Merge tag 'amd-drm-next-6.15-2025-03-21' of https://gitlab.freedesktop.org/agd5f/linux into drm-next

amd-drm-next-6.15-2025-03-21:

amdgpu:
- Refine nomodeset handling
- RAS fixes
- DCN 3.x fixes
- DMUB fixes
- eDP fixes
- SMU 14.0.2 fixes
- SMU 13.0.6 fixes
- SMU 13.0.12 fixes
- SDMA engine reset fixes
- Enforce Isolation fixes
- Runtime workload profile ref count fixes
- Documentation fixes
- SR-IOV fixes
- MES fixes
- GC 11.5 cleaner shader support
- SDMA VM invalidation fixes
- IP discovery improvements for GC based chips

amdkfd:
- Dequeue wait count fixes
- Precise memops fixes

radeon:
- Code cleanup

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

From: Alex Deucher <alexander.deucher@amd.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20250321210909.2809595-1-alexander.deucher@amd.com

73 files changed, 1248 insertions, 1052 deletions

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu.h b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
index 2a9a41f4e748..6d83ccfa42ee 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
@@ -1194,9 +1194,15 @@ struct amdgpu_device {
 	bool                            debug_exp_resets;
 	bool                            debug_disable_gpu_ring_reset;
 
-	bool				enforce_isolation[MAX_XCP];
-	/* Added this mutex for cleaner shader isolation between GFX and compute processes */
+	/* Protection for the following isolation structure */
 	struct mutex                    enforce_isolation_mutex;
+	bool				enforce_isolation[MAX_XCP];
+	struct amdgpu_isolation {
+		void			*owner;
+		struct dma_fence	*spearhead;
+		struct amdgpu_sync	active;
+		struct amdgpu_sync	prev;
+	} isolation[MAX_XCP];
 
 	struct amdgpu_init_level *init_lvl;
 
@@ -1482,6 +1488,9 @@ void amdgpu_device_pcie_port_wreg(struct amdgpu_device *adev,
 struct dma_fence *amdgpu_device_get_gang(struct amdgpu_device *adev);
 struct dma_fence *amdgpu_device_switch_gang(struct amdgpu_device *adev,
 					    struct dma_fence *gang);
+struct dma_fence *amdgpu_device_enforce_isolation(struct amdgpu_device *adev,
+						  struct amdgpu_ring *ring,
+						  struct amdgpu_job *job);
 bool amdgpu_device_has_display_hardware(struct amdgpu_device *adev);
 ssize_t amdgpu_get_soft_full_reset_mask(struct amdgpu_ring *ring);
 ssize_t amdgpu_show_reset_mask(char *buf, uint32_t supported_reset);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_aca.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_aca.c
index ffd4c64e123c..dc47f5fd4ea1 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_aca.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_aca.c
@@ -391,6 +391,7 @@ static void aca_banks_generate_cper(struct amdgpu_device *adev,
 {
 	struct aca_bank_node *node;
 	struct aca_bank *bank;
+	int r;
 
 	if (!adev->cper.enabled)
 		return;
@@ -402,11 +403,27 @@ static void aca_banks_generate_cper(struct amdgpu_device *adev,
 
 	/* UEs must be encoded into separate CPER entries */
 	if (type == ACA_SMU_TYPE_UE) {
+		struct aca_banks de_banks;
+
+		aca_banks_init(&de_banks);
 		list_for_each_entry(node, &banks->list, node) {
 			bank = &node->bank;
-			if (amdgpu_cper_generate_ue_record(adev, bank))
-				dev_warn(adev->dev, "fail to generate ue cper records\n");
+			if (bank->aca_err_type == ACA_ERROR_TYPE_DEFERRED) {
+				r = aca_banks_add_bank(&de_banks, bank);
+				if (r)
+					dev_warn(adev->dev, "fail to add de banks, ret = %d\n", r);
+			} else {
+				if (amdgpu_cper_generate_ue_record(adev, bank))
+					dev_warn(adev->dev, "fail to generate ue cper records\n");
+			}
+		}
+
+		if (!list_empty(&de_banks.list)) {
+			if (amdgpu_cper_generate_ce_records(adev, &de_banks, de_banks.nr_banks))
+				dev_warn(adev->dev, "fail to generate de cper records\n");
 		}
+
+		aca_banks_release(&de_banks);
 	} else {
 		/*
 		 * SMU_TYPE_CE banks are combined into 1 CPER entries,
@@ -541,6 +558,10 @@ static int __aca_get_error_data(struct amdgpu_device *adev, struct aca_handle *h
 	if (ret)
 		return ret;
 
+	/* DEs may contain in CEs or UEs */
+	if (type != ACA_ERROR_TYPE_DEFERRED)
+		aca_log_aca_error(handle, ACA_ERROR_TYPE_DEFERRED, err_data);
+
 	return aca_log_aca_error(handle, type, err_data);
 }
 
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_aca.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_aca.h
index 6f62e5d80ed6..6b180f1b33fd 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_aca.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_aca.h
@@ -76,11 +76,17 @@ struct ras_query_context;
 #define mmSMNAID_XCD1_MCA_SMU		0x38430400	/* SMN AID XCD1 */
 #define mmSMNXCD_XCD0_MCA_SMU		0x40430400	/* SMN XCD XCD0 */
 
-#define ACA_BANK_ERR_CE_DE_DECODE(bank)                                  \
-	((ACA_REG__STATUS__POISON((bank)->regs[ACA_REG_IDX_STATUS]) ||   \
-	  ACA_REG__STATUS__DEFERRED((bank)->regs[ACA_REG_IDX_STATUS])) ? \
-		ACA_ERROR_TYPE_DEFERRED :                                \
-		ACA_ERROR_TYPE_CE)
+#define ACA_BANK_ERR_IS_DEFFERED(bank)                                \
+	(ACA_REG__STATUS__POISON((bank)->regs[ACA_REG_IDX_STATUS]) || \
+	 ACA_REG__STATUS__DEFERRED((bank)->regs[ACA_REG_IDX_STATUS]))
+
+#define ACA_BANK_ERR_CE_DE_DECODE(bank)                             \
+	(ACA_BANK_ERR_IS_DEFFERED(bank) ? ACA_ERROR_TYPE_DEFERRED : \
+					  ACA_ERROR_TYPE_CE)
+
+#define ACA_BANK_ERR_UE_DE_DECODE(bank)                             \
+	(ACA_BANK_ERR_IS_DEFFERED(bank) ? ACA_ERROR_TYPE_DEFERRED : \
+					  ACA_ERROR_TYPE_UE)
 
 enum aca_reg_idx {
 	ACA_REG_IDX_CTL			= 0,
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c
index 2ac6d4fa0601..d2ec4130a316 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c
@@ -491,7 +491,7 @@ static int vm_update_pds(struct amdgpu_vm *vm, struct amdgpu_sync *sync)
 	if (ret)
 		return ret;
 
-	return amdgpu_sync_fence(sync, vm->last_update);
+	return amdgpu_sync_fence(sync, vm->last_update, GFP_KERNEL);
 }
 
 static uint64_t get_pte_flags(struct amdgpu_device *adev, struct kgd_mem *mem)
@@ -1249,7 +1249,7 @@ static int unmap_bo_from_gpuvm(struct kgd_mem *mem,
 
 	(void)amdgpu_vm_clear_freed(adev, vm, &bo_va->last_pt_update);
 
-	(void)amdgpu_sync_fence(sync, bo_va->last_pt_update);
+	(void)amdgpu_sync_fence(sync, bo_va->last_pt_update, GFP_KERNEL);
 
 	return 0;
 }
@@ -1273,7 +1273,7 @@ static int update_gpuvm_pte(struct kgd_mem *mem,
 		return ret;
 	}
 
-	return amdgpu_sync_fence(sync, bo_va->last_pt_update);
+	return amdgpu_sync_fence(sync, bo_va->last_pt_update, GFP_KERNEL);
 }
 
 static int map_bo_to_gpuvm(struct kgd_mem *mem,
@@ -2913,7 +2913,7 @@ int amdgpu_amdkfd_gpuvm_restore_process_bos(void *info, struct dma_fence __rcu *
 		}
 		dma_resv_for_each_fence(&cursor, bo->tbo.base.resv,
 					DMA_RESV_USAGE_KERNEL, fence) {
-			ret = amdgpu_sync_fence(&sync_obj, fence);
+			ret = amdgpu_sync_fence(&sync_obj, fence, GFP_KERNEL);
 			if (ret) {
 				pr_debug("Memory eviction: Sync BO fence failed. Try again\n");
 				goto validate_map_fail;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_cper.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_cper.c
index 3f291b30b79f..360e07a5c7c1 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_cper.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_cper.c
@@ -455,10 +455,10 @@ calc:
 		return umin(rec_len, chunk);
 }
 
-void amdgpu_cper_ring_write(struct amdgpu_ring *ring,
-					      void *src, int count)
+void amdgpu_cper_ring_write(struct amdgpu_ring *ring, void *src, int count)
 {
 	u64 pos, wptr_old, rptr = *ring->rptr_cpu_addr & ring->ptr_mask;
+	int rec_cnt_dw = count >> 2;
 	u32 chunk, ent_sz;
 	u8 *s = (u8 *)src;
 
@@ -485,6 +485,9 @@ void amdgpu_cper_ring_write(struct amdgpu_ring *ring,
 		s += chunk;
 	}
 
+	if (ring->count_dw < rec_cnt_dw)
+		ring->count_dw = 0;
+
 	/* the buffer is overflow, adjust rptr */
 	if (((wptr_old < rptr) && (rptr <= ring->wptr)) ||
 	    ((ring->wptr < wptr_old) && (wptr_old < rptr)) ||
@@ -501,12 +504,10 @@ void amdgpu_cper_ring_write(struct amdgpu_ring *ring,
 			pos = rptr;
 		} while (!amdgpu_cper_is_hdr(ring, rptr));
 	}
-	mutex_unlock(&ring->adev->cper.ring_lock);
 
-	if (ring->count_dw >= (count >> 2))
-		ring->count_dw -= (count >> 2);
-	else
-		ring->count_dw = 0;
+	if (ring->count_dw >= rec_cnt_dw)
+		ring->count_dw -= rec_cnt_dw;
+	mutex_unlock(&ring->adev->cper.ring_lock);
 }
 
 static u64 amdgpu_cper_ring_get_rptr(struct amdgpu_ring *ring)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
index 5cc5f59e3018..82df06a72ee0 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
@@ -428,7 +428,7 @@ static int amdgpu_cs_p2_dependencies(struct amdgpu_cs_parser *p,
 			dma_fence_put(old);
 		}
 
-		r = amdgpu_sync_fence(&p->sync, fence);
+		r = amdgpu_sync_fence(&p->sync, fence, GFP_KERNEL);
 		dma_fence_put(fence);
 		if (r)
 			return r;
@@ -450,7 +450,7 @@ static int amdgpu_syncobj_lookup_and_add(struct amdgpu_cs_parser *p,
 		return r;
 	}
 
-	r = amdgpu_sync_fence(&p->sync, fence);
+	r = amdgpu_sync_fence(&p->sync, fence, GFP_KERNEL);
 	dma_fence_put(fence);
 	return r;
 }
@@ -1111,7 +1111,7 @@ static int amdgpu_cs_vm_handling(struct amdgpu_cs_parser *p)
 			struct drm_gpu_scheduler *sched = entity->rq->sched;
 			struct amdgpu_ring *ring = to_amdgpu_ring(sched);
 
-			if (amdgpu_vmid_uses_reserved(adev, vm, ring->vm_hub))
+			if (amdgpu_vmid_uses_reserved(vm, ring->vm_hub))
 				return -EINVAL;
 		}
 	}
@@ -1124,7 +1124,8 @@ static int amdgpu_cs_vm_handling(struct amdgpu_cs_parser *p)
 	if (r)
 		return r;
 
-	r = amdgpu_sync_fence(&p->sync, fpriv->prt_va->last_pt_update);
+	r = amdgpu_sync_fence(&p->sync, fpriv->prt_va->last_pt_update,
+			      GFP_KERNEL);
 	if (r)
 		return r;
 
@@ -1135,7 +1136,8 @@ static int amdgpu_cs_vm_handling(struct amdgpu_cs_parser *p)
 		if (r)
 			return r;
 
-		r = amdgpu_sync_fence(&p->sync, bo_va->last_pt_update);
+		r = amdgpu_sync_fence(&p->sync, bo_va->last_pt_update,
+				      GFP_KERNEL);
 		if (r)
 			return r;
 	}
@@ -1154,7 +1156,8 @@ static int amdgpu_cs_vm_handling(struct amdgpu_cs_parser *p)
 		if (r)
 			return r;
 
-		r = amdgpu_sync_fence(&p->sync, bo_va->last_pt_update);
+		r = amdgpu_sync_fence(&p->sync, bo_va->last_pt_update,
+				      GFP_KERNEL);
 		if (r)
 			return r;
 	}
@@ -1167,7 +1170,7 @@ static int amdgpu_cs_vm_handling(struct amdgpu_cs_parser *p)
 	if (r)
 		return r;
 
-	r = amdgpu_sync_fence(&p->sync, vm->last_update);
+	r = amdgpu_sync_fence(&p->sync, vm->last_update, GFP_KERNEL);
 	if (r)
 		return r;
 
@@ -1248,7 +1251,8 @@ static int amdgpu_cs_sync_rings(struct amdgpu_cs_parser *p)
 			continue;
 		}
 
-		r = amdgpu_sync_fence(&p->gang_leader->explicit_sync, fence);
+		r = amdgpu_sync_fence(&p->gang_leader->explicit_sync, fence,
+				      GFP_KERNEL);
 		dma_fence_put(fence);
 		if (r)
 			return r;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
index 17ce1677378e..6ebf6179064b 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
@@ -227,6 +227,24 @@ static ssize_t amdgpu_device_get_pcie_replay_count(struct device *dev,
 static DEVICE_ATTR(pcie_replay_count, 0444,
 		amdgpu_device_get_pcie_replay_count, NULL);
 
+static int amdgpu_device_attr_sysfs_init(struct amdgpu_device *adev)
+{
+	int ret = 0;
+
+	if (!amdgpu_sriov_vf(adev))
+		ret = sysfs_create_file(&adev->dev->kobj,
+					&dev_attr_pcie_replay_count.attr);
+
+	return ret;
+}
+
+static void amdgpu_device_attr_sysfs_fini(struct amdgpu_device *adev)
+{
+	if (!amdgpu_sriov_vf(adev))
+		sysfs_remove_file(&adev->dev->kobj,
+				  &dev_attr_pcie_replay_count.attr);
+}
+
 static ssize_t amdgpu_sysfs_reg_state_get(struct file *f, struct kobject *kobj,
 					  struct bin_attribute *attr, char *buf,
 					  loff_t ppos, size_t count)
@@ -4172,11 +4190,6 @@ static bool amdgpu_device_check_iommu_remap(struct amdgpu_device *adev)
 }
 #endif
 
-static const struct attribute *amdgpu_dev_attributes[] = {
-	&dev_attr_pcie_replay_count.attr,
-	NULL
-};
-
 static void amdgpu_device_set_mcbp(struct amdgpu_device *adev)
 {
 	if (amdgpu_mcbp == 1)
@@ -4281,7 +4294,14 @@ int amdgpu_device_init(struct amdgpu_device *adev,
 	mutex_init(&adev->gfx.reset_sem_mutex);
 	/* Initialize the mutex for cleaner shader isolation between GFX and compute processes */
 	mutex_init(&adev->enforce_isolation_mutex);
+	for (i = 0; i < MAX_XCP; ++i) {
+		adev->isolation[i].spearhead = dma_fence_get_stub();
+		amdgpu_sync_create(&adev->isolation[i].active);
+		amdgpu_sync_create(&adev->isolation[i].prev);
+	}
 	mutex_init(&adev->gfx.kfd_sch_mutex);
+	mutex_init(&adev->gfx.workload_profile_mutex);
+	mutex_init(&adev->vcn.workload_profile_mutex);
 
 	amdgpu_device_init_apu_flags(adev);
 
@@ -4399,10 +4419,17 @@ int amdgpu_device_init(struct amdgpu_device *adev,
 	if (r)
 		return r;
 
-	/* Get rid of things like offb */
-	r = aperture_remove_conflicting_pci_devices(adev->pdev, amdgpu_kms_driver.name);
-	if (r)
-		return r;
+	/*
+	 * No need to remove conflicting FBs for non-display class devices.
+	 * This prevents the sysfb from being freed accidently.
+	 */
+	if ((pdev->class >> 8) == PCI_CLASS_DISPLAY_VGA ||
+	    (pdev->class >> 8) == PCI_CLASS_DISPLAY_OTHER) {
+		/* Get rid of things like offb */
+		r = aperture_remove_conflicting_pci_devices(adev->pdev, amdgpu_kms_driver.name);
+		if (r)
+			return r;
+	}
 
 	/* Enable TMZ based on IP_VERSION */
 	amdgpu_gmc_tmz_set(adev);
@@ -4613,7 +4640,7 @@ fence_driver_init:
 	} else
 		adev->ucode_sysfs_en = true;
 
-	r = sysfs_create_files(&adev->dev->kobj, amdgpu_dev_attributes);
+	r = amdgpu_device_attr_sysfs_init(adev);
 	if (r)
 		dev_err(adev->dev, "Could not create amdgpu device attr\n");
 
@@ -4750,7 +4777,7 @@ void amdgpu_device_fini_hw(struct amdgpu_device *adev)
 		amdgpu_pm_sysfs_fini(adev);
 	if (adev->ucode_sysfs_en)
 		amdgpu_ucode_sysfs_fini(adev);
-	sysfs_remove_files(&adev->dev->kobj, amdgpu_dev_attributes);
+	amdgpu_device_attr_sysfs_fini(adev);
 	amdgpu_fru_sysfs_fini(adev);
 
 	amdgpu_reg_state_sysfs_fini(adev);
@@ -4777,7 +4804,7 @@ void amdgpu_device_fini_hw(struct amdgpu_device *adev)
 
 void amdgpu_device_fini_sw(struct amdgpu_device *adev)
 {
-	int idx;
+	int i, idx;
 	bool px;
 
 	amdgpu_device_ip_fini(adev);
@@ -4785,6 +4812,11 @@ void amdgpu_device_fini_sw(struct amdgpu_device *adev)
 	amdgpu_ucode_release(&adev->firmware.gpu_info_fw);
 	adev->accel_working = false;
 	dma_fence_put(rcu_dereference_protected(adev->gang_submit, true));
+	for (i = 0; i < MAX_XCP; ++i) {
+		dma_fence_put(adev->isolation[i].spearhead);
+		amdgpu_sync_free(&adev->isolation[i].active);
+		amdgpu_sync_free(&adev->isolation[i].prev);
+	}
 
 	amdgpu_reset_fini(adev);
 
@@ -4800,6 +4832,9 @@ void amdgpu_device_fini_sw(struct amdgpu_device *adev)
 	kfree(adev->fru_info);
 	adev->fru_info = NULL;
 
+	kfree(adev->xcp_mgr);
+	adev->xcp_mgr = NULL;
+
 	px = amdgpu_device_supports_px(adev_to_drm(adev));
 
 	if (px || (!dev_is_removable(&adev->pdev->dev) &&
@@ -5331,6 +5366,7 @@ static int amdgpu_device_reset_sriov(struct amdgpu_device *adev,
 	if (amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 2) ||
 	    amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 3) ||
 	    amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 4) ||
+	    amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 5, 0) ||
 	    amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(11, 0, 3))
 		amdgpu_ras_resume(adev);
 
@@ -6903,22 +6939,117 @@ struct dma_fence *amdgpu_device_switch_gang(struct amdgpu_device *adev,
 {
 	struct dma_fence *old = NULL;
 
+	dma_fence_get(gang);
 	do {
 		dma_fence_put(old);
 		old = amdgpu_device_get_gang(adev);
 		if (old == gang)
 			break;
 
-		if (!dma_fence_is_signaled(old))
+		if (!dma_fence_is_signaled(old)) {
+			dma_fence_put(gang);
 			return old;
+		}
 
 	} while (cmpxchg((struct dma_fence __force **)&adev->gang_submit,
 			 old, gang) != old);
 
+	/*
+	 * Drop it once for the exchanged reference in adev and once for the
+	 * thread local reference acquired in amdgpu_device_get_gang().
+	 */
+	dma_fence_put(old);
 	dma_fence_put(old);
 	return NULL;
 }
 
+/**
+ * amdgpu_device_enforce_isolation - enforce HW isolation
+ * @adev: the amdgpu device pointer
+ * @ring: the HW ring the job is supposed to run on
+ * @job: the job which is about to be pushed to the HW ring
+ *
+ * Makes sure that only one client at a time can use the GFX block.
+ * Returns: The dependency to wait on before the job can be pushed to the HW.
+ * The function is called multiple times until NULL is returned.
+ */
+struct dma_fence *amdgpu_device_enforce_isolation(struct amdgpu_device *adev,
+						  struct amdgpu_ring *ring,
+						  struct amdgpu_job *job)
+{
+	struct amdgpu_isolation *isolation = &adev->isolation[ring->xcp_id];
+	struct drm_sched_fence *f = job->base.s_fence;
+	struct dma_fence *dep;
+	void *owner;
+	int r;
+
+	/*
+	 * For now enforce isolation only for the GFX block since we only need
+	 * the cleaner shader on those rings.
+	 */
+	if (ring->funcs->type != AMDGPU_RING_TYPE_GFX &&
+	    ring->funcs->type != AMDGPU_RING_TYPE_COMPUTE)
+		return NULL;
+
+	/*
+	 * All submissions where enforce isolation is false are handled as if
+	 * they come from a single client. Use ~0l as the owner to distinct it
+	 * from kernel submissions where the owner is NULL.
+	 */
+	owner = job->enforce_isolation ? f->owner : (void *)~0l;
+
+	mutex_lock(&adev->enforce_isolation_mutex);
+
+	/*
+	 * The "spearhead" submission is the first one which changes the
+	 * ownership to its client. We always need to wait for it to be
+	 * pushed to the HW before proceeding with anything.
+	 */
+	if (&f->scheduled != isolation->spearhead &&
+	    !dma_fence_is_signaled(isolation->spearhead)) {
+		dep = isolation->spearhead;
+		goto out_grab_ref;
+	}
+
+	if (isolation->owner != owner) {
+
+		/*
+		 * Wait for any gang to be assembled before switching to a
+		 * different owner or otherwise we could deadlock the
+		 * submissions.
+		 */
+		if (!job->gang_submit) {
+			dep = amdgpu_device_get_gang(adev);
+			if (!dma_fence_is_signaled(dep))
+				goto out_return_dep;
+			dma_fence_put(dep);
+		}
+
+		dma_fence_put(isolation->spearhead);
+		isolation->spearhead = dma_fence_get(&f->scheduled);
+		amdgpu_sync_move(&isolation->active, &isolation->prev);
+		trace_amdgpu_isolation(isolation->owner, owner);
+		isolation->owner = owner;
+	}
+
+	/*
+	 * Specifying the ring here helps to pipeline submissions even when
+	 * isolation is enabled. If that is not desired for testing NULL can be
+	 * used instead of the ring to enforce a CPU round trip while switching
+	 * between clients.
+	 */
+	dep = amdgpu_sync_peek_fence(&isolation->prev, ring);
+	r = amdgpu_sync_fence(&isolation->active, &f->finished, GFP_NOWAIT);
+	if (r)
+		DRM_WARN("OOM tracking isolation\n");
+
+out_grab_ref:
+	dma_fence_get(dep);
+out_return_dep:
+	mutex_unlock(&adev->enforce_isolation_mutex);
+	return dep;
+}
+
 bool amdgpu_device_has_display_hardware(struct amdgpu_device *adev)
 {
 	switch (adev->asic_type) {
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.c
index 967a992829bd..dc2713ec95a5 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.c
@@ -113,8 +113,13 @@
 #include "amdgpu_isp.h"
 #endif
 
-#define FIRMWARE_IP_DISCOVERY "amdgpu/ip_discovery.bin"
-MODULE_FIRMWARE(FIRMWARE_IP_DISCOVERY);
+MODULE_FIRMWARE("amdgpu/ip_discovery.bin");
+MODULE_FIRMWARE("amdgpu/vega10_ip_discovery.bin");
+MODULE_FIRMWARE("amdgpu/vega12_ip_discovery.bin");
+MODULE_FIRMWARE("amdgpu/vega20_ip_discovery.bin");
+MODULE_FIRMWARE("amdgpu/raven_ip_discovery.bin");
+MODULE_FIRMWARE("amdgpu/raven2_ip_discovery.bin");
+MODULE_FIRMWARE("amdgpu/picasso_ip_discovery.bin");
 
 #define mmIP_DISCOVERY_VERSION  0x16A00
 #define mmRCC_CONFIG_MEMSIZE	0xde3
@@ -297,21 +302,13 @@ static int amdgpu_discovery_read_binary_from_mem(struct amdgpu_device *adev,
 	return ret;
 }
 
-static int amdgpu_discovery_read_binary_from_file(struct amdgpu_device *adev, uint8_t *binary)
+static int amdgpu_discovery_read_binary_from_file(struct amdgpu_device *adev,
+							uint8_t *binary,
+							const char *fw_name)
 {
 	const struct firmware *fw;
-	const char *fw_name;
 	int r;
 
-	switch (amdgpu_discovery) {
-	case 2:
-		fw_name = FIRMWARE_IP_DISCOVERY;
-		break;
-	default:
-		dev_warn(adev->dev, "amdgpu_discovery is not set properly\n");
-		return -EINVAL;
-	}
-
 	r = request_firmware(&fw, fw_name, adev->dev);
 	if (r) {
 		dev_err(adev->dev, "can't load firmware \"%s\"\n",
@@ -404,10 +401,39 @@ static int amdgpu_discovery_verify_npsinfo(struct amdgpu_device *adev,
 	return 0;
 }
 
+static const char *amdgpu_discovery_get_fw_name(struct amdgpu_device *adev)
+{
+	if (amdgpu_discovery == 2)
+		return "amdgpu/ip_discovery.bin";
+
+	switch (adev->asic_type) {
+	case CHIP_VEGA10:
+		return "amdgpu/vega10_ip_discovery.bin";
+	case CHIP_VEGA12:
+		return "amdgpu/vega12_ip_discovery.bin";
+	case CHIP_RAVEN:
+		if (adev->apu_flags & AMD_APU_IS_RAVEN2)
+			return "amdgpu/raven2_ip_discovery.bin";
+		else if (adev->apu_flags & AMD_APU_IS_PICASSO)
+			return "amdgpu/picasso_ip_discovery.bin";
+		else
+			return "amdgpu/raven_ip_discovery.bin";
+	case CHIP_VEGA20:
+		return "amdgpu/vega20_ip_discovery.bin";
+	case CHIP_ARCTURUS:
+		return "amdgpu/arcturus_ip_discovery.bin";
+	case CHIP_ALDEBARAN:
+		return "amdgpu/aldebaran_ip_discovery.bin";
+	default:
+		return NULL;
+	}
+}
+
 static int amdgpu_discovery_init(struct amdgpu_device *adev)
 {
 	struct table_info *info;
 	struct binary_header *bhdr;
+	const char *fw_name;
 	uint16_t offset;
 	uint16_t size;
 	uint16_t checksum;
@@ -419,9 +445,10 @@ static int amdgpu_discovery_init(struct amdgpu_device *adev)
 		return -ENOMEM;
 
 	/* Read from file if it is the preferred option */
-	if (amdgpu_discovery == 2) {
+	fw_name = amdgpu_discovery_get_fw_name(adev);
+	if (fw_name != NULL) {
 		dev_info(adev->dev, "use ip discovery information from file");
-		r = amdgpu_discovery_read_binary_from_file(adev, adev->mman.discovery_bin);
+		r = amdgpu_discovery_read_binary_from_file(adev, adev->mman.discovery_bin, fw_name);
 
 		if (r) {
 			dev_err(adev->dev, "failed to read ip discovery binary from file\n");
@@ -1290,6 +1317,7 @@ static int amdgpu_discovery_reg_base_init(struct amdgpu_device *adev)
 	uint16_t die_offset;
 	uint16_t ip_offset;
 	uint16_t num_dies;
+	uint32_t wafl_ver;
 	uint16_t num_ips;
 	uint16_t hw_id;
 	uint8_t inst;
@@ -1303,6 +1331,7 @@ static int amdgpu_discovery_reg_base_init(struct amdgpu_device *adev)
 		return r;
 	}
 
+	wafl_ver = 0;
 	adev->gfx.xcc_mask = 0;
 	adev->sdma.sdma_mask = 0;
 	adev->vcn.inst_mask = 0;
@@ -1403,6 +1432,10 @@ static int amdgpu_discovery_reg_base_init(struct amdgpu_device *adev)
 				adev->gfx.xcc_mask |=
 					(1U << ip->instance_number);
 
+			if (!wafl_ver && le16_to_cpu(ip->hw_id) == WAFLC_HWID)
+				wafl_ver = IP_VERSION_FULL(ip->major, ip->minor,
+							   ip->revision, 0, 0);
+
 			for (k = 0; k < num_base_address; k++) {
 				/*
 				 * convert the endianness of base addresses in place,
@@ -1468,6 +1501,9 @@ next_ip:
 		}
 	}
 
+	if (wafl_ver && !adev->ip_versions[XGMI_HWIP][0])
+		adev->ip_versions[XGMI_HWIP][0] = wafl_ver;
+
 	return 0;
 }
 
@@ -2511,6 +2547,38 @@ int amdgpu_discovery_set_ip_blocks(struct amdgpu_device *adev)
 
 	switch (adev->asic_type) {
 	case CHIP_VEGA10:
+	case CHIP_VEGA12:
+	case CHIP_RAVEN:
+	case CHIP_VEGA20:
+	case CHIP_ARCTURUS:
+	case CHIP_ALDEBARAN:
+		/* this is not fatal.  We have a fallback below
+		 * if the new firmwares are not present. some of
+		 * this will be overridden below to keep things
+		 * consistent with the current behavior.
+		 */
+		r = amdgpu_discovery_reg_base_init(adev);
+		if (!r) {
+			amdgpu_discovery_harvest_ip(adev);
+			amdgpu_discovery_get_gfx_info(adev);
+			amdgpu_discovery_get_mall_info(adev);
+			amdgpu_discovery_get_vcn_info(adev);
+		}
+		break;
+	default:
+		r = amdgpu_discovery_reg_base_init(adev);
+		if (r)
+			return -EINVAL;
+
+		amdgpu_discovery_harvest_ip(adev);
+		amdgpu_discovery_get_gfx_info(adev);
+		amdgpu_discovery_get_mall_info(adev);
+		amdgpu_discovery_get_vcn_info(adev);
+		break;
+	}
+
+	switch (adev->asic_type) {
+	case CHIP_VEGA10:
 		vega10_reg_base_init(adev);
 		adev->sdma.num_instances = 2;
 		adev->gmc.num_umc = 4;
@@ -2673,14 +2741,6 @@ int amdgpu_discovery_set_ip_blocks(struct amdgpu_device *adev)
 		adev->ip_versions[XGMI_HWIP][0] = IP_VERSION(6, 1, 0);
 		break;
 	default:
-		r = amdgpu_discovery_reg_base_init(adev);
-		if (r)
-			return -EINVAL;
-
-		amdgpu_discovery_harvest_ip(adev);
-		amdgpu_discovery_get_gfx_info(adev);
-		amdgpu_discovery_get_mall_info(adev);
-		amdgpu_discovery_get_vcn_info(adev);
 		break;
 	}
 
@@ -2772,10 +2832,6 @@ int amdgpu_discovery_set_ip_blocks(struct amdgpu_device *adev)
 		break;
 	}
 
-	if (amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 3) ||
-	    amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 4))
-		adev->ip_versions[XGMI_HWIP][0] = IP_VERSION(6, 4, 0);
-
 	/* set NBIO version */
 	switch (amdgpu_ip_version(adev, NBIO_HWIP, 0)) {
 	case IP_VERSION(6, 1, 0):
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
index 653f2bc77530..23cfce5aa1fc 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
@@ -139,6 +139,7 @@ enum AMDGPU_DEBUG_MASK {
 	AMDGPU_DEBUG_ENABLE_RAS_ACA = BIT(4),
 	AMDGPU_DEBUG_ENABLE_EXP_RESETS = BIT(5),
 	AMDGPU_DEBUG_DISABLE_GPU_RING_RESET = BIT(6),
+	AMDGPU_DEBUG_SMU_POOL = BIT(7),
 };
 
 unsigned int amdgpu_vram_limit = UINT_MAX;
@@ -176,6 +177,7 @@ uint amdgpu_sdma_phase_quantum = 32;
 char *amdgpu_disable_cu;
 char *amdgpu_virtual_display;
 bool enforce_isolation;
+int amdgpu_modeset = -1;
 
 /* Specifies the default granularity for SVM, used in buffer
  * migration and restoration of backing memory when handling
@@ -1038,6 +1040,13 @@ module_param(enforce_isolation, bool, 0444);
 MODULE_PARM_DESC(enforce_isolation, "enforce process isolation between graphics and compute . enforce_isolation = on");
 
 /**
+ * DOC: modeset (int)
+ * Override nomodeset (1 = override, -1 = auto). The default is -1 (auto).
+ */
+MODULE_PARM_DESC(modeset, "Override nomodeset (1 = enable, -1 = auto)");
+module_param_named(modeset, amdgpu_modeset, int, 0444);
+
+/**
  * DOC: seamless (int)
  * Seamless boot will keep the image on the screen during the boot process.
  */
@@ -1053,6 +1062,11 @@ module_param_named(seamless, amdgpu_seamless, int, 0444);
  *   limits the VRAM size reported to ROCm applications to the visible
  *   size, usually 256MB.
  * - 0x4: Disable GPU soft recovery, always do a full reset
+ * - 0x8: Use VRAM for firmware loading
+ * - 0x10: Enable ACA based RAS logging
+ * - 0x20: Enable experimental resets
+ * - 0x40: Disable ring resets
+ * - 0x80: Use VRAM for SMU pool
  */
 MODULE_PARM_DESC(debug_mask, "debug options for amdgpu, disabled by default");
 module_param_named_unsafe(debug_mask, amdgpu_debug_mask, uint, 0444);
@@ -2230,6 +2244,10 @@ static void amdgpu_init_debug_options(struct amdgpu_device *adev)
 		pr_info("debug: ring reset disabled\n");
 		adev->debug_disable_gpu_ring_reset = true;
 	}
+	if (amdgpu_debug_mask & AMDGPU_DEBUG_SMU_POOL) {
+		pr_info("debug: use vram for smu pool\n");
+		adev->pm.smu_debug_mask |= SMU_DEBUG_POOL_USE_VRAM;
+	}
 }
 
 static unsigned long amdgpu_fix_asic_type(struct pci_dev *pdev, unsigned long flags)
@@ -2257,6 +2275,12 @@ static int amdgpu_pci_probe(struct pci_dev *pdev,
 	int ret, retry = 0, i;
 	bool supports_atomic = false;
 
+	if ((pdev->class >> 8) == PCI_CLASS_DISPLAY_VGA ||
+	    (pdev->class >> 8) == PCI_CLASS_DISPLAY_OTHER) {
+		if (drm_firmware_drivers_only() && amdgpu_modeset == -1)
+			return -EINVAL;
+	}
+
 	/* skip devices which are owned by radeon */
 	for (i = 0; i < ARRAY_SIZE(amdgpu_unsupported_pciidlist); i++) {
 		if (amdgpu_unsupported_pciidlist[i] == pdev->device)
@@ -2990,9 +3014,6 @@ static int __init amdgpu_init(void)
 {
 	int r;
 
-	if (drm_firmware_drivers_only())
-		return -EINVAL;
-
 	r = amdgpu_sync_init();
 	if (r)
 		goto error_sync;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c
index 984e6ff6e463..72af5e5a894a 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c
@@ -1665,15 +1665,8 @@ static ssize_t amdgpu_gfx_set_enforce_isolation(struct device *dev,
 	}
 
 	mutex_lock(&adev->enforce_isolation_mutex);
-	for (i = 0; i < num_partitions; i++) {
-		if (adev->enforce_isolation[i] && !partition_values[i])
-			/* Going from enabled to disabled */
-			amdgpu_vmid_free_reserved(adev, AMDGPU_GFXHUB(i));
-		else if (!adev->enforce_isolation[i] && partition_values[i])
-			/* Going from disabled to enabled */
-			amdgpu_vmid_alloc_reserved(adev, AMDGPU_GFXHUB(i));
+	for (i = 0; i < num_partitions; i++)
 		adev->enforce_isolation[i] = partition_values[i];
-	}
 	mutex_unlock(&adev->enforce_isolation_mutex);
 
 	amdgpu_mes_update_enforce_isolation(adev);
@@ -2160,11 +2153,16 @@ void amdgpu_gfx_profile_idle_work_handler(struct work_struct *work)
 	for (i = 0; i < (AMDGPU_MAX_COMPUTE_RINGS * AMDGPU_MAX_GC_INSTANCES); ++i)
 		fences += amdgpu_fence_count_emitted(&adev->gfx.compute_ring[i]);
 	if (!fences && !atomic_read(&adev->gfx.total_submission_cnt)) {
-		r = amdgpu_dpm_switch_power_profile(adev, profile, false);
-		if (r)
-			dev_warn(adev->dev, "(%d) failed to disable %s power profile mode\n", r,
-				 profile == PP_SMC_POWER_PROFILE_FULLSCREEN3D ?
-				 "fullscreen 3D" : "compute");
+		mutex_lock(&adev->gfx.workload_profile_mutex);
+		if (adev->gfx.workload_profile_active) {
+			r = amdgpu_dpm_switch_power_profile(adev, profile, false);
+			if (r)
+				dev_warn(adev->dev, "(%d) failed to disable %s power profile mode\n", r,
+					 profile == PP_SMC_POWER_PROFILE_FULLSCREEN3D ?
+					 "fullscreen 3D" : "compute");
+			adev->gfx.workload_profile_active = false;
+		}
+		mutex_unlock(&adev->gfx.workload_profile_mutex);
 	} else {
 		schedule_delayed_work(&adev->gfx.idle_work, GFX_PROFILE_IDLE_TIMEOUT);
 	}
@@ -2183,13 +2181,25 @@ void amdgpu_gfx_profile_ring_begin_use(struct amdgpu_ring *ring)
 
 	atomic_inc(&adev->gfx.total_submission_cnt);
 
-	if (!cancel_delayed_work_sync(&adev->gfx.idle_work)) {
+	cancel_delayed_work_sync(&adev->gfx.idle_work);
+
+	/* We can safely return early here because we've cancelled the
+	 * the delayed work so there is no one else to set it to false
+	 * and we don't care if someone else sets it to true.
+	 */
+	if (adev->gfx.workload_profile_active)
+		return;
+
+	mutex_lock(&adev->gfx.workload_profile_mutex);
+	if (!adev->gfx.workload_profile_active) {
 		r = amdgpu_dpm_switch_power_profile(adev, profile, true);
 		if (r)
 			dev_warn(adev->dev, "(%d) failed to disable %s power profile mode\n", r,
 				 profile == PP_SMC_POWER_PROFILE_FULLSCREEN3D ?
 				 "fullscreen 3D" : "compute");
+		adev->gfx.workload_profile_active = true;
 	}
+	mutex_unlock(&adev->gfx.workload_profile_mutex);
 }
 
 void amdgpu_gfx_profile_ring_end_use(struct amdgpu_ring *ring)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.h
index 4b1675f79caa..87e862188766 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.h
@@ -482,6 +482,8 @@ struct amdgpu_gfx {
 
 	atomic_t			total_submission_cnt;
 	struct delayed_work		idle_work;
+	bool				workload_profile_active;
+	struct mutex                    workload_profile_mutex;
 };
 
 struct amdgpu_gfx_ras_reg_entry {
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c
index 4eefa17fa39b..464625282872 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c
@@ -573,6 +573,7 @@ int amdgpu_gmc_allocate_vm_inv_eng(struct amdgpu_device *adev)
 	unsigned vm_inv_engs[AMDGPU_MAX_VMHUBS] = {0};
 	unsigned i;
 	unsigned vmhub, inv_eng;
+	struct amdgpu_ring *shared_ring;
 
 	/* init the vm inv eng for all vmhubs */
 	for_each_set_bit(i, adev->vmhubs_mask, AMDGPU_MAX_VMHUBS) {
@@ -595,6 +596,10 @@ int amdgpu_gmc_allocate_vm_inv_eng(struct amdgpu_device *adev)
 		    ring == &adev->cper.ring_buf)
 			continue;
 
+		/* Skip if the ring is a shared ring */
+		if (amdgpu_sdma_is_shared_inv_eng(adev, ring))
+			continue;
+
 		inv_eng = ffs(vm_inv_engs[vmhub]);
 		if (!inv_eng) {
 			dev_err(adev->dev, "no VM inv eng for ring %s\n",
@@ -607,6 +612,21 @@ int amdgpu_gmc_allocate_vm_inv_eng(struct amdgpu_device *adev)
 
 		dev_info(adev->dev, "ring %s uses VM inv eng %u on hub %u\n",
 			 ring->name, ring->vm_inv_eng, ring->vm_hub);
+		/* SDMA has a special packet which allows it to use the same
+		 * invalidation engine for all the rings in one instance.
+		 * Therefore, we do not allocate a separate VM invalidation engine
+		 * for SDMA page rings. Instead, they share the VM invalidation
+		 * engine with the SDMA gfx ring. This change ensures efficient
+		 * resource management and avoids the issue of insufficient VM
+		 * invalidation engines.
+		 */
+		shared_ring = amdgpu_sdma_get_shared_ring(adev, ring);
+		if (shared_ring) {
+			shared_ring->vm_inv_eng = ring->vm_inv_eng;
+			dev_info(adev->dev, "ring %s shares VM invalidation engine %u with ring %s on hub %u\n",
+					ring->name, ring->vm_inv_eng, shared_ring->name, ring->vm_hub);
+			continue;
+		}
 	}
 
 	return 0;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ids.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ids.c
index 8e712a11aba5..4c4e087230ac 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ids.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ids.c
@@ -209,7 +209,7 @@ static int amdgpu_vmid_grab_idle(struct amdgpu_ring *ring,
 		return 0;
 	}
 
-	fences = kmalloc_array(id_mgr->num_ids, sizeof(void *), GFP_KERNEL);
+	fences = kmalloc_array(id_mgr->num_ids, sizeof(void *), GFP_NOWAIT);
 	if (!fences)
 		return -ENOMEM;
 
@@ -287,46 +287,34 @@ static int amdgpu_vmid_grab_reserved(struct amdgpu_vm *vm,
 	    (*id)->flushed_updates < updates ||
 	    !(*id)->last_flush ||
 	    ((*id)->last_flush->context != fence_context &&
-	     !dma_fence_is_signaled((*id)->last_flush))) {
+	     !dma_fence_is_signaled((*id)->last_flush)))
+		needs_flush = true;
+
+	if ((*id)->owner != vm->immediate.fence_context ||
+	    (!adev->vm_manager.concurrent_flush && needs_flush)) {
 		struct dma_fence *tmp;
 
-		/* Wait for the gang to be assembled before using a
-		 * reserved VMID or otherwise the gang could deadlock.
+		/* Don't use per engine and per process VMID at the
+		 * same time
 		 */
-		tmp = amdgpu_device_get_gang(adev);
-		if (!dma_fence_is_signaled(tmp) && tmp != job->gang_submit) {
+		if (adev->vm_manager.concurrent_flush)
+			ring = NULL;
+
+		/* to prevent one context starved by another context */
+		(*id)->pd_gpu_addr = 0;
+		tmp = amdgpu_sync_peek_fence(&(*id)->active, ring);
+		if (tmp) {
 			*id = NULL;
-			*fence = tmp;
+			*fence = dma_fence_get(tmp);
 			return 0;
 		}
-		dma_fence_put(tmp);
-
-		/* Make sure the id is owned by the gang before proceeding */
-		if (!job->gang_submit ||
-		    (*id)->owner != vm->immediate.fence_context) {
-
-			/* Don't use per engine and per process VMID at the
-			 * same time
-			 */
-			if (adev->vm_manager.concurrent_flush)
-				ring = NULL;
-
-			/* to prevent one context starved by another context */
-			(*id)->pd_gpu_addr = 0;
-			tmp = amdgpu_sync_peek_fence(&(*id)->active, ring);
-			if (tmp) {
-				*id = NULL;
-				*fence = dma_fence_get(tmp);
-				return 0;
-			}
-		}
-		needs_flush = true;
 	}
 
 	/* Good we can use this VMID. Remember this submission as
 	* user of the VMID.
 	*/
-	r = amdgpu_sync_fence(&(*id)->active, &job->base.s_fence->finished);
+	r = amdgpu_sync_fence(&(*id)->active, &job->base.s_fence->finished,
+			      GFP_NOWAIT);
 	if (r)
 		return r;
 
@@ -385,7 +373,8 @@ static int amdgpu_vmid_grab_used(struct amdgpu_vm *vm,
 		 * user of the VMID.
 		 */
 		r = amdgpu_sync_fence(&(*id)->active,
-				      &job->base.s_fence->finished);
+				      &job->base.s_fence->finished,
+				      GFP_NOWAIT);
 		if (r)
 			return r;
 
@@ -422,7 +411,7 @@ int amdgpu_vmid_grab(struct amdgpu_vm *vm, struct amdgpu_ring *ring,
 	if (r || !idle)
 		goto error;
 
-	if (amdgpu_vmid_uses_reserved(adev, vm, vmhub)) {
+	if (amdgpu_vmid_uses_reserved(vm, vmhub)) {
 		r = amdgpu_vmid_grab_reserved(vm, ring, job, &id, fence);
 		if (r || !id)
 			goto error;
@@ -437,7 +426,8 @@ int amdgpu_vmid_grab(struct amdgpu_vm *vm, struct amdgpu_ring *ring,
 
 			/* Remember this submission as user of the VMID */
 			r = amdgpu_sync_fence(&id->active,
-					      &job->base.s_fence->finished);
+					      &job->base.s_fence->finished,
+					      GFP_NOWAIT);
 			if (r)
 				goto error;
 
@@ -474,19 +464,14 @@ error:
 
 /*
  * amdgpu_vmid_uses_reserved - check if a VM will use a reserved VMID
- * @adev: amdgpu_device pointer
  * @vm: the VM to check
  * @vmhub: the VMHUB which will be used
  *
  * Returns: True if the VM will use a reserved VMID.
  */
-bool amdgpu_vmid_uses_reserved(struct amdgpu_device *adev,
-			       struct amdgpu_vm *vm, unsigned int vmhub)
+bool amdgpu_vmid_uses_reserved(struct amdgpu_vm *vm, unsigned int vmhub)
 {
-	return vm->reserved_vmid[vmhub] ||
-		(adev->enforce_isolation[(vm->root.bo->xcp_id != AMDGPU_XCP_NO_PARTITION) ?
-					 vm->root.bo->xcp_id : 0] &&
-		 AMDGPU_IS_GFXHUB(vmhub));
+	return vm->reserved_vmid[vmhub];
 }
 
 int amdgpu_vmid_alloc_reserved(struct amdgpu_device *adev,
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ids.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_ids.h
index 4012fb2dd08a..240fa6751260 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ids.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ids.h
@@ -78,8 +78,7 @@ void amdgpu_pasid_free_delayed(struct dma_resv *resv,
 
 bool amdgpu_vmid_had_gpu_reset(struct amdgpu_device *adev,
 			       struct amdgpu_vmid *id);
-bool amdgpu_vmid_uses_reserved(struct amdgpu_device *adev,
-			       struct amdgpu_vm *vm, unsigned int vmhub);
+bool amdgpu_vmid_uses_reserved(struct amdgpu_vm *vm, unsigned int vmhub);
 int amdgpu_vmid_alloc_reserved(struct amdgpu_device *adev,
 				unsigned vmhub);
 void amdgpu_vmid_free_reserved(struct amdgpu_device *adev,
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c
index 935df2cdcc16..acb21fc8b3ce 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c
@@ -361,17 +361,24 @@ amdgpu_job_prepare_job(struct drm_sched_job *sched_job,
 {
 	struct amdgpu_ring *ring = to_amdgpu_ring(s_entity->rq->sched);
 	struct amdgpu_job *job = to_amdgpu_job(sched_job);
-	struct dma_fence *fence = NULL;
+	struct dma_fence *fence;
 	int r;
 
 	r = drm_sched_entity_error(s_entity);
 	if (r)
 		goto error;
 
-	if (job->gang_submit)
+	if (job->gang_submit) {
 		fence = amdgpu_device_switch_gang(ring->adev, job->gang_submit);
+		if (fence)
+			return fence;
+	}
+
+	fence = amdgpu_device_enforce_isolation(ring->adev, ring, job);
+	if (fence)
+		return fence;
 
-	if (!fence && job->vm && !job->vmid) {
+	if (job->vm && !job->vmid) {
 		r = amdgpu_vmid_grab(job->vm, ring, job, &fence);
 		if (r) {
 			dev_err(ring->adev->dev, "Error getting VM ID (%d)\n", r);
@@ -384,9 +391,10 @@ amdgpu_job_prepare_job(struct drm_sched_job *sched_job,
 		 */
 		if (!fence)
 			job->vm = NULL;
+		return fence;
 	}
 
-	return fence;
+	return NULL;
 
 error:
 	dma_fence_set_error(&job->base.s_fence->finished, r);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_mes.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_mes.c
index 709c11cbeabd..85f774063f9b 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_mes.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_mes.c
@@ -145,9 +145,8 @@ int amdgpu_mes_init(struct amdgpu_device *adev)
 	adev->mes.vmid_mask_gfxhub = 0xffffff00;
 
 	for (i = 0; i < AMDGPU_MES_MAX_COMPUTE_PIPES; i++) {
-		/* use only 1st MEC pipes */
-		if (i >= adev->gfx.mec.num_pipe_per_mec)
-			continue;
+		if (i >= (adev->gfx.mec.num_pipe_per_mec * adev->gfx.mec.num_mec))
+			break;
 		adev->mes.compute_hqd_mask[i] = 0xc;
 	}
 
@@ -155,14 +154,9 @@ int amdgpu_mes_init(struct amdgpu_device *adev)
 		adev->mes.gfx_hqd_mask[i] = i ? 0 : 0xfffffffe;
 
 	for (i = 0; i < AMDGPU_MES_MAX_SDMA_PIPES; i++) {
-		if (amdgpu_ip_version(adev, SDMA0_HWIP, 0) <
-		    IP_VERSION(6, 0, 0))
-			adev->mes.sdma_hqd_mask[i] = i ? 0 : 0x3fc;
-		/* zero sdma_hqd_mask for non-existent engine */
-		else if (adev->sdma.num_instances == 1)
-			adev->mes.sdma_hqd_mask[i] = i ? 0 : 0xfc;
-		else
-			adev->mes.sdma_hqd_mask[i] = 0xfc;
+		if (i >= adev->sdma.num_instances)
+			break;
+		adev->mes.sdma_hqd_mask[i] = 0xfc;
 	}
 
 	for (i = 0; i < AMDGPU_MAX_MES_PIPES; i++) {
@@ -1336,14 +1330,14 @@ int amdgpu_mes_ctx_map_meta_data(struct amdgpu_device *adev,
 		DRM_ERROR("failed to do vm_bo_update on meta data\n");
 		goto error_del_bo_va;
 	}
-	amdgpu_sync_fence(&sync, bo_va->last_pt_update);
+	amdgpu_sync_fence(&sync, bo_va->last_pt_update, GFP_KERNEL);
 
 	r = amdgpu_vm_update_pdes(adev, vm, false);
 	if (r) {
 		DRM_ERROR("failed to update pdes on meta data\n");
 		goto error_del_bo_va;
 	}
-	amdgpu_sync_fence(&sync, vm->last_update);
+	amdgpu_sync_fence(&sync, vm->last_update, GFP_KERNEL);
 
 	amdgpu_sync_wait(&sync, false);
 	drm_exec_fini(&exec);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c
index 0f2eb69ad715..d54bb1377262 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c
@@ -153,6 +153,9 @@ static int psp_init_sriov_microcode(struct psp_context *psp)
 		adev->virt.autoload_ucode_id = AMDGPU_UCODE_ID_CP_MES1_DATA;
 		ret = psp_init_cap_microcode(psp, ucode_prefix);
 		break;
+	case IP_VERSION(13, 0, 12):
+		ret = psp_init_ta_microcode(psp, ucode_prefix);
+		break;
 	default:
 		return -EINVAL;
 	}
@@ -1861,6 +1864,7 @@ int psp_ras_initialize(struct psp_context *psp)
 	if (adev->gmc.gmc_funcs->query_mem_partition_mode)
 		ras_cmd->ras_in_message.init_flags.nps_mode =
 			adev->gmc.gmc_funcs->query_mem_partition_mode(adev);
+	ras_cmd->ras_in_message.init_flags.active_umc_mask = adev->umc.active_mask;
 
 	ret = psp_ta_load(psp, &psp->ras_context.context);
 
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c
index cfec29835634..bed2603ae4c4 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c
@@ -3473,6 +3473,13 @@ int amdgpu_ras_init_badpage_info(struct amdgpu_device *adev)
 				adev, control->bad_channel_bitmap);
 			con->update_channel_flag = false;
 		}
+
+		/* The format action is only applied to new ASICs */
+		if (IP_VERSION_MAJ(amdgpu_ip_version(adev, UMC_HWIP, 0)) >= 12 &&
+		    control->tbl_hdr.version < RAS_TABLE_VER_V3)
+			if (!amdgpu_ras_eeprom_reset_table(control))
+				if (amdgpu_ras_save_bad_pages(adev, NULL))
+					dev_warn(adev->dev, "Failed to format RAS EEPROM data in V3 version!\n");
 	}
 
 	return ret;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.c
index 3597ecd9baca..0ea7cfaf3587 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.c
@@ -161,6 +161,7 @@ static bool __is_ras_eeprom_supported(struct amdgpu_device *adev)
 	case IP_VERSION(13, 0, 10):
 		return true;
 	case IP_VERSION(13, 0, 6):
+	case IP_VERSION(13, 0, 12):
 	case IP_VERSION(13, 0, 14):
 		return (adev->gmc.is_app_apu) ? false : true;
 	default:
@@ -223,6 +224,7 @@ static bool __get_eeprom_i2c_addr(struct amdgpu_device *adev,
 		return true;
 	case IP_VERSION(13, 0, 6):
 	case IP_VERSION(13, 0, 10):
+	case IP_VERSION(13, 0, 12):
 	case IP_VERSION(13, 0, 14):
 		control->i2c_address = EEPROM_I2C_MADDR_4;
 		return true;
@@ -413,9 +415,11 @@ static void amdgpu_ras_set_eeprom_table_version(struct amdgpu_ras_eeprom_control
 
 	switch (amdgpu_ip_version(adev, UMC_HWIP, 0)) {
 	case IP_VERSION(8, 10, 0):
-	case IP_VERSION(12, 0, 0):
 		hdr->version = RAS_TABLE_VER_V2_1;
 		return;
+	case IP_VERSION(12, 0, 0):
+		hdr->version = RAS_TABLE_VER_V3;
+		return;
 	default:
 		hdr->version = RAS_TABLE_VER_V1;
 		return;
@@ -443,7 +447,7 @@ int amdgpu_ras_eeprom_reset_table(struct amdgpu_ras_eeprom_control *control)
 	hdr->header = RAS_TABLE_HDR_VAL;
 	amdgpu_ras_set_eeprom_table_version(control);
 
-	if (hdr->version == RAS_TABLE_VER_V2_1) {
+	if (hdr->version >= RAS_TABLE_VER_V2_1) {
 		hdr->first_rec_offset = RAS_RECORD_START_V2_1;
 		hdr->tbl_size = RAS_TABLE_HEADER_SIZE +
 				RAS_TABLE_V2_1_INFO_SIZE;
@@ -461,7 +465,7 @@ int amdgpu_ras_eeprom_reset_table(struct amdgpu_ras_eeprom_control *control)
 	}
 
 	csum = __calc_hdr_byte_sum(control);
-	if (hdr->version == RAS_TABLE_VER_V2_1)
+	if (hdr->version >= RAS_TABLE_VER_V2_1)
 		csum += __calc_ras_info_byte_sum(control);
 	csum = -csum;
 	hdr->checksum = csum;
@@ -757,7 +761,7 @@ amdgpu_ras_eeprom_update_header(struct amdgpu_ras_eeprom_control *control)
 			"Saved bad pages %d reaches threshold value %d\n",
 			control->ras_num_bad_pages, ras->bad_page_cnt_threshold);
 		control->tbl_hdr.header = RAS_TABLE_HDR_BAD;
-		if (control->tbl_hdr.version == RAS_TABLE_VER_V2_1) {
+		if (control->tbl_hdr.version >= RAS_TABLE_VER_V2_1) {
 			control->tbl_rai.rma_status = GPU_RETIRED__ECC_REACH_THRESHOLD;
 			control->tbl_rai.health_percent = 0;
 		}
@@ -770,7 +774,7 @@ amdgpu_ras_eeprom_update_header(struct amdgpu_ras_eeprom_control *control)
 		amdgpu_dpm_send_rma_reason(adev);
 	}
 
-	if (control->tbl_hdr.version == RAS_TABLE_VER_V2_1)
+	if (control->tbl_hdr.version >= RAS_TABLE_VER_V2_1)
 		control->tbl_hdr.tbl_size = RAS_TABLE_HEADER_SIZE +
 					    RAS_TABLE_V2_1_INFO_SIZE +
 					    control->ras_num_recs * RAS_TABLE_RECORD_SIZE;
@@ -810,7 +814,7 @@ amdgpu_ras_eeprom_update_header(struct amdgpu_ras_eeprom_control *control)
 	 * now calculate gpu health percent
 	 */
 	if (amdgpu_bad_page_threshold != 0 &&
-	    control->tbl_hdr.version == RAS_TABLE_VER_V2_1 &&
+	    control->tbl_hdr.version >= RAS_TABLE_VER_V2_1 &&
 	    control->ras_num_bad_pages <= ras->bad_page_cnt_threshold)
 		control->tbl_rai.health_percent = ((ras->bad_page_cnt_threshold -
 						   control->ras_num_bad_pages) * 100) /
@@ -823,7 +827,7 @@ amdgpu_ras_eeprom_update_header(struct amdgpu_ras_eeprom_control *control)
 		csum += *pp;
 
 	csum += __calc_hdr_byte_sum(control);
-	if (control->tbl_hdr.version == RAS_TABLE_VER_V2_1)
+	if (control->tbl_hdr.version >= RAS_TABLE_VER_V2_1)
 		csum += __calc_ras_info_byte_sum(control);
 	/* avoid sign extension when assigning to "checksum" */
 	csum = -csum;
@@ -1040,7 +1044,7 @@ uint32_t amdgpu_ras_eeprom_max_record_count(struct amdgpu_ras_eeprom_control *co
 	/* get available eeprom table version first before eeprom table init */
 	amdgpu_ras_set_eeprom_table_version(control);
 
-	if (control->tbl_hdr.version == RAS_TABLE_VER_V2_1)
+	if (control->tbl_hdr.version >= RAS_TABLE_VER_V2_1)
 		return RAS_MAX_RECORD_COUNT_V2_1;
 	else
 		return RAS_MAX_RECORD_COUNT;
@@ -1285,7 +1289,7 @@ static int __verify_ras_table_checksum(struct amdgpu_ras_eeprom_control *control
 	int buf_size, res;
 	u8  csum, *buf, *pp;
 
-	if (control->tbl_hdr.version == RAS_TABLE_VER_V2_1)
+	if (control->tbl_hdr.version >= RAS_TABLE_VER_V2_1)
 		buf_size = RAS_TABLE_HEADER_SIZE +
 			   RAS_TABLE_V2_1_INFO_SIZE +
 			   control->ras_num_recs * RAS_TABLE_RECORD_SIZE;
@@ -1388,7 +1392,7 @@ int amdgpu_ras_eeprom_init(struct amdgpu_ras_eeprom_control *control)
 
 	__decode_table_header_from_buf(hdr, buf);
 
-	if (hdr->version == RAS_TABLE_VER_V2_1) {
+	if (hdr->version >= RAS_TABLE_VER_V2_1) {
 		control->ras_num_recs = RAS_NUM_RECS_V2_1(hdr);
 		control->ras_record_offset = RAS_RECORD_START_V2_1;
 		control->ras_max_record_count = RAS_MAX_RECORD_COUNT_V2_1;
@@ -1428,7 +1432,7 @@ int amdgpu_ras_eeprom_check(struct amdgpu_ras_eeprom_control *control)
 		DRM_DEBUG_DRIVER("Found existing EEPROM table with %d records",
 				 control->ras_num_bad_pages);
 
-		if (hdr->version == RAS_TABLE_VER_V2_1) {
+		if (hdr->version >= RAS_TABLE_VER_V2_1) {
 			res = __read_table_ras_info(control);
 			if (res)
 				return res;
@@ -1448,7 +1452,7 @@ int amdgpu_ras_eeprom_check(struct amdgpu_ras_eeprom_control *control)
 					ras->bad_page_cnt_threshold);
 	} else if (hdr->header == RAS_TABLE_HDR_BAD &&
 		   amdgpu_bad_page_threshold != 0) {
-		if (hdr->version == RAS_TABLE_VER_V2_1) {
+		if (hdr->version >= RAS_TABLE_VER_V2_1) {
 			res = __read_table_ras_info(control);
 			if (res)
 				return res;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.h
index 13f7eda9a696..ec6d7ea37ad0 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.h
@@ -28,6 +28,7 @@
 
 #define RAS_TABLE_VER_V1           0x00010000
 #define RAS_TABLE_VER_V2_1         0x00021000
+#define RAS_TABLE_VER_V3           0x00030000
 
 struct amdgpu_device;
 
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h
index b4fd1e17205e..bb2b66385223 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h
@@ -37,7 +37,7 @@ struct amdgpu_job;
 struct amdgpu_vm;
 
 /* max number of rings */
-#define AMDGPU_MAX_RINGS		133
+#define AMDGPU_MAX_RINGS		149
 #define AMDGPU_MAX_HWIP_RINGS		64
 #define AMDGPU_MAX_GFX_RINGS		2
 #define AMDGPU_MAX_SW_GFX_RINGS         2
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_sdma.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_sdma.c
index 3a4cef896018..529c9696c2f3 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_sdma.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_sdma.c
@@ -504,6 +504,39 @@ void amdgpu_sdma_sysfs_reset_mask_fini(struct amdgpu_device *adev)
 	}
 }
 
+struct amdgpu_ring *amdgpu_sdma_get_shared_ring(struct amdgpu_device *adev, struct amdgpu_ring *ring)
+{
+	if (adev->sdma.has_page_queue &&
+	    (ring->me < adev->sdma.num_instances) &&
+	    (ring == &adev->sdma.instance[ring->me].ring))
+		return &adev->sdma.instance[ring->me].page;
+	else
+		return NULL;
+}
+
+/**
+* amdgpu_sdma_is_shared_inv_eng - Check if a ring is an SDMA ring that shares a VM invalidation engine
+* @adev: Pointer to the AMDGPU device structure
+* @ring: Pointer to the ring structure to check
+*
+* This function checks if the given ring is an SDMA ring that shares a VM invalidation engine.
+* It returns true if the ring is such an SDMA ring, false otherwise.
+*/
+bool amdgpu_sdma_is_shared_inv_eng(struct amdgpu_device *adev, struct amdgpu_ring *ring)
+{
+	int i = ring->me;
+
+	if (!adev->sdma.has_page_queue || i >= adev->sdma.num_instances)
+		return false;
+
+	if (amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 3) ||
+	    amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 4) ||
+	    amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 5, 0))
+		return (ring == &adev->sdma.instance[i].page);
+	else
+		return false;
+}
+
 /**
  * amdgpu_sdma_register_on_reset_callbacks - Register SDMA reset callbacks
  * @funcs: Pointer to the callback structure containing pre_reset and post_reset functions
@@ -532,7 +565,6 @@ void amdgpu_sdma_register_on_reset_callbacks(struct amdgpu_device *adev, struct
  * amdgpu_sdma_reset_engine - Reset a specific SDMA engine
  * @adev: Pointer to the AMDGPU device
  * @instance_id: ID of the SDMA engine instance to reset
- * @suspend_user_queues: check if suspend user queue.
  *
  * This function performs the following steps:
  * 1. Calls all registered pre_reset callbacks to allow KFD and AMDGPU to save their state.
@@ -541,22 +573,16 @@ void amdgpu_sdma_register_on_reset_callbacks(struct amdgpu_device *adev, struct
  *
  * Returns: 0 on success, or a negative error code on failure.
  */
-int amdgpu_sdma_reset_engine(struct amdgpu_device *adev, uint32_t instance_id, bool suspend_user_queues)
+int amdgpu_sdma_reset_engine(struct amdgpu_device *adev, uint32_t instance_id)
 {
 	struct sdma_on_reset_funcs *funcs;
 	int ret = 0;
-	struct amdgpu_sdma_instance *sdma_instance = &adev->sdma.instance[instance_id];;
+	struct amdgpu_sdma_instance *sdma_instance = &adev->sdma.instance[instance_id];
 	struct amdgpu_ring *gfx_ring = &sdma_instance->ring;
 	struct amdgpu_ring *page_ring = &sdma_instance->page;
 	bool gfx_sched_stopped = false, page_sched_stopped = false;
 
-	/* Suspend KFD if suspend_user_queues is true.
-	 * prevent the destruction of in-flight healthy user queue packets and
-	 * avoid race conditions between KFD and KGD during the reset process.
-	 */
-	if (suspend_user_queues)
-		amdgpu_amdkfd_suspend(adev, false);
-
+	mutex_lock(&sdma_instance->engine_reset_mutex);
 	/* Stop the scheduler's work queue for the GFX and page rings if they are running.
 	* This ensures that no new tasks are submitted to the queues while
 	* the reset is in progress.
@@ -609,7 +635,7 @@ exit:
 	 * if they were stopped by this function. This allows new tasks
 	 * to be submitted to the queues after the reset is complete.
 	 */
-	if (ret) {
+	if (!ret) {
 		if (gfx_sched_stopped && amdgpu_ring_sched_ready(gfx_ring)) {
 			drm_sched_wqueue_start(&gfx_ring->sched);
 		}
@@ -617,9 +643,7 @@ exit:
 			drm_sched_wqueue_start(&page_ring->sched);
 		}
 	}
-
-	if (suspend_user_queues)
-		amdgpu_amdkfd_resume(adev, false);
+	mutex_unlock(&sdma_instance->engine_reset_mutex);
 
 	return ret;
 }
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_sdma.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_sdma.h
index 965169320065..47d56fd0589f 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_sdma.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_sdma.h
@@ -64,6 +64,11 @@ struct amdgpu_sdma_instance {
 	struct amdgpu_bo	*sdma_fw_obj;
 	uint64_t		sdma_fw_gpu_addr;
 	uint32_t		*sdma_fw_ptr;
+	struct mutex		engine_reset_mutex;
+	/* track guilty state of GFX and PAGE queues */
+	bool			gfx_guilty;
+	bool			page_guilty;
+
 };
 
 enum amdgpu_sdma_ras_memory_id {
@@ -126,9 +131,6 @@ struct amdgpu_sdma {
 	uint32_t		*ip_dump;
 	uint32_t 		supported_reset;
 	struct list_head	reset_callback_list;
-	/* track guilty state of GFX and PAGE queues */
-	bool gfx_guilty;
-	bool page_guilty;
 };
 
 /*
@@ -169,7 +171,7 @@ struct amdgpu_buffer_funcs {
 };
 
 void amdgpu_sdma_register_on_reset_callbacks(struct amdgpu_device *adev, struct sdma_on_reset_funcs *funcs);
-int amdgpu_sdma_reset_engine(struct amdgpu_device *adev, uint32_t instance_id, bool suspend_user_queues);
+int amdgpu_sdma_reset_engine(struct amdgpu_device *adev, uint32_t instance_id);
 
 #define amdgpu_emit_copy_buffer(adev, ib, s, d, b, t) (adev)->mman.buffer_funcs->emit_copy_buffer((ib),  (s), (d), (b), (t))
 #define amdgpu_emit_fill_buffer(adev, ib, s, d, b) (adev)->mman.buffer_funcs->emit_fill_buffer((ib), (s), (d), (b))
@@ -194,4 +196,7 @@ int amdgpu_sdma_ras_sw_init(struct amdgpu_device *adev);
 void amdgpu_debugfs_sdma_sched_mask_init(struct amdgpu_device *adev);
 int amdgpu_sdma_sysfs_reset_mask_init(struct amdgpu_device *adev);
 void amdgpu_sdma_sysfs_reset_mask_fini(struct amdgpu_device *adev);
+bool amdgpu_sdma_is_shared_inv_eng(struct amdgpu_device *adev, struct amdgpu_ring *ring);
+struct amdgpu_ring *amdgpu_sdma_get_shared_ring(struct amdgpu_device *adev,
+	struct amdgpu_ring *ring);
 #endif
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_sync.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_sync.c
index c586ab4c911b..5576ed0b508f 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_sync.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_sync.c
@@ -135,11 +135,16 @@ static bool amdgpu_sync_add_later(struct amdgpu_sync *sync, struct dma_fence *f)
 	struct amdgpu_sync_entry *e;
 
 	hash_for_each_possible(sync->fences, e, node, f->context) {
-		if (unlikely(e->fence->context != f->context))
-			continue;
+		if (dma_fence_is_signaled(e->fence)) {
+			dma_fence_put(e->fence);
+			e->fence = dma_fence_get(f);
+			return true;
+		}
 
-		amdgpu_sync_keep_later(&e->fence, f);
-		return true;
+		if (likely(e->fence->context == f->context)) {
+			amdgpu_sync_keep_later(&e->fence, f);
+			return true;
+		}
 	}
 	return false;
 }
@@ -149,10 +154,12 @@ static bool amdgpu_sync_add_later(struct amdgpu_sync *sync, struct dma_fence *f)
  *
  * @sync: sync object to add fence to
  * @f: fence to sync to
+ * @flags: memory allocation flags to use when allocating sync entry
  *
  * Add the fence to the sync object.
  */
-int amdgpu_sync_fence(struct amdgpu_sync *sync, struct dma_fence *f)
+int amdgpu_sync_fence(struct amdgpu_sync *sync, struct dma_fence *f,
+		      gfp_t flags)
 {
 	struct amdgpu_sync_entry *e;
 
@@ -162,7 +169,7 @@ int amdgpu_sync_fence(struct amdgpu_sync *sync, struct dma_fence *f)
 	if (amdgpu_sync_add_later(sync, f))
 		return 0;
 
-	e = kmem_cache_alloc(amdgpu_sync_slab, GFP_KERNEL);
+	e = kmem_cache_alloc(amdgpu_sync_slab, flags);
 	if (!e)
 		return -ENOMEM;
 
@@ -249,7 +256,7 @@ int amdgpu_sync_resv(struct amdgpu_device *adev, struct amdgpu_sync *sync,
 			struct dma_fence *tmp = dma_fence_chain_contained(f);
 
 			if (amdgpu_sync_test_fence(adev, mode, owner, tmp)) {
-				r = amdgpu_sync_fence(sync, f);
+				r = amdgpu_sync_fence(sync, f, GFP_KERNEL);
 				dma_fence_put(f);
 				if (r)
 					return r;
@@ -281,7 +288,7 @@ int amdgpu_sync_kfd(struct amdgpu_sync *sync, struct dma_resv *resv)
 		if (fence_owner != AMDGPU_FENCE_OWNER_KFD)
 			continue;
 
-		r = amdgpu_sync_fence(sync, f);
+		r = amdgpu_sync_fence(sync, f, GFP_KERNEL);
 		if (r)
 			break;
 	}
@@ -388,7 +395,7 @@ int amdgpu_sync_clone(struct amdgpu_sync *source, struct amdgpu_sync *clone)
 	hash_for_each_safe(source->fences, i, tmp, e, node) {
 		f = e->fence;
 		if (!dma_fence_is_signaled(f)) {
-			r = amdgpu_sync_fence(clone, f);
+			r = amdgpu_sync_fence(clone, f, GFP_KERNEL);
 			if (r)
 				return r;
 		} else {
@@ -400,6 +407,25 @@ int amdgpu_sync_clone(struct amdgpu_sync *source, struct amdgpu_sync *clone)
 }
 
 /**
+ * amdgpu_sync_move - move all fences from src to dst
+ *
+ * @src: source of the fences, empty after function
+ * @dst: destination for the fences
+ *
+ * Moves all fences from source to destination. All fences in destination are
+ * freed and source is empty after the function call.
+ */
+void amdgpu_sync_move(struct amdgpu_sync *src, struct amdgpu_sync *dst)
+{
+	unsigned int i;
+
+	amdgpu_sync_free(dst);
+
+	for (i = 0; i < HASH_SIZE(src->fences); ++i)
+		hlist_move_list(&src->fences[i], &dst->fences[i]);
+}
+
+/**
  * amdgpu_sync_push_to_job - push fences into job
  * @sync: sync object to get the fences from
  * @job: job to push the fences into
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_sync.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_sync.h
index e3272dce798d..51eb4382c91e 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_sync.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_sync.h
@@ -47,7 +47,8 @@ struct amdgpu_sync {
 };
 
 void amdgpu_sync_create(struct amdgpu_sync *sync);
-int amdgpu_sync_fence(struct amdgpu_sync *sync, struct dma_fence *f);
+int amdgpu_sync_fence(struct amdgpu_sync *sync, struct dma_fence *f,
+		      gfp_t flags);
 int amdgpu_sync_resv(struct amdgpu_device *adev, struct amdgpu_sync *sync,
 		     struct dma_resv *resv, enum amdgpu_sync_mode mode,
 		     void *owner);
@@ -56,6 +57,7 @@ struct dma_fence *amdgpu_sync_peek_fence(struct amdgpu_sync *sync,
 				     struct amdgpu_ring *ring);
 struct dma_fence *amdgpu_sync_get_fence(struct amdgpu_sync *sync);
 int amdgpu_sync_clone(struct amdgpu_sync *source, struct amdgpu_sync *clone);
+void amdgpu_sync_move(struct amdgpu_sync *src, struct amdgpu_sync *dst);
 int amdgpu_sync_push_to_job(struct amdgpu_sync *sync, struct amdgpu_job *job);
 int amdgpu_sync_wait(struct amdgpu_sync *sync, bool intr);
 void amdgpu_sync_free(struct amdgpu_sync *sync);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_trace.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_trace.h
index 383fce40d4dd..11dd2e0f7979 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_trace.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_trace.h
@@ -457,6 +457,38 @@ DEFINE_EVENT(amdgpu_pasid, amdgpu_pasid_freed,
 	    TP_ARGS(pasid)
 );
 
+TRACE_EVENT(amdgpu_isolation,
+	    TP_PROTO(void *prev, void *next),
+	    TP_ARGS(prev, next),
+	    TP_STRUCT__entry(
+			     __field(void *, prev)
+			     __field(void *, next)
+			     ),
+
+	    TP_fast_assign(
+			   __entry->prev = prev;
+			   __entry->next = next;
+			   ),
+	    TP_printk("prev=%p, next=%p",
+		      __entry->prev,
+		      __entry->next)
+);
+
+TRACE_EVENT(amdgpu_cleaner_shader,
+	    TP_PROTO(struct amdgpu_ring *ring, struct dma_fence *fence),
+	    TP_ARGS(ring, fence),
+	    TP_STRUCT__entry(
+			     __string(ring, ring->name)
+			     __field(u64, seqno)
+			     ),
+
+	    TP_fast_assign(
+			   __assign_str(ring);
+			   __entry->seqno = fence->seqno;
+			   ),
+	    TP_printk("ring=%s, seqno=%Lu", __get_str(ring), __entry->seqno)
+);
+
 TRACE_EVENT(amdgpu_bo_list_set,
 	    TP_PROTO(struct amdgpu_bo_list *list, struct amdgpu_bo *bo),
 	    TP_ARGS(list, bo),
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c
index 8d8b39e6d197..1991dd3d1056 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c
@@ -438,10 +438,15 @@ static void amdgpu_vcn_idle_work_handler(struct work_struct *work)
 
 	if (!fences && !atomic_read(&vcn_inst->total_submission_cnt)) {
 		vcn_inst->set_pg_state(vcn_inst, AMD_PG_STATE_GATE);
-		r = amdgpu_dpm_switch_power_profile(adev, PP_SMC_POWER_PROFILE_VIDEO,
-						    false);
-		if (r)
-			dev_warn(adev->dev, "(%d) failed to disable video power profile mode\n", r);
+		mutex_lock(&adev->vcn.workload_profile_mutex);
+		if (adev->vcn.workload_profile_active) {
+			r = amdgpu_dpm_switch_power_profile(adev, PP_SMC_POWER_PROFILE_VIDEO,
+							    false);
+			if (r)
+				dev_warn(adev->dev, "(%d) failed to disable video power profile mode\n", r);
+			adev->vcn.workload_profile_active = false;
+		}
+		mutex_unlock(&adev->vcn.workload_profile_mutex);
 	} else {
 		schedule_delayed_work(&vcn_inst->idle_work, VCN_IDLE_TIMEOUT);
 	}
@@ -455,13 +460,26 @@ void amdgpu_vcn_ring_begin_use(struct amdgpu_ring *ring)
 
 	atomic_inc(&vcn_inst->total_submission_cnt);
 
-	if (!cancel_delayed_work_sync(&vcn_inst->idle_work)) {
+	cancel_delayed_work_sync(&vcn_inst->idle_work);
+
+	/* We can safely return early here because we've cancelled the
+	 * the delayed work so there is no one else to set it to false
+	 * and we don't care if someone else sets it to true.
+	 */
+	if (adev->vcn.workload_profile_active)
+		goto pg_lock;
+
+	mutex_lock(&adev->vcn.workload_profile_mutex);
+	if (!adev->vcn.workload_profile_active) {
 		r = amdgpu_dpm_switch_power_profile(adev, PP_SMC_POWER_PROFILE_VIDEO,
-				true);
+						    true);
 		if (r)
 			dev_warn(adev->dev, "(%d) failed to switch to video power profile mode\n", r);
+		adev->vcn.workload_profile_active = true;
 	}
+	mutex_unlock(&adev->vcn.workload_profile_mutex);
 
+pg_lock:
 	mutex_lock(&vcn_inst->vcn_pg_lock);
 	vcn_inst->set_pg_state(vcn_inst, AMD_PG_STATE_UNGATE);
 
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h
index 26c9c2d90f45..cdcdae7f71ce 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h
@@ -358,6 +358,9 @@ struct amdgpu_vcn {
 
 	bool			per_inst_fw;
 	unsigned		fw_version;
+
+	bool			workload_profile_active;
+	struct mutex            workload_profile_mutex;
 };
 
 struct amdgpu_fw_shared_rb_ptrs_struct {
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c
index 1c0fd95c3820..ce52b4d75e94 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c
@@ -754,6 +754,7 @@ int amdgpu_vm_flush(struct amdgpu_ring *ring, struct amdgpu_job *job,
 		    bool need_pipe_sync)
 {
 	struct amdgpu_device *adev = ring->adev;
+	struct amdgpu_isolation *isolation = &adev->isolation[ring->xcp_id];
 	unsigned vmhub = ring->vm_hub;
 	struct amdgpu_vmid_mgr *id_mgr = &adev->vm_manager.id_mgr[vmhub];
 	struct amdgpu_vmid *id = &id_mgr->ids[job->vmid];
@@ -761,8 +762,9 @@ int amdgpu_vm_flush(struct amdgpu_ring *ring, struct amdgpu_job *job,
 	bool gds_switch_needed = ring->funcs->emit_gds_switch &&
 		job->gds_switch_needed;
 	bool vm_flush_needed = job->vm_needs_flush;
-	struct dma_fence *fence = NULL;
+	bool cleaner_shader_needed = false;
 	bool pasid_mapping_needed = false;
+	struct dma_fence *fence = NULL;
 	unsigned int patch;
 	int r;
 
@@ -785,8 +787,12 @@ int amdgpu_vm_flush(struct amdgpu_ring *ring, struct amdgpu_job *job,
 	pasid_mapping_needed &= adev->gmc.gmc_funcs->emit_pasid_mapping &&
 		ring->funcs->emit_wreg;
 
+	cleaner_shader_needed = adev->gfx.enable_cleaner_shader &&
+		ring->funcs->emit_cleaner_shader && job->base.s_fence &&
+		&job->base.s_fence->scheduled == isolation->spearhead;
+
 	if (!vm_flush_needed && !gds_switch_needed && !need_pipe_sync &&
-	    !(job->enforce_isolation && !job->vmid))
+	    !cleaner_shader_needed)
 		return 0;
 
 	amdgpu_ring_ib_begin(ring);
@@ -797,9 +803,7 @@ int amdgpu_vm_flush(struct amdgpu_ring *ring, struct amdgpu_job *job,
 	if (need_pipe_sync)
 		amdgpu_ring_emit_pipeline_sync(ring);
 
-	if (adev->gfx.enable_cleaner_shader &&
-	    ring->funcs->emit_cleaner_shader &&
-	    job->enforce_isolation)
+	if (cleaner_shader_needed)
 		ring->funcs->emit_cleaner_shader(ring);
 
 	if (vm_flush_needed) {
@@ -821,7 +825,7 @@ int amdgpu_vm_flush(struct amdgpu_ring *ring, struct amdgpu_job *job,
 					    job->oa_size);
 	}
 
-	if (vm_flush_needed || pasid_mapping_needed) {
+	if (vm_flush_needed || pasid_mapping_needed || cleaner_shader_needed) {
 		r = amdgpu_fence_emit(ring, &fence, NULL, 0);
 		if (r)
 			return r;
@@ -843,6 +847,18 @@ int amdgpu_vm_flush(struct amdgpu_ring *ring, struct amdgpu_job *job,
 		id->pasid_mapping = dma_fence_get(fence);
 		mutex_unlock(&id_mgr->lock);
 	}
+
+	/*
+	 * Make sure that all other submissions wait for the cleaner shader to
+	 * finish before we push them to the HW.
+	 */
+	if (cleaner_shader_needed) {
+		trace_amdgpu_cleaner_shader(ring, fence);
+		mutex_lock(&adev->enforce_isolation_mutex);
+		dma_fence_put(isolation->spearhead);
+		isolation->spearhead = dma_fence_get(fence);
+		mutex_unlock(&adev->enforce_isolation_mutex);
+	}
 	dma_fence_put(fence);
 
 	amdgpu_ring_patch_cond_exec(ring, patch);
diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v11_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v11_0.c
index b9f6d89dafb2..d8772cd6db63 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v11_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v11_0.c
@@ -1626,6 +1626,20 @@ static int gfx_v11_0_sw_init(struct amdgpu_ip_block *ip_block)
 			}
 		}
 		break;
+	case IP_VERSION(11, 5, 0):
+	case IP_VERSION(11, 5, 1):
+		adev->gfx.cleaner_shader_ptr = gfx_11_0_3_cleaner_shader_hex;
+		adev->gfx.cleaner_shader_size = sizeof(gfx_11_0_3_cleaner_shader_hex);
+		if (adev->gfx.mec_fw_version >= 26 &&
+		    adev->mes.fw_version[0] >= 114) {
+			adev->gfx.enable_cleaner_shader = true;
+			r = amdgpu_gfx_cleaner_shader_sw_init(adev, adev->gfx.cleaner_shader_size);
+			if (r) {
+				adev->gfx.enable_cleaner_shader = false;
+				dev_err(adev->dev, "Failed to initialize cleaner shader\n");
+			}
+		}
+		break;
 	default:
 		adev->gfx.enable_cleaner_shader = false;
 		break;
diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v12_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v12_0.c
index da364e04f09c..dceb5ad38862 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v12_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v12_0.c
@@ -2637,7 +2637,6 @@ static int gfx_v12_0_cp_gfx_resume(struct amdgpu_device *adev)
 	u32 tmp;
 	u32 rb_bufsz;
 	u64 rb_addr, rptr_addr, wptr_gpu_addr;
-	u32 i;
 
 	/* Set the write pointer delay */
 	WREG32_SOC15(GC, 0, regCP_RB_WPTR_DELAY, 0);
@@ -2692,12 +2691,6 @@ static int gfx_v12_0_cp_gfx_resume(struct amdgpu_device *adev)
 
 	/* start the ring */
 	gfx_v12_0_cp_gfx_start(adev);
-
-	for (i = 0; i < adev->gfx.num_gfx_rings; i++) {
-		ring = &adev->gfx.gfx_ring[i];
-		ring->sched.ready = true;
-	}
-
 	return 0;
 }
 
@@ -3037,10 +3030,6 @@ static int gfx_v12_0_cp_async_gfx_ring_resume(struct amdgpu_device *adev)
 	if (r)
 		goto done;
 
-	for (i = 0; i < adev->gfx.num_gfx_rings; i++) {
-		ring = &adev->gfx.gfx_ring[i];
-		ring->sched.ready = true;
-	}
 done:
 	return r;
 }
diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v9_4_3.c b/drivers/gpu/drm/amd/amdgpu/gfx_v9_4_3.c
index efe45e4edfd7..736398b0d16d 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v9_4_3.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v9_4_3.c
@@ -867,9 +867,8 @@ static int gfx_v9_4_3_aca_bank_parser(struct aca_handle *handle,
 
 	switch (type) {
 	case ACA_SMU_TYPE_UE:
-		bank->aca_err_type = ACA_ERROR_TYPE_UE;
-		ret = aca_error_cache_log_bank_error(handle, &info,
-						     ACA_ERROR_TYPE_UE, 1ULL);
+		bank->aca_err_type = ACA_BANK_ERR_UE_DE_DECODE(bank);
+		ret = aca_error_cache_log_bank_error(handle, &info, bank->aca_err_type, 1ULL);
 		break;
 	case ACA_SMU_TYPE_CE:
 		bank->aca_err_type = ACA_BANK_ERR_CE_DE_DECODE(bank);
diff --git a/drivers/gpu/drm/amd/amdgpu/sdma_v4_4_2.c b/drivers/gpu/drm/amd/amdgpu/sdma_v4_4_2.c
index fd34dc138081..dc94d58d33a6 100644
--- a/drivers/gpu/drm/amd/amdgpu/sdma_v4_4_2.c
+++ b/drivers/gpu/drm/amd/amdgpu/sdma_v4_4_2.c
@@ -31,6 +31,7 @@
 #include "amdgpu_ucode.h"
 #include "amdgpu_trace.h"
 #include "amdgpu_reset.h"
+#include "gc/gc_9_0_sh_mask.h"
 
 #include "sdma/sdma_4_4_2_offset.h"
 #include "sdma/sdma_4_4_2_sh_mask.h"
@@ -672,12 +673,11 @@ static uint32_t sdma_v4_4_2_rb_cntl(struct amdgpu_ring *ring, uint32_t rb_cntl)
  * @adev: amdgpu_device pointer
  * @i: instance to resume
  * @restore: used to restore wptr when restart
- * @guilty: boolean indicating whether this queue is the guilty one (caused the timeout/error)
  *
  * Set up the gfx DMA ring buffers and enable them.
  * Returns 0 for success, error for failure.
  */
-static void sdma_v4_4_2_gfx_resume(struct amdgpu_device *adev, unsigned int i, bool restore, bool guilty)
+static void sdma_v4_4_2_gfx_resume(struct amdgpu_device *adev, unsigned int i, bool restore)
 {
 	struct amdgpu_ring *ring = &adev->sdma.instance[i].ring;
 	u32 rb_cntl, ib_cntl, wptr_poll_cntl;
@@ -714,7 +714,7 @@ static void sdma_v4_4_2_gfx_resume(struct amdgpu_device *adev, unsigned int i, b
 	/* For the guilty queue, set RPTR to the current wptr to skip bad commands,
 	 * It is not a guilty queue, restore cache_rptr and continue execution.
 	 */
-	if (guilty)
+	if (adev->sdma.instance[i].gfx_guilty)
 		rwptr = ring->wptr;
 	else
 		rwptr = ring->cached_rptr;
@@ -779,12 +779,11 @@ static void sdma_v4_4_2_gfx_resume(struct amdgpu_device *adev, unsigned int i, b
  * @adev: amdgpu_device pointer
  * @i: instance to resume
  * @restore: boolean to say restore needed or not
- * @guilty: boolean indicating whether this queue is the guilty one (caused the timeout/error)
  *
  * Set up the page DMA ring buffers and enable them.
  * Returns 0 for success, error for failure.
  */
-static void sdma_v4_4_2_page_resume(struct amdgpu_device *adev, unsigned int i, bool restore, bool guilty)
+static void sdma_v4_4_2_page_resume(struct amdgpu_device *adev, unsigned int i, bool restore)
 {
 	struct amdgpu_ring *ring = &adev->sdma.instance[i].page;
 	u32 rb_cntl, ib_cntl, wptr_poll_cntl;
@@ -803,7 +802,7 @@ static void sdma_v4_4_2_page_resume(struct amdgpu_device *adev, unsigned int i,
 	/* For the guilty queue, set RPTR to the current wptr to skip bad commands,
 	 * It is not a guilty queue, restore cache_rptr and continue execution.
 	 */
-	if (guilty)
+	if (adev->sdma.instance[i].page_guilty)
 		rwptr = ring->wptr;
 	else
 		rwptr = ring->cached_rptr;
@@ -989,9 +988,9 @@ static int sdma_v4_4_2_inst_start(struct amdgpu_device *adev,
 		uint32_t temp;
 
 		WREG32_SDMA(i, regSDMA_SEM_WAIT_FAIL_TIMER_CNTL, 0);
-		sdma_v4_4_2_gfx_resume(adev, i, restore, adev->sdma.gfx_guilty);
+		sdma_v4_4_2_gfx_resume(adev, i, restore);
 		if (adev->sdma.has_page_queue)
-			sdma_v4_4_2_page_resume(adev, i, restore, adev->sdma.page_guilty);
+			sdma_v4_4_2_page_resume(adev, i, restore);
 
 		/* set utc l1 enable flag always to 1 */
 		temp = RREG32_SDMA(i, regSDMA_CNTL);
@@ -1292,21 +1291,71 @@ static void sdma_v4_4_2_ring_emit_pipeline_sync(struct amdgpu_ring *ring)
 			       seq, 0xffffffff, 4);
 }
 
-
-/**
- * sdma_v4_4_2_ring_emit_vm_flush - vm flush using sDMA
+/*
+ * sdma_v4_4_2_get_invalidate_req - Construct the VM_INVALIDATE_ENG0_REQ register value
+ * @vmid: The VMID to invalidate
+ * @flush_type: The type of flush (0 = legacy, 1 = lightweight, 2 = heavyweight)
  *
- * @ring: amdgpu_ring pointer
- * @vmid: vmid number to use
- * @pd_addr: address
+ * This function constructs the VM_INVALIDATE_ENG0_REQ register value for the specified VMID
+ * and flush type. It ensures that all relevant page table cache levels (L1 PTEs, L2 PTEs, and
+ * L2 PDEs) are invalidated.
+ */
+static uint32_t sdma_v4_4_2_get_invalidate_req(unsigned int vmid,
+						uint32_t flush_type)
+{
+	u32 req = 0;
+
+	req = REG_SET_FIELD(req, VM_INVALIDATE_ENG0_REQ,
+			    PER_VMID_INVALIDATE_REQ, 1 << vmid);
+	req = REG_SET_FIELD(req, VM_INVALIDATE_ENG0_REQ, FLUSH_TYPE, flush_type);
+	req = REG_SET_FIELD(req, VM_INVALIDATE_ENG0_REQ, INVALIDATE_L2_PTES, 1);
+	req = REG_SET_FIELD(req, VM_INVALIDATE_ENG0_REQ, INVALIDATE_L2_PDE0, 1);
+	req = REG_SET_FIELD(req, VM_INVALIDATE_ENG0_REQ, INVALIDATE_L2_PDE1, 1);
+	req = REG_SET_FIELD(req, VM_INVALIDATE_ENG0_REQ, INVALIDATE_L2_PDE2, 1);
+	req = REG_SET_FIELD(req, VM_INVALIDATE_ENG0_REQ, INVALIDATE_L1_PTES, 1);
+	req = REG_SET_FIELD(req, VM_INVALIDATE_ENG0_REQ,
+			    CLEAR_PROTECTION_FAULT_STATUS_ADDR,	0);
+
+	return req;
+}
+
+/*
+ * sdma_v4_4_2_ring_emit_vm_flush - Emit VM flush commands for SDMA
+ * @ring: The SDMA ring
+ * @vmid: The VMID to flush
+ * @pd_addr: The page directory address
  *
- * Update the page table base and flush the VM TLB
- * using sDMA.
+ * This function emits the necessary register writes and waits to perform a VM flush for the
+ * specified VMID. It updates the PTB address registers and issues a VM invalidation request
+ * using the specified VM invalidation engine.
  */
 static void sdma_v4_4_2_ring_emit_vm_flush(struct amdgpu_ring *ring,
-					 unsigned vmid, uint64_t pd_addr)
+					    unsigned int vmid, uint64_t pd_addr)
 {
-	amdgpu_gmc_emit_flush_gpu_tlb(ring, vmid, pd_addr);
+	struct amdgpu_device *adev = ring->adev;
+	uint32_t req = sdma_v4_4_2_get_invalidate_req(vmid, 0);
+	unsigned int eng = ring->vm_inv_eng;
+	struct amdgpu_vmhub *hub = &adev->vmhub[ring->vm_hub];
+
+	amdgpu_ring_emit_wreg(ring, hub->ctx0_ptb_addr_lo32 +
+				(hub->ctx_addr_distance * vmid),
+				lower_32_bits(pd_addr));
+
+	amdgpu_ring_emit_wreg(ring, hub->ctx0_ptb_addr_hi32 +
+				(hub->ctx_addr_distance * vmid),
+				upper_32_bits(pd_addr));
+	/*
+	 * Construct and emit the VM invalidation packet
+	 */
+	amdgpu_ring_write(ring,
+		SDMA_PKT_VM_INVALIDATION_HEADER_OP(SDMA_OP_VM_INVALIDATE) |
+		SDMA_PKT_VM_INVALIDATION_HEADER_SUB_OP(SDMA_SUBOP_VM_INVALIDATE) |
+		SDMA_PKT_VM_INVALIDATION_HEADER_XCC0_ENG_ID(0x1f) |
+		SDMA_PKT_VM_INVALIDATION_HEADER_XCC1_ENG_ID(0x1f) |
+		SDMA_PKT_VM_INVALIDATION_HEADER_MMHUB_ENG_ID(eng));
+	amdgpu_ring_write(ring, SDMA_PKT_VM_INVALIDATION_INVALIDATEREQ_INVALIDATEREQ(req));
+	amdgpu_ring_write(ring, 0);
+	amdgpu_ring_write(ring, SDMA_PKT_VM_INVALIDATION_ADDRESSRANGEHI_INVALIDATEACK(BIT(vmid)));
 }
 
 static void sdma_v4_4_2_ring_emit_wreg(struct amdgpu_ring *ring,
@@ -1445,6 +1494,11 @@ static int sdma_v4_4_2_sw_init(struct amdgpu_ip_block *ip_block)
 	}
 
 	for (i = 0; i < adev->sdma.num_instances; i++) {
+		mutex_init(&adev->sdma.instance[i].engine_reset_mutex);
+		/* Initialize guilty flags for GFX and PAGE queues */
+		adev->sdma.instance[i].gfx_guilty = false;
+		adev->sdma.instance[i].page_guilty = false;
+
 		ring = &adev->sdma.instance[i].ring;
 		ring->ring_obj = NULL;
 		ring->use_doorbell = true;
@@ -1506,9 +1560,6 @@ static int sdma_v4_4_2_sw_init(struct amdgpu_ip_block *ip_block)
 	r = amdgpu_sdma_sysfs_reset_mask_init(adev);
 	if (r)
 		return r;
-	/* Initialize guilty flags for GFX and PAGE queues */
-	adev->sdma.gfx_guilty = false;
-	adev->sdma.page_guilty = false;
 
 	return r;
 }
@@ -1666,7 +1717,16 @@ static int sdma_v4_4_2_reset_queue(struct amdgpu_ring *ring, unsigned int vmid)
 {
 	struct amdgpu_device *adev = ring->adev;
 	u32 id = GET_INST(SDMA0, ring->me);
-	return amdgpu_sdma_reset_engine(adev, id, true);
+	int r;
+
+	if (!(adev->sdma.supported_reset & AMDGPU_RESET_TYPE_PER_QUEUE))
+		return -EOPNOTSUPP;
+
+	amdgpu_amdkfd_suspend(adev, false);
+	r = amdgpu_sdma_reset_engine(adev, id);
+	amdgpu_amdkfd_resume(adev, false);
+
+	return r;
 }
 
 static int sdma_v4_4_2_stop_queue(struct amdgpu_device *adev, uint32_t instance_id)
@@ -1679,9 +1739,11 @@ static int sdma_v4_4_2_stop_queue(struct amdgpu_device *adev, uint32_t instance_
 		return -EINVAL;
 
 	/* Check if this queue is the guilty one */
-	adev->sdma.gfx_guilty = sdma_v4_4_2_is_queue_selected(adev, instance_id, false);
+	adev->sdma.instance[instance_id].gfx_guilty =
+		sdma_v4_4_2_is_queue_selected(adev, instance_id, false);
 	if (adev->sdma.has_page_queue)
-		adev->sdma.page_guilty = sdma_v4_4_2_is_queue_selected(adev, instance_id, true);
+		adev->sdma.instance[instance_id].page_guilty =
+			sdma_v4_4_2_is_queue_selected(adev, instance_id, true);
 
 	/* Cache the rptr before reset, after the reset,
 	* all of the registers will be reset to 0
@@ -2115,8 +2177,7 @@ static const struct amdgpu_ring_funcs sdma_v4_4_2_ring_funcs = {
 		3 + /* hdp invalidate */
 		6 + /* sdma_v4_4_2_ring_emit_pipeline_sync */
 		/* sdma_v4_4_2_ring_emit_vm_flush */
-		SOC15_FLUSH_GPU_TLB_NUM_WREG * 3 +
-		SOC15_FLUSH_GPU_TLB_NUM_REG_WAIT * 6 +
+		4 + 2 * 3 +
 		10 + 10 + 10, /* sdma_v4_4_2_ring_emit_fence x3 for user fence, vm fence */
 	.emit_ib_size = 7 + 6, /* sdma_v4_4_2_ring_emit_ib */
 	.emit_ib = sdma_v4_4_2_ring_emit_ib,
@@ -2148,8 +2209,7 @@ static const struct amdgpu_ring_funcs sdma_v4_4_2_page_ring_funcs = {
 		3 + /* hdp invalidate */
 		6 + /* sdma_v4_4_2_ring_emit_pipeline_sync */
 		/* sdma_v4_4_2_ring_emit_vm_flush */
-		SOC15_FLUSH_GPU_TLB_NUM_WREG * 3 +
-		SOC15_FLUSH_GPU_TLB_NUM_REG_WAIT * 6 +
+		4 + 2 * 3 +
 		10 + 10 + 10, /* sdma_v4_4_2_ring_emit_fence x3 for user fence, vm fence */
 	.emit_ib_size = 7 + 6, /* sdma_v4_4_2_ring_emit_ib */
 	.emit_ib = sdma_v4_4_2_ring_emit_ib,
@@ -2347,6 +2407,9 @@ static void sdma_v4_4_2_set_vm_pte_funcs(struct amdgpu_device *adev)
  */
 static void sdma_v4_4_2_update_reset_mask(struct amdgpu_device *adev)
 {
+	/* per queue reset not supported for SRIOV */
+	if (amdgpu_sriov_vf(adev))
+		return;
 
 	/*
 	 * the user queue relies on MEC fw and pmfw when the sdma queue do reset.
diff --git a/drivers/gpu/drm/amd/amdgpu/ta_ras_if.h b/drivers/gpu/drm/amd/amdgpu/ta_ras_if.h
index 64891f099366..a3b5fda22432 100644
--- a/drivers/gpu/drm/amd/amdgpu/ta_ras_if.h
+++ b/drivers/gpu/drm/amd/amdgpu/ta_ras_if.h
@@ -151,6 +151,7 @@ struct ta_ras_init_flags {
 	uint16_t xcc_mask;
 	uint8_t channel_dis_num;
 	uint8_t nps_mode;
+	uint32_t active_umc_mask;
 };
 
 struct ta_ras_mca_addr {
diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c b/drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c
index ff03436698a4..d716510b8dd6 100644
--- a/drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c
+++ b/drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c
@@ -147,10 +147,15 @@ static void vcn_v2_5_idle_work_handler(struct work_struct *work)
 	if (!fences && !atomic_read(&adev->vcn.inst[0].total_submission_cnt)) {
 		amdgpu_device_ip_set_powergating_state(adev, AMD_IP_BLOCK_TYPE_VCN,
 						       AMD_PG_STATE_GATE);
-		r = amdgpu_dpm_switch_power_profile(adev, PP_SMC_POWER_PROFILE_VIDEO,
-						    false);
-		if (r)
-			dev_warn(adev->dev, "(%d) failed to disable video power profile mode\n", r);
+		mutex_lock(&adev->vcn.workload_profile_mutex);
+		if (adev->vcn.workload_profile_active) {
+			r = amdgpu_dpm_switch_power_profile(adev, PP_SMC_POWER_PROFILE_VIDEO,
+							    false);
+			if (r)
+				dev_warn(adev->dev, "(%d) failed to disable video power profile mode\n", r);
+			adev->vcn.workload_profile_active = false;
+		}
+		mutex_unlock(&adev->vcn.workload_profile_mutex);
 	} else {
 		schedule_delayed_work(&adev->vcn.inst[0].idle_work, VCN_IDLE_TIMEOUT);
 	}
@@ -164,13 +169,26 @@ static void vcn_v2_5_ring_begin_use(struct amdgpu_ring *ring)
 
 	atomic_inc(&adev->vcn.inst[0].total_submission_cnt);
 
-	if (!cancel_delayed_work_sync(&adev->vcn.inst[0].idle_work)) {
+	cancel_delayed_work_sync(&adev->vcn.inst[0].idle_work);
+
+	/* We can safely return early here because we've cancelled the
+	 * the delayed work so there is no one else to set it to false
+	 * and we don't care if someone else sets it to true.
+	 */
+	if (adev->vcn.workload_profile_active)
+		goto pg_lock;
+
+	mutex_lock(&adev->vcn.workload_profile_mutex);
+	if (!adev->vcn.workload_profile_active) {
 		r = amdgpu_dpm_switch_power_profile(adev, PP_SMC_POWER_PROFILE_VIDEO,
 						    true);
 		if (r)
 			dev_warn(adev->dev, "(%d) failed to switch to video power profile mode\n", r);
+		adev->vcn.workload_profile_active = true;
 	}
+	mutex_unlock(&adev->vcn.workload_profile_mutex);
 
+pg_lock:
 	mutex_lock(&adev->vcn.inst[0].vcn_pg_lock);
 	amdgpu_device_ip_set_powergating_state(adev, AMD_IP_BLOCK_TYPE_VCN,
 					       AMD_PG_STATE_UNGATE);
diff --git a/drivers/gpu/drm/amd/amdgpu/vega10_sdma_pkt_open.h b/drivers/gpu/drm/amd/amdgpu/vega10_sdma_pkt_open.h
index 8de4ccce5e38..3ca8a417c6d8 100644
--- a/drivers/gpu/drm/amd/amdgpu/vega10_sdma_pkt_open.h
+++ b/drivers/gpu/drm/amd/amdgpu/vega10_sdma_pkt_open.h
@@ -64,6 +64,9 @@
 #define HEADER_BARRIER  5
 #define SDMA_OP_AQL_COPY  0
 #define SDMA_OP_AQL_BARRIER_OR  0
+/* vm invalidation is only available for GC9.4.3/GC9.4.4/GC9.5.0 */
+#define SDMA_OP_VM_INVALIDATE 8
+#define SDMA_SUBOP_VM_INVALIDATE 4
 
 /*define for op field*/
 #define SDMA_PKT_HEADER_op_offset 0
@@ -3331,5 +3334,72 @@
 #define SDMA_AQL_PKT_BARRIER_OR_COMPLETION_SIGNAL_HI_completion_signal_63_32_shift  0
 #define SDMA_AQL_PKT_BARRIER_OR_COMPLETION_SIGNAL_HI_COMPLETION_SIGNAL_63_32(x) (((x) & SDMA_AQL_PKT_BARRIER_OR_COMPLETION_SIGNAL_HI_completion_signal_63_32_mask) << SDMA_AQL_PKT_BARRIER_OR_COMPLETION_SIGNAL_HI_completion_signal_63_32_shift)
 
+/*
+** Definitions for SDMA_PKT_VM_INVALIDATION packet
+*/
+
+/*define for HEADER word*/
+/*define for op field*/
+#define SDMA_PKT_VM_INVALIDATION_HEADER_op_offset 0
+#define SDMA_PKT_VM_INVALIDATION_HEADER_op_mask   0x000000FF
+#define SDMA_PKT_VM_INVALIDATION_HEADER_op_shift  0
+#define SDMA_PKT_VM_INVALIDATION_HEADER_OP(x) ((x & SDMA_PKT_VM_INVALIDATION_HEADER_op_mask) << SDMA_PKT_VM_INVALIDATION_HEADER_op_shift)
+
+/*define for sub_op field*/
+#define SDMA_PKT_VM_INVALIDATION_HEADER_sub_op_offset 0
+#define SDMA_PKT_VM_INVALIDATION_HEADER_sub_op_mask   0x000000FF
+#define SDMA_PKT_VM_INVALIDATION_HEADER_sub_op_shift  8
+#define SDMA_PKT_VM_INVALIDATION_HEADER_SUB_OP(x) ((x & SDMA_PKT_VM_INVALIDATION_HEADER_sub_op_mask) << SDMA_PKT_VM_INVALIDATION_HEADER_sub_op_shift)
+
+/*define for xcc0_eng_id field*/
+#define SDMA_PKT_VM_INVALIDATION_HEADER_xcc0_eng_id_offset 0
+#define SDMA_PKT_VM_INVALIDATION_HEADER_xcc0_eng_id_mask   0x0000001F
+#define SDMA_PKT_VM_INVALIDATION_HEADER_xcc0_eng_id_shift  16
+#define SDMA_PKT_VM_INVALIDATION_HEADER_XCC0_ENG_ID(x) ((x & SDMA_PKT_VM_INVALIDATION_HEADER_xcc0_eng_id_mask) << SDMA_PKT_VM_INVALIDATION_HEADER_xcc0_eng_id_shift)
+
+/*define for xcc1_eng_id field*/
+#define SDMA_PKT_VM_INVALIDATION_HEADER_xcc1_eng_id_offset 0
+#define SDMA_PKT_VM_INVALIDATION_HEADER_xcc1_eng_id_mask   0x0000001F
+#define SDMA_PKT_VM_INVALIDATION_HEADER_xcc1_eng_id_shift  21
+#define SDMA_PKT_VM_INVALIDATION_HEADER_XCC1_ENG_ID(x) ((x & SDMA_PKT_VM_INVALIDATION_HEADER_xcc1_eng_id_mask) << SDMA_PKT_VM_INVALIDATION_HEADER_xcc1_eng_id_shift)
+
+/*define for mmhub_eng_id field*/
+#define SDMA_PKT_VM_INVALIDATION_HEADER_mmhub_eng_id_offset 0
+#define SDMA_PKT_VM_INVALIDATION_HEADER_mmhub_eng_id_mask   0x0000001F
+#define SDMA_PKT_VM_INVALIDATION_HEADER_mmhub_eng_id_shift  26
+#define SDMA_PKT_VM_INVALIDATION_HEADER_MMHUB_ENG_ID(x) ((x & SDMA_PKT_VM_INVALIDATION_HEADER_mmhub_eng_id_mask) << SDMA_PKT_VM_INVALIDATION_HEADER_mmhub_eng_id_shift)
+
+/*define for INVALIDATEREQ word*/
+/*define for invalidatereq field*/
+#define SDMA_PKT_VM_INVALIDATION_INVALIDATEREQ_invalidatereq_offset 1
+#define SDMA_PKT_VM_INVALIDATION_INVALIDATEREQ_invalidatereq_mask   0xFFFFFFFF
+#define SDMA_PKT_VM_INVALIDATION_INVALIDATEREQ_invalidatereq_shift  0
+#define SDMA_PKT_VM_INVALIDATION_INVALIDATEREQ_INVALIDATEREQ(x) ((x & SDMA_PKT_VM_INVALIDATION_INVALIDATEREQ_invalidatereq_mask) << SDMA_PKT_VM_INVALIDATION_INVALIDATEREQ_invalidatereq_shift)
+
+/*define for ADDRESSRANGELO word*/
+/*define for addressrangelo field*/
+#define SDMA_PKT_VM_INVALIDATION_ADDRESSRANGELO_addressrangelo_offset 2
+#define SDMA_PKT_VM_INVALIDATION_ADDRESSRANGELO_addressrangelo_mask   0xFFFFFFFF
+#define SDMA_PKT_VM_INVALIDATION_ADDRESSRANGELO_addressrangelo_shift  0
+#define SDMA_PKT_VM_INVALIDATION_ADDRESSRANGELO_ADDRESSRANGELO(x) ((x & SDMA_PKT_VM_INVALIDATION_ADDRESSRANGELO_addressrangelo_mask) << SDMA_PKT_VM_INVALIDATION_ADDRESSRANGELO_addressrangelo_shift)
+
+/*define for ADDRESSRANGEHI word*/
+/*define for invalidateack field*/
+#define SDMA_PKT_VM_INVALIDATION_ADDRESSRANGEHI_invalidateack_offset 3
+#define SDMA_PKT_VM_INVALIDATION_ADDRESSRANGEHI_invalidateack_mask   0x0000FFFF
+#define SDMA_PKT_VM_INVALIDATION_ADDRESSRANGEHI_invalidateack_shift  0
+#define SDMA_PKT_VM_INVALIDATION_ADDRESSRANGEHI_INVALIDATEACK(x) ((x & SDMA_PKT_VM_INVALIDATION_ADDRESSRANGEHI_invalidateack_mask) << SDMA_PKT_VM_INVALIDATION_ADDRESSRANGEHI_invalidateack_shift)
+
+/*define for addressrangehi field*/
+#define SDMA_PKT_VM_INVALIDATION_ADDRESSRANGEHI_addressrangehi_offset 3
+#define SDMA_PKT_VM_INVALIDATION_ADDRESSRANGEHI_addressrangehi_mask   0x0000001F
+#define SDMA_PKT_VM_INVALIDATION_ADDRESSRANGEHI_addressrangehi_shift  16
+#define SDMA_PKT_VM_INVALIDATION_ADDRESSRANGEHI_ADDRESSRANGEHI(x) ((x & SDMA_PKT_VM_INVALIDATION_ADDRESSRANGEHI_addressrangehi_mask) << SDMA_PKT_VM_INVALIDATION_ADDRESSRANGEHI_addressrangehi_shift)
+
+/*define for reserved field*/
+#define SDMA_PKT_VM_INVALIDATION_ADDRESSRANGEHI_reserved_offset 3
+#define SDMA_PKT_VM_INVALIDATION_ADDRESSRANGEHI_reserved_mask   0x000001FF
+#define SDMA_PKT_VM_INVALIDATION_ADDRESSRANGEHI_reserved_shift  23
+#define SDMA_PKT_VM_INVALIDATION_ADDRESSRANGEHI_RESERVED(x) ((x & SDMA_PKT_VM_INVALIDATION_ADDRESSRANGEHI_reserved_mask) << SDMA_PKT_VM_INVALIDATION_ADDRESSRANGEHI_reserved_shift)
 
 #endif /* __SDMA_PKT_OPEN_H_ */
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.c b/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.c
index 2ed003d3ff0e..c610e172a2b8 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.c
@@ -2310,7 +2310,7 @@ static int reset_hung_queues_sdma(struct device_queue_manager *dqm)
 				continue;
 
 			/* Reset engine and check. */
-			if (amdgpu_sdma_reset_engine(dqm->dev->adev, i, false) ||
+			if (amdgpu_sdma_reset_engine(dqm->dev->adev, i) ||
 			    dqm->dev->kfd2kgd->hqd_sdma_get_doorbell(dqm->dev->adev, i, j) ||
 			    !set_sdma_queue_as_reset(dqm, doorbell_off)) {
 				r = -ENOTRECOVERABLE;
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_packet_manager.c b/drivers/gpu/drm/amd/amdkfd/kfd_packet_manager.c
index 3f574d82b5fc..271c567242ab 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_packet_manager.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_packet_manager.c
@@ -418,6 +418,10 @@ int pm_config_dequeue_wait_counts(struct packet_manager *pm,
 	    !pm->pmf->config_dequeue_wait_counts_size)
 		return 0;
 
+	if (cmd == KFD_DEQUEUE_WAIT_INIT && (KFD_GC_VERSION(pm->dqm->dev) < IP_VERSION(9, 4, 1) ||
+	   KFD_GC_VERSION(pm->dqm->dev) >= IP_VERSION(10, 0, 0)))
+		return 0;
+
 	size = pm->pmf->config_dequeue_wait_counts_size;
 
 	mutex_lock(&pm->lock);
@@ -436,16 +440,16 @@ int pm_config_dequeue_wait_counts(struct packet_manager *pm,
 
 		retval = pm->pmf->config_dequeue_wait_counts(pm, buffer,
 							     cmd, value);
-		if (!retval)
+		if (!retval) {
 			retval = kq_submit_packet(pm->priv_queue);
-		else
+
+			/* If default value is modified, cache that in dqm->wait_times */
+			if (!retval && cmd == KFD_DEQUEUE_WAIT_INIT)
+				update_dqm_wait_times(pm->dqm);
+		} else {
 			kq_rollback_packet(pm->priv_queue);
+		}
 	}
-
-	/* If default value is modified, cache that value in dqm->wait_times */
-	if (!retval && cmd == KFD_DEQUEUE_WAIT_INIT)
-		update_dqm_wait_times(pm->dqm);
-
 out:
 	mutex_unlock(&pm->lock);
 	return retval;
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_packet_manager_v9.c b/drivers/gpu/drm/amd/amdkfd/kfd_packet_manager_v9.c
index d440df602393..2893fd5e5d00 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_packet_manager_v9.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_packet_manager_v9.c
@@ -310,6 +310,13 @@ static inline void pm_build_dequeue_wait_counts_packet_info(struct packet_manage
 		reg_data);
 }
 
+/* pm_config_dequeue_wait_counts_v9: Builds WRITE_DATA packet with
+ *    register/value for configuring dequeue wait counts
+ *
+ * @return: -ve for failure and 0 for success and buffer is
+ *  filled in with packet
+ *
+ **/
 static int pm_config_dequeue_wait_counts_v9(struct packet_manager *pm,
 		uint32_t *buffer,
 		enum kfd_config_dequeue_wait_counts_cmd cmd,
@@ -321,24 +328,25 @@ static int pm_config_dequeue_wait_counts_v9(struct packet_manager *pm,
 
 	switch (cmd) {
 	case KFD_DEQUEUE_WAIT_INIT: {
-		uint32_t sch_wave = 0, que_sleep = 0;
-		/* Reduce CP_IQ_WAIT_TIME2.QUE_SLEEP to 0x1 from default 0x40.
+		uint32_t sch_wave = 0, que_sleep = 1;
+
+		/* For all gfx9 ASICs > gfx941,
+		 * Reduce CP_IQ_WAIT_TIME2.QUE_SLEEP to 0x1 from default 0x40.
 		 * On a 1GHz machine this is roughly 1 microsecond, which is
 		 * about how long it takes to load data out of memory during
 		 * queue connect
 		 * QUE_SLEEP: Wait Count for Dequeue Retry.
+		 *
+		 * Set CWSR grace period to 1x1000 cycle for GFX9.4.3 APU
 		 */
-		if (KFD_GC_VERSION(pm->dqm->dev) >= IP_VERSION(9, 4, 1) &&
-		    KFD_GC_VERSION(pm->dqm->dev) < IP_VERSION(10, 0, 0)) {
-			que_sleep = 1;
-
-			/* Set CWSR grace period to 1x1000 cycle for GFX9.4.3 APU */
-			if (amdgpu_emu_mode == 0 && pm->dqm->dev->adev->gmc.is_app_apu &&
-			(KFD_GC_VERSION(pm->dqm->dev) == IP_VERSION(9, 4, 3)))
-				sch_wave = 1;
-		} else {
-			return 0;
-		}
+		if (KFD_GC_VERSION(pm->dqm->dev) < IP_VERSION(9, 4, 1) ||
+		    KFD_GC_VERSION(pm->dqm->dev) >= IP_VERSION(10, 0, 0))
+			return -EPERM;
+
+		if (amdgpu_emu_mode == 0 && pm->dqm->dev->adev->gmc.is_app_apu &&
+		    (KFD_GC_VERSION(pm->dqm->dev) == IP_VERSION(9, 4, 3)))
+			sch_wave = 1;
+
 		pm_build_dequeue_wait_counts_packet_info(pm, sch_wave, que_sleep,
 			&reg_offset, &reg_data);
 
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_topology.c b/drivers/gpu/drm/amd/amdkfd/kfd_topology.c
index 27e7356eed6f..e477d7509646 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_topology.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_topology.c
@@ -2006,10 +2006,6 @@ static void kfd_topology_set_capabilities(struct kfd_topology_device *dev)
 		dev->node_props.debug_prop |= HSA_DBG_WATCH_ADDR_MASK_LO_BIT_GFX10 |
 					HSA_DBG_WATCH_ADDR_MASK_HI_BIT;
 
-		if (KFD_GC_VERSION(dev->gpu) >= IP_VERSION(11, 0, 0))
-			dev->node_props.capability |=
-				HSA_CAP_TRAP_DEBUG_PRECISE_MEMORY_OPERATIONS_SUPPORTED;
-
 		if (KFD_GC_VERSION(dev->gpu) >= IP_VERSION(12, 0, 0))
 			dev->node_props.capability |=
 				HSA_CAP_TRAP_DEBUG_PRECISE_ALU_OPERATIONS_SUPPORTED;
diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c
index 4f1cf8afe25f..bae83a129b5f 100644
--- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c
+++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c
@@ -1752,7 +1752,7 @@ static void retrieve_dmi_info(struct amdgpu_display_manager *dm, struct dc_init_
 	}
 	if (quirk_entries.support_edp0_on_dp1) {
 		init_data->flags.support_edp0_on_dp1 = true;
-		drm_info(dev, "aux_hpd_discon_quirk attached\n");
+		drm_info(dev, "support_edp0_on_dp1 attached\n");
 	}
 }
 
diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn315/dcn315_clk_mgr.c b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn315/dcn315_clk_mgr.c
index 19a15acd1509..e4d22f74f986 100644
--- a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn315/dcn315_clk_mgr.c
+++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn315/dcn315_clk_mgr.c
@@ -130,7 +130,7 @@ static void dcn315_update_clocks(struct clk_mgr *clk_mgr_base,
 	struct clk_mgr_internal *clk_mgr = TO_CLK_MGR_INTERNAL(clk_mgr_base);
 	struct dc_clocks *new_clocks = &context->bw_ctx.bw.dcn.clk;
 	struct dc *dc = clk_mgr_base->ctx->dc;
-	int display_count;
+	int display_count = 0;
 	bool update_dppclk = false;
 	bool update_dispclk = false;
 	bool dpp_clock_lowered = false;
@@ -202,15 +202,19 @@ static void dcn315_update_clocks(struct clk_mgr *clk_mgr_base,
 		update_dppclk = true;
 	}
 
-	if (should_set_clock(safe_to_lower, new_clocks->dispclk_khz, clk_mgr_base->clks.dispclk_khz)) {
-		/* No need to apply the w/a if we haven't taken over from bios yet */
-		if (clk_mgr_base->clks.dispclk_khz)
-			dcn315_disable_otg_wa(clk_mgr_base, context, true);
+	if (should_set_clock(safe_to_lower, new_clocks->dispclk_khz, clk_mgr_base->clks.dispclk_khz) &&
+	    (new_clocks->dispclk_khz > 0 || (safe_to_lower && display_count == 0))) {
+		int requested_dispclk_khz = new_clocks->dispclk_khz;
 
+		dcn315_disable_otg_wa(clk_mgr_base, context, true);
+
+		/* Clamp the requested clock to PMFW based on their limit. */
+		if (dc->debug.min_disp_clk_khz > 0 && requested_dispclk_khz < dc->debug.min_disp_clk_khz)
+			requested_dispclk_khz = dc->debug.min_disp_clk_khz;
+
+		dcn315_smu_set_dispclk(clk_mgr, requested_dispclk_khz);
 		clk_mgr_base->clks.dispclk_khz = new_clocks->dispclk_khz;
-		dcn315_smu_set_dispclk(clk_mgr, clk_mgr_base->clks.dispclk_khz);
-		if (clk_mgr_base->clks.dispclk_khz)
-			dcn315_disable_otg_wa(clk_mgr_base, context, false);
+		dcn315_disable_otg_wa(clk_mgr_base, context, false);
 
 		update_dispclk = true;
 	}
diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn316/dcn316_clk_mgr.c b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn316/dcn316_clk_mgr.c
index 4b19d9cf27ce..49efea0c8fcf 100644
--- a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn316/dcn316_clk_mgr.c
+++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn316/dcn316_clk_mgr.c
@@ -140,7 +140,7 @@ static void dcn316_update_clocks(struct clk_mgr *clk_mgr_base,
 	struct clk_mgr_internal *clk_mgr = TO_CLK_MGR_INTERNAL(clk_mgr_base);
 	struct dc_clocks *new_clocks = &context->bw_ctx.bw.dcn.clk;
 	struct dc *dc = clk_mgr_base->ctx->dc;
-	int display_count;
+	int display_count = 0;
 	bool update_dppclk = false;
 	bool update_dispclk = false;
 	bool dpp_clock_lowered = false;
@@ -209,11 +209,18 @@ static void dcn316_update_clocks(struct clk_mgr *clk_mgr_base,
 		update_dppclk = true;
 	}
 
-	if (should_set_clock(safe_to_lower, new_clocks->dispclk_khz, clk_mgr_base->clks.dispclk_khz)) {
+	if (should_set_clock(safe_to_lower, new_clocks->dispclk_khz, clk_mgr_base->clks.dispclk_khz) &&
+	    (new_clocks->dispclk_khz > 0 || (safe_to_lower && display_count == 0))) {
+		int requested_dispclk_khz = new_clocks->dispclk_khz;
+
 		dcn316_disable_otg_wa(clk_mgr_base, context, safe_to_lower, true);
 
+		/* Clamp the requested clock to PMFW based on their limit. */
+		if (dc->debug.min_disp_clk_khz > 0 && requested_dispclk_khz < dc->debug.min_disp_clk_khz)
+			requested_dispclk_khz = dc->debug.min_disp_clk_khz;
+
+		dcn316_smu_set_dispclk(clk_mgr, requested_dispclk_khz);
 		clk_mgr_base->clks.dispclk_khz = new_clocks->dispclk_khz;
-		dcn316_smu_set_dispclk(clk_mgr, clk_mgr_base->clks.dispclk_khz);
 		dcn316_disable_otg_wa(clk_mgr_base, context, safe_to_lower, false);
 
 		update_dispclk = true;
diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn35/dcn35_clk_mgr.c b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn35/dcn35_clk_mgr.c
index af722519a1fa..142de8938d7c 100644
--- a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn35/dcn35_clk_mgr.c
+++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn35/dcn35_clk_mgr.c
@@ -204,6 +204,8 @@ static void dcn35_disable_otg_wa(struct clk_mgr *clk_mgr_base, struct dc_state *
 		struct link_encoder *new_pipe_link_enc = new_pipe->link_res.dio_link_enc;
 		struct link_encoder *pipe_link_enc = pipe->link_res.dio_link_enc;
 		bool stream_changed_otg_dig_on = false;
+		bool has_active_hpo = false;
+
 		if (pipe->top_pipe || pipe->prev_odm_pipe)
 			continue;
 
@@ -225,20 +227,15 @@ static void dcn35_disable_otg_wa(struct clk_mgr *clk_mgr_base, struct dc_state *
 		new_pipe->stream_res.stream_enc->funcs->is_fifo_enabled &&
 		new_pipe->stream_res.stream_enc->funcs->is_fifo_enabled(new_pipe->stream_res.stream_enc);
 
-		bool has_active_hpo = false;
-
 		if (old_pipe->stream && new_pipe->stream && old_pipe->stream == new_pipe->stream) {
 			has_active_hpo =  dccg->ctx->dc->link_srv->dp_is_128b_132b_signal(old_pipe) &&
 			dccg->ctx->dc->link_srv->dp_is_128b_132b_signal(new_pipe);
 
-		 }
-
-
-		if (!has_active_hpo && !dccg->ctx->dc->link_srv->dp_is_128b_132b_signal(pipe) &&
-					(pipe->stream && (pipe->stream->dpms_off || dc_is_virtual_signal(pipe->stream->signal) ||
-					!pipe_link_enc) && !stream_changed_otg_dig_on)) {
-
+		}
 
+		if (!has_active_hpo && !stream_changed_otg_dig_on && pipe->stream &&
+		    (pipe->stream->dpms_off || dc_is_virtual_signal(pipe->stream->signal) || !pipe_link_enc) &&
+		    !dccg->ctx->dc->link_srv->dp_is_128b_132b_signal(pipe)) {
 			/* This w/a should not trigger when we have a dig active */
 			if (disable) {
 				if (pipe->stream_res.tg && pipe->stream_res.tg->funcs->immediate_disable_crtc)
diff --git a/drivers/gpu/drm/amd/display/dc/core/dc_link_exports.c b/drivers/gpu/drm/amd/display/dc/core/dc_link_exports.c
index 261c3bc4d46e..71e15da4bb69 100644
--- a/drivers/gpu/drm/amd/display/dc/core/dc_link_exports.c
+++ b/drivers/gpu/drm/amd/display/dc/core/dc_link_exports.c
@@ -370,15 +370,10 @@ bool dc_link_should_enable_fec(const struct dc_link *link)
 	return link->dc->link_srv->dp_should_enable_fec(link);
 }
 
-int dc_link_dp_dpia_handle_usb4_bandwidth_allocation_for_link(
+void dc_link_dp_dpia_handle_usb4_bandwidth_allocation_for_link(
 		struct dc_link *link, int peak_bw)
 {
-	return link->dc->link_srv->dpia_handle_usb4_bandwidth_allocation_for_link(link, peak_bw);
-}
-
-void dc_link_handle_usb4_bw_alloc_response(struct dc_link *link, uint8_t bw, uint8_t result)
-{
-	link->dc->link_srv->dpia_handle_bw_alloc_response(link, bw, result);
+	link->dc->link_srv->dpia_handle_usb4_bandwidth_allocation_for_link(link, peak_bw);
 }
 
 bool dc_link_check_link_loss_status(
diff --git a/drivers/gpu/drm/amd/display/dc/dc.h b/drivers/gpu/drm/amd/display/dc/dc.h
index 67e1bb6fa335..7c2ee0526926 100644
--- a/drivers/gpu/drm/amd/display/dc/dc.h
+++ b/drivers/gpu/drm/amd/display/dc/dc.h
@@ -53,7 +53,7 @@ struct aux_payload;
 struct set_config_cmd_payload;
 struct dmub_notification;
 
-#define DC_VER "3.2.324"
+#define DC_VER "3.2.325"
 
 /**
  * MAX_SURFACES - representative of the upper bound of surfaces that can be piped to a single CRTC
@@ -2354,19 +2354,6 @@ unsigned int dc_dp_trace_get_link_loss_count(struct dc_link *link);
 void dc_link_set_usb4_req_bw_req(struct dc_link *link, int req_bw);
 
 /*
- * Handle function for when the status of the Request above is complete.
- * We will find out the result of allocating on CM and update structs.
- *
- * @link: pointer to the dc_link struct instance
- * @bw: Allocated or Estimated BW depending on the result
- * @result: Response type
- *
- * return: none
- */
-void dc_link_handle_usb4_bw_alloc_response(struct dc_link *link,
-		uint8_t bw, uint8_t result);
-
-/*
  * Handle the USB4 BW Allocation related functionality here:
  * Plug => Try to allocate max bw from timing parameters supported by the sink
  * Unplug => de-allocate bw
@@ -2374,9 +2361,8 @@ void dc_link_handle_usb4_bw_alloc_response(struct dc_link *link,
  * @link: pointer to the dc_link struct instance
  * @peak_bw: Peak bw used by the link/sink
  *
- * return: allocated bw else return 0
  */
-int dc_link_dp_dpia_handle_usb4_bandwidth_allocation_for_link(
+void dc_link_dp_dpia_handle_usb4_bandwidth_allocation_for_link(
 		struct dc_link *link, int peak_bw);
 
 /*
diff --git a/drivers/gpu/drm/amd/display/dc/dc_dmub_srv.c b/drivers/gpu/drm/amd/display/dc/dc_dmub_srv.c
index ca6da53f45ad..614e03bfd598 100644
--- a/drivers/gpu/drm/amd/display/dc/dc_dmub_srv.c
+++ b/drivers/gpu/drm/amd/display/dc/dc_dmub_srv.c
@@ -70,28 +70,20 @@ void dc_dmub_srv_destroy(struct dc_dmub_srv **dmub_srv)
 	}
 }
 
-bool dc_dmub_srv_wait_for_pending(struct dc_dmub_srv *dc_dmub_srv)
+void dc_dmub_srv_wait_idle(struct dc_dmub_srv *dc_dmub_srv)
 {
-	struct dmub_srv *dmub;
-	struct dc_context *dc_ctx;
+	struct dmub_srv *dmub = dc_dmub_srv->dmub;
+	struct dc_context *dc_ctx = dc_dmub_srv->ctx;
 	enum dmub_status status;
 
-	if (!dc_dmub_srv || !dc_dmub_srv->dmub)
-		return false;
-
-	dc_ctx = dc_dmub_srv->ctx;
-	dmub = dc_dmub_srv->dmub;
-
 	do {
-		status = dmub_srv_wait_for_pending(dmub, 100000);
+		status = dmub_srv_wait_for_idle(dmub, 100000);
 	} while (dc_dmub_srv->ctx->dc->debug.disable_timeout && status != DMUB_STATUS_OK);
 
 	if (status != DMUB_STATUS_OK) {
 		DC_ERROR("Error waiting for DMUB idle: status=%d\n", status);
 		dc_dmub_srv_log_diagnostic_data(dc_dmub_srv);
 	}
-
-	return status == DMUB_STATUS_OK;
 }
 
 void dc_dmub_srv_clear_inbox0_ack(struct dc_dmub_srv *dc_dmub_srv)
@@ -134,49 +126,7 @@ void dc_dmub_srv_send_inbox0_cmd(struct dc_dmub_srv *dc_dmub_srv,
 	}
 }
 
-static bool dc_dmub_srv_reg_cmd_list_queue_execute(struct dc_dmub_srv *dc_dmub_srv,
-		unsigned int count,
-		union dmub_rb_cmd *cmd_list)
-{
-	struct dc_context *dc_ctx;
-	struct dmub_srv *dmub;
-	enum dmub_status status = DMUB_STATUS_OK;
-	int i;
-
-	if (!dc_dmub_srv || !dc_dmub_srv->dmub)
-		return false;
-
-	dc_ctx = dc_dmub_srv->ctx;
-	dmub = dc_dmub_srv->dmub;
-
-	for (i = 0 ; i < count; i++) {
-		/* confirm no messages pending */
-		do {
-			status = dmub_srv_wait_for_idle(dmub, 100000);
-		} while (dc_dmub_srv->ctx->dc->debug.disable_timeout && status != DMUB_STATUS_OK);
-
-		/* queue command */
-		if (status == DMUB_STATUS_OK)
-			status = dmub_srv_reg_cmd_execute(dmub, &cmd_list[i]);
-
-		/* check for errors */
-		if (status != DMUB_STATUS_OK) {
-			break;
-		}
-	}
-
-	if (status != DMUB_STATUS_OK) {
-		if (status != DMUB_STATUS_POWER_STATE_D3) {
-			DC_ERROR("Error starting DMUB execution: status=%d\n", status);
-			dc_dmub_srv_log_diagnostic_data(dc_dmub_srv);
-		}
-		return false;
-	}
-
-	return true;
-}
-
-static bool dc_dmub_srv_fb_cmd_list_queue_execute(struct dc_dmub_srv *dc_dmub_srv,
+bool dc_dmub_srv_cmd_list_queue_execute(struct dc_dmub_srv *dc_dmub_srv,
 		unsigned int count,
 		union dmub_rb_cmd *cmd_list)
 {
@@ -193,16 +143,11 @@ static bool dc_dmub_srv_fb_cmd_list_queue_execute(struct dc_dmub_srv *dc_dmub_sr
 
 	for (i = 0 ; i < count; i++) {
 		// Queue command
-		if (!cmd_list[i].cmd_common.header.multi_cmd_pending ||
-				dmub_rb_num_free(&dmub->inbox1.rb) >= count - i) {
-			status = dmub_srv_fb_cmd_queue(dmub, &cmd_list[i]);
-		} else {
-			status = DMUB_STATUS_QUEUE_FULL;
-		}
+		status = dmub_srv_cmd_queue(dmub, &cmd_list[i]);
 
 		if (status == DMUB_STATUS_QUEUE_FULL) {
 			/* Execute and wait for queue to become empty again. */
-			status = dmub_srv_fb_cmd_execute(dmub);
+			status = dmub_srv_cmd_execute(dmub);
 			if (status == DMUB_STATUS_POWER_STATE_D3)
 				return false;
 
@@ -211,7 +156,7 @@ static bool dc_dmub_srv_fb_cmd_list_queue_execute(struct dc_dmub_srv *dc_dmub_sr
 			} while (dc_dmub_srv->ctx->dc->debug.disable_timeout && status != DMUB_STATUS_OK);
 
 			/* Requeue the command. */
-			status = dmub_srv_fb_cmd_queue(dmub, &cmd_list[i]);
+			status = dmub_srv_cmd_queue(dmub, &cmd_list[i]);
 		}
 
 		if (status != DMUB_STATUS_OK) {
@@ -223,7 +168,7 @@ static bool dc_dmub_srv_fb_cmd_list_queue_execute(struct dc_dmub_srv *dc_dmub_sr
 		}
 	}
 
-	status = dmub_srv_fb_cmd_execute(dmub);
+	status = dmub_srv_cmd_execute(dmub);
 	if (status != DMUB_STATUS_OK) {
 		if (status != DMUB_STATUS_POWER_STATE_D3) {
 			DC_ERROR("Error starting DMUB execution: status=%d\n", status);
@@ -235,23 +180,6 @@ static bool dc_dmub_srv_fb_cmd_list_queue_execute(struct dc_dmub_srv *dc_dmub_sr
 	return true;
 }
 
-bool dc_dmub_srv_cmd_list_queue_execute(struct dc_dmub_srv *dc_dmub_srv,
-		unsigned int count,
-		union dmub_rb_cmd *cmd_list)
-{
-	bool res = false;
-
-	if (dc_dmub_srv && dc_dmub_srv->dmub) {
-		if (dc_dmub_srv->dmub->inbox_type == DMUB_CMD_INTERFACE_REG) {
-			res = dc_dmub_srv_reg_cmd_list_queue_execute(dc_dmub_srv, count, cmd_list);
-		} else {
-			res = dc_dmub_srv_fb_cmd_list_queue_execute(dc_dmub_srv, count, cmd_list);
-		}
-	}
-
-	return res;
-}
-
 bool dc_dmub_srv_wait_for_idle(struct dc_dmub_srv *dc_dmub_srv,
 		enum dm_dmub_wait_type wait_type,
 		union dmub_rb_cmd *cmd_list)
@@ -274,8 +202,7 @@ bool dc_dmub_srv_wait_for_idle(struct dc_dmub_srv *dc_dmub_srv,
 			DC_LOG_DEBUG("No reply for DMUB command: status=%d\n", status);
 			if (!dmub->debug.timeout_info.timeout_occured) {
 				dmub->debug.timeout_info.timeout_occured = true;
-				if (cmd_list)
-					dmub->debug.timeout_info.timeout_cmd = *cmd_list;
+				dmub->debug.timeout_info.timeout_cmd = *cmd_list;
 				dmub->debug.timeout_info.timestamp = dm_get_timestamp(dc_dmub_srv->ctx);
 			}
 			dc_dmub_srv_log_diagnostic_data(dc_dmub_srv);
@@ -283,9 +210,8 @@ bool dc_dmub_srv_wait_for_idle(struct dc_dmub_srv *dc_dmub_srv,
 		}
 
 		// Copy data back from ring buffer into command
-		if (wait_type == DM_DMUB_WAIT_TYPE_WAIT_WITH_REPLY && cmd_list) {
-			dmub_srv_cmd_get_response(dc_dmub_srv->dmub, cmd_list);
-		}
+		if (wait_type == DM_DMUB_WAIT_TYPE_WAIT_WITH_REPLY)
+			dmub_rb_get_return_data(&dmub->inbox1_rb, cmd_list);
 	}
 
 	return true;
@@ -298,10 +224,74 @@ bool dc_dmub_srv_cmd_run(struct dc_dmub_srv *dc_dmub_srv, union dmub_rb_cmd *cmd
 
 bool dc_dmub_srv_cmd_run_list(struct dc_dmub_srv *dc_dmub_srv, unsigned int count, union dmub_rb_cmd *cmd_list, enum dm_dmub_wait_type wait_type)
 {
-	if (!dc_dmub_srv_cmd_list_queue_execute(dc_dmub_srv, count, cmd_list))
+	struct dc_context *dc_ctx;
+	struct dmub_srv *dmub;
+	enum dmub_status status;
+	int i;
+
+	if (!dc_dmub_srv || !dc_dmub_srv->dmub)
+		return false;
+
+	dc_ctx = dc_dmub_srv->ctx;
+	dmub = dc_dmub_srv->dmub;
+
+	for (i = 0 ; i < count; i++) {
+		// Queue command
+		status = dmub_srv_cmd_queue(dmub, &cmd_list[i]);
+
+		if (status == DMUB_STATUS_QUEUE_FULL) {
+			/* Execute and wait for queue to become empty again. */
+			status = dmub_srv_cmd_execute(dmub);
+			if (status == DMUB_STATUS_POWER_STATE_D3)
+				return false;
+
+			status = dmub_srv_wait_for_idle(dmub, 100000);
+			if (status != DMUB_STATUS_OK)
+				return false;
+
+			/* Requeue the command. */
+			status = dmub_srv_cmd_queue(dmub, &cmd_list[i]);
+		}
+
+		if (status != DMUB_STATUS_OK) {
+			if (status != DMUB_STATUS_POWER_STATE_D3) {
+				DC_ERROR("Error queueing DMUB command: status=%d\n", status);
+				dc_dmub_srv_log_diagnostic_data(dc_dmub_srv);
+			}
+			return false;
+		}
+	}
+
+	status = dmub_srv_cmd_execute(dmub);
+	if (status != DMUB_STATUS_OK) {
+		if (status != DMUB_STATUS_POWER_STATE_D3) {
+			DC_ERROR("Error starting DMUB execution: status=%d\n", status);
+			dc_dmub_srv_log_diagnostic_data(dc_dmub_srv);
+		}
 		return false;
+	}
+
+	// Wait for DMUB to process command
+	if (wait_type != DM_DMUB_WAIT_TYPE_NO_WAIT) {
+		if (dc_dmub_srv->ctx->dc->debug.disable_timeout) {
+			do {
+				status = dmub_srv_wait_for_idle(dmub, 100000);
+			} while (status != DMUB_STATUS_OK);
+		} else
+			status = dmub_srv_wait_for_idle(dmub, 100000);
 
-	return dc_dmub_srv_wait_for_idle(dc_dmub_srv, wait_type, cmd_list);
+		if (status != DMUB_STATUS_OK) {
+			DC_LOG_DEBUG("No reply for DMUB command: status=%d\n", status);
+			dc_dmub_srv_log_diagnostic_data(dc_dmub_srv);
+			return false;
+		}
+
+		// Copy data back from ring buffer into command
+		if (wait_type == DM_DMUB_WAIT_TYPE_WAIT_WITH_REPLY)
+			dmub_rb_get_return_data(&dmub->inbox1_rb, cmd_list);
+	}
+
+	return true;
 }
 
 bool dc_dmub_srv_optimized_init_done(struct dc_dmub_srv *dc_dmub_srv)
@@ -1253,7 +1243,7 @@ static void dc_dmub_srv_notify_idle(const struct dc *dc, bool allow_idle)
 			ips_fw->signals.bits.ips1_commit,
 			ips_fw->signals.bits.ips2_commit);
 
-		dc_dmub_srv_wait_for_idle(dc->ctx->dmub_srv, DM_DMUB_WAIT_TYPE_WAIT, NULL);
+		dc_dmub_srv_wait_idle(dc->ctx->dmub_srv);
 
 		memset(&new_signals, 0, sizeof(new_signals));
 
@@ -1410,7 +1400,7 @@ static void dc_dmub_srv_exit_low_power_state(const struct dc *dc)
 					ips_fw->signals.bits.ips1_commit,
 					ips_fw->signals.bits.ips2_commit);
 
-				dmub_srv_sync_inboxes(dc->ctx->dmub_srv->dmub);
+				dmub_srv_sync_inbox1(dc->ctx->dmub_srv->dmub);
 			}
 		}
 
@@ -1664,8 +1654,7 @@ void dc_dmub_srv_fams2_update_config(struct dc *dc,
 	/* fill in generic command header */
 	global_cmd->header.type = DMUB_CMD__FW_ASSISTED_MCLK_SWITCH;
 	global_cmd->header.sub_type = DMUB_CMD__FAMS2_CONFIG;
-	global_cmd->header.payload_bytes =
-			sizeof(struct dmub_rb_cmd_fams2) - sizeof(struct dmub_cmd_header);
+	global_cmd->header.payload_bytes = sizeof(struct dmub_rb_cmd_fams2) - sizeof(struct dmub_cmd_header);
 
 	if (enable) {
 		/* send global configuration parameters */
@@ -1684,13 +1673,11 @@ void dc_dmub_srv_fams2_update_config(struct dc *dc,
 			/* configure command header */
 			stream_base_cmd->header.type = DMUB_CMD__FW_ASSISTED_MCLK_SWITCH;
 			stream_base_cmd->header.sub_type = DMUB_CMD__FAMS2_CONFIG;
-			stream_base_cmd->header.payload_bytes =
-					sizeof(struct dmub_rb_cmd_fams2) - sizeof(struct dmub_cmd_header);
+			stream_base_cmd->header.payload_bytes = sizeof(struct dmub_rb_cmd_fams2) - sizeof(struct dmub_cmd_header);
 			stream_base_cmd->header.multi_cmd_pending = 1;
 			stream_sub_state_cmd->header.type = DMUB_CMD__FW_ASSISTED_MCLK_SWITCH;
 			stream_sub_state_cmd->header.sub_type = DMUB_CMD__FAMS2_CONFIG;
-			stream_sub_state_cmd->header.payload_bytes =
-					sizeof(struct dmub_rb_cmd_fams2) - sizeof(struct dmub_cmd_header);
+			stream_sub_state_cmd->header.payload_bytes = sizeof(struct dmub_rb_cmd_fams2) - sizeof(struct dmub_cmd_header);
 			stream_sub_state_cmd->header.multi_cmd_pending = 1;
 			/* copy stream static base state */
 			memcpy(&stream_base_cmd->config,
@@ -1736,8 +1723,7 @@ void dc_dmub_srv_fams2_drr_update(struct dc *dc,
 	cmd.fams2_drr_update.dmub_optc_state_req.v_total_mid_frame_num = vtotal_mid_frame_num;
 	cmd.fams2_drr_update.dmub_optc_state_req.program_manual_trigger = program_manual_trigger;
 
-	cmd.fams2_drr_update.header.payload_bytes =
-			sizeof(cmd.fams2_drr_update) - sizeof(cmd.fams2_drr_update.header);
+	cmd.fams2_drr_update.header.payload_bytes = sizeof(cmd.fams2_drr_update) - sizeof(cmd.fams2_drr_update.header);
 
 	dm_execute_dmub_cmd(dc->ctx, &cmd, DM_DMUB_WAIT_TYPE_WAIT);
 }
@@ -1773,8 +1759,7 @@ void dc_dmub_srv_fams2_passthrough_flip(
 		/* build command header */
 		cmds[num_cmds].fams2_flip.header.type = DMUB_CMD__FW_ASSISTED_MCLK_SWITCH;
 		cmds[num_cmds].fams2_flip.header.sub_type = DMUB_CMD__FAMS2_FLIP;
-		cmds[num_cmds].fams2_flip.header.payload_bytes =
-				sizeof(struct dmub_rb_cmd_fams2_flip) - sizeof(struct dmub_cmd_header);
+		cmds[num_cmds].fams2_flip.header.payload_bytes = sizeof(struct dmub_rb_cmd_fams2_flip);
 
 		/* for chaining multiple commands, all but last command should set to 1 */
 		cmds[num_cmds].fams2_flip.header.multi_cmd_pending = 1;
diff --git a/drivers/gpu/drm/amd/display/dc/dc_dmub_srv.h b/drivers/gpu/drm/amd/display/dc/dc_dmub_srv.h
index ada5c2fb2db3..a636f4c3f01d 100644
--- a/drivers/gpu/drm/amd/display/dc/dc_dmub_srv.h
+++ b/drivers/gpu/drm/amd/display/dc/dc_dmub_srv.h
@@ -58,7 +58,7 @@ struct dc_dmub_srv {
 	bool needs_idle_wake;
 };
 
-bool dc_dmub_srv_wait_for_pending(struct dc_dmub_srv *dc_dmub_srv);
+void dc_dmub_srv_wait_idle(struct dc_dmub_srv *dc_dmub_srv);
 
 bool dc_dmub_srv_optimized_init_done(struct dc_dmub_srv *dc_dmub_srv);
 
diff --git a/drivers/gpu/drm/amd/display/dc/dc_helper.c b/drivers/gpu/drm/amd/display/dc/dc_helper.c
index 72b87b78da0e..8f077e15b4f0 100644
--- a/drivers/gpu/drm/amd/display/dc/dc_helper.c
+++ b/drivers/gpu/drm/amd/display/dc/dc_helper.c
@@ -682,7 +682,7 @@ void reg_sequence_wait_done(const struct dc_context *ctx)
 	if (offload &&
 	    ctx->dc->debug.dmub_offload_enabled &&
 	    !ctx->dc->debug.dmcub_emulation) {
-		dc_dmub_srv_wait_for_idle(ctx->dmub_srv, DM_DMUB_WAIT_TYPE_WAIT, NULL);
+		dc_dmub_srv_wait_idle(ctx->dmub_srv);
 	}
 }
 
diff --git a/drivers/gpu/drm/amd/display/dc/dc_types.h b/drivers/gpu/drm/amd/display/dc/dc_types.h
index acd3b373a18e..83ffaae9f439 100644
--- a/drivers/gpu/drm/amd/display/dc/dc_types.h
+++ b/drivers/gpu/drm/amd/display/dc/dc_types.h
@@ -1224,7 +1224,6 @@ struct dc_dpia_bw_alloc {
 	int bw_granularity;    // BW Granularity
 	int dp_overhead;       // DP overhead in dp tunneling
 	bool bw_alloc_enabled; // The BW Alloc Mode Support is turned ON for all 3:  DP-Tx & Dpia & CM
-	bool response_ready;   // Response ready from the CM side
 	uint8_t nrd_max_lane_count; // Non-reduced max lane count
 	uint8_t nrd_max_link_rate; // Non-reduced max link rate
 };
diff --git a/drivers/gpu/drm/amd/display/dc/dce/dmub_abm_lcd.c b/drivers/gpu/drm/amd/display/dc/dce/dmub_abm_lcd.c
index a641ae04450c..0d7e7f3b81a1 100644
--- a/drivers/gpu/drm/amd/display/dc/dce/dmub_abm_lcd.c
+++ b/drivers/gpu/drm/amd/display/dc/dce/dmub_abm_lcd.c
@@ -240,8 +240,7 @@ bool dmub_abm_save_restore(
 	cmd.abm_save_restore.abm_init_config_data.version = DMUB_CMD_ABM_CONTROL_VERSION_1;
 	cmd.abm_save_restore.abm_init_config_data.panel_mask = panel_mask;
 
-	cmd.abm_save_restore.header.payload_bytes =
-			sizeof(struct dmub_rb_cmd_abm_save_restore) - sizeof(struct dmub_cmd_header);
+	cmd.abm_save_restore.header.payload_bytes = sizeof(struct dmub_rb_cmd_abm_save_restore);
 
 	dc_wake_and_execute_dmub_cmd(dc, &cmd, DM_DMUB_WAIT_TYPE_WAIT);
 
diff --git a/drivers/gpu/drm/amd/display/dc/dce/dmub_hw_lock_mgr.c b/drivers/gpu/drm/amd/display/dc/dce/dmub_hw_lock_mgr.c
index a262598aa676..d37ecfdde4f1 100644
--- a/drivers/gpu/drm/amd/display/dc/dce/dmub_hw_lock_mgr.c
+++ b/drivers/gpu/drm/amd/display/dc/dce/dmub_hw_lock_mgr.c
@@ -73,5 +73,16 @@ bool should_use_dmub_lock(struct dc_link *link)
 	if (link->replay_settings.replay_feature_enabled)
 		return true;
 
+	/* only use HW lock for PSR1 on single eDP */
+	if (link->psr_settings.psr_version == DC_PSR_VERSION_1) {
+		struct dc_link *edp_links[MAX_NUM_EDP];
+		int edp_num;
+
+		dc_get_edp_links(link->dc, edp_links, &edp_num);
+
+		if (edp_num == 1)
+			return true;
+	}
+
 	return false;
 }
diff --git a/drivers/gpu/drm/amd/display/dc/dce/dmub_replay.c b/drivers/gpu/drm/amd/display/dc/dce/dmub_replay.c
index fcd3d86ad517..c31e4f26a305 100644
--- a/drivers/gpu/drm/amd/display/dc/dce/dmub_replay.c
+++ b/drivers/gpu/drm/amd/display/dc/dce/dmub_replay.c
@@ -280,9 +280,7 @@ static void dmub_replay_set_power_opt_and_coasting_vtotal(struct dmub_replay *dm
 	memset(&cmd, 0, sizeof(cmd));
 	pCmd->header.type = DMUB_CMD__REPLAY;
 	pCmd->header.sub_type = DMUB_CMD__REPLAY_SET_POWER_OPT_AND_COASTING_VTOTAL;
-	pCmd->header.payload_bytes =
-			sizeof(struct dmub_rb_cmd_replay_set_power_opt_and_coasting_vtotal) -
-			sizeof(struct dmub_cmd_header);
+	pCmd->header.payload_bytes = sizeof(struct dmub_rb_cmd_replay_set_power_opt_and_coasting_vtotal);
 	pCmd->replay_set_power_opt_data.power_opt = power_opt;
 	pCmd->replay_set_power_opt_data.panel_inst = panel_inst;
 	pCmd->replay_set_coasting_vtotal_data.coasting_vtotal = (coasting_vtotal & 0xFFFF);
@@ -321,8 +319,7 @@ static void dmub_replay_send_cmd(struct dmub_replay *dmub,
 		cmd.replay_set_timing_sync.header.sub_type =
 			DMUB_CMD__REPLAY_SET_TIMING_SYNC_SUPPORTED;
 		cmd.replay_set_timing_sync.header.payload_bytes =
-			sizeof(struct dmub_rb_cmd_replay_set_timing_sync) -
-			sizeof(struct dmub_cmd_header);
+			sizeof(struct dmub_rb_cmd_replay_set_timing_sync);
 		//Cmd Body
 		cmd.replay_set_timing_sync.replay_set_timing_sync_data.panel_inst =
 						cmd_element->sync_data.panel_inst;
@@ -334,8 +331,7 @@ static void dmub_replay_send_cmd(struct dmub_replay *dmub,
 		cmd.replay_set_frameupdate_timer.header.sub_type =
 			DMUB_CMD__REPLAY_SET_RESIDENCY_FRAMEUPDATE_TIMER;
 		cmd.replay_set_frameupdate_timer.header.payload_bytes =
-			sizeof(struct dmub_rb_cmd_replay_set_frameupdate_timer) -
-			sizeof(struct dmub_cmd_header);
+			sizeof(struct dmub_rb_cmd_replay_set_frameupdate_timer);
 		//Cmd Body
 		cmd.replay_set_frameupdate_timer.data.panel_inst =
 						cmd_element->panel_inst;
@@ -349,8 +345,7 @@ static void dmub_replay_send_cmd(struct dmub_replay *dmub,
 		cmd.replay_set_pseudo_vtotal.header.sub_type =
 			DMUB_CMD__REPLAY_SET_PSEUDO_VTOTAL;
 		cmd.replay_set_pseudo_vtotal.header.payload_bytes =
-			sizeof(struct dmub_rb_cmd_replay_set_pseudo_vtotal) -
-			sizeof(struct dmub_cmd_header);
+			sizeof(struct dmub_rb_cmd_replay_set_pseudo_vtotal);
 		//Cmd Body
 		cmd.replay_set_pseudo_vtotal.data.panel_inst =
 			cmd_element->pseudo_vtotal_data.panel_inst;
@@ -362,8 +357,7 @@ static void dmub_replay_send_cmd(struct dmub_replay *dmub,
 		cmd.replay_disabled_adaptive_sync_sdp.header.sub_type =
 			DMUB_CMD__REPLAY_DISABLED_ADAPTIVE_SYNC_SDP;
 		cmd.replay_disabled_adaptive_sync_sdp.header.payload_bytes =
-			sizeof(struct dmub_rb_cmd_replay_disabled_adaptive_sync_sdp) -
-			sizeof(struct dmub_cmd_header);
+			sizeof(struct dmub_rb_cmd_replay_disabled_adaptive_sync_sdp);
 		//Cmd Body
 		cmd.replay_disabled_adaptive_sync_sdp.data.panel_inst =
 			cmd_element->disabled_adaptive_sync_sdp_data.panel_inst;
@@ -375,8 +369,7 @@ static void dmub_replay_send_cmd(struct dmub_replay *dmub,
 		cmd.replay_set_general_cmd.header.sub_type =
 			DMUB_CMD__REPLAY_SET_GENERAL_CMD;
 		cmd.replay_set_general_cmd.header.payload_bytes =
-			sizeof(struct dmub_rb_cmd_replay_set_general_cmd) -
-			sizeof(struct dmub_cmd_header);
+			sizeof(struct dmub_rb_cmd_replay_set_general_cmd);
 		//Cmd Body
 		cmd.replay_set_general_cmd.data.panel_inst =
 			cmd_element->set_general_cmd_data.panel_inst;
diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn35/dcn35_fpu.c b/drivers/gpu/drm/amd/display/dc/dml/dcn35/dcn35_fpu.c
index e8efffcc69a1..92f0a099d089 100644
--- a/drivers/gpu/drm/amd/display/dc/dml/dcn35/dcn35_fpu.c
+++ b/drivers/gpu/drm/amd/display/dc/dml/dcn35/dcn35_fpu.c
@@ -159,7 +159,7 @@ struct _vcs_dpi_soc_bounding_box_st dcn3_5_soc = {
 			.dppclk_mhz = 1200.0,
 			.phyclk_mhz = 810.0,
 			.phyclk_d18_mhz = 667.0,
-			.dscclk_mhz = 417.0,
+			.dscclk_mhz = 400.0,
 			.dtbclk_mhz = 600.0,
 		},
 	},
diff --git a/drivers/gpu/drm/amd/display/dc/dml2/display_mode_core.c b/drivers/gpu/drm/amd/display/dc/dml2/display_mode_core.c
index 84a2de9a76d4..7ae9c0ba0c9e 100644
--- a/drivers/gpu/drm/amd/display/dc/dml2/display_mode_core.c
+++ b/drivers/gpu/drm/amd/display/dc/dml2/display_mode_core.c
@@ -32,6 +32,7 @@
 
 #define DML2_MAX_FMT_420_BUFFER_WIDTH 4096
 #define TB_BORROWED_MAX 400
+#define DML_MAX_VSTARTUP_START 1023
 
 // ---------------------------
 //  Declaration Begins
@@ -6210,6 +6211,7 @@ static dml_uint_t CalculateMaxVStartup(
 	dml_print("DML::%s: vblank_avail = %u\n", __func__, vblank_avail);
 	dml_print("DML::%s: max_vstartup_lines = %u\n", __func__, max_vstartup_lines);
 #endif
+	max_vstartup_lines = (dml_uint_t) dml_min(max_vstartup_lines, DML_MAX_VSTARTUP_START);
 	return max_vstartup_lines;
 }
 
diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml2_translation_helper.c b/drivers/gpu/drm/amd/display/dc/dml2/dml2_translation_helper.c
index 2061d43b92e1..70c39df62533 100644
--- a/drivers/gpu/drm/amd/display/dc/dml2/dml2_translation_helper.c
+++ b/drivers/gpu/drm/amd/display/dc/dml2/dml2_translation_helper.c
@@ -590,11 +590,11 @@ void dml2_init_soc_states(struct dml2_context *dml2, const struct dc *in_dc,
 			p->out_states->state_array[i].dtbclk_mhz = max_dtbclk_mhz;
 			p->out_states->state_array[i].phyclk_mhz = max_phyclk_mhz;
 
-			p->out_states->state_array[i].dscclk_mhz = max_dispclk_mhz / 3.0;
 			p->out_states->state_array[i].phyclk_mhz = max_phyclk_mhz;
 			p->out_states->state_array[i].dtbclk_mhz = max_dtbclk_mhz;
 
 			/* Dependent states. */
+			p->out_states->state_array[i].dscclk_mhz = p->in_states->state_array[i].dscclk_mhz;
 			p->out_states->state_array[i].dram_speed_mts = p->in_states->state_array[i].dram_speed_mts;
 			p->out_states->state_array[i].fabricclk_mhz = p->in_states->state_array[i].fabricclk_mhz;
 			p->out_states->state_array[i].socclk_mhz = p->in_states->state_array[i].socclk_mhz;
diff --git a/drivers/gpu/drm/amd/display/dc/inc/link.h b/drivers/gpu/drm/amd/display/dc/inc/link.h
index fd1f9d3db039..2948a696ee12 100644
--- a/drivers/gpu/drm/amd/display/dc/inc/link.h
+++ b/drivers/gpu/drm/amd/display/dc/inc/link.h
@@ -218,10 +218,8 @@ struct link_service {
 
 
 	/*************************** DP DPIA/PHY ******************************/
-	int (*dpia_handle_usb4_bandwidth_allocation_for_link)(
+	void (*dpia_handle_usb4_bandwidth_allocation_for_link)(
 			struct dc_link *link, int peak_bw);
-	void (*dpia_handle_bw_alloc_response)(
-			struct dc_link *link, uint8_t bw, uint8_t result);
 	void (*dp_set_drive_settings)(
 		struct dc_link *link,
 		const struct link_resource *link_res,
diff --git a/drivers/gpu/drm/amd/display/dc/link/link_dpms.c b/drivers/gpu/drm/amd/display/dc/link/link_dpms.c
index 321fd1785370..268626e73c54 100644
--- a/drivers/gpu/drm/amd/display/dc/link/link_dpms.c
+++ b/drivers/gpu/drm/amd/display/dc/link/link_dpms.c
@@ -2291,22 +2291,7 @@ static bool allocate_usb4_bandwidth_for_stream(struct dc_stream_state *stream, i
 	link->dpia_bw_alloc_config.dp_overhead = link_dp_dpia_get_dp_overhead_in_dp_tunneling(link);
 	req_bw += link->dpia_bw_alloc_config.dp_overhead;
 
-	if (link_dp_dpia_allocate_usb4_bandwidth_for_stream(link, req_bw)) {
-		if (req_bw <= link->dpia_bw_alloc_config.allocated_bw) {
-			DC_LOG_DEBUG("%s, Success in allocate bw for link(%d), allocated_bw(%d), dp_overhead(%d)\n",
-					__func__, link->link_index, link->dpia_bw_alloc_config.allocated_bw,
-					link->dpia_bw_alloc_config.dp_overhead);
-		} else {
-			// Cannot get the required bandwidth.
-			DC_LOG_ERROR("%s, Failed to allocate bw for link(%d), allocated_bw(%d), dp_overhead(%d)\n",
-					__func__, link->link_index, link->dpia_bw_alloc_config.allocated_bw,
-					link->dpia_bw_alloc_config.dp_overhead);
-			return false;
-		}
-	} else {
-		DC_LOG_DEBUG("%s, usb4 request bw timeout\n", __func__);
-		return false;
-	}
+	link_dp_dpia_allocate_usb4_bandwidth_for_stream(link, req_bw);
 
 	if (stream->signal == SIGNAL_TYPE_DISPLAY_PORT_MST) {
 		int i = 0;
diff --git a/drivers/gpu/drm/amd/display/dc/link/link_factory.c b/drivers/gpu/drm/amd/display/dc/link/link_factory.c
index a7877d57a00f..f6b6b19e7481 100644
--- a/drivers/gpu/drm/amd/display/dc/link/link_factory.c
+++ b/drivers/gpu/drm/amd/display/dc/link/link_factory.c
@@ -175,7 +175,6 @@ static void construct_link_service_dp_phy_or_dpia(struct link_service *link_srv)
 {
 	link_srv->dpia_handle_usb4_bandwidth_allocation_for_link =
 			dpia_handle_usb4_bandwidth_allocation_for_link;
-	link_srv->dpia_handle_bw_alloc_response = dpia_handle_bw_alloc_response;
 	link_srv->dp_set_drive_settings = dp_set_drive_settings;
 	link_srv->dpcd_write_rx_power_ctrl = dpcd_write_rx_power_ctrl;
 }
diff --git a/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_dpia.c b/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_dpia.c
index c149210096ac..0d123e647652 100644
--- a/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_dpia.c
+++ b/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_dpia.c
@@ -92,7 +92,6 @@ bool dpia_query_hpd_status(struct dc_link *link)
 
 	/* prepare QUERY_HPD command */
 	cmd.query_hpd.header.type = DMUB_CMD__QUERY_HPD_STATE;
-	cmd.query_hpd.header.payload_bytes = sizeof(cmd.query_hpd.data);
 	cmd.query_hpd.data.instance = link->link_id.enum_id - ENUM_ID_1;
 	cmd.query_hpd.data.ch_type = AUX_CHANNEL_DPIA;
 
diff --git a/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_dpia_bw.c b/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_dpia_bw.c
index a5541b8fc95b..a254ead2f7e8 100644
--- a/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_dpia_bw.c
+++ b/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_dpia_bw.c
@@ -24,7 +24,7 @@
  *
  */
 /*********************************************************************/
-//				USB4 DPIA BANDWIDTH ALLOCATION LOGIC
+// USB4 DPIA BANDWIDTH ALLOCATION LOGIC
 /*********************************************************************/
 #include "link_dp_dpia_bw.h"
 #include "link_dpcd.h"
@@ -36,7 +36,7 @@
 #define Kbps_TO_Gbps (1000 * 1000)
 
 // ------------------------------------------------------------------
-//					PRIVATE FUNCTIONS
+// PRIVATE FUNCTIONS
 // ------------------------------------------------------------------
 /*
  * Always Check the following:
@@ -44,11 +44,11 @@
  *  - Is HPD HIGH?
  *  - Is BW Allocation Support Mode enabled on DP-Tx?
  */
-static bool get_bw_alloc_proceed_flag(struct dc_link *tmp)
+static bool link_dp_is_bw_alloc_available(struct dc_link *link)
 {
-	return (tmp && DISPLAY_ENDPOINT_USB4_DPIA == tmp->ep_type
-			&& tmp->hpd_status
-			&& tmp->dpia_bw_alloc_config.bw_alloc_enabled);
+	return (link && link->ep_type == DISPLAY_ENDPOINT_USB4_DPIA
+		&& link->hpd_status
+		&& link->dpia_bw_alloc_config.bw_alloc_enabled);
 }
 
 static void reset_bw_alloc_struct(struct dc_link *link)
@@ -60,7 +60,6 @@ static void reset_bw_alloc_struct(struct dc_link *link)
 	link->dpia_bw_alloc_config.estimated_bw = 0;
 	link->dpia_bw_alloc_config.bw_granularity = 0;
 	link->dpia_bw_alloc_config.dp_overhead = 0;
-	link->dpia_bw_alloc_config.response_ready = false;
 	link->dpia_bw_alloc_config.nrd_max_lane_count = 0;
 	link->dpia_bw_alloc_config.nrd_max_link_rate = 0;
 	for (int i = 0; i < MAX_SINKS_PER_LINK; i++)
@@ -243,20 +242,20 @@ static int get_host_router_total_dp_tunnel_bw(const struct dc *dc, uint8_t hr_in
 static void dpia_bw_alloc_unplug(struct dc_link *link)
 {
 	if (link) {
-		DC_LOG_DEBUG("%s: resetting bw alloc config for link(%d)\n",
+		DC_LOG_DEBUG("%s: resetting BW alloc config for link(%d)\n",
 			__func__, link->link_index);
 		reset_bw_alloc_struct(link);
 	}
 }
 
-static void set_usb4_req_bw_req(struct dc_link *link, int req_bw)
+static void link_dpia_send_bw_alloc_request(struct dc_link *link, int req_bw)
 {
 	uint8_t requested_bw;
 	uint32_t temp;
 
 	/* Error check whether request bw greater than allocated */
 	if (req_bw > link->dpia_bw_alloc_config.estimated_bw) {
-		DC_LOG_ERROR("%s: Request bw greater than estimated bw for link(%d)\n",
+		DC_LOG_ERROR("%s: Request BW greater than estimated BW for link(%d)\n",
 			__func__, link->link_index);
 		req_bw = link->dpia_bw_alloc_config.estimated_bw;
 	}
@@ -271,32 +270,17 @@ static void set_usb4_req_bw_req(struct dc_link *link, int req_bw)
 	/* Error check whether requested and allocated are equal */
 	req_bw = requested_bw * (Kbps_TO_Gbps / link->dpia_bw_alloc_config.bw_granularity);
 	if (req_bw && (req_bw == link->dpia_bw_alloc_config.allocated_bw)) {
-		DC_LOG_ERROR("%s: Request bw equals to allocated bw for link(%d)\n",
+		DC_LOG_ERROR("%s: Request BW equals to allocated BW for link(%d)\n",
 			__func__, link->link_index);
 	}
 
-	link->dpia_bw_alloc_config.response_ready = false; // Reset flag
-	core_link_write_dpcd(
-		link,
-		REQUESTED_BW,
+	core_link_write_dpcd(link, REQUESTED_BW,
 		&requested_bw,
 		sizeof(uint8_t));
 }
 
-/*
- * Return the response_ready flag from dc_link struct
- *
- * @link: pointer to the dc_link struct instance
- *
- * return: response_ready flag from dc_link struct
- */
-static bool get_cm_response_ready_flag(struct dc_link *link)
-{
-	return link->dpia_bw_alloc_config.response_ready;
-}
-
 // ------------------------------------------------------------------
-//					PUBLIC FUNCTIONS
+// PUBLIC FUNCTIONS
 // ------------------------------------------------------------------
 bool link_dp_dpia_set_dptx_usb4_bw_alloc_support(struct dc_link *link)
 {
@@ -370,9 +354,15 @@ void link_dp_dpia_handle_bw_alloc_status(struct dc_link *link, uint8_t status)
 		DC_LOG_DEBUG("%s: BW Allocation request succeeded on link(%d)",
 				__func__, link->link_index);
 	} else if (status & DP_TUNNELING_BW_REQUEST_FAILED) {
+		link->dpia_bw_alloc_config.estimated_bw = get_estimated_bw(link);
+
 		DC_LOG_DEBUG("%s: BW Allocation request failed on link(%d)  allocated/estimated BW=%d",
 				__func__, link->link_index, link->dpia_bw_alloc_config.estimated_bw);
+
+		link_dpia_send_bw_alloc_request(link, link->dpia_bw_alloc_config.estimated_bw);
 	} else if (status & DP_TUNNELING_ESTIMATED_BW_CHANGED) {
+		link->dpia_bw_alloc_config.estimated_bw = get_estimated_bw(link);
+
 		DC_LOG_DEBUG("%s: Estimated BW changed on link(%d)  new estimated BW=%d",
 				__func__, link->link_index, link->dpia_bw_alloc_config.estimated_bw);
 	}
@@ -382,141 +372,36 @@ void link_dp_dpia_handle_bw_alloc_status(struct dc_link *link, uint8_t status)
 		&status, sizeof(status));
 }
 
-void dpia_handle_bw_alloc_response(struct dc_link *link, uint8_t bw, uint8_t result)
+/*
+ * Handle the DP Bandwidth allocation for DPIA
+ *
+ */
+void dpia_handle_usb4_bandwidth_allocation_for_link(struct dc_link *link, int peak_bw)
 {
-	int bw_needed = 0;
-	int estimated = 0;
-
-	if (!get_bw_alloc_proceed_flag((link)))
-		return;
-
-	switch (result) {
-
-	case DPIA_BW_REQ_FAILED:
-
-		/*
-		 * Ideally, we shouldn't run into this case as we always validate available
-		 * bandwidth and request within that limit
-		 */
-		estimated = bw * (Kbps_TO_Gbps / link->dpia_bw_alloc_config.bw_granularity);
-
-		DC_LOG_ERROR("%s: BW REQ FAILURE for DP-TX Request for link(%d)\n",
-			__func__, link->link_index);
-		DC_LOG_ERROR("%s: current estimated_bw(%d), new estimated_bw(%d)\n",
-			__func__, link->dpia_bw_alloc_config.estimated_bw, estimated);
-
-		/* Update the new Estimated BW value updated by CM */
-		link->dpia_bw_alloc_config.estimated_bw = estimated;
-
-		/* Allocate the previously requested bandwidth */
-		set_usb4_req_bw_req(link, link->dpia_bw_alloc_config.estimated_bw);
-
-		/*
-		 * If FAIL then it is either:
-		 * 1. Due to DP-Tx trying to allocate more than available i.e. it failed locally
-		 *    => get estimated and allocate that
-		 * 2. Due to the fact that DP-Tx tried to allocated ESTIMATED BW and failed then
-		 *    CM will have to update 0xE0023 with new ESTIMATED BW value.
-		 */
-		break;
-
-	case DPIA_BW_REQ_SUCCESS:
-
-		bw_needed = bw * (Kbps_TO_Gbps / link->dpia_bw_alloc_config.bw_granularity);
-
-		DC_LOG_DEBUG("%s: BW REQ SUCCESS for DP-TX Request for link(%d)\n",
-			__func__, link->link_index);
-		DC_LOG_DEBUG("%s: current allocated_bw(%d), new allocated_bw(%d)\n",
-			__func__, link->dpia_bw_alloc_config.allocated_bw, bw_needed);
-
-		link->dpia_bw_alloc_config.allocated_bw = bw_needed;
+	if (link && link->ep_type == DISPLAY_ENDPOINT_USB4_DPIA && link->dpia_bw_alloc_config.bw_alloc_enabled) {
+		//1. Hot Plug
+		if (link->hpd_status && peak_bw > 0) {
+			// If DP over USB4 then we need to check BW allocation
+			link->dpia_bw_alloc_config.link_max_bw = peak_bw;
 
-		link->dpia_bw_alloc_config.response_ready = true;
-		break;
-
-	case DPIA_EST_BW_CHANGED:
-
-		estimated = bw * (Kbps_TO_Gbps / link->dpia_bw_alloc_config.bw_granularity);
-
-		DC_LOG_DEBUG("%s: ESTIMATED BW CHANGED for link(%d)\n",
-			__func__, link->link_index);
-		DC_LOG_DEBUG("%s: current estimated_bw(%d), new estimated_bw(%d)\n",
-			__func__, link->dpia_bw_alloc_config.estimated_bw, estimated);
-
-		link->dpia_bw_alloc_config.estimated_bw = estimated;
-		break;
-
-	case DPIA_BW_ALLOC_CAPS_CHANGED:
-
-		DC_LOG_ERROR("%s: BW ALLOC CAPABILITY CHANGED to Disabled for link(%d)\n",
-			__func__, link->link_index);
-		link->dpia_bw_alloc_config.bw_alloc_enabled = false;
-		break;
+			link_dpia_send_bw_alloc_request(link, peak_bw);
+		}
+		//2. Cold Unplug
+		else if (!link->hpd_status)
+			dpia_bw_alloc_unplug(link);
 	}
 }
-int dpia_handle_usb4_bandwidth_allocation_for_link(struct dc_link *link, int peak_bw)
-{
-	int ret = 0;
-	uint8_t timeout = 10;
-
-	if (!(link && DISPLAY_ENDPOINT_USB4_DPIA == link->ep_type
-			&& link->dpia_bw_alloc_config.bw_alloc_enabled))
-		goto out;
-
-	//1. Hot Plug
-	if (link->hpd_status && peak_bw > 0) {
-
-		// If DP over USB4 then we need to check BW allocation
-		link->dpia_bw_alloc_config.link_max_bw = peak_bw;
-		set_usb4_req_bw_req(link, link->dpia_bw_alloc_config.link_max_bw);
 
-		do {
-			if (timeout > 0)
-				timeout--;
-			else
-				break;
-			msleep(10);
-		} while (!get_cm_response_ready_flag(link));
-
-		if (!timeout)
-			ret = 0;// ERROR TIMEOUT waiting for response for allocating bw
-		else if (link->dpia_bw_alloc_config.allocated_bw > 0)
-			ret = link->dpia_bw_alloc_config.allocated_bw;
-	}
-	//2. Cold Unplug
-	else if (!link->hpd_status)
-		dpia_bw_alloc_unplug(link);
-
-out:
-	return ret;
-}
-bool link_dp_dpia_allocate_usb4_bandwidth_for_stream(struct dc_link *link, int req_bw)
+void link_dp_dpia_allocate_usb4_bandwidth_for_stream(struct dc_link *link, int req_bw)
 {
-	bool ret = false;
-	uint8_t timeout = 10;
-
 	DC_LOG_DEBUG("%s: ENTER: link(%d), hpd_status(%d), current allocated_bw(%d), req_bw(%d)\n",
 		__func__, link->link_index, link->hpd_status,
 		link->dpia_bw_alloc_config.allocated_bw, req_bw);
 
-	if (!get_bw_alloc_proceed_flag(link))
-		goto out;
-
-	set_usb4_req_bw_req(link, req_bw);
-	do {
-		if (timeout > 0)
-			timeout--;
-		else
-			break;
-		msleep(10);
-	} while (!get_cm_response_ready_flag(link));
-
-	if (timeout)
-		ret = true;
-
-out:
-	DC_LOG_DEBUG("%s: EXIT: timeout(%d), ret(%d)\n", __func__, timeout, ret);
-	return ret;
+	if (link_dp_is_bw_alloc_available(link))
+		link_dpia_send_bw_alloc_request(link, req_bw);
+	else
+		DC_LOG_DEBUG("%s:  Not able to send the BW Allocation request", __func__);
 }
 
 bool dpia_validate_usb4_bw(struct dc_link **link, int *bw_needed_per_dpia, const unsigned int num_dpias)
@@ -567,7 +452,7 @@ int link_dp_dpia_get_dp_overhead_in_dp_tunneling(struct dc_link *link)
 {
 	int dp_overhead = 0, link_mst_overhead = 0;
 
-	if (!get_bw_alloc_proceed_flag((link)))
+	if (!link_dp_is_bw_alloc_available(link))
 		return dp_overhead;
 
 	/* if its mst link, add MTPH overhead */
diff --git a/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_dpia_bw.h b/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_dpia_bw.h
index 1b240a2f6ce0..6df9b946b00f 100644
--- a/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_dpia_bw.h
+++ b/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_dpia_bw.h
@@ -59,9 +59,8 @@ bool link_dp_dpia_set_dptx_usb4_bw_alloc_support(struct dc_link *link);
  * @link: pointer to the dc_link struct instance
  * @req_bw: Bw requested by the stream
  *
- * return: true if allocated successfully
  */
-bool link_dp_dpia_allocate_usb4_bandwidth_for_stream(struct dc_link *link, int req_bw);
+void link_dp_dpia_allocate_usb4_bandwidth_for_stream(struct dc_link *link, int req_bw);
 
 /*
  * Handle the USB4 BW Allocation related functionality here:
@@ -71,21 +70,8 @@ bool link_dp_dpia_allocate_usb4_bandwidth_for_stream(struct dc_link *link, int r
  * @link: pointer to the dc_link struct instance
  * @peak_bw: Peak bw used by the link/sink
  *
- * return: allocated bw else return 0
  */
-int dpia_handle_usb4_bandwidth_allocation_for_link(struct dc_link *link, int peak_bw);
-
-/*
- * Handle function for when the status of the Request above is complete.
- * We will find out the result of allocating on CM and update structs.
- *
- * @link: pointer to the dc_link struct instance
- * @bw: Allocated or Estimated BW depending on the result
- * @result: Response type
- *
- * return: none
- */
-void dpia_handle_bw_alloc_response(struct dc_link *link, uint8_t bw, uint8_t result);
+void dpia_handle_usb4_bandwidth_allocation_for_link(struct dc_link *link, int peak_bw);
 
 /*
  * Handle the validation of total BW here and confirm that the bw used by each
diff --git a/drivers/gpu/drm/amd/display/dmub/dmub_srv.h b/drivers/gpu/drm/amd/display/dmub/dmub_srv.h
index 203e3a440845..4e0efff92dca 100644
--- a/drivers/gpu/drm/amd/display/dmub/dmub_srv.h
+++ b/drivers/gpu/drm/amd/display/dmub/dmub_srv.h
@@ -51,8 +51,8 @@
  * for the cache windows.
  *
  * The call to dmub_srv_hw_init() programs the DMCUB registers to prepare
- * for command submission. Commands can be queued via dmub_srv_fb_cmd_queue()
- * and executed via dmub_srv_fb_cmd_execute().
+ * for command submission. Commands can be queued via dmub_srv_cmd_queue()
+ * and executed via dmub_srv_cmd_execute().
  *
  * If the queue is full the dmub_srv_wait_for_idle() call can be used to
  * wait until the queue has been cleared.
@@ -170,13 +170,6 @@ enum dmub_srv_power_state_type {
 	DMUB_POWER_STATE_D3 = 8
 };
 
-/* enum dmub_inbox_cmd_interface type - defines default interface for host->dmub commands */
-enum dmub_inbox_cmd_interface_type {
-	DMUB_CMD_INTERFACE_DEFAULT = 0,
-	DMUB_CMD_INTERFACE_FB = 1,
-	DMUB_CMD_INTERFACE_REG = 2,
-};
-
 /**
  * struct dmub_region - dmub hw memory region
  * @base: base address for region, must be 256 byte aligned
@@ -356,21 +349,6 @@ struct dmub_diagnostic_data {
 	uint8_t is_cw6_enabled : 1;
 };
 
-struct dmub_srv_inbox {
-	/* generic status */
-	uint64_t num_submitted;
-	uint64_t num_reported;
-	union {
-		/* frame buffer mailbox status */
-		struct dmub_rb rb;
-		/* register mailbox status */
-		struct {
-			bool is_pending;
-			bool is_multi_pending;
-		};
-	};
-};
-
 /**
  * struct dmub_srv_base_funcs - Driver specific base callbacks
  */
@@ -484,21 +462,18 @@ struct dmub_srv_hw_funcs {
 	void (*init_reg_offsets)(struct dmub_srv *dmub, struct dc_context *ctx);
 
 	void (*subvp_save_surf_addr)(struct dmub_srv *dmub, const struct dc_plane_address *addr, uint8_t subvp_index);
-
 	void (*send_reg_inbox0_cmd_msg)(struct dmub_srv *dmub,
 			union dmub_rb_cmd *cmd);
 	uint32_t (*read_reg_inbox0_rsp_int_status)(struct dmub_srv *dmub);
 	void (*read_reg_inbox0_cmd_rsp)(struct dmub_srv *dmub,
 			union dmub_rb_cmd *cmd);
 	void (*write_reg_inbox0_rsp_int_ack)(struct dmub_srv *dmub);
-	void (*clear_reg_inbox0_rsp_int_ack)(struct dmub_srv *dmub);
-	void (*enable_reg_inbox0_rsp_int)(struct dmub_srv *dmub, bool enable);
-
 	uint32_t (*read_reg_outbox0_rdy_int_status)(struct dmub_srv *dmub);
 	void (*write_reg_outbox0_rdy_int_ack)(struct dmub_srv *dmub);
 	void (*read_reg_outbox0_msg)(struct dmub_srv *dmub, uint32_t *msg);
 	void (*write_reg_outbox0_rsp)(struct dmub_srv *dmub, uint32_t *rsp);
 	uint32_t (*read_reg_outbox0_rsp_int_status)(struct dmub_srv *dmub);
+	void (*enable_reg_inbox0_rsp_int)(struct dmub_srv *dmub, bool enable);
 	void (*enable_reg_outbox0_rdy_int)(struct dmub_srv *dmub, bool enable);
 };
 
@@ -518,7 +493,6 @@ struct dmub_srv_create_params {
 	enum dmub_asic asic;
 	uint32_t fw_version;
 	bool is_virtual;
-	enum dmub_inbox_cmd_interface_type inbox_type;
 };
 
 /**
@@ -547,9 +521,8 @@ struct dmub_srv {
 	const struct dmub_srv_dcn401_regs *regs_dcn401;
 	struct dmub_srv_base_funcs funcs;
 	struct dmub_srv_hw_funcs hw_funcs;
-	struct dmub_srv_inbox inbox1;
+	struct dmub_rb inbox1_rb;
 	uint32_t inbox1_last_wptr;
-	struct dmub_srv_inbox reg_inbox0;
 	/**
 	 * outbox1_rb is accessed without locks (dal & dc)
 	 * and to be used only in dmub_srv_stat_get_notification()
@@ -569,7 +542,6 @@ struct dmub_srv {
 	struct dmub_fw_meta_info meta_info;
 	struct dmub_feature_caps feature_caps;
 	struct dmub_visual_confirm_color visual_confirm_color;
-	enum dmub_inbox_cmd_interface_type inbox_type;
 
 	enum dmub_srv_power_state_type power_state;
 	struct dmub_diagnostic_data debug;
@@ -727,7 +699,19 @@ enum dmub_status dmub_srv_hw_init(struct dmub_srv *dmub,
 enum dmub_status dmub_srv_hw_reset(struct dmub_srv *dmub);
 
 /**
- * dmub_srv_fb_cmd_queue() - queues a command to the DMUB
+ * dmub_srv_sync_inbox1() - sync sw state with hw state
+ * @dmub: the dmub service
+ *
+ * Sync sw state with hw state when resume from S0i3
+ *
+ * Return:
+ *   DMUB_STATUS_OK - success
+ *   DMUB_STATUS_INVALID - unspecified error
+ */
+enum dmub_status dmub_srv_sync_inbox1(struct dmub_srv *dmub);
+
+/**
+ * dmub_srv_cmd_queue() - queues a command to the DMUB
  * @dmub: the dmub service
  * @cmd: the command to queue
  *
@@ -739,11 +723,11 @@ enum dmub_status dmub_srv_hw_reset(struct dmub_srv *dmub);
  *   DMUB_STATUS_QUEUE_FULL - no remaining room in queue
  *   DMUB_STATUS_INVALID - unspecified error
  */
-enum dmub_status dmub_srv_fb_cmd_queue(struct dmub_srv *dmub,
+enum dmub_status dmub_srv_cmd_queue(struct dmub_srv *dmub,
 				    const union dmub_rb_cmd *cmd);
 
 /**
- * dmub_srv_fb_cmd_execute() - Executes a queued sequence to the dmub
+ * dmub_srv_cmd_execute() - Executes a queued sequence to the dmub
  * @dmub: the dmub service
  *
  * Begins execution of queued commands on the dmub.
@@ -752,7 +736,7 @@ enum dmub_status dmub_srv_fb_cmd_queue(struct dmub_srv *dmub,
  *   DMUB_STATUS_OK - success
  *   DMUB_STATUS_INVALID - unspecified error
  */
-enum dmub_status dmub_srv_fb_cmd_execute(struct dmub_srv *dmub);
+enum dmub_status dmub_srv_cmd_execute(struct dmub_srv *dmub);
 
 /**
  * dmub_srv_wait_for_hw_pwr_up() - Waits for firmware hardware power up is completed
@@ -811,23 +795,6 @@ enum dmub_status dmub_srv_wait_for_phy_init(struct dmub_srv *dmub,
 					    uint32_t timeout_us);
 
 /**
- * dmub_srv_wait_for_pending() - Re-entrant wait for messages currently pending
- * @dmub: the dmub service
- * @timeout_us: the maximum number of microseconds to wait
- *
- * Waits until the commands queued prior to this call are complete.
- * If interfaces remain busy due to additional work being submitted
- * concurrently, this function will not continue to wait.
- *
- * Return:
- *   DMUB_STATUS_OK - success
- *   DMUB_STATUS_TIMEOUT - wait for buffer to flush timed out
- *   DMUB_STATUS_INVALID - unspecified error
- */
-enum dmub_status dmub_srv_wait_for_pending(struct dmub_srv *dmub,
-					uint32_t timeout_us);
-
-/**
  * dmub_srv_wait_for_idle() - Waits for the DMUB to be idle
  * @dmub: the dmub service
  * @timeout_us: the maximum number of microseconds to wait
@@ -925,6 +892,9 @@ enum dmub_status dmub_srv_get_fw_boot_status(struct dmub_srv *dmub,
 enum dmub_status dmub_srv_get_fw_boot_option(struct dmub_srv *dmub,
 					     union dmub_fw_boot_options *option);
 
+enum dmub_status dmub_srv_cmd_with_reply_data(struct dmub_srv *dmub,
+					      union dmub_rb_cmd *cmd);
+
 enum dmub_status dmub_srv_set_skip_panel_power_sequence(struct dmub_srv *dmub,
 					     bool skip);
 
@@ -989,51 +959,35 @@ enum dmub_status dmub_srv_clear_inbox0_ack(struct dmub_srv *dmub);
 void dmub_srv_subvp_save_surf_addr(struct dmub_srv *dmub, const struct dc_plane_address *addr, uint8_t subvp_index);
 
 /**
- * dmub_srv_set_power_state() - Track DC power state in dmub_srv
+ * dmub_srv_send_reg_inbox0_cmd() - send a dmub command and wait for the command
+ * being processed by DMUB.
  * @dmub: The dmub service
- * @power_state: DC power state setting
- *
- * Store DC power state in dmub_srv.  If dmub_srv is in D3, then don't send messages to DMUB
+ * @cmd: The dmub command being sent. If with_replay is true, the function will
+ * update cmd with replied data.
+ * @with_reply: true if DMUB reply needs to be copied back to cmd. false if the
+ * cmd doesn't need to be replied.
+ * @timeout_us: timeout in microseconds.
  *
  * Return:
- *   void
+ * DMUB_STATUS_OK - success
+ * DMUB_STATUS_TIMEOUT - DMUB fails to process the command within the timeout
+ * interval.
  */
-void dmub_srv_set_power_state(struct dmub_srv *dmub, enum dmub_srv_power_state_type dmub_srv_power_state);
+enum dmub_status dmub_srv_send_reg_inbox0_cmd(
+		struct dmub_srv *dmub,
+		union dmub_rb_cmd *cmd,
+		bool with_reply, uint32_t timeout_us);
 
 /**
- * dmub_srv_reg_cmd_execute() - Executes provided command to the dmub
- * @dmub: the dmub service
- * @cmd: the command packet to be executed
- *
- * Executes a single command for the dmub.
- *
- * Return:
- *   DMUB_STATUS_OK - success
- *   DMUB_STATUS_INVALID - unspecified error
- */
-enum dmub_status dmub_srv_reg_cmd_execute(struct dmub_srv *dmub, union dmub_rb_cmd *cmd);
-
-
-/**
- * dmub_srv_cmd_get_response() - Copies return data for command into buffer
- * @dmub: the dmub service
- * @cmd_rsp: response buffer
- *
- * Copies return data for command into buffer
- */
-void dmub_srv_cmd_get_response(struct dmub_srv *dmub,
-		union dmub_rb_cmd *cmd_rsp);
-
-/**
- * dmub_srv_sync_inboxes() - Sync inbox state
- * @dmub: the dmub service
+ * dmub_srv_set_power_state() - Track DC power state in dmub_srv
+ * @dmub: The dmub service
+ * @power_state: DC power state setting
  *
- * Sync inbox state
+ * Store DC power state in dmub_srv.  If dmub_srv is in D3, then don't send messages to DMUB
  *
  * Return:
- *   DMUB_STATUS_OK - success
- *   DMUB_STATUS_INVALID - unspecified error
+ *   void
  */
-enum dmub_status dmub_srv_sync_inboxes(struct dmub_srv *dmub);
+void dmub_srv_set_power_state(struct dmub_srv *dmub, enum dmub_srv_power_state_type dmub_srv_power_state);
 
 #endif /* _DMUB_SRV_H_ */
diff --git a/drivers/gpu/drm/amd/display/dmub/src/dmub_dcn401.c b/drivers/gpu/drm/amd/display/dmub/src/dmub_dcn401.c
index 731ca9b6a6cf..e67f7c4784eb 100644
--- a/drivers/gpu/drm/amd/display/dmub/src/dmub_dcn401.c
+++ b/drivers/gpu/drm/amd/display/dmub/src/dmub_dcn401.c
@@ -517,69 +517,28 @@ void dmub_dcn401_send_reg_inbox0_cmd_msg(struct dmub_srv *dmub,
 		union dmub_rb_cmd *cmd)
 {
 	uint32_t *dwords = (uint32_t *)cmd;
-	int32_t payload_size_bytes = cmd->cmd_common.header.payload_bytes;
-	uint32_t msg_index;
-	static_assert(sizeof(*cmd) == 64, "DMUB command size mismatch");
-
-	/* read remaining data based on payload size */
-	for (msg_index = 0; msg_index < 15; msg_index++) {
-		if (payload_size_bytes <= msg_index * 4) {
-			break;
-		}
 
-		switch (msg_index) {
-		case 0:
-			REG_WRITE(DMCUB_REG_INBOX0_MSG0, dwords[msg_index + 1]);
-			break;
-		case 1:
-			REG_WRITE(DMCUB_REG_INBOX0_MSG1, dwords[msg_index + 1]);
-			break;
-		case 2:
-			REG_WRITE(DMCUB_REG_INBOX0_MSG2, dwords[msg_index + 1]);
-			break;
-		case 3:
-			REG_WRITE(DMCUB_REG_INBOX0_MSG3, dwords[msg_index + 1]);
-			break;
-		case 4:
-			REG_WRITE(DMCUB_REG_INBOX0_MSG4, dwords[msg_index + 1]);
-			break;
-		case 5:
-			REG_WRITE(DMCUB_REG_INBOX0_MSG5, dwords[msg_index + 1]);
-			break;
-		case 6:
-			REG_WRITE(DMCUB_REG_INBOX0_MSG6, dwords[msg_index + 1]);
-			break;
-		case 7:
-			REG_WRITE(DMCUB_REG_INBOX0_MSG7, dwords[msg_index + 1]);
-			break;
-		case 8:
-			REG_WRITE(DMCUB_REG_INBOX0_MSG8, dwords[msg_index + 1]);
-			break;
-		case 9:
-			REG_WRITE(DMCUB_REG_INBOX0_MSG9, dwords[msg_index + 1]);
-			break;
-		case 10:
-			REG_WRITE(DMCUB_REG_INBOX0_MSG10, dwords[msg_index + 1]);
-			break;
-		case 11:
-			REG_WRITE(DMCUB_REG_INBOX0_MSG11, dwords[msg_index + 1]);
-			break;
-		case 12:
-			REG_WRITE(DMCUB_REG_INBOX0_MSG12, dwords[msg_index + 1]);
-			break;
-		case 13:
-			REG_WRITE(DMCUB_REG_INBOX0_MSG13, dwords[msg_index + 1]);
-			break;
-		case 14:
-			REG_WRITE(DMCUB_REG_INBOX0_MSG14, dwords[msg_index + 1]);
-			break;
-		}
-	}
+	static_assert(sizeof(*cmd) == 64, "DMUB command size mismatch");
 
+	REG_WRITE(DMCUB_REG_INBOX0_MSG0, dwords[0]);
+	REG_WRITE(DMCUB_REG_INBOX0_MSG1, dwords[1]);
+	REG_WRITE(DMCUB_REG_INBOX0_MSG2, dwords[2]);
+	REG_WRITE(DMCUB_REG_INBOX0_MSG3, dwords[3]);
+	REG_WRITE(DMCUB_REG_INBOX0_MSG4, dwords[4]);
+	REG_WRITE(DMCUB_REG_INBOX0_MSG5, dwords[5]);
+	REG_WRITE(DMCUB_REG_INBOX0_MSG6, dwords[6]);
+	REG_WRITE(DMCUB_REG_INBOX0_MSG7, dwords[7]);
+	REG_WRITE(DMCUB_REG_INBOX0_MSG8, dwords[8]);
+	REG_WRITE(DMCUB_REG_INBOX0_MSG9, dwords[9]);
+	REG_WRITE(DMCUB_REG_INBOX0_MSG10, dwords[10]);
+	REG_WRITE(DMCUB_REG_INBOX0_MSG11, dwords[11]);
+	REG_WRITE(DMCUB_REG_INBOX0_MSG12, dwords[12]);
+	REG_WRITE(DMCUB_REG_INBOX0_MSG13, dwords[13]);
+	REG_WRITE(DMCUB_REG_INBOX0_MSG14, dwords[14]);
 	/* writing to INBOX RDY register will trigger DMUB REG INBOX0 RDY
 	 * interrupt.
 	 */
-	REG_WRITE(DMCUB_REG_INBOX0_RDY, dwords[0]);
+	REG_WRITE(DMCUB_REG_INBOX0_RDY, dwords[15]);
 }
 
 uint32_t dmub_dcn401_read_reg_inbox0_rsp_int_status(struct dmub_srv *dmub)
@@ -597,39 +556,30 @@ void dmub_dcn401_read_reg_inbox0_cmd_rsp(struct dmub_srv *dmub,
 
 	static_assert(sizeof(*cmd) == 64, "DMUB command size mismatch");
 
-	dwords[0] = REG_READ(DMCUB_REG_INBOX0_RSP);
-	dwords[1] = REG_READ(DMCUB_REG_INBOX0_MSG0);
-	dwords[2] = REG_READ(DMCUB_REG_INBOX0_MSG1);
-	dwords[3] = REG_READ(DMCUB_REG_INBOX0_MSG2);
-	dwords[4] = REG_READ(DMCUB_REG_INBOX0_MSG3);
-	dwords[5] = REG_READ(DMCUB_REG_INBOX0_MSG4);
-	dwords[6] = REG_READ(DMCUB_REG_INBOX0_MSG5);
-	dwords[7] = REG_READ(DMCUB_REG_INBOX0_MSG6);
-	dwords[8] = REG_READ(DMCUB_REG_INBOX0_MSG7);
-	dwords[9] = REG_READ(DMCUB_REG_INBOX0_MSG8);
-	dwords[10] = REG_READ(DMCUB_REG_INBOX0_MSG9);
-	dwords[11] = REG_READ(DMCUB_REG_INBOX0_MSG10);
-	dwords[12] = REG_READ(DMCUB_REG_INBOX0_MSG11);
-	dwords[13] = REG_READ(DMCUB_REG_INBOX0_MSG12);
-	dwords[14] = REG_READ(DMCUB_REG_INBOX0_MSG13);
-	dwords[15] = REG_READ(DMCUB_REG_INBOX0_MSG14);
+	dwords[0] = REG_READ(DMCUB_REG_INBOX0_MSG0);
+	dwords[1] = REG_READ(DMCUB_REG_INBOX0_MSG1);
+	dwords[2] = REG_READ(DMCUB_REG_INBOX0_MSG2);
+	dwords[3] = REG_READ(DMCUB_REG_INBOX0_MSG3);
+	dwords[4] = REG_READ(DMCUB_REG_INBOX0_MSG4);
+	dwords[5] = REG_READ(DMCUB_REG_INBOX0_MSG5);
+	dwords[6] = REG_READ(DMCUB_REG_INBOX0_MSG6);
+	dwords[7] = REG_READ(DMCUB_REG_INBOX0_MSG7);
+	dwords[8] = REG_READ(DMCUB_REG_INBOX0_MSG8);
+	dwords[9] = REG_READ(DMCUB_REG_INBOX0_MSG9);
+	dwords[10] = REG_READ(DMCUB_REG_INBOX0_MSG10);
+	dwords[11] = REG_READ(DMCUB_REG_INBOX0_MSG11);
+	dwords[12] = REG_READ(DMCUB_REG_INBOX0_MSG12);
+	dwords[13] = REG_READ(DMCUB_REG_INBOX0_MSG13);
+	dwords[14] = REG_READ(DMCUB_REG_INBOX0_MSG14);
+	dwords[15] = REG_READ(DMCUB_REG_INBOX0_RSP);
 }
 
 void dmub_dcn401_write_reg_inbox0_rsp_int_ack(struct dmub_srv *dmub)
 {
 	REG_UPDATE(HOST_INTERRUPT_CSR, HOST_REG_INBOX0_RSP_INT_ACK, 1);
-}
-
-void dmub_dcn401_clear_reg_inbox0_rsp_int_ack(struct dmub_srv *dmub)
-{
 	REG_UPDATE(HOST_INTERRUPT_CSR, HOST_REG_INBOX0_RSP_INT_ACK, 0);
 }
 
-void dmub_dcn401_enable_reg_inbox0_rsp_int(struct dmub_srv *dmub, bool enable)
-{
-	REG_UPDATE(HOST_INTERRUPT_CSR, HOST_REG_INBOX0_RSP_INT_EN, enable ? 1:0);
-}
-
 void dmub_dcn401_write_reg_outbox0_rdy_int_ack(struct dmub_srv *dmub)
 {
 	REG_UPDATE(HOST_INTERRUPT_CSR, HOST_REG_OUTBOX0_RDY_INT_ACK, 1);
@@ -654,6 +604,11 @@ uint32_t dmub_dcn401_read_reg_outbox0_rsp_int_status(struct dmub_srv *dmub)
 	return status;
 }
 
+void dmub_dcn401_enable_reg_inbox0_rsp_int(struct dmub_srv *dmub, bool enable)
+{
+	REG_UPDATE(HOST_INTERRUPT_CSR, HOST_REG_INBOX0_RSP_INT_EN, enable ? 1:0);
+}
+
 void dmub_dcn401_enable_reg_outbox0_rdy_int(struct dmub_srv *dmub, bool enable)
 {
 	REG_UPDATE(HOST_INTERRUPT_CSR, HOST_REG_OUTBOX0_RDY_INT_EN, enable ? 1:0);
diff --git a/drivers/gpu/drm/amd/display/dmub/src/dmub_dcn401.h b/drivers/gpu/drm/amd/display/dmub/src/dmub_dcn401.h
index 88c3a44d67d9..c35be52676f6 100644
--- a/drivers/gpu/drm/amd/display/dmub/src/dmub_dcn401.h
+++ b/drivers/gpu/drm/amd/display/dmub/src/dmub_dcn401.h
@@ -277,13 +277,11 @@ uint32_t dmub_dcn401_read_reg_inbox0_rsp_int_status(struct dmub_srv *dmub);
 void dmub_dcn401_read_reg_inbox0_cmd_rsp(struct dmub_srv *dmub,
 		union dmub_rb_cmd *cmd);
 void dmub_dcn401_write_reg_inbox0_rsp_int_ack(struct dmub_srv *dmub);
-void dmub_dcn401_clear_reg_inbox0_rsp_int_ack(struct dmub_srv *dmub);
-void dmub_dcn401_enable_reg_inbox0_rsp_int(struct dmub_srv *dmub, bool enable);
-
 void dmub_dcn401_write_reg_outbox0_rdy_int_ack(struct dmub_srv *dmub);
 void dmub_dcn401_read_reg_outbox0_msg(struct dmub_srv *dmub, uint32_t *msg);
 void dmub_dcn401_write_reg_outbox0_rsp(struct dmub_srv *dmub, uint32_t *msg);
 uint32_t dmub_dcn401_read_reg_outbox0_rsp_int_status(struct dmub_srv *dmub);
+void dmub_dcn401_enable_reg_inbox0_rsp_int(struct dmub_srv *dmub, bool enable);
 void dmub_dcn401_enable_reg_outbox0_rdy_int(struct dmub_srv *dmub, bool enable);
 uint32_t dmub_dcn401_read_reg_outbox0_rdy_int_status(struct dmub_srv *dmub);
 
diff --git a/drivers/gpu/drm/amd/display/dmub/src/dmub_srv.c b/drivers/gpu/drm/amd/display/dmub/src/dmub_srv.c
index 713576a1f6fa..ae8133816b43 100644
--- a/drivers/gpu/drm/amd/display/dmub/src/dmub_srv.c
+++ b/drivers/gpu/drm/amd/display/dmub/src/dmub_srv.c
@@ -157,9 +157,6 @@ static bool dmub_srv_hw_setup(struct dmub_srv *dmub, enum dmub_asic asic)
 {
 	struct dmub_srv_hw_funcs *funcs = &dmub->hw_funcs;
 
-	/* default to specifying now inbox type */
-	enum dmub_inbox_cmd_interface_type default_inbox_type = DMUB_CMD_INTERFACE_DEFAULT;
-
 	switch (asic) {
 	case DMUB_ASIC_DCN20:
 	case DMUB_ASIC_DCN21:
@@ -398,15 +395,10 @@ static bool dmub_srv_hw_setup(struct dmub_srv *dmub, enum dmub_asic asic)
 
 		funcs->get_current_time = dmub_dcn401_get_current_time;
 		funcs->get_diagnostic_data = dmub_dcn401_get_diagnostic_data;
-
 		funcs->send_reg_inbox0_cmd_msg = dmub_dcn401_send_reg_inbox0_cmd_msg;
 		funcs->read_reg_inbox0_rsp_int_status = dmub_dcn401_read_reg_inbox0_rsp_int_status;
 		funcs->read_reg_inbox0_cmd_rsp = dmub_dcn401_read_reg_inbox0_cmd_rsp;
 		funcs->write_reg_inbox0_rsp_int_ack = dmub_dcn401_write_reg_inbox0_rsp_int_ack;
-		funcs->clear_reg_inbox0_rsp_int_ack = dmub_dcn401_clear_reg_inbox0_rsp_int_ack;
-		funcs->enable_reg_inbox0_rsp_int = dmub_dcn401_enable_reg_inbox0_rsp_int;
-		default_inbox_type = DMUB_CMD_INTERFACE_FB; // still default to FB for now
-
 		funcs->write_reg_outbox0_rdy_int_ack = dmub_dcn401_write_reg_outbox0_rdy_int_ack;
 		funcs->read_reg_outbox0_msg = dmub_dcn401_read_reg_outbox0_msg;
 		funcs->write_reg_outbox0_rsp = dmub_dcn401_write_reg_outbox0_rsp;
@@ -419,20 +411,6 @@ static bool dmub_srv_hw_setup(struct dmub_srv *dmub, enum dmub_asic asic)
 		return false;
 	}
 
-	/* set default inbox type if not overriden */
-	if (dmub->inbox_type == DMUB_CMD_INTERFACE_DEFAULT) {
-		if (default_inbox_type != DMUB_CMD_INTERFACE_DEFAULT) {
-			/* use default inbox type as specified by DCN rev */
-			dmub->inbox_type = default_inbox_type;
-		} else if (funcs->send_reg_inbox0_cmd_msg) {
-			/* prefer reg as default inbox type if present */
-			dmub->inbox_type = DMUB_CMD_INTERFACE_REG;
-		} else {
-			/* use fb as fallback */
-			dmub->inbox_type = DMUB_CMD_INTERFACE_FB;
-		}
-	}
-
 	return true;
 }
 
@@ -448,7 +426,6 @@ enum dmub_status dmub_srv_create(struct dmub_srv *dmub,
 	dmub->asic = params->asic;
 	dmub->fw_version = params->fw_version;
 	dmub->is_virtual = params->is_virtual;
-	dmub->inbox_type = params->inbox_type;
 
 	/* Setup asic dependent hardware funcs. */
 	if (!dmub_srv_hw_setup(dmub, params->asic)) {
@@ -718,7 +695,7 @@ enum dmub_status dmub_srv_hw_init(struct dmub_srv *dmub,
 	inbox1.base = cw4.region.base;
 	inbox1.top = cw4.region.base + DMUB_RB_SIZE;
 	outbox1.base = inbox1.top;
-	outbox1.top = inbox1.top + DMUB_RB_SIZE;
+	outbox1.top = cw4.region.top;
 
 	cw5.offset.quad_part = tracebuff_fb->gpu_addr;
 	cw5.region.base = DMUB_CW5_BASE;
@@ -731,7 +708,7 @@ enum dmub_status dmub_srv_hw_init(struct dmub_srv *dmub,
 	cw6.region.base = DMUB_CW6_BASE;
 	cw6.region.top = cw6.region.base + fw_state_fb->size;
 
-	dmub->fw_state = fw_state_fb->cpu_addr;
+	dmub->fw_state = (void *)((uintptr_t)(fw_state_fb->cpu_addr) + DMUB_DEBUG_FW_STATE_OFFSET);
 
 	region6.offset.quad_part = shared_state_fb->gpu_addr;
 	region6.region.base = DMUB_CW6_BASE;
@@ -760,7 +737,7 @@ enum dmub_status dmub_srv_hw_init(struct dmub_srv *dmub,
 	rb_params.ctx = dmub;
 	rb_params.base_address = mail_fb->cpu_addr;
 	rb_params.capacity = DMUB_RB_SIZE;
-	dmub_rb_init(&dmub->inbox1.rb, &rb_params);
+	dmub_rb_init(&dmub->inbox1_rb, &rb_params);
 
 	// Initialize outbox1 ring buffer
 	rb_params.ctx = dmub;
@@ -791,6 +768,27 @@ enum dmub_status dmub_srv_hw_init(struct dmub_srv *dmub,
 	return DMUB_STATUS_OK;
 }
 
+enum dmub_status dmub_srv_sync_inbox1(struct dmub_srv *dmub)
+{
+	if (!dmub->sw_init)
+		return DMUB_STATUS_INVALID;
+
+	if (dmub->hw_funcs.get_inbox1_rptr && dmub->hw_funcs.get_inbox1_wptr) {
+		uint32_t rptr = dmub->hw_funcs.get_inbox1_rptr(dmub);
+		uint32_t wptr = dmub->hw_funcs.get_inbox1_wptr(dmub);
+
+		if (rptr > dmub->inbox1_rb.capacity || wptr > dmub->inbox1_rb.capacity) {
+			return DMUB_STATUS_HW_FAILURE;
+		} else {
+			dmub->inbox1_rb.rptr = rptr;
+			dmub->inbox1_rb.wrpt = wptr;
+			dmub->inbox1_last_wptr = dmub->inbox1_rb.wrpt;
+		}
+	}
+
+	return DMUB_STATUS_OK;
+}
+
 enum dmub_status dmub_srv_hw_reset(struct dmub_srv *dmub)
 {
 	if (!dmub->sw_init)
@@ -801,13 +799,8 @@ enum dmub_status dmub_srv_hw_reset(struct dmub_srv *dmub)
 
 	/* mailboxes have been reset in hw, so reset the sw state as well */
 	dmub->inbox1_last_wptr = 0;
-	dmub->inbox1.rb.wrpt = 0;
-	dmub->inbox1.rb.rptr = 0;
-	dmub->inbox1.num_reported = 0;
-	dmub->inbox1.num_submitted = 0;
-	dmub->reg_inbox0.num_reported = 0;
-	dmub->reg_inbox0.num_submitted = 0;
-	dmub->reg_inbox0.is_pending = 0;
+	dmub->inbox1_rb.wrpt = 0;
+	dmub->inbox1_rb.rptr = 0;
 	dmub->outbox0_rb.wrpt = 0;
 	dmub->outbox0_rb.rptr = 0;
 	dmub->outbox1_rb.wrpt = 0;
@@ -818,7 +811,7 @@ enum dmub_status dmub_srv_hw_reset(struct dmub_srv *dmub)
 	return DMUB_STATUS_OK;
 }
 
-enum dmub_status dmub_srv_fb_cmd_queue(struct dmub_srv *dmub,
+enum dmub_status dmub_srv_cmd_queue(struct dmub_srv *dmub,
 				    const union dmub_rb_cmd *cmd)
 {
 	if (!dmub->hw_init)
@@ -827,20 +820,18 @@ enum dmub_status dmub_srv_fb_cmd_queue(struct dmub_srv *dmub,
 	if (dmub->power_state != DMUB_POWER_STATE_D0)
 		return DMUB_STATUS_POWER_STATE_D3;
 
-	if (dmub->inbox1.rb.rptr > dmub->inbox1.rb.capacity ||
-	    dmub->inbox1.rb.wrpt > dmub->inbox1.rb.capacity) {
+	if (dmub->inbox1_rb.rptr > dmub->inbox1_rb.capacity ||
+	    dmub->inbox1_rb.wrpt > dmub->inbox1_rb.capacity) {
 		return DMUB_STATUS_HW_FAILURE;
 	}
 
-	if (dmub_rb_push_front(&dmub->inbox1.rb, cmd)) {
-		dmub->inbox1.num_submitted++;
+	if (dmub_rb_push_front(&dmub->inbox1_rb, cmd))
 		return DMUB_STATUS_OK;
-	}
 
 	return DMUB_STATUS_QUEUE_FULL;
 }
 
-enum dmub_status dmub_srv_fb_cmd_execute(struct dmub_srv *dmub)
+enum dmub_status dmub_srv_cmd_execute(struct dmub_srv *dmub)
 {
 	struct dmub_rb flush_rb;
 
@@ -855,13 +846,13 @@ enum dmub_status dmub_srv_fb_cmd_execute(struct dmub_srv *dmub)
 	 * been flushed to framebuffer memory. Otherwise DMCUB might
 	 * read back stale, fully invalid or partially invalid data.
 	 */
-	flush_rb = dmub->inbox1.rb;
+	flush_rb = dmub->inbox1_rb;
 	flush_rb.rptr = dmub->inbox1_last_wptr;
 	dmub_rb_flush_pending(&flush_rb);
 
-		dmub->hw_funcs.set_inbox1_wptr(dmub, dmub->inbox1.rb.wrpt);
+		dmub->hw_funcs.set_inbox1_wptr(dmub, dmub->inbox1_rb.wrpt);
 
-	dmub->inbox1_last_wptr = dmub->inbox1.rb.wrpt;
+	dmub->inbox1_last_wptr = dmub->inbox1_rb.wrpt;
 
 	return DMUB_STATUS_OK;
 }
@@ -919,97 +910,26 @@ enum dmub_status dmub_srv_wait_for_auto_load(struct dmub_srv *dmub,
 	return DMUB_STATUS_TIMEOUT;
 }
 
-static void dmub_srv_update_reg_inbox0_status(struct dmub_srv *dmub)
-{
-	if (dmub->reg_inbox0.is_pending) {
-		dmub->reg_inbox0.is_pending = dmub->hw_funcs.read_reg_inbox0_rsp_int_status &&
-				!dmub->hw_funcs.read_reg_inbox0_rsp_int_status(dmub);
-
-		if (!dmub->reg_inbox0.is_pending) {
-			/* ack the rsp interrupt */
-			if (dmub->hw_funcs.write_reg_inbox0_rsp_int_ack)
-				dmub->hw_funcs.write_reg_inbox0_rsp_int_ack(dmub);
-
-			/* only update the reported count if commands aren't being batched */
-			if (!dmub->reg_inbox0.is_pending && !dmub->reg_inbox0.is_multi_pending) {
-				dmub->reg_inbox0.num_reported = dmub->reg_inbox0.num_submitted;
-			}
-		}
-	}
-}
-
-enum dmub_status dmub_srv_wait_for_pending(struct dmub_srv *dmub,
-					uint32_t timeout_us)
-{
-	uint32_t i;
-	const uint32_t polling_interval_us = 1;
-	struct dmub_srv_inbox scratch_reg_inbox0 = dmub->reg_inbox0;
-	struct dmub_srv_inbox scratch_inbox1 = dmub->inbox1;
-	const volatile struct dmub_srv_inbox *reg_inbox0 = &dmub->reg_inbox0;
-	const volatile struct dmub_srv_inbox *inbox1 = &dmub->inbox1;
-
-	if (!dmub->hw_init ||
-			!dmub->hw_funcs.get_inbox1_wptr)
-		return DMUB_STATUS_INVALID;
-
-	/* take a snapshot of the required mailbox state */
-	scratch_inbox1.rb.wrpt = dmub->hw_funcs.get_inbox1_wptr(dmub);
-
-	for (i = 0; i <= timeout_us; i += polling_interval_us) {
-			scratch_inbox1.rb.rptr = dmub->hw_funcs.get_inbox1_rptr(dmub);
-
-		scratch_reg_inbox0.is_pending = scratch_reg_inbox0.is_pending &&
-				dmub->hw_funcs.read_reg_inbox0_rsp_int_status &&
-				!dmub->hw_funcs.read_reg_inbox0_rsp_int_status(dmub);
-
-		if (scratch_inbox1.rb.rptr > dmub->inbox1.rb.capacity)
-			return DMUB_STATUS_HW_FAILURE;
-
-		/* check current HW state first, but use command submission vs reported as a fallback */
-		if ((dmub_rb_empty(&scratch_inbox1.rb) ||
-				inbox1->num_reported >= scratch_inbox1.num_submitted) &&
-				(!scratch_reg_inbox0.is_pending ||
-				reg_inbox0->num_reported >= scratch_reg_inbox0.num_submitted))
-			return DMUB_STATUS_OK;
-
-		udelay(polling_interval_us);
-	}
-
-	return DMUB_STATUS_TIMEOUT;
-}
-
 enum dmub_status dmub_srv_wait_for_idle(struct dmub_srv *dmub,
 					uint32_t timeout_us)
 {
 	uint32_t i, rptr;
-	const uint32_t polling_interval_us = 1;
 
 	if (!dmub->hw_init)
 		return DMUB_STATUS_INVALID;
 
-	for (i = 0; i < timeout_us; i += polling_interval_us) {
-		/* update inbox1 state */
+	for (i = 0; i <= timeout_us; ++i) {
 			rptr = dmub->hw_funcs.get_inbox1_rptr(dmub);
 
-		if (rptr > dmub->inbox1.rb.capacity)
+		if (rptr > dmub->inbox1_rb.capacity)
 			return DMUB_STATUS_HW_FAILURE;
 
-		if (dmub->inbox1.rb.rptr > rptr) {
-			/* rb wrapped */
-			dmub->inbox1.num_reported += (rptr + dmub->inbox1.rb.capacity - dmub->inbox1.rb.rptr) / DMUB_RB_CMD_SIZE;
-		} else {
-			dmub->inbox1.num_reported += (rptr - dmub->inbox1.rb.rptr) / DMUB_RB_CMD_SIZE;
-		}
-		dmub->inbox1.rb.rptr = rptr;
-
-		/* update reg_inbox0 */
-		dmub_srv_update_reg_inbox0_status(dmub);
+		dmub->inbox1_rb.rptr = rptr;
 
-		/* check for idle */
-		if (dmub_rb_empty(&dmub->inbox1.rb) && !dmub->reg_inbox0.is_pending)
+		if (dmub_rb_empty(&dmub->inbox1_rb))
 			return DMUB_STATUS_OK;
 
-		udelay(polling_interval_us);
+		udelay(1);
 	}
 
 	return DMUB_STATUS_TIMEOUT;
@@ -1120,6 +1040,35 @@ enum dmub_status dmub_srv_set_skip_panel_power_sequence(struct dmub_srv *dmub,
 	return DMUB_STATUS_OK;
 }
 
+enum dmub_status dmub_srv_cmd_with_reply_data(struct dmub_srv *dmub,
+					      union dmub_rb_cmd *cmd)
+{
+	enum dmub_status status = DMUB_STATUS_OK;
+
+	// Queue command
+	status = dmub_srv_cmd_queue(dmub, cmd);
+
+	if (status != DMUB_STATUS_OK)
+		return status;
+
+	// Execute command
+	status = dmub_srv_cmd_execute(dmub);
+
+	if (status != DMUB_STATUS_OK)
+		return status;
+
+	// Wait for DMUB to process command
+	status = dmub_srv_wait_for_idle(dmub, 100000);
+
+	if (status != DMUB_STATUS_OK)
+		return status;
+
+	// Copy data back from ring buffer into command
+	dmub_rb_get_return_data(&dmub->inbox1_rb, cmd);
+
+	return status;
+}
+
 static inline bool dmub_rb_out_trace_buffer_front(struct dmub_rb *rb,
 				 void *entry)
 {
@@ -1211,105 +1160,47 @@ void dmub_srv_subvp_save_surf_addr(struct dmub_srv *dmub, const struct dc_plane_
 	}
 }
 
-void dmub_srv_set_power_state(struct dmub_srv *dmub, enum dmub_srv_power_state_type dmub_srv_power_state)
-{
-	if (!dmub || !dmub->hw_init)
-		return;
-
-	dmub->power_state = dmub_srv_power_state;
-}
 
-enum dmub_status dmub_srv_reg_cmd_execute(struct dmub_srv *dmub, union dmub_rb_cmd *cmd)
+enum dmub_status dmub_srv_send_reg_inbox0_cmd(
+		struct dmub_srv *dmub,
+		union dmub_rb_cmd *cmd,
+		bool with_reply, uint32_t timeout_us)
 {
-	uint32_t num_pending = 0;
-
-	if (!dmub->hw_init)
-		return DMUB_STATUS_INVALID;
-
-	if (dmub->power_state != DMUB_POWER_STATE_D0)
-		return DMUB_STATUS_POWER_STATE_D3;
-
-	if (!dmub->hw_funcs.send_reg_inbox0_cmd_msg ||
-			!dmub->hw_funcs.clear_reg_inbox0_rsp_int_ack)
-		return DMUB_STATUS_INVALID;
-
-	if (dmub->reg_inbox0.num_submitted >= dmub->reg_inbox0.num_reported)
-		num_pending = dmub->reg_inbox0.num_submitted - dmub->reg_inbox0.num_reported;
-	else
-		/* num_submitted wrapped */
-		num_pending = DMUB_REG_INBOX0_RB_MAX_ENTRY -
-				(dmub->reg_inbox0.num_reported - dmub->reg_inbox0.num_submitted);
-
-	if (num_pending >= DMUB_REG_INBOX0_RB_MAX_ENTRY)
-		return DMUB_STATUS_QUEUE_FULL;
+	uint32_t rsp_ready = 0;
+	uint32_t i;
 
-	/* clear last rsp ack and send message */
-	dmub->hw_funcs.clear_reg_inbox0_rsp_int_ack(dmub);
 	dmub->hw_funcs.send_reg_inbox0_cmd_msg(dmub, cmd);
 
-	dmub->reg_inbox0.num_submitted++;
-	dmub->reg_inbox0.is_pending = true;
-	dmub->reg_inbox0.is_multi_pending = cmd->cmd_common.header.multi_cmd_pending;
-
-	return DMUB_STATUS_OK;
-}
-
-void dmub_srv_cmd_get_response(struct dmub_srv *dmub,
-		union dmub_rb_cmd *cmd_rsp)
-{
-	if (dmub) {
-		if (dmub->inbox_type == DMUB_CMD_INTERFACE_REG &&
-				dmub->hw_funcs.read_reg_inbox0_cmd_rsp) {
-			dmub->hw_funcs.read_reg_inbox0_cmd_rsp(dmub, cmd_rsp);
-		} else {
-			dmub_rb_get_return_data(&dmub->inbox1.rb, cmd_rsp);
-		}
+	for (i = 0; i < timeout_us; i++) {
+		rsp_ready = dmub->hw_funcs.read_reg_inbox0_rsp_int_status(dmub);
+		if (rsp_ready)
+			break;
+		udelay(1);
 	}
-}
-
-static enum dmub_status dmub_srv_sync_reg_inbox0(struct dmub_srv *dmub)
-{
-	if (!dmub || !dmub->sw_init)
-		return DMUB_STATUS_INVALID;
-
-	dmub->reg_inbox0.is_pending = 0;
-	dmub->reg_inbox0.is_multi_pending = 0;
-
-	return DMUB_STATUS_OK;
-}
+	if (rsp_ready == 0)
+		return DMUB_STATUS_TIMEOUT;
 
-static enum dmub_status dmub_srv_sync_inbox1(struct dmub_srv *dmub)
-{
-	if (!dmub->sw_init)
-		return DMUB_STATUS_INVALID;
+	if (with_reply)
+		dmub->hw_funcs.read_reg_inbox0_cmd_rsp(dmub, cmd);
 
-	if (dmub->hw_funcs.get_inbox1_rptr && dmub->hw_funcs.get_inbox1_wptr) {
-		uint32_t rptr = dmub->hw_funcs.get_inbox1_rptr(dmub);
-		uint32_t wptr = dmub->hw_funcs.get_inbox1_wptr(dmub);
+	dmub->hw_funcs.write_reg_inbox0_rsp_int_ack(dmub);
 
-		if (rptr > dmub->inbox1.rb.capacity || wptr > dmub->inbox1.rb.capacity) {
-			return DMUB_STATUS_HW_FAILURE;
-		} else {
-			dmub->inbox1.rb.rptr = rptr;
-			dmub->inbox1.rb.wrpt = wptr;
-			dmub->inbox1_last_wptr = dmub->inbox1.rb.wrpt;
-		}
+	/* wait for rsp int status is cleared to initial state before exit */
+	for (; i <= timeout_us; i++) {
+		rsp_ready = dmub->hw_funcs.read_reg_inbox0_rsp_int_status(dmub);
+		if (rsp_ready == 0)
+			break;
+		udelay(1);
 	}
+	ASSERT(rsp_ready == 0);
 
 	return DMUB_STATUS_OK;
 }
 
-enum dmub_status dmub_srv_sync_inboxes(struct dmub_srv *dmub)
+void dmub_srv_set_power_state(struct dmub_srv *dmub, enum dmub_srv_power_state_type dmub_srv_power_state)
 {
-	enum dmub_status status;
-
-	status = dmub_srv_sync_reg_inbox0(dmub);
-	if (status != DMUB_STATUS_OK)
-		return status;
-
-	status = dmub_srv_sync_inbox1(dmub);
-	if (status != DMUB_STATUS_OK)
-		return status;
+	if (!dmub || !dmub->hw_init)
+		return;
 
-	return DMUB_STATUS_OK;
+	dmub->power_state = dmub_srv_power_state;
 }
diff --git a/drivers/gpu/drm/amd/include/amd_shared.h b/drivers/gpu/drm/amd/include/amd_shared.h
index 485b713cfad0..4c95b885d1d0 100644
--- a/drivers/gpu/drm/amd/include/amd_shared.h
+++ b/drivers/gpu/drm/amd/include/amd_shared.h
@@ -358,6 +358,18 @@ enum DC_DEBUG_MASK {
 	 * @DC_DISABLE_CUSTOM_BRIGHTNESS_CURVE: If set, disable support for custom brightness curves
 	 */
 	DC_DISABLE_CUSTOM_BRIGHTNESS_CURVE = 0x40000,
+
+	/**
+	 * @DC_HDCP_LC_FORCE_FW_ENABLE: If set, use HDCP Locality Check FW
+	 * path regardless of reported HW capabilities.
+	 */
+	DC_HDCP_LC_FORCE_FW_ENABLE = 0x80000,
+
+	/**
+	 * @DC_HDCP_LC_ENABLE_SW_FALLBACK If set, upon HDCP Locality Check FW
+	 * path failure, retry using legacy SW path.
+	 */
+	DC_HDCP_LC_ENABLE_SW_FALLBACK = 0x100000,
 };
 
 enum amd_dpm_forced_level;
diff --git a/drivers/gpu/drm/amd/pm/inc/amdgpu_dpm.h b/drivers/gpu/drm/amd/pm/inc/amdgpu_dpm.h
index 9fb26b5c8ae7..f93d287dbf13 100644
--- a/drivers/gpu/drm/amd/pm/inc/amdgpu_dpm.h
+++ b/drivers/gpu/drm/amd/pm/inc/amdgpu_dpm.h
@@ -295,7 +295,8 @@ enum ip_power_state {
 };
 
 /* Used to mask smu debug modes */
-#define SMU_DEBUG_HALT_ON_ERROR		0x1
+#define SMU_DEBUG_HALT_ON_ERROR		BIT(0)
+#define SMU_DEBUG_POOL_USE_VRAM		BIT(1)
 
 #define MAX_SMU_I2C_BUSES       2
 
diff --git a/drivers/gpu/drm/amd/pm/swsmu/amdgpu_smu.c b/drivers/gpu/drm/amd/pm/swsmu/amdgpu_smu.c
index d1aa6075c4f4..033c3229b555 100644
--- a/drivers/gpu/drm/amd/pm/swsmu/amdgpu_smu.c
+++ b/drivers/gpu/drm/amd/pm/swsmu/amdgpu_smu.c
@@ -1027,7 +1027,10 @@ static int smu_alloc_memory_pool(struct smu_context *smu)
 
 	memory_pool->size = pool_size;
 	memory_pool->align = PAGE_SIZE;
-	memory_pool->domain = AMDGPU_GEM_DOMAIN_GTT;
+	memory_pool->domain =
+		(adev->pm.smu_debug_mask & SMU_DEBUG_POOL_USE_VRAM) ?
+			AMDGPU_GEM_DOMAIN_VRAM :
+			AMDGPU_GEM_DOMAIN_GTT;
 
 	switch (pool_size) {
 	case SMU_MEMORY_POOL_SIZE_256_MB:
diff --git a/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_12_ppt.c b/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_12_ppt.c
index 51078accc0f8..238bd71baa6d 100644
--- a/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_12_ppt.c
+++ b/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_12_ppt.c
@@ -478,8 +478,8 @@ ssize_t smu_v13_0_12_get_gpu_metrics(struct smu_context *smu, void **table)
 		}
 	}
 
-	gpu_metrics->xgmi_link_width = SMUQ10_ROUND(metrics->XgmiWidth);
-	gpu_metrics->xgmi_link_speed = SMUQ10_ROUND(metrics->XgmiBitrate);
+	gpu_metrics->xgmi_link_width = metrics->XgmiWidth;
+	gpu_metrics->xgmi_link_speed = metrics->XgmiBitrate;
 
 	gpu_metrics->firmware_timestamp = metrics->Timestamp;
 
diff --git a/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_6_ppt.c b/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_6_ppt.c
index 1e1d8989c77a..682646068000 100644
--- a/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_6_ppt.c
+++ b/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_6_ppt.c
@@ -231,7 +231,11 @@ static const struct cmn2asic_mapping smu_v13_0_6_feature_mask_map[SMU_FEATURE_CO
 	SMU_13_0_6_FEA_MAP(SMU_FEATURE_FW_CTF_BIT, 			FEATURE_FW_CTF),
 	SMU_13_0_6_FEA_MAP(SMU_FEATURE_THERMAL_BIT, 			FEATURE_THERMAL),
 	SMU_13_0_6_FEA_MAP(SMU_FEATURE_XGMI_PER_LINK_PWR_DWN_BIT,	FEATURE_XGMI_PER_LINK_PWR_DOWN),
-	SMU_13_0_6_FEA_MAP(SMU_FEATURE_DF_CSTATE_BIT, 			FEATURE_DF_CSTATE),
+	SMU_13_0_6_FEA_MAP(SMU_FEATURE_DF_CSTATE_BIT,			FEATURE_DF_CSTATE),
+	SMU_13_0_6_FEA_MAP(SMU_FEATURE_DS_VCN_BIT,			FEATURE_DS_VCN),
+	SMU_13_0_6_FEA_MAP(SMU_FEATURE_DS_MP1CLK_BIT,			FEATURE_DS_MP1CLK),
+	SMU_13_0_6_FEA_MAP(SMU_FEATURE_DS_MPIOCLK_BIT,			FEATURE_DS_MPIOCLK),
+	SMU_13_0_6_FEA_MAP(SMU_FEATURE_DS_MP0CLK_BIT,			FEATURE_DS_MP0CLK),
 };
 
 #define TABLE_PMSTATUSLOG             0
@@ -2682,8 +2686,8 @@ static ssize_t smu_v13_0_6_get_gpu_metrics(struct smu_context *smu, void **table
 		}
 	}
 
-	gpu_metrics->xgmi_link_width = SMUQ10_ROUND(GET_METRIC_FIELD(XgmiWidth, version));
-	gpu_metrics->xgmi_link_speed = SMUQ10_ROUND(GET_METRIC_FIELD(XgmiBitrate, version));
+	gpu_metrics->xgmi_link_width = GET_METRIC_FIELD(XgmiWidth, version);
+	gpu_metrics->xgmi_link_speed = GET_METRIC_FIELD(XgmiBitrate, version);
 
 	gpu_metrics->firmware_timestamp = GET_METRIC_FIELD(Timestamp, version);
 
diff --git a/drivers/gpu/drm/amd/pm/swsmu/smu14/smu_v14_0_2_ppt.c b/drivers/gpu/drm/amd/pm/swsmu/smu14/smu_v14_0_2_ppt.c
index 21d1b34ada05..f7cfe1f35cae 100644
--- a/drivers/gpu/drm/amd/pm/swsmu/smu14/smu_v14_0_2_ppt.c
+++ b/drivers/gpu/drm/amd/pm/swsmu/smu14/smu_v14_0_2_ppt.c
@@ -1203,16 +1203,9 @@ static int smu_v14_0_2_print_clk_levels(struct smu_context *smu,
 							 PP_OD_FEATURE_GFXCLK_BIT))
 			break;
 
-		PPTable_t *pptable = smu->smu_table.driver_pptable;
-		const OverDriveLimits_t * const overdrive_upperlimits =
-					&pptable->SkuTable.OverDriveLimitsBasicMax;
-		const OverDriveLimits_t * const overdrive_lowerlimits =
-					&pptable->SkuTable.OverDriveLimitsBasicMin;
-
 		size += sysfs_emit_at(buf, size, "OD_SCLK_OFFSET:\n");
-		size += sysfs_emit_at(buf, size, "0: %dMhz\n1: %uMhz\n",
-					overdrive_lowerlimits->GfxclkFoffset,
-					overdrive_upperlimits->GfxclkFoffset);
+		size += sysfs_emit_at(buf, size, "%dMhz\n",
+					od_table->OverDriveTable.GfxclkFoffset);
 		break;
 
 	case SMU_OD_MCLK:
@@ -1347,12 +1340,8 @@ static int smu_v14_0_2_print_clk_levels(struct smu_context *smu,
 
 		if (smu_v14_0_2_is_od_feature_supported(smu, PP_OD_FEATURE_GFXCLK_BIT)) {
 			smu_v14_0_2_get_od_setting_limits(smu,
-							  PP_OD_FEATURE_GFXCLK_FMIN,
-							  &min_value,
-							  NULL);
-			smu_v14_0_2_get_od_setting_limits(smu,
 							  PP_OD_FEATURE_GFXCLK_FMAX,
-							  NULL,
+							  &min_value,
 							  &max_value);
 			size += sysfs_emit_at(buf, size, "SCLK_OFFSET: %7dMhz %10uMhz\n",
 					      min_value, max_value);
@@ -2460,36 +2449,24 @@ static int smu_v14_0_2_od_edit_dpm_table(struct smu_context *smu,
 			return -ENOTSUPP;
 		}
 
-		for (i = 0; i < size; i += 2) {
-			if (i + 2 > size) {
-				dev_info(adev->dev, "invalid number of input parameters %d\n", size);
-				return -EINVAL;
-			}
-
-			switch (input[i]) {
-			case 1:
-				smu_v14_0_2_get_od_setting_limits(smu,
-								  PP_OD_FEATURE_GFXCLK_FMAX,
-								  &minimum,
-								  &maximum);
-				if (input[i + 1] < minimum ||
-				    input[i + 1] > maximum) {
-					dev_info(adev->dev, "GfxclkFmax (%ld) must be within [%u, %u]!\n",
-						input[i + 1], minimum, maximum);
-					return -EINVAL;
-				}
-
-				od_table->OverDriveTable.GfxclkFoffset = input[i + 1];
-				od_table->OverDriveTable.FeatureCtrlMask |= 1U << PP_OD_FEATURE_GFXCLK_BIT;
-				break;
+		if (size != 1) {
+			dev_info(adev->dev, "invalid number of input parameters %d\n", size);
+			return -EINVAL;
+		}
 
-			default:
-				dev_info(adev->dev, "Invalid SCLK_VDDC_TABLE index: %ld\n", input[i]);
-				dev_info(adev->dev, "Supported indices: [0:min,1:max]\n");
-				return -EINVAL;
-			}
+		smu_v14_0_2_get_od_setting_limits(smu,
+						  PP_OD_FEATURE_GFXCLK_FMAX,
+						  &minimum,
+						  &maximum);
+		if (input[0] < minimum ||
+		    input[0] > maximum) {
+			dev_info(adev->dev, "GfxclkFoffset must be within [%d, %u]!\n",
+				 minimum, maximum);
+			return -EINVAL;
 		}
 
+		od_table->OverDriveTable.GfxclkFoffset = input[0];
+		od_table->OverDriveTable.FeatureCtrlMask |= 1U << PP_OD_FEATURE_GFXCLK_BIT;
 		break;
 
 	case PP_OD_EDIT_MCLK_VDDC_TABLE:
diff --git a/drivers/gpu/drm/radeon/radeon_uvd.c b/drivers/gpu/drm/radeon/radeon_uvd.c
index 058a1c8451b2..ded5747a58d1 100644
--- a/drivers/gpu/drm/radeon/radeon_uvd.c
+++ b/drivers/gpu/drm/radeon/radeon_uvd.c
@@ -961,7 +961,7 @@ int radeon_uvd_calc_upll_dividers(struct radeon_device *rdev,
 	unsigned optimal_score = ~0;
 
 	/* loop through vco from low to high */
-	vco_min = max(max(vco_min, vclk), dclk);
+	vco_min = max3(vco_min, vclk, dclk);
 	for (vco_freq = vco_min; vco_freq <= vco_max; vco_freq += 100) {
 
 		uint64_t fb_div = (uint64_t)vco_freq * fb_factor;