Merge tag 'amd-drm-next-6.2-2022-11-18' of https://gitlab.freedesktop.org/agd5f/linux into drm-next

amd-drm-next-6.2-2022-11-18: amdgpu: - SR-IOV fixes - Clean up DC checks - DCN 3.2.x fixes - DCN 3.1.x fixes - Don't enable degamma on asics which don't support it - IP discovery fixes - BACO fixes - Fix vbios allocation handling when vkms is enabled - Drop buggy tdr advanced mode GPU reset handling - Fix the build when DCN is not set in kconfig - MST DSC fixes - Userptr fixes - FRU and RAS EEPROM fixes - VCN 4.x RAS support - Aldrebaran CU occupancy reporting fix - PSP ring cleanup amdkfd: - Memory limit fix - Enable cooperative launch on gfx 10.3 amd-drm-next-6.2-2022-11-11: amdgpu: - SMU 13.x updates - GPUVM TLB race fix - DCN 3.1.4 updates - DCN 3.2.x updates - PSR fixes - Kerneldoc fix - Vega10 fan fix - GPUVM locking fixes in error pathes - BACO fix for Beige Goby - EEPROM I2C address cleanup - GFXOFF fix - Fix DC memory leak in error pathes - Flexible array updates - Mtype fix for GPUVM PTEs - Move Kconfig into amdgpu directory - SR-IOV updates - Fix possible memory leak in CS IOCTL error path amdkfd: - Fix possible memory overrun - CRIU fixes radeon: - ACPI ref count fix - HDA audio notifier support - Move Kconfig into radeon directory UAPI: - Add new GEM_CREATE flags to help to transition more KFD functionality to the DRM UAPI. These are used internally in the driver to align location based memory coherency requirements from memory allocated in the KFD with how we manage GPUVM PTEs. They are currently blocked in the GEM_CREATE IOCTL as we don't have a user right now. They are just used internally in the kernel driver for now for existing KFD memory allocations. So a change to the UAPI header, but no functional change in the UAPI. From: Alex Deucher <alexander.deucher@amd.com> Link: https://patchwork.freedesktop.org/patch/msgid/20221118170807.6505-1-alexander.deucher@amd.com Signed-off-by: Dave Airlie <airlied@redhat.com>
author: Dave Airlie <airlied@redhat.com> 2022-11-22 13:41:11 +1000
committer: Dave Airlie <airlied@redhat.com> 2022-11-22 13:41:11 +1000
commit: fc58764bbf602b65a6f63c53e5fd6feae76c510c (patch)
tree: 03f5448cf5d742b8fd2980e86a33636026557ac6 /drivers/gpu/drm/amd/amdgpu
parent: 819683a1fc2f7e64017d50caf539e7bafcb37b81 (diff)
parent: aec3bb3a01de09058fbebed4821ed7d07e1ed994 (diff)
66 files changed, 1035 insertions, 789 deletions
diff --git a/drivers/gpu/drm/amd/amdgpu/Kconfig b/drivers/gpu/drm/amd/amdgpu/Kconfig
index 7777d55275de..5fcd510f1abb 100644
--- a/drivers/gpu/drm/amd/amdgpu/Kconfig
+++ b/drivers/gpu/drm/amd/amdgpu/Kconfig
@@ -1,4 +1,33 @@
 # SPDX-License-Identifier: MIT
+
+config DRM_AMDGPU
+	tristate "AMD GPU"
+	depends on DRM && PCI && MMU
+	select FW_LOADER
+	select DRM_DISPLAY_DP_HELPER
+	select DRM_DISPLAY_HDMI_HELPER
+	select DRM_DISPLAY_HELPER
+	select DRM_KMS_HELPER
+	select DRM_SCHED
+	select DRM_TTM
+	select DRM_TTM_HELPER
+	select POWER_SUPPLY
+	select HWMON
+	select BACKLIGHT_CLASS_DEVICE
+	select INTERVAL_TREE
+	select DRM_BUDDY
+	# amdgpu depends on ACPI_VIDEO when ACPI is enabled, for select to work
+	# ACPI_VIDEO's dependencies must also be selected.
+	select INPUT if ACPI
+	select ACPI_VIDEO if ACPI
+	# On x86 ACPI_VIDEO also needs ACPI_WMI
+	select X86_PLATFORM_DEVICES if ACPI && X86
+	select ACPI_WMI if ACPI && X86
+	help
+	  Choose this option if you have a recent AMD Radeon graphics card.
+
+	  If M is selected, the module will be called amdgpu.
+
 config DRM_AMDGPU_SI
 	bool "Enable amdgpu support for SI parts"
 	depends on DRM_AMDGPU
diff --git a/drivers/gpu/drm/amd/amdgpu/Makefile b/drivers/gpu/drm/amd/amdgpu/Makefile
index 6ad39cf71bdd..712075a491f2 100644
--- a/drivers/gpu/drm/amd/amdgpu/Makefile
+++ b/drivers/gpu/drm/amd/amdgpu/Makefile
@@ -250,7 +250,7 @@ endif
 amdgpu-$(CONFIG_COMPAT) += amdgpu_ioc32.o
 amdgpu-$(CONFIG_VGA_SWITCHEROO) += amdgpu_atpx_handler.o
 amdgpu-$(CONFIG_ACPI) += amdgpu_acpi.o
-amdgpu-$(CONFIG_HMM_MIRROR) += amdgpu_mn.o
+amdgpu-$(CONFIG_HMM_MIRROR) += amdgpu_hmm.o
 
 include $(FULL_AMD_PATH)/pm/Makefile
 
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu.h b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
index 9999c18e7d8e..6b74df446694 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
@@ -82,7 +82,6 @@
 #include "amdgpu_vce.h"
 #include "amdgpu_vcn.h"
 #include "amdgpu_jpeg.h"
-#include "amdgpu_mn.h"
 #include "amdgpu_gmc.h"
 #include "amdgpu_gfx.h"
 #include "amdgpu_sdma.h"
@@ -1065,6 +1064,7 @@ struct amdgpu_device {
 	struct work_struct		reset_work;
 
 	bool                            job_hang;
+	bool                            dc_enabled;
 };
 
 static inline struct amdgpu_device *drm_to_adev(struct drm_device *ddev)
@@ -1122,6 +1122,8 @@ void amdgpu_device_indirect_wreg64(struct amdgpu_device *adev,
 bool amdgpu_device_asic_has_dc_support(enum amd_asic_type asic_type);
 bool amdgpu_device_has_dc_support(struct amdgpu_device *adev);
 
+void amdgpu_device_set_sriov_virtual_display(struct amdgpu_device *adev);
+
 int amdgpu_device_pre_asic_reset(struct amdgpu_device *adev,
 				 struct amdgpu_reset_context *reset_context);
 
@@ -1295,6 +1297,7 @@ void amdgpu_device_pcie_port_wreg(struct amdgpu_device *adev,
 				u32 reg, u32 v);
 struct dma_fence *amdgpu_device_switch_gang(struct amdgpu_device *adev,
 					    struct dma_fence *gang);
+bool amdgpu_device_has_display_hardware(struct amdgpu_device *adev);
 
 /* atpx handler */
 #if defined(CONFIG_VGA_SWITCHEROO)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_acpi.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_acpi.c
index b14800ac179e..57b5e11446c6 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_acpi.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_acpi.c
@@ -847,7 +847,7 @@ int amdgpu_acpi_init(struct amdgpu_device *adev)
 	struct amdgpu_atif *atif = &amdgpu_acpi_priv.atif;
 
 	if (atif->notifications.brightness_change) {
-		if (amdgpu_device_has_dc_support(adev)) {
+		if (adev->dc_enabled) {
 #if defined(CONFIG_DRM_AMD_DC)
 			struct amdgpu_display_manager *dm = &adev->dm;
 
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_aldebaran.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_aldebaran.c
index c8935d718207..4485bb29bec9 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_aldebaran.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_aldebaran.c
@@ -41,5 +41,6 @@ const struct kfd2kgd_calls aldebaran_kfd2kgd = {
 	.get_atc_vmid_pasid_mapping_info =
 				kgd_gfx_v9_get_atc_vmid_pasid_mapping_info,
 	.set_vm_context_page_table_base = kgd_gfx_v9_set_vm_context_page_table_base,
+	.get_cu_occupancy = kgd_gfx_v9_get_cu_occupancy,
 	.program_trap_handler_settings = kgd_gfx_v9_program_trap_handler_settings
 };
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c
index ba72a910d0d5..3a763916a5a1 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c
@@ -29,6 +29,7 @@
 #include "amdgpu_object.h"
 #include "amdgpu_gem.h"
 #include "amdgpu_vm.h"
+#include "amdgpu_hmm.h"
 #include "amdgpu_amdkfd.h"
 #include "amdgpu_dma_buf.h"
 #include <uapi/linux/kfd_ioctl.h>
@@ -171,9 +172,7 @@ int amdgpu_amdkfd_reserve_mem_limit(struct amdgpu_device *adev,
 	    (kfd_mem_limit.ttm_mem_used + ttm_mem_needed >
 	     kfd_mem_limit.max_ttm_mem_limit) ||
 	    (adev && adev->kfd.vram_used + vram_needed >
-	     adev->gmc.real_vram_size -
-	     atomic64_read(&adev->vram_pin_size) -
-	     reserved_for_pt)) {
+	     adev->gmc.real_vram_size - reserved_for_pt)) {
 		ret = -ENOMEM;
 		goto release;
 	}
@@ -405,63 +404,15 @@ static int vm_update_pds(struct amdgpu_vm *vm, struct amdgpu_sync *sync)
 
 static uint64_t get_pte_flags(struct amdgpu_device *adev, struct kgd_mem *mem)
 {
-	struct amdgpu_device *bo_adev = amdgpu_ttm_adev(mem->bo->tbo.bdev);
-	bool coherent = mem->alloc_flags & KFD_IOC_ALLOC_MEM_FLAGS_COHERENT;
-	bool uncached = mem->alloc_flags & KFD_IOC_ALLOC_MEM_FLAGS_UNCACHED;
-	uint32_t mapping_flags;
-	uint64_t pte_flags;
-	bool snoop = false;
+	uint32_t mapping_flags = AMDGPU_VM_PAGE_READABLE |
+				 AMDGPU_VM_MTYPE_DEFAULT;
 
-	mapping_flags = AMDGPU_VM_PAGE_READABLE;
 	if (mem->alloc_flags & KFD_IOC_ALLOC_MEM_FLAGS_WRITABLE)
 		mapping_flags |= AMDGPU_VM_PAGE_WRITEABLE;
 	if (mem->alloc_flags & KFD_IOC_ALLOC_MEM_FLAGS_EXECUTABLE)
 		mapping_flags |= AMDGPU_VM_PAGE_EXECUTABLE;
 
-	switch (adev->ip_versions[GC_HWIP][0]) {
-	case IP_VERSION(9, 4, 1):
-	case IP_VERSION(9, 4, 2):
-		if (mem->alloc_flags & KFD_IOC_ALLOC_MEM_FLAGS_VRAM) {
-			if (bo_adev == adev) {
-				if (uncached)
-					mapping_flags |= AMDGPU_VM_MTYPE_UC;
-				else if (coherent)
-					mapping_flags |= AMDGPU_VM_MTYPE_CC;
-				else
-					mapping_flags |= AMDGPU_VM_MTYPE_RW;
-				if ((adev->ip_versions[GC_HWIP][0] == IP_VERSION(9, 4, 2)) &&
-				    adev->gmc.xgmi.connected_to_cpu)
-					snoop = true;
-			} else {
-				if (uncached || coherent)
-					mapping_flags |= AMDGPU_VM_MTYPE_UC;
-				else
-					mapping_flags |= AMDGPU_VM_MTYPE_NC;
-				if (amdgpu_xgmi_same_hive(adev, bo_adev))
-					snoop = true;
-			}
-		} else {
-			if (uncached || coherent)
-				mapping_flags |= AMDGPU_VM_MTYPE_UC;
-			else
-				mapping_flags |= AMDGPU_VM_MTYPE_NC;
-			snoop = true;
-		}
-		break;
-	default:
-		if (uncached || coherent)
-			mapping_flags |= AMDGPU_VM_MTYPE_UC;
-		else
-			mapping_flags |= AMDGPU_VM_MTYPE_NC;
-
-		if (!(mem->alloc_flags & KFD_IOC_ALLOC_MEM_FLAGS_VRAM))
-			snoop = true;
-	}
-
-	pte_flags = amdgpu_gem_va_map_flags(adev, mapping_flags);
-	pte_flags |= snoop ? AMDGPU_PTE_SNOOPED : 0;
-
-	return pte_flags;
+	return amdgpu_gem_va_map_flags(adev, mapping_flags);
 }
 
 /**
@@ -988,6 +939,7 @@ static int init_user_pages(struct kgd_mem *mem, uint64_t user_addr,
 	struct amdkfd_process_info *process_info = mem->process_info;
 	struct amdgpu_bo *bo = mem->bo;
 	struct ttm_operation_ctx ctx = { true, false };
+	struct hmm_range *range;
 	int ret = 0;
 
 	mutex_lock(&process_info->lock);
@@ -998,7 +950,7 @@ static int init_user_pages(struct kgd_mem *mem, uint64_t user_addr,
 		goto out;
 	}
 
-	ret = amdgpu_mn_register(bo, user_addr);
+	ret = amdgpu_hmm_register(bo, user_addr);
 	if (ret) {
 		pr_err("%s: Failed to register MMU notifier: %d\n",
 		       __func__, ret);
@@ -1017,7 +969,7 @@ static int init_user_pages(struct kgd_mem *mem, uint64_t user_addr,
 		return 0;
 	}
 
-	ret = amdgpu_ttm_tt_get_user_pages(bo, bo->tbo.ttm->pages);
+	ret = amdgpu_ttm_tt_get_user_pages(bo, bo->tbo.ttm->pages, &range);
 	if (ret) {
 		pr_err("%s: Failed to get user pages: %d\n", __func__, ret);
 		goto unregister_out;
@@ -1035,10 +987,10 @@ static int init_user_pages(struct kgd_mem *mem, uint64_t user_addr,
 	amdgpu_bo_unreserve(bo);
 
 release_out:
-	amdgpu_ttm_tt_get_user_pages_done(bo->tbo.ttm);
+	amdgpu_ttm_tt_get_user_pages_done(bo->tbo.ttm, range);
 unregister_out:
 	if (ret)
-		amdgpu_mn_unregister(bo);
+		amdgpu_hmm_unregister(bo);
 out:
 	mutex_unlock(&process_info->lock);
 	return ret;
@@ -1673,6 +1625,11 @@ int amdgpu_amdkfd_gpuvm_alloc_memory_of_gpu(
 		}
 	}
 
+	if (flags & KFD_IOC_ALLOC_MEM_FLAGS_COHERENT)
+		alloc_flags |= AMDGPU_GEM_CREATE_COHERENT;
+	if (flags & KFD_IOC_ALLOC_MEM_FLAGS_UNCACHED)
+		alloc_flags |= AMDGPU_GEM_CREATE_UNCACHED;
+
 	*mem = kzalloc(sizeof(struct kgd_mem), GFP_KERNEL);
 	if (!*mem) {
 		ret = -ENOMEM;
@@ -1817,7 +1774,7 @@ int amdgpu_amdkfd_gpuvm_free_memory_of_gpu(
 	mutex_unlock(&process_info->lock);
 
 	/* No more MMU notifiers */
-	amdgpu_mn_unregister(mem->bo);
+	amdgpu_hmm_unregister(mem->bo);
 
 	ret = reserve_bo_and_cond_vms(mem, NULL, BO_VM_ALL, &ctx);
 	if (unlikely(ret))
@@ -2362,6 +2319,8 @@ static int update_invalid_user_pages(struct amdkfd_process_info *process_info,
 	/* Go through userptr_inval_list and update any invalid user_pages */
 	list_for_each_entry(mem, &process_info->userptr_inval_list,
 			    validate_list.head) {
+		struct hmm_range *range;
+
 		invalid = atomic_read(&mem->invalid);
 		if (!invalid)
 			/* BO hasn't been invalidated since the last
@@ -2372,7 +2331,8 @@ static int update_invalid_user_pages(struct amdkfd_process_info *process_info,
 		bo = mem->bo;
 
 		/* Get updated user pages */
-		ret = amdgpu_ttm_tt_get_user_pages(bo, bo->tbo.ttm->pages);
+		ret = amdgpu_ttm_tt_get_user_pages(bo, bo->tbo.ttm->pages,
+						   &range);
 		if (ret) {
 			pr_debug("Failed %d to get user pages\n", ret);
 
@@ -2391,7 +2351,7 @@ static int update_invalid_user_pages(struct amdkfd_process_info *process_info,
 			 * FIXME: Cannot ignore the return code, must hold
 			 * notifier_lock
 			 */
-			amdgpu_ttm_tt_get_user_pages_done(bo->tbo.ttm);
+			amdgpu_ttm_tt_get_user_pages_done(bo->tbo.ttm, range);
 		}
 
 		/* Mark the BO as valid unless it was invalidated
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_atomfirmware.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_atomfirmware.c
index b81b77a9efa6..9b97fa39d47a 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_atomfirmware.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_atomfirmware.c
@@ -101,39 +101,97 @@ void amdgpu_atomfirmware_scratch_regs_init(struct amdgpu_device *adev)
 	}
 }
 
+static int amdgpu_atomfirmware_allocate_fb_v2_1(struct amdgpu_device *adev,
+	struct vram_usagebyfirmware_v2_1 *fw_usage, int *usage_bytes)
+{
+	uint32_t start_addr, fw_size, drv_size;
+
+	start_addr = le32_to_cpu(fw_usage->start_address_in_kb);
+	fw_size = le16_to_cpu(fw_usage->used_by_firmware_in_kb);
+	drv_size = le16_to_cpu(fw_usage->used_by_driver_in_kb);
+
+	DRM_DEBUG("atom firmware v2_1 requested %08x %dkb fw %dkb drv\n",
+			  start_addr,
+			  fw_size,
+			  drv_size);
+
+	if ((start_addr & ATOM_VRAM_OPERATION_FLAGS_MASK) ==
+		(uint32_t)(ATOM_VRAM_BLOCK_SRIOV_MSG_SHARE_RESERVATION <<
+		ATOM_VRAM_OPERATION_FLAGS_SHIFT)) {
+		/* Firmware request VRAM reservation for SR-IOV */
+		adev->mman.fw_vram_usage_start_offset = (start_addr &
+			(~ATOM_VRAM_OPERATION_FLAGS_MASK)) << 10;
+		adev->mman.fw_vram_usage_size = fw_size << 10;
+		/* Use the default scratch size */
+		*usage_bytes = 0;
+	} else {
+		*usage_bytes = drv_size << 10;
+	}
+	return 0;
+}
+
+static int amdgpu_atomfirmware_allocate_fb_v2_2(struct amdgpu_device *adev,
+		struct vram_usagebyfirmware_v2_2 *fw_usage, int *usage_bytes)
+{
+	uint32_t fw_start_addr, fw_size, drv_start_addr, drv_size;
+
+	fw_start_addr = le32_to_cpu(fw_usage->fw_region_start_address_in_kb);
+	fw_size = le16_to_cpu(fw_usage->used_by_firmware_in_kb);
+
+	drv_start_addr = le32_to_cpu(fw_usage->driver_region0_start_address_in_kb);
+	drv_size = le32_to_cpu(fw_usage->used_by_driver_region0_in_kb);
+
+	DRM_DEBUG("atom requested fw start at %08x %dkb and drv start at %08x %dkb\n",
+			  fw_start_addr,
+			  fw_size,
+			  drv_start_addr,
+			  drv_size);
+
+	if ((fw_start_addr & (ATOM_VRAM_BLOCK_NEEDS_NO_RESERVATION << 30)) == 0) {
+		/* Firmware request VRAM reservation for SR-IOV */
+		adev->mman.fw_vram_usage_start_offset = (fw_start_addr &
+			(~ATOM_VRAM_OPERATION_FLAGS_MASK)) << 10;
+		adev->mman.fw_vram_usage_size = fw_size << 10;
+	}
+
+	if ((drv_start_addr & (ATOM_VRAM_BLOCK_NEEDS_NO_RESERVATION << 30)) == 0) {
+		/* driver request VRAM reservation for SR-IOV */
+		adev->mman.drv_vram_usage_start_offset = (drv_start_addr &
+			(~ATOM_VRAM_OPERATION_FLAGS_MASK)) << 10;
+		adev->mman.drv_vram_usage_size = drv_size << 10;
+	}
+
+	*usage_bytes = 0;
+	return 0;
+}
+
 int amdgpu_atomfirmware_allocate_fb_scratch(struct amdgpu_device *adev)
 {
 	struct atom_context *ctx = adev->mode_info.atom_context;
 	int index = get_index_into_master_table(atom_master_list_of_data_tables_v2_1,
 						vram_usagebyfirmware);
-	struct vram_usagebyfirmware_v2_1 *firmware_usage;
-	uint32_t start_addr, size;
+	struct vram_usagebyfirmware_v2_1 *fw_usage_v2_1;
+	struct vram_usagebyfirmware_v2_2 *fw_usage_v2_2;
 	uint16_t data_offset;
+	uint8_t frev, crev;
 	int usage_bytes = 0;
 
-	if (amdgpu_atom_parse_data_header(ctx, index, NULL, NULL, NULL, &data_offset)) {
-		firmware_usage = (struct vram_usagebyfirmware_v2_1 *)(ctx->bios + data_offset);
-		DRM_DEBUG("atom firmware requested %08x %dkb fw %dkb drv\n",
-			  le32_to_cpu(firmware_usage->start_address_in_kb),
-			  le16_to_cpu(firmware_usage->used_by_firmware_in_kb),
-			  le16_to_cpu(firmware_usage->used_by_driver_in_kb));
-
-		start_addr = le32_to_cpu(firmware_usage->start_address_in_kb);
-		size = le16_to_cpu(firmware_usage->used_by_firmware_in_kb);
-
-		if ((uint32_t)(start_addr & ATOM_VRAM_OPERATION_FLAGS_MASK) ==
-			(uint32_t)(ATOM_VRAM_BLOCK_SRIOV_MSG_SHARE_RESERVATION <<
-			ATOM_VRAM_OPERATION_FLAGS_SHIFT)) {
-			/* Firmware request VRAM reservation for SR-IOV */
-			adev->mman.fw_vram_usage_start_offset = (start_addr &
-				(~ATOM_VRAM_OPERATION_FLAGS_MASK)) << 10;
-			adev->mman.fw_vram_usage_size = size << 10;
-			/* Use the default scratch size */
-			usage_bytes = 0;
-		} else {
-			usage_bytes = le16_to_cpu(firmware_usage->used_by_driver_in_kb) << 10;
+	if (amdgpu_atom_parse_data_header(ctx, index, NULL, &frev, &crev, &data_offset)) {
+		if (frev == 2 && crev == 1) {
+			fw_usage_v2_1 =
+				(struct vram_usagebyfirmware_v2_1 *)(ctx->bios + data_offset);
+			amdgpu_atomfirmware_allocate_fb_v2_1(adev,
+					fw_usage_v2_1,
+					&usage_bytes);
+		} else if (frev >= 2 && crev >= 2) {
+			fw_usage_v2_2 =
+				(struct vram_usagebyfirmware_v2_2 *)(ctx->bios + data_offset);
+			amdgpu_atomfirmware_allocate_fb_v2_2(adev,
+					fw_usage_v2_2,
+					&usage_bytes);
 		}
 	}
+
 	ctx->scratch_size_bytes = 0;
 	if (usage_bytes == 0)
 		usage_bytes = 20 * 1024;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_bo_list.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_bo_list.c
index 2168163aad2d..252a876b0725 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_bo_list.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_bo_list.c
@@ -209,6 +209,7 @@ void amdgpu_bo_list_get_list(struct amdgpu_bo_list *list,
 			list_add_tail(&e->tv.head, &bucket[priority]);
 
 		e->user_pages = NULL;
+		e->range = NULL;
 	}
 
 	/* Connect the sorted buckets in the output list. */
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_bo_list.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_bo_list.h
index 9caea1688fc3..e4d78491bcc7 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_bo_list.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_bo_list.h
@@ -26,6 +26,8 @@
 #include <drm/ttm/ttm_execbuf_util.h>
 #include <drm/amdgpu_drm.h>
 
+struct hmm_range;
+
 struct amdgpu_device;
 struct amdgpu_bo;
 struct amdgpu_bo_va;
@@ -36,6 +38,7 @@ struct amdgpu_bo_list_entry {
 	struct amdgpu_bo_va		*bo_va;
 	uint32_t			priority;
 	struct page			**user_pages;
+	struct hmm_range		*range;
 	bool				user_invalidated;
 };
 
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
index 0528c2b1db6e..459150460815 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
@@ -287,8 +287,10 @@ static int amdgpu_cs_pass1(struct amdgpu_cs_parser *p,
 		}
 	}
 
-	if (!p->gang_size)
-		return -EINVAL;
+	if (!p->gang_size) {
+		ret = -EINVAL;
+		goto free_partial_kdata;
+	}
 
 	for (i = 0; i < p->gang_size; ++i) {
 		ret = amdgpu_job_alloc(p->adev, vm, p->entities[i], vm,
@@ -917,7 +919,7 @@ static int amdgpu_cs_parser_bos(struct amdgpu_cs_parser *p,
 			goto out_free_user_pages;
 		}
 
-		r = amdgpu_ttm_tt_get_user_pages(bo, e->user_pages);
+		r = amdgpu_ttm_tt_get_user_pages(bo, e->user_pages, &e->range);
 		if (r) {
 			kvfree(e->user_pages);
 			e->user_pages = NULL;
@@ -995,10 +997,12 @@ out_free_user_pages:
 
 		if (!e->user_pages)
 			continue;
-		amdgpu_ttm_tt_get_user_pages_done(bo->tbo.ttm);
+		amdgpu_ttm_tt_get_user_pages_done(bo->tbo.ttm, e->range);
 		kvfree(e->user_pages);
 		e->user_pages = NULL;
+		e->range = NULL;
 	}
+	mutex_unlock(&p->bo_list->bo_list_mutex);
 	return r;
 }
 
@@ -1270,7 +1274,8 @@ static int amdgpu_cs_submit(struct amdgpu_cs_parser *p,
 	amdgpu_bo_list_for_each_userptr_entry(e, p->bo_list) {
 		struct amdgpu_bo *bo = ttm_to_amdgpu_bo(e->tv.bo);
 
-		r |= !amdgpu_ttm_tt_get_user_pages_done(bo->tbo.ttm);
+		r |= !amdgpu_ttm_tt_get_user_pages_done(bo->tbo.ttm, e->range);
+		e->range = NULL;
 	}
 	if (r) {
 		r = -EAGAIN;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c
index de61a85c4b02..0f16d3c09309 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c
@@ -1969,7 +1969,7 @@ int amdgpu_debugfs_init(struct amdgpu_device *adev)
 	amdgpu_ta_if_debugfs_init(adev);
 
 #if defined(CONFIG_DRM_AMD_DC)
-	if (amdgpu_device_has_dc_support(adev))
+	if (adev->dc_enabled)
 		dtn_debugfs_init(adev);
 #endif
 
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
index ad4e78728733..b2b1c66bfe39 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
@@ -1916,6 +1916,16 @@ static void amdgpu_device_enable_virtual_display(struct amdgpu_device *adev)
 	}
 }
 
+void amdgpu_device_set_sriov_virtual_display(struct amdgpu_device *adev)
+{
+	if (amdgpu_sriov_vf(adev) && !adev->enable_virtual_display) {
+		adev->mode_info.num_crtc = 1;
+		adev->enable_virtual_display = true;
+		DRM_INFO("virtual_display:%d, num_crtc:%d\n",
+			 adev->enable_virtual_display, adev->mode_info.num_crtc);
+	}
+}
+
 /**
  * amdgpu_device_parse_gpu_info_fw - parse gpu info firmware
  *
@@ -3348,8 +3358,7 @@ bool amdgpu_device_asic_has_dc_support(enum amd_asic_type asic_type)
  */
 bool amdgpu_device_has_dc_support(struct amdgpu_device *adev)
 {
-	if (amdgpu_sriov_vf(adev) ||
-	    adev->enable_virtual_display ||
+	if (adev->enable_virtual_display ||
 	    (adev->harvest_ip_mask & AMD_HARVEST_IP_DMU_MASK))
 		return false;
 
@@ -4216,25 +4225,27 @@ int amdgpu_device_resume(struct drm_device *dev, bool fbcon)
 
 	amdgpu_ras_resume(adev);
 
-	/*
-	 * Most of the connector probing functions try to acquire runtime pm
-	 * refs to ensure that the GPU is powered on when connector polling is
-	 * performed. Since we're calling this from a runtime PM callback,
-	 * trying to acquire rpm refs will cause us to deadlock.
-	 *
-	 * Since we're guaranteed to be holding the rpm lock, it's safe to
-	 * temporarily disable the rpm helpers so this doesn't deadlock us.
-	 */
+	if (adev->mode_info.num_crtc) {
+		/*
+		 * Most of the connector probing functions try to acquire runtime pm
+		 * refs to ensure that the GPU is powered on when connector polling is
+		 * performed. Since we're calling this from a runtime PM callback,
+		 * trying to acquire rpm refs will cause us to deadlock.
+		 *
+		 * Since we're guaranteed to be holding the rpm lock, it's safe to
+		 * temporarily disable the rpm helpers so this doesn't deadlock us.
+		 */
 #ifdef CONFIG_PM
-	dev->dev->power.disable_depth++;
+		dev->dev->power.disable_depth++;
 #endif
-	if (!amdgpu_device_has_dc_support(adev))
-		drm_helper_hpd_irq_event(dev);
-	else
-		drm_kms_helper_hotplug_event(dev);
+		if (!adev->dc_enabled)
+			drm_helper_hpd_irq_event(dev);
+		else
+			drm_kms_helper_hotplug_event(dev);
 #ifdef CONFIG_PM
-	dev->dev->power.disable_depth--;
+		dev->dev->power.disable_depth--;
 #endif
+	}
 	adev->in_suspend = false;
 
 	if (amdgpu_acpi_smart_shift_update(dev, AMDGPU_SS_DEV_D0))
@@ -4583,6 +4594,10 @@ bool amdgpu_device_should_recover_gpu(struct amdgpu_device *adev)
 	if (amdgpu_gpu_recovery == 0)
 		goto disabled;
 
+	/* Skip soft reset check in fatal error mode */
+	if (!amdgpu_ras_is_poison_mode_supported(adev))
+		return true;
+
 	if (!amdgpu_device_ip_check_soft_reset(adev)) {
 		dev_info(adev->dev,"Timeout, but no hardware hang detected.\n");
 		return false;
@@ -5078,94 +5093,6 @@ static int amdgpu_device_suspend_display_audio(struct amdgpu_device *adev)
 	return 0;
 }
 
-static void amdgpu_device_recheck_guilty_jobs(
-	struct amdgpu_device *adev, struct list_head *device_list_handle,
-	struct amdgpu_reset_context *reset_context)
-{
-	int i, r = 0;
-
-	for (i = 0; i < AMDGPU_MAX_RINGS; ++i) {
-		struct amdgpu_ring *ring = adev->rings[i];
-		int ret = 0;
-		struct drm_sched_job *s_job;
-
-		if (!ring || !ring->sched.thread)
-			continue;
-
-		s_job = list_first_entry_or_null(&ring->sched.pending_list,
-				struct drm_sched_job, list);
-		if (s_job == NULL)
-			continue;
-
-		/* clear job's guilty and depend the folowing step to decide the real one */
-		drm_sched_reset_karma(s_job);
-		drm_sched_resubmit_jobs_ext(&ring->sched, 1);
-
-		if (!s_job->s_fence->parent) {
-			DRM_WARN("Failed to get a HW fence for job!");
-			continue;
-		}
-
-		ret = dma_fence_wait_timeout(s_job->s_fence->parent, false, ring->sched.timeout);
-		if (ret == 0) { /* timeout */
-			DRM_ERROR("Found the real bad job! ring:%s, job_id:%llx\n",
-						ring->sched.name, s_job->id);
-
-
-			amdgpu_fence_driver_isr_toggle(adev, true);
-
-			/* Clear this failed job from fence array */
-			amdgpu_fence_driver_clear_job_fences(ring);
-
-			amdgpu_fence_driver_isr_toggle(adev, false);
-
-			/* Since the job won't signal and we go for
-			 * another resubmit drop this parent pointer
-			 */
-			dma_fence_put(s_job->s_fence->parent);
-			s_job->s_fence->parent = NULL;
-
-			/* set guilty */
-			drm_sched_increase_karma(s_job);
-			amdgpu_reset_prepare_hwcontext(adev, reset_context);
-retry:
-			/* do hw reset */
-			if (amdgpu_sriov_vf(adev)) {
-				amdgpu_virt_fini_data_exchange(adev);
-				r = amdgpu_device_reset_sriov(adev, false);
-				if (r)
-					adev->asic_reset_res = r;
-			} else {
-				clear_bit(AMDGPU_SKIP_HW_RESET,
-					  &reset_context->flags);
-				r = amdgpu_do_asic_reset(device_list_handle,
-							 reset_context);
-				if (r && r == -EAGAIN)
-					goto retry;
-			}
-
-			/*
-			 * add reset counter so that the following
-			 * resubmitted job could flush vmid
-			 */
-			atomic_inc(&adev->gpu_reset_counter);
-			continue;
-		}
-
-		/* got the hw fence, signal finished fence */
-		atomic_dec(ring->sched.score);
-		dma_fence_get(&s_job->s_fence->finished);
-		dma_fence_signal(&s_job->s_fence->finished);
-		dma_fence_put(&s_job->s_fence->finished);
-
-		/* remove node from list and free the job */
-		spin_lock(&ring->sched.job_list_lock);
-		list_del_init(&s_job->list);
-		spin_unlock(&ring->sched.job_list_lock);
-		ring->sched.ops->free_job(s_job);
-	}
-}
-
 static inline void amdgpu_device_stop_pending_resets(struct amdgpu_device *adev)
 {
 	struct amdgpu_ras *con = amdgpu_ras_get_context(adev);
@@ -5186,7 +5113,6 @@ static inline void amdgpu_device_stop_pending_resets(struct amdgpu_device *adev)
 
 }
 
-
 /**
  * amdgpu_device_gpu_recover - reset the asic and recover scheduler
  *
@@ -5209,7 +5135,6 @@ int amdgpu_device_gpu_recover(struct amdgpu_device *adev,
 	int i, r = 0;
 	bool need_emergency_restart = false;
 	bool audio_suspended = false;
-	int tmp_vram_lost_counter;
 	bool gpu_reset_for_dev_remove = false;
 
 	gpu_reset_for_dev_remove =
@@ -5355,7 +5280,6 @@ retry:	/* Rest of adevs pre asic reset from XGMI hive. */
 		amdgpu_device_stop_pending_resets(tmp_adev);
 	}
 
-	tmp_vram_lost_counter = atomic_read(&((adev)->vram_lost_counter));
 	/* Actual ASIC resets if needed.*/
 	/* Host driver will handle XGMI hive reset for SRIOV */
 	if (amdgpu_sriov_vf(adev)) {
@@ -5380,29 +5304,13 @@ skip_hw_reset:
 	/* Post ASIC reset for all devs .*/
 	list_for_each_entry(tmp_adev, device_list_handle, reset_list) {
 
-		/*
-		 * Sometimes a later bad compute job can block a good gfx job as gfx
-		 * and compute ring share internal GC HW mutually. We add an additional
-		 * guilty jobs recheck step to find the real guilty job, it synchronously
-		 * submits and pends for the first job being signaled. If it gets timeout,
-		 * we identify it as a real guilty job.
-		 */
-		if (amdgpu_gpu_recovery == 2 &&
-			!(tmp_vram_lost_counter < atomic_read(&adev->vram_lost_counter)))
-			amdgpu_device_recheck_guilty_jobs(
-				tmp_adev, device_list_handle, reset_context);
-
 		for (i = 0; i < AMDGPU_MAX_RINGS; ++i) {
 			struct amdgpu_ring *ring = tmp_adev->rings[i];
 
 			if (!ring || !ring->sched.thread)
 				continue;
 
-			/* No point to resubmit jobs if we didn't HW reset*/
-			if (!tmp_adev->asic_reset_res && !job_signaled)
-				drm_sched_resubmit_jobs(&ring->sched);
-
-			drm_sched_start(&ring->sched, !tmp_adev->asic_reset_res);
+			drm_sched_start(&ring->sched, true);
 		}
 
 		if (adev->enable_mes && adev->ip_versions[GC_HWIP][0] != IP_VERSION(11, 0, 3))
@@ -5444,6 +5352,8 @@ skip_sched_resume:
 			amdgpu_device_resume_display_audio(tmp_adev);
 
 		amdgpu_device_unset_mp1_state(tmp_adev);
+
+		amdgpu_ras_set_error_query_ready(tmp_adev, true);
 	}
 
 recover_end:
@@ -5855,8 +5765,6 @@ void amdgpu_pci_resume(struct pci_dev *pdev)
 		if (!ring || !ring->sched.thread)
 			continue;
 
-
-		drm_sched_resubmit_jobs(&ring->sched);
 		drm_sched_start(&ring->sched, true);
 	}
 
@@ -6047,3 +5955,44 @@ struct dma_fence *amdgpu_device_switch_gang(struct amdgpu_device *adev,
 	dma_fence_put(old);
 	return NULL;
 }
+
+bool amdgpu_device_has_display_hardware(struct amdgpu_device *adev)
+{
+	switch (adev->asic_type) {
+#ifdef CONFIG_DRM_AMDGPU_SI
+	case CHIP_HAINAN:
+#endif
+	case CHIP_TOPAZ:
+		/* chips with no display hardware */
+		return false;
+#ifdef CONFIG_DRM_AMDGPU_SI
+	case CHIP_TAHITI:
+	case CHIP_PITCAIRN:
+	case CHIP_VERDE:
+	case CHIP_OLAND:
+#endif
+#ifdef CONFIG_DRM_AMDGPU_CIK
+	case CHIP_BONAIRE:
+	case CHIP_HAWAII:
+	case CHIP_KAVERI:
+	case CHIP_KABINI:
+	case CHIP_MULLINS:
+#endif
+	case CHIP_TONGA:
+	case CHIP_FIJI:
+	case CHIP_POLARIS10:
+	case CHIP_POLARIS11:
+	case CHIP_POLARIS12:
+	case CHIP_VEGAM:
+	case CHIP_CARRIZO:
+	case CHIP_STONEY:
+		/* chips with display hardware */
+		return true;
+	default:
+		/* IP discovery */
+		if (!adev->ip_versions[DCE_HWIP][0] ||
+		    (adev->harvest_ip_mask & AMD_HARVEST_IP_DMU_MASK))
+			return false;
+		return true;
+	}
+}
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.c
index 3993e6134914..6b48178455bc 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.c
@@ -305,8 +305,13 @@ static int amdgpu_discovery_init(struct amdgpu_device *adev)
 		goto out;
 	}
 
-	if (!amdgpu_discovery_verify_binary_signature(adev->mman.discovery_bin)) {
-		dev_warn(adev->dev, "get invalid ip discovery binary signature from vram\n");
+	if (!amdgpu_discovery_verify_binary_signature(adev->mman.discovery_bin) || amdgpu_discovery == 2) {
+		/* ignore the discovery binary from vram if discovery=2 in kernel module parameter */
+		if (amdgpu_discovery == 2)
+			dev_info(adev->dev,"force read ip discovery binary from file");
+		else
+			dev_warn(adev->dev, "get invalid ip discovery binary signature from vram\n");
+
 		/* retry read ip discovery binary from file */
 		r = amdgpu_discovery_read_binary_from_file(adev, adev->mman.discovery_bin);
 		if (r) {
@@ -1697,9 +1702,15 @@ static int amdgpu_discovery_set_smu_ip_blocks(struct amdgpu_device *adev)
 	return 0;
 }
 
+static void amdgpu_discovery_set_sriov_display(struct amdgpu_device *adev)
+{
+	amdgpu_device_set_sriov_virtual_display(adev);
+	amdgpu_device_ip_block_add(adev, &amdgpu_vkms_ip_block);
+}
+
 static int amdgpu_discovery_set_display_ip_blocks(struct amdgpu_device *adev)
 {
-	if (adev->enable_virtual_display || amdgpu_sriov_vf(adev)) {
+	if (adev->enable_virtual_display) {
 		amdgpu_device_ip_block_add(adev, &amdgpu_vkms_ip_block);
 		return 0;
 	}
@@ -1727,7 +1738,10 @@ static int amdgpu_discovery_set_display_ip_blocks(struct amdgpu_device *adev)
 		case IP_VERSION(3, 1, 6):
 		case IP_VERSION(3, 2, 0):
 		case IP_VERSION(3, 2, 1):
-			amdgpu_device_ip_block_add(adev, &dm_ip_block);
+			if (amdgpu_sriov_vf(adev))
+				amdgpu_discovery_set_sriov_display(adev);
+			else
+				amdgpu_device_ip_block_add(adev, &dm_ip_block);
 			break;
 		default:
 			dev_err(adev->dev,
@@ -1740,7 +1754,10 @@ static int amdgpu_discovery_set_display_ip_blocks(struct amdgpu_device *adev)
 		case IP_VERSION(12, 0, 0):
 		case IP_VERSION(12, 0, 1):
 		case IP_VERSION(12, 1, 0):
-			amdgpu_device_ip_block_add(adev, &dm_ip_block);
+			if (amdgpu_sriov_vf(adev))
+				amdgpu_discovery_set_sriov_display(adev);
+			else
+				amdgpu_device_ip_block_add(adev, &dm_ip_block);
 			break;
 		default:
 			dev_err(adev->dev,
@@ -2161,6 +2178,7 @@ int amdgpu_discovery_set_ip_blocks(struct amdgpu_device *adev)
 		break;
 	case IP_VERSION(10, 3, 1):
 		adev->family = AMDGPU_FAMILY_VGH;
+		adev->apu_flags |= AMD_APU_IS_VANGOGH;
 		break;
 	case IP_VERSION(10, 3, 3):
 		adev->family = AMDGPU_FAMILY_YC;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_display.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_display.c
index 311a8ea6f065..b22471b3bd63 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_display.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_display.c
@@ -44,6 +44,41 @@
 #include <drm/drm_fourcc.h>
 #include <drm/drm_vblank.h>
 
+/**
+ * amdgpu_display_hotplug_work_func - work handler for display hotplug event
+ *
+ * @work: work struct pointer
+ *
+ * This is the hotplug event work handler (all ASICs).
+ * The work gets scheduled from the IRQ handler if there
+ * was a hotplug interrupt.  It walks through the connector table
+ * and calls hotplug handler for each connector. After this, it sends
+ * a DRM hotplug event to alert userspace.
+ *
+ * This design approach is required in order to defer hotplug event handling
+ * from the IRQ handler to a work handler because hotplug handler has to use
+ * mutexes which cannot be locked in an IRQ handler (since &mutex_lock may
+ * sleep).
+ */
+void amdgpu_display_hotplug_work_func(struct work_struct *work)
+{
+	struct amdgpu_device *adev = container_of(work, struct amdgpu_device,
+						  hotplug_work);
+	struct drm_device *dev = adev_to_drm(adev);
+	struct drm_mode_config *mode_config = &dev->mode_config;
+	struct drm_connector *connector;
+	struct drm_connector_list_iter iter;
+
+	mutex_lock(&mode_config->mutex);
+	drm_connector_list_iter_begin(dev, &iter);
+	drm_for_each_connector_iter(connector, &iter)
+		amdgpu_connector_hotplug(connector);
+	drm_connector_list_iter_end(&iter);
+	mutex_unlock(&mode_config->mutex);
+	/* Just fire off a uevent and let userspace tell us what to do */
+	drm_helper_hpd_irq_event(dev);
+}
+
 static int amdgpu_display_framebuffer_init(struct drm_device *dev,
 					   struct amdgpu_framebuffer *rfb,
 					   const struct drm_mode_fb_cmd2 *mode_cmd,
@@ -514,7 +549,7 @@ uint32_t amdgpu_display_supported_domains(struct amdgpu_device *adev,
 	 */
 	if ((bo_flags & AMDGPU_GEM_CREATE_CPU_GTT_USWC) &&
 	    amdgpu_bo_support_uswc(bo_flags) &&
-	    amdgpu_device_asic_has_dc_support(adev->asic_type) &&
+	    adev->dc_enabled &&
 	    adev->mode_info.gpu_vm_support)
 		domain |= AMDGPU_GEM_DOMAIN_GTT;
 #endif
@@ -1280,7 +1315,7 @@ int amdgpu_display_modeset_create_props(struct amdgpu_device *adev)
 					 "dither",
 					 amdgpu_dither_enum_list, sz);
 
-	if (amdgpu_device_has_dc_support(adev)) {
+	if (adev->dc_enabled) {
 		adev->mode_info.abm_level_property =
 			drm_property_create_range(adev_to_drm(adev), 0,
 						  "abm level", 0, 4);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_display.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_display.h
index 560352f7c317..9d19940f73c8 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_display.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_display.h
@@ -35,6 +35,7 @@
 #define amdgpu_display_add_encoder(adev, e, s, c) (adev)->mode_info.funcs->add_encoder((adev), (e), (s), (c))
 #define amdgpu_display_add_connector(adev, ci, sd, ct, ib, coi, h, r) (adev)->mode_info.funcs->add_connector((adev), (ci), (sd), (ct), (ib), (coi), (h), (r))
 
+void amdgpu_display_hotplug_work_func(struct work_struct *work);
 void amdgpu_display_update_priority(struct amdgpu_device *adev);
 uint32_t amdgpu_display_supported_domains(struct amdgpu_device *adev,
 					  uint64_t bo_flags);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_dma_buf.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_dma_buf.c
index 7bd8e33b14be..271e30e34d93 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_dma_buf.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_dma_buf.c
@@ -328,7 +328,9 @@ amdgpu_dma_buf_create_obj(struct drm_device *dev, struct dma_buf *dma_buf)
 	if (dma_buf->ops == &amdgpu_dmabuf_ops) {
 		struct amdgpu_bo *other = gem_to_amdgpu_bo(dma_buf->priv);
 
-		flags |= other->flags & AMDGPU_GEM_CREATE_CPU_GTT_USWC;
+		flags |= other->flags & (AMDGPU_GEM_CREATE_CPU_GTT_USWC |
+					 AMDGPU_GEM_CREATE_COHERENT |
+					 AMDGPU_GEM_CREATE_UNCACHED);
 	}
 
 	ret = amdgpu_gem_object_create(adev, dma_buf->size, PAGE_SIZE,
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
index 6b35bb948f96..5697d456d7d0 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
@@ -534,7 +534,7 @@ module_param_named(compute_multipipe, amdgpu_compute_multipipe, int, 0444);
  * DOC: gpu_recovery (int)
  * 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, (2 = advanced tdr mode, 1 = enable, 0 = disable, -1 = auto)");
+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);
 
 /**
@@ -1925,9 +1925,6 @@ static const struct pci_device_id pciidlist[] = {
 	{0x1002, 0x73AF, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_SIENNA_CICHLID},
 	{0x1002, 0x73BF, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_SIENNA_CICHLID},
 
-	/* Van Gogh */
-	{0x1002, 0x163F, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_VANGOGH|AMD_IS_APU},
-
 	/* Yellow Carp */
 	{0x1002, 0x164D, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_YELLOW_CARP|AMD_IS_APU},
 	{0x1002, 0x1681, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_YELLOW_CARP|AMD_IS_APU},
@@ -2472,7 +2469,7 @@ static int amdgpu_runtime_idle_check_display(struct device *dev)
 		if (ret)
 			return ret;
 
-		if (amdgpu_device_has_dc_support(adev)) {
+		if (adev->dc_enabled) {
 			struct drm_crtc *crtc;
 
 			drm_for_each_crtc(crtc, drm_dev) {
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_eeprom.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_eeprom.c
index 4d9eb0137f8c..7d2a908438e9 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_eeprom.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_eeprom.c
@@ -79,13 +79,15 @@
  * That is, for an I2C EEPROM driver everything is controlled by
  * the "eeprom_addr".
  *
+ * See also top of amdgpu_ras_eeprom.c.
+ *
  * P.S. If you need to write, lock and read the Identification Page,
  * (M24M02-DR device only, which we do not use), change the "7" to
  * "0xF" in the macro below, and let the client set bit 20 to 1 in
  * "eeprom_addr", and set A10 to 0 to write into it, and A10 and A1 to
  * 1 to lock it permanently.
  */
-#define MAKE_I2C_ADDR(_aa) ((0xA << 3) | (((_aa) >> 16) & 7))
+#define MAKE_I2C_ADDR(_aa) ((0xA << 3) | (((_aa) >> 16) & 0xF))
 
 static int __amdgpu_eeprom_xfer(struct i2c_adapter *i2c_adap, u32 eeprom_addr,
 				u8 *eeprom_buf, u16 buf_size, bool read)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_fru_eeprom.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_fru_eeprom.c
index e325150879df..2c38ac7bc643 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_fru_eeprom.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_fru_eeprom.c
@@ -29,9 +29,10 @@
 #include "amdgpu_fru_eeprom.h"
 #include "amdgpu_eeprom.h"
 
-#define FRU_EEPROM_MADDR        0x60000
+#define FRU_EEPROM_MADDR_6      0x60000
+#define FRU_EEPROM_MADDR_8      0x80000
 
-static bool is_fru_eeprom_supported(struct amdgpu_device *adev)
+static bool is_fru_eeprom_supported(struct amdgpu_device *adev, u32 *fru_addr)
 {
 	/* Only server cards have the FRU EEPROM
 	 * TODO: See if we can figure this out dynamically instead of
@@ -45,6 +46,11 @@ static bool is_fru_eeprom_supported(struct amdgpu_device *adev)
 	if (amdgpu_sriov_vf(adev))
 		return false;
 
+	/* The default I2C EEPROM address of the FRU.
+	 */
+	if (fru_addr)
+		*fru_addr = FRU_EEPROM_MADDR_8;
+
 	/* VBIOS is of the format ###-DXXXYYYY-##. For SKU identification,
 	 * we can use just the "DXXX" portion. If there were more models, we
 	 * could convert the 3 characters to a hex integer and use a switch
@@ -57,21 +63,29 @@ static bool is_fru_eeprom_supported(struct amdgpu_device *adev)
 		if (strnstr(atom_ctx->vbios_version, "D161",
 			    sizeof(atom_ctx->vbios_version)) ||
 		    strnstr(atom_ctx->vbios_version, "D163",
-			    sizeof(atom_ctx->vbios_version)))
+			    sizeof(atom_ctx->vbios_version))) {
+			*fru_addr = FRU_EEPROM_MADDR_6;
 			return true;
-		else
+		} else {
 			return false;
+		}
 	case CHIP_ALDEBARAN:
-		/* All Aldebaran SKUs have the FRU */
+		/* All Aldebaran SKUs have an FRU */
+		if (!strnstr(atom_ctx->vbios_version, "D673",
+			     sizeof(atom_ctx->vbios_version)))
+			if (fru_addr)
+				*fru_addr = FRU_EEPROM_MADDR_6;
 		return true;
 	case CHIP_SIENNA_CICHLID:
 		if (strnstr(atom_ctx->vbios_version, "D603",
-		    sizeof(atom_ctx->vbios_version))) {
+			    sizeof(atom_ctx->vbios_version))) {
 			if (strnstr(atom_ctx->vbios_version, "D603GLXE",
-			    sizeof(atom_ctx->vbios_version)))
+				    sizeof(atom_ctx->vbios_version))) {
 				return false;
-			else
+			} else {
+				*fru_addr = FRU_EEPROM_MADDR_6;
 				return true;
+			}
 		} else {
 			return false;
 		}
@@ -80,41 +94,14 @@ static bool is_fru_eeprom_supported(struct amdgpu_device *adev)
 	}
 }
 
-static int amdgpu_fru_read_eeprom(struct amdgpu_device *adev, uint32_t addrptr,
-				  unsigned char *buf, size_t buf_size)
-{
-	int ret;
-	u8 size;
-
-	ret = amdgpu_eeprom_read(adev->pm.fru_eeprom_i2c_bus, addrptr, buf, 1);
-	if (ret < 1) {
-		DRM_WARN("FRU: Failed to get size field");
-		return ret;
-	}
-
-	/* The size returned by the i2c requires subtraction of 0xC0 since the
-	 * size apparently always reports as 0xC0+actual size.
-	 */
-	size = buf[0] & 0x3F;
-	size = min_t(size_t, size, buf_size);
-
-	ret = amdgpu_eeprom_read(adev->pm.fru_eeprom_i2c_bus, addrptr + 1,
-				 buf, size);
-	if (ret < 1) {
-		DRM_WARN("FRU: Failed to get data field");
-		return ret;
-	}
-
-	return size;
-}
-
 int amdgpu_fru_get_product_info(struct amdgpu_device *adev)
 {
-	unsigned char buf[AMDGPU_PRODUCT_NAME_LEN];
-	u32 addrptr;
+	unsigned char buf[8], *pia;
+	u32 addr, fru_addr;
 	int size, len;
+	u8 csum;
 
-	if (!is_fru_eeprom_supported(adev))
+	if (!is_fru_eeprom_supported(adev, &fru_addr))
 		return 0;
 
 	/* If algo exists, it means that the i2c_adapter's initialized */
@@ -123,88 +110,102 @@ int amdgpu_fru_get_product_info(struct amdgpu_device *adev)
 		return -ENODEV;
 	}
 
-	/* There's a lot of repetition here. This is due to the FRU having
-	 * variable-length fields. To get the information, we have to find the
-	 * size of each field, and then keep reading along and reading along
-	 * until we get all of the data that we want. We use addrptr to track
-	 * the address as we go
-	 */
-
-	/* The first fields are all of size 1-byte, from 0-7 are offsets that
-	 * contain information that isn't useful to us.
-	 * Bytes 8-a are all 1-byte and refer to the size of the entire struct,
-	 * and the language field, so just start from 0xb, manufacturer size
-	 */
-	addrptr = FRU_EEPROM_MADDR + 0xb;
-	size = amdgpu_fru_read_eeprom(adev, addrptr, buf, sizeof(buf));
-	if (size < 1) {
-		DRM_ERROR("Failed to read FRU Manufacturer, ret:%d", size);
-		return -EINVAL;
+	/* Read the IPMI Common header */
+	len = amdgpu_eeprom_read(adev->pm.fru_eeprom_i2c_bus, fru_addr, buf,
+				 sizeof(buf));
+	if (len != 8) {
+		DRM_ERROR("Couldn't read the IPMI Common Header: %d", len);
+		return len < 0 ? len : -EIO;
 	}
 
-	/* Increment the addrptr by the size of the field, and 1 due to the
-	 * size field being 1 byte. This pattern continues below.
-	 */
-	addrptr += size + 1;
-	size = amdgpu_fru_read_eeprom(adev, addrptr, buf, sizeof(buf));
-	if (size < 1) {
-		DRM_ERROR("Failed to read FRU product name, ret:%d", size);
-		return -EINVAL;
+	if (buf[0] != 1) {
+		DRM_ERROR("Bad IPMI Common Header version: 0x%02x", buf[0]);
+		return -EIO;
 	}
 
-	len = size;
-	if (len >= AMDGPU_PRODUCT_NAME_LEN) {
-		DRM_WARN("FRU Product Name is larger than %d characters. This is likely a mistake",
-				AMDGPU_PRODUCT_NAME_LEN);
-		len = AMDGPU_PRODUCT_NAME_LEN - 1;
-	}
-	memcpy(adev->product_name, buf, len);
-	adev->product_name[len] = '\0';
-
-	addrptr += size + 1;
-	size = amdgpu_fru_read_eeprom(adev, addrptr, buf, sizeof(buf));
-	if (size < 1) {
-		DRM_ERROR("Failed to read FRU product number, ret:%d", size);
-		return -EINVAL;
+	for (csum = 0; len > 0; len--)
+		csum += buf[len - 1];
+	if (csum) {
+		DRM_ERROR("Bad IPMI Common Header checksum: 0x%02x", csum);
+		return -EIO;
 	}
 
-	len = size;
-	/* Product number should only be 16 characters. Any more,
-	 * and something could be wrong. Cap it at 16 to be safe
-	 */
-	if (len >= sizeof(adev->product_number)) {
-		DRM_WARN("FRU Product Number is larger than 16 characters. This is likely a mistake");
-		len = sizeof(adev->product_number) - 1;
-	}
-	memcpy(adev->product_number, buf, len);
-	adev->product_number[len] = '\0';
+	/* Get the offset to the Product Info Area (PIA). */
+	addr = buf[4] * 8;
+	if (!addr)
+		return 0;
 
-	addrptr += size + 1;
-	size = amdgpu_fru_read_eeprom(adev, addrptr, buf, sizeof(buf));
+	/* Get the absolute address to the PIA. */
+	addr += fru_addr;
 
-	if (size < 1) {
-		DRM_ERROR("Failed to read FRU product version, ret:%d", size);
-		return -EINVAL;
+	/* Read the header of the PIA. */
+	len = amdgpu_eeprom_read(adev->pm.fru_eeprom_i2c_bus, addr, buf, 3);
+	if (len != 3) {
+		DRM_ERROR("Couldn't read the Product Info Area header: %d", len);
+		return len < 0 ? len : -EIO;
 	}
 
-	addrptr += size + 1;
-	size = amdgpu_fru_read_eeprom(adev, addrptr, buf, sizeof(buf));
+	if (buf[0] != 1) {
+		DRM_ERROR("Bad IPMI Product Info Area version: 0x%02x", buf[0]);
+		return -EIO;
+	}
 
-	if (size < 1) {
-		DRM_ERROR("Failed to read FRU serial number, ret:%d", size);
-		return -EINVAL;
+	size = buf[1] * 8;
+	pia = kzalloc(size, GFP_KERNEL);
+	if (!pia)
+		return -ENOMEM;
+
+	/* Read the whole PIA. */
+	len = amdgpu_eeprom_read(adev->pm.fru_eeprom_i2c_bus, addr, pia, size);
+	if (len != size) {
+		kfree(pia);
+		DRM_ERROR("Couldn't read the Product Info Area: %d", len);
+		return len < 0 ? len : -EIO;
 	}
 
-	len = size;
-	/* Serial number should only be 16 characters. Any more,
-	 * and something could be wrong. Cap it at 16 to be safe
-	 */
-	if (len >= sizeof(adev->serial)) {
-		DRM_WARN("FRU Serial Number is larger than 16 characters. This is likely a mistake");
-		len = sizeof(adev->serial) - 1;
+	for (csum = 0; size > 0; size--)
+		csum += pia[size - 1];
+	if (csum) {
+		DRM_ERROR("Bad Product Info Area checksum: 0x%02x", csum);
+		return -EIO;
 	}
-	memcpy(adev->serial, buf, len);
-	adev->serial[len] = '\0';
 
+	/* Now extract useful information from the PIA.
+	 *
+	 * Skip the Manufacturer Name at [3] and go directly to
+	 * the Product Name field.
+	 */
+	addr = 3 + 1 + (pia[3] & 0x3F);
+	if (addr + 1 >= len)
+		goto Out;
+	memcpy(adev->product_name, pia + addr + 1,
+	       min_t(size_t,
+		     sizeof(adev->product_name),
+		     pia[addr] & 0x3F));
+	adev->product_name[sizeof(adev->product_name) - 1] = '\0';
+
+	/* Go to the Product Part/Model Number field. */
+	addr += 1 + (pia[addr] & 0x3F);
+	if (addr + 1 >= len)
+		goto Out;
+	memcpy(adev->product_number, pia + addr + 1,
+	       min_t(size_t,
+		     sizeof(adev->product_number),
+		     pia[addr] & 0x3F));
+	adev->product_number[sizeof(adev->product_number) - 1] = '\0';
+
+	/* Go to the Product Version field. */
+	addr += 1 + (pia[addr] & 0x3F);
+
+	/* Go to the Product Serial Number field. */
+	addr += 1 + (pia[addr] & 0x3F);
+	if (addr + 1 >= len)
+		goto Out;
+	memcpy(adev->serial, pia + addr + 1, min_t(size_t,
+						   sizeof(adev->serial),
+						   pia[addr] & 0x3F));
+	adev->serial[sizeof(adev->serial) - 1] = '\0';
+Out:
+	kfree(pia);
 	return 0;
 }
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c
index 8ef31d687ef3..a0780a4e3e61 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c
@@ -38,6 +38,7 @@
 #include "amdgpu.h"
 #include "amdgpu_display.h"
 #include "amdgpu_dma_buf.h"
+#include "amdgpu_hmm.h"
 #include "amdgpu_xgmi.h"
 
 static const struct drm_gem_object_funcs amdgpu_gem_object_funcs;
@@ -87,7 +88,7 @@ static void amdgpu_gem_object_free(struct drm_gem_object *gobj)
 	struct amdgpu_bo *robj = gem_to_amdgpu_bo(gobj);
 
 	if (robj) {
-		amdgpu_mn_unregister(robj);
+		amdgpu_hmm_unregister(robj);
 		amdgpu_bo_unref(&robj);
 	}
 }
@@ -378,6 +379,7 @@ int amdgpu_gem_userptr_ioctl(struct drm_device *dev, void *data,
 	struct amdgpu_device *adev = drm_to_adev(dev);
 	struct drm_amdgpu_gem_userptr *args = data;
 	struct drm_gem_object *gobj;
+	struct hmm_range *range;
 	struct amdgpu_bo *bo;
 	uint32_t handle;
 	int r;
@@ -413,14 +415,13 @@ int amdgpu_gem_userptr_ioctl(struct drm_device *dev, void *data,
 	if (r)
 		goto release_object;
 
-	if (args->flags & AMDGPU_GEM_USERPTR_REGISTER) {
-		r = amdgpu_mn_register(bo, args->addr);
-		if (r)
-			goto release_object;
-	}
+	r = amdgpu_hmm_register(bo, args->addr);
+	if (r)
+		goto release_object;
 
 	if (args->flags & AMDGPU_GEM_USERPTR_VALIDATE) {
-		r = amdgpu_ttm_tt_get_user_pages(bo, bo->tbo.ttm->pages);
+		r = amdgpu_ttm_tt_get_user_pages(bo, bo->tbo.ttm->pages,
+						 &range);
 		if (r)
 			goto release_object;
 
@@ -443,7 +444,7 @@ int amdgpu_gem_userptr_ioctl(struct drm_device *dev, void *data,
 
 user_pages_done:
 	if (args->flags & AMDGPU_GEM_USERPTR_VALIDATE)
-		amdgpu_ttm_tt_get_user_pages_done(bo->tbo.ttm);
+		amdgpu_ttm_tt_get_user_pages_done(bo->tbo.ttm, range);
 
 release_object:
 	drm_gem_object_put(gobj);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c
index 9546adc8a76f..23692e5d4d13 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c
@@ -583,10 +583,14 @@ void amdgpu_gfx_off_ctrl(struct amdgpu_device *adev, bool enable)
 		if (adev->gfx.gfx_off_req_count == 0 &&
 		    !adev->gfx.gfx_off_state) {
 			/* If going to s2idle, no need to wait */
-			if (adev->in_s0ix)
-				delay = GFX_OFF_NO_DELAY;
-			schedule_delayed_work(&adev->gfx.gfx_off_delay_work,
+			if (adev->in_s0ix) {
+				if (!amdgpu_dpm_set_powergating_by_smu(adev,
+						AMD_IP_BLOCK_TYPE_GFX, true))
+					adev->gfx.gfx_off_state = true;
+			} else {
+				schedule_delayed_work(&adev->gfx.gfx_off_delay_work,
 					      delay);
+			}
 		}
 	} else {
 		if (adev->gfx.gfx_off_req_count == 0) {
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c
index 9c0d9baab4e2..4365ede42855 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c
@@ -657,7 +657,7 @@ void amdgpu_gmc_get_vbios_allocations(struct amdgpu_device *adev)
 	}
 
 	if (amdgpu_sriov_vf(adev) ||
-	    !amdgpu_device_ip_get_ip_block(adev, AMD_IP_BLOCK_TYPE_DCE)) {
+	    !amdgpu_device_has_display_hardware(adev)) {
 		size = 0;
 	} else {
 		size = amdgpu_gmc_get_vbios_fb_size(adev);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_mn.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_hmm.c
index b86c0b8252a5..a48ea62b12b0 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_mn.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_hmm.c
@@ -49,9 +49,10 @@
 
 #include "amdgpu.h"
 #include "amdgpu_amdkfd.h"
+#include "amdgpu_hmm.h"
 
 /**
- * amdgpu_mn_invalidate_gfx - callback to notify about mm change
+ * amdgpu_hmm_invalidate_gfx - callback to notify about mm change
  *
  * @mni: the range (mm) is about to update
  * @range: details on the invalidation
@@ -60,9 +61,9 @@
  * Block for operations on BOs to finish and mark pages as accessed and
  * potentially dirty.
  */
-static bool amdgpu_mn_invalidate_gfx(struct mmu_interval_notifier *mni,
-				     const struct mmu_notifier_range *range,
-				     unsigned long cur_seq)
+static bool amdgpu_hmm_invalidate_gfx(struct mmu_interval_notifier *mni,
+				      const struct mmu_notifier_range *range,
+				      unsigned long cur_seq)
 {
 	struct amdgpu_bo *bo = container_of(mni, struct amdgpu_bo, notifier);
 	struct amdgpu_device *adev = amdgpu_ttm_adev(bo->tbo.bdev);
@@ -83,12 +84,12 @@ static bool amdgpu_mn_invalidate_gfx(struct mmu_interval_notifier *mni,
 	return true;
 }
 
-static const struct mmu_interval_notifier_ops amdgpu_mn_gfx_ops = {
-	.invalidate = amdgpu_mn_invalidate_gfx,
+static const struct mmu_interval_notifier_ops amdgpu_hmm_gfx_ops = {
+	.invalidate = amdgpu_hmm_invalidate_gfx,
 };
 
 /**
- * amdgpu_mn_invalidate_hsa - callback to notify about mm change
+ * amdgpu_hmm_invalidate_hsa - callback to notify about mm change
  *
  * @mni: the range (mm) is about to update
  * @range: details on the invalidation
@@ -97,9 +98,9 @@ static const struct mmu_interval_notifier_ops amdgpu_mn_gfx_ops = {
  * We temporarily evict the BO attached to this range. This necessitates
  * evicting all user-mode queues of the process.
  */
-static bool amdgpu_mn_invalidate_hsa(struct mmu_interval_notifier *mni,
-				     const struct mmu_notifier_range *range,
-				     unsigned long cur_seq)
+static bool amdgpu_hmm_invalidate_hsa(struct mmu_interval_notifier *mni,
+				      const struct mmu_notifier_range *range,
+				      unsigned long cur_seq)
 {
 	struct amdgpu_bo *bo = container_of(mni, struct amdgpu_bo, notifier);
 	struct amdgpu_device *adev = amdgpu_ttm_adev(bo->tbo.bdev);
@@ -117,12 +118,12 @@ static bool amdgpu_mn_invalidate_hsa(struct mmu_interval_notifier *mni,
 	return true;
 }
 
-static const struct mmu_interval_notifier_ops amdgpu_mn_hsa_ops = {
-	.invalidate = amdgpu_mn_invalidate_hsa,
+static const struct mmu_interval_notifier_ops amdgpu_hmm_hsa_ops = {
+	.invalidate = amdgpu_hmm_invalidate_hsa,
 };
 
 /**
- * amdgpu_mn_register - register a BO for notifier updates
+ * amdgpu_hmm_register - register a BO for notifier updates
  *
  * @bo: amdgpu buffer object
  * @addr: userptr addr we should monitor
@@ -130,25 +131,25 @@ static const struct mmu_interval_notifier_ops amdgpu_mn_hsa_ops = {
  * Registers a mmu_notifier for the given BO at the specified address.
  * Returns 0 on success, -ERRNO if anything goes wrong.
  */
-int amdgpu_mn_register(struct amdgpu_bo *bo, unsigned long addr)
+int amdgpu_hmm_register(struct amdgpu_bo *bo, unsigned long addr)
 {
 	if (bo->kfd_bo)
 		return mmu_interval_notifier_insert(&bo->notifier, current->mm,
 						    addr, amdgpu_bo_size(bo),
-						    &amdgpu_mn_hsa_ops);
+						    &amdgpu_hmm_hsa_ops);
 	return mmu_interval_notifier_insert(&bo->notifier, current->mm, addr,
 					    amdgpu_bo_size(bo),
-					    &amdgpu_mn_gfx_ops);
+					    &amdgpu_hmm_gfx_ops);
 }
 
 /**
- * amdgpu_mn_unregister - unregister a BO for notifier updates
+ * amdgpu_hmm_unregister - unregister a BO for notifier updates
  *
  * @bo: amdgpu buffer object
  *
  * Remove any registration of mmu notifier updates from the buffer object.
  */
-void amdgpu_mn_unregister(struct amdgpu_bo *bo)
+void amdgpu_hmm_unregister(struct amdgpu_bo *bo)
 {
 	if (!bo->notifier.mm)
 		return;
@@ -157,10 +158,9 @@ void amdgpu_mn_unregister(struct amdgpu_bo *bo)
 }
 
 int amdgpu_hmm_range_get_pages(struct mmu_interval_notifier *notifier,
-			       struct mm_struct *mm, struct page **pages,
-			       uint64_t start, uint64_t npages,
-			       struct hmm_range **phmm_range, bool readonly,
-			       bool mmap_locked, void *owner)
+			       uint64_t start, uint64_t npages, bool readonly,
+			       void *owner, struct page **pages,
+			       struct hmm_range **phmm_range)
 {
 	struct hmm_range *hmm_range;
 	unsigned long timeout;
@@ -193,14 +193,7 @@ int amdgpu_hmm_range_get_pages(struct mmu_interval_notifier *notifier,
 
 retry:
 	hmm_range->notifier_seq = mmu_interval_read_begin(notifier);
-
-	if (likely(!mmap_locked))
-		mmap_read_lock(mm);
-
 	r = hmm_range_fault(hmm_range);
-
-	if (likely(!mmap_locked))
-		mmap_read_unlock(mm);
 	if (unlikely(r)) {
 		/*
 		 * FIXME: This timeout should encompass the retry from
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_mn.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_hmm.h
index 14a3c1864085..13ed94d3b01b 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_mn.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_hmm.h
@@ -31,23 +31,22 @@
 #include <linux/interval_tree.h>
 
 int amdgpu_hmm_range_get_pages(struct mmu_interval_notifier *notifier,
-			       struct mm_struct *mm, struct page **pages,
-			       uint64_t start, uint64_t npages,
-			       struct hmm_range **phmm_range, bool readonly,
-			       bool mmap_locked, void *owner);
+			       uint64_t start, uint64_t npages, bool readonly,
+			       void *owner, struct page **pages,
+			       struct hmm_range **phmm_range);
 int amdgpu_hmm_range_get_pages_done(struct hmm_range *hmm_range);
 
 #if defined(CONFIG_HMM_MIRROR)
-int amdgpu_mn_register(struct amdgpu_bo *bo, unsigned long addr);
-void amdgpu_mn_unregister(struct amdgpu_bo *bo);
+int amdgpu_hmm_register(struct amdgpu_bo *bo, unsigned long addr);
+void amdgpu_hmm_unregister(struct amdgpu_bo *bo);
 #else
-static inline int amdgpu_mn_register(struct amdgpu_bo *bo, unsigned long addr)
+static inline int amdgpu_hmm_register(struct amdgpu_bo *bo, unsigned long addr)
 {
 	DRM_WARN_ONCE("HMM_MIRROR kernel config option is not enabled, "
 		      "add CONFIG_ZONE_DEVICE=y in config file to fix this\n");
 	return -ENODEV;
 }
-static inline void amdgpu_mn_unregister(struct amdgpu_bo *bo) {}
+static inline void amdgpu_hmm_unregister(struct amdgpu_bo *bo) {}
 #endif
 
 #endif
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_irq.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_irq.c
index 89011bae7588..a6aef488a822 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_irq.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_irq.c
@@ -101,41 +101,6 @@ const char *soc15_ih_clientid_name[] = {
 };
 
 /**
- * amdgpu_hotplug_work_func - work handler for display hotplug event
- *
- * @work: work struct pointer
- *
- * This is the hotplug event work handler (all ASICs).
- * The work gets scheduled from the IRQ handler if there
- * was a hotplug interrupt.  It walks through the connector table
- * and calls hotplug handler for each connector. After this, it sends
- * a DRM hotplug event to alert userspace.
- *
- * This design approach is required in order to defer hotplug event handling
- * from the IRQ handler to a work handler because hotplug handler has to use
- * mutexes which cannot be locked in an IRQ handler (since &mutex_lock may
- * sleep).
- */
-static void amdgpu_hotplug_work_func(struct work_struct *work)
-{
-	struct amdgpu_device *adev = container_of(work, struct amdgpu_device,
-						  hotplug_work);
-	struct drm_device *dev = adev_to_drm(adev);
-	struct drm_mode_config *mode_config = &dev->mode_config;
-	struct drm_connector *connector;
-	struct drm_connector_list_iter iter;
-
-	mutex_lock(&mode_config->mutex);
-	drm_connector_list_iter_begin(dev, &iter);
-	drm_for_each_connector_iter(connector, &iter)
-		amdgpu_connector_hotplug(connector);
-	drm_connector_list_iter_end(&iter);
-	mutex_unlock(&mode_config->mutex);
-	/* Just fire off a uevent and let userspace tell us what to do */
-	drm_helper_hpd_irq_event(dev);
-}
-
-/**
  * amdgpu_irq_disable_all - disable *all* interrupts
  *
  * @adev: amdgpu device pointer
@@ -317,21 +282,6 @@ int amdgpu_irq_init(struct amdgpu_device *adev)
 		}
 	}
 
-	if (!amdgpu_device_has_dc_support(adev)) {
-		if (!adev->enable_virtual_display)
-			/* Disable vblank IRQs aggressively for power-saving */
-			/* XXX: can this be enabled for DC? */
-			adev_to_drm(adev)->vblank_disable_immediate = true;
-
-		r = drm_vblank_init(adev_to_drm(adev), adev->mode_info.num_crtc);
-		if (r)
-			return r;
-
-		/* Pre-DCE11 */
-		INIT_WORK(&adev->hotplug_work,
-				amdgpu_hotplug_work_func);
-	}
-
 	INIT_WORK(&adev->irq.ih1_work, amdgpu_irq_handle_ih1);
 	INIT_WORK(&adev->irq.ih2_work, amdgpu_irq_handle_ih2);
 	INIT_WORK(&adev->irq.ih_soft_work, amdgpu_irq_handle_ih_soft);
@@ -345,11 +295,8 @@ int amdgpu_irq_init(struct amdgpu_device *adev)
 	/* PCI devices require shared interrupts. */
 	r = request_irq(irq, amdgpu_irq_handler, IRQF_SHARED, adev_to_drm(adev)->driver->name,
 			adev_to_drm(adev));
-	if (r) {
-		if (!amdgpu_device_has_dc_support(adev))
-			flush_work(&adev->hotplug_work);
+	if (r)
 		return r;
-	}
 	adev->irq.installed = true;
 	adev->irq.irq = irq;
 	adev_to_drm(adev)->max_vblank_count = 0x00ffffff;
@@ -366,9 +313,6 @@ void amdgpu_irq_fini_hw(struct amdgpu_device *adev)
 		adev->irq.installed = false;
 		if (adev->irq.msi_enabled)
 			pci_free_irq_vectors(adev->pdev);
-
-		if (!amdgpu_device_has_dc_support(adev))
-			flush_work(&adev->hotplug_work);
 	}
 
 	amdgpu_ih_ring_fini(adev, &adev->irq.ih_soft);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.c
index de182bfcf85f..6f81ed4fb0d9 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.c
@@ -235,3 +235,20 @@ int amdgpu_jpeg_process_poison_irq(struct amdgpu_device *adev,
 
 	return 0;
 }
+
+void jpeg_set_ras_funcs(struct amdgpu_device *adev)
+{
+	if (!adev->jpeg.ras)
+		return;
+
+	amdgpu_ras_register_ras_block(adev, &adev->jpeg.ras->ras_block);
+
+	strcpy(adev->jpeg.ras->ras_block.ras_comm.name, "jpeg");
+	adev->jpeg.ras->ras_block.ras_comm.block = AMDGPU_RAS_BLOCK__JPEG;
+	adev->jpeg.ras->ras_block.ras_comm.type = AMDGPU_RAS_ERROR__POISON;
+	adev->jpeg.ras_if = &adev->jpeg.ras->ras_block.ras_comm;
+
+	/* If don't define special ras_late_init function, use default ras_late_init */
+	if (!adev->jpeg.ras->ras_block.ras_late_init)
+		adev->jpeg.ras->ras_block.ras_late_init = amdgpu_ras_block_late_init;
+}
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.h
index 635dca59a70a..e8ca3e32ad52 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.h
@@ -72,5 +72,6 @@ int amdgpu_jpeg_dec_ring_test_ib(struct amdgpu_ring *ring, long timeout);
 int amdgpu_jpeg_process_poison_irq(struct amdgpu_device *adev,
 				struct amdgpu_irq_src *source,
 				struct amdgpu_iv_entry *entry);
+void jpeg_set_ras_funcs(struct amdgpu_device *adev);
 
 #endif /*__AMDGPU_JPEG_H__*/
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c
index 2fcb5bfbef89..98dbf4e5aae9 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c
@@ -52,6 +52,32 @@ static int psp_load_smu_fw(struct psp_context *psp);
 static int psp_rap_terminate(struct psp_context *psp);
 static int psp_securedisplay_terminate(struct psp_context *psp);
 
+static int psp_ring_init(struct psp_context *psp,
+			 enum psp_ring_type ring_type)
+{
+	int ret = 0;
+	struct psp_ring *ring;
+	struct amdgpu_device *adev = psp->adev;
+
+	ring = &psp->km_ring;
+
+	ring->ring_type = ring_type;
+
+	/* allocate 4k Page of Local Frame Buffer memory for ring */
+	ring->ring_size = 0x1000;
+	ret = amdgpu_bo_create_kernel(adev, ring->ring_size, PAGE_SIZE,
+				      AMDGPU_GEM_DOMAIN_VRAM,
+				      &adev->firmware.rbuf,
+				      &ring->ring_mem_mc_addr,
+				      (void **)&ring->ring_mem);
+	if (ret) {
+		ring->ring_size = 0;
+		return ret;
+	}
+
+	return 0;
+}
+
 /*
  * Due to DF Cstate management centralized to PMFW, the firmware
  * loading sequence will be updated as below:
@@ -1526,11 +1552,6 @@ int psp_ras_initialize(struct psp_context *psp)
 	if (amdgpu_sriov_vf(adev))
 		return 0;
 
-	if (psp->ras_context.context.initialized) {
-		dev_warn(adev->dev, "RAS WARN: TA has already been loaded\n");
-		return 0;
-	}
-
 	if (!adev->psp.ras_context.context.bin_desc.size_bytes ||
 	    !adev->psp.ras_context.context.bin_desc.start_addr) {
 		dev_info(adev->dev, "RAS: optional ras ta ucode is not available\n");
@@ -1602,6 +1623,9 @@ int psp_ras_initialize(struct psp_context *psp)
 	else {
 		if (ras_cmd->ras_status)
 			dev_warn(psp->adev->dev, "RAS Init Status: 0x%X\n", ras_cmd->ras_status);
+
+		/* fail to load RAS TA */
+		psp->ras_context.context.initialized = false;
 	}
 
 	return ret;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.h
index cbd4194a2883..cf4f60c66122 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.h
@@ -118,7 +118,6 @@ struct psp_funcs
 	int (*bootloader_load_dbg_drv)(struct psp_context *psp);
 	int (*bootloader_load_ras_drv)(struct psp_context *psp);
 	int (*bootloader_load_sos)(struct psp_context *psp);
-	int (*ring_init)(struct psp_context *psp, enum psp_ring_type ring_type);
 	int (*ring_create)(struct psp_context *psp,
 			   enum psp_ring_type ring_type);
 	int (*ring_stop)(struct psp_context *psp,
@@ -396,7 +395,6 @@ struct amdgpu_psp_funcs {
 };
 
 
-#define psp_ring_init(psp, type) (psp)->funcs->ring_init((psp), (type))
 #define psp_ring_create(psp, type) (psp)->funcs->ring_create((psp), (type))
 #define psp_ring_stop(psp, type) (psp)->funcs->ring_stop((psp), (type))
 #define psp_ring_destroy(psp, type) ((psp)->funcs->ring_destroy((psp), (type)))
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c
index 693bce07eb46..077404a9c935 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c
@@ -1948,7 +1948,12 @@ static void amdgpu_ras_do_recovery(struct work_struct *work)
 
 		reset_context.method = AMD_RESET_METHOD_NONE;
 		reset_context.reset_req_dev = adev;
-		clear_bit(AMDGPU_NEED_FULL_RESET, &reset_context.flags);
+
+		/* Perform full reset in fatal error mode */
+		if (!amdgpu_ras_is_poison_mode_supported(ras->adev))
+			set_bit(AMDGPU_NEED_FULL_RESET, &reset_context.flags);
+		else
+			clear_bit(AMDGPU_NEED_FULL_RESET, &reset_context.flags);
 
 		amdgpu_device_gpu_recover(ras->adev, NULL, &reset_context);
 	}
@@ -2343,7 +2348,8 @@ static void amdgpu_ras_check_supported(struct amdgpu_device *adev)
 				adev->ras_hw_enabled |= ~(1 << AMDGPU_RAS_BLOCK__UMC |
 							    1 << AMDGPU_RAS_BLOCK__DF);
 
-				if (adev->ip_versions[VCN_HWIP][0] == IP_VERSION(2, 6, 0))
+				if (adev->ip_versions[VCN_HWIP][0] == IP_VERSION(2, 6, 0) ||
+				    adev->ip_versions[VCN_HWIP][0] == IP_VERSION(4, 0, 0))
 					adev->ras_hw_enabled |= (1 << AMDGPU_RAS_BLOCK__VCN |
 							1 << AMDGPU_RAS_BLOCK__JPEG);
 				else
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.c
index 7268ae65c140..2d9f3f4cd79e 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.c
@@ -33,12 +33,29 @@
 
 #include "amdgpu_reset.h"
 
-#define EEPROM_I2C_MADDR_VEGA20         0x0
-#define EEPROM_I2C_MADDR_ARCTURUS       0x40000
-#define EEPROM_I2C_MADDR_ARCTURUS_D342  0x0
-#define EEPROM_I2C_MADDR_SIENNA_CICHLID 0x0
-#define EEPROM_I2C_MADDR_ALDEBARAN      0x0
-#define EEPROM_I2C_MADDR_54H            (0x54UL << 16)
+/* These are memory addresses as would be seen by one or more EEPROM
+ * chips strung on the I2C bus, usually by manipulating pins 1-3 of a
+ * set of EEPROM devices. They form a continuous memory space.
+ *
+ * The I2C device address includes the device type identifier, 1010b,
+ * which is a reserved value and indicates that this is an I2C EEPROM
+ * device. It also includes the top 3 bits of the 19 bit EEPROM memory
+ * address, namely bits 18, 17, and 16. This makes up the 7 bit
+ * address sent on the I2C bus with bit 0 being the direction bit,
+ * which is not represented here, and sent by the hardware directly.
+ *
+ * For instance,
+ *   50h = 1010000b => device type identifier 1010b, bits 18:16 = 000b, address 0.
+ *   54h = 1010100b => --"--, bits 18:16 = 100b, address 40000h.
+ *   56h = 1010110b => --"--, bits 18:16 = 110b, address 60000h.
+ * Depending on the size of the I2C EEPROM device(s), bits 18:16 may
+ * address memory in a device or a device on the I2C bus, depending on
+ * the status of pins 1-3. See top of amdgpu_eeprom.c.
+ *
+ * The RAS table lives either at address 0 or address 40000h of EEPROM.
+ */
+#define EEPROM_I2C_MADDR_0      0x0
+#define EEPROM_I2C_MADDR_4      0x40000
 
 /*
  * The 2 macros bellow represent the actual size in bytes that
@@ -117,9 +134,9 @@ static bool __get_eeprom_i2c_addr_arct(struct amdgpu_device *adev,
 	if (strnstr(atom_ctx->vbios_version,
 	            "D342",
 		    sizeof(atom_ctx->vbios_version)))
-		control->i2c_address = EEPROM_I2C_MADDR_ARCTURUS_D342;
+		control->i2c_address = EEPROM_I2C_MADDR_0;
 	else
-		control->i2c_address = EEPROM_I2C_MADDR_ARCTURUS;
+		control->i2c_address = EEPROM_I2C_MADDR_4;
 
 	return true;
 }
@@ -130,7 +147,7 @@ static bool __get_eeprom_i2c_addr_ip_discovery(struct amdgpu_device *adev,
 	switch (adev->ip_versions[MP1_HWIP][0]) {
 	case IP_VERSION(13, 0, 0):
 	case IP_VERSION(13, 0, 10):
-		control->i2c_address = EEPROM_I2C_MADDR_54H;
+		control->i2c_address = EEPROM_I2C_MADDR_4;
 		return true;
 	default:
 		return false;
@@ -140,6 +157,7 @@ static bool __get_eeprom_i2c_addr_ip_discovery(struct amdgpu_device *adev,
 static bool __get_eeprom_i2c_addr(struct amdgpu_device *adev,
 				  struct amdgpu_ras_eeprom_control *control)
 {
+	struct atom_context *atom_ctx = adev->mode_info.atom_context;
 	u8 i2c_addr;
 
 	if (!control)
@@ -162,18 +180,22 @@ static bool __get_eeprom_i2c_addr(struct amdgpu_device *adev,
 
 	switch (adev->asic_type) {
 	case CHIP_VEGA20:
-		control->i2c_address = EEPROM_I2C_MADDR_VEGA20;
+		control->i2c_address = EEPROM_I2C_MADDR_0;
 		break;
 
 	case CHIP_ARCTURUS:
 		return __get_eeprom_i2c_addr_arct(adev, control);
 
 	case CHIP_SIENNA_CICHLID:
-		control->i2c_address = EEPROM_I2C_MADDR_SIENNA_CICHLID;
+		control->i2c_address = EEPROM_I2C_MADDR_0;
 		break;
 
 	case CHIP_ALDEBARAN:
-		control->i2c_address = EEPROM_I2C_MADDR_ALDEBARAN;
+		if (strnstr(atom_ctx->vbios_version, "D673",
+			    sizeof(atom_ctx->vbios_version)))
+			control->i2c_address = EEPROM_I2C_MADDR_4;
+		else
+			control->i2c_address = EEPROM_I2C_MADDR_0;
 		break;
 
 	case CHIP_IP_DISCOVERY:
@@ -185,7 +207,7 @@ static bool __get_eeprom_i2c_addr(struct amdgpu_device *adev,
 
 	switch (adev->ip_versions[MP1_HWIP][0]) {
 	case IP_VERSION(13, 0, 0):
-		control->i2c_address = EEPROM_I2C_MADDR_54H;
+		control->i2c_address = EEPROM_I2C_MADDR_4;
 		break;
 
 	default:
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
index 0ba87ca6f318..7b5074e776f4 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
@@ -58,6 +58,7 @@
 #include "amdgpu_amdkfd.h"
 #include "amdgpu_sdma.h"
 #include "amdgpu_ras.h"
+#include "amdgpu_hmm.h"
 #include "amdgpu_atomfirmware.h"
 #include "amdgpu_res_cursor.h"
 #include "bif/bif_4_1_d.h"
@@ -634,9 +635,6 @@ struct amdgpu_ttm_tt {
 	struct task_struct	*usertask;
 	uint32_t		userflags;
 	bool			bound;
-#if IS_ENABLED(CONFIG_DRM_AMDGPU_USERPTR)
-	struct hmm_range	*range;
-#endif
 };
 
 #define ttm_to_amdgpu_ttm_tt(ptr)	container_of(ptr, struct amdgpu_ttm_tt, ttm)
@@ -649,7 +647,8 @@ struct amdgpu_ttm_tt {
  * Calling function must call amdgpu_ttm_tt_userptr_range_done() once and only
  * once afterwards to stop HMM tracking
  */
-int amdgpu_ttm_tt_get_user_pages(struct amdgpu_bo *bo, struct page **pages)
+int amdgpu_ttm_tt_get_user_pages(struct amdgpu_bo *bo, struct page **pages,
+				 struct hmm_range **range)
 {
 	struct ttm_tt *ttm = bo->tbo.ttm;
 	struct amdgpu_ttm_tt *gtt = ttm_to_amdgpu_ttm_tt(ttm);
@@ -659,16 +658,15 @@ int amdgpu_ttm_tt_get_user_pages(struct amdgpu_bo *bo, struct page **pages)
 	bool readonly;
 	int r = 0;
 
+	/* Make sure get_user_pages_done() can cleanup gracefully */
+	*range = NULL;
+
 	mm = bo->notifier.mm;
 	if (unlikely(!mm)) {
 		DRM_DEBUG_DRIVER("BO is not registered?\n");
 		return -EFAULT;
 	}
 
-	/* Another get_user_pages is running at the same time?? */
-	if (WARN_ON(gtt->range))
-		return -EFAULT;
-
 	if (!mmget_not_zero(mm)) /* Happens during process shutdown */
 		return -ESRCH;
 
@@ -685,9 +683,8 @@ int amdgpu_ttm_tt_get_user_pages(struct amdgpu_bo *bo, struct page **pages)
 	}
 
 	readonly = amdgpu_ttm_tt_is_readonly(ttm);
-	r = amdgpu_hmm_range_get_pages(&bo->notifier, mm, pages, start,
-				       ttm->num_pages, &gtt->range, readonly,
-				       true, NULL);
+	r = amdgpu_hmm_range_get_pages(&bo->notifier, start, ttm->num_pages,
+				       readonly, NULL, pages, range);
 out_unlock:
 	mmap_read_unlock(mm);
 	if (r)
@@ -704,30 +701,24 @@ out_unlock:
  *
  * Returns: true if pages are still valid
  */
-bool amdgpu_ttm_tt_get_user_pages_done(struct ttm_tt *ttm)
+bool amdgpu_ttm_tt_get_user_pages_done(struct ttm_tt *ttm,
+				       struct hmm_range *range)
 {
 	struct amdgpu_ttm_tt *gtt = ttm_to_amdgpu_ttm_tt(ttm);
-	bool r = false;
 
-	if (!gtt || !gtt->userptr)
+	if (!gtt || !gtt->userptr || !range)
 		return false;
 
 	DRM_DEBUG_DRIVER("user_pages_done 0x%llx pages 0x%x\n",
 		gtt->userptr, ttm->num_pages);
 
-	WARN_ONCE(!gtt->range || !gtt->range->hmm_pfns,
-		"No user pages to check\n");
-
-	if (gtt->range) {
-		/*
-		 * FIXME: Must always hold notifier_lock for this, and must
-		 * not ignore the return code.
-		 */
-		r = amdgpu_hmm_range_get_pages_done(gtt->range);
-		gtt->range = NULL;
-	}
+	WARN_ONCE(!range->hmm_pfns, "No user pages to check\n");
 
-	return !r;
+	/*
+	 * FIXME: Must always hold notifier_lock for this, and must
+	 * not ignore the return code.
+	 */
+	return !amdgpu_hmm_range_get_pages_done(range);
 }
 #endif
 
@@ -804,20 +795,6 @@ static void amdgpu_ttm_tt_unpin_userptr(struct ttm_device *bdev,
 	/* unmap the pages mapped to the device */
 	dma_unmap_sgtable(adev->dev, ttm->sg, direction, 0);
 	sg_free_table(ttm->sg);
-
-#if IS_ENABLED(CONFIG_DRM_AMDGPU_USERPTR)
-	if (gtt->range) {
-		unsigned long i;
-
-		for (i = 0; i < ttm->num_pages; i++) {
-			if (ttm->pages[i] !=
-			    hmm_pfn_to_page(gtt->range->hmm_pfns[i]))
-				break;
-		}
-
-		WARN((i == ttm->num_pages), "Missing get_user_page_done\n");
-	}
-#endif
 }
 
 static void amdgpu_ttm_gart_bind(struct amdgpu_device *adev,
@@ -1168,8 +1145,9 @@ int amdgpu_ttm_tt_get_userptr(const struct ttm_buffer_object *tbo,
  * @addr:  The address in the current tasks VM space to use
  * @flags: Requirements of userptr object.
  *
- * Called by amdgpu_gem_userptr_ioctl() to bind userptr pages
- * to current task
+ * Called by amdgpu_gem_userptr_ioctl() and kfd_ioctl_alloc_memory_of_gpu() to
+ * bind userptr pages to current task and by kfd_ioctl_acquire_vm() to
+ * initialize GPU VM for a KFD process.
  */
 int amdgpu_ttm_tt_set_userptr(struct ttm_buffer_object *bo,
 			      uint64_t addr, uint32_t flags)
@@ -1553,6 +1531,23 @@ static void amdgpu_ttm_fw_reserve_vram_fini(struct amdgpu_device *adev)
 		NULL, &adev->mman.fw_vram_usage_va);
 }
 
+/*
+ * Driver Reservation functions
+ */
+/**
+ * amdgpu_ttm_drv_reserve_vram_fini - free drv reserved vram
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * free drv reserved vram if it has been reserved.
+ */
+static void amdgpu_ttm_drv_reserve_vram_fini(struct amdgpu_device *adev)
+{
+	amdgpu_bo_free_kernel(&adev->mman.drv_vram_usage_reserved_bo,
+						  NULL,
+						  NULL);
+}
+
 /**
  * amdgpu_ttm_fw_reserve_vram_init - create bo vram reservation from fw
  *
@@ -1579,6 +1574,31 @@ static int amdgpu_ttm_fw_reserve_vram_init(struct amdgpu_device *adev)
 					  &adev->mman.fw_vram_usage_va);
 }
 
+/**
+ * amdgpu_ttm_drv_reserve_vram_init - create bo vram reservation from driver
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * create bo vram reservation from drv.
+ */
+static int amdgpu_ttm_drv_reserve_vram_init(struct amdgpu_device *adev)
+{
+	uint64_t vram_size = adev->gmc.visible_vram_size;
+
+	adev->mman.drv_vram_usage_reserved_bo = NULL;
+
+	if (adev->mman.drv_vram_usage_size == 0 ||
+	    adev->mman.drv_vram_usage_size > vram_size)
+		return 0;
+
+	return amdgpu_bo_create_kernel_at(adev,
+					  adev->mman.drv_vram_usage_start_offset,
+					  adev->mman.drv_vram_usage_size,
+					  AMDGPU_GEM_DOMAIN_VRAM,
+					  &adev->mman.drv_vram_usage_reserved_bo,
+					  NULL);
+}
+
 /*
  * Memoy training reservation functions
  */
@@ -1747,6 +1767,14 @@ int amdgpu_ttm_init(struct amdgpu_device *adev)
 	}
 
 	/*
+	 *The reserved vram for driver must be pinned to the specified
+	 *place on the VRAM, so reserve it early.
+	 */
+	r = amdgpu_ttm_drv_reserve_vram_init(adev);
+	if (r)
+		return r;
+
+	/*
 	 * only NAVI10 and onwards ASIC support for IP discovery.
 	 * If IP discovery enabled, a block of memory should be
 	 * reserved for IP discovey.
@@ -1871,6 +1899,7 @@ void amdgpu_ttm_fini(struct amdgpu_device *adev)
 	amdgpu_bo_free_kernel(&adev->mman.sdma_access_bo, NULL,
 					&adev->mman.sdma_access_ptr);
 	amdgpu_ttm_fw_reserve_vram_fini(adev);
+	amdgpu_ttm_drv_reserve_vram_fini(adev);
 
 	if (drm_dev_enter(adev_to_drm(adev), &idx)) {
 
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.h
index 6a70818039dd..b391c8d076ff 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.h
@@ -39,6 +39,8 @@
 
 #define AMDGPU_POISON	0xd0bed0be
 
+struct hmm_range;
+
 struct amdgpu_gtt_mgr {
 	struct ttm_resource_manager manager;
 	struct drm_mm mm;
@@ -84,6 +86,11 @@ struct amdgpu_mman {
 	struct amdgpu_bo	*fw_vram_usage_reserved_bo;
 	void		*fw_vram_usage_va;
 
+	/* driver VRAM reservation */
+	u64		drv_vram_usage_start_offset;
+	u64		drv_vram_usage_size;
+	struct amdgpu_bo	*drv_vram_usage_reserved_bo;
+
 	/* PAGE_SIZE'd BO for process memory r/w over SDMA. */
 	struct amdgpu_bo	*sdma_access_bo;
 	void			*sdma_access_ptr;
@@ -149,15 +156,19 @@ void amdgpu_ttm_recover_gart(struct ttm_buffer_object *tbo);
 uint64_t amdgpu_ttm_domain_start(struct amdgpu_device *adev, uint32_t type);
 
 #if IS_ENABLED(CONFIG_DRM_AMDGPU_USERPTR)
-int amdgpu_ttm_tt_get_user_pages(struct amdgpu_bo *bo, struct page **pages);
-bool amdgpu_ttm_tt_get_user_pages_done(struct ttm_tt *ttm);
+int amdgpu_ttm_tt_get_user_pages(struct amdgpu_bo *bo, struct page **pages,
+				 struct hmm_range **range);
+bool amdgpu_ttm_tt_get_user_pages_done(struct ttm_tt *ttm,
+				       struct hmm_range *range);
 #else
 static inline int amdgpu_ttm_tt_get_user_pages(struct amdgpu_bo *bo,
-					       struct page **pages)
+					       struct page **pages,
+					       struct hmm_range **range)
 {
 	return -EPERM;
 }
-static inline bool amdgpu_ttm_tt_get_user_pages_done(struct ttm_tt *ttm)
+static inline bool amdgpu_ttm_tt_get_user_pages_done(struct ttm_tt *ttm,
+						     struct hmm_range *range)
 {
 	return false;
 }
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c
index 3449145ab2bc..33f3415096f7 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c
@@ -1252,3 +1252,20 @@ int amdgpu_vcn_process_poison_irq(struct amdgpu_device *adev,
 
 	return 0;
 }
+
+void amdgpu_vcn_set_ras_funcs(struct amdgpu_device *adev)
+{
+	if (!adev->vcn.ras)
+		return;
+
+	amdgpu_ras_register_ras_block(adev, &adev->vcn.ras->ras_block);
+
+	strcpy(adev->vcn.ras->ras_block.ras_comm.name, "vcn");
+	adev->vcn.ras->ras_block.ras_comm.block = AMDGPU_RAS_BLOCK__VCN;
+	adev->vcn.ras->ras_block.ras_comm.type = AMDGPU_RAS_ERROR__POISON;
+	adev->vcn.ras_if = &adev->vcn.ras->ras_block.ras_comm;
+
+	/* If don't define special ras_late_init function, use default ras_late_init */
+	if (!adev->vcn.ras->ras_block.ras_late_init)
+		adev->vcn.ras->ras_block.ras_late_init = amdgpu_ras_block_late_init;
+}
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h
index 253ea6b159df..dbb8d68a30c6 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h
@@ -399,5 +399,6 @@ void amdgpu_debugfs_vcn_fwlog_init(struct amdgpu_device *adev,
 int amdgpu_vcn_process_poison_irq(struct amdgpu_device *adev,
 			struct amdgpu_irq_src *source,
 			struct amdgpu_iv_entry *entry);
+void amdgpu_vcn_set_ras_funcs(struct amdgpu_device *adev);
 
 #endif
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vkms.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vkms.c
index 9b81b6867db3..7cf99f752e01 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vkms.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vkms.c
@@ -511,6 +511,10 @@ static int amdgpu_vkms_sw_init(void *handle)
 			return r;
 	}
 
+	r = drm_vblank_init(adev_to_drm(adev), adev->mode_info.num_crtc);
+	if (r)
+		return r;
+
 	drm_kms_helper_poll_init(adev_to_drm(adev));
 
 	adev->mode_info.mode_config_initialized = true;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c
index 2291aa14d888..003aa9e47085 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c
@@ -143,32 +143,6 @@ int amdgpu_vm_set_pasid(struct amdgpu_device *adev, struct amdgpu_vm *vm,
 	return 0;
 }
 
-/*
- * vm eviction_lock can be taken in MMU notifiers. Make sure no reclaim-FS
- * happens while holding this lock anywhere to prevent deadlocks when
- * an MMU notifier runs in reclaim-FS context.
- */
-static inline void amdgpu_vm_eviction_lock(struct amdgpu_vm *vm)
-{
-	mutex_lock(&vm->eviction_lock);
-	vm->saved_flags = memalloc_noreclaim_save();
-}
-
-static inline int amdgpu_vm_eviction_trylock(struct amdgpu_vm *vm)
-{
-	if (mutex_trylock(&vm->eviction_lock)) {
-		vm->saved_flags = memalloc_noreclaim_save();
-		return 1;
-	}
-	return 0;
-}
-
-static inline void amdgpu_vm_eviction_unlock(struct amdgpu_vm *vm)
-{
-	memalloc_noreclaim_restore(vm->saved_flags);
-	mutex_unlock(&vm->eviction_lock);
-}
-
 /**
  * amdgpu_vm_bo_evicted - vm_bo is evicted
  *
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h
index 83acb7bd80fe..6546e786bf00 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h
@@ -492,7 +492,48 @@ void amdgpu_debugfs_vm_bo_info(struct amdgpu_vm *vm, struct seq_file *m);
  */
 static inline uint64_t amdgpu_vm_tlb_seq(struct amdgpu_vm *vm)
 {
+	unsigned long flags;
+	spinlock_t *lock;
+
+	/*
+	 * Workaround to stop racing between the fence signaling and handling
+	 * the cb. The lock is static after initially setting it up, just make
+	 * sure that the dma_fence structure isn't freed up.
+	 */
+	rcu_read_lock();
+	lock = vm->last_tlb_flush->lock;
+	rcu_read_unlock();
+
+	spin_lock_irqsave(lock, flags);
+	spin_unlock_irqrestore(lock, flags);
+
 	return atomic64_read(&vm->tlb_seq);
 }
 
+/*
+ * vm eviction_lock can be taken in MMU notifiers. Make sure no reclaim-FS
+ * happens while holding this lock anywhere to prevent deadlocks when
+ * an MMU notifier runs in reclaim-FS context.
+ */
+static inline void amdgpu_vm_eviction_lock(struct amdgpu_vm *vm)
+{
+	mutex_lock(&vm->eviction_lock);
+	vm->saved_flags = memalloc_noreclaim_save();
+}
+
+static inline bool amdgpu_vm_eviction_trylock(struct amdgpu_vm *vm)
+{
+	if (mutex_trylock(&vm->eviction_lock)) {
+		vm->saved_flags = memalloc_noreclaim_save();
+		return true;
+	}
+	return false;
+}
+
+static inline void amdgpu_vm_eviction_unlock(struct amdgpu_vm *vm)
+{
+	memalloc_noreclaim_restore(vm->saved_flags);
+	mutex_unlock(&vm->eviction_lock);
+}
+
 #endif
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm_pt.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm_pt.c
index 358b91243e37..b5f3bba851db 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm_pt.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm_pt.c
@@ -597,7 +597,9 @@ static int amdgpu_vm_pt_alloc(struct amdgpu_device *adev,
 	if (entry->bo)
 		return 0;
 
+	amdgpu_vm_eviction_unlock(vm);
 	r = amdgpu_vm_pt_create(adev, vm, cursor->level, immediate, &pt);
+	amdgpu_vm_eviction_lock(vm);
 	if (r)
 		return r;
 
diff --git a/drivers/gpu/drm/amd/amdgpu/dce_v10_0.c b/drivers/gpu/drm/amd/amdgpu/dce_v10_0.c
index 90f87b2d985b..248f1a4e915f 100644
--- a/drivers/gpu/drm/amd/amdgpu/dce_v10_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/dce_v10_0.c
@@ -2829,6 +2829,17 @@ static int dce_v10_0_sw_init(void *handle)
 	if (r)
 		return r;
 
+	/* Disable vblank IRQs aggressively for power-saving */
+	/* XXX: can this be enabled for DC? */
+	adev_to_drm(adev)->vblank_disable_immediate = true;
+
+	r = drm_vblank_init(adev_to_drm(adev), adev->mode_info.num_crtc);
+	if (r)
+		return r;
+
+	INIT_WORK(&adev->hotplug_work,
+		  amdgpu_display_hotplug_work_func);
+
 	drm_kms_helper_poll_init(adev_to_drm(adev));
 
 	adev->mode_info.mode_config_initialized = true;
@@ -2891,6 +2902,8 @@ static int dce_v10_0_hw_fini(void *handle)
 
 	dce_v10_0_pageflip_interrupt_fini(adev);
 
+	flush_work(&adev->hotplug_work);
+
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/amd/amdgpu/dce_v11_0.c b/drivers/gpu/drm/amd/amdgpu/dce_v11_0.c
index 0352de72c886..cd9c19060d89 100644
--- a/drivers/gpu/drm/amd/amdgpu/dce_v11_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/dce_v11_0.c
@@ -2948,6 +2948,17 @@ static int dce_v11_0_sw_init(void *handle)
 	if (r)
 		return r;
 
+	/* Disable vblank IRQs aggressively for power-saving */
+	/* XXX: can this be enabled for DC? */
+	adev_to_drm(adev)->vblank_disable_immediate = true;
+
+	r = drm_vblank_init(adev_to_drm(adev), adev->mode_info.num_crtc);
+	if (r)
+		return r;
+
+	INIT_WORK(&adev->hotplug_work,
+		  amdgpu_display_hotplug_work_func);
+
 	drm_kms_helper_poll_init(adev_to_drm(adev));
 
 	adev->mode_info.mode_config_initialized = true;
@@ -3021,6 +3032,8 @@ static int dce_v11_0_hw_fini(void *handle)
 
 	dce_v11_0_pageflip_interrupt_fini(adev);
 
+	flush_work(&adev->hotplug_work);
+
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/amd/amdgpu/dce_v6_0.c b/drivers/gpu/drm/amd/amdgpu/dce_v6_0.c
index 07bd16e82046..76323deecc58 100644
--- a/drivers/gpu/drm/amd/amdgpu/dce_v6_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/dce_v6_0.c
@@ -2706,6 +2706,18 @@ static int dce_v6_0_sw_init(void *handle)
 	if (r)
 		return r;
 
+	/* Disable vblank IRQs aggressively for power-saving */
+	/* XXX: can this be enabled for DC? */
+	adev_to_drm(adev)->vblank_disable_immediate = true;
+
+	r = drm_vblank_init(adev_to_drm(adev), adev->mode_info.num_crtc);
+	if (r)
+		return r;
+
+	/* Pre-DCE11 */
+	INIT_WORK(&adev->hotplug_work,
+		  amdgpu_display_hotplug_work_func);
+
 	drm_kms_helper_poll_init(adev_to_drm(adev));
 
 	return r;
@@ -2764,6 +2776,8 @@ static int dce_v6_0_hw_fini(void *handle)
 
 	dce_v6_0_pageflip_interrupt_fini(adev);
 
+	flush_work(&adev->hotplug_work);
+
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/amd/amdgpu/dce_v8_0.c b/drivers/gpu/drm/amd/amdgpu/dce_v8_0.c
index d73df100f2b3..01cf3ab111cb 100644
--- a/drivers/gpu/drm/amd/amdgpu/dce_v8_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/dce_v8_0.c
@@ -2730,6 +2730,18 @@ static int dce_v8_0_sw_init(void *handle)
 	if (r)
 		return r;
 
+	/* Disable vblank IRQs aggressively for power-saving */
+	/* XXX: can this be enabled for DC? */
+	adev_to_drm(adev)->vblank_disable_immediate = true;
+
+	r = drm_vblank_init(adev_to_drm(adev), adev->mode_info.num_crtc);
+	if (r)
+		return r;
+
+	/* Pre-DCE11 */
+	INIT_WORK(&adev->hotplug_work,
+		  amdgpu_display_hotplug_work_func);
+
 	drm_kms_helper_poll_init(adev_to_drm(adev));
 
 	adev->mode_info.mode_config_initialized = true;
@@ -2790,6 +2802,8 @@ static int dce_v8_0_hw_fini(void *handle)
 
 	dce_v8_0_pageflip_interrupt_fini(adev);
 
+	flush_work(&adev->hotplug_work);
+
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v11_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v11_0.c
index 9447999a3a48..9d2c6523f546 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v11_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v11_0.c
@@ -4392,7 +4392,6 @@ static int gfx_v11_0_hw_fini(void *handle)
 {
 	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
 	int r;
-	uint32_t tmp;
 
 	amdgpu_irq_put(adev, &adev->gfx.priv_reg_irq, 0);
 	amdgpu_irq_put(adev, &adev->gfx.priv_inst_irq, 0);
@@ -4411,15 +4410,14 @@ static int gfx_v11_0_hw_fini(void *handle)
 		amdgpu_mes_kiq_hw_fini(adev);
 	}
 
-	if (amdgpu_sriov_vf(adev)) {
-		gfx_v11_0_cp_gfx_enable(adev, false);
-		/* Program KIQ position of RLC_CP_SCHEDULERS during destroy */
-		tmp = RREG32_SOC15(GC, 0, regRLC_CP_SCHEDULERS);
-		tmp &= 0xffffff00;
-		WREG32_SOC15(GC, 0, regRLC_CP_SCHEDULERS, tmp);
-
+	if (amdgpu_sriov_vf(adev))
+		/* Remove the steps disabling CPG and clearing KIQ position,
+		 * so that CP could perform IDLE-SAVE during switch. Those
+		 * steps are necessary to avoid a DMAR error in gfx9 but it is
+		 * not reproduced on gfx11.
+		 */
 		return 0;
-	}
+
 	gfx_v11_0_cp_enable(adev, false);
 	gfx_v11_0_enable_gui_idle_interrupt(adev, false);
 
diff --git a/drivers/gpu/drm/amd/amdgpu/gfxhub_v3_0_3.c b/drivers/gpu/drm/amd/amdgpu/gfxhub_v3_0_3.c
index 716ae6f2aefe..080ff11ca305 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfxhub_v3_0_3.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfxhub_v3_0_3.c
@@ -357,18 +357,6 @@ static void gfxhub_v3_0_3_program_invalidation(struct amdgpu_device *adev)
 
 static int gfxhub_v3_0_3_gart_enable(struct amdgpu_device *adev)
 {
-	if (amdgpu_sriov_vf(adev)) {
-		/*
-		 * GCMC_VM_FB_LOCATION_BASE/TOP is NULL for VF, becuase they are
-		 * VF copy registers so vbios post doesn't program them, for
-		 * SRIOV driver need to program them
-		 */
-		WREG32_SOC15(GC, 0, regGCMC_VM_FB_LOCATION_BASE,
-			     adev->gmc.vram_start >> 24);
-		WREG32_SOC15(GC, 0, regGCMC_VM_FB_LOCATION_TOP,
-			     adev->gmc.vram_end >> 24);
-	}
-
 	/* GART Enable. */
 	gfxhub_v3_0_3_init_gart_aperture_regs(adev);
 	gfxhub_v3_0_3_init_system_aperture_regs(adev);
diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c
index 657e53708248..21e46817d82d 100644
--- a/drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c
@@ -608,6 +608,8 @@ static void gmc_v10_0_get_vm_pte(struct amdgpu_device *adev,
 				 struct amdgpu_bo_va_mapping *mapping,
 				 uint64_t *flags)
 {
+	struct amdgpu_bo *bo = mapping->bo_va->base.bo;
+
 	*flags &= ~AMDGPU_PTE_EXECUTABLE;
 	*flags |= mapping->flags & AMDGPU_PTE_EXECUTABLE;
 
@@ -624,6 +626,11 @@ static void gmc_v10_0_get_vm_pte(struct amdgpu_device *adev,
 		*flags |= AMDGPU_PTE_SYSTEM;
 		*flags &= ~AMDGPU_PTE_VALID;
 	}
+
+	if (bo && bo->flags & (AMDGPU_GEM_CREATE_COHERENT |
+			       AMDGPU_GEM_CREATE_UNCACHED))
+		*flags = (*flags & ~AMDGPU_PTE_MTYPE_NV10_MASK) |
+			 AMDGPU_PTE_MTYPE_NV10(MTYPE_UC);
 }
 
 static unsigned gmc_v10_0_get_vbios_fb_size(struct amdgpu_device *adev)
diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v11_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v11_0.c
index 66dfb574cc7d..96e52ec0fb69 100644
--- a/drivers/gpu/drm/amd/amdgpu/gmc_v11_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gmc_v11_0.c
@@ -503,6 +503,8 @@ static void gmc_v11_0_get_vm_pte(struct amdgpu_device *adev,
 				 struct amdgpu_bo_va_mapping *mapping,
 				 uint64_t *flags)
 {
+	struct amdgpu_bo *bo = mapping->bo_va->base.bo;
+
 	*flags &= ~AMDGPU_PTE_EXECUTABLE;
 	*flags |= mapping->flags & AMDGPU_PTE_EXECUTABLE;
 
@@ -519,6 +521,11 @@ static void gmc_v11_0_get_vm_pte(struct amdgpu_device *adev,
 		*flags |= AMDGPU_PTE_SYSTEM;
 		*flags &= ~AMDGPU_PTE_VALID;
 	}
+
+	if (bo && bo->flags & (AMDGPU_GEM_CREATE_COHERENT |
+			       AMDGPU_GEM_CREATE_UNCACHED))
+		*flags = (*flags & ~AMDGPU_PTE_MTYPE_NV10_MASK) |
+			 AMDGPU_PTE_MTYPE_NV10(MTYPE_UC);
 }
 
 static unsigned gmc_v11_0_get_vbios_fb_size(struct amdgpu_device *adev)
@@ -551,7 +558,10 @@ static void gmc_v11_0_set_umc_funcs(struct amdgpu_device *adev)
 		adev->umc.node_inst_num = adev->gmc.num_umc;
 		adev->umc.max_ras_err_cnt_per_query = UMC_V8_10_TOTAL_CHANNEL_NUM(adev);
 		adev->umc.channel_offs = UMC_V8_10_PER_CHANNEL_OFFSET;
-		adev->umc.channel_idx_tbl = &umc_v8_10_channel_idx_tbl[0][0][0];
+		if (adev->umc.node_inst_num == 4)
+			adev->umc.channel_idx_tbl = &umc_v8_10_channel_idx_tbl_ext0[0][0][0];
+		else
+			adev->umc.channel_idx_tbl = &umc_v8_10_channel_idx_tbl[0][0][0];
 		adev->umc.ras = &umc_v8_10_ras;
 		break;
 	case IP_VERSION(8, 11, 0):
diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c
index 67ca16a8027c..50386eb2eec8 100644
--- a/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c
@@ -1113,6 +1113,74 @@ static void gmc_v9_0_get_vm_pde(struct amdgpu_device *adev, int level,
 	}
 }
 
+static void gmc_v9_0_get_coherence_flags(struct amdgpu_device *adev,
+					 struct amdgpu_bo *bo,
+					 struct amdgpu_bo_va_mapping *mapping,
+					 uint64_t *flags)
+{
+	struct amdgpu_device *bo_adev = amdgpu_ttm_adev(bo->tbo.bdev);
+	bool is_vram = bo->tbo.resource->mem_type == TTM_PL_VRAM;
+	bool coherent = bo->flags & AMDGPU_GEM_CREATE_COHERENT;
+	bool uncached = bo->flags & AMDGPU_GEM_CREATE_UNCACHED;
+	unsigned int mtype;
+	bool snoop = false;
+
+	switch (adev->ip_versions[GC_HWIP][0]) {
+	case IP_VERSION(9, 4, 1):
+	case IP_VERSION(9, 4, 2):
+		if (is_vram) {
+			if (bo_adev == adev) {
+				if (uncached)
+					mtype = MTYPE_UC;
+				else if (coherent)
+					mtype = MTYPE_CC;
+				else
+					mtype = MTYPE_RW;
+				/* FIXME: is this still needed? Or does
+				 * amdgpu_ttm_tt_pde_flags already handle this?
+				 */
+				if (adev->ip_versions[GC_HWIP][0] ==
+					IP_VERSION(9, 4, 2) &&
+				    adev->gmc.xgmi.connected_to_cpu)
+					snoop = true;
+			} else {
+				if (uncached || coherent)
+					mtype = MTYPE_UC;
+				else
+					mtype = MTYPE_NC;
+				if (mapping->bo_va->is_xgmi)
+					snoop = true;
+			}
+		} else {
+			if (uncached || coherent)
+				mtype = MTYPE_UC;
+			else
+				mtype = MTYPE_NC;
+			/* FIXME: is this still needed? Or does
+			 * amdgpu_ttm_tt_pde_flags already handle this?
+			 */
+			snoop = true;
+		}
+		break;
+	default:
+		if (uncached || coherent)
+			mtype = MTYPE_UC;
+		else
+			mtype = MTYPE_NC;
+
+		/* FIXME: is this still needed? Or does
+		 * amdgpu_ttm_tt_pde_flags already handle this?
+		 */
+		if (!is_vram)
+			snoop = true;
+	}
+
+	if (mtype != MTYPE_NC)
+		*flags = (*flags & ~AMDGPU_PTE_MTYPE_VG10_MASK) |
+			 AMDGPU_PTE_MTYPE_VG10(mtype);
+	*flags |= snoop ? AMDGPU_PTE_SNOOPED : 0;
+}
+
 static void gmc_v9_0_get_vm_pte(struct amdgpu_device *adev,
 				struct amdgpu_bo_va_mapping *mapping,
 				uint64_t *flags)
@@ -1128,14 +1196,9 @@ static void gmc_v9_0_get_vm_pte(struct amdgpu_device *adev,
 		*flags &= ~AMDGPU_PTE_VALID;
 	}
 
-	if ((adev->ip_versions[GC_HWIP][0] == IP_VERSION(9, 4, 1) ||
-	     adev->ip_versions[GC_HWIP][0] == IP_VERSION(9, 4, 2)) &&
-	    !(*flags & AMDGPU_PTE_SYSTEM) &&
-	    mapping->bo_va->is_xgmi)
-		*flags |= AMDGPU_PTE_SNOOPED;
-
-	if (adev->ip_versions[GC_HWIP][0] == IP_VERSION(9, 4, 2))
-		*flags |= mapping->flags & AMDGPU_PTE_SNOOPED;
+	if (mapping->bo_va->base.bo)
+		gmc_v9_0_get_coherence_flags(adev, mapping->bo_va->base.bo,
+					     mapping, flags);
 }
 
 static unsigned gmc_v9_0_get_vbios_fb_size(struct amdgpu_device *adev)
diff --git a/drivers/gpu/drm/amd/amdgpu/jpeg_v2_5.c b/drivers/gpu/drm/amd/amdgpu/jpeg_v2_5.c
index f87d0f6ffc93..f2b743a93915 100644
--- a/drivers/gpu/drm/amd/amdgpu/jpeg_v2_5.c
+++ b/drivers/gpu/drm/amd/amdgpu/jpeg_v2_5.c
@@ -807,16 +807,5 @@ static void jpeg_v2_5_set_ras_funcs(struct amdgpu_device *adev)
 		break;
 	}
 
-	if (adev->jpeg.ras) {
-		amdgpu_ras_register_ras_block(adev, &adev->jpeg.ras->ras_block);
-
-		strcpy(adev->jpeg.ras->ras_block.ras_comm.name, "jpeg");
-		adev->jpeg.ras->ras_block.ras_comm.block = AMDGPU_RAS_BLOCK__JPEG;
-		adev->jpeg.ras->ras_block.ras_comm.type = AMDGPU_RAS_ERROR__POISON;
-		adev->jpeg.ras_if = &adev->jpeg.ras->ras_block.ras_comm;
-
-		/* If don't define special ras_late_init function, use default ras_late_init */
-		if (!adev->jpeg.ras->ras_block.ras_late_init)
-			adev->jpeg.ras->ras_block.ras_late_init = amdgpu_ras_block_late_init;
-	}
+	jpeg_set_ras_funcs(adev);
 }
diff --git a/drivers/gpu/drm/amd/amdgpu/jpeg_v4_0.c b/drivers/gpu/drm/amd/amdgpu/jpeg_v4_0.c
index 63b0d0b810ec..3beb731b2ce5 100644
--- a/drivers/gpu/drm/amd/amdgpu/jpeg_v4_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/jpeg_v4_0.c
@@ -27,6 +27,7 @@
 #include "soc15.h"
 #include "soc15d.h"
 #include "jpeg_v2_0.h"
+#include "jpeg_v4_0.h"
 
 #include "vcn/vcn_4_0_0_offset.h"
 #include "vcn/vcn_4_0_0_sh_mask.h"
@@ -38,6 +39,7 @@ static void jpeg_v4_0_set_dec_ring_funcs(struct amdgpu_device *adev);
 static void jpeg_v4_0_set_irq_funcs(struct amdgpu_device *adev);
 static int jpeg_v4_0_set_powergating_state(void *handle,
 				enum amd_powergating_state state);
+static void jpeg_v4_0_set_ras_funcs(struct amdgpu_device *adev);
 
 /**
  * jpeg_v4_0_early_init - set function pointers
@@ -55,6 +57,7 @@ static int jpeg_v4_0_early_init(void *handle)
 
 	jpeg_v4_0_set_dec_ring_funcs(adev);
 	jpeg_v4_0_set_irq_funcs(adev);
+	jpeg_v4_0_set_ras_funcs(adev);
 
 	return 0;
 }
@@ -78,6 +81,18 @@ static int jpeg_v4_0_sw_init(void *handle)
 	if (r)
 		return r;
 
+	/* JPEG DJPEG POISON EVENT */
+	r = amdgpu_irq_add_id(adev, SOC15_IH_CLIENTID_VCN,
+			VCN_4_0__SRCID_DJPEG0_POISON, &adev->jpeg.inst->irq);
+	if (r)
+		return r;
+
+	/* JPEG EJPEG POISON EVENT */
+	r = amdgpu_irq_add_id(adev, SOC15_IH_CLIENTID_VCN,
+			VCN_4_0__SRCID_EJPEG0_POISON, &adev->jpeg.inst->irq);
+	if (r)
+		return r;
+
 	r = amdgpu_jpeg_sw_init(adev);
 	if (r)
 		return r;
@@ -167,6 +182,8 @@ static int jpeg_v4_0_hw_fini(void *handle)
 	      RREG32_SOC15(JPEG, 0, regUVD_JRBC_STATUS))
 		jpeg_v4_0_set_powergating_state(adev, AMD_PG_STATE_GATE);
 
+	amdgpu_irq_put(adev, &adev->jpeg.inst->irq, 0);
+
 	return 0;
 }
 
@@ -524,6 +541,10 @@ static int jpeg_v4_0_process_interrupt(struct amdgpu_device *adev,
 	case VCN_4_0__SRCID__JPEG_DECODE:
 		amdgpu_fence_process(&adev->jpeg.inst->ring_dec);
 		break;
+	case VCN_4_0__SRCID_DJPEG0_POISON:
+	case VCN_4_0__SRCID_EJPEG0_POISON:
+		amdgpu_jpeg_process_poison_irq(adev, source, entry);
+		break;
 	default:
 		DRM_DEV_ERROR(adev->dev, "Unhandled interrupt: %d %d\n",
 			  entry->src_id, entry->src_data[0]);
@@ -607,3 +628,63 @@ const struct amdgpu_ip_block_version jpeg_v4_0_ip_block = {
 	.rev = 0,
 	.funcs = &jpeg_v4_0_ip_funcs,
 };
+
+static uint32_t jpeg_v4_0_query_poison_by_instance(struct amdgpu_device *adev,
+		uint32_t instance, uint32_t sub_block)
+{
+	uint32_t poison_stat = 0, reg_value = 0;
+
+	switch (sub_block) {
+	case AMDGPU_JPEG_V4_0_JPEG0:
+		reg_value = RREG32_SOC15(JPEG, instance, regUVD_RAS_JPEG0_STATUS);
+		poison_stat = REG_GET_FIELD(reg_value, UVD_RAS_JPEG0_STATUS, POISONED_PF);
+		break;
+	case AMDGPU_JPEG_V4_0_JPEG1:
+		reg_value = RREG32_SOC15(JPEG, instance, regUVD_RAS_JPEG1_STATUS);
+		poison_stat = REG_GET_FIELD(reg_value, UVD_RAS_JPEG1_STATUS, POISONED_PF);
+		break;
+	default:
+		break;
+	}
+
+	if (poison_stat)
+		dev_info(adev->dev, "Poison detected in JPEG%d sub_block%d\n",
+			instance, sub_block);
+
+	return poison_stat;
+}
+
+static bool jpeg_v4_0_query_ras_poison_status(struct amdgpu_device *adev)
+{
+	uint32_t inst = 0, sub = 0, poison_stat = 0;
+
+	for (inst = 0; inst < adev->jpeg.num_jpeg_inst; inst++)
+		for (sub = 0; sub < AMDGPU_JPEG_V4_0_MAX_SUB_BLOCK; sub++)
+			poison_stat +=
+				jpeg_v4_0_query_poison_by_instance(adev, inst, sub);
+
+	return !!poison_stat;
+}
+
+const struct amdgpu_ras_block_hw_ops jpeg_v4_0_ras_hw_ops = {
+	.query_poison_status = jpeg_v4_0_query_ras_poison_status,
+};
+
+static struct amdgpu_jpeg_ras jpeg_v4_0_ras = {
+	.ras_block = {
+		.hw_ops = &jpeg_v4_0_ras_hw_ops,
+	},
+};
+
+static void jpeg_v4_0_set_ras_funcs(struct amdgpu_device *adev)
+{
+	switch (adev->ip_versions[JPEG_HWIP][0]) {
+	case IP_VERSION(4, 0, 0):
+		adev->jpeg.ras = &jpeg_v4_0_ras;
+		break;
+	default:
+		break;
+	}
+
+	jpeg_set_ras_funcs(adev);
+}
diff --git a/drivers/gpu/drm/amd/amdgpu/jpeg_v4_0.h b/drivers/gpu/drm/amd/amdgpu/jpeg_v4_0.h
index f1ed6ccfedca..07d36c2abd6b 100644
--- a/drivers/gpu/drm/amd/amdgpu/jpeg_v4_0.h
+++ b/drivers/gpu/drm/amd/amdgpu/jpeg_v4_0.h
@@ -24,6 +24,13 @@
 #ifndef __JPEG_V4_0_H__
 #define __JPEG_V4_0_H__
 
+enum amdgpu_jpeg_v4_0_sub_block {
+	AMDGPU_JPEG_V4_0_JPEG0 = 0,
+	AMDGPU_JPEG_V4_0_JPEG1,
+
+	AMDGPU_JPEG_V4_0_MAX_SUB_BLOCK,
+};
+
 extern const struct amdgpu_ip_block_version jpeg_v4_0_ip_block;
 
 #endif /* __JPEG_V4_0_H__ */
diff --git a/drivers/gpu/drm/amd/amdgpu/mes_v11_0.c b/drivers/gpu/drm/amd/amdgpu/mes_v11_0.c
index 1395453a0662..8d9c1e841353 100644
--- a/drivers/gpu/drm/amd/amdgpu/mes_v11_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/mes_v11_0.c
@@ -1253,7 +1253,9 @@ static int mes_v11_0_kiq_hw_fini(struct amdgpu_device *adev)
 	if (adev->mes.ring.sched.ready)
 		mes_v11_0_kiq_dequeue_sched(adev);
 
-	mes_v11_0_enable(adev, false);
+	if (!amdgpu_sriov_vf(adev))
+		mes_v11_0_enable(adev, false);
+
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/amd/amdgpu/psp_v10_0.c b/drivers/gpu/drm/amd/amdgpu/psp_v10_0.c
index ed2293686f0d..9de46fa8f46c 100644
--- a/drivers/gpu/drm/amd/amdgpu/psp_v10_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/psp_v10_0.c
@@ -126,32 +126,6 @@ out:
 	return err;
 }
 
-static int psp_v10_0_ring_init(struct psp_context *psp,
-			       enum psp_ring_type ring_type)
-{
-	int ret = 0;
-	struct psp_ring *ring;
-	struct amdgpu_device *adev = psp->adev;
-
-	ring = &psp->km_ring;
-
-	ring->ring_type = ring_type;
-
-	/* allocate 4k Page of Local Frame Buffer memory for ring */
-	ring->ring_size = 0x1000;
-	ret = amdgpu_bo_create_kernel(adev, ring->ring_size, PAGE_SIZE,
-				      AMDGPU_GEM_DOMAIN_VRAM,
-				      &adev->firmware.rbuf,
-				      &ring->ring_mem_mc_addr,
-				      (void **)&ring->ring_mem);
-	if (ret) {
-		ring->ring_size = 0;
-		return ret;
-	}
-
-	return 0;
-}
-
 static int psp_v10_0_ring_create(struct psp_context *psp,
 				 enum psp_ring_type ring_type)
 {
@@ -245,7 +219,6 @@ static void psp_v10_0_ring_set_wptr(struct psp_context *psp, uint32_t value)
 
 static const struct psp_funcs psp_v10_0_funcs = {
 	.init_microcode = psp_v10_0_init_microcode,
-	.ring_init = psp_v10_0_ring_init,
 	.ring_create = psp_v10_0_ring_create,
 	.ring_stop = psp_v10_0_ring_stop,
 	.ring_destroy = psp_v10_0_ring_destroy,
diff --git a/drivers/gpu/drm/amd/amdgpu/psp_v11_0.c b/drivers/gpu/drm/amd/amdgpu/psp_v11_0.c
index 9518b4394a6e..bd3e3e23a939 100644
--- a/drivers/gpu/drm/amd/amdgpu/psp_v11_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/psp_v11_0.c
@@ -360,32 +360,6 @@ static int psp_v11_0_bootloader_load_sos(struct psp_context *psp)
 	return ret;
 }
 
-static int psp_v11_0_ring_init(struct psp_context *psp,
-			      enum psp_ring_type ring_type)
-{
-	int ret = 0;
-	struct psp_ring *ring;
-	struct amdgpu_device *adev = psp->adev;
-
-	ring = &psp->km_ring;
-
-	ring->ring_type = ring_type;
-
-	/* allocate 4k Page of Local Frame Buffer memory for ring */
-	ring->ring_size = 0x1000;
-	ret = amdgpu_bo_create_kernel(adev, ring->ring_size, PAGE_SIZE,
-				      AMDGPU_GEM_DOMAIN_VRAM,
-				      &adev->firmware.rbuf,
-				      &ring->ring_mem_mc_addr,
-				      (void **)&ring->ring_mem);
-	if (ret) {
-		ring->ring_size = 0;
-		return ret;
-	}
-
-	return 0;
-}
-
 static int psp_v11_0_ring_stop(struct psp_context *psp,
 			      enum psp_ring_type ring_type)
 {
@@ -779,7 +753,6 @@ static const struct psp_funcs psp_v11_0_funcs = {
 	.bootloader_load_spl = psp_v11_0_bootloader_load_spl,
 	.bootloader_load_sysdrv = psp_v11_0_bootloader_load_sysdrv,
 	.bootloader_load_sos = psp_v11_0_bootloader_load_sos,
-	.ring_init = psp_v11_0_ring_init,
 	.ring_create = psp_v11_0_ring_create,
 	.ring_stop = psp_v11_0_ring_stop,
 	.ring_destroy = psp_v11_0_ring_destroy,
diff --git a/drivers/gpu/drm/amd/amdgpu/psp_v11_0_8.c b/drivers/gpu/drm/amd/amdgpu/psp_v11_0_8.c
index ff13e1beb49b..5697760a819b 100644
--- a/drivers/gpu/drm/amd/amdgpu/psp_v11_0_8.c
+++ b/drivers/gpu/drm/amd/amdgpu/psp_v11_0_8.c
@@ -28,32 +28,6 @@
 
 #include "mp/mp_11_0_8_offset.h"
 
-static int psp_v11_0_8_ring_init(struct psp_context *psp,
-			      enum psp_ring_type ring_type)
-{
-	int ret = 0;
-	struct psp_ring *ring;
-	struct amdgpu_device *adev = psp->adev;
-
-	ring = &psp->km_ring;
-
-	ring->ring_type = ring_type;
-
-	/* allocate 4k Page of Local Frame Buffer memory for ring */
-	ring->ring_size = 0x1000;
-	ret = amdgpu_bo_create_kernel(adev, ring->ring_size, PAGE_SIZE,
-				      AMDGPU_GEM_DOMAIN_VRAM,
-				      &adev->firmware.rbuf,
-				      &ring->ring_mem_mc_addr,
-				      (void **)&ring->ring_mem);
-	if (ret) {
-		ring->ring_size = 0;
-		return ret;
-	}
-
-	return 0;
-}
-
 static int psp_v11_0_8_ring_stop(struct psp_context *psp,
 			       enum psp_ring_type ring_type)
 {
@@ -194,7 +168,6 @@ static void psp_v11_0_8_ring_set_wptr(struct psp_context *psp, uint32_t value)
 }
 
 static const struct psp_funcs psp_v11_0_8_funcs = {
-	.ring_init = psp_v11_0_8_ring_init,
 	.ring_create = psp_v11_0_8_ring_create,
 	.ring_stop = psp_v11_0_8_ring_stop,
 	.ring_destroy = psp_v11_0_8_ring_destroy,
diff --git a/drivers/gpu/drm/amd/amdgpu/psp_v12_0.c b/drivers/gpu/drm/amd/amdgpu/psp_v12_0.c
index 0b2ac418e4ac..8ed2281b6557 100644
--- a/drivers/gpu/drm/amd/amdgpu/psp_v12_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/psp_v12_0.c
@@ -236,34 +236,6 @@ static void psp_v12_0_reroute_ih(struct psp_context *psp)
 		     0x80000000, 0x8000FFFF, false);
 }
 
-static int psp_v12_0_ring_init(struct psp_context *psp,
-			      enum psp_ring_type ring_type)
-{
-	int ret = 0;
-	struct psp_ring *ring;
-	struct amdgpu_device *adev = psp->adev;
-
-	psp_v12_0_reroute_ih(psp);
-
-	ring = &psp->km_ring;
-
-	ring->ring_type = ring_type;
-
-	/* allocate 4k Page of Local Frame Buffer memory for ring */
-	ring->ring_size = 0x1000;
-	ret = amdgpu_bo_create_kernel(adev, ring->ring_size, PAGE_SIZE,
-				      AMDGPU_GEM_DOMAIN_VRAM,
-				      &adev->firmware.rbuf,
-				      &ring->ring_mem_mc_addr,
-				      (void **)&ring->ring_mem);
-	if (ret) {
-		ring->ring_size = 0;
-		return ret;
-	}
-
-	return 0;
-}
-
 static int psp_v12_0_ring_create(struct psp_context *psp,
 				enum psp_ring_type ring_type)
 {
@@ -272,6 +244,8 @@ static int psp_v12_0_ring_create(struct psp_context *psp,
 	struct psp_ring *ring = &psp->km_ring;
 	struct amdgpu_device *adev = psp->adev;
 
+	psp_v12_0_reroute_ih(psp);
+
 	if (amdgpu_sriov_vf(psp->adev)) {
 		/* Write low address of the ring to C2PMSG_102 */
 		psp_ring_reg = lower_32_bits(ring->ring_mem_mc_addr);
@@ -425,7 +399,6 @@ static const struct psp_funcs psp_v12_0_funcs = {
 	.init_microcode = psp_v12_0_init_microcode,
 	.bootloader_load_sysdrv = psp_v12_0_bootloader_load_sysdrv,
 	.bootloader_load_sos = psp_v12_0_bootloader_load_sos,
-	.ring_init = psp_v12_0_ring_init,
 	.ring_create = psp_v12_0_ring_create,
 	.ring_stop = psp_v12_0_ring_stop,
 	.ring_destroy = psp_v12_0_ring_destroy,
diff --git a/drivers/gpu/drm/amd/amdgpu/psp_v13_0.c b/drivers/gpu/drm/amd/amdgpu/psp_v13_0.c
index 21d822b1d589..ee27bfaba6fd 100644
--- a/drivers/gpu/drm/amd/amdgpu/psp_v13_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/psp_v13_0.c
@@ -45,6 +45,7 @@ MODULE_FIRMWARE("amdgpu/psp_13_0_0_ta.bin");
 MODULE_FIRMWARE("amdgpu/psp_13_0_7_sos.bin");
 MODULE_FIRMWARE("amdgpu/psp_13_0_7_ta.bin");
 MODULE_FIRMWARE("amdgpu/psp_13_0_10_sos.bin");
+MODULE_FIRMWARE("amdgpu/psp_13_0_10_ta.bin");
 
 /* For large FW files the time to complete can be very long */
 #define USBC_PD_POLLING_LIMIT_S 240
@@ -267,32 +268,6 @@ static int psp_v13_0_bootloader_load_sos(struct psp_context *psp)
 	return ret;
 }
 
-static int psp_v13_0_ring_init(struct psp_context *psp,
-			      enum psp_ring_type ring_type)
-{
-	int ret = 0;
-	struct psp_ring *ring;
-	struct amdgpu_device *adev = psp->adev;
-
-	ring = &psp->km_ring;
-
-	ring->ring_type = ring_type;
-
-	/* allocate 4k Page of Local Frame Buffer memory for ring */
-	ring->ring_size = 0x1000;
-	ret = amdgpu_bo_create_kernel(adev, ring->ring_size, PAGE_SIZE,
-				      AMDGPU_GEM_DOMAIN_VRAM,
-				      &adev->firmware.rbuf,
-				      &ring->ring_mem_mc_addr,
-				      (void **)&ring->ring_mem);
-	if (ret) {
-		ring->ring_size = 0;
-		return ret;
-	}
-
-	return 0;
-}
-
 static int psp_v13_0_ring_stop(struct psp_context *psp,
 			       enum psp_ring_type ring_type)
 {
@@ -728,7 +703,6 @@ static const struct psp_funcs psp_v13_0_funcs = {
 	.bootloader_load_dbg_drv = psp_v13_0_bootloader_load_dbg_drv,
 	.bootloader_load_ras_drv = psp_v13_0_bootloader_load_ras_drv,
 	.bootloader_load_sos = psp_v13_0_bootloader_load_sos,
-	.ring_init = psp_v13_0_ring_init,
 	.ring_create = psp_v13_0_ring_create,
 	.ring_stop = psp_v13_0_ring_stop,
 	.ring_destroy = psp_v13_0_ring_destroy,
diff --git a/drivers/gpu/drm/amd/amdgpu/psp_v13_0_4.c b/drivers/gpu/drm/amd/amdgpu/psp_v13_0_4.c
index 321089dfa7db..9d4e24e518e8 100644
--- a/drivers/gpu/drm/amd/amdgpu/psp_v13_0_4.c
+++ b/drivers/gpu/drm/amd/amdgpu/psp_v13_0_4.c
@@ -199,32 +199,6 @@ static int psp_v13_0_4_bootloader_load_sos(struct psp_context *psp)
 	return ret;
 }
 
-static int psp_v13_0_4_ring_init(struct psp_context *psp,
-			      enum psp_ring_type ring_type)
-{
-	int ret = 0;
-	struct psp_ring *ring;
-	struct amdgpu_device *adev = psp->adev;
-
-	ring = &psp->km_ring;
-
-	ring->ring_type = ring_type;
-
-	/* allocate 4k Page of Local Frame Buffer memory for ring */
-	ring->ring_size = 0x1000;
-	ret = amdgpu_bo_create_kernel(adev, ring->ring_size, PAGE_SIZE,
-				      AMDGPU_GEM_DOMAIN_VRAM,
-				      &adev->firmware.rbuf,
-				      &ring->ring_mem_mc_addr,
-				      (void **)&ring->ring_mem);
-	if (ret) {
-		ring->ring_size = 0;
-		return ret;
-	}
-
-	return 0;
-}
-
 static int psp_v13_0_4_ring_stop(struct psp_context *psp,
 			       enum psp_ring_type ring_type)
 {
@@ -373,7 +347,6 @@ static const struct psp_funcs psp_v13_0_4_funcs = {
 	.bootloader_load_intf_drv = psp_v13_0_4_bootloader_load_intf_drv,
 	.bootloader_load_dbg_drv = psp_v13_0_4_bootloader_load_dbg_drv,
 	.bootloader_load_sos = psp_v13_0_4_bootloader_load_sos,
-	.ring_init = psp_v13_0_4_ring_init,
 	.ring_create = psp_v13_0_4_ring_create,
 	.ring_stop = psp_v13_0_4_ring_stop,
 	.ring_destroy = psp_v13_0_4_ring_destroy,
diff --git a/drivers/gpu/drm/amd/amdgpu/psp_v3_1.c b/drivers/gpu/drm/amd/amdgpu/psp_v3_1.c
index 01f3bcc62a6c..157147c6c94e 100644
--- a/drivers/gpu/drm/amd/amdgpu/psp_v3_1.c
+++ b/drivers/gpu/drm/amd/amdgpu/psp_v3_1.c
@@ -160,32 +160,6 @@ static int psp_v3_1_bootloader_load_sos(struct psp_context *psp)
 	return ret;
 }
 
-static int psp_v3_1_ring_init(struct psp_context *psp,
-			      enum psp_ring_type ring_type)
-{
-	int ret = 0;
-	struct psp_ring *ring;
-	struct amdgpu_device *adev = psp->adev;
-
-	ring = &psp->km_ring;
-
-	ring->ring_type = ring_type;
-
-	/* allocate 4k Page of Local Frame Buffer memory for ring */
-	ring->ring_size = 0x1000;
-	ret = amdgpu_bo_create_kernel(adev, ring->ring_size, PAGE_SIZE,
-				      AMDGPU_GEM_DOMAIN_VRAM,
-				      &adev->firmware.rbuf,
-				      &ring->ring_mem_mc_addr,
-				      (void **)&ring->ring_mem);
-	if (ret) {
-		ring->ring_size = 0;
-		return ret;
-	}
-
-	return 0;
-}
-
 static void psp_v3_1_reroute_ih(struct psp_context *psp)
 {
 	struct amdgpu_device *adev = psp->adev;
@@ -401,7 +375,6 @@ static const struct psp_funcs psp_v3_1_funcs = {
 	.init_microcode = psp_v3_1_init_microcode,
 	.bootloader_load_sysdrv = psp_v3_1_bootloader_load_sysdrv,
 	.bootloader_load_sos = psp_v3_1_bootloader_load_sos,
-	.ring_init = psp_v3_1_ring_init,
 	.ring_create = psp_v3_1_ring_create,
 	.ring_stop = psp_v3_1_ring_stop,
 	.ring_destroy = psp_v3_1_ring_destroy,
diff --git a/drivers/gpu/drm/amd/amdgpu/soc21.c b/drivers/gpu/drm/amd/amdgpu/soc21.c
index 1d4013ed0d10..b258e9aa0558 100644
--- a/drivers/gpu/drm/amd/amdgpu/soc21.c
+++ b/drivers/gpu/drm/amd/amdgpu/soc21.c
@@ -322,6 +322,7 @@ soc21_asic_reset_method(struct amdgpu_device *adev)
 	switch (adev->ip_versions[MP1_HWIP][0]) {
 	case IP_VERSION(13, 0, 0):
 	case IP_VERSION(13, 0, 7):
+	case IP_VERSION(13, 0, 10):
 		return AMD_RESET_METHOD_MODE1;
 	case IP_VERSION(13, 0, 4):
 		return AMD_RESET_METHOD_MODE2;
diff --git a/drivers/gpu/drm/amd/amdgpu/umc_v8_10.c b/drivers/gpu/drm/amd/amdgpu/umc_v8_10.c
index 91235df54e22..b7da4528cf0a 100644
--- a/drivers/gpu/drm/amd/amdgpu/umc_v8_10.c
+++ b/drivers/gpu/drm/amd/amdgpu/umc_v8_10.c
@@ -46,6 +46,16 @@ const struct channelnum_map_colbit umc_v8_10_channelnum_map_colbit_table[] = {
 };
 
 const uint32_t
+	umc_v8_10_channel_idx_tbl_ext0[]
+				[UMC_V8_10_UMC_INSTANCE_NUM]
+				[UMC_V8_10_CHANNEL_INSTANCE_NUM] = {
+	   {{1,   5}, {7,  3}},
+	   {{14, 15}, {13, 12}},
+	   {{10, 11}, {9,  8}},
+	   {{6,   2}, {0,  4}}
+	};
+
+const uint32_t
 	umc_v8_10_channel_idx_tbl[]
 				[UMC_V8_10_UMC_INSTANCE_NUM]
 				[UMC_V8_10_CHANNEL_INSTANCE_NUM] = {
diff --git a/drivers/gpu/drm/amd/amdgpu/umc_v8_10.h b/drivers/gpu/drm/amd/amdgpu/umc_v8_10.h
index 849ede88e111..25eaf4af5fcf 100644
--- a/drivers/gpu/drm/amd/amdgpu/umc_v8_10.h
+++ b/drivers/gpu/drm/amd/amdgpu/umc_v8_10.h
@@ -66,5 +66,9 @@ extern const uint32_t
 				[UMC_V8_10_UMC_INSTANCE_NUM]
 				[UMC_V8_10_CHANNEL_INSTANCE_NUM];
 
+extern const uint32_t
+	umc_v8_10_channel_idx_tbl_ext0[]
+				[UMC_V8_10_UMC_INSTANCE_NUM]
+				[UMC_V8_10_CHANNEL_INSTANCE_NUM];
 #endif
 
diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c b/drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c
index 8a7006d62a87..ce8374ee824d 100644
--- a/drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c
+++ b/drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c
@@ -2002,16 +2002,5 @@ static void vcn_v2_5_set_ras_funcs(struct amdgpu_device *adev)
 		break;
 	}
 
-	if (adev->vcn.ras) {
-		amdgpu_ras_register_ras_block(adev, &adev->vcn.ras->ras_block);
-
-		strcpy(adev->vcn.ras->ras_block.ras_comm.name, "vcn");
-		adev->vcn.ras->ras_block.ras_comm.block = AMDGPU_RAS_BLOCK__VCN;
-		adev->vcn.ras->ras_block.ras_comm.type = AMDGPU_RAS_ERROR__POISON;
-		adev->vcn.ras_if = &adev->vcn.ras->ras_block.ras_comm;
-
-		/* If don't define special ras_late_init function, use default ras_late_init */
-		if (!adev->vcn.ras->ras_block.ras_late_init)
-			adev->vcn.ras->ras_block.ras_late_init = amdgpu_ras_block_late_init;
-	}
+	amdgpu_vcn_set_ras_funcs(adev);
 }
diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v4_0.c b/drivers/gpu/drm/amd/amdgpu/vcn_v4_0.c
index 897a5ce9c9da..403d054cf51b 100644
--- a/drivers/gpu/drm/amd/amdgpu/vcn_v4_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/vcn_v4_0.c
@@ -31,6 +31,7 @@
 #include "soc15_hw_ip.h"
 #include "vcn_v2_0.h"
 #include "mmsch_v4_0.h"
+#include "vcn_v4_0.h"
 
 #include "vcn/vcn_4_0_0_offset.h"
 #include "vcn/vcn_4_0_0_sh_mask.h"
@@ -64,6 +65,7 @@ static int vcn_v4_0_set_powergating_state(void *handle,
 static int vcn_v4_0_pause_dpg_mode(struct amdgpu_device *adev,
         int inst_idx, struct dpg_pause_state *new_state);
 static void vcn_v4_0_unified_ring_set_wptr(struct amdgpu_ring *ring);
+static void vcn_v4_0_set_ras_funcs(struct amdgpu_device *adev);
 
 /**
  * vcn_v4_0_early_init - set function pointers
@@ -84,6 +86,7 @@ static int vcn_v4_0_early_init(void *handle)
 
 	vcn_v4_0_set_unified_ring_funcs(adev);
 	vcn_v4_0_set_irq_funcs(adev);
+	vcn_v4_0_set_ras_funcs(adev);
 
 	return 0;
 }
@@ -132,6 +135,12 @@ static int vcn_v4_0_sw_init(void *handle)
 		if (r)
 			return r;
 
+		/* VCN POISON TRAP */
+		r = amdgpu_irq_add_id(adev, amdgpu_ih_clientid_vcns[i],
+				VCN_4_0__SRCID_UVD_POISON, &adev->vcn.inst[i].irq);
+		if (r)
+			return r;
+
 		ring = &adev->vcn.inst[i].ring_enc[0];
 		ring->use_doorbell = true;
 		if (amdgpu_sriov_vf(adev))
@@ -296,6 +305,7 @@ static int vcn_v4_0_hw_fini(void *handle)
 			}
 		}
 
+		amdgpu_irq_put(adev, &adev->vcn.inst[i].irq, 0);
 	}
 
 	return 0;
@@ -1939,6 +1949,9 @@ static int vcn_v4_0_process_interrupt(struct amdgpu_device *adev, struct amdgpu_
 	case VCN_4_0__SRCID__UVD_ENC_GENERAL_PURPOSE:
 		amdgpu_fence_process(&adev->vcn.inst[ip_instance].ring_enc[0]);
 		break;
+	case VCN_4_0__SRCID_UVD_POISON:
+		amdgpu_vcn_process_poison_irq(adev, source, entry);
+		break;
 	default:
 		DRM_ERROR("Unhandled interrupt: %d %d\n",
 			  entry->src_id, entry->src_data[0]);
@@ -2001,3 +2014,60 @@ const struct amdgpu_ip_block_version vcn_v4_0_ip_block =
 	.rev = 0,
 	.funcs = &vcn_v4_0_ip_funcs,
 };
+
+static uint32_t vcn_v4_0_query_poison_by_instance(struct amdgpu_device *adev,
+			uint32_t instance, uint32_t sub_block)
+{
+	uint32_t poison_stat = 0, reg_value = 0;
+
+	switch (sub_block) {
+	case AMDGPU_VCN_V4_0_VCPU_VCODEC:
+		reg_value = RREG32_SOC15(VCN, instance, regUVD_RAS_VCPU_VCODEC_STATUS);
+		poison_stat = REG_GET_FIELD(reg_value, UVD_RAS_VCPU_VCODEC_STATUS, POISONED_PF);
+		break;
+	default:
+		break;
+	}
+
+	if (poison_stat)
+		dev_info(adev->dev, "Poison detected in VCN%d, sub_block%d\n",
+			instance, sub_block);
+
+	return poison_stat;
+}
+
+static bool vcn_v4_0_query_ras_poison_status(struct amdgpu_device *adev)
+{
+	uint32_t inst, sub;
+	uint32_t poison_stat = 0;
+
+	for (inst = 0; inst < adev->vcn.num_vcn_inst; inst++)
+		for (sub = 0; sub < AMDGPU_VCN_V4_0_MAX_SUB_BLOCK; sub++)
+			poison_stat +=
+				vcn_v4_0_query_poison_by_instance(adev, inst, sub);
+
+	return !!poison_stat;
+}
+
+const struct amdgpu_ras_block_hw_ops vcn_v4_0_ras_hw_ops = {
+	.query_poison_status = vcn_v4_0_query_ras_poison_status,
+};
+
+static struct amdgpu_vcn_ras vcn_v4_0_ras = {
+	.ras_block = {
+		.hw_ops = &vcn_v4_0_ras_hw_ops,
+	},
+};
+
+static void vcn_v4_0_set_ras_funcs(struct amdgpu_device *adev)
+{
+	switch (adev->ip_versions[VCN_HWIP][0]) {
+	case IP_VERSION(4, 0, 0):
+		adev->vcn.ras = &vcn_v4_0_ras;
+		break;
+	default:
+		break;
+	}
+
+	amdgpu_vcn_set_ras_funcs(adev);
+}
diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v4_0.h b/drivers/gpu/drm/amd/amdgpu/vcn_v4_0.h
index 7c5c9d91bb52..7d3d11f40f27 100644
--- a/drivers/gpu/drm/amd/amdgpu/vcn_v4_0.h
+++ b/drivers/gpu/drm/amd/amdgpu/vcn_v4_0.h
@@ -24,6 +24,12 @@
 #ifndef __VCN_V4_0_H__
 #define __VCN_V4_0_H__
 
+enum amdgpu_vcn_v4_0_sub_block {
+	AMDGPU_VCN_V4_0_VCPU_VCODEC = 0,
+
+	AMDGPU_VCN_V4_0_MAX_SUB_BLOCK,
+};
+
 extern const struct amdgpu_ip_block_version vcn_v4_0_ip_block;
 
 #endif /* __VCN_V4_0_H__ */
diff --git a/drivers/gpu/drm/amd/amdgpu/vi.c b/drivers/gpu/drm/amd/amdgpu/vi.c
index f6ffd7c96ff9..12ef782eb478 100644
--- a/drivers/gpu/drm/amd/amdgpu/vi.c
+++ b/drivers/gpu/drm/amd/amdgpu/vi.c
@@ -2111,6 +2111,8 @@ void vi_set_virt_ops(struct amdgpu_device *adev)
 
 int vi_set_ip_blocks(struct amdgpu_device *adev)
 {
+	amdgpu_device_set_sriov_virtual_display(adev);
+
 	switch (adev->asic_type) {
 	case CHIP_TOPAZ:
 		/* topaz has no DCE, UVD, VCE */
@@ -2130,7 +2132,7 @@ int vi_set_ip_blocks(struct amdgpu_device *adev)
 		amdgpu_device_ip_block_add(adev, &gfx_v8_0_ip_block);
 		amdgpu_device_ip_block_add(adev, &sdma_v3_0_ip_block);
 		amdgpu_device_ip_block_add(adev, &pp_smu_ip_block);
-		if (adev->enable_virtual_display || amdgpu_sriov_vf(adev))
+		if (adev->enable_virtual_display)
 			amdgpu_device_ip_block_add(adev, &amdgpu_vkms_ip_block);
 #if defined(CONFIG_DRM_AMD_DC)
 		else if (amdgpu_device_has_dc_support(adev))
@@ -2150,7 +2152,7 @@ int vi_set_ip_blocks(struct amdgpu_device *adev)
 		amdgpu_device_ip_block_add(adev, &gfx_v8_0_ip_block);
 		amdgpu_device_ip_block_add(adev, &sdma_v3_0_ip_block);
 		amdgpu_device_ip_block_add(adev, &pp_smu_ip_block);
-		if (adev->enable_virtual_display || amdgpu_sriov_vf(adev))
+		if (adev->enable_virtual_display)
 			amdgpu_device_ip_block_add(adev, &amdgpu_vkms_ip_block);
 #if defined(CONFIG_DRM_AMD_DC)
 		else if (amdgpu_device_has_dc_support(adev))
author	Dave Airlie <airlied@redhat.com>	2022-11-22 13:41:11 +1000
committer	Dave Airlie <airlied@redhat.com>	2022-11-22 13:41:11 +1000
commit	fc58764bbf602b65a6f63c53e5fd6feae76c510c (patch)
tree	03f5448cf5d742b8fd2980e86a33636026557ac6 /drivers/gpu/drm/amd/amdgpu
parent	819683a1fc2f7e64017d50caf539e7bafcb37b81 (diff)
parent	aec3bb3a01de09058fbebed4821ed7d07e1ed994 (diff)