Merge branch 'drm-next-4.11' of git://people.freedesktop.org/~agd5f/linux into drm-next

Some additional fixes for 4.11.  Delayed a bit due to Chinese New Year. Highlights:
- Powerplay fixes
- VCE and UVD powergating fixes
- Clean up amdgpu SI gfx code to match CI and VI
- Misc bug fixes

* 'drm-next-4.11' of git://people.freedesktop.org/~agd5f/linux: (30 commits)
  drm/amdgpu: report the number of bytes moved at buffer creation
  drm/amdgpu: fix a potential deadlock in amdgpu_bo_create_restricted()
  drm/amdgpu: add support for new smc firmware on polaris
  drm/amd/powerplay: refine code to avoid potential bug that the memory not cleared.
  drm/amdgpu: shut up #warning for compile testing
  drm/amdgpu/virt: fix double kfree on bo_va
  drm/radeon: remove some dead code
  drm/radeon: avoid kernel segfault in vce when gpu fails to resume
  drm/amd/powerplay: set fan speed to max in profile peak mode only.
  drm/amd/gfx6: update gb_addr_config
  drm/amdgpu: update HAINAN_GB_ADDR_CONFIG_GOLDEN
  drm/amdgpu: update VERDE_GB_ADDR_CONFIG_GOLDEN
  drm/amdgpu: refine si_read_register
  drm/amdgpu/gfx6: clean up spi configuration
  drm/amdgpu/gfx6: clean up cu configuration
  drm/amdgpu/gfx6: clean up rb configuration
  drm/amdgpu: refine vce3.0 code and related powerplay pg code.
  drm/amdgpu: move subfunctions to the front of vce_v2_0.c.
  drm/amdgpu: enable vce pg feature on Kv.
  drm/amdgpu: refine code for VCE2.0 and related dpm code.
  ...

27 files changed, 682 insertions, 633 deletions

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu.h b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
index 94a64e3..96ad796 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
@@ -1709,6 +1709,7 @@ int amdgpu_cs_parser_init(struct amdgpu_cs_parser *p, void *data);
 int amdgpu_cs_get_ring(struct amdgpu_device *adev, u32 ip_type,
 		       u32 ip_instance, u32 ring,
 		       struct amdgpu_ring **out_ring);
+void amdgpu_cs_report_moved_bytes(struct amdgpu_device *adev, u64 num_bytes);
 void amdgpu_ttm_placement_from_domain(struct amdgpu_bo *abo, u32 domain);
 bool amdgpu_ttm_bo_is_amdgpu_bo(struct ttm_buffer_object *bo);
 int amdgpu_ttm_tt_get_user_pages(struct ttm_tt *ttm, struct page **pages);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_cgs.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_cgs.c
index a5df1ef..fa34dca 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_cgs.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_cgs.c
@@ -850,16 +850,37 @@ static int amdgpu_cgs_get_firmware_info(struct cgs_device *cgs_device,
 				strcpy(fw_name, "amdgpu/fiji_smc.bin");
 				break;
 			case CHIP_POLARIS11:
-				if (type == CGS_UCODE_ID_SMU)
-					strcpy(fw_name, "amdgpu/polaris11_smc.bin");
-				else if (type == CGS_UCODE_ID_SMU_SK)
+				if (type == CGS_UCODE_ID_SMU) {
+					if (((adev->pdev->device == 0x67ef) &&
+					     ((adev->pdev->revision == 0xe0) ||
+					      (adev->pdev->revision == 0xe2) ||
+					      (adev->pdev->revision == 0xe5))) ||
+					    ((adev->pdev->device == 0x67ff) &&
+					     ((adev->pdev->revision == 0xcf) ||
+					      (adev->pdev->revision == 0xef) ||
+					      (adev->pdev->revision == 0xff))))
+						strcpy(fw_name, "amdgpu/polaris11_k_smc.bin");
+					else
+						strcpy(fw_name, "amdgpu/polaris11_smc.bin");
+				} else if (type == CGS_UCODE_ID_SMU_SK) {
 					strcpy(fw_name, "amdgpu/polaris11_smc_sk.bin");
+				}
 				break;
 			case CHIP_POLARIS10:
-				if (type == CGS_UCODE_ID_SMU)
-					strcpy(fw_name, "amdgpu/polaris10_smc.bin");
-				else if (type == CGS_UCODE_ID_SMU_SK)
+				if (type == CGS_UCODE_ID_SMU) {
+					if ((adev->pdev->device == 0x67df) &&
+					    ((adev->pdev->revision == 0xe0) ||
+					     (adev->pdev->revision == 0xe3) ||
+					     (adev->pdev->revision == 0xe4) ||
+					     (adev->pdev->revision == 0xe5) ||
+					     (adev->pdev->revision == 0xe7) ||
+					     (adev->pdev->revision == 0xef)))
+						strcpy(fw_name, "amdgpu/polaris10_k_smc.bin");
+					else
+						strcpy(fw_name, "amdgpu/polaris10_smc.bin");
+				} else if (type == CGS_UCODE_ID_SMU_SK) {
 					strcpy(fw_name, "amdgpu/polaris10_smc_sk.bin");
+				}
 				break;
 			case CHIP_POLARIS12:
 				strcpy(fw_name, "amdgpu/polaris12_smc.bin");
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
index cf2e8c4..57301f5 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
@@ -344,8 +344,7 @@ static u64 amdgpu_cs_get_threshold_for_moves(struct amdgpu_device *adev)
  * submission. This can result in a debt that can stop buffer migrations
  * temporarily.
  */
-static void amdgpu_cs_report_moved_bytes(struct amdgpu_device *adev,
-					 u64 num_bytes)
+void amdgpu_cs_report_moved_bytes(struct amdgpu_device *adev, u64 num_bytes)
 {
 	spin_lock(&adev->mm_stats.lock);
 	adev->mm_stats.accum_us -= bytes_to_us(adev, num_bytes);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c
index 9bd1b4e..ec7037a 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c
@@ -487,67 +487,50 @@ static int amdgpu_gem_va_check(void *param, struct amdgpu_bo *bo)
  *
  * @adev: amdgpu_device pointer
  * @bo_va: bo_va to update
+ * @list: validation list
+ * @operation: map or unmap
  *
- * Update the bo_va directly after setting it's address. Errors are not
+ * Update the bo_va directly after setting its address. Errors are not
  * vital here, so they are not reported back to userspace.
  */
 static void amdgpu_gem_va_update_vm(struct amdgpu_device *adev,
 				    struct amdgpu_bo_va *bo_va,
+				    struct list_head *list,
 				    uint32_t operation)
 {
-	struct ttm_validate_buffer tv, *entry;
-	struct amdgpu_bo_list_entry vm_pd;
-	struct ww_acquire_ctx ticket;
-	struct list_head list, duplicates;
-	int r;
-
-	INIT_LIST_HEAD(&list);
-	INIT_LIST_HEAD(&duplicates);
-
-	tv.bo = &bo_va->bo->tbo;
-	tv.shared = true;
-	list_add(&tv.head, &list);
+	struct ttm_validate_buffer *entry;
+	int r = -ERESTARTSYS;
 
-	amdgpu_vm_get_pd_bo(bo_va->vm, &list, &vm_pd);
-
-	/* Provide duplicates to avoid -EALREADY */
-	r = ttm_eu_reserve_buffers(&ticket, &list, true, &duplicates);
-	if (r)
-		goto error_print;
-
-	list_for_each_entry(entry, &list, head) {
+	list_for_each_entry(entry, list, head) {
 		struct amdgpu_bo *bo =
 			container_of(entry->bo, struct amdgpu_bo, tbo);
 
 		/* if anything is swapped out don't swap it in here,
 		   just abort and wait for the next CS */
 		if (!amdgpu_bo_gpu_accessible(bo))
-			goto error_unreserve;
+			goto error;
 
 		if (bo->shadow && !amdgpu_bo_gpu_accessible(bo->shadow))
-			goto error_unreserve;
+			goto error;
 	}
 
 	r = amdgpu_vm_validate_pt_bos(adev, bo_va->vm, amdgpu_gem_va_check,
 				      NULL);
 	if (r)
-		goto error_unreserve;
+		goto error;
 
 	r = amdgpu_vm_update_page_directory(adev, bo_va->vm);
 	if (r)
-		goto error_unreserve;
+		goto error;
 
 	r = amdgpu_vm_clear_freed(adev, bo_va->vm);
 	if (r)
-		goto error_unreserve;
+		goto error;
 
 	if (operation == AMDGPU_VA_OP_MAP)
 		r = amdgpu_vm_bo_update(adev, bo_va, false);
 
-error_unreserve:
-	ttm_eu_backoff_reservation(&ticket, &list);
-
-error_print:
+error:
 	if (r && r != -ERESTARTSYS)
 		DRM_ERROR("Couldn't update BO_VA (%d)\n", r);
 }
@@ -564,7 +547,7 @@ int amdgpu_gem_va_ioctl(struct drm_device *dev, void *data,
 	struct amdgpu_bo_list_entry vm_pd;
 	struct ttm_validate_buffer tv;
 	struct ww_acquire_ctx ticket;
-	struct list_head list, duplicates;
+	struct list_head list;
 	uint32_t invalid_flags, va_flags = 0;
 	int r = 0;
 
@@ -602,14 +585,13 @@ int amdgpu_gem_va_ioctl(struct drm_device *dev, void *data,
 		return -ENOENT;
 	abo = gem_to_amdgpu_bo(gobj);
 	INIT_LIST_HEAD(&list);
-	INIT_LIST_HEAD(&duplicates);
 	tv.bo = &abo->tbo;
-	tv.shared = true;
+	tv.shared = false;
 	list_add(&tv.head, &list);
 
 	amdgpu_vm_get_pd_bo(&fpriv->vm, &list, &vm_pd);
 
-	r = ttm_eu_reserve_buffers(&ticket, &list, true, &duplicates);
+	r = ttm_eu_reserve_buffers(&ticket, &list, true, NULL);
 	if (r) {
 		drm_gem_object_unreference_unlocked(gobj);
 		return r;
@@ -640,10 +622,10 @@ int amdgpu_gem_va_ioctl(struct drm_device *dev, void *data,
 	default:
 		break;
 	}
-	ttm_eu_backoff_reservation(&ticket, &list);
 	if (!r && !(args->flags & AMDGPU_VM_DELAY_UPDATE) &&
 	    !amdgpu_vm_debug)
-		amdgpu_gem_va_update_vm(adev, bo_va, args->operation);
+		amdgpu_gem_va_update_vm(adev, bo_va, &list, args->operation);
+	ttm_eu_backoff_reservation(&ticket, &list);
 
 	drm_gem_object_unreference_unlocked(gobj);
 	return r;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c
index d1aa291..be80a4a 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c
@@ -323,6 +323,7 @@ int amdgpu_bo_create_restricted(struct amdgpu_device *adev,
 	struct amdgpu_bo *bo;
 	enum ttm_bo_type type;
 	unsigned long page_align;
+	u64 initial_bytes_moved;
 	size_t acc_size;
 	int r;
 
@@ -374,8 +375,10 @@ int amdgpu_bo_create_restricted(struct amdgpu_device *adev,
 	 * See https://bugs.freedesktop.org/show_bug.cgi?id=88758
 	 */
 
+#ifndef CONFIG_COMPILE_TEST
 #warning Please enable CONFIG_MTRR and CONFIG_X86_PAT for better performance \
 	 thanks to write-combining
+#endif
 
 	if (bo->flags & AMDGPU_GEM_CREATE_CPU_GTT_USWC)
 		DRM_INFO_ONCE("Please enable CONFIG_MTRR and CONFIG_X86_PAT for "
@@ -399,12 +402,20 @@ int amdgpu_bo_create_restricted(struct amdgpu_device *adev,
 		locked = ww_mutex_trylock(&bo->tbo.ttm_resv.lock);
 		WARN_ON(!locked);
 	}
+
+	initial_bytes_moved = atomic64_read(&adev->num_bytes_moved);
 	r = ttm_bo_init(&adev->mman.bdev, &bo->tbo, size, type,
 			&bo->placement, page_align, !kernel, NULL,
 			acc_size, sg, resv ? resv : &bo->tbo.ttm_resv,
 			&amdgpu_ttm_bo_destroy);
-	if (unlikely(r != 0))
+	amdgpu_cs_report_moved_bytes(adev,
+		atomic64_read(&adev->num_bytes_moved) - initial_bytes_moved);
+
+	if (unlikely(r != 0)) {
+		if (!resv)
+			ww_mutex_unlock(&bo->tbo.resv->lock);
 		return r;
+	}
 
 	bo->tbo.priority = ilog2(bo->tbo.num_pages);
 	if (kernel)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_pm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_pm.c
index a61882d..95e026a 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_pm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_pm.c
@@ -1142,12 +1142,22 @@ void amdgpu_dpm_enable_vce(struct amdgpu_device *adev, bool enable)
 			/* XXX select vce level based on ring/task */
 			adev->pm.dpm.vce_level = AMD_VCE_LEVEL_AC_ALL;
 			mutex_unlock(&adev->pm.mutex);
+			amdgpu_pm_compute_clocks(adev);
+			amdgpu_set_powergating_state(adev, AMD_IP_BLOCK_TYPE_VCE,
+							AMD_PG_STATE_UNGATE);
+			amdgpu_set_clockgating_state(adev, AMD_IP_BLOCK_TYPE_VCE,
+							AMD_CG_STATE_UNGATE);
 		} else {
+			amdgpu_set_powergating_state(adev, AMD_IP_BLOCK_TYPE_VCE,
+							AMD_PG_STATE_GATE);
+			amdgpu_set_clockgating_state(adev, AMD_IP_BLOCK_TYPE_VCE,
+							AMD_CG_STATE_GATE);
 			mutex_lock(&adev->pm.mutex);
 			adev->pm.dpm.vce_active = false;
 			mutex_unlock(&adev->pm.mutex);
+			amdgpu_pm_compute_clocks(adev);
 		}
-		amdgpu_pm_compute_clocks(adev);
+
 	}
 }
 
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_uvd.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_uvd.c
index 6f62ac4..6d6ab7f 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_uvd.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_uvd.c
@@ -1113,6 +1113,11 @@ static void amdgpu_uvd_idle_work_handler(struct work_struct *work)
 			amdgpu_dpm_enable_uvd(adev, false);
 		} else {
 			amdgpu_asic_set_uvd_clocks(adev, 0, 0);
+			/* shutdown the UVD block */
+			amdgpu_set_powergating_state(adev, AMD_IP_BLOCK_TYPE_UVD,
+							    AMD_PG_STATE_GATE);
+			amdgpu_set_clockgating_state(adev, AMD_IP_BLOCK_TYPE_UVD,
+							    AMD_CG_STATE_GATE);
 		}
 	} else {
 		schedule_delayed_work(&adev->uvd.idle_work, UVD_IDLE_TIMEOUT);
@@ -1129,6 +1134,10 @@ void amdgpu_uvd_ring_begin_use(struct amdgpu_ring *ring)
 			amdgpu_dpm_enable_uvd(adev, true);
 		} else {
 			amdgpu_asic_set_uvd_clocks(adev, 53300, 40000);
+			amdgpu_set_clockgating_state(adev, AMD_IP_BLOCK_TYPE_UVD,
+							    AMD_CG_STATE_UNGATE);
+			amdgpu_set_powergating_state(adev, AMD_IP_BLOCK_TYPE_UVD,
+							    AMD_PG_STATE_UNGATE);
 		}
 	}
 }
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vce.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vce.c
index 79bc9c7..e2c0678 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vce.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vce.c
@@ -321,6 +321,10 @@ static void amdgpu_vce_idle_work_handler(struct work_struct *work)
 			amdgpu_dpm_enable_vce(adev, false);
 		} else {
 			amdgpu_asic_set_vce_clocks(adev, 0, 0);
+			amdgpu_set_powergating_state(adev, AMD_IP_BLOCK_TYPE_VCE,
+							    AMD_PG_STATE_GATE);
+			amdgpu_set_clockgating_state(adev, AMD_IP_BLOCK_TYPE_VCE,
+							    AMD_CG_STATE_GATE);
 		}
 	} else {
 		schedule_delayed_work(&adev->vce.idle_work, VCE_IDLE_TIMEOUT);
@@ -346,6 +350,11 @@ void amdgpu_vce_ring_begin_use(struct amdgpu_ring *ring)
 			amdgpu_dpm_enable_vce(adev, true);
 		} else {
 			amdgpu_asic_set_vce_clocks(adev, 53300, 40000);
+			amdgpu_set_clockgating_state(adev, AMD_IP_BLOCK_TYPE_VCE,
+							    AMD_CG_STATE_UNGATE);
+			amdgpu_set_powergating_state(adev, AMD_IP_BLOCK_TYPE_VCE,
+							    AMD_PG_STATE_UNGATE);
+
 		}
 	}
 	mutex_unlock(&adev->vce.idle_mutex);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c
index 3fd951c..dcfb7df 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c
@@ -83,7 +83,6 @@ int amdgpu_map_static_csa(struct amdgpu_device *adev, struct amdgpu_vm *vm)
 		DRM_ERROR("failed to do bo_map on static CSA, err=%d\n", r);
 		amdgpu_vm_bo_rmv(adev, bo_va);
 		ttm_eu_backoff_reservation(&ticket, &list);
-		kfree(bo_va);
 		return r;
 	}
 
diff --git a/drivers/gpu/drm/amd/amdgpu/ci_dpm.c b/drivers/gpu/drm/amd/amdgpu/ci_dpm.c
index 9498e78..f97ecb4 100644
--- a/drivers/gpu/drm/amd/amdgpu/ci_dpm.c
+++ b/drivers/gpu/drm/amd/amdgpu/ci_dpm.c
@@ -2210,7 +2210,6 @@ static void ci_clear_vc(struct amdgpu_device *adev)
 
 static int ci_upload_firmware(struct amdgpu_device *adev)
 {
-	struct ci_power_info *pi = ci_get_pi(adev);
 	int i, ret;
 
 	if (amdgpu_ci_is_smc_running(adev)) {
@@ -2227,7 +2226,7 @@ static int ci_upload_firmware(struct amdgpu_device *adev)
 	amdgpu_ci_stop_smc_clock(adev);
 	amdgpu_ci_reset_smc(adev);
 
-	ret = amdgpu_ci_load_smc_ucode(adev, pi->sram_end);
+	ret = amdgpu_ci_load_smc_ucode(adev, SMC_RAM_END);
 
 	return ret;
 
@@ -4257,12 +4256,6 @@ static int ci_update_vce_dpm(struct amdgpu_device *adev,
 
 	if (amdgpu_current_state->evclk != amdgpu_new_state->evclk) {
 		if (amdgpu_new_state->evclk) {
-			/* turn the clocks on when encoding */
-			ret = amdgpu_set_clockgating_state(adev, AMD_IP_BLOCK_TYPE_VCE,
-							    AMD_CG_STATE_UNGATE);
-			if (ret)
-				return ret;
-
 			pi->smc_state_table.VceBootLevel = ci_get_vce_boot_level(adev);
 			tmp = RREG32_SMC(ixDPM_TABLE_475);
 			tmp &= ~DPM_TABLE_475__VceBootLevel_MASK;
@@ -4274,9 +4267,6 @@ static int ci_update_vce_dpm(struct amdgpu_device *adev,
 			ret = ci_enable_vce_dpm(adev, false);
 			if (ret)
 				return ret;
-			/* turn the clocks off when not encoding */
-			ret = amdgpu_set_clockgating_state(adev, AMD_IP_BLOCK_TYPE_VCE,
-							    AMD_CG_STATE_GATE);
 		}
 	}
 	return ret;
@@ -6278,13 +6268,13 @@ static int ci_dpm_sw_init(void *handle)
 	adev->pm.current_mclk = adev->clock.default_mclk;
 	adev->pm.int_thermal_type = THERMAL_TYPE_NONE;
 
-	if (amdgpu_dpm == 0)
-		return 0;
-
 	ret = ci_dpm_init_microcode(adev);
 	if (ret)
 		return ret;
 
+	if (amdgpu_dpm == 0)
+		return 0;
+
 	INIT_WORK(&adev->pm.dpm.thermal.work, amdgpu_dpm_thermal_work_handler);
 	mutex_lock(&adev->pm.mutex);
 	ret = ci_dpm_init(adev);
@@ -6328,8 +6318,15 @@ static int ci_dpm_hw_init(void *handle)
 
 	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
 
-	if (!amdgpu_dpm)
+	if (!amdgpu_dpm) {
+		ret = ci_upload_firmware(adev);
+		if (ret) {
+			DRM_ERROR("ci_upload_firmware failed\n");
+			return ret;
+		}
+		ci_dpm_start_smc(adev);
 		return 0;
+	}
 
 	mutex_lock(&adev->pm.mutex);
 	ci_dpm_setup_asic(adev);
@@ -6351,6 +6348,8 @@ static int ci_dpm_hw_fini(void *handle)
 		mutex_lock(&adev->pm.mutex);
 		ci_dpm_disable(adev);
 		mutex_unlock(&adev->pm.mutex);
+	} else {
+		ci_dpm_stop_smc(adev);
 	}
 
 	return 0;
diff --git a/drivers/gpu/drm/amd/amdgpu/cik.c b/drivers/gpu/drm/amd/amdgpu/cik.c
index 7da688b..7c39b538 100644
--- a/drivers/gpu/drm/amd/amdgpu/cik.c
+++ b/drivers/gpu/drm/amd/amdgpu/cik.c
@@ -1722,8 +1722,8 @@ static int cik_common_early_init(void *handle)
 			  AMD_PG_SUPPORT_GFX_SMG |
 			  AMD_PG_SUPPORT_GFX_DMG |*/
 			AMD_PG_SUPPORT_UVD |
-			/*AMD_PG_SUPPORT_VCE |
-			  AMD_PG_SUPPORT_CP |
+			AMD_PG_SUPPORT_VCE |
+			/*  AMD_PG_SUPPORT_CP |
 			  AMD_PG_SUPPORT_GDS |
 			  AMD_PG_SUPPORT_RLC_SMU_HS |
 			  AMD_PG_SUPPORT_ACP |
diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v6_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v6_0.c
index c998f6a..2086e7e 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v6_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v6_0.c
@@ -1325,21 +1325,19 @@ static u32 gfx_v6_0_create_bitmask(u32 bit_width)
 	return (u32)(((u64)1 << bit_width) - 1);
 }
 
-static u32 gfx_v6_0_get_rb_disabled(struct amdgpu_device *adev,
-				    u32 max_rb_num_per_se,
-				    u32 sh_per_se)
+static u32 gfx_v6_0_get_rb_active_bitmap(struct amdgpu_device *adev)
 {
 	u32 data, mask;
 
-	data = RREG32(mmCC_RB_BACKEND_DISABLE);
-	data &= CC_RB_BACKEND_DISABLE__BACKEND_DISABLE_MASK;
-	data |= RREG32(mmGC_USER_RB_BACKEND_DISABLE);
+	data = RREG32(mmCC_RB_BACKEND_DISABLE) |
+		RREG32(mmGC_USER_RB_BACKEND_DISABLE);
 
-	data >>= CC_RB_BACKEND_DISABLE__BACKEND_DISABLE__SHIFT;
+	data = REG_GET_FIELD(data, GC_USER_RB_BACKEND_DISABLE, BACKEND_DISABLE);
 
-	mask = gfx_v6_0_create_bitmask(max_rb_num_per_se / sh_per_se);
+	mask = gfx_v6_0_create_bitmask(adev->gfx.config.max_backends_per_se/
+					adev->gfx.config.max_sh_per_se);
 
-	return data & mask;
+	return ~data & mask;
 }
 
 static void gfx_v6_0_raster_config(struct amdgpu_device *adev, u32 *rconf)
@@ -1468,68 +1466,55 @@ static void gfx_v6_0_write_harvested_raster_configs(struct amdgpu_device *adev,
 	gfx_v6_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff);
 }
 
-static void gfx_v6_0_setup_rb(struct amdgpu_device *adev,
-			      u32 se_num, u32 sh_per_se,
-			      u32 max_rb_num_per_se)
+static void gfx_v6_0_setup_rb(struct amdgpu_device *adev)
 {
 	int i, j;
-	u32 data, mask;
-	u32 disabled_rbs = 0;
-	u32 enabled_rbs = 0;
+	u32 data;
+	u32 raster_config = 0;
+	u32 active_rbs = 0;
+	u32 rb_bitmap_width_per_sh = adev->gfx.config.max_backends_per_se /
+					adev->gfx.config.max_sh_per_se;
 	unsigned num_rb_pipes;
 
 	mutex_lock(&adev->grbm_idx_mutex);
-	for (i = 0; i < se_num; i++) {
-		for (j = 0; j < sh_per_se; j++) {
+	for (i = 0; i < adev->gfx.config.max_shader_engines; i++) {
+		for (j = 0; j < adev->gfx.config.max_sh_per_se; j++) {
 			gfx_v6_0_select_se_sh(adev, i, j, 0xffffffff);
-			data = gfx_v6_0_get_rb_disabled(adev, max_rb_num_per_se, sh_per_se);
-			disabled_rbs |= data << ((i * sh_per_se + j) * 2);
+			data = gfx_v6_0_get_rb_active_bitmap(adev);
+			active_rbs |= data << ((i * adev->gfx.config.max_sh_per_se + j) *
+					rb_bitmap_width_per_sh);
 		}
 	}
 	gfx_v6_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff);
-	mutex_unlock(&adev->grbm_idx_mutex);
-
-	mask = 1;
-	for (i = 0; i < max_rb_num_per_se * se_num; i++) {
-		if (!(disabled_rbs & mask))
-			enabled_rbs |= mask;
-		mask <<= 1;
-	}
 
-	adev->gfx.config.backend_enable_mask = enabled_rbs;
-	adev->gfx.config.num_rbs = hweight32(enabled_rbs);
+	adev->gfx.config.backend_enable_mask = active_rbs;
+	adev->gfx.config.num_rbs = hweight32(active_rbs);
 
 	num_rb_pipes = min_t(unsigned, adev->gfx.config.max_backends_per_se *
 			     adev->gfx.config.max_shader_engines, 16);
 
-	mutex_lock(&adev->grbm_idx_mutex);
-	for (i = 0; i < se_num; i++) {
-		gfx_v6_0_select_se_sh(adev, i, 0xffffffff, 0xffffffff);
-		data = 0;
-		for (j = 0; j < sh_per_se; j++) {
-			switch (enabled_rbs & 3) {
-			case 1:
-				data |= (RASTER_CONFIG_RB_MAP_0 << (i * sh_per_se + j) * 2);
-				break;
-			case 2:
-				data |= (RASTER_CONFIG_RB_MAP_3 << (i * sh_per_se + j) * 2);
-				break;
-			case 3:
-			default:
-				data |= (RASTER_CONFIG_RB_MAP_2 << (i * sh_per_se + j) * 2);
-				break;
-			}
-			enabled_rbs >>= 2;
-		}
-		gfx_v6_0_raster_config(adev, &data);
+	gfx_v6_0_raster_config(adev, &raster_config);
 
-		if (!adev->gfx.config.backend_enable_mask ||
-				adev->gfx.config.num_rbs >= num_rb_pipes)
-			WREG32(mmPA_SC_RASTER_CONFIG, data);
-		else
-			gfx_v6_0_write_harvested_raster_configs(adev, data,
-								adev->gfx.config.backend_enable_mask,
-								num_rb_pipes);
+	if (!adev->gfx.config.backend_enable_mask ||
+			adev->gfx.config.num_rbs >= num_rb_pipes) {
+		WREG32(mmPA_SC_RASTER_CONFIG, raster_config);
+	} else {
+		gfx_v6_0_write_harvested_raster_configs(adev, raster_config,
+							adev->gfx.config.backend_enable_mask,
+							num_rb_pipes);
+	}
+
+	/* cache the values for userspace */
+	for (i = 0; i < adev->gfx.config.max_shader_engines; i++) {
+		for (j = 0; j < adev->gfx.config.max_sh_per_se; j++) {
+			gfx_v6_0_select_se_sh(adev, i, j, 0xffffffff);
+			adev->gfx.config.rb_config[i][j].rb_backend_disable =
+				RREG32(mmCC_RB_BACKEND_DISABLE);
+			adev->gfx.config.rb_config[i][j].user_rb_backend_disable =
+				RREG32(mmGC_USER_RB_BACKEND_DISABLE);
+			adev->gfx.config.rb_config[i][j].raster_config =
+				RREG32(mmPA_SC_RASTER_CONFIG);
+		}
 	}
 	gfx_v6_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff);
 	mutex_unlock(&adev->grbm_idx_mutex);
@@ -1540,36 +1525,44 @@ static void gmc_v6_0_init_compute_vmid(struct amdgpu_device *adev)
 }
 */
 
-static u32 gfx_v6_0_get_cu_enabled(struct amdgpu_device *adev, u32 cu_per_sh)
+static void gfx_v6_0_set_user_cu_inactive_bitmap(struct amdgpu_device *adev,
+						 u32 bitmap)
 {
-	u32 data, mask;
+	u32 data;
 
-	data = RREG32(mmCC_GC_SHADER_ARRAY_CONFIG);
-	data &= CC_GC_SHADER_ARRAY_CONFIG__INACTIVE_CUS_MASK;
-	data |= RREG32(mmGC_USER_SHADER_ARRAY_CONFIG);
+	if (!bitmap)
+		return;
 
-	data >>= CC_GC_SHADER_ARRAY_CONFIG__INACTIVE_CUS__SHIFT;
+	data = bitmap << GC_USER_SHADER_ARRAY_CONFIG__INACTIVE_CUS__SHIFT;
+	data &= GC_USER_SHADER_ARRAY_CONFIG__INACTIVE_CUS_MASK;
 
-	mask = gfx_v6_0_create_bitmask(cu_per_sh);
+	WREG32(mmGC_USER_SHADER_ARRAY_CONFIG, data);
+}
 
-	return ~data & mask;
+static u32 gfx_v6_0_get_cu_enabled(struct amdgpu_device *adev)
+{
+	u32 data, mask;
+
+	data = RREG32(mmCC_GC_SHADER_ARRAY_CONFIG) |
+		RREG32(mmGC_USER_SHADER_ARRAY_CONFIG);
+
+	mask = gfx_v6_0_create_bitmask(adev->gfx.config.max_cu_per_sh);
+	return ~REG_GET_FIELD(data, CC_GC_SHADER_ARRAY_CONFIG, INACTIVE_CUS) & mask;
 }
 
 
-static void gfx_v6_0_setup_spi(struct amdgpu_device *adev,
-			 u32 se_num, u32 sh_per_se,
-			 u32 cu_per_sh)
+static void gfx_v6_0_setup_spi(struct amdgpu_device *adev)
 {
 	int i, j, k;
 	u32 data, mask;
 	u32 active_cu = 0;
 
 	mutex_lock(&adev->grbm_idx_mutex);
-	for (i = 0; i < se_num; i++) {
-		for (j = 0; j < sh_per_se; j++) {
+	for (i = 0; i < adev->gfx.config.max_shader_engines; i++) {
+		for (j = 0; j < adev->gfx.config.max_sh_per_se; j++) {
 			gfx_v6_0_select_se_sh(adev, i, j, 0xffffffff);
 			data = RREG32(mmSPI_STATIC_THREAD_MGMT_3);
-			active_cu = gfx_v6_0_get_cu_enabled(adev, cu_per_sh);
+			active_cu = gfx_v6_0_get_cu_enabled(adev);
 
 			mask = 1;
 			for (k = 0; k < 16; k++) {
@@ -1717,6 +1710,9 @@ static void gfx_v6_0_gpu_init(struct amdgpu_device *adev)
 		gb_addr_config |= 2 << GB_ADDR_CONFIG__ROW_SIZE__SHIFT;
 		break;
 	}
+	gb_addr_config &= ~GB_ADDR_CONFIG__NUM_SHADER_ENGINES_MASK;
+	if (adev->gfx.config.max_shader_engines == 2)
+		gb_addr_config |= 1 << GB_ADDR_CONFIG__NUM_SHADER_ENGINES__SHIFT;
 	adev->gfx.config.gb_addr_config = gb_addr_config;
 
 	WREG32(mmGB_ADDR_CONFIG, gb_addr_config);
@@ -1735,13 +1731,9 @@ static void gfx_v6_0_gpu_init(struct amdgpu_device *adev)
 #endif
 	gfx_v6_0_tiling_mode_table_init(adev);
 
-	gfx_v6_0_setup_rb(adev, adev->gfx.config.max_shader_engines,
-		    adev->gfx.config.max_sh_per_se,
-		    adev->gfx.config.max_backends_per_se);
+	gfx_v6_0_setup_rb(adev);
 
-	gfx_v6_0_setup_spi(adev, adev->gfx.config.max_shader_engines,
-		     adev->gfx.config.max_sh_per_se,
-		     adev->gfx.config.max_cu_per_sh);
+	gfx_v6_0_setup_spi(adev);
 
 	gfx_v6_0_get_cu_info(adev);
 
@@ -2941,61 +2933,16 @@ static void gfx_v6_0_enable_gfx_cgpg(struct amdgpu_device *adev,
 	}
 }
 
-static u32 gfx_v6_0_get_cu_active_bitmap(struct amdgpu_device *adev,
-					 u32 se, u32 sh)
-{
-
-	u32 mask = 0, tmp, tmp1;
-	int i;
-
-	mutex_lock(&adev->grbm_idx_mutex);
-	gfx_v6_0_select_se_sh(adev, se, sh, 0xffffffff);
-	tmp = RREG32(mmCC_GC_SHADER_ARRAY_CONFIG);
-	tmp1 = RREG32(mmGC_USER_SHADER_ARRAY_CONFIG);
-	gfx_v6_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff);
-	mutex_unlock(&adev->grbm_idx_mutex);
-
-	tmp &= 0xffff0000;
-
-	tmp |= tmp1;
-	tmp >>= 16;
-
-	for (i = 0; i < adev->gfx.config.max_cu_per_sh; i ++) {
-		mask <<= 1;
-		mask |= 1;
-	}
-
-	return (~tmp) & mask;
-}
-
 static void gfx_v6_0_init_ao_cu_mask(struct amdgpu_device *adev)
 {
-	u32 i, j, k, active_cu_number = 0;
-
-	u32 mask, counter, cu_bitmap;
-	u32 tmp = 0;
-
-	for (i = 0; i < adev->gfx.config.max_shader_engines; i++) {
-		for (j = 0; j < adev->gfx.config.max_sh_per_se; j++) {
-			mask = 1;
-			cu_bitmap = 0;
-			counter  = 0;
-			for (k = 0; k < adev->gfx.config.max_cu_per_sh; k++) {
-				if (gfx_v6_0_get_cu_active_bitmap(adev, i, j) & mask) {
-					if (counter < 2)
-						cu_bitmap |= mask;
-					counter++;
-				}
-				mask <<= 1;
-			}
+	u32 tmp;
 
-			active_cu_number += counter;
-			tmp |= (cu_bitmap << (i * 16 + j * 8));
-		}
-	}
+	WREG32(mmRLC_PG_ALWAYS_ON_CU_MASK, adev->gfx.cu_info.ao_cu_mask);
 
-	WREG32(mmRLC_PG_AO_CU_MASK, tmp);
-	WREG32_FIELD(RLC_MAX_PG_CU, MAX_POWERED_UP_CU, active_cu_number);
+	tmp = RREG32(mmRLC_MAX_PG_CU);
+	tmp &= ~RLC_MAX_PG_CU__MAX_POWERED_UP_CU_MASK;
+	tmp |= (adev->gfx.cu_info.number << RLC_MAX_PG_CU__MAX_POWERED_UP_CU__SHIFT);
+	WREG32(mmRLC_MAX_PG_CU, tmp);
 }
 
 static void gfx_v6_0_enable_gfx_static_mgpg(struct amdgpu_device *adev,
@@ -3770,18 +3717,26 @@ static void gfx_v6_0_get_cu_info(struct amdgpu_device *adev)
 	int i, j, k, counter, active_cu_number = 0;
 	u32 mask, bitmap, ao_bitmap, ao_cu_mask = 0;
 	struct amdgpu_cu_info *cu_info = &adev->gfx.cu_info;
+	unsigned disable_masks[4 * 2];
 
 	memset(cu_info, 0, sizeof(*cu_info));
 
+	amdgpu_gfx_parse_disable_cu(disable_masks, 4, 2);
+
+	mutex_lock(&adev->grbm_idx_mutex);
 	for (i = 0; i < adev->gfx.config.max_shader_engines; i++) {
 		for (j = 0; j < adev->gfx.config.max_sh_per_se; j++) {
 			mask = 1;
 			ao_bitmap = 0;
 			counter = 0;
-			bitmap = gfx_v6_0_get_cu_active_bitmap(adev, i, j);
+			gfx_v6_0_select_se_sh(adev, i, j, 0xffffffff);
+			if (i < 4 && j < 2)
+				gfx_v6_0_set_user_cu_inactive_bitmap(
+					adev, disable_masks[i * 2 + j]);
+			bitmap = gfx_v6_0_get_cu_enabled(adev);
 			cu_info->bitmap[i][j] = bitmap;
 
-			for (k = 0; k < adev->gfx.config.max_cu_per_sh; k ++) {
+			for (k = 0; k < 16; k++) {
 				if (bitmap & mask) {
 					if (counter < 2)
 						ao_bitmap |= mask;
@@ -3794,6 +3749,9 @@ static void gfx_v6_0_get_cu_info(struct amdgpu_device *adev)
 		}
 	}
 
+	gfx_v6_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff);
+	mutex_unlock(&adev->grbm_idx_mutex);
+
 	cu_info->number = active_cu_number;
 	cu_info->ao_cu_mask = ao_cu_mask;
 }
diff --git a/drivers/gpu/drm/amd/amdgpu/kv_dpm.c b/drivers/gpu/drm/amd/amdgpu/kv_dpm.c
index 8785ca5..f5a343c 100644
--- a/drivers/gpu/drm/amd/amdgpu/kv_dpm.c
+++ b/drivers/gpu/drm/amd/amdgpu/kv_dpm.c
@@ -1550,11 +1550,6 @@ static int kv_update_vce_dpm(struct amdgpu_device *adev,
 
 	if (amdgpu_new_state->evclk > 0 && amdgpu_current_state->evclk == 0) {
 		kv_dpm_powergate_vce(adev, false);
-		/* turn the clocks on when encoding */
-		ret = amdgpu_set_clockgating_state(adev, AMD_IP_BLOCK_TYPE_VCE,
-						    AMD_CG_STATE_UNGATE);
-		if (ret)
-			return ret;
 		if (pi->caps_stable_p_state)
 			pi->vce_boot_level = table->count - 1;
 		else
@@ -1573,15 +1568,9 @@ static int kv_update_vce_dpm(struct amdgpu_device *adev,
 			amdgpu_kv_send_msg_to_smc_with_parameter(adev,
 							  PPSMC_MSG_VCEDPM_SetEnabledMask,
 							  (1 << pi->vce_boot_level));
-
 		kv_enable_vce_dpm(adev, true);
 	} else if (amdgpu_new_state->evclk == 0 && amdgpu_current_state->evclk > 0) {
 		kv_enable_vce_dpm(adev, false);
-		/* turn the clocks off when not encoding */
-		ret = amdgpu_set_clockgating_state(adev, AMD_IP_BLOCK_TYPE_VCE,
-						    AMD_CG_STATE_GATE);
-		if (ret)
-			return ret;
 		kv_dpm_powergate_vce(adev, true);
 	}
 
@@ -1688,70 +1677,44 @@ static void kv_dpm_powergate_uvd(struct amdgpu_device *adev, bool gate)
 	struct kv_power_info *pi = kv_get_pi(adev);
 	int ret;
 
-	if (pi->uvd_power_gated == gate)
-		return;
-
 	pi->uvd_power_gated = gate;
 
 	if (gate) {
-		if (pi->caps_uvd_pg) {
-			/* disable clockgating so we can properly shut down the block */
-			ret = amdgpu_set_clockgating_state(adev, AMD_IP_BLOCK_TYPE_UVD,
-							    AMD_CG_STATE_UNGATE);
-			/* shutdown the UVD block */
-			ret = amdgpu_set_powergating_state(adev, AMD_IP_BLOCK_TYPE_UVD,
-							    AMD_PG_STATE_GATE);
-			/* XXX: check for errors */
-		}
+		/* stop the UVD block */
+		ret = amdgpu_set_powergating_state(adev, AMD_IP_BLOCK_TYPE_UVD,
+							AMD_PG_STATE_GATE);
 		kv_update_uvd_dpm(adev, gate);
 		if (pi->caps_uvd_pg)
 			/* power off the UVD block */
 			amdgpu_kv_notify_message_to_smu(adev, PPSMC_MSG_UVDPowerOFF);
 	} else {
-		if (pi->caps_uvd_pg) {
+		if (pi->caps_uvd_pg)
 			/* power on the UVD block */
 			amdgpu_kv_notify_message_to_smu(adev, PPSMC_MSG_UVDPowerON);
 			/* re-init the UVD block */
-			ret = amdgpu_set_powergating_state(adev, AMD_IP_BLOCK_TYPE_UVD,
-							    AMD_PG_STATE_UNGATE);
-			/* enable clockgating. hw will dynamically gate/ungate clocks on the fly */
-			ret = amdgpu_set_clockgating_state(adev, AMD_IP_BLOCK_TYPE_UVD,
-							    AMD_CG_STATE_GATE);
-			/* XXX: check for errors */
-		}
 		kv_update_uvd_dpm(adev, gate);
+
+		ret = amdgpu_set_powergating_state(adev, AMD_IP_BLOCK_TYPE_UVD,
+							AMD_PG_STATE_UNGATE);
 	}
 }
 
 static void kv_dpm_powergate_vce(struct amdgpu_device *adev, bool gate)
 {
 	struct kv_power_info *pi = kv_get_pi(adev);
-	int ret;
 
 	if (pi->vce_power_gated == gate)
 		return;
 
 	pi->vce_power_gated = gate;
 
-	if (gate) {
-		if (pi->caps_vce_pg) {
-			/* shutdown the VCE block */
-			ret = amdgpu_set_powergating_state(adev, AMD_IP_BLOCK_TYPE_VCE,
-							    AMD_PG_STATE_GATE);
-			/* XXX: check for errors */
-			/* power off the VCE block */
-			amdgpu_kv_notify_message_to_smu(adev, PPSMC_MSG_VCEPowerOFF);
-		}
-	} else {
-		if (pi->caps_vce_pg) {
-			/* power on the VCE block */
-			amdgpu_kv_notify_message_to_smu(adev, PPSMC_MSG_VCEPowerON);
-			/* re-init the VCE block */
-			ret = amdgpu_set_powergating_state(adev, AMD_IP_BLOCK_TYPE_VCE,
-							    AMD_PG_STATE_UNGATE);
-			/* XXX: check for errors */
-		}
-	}
+	if (!pi->caps_vce_pg)
+		return;
+
+	if (gate)
+		amdgpu_kv_notify_message_to_smu(adev, PPSMC_MSG_VCEPowerOFF);
+	else
+		amdgpu_kv_notify_message_to_smu(adev, PPSMC_MSG_VCEPowerON);
 }
 
 static void kv_dpm_powergate_samu(struct amdgpu_device *adev, bool gate)
@@ -3009,8 +2972,7 @@ static int kv_dpm_late_init(void *handle)
 
 	kv_dpm_powergate_acp(adev, true);
 	kv_dpm_powergate_samu(adev, true);
-	kv_dpm_powergate_vce(adev, true);
-	kv_dpm_powergate_uvd(adev, true);
+
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/amd/amdgpu/si.c b/drivers/gpu/drm/amd/amdgpu/si.c
index da46992..b71e3fa 100644
--- a/drivers/gpu/drm/amd/amdgpu/si.c
+++ b/drivers/gpu/drm/amd/amdgpu/si.c
@@ -1010,24 +1010,81 @@ static struct amdgpu_allowed_register_entry si_allowed_read_registers[] = {
 	{PA_SC_RASTER_CONFIG, false, true},
 };
 
-static uint32_t si_read_indexed_register(struct amdgpu_device *adev,
-					  u32 se_num, u32 sh_num,
-					  u32 reg_offset)
+static uint32_t si_get_register_value(struct amdgpu_device *adev,
+				      bool indexed, u32 se_num,
+				      u32 sh_num, u32 reg_offset)
 {
-	uint32_t val;
+	if (indexed) {
+		uint32_t val;
+		unsigned se_idx = (se_num == 0xffffffff) ? 0 : se_num;
+		unsigned sh_idx = (sh_num == 0xffffffff) ? 0 : sh_num;
+
+		switch (reg_offset) {
+		case mmCC_RB_BACKEND_DISABLE:
+			return adev->gfx.config.rb_config[se_idx][sh_idx].rb_backend_disable;
+		case mmGC_USER_RB_BACKEND_DISABLE:
+			return adev->gfx.config.rb_config[se_idx][sh_idx].user_rb_backend_disable;
+		case mmPA_SC_RASTER_CONFIG:
+			return adev->gfx.config.rb_config[se_idx][sh_idx].raster_config;
+		}
 
-	mutex_lock(&adev->grbm_idx_mutex);
-	if (se_num != 0xffffffff || sh_num != 0xffffffff)
-		amdgpu_gfx_select_se_sh(adev, se_num, sh_num, 0xffffffff);
+		mutex_lock(&adev->grbm_idx_mutex);
+		if (se_num != 0xffffffff || sh_num != 0xffffffff)
+			amdgpu_gfx_select_se_sh(adev, se_num, sh_num, 0xffffffff);
 
-	val = RREG32(reg_offset);
+		val = RREG32(reg_offset);
 
-	if (se_num != 0xffffffff || sh_num != 0xffffffff)
-		amdgpu_gfx_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff);
-	mutex_unlock(&adev->grbm_idx_mutex);
-	return val;
+		if (se_num != 0xffffffff || sh_num != 0xffffffff)
+			amdgpu_gfx_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff);
+		mutex_unlock(&adev->grbm_idx_mutex);
+		return val;
+	} else {
+		unsigned idx;
+
+		switch (reg_offset) {
+		case mmGB_ADDR_CONFIG:
+			return adev->gfx.config.gb_addr_config;
+		case mmMC_ARB_RAMCFG:
+			return adev->gfx.config.mc_arb_ramcfg;
+		case mmGB_TILE_MODE0:
+		case mmGB_TILE_MODE1:
+		case mmGB_TILE_MODE2:
+		case mmGB_TILE_MODE3:
+		case mmGB_TILE_MODE4:
+		case mmGB_TILE_MODE5:
+		case mmGB_TILE_MODE6:
+		case mmGB_TILE_MODE7:
+		case mmGB_TILE_MODE8:
+		case mmGB_TILE_MODE9:
+		case mmGB_TILE_MODE10:
+		case mmGB_TILE_MODE11:
+		case mmGB_TILE_MODE12:
+		case mmGB_TILE_MODE13:
+		case mmGB_TILE_MODE14:
+		case mmGB_TILE_MODE15:
+		case mmGB_TILE_MODE16:
+		case mmGB_TILE_MODE17:
+		case mmGB_TILE_MODE18:
+		case mmGB_TILE_MODE19:
+		case mmGB_TILE_MODE20:
+		case mmGB_TILE_MODE21:
+		case mmGB_TILE_MODE22:
+		case mmGB_TILE_MODE23:
+		case mmGB_TILE_MODE24:
+		case mmGB_TILE_MODE25:
+		case mmGB_TILE_MODE26:
+		case mmGB_TILE_MODE27:
+		case mmGB_TILE_MODE28:
+		case mmGB_TILE_MODE29:
+		case mmGB_TILE_MODE30:
+		case mmGB_TILE_MODE31:
+			idx = (reg_offset - mmGB_TILE_MODE0);
+			return adev->gfx.config.tile_mode_array[idx];
+		default:
+			return RREG32(reg_offset);
+		}
+	}
 }
-
 static int si_read_register(struct amdgpu_device *adev, u32 se_num,
 			     u32 sh_num, u32 reg_offset, u32 *value)
 {
@@ -1039,10 +1096,9 @@ static int si_read_register(struct amdgpu_device *adev, u32 se_num,
 			continue;
 
 		if (!si_allowed_read_registers[i].untouched)
-			*value = si_allowed_read_registers[i].grbm_indexed ?
-				 si_read_indexed_register(adev, se_num,
-							   sh_num, reg_offset) :
-				 RREG32(reg_offset);
+			*value = si_get_register_value(adev,
+						si_allowed_read_registers[i].grbm_indexed,
+						se_num, sh_num, reg_offset);
 		return 0;
 	}
 	return -EINVAL;
diff --git a/drivers/gpu/drm/amd/amdgpu/si_enums.h b/drivers/gpu/drm/amd/amdgpu/si_enums.h
index fde2086..dc9e0e6 100644
--- a/drivers/gpu/drm/amd/amdgpu/si_enums.h
+++ b/drivers/gpu/drm/amd/amdgpu/si_enums.h
@@ -143,8 +143,8 @@
 #define RLC_CLEAR_STATE_DESCRIPTOR_OFFSET    0x3D
 
 #define TAHITI_GB_ADDR_CONFIG_GOLDEN        0x12011003
-#define VERDE_GB_ADDR_CONFIG_GOLDEN         0x12010002
-#define HAINAN_GB_ADDR_CONFIG_GOLDEN        0x02010001
+#define VERDE_GB_ADDR_CONFIG_GOLDEN         0x02010002
+#define HAINAN_GB_ADDR_CONFIG_GOLDEN        0x02011003
 
 #define PACKET3(op, n)  ((RADEON_PACKET_TYPE3 << 30) |                  \
                          (((op) & 0xFF) << 8) |                         \
diff --git a/drivers/gpu/drm/amd/amdgpu/uvd_v4_2.c b/drivers/gpu/drm/amd/amdgpu/uvd_v4_2.c
index 7fb9137..f15df99 100644
--- a/drivers/gpu/drm/amd/amdgpu/uvd_v4_2.c
+++ b/drivers/gpu/drm/amd/amdgpu/uvd_v4_2.c
@@ -159,9 +159,6 @@ static int uvd_v4_2_hw_init(void *handle)
 
 	uvd_v4_2_enable_mgcg(adev, true);
 	amdgpu_asic_set_uvd_clocks(adev, 10000, 10000);
-	r = uvd_v4_2_start(adev);
-	if (r)
-		goto done;
 
 	ring->ready = true;
 	r = amdgpu_ring_test_ring(ring);
@@ -198,7 +195,6 @@ static int uvd_v4_2_hw_init(void *handle)
 	amdgpu_ring_commit(ring);
 
 done:
-
 	if (!r)
 		DRM_INFO("UVD initialized successfully.\n");
 
@@ -217,7 +213,9 @@ static int uvd_v4_2_hw_fini(void *handle)
 	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
 	struct amdgpu_ring *ring = &adev->uvd.ring;
 
-	uvd_v4_2_stop(adev);
+	if (RREG32(mmUVD_STATUS) != 0)
+		uvd_v4_2_stop(adev);
+
 	ring->ready = false;
 
 	return 0;
@@ -267,37 +265,26 @@ static int uvd_v4_2_start(struct amdgpu_device *adev)
 	struct amdgpu_ring *ring = &adev->uvd.ring;
 	uint32_t rb_bufsz;
 	int i, j, r;
+	u32 tmp;
 	/* disable byte swapping */
 	u32 lmi_swap_cntl = 0;
 	u32 mp_swap_cntl = 0;
 
-	WREG32(mmUVD_CGC_GATE, 0);
-	uvd_v4_2_set_dcm(adev, true);
+	/* set uvd busy */
+	WREG32_P(mmUVD_STATUS, 1<<2, ~(1<<2));
 
-	uvd_v4_2_mc_resume(adev);
-
-	/* disable interupt */
-	WREG32_P(mmUVD_MASTINT_EN, 0, ~(1 << 1));
-
-	/* Stall UMC and register bus before resetting VCPU */
-	WREG32_P(mmUVD_LMI_CTRL2, 1 << 8, ~(1 << 8));
-	mdelay(1);
-
-	/* put LMI, VCPU, RBC etc... into reset */
-	WREG32(mmUVD_SOFT_RESET, UVD_SOFT_RESET__LMI_SOFT_RESET_MASK |
-		UVD_SOFT_RESET__VCPU_SOFT_RESET_MASK | UVD_SOFT_RESET__LBSI_SOFT_RESET_MASK |
-		UVD_SOFT_RESET__RBC_SOFT_RESET_MASK | UVD_SOFT_RESET__CSM_SOFT_RESET_MASK |
-		UVD_SOFT_RESET__CXW_SOFT_RESET_MASK | UVD_SOFT_RESET__TAP_SOFT_RESET_MASK |
-		UVD_SOFT_RESET__LMI_UMC_SOFT_RESET_MASK);
-	mdelay(5);
+	uvd_v4_2_set_dcm(adev, true);
+	WREG32(mmUVD_CGC_GATE, 0);
 
 	/* take UVD block out of reset */
 	WREG32_P(mmSRBM_SOFT_RESET, 0, ~SRBM_SOFT_RESET__SOFT_RESET_UVD_MASK);
 	mdelay(5);
 
-	/* initialize UVD memory controller */
-	WREG32(mmUVD_LMI_CTRL, 0x40 | (1 << 8) | (1 << 13) |
-			     (1 << 21) | (1 << 9) | (1 << 20));
+	/* enable VCPU clock */
+	WREG32(mmUVD_VCPU_CNTL,  1 << 9);
+
+	/* disable interupt */
+	WREG32_P(mmUVD_MASTINT_EN, 0, ~(1 << 1));
 
 #ifdef __BIG_ENDIAN
 	/* swap (8 in 32) RB and IB */
@@ -306,6 +293,11 @@ static int uvd_v4_2_start(struct amdgpu_device *adev)
 #endif
 	WREG32(mmUVD_LMI_SWAP_CNTL, lmi_swap_cntl);
 	WREG32(mmUVD_MP_SWAP_CNTL, mp_swap_cntl);
+	/* initialize UVD memory controller */
+	WREG32(mmUVD_LMI_CTRL, 0x203108);
+
+	tmp = RREG32(mmUVD_MPC_CNTL);
+	WREG32(mmUVD_MPC_CNTL, tmp | 0x10);
 
 	WREG32(mmUVD_MPC_SET_MUXA0, 0x40c2040);
 	WREG32(mmUVD_MPC_SET_MUXA1, 0x0);
@@ -314,18 +306,20 @@ static int uvd_v4_2_start(struct amdgpu_device *adev)
 	WREG32(mmUVD_MPC_SET_ALU, 0);
 	WREG32(mmUVD_MPC_SET_MUX, 0x88);
 
-	/* take all subblocks out of reset, except VCPU */
-	WREG32(mmUVD_SOFT_RESET, UVD_SOFT_RESET__VCPU_SOFT_RESET_MASK);
-	mdelay(5);
+	uvd_v4_2_mc_resume(adev);
 
-	/* enable VCPU clock */
-	WREG32(mmUVD_VCPU_CNTL,  1 << 9);
+	tmp = RREG32_UVD_CTX(ixUVD_LMI_CACHE_CTRL);
+	WREG32_UVD_CTX(ixUVD_LMI_CACHE_CTRL, tmp & (~0x10));
 
 	/* enable UMC */
 	WREG32_P(mmUVD_LMI_CTRL2, 0, ~(1 << 8));
 
-	/* boot up the VCPU */
-	WREG32(mmUVD_SOFT_RESET, 0);
+	WREG32_P(mmUVD_SOFT_RESET, 0, ~UVD_SOFT_RESET__LMI_SOFT_RESET_MASK);
+
+	WREG32_P(mmUVD_SOFT_RESET, 0, ~UVD_SOFT_RESET__LMI_UMC_SOFT_RESET_MASK);
+
+	WREG32_P(mmUVD_SOFT_RESET, 0, ~UVD_SOFT_RESET__VCPU_SOFT_RESET_MASK);
+
 	mdelay(10);
 
 	for (i = 0; i < 10; ++i) {
@@ -357,6 +351,8 @@ static int uvd_v4_2_start(struct amdgpu_device *adev)
 	/* enable interupt */
 	WREG32_P(mmUVD_MASTINT_EN, 3<<1, ~(3 << 1));
 
+	WREG32_P(mmUVD_STATUS, 0, ~(1<<2));
+
 	/* force RBC into idle state */
 	WREG32(mmUVD_RBC_RB_CNTL, 0x11010101);
 
@@ -393,22 +389,54 @@ static int uvd_v4_2_start(struct amdgpu_device *adev)
  */
 static void uvd_v4_2_stop(struct amdgpu_device *adev)
 {
-	/* force RBC into idle state */
+	uint32_t i, j;
+	uint32_t status;
+
 	WREG32(mmUVD_RBC_RB_CNTL, 0x11010101);
 
+	for (i = 0; i < 10; ++i) {
+		for (j = 0; j < 100; ++j) {
+			status = RREG32(mmUVD_STATUS);
+			if (status & 2)
+				break;
+			mdelay(1);
+		}
+		break;
+	}
+
+	for (i = 0; i < 10; ++i) {
+		for (j = 0; j < 100; ++j) {
+			status = RREG32(mmUVD_LMI_STATUS);
+			if (status & 0xf)
+				break;
+			mdelay(1);
+		}
+		break;
+	}
+
 	/* Stall UMC and register bus before resetting VCPU */
 	WREG32_P(mmUVD_LMI_CTRL2, 1 << 8, ~(1 << 8));
-	mdelay(1);
 
-	/* put VCPU into reset */
-	WREG32(mmUVD_SOFT_RESET, UVD_SOFT_RESET__VCPU_SOFT_RESET_MASK);
-	mdelay(5);
+	for (i = 0; i < 10; ++i) {
+		for (j = 0; j < 100; ++j) {
+			status = RREG32(mmUVD_LMI_STATUS);
+			if (status & 0x240)
+				break;
+			mdelay(1);
+		}
+		break;
+	}
 
-	/* disable VCPU clock */
-	WREG32(mmUVD_VCPU_CNTL, 0x0);
+	WREG32_P(0x3D49, 0, ~(1 << 2));
 
-	/* Unstall UMC and register bus */
-	WREG32_P(mmUVD_LMI_CTRL2, 0, ~(1 << 8));
+	WREG32_P(mmUVD_VCPU_CNTL, 0, ~(1 << 9));
+
+	/* put LMI, VCPU, RBC etc... into reset */
+	WREG32(mmUVD_SOFT_RESET, UVD_SOFT_RESET__LMI_SOFT_RESET_MASK |
+		UVD_SOFT_RESET__VCPU_SOFT_RESET_MASK |
+		UVD_SOFT_RESET__LMI_UMC_SOFT_RESET_MASK);
+
+	WREG32(mmUVD_STATUS, 0);
 
 	uvd_v4_2_set_dcm(adev, false);
 }
@@ -694,8 +722,24 @@ static int uvd_v4_2_set_powergating_state(void *handle,
 
 	if (state == AMD_PG_STATE_GATE) {
 		uvd_v4_2_stop(adev);
+		if (adev->pg_flags & AMD_PG_SUPPORT_UVD && amdgpu_dpm == 0) {
+			if (!(RREG32_SMC(ixCURRENT_PG_STATUS) & 0x4)) {
+				WREG32(mmUVD_PGFSM_CONFIG, (UVD_PGFSM_CONFIG__UVD_PGFSM_FSM_ADDR_MASK   |
+							UVD_PGFSM_CONFIG__UVD_PGFSM_POWER_DOWN_MASK |
+							UVD_PGFSM_CONFIG__UVD_PGFSM_P1_SELECT_MASK));
+				mdelay(20);
+			}
+		}
 		return 0;
 	} else {
+		if (adev->pg_flags & AMD_PG_SUPPORT_UVD && amdgpu_dpm == 0) {
+			if (RREG32_SMC(ixCURRENT_PG_STATUS) & 0x4) {
+				WREG32(mmUVD_PGFSM_CONFIG, (UVD_PGFSM_CONFIG__UVD_PGFSM_FSM_ADDR_MASK   |
+						UVD_PGFSM_CONFIG__UVD_PGFSM_POWER_UP_MASK |
+						UVD_PGFSM_CONFIG__UVD_PGFSM_P1_SELECT_MASK));
+				mdelay(30);
+			}
+		}
 		return uvd_v4_2_start(adev);
 	}
 }
diff --git a/drivers/gpu/drm/amd/amdgpu/uvd_v5_0.c b/drivers/gpu/drm/amd/amdgpu/uvd_v5_0.c
index 9b49824..46e7151 100644
--- a/drivers/gpu/drm/amd/amdgpu/uvd_v5_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/uvd_v5_0.c
@@ -152,9 +152,9 @@ static int uvd_v5_0_hw_init(void *handle)
 	uint32_t tmp;
 	int r;
 
-	r = uvd_v5_0_start(adev);
-	if (r)
-		goto done;
+	amdgpu_asic_set_uvd_clocks(adev, 10000, 10000);
+	uvd_v5_0_set_clockgating_state(adev, AMD_CG_STATE_UNGATE);
+	uvd_v5_0_enable_mgcg(adev, true);
 
 	ring->ready = true;
 	r = amdgpu_ring_test_ring(ring);
@@ -189,11 +189,13 @@ static int uvd_v5_0_hw_init(void *handle)
 	amdgpu_ring_write(ring, 3);
 
 	amdgpu_ring_commit(ring);
+
 done:
 	if (!r)
 		DRM_INFO("UVD initialized successfully.\n");
 
 	return r;
+
 }
 
 /**
@@ -208,7 +210,9 @@ static int uvd_v5_0_hw_fini(void *handle)
 	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
 	struct amdgpu_ring *ring = &adev->uvd.ring;
 
-	uvd_v5_0_stop(adev);
+	if (RREG32(mmUVD_STATUS) != 0)
+		uvd_v5_0_stop(adev);
+
 	ring->ready = false;
 
 	return 0;
@@ -310,10 +314,6 @@ static int uvd_v5_0_start(struct amdgpu_device *adev)
 
 	uvd_v5_0_mc_resume(adev);
 
-	amdgpu_asic_set_uvd_clocks(adev, 10000, 10000);
-	uvd_v5_0_set_clockgating_state(adev, AMD_CG_STATE_UNGATE);
-	uvd_v5_0_enable_mgcg(adev, true);
-
 	/* disable interupt */
 	WREG32_P(mmUVD_MASTINT_EN, 0, ~(1 << 1));
 
@@ -456,6 +456,8 @@ static void uvd_v5_0_stop(struct amdgpu_device *adev)
 
 	/* Unstall UMC and register bus */
 	WREG32_P(mmUVD_LMI_CTRL2, 0, ~(1 << 8));
+
+	WREG32(mmUVD_STATUS, 0);
 }
 
 /**
@@ -792,9 +794,6 @@ static int uvd_v5_0_set_clockgating_state(void *handle,
 	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
 	bool enable = (state == AMD_CG_STATE_GATE) ? true : false;
 
-	if (!(adev->cg_flags & AMD_CG_SUPPORT_UVD_MGCG))
-		return 0;
-
 	if (enable) {
 		/* wait for STATUS to clear */
 		if (uvd_v5_0_wait_for_idle(handle))
@@ -824,9 +823,6 @@ static int uvd_v5_0_set_powergating_state(void *handle,
 	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
 	int ret = 0;
 
-	if (!(adev->pg_flags & AMD_PG_SUPPORT_UVD))
-		return 0;
-
 	if (state == AMD_PG_STATE_GATE) {
 		uvd_v5_0_stop(adev);
 		adev->uvd.is_powergated = true;
diff --git a/drivers/gpu/drm/amd/amdgpu/uvd_v6_0.c b/drivers/gpu/drm/amd/amdgpu/uvd_v6_0.c
index de7e035..af83ab8 100644
--- a/drivers/gpu/drm/amd/amdgpu/uvd_v6_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/uvd_v6_0.c
@@ -155,9 +155,9 @@ static int uvd_v6_0_hw_init(void *handle)
 	uint32_t tmp;
 	int r;
 
-	r = uvd_v6_0_start(adev);
-	if (r)
-		goto done;
+	amdgpu_asic_set_uvd_clocks(adev, 10000, 10000);
+	uvd_v6_0_set_clockgating_state(adev, AMD_CG_STATE_UNGATE);
+	uvd_v6_0_enable_mgcg(adev, true);
 
 	ring->ready = true;
 	r = amdgpu_ring_test_ring(ring);
@@ -212,7 +212,9 @@ static int uvd_v6_0_hw_fini(void *handle)
 	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
 	struct amdgpu_ring *ring = &adev->uvd.ring;
 
-	uvd_v6_0_stop(adev);
+	if (RREG32(mmUVD_STATUS) != 0)
+		uvd_v6_0_stop(adev);
+
 	ring->ready = false;
 
 	return 0;
@@ -397,9 +399,6 @@ static int uvd_v6_0_start(struct amdgpu_device *adev)
 	lmi_swap_cntl = 0;
 	mp_swap_cntl = 0;
 
-	amdgpu_asic_set_uvd_clocks(adev, 10000, 10000);
-	uvd_v6_0_set_clockgating_state(adev, AMD_CG_STATE_UNGATE);
-	uvd_v6_0_enable_mgcg(adev, true);
 	uvd_v6_0_mc_resume(adev);
 
 	/* disable interupt */
@@ -554,6 +553,8 @@ static void uvd_v6_0_stop(struct amdgpu_device *adev)
 
 	/* Unstall UMC and register bus */
 	WREG32_P(mmUVD_LMI_CTRL2, 0, ~(1 << 8));
+
+	WREG32(mmUVD_STATUS, 0);
 }
 
 /**
@@ -1018,9 +1019,6 @@ static int uvd_v6_0_set_clockgating_state(void *handle,
 	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
 	bool enable = (state == AMD_CG_STATE_GATE) ? true : false;
 
-	if (!(adev->cg_flags & AMD_CG_SUPPORT_UVD_MGCG))
-		return 0;
-
 	if (enable) {
 		/* wait for STATUS to clear */
 		if (uvd_v6_0_wait_for_idle(handle))
@@ -1049,9 +1047,6 @@ static int uvd_v6_0_set_powergating_state(void *handle,
 	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
 	int ret = 0;
 
-	if (!(adev->pg_flags & AMD_PG_SUPPORT_UVD))
-		return 0;
-
 	WREG32(mmUVD_POWER_STATUS, UVD_POWER_STATUS__UVD_PG_EN_MASK);
 
 	if (state == AMD_PG_STATE_GATE) {
diff --git a/drivers/gpu/drm/amd/amdgpu/vce_v2_0.c b/drivers/gpu/drm/amd/amdgpu/vce_v2_0.c
index 38ed903..9ea9934 100644
--- a/drivers/gpu/drm/amd/amdgpu/vce_v2_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/vce_v2_0.c
@@ -42,10 +42,9 @@
 #define VCE_V2_0_DATA_SIZE	(23552 * AMDGPU_MAX_VCE_HANDLES)
 #define VCE_STATUS_VCPU_REPORT_FW_LOADED_MASK	0x02
 
-static void vce_v2_0_mc_resume(struct amdgpu_device *adev);
 static void vce_v2_0_set_ring_funcs(struct amdgpu_device *adev);
 static void vce_v2_0_set_irq_funcs(struct amdgpu_device *adev);
-static int vce_v2_0_wait_for_idle(void *handle);
+
 /**
  * vce_v2_0_ring_get_rptr - get read pointer
  *
@@ -140,6 +139,86 @@ static int vce_v2_0_firmware_loaded(struct amdgpu_device *adev)
 	return -ETIMEDOUT;
 }
 
+static void vce_v2_0_disable_cg(struct amdgpu_device *adev)
+{
+	WREG32(mmVCE_CGTT_CLK_OVERRIDE, 7);
+}
+
+static void vce_v2_0_init_cg(struct amdgpu_device *adev)
+{
+	u32 tmp;
+
+	tmp = RREG32(mmVCE_CLOCK_GATING_A);
+	tmp &= ~0xfff;
+	tmp |= ((0 << 0) | (4 << 4));
+	tmp |= 0x40000;
+	WREG32(mmVCE_CLOCK_GATING_A, tmp);
+
+	tmp = RREG32(mmVCE_UENC_CLOCK_GATING);
+	tmp &= ~0xfff;
+	tmp |= ((0 << 0) | (4 << 4));
+	WREG32(mmVCE_UENC_CLOCK_GATING, tmp);
+
+	tmp = RREG32(mmVCE_CLOCK_GATING_B);
+	tmp |= 0x10;
+	tmp &= ~0x100000;
+	WREG32(mmVCE_CLOCK_GATING_B, tmp);
+}
+
+static void vce_v2_0_mc_resume(struct amdgpu_device *adev)
+{
+	uint64_t addr = adev->vce.gpu_addr;
+	uint32_t size;
+
+	WREG32_P(mmVCE_CLOCK_GATING_A, 0, ~(1 << 16));
+	WREG32_P(mmVCE_UENC_CLOCK_GATING, 0x1FF000, ~0xFF9FF000);
+	WREG32_P(mmVCE_UENC_REG_CLOCK_GATING, 0x3F, ~0x3F);
+	WREG32(mmVCE_CLOCK_GATING_B, 0xf7);
+
+	WREG32(mmVCE_LMI_CTRL, 0x00398000);
+	WREG32_P(mmVCE_LMI_CACHE_CTRL, 0x0, ~0x1);
+	WREG32(mmVCE_LMI_SWAP_CNTL, 0);
+	WREG32(mmVCE_LMI_SWAP_CNTL1, 0);
+	WREG32(mmVCE_LMI_VM_CTRL, 0);
+
+	addr += AMDGPU_VCE_FIRMWARE_OFFSET;
+	size = VCE_V2_0_FW_SIZE;
+	WREG32(mmVCE_VCPU_CACHE_OFFSET0, addr & 0x7fffffff);
+	WREG32(mmVCE_VCPU_CACHE_SIZE0, size);
+
+	addr += size;
+	size = VCE_V2_0_STACK_SIZE;
+	WREG32(mmVCE_VCPU_CACHE_OFFSET1, addr & 0x7fffffff);
+	WREG32(mmVCE_VCPU_CACHE_SIZE1, size);
+
+	addr += size;
+	size = VCE_V2_0_DATA_SIZE;
+	WREG32(mmVCE_VCPU_CACHE_OFFSET2, addr & 0x7fffffff);
+	WREG32(mmVCE_VCPU_CACHE_SIZE2, size);
+
+	WREG32_P(mmVCE_LMI_CTRL2, 0x0, ~0x100);
+	WREG32_FIELD(VCE_SYS_INT_EN, VCE_SYS_INT_TRAP_INTERRUPT_EN, 1);
+}
+
+static bool vce_v2_0_is_idle(void *handle)
+{
+	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+
+	return !(RREG32(mmSRBM_STATUS2) & SRBM_STATUS2__VCE_BUSY_MASK);
+}
+
+static int vce_v2_0_wait_for_idle(void *handle)
+{
+	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	unsigned i;
+
+	for (i = 0; i < adev->usec_timeout; i++) {
+		if (vce_v2_0_is_idle(handle))
+			return 0;
+	}
+	return -ETIMEDOUT;
+}
+
 /**
  * vce_v2_0_start - start VCE block
  *
@@ -152,11 +231,14 @@ static int vce_v2_0_start(struct amdgpu_device *adev)
 	struct amdgpu_ring *ring;
 	int r;
 
-	vce_v2_0_mc_resume(adev);
-
 	/* set BUSY flag */
 	WREG32_P(mmVCE_STATUS, 1, ~1);
 
+	vce_v2_0_init_cg(adev);
+	vce_v2_0_disable_cg(adev);
+
+	vce_v2_0_mc_resume(adev);
+
 	ring = &adev->vce.ring[0];
 	WREG32(mmVCE_RB_RPTR, ring->wptr);
 	WREG32(mmVCE_RB_WPTR, ring->wptr);
@@ -189,6 +271,145 @@ static int vce_v2_0_start(struct amdgpu_device *adev)
 	return 0;
 }
 
+static int vce_v2_0_stop(struct amdgpu_device *adev)
+{
+	int i, j;
+	int status;
+
+	if (vce_v2_0_lmi_clean(adev)) {
+		DRM_INFO("vce is not idle \n");
+		return 0;
+	}
+/*
+	for (i = 0; i < 10; ++i) {
+		for (j = 0; j < 100; ++j) {
+			status = RREG32(mmVCE_FW_REG_STATUS);
+			if (!(status & 1))
+				break;
+			mdelay(1);
+		}
+		break;
+	}
+*/
+	if (vce_v2_0_wait_for_idle(adev)) {
+		DRM_INFO("VCE is busy, Can't set clock gateing");
+		return 0;
+	}
+
+	/* Stall UMC and register bus before resetting VCPU */
+	WREG32_P(mmVCE_LMI_CTRL2, 1 << 8, ~(1 << 8));
+
+	for (i = 0; i < 10; ++i) {
+		for (j = 0; j < 100; ++j) {
+			status = RREG32(mmVCE_LMI_STATUS);
+			if (status & 0x240)
+				break;
+			mdelay(1);
+		}
+		break;
+	}
+
+	WREG32_P(mmVCE_VCPU_CNTL, 0, ~0x80001);
+
+	/* put LMI, VCPU, RBC etc... into reset */
+	WREG32_P(mmVCE_SOFT_RESET, 1, ~0x1);
+
+	WREG32(mmVCE_STATUS, 0);
+
+	return 0;
+}
+
+static void vce_v2_0_set_sw_cg(struct amdgpu_device *adev, bool gated)
+{
+	u32 tmp;
+
+	if (gated) {
+		tmp = RREG32(mmVCE_CLOCK_GATING_B);
+		tmp |= 0xe70000;
+		WREG32(mmVCE_CLOCK_GATING_B, tmp);
+
+		tmp = RREG32(mmVCE_UENC_CLOCK_GATING);
+		tmp |= 0xff000000;
+		WREG32(mmVCE_UENC_CLOCK_GATING, tmp);
+
+		tmp = RREG32(mmVCE_UENC_REG_CLOCK_GATING);
+		tmp &= ~0x3fc;
+		WREG32(mmVCE_UENC_REG_CLOCK_GATING, tmp);
+
+		WREG32(mmVCE_CGTT_CLK_OVERRIDE, 0);
+	} else {
+		tmp = RREG32(mmVCE_CLOCK_GATING_B);
+		tmp |= 0xe7;
+		tmp &= ~0xe70000;
+		WREG32(mmVCE_CLOCK_GATING_B, tmp);
+
+		tmp = RREG32(mmVCE_UENC_CLOCK_GATING);
+		tmp |= 0x1fe000;
+		tmp &= ~0xff000000;
+		WREG32(mmVCE_UENC_CLOCK_GATING, tmp);
+
+		tmp = RREG32(mmVCE_UENC_REG_CLOCK_GATING);
+		tmp |= 0x3fc;
+		WREG32(mmVCE_UENC_REG_CLOCK_GATING, tmp);
+	}
+}
+
+static void vce_v2_0_set_dyn_cg(struct amdgpu_device *adev, bool gated)
+{
+	u32 orig, tmp;
+
+/* LMI_MC/LMI_UMC always set in dynamic,
+ * set {CGC_*_GATE_MODE, CGC_*_SW_GATE} = {0, 0}
+ */
+	tmp = RREG32(mmVCE_CLOCK_GATING_B);
+	tmp &= ~0x00060006;
+
+/* Exception for ECPU, IH, SEM, SYS blocks needs to be turned on/off by SW */
+	if (gated) {
+		tmp |= 0xe10000;
+		WREG32(mmVCE_CLOCK_GATING_B, tmp);
+	} else {
+		tmp |= 0xe1;
+		tmp &= ~0xe10000;
+		WREG32(mmVCE_CLOCK_GATING_B, tmp);
+	}
+
+	orig = tmp = RREG32(mmVCE_UENC_CLOCK_GATING);
+	tmp &= ~0x1fe000;
+	tmp &= ~0xff000000;
+	if (tmp != orig)
+		WREG32(mmVCE_UENC_CLOCK_GATING, tmp);
+
+	orig = tmp = RREG32(mmVCE_UENC_REG_CLOCK_GATING);
+	tmp &= ~0x3fc;
+	if (tmp != orig)
+		WREG32(mmVCE_UENC_REG_CLOCK_GATING, tmp);
+
+	/* set VCE_UENC_REG_CLOCK_GATING always in dynamic mode */
+	WREG32(mmVCE_UENC_REG_CLOCK_GATING, 0x00);
+
+	if(gated)
+		WREG32(mmVCE_CGTT_CLK_OVERRIDE, 0);
+}
+
+static void vce_v2_0_enable_mgcg(struct amdgpu_device *adev, bool enable,
+								bool sw_cg)
+{
+	if (enable && (adev->cg_flags & AMD_CG_SUPPORT_VCE_MGCG)) {
+		if (sw_cg)
+			vce_v2_0_set_sw_cg(adev, true);
+		else
+			vce_v2_0_set_dyn_cg(adev, true);
+	} else {
+		vce_v2_0_disable_cg(adev);
+
+		if (sw_cg)
+			vce_v2_0_set_sw_cg(adev, false);
+		else
+			vce_v2_0_set_dyn_cg(adev, false);
+	}
+}
+
 static int vce_v2_0_early_init(void *handle)
 {
 	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
@@ -254,11 +475,8 @@ static int vce_v2_0_hw_init(void *handle)
 	int r, i;
 	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
 
-	r = vce_v2_0_start(adev);
-	/* this error mean vcpu not in running state, so just skip ring test, not stop driver initialize */
-	if (r)
-		return 0;
-
+	amdgpu_asic_set_vce_clocks(adev, 10000, 10000);
+	vce_v2_0_enable_mgcg(adev, true, false);
 	for (i = 0; i < adev->vce.num_rings; i++)
 		adev->vce.ring[i].ready = false;
 
@@ -312,190 +530,6 @@ static int vce_v2_0_resume(void *handle)
 	return r;
 }
 
-static void vce_v2_0_set_sw_cg(struct amdgpu_device *adev, bool gated)
-{
-	u32 tmp;
-
-	if (gated) {
-		tmp = RREG32(mmVCE_CLOCK_GATING_B);
-		tmp |= 0xe70000;
-		WREG32(mmVCE_CLOCK_GATING_B, tmp);
-
-		tmp = RREG32(mmVCE_UENC_CLOCK_GATING);
-		tmp |= 0xff000000;
-		WREG32(mmVCE_UENC_CLOCK_GATING, tmp);
-
-		tmp = RREG32(mmVCE_UENC_REG_CLOCK_GATING);
-		tmp &= ~0x3fc;
-		WREG32(mmVCE_UENC_REG_CLOCK_GATING, tmp);
-
-		WREG32(mmVCE_CGTT_CLK_OVERRIDE, 0);
-	} else {
-		tmp = RREG32(mmVCE_CLOCK_GATING_B);
-		tmp |= 0xe7;
-		tmp &= ~0xe70000;
-		WREG32(mmVCE_CLOCK_GATING_B, tmp);
-
-		tmp = RREG32(mmVCE_UENC_CLOCK_GATING);
-		tmp |= 0x1fe000;
-		tmp &= ~0xff000000;
-		WREG32(mmVCE_UENC_CLOCK_GATING, tmp);
-
-		tmp = RREG32(mmVCE_UENC_REG_CLOCK_GATING);
-		tmp |= 0x3fc;
-		WREG32(mmVCE_UENC_REG_CLOCK_GATING, tmp);
-	}
-}
-
-static void vce_v2_0_set_dyn_cg(struct amdgpu_device *adev, bool gated)
-{
-	if (vce_v2_0_wait_for_idle(adev)) {
-		DRM_INFO("VCE is busy, Can't set clock gateing");
-		return;
-	}
-
-	WREG32_P(mmVCE_LMI_CTRL2, 0x100, ~0x100);
-
-	if (vce_v2_0_lmi_clean(adev)) {
-		DRM_INFO("LMI is busy, Can't set clock gateing");
-		return;
-	}
-
-	WREG32_P(mmVCE_VCPU_CNTL, 0, ~VCE_VCPU_CNTL__CLK_EN_MASK);
-	WREG32_P(mmVCE_SOFT_RESET,
-		 VCE_SOFT_RESET__ECPU_SOFT_RESET_MASK,
-		 ~VCE_SOFT_RESET__ECPU_SOFT_RESET_MASK);
-	WREG32(mmVCE_STATUS, 0);
-
-	if (gated)
-		WREG32(mmVCE_CGTT_CLK_OVERRIDE, 0);
-	/* LMI_MC/LMI_UMC always set in dynamic, set {CGC_*_GATE_MODE, CGC_*_SW_GATE} = {0, 0} */
-	if (gated) {
-		/* Force CLOCK OFF , set {CGC_*_GATE_MODE, CGC_*_SW_GATE} = {*, 1} */
-		WREG32(mmVCE_CLOCK_GATING_B, 0xe90010);
-	} else {
-		/* Force CLOCK ON, set {CGC_*_GATE_MODE, CGC_*_SW_GATE} = {1, 0} */
-		WREG32(mmVCE_CLOCK_GATING_B, 0x800f1);
-	}
-
-	/* Set VCE_UENC_CLOCK_GATING always in dynamic mode {*_FORCE_ON, *_FORCE_OFF} = {0, 0}*/;
-	WREG32(mmVCE_UENC_CLOCK_GATING, 0x40);
-
-	/* set VCE_UENC_REG_CLOCK_GATING always in dynamic mode */
-	WREG32(mmVCE_UENC_REG_CLOCK_GATING, 0x00);
-
-	WREG32_P(mmVCE_LMI_CTRL2, 0, ~0x100);
-	if(!gated) {
-		WREG32_P(mmVCE_VCPU_CNTL, VCE_VCPU_CNTL__CLK_EN_MASK, ~VCE_VCPU_CNTL__CLK_EN_MASK);
-		mdelay(100);
-		WREG32_P(mmVCE_SOFT_RESET, 0, ~VCE_SOFT_RESET__ECPU_SOFT_RESET_MASK);
-
-		vce_v2_0_firmware_loaded(adev);
-		WREG32_P(mmVCE_STATUS, 0, ~VCE_STATUS__JOB_BUSY_MASK);
-	}
-}
-
-static void vce_v2_0_disable_cg(struct amdgpu_device *adev)
-{
-	WREG32(mmVCE_CGTT_CLK_OVERRIDE, 7);
-}
-
-static void vce_v2_0_enable_mgcg(struct amdgpu_device *adev, bool enable)
-{
-	bool sw_cg = false;
-
-	if (enable && (adev->cg_flags & AMD_CG_SUPPORT_VCE_MGCG)) {
-		if (sw_cg)
-			vce_v2_0_set_sw_cg(adev, true);
-		else
-			vce_v2_0_set_dyn_cg(adev, true);
-	} else {
-		vce_v2_0_disable_cg(adev);
-
-		if (sw_cg)
-			vce_v2_0_set_sw_cg(adev, false);
-		else
-			vce_v2_0_set_dyn_cg(adev, false);
-	}
-}
-
-static void vce_v2_0_init_cg(struct amdgpu_device *adev)
-{
-	u32 tmp;
-
-	tmp = RREG32(mmVCE_CLOCK_GATING_A);
-	tmp &= ~0xfff;
-	tmp |= ((0 << 0) | (4 << 4));
-	tmp |= 0x40000;
-	WREG32(mmVCE_CLOCK_GATING_A, tmp);
-
-	tmp = RREG32(mmVCE_UENC_CLOCK_GATING);
-	tmp &= ~0xfff;
-	tmp |= ((0 << 0) | (4 << 4));
-	WREG32(mmVCE_UENC_CLOCK_GATING, tmp);
-
-	tmp = RREG32(mmVCE_CLOCK_GATING_B);
-	tmp |= 0x10;
-	tmp &= ~0x100000;
-	WREG32(mmVCE_CLOCK_GATING_B, tmp);
-}
-
-static void vce_v2_0_mc_resume(struct amdgpu_device *adev)
-{
-	uint64_t addr = adev->vce.gpu_addr;
-	uint32_t size;
-
-	WREG32_P(mmVCE_CLOCK_GATING_A, 0, ~(1 << 16));
-	WREG32_P(mmVCE_UENC_CLOCK_GATING, 0x1FF000, ~0xFF9FF000);
-	WREG32_P(mmVCE_UENC_REG_CLOCK_GATING, 0x3F, ~0x3F);
-	WREG32(mmVCE_CLOCK_GATING_B, 0xf7);
-
-	WREG32(mmVCE_LMI_CTRL, 0x00398000);
-	WREG32_P(mmVCE_LMI_CACHE_CTRL, 0x0, ~0x1);
-	WREG32(mmVCE_LMI_SWAP_CNTL, 0);
-	WREG32(mmVCE_LMI_SWAP_CNTL1, 0);
-	WREG32(mmVCE_LMI_VM_CTRL, 0);
-
-	addr += AMDGPU_VCE_FIRMWARE_OFFSET;
-	size = VCE_V2_0_FW_SIZE;
-	WREG32(mmVCE_VCPU_CACHE_OFFSET0, addr & 0x7fffffff);
-	WREG32(mmVCE_VCPU_CACHE_SIZE0, size);
-
-	addr += size;
-	size = VCE_V2_0_STACK_SIZE;
-	WREG32(mmVCE_VCPU_CACHE_OFFSET1, addr & 0x7fffffff);
-	WREG32(mmVCE_VCPU_CACHE_SIZE1, size);
-
-	addr += size;
-	size = VCE_V2_0_DATA_SIZE;
-	WREG32(mmVCE_VCPU_CACHE_OFFSET2, addr & 0x7fffffff);
-	WREG32(mmVCE_VCPU_CACHE_SIZE2, size);
-
-	WREG32_P(mmVCE_LMI_CTRL2, 0x0, ~0x100);
-	WREG32_FIELD(VCE_SYS_INT_EN, VCE_SYS_INT_TRAP_INTERRUPT_EN, 1);
-
-	vce_v2_0_init_cg(adev);
-}
-
-static bool vce_v2_0_is_idle(void *handle)
-{
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-
-	return !(RREG32(mmSRBM_STATUS2) & SRBM_STATUS2__VCE_BUSY_MASK);
-}
-
-static int vce_v2_0_wait_for_idle(void *handle)
-{
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-	unsigned i;
-
-	for (i = 0; i < adev->usec_timeout; i++) {
-		if (vce_v2_0_is_idle(handle))
-			return 0;
-	}
-	return -ETIMEDOUT;
-}
-
 static int vce_v2_0_soft_reset(void *handle)
 {
 	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
@@ -539,33 +573,20 @@ static int vce_v2_0_process_interrupt(struct amdgpu_device *adev,
 	return 0;
 }
 
-static void vce_v2_0_set_bypass_mode(struct amdgpu_device *adev, bool enable)
-{
-	u32 tmp = RREG32_SMC(ixGCK_DFS_BYPASS_CNTL);
-
-	if (enable)
-		tmp |= GCK_DFS_BYPASS_CNTL__BYPASSECLK_MASK;
-	else
-		tmp &= ~GCK_DFS_BYPASS_CNTL__BYPASSECLK_MASK;
-
-	WREG32_SMC(ixGCK_DFS_BYPASS_CNTL, tmp);
-}
-
-
 static int vce_v2_0_set_clockgating_state(void *handle,
 					  enum amd_clockgating_state state)
 {
 	bool gate = false;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-	bool enable = (state == AMD_CG_STATE_GATE) ? true : false;
-
+	bool sw_cg = false;
 
-	vce_v2_0_set_bypass_mode(adev, enable);
+	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
 
-	if (state == AMD_CG_STATE_GATE)
+	if (state == AMD_CG_STATE_GATE) {
 		gate = true;
+		sw_cg = true;
+	}
 
-	vce_v2_0_enable_mgcg(adev, gate);
+	vce_v2_0_enable_mgcg(adev, gate, sw_cg);
 
 	return 0;
 }
@@ -582,12 +603,8 @@ static int vce_v2_0_set_powergating_state(void *handle,
 	 */
 	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
 
-	if (!(adev->pg_flags & AMD_PG_SUPPORT_VCE))
-		return 0;
-
 	if (state == AMD_PG_STATE_GATE)
-		/* XXX do we need a vce_v2_0_stop()? */
-		return 0;
+		return vce_v2_0_stop(adev);
 	else
 		return vce_v2_0_start(adev);
 }
diff --git a/drivers/gpu/drm/amd/amdgpu/vce_v3_0.c b/drivers/gpu/drm/amd/amdgpu/vce_v3_0.c
index 8db2655..a8c40ee 100644
--- a/drivers/gpu/drm/amd/amdgpu/vce_v3_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/vce_v3_0.c
@@ -230,10 +230,6 @@ static int vce_v3_0_start(struct amdgpu_device *adev)
 	struct amdgpu_ring *ring;
 	int idx, r;
 
-	vce_v3_0_override_vce_clock_gating(adev, true);
-	if (!(adev->flags & AMD_IS_APU))
-		amdgpu_asic_set_vce_clocks(adev, 10000, 10000);
-
 	ring = &adev->vce.ring[0];
 	WREG32(mmVCE_RB_RPTR, ring->wptr);
 	WREG32(mmVCE_RB_WPTR, ring->wptr);
@@ -436,9 +432,9 @@ static int vce_v3_0_hw_init(void *handle)
 	int r, i;
 	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
 
-	r = vce_v3_0_start(adev);
-	if (r)
-		return r;
+	vce_v3_0_override_vce_clock_gating(adev, true);
+	if (!(adev->flags & AMD_IS_APU))
+		amdgpu_asic_set_vce_clocks(adev, 10000, 10000);
 
 	for (i = 0; i < adev->vce.num_rings; i++)
 		adev->vce.ring[i].ready = false;
@@ -766,12 +762,11 @@ static int vce_v3_0_set_powergating_state(void *handle,
 	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
 	int ret = 0;
 
-	if (!(adev->pg_flags & AMD_PG_SUPPORT_VCE))
-		return 0;
-
 	if (state == AMD_PG_STATE_GATE) {
+		ret = vce_v3_0_stop(adev);
+		if (ret)
+			goto out;
 		adev->vce.is_powergated = true;
-		/* XXX do we need a vce_v3_0_stop()? */
 	} else {
 		ret = vce_v3_0_start(adev);
 		if (ret)
diff --git a/drivers/gpu/drm/amd/include/asic_reg/smu/smu_7_0_1_d.h b/drivers/gpu/drm/amd/include/asic_reg/smu/smu_7_0_1_d.h
index f9fd2ea..dbc2e72 100644
--- a/drivers/gpu/drm/amd/include/asic_reg/smu/smu_7_0_1_d.h
+++ b/drivers/gpu/drm/amd/include/asic_reg/smu/smu_7_0_1_d.h
@@ -1310,5 +1310,6 @@
 #define ixROM_SW_DATA_62                                                        0xc060012c
 #define ixROM_SW_DATA_63                                                        0xc0600130
 #define ixROM_SW_DATA_64                                                        0xc0600134
+#define ixCURRENT_PG_STATUS                                                     0xc020029c
 
 #endif /* SMU_7_0_1_D_H */
diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/cz_clockpowergating.c b/drivers/gpu/drm/amd/powerplay/hwmgr/cz_clockpowergating.c
index 3eccac7..b33935f 100644
--- a/drivers/gpu/drm/amd/powerplay/hwmgr/cz_clockpowergating.c
+++ b/drivers/gpu/drm/amd/powerplay/hwmgr/cz_clockpowergating.c
@@ -161,28 +161,25 @@ int cz_dpm_powergate_uvd(struct pp_hwmgr *hwmgr, bool bgate)
 {
 	struct cz_hwmgr *cz_hwmgr = (struct cz_hwmgr *)(hwmgr->backend);
 
-	if (cz_hwmgr->uvd_power_gated == bgate)
-		return 0;
-
 	cz_hwmgr->uvd_power_gated = bgate;
 
 	if (bgate) {
-		cgs_set_clockgating_state(hwmgr->device,
-						AMD_IP_BLOCK_TYPE_UVD,
-						AMD_CG_STATE_GATE);
 		cgs_set_powergating_state(hwmgr->device,
 						AMD_IP_BLOCK_TYPE_UVD,
 						AMD_PG_STATE_GATE);
+		cgs_set_clockgating_state(hwmgr->device,
+						AMD_IP_BLOCK_TYPE_UVD,
+						AMD_CG_STATE_GATE);
 		cz_dpm_update_uvd_dpm(hwmgr, true);
 		cz_dpm_powerdown_uvd(hwmgr);
 	} else {
 		cz_dpm_powerup_uvd(hwmgr);
-		cgs_set_powergating_state(hwmgr->device,
-						AMD_IP_BLOCK_TYPE_UVD,
-						AMD_CG_STATE_UNGATE);
 		cgs_set_clockgating_state(hwmgr->device,
 						AMD_IP_BLOCK_TYPE_UVD,
 						AMD_PG_STATE_UNGATE);
+		cgs_set_powergating_state(hwmgr->device,
+						AMD_IP_BLOCK_TYPE_UVD,
+						AMD_CG_STATE_UNGATE);
 		cz_dpm_update_uvd_dpm(hwmgr, false);
 	}
 
@@ -193,47 +190,34 @@ int cz_dpm_powergate_vce(struct pp_hwmgr *hwmgr, bool bgate)
 {
 	struct cz_hwmgr *cz_hwmgr = (struct cz_hwmgr *)(hwmgr->backend);
 
-	if (phm_cap_enabled(hwmgr->platform_descriptor.platformCaps,
-					PHM_PlatformCaps_VCEPowerGating)) {
-		if (cz_hwmgr->vce_power_gated != bgate) {
-			if (bgate) {
-				cgs_set_clockgating_state(
-							hwmgr->device,
-							AMD_IP_BLOCK_TYPE_VCE,
-							AMD_CG_STATE_GATE);
-				cgs_set_powergating_state(
-							hwmgr->device,
-							AMD_IP_BLOCK_TYPE_VCE,
-							AMD_PG_STATE_GATE);
-				cz_enable_disable_vce_dpm(hwmgr, false);
-				cz_dpm_powerdown_vce(hwmgr);
-				cz_hwmgr->vce_power_gated = true;
-			} else {
-				cz_dpm_powerup_vce(hwmgr);
-				cz_hwmgr->vce_power_gated = false;
-				cgs_set_powergating_state(
-							hwmgr->device,
-							AMD_IP_BLOCK_TYPE_VCE,
-							AMD_CG_STATE_UNGATE);
-				cgs_set_clockgating_state(
-							hwmgr->device,
-							AMD_IP_BLOCK_TYPE_VCE,
-							AMD_PG_STATE_UNGATE);
-				cz_dpm_update_vce_dpm(hwmgr);
-				cz_enable_disable_vce_dpm(hwmgr, true);
-				return 0;
-			}
-		}
+	if (bgate) {
+		cgs_set_powergating_state(
+					hwmgr->device,
+					AMD_IP_BLOCK_TYPE_VCE,
+					AMD_PG_STATE_GATE);
+		cgs_set_clockgating_state(
+					hwmgr->device,
+					AMD_IP_BLOCK_TYPE_VCE,
+					AMD_CG_STATE_GATE);
+		cz_enable_disable_vce_dpm(hwmgr, false);
+		cz_dpm_powerdown_vce(hwmgr);
+		cz_hwmgr->vce_power_gated = true;
 	} else {
-		cz_hwmgr->vce_power_gated = bgate;
+		cz_dpm_powerup_vce(hwmgr);
+		cz_hwmgr->vce_power_gated = false;
+		cgs_set_clockgating_state(
+					hwmgr->device,
+					AMD_IP_BLOCK_TYPE_VCE,
+					AMD_PG_STATE_UNGATE);
+		cgs_set_powergating_state(
+					hwmgr->device,
+					AMD_IP_BLOCK_TYPE_VCE,
+					AMD_CG_STATE_UNGATE);
 		cz_dpm_update_vce_dpm(hwmgr);
-		cz_enable_disable_vce_dpm(hwmgr, !bgate);
+		cz_enable_disable_vce_dpm(hwmgr, true);
 		return 0;
 	}
 
-	if (!cz_hwmgr->vce_power_gated)
-		cz_dpm_update_vce_dpm(hwmgr);
-
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/smu7_clockpowergating.c b/drivers/gpu/drm/amd/powerplay/hwmgr/smu7_clockpowergating.c
index a1fc4fc..8cf71f3 100644
--- a/drivers/gpu/drm/amd/powerplay/hwmgr/smu7_clockpowergating.c
+++ b/drivers/gpu/drm/amd/powerplay/hwmgr/smu7_clockpowergating.c
@@ -147,22 +147,22 @@ int smu7_powergate_uvd(struct pp_hwmgr *hwmgr, bool bgate)
 	data->uvd_power_gated = bgate;
 
 	if (bgate) {
-		cgs_set_clockgating_state(hwmgr->device,
-				AMD_IP_BLOCK_TYPE_UVD,
-				AMD_CG_STATE_GATE);
 		cgs_set_powergating_state(hwmgr->device,
 						AMD_IP_BLOCK_TYPE_UVD,
 						AMD_PG_STATE_GATE);
+		cgs_set_clockgating_state(hwmgr->device,
+				AMD_IP_BLOCK_TYPE_UVD,
+				AMD_CG_STATE_GATE);
 		smu7_update_uvd_dpm(hwmgr, true);
 		smu7_powerdown_uvd(hwmgr);
 	} else {
 		smu7_powerup_uvd(hwmgr);
-		cgs_set_powergating_state(hwmgr->device,
-						AMD_IP_BLOCK_TYPE_UVD,
-						AMD_CG_STATE_UNGATE);
 		cgs_set_clockgating_state(hwmgr->device,
 				AMD_IP_BLOCK_TYPE_UVD,
 				AMD_CG_STATE_UNGATE);
+		cgs_set_powergating_state(hwmgr->device,
+						AMD_IP_BLOCK_TYPE_UVD,
+						AMD_CG_STATE_UNGATE);
 		smu7_update_uvd_dpm(hwmgr, false);
 	}
 
@@ -173,12 +173,12 @@ int smu7_powergate_vce(struct pp_hwmgr *hwmgr, bool bgate)
 {
 	struct smu7_hwmgr *data = (struct smu7_hwmgr *)(hwmgr->backend);
 
-	if (data->vce_power_gated == bgate)
-		return 0;
-
 	data->vce_power_gated = bgate;
 
 	if (bgate) {
+		cgs_set_powergating_state(hwmgr->device,
+						AMD_IP_BLOCK_TYPE_VCE,
+						AMD_PG_STATE_UNGATE);
 		cgs_set_clockgating_state(hwmgr->device,
 				AMD_IP_BLOCK_TYPE_VCE,
 				AMD_CG_STATE_GATE);
@@ -186,10 +186,13 @@ int smu7_powergate_vce(struct pp_hwmgr *hwmgr, bool bgate)
 		smu7_powerdown_vce(hwmgr);
 	} else {
 		smu7_powerup_vce(hwmgr);
-		smu7_update_vce_dpm(hwmgr, false);
 		cgs_set_clockgating_state(hwmgr->device,
 				AMD_IP_BLOCK_TYPE_VCE,
 				AMD_CG_STATE_UNGATE);
+		cgs_set_powergating_state(hwmgr->device,
+						AMD_IP_BLOCK_TYPE_VCE,
+						AMD_PG_STATE_UNGATE);
+		smu7_update_vce_dpm(hwmgr, false);
 	}
 	return 0;
 }
diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/smu7_hwmgr.c b/drivers/gpu/drm/amd/powerplay/hwmgr/smu7_hwmgr.c
index 0a6c833..b1de9e8 100644
--- a/drivers/gpu/drm/amd/powerplay/hwmgr/smu7_hwmgr.c
+++ b/drivers/gpu/drm/amd/powerplay/hwmgr/smu7_hwmgr.c
@@ -2624,6 +2624,7 @@ static int smu7_force_dpm_level(struct pp_hwmgr *hwmgr,
 		smu7_force_clock_level(hwmgr, PP_SCLK, 1<<sclk_mask);
 		smu7_force_clock_level(hwmgr, PP_MCLK, 1<<mclk_mask);
 		smu7_force_clock_level(hwmgr, PP_PCIE, 1<<pcie_mask);
+
 		break;
 	case AMD_DPM_FORCED_LEVEL_MANUAL:
 		hwmgr->dpm_level = level;
@@ -2633,9 +2634,9 @@ static int smu7_force_dpm_level(struct pp_hwmgr *hwmgr,
 		break;
 	}
 
-	if (level & (AMD_DPM_FORCED_LEVEL_PROFILE_PEAK | AMD_DPM_FORCED_LEVEL_HIGH))
+	if (level == AMD_DPM_FORCED_LEVEL_PROFILE_PEAK && hwmgr->saved_dpm_level != AMD_DPM_FORCED_LEVEL_PROFILE_PEAK)
 		smu7_fan_ctrl_set_fan_speed_percent(hwmgr, 100);
-	else
+	else if (level != AMD_DPM_FORCED_LEVEL_PROFILE_PEAK && hwmgr->saved_dpm_level == AMD_DPM_FORCED_LEVEL_PROFILE_PEAK)
 		smu7_fan_ctrl_reset_fan_speed_to_default(hwmgr);
 
 	return 0;
@@ -4397,16 +4398,14 @@ static int smu7_get_sclks(struct pp_hwmgr *hwmgr, struct amd_pp_clocks *clocks)
 		if (table_info == NULL || table_info->vdd_dep_on_sclk == NULL)
 			return -EINVAL;
 		dep_sclk_table = table_info->vdd_dep_on_sclk;
-		for (i = 0; i < dep_sclk_table->count; i++) {
+		for (i = 0; i < dep_sclk_table->count; i++)
 			clocks->clock[i] = dep_sclk_table->entries[i].clk;
-			clocks->count++;
-		}
+		clocks->count = dep_sclk_table->count;
 	} else if (hwmgr->pp_table_version == PP_TABLE_V0) {
 		sclk_table = hwmgr->dyn_state.vddc_dependency_on_sclk;
-		for (i = 0; i < sclk_table->count; i++) {
+		for (i = 0; i < sclk_table->count; i++)
 			clocks->clock[i] = sclk_table->entries[i].clk;
-			clocks->count++;
-		}
+		clocks->count = sclk_table->count;
 	}
 
 	return 0;
@@ -4440,14 +4439,13 @@ static int smu7_get_mclks(struct pp_hwmgr *hwmgr, struct amd_pp_clocks *clocks)
 			clocks->clock[i] = dep_mclk_table->entries[i].clk;
 			clocks->latency[i] = smu7_get_mem_latency(hwmgr,
 						dep_mclk_table->entries[i].clk);
-			clocks->count++;
 		}
+		clocks->count = dep_mclk_table->count;
 	} else if (hwmgr->pp_table_version == PP_TABLE_V0) {
 		mclk_table = hwmgr->dyn_state.vddc_dependency_on_mclk;
-		for (i = 0; i < mclk_table->count; i++) {
+		for (i = 0; i < mclk_table->count; i++)
 			clocks->clock[i] = mclk_table->entries[i].clk;
-			clocks->count++;
-		}
+		clocks->count = mclk_table->count;
 	}
 	return 0;
 }
diff --git a/drivers/gpu/drm/amd/powerplay/smumgr/smumgr.c b/drivers/gpu/drm/amd/powerplay/smumgr/smumgr.c
index 60c3692..c0956a4 100644
--- a/drivers/gpu/drm/amd/powerplay/smumgr/smumgr.c
+++ b/drivers/gpu/drm/amd/powerplay/smumgr/smumgr.c
@@ -37,8 +37,10 @@ MODULE_FIRMWARE("amdgpu/tonga_k_smc.bin");
 MODULE_FIRMWARE("amdgpu/fiji_smc.bin");
 MODULE_FIRMWARE("amdgpu/polaris10_smc.bin");
 MODULE_FIRMWARE("amdgpu/polaris10_smc_sk.bin");
+MODULE_FIRMWARE("amdgpu/polaris10_k_smc.bin");
 MODULE_FIRMWARE("amdgpu/polaris11_smc.bin");
 MODULE_FIRMWARE("amdgpu/polaris11_smc_sk.bin");
+MODULE_FIRMWARE("amdgpu/polaris11_k_smc.bin");
 MODULE_FIRMWARE("amdgpu/polaris12_smc.bin");
 
 
diff --git a/drivers/gpu/drm/radeon/radeon_bios.c b/drivers/gpu/drm/radeon/radeon_bios.c
index 00cfb5d..04c0ed4 100644
--- a/drivers/gpu/drm/radeon/radeon_bios.c
+++ b/drivers/gpu/drm/radeon/radeon_bios.c
@@ -638,10 +638,8 @@ static bool radeon_acpi_vfct_bios(struct radeon_device *rdev)
 					     vhdr->ImageLength,
 					     GFP_KERNEL);
 
-			if (!rdev->bios) {
-				kfree(rdev->bios);
+			if (!rdev->bios)
 				return false;
-			}
 			return true;
 		}
 	}
diff --git a/drivers/gpu/drm/radeon/vce_v1_0.c b/drivers/gpu/drm/radeon/vce_v1_0.c
index a01efe3..f541a4b 100644
--- a/drivers/gpu/drm/radeon/vce_v1_0.c
+++ b/drivers/gpu/drm/radeon/vce_v1_0.c
@@ -196,7 +196,7 @@ int vce_v1_0_load_fw(struct radeon_device *rdev, uint32_t *data)
 	memset(&data[5], 0, 44);
 	memcpy(&data[16], &sign[1], rdev->vce_fw->size - sizeof(*sign));
 
-	data += le32_to_cpu(data[4]) / 4;
+	data += (le32_to_cpu(sign->len) + 64) / 4;
 	data[0] = sign->val[i].sigval[0];
 	data[1] = sign->val[i].sigval[1];
 	data[2] = sign->val[i].sigval[2];