summaryrefslogtreecommitdiff
path: root/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
diff options
context:
space:
mode:
Diffstat (limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c')
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c455
1 files changed, 315 insertions, 140 deletions
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
index 38686203bea6..2dfbddcef9ab 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
@@ -23,7 +23,7 @@
*/
#include <drm/amdgpu_drm.h>
-#include <drm/drm_client_setup.h>
+#include <drm/clients/drm_client_setup.h>
#include <drm/drm_drv.h>
#include <drm/drm_fbdev_ttm.h>
#include <drm/drm_gem.h>
@@ -51,6 +51,8 @@
#include "amdgpu_reset.h"
#include "amdgpu_sched.h"
#include "amdgpu_xgmi.h"
+#include "amdgpu_userq.h"
+#include "amdgpu_userq_fence.h"
#include "../amdxcp/amdgpu_xcp_drv.h"
/*
@@ -119,9 +121,14 @@
* - 3.57.0 - Compute tunneling on GFX10+
* - 3.58.0 - Add GFX12 DCC support
* - 3.59.0 - Cleared VRAM
+ * - 3.60.0 - Add AMDGPU_TILING_GFX12_DCC_WRITE_COMPRESS_DISABLE (Vulkan requirement)
+ * - 3.61.0 - Contains fix for RV/PCO compute queues
+ * - 3.62.0 - Add AMDGPU_IDS_FLAGS_MODE_PF, AMDGPU_IDS_FLAGS_MODE_VF & AMDGPU_IDS_FLAGS_MODE_PT
+ * - 3.63.0 - GFX12 display DCC supports 256B max compressed block size
+ * - 3.64.0 - Userq IP support query
*/
#define KMS_DRIVER_MAJOR 3
-#define KMS_DRIVER_MINOR 59
+#define KMS_DRIVER_MINOR 64
#define KMS_DRIVER_PATCHLEVEL 0
/*
@@ -134,6 +141,11 @@ enum AMDGPU_DEBUG_MASK {
AMDGPU_DEBUG_USE_VRAM_FW_BUF = BIT(3),
AMDGPU_DEBUG_ENABLE_RAS_ACA = BIT(4),
AMDGPU_DEBUG_ENABLE_EXP_RESETS = BIT(5),
+ AMDGPU_DEBUG_DISABLE_GPU_RING_RESET = BIT(6),
+ AMDGPU_DEBUG_SMU_POOL = BIT(7),
+ AMDGPU_DEBUG_VM_USERPTR = BIT(8),
+ AMDGPU_DEBUG_DISABLE_RAS_CE_LOG = BIT(9),
+ AMDGPU_DEBUG_ENABLE_CE_CS = BIT(10)
};
unsigned int amdgpu_vram_limit = UINT_MAX;
@@ -170,7 +182,8 @@ uint amdgpu_pg_mask = 0xffffffff;
uint amdgpu_sdma_phase_quantum = 32;
char *amdgpu_disable_cu;
char *amdgpu_virtual_display;
-bool enforce_isolation;
+int amdgpu_enforce_isolation = -1;
+int amdgpu_modeset = -1;
/* Specifies the default granularity for SVM, used in buffer
* migration and restoration of backing memory when handling
@@ -231,6 +244,8 @@ int amdgpu_agp = -1; /* auto */
int amdgpu_wbrf = -1;
int amdgpu_damage_clips = -1; /* auto */
int amdgpu_umsch_mm_fwlog;
+int amdgpu_rebar = -1; /* auto */
+int amdgpu_user_queue = -1;
DECLARE_DYNDBG_CLASSMAP(drm_debug_classes, DD_CLASS_TYPE_DISJOINT_BITS, 0,
"DRM_UT_CORE",
@@ -280,7 +295,8 @@ module_param_named(gartsize, amdgpu_gart_size, uint, 0600);
/**
* DOC: gttsize (int)
* Restrict the size of GTT domain (for userspace use) in MiB for testing.
- * The default is -1 (Use 1/2 RAM, minimum value is 3GB).
+ * The default is -1 (Use value specified by TTM).
+ * This parameter is deprecated and will be removed in the future.
*/
MODULE_PARM_DESC(gttsize, "Size of the GTT userspace domain in megabytes (-1 = auto)");
module_param_named(gttsize, amdgpu_gtt_size, int, 0600);
@@ -296,7 +312,7 @@ module_param_named(moverate, amdgpu_moverate, int, 0600);
* DOC: audio (int)
* Set HDMI/DPAudio. Only affects non-DC display handling. The default is -1 (Enabled), set 0 to disabled it.
*/
-MODULE_PARM_DESC(audio, "Audio enable (-1 = auto, 0 = disable, 1 = enable)");
+MODULE_PARM_DESC(audio, "HDMI/DP Audio enable for non DC displays (-1 = auto, 0 = disable, 1 = enable)");
module_param_named(audio, amdgpu_audio, int, 0444);
/**
@@ -338,22 +354,16 @@ module_param_named(svm_default_granularity, amdgpu_svm_default_granularity, uint
* DOC: lockup_timeout (string)
* Set GPU scheduler timeout value in ms.
*
- * The format can be [Non-Compute] or [GFX,Compute,SDMA,Video]. That is there can be one or
- * multiple values specified. 0 and negative values are invalidated. They will be adjusted
- * to the default timeout.
+ * The format can be [single value] for setting all timeouts at once or
+ * [GFX,Compute,SDMA,Video] to set individual timeouts.
+ * Negative values mean infinity.
*
- * - With one value specified, the setting will apply to all non-compute jobs.
- * - With multiple values specified, the first one will be for GFX.
- * The second one is for Compute. The third and fourth ones are
- * for SDMA and Video.
- *
- * By default(with no lockup_timeout settings), the timeout for all non-compute(GFX, SDMA and Video)
- * jobs is 10000. The timeout for compute is 60000.
+ * By default(with no lockup_timeout settings), the timeout for all queues is 2000.
*/
-MODULE_PARM_DESC(lockup_timeout, "GPU lockup timeout in ms (default: for bare metal 10000 for non-compute jobs and 60000 for compute jobs; "
- "for passthrough or sriov, 10000 for all jobs. 0: keep default value. negative: infinity timeout), format: for bare metal [Non-Compute] or [GFX,Compute,SDMA,Video]; "
- "for passthrough or sriov [all jobs] or [GFX,Compute,SDMA,Video].");
-module_param_string(lockup_timeout, amdgpu_lockup_timeout, sizeof(amdgpu_lockup_timeout), 0444);
+MODULE_PARM_DESC(lockup_timeout,
+ "GPU lockup timeout in ms (default: 2000. 0: keep default value. negative: infinity timeout), format: [single value for all] or [GFX,Compute,SDMA,Video].");
+module_param_string(lockup_timeout, amdgpu_lockup_timeout,
+ sizeof(amdgpu_lockup_timeout), 0444);
/**
* DOC: dpm (int)
@@ -399,7 +409,7 @@ module_param_named(runpm, amdgpu_runtime_pm, int, 0444);
* the kernel log for the list of IPs on the asic. The default is 0xffffffff (enable all blocks on a device).
*/
MODULE_PARM_DESC(ip_block_mask, "IP Block Mask (all blocks enabled (default))");
-module_param_named(ip_block_mask, amdgpu_ip_block_mask, uint, 0444);
+module_param_named_unsafe(ip_block_mask, amdgpu_ip_block_mask, uint, 0444);
/**
* DOC: bapm (int)
@@ -457,7 +467,7 @@ module_param_named(vm_update_mode, amdgpu_vm_update_mode, int, 0444);
* Enable experimental hw support (1 = enable). The default is 0 (disabled).
*/
MODULE_PARM_DESC(exp_hw_support, "experimental hw support (1 = enable, 0 = disable (default))");
-module_param_named(exp_hw_support, amdgpu_exp_hw_support, int, 0444);
+module_param_named_unsafe(exp_hw_support, amdgpu_exp_hw_support, int, 0444);
/**
* DOC: dc (int)
@@ -568,14 +578,14 @@ module_param_named(compute_multipipe, amdgpu_compute_multipipe, int, 0444);
* Set to enable GPU recovery mechanism (1 = enable, 0 = disable). The default is -1 (auto, disabled except SRIOV).
*/
MODULE_PARM_DESC(gpu_recovery, "Enable GPU recovery mechanism, (1 = enable, 0 = disable, -1 = auto)");
-module_param_named(gpu_recovery, amdgpu_gpu_recovery, int, 0444);
+module_param_named_unsafe(gpu_recovery, amdgpu_gpu_recovery, int, 0444);
/**
* DOC: emu_mode (int)
* Set value 1 to enable emulation mode. This is only needed when running on an emulator. The default is 0 (disabled).
*/
MODULE_PARM_DESC(emu_mode, "Emulation mode, (1 = enable, 0 = disable)");
-module_param_named(emu_mode, amdgpu_emu_mode, int, 0444);
+module_param_named_unsafe(emu_mode, amdgpu_emu_mode, int, 0444);
/**
* DOC: ras_enable (int)
@@ -608,39 +618,39 @@ module_param_named(timeout_period, amdgpu_watchdog_timer.period, uint, 0644);
/**
* DOC: si_support (int)
- * Set SI support driver. This parameter works after set config CONFIG_DRM_AMDGPU_SI. For SI asic, when radeon driver is enabled,
- * set value 0 to use radeon driver, while set value 1 to use amdgpu driver. The default is using radeon driver when it available,
- * otherwise using amdgpu driver.
- */
+ * 1 = enabled, 0 = disabled, -1 = default
+ *
+ * SI (Southern Islands) are first generation GCN GPUs, supported by both
+ * drivers: radeon (old) and amdgpu (new). This parameter controls whether
+ * amdgpu should support SI.
+ * By default, SI dedicated GPUs are supported by amdgpu.
+ * Only relevant when CONFIG_DRM_AMDGPU_SI is enabled to build SI support in amdgpu.
+ * See also radeon.si_support which should be disabled when amdgpu.si_support is
+ * enabled, and vice versa.
+ */
+int amdgpu_si_support = -1;
#ifdef CONFIG_DRM_AMDGPU_SI
-
-#if IS_ENABLED(CONFIG_DRM_RADEON) || IS_ENABLED(CONFIG_DRM_RADEON_MODULE)
-int amdgpu_si_support;
-MODULE_PARM_DESC(si_support, "SI support (1 = enabled, 0 = disabled (default))");
-#else
-int amdgpu_si_support = 1;
-MODULE_PARM_DESC(si_support, "SI support (1 = enabled (default), 0 = disabled)");
-#endif
-
+MODULE_PARM_DESC(si_support, "SI support (1 = enabled, 0 = disabled, -1 = default)");
module_param_named(si_support, amdgpu_si_support, int, 0444);
#endif
/**
* DOC: cik_support (int)
- * Set CIK support driver. This parameter works after set config CONFIG_DRM_AMDGPU_CIK. For CIK asic, when radeon driver is enabled,
- * set value 0 to use radeon driver, while set value 1 to use amdgpu driver. The default is using radeon driver when it available,
- * otherwise using amdgpu driver.
- */
+ * 1 = enabled, 0 = disabled, -1 = default
+ *
+ * CIK (Sea Islands) are second generation GCN GPUs, supported by both
+ * drivers: radeon (old) and amdgpu (new). This parameter controls whether
+ * amdgpu should support CIK.
+ * By default:
+ * - CIK dedicated GPUs are supported by amdgpu.
+ * - CIK APUs are supported by radeon (except when radeon is not built).
+ * Only relevant when CONFIG_DRM_AMDGPU_CIK is enabled to build CIK support in amdgpu.
+ * See also radeon.cik_support which should be disabled when amdgpu.cik_support is
+ * enabled, and vice versa.
+ */
+int amdgpu_cik_support = -1;
#ifdef CONFIG_DRM_AMDGPU_CIK
-
-#if IS_ENABLED(CONFIG_DRM_RADEON) || IS_ENABLED(CONFIG_DRM_RADEON_MODULE)
-int amdgpu_cik_support;
-MODULE_PARM_DESC(cik_support, "CIK support (1 = enabled, 0 = disabled (default))");
-#else
-int amdgpu_cik_support = 1;
-MODULE_PARM_DESC(cik_support, "CIK support (1 = enabled (default), 0 = disabled)");
-#endif
-
+MODULE_PARM_DESC(cik_support, "CIK support (1 = enabled, 0 = disabled, -1 = default)");
module_param_named(cik_support, amdgpu_cik_support, int, 0444);
#endif
@@ -730,7 +740,7 @@ module_param_named(noretry, amdgpu_noretry, int, 0644);
*/
MODULE_PARM_DESC(force_asic_type,
"A non negative value used to specify the asic type for all supported GPUs");
-module_param_named(force_asic_type, amdgpu_force_asic_type, int, 0444);
+module_param_named_unsafe(force_asic_type, amdgpu_force_asic_type, int, 0444);
/**
* DOC: use_xgmi_p2p (int)
@@ -749,7 +759,7 @@ module_param_named(use_xgmi_p2p, amdgpu_use_xgmi_p2p, int, 0444);
* assigns queues to HQDs.
*/
int sched_policy = KFD_SCHED_POLICY_HWS;
-module_param(sched_policy, int, 0444);
+module_param_unsafe(sched_policy, int, 0444);
MODULE_PARM_DESC(sched_policy,
"Scheduling policy (0 = HWS (Default), 1 = HWS without over-subscription, 2 = Non-HWS (Used for debugging only)");
@@ -799,7 +809,7 @@ MODULE_PARM_DESC(send_sigterm,
* Setting 1 enables halt on hang.
*/
int halt_if_hws_hang;
-module_param(halt_if_hws_hang, int, 0644);
+module_param_unsafe(halt_if_hws_hang, int, 0644);
MODULE_PARM_DESC(halt_if_hws_hang, "Halt if HWS hang is detected (0 = off (default), 1 = on)");
/**
@@ -808,7 +818,7 @@ MODULE_PARM_DESC(halt_if_hws_hang, "Halt if HWS hang is detected (0 = off (defau
* check says. Default value: false (rely on MEC2 firmware version check).
*/
bool hws_gws_support;
-module_param(hws_gws_support, bool, 0444);
+module_param_unsafe(hws_gws_support, bool, 0444);
MODULE_PARM_DESC(hws_gws_support, "Assume MEC2 FW supports GWS barriers (false = rely on FW version check (Default), true = force supported)");
/**
@@ -841,7 +851,7 @@ MODULE_PARM_DESC(no_system_mem_limit, "disable system memory limit (false = defa
*/
int amdgpu_no_queue_eviction_on_vm_fault;
MODULE_PARM_DESC(no_queue_eviction_on_vm_fault, "No queue eviction on VM fault (0 = queue eviction, 1 = no queue eviction)");
-module_param_named(no_queue_eviction_on_vm_fault, amdgpu_no_queue_eviction_on_vm_fault, int, 0444);
+module_param_named_unsafe(no_queue_eviction_on_vm_fault, amdgpu_no_queue_eviction_on_vm_fault, int, 0444);
#endif
/**
@@ -849,7 +859,7 @@ module_param_named(no_queue_eviction_on_vm_fault, amdgpu_no_queue_eviction_on_vm
*/
int amdgpu_mtype_local;
MODULE_PARM_DESC(mtype_local, "MTYPE for local memory (0 = MTYPE_RW (default), 1 = MTYPE_NC, 2 = MTYPE_CC)");
-module_param_named(mtype_local, amdgpu_mtype_local, int, 0444);
+module_param_named_unsafe(mtype_local, amdgpu_mtype_local, int, 0444);
/**
* DOC: pcie_p2p (bool)
@@ -871,7 +881,7 @@ module_param_named(dcfeaturemask, amdgpu_dc_feature_mask, uint, 0444);
/**
* DOC: dcdebugmask (uint)
- * Override display features enabled. See enum DC_DEBUG_MASK in drivers/gpu/drm/amd/include/amd_shared.h.
+ * Display debug options. See enum DC_DEBUG_MASK in drivers/gpu/drm/amd/include/amd_shared.h.
*/
MODULE_PARM_DESC(dcdebugmask, "all debug options disabled (default))");
module_param_named(dcdebugmask, amdgpu_dc_debug_mask, uint, 0444);
@@ -945,7 +955,7 @@ module_param_named(tmz, amdgpu_tmz, int, 0444);
*/
MODULE_PARM_DESC(
freesync_video,
- "Enable freesync modesetting optimization feature (0 = off (default), 1 = on)");
+ "Adds additional modes via VRR for refresh changes without a full modeset (0 = off (default), 1 = on)");
module_param_named(freesync_video, amdgpu_freesync_vid_mode, uint, 0444);
/**
@@ -953,7 +963,7 @@ module_param_named(freesync_video, amdgpu_freesync_vid_mode, uint, 0444);
* GPU reset method (-1 = auto (default), 0 = legacy, 1 = mode0, 2 = mode1, 3 = mode2, 4 = baco)
*/
MODULE_PARM_DESC(reset_method, "GPU reset method (-1 = auto (default), 0 = legacy, 1 = mode0, 2 = mode1, 3 = mode2, 4 = baco/bamaco)");
-module_param_named(reset_method, amdgpu_reset_method, int, 0644);
+module_param_named_unsafe(reset_method, amdgpu_reset_method, int, 0644);
/**
* DOC: bad_page_threshold (int) Bad page threshold is specifies the
@@ -961,7 +971,7 @@ module_param_named(reset_method, amdgpu_reset_method, int, 0644);
* result in the GPU entering bad status when the number of total
* faulty pages by ECC exceeds the threshold value.
*/
-MODULE_PARM_DESC(bad_page_threshold, "Bad page threshold(-1 = ignore threshold (default value), 0 = disable bad page retirement, -2 = driver sets threshold)");
+MODULE_PARM_DESC(bad_page_threshold, "Bad page threshold(-1 = ignore threshold (default value), 0 = disable bad page retirement, -2 = threshold determined by a formula, 0 < threshold < max records, user-defined threshold)");
module_param_named(bad_page_threshold, amdgpu_bad_page_threshold, int, 0444);
MODULE_PARM_DESC(num_kcq, "number of kernel compute queue user want to setup (8 if set to greater than 8 or less than 0, only affect gfx 8+)");
@@ -1025,11 +1035,20 @@ module_param_named(user_partt_mode, amdgpu_user_partt_mode, uint, 0444);
/**
- * DOC: enforce_isolation (bool)
- * enforce process isolation between graphics and compute via using the same reserved vmid.
+ * DOC: enforce_isolation (int)
+ * enforce process isolation between graphics and compute.
+ * (-1 = auto, 0 = disable, 1 = enable, 2 = enable legacy mode, 3 = enable without cleaner shader)
*/
-module_param(enforce_isolation, bool, 0444);
-MODULE_PARM_DESC(enforce_isolation, "enforce process isolation between graphics and compute . enforce_isolation = on");
+module_param_named(enforce_isolation, amdgpu_enforce_isolation, int, 0444);
+MODULE_PARM_DESC(enforce_isolation,
+"enforce process isolation between graphics and compute. (-1 = auto, 0 = disable, 1 = enable, 2 = enable legacy mode, 3 = enable without cleaner shader)");
+
+/**
+ * DOC: modeset (int)
+ * Override nomodeset (1 = override, -1 = auto). The default is -1 (auto).
+ */
+MODULE_PARM_DESC(modeset, "Override nomodeset (1 = enable, -1 = auto)");
+module_param_named(modeset, amdgpu_modeset, int, 0444);
/**
* DOC: seamless (int)
@@ -1047,9 +1066,14 @@ module_param_named(seamless, amdgpu_seamless, int, 0444);
* limits the VRAM size reported to ROCm applications to the visible
* size, usually 256MB.
* - 0x4: Disable GPU soft recovery, always do a full reset
+ * - 0x8: Use VRAM for firmware loading
+ * - 0x10: Enable ACA based RAS logging
+ * - 0x20: Enable experimental resets
+ * - 0x40: Disable ring resets
+ * - 0x80: Use VRAM for SMU pool
*/
MODULE_PARM_DESC(debug_mask, "debug options for amdgpu, disabled by default");
-module_param_named(debug_mask, amdgpu_debug_mask, uint, 0444);
+module_param_named_unsafe(debug_mask, amdgpu_debug_mask, uint, 0444);
/**
* DOC: agp (int)
@@ -1076,6 +1100,28 @@ MODULE_PARM_DESC(wbrf,
"Enable Wifi RFI interference mitigation (0 = disabled, 1 = enabled, -1 = auto(default)");
module_param_named(wbrf, amdgpu_wbrf, int, 0444);
+/**
+ * DOC: rebar (int)
+ * Allow BAR resizing. Disable this to prevent the driver from attempting
+ * to resize the BAR if the GPU supports it and there is available MMIO space.
+ * Note that this just prevents the driver from resizing the BAR. The BIOS
+ * may have already resized the BAR at boot time.
+ */
+MODULE_PARM_DESC(rebar, "Resizable BAR (-1 = auto (default), 0 = disable, 1 = enable)");
+module_param_named(rebar, amdgpu_rebar, int, 0444);
+
+/**
+ * DOC: user_queue (int)
+ * Enable user queues on systems that support user queues. Possible values:
+ *
+ * - -1 = auto (ASIC specific default)
+ * - 0 = user queues disabled
+ * - 1 = user queues enabled and kernel queues enabled (if supported)
+ * - 2 = user queues enabled and kernel queues disabled
+ */
+MODULE_PARM_DESC(user_queue, "Enable user queues (-1 = auto (default), 0 = disable, 1 = enable, 2 = enable UQs and disable KQs)");
+module_param_named(user_queue, amdgpu_user_queue, int, 0444);
+
/* These devices are not supported by amdgpu.
* They are supported by the mach64, r128, radeon drivers
*/
@@ -1789,7 +1835,6 @@ static const u16 amdgpu_unsupported_pciidlist[] = {
};
static const struct pci_device_id pciidlist[] = {
-#ifdef CONFIG_DRM_AMDGPU_SI
{0x1002, 0x6780, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_TAHITI},
{0x1002, 0x6784, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_TAHITI},
{0x1002, 0x6788, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_TAHITI},
@@ -1862,8 +1907,6 @@ static const struct pci_device_id pciidlist[] = {
{0x1002, 0x6665, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_HAINAN|AMD_IS_MOBILITY},
{0x1002, 0x6667, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_HAINAN|AMD_IS_MOBILITY},
{0x1002, 0x666F, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_HAINAN|AMD_IS_MOBILITY},
-#endif
-#ifdef CONFIG_DRM_AMDGPU_CIK
/* Kaveri */
{0x1002, 0x1304, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_KAVERI|AMD_IS_MOBILITY|AMD_IS_APU},
{0x1002, 0x1305, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_KAVERI|AMD_IS_APU},
@@ -1946,7 +1989,6 @@ static const struct pci_device_id pciidlist[] = {
{0x1002, 0x985D, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_MULLINS|AMD_IS_MOBILITY|AMD_IS_APU},
{0x1002, 0x985E, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_MULLINS|AMD_IS_MOBILITY|AMD_IS_APU},
{0x1002, 0x985F, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_MULLINS|AMD_IS_MOBILITY|AMD_IS_APU},
-#endif
/* topaz */
{0x1002, 0x6900, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_TOPAZ},
{0x1002, 0x6901, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_TOPAZ},
@@ -2125,6 +2167,11 @@ static const struct pci_device_id pciidlist[] = {
{0x1002, 0x7410, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_ALDEBARAN},
/* CYAN_SKILLFISH */
+ {0x1002, 0x13DB, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_CYAN_SKILLFISH|AMD_IS_APU},
+ {0x1002, 0x13F9, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_CYAN_SKILLFISH|AMD_IS_APU},
+ {0x1002, 0x13FA, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_CYAN_SKILLFISH|AMD_IS_APU},
+ {0x1002, 0x13FB, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_CYAN_SKILLFISH|AMD_IS_APU},
+ {0x1002, 0x13FC, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_CYAN_SKILLFISH|AMD_IS_APU},
{0x1002, 0x13FE, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_CYAN_SKILLFISH|AMD_IS_APU},
{0x1002, 0x143F, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_CYAN_SKILLFISH|AMD_IS_APU},
@@ -2181,7 +2228,6 @@ static void amdgpu_get_secondary_funcs(struct amdgpu_device *adev)
adev->pdev->bus->number, i);
if (p) {
pm_runtime_get_sync(&p->dev);
- pm_runtime_mark_last_busy(&p->dev);
pm_runtime_put_autosuspend(&p->dev);
pci_dev_put(p);
}
@@ -2219,6 +2265,29 @@ static void amdgpu_init_debug_options(struct amdgpu_device *adev)
pr_info("debug: enable experimental reset features\n");
adev->debug_exp_resets = true;
}
+
+ if (amdgpu_debug_mask & AMDGPU_DEBUG_DISABLE_GPU_RING_RESET) {
+ pr_info("debug: ring reset disabled\n");
+ adev->debug_disable_gpu_ring_reset = true;
+ }
+ if (amdgpu_debug_mask & AMDGPU_DEBUG_SMU_POOL) {
+ pr_info("debug: use vram for smu pool\n");
+ adev->pm.smu_debug_mask |= SMU_DEBUG_POOL_USE_VRAM;
+ }
+ if (amdgpu_debug_mask & AMDGPU_DEBUG_VM_USERPTR) {
+ pr_info("debug: VM mode debug for userptr is enabled\n");
+ adev->debug_vm_userptr = true;
+ }
+
+ if (amdgpu_debug_mask & AMDGPU_DEBUG_DISABLE_RAS_CE_LOG) {
+ pr_info("debug: disable kernel logs of correctable errors\n");
+ adev->debug_disable_ce_logs = true;
+ }
+
+ if (amdgpu_debug_mask & AMDGPU_DEBUG_ENABLE_CE_CS) {
+ pr_info("debug: allowing command submission to CE engine\n");
+ adev->debug_enable_ce_cs = true;
+ }
}
static unsigned long amdgpu_fix_asic_type(struct pci_dev *pdev, unsigned long flags)
@@ -2237,6 +2306,72 @@ static unsigned long amdgpu_fix_asic_type(struct pci_dev *pdev, unsigned long fl
return flags;
}
+static bool amdgpu_support_enabled(struct device *dev,
+ const enum amd_asic_type family)
+{
+ const char *gen;
+ const char *param;
+ int module_param = -1;
+ bool radeon_support_built = IS_ENABLED(CONFIG_DRM_RADEON);
+ bool amdgpu_support_built = false;
+ bool support_by_default = false;
+
+ switch (family) {
+ case CHIP_TAHITI:
+ case CHIP_PITCAIRN:
+ case CHIP_VERDE:
+ case CHIP_OLAND:
+ case CHIP_HAINAN:
+ gen = "SI";
+ param = "si_support";
+ module_param = amdgpu_si_support;
+ amdgpu_support_built = IS_ENABLED(CONFIG_DRM_AMDGPU_SI);
+ support_by_default = true;
+ break;
+
+ case CHIP_BONAIRE:
+ case CHIP_HAWAII:
+ support_by_default = true;
+ fallthrough;
+ case CHIP_KAVERI:
+ case CHIP_KABINI:
+ case CHIP_MULLINS:
+ gen = "CIK";
+ param = "cik_support";
+ module_param = amdgpu_cik_support;
+ amdgpu_support_built = IS_ENABLED(CONFIG_DRM_AMDGPU_CIK);
+ break;
+
+ default:
+ /* All other chips are supported by amdgpu only */
+ return true;
+ }
+
+ if (!amdgpu_support_built) {
+ dev_info(dev, "amdgpu built without %s support\n", gen);
+ return false;
+ }
+
+ if ((module_param == -1 && (support_by_default || !radeon_support_built)) ||
+ module_param == 1) {
+ if (radeon_support_built)
+ dev_info(dev, "%s support provided by amdgpu.\n"
+ "Use radeon.%s=1 amdgpu.%s=0 to override.\n",
+ gen, param, param);
+
+ return true;
+ }
+
+ if (radeon_support_built)
+ dev_info(dev, "%s support provided by radeon.\n"
+ "Use radeon.%s=0 amdgpu.%s=1 to override.\n",
+ gen, param, param);
+ else if (module_param == 0)
+ dev_info(dev, "%s support disabled by module param\n", gen);
+
+ return false;
+}
+
static int amdgpu_pci_probe(struct pci_dev *pdev,
const struct pci_device_id *ent)
{
@@ -2246,6 +2381,12 @@ static int amdgpu_pci_probe(struct pci_dev *pdev,
int ret, retry = 0, i;
bool supports_atomic = false;
+ if ((pdev->class >> 8) == PCI_CLASS_DISPLAY_VGA ||
+ (pdev->class >> 8) == PCI_CLASS_DISPLAY_OTHER) {
+ if (drm_firmware_drivers_only() && amdgpu_modeset == -1)
+ return -EINVAL;
+ }
+
/* skip devices which are owned by radeon */
for (i = 0; i < ARRAY_SIZE(amdgpu_unsupported_pciidlist); i++) {
if (amdgpu_unsupported_pciidlist[i] == pdev->device)
@@ -2256,7 +2397,7 @@ static int amdgpu_pci_probe(struct pci_dev *pdev,
amdgpu_aspm = 0;
if (amdgpu_virtual_display ||
- amdgpu_device_asic_has_dc_support(flags & AMD_ASIC_MASK))
+ amdgpu_device_asic_has_dc_support(pdev, flags & AMD_ASIC_MASK))
supports_atomic = true;
if ((flags & AMD_EXP_HW_SUPPORT) && !amdgpu_exp_hw_support) {
@@ -2278,40 +2419,8 @@ static int amdgpu_pci_probe(struct pci_dev *pdev,
return -ENOTSUPP;
}
-#ifdef CONFIG_DRM_AMDGPU_SI
- if (!amdgpu_si_support) {
- switch (flags & AMD_ASIC_MASK) {
- case CHIP_TAHITI:
- case CHIP_PITCAIRN:
- case CHIP_VERDE:
- case CHIP_OLAND:
- case CHIP_HAINAN:
- dev_info(&pdev->dev,
- "SI support provided by radeon.\n");
- dev_info(&pdev->dev,
- "Use radeon.si_support=0 amdgpu.si_support=1 to override.\n"
- );
- return -ENODEV;
- }
- }
-#endif
-#ifdef CONFIG_DRM_AMDGPU_CIK
- if (!amdgpu_cik_support) {
- switch (flags & AMD_ASIC_MASK) {
- case CHIP_KAVERI:
- case CHIP_BONAIRE:
- case CHIP_HAWAII:
- case CHIP_KABINI:
- case CHIP_MULLINS:
- dev_info(&pdev->dev,
- "CIK support provided by radeon.\n");
- dev_info(&pdev->dev,
- "Use radeon.cik_support=0 amdgpu.cik_support=1 to override.\n"
- );
- return -ENODEV;
- }
- }
-#endif
+ if (!amdgpu_support_enabled(&pdev->dev, flags & AMD_ASIC_MASK))
+ return -ENODEV;
adev = devm_drm_dev_alloc(&pdev->dev, &amdgpu_kms_driver, typeof(*adev), ddev);
if (IS_ERR(adev))
@@ -2378,10 +2487,10 @@ retry_init:
if (adev->pm.rpm_mode != AMDGPU_RUNPM_NONE) {
/* only need to skip on ATPX */
- if (amdgpu_device_supports_px(ddev))
+ if (amdgpu_device_supports_px(adev))
dev_pm_set_driver_flags(ddev->dev, DPM_FLAG_NO_DIRECT_COMPLETE);
/* we want direct complete for BOCO */
- if (amdgpu_device_supports_boco(ddev))
+ if (amdgpu_device_supports_boco(adev))
dev_pm_set_driver_flags(ddev->dev, DPM_FLAG_SMART_PREPARE |
DPM_FLAG_SMART_SUSPEND |
DPM_FLAG_MAY_SKIP_RESUME);
@@ -2390,7 +2499,6 @@ retry_init:
pm_runtime_allow(ddev->dev);
- pm_runtime_mark_last_busy(ddev->dev);
pm_runtime_put_autosuspend(ddev->dev);
pci_wake_from_d3(pdev, TRUE);
@@ -2414,9 +2522,9 @@ retry_init:
* into D0 state. Then there will be a PMFW-aware D-state
* transition(D0->D3) on runpm suspend.
*/
- if (amdgpu_device_supports_baco(ddev) &&
+ if (amdgpu_device_supports_baco(adev) &&
!(adev->flags & AMD_IS_APU) &&
- (adev->asic_type >= CHIP_NAVI10))
+ adev->asic_type >= CHIP_NAVI10)
amdgpu_get_secondary_funcs(adev);
}
@@ -2433,6 +2541,7 @@ amdgpu_pci_remove(struct pci_dev *pdev)
struct drm_device *dev = pci_get_drvdata(pdev);
struct amdgpu_device *adev = drm_to_adev(dev);
+ amdgpu_ras_eeprom_check_and_recover(adev);
amdgpu_xcp_dev_unplug(adev);
amdgpu_gmc_prepare_nps_mode_change(adev);
drm_dev_unplug(dev);
@@ -2462,6 +2571,10 @@ amdgpu_pci_shutdown(struct pci_dev *pdev)
if (amdgpu_ras_intr_triggered())
return;
+ /* device maybe not resumed here, return immediately in this case */
+ if (adev->in_s4 && adev->in_suspend)
+ return;
+
/* if we are running in a VM, make sure the device
* torn down properly on reboot/shutdown.
* unfortunately we can't detect certain
@@ -2469,7 +2582,8 @@ amdgpu_pci_shutdown(struct pci_dev *pdev)
*/
if (!amdgpu_passthrough(adev))
adev->mp1_state = PP_MP1_STATE_UNLOAD;
- amdgpu_device_ip_suspend(adev);
+ amdgpu_device_prepare(dev);
+ amdgpu_device_suspend(dev, true);
adev->mp1_state = PP_MP1_STATE_NONE;
}
@@ -2478,11 +2592,14 @@ static int amdgpu_pmops_prepare(struct device *dev)
struct drm_device *drm_dev = dev_get_drvdata(dev);
struct amdgpu_device *adev = drm_to_adev(drm_dev);
+ /* device maybe not resumed here, return immediately in this case */
+ if (adev->in_s4 && adev->in_suspend)
+ return 0;
+
/* Return a positive number here so
* DPM_FLAG_SMART_SUSPEND works properly
*/
- if (amdgpu_device_supports_boco(drm_dev) &&
- pm_runtime_suspended(dev))
+ if (amdgpu_device_supports_boco(adev) && pm_runtime_suspended(dev))
return 1;
/* if we will not support s3 or s2i for the device
@@ -2497,7 +2614,7 @@ static int amdgpu_pmops_prepare(struct device *dev)
static void amdgpu_pmops_complete(struct device *dev)
{
- /* nothing to do */
+ amdgpu_device_complete(dev_get_drvdata(dev));
}
static int amdgpu_pmops_suspend(struct device *dev)
@@ -2509,8 +2626,24 @@ static int amdgpu_pmops_suspend(struct device *dev)
adev->in_s0ix = true;
else if (amdgpu_acpi_is_s3_active(adev))
adev->in_s3 = true;
- if (!adev->in_s0ix && !adev->in_s3)
+ if (!adev->in_s0ix && !adev->in_s3) {
+#if IS_ENABLED(CONFIG_SUSPEND)
+ /* don't allow going deep first time followed by s2idle the next time */
+ if (adev->last_suspend_state != PM_SUSPEND_ON &&
+ adev->last_suspend_state != pm_suspend_target_state) {
+ drm_err_once(drm_dev, "Unsupported suspend state %d\n",
+ pm_suspend_target_state);
+ return -EINVAL;
+ }
+#endif
return 0;
+ }
+
+#if IS_ENABLED(CONFIG_SUSPEND)
+ /* cache the state last used for suspend */
+ adev->last_suspend_state = pm_suspend_target_state;
+#endif
+
return amdgpu_device_suspend(drm_dev, true);
}
@@ -2518,9 +2651,14 @@ static int amdgpu_pmops_suspend_noirq(struct device *dev)
{
struct drm_device *drm_dev = dev_get_drvdata(dev);
struct amdgpu_device *adev = drm_to_adev(drm_dev);
+ int r;
- if (amdgpu_acpi_should_gpu_reset(adev))
- return amdgpu_asic_reset(adev);
+ if (amdgpu_acpi_should_gpu_reset(adev)) {
+ amdgpu_device_lock_reset_domain(adev->reset_domain);
+ r = amdgpu_asic_reset(adev);
+ amdgpu_device_unlock_reset_domain(adev->reset_domain);
+ return r;
+ }
return 0;
}
@@ -2552,9 +2690,7 @@ static int amdgpu_pmops_freeze(struct device *dev)
struct amdgpu_device *adev = drm_to_adev(drm_dev);
int r;
- adev->in_s4 = true;
r = amdgpu_device_suspend(drm_dev, true);
- adev->in_s4 = false;
if (r)
return r;
@@ -2567,12 +2703,21 @@ static int amdgpu_pmops_thaw(struct device *dev)
{
struct drm_device *drm_dev = dev_get_drvdata(dev);
+ /* do not resume device if it's normal hibernation */
+ if (!pm_hibernate_is_recovering() && !pm_hibernation_mode_is_suspend())
+ return 0;
+
return amdgpu_device_resume(drm_dev, true);
}
static int amdgpu_pmops_poweroff(struct device *dev)
{
struct drm_device *drm_dev = dev_get_drvdata(dev);
+ struct amdgpu_device *adev = drm_to_adev(drm_dev);
+
+ /* device maybe not resumed here, return immediately in this case */
+ if (adev->in_s4 && adev->in_suspend)
+ return 0;
return amdgpu_device_suspend(drm_dev, true);
}
@@ -2651,6 +2796,15 @@ static int amdgpu_runtime_idle_check_display(struct device *dev)
return 0;
}
+static int amdgpu_runtime_idle_check_userq(struct device *dev)
+{
+ struct pci_dev *pdev = to_pci_dev(dev);
+ struct drm_device *drm_dev = pci_get_drvdata(pdev);
+ struct amdgpu_device *adev = drm_to_adev(drm_dev);
+
+ return xa_empty(&adev->userq_doorbell_xa) ? 0 : -EBUSY;
+}
+
static int amdgpu_pmops_runtime_suspend(struct device *dev)
{
struct pci_dev *pdev = to_pci_dev(dev);
@@ -2666,6 +2820,9 @@ static int amdgpu_pmops_runtime_suspend(struct device *dev)
ret = amdgpu_runtime_idle_check_display(dev);
if (ret)
return ret;
+ ret = amdgpu_runtime_idle_check_userq(dev);
+ if (ret)
+ return ret;
/* wait for all rings to drain before suspending */
for (i = 0; i < AMDGPU_MAX_RINGS; i++) {
@@ -2719,7 +2876,7 @@ static int amdgpu_pmops_runtime_suspend(struct device *dev)
/* nothing to do */
} else if ((adev->pm.rpm_mode == AMDGPU_RUNPM_BACO) ||
(adev->pm.rpm_mode == AMDGPU_RUNPM_BAMACO)) {
- amdgpu_device_baco_enter(drm_dev);
+ amdgpu_device_baco_enter(adev);
}
dev_dbg(&pdev->dev, "asic/device is runtime suspended\n");
@@ -2760,7 +2917,7 @@ static int amdgpu_pmops_runtime_resume(struct device *dev)
pci_set_master(pdev);
} else if ((adev->pm.rpm_mode == AMDGPU_RUNPM_BACO) ||
(adev->pm.rpm_mode == AMDGPU_RUNPM_BAMACO)) {
- amdgpu_device_baco_exit(drm_dev);
+ amdgpu_device_baco_exit(adev);
}
ret = amdgpu_device_resume(drm_dev, false);
if (ret) {
@@ -2787,12 +2944,32 @@ static int amdgpu_pmops_runtime_idle(struct device *dev)
}
ret = amdgpu_runtime_idle_check_display(dev);
+ if (ret)
+ goto done;
- pm_runtime_mark_last_busy(dev);
+ ret = amdgpu_runtime_idle_check_userq(dev);
+done:
pm_runtime_autosuspend(dev);
return ret;
}
+static int amdgpu_drm_release(struct inode *inode, struct file *filp)
+{
+ struct drm_file *file_priv = filp->private_data;
+ struct amdgpu_fpriv *fpriv = file_priv->driver_priv;
+ struct drm_device *dev = file_priv->minor->dev;
+ int idx;
+
+ if (fpriv && drm_dev_enter(dev, &idx)) {
+ fpriv->evf_mgr.fd_closing = true;
+ amdgpu_eviction_fence_destroy(&fpriv->evf_mgr);
+ amdgpu_userq_mgr_fini(&fpriv->userq_mgr);
+ drm_dev_exit(idx);
+ }
+
+ return drm_release(inode, filp);
+}
+
long amdgpu_drm_ioctl(struct file *filp,
unsigned int cmd, unsigned long arg)
{
@@ -2807,22 +2984,21 @@ long amdgpu_drm_ioctl(struct file *filp,
ret = drm_ioctl(filp, cmd, arg);
- pm_runtime_mark_last_busy(dev->dev);
out:
pm_runtime_put_autosuspend(dev->dev);
return ret;
}
static const struct dev_pm_ops amdgpu_pm_ops = {
- .prepare = amdgpu_pmops_prepare,
- .complete = amdgpu_pmops_complete,
- .suspend = amdgpu_pmops_suspend,
- .suspend_noirq = amdgpu_pmops_suspend_noirq,
- .resume = amdgpu_pmops_resume,
- .freeze = amdgpu_pmops_freeze,
- .thaw = amdgpu_pmops_thaw,
- .poweroff = amdgpu_pmops_poweroff,
- .restore = amdgpu_pmops_restore,
+ .prepare = pm_sleep_ptr(amdgpu_pmops_prepare),
+ .complete = pm_sleep_ptr(amdgpu_pmops_complete),
+ .suspend = pm_sleep_ptr(amdgpu_pmops_suspend),
+ .suspend_noirq = pm_sleep_ptr(amdgpu_pmops_suspend_noirq),
+ .resume = pm_sleep_ptr(amdgpu_pmops_resume),
+ .freeze = pm_sleep_ptr(amdgpu_pmops_freeze),
+ .thaw = pm_sleep_ptr(amdgpu_pmops_thaw),
+ .poweroff = pm_sleep_ptr(amdgpu_pmops_poweroff),
+ .restore = pm_sleep_ptr(amdgpu_pmops_restore),
.runtime_suspend = amdgpu_pmops_runtime_suspend,
.runtime_resume = amdgpu_pmops_runtime_resume,
.runtime_idle = amdgpu_pmops_runtime_idle,
@@ -2844,7 +3020,7 @@ static const struct file_operations amdgpu_driver_kms_fops = {
.owner = THIS_MODULE,
.open = drm_open,
.flush = amdgpu_flush,
- .release = drm_release,
+ .release = amdgpu_drm_release,
.unlocked_ioctl = amdgpu_drm_ioctl,
.mmap = drm_gem_mmap,
.poll = drm_poll,
@@ -2891,6 +3067,10 @@ const struct drm_ioctl_desc amdgpu_ioctls_kms[] = {
DRM_IOCTL_DEF_DRV(AMDGPU_GEM_VA, amdgpu_gem_va_ioctl, DRM_AUTH|DRM_RENDER_ALLOW),
DRM_IOCTL_DEF_DRV(AMDGPU_GEM_OP, amdgpu_gem_op_ioctl, DRM_AUTH|DRM_RENDER_ALLOW),
DRM_IOCTL_DEF_DRV(AMDGPU_GEM_USERPTR, amdgpu_gem_userptr_ioctl, DRM_AUTH|DRM_RENDER_ALLOW),
+ DRM_IOCTL_DEF_DRV(AMDGPU_USERQ, amdgpu_userq_ioctl, DRM_AUTH|DRM_RENDER_ALLOW),
+ DRM_IOCTL_DEF_DRV(AMDGPU_USERQ_SIGNAL, amdgpu_userq_signal_ioctl, DRM_AUTH|DRM_RENDER_ALLOW),
+ DRM_IOCTL_DEF_DRV(AMDGPU_USERQ_WAIT, amdgpu_userq_wait_ioctl, DRM_AUTH|DRM_RENDER_ALLOW),
+ DRM_IOCTL_DEF_DRV(AMDGPU_GEM_LIST_HANDLES, amdgpu_gem_list_handles_ioctl, DRM_AUTH|DRM_RENDER_ALLOW),
};
static const struct drm_driver amdgpu_kms_driver = {
@@ -2916,7 +3096,6 @@ static const struct drm_driver amdgpu_kms_driver = {
.name = DRIVER_NAME,
.desc = DRIVER_DESC,
- .date = DRIVER_DATE,
.major = KMS_DRIVER_MAJOR,
.minor = KMS_DRIVER_MINOR,
.patchlevel = KMS_DRIVER_PATCHLEVEL,
@@ -2940,7 +3119,6 @@ const struct drm_driver amdgpu_partition_driver = {
.name = DRIVER_NAME,
.desc = DRIVER_DESC,
- .date = DRIVER_DATE,
.major = KMS_DRIVER_MAJOR,
.minor = KMS_DRIVER_MINOR,
.patchlevel = KMS_DRIVER_PATCHLEVEL,
@@ -2966,7 +3144,7 @@ static struct pci_driver amdgpu_kms_pci_driver = {
.probe = amdgpu_pci_probe,
.remove = amdgpu_pci_remove,
.shutdown = amdgpu_pci_shutdown,
- .driver.pm = &amdgpu_pm_ops,
+ .driver.pm = pm_ptr(&amdgpu_pm_ops),
.err_handler = &amdgpu_pci_err_handler,
.dev_groups = amdgpu_sysfs_groups,
};
@@ -2975,14 +3153,11 @@ static int __init amdgpu_init(void)
{
int r;
- if (drm_firmware_drivers_only())
- return -EINVAL;
-
r = amdgpu_sync_init();
if (r)
goto error_sync;
- r = amdgpu_fence_slab_init();
+ r = amdgpu_userq_fence_slab_init();
if (r)
goto error_fence;
@@ -3016,7 +3191,7 @@ static void __exit amdgpu_exit(void)
amdgpu_unregister_atpx_handler();
amdgpu_acpi_release();
amdgpu_sync_fini();
- amdgpu_fence_slab_fini();
+ amdgpu_userq_fence_slab_fini();
mmu_notifier_synchronize();
amdgpu_xcp_drv_release();
}
generated by cgit (git 2.34.1) at 2025-12-22 11:54:06 +0000