Merge branch 'drm-next-4.15' of git://people.freedesktop.org/~agd5f/linux into drm-next

First feature pull for 4.15.  Highlights:
- Per VM BO support
- Lots of powerplay cleanups
- Powerplay support for CI
- pasid mgr for kfd
- interrupt infrastructure for recoverable page faults
- SR-IOV fixes
- initial GPU reset for vega10
- prime mmap support
- ttm page table debugging improvements
- lots of bug fixes

* 'drm-next-4.15' of git://people.freedesktop.org/~agd5f/linux: (232 commits)
  drm/amdgpu: clarify license in amdgpu_trace_points.c
  drm/amdgpu: Add gem_prime_mmap support
  drm/amd/powerplay: delete dead code in smumgr
  drm/amd/powerplay: delete SMUM_FIELD_MASK
  drm/amd/powerplay: delete SMUM_WAIT_INDIRECT_FIELD
  drm/amd/powerplay: delete SMUM_READ_FIELD
  drm/amd/powerplay: delete SMUM_SET_FIELD
  drm/amd/powerplay: delete SMUM_READ_VFPF_INDIRECT_FIELD
  drm/amd/powerplay: delete SMUM_WRITE_VFPF_INDIRECT_FIELD
  drm/amd/powerplay: delete SMUM_WRITE_FIELD
  drm/amd/powerplay: delete SMU_WRITE_INDIRECT_FIELD
  drm/amd/powerplay: move macros to hwmgr.h
  drm/amd/powerplay: move PHM_WAIT_VFPF_INDIRECT_FIELD to hwmgr.h
  drm/amd/powerplay: move SMUM_WAIT_VFPF_INDIRECT_FIELD_UNEQUAL to hwmgr.h
  drm/amd/powerplay: move SMUM_WAIT_INDIRECT_FIELD_UNEQUAL to hwmgr.h
  drm/amd/powerplay: add new helper functions in hwmgr.h
  drm/amd/powerplay: use SMU_IND_INDEX/DATA_11 pair
  drm/amd/powerplay: refine powerplay code.
  drm/amd/powerplay: delete dead code in hwmgr.h
  drm/amd/powerplay: refine interface in struct pp_smumgr_func
  ...

drivers/gpu/drm/Kconfig

@@ -184,6 +184,7 @@ config DRM_AMDGPU
 	select BACKLIGHT_CLASS_DEVICE
 	select BACKLIGHT_LCD_SUPPORT
 	select INTERVAL_TREE
+	select CHASH
 	help
 	  Choose this option if you have a recent AMD Radeon graphics card.
 
@@ -191,6 +192,8 @@ config DRM_AMDGPU
 
 source "drivers/gpu/drm/amd/amdgpu/Kconfig"
 
+source "drivers/gpu/drm/amd/lib/Kconfig"
+
 source "drivers/gpu/drm/nouveau/Kconfig"
 
 source "drivers/gpu/drm/i915/Kconfig"

drivers/gpu/drm/Makefile

@@ -50,6 +50,7 @@ obj-$(CONFIG_DRM_ARM)	+= arm/
 obj-$(CONFIG_DRM_TTM)	+= ttm/
 obj-$(CONFIG_DRM_TDFX)	+= tdfx/
 obj-$(CONFIG_DRM_R128)	+= r128/
+obj-y			+= amd/lib/
 obj-$(CONFIG_HSA_AMD) += amd/amdkfd/
 obj-$(CONFIG_DRM_RADEON)+= radeon/
 obj-$(CONFIG_DRM_AMDGPU)+= amd/amdgpu/

drivers/gpu/drm/amd/amdgpu/Makefile

@@ -133,5 +133,3 @@ include $(FULL_AMD_PATH)/powerplay/Makefile
 amdgpu-y += $(AMD_POWERPLAY_FILES)
 
 obj-$(CONFIG_DRM_AMDGPU)+= amdgpu.o
-
-CFLAGS_amdgpu_trace_points.o := -I$(src)

drivers/gpu/drm/amd/amdgpu/amdgpu.h

@@ -65,6 +65,7 @@
 #include "amdgpu_uvd.h"
 #include "amdgpu_vce.h"
 #include "amdgpu_vcn.h"
+#include "amdgpu_mn.h"
 
 #include "gpu_scheduler.h"
 #include "amdgpu_virt.h"
@@ -91,7 +92,7 @@ extern int amdgpu_dpm;
 extern int amdgpu_fw_load_type;
 extern int amdgpu_aspm;
 extern int amdgpu_runtime_pm;
-extern unsigned amdgpu_ip_block_mask;
+extern uint amdgpu_ip_block_mask;
 extern int amdgpu_bapm;
 extern int amdgpu_deep_color;
 extern int amdgpu_vm_size;
@@ -104,14 +105,14 @@ extern int amdgpu_sched_jobs;
 extern int amdgpu_sched_hw_submission;
 extern int amdgpu_no_evict;
 extern int amdgpu_direct_gma_size;
-extern unsigned amdgpu_pcie_gen_cap;
-extern unsigned amdgpu_pcie_lane_cap;
-extern unsigned amdgpu_cg_mask;
-extern unsigned amdgpu_pg_mask;
-extern unsigned amdgpu_sdma_phase_quantum;
+extern uint amdgpu_pcie_gen_cap;
+extern uint amdgpu_pcie_lane_cap;
+extern uint amdgpu_cg_mask;
+extern uint amdgpu_pg_mask;
+extern uint amdgpu_sdma_phase_quantum;
 extern char *amdgpu_disable_cu;
 extern char *amdgpu_virtual_display;
-extern unsigned amdgpu_pp_feature_mask;
+extern uint amdgpu_pp_feature_mask;
 extern int amdgpu_vram_page_split;
 extern int amdgpu_ngg;
 extern int amdgpu_prim_buf_per_se;
@@ -178,6 +179,7 @@ struct amdgpu_cs_parser;
 struct amdgpu_job;
 struct amdgpu_irq_src;
 struct amdgpu_fpriv;
+struct amdgpu_bo_va_mapping;
 
 enum amdgpu_cp_irq {
 	AMDGPU_CP_IRQ_GFX_EOP = 0,
@@ -292,14 +294,25 @@ struct amdgpu_buffer_funcs {
 
 /* provided by hw blocks that can write ptes, e.g., sdma */
 struct amdgpu_vm_pte_funcs {
+	/* number of dw to reserve per operation */
+	unsigned	copy_pte_num_dw;
+
 	/* copy pte entries from GART */
 	void (*copy_pte)(struct amdgpu_ib *ib,
 			 uint64_t pe, uint64_t src,
 			 unsigned count);
+
 	/* write pte one entry at a time with addr mapping */
 	void (*write_pte)(struct amdgpu_ib *ib, uint64_t pe,
 			  uint64_t value, unsigned count,
 			  uint32_t incr);
+
+	/* maximum nums of PTEs/PDEs in a single operation */
+	uint32_t	set_max_nums_pte_pde;
+
+	/* number of dw to reserve per operation */
+	unsigned	set_pte_pde_num_dw;
+
 	/* for linear pte/pde updates without addr mapping */
 	void (*set_pte_pde)(struct amdgpu_ib *ib,
 			    uint64_t pe,
@@ -332,6 +345,7 @@ struct amdgpu_gart_funcs {
 struct amdgpu_ih_funcs {
 	/* ring read/write ptr handling, called from interrupt context */
 	u32 (*get_wptr)(struct amdgpu_device *adev);
+	bool (*prescreen_iv)(struct amdgpu_device *adev);
 	void (*decode_iv)(struct amdgpu_device *adev,
 			  struct amdgpu_iv_entry *entry);
 	void (*set_rptr)(struct amdgpu_device *adev);
@@ -399,6 +413,7 @@ void amdgpu_gem_prime_unpin(struct drm_gem_object *obj);
 struct reservation_object *amdgpu_gem_prime_res_obj(struct drm_gem_object *);
 void *amdgpu_gem_prime_vmap(struct drm_gem_object *obj);
 void amdgpu_gem_prime_vunmap(struct drm_gem_object *obj, void *vaddr);
+int amdgpu_gem_prime_mmap(struct drm_gem_object *obj, struct vm_area_struct *vma);
 int amdgpu_gem_debugfs_init(struct amdgpu_device *adev);
 
 /* sub-allocation manager, it has to be protected by another lock.
@@ -455,9 +470,10 @@ struct amdgpu_sa_bo {
  */
 void amdgpu_gem_force_release(struct amdgpu_device *adev);
 int amdgpu_gem_object_create(struct amdgpu_device *adev, unsigned long size,
-				int alignment, u32 initial_domain,
-				u64 flags, bool kernel,
-				struct drm_gem_object **obj);
+			     int alignment, u32 initial_domain,
+			     u64 flags, bool kernel,
+			     struct reservation_object *resv,
+			     struct drm_gem_object **obj);
 
 int amdgpu_mode_dumb_create(struct drm_file *file_priv,
 			    struct drm_device *dev,
@@ -731,8 +747,8 @@ struct amdgpu_ctx_mgr {
 struct amdgpu_ctx *amdgpu_ctx_get(struct amdgpu_fpriv *fpriv, uint32_t id);
 int amdgpu_ctx_put(struct amdgpu_ctx *ctx);
 
-uint64_t amdgpu_ctx_add_fence(struct amdgpu_ctx *ctx, struct amdgpu_ring *ring,
-			      struct dma_fence *fence);
+int amdgpu_ctx_add_fence(struct amdgpu_ctx *ctx, struct amdgpu_ring *ring,
+			      struct dma_fence *fence, uint64_t *seq);
 struct dma_fence *amdgpu_ctx_get_fence(struct amdgpu_ctx *ctx,
 				   struct amdgpu_ring *ring, uint64_t seq);
 
@@ -1014,7 +1030,6 @@ struct amdgpu_gfx {
 	/* reset mask */
 	uint32_t                        grbm_soft_reset;
 	uint32_t                        srbm_soft_reset;
-	bool                            in_reset;
 	/* s3/s4 mask */
 	bool                            in_suspend;
 	/* NGG */
@@ -1056,6 +1071,7 @@ struct amdgpu_cs_parser {
 	/* buffer objects */
 	struct ww_acquire_ctx		ticket;
 	struct amdgpu_bo_list		*bo_list;
+	struct amdgpu_mn		*mn;
 	struct amdgpu_bo_list_entry	vm_pd;
 	struct list_head		validated;
 	struct dma_fence		*fence;
@@ -1183,6 +1199,9 @@ struct amdgpu_firmware {
 
 	/* gpu info firmware data pointer */
 	const struct firmware *gpu_info_fw;
+
+	void *fw_buf_ptr;
+	uint64_t fw_buf_mc;
 };
 
 /*
@@ -1196,20 +1215,6 @@ void amdgpu_benchmark(struct amdgpu_device *adev, int test_number);
  */
 void amdgpu_test_moves(struct amdgpu_device *adev);
 
-/*
- * MMU Notifier
- */
-#if defined(CONFIG_MMU_NOTIFIER)
-int amdgpu_mn_register(struct amdgpu_bo *bo, unsigned long addr);
-void amdgpu_mn_unregister(struct amdgpu_bo *bo);
-#else
-static inline int amdgpu_mn_register(struct amdgpu_bo *bo, unsigned long addr)
-{
-	return -ENODEV;
-}
-static inline void amdgpu_mn_unregister(struct amdgpu_bo *bo) {}
-#endif
-
 /*
  * Debugfs
  */
@@ -1592,6 +1597,7 @@ struct amdgpu_device {
 
 	/* record last mm index being written through WREG32*/
 	unsigned long last_mm_index;
+	bool                            in_sriov_reset;
 };
 
 static inline struct amdgpu_device *amdgpu_ttm_adev(struct ttm_bo_device *bdev)
@@ -1759,6 +1765,7 @@ amdgpu_get_sdma_instance(struct amdgpu_ring *ring)
 #define amdgpu_ring_init_cond_exec(r) (r)->funcs->init_cond_exec((r))
 #define amdgpu_ring_patch_cond_exec(r,o) (r)->funcs->patch_cond_exec((r),(o))
 #define amdgpu_ih_get_wptr(adev) (adev)->irq.ih_funcs->get_wptr((adev))
+#define amdgpu_ih_prescreen_iv(adev) (adev)->irq.ih_funcs->prescreen_iv((adev))
 #define amdgpu_ih_decode_iv(adev, iv) (adev)->irq.ih_funcs->decode_iv((adev), (iv))
 #define amdgpu_ih_set_rptr(adev) (adev)->irq.ih_funcs->set_rptr((adev))
 #define amdgpu_display_vblank_get_counter(adev, crtc) (adev)->mode_info.funcs->vblank_get_counter((adev), (crtc))
@@ -1791,18 +1798,6 @@ void amdgpu_cs_report_moved_bytes(struct amdgpu_device *adev, u64 num_bytes,
 				  u64 num_vis_bytes);
 void amdgpu_ttm_placement_from_domain(struct amdgpu_bo *abo, u32 domain);
 bool amdgpu_ttm_bo_is_amdgpu_bo(struct ttm_buffer_object *bo);
-int amdgpu_ttm_tt_get_user_pages(struct ttm_tt *ttm, struct page **pages);
-int amdgpu_ttm_tt_set_userptr(struct ttm_tt *ttm, uint64_t addr,
-				     uint32_t flags);
-bool amdgpu_ttm_tt_has_userptr(struct ttm_tt *ttm);
-struct mm_struct *amdgpu_ttm_tt_get_usermm(struct ttm_tt *ttm);
-bool amdgpu_ttm_tt_affect_userptr(struct ttm_tt *ttm, unsigned long start,
-				  unsigned long end);
-bool amdgpu_ttm_tt_userptr_invalidated(struct ttm_tt *ttm,
-				       int *last_invalidated);
-bool amdgpu_ttm_tt_is_readonly(struct ttm_tt *ttm);
-uint64_t amdgpu_ttm_tt_pte_flags(struct amdgpu_device *adev, struct ttm_tt *ttm,
-				 struct ttm_mem_reg *mem);
 void amdgpu_vram_location(struct amdgpu_device *adev, struct amdgpu_mc *mc, u64 base);
 void amdgpu_gart_location(struct amdgpu_device *adev, struct amdgpu_mc *mc);
 void amdgpu_ttm_set_active_vram_size(struct amdgpu_device *adev, u64 size);
@@ -1885,10 +1880,9 @@ static inline int amdgpu_acpi_init(struct amdgpu_device *adev) { return 0; }
 static inline void amdgpu_acpi_fini(struct amdgpu_device *adev) { }
 #endif
 
-struct amdgpu_bo_va_mapping *
-amdgpu_cs_find_mapping(struct amdgpu_cs_parser *parser,
-		       uint64_t addr, struct amdgpu_bo **bo);
-int amdgpu_cs_sysvm_access_required(struct amdgpu_cs_parser *parser);
+int amdgpu_cs_find_mapping(struct amdgpu_cs_parser *parser,
+			   uint64_t addr, struct amdgpu_bo **bo,
+			   struct amdgpu_bo_va_mapping **mapping);
 
 #include "amdgpu_object.h"
 #endif

drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v7.c

@@ -169,6 +169,8 @@ static const struct kfd2kgd_calls kfd2kgd = {
 	.get_vmem_size = get_vmem_size,
 	.get_gpu_clock_counter = get_gpu_clock_counter,
 	.get_max_engine_clock_in_mhz = get_max_engine_clock_in_mhz,
+	.alloc_pasid = amdgpu_vm_alloc_pasid,
+	.free_pasid = amdgpu_vm_free_pasid,
 	.program_sh_mem_settings = kgd_program_sh_mem_settings,
 	.set_pasid_vmid_mapping = kgd_set_pasid_vmid_mapping,
 	.init_pipeline = kgd_init_pipeline,

drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v8.c

@@ -128,6 +128,8 @@ static const struct kfd2kgd_calls kfd2kgd = {
 	.get_vmem_size = get_vmem_size,
 	.get_gpu_clock_counter = get_gpu_clock_counter,
 	.get_max_engine_clock_in_mhz = get_max_engine_clock_in_mhz,
+	.alloc_pasid = amdgpu_vm_alloc_pasid,
+	.free_pasid = amdgpu_vm_free_pasid,
 	.program_sh_mem_settings = kgd_program_sh_mem_settings,
 	.set_pasid_vmid_mapping = kgd_set_pasid_vmid_mapping,
 	.init_pipeline = kgd_init_pipeline,

drivers/gpu/drm/amd/amdgpu/amdgpu_cgs.c

@@ -45,7 +45,6 @@ struct amdgpu_cgs_device {
 static int amdgpu_cgs_alloc_gpu_mem(struct cgs_device *cgs_device,
 				    enum cgs_gpu_mem_type type,
 				    uint64_t size, uint64_t align,
-				    uint64_t min_offset, uint64_t max_offset,
 				    cgs_handle_t *handle)
 {
 	CGS_FUNC_ADEV;
@@ -53,13 +52,6 @@ static int amdgpu_cgs_alloc_gpu_mem(struct cgs_device *cgs_device,
 	int ret = 0;
 	uint32_t domain = 0;
 	struct amdgpu_bo *obj;
-	struct ttm_placement placement;
-	struct ttm_place place;
-
-	if (min_offset > max_offset) {
-		BUG_ON(1);
-		return -EINVAL;
-	}
 
 	/* fail if the alignment is not a power of 2 */
 	if (((align != 1) && (align & (align - 1)))
@@ -73,41 +65,19 @@ static int amdgpu_cgs_alloc_gpu_mem(struct cgs_device *cgs_device,
 		flags = AMDGPU_GEM_CREATE_CPU_ACCESS_REQUIRED |
 			AMDGPU_GEM_CREATE_VRAM_CONTIGUOUS;
 		domain = AMDGPU_GEM_DOMAIN_VRAM;
-		if (max_offset > adev->mc.real_vram_size)
-			return -EINVAL;
-		place.fpfn = min_offset >> PAGE_SHIFT;
-		place.lpfn = max_offset >> PAGE_SHIFT;
-		place.flags = TTM_PL_FLAG_WC | TTM_PL_FLAG_UNCACHED |
-			TTM_PL_FLAG_VRAM;
 		break;
 	case CGS_GPU_MEM_TYPE__INVISIBLE_CONTIG_FB:
 	case CGS_GPU_MEM_TYPE__INVISIBLE_FB:
 		flags = AMDGPU_GEM_CREATE_NO_CPU_ACCESS |
 			AMDGPU_GEM_CREATE_VRAM_CONTIGUOUS;
 		domain = AMDGPU_GEM_DOMAIN_VRAM;
-		if (adev->mc.visible_vram_size < adev->mc.real_vram_size) {
-			place.fpfn =
-				max(min_offset, adev->mc.visible_vram_size) >> PAGE_SHIFT;
-			place.lpfn =
-				min(max_offset, adev->mc.real_vram_size) >> PAGE_SHIFT;
-			place.flags = TTM_PL_FLAG_WC | TTM_PL_FLAG_UNCACHED |
-				TTM_PL_FLAG_VRAM;
-		}
-
 		break;
 	case CGS_GPU_MEM_TYPE__GART_CACHEABLE:
 		domain = AMDGPU_GEM_DOMAIN_GTT;
-		place.fpfn = min_offset >> PAGE_SHIFT;
-		place.lpfn = max_offset >> PAGE_SHIFT;
-		place.flags = TTM_PL_FLAG_CACHED | TTM_PL_FLAG_TT;
 		break;
 	case CGS_GPU_MEM_TYPE__GART_WRITECOMBINE:
 		flags = AMDGPU_GEM_CREATE_CPU_GTT_USWC;
 		domain = AMDGPU_GEM_DOMAIN_GTT;
-		place.fpfn = min_offset >> PAGE_SHIFT;
-		place.lpfn = max_offset >> PAGE_SHIFT;
-		place.flags = TTM_PL_FLAG_WC | TTM_PL_FLAG_TT |
-			TTM_PL_FLAG_UNCACHED;
 		break;
 	default:
 		return -EINVAL;
@@ -116,15 +86,8 @@ static int amdgpu_cgs_alloc_gpu_mem(struct cgs_device *cgs_device,
 
 	*handle = 0;
 
-	placement.placement = &place;
-	placement.num_placement = 1;
-	placement.busy_placement = &place;
-	placement.num_busy_placement = 1;
-
-	ret = amdgpu_bo_create_restricted(adev, size, PAGE_SIZE,
-					  true, domain, flags,
-					  NULL, &placement, NULL,
-					  0, &obj);
+	ret = amdgpu_bo_create(adev, size, align, true, domain, flags,
+			       NULL, NULL, 0, &obj);
 	if (ret) {
 		DRM_ERROR("(%d) bo create failed\n", ret);
 		return ret;
@@ -155,19 +118,14 @@ static int amdgpu_cgs_gmap_gpu_mem(struct cgs_device *cgs_device, cgs_handle_t h
 				   uint64_t *mcaddr)
 {
 	int r;
-	u64 min_offset, max_offset;
 	struct amdgpu_bo *obj = (struct amdgpu_bo *)handle;
 
 	WARN_ON_ONCE(obj->placement.num_placement > 1);
 
-	min_offset = obj->placements[0].fpfn << PAGE_SHIFT;
-	max_offset = obj->placements[0].lpfn << PAGE_SHIFT;
-
 	r = amdgpu_bo_reserve(obj, true);
 	if (unlikely(r != 0))
 		return r;
-	r = amdgpu_bo_pin_restricted(obj, obj->preferred_domains,
-				     min_offset, max_offset, mcaddr);
+	r = amdgpu_bo_pin(obj, obj->preferred_domains, mcaddr);
 	amdgpu_bo_unreserve(obj);
 	return r;
 }
@@ -675,6 +633,85 @@ static int amdgpu_cgs_get_firmware_info(struct cgs_device *cgs_device,
 
 		if (!adev->pm.fw) {
 			switch (adev->asic_type) {
+			case CHIP_TAHITI:
+				strcpy(fw_name, "radeon/tahiti_smc.bin");
+				break;
+			case CHIP_PITCAIRN:
+				if ((adev->pdev->revision == 0x81) &&
+				    ((adev->pdev->device == 0x6810) ||
+				    (adev->pdev->device == 0x6811))) {
+					info->is_kicker = true;
+					strcpy(fw_name, "radeon/pitcairn_k_smc.bin");
+				} else {
+					strcpy(fw_name, "radeon/pitcairn_smc.bin");
+				}
+				break;
+			case CHIP_VERDE:
+				if (((adev->pdev->device == 0x6820) &&
+					((adev->pdev->revision == 0x81) ||
+					(adev->pdev->revision == 0x83))) ||
+				    ((adev->pdev->device == 0x6821) &&
+					((adev->pdev->revision == 0x83) ||
+					(adev->pdev->revision == 0x87))) ||
+				    ((adev->pdev->revision == 0x87) &&
+					((adev->pdev->device == 0x6823) ||
+					(adev->pdev->device == 0x682b)))) {
+					info->is_kicker = true;
+					strcpy(fw_name, "radeon/verde_k_smc.bin");
+				} else {
+					strcpy(fw_name, "radeon/verde_smc.bin");
+				}
+				break;
+			case CHIP_OLAND:
+				if (((adev->pdev->revision == 0x81) &&
+					((adev->pdev->device == 0x6600) ||
+					(adev->pdev->device == 0x6604) ||
+					(adev->pdev->device == 0x6605) ||
+					(adev->pdev->device == 0x6610))) ||
+				    ((adev->pdev->revision == 0x83) &&
+					(adev->pdev->device == 0x6610))) {
+					info->is_kicker = true;
+					strcpy(fw_name, "radeon/oland_k_smc.bin");
+				} else {
+					strcpy(fw_name, "radeon/oland_smc.bin");
+				}
+				break;
+			case CHIP_HAINAN:
+				if (((adev->pdev->revision == 0x81) &&
+					(adev->pdev->device == 0x6660)) ||
+				    ((adev->pdev->revision == 0x83) &&
+					((adev->pdev->device == 0x6660) ||
+					(adev->pdev->device == 0x6663) ||
+					(adev->pdev->device == 0x6665) ||
+					 (adev->pdev->device == 0x6667)))) {
+					info->is_kicker = true;
+					strcpy(fw_name, "radeon/hainan_k_smc.bin");
+				} else if ((adev->pdev->revision == 0xc3) &&
+					 (adev->pdev->device == 0x6665)) {
+					info->is_kicker = true;
+					strcpy(fw_name, "radeon/banks_k_2_smc.bin");
+				} else {
+					strcpy(fw_name, "radeon/hainan_smc.bin");
+				}
+				break;
+			case CHIP_BONAIRE:
+				if ((adev->pdev->revision == 0x80) ||
+					(adev->pdev->revision == 0x81) ||
+					(adev->pdev->device == 0x665f)) {
+					info->is_kicker = true;
+					strcpy(fw_name, "radeon/bonaire_k_smc.bin");
+				} else {
+					strcpy(fw_name, "radeon/bonaire_smc.bin");
+				}
+				break;
+			case CHIP_HAWAII:
+				if (adev->pdev->revision == 0x80) {
+					info->is_kicker = true;
+					strcpy(fw_name, "radeon/hawaii_k_smc.bin");
+				} else {
+					strcpy(fw_name, "radeon/hawaii_smc.bin");
+				}
+				break;
 			case CHIP_TOPAZ:
 				if (((adev->pdev->device == 0x6900) && (adev->pdev->revision == 0x81)) ||
 				    ((adev->pdev->device == 0x6900) && (adev->pdev->revision == 0x83)) ||
@@ -838,6 +875,9 @@ static int amdgpu_cgs_query_system_info(struct cgs_device *cgs_device,
 	case CGS_SYSTEM_INFO_PCIE_SUB_SYS_VENDOR_ID:
 		sys_info->value = adev->pdev->subsystem_vendor;
 		break;
+	case CGS_SYSTEM_INFO_PCIE_BUS_DEVFN:
+		sys_info->value = adev->pdev->devfn;
+		break;
 	default:
 		return -ENODEV;
 	}

drivers/gpu/drm/amd/amdgpu/amdgpu_connectors.c

@@ -346,10 +346,8 @@ static void amdgpu_connector_free_edid(struct drm_connector *connector)
 {
 	struct amdgpu_connector *amdgpu_connector = to_amdgpu_connector(connector);
 
-	if (amdgpu_connector->edid) {
-		kfree(amdgpu_connector->edid);
-		amdgpu_connector->edid = NULL;
-	}
+	kfree(amdgpu_connector->edid);
+	amdgpu_connector->edid = NULL;
 }
 
 static int amdgpu_connector_ddc_get_modes(struct drm_connector *connector)

drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c

@@ -473,11 +473,16 @@ static int amdgpu_cs_list_validate(struct amdgpu_cs_parser *p,
 			return -EPERM;
 
 		/* Check if we have user pages and nobody bound the BO already */
-		if (lobj->user_pages && bo->tbo.ttm->state != tt_bound) {
-			size_t size = sizeof(struct page *);
-
-			size *= bo->tbo.ttm->num_pages;
-			memcpy(bo->tbo.ttm->pages, lobj->user_pages, size);
+		if (amdgpu_ttm_tt_userptr_needs_pages(bo->tbo.ttm) &&
+		    lobj->user_pages) {
+			amdgpu_ttm_placement_from_domain(bo,
+							 AMDGPU_GEM_DOMAIN_CPU);
+			r = ttm_bo_validate(&bo->tbo, &bo->placement, true,
+					    false);
+			if (r)
+				return r;
+			amdgpu_ttm_tt_set_user_pages(bo->tbo.ttm,
+						     lobj->user_pages);
 			binding_userptr = true;
 		}
 
@@ -502,7 +507,6 @@ static int amdgpu_cs_parser_bos(struct amdgpu_cs_parser *p,
 	struct amdgpu_fpriv *fpriv = p->filp->driver_priv;
 	struct amdgpu_bo_list_entry *e;
 	struct list_head duplicates;
-	bool need_mmap_lock = false;
 	unsigned i, tries = 10;
 	int r;
 
@@ -510,9 +514,9 @@ static int amdgpu_cs_parser_bos(struct amdgpu_cs_parser *p,
 
 	p->bo_list = amdgpu_bo_list_get(fpriv, cs->in.bo_list_handle);
 	if (p->bo_list) {
-		need_mmap_lock = p->bo_list->first_userptr !=
-			p->bo_list->num_entries;
 		amdgpu_bo_list_get_list(p->bo_list, &p->validated);
+		if (p->bo_list->first_userptr != p->bo_list->num_entries)
+			p->mn = amdgpu_mn_get(p->adev);
 	}
 
 	INIT_LIST_HEAD(&duplicates);
@@ -521,9 +525,6 @@ static int amdgpu_cs_parser_bos(struct amdgpu_cs_parser *p,
 	if (p->uf_entry.robj)
 		list_add(&p->uf_entry.tv.head, &p->validated);
 
-	if (need_mmap_lock)
-		down_read(&current->mm->mmap_sem);
-
 	while (1) {
 		struct list_head need_pages;
 		unsigned i;
@@ -543,23 +544,25 @@ static int amdgpu_cs_parser_bos(struct amdgpu_cs_parser *p,
 		INIT_LIST_HEAD(&need_pages);
 		for (i = p->bo_list->first_userptr;
 		     i < p->bo_list->num_entries; ++i) {
+			struct amdgpu_bo *bo;
 
 			e = &p->bo_list->array[i];
+			bo = e->robj;
 
-			if (amdgpu_ttm_tt_userptr_invalidated(e->robj->tbo.ttm,
+			if (amdgpu_ttm_tt_userptr_invalidated(bo->tbo.ttm,
 				 &e->user_invalidated) && e->user_pages) {
 
 				/* We acquired a page array, but somebody
 				 * invalidated it. Free it and try again
 				 */
 				release_pages(e->user_pages,
-					      e->robj->tbo.ttm->num_pages,
+					      bo->tbo.ttm->num_pages,
 					      false);
 				kvfree(e->user_pages);
 				e->user_pages = NULL;
 			}
 
-			if (e->robj->tbo.ttm->state != tt_bound &&
+			if (amdgpu_ttm_tt_userptr_needs_pages(bo->tbo.ttm) &&
 			    !e->user_pages) {
 				list_del(&e->tv.head);
 				list_add(&e->tv.head, &need_pages);
@@ -636,9 +639,6 @@ static int amdgpu_cs_parser_bos(struct amdgpu_cs_parser *p,
 
 	amdgpu_cs_report_moved_bytes(p->adev, p->bytes_moved,
 				     p->bytes_moved_vis);
-	fpriv->vm.last_eviction_counter =
-		atomic64_read(&p->adev->num_evictions);
-
 	if (p->bo_list) {
 		struct amdgpu_bo *gds = p->bo_list->gds_obj;
 		struct amdgpu_bo *gws = p->bo_list->gws_obj;
@@ -679,9 +679,6 @@ static int amdgpu_cs_parser_bos(struct amdgpu_cs_parser *p,
 
 error_free_pages:
 
-	if (need_mmap_lock)
-		up_read(&current->mm->mmap_sem);
-
 	if (p->bo_list) {
 		for (i = p->bo_list->first_userptr;
 		     i < p->bo_list->num_entries; ++i) {
@@ -728,11 +725,7 @@ static void amdgpu_cs_parser_fini(struct amdgpu_cs_parser *parser, int error,
 {
 	unsigned i;
 
-	if (!error)
-		ttm_eu_fence_buffer_objects(&parser->ticket,
-					    &parser->validated,
-					    parser->fence);
-	else if (backoff)
+	if (error && backoff)
 		ttm_eu_backoff_reservation(&parser->ticket,
 					   &parser->validated);
 
@@ -768,10 +761,6 @@ static int amdgpu_bo_vm_update_pte(struct amdgpu_cs_parser *p)
 	if (r)
 		return r;
 
-	r = amdgpu_sync_fence(adev, &p->job->sync, vm->last_dir_update);
-	if (r)
-		return r;
-
 	r = amdgpu_vm_clear_freed(adev, vm, NULL);
 	if (r)
 		return r;
@@ -825,7 +814,13 @@ static int amdgpu_bo_vm_update_pte(struct amdgpu_cs_parser *p)
 
 	}
 
-	r = amdgpu_vm_clear_moved(adev, vm, &p->job->sync);
+	r = amdgpu_vm_handle_moved(adev, vm);
+	if (r)
+		return r;
+
+	r = amdgpu_sync_fence(adev, &p->job->sync, vm->last_update);
+	if (r)
+		return r;
 
 	if (amdgpu_vm_debug && p->bo_list) {
 		/* Invalidate all BOs to test for userspace bugs */
@@ -835,7 +830,7 @@ static int amdgpu_bo_vm_update_pte(struct amdgpu_cs_parser *p)
 			if (!bo)
 				continue;
 
-			amdgpu_vm_bo_invalidate(adev, bo);
+			amdgpu_vm_bo_invalidate(adev, bo, false);
 		}
 	}
 
@@ -860,7 +855,7 @@ static int amdgpu_cs_ib_vm_chunk(struct amdgpu_device *adev,
 	}
 
 	if (p->job->vm) {
-		p->job->vm_pd_addr = amdgpu_bo_gpu_offset(vm->root.bo);
+		p->job->vm_pd_addr = amdgpu_bo_gpu_offset(vm->root.base.bo);
 
 		r = amdgpu_bo_vm_update_pte(p);
 		if (r)
@@ -928,11 +923,11 @@ static int amdgpu_cs_ib_fill(struct amdgpu_device *adev,
 			uint64_t offset;
 			uint8_t *kptr;
 
-			m = amdgpu_cs_find_mapping(parser, chunk_ib->va_start,
-						   &aobj);
-			if (!aobj) {
+			r = amdgpu_cs_find_mapping(parser, chunk_ib->va_start,
+						   &aobj, &m);
+			if (r) {
 				DRM_ERROR("IB va_start is invalid\n");
-				return -EINVAL;
+				return r;
 			}
 
 			if ((chunk_ib->va_start + chunk_ib->ib_bytes) >
@@ -1133,14 +1128,31 @@ static int amdgpu_cs_submit(struct amdgpu_cs_parser *p,
 	struct amdgpu_ring *ring = p->job->ring;
 	struct amd_sched_entity *entity = &p->ctx->rings[ring->idx].entity;
 	struct amdgpu_job *job;
+	unsigned i;
+	uint64_t seq;
+
 	int r;
 
+	amdgpu_mn_lock(p->mn);
+	if (p->bo_list) {
+		for (i = p->bo_list->first_userptr;
+		     i < p->bo_list->num_entries; ++i) {
+			struct amdgpu_bo *bo = p->bo_list->array[i].robj;
+
+			if (amdgpu_ttm_tt_userptr_needs_pages(bo->tbo.ttm)) {
+				amdgpu_mn_unlock(p->mn);
+				return -ERESTARTSYS;
+			}
+		}
+	}
+
 	job = p->job;
 	p->job = NULL;
 
 	r = amd_sched_job_init(&job->base, &ring->sched, entity, p->filp);
 	if (r) {
 		amdgpu_job_free(job);
+		amdgpu_mn_unlock(p->mn);
 		return r;
 	}
 
@@ -1148,14 +1160,28 @@ static int amdgpu_cs_submit(struct amdgpu_cs_parser *p,
 	job->fence_ctx = entity->fence_context;
 	p->fence = dma_fence_get(&job->base.s_fence->finished);
 
+	r = amdgpu_ctx_add_fence(p->ctx, ring, p->fence, &seq);
+	if (r) {
+		dma_fence_put(p->fence);
+		dma_fence_put(&job->base.s_fence->finished);
+		amdgpu_job_free(job);
+		amdgpu_mn_unlock(p->mn);
+		return r;
+	}
+
 	amdgpu_cs_post_dependencies(p);
 
-	cs->out.handle = amdgpu_ctx_add_fence(p->ctx, ring, p->fence);
-	job->uf_sequence = cs->out.handle;
+	cs->out.handle = seq;
+	job->uf_sequence = seq;
+
 	amdgpu_job_free_resources(job);
 
 	trace_amdgpu_cs_ioctl(job);
 	amd_sched_entity_push_job(&job->base);
+
+	ttm_eu_fence_buffer_objects(&p->ticket, &p->validated, p->fence);
+	amdgpu_mn_unlock(p->mn);
+
 	return 0;
 }
 
@@ -1383,6 +1409,7 @@ static int amdgpu_cs_wait_any_fence(struct amdgpu_device *adev,
 			array[i] = fence;
 		} else { /* NULL, the fence has been already signaled */
 			r = 1;
+			first = i;
 			goto out;
 		}
 	}
@@ -1462,78 +1489,36 @@ int amdgpu_cs_wait_fences_ioctl(struct drm_device *dev, void *data,
  * virtual memory address. Returns allocation structure when found, NULL
  * otherwise.
  */
-struct amdgpu_bo_va_mapping *
-amdgpu_cs_find_mapping(struct amdgpu_cs_parser *parser,
-		       uint64_t addr, struct amdgpu_bo **bo)
+int amdgpu_cs_find_mapping(struct amdgpu_cs_parser *parser,
+			   uint64_t addr, struct amdgpu_bo **bo,
+			   struct amdgpu_bo_va_mapping **map)
 {
+	struct amdgpu_fpriv *fpriv = parser->filp->driver_priv;
+	struct amdgpu_vm *vm = &fpriv->vm;
 	struct amdgpu_bo_va_mapping *mapping;
-	unsigned i;
-
-	if (!parser->bo_list)
-		return NULL;
+	int r;
 
 	addr /= AMDGPU_GPU_PAGE_SIZE;
 
-	for (i = 0; i < parser->bo_list->num_entries; i++) {
-		struct amdgpu_bo_list_entry *lobj;
-
-		lobj = &parser->bo_list->array[i];
-		if (!lobj->bo_va)
-			continue;
-
-		list_for_each_entry(mapping, &lobj->bo_va->valids, list) {
-			if (mapping->start > addr ||
-			    addr > mapping->last)
-				continue;
-
-			*bo = lobj->bo_va->base.bo;
-			return mapping;
-		}
-
-		list_for_each_entry(mapping, &lobj->bo_va->invalids, list) {
-			if (mapping->start > addr ||
-			    addr > mapping->last)
-				continue;
+	mapping = amdgpu_vm_bo_lookup_mapping(vm, addr);
+	if (!mapping || !mapping->bo_va || !mapping->bo_va->base.bo)
+		return -EINVAL;
 
-			*bo = lobj->bo_va->base.bo;
-			return mapping;
-		}
-	}
+	*bo = mapping->bo_va->base.bo;
+	*map = mapping;
 
-	return NULL;
-}
+	/* Double check that the BO is reserved by this CS */
+	if (READ_ONCE((*bo)->tbo.resv->lock.ctx) != &parser->ticket)
+		return -EINVAL;
 
-/**
- * amdgpu_cs_sysvm_access_required - make BOs accessible by the system VM
- *
- * @parser: command submission parser context
- *
- * Helper for UVD/VCE VM emulation, make sure BOs are accessible by the system VM.
- */
-int amdgpu_cs_sysvm_access_required(struct amdgpu_cs_parser *parser)
-{
-	unsigned i;
-	int r;
+	r = amdgpu_ttm_bind(&(*bo)->tbo, &(*bo)->tbo.mem);
+	if (unlikely(r))
+		return r;
 
-	if (!parser->bo_list)
+	if ((*bo)->flags & AMDGPU_GEM_CREATE_VRAM_CONTIGUOUS)
 		return 0;
 
-	for (i = 0; i < parser->bo_list->num_entries; i++) {
-		struct amdgpu_bo *bo = parser->bo_list->array[i].robj;
-
-		r = amdgpu_ttm_bind(&bo->tbo, &bo->tbo.mem);
-		if (unlikely(r))
-			return r;
-
-		if (bo->flags & AMDGPU_GEM_CREATE_VRAM_CONTIGUOUS)
-			continue;
-
-		bo->flags |= AMDGPU_GEM_CREATE_VRAM_CONTIGUOUS;
-		amdgpu_ttm_placement_from_domain(bo, bo->allowed_domains);
-		r = ttm_bo_validate(&bo->tbo, &bo->placement, false, false);
-		if (unlikely(r))
-			return r;
-	}
-
-	return 0;
+	(*bo)->flags |= AMDGPU_GEM_CREATE_VRAM_CONTIGUOUS;
+	amdgpu_ttm_placement_from_domain(*bo, (*bo)->allowed_domains);
+	return ttm_bo_validate(&(*bo)->tbo, &(*bo)->placement, false, false);
 }

drivers/gpu/drm/amd/amdgpu/amdgpu_ctx.c

@@ -246,8 +246,8 @@ int amdgpu_ctx_put(struct amdgpu_ctx *ctx)
 	return 0;
 }
 
-uint64_t amdgpu_ctx_add_fence(struct amdgpu_ctx *ctx, struct amdgpu_ring *ring,
-			      struct dma_fence *fence)
+int amdgpu_ctx_add_fence(struct amdgpu_ctx *ctx, struct amdgpu_ring *ring,
+			      struct dma_fence *fence, uint64_t* handler)
 {
 	struct amdgpu_ctx_ring *cring = & ctx->rings[ring->idx];
 	uint64_t seq = cring->sequence;
@@ -258,9 +258,9 @@ uint64_t amdgpu_ctx_add_fence(struct amdgpu_ctx *ctx, struct amdgpu_ring *ring,
 	other = cring->fences[idx];
 	if (other) {
 		signed long r;
-		r = dma_fence_wait_timeout(other, false, MAX_SCHEDULE_TIMEOUT);
+		r = dma_fence_wait_timeout(other, true, MAX_SCHEDULE_TIMEOUT);
 		if (r < 0)
-			DRM_ERROR("Error (%ld) waiting for fence!\n", r);
+			return r;
 	}
 
 	dma_fence_get(fence);
@@ -271,8 +271,10 @@ uint64_t amdgpu_ctx_add_fence(struct amdgpu_ctx *ctx, struct amdgpu_ring *ring,
 	spin_unlock(&ctx->ring_lock);
 
 	dma_fence_put(other);
+	if (handler)
+		*handler = seq;
 
-	return seq;
+	return 0;
 }
 
 struct dma_fence *amdgpu_ctx_get_fence(struct amdgpu_ctx *ctx,

drivers/gpu/drm/amd/amdgpu/amdgpu_device.c

@@ -65,6 +65,7 @@ MODULE_FIRMWARE("amdgpu/raven_gpu_info.bin");
 static int amdgpu_debugfs_regs_init(struct amdgpu_device *adev);
 static void amdgpu_debugfs_regs_cleanup(struct amdgpu_device *adev);
 static int amdgpu_debugfs_test_ib_ring_init(struct amdgpu_device *adev);
+static int amdgpu_debugfs_vbios_dump_init(struct amdgpu_device *adev);
 
 static const char *amdgpu_asic_name[] = {
 	"TAHITI",
@@ -402,6 +403,15 @@ void amdgpu_pci_config_reset(struct amdgpu_device *adev)
  */
 static int amdgpu_doorbell_init(struct amdgpu_device *adev)
 {
+	/* No doorbell on SI hardware generation */
+	if (adev->asic_type < CHIP_BONAIRE) {
+		adev->doorbell.base = 0;
+		adev->doorbell.size = 0;
+		adev->doorbell.num_doorbells = 0;
+		adev->doorbell.ptr = NULL;
+		return 0;
+	}
+
 	/* doorbell bar mapping */
 	adev->doorbell.base = pci_resource_start(adev->pdev, 2);
 	adev->doorbell.size = pci_resource_len(adev->pdev, 2);
@@ -887,6 +897,20 @@ static uint32_t cail_ioreg_read(struct card_info *info, uint32_t reg)
 	return r;
 }
 
+static ssize_t amdgpu_atombios_get_vbios_version(struct device *dev,
+						 struct device_attribute *attr,
+						 char *buf)
+{
+	struct drm_device *ddev = dev_get_drvdata(dev);
+	struct amdgpu_device *adev = ddev->dev_private;
+	struct atom_context *ctx = adev->mode_info.atom_context;
+
+	return snprintf(buf, PAGE_SIZE, "%s\n", ctx->vbios_version);
+}
+
+static DEVICE_ATTR(vbios_version, 0444, amdgpu_atombios_get_vbios_version,
+		   NULL);
+
 /**
  * amdgpu_atombios_fini - free the driver info and callbacks for atombios
  *
@@ -906,6 +930,7 @@ static void amdgpu_atombios_fini(struct amdgpu_device *adev)
 	adev->mode_info.atom_context = NULL;
 	kfree(adev->mode_info.atom_card_info);
 	adev->mode_info.atom_card_info = NULL;
+	device_remove_file(adev->dev, &dev_attr_vbios_version);
 }
 
 /**
@@ -922,6 +947,7 @@ static int amdgpu_atombios_init(struct amdgpu_device *adev)
 {
 	struct card_info *atom_card_info =
 	    kzalloc(sizeof(struct card_info), GFP_KERNEL);
+	int ret;
 
 	if (!atom_card_info)
 		return -ENOMEM;
@@ -958,6 +984,13 @@ static int amdgpu_atombios_init(struct amdgpu_device *adev)
 		amdgpu_atombios_scratch_regs_init(adev);
 		amdgpu_atombios_allocate_fb_scratch(adev);
 	}
+
+	ret = device_create_file(adev->dev, &dev_attr_vbios_version);
+	if (ret) {
+		DRM_ERROR("Failed to create device file for VBIOS version\n");
+		return ret;
+	}
+
 	return 0;
 }
 
@@ -1757,10 +1790,8 @@ static int amdgpu_fini(struct amdgpu_device *adev)
 		adev->ip_blocks[i].status.late_initialized = false;
 	}
 
-	if (amdgpu_sriov_vf(adev)) {
-		amdgpu_bo_free_kernel(&adev->virt.csa_obj, &adev->virt.csa_vmid0_addr, NULL);
+	if (amdgpu_sriov_vf(adev))
 		amdgpu_virt_release_full_gpu(adev, false);
-	}
 
 	return 0;
 }
@@ -2051,9 +2082,8 @@ int amdgpu_device_init(struct amdgpu_device *adev,
 	DRM_INFO("register mmio base: 0x%08X\n", (uint32_t)adev->rmmio_base);
 	DRM_INFO("register mmio size: %u\n", (unsigned)adev->rmmio_size);
 
-	if (adev->asic_type >= CHIP_BONAIRE)
-		/* doorbell bar mapping */
-		amdgpu_doorbell_init(adev);
+	/* doorbell bar mapping */
+	amdgpu_doorbell_init(adev);
 
 	/* io port mapping */
 	for (i = 0; i < DEVICE_COUNT_RESOURCE; i++) {
@@ -2201,6 +2231,10 @@ int amdgpu_device_init(struct amdgpu_device *adev,
 	if (r)
 		DRM_ERROR("registering firmware debugfs failed (%d).\n", r);
 
+	r = amdgpu_debugfs_vbios_dump_init(adev);
+	if (r)
+		DRM_ERROR("Creating vbios dump debugfs failed (%d).\n", r);
+
 	if ((amdgpu_testing & 1)) {
 		if (adev->accel_working)
 			amdgpu_test_moves(adev);
@@ -2276,8 +2310,7 @@ void amdgpu_device_fini(struct amdgpu_device *adev)
 	adev->rio_mem = NULL;
 	iounmap(adev->rmmio);
 	adev->rmmio = NULL;
-	if (adev->asic_type >= CHIP_BONAIRE)
-		amdgpu_doorbell_fini(adev);
+	amdgpu_doorbell_fini(adev);
 	amdgpu_debugfs_regs_cleanup(adev);
 }
 
@@ -2546,7 +2579,8 @@ static bool amdgpu_need_full_reset(struct amdgpu_device *adev)
 		if ((adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_GMC) ||
 		    (adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_SMC) ||
 		    (adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_ACP) ||
-		    (adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_DCE)) {
+		    (adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_DCE) ||
+		     adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_PSP) {
 			if (adev->ip_blocks[i].status.hang) {
 				DRM_INFO("Some block need full reset!\n");
 				return true;
@@ -2654,7 +2688,7 @@ int amdgpu_sriov_gpu_reset(struct amdgpu_device *adev, struct amdgpu_job *job)
 
 	mutex_lock(&adev->virt.lock_reset);
 	atomic_inc(&adev->gpu_reset_counter);
-	adev->gfx.in_reset = true;
+	adev->in_sriov_reset = true;
 
 	/* block TTM */
 	resched = ttm_bo_lock_delayed_workqueue(&adev->mman.bdev);
@@ -2765,7 +2799,7 @@ int amdgpu_sriov_gpu_reset(struct amdgpu_device *adev, struct amdgpu_job *job)
 		dev_info(adev->dev, "GPU reset successed!\n");
 	}
 
-	adev->gfx.in_reset = false;
+	adev->in_sriov_reset = false;
 	mutex_unlock(&adev->virt.lock_reset);
 	return r;
 }
@@ -3463,10 +3497,7 @@ static ssize_t amdgpu_debugfs_sensor_read(struct file *f, char __user *buf,
 
 	valuesize = sizeof(values);
 	if (adev->powerplay.pp_funcs && adev->powerplay.pp_funcs->read_sensor)
-		r = adev->powerplay.pp_funcs->read_sensor(adev->powerplay.pp_handle, idx, &values[0], &valuesize);
-	else if (adev->pm.funcs && adev->pm.funcs->read_sensor)
-		r = adev->pm.funcs->read_sensor(adev, idx, &values[0],
-						&valuesize);
+		r = amdgpu_dpm_read_sensor(adev, idx, &values[0], &valuesize);
 	else
 		return -EINVAL;
 
@@ -3754,6 +3785,28 @@ int amdgpu_debugfs_init(struct drm_minor *minor)
 {
 	return 0;
 }
+
+static int amdgpu_debugfs_get_vbios_dump(struct seq_file *m, void *data)
+{
+	struct drm_info_node *node = (struct drm_info_node *) m->private;
+	struct drm_device *dev = node->minor->dev;
+	struct amdgpu_device *adev = dev->dev_private;
+
+	seq_write(m, adev->bios, adev->bios_size);
+	return 0;
+}
+
+static const struct drm_info_list amdgpu_vbios_dump_list[] = {
+		{"amdgpu_vbios",
+		 amdgpu_debugfs_get_vbios_dump,
+		 0, NULL},
+};
+
+static int amdgpu_debugfs_vbios_dump_init(struct amdgpu_device *adev)
+{
+	return amdgpu_debugfs_add_files(adev,
+					amdgpu_vbios_dump_list, 1);
+}
 #else
 static int amdgpu_debugfs_test_ib_ring_init(struct amdgpu_device *adev)
 {
@@ -3763,5 +3816,9 @@ static int amdgpu_debugfs_regs_init(struct amdgpu_device *adev)
 {
 	return 0;
 }
+static int amdgpu_debugfs_vbios_dump_init(struct amdgpu_device *adev)
+{
+	return 0;
+}
 static void amdgpu_debugfs_regs_cleanup(struct amdgpu_device *adev) { }
 #endif

drivers/gpu/drm/amd/amdgpu/amdgpu_dpm.c

@@ -960,8 +960,10 @@ u8 amdgpu_encode_pci_lane_width(u32 lanes)
 }
 
 struct amd_vce_state*
-amdgpu_get_vce_clock_state(struct amdgpu_device *adev, unsigned idx)
+amdgpu_get_vce_clock_state(void *handle, u32 idx)
 {
+	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+
 	if (idx < adev->pm.dpm.num_of_vce_states)
 		return &adev->pm.dpm.vce_states[idx];
 

drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c

@@ -69,9 +69,10 @@
  * - 3.17.0 - Add AMDGPU_NUM_VRAM_CPU_PAGE_FAULTS.
  * - 3.18.0 - Export gpu always on cu bitmap
  * - 3.19.0 - Add support for UVD MJPEG decode
+ * - 3.20.0 - Add support for local BOs
  */
 #define KMS_DRIVER_MAJOR	3
-#define KMS_DRIVER_MINOR	19
+#define KMS_DRIVER_MINOR	20
 #define KMS_DRIVER_PATCHLEVEL	0
 
 int amdgpu_vram_limit = 0;
@@ -91,7 +92,7 @@ int amdgpu_dpm = -1;
 int amdgpu_fw_load_type = -1;
 int amdgpu_aspm = -1;
 int amdgpu_runtime_pm = -1;
-unsigned amdgpu_ip_block_mask = 0xffffffff;
+uint amdgpu_ip_block_mask = 0xffffffff;
 int amdgpu_bapm = -1;
 int amdgpu_deep_color = 0;
 int amdgpu_vm_size = -1;
@@ -106,14 +107,14 @@ int amdgpu_sched_jobs = 32;
 int amdgpu_sched_hw_submission = 2;
 int amdgpu_no_evict = 0;
 int amdgpu_direct_gma_size = 0;
-unsigned amdgpu_pcie_gen_cap = 0;
-unsigned amdgpu_pcie_lane_cap = 0;
-unsigned amdgpu_cg_mask = 0xffffffff;
-unsigned amdgpu_pg_mask = 0xffffffff;
-unsigned amdgpu_sdma_phase_quantum = 32;
+uint amdgpu_pcie_gen_cap = 0;
+uint amdgpu_pcie_lane_cap = 0;
+uint amdgpu_cg_mask = 0xffffffff;
+uint amdgpu_pg_mask = 0xffffffff;
+uint amdgpu_sdma_phase_quantum = 32;
 char *amdgpu_disable_cu = NULL;
 char *amdgpu_virtual_display = NULL;
-unsigned amdgpu_pp_feature_mask = 0xffffffff;
+uint amdgpu_pp_feature_mask = 0xffffffff;
 int amdgpu_ngg = 0;
 int amdgpu_prim_buf_per_se = 0;
 int amdgpu_pos_buf_per_se = 0;
@@ -608,6 +609,8 @@ amdgpu_pci_remove(struct pci_dev *pdev)
 
 	drm_dev_unregister(dev);
 	drm_dev_unref(dev);
+	pci_disable_device(pdev);
+	pci_set_drvdata(pdev, NULL);
 }
 
 static void
@@ -852,6 +855,7 @@ static struct drm_driver kms_driver = {
 	.gem_prime_import_sg_table = amdgpu_gem_prime_import_sg_table,
 	.gem_prime_vmap = amdgpu_gem_prime_vmap,
 	.gem_prime_vunmap = amdgpu_gem_prime_vunmap,
+	.gem_prime_mmap = amdgpu_gem_prime_mmap,
 
 	.name = DRIVER_NAME,
 	.desc = DRIVER_DESC,

drivers/gpu/drm/amd/amdgpu/amdgpu_fb.c

@@ -149,7 +149,7 @@ static int amdgpufb_create_pinned_object(struct amdgpu_fbdev *rfbdev,
 				       AMDGPU_GEM_CREATE_CPU_ACCESS_REQUIRED |
 				       AMDGPU_GEM_CREATE_VRAM_CONTIGUOUS |
 				       AMDGPU_GEM_CREATE_VRAM_CLEARED,
-				       true, &gobj);
+				       true, NULL, &gobj);
 	if (ret) {
 		pr_err("failed to allocate framebuffer (%d)\n", aligned_size);
 		return -ENOMEM;
@@ -303,10 +303,10 @@ static int amdgpu_fbdev_destroy(struct drm_device *dev, struct amdgpu_fbdev *rfb
 	if (rfb->obj) {
 		amdgpufb_destroy_pinned_object(rfb->obj);
 		rfb->obj = NULL;
+		drm_framebuffer_unregister_private(&rfb->base);
+		drm_framebuffer_cleanup(&rfb->base);
 	}
 	drm_fb_helper_fini(&rfbdev->helper);
-	drm_framebuffer_unregister_private(&rfb->base);
-	drm_framebuffer_cleanup(&rfb->base);
 
 	return 0;
 }

drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c

@@ -44,11 +44,12 @@ void amdgpu_gem_object_free(struct drm_gem_object *gobj)
 }
 
 int amdgpu_gem_object_create(struct amdgpu_device *adev, unsigned long size,
-				int alignment, u32 initial_domain,
-				u64 flags, bool kernel,
-				struct drm_gem_object **obj)
+			     int alignment, u32 initial_domain,
+			     u64 flags, bool kernel,
+			     struct reservation_object *resv,
+			     struct drm_gem_object **obj)
 {
-	struct amdgpu_bo *robj;
+	struct amdgpu_bo *bo;
 	int r;
 
 	*obj = NULL;
@@ -59,7 +60,7 @@ int amdgpu_gem_object_create(struct amdgpu_device *adev, unsigned long size,
 
 retry:
 	r = amdgpu_bo_create(adev, size, alignment, kernel, initial_domain,
-			     flags, NULL, NULL, 0, &robj);
+			     flags, NULL, resv, 0, &bo);
 	if (r) {
 		if (r != -ERESTARTSYS) {
 			if (initial_domain == AMDGPU_GEM_DOMAIN_VRAM) {
@@ -71,7 +72,7 @@ int amdgpu_gem_object_create(struct amdgpu_device *adev, unsigned long size,
 		}
 		return r;
 	}
-	*obj = &robj->gem_base;
+	*obj = &bo->gem_base;
 
 	return 0;
 }
@@ -112,7 +113,17 @@ int amdgpu_gem_object_open(struct drm_gem_object *obj,
 	struct amdgpu_fpriv *fpriv = file_priv->driver_priv;
 	struct amdgpu_vm *vm = &fpriv->vm;
 	struct amdgpu_bo_va *bo_va;
+	struct mm_struct *mm;
 	int r;
+
+	mm = amdgpu_ttm_tt_get_usermm(abo->tbo.ttm);
+	if (mm && mm != current->mm)
+		return -EPERM;
+
+	if (abo->flags & AMDGPU_GEM_CREATE_VM_ALWAYS_VALID &&
+	    abo->tbo.resv != vm->root.base.bo->tbo.resv)
+		return -EPERM;
+
 	r = amdgpu_bo_reserve(abo, false);
 	if (r)
 		return r;
@@ -127,35 +138,6 @@ int amdgpu_gem_object_open(struct drm_gem_object *obj,
 	return 0;
 }
 
-static int amdgpu_gem_vm_check(void *param, struct amdgpu_bo *bo)
-{
-	/* if anything is swapped out don't swap it in here,
-	   just abort and wait for the next CS */
-	if (!amdgpu_bo_gpu_accessible(bo))
-		return -ERESTARTSYS;
-
-	if (bo->shadow && !amdgpu_bo_gpu_accessible(bo->shadow))
-		return -ERESTARTSYS;
-
-	return 0;
-}
-
-static bool amdgpu_gem_vm_ready(struct amdgpu_device *adev,
-				struct amdgpu_vm *vm,
-				struct list_head *list)
-{
-	struct ttm_validate_buffer *entry;
-
-	list_for_each_entry(entry, list, head) {
-		struct amdgpu_bo *bo =
-			container_of(entry->bo, struct amdgpu_bo, tbo);
-		if (amdgpu_gem_vm_check(NULL, bo))
-			return false;
-	}
-
-	return !amdgpu_vm_validate_pt_bos(adev, vm, amdgpu_gem_vm_check, NULL);
-}
-
 void amdgpu_gem_object_close(struct drm_gem_object *obj,
 			     struct drm_file *file_priv)
 {
@@ -165,13 +147,14 @@ void amdgpu_gem_object_close(struct drm_gem_object *obj,
 	struct amdgpu_vm *vm = &fpriv->vm;
 
 	struct amdgpu_bo_list_entry vm_pd;
-	struct list_head list;
+	struct list_head list, duplicates;
 	struct ttm_validate_buffer tv;
 	struct ww_acquire_ctx ticket;
 	struct amdgpu_bo_va *bo_va;
 	int r;
 
 	INIT_LIST_HEAD(&list);
+	INIT_LIST_HEAD(&duplicates);
 
 	tv.bo = &bo->tbo;
 	tv.shared = true;
@@ -179,7 +162,7 @@ void amdgpu_gem_object_close(struct drm_gem_object *obj,
 
 	amdgpu_vm_get_pd_bo(vm, &list, &vm_pd);
 
-	r = ttm_eu_reserve_buffers(&ticket, &list, false, NULL);
+	r = ttm_eu_reserve_buffers(&ticket, &list, false, &duplicates);
 	if (r) {
 		dev_err(adev->dev, "leaking bo va because "
 			"we fail to reserve bo (%d)\n", r);
@@ -189,7 +172,7 @@ void amdgpu_gem_object_close(struct drm_gem_object *obj,
 	if (bo_va && --bo_va->ref_count == 0) {
 		amdgpu_vm_bo_rmv(adev, bo_va);
 
-		if (amdgpu_gem_vm_ready(adev, vm, &list)) {
+		if (amdgpu_vm_ready(vm)) {
 			struct dma_fence *fence = NULL;
 
 			r = amdgpu_vm_clear_freed(adev, vm, &fence);
@@ -214,18 +197,22 @@ int amdgpu_gem_create_ioctl(struct drm_device *dev, void *data,
 			    struct drm_file *filp)
 {
 	struct amdgpu_device *adev = dev->dev_private;
+	struct amdgpu_fpriv *fpriv = filp->driver_priv;
+	struct amdgpu_vm *vm = &fpriv->vm;
 	union drm_amdgpu_gem_create *args = data;
+	uint64_t flags = args->in.domain_flags;
 	uint64_t size = args->in.bo_size;
+	struct reservation_object *resv = NULL;
 	struct drm_gem_object *gobj;
 	uint32_t handle;
-	bool kernel = false;
 	int r;
 
 	/* reject invalid gem flags */
-	if (args->in.domain_flags & ~(AMDGPU_GEM_CREATE_CPU_ACCESS_REQUIRED |
-				      AMDGPU_GEM_CREATE_NO_CPU_ACCESS |
-				      AMDGPU_GEM_CREATE_CPU_GTT_USWC |
-				      AMDGPU_GEM_CREATE_VRAM_CLEARED))
+	if (flags & ~(AMDGPU_GEM_CREATE_CPU_ACCESS_REQUIRED |
+		      AMDGPU_GEM_CREATE_NO_CPU_ACCESS |
+		      AMDGPU_GEM_CREATE_CPU_GTT_USWC |
+		      AMDGPU_GEM_CREATE_VRAM_CLEARED |
+		      AMDGPU_GEM_CREATE_VM_ALWAYS_VALID))
 		return -EINVAL;
 
 	/* reject invalid gem domains */
@@ -240,7 +227,7 @@ int amdgpu_gem_create_ioctl(struct drm_device *dev, void *data,
 	/* create a gem object to contain this object in */
 	if (args->in.domains & (AMDGPU_GEM_DOMAIN_GDS |
 	    AMDGPU_GEM_DOMAIN_GWS | AMDGPU_GEM_DOMAIN_OA)) {
-		kernel = true;
+		flags |= AMDGPU_GEM_CREATE_NO_CPU_ACCESS;
 		if (args->in.domains == AMDGPU_GEM_DOMAIN_GDS)
 			size = size << AMDGPU_GDS_SHIFT;
 		else if (args->in.domains == AMDGPU_GEM_DOMAIN_GWS)
@@ -252,10 +239,25 @@ int amdgpu_gem_create_ioctl(struct drm_device *dev, void *data,
 	}
 	size = roundup(size, PAGE_SIZE);
 
+	if (flags & AMDGPU_GEM_CREATE_VM_ALWAYS_VALID) {
+		r = amdgpu_bo_reserve(vm->root.base.bo, false);
+		if (r)
+			return r;
+
+		resv = vm->root.base.bo->tbo.resv;
+	}
+
 	r = amdgpu_gem_object_create(adev, size, args->in.alignment,
 				     (u32)(0xffffffff & args->in.domains),
-				     args->in.domain_flags,
-				     kernel, &gobj);
+				     flags, false, resv, &gobj);
+	if (flags & AMDGPU_GEM_CREATE_VM_ALWAYS_VALID) {
+		if (!r) {
+			struct amdgpu_bo *abo = gem_to_amdgpu_bo(gobj);
+
+			abo->parent = amdgpu_bo_ref(vm->root.base.bo);
+		}
+		amdgpu_bo_unreserve(vm->root.base.bo);
+	}
 	if (r)
 		return r;
 
@@ -297,9 +299,8 @@ int amdgpu_gem_userptr_ioctl(struct drm_device *dev, void *data,
 	}
 
 	/* create a gem object to contain this object in */
-	r = amdgpu_gem_object_create(adev, args->size, 0,
-				     AMDGPU_GEM_DOMAIN_CPU, 0,
-				     0, &gobj);
+	r = amdgpu_gem_object_create(adev, args->size, 0, AMDGPU_GEM_DOMAIN_CPU,
+				     0, 0, NULL, &gobj);
 	if (r)
 		return r;
 
@@ -317,8 +318,6 @@ int amdgpu_gem_userptr_ioctl(struct drm_device *dev, void *data,
 	}
 
 	if (args->flags & AMDGPU_GEM_USERPTR_VALIDATE) {
-		down_read(&current->mm->mmap_sem);
-
 		r = amdgpu_ttm_tt_get_user_pages(bo->tbo.ttm,
 						 bo->tbo.ttm->pages);
 		if (r)
@@ -333,8 +332,6 @@ int amdgpu_gem_userptr_ioctl(struct drm_device *dev, void *data,
 		amdgpu_bo_unreserve(bo);
 		if (r)
 			goto free_pages;
-
-		up_read(&current->mm->mmap_sem);
 	}
 
 	r = drm_gem_handle_create(filp, gobj, &handle);
@@ -511,10 +508,10 @@ static void amdgpu_gem_va_update_vm(struct amdgpu_device *adev,
 				    struct list_head *list,
 				    uint32_t operation)
 {
-	int r = -ERESTARTSYS;
+	int r;
 
-	if (!amdgpu_gem_vm_ready(adev, vm, list))
-		goto error;
+	if (!amdgpu_vm_ready(vm))
+		return;
 
 	r = amdgpu_vm_update_directories(adev, vm);
 	if (r)
@@ -551,7 +548,7 @@ int amdgpu_gem_va_ioctl(struct drm_device *dev, void *data,
 	struct amdgpu_bo_list_entry vm_pd;
 	struct ttm_validate_buffer tv;
 	struct ww_acquire_ctx ticket;
-	struct list_head list;
+	struct list_head list, duplicates;
 	uint64_t va_flags;
 	int r = 0;
 
@@ -587,6 +584,7 @@ int amdgpu_gem_va_ioctl(struct drm_device *dev, void *data,
 	}
 
 	INIT_LIST_HEAD(&list);
+	INIT_LIST_HEAD(&duplicates);
 	if ((args->operation != AMDGPU_VA_OP_CLEAR) &&
 	    !(args->flags & AMDGPU_VM_PAGE_PRT)) {
 		gobj = drm_gem_object_lookup(filp, args->handle);
@@ -603,7 +601,7 @@ int amdgpu_gem_va_ioctl(struct drm_device *dev, void *data,
 
 	amdgpu_vm_get_pd_bo(&fpriv->vm, &list, &vm_pd);
 
-	r = ttm_eu_reserve_buffers(&ticket, &list, true, NULL);
+	r = ttm_eu_reserve_buffers(&ticket, &list, true, &duplicates);
 	if (r)
 		goto error_unref;
 
@@ -669,6 +667,7 @@ int amdgpu_gem_va_ioctl(struct drm_device *dev, void *data,
 int amdgpu_gem_op_ioctl(struct drm_device *dev, void *data,
 			struct drm_file *filp)
 {
+	struct amdgpu_device *adev = dev->dev_private;
 	struct drm_amdgpu_gem_op *args = data;
 	struct drm_gem_object *gobj;
 	struct amdgpu_bo *robj;
@@ -716,6 +715,9 @@ int amdgpu_gem_op_ioctl(struct drm_device *dev, void *data,
 		if (robj->allowed_domains == AMDGPU_GEM_DOMAIN_VRAM)
 			robj->allowed_domains |= AMDGPU_GEM_DOMAIN_GTT;
 
+		if (robj->flags & AMDGPU_GEM_CREATE_VM_ALWAYS_VALID)
+			amdgpu_vm_bo_invalidate(adev, robj, true);
+
 		amdgpu_bo_unreserve(robj);
 		break;
 	default:
@@ -745,8 +747,7 @@ int amdgpu_mode_dumb_create(struct drm_file *file_priv,
 	r = amdgpu_gem_object_create(adev, args->size, 0,
 				     AMDGPU_GEM_DOMAIN_VRAM,
 				     AMDGPU_GEM_CREATE_CPU_ACCESS_REQUIRED,
-				     ttm_bo_type_device,
-				     &gobj);
+				     false, NULL, &gobj);
 	if (r)
 		return -ENOMEM;
 

drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c

@@ -260,8 +260,13 @@ int amdgpu_gfx_compute_mqd_sw_init(struct amdgpu_device *adev,
 	/* create MQD for KIQ */
 	ring = &adev->gfx.kiq.ring;
 	if (!ring->mqd_obj) {
+		/* originaly the KIQ MQD is put in GTT domain, but for SRIOV VRAM domain is a must
+		 * otherwise hypervisor trigger SAVE_VF fail after driver unloaded which mean MQD
+		 * deallocated and gart_unbind, to strict diverage we decide to use VRAM domain for
+		 * KIQ MQD no matter SRIOV or Bare-metal
+		 */
 		r = amdgpu_bo_create_kernel(adev, mqd_size, PAGE_SIZE,
-					    AMDGPU_GEM_DOMAIN_GTT, &ring->mqd_obj,
+					    AMDGPU_GEM_DOMAIN_VRAM, &ring->mqd_obj,
 					    &ring->mqd_gpu_addr, &ring->mqd_ptr);
 		if (r) {
 			dev_warn(adev->dev, "failed to create ring mqd ob (%d)", r);

drivers/gpu/drm/amd/amdgpu/amdgpu_ih.c

@@ -169,6 +169,12 @@ int amdgpu_ih_process(struct amdgpu_device *adev)
 	while (adev->irq.ih.rptr != wptr) {
 		u32 ring_index = adev->irq.ih.rptr >> 2;
 
+		/* Prescreening of high-frequency interrupts */
+		if (!amdgpu_ih_prescreen_iv(adev)) {
+			adev->irq.ih.rptr &= adev->irq.ih.ptr_mask;
+			continue;
+		}
+
 		/* Before dispatching irq to IP blocks, send it to amdkfd */
 		amdgpu_amdkfd_interrupt(adev,
 				(const void *) &adev->irq.ih.ring[ring_index]);
@@ -190,3 +196,79 @@ int amdgpu_ih_process(struct amdgpu_device *adev)
 
 	return IRQ_HANDLED;
 }
+
+/**
+ * amdgpu_ih_add_fault - Add a page fault record
+ *
+ * @adev: amdgpu device pointer
+ * @key: 64-bit encoding of PASID and address
+ *
+ * This should be called when a retry page fault interrupt is
+ * received. If this is a new page fault, it will be added to a hash
+ * table. The return value indicates whether this is a new fault, or
+ * a fault that was already known and is already being handled.
+ *
+ * If there are too many pending page faults, this will fail. Retry
+ * interrupts should be ignored in this case until there is enough
+ * free space.
+ *
+ * Returns 0 if the fault was added, 1 if the fault was already known,
+ * -ENOSPC if there are too many pending faults.
+ */
+int amdgpu_ih_add_fault(struct amdgpu_device *adev, u64 key)
+{
+	unsigned long flags;
+	int r = -ENOSPC;
+
+	if (WARN_ON_ONCE(!adev->irq.ih.faults))
+		/* Should be allocated in <IP>_ih_sw_init on GPUs that
+		 * support retry faults and require retry filtering.
+		 */
+		return r;
+
+	spin_lock_irqsave(&adev->irq.ih.faults->lock, flags);
+
+	/* Only let the hash table fill up to 50% for best performance */
+	if (adev->irq.ih.faults->count >= (1 << (AMDGPU_PAGEFAULT_HASH_BITS-1)))
+		goto unlock_out;
+
+	r = chash_table_copy_in(&adev->irq.ih.faults->hash, key, NULL);
+	if (!r)
+		adev->irq.ih.faults->count++;
+
+	/* chash_table_copy_in should never fail unless we're losing count */
+	WARN_ON_ONCE(r < 0);
+
+unlock_out:
+	spin_unlock_irqrestore(&adev->irq.ih.faults->lock, flags);
+	return r;
+}
+
+/**
+ * amdgpu_ih_clear_fault - Remove a page fault record
+ *
+ * @adev: amdgpu device pointer
+ * @key: 64-bit encoding of PASID and address
+ *
+ * This should be called when a page fault has been handled. Any
+ * future interrupt with this key will be processed as a new
+ * page fault.
+ */
+void amdgpu_ih_clear_fault(struct amdgpu_device *adev, u64 key)
+{
+	unsigned long flags;
+	int r;
+
+	if (!adev->irq.ih.faults)
+		return;
+
+	spin_lock_irqsave(&adev->irq.ih.faults->lock, flags);
+
+	r = chash_table_remove(&adev->irq.ih.faults->hash, key, NULL);
+	if (!WARN_ON_ONCE(r < 0)) {
+		adev->irq.ih.faults->count--;
+		WARN_ON_ONCE(adev->irq.ih.faults->count < 0);
+	}
+
+	spin_unlock_irqrestore(&adev->irq.ih.faults->lock, flags);
+}

drivers/gpu/drm/amd/amdgpu/amdgpu_ih.h

@@ -24,6 +24,8 @@
 #ifndef __AMDGPU_IH_H__
 #define __AMDGPU_IH_H__
 
+#include <linux/chash.h>
+
 struct amdgpu_device;
  /*
   * vega10+ IH clients
@@ -69,6 +71,13 @@ enum amdgpu_ih_clientid
 
 #define AMDGPU_IH_CLIENTID_LEGACY 0
 
+#define AMDGPU_PAGEFAULT_HASH_BITS 8
+struct amdgpu_retryfault_hashtable {
+	DECLARE_CHASH_TABLE(hash, AMDGPU_PAGEFAULT_HASH_BITS, 8, 0);
+	spinlock_t	lock;
+	int		count;
+};
+
 /*
  * R6xx+ IH ring
  */
@@ -87,6 +96,7 @@ struct amdgpu_ih_ring {
 	bool			use_doorbell;
 	bool			use_bus_addr;
 	dma_addr_t		rb_dma_addr; /* only used when use_bus_addr = true */
+	struct amdgpu_retryfault_hashtable *faults;
 };
 
 #define AMDGPU_IH_SRC_DATA_MAX_SIZE_DW 4
@@ -109,5 +119,7 @@ int amdgpu_ih_ring_init(struct amdgpu_device *adev, unsigned ring_size,
 			bool use_bus_addr);
 void amdgpu_ih_ring_fini(struct amdgpu_device *adev);
 int amdgpu_ih_process(struct amdgpu_device *adev);
+int amdgpu_ih_add_fault(struct amdgpu_device *adev, u64 key);
+void amdgpu_ih_clear_fault(struct amdgpu_device *adev, u64 key);
 
 #endif

drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c

@@ -825,7 +825,7 @@ int amdgpu_driver_open_kms(struct drm_device *dev, struct drm_file *file_priv)
 	}
 
 	r = amdgpu_vm_init(adev, &fpriv->vm,
-			   AMDGPU_VM_CONTEXT_GFX);
+			   AMDGPU_VM_CONTEXT_GFX, 0);
 	if (r) {
 		kfree(fpriv);
 		goto out_suspend;
@@ -841,8 +841,11 @@ int amdgpu_driver_open_kms(struct drm_device *dev, struct drm_file *file_priv)
 
 	if (amdgpu_sriov_vf(adev)) {
 		r = amdgpu_map_static_csa(adev, &fpriv->vm, &fpriv->csa_va);
-		if (r)
+		if (r) {
+			amdgpu_vm_fini(adev, &fpriv->vm);
+			kfree(fpriv);
 			goto out_suspend;
+		}
 	}
 
 	mutex_init(&fpriv->bo_list_lock);

drivers/gpu/drm/amd/amdgpu/amdgpu_mn.c

@@ -50,8 +50,10 @@ struct amdgpu_mn {
 	struct hlist_node	node;
 
 	/* objects protected by lock */
-	struct mutex		lock;
+	struct rw_semaphore	lock;
 	struct rb_root_cached	objects;
+	struct mutex		read_lock;
+	atomic_t		recursion;
 };
 
 struct amdgpu_mn_node {
@@ -74,7 +76,7 @@ static void amdgpu_mn_destroy(struct work_struct *work)
 	struct amdgpu_bo *bo, *next_bo;
 
 	mutex_lock(&adev->mn_lock);
-	mutex_lock(&rmn->lock);
+	down_write(&rmn->lock);
 	hash_del(&rmn->node);
 	rbtree_postorder_for_each_entry_safe(node, next_node,
 					     &rmn->objects.rb_root, it.rb) {
@@ -84,7 +86,7 @@ static void amdgpu_mn_destroy(struct work_struct *work)
 		}
 		kfree(node);
 	}
-	mutex_unlock(&rmn->lock);
+	up_write(&rmn->lock);
 	mutex_unlock(&adev->mn_lock);
 	mmu_notifier_unregister_no_release(&rmn->mn, rmn->mm);
 	kfree(rmn);
@@ -106,6 +108,53 @@ static void amdgpu_mn_release(struct mmu_notifier *mn,
 	schedule_work(&rmn->work);
 }
 
+
+/**
+ * amdgpu_mn_lock - take the write side lock for this mn
+ */
+void amdgpu_mn_lock(struct amdgpu_mn *mn)
+{
+	if (mn)
+		down_write(&mn->lock);
+}
+
+/**
+ * amdgpu_mn_unlock - drop the write side lock for this mn
+ */
+void amdgpu_mn_unlock(struct amdgpu_mn *mn)
+{
+	if (mn)
+		up_write(&mn->lock);
+}
+
+/**
+ * amdgpu_mn_read_lock - take the rmn read lock
+ *
+ * @rmn: our notifier
+ *
+ * Take the rmn read side lock.
+ */
+static void amdgpu_mn_read_lock(struct amdgpu_mn *rmn)
+{
+	mutex_lock(&rmn->read_lock);
+	if (atomic_inc_return(&rmn->recursion) == 1)
+		down_read_non_owner(&rmn->lock);
+	mutex_unlock(&rmn->read_lock);
+}
+
+/**
+ * amdgpu_mn_read_unlock - drop the rmn read lock
+ *
+ * @rmn: our notifier
+ *
+ * Drop the rmn read side lock.
+ */
+static void amdgpu_mn_read_unlock(struct amdgpu_mn *rmn)
+{
+	if (atomic_dec_return(&rmn->recursion) == 0)
+		up_read_non_owner(&rmn->lock);
+}
+
 /**
  * amdgpu_mn_invalidate_node - unmap all BOs of a node
  *
@@ -126,23 +175,12 @@ static void amdgpu_mn_invalidate_node(struct amdgpu_mn_node *node,
 		if (!amdgpu_ttm_tt_affect_userptr(bo->tbo.ttm, start, end))
 			continue;
 
-		r = amdgpu_bo_reserve(bo, true);
-		if (r) {
-			DRM_ERROR("(%ld) failed to reserve user bo\n", r);
-			continue;
-		}
-
 		r = reservation_object_wait_timeout_rcu(bo->tbo.resv,
 			true, false, MAX_SCHEDULE_TIMEOUT);
 		if (r <= 0)
 			DRM_ERROR("(%ld) failed to wait for user bo\n", r);
 
-		amdgpu_ttm_placement_from_domain(bo, AMDGPU_GEM_DOMAIN_CPU);
-		r = ttm_bo_validate(&bo->tbo, &bo->placement, false, false);
-		if (r)
-			DRM_ERROR("(%ld) failed to validate user bo\n", r);
-
-		amdgpu_bo_unreserve(bo);
+		amdgpu_ttm_tt_mark_user_pages(bo->tbo.ttm);
 	}
 }
 
@@ -168,7 +206,7 @@ static void amdgpu_mn_invalidate_range_start(struct mmu_notifier *mn,
 	/* notification is exclusive, but interval is inclusive */
 	end -= 1;
 
-	mutex_lock(&rmn->lock);
+	amdgpu_mn_read_lock(rmn);
 
 	it = interval_tree_iter_first(&rmn->objects, start, end);
 	while (it) {
@@ -179,13 +217,32 @@ static void amdgpu_mn_invalidate_range_start(struct mmu_notifier *mn,
 
 		amdgpu_mn_invalidate_node(node, start, end);
 	}
+}
 
-	mutex_unlock(&rmn->lock);
+/**
+ * amdgpu_mn_invalidate_range_end - callback to notify about mm change
+ *
+ * @mn: our notifier
+ * @mn: the mm this callback is about
+ * @start: start of updated range
+ * @end: end of updated range
+ *
+ * Release the lock again to allow new command submissions.
+ */
+static void amdgpu_mn_invalidate_range_end(struct mmu_notifier *mn,
+					   struct mm_struct *mm,
+					   unsigned long start,
+					   unsigned long end)
+{
+	struct amdgpu_mn *rmn = container_of(mn, struct amdgpu_mn, mn);
+
+	amdgpu_mn_read_unlock(rmn);
 }
 
 static const struct mmu_notifier_ops amdgpu_mn_ops = {
 	.release = amdgpu_mn_release,
 	.invalidate_range_start = amdgpu_mn_invalidate_range_start,
+	.invalidate_range_end = amdgpu_mn_invalidate_range_end,
 };
 
 /**
@@ -195,7 +252,7 @@ static const struct mmu_notifier_ops amdgpu_mn_ops = {
  *
  * Creates a notifier context for current->mm.
  */
-static struct amdgpu_mn *amdgpu_mn_get(struct amdgpu_device *adev)
+struct amdgpu_mn *amdgpu_mn_get(struct amdgpu_device *adev)
 {
 	struct mm_struct *mm = current->mm;
 	struct amdgpu_mn *rmn;
@@ -220,8 +277,10 @@ static struct amdgpu_mn *amdgpu_mn_get(struct amdgpu_device *adev)
 	rmn->adev = adev;
 	rmn->mm = mm;
 	rmn->mn.ops = &amdgpu_mn_ops;
-	mutex_init(&rmn->lock);
+	init_rwsem(&rmn->lock);
 	rmn->objects = RB_ROOT_CACHED;
+	mutex_init(&rmn->read_lock);
+	atomic_set(&rmn->recursion, 0);
 
 	r = __mmu_notifier_register(&rmn->mn, mm);
 	if (r)
@@ -267,7 +326,7 @@ int amdgpu_mn_register(struct amdgpu_bo *bo, unsigned long addr)
 
 	INIT_LIST_HEAD(&bos);
 
-	mutex_lock(&rmn->lock);
+	down_write(&rmn->lock);
 
 	while ((it = interval_tree_iter_first(&rmn->objects, addr, end))) {
 		kfree(node);
@@ -281,7 +340,7 @@ int amdgpu_mn_register(struct amdgpu_bo *bo, unsigned long addr)
 	if (!node) {
 		node = kmalloc(sizeof(struct amdgpu_mn_node), GFP_KERNEL);
 		if (!node) {
-			mutex_unlock(&rmn->lock);
+			up_write(&rmn->lock);
 			return -ENOMEM;
 		}
 	}
@@ -296,7 +355,7 @@ int amdgpu_mn_register(struct amdgpu_bo *bo, unsigned long addr)
 
 	interval_tree_insert(&node->it, &rmn->objects);
 
-	mutex_unlock(&rmn->lock);
+	up_write(&rmn->lock);
 
 	return 0;
 }
@@ -322,7 +381,7 @@ void amdgpu_mn_unregister(struct amdgpu_bo *bo)
 		return;
 	}
 
-	mutex_lock(&rmn->lock);
+	down_write(&rmn->lock);
 
 	/* save the next list entry for later */
 	head = bo->mn_list.next;
@@ -337,6 +396,7 @@ void amdgpu_mn_unregister(struct amdgpu_bo *bo)
 		kfree(node);
 	}
 
-	mutex_unlock(&rmn->lock);
+	up_write(&rmn->lock);
 	mutex_unlock(&adev->mn_lock);
 }
+

drivers/gpu/drm/amd/powerplay/eventmgr/psm.h → drivers/gpu/drm/amd/amdgpu/amdgpu_mn.h

 /*
- * Copyright 2015 Advanced Micro Devices, Inc.
+ * Copyright 2017 Advanced Micro Devices, Inc.
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
@@ -19,20 +19,34 @@
  * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
  * OTHER DEALINGS IN THE SOFTWARE.
  *
+ * Authors: Christian König
  */
-#include "eventmgr.h"
-#include "eventinit.h"
-#include "eventmanagement.h"
-#include "eventmanager.h"
-#include "power_state.h"
-#include "hardwaremanager.h"
+#ifndef __AMDGPU_MN_H__
+#define __AMDGPU_MN_H__
 
-int psm_get_ui_state(struct pp_eventmgr *eventmgr, enum PP_StateUILabel ui_label, unsigned long *state_id);
-
-int psm_get_state_by_classification(struct pp_eventmgr *eventmgr, enum PP_StateClassificationFlag flag, unsigned long *state_id);
-
-int psm_set_states(struct pp_eventmgr *eventmgr, unsigned long *state_id);
+/*
+ * MMU Notifier
+ */
+struct amdgpu_mn;
 
-int psm_adjust_power_state_dynamic(struct pp_eventmgr *eventmgr, bool skip);
+#if defined(CONFIG_MMU_NOTIFIER)
+void amdgpu_mn_lock(struct amdgpu_mn *mn);
+void amdgpu_mn_unlock(struct amdgpu_mn *mn);
+struct amdgpu_mn *amdgpu_mn_get(struct amdgpu_device *adev);
+int amdgpu_mn_register(struct amdgpu_bo *bo, unsigned long addr);
+void amdgpu_mn_unregister(struct amdgpu_bo *bo);
+#else
+static inline void amdgpu_mn_lock(struct amdgpu_mn *mn) {}
+static inline void amdgpu_mn_unlock(struct amdgpu_mn *mn) {}
+static inline struct amdgpu_mn *amdgpu_mn_get(struct amdgpu_device *adev)
+{
+	return NULL;
+}
+static inline int amdgpu_mn_register(struct amdgpu_bo *bo, unsigned long addr)
+{
+	return -ENODEV;
+}
+static inline void amdgpu_mn_unregister(struct amdgpu_bo *bo) {}
+#endif
 
-int psm_adjust_power_state_static(struct pp_eventmgr *eventmgr, bool skip);
+#endif

drivers/gpu/drm/amd/amdgpu/amdgpu_object.c

@@ -64,11 +64,12 @@ bool amdgpu_ttm_bo_is_amdgpu_bo(struct ttm_buffer_object *bo)
 	return false;
 }
 
-static void amdgpu_ttm_placement_init(struct amdgpu_device *adev,
-				      struct ttm_placement *placement,
-				      struct ttm_place *places,
-				      u32 domain, u64 flags)
+void amdgpu_ttm_placement_from_domain(struct amdgpu_bo *abo, u32 domain)
 {
+	struct amdgpu_device *adev = amdgpu_ttm_adev(abo->tbo.bdev);
+	struct ttm_placement *placement = &abo->placement;
+	struct ttm_place *places = abo->placements;
+	u64 flags = abo->flags;
 	u32 c = 0;
 
 	if (domain & AMDGPU_GEM_DOMAIN_VRAM) {
@@ -151,27 +152,6 @@ static void amdgpu_ttm_placement_init(struct amdgpu_device *adev,
 	placement->busy_placement = places;
 }
 
-void amdgpu_ttm_placement_from_domain(struct amdgpu_bo *abo, u32 domain)
-{
-	struct amdgpu_device *adev = amdgpu_ttm_adev(abo->tbo.bdev);
-
-	amdgpu_ttm_placement_init(adev, &abo->placement, abo->placements,
-				  domain, abo->flags);
-}
-
-static void amdgpu_fill_placement_to_bo(struct amdgpu_bo *bo,
-					struct ttm_placement *placement)
-{
-	BUG_ON(placement->num_placement > (AMDGPU_GEM_DOMAIN_MAX + 1));
-
-	memcpy(bo->placements, placement->placement,
-	       placement->num_placement * sizeof(struct ttm_place));
-	bo->placement.num_placement = placement->num_placement;
-	bo->placement.num_busy_placement = placement->num_busy_placement;
-	bo->placement.placement = bo->placements;
-	bo->placement.busy_placement = bo->placements;
-}
-
 /**
  * amdgpu_bo_create_reserved - create reserved BO for kernel use
  *
@@ -303,14 +283,13 @@ void amdgpu_bo_free_kernel(struct amdgpu_bo **bo, u64 *gpu_addr,
 		*cpu_addr = NULL;
 }
 
-int amdgpu_bo_create_restricted(struct amdgpu_device *adev,
-				unsigned long size, int byte_align,
-				bool kernel, u32 domain, u64 flags,
-				struct sg_table *sg,
-				struct ttm_placement *placement,
-				struct reservation_object *resv,
-				uint64_t init_value,
-				struct amdgpu_bo **bo_ptr)
+static int amdgpu_bo_do_create(struct amdgpu_device *adev,
+			       unsigned long size, int byte_align,
+			       bool kernel, u32 domain, u64 flags,
+			       struct sg_table *sg,
+			       struct reservation_object *resv,
+			       uint64_t init_value,
+			       struct amdgpu_bo **bo_ptr)
 {
 	struct amdgpu_bo *bo;
 	enum ttm_bo_type type;
@@ -384,10 +363,11 @@ int amdgpu_bo_create_restricted(struct amdgpu_device *adev,
 		bo->flags &= ~AMDGPU_GEM_CREATE_CPU_GTT_USWC;
 #endif
 
-	amdgpu_fill_placement_to_bo(bo, placement);
-	/* Kernel allocation are uninterruptible */
+	bo->tbo.bdev = &adev->mman.bdev;
+	amdgpu_ttm_placement_from_domain(bo, domain);
 
 	initial_bytes_moved = atomic64_read(&adev->num_bytes_moved);
+	/* Kernel allocation are uninterruptible */
 	r = ttm_bo_init_reserved(&adev->mman.bdev, &bo->tbo, size, type,
 				 &bo->placement, page_align, !kernel, NULL,
 				 acc_size, sg, resv, &amdgpu_ttm_bo_destroy);
@@ -442,27 +422,17 @@ static int amdgpu_bo_create_shadow(struct amdgpu_device *adev,
 				   unsigned long size, int byte_align,
 				   struct amdgpu_bo *bo)
 {
-	struct ttm_placement placement = {0};
-	struct ttm_place placements[AMDGPU_GEM_DOMAIN_MAX + 1];
 	int r;
 
 	if (bo->shadow)
 		return 0;
 
-	memset(&placements, 0, sizeof(placements));
-	amdgpu_ttm_placement_init(adev, &placement, placements,
-				  AMDGPU_GEM_DOMAIN_GTT,
-				  AMDGPU_GEM_CREATE_CPU_GTT_USWC |
-				  AMDGPU_GEM_CREATE_SHADOW);
-
-	r = amdgpu_bo_create_restricted(adev, size, byte_align, true,
-					AMDGPU_GEM_DOMAIN_GTT,
-					AMDGPU_GEM_CREATE_CPU_GTT_USWC |
-					AMDGPU_GEM_CREATE_SHADOW,
-					NULL, &placement,
-					bo->tbo.resv,
-					0,
-					&bo->shadow);
+	r = amdgpu_bo_do_create(adev, size, byte_align, true,
+				AMDGPU_GEM_DOMAIN_GTT,
+				AMDGPU_GEM_CREATE_CPU_GTT_USWC |
+				AMDGPU_GEM_CREATE_SHADOW,
+				NULL, bo->tbo.resv, 0,
+				&bo->shadow);
 	if (!r) {
 		bo->shadow->parent = amdgpu_bo_ref(bo);
 		mutex_lock(&adev->shadow_list_lock);
@@ -484,18 +454,11 @@ int amdgpu_bo_create(struct amdgpu_device *adev,
 		     uint64_t init_value,
 		     struct amdgpu_bo **bo_ptr)
 {
-	struct ttm_placement placement = {0};
-	struct ttm_place placements[AMDGPU_GEM_DOMAIN_MAX + 1];
 	uint64_t parent_flags = flags & ~AMDGPU_GEM_CREATE_SHADOW;
 	int r;
 
-	memset(&placements, 0, sizeof(placements));
-	amdgpu_ttm_placement_init(adev, &placement, placements,
-				  domain, parent_flags);
-
-	r = amdgpu_bo_create_restricted(adev, size, byte_align, kernel, domain,
-					parent_flags, sg, &placement, resv,
-					init_value, bo_ptr);
+	r = amdgpu_bo_do_create(adev, size, byte_align, kernel, domain,
+				parent_flags, sg, resv, init_value, bo_ptr);
 	if (r)
 		return r;
 
@@ -672,7 +635,6 @@ int amdgpu_bo_pin_restricted(struct amdgpu_bo *bo, u32 domain,
 {
 	struct amdgpu_device *adev = amdgpu_ttm_adev(bo->tbo.bdev);
 	int r, i;
-	unsigned fpfn, lpfn;
 
 	if (amdgpu_ttm_tt_get_usermm(bo->tbo.ttm))
 		return -EPERM;
@@ -704,22 +666,16 @@ int amdgpu_bo_pin_restricted(struct amdgpu_bo *bo, u32 domain,
 	}
 
 	bo->flags |= AMDGPU_GEM_CREATE_VRAM_CONTIGUOUS;
+	/* force to pin into visible video ram */
+	if (!(bo->flags & AMDGPU_GEM_CREATE_NO_CPU_ACCESS))
+		bo->flags |= AMDGPU_GEM_CREATE_CPU_ACCESS_REQUIRED;
 	amdgpu_ttm_placement_from_domain(bo, domain);
 	for (i = 0; i < bo->placement.num_placement; i++) {
-		/* force to pin into visible video ram */
-		if ((bo->placements[i].flags & TTM_PL_FLAG_VRAM) &&
-		    !(bo->flags & AMDGPU_GEM_CREATE_NO_CPU_ACCESS) &&
-		    (!max_offset || max_offset >
-		     adev->mc.visible_vram_size)) {
-			if (WARN_ON_ONCE(min_offset >
-					 adev->mc.visible_vram_size))
-				return -EINVAL;
-			fpfn = min_offset >> PAGE_SHIFT;
-			lpfn = adev->mc.visible_vram_size >> PAGE_SHIFT;
-		} else {
-			fpfn = min_offset >> PAGE_SHIFT;
-			lpfn = max_offset >> PAGE_SHIFT;
-		}
+		unsigned fpfn, lpfn;
+
+		fpfn = min_offset >> PAGE_SHIFT;
+		lpfn = max_offset >> PAGE_SHIFT;
+
 		if (fpfn > bo->placements[i].fpfn)
 			bo->placements[i].fpfn = fpfn;
 		if (!bo->placements[i].lpfn ||
@@ -929,7 +885,7 @@ void amdgpu_bo_move_notify(struct ttm_buffer_object *bo,
 		return;
 
 	abo = container_of(bo, struct amdgpu_bo, tbo);
-	amdgpu_vm_bo_invalidate(adev, abo);
+	amdgpu_vm_bo_invalidate(adev, abo, evict);
 
 	amdgpu_bo_kunmap(abo);
 

drivers/gpu/drm/amd/amdgpu/amdgpu_object.h

@@ -35,6 +35,7 @@
 
 /* bo virtual addresses in a vm */
 struct amdgpu_bo_va_mapping {
+	struct amdgpu_bo_va		*bo_va;
 	struct list_head		list;
 	struct rb_node			rb;
 	uint64_t			start;
@@ -49,12 +50,17 @@ struct amdgpu_bo_va {
 	struct amdgpu_vm_bo_base	base;
 
 	/* protected by bo being reserved */
-	struct dma_fence	        *last_pt_update;
 	unsigned			ref_count;
 
+	/* all other members protected by the VM PD being reserved */
+	struct dma_fence	        *last_pt_update;
+
 	/* mappings for this bo_va */
 	struct list_head		invalids;
 	struct list_head		valids;
+
+	/* If the mappings are cleared or filled */
+	bool				cleared;
 };
 
 struct amdgpu_bo {
@@ -189,14 +195,6 @@ int amdgpu_bo_create(struct amdgpu_device *adev,
 			    struct reservation_object *resv,
 			    uint64_t init_value,
 			    struct amdgpu_bo **bo_ptr);
-int amdgpu_bo_create_restricted(struct amdgpu_device *adev,
-				unsigned long size, int byte_align,
-				bool kernel, u32 domain, u64 flags,
-				struct sg_table *sg,
-				struct ttm_placement *placement,
-			        struct reservation_object *resv,
-				uint64_t init_value,
-				struct amdgpu_bo **bo_ptr);
 int amdgpu_bo_create_reserved(struct amdgpu_device *adev,
 			      unsigned long size, int align,
 			      u32 domain, struct amdgpu_bo **bo_ptr,

drivers/gpu/drm/amd/amdgpu/amdgpu_powerplay.c

@@ -87,17 +87,28 @@ static int amdgpu_pp_early_init(void *handle)
 	case CHIP_OLAND:
 	case CHIP_HAINAN:
 		amd_pp->ip_funcs = &si_dpm_ip_funcs;
+		amd_pp->pp_funcs = &si_dpm_funcs;
 	break;
 #endif
 #ifdef CONFIG_DRM_AMDGPU_CIK
 	case CHIP_BONAIRE:
 	case CHIP_HAWAII:
-		amd_pp->ip_funcs = &ci_dpm_ip_funcs;
+		if (amdgpu_dpm == -1) {
+			amd_pp->ip_funcs = &ci_dpm_ip_funcs;
+			amd_pp->pp_funcs = &ci_dpm_funcs;
+		} else {
+			adev->pp_enabled = true;
+			if (amdgpu_create_pp_handle(adev))
+				return -EINVAL;
+			amd_pp->ip_funcs = &pp_ip_funcs;
+			amd_pp->pp_funcs = &pp_dpm_funcs;
+		}
 		break;
 	case CHIP_KABINI:
 	case CHIP_MULLINS:
 	case CHIP_KAVERI:
 		amd_pp->ip_funcs = &kv_dpm_ip_funcs;
+		amd_pp->pp_funcs = &kv_dpm_funcs;
 		break;
 #endif
 	default:
@@ -128,7 +139,7 @@ static int amdgpu_pp_late_init(void *handle)
 
 	if (adev->pp_enabled && adev->pm.dpm_enabled) {
 		amdgpu_pm_sysfs_init(adev);
-		amdgpu_dpm_dispatch_task(adev, AMD_PP_EVENT_COMPLETE_INIT, NULL, NULL);
+		amdgpu_dpm_dispatch_task(adev, AMD_PP_TASK_COMPLETE_INIT, NULL, NULL);
 	}
 
 	return ret;

drivers/gpu/drm/amd/amdgpu/amdgpu_prime.c

@@ -57,6 +57,40 @@ void amdgpu_gem_prime_vunmap(struct drm_gem_object *obj, void *vaddr)
 	ttm_bo_kunmap(&bo->dma_buf_vmap);
 }
 
+int amdgpu_gem_prime_mmap(struct drm_gem_object *obj, struct vm_area_struct *vma)
+{
+	struct amdgpu_bo *bo = gem_to_amdgpu_bo(obj);
+	struct amdgpu_device *adev = amdgpu_ttm_adev(bo->tbo.bdev);
+	unsigned asize = amdgpu_bo_size(bo);
+	int ret;
+
+	if (!vma->vm_file)
+		return -ENODEV;
+
+	if (adev == NULL)
+		return -ENODEV;
+
+	/* Check for valid size. */
+	if (asize < vma->vm_end - vma->vm_start)
+		return -EINVAL;
+
+	if (amdgpu_ttm_tt_get_usermm(bo->tbo.ttm) ||
+	    (bo->flags & AMDGPU_GEM_CREATE_NO_CPU_ACCESS)) {
+		return -EPERM;
+	}
+	vma->vm_pgoff += amdgpu_bo_mmap_offset(bo) >> PAGE_SHIFT;
+
+	/* prime mmap does not need to check access, so allow here */
+	ret = drm_vma_node_allow(&obj->vma_node, vma->vm_file->private_data);
+	if (ret)
+		return ret;
+
+	ret = ttm_bo_mmap(vma->vm_file, vma, &adev->mman.bdev);
+	drm_vma_node_revoke(&obj->vma_node, vma->vm_file->private_data);
+
+	return ret;
+}
+
 struct drm_gem_object *
 amdgpu_gem_prime_import_sg_table(struct drm_device *dev,
 				 struct dma_buf_attachment *attach,
@@ -136,7 +170,8 @@ struct dma_buf *amdgpu_gem_prime_export(struct drm_device *dev,
 {
 	struct amdgpu_bo *bo = gem_to_amdgpu_bo(gobj);
 
-	if (amdgpu_ttm_tt_get_usermm(bo->tbo.ttm))
+	if (amdgpu_ttm_tt_get_usermm(bo->tbo.ttm) ||
+	    bo->flags & AMDGPU_GEM_CREATE_VM_ALWAYS_VALID)
 		return ERR_PTR(-EPERM);
 
 	return drm_gem_prime_export(dev, gobj, flags);

drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c

@@ -57,21 +57,23 @@ static int psp_sw_init(void *handle)
 		psp->prep_cmd_buf = psp_v3_1_prep_cmd_buf;
 		psp->ring_init = psp_v3_1_ring_init;
 		psp->ring_create = psp_v3_1_ring_create;
+		psp->ring_stop = psp_v3_1_ring_stop;
 		psp->ring_destroy = psp_v3_1_ring_destroy;
 		psp->cmd_submit = psp_v3_1_cmd_submit;
 		psp->compare_sram_data = psp_v3_1_compare_sram_data;
 		psp->smu_reload_quirk = psp_v3_1_smu_reload_quirk;
+		psp->mode1_reset = psp_v3_1_mode1_reset;
 		break;
 	case CHIP_RAVEN:
-#if 0
 		psp->init_microcode = psp_v10_0_init_microcode;
-#endif
 		psp->prep_cmd_buf = psp_v10_0_prep_cmd_buf;
 		psp->ring_init = psp_v10_0_ring_init;
 		psp->ring_create = psp_v10_0_ring_create;
+		psp->ring_stop = psp_v10_0_ring_stop;
 		psp->ring_destroy = psp_v10_0_ring_destroy;
 		psp->cmd_submit = psp_v10_0_cmd_submit;
 		psp->compare_sram_data = psp_v10_0_compare_sram_data;
+		psp->mode1_reset = psp_v10_0_mode1_reset;
 		break;
 	default:
 		return -EINVAL;
@@ -90,6 +92,12 @@ static int psp_sw_init(void *handle)
 
 static int psp_sw_fini(void *handle)
 {
+	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+
+	release_firmware(adev->psp.sos_fw);
+	adev->psp.sos_fw = NULL;
+	release_firmware(adev->psp.asd_fw);
+	adev->psp.asd_fw = NULL;
 	return 0;
 }
 
@@ -253,15 +261,18 @@ static int psp_asd_load(struct psp_context *psp)
 
 static int psp_hw_start(struct psp_context *psp)
 {
+	struct amdgpu_device *adev = psp->adev;
 	int ret;
 
-	ret = psp_bootloader_load_sysdrv(psp);
-	if (ret)
-		return ret;
+	if (!amdgpu_sriov_vf(adev) || !adev->in_sriov_reset) {
+		ret = psp_bootloader_load_sysdrv(psp);
+		if (ret)
+			return ret;
 
-	ret = psp_bootloader_load_sos(psp);
-	if (ret)
-		return ret;
+		ret = psp_bootloader_load_sos(psp);
+		if (ret)
+			return ret;
+	}
 
 	ret = psp_ring_create(psp, PSP_RING_TYPE__KM);
 	if (ret)
@@ -453,6 +464,16 @@ static int psp_hw_fini(void *handle)
 
 static int psp_suspend(void *handle)
 {
+	int ret;
+	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct psp_context *psp = &adev->psp;
+
+	ret = psp_ring_stop(psp, PSP_RING_TYPE__KM);
+	if (ret) {
+		DRM_ERROR("PSP ring stop failed\n");
+		return ret;
+	}
+
 	return 0;
 }
 
@@ -487,6 +508,22 @@ static int psp_resume(void *handle)
 	return ret;
 }
 
+static bool psp_check_reset(void* handle)
+{
+	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+
+	if (adev->flags & AMD_IS_APU)
+		return true;
+
+	return false;
+}
+
+static int psp_reset(void* handle)
+{
+	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	return psp_mode1_reset(&adev->psp);
+}
+
 static bool psp_check_fw_loading_status(struct amdgpu_device *adev,
 					enum AMDGPU_UCODE_ID ucode_type)
 {
@@ -530,8 +567,9 @@ const struct amd_ip_funcs psp_ip_funcs = {
 	.suspend = psp_suspend,
 	.resume = psp_resume,
 	.is_idle = NULL,
+	.check_soft_reset = psp_check_reset,
 	.wait_for_idle = NULL,
-	.soft_reset = NULL,
+	.soft_reset = psp_reset,
 	.set_clockgating_state = psp_set_clockgating_state,
 	.set_powergating_state = psp_set_powergating_state,
 };

drivers/gpu/drm/amd/amdgpu/amdgpu_psp.h

@@ -66,6 +66,8 @@ struct psp_context
 			    struct psp_gfx_cmd_resp *cmd);
 	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,
+			    enum psp_ring_type ring_type);
 	int (*ring_destroy)(struct psp_context *psp,
 			    enum psp_ring_type ring_type);
 	int (*cmd_submit)(struct psp_context *psp, struct amdgpu_firmware_info *ucode,
@@ -74,6 +76,7 @@ struct psp_context
 				  struct amdgpu_firmware_info *ucode,
 				  enum AMDGPU_UCODE_ID ucode_type);
 	bool (*smu_reload_quirk)(struct psp_context *psp);
+	int (*mode1_reset)(struct psp_context *psp);
 
 	/* fence buffer */
 	struct amdgpu_bo 		*fw_pri_bo;
@@ -123,6 +126,7 @@ struct amdgpu_psp_funcs {
 #define psp_prep_cmd_buf(ucode, type) (psp)->prep_cmd_buf((ucode), (type))
 #define psp_ring_init(psp, type) (psp)->ring_init((psp), (type))
 #define psp_ring_create(psp, type) (psp)->ring_create((psp), (type))
+#define psp_ring_stop(psp, type) (psp)->ring_stop((psp), (type))
 #define psp_ring_destroy(psp, type) ((psp)->ring_destroy((psp), (type)))
 #define psp_cmd_submit(psp, ucode, cmd_mc, fence_mc, index) \
 		(psp)->cmd_submit((psp), (ucode), (cmd_mc), (fence_mc), (index))
@@ -136,6 +140,8 @@ struct amdgpu_psp_funcs {
 		((psp)->bootloader_load_sos ? (psp)->bootloader_load_sos((psp)) : 0)
 #define psp_smu_reload_quirk(psp) \
 		((psp)->smu_reload_quirk ? (psp)->smu_reload_quirk((psp)) : false)
+#define psp_mode1_reset(psp) \
+		((psp)->mode1_reset ? (psp)->mode1_reset((psp)) : false)
 
 extern const struct amd_ip_funcs psp_ip_funcs;
 

drivers/gpu/drm/amd/amdgpu/amdgpu_trace.h

@@ -14,62 +14,6 @@
 #define AMDGPU_JOB_GET_TIMELINE_NAME(job) \
 	 job->base.s_fence->finished.ops->get_timeline_name(&job->base.s_fence->finished)
 
-TRACE_EVENT(amdgpu_ttm_tt_populate,
-	    TP_PROTO(struct amdgpu_device *adev, uint64_t dma_address, uint64_t phys_address),
-	    TP_ARGS(adev, dma_address, phys_address),
-	    TP_STRUCT__entry(
-				__field(uint16_t, domain)
-				__field(uint8_t, bus)
-				__field(uint8_t, slot)
-				__field(uint8_t, func)
-				__field(uint64_t, dma)
-				__field(uint64_t, phys)
-			    ),
-	    TP_fast_assign(
-			   __entry->domain = pci_domain_nr(adev->pdev->bus);
-			   __entry->bus = adev->pdev->bus->number;
-			   __entry->slot = PCI_SLOT(adev->pdev->devfn);
-			   __entry->func = PCI_FUNC(adev->pdev->devfn);
-			   __entry->dma = dma_address;
-			   __entry->phys = phys_address;
-			   ),
-	    TP_printk("%04x:%02x:%02x.%x: 0x%llx => 0x%llx",
-		      (unsigned)__entry->domain,
-		      (unsigned)__entry->bus,
-		      (unsigned)__entry->slot,
-		      (unsigned)__entry->func,
-		      (unsigned long long)__entry->dma,
-		      (unsigned long long)__entry->phys)
-);
-
-TRACE_EVENT(amdgpu_ttm_tt_unpopulate,
-	    TP_PROTO(struct amdgpu_device *adev, uint64_t dma_address, uint64_t phys_address),
-	    TP_ARGS(adev, dma_address, phys_address),
-	    TP_STRUCT__entry(
-				__field(uint16_t, domain)
-				__field(uint8_t, bus)
-				__field(uint8_t, slot)
-				__field(uint8_t, func)
-				__field(uint64_t, dma)
-				__field(uint64_t, phys)
-			    ),
-	    TP_fast_assign(
-			   __entry->domain = pci_domain_nr(adev->pdev->bus);
-			   __entry->bus = adev->pdev->bus->number;
-			   __entry->slot = PCI_SLOT(adev->pdev->devfn);
-			   __entry->func = PCI_FUNC(adev->pdev->devfn);
-			   __entry->dma = dma_address;
-			   __entry->phys = phys_address;
-			   ),
-	    TP_printk("%04x:%02x:%02x.%x: 0x%llx => 0x%llx",
-		      (unsigned)__entry->domain,
-		      (unsigned)__entry->bus,
-		      (unsigned)__entry->slot,
-		      (unsigned)__entry->func,
-		      (unsigned long long)__entry->dma,
-		      (unsigned long long)__entry->phys)
-);
-
 TRACE_EVENT(amdgpu_mm_rreg,
 	    TP_PROTO(unsigned did, uint32_t reg, uint32_t value),
 	    TP_ARGS(did, reg, value),
@@ -473,5 +417,5 @@ TRACE_EVENT(amdgpu_ttm_bo_move,
 
 /* This part must be outside protection */
 #undef TRACE_INCLUDE_PATH
-#define TRACE_INCLUDE_PATH .
+#define TRACE_INCLUDE_PATH ../../drivers/gpu/drm/amd/amdgpu
 #include <trace/define_trace.h>

drivers/gpu/drm/amd/amdgpu/amdgpu_trace_points.c

 /* Copyright Red Hat Inc 2010.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
  * Author : Dave Airlie <airlied@redhat.com>
  */
 #include <drm/drmP.h>

drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.h

@@ -24,6 +24,7 @@
 #ifndef __AMDGPU_TTM_H__
 #define __AMDGPU_TTM_H__
 
+#include "amdgpu.h"
 #include "gpu_scheduler.h"
 
 #define AMDGPU_PL_GDS		(TTM_PL_PRIV + 0)
@@ -45,8 +46,7 @@ struct amdgpu_mman {
 	bool				initialized;
 
 #if defined(CONFIG_DEBUG_FS)
-	struct dentry			*vram;
-	struct dentry			*gtt;
+	struct dentry			*debugfs_entries[8];
 #endif
 
 	/* buffer handling */
@@ -82,4 +82,20 @@ bool amdgpu_ttm_is_bound(struct ttm_tt *ttm);
 int amdgpu_ttm_bind(struct ttm_buffer_object *bo, struct ttm_mem_reg *bo_mem);
 int amdgpu_ttm_recover_gart(struct amdgpu_device *adev);
 
+int amdgpu_ttm_tt_get_user_pages(struct ttm_tt *ttm, struct page **pages);
+void amdgpu_ttm_tt_set_user_pages(struct ttm_tt *ttm, struct page **pages);
+void amdgpu_ttm_tt_mark_user_pages(struct ttm_tt *ttm);
+int amdgpu_ttm_tt_set_userptr(struct ttm_tt *ttm, uint64_t addr,
+				     uint32_t flags);
+bool amdgpu_ttm_tt_has_userptr(struct ttm_tt *ttm);
+struct mm_struct *amdgpu_ttm_tt_get_usermm(struct ttm_tt *ttm);
+bool amdgpu_ttm_tt_affect_userptr(struct ttm_tt *ttm, unsigned long start,
+				  unsigned long end);
+bool amdgpu_ttm_tt_userptr_invalidated(struct ttm_tt *ttm,
+				       int *last_invalidated);
+bool amdgpu_ttm_tt_userptr_needs_pages(struct ttm_tt *ttm);
+bool amdgpu_ttm_tt_is_readonly(struct ttm_tt *ttm);
+uint64_t amdgpu_ttm_tt_pte_flags(struct amdgpu_device *adev, struct ttm_tt *ttm,
+				 struct ttm_mem_reg *mem);
+
 #endif

drivers/gpu/drm/amd/amdgpu/amdgpu_ucode.c

@@ -270,12 +270,8 @@ amdgpu_ucode_get_load_type(struct amdgpu_device *adev, int load_type)
 		else
 			return AMDGPU_FW_LOAD_SMU;
 	case CHIP_VEGA10:
-		if (!load_type)
-			return AMDGPU_FW_LOAD_DIRECT;
-		else
-			return AMDGPU_FW_LOAD_PSP;
 	case CHIP_RAVEN:
-		if (load_type != 2)
+		if (!load_type)
 			return AMDGPU_FW_LOAD_DIRECT;
 		else
 			return AMDGPU_FW_LOAD_PSP;
@@ -364,8 +360,6 @@ static int amdgpu_ucode_patch_jt(struct amdgpu_firmware_info *ucode,
 int amdgpu_ucode_init_bo(struct amdgpu_device *adev)
 {
 	struct amdgpu_bo **bo = &adev->firmware.fw_buf;
-	uint64_t fw_mc_addr;
-	void *fw_buf_ptr = NULL;
 	uint64_t fw_offset = 0;
 	int i, err;
 	struct amdgpu_firmware_info *ucode = NULL;
@@ -376,37 +370,39 @@ int amdgpu_ucode_init_bo(struct amdgpu_device *adev)
 		return 0;
 	}
 
-	err = amdgpu_bo_create(adev, adev->firmware.fw_size, PAGE_SIZE, true,
-				amdgpu_sriov_vf(adev) ? AMDGPU_GEM_DOMAIN_VRAM : AMDGPU_GEM_DOMAIN_GTT,
-				AMDGPU_GEM_CREATE_VRAM_CONTIGUOUS,
-				NULL, NULL, 0, bo);
-	if (err) {
-		dev_err(adev->dev, "(%d) Firmware buffer allocate failed\n", err);
-		goto failed;
-	}
+	if (!amdgpu_sriov_vf(adev) || !adev->in_sriov_reset) {
+		err = amdgpu_bo_create(adev, adev->firmware.fw_size, PAGE_SIZE, true,
+					amdgpu_sriov_vf(adev) ? AMDGPU_GEM_DOMAIN_VRAM : AMDGPU_GEM_DOMAIN_GTT,
+					AMDGPU_GEM_CREATE_VRAM_CONTIGUOUS,
+					NULL, NULL, 0, bo);
+		if (err) {
+			dev_err(adev->dev, "(%d) Firmware buffer allocate failed\n", err);
+			goto failed;
+		}
 
-	err = amdgpu_bo_reserve(*bo, false);
-	if (err) {
-		dev_err(adev->dev, "(%d) Firmware buffer reserve failed\n", err);
-		goto failed_reserve;
-	}
+		err = amdgpu_bo_reserve(*bo, false);
+		if (err) {
+			dev_err(adev->dev, "(%d) Firmware buffer reserve failed\n", err);
+			goto failed_reserve;
+		}
 
-	err = amdgpu_bo_pin(*bo, amdgpu_sriov_vf(adev) ? AMDGPU_GEM_DOMAIN_VRAM : AMDGPU_GEM_DOMAIN_GTT,
-				&fw_mc_addr);
-	if (err) {
-		dev_err(adev->dev, "(%d) Firmware buffer pin failed\n", err);
-		goto failed_pin;
-	}
+		err = amdgpu_bo_pin(*bo, amdgpu_sriov_vf(adev) ? AMDGPU_GEM_DOMAIN_VRAM : AMDGPU_GEM_DOMAIN_GTT,
+					&adev->firmware.fw_buf_mc);
+		if (err) {
+			dev_err(adev->dev, "(%d) Firmware buffer pin failed\n", err);
+			goto failed_pin;
+		}
 
-	err = amdgpu_bo_kmap(*bo, &fw_buf_ptr);
-	if (err) {
-		dev_err(adev->dev, "(%d) Firmware buffer kmap failed\n", err);
-		goto failed_kmap;
-	}
+		err = amdgpu_bo_kmap(*bo, &adev->firmware.fw_buf_ptr);
+		if (err) {
+			dev_err(adev->dev, "(%d) Firmware buffer kmap failed\n", err);
+			goto failed_kmap;
+		}
 
-	amdgpu_bo_unreserve(*bo);
+		amdgpu_bo_unreserve(*bo);
+	}
 
-	memset(fw_buf_ptr, 0, adev->firmware.fw_size);
+	memset(adev->firmware.fw_buf_ptr, 0, adev->firmware.fw_size);
 
 	/*
 	 * if SMU loaded firmware, it needn't add SMC, UVD, and VCE
@@ -425,14 +421,14 @@ int amdgpu_ucode_init_bo(struct amdgpu_device *adev)
 		ucode = &adev->firmware.ucode[i];
 		if (ucode->fw) {
 			header = (const struct common_firmware_header *)ucode->fw->data;
-			amdgpu_ucode_init_single_fw(adev, ucode, fw_mc_addr + fw_offset,
-						    (void *)((uint8_t *)fw_buf_ptr + fw_offset));
+			amdgpu_ucode_init_single_fw(adev, ucode, adev->firmware.fw_buf_mc + fw_offset,
+						    adev->firmware.fw_buf_ptr + fw_offset);
 			if (i == AMDGPU_UCODE_ID_CP_MEC1 &&
 			    adev->firmware.load_type != AMDGPU_FW_LOAD_PSP) {
 				const struct gfx_firmware_header_v1_0 *cp_hdr;
 				cp_hdr = (const struct gfx_firmware_header_v1_0 *)ucode->fw->data;
-				amdgpu_ucode_patch_jt(ucode, fw_mc_addr + fw_offset,
-						    fw_buf_ptr + fw_offset);
+				amdgpu_ucode_patch_jt(ucode,  adev->firmware.fw_buf_mc + fw_offset,
+						    adev->firmware.fw_buf_ptr + fw_offset);
 				fw_offset += ALIGN(le32_to_cpu(cp_hdr->jt_size) << 2, PAGE_SIZE);
 			}
 			fw_offset += ALIGN(ucode->ucode_size, PAGE_SIZE);

drivers/gpu/drm/amd/amdgpu/amdgpu_uvd.c

@@ -269,6 +269,7 @@ int amdgpu_uvd_sw_init(struct amdgpu_device *adev)
 
 int amdgpu_uvd_sw_fini(struct amdgpu_device *adev)
 {
+	int i;
 	kfree(adev->uvd.saved_bo);
 
 	amd_sched_entity_fini(&adev->uvd.ring.sched, &adev->uvd.entity);
@@ -279,6 +280,9 @@ int amdgpu_uvd_sw_fini(struct amdgpu_device *adev)
 
 	amdgpu_ring_fini(&adev->uvd.ring);
 
+	for (i = 0; i < AMDGPU_MAX_UVD_ENC_RINGS; ++i)
+		amdgpu_ring_fini(&adev->uvd.ring_enc[i]);
+
 	release_firmware(adev->uvd.fw);
 
 	return 0;
@@ -410,10 +414,10 @@ static int amdgpu_uvd_cs_pass1(struct amdgpu_uvd_cs_ctx *ctx)
 	uint64_t addr = amdgpu_uvd_get_addr_from_ctx(ctx);
 	int r = 0;
 
-	mapping = amdgpu_cs_find_mapping(ctx->parser, addr, &bo);
-	if (mapping == NULL) {
+	r = amdgpu_cs_find_mapping(ctx->parser, addr, &bo, &mapping);
+	if (r) {
 		DRM_ERROR("Can't find BO for addr 0x%08Lx\n", addr);
-		return -EINVAL;
+		return r;
 	}
 
 	if (!ctx->parser->adev->uvd.address_64_bit) {
@@ -737,10 +741,10 @@ static int amdgpu_uvd_cs_pass2(struct amdgpu_uvd_cs_ctx *ctx)
 	uint64_t addr = amdgpu_uvd_get_addr_from_ctx(ctx);
 	int r;
 
-	mapping = amdgpu_cs_find_mapping(ctx->parser, addr, &bo);
-	if (mapping == NULL) {
+	r = amdgpu_cs_find_mapping(ctx->parser, addr, &bo, &mapping);
+	if (r) {
 		DRM_ERROR("Can't find BO for addr 0x%08Lx\n", addr);
-		return -EINVAL;
+		return r;
 	}
 
 	start = amdgpu_bo_gpu_offset(bo);
@@ -917,10 +921,6 @@ int amdgpu_uvd_ring_parse_cs(struct amdgpu_cs_parser *parser, uint32_t ib_idx)
 		return -EINVAL;
 	}
 
-	r = amdgpu_cs_sysvm_access_required(parser);
-	if (r)
-		return r;
-
 	ctx.parser = parser;
 	ctx.buf_sizes = buf_sizes;
 	ctx.ib_idx = ib_idx;

drivers/gpu/drm/amd/amdgpu/amdgpu_vce.c

@@ -559,6 +559,7 @@ static int amdgpu_vce_cs_reloc(struct amdgpu_cs_parser *p, uint32_t ib_idx,
 	struct amdgpu_bo_va_mapping *mapping;
 	struct amdgpu_bo *bo;
 	uint64_t addr;
+	int r;
 
 	if (index == 0xffffffff)
 		index = 0;
@@ -567,11 +568,11 @@ static int amdgpu_vce_cs_reloc(struct amdgpu_cs_parser *p, uint32_t ib_idx,
 	       ((uint64_t)amdgpu_get_ib_value(p, ib_idx, hi)) << 32;
 	addr += ((uint64_t)size) * ((uint64_t)index);
 
-	mapping = amdgpu_cs_find_mapping(p, addr, &bo);
-	if (mapping == NULL) {
+	r = amdgpu_cs_find_mapping(p, addr, &bo, &mapping);
+	if (r) {
 		DRM_ERROR("Can't find BO for addr 0x%010Lx %d %d %d %d\n",
 			  addr, lo, hi, size, index);
-		return -EINVAL;
+		return r;
 	}
 
 	if ((addr + (uint64_t)size) >
@@ -652,10 +653,6 @@ int amdgpu_vce_ring_parse_cs(struct amdgpu_cs_parser *p, uint32_t ib_idx)
 	p->job->vm = NULL;
 	ib->gpu_addr = amdgpu_sa_bo_gpu_addr(ib->sa_bo);
 
-	r = amdgpu_cs_sysvm_access_required(p);
-	if (r)
-		return r;
-
 	while (idx < ib->length_dw) {
 		uint32_t len = amdgpu_get_ib_value(p, ib_idx, idx);
 		uint32_t cmd = amdgpu_get_ib_value(p, ib_idx, idx + 1);

drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h

@@ -25,6 +25,7 @@
 #define __AMDGPU_VM_H__
 
 #include <linux/rbtree.h>
+#include <linux/idr.h>
 
 #include "gpu_scheduler.h"
 #include "amdgpu_sync.h"
@@ -105,17 +106,24 @@ struct amdgpu_vm_bo_base {
 
 	/* protected by spinlock */
 	struct list_head		vm_status;
+
+	/* protected by the BO being reserved */
+	bool				moved;
 };
 
 struct amdgpu_vm_pt {
-	struct amdgpu_bo	*bo;
-	uint64_t		addr;
+	struct amdgpu_vm_bo_base	base;
+	uint64_t			addr;
 
 	/* array of page tables, one for each directory entry */
-	struct amdgpu_vm_pt	*entries;
-	unsigned		last_entry_used;
+	struct amdgpu_vm_pt		*entries;
+	unsigned			last_entry_used;
 };
 
+#define AMDGPU_VM_FAULT(pasid, addr) (((u64)(pasid) << 48) | (addr))
+#define AMDGPU_VM_FAULT_PASID(fault) ((u64)(fault) >> 48)
+#define AMDGPU_VM_FAULT_ADDR(fault)  ((u64)(fault) & 0xfffffffff000ULL)
+
 struct amdgpu_vm {
 	/* tree of virtual addresses mapped */
 	struct rb_root_cached	va;
@@ -123,19 +131,21 @@ struct amdgpu_vm {
 	/* protecting invalidated */
 	spinlock_t		status_lock;
 
+	/* BOs who needs a validation */
+	struct list_head	evicted;
+
+	/* PT BOs which relocated and their parent need an update */
+	struct list_head	relocated;
+
 	/* BOs moved, but not yet updated in the PT */
 	struct list_head	moved;
 
-	/* BOs cleared in the PT because of a move */
-	struct list_head	cleared;
-
 	/* BO mappings freed, but not yet updated in the PT */
 	struct list_head	freed;
 
 	/* contains the page directory */
 	struct amdgpu_vm_pt     root;
-	struct dma_fence	*last_dir_update;
-	uint64_t		last_eviction_counter;
+	struct dma_fence	*last_update;
 
 	/* protecting freed */
 	spinlock_t		freed_lock;
@@ -143,8 +153,9 @@ struct amdgpu_vm {
 	/* Scheduler entity for page table updates */
 	struct amd_sched_entity	entity;
 
-	/* client id */
+	/* client id and PASID (TODO: replace client_id with PASID) */
 	u64                     client_id;
+	unsigned int		pasid;
 	/* dedicated to vm */
 	struct amdgpu_vm_id	*reserved_vmid[AMDGPU_MAX_VMHUBS];
 
@@ -153,6 +164,9 @@ struct amdgpu_vm {
 
 	/* Flag to indicate ATS support from PTE for GFX9 */
 	bool			pte_support_ats;
+
+	/* Up to 128 pending page faults */
+	DECLARE_KFIFO(faults, u64, 128);
 };
 
 struct amdgpu_vm_id {
@@ -215,16 +229,25 @@ struct amdgpu_vm_manager {
 	 * BIT1[= 0] Compute updated by SDMA [= 1] by CPU
 	 */
 	int					vm_update_mode;
+
+	/* PASID to VM mapping, will be used in interrupt context to
+	 * look up VM of a page fault
+	 */
+	struct idr				pasid_idr;
+	spinlock_t				pasid_lock;
 };
 
+int amdgpu_vm_alloc_pasid(unsigned int bits);
+void amdgpu_vm_free_pasid(unsigned int pasid);
 void amdgpu_vm_manager_init(struct amdgpu_device *adev);
 void amdgpu_vm_manager_fini(struct amdgpu_device *adev);
 int amdgpu_vm_init(struct amdgpu_device *adev, struct amdgpu_vm *vm,
-		   int vm_context);
+		   int vm_context, unsigned int pasid);
 void amdgpu_vm_fini(struct amdgpu_device *adev, struct amdgpu_vm *vm);
 void amdgpu_vm_get_pd_bo(struct amdgpu_vm *vm,
 			 struct list_head *validated,
 			 struct amdgpu_bo_list_entry *entry);
+bool amdgpu_vm_ready(struct amdgpu_vm *vm);
 int amdgpu_vm_validate_pt_bos(struct amdgpu_device *adev, struct amdgpu_vm *vm,
 			      int (*callback)(void *p, struct amdgpu_bo *bo),
 			      void *param);
@@ -243,13 +266,13 @@ int amdgpu_vm_update_directories(struct amdgpu_device *adev,
 int amdgpu_vm_clear_freed(struct amdgpu_device *adev,
 			  struct amdgpu_vm *vm,
 			  struct dma_fence **fence);
-int amdgpu_vm_clear_moved(struct amdgpu_device *adev, struct amdgpu_vm *vm,
-			  struct amdgpu_sync *sync);
+int amdgpu_vm_handle_moved(struct amdgpu_device *adev,
+			   struct amdgpu_vm *vm);
 int amdgpu_vm_bo_update(struct amdgpu_device *adev,
 			struct amdgpu_bo_va *bo_va,
 			bool clear);
 void amdgpu_vm_bo_invalidate(struct amdgpu_device *adev,
-			     struct amdgpu_bo *bo);
+			     struct amdgpu_bo *bo, bool evicted);
 struct amdgpu_bo_va *amdgpu_vm_bo_find(struct amdgpu_vm *vm,
 				       struct amdgpu_bo *bo);
 struct amdgpu_bo_va *amdgpu_vm_bo_add(struct amdgpu_device *adev,
@@ -269,6 +292,8 @@ int amdgpu_vm_bo_unmap(struct amdgpu_device *adev,
 int amdgpu_vm_bo_clear_mappings(struct amdgpu_device *adev,
 				struct amdgpu_vm *vm,
 				uint64_t saddr, uint64_t size);
+struct amdgpu_bo_va_mapping *amdgpu_vm_bo_lookup_mapping(struct amdgpu_vm *vm,
+							 uint64_t addr);
 void amdgpu_vm_bo_rmv(struct amdgpu_device *adev,
 		      struct amdgpu_bo_va *bo_va);
 void amdgpu_vm_set_fragment_size(struct amdgpu_device *adev,