name);
+		&zone->kobj, &ttm_mem_zone_kobj_type, &glob->kobj, "%s",
+		zone->name);
 	if (unlikely(ret != 0)) {
 		kobject_put(&zone->kobj);
 		return ret;
diff --git a/drivers/gpu/drm/ttm/ttm_page_alloc.c b/drivers/gpu/drm/ttm/ttm_page_alloc.c
index 09874d695188..025c429050c0 100644
--- a/drivers/gpu/drm/ttm/ttm_page_alloc.c
+++ b/drivers/gpu/drm/ttm/ttm_page_alloc.c
@@ -297,11 +297,12 @@ static void ttm_pool_update_free_locked(struct ttm_page_pool *pool,
  *
  * @pool: to free the pages from
  * @free_all: If set to true will free all pages in pool
- * @gfp: GFP flags.
+ * @use_static: Safe to use static buffer
  **/
 static int ttm_page_pool_free(struct ttm_page_pool *pool, unsigned nr_free,
-			      gfp_t gfp)
+			      bool use_static)
 {
+	static struct page *static_buf[NUM_PAGES_TO_ALLOC];
 	unsigned long irq_flags;
 	struct page *p;
 	struct page **pages_to_free;
@@ -311,7 +312,11 @@ static int ttm_page_pool_free(struct ttm_page_pool *pool, unsigned nr_free,
 	if (NUM_PAGES_TO_ALLOC < nr_free)
 		npages_to_free = NUM_PAGES_TO_ALLOC;
 
-	pages_to_free = kmalloc(npages_to_free * sizeof(struct page *), gfp);
+	if (use_static)
+		pages_to_free = static_buf;
+	else
+		pages_to_free = kmalloc(npages_to_free * sizeof(struct page *),
+					GFP_KERNEL);
 	if (!pages_to_free) {
 		pr_err("Failed to allocate memory for pool free operation\n");
 		return 0;
@@ -374,7 +379,8 @@ restart:
 	if (freed_pages)
 		ttm_pages_put(pages_to_free, freed_pages);
 out:
-	kfree(pages_to_free);
+	if (pages_to_free != static_buf)
+		kfree(pages_to_free);
 	return nr_free;
 }
 
@@ -383,8 +389,6 @@ out:
  *
  * XXX: (dchinner) Deadlock warning!
  *
- * We need to pass sc->gfp_mask to ttm_page_pool_free().
- *
  * This code is crying out for a shrinker per pool....
  */
 static unsigned long
@@ -407,8 +411,8 @@ ttm_pool_shrink_scan(struct shrinker *shrink, struct shrink_control *sc)
 		if (shrink_pages == 0)
 			break;
 		pool = &_manager->pools[(i + pool_offset)%NUM_POOLS];
-		shrink_pages = ttm_page_pool_free(pool, nr_free,
-						  sc->gfp_mask);
+		/* OK to use static buffer since global mutex is held. */
+		shrink_pages = ttm_page_pool_free(pool, nr_free, true);
 		freed += nr_free - shrink_pages;
 	}
 	mutex_unlock(&lock);
@@ -710,7 +714,7 @@ static void ttm_put_pages(struct page **pages, unsigned npages, int flags,
 	}
 	spin_unlock_irqrestore(&pool->lock, irq_flags);
 	if (npages)
-		ttm_page_pool_free(pool, npages, GFP_KERNEL);
+		ttm_page_pool_free(pool, npages, false);
 }
 
 /*
@@ -849,9 +853,9 @@ void ttm_page_alloc_fini(void)
 	pr_info("Finalizing pool allocator\n");
 	ttm_pool_mm_shrink_fini(_manager);
 
+	/* OK to use static buffer since global mutex is no longer used. */
 	for (i = 0; i < NUM_POOLS; ++i)
-		ttm_page_pool_free(&_manager->pools[i], FREE_ALL_PAGES,
-				   GFP_KERNEL);
+		ttm_page_pool_free(&_manager->pools[i], FREE_ALL_PAGES, true);
 
 	kobject_put(&_manager->kobj);
 	_manager = NULL;
diff --git a/drivers/gpu/drm/ttm/ttm_page_alloc_dma.c b/drivers/gpu/drm/ttm/ttm_page_alloc_dma.c
index c96db433f8af..01e1d27eb078 100644
--- a/drivers/gpu/drm/ttm/ttm_page_alloc_dma.c
+++ b/drivers/gpu/drm/ttm/ttm_page_alloc_dma.c
@@ -411,11 +411,12 @@ static void ttm_dma_page_put(struct dma_pool *pool, struct dma_page *d_page)
  *
  * @pool: to free the pages from
  * @nr_free: If set to true will free all pages in pool
- * @gfp: GFP flags.
+ * @use_static: Safe to use static buffer
  **/
 static unsigned ttm_dma_page_pool_free(struct dma_pool *pool, unsigned nr_free,
-				       gfp_t gfp)
+				       bool use_static)
 {
+	static struct page *static_buf[NUM_PAGES_TO_ALLOC];
 	unsigned long irq_flags;
 	struct dma_page *dma_p, *tmp;
 	struct page **pages_to_free;
@@ -432,7 +433,11 @@ static unsigned ttm_dma_page_pool_free(struct dma_pool *pool, unsigned nr_free,
 			 npages_to_free, nr_free);
 	}
 #endif
-	pages_to_free = kmalloc(npages_to_free * sizeof(struct page *), gfp);
+	if (use_static)
+		pages_to_free = static_buf;
+	else
+		pages_to_free = kmalloc(npages_to_free * sizeof(struct page *),
+					GFP_KERNEL);
 
 	if (!pages_to_free) {
 		pr_err("%s: Failed to allocate memory for pool free operation\n",
@@ -502,7 +507,8 @@ restart:
 	if (freed_pages)
 		ttm_dma_pages_put(pool, &d_pages, pages_to_free, freed_pages);
 out:
-	kfree(pages_to_free);
+	if (pages_to_free != static_buf)
+		kfree(pages_to_free);
 	return nr_free;
 }
 
@@ -531,7 +537,8 @@ static void ttm_dma_free_pool(struct device *dev, enum pool_type type)
 		if (pool->type != type)
 			continue;
 		/* Takes a spinlock.. */
-		ttm_dma_page_pool_free(pool, FREE_ALL_PAGES, GFP_KERNEL);
+		/* OK to use static buffer since global mutex is held. */
+		ttm_dma_page_pool_free(pool, FREE_ALL_PAGES, true);
 		WARN_ON(((pool->npages_in_use + pool->npages_free) != 0));
 		/* This code path is called after _all_ references to the
 		 * struct device has been dropped - so nobody should be
@@ -986,7 +993,7 @@ void ttm_dma_unpopulate(struct ttm_dma_tt *ttm_dma, struct device *dev)
 
 	/* shrink pool if necessary (only on !is_cached pools)*/
 	if (npages)
-		ttm_dma_page_pool_free(pool, npages, GFP_KERNEL);
+		ttm_dma_page_pool_free(pool, npages, false);
 	ttm->state = tt_unpopulated;
 }
 EXPORT_SYMBOL_GPL(ttm_dma_unpopulate);
@@ -996,8 +1003,6 @@ EXPORT_SYMBOL_GPL(ttm_dma_unpopulate);
  *
  * XXX: (dchinner) Deadlock warning!
  *
- * We need to pass sc->gfp_mask to ttm_dma_page_pool_free().
- *
  * I'm getting sadder as I hear more pathetical whimpers about needing per-pool
  * shrinkers
  */
@@ -1030,8 +1035,8 @@ ttm_dma_pool_shrink_scan(struct shrinker *shrink, struct shrink_control *sc)
 		if (++idx < pool_offset)
 			continue;
 		nr_free = shrink_pages;
-		shrink_pages = ttm_dma_page_pool_free(p->pool, nr_free,
-						      sc->gfp_mask);
+		/* OK to use static buffer since global mutex is held. */
+		shrink_pages = ttm_dma_page_pool_free(p->pool, nr_free, true);
 		freed += nr_free - shrink_pages;
 
 		pr_debug("%s: (%s:%d) Asked to shrink %d, have %d more to go\n",
diff --git a/drivers/gpu/drm/udl/Kconfig b/drivers/gpu/drm/udl/Kconfig
index f02528686cd5..613ab0622d6e 100644
--- a/drivers/gpu/drm/udl/Kconfig
+++ b/drivers/gpu/drm/udl/Kconfig
@@ -1,8 +1,9 @@
 config DRM_UDL
 	tristate "DisplayLink"
 	depends on DRM
+	depends on USB_SUPPORT
 	depends on USB_ARCH_HAS_HCD
-	select DRM_USB
+	select USB
 	select FB_SYS_FILLRECT
 	select FB_SYS_COPYAREA
 	select FB_SYS_IMAGEBLIT
diff --git a/drivers/gpu/drm/udl/Makefile b/drivers/gpu/drm/udl/Makefile
index 05c7481bfd40..195bcac0b6c8 100644
--- a/drivers/gpu/drm/udl/Makefile
+++ b/drivers/gpu/drm/udl/Makefile
@@ -1,6 +1,6 @@
 
 ccflags-y := -Iinclude/drm
 
-udl-y := udl_drv.o udl_modeset.o udl_connector.o udl_encoder.o udl_main.o udl_fb.o udl_transfer.o udl_gem.o
+udl-y := udl_drv.o udl_modeset.o udl_connector.o udl_encoder.o udl_main.o udl_fb.o udl_transfer.o udl_gem.o udl_dmabuf.o
 
 obj-$(CONFIG_DRM_UDL) := udl.o
diff --git a/drivers/gpu/drm/udl/udl_connector.c b/drivers/gpu/drm/udl/udl_connector.c
index e026a9e2942a..0110d95522f3 100644
--- a/drivers/gpu/drm/udl/udl_connector.c
+++ b/drivers/gpu/drm/udl/udl_connector.c
@@ -34,8 +34,8 @@ static u8 *udl_get_edid(struct udl_device *udl)
 		goto error;
 
 	for (i = 0; i < EDID_LENGTH; i++) {
-		ret = usb_control_msg(udl->ddev->usbdev,
-				      usb_rcvctrlpipe(udl->ddev->usbdev, 0), (0x02),
+		ret = usb_control_msg(udl->udev,
+				      usb_rcvctrlpipe(udl->udev, 0), (0x02),
 				      (0x80 | (0x02 << 5)), i << 8, 0xA1, rbuf, 2,
 				      HZ);
 		if (ret < 1) {
diff --git a/drivers/gpu/drm/udl/udl_dmabuf.c b/drivers/gpu/drm/udl/udl_dmabuf.c
new file mode 100644
index 000000000000..ac8a66b4dfc2
--- /dev/null
+++ b/drivers/gpu/drm/udl/udl_dmabuf.c
@@ -0,0 +1,276 @@
+/*
+ * udl_dmabuf.c
+ *
+ * Copyright (c) 2014 The Chromium OS Authors
+ *
+ * This program is free software; you can redistribute  it and/or modify it
+ * under  the terms of  the GNU General  Public License as published by the
+ * Free Software Foundation;  either version 2 of the  License, or (at your
+ * option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program. If not, see <http://www.gnu.org/licenses/>.
+ */
+
+#include <drm/drmP.h>
+#include "udl_drv.h"
+#include <linux/shmem_fs.h>
+#include <linux/dma-buf.h>
+
+struct udl_drm_dmabuf_attachment {
+	struct sg_table sgt;
+	enum dma_data_direction dir;
+	bool is_mapped;
+};
+
+static int udl_attach_dma_buf(struct dma_buf *dmabuf,
+			      struct device *dev,
+			      struct dma_buf_attachment *attach)
+{
+	struct udl_drm_dmabuf_attachment *udl_attach;
+
+	DRM_DEBUG_PRIME("[DEV:%s] size:%zd\n", dev_name(attach->dev),
+			attach->dmabuf->size);
+
+	udl_attach = kzalloc(sizeof(*udl_attach), GFP_KERNEL);
+	if (!udl_attach)
+		return -ENOMEM;
+
+	udl_attach->dir = DMA_NONE;
+	attach->priv = udl_attach;
+
+	return 0;
+}
+
+static void udl_detach_dma_buf(struct dma_buf *dmabuf,
+			       struct dma_buf_attachment *attach)
+{
+	struct udl_drm_dmabuf_attachment *udl_attach = attach->priv;
+	struct sg_table *sgt;
+
+	if (!udl_attach)
+		return;
+
+	DRM_DEBUG_PRIME("[DEV:%s] size:%zd\n", dev_name(attach->dev),
+			attach->dmabuf->size);
+
+	sgt = &udl_attach->sgt;
+
+	if (udl_attach->dir != DMA_NONE)
+		dma_unmap_sg(attach->dev, sgt->sgl, sgt->nents,
+				udl_attach->dir);
+
+	sg_free_table(sgt);
+	kfree(udl_attach);
+	attach->priv = NULL;
+}
+
+static struct sg_table *udl_map_dma_buf(struct dma_buf_attachment *attach,
+					enum dma_data_direction dir)
+{
+	struct udl_drm_dmabuf_attachment *udl_attach = attach->priv;
+	struct udl_gem_object *obj = to_udl_bo(attach->dmabuf->priv);
+	struct drm_device *dev = obj->base.dev;
+	struct scatterlist *rd, *wr;
+	struct sg_table *sgt = NULL;
+	unsigned int i;
+	int page_count;
+	int nents, ret;
+
+	DRM_DEBUG_PRIME("[DEV:%s] size:%zd dir=%d\n", dev_name(attach->dev),
+			attach->dmabuf->size, dir);
+
+	/* just return current sgt if already requested. */
+	if (udl_attach->dir == dir && udl_attach->is_mapped)
+		return &udl_attach->sgt;
+
+	if (!obj->pages) {
+		ret = udl_gem_get_pages(obj);
+		if (ret) {
+			DRM_ERROR("failed to map pages.\n");
+			return ERR_PTR(ret);
+		}
+	}
+
+	page_count = obj->base.size / PAGE_SIZE;
+	obj->sg = drm_prime_pages_to_sg(obj->pages, page_count);
+	if (IS_ERR(obj->sg)) {
+		DRM_ERROR("failed to allocate sgt.\n");
+		return ERR_CAST(obj->sg);
+	}
+
+	sgt = &udl_attach->sgt;
+
+	ret = sg_alloc_table(sgt, obj->sg->orig_nents, GFP_KERNEL);
+	if (ret) {
+		DRM_ERROR("failed to alloc sgt.\n");
+		return ERR_PTR(-ENOMEM);
+	}
+
+	mutex_lock(&dev->struct_mutex);
+
+	rd = obj->sg->sgl;
+	wr = sgt->sgl;
+	for (i = 0; i < sgt->orig_nents; ++i) {
+		sg_set_page(wr, sg_page(rd), rd->length, rd->offset);
+		rd = sg_next(rd);
+		wr = sg_next(wr);
+	}
+
+	if (dir != DMA_NONE) {
+		nents = dma_map_sg(attach->dev, sgt->sgl, sgt->orig_nents, dir);
+		if (!nents) {
+			DRM_ERROR("failed to map sgl with iommu.\n");
+			sg_free_table(sgt);
+			sgt = ERR_PTR(-EIO);
+			goto err_unlock;
+		}
+	}
+
+	udl_attach->is_mapped = true;
+	udl_attach->dir = dir;
+	attach->priv = udl_attach;
+
+err_unlock:
+	mutex_unlock(&dev->struct_mutex);
+	return sgt;
+}
+
+static void udl_unmap_dma_buf(struct dma_buf_attachment *attach,
+			      struct sg_table *sgt,
+			      enum dma_data_direction dir)
+{
+	/* Nothing to do. */
+	DRM_DEBUG_PRIME("[DEV:%s] size:%zd dir:%d\n", dev_name(attach->dev),
+			attach->dmabuf->size, dir);
+}
+
+static void *udl_dmabuf_kmap(struct dma_buf *dma_buf, unsigned long page_num)
+{
+	/* TODO */
+
+	return NULL;
+}
+
+static void *udl_dmabuf_kmap_atomic(struct dma_buf *dma_buf,
+				    unsigned long page_num)
+{
+	/* TODO */
+
+	return NULL;
+}
+
+static void udl_dmabuf_kunmap(struct dma_buf *dma_buf,
+			      unsigned long page_num, void *addr)
+{
+	/* TODO */
+}
+
+static void udl_dmabuf_kunmap_atomic(struct dma_buf *dma_buf,
+				     unsigned long page_num,
+				     void *addr)
+{
+	/* TODO */
+}
+
+static int udl_dmabuf_mmap(struct dma_buf *dma_buf,
+			   struct vm_area_struct *vma)
+{
+	/* TODO */
+
+	return -EINVAL;
+}
+
+static struct dma_buf_ops udl_dmabuf_ops = {
+	.attach			= udl_attach_dma_buf,
+	.detach			= udl_detach_dma_buf,
+	.map_dma_buf		= udl_map_dma_buf,
+	.unmap_dma_buf		= udl_unmap_dma_buf,
+	.kmap			= udl_dmabuf_kmap,
+	.kmap_atomic		= udl_dmabuf_kmap_atomic,
+	.kunmap			= udl_dmabuf_kunmap,
+	.kunmap_atomic		= udl_dmabuf_kunmap_atomic,
+	.mmap			= udl_dmabuf_mmap,
+	.release		= drm_gem_dmabuf_release,
+};
+
+struct dma_buf *udl_gem_prime_export(struct drm_device *dev,
+				     struct drm_gem_object *obj, int flags)
+{
+	return dma_buf_export(obj, &udl_dmabuf_ops, obj->size, flags, NULL);
+}
+
+static int udl_prime_create(struct drm_device *dev,
+			    size_t size,
+			    struct sg_table *sg,
+			    struct udl_gem_object **obj_p)
+{
+	struct udl_gem_object *obj;
+	int npages;
+
+	npages = size / PAGE_SIZE;
+
+	*obj_p = NULL;
+	obj = udl_gem_alloc_object(dev, npages * PAGE_SIZE);
+	if (!obj)
+		return -ENOMEM;
+
+	obj->sg = sg;
+	obj->pages = drm_malloc_ab(npages, sizeof(struct page *));
+	if (obj->pages == NULL) {
+		DRM_ERROR("obj pages is NULL %d\n", npages);
+		return -ENOMEM;
+	}
+
+	drm_prime_sg_to_page_addr_arrays(sg, obj->pages, NULL, npages);
+
+	*obj_p = obj;
+	return 0;
+}
+
+struct drm_gem_object *udl_gem_prime_import(struct drm_device *dev,
+				struct dma_buf *dma_buf)
+{
+	struct dma_buf_attachment *attach;
+	struct sg_table *sg;
+	struct udl_gem_object *uobj;
+	int ret;
+
+	/* need to attach */
+	get_device(dev->dev);
+	attach = dma_buf_attach(dma_buf, dev->dev);
+	if (IS_ERR(attach)) {
+		put_device(dev->dev);
+		return ERR_CAST(attach);
+	}
+
+	get_dma_buf(dma_buf);
+
+	sg = dma_buf_map_attachment(attach, DMA_BIDIRECTIONAL);
+	if (IS_ERR(sg)) {
+		ret = PTR_ERR(sg);
+		goto fail_detach;
+	}
+
+	ret = udl_prime_create(dev, dma_buf->size, sg, &uobj);
+	if (ret)
+		goto fail_unmap;
+
+	uobj->base.import_attach = attach;
+	uobj->flags = UDL_BO_WC;
+
+	return &uobj->base;
+
+fail_unmap:
+	dma_buf_unmap_attachment(attach, sg, DMA_BIDIRECTIONAL);
+fail_detach:
+	dma_buf_detach(dma_buf, attach);
+	dma_buf_put(dma_buf);
+	put_device(dev->dev);
+	return ERR_PTR(ret);
+}
diff --git a/drivers/gpu/drm/udl/udl_drv.c b/drivers/gpu/drm/udl/udl_drv.c
index 3ddd6cd98ac1..d5728ec85254 100644
--- a/drivers/gpu/drm/udl/udl_drv.c
+++ b/drivers/gpu/drm/udl/udl_drv.c
@@ -7,48 +7,13 @@
  */
 
 #include <linux/module.h>
-#include <drm/drm_usb.h>
+#include <drm/drmP.h>
 #include <drm/drm_crtc_helper.h>
 #include "udl_drv.h"
 
-static struct drm_driver driver;
-
-/*
- * There are many DisplayLink-based graphics products, all with unique PIDs.
- * So we match on DisplayLink's VID + Vendor-Defined Interface Class (0xff)
- * We also require a match on SubClass (0x00) and Protocol (0x00),
- * which is compatible with all known USB 2.0 era graphics chips and firmware,
- * but allows DisplayLink to increment those for any future incompatible chips
- */
-static struct usb_device_id id_table[] = {
-	{.idVendor = 0x17e9, .bInterfaceClass = 0xff,
-	 .bInterfaceSubClass = 0x00,
-	 .bInterfaceProtocol = 0x00,
-	 .match_flags = USB_DEVICE_ID_MATCH_VENDOR |
-			USB_DEVICE_ID_MATCH_INT_CLASS |
-			USB_DEVICE_ID_MATCH_INT_SUBCLASS |
-			USB_DEVICE_ID_MATCH_INT_PROTOCOL,},
-	{},
-};
-MODULE_DEVICE_TABLE(usb, id_table);
-
-MODULE_LICENSE("GPL");
-
-static int udl_usb_probe(struct usb_interface *interface,
-			 const struct usb_device_id *id)
+static int udl_driver_set_busid(struct drm_device *d, struct drm_master *m)
 {
-	return drm_get_usb_dev(interface, id, &driver);
-}
-
-static void udl_usb_disconnect(struct usb_interface *interface)
-{
-	struct drm_device *dev = usb_get_intfdata(interface);
-
-	drm_kms_helper_poll_disable(dev);
-	drm_connector_unplug_all(dev);
-	udl_fbdev_unplug(dev);
-	udl_drop_usb(dev);
-	drm_unplug_dev(dev);
+	return 0;
 }
 
 static const struct vm_operations_struct udl_gem_vm_ops = {
@@ -75,6 +40,7 @@ static struct drm_driver driver = {
 	.driver_features = DRIVER_MODESET | DRIVER_GEM | DRIVER_PRIME,
 	.load = udl_driver_load,
 	.unload = udl_driver_unload,
+	.set_busid = udl_driver_set_busid,
 
 	/* gem hooks */
 	.gem_free_object = udl_gem_free_object,
@@ -85,7 +51,9 @@ static struct drm_driver driver = {
 	.dumb_destroy = drm_gem_dumb_destroy,
 	.fops = &udl_driver_fops,
 
+	.prime_handle_to_fd = drm_gem_prime_handle_to_fd,
 	.prime_fd_to_handle = drm_gem_prime_fd_to_handle,
+	.gem_prime_export = udl_gem_prime_export,
 	.gem_prime_import = udl_gem_prime_import,
 
 	.name = DRIVER_NAME,
@@ -96,6 +64,61 @@ static struct drm_driver driver = {
 	.patchlevel = DRIVER_PATCHLEVEL,
 };
 
+static int udl_usb_probe(struct usb_interface *interface,
+			 const struct usb_device_id *id)
+{
+	struct usb_device *udev = interface_to_usbdev(interface);
+	struct drm_device *dev;
+	int r;
+
+	dev = drm_dev_alloc(&driver, &interface->dev);
+	if (!dev)
+		return -ENOMEM;
+
+	r = drm_dev_register(dev, (unsigned long)udev);
+	if (r)
+		goto err_free;
+
+	usb_set_intfdata(interface, dev);
+	DRM_INFO("Initialized udl on minor %d\n", dev->primary->index);
+
+	return 0;
+
+err_free:
+	drm_dev_unref(dev);
+	return r;
+}
+
+static void udl_usb_disconnect(struct usb_interface *interface)
+{
+	struct drm_device *dev = usb_get_intfdata(interface);
+
+	drm_kms_helper_poll_disable(dev);
+	drm_connector_unplug_all(dev);
+	udl_fbdev_unplug(dev);
+	udl_drop_usb(dev);
+	drm_unplug_dev(dev);
+}
+
+/*
+ * There are many DisplayLink-based graphics products, all with unique PIDs.
+ * So we match on DisplayLink's VID + Vendor-Defined Interface Class (0xff)
+ * We also require a match on SubClass (0x00) and Protocol (0x00),
+ * which is compatible with all known USB 2.0 era graphics chips and firmware,
+ * but allows DisplayLink to increment those for any future incompatible chips
+ */
+static struct usb_device_id id_table[] = {
+	{.idVendor = 0x17e9, .bInterfaceClass = 0xff,
+	 .bInterfaceSubClass = 0x00,
+	 .bInterfaceProtocol = 0x00,
+	 .match_flags = USB_DEVICE_ID_MATCH_VENDOR |
+			USB_DEVICE_ID_MATCH_INT_CLASS |
+			USB_DEVICE_ID_MATCH_INT_SUBCLASS |
+			USB_DEVICE_ID_MATCH_INT_PROTOCOL,},
+	{},
+};
+MODULE_DEVICE_TABLE(usb, id_table);
+
 static struct usb_driver udl_driver = {
 	.name = "udl",
 	.probe = udl_usb_probe,
@@ -105,13 +128,14 @@ static struct usb_driver udl_driver = {
 
 static int __init udl_init(void)
 {
-	return drm_usb_init(&driver, &udl_driver);
+	return usb_register(&udl_driver);
 }
 
 static void __exit udl_exit(void)
 {
-	drm_usb_exit(&driver, &udl_driver);
+	usb_deregister(&udl_driver);
 }
 
 module_init(udl_init);
 module_exit(udl_exit);
+MODULE_LICENSE("GPL");
diff --git a/drivers/gpu/drm/udl/udl_drv.h b/drivers/gpu/drm/udl/udl_drv.h
index 1fbf7b357f16..80adbac82bde 100644
--- a/drivers/gpu/drm/udl/udl_drv.h
+++ b/drivers/gpu/drm/udl/udl_drv.h
@@ -15,6 +15,7 @@
 #define UDL_DRV_H
 
 #include <linux/usb.h>
+#include <drm/drm_gem.h>
 
 #define DRIVER_NAME		"udl"
 #define DRIVER_DESC		"DisplayLink"
@@ -24,6 +25,9 @@
 #define DRIVER_MINOR		0
 #define DRIVER_PATCHLEVEL	1
 
+#define UDL_BO_CACHEABLE		(1 << 0)
+#define UDL_BO_WC		(1 << 1)
+
 struct udl_device;
 
 struct urb_node {
@@ -47,6 +51,7 @@ struct udl_fbdev;
 struct udl_device {
 	struct device *dev;
 	struct drm_device *ddev;
+	struct usb_device *udev;
 
 	int sku_pixel_limit;
 
@@ -67,6 +72,7 @@ struct udl_gem_object {
 	struct page **pages;
 	void *vmapping;
 	struct sg_table *sg;
+	unsigned int flags;
 };
 
 #define to_udl_bo(x) container_of(x, struct udl_gem_object, base)
@@ -118,9 +124,13 @@ int udl_gem_mmap(struct drm_file *file_priv, struct drm_device *dev,
 void udl_gem_free_object(struct drm_gem_object *gem_obj);
 struct udl_gem_object *udl_gem_alloc_object(struct drm_device *dev,
 					    size_t size);
+struct dma_buf *udl_gem_prime_export(struct drm_device *dev,
+				     struct drm_gem_object *obj, int flags);
 struct drm_gem_object *udl_gem_prime_import(struct drm_device *dev,
 				struct dma_buf *dma_buf);
 
+int udl_gem_get_pages(struct udl_gem_object *obj);
+void udl_gem_put_pages(struct udl_gem_object *obj);
 int udl_gem_vmap(struct udl_gem_object *obj);
 void udl_gem_vunmap(struct udl_gem_object *obj);
 int udl_drm_gem_mmap(struct file *filp, struct vm_area_struct *vma);
diff --git a/drivers/gpu/drm/udl/udl_fb.c b/drivers/gpu/drm/udl/udl_fb.c
index d1da339843ca..8cbcb4589bd3 100644
--- a/drivers/gpu/drm/udl/udl_fb.c
+++ b/drivers/gpu/drm/udl/udl_fb.c
@@ -472,7 +472,8 @@ udl_framebuffer_init(struct drm_device *dev,
 static int udlfb_create(struct drm_fb_helper *helper,
 			struct drm_fb_helper_surface_size *sizes)
 {
-	struct udl_fbdev *ufbdev = (struct udl_fbdev *)helper;
+	struct udl_fbdev *ufbdev =
+		container_of(helper, struct udl_fbdev, helper);
 	struct drm_device *dev = ufbdev->helper.dev;
 	struct fb_info *info;
 	struct device *device = dev->dev;
diff --git a/drivers/gpu/drm/udl/udl_gem.c b/drivers/gpu/drm/udl/udl_gem.c
index 8044f5fb7c49..2a0a784ab6ee 100644
--- a/drivers/gpu/drm/udl/udl_gem.c
+++ b/drivers/gpu/drm/udl/udl_gem.c
@@ -25,6 +25,7 @@ struct udl_gem_object *udl_gem_alloc_object(struct drm_device *dev,
 		return NULL;
 	}
 
+	obj->flags = UDL_BO_CACHEABLE;
 	return obj;
 }
 
@@ -56,6 +57,23 @@ udl_gem_create(struct drm_file *file,
 	return 0;
 }
 
+static void update_vm_cache_attr(struct udl_gem_object *obj,
+				 struct vm_area_struct *vma)
+{
+	DRM_DEBUG_KMS("flags = 0x%x\n", obj->flags);
+
+	/* non-cacheable as default. */
+	if (obj->flags & UDL_BO_CACHEABLE) {
+		vma->vm_page_prot = vm_get_page_prot(vma->vm_flags);
+	} else if (obj->flags & UDL_BO_WC) {
+		vma->vm_page_prot =
+			pgprot_writecombine(vm_get_page_prot(vma->vm_flags));
+	} else {
+		vma->vm_page_prot =
+			pgprot_noncached(vm_get_page_prot(vma->vm_flags));
+	}
+}
+
 int udl_dumb_create(struct drm_file *file,
 		    struct drm_device *dev,
 		    struct drm_mode_create_dumb *args)
@@ -77,6 +95,8 @@ int udl_drm_gem_mmap(struct file *filp, struct vm_area_struct *vma)
 	vma->vm_flags &= ~VM_PFNMAP;
 	vma->vm_flags |= VM_MIXEDMAP;
 
+	update_vm_cache_attr(to_udl_bo(vma->vm_private_data), vma);
+
 	return ret;
 }
 
@@ -107,7 +127,7 @@ int udl_gem_fault(struct vm_area_struct *vma, struct vm_fault *vmf)
 	}
 }
 
-static int udl_gem_get_pages(struct udl_gem_object *obj)
+int udl_gem_get_pages(struct udl_gem_object *obj)
 {
 	struct page **pages;
 
@@ -123,7 +143,7 @@ static int udl_gem_get_pages(struct udl_gem_object *obj)
 	return 0;
 }
 
-static void udl_gem_put_pages(struct udl_gem_object *obj)
+void udl_gem_put_pages(struct udl_gem_object *obj)
 {
 	if (obj->base.import_attach) {
 		drm_free_large(obj->pages);
@@ -164,8 +184,7 @@ void udl_gem_vunmap(struct udl_gem_object *obj)
 		return;
 	}
 
-	if (obj->vmapping)
-		vunmap(obj->vmapping);
+	vunmap(obj->vmapping);
 
 	udl_gem_put_pages(obj);
 }
@@ -220,73 +239,3 @@ unlock:
 	mutex_unlock(&dev->struct_mutex);
 	return ret;
 }
-
-static int udl_prime_create(struct drm_device *dev,
-			    size_t size,
-			    struct sg_table *sg,
-			    struct udl_gem_object **obj_p)
-{
-	struct udl_gem_object *obj;
-	int npages;
-
-	npages = size / PAGE_SIZE;
-
-	*obj_p = NULL;
-	obj = udl_gem_alloc_object(dev, npages * PAGE_SIZE);
-	if (!obj)
-		return -ENOMEM;
-
-	obj->sg = sg;
-	obj->pages = drm_malloc_ab(npages, sizeof(struct page *));
-	if (obj->pages == NULL) {
-		DRM_ERROR("obj pages is NULL %d\n", npages);
-		return -ENOMEM;
-	}
-
-	drm_prime_sg_to_page_addr_arrays(sg, obj->pages, NULL, npages);
-
-	*obj_p = obj;
-	return 0;
-}
-
-struct drm_gem_object *udl_gem_prime_import(struct drm_device *dev,
-				struct dma_buf *dma_buf)
-{
-	struct dma_buf_attachment *attach;
-	struct sg_table *sg;
-	struct udl_gem_object *uobj;
-	int ret;
-
-	/* need to attach */
-	get_device(dev->dev);
-	attach = dma_buf_attach(dma_buf, dev->dev);
-	if (IS_ERR(attach)) {
-		put_device(dev->dev);
-		return ERR_CAST(attach);
-	}
-
-	get_dma_buf(dma_buf);
-
-	sg = dma_buf_map_attachment(attach, DMA_BIDIRECTIONAL);
-	if (IS_ERR(sg)) {
-		ret = PTR_ERR(sg);
-		goto fail_detach;
-	}
-
-	ret = udl_prime_create(dev, dma_buf->size, sg, &uobj);
-	if (ret) {
-		goto fail_unmap;
-	}
-
-	uobj->base.import_attach = attach;
-
-	return &uobj->base;
-
-fail_unmap:
-	dma_buf_unmap_attachment(attach, sg, DMA_BIDIRECTIONAL);
-fail_detach:
-	dma_buf_detach(dma_buf, attach);
-	dma_buf_put(dma_buf);
-	put_device(dev->dev);
-	return ERR_PTR(ret);
-}
diff --git a/drivers/gpu/drm/udl/udl_main.c b/drivers/gpu/drm/udl/udl_main.c
index 42795674bc07..33dbfb2c4748 100644
--- a/drivers/gpu/drm/udl/udl_main.c
+++ b/drivers/gpu/drm/udl/udl_main.c
@@ -202,7 +202,7 @@ static int udl_alloc_urb_list(struct drm_device *dev, int count, size_t size)
 		}
 		unode->urb = urb;
 
-		buf = usb_alloc_coherent(udl->ddev->usbdev, MAX_TRANSFER, GFP_KERNEL,
+		buf = usb_alloc_coherent(udl->udev, MAX_TRANSFER, GFP_KERNEL,
 					 &urb->transfer_dma);
 		if (!buf) {
 			kfree(unode);
@@ -211,7 +211,7 @@ static int udl_alloc_urb_list(struct drm_device *dev, int count, size_t size)
 		}
 
 		/* urb->transfer_buffer_length set to actual before submit */
-		usb_fill_bulk_urb(urb, udl->ddev->usbdev, usb_sndbulkpipe(udl->ddev->usbdev, 1),
+		usb_fill_bulk_urb(urb, udl->udev, usb_sndbulkpipe(udl->udev, 1),
 			buf, size, udl_urb_completion, unode);
 		urb->transfer_flags |= URB_NO_TRANSFER_DMA_MAP;
 
@@ -282,6 +282,7 @@ int udl_submit_urb(struct drm_device *dev, struct urb *urb, size_t len)
 
 int udl_driver_load(struct drm_device *dev, unsigned long flags)
 {
+	struct usb_device *udev = (void*)flags;
 	struct udl_device *udl;
 	int ret = -ENOMEM;
 
@@ -290,10 +291,11 @@ int udl_driver_load(struct drm_device *dev, unsigned long flags)
 	if (!udl)
 		return -ENOMEM;
 
+	udl->udev = udev;
 	udl->ddev = dev;
 	dev->dev_private = udl;
 
-	if (!udl_parse_vendor_descriptor(dev, dev->usbdev)) {
+	if (!udl_parse_vendor_descriptor(dev, udl->udev)) {
 		ret = -ENODEV;
 		DRM_ERROR("firmware not recognized. Assume incompatible device\n");
 		goto err;
diff --git a/drivers/gpu/drm/udl/udl_modeset.c b/drivers/gpu/drm/udl/udl_modeset.c
index dc145d320b25..1701f1dfb23f 100644
--- a/drivers/gpu/drm/udl/udl_modeset.c
+++ b/drivers/gpu/drm/udl/udl_modeset.c
@@ -14,6 +14,7 @@
 #include <drm/drmP.h>
 #include <drm/drm_crtc.h>
 #include <drm/drm_crtc_helper.h>
+#include <drm/drm_plane_helper.h>
 #include "udl_drv.h"
 
 /*
diff --git a/drivers/gpu/drm/via/via_dma.c b/drivers/gpu/drm/via/via_dma.c
index 6fc0648dd37f..d17d8f245c1a 100644
--- a/drivers/gpu/drm/via/via_dma.c
+++ b/drivers/gpu/drm/via/via_dma.c
@@ -161,7 +161,7 @@ int via_dma_cleanup(struct drm_device *dev)
 		if (dev_priv->ring.virtual_start) {
 			via_cmdbuf_reset(dev_priv);
 
-			drm_core_ioremapfree(&dev_priv->ring.map, dev);
+			drm_legacy_ioremapfree(&dev_priv->ring.map, dev);
 			dev_priv->ring.virtual_start = NULL;
 		}
 
@@ -200,7 +200,7 @@ static int via_initialize(struct drm_device *dev,
 	dev_priv->ring.map.flags = 0;
 	dev_priv->ring.map.mtrr = 0;
 
-	drm_core_ioremap(&dev_priv->ring.map, dev);
+	drm_legacy_ioremap(&dev_priv->ring.map, dev);
 
 	if (dev_priv->ring.map.handle == NULL) {
 		via_dma_cleanup(dev);
diff --git a/drivers/gpu/drm/via/via_drv.c b/drivers/gpu/drm/via/via_drv.c
index 50abc2adfaee..ed8aa8ff861a 100644
--- a/drivers/gpu/drm/via/via_drv.c
+++ b/drivers/gpu/drm/via/via_drv.c
@@ -62,7 +62,7 @@ static const struct file_operations via_driver_fops = {
 	.open = drm_open,
 	.release = drm_release,
 	.unlocked_ioctl = drm_ioctl,
-	.mmap = drm_mmap,
+	.mmap = drm_legacy_mmap,
 	.poll = drm_poll,
 #ifdef CONFIG_COMPAT
 	.compat_ioctl = drm_compat_ioctl,
@@ -79,6 +79,7 @@ static struct drm_driver driver = {
 	.open = via_driver_open,
 	.preclose = via_reclaim_buffers_locked,
 	.postclose = via_driver_postclose,
+	.set_busid = drm_pci_set_busid,
 	.context_dtor = via_final_context,
 	.get_vblank_counter = via_get_vblank_counter,
 	.enable_vblank = via_enable_vblank,
diff --git a/drivers/gpu/drm/via/via_drv.h b/drivers/gpu/drm/via/via_drv.h
index ad0273256beb..ef8c500b4a00 100644
--- a/drivers/gpu/drm/via/via_drv.h
+++ b/drivers/gpu/drm/via/via_drv.h
@@ -25,6 +25,8 @@
 #define _VIA_DRV_H_
 
 #include <drm/drm_mm.h>
+#include <drm/drm_legacy.h>
+
 #define DRIVER_AUTHOR	"Various"
 
 #define DRIVER_NAME		"via"
diff --git a/drivers/gpu/drm/via/via_map.c b/drivers/gpu/drm/via/via_map.c
index d0ab3fb32acd..0b3522dba6e8 100644
--- a/drivers/gpu/drm/via/via_map.c
+++ b/drivers/gpu/drm/via/via_map.c
@@ -31,7 +31,7 @@ static int via_do_init_map(struct drm_device *dev, drm_via_init_t *init)
 
 	DRM_DEBUG("\n");
 
-	dev_priv->sarea = drm_getsarea(dev);
+	dev_priv->sarea = drm_legacy_getsarea(dev);
 	if (!dev_priv->sarea) {
 		DRM_ERROR("could not find sarea!\n");
 		dev->dev_private = (void *)dev_priv;
@@ -39,14 +39,14 @@ static int via_do_init_map(struct drm_device *dev, drm_via_init_t *init)
 		return -EINVAL;
 	}
 
-	dev_priv->fb = drm_core_findmap(dev, init->fb_offset);
+	dev_priv->fb = drm_legacy_findmap(dev, init->fb_offset);
 	if (!dev_priv->fb) {
 		DRM_ERROR("could not find framebuffer!\n");
 		dev->dev_private = (void *)dev_priv;
 		via_do_cleanup_map(dev);
 		return -EINVAL;
 	}
-	dev_priv->mmio = drm_core_findmap(dev, init->mmio_offset);
+	dev_priv->mmio = drm_legacy_findmap(dev, init->mmio_offset);
 	if (!dev_priv->mmio) {
 		DRM_ERROR("could not find mmio region!\n");
 		dev->dev_private = (void *)dev_priv;
diff --git a/drivers/gpu/drm/via/via_mm.c b/drivers/gpu/drm/via/via_mm.c
index d70b1e1544bf..4f20742e7788 100644
--- a/drivers/gpu/drm/via/via_mm.c
+++ b/drivers/gpu/drm/via/via_mm.c
@@ -211,12 +211,12 @@ void via_reclaim_buffers_locked(struct drm_device *dev,
 	if (!(file->minor->master && file->master->lock.hw_lock))
 		return;
 
-	drm_idlelock_take(&file->master->lock);
+	drm_legacy_idlelock_take(&file->master->lock);
 
 	mutex_lock(&dev->struct_mutex);
 	if (list_empty(&file_priv->obj_list)) {
 		mutex_unlock(&dev->struct_mutex);
-		drm_idlelock_release(&file->master->lock);
+		drm_legacy_idlelock_release(&file->master->lock);
 
 		return;
 	}
@@ -231,7 +231,7 @@ void via_reclaim_buffers_locked(struct drm_device *dev,
 	}
 	mutex_unlock(&dev->struct_mutex);
 
-	drm_idlelock_release(&file->master->lock);
+	drm_legacy_idlelock_release(&file->master->lock);
 
 	return;
 }
diff --git a/drivers/gpu/drm/via/via_verifier.c b/drivers/gpu/drm/via/via_verifier.c
index 9dbc92bd1512..0677bbf4ec7e 100644
--- a/drivers/gpu/drm/via/via_verifier.c
+++ b/drivers/gpu/drm/via/via_verifier.c
@@ -31,6 +31,7 @@
 #include "via_3d_reg.h"
 #include <drm/drmP.h>
 #include <drm/via_drm.h>
+#include <drm/drm_legacy.h>
 #include "via_verifier.h"
 #include "via_drv.h"
 
diff --git a/drivers/gpu/drm/vmwgfx/svga_reg.h b/drivers/gpu/drm/vmwgfx/svga_reg.h
index 11323dd5196f..e4259c2c1acc 100644
--- a/drivers/gpu/drm/vmwgfx/svga_reg.h
+++ b/drivers/gpu/drm/vmwgfx/svga_reg.h
@@ -35,7 +35,6 @@
 /*
  * PCI device IDs.
  */
-#define PCI_VENDOR_ID_VMWARE            0x15AD
 #define PCI_DEVICE_ID_VMWARE_SVGA2      0x0405
 
 /*
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_buffer.c b/drivers/gpu/drm/vmwgfx/vmwgfx_buffer.c
index 6327cfc36805..cff2bf9db9d2 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_buffer.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_buffer.c
@@ -30,66 +30,101 @@
 #include <drm/ttm/ttm_placement.h>
 #include <drm/ttm/ttm_page_alloc.h>
 
-static uint32_t vram_placement_flags = TTM_PL_FLAG_VRAM |
-	TTM_PL_FLAG_CACHED;
-
-static uint32_t vram_ne_placement_flags = TTM_PL_FLAG_VRAM |
-	TTM_PL_FLAG_CACHED |
-	TTM_PL_FLAG_NO_EVICT;
+static struct ttm_place vram_placement_flags = {
+	.fpfn = 0,
+	.lpfn = 0,
+	.flags = TTM_PL_FLAG_VRAM | TTM_PL_FLAG_CACHED
+};
 
-static uint32_t sys_placement_flags = TTM_PL_FLAG_SYSTEM |
-	TTM_PL_FLAG_CACHED;
+static struct ttm_place vram_ne_placement_flags = {
+	.fpfn = 0,
+	.lpfn = 0,
+	.flags = TTM_PL_FLAG_VRAM | TTM_PL_FLAG_CACHED | TTM_PL_FLAG_NO_EVICT
+};
 
-static uint32_t sys_ne_placement_flags = TTM_PL_FLAG_SYSTEM |
-	TTM_PL_FLAG_CACHED |
-	TTM_PL_FLAG_NO_EVICT;
+static struct ttm_place sys_placement_flags = {
+	.fpfn = 0,
+	.lpfn = 0,
+	.flags = TTM_PL_FLAG_SYSTEM | TTM_PL_FLAG_CACHED
+};
 
-static uint32_t gmr_placement_flags = VMW_PL_FLAG_GMR |
-	TTM_PL_FLAG_CACHED;
+static struct ttm_place sys_ne_placement_flags = {
+	.fpfn = 0,
+	.lpfn = 0,
+	.flags = TTM_PL_FLAG_SYSTEM | TTM_PL_FLAG_CACHED | TTM_PL_FLAG_NO_EVICT
+};
 
-static uint32_t gmr_ne_placement_flags = VMW_PL_FLAG_GMR |
-	TTM_PL_FLAG_CACHED |
-	TTM_PL_FLAG_NO_EVICT;
+static struct ttm_place gmr_placement_flags = {
+	.fpfn = 0,
+	.lpfn = 0,
+	.flags = VMW_PL_FLAG_GMR | TTM_PL_FLAG_CACHED
+};
 
-static uint32_t mob_placement_flags = VMW_PL_FLAG_MOB |
-	TTM_PL_FLAG_CACHED;
+static struct ttm_place gmr_ne_placement_flags = {
+	.fpfn = 0,
+	.lpfn = 0,
+	.flags = VMW_PL_FLAG_GMR | TTM_PL_FLAG_CACHED | TTM_PL_FLAG_NO_EVICT
+};
 
-struct ttm_placement vmw_vram_placement = {
+static struct ttm_place mob_placement_flags = {
 	.fpfn = 0,
 	.lpfn = 0,
+	.flags = VMW_PL_FLAG_MOB | TTM_PL_FLAG_CACHED
+};
+
+struct ttm_placement vmw_vram_placement = {
 	.num_placement = 1,
 	.placement = &vram_placement_flags,
 	.num_busy_placement = 1,
 	.busy_placement = &vram_placement_flags
 };
 
-static uint32_t vram_gmr_placement_flags[] = {
-	TTM_PL_FLAG_VRAM | TTM_PL_FLAG_CACHED,
-	VMW_PL_FLAG_GMR | TTM_PL_FLAG_CACHED
+static struct ttm_place vram_gmr_placement_flags[] = {
+	{
+		.fpfn = 0,
+		.lpfn = 0,
+		.flags = TTM_PL_FLAG_VRAM | TTM_PL_FLAG_CACHED
+	}, {
+		.fpfn = 0,
+		.lpfn = 0,
+		.flags = VMW_PL_FLAG_GMR | TTM_PL_FLAG_CACHED
+	}
 };
 
-static uint32_t gmr_vram_placement_flags[] = {
-	VMW_PL_FLAG_GMR | TTM_PL_FLAG_CACHED,
-	TTM_PL_FLAG_VRAM | TTM_PL_FLAG_CACHED
+static struct ttm_place gmr_vram_placement_flags[] = {
+	{
+		.fpfn = 0,
+		.lpfn = 0,
+		.flags = VMW_PL_FLAG_GMR | TTM_PL_FLAG_CACHED
+	}, {
+		.fpfn = 0,
+		.lpfn = 0,
+		.flags = TTM_PL_FLAG_VRAM | TTM_PL_FLAG_CACHED
+	}
 };
 
 struct ttm_placement vmw_vram_gmr_placement = {
-	.fpfn = 0,
-	.lpfn = 0,
 	.num_placement = 2,
 	.placement = vram_gmr_placement_flags,
 	.num_busy_placement = 1,
 	.busy_placement = &gmr_placement_flags
 };
 
-static uint32_t vram_gmr_ne_placement_flags[] = {
-	TTM_PL_FLAG_VRAM | TTM_PL_FLAG_CACHED | TTM_PL_FLAG_NO_EVICT,
-	VMW_PL_FLAG_GMR | TTM_PL_FLAG_CACHED | TTM_PL_FLAG_NO_EVICT
+static struct ttm_place vram_gmr_ne_placement_flags[] = {
+	{
+		.fpfn = 0,
+		.lpfn = 0,
+		.flags = TTM_PL_FLAG_VRAM | TTM_PL_FLAG_CACHED |
+			 TTM_PL_FLAG_NO_EVICT
+	}, {
+		.fpfn = 0,
+		.lpfn = 0,
+		.flags = VMW_PL_FLAG_GMR | TTM_PL_FLAG_CACHED |
+			 TTM_PL_FLAG_NO_EVICT
+	}
 };
 
 struct ttm_placement vmw_vram_gmr_ne_placement = {
-	.fpfn = 0,
-	.lpfn = 0,
 	.num_placement = 2,
 	.placement = vram_gmr_ne_placement_flags,
 	.num_busy_placement = 1,
@@ -97,8 +132,6 @@ struct ttm_placement vmw_vram_gmr_ne_placement = {
 };
 
 struct ttm_placement vmw_vram_sys_placement = {
-	.fpfn = 0,
-	.lpfn = 0,
 	.num_placement = 1,
 	.placement = &vram_placement_flags,
 	.num_busy_placement = 1,
@@ -106,8 +139,6 @@ struct ttm_placement vmw_vram_sys_placement = {
 };
 
 struct ttm_placement vmw_vram_ne_placement = {
-	.fpfn = 0,
-	.lpfn = 0,
 	.num_placement = 1,
 	.placement = &vram_ne_placement_flags,
 	.num_busy_placement = 1,
@@ -115,8 +146,6 @@ struct ttm_placement vmw_vram_ne_placement = {
 };
 
 struct ttm_placement vmw_sys_placement = {
-	.fpfn = 0,
-	.lpfn = 0,
 	.num_placement = 1,
 	.placement = &sys_placement_flags,
 	.num_busy_placement = 1,
@@ -124,24 +153,33 @@ struct ttm_placement vmw_sys_placement = {
 };
 
 struct ttm_placement vmw_sys_ne_placement = {
-	.fpfn = 0,
-	.lpfn = 0,
 	.num_placement = 1,
 	.placement = &sys_ne_placement_flags,
 	.num_busy_placement = 1,
 	.busy_placement = &sys_ne_placement_flags
 };
 
-static uint32_t evictable_placement_flags[] = {
-	TTM_PL_FLAG_SYSTEM | TTM_PL_FLAG_CACHED,
-	TTM_PL_FLAG_VRAM | TTM_PL_FLAG_CACHED,
-	VMW_PL_FLAG_GMR | TTM_PL_FLAG_CACHED,
-	VMW_PL_FLAG_MOB | TTM_PL_FLAG_CACHED
+static struct ttm_place evictable_placement_flags[] = {
+	{
+		.fpfn = 0,
+		.lpfn = 0,
+		.flags = TTM_PL_FLAG_SYSTEM | TTM_PL_FLAG_CACHED
+	}, {
+		.fpfn = 0,
+		.lpfn = 0,
+		.flags = TTM_PL_FLAG_VRAM | TTM_PL_FLAG_CACHED
+	}, {
+		.fpfn = 0,
+		.lpfn = 0,
+		.flags = VMW_PL_FLAG_GMR | TTM_PL_FLAG_CACHED
+	}, {
+		.fpfn = 0,
+		.lpfn = 0,
+		.flags = VMW_PL_FLAG_MOB | TTM_PL_FLAG_CACHED
+	}
 };
 
 struct ttm_placement vmw_evictable_placement = {
-	.fpfn = 0,
-	.lpfn = 0,
 	.num_placement = 4,
 	.placement = evictable_placement_flags,
 	.num_busy_placement = 1,
@@ -149,8 +187,6 @@ struct ttm_placement vmw_evictable_placement = {
 };
 
 struct ttm_placement vmw_srf_placement = {
-	.fpfn = 0,
-	.lpfn = 0,
 	.num_placement = 1,
 	.num_busy_placement = 2,
 	.placement = &gmr_placement_flags,
@@ -158,8 +194,6 @@ struct ttm_placement vmw_srf_placement = {
 };
 
 struct ttm_placement vmw_mob_placement = {
-	.fpfn = 0,
-	.lpfn = 0,
 	.num_placement = 1,
 	.num_busy_placement = 1,
 	.placement = &mob_placement_flags,
@@ -768,44 +802,6 @@ static int vmw_ttm_fault_reserve_notify(struct ttm_buffer_object *bo)
 }
 
 /**
- * FIXME: We're using the old vmware polling method to sync.
- * Do this with fences instead.
- */
-
-static void *vmw_sync_obj_ref(void *sync_obj)
-{
-
-	return (void *)
-		vmw_fence_obj_reference((struct vmw_fence_obj *) sync_obj);
-}
-
-static void vmw_sync_obj_unref(void **sync_obj)
-{
-	vmw_fence_obj_unreference((struct vmw_fence_obj **) sync_obj);
-}
-
-static int vmw_sync_obj_flush(void *sync_obj)
-{
-	vmw_fence_obj_flush((struct vmw_fence_obj *) sync_obj);
-	return 0;
-}
-
-static bool vmw_sync_obj_signaled(void *sync_obj)
-{
-	return	vmw_fence_obj_signaled((struct vmw_fence_obj *) sync_obj,
-				       DRM_VMW_FENCE_FLAG_EXEC);
-
-}
-
-static int vmw_sync_obj_wait(void *sync_obj, bool lazy, bool interruptible)
-{
-	return vmw_fence_obj_wait((struct vmw_fence_obj *) sync_obj,
-				  DRM_VMW_FENCE_FLAG_EXEC,
-				  lazy, interruptible,
-				  VMW_FENCE_WAIT_TIMEOUT);
-}
-
-/**
  * vmw_move_notify - TTM move_notify_callback
  *
  * @bo:             The TTM buffer object about to move.
@@ -829,11 +825,7 @@ static void vmw_move_notify(struct ttm_buffer_object *bo,
  */
 static void vmw_swap_notify(struct ttm_buffer_object *bo)
 {
-	struct ttm_bo_device *bdev = bo->bdev;
-
-	spin_lock(&bdev->fence_lock);
 	ttm_bo_wait(bo, false, false, false);
-	spin_unlock(&bdev->fence_lock);
 }
 
 
@@ -846,11 +838,6 @@ struct ttm_bo_driver vmw_bo_driver = {
 	.evict_flags = vmw_evict_flags,
 	.move = NULL,
 	.verify_access = vmw_verify_access,
-	.sync_obj_signaled = vmw_sync_obj_signaled,
-	.sync_obj_wait = vmw_sync_obj_wait,
-	.sync_obj_flush = vmw_sync_obj_flush,
-	.sync_obj_unref = vmw_sync_obj_unref,
-	.sync_obj_ref = vmw_sync_obj_ref,
 	.move_notify = vmw_move_notify,
 	.swap_notify = vmw_swap_notify,
 	.fault_reserve_notify = &vmw_ttm_fault_reserve_notify,
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_cmdbuf_res.c b/drivers/gpu/drm/vmwgfx/vmwgfx_cmdbuf_res.c
index bfeb4b1f2acf..21e9b7f8dad0 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_cmdbuf_res.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_cmdbuf_res.c
@@ -246,7 +246,8 @@ int vmw_cmdbuf_res_remove(struct vmw_cmdbuf_res_manager *man,
 	struct drm_hash_item *hash;
 	int ret;
 
-	ret = drm_ht_find_item(&man->resources, user_key, &hash);
+	ret = drm_ht_find_item(&man->resources, user_key | (res_type << 24),
+			       &hash);
 	if (likely(ret != 0))
 		return -EINVAL;
 
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_dmabuf.c b/drivers/gpu/drm/vmwgfx/vmwgfx_dmabuf.c
index ed1d51006ab1..914b375763dc 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_dmabuf.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_dmabuf.c
@@ -198,13 +198,19 @@ int vmw_dmabuf_to_start_of_vram(struct vmw_private *dev_priv,
 {
 	struct ttm_buffer_object *bo = &buf->base;
 	struct ttm_placement placement;
+	struct ttm_place place;
 	int ret = 0;
 
 	if (pin)
-		placement = vmw_vram_ne_placement;
+		place = vmw_vram_ne_placement.placement[0];
 	else
-		placement = vmw_vram_placement;
-	placement.lpfn = bo->num_pages;
+		place = vmw_vram_placement.placement[0];
+	place.lpfn = bo->num_pages;
+
+	placement.num_placement = 1;
+	placement.placement = &place;
+	placement.num_busy_placement = 1;
+	placement.busy_placement = &place;
 
 	ret = ttm_write_lock(&dev_priv->reservation_sem, interruptible);
 	if (unlikely(ret != 0))
@@ -293,21 +299,23 @@ void vmw_bo_get_guest_ptr(const struct ttm_buffer_object *bo,
  */
 void vmw_bo_pin(struct ttm_buffer_object *bo, bool pin)
 {
-	uint32_t pl_flags;
+	struct ttm_place pl;
 	struct ttm_placement placement;
 	uint32_t old_mem_type = bo->mem.mem_type;
 	int ret;
 
 	lockdep_assert_held(&bo->resv->lock.base);
 
-	pl_flags = TTM_PL_FLAG_VRAM | VMW_PL_FLAG_GMR | VMW_PL_FLAG_MOB
+	pl.fpfn = 0;
+	pl.lpfn = 0;
+	pl.flags = TTM_PL_FLAG_VRAM | VMW_PL_FLAG_GMR | VMW_PL_FLAG_MOB
 		| TTM_PL_FLAG_SYSTEM | TTM_PL_FLAG_CACHED;
 	if (pin)
-		pl_flags |= TTM_PL_FLAG_NO_EVICT;
+		pl.flags |= TTM_PL_FLAG_NO_EVICT;
 
 	memset(&placement, 0, sizeof(placement));
 	placement.num_placement = 1;
-	placement.placement = &pl_flags;
+	placement.placement = &pl;
 
 	ret = ttm_bo_validate(bo, &placement, false, true);
 
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c
index 18b54acacfbb..7b5d22110f25 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c
@@ -688,7 +688,11 @@ static int vmw_driver_load(struct drm_device *dev, unsigned long chipset)
 		goto out_err0;
 	}
 
-	if (unlikely(dev_priv->prim_bb_mem < dev_priv->vram_size))
+	/*
+	 * Limit back buffer size to VRAM size.  Remove this once
+	 * screen targets are implemented.
+	 */
+	if (dev_priv->prim_bb_mem > dev_priv->vram_size)
 		dev_priv->prim_bb_mem = dev_priv->vram_size;
 
 	mutex_unlock(&dev_priv->hw_mutex);
@@ -885,8 +889,7 @@ static int vmw_driver_unload(struct drm_device *dev)
 
 	if (dev_priv->ctx.res_ht_initialized)
 		drm_ht_remove(&dev_priv->ctx.res_ht);
-	if (dev_priv->ctx.cmd_bounce)
-		vfree(dev_priv->ctx.cmd_bounce);
+	vfree(dev_priv->ctx.cmd_bounce);
 	if (dev_priv->enable_fb) {
 		vmw_fb_close(dev_priv);
 		vmw_kms_restore_vga(dev_priv);
@@ -1059,8 +1062,12 @@ static long vmw_generic_ioctl(struct file *filp, unsigned int cmd,
 
 	vmaster = vmw_master_check(dev, file_priv, flags);
 	if (unlikely(IS_ERR(vmaster))) {
-		DRM_INFO("IOCTL ERROR %d\n", nr);
-		return PTR_ERR(vmaster);
+		ret = PTR_ERR(vmaster);
+
+		if (ret != -ERESTARTSYS)
+			DRM_INFO("IOCTL ERROR Command %d, Error %ld.\n",
+				 nr, ret);
+		return ret;
 	}
 
 	ret = ioctl_func(filp, cmd, arg);
@@ -1418,6 +1425,7 @@ static struct drm_driver driver = {
 	.open = vmw_driver_open,
 	.preclose = vmw_preclose,
 	.postclose = vmw_postclose,
+	.set_busid = drm_pci_set_busid,
 
 	.dumb_create = vmw_dumb_create,
 	.dumb_map_offset = vmw_dumb_map_offset,
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h
index 99f731757c4b..4ee799b43d5d 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h
@@ -342,7 +342,6 @@ struct vmw_sw_context{
 	uint32_t *cmd_bounce;
 	uint32_t cmd_bounce_size;
 	struct list_head resource_list;
-	uint32_t fence_flags;
 	struct ttm_buffer_object *cur_query_bo;
 	struct list_head res_relocations;
 	uint32_t *buf_start;
@@ -704,6 +703,7 @@ extern void *vmw_fifo_reserve(struct vmw_private *dev_priv, uint32_t bytes);
 extern void vmw_fifo_commit(struct vmw_private *dev_priv, uint32_t bytes);
 extern int vmw_fifo_send_fence(struct vmw_private *dev_priv,
 			       uint32_t *seqno);
+extern void vmw_fifo_ping_host_locked(struct vmw_private *, uint32_t reason);
 extern void vmw_fifo_ping_host(struct vmw_private *dev_priv, uint32_t reason);
 extern bool vmw_fifo_have_3d(struct vmw_private *dev_priv);
 extern bool vmw_fifo_have_pitchlock(struct vmw_private *dev_priv);
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_execbuf.c b/drivers/gpu/drm/vmwgfx/vmwgfx_execbuf.c
index 36b871686d3c..33176d05db35 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_execbuf.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_execbuf.c
@@ -346,13 +346,11 @@ static int vmw_bo_to_validate_list(struct vmw_sw_context *sw_context,
 		++sw_context->cur_val_buf;
 		val_buf = &vval_buf->base;
 		val_buf->bo = ttm_bo_reference(bo);
-		val_buf->reserved = false;
+		val_buf->shared = false;
 		list_add_tail(&val_buf->head, &sw_context->validate_nodes);
 		vval_buf->validate_as_mob = validate_as_mob;
 	}
 
-	sw_context->fence_flags |= DRM_VMW_FENCE_FLAG_EXEC;
-
 	if (p_val_node)
 		*p_val_node = val_node;
 
@@ -2337,13 +2335,9 @@ int vmw_execbuf_fence_commands(struct drm_file *file_priv,
 
 	if (p_handle != NULL)
 		ret = vmw_user_fence_create(file_priv, dev_priv->fman,
-					    sequence,
-					    DRM_VMW_FENCE_FLAG_EXEC,
-					    p_fence, p_handle);
+					    sequence, p_fence, p_handle);
 	else
-		ret = vmw_fence_create(dev_priv->fman, sequence,
-				       DRM_VMW_FENCE_FLAG_EXEC,
-				       p_fence);
+		ret = vmw_fence_create(dev_priv->fman, sequence, p_fence);
 
 	if (unlikely(ret != 0 && !synced)) {
 		(void) vmw_fallback_wait(dev_priv, false, false,
@@ -2395,7 +2389,7 @@ vmw_execbuf_copy_fence_user(struct vmw_private *dev_priv,
 		BUG_ON(fence == NULL);
 
 		fence_rep.handle = fence_handle;
-		fence_rep.seqno = fence->seqno;
+		fence_rep.seqno = fence->base.seqno;
 		vmw_update_seqno(dev_priv, &dev_priv->fifo);
 		fence_rep.passed_seqno = dev_priv->last_read_seqno;
 	}
@@ -2416,8 +2410,7 @@ vmw_execbuf_copy_fence_user(struct vmw_private *dev_priv,
 		ttm_ref_object_base_unref(vmw_fp->tfile,
 					  fence_handle, TTM_REF_USAGE);
 		DRM_ERROR("Fence copy error. Syncing.\n");
-		(void) vmw_fence_obj_wait(fence, fence->signal_mask,
-					  false, false,
+		(void) vmw_fence_obj_wait(fence, false, false,
 					  VMW_FENCE_WAIT_TIMEOUT);
 	}
 }
@@ -2469,7 +2462,6 @@ int vmw_execbuf_process(struct drm_file *file_priv,
 	sw_context->fp = vmw_fpriv(file_priv);
 	sw_context->cur_reloc = 0;
 	sw_context->cur_val_buf = 0;
-	sw_context->fence_flags = 0;
 	INIT_LIST_HEAD(&sw_context->resource_list);
 	sw_context->cur_query_bo = dev_priv->pinned_bo;
 	sw_context->last_query_ctx = NULL;
@@ -2495,7 +2487,8 @@ int vmw_execbuf_process(struct drm_file *file_priv,
 	if (unlikely(ret != 0))
 		goto out_err_nores;
 
-	ret = ttm_eu_reserve_buffers(&ticket, &sw_context->validate_nodes);
+	ret = ttm_eu_reserve_buffers(&ticket, &sw_context->validate_nodes,
+				     true, NULL);
 	if (unlikely(ret != 0))
 		goto out_err;
 
@@ -2678,15 +2671,15 @@ void __vmw_execbuf_release_pinned_bo(struct vmw_private *dev_priv,
 	INIT_LIST_HEAD(&validate_list);
 
 	pinned_val.bo = ttm_bo_reference(dev_priv->pinned_bo);
+	pinned_val.shared = false;
 	list_add_tail(&pinned_val.head, &validate_list);
 
 	query_val.bo = ttm_bo_reference(dev_priv->dummy_query_bo);
+	query_val.shared = false;
 	list_add_tail(&query_val.head, &validate_list);
 
-	do {
-		ret = ttm_eu_reserve_buffers(&ticket, &validate_list);
-	} while (ret == -ERESTARTSYS);
-
+	ret = ttm_eu_reserve_buffers(&ticket, &validate_list,
+				     false, NULL);
 	if (unlikely(ret != 0)) {
 		vmw_execbuf_unpin_panic(dev_priv);
 		goto out_no_reserve;
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_fb.c b/drivers/gpu/drm/vmwgfx/vmwgfx_fb.c
index b031b48dbb3c..0a474f391fad 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_fb.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_fb.c
@@ -374,10 +374,16 @@ static int vmw_fb_create_bo(struct vmw_private *vmw_priv,
 			    size_t size, struct vmw_dma_buffer **out)
 {
 	struct vmw_dma_buffer *vmw_bo;
-	struct ttm_placement ne_placement = vmw_vram_ne_placement;
+	struct ttm_place ne_place = vmw_vram_ne_placement.placement[0];
+	struct ttm_placement ne_placement;
 	int ret;
 
-	ne_placement.lpfn = (size + PAGE_SIZE - 1) >> PAGE_SHIFT;
+	ne_placement.num_placement = 1;
+	ne_placement.placement = &ne_place;
+	ne_placement.num_busy_placement = 1;
+	ne_placement.busy_placement = &ne_place;
+
+	ne_place.lpfn = (size + PAGE_SIZE - 1) >> PAGE_SHIFT;
 
 	(void) ttm_write_lock(&vmw_priv->reservation_sem, false);
 
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_fence.c b/drivers/gpu/drm/vmwgfx/vmwgfx_fence.c
index 436b013b4231..b7594cb758af 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_fence.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_fence.c
@@ -35,7 +35,7 @@ struct vmw_fence_manager {
 	struct vmw_private *dev_priv;
 	spinlock_t lock;
 	struct list_head fence_list;
-	struct work_struct work;
+	struct work_struct work, ping_work;
 	u32 user_fence_size;
 	u32 fence_size;
 	u32 event_fence_action_size;
@@ -46,6 +46,7 @@ struct vmw_fence_manager {
 	bool goal_irq_on; /* Protected by @goal_irq_mutex */
 	bool seqno_valid; /* Protected by @lock, and may not be set to true
 			     without the @goal_irq_mutex held. */
+	unsigned ctx;
 };
 
 struct vmw_user_fence {
@@ -80,6 +81,12 @@ struct vmw_event_fence_action {
 	uint32_t *tv_usec;
 };
 
+static struct vmw_fence_manager *
+fman_from_fence(struct vmw_fence_obj *fence)
+{
+	return container_of(fence->base.lock, struct vmw_fence_manager, lock);
+}
+
 /**
  * Note on fencing subsystem usage of irqs:
  * Typically the vmw_fences_update function is called
@@ -102,25 +109,143 @@ struct vmw_event_fence_action {
  * objects with actions attached to them.
  */
 
-static void vmw_fence_obj_destroy_locked(struct kref *kref)
+static void vmw_fence_obj_destroy(struct fence *f)
 {
 	struct vmw_fence_obj *fence =
-		container_of(kref, struct vmw_fence_obj, kref);
+		container_of(f, struct vmw_fence_obj, base);
 
-	struct vmw_fence_manager *fman = fence->fman;
-	unsigned int num_fences;
+	struct vmw_fence_manager *fman = fman_from_fence(fence);
+	unsigned long irq_flags;
 
+	spin_lock_irqsave(&fman->lock, irq_flags);
 	list_del_init(&fence->head);
-	num_fences = --fman->num_fence_objects;
-	spin_unlock_irq(&fman->lock);
-	if (fence->destroy)
-		fence->destroy(fence);
-	else
-		kfree(fence);
+	--fman->num_fence_objects;
+	spin_unlock_irqrestore(&fman->lock, irq_flags);
+	fence->destroy(fence);
+}
 
-	spin_lock_irq(&fman->lock);
+static const char *vmw_fence_get_driver_name(struct fence *f)
+{
+	return "vmwgfx";
+}
+
+static const char *vmw_fence_get_timeline_name(struct fence *f)
+{
+	return "svga";
+}
+
+static void vmw_fence_ping_func(struct work_struct *work)
+{
+	struct vmw_fence_manager *fman =
+		container_of(work, struct vmw_fence_manager, ping_work);
+
+	vmw_fifo_ping_host(fman->dev_priv, SVGA_SYNC_GENERIC);
+}
+
+static bool vmw_fence_enable_signaling(struct fence *f)
+{
+	struct vmw_fence_obj *fence =
+		container_of(f, struct vmw_fence_obj, base);
+
+	struct vmw_fence_manager *fman = fman_from_fence(fence);
+	struct vmw_private *dev_priv = fman->dev_priv;
+
+	__le32 __iomem *fifo_mem = dev_priv->mmio_virt;
+	u32 seqno = ioread32(fifo_mem + SVGA_FIFO_FENCE);
+	if (seqno - fence->base.seqno < VMW_FENCE_WRAP)
+		return false;
+
+	if (mutex_trylock(&dev_priv->hw_mutex)) {
+		vmw_fifo_ping_host_locked(dev_priv, SVGA_SYNC_GENERIC);
+		mutex_unlock(&dev_priv->hw_mutex);
+	} else
+		schedule_work(&fman->ping_work);
+
+	return true;
+}
+
+struct vmwgfx_wait_cb {
+	struct fence_cb base;
+	struct task_struct *task;
+};
+
+static void
+vmwgfx_wait_cb(struct fence *fence, struct fence_cb *cb)
+{
+	struct vmwgfx_wait_cb *wait =
+		container_of(cb, struct vmwgfx_wait_cb, base);
+
+	wake_up_process(wait->task);
+}
+
+static void __vmw_fences_update(struct vmw_fence_manager *fman);
+
+static long vmw_fence_wait(struct fence *f, bool intr, signed long timeout)
+{
+	struct vmw_fence_obj *fence =
+		container_of(f, struct vmw_fence_obj, base);
+
+	struct vmw_fence_manager *fman = fman_from_fence(fence);
+	struct vmw_private *dev_priv = fman->dev_priv;
+	struct vmwgfx_wait_cb cb;
+	long ret = timeout;
+	unsigned long irq_flags;
+
+	if (likely(vmw_fence_obj_signaled(fence)))
+		return timeout;
+
+	vmw_fifo_ping_host(dev_priv, SVGA_SYNC_GENERIC);
+	vmw_seqno_waiter_add(dev_priv);
+
+	spin_lock_irqsave(f->lock, irq_flags);
+
+	if (intr && signal_pending(current)) {
+		ret = -ERESTARTSYS;
+		goto out;
+	}
+
+	cb.base.func = vmwgfx_wait_cb;
+	cb.task = current;
+	list_add(&cb.base.node, &f->cb_list);
+
+	while (ret > 0) {
+		__vmw_fences_update(fman);
+		if (test_bit(FENCE_FLAG_SIGNALED_BIT, &f->flags))
+			break;
+
+		if (intr)
+			__set_current_state(TASK_INTERRUPTIBLE);
+		else
+			__set_current_state(TASK_UNINTERRUPTIBLE);
+		spin_unlock_irqrestore(f->lock, irq_flags);
+
+		ret = schedule_timeout(ret);
+
+		spin_lock_irqsave(f->lock, irq_flags);
+		if (ret > 0 && intr && signal_pending(current))
+			ret = -ERESTARTSYS;
+	}
+
+	if (!list_empty(&cb.base.node))
+		list_del(&cb.base.node);
+	__set_current_state(TASK_RUNNING);
+
+out:
+	spin_unlock_irqrestore(f->lock, irq_flags);
+
+	vmw_seqno_waiter_remove(dev_priv);
+
+	return ret;
 }
 
+static struct fence_ops vmw_fence_ops = {
+	.get_driver_name = vmw_fence_get_driver_name,
+	.get_timeline_name = vmw_fence_get_timeline_name,
+	.enable_signaling = vmw_fence_enable_signaling,
+	.wait = vmw_fence_wait,
+	.release = vmw_fence_obj_destroy,
+};
+
 
 /**
  * Execute signal actions on fences recently signaled.
@@ -180,12 +305,14 @@ struct vmw_fence_manager *vmw_fence_manager_init(struct vmw_private *dev_priv)
 	INIT_LIST_HEAD(&fman->fence_list);
 	INIT_LIST_HEAD(&fman->cleanup_list);
 	INIT_WORK(&fman->work, &vmw_fence_work_func);
+	INIT_WORK(&fman->ping_work, &vmw_fence_ping_func);
 	fman->fifo_down = true;
 	fman->user_fence_size = ttm_round_pot(sizeof(struct vmw_user_fence));
 	fman->fence_size = ttm_round_pot(sizeof(struct vmw_fence_obj));
 	fman->event_fence_action_size =
 		ttm_round_pot(sizeof(struct vmw_event_fence_action));
 	mutex_init(&fman->goal_irq_mutex);
+	fman->ctx = fence_context_alloc(1);
 
 	return fman;
 }
@@ -196,6 +323,7 @@ void vmw_fence_manager_takedown(struct vmw_fence_manager *fman)
 	bool lists_empty;
 
 	(void) cancel_work_sync(&fman->work);
+	(void) cancel_work_sync(&fman->ping_work);
 
 	spin_lock_irqsave(&fman->lock, irq_flags);
 	lists_empty = list_empty(&fman->fence_list) &&
@@ -207,23 +335,16 @@ void vmw_fence_manager_takedown(struct vmw_fence_manager *fman)
 }
 
 static int vmw_fence_obj_init(struct vmw_fence_manager *fman,
-			      struct vmw_fence_obj *fence,
-			      u32 seqno,
-			      uint32_t mask,
+			      struct vmw_fence_obj *fence, u32 seqno,
 			      void (*destroy) (struct vmw_fence_obj *fence))
 {
 	unsigned long irq_flags;
-	unsigned int num_fences;
 	int ret = 0;
 
-	fence->seqno = seqno;
+	fence_init(&fence->base, &vmw_fence_ops, &fman->lock,
+		   fman->ctx, seqno);
 	INIT_LIST_HEAD(&fence->seq_passed_actions);
-	fence->fman = fman;
-	fence->signaled = 0;
-	fence->signal_mask = mask;
-	kref_init(&fence->kref);
 	fence->destroy = destroy;
-	init_waitqueue_head(&fence->queue);
 
 	spin_lock_irqsave(&fman->lock, irq_flags);
 	if (unlikely(fman->fifo_down)) {
@@ -231,7 +352,7 @@ static int vmw_fence_obj_init(struct vmw_fence_manager *fman,
 		goto out_unlock;
 	}
 	list_add_tail(&fence->head, &fman->fence_list);
-	num_fences = ++fman->num_fence_objects;
+	++fman->num_fence_objects;
 
 out_unlock:
 	spin_unlock_irqrestore(&fman->lock, irq_flags);
@@ -239,38 +360,6 @@ out_unlock:
 
 }
 
-struct vmw_fence_obj *vmw_fence_obj_reference(struct vmw_fence_obj *fence)
-{
-	if (unlikely(fence == NULL))
-		return NULL;
-
-	kref_get(&fence->kref);
-	return fence;
-}
-
-/**
- * vmw_fence_obj_unreference
- *
- * Note that this function may not be entered with disabled irqs since
- * it may re-enable them in the destroy function.
- *
- */
-void vmw_fence_obj_unreference(struct vmw_fence_obj **fence_p)
-{
-	struct vmw_fence_obj *fence = *fence_p;
-	struct vmw_fence_manager *fman;
-
-	if (unlikely(fence == NULL))
-		return;
-
-	fman = fence->fman;
-	*fence_p = NULL;
-	spin_lock_irq(&fman->lock);
-	BUG_ON(atomic_read(&fence->kref.refcount) == 0);
-	kref_put(&fence->kref, vmw_fence_obj_destroy_locked);
-	spin_unlock_irq(&fman->lock);
-}
-
 static void vmw_fences_perform_actions(struct vmw_fence_manager *fman,
 				struct list_head *list)
 {
@@ -326,7 +415,7 @@ static bool vmw_fence_goal_new_locked(struct vmw_fence_manager *fman,
 	list_for_each_entry(fence, &fman->fence_list, head) {
 		if (!list_empty(&fence->seq_passed_actions)) {
 			fman->seqno_valid = true;
-			iowrite32(fence->seqno,
+			iowrite32(fence->base.seqno,
 				  fifo_mem + SVGA_FIFO_FENCE_GOAL);
 			break;
 		}
@@ -353,27 +442,27 @@ static bool vmw_fence_goal_new_locked(struct vmw_fence_manager *fman,
  */
 static bool vmw_fence_goal_check_locked(struct vmw_fence_obj *fence)
 {
+	struct vmw_fence_manager *fman = fman_from_fence(fence);
 	u32 goal_seqno;
 	__le32 __iomem *fifo_mem;
 
-	if (fence->signaled & DRM_VMW_FENCE_FLAG_EXEC)
+	if (fence_is_signaled_locked(&fence->base))
 		return false;
 
-	fifo_mem = fence->fman->dev_priv->mmio_virt;
+	fifo_mem = fman->dev_priv->mmio_virt;
 	goal_seqno = ioread32(fifo_mem + SVGA_FIFO_FENCE_GOAL);
-	if (likely(fence->fman->seqno_valid &&
-		   goal_seqno - fence->seqno < VMW_FENCE_WRAP))
+	if (likely(fman->seqno_valid &&
+		   goal_seqno - fence->base.seqno < VMW_FENCE_WRAP))
 		return false;
 
-	iowrite32(fence->seqno, fifo_mem + SVGA_FIFO_FENCE_GOAL);
-	fence->fman->seqno_valid = true;
+	iowrite32(fence->base.seqno, fifo_mem + SVGA_FIFO_FENCE_GOAL);
+	fman->seqno_valid = true;
 
 	return true;
 }
 
-void vmw_fences_update(struct vmw_fence_manager *fman)
+static void __vmw_fences_update(struct vmw_fence_manager *fman)
 {
-	unsigned long flags;
 	struct vmw_fence_obj *fence, *next_fence;
 	struct list_head action_list;
 	bool needs_rerun;
@@ -382,32 +471,25 @@ void vmw_fences_update(struct vmw_fence_manager *fman)
 
 	seqno = ioread32(fifo_mem + SVGA_FIFO_FENCE);
 rerun:
-	spin_lock_irqsave(&fman->lock, flags);
 	list_for_each_entry_safe(fence, next_fence, &fman->fence_list, head) {
-		if (seqno - fence->seqno < VMW_FENCE_WRAP) {
+		if (seqno - fence->base.seqno < VMW_FENCE_WRAP) {
 			list_del_init(&fence->head);
-			fence->signaled |= DRM_VMW_FENCE_FLAG_EXEC;
+			fence_signal_locked(&fence->base);
 			INIT_LIST_HEAD(&action_list);
 			list_splice_init(&fence->seq_passed_actions,
 					 &action_list);
 			vmw_fences_perform_actions(fman, &action_list);
-			wake_up_all(&fence->queue);
 		} else
 			break;
 	}
 
-	needs_rerun = vmw_fence_goal_new_locked(fman, seqno);
-
-	if (!list_empty(&fman->cleanup_list))
-		(void) schedule_work(&fman->work);
-	spin_unlock_irqrestore(&fman->lock, flags);
-
 	/*
 	 * Rerun if the fence goal seqno was updated, and the
 	 * hardware might have raced with that update, so that
 	 * we missed a fence_goal irq.
 	 */
 
+	needs_rerun = vmw_fence_goal_new_locked(fman, seqno);
 	if (unlikely(needs_rerun)) {
 		new_seqno = ioread32(fifo_mem + SVGA_FIFO_FENCE);
 		if (new_seqno != seqno) {
@@ -415,107 +497,69 @@ rerun:
 			goto rerun;
 		}
 	}
+
+	if (!list_empty(&fman->cleanup_list))
+		(void) schedule_work(&fman->work);
 }
 
-bool vmw_fence_obj_signaled(struct vmw_fence_obj *fence,
-			    uint32_t flags)
+void vmw_fences_update(struct vmw_fence_manager *fman)
 {
-	struct vmw_fence_manager *fman = fence->fman;
 	unsigned long irq_flags;
-	uint32_t signaled;
 
 	spin_lock_irqsave(&fman->lock, irq_flags);
-	signaled = fence->signaled;
+	__vmw_fences_update(fman);
 	spin_unlock_irqrestore(&fman->lock, irq_flags);
+}
 
-	flags &= fence->signal_mask;
-	if ((signaled & flags) == flags)
-		return 1;
+bool vmw_fence_obj_signaled(struct vmw_fence_obj *fence)
+{
+	struct vmw_fence_manager *fman = fman_from_fence(fence);
 
-	if ((signaled & DRM_VMW_FENCE_FLAG_EXEC) == 0)
-		vmw_fences_update(fman);
+	if (test_bit(FENCE_FLAG_SIGNALED_BIT, &fence->base.flags))
+		return 1;
 
-	spin_lock_irqsave(&fman->lock, irq_flags);
-	signaled = fence->signaled;
-	spin_unlock_irqrestore(&fman->lock, irq_flags);
+	vmw_fences_update(fman);
 
-	return ((signaled & flags) == flags);
+	return fence_is_signaled(&fence->base);
 }
 
-int vmw_fence_obj_wait(struct vmw_fence_obj *fence,
-		       uint32_t flags, bool lazy,
+int vmw_fence_obj_wait(struct vmw_fence_obj *fence, bool lazy,
 		       bool interruptible, unsigned long timeout)
 {
-	struct vmw_private *dev_priv = fence->fman->dev_priv;
-	long ret;
+	long ret = fence_wait_timeout(&fence->base, interruptible, timeout);
 
-	if (likely(vmw_fence_obj_signaled(fence, flags)))
+	if (likely(ret > 0))
 		return 0;
-
-	vmw_fifo_ping_host(dev_priv, SVGA_SYNC_GENERIC);
-	vmw_seqno_waiter_add(dev_priv);
-
-	if (interruptible)
-		ret = wait_event_interruptible_timeout
-			(fence->queue,
-			 vmw_fence_obj_signaled(fence, flags),
-			 timeout);
+	else if (ret == 0)
+		return -EBUSY;
 	else
-		ret = wait_event_timeout
-			(fence->queue,
-			 vmw_fence_obj_signaled(fence, flags),
-			 timeout);
-
-	vmw_seqno_waiter_remove(dev_priv);
-
-	if (unlikely(ret == 0))
-		ret = -EBUSY;
-	else if (likely(ret > 0))
-		ret = 0;
-
-	return ret;
+		return ret;
 }
 
 void vmw_fence_obj_flush(struct vmw_fence_obj *fence)
 {
-	struct vmw_private *dev_priv = fence->fman->dev_priv;
+	struct vmw_private *dev_priv = fman_from_fence(fence)->dev_priv;
 
 	vmw_fifo_ping_host(dev_priv, SVGA_SYNC_GENERIC);
 }
 
 static void vmw_fence_destroy(struct vmw_fence_obj *fence)
 {
-	struct vmw_fence_manager *fman = fence->fman;
-
-	kfree(fence);
-	/*
-	 * Free kernel space accounting.
-	 */
-	ttm_mem_global_free(vmw_mem_glob(fman->dev_priv),
-			    fman->fence_size);
+	fence_free(&fence->base);
 }
 
 int vmw_fence_create(struct vmw_fence_manager *fman,
 		     uint32_t seqno,
-		     uint32_t mask,
 		     struct vmw_fence_obj **p_fence)
 {
-	struct ttm_mem_global *mem_glob = vmw_mem_glob(fman->dev_priv);
 	struct vmw_fence_obj *fence;
 	int ret;
 
-	ret = ttm_mem_global_alloc(mem_glob, fman->fence_size,
-				   false, false);
-	if (unlikely(ret != 0))
-		return ret;
-
 	fence = kzalloc(sizeof(*fence), GFP_KERNEL);
-	if (unlikely(fence == NULL)) {
-		ret = -ENOMEM;
-		goto out_no_object;
-	}
+	if (unlikely(fence == NULL))
+		return -ENOMEM;
 
-	ret = vmw_fence_obj_init(fman, fence, seqno, mask,
+	ret = vmw_fence_obj_init(fman, fence, seqno,
 				 vmw_fence_destroy);
 	if (unlikely(ret != 0))
 		goto out_err_init;
@@ -525,8 +569,6 @@ int vmw_fence_create(struct vmw_fence_manager *fman,
 
 out_err_init:
 	kfree(fence);
-out_no_object:
-	ttm_mem_global_free(mem_glob, fman->fence_size);
 	return ret;
 }
 
@@ -535,7 +577,7 @@ static void vmw_user_fence_destroy(struct vmw_fence_obj *fence)
 {
 	struct vmw_user_fence *ufence =
 		container_of(fence, struct vmw_user_fence, fence);
-	struct vmw_fence_manager *fman = fence->fman;
+	struct vmw_fence_manager *fman = fman_from_fence(fence);
 
 	ttm_base_object_kfree(ufence, base);
 	/*
@@ -559,7 +601,6 @@ static void vmw_user_fence_base_release(struct ttm_base_object **p_base)
 int vmw_user_fence_create(struct drm_file *file_priv,
 			  struct vmw_fence_manager *fman,
 			  uint32_t seqno,
-			  uint32_t mask,
 			  struct vmw_fence_obj **p_fence,
 			  uint32_t *p_handle)
 {
@@ -586,7 +627,7 @@ int vmw_user_fence_create(struct drm_file *file_priv,
 	}
 
 	ret = vmw_fence_obj_init(fman, &ufence->fence, seqno,
-				 mask, vmw_user_fence_destroy);
+				 vmw_user_fence_destroy);
 	if (unlikely(ret != 0)) {
 		kfree(ufence);
 		goto out_no_object;
@@ -629,7 +670,6 @@ out_no_object:
 
 void vmw_fence_fifo_down(struct vmw_fence_manager *fman)
 {
-	unsigned long irq_flags;
 	struct list_head action_list;
 	int ret;
 
@@ -638,35 +678,32 @@ void vmw_fence_fifo_down(struct vmw_fence_manager *fman)
 	 * restart when we've released the fman->lock.
 	 */
 
-	spin_lock_irqsave(&fman->lock, irq_flags);
+	spin_lock_irq(&fman->lock);
 	fman->fifo_down = true;
 	while (!list_empty(&fman->fence_list)) {
 		struct vmw_fence_obj *fence =
 			list_entry(fman->fence_list.prev, struct vmw_fence_obj,
 				   head);
-		kref_get(&fence->kref);
+		fence_get(&fence->base);
 		spin_unlock_irq(&fman->lock);
 
-		ret = vmw_fence_obj_wait(fence, fence->signal_mask,
-					 false, false,
+		ret = vmw_fence_obj_wait(fence, false, false,
 					 VMW_FENCE_WAIT_TIMEOUT);
 
 		if (unlikely(ret != 0)) {
 			list_del_init(&fence->head);
-			fence->signaled |= DRM_VMW_FENCE_FLAG_EXEC;
+			fence_signal(&fence->base);
 			INIT_LIST_HEAD(&action_list);
 			list_splice_init(&fence->seq_passed_actions,
 					 &action_list);
 			vmw_fences_perform_actions(fman, &action_list);
-			wake_up_all(&fence->queue);
 		}
 
-		spin_lock_irq(&fman->lock);
-
 		BUG_ON(!list_empty(&fence->head));
-		kref_put(&fence->kref, vmw_fence_obj_destroy_locked);
+		fence_put(&fence->base);
+		spin_lock_irq(&fman->lock);
 	}
-	spin_unlock_irqrestore(&fman->lock, irq_flags);
+	spin_unlock_irq(&fman->lock);
 }
 
 void vmw_fence_fifo_up(struct vmw_fence_manager *fman)
@@ -716,14 +753,14 @@ int vmw_fence_obj_wait_ioctl(struct drm_device *dev, void *data,
 
 	timeout = jiffies;
 	if (time_after_eq(timeout, (unsigned long)arg->kernel_cookie)) {
-		ret = ((vmw_fence_obj_signaled(fence, arg->flags)) ?
+		ret = ((vmw_fence_obj_signaled(fence)) ?
 		       0 : -EBUSY);
 		goto out;
 	}
 
 	timeout = (unsigned long)arg->kernel_cookie - timeout;
 
-	ret = vmw_fence_obj_wait(fence, arg->flags, arg->lazy, true, timeout);
+	ret = vmw_fence_obj_wait(fence, arg->lazy, true, timeout);
 
 out:
 	ttm_base_object_unref(&base);
@@ -758,12 +795,12 @@ int vmw_fence_obj_signaled_ioctl(struct drm_device *dev, void *data,
 	}
 
 	fence = &(container_of(base, struct vmw_user_fence, base)->fence);
-	fman = fence->fman;
+	fman = fman_from_fence(fence);
 
-	arg->signaled = vmw_fence_obj_signaled(fence, arg->flags);
-	spin_lock_irq(&fman->lock);
+	arg->signaled = vmw_fence_obj_signaled(fence);
 
-	arg->signaled_flags = fence->signaled;
+	arg->signaled_flags = arg->flags;
+	spin_lock_irq(&fman->lock);
 	arg->passed_seqno = dev_priv->last_read_seqno;
 	spin_unlock_irq(&fman->lock);
 
@@ -876,7 +913,7 @@ static void vmw_event_fence_action_cleanup(struct vmw_fence_action *action)
 {
 	struct vmw_event_fence_action *eaction =
 		container_of(action, struct vmw_event_fence_action, action);
-	struct vmw_fence_manager *fman = eaction->fence->fman;
+	struct vmw_fence_manager *fman = fman_from_fence(eaction->fence);
 	unsigned long irq_flags;
 
 	spin_lock_irqsave(&fman->lock, irq_flags);
@@ -900,7 +937,7 @@ static void vmw_event_fence_action_cleanup(struct vmw_fence_action *action)
 static void vmw_fence_obj_add_action(struct vmw_fence_obj *fence,
 			      struct vmw_fence_action *action)
 {
-	struct vmw_fence_manager *fman = fence->fman;
+	struct vmw_fence_manager *fman = fman_from_fence(fence);
 	unsigned long irq_flags;
 	bool run_update = false;
 
@@ -908,7 +945,7 @@ static void vmw_fence_obj_add_action(struct vmw_fence_obj *fence,
 	spin_lock_irqsave(&fman->lock, irq_flags);
 
 	fman->pending_actions[action->type]++;
-	if (fence->signaled & DRM_VMW_FENCE_FLAG_EXEC) {
+	if (fence_is_signaled_locked(&fence->base)) {
 		struct list_head action_list;
 
 		INIT_LIST_HEAD(&action_list);
@@ -960,7 +997,7 @@ int vmw_event_fence_action_queue(struct drm_file *file_priv,
 				 bool interruptible)
 {
 	struct vmw_event_fence_action *eaction;
-	struct vmw_fence_manager *fman = fence->fman;
+	struct vmw_fence_manager *fman = fman_from_fence(fence);
 	struct vmw_fpriv *vmw_fp = vmw_fpriv(file_priv);
 	unsigned long irq_flags;
 
@@ -1000,7 +1037,8 @@ static int vmw_event_fence_action_create(struct drm_file *file_priv,
 				  bool interruptible)
 {
 	struct vmw_event_fence_pending *event;
-	struct drm_device *dev = fence->fman->dev_priv->dev;
+	struct vmw_fence_manager *fman = fman_from_fence(fence);
+	struct drm_device *dev = fman->dev_priv->dev;
 	unsigned long irq_flags;
 	int ret;
 
@@ -1049,6 +1087,8 @@ static int vmw_event_fence_action_create(struct drm_file *file_priv,
 	if (ret != 0)
 		goto out_no_queue;
 
+	return 0;
+
 out_no_queue:
 	event->base.destroy(&event->base);
 out_no_event:
@@ -1124,17 +1164,10 @@ int vmw_fence_event_ioctl(struct drm_device *dev, void *data,
 
 	BUG_ON(fence == NULL);
 
-	if (arg->flags & DRM_VMW_FE_FLAG_REQ_TIME)
-		ret = vmw_event_fence_action_create(file_priv, fence,
-						    arg->flags,
-						    arg->user_data,
-						    true);
-	else
-		ret = vmw_event_fence_action_create(file_priv, fence,
-						    arg->flags,
-						    arg->user_data,
-						    true);
-
+	ret = vmw_event_fence_action_create(file_priv, fence,
+					    arg->flags,
+					    arg->user_data,
+					    true);
 	if (unlikely(ret != 0)) {
 		if (ret != -ERESTARTSYS)
 			DRM_ERROR("Failed to attach event to fence.\n");
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_fence.h b/drivers/gpu/drm/vmwgfx/vmwgfx_fence.h
index faf2e7873860..26a4add39208 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_fence.h
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_fence.h
@@ -27,6 +27,8 @@
 
 #ifndef _VMWGFX_FENCE_H_
 
+#include <linux/fence.h>
+
 #define VMW_FENCE_WAIT_TIMEOUT (5*HZ)
 
 struct vmw_private;
@@ -50,16 +52,11 @@ struct vmw_fence_action {
 };
 
 struct vmw_fence_obj {
-	struct kref kref;
-	u32 seqno;
+	struct fence base;
 
-	struct vmw_fence_manager *fman;
 	struct list_head head;
-	uint32_t signaled;
-	uint32_t signal_mask;
 	struct list_head seq_passed_actions;
 	void (*destroy)(struct vmw_fence_obj *fence);
-	wait_queue_head_t queue;
 };
 
 extern struct vmw_fence_manager *
@@ -67,17 +64,29 @@ vmw_fence_manager_init(struct vmw_private *dev_priv);
 
 extern void vmw_fence_manager_takedown(struct vmw_fence_manager *fman);
 
-extern void vmw_fence_obj_unreference(struct vmw_fence_obj **fence_p);
+static inline void
+vmw_fence_obj_unreference(struct vmw_fence_obj **fence_p)
+{
+	struct vmw_fence_obj *fence = *fence_p;
+
+	*fence_p = NULL;
+	if (fence)
+		fence_put(&fence->base);
+}
 
-extern struct vmw_fence_obj *
-vmw_fence_obj_reference(struct vmw_fence_obj *fence);
+static inline struct vmw_fence_obj *
+vmw_fence_obj_reference(struct vmw_fence_obj *fence)
+{
+	if (fence)
+		fence_get(&fence->base);
+	return fence;
+}
 
 extern void vmw_fences_update(struct vmw_fence_manager *fman);
 
-extern bool vmw_fence_obj_signaled(struct vmw_fence_obj *fence,
-				   uint32_t flags);
+extern bool vmw_fence_obj_signaled(struct vmw_fence_obj *fence);
 
-extern int vmw_fence_obj_wait(struct vmw_fence_obj *fence, uint32_t flags,
+extern int vmw_fence_obj_wait(struct vmw_fence_obj *fence,
 			      bool lazy,
 			      bool interruptible, unsigned long timeout);
 
@@ -85,13 +94,11 @@ extern void vmw_fence_obj_flush(struct vmw_fence_obj *fence);
 
 extern int vmw_fence_create(struct vmw_fence_manager *fman,
 			    uint32_t seqno,
-			    uint32_t mask,
 			    struct vmw_fence_obj **p_fence);
 
 extern int vmw_user_fence_create(struct drm_file *file_priv,
 				 struct vmw_fence_manager *fman,
 				 uint32_t sequence,
-				 uint32_t mask,
 				 struct vmw_fence_obj **p_fence,
 				 uint32_t *p_handle);
 
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_fifo.c b/drivers/gpu/drm/vmwgfx/vmwgfx_fifo.c
index 6eae14d2a3f7..09e10aefcd8e 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_fifo.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_fifo.c
@@ -160,16 +160,21 @@ int vmw_fifo_init(struct vmw_private *dev_priv, struct vmw_fifo_state *fifo)
 	return vmw_fifo_send_fence(dev_priv, &dummy);
 }
 
-void vmw_fifo_ping_host(struct vmw_private *dev_priv, uint32_t reason)
+void vmw_fifo_ping_host_locked(struct vmw_private *dev_priv, uint32_t reason)
 {
 	__le32 __iomem *fifo_mem = dev_priv->mmio_virt;
 
-	mutex_lock(&dev_priv->hw_mutex);
-
 	if (unlikely(ioread32(fifo_mem + SVGA_FIFO_BUSY) == 0)) {
 		iowrite32(1, fifo_mem + SVGA_FIFO_BUSY);
 		vmw_write(dev_priv, SVGA_REG_SYNC, reason);
 	}
+}
+
+void vmw_fifo_ping_host(struct vmw_private *dev_priv, uint32_t reason)
+{
+	mutex_lock(&dev_priv->hw_mutex);
+
+	vmw_fifo_ping_host_locked(dev_priv, reason);
 
 	mutex_unlock(&dev_priv->hw_mutex);
 }
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_gmrid_manager.c b/drivers/gpu/drm/vmwgfx/vmwgfx_gmrid_manager.c
index 26f8bdde3529..170b61be1e4e 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_gmrid_manager.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_gmrid_manager.c
@@ -46,8 +46,7 @@ struct vmwgfx_gmrid_man {
 
 static int vmw_gmrid_man_get_node(struct ttm_mem_type_manager *man,
 				  struct ttm_buffer_object *bo,
-				  struct ttm_placement *placement,
-				  uint32_t flags,
+				  const struct ttm_place *place,
 				  struct ttm_mem_reg *mem)
 {
 	struct vmwgfx_gmrid_man *gman =
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_kms.c b/drivers/gpu/drm/vmwgfx/vmwgfx_kms.c
index d2bc2b03d4c6..3725b521d931 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_kms.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_kms.c
@@ -187,7 +187,7 @@ int vmw_du_crtc_cursor_set(struct drm_crtc *crtc, struct drm_file *file_priv,
 	 * can do this since the caller in the drm core doesn't check anything
 	 * which is protected by any looks.
 	 */
-	drm_modeset_unlock(&crtc->mutex);
+	drm_modeset_unlock_crtc(crtc);
 	drm_modeset_lock_all(dev_priv->dev);
 
 	/* A lot of the code assumes this */
@@ -252,7 +252,7 @@ int vmw_du_crtc_cursor_set(struct drm_crtc *crtc, struct drm_file *file_priv,
 	ret = 0;
 out:
 	drm_modeset_unlock_all(dev_priv->dev);
-	drm_modeset_lock(&crtc->mutex, NULL);
+	drm_modeset_lock_crtc(crtc, crtc->cursor);
 
 	return ret;
 }
@@ -273,7 +273,7 @@ int vmw_du_crtc_cursor_move(struct drm_crtc *crtc, int x, int y)
 	 * can do this since the caller in the drm core doesn't check anything
 	 * which is protected by any looks.
 	 */
-	drm_modeset_unlock(&crtc->mutex);
+	drm_modeset_unlock_crtc(crtc);
 	drm_modeset_lock_all(dev_priv->dev);
 
 	vmw_cursor_update_position(dev_priv, shown,
@@ -281,7 +281,7 @@ int vmw_du_crtc_cursor_move(struct drm_crtc *crtc, int x, int y)
 				   du->cursor_y + du->hotspot_y);
 
 	drm_modeset_unlock_all(dev_priv->dev);
-	drm_modeset_lock(&crtc->mutex, NULL);
+	drm_modeset_lock_crtc(crtc, crtc->cursor);
 
 	return 0;
 }
@@ -1950,6 +1950,14 @@ int vmw_du_connector_fill_modes(struct drm_connector *connector,
 		DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_PVSYNC)
 	};
 	int i;
+	u32 assumed_bpp = 2;
+
+	/*
+	 * If using screen objects, then assume 32-bpp because that's what the
+	 * SVGA device is assuming
+	 */
+	if (dev_priv->sou_priv)
+		assumed_bpp = 4;
 
 	/* Add preferred mode */
 	{
@@ -1960,8 +1968,9 @@ int vmw_du_connector_fill_modes(struct drm_connector *connector,
 		mode->vdisplay = du->pref_height;
 		vmw_guess_mode_timing(mode);
 
-		if (vmw_kms_validate_mode_vram(dev_priv, mode->hdisplay * 2,
-					       mode->vdisplay)) {
+		if (vmw_kms_validate_mode_vram(dev_priv,
+						mode->hdisplay * assumed_bpp,
+						mode->vdisplay)) {
 			drm_mode_probed_add(connector, mode);
 		} else {
 			drm_mode_destroy(dev, mode);
@@ -1983,7 +1992,8 @@ int vmw_du_connector_fill_modes(struct drm_connector *connector,
 		    bmode->vdisplay > max_height)
 			continue;
 
-		if (!vmw_kms_validate_mode_vram(dev_priv, bmode->hdisplay * 2,
+		if (!vmw_kms_validate_mode_vram(dev_priv,
+						bmode->hdisplay * assumed_bpp,
 						bmode->vdisplay))
 			continue;
 
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_ldu.c b/drivers/gpu/drm/vmwgfx/vmwgfx_ldu.c
index 15e185ae4c99..5c289f748ab4 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_ldu.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_ldu.c
@@ -26,6 +26,7 @@
  **************************************************************************/
 
 #include "vmwgfx_kms.h"
+#include <drm/drm_plane_helper.h>
 
 
 #define vmw_crtc_to_ldu(x) \
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_resource.c b/drivers/gpu/drm/vmwgfx/vmwgfx_resource.c
index a432c0db257c..210ef15b1d09 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_resource.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_resource.c
@@ -133,6 +133,7 @@ static void vmw_resource_release(struct kref *kref)
 			struct ttm_validate_buffer val_buf;
 
 			val_buf.bo = bo;
+			val_buf.shared = false;
 			res->func->unbind(res, false, &val_buf);
 		}
 		res->backup_dirty = false;
@@ -429,7 +430,7 @@ int vmw_dmabuf_init(struct vmw_private *dev_priv,
 	ret = ttm_bo_init(bdev, &vmw_bo->base, size,
 			  ttm_bo_type_device, placement,
 			  0, interruptible,
-			  NULL, acc_size, NULL, bo_free);
+			  NULL, acc_size, NULL, NULL, bo_free);
 	return ret;
 }
 
@@ -567,13 +568,18 @@ static int vmw_user_dmabuf_synccpu_grab(struct vmw_user_dma_buffer *user_bo,
 	int ret;
 
 	if (flags & drm_vmw_synccpu_allow_cs) {
-		struct ttm_bo_device *bdev = bo->bdev;
+		bool nonblock = !!(flags & drm_vmw_synccpu_dontblock);
+		long lret;
 
-		spin_lock(&bdev->fence_lock);
-		ret = ttm_bo_wait(bo, false, true,
-				  !!(flags & drm_vmw_synccpu_dontblock));
-		spin_unlock(&bdev->fence_lock);
-		return ret;
+		if (nonblock)
+			return reservation_object_test_signaled_rcu(bo->resv, true) ? 0 : -EBUSY;
+
+		lret = reservation_object_wait_timeout_rcu(bo->resv, true, true, MAX_SCHEDULE_TIMEOUT);
+		if (!lret)
+			return -EBUSY;
+		else if (lret < 0)
+			return lret;
+		return 0;
 	}
 
 	ret = ttm_bo_synccpu_write_grab
@@ -1214,8 +1220,9 @@ vmw_resource_check_buffer(struct vmw_resource *res,
 
 	INIT_LIST_HEAD(&val_list);
 	val_buf->bo = ttm_bo_reference(&res->backup->base);
+	val_buf->shared = false;
 	list_add_tail(&val_buf->head, &val_list);
-	ret = ttm_eu_reserve_buffers(NULL, &val_list);
+	ret = ttm_eu_reserve_buffers(NULL, &val_list, interruptible, NULL);
 	if (unlikely(ret != 0))
 		goto out_no_reserve;
 
@@ -1307,6 +1314,7 @@ int vmw_resource_do_evict(struct vmw_resource *res, bool interruptible)
 	BUG_ON(!func->may_evict);
 
 	val_buf.bo = NULL;
+	val_buf.shared = false;
 	ret = vmw_resource_check_buffer(res, interruptible, &val_buf);
 	if (unlikely(ret != 0))
 		return ret;
@@ -1352,6 +1360,7 @@ int vmw_resource_validate(struct vmw_resource *res)
 		return 0;
 
 	val_buf.bo = NULL;
+	val_buf.shared = false;
 	if (res->backup)
 		val_buf.bo = &res->backup->base;
 	do {
@@ -1419,25 +1428,16 @@ void vmw_fence_single_bo(struct ttm_buffer_object *bo,
 			 struct vmw_fence_obj *fence)
 {
 	struct ttm_bo_device *bdev = bo->bdev;
-	struct ttm_bo_driver *driver = bdev->driver;
-	struct vmw_fence_obj *old_fence_obj;
+
 	struct vmw_private *dev_priv =
 		container_of(bdev, struct vmw_private, bdev);
 
-	if (fence == NULL)
+	if (fence == NULL) {
 		vmw_execbuf_fence_commands(NULL, dev_priv, &fence, NULL);
-	else
-		driver->sync_obj_ref(fence);
-
-	spin_lock(&bdev->fence_lock);
-
-	old_fence_obj = bo->sync_obj;
-	bo->sync_obj = fence;
-
-	spin_unlock(&bdev->fence_lock);
-
-	if (old_fence_obj)
-		vmw_fence_obj_unreference(&old_fence_obj);
+		reservation_object_add_excl_fence(bo->resv, &fence->base);
+		fence_put(&fence->base);
+	} else
+		reservation_object_add_excl_fence(bo->resv, &fence->base);
 }
 
 /**
@@ -1475,10 +1475,10 @@ void vmw_resource_move_notify(struct ttm_buffer_object *bo,
 
 	if (mem->mem_type != VMW_PL_MOB) {
 		struct vmw_resource *res, *n;
-		struct ttm_bo_device *bdev = bo->bdev;
 		struct ttm_validate_buffer val_buf;
 
 		val_buf.bo = bo;
+		val_buf.shared = false;
 
 		list_for_each_entry_safe(res, n, &dma_buf->res_list, mob_head) {
 
@@ -1491,9 +1491,7 @@ void vmw_resource_move_notify(struct ttm_buffer_object *bo,
 			list_del_init(&res->mob_head);
 		}
 
-		spin_lock(&bdev->fence_lock);
 		(void) ttm_bo_wait(bo, false, false, false);
-		spin_unlock(&bdev->fence_lock);
 	}
 }
 
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_scrn.c b/drivers/gpu/drm/vmwgfx/vmwgfx_scrn.c
index b295463a60b3..7dc591d04d9a 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_scrn.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_scrn.c
@@ -26,6 +26,7 @@
  **************************************************************************/
 
 #include "vmwgfx_kms.h"
+#include <drm/drm_plane_helper.h>
 
 
 #define vmw_crtc_to_sou(x) \
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_shader.c b/drivers/gpu/drm/vmwgfx/vmwgfx_shader.c
index 8719fb3cccc9..6a4584a43aa6 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_shader.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_shader.c
@@ -198,7 +198,7 @@ static int vmw_gb_shader_bind(struct vmw_resource *res,
 	cmd->header.size = sizeof(cmd->body);
 	cmd->body.shid = res->id;
 	cmd->body.mobid = bo->mem.start;
-	cmd->body.offsetInBytes = 0;
+	cmd->body.offsetInBytes = res->backup_offset;
 	res->backup_dirty = false;
 	vmw_fifo_commit(dev_priv, sizeof(*cmd));
 
diff --git a/drivers/gpu/host1x/cdma.c b/drivers/gpu/host1x/cdma.c
index 3995255b16c7..5a8c8d55317a 100644
--- a/drivers/gpu/host1x/cdma.c
+++ b/drivers/gpu/host1x/cdma.c
@@ -97,7 +97,7 @@ fail:
 static void host1x_pushbuffer_push(struct push_buffer *pb, u32 op1, u32 op2)
 {
 	u32 pos = pb->pos;
-	u32 *p = (u32 *)((u32)pb->mapped + pos);
+	u32 *p = (u32 *)((void *)pb->mapped + pos);
 	WARN_ON(pos == pb->fence);
 	*(p++) = op1;
 	*(p++) = op2;
diff --git a/drivers/gpu/host1x/cdma.h b/drivers/gpu/host1x/cdma.h
index 313c4b784348..470087af8fe5 100644
--- a/drivers/gpu/host1x/cdma.h
+++ b/drivers/gpu/host1x/cdma.h
@@ -42,7 +42,7 @@ struct host1x_job;
  */
 
 struct push_buffer {
-	u32 *mapped;			/* mapped pushbuffer memory */
+	void *mapped;			/* mapped pushbuffer memory */
 	dma_addr_t phys;		/* physical address of pushbuffer */
 	u32 fence;			/* index we've written */
 	u32 pos;			/* index to write to */
diff --git a/drivers/gpu/host1x/hw/cdma_hw.c b/drivers/gpu/host1x/hw/cdma_hw.c
index 6b09b71940c2..305ea8f3382d 100644
--- a/drivers/gpu/host1x/hw/cdma_hw.c
+++ b/drivers/gpu/host1x/hw/cdma_hw.c
@@ -26,11 +26,11 @@
 #include "../debug.h"
 
 /*
- * Put the restart at the end of pushbuffer memor
+ * Put the restart at the end of pushbuffer memory
  */
 static void push_buffer_init(struct push_buffer *pb)
 {
-	*(pb->mapped + (pb->size_bytes >> 2)) = host1x_opcode_restart(0);
+	*(u32 *)(pb->mapped + pb->size_bytes) = host1x_opcode_restart(0);
 }
 
 /*
@@ -51,11 +51,11 @@ static void cdma_timeout_cpu_incr(struct host1x_cdma *cdma, u32 getptr,
 
 	/* NOP all the PB slots */
 	while (nr_slots--) {
-		u32 *p = (u32 *)((u32)pb->mapped + getptr);
+		u32 *p = (u32 *)(pb->mapped + getptr);
 		*(p++) = HOST1X_OPCODE_NOP;
 		*(p++) = HOST1X_OPCODE_NOP;
-		dev_dbg(host1x->dev, "%s: NOP at %#llx\n", __func__,
-			(u64)pb->phys + getptr);
+		dev_dbg(host1x->dev, "%s: NOP at %pad+%#x\n", __func__,
+			&pb->phys, getptr);
 		getptr = (getptr + 8) & (pb->size_bytes - 1);
 	}
 	wmb();
diff --git a/drivers/gpu/host1x/hw/channel_hw.c b/drivers/gpu/host1x/hw/channel_hw.c
index 4608257ab656..946c332c3906 100644
--- a/drivers/gpu/host1x/hw/channel_hw.c
+++ b/drivers/gpu/host1x/hw/channel_hw.c
@@ -32,6 +32,7 @@
 static void trace_write_gather(struct host1x_cdma *cdma, struct host1x_bo *bo,
 			       u32 offset, u32 words)
 {
+	struct device *dev = cdma_to_channel(cdma)->dev;
 	void *mem = NULL;
 
 	if (host1x_debug_trace_cmdbuf)
@@ -44,11 +45,14 @@ static void trace_write_gather(struct host1x_cdma *cdma, struct host1x_bo *bo,
 		 * of how much you can output to ftrace at once.
 		 */
 		for (i = 0; i < words; i += TRACE_MAX_LENGTH) {
-			trace_host1x_cdma_push_gather(
-				dev_name(cdma_to_channel(cdma)->dev),
-				(u32)bo, min(words - i, TRACE_MAX_LENGTH),
-				offset + i * sizeof(u32), mem);
+			u32 num_words = min(words - i, TRACE_MAX_LENGTH);
+			offset += i * sizeof(u32);
+
+			trace_host1x_cdma_push_gather(dev_name(dev), bo,
+						      num_words, offset,
+						      mem);
 		}
+
 		host1x_bo_munmap(bo, mem);
 	}
 }
diff --git a/drivers/gpu/host1x/hw/debug_hw.c b/drivers/gpu/host1x/hw/debug_hw.c
index f72c873eff81..791de9351eeb 100644
--- a/drivers/gpu/host1x/hw/debug_hw.c
+++ b/drivers/gpu/host1x/hw/debug_hw.c
@@ -163,8 +163,8 @@ static void show_channel_gathers(struct output *o, struct host1x_cdma *cdma)
 				continue;
 			}
 
-			host1x_debug_output(o, "    GATHER at %#llx+%04x, %d words\n",
-					    (u64)g->base, g->offset, g->words);
+			host1x_debug_output(o, "    GATHER at %pad+%#x, %d words\n",
+					    &g->base, g->offset, g->words);
 
 			show_gather(o, g->base + g->offset, g->words, cdma,
 				    g->base, mapped);
diff --git a/drivers/gpu/host1x/job.h b/drivers/gpu/host1x/job.h
index 33a697d6dcef..8b3c15df0660 100644
--- a/drivers/gpu/host1x/job.h
+++ b/drivers/gpu/host1x/job.h
@@ -23,7 +23,7 @@ struct host1x_job_gather {
 	u32 words;
 	dma_addr_t base;
 	struct host1x_bo *bo;
-	int offset;
+	u32 offset;
 	bool handled;
 };
 
diff --git a/drivers/gpu/host1x/mipi.c b/drivers/gpu/host1x/mipi.c
index 9882ea122024..fbc6ee6ca337 100644
--- a/drivers/gpu/host1x/mipi.c
+++ b/drivers/gpu/host1x/mipi.c
@@ -49,35 +49,47 @@
 #define MIPI_CAL_CONFIG_DSIC		0x10
 #define MIPI_CAL_CONFIG_DSID		0x11
 
+#define MIPI_CAL_CONFIG_DSIAB_CLK	0x19
+#define MIPI_CAL_CONFIG_DSICD_CLK	0x1a
+#define MIPI_CAL_CONFIG_CSIAB_CLK	0x1b
+#define MIPI_CAL_CONFIG_CSICD_CLK	0x1c
+#define MIPI_CAL_CONFIG_CSIE_CLK	0x1d
+
+/* for data and clock lanes */
 #define MIPI_CAL_CONFIG_SELECT		(1 << 21)
+
+/* for data lanes */
 #define MIPI_CAL_CONFIG_HSPDOS(x)	(((x) & 0x1f) << 16)
 #define MIPI_CAL_CONFIG_HSPUOS(x)	(((x) & 0x1f) <<  8)
 #define MIPI_CAL_CONFIG_TERMOS(x)	(((x) & 0x1f) <<  0)
 
+/* for clock lanes */
+#define MIPI_CAL_CONFIG_HSCLKPDOSD(x)	(((x) & 0x1f) <<  8)
+#define MIPI_CAL_CONFIG_HSCLKPUOSD(x)	(((x) & 0x1f) <<  0)
+
 #define MIPI_CAL_BIAS_PAD_CFG0		0x16
 #define MIPI_CAL_BIAS_PAD_PDVCLAMP	(1 << 1)
 #define MIPI_CAL_BIAS_PAD_E_VCLAMP_REF	(1 << 0)
 
 #define MIPI_CAL_BIAS_PAD_CFG1		0x17
+#define MIPI_CAL_BIAS_PAD_DRV_DN_REF(x) (((x) & 0x7) << 16)
 
 #define MIPI_CAL_BIAS_PAD_CFG2		0x18
 #define MIPI_CAL_BIAS_PAD_PDVREG	(1 << 1)
 
-static const struct module {
-	unsigned long reg;
-} modules[] = {
-	{ .reg = MIPI_CAL_CONFIG_CSIA },
-	{ .reg = MIPI_CAL_CONFIG_CSIB },
-	{ .reg = MIPI_CAL_CONFIG_CSIC },
-	{ .reg = MIPI_CAL_CONFIG_CSID },
-	{ .reg = MIPI_CAL_CONFIG_CSIE },
-	{ .reg = MIPI_CAL_CONFIG_DSIA },
-	{ .reg = MIPI_CAL_CONFIG_DSIB },
-	{ .reg = MIPI_CAL_CONFIG_DSIC },
-	{ .reg = MIPI_CAL_CONFIG_DSID },
+struct tegra_mipi_pad {
+	unsigned long data;
+	unsigned long clk;
+};
+
+struct tegra_mipi_soc {
+	bool has_clk_lane;
+	const struct tegra_mipi_pad *pads;
+	unsigned int num_pads;
 };
 
 struct tegra_mipi {
+	const struct tegra_mipi_soc *soc;
 	void __iomem *regs;
 	struct mutex lock;
 	struct clk *clk;
@@ -90,16 +102,16 @@ struct tegra_mipi_device {
 	unsigned long pads;
 };
 
-static inline unsigned long tegra_mipi_readl(struct tegra_mipi *mipi,
-					     unsigned long reg)
+static inline u32 tegra_mipi_readl(struct tegra_mipi *mipi,
+				   unsigned long offset)
 {
-	return readl(mipi->regs + (reg << 2));
+	return readl(mipi->regs + (offset << 2));
 }
 
-static inline void tegra_mipi_writel(struct tegra_mipi *mipi,
-				     unsigned long value, unsigned long reg)
+static inline void tegra_mipi_writel(struct tegra_mipi *mipi, u32 value,
+				     unsigned long offset)
 {
-	writel(value, mipi->regs + (reg << 2));
+	writel(value, mipi->regs + (offset << 2));
 }
 
 struct tegra_mipi_device *tegra_mipi_request(struct device *device)
@@ -117,36 +129,35 @@ struct tegra_mipi_device *tegra_mipi_request(struct device *device)
 
 	dev = kzalloc(sizeof(*dev), GFP_KERNEL);
 	if (!dev) {
-		of_node_put(args.np);
 		err = -ENOMEM;
 		goto out;
 	}
 
 	dev->pdev = of_find_device_by_node(args.np);
 	if (!dev->pdev) {
-		of_node_put(args.np);
 		err = -ENODEV;
 		goto free;
 	}
 
-	of_node_put(args.np);
-
 	dev->mipi = platform_get_drvdata(dev->pdev);
 	if (!dev->mipi) {
 		err = -EPROBE_DEFER;
-		goto pdev_put;
+		goto put;
 	}
 
+	of_node_put(args.np);
+
 	dev->pads = args.args[0];
 	dev->device = device;
 
 	return dev;
 
-pdev_put:
+put:
 	platform_device_put(dev->pdev);
 free:
 	kfree(dev);
 out:
+	of_node_put(args.np);
 	return ERR_PTR(err);
 }
 EXPORT_SYMBOL(tegra_mipi_request);
@@ -161,7 +172,7 @@ EXPORT_SYMBOL(tegra_mipi_free);
 static int tegra_mipi_wait(struct tegra_mipi *mipi)
 {
 	unsigned long timeout = jiffies + msecs_to_jiffies(250);
-	unsigned long value;
+	u32 value;
 
 	while (time_before(jiffies, timeout)) {
 		value = tegra_mipi_readl(mipi, MIPI_CAL_STATUS);
@@ -177,8 +188,9 @@ static int tegra_mipi_wait(struct tegra_mipi *mipi)
 
 int tegra_mipi_calibrate(struct tegra_mipi_device *device)
 {
-	unsigned long value;
+	const struct tegra_mipi_soc *soc = device->mipi->soc;
 	unsigned int i;
+	u32 value;
 	int err;
 
 	err = clk_enable(device->mipi->clk);
@@ -192,23 +204,35 @@ int tegra_mipi_calibrate(struct tegra_mipi_device *device)
 	value |= MIPI_CAL_BIAS_PAD_E_VCLAMP_REF;
 	tegra_mipi_writel(device->mipi, value, MIPI_CAL_BIAS_PAD_CFG0);
 
+	tegra_mipi_writel(device->mipi, MIPI_CAL_BIAS_PAD_DRV_DN_REF(2),
+			  MIPI_CAL_BIAS_PAD_CFG1);
+
 	value = tegra_mipi_readl(device->mipi, MIPI_CAL_BIAS_PAD_CFG2);
 	value &= ~MIPI_CAL_BIAS_PAD_PDVREG;
 	tegra_mipi_writel(device->mipi, value, MIPI_CAL_BIAS_PAD_CFG2);
 
-	for (i = 0; i < ARRAY_SIZE(modules); i++) {
-		if (device->pads & BIT(i))
-			value = MIPI_CAL_CONFIG_SELECT |
-				MIPI_CAL_CONFIG_HSPDOS(0) |
-				MIPI_CAL_CONFIG_HSPUOS(4) |
-				MIPI_CAL_CONFIG_TERMOS(5);
-		else
-			value = 0;
+	for (i = 0; i < soc->num_pads; i++) {
+		u32 clk = 0, data = 0;
+
+		if (device->pads & BIT(i)) {
+			data = MIPI_CAL_CONFIG_SELECT |
+			       MIPI_CAL_CONFIG_HSPDOS(0) |
+			       MIPI_CAL_CONFIG_HSPUOS(4) |
+			       MIPI_CAL_CONFIG_TERMOS(5);
+			clk = MIPI_CAL_CONFIG_SELECT |
+			      MIPI_CAL_CONFIG_HSCLKPDOSD(0) |
+			      MIPI_CAL_CONFIG_HSCLKPUOSD(4);
+		}
 
-		tegra_mipi_writel(device->mipi, value, modules[i].reg);
+		tegra_mipi_writel(device->mipi, data, soc->pads[i].data);
+
+		if (soc->has_clk_lane)
+			tegra_mipi_writel(device->mipi, clk, soc->pads[i].clk);
 	}
 
-	tegra_mipi_writel(device->mipi, MIPI_CAL_CTRL_START, MIPI_CAL_CTRL);
+	value = tegra_mipi_readl(device->mipi, MIPI_CAL_CTRL);
+	value |= MIPI_CAL_CTRL_START;
+	tegra_mipi_writel(device->mipi, value, MIPI_CAL_CTRL);
 
 	err = tegra_mipi_wait(device->mipi);
 
@@ -219,16 +243,63 @@ int tegra_mipi_calibrate(struct tegra_mipi_device *device)
 }
 EXPORT_SYMBOL(tegra_mipi_calibrate);
 
+static const struct tegra_mipi_pad tegra114_mipi_pads[] = {
+	{ .data = MIPI_CAL_CONFIG_CSIA },
+	{ .data = MIPI_CAL_CONFIG_CSIB },
+	{ .data = MIPI_CAL_CONFIG_CSIC },
+	{ .data = MIPI_CAL_CONFIG_CSID },
+	{ .data = MIPI_CAL_CONFIG_CSIE },
+	{ .data = MIPI_CAL_CONFIG_DSIA },
+	{ .data = MIPI_CAL_CONFIG_DSIB },
+	{ .data = MIPI_CAL_CONFIG_DSIC },
+	{ .data = MIPI_CAL_CONFIG_DSID },
+};
+
+static const struct tegra_mipi_soc tegra114_mipi_soc = {
+	.has_clk_lane = false,
+	.pads = tegra114_mipi_pads,
+	.num_pads = ARRAY_SIZE(tegra114_mipi_pads),
+};
+
+static const struct tegra_mipi_pad tegra124_mipi_pads[] = {
+	{ .data = MIPI_CAL_CONFIG_CSIA, .clk = MIPI_CAL_CONFIG_CSIAB_CLK },
+	{ .data = MIPI_CAL_CONFIG_CSIB, .clk = MIPI_CAL_CONFIG_CSIAB_CLK },
+	{ .data = MIPI_CAL_CONFIG_CSIC, .clk = MIPI_CAL_CONFIG_CSICD_CLK },
+	{ .data = MIPI_CAL_CONFIG_CSID, .clk = MIPI_CAL_CONFIG_CSICD_CLK },
+	{ .data = MIPI_CAL_CONFIG_CSIE, .clk = MIPI_CAL_CONFIG_CSIE_CLK },
+	{ .data = MIPI_CAL_CONFIG_DSIA, .clk = MIPI_CAL_CONFIG_DSIAB_CLK },
+	{ .data = MIPI_CAL_CONFIG_DSIB, .clk = MIPI_CAL_CONFIG_DSIAB_CLK },
+};
+
+static const struct tegra_mipi_soc tegra124_mipi_soc = {
+	.has_clk_lane = true,
+	.pads = tegra124_mipi_pads,
+	.num_pads = ARRAY_SIZE(tegra124_mipi_pads),
+};
+
+static struct of_device_id tegra_mipi_of_match[] = {
+	{ .compatible = "nvidia,tegra114-mipi", .data = &tegra114_mipi_soc },
+	{ .compatible = "nvidia,tegra124-mipi", .data = &tegra124_mipi_soc },
+	{ },
+};
+
 static int tegra_mipi_probe(struct platform_device *pdev)
 {
+	const struct of_device_id *match;
 	struct tegra_mipi *mipi;
 	struct resource *res;
 	int err;
 
+	match = of_match_node(tegra_mipi_of_match, pdev->dev.of_node);
+	if (!match)
+		return -ENODEV;
+
 	mipi = devm_kzalloc(&pdev->dev, sizeof(*mipi), GFP_KERNEL);
 	if (!mipi)
 		return -ENOMEM;
 
+	mipi->soc = match->data;
+
 	res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
 	mipi->regs = devm_ioremap_resource(&pdev->dev, res);
 	if (IS_ERR(mipi->regs))
@@ -260,11 +331,6 @@ static int tegra_mipi_remove(struct platform_device *pdev)
 	return 0;
 }
 
-static struct of_device_id tegra_mipi_of_match[] = {
-	{ .compatible = "nvidia,tegra114-mipi", },
-	{ },
-};
-
 struct platform_driver tegra_mipi_driver = {
 	.driver = {
 		.name = "tegra-mipi",
diff --git a/drivers/gpu/ipu-v3/Kconfig b/drivers/gpu/ipu-v3/Kconfig
index 2f228a2f2a48..aefdff95356d 100644
--- a/drivers/gpu/ipu-v3/Kconfig
+++ b/drivers/gpu/ipu-v3/Kconfig
@@ -1,7 +1,8 @@
 config IMX_IPUV3_CORE
 	tristate "IPUv3 core support"
-	depends on SOC_IMX5 || SOC_IMX6Q || SOC_IMX6SL || ARCH_MULTIPLATFORM
+	depends on SOC_IMX5 || SOC_IMX6Q || ARCH_MULTIPLATFORM
 	depends on RESET_CONTROLLER
+	select GENERIC_IRQ_CHIP
 	help
 	  Choose this if you have a i.MX5/6 system and want to use the Image
 	  Processing Unit. This option only enables IPU base support.
diff --git a/drivers/gpu/ipu-v3/Makefile b/drivers/gpu/ipu-v3/Makefile
index 1887972b4ac2..107ec236a4a6 100644
--- a/drivers/gpu/ipu-v3/Makefile
+++ b/drivers/gpu/ipu-v3/Makefile
@@ -1,3 +1,4 @@
 obj-$(CONFIG_IMX_IPUV3_CORE) += imx-ipu-v3.o
 
-imx-ipu-v3-objs := ipu-common.o ipu-dc.o ipu-di.o ipu-dp.o ipu-dmfc.o ipu-smfc.o
+imx-ipu-v3-objs := ipu-common.o ipu-cpmem.o ipu-csi.o ipu-dc.o ipu-di.o \
+		ipu-dp.o ipu-dmfc.o ipu-ic.o ipu-smfc.o
diff --git a/drivers/gpu/ipu-v3/ipu-common.c b/drivers/gpu/ipu-v3/ipu-common.c
index 04e7b2eafbdd..f707d25ae78f 100644
--- a/drivers/gpu/ipu-v3/ipu-common.c
+++ b/drivers/gpu/ipu-v3/ipu-common.c
@@ -44,17 +44,6 @@ static inline void ipu_cm_write(struct ipu_soc *ipu, u32 value, unsigned offset)
 	writel(value, ipu->cm_reg + offset);
 }
 
-static inline u32 ipu_idmac_read(struct ipu_soc *ipu, unsigned offset)
-{
-	return readl(ipu->idmac_reg + offset);
-}
-
-static inline void ipu_idmac_write(struct ipu_soc *ipu, u32 value,
-		unsigned offset)
-{
-	writel(value, ipu->idmac_reg + offset);
-}
-
 void ipu_srm_dp_sync_update(struct ipu_soc *ipu)
 {
 	u32 val;
@@ -65,457 +54,184 @@ void ipu_srm_dp_sync_update(struct ipu_soc *ipu)
 }
 EXPORT_SYMBOL_GPL(ipu_srm_dp_sync_update);
 
-struct ipu_ch_param __iomem *ipu_get_cpmem(struct ipuv3_channel *channel)
-{
-	struct ipu_soc *ipu = channel->ipu;
-
-	return ipu->cpmem_base + channel->num;
-}
-EXPORT_SYMBOL_GPL(ipu_get_cpmem);
-
-void ipu_cpmem_set_high_priority(struct ipuv3_channel *channel)
-{
-	struct ipu_soc *ipu = channel->ipu;
-	struct ipu_ch_param __iomem *p = ipu_get_cpmem(channel);
-	u32 val;
-
-	if (ipu->ipu_type == IPUV3EX)
-		ipu_ch_param_write_field(p, IPU_FIELD_ID, 1);
-
-	val = ipu_idmac_read(ipu, IDMAC_CHA_PRI(channel->num));
-	val |= 1 << (channel->num % 32);
-	ipu_idmac_write(ipu, val, IDMAC_CHA_PRI(channel->num));
-};
-EXPORT_SYMBOL_GPL(ipu_cpmem_set_high_priority);
-
-void ipu_ch_param_write_field(struct ipu_ch_param __iomem *base, u32 wbs, u32 v)
-{
-	u32 bit = (wbs >> 8) % 160;
-	u32 size = wbs & 0xff;
-	u32 word = (wbs >> 8) / 160;
-	u32 i = bit / 32;
-	u32 ofs = bit % 32;
-	u32 mask = (1 << size) - 1;
-	u32 val;
-
-	pr_debug("%s %d %d %d\n", __func__, word, bit , size);
-
-	val = readl(&base->word[word].data[i]);
-	val &= ~(mask << ofs);
-	val |= v << ofs;
-	writel(val, &base->word[word].data[i]);
-
-	if ((bit + size - 1) / 32 > i) {
-		val = readl(&base->word[word].data[i + 1]);
-		val &= ~(mask >> (ofs ? (32 - ofs) : 0));
-		val |= v >> (ofs ? (32 - ofs) : 0);
-		writel(val, &base->word[word].data[i + 1]);
-	}
-}
-EXPORT_SYMBOL_GPL(ipu_ch_param_write_field);
-
-u32 ipu_ch_param_read_field(struct ipu_ch_param __iomem *base, u32 wbs)
-{
-	u32 bit = (wbs >> 8) % 160;
-	u32 size = wbs & 0xff;
-	u32 word = (wbs >> 8) / 160;
-	u32 i = bit / 32;
-	u32 ofs = bit % 32;
-	u32 mask = (1 << size) - 1;
-	u32 val = 0;
-
-	pr_debug("%s %d %d %d\n", __func__, word, bit , size);
-
-	val = (readl(&base->word[word].data[i]) >> ofs) & mask;
-
-	if ((bit + size - 1) / 32 > i) {
-		u32 tmp;
-		tmp = readl(&base->word[word].data[i + 1]);
-		tmp &= mask >> (ofs ? (32 - ofs) : 0);
-		val |= tmp << (ofs ? (32 - ofs) : 0);
-	}
-
-	return val;
-}
-EXPORT_SYMBOL_GPL(ipu_ch_param_read_field);
-
-int ipu_cpmem_set_format_rgb(struct ipu_ch_param __iomem *p,
-		const struct ipu_rgb *rgb)
-{
-	int bpp = 0, npb = 0, ro, go, bo, to;
-
-	ro = rgb->bits_per_pixel - rgb->red.length - rgb->red.offset;
-	go = rgb->bits_per_pixel - rgb->green.length - rgb->green.offset;
-	bo = rgb->bits_per_pixel - rgb->blue.length - rgb->blue.offset;
-	to = rgb->bits_per_pixel - rgb->transp.length - rgb->transp.offset;
-
-	ipu_ch_param_write_field(p, IPU_FIELD_WID0, rgb->red.length - 1);
-	ipu_ch_param_write_field(p, IPU_FIELD_OFS0, ro);
-	ipu_ch_param_write_field(p, IPU_FIELD_WID1, rgb->green.length - 1);
-	ipu_ch_param_write_field(p, IPU_FIELD_OFS1, go);
-	ipu_ch_param_write_field(p, IPU_FIELD_WID2, rgb->blue.length - 1);
-	ipu_ch_param_write_field(p, IPU_FIELD_OFS2, bo);
-
-	if (rgb->transp.length) {
-		ipu_ch_param_write_field(p, IPU_FIELD_WID3,
-				rgb->transp.length - 1);
-		ipu_ch_param_write_field(p, IPU_FIELD_OFS3, to);
-	} else {
-		ipu_ch_param_write_field(p, IPU_FIELD_WID3, 7);
-		ipu_ch_param_write_field(p, IPU_FIELD_OFS3,
-				rgb->bits_per_pixel);
-	}
-
-	switch (rgb->bits_per_pixel) {
-	case 32:
-		bpp = 0;
-		npb = 15;
-		break;
-	case 24:
-		bpp = 1;
-		npb = 19;
-		break;
-	case 16:
-		bpp = 3;
-		npb = 31;
-		break;
-	case 8:
-		bpp = 5;
-		npb = 63;
-		break;
-	default:
-		return -EINVAL;
-	}
-	ipu_ch_param_write_field(p, IPU_FIELD_BPP, bpp);
-	ipu_ch_param_write_field(p, IPU_FIELD_NPB, npb);
-	ipu_ch_param_write_field(p, IPU_FIELD_PFS, 7); /* rgb mode */
-
-	return 0;
-}
-EXPORT_SYMBOL_GPL(ipu_cpmem_set_format_rgb);
-
-int ipu_cpmem_set_format_passthrough(struct ipu_ch_param __iomem *p,
-		int width)
+enum ipu_color_space ipu_drm_fourcc_to_colorspace(u32 drm_fourcc)
 {
-	int bpp = 0, npb = 0;
-
-	switch (width) {
-	case 32:
-		bpp = 0;
-		npb = 15;
-		break;
-	case 24:
-		bpp = 1;
-		npb = 19;
-		break;
-	case 16:
-		bpp = 3;
-		npb = 31;
-		break;
-	case 8:
-		bpp = 5;
-		npb = 63;
-		break;
+	switch (drm_fourcc) {
+	case DRM_FORMAT_RGB565:
+	case DRM_FORMAT_BGR565:
+	case DRM_FORMAT_RGB888:
+	case DRM_FORMAT_BGR888:
+	case DRM_FORMAT_XRGB8888:
+	case DRM_FORMAT_XBGR8888:
+	case DRM_FORMAT_RGBX8888:
+	case DRM_FORMAT_BGRX8888:
+	case DRM_FORMAT_ARGB8888:
+	case DRM_FORMAT_ABGR8888:
+	case DRM_FORMAT_RGBA8888:
+	case DRM_FORMAT_BGRA8888:
+		return IPUV3_COLORSPACE_RGB;
+	case DRM_FORMAT_YUYV:
+	case DRM_FORMAT_UYVY:
+	case DRM_FORMAT_YUV420:
+	case DRM_FORMAT_YVU420:
+	case DRM_FORMAT_YUV422:
+	case DRM_FORMAT_YVU422:
+	case DRM_FORMAT_NV12:
+	case DRM_FORMAT_NV21:
+	case DRM_FORMAT_NV16:
+	case DRM_FORMAT_NV61:
+		return IPUV3_COLORSPACE_YUV;
 	default:
-		return -EINVAL;
+		return IPUV3_COLORSPACE_UNKNOWN;
 	}
-
-	ipu_ch_param_write_field(p, IPU_FIELD_BPP, bpp);
-	ipu_ch_param_write_field(p, IPU_FIELD_NPB, npb);
-	ipu_ch_param_write_field(p, IPU_FIELD_PFS, 6); /* raw mode */
-
-	return 0;
 }
-EXPORT_SYMBOL_GPL(ipu_cpmem_set_format_passthrough);
+EXPORT_SYMBOL_GPL(ipu_drm_fourcc_to_colorspace);
 
-void ipu_cpmem_set_yuv_interleaved(struct ipu_ch_param __iomem *p,
-				   u32 pixel_format)
+enum ipu_color_space ipu_pixelformat_to_colorspace(u32 pixelformat)
 {
-	switch (pixel_format) {
+	switch (pixelformat) {
+	case V4L2_PIX_FMT_YUV420:
+	case V4L2_PIX_FMT_YVU420:
+	case V4L2_PIX_FMT_YUV422P:
 	case V4L2_PIX_FMT_UYVY:
-		ipu_ch_param_write_field(p, IPU_FIELD_BPP, 3);    /* bits/pixel */
-		ipu_ch_param_write_field(p, IPU_FIELD_PFS, 0xA);  /* pix format */
-		ipu_ch_param_write_field(p, IPU_FIELD_NPB, 31);   /* burst size */
-		break;
 	case V4L2_PIX_FMT_YUYV:
-		ipu_ch_param_write_field(p, IPU_FIELD_BPP, 3);    /* bits/pixel */
-		ipu_ch_param_write_field(p, IPU_FIELD_PFS, 0x8);  /* pix format */
-		ipu_ch_param_write_field(p, IPU_FIELD_NPB, 31);   /* burst size */
-		break;
+	case V4L2_PIX_FMT_NV12:
+	case V4L2_PIX_FMT_NV21:
+	case V4L2_PIX_FMT_NV16:
+	case V4L2_PIX_FMT_NV61:
+		return IPUV3_COLORSPACE_YUV;
+	case V4L2_PIX_FMT_RGB32:
+	case V4L2_PIX_FMT_BGR32:
+	case V4L2_PIX_FMT_RGB24:
+	case V4L2_PIX_FMT_BGR24:
+	case V4L2_PIX_FMT_RGB565:
+		return IPUV3_COLORSPACE_RGB;
+	default:
+		return IPUV3_COLORSPACE_UNKNOWN;
 	}
 }
-EXPORT_SYMBOL_GPL(ipu_cpmem_set_yuv_interleaved);
+EXPORT_SYMBOL_GPL(ipu_pixelformat_to_colorspace);
 
-void ipu_cpmem_set_yuv_planar_full(struct ipu_ch_param __iomem *p,
-		u32 pixel_format, int stride, int u_offset, int v_offset)
+bool ipu_pixelformat_is_planar(u32 pixelformat)
 {
-	switch (pixel_format) {
+	switch (pixelformat) {
 	case V4L2_PIX_FMT_YUV420:
-		ipu_ch_param_write_field(p, IPU_FIELD_SLUV, (stride / 2) - 1);
-		ipu_ch_param_write_field(p, IPU_FIELD_UBO, u_offset / 8);
-		ipu_ch_param_write_field(p, IPU_FIELD_VBO, v_offset / 8);
-		break;
 	case V4L2_PIX_FMT_YVU420:
-		ipu_ch_param_write_field(p, IPU_FIELD_SLUV, (stride / 2) - 1);
-		ipu_ch_param_write_field(p, IPU_FIELD_UBO, v_offset / 8);
-		ipu_ch_param_write_field(p, IPU_FIELD_VBO, u_offset / 8);
-		break;
+	case V4L2_PIX_FMT_YUV422P:
+	case V4L2_PIX_FMT_NV12:
+	case V4L2_PIX_FMT_NV21:
+	case V4L2_PIX_FMT_NV16:
+	case V4L2_PIX_FMT_NV61:
+		return true;
 	}
-}
-EXPORT_SYMBOL_GPL(ipu_cpmem_set_yuv_planar_full);
-
-void ipu_cpmem_set_yuv_planar(struct ipu_ch_param __iomem *p, u32 pixel_format,
-		int stride, int height)
-{
-	int u_offset, v_offset;
-	int uv_stride = 0;
 
-	switch (pixel_format) {
-	case V4L2_PIX_FMT_YUV420:
-	case V4L2_PIX_FMT_YVU420:
-		uv_stride = stride / 2;
-		u_offset = stride * height;
-		v_offset = u_offset + (uv_stride * height / 2);
-		ipu_cpmem_set_yuv_planar_full(p, pixel_format, stride,
-				u_offset, v_offset);
-		break;
-	}
+	return false;
 }
-EXPORT_SYMBOL_GPL(ipu_cpmem_set_yuv_planar);
-
-static const struct ipu_rgb def_rgb_32 = {
-	.red	= { .offset = 16, .length = 8, },
-	.green	= { .offset =  8, .length = 8, },
-	.blue	= { .offset =  0, .length = 8, },
-	.transp = { .offset = 24, .length = 8, },
-	.bits_per_pixel = 32,
-};
-
-static const struct ipu_rgb def_bgr_32 = {
-	.red	= { .offset =  0, .length = 8, },
-	.green	= { .offset =  8, .length = 8, },
-	.blue	= { .offset = 16, .length = 8, },
-	.transp = { .offset = 24, .length = 8, },
-	.bits_per_pixel = 32,
-};
-
-static const struct ipu_rgb def_rgb_24 = {
-	.red	= { .offset = 16, .length = 8, },
-	.green	= { .offset =  8, .length = 8, },
-	.blue	= { .offset =  0, .length = 8, },
-	.transp = { .offset =  0, .length = 0, },
-	.bits_per_pixel = 24,
-};
-
-static const struct ipu_rgb def_bgr_24 = {
-	.red	= { .offset =  0, .length = 8, },
-	.green	= { .offset =  8, .length = 8, },
-	.blue	= { .offset = 16, .length = 8, },
-	.transp = { .offset =  0, .length = 0, },
-	.bits_per_pixel = 24,
-};
-
-static const struct ipu_rgb def_rgb_16 = {
-	.red	= { .offset = 11, .length = 5, },
-	.green	= { .offset =  5, .length = 6, },
-	.blue	= { .offset =  0, .length = 5, },
-	.transp = { .offset =  0, .length = 0, },
-	.bits_per_pixel = 16,
-};
-
-static const struct ipu_rgb def_bgr_16 = {
-	.red	= { .offset =  0, .length = 5, },
-	.green	= { .offset =  5, .length = 6, },
-	.blue	= { .offset = 11, .length = 5, },
-	.transp = { .offset =  0, .length = 0, },
-	.bits_per_pixel = 16,
-};
-
-#define Y_OFFSET(pix, x, y)	((x) + pix->width * (y))
-#define U_OFFSET(pix, x, y)	((pix->width * pix->height) + \
-					(pix->width * (y) / 4) + (x) / 2)
-#define V_OFFSET(pix, x, y)	((pix->width * pix->height) + \
-					(pix->width * pix->height / 4) + \
-					(pix->width * (y) / 4) + (x) / 2)
+EXPORT_SYMBOL_GPL(ipu_pixelformat_is_planar);
 
-int ipu_cpmem_set_fmt(struct ipu_ch_param __iomem *cpmem, u32 drm_fourcc)
+enum ipu_color_space ipu_mbus_code_to_colorspace(u32 mbus_code)
 {
-	switch (drm_fourcc) {
-	case DRM_FORMAT_YUV420:
-	case DRM_FORMAT_YVU420:
-		/* pix format */
-		ipu_ch_param_write_field(cpmem, IPU_FIELD_PFS, 2);
-		/* burst size */
-		ipu_ch_param_write_field(cpmem, IPU_FIELD_NPB, 63);
-		break;
-	case DRM_FORMAT_UYVY:
-		/* bits/pixel */
-		ipu_ch_param_write_field(cpmem, IPU_FIELD_BPP, 3);
-		/* pix format */
-		ipu_ch_param_write_field(cpmem, IPU_FIELD_PFS, 0xA);
-		/* burst size */
-		ipu_ch_param_write_field(cpmem, IPU_FIELD_NPB, 31);
-		break;
-	case DRM_FORMAT_YUYV:
-		/* bits/pixel */
-		ipu_ch_param_write_field(cpmem, IPU_FIELD_BPP, 3);
-		/* pix format */
-		ipu_ch_param_write_field(cpmem, IPU_FIELD_PFS, 0x8);
-		/* burst size */
-		ipu_ch_param_write_field(cpmem, IPU_FIELD_NPB, 31);
-		break;
-	case DRM_FORMAT_ABGR8888:
-	case DRM_FORMAT_XBGR8888:
-		ipu_cpmem_set_format_rgb(cpmem, &def_bgr_32);
-		break;
-	case DRM_FORMAT_ARGB8888:
-	case DRM_FORMAT_XRGB8888:
-		ipu_cpmem_set_format_rgb(cpmem, &def_rgb_32);
-		break;
-	case DRM_FORMAT_BGR888:
-		ipu_cpmem_set_format_rgb(cpmem, &def_bgr_24);
-		break;
-	case DRM_FORMAT_RGB888:
-		ipu_cpmem_set_format_rgb(cpmem, &def_rgb_24);
-		break;
-	case DRM_FORMAT_RGB565:
-		ipu_cpmem_set_format_rgb(cpmem, &def_rgb_16);
-		break;
-	case DRM_FORMAT_BGR565:
-		ipu_cpmem_set_format_rgb(cpmem, &def_bgr_16);
-		break;
+	switch (mbus_code & 0xf000) {
+	case 0x1000:
+		return IPUV3_COLORSPACE_RGB;
+	case 0x2000:
+		return IPUV3_COLORSPACE_YUV;
 	default:
-		return -EINVAL;
+		return IPUV3_COLORSPACE_UNKNOWN;
 	}
-
-	return 0;
 }
-EXPORT_SYMBOL_GPL(ipu_cpmem_set_fmt);
+EXPORT_SYMBOL_GPL(ipu_mbus_code_to_colorspace);
 
-/*
- * The V4L2 spec defines packed RGB formats in memory byte order, which from
- * point of view of the IPU corresponds to little-endian words with the first
- * component in the least significant bits.
- * The DRM pixel formats and IPU internal representation are ordered the other
- * way around, with the first named component ordered at the most significant
- * bits. Further, V4L2 formats are not well defined:
- *     http://linuxtv.org/downloads/v4l-dvb-apis/packed-rgb.html
- * We choose the interpretation which matches GStreamer behavior.
- */
-static int v4l2_pix_fmt_to_drm_fourcc(u32 pixelformat)
+int ipu_stride_to_bytes(u32 pixel_stride, u32 pixelformat)
 {
 	switch (pixelformat) {
-	case V4L2_PIX_FMT_RGB565:
+	case V4L2_PIX_FMT_YUV420:
+	case V4L2_PIX_FMT_YVU420:
+	case V4L2_PIX_FMT_YUV422P:
+	case V4L2_PIX_FMT_NV12:
+	case V4L2_PIX_FMT_NV21:
+	case V4L2_PIX_FMT_NV16:
+	case V4L2_PIX_FMT_NV61:
 		/*
-		 * Here we choose the 'corrected' interpretation of RGBP, a
-		 * little-endian 16-bit word with the red component at the most
-		 * significant bits:
-		 * g[2:0]b[4:0] r[4:0]g[5:3] <=> [16:0] R:G:B
+		 * for the planar YUV formats, the stride passed to
+		 * cpmem must be the stride in bytes of the Y plane.
+		 * And all the planar YUV formats have an 8-bit
+		 * Y component.
 		 */
-		return DRM_FORMAT_RGB565;
+		return (8 * pixel_stride) >> 3;
+	case V4L2_PIX_FMT_RGB565:
+	case V4L2_PIX_FMT_YUYV:
+	case V4L2_PIX_FMT_UYVY:
+		return (16 * pixel_stride) >> 3;
 	case V4L2_PIX_FMT_BGR24:
-		/* B G R <=> [24:0] R:G:B */
-		return DRM_FORMAT_RGB888;
 	case V4L2_PIX_FMT_RGB24:
-		/* R G B <=> [24:0] B:G:R */
-		return DRM_FORMAT_BGR888;
+		return (24 * pixel_stride) >> 3;
 	case V4L2_PIX_FMT_BGR32:
-		/* B G R A <=> [32:0] A:B:G:R */
-		return DRM_FORMAT_XRGB8888;
 	case V4L2_PIX_FMT_RGB32:
-		/* R G B A <=> [32:0] A:B:G:R */
-		return DRM_FORMAT_XBGR8888;
-	case V4L2_PIX_FMT_UYVY:
-		return DRM_FORMAT_UYVY;
-	case V4L2_PIX_FMT_YUYV:
-		return DRM_FORMAT_YUYV;
-	case V4L2_PIX_FMT_YUV420:
-		return DRM_FORMAT_YUV420;
-	case V4L2_PIX_FMT_YVU420:
-		return DRM_FORMAT_YVU420;
+		return (32 * pixel_stride) >> 3;
+	default:
+		break;
 	}
 
 	return -EINVAL;
 }
+EXPORT_SYMBOL_GPL(ipu_stride_to_bytes);
 
-enum ipu_color_space ipu_drm_fourcc_to_colorspace(u32 drm_fourcc)
+int ipu_degrees_to_rot_mode(enum ipu_rotate_mode *mode, int degrees,
+			    bool hflip, bool vflip)
 {
-	switch (drm_fourcc) {
-	case DRM_FORMAT_RGB565:
-	case DRM_FORMAT_BGR565:
-	case DRM_FORMAT_RGB888:
-	case DRM_FORMAT_BGR888:
-	case DRM_FORMAT_XRGB8888:
-	case DRM_FORMAT_XBGR8888:
-	case DRM_FORMAT_RGBX8888:
-	case DRM_FORMAT_BGRX8888:
-	case DRM_FORMAT_ARGB8888:
-	case DRM_FORMAT_ABGR8888:
-	case DRM_FORMAT_RGBA8888:
-	case DRM_FORMAT_BGRA8888:
-		return IPUV3_COLORSPACE_RGB;
-	case DRM_FORMAT_YUYV:
-	case DRM_FORMAT_UYVY:
-	case DRM_FORMAT_YUV420:
-	case DRM_FORMAT_YVU420:
-		return IPUV3_COLORSPACE_YUV;
+	u32 r90, vf, hf;
+
+	switch (degrees) {
+	case 0:
+		vf = hf = r90 = 0;
+		break;
+	case 90:
+		vf = hf = 0;
+		r90 = 1;
+		break;
+	case 180:
+		vf = hf = 1;
+		r90 = 0;
+		break;
+	case 270:
+		vf = hf = r90 = 1;
+		break;
 	default:
-		return IPUV3_COLORSPACE_UNKNOWN;
+		return -EINVAL;
 	}
-}
-EXPORT_SYMBOL_GPL(ipu_drm_fourcc_to_colorspace);
 
-int ipu_cpmem_set_image(struct ipu_ch_param __iomem *cpmem,
-		struct ipu_image *image)
-{
-	struct v4l2_pix_format *pix = &image->pix;
-	int y_offset, u_offset, v_offset;
+	hf ^= (u32)hflip;
+	vf ^= (u32)vflip;
 
-	pr_debug("%s: resolution: %dx%d stride: %d\n",
-			__func__, pix->width, pix->height,
-			pix->bytesperline);
+	*mode = (enum ipu_rotate_mode)((r90 << 2) | (hf << 1) | vf);
+	return 0;
+}
+EXPORT_SYMBOL_GPL(ipu_degrees_to_rot_mode);
 
-	ipu_cpmem_set_resolution(cpmem, image->rect.width,
-			image->rect.height);
-	ipu_cpmem_set_stride(cpmem, pix->bytesperline);
+int ipu_rot_mode_to_degrees(int *degrees, enum ipu_rotate_mode mode,
+			    bool hflip, bool vflip)
+{
+	u32 r90, vf, hf;
 
-	ipu_cpmem_set_fmt(cpmem, v4l2_pix_fmt_to_drm_fourcc(pix->pixelformat));
+	r90 = ((u32)mode >> 2) & 0x1;
+	hf = ((u32)mode >> 1) & 0x1;
+	vf = ((u32)mode >> 0) & 0x1;
+	hf ^= (u32)hflip;
+	vf ^= (u32)vflip;
 
-	switch (pix->pixelformat) {
-	case V4L2_PIX_FMT_YUV420:
-	case V4L2_PIX_FMT_YVU420:
-		y_offset = Y_OFFSET(pix, image->rect.left, image->rect.top);
-		u_offset = U_OFFSET(pix, image->rect.left,
-				image->rect.top) - y_offset;
-		v_offset = V_OFFSET(pix, image->rect.left,
-				image->rect.top) - y_offset;
-
-		ipu_cpmem_set_yuv_planar_full(cpmem, pix->pixelformat,
-				pix->bytesperline, u_offset, v_offset);
-		ipu_cpmem_set_buffer(cpmem, 0, image->phys + y_offset);
+	switch ((enum ipu_rotate_mode)((r90 << 2) | (hf << 1) | vf)) {
+	case IPU_ROTATE_NONE:
+		*degrees = 0;
 		break;
-	case V4L2_PIX_FMT_UYVY:
-	case V4L2_PIX_FMT_YUYV:
-		ipu_cpmem_set_buffer(cpmem, 0, image->phys +
-				image->rect.left * 2 +
-				image->rect.top * image->pix.bytesperline);
+	case IPU_ROTATE_90_RIGHT:
+		*degrees = 90;
 		break;
-	case V4L2_PIX_FMT_RGB32:
-	case V4L2_PIX_FMT_BGR32:
-		ipu_cpmem_set_buffer(cpmem, 0, image->phys +
-				image->rect.left * 4 +
-				image->rect.top * image->pix.bytesperline);
+	case IPU_ROTATE_180:
+		*degrees = 180;
 		break;
-	case V4L2_PIX_FMT_RGB565:
-		ipu_cpmem_set_buffer(cpmem, 0, image->phys +
-				image->rect.left * 2 +
-				image->rect.top * image->pix.bytesperline);
-		break;
-	case V4L2_PIX_FMT_RGB24:
-	case V4L2_PIX_FMT_BGR24:
-		ipu_cpmem_set_buffer(cpmem, 0, image->phys +
-				image->rect.left * 3 +
-				image->rect.top * image->pix.bytesperline);
+	case IPU_ROTATE_90_LEFT:
+		*degrees = 270;
 		break;
 	default:
 		return -EINVAL;
@@ -523,27 +239,7 @@ int ipu_cpmem_set_image(struct ipu_ch_param __iomem *cpmem,
 
 	return 0;
 }
-EXPORT_SYMBOL_GPL(ipu_cpmem_set_image);
-
-enum ipu_color_space ipu_pixelformat_to_colorspace(u32 pixelformat)
-{
-	switch (pixelformat) {
-	case V4L2_PIX_FMT_YUV420:
-	case V4L2_PIX_FMT_YVU420:
-	case V4L2_PIX_FMT_UYVY:
-	case V4L2_PIX_FMT_YUYV:
-		return IPUV3_COLORSPACE_YUV;
-	case V4L2_PIX_FMT_RGB32:
-	case V4L2_PIX_FMT_BGR32:
-	case V4L2_PIX_FMT_RGB24:
-	case V4L2_PIX_FMT_BGR24:
-	case V4L2_PIX_FMT_RGB565:
-		return IPUV3_COLORSPACE_RGB;
-	default:
-		return IPUV3_COLORSPACE_UNKNOWN;
-	}
-}
-EXPORT_SYMBOL_GPL(ipu_pixelformat_to_colorspace);
+EXPORT_SYMBOL_GPL(ipu_rot_mode_to_degrees);
 
 struct ipuv3_channel *ipu_idmac_get(struct ipu_soc *ipu, unsigned num)
 {
@@ -587,7 +283,26 @@ void ipu_idmac_put(struct ipuv3_channel *channel)
 }
 EXPORT_SYMBOL_GPL(ipu_idmac_put);
 
-#define idma_mask(ch)			(1 << (ch & 0x1f))
+#define idma_mask(ch)			(1 << ((ch) & 0x1f))
+
+/*
+ * This is an undocumented feature, a write one to a channel bit in
+ * IPU_CHA_CUR_BUF and IPU_CHA_TRIPLE_CUR_BUF will reset the channel's
+ * internal current buffer pointer so that transfers start from buffer
+ * 0 on the next channel enable (that's the theory anyway, the imx6 TRM
+ * only says these are read-only registers). This operation is required
+ * for channel linking to work correctly, for instance video capture
+ * pipelines that carry out image rotations will fail after the first
+ * streaming unless this function is called for each channel before
+ * re-enabling the channels.
+ */
+static void __ipu_idmac_reset_current_buffer(struct ipuv3_channel *channel)
+{
+	struct ipu_soc *ipu = channel->ipu;
+	unsigned int chno = channel->num;
+
+	ipu_cm_write(ipu, idma_mask(chno), IPU_CHA_CUR_BUF(chno));
+}
 
 void ipu_idmac_set_double_buffer(struct ipuv3_channel *channel,
 		bool doublebuffer)
@@ -605,10 +320,81 @@ void ipu_idmac_set_double_buffer(struct ipuv3_channel *channel,
 		reg &= ~idma_mask(channel->num);
 	ipu_cm_write(ipu, reg, IPU_CHA_DB_MODE_SEL(channel->num));
 
+	__ipu_idmac_reset_current_buffer(channel);
+
 	spin_unlock_irqrestore(&ipu->lock, flags);
 }
 EXPORT_SYMBOL_GPL(ipu_idmac_set_double_buffer);
 
+static const struct {
+	int chnum;
+	u32 reg;
+	int shift;
+} idmac_lock_en_info[] = {
+	{ .chnum =  5, .reg = IDMAC_CH_LOCK_EN_1, .shift =  0, },
+	{ .chnum = 11, .reg = IDMAC_CH_LOCK_EN_1, .shift =  2, },
+	{ .chnum = 12, .reg = IDMAC_CH_LOCK_EN_1, .shift =  4, },
+	{ .chnum = 14, .reg = IDMAC_CH_LOCK_EN_1, .shift =  6, },
+	{ .chnum = 15, .reg = IDMAC_CH_LOCK_EN_1, .shift =  8, },
+	{ .chnum = 20, .reg = IDMAC_CH_LOCK_EN_1, .shift = 10, },
+	{ .chnum = 21, .reg = IDMAC_CH_LOCK_EN_1, .shift = 12, },
+	{ .chnum = 22, .reg = IDMAC_CH_LOCK_EN_1, .shift = 14, },
+	{ .chnum = 23, .reg = IDMAC_CH_LOCK_EN_1, .shift = 16, },
+	{ .chnum = 27, .reg = IDMAC_CH_LOCK_EN_1, .shift = 18, },
+	{ .chnum = 28, .reg = IDMAC_CH_LOCK_EN_1, .shift = 20, },
+	{ .chnum = 45, .reg = IDMAC_CH_LOCK_EN_2, .shift =  0, },
+	{ .chnum = 46, .reg = IDMAC_CH_LOCK_EN_2, .shift =  2, },
+	{ .chnum = 47, .reg = IDMAC_CH_LOCK_EN_2, .shift =  4, },
+	{ .chnum = 48, .reg = IDMAC_CH_LOCK_EN_2, .shift =  6, },
+	{ .chnum = 49, .reg = IDMAC_CH_LOCK_EN_2, .shift =  8, },
+	{ .chnum = 50, .reg = IDMAC_CH_LOCK_EN_2, .shift = 10, },
+};
+
+int ipu_idmac_lock_enable(struct ipuv3_channel *channel, int num_bursts)
+{
+	struct ipu_soc *ipu = channel->ipu;
+	unsigned long flags;
+	u32 bursts, regval;
+	int i;
+
+	switch (num_bursts) {
+	case 0:
+	case 1:
+		bursts = 0x00; /* locking disabled */
+		break;
+	case 2:
+		bursts = 0x01;
+		break;
+	case 4:
+		bursts = 0x02;
+		break;
+	case 8:
+		bursts = 0x03;
+		break;
+	default:
+		return -EINVAL;
+	}
+
+	for (i = 0; i < ARRAY_SIZE(idmac_lock_en_info); i++) {
+		if (channel->num == idmac_lock_en_info[i].chnum)
+			break;
+	}
+	if (i >= ARRAY_SIZE(idmac_lock_en_info))
+		return -EINVAL;
+
+	spin_lock_irqsave(&ipu->lock, flags);
+
+	regval = ipu_idmac_read(ipu, idmac_lock_en_info[i].reg);
+	regval &= ~(0x03 << idmac_lock_en_info[i].shift);
+	regval |= (bursts << idmac_lock_en_info[i].shift);
+	ipu_idmac_write(ipu, regval, idmac_lock_en_info[i].reg);
+
+	spin_unlock_irqrestore(&ipu->lock, flags);
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(ipu_idmac_lock_enable);
+
 int ipu_module_enable(struct ipu_soc *ipu, u32 mask)
 {
 	unsigned long lock_flags;
@@ -661,30 +447,6 @@ int ipu_module_disable(struct ipu_soc *ipu, u32 mask)
 }
 EXPORT_SYMBOL_GPL(ipu_module_disable);
 
-int ipu_csi_enable(struct ipu_soc *ipu, int csi)
-{
-	return ipu_module_enable(ipu, csi ? IPU_CONF_CSI1_EN : IPU_CONF_CSI0_EN);
-}
-EXPORT_SYMBOL_GPL(ipu_csi_enable);
-
-int ipu_csi_disable(struct ipu_soc *ipu, int csi)
-{
-	return ipu_module_disable(ipu, csi ? IPU_CONF_CSI1_EN : IPU_CONF_CSI0_EN);
-}
-EXPORT_SYMBOL_GPL(ipu_csi_disable);
-
-int ipu_smfc_enable(struct ipu_soc *ipu)
-{
-	return ipu_module_enable(ipu, IPU_CONF_SMFC_EN);
-}
-EXPORT_SYMBOL_GPL(ipu_smfc_enable);
-
-int ipu_smfc_disable(struct ipu_soc *ipu)
-{
-	return ipu_module_disable(ipu, IPU_CONF_SMFC_EN);
-}
-EXPORT_SYMBOL_GPL(ipu_smfc_disable);
-
 int ipu_idmac_get_current_buffer(struct ipuv3_channel *channel)
 {
 	struct ipu_soc *ipu = channel->ipu;
@@ -694,6 +456,30 @@ int ipu_idmac_get_current_buffer(struct ipuv3_channel *channel)
 }
 EXPORT_SYMBOL_GPL(ipu_idmac_get_current_buffer);
 
+bool ipu_idmac_buffer_is_ready(struct ipuv3_channel *channel, u32 buf_num)
+{
+	struct ipu_soc *ipu = channel->ipu;
+	unsigned long flags;
+	u32 reg = 0;
+
+	spin_lock_irqsave(&ipu->lock, flags);
+	switch (buf_num) {
+	case 0:
+		reg = ipu_cm_read(ipu, IPU_CHA_BUF0_RDY(channel->num));
+		break;
+	case 1:
+		reg = ipu_cm_read(ipu, IPU_CHA_BUF1_RDY(channel->num));
+		break;
+	case 2:
+		reg = ipu_cm_read(ipu, IPU_CHA_BUF2_RDY(channel->num));
+		break;
+	}
+	spin_unlock_irqrestore(&ipu->lock, flags);
+
+	return ((reg & idma_mask(channel->num)) != 0);
+}
+EXPORT_SYMBOL_GPL(ipu_idmac_buffer_is_ready);
+
 void ipu_idmac_select_buffer(struct ipuv3_channel *channel, u32 buf_num)
 {
 	struct ipu_soc *ipu = channel->ipu;
@@ -712,6 +498,34 @@ void ipu_idmac_select_buffer(struct ipuv3_channel *channel, u32 buf_num)
 }
 EXPORT_SYMBOL_GPL(ipu_idmac_select_buffer);
 
+void ipu_idmac_clear_buffer(struct ipuv3_channel *channel, u32 buf_num)
+{
+	struct ipu_soc *ipu = channel->ipu;
+	unsigned int chno = channel->num;
+	unsigned long flags;
+
+	spin_lock_irqsave(&ipu->lock, flags);
+
+	ipu_cm_write(ipu, 0xF0300000, IPU_GPR); /* write one to clear */
+	switch (buf_num) {
+	case 0:
+		ipu_cm_write(ipu, idma_mask(chno), IPU_CHA_BUF0_RDY(chno));
+		break;
+	case 1:
+		ipu_cm_write(ipu, idma_mask(chno), IPU_CHA_BUF1_RDY(chno));
+		break;
+	case 2:
+		ipu_cm_write(ipu, idma_mask(chno), IPU_CHA_BUF2_RDY(chno));
+		break;
+	default:
+		break;
+	}
+	ipu_cm_write(ipu, 0x0, IPU_GPR); /* write one to set */
+
+	spin_unlock_irqrestore(&ipu->lock, flags);
+}
+EXPORT_SYMBOL_GPL(ipu_idmac_clear_buffer);
+
 int ipu_idmac_enable_channel(struct ipuv3_channel *channel)
 {
 	struct ipu_soc *ipu = channel->ipu;
@@ -782,6 +596,8 @@ int ipu_idmac_disable_channel(struct ipuv3_channel *channel)
 	val &= ~idma_mask(channel->num);
 	ipu_idmac_write(ipu, val, IDMAC_CHA_EN(channel->num));
 
+	__ipu_idmac_reset_current_buffer(channel);
+
 	/* Set channel buffers NOT to be ready */
 	ipu_cm_write(ipu, 0xf0000000, IPU_GPR); /* write one to clear */
 
@@ -810,6 +626,31 @@ int ipu_idmac_disable_channel(struct ipuv3_channel *channel)
 }
 EXPORT_SYMBOL_GPL(ipu_idmac_disable_channel);
 
+/*
+ * The imx6 rev. D TRM says that enabling the WM feature will increase
+ * a channel's priority. Refer to Table 36-8 Calculated priority value.
+ * The sub-module that is the sink or source for the channel must enable
+ * watermark signal for this to take effect (SMFC_WM for instance).
+ */
+void ipu_idmac_enable_watermark(struct ipuv3_channel *channel, bool enable)
+{
+	struct ipu_soc *ipu = channel->ipu;
+	unsigned long flags;
+	u32 val;
+
+	spin_lock_irqsave(&ipu->lock, flags);
+
+	val = ipu_idmac_read(ipu, IDMAC_WM_EN(channel->num));
+	if (enable)
+		val |= 1 << (channel->num % 32);
+	else
+		val &= ~(1 << (channel->num % 32));
+	ipu_idmac_write(ipu, val, IDMAC_WM_EN(channel->num));
+
+	spin_unlock_irqrestore(&ipu->lock, flags);
+}
+EXPORT_SYMBOL_GPL(ipu_idmac_enable_watermark);
+
 static int ipu_memory_reset(struct ipu_soc *ipu)
 {
 	unsigned long timeout;
@@ -826,12 +667,66 @@ static int ipu_memory_reset(struct ipu_soc *ipu)
 	return 0;
 }
 
+/*
+ * Set the source mux for the given CSI. Selects either parallel or
+ * MIPI CSI2 sources.
+ */
+void ipu_set_csi_src_mux(struct ipu_soc *ipu, int csi_id, bool mipi_csi2)
+{
+	unsigned long flags;
+	u32 val, mask;
+
+	mask = (csi_id == 1) ? IPU_CONF_CSI1_DATA_SOURCE :
+		IPU_CONF_CSI0_DATA_SOURCE;
+
+	spin_lock_irqsave(&ipu->lock, flags);
+
+	val = ipu_cm_read(ipu, IPU_CONF);
+	if (mipi_csi2)
+		val |= mask;
+	else
+		val &= ~mask;
+	ipu_cm_write(ipu, val, IPU_CONF);
+
+	spin_unlock_irqrestore(&ipu->lock, flags);
+}
+EXPORT_SYMBOL_GPL(ipu_set_csi_src_mux);
+
+/*
+ * Set the source mux for the IC. Selects either CSI[01] or the VDI.
+ */
+void ipu_set_ic_src_mux(struct ipu_soc *ipu, int csi_id, bool vdi)
+{
+	unsigned long flags;
+	u32 val;
+
+	spin_lock_irqsave(&ipu->lock, flags);
+
+	val = ipu_cm_read(ipu, IPU_CONF);
+	if (vdi) {
+		val |= IPU_CONF_IC_INPUT;
+	} else {
+		val &= ~IPU_CONF_IC_INPUT;
+		if (csi_id == 1)
+			val |= IPU_CONF_CSI_SEL;
+		else
+			val &= ~IPU_CONF_CSI_SEL;
+	}
+	ipu_cm_write(ipu, val, IPU_CONF);
+
+	spin_unlock_irqrestore(&ipu->lock, flags);
+}
+EXPORT_SYMBOL_GPL(ipu_set_ic_src_mux);
+
 struct ipu_devtype {
 	const char *name;
 	unsigned long cm_ofs;
 	unsigned long cpmem_ofs;
 	unsigned long srm_ofs;
 	unsigned long tpm_ofs;
+	unsigned long csi0_ofs;
+	unsigned long csi1_ofs;
+	unsigned long ic_ofs;
 	unsigned long disp0_ofs;
 	unsigned long disp1_ofs;
 	unsigned long dc_tmpl_ofs;
@@ -845,6 +740,9 @@ static struct ipu_devtype ipu_type_imx51 = {
 	.cpmem_ofs = 0x1f000000,
 	.srm_ofs = 0x1f040000,
 	.tpm_ofs = 0x1f060000,
+	.csi0_ofs = 0x1f030000,
+	.csi1_ofs = 0x1f038000,
+	.ic_ofs = 0x1f020000,
 	.disp0_ofs = 0x1e040000,
 	.disp1_ofs = 0x1e048000,
 	.dc_tmpl_ofs = 0x1f080000,
@@ -858,6 +756,9 @@ static struct ipu_devtype ipu_type_imx53 = {
 	.cpmem_ofs = 0x07000000,
 	.srm_ofs = 0x07040000,
 	.tpm_ofs = 0x07060000,
+	.csi0_ofs = 0x07030000,
+	.csi1_ofs = 0x07038000,
+	.ic_ofs = 0x07020000,
 	.disp0_ofs = 0x06040000,
 	.disp1_ofs = 0x06048000,
 	.dc_tmpl_ofs = 0x07080000,
@@ -871,6 +772,9 @@ static struct ipu_devtype ipu_type_imx6q = {
 	.cpmem_ofs = 0x00300000,
 	.srm_ofs = 0x00340000,
 	.tpm_ofs = 0x00360000,
+	.csi0_ofs = 0x00230000,
+	.csi1_ofs = 0x00238000,
+	.ic_ofs = 0x00220000,
 	.disp0_ofs = 0x00240000,
 	.disp1_ofs = 0x00248000,
 	.dc_tmpl_ofs = 0x00380000,
@@ -895,8 +799,36 @@ static int ipu_submodules_init(struct ipu_soc *ipu,
 	struct device *dev = &pdev->dev;
 	const struct ipu_devtype *devtype = ipu->devtype;
 
+	ret = ipu_cpmem_init(ipu, dev, ipu_base + devtype->cpmem_ofs);
+	if (ret) {
+		unit = "cpmem";
+		goto err_cpmem;
+	}
+
+	ret = ipu_csi_init(ipu, dev, 0, ipu_base + devtype->csi0_ofs,
+			   IPU_CONF_CSI0_EN, ipu_clk);
+	if (ret) {
+		unit = "csi0";
+		goto err_csi_0;
+	}
+
+	ret = ipu_csi_init(ipu, dev, 1, ipu_base + devtype->csi1_ofs,
+			   IPU_CONF_CSI1_EN, ipu_clk);
+	if (ret) {
+		unit = "csi1";
+		goto err_csi_1;
+	}
+
+	ret = ipu_ic_init(ipu, dev,
+			  ipu_base + devtype->ic_ofs,
+			  ipu_base + devtype->tpm_ofs);
+	if (ret) {
+		unit = "ic";
+		goto err_ic;
+	}
+
 	ret = ipu_di_init(ipu, dev, 0, ipu_base + devtype->disp0_ofs,
-			IPU_CONF_DI0_EN, ipu_clk);
+			  IPU_CONF_DI0_EN, ipu_clk);
 	if (ret) {
 		unit = "di0";
 		goto err_di_0;
@@ -949,6 +881,14 @@ err_dc:
 err_di_1:
 	ipu_di_exit(ipu, 0);
 err_di_0:
+	ipu_ic_exit(ipu);
+err_ic:
+	ipu_csi_exit(ipu, 1);
+err_csi_1:
+	ipu_csi_exit(ipu, 0);
+err_csi_0:
+	ipu_cpmem_exit(ipu);
+err_cpmem:
 	dev_err(&pdev->dev, "init %s failed with %d\n", unit, ret);
 	return ret;
 }
@@ -1025,6 +965,10 @@ static void ipu_submodules_exit(struct ipu_soc *ipu)
 	ipu_dc_exit(ipu);
 	ipu_di_exit(ipu, 1);
 	ipu_di_exit(ipu, 0);
+	ipu_ic_exit(ipu);
+	ipu_csi_exit(ipu, 1);
+	ipu_csi_exit(ipu, 0);
+	ipu_cpmem_exit(ipu);
 }
 
 static int platform_remove_devices_fn(struct device *dev, void *unused)
@@ -1116,8 +1060,10 @@ static int ipu_add_client_devices(struct ipu_soc *ipu, unsigned long ipu_base)
 				id++, &reg->pdata, sizeof(reg->pdata));
 		}
 
-		if (IS_ERR(pdev))
+		if (IS_ERR(pdev)) {
+			ret = PTR_ERR(pdev);
 			goto err_register;
+		}
 	}
 
 	return 0;
@@ -1201,6 +1147,44 @@ static void ipu_irq_exit(struct ipu_soc *ipu)
 	irq_domain_remove(ipu->domain);
 }
 
+void ipu_dump(struct ipu_soc *ipu)
+{
+	int i;
+
+	dev_dbg(ipu->dev, "IPU_CONF = \t0x%08X\n",
+		ipu_cm_read(ipu, IPU_CONF));
+	dev_dbg(ipu->dev, "IDMAC_CONF = \t0x%08X\n",
+		ipu_idmac_read(ipu, IDMAC_CONF));
+	dev_dbg(ipu->dev, "IDMAC_CHA_EN1 = \t0x%08X\n",
+		ipu_idmac_read(ipu, IDMAC_CHA_EN(0)));
+	dev_dbg(ipu->dev, "IDMAC_CHA_EN2 = \t0x%08X\n",
+		ipu_idmac_read(ipu, IDMAC_CHA_EN(32)));
+	dev_dbg(ipu->dev, "IDMAC_CHA_PRI1 = \t0x%08X\n",
+		ipu_idmac_read(ipu, IDMAC_CHA_PRI(0)));
+	dev_dbg(ipu->dev, "IDMAC_CHA_PRI2 = \t0x%08X\n",
+		ipu_idmac_read(ipu, IDMAC_CHA_PRI(32)));
+	dev_dbg(ipu->dev, "IDMAC_BAND_EN1 = \t0x%08X\n",
+		ipu_idmac_read(ipu, IDMAC_BAND_EN(0)));
+	dev_dbg(ipu->dev, "IDMAC_BAND_EN2 = \t0x%08X\n",
+		ipu_idmac_read(ipu, IDMAC_BAND_EN(32)));
+	dev_dbg(ipu->dev, "IPU_CHA_DB_MODE_SEL0 = \t0x%08X\n",
+		ipu_cm_read(ipu, IPU_CHA_DB_MODE_SEL(0)));
+	dev_dbg(ipu->dev, "IPU_CHA_DB_MODE_SEL1 = \t0x%08X\n",
+		ipu_cm_read(ipu, IPU_CHA_DB_MODE_SEL(32)));
+	dev_dbg(ipu->dev, "IPU_FS_PROC_FLOW1 = \t0x%08X\n",
+		ipu_cm_read(ipu, IPU_FS_PROC_FLOW1));
+	dev_dbg(ipu->dev, "IPU_FS_PROC_FLOW2 = \t0x%08X\n",
+		ipu_cm_read(ipu, IPU_FS_PROC_FLOW2));
+	dev_dbg(ipu->dev, "IPU_FS_PROC_FLOW3 = \t0x%08X\n",
+		ipu_cm_read(ipu, IPU_FS_PROC_FLOW3));
+	dev_dbg(ipu->dev, "IPU_FS_DISP_FLOW1 = \t0x%08X\n",
+		ipu_cm_read(ipu, IPU_FS_DISP_FLOW1));
+	for (i = 0; i < 15; i++)
+		dev_dbg(ipu->dev, "IPU_INT_CTRL(%d) = \t%08X\n", i,
+			ipu_cm_read(ipu, IPU_INT_CTRL(i)));
+}
+EXPORT_SYMBOL_GPL(ipu_dump);
+
 static int ipu_probe(struct platform_device *pdev)
 {
 	const struct of_device_id *of_id =
@@ -1243,6 +1227,12 @@ static int ipu_probe(struct platform_device *pdev)
 			ipu_base + devtype->cm_ofs + IPU_CM_IDMAC_REG_OFS);
 	dev_dbg(&pdev->dev, "cpmem:    0x%08lx\n",
 			ipu_base + devtype->cpmem_ofs);
+	dev_dbg(&pdev->dev, "csi0:    0x%08lx\n",
+			ipu_base + devtype->csi0_ofs);
+	dev_dbg(&pdev->dev, "csi1:    0x%08lx\n",
+			ipu_base + devtype->csi1_ofs);
+	dev_dbg(&pdev->dev, "ic:      0x%08lx\n",
+			ipu_base + devtype->ic_ofs);
 	dev_dbg(&pdev->dev, "disp0:    0x%08lx\n",
 			ipu_base + devtype->disp0_ofs);
 	dev_dbg(&pdev->dev, "disp1:    0x%08lx\n",
@@ -1265,10 +1255,8 @@ static int ipu_probe(struct platform_device *pdev)
 	ipu->idmac_reg = devm_ioremap(&pdev->dev,
 			ipu_base + devtype->cm_ofs + IPU_CM_IDMAC_REG_OFS,
 			PAGE_SIZE);
-	ipu->cpmem_base = devm_ioremap(&pdev->dev,
-			ipu_base + devtype->cpmem_ofs, PAGE_SIZE);
 
-	if (!ipu->cm_reg || !ipu->idmac_reg || !ipu->cpmem_base)
+	if (!ipu->cm_reg || !ipu->idmac_reg)
 		return -ENOMEM;
 
 	ipu->clk = devm_clk_get(&pdev->dev, "bus");
diff --git a/drivers/gpu/ipu-v3/ipu-cpmem.c b/drivers/gpu/ipu-v3/ipu-cpmem.c
new file mode 100644
index 000000000000..3bf05bc4ab67
--- /dev/null
+++ b/drivers/gpu/ipu-v3/ipu-cpmem.c
@@ -0,0 +1,764 @@
+/*
+ * Copyright (C) 2012 Mentor Graphics Inc.
+ * Copyright 2005-2012 Freescale Semiconductor, Inc. All Rights Reserved.
+ *
+ * The code contained herein is licensed under the GNU General Public
+ * License. You may obtain a copy of the GNU General Public License
+ * Version 2 or later at the following locations:
+ *
+ * http://www.opensource.org/licenses/gpl-license.html
+ * http://www.gnu.org/copyleft/gpl.html
+ */
+#include <linux/types.h>
+#include <linux/bitrev.h>
+#include <linux/io.h>
+#include <drm/drm_fourcc.h>
+#include "ipu-prv.h"
+
+struct ipu_cpmem_word {
+	u32 data[5];
+	u32 res[3];
+};
+
+struct ipu_ch_param {
+	struct ipu_cpmem_word word[2];
+};
+
+struct ipu_cpmem {
+	struct ipu_ch_param __iomem *base;
+	u32 module;
+	spinlock_t lock;
+	int use_count;
+	struct ipu_soc *ipu;
+};
+
+#define IPU_CPMEM_WORD(word, ofs, size) ((((word) * 160 + (ofs)) << 8) | (size))
+
+#define IPU_FIELD_UBO		IPU_CPMEM_WORD(0, 46, 22)
+#define IPU_FIELD_VBO		IPU_CPMEM_WORD(0, 68, 22)
+#define IPU_FIELD_IOX		IPU_CPMEM_WORD(0, 90, 4)
+#define IPU_FIELD_RDRW		IPU_CPMEM_WORD(0, 94, 1)
+#define IPU_FIELD_SO		IPU_CPMEM_WORD(0, 113, 1)
+#define IPU_FIELD_SLY		IPU_CPMEM_WORD(1, 102, 14)
+#define IPU_FIELD_SLUV		IPU_CPMEM_WORD(1, 128, 14)
+
+#define IPU_FIELD_XV		IPU_CPMEM_WORD(0, 0, 10)
+#define IPU_FIELD_YV		IPU_CPMEM_WORD(0, 10, 9)
+#define IPU_FIELD_XB		IPU_CPMEM_WORD(0, 19, 13)
+#define IPU_FIELD_YB		IPU_CPMEM_WORD(0, 32, 12)
+#define IPU_FIELD_NSB_B		IPU_CPMEM_WORD(0, 44, 1)
+#define IPU_FIELD_CF		IPU_CPMEM_WORD(0, 45, 1)
+#define IPU_FIELD_SX		IPU_CPMEM_WORD(0, 46, 12)
+#define IPU_FIELD_SY		IPU_CPMEM_WORD(0, 58, 11)
+#define IPU_FIELD_NS		IPU_CPMEM_WORD(0, 69, 10)
+#define IPU_FIELD_SDX		IPU_CPMEM_WORD(0, 79, 7)
+#define IPU_FIELD_SM		IPU_CPMEM_WORD(0, 86, 10)
+#define IPU_FIELD_SCC		IPU_CPMEM_WORD(0, 96, 1)
+#define IPU_FIELD_SCE		IPU_CPMEM_WORD(0, 97, 1)
+#define IPU_FIELD_SDY		IPU_CPMEM_WORD(0, 98, 7)
+#define IPU_FIELD_SDRX		IPU_CPMEM_WORD(0, 105, 1)
+#define IPU_FIELD_SDRY		IPU_CPMEM_WORD(0, 106, 1)
+#define IPU_FIELD_BPP		IPU_CPMEM_WORD(0, 107, 3)
+#define IPU_FIELD_DEC_SEL	IPU_CPMEM_WORD(0, 110, 2)
+#define IPU_FIELD_DIM		IPU_CPMEM_WORD(0, 112, 1)
+#define IPU_FIELD_BNDM		IPU_CPMEM_WORD(0, 114, 3)
+#define IPU_FIELD_BM		IPU_CPMEM_WORD(0, 117, 2)
+#define IPU_FIELD_ROT		IPU_CPMEM_WORD(0, 119, 1)
+#define IPU_FIELD_ROT_HF_VF	IPU_CPMEM_WORD(0, 119, 3)
+#define IPU_FIELD_HF		IPU_CPMEM_WORD(0, 120, 1)
+#define IPU_FIELD_VF		IPU_CPMEM_WORD(0, 121, 1)
+#define IPU_FIELD_THE		IPU_CPMEM_WORD(0, 122, 1)
+#define IPU_FIELD_CAP		IPU_CPMEM_WORD(0, 123, 1)
+#define IPU_FIELD_CAE		IPU_CPMEM_WORD(0, 124, 1)
+#define IPU_FIELD_FW		IPU_CPMEM_WORD(0, 125, 13)
+#define IPU_FIELD_FH		IPU_CPMEM_WORD(0, 138, 12)
+#define IPU_FIELD_EBA0		IPU_CPMEM_WORD(1, 0, 29)
+#define IPU_FIELD_EBA1		IPU_CPMEM_WORD(1, 29, 29)
+#define IPU_FIELD_ILO		IPU_CPMEM_WORD(1, 58, 20)
+#define IPU_FIELD_NPB		IPU_CPMEM_WORD(1, 78, 7)
+#define IPU_FIELD_PFS		IPU_CPMEM_WORD(1, 85, 4)
+#define IPU_FIELD_ALU		IPU_CPMEM_WORD(1, 89, 1)
+#define IPU_FIELD_ALBM		IPU_CPMEM_WORD(1, 90, 3)
+#define IPU_FIELD_ID		IPU_CPMEM_WORD(1, 93, 2)
+#define IPU_FIELD_TH		IPU_CPMEM_WORD(1, 95, 7)
+#define IPU_FIELD_SL		IPU_CPMEM_WORD(1, 102, 14)
+#define IPU_FIELD_WID0		IPU_CPMEM_WORD(1, 116, 3)
+#define IPU_FIELD_WID1		IPU_CPMEM_WORD(1, 119, 3)
+#define IPU_FIELD_WID2		IPU_CPMEM_WORD(1, 122, 3)
+#define IPU_FIELD_WID3		IPU_CPMEM_WORD(1, 125, 3)
+#define IPU_FIELD_OFS0		IPU_CPMEM_WORD(1, 128, 5)
+#define IPU_FIELD_OFS1		IPU_CPMEM_WORD(1, 133, 5)
+#define IPU_FIELD_OFS2		IPU_CPMEM_WORD(1, 138, 5)
+#define IPU_FIELD_OFS3		IPU_CPMEM_WORD(1, 143, 5)
+#define IPU_FIELD_SXYS		IPU_CPMEM_WORD(1, 148, 1)
+#define IPU_FIELD_CRE		IPU_CPMEM_WORD(1, 149, 1)
+#define IPU_FIELD_DEC_SEL2	IPU_CPMEM_WORD(1, 150, 1)
+
+static inline struct ipu_ch_param __iomem *
+ipu_get_cpmem(struct ipuv3_channel *ch)
+{
+	struct ipu_cpmem *cpmem = ch->ipu->cpmem_priv;
+
+	return cpmem->base + ch->num;
+}
+
+static void ipu_ch_param_write_field(struct ipuv3_channel *ch, u32 wbs, u32 v)
+{
+	struct ipu_ch_param __iomem *base = ipu_get_cpmem(ch);
+	u32 bit = (wbs >> 8) % 160;
+	u32 size = wbs & 0xff;
+	u32 word = (wbs >> 8) / 160;
+	u32 i = bit / 32;
+	u32 ofs = bit % 32;
+	u32 mask = (1 << size) - 1;
+	u32 val;
+
+	pr_debug("%s %d %d %d\n", __func__, word, bit , size);
+
+	val = readl(&base->word[word].data[i]);
+	val &= ~(mask << ofs);
+	val |= v << ofs;
+	writel(val, &base->word[word].data[i]);
+
+	if ((bit + size - 1) / 32 > i) {
+		val = readl(&base->word[word].data[i + 1]);
+		val &= ~(mask >> (ofs ? (32 - ofs) : 0));
+		val |= v >> (ofs ? (32 - ofs) : 0);
+		writel(val, &base->word[word].data[i + 1]);
+	}
+}
+
+static u32 ipu_ch_param_read_field(struct ipuv3_channel *ch, u32 wbs)
+{
+	struct ipu_ch_param __iomem *base = ipu_get_cpmem(ch);
+	u32 bit = (wbs >> 8) % 160;
+	u32 size = wbs & 0xff;
+	u32 word = (wbs >> 8) / 160;
+	u32 i = bit / 32;
+	u32 ofs = bit % 32;
+	u32 mask = (1 << size) - 1;
+	u32 val = 0;
+
+	pr_debug("%s %d %d %d\n", __func__, word, bit , size);
+
+	val = (readl(&base->word[word].data[i]) >> ofs) & mask;
+
+	if ((bit + size - 1) / 32 > i) {
+		u32 tmp;
+
+		tmp = readl(&base->word[word].data[i + 1]);
+		tmp &= mask >> (ofs ? (32 - ofs) : 0);
+		val |= tmp << (ofs ? (32 - ofs) : 0);
+	}
+
+	return val;
+}
+
+/*
+ * The V4L2 spec defines packed RGB formats in memory byte order, which from
+ * point of view of the IPU corresponds to little-endian words with the first
+ * component in the least significant bits.
+ * The DRM pixel formats and IPU internal representation are ordered the other
+ * way around, with the first named component ordered at the most significant
+ * bits. Further, V4L2 formats are not well defined:
+ *     http://linuxtv.org/downloads/v4l-dvb-apis/packed-rgb.html
+ * We choose the interpretation which matches GStreamer behavior.
+ */
+static int v4l2_pix_fmt_to_drm_fourcc(u32 pixelformat)
+{
+	switch (pixelformat) {
+	case V4L2_PIX_FMT_RGB565:
+		/*
+		 * Here we choose the 'corrected' interpretation of RGBP, a
+		 * little-endian 16-bit word with the red component at the most
+		 * significant bits:
+		 * g[2:0]b[4:0] r[4:0]g[5:3] <=> [16:0] R:G:B
+		 */
+		return DRM_FORMAT_RGB565;
+	case V4L2_PIX_FMT_BGR24:
+		/* B G R <=> [24:0] R:G:B */
+		return DRM_FORMAT_RGB888;
+	case V4L2_PIX_FMT_RGB24:
+		/* R G B <=> [24:0] B:G:R */
+		return DRM_FORMAT_BGR888;
+	case V4L2_PIX_FMT_BGR32:
+		/* B G R A <=> [32:0] A:B:G:R */
+		return DRM_FORMAT_XRGB8888;
+	case V4L2_PIX_FMT_RGB32:
+		/* R G B A <=> [32:0] A:B:G:R */
+		return DRM_FORMAT_XBGR8888;
+	case V4L2_PIX_FMT_UYVY:
+		return DRM_FORMAT_UYVY;
+	case V4L2_PIX_FMT_YUYV:
+		return DRM_FORMAT_YUYV;
+	case V4L2_PIX_FMT_YUV420:
+		return DRM_FORMAT_YUV420;
+	case V4L2_PIX_FMT_YUV422P:
+		return DRM_FORMAT_YUV422;
+	case V4L2_PIX_FMT_YVU420:
+		return DRM_FORMAT_YVU420;
+	case V4L2_PIX_FMT_NV12:
+		return DRM_FORMAT_NV12;
+	case V4L2_PIX_FMT_NV16:
+		return DRM_FORMAT_NV16;
+	}
+
+	return -EINVAL;
+}
+
+void ipu_cpmem_zero(struct ipuv3_channel *ch)
+{
+	struct ipu_ch_param __iomem *p = ipu_get_cpmem(ch);
+	void __iomem *base = p;
+	int i;
+
+	for (i = 0; i < sizeof(*p) / sizeof(u32); i++)
+		writel(0, base + i * sizeof(u32));
+}
+EXPORT_SYMBOL_GPL(ipu_cpmem_zero);
+
+void ipu_cpmem_set_resolution(struct ipuv3_channel *ch, int xres, int yres)
+{
+	ipu_ch_param_write_field(ch, IPU_FIELD_FW, xres - 1);
+	ipu_ch_param_write_field(ch, IPU_FIELD_FH, yres - 1);
+}
+EXPORT_SYMBOL_GPL(ipu_cpmem_set_resolution);
+
+void ipu_cpmem_set_stride(struct ipuv3_channel *ch, int stride)
+{
+	ipu_ch_param_write_field(ch, IPU_FIELD_SLY, stride - 1);
+}
+EXPORT_SYMBOL_GPL(ipu_cpmem_set_stride);
+
+void ipu_cpmem_set_high_priority(struct ipuv3_channel *ch)
+{
+	struct ipu_soc *ipu = ch->ipu;
+	u32 val;
+
+	if (ipu->ipu_type == IPUV3EX)
+		ipu_ch_param_write_field(ch, IPU_FIELD_ID, 1);
+
+	val = ipu_idmac_read(ipu, IDMAC_CHA_PRI(ch->num));
+	val |= 1 << (ch->num % 32);
+	ipu_idmac_write(ipu, val, IDMAC_CHA_PRI(ch->num));
+};
+EXPORT_SYMBOL_GPL(ipu_cpmem_set_high_priority);
+
+void ipu_cpmem_set_buffer(struct ipuv3_channel *ch, int bufnum, dma_addr_t buf)
+{
+	if (bufnum)
+		ipu_ch_param_write_field(ch, IPU_FIELD_EBA1, buf >> 3);
+	else
+		ipu_ch_param_write_field(ch, IPU_FIELD_EBA0, buf >> 3);
+}
+EXPORT_SYMBOL_GPL(ipu_cpmem_set_buffer);
+
+void ipu_cpmem_interlaced_scan(struct ipuv3_channel *ch, int stride)
+{
+	ipu_ch_param_write_field(ch, IPU_FIELD_SO, 1);
+	ipu_ch_param_write_field(ch, IPU_FIELD_ILO, stride / 8);
+	ipu_ch_param_write_field(ch, IPU_FIELD_SLY, (stride * 2) - 1);
+};
+EXPORT_SYMBOL_GPL(ipu_cpmem_interlaced_scan);
+
+void ipu_cpmem_set_axi_id(struct ipuv3_channel *ch, u32 id)
+{
+	id &= 0x3;
+	ipu_ch_param_write_field(ch, IPU_FIELD_ID, id);
+}
+EXPORT_SYMBOL_GPL(ipu_cpmem_set_axi_id);
+
+void ipu_cpmem_set_burstsize(struct ipuv3_channel *ch, int burstsize)
+{
+	ipu_ch_param_write_field(ch, IPU_FIELD_NPB, burstsize - 1);
+};
+EXPORT_SYMBOL_GPL(ipu_cpmem_set_burstsize);
+
+void ipu_cpmem_set_block_mode(struct ipuv3_channel *ch)
+{
+	ipu_ch_param_write_field(ch, IPU_FIELD_BM, 1);
+}
+EXPORT_SYMBOL_GPL(ipu_cpmem_set_block_mode);
+
+void ipu_cpmem_set_rotation(struct ipuv3_channel *ch,
+			    enum ipu_rotate_mode rot)
+{
+	u32 temp_rot = bitrev8(rot) >> 5;
+
+	ipu_ch_param_write_field(ch, IPU_FIELD_ROT_HF_VF, temp_rot);
+}
+EXPORT_SYMBOL_GPL(ipu_cpmem_set_rotation);
+
+int ipu_cpmem_set_format_rgb(struct ipuv3_channel *ch,
+			     const struct ipu_rgb *rgb)
+{
+	int bpp = 0, npb = 0, ro, go, bo, to;
+
+	ro = rgb->bits_per_pixel - rgb->red.length - rgb->red.offset;
+	go = rgb->bits_per_pixel - rgb->green.length - rgb->green.offset;
+	bo = rgb->bits_per_pixel - rgb->blue.length - rgb->blue.offset;
+	to = rgb->bits_per_pixel - rgb->transp.length - rgb->transp.offset;
+
+	ipu_ch_param_write_field(ch, IPU_FIELD_WID0, rgb->red.length - 1);
+	ipu_ch_param_write_field(ch, IPU_FIELD_OFS0, ro);
+	ipu_ch_param_write_field(ch, IPU_FIELD_WID1, rgb->green.length - 1);
+	ipu_ch_param_write_field(ch, IPU_FIELD_OFS1, go);
+	ipu_ch_param_write_field(ch, IPU_FIELD_WID2, rgb->blue.length - 1);
+	ipu_ch_param_write_field(ch, IPU_FIELD_OFS2, bo);
+
+	if (rgb->transp.length) {
+		ipu_ch_param_write_field(ch, IPU_FIELD_WID3,
+				rgb->transp.length - 1);
+		ipu_ch_param_write_field(ch, IPU_FIELD_OFS3, to);
+	} else {
+		ipu_ch_param_write_field(ch, IPU_FIELD_WID3, 7);
+		ipu_ch_param_write_field(ch, IPU_FIELD_OFS3,
+				rgb->bits_per_pixel);
+	}
+
+	switch (rgb->bits_per_pixel) {
+	case 32:
+		bpp = 0;
+		npb = 15;
+		break;
+	case 24:
+		bpp = 1;
+		npb = 19;
+		break;
+	case 16:
+		bpp = 3;
+		npb = 31;
+		break;
+	case 8:
+		bpp = 5;
+		npb = 63;
+		break;
+	default:
+		return -EINVAL;
+	}
+	ipu_ch_param_write_field(ch, IPU_FIELD_BPP, bpp);
+	ipu_ch_param_write_field(ch, IPU_FIELD_NPB, npb);
+	ipu_ch_param_write_field(ch, IPU_FIELD_PFS, 7); /* rgb mode */
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(ipu_cpmem_set_format_rgb);
+
+int ipu_cpmem_set_format_passthrough(struct ipuv3_channel *ch, int width)
+{
+	int bpp = 0, npb = 0;
+
+	switch (width) {
+	case 32:
+		bpp = 0;
+		npb = 15;
+		break;
+	case 24:
+		bpp = 1;
+		npb = 19;
+		break;
+	case 16:
+		bpp = 3;
+		npb = 31;
+		break;
+	case 8:
+		bpp = 5;
+		npb = 63;
+		break;
+	default:
+		return -EINVAL;
+	}
+
+	ipu_ch_param_write_field(ch, IPU_FIELD_BPP, bpp);
+	ipu_ch_param_write_field(ch, IPU_FIELD_NPB, npb);
+	ipu_ch_param_write_field(ch, IPU_FIELD_PFS, 6); /* raw mode */
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(ipu_cpmem_set_format_passthrough);
+
+void ipu_cpmem_set_yuv_interleaved(struct ipuv3_channel *ch, u32 pixel_format)
+{
+	switch (pixel_format) {
+	case V4L2_PIX_FMT_UYVY:
+		ipu_ch_param_write_field(ch, IPU_FIELD_BPP, 3); /* bits/pixel */
+		ipu_ch_param_write_field(ch, IPU_FIELD_PFS, 0xA);/* pix fmt */
+		ipu_ch_param_write_field(ch, IPU_FIELD_NPB, 31);/* burst size */
+		break;
+	case V4L2_PIX_FMT_YUYV:
+		ipu_ch_param_write_field(ch, IPU_FIELD_BPP, 3); /* bits/pixel */
+		ipu_ch_param_write_field(ch, IPU_FIELD_PFS, 0x8);/* pix fmt */
+		ipu_ch_param_write_field(ch, IPU_FIELD_NPB, 31);/* burst size */
+		break;
+	}
+}
+EXPORT_SYMBOL_GPL(ipu_cpmem_set_yuv_interleaved);
+
+void ipu_cpmem_set_yuv_planar_full(struct ipuv3_channel *ch,
+				   u32 pixel_format, int stride,
+				   int u_offset, int v_offset)
+{
+	switch (pixel_format) {
+	case V4L2_PIX_FMT_YUV420:
+	case V4L2_PIX_FMT_YUV422P:
+		ipu_ch_param_write_field(ch, IPU_FIELD_SLUV, (stride / 2) - 1);
+		ipu_ch_param_write_field(ch, IPU_FIELD_UBO, u_offset / 8);
+		ipu_ch_param_write_field(ch, IPU_FIELD_VBO, v_offset / 8);
+		break;
+	case V4L2_PIX_FMT_YVU420:
+		ipu_ch_param_write_field(ch, IPU_FIELD_SLUV, (stride / 2) - 1);
+		ipu_ch_param_write_field(ch, IPU_FIELD_UBO, v_offset / 8);
+		ipu_ch_param_write_field(ch, IPU_FIELD_VBO, u_offset / 8);
+		break;
+	case V4L2_PIX_FMT_NV12:
+	case V4L2_PIX_FMT_NV16:
+		ipu_ch_param_write_field(ch, IPU_FIELD_SLUV, stride - 1);
+		ipu_ch_param_write_field(ch, IPU_FIELD_UBO, u_offset / 8);
+		ipu_ch_param_write_field(ch, IPU_FIELD_VBO, u_offset / 8);
+		break;
+	}
+}
+EXPORT_SYMBOL_GPL(ipu_cpmem_set_yuv_planar_full);
+
+void ipu_cpmem_set_yuv_planar(struct ipuv3_channel *ch,
+			      u32 pixel_format, int stride, int height)
+{
+	int u_offset, v_offset;
+	int uv_stride = 0;
+
+	switch (pixel_format) {
+	case V4L2_PIX_FMT_YUV420:
+	case V4L2_PIX_FMT_YVU420:
+		uv_stride = stride / 2;
+		u_offset = stride * height;
+		v_offset = u_offset + (uv_stride * height / 2);
+		ipu_cpmem_set_yuv_planar_full(ch, pixel_format, stride,
+					      u_offset, v_offset);
+		break;
+	case V4L2_PIX_FMT_YUV422P:
+		uv_stride = stride / 2;
+		u_offset = stride * height;
+		v_offset = u_offset + (uv_stride * height);
+		ipu_cpmem_set_yuv_planar_full(ch, pixel_format, stride,
+					      u_offset, v_offset);
+		break;
+	case V4L2_PIX_FMT_NV12:
+	case V4L2_PIX_FMT_NV16:
+		u_offset = stride * height;
+		ipu_cpmem_set_yuv_planar_full(ch, pixel_format, stride,
+					      u_offset, 0);
+		break;
+	}
+}
+EXPORT_SYMBOL_GPL(ipu_cpmem_set_yuv_planar);
+
+static const struct ipu_rgb def_rgb_32 = {
+	.red	= { .offset = 16, .length = 8, },
+	.green	= { .offset =  8, .length = 8, },
+	.blue	= { .offset =  0, .length = 8, },
+	.transp = { .offset = 24, .length = 8, },
+	.bits_per_pixel = 32,
+};
+
+static const struct ipu_rgb def_bgr_32 = {
+	.red	= { .offset =  0, .length = 8, },
+	.green	= { .offset =  8, .length = 8, },
+	.blue	= { .offset = 16, .length = 8, },
+	.transp = { .offset = 24, .length = 8, },
+	.bits_per_pixel = 32,
+};
+
+static const struct ipu_rgb def_rgb_24 = {
+	.red	= { .offset = 16, .length = 8, },
+	.green	= { .offset =  8, .length = 8, },
+	.blue	= { .offset =  0, .length = 8, },
+	.transp = { .offset =  0, .length = 0, },
+	.bits_per_pixel = 24,
+};
+
+static const struct ipu_rgb def_bgr_24 = {
+	.red	= { .offset =  0, .length = 8, },
+	.green	= { .offset =  8, .length = 8, },
+	.blue	= { .offset = 16, .length = 8, },
+	.transp = { .offset =  0, .length = 0, },
+	.bits_per_pixel = 24,
+};
+
+static const struct ipu_rgb def_rgb_16 = {
+	.red	= { .offset = 11, .length = 5, },
+	.green	= { .offset =  5, .length = 6, },
+	.blue	= { .offset =  0, .length = 5, },
+	.transp = { .offset =  0, .length = 0, },
+	.bits_per_pixel = 16,
+};
+
+static const struct ipu_rgb def_bgr_16 = {
+	.red	= { .offset =  0, .length = 5, },
+	.green	= { .offset =  5, .length = 6, },
+	.blue	= { .offset = 11, .length = 5, },
+	.transp = { .offset =  0, .length = 0, },
+	.bits_per_pixel = 16,
+};
+
+#define Y_OFFSET(pix, x, y)	((x) + pix->width * (y))
+#define U_OFFSET(pix, x, y)	((pix->width * pix->height) +		\
+				 (pix->width * (y) / 4) + (x) / 2)
+#define V_OFFSET(pix, x, y)	((pix->width * pix->height) +		\
+				 (pix->width * pix->height / 4) +	\
+				 (pix->width * (y) / 4) + (x) / 2)
+#define U2_OFFSET(pix, x, y)	((pix->width * pix->height) +		\
+				 (pix->width * (y) / 2) + (x) / 2)
+#define V2_OFFSET(pix, x, y)	((pix->width * pix->height) +		\
+				 (pix->width * pix->height / 2) +	\
+				 (pix->width * (y) / 2) + (x) / 2)
+#define UV_OFFSET(pix, x, y)	((pix->width * pix->height) +	\
+				 (pix->width * (y) / 2) + (x))
+#define UV2_OFFSET(pix, x, y)	((pix->width * pix->height) +	\
+				 (pix->width * y) + (x))
+
+int ipu_cpmem_set_fmt(struct ipuv3_channel *ch, u32 drm_fourcc)
+{
+	switch (drm_fourcc) {
+	case DRM_FORMAT_YUV420:
+	case DRM_FORMAT_YVU420:
+		/* pix format */
+		ipu_ch_param_write_field(ch, IPU_FIELD_PFS, 2);
+		/* burst size */
+		ipu_ch_param_write_field(ch, IPU_FIELD_NPB, 31);
+		break;
+	case DRM_FORMAT_YUV422:
+	case DRM_FORMAT_YVU422:
+		/* pix format */
+		ipu_ch_param_write_field(ch, IPU_FIELD_PFS, 1);
+		/* burst size */
+		ipu_ch_param_write_field(ch, IPU_FIELD_NPB, 31);
+		break;
+	case DRM_FORMAT_NV12:
+		/* pix format */
+		ipu_ch_param_write_field(ch, IPU_FIELD_PFS, 4);
+		/* burst size */
+		ipu_ch_param_write_field(ch, IPU_FIELD_NPB, 31);
+		break;
+	case DRM_FORMAT_NV16:
+		/* pix format */
+		ipu_ch_param_write_field(ch, IPU_FIELD_PFS, 3);
+		/* burst size */
+		ipu_ch_param_write_field(ch, IPU_FIELD_NPB, 31);
+		break;
+	case DRM_FORMAT_UYVY:
+		/* bits/pixel */
+		ipu_ch_param_write_field(ch, IPU_FIELD_BPP, 3);
+		/* pix format */
+		ipu_ch_param_write_field(ch, IPU_FIELD_PFS, 0xA);
+		/* burst size */
+		ipu_ch_param_write_field(ch, IPU_FIELD_NPB, 31);
+		break;
+	case DRM_FORMAT_YUYV:
+		/* bits/pixel */
+		ipu_ch_param_write_field(ch, IPU_FIELD_BPP, 3);
+		/* pix format */
+		ipu_ch_param_write_field(ch, IPU_FIELD_PFS, 0x8);
+		/* burst size */
+		ipu_ch_param_write_field(ch, IPU_FIELD_NPB, 31);
+		break;
+	case DRM_FORMAT_ABGR8888:
+	case DRM_FORMAT_XBGR8888:
+		ipu_cpmem_set_format_rgb(ch, &def_bgr_32);
+		break;
+	case DRM_FORMAT_ARGB8888:
+	case DRM_FORMAT_XRGB8888:
+		ipu_cpmem_set_format_rgb(ch, &def_rgb_32);
+		break;
+	case DRM_FORMAT_BGR888:
+		ipu_cpmem_set_format_rgb(ch, &def_bgr_24);
+		break;
+	case DRM_FORMAT_RGB888:
+		ipu_cpmem_set_format_rgb(ch, &def_rgb_24);
+		break;
+	case DRM_FORMAT_RGB565:
+		ipu_cpmem_set_format_rgb(ch, &def_rgb_16);
+		break;
+	case DRM_FORMAT_BGR565:
+		ipu_cpmem_set_format_rgb(ch, &def_bgr_16);
+		break;
+	default:
+		return -EINVAL;
+	}
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(ipu_cpmem_set_fmt);
+
+int ipu_cpmem_set_image(struct ipuv3_channel *ch, struct ipu_image *image)
+{
+	struct v4l2_pix_format *pix = &image->pix;
+	int offset, u_offset, v_offset;
+
+	pr_debug("%s: resolution: %dx%d stride: %d\n",
+		 __func__, pix->width, pix->height,
+		 pix->bytesperline);
+
+	ipu_cpmem_set_resolution(ch, image->rect.width, image->rect.height);
+	ipu_cpmem_set_stride(ch, pix->bytesperline);
+
+	ipu_cpmem_set_fmt(ch, v4l2_pix_fmt_to_drm_fourcc(pix->pixelformat));
+
+	switch (pix->pixelformat) {
+	case V4L2_PIX_FMT_YUV420:
+	case V4L2_PIX_FMT_YVU420:
+		offset = Y_OFFSET(pix, image->rect.left, image->rect.top);
+		u_offset = U_OFFSET(pix, image->rect.left,
+				    image->rect.top) - offset;
+		v_offset = V_OFFSET(pix, image->rect.left,
+				    image->rect.top) - offset;
+
+		ipu_cpmem_set_yuv_planar_full(ch, pix->pixelformat,
+					      pix->bytesperline,
+					      u_offset, v_offset);
+		break;
+	case V4L2_PIX_FMT_YUV422P:
+		offset = Y_OFFSET(pix, image->rect.left, image->rect.top);
+		u_offset = U2_OFFSET(pix, image->rect.left,
+				     image->rect.top) - offset;
+		v_offset = V2_OFFSET(pix, image->rect.left,
+				     image->rect.top) - offset;
+
+		ipu_cpmem_set_yuv_planar_full(ch, pix->pixelformat,
+					      pix->bytesperline,
+					      u_offset, v_offset);
+		break;
+	case V4L2_PIX_FMT_NV12:
+		offset = Y_OFFSET(pix, image->rect.left, image->rect.top);
+		u_offset = UV_OFFSET(pix, image->rect.left,
+				     image->rect.top) - offset;
+		v_offset = 0;
+
+		ipu_cpmem_set_yuv_planar_full(ch, pix->pixelformat,
+					      pix->bytesperline,
+					      u_offset, v_offset);
+		break;
+	case V4L2_PIX_FMT_NV16:
+		offset = Y_OFFSET(pix, image->rect.left, image->rect.top);
+		u_offset = UV2_OFFSET(pix, image->rect.left,
+				      image->rect.top) - offset;
+		v_offset = 0;
+
+		ipu_cpmem_set_yuv_planar_full(ch, pix->pixelformat,
+					      pix->bytesperline,
+					      u_offset, v_offset);
+		break;
+	case V4L2_PIX_FMT_UYVY:
+	case V4L2_PIX_FMT_YUYV:
+	case V4L2_PIX_FMT_RGB565:
+		offset = image->rect.left * 2 +
+			image->rect.top * pix->bytesperline;
+		break;
+	case V4L2_PIX_FMT_RGB32:
+	case V4L2_PIX_FMT_BGR32:
+		offset = image->rect.left * 4 +
+			image->rect.top * pix->bytesperline;
+		break;
+	case V4L2_PIX_FMT_RGB24:
+	case V4L2_PIX_FMT_BGR24:
+		offset = image->rect.left * 3 +
+			image->rect.top * pix->bytesperline;
+		break;
+	default:
+		return -EINVAL;
+	}
+
+	ipu_cpmem_set_buffer(ch, 0, image->phys0 + offset);
+	ipu_cpmem_set_buffer(ch, 1, image->phys1 + offset);
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(ipu_cpmem_set_image);
+
+void ipu_cpmem_dump(struct ipuv3_channel *ch)
+{
+	struct ipu_ch_param __iomem *p = ipu_get_cpmem(ch);
+	struct ipu_soc *ipu = ch->ipu;
+	int chno = ch->num;
+
+	dev_dbg(ipu->dev, "ch %d word 0 - %08X %08X %08X %08X %08X\n", chno,
+		readl(&p->word[0].data[0]),
+		readl(&p->word[0].data[1]),
+		readl(&p->word[0].data[2]),
+		readl(&p->word[0].data[3]),
+		readl(&p->word[0].data[4]));
+	dev_dbg(ipu->dev, "ch %d word 1 - %08X %08X %08X %08X %08X\n", chno,
+		readl(&p->word[1].data[0]),
+		readl(&p->word[1].data[1]),
+		readl(&p->word[1].data[2]),
+		readl(&p->word[1].data[3]),
+		readl(&p->word[1].data[4]));
+	dev_dbg(ipu->dev, "PFS 0x%x, ",
+		 ipu_ch_param_read_field(ch, IPU_FIELD_PFS));
+	dev_dbg(ipu->dev, "BPP 0x%x, ",
+		ipu_ch_param_read_field(ch, IPU_FIELD_BPP));
+	dev_dbg(ipu->dev, "NPB 0x%x\n",
+		 ipu_ch_param_read_field(ch, IPU_FIELD_NPB));
+
+	dev_dbg(ipu->dev, "FW %d, ",
+		 ipu_ch_param_read_field(ch, IPU_FIELD_FW));
+	dev_dbg(ipu->dev, "FH %d, ",
+		 ipu_ch_param_read_field(ch, IPU_FIELD_FH));
+	dev_dbg(ipu->dev, "EBA0 0x%x\n",
+		 ipu_ch_param_read_field(ch, IPU_FIELD_EBA0) << 3);
+	dev_dbg(ipu->dev, "EBA1 0x%x\n",
+		 ipu_ch_param_read_field(ch, IPU_FIELD_EBA1) << 3);
+	dev_dbg(ipu->dev, "Stride %d\n",
+		 ipu_ch_param_read_field(ch, IPU_FIELD_SL));
+	dev_dbg(ipu->dev, "scan_order %d\n",
+		 ipu_ch_param_read_field(ch, IPU_FIELD_SO));
+	dev_dbg(ipu->dev, "uv_stride %d\n",
+		 ipu_ch_param_read_field(ch, IPU_FIELD_SLUV));
+	dev_dbg(ipu->dev, "u_offset 0x%x\n",
+		 ipu_ch_param_read_field(ch, IPU_FIELD_UBO) << 3);
+	dev_dbg(ipu->dev, "v_offset 0x%x\n",
+		 ipu_ch_param_read_field(ch, IPU_FIELD_VBO) << 3);
+
+	dev_dbg(ipu->dev, "Width0 %d+1, ",
+		 ipu_ch_param_read_field(ch, IPU_FIELD_WID0));
+	dev_dbg(ipu->dev, "Width1 %d+1, ",
+		 ipu_ch_param_read_field(ch, IPU_FIELD_WID1));
+	dev_dbg(ipu->dev, "Width2 %d+1, ",
+		 ipu_ch_param_read_field(ch, IPU_FIELD_WID2));
+	dev_dbg(ipu->dev, "Width3 %d+1, ",
+		 ipu_ch_param_read_field(ch, IPU_FIELD_WID3));
+	dev_dbg(ipu->dev, "Offset0 %d, ",
+		 ipu_ch_param_read_field(ch, IPU_FIELD_OFS0));
+	dev_dbg(ipu->dev, "Offset1 %d, ",
+		 ipu_ch_param_read_field(ch, IPU_FIELD_OFS1));
+	dev_dbg(ipu->dev, "Offset2 %d, ",
+		 ipu_ch_param_read_field(ch, IPU_FIELD_OFS2));
+	dev_dbg(ipu->dev, "Offset3 %d\n",
+		 ipu_ch_param_read_field(ch, IPU_FIELD_OFS3));
+}
+EXPORT_SYMBOL_GPL(ipu_cpmem_dump);
+
+int ipu_cpmem_init(struct ipu_soc *ipu, struct device *dev, unsigned long base)
+{
+	struct ipu_cpmem *cpmem;
+
+	cpmem = devm_kzalloc(dev, sizeof(*cpmem), GFP_KERNEL);
+	if (!cpmem)
+		return -ENOMEM;
+
+	ipu->cpmem_priv = cpmem;
+
+	spin_lock_init(&cpmem->lock);
+	cpmem->base = devm_ioremap(dev, base, SZ_128K);
+	if (!cpmem->base)
+		return -ENOMEM;
+
+	dev_dbg(dev, "CPMEM base: 0x%08lx remapped to %p\n",
+		base, cpmem->base);
+	cpmem->ipu = ipu;
+
+	return 0;
+}
+
+void ipu_cpmem_exit(struct ipu_soc *ipu)
+{
+}
diff --git a/drivers/gpu/ipu-v3/ipu-csi.c b/drivers/gpu/ipu-v3/ipu-csi.c
new file mode 100644
index 000000000000..752cdd2da89a
--- /dev/null
+++ b/drivers/gpu/ipu-v3/ipu-csi.c
@@ -0,0 +1,741 @@
+/*
+ * Copyright (C) 2012-2014 Mentor Graphics Inc.
+ * Copyright (C) 2005-2009 Freescale Semiconductor, Inc.
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License as published by the
+ * Free Software Foundation; either version 2 of the License, or (at your
+ * option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful, but
+ * WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY
+ * or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License
+ * for more details.
+ */
+#include <linux/export.h>
+#include <linux/module.h>
+#include <linux/types.h>
+#include <linux/errno.h>
+#include <linux/delay.h>
+#include <linux/io.h>
+#include <linux/err.h>
+#include <linux/platform_device.h>
+#include <linux/videodev2.h>
+#include <uapi/linux/v4l2-mediabus.h>
+#include <linux/clk.h>
+#include <linux/clk-provider.h>
+#include <linux/clkdev.h>
+
+#include "ipu-prv.h"
+
+struct ipu_csi {
+	void __iomem *base;
+	int id;
+	u32 module;
+	struct clk *clk_ipu;	/* IPU bus clock */
+	spinlock_t lock;
+	bool inuse;
+	struct ipu_soc *ipu;
+};
+
+/* CSI Register Offsets */
+#define CSI_SENS_CONF		0x0000
+#define CSI_SENS_FRM_SIZE	0x0004
+#define CSI_ACT_FRM_SIZE	0x0008
+#define CSI_OUT_FRM_CTRL	0x000c
+#define CSI_TST_CTRL		0x0010
+#define CSI_CCIR_CODE_1		0x0014
+#define CSI_CCIR_CODE_2		0x0018
+#define CSI_CCIR_CODE_3		0x001c
+#define CSI_MIPI_DI		0x0020
+#define CSI_SKIP		0x0024
+#define CSI_CPD_CTRL		0x0028
+#define CSI_CPD_RC(n)		(0x002c + ((n)*4))
+#define CSI_CPD_RS(n)		(0x004c + ((n)*4))
+#define CSI_CPD_GRC(n)		(0x005c + ((n)*4))
+#define CSI_CPD_GRS(n)		(0x007c + ((n)*4))
+#define CSI_CPD_GBC(n)		(0x008c + ((n)*4))
+#define CSI_CPD_GBS(n)		(0x00Ac + ((n)*4))
+#define CSI_CPD_BC(n)		(0x00Bc + ((n)*4))
+#define CSI_CPD_BS(n)		(0x00Dc + ((n)*4))
+#define CSI_CPD_OFFSET1		0x00ec
+#define CSI_CPD_OFFSET2		0x00f0
+
+/* CSI Register Fields */
+#define CSI_SENS_CONF_DATA_FMT_SHIFT		8
+#define CSI_SENS_CONF_DATA_FMT_MASK		0x00000700
+#define CSI_SENS_CONF_DATA_FMT_RGB_YUV444	0L
+#define CSI_SENS_CONF_DATA_FMT_YUV422_YUYV	1L
+#define CSI_SENS_CONF_DATA_FMT_YUV422_UYVY	2L
+#define CSI_SENS_CONF_DATA_FMT_BAYER		3L
+#define CSI_SENS_CONF_DATA_FMT_RGB565		4L
+#define CSI_SENS_CONF_DATA_FMT_RGB555		5L
+#define CSI_SENS_CONF_DATA_FMT_RGB444		6L
+#define CSI_SENS_CONF_DATA_FMT_JPEG		7L
+
+#define CSI_SENS_CONF_VSYNC_POL_SHIFT		0
+#define CSI_SENS_CONF_HSYNC_POL_SHIFT		1
+#define CSI_SENS_CONF_DATA_POL_SHIFT		2
+#define CSI_SENS_CONF_PIX_CLK_POL_SHIFT		3
+#define CSI_SENS_CONF_SENS_PRTCL_MASK		0x00000070
+#define CSI_SENS_CONF_SENS_PRTCL_SHIFT		4
+#define CSI_SENS_CONF_PACK_TIGHT_SHIFT		7
+#define CSI_SENS_CONF_DATA_WIDTH_SHIFT		11
+#define CSI_SENS_CONF_EXT_VSYNC_SHIFT		15
+#define CSI_SENS_CONF_DIVRATIO_SHIFT		16
+
+#define CSI_SENS_CONF_DIVRATIO_MASK		0x00ff0000
+#define CSI_SENS_CONF_DATA_DEST_SHIFT		24
+#define CSI_SENS_CONF_DATA_DEST_MASK		0x07000000
+#define CSI_SENS_CONF_JPEG8_EN_SHIFT		27
+#define CSI_SENS_CONF_JPEG_EN_SHIFT		28
+#define CSI_SENS_CONF_FORCE_EOF_SHIFT		29
+#define CSI_SENS_CONF_DATA_EN_POL_SHIFT		31
+
+#define CSI_DATA_DEST_IC			2
+#define CSI_DATA_DEST_IDMAC			4
+
+#define CSI_CCIR_ERR_DET_EN			0x01000000
+#define CSI_HORI_DOWNSIZE_EN			0x80000000
+#define CSI_VERT_DOWNSIZE_EN			0x40000000
+#define CSI_TEST_GEN_MODE_EN			0x01000000
+
+#define CSI_HSC_MASK				0x1fff0000
+#define CSI_HSC_SHIFT				16
+#define CSI_VSC_MASK				0x00000fff
+#define CSI_VSC_SHIFT				0
+
+#define CSI_TEST_GEN_R_MASK			0x000000ff
+#define CSI_TEST_GEN_R_SHIFT			0
+#define CSI_TEST_GEN_G_MASK			0x0000ff00
+#define CSI_TEST_GEN_G_SHIFT			8
+#define CSI_TEST_GEN_B_MASK			0x00ff0000
+#define CSI_TEST_GEN_B_SHIFT			16
+
+#define CSI_MAX_RATIO_SKIP_SMFC_MASK		0x00000007
+#define CSI_MAX_RATIO_SKIP_SMFC_SHIFT		0
+#define CSI_SKIP_SMFC_MASK			0x000000f8
+#define CSI_SKIP_SMFC_SHIFT			3
+#define CSI_ID_2_SKIP_MASK			0x00000300
+#define CSI_ID_2_SKIP_SHIFT			8
+
+#define CSI_COLOR_FIRST_ROW_MASK		0x00000002
+#define CSI_COLOR_FIRST_COMP_MASK		0x00000001
+
+/* MIPI CSI-2 data types */
+#define MIPI_DT_YUV420		0x18 /* YYY.../UYVY.... */
+#define MIPI_DT_YUV420_LEGACY	0x1a /* UYY.../VYY...   */
+#define MIPI_DT_YUV422		0x1e /* UYVY...         */
+#define MIPI_DT_RGB444		0x20
+#define MIPI_DT_RGB555		0x21
+#define MIPI_DT_RGB565		0x22
+#define MIPI_DT_RGB666		0x23
+#define MIPI_DT_RGB888		0x24
+#define MIPI_DT_RAW6		0x28
+#define MIPI_DT_RAW7		0x29
+#define MIPI_DT_RAW8		0x2a
+#define MIPI_DT_RAW10		0x2b
+#define MIPI_DT_RAW12		0x2c
+#define MIPI_DT_RAW14		0x2d
+
+/*
+ * Bitfield of CSI bus signal polarities and modes.
+ */
+struct ipu_csi_bus_config {
+	unsigned data_width:4;
+	unsigned clk_mode:3;
+	unsigned ext_vsync:1;
+	unsigned vsync_pol:1;
+	unsigned hsync_pol:1;
+	unsigned pixclk_pol:1;
+	unsigned data_pol:1;
+	unsigned sens_clksrc:1;
+	unsigned pack_tight:1;
+	unsigned force_eof:1;
+	unsigned data_en_pol:1;
+
+	unsigned data_fmt;
+	unsigned mipi_dt;
+};
+
+/*
+ * Enumeration of CSI data bus widths.
+ */
+enum ipu_csi_data_width {
+	IPU_CSI_DATA_WIDTH_4   = 0,
+	IPU_CSI_DATA_WIDTH_8   = 1,
+	IPU_CSI_DATA_WIDTH_10  = 3,
+	IPU_CSI_DATA_WIDTH_12  = 5,
+	IPU_CSI_DATA_WIDTH_16  = 9,
+};
+
+/*
+ * Enumeration of CSI clock modes.
+ */
+enum ipu_csi_clk_mode {
+	IPU_CSI_CLK_MODE_GATED_CLK,
+	IPU_CSI_CLK_MODE_NONGATED_CLK,
+	IPU_CSI_CLK_MODE_CCIR656_PROGRESSIVE,
+	IPU_CSI_CLK_MODE_CCIR656_INTERLACED,
+	IPU_CSI_CLK_MODE_CCIR1120_PROGRESSIVE_DDR,
+	IPU_CSI_CLK_MODE_CCIR1120_PROGRESSIVE_SDR,
+	IPU_CSI_CLK_MODE_CCIR1120_INTERLACED_DDR,
+	IPU_CSI_CLK_MODE_CCIR1120_INTERLACED_SDR,
+};
+
+static inline u32 ipu_csi_read(struct ipu_csi *csi, unsigned offset)
+{
+	return readl(csi->base + offset);
+}
+
+static inline void ipu_csi_write(struct ipu_csi *csi, u32 value,
+				 unsigned offset)
+{
+	writel(value, csi->base + offset);
+}
+
+/*
+ * Set mclk division ratio for generating test mode mclk. Only used
+ * for test generator.
+ */
+static int ipu_csi_set_testgen_mclk(struct ipu_csi *csi, u32 pixel_clk,
+					u32 ipu_clk)
+{
+	u32 temp;
+	u32 div_ratio;
+
+	div_ratio = (ipu_clk / pixel_clk) - 1;
+
+	if (div_ratio > 0xFF || div_ratio < 0) {
+		dev_err(csi->ipu->dev,
+			"value of pixel_clk extends normal range\n");
+		return -EINVAL;
+	}
+
+	temp = ipu_csi_read(csi, CSI_SENS_CONF);
+	temp &= ~CSI_SENS_CONF_DIVRATIO_MASK;
+	ipu_csi_write(csi, temp | (div_ratio << CSI_SENS_CONF_DIVRATIO_SHIFT),
+			  CSI_SENS_CONF);
+
+	return 0;
+}
+
+/*
+ * Find the CSI data format and data width for the given V4L2 media
+ * bus pixel format code.
+ */
+static int mbus_code_to_bus_cfg(struct ipu_csi_bus_config *cfg, u32 mbus_code)
+{
+	switch (mbus_code) {
+	case MEDIA_BUS_FMT_BGR565_2X8_BE:
+	case MEDIA_BUS_FMT_BGR565_2X8_LE:
+	case MEDIA_BUS_FMT_RGB565_2X8_BE:
+	case MEDIA_BUS_FMT_RGB565_2X8_LE:
+		cfg->data_fmt = CSI_SENS_CONF_DATA_FMT_RGB565;
+		cfg->mipi_dt = MIPI_DT_RGB565;
+		cfg->data_width = IPU_CSI_DATA_WIDTH_8;
+		break;
+	case MEDIA_BUS_FMT_RGB444_2X8_PADHI_BE:
+	case MEDIA_BUS_FMT_RGB444_2X8_PADHI_LE:
+		cfg->data_fmt = CSI_SENS_CONF_DATA_FMT_RGB444;
+		cfg->mipi_dt = MIPI_DT_RGB444;
+		cfg->data_width = IPU_CSI_DATA_WIDTH_8;
+		break;
+	case MEDIA_BUS_FMT_RGB555_2X8_PADHI_BE:
+	case MEDIA_BUS_FMT_RGB555_2X8_PADHI_LE:
+		cfg->data_fmt = CSI_SENS_CONF_DATA_FMT_RGB555;
+		cfg->mipi_dt = MIPI_DT_RGB555;
+		cfg->data_width = IPU_CSI_DATA_WIDTH_8;
+		break;
+	case MEDIA_BUS_FMT_UYVY8_2X8:
+		cfg->data_fmt = CSI_SENS_CONF_DATA_FMT_YUV422_UYVY;
+		cfg->mipi_dt = MIPI_DT_YUV422;
+		cfg->data_width = IPU_CSI_DATA_WIDTH_8;
+		break;
+	case MEDIA_BUS_FMT_YUYV8_2X8:
+		cfg->data_fmt = CSI_SENS_CONF_DATA_FMT_YUV422_YUYV;
+		cfg->mipi_dt = MIPI_DT_YUV422;
+		cfg->data_width = IPU_CSI_DATA_WIDTH_8;
+		break;
+	case MEDIA_BUS_FMT_UYVY8_1X16:
+		cfg->data_fmt = CSI_SENS_CONF_DATA_FMT_YUV422_UYVY;
+		cfg->mipi_dt = MIPI_DT_YUV422;
+		cfg->data_width = IPU_CSI_DATA_WIDTH_16;
+		break;
+	case MEDIA_BUS_FMT_YUYV8_1X16:
+		cfg->data_fmt = CSI_SENS_CONF_DATA_FMT_YUV422_YUYV;
+		cfg->mipi_dt = MIPI_DT_YUV422;
+		cfg->data_width = IPU_CSI_DATA_WIDTH_16;
+		break;
+	case MEDIA_BUS_FMT_SBGGR8_1X8:
+	case MEDIA_BUS_FMT_SGBRG8_1X8:
+	case MEDIA_BUS_FMT_SGRBG8_1X8:
+	case MEDIA_BUS_FMT_SRGGB8_1X8:
+		cfg->data_fmt = CSI_SENS_CONF_DATA_FMT_BAYER;
+		cfg->mipi_dt = MIPI_DT_RAW8;
+		cfg->data_width = IPU_CSI_DATA_WIDTH_8;
+		break;
+	case MEDIA_BUS_FMT_SBGGR10_DPCM8_1X8:
+	case MEDIA_BUS_FMT_SGBRG10_DPCM8_1X8:
+	case MEDIA_BUS_FMT_SGRBG10_DPCM8_1X8:
+	case MEDIA_BUS_FMT_SRGGB10_DPCM8_1X8:
+	case MEDIA_BUS_FMT_SBGGR10_2X8_PADHI_BE:
+	case MEDIA_BUS_FMT_SBGGR10_2X8_PADHI_LE:
+	case MEDIA_BUS_FMT_SBGGR10_2X8_PADLO_BE:
+	case MEDIA_BUS_FMT_SBGGR10_2X8_PADLO_LE:
+		cfg->data_fmt = CSI_SENS_CONF_DATA_FMT_BAYER;
+		cfg->mipi_dt = MIPI_DT_RAW10;
+		cfg->data_width = IPU_CSI_DATA_WIDTH_8;
+		break;
+	case MEDIA_BUS_FMT_SBGGR10_1X10:
+	case MEDIA_BUS_FMT_SGBRG10_1X10:
+	case MEDIA_BUS_FMT_SGRBG10_1X10:
+	case MEDIA_BUS_FMT_SRGGB10_1X10:
+		cfg->data_fmt = CSI_SENS_CONF_DATA_FMT_BAYER;
+		cfg->mipi_dt = MIPI_DT_RAW10;
+		cfg->data_width = IPU_CSI_DATA_WIDTH_10;
+		break;
+	case MEDIA_BUS_FMT_SBGGR12_1X12:
+	case MEDIA_BUS_FMT_SGBRG12_1X12:
+	case MEDIA_BUS_FMT_SGRBG12_1X12:
+	case MEDIA_BUS_FMT_SRGGB12_1X12:
+		cfg->data_fmt = CSI_SENS_CONF_DATA_FMT_BAYER;
+		cfg->mipi_dt = MIPI_DT_RAW12;
+		cfg->data_width = IPU_CSI_DATA_WIDTH_12;
+		break;
+	case MEDIA_BUS_FMT_JPEG_1X8:
+		/* TODO */
+		cfg->data_fmt = CSI_SENS_CONF_DATA_FMT_JPEG;
+		cfg->mipi_dt = MIPI_DT_RAW8;
+		cfg->data_width = IPU_CSI_DATA_WIDTH_8;
+		break;
+	default:
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+/*
+ * Fill a CSI bus config struct from mbus_config and mbus_framefmt.
+ */
+static void fill_csi_bus_cfg(struct ipu_csi_bus_config *csicfg,
+				 struct v4l2_mbus_config *mbus_cfg,
+				 struct v4l2_mbus_framefmt *mbus_fmt)
+{
+	memset(csicfg, 0, sizeof(*csicfg));
+
+	mbus_code_to_bus_cfg(csicfg, mbus_fmt->code);
+
+	switch (mbus_cfg->type) {
+	case V4L2_MBUS_PARALLEL:
+		csicfg->ext_vsync = 1;
+		csicfg->vsync_pol = (mbus_cfg->flags &
+				     V4L2_MBUS_VSYNC_ACTIVE_LOW) ? 1 : 0;
+		csicfg->hsync_pol = (mbus_cfg->flags &
+				     V4L2_MBUS_HSYNC_ACTIVE_LOW) ? 1 : 0;
+		csicfg->pixclk_pol = (mbus_cfg->flags &
+				      V4L2_MBUS_PCLK_SAMPLE_FALLING) ? 1 : 0;
+		csicfg->clk_mode = IPU_CSI_CLK_MODE_GATED_CLK;
+		break;
+	case V4L2_MBUS_BT656:
+		csicfg->ext_vsync = 0;
+		if (V4L2_FIELD_HAS_BOTH(mbus_fmt->field))
+			csicfg->clk_mode = IPU_CSI_CLK_MODE_CCIR656_INTERLACED;
+		else
+			csicfg->clk_mode = IPU_CSI_CLK_MODE_CCIR656_PROGRESSIVE;
+		break;
+	case V4L2_MBUS_CSI2:
+		/*
+		 * MIPI CSI-2 requires non gated clock mode, all other
+		 * parameters are not applicable for MIPI CSI-2 bus.
+		 */
+		csicfg->clk_mode = IPU_CSI_CLK_MODE_NONGATED_CLK;
+		break;
+	default:
+		/* will never get here, keep compiler quiet */
+		break;
+	}
+}
+
+int ipu_csi_init_interface(struct ipu_csi *csi,
+			   struct v4l2_mbus_config *mbus_cfg,
+			   struct v4l2_mbus_framefmt *mbus_fmt)
+{
+	struct ipu_csi_bus_config cfg;
+	unsigned long flags;
+	u32 data = 0;
+
+	fill_csi_bus_cfg(&cfg, mbus_cfg, mbus_fmt);
+
+	/* Set the CSI_SENS_CONF register remaining fields */
+	data |= cfg.data_width << CSI_SENS_CONF_DATA_WIDTH_SHIFT |
+		cfg.data_fmt << CSI_SENS_CONF_DATA_FMT_SHIFT |
+		cfg.data_pol << CSI_SENS_CONF_DATA_POL_SHIFT |
+		cfg.vsync_pol << CSI_SENS_CONF_VSYNC_POL_SHIFT |
+		cfg.hsync_pol << CSI_SENS_CONF_HSYNC_POL_SHIFT |
+		cfg.pixclk_pol << CSI_SENS_CONF_PIX_CLK_POL_SHIFT |
+		cfg.ext_vsync << CSI_SENS_CONF_EXT_VSYNC_SHIFT |
+		cfg.clk_mode << CSI_SENS_CONF_SENS_PRTCL_SHIFT |
+		cfg.pack_tight << CSI_SENS_CONF_PACK_TIGHT_SHIFT |
+		cfg.force_eof << CSI_SENS_CONF_FORCE_EOF_SHIFT |
+		cfg.data_en_pol << CSI_SENS_CONF_DATA_EN_POL_SHIFT;
+
+	spin_lock_irqsave(&csi->lock, flags);
+
+	ipu_csi_write(csi, data, CSI_SENS_CONF);
+
+	/* Setup sensor frame size */
+	ipu_csi_write(csi,
+		      (mbus_fmt->width - 1) | ((mbus_fmt->height - 1) << 16),
+		      CSI_SENS_FRM_SIZE);
+
+	/* Set CCIR registers */
+
+	switch (cfg.clk_mode) {
+	case IPU_CSI_CLK_MODE_CCIR656_PROGRESSIVE:
+		ipu_csi_write(csi, 0x40030, CSI_CCIR_CODE_1);
+		ipu_csi_write(csi, 0xFF0000, CSI_CCIR_CODE_3);
+		break;
+	case IPU_CSI_CLK_MODE_CCIR656_INTERLACED:
+		if (mbus_fmt->width == 720 && mbus_fmt->height == 576) {
+			/*
+			 * PAL case
+			 *
+			 * Field0BlankEnd = 0x6, Field0BlankStart = 0x2,
+			 * Field0ActiveEnd = 0x4, Field0ActiveStart = 0
+			 * Field1BlankEnd = 0x7, Field1BlankStart = 0x3,
+			 * Field1ActiveEnd = 0x5, Field1ActiveStart = 0x1
+			 */
+			ipu_csi_write(csi, 0x40596 | CSI_CCIR_ERR_DET_EN,
+					  CSI_CCIR_CODE_1);
+			ipu_csi_write(csi, 0xD07DF, CSI_CCIR_CODE_2);
+			ipu_csi_write(csi, 0xFF0000, CSI_CCIR_CODE_3);
+
+		} else if (mbus_fmt->width == 720 && mbus_fmt->height == 480) {
+			/*
+			 * NTSC case
+			 *
+			 * Field0BlankEnd = 0x7, Field0BlankStart = 0x3,
+			 * Field0ActiveEnd = 0x5, Field0ActiveStart = 0x1
+			 * Field1BlankEnd = 0x6, Field1BlankStart = 0x2,
+			 * Field1ActiveEnd = 0x4, Field1ActiveStart = 0
+			 */
+			ipu_csi_write(csi, 0xD07DF | CSI_CCIR_ERR_DET_EN,
+					  CSI_CCIR_CODE_1);
+			ipu_csi_write(csi, 0x40596, CSI_CCIR_CODE_2);
+			ipu_csi_write(csi, 0xFF0000, CSI_CCIR_CODE_3);
+		} else {
+			dev_err(csi->ipu->dev,
+				"Unsupported CCIR656 interlaced video mode\n");
+			spin_unlock_irqrestore(&csi->lock, flags);
+			return -EINVAL;
+		}
+		break;
+	case IPU_CSI_CLK_MODE_CCIR1120_PROGRESSIVE_DDR:
+	case IPU_CSI_CLK_MODE_CCIR1120_PROGRESSIVE_SDR:
+	case IPU_CSI_CLK_MODE_CCIR1120_INTERLACED_DDR:
+	case IPU_CSI_CLK_MODE_CCIR1120_INTERLACED_SDR:
+		ipu_csi_write(csi, 0x40030 | CSI_CCIR_ERR_DET_EN,
+				   CSI_CCIR_CODE_1);
+		ipu_csi_write(csi, 0xFF0000, CSI_CCIR_CODE_3);
+		break;
+	case IPU_CSI_CLK_MODE_GATED_CLK:
+	case IPU_CSI_CLK_MODE_NONGATED_CLK:
+		ipu_csi_write(csi, 0, CSI_CCIR_CODE_1);
+		break;
+	}
+
+	dev_dbg(csi->ipu->dev, "CSI_SENS_CONF = 0x%08X\n",
+		ipu_csi_read(csi, CSI_SENS_CONF));
+	dev_dbg(csi->ipu->dev, "CSI_ACT_FRM_SIZE = 0x%08X\n",
+		ipu_csi_read(csi, CSI_ACT_FRM_SIZE));
+
+	spin_unlock_irqrestore(&csi->lock, flags);
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(ipu_csi_init_interface);
+
+bool ipu_csi_is_interlaced(struct ipu_csi *csi)
+{
+	unsigned long flags;
+	u32 sensor_protocol;
+
+	spin_lock_irqsave(&csi->lock, flags);
+	sensor_protocol =
+		(ipu_csi_read(csi, CSI_SENS_CONF) &
+		 CSI_SENS_CONF_SENS_PRTCL_MASK) >>
+		CSI_SENS_CONF_SENS_PRTCL_SHIFT;
+	spin_unlock_irqrestore(&csi->lock, flags);
+
+	switch (sensor_protocol) {
+	case IPU_CSI_CLK_MODE_GATED_CLK:
+	case IPU_CSI_CLK_MODE_NONGATED_CLK:
+	case IPU_CSI_CLK_MODE_CCIR656_PROGRESSIVE:
+	case IPU_CSI_CLK_MODE_CCIR1120_PROGRESSIVE_DDR:
+	case IPU_CSI_CLK_MODE_CCIR1120_PROGRESSIVE_SDR:
+		return false;
+	case IPU_CSI_CLK_MODE_CCIR656_INTERLACED:
+	case IPU_CSI_CLK_MODE_CCIR1120_INTERLACED_DDR:
+	case IPU_CSI_CLK_MODE_CCIR1120_INTERLACED_SDR:
+		return true;
+	default:
+		dev_err(csi->ipu->dev,
+			"CSI %d sensor protocol unsupported\n", csi->id);
+		return false;
+	}
+}
+EXPORT_SYMBOL_GPL(ipu_csi_is_interlaced);
+
+void ipu_csi_get_window(struct ipu_csi *csi, struct v4l2_rect *w)
+{
+	unsigned long flags;
+	u32 reg;
+
+	spin_lock_irqsave(&csi->lock, flags);
+
+	reg = ipu_csi_read(csi, CSI_ACT_FRM_SIZE);
+	w->width = (reg & 0xFFFF) + 1;
+	w->height = (reg >> 16 & 0xFFFF) + 1;
+
+	reg = ipu_csi_read(csi, CSI_OUT_FRM_CTRL);
+	w->left = (reg & CSI_HSC_MASK) >> CSI_HSC_SHIFT;
+	w->top = (reg & CSI_VSC_MASK) >> CSI_VSC_SHIFT;
+
+	spin_unlock_irqrestore(&csi->lock, flags);
+}
+EXPORT_SYMBOL_GPL(ipu_csi_get_window);
+
+void ipu_csi_set_window(struct ipu_csi *csi, struct v4l2_rect *w)
+{
+	unsigned long flags;
+	u32 reg;
+
+	spin_lock_irqsave(&csi->lock, flags);
+
+	ipu_csi_write(csi, (w->width - 1) | ((w->height - 1) << 16),
+			  CSI_ACT_FRM_SIZE);
+
+	reg = ipu_csi_read(csi, CSI_OUT_FRM_CTRL);
+	reg &= ~(CSI_HSC_MASK | CSI_VSC_MASK);
+	reg |= ((w->top << CSI_VSC_SHIFT) | (w->left << CSI_HSC_SHIFT));
+	ipu_csi_write(csi, reg, CSI_OUT_FRM_CTRL);
+
+	spin_unlock_irqrestore(&csi->lock, flags);
+}
+EXPORT_SYMBOL_GPL(ipu_csi_set_window);
+
+void ipu_csi_set_test_generator(struct ipu_csi *csi, bool active,
+				u32 r_value, u32 g_value, u32 b_value,
+				u32 pix_clk)
+{
+	unsigned long flags;
+	u32 ipu_clk = clk_get_rate(csi->clk_ipu);
+	u32 temp;
+
+	spin_lock_irqsave(&csi->lock, flags);
+
+	temp = ipu_csi_read(csi, CSI_TST_CTRL);
+
+	if (active == false) {
+		temp &= ~CSI_TEST_GEN_MODE_EN;
+		ipu_csi_write(csi, temp, CSI_TST_CTRL);
+	} else {
+		/* Set sensb_mclk div_ratio */
+		ipu_csi_set_testgen_mclk(csi, pix_clk, ipu_clk);
+
+		temp &= ~(CSI_TEST_GEN_R_MASK | CSI_TEST_GEN_G_MASK |
+			  CSI_TEST_GEN_B_MASK);
+		temp |= CSI_TEST_GEN_MODE_EN;
+		temp |= (r_value << CSI_TEST_GEN_R_SHIFT) |
+			(g_value << CSI_TEST_GEN_G_SHIFT) |
+			(b_value << CSI_TEST_GEN_B_SHIFT);
+		ipu_csi_write(csi, temp, CSI_TST_CTRL);
+	}
+
+	spin_unlock_irqrestore(&csi->lock, flags);
+}
+EXPORT_SYMBOL_GPL(ipu_csi_set_test_generator);
+
+int ipu_csi_set_mipi_datatype(struct ipu_csi *csi, u32 vc,
+			      struct v4l2_mbus_framefmt *mbus_fmt)
+{
+	struct ipu_csi_bus_config cfg;
+	unsigned long flags;
+	u32 temp;
+
+	if (vc > 3)
+		return -EINVAL;
+
+	mbus_code_to_bus_cfg(&cfg, mbus_fmt->code);
+
+	spin_lock_irqsave(&csi->lock, flags);
+
+	temp = ipu_csi_read(csi, CSI_MIPI_DI);
+	temp &= ~(0xff << (vc * 8));
+	temp |= (cfg.mipi_dt << (vc * 8));
+	ipu_csi_write(csi, temp, CSI_MIPI_DI);
+
+	spin_unlock_irqrestore(&csi->lock, flags);
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(ipu_csi_set_mipi_datatype);
+
+int ipu_csi_set_skip_smfc(struct ipu_csi *csi, u32 skip,
+			  u32 max_ratio, u32 id)
+{
+	unsigned long flags;
+	u32 temp;
+
+	if (max_ratio > 5 || id > 3)
+		return -EINVAL;
+
+	spin_lock_irqsave(&csi->lock, flags);
+
+	temp = ipu_csi_read(csi, CSI_SKIP);
+	temp &= ~(CSI_MAX_RATIO_SKIP_SMFC_MASK | CSI_ID_2_SKIP_MASK |
+		  CSI_SKIP_SMFC_MASK);
+	temp |= (max_ratio << CSI_MAX_RATIO_SKIP_SMFC_SHIFT) |
+		(id << CSI_ID_2_SKIP_SHIFT) |
+		(skip << CSI_SKIP_SMFC_SHIFT);
+	ipu_csi_write(csi, temp, CSI_SKIP);
+
+	spin_unlock_irqrestore(&csi->lock, flags);
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(ipu_csi_set_skip_smfc);
+
+int ipu_csi_set_dest(struct ipu_csi *csi, enum ipu_csi_dest csi_dest)
+{
+	unsigned long flags;
+	u32 csi_sens_conf, dest;
+
+	if (csi_dest == IPU_CSI_DEST_IDMAC)
+		dest = CSI_DATA_DEST_IDMAC;
+	else
+		dest = CSI_DATA_DEST_IC; /* IC or VDIC */
+
+	spin_lock_irqsave(&csi->lock, flags);
+
+	csi_sens_conf = ipu_csi_read(csi, CSI_SENS_CONF);
+	csi_sens_conf &= ~CSI_SENS_CONF_DATA_DEST_MASK;
+	csi_sens_conf |= (dest << CSI_SENS_CONF_DATA_DEST_SHIFT);
+	ipu_csi_write(csi, csi_sens_conf, CSI_SENS_CONF);
+
+	spin_unlock_irqrestore(&csi->lock, flags);
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(ipu_csi_set_dest);
+
+int ipu_csi_enable(struct ipu_csi *csi)
+{
+	ipu_module_enable(csi->ipu, csi->module);
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(ipu_csi_enable);
+
+int ipu_csi_disable(struct ipu_csi *csi)
+{
+	ipu_module_disable(csi->ipu, csi->module);
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(ipu_csi_disable);
+
+struct ipu_csi *ipu_csi_get(struct ipu_soc *ipu, int id)
+{
+	unsigned long flags;
+	struct ipu_csi *csi, *ret;
+
+	if (id > 1)
+		return ERR_PTR(-EINVAL);
+
+	csi = ipu->csi_priv[id];
+	ret = csi;
+
+	spin_lock_irqsave(&csi->lock, flags);
+
+	if (csi->inuse) {
+		ret = ERR_PTR(-EBUSY);
+		goto unlock;
+	}
+
+	csi->inuse = true;
+unlock:
+	spin_unlock_irqrestore(&csi->lock, flags);
+	return ret;
+}
+EXPORT_SYMBOL_GPL(ipu_csi_get);
+
+void ipu_csi_put(struct ipu_csi *csi)
+{
+	unsigned long flags;
+
+	spin_lock_irqsave(&csi->lock, flags);
+	csi->inuse = false;
+	spin_unlock_irqrestore(&csi->lock, flags);
+}
+EXPORT_SYMBOL_GPL(ipu_csi_put);
+
+int ipu_csi_init(struct ipu_soc *ipu, struct device *dev, int id,
+		 unsigned long base, u32 module, struct clk *clk_ipu)
+{
+	struct ipu_csi *csi;
+
+	if (id > 1)
+		return -ENODEV;
+
+	csi = devm_kzalloc(dev, sizeof(*csi), GFP_KERNEL);
+	if (!csi)
+		return -ENOMEM;
+
+	ipu->csi_priv[id] = csi;
+
+	spin_lock_init(&csi->lock);
+	csi->module = module;
+	csi->id = id;
+	csi->clk_ipu = clk_ipu;
+	csi->base = devm_ioremap(dev, base, PAGE_SIZE);
+	if (!csi->base)
+		return -ENOMEM;
+
+	dev_dbg(dev, "CSI%d base: 0x%08lx remapped to %p\n",
+		id, base, csi->base);
+	csi->ipu = ipu;
+
+	return 0;
+}
+
+void ipu_csi_exit(struct ipu_soc *ipu, int id)
+{
+}
+
+void ipu_csi_dump(struct ipu_csi *csi)
+{
+	dev_dbg(csi->ipu->dev, "CSI_SENS_CONF:     %08x\n",
+		ipu_csi_read(csi, CSI_SENS_CONF));
+	dev_dbg(csi->ipu->dev, "CSI_SENS_FRM_SIZE: %08x\n",
+		ipu_csi_read(csi, CSI_SENS_FRM_SIZE));
+	dev_dbg(csi->ipu->dev, "CSI_ACT_FRM_SIZE:  %08x\n",
+		ipu_csi_read(csi, CSI_ACT_FRM_SIZE));
+	dev_dbg(csi->ipu->dev, "CSI_OUT_FRM_CTRL:  %08x\n",
+		ipu_csi_read(csi, CSI_OUT_FRM_CTRL));
+	dev_dbg(csi->ipu->dev, "CSI_TST_CTRL:      %08x\n",
+		ipu_csi_read(csi, CSI_TST_CTRL));
+	dev_dbg(csi->ipu->dev, "CSI_CCIR_CODE_1:   %08x\n",
+		ipu_csi_read(csi, CSI_CCIR_CODE_1));
+	dev_dbg(csi->ipu->dev, "CSI_CCIR_CODE_2:   %08x\n",
+		ipu_csi_read(csi, CSI_CCIR_CODE_2));
+	dev_dbg(csi->ipu->dev, "CSI_CCIR_CODE_3:   %08x\n",
+		ipu_csi_read(csi, CSI_CCIR_CODE_3));
+	dev_dbg(csi->ipu->dev, "CSI_MIPI_DI:       %08x\n",
+		ipu_csi_read(csi, CSI_MIPI_DI));
+	dev_dbg(csi->ipu->dev, "CSI_SKIP:          %08x\n",
+		ipu_csi_read(csi, CSI_SKIP));
+}
+EXPORT_SYMBOL_GPL(ipu_csi_dump);
diff --git a/drivers/gpu/ipu-v3/ipu-ic.c b/drivers/gpu/ipu-v3/ipu-ic.c
new file mode 100644
index 000000000000..ad75588e1629
--- /dev/null
+++ b/drivers/gpu/ipu-v3/ipu-ic.c
@@ -0,0 +1,778 @@
+/*
+ * Copyright (C) 2012-2014 Mentor Graphics Inc.
+ * Copyright 2005-2012 Freescale Semiconductor, Inc. All Rights Reserved.
+ *
+ * The code contained herein is licensed under the GNU General Public
+ * License. You may obtain a copy of the GNU General Public License
+ * Version 2 or later at the following locations:
+ *
+ * http://www.opensource.org/licenses/gpl-license.html
+ * http://www.gnu.org/copyleft/gpl.html
+ */
+
+#include <linux/types.h>
+#include <linux/init.h>
+#include <linux/errno.h>
+#include <linux/spinlock.h>
+#include <linux/bitrev.h>
+#include <linux/io.h>
+#include <linux/err.h>
+#include "ipu-prv.h"
+
+/* IC Register Offsets */
+#define IC_CONF                 0x0000
+#define IC_PRP_ENC_RSC          0x0004
+#define IC_PRP_VF_RSC           0x0008
+#define IC_PP_RSC               0x000C
+#define IC_CMBP_1               0x0010
+#define IC_CMBP_2               0x0014
+#define IC_IDMAC_1              0x0018
+#define IC_IDMAC_2              0x001C
+#define IC_IDMAC_3              0x0020
+#define IC_IDMAC_4              0x0024
+
+/* IC Register Fields */
+#define IC_CONF_PRPENC_EN       (1 << 0)
+#define IC_CONF_PRPENC_CSC1     (1 << 1)
+#define IC_CONF_PRPENC_ROT_EN   (1 << 2)
+#define IC_CONF_PRPVF_EN        (1 << 8)
+#define IC_CONF_PRPVF_CSC1      (1 << 9)
+#define IC_CONF_PRPVF_CSC2      (1 << 10)
+#define IC_CONF_PRPVF_CMB       (1 << 11)
+#define IC_CONF_PRPVF_ROT_EN    (1 << 12)
+#define IC_CONF_PP_EN           (1 << 16)
+#define IC_CONF_PP_CSC1         (1 << 17)
+#define IC_CONF_PP_CSC2         (1 << 18)
+#define IC_CONF_PP_CMB          (1 << 19)
+#define IC_CONF_PP_ROT_EN       (1 << 20)
+#define IC_CONF_IC_GLB_LOC_A    (1 << 28)
+#define IC_CONF_KEY_COLOR_EN    (1 << 29)
+#define IC_CONF_RWS_EN          (1 << 30)
+#define IC_CONF_CSI_MEM_WR_EN   (1 << 31)
+
+#define IC_IDMAC_1_CB0_BURST_16         (1 << 0)
+#define IC_IDMAC_1_CB1_BURST_16         (1 << 1)
+#define IC_IDMAC_1_CB2_BURST_16         (1 << 2)
+#define IC_IDMAC_1_CB3_BURST_16         (1 << 3)
+#define IC_IDMAC_1_CB4_BURST_16         (1 << 4)
+#define IC_IDMAC_1_CB5_BURST_16         (1 << 5)
+#define IC_IDMAC_1_CB6_BURST_16         (1 << 6)
+#define IC_IDMAC_1_CB7_BURST_16         (1 << 7)
+#define IC_IDMAC_1_PRPENC_ROT_MASK      (0x7 << 11)
+#define IC_IDMAC_1_PRPENC_ROT_OFFSET    11
+#define IC_IDMAC_1_PRPVF_ROT_MASK       (0x7 << 14)
+#define IC_IDMAC_1_PRPVF_ROT_OFFSET     14
+#define IC_IDMAC_1_PP_ROT_MASK          (0x7 << 17)
+#define IC_IDMAC_1_PP_ROT_OFFSET        17
+#define IC_IDMAC_1_PP_FLIP_RS           (1 << 22)
+#define IC_IDMAC_1_PRPVF_FLIP_RS        (1 << 21)
+#define IC_IDMAC_1_PRPENC_FLIP_RS       (1 << 20)
+
+#define IC_IDMAC_2_PRPENC_HEIGHT_MASK   (0x3ff << 0)
+#define IC_IDMAC_2_PRPENC_HEIGHT_OFFSET 0
+#define IC_IDMAC_2_PRPVF_HEIGHT_MASK    (0x3ff << 10)
+#define IC_IDMAC_2_PRPVF_HEIGHT_OFFSET  10
+#define IC_IDMAC_2_PP_HEIGHT_MASK       (0x3ff << 20)
+#define IC_IDMAC_2_PP_HEIGHT_OFFSET     20
+
+#define IC_IDMAC_3_PRPENC_WIDTH_MASK    (0x3ff << 0)
+#define IC_IDMAC_3_PRPENC_WIDTH_OFFSET  0
+#define IC_IDMAC_3_PRPVF_WIDTH_MASK     (0x3ff << 10)
+#define IC_IDMAC_3_PRPVF_WIDTH_OFFSET   10
+#define IC_IDMAC_3_PP_WIDTH_MASK        (0x3ff << 20)
+#define IC_IDMAC_3_PP_WIDTH_OFFSET      20
+
+struct ic_task_regoffs {
+	u32 rsc;
+	u32 tpmem_csc[2];
+};
+
+struct ic_task_bitfields {
+	u32 ic_conf_en;
+	u32 ic_conf_rot_en;
+	u32 ic_conf_cmb_en;
+	u32 ic_conf_csc1_en;
+	u32 ic_conf_csc2_en;
+	u32 ic_cmb_galpha_bit;
+};
+
+static const struct ic_task_regoffs ic_task_reg[IC_NUM_TASKS] = {
+	[IC_TASK_ENCODER] = {
+		.rsc = IC_PRP_ENC_RSC,
+		.tpmem_csc = {0x2008, 0},
+	},
+	[IC_TASK_VIEWFINDER] = {
+		.rsc = IC_PRP_VF_RSC,
+		.tpmem_csc = {0x4028, 0x4040},
+	},
+	[IC_TASK_POST_PROCESSOR] = {
+		.rsc = IC_PP_RSC,
+		.tpmem_csc = {0x6060, 0x6078},
+	},
+};
+
+static const struct ic_task_bitfields ic_task_bit[IC_NUM_TASKS] = {
+	[IC_TASK_ENCODER] = {
+		.ic_conf_en = IC_CONF_PRPENC_EN,
+		.ic_conf_rot_en = IC_CONF_PRPENC_ROT_EN,
+		.ic_conf_cmb_en = 0,    /* NA */
+		.ic_conf_csc1_en = IC_CONF_PRPENC_CSC1,
+		.ic_conf_csc2_en = 0,   /* NA */
+		.ic_cmb_galpha_bit = 0, /* NA */
+	},
+	[IC_TASK_VIEWFINDER] = {
+		.ic_conf_en = IC_CONF_PRPVF_EN,
+		.ic_conf_rot_en = IC_CONF_PRPVF_ROT_EN,
+		.ic_conf_cmb_en = IC_CONF_PRPVF_CMB,
+		.ic_conf_csc1_en = IC_CONF_PRPVF_CSC1,
+		.ic_conf_csc2_en = IC_CONF_PRPVF_CSC2,
+		.ic_cmb_galpha_bit = 0,
+	},
+	[IC_TASK_POST_PROCESSOR] = {
+		.ic_conf_en = IC_CONF_PP_EN,
+		.ic_conf_rot_en = IC_CONF_PP_ROT_EN,
+		.ic_conf_cmb_en = IC_CONF_PP_CMB,
+		.ic_conf_csc1_en = IC_CONF_PP_CSC1,
+		.ic_conf_csc2_en = IC_CONF_PP_CSC2,
+		.ic_cmb_galpha_bit = 8,
+	},
+};
+
+struct ipu_ic_priv;
+
+struct ipu_ic {
+	enum ipu_ic_task task;
+	const struct ic_task_regoffs *reg;
+	const struct ic_task_bitfields *bit;
+
+	enum ipu_color_space in_cs, g_in_cs;
+	enum ipu_color_space out_cs;
+	bool graphics;
+	bool rotation;
+	bool in_use;
+
+	struct ipu_ic_priv *priv;
+};
+
+struct ipu_ic_priv {
+	void __iomem *base;
+	void __iomem *tpmem_base;
+	spinlock_t lock;
+	struct ipu_soc *ipu;
+	int use_count;
+	struct ipu_ic task[IC_NUM_TASKS];
+};
+
+static inline u32 ipu_ic_read(struct ipu_ic *ic, unsigned offset)
+{
+	return readl(ic->priv->base + offset);
+}
+
+static inline void ipu_ic_write(struct ipu_ic *ic, u32 value, unsigned offset)
+{
+	writel(value, ic->priv->base + offset);
+}
+
+struct ic_csc_params {
+	s16 coeff[3][3];	/* signed 9-bit integer coefficients */
+	s16 offset[3];		/* signed 11+2-bit fixed point offset */
+	u8 scale:2;		/* scale coefficients * 2^(scale-1) */
+	bool sat:1;		/* saturate to (16, 235(Y) / 240(U, V)) */
+};
+
+/*
+ * Y = R *  .299 + G *  .587 + B *  .114;
+ * U = R * -.169 + G * -.332 + B *  .500 + 128.;
+ * V = R *  .500 + G * -.419 + B * -.0813 + 128.;
+ */
+static const struct ic_csc_params ic_csc_rgb2ycbcr = {
+	.coeff = {
+		{ 77, 150, 29 },
+		{ 469, 427, 128 },
+		{ 128, 405, 491 },
+	},
+	.offset = { 0, 512, 512 },
+	.scale = 1,
+};
+
+/* transparent RGB->RGB matrix for graphics combining */
+static const struct ic_csc_params ic_csc_rgb2rgb = {
+	.coeff = {
+		{ 128, 0, 0 },
+		{ 0, 128, 0 },
+		{ 0, 0, 128 },
+	},
+	.scale = 2,
+};
+
+/*
+ * R = (1.164 * (Y - 16)) + (1.596 * (Cr - 128));
+ * G = (1.164 * (Y - 16)) - (0.392 * (Cb - 128)) - (0.813 * (Cr - 128));
+ * B = (1.164 * (Y - 16)) + (2.017 * (Cb - 128);
+ */
+static const struct ic_csc_params ic_csc_ycbcr2rgb = {
+	.coeff = {
+		{ 149, 0, 204 },
+		{ 149, 462, 408 },
+		{ 149, 255, 0 },
+	},
+	.offset = { -446, 266, -554 },
+	.scale = 2,
+};
+
+static int init_csc(struct ipu_ic *ic,
+		    enum ipu_color_space inf,
+		    enum ipu_color_space outf,
+		    int csc_index)
+{
+	struct ipu_ic_priv *priv = ic->priv;
+	const struct ic_csc_params *params;
+	u32 __iomem *base;
+	const u16 (*c)[3];
+	const u16 *a;
+	u32 param;
+
+	base = (u32 __iomem *)
+		(priv->tpmem_base + ic->reg->tpmem_csc[csc_index]);
+
+	if (inf == IPUV3_COLORSPACE_YUV && outf == IPUV3_COLORSPACE_RGB)
+		params = &ic_csc_ycbcr2rgb;
+	else if (inf == IPUV3_COLORSPACE_RGB && outf == IPUV3_COLORSPACE_YUV)
+		params = &ic_csc_rgb2ycbcr;
+	else if (inf == IPUV3_COLORSPACE_RGB && outf == IPUV3_COLORSPACE_RGB)
+		params = &ic_csc_rgb2rgb;
+	else {
+		dev_err(priv->ipu->dev, "Unsupported color space conversion\n");
+		return -EINVAL;
+	}
+
+	/* Cast to unsigned */
+	c = (const u16 (*)[3])params->coeff;
+	a = (const u16 *)params->offset;
+
+	param = ((a[0] & 0x1f) << 27) | ((c[0][0] & 0x1ff) << 18) |
+		((c[1][1] & 0x1ff) << 9) | (c[2][2] & 0x1ff);
+	writel(param, base++);
+
+	param = ((a[0] & 0x1fe0) >> 5) | (params->scale << 8) |
+		(params->sat << 9);
+	writel(param, base++);
+
+	param = ((a[1] & 0x1f) << 27) | ((c[0][1] & 0x1ff) << 18) |
+		((c[1][0] & 0x1ff) << 9) | (c[2][0] & 0x1ff);
+	writel(param, base++);
+
+	param = ((a[1] & 0x1fe0) >> 5);
+	writel(param, base++);
+
+	param = ((a[2] & 0x1f) << 27) | ((c[0][2] & 0x1ff) << 18) |
+		((c[1][2] & 0x1ff) << 9) | (c[2][1] & 0x1ff);
+	writel(param, base++);
+
+	param = ((a[2] & 0x1fe0) >> 5);
+	writel(param, base++);
+
+	return 0;
+}
+
+static int calc_resize_coeffs(struct ipu_ic *ic,
+			      u32 in_size, u32 out_size,
+			      u32 *resize_coeff,
+			      u32 *downsize_coeff)
+{
+	struct ipu_ic_priv *priv = ic->priv;
+	struct ipu_soc *ipu = priv->ipu;
+	u32 temp_size, temp_downsize;
+
+	/*
+	 * Input size cannot be more than 4096, and output size cannot
+	 * be more than 1024
+	 */
+	if (in_size > 4096) {
+		dev_err(ipu->dev, "Unsupported resize (in_size > 4096)\n");
+		return -EINVAL;
+	}
+	if (out_size > 1024) {
+		dev_err(ipu->dev, "Unsupported resize (out_size > 1024)\n");
+		return -EINVAL;
+	}
+
+	/* Cannot downsize more than 8:1 */
+	if ((out_size << 3) < in_size) {
+		dev_err(ipu->dev, "Unsupported downsize\n");
+		return -EINVAL;
+	}
+
+	/* Compute downsizing coefficient */
+	temp_downsize = 0;
+	temp_size = in_size;
+	while (((temp_size > 1024) || (temp_size >= out_size * 2)) &&
+	       (temp_downsize < 2)) {
+		temp_size >>= 1;
+		temp_downsize++;
+	}
+	*downsize_coeff = temp_downsize;
+
+	/*
+	 * compute resizing coefficient using the following equation:
+	 * resize_coeff = M * (SI - 1) / (SO - 1)
+	 * where M = 2^13, SI = input size, SO = output size
+	 */
+	*resize_coeff = (8192L * (temp_size - 1)) / (out_size - 1);
+	if (*resize_coeff >= 16384L) {
+		dev_err(ipu->dev, "Warning! Overflow on resize coeff.\n");
+		*resize_coeff = 0x3FFF;
+	}
+
+	return 0;
+}
+
+void ipu_ic_task_enable(struct ipu_ic *ic)
+{
+	struct ipu_ic_priv *priv = ic->priv;
+	unsigned long flags;
+	u32 ic_conf;
+
+	spin_lock_irqsave(&priv->lock, flags);
+
+	ic_conf = ipu_ic_read(ic, IC_CONF);
+
+	ic_conf |= ic->bit->ic_conf_en;
+
+	if (ic->rotation)
+		ic_conf |= ic->bit->ic_conf_rot_en;
+
+	if (ic->in_cs != ic->out_cs)
+		ic_conf |= ic->bit->ic_conf_csc1_en;
+
+	if (ic->graphics) {
+		ic_conf |= ic->bit->ic_conf_cmb_en;
+		ic_conf |= ic->bit->ic_conf_csc1_en;
+
+		if (ic->g_in_cs != ic->out_cs)
+			ic_conf |= ic->bit->ic_conf_csc2_en;
+	}
+
+	ipu_ic_write(ic, ic_conf, IC_CONF);
+
+	spin_unlock_irqrestore(&priv->lock, flags);
+}
+EXPORT_SYMBOL_GPL(ipu_ic_task_enable);
+
+void ipu_ic_task_disable(struct ipu_ic *ic)
+{
+	struct ipu_ic_priv *priv = ic->priv;
+	unsigned long flags;
+	u32 ic_conf;
+
+	spin_lock_irqsave(&priv->lock, flags);
+
+	ic_conf = ipu_ic_read(ic, IC_CONF);
+
+	ic_conf &= ~(ic->bit->ic_conf_en |
+		     ic->bit->ic_conf_csc1_en |
+		     ic->bit->ic_conf_rot_en);
+	if (ic->bit->ic_conf_csc2_en)
+		ic_conf &= ~ic->bit->ic_conf_csc2_en;
+	if (ic->bit->ic_conf_cmb_en)
+		ic_conf &= ~ic->bit->ic_conf_cmb_en;
+
+	ipu_ic_write(ic, ic_conf, IC_CONF);
+
+	ic->rotation = ic->graphics = false;
+
+	spin_unlock_irqrestore(&priv->lock, flags);
+}
+EXPORT_SYMBOL_GPL(ipu_ic_task_disable);
+
+int ipu_ic_task_graphics_init(struct ipu_ic *ic,
+			      enum ipu_color_space in_g_cs,
+			      bool galpha_en, u32 galpha,
+			      bool colorkey_en, u32 colorkey)
+{
+	struct ipu_ic_priv *priv = ic->priv;
+	unsigned long flags;
+	u32 reg, ic_conf;
+	int ret = 0;
+
+	if (ic->task == IC_TASK_ENCODER)
+		return -EINVAL;
+
+	spin_lock_irqsave(&priv->lock, flags);
+
+	ic_conf = ipu_ic_read(ic, IC_CONF);
+
+	if (!(ic_conf & ic->bit->ic_conf_csc1_en)) {
+		/* need transparent CSC1 conversion */
+		ret = init_csc(ic, IPUV3_COLORSPACE_RGB,
+			       IPUV3_COLORSPACE_RGB, 0);
+		if (ret)
+			goto unlock;
+	}
+
+	ic->g_in_cs = in_g_cs;
+
+	if (ic->g_in_cs != ic->out_cs) {
+		ret = init_csc(ic, ic->g_in_cs, ic->out_cs, 1);
+		if (ret)
+			goto unlock;
+	}
+
+	if (galpha_en) {
+		ic_conf |= IC_CONF_IC_GLB_LOC_A;
+		reg = ipu_ic_read(ic, IC_CMBP_1);
+		reg &= ~(0xff << ic->bit->ic_cmb_galpha_bit);
+		reg |= (galpha << ic->bit->ic_cmb_galpha_bit);
+		ipu_ic_write(ic, reg, IC_CMBP_1);
+	} else
+		ic_conf &= ~IC_CONF_IC_GLB_LOC_A;
+
+	if (colorkey_en) {
+		ic_conf |= IC_CONF_KEY_COLOR_EN;
+		ipu_ic_write(ic, colorkey, IC_CMBP_2);
+	} else
+		ic_conf &= ~IC_CONF_KEY_COLOR_EN;
+
+	ipu_ic_write(ic, ic_conf, IC_CONF);
+
+	ic->graphics = true;
+unlock:
+	spin_unlock_irqrestore(&priv->lock, flags);
+	return ret;
+}
+EXPORT_SYMBOL_GPL(ipu_ic_task_graphics_init);
+
+int ipu_ic_task_init(struct ipu_ic *ic,
+		     int in_width, int in_height,
+		     int out_width, int out_height,
+		     enum ipu_color_space in_cs,
+		     enum ipu_color_space out_cs)
+{
+	struct ipu_ic_priv *priv = ic->priv;
+	u32 reg, downsize_coeff, resize_coeff;
+	unsigned long flags;
+	int ret = 0;
+
+	/* Setup vertical resizing */
+	ret = calc_resize_coeffs(ic, in_height, out_height,
+				 &resize_coeff, &downsize_coeff);
+	if (ret)
+		return ret;
+
+	reg = (downsize_coeff << 30) | (resize_coeff << 16);
+
+	/* Setup horizontal resizing */
+	ret = calc_resize_coeffs(ic, in_width, out_width,
+				 &resize_coeff, &downsize_coeff);
+	if (ret)
+		return ret;
+
+	reg |= (downsize_coeff << 14) | resize_coeff;
+
+	spin_lock_irqsave(&priv->lock, flags);
+
+	ipu_ic_write(ic, reg, ic->reg->rsc);
+
+	/* Setup color space conversion */
+	ic->in_cs = in_cs;
+	ic->out_cs = out_cs;
+
+	if (ic->in_cs != ic->out_cs) {
+		ret = init_csc(ic, ic->in_cs, ic->out_cs, 0);
+		if (ret)
+			goto unlock;
+	}
+
+unlock:
+	spin_unlock_irqrestore(&priv->lock, flags);
+	return ret;
+}
+EXPORT_SYMBOL_GPL(ipu_ic_task_init);
+
+int ipu_ic_task_idma_init(struct ipu_ic *ic, struct ipuv3_channel *channel,
+			  u32 width, u32 height, int burst_size,
+			  enum ipu_rotate_mode rot)
+{
+	struct ipu_ic_priv *priv = ic->priv;
+	struct ipu_soc *ipu = priv->ipu;
+	u32 ic_idmac_1, ic_idmac_2, ic_idmac_3;
+	u32 temp_rot = bitrev8(rot) >> 5;
+	bool need_hor_flip = false;
+	unsigned long flags;
+	int ret = 0;
+
+	if ((burst_size != 8) && (burst_size != 16)) {
+		dev_err(ipu->dev, "Illegal burst length for IC\n");
+		return -EINVAL;
+	}
+
+	width--;
+	height--;
+
+	if (temp_rot & 0x2)	/* Need horizontal flip */
+		need_hor_flip = true;
+
+	spin_lock_irqsave(&priv->lock, flags);
+
+	ic_idmac_1 = ipu_ic_read(ic, IC_IDMAC_1);
+	ic_idmac_2 = ipu_ic_read(ic, IC_IDMAC_2);
+	ic_idmac_3 = ipu_ic_read(ic, IC_IDMAC_3);
+
+	switch (channel->num) {
+	case IPUV3_CHANNEL_IC_PP_MEM:
+		if (burst_size == 16)
+			ic_idmac_1 |= IC_IDMAC_1_CB2_BURST_16;
+		else
+			ic_idmac_1 &= ~IC_IDMAC_1_CB2_BURST_16;
+
+		if (need_hor_flip)
+			ic_idmac_1 |= IC_IDMAC_1_PP_FLIP_RS;
+		else
+			ic_idmac_1 &= ~IC_IDMAC_1_PP_FLIP_RS;
+
+		ic_idmac_2 &= ~IC_IDMAC_2_PP_HEIGHT_MASK;
+		ic_idmac_2 |= height << IC_IDMAC_2_PP_HEIGHT_OFFSET;
+
+		ic_idmac_3 &= ~IC_IDMAC_3_PP_WIDTH_MASK;
+		ic_idmac_3 |= width << IC_IDMAC_3_PP_WIDTH_OFFSET;
+		break;
+	case IPUV3_CHANNEL_MEM_IC_PP:
+		if (burst_size == 16)
+			ic_idmac_1 |= IC_IDMAC_1_CB5_BURST_16;
+		else
+			ic_idmac_1 &= ~IC_IDMAC_1_CB5_BURST_16;
+		break;
+	case IPUV3_CHANNEL_MEM_ROT_PP:
+		ic_idmac_1 &= ~IC_IDMAC_1_PP_ROT_MASK;
+		ic_idmac_1 |= temp_rot << IC_IDMAC_1_PP_ROT_OFFSET;
+		break;
+	case IPUV3_CHANNEL_MEM_IC_PRP_VF:
+		if (burst_size == 16)
+			ic_idmac_1 |= IC_IDMAC_1_CB6_BURST_16;
+		else
+			ic_idmac_1 &= ~IC_IDMAC_1_CB6_BURST_16;
+		break;
+	case IPUV3_CHANNEL_IC_PRP_ENC_MEM:
+		if (burst_size == 16)
+			ic_idmac_1 |= IC_IDMAC_1_CB0_BURST_16;
+		else
+			ic_idmac_1 &= ~IC_IDMAC_1_CB0_BURST_16;
+
+		if (need_hor_flip)
+			ic_idmac_1 |= IC_IDMAC_1_PRPENC_FLIP_RS;
+		else
+			ic_idmac_1 &= ~IC_IDMAC_1_PRPENC_FLIP_RS;
+
+		ic_idmac_2 &= ~IC_IDMAC_2_PRPENC_HEIGHT_MASK;
+		ic_idmac_2 |= height << IC_IDMAC_2_PRPENC_HEIGHT_OFFSET;
+
+		ic_idmac_3 &= ~IC_IDMAC_3_PRPENC_WIDTH_MASK;
+		ic_idmac_3 |= width << IC_IDMAC_3_PRPENC_WIDTH_OFFSET;
+		break;
+	case IPUV3_CHANNEL_MEM_ROT_ENC:
+		ic_idmac_1 &= ~IC_IDMAC_1_PRPENC_ROT_MASK;
+		ic_idmac_1 |= temp_rot << IC_IDMAC_1_PRPENC_ROT_OFFSET;
+		break;
+	case IPUV3_CHANNEL_IC_PRP_VF_MEM:
+		if (burst_size == 16)
+			ic_idmac_1 |= IC_IDMAC_1_CB1_BURST_16;
+		else
+			ic_idmac_1 &= ~IC_IDMAC_1_CB1_BURST_16;
+
+		if (need_hor_flip)
+			ic_idmac_1 |= IC_IDMAC_1_PRPVF_FLIP_RS;
+		else
+			ic_idmac_1 &= ~IC_IDMAC_1_PRPVF_FLIP_RS;
+
+		ic_idmac_2 &= ~IC_IDMAC_2_PRPVF_HEIGHT_MASK;
+		ic_idmac_2 |= height << IC_IDMAC_2_PRPVF_HEIGHT_OFFSET;
+
+		ic_idmac_3 &= ~IC_IDMAC_3_PRPVF_WIDTH_MASK;
+		ic_idmac_3 |= width << IC_IDMAC_3_PRPVF_WIDTH_OFFSET;
+		break;
+	case IPUV3_CHANNEL_MEM_ROT_VF:
+		ic_idmac_1 &= ~IC_IDMAC_1_PRPVF_ROT_MASK;
+		ic_idmac_1 |= temp_rot << IC_IDMAC_1_PRPVF_ROT_OFFSET;
+		break;
+	case IPUV3_CHANNEL_G_MEM_IC_PRP_VF:
+		if (burst_size == 16)
+			ic_idmac_1 |= IC_IDMAC_1_CB3_BURST_16;
+		else
+			ic_idmac_1 &= ~IC_IDMAC_1_CB3_BURST_16;
+		break;
+	case IPUV3_CHANNEL_G_MEM_IC_PP:
+		if (burst_size == 16)
+			ic_idmac_1 |= IC_IDMAC_1_CB4_BURST_16;
+		else
+			ic_idmac_1 &= ~IC_IDMAC_1_CB4_BURST_16;
+		break;
+	case IPUV3_CHANNEL_VDI_MEM_IC_VF:
+		if (burst_size == 16)
+			ic_idmac_1 |= IC_IDMAC_1_CB7_BURST_16;
+		else
+			ic_idmac_1 &= ~IC_IDMAC_1_CB7_BURST_16;
+		break;
+	default:
+		goto unlock;
+	}
+
+	ipu_ic_write(ic, ic_idmac_1, IC_IDMAC_1);
+	ipu_ic_write(ic, ic_idmac_2, IC_IDMAC_2);
+	ipu_ic_write(ic, ic_idmac_3, IC_IDMAC_3);
+
+	if (rot >= IPU_ROTATE_90_RIGHT)
+		ic->rotation = true;
+
+unlock:
+	spin_unlock_irqrestore(&priv->lock, flags);
+	return ret;
+}
+EXPORT_SYMBOL_GPL(ipu_ic_task_idma_init);
+
+int ipu_ic_enable(struct ipu_ic *ic)
+{
+	struct ipu_ic_priv *priv = ic->priv;
+	unsigned long flags;
+	u32 module = IPU_CONF_IC_EN;
+
+	spin_lock_irqsave(&priv->lock, flags);
+
+	if (ic->rotation)
+		module |= IPU_CONF_ROT_EN;
+
+	if (!priv->use_count)
+		ipu_module_enable(priv->ipu, module);
+
+	priv->use_count++;
+
+	spin_unlock_irqrestore(&priv->lock, flags);
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(ipu_ic_enable);
+
+int ipu_ic_disable(struct ipu_ic *ic)
+{
+	struct ipu_ic_priv *priv = ic->priv;
+	unsigned long flags;
+	u32 module = IPU_CONF_IC_EN | IPU_CONF_ROT_EN;
+
+	spin_lock_irqsave(&priv->lock, flags);
+
+	priv->use_count--;
+
+	if (!priv->use_count)
+		ipu_module_disable(priv->ipu, module);
+
+	if (priv->use_count < 0)
+		priv->use_count = 0;
+
+	spin_unlock_irqrestore(&priv->lock, flags);
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(ipu_ic_disable);
+
+struct ipu_ic *ipu_ic_get(struct ipu_soc *ipu, enum ipu_ic_task task)
+{
+	struct ipu_ic_priv *priv = ipu->ic_priv;
+	unsigned long flags;
+	struct ipu_ic *ic, *ret;
+
+	if (task >= IC_NUM_TASKS)
+		return ERR_PTR(-EINVAL);
+
+	ic = &priv->task[task];
+
+	spin_lock_irqsave(&priv->lock, flags);
+
+	if (ic->in_use) {
+		ret = ERR_PTR(-EBUSY);
+		goto unlock;
+	}
+
+	ic->in_use = true;
+	ret = ic;
+
+unlock:
+	spin_unlock_irqrestore(&priv->lock, flags);
+	return ret;
+}
+EXPORT_SYMBOL_GPL(ipu_ic_get);
+
+void ipu_ic_put(struct ipu_ic *ic)
+{
+	struct ipu_ic_priv *priv = ic->priv;
+	unsigned long flags;
+
+	spin_lock_irqsave(&priv->lock, flags);
+	ic->in_use = false;
+	spin_unlock_irqrestore(&priv->lock, flags);
+}
+EXPORT_SYMBOL_GPL(ipu_ic_put);
+
+int ipu_ic_init(struct ipu_soc *ipu, struct device *dev,
+		unsigned long base, unsigned long tpmem_base)
+{
+	struct ipu_ic_priv *priv;
+	int i;
+
+	priv = devm_kzalloc(dev, sizeof(*priv), GFP_KERNEL);
+	if (!priv)
+		return -ENOMEM;
+
+	ipu->ic_priv = priv;
+
+	spin_lock_init(&priv->lock);
+	priv->base = devm_ioremap(dev, base, PAGE_SIZE);
+	if (!priv->base)
+		return -ENOMEM;
+	priv->tpmem_base = devm_ioremap(dev, tpmem_base, SZ_64K);
+	if (!priv->tpmem_base)
+		return -ENOMEM;
+
+	dev_dbg(dev, "IC base: 0x%08lx remapped to %p\n", base, priv->base);
+
+	priv->ipu = ipu;
+
+	for (i = 0; i < IC_NUM_TASKS; i++) {
+		priv->task[i].task = i;
+		priv->task[i].priv = priv;
+		priv->task[i].reg = &ic_task_reg[i];
+		priv->task[i].bit = &ic_task_bit[i];
+	}
+
+	return 0;
+}
+
+void ipu_ic_exit(struct ipu_soc *ipu)
+{
+}
+
+void ipu_ic_dump(struct ipu_ic *ic)
+{
+	struct ipu_ic_priv *priv = ic->priv;
+	struct ipu_soc *ipu = priv->ipu;
+
+	dev_dbg(ipu->dev, "IC_CONF = \t0x%08X\n",
+		ipu_ic_read(ic, IC_CONF));
+	dev_dbg(ipu->dev, "IC_PRP_ENC_RSC = \t0x%08X\n",
+		ipu_ic_read(ic, IC_PRP_ENC_RSC));
+	dev_dbg(ipu->dev, "IC_PRP_VF_RSC = \t0x%08X\n",
+		ipu_ic_read(ic, IC_PRP_VF_RSC));
+	dev_dbg(ipu->dev, "IC_PP_RSC = \t0x%08X\n",
+		ipu_ic_read(ic, IC_PP_RSC));
+	dev_dbg(ipu->dev, "IC_CMBP_1 = \t0x%08X\n",
+		ipu_ic_read(ic, IC_CMBP_1));
+	dev_dbg(ipu->dev, "IC_CMBP_2 = \t0x%08X\n",
+		ipu_ic_read(ic, IC_CMBP_2));
+	dev_dbg(ipu->dev, "IC_IDMAC_1 = \t0x%08X\n",
+		ipu_ic_read(ic, IC_IDMAC_1));
+	dev_dbg(ipu->dev, "IC_IDMAC_2 = \t0x%08X\n",
+		ipu_ic_read(ic, IC_IDMAC_2));
+	dev_dbg(ipu->dev, "IC_IDMAC_3 = \t0x%08X\n",
+		ipu_ic_read(ic, IC_IDMAC_3));
+	dev_dbg(ipu->dev, "IC_IDMAC_4 = \t0x%08X\n",
+		ipu_ic_read(ic, IC_IDMAC_4));
+}
+EXPORT_SYMBOL_GPL(ipu_ic_dump);
diff --git a/drivers/gpu/ipu-v3/ipu-prv.h b/drivers/gpu/ipu-v3/ipu-prv.h
index c93f50ec04f7..bfb1e8a4483f 100644
--- a/drivers/gpu/ipu-v3/ipu-prv.h
+++ b/drivers/gpu/ipu-v3/ipu-prv.h
@@ -24,23 +24,6 @@ struct ipu_soc;
 
 #include <video/imx-ipu-v3.h>
 
-#define IPUV3_CHANNEL_CSI0			 0
-#define IPUV3_CHANNEL_CSI1			 1
-#define IPUV3_CHANNEL_CSI2			 2
-#define IPUV3_CHANNEL_CSI3			 3
-#define IPUV3_CHANNEL_MEM_BG_SYNC		23
-#define IPUV3_CHANNEL_MEM_FG_SYNC		27
-#define IPUV3_CHANNEL_MEM_DC_SYNC		28
-#define IPUV3_CHANNEL_MEM_FG_SYNC_ALPHA		31
-#define IPUV3_CHANNEL_MEM_DC_ASYNC		41
-#define IPUV3_CHANNEL_ROT_ENC_MEM		45
-#define IPUV3_CHANNEL_ROT_VF_MEM		46
-#define IPUV3_CHANNEL_ROT_PP_MEM		47
-#define IPUV3_CHANNEL_ROT_ENC_MEM_OUT		48
-#define IPUV3_CHANNEL_ROT_VF_MEM_OUT		49
-#define IPUV3_CHANNEL_ROT_PP_MEM_OUT		50
-#define IPUV3_CHANNEL_MEM_BG_SYNC_ALPHA		51
-
 #define IPU_MCU_T_DEFAULT	8
 #define IPU_CM_IDMAC_REG_OFS	0x00008000
 #define IPU_CM_IC_REG_OFS	0x00020000
@@ -85,6 +68,7 @@ struct ipu_soc;
 #define IPU_DISP_TASK_STAT		IPU_CM_REG(0x0254)
 #define IPU_CHA_BUF0_RDY(ch)		IPU_CM_REG(0x0268 + 4 * ((ch) / 32))
 #define IPU_CHA_BUF1_RDY(ch)		IPU_CM_REG(0x0270 + 4 * ((ch) / 32))
+#define IPU_CHA_BUF2_RDY(ch)		IPU_CM_REG(0x0288 + 4 * ((ch) / 32))
 #define IPU_ALT_CHA_BUF0_RDY(ch)	IPU_CM_REG(0x0278 + 4 * ((ch) / 32))
 #define IPU_ALT_CHA_BUF1_RDY(ch)	IPU_CM_REG(0x0280 + 4 * ((ch) / 32))
 
@@ -148,9 +132,12 @@ struct ipuv3_channel {
 	struct ipu_soc *ipu;
 };
 
+struct ipu_cpmem;
+struct ipu_csi;
 struct ipu_dc_priv;
 struct ipu_dmfc_priv;
 struct ipu_di;
+struct ipu_ic_priv;
 struct ipu_smfc_priv;
 
 struct ipu_devtype;
@@ -164,7 +151,6 @@ struct ipu_soc {
 
 	void __iomem		*cm_reg;
 	void __iomem		*idmac_reg;
-	struct ipu_ch_param __iomem	*cpmem_base;
 
 	int			usecount;
 
@@ -176,13 +162,27 @@ struct ipu_soc {
 	int			irq_err;
 	struct irq_domain	*domain;
 
+	struct ipu_cpmem	*cpmem_priv;
 	struct ipu_dc_priv	*dc_priv;
 	struct ipu_dp_priv	*dp_priv;
 	struct ipu_dmfc_priv	*dmfc_priv;
 	struct ipu_di		*di_priv[2];
+	struct ipu_csi		*csi_priv[2];
+	struct ipu_ic_priv	*ic_priv;
 	struct ipu_smfc_priv	*smfc_priv;
 };
 
+static inline u32 ipu_idmac_read(struct ipu_soc *ipu, unsigned offset)
+{
+	return readl(ipu->idmac_reg + offset);
+}
+
+static inline void ipu_idmac_write(struct ipu_soc *ipu, u32 value,
+				   unsigned offset)
+{
+	writel(value, ipu->idmac_reg + offset);
+}
+
 void ipu_srm_dp_sync_update(struct ipu_soc *ipu);
 
 int ipu_module_enable(struct ipu_soc *ipu, u32 mask);
@@ -191,6 +191,14 @@ int ipu_module_disable(struct ipu_soc *ipu, u32 mask);
 bool ipu_idmac_channel_busy(struct ipu_soc *ipu, unsigned int chno);
 int ipu_wait_interrupt(struct ipu_soc *ipu, int irq, int ms);
 
+int ipu_csi_init(struct ipu_soc *ipu, struct device *dev, int id,
+		 unsigned long base, u32 module, struct clk *clk_ipu);
+void ipu_csi_exit(struct ipu_soc *ipu, int id);
+
+int ipu_ic_init(struct ipu_soc *ipu, struct device *dev,
+		unsigned long base, unsigned long tpmem_base);
+void ipu_ic_exit(struct ipu_soc *ipu);
+
 int ipu_di_init(struct ipu_soc *ipu, struct device *dev, int id,
 		unsigned long base, u32 module, struct clk *ipu_clk);
 void ipu_di_exit(struct ipu_soc *ipu, int id);
diff --git a/drivers/gpu/ipu-v3/ipu-smfc.c b/drivers/gpu/ipu-v3/ipu-smfc.c
index e4f85ad286fc..4ef910991413 100644
--- a/drivers/gpu/ipu-v3/ipu-smfc.c
+++ b/drivers/gpu/ipu-v3/ipu-smfc.c
@@ -8,7 +8,6 @@
  * http://www.opensource.org/licenses/gpl-license.html
  * http://www.gnu.org/copyleft/gpl.html
  */
-#define DEBUG
 #include <linux/export.h>
 #include <linux/types.h>
 #include <linux/init.h>
@@ -21,9 +20,18 @@
 
 #include "ipu-prv.h"
 
+struct ipu_smfc {
+	struct ipu_smfc_priv *priv;
+	int chno;
+	bool inuse;
+};
+
 struct ipu_smfc_priv {
 	void __iomem *base;
 	spinlock_t lock;
+	struct ipu_soc *ipu;
+	struct ipu_smfc channel[4];
+	int use_count;
 };
 
 /*SMFC Registers */
@@ -31,63 +39,166 @@ struct ipu_smfc_priv {
 #define SMFC_WMC	0x0004
 #define SMFC_BS		0x0008
 
-int ipu_smfc_set_burstsize(struct ipu_soc *ipu, int channel, int burstsize)
+int ipu_smfc_set_burstsize(struct ipu_smfc *smfc, int burstsize)
 {
-	struct ipu_smfc_priv *smfc = ipu->smfc_priv;
+	struct ipu_smfc_priv *priv = smfc->priv;
 	unsigned long flags;
 	u32 val, shift;
 
-	spin_lock_irqsave(&smfc->lock, flags);
+	spin_lock_irqsave(&priv->lock, flags);
 
-	shift = channel * 4;
-	val = readl(smfc->base + SMFC_BS);
+	shift = smfc->chno * 4;
+	val = readl(priv->base + SMFC_BS);
 	val &= ~(0xf << shift);
 	val |= burstsize << shift;
-	writel(val, smfc->base + SMFC_BS);
+	writel(val, priv->base + SMFC_BS);
 
-	spin_unlock_irqrestore(&smfc->lock, flags);
+	spin_unlock_irqrestore(&priv->lock, flags);
 
 	return 0;
 }
 EXPORT_SYMBOL_GPL(ipu_smfc_set_burstsize);
 
-int ipu_smfc_map_channel(struct ipu_soc *ipu, int channel, int csi_id, int mipi_id)
+int ipu_smfc_map_channel(struct ipu_smfc *smfc, int csi_id, int mipi_id)
 {
-	struct ipu_smfc_priv *smfc = ipu->smfc_priv;
+	struct ipu_smfc_priv *priv = smfc->priv;
 	unsigned long flags;
 	u32 val, shift;
 
-	spin_lock_irqsave(&smfc->lock, flags);
+	spin_lock_irqsave(&priv->lock, flags);
 
-	shift = channel * 3;
-	val = readl(smfc->base + SMFC_MAP);
+	shift = smfc->chno * 3;
+	val = readl(priv->base + SMFC_MAP);
 	val &= ~(0x7 << shift);
 	val |= ((csi_id << 2) | mipi_id) << shift;
-	writel(val, smfc->base + SMFC_MAP);
+	writel(val, priv->base + SMFC_MAP);
 
-	spin_unlock_irqrestore(&smfc->lock, flags);
+	spin_unlock_irqrestore(&priv->lock, flags);
 
 	return 0;
 }
 EXPORT_SYMBOL_GPL(ipu_smfc_map_channel);
 
+int ipu_smfc_set_watermark(struct ipu_smfc *smfc, u32 set_level, u32 clr_level)
+{
+	struct ipu_smfc_priv *priv = smfc->priv;
+	unsigned long flags;
+	u32 val, shift;
+
+	spin_lock_irqsave(&priv->lock, flags);
+
+	shift = smfc->chno * 6 + (smfc->chno > 1 ? 4 : 0);
+	val = readl(priv->base + SMFC_WMC);
+	val &= ~(0x3f << shift);
+	val |= ((clr_level << 3) | set_level) << shift;
+	writel(val, priv->base + SMFC_WMC);
+
+	spin_unlock_irqrestore(&priv->lock, flags);
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(ipu_smfc_set_watermark);
+
+int ipu_smfc_enable(struct ipu_smfc *smfc)
+{
+	struct ipu_smfc_priv *priv = smfc->priv;
+	unsigned long flags;
+
+	spin_lock_irqsave(&priv->lock, flags);
+
+	if (!priv->use_count)
+		ipu_module_enable(priv->ipu, IPU_CONF_SMFC_EN);
+
+	priv->use_count++;
+
+	spin_unlock_irqrestore(&priv->lock, flags);
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(ipu_smfc_enable);
+
+int ipu_smfc_disable(struct ipu_smfc *smfc)
+{
+	struct ipu_smfc_priv *priv = smfc->priv;
+	unsigned long flags;
+
+	spin_lock_irqsave(&priv->lock, flags);
+
+	priv->use_count--;
+
+	if (!priv->use_count)
+		ipu_module_disable(priv->ipu, IPU_CONF_SMFC_EN);
+
+	if (priv->use_count < 0)
+		priv->use_count = 0;
+
+	spin_unlock_irqrestore(&priv->lock, flags);
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(ipu_smfc_disable);
+
+struct ipu_smfc *ipu_smfc_get(struct ipu_soc *ipu, unsigned int chno)
+{
+	struct ipu_smfc_priv *priv = ipu->smfc_priv;
+	struct ipu_smfc *smfc, *ret;
+	unsigned long flags;
+
+	if (chno >= 4)
+		return ERR_PTR(-EINVAL);
+
+	smfc = &priv->channel[chno];
+	ret = smfc;
+
+	spin_lock_irqsave(&priv->lock, flags);
+
+	if (smfc->inuse) {
+		ret = ERR_PTR(-EBUSY);
+		goto unlock;
+	}
+
+	smfc->inuse = true;
+unlock:
+	spin_unlock_irqrestore(&priv->lock, flags);
+	return ret;
+}
+EXPORT_SYMBOL_GPL(ipu_smfc_get);
+
+void ipu_smfc_put(struct ipu_smfc *smfc)
+{
+	struct ipu_smfc_priv *priv = smfc->priv;
+	unsigned long flags;
+
+	spin_lock_irqsave(&priv->lock, flags);
+	smfc->inuse = false;
+	spin_unlock_irqrestore(&priv->lock, flags);
+}
+EXPORT_SYMBOL_GPL(ipu_smfc_put);
+
 int ipu_smfc_init(struct ipu_soc *ipu, struct device *dev,
 		  unsigned long base)
 {
-	struct ipu_smfc_priv *smfc;
+	struct ipu_smfc_priv *priv;
+	int i;
 
-	smfc = devm_kzalloc(dev, sizeof(*smfc), GFP_KERNEL);
-	if (!smfc)
+	priv = devm_kzalloc(dev, sizeof(*priv), GFP_KERNEL);
+	if (!priv)
 		return -ENOMEM;
 
-	ipu->smfc_priv = smfc;
-	spin_lock_init(&smfc->lock);
+	ipu->smfc_priv = priv;
+	spin_lock_init(&priv->lock);
+	priv->ipu = ipu;
 
-	smfc->base = devm_ioremap(dev, base, PAGE_SIZE);
-	if (!smfc->base)
+	priv->base = devm_ioremap(dev, base, PAGE_SIZE);
+	if (!priv->base)
 		return -ENOMEM;
 
-	pr_debug("%s: ioremap 0x%08lx -> %p\n", __func__, base, smfc->base);
+	for (i = 0; i < 4; i++) {
+		priv->channel[i].priv = priv;
+		priv->channel[i].chno = i;
+	}
+
+	pr_debug("%s: ioremap 0x%08lx -> %p\n", __func__, base, priv->base);
 
 	return 0;
 }
diff --git a/drivers/gpu/vga/vgaarb.c b/drivers/gpu/vga/vgaarb.c
index 77711623b973..7bcbf863656e 100644
--- a/drivers/gpu/vga/vgaarb.c
+++ b/drivers/gpu/vga/vgaarb.c
@@ -400,7 +400,6 @@ int vga_get(struct pci_dev *pdev, unsigned int rsrc, int interruptible)
 		}
 		schedule();
 		remove_wait_queue(&vga_wait_queue, &wait);
-		set_current_state(TASK_RUNNING);
 	}
 	return rc;
 }