Merge tag 'amd-drm-fixes-6.11-2024-08-14' of https://gitlab.freedesktop.org/agd5f/linux into drm-fixes

amd-drm-fixes-6.11-2024-08-14:

amdgpu:
- Fix MES ring buffer overflow
- DCN 3.5 fix
- DCN 3.2.1 fix
- DP MST fix
- Cursor fixes
- JPEG fixes
- Context ops validation
- MES 12 fixes
- VCN 5.0 fix
- HDP fix

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

From: Alex Deucher <alexander.deucher@amd.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20240814213846.1331827-1-alexander.deucher@amd.com

26 files changed, 420 insertions, 252 deletions

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
index 9aa952f258cf..6dfdff58bffd 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
@@ -1057,6 +1057,9 @@ static int amdgpu_cs_patch_ibs(struct amdgpu_cs_parser *p,
 			r = amdgpu_ring_parse_cs(ring, p, job, ib);
 			if (r)
 				return r;
+
+			if (ib->sa_bo)
+				ib->gpu_addr =  amdgpu_sa_bo_gpu_addr(ib->sa_bo);
 		} else {
 			ib->ptr = (uint32_t *)kptr;
 			r = amdgpu_ring_patch_cs_in_place(ring, p, job, ib);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ctx.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ctx.c
index 5cb33ac99f70..c43d1b6e5d66 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ctx.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ctx.c
@@ -685,16 +685,24 @@ int amdgpu_ctx_ioctl(struct drm_device *dev, void *data,
 
 	switch (args->in.op) {
 	case AMDGPU_CTX_OP_ALLOC_CTX:
+		if (args->in.flags)
+			return -EINVAL;
 		r = amdgpu_ctx_alloc(adev, fpriv, filp, priority, &id);
 		args->out.alloc.ctx_id = id;
 		break;
 	case AMDGPU_CTX_OP_FREE_CTX:
+		if (args->in.flags)
+			return -EINVAL;
 		r = amdgpu_ctx_free(fpriv, id);
 		break;
 	case AMDGPU_CTX_OP_QUERY_STATE:
+		if (args->in.flags)
+			return -EINVAL;
 		r = amdgpu_ctx_query(adev, fpriv, id, &args->out);
 		break;
 	case AMDGPU_CTX_OP_QUERY_STATE2:
+		if (args->in.flags)
+			return -EINVAL;
 		r = amdgpu_ctx_query2(adev, fpriv, id, &args->out);
 		break;
 	case AMDGPU_CTX_OP_GET_STABLE_PSTATE:
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c
index 82452606ae6c..c770cb201e64 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c
@@ -509,6 +509,16 @@ int amdgpu_gfx_disable_kcq(struct amdgpu_device *adev, int xcc_id)
 	int i, r = 0;
 	int j;
 
+	if (adev->enable_mes) {
+		for (i = 0; i < adev->gfx.num_compute_rings; i++) {
+			j = i + xcc_id * adev->gfx.num_compute_rings;
+			amdgpu_mes_unmap_legacy_queue(adev,
+						   &adev->gfx.compute_ring[j],
+						   RESET_QUEUES, 0, 0);
+		}
+		return 0;
+	}
+
 	if (!kiq->pmf || !kiq->pmf->kiq_unmap_queues)
 		return -EINVAL;
 
@@ -551,6 +561,18 @@ int amdgpu_gfx_disable_kgq(struct amdgpu_device *adev, int xcc_id)
 	int i, r = 0;
 	int j;
 
+	if (adev->enable_mes) {
+		if (amdgpu_gfx_is_master_xcc(adev, xcc_id)) {
+			for (i = 0; i < adev->gfx.num_gfx_rings; i++) {
+				j = i + xcc_id * adev->gfx.num_gfx_rings;
+				amdgpu_mes_unmap_legacy_queue(adev,
+						      &adev->gfx.gfx_ring[j],
+						      PREEMPT_QUEUES, 0, 0);
+			}
+		}
+		return 0;
+	}
+
 	if (!kiq->pmf || !kiq->pmf->kiq_unmap_queues)
 		return -EINVAL;
 
@@ -995,7 +1017,7 @@ uint32_t amdgpu_kiq_rreg(struct amdgpu_device *adev, uint32_t reg, uint32_t xcc_
 	if (amdgpu_device_skip_hw_access(adev))
 		return 0;
 
-	if (adev->mes.ring.sched.ready)
+	if (adev->mes.ring[0].sched.ready)
 		return amdgpu_mes_rreg(adev, reg);
 
 	BUG_ON(!ring->funcs->emit_rreg);
@@ -1065,7 +1087,7 @@ void amdgpu_kiq_wreg(struct amdgpu_device *adev, uint32_t reg, uint32_t v, uint3
 	if (amdgpu_device_skip_hw_access(adev))
 		return;
 
-	if (adev->mes.ring.sched.ready) {
+	if (adev->mes.ring[0].sched.ready) {
 		amdgpu_mes_wreg(adev, reg, v);
 		return;
 	}
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c
index c02659025656..b49b3650fd62 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c
@@ -589,7 +589,8 @@ int amdgpu_gmc_allocate_vm_inv_eng(struct amdgpu_device *adev)
 		ring = adev->rings[i];
 		vmhub = ring->vm_hub;
 
-		if (ring == &adev->mes.ring ||
+		if (ring == &adev->mes.ring[0] ||
+		    ring == &adev->mes.ring[1] ||
 		    ring == &adev->umsch_mm.ring)
 			continue;
 
@@ -761,7 +762,7 @@ void amdgpu_gmc_fw_reg_write_reg_wait(struct amdgpu_device *adev,
 	unsigned long flags;
 	uint32_t seq;
 
-	if (adev->mes.ring.sched.ready) {
+	if (adev->mes.ring[0].sched.ready) {
 		amdgpu_mes_reg_write_reg_wait(adev, reg0, reg1,
 					      ref, mask);
 		return;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_mes.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_mes.c
index dac88d2dd70d..1cb1ec7beefe 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_mes.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_mes.c
@@ -135,9 +135,11 @@ int amdgpu_mes_init(struct amdgpu_device *adev)
 	idr_init(&adev->mes.queue_id_idr);
 	ida_init(&adev->mes.doorbell_ida);
 	spin_lock_init(&adev->mes.queue_id_lock);
-	spin_lock_init(&adev->mes.ring_lock);
 	mutex_init(&adev->mes.mutex_hidden);
 
+	for (i = 0; i < AMDGPU_MAX_MES_PIPES; i++)
+		spin_lock_init(&adev->mes.ring_lock[i]);
+
 	adev->mes.total_max_queue = AMDGPU_FENCE_MES_QUEUE_ID_MASK;
 	adev->mes.vmid_mask_mmhub = 0xffffff00;
 	adev->mes.vmid_mask_gfxhub = 0xffffff00;
@@ -163,36 +165,38 @@ int amdgpu_mes_init(struct amdgpu_device *adev)
 			adev->mes.sdma_hqd_mask[i] = 0xfc;
 	}
 
-	r = amdgpu_device_wb_get(adev, &adev->mes.sch_ctx_offs);
-	if (r) {
-		dev_err(adev->dev,
-			"(%d) ring trail_fence_offs wb alloc failed\n", r);
-		goto error_ids;
-	}
-	adev->mes.sch_ctx_gpu_addr =
-		adev->wb.gpu_addr + (adev->mes.sch_ctx_offs * 4);
-	adev->mes.sch_ctx_ptr =
-		(uint64_t *)&adev->wb.wb[adev->mes.sch_ctx_offs];
+	for (i = 0; i < AMDGPU_MAX_MES_PIPES; i++) {
+		r = amdgpu_device_wb_get(adev, &adev->mes.sch_ctx_offs[i]);
+		if (r) {
+			dev_err(adev->dev,
+				"(%d) ring trail_fence_offs wb alloc failed\n",
+				r);
+			goto error;
+		}
+		adev->mes.sch_ctx_gpu_addr[i] =
+			adev->wb.gpu_addr + (adev->mes.sch_ctx_offs[i] * 4);
+		adev->mes.sch_ctx_ptr[i] =
+			(uint64_t *)&adev->wb.wb[adev->mes.sch_ctx_offs[i]];
 
-	r = amdgpu_device_wb_get(adev, &adev->mes.query_status_fence_offs);
-	if (r) {
-		amdgpu_device_wb_free(adev, adev->mes.sch_ctx_offs);
-		dev_err(adev->dev,
-			"(%d) query_status_fence_offs wb alloc failed\n", r);
-		goto error_ids;
+		r = amdgpu_device_wb_get(adev,
+				 &adev->mes.query_status_fence_offs[i]);
+		if (r) {
+			dev_err(adev->dev,
+			      "(%d) query_status_fence_offs wb alloc failed\n",
+			      r);
+			goto error;
+		}
+		adev->mes.query_status_fence_gpu_addr[i] = adev->wb.gpu_addr +
+			(adev->mes.query_status_fence_offs[i] * 4);
+		adev->mes.query_status_fence_ptr[i] =
+			(uint64_t *)&adev->wb.wb[adev->mes.query_status_fence_offs[i]];
 	}
-	adev->mes.query_status_fence_gpu_addr =
-		adev->wb.gpu_addr + (adev->mes.query_status_fence_offs * 4);
-	adev->mes.query_status_fence_ptr =
-		(uint64_t *)&adev->wb.wb[adev->mes.query_status_fence_offs];
 
 	r = amdgpu_device_wb_get(adev, &adev->mes.read_val_offs);
 	if (r) {
-		amdgpu_device_wb_free(adev, adev->mes.sch_ctx_offs);
-		amdgpu_device_wb_free(adev, adev->mes.query_status_fence_offs);
 		dev_err(adev->dev,
 			"(%d) read_val_offs alloc failed\n", r);
-		goto error_ids;
+		goto error;
 	}
 	adev->mes.read_val_gpu_addr =
 		adev->wb.gpu_addr + (adev->mes.read_val_offs * 4);
@@ -212,10 +216,16 @@ int amdgpu_mes_init(struct amdgpu_device *adev)
 error_doorbell:
 	amdgpu_mes_doorbell_free(adev);
 error:
-	amdgpu_device_wb_free(adev, adev->mes.sch_ctx_offs);
-	amdgpu_device_wb_free(adev, adev->mes.query_status_fence_offs);
-	amdgpu_device_wb_free(adev, adev->mes.read_val_offs);
-error_ids:
+	for (i = 0; i < AMDGPU_MAX_MES_PIPES; i++) {
+		if (adev->mes.sch_ctx_ptr[i])
+			amdgpu_device_wb_free(adev, adev->mes.sch_ctx_offs[i]);
+		if (adev->mes.query_status_fence_ptr[i])
+			amdgpu_device_wb_free(adev,
+				      adev->mes.query_status_fence_offs[i]);
+	}
+	if (adev->mes.read_val_ptr)
+		amdgpu_device_wb_free(adev, adev->mes.read_val_offs);
+
 	idr_destroy(&adev->mes.pasid_idr);
 	idr_destroy(&adev->mes.gang_id_idr);
 	idr_destroy(&adev->mes.queue_id_idr);
@@ -226,13 +236,22 @@ error_ids:
 
 void amdgpu_mes_fini(struct amdgpu_device *adev)
 {
+	int i;
+
 	amdgpu_bo_free_kernel(&adev->mes.event_log_gpu_obj,
 			      &adev->mes.event_log_gpu_addr,
 			      &adev->mes.event_log_cpu_addr);
 
-	amdgpu_device_wb_free(adev, adev->mes.sch_ctx_offs);
-	amdgpu_device_wb_free(adev, adev->mes.query_status_fence_offs);
-	amdgpu_device_wb_free(adev, adev->mes.read_val_offs);
+	for (i = 0; i < AMDGPU_MAX_MES_PIPES; i++) {
+		if (adev->mes.sch_ctx_ptr[i])
+			amdgpu_device_wb_free(adev, adev->mes.sch_ctx_offs[i]);
+		if (adev->mes.query_status_fence_ptr[i])
+			amdgpu_device_wb_free(adev,
+				      adev->mes.query_status_fence_offs[i]);
+	}
+	if (adev->mes.read_val_ptr)
+		amdgpu_device_wb_free(adev, adev->mes.read_val_offs);
+
 	amdgpu_mes_doorbell_free(adev);
 
 	idr_destroy(&adev->mes.pasid_idr);
@@ -1499,7 +1518,7 @@ int amdgpu_mes_init_microcode(struct amdgpu_device *adev, int pipe)
 
 	amdgpu_ucode_ip_version_decode(adev, GC_HWIP, ucode_prefix,
 				       sizeof(ucode_prefix));
-	if (adev->enable_uni_mes && pipe == AMDGPU_MES_SCHED_PIPE) {
+	if (adev->enable_uni_mes) {
 		snprintf(fw_name, sizeof(fw_name),
 			 "amdgpu/%s_uni_mes.bin", ucode_prefix);
 	} else if (amdgpu_ip_version(adev, GC_HWIP, 0) >= IP_VERSION(11, 0, 0) &&
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_mes.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_mes.h
index 2d659c612f03..0bc837dab578 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_mes.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_mes.h
@@ -82,8 +82,8 @@ struct amdgpu_mes {
 	uint64_t                        default_process_quantum;
 	uint64_t                        default_gang_quantum;
 
-	struct amdgpu_ring              ring;
-	spinlock_t                      ring_lock;
+	struct amdgpu_ring              ring[AMDGPU_MAX_MES_PIPES];
+	spinlock_t                      ring_lock[AMDGPU_MAX_MES_PIPES];
 
 	const struct firmware           *fw[AMDGPU_MAX_MES_PIPES];
 
@@ -112,12 +112,12 @@ struct amdgpu_mes {
 	uint32_t                        gfx_hqd_mask[AMDGPU_MES_MAX_GFX_PIPES];
 	uint32_t                        sdma_hqd_mask[AMDGPU_MES_MAX_SDMA_PIPES];
 	uint32_t                        aggregated_doorbells[AMDGPU_MES_PRIORITY_NUM_LEVELS];
-	uint32_t                        sch_ctx_offs;
-	uint64_t			sch_ctx_gpu_addr;
-	uint64_t			*sch_ctx_ptr;
-	uint32_t			query_status_fence_offs;
-	uint64_t			query_status_fence_gpu_addr;
-	uint64_t			*query_status_fence_ptr;
+	uint32_t                        sch_ctx_offs[AMDGPU_MAX_MES_PIPES];
+	uint64_t			sch_ctx_gpu_addr[AMDGPU_MAX_MES_PIPES];
+	uint64_t			*sch_ctx_ptr[AMDGPU_MAX_MES_PIPES];
+	uint32_t			query_status_fence_offs[AMDGPU_MAX_MES_PIPES];
+	uint64_t			query_status_fence_gpu_addr[AMDGPU_MAX_MES_PIPES];
+	uint64_t			*query_status_fence_ptr[AMDGPU_MAX_MES_PIPES];
 	uint32_t                        read_val_offs;
 	uint64_t			read_val_gpu_addr;
 	uint32_t			*read_val_ptr;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.c
index ad49cecb20b8..e6344a6b0a9f 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.c
@@ -212,6 +212,8 @@ int amdgpu_ring_init(struct amdgpu_device *adev, struct amdgpu_ring *ring,
 	 */
 	if (ring->funcs->type == AMDGPU_RING_TYPE_KIQ)
 		sched_hw_submission = max(sched_hw_submission, 256);
+	if (ring->funcs->type == AMDGPU_RING_TYPE_MES)
+		sched_hw_submission = 8;
 	else if (ring == &adev->sdma.instance[0].page)
 		sched_hw_submission = 256;
 
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h
index 1a5439abd1a0..c87d68d4be53 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h
@@ -461,8 +461,11 @@ struct amdgpu_vcn5_fw_shared {
 	struct amdgpu_fw_shared_unified_queue_struct sq;
 	uint8_t pad1[8];
 	struct amdgpu_fw_shared_fw_logging fw_log;
+	uint8_t pad2[20];
 	struct amdgpu_fw_shared_rb_setup rb_setup;
-	uint8_t pad2[4];
+	struct amdgpu_fw_shared_smu_interface_info smu_dpm_interface;
+	struct amdgpu_fw_shared_drm_key_wa drm_key_wa;
+	uint8_t pad3[9];
 };
 
 #define VCN_BLOCK_ENCODE_DISABLE_MASK 0x80
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c
index 111c380f929b..b287a82e6177 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c
@@ -858,7 +858,7 @@ void amdgpu_virt_post_reset(struct amdgpu_device *adev)
 		adev->gfx.is_poweron = false;
 	}
 
-	adev->mes.ring.sched.ready = false;
+	adev->mes.ring[0].sched.ready = false;
 }
 
 bool amdgpu_virt_fw_load_skip_check(struct amdgpu_device *adev, uint32_t ucode_id)
diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v12_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v12_0.c
index 506fa8003388..2c611b8577a7 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v12_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v12_0.c
@@ -3546,33 +3546,9 @@ static int gfx_v12_0_hw_init(void *handle)
 	return r;
 }
 
-static int gfx_v12_0_kiq_disable_kgq(struct amdgpu_device *adev)
-{
-	struct amdgpu_kiq *kiq = &adev->gfx.kiq[0];
-	struct amdgpu_ring *kiq_ring = &kiq->ring;
-	int i, r = 0;
-
-	if (!kiq->pmf || !kiq->pmf->kiq_unmap_queues)
-		return -EINVAL;
-
-	if (amdgpu_ring_alloc(kiq_ring, kiq->pmf->unmap_queues_size *
-					adev->gfx.num_gfx_rings))
-		return -ENOMEM;
-
-	for (i = 0; i < adev->gfx.num_gfx_rings; i++)
-		kiq->pmf->kiq_unmap_queues(kiq_ring, &adev->gfx.gfx_ring[i],
-					   PREEMPT_QUEUES, 0, 0);
-
-	if (adev->gfx.kiq[0].ring.sched.ready)
-		r = amdgpu_ring_test_helper(kiq_ring);
-
-	return r;
-}
-
 static int gfx_v12_0_hw_fini(void *handle)
 {
 	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-	int r;
 	uint32_t tmp;
 
 	amdgpu_irq_put(adev, &adev->gfx.priv_reg_irq, 0);
@@ -3580,8 +3556,7 @@ static int gfx_v12_0_hw_fini(void *handle)
 
 	if (!adev->no_hw_access) {
 		if (amdgpu_async_gfx_ring) {
-			r = gfx_v12_0_kiq_disable_kgq(adev);
-			if (r)
+			if (amdgpu_gfx_disable_kgq(adev, 0))
 				DRM_ERROR("KGQ disable failed\n");
 		}
 
diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v11_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v11_0.c
index b88a6fa173b3..2797fd84432b 100644
--- a/drivers/gpu/drm/amd/amdgpu/gmc_v11_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gmc_v11_0.c
@@ -231,7 +231,7 @@ static void gmc_v11_0_flush_gpu_tlb(struct amdgpu_device *adev, uint32_t vmid,
 	/* This is necessary for SRIOV as well as for GFXOFF to function
 	 * properly under bare metal
 	 */
-	if ((adev->gfx.kiq[0].ring.sched.ready || adev->mes.ring.sched.ready) &&
+	if ((adev->gfx.kiq[0].ring.sched.ready || adev->mes.ring[0].sched.ready) &&
 	    (amdgpu_sriov_runtime(adev) || !amdgpu_sriov_vf(adev))) {
 		amdgpu_gmc_fw_reg_write_reg_wait(adev, req, ack, inv_req,
 						 1 << vmid, GET_INST(GC, 0));
diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v12_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v12_0.c
index 26efce9aa410..edcb5351f8cc 100644
--- a/drivers/gpu/drm/amd/amdgpu/gmc_v12_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gmc_v12_0.c
@@ -299,7 +299,7 @@ static void gmc_v12_0_flush_gpu_tlb(struct amdgpu_device *adev, uint32_t vmid,
 	/* This is necessary for SRIOV as well as for GFXOFF to function
 	 * properly under bare metal
 	 */
-	if ((adev->gfx.kiq[0].ring.sched.ready || adev->mes.ring.sched.ready) &&
+	if ((adev->gfx.kiq[0].ring.sched.ready || adev->mes.ring[0].sched.ready) &&
 	    (amdgpu_sriov_runtime(adev) || !amdgpu_sriov_vf(adev))) {
 		struct amdgpu_vmhub *hub = &adev->vmhub[vmhub];
 		const unsigned eng = 17;
diff --git a/drivers/gpu/drm/amd/amdgpu/jpeg_v2_0.c b/drivers/gpu/drm/amd/amdgpu/jpeg_v2_0.c
index 99adf3625657..98aa3ccd0d20 100644
--- a/drivers/gpu/drm/amd/amdgpu/jpeg_v2_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/jpeg_v2_0.c
@@ -538,11 +538,11 @@ void jpeg_v2_0_dec_ring_emit_ib(struct amdgpu_ring *ring,
 
 	amdgpu_ring_write(ring, PACKETJ(mmUVD_LMI_JRBC_IB_VMID_INTERNAL_OFFSET,
 		0, 0, PACKETJ_TYPE0));
-	amdgpu_ring_write(ring, (vmid | (vmid << 4)));
+	amdgpu_ring_write(ring, (vmid | (vmid << 4) | (vmid << 8)));
 
 	amdgpu_ring_write(ring, PACKETJ(mmUVD_LMI_JPEG_VMID_INTERNAL_OFFSET,
 		0, 0, PACKETJ_TYPE0));
-	amdgpu_ring_write(ring, (vmid | (vmid << 4)));
+	amdgpu_ring_write(ring, (vmid | (vmid << 4) | (vmid << 8)));
 
 	amdgpu_ring_write(ring,	PACKETJ(mmUVD_LMI_JRBC_IB_64BIT_BAR_LOW_INTERNAL_OFFSET,
 		0, 0, PACKETJ_TYPE0));
diff --git a/drivers/gpu/drm/amd/amdgpu/jpeg_v4_0_3.c b/drivers/gpu/drm/amd/amdgpu/jpeg_v4_0_3.c
index ad524ddc9760..6ae5a784e187 100644
--- a/drivers/gpu/drm/amd/amdgpu/jpeg_v4_0_3.c
+++ b/drivers/gpu/drm/amd/amdgpu/jpeg_v4_0_3.c
@@ -23,6 +23,7 @@
 
 #include "amdgpu.h"
 #include "amdgpu_jpeg.h"
+#include "amdgpu_cs.h"
 #include "soc15.h"
 #include "soc15d.h"
 #include "jpeg_v4_0_3.h"
@@ -782,11 +783,15 @@ void jpeg_v4_0_3_dec_ring_emit_ib(struct amdgpu_ring *ring,
 
 	amdgpu_ring_write(ring, PACKETJ(regUVD_LMI_JRBC_IB_VMID_INTERNAL_OFFSET,
 		0, 0, PACKETJ_TYPE0));
-	amdgpu_ring_write(ring, (vmid | (vmid << 4)));
+
+	if (ring->funcs->parse_cs)
+		amdgpu_ring_write(ring, 0);
+	else
+		amdgpu_ring_write(ring, (vmid | (vmid << 4) | (vmid << 8)));
 
 	amdgpu_ring_write(ring, PACKETJ(regUVD_LMI_JPEG_VMID_INTERNAL_OFFSET,
 		0, 0, PACKETJ_TYPE0));
-	amdgpu_ring_write(ring, (vmid | (vmid << 4)));
+	amdgpu_ring_write(ring, (vmid | (vmid << 4) | (vmid << 8)));
 
 	amdgpu_ring_write(ring,	PACKETJ(regUVD_LMI_JRBC_IB_64BIT_BAR_LOW_INTERNAL_OFFSET,
 		0, 0, PACKETJ_TYPE0));
@@ -1084,6 +1089,7 @@ static const struct amdgpu_ring_funcs jpeg_v4_0_3_dec_ring_vm_funcs = {
 	.get_rptr = jpeg_v4_0_3_dec_ring_get_rptr,
 	.get_wptr = jpeg_v4_0_3_dec_ring_get_wptr,
 	.set_wptr = jpeg_v4_0_3_dec_ring_set_wptr,
+	.parse_cs = jpeg_v4_0_3_dec_ring_parse_cs,
 	.emit_frame_size =
 		SOC15_FLUSH_GPU_TLB_NUM_WREG * 6 +
 		SOC15_FLUSH_GPU_TLB_NUM_REG_WAIT * 8 +
@@ -1248,3 +1254,56 @@ static void jpeg_v4_0_3_set_ras_funcs(struct amdgpu_device *adev)
 {
 	adev->jpeg.ras = &jpeg_v4_0_3_ras;
 }
+
+/**
+ * jpeg_v4_0_3_dec_ring_parse_cs - command submission parser
+ *
+ * @parser: Command submission parser context
+ * @job: the job to parse
+ * @ib: the IB to parse
+ *
+ * Parse the command stream, return -EINVAL for invalid packet,
+ * 0 otherwise
+ */
+int jpeg_v4_0_3_dec_ring_parse_cs(struct amdgpu_cs_parser *parser,
+			     struct amdgpu_job *job,
+			     struct amdgpu_ib *ib)
+{
+	uint32_t i, reg, res, cond, type;
+	struct amdgpu_device *adev = parser->adev;
+
+	for (i = 0; i < ib->length_dw ; i += 2) {
+		reg  = CP_PACKETJ_GET_REG(ib->ptr[i]);
+		res  = CP_PACKETJ_GET_RES(ib->ptr[i]);
+		cond = CP_PACKETJ_GET_COND(ib->ptr[i]);
+		type = CP_PACKETJ_GET_TYPE(ib->ptr[i]);
+
+		if (res) /* only support 0 at the moment */
+			return -EINVAL;
+
+		switch (type) {
+		case PACKETJ_TYPE0:
+			if (cond != PACKETJ_CONDITION_CHECK0 || reg < JPEG_REG_RANGE_START || reg > JPEG_REG_RANGE_END) {
+				dev_err(adev->dev, "Invalid packet [0x%08x]!\n", ib->ptr[i]);
+				return -EINVAL;
+			}
+			break;
+		case PACKETJ_TYPE3:
+			if (cond != PACKETJ_CONDITION_CHECK3 || reg < JPEG_REG_RANGE_START || reg > JPEG_REG_RANGE_END) {
+				dev_err(adev->dev, "Invalid packet [0x%08x]!\n", ib->ptr[i]);
+				return -EINVAL;
+			}
+			break;
+		case PACKETJ_TYPE6:
+			if (ib->ptr[i] == CP_PACKETJ_NOP)
+				continue;
+			dev_err(adev->dev, "Invalid packet [0x%08x]!\n", ib->ptr[i]);
+			return -EINVAL;
+		default:
+			dev_err(adev->dev, "Unknown packet type %d !\n", type);
+			return -EINVAL;
+		}
+	}
+
+	return 0;
+}
diff --git a/drivers/gpu/drm/amd/amdgpu/jpeg_v4_0_3.h b/drivers/gpu/drm/amd/amdgpu/jpeg_v4_0_3.h
index 747a3e5f6856..71c54b294e15 100644
--- a/drivers/gpu/drm/amd/amdgpu/jpeg_v4_0_3.h
+++ b/drivers/gpu/drm/amd/amdgpu/jpeg_v4_0_3.h
@@ -46,6 +46,9 @@
 
 #define JRBC_DEC_EXTERNAL_REG_WRITE_ADDR				0x18000
 
+#define JPEG_REG_RANGE_START						0x4000
+#define JPEG_REG_RANGE_END						0x41c2
+
 extern const struct amdgpu_ip_block_version jpeg_v4_0_3_ip_block;
 
 void jpeg_v4_0_3_dec_ring_emit_ib(struct amdgpu_ring *ring,
@@ -62,5 +65,7 @@ void jpeg_v4_0_3_dec_ring_insert_end(struct amdgpu_ring *ring);
 void jpeg_v4_0_3_dec_ring_emit_wreg(struct amdgpu_ring *ring, uint32_t reg, uint32_t val);
 void jpeg_v4_0_3_dec_ring_emit_reg_wait(struct amdgpu_ring *ring, uint32_t reg,
 					uint32_t val, uint32_t mask);
-
+int jpeg_v4_0_3_dec_ring_parse_cs(struct amdgpu_cs_parser *parser,
+				  struct amdgpu_job *job,
+				  struct amdgpu_ib *ib);
 #endif /* __JPEG_V4_0_3_H__ */
diff --git a/drivers/gpu/drm/amd/amdgpu/jpeg_v5_0_0.c b/drivers/gpu/drm/amd/amdgpu/jpeg_v5_0_0.c
index d694a276498a..f4daff90c770 100644
--- a/drivers/gpu/drm/amd/amdgpu/jpeg_v5_0_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/jpeg_v5_0_0.c
@@ -646,6 +646,7 @@ static const struct amdgpu_ring_funcs jpeg_v5_0_0_dec_ring_vm_funcs = {
 	.get_rptr = jpeg_v5_0_0_dec_ring_get_rptr,
 	.get_wptr = jpeg_v5_0_0_dec_ring_get_wptr,
 	.set_wptr = jpeg_v5_0_0_dec_ring_set_wptr,
+	.parse_cs = jpeg_v4_0_3_dec_ring_parse_cs,
 	.emit_frame_size =
 		SOC15_FLUSH_GPU_TLB_NUM_WREG * 6 +
 		SOC15_FLUSH_GPU_TLB_NUM_REG_WAIT * 8 +
diff --git a/drivers/gpu/drm/amd/amdgpu/mes_v11_0.c b/drivers/gpu/drm/amd/amdgpu/mes_v11_0.c
index f9343642ae7e..2ea8223eb969 100644
--- a/drivers/gpu/drm/amd/amdgpu/mes_v11_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/mes_v11_0.c
@@ -162,13 +162,13 @@ static int mes_v11_0_submit_pkt_and_poll_completion(struct amdgpu_mes *mes,
 	union MESAPI__QUERY_MES_STATUS mes_status_pkt;
 	signed long timeout = 3000000; /* 3000 ms */
 	struct amdgpu_device *adev = mes->adev;
-	struct amdgpu_ring *ring = &mes->ring;
+	struct amdgpu_ring *ring = &mes->ring[0];
 	struct MES_API_STATUS *api_status;
 	union MESAPI__MISC *x_pkt = pkt;
 	const char *op_str, *misc_op_str;
 	unsigned long flags;
 	u64 status_gpu_addr;
-	u32 status_offset;
+	u32 seq, status_offset;
 	u64 *status_ptr;
 	signed long r;
 	int ret;
@@ -191,11 +191,18 @@ static int mes_v11_0_submit_pkt_and_poll_completion(struct amdgpu_mes *mes,
 	status_ptr = (u64 *)&adev->wb.wb[status_offset];
 	*status_ptr = 0;
 
-	spin_lock_irqsave(&mes->ring_lock, flags);
+	spin_lock_irqsave(&mes->ring_lock[0], flags);
 	r = amdgpu_ring_alloc(ring, (size + sizeof(mes_status_pkt)) / 4);
 	if (r)
 		goto error_unlock_free;
 
+	seq = ++ring->fence_drv.sync_seq;
+	r = amdgpu_fence_wait_polling(ring,
+				      seq - ring->fence_drv.num_fences_mask,
+				      timeout);
+	if (r < 1)
+		goto error_undo;
+
 	api_status = (struct MES_API_STATUS *)((char *)pkt + api_status_off);
 	api_status->api_completion_fence_addr = status_gpu_addr;
 	api_status->api_completion_fence_value = 1;
@@ -208,14 +215,13 @@ static int mes_v11_0_submit_pkt_and_poll_completion(struct amdgpu_mes *mes,
 	mes_status_pkt.header.dwsize = API_FRAME_SIZE_IN_DWORDS;
 	mes_status_pkt.api_status.api_completion_fence_addr =
 		ring->fence_drv.gpu_addr;
-	mes_status_pkt.api_status.api_completion_fence_value =
-		++ring->fence_drv.sync_seq;
+	mes_status_pkt.api_status.api_completion_fence_value = seq;
 
 	amdgpu_ring_write_multiple(ring, &mes_status_pkt,
 				   sizeof(mes_status_pkt) / 4);
 
 	amdgpu_ring_commit(ring);
-	spin_unlock_irqrestore(&mes->ring_lock, flags);
+	spin_unlock_irqrestore(&mes->ring_lock[0], flags);
 
 	op_str = mes_v11_0_get_op_string(x_pkt);
 	misc_op_str = mes_v11_0_get_misc_op_string(x_pkt);
@@ -229,7 +235,7 @@ static int mes_v11_0_submit_pkt_and_poll_completion(struct amdgpu_mes *mes,
 		dev_dbg(adev->dev, "MES msg=%d was emitted\n",
 			x_pkt->header.opcode);
 
-	r = amdgpu_fence_wait_polling(ring, ring->fence_drv.sync_seq, timeout);
+	r = amdgpu_fence_wait_polling(ring, seq, timeout);
 	if (r < 1 || !*status_ptr) {
 
 		if (misc_op_str)
@@ -252,8 +258,12 @@ static int mes_v11_0_submit_pkt_and_poll_completion(struct amdgpu_mes *mes,
 	amdgpu_device_wb_free(adev, status_offset);
 	return 0;
 
+error_undo:
+	dev_err(adev->dev, "MES ring buffer is full.\n");
+	amdgpu_ring_undo(ring);
+
 error_unlock_free:
-	spin_unlock_irqrestore(&mes->ring_lock, flags);
+	spin_unlock_irqrestore(&mes->ring_lock[0], flags);
 
 error_wb_free:
 	amdgpu_device_wb_free(adev, status_offset);
@@ -512,9 +522,9 @@ static int mes_v11_0_set_hw_resources(struct amdgpu_mes *mes)
 	mes_set_hw_res_pkt.vmid_mask_gfxhub = mes->vmid_mask_gfxhub;
 	mes_set_hw_res_pkt.gds_size = adev->gds.gds_size;
 	mes_set_hw_res_pkt.paging_vmid = 0;
-	mes_set_hw_res_pkt.g_sch_ctx_gpu_mc_ptr = mes->sch_ctx_gpu_addr;
+	mes_set_hw_res_pkt.g_sch_ctx_gpu_mc_ptr = mes->sch_ctx_gpu_addr[0];
 	mes_set_hw_res_pkt.query_status_fence_gpu_mc_ptr =
-		mes->query_status_fence_gpu_addr;
+		mes->query_status_fence_gpu_addr[0];
 
 	for (i = 0; i < MAX_COMPUTE_PIPES; i++)
 		mes_set_hw_res_pkt.compute_hqd_mask[i] =
@@ -1015,7 +1025,7 @@ static int mes_v11_0_kiq_enable_queue(struct amdgpu_device *adev)
 		return r;
 	}
 
-	kiq->pmf->kiq_map_queues(kiq_ring, &adev->mes.ring);
+	kiq->pmf->kiq_map_queues(kiq_ring, &adev->mes.ring[0]);
 
 	return amdgpu_ring_test_helper(kiq_ring);
 }
@@ -1029,7 +1039,7 @@ static int mes_v11_0_queue_init(struct amdgpu_device *adev,
 	if (pipe == AMDGPU_MES_KIQ_PIPE)
 		ring = &adev->gfx.kiq[0].ring;
 	else if (pipe == AMDGPU_MES_SCHED_PIPE)
-		ring = &adev->mes.ring;
+		ring = &adev->mes.ring[0];
 	else
 		BUG();
 
@@ -1071,7 +1081,7 @@ static int mes_v11_0_ring_init(struct amdgpu_device *adev)
 {
 	struct amdgpu_ring *ring;
 
-	ring = &adev->mes.ring;
+	ring = &adev->mes.ring[0];
 
 	ring->funcs = &mes_v11_0_ring_funcs;
 
@@ -1124,7 +1134,7 @@ static int mes_v11_0_mqd_sw_init(struct amdgpu_device *adev,
 	if (pipe == AMDGPU_MES_KIQ_PIPE)
 		ring = &adev->gfx.kiq[0].ring;
 	else if (pipe == AMDGPU_MES_SCHED_PIPE)
-		ring = &adev->mes.ring;
+		ring = &adev->mes.ring[0];
 	else
 		BUG();
 
@@ -1200,9 +1210,6 @@ static int mes_v11_0_sw_fini(void *handle)
 	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
 	int pipe;
 
-	amdgpu_device_wb_free(adev, adev->mes.sch_ctx_offs);
-	amdgpu_device_wb_free(adev, adev->mes.query_status_fence_offs);
-
 	for (pipe = 0; pipe < AMDGPU_MAX_MES_PIPES; pipe++) {
 		kfree(adev->mes.mqd_backup[pipe]);
 
@@ -1216,12 +1223,12 @@ static int mes_v11_0_sw_fini(void *handle)
 			      &adev->gfx.kiq[0].ring.mqd_gpu_addr,
 			      &adev->gfx.kiq[0].ring.mqd_ptr);
 
-	amdgpu_bo_free_kernel(&adev->mes.ring.mqd_obj,
-			      &adev->mes.ring.mqd_gpu_addr,
-			      &adev->mes.ring.mqd_ptr);
+	amdgpu_bo_free_kernel(&adev->mes.ring[0].mqd_obj,
+			      &adev->mes.ring[0].mqd_gpu_addr,
+			      &adev->mes.ring[0].mqd_ptr);
 
 	amdgpu_ring_fini(&adev->gfx.kiq[0].ring);
-	amdgpu_ring_fini(&adev->mes.ring);
+	amdgpu_ring_fini(&adev->mes.ring[0]);
 
 	if (adev->firmware.load_type == AMDGPU_FW_LOAD_DIRECT) {
 		mes_v11_0_free_ucode_buffers(adev, AMDGPU_MES_KIQ_PIPE);
@@ -1332,9 +1339,9 @@ failure:
 
 static int mes_v11_0_kiq_hw_fini(struct amdgpu_device *adev)
 {
-	if (adev->mes.ring.sched.ready) {
-		mes_v11_0_kiq_dequeue(&adev->mes.ring);
-		adev->mes.ring.sched.ready = false;
+	if (adev->mes.ring[0].sched.ready) {
+		mes_v11_0_kiq_dequeue(&adev->mes.ring[0]);
+		adev->mes.ring[0].sched.ready = false;
 	}
 
 	if (amdgpu_sriov_vf(adev)) {
@@ -1352,7 +1359,7 @@ static int mes_v11_0_hw_init(void *handle)
 	int r;
 	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
 
-	if (adev->mes.ring.sched.ready)
+	if (adev->mes.ring[0].sched.ready)
 		goto out;
 
 	if (!adev->enable_mes_kiq) {
@@ -1397,7 +1404,7 @@ out:
 	 * with MES enabled.
 	 */
 	adev->gfx.kiq[0].ring.sched.ready = false;
-	adev->mes.ring.sched.ready = true;
+	adev->mes.ring[0].sched.ready = true;
 
 	return 0;
 
diff --git a/drivers/gpu/drm/amd/amdgpu/mes_v12_0.c b/drivers/gpu/drm/amd/amdgpu/mes_v12_0.c
index 0713bc3eb263..e39a58d262c9 100644
--- a/drivers/gpu/drm/amd/amdgpu/mes_v12_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/mes_v12_0.c
@@ -142,19 +142,20 @@ static const char *mes_v12_0_get_misc_op_string(union MESAPI__MISC *x_pkt)
 }
 
 static int mes_v12_0_submit_pkt_and_poll_completion(struct amdgpu_mes *mes,
-						    void *pkt, int size,
-						    int api_status_off)
+					    int pipe, void *pkt, int size,
+					    int api_status_off)
 {
 	union MESAPI__QUERY_MES_STATUS mes_status_pkt;
 	signed long timeout = 3000000; /* 3000 ms */
 	struct amdgpu_device *adev = mes->adev;
-	struct amdgpu_ring *ring = &mes->ring;
+	struct amdgpu_ring *ring = &mes->ring[pipe];
+	spinlock_t *ring_lock = &mes->ring_lock[pipe];
 	struct MES_API_STATUS *api_status;
 	union MESAPI__MISC *x_pkt = pkt;
 	const char *op_str, *misc_op_str;
 	unsigned long flags;
 	u64 status_gpu_addr;
-	u32 status_offset;
+	u32 seq, status_offset;
 	u64 *status_ptr;
 	signed long r;
 	int ret;
@@ -177,11 +178,18 @@ static int mes_v12_0_submit_pkt_and_poll_completion(struct amdgpu_mes *mes,
 	status_ptr = (u64 *)&adev->wb.wb[status_offset];
 	*status_ptr = 0;
 
-	spin_lock_irqsave(&mes->ring_lock, flags);
+	spin_lock_irqsave(ring_lock, flags);
 	r = amdgpu_ring_alloc(ring, (size + sizeof(mes_status_pkt)) / 4);
 	if (r)
 		goto error_unlock_free;
 
+	seq = ++ring->fence_drv.sync_seq;
+	r = amdgpu_fence_wait_polling(ring,
+				      seq - ring->fence_drv.num_fences_mask,
+				      timeout);
+	if (r < 1)
+		goto error_undo;
+
 	api_status = (struct MES_API_STATUS *)((char *)pkt + api_status_off);
 	api_status->api_completion_fence_addr = status_gpu_addr;
 	api_status->api_completion_fence_value = 1;
@@ -194,39 +202,39 @@ static int mes_v12_0_submit_pkt_and_poll_completion(struct amdgpu_mes *mes,
 	mes_status_pkt.header.dwsize = API_FRAME_SIZE_IN_DWORDS;
 	mes_status_pkt.api_status.api_completion_fence_addr =
 		ring->fence_drv.gpu_addr;
-	mes_status_pkt.api_status.api_completion_fence_value =
-		++ring->fence_drv.sync_seq;
+	mes_status_pkt.api_status.api_completion_fence_value = seq;
 
 	amdgpu_ring_write_multiple(ring, &mes_status_pkt,
 				   sizeof(mes_status_pkt) / 4);
 
 	amdgpu_ring_commit(ring);
-	spin_unlock_irqrestore(&mes->ring_lock, flags);
+	spin_unlock_irqrestore(ring_lock, flags);
 
 	op_str = mes_v12_0_get_op_string(x_pkt);
 	misc_op_str = mes_v12_0_get_misc_op_string(x_pkt);
 
 	if (misc_op_str)
-		dev_dbg(adev->dev, "MES msg=%s (%s) was emitted\n", op_str,
-			misc_op_str);
+		dev_dbg(adev->dev, "MES(%d) msg=%s (%s) was emitted\n",
+			pipe, op_str, misc_op_str);
 	else if (op_str)
-		dev_dbg(adev->dev, "MES msg=%s was emitted\n", op_str);
+		dev_dbg(adev->dev, "MES(%d) msg=%s was emitted\n",
+			pipe, op_str);
 	else
-		dev_dbg(adev->dev, "MES msg=%d was emitted\n",
-			x_pkt->header.opcode);
+		dev_dbg(adev->dev, "MES(%d) msg=%d was emitted\n",
+			pipe, x_pkt->header.opcode);
 
-	r = amdgpu_fence_wait_polling(ring, ring->fence_drv.sync_seq, timeout);
+	r = amdgpu_fence_wait_polling(ring, seq, timeout);
 	if (r < 1 || !*status_ptr) {
 
 		if (misc_op_str)
-			dev_err(adev->dev, "MES failed to respond to msg=%s (%s)\n",
-				op_str, misc_op_str);
+			dev_err(adev->dev, "MES(%d) failed to respond to msg=%s (%s)\n",
+				pipe, op_str, misc_op_str);
 		else if (op_str)
-			dev_err(adev->dev, "MES failed to respond to msg=%s\n",
-				op_str);
+			dev_err(adev->dev, "MES(%d) failed to respond to msg=%s\n",
+				pipe, op_str);
 		else
-			dev_err(adev->dev, "MES failed to respond to msg=%d\n",
-				x_pkt->header.opcode);
+			dev_err(adev->dev, "MES(%d) failed to respond to msg=%d\n",
+				pipe, x_pkt->header.opcode);
 
 		while (halt_if_hws_hang)
 			schedule();
@@ -238,8 +246,12 @@ static int mes_v12_0_submit_pkt_and_poll_completion(struct amdgpu_mes *mes,
 	amdgpu_device_wb_free(adev, status_offset);
 	return 0;
 
+error_undo:
+	dev_err(adev->dev, "MES ring buffer is full.\n");
+	amdgpu_ring_undo(ring);
+
 error_unlock_free:
-	spin_unlock_irqrestore(&mes->ring_lock, flags);
+	spin_unlock_irqrestore(ring_lock, flags);
 
 error_wb_free:
 	amdgpu_device_wb_free(adev, status_offset);
@@ -254,6 +266,8 @@ static int convert_to_mes_queue_type(int queue_type)
 		return MES_QUEUE_TYPE_COMPUTE;
 	else if (queue_type == AMDGPU_RING_TYPE_SDMA)
 		return MES_QUEUE_TYPE_SDMA;
+	else if (queue_type == AMDGPU_RING_TYPE_MES)
+		return MES_QUEUE_TYPE_SCHQ;
 	else
 		BUG();
 	return -1;
@@ -311,6 +325,7 @@ static int mes_v12_0_add_hw_queue(struct amdgpu_mes *mes,
 	mes_add_queue_pkt.gds_size = input->queue_size;
 
 	return mes_v12_0_submit_pkt_and_poll_completion(mes,
+			AMDGPU_MES_SCHED_PIPE,
 			&mes_add_queue_pkt, sizeof(mes_add_queue_pkt),
 			offsetof(union MESAPI__ADD_QUEUE, api_status));
 }
@@ -330,6 +345,7 @@ static int mes_v12_0_remove_hw_queue(struct amdgpu_mes *mes,
 	mes_remove_queue_pkt.gang_context_addr = input->gang_context_addr;
 
 	return mes_v12_0_submit_pkt_and_poll_completion(mes,
+			AMDGPU_MES_SCHED_PIPE,
 			&mes_remove_queue_pkt, sizeof(mes_remove_queue_pkt),
 			offsetof(union MESAPI__REMOVE_QUEUE, api_status));
 }
@@ -338,6 +354,7 @@ static int mes_v12_0_map_legacy_queue(struct amdgpu_mes *mes,
 				      struct mes_map_legacy_queue_input *input)
 {
 	union MESAPI__ADD_QUEUE mes_add_queue_pkt;
+	int pipe;
 
 	memset(&mes_add_queue_pkt, 0, sizeof(mes_add_queue_pkt));
 
@@ -354,7 +371,12 @@ static int mes_v12_0_map_legacy_queue(struct amdgpu_mes *mes,
 		convert_to_mes_queue_type(input->queue_type);
 	mes_add_queue_pkt.map_legacy_kq = 1;
 
-	return mes_v12_0_submit_pkt_and_poll_completion(mes,
+	if (mes->adev->enable_uni_mes)
+		pipe = AMDGPU_MES_KIQ_PIPE;
+	else
+		pipe = AMDGPU_MES_SCHED_PIPE;
+
+	return mes_v12_0_submit_pkt_and_poll_completion(mes, pipe,
 			&mes_add_queue_pkt, sizeof(mes_add_queue_pkt),
 			offsetof(union MESAPI__ADD_QUEUE, api_status));
 }
@@ -363,6 +385,7 @@ static int mes_v12_0_unmap_legacy_queue(struct amdgpu_mes *mes,
 			struct mes_unmap_legacy_queue_input *input)
 {
 	union MESAPI__REMOVE_QUEUE mes_remove_queue_pkt;
+	int pipe;
 
 	memset(&mes_remove_queue_pkt, 0, sizeof(mes_remove_queue_pkt));
 
@@ -387,7 +410,12 @@ static int mes_v12_0_unmap_legacy_queue(struct amdgpu_mes *mes,
 			convert_to_mes_queue_type(input->queue_type);
 	}
 
-	return mes_v12_0_submit_pkt_and_poll_completion(mes,
+	if (mes->adev->enable_uni_mes)
+		pipe = AMDGPU_MES_KIQ_PIPE;
+	else
+		pipe = AMDGPU_MES_SCHED_PIPE;
+
+	return mes_v12_0_submit_pkt_and_poll_completion(mes, pipe,
 			&mes_remove_queue_pkt, sizeof(mes_remove_queue_pkt),
 			offsetof(union MESAPI__REMOVE_QUEUE, api_status));
 }
@@ -404,7 +432,7 @@ static int mes_v12_0_resume_gang(struct amdgpu_mes *mes,
 	return 0;
 }
 
-static int mes_v12_0_query_sched_status(struct amdgpu_mes *mes)
+static int mes_v12_0_query_sched_status(struct amdgpu_mes *mes, int pipe)
 {
 	union MESAPI__QUERY_MES_STATUS mes_status_pkt;
 
@@ -414,7 +442,7 @@ static int mes_v12_0_query_sched_status(struct amdgpu_mes *mes)
 	mes_status_pkt.header.opcode = MES_SCH_API_QUERY_SCHEDULER_STATUS;
 	mes_status_pkt.header.dwsize = API_FRAME_SIZE_IN_DWORDS;
 
-	return mes_v12_0_submit_pkt_and_poll_completion(mes,
+	return mes_v12_0_submit_pkt_and_poll_completion(mes, pipe,
 			&mes_status_pkt, sizeof(mes_status_pkt),
 			offsetof(union MESAPI__QUERY_MES_STATUS, api_status));
 }
@@ -423,6 +451,7 @@ static int mes_v12_0_misc_op(struct amdgpu_mes *mes,
 			     struct mes_misc_op_input *input)
 {
 	union MESAPI__MISC misc_pkt;
+	int pipe;
 
 	memset(&misc_pkt, 0, sizeof(misc_pkt));
 
@@ -475,12 +504,17 @@ static int mes_v12_0_misc_op(struct amdgpu_mes *mes,
 		return -EINVAL;
 	}
 
-	return mes_v12_0_submit_pkt_and_poll_completion(mes,
+	if (mes->adev->enable_uni_mes)
+		pipe = AMDGPU_MES_KIQ_PIPE;
+	else
+		pipe = AMDGPU_MES_SCHED_PIPE;
+
+	return mes_v12_0_submit_pkt_and_poll_completion(mes, pipe,
 			&misc_pkt, sizeof(misc_pkt),
 			offsetof(union MESAPI__MISC, api_status));
 }
 
-static int mes_v12_0_set_hw_resources_1(struct amdgpu_mes *mes)
+static int mes_v12_0_set_hw_resources_1(struct amdgpu_mes *mes, int pipe)
 {
 	union MESAPI_SET_HW_RESOURCES_1 mes_set_hw_res_1_pkt;
 
@@ -491,12 +525,12 @@ static int mes_v12_0_set_hw_resources_1(struct amdgpu_mes *mes)
 	mes_set_hw_res_1_pkt.header.dwsize = API_FRAME_SIZE_IN_DWORDS;
 	mes_set_hw_res_1_pkt.mes_kiq_unmap_timeout = 100;
 
-	return mes_v12_0_submit_pkt_and_poll_completion(mes,
+	return mes_v12_0_submit_pkt_and_poll_completion(mes, pipe,
 			&mes_set_hw_res_1_pkt, sizeof(mes_set_hw_res_1_pkt),
 			offsetof(union MESAPI_SET_HW_RESOURCES_1, api_status));
 }
 
-static int mes_v12_0_set_hw_resources(struct amdgpu_mes *mes)
+static int mes_v12_0_set_hw_resources(struct amdgpu_mes *mes, int pipe)
 {
 	int i;
 	struct amdgpu_device *adev = mes->adev;
@@ -508,27 +542,33 @@ static int mes_v12_0_set_hw_resources(struct amdgpu_mes *mes)
 	mes_set_hw_res_pkt.header.opcode = MES_SCH_API_SET_HW_RSRC;
 	mes_set_hw_res_pkt.header.dwsize = API_FRAME_SIZE_IN_DWORDS;
 
-	mes_set_hw_res_pkt.vmid_mask_mmhub = mes->vmid_mask_mmhub;
-	mes_set_hw_res_pkt.vmid_mask_gfxhub = mes->vmid_mask_gfxhub;
-	mes_set_hw_res_pkt.gds_size = adev->gds.gds_size;
-	mes_set_hw_res_pkt.paging_vmid = 0;
-	mes_set_hw_res_pkt.g_sch_ctx_gpu_mc_ptr = mes->sch_ctx_gpu_addr;
-	mes_set_hw_res_pkt.query_status_fence_gpu_mc_ptr =
-		mes->query_status_fence_gpu_addr;
-
-	for (i = 0; i < MAX_COMPUTE_PIPES; i++)
-		mes_set_hw_res_pkt.compute_hqd_mask[i] =
-			mes->compute_hqd_mask[i];
-
-	for (i = 0; i < MAX_GFX_PIPES; i++)
-		mes_set_hw_res_pkt.gfx_hqd_mask[i] = mes->gfx_hqd_mask[i];
-
-	for (i = 0; i < MAX_SDMA_PIPES; i++)
-		mes_set_hw_res_pkt.sdma_hqd_mask[i] = mes->sdma_hqd_mask[i];
+	if (pipe == AMDGPU_MES_SCHED_PIPE) {
+		mes_set_hw_res_pkt.vmid_mask_mmhub = mes->vmid_mask_mmhub;
+		mes_set_hw_res_pkt.vmid_mask_gfxhub = mes->vmid_mask_gfxhub;
+		mes_set_hw_res_pkt.gds_size = adev->gds.gds_size;
+		mes_set_hw_res_pkt.paging_vmid = 0;
+
+		for (i = 0; i < MAX_COMPUTE_PIPES; i++)
+			mes_set_hw_res_pkt.compute_hqd_mask[i] =
+				mes->compute_hqd_mask[i];
+
+		for (i = 0; i < MAX_GFX_PIPES; i++)
+			mes_set_hw_res_pkt.gfx_hqd_mask[i] =
+				mes->gfx_hqd_mask[i];
+
+		for (i = 0; i < MAX_SDMA_PIPES; i++)
+			mes_set_hw_res_pkt.sdma_hqd_mask[i] =
+				mes->sdma_hqd_mask[i];
+
+		for (i = 0; i < AMD_PRIORITY_NUM_LEVELS; i++)
+			mes_set_hw_res_pkt.aggregated_doorbells[i] =
+				mes->aggregated_doorbells[i];
+	}
 
-	for (i = 0; i < AMD_PRIORITY_NUM_LEVELS; i++)
-		mes_set_hw_res_pkt.aggregated_doorbells[i] =
-			mes->aggregated_doorbells[i];
+	mes_set_hw_res_pkt.g_sch_ctx_gpu_mc_ptr =
+		mes->sch_ctx_gpu_addr[pipe];
+	mes_set_hw_res_pkt.query_status_fence_gpu_mc_ptr =
+		mes->query_status_fence_gpu_addr[pipe];
 
 	for (i = 0; i < 5; i++) {
 		mes_set_hw_res_pkt.gc_base[i] = adev->reg_offset[GC_HWIP][0][i];
@@ -556,7 +596,7 @@ static int mes_v12_0_set_hw_resources(struct amdgpu_mes *mes)
 		mes_set_hw_res_pkt.event_intr_history_gpu_mc_ptr = mes->event_log_gpu_addr;
 	}
 
-	return mes_v12_0_submit_pkt_and_poll_completion(mes,
+	return mes_v12_0_submit_pkt_and_poll_completion(mes, pipe,
 			&mes_set_hw_res_pkt, sizeof(mes_set_hw_res_pkt),
 			offsetof(union MESAPI_SET_HW_RESOURCES, api_status));
 }
@@ -734,16 +774,11 @@ static void mes_v12_0_enable(struct amdgpu_device *adev, bool enable)
 	if (enable) {
 		data = RREG32_SOC15(GC, 0, regCP_MES_CNTL);
 		data = REG_SET_FIELD(data, CP_MES_CNTL, MES_PIPE0_RESET, 1);
-		data = REG_SET_FIELD(data, CP_MES_CNTL, MES_PIPE1_RESET,
-		       (!adev->enable_uni_mes && adev->enable_mes_kiq) ? 1 : 0);
+		data = REG_SET_FIELD(data, CP_MES_CNTL, MES_PIPE1_RESET, 1);
 		WREG32_SOC15(GC, 0, regCP_MES_CNTL, data);
 
 		mutex_lock(&adev->srbm_mutex);
 		for (pipe = 0; pipe < AMDGPU_MAX_MES_PIPES; pipe++) {
-			if ((!adev->enable_mes_kiq || adev->enable_uni_mes) &&
-			    pipe == AMDGPU_MES_KIQ_PIPE)
-				continue;
-
 			soc21_grbm_select(adev, 3, pipe, 0, 0);
 
 			ucode_addr = adev->mes.uc_start_addr[pipe] >> 2;
@@ -757,8 +792,7 @@ static void mes_v12_0_enable(struct amdgpu_device *adev, bool enable)
 
 		/* unhalt MES and activate pipe0 */
 		data = REG_SET_FIELD(0, CP_MES_CNTL, MES_PIPE0_ACTIVE, 1);
-		data = REG_SET_FIELD(data, CP_MES_CNTL, MES_PIPE1_ACTIVE,
-		       (!adev->enable_uni_mes && adev->enable_mes_kiq) ? 1 : 0);
+		data = REG_SET_FIELD(data, CP_MES_CNTL, MES_PIPE1_ACTIVE, 1);
 		WREG32_SOC15(GC, 0, regCP_MES_CNTL, data);
 
 		if (amdgpu_emu_mode)
@@ -774,8 +808,7 @@ static void mes_v12_0_enable(struct amdgpu_device *adev, bool enable)
 		data = REG_SET_FIELD(data, CP_MES_CNTL,
 				     MES_INVALIDATE_ICACHE, 1);
 		data = REG_SET_FIELD(data, CP_MES_CNTL, MES_PIPE0_RESET, 1);
-		data = REG_SET_FIELD(data, CP_MES_CNTL, MES_PIPE1_RESET,
-		       (!adev->enable_uni_mes && adev->enable_mes_kiq) ? 1 : 0);
+		data = REG_SET_FIELD(data, CP_MES_CNTL, MES_PIPE1_RESET, 1);
 		data = REG_SET_FIELD(data, CP_MES_CNTL, MES_HALT, 1);
 		WREG32_SOC15(GC, 0, regCP_MES_CNTL, data);
 	}
@@ -790,10 +823,6 @@ static void mes_v12_0_set_ucode_start_addr(struct amdgpu_device *adev)
 
 	mutex_lock(&adev->srbm_mutex);
 	for (pipe = 0; pipe < AMDGPU_MAX_MES_PIPES; pipe++) {
-		if ((!adev->enable_mes_kiq || adev->enable_uni_mes) &&
-		    pipe == AMDGPU_MES_KIQ_PIPE)
-			continue;
-
 		/* me=3, queue=0 */
 		soc21_grbm_select(adev, 3, pipe, 0, 0);
 
@@ -1085,7 +1114,7 @@ static int mes_v12_0_kiq_enable_queue(struct amdgpu_device *adev)
 		return r;
 	}
 
-	kiq->pmf->kiq_map_queues(kiq_ring, &adev->mes.ring);
+	kiq->pmf->kiq_map_queues(kiq_ring, &adev->mes.ring[0]);
 
 	r = amdgpu_ring_test_ring(kiq_ring);
 	if (r) {
@@ -1101,14 +1130,12 @@ static int mes_v12_0_queue_init(struct amdgpu_device *adev,
 	struct amdgpu_ring *ring;
 	int r;
 
-	if (pipe == AMDGPU_MES_KIQ_PIPE)
+	if (!adev->enable_uni_mes && pipe == AMDGPU_MES_KIQ_PIPE)
 		ring = &adev->gfx.kiq[0].ring;
-	else if (pipe == AMDGPU_MES_SCHED_PIPE)
-		ring = &adev->mes.ring;
 	else
-		BUG();
+		ring = &adev->mes.ring[pipe];
 
-	if ((pipe == AMDGPU_MES_SCHED_PIPE) &&
+	if ((adev->enable_uni_mes || pipe == AMDGPU_MES_SCHED_PIPE) &&
 	    (amdgpu_in_reset(adev) || adev->in_suspend)) {
 		*(ring->wptr_cpu_addr) = 0;
 		*(ring->rptr_cpu_addr) = 0;
@@ -1120,13 +1147,12 @@ static int mes_v12_0_queue_init(struct amdgpu_device *adev,
 		return r;
 
 	if (pipe == AMDGPU_MES_SCHED_PIPE) {
-		if (adev->enable_uni_mes) {
-			mes_v12_0_queue_init_register(ring);
-		} else {
+		if (adev->enable_uni_mes)
+			r = amdgpu_mes_map_legacy_queue(adev, ring);
+		else
 			r = mes_v12_0_kiq_enable_queue(adev);
-			if (r)
-				return r;
-		}
+		if (r)
+			return r;
 	} else {
 		mes_v12_0_queue_init_register(ring);
 	}
@@ -1146,25 +1172,29 @@ static int mes_v12_0_queue_init(struct amdgpu_device *adev,
 	return 0;
 }
 
-static int mes_v12_0_ring_init(struct amdgpu_device *adev)
+static int mes_v12_0_ring_init(struct amdgpu_device *adev, int pipe)
 {
 	struct amdgpu_ring *ring;
 
-	ring = &adev->mes.ring;
+	ring = &adev->mes.ring[pipe];
 
 	ring->funcs = &mes_v12_0_ring_funcs;
 
 	ring->me = 3;
-	ring->pipe = 0;
+	ring->pipe = pipe;
 	ring->queue = 0;
 
 	ring->ring_obj = NULL;
 	ring->use_doorbell = true;
-	ring->doorbell_index = adev->doorbell_index.mes_ring0 << 1;
-	ring->eop_gpu_addr = adev->mes.eop_gpu_addr[AMDGPU_MES_SCHED_PIPE];
+	ring->eop_gpu_addr = adev->mes.eop_gpu_addr[pipe];
 	ring->no_scheduler = true;
 	sprintf(ring->name, "mes_%d.%d.%d", ring->me, ring->pipe, ring->queue);
 
+	if (pipe == AMDGPU_MES_SCHED_PIPE)
+		ring->doorbell_index = adev->doorbell_index.mes_ring0 << 1;
+	else
+		ring->doorbell_index = adev->doorbell_index.mes_ring1 << 1;
+
 	return amdgpu_ring_init(adev, ring, 1024, NULL, 0,
 				AMDGPU_RING_PRIO_DEFAULT, NULL);
 }
@@ -1178,7 +1208,7 @@ static int mes_v12_0_kiq_ring_init(struct amdgpu_device *adev)
 	ring = &adev->gfx.kiq[0].ring;
 
 	ring->me = 3;
-	ring->pipe = adev->enable_uni_mes ? 0 : 1;
+	ring->pipe = 1;
 	ring->queue = 0;
 
 	ring->adev = NULL;
@@ -1200,12 +1230,10 @@ static int mes_v12_0_mqd_sw_init(struct amdgpu_device *adev,
 	int r, mqd_size = sizeof(struct v12_compute_mqd);
 	struct amdgpu_ring *ring;
 
-	if (pipe == AMDGPU_MES_KIQ_PIPE)
+	if (!adev->enable_uni_mes && pipe == AMDGPU_MES_KIQ_PIPE)
 		ring = &adev->gfx.kiq[0].ring;
-	else if (pipe == AMDGPU_MES_SCHED_PIPE)
-		ring = &adev->mes.ring;
 	else
-		BUG();
+		ring = &adev->mes.ring[pipe];
 
 	if (ring->mqd_obj)
 		return 0;
@@ -1246,9 +1274,6 @@ static int mes_v12_0_sw_init(void *handle)
 		return r;
 
 	for (pipe = 0; pipe < AMDGPU_MAX_MES_PIPES; pipe++) {
-		if (!adev->enable_mes_kiq && pipe == AMDGPU_MES_KIQ_PIPE)
-			continue;
-
 		r = mes_v12_0_allocate_eop_buf(adev, pipe);
 		if (r)
 			return r;
@@ -1256,18 +1281,15 @@ static int mes_v12_0_sw_init(void *handle)
 		r = mes_v12_0_mqd_sw_init(adev, pipe);
 		if (r)
 			return r;
-	}
 
-	if (adev->enable_mes_kiq) {
-		r = mes_v12_0_kiq_ring_init(adev);
+		if (!adev->enable_uni_mes && pipe == AMDGPU_MES_KIQ_PIPE)
+			r = mes_v12_0_kiq_ring_init(adev);
+		else
+			r = mes_v12_0_ring_init(adev, pipe);
 		if (r)
 			return r;
 	}
 
-	r = mes_v12_0_ring_init(adev);
-	if (r)
-		return r;
-
 	return 0;
 }
 
@@ -1276,9 +1298,6 @@ static int mes_v12_0_sw_fini(void *handle)
 	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
 	int pipe;
 
-	amdgpu_device_wb_free(adev, adev->mes.sch_ctx_offs);
-	amdgpu_device_wb_free(adev, adev->mes.query_status_fence_offs);
-
 	for (pipe = 0; pipe < AMDGPU_MAX_MES_PIPES; pipe++) {
 		kfree(adev->mes.mqd_backup[pipe]);
 
@@ -1286,18 +1305,21 @@ static int mes_v12_0_sw_fini(void *handle)
 				      &adev->mes.eop_gpu_addr[pipe],
 				      NULL);
 		amdgpu_ucode_release(&adev->mes.fw[pipe]);
-	}
-
-	amdgpu_bo_free_kernel(&adev->gfx.kiq[0].ring.mqd_obj,
-			      &adev->gfx.kiq[0].ring.mqd_gpu_addr,
-			      &adev->gfx.kiq[0].ring.mqd_ptr);
 
-	amdgpu_bo_free_kernel(&adev->mes.ring.mqd_obj,
-			      &adev->mes.ring.mqd_gpu_addr,
-			      &adev->mes.ring.mqd_ptr);
+		if (adev->enable_uni_mes || pipe == AMDGPU_MES_SCHED_PIPE) {
+			amdgpu_bo_free_kernel(&adev->mes.ring[pipe].mqd_obj,
+					      &adev->mes.ring[pipe].mqd_gpu_addr,
+					      &adev->mes.ring[pipe].mqd_ptr);
+			amdgpu_ring_fini(&adev->mes.ring[pipe]);
+		}
+	}
 
-	amdgpu_ring_fini(&adev->gfx.kiq[0].ring);
-	amdgpu_ring_fini(&adev->mes.ring);
+	if (!adev->enable_uni_mes) {
+		amdgpu_bo_free_kernel(&adev->gfx.kiq[0].ring.mqd_obj,
+				      &adev->gfx.kiq[0].ring.mqd_gpu_addr,
+				      &adev->gfx.kiq[0].ring.mqd_ptr);
+		amdgpu_ring_fini(&adev->gfx.kiq[0].ring);
+	}
 
 	if (adev->firmware.load_type == AMDGPU_FW_LOAD_DIRECT) {
 		mes_v12_0_free_ucode_buffers(adev, AMDGPU_MES_KIQ_PIPE);
@@ -1341,7 +1363,7 @@ static void mes_v12_0_kiq_dequeue_sched(struct amdgpu_device *adev)
 	soc21_grbm_select(adev, 0, 0, 0, 0);
 	mutex_unlock(&adev->srbm_mutex);
 
-	adev->mes.ring.sched.ready = false;
+	adev->mes.ring[0].sched.ready = false;
 }
 
 static void mes_v12_0_kiq_setting(struct amdgpu_ring *ring)
@@ -1362,10 +1384,10 @@ static int mes_v12_0_kiq_hw_init(struct amdgpu_device *adev)
 {
 	int r = 0;
 
-	mes_v12_0_kiq_setting(&adev->gfx.kiq[0].ring);
-
 	if (adev->enable_uni_mes)
-		return mes_v12_0_hw_init(adev);
+		mes_v12_0_kiq_setting(&adev->mes.ring[AMDGPU_MES_KIQ_PIPE]);
+	else
+		mes_v12_0_kiq_setting(&adev->gfx.kiq[0].ring);
 
 	if (adev->firmware.load_type == AMDGPU_FW_LOAD_DIRECT) {
 
@@ -1392,6 +1414,14 @@ static int mes_v12_0_kiq_hw_init(struct amdgpu_device *adev)
 	if (r)
 		goto failure;
 
+	if (adev->enable_uni_mes) {
+		r = mes_v12_0_set_hw_resources(&adev->mes, AMDGPU_MES_KIQ_PIPE);
+		if (r)
+			goto failure;
+
+		mes_v12_0_set_hw_resources_1(&adev->mes, AMDGPU_MES_KIQ_PIPE);
+	}
+
 	r = mes_v12_0_hw_init(adev);
 	if (r)
 		goto failure;
@@ -1405,9 +1435,15 @@ failure:
 
 static int mes_v12_0_kiq_hw_fini(struct amdgpu_device *adev)
 {
-	if (adev->mes.ring.sched.ready) {
-		mes_v12_0_kiq_dequeue_sched(adev);
-		adev->mes.ring.sched.ready = false;
+	if (adev->mes.ring[0].sched.ready) {
+		if (adev->enable_uni_mes)
+			amdgpu_mes_unmap_legacy_queue(adev,
+				      &adev->mes.ring[AMDGPU_MES_SCHED_PIPE],
+				      RESET_QUEUES, 0, 0);
+		else
+			mes_v12_0_kiq_dequeue_sched(adev);
+
+		adev->mes.ring[0].sched.ready = false;
 	}
 
 	mes_v12_0_enable(adev, false);
@@ -1420,10 +1456,10 @@ static int mes_v12_0_hw_init(void *handle)
 	int r;
 	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
 
-	if (adev->mes.ring.sched.ready)
+	if (adev->mes.ring[0].sched.ready)
 		goto out;
 
-	if (!adev->enable_mes_kiq || adev->enable_uni_mes) {
+	if (!adev->enable_mes_kiq) {
 		if (adev->firmware.load_type == AMDGPU_FW_LOAD_DIRECT) {
 			r = mes_v12_0_load_microcode(adev,
 					     AMDGPU_MES_SCHED_PIPE, true);
@@ -1443,23 +1479,23 @@ static int mes_v12_0_hw_init(void *handle)
 		mes_v12_0_enable(adev, true);
 	}
 
+	/* Enable the MES to handle doorbell ring on unmapped queue */
+	mes_v12_0_enable_unmapped_doorbell_handling(&adev->mes, true);
+
 	r = mes_v12_0_queue_init(adev, AMDGPU_MES_SCHED_PIPE);
 	if (r)
 		goto failure;
 
-	r = mes_v12_0_set_hw_resources(&adev->mes);
+	r = mes_v12_0_set_hw_resources(&adev->mes, AMDGPU_MES_SCHED_PIPE);
 	if (r)
 		goto failure;
 
 	if (adev->enable_uni_mes)
-		mes_v12_0_set_hw_resources_1(&adev->mes);
+		mes_v12_0_set_hw_resources_1(&adev->mes, AMDGPU_MES_SCHED_PIPE);
 
 	mes_v12_0_init_aggregated_doorbell(&adev->mes);
 
-	/* Enable the MES to handle doorbell ring on unmapped queue */
-	mes_v12_0_enable_unmapped_doorbell_handling(&adev->mes, true);
-
-	r = mes_v12_0_query_sched_status(&adev->mes);
+	r = mes_v12_0_query_sched_status(&adev->mes, AMDGPU_MES_SCHED_PIPE);
 	if (r) {
 		DRM_ERROR("MES is busy\n");
 		goto failure;
@@ -1472,7 +1508,7 @@ out:
 	 * with MES enabled.
 	 */
 	adev->gfx.kiq[0].ring.sched.ready = false;
-	adev->mes.ring.sched.ready = true;
+	adev->mes.ring[0].sched.ready = true;
 
 	return 0;
 
@@ -1515,17 +1551,7 @@ static int mes_v12_0_early_init(void *handle)
 	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
 	int pipe, r;
 
-	if (adev->enable_uni_mes) {
-		r = amdgpu_mes_init_microcode(adev, AMDGPU_MES_SCHED_PIPE);
-		if (!r)
-			return 0;
-
-		adev->enable_uni_mes = false;
-	}
-
 	for (pipe = 0; pipe < AMDGPU_MAX_MES_PIPES; pipe++) {
-		if (!adev->enable_mes_kiq && pipe == AMDGPU_MES_KIQ_PIPE)
-			continue;
 		r = amdgpu_mes_init_microcode(adev, pipe);
 		if (r)
 			return r;
diff --git a/drivers/gpu/drm/amd/amdgpu/soc15d.h b/drivers/gpu/drm/amd/amdgpu/soc15d.h
index 2357ff39323f..e74e1983da53 100644
--- a/drivers/gpu/drm/amd/amdgpu/soc15d.h
+++ b/drivers/gpu/drm/amd/amdgpu/soc15d.h
@@ -76,6 +76,12 @@
 			 ((cond & 0xF) << 24) |				\
 			 ((type & 0xF) << 28))
 
+#define CP_PACKETJ_NOP		0x60000000
+#define CP_PACKETJ_GET_REG(x)  ((x) & 0x3FFFF)
+#define CP_PACKETJ_GET_RES(x)  (((x) >> 18) & 0x3F)
+#define CP_PACKETJ_GET_COND(x) (((x) >> 24) & 0xF)
+#define CP_PACKETJ_GET_TYPE(x) (((x) >> 28) & 0xF)
+
 /* Packet 3 types */
 #define	PACKET3_NOP					0x10
 #define	PACKET3_SET_BASE				0x11
diff --git a/drivers/gpu/drm/amd/amdgpu/soc24.c b/drivers/gpu/drm/amd/amdgpu/soc24.c
index 7d641d0dadba..b0c3678cfb31 100644
--- a/drivers/gpu/drm/amd/amdgpu/soc24.c
+++ b/drivers/gpu/drm/amd/amdgpu/soc24.c
@@ -406,6 +406,7 @@ static int soc24_common_early_init(void *handle)
 			AMD_CG_SUPPORT_ATHUB_MGCG |
 			AMD_CG_SUPPORT_ATHUB_LS |
 			AMD_CG_SUPPORT_MC_MGCG |
+			AMD_CG_SUPPORT_HDP_SD |
 			AMD_CG_SUPPORT_MC_LS;
 		adev->pg_flags = AMD_PG_SUPPORT_VCN |
 			AMD_PG_SUPPORT_JPEG |
@@ -424,6 +425,7 @@ static int soc24_common_early_init(void *handle)
 			AMD_CG_SUPPORT_ATHUB_MGCG |
 			AMD_CG_SUPPORT_ATHUB_LS |
 			AMD_CG_SUPPORT_MC_MGCG |
+			AMD_CG_SUPPORT_HDP_SD |
 			AMD_CG_SUPPORT_MC_LS;
 
 		adev->pg_flags = AMD_PG_SUPPORT_VCN |
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 7e7929f24ae4..983a977632ff 100644
--- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c
+++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c
@@ -2893,6 +2893,9 @@ static int dm_suspend(void *handle)
 
 	hpd_rx_irq_work_suspend(dm);
 
+	if (adev->dm.dc->caps.ips_support)
+		dc_allow_idle_optimizations(adev->dm.dc, true);
+
 	dc_set_power_state(dm->dc, DC_ACPI_CM_POWER_STATE_D3);
 	dc_dmub_srv_set_power_state(dm->dc->ctx->dmub_srv, DC_ACPI_CM_POWER_STATE_D3);
 
diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c
index 915eb2c08ece..2e9f6da1acdc 100644
--- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c
+++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c
@@ -804,12 +804,25 @@ struct dsc_mst_fairness_params {
 };
 
 #if defined(CONFIG_DRM_AMD_DC_FP)
-static int kbps_to_peak_pbn(int kbps)
+static uint16_t get_fec_overhead_multiplier(struct dc_link *dc_link)
+{
+	u8 link_coding_cap;
+	uint16_t fec_overhead_multiplier_x1000 = PBN_FEC_OVERHEAD_MULTIPLIER_8B_10B;
+
+	link_coding_cap = dc_link_dp_mst_decide_link_encoding_format(dc_link);
+	if (link_coding_cap == DP_128b_132b_ENCODING)
+		fec_overhead_multiplier_x1000 = PBN_FEC_OVERHEAD_MULTIPLIER_128B_132B;
+
+	return fec_overhead_multiplier_x1000;
+}
+
+static int kbps_to_peak_pbn(int kbps, uint16_t fec_overhead_multiplier_x1000)
 {
 	u64 peak_kbps = kbps;
 
 	peak_kbps *= 1006;
-	peak_kbps = div_u64(peak_kbps, 1000);
+	peak_kbps *= fec_overhead_multiplier_x1000;
+	peak_kbps = div_u64(peak_kbps, 1000 * 1000);
 	return (int) DIV64_U64_ROUND_UP(peak_kbps * 64, (54 * 8 * 1000));
 }
 
@@ -910,11 +923,12 @@ static int increase_dsc_bpp(struct drm_atomic_state *state,
 	int link_timeslots_used;
 	int fair_pbn_alloc;
 	int ret = 0;
+	uint16_t fec_overhead_multiplier_x1000 = get_fec_overhead_multiplier(dc_link);
 
 	for (i = 0; i < count; i++) {
 		if (vars[i + k].dsc_enabled) {
 			initial_slack[i] =
-			kbps_to_peak_pbn(params[i].bw_range.max_kbps) - vars[i + k].pbn;
+			kbps_to_peak_pbn(params[i].bw_range.max_kbps, fec_overhead_multiplier_x1000) - vars[i + k].pbn;
 			bpp_increased[i] = false;
 			remaining_to_increase += 1;
 		} else {
@@ -1010,6 +1024,7 @@ static int try_disable_dsc(struct drm_atomic_state *state,
 	int next_index;
 	int remaining_to_try = 0;
 	int ret;
+	uint16_t fec_overhead_multiplier_x1000 = get_fec_overhead_multiplier(dc_link);
 
 	for (i = 0; i < count; i++) {
 		if (vars[i + k].dsc_enabled
@@ -1039,7 +1054,7 @@ static int try_disable_dsc(struct drm_atomic_state *state,
 		if (next_index == -1)
 			break;
 
-		vars[next_index].pbn = kbps_to_peak_pbn(params[next_index].bw_range.stream_kbps);
+		vars[next_index].pbn = kbps_to_peak_pbn(params[next_index].bw_range.stream_kbps, fec_overhead_multiplier_x1000);
 		ret = drm_dp_atomic_find_time_slots(state,
 						    params[next_index].port->mgr,
 						    params[next_index].port,
@@ -1052,8 +1067,7 @@ static int try_disable_dsc(struct drm_atomic_state *state,
 			vars[next_index].dsc_enabled = false;
 			vars[next_index].bpp_x16 = 0;
 		} else {
-			vars[next_index].pbn = kbps_to_peak_pbn(
-				params[next_index].bw_range.max_kbps);
+			vars[next_index].pbn = kbps_to_peak_pbn(params[next_index].bw_range.stream_kbps, fec_overhead_multiplier_x1000);
 			ret = drm_dp_atomic_find_time_slots(state,
 							    params[next_index].port->mgr,
 							    params[next_index].port,
@@ -1082,6 +1096,7 @@ static int compute_mst_dsc_configs_for_link(struct drm_atomic_state *state,
 	int count = 0;
 	int i, k, ret;
 	bool debugfs_overwrite = false;
+	uint16_t fec_overhead_multiplier_x1000 = get_fec_overhead_multiplier(dc_link);
 
 	memset(params, 0, sizeof(params));
 
@@ -1146,7 +1161,7 @@ static int compute_mst_dsc_configs_for_link(struct drm_atomic_state *state,
 	/* Try no compression */
 	for (i = 0; i < count; i++) {
 		vars[i + k].aconnector = params[i].aconnector;
-		vars[i + k].pbn = kbps_to_peak_pbn(params[i].bw_range.stream_kbps);
+		vars[i + k].pbn = kbps_to_peak_pbn(params[i].bw_range.stream_kbps, fec_overhead_multiplier_x1000);
 		vars[i + k].dsc_enabled = false;
 		vars[i + k].bpp_x16 = 0;
 		ret = drm_dp_atomic_find_time_slots(state, params[i].port->mgr, params[i].port,
@@ -1165,7 +1180,7 @@ static int compute_mst_dsc_configs_for_link(struct drm_atomic_state *state,
 	/* Try max compression */
 	for (i = 0; i < count; i++) {
 		if (params[i].compression_possible && params[i].clock_force_enable != DSC_CLK_FORCE_DISABLE) {
-			vars[i + k].pbn = kbps_to_peak_pbn(params[i].bw_range.min_kbps);
+			vars[i + k].pbn = kbps_to_peak_pbn(params[i].bw_range.min_kbps, fec_overhead_multiplier_x1000);
 			vars[i + k].dsc_enabled = true;
 			vars[i + k].bpp_x16 = params[i].bw_range.min_target_bpp_x16;
 			ret = drm_dp_atomic_find_time_slots(state, params[i].port->mgr,
@@ -1173,7 +1188,7 @@ static int compute_mst_dsc_configs_for_link(struct drm_atomic_state *state,
 			if (ret < 0)
 				return ret;
 		} else {
-			vars[i + k].pbn = kbps_to_peak_pbn(params[i].bw_range.stream_kbps);
+			vars[i + k].pbn = kbps_to_peak_pbn(params[i].bw_range.stream_kbps, fec_overhead_multiplier_x1000);
 			vars[i + k].dsc_enabled = false;
 			vars[i + k].bpp_x16 = 0;
 			ret = drm_dp_atomic_find_time_slots(state, params[i].port->mgr,
diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.h b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.h
index fa84d34b7373..600d6e221011 100644
--- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.h
+++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.h
@@ -46,6 +46,9 @@
 #define SYNAPTICS_CASCADED_HUB_ID  0x5A
 #define IS_SYNAPTICS_CASCADED_PANAMERA(devName, data) ((IS_SYNAPTICS_PANAMERA(devName) && ((int)data[2] == SYNAPTICS_CASCADED_HUB_ID)) ? 1 : 0)
 
+#define PBN_FEC_OVERHEAD_MULTIPLIER_8B_10B     1031
+#define PBN_FEC_OVERHEAD_MULTIPLIER_128B_132B  1000
+
 enum mst_msg_ready_type {
 	NONE_MSG_RDY_EVENT = 0,
 	DOWN_REP_MSG_RDY_EVENT = 1,
diff --git a/drivers/gpu/drm/amd/display/dc/hwss/dcn10/dcn10_hwseq.c b/drivers/gpu/drm/amd/display/dc/hwss/dcn10/dcn10_hwseq.c
index ff03b1d98aa7..14a902ff3b8a 100644
--- a/drivers/gpu/drm/amd/display/dc/hwss/dcn10/dcn10_hwseq.c
+++ b/drivers/gpu/drm/amd/display/dc/hwss/dcn10/dcn10_hwseq.c
@@ -3589,7 +3589,7 @@ void dcn10_set_cursor_position(struct pipe_ctx *pipe_ctx)
 						(int)hubp->curs_attr.width || pos_cpy.x
 						<= (int)hubp->curs_attr.width +
 						pipe_ctx->plane_state->src_rect.x) {
-						pos_cpy.x = temp_x + viewport_width;
+						pos_cpy.x = 2 * viewport_width - temp_x;
 					}
 				}
 			} else {
@@ -3682,7 +3682,7 @@ void dcn10_set_cursor_position(struct pipe_ctx *pipe_ctx)
 						(int)hubp->curs_attr.width || pos_cpy.x
 						<= (int)hubp->curs_attr.width +
 						pipe_ctx->plane_state->src_rect.x) {
-						pos_cpy.x = 2 * viewport_width - temp_x;
+						pos_cpy.x = temp_x + viewport_width;
 					}
 				}
 			} else {
diff --git a/drivers/gpu/drm/amd/display/dc/resource/dcn321/dcn321_resource.c b/drivers/gpu/drm/amd/display/dc/resource/dcn321/dcn321_resource.c
index 9a3cc0514a36..8e0588b1cf30 100644
--- a/drivers/gpu/drm/amd/display/dc/resource/dcn321/dcn321_resource.c
+++ b/drivers/gpu/drm/amd/display/dc/resource/dcn321/dcn321_resource.c
@@ -1778,6 +1778,9 @@ static bool dcn321_resource_construct(
 	dc->caps.color.mpc.ogam_rom_caps.hlg = 0;
 	dc->caps.color.mpc.ocsc = 1;
 
+	/* Use pipe context based otg sync logic */
+	dc->config.use_pipe_ctx_sync_logic = true;
+
 	dc->config.dc_mode_clk_limit_support = true;
 	dc->config.enable_windowed_mpo_odm = true;
 	/* read VBIOS LTTPR caps */
diff --git a/drivers/gpu/drm/amd/include/mes_v12_api_def.h b/drivers/gpu/drm/amd/include/mes_v12_api_def.h
index 4cf2c9f30b3d..101e2fe962c6 100644
--- a/drivers/gpu/drm/amd/include/mes_v12_api_def.h
+++ b/drivers/gpu/drm/amd/include/mes_v12_api_def.h
@@ -97,6 +97,7 @@ enum MES_QUEUE_TYPE {
 	MES_QUEUE_TYPE_SDMA,
 
 	MES_QUEUE_TYPE_MAX,
+	MES_QUEUE_TYPE_SCHQ = MES_QUEUE_TYPE_MAX,
 };
 
 struct MES_API_STATUS {
@@ -242,8 +243,12 @@ union MESAPI_SET_HW_RESOURCES {
 				uint32_t send_write_data : 1;
 				uint32_t os_tdr_timeout_override : 1;
 				uint32_t use_rs64mem_for_proc_gang_ctx : 1;
+				uint32_t halt_on_misaligned_access : 1;
+				uint32_t use_add_queue_unmap_flag_addr : 1;
+				uint32_t enable_mes_sch_stb_log : 1;
+				uint32_t limit_single_process : 1;
 				uint32_t unmapped_doorbell_handling: 2;
-				uint32_t reserved : 15;
+				uint32_t reserved : 11;
 			};
 			uint32_t uint32_all;
 		};