Merge branch 'drm-next-5.2' of git://people.freedesktop.org/~agd5f/linux into drm-next

- Fixes for powerplay custom profiles
- DC bandwidth clean ups and fixes
- RAS fixes for vega20
- DC atomic resume fix
- Better plane handling in DC
- Freesync improvements
- Misc bug fixes and cleanups

Signed-off-by: Dave Airlie <[e-mail]>
From: Alex Deucher <[e-mail]>
Link: https://patchwork.freedesktop.org/patch/msgid/[e-mail]

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

···
 	/* For pre-DCE11. DCE11 and later are in "struct amdgpu_device->dm" */
 	struct work_struct		hotplug_work;
 	struct amdgpu_irq_src		crtc_irq;
+	struct amdgpu_irq_src		vupdate_irq;
 	struct amdgpu_irq_src		pageflip_irq;
 	struct amdgpu_irq_src		hpd_irq;
 

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

···
 	mutex_init(&adev->virt.vf_errors.lock);
 	hash_init(adev->mn_hash);
 	mutex_init(&adev->lock_reset);
+	mutex_init(&adev->virt.dpm_mutex);
 
 	amdgpu_device_check_arguments(adev);
 

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

···
 #include <drm/amdgpu_drm.h>
 #include "amdgpu.h"
 #include "amdgpu_display.h"
+#include "amdgpu_xgmi.h"
 
 void amdgpu_gem_object_free(struct drm_gem_object *gobj)
 {
···
 	struct amdgpu_device *adev = dev->dev_private;
 	struct drm_amdgpu_gem_op *args = data;
 	struct drm_gem_object *gobj;
+	struct amdgpu_vm_bo_base *base;
 	struct amdgpu_bo *robj;
 	int r;
 
···
 			amdgpu_bo_unreserve(robj);
 			break;
 		}
+		for (base = robj->vm_bo; base; base = base->next)
+			if (amdgpu_xgmi_same_hive(amdgpu_ttm_adev(robj->tbo.bdev),
+				amdgpu_ttm_adev(base->vm->root.base.bo->tbo.bdev))) {
+				r = -EINVAL;
+				amdgpu_bo_unreserve(robj);
+				goto out;
+			}
+
+
 		robj->preferred_domains = args->value & (AMDGPU_GEM_DOMAIN_VRAM |
 							AMDGPU_GEM_DOMAIN_GTT |
 							AMDGPU_GEM_DOMAIN_CPU);

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

···
 		if (adev->pm.dpm_enabled) {
 			dev_info.max_engine_clock = amdgpu_dpm_get_sclk(adev, false) * 10;
 			dev_info.max_memory_clock = amdgpu_dpm_get_mclk(adev, false) * 10;
+		} else if (amdgpu_sriov_vf(adev) && amdgim_is_hwperf(adev) &&
+			   adev->virt.ops->get_pp_clk) {
+			dev_info.max_engine_clock = amdgpu_virt_get_sclk(adev, false) * 10;
+			dev_info.max_memory_clock = amdgpu_virt_get_mclk(adev, false) * 10;
 		} else {
 			dev_info.max_engine_clock = adev->clock.default_sclk * 10;
 			dev_info.max_memory_clock = adev->clock.default_mclk * 10;

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

···
 #define to_amdgpu_encoder(x) container_of(x, struct amdgpu_encoder, base)
 #define to_amdgpu_framebuffer(x) container_of(x, struct amdgpu_framebuffer, base)
 
-#define to_dm_plane_state(x)	container_of(x, struct dm_plane_state, base);
+#define to_dm_plane_state(x)	container_of(x, struct dm_plane_state, base)
 
 #define AMDGPU_MAX_HPD_PINS 6
 #define AMDGPU_MAX_CRTCS 6
···
 	struct amdgpu_flip_work *pflip_works;
 	enum amdgpu_flip_status pflip_status;
 	int deferred_flip_completion;
-	u64 last_flip_vblank;
+	u32 last_flip_vblank;
 	/* pll sharing */
 	struct amdgpu_atom_ss ss;
 	bool ss_enabled;

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

···
 		goto fail;
 	}
 
+        if (amdgpu_sriov_vf(adev)) {
+                if (amdgim_is_hwperf(adev) &&
+                    adev->virt.ops->force_dpm_level) {
+                        mutex_lock(&adev->pm.mutex);
+                        adev->virt.ops->force_dpm_level(adev, level);
+                        mutex_unlock(&adev->pm.mutex);
+                        return count;
+                } else {
+                        return -EINVAL;
+		}
+        }
+
 	if (current_level == level)
 		return count;
 
···
 {
 	struct drm_device *ddev = dev_get_drvdata(dev);
 	struct amdgpu_device *adev = ddev->dev_private;
+
+	if (amdgpu_sriov_vf(adev) && amdgim_is_hwperf(adev) &&
+	    adev->virt.ops->get_pp_clk)
+		return adev->virt.ops->get_pp_clk(adev, PP_SCLK, buf);
 
 	if (is_support_sw_smu(adev))
 		return smu_print_clk_levels(&adev->smu, PP_SCLK, buf);

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

···
 	int ret;
 
 	/*
-	 * Allocate 3M memory aligned to 1M from Frame Buffer (local
-	 * physical).
+	 * According to HW engineer, they prefer the TMR address be "naturally
+	 * aligned" , e.g. the start address be an integer divide of TMR size.
 	 *
 	 * Note: this memory need be reserved till the driver
 	 * uninitializes.
 	 */
-	ret = amdgpu_bo_create_kernel(psp->adev, PSP_TMR_SIZE, 0x100000,
+	ret = amdgpu_bo_create_kernel(psp->adev, PSP_TMR_SIZE, PSP_TMR_SIZE,
 				      AMDGPU_GEM_DOMAIN_VRAM,
 				      &psp->tmr_bo, &psp->tmr_mc_addr, &psp->tmr_buf);
 

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

···
 	return 0;
 }
 
+/* Only used in device probe stage and called only once. */
+int amdgpu_ras_feature_enable_on_boot(struct amdgpu_device *adev,
+		struct ras_common_if *head, bool enable)
+{
+	struct amdgpu_ras *con = amdgpu_ras_get_context(adev);
+	int ret;
+
+	if (!con)
+		return -EINVAL;
+
+	if (con->flags & AMDGPU_RAS_FLAG_INIT_BY_VBIOS) {
+		/* If ras is enabled by vbios, we set up ras object first in
+		 * both case. For enable, that is all what we need do. For
+		 * disable, we need perform a ras TA disable cmd after that.
+		 */
+		ret = __amdgpu_ras_feature_enable(adev, head, 1);
+		if (ret)
+			return ret;
+
+		if (!enable)
+			ret = amdgpu_ras_feature_enable(adev, head, 0);
+	} else
+		ret = amdgpu_ras_feature_enable(adev, head, enable);
+
+	return ret;
+}
+
 static int amdgpu_ras_disable_all_features(struct amdgpu_device *adev,
 		bool bypass)
 {
···
 	struct amdgpu_ras *con = amdgpu_ras_get_context(adev);
 	int ras_block_count = AMDGPU_RAS_BLOCK_COUNT;
 	int i;
+	const enum amdgpu_ras_error_type default_ras_type =
+		AMDGPU_RAS_ERROR__NONE;
 
 	for (i = 0; i < ras_block_count; i++) {
 		struct ras_common_if head = {
 			.block = i,
-			.type = AMDGPU_RAS_ERROR__MULTI_UNCORRECTABLE,
+			.type = default_ras_type,
 			.sub_block_index = 0,
 		};
 		strcpy(head.name, ras_block_str(i));
···
 
 	amdgpu_ras_mask &= AMDGPU_RAS_BLOCK_MASK;
 
-	if (con->flags & AMDGPU_RAS_FLAG_INIT_BY_VBIOS)
-		amdgpu_ras_enable_all_features(adev, 1);
-
 	if (amdgpu_ras_fs_init(adev))
 		goto fs_out;
 
···
 	if (!con)
 		return;
 
-	/* We enable ras on all hw_supported block, but as boot parameter might
-	 * disable some of them and one or more IP has not implemented yet.
-	 * So we disable them on behalf.
-	 */
 	if (con->flags & AMDGPU_RAS_FLAG_INIT_BY_VBIOS) {
+		/* Set up all other IPs which are not implemented. There is a
+		 * tricky thing that IP's actual ras error type should be
+		 * MULTI_UNCORRECTABLE, but as driver does not handle it, so
+		 * ERROR_NONE make sense anyway.
+		 */
+		amdgpu_ras_enable_all_features(adev, 1);
+
+		/* We enable ras on all hw_supported block, but as boot
+		 * parameter might disable some of them and one or more IP has
+		 * not implemented yet. So we disable them on behalf.
+		 */
 		list_for_each_entry_safe(obj, tmp, &con->head, node) {
 			if (!amdgpu_ras_is_supported(adev, obj->head.block)) {
 				amdgpu_ras_feature_enable(adev, &obj->head, 0);
 				/* there should be no any reference. */
 				WARN_ON(alive_obj(obj));
 			}
-		};
+		}
 	}
 }
 

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

···
 int amdgpu_ras_feature_enable(struct amdgpu_device *adev,
 		struct ras_common_if *head, bool enable);
 
+int amdgpu_ras_feature_enable_on_boot(struct amdgpu_device *adev,
+		struct ras_common_if *head, bool enable);
+
 int amdgpu_ras_sysfs_create(struct amdgpu_device *adev,
 		struct ras_fs_if *head);
 

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

···
 #define AMDGPU_MAX_SDMA_INSTANCES		2
 
 enum amdgpu_sdma_irq {
-	AMDGPU_SDMA_IRQ_TRAP0 = 0,
-	AMDGPU_SDMA_IRQ_TRAP1,
-	AMDGPU_SDMA_IRQ_ECC0,
-	AMDGPU_SDMA_IRQ_ECC1,
-
+	AMDGPU_SDMA_IRQ_INSTANCE0  = 0,
+	AMDGPU_SDMA_IRQ_INSTANCE1,
 	AMDGPU_SDMA_IRQ_LAST
 };
 

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

···
 	}
 }
 
+static uint32_t parse_clk(char *buf, bool min)
+{
+        char *ptr = buf;
+        uint32_t clk = 0;
+
+        do {
+                ptr = strchr(ptr, ':');
+                if (!ptr)
+                        break;
+                ptr+=2;
+                clk = simple_strtoul(ptr, NULL, 10);
+        } while (!min);
+
+        return clk * 100;
+}
+
+uint32_t amdgpu_virt_get_sclk(struct amdgpu_device *adev, bool lowest)
+{
+	char *buf = NULL;
+	uint32_t clk = 0;
+
+	buf = kzalloc(PAGE_SIZE, GFP_KERNEL);
+	if (!buf)
+		return -ENOMEM;
+
+	adev->virt.ops->get_pp_clk(adev, PP_SCLK, buf);
+	clk = parse_clk(buf, lowest);
+
+	kfree(buf);
+
+	return clk;
+}
+
+uint32_t amdgpu_virt_get_mclk(struct amdgpu_device *adev, bool lowest)
+{
+	char *buf = NULL;
+	uint32_t clk = 0;
+
+	buf = kzalloc(PAGE_SIZE, GFP_KERNEL);
+	if (!buf)
+		return -ENOMEM;
+
+	adev->virt.ops->get_pp_clk(adev, PP_MCLK, buf);
+	clk = parse_clk(buf, lowest);
+
+	kfree(buf);
+
+	return clk;
+}
 

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

···
 	int (*reset_gpu)(struct amdgpu_device *adev);
 	int (*wait_reset)(struct amdgpu_device *adev);
 	void (*trans_msg)(struct amdgpu_device *adev, u32 req, u32 data1, u32 data2, u32 data3);
+	int (*get_pp_clk)(struct amdgpu_device *adev, u32 type, char *buf);
+	int (*force_dpm_level)(struct amdgpu_device *adev, u32 level);
 };
 
 /*
···
 	AMDGIM_FEATURE_GIM_LOAD_UCODES   = 0x2,
 	/* VRAM LOST by GIM */
 	AMDGIM_FEATURE_GIM_FLR_VRAMLOST = 0x4,
+	/* HW PERF SIM in GIM */
+	AMDGIM_FEATURE_HW_PERF_SIMULATION = (1 << 3),
 };
 
 struct amd_sriov_msg_pf2vf_info_header {
···
 	struct amdgpu_vf_error_buffer   vf_errors;
 	struct amdgpu_virt_fw_reserve	fw_reserve;
 	uint32_t gim_feature;
+	/* protect DPM events to GIM */
+	struct mutex                    dpm_mutex;
 };
 
 #define amdgpu_sriov_enabled(adev) \
···
 #endif
 }
 
+#define amdgim_is_hwperf(adev) \
+	((adev)->virt.gim_feature & AMDGIM_FEATURE_HW_PERF_SIMULATION)
+
 bool amdgpu_virt_mmio_blocked(struct amdgpu_device *adev);
 void amdgpu_virt_init_setting(struct amdgpu_device *adev);
 uint32_t amdgpu_virt_kiq_rreg(struct amdgpu_device *adev, uint32_t reg);
···
 					unsigned int key,
 					unsigned int chksum);
 void amdgpu_virt_init_data_exchange(struct amdgpu_device *adev);
+uint32_t amdgpu_virt_get_sclk(struct amdgpu_device *adev, bool lowest);
+uint32_t amdgpu_virt_get_mclk(struct amdgpu_device *adev, bool lowest);
 
 #endif

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

···
 
 	addr = 0;
 	if (ats_entries) {
-		uint64_t ats_value;
+		uint64_t value = 0, flags;
 
-		ats_value = AMDGPU_PTE_DEFAULT_ATC;
-		if (level != AMDGPU_VM_PTB)
-			ats_value |= AMDGPU_PDE_PTE;
+		flags = AMDGPU_PTE_DEFAULT_ATC;
+		if (level != AMDGPU_VM_PTB) {
+			/* Handle leaf PDEs as PTEs */
+			flags |= AMDGPU_PDE_PTE;
+			amdgpu_gmc_get_vm_pde(adev, level, &value, &flags);
+		}
 
 		r = vm->update_funcs->update(&params, bo, addr, 0, ats_entries,
-					     0, ats_value);
+					     value, flags);
 		if (r)
 			return r;
 
···
 	}
 
 	if (entries) {
-		uint64_t value = 0;
+		uint64_t value = 0, flags = 0;
 
-		/* Workaround for fault priority problem on GMC9 */
-		if (level == AMDGPU_VM_PTB &&
-		    adev->asic_type >= CHIP_VEGA10)
-			value = AMDGPU_PTE_EXECUTABLE;
+		if (adev->asic_type >= CHIP_VEGA10) {
+			if (level != AMDGPU_VM_PTB) {
+				/* Handle leaf PDEs as PTEs */
+				flags |= AMDGPU_PDE_PTE;
+				amdgpu_gmc_get_vm_pde(adev, level,
+						      &value, &flags);
+			} else {
+				/* Workaround for fault priority problem on GMC9 */
+				flags = AMDGPU_PTE_EXECUTABLE;
+			}
+		}
 
 		r = vm->update_funcs->update(&params, bo, addr, 0, entries,
-					     0, value);
+					     value, flags);
 		if (r)
 			return r;
 	}
···
 	INIT_LIST_HEAD(&bo_va->valids);
 	INIT_LIST_HEAD(&bo_va->invalids);
 
-	if (bo && amdgpu_xgmi_same_hive(adev, amdgpu_ttm_adev(bo->tbo.bdev))) {
+	if (bo && amdgpu_xgmi_same_hive(adev, amdgpu_ttm_adev(bo->tbo.bdev)) &&
+	    (bo->preferred_domains & AMDGPU_GEM_DOMAIN_VRAM)) {
 		bo_va->is_xgmi = true;
 		mutex_lock(&adev->vm_manager.lock_pstate);
 		/* Power up XGMI if it can be potentially used */

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

···
 	const struct amdgpu_vm_pte_funcs	*vm_pte_funcs;
 	struct drm_sched_rq			*vm_pte_rqs[AMDGPU_MAX_RINGS];
 	unsigned				vm_pte_num_rqs;
+	struct amdgpu_ring			*page_fault;
 
 	/* partial resident texture handling */
 	spinlock_t				prt_lock;

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

···
 		r = amdgpu_ring_init(adev, ring, 1024,
 				     &adev->sdma.trap_irq,
 				     (i == 0) ?
-				     AMDGPU_SDMA_IRQ_TRAP0 :
-				     AMDGPU_SDMA_IRQ_TRAP1);
+				     AMDGPU_SDMA_IRQ_INSTANCE0 :
+				     AMDGPU_SDMA_IRQ_INSTANCE1);
 		if (r)
 			return r;
 	}
···
 	u32 sdma_cntl;
 
 	switch (type) {
-	case AMDGPU_SDMA_IRQ_TRAP0:
+	case AMDGPU_SDMA_IRQ_INSTANCE0:
 		switch (state) {
 		case AMDGPU_IRQ_STATE_DISABLE:
 			sdma_cntl = RREG32(mmSDMA0_CNTL + SDMA0_REGISTER_OFFSET);
···
 			break;
 		}
 		break;
-	case AMDGPU_SDMA_IRQ_TRAP1:
+	case AMDGPU_SDMA_IRQ_INSTANCE1:
 		switch (state) {
 		case AMDGPU_IRQ_STATE_DISABLE:
 			sdma_cntl = RREG32(mmSDMA0_CNTL + SDMA1_REGISTER_OFFSET);

drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c

···
 	int r;
 
 	if (!amdgpu_ras_is_supported(adev, AMDGPU_RAS_BLOCK__GFX)) {
-		amdgpu_ras_feature_enable(adev, &ras_block, 0);
+		amdgpu_ras_feature_enable_on_boot(adev, &ras_block, 0);
 		return 0;
 	}
 
···
 
 	**ras_if = ras_block;
 
-	r = amdgpu_ras_feature_enable(adev, *ras_if, 1);
+	r = amdgpu_ras_feature_enable_on_boot(adev, *ras_if, 1);
 	if (r)
 		goto feature;
 
···
 				  struct amdgpu_irq_src *source,
 				  struct amdgpu_iv_entry *entry)
 {
+	struct ras_common_if *ras_if = adev->gfx.ras_if;
 	struct ras_dispatch_if ih_data = {
-		.head = *adev->gfx.ras_if,
 		.entry = entry,
 	};
+
+	if (!ras_if)
+		return 0;
+
+	ih_data.head = *ras_if;
+
 	DRM_ERROR("CP ECC ERROR IRQ\n");
 	amdgpu_ras_interrupt_dispatch(adev, &ih_data);
 	return 0;

drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c

···
 		struct amdgpu_irq_src *source,
 		struct amdgpu_iv_entry *entry)
 {
+	struct ras_common_if *ras_if = adev->gmc.ras_if;
 	struct ras_dispatch_if ih_data = {
-		.head = *adev->gmc.ras_if,
 		.entry = entry,
 	};
+
+	if (!ras_if)
+		return 0;
+
+	ih_data.head = *ras_if;
+
 	amdgpu_ras_interrupt_dispatch(adev, &ih_data);
 	return 0;
 }
···
 	int r;
 
 	if (!amdgpu_ras_is_supported(adev, AMDGPU_RAS_BLOCK__UMC)) {
-		amdgpu_ras_feature_enable(adev, &ras_block, 0);
+		amdgpu_ras_feature_enable_on_boot(adev, &ras_block, 0);
 		return 0;
 	}
 	/* handle resume path. */
···
 
 	**ras_if = ras_block;
 
-	r = amdgpu_ras_feature_enable(adev, *ras_if, 1);
+	r = amdgpu_ras_feature_enable_on_boot(adev, *ras_if, 1);
 	if (r)
 		goto feature;
 

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

···
 	xgpu_ai_mailbox_set_valid(adev, false);
 }
 
+static int xgpu_ai_get_pp_clk(struct amdgpu_device *adev, u32 type, char *buf)
+{
+        int r = 0;
+        u32 req, val, size;
+
+        if (!amdgim_is_hwperf(adev) || buf == NULL)
+                return -EBADRQC;
+
+        switch(type) {
+        case PP_SCLK:
+                req = IDH_IRQ_GET_PP_SCLK;
+                break;
+        case PP_MCLK:
+                req = IDH_IRQ_GET_PP_MCLK;
+                break;
+        default:
+                return -EBADRQC;
+        }
+
+        mutex_lock(&adev->virt.dpm_mutex);
+
+        xgpu_ai_mailbox_trans_msg(adev, req, 0, 0, 0);
+
+        r = xgpu_ai_poll_msg(adev, IDH_SUCCESS);
+        if (!r && adev->fw_vram_usage.va != NULL) {
+                val = RREG32_NO_KIQ(
+                        SOC15_REG_OFFSET(NBIO, 0,
+                                         mmBIF_BX_PF0_MAILBOX_MSGBUF_RCV_DW1));
+                size = strnlen((((char *)adev->virt.fw_reserve.p_pf2vf) +
+                                val), PAGE_SIZE);
+
+                if (size < PAGE_SIZE)
+                        strcpy(buf,((char *)adev->virt.fw_reserve.p_pf2vf + val));
+                else
+                        size = 0;
+
+                r = size;
+                goto out;
+        }
+
+        r = xgpu_ai_poll_msg(adev, IDH_FAIL);
+        if(r)
+                pr_info("%s DPM request failed",
+                        (type == PP_SCLK)? "SCLK" : "MCLK");
+
+out:
+        mutex_unlock(&adev->virt.dpm_mutex);
+        return r;
+}
+
+static int xgpu_ai_force_dpm_level(struct amdgpu_device *adev, u32 level)
+{
+        int r = 0;
+        u32 req = IDH_IRQ_FORCE_DPM_LEVEL;
+
+        if (!amdgim_is_hwperf(adev))
+                return -EBADRQC;
+
+        mutex_lock(&adev->virt.dpm_mutex);
+        xgpu_ai_mailbox_trans_msg(adev, req, level, 0, 0);
+
+        r = xgpu_ai_poll_msg(adev, IDH_SUCCESS);
+        if (!r)
+                goto out;
+
+        r = xgpu_ai_poll_msg(adev, IDH_FAIL);
+        if (!r)
+                pr_info("DPM request failed");
+        else
+                pr_info("Mailbox is broken");
+
+out:
+        mutex_unlock(&adev->virt.dpm_mutex);
+        return r;
+}
+
 static int xgpu_ai_send_access_requests(struct amdgpu_device *adev,
 					enum idh_request req)
 {
···
 	.reset_gpu = xgpu_ai_request_reset,
 	.wait_reset = NULL,
 	.trans_msg = xgpu_ai_mailbox_trans_msg,
+	.get_pp_clk = xgpu_ai_get_pp_clk,
+	.force_dpm_level = xgpu_ai_force_dpm_level,
 };

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

···
 	IDH_REL_GPU_FINI_ACCESS,
 	IDH_REQ_GPU_RESET_ACCESS,
 
+	IDH_IRQ_FORCE_DPM_LEVEL = 10,
+	IDH_IRQ_GET_PP_SCLK,
+	IDH_IRQ_GET_PP_MCLK,
+
 	IDH_LOG_VF_ERROR       = 200,
 };
 
···
 	IDH_READY_TO_ACCESS_GPU,
 	IDH_FLR_NOTIFICATION,
 	IDH_FLR_NOTIFICATION_CMPL,
+	IDH_SUCCESS,
+	IDH_FAIL,
 	IDH_EVENT_MAX
 };
 

drivers/gpu/drm/amd/amdgpu/psp_v11_0.c

···
 	for (i = 0; i < topology_info_input->num_nodes; i++) {
 		topology_info_input->nodes[i].node_id = topology->nodes[i].node_id;
 		topology_info_input->nodes[i].num_hops = topology->nodes[i].num_hops;
-		topology_info_input->nodes[i].is_sharing_enabled = topology->nodes[i].is_sharing_enabled;
+		topology_info_input->nodes[i].is_sharing_enabled = 1;
 		topology_info_input->nodes[i].sdma_engine = topology->nodes[i].sdma_engine;
 	}
 

drivers/gpu/drm/amd/amdgpu/sdma_v2_4.c

···
 		r = amdgpu_ring_init(adev, ring, 1024,
 				     &adev->sdma.trap_irq,
 				     (i == 0) ?
-				     AMDGPU_SDMA_IRQ_TRAP0 :
-				     AMDGPU_SDMA_IRQ_TRAP1);
+				     AMDGPU_SDMA_IRQ_INSTANCE0 :
+				     AMDGPU_SDMA_IRQ_INSTANCE1);
 		if (r)
 			return r;
 	}
···
 	u32 sdma_cntl;
 
 	switch (type) {
-	case AMDGPU_SDMA_IRQ_TRAP0:
+	case AMDGPU_SDMA_IRQ_INSTANCE0:
 		switch (state) {
 		case AMDGPU_IRQ_STATE_DISABLE:
 			sdma_cntl = RREG32(mmSDMA0_CNTL + SDMA0_REGISTER_OFFSET);
···
 			break;
 		}
 		break;
-	case AMDGPU_SDMA_IRQ_TRAP1:
+	case AMDGPU_SDMA_IRQ_INSTANCE1:
 		switch (state) {
 		case AMDGPU_IRQ_STATE_DISABLE:
 			sdma_cntl = RREG32(mmSDMA0_CNTL + SDMA1_REGISTER_OFFSET);

drivers/gpu/drm/amd/amdgpu/sdma_v3_0.c

···
 		r = amdgpu_ring_init(adev, ring, 1024,
 				     &adev->sdma.trap_irq,
 				     (i == 0) ?
-				     AMDGPU_SDMA_IRQ_TRAP0 :
-				     AMDGPU_SDMA_IRQ_TRAP1);
+				     AMDGPU_SDMA_IRQ_INSTANCE0 :
+				     AMDGPU_SDMA_IRQ_INSTANCE1);
 		if (r)
 			return r;
 	}
···
 	u32 sdma_cntl;
 
 	switch (type) {
-	case AMDGPU_SDMA_IRQ_TRAP0:
+	case AMDGPU_SDMA_IRQ_INSTANCE0:
 		switch (state) {
 		case AMDGPU_IRQ_STATE_DISABLE:
 			sdma_cntl = RREG32(mmSDMA0_CNTL + SDMA0_REGISTER_OFFSET);
···
 			break;
 		}
 		break;
-	case AMDGPU_SDMA_IRQ_TRAP1:
+	case AMDGPU_SDMA_IRQ_INSTANCE1:
 		switch (state) {
 		case AMDGPU_IRQ_STATE_DISABLE:
 			sdma_cntl = RREG32(mmSDMA0_CNTL + SDMA1_REGISTER_OFFSET);

drivers/gpu/drm/amd/amdgpu/sdma_v4_0.c

···
 	int r;
 
 	if (!amdgpu_ras_is_supported(adev, AMDGPU_RAS_BLOCK__SDMA)) {
-		amdgpu_ras_feature_enable(adev, &ras_block, 0);
+		amdgpu_ras_feature_enable_on_boot(adev, &ras_block, 0);
 		return 0;
 	}
 
···
 
 	**ras_if = ras_block;
 
-	r = amdgpu_ras_feature_enable(adev, *ras_if, 1);
+	r = amdgpu_ras_feature_enable_on_boot(adev, *ras_if, 1);
 	if (r)
 		goto feature;
 
···
 	if (r)
 		goto sysfs;
 resume:
-	r = amdgpu_irq_get(adev, &adev->sdma.ecc_irq, AMDGPU_SDMA_IRQ_ECC0);
+	r = amdgpu_irq_get(adev, &adev->sdma.ecc_irq, AMDGPU_SDMA_IRQ_INSTANCE0);
 	if (r)
 		goto irq;
 
-	r = amdgpu_irq_get(adev, &adev->sdma.ecc_irq, AMDGPU_SDMA_IRQ_ECC1);
+	r = amdgpu_irq_get(adev, &adev->sdma.ecc_irq, AMDGPU_SDMA_IRQ_INSTANCE1);
 	if (r) {
-		amdgpu_irq_put(adev, &adev->sdma.ecc_irq, AMDGPU_SDMA_IRQ_ECC0);
+		amdgpu_irq_put(adev, &adev->sdma.ecc_irq, AMDGPU_SDMA_IRQ_INSTANCE0);
 		goto irq;
 	}
 
···
 		r = amdgpu_ring_init(adev, ring, 1024,
 				     &adev->sdma.trap_irq,
 				     (i == 0) ?
-				     AMDGPU_SDMA_IRQ_TRAP0 :
-				     AMDGPU_SDMA_IRQ_TRAP1);
+				     AMDGPU_SDMA_IRQ_INSTANCE0 :
+				     AMDGPU_SDMA_IRQ_INSTANCE1);
 		if (r)
 			return r;
 
···
 			r = amdgpu_ring_init(adev, ring, 1024,
 					     &adev->sdma.trap_irq,
 					     (i == 0) ?
-					     AMDGPU_SDMA_IRQ_TRAP0 :
-					     AMDGPU_SDMA_IRQ_TRAP1);
+					     AMDGPU_SDMA_IRQ_INSTANCE0 :
+					     AMDGPU_SDMA_IRQ_INSTANCE1);
 			if (r)
 				return r;
 		}
···
 	if (amdgpu_sriov_vf(adev))
 		return 0;
 
-	amdgpu_irq_put(adev, &adev->sdma.ecc_irq, AMDGPU_SDMA_IRQ_ECC0);
-	amdgpu_irq_put(adev, &adev->sdma.ecc_irq, AMDGPU_SDMA_IRQ_ECC1);
+	amdgpu_irq_put(adev, &adev->sdma.ecc_irq, AMDGPU_SDMA_IRQ_INSTANCE0);
+	amdgpu_irq_put(adev, &adev->sdma.ecc_irq, AMDGPU_SDMA_IRQ_INSTANCE1);
 
 	sdma_v4_0_ctx_switch_enable(adev, false);
 	sdma_v4_0_enable(adev, false);
···
 					unsigned type,
 					enum amdgpu_interrupt_state state)
 {
-	unsigned int instance = (type == AMDGPU_SDMA_IRQ_TRAP0) ? 0 : 1;
 	u32 sdma_cntl;
 
-	sdma_cntl = RREG32_SDMA(instance, mmSDMA0_CNTL);
+	sdma_cntl = RREG32_SDMA(type, mmSDMA0_CNTL);
 	sdma_cntl = REG_SET_FIELD(sdma_cntl, SDMA0_CNTL, TRAP_ENABLE,
 		       state == AMDGPU_IRQ_STATE_ENABLE ? 1 : 0);
-	WREG32_SDMA(instance, mmSDMA0_CNTL, sdma_cntl);
+	WREG32_SDMA(type, mmSDMA0_CNTL, sdma_cntl);
 
 	return 0;
 }
···
 				      struct amdgpu_irq_src *source,
 				      struct amdgpu_iv_entry *entry)
 {
+	struct ras_common_if *ras_if = adev->sdma.ras_if;
 	struct ras_dispatch_if ih_data = {
-		.head = *adev->sdma.ras_if,
 		.entry = entry,
 	};
+
+	if (!ras_if)
+		return 0;
+
+	ih_data.head = *ras_if;
+
 	amdgpu_ras_interrupt_dispatch(adev, &ih_data);
 	return 0;
 }
···
 {
 	u32 sdma_edc_config;
 
-	u32 reg_offset = (type == AMDGPU_SDMA_IRQ_ECC0) ?
+	u32 reg_offset = (type == AMDGPU_SDMA_IRQ_INSTANCE0) ?
 		sdma_v4_0_get_reg_offset(adev, 0, mmSDMA0_EDC_CONFIG) :
 		sdma_v4_0_get_reg_offset(adev, 1, mmSDMA0_EDC_CONFIG);
 
···
 				&sched->sched_rq[DRM_SCHED_PRIORITY_KERNEL];
 		}
 		adev->vm_manager.vm_pte_num_rqs = adev->sdma.num_instances - 1;
+		adev->vm_manager.page_fault = &adev->sdma.instance[0].page;
 	} else {
 		for (i = 0; i < adev->sdma.num_instances; i++) {
 			sched = &adev->sdma.instance[i].ring.sched;

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

···
 		r = amdgpu_ring_init(adev, ring, 1024,
 				     &adev->sdma.trap_irq,
 				     (i == 0) ?
-				     AMDGPU_SDMA_IRQ_TRAP0 :
-				     AMDGPU_SDMA_IRQ_TRAP1);
+				     AMDGPU_SDMA_IRQ_INSTANCE0 :
+				     AMDGPU_SDMA_IRQ_INSTANCE1);
 		if (r)
 			return r;
 	}
···
 	u32 sdma_cntl;
 
 	switch (type) {
-	case AMDGPU_SDMA_IRQ_TRAP0:
+	case AMDGPU_SDMA_IRQ_INSTANCE0:
 		switch (state) {
 		case AMDGPU_IRQ_STATE_DISABLE:
 			sdma_cntl = RREG32(DMA_CNTL + DMA0_REGISTER_OFFSET);
···
 			break;
 		}
 		break;
-	case AMDGPU_SDMA_IRQ_TRAP1:
+	case AMDGPU_SDMA_IRQ_INSTANCE1:
 		switch (state) {
 		case AMDGPU_IRQ_STATE_DISABLE:
 			sdma_cntl = RREG32(DMA_CNTL + DMA1_REGISTER_OFFSET);

drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c

···
 
 static int amdgpu_dm_plane_init(struct amdgpu_display_manager *dm,
 				struct drm_plane *plane,
-				unsigned long possible_crtcs);
+				unsigned long possible_crtcs,
+				const struct dc_plane_cap *plane_cap);
 static int amdgpu_dm_crtc_init(struct amdgpu_display_manager *dm,
 			       struct drm_plane *plane,
 			       uint32_t link_index);
···
 	return NULL;
 }
 
+static inline bool amdgpu_dm_vrr_active(struct dm_crtc_state *dm_state)
+{
+	return dm_state->freesync_config.state == VRR_STATE_ACTIVE_VARIABLE ||
+	       dm_state->freesync_config.state == VRR_STATE_ACTIVE_FIXED;
+}
+
 static void dm_pflip_high_irq(void *interrupt_params)
 {
 	struct amdgpu_crtc *amdgpu_crtc;
 	struct common_irq_params *irq_params = interrupt_params;
 	struct amdgpu_device *adev = irq_params->adev;
 	unsigned long flags;
+	struct drm_pending_vblank_event *e;
+	struct dm_crtc_state *acrtc_state;
+	uint32_t vpos, hpos, v_blank_start, v_blank_end;
+	bool vrr_active;
 
 	amdgpu_crtc = get_crtc_by_otg_inst(adev, irq_params->irq_src - IRQ_TYPE_PFLIP);
 
···
 		return;
 	}
 
-	/* Update to correct count(s) if racing with vblank irq */
-	amdgpu_crtc->last_flip_vblank = drm_crtc_accurate_vblank_count(&amdgpu_crtc->base);
+	/* page flip completed. */
+	e = amdgpu_crtc->event;
+	amdgpu_crtc->event = NULL;
 
-	/* wake up userspace */
-	if (amdgpu_crtc->event) {
-		drm_crtc_send_vblank_event(&amdgpu_crtc->base, amdgpu_crtc->event);
-
-		/* page flip completed. clean up */
-		amdgpu_crtc->event = NULL;
-
-	} else
+	if (!e)
 		WARN_ON(1);
+
+	acrtc_state = to_dm_crtc_state(amdgpu_crtc->base.state);
+	vrr_active = amdgpu_dm_vrr_active(acrtc_state);
+
+	/* Fixed refresh rate, or VRR scanout position outside front-porch? */
+	if (!vrr_active ||
+	    !dc_stream_get_scanoutpos(acrtc_state->stream, &v_blank_start,
+				      &v_blank_end, &hpos, &vpos) ||
+	    (vpos < v_blank_start)) {
+		/* Update to correct count and vblank timestamp if racing with
+		 * vblank irq. This also updates to the correct vblank timestamp
+		 * even in VRR mode, as scanout is past the front-porch atm.
+		 */
+		drm_crtc_accurate_vblank_count(&amdgpu_crtc->base);
+
+		/* Wake up userspace by sending the pageflip event with proper
+		 * count and timestamp of vblank of flip completion.
+		 */
+		if (e) {
+			drm_crtc_send_vblank_event(&amdgpu_crtc->base, e);
+
+			/* Event sent, so done with vblank for this flip */
+			drm_crtc_vblank_put(&amdgpu_crtc->base);
+		}
+	} else if (e) {
+		/* VRR active and inside front-porch: vblank count and
+		 * timestamp for pageflip event will only be up to date after
+		 * drm_crtc_handle_vblank() has been executed from late vblank
+		 * irq handler after start of back-porch (vline 0). We queue the
+		 * pageflip event for send-out by drm_crtc_handle_vblank() with
+		 * updated timestamp and count, once it runs after us.
+		 *
+		 * We need to open-code this instead of using the helper
+		 * drm_crtc_arm_vblank_event(), as that helper would
+		 * call drm_crtc_accurate_vblank_count(), which we must
+		 * not call in VRR mode while we are in front-porch!
+		 */
+
+		/* sequence will be replaced by real count during send-out. */
+		e->sequence = drm_crtc_vblank_count(&amdgpu_crtc->base);
+		e->pipe = amdgpu_crtc->crtc_id;
+
+		list_add_tail(&e->base.link, &adev->ddev->vblank_event_list);
+		e = NULL;
+	}
+
+	/* Keep track of vblank of this flip for flip throttling. We use the
+	 * cooked hw counter, as that one incremented at start of this vblank
+	 * of pageflip completion, so last_flip_vblank is the forbidden count
+	 * for queueing new pageflips if vsync + VRR is enabled.
+	 */
+	amdgpu_crtc->last_flip_vblank = amdgpu_get_vblank_counter_kms(adev->ddev,
+							amdgpu_crtc->crtc_id);
 
 	amdgpu_crtc->pflip_status = AMDGPU_FLIP_NONE;
 	spin_unlock_irqrestore(&adev->ddev->event_lock, flags);
 
-	DRM_DEBUG_DRIVER("%s - crtc :%d[%p], pflip_stat:AMDGPU_FLIP_NONE\n",
-					__func__, amdgpu_crtc->crtc_id, amdgpu_crtc);
+	DRM_DEBUG_DRIVER("crtc:%d[%p], pflip_stat:AMDGPU_FLIP_NONE, vrr[%d]-fp %d\n",
+			 amdgpu_crtc->crtc_id, amdgpu_crtc,
+			 vrr_active, (int) !e);
+}
 
-	drm_crtc_vblank_put(&amdgpu_crtc->base);
+static void dm_vupdate_high_irq(void *interrupt_params)
+{
+	struct common_irq_params *irq_params = interrupt_params;
+	struct amdgpu_device *adev = irq_params->adev;
+	struct amdgpu_crtc *acrtc;
+	struct dm_crtc_state *acrtc_state;
+
+	acrtc = get_crtc_by_otg_inst(adev, irq_params->irq_src - IRQ_TYPE_VUPDATE);
+
+	if (acrtc) {
+		acrtc_state = to_dm_crtc_state(acrtc->base.state);
+
+		DRM_DEBUG_DRIVER("crtc:%d, vupdate-vrr:%d\n", acrtc->crtc_id,
+				 amdgpu_dm_vrr_active(acrtc_state));
+
+		/* Core vblank handling is done here after end of front-porch in
+		 * vrr mode, as vblank timestamping will give valid results
+		 * while now done after front-porch. This will also deliver
+		 * page-flip completion events that have been queued to us
+		 * if a pageflip happened inside front-porch.
+		 */
+		if (amdgpu_dm_vrr_active(acrtc_state))
+			drm_crtc_handle_vblank(&acrtc->base);
+	}
 }
 
 static void dm_crtc_high_irq(void *interrupt_params)
···
 	acrtc = get_crtc_by_otg_inst(adev, irq_params->irq_src - IRQ_TYPE_VBLANK);
 
 	if (acrtc) {
-		drm_crtc_handle_vblank(&acrtc->base);
-		amdgpu_dm_crtc_handle_crc_irq(&acrtc->base);
-
 		acrtc_state = to_dm_crtc_state(acrtc->base.state);
+
+		DRM_DEBUG_DRIVER("crtc:%d, vupdate-vrr:%d\n", acrtc->crtc_id,
+				 amdgpu_dm_vrr_active(acrtc_state));
+
+		/* Core vblank handling at start of front-porch is only possible
+		 * in non-vrr mode, as only there vblank timestamping will give
+		 * valid results while done in front-porch. Otherwise defer it
+		 * to dm_vupdate_high_irq after end of front-porch.
+		 */
+		if (!amdgpu_dm_vrr_active(acrtc_state))
+			drm_crtc_handle_vblank(&acrtc->base);
+
+		/* Following stuff must happen at start of vblank, for crc
+		 * computation and below-the-range btr support in vrr mode.
+		 */
+		amdgpu_dm_crtc_handle_crc_irq(&acrtc->base);
 
 		if (acrtc_state->stream &&
 		    acrtc_state->vrr_params.supported &&
···
 	struct drm_plane *plane;
 	struct drm_plane_state *new_plane_state;
 	struct dm_plane_state *dm_new_plane_state;
+	struct dm_atomic_state *dm_state = to_dm_atomic_state(dm->atomic_obj.state);
 	enum dc_connection_type new_connection_type = dc_connection_none;
 	int i;
+
+	/* Recreate dc_state - DC invalidates it when setting power state to S3. */
+	dc_release_state(dm_state->context);
+	dm_state->context = dc_create_state(dm->dc);
+	/* TODO: Remove dc_state->dccg, use dc->dccg directly. */
+	dc_resource_state_construct(dm->dc, dm_state->context);
 
 	/* power on hardware */
 	dc_set_power_state(dm->dc, DC_ACPI_CM_POWER_STATE_D0);
···
 				dm_crtc_high_irq, c_irq_params);
 	}
 
+	/* Use VUPDATE interrupt */
+	for (i = VISLANDS30_IV_SRCID_D1_V_UPDATE_INT; i <= VISLANDS30_IV_SRCID_D6_V_UPDATE_INT; i += 2) {
+		r = amdgpu_irq_add_id(adev, client_id, i, &adev->vupdate_irq);
+		if (r) {
+			DRM_ERROR("Failed to add vupdate irq id!\n");
+			return r;
+		}
+
+		int_params.int_context = INTERRUPT_HIGH_IRQ_CONTEXT;
+		int_params.irq_source =
+			dc_interrupt_to_irq_source(dc, i, 0);
+
+		c_irq_params = &adev->dm.vupdate_params[int_params.irq_source - DC_IRQ_SOURCE_VUPDATE1];
+
+		c_irq_params->adev = adev;
+		c_irq_params->irq_src = int_params.irq_source;
+
+		amdgpu_dm_irq_register_interrupt(adev, &int_params,
+				dm_vupdate_high_irq, c_irq_params);
+	}
+
 	/* Use GRPH_PFLIP interrupt */
 	for (i = VISLANDS30_IV_SRCID_D1_GRPH_PFLIP;
 			i <= VISLANDS30_IV_SRCID_D6_GRPH_PFLIP; i += 2) {
···
 
 		amdgpu_dm_irq_register_interrupt(adev, &int_params,
 				dm_crtc_high_irq, c_irq_params);
+	}
+
+	/* Use VUPDATE_NO_LOCK interrupt on DCN, which seems to correspond to
+	 * the regular VUPDATE interrupt on DCE. We want DC_IRQ_SOURCE_VUPDATEx
+	 * to trigger at end of each vblank, regardless of state of the lock,
+	 * matching DCE behaviour.
+	 */
+	for (i = DCN_1_0__SRCID__OTG0_IHC_V_UPDATE_NO_LOCK_INTERRUPT;
+	     i <= DCN_1_0__SRCID__OTG0_IHC_V_UPDATE_NO_LOCK_INTERRUPT + adev->mode_info.num_crtc - 1;
+	     i++) {
+		r = amdgpu_irq_add_id(adev, SOC15_IH_CLIENTID_DCE, i, &adev->vupdate_irq);
+
+		if (r) {
+			DRM_ERROR("Failed to add vupdate irq id!\n");
+			return r;
+		}
+
+		int_params.int_context = INTERRUPT_HIGH_IRQ_CONTEXT;
+		int_params.irq_source =
+			dc_interrupt_to_irq_source(dc, i, 0);
+
+		c_irq_params = &adev->dm.vupdate_params[int_params.irq_source - DC_IRQ_SOURCE_VUPDATE1];
+
+		c_irq_params->adev = adev;
+		c_irq_params->irq_src = int_params.irq_source;
+
+		amdgpu_dm_irq_register_interrupt(adev, &int_params,
+				dm_vupdate_high_irq, c_irq_params);
 	}
 
 	/* Use GRPH_PFLIP interrupt */
···
 
 	__drm_atomic_helper_private_obj_duplicate_state(obj, &new_state->base);
 
-	new_state->context = dc_create_state();
+	old_state = to_dm_atomic_state(obj->state);
+
+	if (old_state && old_state->context)
+		new_state->context = dc_copy_state(old_state->context);
+
 	if (!new_state->context) {
 		kfree(new_state);
 		return NULL;
 	}
-
-	old_state = to_dm_atomic_state(obj->state);
-	if (old_state && old_state->context)
-		dc_resource_state_copy_construct(old_state->context,
-						 new_state->context);
 
 	return &new_state->base;
 }
···
 	if (!state)
 		return -ENOMEM;
 
-	state->context = dc_create_state();
+	state->context = dc_create_state(adev->dm.dc);
 	if (!state->context) {
 		kfree(state);
 		return -ENOMEM;
···
 
 static int initialize_plane(struct amdgpu_display_manager *dm,
 			    struct amdgpu_mode_info *mode_info, int plane_id,
-			    enum drm_plane_type plane_type)
+			    enum drm_plane_type plane_type,
+			    const struct dc_plane_cap *plane_cap)
 {
 	struct drm_plane *plane;
 	unsigned long possible_crtcs;
···
 	if (plane_id >= dm->dc->caps.max_streams)
 		possible_crtcs = 0xff;
 
-	ret = amdgpu_dm_plane_init(dm, plane, possible_crtcs);
+	ret = amdgpu_dm_plane_init(dm, plane, possible_crtcs, plane_cap);
 
 	if (ret) {
 		DRM_ERROR("KMS: Failed to initialize plane\n");
···
 	struct amdgpu_encoder *aencoder = NULL;
 	struct amdgpu_mode_info *mode_info = &adev->mode_info;
 	uint32_t link_cnt;
-	int32_t overlay_planes, primary_planes;
+	int32_t primary_planes;
 	enum dc_connection_type new_connection_type = dc_connection_none;
+	const struct dc_plane_cap *plane;
 
 	link_cnt = dm->dc->caps.max_links;
 	if (amdgpu_dm_mode_config_init(dm->adev)) {
 		DRM_ERROR("DM: Failed to initialize mode config\n");
 		return -EINVAL;
 	}
-
-	/*
-	 * Determine the number of overlay planes supported.
-	 * Only support DCN for now, and cap so we don't encourage
-	 * userspace to use up all the planes.
-	 */
-	overlay_planes = 0;
-
-	for (i = 0; i < dm->dc->caps.max_planes; ++i) {
-		struct dc_plane_cap *plane = &dm->dc->caps.planes[i];
-
-		if (plane->type == DC_PLANE_TYPE_DCN_UNIVERSAL &&
-		    plane->blends_with_above && plane->blends_with_below &&
-		    plane->supports_argb8888)
-			overlay_planes += 1;
-	}
-
-	overlay_planes = min(overlay_planes, 1);
 
 	/* There is one primary plane per CRTC */
 	primary_planes = dm->dc->caps.max_streams;
···
 	 * Order is reversed to match iteration order in atomic check.
 	 */
 	for (i = (primary_planes - 1); i >= 0; i--) {
+		plane = &dm->dc->caps.planes[i];
+
 		if (initialize_plane(dm, mode_info, i,
-				     DRM_PLANE_TYPE_PRIMARY)) {
+				     DRM_PLANE_TYPE_PRIMARY, plane)) {
 			DRM_ERROR("KMS: Failed to initialize primary plane\n");
 			goto fail;
 		}
···
 	 * These planes have a higher DRM index than the primary planes since
 	 * they should be considered as having a higher z-order.
 	 * Order is reversed to match iteration order in atomic check.
+	 *
+	 * Only support DCN for now, and only expose one so we don't encourage
+	 * userspace to use up all the pipes.
 	 */
-	for (i = (overlay_planes - 1); i >= 0; i--) {
+	for (i = 0; i < dm->dc->caps.max_planes; ++i) {
+		struct dc_plane_cap *plane = &dm->dc->caps.planes[i];
+
+		if (plane->type != DC_PLANE_TYPE_DCN_UNIVERSAL)
+			continue;
+
+		if (!plane->blends_with_above || !plane->blends_with_below)
+			continue;
+
+		if (!plane->supports_argb8888)
+			continue;
+
 		if (initialize_plane(dm, NULL, primary_planes + i,
-				     DRM_PLANE_TYPE_OVERLAY)) {
+				     DRM_PLANE_TYPE_OVERLAY, plane)) {
 			DRM_ERROR("KMS: Failed to initialize overlay plane\n");
 			goto fail;
 		}
+
+		/* Only create one overlay plane. */
+		break;
 	}
 
 	for (i = 0; i < dm->dc->caps.max_streams; i++)
···
 	}
 }
 
+static int
+fill_plane_color_attributes(const struct drm_plane_state *plane_state,
+			    const struct dc_plane_state *dc_plane_state,
+			    enum dc_color_space *color_space)
+{
+	bool full_range;
+
+	*color_space = COLOR_SPACE_SRGB;
+
+	/* DRM color properties only affect non-RGB formats. */
+	if (dc_plane_state->format < SURFACE_PIXEL_FORMAT_VIDEO_BEGIN)
+		return 0;
+
+	full_range = (plane_state->color_range == DRM_COLOR_YCBCR_FULL_RANGE);
+
+	switch (plane_state->color_encoding) {
+	case DRM_COLOR_YCBCR_BT601:
+		if (full_range)
+			*color_space = COLOR_SPACE_YCBCR601;
+		else
+			*color_space = COLOR_SPACE_YCBCR601_LIMITED;
+		break;
+
+	case DRM_COLOR_YCBCR_BT709:
+		if (full_range)
+			*color_space = COLOR_SPACE_YCBCR709;
+		else
+			*color_space = COLOR_SPACE_YCBCR709_LIMITED;
+		break;
+
+	case DRM_COLOR_YCBCR_BT2020:
+		if (full_range)
+			*color_space = COLOR_SPACE_2020_YCBCR;
+		else
+			return -EINVAL;
+		break;
+
+	default:
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
 static int fill_plane_attributes(struct amdgpu_device *adev,
 				 struct dc_plane_state *dc_plane_state,
 				 struct drm_plane_state *plane_state,
···
 		dc_plane_state,
 		amdgpu_fb);
 
+	if (ret)
+		return ret;
+
+	ret = fill_plane_color_attributes(plane_state, dc_plane_state,
+					  &dc_plane_state->color_space);
 	if (ret)
 		return ret;
 
···
 	return &state->base;
 }
 
+static inline int dm_set_vupdate_irq(struct drm_crtc *crtc, bool enable)
+{
+	enum dc_irq_source irq_source;
+	struct amdgpu_crtc *acrtc = to_amdgpu_crtc(crtc);
+	struct amdgpu_device *adev = crtc->dev->dev_private;
+	int rc;
+
+	irq_source = IRQ_TYPE_VUPDATE + acrtc->otg_inst;
+
+	rc = dc_interrupt_set(adev->dm.dc, irq_source, enable) ? 0 : -EBUSY;
+
+	DRM_DEBUG_DRIVER("crtc %d - vupdate irq %sabling: r=%d\n",
+			 acrtc->crtc_id, enable ? "en" : "dis", rc);
+	return rc;
+}
 
 static inline int dm_set_vblank(struct drm_crtc *crtc, bool enable)
 {
 	enum dc_irq_source irq_source;
 	struct amdgpu_crtc *acrtc = to_amdgpu_crtc(crtc);
 	struct amdgpu_device *adev = crtc->dev->dev_private;
+	struct dm_crtc_state *acrtc_state = to_dm_crtc_state(crtc->state);
+	int rc = 0;
+
+	if (enable) {
+		/* vblank irq on -> Only need vupdate irq in vrr mode */
+		if (amdgpu_dm_vrr_active(acrtc_state))
+			rc = dm_set_vupdate_irq(crtc, true);
+	} else {
+		/* vblank irq off -> vupdate irq off */
+		rc = dm_set_vupdate_irq(crtc, false);
+	}
+
+	if (rc)
+		return rc;
 
 	irq_source = IRQ_TYPE_VBLANK + acrtc->otg_inst;
 	return dc_interrupt_set(adev->dm.dc, irq_source, enable) ? 0 : -EBUSY;
···
 
 static int amdgpu_dm_plane_init(struct amdgpu_display_manager *dm,
 				struct drm_plane *plane,
-				unsigned long possible_crtcs)
+				unsigned long possible_crtcs,
+				const struct dc_plane_cap *plane_cap)
 {
 	int res = -EPERM;
 
···
 		break;
 	}
 
-	/* TODO: Check DC plane caps explicitly here for adding propertes */
-	if (plane->type == DRM_PLANE_TYPE_OVERLAY) {
+	if (plane->type == DRM_PLANE_TYPE_OVERLAY &&
+	    plane_cap && plane_cap->per_pixel_alpha) {
 		unsigned int blend_caps = BIT(DRM_MODE_BLEND_PIXEL_NONE) |
 					  BIT(DRM_MODE_BLEND_PREMULTI);
 
···
 		goto fail;
 
 	cursor_plane->type = DRM_PLANE_TYPE_CURSOR;
-	res = amdgpu_dm_plane_init(dm, cursor_plane, 0);
+	res = amdgpu_dm_plane_init(dm, cursor_plane, 0, NULL);
 
 	acrtc = kzalloc(sizeof(struct amdgpu_crtc), GFP_KERNEL);
 	if (!acrtc)
···
 {
 	struct mod_vrr_params vrr_params = new_crtc_state->vrr_params;
 	struct dc_info_packet vrr_infopacket = {0};
-	struct mod_freesync_config config = new_crtc_state->freesync_config;
 
 	if (!new_stream)
 		return;
···
 
 	if (!new_stream->timing.h_total || !new_stream->timing.v_total)
 		return;
-
-	if (new_crtc_state->vrr_supported &&
-	    config.min_refresh_in_uhz &&
-	    config.max_refresh_in_uhz) {
-		config.state = new_crtc_state->base.vrr_enabled ?
-			VRR_STATE_ACTIVE_VARIABLE :
-			VRR_STATE_INACTIVE;
-	} else {
-		config.state = VRR_STATE_UNSUPPORTED;
-	}
-
-	mod_freesync_build_vrr_params(dm->freesync_module,
-				      new_stream,
-				      &config, &vrr_params);
 
 	if (surface) {
 		mod_freesync_handle_preflip(
···
 			      (int)vrr_params.state);
 }
 
+static void pre_update_freesync_state_on_stream(
+	struct amdgpu_display_manager *dm,
+	struct dm_crtc_state *new_crtc_state)
+{
+	struct dc_stream_state *new_stream = new_crtc_state->stream;
+	struct mod_vrr_params vrr_params = new_crtc_state->vrr_params;
+	struct mod_freesync_config config = new_crtc_state->freesync_config;
+
+	if (!new_stream)
+		return;
+
+	/*
+	 * TODO: Determine why min/max totals and vrefresh can be 0 here.
+	 * For now it's sufficient to just guard against these conditions.
+	 */
+	if (!new_stream->timing.h_total || !new_stream->timing.v_total)
+		return;
+
+	if (new_crtc_state->vrr_supported &&
+	    config.min_refresh_in_uhz &&
+	    config.max_refresh_in_uhz) {
+		config.state = new_crtc_state->base.vrr_enabled ?
+			VRR_STATE_ACTIVE_VARIABLE :
+			VRR_STATE_INACTIVE;
+	} else {
+		config.state = VRR_STATE_UNSUPPORTED;
+	}
+
+	mod_freesync_build_vrr_params(dm->freesync_module,
+				      new_stream,
+				      &config, &vrr_params);
+
+	new_crtc_state->freesync_timing_changed |=
+		(memcmp(&new_crtc_state->vrr_params.adjust,
+			&vrr_params.adjust,
+			sizeof(vrr_params.adjust)) != 0);
+
+	new_crtc_state->vrr_params = vrr_params;
+}
+
+static void amdgpu_dm_handle_vrr_transition(struct dm_crtc_state *old_state,
+					    struct dm_crtc_state *new_state)
+{
+	bool old_vrr_active = amdgpu_dm_vrr_active(old_state);
+	bool new_vrr_active = amdgpu_dm_vrr_active(new_state);
+
+	if (!old_vrr_active && new_vrr_active) {
+		/* Transition VRR inactive -> active:
+		 * While VRR is active, we must not disable vblank irq, as a
+		 * reenable after disable would compute bogus vblank/pflip
+		 * timestamps if it likely happened inside display front-porch.
+		 *
+		 * We also need vupdate irq for the actual core vblank handling
+		 * at end of vblank.
+		 */
+		dm_set_vupdate_irq(new_state->base.crtc, true);
+		drm_crtc_vblank_get(new_state->base.crtc);
+		DRM_DEBUG_DRIVER("%s: crtc=%u VRR off->on: Get vblank ref\n",
+				 __func__, new_state->base.crtc->base.id);
+	} else if (old_vrr_active && !new_vrr_active) {
+		/* Transition VRR active -> inactive:
+		 * Allow vblank irq disable again for fixed refresh rate.
+		 */
+		dm_set_vupdate_irq(new_state->base.crtc, false);
+		drm_crtc_vblank_put(new_state->base.crtc);
+		DRM_DEBUG_DRIVER("%s: crtc=%u VRR on->off: Drop vblank ref\n",
+				 __func__, new_state->base.crtc->base.id);
+	}
+}
+
 static void amdgpu_dm_commit_planes(struct drm_atomic_state *state,
 				    struct dc_state *dc_state,
 				    struct drm_device *dev,
···
 	unsigned long flags;
 	struct amdgpu_bo *abo;
 	uint64_t tiling_flags;
-	uint32_t target, target_vblank;
-	uint64_t last_flip_vblank;
-	bool vrr_active = acrtc_state->freesync_config.state == VRR_STATE_ACTIVE_VARIABLE;
+	uint32_t target_vblank, last_flip_vblank;
+	bool vrr_active = amdgpu_dm_vrr_active(acrtc_state);
 	bool pflip_present = false;
-
 	struct {
 		struct dc_surface_update surface_updates[MAX_SURFACES];
 		struct dc_plane_info plane_infos[MAX_SURFACES];
···
 		bundle->scaling_infos[planes_count].clip_rect = dc_plane->clip_rect;
 		bundle->surface_updates[planes_count].scaling_info = &bundle->scaling_infos[planes_count];
 
+		fill_plane_color_attributes(
+			new_plane_state, dc_plane,
+			&bundle->plane_infos[planes_count].color_space);
 
-		bundle->plane_infos[planes_count].color_space = dc_plane->color_space;
 		bundle->plane_infos[planes_count].format = dc_plane->format;
 		bundle->plane_infos[planes_count].plane_size = dc_plane->plane_size;
 		bundle->plane_infos[planes_count].rotation = dc_plane->rotation;
···
 			 * clients using the GLX_OML_sync_control extension or
 			 * DRI3/Present extension with defined target_msc.
 			 */
-			last_flip_vblank = drm_crtc_vblank_count(pcrtc);
+			last_flip_vblank = amdgpu_get_vblank_counter_kms(dm->ddev, acrtc_attach->crtc_id);
 		}
 		else {
 			/* For variable refresh rate mode only:
···
 			spin_unlock_irqrestore(&pcrtc->dev->event_lock, flags);
 		}
 
-		target = (uint32_t)last_flip_vblank + wait_for_vblank;
-
-		/* Prepare wait for target vblank early - before the fence-waits */
-		target_vblank = target - (uint32_t)drm_crtc_vblank_count(pcrtc) +
-				amdgpu_get_vblank_counter_kms(pcrtc->dev, acrtc_attach->crtc_id);
+		target_vblank = last_flip_vblank + wait_for_vblank;
 
 		/*
 		 * Wait until we're out of the vertical blank period before the one
···
 		dc_state = dm_state->context;
 	} else {
 		/* No state changes, retain current state. */
-		dc_state_temp = dc_create_state();
+		dc_state_temp = dc_create_state(dm->dc);
 		ASSERT(dc_state_temp);
 		dc_state = dc_state_temp;
 		dc_resource_state_copy_construct_current(dm->dc, dc_state);
···
 		mutex_unlock(&dm->dc_lock);
 	}
 
+	/* Update freesync state before amdgpu_dm_handle_vrr_transition(). */
+	for_each_new_crtc_in_state(state, crtc, new_crtc_state, i) {
+		dm_new_crtc_state = to_dm_crtc_state(new_crtc_state);
+		pre_update_freesync_state_on_stream(dm, dm_new_crtc_state);
+	}
+
 	for_each_oldnew_crtc_in_state(state, crtc, old_crtc_state,
 			new_crtc_state, i) {
 		/*
···
 
 		dm_new_crtc_state = to_dm_crtc_state(new_crtc_state);
 		dm_old_crtc_state = to_dm_crtc_state(old_crtc_state);
+
+		/* Handle vrr on->off / off->on transitions */
+		amdgpu_dm_handle_vrr_transition(dm_old_crtc_state,
+						dm_new_crtc_state);
+
 		modeset_needed = modeset_required(
 				new_crtc_state,
 				dm_new_crtc_state->stream,
···
 	}
 
 	for_each_oldnew_crtc_in_state(state, crtc, old_crtc_state, new_crtc_state, i) {
-		struct dc_stream_update stream_update = { 0 };
+		struct dc_stream_update stream_update;
+
+		memset(&stream_update, 0, sizeof(stream_update));
 
 		new_dm_crtc_state = to_dm_crtc_state(new_crtc_state);
 		old_dm_crtc_state = to_dm_crtc_state(old_crtc_state);

drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.h

···
 	struct common_irq_params
 	vblank_params[DC_IRQ_SOURCE_VBLANK6 - DC_IRQ_SOURCE_VBLANK1 + 1];
 
+	/**
+	 * @vupdate_params:
+	 *
+	 * Vertical update IRQ parameters, passed to registered handlers when
+	 * triggered.
+	 */
+	struct common_irq_params
+	vupdate_params[DC_IRQ_SOURCE_VUPDATE6 - DC_IRQ_SOURCE_VUPDATE1 + 1];
+
 	spinlock_t irq_handler_list_table_lock;
 
 	struct backlight_device *backlight_dev;

drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_debugfs.c

···
 	{"amdgpu_target_backlight_pwm", &target_backlight_read},
 };
 
+/*
+ * Sets the DC visual confirm debug option from the given string.
+ * Example usage: echo 1 > /sys/kernel/debug/dri/0/amdgpu_visual_confirm
+ */
+static int visual_confirm_set(void *data, u64 val)
+{
+	struct amdgpu_device *adev = data;
+
+	adev->dm.dc->debug.visual_confirm = (enum visual_confirm)val;
+
+	return 0;
+}
+
+/*
+ * Reads the DC visual confirm debug option value into the given buffer.
+ * Example usage: cat /sys/kernel/debug/dri/0/amdgpu_dm_visual_confirm
+ */
+static int visual_confirm_get(void *data, u64 *val)
+{
+	struct amdgpu_device *adev = data;
+
+	*val = adev->dm.dc->debug.visual_confirm;
+
+	return 0;
+}
+
+DEFINE_DEBUGFS_ATTRIBUTE(visual_confirm_fops, visual_confirm_get,
+			 visual_confirm_set, "%llu\n");
+
 int dtn_debugfs_init(struct amdgpu_device *adev)
 {
 	static const struct file_operations dtn_log_fops = {
···
 		adev,
 		&dtn_log_fops);
 
-	return PTR_ERR_OR_ZERO(ent);
+	if (IS_ERR(ent))
+		return PTR_ERR(ent);
+
+	ent = debugfs_create_file_unsafe("amdgpu_dm_visual_confirm", 0644, root,
+					 adev, &visual_confirm_fops);
+	if (IS_ERR(ent))
+		return PTR_ERR(ent);
+
+	return 0;
 }

drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_irq.c

···
 		__func__);
 }
 
+static int amdgpu_dm_set_vupdate_irq_state(struct amdgpu_device *adev,
+					   struct amdgpu_irq_src *source,
+					   unsigned int crtc_id,
+					   enum amdgpu_interrupt_state state)
+{
+	return dm_irq_state(
+		adev,
+		source,
+		crtc_id,
+		state,
+		IRQ_TYPE_VUPDATE,
+		__func__);
+}
+
 static const struct amdgpu_irq_src_funcs dm_crtc_irq_funcs = {
 	.set = amdgpu_dm_set_crtc_irq_state,
+	.process = amdgpu_dm_irq_handler,
+};
+
+static const struct amdgpu_irq_src_funcs dm_vupdate_irq_funcs = {
+	.set = amdgpu_dm_set_vupdate_irq_state,
 	.process = amdgpu_dm_irq_handler,
 };
 
···
 
 	adev->crtc_irq.num_types = adev->mode_info.num_crtc;
 	adev->crtc_irq.funcs = &dm_crtc_irq_funcs;
+
+	adev->vupdate_irq.num_types = adev->mode_info.num_crtc;
+	adev->vupdate_irq.funcs = &dm_vupdate_irq_funcs;
 
 	adev->pageflip_irq.num_types = adev->mode_info.num_crtc;
 	adev->pageflip_irq.funcs = &dm_pageflip_irq_funcs;

drivers/gpu/drm/amd/display/dc/calcs/dcn_calcs.c

···
 		v->fabric_and_dram_bandwidth = v->fabric_and_dram_bandwidth_vnom0p8;
 		dispclkdppclkdcfclk_deep_sleep_prefetch_parameters_watermarks_and_performance_calculation(v);
 
-		context->bw.dcn.watermarks.b.cstate_pstate.cstate_exit_ns =
+		context->bw_ctx.bw.dcn.watermarks.b.cstate_pstate.cstate_exit_ns =
 			v->stutter_exit_watermark * 1000;
-		context->bw.dcn.watermarks.b.cstate_pstate.cstate_enter_plus_exit_ns =
+		context->bw_ctx.bw.dcn.watermarks.b.cstate_pstate.cstate_enter_plus_exit_ns =
 				v->stutter_enter_plus_exit_watermark * 1000;
-		context->bw.dcn.watermarks.b.cstate_pstate.pstate_change_ns =
+		context->bw_ctx.bw.dcn.watermarks.b.cstate_pstate.pstate_change_ns =
 				v->dram_clock_change_watermark * 1000;
-		context->bw.dcn.watermarks.b.pte_meta_urgent_ns = v->ptemeta_urgent_watermark * 1000;
-		context->bw.dcn.watermarks.b.urgent_ns = v->urgent_watermark * 1000;
+		context->bw_ctx.bw.dcn.watermarks.b.pte_meta_urgent_ns = v->ptemeta_urgent_watermark * 1000;
+		context->bw_ctx.bw.dcn.watermarks.b.urgent_ns = v->urgent_watermark * 1000;
 
 		v->dcfclk_per_state[1] = v->dcfclkv_nom0p8;
 		v->dcfclk_per_state[0] = v->dcfclkv_nom0p8;
 		v->dcfclk = v->dcfclkv_nom0p8;
 		dispclkdppclkdcfclk_deep_sleep_prefetch_parameters_watermarks_and_performance_calculation(v);
 
-		context->bw.dcn.watermarks.c.cstate_pstate.cstate_exit_ns =
+		context->bw_ctx.bw.dcn.watermarks.c.cstate_pstate.cstate_exit_ns =
 			v->stutter_exit_watermark * 1000;
-		context->bw.dcn.watermarks.c.cstate_pstate.cstate_enter_plus_exit_ns =
+		context->bw_ctx.bw.dcn.watermarks.c.cstate_pstate.cstate_enter_plus_exit_ns =
 				v->stutter_enter_plus_exit_watermark * 1000;
-		context->bw.dcn.watermarks.c.cstate_pstate.pstate_change_ns =
+		context->bw_ctx.bw.dcn.watermarks.c.cstate_pstate.pstate_change_ns =
 				v->dram_clock_change_watermark * 1000;
-		context->bw.dcn.watermarks.c.pte_meta_urgent_ns = v->ptemeta_urgent_watermark * 1000;
-		context->bw.dcn.watermarks.c.urgent_ns = v->urgent_watermark * 1000;
+		context->bw_ctx.bw.dcn.watermarks.c.pte_meta_urgent_ns = v->ptemeta_urgent_watermark * 1000;
+		context->bw_ctx.bw.dcn.watermarks.c.urgent_ns = v->urgent_watermark * 1000;
 	}
 
 	if (v->voltage_level < 3) {
···
 		v->dcfclk = v->dcfclkv_max0p9;
 		dispclkdppclkdcfclk_deep_sleep_prefetch_parameters_watermarks_and_performance_calculation(v);
 
-		context->bw.dcn.watermarks.d.cstate_pstate.cstate_exit_ns =
+		context->bw_ctx.bw.dcn.watermarks.d.cstate_pstate.cstate_exit_ns =
 			v->stutter_exit_watermark * 1000;
-		context->bw.dcn.watermarks.d.cstate_pstate.cstate_enter_plus_exit_ns =
+		context->bw_ctx.bw.dcn.watermarks.d.cstate_pstate.cstate_enter_plus_exit_ns =
 				v->stutter_enter_plus_exit_watermark * 1000;
-		context->bw.dcn.watermarks.d.cstate_pstate.pstate_change_ns =
+		context->bw_ctx.bw.dcn.watermarks.d.cstate_pstate.pstate_change_ns =
 				v->dram_clock_change_watermark * 1000;
-		context->bw.dcn.watermarks.d.pte_meta_urgent_ns = v->ptemeta_urgent_watermark * 1000;
-		context->bw.dcn.watermarks.d.urgent_ns = v->urgent_watermark * 1000;
+		context->bw_ctx.bw.dcn.watermarks.d.pte_meta_urgent_ns = v->ptemeta_urgent_watermark * 1000;
+		context->bw_ctx.bw.dcn.watermarks.d.urgent_ns = v->urgent_watermark * 1000;
 	}
 
 	v->fabric_and_dram_bandwidth_per_state[2] = v->fabric_and_dram_bandwidth_vnom0p8;
···
 	v->dcfclk = v->dcfclk_per_state[v->voltage_level];
 	dispclkdppclkdcfclk_deep_sleep_prefetch_parameters_watermarks_and_performance_calculation(v);
 
-	context->bw.dcn.watermarks.a.cstate_pstate.cstate_exit_ns =
+	context->bw_ctx.bw.dcn.watermarks.a.cstate_pstate.cstate_exit_ns =
 		v->stutter_exit_watermark * 1000;
-	context->bw.dcn.watermarks.a.cstate_pstate.cstate_enter_plus_exit_ns =
+	context->bw_ctx.bw.dcn.watermarks.a.cstate_pstate.cstate_enter_plus_exit_ns =
 			v->stutter_enter_plus_exit_watermark * 1000;
-	context->bw.dcn.watermarks.a.cstate_pstate.pstate_change_ns =
+	context->bw_ctx.bw.dcn.watermarks.a.cstate_pstate.pstate_change_ns =
 			v->dram_clock_change_watermark * 1000;
-	context->bw.dcn.watermarks.a.pte_meta_urgent_ns = v->ptemeta_urgent_watermark * 1000;
-	context->bw.dcn.watermarks.a.urgent_ns = v->urgent_watermark * 1000;
+	context->bw_ctx.bw.dcn.watermarks.a.pte_meta_urgent_ns = v->ptemeta_urgent_watermark * 1000;
+	context->bw_ctx.bw.dcn.watermarks.a.urgent_ns = v->urgent_watermark * 1000;
 	if (v->voltage_level >= 2) {
-		context->bw.dcn.watermarks.b = context->bw.dcn.watermarks.a;
-		context->bw.dcn.watermarks.c = context->bw.dcn.watermarks.a;
+		context->bw_ctx.bw.dcn.watermarks.b = context->bw_ctx.bw.dcn.watermarks.a;
+		context->bw_ctx.bw.dcn.watermarks.c = context->bw_ctx.bw.dcn.watermarks.a;
 	}
 	if (v->voltage_level >= 3)
-		context->bw.dcn.watermarks.d = context->bw.dcn.watermarks.a;
+		context->bw_ctx.bw.dcn.watermarks.d = context->bw_ctx.bw.dcn.watermarks.a;
 }
 #endif
 
···
 				dc->debug.sr_enter_plus_exit_time_dpm0_ns / 1000.0f;
 		if (dc->debug.sr_exit_time_dpm0_ns)
 			v->sr_exit_time =  dc->debug.sr_exit_time_dpm0_ns / 1000.0f;
-		dc->dml.soc.sr_enter_plus_exit_time_us = v->sr_enter_plus_exit_time;
-		dc->dml.soc.sr_exit_time_us = v->sr_exit_time;
+		context->bw_ctx.dml.soc.sr_enter_plus_exit_time_us = v->sr_enter_plus_exit_time;
+		context->bw_ctx.dml.soc.sr_exit_time_us = v->sr_exit_time;
 		mode_support_and_system_configuration(v);
 	}
 
···
 		 */
 		dispclkdppclkdcfclk_deep_sleep_prefetch_parameters_watermarks_and_performance_calculation(v);
 
-		context->bw.dcn.watermarks.a.cstate_pstate.cstate_exit_ns =
+		context->bw_ctx.bw.dcn.watermarks.a.cstate_pstate.cstate_exit_ns =
 			v->stutter_exit_watermark * 1000;
-		context->bw.dcn.watermarks.a.cstate_pstate.cstate_enter_plus_exit_ns =
+		context->bw_ctx.bw.dcn.watermarks.a.cstate_pstate.cstate_enter_plus_exit_ns =
 				v->stutter_enter_plus_exit_watermark * 1000;
-		context->bw.dcn.watermarks.a.cstate_pstate.pstate_change_ns =
+		context->bw_ctx.bw.dcn.watermarks.a.cstate_pstate.pstate_change_ns =
 				v->dram_clock_change_watermark * 1000;
-		context->bw.dcn.watermarks.a.pte_meta_urgent_ns = v->ptemeta_urgent_watermark * 1000;
-		context->bw.dcn.watermarks.a.urgent_ns = v->urgent_watermark * 1000;
-		context->bw.dcn.watermarks.b = context->bw.dcn.watermarks.a;
-		context->bw.dcn.watermarks.c = context->bw.dcn.watermarks.a;
-		context->bw.dcn.watermarks.d = context->bw.dcn.watermarks.a;
+		context->bw_ctx.bw.dcn.watermarks.a.pte_meta_urgent_ns = v->ptemeta_urgent_watermark * 1000;
+		context->bw_ctx.bw.dcn.watermarks.a.urgent_ns = v->urgent_watermark * 1000;
+		context->bw_ctx.bw.dcn.watermarks.b = context->bw_ctx.bw.dcn.watermarks.a;
+		context->bw_ctx.bw.dcn.watermarks.c = context->bw_ctx.bw.dcn.watermarks.a;
+		context->bw_ctx.bw.dcn.watermarks.d = context->bw_ctx.bw.dcn.watermarks.a;
 
-		context->bw.dcn.clk.fclk_khz = (int)(bw_consumed * 1000000 /
+		context->bw_ctx.bw.dcn.clk.fclk_khz = (int)(bw_consumed * 1000000 /
 				(ddr4_dram_factor_single_Channel * v->number_of_channels));
 		if (bw_consumed == v->fabric_and_dram_bandwidth_vmin0p65) {
-			context->bw.dcn.clk.fclk_khz = (int)(bw_consumed * 1000000 / 32);
+			context->bw_ctx.bw.dcn.clk.fclk_khz = (int)(bw_consumed * 1000000 / 32);
 		}
 
-		context->bw.dcn.clk.dcfclk_deep_sleep_khz = (int)(v->dcf_clk_deep_sleep * 1000);
-		context->bw.dcn.clk.dcfclk_khz = (int)(v->dcfclk * 1000);
+		context->bw_ctx.bw.dcn.clk.dcfclk_deep_sleep_khz = (int)(v->dcf_clk_deep_sleep * 1000);
+		context->bw_ctx.bw.dcn.clk.dcfclk_khz = (int)(v->dcfclk * 1000);
 
-		context->bw.dcn.clk.dispclk_khz = (int)(v->dispclk * 1000);
+		context->bw_ctx.bw.dcn.clk.dispclk_khz = (int)(v->dispclk * 1000);
 		if (dc->debug.max_disp_clk == true)
-			context->bw.dcn.clk.dispclk_khz = (int)(dc->dcn_soc->max_dispclk_vmax0p9 * 1000);
+			context->bw_ctx.bw.dcn.clk.dispclk_khz = (int)(dc->dcn_soc->max_dispclk_vmax0p9 * 1000);
 
-		if (context->bw.dcn.clk.dispclk_khz <
+		if (context->bw_ctx.bw.dcn.clk.dispclk_khz <
 				dc->debug.min_disp_clk_khz) {
-			context->bw.dcn.clk.dispclk_khz =
+			context->bw_ctx.bw.dcn.clk.dispclk_khz =
 					dc->debug.min_disp_clk_khz;
 		}
 
-		context->bw.dcn.clk.dppclk_khz = context->bw.dcn.clk.dispclk_khz / v->dispclk_dppclk_ratio;
-		context->bw.dcn.clk.phyclk_khz = v->phyclk_per_state[v->voltage_level];
+		context->bw_ctx.bw.dcn.clk.dppclk_khz = context->bw_ctx.bw.dcn.clk.dispclk_khz / v->dispclk_dppclk_ratio;
+		context->bw_ctx.bw.dcn.clk.phyclk_khz = v->phyclk_per_state[v->voltage_level];
 		switch (v->voltage_level) {
 		case 0:
-			context->bw.dcn.clk.max_supported_dppclk_khz =
+			context->bw_ctx.bw.dcn.clk.max_supported_dppclk_khz =
 					(int)(dc->dcn_soc->max_dppclk_vmin0p65 * 1000);
 			break;
 		case 1:
-			context->bw.dcn.clk.max_supported_dppclk_khz =
+			context->bw_ctx.bw.dcn.clk.max_supported_dppclk_khz =
 					(int)(dc->dcn_soc->max_dppclk_vmid0p72 * 1000);
 			break;
 		case 2:
-			context->bw.dcn.clk.max_supported_dppclk_khz =
+			context->bw_ctx.bw.dcn.clk.max_supported_dppclk_khz =
 					(int)(dc->dcn_soc->max_dppclk_vnom0p8 * 1000);
 			break;
 		default:
-			context->bw.dcn.clk.max_supported_dppclk_khz =
+			context->bw_ctx.bw.dcn.clk.max_supported_dppclk_khz =
 					(int)(dc->dcn_soc->max_dppclk_vmax0p9 * 1000);
 			break;
 		}
···
 
 	if (v->voltage_level == 0) {
 
-		dc->dml.soc.sr_enter_plus_exit_time_us =
+		context->bw_ctx.dml.soc.sr_enter_plus_exit_time_us =
 				dc->dcn_soc->sr_enter_plus_exit_time;
-		dc->dml.soc.sr_exit_time_us = dc->dcn_soc->sr_exit_time;
+		context->bw_ctx.dml.soc.sr_exit_time_us = dc->dcn_soc->sr_exit_time;
 	}
 
 	/*

drivers/gpu/drm/amd/display/dc/core/dc.c

···
 			cust_pattern_size);
 }
 
+uint32_t dc_link_bandwidth_kbps(
+	const struct dc_link *link,
+	const struct dc_link_settings *link_setting)
+{
+	uint32_t link_bw_kbps = link_setting->link_rate * LINK_RATE_REF_FREQ_IN_KHZ; /* bytes per sec */
+
+	link_bw_kbps *= 8;   /* 8 bits per byte*/
+	link_bw_kbps *= link_setting->lane_count;
+
+	return link_bw_kbps;
+
+}
+
+const struct dc_link_settings *dc_link_get_verified_link_cap(
+		const struct dc_link *link)
+{
+	if (link->preferred_link_setting.lane_count != LANE_COUNT_UNKNOWN &&
+			link->preferred_link_setting.link_rate != LINK_RATE_UNKNOWN)
+		return &link->preferred_link_setting;
+	return &link->verified_link_cap;
+}
+
 static void destruct(struct dc *dc)
 {
 	dc_release_state(dc->current_state);
···
 #endif
 
 	enum dce_version dc_version = DCE_VERSION_UNKNOWN;
+	dc->config = init_params->flags;
+
 	memcpy(&dc->bb_overrides, &init_params->bb_overrides, sizeof(dc->bb_overrides));
 
 	dc_dceip = kzalloc(sizeof(*dc_dceip), GFP_KERNEL);
···
 	dc_ctx->dc_sink_id_count = 0;
 	dc_ctx->dc_stream_id_count = 0;
 	dc->ctx = dc_ctx;
-
-	dc->current_state = dc_create_state();
-
-	if (!dc->current_state) {
-		dm_error("%s: failed to create validate ctx\n", __func__);
-		goto fail;
-	}
 
 	/* Create logger */
 
···
 	if (!dc->res_pool)
 		goto fail;
 
+	/* Creation of current_state must occur after dc->dml
+	 * is initialized in dc_create_resource_pool because
+	 * on creation it copies the contents of dc->dml
+	 */
+
+	dc->current_state = dc_create_state(dc);
+
+	if (!dc->current_state) {
+		dm_error("%s: failed to create validate ctx\n", __func__);
+		goto fail;
+	}
+
 	dc_resource_state_construct(dc, dc->current_state);
 
 	if (!create_links(dc, init_params->num_virtual_links))
···
 static void disable_dangling_plane(struct dc *dc, struct dc_state *context)
 {
 	int i, j;
-	struct dc_state *dangling_context = dc_create_state();
+	struct dc_state *dangling_context = dc_create_state(dc);
 	struct dc_state *current_ctx;
 
 	if (dangling_context == NULL)
···
 
 	if (dc->res_pool->dmcu != NULL)
 		dc->versions.dmcu_version = dc->res_pool->dmcu->dmcu_version;
-
-	dc->config = init_params->flags;
 
 	dc->build_id = DC_BUILD_ID;
 
···
 	return true;
 }
 
-struct dc_state *dc_create_state(void)
+struct dc_state *dc_create_state(struct dc *dc)
 {
 	struct dc_state *context = kzalloc(sizeof(struct dc_state),
 					   GFP_KERNEL);
 
 	if (!context)
 		return NULL;
+	/* Each context must have their own instance of VBA and in order to
+	 * initialize and obtain IP and SOC the base DML instance from DC is
+	 * initially copied into every context
+	 */
+#ifdef CONFIG_DRM_AMD_DC_DCN1_0
+	memcpy(&context->bw_ctx.dml, &dc->dml, sizeof(struct display_mode_lib));
+#endif
 
 	kref_init(&context->refcount);
+
 	return context;
+}
+
+struct dc_state *dc_copy_state(struct dc_state *src_ctx)
+{
+	int i, j;
+	struct dc_state *new_ctx = kzalloc(sizeof(struct dc_state),
+					   GFP_KERNEL);
+
+	if (!new_ctx)
+		return NULL;
+
+	memcpy(new_ctx, src_ctx, sizeof(struct dc_state));
+
+	for (i = 0; i < MAX_PIPES; i++) {
+			struct pipe_ctx *cur_pipe = &new_ctx->res_ctx.pipe_ctx[i];
+
+			if (cur_pipe->top_pipe)
+				cur_pipe->top_pipe =  &new_ctx->res_ctx.pipe_ctx[cur_pipe->top_pipe->pipe_idx];
+
+			if (cur_pipe->bottom_pipe)
+				cur_pipe->bottom_pipe = &new_ctx->res_ctx.pipe_ctx[cur_pipe->bottom_pipe->pipe_idx];
+
+	}
+
+	for (i = 0; i < new_ctx->stream_count; i++) {
+			dc_stream_retain(new_ctx->streams[i]);
+			for (j = 0; j < new_ctx->stream_status[i].plane_count; j++)
+				dc_plane_state_retain(
+					new_ctx->stream_status[i].plane_states[j]);
+	}
+
+	kref_init(&new_ctx->refcount);
+
+	return new_ctx;
 }
 
 void dc_retain_state(struct dc_state *context)
···
 	if (update_type >= UPDATE_TYPE_FULL) {
 
 		/* initialize scratch memory for building context */
-		context = dc_create_state();
+		context = dc_create_state(dc);
 		if (context == NULL) {
 			DC_ERROR("Failed to allocate new validate context!\n");
 			return;
···
 
 void get_clock_requirements_for_state(struct dc_state *state, struct AsicStateEx *info)
 {
-	info->displayClock				= (unsigned int)state->bw.dcn.clk.dispclk_khz;
-	info->engineClock				= (unsigned int)state->bw.dcn.clk.dcfclk_khz;
-	info->memoryClock				= (unsigned int)state->bw.dcn.clk.dramclk_khz;
-	info->maxSupportedDppClock		= (unsigned int)state->bw.dcn.clk.max_supported_dppclk_khz;
-	info->dppClock					= (unsigned int)state->bw.dcn.clk.dppclk_khz;
-	info->socClock					= (unsigned int)state->bw.dcn.clk.socclk_khz;
-	info->dcfClockDeepSleep			= (unsigned int)state->bw.dcn.clk.dcfclk_deep_sleep_khz;
-	info->fClock					= (unsigned int)state->bw.dcn.clk.fclk_khz;
-	info->phyClock					= (unsigned int)state->bw.dcn.clk.phyclk_khz;
+	info->displayClock				= (unsigned int)state->bw_ctx.bw.dcn.clk.dispclk_khz;
+	info->engineClock				= (unsigned int)state->bw_ctx.bw.dcn.clk.dcfclk_khz;
+	info->memoryClock				= (unsigned int)state->bw_ctx.bw.dcn.clk.dramclk_khz;
+	info->maxSupportedDppClock		= (unsigned int)state->bw_ctx.bw.dcn.clk.max_supported_dppclk_khz;
+	info->dppClock					= (unsigned int)state->bw_ctx.bw.dcn.clk.dppclk_khz;
+	info->socClock					= (unsigned int)state->bw_ctx.bw.dcn.clk.socclk_khz;
+	info->dcfClockDeepSleep			= (unsigned int)state->bw_ctx.bw.dcn.clk.dcfclk_deep_sleep_khz;
+	info->fClock					= (unsigned int)state->bw_ctx.bw.dcn.clk.fclk_khz;
+	info->phyClock					= (unsigned int)state->bw_ctx.bw.dcn.clk.phyclk_khz;
 }

drivers/gpu/drm/amd/display/dc/core/dc_debug.c

···
 	DC_LOGGER_INIT(dc->ctx->logger);
 	CLOCK_TRACE("Current: dispclk_khz:%d  max_dppclk_khz:%d  dcfclk_khz:%d\n"
 			"dcfclk_deep_sleep_khz:%d  fclk_khz:%d  socclk_khz:%d\n",
-			context->bw.dcn.clk.dispclk_khz,
-			context->bw.dcn.clk.dppclk_khz,
-			context->bw.dcn.clk.dcfclk_khz,
-			context->bw.dcn.clk.dcfclk_deep_sleep_khz,
-			context->bw.dcn.clk.fclk_khz,
-			context->bw.dcn.clk.socclk_khz);
+			context->bw_ctx.bw.dcn.clk.dispclk_khz,
+			context->bw_ctx.bw.dcn.clk.dppclk_khz,
+			context->bw_ctx.bw.dcn.clk.dcfclk_khz,
+			context->bw_ctx.bw.dcn.clk.dcfclk_deep_sleep_khz,
+			context->bw_ctx.bw.dcn.clk.fclk_khz,
+			context->bw_ctx.bw.dcn.clk.socclk_khz);
 	CLOCK_TRACE("Calculated: dispclk_khz:%d  max_dppclk_khz:%d  dcfclk_khz:%d\n"
 			"dcfclk_deep_sleep_khz:%d  fclk_khz:%d  socclk_khz:%d\n",
-			context->bw.dcn.clk.dispclk_khz,
-			context->bw.dcn.clk.dppclk_khz,
-			context->bw.dcn.clk.dcfclk_khz,
-			context->bw.dcn.clk.dcfclk_deep_sleep_khz,
-			context->bw.dcn.clk.fclk_khz,
-			context->bw.dcn.clk.socclk_khz);
+			context->bw_ctx.bw.dcn.clk.dispclk_khz,
+			context->bw_ctx.bw.dcn.clk.dppclk_khz,
+			context->bw_ctx.bw.dcn.clk.dcfclk_khz,
+			context->bw_ctx.bw.dcn.clk.dcfclk_deep_sleep_khz,
+			context->bw_ctx.bw.dcn.clk.fclk_khz,
+			context->bw_ctx.bw.dcn.clk.socclk_khz);
 #endif
 }

drivers/gpu/drm/amd/display/dc/core/dc_link.c

···
  ******************************************************************************/
 
 enum {
-	LINK_RATE_REF_FREQ_IN_MHZ = 27,
 	PEAK_FACTOR_X1000 = 1006,
 	/*
 	* Some receivers fail to train on first try and are good
···
 	/* get link settings for video mode timing */
 	decide_link_settings(stream, &link_settings);
 
+	/* If link settings are different than current and link already enabled
+	 * then need to disable before programming to new rate.
+	 */
+	if (link->link_status.link_active &&
+		(link->cur_link_settings.lane_count != link_settings.lane_count ||
+		 link->cur_link_settings.link_rate != link_settings.link_rate)) {
+		dp_disable_link_phy(link, pipe_ctx->stream->signal);
+	}
+
 	pipe_ctx->stream_res.pix_clk_params.requested_sym_clk =
 			link_settings.link_rate * LINK_RATE_REF_FREQ_IN_KHZ;
-	state->dccg->funcs->update_clocks(state->dccg, state, false);
+	state->clk_mgr->funcs->update_clocks(state->clk_mgr, state, false);
 
 	dp_enable_link_phy(
 		link,
···
 		return false;
 	}
 
-	if (get_timing_pixel_clock_100hz(timing) > (dongle_caps->dp_hdmi_max_pixel_clk * 10))
+	if (get_timing_pixel_clock_100hz(timing) > (dongle_caps->dp_hdmi_max_pixel_clk_in_khz * 10))
 		return false;
 
 	return true;
···
 
 static struct fixed31_32 get_pbn_per_slot(struct dc_stream_state *stream)
 {
-	struct dc_link_settings *link_settings =
-			&stream->link->cur_link_settings;
-	uint32_t link_rate_in_mbps =
-			link_settings->link_rate * LINK_RATE_REF_FREQ_IN_MHZ;
-	struct fixed31_32 mbps = dc_fixpt_from_int(
-			link_rate_in_mbps * link_settings->lane_count);
+	struct fixed31_32 mbytes_per_sec;
+	uint32_t link_rate_in_mbytes_per_sec = dc_link_bandwidth_kbps(stream->link, &stream->link->cur_link_settings);
+	link_rate_in_mbytes_per_sec /= 8000; /* Kbits to MBytes */
 
-	return dc_fixpt_div_int(mbps, 54);
+	mbytes_per_sec = dc_fixpt_from_int(link_rate_in_mbytes_per_sec);
+
+	return dc_fixpt_div_int(mbytes_per_sec, 54);
 }
 
 static int get_color_depth(enum dc_color_depth color_depth)
···
 	uint32_t denominator;
 
 	bpc = get_color_depth(pipe_ctx->stream_res.pix_clk_params.color_depth);
-	kbps = pipe_ctx->stream_res.pix_clk_params.requested_pix_clk_100hz / 10 * bpc * 3;
+	kbps = dc_bandwidth_in_kbps_from_timing(&pipe_ctx->stream->timing);
 
 	/*
 	 * margin 5300ppm + 300ppm ~ 0.6% as per spec, factor is 1.006
···
 	}
 }
 
+uint32_t dc_bandwidth_in_kbps_from_timing(
+	const struct dc_crtc_timing *timing)
+{
+	uint32_t bits_per_channel = 0;
+	uint32_t kbps;
+
+	switch (timing->display_color_depth) {
+	case COLOR_DEPTH_666:
+		bits_per_channel = 6;
+		break;
+	case COLOR_DEPTH_888:
+		bits_per_channel = 8;
+		break;
+	case COLOR_DEPTH_101010:
+		bits_per_channel = 10;
+		break;
+	case COLOR_DEPTH_121212:
+		bits_per_channel = 12;
+		break;
+	case COLOR_DEPTH_141414:
+		bits_per_channel = 14;
+		break;
+	case COLOR_DEPTH_161616:
+		bits_per_channel = 16;
+		break;
+	default:
+		break;
+	}
+
+	ASSERT(bits_per_channel != 0);
+
+	kbps = timing->pix_clk_100hz / 10;
+	kbps *= bits_per_channel;
+
+	if (timing->flags.Y_ONLY != 1) {
+		/*Only YOnly make reduce bandwidth by 1/3 compares to RGB*/
+		kbps *= 3;
+		if (timing->pixel_encoding == PIXEL_ENCODING_YCBCR420)
+			kbps /= 2;
+		else if (timing->pixel_encoding == PIXEL_ENCODING_YCBCR422)
+			kbps = kbps * 2 / 3;
+	}
+
+	return kbps;
+
+}

drivers/gpu/drm/amd/display/dc/core/dc_link_dp.c

···
 	return true;
 }
 
-static uint32_t bandwidth_in_kbps_from_timing(
-	const struct dc_crtc_timing *timing)
-{
-	uint32_t bits_per_channel = 0;
-	uint32_t kbps;
-
-	switch (timing->display_color_depth) {
-	case COLOR_DEPTH_666:
-		bits_per_channel = 6;
-		break;
-	case COLOR_DEPTH_888:
-		bits_per_channel = 8;
-		break;
-	case COLOR_DEPTH_101010:
-		bits_per_channel = 10;
-		break;
-	case COLOR_DEPTH_121212:
-		bits_per_channel = 12;
-		break;
-	case COLOR_DEPTH_141414:
-		bits_per_channel = 14;
-		break;
-	case COLOR_DEPTH_161616:
-		bits_per_channel = 16;
-		break;
-	default:
-		break;
-	}
-
-	ASSERT(bits_per_channel != 0);
-
-	kbps = timing->pix_clk_100hz / 10;
-	kbps *= bits_per_channel;
-
-	if (timing->flags.Y_ONLY != 1) {
-		/*Only YOnly make reduce bandwidth by 1/3 compares to RGB*/
-		kbps *= 3;
-		if (timing->pixel_encoding == PIXEL_ENCODING_YCBCR420)
-			kbps /= 2;
-		else if (timing->pixel_encoding == PIXEL_ENCODING_YCBCR422)
-			kbps = kbps * 2 / 3;
-	}
-
-	return kbps;
-
-}
-
-static uint32_t bandwidth_in_kbps_from_link_settings(
-	const struct dc_link_settings *link_setting)
-{
-	uint32_t link_rate_in_kbps = link_setting->link_rate *
-		LINK_RATE_REF_FREQ_IN_KHZ;
-
-	uint32_t lane_count  = link_setting->lane_count;
-	uint32_t kbps = link_rate_in_kbps;
-
-	kbps *= lane_count;
-	kbps *= 8;   /* 8 bits per byte*/
-
-	return kbps;
-
-}
-
 bool dp_validate_mode_timing(
 	struct dc_link *link,
 	const struct dc_crtc_timing *timing)
···
 		return true;
 
 	/* We always use verified link settings */
-	link_setting = &link->verified_link_cap;
+	link_setting = dc_link_get_verified_link_cap(link);
 
 	/* TODO: DYNAMIC_VALIDATION needs to be implemented */
 	/*if (flags.DYNAMIC_VALIDATION == 1 &&
···
 		link_setting = &link->verified_link_cap;
 	*/
 
-	req_bw = bandwidth_in_kbps_from_timing(timing);
-	max_bw = bandwidth_in_kbps_from_link_settings(link_setting);
+	req_bw = dc_bandwidth_in_kbps_from_timing(timing);
+	max_bw = dc_link_bandwidth_kbps(link, link_setting);
 
 	if (req_bw <= max_bw) {
 		/* remember the biggest mode here, during
···
 	 */
 	while (current_link_setting.link_rate <=
 			link->verified_link_cap.link_rate) {
-		link_bw = bandwidth_in_kbps_from_link_settings(
+		link_bw = dc_link_bandwidth_kbps(
+				link,
 				&current_link_setting);
 		if (req_bw <= link_bw) {
 			*link_setting = current_link_setting;
···
 	 */
 	while (current_link_setting.link_rate <=
 			link->verified_link_cap.link_rate) {
-		link_bw = bandwidth_in_kbps_from_link_settings(
+		link_bw = dc_link_bandwidth_kbps(
+				link,
 				&current_link_setting);
 		if (req_bw <= link_bw) {
 			*link_setting = current_link_setting;
···
 	struct dc_link *link;
 	uint32_t req_bw;
 
-	req_bw = bandwidth_in_kbps_from_timing(&stream->timing);
+	req_bw = dc_bandwidth_in_kbps_from_timing(&stream->timing);
 
 	link = stream->link;
 
···
 					hdmi_caps = {.raw = det_caps[3] };
 				union dwnstream_port_caps_byte2
 					hdmi_color_caps = {.raw = det_caps[2] };
-				link->dpcd_caps.dongle_caps.dp_hdmi_max_pixel_clk =
-					det_caps[1] * 25000;
+				link->dpcd_caps.dongle_caps.dp_hdmi_max_pixel_clk_in_khz =
+					det_caps[1] * 2500;
 
 				link->dpcd_caps.dongle_caps.is_dp_hdmi_s3d_converter =
 					hdmi_caps.bits.FRAME_SEQ_TO_FRAME_PACK;
···
 					translate_dpcd_max_bpc(
 						hdmi_color_caps.bits.MAX_BITS_PER_COLOR_COMPONENT);
 
-				if (link->dpcd_caps.dongle_caps.dp_hdmi_max_pixel_clk != 0)
+				if (link->dpcd_caps.dongle_caps.dp_hdmi_max_pixel_clk_in_khz != 0)
 					link->dpcd_caps.dongle_caps.extendedCapValid = true;
 			}
 

drivers/gpu/drm/amd/display/dc/core/dc_resource.c

···
 bool dc_res_is_odm_head_pipe(struct pipe_ctx *pipe_ctx)
 {
 	struct pipe_ctx *top_pipe = pipe_ctx->top_pipe;
-	bool result = false;
 
+	if (!top_pipe)
+		return false;
 	if (top_pipe && top_pipe->stream_res.opp == pipe_ctx->stream_res.opp)
 		return false;
 
-	while (top_pipe) {
-		if (!top_pipe->top_pipe && top_pipe->stream_res.opp != pipe_ctx->stream_res.opp)
-			result = true;
-		top_pipe = top_pipe->top_pipe;
-	}
-
-	return result;
+	return true;
 }
 
 bool dc_remove_plane_from_context(
···
 		const struct dc *dc,
 		struct dc_state *dst_ctx)
 {
-	dst_ctx->dccg = dc->res_pool->clk_mgr;
+	dst_ctx->clk_mgr = dc->res_pool->clk_mgr;
 }
 
 /**

drivers/gpu/drm/amd/display/dc/dc.h

···
 #include "inc/hw/dmcu.h"
 #include "dml/display_mode_lib.h"
 
-#define DC_VER "3.2.24"
+#define DC_VER "3.2.25"
 
 #define MAX_SURFACES 3
 #define MAX_PLANES 6
···
 bool dc_commit_state(struct dc *dc, struct dc_state *context);
 
 
-struct dc_state *dc_create_state(void);
+struct dc_state *dc_create_state(struct dc *dc);
+struct dc_state *dc_copy_state(struct dc_state *src_ctx);
 void dc_retain_state(struct dc_state *context);
 void dc_release_state(struct dc_state *context);
 

drivers/gpu/drm/amd/display/dc/dc_link.h

···
 			const struct link_training_settings *p_link_settings,
 			const unsigned char *p_custom_pattern,
 			unsigned int cust_pattern_size);
+uint32_t dc_link_bandwidth_kbps(
+	const struct dc_link *link,
+	const struct dc_link_settings *link_setting);
+
+const struct dc_link_settings *dc_link_get_verified_link_cap(
+		const struct dc_link *link);
 
 bool dc_submit_i2c(
 		struct dc *dc,
 		uint32_t link_index,
 		struct i2c_command *cmd);
 
+uint32_t dc_bandwidth_in_kbps_from_timing(
+	const struct dc_crtc_timing *timing);
 #endif /* DC_LINK_H_ */

drivers/gpu/drm/amd/display/dc/dc_types.h

···
 	bool is_dp_hdmi_ycbcr422_converter;
 	bool is_dp_hdmi_ycbcr420_converter;
 	uint32_t dp_hdmi_max_bpc;
-	uint32_t dp_hdmi_max_pixel_clk;
+	uint32_t dp_hdmi_max_pixel_clk_in_khz;
 };
 /* Scaling format */
 enum scaling_transformation {

drivers/gpu/drm/amd/display/dc/dce/dce_clk_mgr.c

···
 	 * all required clocks
 	 */
 	for (i = clk_mgr_dce->max_clks_state; i >= DM_PP_CLOCKS_STATE_ULTRA_LOW; i--)
-		if (context->bw.dce.dispclk_khz >
+		if (context->bw_ctx.bw.dce.dispclk_khz >
 				clk_mgr_dce->max_clks_by_state[i].display_clk_khz
 			|| max_pix_clk >
 				clk_mgr_dce->max_clks_by_state[i].pixel_clk_khz)
···
 	if (low_req_clk > clk_mgr_dce->max_clks_state) {
 		/* set max clock state for high phyclock, invalid on exceeding display clock */
 		if (clk_mgr_dce->max_clks_by_state[clk_mgr_dce->max_clks_state].display_clk_khz
-				< context->bw.dce.dispclk_khz)
+				< context->bw_ctx.bw.dce.dispclk_khz)
 			low_req_clk = DM_PP_CLOCKS_STATE_INVALID;
 		else
 			low_req_clk = clk_mgr_dce->max_clks_state;
···
 	struct dm_pp_display_configuration *pp_display_cfg = &context->pp_display_cfg;
 
 	pp_display_cfg->all_displays_in_sync =
-		context->bw.dce.all_displays_in_sync;
+		context->bw_ctx.bw.dce.all_displays_in_sync;
 	pp_display_cfg->nb_pstate_switch_disable =
-			context->bw.dce.nbp_state_change_enable == false;
+			context->bw_ctx.bw.dce.nbp_state_change_enable == false;
 	pp_display_cfg->cpu_cc6_disable =
-			context->bw.dce.cpuc_state_change_enable == false;
+			context->bw_ctx.bw.dce.cpuc_state_change_enable == false;
 	pp_display_cfg->cpu_pstate_disable =
-			context->bw.dce.cpup_state_change_enable == false;
+			context->bw_ctx.bw.dce.cpup_state_change_enable == false;
 	pp_display_cfg->cpu_pstate_separation_time =
-			context->bw.dce.blackout_recovery_time_us;
+			context->bw_ctx.bw.dce.blackout_recovery_time_us;
 
-	pp_display_cfg->min_memory_clock_khz = context->bw.dce.yclk_khz
+	pp_display_cfg->min_memory_clock_khz = context->bw_ctx.bw.dce.yclk_khz
 		/ MEMORY_TYPE_MULTIPLIER_CZ;
 
 	pp_display_cfg->min_engine_clock_khz = determine_sclk_from_bounding_box(
 			dc,
-			context->bw.dce.sclk_khz);
+			context->bw_ctx.bw.dce.sclk_khz);
 
 	/*
 	 * As workaround for >4x4K lightup set dcfclock to min_engine_clock value.
···
 			pp_display_cfg->min_engine_clock_khz : 0;
 
 	pp_display_cfg->min_engine_clock_deep_sleep_khz
-			= context->bw.dce.sclk_deep_sleep_khz;
+			= context->bw_ctx.bw.dce.sclk_deep_sleep_khz;
 
 	pp_display_cfg->avail_mclk_switch_time_us =
 						dce110_get_min_vblank_time_us(context);
···
 {
 	struct dce_clk_mgr *clk_mgr_dce = TO_DCE_CLK_MGR(clk_mgr);
 	struct dm_pp_power_level_change_request level_change_req;
-	int patched_disp_clk = context->bw.dce.dispclk_khz;
+	int patched_disp_clk = context->bw_ctx.bw.dce.dispclk_khz;
 
 	/*TODO: W/A for dal3 linux, investigate why this works */
 	if (!clk_mgr_dce->dfs_bypass_active)
···
 {
 	struct dce_clk_mgr *clk_mgr_dce = TO_DCE_CLK_MGR(clk_mgr);
 	struct dm_pp_power_level_change_request level_change_req;
-	int patched_disp_clk = context->bw.dce.dispclk_khz;
+	int patched_disp_clk = context->bw_ctx.bw.dce.dispclk_khz;
 
 	/*TODO: W/A for dal3 linux, investigate why this works */
 	if (!clk_mgr_dce->dfs_bypass_active)
···
 	}
 
 	if (should_set_clock(safe_to_lower, patched_disp_clk, clk_mgr->clks.dispclk_khz)) {
-		context->bw.dce.dispclk_khz = dce_set_clock(clk_mgr, patched_disp_clk);
+		context->bw_ctx.bw.dce.dispclk_khz = dce_set_clock(clk_mgr, patched_disp_clk);
 		clk_mgr->clks.dispclk_khz = patched_disp_clk;
 	}
 	dce11_pplib_apply_display_requirements(clk_mgr->ctx->dc, context);
···
 {
 	struct dce_clk_mgr *clk_mgr_dce = TO_DCE_CLK_MGR(clk_mgr);
 	struct dm_pp_power_level_change_request level_change_req;
-	int patched_disp_clk = context->bw.dce.dispclk_khz;
+	int patched_disp_clk = context->bw_ctx.bw.dce.dispclk_khz;
 
 	/*TODO: W/A for dal3 linux, investigate why this works */
 	if (!clk_mgr_dce->dfs_bypass_active)
···
 	struct dce_clk_mgr *clk_mgr_dce = TO_DCE_CLK_MGR(clk_mgr);
 	struct dm_pp_clock_for_voltage_req clock_voltage_req = {0};
 	int max_pix_clk = get_max_pixel_clock_for_all_paths(context);
-	int patched_disp_clk = context->bw.dce.dispclk_khz;
+	int patched_disp_clk = context->bw_ctx.bw.dce.dispclk_khz;
 
 	/*TODO: W/A for dal3 linux, investigate why this works */
 	if (!clk_mgr_dce->dfs_bypass_active)

drivers/gpu/drm/amd/display/dc/dce100/dce100_resource.c

···
 
 	if (at_least_one_pipe) {
 		/* TODO implement when needed but for now hardcode max value*/
-		context->bw.dce.dispclk_khz = 681000;
-		context->bw.dce.yclk_khz = 250000 * MEMORY_TYPE_MULTIPLIER_CZ;
+		context->bw_ctx.bw.dce.dispclk_khz = 681000;
+		context->bw_ctx.bw.dce.yclk_khz = 250000 * MEMORY_TYPE_MULTIPLIER_CZ;
 	} else {
-		context->bw.dce.dispclk_khz = 0;
-		context->bw.dce.yclk_khz = 0;
+		context->bw_ctx.bw.dce.dispclk_khz = 0;
+		context->bw_ctx.bw.dce.yclk_khz = 0;
 	}
 
 	return true;

drivers/gpu/drm/amd/display/dc/dce110/dce110_hw_sequencer.c

···
 	if (pipe_ctx->stream->signal == SIGNAL_TYPE_DISPLAY_PORT ||
 			pipe_ctx->stream->signal == SIGNAL_TYPE_DISPLAY_PORT_MST) {
 		audio_output->pll_info.dp_dto_source_clock_in_khz =
-				state->dccg->funcs->get_dp_ref_clk_frequency(
-						state->dccg);
+				state->clk_mgr->funcs->get_dp_ref_clk_frequency(
+						state->clk_mgr);
 	}
 
 	audio_output->pll_info.feed_back_divider =
···
 			dc->bw_vbios->blackout_duration, pipe_ctx->stream);
 		pipe_ctx->plane_res.mi->funcs->mem_input_program_display_marks(
 			pipe_ctx->plane_res.mi,
-			context->bw.dce.nbp_state_change_wm_ns[num_pipes],
-			context->bw.dce.stutter_exit_wm_ns[num_pipes],
-			context->bw.dce.stutter_entry_wm_ns[num_pipes],
-			context->bw.dce.urgent_wm_ns[num_pipes],
+			context->bw_ctx.bw.dce.nbp_state_change_wm_ns[num_pipes],
+			context->bw_ctx.bw.dce.stutter_exit_wm_ns[num_pipes],
+			context->bw_ctx.bw.dce.stutter_entry_wm_ns[num_pipes],
+			context->bw_ctx.bw.dce.urgent_wm_ns[num_pipes],
 			total_dest_line_time_ns);
 		if (i == underlay_idx) {
 			num_pipes++;
 			pipe_ctx->plane_res.mi->funcs->mem_input_program_chroma_display_marks(
 				pipe_ctx->plane_res.mi,
-				context->bw.dce.nbp_state_change_wm_ns[num_pipes],
-				context->bw.dce.stutter_exit_wm_ns[num_pipes],
-				context->bw.dce.urgent_wm_ns[num_pipes],
+				context->bw_ctx.bw.dce.nbp_state_change_wm_ns[num_pipes],
+				context->bw_ctx.bw.dce.stutter_exit_wm_ns[num_pipes],
+				context->bw_ctx.bw.dce.urgent_wm_ns[num_pipes],
 				total_dest_line_time_ns);
 		}
 		num_pipes++;

drivers/gpu/drm/amd/display/dc/dce110/dce110_resource.c

···
 			dc->bw_vbios,
 			context->res_ctx.pipe_ctx,
 			dc->res_pool->pipe_count,
-			&context->bw.dce))
+			&context->bw_ctx.bw.dce))
 		result =  true;
 
 	if (!result)
···
 			context->streams[0]->timing.v_addressable,
 			context->streams[0]->timing.pix_clk_100hz / 10);
 
-	if (memcmp(&dc->current_state->bw.dce,
-			&context->bw.dce, sizeof(context->bw.dce))) {
+	if (memcmp(&dc->current_state->bw_ctx.bw.dce,
+			&context->bw_ctx.bw.dce, sizeof(context->bw_ctx.bw.dce))) {
 
 		DC_LOG_BANDWIDTH_CALCS(
 			"%s: finish,\n"
···
 			"sclk: %d sclk_sleep: %d yclk: %d blackout_recovery_time_us: %d\n"
 			,
 			__func__,
-			context->bw.dce.nbp_state_change_wm_ns[0].b_mark,
-			context->bw.dce.nbp_state_change_wm_ns[0].a_mark,
-			context->bw.dce.urgent_wm_ns[0].b_mark,
-			context->bw.dce.urgent_wm_ns[0].a_mark,
-			context->bw.dce.stutter_exit_wm_ns[0].b_mark,
-			context->bw.dce.stutter_exit_wm_ns[0].a_mark,
-			context->bw.dce.nbp_state_change_wm_ns[1].b_mark,
-			context->bw.dce.nbp_state_change_wm_ns[1].a_mark,
-			context->bw.dce.urgent_wm_ns[1].b_mark,
-			context->bw.dce.urgent_wm_ns[1].a_mark,
-			context->bw.dce.stutter_exit_wm_ns[1].b_mark,
-			context->bw.dce.stutter_exit_wm_ns[1].a_mark,
-			context->bw.dce.nbp_state_change_wm_ns[2].b_mark,
-			context->bw.dce.nbp_state_change_wm_ns[2].a_mark,
-			context->bw.dce.urgent_wm_ns[2].b_mark,
-			context->bw.dce.urgent_wm_ns[2].a_mark,
-			context->bw.dce.stutter_exit_wm_ns[2].b_mark,
-			context->bw.dce.stutter_exit_wm_ns[2].a_mark,
-			context->bw.dce.stutter_mode_enable,
-			context->bw.dce.cpuc_state_change_enable,
-			context->bw.dce.cpup_state_change_enable,
-			context->bw.dce.nbp_state_change_enable,
-			context->bw.dce.all_displays_in_sync,
-			context->bw.dce.dispclk_khz,
-			context->bw.dce.sclk_khz,
-			context->bw.dce.sclk_deep_sleep_khz,
-			context->bw.dce.yclk_khz,
-			context->bw.dce.blackout_recovery_time_us);
+			context->bw_ctx.bw.dce.nbp_state_change_wm_ns[0].b_mark,
+			context->bw_ctx.bw.dce.nbp_state_change_wm_ns[0].a_mark,
+			context->bw_ctx.bw.dce.urgent_wm_ns[0].b_mark,
+			context->bw_ctx.bw.dce.urgent_wm_ns[0].a_mark,
+			context->bw_ctx.bw.dce.stutter_exit_wm_ns[0].b_mark,
+			context->bw_ctx.bw.dce.stutter_exit_wm_ns[0].a_mark,
+			context->bw_ctx.bw.dce.nbp_state_change_wm_ns[1].b_mark,
+			context->bw_ctx.bw.dce.nbp_state_change_wm_ns[1].a_mark,
+			context->bw_ctx.bw.dce.urgent_wm_ns[1].b_mark,
+			context->bw_ctx.bw.dce.urgent_wm_ns[1].a_mark,
+			context->bw_ctx.bw.dce.stutter_exit_wm_ns[1].b_mark,
+			context->bw_ctx.bw.dce.stutter_exit_wm_ns[1].a_mark,
+			context->bw_ctx.bw.dce.nbp_state_change_wm_ns[2].b_mark,
+			context->bw_ctx.bw.dce.nbp_state_change_wm_ns[2].a_mark,
+			context->bw_ctx.bw.dce.urgent_wm_ns[2].b_mark,
+			context->bw_ctx.bw.dce.urgent_wm_ns[2].a_mark,
+			context->bw_ctx.bw.dce.stutter_exit_wm_ns[2].b_mark,
+			context->bw_ctx.bw.dce.stutter_exit_wm_ns[2].a_mark,
+			context->bw_ctx.bw.dce.stutter_mode_enable,
+			context->bw_ctx.bw.dce.cpuc_state_change_enable,
+			context->bw_ctx.bw.dce.cpup_state_change_enable,
+			context->bw_ctx.bw.dce.nbp_state_change_enable,
+			context->bw_ctx.bw.dce.all_displays_in_sync,
+			context->bw_ctx.bw.dce.dispclk_khz,
+			context->bw_ctx.bw.dce.sclk_khz,
+			context->bw_ctx.bw.dce.sclk_deep_sleep_khz,
+			context->bw_ctx.bw.dce.yclk_khz,
+			context->bw_ctx.bw.dce.blackout_recovery_time_us);
 	}
 	return result;
 }

drivers/gpu/drm/amd/display/dc/dce112/dce112_resource.c

···
 			dc->bw_vbios,
 			context->res_ctx.pipe_ctx,
 			dc->res_pool->pipe_count,
-			&context->bw.dce))
+			&context->bw_ctx.bw.dce))
 		result = true;
 
 	if (!result)
···
 			"%s: Bandwidth validation failed!",
 			__func__);
 
-	if (memcmp(&dc->current_state->bw.dce,
-			&context->bw.dce, sizeof(context->bw.dce))) {
+	if (memcmp(&dc->current_state->bw_ctx.bw.dce,
+			&context->bw_ctx.bw.dce, sizeof(context->bw_ctx.bw.dce))) {
 
 		DC_LOG_BANDWIDTH_CALCS(
 			"%s: finish,\n"
···
 			"sclk: %d sclk_sleep: %d yclk: %d blackout_recovery_time_us: %d\n"
 			,
 			__func__,
-			context->bw.dce.nbp_state_change_wm_ns[0].b_mark,
-			context->bw.dce.nbp_state_change_wm_ns[0].a_mark,
-			context->bw.dce.urgent_wm_ns[0].b_mark,
-			context->bw.dce.urgent_wm_ns[0].a_mark,
-			context->bw.dce.stutter_exit_wm_ns[0].b_mark,
-			context->bw.dce.stutter_exit_wm_ns[0].a_mark,
-			context->bw.dce.nbp_state_change_wm_ns[1].b_mark,
-			context->bw.dce.nbp_state_change_wm_ns[1].a_mark,
-			context->bw.dce.urgent_wm_ns[1].b_mark,
-			context->bw.dce.urgent_wm_ns[1].a_mark,
-			context->bw.dce.stutter_exit_wm_ns[1].b_mark,
-			context->bw.dce.stutter_exit_wm_ns[1].a_mark,
-			context->bw.dce.nbp_state_change_wm_ns[2].b_mark,
-			context->bw.dce.nbp_state_change_wm_ns[2].a_mark,
-			context->bw.dce.urgent_wm_ns[2].b_mark,
-			context->bw.dce.urgent_wm_ns[2].a_mark,
-			context->bw.dce.stutter_exit_wm_ns[2].b_mark,
-			context->bw.dce.stutter_exit_wm_ns[2].a_mark,
-			context->bw.dce.stutter_mode_enable,
-			context->bw.dce.cpuc_state_change_enable,
-			context->bw.dce.cpup_state_change_enable,
-			context->bw.dce.nbp_state_change_enable,
-			context->bw.dce.all_displays_in_sync,
-			context->bw.dce.dispclk_khz,
-			context->bw.dce.sclk_khz,
-			context->bw.dce.sclk_deep_sleep_khz,
-			context->bw.dce.yclk_khz,
-			context->bw.dce.blackout_recovery_time_us);
+			context->bw_ctx.bw.dce.nbp_state_change_wm_ns[0].b_mark,
+			context->bw_ctx.bw.dce.nbp_state_change_wm_ns[0].a_mark,
+			context->bw_ctx.bw.dce.urgent_wm_ns[0].b_mark,
+			context->bw_ctx.bw.dce.urgent_wm_ns[0].a_mark,
+			context->bw_ctx.bw.dce.stutter_exit_wm_ns[0].b_mark,
+			context->bw_ctx.bw.dce.stutter_exit_wm_ns[0].a_mark,
+			context->bw_ctx.bw.dce.nbp_state_change_wm_ns[1].b_mark,
+			context->bw_ctx.bw.dce.nbp_state_change_wm_ns[1].a_mark,
+			context->bw_ctx.bw.dce.urgent_wm_ns[1].b_mark,
+			context->bw_ctx.bw.dce.urgent_wm_ns[1].a_mark,
+			context->bw_ctx.bw.dce.stutter_exit_wm_ns[1].b_mark,
+			context->bw_ctx.bw.dce.stutter_exit_wm_ns[1].a_mark,
+			context->bw_ctx.bw.dce.nbp_state_change_wm_ns[2].b_mark,
+			context->bw_ctx.bw.dce.nbp_state_change_wm_ns[2].a_mark,
+			context->bw_ctx.bw.dce.urgent_wm_ns[2].b_mark,
+			context->bw_ctx.bw.dce.urgent_wm_ns[2].a_mark,
+			context->bw_ctx.bw.dce.stutter_exit_wm_ns[2].b_mark,
+			context->bw_ctx.bw.dce.stutter_exit_wm_ns[2].a_mark,
+			context->bw_ctx.bw.dce.stutter_mode_enable,
+			context->bw_ctx.bw.dce.cpuc_state_change_enable,
+			context->bw_ctx.bw.dce.cpup_state_change_enable,
+			context->bw_ctx.bw.dce.nbp_state_change_enable,
+			context->bw_ctx.bw.dce.all_displays_in_sync,
+			context->bw_ctx.bw.dce.dispclk_khz,
+			context->bw_ctx.bw.dce.sclk_khz,
+			context->bw_ctx.bw.dce.sclk_deep_sleep_khz,
+			context->bw_ctx.bw.dce.yclk_khz,
+			context->bw_ctx.bw.dce.blackout_recovery_time_us);
 	}
 	return result;
 }

drivers/gpu/drm/amd/display/dc/dce80/dce80_resource.c

···
 
 	if (at_least_one_pipe) {
 		/* TODO implement when needed but for now hardcode max value*/
-		context->bw.dce.dispclk_khz = 681000;
-		context->bw.dce.yclk_khz = 250000 * MEMORY_TYPE_MULTIPLIER_CZ;
+		context->bw_ctx.bw.dce.dispclk_khz = 681000;
+		context->bw_ctx.bw.dce.yclk_khz = 250000 * MEMORY_TYPE_MULTIPLIER_CZ;
 	} else {
-		context->bw.dce.dispclk_khz = 0;
-		context->bw.dce.yclk_khz = 0;
+		context->bw_ctx.bw.dce.dispclk_khz = 0;
+		context->bw_ctx.bw.dce.yclk_khz = 0;
 	}
 
 	return true;

drivers/gpu/drm/amd/display/dc/dcn10/dcn10_clk_mgr.c

···
 {
 	struct dc *dc = clk_mgr->ctx->dc;
 	struct dc_debug_options *debug = &dc->debug;
-	struct dc_clocks *new_clocks = &context->bw.dcn.clk;
+	struct dc_clocks *new_clocks = &context->bw_ctx.bw.dcn.clk;
 	struct pp_smu_display_requirement_rv *smu_req_cur =
 			&dc->res_pool->pp_smu_req;
 	struct pp_smu_display_requirement_rv smu_req = *smu_req_cur;

drivers/gpu/drm/amd/display/dc/dcn10/dcn10_dpp.c

···
 	default:
 		break;
 	}
+
+	/* Set default color space based on format if none is given. */
+	color_space = input_color_space ? input_color_space : color_space;
+
 	REG_SET(CNVC_SURFACE_PIXEL_FORMAT, 0,
 			CNVC_SURFACE_PIXEL_FORMAT, pixel_format);
 	REG_UPDATE(FORMAT_CONTROL, FORMAT_CONTROL__ALPHA_EN, alpha_en);
···
 		for (i = 0; i < 12; i++)
 			tbl_entry.regval[i] = input_csc_color_matrix.matrix[i];
 
-		tbl_entry.color_space = input_color_space;
+		tbl_entry.color_space = color_space;
 
 		if (color_space >= COLOR_SPACE_YCBCR601)
 			select = INPUT_CSC_SELECT_ICSC;

drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer.c

···
 
 	DTN_INFO("\nCALCULATED Clocks: dcfclk_khz:%d  dcfclk_deep_sleep_khz:%d  dispclk_khz:%d\n"
 		"dppclk_khz:%d  max_supported_dppclk_khz:%d  fclk_khz:%d  socclk_khz:%d\n\n",
-			dc->current_state->bw.dcn.clk.dcfclk_khz,
-			dc->current_state->bw.dcn.clk.dcfclk_deep_sleep_khz,
-			dc->current_state->bw.dcn.clk.dispclk_khz,
-			dc->current_state->bw.dcn.clk.dppclk_khz,
-			dc->current_state->bw.dcn.clk.max_supported_dppclk_khz,
-			dc->current_state->bw.dcn.clk.fclk_khz,
-			dc->current_state->bw.dcn.clk.socclk_khz);
+			dc->current_state->bw_ctx.bw.dcn.clk.dcfclk_khz,
+			dc->current_state->bw_ctx.bw.dcn.clk.dcfclk_deep_sleep_khz,
+			dc->current_state->bw_ctx.bw.dcn.clk.dispclk_khz,
+			dc->current_state->bw_ctx.bw.dcn.clk.dppclk_khz,
+			dc->current_state->bw_ctx.bw.dcn.clk.max_supported_dppclk_khz,
+			dc->current_state->bw_ctx.bw.dcn.clk.fclk_khz,
+			dc->current_state->bw_ctx.bw.dcn.clk.socclk_khz);
 
 	log_mpc_crc(dc, log_ctx);
 
···
 		 * to non-preferred front end. If pipe_ctx->stream is not NULL,
 		 * we will use the pipe, so don't disable
 		 */
-		if (pipe_ctx->stream != NULL)
+		if (pipe_ctx->stream != NULL && can_apply_seamless_boot)
 			continue;
-
-		if (tg->funcs->is_tg_enabled(tg))
-			tg->funcs->lock(tg);
 
 		/* Blank controller using driver code instead of
 		 * command table.
 		 */
 		if (tg->funcs->is_tg_enabled(tg)) {
+			tg->funcs->lock(tg);
 			tg->funcs->set_blank(tg, true);
 			hwss_wait_for_blank_complete(tg);
 		}
···
 
 	switch (top_pipe_ctx->plane_res.scl_data.format) {
 	case PIXEL_FORMAT_ARGB2101010:
-		if (top_pipe_ctx->stream->out_transfer_func->tf == TRANSFER_FUNCTION_UNITY) {
+		if (top_pipe_ctx->stream->out_transfer_func->tf == TRANSFER_FUNCTION_PQ) {
 			/* HDR10, ARGB2101010 - set boarder color to red */
 			color->color_r_cr = color_value;
 		}
···
 			plane_state->format,
 			EXPANSION_MODE_ZERO,
 			plane_state->input_csc_color_matrix,
-			COLOR_SPACE_YCBCR601_LIMITED);
+			plane_state->color_space);
 
 	//set scale and bias registers
 	build_prescale_params(&bns_params, plane_state);
···
 	 * divided by 2
 	 */
 	if (plane_state->update_flags.bits.full_update) {
-		bool should_divided_by_2 = context->bw.dcn.clk.dppclk_khz <=
+		bool should_divided_by_2 = context->bw_ctx.bw.dcn.clk.dppclk_khz <=
 				dc->res_pool->clk_mgr->clks.dispclk_khz / 2;
 
 		dpp->funcs->dpp_dppclk_control(
···
 	if (pipe_ctx->stream->cursor_attributes.address.quad_part != 0) {
 		dc->hwss.set_cursor_position(pipe_ctx);
 		dc->hwss.set_cursor_attribute(pipe_ctx);
+
+		if (dc->hwss.set_cursor_sdr_white_level)
+			dc->hwss.set_cursor_sdr_white_level(pipe_ctx);
 	}
 
 	if (plane_state->update_flags.bits.full_update) {
···
 	int i;
 	struct timing_generator *tg;
 	bool removed_pipe[4] = { false };
+	bool interdependent_update = false;
 	struct pipe_ctx *top_pipe_to_program =
 			find_top_pipe_for_stream(dc, context, stream);
 	DC_LOGGER_INIT(dc->ctx->logger);
···
 
 	tg = top_pipe_to_program->stream_res.tg;
 
-	dcn10_pipe_control_lock(dc, top_pipe_to_program, true);
+	interdependent_update = top_pipe_to_program->plane_state &&
+		top_pipe_to_program->plane_state->update_flags.bits.full_update;
+
+	if (interdependent_update)
+		lock_all_pipes(dc, context, true);
+	else
+		dcn10_pipe_control_lock(dc, top_pipe_to_program, true);
 
 	if (num_planes == 0) {
 		/* OTG blank before remove all front end */
···
 		 */
 		if (pipe_ctx->plane_state && !old_pipe_ctx->plane_state) {
 			if (old_pipe_ctx->stream_res.tg == tg &&
-				old_pipe_ctx->plane_res.hubp &&
-				old_pipe_ctx->plane_res.hubp->opp_id != 0xf) {
+			    old_pipe_ctx->plane_res.hubp &&
+			    old_pipe_ctx->plane_res.hubp->opp_id != 0xf)
 				dcn10_disable_plane(dc, old_pipe_ctx);
-				/*
-				 * power down fe will unlock when calling reset, need
-				 * to lock it back here. Messy, need rework.
-				 */
-				pipe_ctx->stream_res.tg->funcs->lock(pipe_ctx->stream_res.tg);
-			}
 		}
 
 		if ((!pipe_ctx->plane_state ||
···
 	if (num_planes > 0)
 		program_all_pipe_in_tree(dc, top_pipe_to_program, context);
 
-	dcn10_pipe_control_lock(dc, top_pipe_to_program, false);
-
-	if (top_pipe_to_program->plane_state &&
-			top_pipe_to_program->plane_state->update_flags.bits.full_update)
+	if (interdependent_update)
 		for (i = 0; i < dc->res_pool->pipe_count; i++) {
 			struct pipe_ctx *pipe_ctx = &context->res_ctx.pipe_ctx[i];
-			tg = pipe_ctx->stream_res.tg;
 			/* Skip inactive pipes and ones already updated */
-			if (!pipe_ctx->stream || pipe_ctx->stream == stream
-					|| !pipe_ctx->plane_state
-					|| !tg->funcs->is_tg_enabled(tg))
+			if (!pipe_ctx->stream || pipe_ctx->stream == stream ||
+			    !pipe_ctx->plane_state || !tg->funcs->is_tg_enabled(tg))
 				continue;
-
-			tg->funcs->lock(tg);
 
 			pipe_ctx->plane_res.hubp->funcs->hubp_setup_interdependent(
 				pipe_ctx->plane_res.hubp,
 				&pipe_ctx->dlg_regs,
 				&pipe_ctx->ttu_regs);
-
-			tg->funcs->unlock(tg);
 		}
+
+	if (interdependent_update)
+		lock_all_pipes(dc, context, false);
+	else
+		dcn10_pipe_control_lock(dc, top_pipe_to_program, false);
 
 	if (num_planes == 0)
 		false_optc_underflow_wa(dc, stream, tg);
···
 
 	if (!IS_FPGA_MAXIMUS_DC(dc->ctx->dce_environment)) {
 		if (context->stream_count == 0)
-			context->bw.dcn.clk.phyclk_khz = 0;
+			context->bw_ctx.bw.dcn.clk.phyclk_khz = 0;
 
 		dc->res_pool->clk_mgr->funcs->update_clocks(
 				dc->res_pool->clk_mgr,
···
 	}
 
 	hubbub1_program_watermarks(dc->res_pool->hubbub,
-			&context->bw.dcn.watermarks,
+			&context->bw_ctx.bw.dcn.watermarks,
 			dc->res_pool->ref_clocks.dchub_ref_clock_inKhz / 1000,
 			true);
 	dcn10_stereo_hw_frame_pack_wa(dc, context);
···
 
 	if (!IS_FPGA_MAXIMUS_DC(dc->ctx->dce_environment)) {
 		if (context->stream_count == 0)
-			context->bw.dcn.clk.phyclk_khz = 0;
+			context->bw_ctx.bw.dcn.clk.phyclk_khz = 0;
 
 		dc->res_pool->clk_mgr->funcs->update_clocks(
 				dc->res_pool->clk_mgr,
···
 	}
 
 	hubbub1_program_watermarks(dc->res_pool->hubbub,
-			&context->bw.dcn.watermarks,
+			&context->bw_ctx.bw.dcn.watermarks,
 			dc->res_pool->ref_clocks.dchub_ref_clock_inKhz / 1000,
 			true);
 	dcn10_stereo_hw_frame_pack_wa(dc, context);
···
 		.rotation = pipe_ctx->plane_state->rotation,
 		.mirror = pipe_ctx->plane_state->horizontal_mirror
 	};
+	uint32_t x_plane = pipe_ctx->plane_state->dst_rect.x;
+	uint32_t y_plane = pipe_ctx->plane_state->dst_rect.y;
+	uint32_t x_offset = min(x_plane, pos_cpy.x);
+	uint32_t y_offset = min(y_plane, pos_cpy.y);
 
-	pos_cpy.x_hotspot += pipe_ctx->plane_state->dst_rect.x;
-	pos_cpy.y_hotspot += pipe_ctx->plane_state->dst_rect.y;
+	pos_cpy.x -= x_offset;
+	pos_cpy.y -= y_offset;
+	pos_cpy.x_hotspot += (x_plane - x_offset);
+	pos_cpy.y_hotspot += (y_plane - y_offset);
 
 	if (pipe_ctx->plane_state->address.type
 			== PLN_ADDR_TYPE_VIDEO_PROGRESSIVE)
···
 			optc->dlg_otg_param.vstartup_start + 1;
 
 	return vertical_line_start;
+}
+
+void lock_all_pipes(struct dc *dc,
+	struct dc_state *context,
+	bool lock)
+{
+	struct pipe_ctx *pipe_ctx;
+	struct timing_generator *tg;
+	int i;
+
+	for (i = 0; i < dc->res_pool->pipe_count; i++) {
+		pipe_ctx = &context->res_ctx.pipe_ctx[i];
+		tg = pipe_ctx->stream_res.tg;
+		/*
+		 * Only lock the top pipe's tg to prevent redundant
+		 * (un)locking. Also skip if pipe is disabled.
+		 */
+		if (pipe_ctx->top_pipe ||
+		    !pipe_ctx->stream || !pipe_ctx->plane_state ||
+		    !tg->funcs->is_tg_enabled(tg))
+			continue;
+
+		if (lock)
+			tg->funcs->lock(tg);
+		else
+			tg->funcs->unlock(tg);
+	}
 }
 
 static void calc_vupdate_position(

drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer.h

···
 
 int get_vupdate_offset_from_vsync(struct pipe_ctx *pipe_ctx);
 
+void lock_all_pipes(struct dc *dc,
+	struct dc_state *context,
+	bool lock);
+
 #endif /* __DC_HWSS_DCN10_H__ */

drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer_debug.c

···
 	chars_printed = snprintf_count(pBuf, bufSize, "dcfclk,dcfclk_deep_sleep,dispclk,"
 		"dppclk,fclk,socclk\n"
 		"%d,%d,%d,%d,%d,%d\n",
-		dc->current_state->bw.dcn.clk.dcfclk_khz,
-		dc->current_state->bw.dcn.clk.dcfclk_deep_sleep_khz,
-		dc->current_state->bw.dcn.clk.dispclk_khz,
-		dc->current_state->bw.dcn.clk.dppclk_khz,
-		dc->current_state->bw.dcn.clk.fclk_khz,
-		dc->current_state->bw.dcn.clk.socclk_khz);
+		dc->current_state->bw_ctx.bw.dcn.clk.dcfclk_khz,
+		dc->current_state->bw_ctx.bw.dcn.clk.dcfclk_deep_sleep_khz,
+		dc->current_state->bw_ctx.bw.dcn.clk.dispclk_khz,
+		dc->current_state->bw_ctx.bw.dcn.clk.dppclk_khz,
+		dc->current_state->bw_ctx.bw.dcn.clk.fclk_khz,
+		dc->current_state->bw_ctx.bw.dcn.clk.socclk_khz);
 
 	remaining_buffer -= chars_printed;
 	pBuf += chars_printed;

drivers/gpu/drm/amd/display/dc/dcn10/dcn10_resource.c

···
 			continue;
 
 		if (context->stream_status[i].plane_count > 2)
-			return false;
+			return DC_FAIL_UNSUPPORTED_1;
 
 		for (j = 0; j < context->stream_status[i].plane_count; j++) {
 			struct dc_plane_state *plane =

drivers/gpu/drm/amd/display/dc/inc/core_status.h

···
 	DC_FAIL_BANDWIDTH_VALIDATE = 13, /* BW and Watermark validation */
 	DC_FAIL_SCALING = 14,
 	DC_FAIL_DP_LINK_TRAINING = 15,
-
+	DC_FAIL_UNSUPPORTED_1 = 18,
 	DC_ERROR_UNEXPECTED = -1
 };
 

drivers/gpu/drm/amd/display/dc/inc/core_types.h

···
 	struct dcn_watermark_set watermarks;
 };
 
-union bw_context {
+union bw_output {
 	struct dcn_bw_output dcn;
 	struct dce_bw_output dce;
 };
 
+struct bw_context {
+	union bw_output bw;
+	struct display_mode_lib dml;
+};
 /**
  * struct dc_state - The full description of a state requested by a user
  *
  * @streams: Stream properties
  * @stream_status: The planes on a given stream
  * @res_ctx: Persistent state of resources
- * @bw: The output from bandwidth and watermark calculations
+ * @bw_ctx: The output from bandwidth and watermark calculations and the DML
  * @pp_display_cfg: PowerPlay clocks and settings
  * @dcn_bw_vars: non-stack memory to support bandwidth calculations
  *
···
 
 	struct resource_context res_ctx;
 
-	union bw_context bw;
+	struct bw_context bw_ctx;
 
 	/* Note: these are big structures, do *not* put on stack! */
 	struct dm_pp_display_configuration pp_display_cfg;
···
 	struct dcn_bw_internal_vars dcn_bw_vars;
 #endif
 
-	struct clk_mgr *dccg;
+	struct clk_mgr *clk_mgr;
 
 	struct {
 		bool full_update_needed : 1;

drivers/gpu/drm/amd/display/dc/irq/dce110/irq_service_dce110.c

···
 	.ack = NULL
 };
 
+static const struct irq_source_info_funcs vupdate_irq_info_funcs = {
+	.set = NULL,
+	.ack = NULL
+};
+
 #define hpd_int_entry(reg_num)\
 	[DC_IRQ_SOURCE_HPD1 + reg_num] = {\
 		.enable_reg = mmHPD ## reg_num ## _DC_HPD_INT_CONTROL,\
···
 		CRTC_V_UPDATE_INT_STATUS__CRTC_V_UPDATE_INT_CLEAR_MASK,\
 		.ack_value =\
 		CRTC_V_UPDATE_INT_STATUS__CRTC_V_UPDATE_INT_CLEAR_MASK,\
-		.funcs = &vblank_irq_info_funcs\
+		.funcs = &vupdate_irq_info_funcs\
 	}
 
 #define vblank_int_entry(reg_num)\

drivers/gpu/drm/amd/display/dc/irq/dce120/irq_service_dce120.c

···
 	.ack = NULL
 };
 
+static const struct irq_source_info_funcs vupdate_irq_info_funcs = {
+	.set = NULL,
+	.ack = NULL
+};
+
 #define BASE_INNER(seg) \
 	DCE_BASE__INST0_SEG ## seg
 
···
 		IRQ_REG_ENTRY(CRTC, reg_num,\
 			CRTC_INTERRUPT_CONTROL, CRTC_V_UPDATE_INT_MSK,\
 			CRTC_V_UPDATE_INT_STATUS, CRTC_V_UPDATE_INT_CLEAR),\
-		.funcs = &vblank_irq_info_funcs\
+		.funcs = &vupdate_irq_info_funcs\
 	}
 
 #define vblank_int_entry(reg_num)\

drivers/gpu/drm/amd/display/dc/irq/dce80/irq_service_dce80.c

···
 	.ack = NULL
 };
 
+static const struct irq_source_info_funcs vupdate_irq_info_funcs = {
+	.set = NULL,
+	.ack = NULL
+};
 
 #define hpd_int_entry(reg_num)\
 	[DC_IRQ_SOURCE_INVALID + reg_num] = {\
···
 		CRTC_V_UPDATE_INT_STATUS__CRTC_V_UPDATE_INT_CLEAR_MASK,\
 		.ack_value =\
 		CRTC_V_UPDATE_INT_STATUS__CRTC_V_UPDATE_INT_CLEAR_MASK,\
-		.funcs = &vblank_irq_info_funcs\
+		.funcs = &vupdate_irq_info_funcs\
 	}
 
 #define vblank_int_entry(reg_num)\

drivers/gpu/drm/amd/display/dc/irq/dcn10/irq_service_dcn10.c

···
 		return DC_IRQ_SOURCE_VBLANK5;
 	case DCN_1_0__SRCID__DC_D6_OTG_VSTARTUP:
 		return DC_IRQ_SOURCE_VBLANK6;
+	case DCN_1_0__SRCID__OTG0_IHC_V_UPDATE_NO_LOCK_INTERRUPT:
+		return DC_IRQ_SOURCE_VUPDATE1;
+	case DCN_1_0__SRCID__OTG1_IHC_V_UPDATE_NO_LOCK_INTERRUPT:
+		return DC_IRQ_SOURCE_VUPDATE2;
+	case DCN_1_0__SRCID__OTG2_IHC_V_UPDATE_NO_LOCK_INTERRUPT:
+		return DC_IRQ_SOURCE_VUPDATE3;
+	case DCN_1_0__SRCID__OTG3_IHC_V_UPDATE_NO_LOCK_INTERRUPT:
+		return DC_IRQ_SOURCE_VUPDATE4;
+	case DCN_1_0__SRCID__OTG4_IHC_V_UPDATE_NO_LOCK_INTERRUPT:
+		return DC_IRQ_SOURCE_VUPDATE5;
+	case DCN_1_0__SRCID__OTG5_IHC_V_UPDATE_NO_LOCK_INTERRUPT:
+		return DC_IRQ_SOURCE_VUPDATE6;
 	case DCN_1_0__SRCID__HUBP0_FLIP_INTERRUPT:
 		return DC_IRQ_SOURCE_PFLIP1;
 	case DCN_1_0__SRCID__HUBP1_FLIP_INTERRUPT:
···
 	.ack = NULL
 };
 
+static const struct irq_source_info_funcs vupdate_no_lock_irq_info_funcs = {
+	.set = NULL,
+	.ack = NULL
+};
+
 #define BASE_INNER(seg) \
 	DCE_BASE__INST0_SEG ## seg
 
···
 		.funcs = &pflip_irq_info_funcs\
 	}
 
-#define vupdate_int_entry(reg_num)\
+/* vupdate_no_lock_int_entry maps to DC_IRQ_SOURCE_VUPDATEx, to match semantic
+ * of DCE's DC_IRQ_SOURCE_VUPDATEx.
+ */
+#define vupdate_no_lock_int_entry(reg_num)\
 	[DC_IRQ_SOURCE_VUPDATE1 + reg_num] = {\
 		IRQ_REG_ENTRY(OTG, reg_num,\
-			OTG_GLOBAL_SYNC_STATUS, VUPDATE_INT_EN,\
-			OTG_GLOBAL_SYNC_STATUS, VUPDATE_EVENT_CLEAR),\
-		.funcs = &vblank_irq_info_funcs\
+			OTG_GLOBAL_SYNC_STATUS, VUPDATE_NO_LOCK_INT_EN,\
+			OTG_GLOBAL_SYNC_STATUS, VUPDATE_NO_LOCK_EVENT_CLEAR),\
+		.funcs = &vupdate_no_lock_irq_info_funcs\
 	}
 
 #define vblank_int_entry(reg_num)\
···
 	dc_underflow_int_entry(6),
 	[DC_IRQ_SOURCE_DMCU_SCP] = dummy_irq_entry(),
 	[DC_IRQ_SOURCE_VBIOS_SW] = dummy_irq_entry(),
-	vupdate_int_entry(0),
-	vupdate_int_entry(1),
-	vupdate_int_entry(2),
-	vupdate_int_entry(3),
-	vupdate_int_entry(4),
-	vupdate_int_entry(5),
+	vupdate_no_lock_int_entry(0),
+	vupdate_no_lock_int_entry(1),
+	vupdate_no_lock_int_entry(2),
+	vupdate_no_lock_int_entry(3),
+	vupdate_no_lock_int_entry(4),
+	vupdate_no_lock_int_entry(5),
 	vblank_int_entry(0),
 	vblank_int_entry(1),
 	vblank_int_entry(2),

drivers/gpu/drm/amd/display/modules/power/power_helpers.c

···
 
 /* Possible ABM 2.2 Min Reduction configs from least aggressive to most aggressive
  *  0    1     2     3     4     5     6     7     8     9     10    11   12
- * 100  100   100   100   100   100   100   90.2  85.1  80.0  80.0  75.3  75.3 %
+ * 100  100   100   100   100   100   100   100  100  92.2  83.1  75.3  75.3 %
  */
 static const unsigned char min_reduction_table_v_2_2[13] = {
-0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xe6, 0xd9, 0xcc, 0xcc, 0xc0, 0xc0};
+0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xeb, 0xd4, 0xc0, 0xc0};
 
 /* Possible ABM 2.2 Max Reduction configs from least aggressive to most aggressive
  *  0    1     2     3     4     5     6     7     8     9     10    11   12

drivers/gpu/drm/amd/powerplay/amdgpu_smu.c

···
 	return ret;
 }
 
-int smu_update_table(struct smu_context *smu, uint32_t table_id,
+int smu_update_table_with_arg(struct smu_context *smu, uint16_t table_id, uint16_t exarg,
 		     void *table_data, bool drv2smu)
 {
 	struct smu_table_context *smu_table = &smu->smu_table;
 	struct smu_table *table = NULL;
 	int ret = 0;
+	uint32_t table_index;
 
 	if (!table_data || table_id >= smu_table->table_count)
 		return -EINVAL;
+
+	table_index = (exarg << 16) | table_id;
 
 	table = &smu_table->tables[table_id];
 
···
 	ret = smu_send_smc_msg_with_param(smu, drv2smu ?
 					  SMU_MSG_TransferTableDram2Smu :
 					  SMU_MSG_TransferTableSmu2Dram,
-					  table_id);
+					  table_index);
 	if (ret)
 		return ret;
 

drivers/gpu/drm/amd/powerplay/hwmgr/smu7_hwmgr.c

···
 #define PCIE_BUS_CLK                10000
 #define TCLK                        (PCIE_BUS_CLK / 10)
 
-static const struct profile_mode_setting smu7_profiling[7] =
+static struct profile_mode_setting smu7_profiling[7] =
 					{{0, 0, 0, 0, 0, 0, 0, 0},
 					 {1, 0, 100, 30, 1, 0, 100, 10},
 					 {1, 10, 0, 30, 0, 0, 0, 0},
···
 	mode = input[size];
 	switch (mode) {
 	case PP_SMC_POWER_PROFILE_CUSTOM:
-		if (size < 8)
+		if (size < 8 && size != 0)
 			return -EINVAL;
-
-		tmp.bupdate_sclk = input[0];
-		tmp.sclk_up_hyst = input[1];
-		tmp.sclk_down_hyst = input[2];
-		tmp.sclk_activity = input[3];
-		tmp.bupdate_mclk = input[4];
-		tmp.mclk_up_hyst = input[5];
-		tmp.mclk_down_hyst = input[6];
-		tmp.mclk_activity = input[7];
+		/* If only CUSTOM is passed in, use the saved values. Check
+		 * that we actually have a CUSTOM profile by ensuring that
+		 * the "use sclk" or the "use mclk" bits are set
+		 */
+		tmp = smu7_profiling[PP_SMC_POWER_PROFILE_CUSTOM];
+		if (size == 0) {
+			if (tmp.bupdate_sclk == 0 && tmp.bupdate_mclk == 0)
+				return -EINVAL;
+		} else {
+			tmp.bupdate_sclk = input[0];
+			tmp.sclk_up_hyst = input[1];
+			tmp.sclk_down_hyst = input[2];
+			tmp.sclk_activity = input[3];
+			tmp.bupdate_mclk = input[4];
+			tmp.mclk_up_hyst = input[5];
+			tmp.mclk_down_hyst = input[6];
+			tmp.mclk_activity = input[7];
+			smu7_profiling[PP_SMC_POWER_PROFILE_CUSTOM] = tmp;
+		}
 		if (!smum_update_dpm_settings(hwmgr, &tmp)) {
 			memcpy(&data->current_profile_setting, &tmp, sizeof(struct profile_mode_setting));
 			hwmgr->power_profile_mode = mode;

drivers/gpu/drm/amd/powerplay/hwmgr/vega10_hwmgr.c

···
 
 	vega10_setup_default_pcie_table(hwmgr);
 
+	/* Zero out the saved copy of the CUSTOM profile
+	 * This will be checked when trying to set the profile
+	 * and will require that new values be passed in
+	 */
+	data->custom_profile_mode[0] = 0;
+	data->custom_profile_mode[1] = 0;
+	data->custom_profile_mode[2] = 0;
+	data->custom_profile_mode[3] = 0;
+
 	/* save a copy of the default DPM table */
 	memcpy(&(data->golden_dpm_table), &(data->dpm_table),
 			sizeof(struct vega10_dpm_table));
···
 	uint8_t min_active_level;
 	uint32_t power_profile_mode = input[size];
 
-	smum_send_msg_to_smc_with_parameter(hwmgr, PPSMC_MSG_SetWorkloadMask,
-						1 << power_profile_mode);
-
 	if (power_profile_mode == PP_SMC_POWER_PROFILE_CUSTOM) {
-		if (size == 0 || size > 4)
+		if (size != 0 && size != 4)
 			return -EINVAL;
+
+		/* If size = 0 and the CUSTOM profile has been set already
+		 * then just apply the profile. The copy stored in the hwmgr
+		 * is zeroed out on init
+		 */
+		if (size == 0) {
+			if (data->custom_profile_mode[0] != 0)
+				goto out;
+			else
+				return -EINVAL;
+		}
 
 		data->custom_profile_mode[0] = busy_set_point = input[0];
 		data->custom_profile_mode[1] = FPS = input[1];
···
 					use_rlc_busy << 16 | min_active_level<<24);
 	}
 
+out:
+	smum_send_msg_to_smc_with_parameter(hwmgr, PPSMC_MSG_SetWorkloadMask,
+						1 << power_profile_mode);
 	hwmgr->power_profile_mode = power_profile_mode;
 
 	return 0;

drivers/gpu/drm/amd/powerplay/hwmgr/vega20_hwmgr.c

···
 	hwmgr->platform_descriptor.clockStep.memoryClock = 500;
 
 	data->total_active_cus = adev->gfx.cu_info.number;
+	data->is_custom_profile_set = false;
 
 	return 0;
 }
···
 	}
 
 	if (power_profile_mode == PP_SMC_POWER_PROFILE_CUSTOM) {
-		if (size < 10)
+		struct vega20_hwmgr *data =
+			(struct vega20_hwmgr *)(hwmgr->backend);
+		if (size == 0 && !data->is_custom_profile_set)
+			return -EINVAL;
+		if (size < 10 && size != 0)
 			return -EINVAL;
 
 		result = vega20_get_activity_monitor_coeff(hwmgr,
···
 		PP_ASSERT_WITH_CODE(!result,
 				"[SetPowerProfile] Failed to get activity monitor!",
 				return result);
+
+		/* If size==0, then we want to apply the already-configured
+		 * CUSTOM profile again. Just apply it, since we checked its
+		 * validity above
+		 */
+		if (size == 0)
+			goto out;
 
 		switch (input[0]) {
 		case 0: /* Gfxclk */
···
 		result = vega20_set_activity_monitor_coeff(hwmgr,
 				(uint8_t *)(&activity_monitor),
 				WORKLOAD_PPLIB_CUSTOM_BIT);
+		data->is_custom_profile_set = true;
 		PP_ASSERT_WITH_CODE(!result,
 				"[SetPowerProfile] Failed to set activity monitor!",
 				return result);
 	}
 
+out:
 	/* conv PP_SMC_POWER_PROFILE* to WORKLOAD_PPLIB_*_BIT */
 	workload_type =
 		conv_power_profile_to_pplib_workload(power_profile_mode);

drivers/gpu/drm/amd/powerplay/hwmgr/vega20_hwmgr.h

···
 	bool                           pcie_parameters_override;
 	uint32_t                       pcie_gen_level1;
 	uint32_t                       pcie_width_level1;
+
+	bool                           is_custom_profile_set;
 };
 
 #define VEGA20_DPM2_NEAR_TDP_DEC                      10

drivers/gpu/drm/amd/powerplay/inc/amdgpu_smu.h

···
 					       struct dm_pp_wm_sets_with_clock_ranges_soc15 *clock_ranges);
 	int (*set_od8_default_settings)(struct smu_context *smu,
 					bool initialize);
-	int (*get_activity_monitor_coeff)(struct smu_context *smu,
-				      uint8_t *table,
-				      uint16_t workload_type);
-	int (*set_activity_monitor_coeff)(struct smu_context *smu,
-				      uint8_t *table,
-				      uint16_t workload_type);
 	int (*conv_power_profile_to_pplib_workload)(int power_profile);
 	int (*get_power_profile_mode)(struct smu_context *smu, char *buf);
 	int (*set_power_profile_mode)(struct smu_context *smu, long *input, uint32_t size);
···
 extern int smu_feature_is_supported(struct smu_context *smu, int feature_id);
 extern int smu_feature_set_supported(struct smu_context *smu, int feature_id, bool enable);
 
-int smu_update_table(struct smu_context *smu, uint32_t table_id,
+int smu_update_table_with_arg(struct smu_context *smu, uint16_t table_id, uint16_t exarg,
 		     void *table_data, bool drv2smu);
+#define smu_update_table(smu, table_id, table_data, drv2smu) \
+	smu_update_table_with_arg((smu), (table_id), 0, (table_data), (drv2smu))
+
 bool is_support_sw_smu(struct amdgpu_device *adev);
 int smu_reset(struct smu_context *smu);
 int smu_common_read_sensor(struct smu_context *smu, enum amd_pp_sensors sensor,

drivers/gpu/drm/amd/powerplay/smu_v11_0.c

···
 	return 0;
 }
 
-static int smu_v11_0_set_activity_monitor_coeff(struct smu_context *smu,
-				      uint8_t *table, uint16_t workload_type)
-{
-	int ret = 0;
-	memcpy(smu->smu_table.tables[TABLE_ACTIVITY_MONITOR_COEFF].cpu_addr,
-	       table, smu->smu_table.tables[TABLE_ACTIVITY_MONITOR_COEFF].size);
-	ret = smu_send_smc_msg_with_param(smu, SMU_MSG_SetDriverDramAddrHigh,
-					  upper_32_bits(smu->smu_table.tables[TABLE_ACTIVITY_MONITOR_COEFF].mc_address));
-	if (ret) {
-		pr_err("[%s] Attempt to Set Dram Addr High Failed!", __func__);
-		return ret;
-	}
-	ret = smu_send_smc_msg_with_param(smu, SMU_MSG_SetDriverDramAddrLow,
-					  lower_32_bits(smu->smu_table.tables[TABLE_ACTIVITY_MONITOR_COEFF].mc_address));
-	if (ret) {
-		pr_err("[%s] Attempt to Set Dram Addr Low Failed!", __func__);
-		return ret;
-	}
-	ret = smu_send_smc_msg_with_param(smu, SMU_MSG_TransferTableSmu2Dram,
-					  TABLE_ACTIVITY_MONITOR_COEFF | (workload_type << 16));
-	if (ret) {
-		pr_err("[%s] Attempt to Transfer Table From SMU Failed!", __func__);
-		return ret;
-	}
-
-	return ret;
-}
-
-static int smu_v11_0_get_activity_monitor_coeff(struct smu_context *smu,
-				      uint8_t *table, uint16_t workload_type)
-{
-	int ret = 0;
-	ret = smu_send_smc_msg_with_param(smu, SMU_MSG_SetDriverDramAddrHigh,
-					  upper_32_bits(smu->smu_table.tables[TABLE_ACTIVITY_MONITOR_COEFF].mc_address));
-	if (ret) {
-		pr_err("[%s] Attempt to Set Dram Addr High Failed!", __func__);
-		return ret;
-	}
-
-	ret = smu_send_smc_msg_with_param(smu, SMU_MSG_SetDriverDramAddrLow,
-					  lower_32_bits(smu->smu_table.tables[TABLE_ACTIVITY_MONITOR_COEFF].mc_address));
-	if (ret) {
-		pr_err("[%s] Attempt to Set Dram Addr Low Failed!", __func__);
-		return ret;
-	}
-
-	ret = smu_send_smc_msg_with_param(smu, SMU_MSG_TransferTableSmu2Dram,
-					  TABLE_ACTIVITY_MONITOR_COEFF | (workload_type << 16));
-	if (ret) {
-		pr_err("[%s] Attempt to Transfer Table From SMU Failed!", __func__);
-		return ret;
-	}
-
-	return ret;
-}
-
 static int smu_v11_0_conv_power_profile_to_pplib_workload(int power_profile)
 {
 	int pplib_workload = 0;
···
 	for (i = 0; i <= PP_SMC_POWER_PROFILE_CUSTOM; i++) {
 		/* conv PP_SMC_POWER_PROFILE* to WORKLOAD_PPLIB_*_BIT */
 		workload_type = smu_v11_0_conv_power_profile_to_pplib_workload(i);
-		result = smu_v11_0_get_activity_monitor_coeff(smu,
-							      (uint8_t *)(&activity_monitor),
-							      workload_type);
+		result = smu_update_table_with_arg(smu, TABLE_ACTIVITY_MONITOR_COEFF,
+						   workload_type, &activity_monitor, false);
 		if (result) {
 			pr_err("[%s] Failed to get activity monitor!", __func__);
 			return result;
···
 static int smu_v11_0_set_power_profile_mode(struct smu_context *smu, long *input, uint32_t size)
 {
 	DpmActivityMonitorCoeffInt_t activity_monitor;
-	int workload_type, ret = 0;
+	int workload_type = 0, ret = 0;
 
 	smu->power_profile_mode = input[size];
 
···
 	}
 
 	if (smu->power_profile_mode == PP_SMC_POWER_PROFILE_CUSTOM) {
-		ret = smu_v11_0_get_activity_monitor_coeff(smu,
-							   (uint8_t *)(&activity_monitor),
-							   WORKLOAD_PPLIB_CUSTOM_BIT);
+		ret = smu_update_table_with_arg(smu, TABLE_ACTIVITY_MONITOR_COEFF,
+						WORKLOAD_PPLIB_CUSTOM_BIT, &activity_monitor, false);
 		if (ret) {
 			pr_err("[%s] Failed to get activity monitor!", __func__);
 			return ret;
···
 			break;
 		}
 
-		ret = smu_v11_0_set_activity_monitor_coeff(smu,
-							   (uint8_t *)(&activity_monitor),
-							   WORKLOAD_PPLIB_CUSTOM_BIT);
+		ret = smu_update_table_with_arg(smu, TABLE_ACTIVITY_MONITOR_COEFF,
+						WORKLOAD_PPLIB_COMPUTE_BIT, &activity_monitor, true);
 		if (ret) {
 			pr_err("[%s] Failed to set activity monitor!", __func__);
 			return ret;
···
 	.get_sclk = smu_v11_0_dpm_get_sclk,
 	.get_mclk = smu_v11_0_dpm_get_mclk,
 	.set_od8_default_settings = smu_v11_0_set_od8_default_settings,
-	.get_activity_monitor_coeff = smu_v11_0_get_activity_monitor_coeff,
-	.set_activity_monitor_coeff = smu_v11_0_set_activity_monitor_coeff,
 	.conv_power_profile_to_pplib_workload = smu_v11_0_conv_power_profile_to_pplib_workload,
 	.get_power_profile_mode = smu_v11_0_get_power_profile_mode,
 	.set_power_profile_mode = smu_v11_0_set_power_profile_mode,