Merge tag 'amd-drm-next-5.17-2021-12-02' of https://gitlab.freedesktop.org/agd5f/linux into drm-next

amd-drm-next-5.17-2021-12-02:

amdgpu:
- Use generic drm fb helpers
- PSR fixes
- Rework DCN3.1 clkmgr
- DPCD 1.3 fixes
- Misc display fixes can cleanups
- Clock query fixes for APUs
- LTTPR fixes
- DSC fixes
- Misc PM fixes
- RAS fixes
- OLED backlight fix
- SRIOV fixes
- Add STB (Smart Trace Buffer) for supported dGPUs
- IH rework
- Enable seamless boot for DCN3.01

amdkfd:
- Rework more stuff around IP discovery enumeration
- Further clean up of interfaces with amdgpu
- SVM fixes

radeon:
- Indentation fixes

UAPI:
- Add a new KFD header that defines some of the sysfs bitfields and enums that userspace has been using for a while
  The corresponding bit-fields and enums in user mode are defined in
  https://github.com/RadeonOpenCompute/ROCT-Thunk-Interface/blob/master/include/hsakmttypes.h

Signed-off-by: Dave Airlie <airlied@redhat.com>

# Conflicts:
#	drivers/gpu/drm/amd/pm/swsmu/smu_cmn.c
From: Alex Deucher <alexander.deucher@amd.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20211202191643.5970-1-alexander.deucher@amd.com

42 files changed, 961 insertions, 1180 deletions

diff --git a/drivers/gpu/drm/amd/amdgpu/Makefile b/drivers/gpu/drm/amd/amdgpu/Makefile
index 653726588956..7fedbb725e17 100644
--- a/drivers/gpu/drm/amd/amdgpu/Makefile
+++ b/drivers/gpu/drm/amd/amdgpu/Makefile
@@ -45,7 +45,7 @@ amdgpu-y += amdgpu_device.o amdgpu_kms.o \
 	amdgpu_atombios.o atombios_crtc.o amdgpu_connectors.o \
 	atom.o amdgpu_fence.o amdgpu_ttm.o amdgpu_object.o amdgpu_gart.o \
 	amdgpu_encoders.o amdgpu_display.o amdgpu_i2c.o \
-	amdgpu_fb.o amdgpu_gem.o amdgpu_ring.o \
+	amdgpu_gem.o amdgpu_ring.o \
 	amdgpu_cs.o amdgpu_bios.o amdgpu_benchmark.o amdgpu_test.o \
 	atombios_dp.o amdgpu_afmt.o amdgpu_trace_points.o \
 	atombios_encoders.o amdgpu_sa.o atombios_i2c.o \
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c
index 7077f21f0021..46cf48b3904a 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c
@@ -72,7 +72,7 @@ void amdgpu_amdkfd_device_probe(struct amdgpu_device *adev)
 	if (!kfd_initialized)
 		return;
 
-	adev->kfd.dev = kgd2kfd_probe((struct kgd_dev *)adev, vf);
+	adev->kfd.dev = kgd2kfd_probe(adev, vf);
 
 	if (adev->kfd.dev)
 		amdgpu_amdkfd_total_mem_size += adev->gmc.real_vram_size;
@@ -233,19 +233,16 @@ int amdgpu_amdkfd_post_reset(struct amdgpu_device *adev)
 	return r;
 }
 
-void amdgpu_amdkfd_gpu_reset(struct kgd_dev *kgd)
+void amdgpu_amdkfd_gpu_reset(struct amdgpu_device *adev)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)kgd;
-
 	if (amdgpu_device_should_recover_gpu(adev))
 		amdgpu_device_gpu_recover(adev, NULL);
 }
 
-int amdgpu_amdkfd_alloc_gtt_mem(struct kgd_dev *kgd, size_t size,
+int amdgpu_amdkfd_alloc_gtt_mem(struct amdgpu_device *adev, size_t size,
 				void **mem_obj, uint64_t *gpu_addr,
 				void **cpu_ptr, bool cp_mqd_gfx9)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)kgd;
 	struct amdgpu_bo *bo = NULL;
 	struct amdgpu_bo_param bp;
 	int r;
@@ -314,7 +311,7 @@ allocate_mem_reserve_bo_failed:
 	return r;
 }
 
-void amdgpu_amdkfd_free_gtt_mem(struct kgd_dev *kgd, void *mem_obj)
+void amdgpu_amdkfd_free_gtt_mem(struct amdgpu_device *adev, void *mem_obj)
 {
 	struct amdgpu_bo *bo = (struct amdgpu_bo *) mem_obj;
 
@@ -325,10 +322,9 @@ void amdgpu_amdkfd_free_gtt_mem(struct kgd_dev *kgd, void *mem_obj)
 	amdgpu_bo_unref(&(bo));
 }
 
-int amdgpu_amdkfd_alloc_gws(struct kgd_dev *kgd, size_t size,
+int amdgpu_amdkfd_alloc_gws(struct amdgpu_device *adev, size_t size,
 				void **mem_obj)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)kgd;
 	struct amdgpu_bo *bo = NULL;
 	struct amdgpu_bo_user *ubo;
 	struct amdgpu_bo_param bp;
@@ -355,18 +351,16 @@ int amdgpu_amdkfd_alloc_gws(struct kgd_dev *kgd, size_t size,
 	return 0;
 }
 
-void amdgpu_amdkfd_free_gws(struct kgd_dev *kgd, void *mem_obj)
+void amdgpu_amdkfd_free_gws(struct amdgpu_device *adev, void *mem_obj)
 {
 	struct amdgpu_bo *bo = (struct amdgpu_bo *)mem_obj;
 
 	amdgpu_bo_unref(&bo);
 }
 
-uint32_t amdgpu_amdkfd_get_fw_version(struct kgd_dev *kgd,
+uint32_t amdgpu_amdkfd_get_fw_version(struct amdgpu_device *adev,
 				      enum kgd_engine_type type)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)kgd;
-
 	switch (type) {
 	case KGD_ENGINE_PFP:
 		return adev->gfx.pfp_fw_version;
@@ -399,11 +393,9 @@ uint32_t amdgpu_amdkfd_get_fw_version(struct kgd_dev *kgd,
 	return 0;
 }
 
-void amdgpu_amdkfd_get_local_mem_info(struct kgd_dev *kgd,
+void amdgpu_amdkfd_get_local_mem_info(struct amdgpu_device *adev,
 				      struct kfd_local_mem_info *mem_info)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)kgd;
-
 	memset(mem_info, 0, sizeof(*mem_info));
 
 	mem_info->local_mem_size_public = adev->gmc.visible_vram_size;
@@ -428,19 +420,15 @@ void amdgpu_amdkfd_get_local_mem_info(struct kgd_dev *kgd,
 		mem_info->mem_clk_max = 100;
 }
 
-uint64_t amdgpu_amdkfd_get_gpu_clock_counter(struct kgd_dev *kgd)
+uint64_t amdgpu_amdkfd_get_gpu_clock_counter(struct amdgpu_device *adev)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)kgd;
-
 	if (adev->gfx.funcs->get_gpu_clock_counter)
 		return adev->gfx.funcs->get_gpu_clock_counter(adev);
 	return 0;
 }
 
-uint32_t amdgpu_amdkfd_get_max_engine_clock_in_mhz(struct kgd_dev *kgd)
+uint32_t amdgpu_amdkfd_get_max_engine_clock_in_mhz(struct amdgpu_device *adev)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)kgd;
-
 	/* the sclk is in quantas of 10kHz */
 	if (amdgpu_sriov_vf(adev))
 		return adev->clock.default_sclk / 100;
@@ -450,9 +438,8 @@ uint32_t amdgpu_amdkfd_get_max_engine_clock_in_mhz(struct kgd_dev *kgd)
 		return 100;
 }
 
-void amdgpu_amdkfd_get_cu_info(struct kgd_dev *kgd, struct kfd_cu_info *cu_info)
+void amdgpu_amdkfd_get_cu_info(struct amdgpu_device *adev, struct kfd_cu_info *cu_info)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)kgd;
 	struct amdgpu_cu_info acu_info = adev->gfx.cu_info;
 
 	memset(cu_info, 0, sizeof(*cu_info));
@@ -473,13 +460,12 @@ void amdgpu_amdkfd_get_cu_info(struct kgd_dev *kgd, struct kfd_cu_info *cu_info)
 	cu_info->lds_size = acu_info.lds_size;
 }
 
-int amdgpu_amdkfd_get_dmabuf_info(struct kgd_dev *kgd, int dma_buf_fd,
-				  struct kgd_dev **dma_buf_kgd,
+int amdgpu_amdkfd_get_dmabuf_info(struct amdgpu_device *adev, int dma_buf_fd,
+				  struct amdgpu_device **dmabuf_adev,
 				  uint64_t *bo_size, void *metadata_buffer,
 				  size_t buffer_size, uint32_t *metadata_size,
 				  uint32_t *flags)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)kgd;
 	struct dma_buf *dma_buf;
 	struct drm_gem_object *obj;
 	struct amdgpu_bo *bo;
@@ -507,8 +493,8 @@ int amdgpu_amdkfd_get_dmabuf_info(struct kgd_dev *kgd, int dma_buf_fd,
 		goto out_put;
 
 	r = 0;
-	if (dma_buf_kgd)
-		*dma_buf_kgd = (struct kgd_dev *)adev;
+	if (dmabuf_adev)
+		*dmabuf_adev = adev;
 	if (bo_size)
 		*bo_size = amdgpu_bo_size(bo);
 	if (metadata_buffer)
@@ -528,32 +514,18 @@ out_put:
 	return r;
 }
 
-uint64_t amdgpu_amdkfd_get_vram_usage(struct kgd_dev *kgd)
+uint64_t amdgpu_amdkfd_get_vram_usage(struct amdgpu_device *adev)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)kgd;
 	struct ttm_resource_manager *vram_man = ttm_manager_type(&adev->mman.bdev, TTM_PL_VRAM);
 
 	return amdgpu_vram_mgr_usage(vram_man);
 }
 
-uint64_t amdgpu_amdkfd_get_hive_id(struct kgd_dev *kgd)
-{
-	struct amdgpu_device *adev = (struct amdgpu_device *)kgd;
-
-	return adev->gmc.xgmi.hive_id;
-}
-
-uint64_t amdgpu_amdkfd_get_unique_id(struct kgd_dev *kgd)
-{
-	struct amdgpu_device *adev = (struct amdgpu_device *)kgd;
-
-	return adev->unique_id;
-}
-
-uint8_t amdgpu_amdkfd_get_xgmi_hops_count(struct kgd_dev *dst, struct kgd_dev *src)
+uint8_t amdgpu_amdkfd_get_xgmi_hops_count(struct amdgpu_device *dst,
+					  struct amdgpu_device *src)
 {
-	struct amdgpu_device *peer_adev = (struct amdgpu_device *)src;
-	struct amdgpu_device *adev = (struct amdgpu_device *)dst;
+	struct amdgpu_device *peer_adev = src;
+	struct amdgpu_device *adev = dst;
 	int ret = amdgpu_xgmi_get_hops_count(adev, peer_adev);
 
 	if (ret < 0) {
@@ -565,16 +537,18 @@ uint8_t amdgpu_amdkfd_get_xgmi_hops_count(struct kgd_dev *dst, struct kgd_dev *s
 	return  (uint8_t)ret;
 }
 
-int amdgpu_amdkfd_get_xgmi_bandwidth_mbytes(struct kgd_dev *dst, struct kgd_dev *src, bool is_min)
+int amdgpu_amdkfd_get_xgmi_bandwidth_mbytes(struct amdgpu_device *dst,
+					    struct amdgpu_device *src,
+					    bool is_min)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)dst, *peer_adev;
+	struct amdgpu_device *adev = dst, *peer_adev;
 	int num_links;
 
 	if (adev->asic_type != CHIP_ALDEBARAN)
 		return 0;
 
 	if (src)
-		peer_adev = (struct amdgpu_device *)src;
+		peer_adev = src;
 
 	/* num links returns 0 for indirect peers since indirect route is unknown. */
 	num_links = is_min ? 1 : amdgpu_xgmi_get_num_links(adev, peer_adev);
@@ -589,9 +563,8 @@ int amdgpu_amdkfd_get_xgmi_bandwidth_mbytes(struct kgd_dev *dst, struct kgd_dev
 	return (num_links * 16 * 25000)/BITS_PER_BYTE;
 }
 
-int amdgpu_amdkfd_get_pcie_bandwidth_mbytes(struct kgd_dev *dev, bool is_min)
+int amdgpu_amdkfd_get_pcie_bandwidth_mbytes(struct amdgpu_device *adev, bool is_min)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)dev;
 	int num_lanes_shift = (is_min ? ffs(adev->pm.pcie_mlw_mask) :
 							fls(adev->pm.pcie_mlw_mask)) - 1;
 	int gen_speed_shift = (is_min ? ffs(adev->pm.pcie_gen_mask &
@@ -647,39 +620,11 @@ int amdgpu_amdkfd_get_pcie_bandwidth_mbytes(struct kgd_dev *dev, bool is_min)
 	return (num_lanes_factor * gen_speed_mbits_factor)/BITS_PER_BYTE;
 }
 
-uint64_t amdgpu_amdkfd_get_mmio_remap_phys_addr(struct kgd_dev *kgd)
-{
-	struct amdgpu_device *adev = (struct amdgpu_device *)kgd;
-
-	return adev->rmmio_remap.bus_addr;
-}
-
-uint32_t amdgpu_amdkfd_get_num_gws(struct kgd_dev *kgd)
-{
-	struct amdgpu_device *adev = (struct amdgpu_device *)kgd;
-
-	return adev->gds.gws_size;
-}
-
-uint32_t amdgpu_amdkfd_get_asic_rev_id(struct kgd_dev *kgd)
-{
-	struct amdgpu_device *adev = (struct amdgpu_device *)kgd;
-
-	return adev->rev_id;
-}
-
-int amdgpu_amdkfd_get_noretry(struct kgd_dev *kgd)
-{
-	struct amdgpu_device *adev = (struct amdgpu_device *)kgd;
-
-	return adev->gmc.noretry;
-}
-
-int amdgpu_amdkfd_submit_ib(struct kgd_dev *kgd, enum kgd_engine_type engine,
+int amdgpu_amdkfd_submit_ib(struct amdgpu_device *adev,
+				enum kgd_engine_type engine,
 				uint32_t vmid, uint64_t gpu_addr,
 				uint32_t *ib_cmd, uint32_t ib_len)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)kgd;
 	struct amdgpu_job *job;
 	struct amdgpu_ib *ib;
 	struct amdgpu_ring *ring;
@@ -730,10 +675,8 @@ err:
 	return ret;
 }
 
-void amdgpu_amdkfd_set_compute_idle(struct kgd_dev *kgd, bool idle)
+void amdgpu_amdkfd_set_compute_idle(struct amdgpu_device *adev, bool idle)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)kgd;
-
 	amdgpu_dpm_switch_power_profile(adev,
 					PP_SMC_POWER_PROFILE_COMPUTE,
 					!idle);
@@ -747,10 +690,9 @@ bool amdgpu_amdkfd_is_kfd_vmid(struct amdgpu_device *adev, u32 vmid)
 	return false;
 }
 
-int amdgpu_amdkfd_flush_gpu_tlb_vmid(struct kgd_dev *kgd, uint16_t vmid)
+int amdgpu_amdkfd_flush_gpu_tlb_vmid(struct amdgpu_device *adev,
+				     uint16_t vmid)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)kgd;
-
 	if (adev->family == AMDGPU_FAMILY_AI) {
 		int i;
 
@@ -763,10 +705,9 @@ int amdgpu_amdkfd_flush_gpu_tlb_vmid(struct kgd_dev *kgd, uint16_t vmid)
 	return 0;
 }
 
-int amdgpu_amdkfd_flush_gpu_tlb_pasid(struct kgd_dev *kgd, uint16_t pasid,
-				      enum TLB_FLUSH_TYPE flush_type)
+int amdgpu_amdkfd_flush_gpu_tlb_pasid(struct amdgpu_device *adev,
+				      uint16_t pasid, enum TLB_FLUSH_TYPE flush_type)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)kgd;
 	bool all_hub = false;
 
 	if (adev->family == AMDGPU_FAMILY_AI)
@@ -775,21 +716,18 @@ int amdgpu_amdkfd_flush_gpu_tlb_pasid(struct kgd_dev *kgd, uint16_t pasid,
 	return amdgpu_gmc_flush_gpu_tlb_pasid(adev, pasid, flush_type, all_hub);
 }
 
-bool amdgpu_amdkfd_have_atomics_support(struct kgd_dev *kgd)
+bool amdgpu_amdkfd_have_atomics_support(struct amdgpu_device *adev)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)kgd;
-
 	return adev->have_atomics_support;
 }
 
-void amdgpu_amdkfd_ras_poison_consumption_handler(struct kgd_dev *kgd)
+void amdgpu_amdkfd_ras_poison_consumption_handler(struct amdgpu_device *adev)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)kgd;
 	struct ras_err_data err_data = {0, 0, 0, NULL};
 
 	/* CPU MCA will handle page retirement if connected_to_cpu is 1 */
 	if (!adev->gmc.xgmi.connected_to_cpu)
 		amdgpu_umc_process_ras_data_cb(adev, &err_data, NULL);
 	else
-		amdgpu_amdkfd_gpu_reset(kgd);
+		amdgpu_amdkfd_gpu_reset(adev);
 }
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h
index a15a4787c7ee..fcbc8a9c9e06 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h
@@ -144,14 +144,16 @@ void amdgpu_amdkfd_interrupt(struct amdgpu_device *adev,
 void amdgpu_amdkfd_device_probe(struct amdgpu_device *adev);
 void amdgpu_amdkfd_device_init(struct amdgpu_device *adev);
 void amdgpu_amdkfd_device_fini_sw(struct amdgpu_device *adev);
-int amdgpu_amdkfd_submit_ib(struct kgd_dev *kgd, enum kgd_engine_type engine,
+int amdgpu_amdkfd_submit_ib(struct amdgpu_device *adev,
+				enum kgd_engine_type engine,
 				uint32_t vmid, uint64_t gpu_addr,
 				uint32_t *ib_cmd, uint32_t ib_len);
-void amdgpu_amdkfd_set_compute_idle(struct kgd_dev *kgd, bool idle);
-bool amdgpu_amdkfd_have_atomics_support(struct kgd_dev *kgd);
-int amdgpu_amdkfd_flush_gpu_tlb_vmid(struct kgd_dev *kgd, uint16_t vmid);
-int amdgpu_amdkfd_flush_gpu_tlb_pasid(struct kgd_dev *kgd, uint16_t pasid,
-				      enum TLB_FLUSH_TYPE flush_type);
+void amdgpu_amdkfd_set_compute_idle(struct amdgpu_device *adev, bool idle);
+bool amdgpu_amdkfd_have_atomics_support(struct amdgpu_device *adev);
+int amdgpu_amdkfd_flush_gpu_tlb_vmid(struct amdgpu_device *adev,
+				uint16_t vmid);
+int amdgpu_amdkfd_flush_gpu_tlb_pasid(struct amdgpu_device *adev,
+				uint16_t pasid, enum TLB_FLUSH_TYPE flush_type);
 
 bool amdgpu_amdkfd_is_kfd_vmid(struct amdgpu_device *adev, u32 vmid);
 
@@ -159,7 +161,7 @@ int amdgpu_amdkfd_pre_reset(struct amdgpu_device *adev);
 
 int amdgpu_amdkfd_post_reset(struct amdgpu_device *adev);
 
-void amdgpu_amdkfd_gpu_reset(struct kgd_dev *kgd);
+void amdgpu_amdkfd_gpu_reset(struct amdgpu_device *adev);
 
 int amdgpu_queue_mask_bit_to_set_resource_bit(struct amdgpu_device *adev,
 					int queue_bit);
@@ -198,37 +200,36 @@ int amdgpu_amdkfd_evict_userptr(struct kgd_mem *mem, struct mm_struct *mm)
 }
 #endif
 /* Shared API */
-int amdgpu_amdkfd_alloc_gtt_mem(struct kgd_dev *kgd, size_t size,
+int amdgpu_amdkfd_alloc_gtt_mem(struct amdgpu_device *adev, size_t size,
 				void **mem_obj, uint64_t *gpu_addr,
 				void **cpu_ptr, bool mqd_gfx9);
-void amdgpu_amdkfd_free_gtt_mem(struct kgd_dev *kgd, void *mem_obj);
-int amdgpu_amdkfd_alloc_gws(struct kgd_dev *kgd, size_t size, void **mem_obj);
-void amdgpu_amdkfd_free_gws(struct kgd_dev *kgd, void *mem_obj);
+void amdgpu_amdkfd_free_gtt_mem(struct amdgpu_device *adev, void *mem_obj);
+int amdgpu_amdkfd_alloc_gws(struct amdgpu_device *adev, size_t size,
+				void **mem_obj);
+void amdgpu_amdkfd_free_gws(struct amdgpu_device *adev, void *mem_obj);
 int amdgpu_amdkfd_add_gws_to_process(void *info, void *gws, struct kgd_mem **mem);
 int amdgpu_amdkfd_remove_gws_from_process(void *info, void *mem);
-uint32_t amdgpu_amdkfd_get_fw_version(struct kgd_dev *kgd,
+uint32_t amdgpu_amdkfd_get_fw_version(struct amdgpu_device *adev,
 				      enum kgd_engine_type type);
-void amdgpu_amdkfd_get_local_mem_info(struct kgd_dev *kgd,
+void amdgpu_amdkfd_get_local_mem_info(struct amdgpu_device *adev,
 				      struct kfd_local_mem_info *mem_info);
-uint64_t amdgpu_amdkfd_get_gpu_clock_counter(struct kgd_dev *kgd);
+uint64_t amdgpu_amdkfd_get_gpu_clock_counter(struct amdgpu_device *adev);
 
-uint32_t amdgpu_amdkfd_get_max_engine_clock_in_mhz(struct kgd_dev *kgd);
-void amdgpu_amdkfd_get_cu_info(struct kgd_dev *kgd, struct kfd_cu_info *cu_info);
-int amdgpu_amdkfd_get_dmabuf_info(struct kgd_dev *kgd, int dma_buf_fd,
-				  struct kgd_dev **dmabuf_kgd,
+uint32_t amdgpu_amdkfd_get_max_engine_clock_in_mhz(struct amdgpu_device *adev);
+void amdgpu_amdkfd_get_cu_info(struct amdgpu_device *adev,
+			       struct kfd_cu_info *cu_info);
+int amdgpu_amdkfd_get_dmabuf_info(struct amdgpu_device *adev, int dma_buf_fd,
+				  struct amdgpu_device **dmabuf_adev,
 				  uint64_t *bo_size, void *metadata_buffer,
 				  size_t buffer_size, uint32_t *metadata_size,
 				  uint32_t *flags);
-uint64_t amdgpu_amdkfd_get_vram_usage(struct kgd_dev *kgd);
-uint64_t amdgpu_amdkfd_get_hive_id(struct kgd_dev *kgd);
-uint64_t amdgpu_amdkfd_get_unique_id(struct kgd_dev *kgd);
-uint64_t amdgpu_amdkfd_get_mmio_remap_phys_addr(struct kgd_dev *kgd);
-uint32_t amdgpu_amdkfd_get_num_gws(struct kgd_dev *kgd);
-uint32_t amdgpu_amdkfd_get_asic_rev_id(struct kgd_dev *kgd);
-int amdgpu_amdkfd_get_noretry(struct kgd_dev *kgd);
-uint8_t amdgpu_amdkfd_get_xgmi_hops_count(struct kgd_dev *dst, struct kgd_dev *src);
-int amdgpu_amdkfd_get_xgmi_bandwidth_mbytes(struct kgd_dev *dst, struct kgd_dev *src, bool is_min);
-int amdgpu_amdkfd_get_pcie_bandwidth_mbytes(struct kgd_dev *dev, bool is_min);
+uint64_t amdgpu_amdkfd_get_vram_usage(struct amdgpu_device *adev);
+uint8_t amdgpu_amdkfd_get_xgmi_hops_count(struct amdgpu_device *dst,
+					  struct amdgpu_device *src);
+int amdgpu_amdkfd_get_xgmi_bandwidth_mbytes(struct amdgpu_device *dst,
+					    struct amdgpu_device *src,
+					    bool is_min);
+int amdgpu_amdkfd_get_pcie_bandwidth_mbytes(struct amdgpu_device *adev, bool is_min);
 
 /* Read user wptr from a specified user address space with page fault
  * disabled. The memory must be pinned and mapped to the hardware when
@@ -258,45 +259,54 @@ int amdgpu_amdkfd_get_pcie_bandwidth_mbytes(struct kgd_dev *dev, bool is_min);
 	(&((struct amdgpu_fpriv *)					\
 		((struct drm_file *)(drm_priv))->driver_priv)->vm)
 
-int amdgpu_amdkfd_gpuvm_acquire_process_vm(struct kgd_dev *kgd,
+int amdgpu_amdkfd_gpuvm_acquire_process_vm(struct amdgpu_device *adev,
 					struct file *filp, u32 pasid,
 					void **process_info,
 					struct dma_fence **ef);
-void amdgpu_amdkfd_gpuvm_release_process_vm(struct kgd_dev *kgd, void *drm_priv);
+void amdgpu_amdkfd_gpuvm_release_process_vm(struct amdgpu_device *adev,
+					void *drm_priv);
 uint64_t amdgpu_amdkfd_gpuvm_get_process_page_dir(void *drm_priv);
 int amdgpu_amdkfd_gpuvm_alloc_memory_of_gpu(
-		struct kgd_dev *kgd, uint64_t va, uint64_t size,
+		struct amdgpu_device *adev, uint64_t va, uint64_t size,
 		void *drm_priv, struct kgd_mem **mem,
 		uint64_t *offset, uint32_t flags);
 int amdgpu_amdkfd_gpuvm_free_memory_of_gpu(
-		struct kgd_dev *kgd, struct kgd_mem *mem, void *drm_priv,
+		struct amdgpu_device *adev, struct kgd_mem *mem, void *drm_priv,
 		uint64_t *size);
 int amdgpu_amdkfd_gpuvm_map_memory_to_gpu(
-		struct kgd_dev *kgd, struct kgd_mem *mem, void *drm_priv, bool *table_freed);
+		struct amdgpu_device *adev, struct kgd_mem *mem, void *drm_priv,
+		bool *table_freed);
 int amdgpu_amdkfd_gpuvm_unmap_memory_from_gpu(
-		struct kgd_dev *kgd, struct kgd_mem *mem, void *drm_priv);
+		struct amdgpu_device *adev, struct kgd_mem *mem, void *drm_priv);
 int amdgpu_amdkfd_gpuvm_sync_memory(
-		struct kgd_dev *kgd, struct kgd_mem *mem, bool intr);
-int amdgpu_amdkfd_gpuvm_map_gtt_bo_to_kernel(struct kgd_dev *kgd,
+		struct amdgpu_device *adev, struct kgd_mem *mem, bool intr);
+int amdgpu_amdkfd_gpuvm_map_gtt_bo_to_kernel(struct amdgpu_device *adev,
 		struct kgd_mem *mem, void **kptr, uint64_t *size);
-void amdgpu_amdkfd_gpuvm_unmap_gtt_bo_from_kernel(struct kgd_dev *kgd, struct kgd_mem *mem);
+void amdgpu_amdkfd_gpuvm_unmap_gtt_bo_from_kernel(struct amdgpu_device *adev,
+		struct kgd_mem *mem);
 
 int amdgpu_amdkfd_gpuvm_restore_process_bos(void *process_info,
 					    struct dma_fence **ef);
-int amdgpu_amdkfd_gpuvm_get_vm_fault_info(struct kgd_dev *kgd,
+int amdgpu_amdkfd_gpuvm_get_vm_fault_info(struct amdgpu_device *adev,
 					      struct kfd_vm_fault_info *info);
-int amdgpu_amdkfd_gpuvm_import_dmabuf(struct kgd_dev *kgd,
+int amdgpu_amdkfd_gpuvm_import_dmabuf(struct amdgpu_device *adev,
 				      struct dma_buf *dmabuf,
 				      uint64_t va, void *drm_priv,
 				      struct kgd_mem **mem, uint64_t *size,
 				      uint64_t *mmap_offset);
-int amdgpu_amdkfd_get_tile_config(struct kgd_dev *kgd,
+int amdgpu_amdkfd_get_tile_config(struct amdgpu_device *adev,
 				struct tile_config *config);
-void amdgpu_amdkfd_ras_poison_consumption_handler(struct kgd_dev *kgd);
+void amdgpu_amdkfd_ras_poison_consumption_handler(struct amdgpu_device *adev);
 #if IS_ENABLED(CONFIG_HSA_AMD)
 void amdgpu_amdkfd_gpuvm_init_mem_limits(void);
 void amdgpu_amdkfd_gpuvm_destroy_cb(struct amdgpu_device *adev,
 				struct amdgpu_vm *vm);
+
+/**
+ * @amdgpu_amdkfd_release_notify() - Notify KFD when GEM object is released
+ *
+ * Allows KFD to release its resources associated with the GEM object.
+ */
 void amdgpu_amdkfd_release_notify(struct amdgpu_bo *bo);
 void amdgpu_amdkfd_reserve_system_mem(uint64_t size);
 #else
@@ -324,7 +334,7 @@ int kgd2kfd_schedule_evict_and_restore_process(struct mm_struct *mm,
 #if IS_ENABLED(CONFIG_HSA_AMD)
 int kgd2kfd_init(void);
 void kgd2kfd_exit(void);
-struct kfd_dev *kgd2kfd_probe(struct kgd_dev *kgd, bool vf);
+struct kfd_dev *kgd2kfd_probe(struct amdgpu_device *adev, bool vf);
 bool kgd2kfd_device_init(struct kfd_dev *kfd,
 			 struct drm_device *ddev,
 			 const struct kgd2kfd_shared_resources *gpu_resources);
@@ -348,7 +358,7 @@ static inline void kgd2kfd_exit(void)
 }
 
 static inline
-struct kfd_dev *kgd2kfd_probe(struct kgd_dev *kgd, bool vf)
+struct kfd_dev *kgd2kfd_probe(struct amdgpu_device *adev, bool vf)
 {
 	return NULL;
 }
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_arcturus.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_arcturus.c
index 5a7f680bcb3f..abe93b3ff765 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_arcturus.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_arcturus.c
@@ -57,11 +57,6 @@
 		(*dump)[i++][1] = RREG32(addr);		\
 	} while (0)
 
-static inline struct amdgpu_device *get_amdgpu_device(struct kgd_dev *kgd)
-{
-	return (struct amdgpu_device *)kgd;
-}
-
 static inline struct v9_sdma_mqd *get_sdma_mqd(void *mqd)
 {
 	return (struct v9_sdma_mqd *)mqd;
@@ -123,10 +118,9 @@ static uint32_t get_sdma_rlc_reg_offset(struct amdgpu_device *adev,
 	return sdma_rlc_reg_offset;
 }
 
-int kgd_arcturus_hqd_sdma_load(struct kgd_dev *kgd, void *mqd,
+int kgd_arcturus_hqd_sdma_load(struct amdgpu_device *adev, void *mqd,
 			     uint32_t __user *wptr, struct mm_struct *mm)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	struct v9_sdma_mqd *m;
 	uint32_t sdma_rlc_reg_offset;
 	unsigned long end_jiffies;
@@ -193,11 +187,10 @@ int kgd_arcturus_hqd_sdma_load(struct kgd_dev *kgd, void *mqd,
 	return 0;
 }
 
-int kgd_arcturus_hqd_sdma_dump(struct kgd_dev *kgd,
+int kgd_arcturus_hqd_sdma_dump(struct amdgpu_device *adev,
 			     uint32_t engine_id, uint32_t queue_id,
 			     uint32_t (**dump)[2], uint32_t *n_regs)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	uint32_t sdma_rlc_reg_offset = get_sdma_rlc_reg_offset(adev,
 			engine_id, queue_id);
 	uint32_t i = 0, reg;
@@ -225,9 +218,9 @@ int kgd_arcturus_hqd_sdma_dump(struct kgd_dev *kgd,
 	return 0;
 }
 
-bool kgd_arcturus_hqd_sdma_is_occupied(struct kgd_dev *kgd, void *mqd)
+bool kgd_arcturus_hqd_sdma_is_occupied(struct amdgpu_device *adev,
+				void *mqd)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	struct v9_sdma_mqd *m;
 	uint32_t sdma_rlc_reg_offset;
 	uint32_t sdma_rlc_rb_cntl;
@@ -244,10 +237,9 @@ bool kgd_arcturus_hqd_sdma_is_occupied(struct kgd_dev *kgd, void *mqd)
 	return false;
 }
 
-int kgd_arcturus_hqd_sdma_destroy(struct kgd_dev *kgd, void *mqd,
+int kgd_arcturus_hqd_sdma_destroy(struct amdgpu_device *adev, void *mqd,
 				unsigned int utimeout)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	struct v9_sdma_mqd *m;
 	uint32_t sdma_rlc_reg_offset;
 	uint32_t temp;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_arcturus.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_arcturus.h
index ce08131b7b5f..756c1a5679c0 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_arcturus.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_arcturus.h
@@ -20,11 +20,12 @@
  * OTHER DEALINGS IN THE SOFTWARE.
  */
 
-int kgd_arcturus_hqd_sdma_load(struct kgd_dev *kgd, void *mqd,
+int kgd_arcturus_hqd_sdma_load(struct amdgpu_device *adev, void *mqd,
 			     uint32_t __user *wptr, struct mm_struct *mm);
-int kgd_arcturus_hqd_sdma_dump(struct kgd_dev *kgd,
+int kgd_arcturus_hqd_sdma_dump(struct amdgpu_device *adev,
 			     uint32_t engine_id, uint32_t queue_id,
 			     uint32_t (**dump)[2], uint32_t *n_regs);
-bool kgd_arcturus_hqd_sdma_is_occupied(struct kgd_dev *kgd, void *mqd);
-int kgd_arcturus_hqd_sdma_destroy(struct kgd_dev *kgd, void *mqd,
+bool kgd_arcturus_hqd_sdma_is_occupied(struct amdgpu_device *adev,
+				void *mqd);
+int kgd_arcturus_hqd_sdma_destroy(struct amdgpu_device *adev, void *mqd,
 				unsigned int utimeout);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v10.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v10.c
index 960acf68150a..7b7f4b2764c1 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v10.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v10.c
@@ -39,37 +39,26 @@ enum hqd_dequeue_request_type {
 	SAVE_WAVES
 };
 
-static inline struct amdgpu_device *get_amdgpu_device(struct kgd_dev *kgd)
-{
-	return (struct amdgpu_device *)kgd;
-}
-
-static void lock_srbm(struct kgd_dev *kgd, uint32_t mec, uint32_t pipe,
+static void lock_srbm(struct amdgpu_device *adev, uint32_t mec, uint32_t pipe,
 			uint32_t queue, uint32_t vmid)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
-
 	mutex_lock(&adev->srbm_mutex);
 	nv_grbm_select(adev, mec, pipe, queue, vmid);
 }
 
-static void unlock_srbm(struct kgd_dev *kgd)
+static void unlock_srbm(struct amdgpu_device *adev)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
-
 	nv_grbm_select(adev, 0, 0, 0, 0);
 	mutex_unlock(&adev->srbm_mutex);
 }
 
-static void acquire_queue(struct kgd_dev *kgd, uint32_t pipe_id,
+static void acquire_queue(struct amdgpu_device *adev, uint32_t pipe_id,
 				uint32_t queue_id)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
-
 	uint32_t mec = (pipe_id / adev->gfx.mec.num_pipe_per_mec) + 1;
 	uint32_t pipe = (pipe_id % adev->gfx.mec.num_pipe_per_mec);
 
-	lock_srbm(kgd, mec, pipe, queue_id, 0);
+	lock_srbm(adev, mec, pipe, queue_id, 0);
 }
 
 static uint64_t get_queue_mask(struct amdgpu_device *adev,
@@ -81,33 +70,29 @@ static uint64_t get_queue_mask(struct amdgpu_device *adev,
 	return 1ull << bit;
 }
 
-static void release_queue(struct kgd_dev *kgd)
+static void release_queue(struct amdgpu_device *adev)
 {
-	unlock_srbm(kgd);
+	unlock_srbm(adev);
 }
 
-static void kgd_program_sh_mem_settings(struct kgd_dev *kgd, uint32_t vmid,
+static void kgd_program_sh_mem_settings(struct amdgpu_device *adev, uint32_t vmid,
 					uint32_t sh_mem_config,
 					uint32_t sh_mem_ape1_base,
 					uint32_t sh_mem_ape1_limit,
 					uint32_t sh_mem_bases)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
-
-	lock_srbm(kgd, 0, 0, 0, vmid);
+	lock_srbm(adev, 0, 0, 0, vmid);
 
 	WREG32_SOC15(GC, 0, mmSH_MEM_CONFIG, sh_mem_config);
 	WREG32_SOC15(GC, 0, mmSH_MEM_BASES, sh_mem_bases);
 	/* APE1 no longer exists on GFX9 */
 
-	unlock_srbm(kgd);
+	unlock_srbm(adev);
 }
 
-static int kgd_set_pasid_vmid_mapping(struct kgd_dev *kgd, u32 pasid,
+static int kgd_set_pasid_vmid_mapping(struct amdgpu_device *adev, u32 pasid,
 					unsigned int vmid)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
-
 	/*
 	 * We have to assume that there is no outstanding mapping.
 	 * The ATC_VMID_PASID_MAPPING_UPDATE_STATUS bit could be 0 because
@@ -150,22 +135,21 @@ static int kgd_set_pasid_vmid_mapping(struct kgd_dev *kgd, u32 pasid,
  * but still works
  */
 
-static int kgd_init_interrupts(struct kgd_dev *kgd, uint32_t pipe_id)
+static int kgd_init_interrupts(struct amdgpu_device *adev, uint32_t pipe_id)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	uint32_t mec;
 	uint32_t pipe;
 
 	mec = (pipe_id / adev->gfx.mec.num_pipe_per_mec) + 1;
 	pipe = (pipe_id % adev->gfx.mec.num_pipe_per_mec);
 
-	lock_srbm(kgd, mec, pipe, 0, 0);
+	lock_srbm(adev, mec, pipe, 0, 0);
 
 	WREG32_SOC15(GC, 0, mmCPC_INT_CNTL,
 		CP_INT_CNTL_RING0__TIME_STAMP_INT_ENABLE_MASK |
 		CP_INT_CNTL_RING0__OPCODE_ERROR_INT_ENABLE_MASK);
 
-	unlock_srbm(kgd);
+	unlock_srbm(adev);
 
 	return 0;
 }
@@ -218,12 +202,11 @@ static inline struct v10_sdma_mqd *get_sdma_mqd(void *mqd)
 	return (struct v10_sdma_mqd *)mqd;
 }
 
-static int kgd_hqd_load(struct kgd_dev *kgd, void *mqd, uint32_t pipe_id,
-			uint32_t queue_id, uint32_t __user *wptr,
-			uint32_t wptr_shift, uint32_t wptr_mask,
-			struct mm_struct *mm)
+static int kgd_hqd_load(struct amdgpu_device *adev, void *mqd,
+			uint32_t pipe_id, uint32_t queue_id,
+			uint32_t __user *wptr, uint32_t wptr_shift,
+			uint32_t wptr_mask, struct mm_struct *mm)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	struct v10_compute_mqd *m;
 	uint32_t *mqd_hqd;
 	uint32_t reg, hqd_base, data;
@@ -231,7 +214,7 @@ static int kgd_hqd_load(struct kgd_dev *kgd, void *mqd, uint32_t pipe_id,
 	m = get_mqd(mqd);
 
 	pr_debug("Load hqd of pipe %d queue %d\n", pipe_id, queue_id);
-	acquire_queue(kgd, pipe_id, queue_id);
+	acquire_queue(adev, pipe_id, queue_id);
 
 	/* HQD registers extend from CP_MQD_BASE_ADDR to CP_HQD_EOP_WPTR_MEM. */
 	mqd_hqd = &m->cp_mqd_base_addr_lo;
@@ -296,16 +279,15 @@ static int kgd_hqd_load(struct kgd_dev *kgd, void *mqd, uint32_t pipe_id,
 	data = REG_SET_FIELD(m->cp_hqd_active, CP_HQD_ACTIVE, ACTIVE, 1);
 	WREG32_SOC15(GC, 0, mmCP_HQD_ACTIVE, data);
 
-	release_queue(kgd);
+	release_queue(adev);
 
 	return 0;
 }
 
-static int kgd_hiq_mqd_load(struct kgd_dev *kgd, void *mqd,
+static int kgd_hiq_mqd_load(struct amdgpu_device *adev, void *mqd,
 			    uint32_t pipe_id, uint32_t queue_id,
 			    uint32_t doorbell_off)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	struct amdgpu_ring *kiq_ring = &adev->gfx.kiq.ring;
 	struct v10_compute_mqd *m;
 	uint32_t mec, pipe;
@@ -313,7 +295,7 @@ static int kgd_hiq_mqd_load(struct kgd_dev *kgd, void *mqd,
 
 	m = get_mqd(mqd);
 
-	acquire_queue(kgd, pipe_id, queue_id);
+	acquire_queue(adev, pipe_id, queue_id);
 
 	mec = (pipe_id / adev->gfx.mec.num_pipe_per_mec) + 1;
 	pipe = (pipe_id % adev->gfx.mec.num_pipe_per_mec);
@@ -349,16 +331,15 @@ static int kgd_hiq_mqd_load(struct kgd_dev *kgd, void *mqd,
 
 out_unlock:
 	spin_unlock(&adev->gfx.kiq.ring_lock);
-	release_queue(kgd);
+	release_queue(adev);
 
 	return r;
 }
 
-static int kgd_hqd_dump(struct kgd_dev *kgd,
+static int kgd_hqd_dump(struct amdgpu_device *adev,
 			uint32_t pipe_id, uint32_t queue_id,
 			uint32_t (**dump)[2], uint32_t *n_regs)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	uint32_t i = 0, reg;
 #define HQD_N_REGS 56
 #define DUMP_REG(addr) do {				\
@@ -372,13 +353,13 @@ static int kgd_hqd_dump(struct kgd_dev *kgd,
 	if (*dump == NULL)
 		return -ENOMEM;
 
-	acquire_queue(kgd, pipe_id, queue_id);
+	acquire_queue(adev, pipe_id, queue_id);
 
 	for (reg = SOC15_REG_OFFSET(GC, 0, mmCP_MQD_BASE_ADDR);
 	     reg <= SOC15_REG_OFFSET(GC, 0, mmCP_HQD_PQ_WPTR_HI); reg++)
 		DUMP_REG(reg);
 
-	release_queue(kgd);
+	release_queue(adev);
 
 	WARN_ON_ONCE(i != HQD_N_REGS);
 	*n_regs = i;
@@ -386,10 +367,9 @@ static int kgd_hqd_dump(struct kgd_dev *kgd,
 	return 0;
 }
 
-static int kgd_hqd_sdma_load(struct kgd_dev *kgd, void *mqd,
+static int kgd_hqd_sdma_load(struct amdgpu_device *adev, void *mqd,
 			     uint32_t __user *wptr, struct mm_struct *mm)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	struct v10_sdma_mqd *m;
 	uint32_t sdma_rlc_reg_offset;
 	unsigned long end_jiffies;
@@ -456,11 +436,10 @@ static int kgd_hqd_sdma_load(struct kgd_dev *kgd, void *mqd,
 	return 0;
 }
 
-static int kgd_hqd_sdma_dump(struct kgd_dev *kgd,
+static int kgd_hqd_sdma_dump(struct amdgpu_device *adev,
 			     uint32_t engine_id, uint32_t queue_id,
 			     uint32_t (**dump)[2], uint32_t *n_regs)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	uint32_t sdma_rlc_reg_offset = get_sdma_rlc_reg_offset(adev,
 			engine_id, queue_id);
 	uint32_t i = 0, reg;
@@ -488,15 +467,15 @@ static int kgd_hqd_sdma_dump(struct kgd_dev *kgd,
 	return 0;
 }
 
-static bool kgd_hqd_is_occupied(struct kgd_dev *kgd, uint64_t queue_address,
-				uint32_t pipe_id, uint32_t queue_id)
+static bool kgd_hqd_is_occupied(struct amdgpu_device *adev,
+				uint64_t queue_address, uint32_t pipe_id,
+				uint32_t queue_id)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	uint32_t act;
 	bool retval = false;
 	uint32_t low, high;
 
-	acquire_queue(kgd, pipe_id, queue_id);
+	acquire_queue(adev, pipe_id, queue_id);
 	act = RREG32_SOC15(GC, 0, mmCP_HQD_ACTIVE);
 	if (act) {
 		low = lower_32_bits(queue_address >> 8);
@@ -506,13 +485,12 @@ static bool kgd_hqd_is_occupied(struct kgd_dev *kgd, uint64_t queue_address,
 		   high == RREG32_SOC15(GC, 0, mmCP_HQD_PQ_BASE_HI))
 			retval = true;
 	}
-	release_queue(kgd);
+	release_queue(adev);
 	return retval;
 }
 
-static bool kgd_hqd_sdma_is_occupied(struct kgd_dev *kgd, void *mqd)
+static bool kgd_hqd_sdma_is_occupied(struct amdgpu_device *adev, void *mqd)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	struct v10_sdma_mqd *m;
 	uint32_t sdma_rlc_reg_offset;
 	uint32_t sdma_rlc_rb_cntl;
@@ -529,12 +507,11 @@ static bool kgd_hqd_sdma_is_occupied(struct kgd_dev *kgd, void *mqd)
 	return false;
 }
 
-static int kgd_hqd_destroy(struct kgd_dev *kgd, void *mqd,
+static int kgd_hqd_destroy(struct amdgpu_device *adev, void *mqd,
 				enum kfd_preempt_type reset_type,
 				unsigned int utimeout, uint32_t pipe_id,
 				uint32_t queue_id)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	enum hqd_dequeue_request_type type;
 	unsigned long end_jiffies;
 	uint32_t temp;
@@ -548,7 +525,7 @@ static int kgd_hqd_destroy(struct kgd_dev *kgd, void *mqd,
 	int retry;
 #endif
 
-	acquire_queue(kgd, pipe_id, queue_id);
+	acquire_queue(adev, pipe_id, queue_id);
 
 	if (m->cp_hqd_vmid == 0)
 		WREG32_FIELD15(GC, 0, RLC_CP_SCHEDULERS, scheduler1, 0);
@@ -633,20 +610,19 @@ loop:
 			break;
 		if (time_after(jiffies, end_jiffies)) {
 			pr_err("cp queue preemption time out.\n");
-			release_queue(kgd);
+			release_queue(adev);
 			return -ETIME;
 		}
 		usleep_range(500, 1000);
 	}
 
-	release_queue(kgd);
+	release_queue(adev);
 	return 0;
 }
 
-static int kgd_hqd_sdma_destroy(struct kgd_dev *kgd, void *mqd,
+static int kgd_hqd_sdma_destroy(struct amdgpu_device *adev, void *mqd,
 				unsigned int utimeout)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	struct v10_sdma_mqd *m;
 	uint32_t sdma_rlc_reg_offset;
 	uint32_t temp;
@@ -683,11 +659,10 @@ static int kgd_hqd_sdma_destroy(struct kgd_dev *kgd, void *mqd,
 	return 0;
 }
 
-static bool get_atc_vmid_pasid_mapping_info(struct kgd_dev *kgd,
+static bool get_atc_vmid_pasid_mapping_info(struct amdgpu_device *adev,
 					uint8_t vmid, uint16_t *p_pasid)
 {
 	uint32_t value;
-	struct amdgpu_device *adev = (struct amdgpu_device *) kgd;
 
 	value = RREG32(SOC15_REG_OFFSET(ATHUB, 0, mmATC_VMID0_PASID_MAPPING)
 		     + vmid);
@@ -696,12 +671,12 @@ static bool get_atc_vmid_pasid_mapping_info(struct kgd_dev *kgd,
 	return !!(value & ATC_VMID0_PASID_MAPPING__VALID_MASK);
 }
 
-static int kgd_address_watch_disable(struct kgd_dev *kgd)
+static int kgd_address_watch_disable(struct amdgpu_device *adev)
 {
 	return 0;
 }
 
-static int kgd_address_watch_execute(struct kgd_dev *kgd,
+static int kgd_address_watch_execute(struct amdgpu_device *adev,
 					unsigned int watch_point_id,
 					uint32_t cntl_val,
 					uint32_t addr_hi,
@@ -710,11 +685,10 @@ static int kgd_address_watch_execute(struct kgd_dev *kgd,
 	return 0;
 }
 
-static int kgd_wave_control_execute(struct kgd_dev *kgd,
+static int kgd_wave_control_execute(struct amdgpu_device *adev,
 					uint32_t gfx_index_val,
 					uint32_t sq_cmd)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	uint32_t data = 0;
 
 	mutex_lock(&adev->grbm_idx_mutex);
@@ -735,18 +709,16 @@ static int kgd_wave_control_execute(struct kgd_dev *kgd,
 	return 0;
 }
 
-static uint32_t kgd_address_watch_get_offset(struct kgd_dev *kgd,
+static uint32_t kgd_address_watch_get_offset(struct amdgpu_device *adev,
 					unsigned int watch_point_id,
 					unsigned int reg_offset)
 {
 	return 0;
 }
 
-static void set_vm_context_page_table_base(struct kgd_dev *kgd, uint32_t vmid,
-		uint64_t page_table_base)
+static void set_vm_context_page_table_base(struct amdgpu_device *adev,
+		uint32_t vmid, uint64_t page_table_base)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
-
 	if (!amdgpu_amdkfd_is_kfd_vmid(adev, vmid)) {
 		pr_err("trying to set page table base for wrong VMID %u\n",
 		       vmid);
@@ -757,12 +729,10 @@ static void set_vm_context_page_table_base(struct kgd_dev *kgd, uint32_t vmid,
 	adev->gfxhub.funcs->setup_vm_pt_regs(adev, vmid, page_table_base);
 }
 
-static void program_trap_handler_settings(struct kgd_dev *kgd,
+static void program_trap_handler_settings(struct amdgpu_device *adev,
 		uint32_t vmid, uint64_t tba_addr, uint64_t tma_addr)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
-
-	lock_srbm(kgd, 0, 0, 0, vmid);
+	lock_srbm(adev, 0, 0, 0, vmid);
 
 	/*
 	 * Program TBA registers
@@ -781,7 +751,7 @@ static void program_trap_handler_settings(struct kgd_dev *kgd,
 	WREG32(SOC15_REG_OFFSET(GC, 0, mmSQ_SHADER_TMA_HI),
 			upper_32_bits(tma_addr >> 8));
 
-	unlock_srbm(kgd);
+	unlock_srbm(adev);
 }
 
 const struct kfd2kgd_calls gfx_v10_kfd2kgd = {
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v10_3.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v10_3.c
index dac0d751d5af..1f37d3574001 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v10_3.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v10_3.c
@@ -38,37 +38,26 @@ enum hqd_dequeue_request_type {
 	SAVE_WAVES
 };
 
-static inline struct amdgpu_device *get_amdgpu_device(struct kgd_dev *kgd)
-{
-	return (struct amdgpu_device *)kgd;
-}
-
-static void lock_srbm(struct kgd_dev *kgd, uint32_t mec, uint32_t pipe,
+static void lock_srbm(struct amdgpu_device *adev, uint32_t mec, uint32_t pipe,
 			uint32_t queue, uint32_t vmid)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
-
 	mutex_lock(&adev->srbm_mutex);
 	nv_grbm_select(adev, mec, pipe, queue, vmid);
 }
 
-static void unlock_srbm(struct kgd_dev *kgd)
+static void unlock_srbm(struct amdgpu_device *adev)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
-
 	nv_grbm_select(adev, 0, 0, 0, 0);
 	mutex_unlock(&adev->srbm_mutex);
 }
 
-static void acquire_queue(struct kgd_dev *kgd, uint32_t pipe_id,
+static void acquire_queue(struct amdgpu_device *adev, uint32_t pipe_id,
 				uint32_t queue_id)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
-
 	uint32_t mec = (pipe_id / adev->gfx.mec.num_pipe_per_mec) + 1;
 	uint32_t pipe = (pipe_id % adev->gfx.mec.num_pipe_per_mec);
 
-	lock_srbm(kgd, mec, pipe, queue_id, 0);
+	lock_srbm(adev, mec, pipe, queue_id, 0);
 }
 
 static uint64_t get_queue_mask(struct amdgpu_device *adev,
@@ -80,34 +69,30 @@ static uint64_t get_queue_mask(struct amdgpu_device *adev,
 	return 1ull << bit;
 }
 
-static void release_queue(struct kgd_dev *kgd)
+static void release_queue(struct amdgpu_device *adev)
 {
-	unlock_srbm(kgd);
+	unlock_srbm(adev);
 }
 
-static void program_sh_mem_settings_v10_3(struct kgd_dev *kgd, uint32_t vmid,
+static void program_sh_mem_settings_v10_3(struct amdgpu_device *adev, uint32_t vmid,
 					uint32_t sh_mem_config,
 					uint32_t sh_mem_ape1_base,
 					uint32_t sh_mem_ape1_limit,
 					uint32_t sh_mem_bases)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
-
-	lock_srbm(kgd, 0, 0, 0, vmid);
+	lock_srbm(adev, 0, 0, 0, vmid);
 
 	WREG32_SOC15(GC, 0, mmSH_MEM_CONFIG, sh_mem_config);
 	WREG32_SOC15(GC, 0, mmSH_MEM_BASES, sh_mem_bases);
 	/* APE1 no longer exists on GFX9 */
 
-	unlock_srbm(kgd);
+	unlock_srbm(adev);
 }
 
 /* ATC is defeatured on Sienna_Cichlid */
-static int set_pasid_vmid_mapping_v10_3(struct kgd_dev *kgd, unsigned int pasid,
+static int set_pasid_vmid_mapping_v10_3(struct amdgpu_device *adev, unsigned int pasid,
 					unsigned int vmid)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
-
 	uint32_t value = pasid << IH_VMID_0_LUT__PASID__SHIFT;
 
 	/* Mapping vmid to pasid also for IH block */
@@ -118,22 +103,21 @@ static int set_pasid_vmid_mapping_v10_3(struct kgd_dev *kgd, unsigned int pasid,
 	return 0;
 }
 
-static int init_interrupts_v10_3(struct kgd_dev *kgd, uint32_t pipe_id)
+static int init_interrupts_v10_3(struct amdgpu_device *adev, uint32_t pipe_id)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	uint32_t mec;
 	uint32_t pipe;
 
 	mec = (pipe_id / adev->gfx.mec.num_pipe_per_mec) + 1;
 	pipe = (pipe_id % adev->gfx.mec.num_pipe_per_mec);
 
-	lock_srbm(kgd, mec, pipe, 0, 0);
+	lock_srbm(adev, mec, pipe, 0, 0);
 
 	WREG32_SOC15(GC, 0, mmCPC_INT_CNTL,
 		CP_INT_CNTL_RING0__TIME_STAMP_INT_ENABLE_MASK |
 		CP_INT_CNTL_RING0__OPCODE_ERROR_INT_ENABLE_MASK);
 
-	unlock_srbm(kgd);
+	unlock_srbm(adev);
 
 	return 0;
 }
@@ -188,12 +172,11 @@ static inline struct v10_sdma_mqd *get_sdma_mqd(void *mqd)
 	return (struct v10_sdma_mqd *)mqd;
 }
 
-static int hqd_load_v10_3(struct kgd_dev *kgd, void *mqd, uint32_t pipe_id,
-			uint32_t queue_id, uint32_t __user *wptr,
-			uint32_t wptr_shift, uint32_t wptr_mask,
-			struct mm_struct *mm)
+static int hqd_load_v10_3(struct amdgpu_device *adev, void *mqd,
+			uint32_t pipe_id, uint32_t queue_id,
+			uint32_t __user *wptr, uint32_t wptr_shift,
+			uint32_t wptr_mask, struct mm_struct *mm)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	struct v10_compute_mqd *m;
 	uint32_t *mqd_hqd;
 	uint32_t reg, hqd_base, data;
@@ -201,7 +184,7 @@ static int hqd_load_v10_3(struct kgd_dev *kgd, void *mqd, uint32_t pipe_id,
 	m = get_mqd(mqd);
 
 	pr_debug("Load hqd of pipe %d queue %d\n", pipe_id, queue_id);
-	acquire_queue(kgd, pipe_id, queue_id);
+	acquire_queue(adev, pipe_id, queue_id);
 
 	/* HIQ is set during driver init period with vmid set to 0*/
 	if (m->cp_hqd_vmid == 0) {
@@ -281,16 +264,15 @@ static int hqd_load_v10_3(struct kgd_dev *kgd, void *mqd, uint32_t pipe_id,
 	data = REG_SET_FIELD(m->cp_hqd_active, CP_HQD_ACTIVE, ACTIVE, 1);
 	WREG32_SOC15(GC, 0, mmCP_HQD_ACTIVE, data);
 
-	release_queue(kgd);
+	release_queue(adev);
 
 	return 0;
 }
 
-static int hiq_mqd_load_v10_3(struct kgd_dev *kgd, void *mqd,
+static int hiq_mqd_load_v10_3(struct amdgpu_device *adev, void *mqd,
 			    uint32_t pipe_id, uint32_t queue_id,
 			    uint32_t doorbell_off)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	struct amdgpu_ring *kiq_ring = &adev->gfx.kiq.ring;
 	struct v10_compute_mqd *m;
 	uint32_t mec, pipe;
@@ -298,7 +280,7 @@ static int hiq_mqd_load_v10_3(struct kgd_dev *kgd, void *mqd,
 
 	m = get_mqd(mqd);
 
-	acquire_queue(kgd, pipe_id, queue_id);
+	acquire_queue(adev, pipe_id, queue_id);
 
 	mec = (pipe_id / adev->gfx.mec.num_pipe_per_mec) + 1;
 	pipe = (pipe_id % adev->gfx.mec.num_pipe_per_mec);
@@ -334,16 +316,15 @@ static int hiq_mqd_load_v10_3(struct kgd_dev *kgd, void *mqd,
 
 out_unlock:
 	spin_unlock(&adev->gfx.kiq.ring_lock);
-	release_queue(kgd);
+	release_queue(adev);
 
 	return r;
 }
 
-static int hqd_dump_v10_3(struct kgd_dev *kgd,
+static int hqd_dump_v10_3(struct amdgpu_device *adev,
 			uint32_t pipe_id, uint32_t queue_id,
 			uint32_t (**dump)[2], uint32_t *n_regs)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	uint32_t i = 0, reg;
 #define HQD_N_REGS 56
 #define DUMP_REG(addr) do {				\
@@ -357,13 +338,13 @@ static int hqd_dump_v10_3(struct kgd_dev *kgd,
 	if (*dump == NULL)
 		return -ENOMEM;
 
-	acquire_queue(kgd, pipe_id, queue_id);
+	acquire_queue(adev, pipe_id, queue_id);
 
 	for (reg = SOC15_REG_OFFSET(GC, 0, mmCP_MQD_BASE_ADDR);
 	     reg <= SOC15_REG_OFFSET(GC, 0, mmCP_HQD_PQ_WPTR_HI); reg++)
 		DUMP_REG(reg);
 
-	release_queue(kgd);
+	release_queue(adev);
 
 	WARN_ON_ONCE(i != HQD_N_REGS);
 	*n_regs = i;
@@ -371,10 +352,9 @@ static int hqd_dump_v10_3(struct kgd_dev *kgd,
 	return 0;
 }
 
-static int hqd_sdma_load_v10_3(struct kgd_dev *kgd, void *mqd,
+static int hqd_sdma_load_v10_3(struct amdgpu_device *adev, void *mqd,
 			     uint32_t __user *wptr, struct mm_struct *mm)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	struct v10_sdma_mqd *m;
 	uint32_t sdma_rlc_reg_offset;
 	unsigned long end_jiffies;
@@ -441,11 +421,10 @@ static int hqd_sdma_load_v10_3(struct kgd_dev *kgd, void *mqd,
 	return 0;
 }
 
-static int hqd_sdma_dump_v10_3(struct kgd_dev *kgd,
+static int hqd_sdma_dump_v10_3(struct amdgpu_device *adev,
 			     uint32_t engine_id, uint32_t queue_id,
 			     uint32_t (**dump)[2], uint32_t *n_regs)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	uint32_t sdma_rlc_reg_offset = get_sdma_rlc_reg_offset(adev,
 			engine_id, queue_id);
 	uint32_t i = 0, reg;
@@ -473,15 +452,15 @@ static int hqd_sdma_dump_v10_3(struct kgd_dev *kgd,
 	return 0;
 }
 
-static bool hqd_is_occupied_v10_3(struct kgd_dev *kgd, uint64_t queue_address,
-				uint32_t pipe_id, uint32_t queue_id)
+static bool hqd_is_occupied_v10_3(struct amdgpu_device *adev,
+				uint64_t queue_address, uint32_t pipe_id,
+				uint32_t queue_id)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	uint32_t act;
 	bool retval = false;
 	uint32_t low, high;
 
-	acquire_queue(kgd, pipe_id, queue_id);
+	acquire_queue(adev, pipe_id, queue_id);
 	act = RREG32_SOC15(GC, 0, mmCP_HQD_ACTIVE);
 	if (act) {
 		low = lower_32_bits(queue_address >> 8);
@@ -491,13 +470,13 @@ static bool hqd_is_occupied_v10_3(struct kgd_dev *kgd, uint64_t queue_address,
 		   high == RREG32_SOC15(GC, 0, mmCP_HQD_PQ_BASE_HI))
 			retval = true;
 	}
-	release_queue(kgd);
+	release_queue(adev);
 	return retval;
 }
 
-static bool hqd_sdma_is_occupied_v10_3(struct kgd_dev *kgd, void *mqd)
+static bool hqd_sdma_is_occupied_v10_3(struct amdgpu_device *adev,
+				void *mqd)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	struct v10_sdma_mqd *m;
 	uint32_t sdma_rlc_reg_offset;
 	uint32_t sdma_rlc_rb_cntl;
@@ -514,18 +493,17 @@ static bool hqd_sdma_is_occupied_v10_3(struct kgd_dev *kgd, void *mqd)
 	return false;
 }
 
-static int hqd_destroy_v10_3(struct kgd_dev *kgd, void *mqd,
+static int hqd_destroy_v10_3(struct amdgpu_device *adev, void *mqd,
 				enum kfd_preempt_type reset_type,
 				unsigned int utimeout, uint32_t pipe_id,
 				uint32_t queue_id)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	enum hqd_dequeue_request_type type;
 	unsigned long end_jiffies;
 	uint32_t temp;
 	struct v10_compute_mqd *m = get_mqd(mqd);
 
-	acquire_queue(kgd, pipe_id, queue_id);
+	acquire_queue(adev, pipe_id, queue_id);
 
 	if (m->cp_hqd_vmid == 0)
 		WREG32_FIELD15(GC, 0, RLC_CP_SCHEDULERS, scheduler1, 0);
@@ -555,20 +533,19 @@ static int hqd_destroy_v10_3(struct kgd_dev *kgd, void *mqd,
 		if (time_after(jiffies, end_jiffies)) {
 			pr_err("cp queue pipe %d queue %d preemption failed\n",
 					pipe_id, queue_id);
-			release_queue(kgd);
+			release_queue(adev);
 			return -ETIME;
 		}
 		usleep_range(500, 1000);
 	}
 
-	release_queue(kgd);
+	release_queue(adev);
 	return 0;
 }
 
-static int hqd_sdma_destroy_v10_3(struct kgd_dev *kgd, void *mqd,
+static int hqd_sdma_destroy_v10_3(struct amdgpu_device *adev, void *mqd,
 				unsigned int utimeout)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	struct v10_sdma_mqd *m;
 	uint32_t sdma_rlc_reg_offset;
 	uint32_t temp;
@@ -606,12 +583,12 @@ static int hqd_sdma_destroy_v10_3(struct kgd_dev *kgd, void *mqd,
 }
 
 
-static int address_watch_disable_v10_3(struct kgd_dev *kgd)
+static int address_watch_disable_v10_3(struct amdgpu_device *adev)
 {
 	return 0;
 }
 
-static int address_watch_execute_v10_3(struct kgd_dev *kgd,
+static int address_watch_execute_v10_3(struct amdgpu_device *adev,
 					unsigned int watch_point_id,
 					uint32_t cntl_val,
 					uint32_t addr_hi,
@@ -620,11 +597,10 @@ static int address_watch_execute_v10_3(struct kgd_dev *kgd,
 	return 0;
 }
 
-static int wave_control_execute_v10_3(struct kgd_dev *kgd,
+static int wave_control_execute_v10_3(struct amdgpu_device *adev,
 					uint32_t gfx_index_val,
 					uint32_t sq_cmd)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	uint32_t data = 0;
 
 	mutex_lock(&adev->grbm_idx_mutex);
@@ -645,28 +621,24 @@ static int wave_control_execute_v10_3(struct kgd_dev *kgd,
 	return 0;
 }
 
-static uint32_t address_watch_get_offset_v10_3(struct kgd_dev *kgd,
+static uint32_t address_watch_get_offset_v10_3(struct amdgpu_device *adev,
 					unsigned int watch_point_id,
 					unsigned int reg_offset)
 {
 	return 0;
 }
 
-static void set_vm_context_page_table_base_v10_3(struct kgd_dev *kgd, uint32_t vmid,
-		uint64_t page_table_base)
+static void set_vm_context_page_table_base_v10_3(struct amdgpu_device *adev,
+		uint32_t vmid, uint64_t page_table_base)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
-
 	/* SDMA is on gfxhub as well for Navi1* series */
 	adev->gfxhub.funcs->setup_vm_pt_regs(adev, vmid, page_table_base);
 }
 
-static void program_trap_handler_settings_v10_3(struct kgd_dev *kgd,
+static void program_trap_handler_settings_v10_3(struct amdgpu_device *adev,
 			uint32_t vmid, uint64_t tba_addr, uint64_t tma_addr)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
-
-	lock_srbm(kgd, 0, 0, 0, vmid);
+	lock_srbm(adev, 0, 0, 0, vmid);
 
 	/*
 	 * Program TBA registers
@@ -685,15 +657,14 @@ static void program_trap_handler_settings_v10_3(struct kgd_dev *kgd,
 	WREG32(SOC15_REG_OFFSET(GC, 0, mmSQ_SHADER_TMA_HI),
 			 upper_32_bits(tma_addr >> 8));
 
-	unlock_srbm(kgd);
+	unlock_srbm(adev);
 }
 
 #if 0
-uint32_t enable_debug_trap_v10_3(struct kgd_dev *kgd,
+uint32_t enable_debug_trap_v10_3(struct amdgpu_device *adev,
 				uint32_t trap_debug_wave_launch_mode,
 				uint32_t vmid)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	uint32_t data = 0;
 	uint32_t orig_wave_cntl_value;
 	uint32_t orig_stall_vmid;
@@ -720,10 +691,8 @@ uint32_t enable_debug_trap_v10_3(struct kgd_dev *kgd,
 	return 0;
 }
 
-uint32_t disable_debug_trap_v10_3(struct kgd_dev *kgd)
+uint32_t disable_debug_trap_v10_3(struct amdgpu_device *adev)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
-
 	mutex_lock(&adev->grbm_idx_mutex);
 
 	WREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_GDBG_TRAP_MASK), 0);
@@ -733,11 +702,10 @@ uint32_t disable_debug_trap_v10_3(struct kgd_dev *kgd)
 	return 0;
 }
 
-uint32_t set_wave_launch_trap_override_v10_3(struct kgd_dev *kgd,
+uint32_t set_wave_launch_trap_override_v10_3(struct amdgpu_device *adev,
 						uint32_t trap_override,
 						uint32_t trap_mask)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	uint32_t data = 0;
 
 	mutex_lock(&adev->grbm_idx_mutex);
@@ -762,11 +730,10 @@ uint32_t set_wave_launch_trap_override_v10_3(struct kgd_dev *kgd,
 	return 0;
 }
 
-uint32_t set_wave_launch_mode_v10_3(struct kgd_dev *kgd,
+uint32_t set_wave_launch_mode_v10_3(struct amdgpu_device *adev,
 					uint8_t wave_launch_mode,
 					uint32_t vmid)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	uint32_t data = 0;
 	bool is_stall_mode;
 	bool is_mode_set;
@@ -805,16 +772,14 @@ uint32_t set_wave_launch_mode_v10_3(struct kgd_dev *kgd,
  *	sem_rearm_wait_time      -- Wait Count for Semaphore re-arm.
  *	deq_retry_wait_time      -- Wait Count for Global Wave Syncs.
  */
-void get_iq_wait_times_v10_3(struct kgd_dev *kgd,
+void get_iq_wait_times_v10_3(struct amdgpu_device *adev,
 					uint32_t *wait_times)
 
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
-
 	*wait_times = RREG32(SOC15_REG_OFFSET(GC, 0, mmCP_IQ_WAIT_TIME2));
 }
 
-void build_grace_period_packet_info_v10_3(struct kgd_dev *kgd,
+void build_grace_period_packet_info_v10_3(struct amdgpu_device *adev,
 						uint32_t wait_times,
 						uint32_t grace_period,
 						uint32_t *reg_offset,
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v7.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v7.c
index b91d27e39bad..36528dad7684 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v7.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v7.c
@@ -82,68 +82,54 @@ union TCP_WATCH_CNTL_BITS {
 	float f32All;
 };
 
-static inline struct amdgpu_device *get_amdgpu_device(struct kgd_dev *kgd)
-{
-	return (struct amdgpu_device *)kgd;
-}
-
-static void lock_srbm(struct kgd_dev *kgd, uint32_t mec, uint32_t pipe,
+static void lock_srbm(struct amdgpu_device *adev, uint32_t mec, uint32_t pipe,
 			uint32_t queue, uint32_t vmid)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	uint32_t value = PIPEID(pipe) | MEID(mec) | VMID(vmid) | QUEUEID(queue);
 
 	mutex_lock(&adev->srbm_mutex);
 	WREG32(mmSRBM_GFX_CNTL, value);
 }
 
-static void unlock_srbm(struct kgd_dev *kgd)
+static void unlock_srbm(struct amdgpu_device *adev)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
-
 	WREG32(mmSRBM_GFX_CNTL, 0);
 	mutex_unlock(&adev->srbm_mutex);
 }
 
-static void acquire_queue(struct kgd_dev *kgd, uint32_t pipe_id,
+static void acquire_queue(struct amdgpu_device *adev, uint32_t pipe_id,
 				uint32_t queue_id)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
-
 	uint32_t mec = (pipe_id / adev->gfx.mec.num_pipe_per_mec) + 1;
 	uint32_t pipe = (pipe_id % adev->gfx.mec.num_pipe_per_mec);
 
-	lock_srbm(kgd, mec, pipe, queue_id, 0);
+	lock_srbm(adev, mec, pipe, queue_id, 0);
 }
 
-static void release_queue(struct kgd_dev *kgd)
+static void release_queue(struct amdgpu_device *adev)
 {
-	unlock_srbm(kgd);
+	unlock_srbm(adev);
 }
 
-static void kgd_program_sh_mem_settings(struct kgd_dev *kgd, uint32_t vmid,
+static void kgd_program_sh_mem_settings(struct amdgpu_device *adev, uint32_t vmid,
 					uint32_t sh_mem_config,
 					uint32_t sh_mem_ape1_base,
 					uint32_t sh_mem_ape1_limit,
 					uint32_t sh_mem_bases)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
-
-	lock_srbm(kgd, 0, 0, 0, vmid);
+	lock_srbm(adev, 0, 0, 0, vmid);
 
 	WREG32(mmSH_MEM_CONFIG, sh_mem_config);
 	WREG32(mmSH_MEM_APE1_BASE, sh_mem_ape1_base);
 	WREG32(mmSH_MEM_APE1_LIMIT, sh_mem_ape1_limit);
 	WREG32(mmSH_MEM_BASES, sh_mem_bases);
 
-	unlock_srbm(kgd);
+	unlock_srbm(adev);
 }
 
-static int kgd_set_pasid_vmid_mapping(struct kgd_dev *kgd, u32 pasid,
+static int kgd_set_pasid_vmid_mapping(struct amdgpu_device *adev, u32 pasid,
 					unsigned int vmid)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
-
 	/*
 	 * We have to assume that there is no outstanding mapping.
 	 * The ATC_VMID_PASID_MAPPING_UPDATE_STATUS bit could be 0 because
@@ -165,21 +151,20 @@ static int kgd_set_pasid_vmid_mapping(struct kgd_dev *kgd, u32 pasid,
 	return 0;
 }
 
-static int kgd_init_interrupts(struct kgd_dev *kgd, uint32_t pipe_id)
+static int kgd_init_interrupts(struct amdgpu_device *adev, uint32_t pipe_id)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	uint32_t mec;
 	uint32_t pipe;
 
 	mec = (pipe_id / adev->gfx.mec.num_pipe_per_mec) + 1;
 	pipe = (pipe_id % adev->gfx.mec.num_pipe_per_mec);
 
-	lock_srbm(kgd, mec, pipe, 0, 0);
+	lock_srbm(adev, mec, pipe, 0, 0);
 
 	WREG32(mmCPC_INT_CNTL, CP_INT_CNTL_RING0__TIME_STAMP_INT_ENABLE_MASK |
 			CP_INT_CNTL_RING0__OPCODE_ERROR_INT_ENABLE_MASK);
 
-	unlock_srbm(kgd);
+	unlock_srbm(adev);
 
 	return 0;
 }
@@ -207,12 +192,11 @@ static inline struct cik_sdma_rlc_registers *get_sdma_mqd(void *mqd)
 	return (struct cik_sdma_rlc_registers *)mqd;
 }
 
-static int kgd_hqd_load(struct kgd_dev *kgd, void *mqd, uint32_t pipe_id,
-			uint32_t queue_id, uint32_t __user *wptr,
-			uint32_t wptr_shift, uint32_t wptr_mask,
-			struct mm_struct *mm)
+static int kgd_hqd_load(struct amdgpu_device *adev, void *mqd,
+			uint32_t pipe_id, uint32_t queue_id,
+			uint32_t __user *wptr, uint32_t wptr_shift,
+			uint32_t wptr_mask, struct mm_struct *mm)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	struct cik_mqd *m;
 	uint32_t *mqd_hqd;
 	uint32_t reg, wptr_val, data;
@@ -220,7 +204,7 @@ static int kgd_hqd_load(struct kgd_dev *kgd, void *mqd, uint32_t pipe_id,
 
 	m = get_mqd(mqd);
 
-	acquire_queue(kgd, pipe_id, queue_id);
+	acquire_queue(adev, pipe_id, queue_id);
 
 	/* HQD registers extend from CP_MQD_BASE_ADDR to CP_MQD_CONTROL. */
 	mqd_hqd = &m->cp_mqd_base_addr_lo;
@@ -239,25 +223,24 @@ static int kgd_hqd_load(struct kgd_dev *kgd, void *mqd, uint32_t pipe_id,
 	 * release srbm_mutex to avoid circular dependency between
 	 * srbm_mutex->mm_sem->reservation_ww_class_mutex->srbm_mutex.
 	 */
-	release_queue(kgd);
+	release_queue(adev);
 	valid_wptr = read_user_wptr(mm, wptr, wptr_val);
-	acquire_queue(kgd, pipe_id, queue_id);
+	acquire_queue(adev, pipe_id, queue_id);
 	if (valid_wptr)
 		WREG32(mmCP_HQD_PQ_WPTR, (wptr_val << wptr_shift) & wptr_mask);
 
 	data = REG_SET_FIELD(m->cp_hqd_active, CP_HQD_ACTIVE, ACTIVE, 1);
 	WREG32(mmCP_HQD_ACTIVE, data);
 
-	release_queue(kgd);
+	release_queue(adev);
 
 	return 0;
 }
 
-static int kgd_hqd_dump(struct kgd_dev *kgd,
+static int kgd_hqd_dump(struct amdgpu_device *adev,
 			uint32_t pipe_id, uint32_t queue_id,
 			uint32_t (**dump)[2], uint32_t *n_regs)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	uint32_t i = 0, reg;
 #define HQD_N_REGS (35+4)
 #define DUMP_REG(addr) do {				\
@@ -271,7 +254,7 @@ static int kgd_hqd_dump(struct kgd_dev *kgd,
 	if (*dump == NULL)
 		return -ENOMEM;
 
-	acquire_queue(kgd, pipe_id, queue_id);
+	acquire_queue(adev, pipe_id, queue_id);
 
 	DUMP_REG(mmCOMPUTE_STATIC_THREAD_MGMT_SE0);
 	DUMP_REG(mmCOMPUTE_STATIC_THREAD_MGMT_SE1);
@@ -281,7 +264,7 @@ static int kgd_hqd_dump(struct kgd_dev *kgd,
 	for (reg = mmCP_MQD_BASE_ADDR; reg <= mmCP_MQD_CONTROL; reg++)
 		DUMP_REG(reg);
 
-	release_queue(kgd);
+	release_queue(adev);
 
 	WARN_ON_ONCE(i != HQD_N_REGS);
 	*n_regs = i;
@@ -289,10 +272,9 @@ static int kgd_hqd_dump(struct kgd_dev *kgd,
 	return 0;
 }
 
-static int kgd_hqd_sdma_load(struct kgd_dev *kgd, void *mqd,
+static int kgd_hqd_sdma_load(struct amdgpu_device *adev, void *mqd,
 			     uint32_t __user *wptr, struct mm_struct *mm)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	struct cik_sdma_rlc_registers *m;
 	unsigned long end_jiffies;
 	uint32_t sdma_rlc_reg_offset;
@@ -345,11 +327,10 @@ static int kgd_hqd_sdma_load(struct kgd_dev *kgd, void *mqd,
 	return 0;
 }
 
-static int kgd_hqd_sdma_dump(struct kgd_dev *kgd,
+static int kgd_hqd_sdma_dump(struct amdgpu_device *adev,
 			     uint32_t engine_id, uint32_t queue_id,
 			     uint32_t (**dump)[2], uint32_t *n_regs)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	uint32_t sdma_offset = engine_id * SDMA1_REGISTER_OFFSET +
 		queue_id * KFD_CIK_SDMA_QUEUE_OFFSET;
 	uint32_t i = 0, reg;
@@ -372,15 +353,15 @@ static int kgd_hqd_sdma_dump(struct kgd_dev *kgd,
 	return 0;
 }
 
-static bool kgd_hqd_is_occupied(struct kgd_dev *kgd, uint64_t queue_address,
-				uint32_t pipe_id, uint32_t queue_id)
+static bool kgd_hqd_is_occupied(struct amdgpu_device *adev,
+				uint64_t queue_address, uint32_t pipe_id,
+				uint32_t queue_id)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	uint32_t act;
 	bool retval = false;
 	uint32_t low, high;
 
-	acquire_queue(kgd, pipe_id, queue_id);
+	acquire_queue(adev, pipe_id, queue_id);
 	act = RREG32(mmCP_HQD_ACTIVE);
 	if (act) {
 		low = lower_32_bits(queue_address >> 8);
@@ -390,13 +371,12 @@ static bool kgd_hqd_is_occupied(struct kgd_dev *kgd, uint64_t queue_address,
 				high == RREG32(mmCP_HQD_PQ_BASE_HI))
 			retval = true;
 	}
-	release_queue(kgd);
+	release_queue(adev);
 	return retval;
 }
 
-static bool kgd_hqd_sdma_is_occupied(struct kgd_dev *kgd, void *mqd)
+static bool kgd_hqd_sdma_is_occupied(struct amdgpu_device *adev, void *mqd)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	struct cik_sdma_rlc_registers *m;
 	uint32_t sdma_rlc_reg_offset;
 	uint32_t sdma_rlc_rb_cntl;
@@ -412,12 +392,11 @@ static bool kgd_hqd_sdma_is_occupied(struct kgd_dev *kgd, void *mqd)
 	return false;
 }
 
-static int kgd_hqd_destroy(struct kgd_dev *kgd, void *mqd,
+static int kgd_hqd_destroy(struct amdgpu_device *adev, void *mqd,
 				enum kfd_preempt_type reset_type,
 				unsigned int utimeout, uint32_t pipe_id,
 				uint32_t queue_id)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	uint32_t temp;
 	enum hqd_dequeue_request_type type;
 	unsigned long flags, end_jiffies;
@@ -426,7 +405,7 @@ static int kgd_hqd_destroy(struct kgd_dev *kgd, void *mqd,
 	if (amdgpu_in_reset(adev))
 		return -EIO;
 
-	acquire_queue(kgd, pipe_id, queue_id);
+	acquire_queue(adev, pipe_id, queue_id);
 	WREG32(mmCP_HQD_PQ_DOORBELL_CONTROL, 0);
 
 	switch (reset_type) {
@@ -504,20 +483,19 @@ loop:
 			break;
 		if (time_after(jiffies, end_jiffies)) {
 			pr_err("cp queue preemption time out\n");
-			release_queue(kgd);
+			release_queue(adev);
 			return -ETIME;
 		}
 		usleep_range(500, 1000);
 	}
 
-	release_queue(kgd);
+	release_queue(adev);
 	return 0;
 }
 
-static int kgd_hqd_sdma_destroy(struct kgd_dev *kgd, void *mqd,
+static int kgd_hqd_sdma_destroy(struct amdgpu_device *adev, void *mqd,
 				unsigned int utimeout)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	struct cik_sdma_rlc_registers *m;
 	uint32_t sdma_rlc_reg_offset;
 	uint32_t temp;
@@ -551,9 +529,8 @@ static int kgd_hqd_sdma_destroy(struct kgd_dev *kgd, void *mqd,
 	return 0;
 }
 
-static int kgd_address_watch_disable(struct kgd_dev *kgd)
+static int kgd_address_watch_disable(struct amdgpu_device *adev)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	union TCP_WATCH_CNTL_BITS cntl;
 	unsigned int i;
 
@@ -571,13 +548,12 @@ static int kgd_address_watch_disable(struct kgd_dev *kgd)
 	return 0;
 }
 
-static int kgd_address_watch_execute(struct kgd_dev *kgd,
+static int kgd_address_watch_execute(struct amdgpu_device *adev,
 					unsigned int watch_point_id,
 					uint32_t cntl_val,
 					uint32_t addr_hi,
 					uint32_t addr_lo)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	union TCP_WATCH_CNTL_BITS cntl;
 
 	cntl.u32All = cntl_val;
@@ -602,11 +578,10 @@ static int kgd_address_watch_execute(struct kgd_dev *kgd,
 	return 0;
 }
 
-static int kgd_wave_control_execute(struct kgd_dev *kgd,
+static int kgd_wave_control_execute(struct amdgpu_device *adev,
 					uint32_t gfx_index_val,
 					uint32_t sq_cmd)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	uint32_t data;
 
 	mutex_lock(&adev->grbm_idx_mutex);
@@ -627,18 +602,17 @@ static int kgd_wave_control_execute(struct kgd_dev *kgd,
 	return 0;
 }
 
-static uint32_t kgd_address_watch_get_offset(struct kgd_dev *kgd,
+static uint32_t kgd_address_watch_get_offset(struct amdgpu_device *adev,
 					unsigned int watch_point_id,
 					unsigned int reg_offset)
 {
 	return watchRegs[watch_point_id * ADDRESS_WATCH_REG_MAX + reg_offset];
 }
 
-static bool get_atc_vmid_pasid_mapping_info(struct kgd_dev *kgd,
+static bool get_atc_vmid_pasid_mapping_info(struct amdgpu_device *adev,
 					uint8_t vmid, uint16_t *p_pasid)
 {
 	uint32_t value;
-	struct amdgpu_device *adev = (struct amdgpu_device *) kgd;
 
 	value = RREG32(mmATC_VMID0_PASID_MAPPING + vmid);
 	*p_pasid = value & ATC_VMID0_PASID_MAPPING__PASID_MASK;
@@ -646,21 +620,17 @@ static bool get_atc_vmid_pasid_mapping_info(struct kgd_dev *kgd,
 	return !!(value & ATC_VMID0_PASID_MAPPING__VALID_MASK);
 }
 
-static void set_scratch_backing_va(struct kgd_dev *kgd,
+static void set_scratch_backing_va(struct amdgpu_device *adev,
 					uint64_t va, uint32_t vmid)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *) kgd;
-
-	lock_srbm(kgd, 0, 0, 0, vmid);
+	lock_srbm(adev, 0, 0, 0, vmid);
 	WREG32(mmSH_HIDDEN_PRIVATE_BASE_VMID, va);
-	unlock_srbm(kgd);
+	unlock_srbm(adev);
 }
 
-static void set_vm_context_page_table_base(struct kgd_dev *kgd, uint32_t vmid,
-			uint64_t page_table_base)
+static void set_vm_context_page_table_base(struct amdgpu_device *adev,
+			uint32_t vmid, uint64_t page_table_base)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
-
 	if (!amdgpu_amdkfd_is_kfd_vmid(adev, vmid)) {
 		pr_err("trying to set page table base for wrong VMID\n");
 		return;
@@ -676,10 +646,8 @@ static void set_vm_context_page_table_base(struct kgd_dev *kgd, uint32_t vmid,
   * @vmid: vmid pointer
   * read vmid from register (CIK).
   */
-static uint32_t read_vmid_from_vmfault_reg(struct kgd_dev *kgd)
+static uint32_t read_vmid_from_vmfault_reg(struct amdgpu_device *adev)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
-
 	uint32_t status = RREG32(mmVM_CONTEXT1_PROTECTION_FAULT_STATUS);
 
 	return REG_GET_FIELD(status, VM_CONTEXT1_PROTECTION_FAULT_STATUS, VMID);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v8.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v8.c
index 5ce0ce704a21..52832cd69a93 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v8.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v8.c
@@ -39,68 +39,54 @@ enum hqd_dequeue_request_type {
 	RESET_WAVES
 };
 
-static inline struct amdgpu_device *get_amdgpu_device(struct kgd_dev *kgd)
-{
-	return (struct amdgpu_device *)kgd;
-}
-
-static void lock_srbm(struct kgd_dev *kgd, uint32_t mec, uint32_t pipe,
+static void lock_srbm(struct amdgpu_device *adev, uint32_t mec, uint32_t pipe,
 			uint32_t queue, uint32_t vmid)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	uint32_t value = PIPEID(pipe) | MEID(mec) | VMID(vmid) | QUEUEID(queue);
 
 	mutex_lock(&adev->srbm_mutex);
 	WREG32(mmSRBM_GFX_CNTL, value);
 }
 
-static void unlock_srbm(struct kgd_dev *kgd)
+static void unlock_srbm(struct amdgpu_device *adev)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
-
 	WREG32(mmSRBM_GFX_CNTL, 0);
 	mutex_unlock(&adev->srbm_mutex);
 }
 
-static void acquire_queue(struct kgd_dev *kgd, uint32_t pipe_id,
+static void acquire_queue(struct amdgpu_device *adev, uint32_t pipe_id,
 				uint32_t queue_id)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
-
 	uint32_t mec = (pipe_id / adev->gfx.mec.num_pipe_per_mec) + 1;
 	uint32_t pipe = (pipe_id % adev->gfx.mec.num_pipe_per_mec);
 
-	lock_srbm(kgd, mec, pipe, queue_id, 0);
+	lock_srbm(adev, mec, pipe, queue_id, 0);
 }
 
-static void release_queue(struct kgd_dev *kgd)
+static void release_queue(struct amdgpu_device *adev)
 {
-	unlock_srbm(kgd);
+	unlock_srbm(adev);
 }
 
-static void kgd_program_sh_mem_settings(struct kgd_dev *kgd, uint32_t vmid,
+static void kgd_program_sh_mem_settings(struct amdgpu_device *adev, uint32_t vmid,
 					uint32_t sh_mem_config,
 					uint32_t sh_mem_ape1_base,
 					uint32_t sh_mem_ape1_limit,
 					uint32_t sh_mem_bases)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
-
-	lock_srbm(kgd, 0, 0, 0, vmid);
+	lock_srbm(adev, 0, 0, 0, vmid);
 
 	WREG32(mmSH_MEM_CONFIG, sh_mem_config);
 	WREG32(mmSH_MEM_APE1_BASE, sh_mem_ape1_base);
 	WREG32(mmSH_MEM_APE1_LIMIT, sh_mem_ape1_limit);
 	WREG32(mmSH_MEM_BASES, sh_mem_bases);
 
-	unlock_srbm(kgd);
+	unlock_srbm(adev);
 }
 
-static int kgd_set_pasid_vmid_mapping(struct kgd_dev *kgd, u32 pasid,
+static int kgd_set_pasid_vmid_mapping(struct amdgpu_device *adev, u32 pasid,
 					unsigned int vmid)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
-
 	/*
 	 * We have to assume that there is no outstanding mapping.
 	 * The ATC_VMID_PASID_MAPPING_UPDATE_STATUS bit could be 0 because
@@ -123,21 +109,20 @@ static int kgd_set_pasid_vmid_mapping(struct kgd_dev *kgd, u32 pasid,
 	return 0;
 }
 
-static int kgd_init_interrupts(struct kgd_dev *kgd, uint32_t pipe_id)
+static int kgd_init_interrupts(struct amdgpu_device *adev, uint32_t pipe_id)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	uint32_t mec;
 	uint32_t pipe;
 
 	mec = (pipe_id / adev->gfx.mec.num_pipe_per_mec) + 1;
 	pipe = (pipe_id % adev->gfx.mec.num_pipe_per_mec);
 
-	lock_srbm(kgd, mec, pipe, 0, 0);
+	lock_srbm(adev, mec, pipe, 0, 0);
 
 	WREG32(mmCPC_INT_CNTL, CP_INT_CNTL_RING0__TIME_STAMP_INT_ENABLE_MASK |
 			CP_INT_CNTL_RING0__OPCODE_ERROR_INT_ENABLE_MASK);
 
-	unlock_srbm(kgd);
+	unlock_srbm(adev);
 
 	return 0;
 }
@@ -165,12 +150,11 @@ static inline struct vi_sdma_mqd *get_sdma_mqd(void *mqd)
 	return (struct vi_sdma_mqd *)mqd;
 }
 
-static int kgd_hqd_load(struct kgd_dev *kgd, void *mqd, uint32_t pipe_id,
-			uint32_t queue_id, uint32_t __user *wptr,
-			uint32_t wptr_shift, uint32_t wptr_mask,
-			struct mm_struct *mm)
+static int kgd_hqd_load(struct amdgpu_device *adev, void *mqd,
+			uint32_t pipe_id, uint32_t queue_id,
+			uint32_t __user *wptr, uint32_t wptr_shift,
+			uint32_t wptr_mask, struct mm_struct *mm)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	struct vi_mqd *m;
 	uint32_t *mqd_hqd;
 	uint32_t reg, wptr_val, data;
@@ -178,7 +162,7 @@ static int kgd_hqd_load(struct kgd_dev *kgd, void *mqd, uint32_t pipe_id,
 
 	m = get_mqd(mqd);
 
-	acquire_queue(kgd, pipe_id, queue_id);
+	acquire_queue(adev, pipe_id, queue_id);
 
 	/* HIQ is set during driver init period with vmid set to 0*/
 	if (m->cp_hqd_vmid == 0) {
@@ -206,7 +190,7 @@ static int kgd_hqd_load(struct kgd_dev *kgd, void *mqd, uint32_t pipe_id,
 	 * on ASICs that do not support context-save.
 	 * EOP writes/reads can start anywhere in the ring.
 	 */
-	if (get_amdgpu_device(kgd)->asic_type != CHIP_TONGA) {
+	if (adev->asic_type != CHIP_TONGA) {
 		WREG32(mmCP_HQD_EOP_RPTR, m->cp_hqd_eop_rptr);
 		WREG32(mmCP_HQD_EOP_WPTR, m->cp_hqd_eop_wptr);
 		WREG32(mmCP_HQD_EOP_WPTR_MEM, m->cp_hqd_eop_wptr_mem);
@@ -226,25 +210,24 @@ static int kgd_hqd_load(struct kgd_dev *kgd, void *mqd, uint32_t pipe_id,
 	 * release srbm_mutex to avoid circular dependency between
 	 * srbm_mutex->mm_sem->reservation_ww_class_mutex->srbm_mutex.
 	 */
-	release_queue(kgd);
+	release_queue(adev);
 	valid_wptr = read_user_wptr(mm, wptr, wptr_val);
-	acquire_queue(kgd, pipe_id, queue_id);
+	acquire_queue(adev, pipe_id, queue_id);
 	if (valid_wptr)
 		WREG32(mmCP_HQD_PQ_WPTR, (wptr_val << wptr_shift) & wptr_mask);
 
 	data = REG_SET_FIELD(m->cp_hqd_active, CP_HQD_ACTIVE, ACTIVE, 1);
 	WREG32(mmCP_HQD_ACTIVE, data);
 
-	release_queue(kgd);
+	release_queue(adev);
 
 	return 0;
 }
 
-static int kgd_hqd_dump(struct kgd_dev *kgd,
+static int kgd_hqd_dump(struct amdgpu_device *adev,
 			uint32_t pipe_id, uint32_t queue_id,
 			uint32_t (**dump)[2], uint32_t *n_regs)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	uint32_t i = 0, reg;
 #define HQD_N_REGS (54+4)
 #define DUMP_REG(addr) do {				\
@@ -258,7 +241,7 @@ static int kgd_hqd_dump(struct kgd_dev *kgd,
 	if (*dump == NULL)
 		return -ENOMEM;
 
-	acquire_queue(kgd, pipe_id, queue_id);
+	acquire_queue(adev, pipe_id, queue_id);
 
 	DUMP_REG(mmCOMPUTE_STATIC_THREAD_MGMT_SE0);
 	DUMP_REG(mmCOMPUTE_STATIC_THREAD_MGMT_SE1);
@@ -268,7 +251,7 @@ static int kgd_hqd_dump(struct kgd_dev *kgd,
 	for (reg = mmCP_MQD_BASE_ADDR; reg <= mmCP_HQD_EOP_DONES; reg++)
 		DUMP_REG(reg);
 
-	release_queue(kgd);
+	release_queue(adev);
 
 	WARN_ON_ONCE(i != HQD_N_REGS);
 	*n_regs = i;
@@ -276,10 +259,9 @@ static int kgd_hqd_dump(struct kgd_dev *kgd,
 	return 0;
 }
 
-static int kgd_hqd_sdma_load(struct kgd_dev *kgd, void *mqd,
+static int kgd_hqd_sdma_load(struct amdgpu_device *adev, void *mqd,
 			     uint32_t __user *wptr, struct mm_struct *mm)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	struct vi_sdma_mqd *m;
 	unsigned long end_jiffies;
 	uint32_t sdma_rlc_reg_offset;
@@ -331,11 +313,10 @@ static int kgd_hqd_sdma_load(struct kgd_dev *kgd, void *mqd,
 	return 0;
 }
 
-static int kgd_hqd_sdma_dump(struct kgd_dev *kgd,
+static int kgd_hqd_sdma_dump(struct amdgpu_device *adev,
 			     uint32_t engine_id, uint32_t queue_id,
 			     uint32_t (**dump)[2], uint32_t *n_regs)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	uint32_t sdma_offset = engine_id * SDMA1_REGISTER_OFFSET +
 		queue_id * KFD_VI_SDMA_QUEUE_OFFSET;
 	uint32_t i = 0, reg;
@@ -367,15 +348,15 @@ static int kgd_hqd_sdma_dump(struct kgd_dev *kgd,
 	return 0;
 }
 
-static bool kgd_hqd_is_occupied(struct kgd_dev *kgd, uint64_t queue_address,
-				uint32_t pipe_id, uint32_t queue_id)
+static bool kgd_hqd_is_occupied(struct amdgpu_device *adev,
+				uint64_t queue_address, uint32_t pipe_id,
+				uint32_t queue_id)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	uint32_t act;
 	bool retval = false;
 	uint32_t low, high;
 
-	acquire_queue(kgd, pipe_id, queue_id);
+	acquire_queue(adev, pipe_id, queue_id);
 	act = RREG32(mmCP_HQD_ACTIVE);
 	if (act) {
 		low = lower_32_bits(queue_address >> 8);
@@ -385,13 +366,12 @@ static bool kgd_hqd_is_occupied(struct kgd_dev *kgd, uint64_t queue_address,
 				high == RREG32(mmCP_HQD_PQ_BASE_HI))
 			retval = true;
 	}
-	release_queue(kgd);
+	release_queue(adev);
 	return retval;
 }
 
-static bool kgd_hqd_sdma_is_occupied(struct kgd_dev *kgd, void *mqd)
+static bool kgd_hqd_sdma_is_occupied(struct amdgpu_device *adev, void *mqd)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	struct vi_sdma_mqd *m;
 	uint32_t sdma_rlc_reg_offset;
 	uint32_t sdma_rlc_rb_cntl;
@@ -407,12 +387,11 @@ static bool kgd_hqd_sdma_is_occupied(struct kgd_dev *kgd, void *mqd)
 	return false;
 }
 
-static int kgd_hqd_destroy(struct kgd_dev *kgd, void *mqd,
+static int kgd_hqd_destroy(struct amdgpu_device *adev, void *mqd,
 				enum kfd_preempt_type reset_type,
 				unsigned int utimeout, uint32_t pipe_id,
 				uint32_t queue_id)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	uint32_t temp;
 	enum hqd_dequeue_request_type type;
 	unsigned long flags, end_jiffies;
@@ -422,7 +401,7 @@ static int kgd_hqd_destroy(struct kgd_dev *kgd, void *mqd,
 	if (amdgpu_in_reset(adev))
 		return -EIO;
 
-	acquire_queue(kgd, pipe_id, queue_id);
+	acquire_queue(adev, pipe_id, queue_id);
 
 	if (m->cp_hqd_vmid == 0)
 		WREG32_FIELD(RLC_CP_SCHEDULERS, scheduler1, 0);
@@ -502,20 +481,19 @@ loop:
 			break;
 		if (time_after(jiffies, end_jiffies)) {
 			pr_err("cp queue preemption time out.\n");
-			release_queue(kgd);
+			release_queue(adev);
 			return -ETIME;
 		}
 		usleep_range(500, 1000);
 	}
 
-	release_queue(kgd);
+	release_queue(adev);
 	return 0;
 }
 
-static int kgd_hqd_sdma_destroy(struct kgd_dev *kgd, void *mqd,
+static int kgd_hqd_sdma_destroy(struct amdgpu_device *adev, void *mqd,
 				unsigned int utimeout)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	struct vi_sdma_mqd *m;
 	uint32_t sdma_rlc_reg_offset;
 	uint32_t temp;
@@ -549,11 +527,10 @@ static int kgd_hqd_sdma_destroy(struct kgd_dev *kgd, void *mqd,
 	return 0;
 }
 
-static bool get_atc_vmid_pasid_mapping_info(struct kgd_dev *kgd,
+static bool get_atc_vmid_pasid_mapping_info(struct amdgpu_device *adev,
 					uint8_t vmid, uint16_t *p_pasid)
 {
 	uint32_t value;
-	struct amdgpu_device *adev = (struct amdgpu_device *) kgd;
 
 	value = RREG32(mmATC_VMID0_PASID_MAPPING + vmid);
 	*p_pasid = value & ATC_VMID0_PASID_MAPPING__PASID_MASK;
@@ -561,12 +538,12 @@ static bool get_atc_vmid_pasid_mapping_info(struct kgd_dev *kgd,
 	return !!(value & ATC_VMID0_PASID_MAPPING__VALID_MASK);
 }
 
-static int kgd_address_watch_disable(struct kgd_dev *kgd)
+static int kgd_address_watch_disable(struct amdgpu_device *adev)
 {
 	return 0;
 }
 
-static int kgd_address_watch_execute(struct kgd_dev *kgd,
+static int kgd_address_watch_execute(struct amdgpu_device *adev,
 					unsigned int watch_point_id,
 					uint32_t cntl_val,
 					uint32_t addr_hi,
@@ -575,11 +552,10 @@ static int kgd_address_watch_execute(struct kgd_dev *kgd,
 	return 0;
 }
 
-static int kgd_wave_control_execute(struct kgd_dev *kgd,
+static int kgd_wave_control_execute(struct amdgpu_device *adev,
 					uint32_t gfx_index_val,
 					uint32_t sq_cmd)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	uint32_t data = 0;
 
 	mutex_lock(&adev->grbm_idx_mutex);
@@ -600,28 +576,24 @@ static int kgd_wave_control_execute(struct kgd_dev *kgd,
 	return 0;
 }
 
-static uint32_t kgd_address_watch_get_offset(struct kgd_dev *kgd,
+static uint32_t kgd_address_watch_get_offset(struct amdgpu_device *adev,
 					unsigned int watch_point_id,
 					unsigned int reg_offset)
 {
 	return 0;
 }
 
-static void set_scratch_backing_va(struct kgd_dev *kgd,
+static void set_scratch_backing_va(struct amdgpu_device *adev,
 					uint64_t va, uint32_t vmid)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *) kgd;
-
-	lock_srbm(kgd, 0, 0, 0, vmid);
+	lock_srbm(adev, 0, 0, 0, vmid);
 	WREG32(mmSH_HIDDEN_PRIVATE_BASE_VMID, va);
-	unlock_srbm(kgd);
+	unlock_srbm(adev);
 }
 
-static void set_vm_context_page_table_base(struct kgd_dev *kgd, uint32_t vmid,
-		uint64_t page_table_base)
+static void set_vm_context_page_table_base(struct amdgpu_device *adev,
+		uint32_t vmid, uint64_t page_table_base)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
-
 	if (!amdgpu_amdkfd_is_kfd_vmid(adev, vmid)) {
 		pr_err("trying to set page table base for wrong VMID\n");
 		return;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v9.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v9.c
index bcc1cbeb8799..ddfe7aff919d 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v9.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v9.c
@@ -46,37 +46,26 @@ enum hqd_dequeue_request_type {
 	SAVE_WAVES
 };
 
-static inline struct amdgpu_device *get_amdgpu_device(struct kgd_dev *kgd)
-{
-	return (struct amdgpu_device *)kgd;
-}
-
-static void lock_srbm(struct kgd_dev *kgd, uint32_t mec, uint32_t pipe,
+static void lock_srbm(struct amdgpu_device *adev, uint32_t mec, uint32_t pipe,
 			uint32_t queue, uint32_t vmid)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
-
 	mutex_lock(&adev->srbm_mutex);
 	soc15_grbm_select(adev, mec, pipe, queue, vmid);
 }
 
-static void unlock_srbm(struct kgd_dev *kgd)
+static void unlock_srbm(struct amdgpu_device *adev)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
-
 	soc15_grbm_select(adev, 0, 0, 0, 0);
 	mutex_unlock(&adev->srbm_mutex);
 }
 
-static void acquire_queue(struct kgd_dev *kgd, uint32_t pipe_id,
+static void acquire_queue(struct amdgpu_device *adev, uint32_t pipe_id,
 				uint32_t queue_id)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
-
 	uint32_t mec = (pipe_id / adev->gfx.mec.num_pipe_per_mec) + 1;
 	uint32_t pipe = (pipe_id % adev->gfx.mec.num_pipe_per_mec);
 
-	lock_srbm(kgd, mec, pipe, queue_id, 0);
+	lock_srbm(adev, mec, pipe, queue_id, 0);
 }
 
 static uint64_t get_queue_mask(struct amdgpu_device *adev,
@@ -88,33 +77,29 @@ static uint64_t get_queue_mask(struct amdgpu_device *adev,
 	return 1ull << bit;
 }
 
-static void release_queue(struct kgd_dev *kgd)
+static void release_queue(struct amdgpu_device *adev)
 {
-	unlock_srbm(kgd);
+	unlock_srbm(adev);
 }
 
-void kgd_gfx_v9_program_sh_mem_settings(struct kgd_dev *kgd, uint32_t vmid,
+void kgd_gfx_v9_program_sh_mem_settings(struct amdgpu_device *adev, uint32_t vmid,
 					uint32_t sh_mem_config,
 					uint32_t sh_mem_ape1_base,
 					uint32_t sh_mem_ape1_limit,
 					uint32_t sh_mem_bases)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
-
-	lock_srbm(kgd, 0, 0, 0, vmid);
+	lock_srbm(adev, 0, 0, 0, vmid);
 
 	WREG32_RLC(SOC15_REG_OFFSET(GC, 0, mmSH_MEM_CONFIG), sh_mem_config);
 	WREG32_RLC(SOC15_REG_OFFSET(GC, 0, mmSH_MEM_BASES), sh_mem_bases);
 	/* APE1 no longer exists on GFX9 */
 
-	unlock_srbm(kgd);
+	unlock_srbm(adev);
 }
 
-int kgd_gfx_v9_set_pasid_vmid_mapping(struct kgd_dev *kgd, u32 pasid,
+int kgd_gfx_v9_set_pasid_vmid_mapping(struct amdgpu_device *adev, u32 pasid,
 					unsigned int vmid)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
-
 	/*
 	 * We have to assume that there is no outstanding mapping.
 	 * The ATC_VMID_PASID_MAPPING_UPDATE_STATUS bit could be 0 because
@@ -171,22 +156,21 @@ int kgd_gfx_v9_set_pasid_vmid_mapping(struct kgd_dev *kgd, u32 pasid,
  * but still works
  */
 
-int kgd_gfx_v9_init_interrupts(struct kgd_dev *kgd, uint32_t pipe_id)
+int kgd_gfx_v9_init_interrupts(struct amdgpu_device *adev, uint32_t pipe_id)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	uint32_t mec;
 	uint32_t pipe;
 
 	mec = (pipe_id / adev->gfx.mec.num_pipe_per_mec) + 1;
 	pipe = (pipe_id % adev->gfx.mec.num_pipe_per_mec);
 
-	lock_srbm(kgd, mec, pipe, 0, 0);
+	lock_srbm(adev, mec, pipe, 0, 0);
 
 	WREG32(SOC15_REG_OFFSET(GC, 0, mmCPC_INT_CNTL),
 		CP_INT_CNTL_RING0__TIME_STAMP_INT_ENABLE_MASK |
 		CP_INT_CNTL_RING0__OPCODE_ERROR_INT_ENABLE_MASK);
 
-	unlock_srbm(kgd);
+	unlock_srbm(adev);
 
 	return 0;
 }
@@ -233,19 +217,18 @@ static inline struct v9_sdma_mqd *get_sdma_mqd(void *mqd)
 	return (struct v9_sdma_mqd *)mqd;
 }
 
-int kgd_gfx_v9_hqd_load(struct kgd_dev *kgd, void *mqd, uint32_t pipe_id,
-			uint32_t queue_id, uint32_t __user *wptr,
-			uint32_t wptr_shift, uint32_t wptr_mask,
-			struct mm_struct *mm)
+int kgd_gfx_v9_hqd_load(struct amdgpu_device *adev, void *mqd,
+			uint32_t pipe_id, uint32_t queue_id,
+			uint32_t __user *wptr, uint32_t wptr_shift,
+			uint32_t wptr_mask, struct mm_struct *mm)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	struct v9_mqd *m;
 	uint32_t *mqd_hqd;
 	uint32_t reg, hqd_base, data;
 
 	m = get_mqd(mqd);
 
-	acquire_queue(kgd, pipe_id, queue_id);
+	acquire_queue(adev, pipe_id, queue_id);
 
 	/* HQD registers extend from CP_MQD_BASE_ADDR to CP_HQD_EOP_WPTR_MEM. */
 	mqd_hqd = &m->cp_mqd_base_addr_lo;
@@ -308,16 +291,15 @@ int kgd_gfx_v9_hqd_load(struct kgd_dev *kgd, void *mqd, uint32_t pipe_id,
 	data = REG_SET_FIELD(m->cp_hqd_active, CP_HQD_ACTIVE, ACTIVE, 1);
 	WREG32_RLC(SOC15_REG_OFFSET(GC, 0, mmCP_HQD_ACTIVE), data);
 
-	release_queue(kgd);
+	release_queue(adev);
 
 	return 0;
 }
 
-int kgd_gfx_v9_hiq_mqd_load(struct kgd_dev *kgd, void *mqd,
+int kgd_gfx_v9_hiq_mqd_load(struct amdgpu_device *adev, void *mqd,
 			    uint32_t pipe_id, uint32_t queue_id,
 			    uint32_t doorbell_off)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	struct amdgpu_ring *kiq_ring = &adev->gfx.kiq.ring;
 	struct v9_mqd *m;
 	uint32_t mec, pipe;
@@ -325,7 +307,7 @@ int kgd_gfx_v9_hiq_mqd_load(struct kgd_dev *kgd, void *mqd,
 
 	m = get_mqd(mqd);
 
-	acquire_queue(kgd, pipe_id, queue_id);
+	acquire_queue(adev, pipe_id, queue_id);
 
 	mec = (pipe_id / adev->gfx.mec.num_pipe_per_mec) + 1;
 	pipe = (pipe_id % adev->gfx.mec.num_pipe_per_mec);
@@ -361,16 +343,15 @@ int kgd_gfx_v9_hiq_mqd_load(struct kgd_dev *kgd, void *mqd,
 
 out_unlock:
 	spin_unlock(&adev->gfx.kiq.ring_lock);
-	release_queue(kgd);
+	release_queue(adev);
 
 	return r;
 }
 
-int kgd_gfx_v9_hqd_dump(struct kgd_dev *kgd,
+int kgd_gfx_v9_hqd_dump(struct amdgpu_device *adev,
 			uint32_t pipe_id, uint32_t queue_id,
 			uint32_t (**dump)[2], uint32_t *n_regs)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	uint32_t i = 0, reg;
 #define HQD_N_REGS 56
 #define DUMP_REG(addr) do {				\
@@ -384,13 +365,13 @@ int kgd_gfx_v9_hqd_dump(struct kgd_dev *kgd,
 	if (*dump == NULL)
 		return -ENOMEM;
 
-	acquire_queue(kgd, pipe_id, queue_id);
+	acquire_queue(adev, pipe_id, queue_id);
 
 	for (reg = SOC15_REG_OFFSET(GC, 0, mmCP_MQD_BASE_ADDR);
 	     reg <= SOC15_REG_OFFSET(GC, 0, mmCP_HQD_PQ_WPTR_HI); reg++)
 		DUMP_REG(reg);
 
-	release_queue(kgd);
+	release_queue(adev);
 
 	WARN_ON_ONCE(i != HQD_N_REGS);
 	*n_regs = i;
@@ -398,10 +379,9 @@ int kgd_gfx_v9_hqd_dump(struct kgd_dev *kgd,
 	return 0;
 }
 
-static int kgd_hqd_sdma_load(struct kgd_dev *kgd, void *mqd,
+static int kgd_hqd_sdma_load(struct amdgpu_device *adev, void *mqd,
 			     uint32_t __user *wptr, struct mm_struct *mm)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	struct v9_sdma_mqd *m;
 	uint32_t sdma_rlc_reg_offset;
 	unsigned long end_jiffies;
@@ -468,11 +448,10 @@ static int kgd_hqd_sdma_load(struct kgd_dev *kgd, void *mqd,
 	return 0;
 }
 
-static int kgd_hqd_sdma_dump(struct kgd_dev *kgd,
+static int kgd_hqd_sdma_dump(struct amdgpu_device *adev,
 			     uint32_t engine_id, uint32_t queue_id,
 			     uint32_t (**dump)[2], uint32_t *n_regs)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	uint32_t sdma_rlc_reg_offset = get_sdma_rlc_reg_offset(adev,
 			engine_id, queue_id);
 	uint32_t i = 0, reg;
@@ -500,15 +479,15 @@ static int kgd_hqd_sdma_dump(struct kgd_dev *kgd,
 	return 0;
 }
 
-bool kgd_gfx_v9_hqd_is_occupied(struct kgd_dev *kgd, uint64_t queue_address,
-				uint32_t pipe_id, uint32_t queue_id)
+bool kgd_gfx_v9_hqd_is_occupied(struct amdgpu_device *adev,
+				uint64_t queue_address, uint32_t pipe_id,
+				uint32_t queue_id)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	uint32_t act;
 	bool retval = false;
 	uint32_t low, high;
 
-	acquire_queue(kgd, pipe_id, queue_id);
+	acquire_queue(adev, pipe_id, queue_id);
 	act = RREG32(SOC15_REG_OFFSET(GC, 0, mmCP_HQD_ACTIVE));
 	if (act) {
 		low = lower_32_bits(queue_address >> 8);
@@ -518,13 +497,12 @@ bool kgd_gfx_v9_hqd_is_occupied(struct kgd_dev *kgd, uint64_t queue_address,
 		   high == RREG32(SOC15_REG_OFFSET(GC, 0, mmCP_HQD_PQ_BASE_HI)))
 			retval = true;
 	}
-	release_queue(kgd);
+	release_queue(adev);
 	return retval;
 }
 
-static bool kgd_hqd_sdma_is_occupied(struct kgd_dev *kgd, void *mqd)
+static bool kgd_hqd_sdma_is_occupied(struct amdgpu_device *adev, void *mqd)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	struct v9_sdma_mqd *m;
 	uint32_t sdma_rlc_reg_offset;
 	uint32_t sdma_rlc_rb_cntl;
@@ -541,12 +519,11 @@ static bool kgd_hqd_sdma_is_occupied(struct kgd_dev *kgd, void *mqd)
 	return false;
 }
 
-int kgd_gfx_v9_hqd_destroy(struct kgd_dev *kgd, void *mqd,
+int kgd_gfx_v9_hqd_destroy(struct amdgpu_device *adev, void *mqd,
 				enum kfd_preempt_type reset_type,
 				unsigned int utimeout, uint32_t pipe_id,
 				uint32_t queue_id)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	enum hqd_dequeue_request_type type;
 	unsigned long end_jiffies;
 	uint32_t temp;
@@ -555,7 +532,7 @@ int kgd_gfx_v9_hqd_destroy(struct kgd_dev *kgd, void *mqd,
 	if (amdgpu_in_reset(adev))
 		return -EIO;
 
-	acquire_queue(kgd, pipe_id, queue_id);
+	acquire_queue(adev, pipe_id, queue_id);
 
 	if (m->cp_hqd_vmid == 0)
 		WREG32_FIELD15_RLC(GC, 0, RLC_CP_SCHEDULERS, scheduler1, 0);
@@ -584,20 +561,19 @@ int kgd_gfx_v9_hqd_destroy(struct kgd_dev *kgd, void *mqd,
 			break;
 		if (time_after(jiffies, end_jiffies)) {
 			pr_err("cp queue preemption time out.\n");
-			release_queue(kgd);
+			release_queue(adev);
 			return -ETIME;
 		}
 		usleep_range(500, 1000);
 	}
 
-	release_queue(kgd);
+	release_queue(adev);
 	return 0;
 }
 
-static int kgd_hqd_sdma_destroy(struct kgd_dev *kgd, void *mqd,
+static int kgd_hqd_sdma_destroy(struct amdgpu_device *adev, void *mqd,
 				unsigned int utimeout)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	struct v9_sdma_mqd *m;
 	uint32_t sdma_rlc_reg_offset;
 	uint32_t temp;
@@ -634,11 +610,10 @@ static int kgd_hqd_sdma_destroy(struct kgd_dev *kgd, void *mqd,
 	return 0;
 }
 
-bool kgd_gfx_v9_get_atc_vmid_pasid_mapping_info(struct kgd_dev *kgd,
+bool kgd_gfx_v9_get_atc_vmid_pasid_mapping_info(struct amdgpu_device *adev,
 					uint8_t vmid, uint16_t *p_pasid)
 {
 	uint32_t value;
-	struct amdgpu_device *adev = (struct amdgpu_device *) kgd;
 
 	value = RREG32(SOC15_REG_OFFSET(ATHUB, 0, mmATC_VMID0_PASID_MAPPING)
 		     + vmid);
@@ -647,12 +622,12 @@ bool kgd_gfx_v9_get_atc_vmid_pasid_mapping_info(struct kgd_dev *kgd,
 	return !!(value & ATC_VMID0_PASID_MAPPING__VALID_MASK);
 }
 
-int kgd_gfx_v9_address_watch_disable(struct kgd_dev *kgd)
+int kgd_gfx_v9_address_watch_disable(struct amdgpu_device *adev)
 {
 	return 0;
 }
 
-int kgd_gfx_v9_address_watch_execute(struct kgd_dev *kgd,
+int kgd_gfx_v9_address_watch_execute(struct amdgpu_device *adev,
 					unsigned int watch_point_id,
 					uint32_t cntl_val,
 					uint32_t addr_hi,
@@ -661,11 +636,10 @@ int kgd_gfx_v9_address_watch_execute(struct kgd_dev *kgd,
 	return 0;
 }
 
-int kgd_gfx_v9_wave_control_execute(struct kgd_dev *kgd,
+int kgd_gfx_v9_wave_control_execute(struct amdgpu_device *adev,
 					uint32_t gfx_index_val,
 					uint32_t sq_cmd)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	uint32_t data = 0;
 
 	mutex_lock(&adev->grbm_idx_mutex);
@@ -686,18 +660,16 @@ int kgd_gfx_v9_wave_control_execute(struct kgd_dev *kgd,
 	return 0;
 }
 
-uint32_t kgd_gfx_v9_address_watch_get_offset(struct kgd_dev *kgd,
+uint32_t kgd_gfx_v9_address_watch_get_offset(struct amdgpu_device *adev,
 					unsigned int watch_point_id,
 					unsigned int reg_offset)
 {
 	return 0;
 }
 
-void kgd_gfx_v9_set_vm_context_page_table_base(struct kgd_dev *kgd,
+void kgd_gfx_v9_set_vm_context_page_table_base(struct amdgpu_device *adev,
 			uint32_t vmid, uint64_t page_table_base)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
-
 	if (!amdgpu_amdkfd_is_kfd_vmid(adev, vmid)) {
 		pr_err("trying to set page table base for wrong VMID %u\n",
 		       vmid);
@@ -804,7 +776,7 @@ static void get_wave_count(struct amdgpu_device *adev, int queue_idx,
  *
  *  Reading registers referenced above involves programming GRBM appropriately
  */
-void kgd_gfx_v9_get_cu_occupancy(struct kgd_dev *kgd, int pasid,
+void kgd_gfx_v9_get_cu_occupancy(struct amdgpu_device *adev, int pasid,
 		int *pasid_wave_cnt, int *max_waves_per_cu)
 {
 	int qidx;
@@ -818,10 +790,8 @@ void kgd_gfx_v9_get_cu_occupancy(struct kgd_dev *kgd, int pasid,
 	int pasid_tmp;
 	int max_queue_cnt;
 	int vmid_wave_cnt = 0;
-	struct amdgpu_device *adev;
 	DECLARE_BITMAP(cp_queue_bitmap, KGD_MAX_QUEUES);
 
-	adev = get_amdgpu_device(kgd);
 	lock_spi_csq_mutexes(adev);
 	soc15_grbm_select(adev, 1, 0, 0, 0);
 
@@ -882,12 +852,10 @@ void kgd_gfx_v9_get_cu_occupancy(struct kgd_dev *kgd, int pasid,
 				adev->gfx.cu_info.max_waves_per_simd;
 }
 
-void kgd_gfx_v9_program_trap_handler_settings(struct kgd_dev *kgd,
+void kgd_gfx_v9_program_trap_handler_settings(struct amdgpu_device *adev,
                         uint32_t vmid, uint64_t tba_addr, uint64_t tma_addr)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
-
-	lock_srbm(kgd, 0, 0, 0, vmid);
+	lock_srbm(adev, 0, 0, 0, vmid);
 
 	/*
 	 * Program TBA registers
@@ -905,7 +873,7 @@ void kgd_gfx_v9_program_trap_handler_settings(struct kgd_dev *kgd,
 	WREG32(SOC15_REG_OFFSET(GC, 0, mmSQ_SHADER_TMA_HI),
 			upper_32_bits(tma_addr >> 8));
 
-	unlock_srbm(kgd);
+	unlock_srbm(adev);
 }
 
 const struct kfd2kgd_calls gfx_v9_kfd2kgd = {
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v9.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v9.h
index c63591106879..24be49df26fd 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v9.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v9.h
@@ -22,48 +22,49 @@
 
 
 
-void kgd_gfx_v9_program_sh_mem_settings(struct kgd_dev *kgd, uint32_t vmid,
+void kgd_gfx_v9_program_sh_mem_settings(struct amdgpu_device *adev, uint32_t vmid,
 		uint32_t sh_mem_config,
 		uint32_t sh_mem_ape1_base, uint32_t sh_mem_ape1_limit,
 		uint32_t sh_mem_bases);
-int kgd_gfx_v9_set_pasid_vmid_mapping(struct kgd_dev *kgd, u32 pasid,
+int kgd_gfx_v9_set_pasid_vmid_mapping(struct amdgpu_device *adev, u32 pasid,
 		unsigned int vmid);
-int kgd_gfx_v9_init_interrupts(struct kgd_dev *kgd, uint32_t pipe_id);
-int kgd_gfx_v9_hqd_load(struct kgd_dev *kgd, void *mqd, uint32_t pipe_id,
+int kgd_gfx_v9_init_interrupts(struct amdgpu_device *adev, uint32_t pipe_id);
+int kgd_gfx_v9_hqd_load(struct amdgpu_device *adev, void *mqd, uint32_t pipe_id,
 			uint32_t queue_id, uint32_t __user *wptr,
 			uint32_t wptr_shift, uint32_t wptr_mask,
 			struct mm_struct *mm);
-int kgd_gfx_v9_hiq_mqd_load(struct kgd_dev *kgd, void *mqd,
+int kgd_gfx_v9_hiq_mqd_load(struct amdgpu_device *adev, void *mqd,
 			    uint32_t pipe_id, uint32_t queue_id,
 			    uint32_t doorbell_off);
-int kgd_gfx_v9_hqd_dump(struct kgd_dev *kgd,
+int kgd_gfx_v9_hqd_dump(struct amdgpu_device *adev,
 			uint32_t pipe_id, uint32_t queue_id,
 			uint32_t (**dump)[2], uint32_t *n_regs);
-bool kgd_gfx_v9_hqd_is_occupied(struct kgd_dev *kgd, uint64_t queue_address,
-		uint32_t pipe_id, uint32_t queue_id);
-int kgd_gfx_v9_hqd_destroy(struct kgd_dev *kgd, void *mqd,
+bool kgd_gfx_v9_hqd_is_occupied(struct amdgpu_device *adev,
+			uint64_t queue_address, uint32_t pipe_id,
+			uint32_t queue_id);
+int kgd_gfx_v9_hqd_destroy(struct amdgpu_device *adev, void *mqd,
 				enum kfd_preempt_type reset_type,
 				unsigned int utimeout, uint32_t pipe_id,
 				uint32_t queue_id);
-int kgd_gfx_v9_address_watch_disable(struct kgd_dev *kgd);
-int kgd_gfx_v9_address_watch_execute(struct kgd_dev *kgd,
+int kgd_gfx_v9_address_watch_disable(struct amdgpu_device *adev);
+int kgd_gfx_v9_address_watch_execute(struct amdgpu_device *adev,
 					unsigned int watch_point_id,
 					uint32_t cntl_val,
 					uint32_t addr_hi,
 					uint32_t addr_lo);
-int kgd_gfx_v9_wave_control_execute(struct kgd_dev *kgd,
+int kgd_gfx_v9_wave_control_execute(struct amdgpu_device *adev,
 					uint32_t gfx_index_val,
 					uint32_t sq_cmd);
-uint32_t kgd_gfx_v9_address_watch_get_offset(struct kgd_dev *kgd,
+uint32_t kgd_gfx_v9_address_watch_get_offset(struct amdgpu_device *adev,
 					unsigned int watch_point_id,
 					unsigned int reg_offset);
 
-bool kgd_gfx_v9_get_atc_vmid_pasid_mapping_info(struct kgd_dev *kgd,
+bool kgd_gfx_v9_get_atc_vmid_pasid_mapping_info(struct amdgpu_device *adev,
 					uint8_t vmid, uint16_t *p_pasid);
 
-void kgd_gfx_v9_set_vm_context_page_table_base(struct kgd_dev *kgd,
+void kgd_gfx_v9_set_vm_context_page_table_base(struct amdgpu_device *adev,
 			uint32_t vmid, uint64_t page_table_base);
-void kgd_gfx_v9_get_cu_occupancy(struct kgd_dev *kgd, int pasid,
+void kgd_gfx_v9_get_cu_occupancy(struct amdgpu_device *adev, int pasid,
 		int *pasid_wave_cnt, int *max_waves_per_cu);
-void kgd_gfx_v9_program_trap_handler_settings(struct kgd_dev *kgd,
+void kgd_gfx_v9_program_trap_handler_settings(struct amdgpu_device *adev,
 		uint32_t vmid, uint64_t tba_addr, uint64_t tma_addr);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c
index 71acd577803e..c5362c23e687 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c
@@ -60,12 +60,6 @@ static const char * const domain_bit_to_string[] = {
 
 static void amdgpu_amdkfd_restore_userptr_worker(struct work_struct *work);
 
-
-static inline struct amdgpu_device *get_amdgpu_device(struct kgd_dev *kgd)
-{
-	return (struct amdgpu_device *)kgd;
-}
-
 static bool kfd_mem_is_attached(struct amdgpu_vm *avm,
 		struct kgd_mem *mem)
 {
@@ -126,8 +120,19 @@ static size_t amdgpu_amdkfd_acc_size(uint64_t size)
 		PAGE_ALIGN(size);
 }
 
+/**
+ * @amdgpu_amdkfd_reserve_mem_limit() - Decrease available memory by size
+ * of buffer including any reserved for control structures
+ *
+ * @adev: Device to which allocated BO belongs to
+ * @size: Size of buffer, in bytes, encapsulated by B0. This should be
+ * equivalent to amdgpu_bo_size(BO)
+ * @alloc_flag: Flag used in allocating a BO as noted above
+ *
+ * Return: returns -ENOMEM in case of error, ZERO otherwise
+ */
 static int amdgpu_amdkfd_reserve_mem_limit(struct amdgpu_device *adev,
-		uint64_t size, u32 domain, bool sg)
+		uint64_t size, u32 alloc_flag)
 {
 	uint64_t reserved_for_pt =
 		ESTIMATE_PT_SIZE(amdgpu_amdkfd_total_mem_size);
@@ -137,20 +142,24 @@ static int amdgpu_amdkfd_reserve_mem_limit(struct amdgpu_device *adev,
 	acc_size = amdgpu_amdkfd_acc_size(size);
 
 	vram_needed = 0;
-	if (domain == AMDGPU_GEM_DOMAIN_GTT) {
-		/* TTM GTT memory */
+	if (alloc_flag & KFD_IOC_ALLOC_MEM_FLAGS_GTT) {
 		system_mem_needed = acc_size + size;
 		ttm_mem_needed = acc_size + size;
-	} else if (domain == AMDGPU_GEM_DOMAIN_CPU && !sg) {
-		/* Userptr */
+	} else if (alloc_flag & KFD_IOC_ALLOC_MEM_FLAGS_VRAM) {
+		system_mem_needed = acc_size;
+		ttm_mem_needed = acc_size;
+		vram_needed = size;
+	} else if (alloc_flag & KFD_IOC_ALLOC_MEM_FLAGS_USERPTR) {
 		system_mem_needed = acc_size + size;
 		ttm_mem_needed = acc_size;
-	} else {
-		/* VRAM and SG */
+	} else if (alloc_flag &
+		   (KFD_IOC_ALLOC_MEM_FLAGS_DOORBELL |
+		    KFD_IOC_ALLOC_MEM_FLAGS_MMIO_REMAP)) {
 		system_mem_needed = acc_size;
 		ttm_mem_needed = acc_size;
-		if (domain == AMDGPU_GEM_DOMAIN_VRAM)
-			vram_needed = size;
+	} else {
+		pr_err("%s: Invalid BO type %#x\n", __func__, alloc_flag);
+		return -ENOMEM;
 	}
 
 	spin_lock(&kfd_mem_limit.mem_limit_lock);
@@ -166,64 +175,72 @@ static int amdgpu_amdkfd_reserve_mem_limit(struct amdgpu_device *adev,
 	    (adev->kfd.vram_used + vram_needed >
 	     adev->gmc.real_vram_size - reserved_for_pt)) {
 		ret = -ENOMEM;
-	} else {
-		kfd_mem_limit.system_mem_used += system_mem_needed;
-		kfd_mem_limit.ttm_mem_used += ttm_mem_needed;
-		adev->kfd.vram_used += vram_needed;
+		goto release;
 	}
 
+	/* Update memory accounting by decreasing available system
+	 * memory, TTM memory and GPU memory as computed above
+	 */
+	adev->kfd.vram_used += vram_needed;
+	kfd_mem_limit.system_mem_used += system_mem_needed;
+	kfd_mem_limit.ttm_mem_used += ttm_mem_needed;
+
+release:
 	spin_unlock(&kfd_mem_limit.mem_limit_lock);
 	return ret;
 }
 
 static void unreserve_mem_limit(struct amdgpu_device *adev,
-		uint64_t size, u32 domain, bool sg)
+		uint64_t size, u32 alloc_flag)
 {
 	size_t acc_size;
 
 	acc_size = amdgpu_amdkfd_acc_size(size);
 
 	spin_lock(&kfd_mem_limit.mem_limit_lock);
-	if (domain == AMDGPU_GEM_DOMAIN_GTT) {
+
+	if (alloc_flag & KFD_IOC_ALLOC_MEM_FLAGS_GTT) {
 		kfd_mem_limit.system_mem_used -= (acc_size + size);
 		kfd_mem_limit.ttm_mem_used -= (acc_size + size);
-	} else if (domain == AMDGPU_GEM_DOMAIN_CPU && !sg) {
+	} else if (alloc_flag & KFD_IOC_ALLOC_MEM_FLAGS_VRAM) {
+		kfd_mem_limit.system_mem_used -= acc_size;
+		kfd_mem_limit.ttm_mem_used -= acc_size;
+		adev->kfd.vram_used -= size;
+	} else if (alloc_flag & KFD_IOC_ALLOC_MEM_FLAGS_USERPTR) {
 		kfd_mem_limit.system_mem_used -= (acc_size + size);
 		kfd_mem_limit.ttm_mem_used -= acc_size;
-	} else {
+	} else if (alloc_flag &
+		   (KFD_IOC_ALLOC_MEM_FLAGS_DOORBELL |
+		    KFD_IOC_ALLOC_MEM_FLAGS_MMIO_REMAP)) {
 		kfd_mem_limit.system_mem_used -= acc_size;
 		kfd_mem_limit.ttm_mem_used -= acc_size;
-		if (domain == AMDGPU_GEM_DOMAIN_VRAM) {
-			adev->kfd.vram_used -= size;
-			WARN_ONCE(adev->kfd.vram_used < 0,
-				  "kfd VRAM memory accounting unbalanced");
-		}
+	} else {
+		pr_err("%s: Invalid BO type %#x\n", __func__, alloc_flag);
+		goto release;
 	}
-	WARN_ONCE(kfd_mem_limit.system_mem_used < 0,
-		  "kfd system memory accounting unbalanced");
+
+	WARN_ONCE(adev->kfd.vram_used < 0,
+		  "KFD VRAM memory accounting unbalanced");
 	WARN_ONCE(kfd_mem_limit.ttm_mem_used < 0,
-		  "kfd TTM memory accounting unbalanced");
+		  "KFD TTM memory accounting unbalanced");
+	WARN_ONCE(kfd_mem_limit.system_mem_used < 0,
+		  "KFD system memory accounting unbalanced");
 
+release:
 	spin_unlock(&kfd_mem_limit.mem_limit_lock);
 }
 
 void amdgpu_amdkfd_release_notify(struct amdgpu_bo *bo)
 {
 	struct amdgpu_device *adev = amdgpu_ttm_adev(bo->tbo.bdev);
-	u32 domain = bo->preferred_domains;
-	bool sg = (bo->preferred_domains == AMDGPU_GEM_DOMAIN_CPU);
-
-	if (bo->flags & AMDGPU_AMDKFD_CREATE_USERPTR_BO) {
-		domain = AMDGPU_GEM_DOMAIN_CPU;
-		sg = false;
-	}
+	u32 alloc_flags = bo->kfd_bo->alloc_flags;
+	u64 size = amdgpu_bo_size(bo);
 
-	unreserve_mem_limit(adev, amdgpu_bo_size(bo), domain, sg);
+	unreserve_mem_limit(adev, size, alloc_flags);
 
 	kfree(bo->kfd_bo);
 }
 
-
 /* amdgpu_amdkfd_remove_eviction_fence - Removes eviction fence from BO's
  *  reservation object.
  *
@@ -646,12 +663,6 @@ kfd_mem_attach_dmabuf(struct amdgpu_device *adev, struct kgd_mem *mem,
 	if (IS_ERR(gobj))
 		return PTR_ERR(gobj);
 
-	/* Import takes an extra reference on the dmabuf. Drop it now to
-	 * avoid leaking it. We only need the one reference in
-	 * kgd_mem->dmabuf.
-	 */
-	dma_buf_put(mem->dmabuf);
-
 	*bo = gem_to_amdgpu_bo(gobj);
 	(*bo)->flags |= AMDGPU_GEM_CREATE_PREEMPTIBLE;
 	(*bo)->parent = amdgpu_bo_ref(mem->bo);
@@ -1278,12 +1289,60 @@ create_evict_fence_fail:
 	return ret;
 }
 
-int amdgpu_amdkfd_gpuvm_acquire_process_vm(struct kgd_dev *kgd,
+/**
+ * amdgpu_amdkfd_gpuvm_pin_bo() - Pins a BO using following criteria
+ * @bo: Handle of buffer object being pinned
+ * @domain: Domain into which BO should be pinned
+ *
+ *   - USERPTR BOs are UNPINNABLE and will return error
+ *   - All other BO types (GTT, VRAM, MMIO and DOORBELL) will have their
+ *     PIN count incremented. It is valid to PIN a BO multiple times
+ *
+ * Return: ZERO if successful in pinning, Non-Zero in case of error.
+ */
+static int amdgpu_amdkfd_gpuvm_pin_bo(struct amdgpu_bo *bo, u32 domain)
+{
+	int ret = 0;
+
+	ret = amdgpu_bo_reserve(bo, false);
+	if (unlikely(ret))
+		return ret;
+
+	ret = amdgpu_bo_pin_restricted(bo, domain, 0, 0);
+	if (ret)
+		pr_err("Error in Pinning BO to domain: %d\n", domain);
+
+	amdgpu_bo_sync_wait(bo, AMDGPU_FENCE_OWNER_KFD, false);
+	amdgpu_bo_unreserve(bo);
+
+	return ret;
+}
+
+/**
+ * amdgpu_amdkfd_gpuvm_unpin_bo() - Unpins BO using following criteria
+ * @bo: Handle of buffer object being unpinned
+ *
+ *   - Is a illegal request for USERPTR BOs and is ignored
+ *   - All other BO types (GTT, VRAM, MMIO and DOORBELL) will have their
+ *     PIN count decremented. Calls to UNPIN must balance calls to PIN
+ */
+static void amdgpu_amdkfd_gpuvm_unpin_bo(struct amdgpu_bo *bo)
+{
+	int ret = 0;
+
+	ret = amdgpu_bo_reserve(bo, false);
+	if (unlikely(ret))
+		return;
+
+	amdgpu_bo_unpin(bo);
+	amdgpu_bo_unreserve(bo);
+}
+
+int amdgpu_amdkfd_gpuvm_acquire_process_vm(struct amdgpu_device *adev,
 					   struct file *filp, u32 pasid,
 					   void **process_info,
 					   struct dma_fence **ef)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	struct amdgpu_fpriv *drv_priv;
 	struct amdgpu_vm *avm;
 	int ret;
@@ -1359,12 +1418,12 @@ void amdgpu_amdkfd_gpuvm_destroy_cb(struct amdgpu_device *adev,
 	}
 }
 
-void amdgpu_amdkfd_gpuvm_release_process_vm(struct kgd_dev *kgd, void *drm_priv)
+void amdgpu_amdkfd_gpuvm_release_process_vm(struct amdgpu_device *adev,
+					    void *drm_priv)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	struct amdgpu_vm *avm;
 
-	if (WARN_ON(!kgd || !drm_priv))
+	if (WARN_ON(!adev || !drm_priv))
 		return;
 
 	avm = drm_priv_to_vm(drm_priv);
@@ -1392,11 +1451,10 @@ uint64_t amdgpu_amdkfd_gpuvm_get_process_page_dir(void *drm_priv)
 }
 
 int amdgpu_amdkfd_gpuvm_alloc_memory_of_gpu(
-		struct kgd_dev *kgd, uint64_t va, uint64_t size,
+		struct amdgpu_device *adev, uint64_t va, uint64_t size,
 		void *drm_priv, struct kgd_mem **mem,
 		uint64_t *offset, uint32_t flags)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	struct amdgpu_vm *avm = drm_priv_to_vm(drm_priv);
 	enum ttm_bo_type bo_type = ttm_bo_type_device;
 	struct sg_table *sg = NULL;
@@ -1460,7 +1518,7 @@ int amdgpu_amdkfd_gpuvm_alloc_memory_of_gpu(
 
 	amdgpu_sync_create(&(*mem)->sync);
 
-	ret = amdgpu_amdkfd_reserve_mem_limit(adev, size, alloc_domain, !!sg);
+	ret = amdgpu_amdkfd_reserve_mem_limit(adev, size, flags);
 	if (ret) {
 		pr_debug("Insufficient memory\n");
 		goto err_reserve_limit;
@@ -1506,17 +1564,29 @@ int amdgpu_amdkfd_gpuvm_alloc_memory_of_gpu(
 	if (offset)
 		*offset = amdgpu_bo_mmap_offset(bo);
 
+	if (flags & (KFD_IOC_ALLOC_MEM_FLAGS_DOORBELL |
+			KFD_IOC_ALLOC_MEM_FLAGS_MMIO_REMAP)) {
+		ret = amdgpu_amdkfd_gpuvm_pin_bo(bo, AMDGPU_GEM_DOMAIN_GTT);
+		if (ret) {
+			pr_err("Pinning MMIO/DOORBELL BO during ALLOC FAILED\n");
+			goto err_pin_bo;
+		}
+		bo->allowed_domains = AMDGPU_GEM_DOMAIN_GTT;
+		bo->preferred_domains = AMDGPU_GEM_DOMAIN_GTT;
+	}
+
 	return 0;
 
 allocate_init_user_pages_failed:
 	remove_kgd_mem_from_kfd_bo_list(*mem, avm->process_info);
+err_pin_bo:
 	drm_vma_node_revoke(&gobj->vma_node, drm_priv);
 err_node_allow:
 	drm_gem_object_put(gobj);
 	/* Don't unreserve system mem limit twice */
 	goto err_reserve_limit;
 err_bo_create:
-	unreserve_mem_limit(adev, size, alloc_domain, !!sg);
+	unreserve_mem_limit(adev, size, flags);
 err_reserve_limit:
 	mutex_destroy(&(*mem)->lock);
 	kfree(*mem);
@@ -1529,7 +1599,7 @@ err:
 }
 
 int amdgpu_amdkfd_gpuvm_free_memory_of_gpu(
-		struct kgd_dev *kgd, struct kgd_mem *mem, void *drm_priv,
+		struct amdgpu_device *adev, struct kgd_mem *mem, void *drm_priv,
 		uint64_t *size)
 {
 	struct amdkfd_process_info *process_info = mem->process_info;
@@ -1542,6 +1612,14 @@ int amdgpu_amdkfd_gpuvm_free_memory_of_gpu(
 	bool is_imported = false;
 
 	mutex_lock(&mem->lock);
+
+	/* Unpin MMIO/DOORBELL BO's that were pinnned during allocation */
+	if (mem->alloc_flags &
+	    (KFD_IOC_ALLOC_MEM_FLAGS_DOORBELL |
+	     KFD_IOC_ALLOC_MEM_FLAGS_MMIO_REMAP)) {
+		amdgpu_amdkfd_gpuvm_unpin_bo(mem->bo);
+	}
+
 	mapped_to_gpu_memory = mem->mapped_to_gpu_memory;
 	is_imported = mem->is_imported;
 	mutex_unlock(&mem->lock);
@@ -1621,10 +1699,9 @@ int amdgpu_amdkfd_gpuvm_free_memory_of_gpu(
 }
 
 int amdgpu_amdkfd_gpuvm_map_memory_to_gpu(
-		struct kgd_dev *kgd, struct kgd_mem *mem,
+		struct amdgpu_device *adev, struct kgd_mem *mem,
 		void *drm_priv, bool *table_freed)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	struct amdgpu_vm *avm = drm_priv_to_vm(drm_priv);
 	int ret;
 	struct amdgpu_bo *bo;
@@ -1751,7 +1828,7 @@ out:
 }
 
 int amdgpu_amdkfd_gpuvm_unmap_memory_from_gpu(
-		struct kgd_dev *kgd, struct kgd_mem *mem, void *drm_priv)
+		struct amdgpu_device *adev, struct kgd_mem *mem, void *drm_priv)
 {
 	struct amdgpu_vm *avm = drm_priv_to_vm(drm_priv);
 	struct amdkfd_process_info *process_info = avm->process_info;
@@ -1812,7 +1889,7 @@ out:
 }
 
 int amdgpu_amdkfd_gpuvm_sync_memory(
-		struct kgd_dev *kgd, struct kgd_mem *mem, bool intr)
+		struct amdgpu_device *adev, struct kgd_mem *mem, bool intr)
 {
 	struct amdgpu_sync sync;
 	int ret;
@@ -1828,7 +1905,7 @@ int amdgpu_amdkfd_gpuvm_sync_memory(
 	return ret;
 }
 
-int amdgpu_amdkfd_gpuvm_map_gtt_bo_to_kernel(struct kgd_dev *kgd,
+int amdgpu_amdkfd_gpuvm_map_gtt_bo_to_kernel(struct amdgpu_device *adev,
 		struct kgd_mem *mem, void **kptr, uint64_t *size)
 {
 	int ret;
@@ -1884,7 +1961,8 @@ bo_reserve_failed:
 	return ret;
 }
 
-void amdgpu_amdkfd_gpuvm_unmap_gtt_bo_from_kernel(struct kgd_dev *kgd, struct kgd_mem *mem)
+void amdgpu_amdkfd_gpuvm_unmap_gtt_bo_from_kernel(struct amdgpu_device *adev,
+						  struct kgd_mem *mem)
 {
 	struct amdgpu_bo *bo = mem->bo;
 
@@ -1894,12 +1972,9 @@ void amdgpu_amdkfd_gpuvm_unmap_gtt_bo_from_kernel(struct kgd_dev *kgd, struct kg
 	amdgpu_bo_unreserve(bo);
 }
 
-int amdgpu_amdkfd_gpuvm_get_vm_fault_info(struct kgd_dev *kgd,
-					      struct kfd_vm_fault_info *mem)
+int amdgpu_amdkfd_gpuvm_get_vm_fault_info(struct amdgpu_device *adev,
+					  struct kfd_vm_fault_info *mem)
 {
-	struct amdgpu_device *adev;
-
-	adev = (struct amdgpu_device *)kgd;
 	if (atomic_read(&adev->gmc.vm_fault_info_updated) == 1) {
 		*mem = *adev->gmc.vm_fault_info;
 		mb();
@@ -1908,13 +1983,12 @@ int amdgpu_amdkfd_gpuvm_get_vm_fault_info(struct kgd_dev *kgd,
 	return 0;
 }
 
-int amdgpu_amdkfd_gpuvm_import_dmabuf(struct kgd_dev *kgd,
+int amdgpu_amdkfd_gpuvm_import_dmabuf(struct amdgpu_device *adev,
 				      struct dma_buf *dma_buf,
 				      uint64_t va, void *drm_priv,
 				      struct kgd_mem **mem, uint64_t *size,
 				      uint64_t *mmap_offset)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)kgd;
 	struct amdgpu_vm *avm = drm_priv_to_vm(drm_priv);
 	struct drm_gem_object *obj;
 	struct amdgpu_bo *bo;
@@ -2541,11 +2615,9 @@ int amdgpu_amdkfd_remove_gws_from_process(void *info, void *mem)
 }
 
 /* Returns GPU-specific tiling mode information */
-int amdgpu_amdkfd_get_tile_config(struct kgd_dev *kgd,
+int amdgpu_amdkfd_get_tile_config(struct amdgpu_device *adev,
 				struct tile_config *config)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)kgd;
-
 	config->gb_addr_config = adev->gfx.config.gb_addr_config;
 	config->tile_config_ptr = adev->gfx.config.tile_mode_array;
 	config->num_tile_configs =
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_atombios.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_atombios.c
index 96b7bb13a2dd..12a6b1c99c93 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_atombios.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_atombios.c
@@ -1569,6 +1569,18 @@ void amdgpu_atombios_scratch_regs_engine_hung(struct amdgpu_device *adev,
 	WREG32(adev->bios_scratch_reg_offset + 3, tmp);
 }
 
+void amdgpu_atombios_scratch_regs_set_backlight_level(struct amdgpu_device *adev,
+						      u32 backlight_level)
+{
+	u32 tmp = RREG32(adev->bios_scratch_reg_offset + 2);
+
+	tmp &= ~ATOM_S2_CURRENT_BL_LEVEL_MASK;
+	tmp |= (backlight_level << ATOM_S2_CURRENT_BL_LEVEL_SHIFT) &
+		ATOM_S2_CURRENT_BL_LEVEL_MASK;
+
+	WREG32(adev->bios_scratch_reg_offset + 2, tmp);
+}
+
 bool amdgpu_atombios_scratch_need_asic_init(struct amdgpu_device *adev)
 {
 	u32 tmp = RREG32(adev->bios_scratch_reg_offset + 7);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_atombios.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_atombios.h
index 8cc0222dba19..27e74b1fc260 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_atombios.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_atombios.h
@@ -185,6 +185,8 @@ bool amdgpu_atombios_has_gpu_virtualization_table(struct amdgpu_device *adev);
 void amdgpu_atombios_scratch_regs_lock(struct amdgpu_device *adev, bool lock);
 void amdgpu_atombios_scratch_regs_engine_hung(struct amdgpu_device *adev,
 					      bool hung);
+void amdgpu_atombios_scratch_regs_set_backlight_level(struct amdgpu_device *adev,
+						      u32 backlight_level);
 bool amdgpu_atombios_scratch_need_asic_init(struct amdgpu_device *adev);
 
 void amdgpu_atombios_copy_swap(u8 *dst, u8 *src, u8 num_bytes, bool to_le);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
index 188accb71249..6acdf7cfe0a4 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
@@ -3687,8 +3687,6 @@ fence_driver_init:
 	/* Get a log2 for easy divisions. */
 	adev->mm_stats.log2_max_MBps = ilog2(max(1u, max_MBps));
 
-	amdgpu_fbdev_init(adev);
-
 	r = amdgpu_pm_sysfs_init(adev);
 	if (r) {
 		adev->pm_sysfs_en = false;
@@ -3846,8 +3844,6 @@ void amdgpu_device_fini_hw(struct amdgpu_device *adev)
 		amdgpu_ucode_sysfs_fini(adev);
 	sysfs_remove_files(&adev->dev->kobj, amdgpu_dev_attributes);
 
-	amdgpu_fbdev_fini(adev);
-
 	amdgpu_device_ip_fini_early(adev);
 
 	amdgpu_irq_fini_hw(adev);
@@ -3942,7 +3938,7 @@ int amdgpu_device_suspend(struct drm_device *dev, bool fbcon)
 	drm_kms_helper_poll_disable(dev);
 
 	if (fbcon)
-		amdgpu_fbdev_set_suspend(adev, 1);
+		drm_fb_helper_set_suspend_unlocked(adev_to_drm(adev)->fb_helper, true);
 
 	cancel_delayed_work_sync(&adev->delayed_init_work);
 
@@ -4019,7 +4015,7 @@ int amdgpu_device_resume(struct drm_device *dev, bool fbcon)
 	flush_delayed_work(&adev->delayed_init_work);
 
 	if (fbcon)
-		amdgpu_fbdev_set_suspend(adev, 0);
+		drm_fb_helper_set_suspend_unlocked(adev_to_drm(adev)->fb_helper, false);
 
 	drm_kms_helper_poll_enable(dev);
 
@@ -4316,7 +4312,6 @@ static int amdgpu_device_reset_sriov(struct amdgpu_device *adev,
 
 	amdgpu_irq_gpu_reset_resume_helper(adev);
 	r = amdgpu_ib_ring_tests(adev);
-	amdgpu_amdkfd_post_reset(adev);
 
 error:
 	if (!r && adev->virt.gim_feature & AMDGIM_FEATURE_GIM_FLR_VRAMLOST) {
@@ -4649,7 +4644,7 @@ int amdgpu_do_asic_reset(struct list_head *device_list_handle,
 				if (r)
 					goto out;
 
-				amdgpu_fbdev_set_suspend(tmp_adev, 0);
+				drm_fb_helper_set_suspend_unlocked(adev_to_drm(tmp_adev)->fb_helper, false);
 
 				/*
 				 * The GPU enters bad state once faulty pages
@@ -5039,7 +5034,7 @@ int amdgpu_device_gpu_recover(struct amdgpu_device *adev,
 		 */
 		amdgpu_unregister_gpu_instance(tmp_adev);
 
-		amdgpu_fbdev_set_suspend(tmp_adev, 1);
+		drm_fb_helper_set_suspend_unlocked(adev_to_drm(adev)->fb_helper, true);
 
 		/* disable ras on ALL IPs */
 		if (!need_emergency_restart &&
@@ -5089,7 +5084,7 @@ retry:	/* Rest of adevs pre asic reset from XGMI hive. */
 
 	tmp_vram_lost_counter = atomic_read(&((adev)->vram_lost_counter));
 	/* Actual ASIC resets if needed.*/
-	/* TODO Implement XGMI hive reset logic for SRIOV */
+	/* Host driver will handle XGMI hive reset for SRIOV */
 	if (amdgpu_sriov_vf(adev)) {
 		r = amdgpu_device_reset_sriov(adev, job ? false : true);
 		if (r)
@@ -5149,8 +5144,8 @@ skip_hw_reset:
 
 skip_sched_resume:
 	list_for_each_entry(tmp_adev, device_list_handle, reset_list) {
-		/* unlock kfd: SRIOV would do it separately */
-		if (!need_emergency_restart && !amdgpu_sriov_vf(tmp_adev))
+		/* unlock kfd */
+		if (!need_emergency_restart)
 	                amdgpu_amdkfd_post_reset(tmp_adev);
 
 		/* kfd_post_reset will do nothing if kfd device is not initialized,
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.c
index 4e3669407518..503995c7ff6c 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.c
@@ -248,8 +248,8 @@ get_from_vram:
 
 	offset = offsetof(struct binary_header, binary_checksum) +
 		sizeof(bhdr->binary_checksum);
-	size = bhdr->binary_size - offset;
-	checksum = bhdr->binary_checksum;
+	size = le16_to_cpu(bhdr->binary_size) - offset;
+	checksum = le16_to_cpu(bhdr->binary_checksum);
 
 	if (!amdgpu_discovery_verify_checksum(adev->mman.discovery_bin + offset,
 					      size, checksum)) {
@@ -270,7 +270,7 @@ get_from_vram:
 	}
 
 	if (!amdgpu_discovery_verify_checksum(adev->mman.discovery_bin + offset,
-					      ihdr->size, checksum)) {
+					      le16_to_cpu(ihdr->size), checksum)) {
 		DRM_ERROR("invalid ip discovery data table checksum\n");
 		r = -EINVAL;
 		goto out;
@@ -282,7 +282,7 @@ get_from_vram:
 	ghdr = (struct gpu_info_header *)(adev->mman.discovery_bin + offset);
 
 	if (!amdgpu_discovery_verify_checksum(adev->mman.discovery_bin + offset,
-				              ghdr->size, checksum)) {
+				              le32_to_cpu(ghdr->size), checksum)) {
 		DRM_ERROR("invalid gc data table checksum\n");
 		r = -EINVAL;
 		goto out;
@@ -489,10 +489,10 @@ void amdgpu_discovery_harvest_ip(struct amdgpu_device *adev)
 			le16_to_cpu(bhdr->table_list[HARVEST_INFO].offset));
 
 	for (i = 0; i < 32; i++) {
-		if (le32_to_cpu(harvest_info->list[i].hw_id) == 0)
+		if (le16_to_cpu(harvest_info->list[i].hw_id) == 0)
 			break;
 
-		switch (le32_to_cpu(harvest_info->list[i].hw_id)) {
+		switch (le16_to_cpu(harvest_info->list[i].hw_id)) {
 		case VCN_HWID:
 			vcn_harvest_count++;
 			if (harvest_info->list[i].number_instance == 0)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_display.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_display.c
index 68108f151dad..18cc7155e667 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_display.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_display.c
@@ -1599,13 +1599,10 @@ int amdgpu_display_suspend_helper(struct amdgpu_device *adev)
 			continue;
 		}
 		robj = gem_to_amdgpu_bo(fb->obj[0]);
-		/* don't unpin kernel fb objects */
-		if (!amdgpu_fbdev_robj_is_fb(adev, robj)) {
-			r = amdgpu_bo_reserve(robj, true);
-			if (r == 0) {
-				amdgpu_bo_unpin(robj);
-				amdgpu_bo_unreserve(robj);
-			}
+		r = amdgpu_bo_reserve(robj, true);
+		if (r == 0) {
+			amdgpu_bo_unpin(robj);
+			amdgpu_bo_unreserve(robj);
 		}
 	}
 	return 0;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
index 438468b82eb6..3a6f125c6dc9 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
@@ -2001,6 +2001,19 @@ retry_init:
 		goto err_pci;
 	}
 
+	/*
+	 * 1. don't init fbdev on hw without DCE
+	 * 2. don't init fbdev if there are no connectors
+	 */
+	if (adev->mode_info.mode_config_initialized &&
+	    !list_empty(&adev_to_drm(adev)->mode_config.connector_list)) {
+		/* select 8 bpp console on low vram cards */
+		if (adev->gmc.real_vram_size <= (32*1024*1024))
+			drm_fbdev_generic_setup(adev_to_drm(adev), 8);
+		else
+			drm_fbdev_generic_setup(adev_to_drm(adev), 32);
+	}
+
 	ret = amdgpu_debugfs_init(adev);
 	if (ret)
 		DRM_ERROR("Creating debugfs files failed (%d).\n", ret);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_fb.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_fb.c
deleted file mode 100644
index cd0acbea75da..000000000000
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_fb.c
+++ /dev/null
@@ -1,388 +0,0 @@
-/*
- * Copyright © 2007 David Airlie
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * the rights to use, copy, modify, merge, publish, distribute, sublicense,
- * and/or sell copies of the Software, and to permit persons to whom the
- * Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice (including the next
- * paragraph) shall be included in all copies or substantial portions of the
- * Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
- * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
- * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
- * DEALINGS IN THE SOFTWARE.
- *
- * Authors:
- *     David Airlie
- */
-
-#include <linux/module.h>
-#include <linux/pm_runtime.h>
-#include <linux/slab.h>
-#include <linux/vga_switcheroo.h>
-
-#include <drm/amdgpu_drm.h>
-#include <drm/drm_crtc.h>
-#include <drm/drm_crtc_helper.h>
-#include <drm/drm_fb_helper.h>
-#include <drm/drm_fourcc.h>
-
-#include "amdgpu.h"
-#include "cikd.h"
-#include "amdgpu_gem.h"
-
-#include "amdgpu_display.h"
-
-/* object hierarchy -
-   this contains a helper + a amdgpu fb
-   the helper contains a pointer to amdgpu framebuffer baseclass.
-*/
-
-static int
-amdgpufb_open(struct fb_info *info, int user)
-{
-	struct drm_fb_helper *fb_helper = info->par;
-	int ret = pm_runtime_get_sync(fb_helper->dev->dev);
-	if (ret < 0 && ret != -EACCES) {
-		pm_runtime_mark_last_busy(fb_helper->dev->dev);
-		pm_runtime_put_autosuspend(fb_helper->dev->dev);
-		return ret;
-	}
-	return 0;
-}
-
-static int
-amdgpufb_release(struct fb_info *info, int user)
-{
-	struct drm_fb_helper *fb_helper = info->par;
-
-	pm_runtime_mark_last_busy(fb_helper->dev->dev);
-	pm_runtime_put_autosuspend(fb_helper->dev->dev);
-	return 0;
-}
-
-static const struct fb_ops amdgpufb_ops = {
-	.owner = THIS_MODULE,
-	DRM_FB_HELPER_DEFAULT_OPS,
-	.fb_open = amdgpufb_open,
-	.fb_release = amdgpufb_release,
-	.fb_fillrect = drm_fb_helper_cfb_fillrect,
-	.fb_copyarea = drm_fb_helper_cfb_copyarea,
-	.fb_imageblit = drm_fb_helper_cfb_imageblit,
-};
-
-
-int amdgpu_align_pitch(struct amdgpu_device *adev, int width, int cpp, bool tiled)
-{
-	int aligned = width;
-	int pitch_mask = 0;
-
-	switch (cpp) {
-	case 1:
-		pitch_mask = 255;
-		break;
-	case 2:
-		pitch_mask = 127;
-		break;
-	case 3:
-	case 4:
-		pitch_mask = 63;
-		break;
-	}
-
-	aligned += pitch_mask;
-	aligned &= ~pitch_mask;
-	return aligned * cpp;
-}
-
-static void amdgpufb_destroy_pinned_object(struct drm_gem_object *gobj)
-{
-	struct amdgpu_bo *abo = gem_to_amdgpu_bo(gobj);
-	int ret;
-
-	ret = amdgpu_bo_reserve(abo, true);
-	if (likely(ret == 0)) {
-		amdgpu_bo_kunmap(abo);
-		amdgpu_bo_unpin(abo);
-		amdgpu_bo_unreserve(abo);
-	}
-	drm_gem_object_put(gobj);
-}
-
-static int amdgpufb_create_pinned_object(struct amdgpu_fbdev *rfbdev,
-					 struct drm_mode_fb_cmd2 *mode_cmd,
-					 struct drm_gem_object **gobj_p)
-{
-	const struct drm_format_info *info;
-	struct amdgpu_device *adev = rfbdev->adev;
-	struct drm_gem_object *gobj = NULL;
-	struct amdgpu_bo *abo = NULL;
-	bool fb_tiled = false; /* useful for testing */
-	u32 tiling_flags = 0, domain;
-	int ret;
-	int aligned_size, size;
-	int height = mode_cmd->height;
-	u32 cpp;
-	u64 flags = AMDGPU_GEM_CREATE_CPU_ACCESS_REQUIRED |
-			       AMDGPU_GEM_CREATE_VRAM_CONTIGUOUS     |
-			       AMDGPU_GEM_CREATE_VRAM_CLEARED;
-
-	info = drm_get_format_info(adev_to_drm(adev), mode_cmd);
-	cpp = info->cpp[0];
-
-	/* need to align pitch with crtc limits */
-	mode_cmd->pitches[0] = amdgpu_align_pitch(adev, mode_cmd->width, cpp,
-						  fb_tiled);
-	domain = amdgpu_display_supported_domains(adev, flags);
-	height = ALIGN(mode_cmd->height, 8);
-	size = mode_cmd->pitches[0] * height;
-	aligned_size = ALIGN(size, PAGE_SIZE);
-	ret = amdgpu_gem_object_create(adev, aligned_size, 0, domain, flags,
-				       ttm_bo_type_device, NULL, &gobj);
-	if (ret) {
-		pr_err("failed to allocate framebuffer (%d)\n", aligned_size);
-		return -ENOMEM;
-	}
-	abo = gem_to_amdgpu_bo(gobj);
-
-	if (fb_tiled)
-		tiling_flags = AMDGPU_TILING_SET(ARRAY_MODE, GRPH_ARRAY_2D_TILED_THIN1);
-
-	ret = amdgpu_bo_reserve(abo, false);
-	if (unlikely(ret != 0))
-		goto out_unref;
-
-	if (tiling_flags) {
-		ret = amdgpu_bo_set_tiling_flags(abo,
-						 tiling_flags);
-		if (ret)
-			dev_err(adev->dev, "FB failed to set tiling flags\n");
-	}
-
-	ret = amdgpu_bo_pin(abo, domain);
-	if (ret) {
-		amdgpu_bo_unreserve(abo);
-		goto out_unref;
-	}
-
-	ret = amdgpu_ttm_alloc_gart(&abo->tbo);
-	if (ret) {
-		amdgpu_bo_unreserve(abo);
-		dev_err(adev->dev, "%p bind failed\n", abo);
-		goto out_unref;
-	}
-
-	ret = amdgpu_bo_kmap(abo, NULL);
-	amdgpu_bo_unreserve(abo);
-	if (ret) {
-		goto out_unref;
-	}
-
-	*gobj_p = gobj;
-	return 0;
-out_unref:
-	amdgpufb_destroy_pinned_object(gobj);
-	*gobj_p = NULL;
-	return ret;
-}
-
-static int amdgpufb_create(struct drm_fb_helper *helper,
-			   struct drm_fb_helper_surface_size *sizes)
-{
-	struct amdgpu_fbdev *rfbdev = (struct amdgpu_fbdev *)helper;
-	struct amdgpu_device *adev = rfbdev->adev;
-	struct fb_info *info;
-	struct drm_framebuffer *fb = NULL;
-	struct drm_mode_fb_cmd2 mode_cmd;
-	struct drm_gem_object *gobj = NULL;
-	struct amdgpu_bo *abo = NULL;
-	int ret;
-
-	memset(&mode_cmd, 0, sizeof(mode_cmd));
-	mode_cmd.width = sizes->surface_width;
-	mode_cmd.height = sizes->surface_height;
-
-	if (sizes->surface_bpp == 24)
-		sizes->surface_bpp = 32;
-
-	mode_cmd.pixel_format = drm_mode_legacy_fb_format(sizes->surface_bpp,
-							  sizes->surface_depth);
-
-	ret = amdgpufb_create_pinned_object(rfbdev, &mode_cmd, &gobj);
-	if (ret) {
-		DRM_ERROR("failed to create fbcon object %d\n", ret);
-		return ret;
-	}
-
-	abo = gem_to_amdgpu_bo(gobj);
-
-	/* okay we have an object now allocate the framebuffer */
-	info = drm_fb_helper_alloc_fbi(helper);
-	if (IS_ERR(info)) {
-		ret = PTR_ERR(info);
-		goto out;
-	}
-
-	ret = amdgpu_display_gem_fb_init(adev_to_drm(adev), &rfbdev->rfb,
-					 &mode_cmd, gobj);
-	if (ret) {
-		DRM_ERROR("failed to initialize framebuffer %d\n", ret);
-		goto out;
-	}
-
-	fb = &rfbdev->rfb.base;
-
-	/* setup helper */
-	rfbdev->helper.fb = fb;
-
-	info->fbops = &amdgpufb_ops;
-
-	info->fix.smem_start = amdgpu_gmc_vram_cpu_pa(adev, abo);
-	info->fix.smem_len = amdgpu_bo_size(abo);
-	info->screen_base = amdgpu_bo_kptr(abo);
-	info->screen_size = amdgpu_bo_size(abo);
-
-	drm_fb_helper_fill_info(info, &rfbdev->helper, sizes);
-
-	/* setup aperture base/size for vesafb takeover */
-	info->apertures->ranges[0].base = adev_to_drm(adev)->mode_config.fb_base;
-	info->apertures->ranges[0].size = adev->gmc.aper_size;
-
-	/* Use default scratch pixmap (info->pixmap.flags = FB_PIXMAP_SYSTEM) */
-
-	if (info->screen_base == NULL) {
-		ret = -ENOSPC;
-		goto out;
-	}
-
-	DRM_INFO("fb mappable at 0x%lX\n",  info->fix.smem_start);
-	DRM_INFO("vram apper at 0x%lX\n",  (unsigned long)adev->gmc.aper_base);
-	DRM_INFO("size %lu\n", (unsigned long)amdgpu_bo_size(abo));
-	DRM_INFO("fb depth is %d\n", fb->format->depth);
-	DRM_INFO("   pitch is %d\n", fb->pitches[0]);
-
-	vga_switcheroo_client_fb_set(adev->pdev, info);
-	return 0;
-
-out:
-	if (fb && ret) {
-		drm_gem_object_put(gobj);
-		drm_framebuffer_unregister_private(fb);
-		drm_framebuffer_cleanup(fb);
-		kfree(fb);
-	}
-	return ret;
-}
-
-static int amdgpu_fbdev_destroy(struct drm_device *dev, struct amdgpu_fbdev *rfbdev)
-{
-	struct amdgpu_framebuffer *rfb = &rfbdev->rfb;
-	int i;
-
-	drm_fb_helper_unregister_fbi(&rfbdev->helper);
-
-	if (rfb->base.obj[0]) {
-		for (i = 0; i < rfb->base.format->num_planes; i++)
-			drm_gem_object_put(rfb->base.obj[0]);
-		amdgpufb_destroy_pinned_object(rfb->base.obj[0]);
-		rfb->base.obj[0] = NULL;
-		drm_framebuffer_unregister_private(&rfb->base);
-		drm_framebuffer_cleanup(&rfb->base);
-	}
-	drm_fb_helper_fini(&rfbdev->helper);
-
-	return 0;
-}
-
-static const struct drm_fb_helper_funcs amdgpu_fb_helper_funcs = {
-	.fb_probe = amdgpufb_create,
-};
-
-int amdgpu_fbdev_init(struct amdgpu_device *adev)
-{
-	struct amdgpu_fbdev *rfbdev;
-	int bpp_sel = 32;
-	int ret;
-
-	/* don't init fbdev on hw without DCE */
-	if (!adev->mode_info.mode_config_initialized)
-		return 0;
-
-	/* don't init fbdev if there are no connectors */
-	if (list_empty(&adev_to_drm(adev)->mode_config.connector_list))
-		return 0;
-
-	/* select 8 bpp console on low vram cards */
-	if (adev->gmc.real_vram_size <= (32*1024*1024))
-		bpp_sel = 8;
-
-	rfbdev = kzalloc(sizeof(struct amdgpu_fbdev), GFP_KERNEL);
-	if (!rfbdev)
-		return -ENOMEM;
-
-	rfbdev->adev = adev;
-	adev->mode_info.rfbdev = rfbdev;
-
-	drm_fb_helper_prepare(adev_to_drm(adev), &rfbdev->helper,
-			      &amdgpu_fb_helper_funcs);
-
-	ret = drm_fb_helper_init(adev_to_drm(adev), &rfbdev->helper);
-	if (ret) {
-		kfree(rfbdev);
-		return ret;
-	}
-
-	/* disable all the possible outputs/crtcs before entering KMS mode */
-	if (!amdgpu_device_has_dc_support(adev) && !amdgpu_virtual_display)
-		drm_helper_disable_unused_functions(adev_to_drm(adev));
-
-	drm_fb_helper_initial_config(&rfbdev->helper, bpp_sel);
-	return 0;
-}
-
-void amdgpu_fbdev_fini(struct amdgpu_device *adev)
-{
-	if (!adev->mode_info.rfbdev)
-		return;
-
-	amdgpu_fbdev_destroy(adev_to_drm(adev), adev->mode_info.rfbdev);
-	kfree(adev->mode_info.rfbdev);
-	adev->mode_info.rfbdev = NULL;
-}
-
-void amdgpu_fbdev_set_suspend(struct amdgpu_device *adev, int state)
-{
-	if (adev->mode_info.rfbdev)
-		drm_fb_helper_set_suspend_unlocked(&adev->mode_info.rfbdev->helper,
-						   state);
-}
-
-int amdgpu_fbdev_total_size(struct amdgpu_device *adev)
-{
-	struct amdgpu_bo *robj;
-	int size = 0;
-
-	if (!adev->mode_info.rfbdev)
-		return 0;
-
-	robj = gem_to_amdgpu_bo(adev->mode_info.rfbdev->rfb.base.obj[0]);
-	size += amdgpu_bo_size(robj);
-	return size;
-}
-
-bool amdgpu_fbdev_robj_is_fb(struct amdgpu_device *adev, struct amdgpu_bo *robj)
-{
-	if (!adev->mode_info.rfbdev)
-		return false;
-	if (robj == gem_to_amdgpu_bo(adev->mode_info.rfbdev->rfb.base.obj[0]))
-		return true;
-	return false;
-}
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c
index a1e63ba4c54a..c0d8f40a5b45 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c
@@ -877,6 +877,32 @@ out:
 	return r;
 }
 
+static int amdgpu_gem_align_pitch(struct amdgpu_device *adev,
+				  int width,
+				  int cpp,
+				  bool tiled)
+{
+	int aligned = width;
+	int pitch_mask = 0;
+
+	switch (cpp) {
+	case 1:
+		pitch_mask = 255;
+		break;
+	case 2:
+		pitch_mask = 127;
+		break;
+	case 3:
+	case 4:
+		pitch_mask = 63;
+		break;
+	}
+
+	aligned += pitch_mask;
+	aligned &= ~pitch_mask;
+	return aligned * cpp;
+}
+
 int amdgpu_mode_dumb_create(struct drm_file *file_priv,
 			    struct drm_device *dev,
 			    struct drm_mode_create_dumb *args)
@@ -885,7 +911,8 @@ int amdgpu_mode_dumb_create(struct drm_file *file_priv,
 	struct drm_gem_object *gobj;
 	uint32_t handle;
 	u64 flags = AMDGPU_GEM_CREATE_CPU_ACCESS_REQUIRED |
-		    AMDGPU_GEM_CREATE_CPU_GTT_USWC;
+		    AMDGPU_GEM_CREATE_CPU_GTT_USWC |
+		    AMDGPU_GEM_CREATE_VRAM_CONTIGUOUS;
 	u32 domain;
 	int r;
 
@@ -897,8 +924,8 @@ int amdgpu_mode_dumb_create(struct drm_file *file_priv,
 	if (adev->mman.buffer_funcs_enabled)
 		flags |= AMDGPU_GEM_CREATE_VRAM_CLEARED;
 
-	args->pitch = amdgpu_align_pitch(adev, args->width,
-					 DIV_ROUND_UP(args->bpp, 8), 0);
+	args->pitch = amdgpu_gem_align_pitch(adev, args->width,
+					     DIV_ROUND_UP(args->bpp, 8), 0);
 	args->size = (u64)args->pitch * args->height;
 	args->size = ALIGN(args->size, PAGE_SIZE);
 	domain = amdgpu_bo_get_preferred_domain(adev,
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ih.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ih.c
index f3d62e196901..0c7963dfacad 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ih.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ih.c
@@ -223,7 +223,7 @@ int amdgpu_ih_wait_on_checkpoint_process(struct amdgpu_device *adev,
  */
 int amdgpu_ih_process(struct amdgpu_device *adev, struct amdgpu_ih_ring *ih)
 {
-	unsigned int count = AMDGPU_IH_MAX_NUM_IVS;
+	unsigned int count;
 	u32 wptr;
 
 	if (!ih->enabled || adev->shutdown)
@@ -232,6 +232,7 @@ int amdgpu_ih_process(struct amdgpu_device *adev, struct amdgpu_ih_ring *ih)
 	wptr = amdgpu_ih_get_wptr(adev, ih);
 
 restart_ih:
+	count  = AMDGPU_IH_MAX_NUM_IVS;
 	DRM_DEBUG("%s: rptr %d, wptr %d\n", __func__, ih->rptr, wptr);
 
 	/* Order reading of wptr vs. reading of IH ring data */
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_irq.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_irq.c
index cc2e0c9cfe0a..4f3c62adccbd 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_irq.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_irq.c
@@ -333,7 +333,6 @@ int amdgpu_irq_init(struct amdgpu_device *adev)
 	if (!amdgpu_device_has_dc_support(adev)) {
 		if (!adev->enable_virtual_display)
 			/* Disable vblank IRQs aggressively for power-saving */
-			/* XXX: can this be enabled for DC? */
 			adev_to_drm(adev)->vblank_disable_immediate = true;
 
 		r = drm_vblank_init(adev_to_drm(adev), adev->mode_info.num_crtc);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_mode.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_mode.h
index 89fb372ed49c..6043bf6fd414 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_mode.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_mode.h
@@ -232,8 +232,6 @@ struct amdgpu_i2c_chan {
 	struct mutex mutex;
 };
 
-struct amdgpu_fbdev;
-
 struct amdgpu_afmt {
 	bool enabled;
 	int offset;
@@ -309,13 +307,6 @@ struct amdgpu_framebuffer {
 	uint64_t address;
 };
 
-struct amdgpu_fbdev {
-	struct drm_fb_helper helper;
-	struct amdgpu_framebuffer rfb;
-	struct list_head fbdev_list;
-	struct amdgpu_device *adev;
-};
-
 struct amdgpu_mode_info {
 	struct atom_context *atom_context;
 	struct card_info *atom_card_info;
@@ -341,8 +332,6 @@ struct amdgpu_mode_info {
 	struct edid *bios_hardcoded_edid;
 	int bios_hardcoded_edid_size;
 
-	/* pointer to fbdev info structure */
-	struct amdgpu_fbdev *rfbdev;
 	/* firmware flags */
 	u32 firmware_flags;
 	/* pointer to backlight encoder */
@@ -631,15 +620,6 @@ bool amdgpu_crtc_get_scanout_position(struct drm_crtc *crtc,
 			int *hpos, ktime_t *stime, ktime_t *etime,
 			const struct drm_display_mode *mode);
 
-/* fbdev layer */
-int amdgpu_fbdev_init(struct amdgpu_device *adev);
-void amdgpu_fbdev_fini(struct amdgpu_device *adev);
-void amdgpu_fbdev_set_suspend(struct amdgpu_device *adev, int state);
-int amdgpu_fbdev_total_size(struct amdgpu_device *adev);
-bool amdgpu_fbdev_robj_is_fb(struct amdgpu_device *adev, struct amdgpu_bo *robj);
-
-int amdgpu_align_pitch(struct amdgpu_device *adev, int width, int bpp, bool tiled);
-
 /* amdgpu_display.c */
 void amdgpu_display_print_display_setup(struct drm_device *dev);
 int amdgpu_display_modeset_create_props(struct amdgpu_device *adev);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c
index 4fcfc2313b8c..3a7b56e57cec 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c
@@ -1032,9 +1032,14 @@ int amdgpu_bo_init(struct amdgpu_device *adev)
 	/* On A+A platform, VRAM can be mapped as WB */
 	if (!adev->gmc.xgmi.connected_to_cpu) {
 		/* reserve PAT memory space to WC for VRAM */
-		arch_io_reserve_memtype_wc(adev->gmc.aper_base,
+		int r = arch_io_reserve_memtype_wc(adev->gmc.aper_base,
 				adev->gmc.aper_size);
 
+		if (r) {
+			DRM_ERROR("Unable to set WC memtype for the aperture base\n");
+			return r;
+		}
+
 		/* Add an MTRR for the VRAM */
 		adev->gmc.vram_mtrr = arch_phys_wc_add(adev->gmc.aper_base,
 				adev->gmc.aper_size);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c
index 08133de21fdd..46910e7b2927 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c
@@ -892,6 +892,38 @@ void amdgpu_ras_mca_query_error_status(struct amdgpu_device *adev,
 	}
 }
 
+static void amdgpu_ras_get_ecc_info(struct amdgpu_device *adev, struct ras_err_data *err_data)
+{
+	struct amdgpu_ras *ras = amdgpu_ras_get_context(adev);
+	int ret = 0;
+
+	/*
+	 * choosing right query method according to
+	 * whether smu support query error information
+	 */
+	ret = smu_get_ecc_info(&adev->smu, (void *)&(ras->umc_ecc));
+	if (ret == -EOPNOTSUPP) {
+		if (adev->umc.ras_funcs &&
+			adev->umc.ras_funcs->query_ras_error_count)
+			adev->umc.ras_funcs->query_ras_error_count(adev, err_data);
+
+		/* umc query_ras_error_address is also responsible for clearing
+		 * error status
+		 */
+		if (adev->umc.ras_funcs &&
+		    adev->umc.ras_funcs->query_ras_error_address)
+			adev->umc.ras_funcs->query_ras_error_address(adev, err_data);
+	} else if (!ret) {
+		if (adev->umc.ras_funcs &&
+			adev->umc.ras_funcs->ecc_info_query_ras_error_count)
+			adev->umc.ras_funcs->ecc_info_query_ras_error_count(adev, err_data);
+
+		if (adev->umc.ras_funcs &&
+			adev->umc.ras_funcs->ecc_info_query_ras_error_address)
+			adev->umc.ras_funcs->ecc_info_query_ras_error_address(adev, err_data);
+	}
+}
+
 /* query/inject/cure begin */
 int amdgpu_ras_query_error_status(struct amdgpu_device *adev,
 				  struct ras_query_if *info)
@@ -905,15 +937,7 @@ int amdgpu_ras_query_error_status(struct amdgpu_device *adev,
 
 	switch (info->head.block) {
 	case AMDGPU_RAS_BLOCK__UMC:
-		if (adev->umc.ras_funcs &&
-		    adev->umc.ras_funcs->query_ras_error_count)
-			adev->umc.ras_funcs->query_ras_error_count(adev, &err_data);
-		/* umc query_ras_error_address is also responsible for clearing
-		 * error status
-		 */
-		if (adev->umc.ras_funcs &&
-		    adev->umc.ras_funcs->query_ras_error_address)
-			adev->umc.ras_funcs->query_ras_error_address(adev, &err_data);
+		amdgpu_ras_get_ecc_info(adev, &err_data);
 		break;
 	case AMDGPU_RAS_BLOCK__SDMA:
 		if (adev->sdma.funcs->query_ras_error_count) {
@@ -1935,9 +1959,11 @@ int amdgpu_ras_save_bad_pages(struct amdgpu_device *adev)
 	if (!con || !con->eh_data)
 		return 0;
 
+	mutex_lock(&con->recovery_lock);
 	control = &con->eeprom_control;
 	data = con->eh_data;
 	save_count = data->count - control->ras_num_recs;
+	mutex_unlock(&con->recovery_lock);
 	/* only new entries are saved */
 	if (save_count > 0) {
 		if (amdgpu_ras_eeprom_append(control,
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.h
index e36f4de9fa55..1c708122d492 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.h
@@ -319,6 +319,19 @@ struct ras_common_if {
 	char name[32];
 };
 
+#define MAX_UMC_CHANNEL_NUM 32
+
+struct ecc_info_per_ch {
+	uint16_t ce_count_lo_chip;
+	uint16_t ce_count_hi_chip;
+	uint64_t mca_umc_status;
+	uint64_t mca_umc_addr;
+};
+
+struct umc_ecc_info {
+	struct ecc_info_per_ch ecc[MAX_UMC_CHANNEL_NUM];
+};
+
 struct amdgpu_ras {
 	/* ras infrastructure */
 	/* for ras itself. */
@@ -358,6 +371,9 @@ struct amdgpu_ras {
 	struct delayed_work ras_counte_delay_work;
 	atomic_t ras_ue_count;
 	atomic_t ras_ce_count;
+
+	/* record umc error info queried from smu */
+	struct umc_ecc_info umc_ecc;
 };
 
 struct ras_fs_data {
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
index c15687ce67c4..fb0d8bffdce2 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
@@ -913,11 +913,6 @@ static int amdgpu_ttm_backend_bind(struct ttm_device *bdev,
 		     ttm->num_pages, bo_mem, ttm);
 	}
 
-	if (bo_mem->mem_type == AMDGPU_PL_GDS ||
-	    bo_mem->mem_type == AMDGPU_PL_GWS ||
-	    bo_mem->mem_type == AMDGPU_PL_OA)
-		return -EINVAL;
-
 	if (bo_mem->mem_type != TTM_PL_TT ||
 	    !amdgpu_gtt_mgr_has_gart_addr(bo_mem)) {
 		gtt->offset = AMDGPU_BO_INVALID_OFFSET;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_umc.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_umc.c
index a90029ee9733..6e4bea012ea4 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_umc.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_umc.c
@@ -94,30 +94,58 @@ int amdgpu_umc_process_ras_data_cb(struct amdgpu_device *adev,
 {
 	struct ras_err_data *err_data = (struct ras_err_data *)ras_error_status;
 	struct amdgpu_ras *con = amdgpu_ras_get_context(adev);
+	int ret = 0;
 
 	kgd2kfd_set_sram_ecc_flag(adev->kfd.dev);
-	if (adev->umc.ras_funcs &&
-	    adev->umc.ras_funcs->query_ras_error_count)
-	    adev->umc.ras_funcs->query_ras_error_count(adev, ras_error_status);
-
-	if (adev->umc.ras_funcs &&
-	    adev->umc.ras_funcs->query_ras_error_address &&
-	    adev->umc.max_ras_err_cnt_per_query) {
-		err_data->err_addr =
-			kcalloc(adev->umc.max_ras_err_cnt_per_query,
-				sizeof(struct eeprom_table_record), GFP_KERNEL);
-
-		/* still call query_ras_error_address to clear error status
-		 * even NOMEM error is encountered
-		 */
-		if(!err_data->err_addr)
-			dev_warn(adev->dev, "Failed to alloc memory for "
-					"umc error address record!\n");
-
-		/* umc query_ras_error_address is also responsible for clearing
-		 * error status
-		 */
-		adev->umc.ras_funcs->query_ras_error_address(adev, ras_error_status);
+	ret = smu_get_ecc_info(&adev->smu, (void *)&(con->umc_ecc));
+	if (ret == -EOPNOTSUPP) {
+		if (adev->umc.ras_funcs &&
+		    adev->umc.ras_funcs->query_ras_error_count)
+		    adev->umc.ras_funcs->query_ras_error_count(adev, ras_error_status);
+
+		if (adev->umc.ras_funcs &&
+		    adev->umc.ras_funcs->query_ras_error_address &&
+		    adev->umc.max_ras_err_cnt_per_query) {
+			err_data->err_addr =
+				kcalloc(adev->umc.max_ras_err_cnt_per_query,
+					sizeof(struct eeprom_table_record), GFP_KERNEL);
+
+			/* still call query_ras_error_address to clear error status
+			 * even NOMEM error is encountered
+			 */
+			if(!err_data->err_addr)
+				dev_warn(adev->dev, "Failed to alloc memory for "
+						"umc error address record!\n");
+
+			/* umc query_ras_error_address is also responsible for clearing
+			 * error status
+			 */
+			adev->umc.ras_funcs->query_ras_error_address(adev, ras_error_status);
+		}
+	} else if (!ret) {
+		if (adev->umc.ras_funcs &&
+		    adev->umc.ras_funcs->ecc_info_query_ras_error_count)
+		    adev->umc.ras_funcs->ecc_info_query_ras_error_count(adev, ras_error_status);
+
+		if (adev->umc.ras_funcs &&
+		    adev->umc.ras_funcs->ecc_info_query_ras_error_address &&
+		    adev->umc.max_ras_err_cnt_per_query) {
+			err_data->err_addr =
+				kcalloc(adev->umc.max_ras_err_cnt_per_query,
+					sizeof(struct eeprom_table_record), GFP_KERNEL);
+
+			/* still call query_ras_error_address to clear error status
+			 * even NOMEM error is encountered
+			 */
+			if(!err_data->err_addr)
+				dev_warn(adev->dev, "Failed to alloc memory for "
+						"umc error address record!\n");
+
+			/* umc query_ras_error_address is also responsible for clearing
+			 * error status
+			 */
+			adev->umc.ras_funcs->ecc_info_query_ras_error_address(adev, ras_error_status);
+		}
 	}
 
 	/* only uncorrectable error needs gpu reset */
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_umc.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_umc.h
index 1f5fe2315236..9e40bade0a68 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_umc.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_umc.h
@@ -49,6 +49,10 @@ struct amdgpu_umc_ras_funcs {
 	void (*query_ras_error_address)(struct amdgpu_device *adev,
 					void *ras_error_status);
 	bool (*query_ras_poison_mode)(struct amdgpu_device *adev);
+	void (*ecc_info_query_ras_error_count)(struct amdgpu_device *adev,
+				      void *ras_error_status);
+	void (*ecc_info_query_ras_error_address)(struct amdgpu_device *adev,
+					void *ras_error_status);
 };
 
 struct amdgpu_umc_funcs {
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c
index 04cf9b207e62..3fc49823f527 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c
@@ -283,17 +283,15 @@ static int amdgpu_virt_init_ras_err_handler_data(struct amdgpu_device *adev)
 
 	*data = kmalloc(sizeof(struct amdgpu_virt_ras_err_handler_data), GFP_KERNEL);
 	if (!*data)
-		return -ENOMEM;
+		goto data_failure;
 
 	bps = kmalloc_array(align_space, sizeof((*data)->bps), GFP_KERNEL);
-	bps_bo = kmalloc_array(align_space, sizeof((*data)->bps_bo), GFP_KERNEL);
+	if (!bps)
+		goto bps_failure;
 
-	if (!bps || !bps_bo) {
-		kfree(bps);
-		kfree(bps_bo);
-		kfree(*data);
-		return -ENOMEM;
-	}
+	bps_bo = kmalloc_array(align_space, sizeof((*data)->bps_bo), GFP_KERNEL);
+	if (!bps_bo)
+		goto bps_bo_failure;
 
 	(*data)->bps = bps;
 	(*data)->bps_bo = bps_bo;
@@ -303,6 +301,13 @@ static int amdgpu_virt_init_ras_err_handler_data(struct amdgpu_device *adev)
 	virt->ras_init_done = true;
 
 	return 0;
+
+bps_bo_failure:
+	kfree(bps);
+bps_failure:
+	kfree(*data);
+data_failure:
+	return -ENOMEM;
 }
 
 static void amdgpu_virt_ras_release_bp(struct amdgpu_device *adev)
diff --git a/drivers/gpu/drm/amd/amdgpu/dce_v8_0.c b/drivers/gpu/drm/amd/amdgpu/dce_v8_0.c
index b200b9e722d9..8318ee8339f1 100644
--- a/drivers/gpu/drm/amd/amdgpu/dce_v8_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/dce_v8_0.c
@@ -2092,22 +2092,18 @@ static int dce_v8_0_pick_dig_encoder(struct drm_encoder *encoder)
 			return 1;
 		else
 			return 0;
-		break;
 	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY1:
 		if (dig->linkb)
 			return 3;
 		else
 			return 2;
-		break;
 	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY2:
 		if (dig->linkb)
 			return 5;
 		else
 			return 4;
-		break;
 	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY3:
 		return 6;
-		break;
 	default:
 		DRM_ERROR("invalid encoder_id: 0x%x\n", amdgpu_encoder->encoder_id);
 		return 0;
diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c
index e7dfeb466a0e..dbe7442fb25c 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c
@@ -7707,8 +7707,19 @@ static uint64_t gfx_v10_0_get_gpu_clock_counter(struct amdgpu_device *adev)
 	switch (adev->ip_versions[GC_HWIP][0]) {
 	case IP_VERSION(10, 3, 1):
 	case IP_VERSION(10, 3, 3):
-		clock = (uint64_t)RREG32_SOC15(SMUIO, 0, mmGOLDEN_TSC_COUNT_LOWER_Vangogh) |
-			((uint64_t)RREG32_SOC15(SMUIO, 0, mmGOLDEN_TSC_COUNT_UPPER_Vangogh) << 32ULL);
+		preempt_disable();
+		clock_hi = RREG32_SOC15_NO_KIQ(SMUIO, 0, mmGOLDEN_TSC_COUNT_UPPER_Vangogh);
+		clock_lo = RREG32_SOC15_NO_KIQ(SMUIO, 0, mmGOLDEN_TSC_COUNT_LOWER_Vangogh);
+		hi_check = RREG32_SOC15_NO_KIQ(SMUIO, 0, mmGOLDEN_TSC_COUNT_UPPER_Vangogh);
+		/* The SMUIO TSC clock frequency is 100MHz, which sets 32-bit carry over
+		 * roughly every 42 seconds.
+		 */
+		if (hi_check != clock_hi) {
+			clock_lo = RREG32_SOC15_NO_KIQ(SMUIO, 0, mmGOLDEN_TSC_COUNT_LOWER_Vangogh);
+			clock_hi = hi_check;
+		}
+		preempt_enable();
+		clock = clock_lo | (clock_hi << 32ULL);
 		break;
 	default:
 		preempt_disable();
diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c
index b4b80f27b894..34478bcc4d09 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c
@@ -140,6 +140,11 @@ MODULE_FIRMWARE("amdgpu/aldebaran_rlc.bin");
 #define mmTCP_CHAN_STEER_5_ARCT								0x0b0c
 #define mmTCP_CHAN_STEER_5_ARCT_BASE_IDX							0
 
+#define mmGOLDEN_TSC_COUNT_UPPER_Renoir                0x0025
+#define mmGOLDEN_TSC_COUNT_UPPER_Renoir_BASE_IDX       1
+#define mmGOLDEN_TSC_COUNT_LOWER_Renoir                0x0026
+#define mmGOLDEN_TSC_COUNT_LOWER_Renoir_BASE_IDX       1
+
 enum ta_ras_gfx_subblock {
 	/*CPC*/
 	TA_RAS_BLOCK__GFX_CPC_INDEX_START = 0,
@@ -4238,19 +4243,38 @@ failed_kiq_read:
 
 static uint64_t gfx_v9_0_get_gpu_clock_counter(struct amdgpu_device *adev)
 {
-	uint64_t clock;
+	uint64_t clock, clock_lo, clock_hi, hi_check;
 
-	amdgpu_gfx_off_ctrl(adev, false);
-	mutex_lock(&adev->gfx.gpu_clock_mutex);
-	if (adev->ip_versions[GC_HWIP][0] == IP_VERSION(9, 0, 1) && amdgpu_sriov_runtime(adev)) {
-		clock = gfx_v9_0_kiq_read_clock(adev);
-	} else {
-		WREG32_SOC15(GC, 0, mmRLC_CAPTURE_GPU_CLOCK_COUNT, 1);
-		clock = (uint64_t)RREG32_SOC15(GC, 0, mmRLC_GPU_CLOCK_COUNT_LSB) |
-			((uint64_t)RREG32_SOC15(GC, 0, mmRLC_GPU_CLOCK_COUNT_MSB) << 32ULL);
+	switch (adev->ip_versions[GC_HWIP][0]) {
+	case IP_VERSION(9, 3, 0):
+		preempt_disable();
+		clock_hi = RREG32_SOC15_NO_KIQ(SMUIO, 0, mmGOLDEN_TSC_COUNT_UPPER_Renoir);
+		clock_lo = RREG32_SOC15_NO_KIQ(SMUIO, 0, mmGOLDEN_TSC_COUNT_LOWER_Renoir);
+		hi_check = RREG32_SOC15_NO_KIQ(SMUIO, 0, mmGOLDEN_TSC_COUNT_UPPER_Renoir);
+		/* The SMUIO TSC clock frequency is 100MHz, which sets 32-bit carry over
+		 * roughly every 42 seconds.
+		 */
+		if (hi_check != clock_hi) {
+			clock_lo = RREG32_SOC15_NO_KIQ(SMUIO, 0, mmGOLDEN_TSC_COUNT_LOWER_Renoir);
+			clock_hi = hi_check;
+		}
+		preempt_enable();
+		clock = clock_lo | (clock_hi << 32ULL);
+		break;
+	default:
+		amdgpu_gfx_off_ctrl(adev, false);
+		mutex_lock(&adev->gfx.gpu_clock_mutex);
+		if (adev->ip_versions[GC_HWIP][0] == IP_VERSION(9, 0, 1) && amdgpu_sriov_runtime(adev)) {
+			clock = gfx_v9_0_kiq_read_clock(adev);
+		} else {
+			WREG32_SOC15(GC, 0, mmRLC_CAPTURE_GPU_CLOCK_COUNT, 1);
+			clock = (uint64_t)RREG32_SOC15(GC, 0, mmRLC_GPU_CLOCK_COUNT_LSB) |
+				((uint64_t)RREG32_SOC15(GC, 0, mmRLC_GPU_CLOCK_COUNT_MSB) << 32ULL);
+		}
+		mutex_unlock(&adev->gfx.gpu_clock_mutex);
+		amdgpu_gfx_off_ctrl(adev, true);
+		break;
 	}
-	mutex_unlock(&adev->gfx.gpu_clock_mutex);
-	amdgpu_gfx_off_ctrl(adev, true);
 	return clock;
 }
 
diff --git a/drivers/gpu/drm/amd/amdgpu/navi10_ih.c b/drivers/gpu/drm/amd/amdgpu/navi10_ih.c
index 1d8414c3fadb..38241cf0e1f1 100644
--- a/drivers/gpu/drm/amd/amdgpu/navi10_ih.c
+++ b/drivers/gpu/drm/amd/amdgpu/navi10_ih.c
@@ -160,6 +160,7 @@ static int navi10_ih_toggle_ring_interrupts(struct amdgpu_device *adev,
 
 	tmp = RREG32(ih_regs->ih_rb_cntl);
 	tmp = REG_SET_FIELD(tmp, IH_RB_CNTL, RB_ENABLE, (enable ? 1 : 0));
+	tmp = REG_SET_FIELD(tmp, IH_RB_CNTL, RB_GPU_TS_ENABLE, 1);
 	/* enable_intr field is only valid in ring0 */
 	if (ih == &adev->irq.ih)
 		tmp = REG_SET_FIELD(tmp, IH_RB_CNTL, ENABLE_INTR, (enable ? 1 : 0));
@@ -275,10 +276,8 @@ static int navi10_ih_enable_ring(struct amdgpu_device *adev,
 	tmp = navi10_ih_rb_cntl(ih, tmp);
 	if (ih == &adev->irq.ih)
 		tmp = REG_SET_FIELD(tmp, IH_RB_CNTL, RPTR_REARM, !!adev->irq.msi_enabled);
-	if (ih == &adev->irq.ih1) {
-		tmp = REG_SET_FIELD(tmp, IH_RB_CNTL, WPTR_OVERFLOW_ENABLE, 0);
+	if (ih == &adev->irq.ih1)
 		tmp = REG_SET_FIELD(tmp, IH_RB_CNTL, RB_FULL_DRAIN_ENABLE, 1);
-	}
 
 	if (amdgpu_sriov_vf(adev) && amdgpu_sriov_reg_indirect_ih(adev)) {
 		if (psp_reg_program(&adev->psp, ih_regs->psp_reg_id, tmp)) {
@@ -319,7 +318,6 @@ static int navi10_ih_irq_init(struct amdgpu_device *adev)
 {
 	struct amdgpu_ih_ring *ih[] = {&adev->irq.ih, &adev->irq.ih1, &adev->irq.ih2};
 	u32 ih_chicken;
-	u32 tmp;
 	int ret;
 	int i;
 
@@ -363,15 +361,6 @@ static int navi10_ih_irq_init(struct amdgpu_device *adev)
 	adev->nbio.funcs->ih_doorbell_range(adev, ih[0]->use_doorbell,
 					    ih[0]->doorbell_index);
 
-	tmp = RREG32_SOC15(OSSSYS, 0, mmIH_STORM_CLIENT_LIST_CNTL);
-	tmp = REG_SET_FIELD(tmp, IH_STORM_CLIENT_LIST_CNTL,
-			    CLIENT18_IS_STORM_CLIENT, 1);
-	WREG32_SOC15(OSSSYS, 0, mmIH_STORM_CLIENT_LIST_CNTL, tmp);
-
-	tmp = RREG32_SOC15(OSSSYS, 0, mmIH_INT_FLOOD_CNTL);
-	tmp = REG_SET_FIELD(tmp, IH_INT_FLOOD_CNTL, FLOOD_CNTL_ENABLE, 1);
-	WREG32_SOC15(OSSSYS, 0, mmIH_INT_FLOOD_CNTL, tmp);
-
 	pci_set_master(adev->pdev);
 
 	/* enable interrupts */
@@ -420,12 +409,19 @@ static u32 navi10_ih_get_wptr(struct amdgpu_device *adev,
 	u32 wptr, tmp;
 	struct amdgpu_ih_regs *ih_regs;
 
-	wptr = le32_to_cpu(*ih->wptr_cpu);
-	ih_regs = &ih->ih_regs;
+	if (ih == &adev->irq.ih) {
+		/* Only ring0 supports writeback. On other rings fall back
+		 * to register-based code with overflow checking below.
+		 */
+		wptr = le32_to_cpu(*ih->wptr_cpu);
 
-	if (!REG_GET_FIELD(wptr, IH_RB_WPTR, RB_OVERFLOW))
-		goto out;
+		if (!REG_GET_FIELD(wptr, IH_RB_WPTR, RB_OVERFLOW))
+			goto out;
+	}
 
+	ih_regs = &ih->ih_regs;
+
+	/* Double check that the overflow wasn't already cleared. */
 	wptr = RREG32_NO_KIQ(ih_regs->ih_rb_wptr);
 	if (!REG_GET_FIELD(wptr, IH_RB_WPTR, RB_OVERFLOW))
 		goto out;
@@ -513,15 +509,11 @@ static int navi10_ih_self_irq(struct amdgpu_device *adev,
 			      struct amdgpu_irq_src *source,
 			      struct amdgpu_iv_entry *entry)
 {
-	uint32_t wptr = cpu_to_le32(entry->src_data[0]);
-
 	switch (entry->ring_id) {
 	case 1:
-		*adev->irq.ih1.wptr_cpu = wptr;
 		schedule_work(&adev->irq.ih1_work);
 		break;
 	case 2:
-		*adev->irq.ih2.wptr_cpu = wptr;
 		schedule_work(&adev->irq.ih2_work);
 		break;
 	default: break;
diff --git a/drivers/gpu/drm/amd/amdgpu/nbio_v2_3.c b/drivers/gpu/drm/amd/amdgpu/nbio_v2_3.c
index 4ecd2b5808ce..ee7cab37dfd5 100644
--- a/drivers/gpu/drm/amd/amdgpu/nbio_v2_3.c
+++ b/drivers/gpu/drm/amd/amdgpu/nbio_v2_3.c
@@ -359,6 +359,10 @@ static void nbio_v2_3_init_registers(struct amdgpu_device *adev)
 
 	if (def != data)
 		WREG32_PCIE(smnPCIE_CONFIG_CNTL, data);
+
+	if (amdgpu_sriov_vf(adev))
+		adev->rmmio_remap.reg_offset = SOC15_REG_OFFSET(NBIO, 0,
+			mmBIF_BX_DEV0_EPF0_VF0_HDP_MEM_COHERENCY_FLUSH_CNTL) << 2;
 }
 
 #define NAVI10_PCIE__LC_L0S_INACTIVITY_DEFAULT		0x00000000 // off by default, no gains over L1
diff --git a/drivers/gpu/drm/amd/amdgpu/nbio_v6_1.c b/drivers/gpu/drm/amd/amdgpu/nbio_v6_1.c
index 0d2d629e2d6a..4bbacf1be25a 100644
--- a/drivers/gpu/drm/amd/amdgpu/nbio_v6_1.c
+++ b/drivers/gpu/drm/amd/amdgpu/nbio_v6_1.c
@@ -276,6 +276,10 @@ static void nbio_v6_1_init_registers(struct amdgpu_device *adev)
 
 	if (def != data)
 		WREG32_PCIE(smnPCIE_CI_CNTL, data);
+
+	if (amdgpu_sriov_vf(adev))
+		adev->rmmio_remap.reg_offset = SOC15_REG_OFFSET(NBIO, 0,
+			mmBIF_BX_DEV0_EPF0_VF0_HDP_MEM_COHERENCY_FLUSH_CNTL) << 2;
 }
 
 static void nbio_v6_1_program_ltr(struct amdgpu_device *adev)
diff --git a/drivers/gpu/drm/amd/amdgpu/nbio_v7_0.c b/drivers/gpu/drm/amd/amdgpu/nbio_v7_0.c
index 3c00666a13e1..37a4039fdfc5 100644
--- a/drivers/gpu/drm/amd/amdgpu/nbio_v7_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/nbio_v7_0.c
@@ -273,7 +273,9 @@ const struct nbio_hdp_flush_reg nbio_v7_0_hdp_flush_reg = {
 
 static void nbio_v7_0_init_registers(struct amdgpu_device *adev)
 {
-
+	if (amdgpu_sriov_vf(adev))
+		adev->rmmio_remap.reg_offset =
+			SOC15_REG_OFFSET(NBIO, 0, mmHDP_MEM_COHERENCY_FLUSH_CNTL) << 2;
 }
 
 const struct amdgpu_nbio_funcs nbio_v7_0_funcs = {
diff --git a/drivers/gpu/drm/amd/amdgpu/nbio_v7_2.c b/drivers/gpu/drm/amd/amdgpu/nbio_v7_2.c
index 8f2a315e7c73..3444332ea110 100644
--- a/drivers/gpu/drm/amd/amdgpu/nbio_v7_2.c
+++ b/drivers/gpu/drm/amd/amdgpu/nbio_v7_2.c
@@ -371,6 +371,10 @@ static void nbio_v7_2_init_registers(struct amdgpu_device *adev)
 		if (def != data)
 			WREG32_PCIE_PORT(SOC15_REG_OFFSET(NBIO, 0, regPCIE_CONFIG_CNTL), data);
 	}
+
+	if (amdgpu_sriov_vf(adev))
+		adev->rmmio_remap.reg_offset = SOC15_REG_OFFSET(NBIO, 0,
+			regBIF_BX_PF0_HDP_MEM_COHERENCY_FLUSH_CNTL) << 2;
 }
 
 const struct amdgpu_nbio_funcs nbio_v7_2_funcs = {
diff --git a/drivers/gpu/drm/amd/amdgpu/nbio_v7_4.c b/drivers/gpu/drm/amd/amdgpu/nbio_v7_4.c
index b8bd03d16dba..dc5e93756fea 100644
--- a/drivers/gpu/drm/amd/amdgpu/nbio_v7_4.c
+++ b/drivers/gpu/drm/amd/amdgpu/nbio_v7_4.c
@@ -362,7 +362,9 @@ const struct nbio_hdp_flush_reg nbio_v7_4_hdp_flush_reg_ald = {
 
 static void nbio_v7_4_init_registers(struct amdgpu_device *adev)
 {
-
+	if (amdgpu_sriov_vf(adev))
+		adev->rmmio_remap.reg_offset = SOC15_REG_OFFSET(NBIO, 0,
+			mmBIF_BX_DEV0_EPF0_VF0_HDP_MEM_COHERENCY_FLUSH_CNTL) << 2;
 }
 
 static void nbio_v7_4_handle_ras_controller_intr_no_bifring(struct amdgpu_device *adev)
@@ -692,6 +694,9 @@ static void nbio_v7_4_program_aspm(struct amdgpu_device *adev)
 {
 	uint32_t def, data;
 
+	if (adev->ip_versions[NBIO_HWIP][0] == IP_VERSION(7, 4, 4))
+		return;
+
 	def = data = RREG32_PCIE(smnPCIE_LC_CNTL);
 	data &= ~PCIE_LC_CNTL__LC_L1_INACTIVITY_MASK;
 	data &= ~PCIE_LC_CNTL__LC_L0S_INACTIVITY_MASK;
diff --git a/drivers/gpu/drm/amd/amdgpu/nv.c b/drivers/gpu/drm/amd/amdgpu/nv.c
index 59eafa31c626..a6659d9ecdd2 100644
--- a/drivers/gpu/drm/amd/amdgpu/nv.c
+++ b/drivers/gpu/drm/amd/amdgpu/nv.c
@@ -731,8 +731,10 @@ static int nv_common_early_init(void *handle)
 #define MMIO_REG_HOLE_OFFSET (0x80000 - PAGE_SIZE)
 	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
 
-	adev->rmmio_remap.reg_offset = MMIO_REG_HOLE_OFFSET;
-	adev->rmmio_remap.bus_addr = adev->rmmio_base + MMIO_REG_HOLE_OFFSET;
+	if (!amdgpu_sriov_vf(adev)) {
+		adev->rmmio_remap.reg_offset = MMIO_REG_HOLE_OFFSET;
+		adev->rmmio_remap.bus_addr = adev->rmmio_base + MMIO_REG_HOLE_OFFSET;
+	}
 	adev->smc_rreg = NULL;
 	adev->smc_wreg = NULL;
 	adev->pcie_rreg = &nv_pcie_rreg;
@@ -1032,7 +1034,7 @@ static int nv_common_hw_init(void *handle)
 	 * for the purpose of expose those registers
 	 * to process space
 	 */
-	if (adev->nbio.funcs->remap_hdp_registers)
+	if (adev->nbio.funcs->remap_hdp_registers && !amdgpu_sriov_vf(adev))
 		adev->nbio.funcs->remap_hdp_registers(adev);
 	/* enable the doorbell aperture */
 	nv_enable_doorbell_aperture(adev, true);
diff --git a/drivers/gpu/drm/amd/amdgpu/soc15.c b/drivers/gpu/drm/amd/amdgpu/soc15.c
index 0c316a2d42ed..de9b55383e9f 100644
--- a/drivers/gpu/drm/amd/amdgpu/soc15.c
+++ b/drivers/gpu/drm/amd/amdgpu/soc15.c
@@ -971,8 +971,10 @@ static int soc15_common_early_init(void *handle)
 #define MMIO_REG_HOLE_OFFSET (0x80000 - PAGE_SIZE)
 	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
 
-	adev->rmmio_remap.reg_offset = MMIO_REG_HOLE_OFFSET;
-	adev->rmmio_remap.bus_addr = adev->rmmio_base + MMIO_REG_HOLE_OFFSET;
+	if (!amdgpu_sriov_vf(adev)) {
+		adev->rmmio_remap.reg_offset = MMIO_REG_HOLE_OFFSET;
+		adev->rmmio_remap.bus_addr = adev->rmmio_base + MMIO_REG_HOLE_OFFSET;
+	}
 	adev->smc_rreg = NULL;
 	adev->smc_wreg = NULL;
 	adev->pcie_rreg = &soc15_pcie_rreg;
@@ -1285,7 +1287,7 @@ static int soc15_common_hw_init(void *handle)
 	 * for the purpose of expose those registers
 	 * to process space
 	 */
-	if (adev->nbio.funcs->remap_hdp_registers)
+	if (adev->nbio.funcs->remap_hdp_registers && !amdgpu_sriov_vf(adev))
 		adev->nbio.funcs->remap_hdp_registers(adev);
 
 	/* enable the doorbell aperture */
diff --git a/drivers/gpu/drm/amd/amdgpu/umc_v6_7.c b/drivers/gpu/drm/amd/amdgpu/umc_v6_7.c
index f7ec3fe134e5..6dd1e19e8d43 100644
--- a/drivers/gpu/drm/amd/amdgpu/umc_v6_7.c
+++ b/drivers/gpu/drm/amd/amdgpu/umc_v6_7.c
@@ -50,6 +50,165 @@ static inline uint32_t get_umc_v6_7_reg_offset(struct amdgpu_device *adev,
 	return adev->umc.channel_offs * ch_inst + UMC_V6_7_INST_DIST * umc_inst;
 }
 
+static inline uint32_t get_umc_v6_7_channel_index(struct amdgpu_device *adev,
+					      uint32_t umc_inst,
+					      uint32_t ch_inst)
+{
+	return adev->umc.channel_idx_tbl[umc_inst * adev->umc.channel_inst_num + ch_inst];
+}
+
+static void umc_v6_7_ecc_info_query_correctable_error_count(struct amdgpu_device *adev,
+						   uint32_t channel_index,
+						   unsigned long *error_count)
+{
+	uint32_t ecc_err_cnt;
+	uint64_t mc_umc_status;
+	struct amdgpu_ras *ras = amdgpu_ras_get_context(adev);
+
+	/*
+	 * select the lower chip and check the error count
+	 * skip add error count, calc error counter only from mca_umc_status
+	 */
+	ecc_err_cnt = ras->umc_ecc.ecc[channel_index].ce_count_lo_chip;
+
+	/*
+	 * select the higher chip and check the err counter
+	 * skip add error count, calc error counter only from mca_umc_status
+	 */
+	ecc_err_cnt = ras->umc_ecc.ecc[channel_index].ce_count_hi_chip;
+
+	/* check for SRAM correctable error
+	  MCUMC_STATUS is a 64 bit register */
+	mc_umc_status = ras->umc_ecc.ecc[channel_index].mca_umc_status;
+	if (REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, Val) == 1 &&
+	    REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, CECC) == 1)
+		*error_count += 1;
+}
+
+static void umc_v6_7_ecc_info_querry_uncorrectable_error_count(struct amdgpu_device *adev,
+						      uint32_t channel_index,
+						      unsigned long *error_count)
+{
+	uint64_t mc_umc_status;
+	struct amdgpu_ras *ras = amdgpu_ras_get_context(adev);
+
+	/* check the MCUMC_STATUS */
+	mc_umc_status = ras->umc_ecc.ecc[channel_index].mca_umc_status;
+	if ((REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, Val) == 1) &&
+	    (REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, Deferred) == 1 ||
+	    REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, UECC) == 1 ||
+	    REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, PCC) == 1 ||
+	    REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, UC) == 1 ||
+	    REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, TCC) == 1))
+		*error_count += 1;
+}
+
+static void umc_v6_7_ecc_info_query_ras_error_count(struct amdgpu_device *adev,
+					   void *ras_error_status)
+{
+	struct ras_err_data *err_data = (struct ras_err_data *)ras_error_status;
+
+	uint32_t umc_inst        = 0;
+	uint32_t ch_inst         = 0;
+	uint32_t umc_reg_offset  = 0;
+	uint32_t channel_index	 = 0;
+
+	/*TODO: driver needs to toggle DF Cstate to ensure
+	 * safe access of UMC registers. Will add the protection */
+	LOOP_UMC_INST_AND_CH(umc_inst, ch_inst) {
+		umc_reg_offset = get_umc_v6_7_reg_offset(adev,
+							 umc_inst,
+							 ch_inst);
+		channel_index = get_umc_v6_7_channel_index(adev,
+							 umc_inst,
+							 ch_inst);
+		umc_v6_7_ecc_info_query_correctable_error_count(adev,
+						      channel_index,
+						      &(err_data->ce_count));
+		umc_v6_7_ecc_info_querry_uncorrectable_error_count(adev,
+							  channel_index,
+							  &(err_data->ue_count));
+	}
+}
+
+static void umc_v6_7_ecc_info_query_error_address(struct amdgpu_device *adev,
+					 struct ras_err_data *err_data,
+					 uint32_t umc_reg_offset,
+					 uint32_t ch_inst,
+					 uint32_t umc_inst)
+{
+	uint64_t mc_umc_status, err_addr, retired_page;
+	struct eeprom_table_record *err_rec;
+	uint32_t channel_index;
+	struct amdgpu_ras *ras = amdgpu_ras_get_context(adev);
+
+	channel_index =
+		adev->umc.channel_idx_tbl[umc_inst * adev->umc.channel_inst_num + ch_inst];
+
+	mc_umc_status = ras->umc_ecc.ecc[channel_index].mca_umc_status;
+
+	if (mc_umc_status == 0)
+		return;
+
+	if (!err_data->err_addr)
+		return;
+
+	err_rec = &err_data->err_addr[err_data->err_addr_cnt];
+
+	/* calculate error address if ue/ce error is detected */
+	if (REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, Val) == 1 &&
+	    (REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, UECC) == 1 ||
+	    REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, CECC) == 1)) {
+
+		err_addr = ras->umc_ecc.ecc[channel_index].mca_umc_addr;
+		err_addr = REG_GET_FIELD(err_addr, MCA_UMC_UMC0_MCUMC_ADDRT0, ErrorAddr);
+
+		/* translate umc channel address to soc pa, 3 parts are included */
+		retired_page = ADDR_OF_8KB_BLOCK(err_addr) |
+				ADDR_OF_256B_BLOCK(channel_index) |
+				OFFSET_IN_256B_BLOCK(err_addr);
+
+		/* we only save ue error information currently, ce is skipped */
+		if (REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, UECC)
+				== 1) {
+			err_rec->address = err_addr;
+			/* page frame address is saved */
+			err_rec->retired_page = retired_page >> AMDGPU_GPU_PAGE_SHIFT;
+			err_rec->ts = (uint64_t)ktime_get_real_seconds();
+			err_rec->err_type = AMDGPU_RAS_EEPROM_ERR_NON_RECOVERABLE;
+			err_rec->cu = 0;
+			err_rec->mem_channel = channel_index;
+			err_rec->mcumc_id = umc_inst;
+
+			err_data->err_addr_cnt++;
+		}
+	}
+}
+
+static void umc_v6_7_ecc_info_query_ras_error_address(struct amdgpu_device *adev,
+					     void *ras_error_status)
+{
+	struct ras_err_data *err_data = (struct ras_err_data *)ras_error_status;
+
+	uint32_t umc_inst        = 0;
+	uint32_t ch_inst         = 0;
+	uint32_t umc_reg_offset  = 0;
+
+	/*TODO: driver needs to toggle DF Cstate to ensure
+	 * safe access of UMC resgisters. Will add the protection
+	 * when firmware interface is ready */
+	LOOP_UMC_INST_AND_CH(umc_inst, ch_inst) {
+		umc_reg_offset = get_umc_v6_7_reg_offset(adev,
+							 umc_inst,
+							 ch_inst);
+		umc_v6_7_ecc_info_query_error_address(adev,
+					     err_data,
+					     umc_reg_offset,
+					     ch_inst,
+					     umc_inst);
+	}
+}
+
 static void umc_v6_7_query_correctable_error_count(struct amdgpu_device *adev,
 						   uint32_t umc_reg_offset,
 						   unsigned long *error_count)
@@ -327,4 +486,6 @@ const struct amdgpu_umc_ras_funcs umc_v6_7_ras_funcs = {
 	.query_ras_error_count = umc_v6_7_query_ras_error_count,
 	.query_ras_error_address = umc_v6_7_query_ras_error_address,
 	.query_ras_poison_mode = umc_v6_7_query_ras_poison_mode,
+	.ecc_info_query_ras_error_count = umc_v6_7_ecc_info_query_ras_error_count,
+	.ecc_info_query_ras_error_address = umc_v6_7_ecc_info_query_ras_error_address,
 };