Merge tag 'drm-misc-next-2020-09-21' of git://anongit.freedesktop.org/drm/drm-misc into drm-next

drm-misc-next for 5.10:

UAPI Changes:

Cross-subsystem Changes:
  - virtio: Merged a PR for patches that will affect drm/virtio

Core Changes:
  - dev: More devm_drm convertions and removal of drm_dev_init
  - atomic: Split out drm_atomic_helper_calc_timestamping_constants of
    drm_atomic_helper_update_legacy_modeset_state
  - ttm: More rework

Driver Changes:
  - i915: selftests improvements
  - panfrost: support for Amlogic SoC
  - vc4: one fix
  - tree-wide: conversions to devm_drm_dev_alloc,
  - ast: simplifications of the atomic modesetting code
  - panfrost: multiple fixes
  - vc4: multiple fixes
Signed-off-by: Dave Airlie <airlied@redhat.com>

From: Maxime Ripard <maxime@cerno.tech>
Link: https://patchwork.freedesktop.org/patch/msgid/20200921152956.2gxnsdgxmwhvjyut@gilmour.lan

165 files changed, 9579 insertions, 2684 deletions

diff --git a/drivers/gpu/drm/Makefile b/drivers/gpu/drm/Makefile
index 2f31579f91d4..81569009f884 100644
--- a/drivers/gpu/drm/Makefile
+++ b/drivers/gpu/drm/Makefile
@@ -100,7 +100,7 @@ obj-$(CONFIG_DRM_MSM) += msm/
 obj-$(CONFIG_DRM_TEGRA) += tegra/
 obj-$(CONFIG_DRM_STM) += stm/
 obj-$(CONFIG_DRM_STI) += sti/
-obj-$(CONFIG_DRM_IMX) += imx/
+obj-y 			+= imx/
 obj-$(CONFIG_DRM_INGENIC) += ingenic/
 obj-$(CONFIG_DRM_MEDIATEK) += mediatek/
 obj-$(CONFIG_DRM_MESON)	+= meson/
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_dma_buf.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_dma_buf.c
index 427a7c5ba93d..957934926b24 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_dma_buf.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_dma_buf.c
@@ -303,7 +303,8 @@ static struct sg_table *amdgpu_dma_buf_map(struct dma_buf_attachment *attach,
 
 	switch (bo->tbo.mem.mem_type) {
 	case TTM_PL_TT:
-		sgt = drm_prime_pages_to_sg(bo->tbo.ttm->pages,
+		sgt = drm_prime_pages_to_sg(obj->dev,
+					    bo->tbo.ttm->pages,
 					    bo->tbo.num_pages);
 		if (IS_ERR(sgt))
 			return sgt;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
index a4b518211b1f..81e4cf869f50 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
@@ -1159,25 +1159,20 @@ static int amdgpu_pci_probe(struct pci_dev *pdev,
 	if (ret)
 		return ret;
 
-	adev = kzalloc(sizeof(*adev), GFP_KERNEL);
-	if (!adev)
-		return -ENOMEM;
+	adev = devm_drm_dev_alloc(&pdev->dev, &kms_driver, typeof(*adev), ddev);
+	if (IS_ERR(adev))
+		return PTR_ERR(adev);
 
 	adev->dev  = &pdev->dev;
 	adev->pdev = pdev;
 	ddev = adev_to_drm(adev);
-	ret = drm_dev_init(ddev, &kms_driver, &pdev->dev);
-	if (ret)
-		goto err_free;
-
-	drmm_add_final_kfree(ddev, adev);
 
 	if (!supports_atomic)
 		ddev->driver_features &= ~DRIVER_ATOMIC;
 
 	ret = pci_enable_device(pdev);
 	if (ret)
-		goto err_free;
+		return ret;
 
 	ddev->pdev = pdev;
 	pci_set_drvdata(pdev, ddev);
@@ -1205,8 +1200,6 @@ retry_init:
 
 err_pci:
 	pci_disable_device(pdev);
-err_free:
-	drm_dev_put(ddev);
 	return ret;
 }
 
@@ -1223,7 +1216,6 @@ amdgpu_pci_remove(struct pci_dev *pdev)
 	amdgpu_driver_unload_kms(dev);
 	pci_disable_device(pdev);
 	pci_set_drvdata(pdev, NULL);
-	drm_dev_put(dev);
 }
 
 static void
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gtt_mgr.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gtt_mgr.c
index 2b1b7c136343..f203e4a6a3f2 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gtt_mgr.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gtt_mgr.c
@@ -95,8 +95,6 @@ int amdgpu_gtt_mgr_init(struct amdgpu_device *adev, uint64_t gtt_size)
 
 	man->use_tt = true;
 	man->func = &amdgpu_gtt_mgr_func;
-	man->available_caching = TTM_PL_MASK_CACHING;
-	man->default_caching = TTM_PL_FLAG_CACHED;
 
 	ttm_resource_manager_init(man, gtt_size >> PAGE_SHIFT);
 
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c
index 887a4da5eefc..ac043baac05d 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c
@@ -136,8 +136,8 @@ void amdgpu_bo_placement_from_domain(struct amdgpu_bo *abo, u32 domain)
 
 		places[c].fpfn = 0;
 		places[c].lpfn = 0;
-		places[c].flags = TTM_PL_FLAG_WC | TTM_PL_FLAG_UNCACHED |
-			TTM_PL_FLAG_VRAM;
+		places[c].mem_type = TTM_PL_VRAM;
+		places[c].flags = TTM_PL_FLAG_WC | TTM_PL_FLAG_UNCACHED;
 
 		if (flags & AMDGPU_GEM_CREATE_CPU_ACCESS_REQUIRED)
 			places[c].lpfn = visible_pfn;
@@ -152,7 +152,8 @@ void amdgpu_bo_placement_from_domain(struct amdgpu_bo *abo, u32 domain)
 	if (domain & AMDGPU_GEM_DOMAIN_GTT) {
 		places[c].fpfn = 0;
 		places[c].lpfn = 0;
-		places[c].flags = TTM_PL_FLAG_TT;
+		places[c].mem_type = TTM_PL_TT;
+		places[c].flags = 0;
 		if (flags & AMDGPU_GEM_CREATE_CPU_GTT_USWC)
 			places[c].flags |= TTM_PL_FLAG_WC |
 				TTM_PL_FLAG_UNCACHED;
@@ -164,7 +165,8 @@ void amdgpu_bo_placement_from_domain(struct amdgpu_bo *abo, u32 domain)
 	if (domain & AMDGPU_GEM_DOMAIN_CPU) {
 		places[c].fpfn = 0;
 		places[c].lpfn = 0;
-		places[c].flags = TTM_PL_FLAG_SYSTEM;
+		places[c].mem_type = TTM_PL_SYSTEM;
+		places[c].flags = 0;
 		if (flags & AMDGPU_GEM_CREATE_CPU_GTT_USWC)
 			places[c].flags |= TTM_PL_FLAG_WC |
 				TTM_PL_FLAG_UNCACHED;
@@ -176,28 +178,32 @@ void amdgpu_bo_placement_from_domain(struct amdgpu_bo *abo, u32 domain)
 	if (domain & AMDGPU_GEM_DOMAIN_GDS) {
 		places[c].fpfn = 0;
 		places[c].lpfn = 0;
-		places[c].flags = TTM_PL_FLAG_UNCACHED | AMDGPU_PL_FLAG_GDS;
+		places[c].mem_type = AMDGPU_PL_GDS;
+		places[c].flags = TTM_PL_FLAG_UNCACHED;
 		c++;
 	}
 
 	if (domain & AMDGPU_GEM_DOMAIN_GWS) {
 		places[c].fpfn = 0;
 		places[c].lpfn = 0;
-		places[c].flags = TTM_PL_FLAG_UNCACHED | AMDGPU_PL_FLAG_GWS;
+		places[c].mem_type = AMDGPU_PL_GWS;
+		places[c].flags = TTM_PL_FLAG_UNCACHED;
 		c++;
 	}
 
 	if (domain & AMDGPU_GEM_DOMAIN_OA) {
 		places[c].fpfn = 0;
 		places[c].lpfn = 0;
-		places[c].flags = TTM_PL_FLAG_UNCACHED | AMDGPU_PL_FLAG_OA;
+		places[c].mem_type = AMDGPU_PL_OA;
+		places[c].flags = TTM_PL_FLAG_UNCACHED;
 		c++;
 	}
 
 	if (!c) {
 		places[c].fpfn = 0;
 		places[c].lpfn = 0;
-		places[c].flags = TTM_PL_MASK_CACHING | TTM_PL_FLAG_SYSTEM;
+		places[c].mem_type = TTM_PL_SYSTEM;
+		places[c].flags = TTM_PL_MASK_CACHING;
 		c++;
 	}
 
@@ -594,7 +600,7 @@ static int amdgpu_bo_do_create(struct amdgpu_device *adev,
 		amdgpu_cs_report_moved_bytes(adev, ctx.bytes_moved, 0);
 
 	if (bp->flags & AMDGPU_GEM_CREATE_VRAM_CLEARED &&
-	    bo->tbo.mem.placement & TTM_PL_FLAG_VRAM) {
+	    bo->tbo.mem.mem_type == TTM_PL_VRAM) {
 		struct dma_fence *fence;
 
 		r = amdgpu_fill_buffer(bo, 0, bo->tbo.base.resv, &fence);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
index 4a85f8cedd77..8039d2399584 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
@@ -63,12 +63,15 @@
 
 #define AMDGPU_TTM_VRAM_MAX_DW_READ	(size_t)128
 
+static int amdgpu_ttm_backend_bind(struct ttm_bo_device *bdev,
+				   struct ttm_tt *ttm,
+				   struct ttm_resource *bo_mem);
+
 static int amdgpu_ttm_init_on_chip(struct amdgpu_device *adev,
 				    unsigned int type,
 				    uint64_t size)
 {
 	return ttm_range_man_init(&adev->mman.bdev, type,
-				  TTM_PL_FLAG_UNCACHED, TTM_PL_FLAG_UNCACHED,
 				  false, size >> PAGE_SHIFT);
 }
 
@@ -88,7 +91,8 @@ static void amdgpu_evict_flags(struct ttm_buffer_object *bo,
 	static const struct ttm_place placements = {
 		.fpfn = 0,
 		.lpfn = 0,
-		.flags = TTM_PL_MASK_CACHING | TTM_PL_FLAG_SYSTEM
+		.mem_type = TTM_PL_SYSTEM,
+		.flags = TTM_PL_MASK_CACHING
 	};
 
 	/* Don't handle scatter gather BOs */
@@ -175,24 +179,6 @@ static int amdgpu_verify_access(struct ttm_buffer_object *bo, struct file *filp)
 }
 
 /**
- * amdgpu_move_null - Register memory for a buffer object
- *
- * @bo: The bo to assign the memory to
- * @new_mem: The memory to be assigned.
- *
- * Assign the memory from new_mem to the memory of the buffer object bo.
- */
-static void amdgpu_move_null(struct ttm_buffer_object *bo,
-			     struct ttm_resource *new_mem)
-{
-	struct ttm_resource *old_mem = &bo->mem;
-
-	BUG_ON(old_mem->mm_node != NULL);
-	*old_mem = *new_mem;
-	new_mem->mm_node = NULL;
-}
-
-/**
  * amdgpu_mm_node_addr - Compute the GPU relative offset of a GTT buffer.
  *
  * @bo: The bo to assign the memory to.
@@ -514,9 +500,9 @@ static int amdgpu_move_blit(struct ttm_buffer_object *bo,
 
 	/* Always block for VM page tables before committing the new location */
 	if (bo->type == ttm_bo_type_kernel)
-		r = ttm_bo_move_accel_cleanup(bo, fence, true, new_mem);
+		r = ttm_bo_move_accel_cleanup(bo, fence, true, false, new_mem);
 	else
-		r = ttm_bo_pipeline_move(bo, fence, evict, new_mem);
+		r = ttm_bo_move_accel_cleanup(bo, fence, evict, true, new_mem);
 	dma_fence_put(fence);
 	return r;
 
@@ -551,7 +537,8 @@ static int amdgpu_move_vram_ram(struct ttm_buffer_object *bo, bool evict,
 	placement.busy_placement = &placements;
 	placements.fpfn = 0;
 	placements.lpfn = 0;
-	placements.flags = TTM_PL_MASK_CACHING | TTM_PL_FLAG_TT;
+	placements.mem_type = TTM_PL_TT;
+	placements.flags = TTM_PL_MASK_CACHING;
 	r = ttm_bo_mem_space(bo, &placement, &tmp_mem, ctx);
 	if (unlikely(r)) {
 		pr_err("Failed to find GTT space for blit from VRAM\n");
@@ -564,8 +551,12 @@ static int amdgpu_move_vram_ram(struct ttm_buffer_object *bo, bool evict,
 		goto out_cleanup;
 	}
 
+	r = ttm_tt_populate(bo->bdev, bo->ttm, ctx);
+	if (unlikely(r))
+		goto out_cleanup;
+
 	/* Bind the memory to the GTT space */
-	r = ttm_tt_bind(bo->ttm, &tmp_mem, ctx);
+	r = amdgpu_ttm_backend_bind(bo->bdev, bo->ttm, &tmp_mem);
 	if (unlikely(r)) {
 		goto out_cleanup;
 	}
@@ -607,7 +598,8 @@ static int amdgpu_move_ram_vram(struct ttm_buffer_object *bo, bool evict,
 	placement.busy_placement = &placements;
 	placements.fpfn = 0;
 	placements.lpfn = 0;
-	placements.flags = TTM_PL_MASK_CACHING | TTM_PL_FLAG_TT;
+	placements.mem_type = TTM_PL_TT;
+	placements.flags = TTM_PL_MASK_CACHING;
 	r = ttm_bo_mem_space(bo, &placement, &tmp_mem, ctx);
 	if (unlikely(r)) {
 		pr_err("Failed to find GTT space for blit to VRAM\n");
@@ -676,7 +668,7 @@ static int amdgpu_bo_move(struct ttm_buffer_object *bo, bool evict,
 	adev = amdgpu_ttm_adev(bo->bdev);
 
 	if (old_mem->mem_type == TTM_PL_SYSTEM && bo->ttm == NULL) {
-		amdgpu_move_null(bo, new_mem);
+		ttm_bo_move_null(bo, new_mem);
 		return 0;
 	}
 	if ((old_mem->mem_type == TTM_PL_TT &&
@@ -684,7 +676,7 @@ static int amdgpu_bo_move(struct ttm_buffer_object *bo, bool evict,
 	    (old_mem->mem_type == TTM_PL_SYSTEM &&
 	     new_mem->mem_type == TTM_PL_TT)) {
 		/* bind is enough */
-		amdgpu_move_null(bo, new_mem);
+		ttm_bo_move_null(bo, new_mem);
 		return 0;
 	}
 	if (old_mem->mem_type == AMDGPU_PL_GDS ||
@@ -694,7 +686,7 @@ static int amdgpu_bo_move(struct ttm_buffer_object *bo, bool evict,
 	    new_mem->mem_type == AMDGPU_PL_GWS ||
 	    new_mem->mem_type == AMDGPU_PL_OA) {
 		/* Nothing to save here */
-		amdgpu_move_null(bo, new_mem);
+		ttm_bo_move_null(bo, new_mem);
 		return 0;
 	}
 
@@ -773,7 +765,7 @@ static int amdgpu_ttm_io_mem_reserve(struct ttm_bo_device *bdev, struct ttm_reso
 			mem->bus.addr = (u8 *)adev->mman.aper_base_kaddr +
 					mem->bus.offset;
 
-		mem->bus.base = adev->gmc.aper_base;
+		mem->bus.offset += adev->gmc.aper_base;
 		mem->bus.is_iomem = true;
 		break;
 	default:
@@ -785,12 +777,13 @@ static int amdgpu_ttm_io_mem_reserve(struct ttm_bo_device *bdev, struct ttm_reso
 static unsigned long amdgpu_ttm_io_mem_pfn(struct ttm_buffer_object *bo,
 					   unsigned long page_offset)
 {
+	struct amdgpu_device *adev = amdgpu_ttm_adev(bo->bdev);
 	uint64_t offset = (page_offset << PAGE_SHIFT);
 	struct drm_mm_node *mm;
 
 	mm = amdgpu_find_mm_node(&bo->mem, &offset);
-	return (bo->mem.bus.base >> PAGE_SHIFT) + mm->start +
-		(offset >> PAGE_SHIFT);
+	offset += adev->gmc.aper_base;
+	return mm->start + (offset >> PAGE_SHIFT);
 }
 
 /**
@@ -824,6 +817,7 @@ struct amdgpu_ttm_tt {
 	uint64_t		userptr;
 	struct task_struct	*usertask;
 	uint32_t		userflags;
+	bool			bound;
 #if IS_ENABLED(CONFIG_DRM_AMDGPU_USERPTR)
 	struct hmm_range	*range;
 #endif
@@ -991,9 +985,10 @@ void amdgpu_ttm_tt_set_user_pages(struct ttm_tt *ttm, struct page **pages)
  *
  * Called by amdgpu_ttm_backend_bind()
  **/
-static int amdgpu_ttm_tt_pin_userptr(struct ttm_tt *ttm)
+static int amdgpu_ttm_tt_pin_userptr(struct ttm_bo_device *bdev,
+				     struct ttm_tt *ttm)
 {
-	struct amdgpu_device *adev = amdgpu_ttm_adev(ttm->bdev);
+	struct amdgpu_device *adev = amdgpu_ttm_adev(bdev);
 	struct amdgpu_ttm_tt *gtt = (void *)ttm;
 	int r;
 
@@ -1028,9 +1023,10 @@ release_sg:
 /**
  * amdgpu_ttm_tt_unpin_userptr - Unpin and unmap userptr pages
  */
-static void amdgpu_ttm_tt_unpin_userptr(struct ttm_tt *ttm)
+static void amdgpu_ttm_tt_unpin_userptr(struct ttm_bo_device *bdev,
+					struct ttm_tt *ttm)
 {
-	struct amdgpu_device *adev = amdgpu_ttm_adev(ttm->bdev);
+	struct amdgpu_device *adev = amdgpu_ttm_adev(bdev);
 	struct amdgpu_ttm_tt *gtt = (void *)ttm;
 
 	int write = !(gtt->userflags & AMDGPU_GEM_USERPTR_READONLY);
@@ -1111,16 +1107,23 @@ gart_bind_fail:
  * Called by ttm_tt_bind() on behalf of ttm_bo_handle_move_mem().
  * This handles binding GTT memory to the device address space.
  */
-static int amdgpu_ttm_backend_bind(struct ttm_tt *ttm,
+static int amdgpu_ttm_backend_bind(struct ttm_bo_device *bdev,
+				   struct ttm_tt *ttm,
 				   struct ttm_resource *bo_mem)
 {
-	struct amdgpu_device *adev = amdgpu_ttm_adev(ttm->bdev);
+	struct amdgpu_device *adev = amdgpu_ttm_adev(bdev);
 	struct amdgpu_ttm_tt *gtt = (void*)ttm;
 	uint64_t flags;
 	int r = 0;
 
+	if (!bo_mem)
+		return -EINVAL;
+
+	if (gtt->bound)
+		return 0;
+
 	if (gtt->userptr) {
-		r = amdgpu_ttm_tt_pin_userptr(ttm);
+		r = amdgpu_ttm_tt_pin_userptr(bdev, ttm);
 		if (r) {
 			DRM_ERROR("failed to pin userptr\n");
 			return r;
@@ -1152,6 +1155,7 @@ static int amdgpu_ttm_backend_bind(struct ttm_tt *ttm,
 	if (r)
 		DRM_ERROR("failed to bind %lu pages at 0x%08llX\n",
 			  ttm->num_pages, gtt->offset);
+	gtt->bound = true;
 	return r;
 }
 
@@ -1191,8 +1195,8 @@ int amdgpu_ttm_alloc_gart(struct ttm_buffer_object *bo)
 		placement.busy_placement = &placements;
 		placements.fpfn = 0;
 		placements.lpfn = adev->gmc.gart_size >> PAGE_SHIFT;
-		placements.flags = (bo->mem.placement & ~TTM_PL_MASK_MEM) |
-			TTM_PL_FLAG_TT;
+		placements.mem_type = TTM_PL_TT;
+		placements.flags = bo->mem.placement;
 
 		r = ttm_bo_mem_space(bo, &placement, &tmp, &ctx);
 		if (unlikely(r))
@@ -1243,15 +1247,19 @@ int amdgpu_ttm_recover_gart(struct ttm_buffer_object *tbo)
  * Called by ttm_tt_unbind() on behalf of ttm_bo_move_ttm() and
  * ttm_tt_destroy().
  */
-static void amdgpu_ttm_backend_unbind(struct ttm_tt *ttm)
+static void amdgpu_ttm_backend_unbind(struct ttm_bo_device *bdev,
+				      struct ttm_tt *ttm)
 {
-	struct amdgpu_device *adev = amdgpu_ttm_adev(ttm->bdev);
+	struct amdgpu_device *adev = amdgpu_ttm_adev(bdev);
 	struct amdgpu_ttm_tt *gtt = (void *)ttm;
 	int r;
 
+	if (!gtt->bound)
+		return;
+
 	/* if the pages have userptr pinning then clear that first */
 	if (gtt->userptr)
-		amdgpu_ttm_tt_unpin_userptr(ttm);
+		amdgpu_ttm_tt_unpin_userptr(bdev, ttm);
 
 	if (gtt->offset == AMDGPU_BO_INVALID_OFFSET)
 		return;
@@ -1261,12 +1269,16 @@ static void amdgpu_ttm_backend_unbind(struct ttm_tt *ttm)
 	if (r)
 		DRM_ERROR("failed to unbind %lu pages at 0x%08llX\n",
 			  gtt->ttm.ttm.num_pages, gtt->offset);
+	gtt->bound = false;
 }
 
-static void amdgpu_ttm_backend_destroy(struct ttm_tt *ttm)
+static void amdgpu_ttm_backend_destroy(struct ttm_bo_device *bdev,
+				       struct ttm_tt *ttm)
 {
 	struct amdgpu_ttm_tt *gtt = (void *)ttm;
 
+	amdgpu_ttm_backend_unbind(bdev, ttm);
+	ttm_tt_destroy_common(bdev, ttm);
 	if (gtt->usertask)
 		put_task_struct(gtt->usertask);
 
@@ -1274,12 +1286,6 @@ static void amdgpu_ttm_backend_destroy(struct ttm_tt *ttm)
 	kfree(gtt);
 }
 
-static struct ttm_backend_func amdgpu_backend_func = {
-	.bind = &amdgpu_ttm_backend_bind,
-	.unbind = &amdgpu_ttm_backend_unbind,
-	.destroy = &amdgpu_ttm_backend_destroy,
-};
-
 /**
  * amdgpu_ttm_tt_create - Create a ttm_tt object for a given BO
  *
@@ -1296,7 +1302,6 @@ static struct ttm_tt *amdgpu_ttm_tt_create(struct ttm_buffer_object *bo,
 	if (gtt == NULL) {
 		return NULL;
 	}
-	gtt->ttm.ttm.func = &amdgpu_backend_func;
 	gtt->gobj = &bo->base;
 
 	/* allocate space for the uninitialized page entries */
@@ -1313,10 +1318,11 @@ static struct ttm_tt *amdgpu_ttm_tt_create(struct ttm_buffer_object *bo,
  * Map the pages of a ttm_tt object to an address space visible
  * to the underlying device.
  */
-static int amdgpu_ttm_tt_populate(struct ttm_tt *ttm,
-			struct ttm_operation_ctx *ctx)
+static int amdgpu_ttm_tt_populate(struct ttm_bo_device *bdev,
+				  struct ttm_tt *ttm,
+				  struct ttm_operation_ctx *ctx)
 {
-	struct amdgpu_device *adev = amdgpu_ttm_adev(ttm->bdev);
+	struct amdgpu_device *adev = amdgpu_ttm_adev(bdev);
 	struct amdgpu_ttm_tt *gtt = (void *)ttm;
 
 	/* user pages are bound by amdgpu_ttm_tt_pin_userptr() */
@@ -1326,7 +1332,7 @@ static int amdgpu_ttm_tt_populate(struct ttm_tt *ttm,
 			return -ENOMEM;
 
 		ttm->page_flags |= TTM_PAGE_FLAG_SG;
-		ttm->state = tt_unbound;
+		ttm_tt_set_populated(ttm);
 		return 0;
 	}
 
@@ -1346,7 +1352,7 @@ static int amdgpu_ttm_tt_populate(struct ttm_tt *ttm,
 		drm_prime_sg_to_page_addr_arrays(ttm->sg, ttm->pages,
 						 gtt->ttm.dma_address,
 						 ttm->num_pages);
-		ttm->state = tt_unbound;
+		ttm_tt_set_populated(ttm);
 		return 0;
 	}
 
@@ -1367,7 +1373,7 @@ static int amdgpu_ttm_tt_populate(struct ttm_tt *ttm,
  * Unmaps pages of a ttm_tt object from the device address space and
  * unpopulates the page array backing it.
  */
-static void amdgpu_ttm_tt_unpopulate(struct ttm_tt *ttm)
+static void amdgpu_ttm_tt_unpopulate(struct ttm_bo_device *bdev, struct ttm_tt *ttm)
 {
 	struct amdgpu_ttm_tt *gtt = (void *)ttm;
 	struct amdgpu_device *adev;
@@ -1391,7 +1397,7 @@ static void amdgpu_ttm_tt_unpopulate(struct ttm_tt *ttm)
 	if (ttm->page_flags & TTM_PAGE_FLAG_SG)
 		return;
 
-	adev = amdgpu_ttm_adev(ttm->bdev);
+	adev = amdgpu_ttm_adev(bdev);
 
 #ifdef CONFIG_SWIOTLB
 	if (adev->need_swiotlb && swiotlb_nr_tbl()) {
@@ -1697,6 +1703,9 @@ static struct ttm_bo_driver amdgpu_bo_driver = {
 	.ttm_tt_create = &amdgpu_ttm_tt_create,
 	.ttm_tt_populate = &amdgpu_ttm_tt_populate,
 	.ttm_tt_unpopulate = &amdgpu_ttm_tt_unpopulate,
+	.ttm_tt_bind = &amdgpu_ttm_backend_bind,
+	.ttm_tt_unbind = &amdgpu_ttm_backend_unbind,
+	.ttm_tt_destroy = &amdgpu_ttm_backend_destroy,
 	.eviction_valuable = amdgpu_ttm_bo_eviction_valuable,
 	.evict_flags = &amdgpu_evict_flags,
 	.move = &amdgpu_bo_move,
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.h
index 68c5e3662b87..a87951b2f06d 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.h
@@ -32,10 +32,6 @@
 #define AMDGPU_PL_GWS		(TTM_PL_PRIV + 1)
 #define AMDGPU_PL_OA		(TTM_PL_PRIV + 2)
 
-#define AMDGPU_PL_FLAG_GDS		(TTM_PL_FLAG_PRIV << 0)
-#define AMDGPU_PL_FLAG_GWS		(TTM_PL_FLAG_PRIV << 1)
-#define AMDGPU_PL_FLAG_OA		(TTM_PL_FLAG_PRIV << 2)
-
 #define AMDGPU_GTT_MAX_TRANSFER_SIZE	512
 #define AMDGPU_GTT_NUM_TRANSFER_WINDOWS	2
 
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vram_mgr.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vram_mgr.c
index b2adc2abc581..01c1171afbe0 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vram_mgr.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vram_mgr.c
@@ -179,9 +179,6 @@ int amdgpu_vram_mgr_init(struct amdgpu_device *adev)
 	struct ttm_resource_manager *man = &mgr->manager;
 	int ret;
 
-	man->available_caching = TTM_PL_FLAG_UNCACHED | TTM_PL_FLAG_WC;
-	man->default_caching = TTM_PL_FLAG_WC;
-
 	ttm_resource_manager_init(man, adev->gmc.real_vram_size >> PAGE_SHIFT);
 
 	man->func = &amdgpu_vram_mgr_func;
diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c
index 57738164625b..bb1bc7f5d149 100644
--- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c
+++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c
@@ -7494,6 +7494,7 @@ static void amdgpu_dm_atomic_commit_tail(struct drm_atomic_state *state)
 	bool mode_set_reset_required = false;
 
 	drm_atomic_helper_update_legacy_modeset_state(dev, state);
+	drm_atomic_helper_calc_timestamping_constants(state);
 
 	dm_state = dm_atomic_get_new_state(state);
 	if (dm_state && dm_state->context) {
diff --git a/drivers/gpu/drm/armada/armada_crtc.c b/drivers/gpu/drm/armada/armada_crtc.c
index 38dfaa46d306..a887b6a5f8bd 100644
--- a/drivers/gpu/drm/armada/armada_crtc.c
+++ b/drivers/gpu/drm/armada/armada_crtc.c
@@ -757,7 +757,7 @@ static int armada_drm_crtc_cursor_move(struct drm_crtc *crtc, int x, int y)
 static void armada_drm_crtc_destroy(struct drm_crtc *crtc)
 {
 	struct armada_crtc *dcrtc = drm_to_armada_crtc(crtc);
-	struct armada_private *priv = crtc->dev->dev_private;
+	struct armada_private *priv = drm_to_armada_dev(crtc->dev);
 
 	if (dcrtc->cursor_obj)
 		drm_gem_object_put(&dcrtc->cursor_obj->obj);
@@ -901,7 +901,7 @@ static int armada_drm_crtc_create(struct drm_device *drm, struct device *dev,
 	struct resource *res, int irq, const struct armada_variant *variant,
 	struct device_node *port)
 {
-	struct armada_private *priv = drm->dev_private;
+	struct armada_private *priv = drm_to_armada_dev(drm);
 	struct armada_crtc *dcrtc;
 	struct drm_plane *primary;
 	void __iomem *base;
diff --git a/drivers/gpu/drm/armada/armada_debugfs.c b/drivers/gpu/drm/armada/armada_debugfs.c
index c6fc2f1d58e9..29f4b52e3c8d 100644
--- a/drivers/gpu/drm/armada/armada_debugfs.c
+++ b/drivers/gpu/drm/armada/armada_debugfs.c
@@ -19,7 +19,7 @@ static int armada_debugfs_gem_linear_show(struct seq_file *m, void *data)
 {
 	struct drm_info_node *node = m->private;
 	struct drm_device *dev = node->minor->dev;
-	struct armada_private *priv = dev->dev_private;
+	struct armada_private *priv = drm_to_armada_dev(dev);
 	struct drm_printer p = drm_seq_file_printer(m);
 
 	mutex_lock(&priv->linear_lock);
diff --git a/drivers/gpu/drm/armada/armada_drm.h b/drivers/gpu/drm/armada/armada_drm.h
index a11bdaccbb33..6a5a87932576 100644
--- a/drivers/gpu/drm/armada/armada_drm.h
+++ b/drivers/gpu/drm/armada/armada_drm.h
@@ -73,6 +73,8 @@ struct armada_private {
 #endif
 };
 
+#define drm_to_armada_dev(dev) container_of(dev, struct armada_private, drm)
+
 int armada_fbdev_init(struct drm_device *);
 void armada_fbdev_fini(struct drm_device *);
 
diff --git a/drivers/gpu/drm/armada/armada_drv.c b/drivers/gpu/drm/armada/armada_drv.c
index 5fc25c3f445c..980d3f1f8f16 100644
--- a/drivers/gpu/drm/armada/armada_drv.c
+++ b/drivers/gpu/drm/armada/armada_drv.c
@@ -87,24 +87,13 @@ static int armada_drm_bind(struct device *dev)
 				     "armada-drm"))
 		return -EBUSY;
 
-	priv = kzalloc(sizeof(*priv), GFP_KERNEL);
-	if (!priv)
-		return -ENOMEM;
-
-	/*
-	 * The drm_device structure must be at the start of
-	 * armada_private for drm_dev_put() to work correctly.
-	 */
-	BUILD_BUG_ON(offsetof(struct armada_private, drm) != 0);
-
-	ret = drm_dev_init(&priv->drm, &armada_drm_driver, dev);
-	if (ret) {
-		dev_err(dev, "[" DRM_NAME ":%s] drm_dev_init failed: %d\n",
-			__func__, ret);
-		kfree(priv);
-		return ret;
+	priv = devm_drm_dev_alloc(dev, &armada_drm_driver,
+				  struct armada_private, drm);
+	if (IS_ERR(priv)) {
+		dev_err(dev, "[" DRM_NAME ":%s] devm_drm_dev_alloc failed: %li\n",
+			__func__, PTR_ERR(priv));
+		return PTR_ERR(priv);
 	}
-	drmm_add_final_kfree(&priv->drm, priv);
 
 	/* Remove early framebuffers */
 	ret = drm_fb_helper_remove_conflicting_framebuffers(NULL,
@@ -117,8 +106,6 @@ static int armada_drm_bind(struct device *dev)
 		return ret;
 	}
 
-	priv->drm.dev_private = priv;
-
 	dev_set_drvdata(dev, &priv->drm);
 
 	/* Mode setting support */
@@ -174,14 +161,13 @@ static int armada_drm_bind(struct device *dev)
  err_kms:
 	drm_mode_config_cleanup(&priv->drm);
 	drm_mm_takedown(&priv->linear);
-	drm_dev_put(&priv->drm);
 	return ret;
 }
 
 static void armada_drm_unbind(struct device *dev)
 {
 	struct drm_device *drm = dev_get_drvdata(dev);
-	struct armada_private *priv = drm->dev_private;
+	struct armada_private *priv = drm_to_armada_dev(drm);
 
 	drm_kms_helper_poll_fini(&priv->drm);
 	armada_fbdev_fini(&priv->drm);
@@ -194,8 +180,6 @@ static void armada_drm_unbind(struct device *dev)
 
 	drm_mode_config_cleanup(&priv->drm);
 	drm_mm_takedown(&priv->linear);
-
-	drm_dev_put(&priv->drm);
 }
 
 static int compare_of(struct device *dev, void *data)
diff --git a/drivers/gpu/drm/armada/armada_fbdev.c b/drivers/gpu/drm/armada/armada_fbdev.c
index 0c4601275507..38f5170c0fea 100644
--- a/drivers/gpu/drm/armada/armada_fbdev.c
+++ b/drivers/gpu/drm/armada/armada_fbdev.c
@@ -117,7 +117,7 @@ static const struct drm_fb_helper_funcs armada_fb_helper_funcs = {
 
 int armada_fbdev_init(struct drm_device *dev)
 {
-	struct armada_private *priv = dev->dev_private;
+	struct armada_private *priv = drm_to_armada_dev(dev);
 	struct drm_fb_helper *fbh;
 	int ret;
 
@@ -151,7 +151,7 @@ int armada_fbdev_init(struct drm_device *dev)
 
 void armada_fbdev_fini(struct drm_device *dev)
 {
-	struct armada_private *priv = dev->dev_private;
+	struct armada_private *priv = drm_to_armada_dev(dev);
 	struct drm_fb_helper *fbh = priv->fbdev;
 
 	if (fbh) {
diff --git a/drivers/gpu/drm/armada/armada_gem.c b/drivers/gpu/drm/armada/armada_gem.c
index a63008ce284d..6654bccd9466 100644
--- a/drivers/gpu/drm/armada/armada_gem.c
+++ b/drivers/gpu/drm/armada/armada_gem.c
@@ -39,7 +39,7 @@ static size_t roundup_gem_size(size_t size)
 void armada_gem_free_object(struct drm_gem_object *obj)
 {
 	struct armada_gem_object *dobj = drm_to_armada_gem(obj);
-	struct armada_private *priv = obj->dev->dev_private;
+	struct armada_private *priv = drm_to_armada_dev(obj->dev);
 
 	DRM_DEBUG_DRIVER("release obj %p\n", dobj);
 
@@ -77,7 +77,7 @@ void armada_gem_free_object(struct drm_gem_object *obj)
 int
 armada_gem_linear_back(struct drm_device *dev, struct armada_gem_object *obj)
 {
-	struct armada_private *priv = dev->dev_private;
+	struct armada_private *priv = drm_to_armada_dev(dev);
 	size_t size = obj->obj.size;
 
 	if (obj->page || obj->linear)
diff --git a/drivers/gpu/drm/armada/armada_overlay.c b/drivers/gpu/drm/armada/armada_overlay.c
index 07f0da4d9ba1..30e01101f59e 100644
--- a/drivers/gpu/drm/armada/armada_overlay.c
+++ b/drivers/gpu/drm/armada/armada_overlay.c
@@ -344,7 +344,7 @@ static int armada_overlay_set_property(struct drm_plane *plane,
 	struct drm_plane_state *state, struct drm_property *property,
 	uint64_t val)
 {
-	struct armada_private *priv = plane->dev->dev_private;
+	struct armada_private *priv = drm_to_armada_dev(plane->dev);
 
 #define K2R(val) (((val) >> 0) & 0xff)
 #define K2G(val) (((val) >> 8) & 0xff)
@@ -412,7 +412,7 @@ static int armada_overlay_get_property(struct drm_plane *plane,
 	const struct drm_plane_state *state, struct drm_property *property,
 	uint64_t *val)
 {
-	struct armada_private *priv = plane->dev->dev_private;
+	struct armada_private *priv = drm_to_armada_dev(plane->dev);
 
 #define C2K(c,s)	(((c) >> (s)) & 0xff)
 #define R2BGR(r,g,b,s)	(C2K(r,s) << 0 | C2K(g,s) << 8 | C2K(b,s) << 16)
@@ -505,7 +505,7 @@ static const struct drm_prop_enum_list armada_drm_colorkey_enum_list[] = {
 
 static int armada_overlay_create_properties(struct drm_device *dev)
 {
-	struct armada_private *priv = dev->dev_private;
+	struct armada_private *priv = drm_to_armada_dev(dev);
 
 	if (priv->colorkey_prop)
 		return 0;
@@ -539,7 +539,7 @@ static int armada_overlay_create_properties(struct drm_device *dev)
 
 int armada_overlay_plane_create(struct drm_device *dev, unsigned long crtcs)
 {
-	struct armada_private *priv = dev->dev_private;
+	struct armada_private *priv = drm_to_armada_dev(dev);
 	struct drm_mode_object *mobj;
 	struct drm_plane *overlay;
 	int ret;
diff --git a/drivers/gpu/drm/aspeed/aspeed_gfx_drv.c b/drivers/gpu/drm/aspeed/aspeed_gfx_drv.c
index 903f4f304647..2b424b2b85cc 100644
--- a/drivers/gpu/drm/aspeed/aspeed_gfx_drv.c
+++ b/drivers/gpu/drm/aspeed/aspeed_gfx_drv.c
@@ -63,15 +63,21 @@ static const struct drm_mode_config_funcs aspeed_gfx_mode_config_funcs = {
 	.atomic_commit		= drm_atomic_helper_commit,
 };
 
-static void aspeed_gfx_setup_mode_config(struct drm_device *drm)
+static int aspeed_gfx_setup_mode_config(struct drm_device *drm)
 {
-	drm_mode_config_init(drm);
+	int ret;
+
+	ret = drmm_mode_config_init(drm);
+	if (ret)
+		return ret;
 
 	drm->mode_config.min_width = 0;
 	drm->mode_config.min_height = 0;
 	drm->mode_config.max_width = 800;
 	drm->mode_config.max_height = 600;
 	drm->mode_config.funcs = &aspeed_gfx_mode_config_funcs;
+
+	return ret;
 }
 
 static irqreturn_t aspeed_gfx_irq_handler(int irq, void *data)
@@ -144,7 +150,9 @@ static int aspeed_gfx_load(struct drm_device *drm)
 	writel(0, priv->base + CRT_CTRL1);
 	writel(0, priv->base + CRT_CTRL2);
 
-	aspeed_gfx_setup_mode_config(drm);
+	ret = aspeed_gfx_setup_mode_config(drm);
+	if (ret < 0)
+		return ret;
 
 	ret = drm_vblank_init(drm, 1);
 	if (ret < 0) {
@@ -179,7 +187,6 @@ static int aspeed_gfx_load(struct drm_device *drm)
 static void aspeed_gfx_unload(struct drm_device *drm)
 {
 	drm_kms_helper_poll_fini(drm);
-	drm_mode_config_cleanup(drm);
 }
 
 DEFINE_DRM_GEM_CMA_FOPS(fops);
diff --git a/drivers/gpu/drm/ast/ast_drv.h b/drivers/gpu/drm/ast/ast_drv.h
index c1af6b725933..467049ca8430 100644
--- a/drivers/gpu/drm/ast/ast_drv.h
+++ b/drivers/gpu/drm/ast/ast_drv.h
@@ -177,6 +177,8 @@ struct ast_private *ast_device_create(struct drm_driver *drv,
 
 #define AST_IO_MM_OFFSET		(0x380)
 
+#define AST_IO_VGAIR1_VREFRESH		BIT(3)
+
 #define __ast_read(x) \
 static inline u##x ast_read##x(struct ast_private *ast, u32 reg) { \
 u##x val = 0;\
diff --git a/drivers/gpu/drm/ast/ast_mode.c b/drivers/gpu/drm/ast/ast_mode.c
index 62fe682a7de6..834a156e3a75 100644
--- a/drivers/gpu/drm/ast/ast_mode.c
+++ b/drivers/gpu/drm/ast/ast_mode.c
@@ -514,6 +514,16 @@ static void ast_set_start_address_crt1(struct ast_private *ast,
 
 }
 
+static void ast_wait_for_vretrace(struct ast_private *ast)
+{
+	unsigned long timeout = jiffies + HZ;
+	u8 vgair1;
+
+	do {
+		vgair1 = ast_io_read8(ast, AST_IO_INPUT_STATUS1_READ);
+	} while (!(vgair1 & AST_IO_VGAIR1_VREFRESH) && time_before(jiffies, timeout));
+}
+
 /*
  * Primary plane
  */
@@ -562,13 +572,24 @@ ast_primary_plane_helper_atomic_update(struct drm_plane *plane,
 	struct drm_plane_state *state = plane->state;
 	struct drm_gem_vram_object *gbo;
 	s64 gpu_addr;
+	struct drm_framebuffer *fb = state->fb;
+	struct drm_framebuffer *old_fb = old_state->fb;
 
-	gbo = drm_gem_vram_of_gem(state->fb->obj[0]);
+	if (!old_fb || (fb->format != old_fb->format)) {
+		struct drm_crtc_state *crtc_state = state->crtc->state;
+		struct ast_crtc_state *ast_crtc_state = to_ast_crtc_state(crtc_state);
+		struct ast_vbios_mode_info *vbios_mode_info = &ast_crtc_state->vbios_mode_info;
+
+		ast_set_color_reg(ast, fb->format);
+		ast_set_vbios_color_reg(ast, fb->format, vbios_mode_info);
+	}
+
+	gbo = drm_gem_vram_of_gem(fb->obj[0]);
 	gpu_addr = drm_gem_vram_offset(gbo);
 	if (drm_WARN_ON_ONCE(dev, gpu_addr < 0))
 		return; /* Bug: we didn't pin the BO to VRAM in prepare_fb. */
 
-	ast_set_offset_reg(ast, state->fb);
+	ast_set_offset_reg(ast, fb);
 	ast_set_start_address_crt1(ast, (u32)gpu_addr);
 
 	ast_set_index_reg_mask(ast, AST_IO_SEQ_PORT, 0x1, 0xdf, 0x00);
@@ -733,6 +754,7 @@ static void ast_crtc_dpms(struct drm_crtc *crtc, int mode)
 static int ast_crtc_helper_atomic_check(struct drm_crtc *crtc,
 					struct drm_crtc_state *state)
 {
+	struct drm_device *dev = crtc->dev;
 	struct ast_crtc_state *ast_state;
 	const struct drm_format_info *format;
 	bool succ;
@@ -743,8 +765,8 @@ static int ast_crtc_helper_atomic_check(struct drm_crtc *crtc,
 	ast_state = to_ast_crtc_state(state);
 
 	format = ast_state->format;
-	if (!format)
-		return 0;
+	if (drm_WARN_ON_ONCE(dev, !format))
+		return -EINVAL; /* BUG: We didn't set format in primary check(). */
 
 	succ = ast_get_vbios_mode_info(format, &state->mode,
 				       &state->adjusted_mode,
@@ -755,39 +777,17 @@ static int ast_crtc_helper_atomic_check(struct drm_crtc *crtc,
 	return 0;
 }
 
-static void ast_crtc_helper_atomic_begin(struct drm_crtc *crtc,
-					 struct drm_crtc_state *old_crtc_state)
-{
-	struct ast_private *ast = to_ast_private(crtc->dev);
-
-	ast_open_key(ast);
-}
-
-static void ast_crtc_helper_atomic_flush(struct drm_crtc *crtc,
-					 struct drm_crtc_state *old_crtc_state)
+static void
+ast_crtc_helper_atomic_enable(struct drm_crtc *crtc,
+			      struct drm_crtc_state *old_crtc_state)
 {
 	struct drm_device *dev = crtc->dev;
 	struct ast_private *ast = to_ast_private(dev);
-	struct ast_crtc_state *ast_state;
-	const struct drm_format_info *format;
-	struct ast_vbios_mode_info *vbios_mode_info;
-	struct drm_display_mode *adjusted_mode;
-
-	ast_state = to_ast_crtc_state(crtc->state);
-
-	format = ast_state->format;
-	if (!format)
-		return;
-
-	vbios_mode_info = &ast_state->vbios_mode_info;
-
-	ast_set_color_reg(ast, format);
-	ast_set_vbios_color_reg(ast, format, vbios_mode_info);
-
-	if (!crtc->state->mode_changed)
-		return;
-
-	adjusted_mode = &crtc->state->adjusted_mode;
+	struct drm_crtc_state *crtc_state = crtc->state;
+	struct ast_crtc_state *ast_crtc_state = to_ast_crtc_state(crtc_state);
+	struct ast_vbios_mode_info *vbios_mode_info =
+		&ast_crtc_state->vbios_mode_info;
+	struct drm_display_mode *adjusted_mode = &crtc_state->adjusted_mode;
 
 	ast_set_vbios_mode_reg(ast, adjusted_mode, vbios_mode_info);
 	ast_set_index_reg(ast, AST_IO_CRTC_PORT, 0xa1, 0x06);
@@ -796,12 +796,7 @@ static void ast_crtc_helper_atomic_flush(struct drm_crtc *crtc,
 	ast_set_dclk_reg(ast, adjusted_mode, vbios_mode_info);
 	ast_set_crtthd_reg(ast);
 	ast_set_sync_reg(ast, adjusted_mode, vbios_mode_info);
-}
 
-static void
-ast_crtc_helper_atomic_enable(struct drm_crtc *crtc,
-			      struct drm_crtc_state *old_crtc_state)
-{
 	ast_crtc_dpms(crtc, DRM_MODE_DPMS_ON);
 }
 
@@ -809,13 +804,32 @@ static void
 ast_crtc_helper_atomic_disable(struct drm_crtc *crtc,
 			       struct drm_crtc_state *old_crtc_state)
 {
+	struct drm_device *dev = crtc->dev;
+	struct ast_private *ast = to_ast_private(dev);
+
 	ast_crtc_dpms(crtc, DRM_MODE_DPMS_OFF);
+
+	/*
+	 * HW cursors require the underlying primary plane and CRTC to
+	 * display a valid mode and image. This is not the case during
+	 * full modeset operations. So we temporarily disable any active
+	 * plane, including the HW cursor. Each plane's atomic_update()
+	 * helper will re-enable it if necessary.
+	 *
+	 * We only do this during *full* modesets. It does not affect
+	 * simple pageflips on the planes.
+	 */
+	drm_atomic_helper_disable_planes_on_crtc(old_crtc_state, false);
+
+	/*
+	 * Ensure that no scanout takes place before reprogramming mode
+	 * and format registers.
+	 */
+	ast_wait_for_vretrace(ast);
 }
 
 static const struct drm_crtc_helper_funcs ast_crtc_helper_funcs = {
 	.atomic_check = ast_crtc_helper_atomic_check,
-	.atomic_begin = ast_crtc_helper_atomic_begin,
-	.atomic_flush = ast_crtc_helper_atomic_flush,
 	.atomic_enable = ast_crtc_helper_atomic_enable,
 	.atomic_disable = ast_crtc_helper_atomic_disable,
 };
@@ -1054,6 +1068,11 @@ static int ast_connector_init(struct drm_device *dev)
  * Mode config
  */
 
+static const struct drm_mode_config_helper_funcs
+ast_mode_config_helper_funcs = {
+	.atomic_commit_tail = drm_atomic_helper_commit_tail_rpm,
+};
+
 static const struct drm_mode_config_funcs ast_mode_config_funcs = {
 	.fb_create = drm_gem_fb_create,
 	.mode_valid = drm_vram_helper_mode_valid,
@@ -1093,6 +1112,8 @@ int ast_mode_config_init(struct ast_private *ast)
 		dev->mode_config.max_height = 1200;
 	}
 
+	dev->mode_config.helper_private = &ast_mode_config_helper_funcs;
+
 	memset(&ast->primary_plane, 0, sizeof(ast->primary_plane));
 	ret = drm_universal_plane_init(dev, &ast->primary_plane, 0x01,
 				       &ast_primary_plane_funcs,
diff --git a/drivers/gpu/drm/bridge/parade-ps8640.c b/drivers/gpu/drm/bridge/parade-ps8640.c
index 9f7b7a9c53c5..7bd0affa057a 100644
--- a/drivers/gpu/drm/bridge/parade-ps8640.c
+++ b/drivers/gpu/drm/bridge/parade-ps8640.c
@@ -65,6 +65,7 @@ struct ps8640 {
 	struct regulator_bulk_data supplies[2];
 	struct gpio_desc *gpio_reset;
 	struct gpio_desc *gpio_powerdown;
+	bool powered;
 };
 
 static inline struct ps8640 *bridge_to_ps8640(struct drm_bridge *e)
@@ -91,13 +92,15 @@ static int ps8640_bridge_vdo_control(struct ps8640 *ps_bridge,
 	return 0;
 }
 
-static void ps8640_pre_enable(struct drm_bridge *bridge)
+static void ps8640_bridge_poweron(struct ps8640 *ps_bridge)
 {
-	struct ps8640 *ps_bridge = bridge_to_ps8640(bridge);
 	struct i2c_client *client = ps_bridge->page[PAGE2_TOP_CNTL];
 	unsigned long timeout;
 	int ret, status;
 
+	if (ps_bridge->powered)
+		return;
+
 	ret = regulator_bulk_enable(ARRAY_SIZE(ps_bridge->supplies),
 				    ps_bridge->supplies);
 	if (ret < 0) {
@@ -152,10 +155,6 @@ static void ps8640_pre_enable(struct drm_bridge *bridge)
 		goto err_regulators_disable;
 	}
 
-	ret = ps8640_bridge_vdo_control(ps_bridge, ENABLE);
-	if (ret)
-		goto err_regulators_disable;
-
 	/* Switch access edp panel's edid through i2c */
 	ret = i2c_smbus_write_byte_data(client, PAGE2_I2C_BYPASS,
 					I2C_BYPASS_EN);
@@ -164,6 +163,8 @@ static void ps8640_pre_enable(struct drm_bridge *bridge)
 		goto err_regulators_disable;
 	}
 
+	ps_bridge->powered = true;
+
 	return;
 
 err_regulators_disable:
@@ -171,12 +172,12 @@ err_regulators_disable:
 			       ps_bridge->supplies);
 }
 
-static void ps8640_post_disable(struct drm_bridge *bridge)
+static void ps8640_bridge_poweroff(struct ps8640 *ps_bridge)
 {
-	struct ps8640 *ps_bridge = bridge_to_ps8640(bridge);
 	int ret;
 
-	ps8640_bridge_vdo_control(ps_bridge, DISABLE);
+	if (!ps_bridge->powered)
+		return;
 
 	gpiod_set_value(ps_bridge->gpio_reset, 1);
 	gpiod_set_value(ps_bridge->gpio_powerdown, 1);
@@ -184,6 +185,28 @@ static void ps8640_post_disable(struct drm_bridge *bridge)
 				     ps_bridge->supplies);
 	if (ret < 0)
 		DRM_ERROR("cannot disable regulators %d\n", ret);
+
+	ps_bridge->powered = false;
+}
+
+static void ps8640_pre_enable(struct drm_bridge *bridge)
+{
+	struct ps8640 *ps_bridge = bridge_to_ps8640(bridge);
+	int ret;
+
+	ps8640_bridge_poweron(ps_bridge);
+
+	ret = ps8640_bridge_vdo_control(ps_bridge, ENABLE);
+	if (ret < 0)
+		ps8640_bridge_poweroff(ps_bridge);
+}
+
+static void ps8640_post_disable(struct drm_bridge *bridge)
+{
+	struct ps8640 *ps_bridge = bridge_to_ps8640(bridge);
+
+	ps8640_bridge_vdo_control(ps_bridge, DISABLE);
+	ps8640_bridge_poweroff(ps_bridge);
 }
 
 static int ps8640_bridge_attach(struct drm_bridge *bridge,
@@ -249,9 +272,34 @@ static struct edid *ps8640_bridge_get_edid(struct drm_bridge *bridge,
 					   struct drm_connector *connector)
 {
 	struct ps8640 *ps_bridge = bridge_to_ps8640(bridge);
+	bool poweroff = !ps_bridge->powered;
+	struct edid *edid;
+
+	/*
+	 * When we end calling get_edid() triggered by an ioctl, i.e
+	 *
+	 *   drm_mode_getconnector (ioctl)
+	 *     -> drm_helper_probe_single_connector_modes
+	 *        -> drm_bridge_connector_get_modes
+	 *           -> ps8640_bridge_get_edid
+	 *
+	 * We need to make sure that what we need is enabled before reading
+	 * EDID, for this chip, we need to do a full poweron, otherwise it will
+	 * fail.
+	 */
+	drm_bridge_chain_pre_enable(bridge);
 
-	return drm_get_edid(connector,
+	edid = drm_get_edid(connector,
 			    ps_bridge->page[PAGE0_DP_CNTL]->adapter);
+
+	/*
+	 * If we call the get_edid() function without having enabled the chip
+	 * before, return the chip to its original power state.
+	 */
+	if (poweroff)
+		drm_bridge_chain_post_disable(bridge);
+
+	return edid;
 }
 
 static const struct drm_bridge_funcs ps8640_bridge_funcs = {
diff --git a/drivers/gpu/drm/bridge/synopsys/dw-mipi-dsi.c b/drivers/gpu/drm/bridge/synopsys/dw-mipi-dsi.c
index d580b2aa4ce9..6b268f9445b3 100644
--- a/drivers/gpu/drm/bridge/synopsys/dw-mipi-dsi.c
+++ b/drivers/gpu/drm/bridge/synopsys/dw-mipi-dsi.c
@@ -89,7 +89,9 @@
 #define VID_MODE_TYPE_NON_BURST_SYNC_EVENTS	0x1
 #define VID_MODE_TYPE_BURST			0x2
 #define VID_MODE_TYPE_MASK			0x3
+#define ENABLE_LOW_POWER_CMD		BIT(15)
 #define VID_MODE_VPG_ENABLE		BIT(16)
+#define VID_MODE_VPG_MODE		BIT(20)
 #define VID_MODE_VPG_HORIZONTAL		BIT(24)
 
 #define DSI_VID_PKT_SIZE		0x3c
@@ -220,6 +222,21 @@
 #define PHY_STATUS_TIMEOUT_US		10000
 #define CMD_PKT_STATUS_TIMEOUT_US	20000
 
+#ifdef CONFIG_DEBUG_FS
+#define VPG_DEFS(name, dsi) \
+	((void __force *)&((*dsi).vpg_defs.name))
+
+#define REGISTER(name, mask, dsi) \
+	{ #name, VPG_DEFS(name, dsi), mask, dsi }
+
+struct debugfs_entries {
+	const char				*name;
+	bool					*reg;
+	u32					mask;
+	struct dw_mipi_dsi			*dsi;
+};
+#endif /* CONFIG_DEBUG_FS */
+
 struct dw_mipi_dsi {
 	struct drm_bridge bridge;
 	struct mipi_dsi_host dsi_host;
@@ -237,9 +254,12 @@ struct dw_mipi_dsi {
 
 #ifdef CONFIG_DEBUG_FS
 	struct dentry *debugfs;
-
-	bool vpg;
-	bool vpg_horizontal;
+	struct debugfs_entries *debugfs_vpg;
+	struct {
+		bool vpg;
+		bool vpg_horizontal;
+		bool vpg_ber_pattern;
+	} vpg_defs;
 #endif /* CONFIG_DEBUG_FS */
 
 	struct dw_mipi_dsi *master; /* dual-dsi master ptr */
@@ -360,13 +380,28 @@ static void dw_mipi_message_config(struct dw_mipi_dsi *dsi,
 	bool lpm = msg->flags & MIPI_DSI_MSG_USE_LPM;
 	u32 val = 0;
 
+	/*
+	 * TODO dw drv improvements
+	 * largest packet sizes during hfp or during vsa/vpb/vfp
+	 * should be computed according to byte lane, lane number and only
+	 * if sending lp cmds in high speed is enable (PHY_TXREQUESTCLKHS)
+	 */
+	dsi_write(dsi, DSI_DPI_LP_CMD_TIM, OUTVACT_LPCMD_TIME(16)
+		  | INVACT_LPCMD_TIME(4));
+
 	if (msg->flags & MIPI_DSI_MSG_REQ_ACK)
 		val |= ACK_RQST_EN;
 	if (lpm)
 		val |= CMD_MODE_ALL_LP;
 
-	dsi_write(dsi, DSI_LPCLK_CTRL, lpm ? 0 : PHY_TXREQUESTCLKHS);
 	dsi_write(dsi, DSI_CMD_MODE_CFG, val);
+
+	val = dsi_read(dsi, DSI_VID_MODE_CFG);
+	if (lpm)
+		val |= ENABLE_LOW_POWER_CMD;
+	else
+		val &= ~ENABLE_LOW_POWER_CMD;
+	dsi_write(dsi, DSI_VID_MODE_CFG, val);
 }
 
 static int dw_mipi_dsi_gen_pkt_hdr_write(struct dw_mipi_dsi *dsi, u32 hdr_val)
@@ -529,9 +564,11 @@ static void dw_mipi_dsi_video_mode_config(struct dw_mipi_dsi *dsi)
 		val |= VID_MODE_TYPE_NON_BURST_SYNC_EVENTS;
 
 #ifdef CONFIG_DEBUG_FS
-	if (dsi->vpg) {
+	if (dsi->vpg_defs.vpg) {
 		val |= VID_MODE_VPG_ENABLE;
-		val |= dsi->vpg_horizontal ? VID_MODE_VPG_HORIZONTAL : 0;
+		val |= dsi->vpg_defs.vpg_horizontal ?
+		       VID_MODE_VPG_HORIZONTAL : 0;
+		val |= dsi->vpg_defs.vpg_ber_pattern ? VID_MODE_VPG_MODE : 0;
 	}
 #endif /* CONFIG_DEBUG_FS */
 
@@ -541,16 +578,22 @@ static void dw_mipi_dsi_video_mode_config(struct dw_mipi_dsi *dsi)
 static void dw_mipi_dsi_set_mode(struct dw_mipi_dsi *dsi,
 				 unsigned long mode_flags)
 {
+	u32 val;
+
 	dsi_write(dsi, DSI_PWR_UP, RESET);
 
 	if (mode_flags & MIPI_DSI_MODE_VIDEO) {
 		dsi_write(dsi, DSI_MODE_CFG, ENABLE_VIDEO_MODE);
 		dw_mipi_dsi_video_mode_config(dsi);
-		dsi_write(dsi, DSI_LPCLK_CTRL, PHY_TXREQUESTCLKHS);
 	} else {
 		dsi_write(dsi, DSI_MODE_CFG, ENABLE_CMD_MODE);
 	}
 
+	val = PHY_TXREQUESTCLKHS;
+	if (dsi->mode_flags & MIPI_DSI_CLOCK_NON_CONTINUOUS)
+		val |= AUTO_CLKLANE_CTRL;
+	dsi_write(dsi, DSI_LPCLK_CTRL, val);
+
 	dsi_write(dsi, DSI_PWR_UP, POWERUP);
 }
 
@@ -562,15 +605,30 @@ static void dw_mipi_dsi_disable(struct dw_mipi_dsi *dsi)
 
 static void dw_mipi_dsi_init(struct dw_mipi_dsi *dsi)
 {
+	const struct dw_mipi_dsi_phy_ops *phy_ops = dsi->plat_data->phy_ops;
+	unsigned int esc_rate; /* in MHz */
+	u32 esc_clk_division;
+	int ret;
+
 	/*
 	 * The maximum permitted escape clock is 20MHz and it is derived from
-	 * lanebyteclk, which is running at "lane_mbps / 8".  Thus we want:
-	 *
-	 *     (lane_mbps >> 3) / esc_clk_division < 20
+	 * lanebyteclk, which is running at "lane_mbps / 8".
+	 */
+	if (phy_ops->get_esc_clk_rate) {
+		ret = phy_ops->get_esc_clk_rate(dsi->plat_data->priv_data,
+						&esc_rate);
+		if (ret)
+			DRM_DEBUG_DRIVER("Phy get_esc_clk_rate() failed\n");
+	} else
+		esc_rate = 20; /* Default to 20MHz */
+
+	/*
+	 * We want :
+	 *     (lane_mbps >> 3) / esc_clk_division < X
 	 * which is:
-	 *     (lane_mbps >> 3) / 20 > esc_clk_division
+	 *     (lane_mbps >> 3) / X > esc_clk_division
 	 */
-	u32 esc_clk_division = (dsi->lane_mbps >> 3) / 20 + 1;
+	esc_clk_division = (dsi->lane_mbps >> 3) / esc_rate + 1;
 
 	dsi_write(dsi, DSI_PWR_UP, RESET);
 
@@ -611,14 +669,6 @@ static void dw_mipi_dsi_dpi_config(struct dw_mipi_dsi *dsi,
 	dsi_write(dsi, DSI_DPI_VCID, DPI_VCID(dsi->channel));
 	dsi_write(dsi, DSI_DPI_COLOR_CODING, color);
 	dsi_write(dsi, DSI_DPI_CFG_POL, val);
-	/*
-	 * TODO dw drv improvements
-	 * largest packet sizes during hfp or during vsa/vpb/vfp
-	 * should be computed according to byte lane, lane number and only
-	 * if sending lp cmds in high speed is enable (PHY_TXREQUESTCLKHS)
-	 */
-	dsi_write(dsi, DSI_DPI_LP_CMD_TIM, OUTVACT_LPCMD_TIME(4)
-		  | INVACT_LPCMD_TIME(4));
 }
 
 static void dw_mipi_dsi_packet_handler_config(struct dw_mipi_dsi *dsi)
@@ -964,6 +1014,66 @@ static const struct drm_bridge_funcs dw_mipi_dsi_bridge_funcs = {
 
 #ifdef CONFIG_DEBUG_FS
 
+static int dw_mipi_dsi_debugfs_write(void *data, u64 val)
+{
+	struct debugfs_entries *vpg = data;
+	struct dw_mipi_dsi *dsi;
+	u32 mode_cfg;
+
+	if (!vpg)
+		return -ENODEV;
+
+	dsi = vpg->dsi;
+
+	*vpg->reg = (bool)val;
+
+	mode_cfg = dsi_read(dsi, DSI_VID_MODE_CFG);
+
+	if (*vpg->reg)
+		mode_cfg |= vpg->mask;
+	else
+		mode_cfg &= ~vpg->mask;
+
+	dsi_write(dsi, DSI_VID_MODE_CFG, mode_cfg);
+
+	return 0;
+}
+
+static int dw_mipi_dsi_debugfs_show(void *data, u64 *val)
+{
+	struct debugfs_entries *vpg = data;
+
+	if (!vpg)
+		return -ENODEV;
+
+	*val = *vpg->reg;
+
+	return 0;
+}
+
+DEFINE_DEBUGFS_ATTRIBUTE(fops_x32, dw_mipi_dsi_debugfs_show,
+			 dw_mipi_dsi_debugfs_write, "%llu\n");
+
+static void debugfs_create_files(void *data)
+{
+	struct dw_mipi_dsi *dsi = data;
+	struct debugfs_entries debugfs[] = {
+		REGISTER(vpg, VID_MODE_VPG_ENABLE, dsi),
+		REGISTER(vpg_horizontal, VID_MODE_VPG_HORIZONTAL, dsi),
+		REGISTER(vpg_ber_pattern, VID_MODE_VPG_MODE, dsi),
+	};
+	int i;
+
+	dsi->debugfs_vpg = kmemdup(debugfs, sizeof(debugfs), GFP_KERNEL);
+	if (!dsi->debugfs_vpg)
+		return;
+
+	for (i = 0; i < ARRAY_SIZE(debugfs); i++)
+		debugfs_create_file(dsi->debugfs_vpg[i].name, 0644,
+				    dsi->debugfs, &dsi->debugfs_vpg[i],
+				    &fops_x32);
+}
+
 static void dw_mipi_dsi_debugfs_init(struct dw_mipi_dsi *dsi)
 {
 	dsi->debugfs = debugfs_create_dir(dev_name(dsi->dev), NULL);
@@ -972,14 +1082,13 @@ static void dw_mipi_dsi_debugfs_init(struct dw_mipi_dsi *dsi)
 		return;
 	}
 
-	debugfs_create_bool("vpg", 0660, dsi->debugfs, &dsi->vpg);
-	debugfs_create_bool("vpg_horizontal", 0660, dsi->debugfs,
-			    &dsi->vpg_horizontal);
+	debugfs_create_files(dsi);
 }
 
 static void dw_mipi_dsi_debugfs_remove(struct dw_mipi_dsi *dsi)
 {
 	debugfs_remove_recursive(dsi->debugfs);
+	kfree(dsi->debugfs_vpg);
 }
 
 #else
diff --git a/drivers/gpu/drm/bridge/tc358775.c b/drivers/gpu/drm/bridge/tc358775.c
index d951cdc58297..2272adcc5b4a 100644
--- a/drivers/gpu/drm/bridge/tc358775.c
+++ b/drivers/gpu/drm/bridge/tc358775.c
@@ -485,7 +485,7 @@ static void tc_bridge_enable(struct drm_bridge *bridge)
 		val |= TC358775_LVCFG_PCLKDIV(DIVIDE_BY_6);
 	} else {
 		val |= TC358775_LVCFG_PCLKDIV(DIVIDE_BY_3);
-	};
+	}
 	d2l_write(tc->i2c, LVCFG, val);
 }
 
diff --git a/drivers/gpu/drm/drm_atomic_helper.c b/drivers/gpu/drm/drm_atomic_helper.c
index 9e1ad493e689..f9170b4b22e7 100644
--- a/drivers/gpu/drm/drm_atomic_helper.c
+++ b/drivers/gpu/drm/drm_atomic_helper.c
@@ -1115,9 +1115,7 @@ disable_outputs(struct drm_device *dev, struct drm_atomic_state *old_state)
  * @old_state: atomic state object with old state structures
  *
  * This function updates all the various legacy modeset state pointers in
- * connectors, encoders and CRTCs. It also updates the timestamping constants
- * used for precise vblank timestamps by calling
- * drm_calc_timestamping_constants().
+ * connectors, encoders and CRTCs.
  *
  * Drivers can use this for building their own atomic commit if they don't have
  * a pure helper-based modeset implementation.
@@ -1186,13 +1184,30 @@ drm_atomic_helper_update_legacy_modeset_state(struct drm_device *dev,
 			crtc->x = new_plane_state->src_x >> 16;
 			crtc->y = new_plane_state->src_y >> 16;
 		}
+	}
+}
+EXPORT_SYMBOL(drm_atomic_helper_update_legacy_modeset_state);
 
+/**
+ * drm_atomic_helper_calc_timestamping_constants - update vblank timestamping constants
+ * @state: atomic state object
+ *
+ * Updates the timestamping constants used for precise vblank timestamps
+ * by calling drm_calc_timestamping_constants() for all enabled crtcs in @state.
+ */
+void drm_atomic_helper_calc_timestamping_constants(struct drm_atomic_state *state)
+{
+	struct drm_crtc_state *new_crtc_state;
+	struct drm_crtc *crtc;
+	int i;
+
+	for_each_new_crtc_in_state(state, crtc, new_crtc_state, i) {
 		if (new_crtc_state->enable)
 			drm_calc_timestamping_constants(crtc,
 							&new_crtc_state->adjusted_mode);
 	}
 }
-EXPORT_SYMBOL(drm_atomic_helper_update_legacy_modeset_state);
+EXPORT_SYMBOL(drm_atomic_helper_calc_timestamping_constants);
 
 static void
 crtc_set_mode(struct drm_device *dev, struct drm_atomic_state *old_state)
@@ -1276,6 +1291,7 @@ void drm_atomic_helper_commit_modeset_disables(struct drm_device *dev,
 	disable_outputs(dev, old_state);
 
 	drm_atomic_helper_update_legacy_modeset_state(dev, old_state);
+	drm_atomic_helper_calc_timestamping_constants(old_state);
 
 	crtc_set_mode(dev, old_state);
 }
diff --git a/drivers/gpu/drm/drm_connector.c b/drivers/gpu/drm/drm_connector.c
index 3d48ad1c3682..717c4e7271b0 100644
--- a/drivers/gpu/drm/drm_connector.c
+++ b/drivers/gpu/drm/drm_connector.c
@@ -2289,7 +2289,7 @@ static struct drm_encoder *drm_connector_get_encoder(struct drm_connector *conne
 
 static bool
 drm_mode_expose_to_userspace(const struct drm_display_mode *mode,
-			     const struct list_head *export_list,
+			     const struct list_head *modes,
 			     const struct drm_file *file_priv)
 {
 	/*
@@ -2305,15 +2305,17 @@ drm_mode_expose_to_userspace(const struct drm_display_mode *mode,
 	 * while preparing the list of user-modes.
 	 */
 	if (!file_priv->aspect_ratio_allowed) {
-		struct drm_display_mode *mode_itr;
+		const struct drm_display_mode *mode_itr;
 
-		list_for_each_entry(mode_itr, export_list, export_head)
-			if (drm_mode_match(mode_itr, mode,
+		list_for_each_entry(mode_itr, modes, head) {
+			if (mode_itr->expose_to_userspace &&
+			    drm_mode_match(mode_itr, mode,
 					   DRM_MODE_MATCH_TIMINGS |
 					   DRM_MODE_MATCH_CLOCK |
 					   DRM_MODE_MATCH_FLAGS |
 					   DRM_MODE_MATCH_3D_FLAGS))
 				return false;
+		}
 	}
 
 	return true;
@@ -2333,7 +2335,6 @@ int drm_mode_getconnector(struct drm_device *dev, void *data,
 	struct drm_mode_modeinfo u_mode;
 	struct drm_mode_modeinfo __user *mode_ptr;
 	uint32_t __user *encoder_ptr;
-	LIST_HEAD(export_list);
 
 	if (!drm_core_check_feature(dev, DRIVER_MODESET))
 		return -EOPNOTSUPP;
@@ -2377,25 +2378,30 @@ int drm_mode_getconnector(struct drm_device *dev, void *data,
 	out_resp->connection = connector->status;
 
 	/* delayed so we get modes regardless of pre-fill_modes state */
-	list_for_each_entry(mode, &connector->modes, head)
-		if (drm_mode_expose_to_userspace(mode, &export_list,
+	list_for_each_entry(mode, &connector->modes, head) {
+		WARN_ON(mode->expose_to_userspace);
+
+		if (drm_mode_expose_to_userspace(mode, &connector->modes,
 						 file_priv)) {
-			list_add_tail(&mode->export_head, &export_list);
+			mode->expose_to_userspace = true;
 			mode_count++;
 		}
+	}
 
 	/*
 	 * This ioctl is called twice, once to determine how much space is
 	 * needed, and the 2nd time to fill it.
-	 * The modes that need to be exposed to the user are maintained in the
-	 * 'export_list'. When the ioctl is called first time to determine the,
-	 * space, the export_list gets filled, to find the no.of modes. In the
-	 * 2nd time, the user modes are filled, one by one from the export_list.
 	 */
 	if ((out_resp->count_modes >= mode_count) && mode_count) {
 		copied = 0;
 		mode_ptr = (struct drm_mode_modeinfo __user *)(unsigned long)out_resp->modes_ptr;
-		list_for_each_entry(mode, &export_list, export_head) {
+		list_for_each_entry(mode, &connector->modes, head) {
+			if (!mode->expose_to_userspace)
+				continue;
+
+			/* Clear the tag for the next time around */
+			mode->expose_to_userspace = false;
+
 			drm_mode_convert_to_umode(&u_mode, mode);
 			/*
 			 * Reset aspect ratio flags of user-mode, if modes with
@@ -2406,13 +2412,26 @@ int drm_mode_getconnector(struct drm_device *dev, void *data,
 			if (copy_to_user(mode_ptr + copied,
 					 &u_mode, sizeof(u_mode))) {
 				ret = -EFAULT;
+
+				/*
+				 * Clear the tag for the rest of
+				 * the modes for the next time around.
+				 */
+				list_for_each_entry_continue(mode, &connector->modes, head)
+					mode->expose_to_userspace = false;
+
 				mutex_unlock(&dev->mode_config.mutex);
 
 				goto out;
 			}
 			copied++;
 		}
+	} else {
+		/* Clear the tag for the next time around */
+		list_for_each_entry(mode, &connector->modes, head)
+			mode->expose_to_userspace = false;
 	}
+
 	out_resp->count_modes = mode_count;
 	mutex_unlock(&dev->mode_config.mutex);
 
diff --git a/drivers/gpu/drm/drm_debugfs_crc.c b/drivers/gpu/drm/drm_debugfs_crc.c
index 5d67a41f7c3a..3dd70d813f69 100644
--- a/drivers/gpu/drm/drm_debugfs_crc.c
+++ b/drivers/gpu/drm/drm_debugfs_crc.c
@@ -144,8 +144,10 @@ static ssize_t crc_control_write(struct file *file, const char __user *ubuf,
 		source[len - 1] = '\0';
 
 	ret = crtc->funcs->verify_crc_source(crtc, source, &values_cnt);
-	if (ret)
+	if (ret) {
+		kfree(source);
 		return ret;
+	}
 
 	spin_lock_irq(&crc->lock);
 
diff --git a/drivers/gpu/drm/drm_dp_mst_topology.c b/drivers/gpu/drm/drm_dp_mst_topology.c
index 54864400015d..e87542533640 100644
--- a/drivers/gpu/drm/drm_dp_mst_topology.c
+++ b/drivers/gpu/drm/drm_dp_mst_topology.c
@@ -1039,6 +1039,8 @@ static bool drm_dp_sideband_parse_reply(struct drm_dp_sideband_msg_rx *raw,
 		return drm_dp_sideband_parse_remote_dpcd_write(raw, msg);
 	case DP_REMOTE_I2C_READ:
 		return drm_dp_sideband_parse_remote_i2c_read_ack(raw, msg);
+	case DP_REMOTE_I2C_WRITE:
+		return true; /* since there's nothing to parse */
 	case DP_ENUM_PATH_RESOURCES:
 		return drm_dp_sideband_parse_enum_path_resources_ack(raw, msg);
 	case DP_ALLOCATE_PAYLOAD:
@@ -5499,29 +5501,29 @@ static bool remote_i2c_read_ok(const struct i2c_msg msgs[], int num)
 		msgs[num - 1].len <= 0xff;
 }
 
-/* I2C device */
-static int drm_dp_mst_i2c_xfer(struct i2c_adapter *adapter, struct i2c_msg *msgs,
-			       int num)
+static bool remote_i2c_write_ok(const struct i2c_msg msgs[], int num)
+{
+	int i;
+
+	for (i = 0; i < num - 1; i++) {
+		if (msgs[i].flags & I2C_M_RD || !(msgs[i].flags & I2C_M_STOP) ||
+		    msgs[i].len > 0xff)
+			return false;
+	}
+
+	return !(msgs[num - 1].flags & I2C_M_RD) && msgs[num - 1].len <= 0xff;
+}
+
+static int drm_dp_mst_i2c_read(struct drm_dp_mst_branch *mstb,
+			       struct drm_dp_mst_port *port,
+			       struct i2c_msg *msgs, int num)
 {
-	struct drm_dp_aux *aux = adapter->algo_data;
-	struct drm_dp_mst_port *port = container_of(aux, struct drm_dp_mst_port, aux);
-	struct drm_dp_mst_branch *mstb;
 	struct drm_dp_mst_topology_mgr *mgr = port->mgr;
 	unsigned int i;
 	struct drm_dp_sideband_msg_req_body msg;
 	struct drm_dp_sideband_msg_tx *txmsg = NULL;
 	int ret;
 
-	mstb = drm_dp_mst_topology_get_mstb_validated(mgr, port->parent);
-	if (!mstb)
-		return -EREMOTEIO;
-
-	if (!remote_i2c_read_ok(msgs, num)) {
-		DRM_DEBUG_KMS("Unsupported I2C transaction for MST device\n");
-		ret = -EIO;
-		goto out;
-	}
-
 	memset(&msg, 0, sizeof(msg));
 	msg.req_type = DP_REMOTE_I2C_READ;
 	msg.u.i2c_read.num_transactions = num - 1;
@@ -5562,6 +5564,78 @@ static int drm_dp_mst_i2c_xfer(struct i2c_adapter *adapter, struct i2c_msg *msgs
 	}
 out:
 	kfree(txmsg);
+	return ret;
+}
+
+static int drm_dp_mst_i2c_write(struct drm_dp_mst_branch *mstb,
+				struct drm_dp_mst_port *port,
+				struct i2c_msg *msgs, int num)
+{
+	struct drm_dp_mst_topology_mgr *mgr = port->mgr;
+	unsigned int i;
+	struct drm_dp_sideband_msg_req_body msg;
+	struct drm_dp_sideband_msg_tx *txmsg = NULL;
+	int ret;
+
+	txmsg = kzalloc(sizeof(*txmsg), GFP_KERNEL);
+	if (!txmsg) {
+		ret = -ENOMEM;
+		goto out;
+	}
+	for (i = 0; i < num; i++) {
+		memset(&msg, 0, sizeof(msg));
+		msg.req_type = DP_REMOTE_I2C_WRITE;
+		msg.u.i2c_write.port_number = port->port_num;
+		msg.u.i2c_write.write_i2c_device_id = msgs[i].addr;
+		msg.u.i2c_write.num_bytes = msgs[i].len;
+		msg.u.i2c_write.bytes = msgs[i].buf;
+
+		memset(txmsg, 0, sizeof(*txmsg));
+		txmsg->dst = mstb;
+
+		drm_dp_encode_sideband_req(&msg, txmsg);
+		drm_dp_queue_down_tx(mgr, txmsg);
+
+		ret = drm_dp_mst_wait_tx_reply(mstb, txmsg);
+		if (ret > 0) {
+			if (txmsg->reply.reply_type == DP_SIDEBAND_REPLY_NAK) {
+				ret = -EREMOTEIO;
+				goto out;
+			}
+		} else {
+			goto out;
+		}
+	}
+	ret = num;
+out:
+	kfree(txmsg);
+	return ret;
+}
+
+/* I2C device */
+static int drm_dp_mst_i2c_xfer(struct i2c_adapter *adapter,
+			       struct i2c_msg *msgs, int num)
+{
+	struct drm_dp_aux *aux = adapter->algo_data;
+	struct drm_dp_mst_port *port =
+		container_of(aux, struct drm_dp_mst_port, aux);
+	struct drm_dp_mst_branch *mstb;
+	struct drm_dp_mst_topology_mgr *mgr = port->mgr;
+	int ret;
+
+	mstb = drm_dp_mst_topology_get_mstb_validated(mgr, port->parent);
+	if (!mstb)
+		return -EREMOTEIO;
+
+	if (remote_i2c_read_ok(msgs, num)) {
+		ret = drm_dp_mst_i2c_read(mstb, port, msgs, num);
+	} else if (remote_i2c_write_ok(msgs, num)) {
+		ret = drm_dp_mst_i2c_write(mstb, port, msgs, num);
+	} else {
+		DRM_DEBUG_KMS("Unsupported I2C transaction for MST device\n");
+		ret = -EIO;
+	}
+
 	drm_dp_mst_topology_put_mstb(mstb);
 	return ret;
 }
diff --git a/drivers/gpu/drm/drm_drv.c b/drivers/gpu/drm/drm_drv.c
index 13068fdf4331..cd162d406078 100644
--- a/drivers/gpu/drm/drm_drv.c
+++ b/drivers/gpu/drm/drm_drv.c
@@ -240,13 +240,13 @@ void drm_minor_release(struct drm_minor *minor)
  * DOC: driver instance overview
  *
  * A device instance for a drm driver is represented by &struct drm_device. This
- * is initialized with drm_dev_init(), usually from bus-specific ->probe()
- * callbacks implemented by the driver. The driver then needs to initialize all
- * the various subsystems for the drm device like memory management, vblank
- * handling, modesetting support and intial output configuration plus obviously
- * initialize all the corresponding hardware bits. Finally when everything is up
- * and running and ready for userspace the device instance can be published
- * using drm_dev_register().
+ * is allocated and initialized with devm_drm_dev_alloc(), usually from
+ * bus-specific ->probe() callbacks implemented by the driver. The driver then
+ * needs to initialize all the various subsystems for the drm device like memory
+ * management, vblank handling, modesetting support and initial output
+ * configuration plus obviously initialize all the corresponding hardware bits.
+ * Finally when everything is up and running and ready for userspace the device
+ * instance can be published using drm_dev_register().
  *
  * There is also deprecated support for initalizing device instances using
  * bus-specific helpers and the &drm_driver.load callback. But due to
@@ -274,7 +274,7 @@ void drm_minor_release(struct drm_minor *minor)
  *
  * The following example shows a typical structure of a DRM display driver.
  * The example focus on the probe() function and the other functions that is
- * almost always present and serves as a demonstration of devm_drm_dev_init().
+ * almost always present and serves as a demonstration of devm_drm_dev_alloc().
  *
  * .. code-block:: c
  *
@@ -294,22 +294,12 @@ void drm_minor_release(struct drm_minor *minor)
  *		struct drm_device *drm;
  *		int ret;
  *
- *		// devm_kzalloc() can't be used here because the drm_device '
- *		// lifetime can exceed the device lifetime if driver unbind
- *		// happens when userspace still has open file descriptors.
- *		priv = kzalloc(sizeof(*priv), GFP_KERNEL);
- *		if (!priv)
- *			return -ENOMEM;
- *
+ *		priv = devm_drm_dev_alloc(&pdev->dev, &driver_drm_driver,
+ *					  struct driver_device, drm);
+ *		if (IS_ERR(priv))
+ *			return PTR_ERR(priv);
  *		drm = &priv->drm;
  *
- *		ret = devm_drm_dev_init(&pdev->dev, drm, &driver_drm_driver);
- *		if (ret) {
- *			kfree(priv);
- *			return ret;
- *		}
- *		drmm_add_final_kfree(drm, priv);
- *
  *		ret = drmm_mode_config_init(drm);
  *		if (ret)
  *			return ret;
@@ -550,9 +540,9 @@ static void drm_fs_inode_free(struct inode *inode)
  * following guidelines apply:
  *
  *  - The entire device initialization procedure should be run from the
- *    &component_master_ops.master_bind callback, starting with drm_dev_init(),
- *    then binding all components with component_bind_all() and finishing with
- *    drm_dev_register().
+ *    &component_master_ops.master_bind callback, starting with
+ *    devm_drm_dev_alloc(), then binding all components with
+ *    component_bind_all() and finishing with drm_dev_register().
  *
  *  - The opaque pointer passed to all components through component_bind_all()
  *    should point at &struct drm_device of the device instance, not some driver
@@ -583,43 +573,9 @@ static void drm_dev_init_release(struct drm_device *dev, void *res)
 	drm_legacy_destroy_members(dev);
 }
 
-/**
- * drm_dev_init - Initialise new DRM device
- * @dev: DRM device
- * @driver: DRM driver
- * @parent: Parent device object
- *
- * Initialize a new DRM device. No device registration is done.
- * Call drm_dev_register() to advertice the device to user space and register it
- * with other core subsystems. This should be done last in the device
- * initialization sequence to make sure userspace can't access an inconsistent
- * state.
- *
- * The initial ref-count of the object is 1. Use drm_dev_get() and
- * drm_dev_put() to take and drop further ref-counts.
- *
- * It is recommended that drivers embed &struct drm_device into their own device
- * structure.
- *
- * Drivers that do not want to allocate their own device struct
- * embedding &struct drm_device can call drm_dev_alloc() instead. For drivers
- * that do embed &struct drm_device it must be placed first in the overall
- * structure, and the overall structure must be allocated using kmalloc(): The
- * drm core's release function unconditionally calls kfree() on the @dev pointer
- * when the final reference is released. To override this behaviour, and so
- * allow embedding of the drm_device inside the driver's device struct at an
- * arbitrary offset, you must supply a &drm_driver.release callback and control
- * the finalization explicitly.
- *
- * Note that drivers must call drmm_add_final_kfree() after this function has
- * completed successfully.
- *
- * RETURNS:
- * 0 on success, or error code on failure.
- */
-int drm_dev_init(struct drm_device *dev,
-		 struct drm_driver *driver,
-		 struct device *parent)
+static int drm_dev_init(struct drm_device *dev,
+			struct drm_driver *driver,
+			struct device *parent)
 {
 	int ret;
 
@@ -699,31 +655,15 @@ err:
 
 	return ret;
 }
-EXPORT_SYMBOL(drm_dev_init);
 
 static void devm_drm_dev_init_release(void *data)
 {
 	drm_dev_put(data);
 }
 
-/**
- * devm_drm_dev_init - Resource managed drm_dev_init()
- * @parent: Parent device object
- * @dev: DRM device
- * @driver: DRM driver
- *
- * Managed drm_dev_init(). The DRM device initialized with this function is
- * automatically put on driver detach using drm_dev_put().
- *
- * Note that drivers must call drmm_add_final_kfree() after this function has
- * completed successfully.
- *
- * RETURNS:
- * 0 on success, or error code on failure.
- */
-int devm_drm_dev_init(struct device *parent,
-		      struct drm_device *dev,
-		      struct drm_driver *driver)
+static int devm_drm_dev_init(struct device *parent,
+			     struct drm_device *dev,
+			     struct drm_driver *driver)
 {
 	int ret;
 
@@ -737,7 +677,6 @@ int devm_drm_dev_init(struct device *parent,
 
 	return ret;
 }
-EXPORT_SYMBOL(devm_drm_dev_init);
 
 void *__devm_drm_dev_alloc(struct device *parent, struct drm_driver *driver,
 			   size_t size, size_t offset)
@@ -767,19 +706,9 @@ EXPORT_SYMBOL(__devm_drm_dev_alloc);
  * @driver: DRM driver to allocate device for
  * @parent: Parent device object
  *
- * Allocate and initialize a new DRM device. No device registration is done.
- * Call drm_dev_register() to advertice the device to user space and register it
- * with other core subsystems. This should be done last in the device
- * initialization sequence to make sure userspace can't access an inconsistent
- * state.
- *
- * The initial ref-count of the object is 1. Use drm_dev_get() and
- * drm_dev_put() to take and drop further ref-counts.
- *
- * Note that for purely virtual devices @parent can be NULL.
- *
- * Drivers that wish to subclass or embed &struct drm_device into their
- * own struct should look at using drm_dev_init() instead.
+ * This is the deprecated version of devm_drm_dev_alloc(), which does not support
+ * subclassing through embedding the struct &drm_device in a driver private
+ * structure, and which does not support automatic cleanup through devres.
  *
  * RETURNS:
  * Pointer to new DRM device, or ERR_PTR on failure.
diff --git a/drivers/gpu/drm/drm_framebuffer.c b/drivers/gpu/drm/drm_framebuffer.c
index df656366a530..2f5b0c2bb0fe 100644
--- a/drivers/gpu/drm/drm_framebuffer.c
+++ b/drivers/gpu/drm/drm_framebuffer.c
@@ -176,8 +176,7 @@ static int framebuffer_check(struct drm_device *dev,
 	int i;
 
 	/* check if the format is supported at all */
-	info = __drm_format_info(r->pixel_format);
-	if (!info) {
+	if (!__drm_format_info(r->pixel_format)) {
 		struct drm_format_name_buf format_name;
 
 		DRM_DEBUG_KMS("bad framebuffer format %s\n",
@@ -186,9 +185,6 @@ static int framebuffer_check(struct drm_device *dev,
 		return -EINVAL;
 	}
 
-	/* now let the driver pick its own format info */
-	info = drm_get_format_info(dev, r);
-
 	if (r->width == 0) {
 		DRM_DEBUG_KMS("bad framebuffer width %u\n", r->width);
 		return -EINVAL;
@@ -199,6 +195,9 @@ static int framebuffer_check(struct drm_device *dev,
 		return -EINVAL;
 	}
 
+	/* now let the driver pick its own format info */
+	info = drm_get_format_info(dev, r);
+
 	for (i = 0; i < info->num_planes; i++) {
 		unsigned int width = fb_plane_width(r->width, info, i);
 		unsigned int height = fb_plane_height(r->height, info, i);
diff --git a/drivers/gpu/drm/drm_gem_shmem_helper.c b/drivers/gpu/drm/drm_gem_shmem_helper.c
index 47d8211221f2..d77c9f8ff26c 100644
--- a/drivers/gpu/drm/drm_gem_shmem_helper.c
+++ b/drivers/gpu/drm/drm_gem_shmem_helper.c
@@ -655,7 +655,7 @@ struct sg_table *drm_gem_shmem_get_sg_table(struct drm_gem_object *obj)
 
 	WARN_ON(shmem->base.import_attach);
 
-	return drm_prime_pages_to_sg(shmem->pages, obj->size >> PAGE_SHIFT);
+	return drm_prime_pages_to_sg(obj->dev, shmem->pages, obj->size >> PAGE_SHIFT);
 }
 EXPORT_SYMBOL_GPL(drm_gem_shmem_get_sg_table);
 
diff --git a/drivers/gpu/drm/drm_gem_ttm_helper.c b/drivers/gpu/drm/drm_gem_ttm_helper.c
index 892b2288a104..0e4fb9ba43ad 100644
--- a/drivers/gpu/drm/drm_gem_ttm_helper.c
+++ b/drivers/gpu/drm/drm_gem_ttm_helper.c
@@ -43,12 +43,9 @@ void drm_gem_ttm_print_info(struct drm_printer *p, unsigned int indent,
 	drm_print_bits(p, bo->mem.placement, plname, ARRAY_SIZE(plname));
 	drm_printf(p, "\n");
 
-	if (bo->mem.bus.is_iomem) {
-		drm_printf_indent(p, indent, "bus.base=%lx\n",
-				  (unsigned long)bo->mem.bus.base);
+	if (bo->mem.bus.is_iomem)
 		drm_printf_indent(p, indent, "bus.offset=%lx\n",
 				  (unsigned long)bo->mem.bus.offset);
-	}
 }
 EXPORT_SYMBOL(drm_gem_ttm_print_info);
 
diff --git a/drivers/gpu/drm/drm_gem_vram_helper.c b/drivers/gpu/drm/drm_gem_vram_helper.c
index 545a877406f4..50cad0e4a92e 100644
--- a/drivers/gpu/drm/drm_gem_vram_helper.c
+++ b/drivers/gpu/drm/drm_gem_vram_helper.c
@@ -97,8 +97,8 @@ static const struct drm_gem_object_funcs drm_gem_vram_object_funcs;
  * hardware's draing engine.
  *
  * To access a buffer object's memory from the DRM driver, call
- * drm_gem_vram_kmap(). It (optionally) maps the buffer into kernel address
- * space and returns the memory address. Use drm_gem_vram_kunmap() to
+ * drm_gem_vram_vmap(). It maps the buffer into kernel address
+ * space and returns the memory address. Use drm_gem_vram_vunmap() to
  * release the mapping.
  */
 
@@ -135,28 +135,28 @@ static void ttm_buffer_object_destroy(struct ttm_buffer_object *bo)
 static void drm_gem_vram_placement(struct drm_gem_vram_object *gbo,
 				   unsigned long pl_flag)
 {
+	u32 invariant_flags = 0;
 	unsigned int i;
 	unsigned int c = 0;
-	u32 invariant_flags = pl_flag & TTM_PL_FLAG_TOPDOWN;
+
+	if (pl_flag & DRM_GEM_VRAM_PL_FLAG_TOPDOWN)
+		pl_flag = TTM_PL_FLAG_TOPDOWN;
 
 	gbo->placement.placement = gbo->placements;
 	gbo->placement.busy_placement = gbo->placements;
 
-	if (pl_flag & TTM_PL_FLAG_VRAM)
+	if (pl_flag & DRM_GEM_VRAM_PL_FLAG_VRAM) {
+		gbo->placements[c].mem_type = TTM_PL_VRAM;
 		gbo->placements[c++].flags = TTM_PL_FLAG_WC |
 					     TTM_PL_FLAG_UNCACHED |
-					     TTM_PL_FLAG_VRAM |
-					     invariant_flags;
-
-	if (pl_flag & TTM_PL_FLAG_SYSTEM)
-		gbo->placements[c++].flags = TTM_PL_MASK_CACHING |
-					     TTM_PL_FLAG_SYSTEM |
 					     invariant_flags;
+	}
 
-	if (!c)
+	if (pl_flag & DRM_GEM_VRAM_PL_FLAG_SYSTEM || !c) {
+		gbo->placements[c].mem_type = TTM_PL_SYSTEM;
 		gbo->placements[c++].flags = TTM_PL_MASK_CACHING |
-					     TTM_PL_FLAG_SYSTEM |
 					     invariant_flags;
+	}
 
 	gbo->placement.num_placement = c;
 	gbo->placement.num_busy_placement = c;
@@ -167,6 +167,10 @@ static void drm_gem_vram_placement(struct drm_gem_vram_object *gbo,
 	}
 }
 
+/*
+ * Note that on error, drm_gem_vram_init will free the buffer object.
+ */
+
 static int drm_gem_vram_init(struct drm_device *dev,
 			     struct drm_gem_vram_object *gbo,
 			     size_t size, unsigned long pg_align)
@@ -176,32 +180,37 @@ static int drm_gem_vram_init(struct drm_device *dev,
 	int ret;
 	size_t acc_size;
 
-	if (WARN_ONCE(!vmm, "VRAM MM not initialized"))
+	if (WARN_ONCE(!vmm, "VRAM MM not initialized")) {
+		kfree(gbo);
 		return -EINVAL;
+	}
 	bdev = &vmm->bdev;
 
 	gbo->bo.base.funcs = &drm_gem_vram_object_funcs;
 
 	ret = drm_gem_object_init(dev, &gbo->bo.base, size);
-	if (ret)
+	if (ret) {
+		kfree(gbo);
 		return ret;
+	}
 
 	acc_size = ttm_bo_dma_acc_size(bdev, size, sizeof(*gbo));
 
 	gbo->bo.bdev = bdev;
-	drm_gem_vram_placement(gbo, TTM_PL_FLAG_VRAM | TTM_PL_FLAG_SYSTEM);
+	drm_gem_vram_placement(gbo, DRM_GEM_VRAM_PL_FLAG_VRAM |
+			       DRM_GEM_VRAM_PL_FLAG_SYSTEM);
 
 	ret = ttm_bo_init(bdev, &gbo->bo, size, ttm_bo_type_device,
 			  &gbo->placement, pg_align, false, acc_size,
 			  NULL, NULL, ttm_buffer_object_destroy);
 	if (ret)
-		goto err_drm_gem_object_release;
+		/*
+		 * A failing ttm_bo_init will call ttm_buffer_object_destroy
+		 * to release gbo->bo.base and kfree gbo.
+		 */
+		return ret;
 
 	return 0;
-
-err_drm_gem_object_release:
-	drm_gem_object_release(&gbo->bo.base);
-	return ret;
 }
 
 /**
@@ -235,13 +244,9 @@ struct drm_gem_vram_object *drm_gem_vram_create(struct drm_device *dev,
 
 	ret = drm_gem_vram_init(dev, gbo, size, pg_align);
 	if (ret < 0)
-		goto err_kfree;
+		return ERR_PTR(ret);
 
 	return gbo;
-
-err_kfree:
-	kfree(gbo);
-	return ERR_PTR(ret);
 }
 EXPORT_SYMBOL(drm_gem_vram_create);
 
@@ -436,39 +441,6 @@ out:
 	return kmap->virtual;
 }
 
-/**
- * drm_gem_vram_kmap() - Maps a GEM VRAM object into kernel address space
- * @gbo:	the GEM VRAM object
- * @map:	establish a mapping if necessary
- * @is_iomem:	returns true if the mapped memory is I/O memory, or false \
-	otherwise; can be NULL
- *
- * This function maps the buffer object into the kernel's address space
- * or returns the current mapping. If the parameter map is false, the
- * function only queries the current mapping, but does not establish a
- * new one.
- *
- * Returns:
- * The buffers virtual address if mapped, or
- * NULL if not mapped, or
- * an ERR_PTR()-encoded error code otherwise.
- */
-void *drm_gem_vram_kmap(struct drm_gem_vram_object *gbo, bool map,
-			bool *is_iomem)
-{
-	int ret;
-	void *virtual;
-
-	ret = ttm_bo_reserve(&gbo->bo, true, false, NULL);
-	if (ret)
-		return ERR_PTR(ret);
-	virtual = drm_gem_vram_kmap_locked(gbo, map, is_iomem);
-	ttm_bo_unreserve(&gbo->bo);
-
-	return virtual;
-}
-EXPORT_SYMBOL(drm_gem_vram_kmap);
-
 static void drm_gem_vram_kunmap_locked(struct drm_gem_vram_object *gbo)
 {
 	if (WARN_ON_ONCE(!gbo->kmap_use_count))
@@ -485,22 +457,6 @@ static void drm_gem_vram_kunmap_locked(struct drm_gem_vram_object *gbo)
 }
 
 /**
- * drm_gem_vram_kunmap() - Unmaps a GEM VRAM object
- * @gbo:	the GEM VRAM object
- */
-void drm_gem_vram_kunmap(struct drm_gem_vram_object *gbo)
-{
-	int ret;
-
-	ret = ttm_bo_reserve(&gbo->bo, false, false, NULL);
-	if (WARN_ONCE(ret, "ttm_bo_reserve_failed(): ret=%d\n", ret))
-		return;
-	drm_gem_vram_kunmap_locked(gbo);
-	ttm_bo_unreserve(&gbo->bo);
-}
-EXPORT_SYMBOL(drm_gem_vram_kunmap);
-
-/**
  * drm_gem_vram_vmap() - Pins and maps a GEM VRAM object into kernel address
  *                       space
  * @gbo:	The GEM VRAM object to map
@@ -511,9 +467,6 @@ EXPORT_SYMBOL(drm_gem_vram_kunmap);
  * permanently. Call drm_gem_vram_vunmap() with the returned address to
  * unmap and unpin the GEM VRAM object.
  *
- * If you have special requirements for the pinning or mapping operations,
- * call drm_gem_vram_pin() and drm_gem_vram_kmap() directly.
- *
  * Returns:
  * The buffer's virtual address on success, or
  * an ERR_PTR()-encoded error code otherwise.
@@ -647,7 +600,7 @@ static bool drm_is_gem_vram(struct ttm_buffer_object *bo)
 static void drm_gem_vram_bo_driver_evict_flags(struct drm_gem_vram_object *gbo,
 					       struct ttm_placement *pl)
 {
-	drm_gem_vram_placement(gbo, TTM_PL_FLAG_SYSTEM);
+	drm_gem_vram_placement(gbo, DRM_GEM_VRAM_PL_FLAG_SYSTEM);
 	*pl = gbo->placement;
 }
 
@@ -967,16 +920,13 @@ static const struct drm_gem_object_funcs drm_gem_vram_object_funcs = {
  * TTM TT
  */
 
-static void backend_func_destroy(struct ttm_tt *tt)
+static void bo_driver_ttm_tt_destroy(struct ttm_bo_device *bdev, struct ttm_tt *tt)
 {
+	ttm_tt_destroy_common(bdev, tt);
 	ttm_tt_fini(tt);
 	kfree(tt);
 }
 
-static struct ttm_backend_func backend_func = {
-	.destroy = backend_func_destroy
-};
-
 /*
  * TTM BO device
  */
@@ -991,8 +941,6 @@ static struct ttm_tt *bo_driver_ttm_tt_create(struct ttm_buffer_object *bo,
 	if (!tt)
 		return NULL;
 
-	tt->func = &backend_func;
-
 	ret = ttm_tt_init(tt, bo, page_flags);
 	if (ret < 0)
 		goto err_ttm_tt_init;
@@ -1042,8 +990,7 @@ static int bo_driver_io_mem_reserve(struct ttm_bo_device *bdev,
 	case TTM_PL_SYSTEM:	/* nothing to do */
 		break;
 	case TTM_PL_VRAM:
-		mem->bus.offset = mem->start << PAGE_SHIFT;
-		mem->bus.base = vmm->vram_base;
+		mem->bus.offset = (mem->start << PAGE_SHIFT) + vmm->vram_base;
 		mem->bus.is_iomem = true;
 		break;
 	default:
@@ -1055,6 +1002,7 @@ static int bo_driver_io_mem_reserve(struct ttm_bo_device *bdev,
 
 static struct ttm_bo_driver bo_driver = {
 	.ttm_tt_create = bo_driver_ttm_tt_create,
+	.ttm_tt_destroy = bo_driver_ttm_tt_destroy,
 	.eviction_valuable = ttm_bo_eviction_valuable,
 	.evict_flags = bo_driver_evict_flags,
 	.move_notify = bo_driver_move_notify,
@@ -1110,9 +1058,7 @@ static int drm_vram_mm_init(struct drm_vram_mm *vmm, struct drm_device *dev,
 		return ret;
 
 	ret = ttm_range_man_init(&vmm->bdev, TTM_PL_VRAM,
-				 TTM_PL_FLAG_UNCACHED | TTM_PL_FLAG_WC,
-				 TTM_PL_FLAG_WC, false,
-				 vram_size >> PAGE_SHIFT);
+				 false, vram_size >> PAGE_SHIFT);
 	if (ret)
 		return ret;
 
diff --git a/drivers/gpu/drm/drm_internal.h b/drivers/gpu/drm/drm_internal.h
index 8e01caaf95cc..b65865c630b0 100644
--- a/drivers/gpu/drm/drm_internal.h
+++ b/drivers/gpu/drm/drm_internal.h
@@ -95,6 +95,7 @@ void drm_minor_release(struct drm_minor *minor);
 
 /* drm_managed.c */
 void drm_managed_release(struct drm_device *dev);
+void drmm_add_final_kfree(struct drm_device *dev, void *container);
 
 /* drm_vblank.c */
 static inline bool drm_vblank_passed(u64 seq, u64 ref)
diff --git a/drivers/gpu/drm/drm_managed.c b/drivers/gpu/drm/drm_managed.c
index 1e1356560c2e..37d7db6223be 100644
--- a/drivers/gpu/drm/drm_managed.c
+++ b/drivers/gpu/drm/drm_managed.c
@@ -27,7 +27,7 @@
  * be done directly with drmm_kmalloc() and the related functions. Everything
  * will be released on the final drm_dev_put() in reverse order of how the
  * release actions have been added and memory has been allocated since driver
- * loading started with drm_dev_init().
+ * loading started with devm_drm_dev_alloc().
  *
  * Note that release actions and managed memory can also be added and removed
  * during the lifetime of the driver, all the functions are fully concurrent
@@ -125,18 +125,6 @@ static void add_dr(struct drm_device *dev, struct drmres *dr)
 		       dr, dr->node.name, (unsigned long) dr->node.size);
 }
 
-/**
- * drmm_add_final_kfree - add release action for the final kfree()
- * @dev: DRM device
- * @container: pointer to the kmalloc allocation containing @dev
- *
- * Since the allocation containing the struct &drm_device must be allocated
- * before it can be initialized with drm_dev_init() there's no way to allocate
- * that memory with drmm_kmalloc(). To side-step this chicken-egg problem the
- * pointer for this final kfree() must be specified by calling this function. It
- * will be released in the final drm_dev_put() for @dev, after all other release
- * actions installed through drmm_add_action() have been processed.
- */
 void drmm_add_final_kfree(struct drm_device *dev, void *container)
 {
 	WARN_ON(dev->managed.final_kfree);
@@ -144,7 +132,6 @@ void drmm_add_final_kfree(struct drm_device *dev, void *container)
 	WARN_ON(dev + 1 > (struct drm_device *) (container + ksize(container)));
 	dev->managed.final_kfree = container;
 }
-EXPORT_SYMBOL(drmm_add_final_kfree);
 
 int __drmm_add_action(struct drm_device *dev,
 		      drmres_release_t action,
diff --git a/drivers/gpu/drm/drm_prime.c b/drivers/gpu/drm/drm_prime.c
index b8c7f068a5a4..11fe9ff76fd5 100644
--- a/drivers/gpu/drm/drm_prime.c
+++ b/drivers/gpu/drm/drm_prime.c
@@ -803,9 +803,11 @@ static const struct dma_buf_ops drm_gem_prime_dmabuf_ops =  {
  *
  * This is useful for implementing &drm_gem_object_funcs.get_sg_table.
  */
-struct sg_table *drm_prime_pages_to_sg(struct page **pages, unsigned int nr_pages)
+struct sg_table *drm_prime_pages_to_sg(struct drm_device *dev,
+				       struct page **pages, unsigned int nr_pages)
 {
 	struct sg_table *sg = NULL;
+	size_t max_segment = 0;
 	int ret;
 
 	sg = kmalloc(sizeof(struct sg_table), GFP_KERNEL);
@@ -814,8 +816,13 @@ struct sg_table *drm_prime_pages_to_sg(struct page **pages, unsigned int nr_page
 		goto out;
 	}
 
-	ret = sg_alloc_table_from_pages(sg, pages, nr_pages, 0,
-				nr_pages << PAGE_SHIFT, GFP_KERNEL);
+	if (dev)
+		max_segment = dma_max_mapping_size(dev->dev);
+	if (max_segment == 0 || max_segment > SCATTERLIST_MAX_SEGMENT)
+		max_segment = SCATTERLIST_MAX_SEGMENT;
+	ret = __sg_alloc_table_from_pages(sg, pages, nr_pages, 0,
+					  nr_pages << PAGE_SHIFT,
+					  max_segment, GFP_KERNEL);
 	if (ret)
 		goto out;
 
diff --git a/drivers/gpu/drm/drm_vblank.c b/drivers/gpu/drm/drm_vblank.c
index b18e1efbbae1..f135b79593dd 100644
--- a/drivers/gpu/drm/drm_vblank.c
+++ b/drivers/gpu/drm/drm_vblank.c
@@ -674,7 +674,7 @@ EXPORT_SYMBOL(drm_calc_timestamping_constants);
  *
  * Note that atomic drivers must call drm_calc_timestamping_constants() before
  * enabling a CRTC. The atomic helpers already take care of that in
- * drm_atomic_helper_update_legacy_modeset_state().
+ * drm_atomic_helper_calc_timestamping_constants().
  *
  * Returns:
  *
@@ -819,7 +819,7 @@ EXPORT_SYMBOL(drm_crtc_vblank_helper_get_vblank_timestamp_internal);
  *
  * Note that atomic drivers must call drm_calc_timestamping_constants() before
  * enabling a CRTC. The atomic helpers already take care of that in
- * drm_atomic_helper_update_legacy_modeset_state().
+ * drm_atomic_helper_calc_timestamping_constants().
  *
  * Returns:
  *
diff --git a/drivers/gpu/drm/etnaviv/etnaviv_gem.c b/drivers/gpu/drm/etnaviv/etnaviv_gem.c
index eaf1949bc2e4..d1533bdc1335 100644
--- a/drivers/gpu/drm/etnaviv/etnaviv_gem.c
+++ b/drivers/gpu/drm/etnaviv/etnaviv_gem.c
@@ -103,7 +103,8 @@ struct page **etnaviv_gem_get_pages(struct etnaviv_gem_object *etnaviv_obj)
 		int npages = etnaviv_obj->base.size >> PAGE_SHIFT;
 		struct sg_table *sgt;
 
-		sgt = drm_prime_pages_to_sg(etnaviv_obj->pages, npages);
+		sgt = drm_prime_pages_to_sg(etnaviv_obj->base.dev,
+					    etnaviv_obj->pages, npages);
 		if (IS_ERR(sgt)) {
 			dev_err(dev->dev, "failed to allocate sgt: %ld\n",
 				PTR_ERR(sgt));
diff --git a/drivers/gpu/drm/etnaviv/etnaviv_gem_prime.c b/drivers/gpu/drm/etnaviv/etnaviv_gem_prime.c
index 6d9e5c3c4dd5..4aa3426a9ba4 100644
--- a/drivers/gpu/drm/etnaviv/etnaviv_gem_prime.c
+++ b/drivers/gpu/drm/etnaviv/etnaviv_gem_prime.c
@@ -19,7 +19,7 @@ struct sg_table *etnaviv_gem_prime_get_sg_table(struct drm_gem_object *obj)
 	if (WARN_ON(!etnaviv_obj->pages))  /* should have already pinned! */
 		return ERR_PTR(-EINVAL);
 
-	return drm_prime_pages_to_sg(etnaviv_obj->pages, npages);
+	return drm_prime_pages_to_sg(obj->dev, etnaviv_obj->pages, npages);
 }
 
 void *etnaviv_gem_prime_vmap(struct drm_gem_object *obj)
diff --git a/drivers/gpu/drm/gma500/framebuffer.c b/drivers/gpu/drm/gma500/framebuffer.c
index da02d7e8a8f5..54d9876b5305 100644
--- a/drivers/gpu/drm/gma500/framebuffer.c
+++ b/drivers/gpu/drm/gma500/framebuffer.c
@@ -164,7 +164,7 @@ static int psbfb_mmap(struct fb_info *info, struct vm_area_struct *vma)
 	return 0;
 }
 
-static struct fb_ops psbfb_ops = {
+static const struct fb_ops psbfb_ops = {
 	.owner = THIS_MODULE,
 	DRM_FB_HELPER_DEFAULT_OPS,
 	.fb_setcolreg = psbfb_setcolreg,
@@ -175,7 +175,7 @@ static struct fb_ops psbfb_ops = {
 	.fb_sync = psbfb_sync,
 };
 
-static struct fb_ops psbfb_roll_ops = {
+static const struct fb_ops psbfb_roll_ops = {
 	.owner = THIS_MODULE,
 	DRM_FB_HELPER_DEFAULT_OPS,
 	.fb_setcolreg = psbfb_setcolreg,
@@ -186,7 +186,7 @@ static struct fb_ops psbfb_roll_ops = {
 	.fb_mmap = psbfb_mmap,
 };
 
-static struct fb_ops psbfb_unaccel_ops = {
+static const struct fb_ops psbfb_unaccel_ops = {
 	.owner = THIS_MODULE,
 	DRM_FB_HELPER_DEFAULT_OPS,
 	.fb_setcolreg = psbfb_setcolreg,
diff --git a/drivers/gpu/drm/i810/i810_dma.c b/drivers/gpu/drm/i810/i810_dma.c
index 303c2d483c6e..88250860f8e4 100644
--- a/drivers/gpu/drm/i810/i810_dma.c
+++ b/drivers/gpu/drm/i810/i810_dma.c
@@ -853,11 +853,11 @@ static void i810_dma_quiescent(struct drm_device *dev)
 	i810_wait_ring(dev, dev_priv->ring.Size - 8);
 }
 
-static int i810_flush_queue(struct drm_device *dev)
+static void i810_flush_queue(struct drm_device *dev)
 {
 	drm_i810_private_t *dev_priv = dev->dev_private;
 	struct drm_device_dma *dma = dev->dma;
-	int i, ret = 0;
+	int i;
 	RING_LOCALS;
 
 	i810_kernel_lost_context(dev);
@@ -882,7 +882,7 @@ static int i810_flush_queue(struct drm_device *dev)
 			DRM_DEBUG("still on client\n");
 	}
 
-	return ret;
+	return;
 }
 
 /* Must be called with the lock held */
diff --git a/drivers/gpu/drm/i915/display/intel_display.c b/drivers/gpu/drm/i915/display/intel_display.c
index efc1d0f33fd7..5a9d933e425a 100644
--- a/drivers/gpu/drm/i915/display/intel_display.c
+++ b/drivers/gpu/drm/i915/display/intel_display.c
@@ -13484,12 +13484,6 @@ encoder_retry:
 		    "hw max bpp: %i, pipe bpp: %i, dithering: %i\n",
 		    base_bpp, pipe_config->pipe_bpp, pipe_config->dither);
 
-	/*
-	 * Make drm_calc_timestamping_constants in
-	 * drm_atomic_helper_update_legacy_modeset_state() happy
-	 */
-	pipe_config->uapi.adjusted_mode = pipe_config->hw.adjusted_mode;
-
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/i915/gem/selftests/huge_pages.c b/drivers/gpu/drm/i915/gem/selftests/huge_pages.c
index 5daf4a2be422..1f35e71429b4 100644
--- a/drivers/gpu/drm/i915/gem/selftests/huge_pages.c
+++ b/drivers/gpu/drm/i915/gem/selftests/huge_pages.c
@@ -1617,7 +1617,7 @@ int i915_gem_huge_page_mock_selftests(void)
 out_put:
 	i915_vm_put(&ppgtt->vm);
 out_unlock:
-	drm_dev_put(&dev_priv->drm);
+	mock_destroy_device(dev_priv);
 	return err;
 }
 
diff --git a/drivers/gpu/drm/i915/gem/selftests/i915_gem_context.c b/drivers/gpu/drm/i915/gem/selftests/i915_gem_context.c
index 99becb86abd3..d3f87dc4eda3 100644
--- a/drivers/gpu/drm/i915/gem/selftests/i915_gem_context.c
+++ b/drivers/gpu/drm/i915/gem/selftests/i915_gem_context.c
@@ -1997,7 +1997,7 @@ int i915_gem_context_mock_selftests(void)
 
 	err = i915_subtests(tests, i915);
 
-	drm_dev_put(&i915->drm);
+	mock_destroy_device(i915);
 	return err;
 }
 
diff --git a/drivers/gpu/drm/i915/gem/selftests/i915_gem_dmabuf.c b/drivers/gpu/drm/i915/gem/selftests/i915_gem_dmabuf.c
index 2a52b92586b9..0845ce1ae37c 100644
--- a/drivers/gpu/drm/i915/gem/selftests/i915_gem_dmabuf.c
+++ b/drivers/gpu/drm/i915/gem/selftests/i915_gem_dmabuf.c
@@ -272,7 +272,7 @@ int i915_gem_dmabuf_mock_selftests(void)
 
 	err = i915_subtests(tests, i915);
 
-	drm_dev_put(&i915->drm);
+	mock_destroy_device(i915);
 	return err;
 }
 
diff --git a/drivers/gpu/drm/i915/gem/selftests/i915_gem_object.c b/drivers/gpu/drm/i915/gem/selftests/i915_gem_object.c
index faa5b6d91795..bf853c40ec65 100644
--- a/drivers/gpu/drm/i915/gem/selftests/i915_gem_object.c
+++ b/drivers/gpu/drm/i915/gem/selftests/i915_gem_object.c
@@ -85,7 +85,7 @@ int i915_gem_object_mock_selftests(void)
 
 	err = i915_subtests(tests, i915);
 
-	drm_dev_put(&i915->drm);
+	mock_destroy_device(i915);
 	return err;
 }
 
diff --git a/drivers/gpu/drm/i915/gem/selftests/i915_gem_phys.c b/drivers/gpu/drm/i915/gem/selftests/i915_gem_phys.c
index a94243dc4c5c..8cee68c6a6dc 100644
--- a/drivers/gpu/drm/i915/gem/selftests/i915_gem_phys.c
+++ b/drivers/gpu/drm/i915/gem/selftests/i915_gem_phys.c
@@ -73,6 +73,6 @@ int i915_gem_phys_mock_selftests(void)
 
 	err = i915_subtests(tests, i915);
 
-	drm_dev_put(&i915->drm);
+	mock_destroy_device(i915);
 	return err;
 }
diff --git a/drivers/gpu/drm/i915/gt/selftest_timeline.c b/drivers/gpu/drm/i915/gt/selftest_timeline.c
index 96d164a3841d..19c2cb166e7c 100644
--- a/drivers/gpu/drm/i915/gt/selftest_timeline.c
+++ b/drivers/gpu/drm/i915/gt/selftest_timeline.c
@@ -158,7 +158,7 @@ out:
 		__mock_hwsp_record(&state, na, NULL);
 	kfree(state.history);
 err_put:
-	drm_dev_put(&i915->drm);
+	mock_destroy_device(i915);
 	return err;
 }
 
diff --git a/drivers/gpu/drm/i915/selftests/i915_gem_evict.c b/drivers/gpu/drm/i915/selftests/i915_gem_evict.c
index 028baae9631f..f88473d396f4 100644
--- a/drivers/gpu/drm/i915/selftests/i915_gem_evict.c
+++ b/drivers/gpu/drm/i915/selftests/i915_gem_evict.c
@@ -536,7 +536,7 @@ int i915_gem_evict_mock_selftests(void)
 	with_intel_runtime_pm(&i915->runtime_pm, wakeref)
 		err = i915_subtests(tests, &i915->gt);
 
-	drm_dev_put(&i915->drm);
+	mock_destroy_device(i915);
 	return err;
 }
 
diff --git a/drivers/gpu/drm/i915/selftests/i915_gem_gtt.c b/drivers/gpu/drm/i915/selftests/i915_gem_gtt.c
index af8205a2bd8f..c53a222e3dec 100644
--- a/drivers/gpu/drm/i915/selftests/i915_gem_gtt.c
+++ b/drivers/gpu/drm/i915/selftests/i915_gem_gtt.c
@@ -1727,7 +1727,7 @@ int i915_gem_gtt_mock_selftests(void)
 	mock_fini_ggtt(ggtt);
 	kfree(ggtt);
 out_put:
-	drm_dev_put(&i915->drm);
+	mock_destroy_device(i915);
 	return err;
 }
 
diff --git a/drivers/gpu/drm/i915/selftests/i915_request.c b/drivers/gpu/drm/i915/selftests/i915_request.c
index 3092ca763789..64bbb8288249 100644
--- a/drivers/gpu/drm/i915/selftests/i915_request.c
+++ b/drivers/gpu/drm/i915/selftests/i915_request.c
@@ -527,7 +527,7 @@ int i915_request_mock_selftests(void)
 	with_intel_runtime_pm(&i915->runtime_pm, wakeref)
 		err = i915_subtests(tests, i915);
 
-	drm_dev_put(&i915->drm);
+	mock_destroy_device(i915);
 
 	return err;
 }
diff --git a/drivers/gpu/drm/i915/selftests/i915_vma.c b/drivers/gpu/drm/i915/selftests/i915_vma.c
index 88c5e9acb84c..1b6125e4c1ac 100644
--- a/drivers/gpu/drm/i915/selftests/i915_vma.c
+++ b/drivers/gpu/drm/i915/selftests/i915_vma.c
@@ -841,7 +841,7 @@ int i915_vma_mock_selftests(void)
 	mock_fini_ggtt(ggtt);
 	kfree(ggtt);
 out_put:
-	drm_dev_put(&i915->drm);
+	mock_destroy_device(i915);
 	return err;
 }
 
diff --git a/drivers/gpu/drm/i915/selftests/intel_memory_region.c b/drivers/gpu/drm/i915/selftests/intel_memory_region.c
index 93a38a323584..334b0648e253 100644
--- a/drivers/gpu/drm/i915/selftests/intel_memory_region.c
+++ b/drivers/gpu/drm/i915/selftests/intel_memory_region.c
@@ -791,7 +791,7 @@ int intel_memory_region_mock_selftests(void)
 
 	intel_memory_region_put(mem);
 out_unref:
-	drm_dev_put(&i915->drm);
+	mock_destroy_device(i915);
 	return err;
 }
 
diff --git a/drivers/gpu/drm/i915/selftests/mock_gem_device.c b/drivers/gpu/drm/i915/selftests/mock_gem_device.c
index 397c313a8b69..b6c42fd872ad 100644
--- a/drivers/gpu/drm/i915/selftests/mock_gem_device.c
+++ b/drivers/gpu/drm/i915/selftests/mock_gem_device.c
@@ -79,8 +79,6 @@ static void mock_device_release(struct drm_device *dev)
 
 out:
 	i915_params_free(&i915->params);
-	put_device(&i915->drm.pdev->dev);
-	i915->drm.pdev = NULL;
 }
 
 static struct drm_driver mock_driver = {
@@ -123,17 +121,10 @@ struct drm_i915_private *mock_gem_device(void)
 #endif
 	struct drm_i915_private *i915;
 	struct pci_dev *pdev;
-	int err;
 
 	pdev = kzalloc(sizeof(*pdev), GFP_KERNEL);
 	if (!pdev)
 		return NULL;
-	i915 = kzalloc(sizeof(*i915), GFP_KERNEL);
-	if (!i915) {
-		kfree(pdev);
-		return NULL;
-	}
-
 	device_initialize(&pdev->dev);
 	pdev->class = PCI_BASE_CLASS_DISPLAY << 16;
 	pdev->dev.release = release_dev;
@@ -144,8 +135,23 @@ struct drm_i915_private *mock_gem_device(void)
 	/* HACK to disable iommu for the fake device; force identity mapping */
 	pdev->dev.iommu = &fake_iommu;
 #endif
+	if (!devres_open_group(&pdev->dev, NULL, GFP_KERNEL)) {
+		put_device(&pdev->dev);
+		return NULL;
+	}
+
+	i915 = devm_drm_dev_alloc(&pdev->dev, &mock_driver,
+				  struct drm_i915_private, drm);
+	if (IS_ERR(i915)) {
+		pr_err("Failed to allocate mock GEM device: err=%ld\n", PTR_ERR(i915));
+		devres_release_group(&pdev->dev, NULL);
+		put_device(&pdev->dev);
+
+		return NULL;
+	}
 
 	pci_set_drvdata(pdev, i915);
+	i915->drm.pdev = pdev;
 
 	dev_pm_domain_set(&pdev->dev, &pm_domain);
 	pm_runtime_enable(&pdev->dev);
@@ -153,16 +159,6 @@ struct drm_i915_private *mock_gem_device(void)
 	if (pm_runtime_enabled(&pdev->dev))
 		WARN_ON(pm_runtime_get_sync(&pdev->dev));
 
-	err = drm_dev_init(&i915->drm, &mock_driver, &pdev->dev);
-	if (err) {
-		pr_err("Failed to initialise mock GEM device: err=%d\n", err);
-		put_device(&pdev->dev);
-		kfree(i915);
-
-		return NULL;
-	}
-	i915->drm.pdev = pdev;
-	drmm_add_final_kfree(&i915->drm, i915);
 
 	i915_params_copy(&i915->params, &i915_modparams);
 
@@ -222,7 +218,15 @@ err_drv:
 	intel_gt_driver_late_release(&i915->gt);
 	intel_memory_regions_driver_release(i915);
 	drm_mode_config_cleanup(&i915->drm);
-	drm_dev_put(&i915->drm);
+	mock_destroy_device(i915);
 
 	return NULL;
 }
+
+void mock_destroy_device(struct drm_i915_private *i915)
+{
+	struct device *dev = i915->drm.dev;
+
+	devres_release_group(dev, NULL);
+	put_device(dev);
+}
diff --git a/drivers/gpu/drm/i915/selftests/mock_gem_device.h b/drivers/gpu/drm/i915/selftests/mock_gem_device.h
index b5dc4e394555..953cfe4fab34 100644
--- a/drivers/gpu/drm/i915/selftests/mock_gem_device.h
+++ b/drivers/gpu/drm/i915/selftests/mock_gem_device.h
@@ -7,4 +7,6 @@ struct drm_i915_private;
 struct drm_i915_private *mock_gem_device(void);
 void mock_device_flush(struct drm_i915_private *i915);
 
+void mock_destroy_device(struct drm_i915_private *i915);
+
 #endif /* !__MOCK_GEM_DEVICE_H__ */
diff --git a/drivers/gpu/drm/imx/Kconfig b/drivers/gpu/drm/imx/Kconfig
index 207bf7409dfb..6231048aa5aa 100644
--- a/drivers/gpu/drm/imx/Kconfig
+++ b/drivers/gpu/drm/imx/Kconfig
@@ -39,3 +39,5 @@ config DRM_IMX_HDMI
 	depends on DRM_IMX
 	help
 	  Choose this if you want to use HDMI on i.MX6.
+
+source "drivers/gpu/drm/imx/dcss/Kconfig"
diff --git a/drivers/gpu/drm/imx/Makefile b/drivers/gpu/drm/imx/Makefile
index 21cdcc2faabc..b644deffe948 100644
--- a/drivers/gpu/drm/imx/Makefile
+++ b/drivers/gpu/drm/imx/Makefile
@@ -9,3 +9,4 @@ obj-$(CONFIG_DRM_IMX_TVE) += imx-tve.o
 obj-$(CONFIG_DRM_IMX_LDB) += imx-ldb.o
 
 obj-$(CONFIG_DRM_IMX_HDMI) += dw_hdmi-imx.o
+obj-$(CONFIG_DRM_IMX_DCSS) += dcss/
diff --git a/drivers/gpu/drm/imx/dcss/Kconfig b/drivers/gpu/drm/imx/dcss/Kconfig
new file mode 100644
index 000000000000..2b17a964ff05
--- /dev/null
+++ b/drivers/gpu/drm/imx/dcss/Kconfig
@@ -0,0 +1,9 @@
+config DRM_IMX_DCSS
+	tristate "i.MX8MQ DCSS"
+	select IMX_IRQSTEER
+	select DRM_KMS_CMA_HELPER
+	select VIDEOMODE_HELPERS
+	depends on DRM && ARCH_MXC && ARM64
+	help
+	  Choose this if you have a NXP i.MX8MQ based system and want to use the
+	  Display Controller Subsystem. This option enables DCSS support.
diff --git a/drivers/gpu/drm/imx/dcss/Makefile b/drivers/gpu/drm/imx/dcss/Makefile
new file mode 100644
index 000000000000..8c7c8da42792
--- /dev/null
+++ b/drivers/gpu/drm/imx/dcss/Makefile
@@ -0,0 +1,6 @@
+imx-dcss-objs := dcss-drv.o dcss-dev.o dcss-blkctl.o dcss-ctxld.o dcss-dtg.o \
+				 dcss-ss.o dcss-dpr.o dcss-scaler.o dcss-kms.o dcss-crtc.o \
+				 dcss-plane.o
+
+obj-$(CONFIG_DRM_IMX_DCSS) += imx-dcss.o
+
diff --git a/drivers/gpu/drm/imx/dcss/dcss-blkctl.c b/drivers/gpu/drm/imx/dcss/dcss-blkctl.c
new file mode 100644
index 000000000000..c9b54bb2692d
--- /dev/null
+++ b/drivers/gpu/drm/imx/dcss/dcss-blkctl.c
@@ -0,0 +1,70 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Copyright 2019 NXP.
+ */
+
+#include <linux/device.h>
+#include <linux/of.h>
+#include <linux/slab.h>
+
+#include "dcss-dev.h"
+
+#define DCSS_BLKCTL_RESET_CTRL		0x00
+#define   B_CLK_RESETN			BIT(0)
+#define   APB_CLK_RESETN		BIT(1)
+#define   P_CLK_RESETN			BIT(2)
+#define   RTR_CLK_RESETN		BIT(4)
+#define DCSS_BLKCTL_CONTROL0		0x10
+#define   HDMI_MIPI_CLK_SEL		BIT(0)
+#define   DISPMIX_REFCLK_SEL_POS	4
+#define   DISPMIX_REFCLK_SEL_MASK	GENMASK(5, 4)
+#define   DISPMIX_PIXCLK_SEL		BIT(8)
+#define   HDMI_SRC_SECURE_EN		BIT(16)
+
+struct dcss_blkctl {
+	struct dcss_dev *dcss;
+	void __iomem *base_reg;
+};
+
+void dcss_blkctl_cfg(struct dcss_blkctl *blkctl)
+{
+	if (blkctl->dcss->hdmi_output)
+		dcss_writel(0, blkctl->base_reg + DCSS_BLKCTL_CONTROL0);
+	else
+		dcss_writel(DISPMIX_PIXCLK_SEL,
+			    blkctl->base_reg + DCSS_BLKCTL_CONTROL0);
+
+	dcss_set(B_CLK_RESETN | APB_CLK_RESETN | P_CLK_RESETN | RTR_CLK_RESETN,
+		 blkctl->base_reg + DCSS_BLKCTL_RESET_CTRL);
+}
+
+int dcss_blkctl_init(struct dcss_dev *dcss, unsigned long blkctl_base)
+{
+	struct dcss_blkctl *blkctl;
+
+	blkctl = kzalloc(sizeof(*blkctl), GFP_KERNEL);
+	if (!blkctl)
+		return -ENOMEM;
+
+	blkctl->base_reg = ioremap(blkctl_base, SZ_4K);
+	if (!blkctl->base_reg) {
+		dev_err(dcss->dev, "unable to remap BLK CTRL base\n");
+		kfree(blkctl);
+		return -ENOMEM;
+	}
+
+	dcss->blkctl = blkctl;
+	blkctl->dcss = dcss;
+
+	dcss_blkctl_cfg(blkctl);
+
+	return 0;
+}
+
+void dcss_blkctl_exit(struct dcss_blkctl *blkctl)
+{
+	if (blkctl->base_reg)
+		iounmap(blkctl->base_reg);
+
+	kfree(blkctl);
+}
diff --git a/drivers/gpu/drm/imx/dcss/dcss-crtc.c b/drivers/gpu/drm/imx/dcss/dcss-crtc.c
new file mode 100644
index 000000000000..36abff0890b2
--- /dev/null
+++ b/drivers/gpu/drm/imx/dcss/dcss-crtc.c
@@ -0,0 +1,219 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Copyright 2019 NXP.
+ */
+
+#include <drm/drm_atomic_helper.h>
+#include <drm/drm_vblank.h>
+#include <linux/platform_device.h>
+#include <linux/pm_runtime.h>
+
+#include "dcss-dev.h"
+#include "dcss-kms.h"
+
+static int dcss_enable_vblank(struct drm_crtc *crtc)
+{
+	struct dcss_crtc *dcss_crtc = container_of(crtc, struct dcss_crtc,
+						   base);
+	struct dcss_dev *dcss = crtc->dev->dev_private;
+
+	dcss_dtg_vblank_irq_enable(dcss->dtg, true);
+
+	dcss_dtg_ctxld_kick_irq_enable(dcss->dtg, true);
+
+	enable_irq(dcss_crtc->irq);
+
+	return 0;
+}
+
+static void dcss_disable_vblank(struct drm_crtc *crtc)
+{
+	struct dcss_crtc *dcss_crtc = container_of(crtc, struct dcss_crtc,
+						   base);
+	struct dcss_dev *dcss = dcss_crtc->base.dev->dev_private;
+
+	disable_irq_nosync(dcss_crtc->irq);
+
+	dcss_dtg_vblank_irq_enable(dcss->dtg, false);
+
+	if (dcss_crtc->disable_ctxld_kick_irq)
+		dcss_dtg_ctxld_kick_irq_enable(dcss->dtg, false);
+}
+
+static const struct drm_crtc_funcs dcss_crtc_funcs = {
+	.set_config = drm_atomic_helper_set_config,
+	.destroy = drm_crtc_cleanup,
+	.page_flip = drm_atomic_helper_page_flip,
+	.reset = drm_atomic_helper_crtc_reset,
+	.atomic_duplicate_state = drm_atomic_helper_crtc_duplicate_state,
+	.atomic_destroy_state = drm_atomic_helper_crtc_destroy_state,
+	.enable_vblank = dcss_enable_vblank,
+	.disable_vblank = dcss_disable_vblank,
+};
+
+static void dcss_crtc_atomic_begin(struct drm_crtc *crtc,
+				   struct drm_crtc_state *old_crtc_state)
+{
+	drm_crtc_vblank_on(crtc);
+}
+
+static void dcss_crtc_atomic_flush(struct drm_crtc *crtc,
+				   struct drm_crtc_state *old_crtc_state)
+{
+	struct dcss_crtc *dcss_crtc = container_of(crtc, struct dcss_crtc,
+						   base);
+	struct dcss_dev *dcss = dcss_crtc->base.dev->dev_private;
+
+	spin_lock_irq(&crtc->dev->event_lock);
+	if (crtc->state->event) {
+		WARN_ON(drm_crtc_vblank_get(crtc));
+		drm_crtc_arm_vblank_event(crtc, crtc->state->event);
+		crtc->state->event = NULL;
+	}
+	spin_unlock_irq(&crtc->dev->event_lock);
+
+	if (dcss_dtg_is_enabled(dcss->dtg))
+		dcss_ctxld_enable(dcss->ctxld);
+}
+
+static void dcss_crtc_atomic_enable(struct drm_crtc *crtc,
+				    struct drm_crtc_state *old_crtc_state)
+{
+	struct dcss_crtc *dcss_crtc = container_of(crtc, struct dcss_crtc,
+						   base);
+	struct dcss_dev *dcss = dcss_crtc->base.dev->dev_private;
+	struct drm_display_mode *mode = &crtc->state->adjusted_mode;
+	struct drm_display_mode *old_mode = &old_crtc_state->adjusted_mode;
+	struct videomode vm;
+
+	drm_display_mode_to_videomode(mode, &vm);
+
+	pm_runtime_get_sync(dcss->dev);
+
+	vm.pixelclock = mode->crtc_clock * 1000;
+
+	dcss_ss_subsam_set(dcss->ss);
+	dcss_dtg_css_set(dcss->dtg);
+
+	if (!drm_mode_equal(mode, old_mode) || !old_crtc_state->active) {
+		dcss_dtg_sync_set(dcss->dtg, &vm);
+		dcss_ss_sync_set(dcss->ss, &vm,
+				 mode->flags & DRM_MODE_FLAG_PHSYNC,
+				 mode->flags & DRM_MODE_FLAG_PVSYNC);
+	}
+
+	dcss_enable_dtg_and_ss(dcss);
+
+	dcss_ctxld_enable(dcss->ctxld);
+
+	/* Allow CTXLD kick interrupt to be disabled when VBLANK is disabled. */
+	dcss_crtc->disable_ctxld_kick_irq = true;
+}
+
+static void dcss_crtc_atomic_disable(struct drm_crtc *crtc,
+				     struct drm_crtc_state *old_crtc_state)
+{
+	struct dcss_crtc *dcss_crtc = container_of(crtc, struct dcss_crtc,
+						   base);
+	struct dcss_dev *dcss = dcss_crtc->base.dev->dev_private;
+	struct drm_display_mode *mode = &crtc->state->adjusted_mode;
+	struct drm_display_mode *old_mode = &old_crtc_state->adjusted_mode;
+
+	drm_atomic_helper_disable_planes_on_crtc(old_crtc_state, false);
+
+	spin_lock_irq(&crtc->dev->event_lock);
+	if (crtc->state->event) {
+		drm_crtc_send_vblank_event(crtc, crtc->state->event);
+		crtc->state->event = NULL;
+	}
+	spin_unlock_irq(&crtc->dev->event_lock);
+
+	dcss_dtg_ctxld_kick_irq_enable(dcss->dtg, true);
+
+	reinit_completion(&dcss->disable_completion);
+
+	dcss_disable_dtg_and_ss(dcss);
+
+	dcss_ctxld_enable(dcss->ctxld);
+
+	if (!drm_mode_equal(mode, old_mode) || !crtc->state->active)
+		if (!wait_for_completion_timeout(&dcss->disable_completion,
+						 msecs_to_jiffies(100)))
+			dev_err(dcss->dev, "Shutting off DTG timed out.\n");
+
+	/*
+	 * Do not shut off CTXLD kick interrupt when shutting VBLANK off. It
+	 * will be needed to commit the last changes, before going to suspend.
+	 */
+	dcss_crtc->disable_ctxld_kick_irq = false;
+
+	drm_crtc_vblank_off(crtc);
+
+	pm_runtime_mark_last_busy(dcss->dev);
+	pm_runtime_put_autosuspend(dcss->dev);
+}
+
+static const struct drm_crtc_helper_funcs dcss_helper_funcs = {
+	.atomic_begin = dcss_crtc_atomic_begin,
+	.atomic_flush = dcss_crtc_atomic_flush,
+	.atomic_enable = dcss_crtc_atomic_enable,
+	.atomic_disable = dcss_crtc_atomic_disable,
+};
+
+static irqreturn_t dcss_crtc_irq_handler(int irq, void *dev_id)
+{
+	struct dcss_crtc *dcss_crtc = dev_id;
+	struct dcss_dev *dcss = dcss_crtc->base.dev->dev_private;
+
+	if (!dcss_dtg_vblank_irq_valid(dcss->dtg))
+		return IRQ_NONE;
+
+	if (dcss_ctxld_is_flushed(dcss->ctxld))
+		drm_crtc_handle_vblank(&dcss_crtc->base);
+
+	dcss_dtg_vblank_irq_clear(dcss->dtg);
+
+	return IRQ_HANDLED;
+}
+
+int dcss_crtc_init(struct dcss_crtc *crtc, struct drm_device *drm)
+{
+	struct dcss_dev *dcss = drm->dev_private;
+	struct platform_device *pdev = to_platform_device(dcss->dev);
+	int ret;
+
+	crtc->plane[0] = dcss_plane_init(drm, drm_crtc_mask(&crtc->base),
+					 DRM_PLANE_TYPE_PRIMARY, 0);
+	if (IS_ERR(crtc->plane[0]))
+		return PTR_ERR(crtc->plane[0]);
+
+	crtc->base.port = dcss->of_port;
+
+	drm_crtc_helper_add(&crtc->base, &dcss_helper_funcs);
+	ret = drm_crtc_init_with_planes(drm, &crtc->base, &crtc->plane[0]->base,
+					NULL, &dcss_crtc_funcs, NULL);
+	if (ret) {
+		dev_err(dcss->dev, "failed to init crtc\n");
+		return ret;
+	}
+
+	crtc->irq = platform_get_irq_byname(pdev, "vblank");
+	if (crtc->irq < 0)
+		return crtc->irq;
+
+	ret = request_irq(crtc->irq, dcss_crtc_irq_handler,
+			  0, "dcss_drm", crtc);
+	if (ret) {
+		dev_err(dcss->dev, "irq request failed with %d.\n", ret);
+		return ret;
+	}
+
+	disable_irq(crtc->irq);
+
+	return 0;
+}
+
+void dcss_crtc_deinit(struct dcss_crtc *crtc, struct drm_device *drm)
+{
+	free_irq(crtc->irq, crtc);
+}
diff --git a/drivers/gpu/drm/imx/dcss/dcss-ctxld.c b/drivers/gpu/drm/imx/dcss/dcss-ctxld.c
new file mode 100644
index 000000000000..3a84cb3209c4
--- /dev/null
+++ b/drivers/gpu/drm/imx/dcss/dcss-ctxld.c
@@ -0,0 +1,424 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Copyright 2019 NXP.
+ */
+
+#include <linux/delay.h>
+#include <linux/dma-mapping.h>
+#include <linux/interrupt.h>
+#include <linux/platform_device.h>
+#include <linux/slab.h>
+
+#include "dcss-dev.h"
+
+#define DCSS_CTXLD_CONTROL_STATUS	0x0
+#define   CTXLD_ENABLE			BIT(0)
+#define   ARB_SEL			BIT(1)
+#define   RD_ERR_EN			BIT(2)
+#define   DB_COMP_EN			BIT(3)
+#define   SB_HP_COMP_EN			BIT(4)
+#define   SB_LP_COMP_EN			BIT(5)
+#define   DB_PEND_SB_REC_EN		BIT(6)
+#define   SB_PEND_DISP_ACTIVE_EN	BIT(7)
+#define   AHB_ERR_EN			BIT(8)
+#define   RD_ERR			BIT(16)
+#define   DB_COMP			BIT(17)
+#define   SB_HP_COMP			BIT(18)
+#define   SB_LP_COMP			BIT(19)
+#define   DB_PEND_SB_REC		BIT(20)
+#define   SB_PEND_DISP_ACTIVE		BIT(21)
+#define   AHB_ERR			BIT(22)
+#define DCSS_CTXLD_DB_BASE_ADDR		0x10
+#define DCSS_CTXLD_DB_COUNT		0x14
+#define DCSS_CTXLD_SB_BASE_ADDR		0x18
+#define DCSS_CTXLD_SB_COUNT		0x1C
+#define   SB_HP_COUNT_POS		0
+#define   SB_HP_COUNT_MASK		0xffff
+#define   SB_LP_COUNT_POS		16
+#define   SB_LP_COUNT_MASK		0xffff0000
+#define DCSS_AHB_ERR_ADDR		0x20
+
+#define CTXLD_IRQ_COMPLETION		(DB_COMP | SB_HP_COMP | SB_LP_COMP)
+#define CTXLD_IRQ_ERROR			(RD_ERR | DB_PEND_SB_REC | AHB_ERR)
+
+/* The following sizes are in context loader entries, 8 bytes each. */
+#define CTXLD_DB_CTX_ENTRIES		1024	/* max 65536 */
+#define CTXLD_SB_LP_CTX_ENTRIES		10240	/* max 65536 */
+#define CTXLD_SB_HP_CTX_ENTRIES		20000	/* max 65536 */
+#define CTXLD_SB_CTX_ENTRIES		(CTXLD_SB_LP_CTX_ENTRIES + \
+					 CTXLD_SB_HP_CTX_ENTRIES)
+
+/* Sizes, in entries, of the DB, SB_HP and SB_LP context regions. */
+static u16 dcss_ctxld_ctx_size[3] = {
+	CTXLD_DB_CTX_ENTRIES,
+	CTXLD_SB_HP_CTX_ENTRIES,
+	CTXLD_SB_LP_CTX_ENTRIES
+};
+
+/* this represents an entry in the context loader map */
+struct dcss_ctxld_item {
+	u32 val;
+	u32 ofs;
+};
+
+#define CTX_ITEM_SIZE			sizeof(struct dcss_ctxld_item)
+
+struct dcss_ctxld {
+	struct device *dev;
+	void __iomem *ctxld_reg;
+	int irq;
+	bool irq_en;
+
+	struct dcss_ctxld_item *db[2];
+	struct dcss_ctxld_item *sb_hp[2];
+	struct dcss_ctxld_item *sb_lp[2];
+
+	dma_addr_t db_paddr[2];
+	dma_addr_t sb_paddr[2];
+
+	u16 ctx_size[2][3]; /* holds the sizes of DB, SB_HP and SB_LP ctx */
+	u8 current_ctx;
+
+	bool in_use;
+	bool armed;
+
+	spinlock_t lock; /* protects concurent access to private data */
+};
+
+static irqreturn_t dcss_ctxld_irq_handler(int irq, void *data)
+{
+	struct dcss_ctxld *ctxld = data;
+	struct dcss_dev *dcss = dcss_drv_dev_to_dcss(ctxld->dev);
+	u32 irq_status;
+
+	irq_status = dcss_readl(ctxld->ctxld_reg + DCSS_CTXLD_CONTROL_STATUS);
+
+	if (irq_status & CTXLD_IRQ_COMPLETION &&
+	    !(irq_status & CTXLD_ENABLE) && ctxld->in_use) {
+		ctxld->in_use = false;
+
+		if (dcss && dcss->disable_callback)
+			dcss->disable_callback(dcss);
+	} else if (irq_status & CTXLD_IRQ_ERROR) {
+		/*
+		 * Except for throwing an error message and clearing the status
+		 * register, there's not much we can do here.
+		 */
+		dev_err(ctxld->dev, "ctxld: error encountered: %08x\n",
+			irq_status);
+		dev_err(ctxld->dev, "ctxld: db=%d, sb_hp=%d, sb_lp=%d\n",
+			ctxld->ctx_size[ctxld->current_ctx ^ 1][CTX_DB],
+			ctxld->ctx_size[ctxld->current_ctx ^ 1][CTX_SB_HP],
+			ctxld->ctx_size[ctxld->current_ctx ^ 1][CTX_SB_LP]);
+	}
+
+	dcss_clr(irq_status & (CTXLD_IRQ_ERROR | CTXLD_IRQ_COMPLETION),
+		 ctxld->ctxld_reg + DCSS_CTXLD_CONTROL_STATUS);
+
+	return IRQ_HANDLED;
+}
+
+static int dcss_ctxld_irq_config(struct dcss_ctxld *ctxld,
+				 struct platform_device *pdev)
+{
+	int ret;
+
+	ctxld->irq = platform_get_irq_byname(pdev, "ctxld");
+	if (ctxld->irq < 0)
+		return ctxld->irq;
+
+	ret = request_irq(ctxld->irq, dcss_ctxld_irq_handler,
+			  0, "dcss_ctxld", ctxld);
+	if (ret) {
+		dev_err(ctxld->dev, "ctxld: irq request failed.\n");
+		return ret;
+	}
+
+	ctxld->irq_en = true;
+
+	return 0;
+}
+
+static void dcss_ctxld_hw_cfg(struct dcss_ctxld *ctxld)
+{
+	dcss_writel(RD_ERR_EN | SB_HP_COMP_EN |
+		    DB_PEND_SB_REC_EN | AHB_ERR_EN | RD_ERR | AHB_ERR,
+		    ctxld->ctxld_reg + DCSS_CTXLD_CONTROL_STATUS);
+}
+
+static void dcss_ctxld_free_ctx(struct dcss_ctxld *ctxld)
+{
+	struct dcss_ctxld_item *ctx;
+	int i;
+
+	for (i = 0; i < 2; i++) {
+		if (ctxld->db[i]) {
+			dma_free_coherent(ctxld->dev,
+					  CTXLD_DB_CTX_ENTRIES * sizeof(*ctx),
+					  ctxld->db[i], ctxld->db_paddr[i]);
+			ctxld->db[i] = NULL;
+			ctxld->db_paddr[i] = 0;
+		}
+
+		if (ctxld->sb_hp[i]) {
+			dma_free_coherent(ctxld->dev,
+					  CTXLD_SB_CTX_ENTRIES * sizeof(*ctx),
+					  ctxld->sb_hp[i], ctxld->sb_paddr[i]);
+			ctxld->sb_hp[i] = NULL;
+			ctxld->sb_paddr[i] = 0;
+		}
+	}
+}
+
+static int dcss_ctxld_alloc_ctx(struct dcss_ctxld *ctxld)
+{
+	struct dcss_ctxld_item *ctx;
+	int i;
+
+	for (i = 0; i < 2; i++) {
+		ctx = dma_alloc_coherent(ctxld->dev,
+					 CTXLD_DB_CTX_ENTRIES * sizeof(*ctx),
+					 &ctxld->db_paddr[i], GFP_KERNEL);
+		if (!ctx)
+			return -ENOMEM;
+
+		ctxld->db[i] = ctx;
+
+		ctx = dma_alloc_coherent(ctxld->dev,
+					 CTXLD_SB_CTX_ENTRIES * sizeof(*ctx),
+					 &ctxld->sb_paddr[i], GFP_KERNEL);
+		if (!ctx)
+			return -ENOMEM;
+
+		ctxld->sb_hp[i] = ctx;
+		ctxld->sb_lp[i] = ctx + CTXLD_SB_HP_CTX_ENTRIES;
+	}
+
+	return 0;
+}
+
+int dcss_ctxld_init(struct dcss_dev *dcss, unsigned long ctxld_base)
+{
+	struct dcss_ctxld *ctxld;
+	int ret;
+
+	ctxld = kzalloc(sizeof(*ctxld), GFP_KERNEL);
+	if (!ctxld)
+		return -ENOMEM;
+
+	dcss->ctxld = ctxld;
+	ctxld->dev = dcss->dev;
+
+	spin_lock_init(&ctxld->lock);
+
+	ret = dcss_ctxld_alloc_ctx(ctxld);
+	if (ret) {
+		dev_err(dcss->dev, "ctxld: cannot allocate context memory.\n");
+		goto err;
+	}
+
+	ctxld->ctxld_reg = ioremap(ctxld_base, SZ_4K);
+	if (!ctxld->ctxld_reg) {
+		dev_err(dcss->dev, "ctxld: unable to remap ctxld base\n");
+		ret = -ENOMEM;
+		goto err;
+	}
+
+	ret = dcss_ctxld_irq_config(ctxld, to_platform_device(dcss->dev));
+	if (ret)
+		goto err_irq;
+
+	dcss_ctxld_hw_cfg(ctxld);
+
+	return 0;
+
+err_irq:
+	iounmap(ctxld->ctxld_reg);
+
+err:
+	dcss_ctxld_free_ctx(ctxld);
+	kfree(ctxld);
+
+	return ret;
+}
+
+void dcss_ctxld_exit(struct dcss_ctxld *ctxld)
+{
+	free_irq(ctxld->irq, ctxld);
+
+	if (ctxld->ctxld_reg)
+		iounmap(ctxld->ctxld_reg);
+
+	dcss_ctxld_free_ctx(ctxld);
+	kfree(ctxld);
+}
+
+static int dcss_ctxld_enable_locked(struct dcss_ctxld *ctxld)
+{
+	int curr_ctx = ctxld->current_ctx;
+	u32 db_base, sb_base, sb_count;
+	u32 sb_hp_cnt, sb_lp_cnt, db_cnt;
+	struct dcss_dev *dcss = dcss_drv_dev_to_dcss(ctxld->dev);
+
+	if (!dcss)
+		return 0;
+
+	dcss_dpr_write_sysctrl(dcss->dpr);
+
+	dcss_scaler_write_sclctrl(dcss->scaler);
+
+	sb_hp_cnt = ctxld->ctx_size[curr_ctx][CTX_SB_HP];
+	sb_lp_cnt = ctxld->ctx_size[curr_ctx][CTX_SB_LP];
+	db_cnt = ctxld->ctx_size[curr_ctx][CTX_DB];
+
+	/* make sure SB_LP context area comes after SB_HP */
+	if (sb_lp_cnt &&
+	    ctxld->sb_lp[curr_ctx] != ctxld->sb_hp[curr_ctx] + sb_hp_cnt) {
+		struct dcss_ctxld_item *sb_lp_adjusted;
+
+		sb_lp_adjusted = ctxld->sb_hp[curr_ctx] + sb_hp_cnt;
+
+		memcpy(sb_lp_adjusted, ctxld->sb_lp[curr_ctx],
+		       sb_lp_cnt * CTX_ITEM_SIZE);
+	}
+
+	db_base = db_cnt ? ctxld->db_paddr[curr_ctx] : 0;
+
+	dcss_writel(db_base, ctxld->ctxld_reg + DCSS_CTXLD_DB_BASE_ADDR);
+	dcss_writel(db_cnt, ctxld->ctxld_reg + DCSS_CTXLD_DB_COUNT);
+
+	if (sb_hp_cnt)
+		sb_count = ((sb_hp_cnt << SB_HP_COUNT_POS) & SB_HP_COUNT_MASK) |
+			   ((sb_lp_cnt << SB_LP_COUNT_POS) & SB_LP_COUNT_MASK);
+	else
+		sb_count = (sb_lp_cnt << SB_HP_COUNT_POS) & SB_HP_COUNT_MASK;
+
+	sb_base = sb_count ? ctxld->sb_paddr[curr_ctx] : 0;
+
+	dcss_writel(sb_base, ctxld->ctxld_reg + DCSS_CTXLD_SB_BASE_ADDR);
+	dcss_writel(sb_count, ctxld->ctxld_reg + DCSS_CTXLD_SB_COUNT);
+
+	/* enable the context loader */
+	dcss_set(CTXLD_ENABLE, ctxld->ctxld_reg + DCSS_CTXLD_CONTROL_STATUS);
+
+	ctxld->in_use = true;
+
+	/*
+	 * Toggle the current context to the alternate one so that any updates
+	 * in the modules' settings take place there.
+	 */
+	ctxld->current_ctx ^= 1;
+
+	ctxld->ctx_size[ctxld->current_ctx][CTX_DB] = 0;
+	ctxld->ctx_size[ctxld->current_ctx][CTX_SB_HP] = 0;
+	ctxld->ctx_size[ctxld->current_ctx][CTX_SB_LP] = 0;
+
+	return 0;
+}
+
+int dcss_ctxld_enable(struct dcss_ctxld *ctxld)
+{
+	spin_lock_irq(&ctxld->lock);
+	ctxld->armed = true;
+	spin_unlock_irq(&ctxld->lock);
+
+	return 0;
+}
+
+void dcss_ctxld_kick(struct dcss_ctxld *ctxld)
+{
+	unsigned long flags;
+
+	spin_lock_irqsave(&ctxld->lock, flags);
+	if (ctxld->armed && !ctxld->in_use) {
+		ctxld->armed = false;
+		dcss_ctxld_enable_locked(ctxld);
+	}
+	spin_unlock_irqrestore(&ctxld->lock, flags);
+}
+
+void dcss_ctxld_write_irqsafe(struct dcss_ctxld *ctxld, u32 ctx_id, u32 val,
+			      u32 reg_ofs)
+{
+	int curr_ctx = ctxld->current_ctx;
+	struct dcss_ctxld_item *ctx[] = {
+		[CTX_DB] = ctxld->db[curr_ctx],
+		[CTX_SB_HP] = ctxld->sb_hp[curr_ctx],
+		[CTX_SB_LP] = ctxld->sb_lp[curr_ctx]
+	};
+	int item_idx = ctxld->ctx_size[curr_ctx][ctx_id];
+
+	if (item_idx + 1 > dcss_ctxld_ctx_size[ctx_id]) {
+		WARN_ON(1);
+		return;
+	}
+
+	ctx[ctx_id][item_idx].val = val;
+	ctx[ctx_id][item_idx].ofs = reg_ofs;
+	ctxld->ctx_size[curr_ctx][ctx_id] += 1;
+}
+
+void dcss_ctxld_write(struct dcss_ctxld *ctxld, u32 ctx_id,
+		      u32 val, u32 reg_ofs)
+{
+	spin_lock_irq(&ctxld->lock);
+	dcss_ctxld_write_irqsafe(ctxld, ctx_id, val, reg_ofs);
+	spin_unlock_irq(&ctxld->lock);
+}
+
+bool dcss_ctxld_is_flushed(struct dcss_ctxld *ctxld)
+{
+	return ctxld->ctx_size[ctxld->current_ctx][CTX_DB] == 0 &&
+		ctxld->ctx_size[ctxld->current_ctx][CTX_SB_HP] == 0 &&
+		ctxld->ctx_size[ctxld->current_ctx][CTX_SB_LP] == 0;
+}
+
+int dcss_ctxld_resume(struct dcss_ctxld *ctxld)
+{
+	dcss_ctxld_hw_cfg(ctxld);
+
+	if (!ctxld->irq_en) {
+		enable_irq(ctxld->irq);
+		ctxld->irq_en = true;
+	}
+
+	return 0;
+}
+
+int dcss_ctxld_suspend(struct dcss_ctxld *ctxld)
+{
+	int ret = 0;
+	unsigned long timeout = jiffies + msecs_to_jiffies(500);
+
+	if (!dcss_ctxld_is_flushed(ctxld)) {
+		dcss_ctxld_kick(ctxld);
+
+		while (!time_after(jiffies, timeout) && ctxld->in_use)
+			msleep(20);
+
+		if (time_after(jiffies, timeout))
+			return -ETIMEDOUT;
+	}
+
+	spin_lock_irq(&ctxld->lock);
+
+	if (ctxld->irq_en) {
+		disable_irq_nosync(ctxld->irq);
+		ctxld->irq_en = false;
+	}
+
+	/* reset context region and sizes */
+	ctxld->current_ctx = 0;
+	ctxld->ctx_size[0][CTX_DB] = 0;
+	ctxld->ctx_size[0][CTX_SB_HP] = 0;
+	ctxld->ctx_size[0][CTX_SB_LP] = 0;
+
+	spin_unlock_irq(&ctxld->lock);
+
+	return ret;
+}
+
+void dcss_ctxld_assert_locked(struct dcss_ctxld *ctxld)
+{
+	lockdep_assert_held(&ctxld->lock);
+}
diff --git a/drivers/gpu/drm/imx/dcss/dcss-dev.c b/drivers/gpu/drm/imx/dcss/dcss-dev.c
new file mode 100644
index 000000000000..c849533ca83e
--- /dev/null
+++ b/drivers/gpu/drm/imx/dcss/dcss-dev.c
@@ -0,0 +1,325 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Copyright 2019 NXP.
+ */
+
+#include <linux/clk.h>
+#include <linux/of_device.h>
+#include <linux/of_graph.h>
+#include <linux/pm_runtime.h>
+#include <linux/slab.h>
+#include <drm/drm_bridge_connector.h>
+#include <drm/drm_device.h>
+#include <drm/drm_modeset_helper.h>
+
+#include "dcss-dev.h"
+#include "dcss-kms.h"
+
+static void dcss_clocks_enable(struct dcss_dev *dcss)
+{
+	clk_prepare_enable(dcss->axi_clk);
+	clk_prepare_enable(dcss->apb_clk);
+	clk_prepare_enable(dcss->rtrm_clk);
+	clk_prepare_enable(dcss->dtrc_clk);
+	clk_prepare_enable(dcss->pix_clk);
+}
+
+static void dcss_clocks_disable(struct dcss_dev *dcss)
+{
+	clk_disable_unprepare(dcss->pix_clk);
+	clk_disable_unprepare(dcss->dtrc_clk);
+	clk_disable_unprepare(dcss->rtrm_clk);
+	clk_disable_unprepare(dcss->apb_clk);
+	clk_disable_unprepare(dcss->axi_clk);
+}
+
+static void dcss_disable_dtg_and_ss_cb(void *data)
+{
+	struct dcss_dev *dcss = data;
+
+	dcss->disable_callback = NULL;
+
+	dcss_ss_shutoff(dcss->ss);
+	dcss_dtg_shutoff(dcss->dtg);
+
+	complete(&dcss->disable_completion);
+}
+
+void dcss_disable_dtg_and_ss(struct dcss_dev *dcss)
+{
+	dcss->disable_callback = dcss_disable_dtg_and_ss_cb;
+}
+
+void dcss_enable_dtg_and_ss(struct dcss_dev *dcss)
+{
+	if (dcss->disable_callback)
+		dcss->disable_callback = NULL;
+
+	dcss_dtg_enable(dcss->dtg);
+	dcss_ss_enable(dcss->ss);
+}
+
+static int dcss_submodules_init(struct dcss_dev *dcss)
+{
+	int ret = 0;
+	u32 base_addr = dcss->start_addr;
+	const struct dcss_type_data *devtype = dcss->devtype;
+
+	dcss_clocks_enable(dcss);
+
+	ret = dcss_blkctl_init(dcss, base_addr + devtype->blkctl_ofs);
+	if (ret)
+		return ret;
+
+	ret = dcss_ctxld_init(dcss, base_addr + devtype->ctxld_ofs);
+	if (ret)
+		goto ctxld_err;
+
+	ret = dcss_dtg_init(dcss, base_addr + devtype->dtg_ofs);
+	if (ret)
+		goto dtg_err;
+
+	ret = dcss_ss_init(dcss, base_addr + devtype->ss_ofs);
+	if (ret)
+		goto ss_err;
+
+	ret = dcss_dpr_init(dcss, base_addr + devtype->dpr_ofs);
+	if (ret)
+		goto dpr_err;
+
+	ret = dcss_scaler_init(dcss, base_addr + devtype->scaler_ofs);
+	if (ret)
+		goto scaler_err;
+
+	dcss_clocks_disable(dcss);
+
+	return 0;
+
+scaler_err:
+	dcss_dpr_exit(dcss->dpr);
+
+dpr_err:
+	dcss_ss_exit(dcss->ss);
+
+ss_err:
+	dcss_dtg_exit(dcss->dtg);
+
+dtg_err:
+	dcss_ctxld_exit(dcss->ctxld);
+
+ctxld_err:
+	dcss_blkctl_exit(dcss->blkctl);
+
+	dcss_clocks_disable(dcss);
+
+	return ret;
+}
+
+static void dcss_submodules_stop(struct dcss_dev *dcss)
+{
+	dcss_clocks_enable(dcss);
+	dcss_scaler_exit(dcss->scaler);
+	dcss_dpr_exit(dcss->dpr);
+	dcss_ss_exit(dcss->ss);
+	dcss_dtg_exit(dcss->dtg);
+	dcss_ctxld_exit(dcss->ctxld);
+	dcss_blkctl_exit(dcss->blkctl);
+	dcss_clocks_disable(dcss);
+}
+
+static int dcss_clks_init(struct dcss_dev *dcss)
+{
+	int i;
+	struct {
+		const char *id;
+		struct clk **clk;
+	} clks[] = {
+		{"apb",   &dcss->apb_clk},
+		{"axi",   &dcss->axi_clk},
+		{"pix",   &dcss->pix_clk},
+		{"rtrm",  &dcss->rtrm_clk},
+		{"dtrc",  &dcss->dtrc_clk},
+	};
+
+	for (i = 0; i < ARRAY_SIZE(clks); i++) {
+		*clks[i].clk = devm_clk_get(dcss->dev, clks[i].id);
+		if (IS_ERR(*clks[i].clk)) {
+			dev_err(dcss->dev, "failed to get %s clock\n",
+				clks[i].id);
+			return PTR_ERR(*clks[i].clk);
+		}
+	}
+
+	return 0;
+}
+
+static void dcss_clks_release(struct dcss_dev *dcss)
+{
+	devm_clk_put(dcss->dev, dcss->dtrc_clk);
+	devm_clk_put(dcss->dev, dcss->rtrm_clk);
+	devm_clk_put(dcss->dev, dcss->pix_clk);
+	devm_clk_put(dcss->dev, dcss->axi_clk);
+	devm_clk_put(dcss->dev, dcss->apb_clk);
+}
+
+struct dcss_dev *dcss_dev_create(struct device *dev, bool hdmi_output)
+{
+	struct platform_device *pdev = to_platform_device(dev);
+	int ret;
+	struct resource *res;
+	struct dcss_dev *dcss;
+	const struct dcss_type_data *devtype;
+
+	devtype = of_device_get_match_data(dev);
+	if (!devtype) {
+		dev_err(dev, "no device match found\n");
+		return ERR_PTR(-ENODEV);
+	}
+
+	res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
+	if (!res) {
+		dev_err(dev, "cannot get memory resource\n");
+		return ERR_PTR(-EINVAL);
+	}
+
+	dcss = kzalloc(sizeof(*dcss), GFP_KERNEL);
+	if (!dcss)
+		return ERR_PTR(-ENOMEM);
+
+	dcss->dev = dev;
+	dcss->devtype = devtype;
+	dcss->hdmi_output = hdmi_output;
+
+	ret = dcss_clks_init(dcss);
+	if (ret) {
+		dev_err(dev, "clocks initialization failed\n");
+		goto err;
+	}
+
+	dcss->of_port = of_graph_get_port_by_id(dev->of_node, 0);
+	if (!dcss->of_port) {
+		dev_err(dev, "no port@0 node in %s\n", dev->of_node->full_name);
+		ret = -ENODEV;
+		goto clks_err;
+	}
+
+	dcss->start_addr = res->start;
+
+	ret = dcss_submodules_init(dcss);
+	if (ret) {
+		dev_err(dev, "submodules initialization failed\n");
+		goto clks_err;
+	}
+
+	init_completion(&dcss->disable_completion);
+
+	pm_runtime_set_autosuspend_delay(dev, 100);
+	pm_runtime_use_autosuspend(dev);
+	pm_runtime_set_suspended(dev);
+	pm_runtime_allow(dev);
+	pm_runtime_enable(dev);
+
+	return dcss;
+
+clks_err:
+	dcss_clks_release(dcss);
+
+err:
+	kfree(dcss);
+
+	return ERR_PTR(ret);
+}
+
+void dcss_dev_destroy(struct dcss_dev *dcss)
+{
+	if (!pm_runtime_suspended(dcss->dev)) {
+		dcss_ctxld_suspend(dcss->ctxld);
+		dcss_clocks_disable(dcss);
+	}
+
+	pm_runtime_disable(dcss->dev);
+
+	dcss_submodules_stop(dcss);
+
+	dcss_clks_release(dcss);
+
+	kfree(dcss);
+}
+
+#ifdef CONFIG_PM_SLEEP
+int dcss_dev_suspend(struct device *dev)
+{
+	struct dcss_dev *dcss = dcss_drv_dev_to_dcss(dev);
+	struct drm_device *ddev = dcss_drv_dev_to_drm(dev);
+	struct dcss_kms_dev *kms = container_of(ddev, struct dcss_kms_dev, base);
+	int ret;
+
+	drm_bridge_connector_disable_hpd(kms->connector);
+
+	drm_mode_config_helper_suspend(ddev);
+
+	if (pm_runtime_suspended(dev))
+		return 0;
+
+	ret = dcss_ctxld_suspend(dcss->ctxld);
+	if (ret)
+		return ret;
+
+	dcss_clocks_disable(dcss);
+
+	return 0;
+}
+
+int dcss_dev_resume(struct device *dev)
+{
+	struct dcss_dev *dcss = dcss_drv_dev_to_dcss(dev);
+	struct drm_device *ddev = dcss_drv_dev_to_drm(dev);
+	struct dcss_kms_dev *kms = container_of(ddev, struct dcss_kms_dev, base);
+
+	if (pm_runtime_suspended(dev)) {
+		drm_mode_config_helper_resume(ddev);
+		return 0;
+	}
+
+	dcss_clocks_enable(dcss);
+
+	dcss_blkctl_cfg(dcss->blkctl);
+
+	dcss_ctxld_resume(dcss->ctxld);
+
+	drm_mode_config_helper_resume(ddev);
+
+	drm_bridge_connector_enable_hpd(kms->connector);
+
+	return 0;
+}
+#endif /* CONFIG_PM_SLEEP */
+
+#ifdef CONFIG_PM
+int dcss_dev_runtime_suspend(struct device *dev)
+{
+	struct dcss_dev *dcss = dcss_drv_dev_to_dcss(dev);
+	int ret;
+
+	ret = dcss_ctxld_suspend(dcss->ctxld);
+	if (ret)
+		return ret;
+
+	dcss_clocks_disable(dcss);
+
+	return 0;
+}
+
+int dcss_dev_runtime_resume(struct device *dev)
+{
+	struct dcss_dev *dcss = dcss_drv_dev_to_dcss(dev);
+
+	dcss_clocks_enable(dcss);
+
+	dcss_blkctl_cfg(dcss->blkctl);
+
+	dcss_ctxld_resume(dcss->ctxld);
+
+	return 0;
+}
+#endif /* CONFIG_PM */
diff --git a/drivers/gpu/drm/imx/dcss/dcss-dev.h b/drivers/gpu/drm/imx/dcss/dcss-dev.h
new file mode 100644
index 000000000000..c642ae17837f
--- /dev/null
+++ b/drivers/gpu/drm/imx/dcss/dcss-dev.h
@@ -0,0 +1,177 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+/*
+ * Copyright 2019 NXP.
+ */
+
+#ifndef __DCSS_PRV_H__
+#define __DCSS_PRV_H__
+
+#include <drm/drm_fourcc.h>
+#include <linux/io.h>
+#include <video/videomode.h>
+
+#define SET			0x04
+#define CLR			0x08
+#define TGL			0x0C
+
+#define dcss_writel(v, c)	writel((v), (c))
+#define dcss_readl(c)		readl(c)
+#define dcss_set(v, c)		writel((v), (c) + SET)
+#define dcss_clr(v, c)		writel((v), (c) + CLR)
+#define dcss_toggle(v, c)	writel((v), (c) + TGL)
+
+static inline void dcss_update(u32 v, u32 m, void __iomem *c)
+{
+	writel((readl(c) & ~(m)) | (v), (c));
+}
+
+#define DCSS_DBG_REG(reg)	{.name = #reg, .ofs = reg}
+
+enum {
+	DCSS_IMX8MQ = 0,
+};
+
+struct dcss_type_data {
+	const char *name;
+	u32 blkctl_ofs;
+	u32 ctxld_ofs;
+	u32 rdsrc_ofs;
+	u32 wrscl_ofs;
+	u32 dtg_ofs;
+	u32 scaler_ofs;
+	u32 ss_ofs;
+	u32 dpr_ofs;
+	u32 dtrc_ofs;
+	u32 dec400d_ofs;
+	u32 hdr10_ofs;
+};
+
+struct dcss_debug_reg {
+	char *name;
+	u32 ofs;
+};
+
+enum dcss_ctxld_ctx_type {
+	CTX_DB,
+	CTX_SB_HP, /* high-priority */
+	CTX_SB_LP, /* low-priority  */
+};
+
+struct dcss_dev {
+	struct device *dev;
+	const struct dcss_type_data *devtype;
+	struct device_node *of_port;
+
+	u32 start_addr;
+
+	struct dcss_blkctl *blkctl;
+	struct dcss_ctxld *ctxld;
+	struct dcss_dpr *dpr;
+	struct dcss_dtg *dtg;
+	struct dcss_ss *ss;
+	struct dcss_hdr10 *hdr10;
+	struct dcss_scaler *scaler;
+	struct dcss_dtrc *dtrc;
+	struct dcss_dec400d *dec400d;
+	struct dcss_wrscl *wrscl;
+	struct dcss_rdsrc *rdsrc;
+
+	struct clk *apb_clk;
+	struct clk *axi_clk;
+	struct clk *pix_clk;
+	struct clk *rtrm_clk;
+	struct clk *dtrc_clk;
+	struct clk *pll_src_clk;
+	struct clk *pll_phy_ref_clk;
+
+	bool hdmi_output;
+
+	void (*disable_callback)(void *data);
+	struct completion disable_completion;
+};
+
+struct dcss_dev *dcss_drv_dev_to_dcss(struct device *dev);
+struct drm_device *dcss_drv_dev_to_drm(struct device *dev);
+struct dcss_dev *dcss_dev_create(struct device *dev, bool hdmi_output);
+void dcss_dev_destroy(struct dcss_dev *dcss);
+int dcss_dev_runtime_suspend(struct device *dev);
+int dcss_dev_runtime_resume(struct device *dev);
+int dcss_dev_suspend(struct device *dev);
+int dcss_dev_resume(struct device *dev);
+void dcss_enable_dtg_and_ss(struct dcss_dev *dcss);
+void dcss_disable_dtg_and_ss(struct dcss_dev *dcss);
+
+/* BLKCTL */
+int dcss_blkctl_init(struct dcss_dev *dcss, unsigned long blkctl_base);
+void dcss_blkctl_cfg(struct dcss_blkctl *blkctl);
+void dcss_blkctl_exit(struct dcss_blkctl *blkctl);
+
+/* CTXLD */
+int dcss_ctxld_init(struct dcss_dev *dcss, unsigned long ctxld_base);
+void dcss_ctxld_exit(struct dcss_ctxld *ctxld);
+void dcss_ctxld_write(struct dcss_ctxld *ctxld, u32 ctx_id,
+		      u32 val, u32 reg_idx);
+int dcss_ctxld_resume(struct dcss_ctxld *dcss_ctxld);
+int dcss_ctxld_suspend(struct dcss_ctxld *dcss_ctxld);
+void dcss_ctxld_write_irqsafe(struct dcss_ctxld *ctlxd, u32 ctx_id, u32 val,
+			      u32 reg_ofs);
+void dcss_ctxld_kick(struct dcss_ctxld *ctxld);
+bool dcss_ctxld_is_flushed(struct dcss_ctxld *ctxld);
+int dcss_ctxld_enable(struct dcss_ctxld *ctxld);
+void dcss_ctxld_register_completion(struct dcss_ctxld *ctxld,
+				    struct completion *dis_completion);
+void dcss_ctxld_assert_locked(struct dcss_ctxld *ctxld);
+
+/* DPR */
+int dcss_dpr_init(struct dcss_dev *dcss, unsigned long dpr_base);
+void dcss_dpr_exit(struct dcss_dpr *dpr);
+void dcss_dpr_write_sysctrl(struct dcss_dpr *dpr);
+void dcss_dpr_set_res(struct dcss_dpr *dpr, int ch_num, u32 xres, u32 yres);
+void dcss_dpr_addr_set(struct dcss_dpr *dpr, int ch_num, u32 luma_base_addr,
+		       u32 chroma_base_addr, u16 pitch);
+void dcss_dpr_enable(struct dcss_dpr *dpr, int ch_num, bool en);
+void dcss_dpr_format_set(struct dcss_dpr *dpr, int ch_num,
+			 const struct drm_format_info *format, u64 modifier);
+void dcss_dpr_set_rotation(struct dcss_dpr *dpr, int ch_num, u32 rotation);
+
+/* DTG */
+int dcss_dtg_init(struct dcss_dev *dcss, unsigned long dtg_base);
+void dcss_dtg_exit(struct dcss_dtg *dtg);
+bool dcss_dtg_vblank_irq_valid(struct dcss_dtg *dtg);
+void dcss_dtg_vblank_irq_enable(struct dcss_dtg *dtg, bool en);
+void dcss_dtg_vblank_irq_clear(struct dcss_dtg *dtg);
+void dcss_dtg_sync_set(struct dcss_dtg *dtg, struct videomode *vm);
+void dcss_dtg_css_set(struct dcss_dtg *dtg);
+void dcss_dtg_enable(struct dcss_dtg *dtg);
+void dcss_dtg_shutoff(struct dcss_dtg *dtg);
+bool dcss_dtg_is_enabled(struct dcss_dtg *dtg);
+void dcss_dtg_ctxld_kick_irq_enable(struct dcss_dtg *dtg, bool en);
+bool dcss_dtg_global_alpha_changed(struct dcss_dtg *dtg, int ch_num, int alpha);
+void dcss_dtg_plane_alpha_set(struct dcss_dtg *dtg, int ch_num,
+			      const struct drm_format_info *format, int alpha);
+void dcss_dtg_plane_pos_set(struct dcss_dtg *dtg, int ch_num,
+			    int px, int py, int pw, int ph);
+void dcss_dtg_ch_enable(struct dcss_dtg *dtg, int ch_num, bool en);
+
+/* SUBSAM */
+int dcss_ss_init(struct dcss_dev *dcss, unsigned long subsam_base);
+void dcss_ss_exit(struct dcss_ss *ss);
+void dcss_ss_enable(struct dcss_ss *ss);
+void dcss_ss_shutoff(struct dcss_ss *ss);
+void dcss_ss_subsam_set(struct dcss_ss *ss);
+void dcss_ss_sync_set(struct dcss_ss *ss, struct videomode *vm,
+		      bool phsync, bool pvsync);
+
+/* SCALER */
+int dcss_scaler_init(struct dcss_dev *dcss, unsigned long scaler_base);
+void dcss_scaler_exit(struct dcss_scaler *scl);
+void dcss_scaler_setup(struct dcss_scaler *scl, int ch_num,
+		       const struct drm_format_info *format,
+		       int src_xres, int src_yres, int dst_xres, int dst_yres,
+		       u32 vrefresh_hz);
+void dcss_scaler_ch_enable(struct dcss_scaler *scl, int ch_num, bool en);
+int dcss_scaler_get_min_max_ratios(struct dcss_scaler *scl, int ch_num,
+				   int *min, int *max);
+void dcss_scaler_write_sclctrl(struct dcss_scaler *scl);
+
+#endif /* __DCSS_PRV_H__ */
diff --git a/drivers/gpu/drm/imx/dcss/dcss-dpr.c b/drivers/gpu/drm/imx/dcss/dcss-dpr.c
new file mode 100644
index 000000000000..df9dab949bf2
--- /dev/null
+++ b/drivers/gpu/drm/imx/dcss/dcss-dpr.c
@@ -0,0 +1,562 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Copyright 2019 NXP.
+ */
+
+#include <linux/device.h>
+#include <linux/slab.h>
+
+#include "dcss-dev.h"
+
+#define DCSS_DPR_SYSTEM_CTRL0			0x000
+#define   RUN_EN				BIT(0)
+#define   SOFT_RESET				BIT(1)
+#define   REPEAT_EN				BIT(2)
+#define   SHADOW_LOAD_EN			BIT(3)
+#define   SW_SHADOW_LOAD_SEL			BIT(4)
+#define   BCMD2AXI_MSTR_ID_CTRL			BIT(16)
+#define DCSS_DPR_IRQ_MASK			0x020
+#define DCSS_DPR_IRQ_MASK_STATUS		0x030
+#define DCSS_DPR_IRQ_NONMASK_STATUS		0x040
+#define   IRQ_DPR_CTRL_DONE			BIT(0)
+#define   IRQ_DPR_RUN				BIT(1)
+#define   IRQ_DPR_SHADOW_LOADED			BIT(2)
+#define   IRQ_AXI_READ_ERR			BIT(3)
+#define   DPR2RTR_YRGB_FIFO_OVFL		BIT(4)
+#define   DPR2RTR_UV_FIFO_OVFL			BIT(5)
+#define   DPR2RTR_FIFO_LD_BUF_RDY_YRGB_ERR	BIT(6)
+#define   DPR2RTR_FIFO_LD_BUF_RDY_UV_ERR	BIT(7)
+#define DCSS_DPR_MODE_CTRL0			0x050
+#define   RTR_3BUF_EN				BIT(0)
+#define   RTR_4LINE_BUF_EN			BIT(1)
+#define   TILE_TYPE_POS				2
+#define   TILE_TYPE_MASK			GENMASK(4, 2)
+#define   YUV_EN				BIT(6)
+#define   COMP_2PLANE_EN			BIT(7)
+#define   PIX_SIZE_POS				8
+#define   PIX_SIZE_MASK				GENMASK(9, 8)
+#define   PIX_LUMA_UV_SWAP			BIT(10)
+#define   PIX_UV_SWAP				BIT(11)
+#define   B_COMP_SEL_POS			12
+#define   B_COMP_SEL_MASK			GENMASK(13, 12)
+#define   G_COMP_SEL_POS			14
+#define   G_COMP_SEL_MASK			GENMASK(15, 14)
+#define   R_COMP_SEL_POS			16
+#define   R_COMP_SEL_MASK			GENMASK(17, 16)
+#define   A_COMP_SEL_POS			18
+#define   A_COMP_SEL_MASK			GENMASK(19, 18)
+#define DCSS_DPR_FRAME_CTRL0			0x070
+#define   HFLIP_EN				BIT(0)
+#define   VFLIP_EN				BIT(1)
+#define   ROT_ENC_POS				2
+#define   ROT_ENC_MASK				GENMASK(3, 2)
+#define   ROT_FLIP_ORDER_EN			BIT(4)
+#define   PITCH_POS				16
+#define   PITCH_MASK				GENMASK(31, 16)
+#define DCSS_DPR_FRAME_1P_CTRL0			0x090
+#define DCSS_DPR_FRAME_1P_PIX_X_CTRL		0x0A0
+#define DCSS_DPR_FRAME_1P_PIX_Y_CTRL		0x0B0
+#define DCSS_DPR_FRAME_1P_BASE_ADDR		0x0C0
+#define DCSS_DPR_FRAME_2P_CTRL0			0x0E0
+#define DCSS_DPR_FRAME_2P_PIX_X_CTRL		0x0F0
+#define DCSS_DPR_FRAME_2P_PIX_Y_CTRL		0x100
+#define DCSS_DPR_FRAME_2P_BASE_ADDR		0x110
+#define DCSS_DPR_STATUS_CTRL0			0x130
+#define   STATUS_MUX_SEL_MASK			GENMASK(2, 0)
+#define   STATUS_SRC_SEL_POS			16
+#define   STATUS_SRC_SEL_MASK			GENMASK(18, 16)
+#define DCSS_DPR_STATUS_CTRL1			0x140
+#define DCSS_DPR_RTRAM_CTRL0			0x200
+#define   NUM_ROWS_ACTIVE			BIT(0)
+#define   THRES_HIGH_POS			1
+#define   THRES_HIGH_MASK			GENMASK(3, 1)
+#define   THRES_LOW_POS				4
+#define   THRES_LOW_MASK			GENMASK(6, 4)
+#define   ABORT_SEL				BIT(7)
+
+enum dcss_tile_type {
+	TILE_LINEAR = 0,
+	TILE_GPU_STANDARD,
+	TILE_GPU_SUPER,
+	TILE_VPU_YUV420,
+	TILE_VPU_VP9,
+};
+
+enum dcss_pix_size {
+	PIX_SIZE_8,
+	PIX_SIZE_16,
+	PIX_SIZE_32,
+};
+
+struct dcss_dpr_ch {
+	struct dcss_dpr *dpr;
+	void __iomem *base_reg;
+	u32 base_ofs;
+
+	struct drm_format_info format;
+	enum dcss_pix_size pix_size;
+	enum dcss_tile_type tile;
+	bool rtram_4line_en;
+	bool rtram_3buf_en;
+
+	u32 frame_ctrl;
+	u32 mode_ctrl;
+	u32 sys_ctrl;
+	u32 rtram_ctrl;
+
+	bool sys_ctrl_chgd;
+
+	int ch_num;
+	int irq;
+};
+
+struct dcss_dpr {
+	struct device *dev;
+	struct dcss_ctxld *ctxld;
+	u32  ctx_id;
+
+	struct dcss_dpr_ch ch[3];
+};
+
+static void dcss_dpr_write(struct dcss_dpr_ch *ch, u32 val, u32 ofs)
+{
+	struct dcss_dpr *dpr = ch->dpr;
+
+	dcss_ctxld_write(dpr->ctxld, dpr->ctx_id, val, ch->base_ofs + ofs);
+}
+
+static int dcss_dpr_ch_init_all(struct dcss_dpr *dpr, unsigned long dpr_base)
+{
+	struct dcss_dpr_ch *ch;
+	int i;
+
+	for (i = 0; i < 3; i++) {
+		ch = &dpr->ch[i];
+
+		ch->base_ofs = dpr_base + i * 0x1000;
+
+		ch->base_reg = ioremap(ch->base_ofs, SZ_4K);
+		if (!ch->base_reg) {
+			dev_err(dpr->dev, "dpr: unable to remap ch %d base\n",
+				i);
+			return -ENOMEM;
+		}
+
+		ch->dpr = dpr;
+		ch->ch_num = i;
+
+		dcss_writel(0xff, ch->base_reg + DCSS_DPR_IRQ_MASK);
+	}
+
+	return 0;
+}
+
+int dcss_dpr_init(struct dcss_dev *dcss, unsigned long dpr_base)
+{
+	struct dcss_dpr *dpr;
+
+	dpr = kzalloc(sizeof(*dpr), GFP_KERNEL);
+	if (!dpr)
+		return -ENOMEM;
+
+	dcss->dpr = dpr;
+	dpr->dev = dcss->dev;
+	dpr->ctxld = dcss->ctxld;
+	dpr->ctx_id = CTX_SB_HP;
+
+	if (dcss_dpr_ch_init_all(dpr, dpr_base)) {
+		int i;
+
+		for (i = 0; i < 3; i++) {
+			if (dpr->ch[i].base_reg)
+				iounmap(dpr->ch[i].base_reg);
+		}
+
+		kfree(dpr);
+
+		return -ENOMEM;
+	}
+
+	return 0;
+}
+
+void dcss_dpr_exit(struct dcss_dpr *dpr)
+{
+	int ch_no;
+
+	/* stop DPR on all channels */
+	for (ch_no = 0; ch_no < 3; ch_no++) {
+		struct dcss_dpr_ch *ch = &dpr->ch[ch_no];
+
+		dcss_writel(0, ch->base_reg + DCSS_DPR_SYSTEM_CTRL0);
+
+		if (ch->base_reg)
+			iounmap(ch->base_reg);
+	}
+
+	kfree(dpr);
+}
+
+static u32 dcss_dpr_x_pix_wide_adjust(struct dcss_dpr_ch *ch, u32 pix_wide,
+				      u32 pix_format)
+{
+	u8 pix_in_64byte_map[3][5] = {
+		/* LIN, GPU_STD, GPU_SUP, VPU_YUV420, VPU_VP9 */
+		{   64,       8,       8,          8,     16}, /* PIX_SIZE_8  */
+		{   32,       8,       8,          8,      8}, /* PIX_SIZE_16 */
+		{   16,       4,       4,          8,      8}, /* PIX_SIZE_32 */
+	};
+	u32 offset;
+	u32 div_64byte_mod, pix_in_64byte;
+
+	pix_in_64byte = pix_in_64byte_map[ch->pix_size][ch->tile];
+
+	div_64byte_mod = pix_wide % pix_in_64byte;
+	offset = (div_64byte_mod == 0) ? 0 : (pix_in_64byte - div_64byte_mod);
+
+	return pix_wide + offset;
+}
+
+static u32 dcss_dpr_y_pix_high_adjust(struct dcss_dpr_ch *ch, u32 pix_high,
+				      u32 pix_format)
+{
+	u8 num_rows_buf = ch->rtram_4line_en ? 4 : 8;
+	u32 offset, pix_y_mod;
+
+	pix_y_mod = pix_high % num_rows_buf;
+	offset = pix_y_mod ? (num_rows_buf - pix_y_mod) : 0;
+
+	return pix_high + offset;
+}
+
+void dcss_dpr_set_res(struct dcss_dpr *dpr, int ch_num, u32 xres, u32 yres)
+{
+	struct dcss_dpr_ch *ch = &dpr->ch[ch_num];
+	u32 pix_format = ch->format.format;
+	u32 gap = DCSS_DPR_FRAME_2P_BASE_ADDR - DCSS_DPR_FRAME_1P_BASE_ADDR;
+	int plane, max_planes = 1;
+	u32 pix_x_wide, pix_y_high;
+
+	if (pix_format == DRM_FORMAT_NV12 ||
+	    pix_format == DRM_FORMAT_NV21)
+		max_planes = 2;
+
+	for (plane = 0; plane < max_planes; plane++) {
+		yres = plane == 1 ? yres >> 1 : yres;
+
+		pix_x_wide = dcss_dpr_x_pix_wide_adjust(ch, xres, pix_format);
+		pix_y_high = dcss_dpr_y_pix_high_adjust(ch, yres, pix_format);
+
+		dcss_dpr_write(ch, pix_x_wide,
+			       DCSS_DPR_FRAME_1P_PIX_X_CTRL + plane * gap);
+		dcss_dpr_write(ch, pix_y_high,
+			       DCSS_DPR_FRAME_1P_PIX_Y_CTRL + plane * gap);
+
+		dcss_dpr_write(ch, 2, DCSS_DPR_FRAME_1P_CTRL0 + plane * gap);
+	}
+}
+
+void dcss_dpr_addr_set(struct dcss_dpr *dpr, int ch_num, u32 luma_base_addr,
+		       u32 chroma_base_addr, u16 pitch)
+{
+	struct dcss_dpr_ch *ch = &dpr->ch[ch_num];
+
+	dcss_dpr_write(ch, luma_base_addr, DCSS_DPR_FRAME_1P_BASE_ADDR);
+
+	dcss_dpr_write(ch, chroma_base_addr, DCSS_DPR_FRAME_2P_BASE_ADDR);
+
+	ch->frame_ctrl &= ~PITCH_MASK;
+	ch->frame_ctrl |= (((u32)pitch << PITCH_POS) & PITCH_MASK);
+}
+
+static void dcss_dpr_argb_comp_sel(struct dcss_dpr_ch *ch, int a_sel, int r_sel,
+				   int g_sel, int b_sel)
+{
+	u32 sel;
+
+	sel = ((a_sel << A_COMP_SEL_POS) & A_COMP_SEL_MASK) |
+	      ((r_sel << R_COMP_SEL_POS) & R_COMP_SEL_MASK) |
+	      ((g_sel << G_COMP_SEL_POS) & G_COMP_SEL_MASK) |
+	      ((b_sel << B_COMP_SEL_POS) & B_COMP_SEL_MASK);
+
+	ch->mode_ctrl &= ~(A_COMP_SEL_MASK | R_COMP_SEL_MASK |
+			   G_COMP_SEL_MASK | B_COMP_SEL_MASK);
+	ch->mode_ctrl |= sel;
+}
+
+static void dcss_dpr_pix_size_set(struct dcss_dpr_ch *ch,
+				  const struct drm_format_info *format)
+{
+	u32 val;
+
+	switch (format->format) {
+	case DRM_FORMAT_NV12:
+	case DRM_FORMAT_NV21:
+		val = PIX_SIZE_8;
+		break;
+
+	case DRM_FORMAT_UYVY:
+	case DRM_FORMAT_VYUY:
+	case DRM_FORMAT_YUYV:
+	case DRM_FORMAT_YVYU:
+		val = PIX_SIZE_16;
+		break;
+
+	default:
+		val = PIX_SIZE_32;
+		break;
+	}
+
+	ch->pix_size = val;
+
+	ch->mode_ctrl &= ~PIX_SIZE_MASK;
+	ch->mode_ctrl |= ((val << PIX_SIZE_POS) & PIX_SIZE_MASK);
+}
+
+static void dcss_dpr_uv_swap(struct dcss_dpr_ch *ch, bool swap)
+{
+	ch->mode_ctrl &= ~PIX_UV_SWAP;
+	ch->mode_ctrl |= (swap ? PIX_UV_SWAP : 0);
+}
+
+static void dcss_dpr_y_uv_swap(struct dcss_dpr_ch *ch, bool swap)
+{
+	ch->mode_ctrl &= ~PIX_LUMA_UV_SWAP;
+	ch->mode_ctrl |= (swap ? PIX_LUMA_UV_SWAP : 0);
+}
+
+static void dcss_dpr_2plane_en(struct dcss_dpr_ch *ch, bool en)
+{
+	ch->mode_ctrl &= ~COMP_2PLANE_EN;
+	ch->mode_ctrl |= (en ? COMP_2PLANE_EN : 0);
+}
+
+static void dcss_dpr_yuv_en(struct dcss_dpr_ch *ch, bool en)
+{
+	ch->mode_ctrl &= ~YUV_EN;
+	ch->mode_ctrl |= (en ? YUV_EN : 0);
+}
+
+void dcss_dpr_enable(struct dcss_dpr *dpr, int ch_num, bool en)
+{
+	struct dcss_dpr_ch *ch = &dpr->ch[ch_num];
+	u32 sys_ctrl;
+
+	sys_ctrl = (en ? REPEAT_EN | RUN_EN : 0);
+
+	if (en) {
+		dcss_dpr_write(ch, ch->mode_ctrl, DCSS_DPR_MODE_CTRL0);
+		dcss_dpr_write(ch, ch->frame_ctrl, DCSS_DPR_FRAME_CTRL0);
+		dcss_dpr_write(ch, ch->rtram_ctrl, DCSS_DPR_RTRAM_CTRL0);
+	}
+
+	if (ch->sys_ctrl != sys_ctrl)
+		ch->sys_ctrl_chgd = true;
+
+	ch->sys_ctrl = sys_ctrl;
+}
+
+struct rgb_comp_sel {
+	u32 drm_format;
+	int a_sel;
+	int r_sel;
+	int g_sel;
+	int b_sel;
+};
+
+static struct rgb_comp_sel comp_sel_map[] = {
+	{DRM_FORMAT_ARGB8888, 3, 2, 1, 0},
+	{DRM_FORMAT_XRGB8888, 3, 2, 1, 0},
+	{DRM_FORMAT_ABGR8888, 3, 0, 1, 2},
+	{DRM_FORMAT_XBGR8888, 3, 0, 1, 2},
+	{DRM_FORMAT_RGBA8888, 0, 3, 2, 1},
+	{DRM_FORMAT_RGBX8888, 0, 3, 2, 1},
+	{DRM_FORMAT_BGRA8888, 0, 1, 2, 3},
+	{DRM_FORMAT_BGRX8888, 0, 1, 2, 3},
+};
+
+static int to_comp_sel(u32 pix_fmt, int *a_sel, int *r_sel, int *g_sel,
+		       int *b_sel)
+{
+	int i;
+
+	for (i = 0; i < ARRAY_SIZE(comp_sel_map); i++) {
+		if (comp_sel_map[i].drm_format == pix_fmt) {
+			*a_sel = comp_sel_map[i].a_sel;
+			*r_sel = comp_sel_map[i].r_sel;
+			*g_sel = comp_sel_map[i].g_sel;
+			*b_sel = comp_sel_map[i].b_sel;
+
+			return 0;
+		}
+	}
+
+	return -1;
+}
+
+static void dcss_dpr_rtram_set(struct dcss_dpr_ch *ch, u32 pix_format)
+{
+	u32 val, mask;
+
+	switch (pix_format) {
+	case DRM_FORMAT_NV21:
+	case DRM_FORMAT_NV12:
+		ch->rtram_3buf_en = true;
+		ch->rtram_4line_en = false;
+		break;
+
+	default:
+		ch->rtram_3buf_en = true;
+		ch->rtram_4line_en = true;
+		break;
+	}
+
+	val = (ch->rtram_4line_en ? RTR_4LINE_BUF_EN : 0);
+	val |= (ch->rtram_3buf_en ? RTR_3BUF_EN : 0);
+	mask = RTR_4LINE_BUF_EN | RTR_3BUF_EN;
+
+	ch->mode_ctrl &= ~mask;
+	ch->mode_ctrl |= (val & mask);
+
+	val = (ch->rtram_4line_en ? 0 : NUM_ROWS_ACTIVE);
+	val |= (3 << THRES_LOW_POS) & THRES_LOW_MASK;
+	val |= (4 << THRES_HIGH_POS) & THRES_HIGH_MASK;
+	mask = THRES_LOW_MASK | THRES_HIGH_MASK | NUM_ROWS_ACTIVE;
+
+	ch->rtram_ctrl &= ~mask;
+	ch->rtram_ctrl |= (val & mask);
+}
+
+static void dcss_dpr_setup_components(struct dcss_dpr_ch *ch,
+				      const struct drm_format_info *format)
+{
+	int a_sel, r_sel, g_sel, b_sel;
+	bool uv_swap, y_uv_swap;
+
+	switch (format->format) {
+	case DRM_FORMAT_YVYU:
+		uv_swap = true;
+		y_uv_swap = true;
+		break;
+
+	case DRM_FORMAT_VYUY:
+	case DRM_FORMAT_NV21:
+		uv_swap = true;
+		y_uv_swap = false;
+		break;
+
+	case DRM_FORMAT_YUYV:
+		uv_swap = false;
+		y_uv_swap = true;
+		break;
+
+	default:
+		uv_swap = false;
+		y_uv_swap = false;
+		break;
+	}
+
+	dcss_dpr_uv_swap(ch, uv_swap);
+
+	dcss_dpr_y_uv_swap(ch, y_uv_swap);
+
+	if (!format->is_yuv) {
+		if (!to_comp_sel(format->format, &a_sel, &r_sel,
+				 &g_sel, &b_sel)) {
+			dcss_dpr_argb_comp_sel(ch, a_sel, r_sel, g_sel, b_sel);
+		} else {
+			dcss_dpr_argb_comp_sel(ch, 3, 2, 1, 0);
+		}
+	} else {
+		dcss_dpr_argb_comp_sel(ch, 0, 0, 0, 0);
+	}
+}
+
+static void dcss_dpr_tile_set(struct dcss_dpr_ch *ch, uint64_t modifier)
+{
+	switch (ch->ch_num) {
+	case 0:
+		switch (modifier) {
+		case DRM_FORMAT_MOD_LINEAR:
+			ch->tile = TILE_LINEAR;
+			break;
+		case DRM_FORMAT_MOD_VIVANTE_TILED:
+			ch->tile = TILE_GPU_STANDARD;
+			break;
+		case DRM_FORMAT_MOD_VIVANTE_SUPER_TILED:
+			ch->tile = TILE_GPU_SUPER;
+			break;
+		default:
+			WARN_ON(1);
+			break;
+		}
+		break;
+	case 1:
+	case 2:
+		ch->tile = TILE_LINEAR;
+		break;
+	default:
+		WARN_ON(1);
+		return;
+	}
+
+	ch->mode_ctrl &= ~TILE_TYPE_MASK;
+	ch->mode_ctrl |= ((ch->tile << TILE_TYPE_POS) & TILE_TYPE_MASK);
+}
+
+void dcss_dpr_format_set(struct dcss_dpr *dpr, int ch_num,
+			 const struct drm_format_info *format, u64 modifier)
+{
+	struct dcss_dpr_ch *ch = &dpr->ch[ch_num];
+
+	ch->format = *format;
+
+	dcss_dpr_yuv_en(ch, format->is_yuv);
+
+	dcss_dpr_pix_size_set(ch, format);
+
+	dcss_dpr_setup_components(ch, format);
+
+	dcss_dpr_2plane_en(ch, format->num_planes == 2);
+
+	dcss_dpr_rtram_set(ch, format->format);
+
+	dcss_dpr_tile_set(ch, modifier);
+}
+
+/* This function will be called from interrupt context. */
+void dcss_dpr_write_sysctrl(struct dcss_dpr *dpr)
+{
+	int chnum;
+
+	dcss_ctxld_assert_locked(dpr->ctxld);
+
+	for (chnum = 0; chnum < 3; chnum++) {
+		struct dcss_dpr_ch *ch = &dpr->ch[chnum];
+
+		if (ch->sys_ctrl_chgd) {
+			dcss_ctxld_write_irqsafe(dpr->ctxld, dpr->ctx_id,
+						 ch->sys_ctrl,
+						 ch->base_ofs +
+						 DCSS_DPR_SYSTEM_CTRL0);
+			ch->sys_ctrl_chgd = false;
+		}
+	}
+}
+
+void dcss_dpr_set_rotation(struct dcss_dpr *dpr, int ch_num, u32 rotation)
+{
+	struct dcss_dpr_ch *ch = &dpr->ch[ch_num];
+
+	ch->frame_ctrl &= ~(HFLIP_EN | VFLIP_EN | ROT_ENC_MASK);
+
+	ch->frame_ctrl |= rotation & DRM_MODE_REFLECT_X ? HFLIP_EN : 0;
+	ch->frame_ctrl |= rotation & DRM_MODE_REFLECT_Y ? VFLIP_EN : 0;
+
+	if (rotation & DRM_MODE_ROTATE_90)
+		ch->frame_ctrl |= 1 << ROT_ENC_POS;
+	else if (rotation & DRM_MODE_ROTATE_180)
+		ch->frame_ctrl |= 2 << ROT_ENC_POS;
+	else if (rotation & DRM_MODE_ROTATE_270)
+		ch->frame_ctrl |= 3 << ROT_ENC_POS;
+}
diff --git a/drivers/gpu/drm/imx/dcss/dcss-drv.c b/drivers/gpu/drm/imx/dcss/dcss-drv.c
new file mode 100644
index 000000000000..8dc2f85c514b
--- /dev/null
+++ b/drivers/gpu/drm/imx/dcss/dcss-drv.c
@@ -0,0 +1,138 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Copyright 2019 NXP.
+ */
+
+#include <linux/module.h>
+#include <linux/kernel.h>
+#include <linux/platform_device.h>
+#include <drm/drm_of.h>
+
+#include "dcss-dev.h"
+#include "dcss-kms.h"
+
+struct dcss_drv {
+	struct dcss_dev *dcss;
+	struct dcss_kms_dev *kms;
+};
+
+struct dcss_dev *dcss_drv_dev_to_dcss(struct device *dev)
+{
+	struct dcss_drv *mdrv = dev_get_drvdata(dev);
+
+	return mdrv ? mdrv->dcss : NULL;
+}
+
+struct drm_device *dcss_drv_dev_to_drm(struct device *dev)
+{
+	struct dcss_drv *mdrv = dev_get_drvdata(dev);
+
+	return mdrv ? &mdrv->kms->base : NULL;
+}
+
+static int dcss_drv_platform_probe(struct platform_device *pdev)
+{
+	struct device *dev = &pdev->dev;
+	struct device_node *remote;
+	struct dcss_drv *mdrv;
+	int err = 0;
+	bool hdmi_output = true;
+
+	if (!dev->of_node)
+		return -ENODEV;
+
+	remote = of_graph_get_remote_node(dev->of_node, 0, 0);
+	if (!remote)
+		return -ENODEV;
+
+	hdmi_output = !of_device_is_compatible(remote, "fsl,imx8mq-nwl-dsi");
+
+	of_node_put(remote);
+
+	mdrv = kzalloc(sizeof(*mdrv), GFP_KERNEL);
+	if (!mdrv)
+		return -ENOMEM;
+
+	mdrv->dcss = dcss_dev_create(dev, hdmi_output);
+	if (IS_ERR(mdrv->dcss)) {
+		err = PTR_ERR(mdrv->dcss);
+		goto err;
+	}
+
+	dev_set_drvdata(dev, mdrv);
+
+	mdrv->kms = dcss_kms_attach(mdrv->dcss);
+	if (IS_ERR(mdrv->kms)) {
+		err = PTR_ERR(mdrv->kms);
+		goto dcss_shutoff;
+	}
+
+	return 0;
+
+dcss_shutoff:
+	dcss_dev_destroy(mdrv->dcss);
+
+	dev_set_drvdata(dev, NULL);
+
+err:
+	kfree(mdrv);
+	return err;
+}
+
+static int dcss_drv_platform_remove(struct platform_device *pdev)
+{
+	struct dcss_drv *mdrv = dev_get_drvdata(&pdev->dev);
+
+	if (!mdrv)
+		return 0;
+
+	dcss_kms_detach(mdrv->kms);
+	dcss_dev_destroy(mdrv->dcss);
+
+	dev_set_drvdata(&pdev->dev, NULL);
+
+	kfree(mdrv);
+
+	return 0;
+}
+
+static struct dcss_type_data dcss_types[] = {
+	[DCSS_IMX8MQ] = {
+		.name = "DCSS_IMX8MQ",
+		.blkctl_ofs = 0x2F000,
+		.ctxld_ofs = 0x23000,
+		.dtg_ofs = 0x20000,
+		.scaler_ofs = 0x1C000,
+		.ss_ofs = 0x1B000,
+		.dpr_ofs = 0x18000,
+	},
+};
+
+static const struct of_device_id dcss_of_match[] = {
+	{ .compatible = "nxp,imx8mq-dcss", .data = &dcss_types[DCSS_IMX8MQ], },
+	{},
+};
+
+MODULE_DEVICE_TABLE(of, dcss_of_match);
+
+static const struct dev_pm_ops dcss_dev_pm = {
+	SET_SYSTEM_SLEEP_PM_OPS(dcss_dev_suspend, dcss_dev_resume)
+	SET_RUNTIME_PM_OPS(dcss_dev_runtime_suspend,
+			   dcss_dev_runtime_resume, NULL)
+};
+
+static struct platform_driver dcss_platform_driver = {
+	.probe	= dcss_drv_platform_probe,
+	.remove	= dcss_drv_platform_remove,
+	.driver	= {
+		.name = "imx-dcss",
+		.of_match_table	= dcss_of_match,
+		.pm = &dcss_dev_pm,
+	},
+};
+
+module_platform_driver(dcss_platform_driver);
+
+MODULE_AUTHOR("Laurentiu Palcu <laurentiu.palcu@nxp.com>");
+MODULE_DESCRIPTION("DCSS driver for i.MX8MQ");
+MODULE_LICENSE("GPL v2");
diff --git a/drivers/gpu/drm/imx/dcss/dcss-dtg.c b/drivers/gpu/drm/imx/dcss/dcss-dtg.c
new file mode 100644
index 000000000000..30de00540f63
--- /dev/null
+++ b/drivers/gpu/drm/imx/dcss/dcss-dtg.c
@@ -0,0 +1,409 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Copyright 2019 NXP.
+ */
+
+#include <linux/clk.h>
+#include <linux/delay.h>
+#include <linux/interrupt.h>
+#include <linux/of.h>
+#include <linux/platform_device.h>
+#include <linux/slab.h>
+
+#include "dcss-dev.h"
+
+#define DCSS_DTG_TC_CONTROL_STATUS			0x00
+#define   CH3_EN					BIT(0)
+#define   CH2_EN					BIT(1)
+#define   CH1_EN					BIT(2)
+#define   OVL_DATA_MODE					BIT(3)
+#define   BLENDER_VIDEO_ALPHA_SEL			BIT(7)
+#define   DTG_START					BIT(8)
+#define   DBY_MODE_EN					BIT(9)
+#define   CH1_ALPHA_SEL					BIT(10)
+#define   CSS_PIX_COMP_SWAP_POS				12
+#define   CSS_PIX_COMP_SWAP_MASK			GENMASK(14, 12)
+#define   DEFAULT_FG_ALPHA_POS				24
+#define   DEFAULT_FG_ALPHA_MASK				GENMASK(31, 24)
+#define DCSS_DTG_TC_DTG					0x04
+#define DCSS_DTG_TC_DISP_TOP				0x08
+#define DCSS_DTG_TC_DISP_BOT				0x0C
+#define DCSS_DTG_TC_CH1_TOP				0x10
+#define DCSS_DTG_TC_CH1_BOT				0x14
+#define DCSS_DTG_TC_CH2_TOP				0x18
+#define DCSS_DTG_TC_CH2_BOT				0x1C
+#define DCSS_DTG_TC_CH3_TOP				0x20
+#define DCSS_DTG_TC_CH3_BOT				0x24
+#define   TC_X_POS					0
+#define   TC_X_MASK					GENMASK(12, 0)
+#define   TC_Y_POS					16
+#define   TC_Y_MASK					GENMASK(28, 16)
+#define DCSS_DTG_TC_CTXLD				0x28
+#define   TC_CTXLD_DB_Y_POS				0
+#define   TC_CTXLD_DB_Y_MASK				GENMASK(12, 0)
+#define   TC_CTXLD_SB_Y_POS				16
+#define   TC_CTXLD_SB_Y_MASK				GENMASK(28, 16)
+#define DCSS_DTG_TC_CH1_BKRND				0x2C
+#define DCSS_DTG_TC_CH2_BKRND				0x30
+#define   BKRND_R_Y_COMP_POS				20
+#define   BKRND_R_Y_COMP_MASK				GENMASK(29, 20)
+#define   BKRND_G_U_COMP_POS				10
+#define   BKRND_G_U_COMP_MASK				GENMASK(19, 10)
+#define   BKRND_B_V_COMP_POS				0
+#define   BKRND_B_V_COMP_MASK				GENMASK(9, 0)
+#define DCSS_DTG_BLENDER_DBY_RANGEINV			0x38
+#define DCSS_DTG_BLENDER_DBY_RANGEMIN			0x3C
+#define DCSS_DTG_BLENDER_DBY_BDP			0x40
+#define DCSS_DTG_BLENDER_BKRND_I			0x44
+#define DCSS_DTG_BLENDER_BKRND_P			0x48
+#define DCSS_DTG_BLENDER_BKRND_T			0x4C
+#define DCSS_DTG_LINE0_INT				0x50
+#define DCSS_DTG_LINE1_INT				0x54
+#define DCSS_DTG_BG_ALPHA_DEFAULT			0x58
+#define DCSS_DTG_INT_STATUS				0x5C
+#define DCSS_DTG_INT_CONTROL				0x60
+#define DCSS_DTG_TC_CH3_BKRND				0x64
+#define DCSS_DTG_INT_MASK				0x68
+#define   LINE0_IRQ					BIT(0)
+#define   LINE1_IRQ					BIT(1)
+#define   LINE2_IRQ					BIT(2)
+#define   LINE3_IRQ					BIT(3)
+#define DCSS_DTG_LINE2_INT				0x6C
+#define DCSS_DTG_LINE3_INT				0x70
+#define DCSS_DTG_DBY_OL					0x74
+#define DCSS_DTG_DBY_BL					0x78
+#define DCSS_DTG_DBY_EL					0x7C
+
+struct dcss_dtg {
+	struct device *dev;
+	struct dcss_ctxld *ctxld;
+	void __iomem *base_reg;
+	u32 base_ofs;
+
+	u32 ctx_id;
+
+	bool in_use;
+
+	u32 dis_ulc_x;
+	u32 dis_ulc_y;
+
+	u32 control_status;
+	u32 alpha;
+	u32 alpha_cfg;
+
+	int ctxld_kick_irq;
+	bool ctxld_kick_irq_en;
+};
+
+static void dcss_dtg_write(struct dcss_dtg *dtg, u32 val, u32 ofs)
+{
+	if (!dtg->in_use)
+		dcss_writel(val, dtg->base_reg + ofs);
+
+	dcss_ctxld_write(dtg->ctxld, dtg->ctx_id,
+			 val, dtg->base_ofs + ofs);
+}
+
+static irqreturn_t dcss_dtg_irq_handler(int irq, void *data)
+{
+	struct dcss_dtg *dtg = data;
+	u32 status;
+
+	status = dcss_readl(dtg->base_reg + DCSS_DTG_INT_STATUS);
+
+	if (!(status & LINE0_IRQ))
+		return IRQ_NONE;
+
+	dcss_ctxld_kick(dtg->ctxld);
+
+	dcss_writel(status & LINE0_IRQ, dtg->base_reg + DCSS_DTG_INT_CONTROL);
+
+	return IRQ_HANDLED;
+}
+
+static int dcss_dtg_irq_config(struct dcss_dtg *dtg,
+			       struct platform_device *pdev)
+{
+	int ret;
+
+	dtg->ctxld_kick_irq = platform_get_irq_byname(pdev, "ctxld_kick");
+	if (dtg->ctxld_kick_irq < 0)
+		return dtg->ctxld_kick_irq;
+
+	dcss_update(0, LINE0_IRQ | LINE1_IRQ,
+		    dtg->base_reg + DCSS_DTG_INT_MASK);
+
+	ret = request_irq(dtg->ctxld_kick_irq, dcss_dtg_irq_handler,
+			  0, "dcss_ctxld_kick", dtg);
+	if (ret) {
+		dev_err(dtg->dev, "dtg: irq request failed.\n");
+		return ret;
+	}
+
+	disable_irq(dtg->ctxld_kick_irq);
+
+	dtg->ctxld_kick_irq_en = false;
+
+	return 0;
+}
+
+int dcss_dtg_init(struct dcss_dev *dcss, unsigned long dtg_base)
+{
+	int ret = 0;
+	struct dcss_dtg *dtg;
+
+	dtg = kzalloc(sizeof(*dtg), GFP_KERNEL);
+	if (!dtg)
+		return -ENOMEM;
+
+	dcss->dtg = dtg;
+	dtg->dev = dcss->dev;
+	dtg->ctxld = dcss->ctxld;
+
+	dtg->base_reg = ioremap(dtg_base, SZ_4K);
+	if (!dtg->base_reg) {
+		dev_err(dcss->dev, "dtg: unable to remap dtg base\n");
+		ret = -ENOMEM;
+		goto err_ioremap;
+	}
+
+	dtg->base_ofs = dtg_base;
+	dtg->ctx_id = CTX_DB;
+
+	dtg->alpha = 255;
+
+	dtg->control_status |= OVL_DATA_MODE | BLENDER_VIDEO_ALPHA_SEL |
+		((dtg->alpha << DEFAULT_FG_ALPHA_POS) & DEFAULT_FG_ALPHA_MASK);
+
+	ret = dcss_dtg_irq_config(dtg, to_platform_device(dcss->dev));
+	if (ret)
+		goto err_irq;
+
+	return 0;
+
+err_irq:
+	iounmap(dtg->base_reg);
+
+err_ioremap:
+	kfree(dtg);
+
+	return ret;
+}
+
+void dcss_dtg_exit(struct dcss_dtg *dtg)
+{
+	free_irq(dtg->ctxld_kick_irq, dtg);
+
+	if (dtg->base_reg)
+		iounmap(dtg->base_reg);
+
+	kfree(dtg);
+}
+
+void dcss_dtg_sync_set(struct dcss_dtg *dtg, struct videomode *vm)
+{
+	struct dcss_dev *dcss = dcss_drv_dev_to_dcss(dtg->dev);
+	u16 dtg_lrc_x, dtg_lrc_y;
+	u16 dis_ulc_x, dis_ulc_y;
+	u16 dis_lrc_x, dis_lrc_y;
+	u32 sb_ctxld_trig, db_ctxld_trig;
+	u32 pixclock = vm->pixelclock;
+	u32 actual_clk;
+
+	dtg_lrc_x = vm->hfront_porch + vm->hback_porch + vm->hsync_len +
+		    vm->hactive - 1;
+	dtg_lrc_y = vm->vfront_porch + vm->vback_porch + vm->vsync_len +
+		    vm->vactive - 1;
+	dis_ulc_x = vm->hsync_len + vm->hback_porch - 1;
+	dis_ulc_y = vm->vsync_len + vm->vfront_porch + vm->vback_porch - 1;
+	dis_lrc_x = vm->hsync_len + vm->hback_porch + vm->hactive - 1;
+	dis_lrc_y = vm->vsync_len + vm->vfront_porch + vm->vback_porch +
+		    vm->vactive - 1;
+
+	clk_disable_unprepare(dcss->pix_clk);
+	clk_set_rate(dcss->pix_clk, vm->pixelclock);
+	clk_prepare_enable(dcss->pix_clk);
+
+	actual_clk = clk_get_rate(dcss->pix_clk);
+	if (pixclock != actual_clk) {
+		dev_info(dtg->dev,
+			 "Pixel clock set to %u kHz instead of %u kHz.\n",
+			 (actual_clk / 1000), (pixclock / 1000));
+	}
+
+	dcss_dtg_write(dtg, ((dtg_lrc_y << TC_Y_POS) | dtg_lrc_x),
+		       DCSS_DTG_TC_DTG);
+	dcss_dtg_write(dtg, ((dis_ulc_y << TC_Y_POS) | dis_ulc_x),
+		       DCSS_DTG_TC_DISP_TOP);
+	dcss_dtg_write(dtg, ((dis_lrc_y << TC_Y_POS) | dis_lrc_x),
+		       DCSS_DTG_TC_DISP_BOT);
+
+	dtg->dis_ulc_x = dis_ulc_x;
+	dtg->dis_ulc_y = dis_ulc_y;
+
+	sb_ctxld_trig = ((0 * dis_lrc_y / 100) << TC_CTXLD_SB_Y_POS) &
+							TC_CTXLD_SB_Y_MASK;
+	db_ctxld_trig = ((99 * dis_lrc_y / 100) << TC_CTXLD_DB_Y_POS) &
+							TC_CTXLD_DB_Y_MASK;
+
+	dcss_dtg_write(dtg, sb_ctxld_trig | db_ctxld_trig, DCSS_DTG_TC_CTXLD);
+
+	/* vblank trigger */
+	dcss_dtg_write(dtg, 0, DCSS_DTG_LINE1_INT);
+
+	/* CTXLD trigger */
+	dcss_dtg_write(dtg, ((90 * dis_lrc_y) / 100) << 16, DCSS_DTG_LINE0_INT);
+}
+
+void dcss_dtg_plane_pos_set(struct dcss_dtg *dtg, int ch_num,
+			    int px, int py, int pw, int ph)
+{
+	u16 p_ulc_x, p_ulc_y;
+	u16 p_lrc_x, p_lrc_y;
+
+	p_ulc_x = dtg->dis_ulc_x + px;
+	p_ulc_y = dtg->dis_ulc_y + py;
+	p_lrc_x = p_ulc_x + pw;
+	p_lrc_y = p_ulc_y + ph;
+
+	if (!px && !py && !pw && !ph) {
+		dcss_dtg_write(dtg, 0, DCSS_DTG_TC_CH1_TOP + 0x8 * ch_num);
+		dcss_dtg_write(dtg, 0, DCSS_DTG_TC_CH1_BOT + 0x8 * ch_num);
+	} else {
+		dcss_dtg_write(dtg, ((p_ulc_y << TC_Y_POS) | p_ulc_x),
+			       DCSS_DTG_TC_CH1_TOP + 0x8 * ch_num);
+		dcss_dtg_write(dtg, ((p_lrc_y << TC_Y_POS) | p_lrc_x),
+			       DCSS_DTG_TC_CH1_BOT + 0x8 * ch_num);
+	}
+}
+
+bool dcss_dtg_global_alpha_changed(struct dcss_dtg *dtg, int ch_num, int alpha)
+{
+	if (ch_num)
+		return false;
+
+	return alpha != dtg->alpha;
+}
+
+void dcss_dtg_plane_alpha_set(struct dcss_dtg *dtg, int ch_num,
+			      const struct drm_format_info *format, int alpha)
+{
+	/* we care about alpha only when channel 0 is concerned */
+	if (ch_num)
+		return;
+
+	/*
+	 * Use global alpha if pixel format does not have alpha channel or the
+	 * user explicitly chose to use global alpha (i.e. alpha is not OPAQUE).
+	 */
+	if (!format->has_alpha || alpha != 255)
+		dtg->alpha_cfg = (alpha << DEFAULT_FG_ALPHA_POS) & DEFAULT_FG_ALPHA_MASK;
+	else /* use per-pixel alpha otherwise */
+		dtg->alpha_cfg = CH1_ALPHA_SEL;
+
+	dtg->alpha = alpha;
+}
+
+void dcss_dtg_css_set(struct dcss_dtg *dtg)
+{
+	dtg->control_status |=
+			(0x5 << CSS_PIX_COMP_SWAP_POS) & CSS_PIX_COMP_SWAP_MASK;
+}
+
+void dcss_dtg_enable(struct dcss_dtg *dtg)
+{
+	dtg->control_status |= DTG_START;
+
+	dtg->control_status &= ~(CH1_ALPHA_SEL | DEFAULT_FG_ALPHA_MASK);
+	dtg->control_status |= dtg->alpha_cfg;
+
+	dcss_dtg_write(dtg, dtg->control_status, DCSS_DTG_TC_CONTROL_STATUS);
+
+	dtg->in_use = true;
+}
+
+void dcss_dtg_shutoff(struct dcss_dtg *dtg)
+{
+	dtg->control_status &= ~DTG_START;
+
+	dcss_writel(dtg->control_status,
+		    dtg->base_reg + DCSS_DTG_TC_CONTROL_STATUS);
+
+	dtg->in_use = false;
+}
+
+bool dcss_dtg_is_enabled(struct dcss_dtg *dtg)
+{
+	return dtg->in_use;
+}
+
+void dcss_dtg_ch_enable(struct dcss_dtg *dtg, int ch_num, bool en)
+{
+	u32 ch_en_map[] = {CH1_EN, CH2_EN, CH3_EN};
+	u32 control_status;
+
+	control_status = dtg->control_status & ~ch_en_map[ch_num];
+	control_status |= en ? ch_en_map[ch_num] : 0;
+
+	control_status &= ~(CH1_ALPHA_SEL | DEFAULT_FG_ALPHA_MASK);
+	control_status |= dtg->alpha_cfg;
+
+	if (dtg->control_status != control_status)
+		dcss_dtg_write(dtg, control_status, DCSS_DTG_TC_CONTROL_STATUS);
+
+	dtg->control_status = control_status;
+}
+
+void dcss_dtg_vblank_irq_enable(struct dcss_dtg *dtg, bool en)
+{
+	u32 status;
+	u32 mask = en ? LINE1_IRQ : 0;
+
+	if (en) {
+		status = dcss_readl(dtg->base_reg + DCSS_DTG_INT_STATUS);
+		dcss_writel(status & LINE1_IRQ,
+			    dtg->base_reg + DCSS_DTG_INT_CONTROL);
+	}
+
+	dcss_update(mask, LINE1_IRQ, dtg->base_reg + DCSS_DTG_INT_MASK);
+}
+
+void dcss_dtg_ctxld_kick_irq_enable(struct dcss_dtg *dtg, bool en)
+{
+	u32 status;
+	u32 mask = en ? LINE0_IRQ : 0;
+
+	if (en) {
+		status = dcss_readl(dtg->base_reg + DCSS_DTG_INT_STATUS);
+
+		if (!dtg->ctxld_kick_irq_en) {
+			dcss_writel(status & LINE0_IRQ,
+				    dtg->base_reg + DCSS_DTG_INT_CONTROL);
+			enable_irq(dtg->ctxld_kick_irq);
+			dtg->ctxld_kick_irq_en = true;
+			dcss_update(mask, LINE0_IRQ,
+				    dtg->base_reg + DCSS_DTG_INT_MASK);
+		}
+
+		return;
+	}
+
+	if (!dtg->ctxld_kick_irq_en)
+		return;
+
+	disable_irq_nosync(dtg->ctxld_kick_irq);
+	dtg->ctxld_kick_irq_en = false;
+
+	dcss_update(mask, LINE0_IRQ, dtg->base_reg + DCSS_DTG_INT_MASK);
+}
+
+void dcss_dtg_vblank_irq_clear(struct dcss_dtg *dtg)
+{
+	dcss_update(LINE1_IRQ, LINE1_IRQ, dtg->base_reg + DCSS_DTG_INT_CONTROL);
+}
+
+bool dcss_dtg_vblank_irq_valid(struct dcss_dtg *dtg)
+{
+	return !!(dcss_readl(dtg->base_reg + DCSS_DTG_INT_STATUS) & LINE1_IRQ);
+}
+
diff --git a/drivers/gpu/drm/imx/dcss/dcss-kms.c b/drivers/gpu/drm/imx/dcss/dcss-kms.c
new file mode 100644
index 000000000000..135a62366ab8
--- /dev/null
+++ b/drivers/gpu/drm/imx/dcss/dcss-kms.c
@@ -0,0 +1,198 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Copyright 2019 NXP.
+ */
+
+#include <drm/drm_atomic.h>
+#include <drm/drm_atomic_helper.h>
+#include <drm/drm_bridge_connector.h>
+#include <drm/drm_drv.h>
+#include <drm/drm_fb_helper.h>
+#include <drm/drm_gem_cma_helper.h>
+#include <drm/drm_gem_framebuffer_helper.h>
+#include <drm/drm_of.h>
+#include <drm/drm_probe_helper.h>
+#include <drm/drm_vblank.h>
+
+#include "dcss-dev.h"
+#include "dcss-kms.h"
+
+DEFINE_DRM_GEM_CMA_FOPS(dcss_cma_fops);
+
+static const struct drm_mode_config_funcs dcss_drm_mode_config_funcs = {
+	.fb_create = drm_gem_fb_create,
+	.output_poll_changed = drm_fb_helper_output_poll_changed,
+	.atomic_check = drm_atomic_helper_check,
+	.atomic_commit = drm_atomic_helper_commit,
+};
+
+static struct drm_driver dcss_kms_driver = {
+	.driver_features	= DRIVER_MODESET | DRIVER_GEM | DRIVER_ATOMIC,
+	.gem_free_object_unlocked = drm_gem_cma_free_object,
+	.gem_vm_ops		= &drm_gem_cma_vm_ops,
+	.dumb_create		= drm_gem_cma_dumb_create,
+
+	.prime_handle_to_fd	= drm_gem_prime_handle_to_fd,
+	.prime_fd_to_handle	= drm_gem_prime_fd_to_handle,
+	.gem_prime_import	= drm_gem_prime_import,
+	.gem_prime_export	= drm_gem_prime_export,
+	.gem_prime_get_sg_table	= drm_gem_cma_prime_get_sg_table,
+	.gem_prime_import_sg_table = drm_gem_cma_prime_import_sg_table,
+	.gem_prime_vmap		= drm_gem_cma_prime_vmap,
+	.gem_prime_vunmap	= drm_gem_cma_prime_vunmap,
+	.gem_prime_mmap		= drm_gem_cma_prime_mmap,
+	.fops			= &dcss_cma_fops,
+	.name			= "imx-dcss",
+	.desc			= "i.MX8MQ Display Subsystem",
+	.date			= "20190917",
+	.major			= 1,
+	.minor			= 0,
+	.patchlevel		= 0,
+};
+
+static const struct drm_mode_config_helper_funcs dcss_mode_config_helpers = {
+	.atomic_commit_tail = drm_atomic_helper_commit_tail_rpm,
+};
+
+static void dcss_kms_mode_config_init(struct dcss_kms_dev *kms)
+{
+	struct drm_mode_config *config = &kms->base.mode_config;
+
+	drm_mode_config_init(&kms->base);
+
+	config->min_width = 1;
+	config->min_height = 1;
+	config->max_width = 4096;
+	config->max_height = 4096;
+	config->allow_fb_modifiers = true;
+	config->normalize_zpos = true;
+
+	config->funcs = &dcss_drm_mode_config_funcs;
+	config->helper_private = &dcss_mode_config_helpers;
+}
+
+static const struct drm_encoder_funcs dcss_kms_simple_encoder_funcs = {
+	.destroy = drm_encoder_cleanup,
+};
+
+static int dcss_kms_bridge_connector_init(struct dcss_kms_dev *kms)
+{
+	struct drm_device *ddev = &kms->base;
+	struct drm_encoder *encoder = &kms->encoder;
+	struct drm_crtc *crtc = (struct drm_crtc *)&kms->crtc;
+	struct drm_panel *panel;
+	struct drm_bridge *bridge;
+	int ret;
+
+	ret = drm_of_find_panel_or_bridge(ddev->dev->of_node, 0, 0,
+					  &panel, &bridge);
+	if (ret)
+		return ret;
+
+	if (!bridge) {
+		dev_err(ddev->dev, "No bridge found %d.\n", ret);
+		return -ENODEV;
+	}
+
+	encoder->possible_crtcs = drm_crtc_mask(crtc);
+
+	ret = drm_encoder_init(&kms->base, encoder,
+			       &dcss_kms_simple_encoder_funcs,
+			       DRM_MODE_ENCODER_NONE, NULL);
+	if (ret) {
+		dev_err(ddev->dev, "Failed initializing encoder %d.\n", ret);
+		return ret;
+	}
+
+	ret = drm_bridge_attach(encoder, bridge, NULL,
+				DRM_BRIDGE_ATTACH_NO_CONNECTOR);
+	if (ret < 0) {
+		dev_err(ddev->dev, "Unable to attach bridge %pOF\n",
+			bridge->of_node);
+		return ret;
+	}
+
+	kms->connector = drm_bridge_connector_init(ddev, encoder);
+	if (IS_ERR(kms->connector)) {
+		dev_err(ddev->dev, "Unable to create bridge connector.\n");
+		return PTR_ERR(kms->connector);
+	}
+
+	drm_connector_attach_encoder(kms->connector, encoder);
+
+	return 0;
+}
+
+struct dcss_kms_dev *dcss_kms_attach(struct dcss_dev *dcss)
+{
+	struct dcss_kms_dev *kms;
+	struct drm_device *drm;
+	struct dcss_crtc *crtc;
+	int ret;
+
+	kms = devm_drm_dev_alloc(dcss->dev, &dcss_kms_driver,
+				 struct dcss_kms_dev, base);
+	if (IS_ERR(kms))
+		return kms;
+
+	drm = &kms->base;
+	crtc = &kms->crtc;
+
+	drm->dev_private = dcss;
+
+	dcss_kms_mode_config_init(kms);
+
+	ret = drm_vblank_init(drm, 1);
+	if (ret)
+		goto cleanup_mode_config;
+
+	drm->irq_enabled = true;
+
+	ret = dcss_kms_bridge_connector_init(kms);
+	if (ret)
+		goto cleanup_mode_config;
+
+	ret = dcss_crtc_init(crtc, drm);
+	if (ret)
+		goto cleanup_mode_config;
+
+	drm_mode_config_reset(drm);
+
+	drm_kms_helper_poll_init(drm);
+
+	drm_bridge_connector_enable_hpd(kms->connector);
+
+	ret = drm_dev_register(drm, 0);
+	if (ret)
+		goto cleanup_crtc;
+
+	drm_fbdev_generic_setup(drm, 32);
+
+	return kms;
+
+cleanup_crtc:
+	drm_bridge_connector_disable_hpd(kms->connector);
+	drm_kms_helper_poll_fini(drm);
+	dcss_crtc_deinit(crtc, drm);
+
+cleanup_mode_config:
+	drm_mode_config_cleanup(drm);
+	drm->dev_private = NULL;
+
+	return ERR_PTR(ret);
+}
+
+void dcss_kms_detach(struct dcss_kms_dev *kms)
+{
+	struct drm_device *drm = &kms->base;
+
+	drm_dev_unregister(drm);
+	drm_bridge_connector_disable_hpd(kms->connector);
+	drm_kms_helper_poll_fini(drm);
+	drm_atomic_helper_shutdown(drm);
+	drm_crtc_vblank_off(&kms->crtc.base);
+	drm->irq_enabled = false;
+	drm_mode_config_cleanup(drm);
+	dcss_crtc_deinit(&kms->crtc, drm);
+	drm->dev_private = NULL;
+}
diff --git a/drivers/gpu/drm/imx/dcss/dcss-kms.h b/drivers/gpu/drm/imx/dcss/dcss-kms.h
new file mode 100644
index 000000000000..dfe5dd99eea3
--- /dev/null
+++ b/drivers/gpu/drm/imx/dcss/dcss-kms.h
@@ -0,0 +1,44 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+/*
+ * Copyright 2019 NXP.
+ */
+
+#ifndef _DCSS_KMS_H_
+#define _DCSS_KMS_H_
+
+#include <drm/drm_encoder.h>
+
+struct dcss_plane {
+	struct drm_plane base;
+
+	int ch_num;
+};
+
+struct dcss_crtc {
+	struct drm_crtc		base;
+	struct drm_crtc_state	*state;
+
+	struct dcss_plane	*plane[3];
+
+	int			irq;
+
+	bool disable_ctxld_kick_irq;
+};
+
+struct dcss_kms_dev {
+	struct drm_device base;
+	struct dcss_crtc crtc;
+	struct drm_encoder encoder;
+	struct drm_connector *connector;
+};
+
+struct dcss_kms_dev *dcss_kms_attach(struct dcss_dev *dcss);
+void dcss_kms_detach(struct dcss_kms_dev *kms);
+int dcss_crtc_init(struct dcss_crtc *crtc, struct drm_device *drm);
+void dcss_crtc_deinit(struct dcss_crtc *crtc, struct drm_device *drm);
+struct dcss_plane *dcss_plane_init(struct drm_device *drm,
+				   unsigned int possible_crtcs,
+				   enum drm_plane_type type,
+				   unsigned int zpos);
+
+#endif
diff --git a/drivers/gpu/drm/imx/dcss/dcss-plane.c b/drivers/gpu/drm/imx/dcss/dcss-plane.c
new file mode 100644
index 000000000000..961d671f171b
--- /dev/null
+++ b/drivers/gpu/drm/imx/dcss/dcss-plane.c
@@ -0,0 +1,405 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Copyright 2019 NXP.
+ */
+
+#include <drm/drm_atomic.h>
+#include <drm/drm_atomic_helper.h>
+#include <drm/drm_fb_cma_helper.h>
+#include <drm/drm_gem_framebuffer_helper.h>
+#include <drm/drm_gem_cma_helper.h>
+
+#include "dcss-dev.h"
+#include "dcss-kms.h"
+
+static const u32 dcss_common_formats[] = {
+	/* RGB */
+	DRM_FORMAT_ARGB8888,
+	DRM_FORMAT_XRGB8888,
+	DRM_FORMAT_ABGR8888,
+	DRM_FORMAT_XBGR8888,
+	DRM_FORMAT_RGBA8888,
+	DRM_FORMAT_RGBX8888,
+	DRM_FORMAT_BGRA8888,
+	DRM_FORMAT_BGRX8888,
+	DRM_FORMAT_XRGB2101010,
+	DRM_FORMAT_XBGR2101010,
+	DRM_FORMAT_RGBX1010102,
+	DRM_FORMAT_BGRX1010102,
+	DRM_FORMAT_ARGB2101010,
+	DRM_FORMAT_ABGR2101010,
+	DRM_FORMAT_RGBA1010102,
+	DRM_FORMAT_BGRA1010102,
+};
+
+static const u64 dcss_video_format_modifiers[] = {
+	DRM_FORMAT_MOD_LINEAR,
+	DRM_FORMAT_MOD_INVALID,
+};
+
+static const u64 dcss_graphics_format_modifiers[] = {
+	DRM_FORMAT_MOD_VIVANTE_TILED,
+	DRM_FORMAT_MOD_VIVANTE_SUPER_TILED,
+	DRM_FORMAT_MOD_LINEAR,
+	DRM_FORMAT_MOD_INVALID,
+};
+
+static inline struct dcss_plane *to_dcss_plane(struct drm_plane *p)
+{
+	return container_of(p, struct dcss_plane, base);
+}
+
+static inline bool dcss_plane_fb_is_linear(const struct drm_framebuffer *fb)
+{
+	return ((fb->flags & DRM_MODE_FB_MODIFIERS) == 0) ||
+	       ((fb->flags & DRM_MODE_FB_MODIFIERS) != 0 &&
+		fb->modifier == DRM_FORMAT_MOD_LINEAR);
+}
+
+static void dcss_plane_destroy(struct drm_plane *plane)
+{
+	struct dcss_plane *dcss_plane = container_of(plane, struct dcss_plane,
+						     base);
+
+	drm_plane_cleanup(plane);
+	kfree(dcss_plane);
+}
+
+static bool dcss_plane_format_mod_supported(struct drm_plane *plane,
+					    u32 format,
+					    u64 modifier)
+{
+	switch (plane->type) {
+	case DRM_PLANE_TYPE_PRIMARY:
+		switch (format) {
+		case DRM_FORMAT_ARGB8888:
+		case DRM_FORMAT_XRGB8888:
+		case DRM_FORMAT_ARGB2101010:
+			return modifier == DRM_FORMAT_MOD_LINEAR ||
+			       modifier == DRM_FORMAT_MOD_VIVANTE_TILED ||
+			       modifier == DRM_FORMAT_MOD_VIVANTE_SUPER_TILED;
+		default:
+			return modifier == DRM_FORMAT_MOD_LINEAR;
+		}
+		break;
+	case DRM_PLANE_TYPE_OVERLAY:
+		return modifier == DRM_FORMAT_MOD_LINEAR;
+	default:
+		return false;
+	}
+}
+
+static const struct drm_plane_funcs dcss_plane_funcs = {
+	.update_plane		= drm_atomic_helper_update_plane,
+	.disable_plane		= drm_atomic_helper_disable_plane,
+	.destroy		= dcss_plane_destroy,
+	.reset			= drm_atomic_helper_plane_reset,
+	.atomic_duplicate_state	= drm_atomic_helper_plane_duplicate_state,
+	.atomic_destroy_state	= drm_atomic_helper_plane_destroy_state,
+	.format_mod_supported	= dcss_plane_format_mod_supported,
+};
+
+static bool dcss_plane_can_rotate(const struct drm_format_info *format,
+				  bool mod_present, u64 modifier,
+				  unsigned int rotation)
+{
+	bool linear_format = !mod_present ||
+			     (mod_present && modifier == DRM_FORMAT_MOD_LINEAR);
+	u32 supported_rotation = DRM_MODE_ROTATE_0;
+
+	if (!format->is_yuv && linear_format)
+		supported_rotation = DRM_MODE_ROTATE_0 | DRM_MODE_ROTATE_180 |
+				     DRM_MODE_REFLECT_MASK;
+	else if (!format->is_yuv &&
+		 modifier == DRM_FORMAT_MOD_VIVANTE_TILED)
+		supported_rotation = DRM_MODE_ROTATE_MASK |
+				     DRM_MODE_REFLECT_MASK;
+	else if (format->is_yuv && linear_format &&
+		 (format->format == DRM_FORMAT_NV12 ||
+		  format->format == DRM_FORMAT_NV21))
+		supported_rotation = DRM_MODE_ROTATE_0 | DRM_MODE_ROTATE_180 |
+				     DRM_MODE_REFLECT_MASK;
+
+	return !!(rotation & supported_rotation);
+}
+
+static bool dcss_plane_is_source_size_allowed(u16 src_w, u16 src_h, u32 pix_fmt)
+{
+	if (src_w < 64 &&
+	    (pix_fmt == DRM_FORMAT_NV12 || pix_fmt == DRM_FORMAT_NV21))
+		return false;
+	else if (src_w < 32 &&
+		 (pix_fmt == DRM_FORMAT_UYVY || pix_fmt == DRM_FORMAT_VYUY ||
+		  pix_fmt == DRM_FORMAT_YUYV || pix_fmt == DRM_FORMAT_YVYU))
+		return false;
+
+	return src_w >= 16 && src_h >= 8;
+}
+
+static int dcss_plane_atomic_check(struct drm_plane *plane,
+				   struct drm_plane_state *state)
+{
+	struct dcss_plane *dcss_plane = to_dcss_plane(plane);
+	struct dcss_dev *dcss = plane->dev->dev_private;
+	struct drm_framebuffer *fb = state->fb;
+	bool is_primary_plane = plane->type == DRM_PLANE_TYPE_PRIMARY;
+	struct drm_gem_cma_object *cma_obj;
+	struct drm_crtc_state *crtc_state;
+	int hdisplay, vdisplay;
+	int min, max;
+	int ret;
+
+	if (!fb || !state->crtc)
+		return 0;
+
+	cma_obj = drm_fb_cma_get_gem_obj(fb, 0);
+	WARN_ON(!cma_obj);
+
+	crtc_state = drm_atomic_get_existing_crtc_state(state->state,
+							state->crtc);
+
+	hdisplay = crtc_state->adjusted_mode.hdisplay;
+	vdisplay = crtc_state->adjusted_mode.vdisplay;
+
+	if (!dcss_plane_is_source_size_allowed(state->src_w >> 16,
+					       state->src_h >> 16,
+					       fb->format->format)) {
+		DRM_DEBUG_KMS("Source plane size is not allowed!\n");
+		return -EINVAL;
+	}
+
+	dcss_scaler_get_min_max_ratios(dcss->scaler, dcss_plane->ch_num,
+				       &min, &max);
+
+	ret = drm_atomic_helper_check_plane_state(state, crtc_state,
+						  min, max, !is_primary_plane,
+						  false);
+	if (ret)
+		return ret;
+
+	if (!state->visible)
+		return 0;
+
+	if (!dcss_plane_can_rotate(fb->format,
+				   !!(fb->flags & DRM_MODE_FB_MODIFIERS),
+				   fb->modifier,
+				   state->rotation)) {
+		DRM_DEBUG_KMS("requested rotation is not allowed!\n");
+		return -EINVAL;
+	}
+
+	if ((state->crtc_x < 0 || state->crtc_y < 0 ||
+	     state->crtc_x + state->crtc_w > hdisplay ||
+	     state->crtc_y + state->crtc_h > vdisplay) &&
+	    !dcss_plane_fb_is_linear(fb)) {
+		DRM_DEBUG_KMS("requested cropping operation is not allowed!\n");
+		return -EINVAL;
+	}
+
+	if ((fb->flags & DRM_MODE_FB_MODIFIERS) &&
+	    !plane->funcs->format_mod_supported(plane,
+				fb->format->format,
+				fb->modifier)) {
+		DRM_DEBUG_KMS("Invalid modifier: %llx", fb->modifier);
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static void dcss_plane_atomic_set_base(struct dcss_plane *dcss_plane)
+{
+	struct drm_plane *plane = &dcss_plane->base;
+	struct drm_plane_state *state = plane->state;
+	struct dcss_dev *dcss = plane->dev->dev_private;
+	struct drm_framebuffer *fb = state->fb;
+	const struct drm_format_info *format = fb->format;
+	struct drm_gem_cma_object *cma_obj = drm_fb_cma_get_gem_obj(fb, 0);
+	unsigned long p1_ba = 0, p2_ba = 0;
+
+	if (!format->is_yuv ||
+	    format->format == DRM_FORMAT_NV12 ||
+	    format->format == DRM_FORMAT_NV21)
+		p1_ba = cma_obj->paddr + fb->offsets[0] +
+			fb->pitches[0] * (state->src.y1 >> 16) +
+			format->char_per_block[0] * (state->src.x1 >> 16);
+	else if (format->format == DRM_FORMAT_UYVY ||
+		 format->format == DRM_FORMAT_VYUY ||
+		 format->format == DRM_FORMAT_YUYV ||
+		 format->format == DRM_FORMAT_YVYU)
+		p1_ba = cma_obj->paddr + fb->offsets[0] +
+			fb->pitches[0] * (state->src.y1 >> 16) +
+			2 * format->char_per_block[0] * (state->src.x1 >> 17);
+
+	if (format->format == DRM_FORMAT_NV12 ||
+	    format->format == DRM_FORMAT_NV21)
+		p2_ba = cma_obj->paddr + fb->offsets[1] +
+			(((fb->pitches[1] >> 1) * (state->src.y1 >> 17) +
+			(state->src.x1 >> 17)) << 1);
+
+	dcss_dpr_addr_set(dcss->dpr, dcss_plane->ch_num, p1_ba, p2_ba,
+			  fb->pitches[0]);
+}
+
+static bool dcss_plane_needs_setup(struct drm_plane_state *state,
+				   struct drm_plane_state *old_state)
+{
+	struct drm_framebuffer *fb = state->fb;
+	struct drm_framebuffer *old_fb = old_state->fb;
+
+	return state->crtc_x != old_state->crtc_x ||
+	       state->crtc_y != old_state->crtc_y ||
+	       state->crtc_w != old_state->crtc_w ||
+	       state->crtc_h != old_state->crtc_h ||
+	       state->src_x  != old_state->src_x  ||
+	       state->src_y  != old_state->src_y  ||
+	       state->src_w  != old_state->src_w  ||
+	       state->src_h  != old_state->src_h  ||
+	       fb->format->format != old_fb->format->format ||
+	       fb->modifier  != old_fb->modifier ||
+	       state->rotation != old_state->rotation;
+}
+
+static void dcss_plane_atomic_update(struct drm_plane *plane,
+				     struct drm_plane_state *old_state)
+{
+	struct drm_plane_state *state = plane->state;
+	struct dcss_plane *dcss_plane = to_dcss_plane(plane);
+	struct dcss_dev *dcss = plane->dev->dev_private;
+	struct drm_framebuffer *fb = state->fb;
+	u32 pixel_format;
+	struct drm_crtc_state *crtc_state;
+	bool modifiers_present;
+	u32 src_w, src_h, dst_w, dst_h;
+	struct drm_rect src, dst;
+	bool enable = true;
+
+	if (!fb || !state->crtc || !state->visible)
+		return;
+
+	pixel_format = state->fb->format->format;
+	crtc_state = state->crtc->state;
+	modifiers_present = !!(fb->flags & DRM_MODE_FB_MODIFIERS);
+
+	if (old_state->fb && !drm_atomic_crtc_needs_modeset(crtc_state) &&
+	    !dcss_plane_needs_setup(state, old_state)) {
+		dcss_plane_atomic_set_base(dcss_plane);
+		return;
+	}
+
+	src = plane->state->src;
+	dst = plane->state->dst;
+
+	/*
+	 * The width and height after clipping.
+	 */
+	src_w = drm_rect_width(&src) >> 16;
+	src_h = drm_rect_height(&src) >> 16;
+	dst_w = drm_rect_width(&dst);
+	dst_h = drm_rect_height(&dst);
+
+	if (plane->type == DRM_PLANE_TYPE_OVERLAY &&
+	    modifiers_present && fb->modifier == DRM_FORMAT_MOD_LINEAR)
+		modifiers_present = false;
+
+	dcss_dpr_format_set(dcss->dpr, dcss_plane->ch_num, state->fb->format,
+			    modifiers_present ? fb->modifier :
+						DRM_FORMAT_MOD_LINEAR);
+	dcss_dpr_set_res(dcss->dpr, dcss_plane->ch_num, src_w, src_h);
+	dcss_dpr_set_rotation(dcss->dpr, dcss_plane->ch_num,
+			      state->rotation);
+
+	dcss_plane_atomic_set_base(dcss_plane);
+
+	dcss_scaler_setup(dcss->scaler, dcss_plane->ch_num,
+			  state->fb->format, src_w, src_h,
+			  dst_w, dst_h,
+			  drm_mode_vrefresh(&crtc_state->mode));
+
+	dcss_dtg_plane_pos_set(dcss->dtg, dcss_plane->ch_num,
+			       dst.x1, dst.y1, dst_w, dst_h);
+	dcss_dtg_plane_alpha_set(dcss->dtg, dcss_plane->ch_num,
+				 fb->format, state->alpha >> 8);
+
+	if (!dcss_plane->ch_num && (state->alpha >> 8) == 0)
+		enable = false;
+
+	dcss_dpr_enable(dcss->dpr, dcss_plane->ch_num, enable);
+	dcss_scaler_ch_enable(dcss->scaler, dcss_plane->ch_num, enable);
+
+	if (!enable)
+		dcss_dtg_plane_pos_set(dcss->dtg, dcss_plane->ch_num,
+				       0, 0, 0, 0);
+
+	dcss_dtg_ch_enable(dcss->dtg, dcss_plane->ch_num, enable);
+}
+
+static void dcss_plane_atomic_disable(struct drm_plane *plane,
+				      struct drm_plane_state *old_state)
+{
+	struct dcss_plane *dcss_plane = to_dcss_plane(plane);
+	struct dcss_dev *dcss = plane->dev->dev_private;
+
+	dcss_dpr_enable(dcss->dpr, dcss_plane->ch_num, false);
+	dcss_scaler_ch_enable(dcss->scaler, dcss_plane->ch_num, false);
+	dcss_dtg_plane_pos_set(dcss->dtg, dcss_plane->ch_num, 0, 0, 0, 0);
+	dcss_dtg_ch_enable(dcss->dtg, dcss_plane->ch_num, false);
+}
+
+static const struct drm_plane_helper_funcs dcss_plane_helper_funcs = {
+	.prepare_fb = drm_gem_fb_prepare_fb,
+	.atomic_check = dcss_plane_atomic_check,
+	.atomic_update = dcss_plane_atomic_update,
+	.atomic_disable = dcss_plane_atomic_disable,
+};
+
+struct dcss_plane *dcss_plane_init(struct drm_device *drm,
+				   unsigned int possible_crtcs,
+				   enum drm_plane_type type,
+				   unsigned int zpos)
+{
+	struct dcss_plane *dcss_plane;
+	const u64 *format_modifiers = dcss_video_format_modifiers;
+	int ret;
+
+	if (zpos > 2)
+		return ERR_PTR(-EINVAL);
+
+	dcss_plane = kzalloc(sizeof(*dcss_plane), GFP_KERNEL);
+	if (!dcss_plane) {
+		DRM_ERROR("failed to allocate plane\n");
+		return ERR_PTR(-ENOMEM);
+	}
+
+	if (type == DRM_PLANE_TYPE_PRIMARY)
+		format_modifiers = dcss_graphics_format_modifiers;
+
+	ret = drm_universal_plane_init(drm, &dcss_plane->base, possible_crtcs,
+				       &dcss_plane_funcs, dcss_common_formats,
+				       ARRAY_SIZE(dcss_common_formats),
+				       format_modifiers, type, NULL);
+	if (ret) {
+		DRM_ERROR("failed to initialize plane\n");
+		kfree(dcss_plane);
+		return ERR_PTR(ret);
+	}
+
+	drm_plane_helper_add(&dcss_plane->base, &dcss_plane_helper_funcs);
+
+	ret = drm_plane_create_zpos_immutable_property(&dcss_plane->base, zpos);
+	if (ret)
+		return ERR_PTR(ret);
+
+	drm_plane_create_rotation_property(&dcss_plane->base,
+					   DRM_MODE_ROTATE_0,
+					   DRM_MODE_ROTATE_0   |
+					   DRM_MODE_ROTATE_90  |
+					   DRM_MODE_ROTATE_180 |
+					   DRM_MODE_ROTATE_270 |
+					   DRM_MODE_REFLECT_X  |
+					   DRM_MODE_REFLECT_Y);
+
+	dcss_plane->ch_num = zpos;
+
+	return dcss_plane;
+}
diff --git a/drivers/gpu/drm/imx/dcss/dcss-scaler.c b/drivers/gpu/drm/imx/dcss/dcss-scaler.c
new file mode 100644
index 000000000000..cd21905de580
--- /dev/null
+++ b/drivers/gpu/drm/imx/dcss/dcss-scaler.c
@@ -0,0 +1,826 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Copyright 2019 NXP.
+ *
+ * Scaling algorithms were contributed by Dzung Hoang <dzung.hoang@nxp.com>
+ */
+
+#include <linux/device.h>
+#include <linux/slab.h>
+
+#include "dcss-dev.h"
+
+#define DCSS_SCALER_CTRL			0x00
+#define   SCALER_EN				BIT(0)
+#define   REPEAT_EN				BIT(4)
+#define   SCALE2MEM_EN				BIT(8)
+#define   MEM2OFIFO_EN				BIT(12)
+#define DCSS_SCALER_OFIFO_CTRL			0x04
+#define   OFIFO_LOW_THRES_POS			0
+#define   OFIFO_LOW_THRES_MASK			GENMASK(9, 0)
+#define   OFIFO_HIGH_THRES_POS			16
+#define   OFIFO_HIGH_THRES_MASK			GENMASK(25, 16)
+#define   UNDERRUN_DETECT_CLR			BIT(26)
+#define   LOW_THRES_DETECT_CLR			BIT(27)
+#define   HIGH_THRES_DETECT_CLR			BIT(28)
+#define   UNDERRUN_DETECT_EN			BIT(29)
+#define   LOW_THRES_DETECT_EN			BIT(30)
+#define   HIGH_THRES_DETECT_EN			BIT(31)
+#define DCSS_SCALER_SDATA_CTRL			0x08
+#define   YUV_EN				BIT(0)
+#define   RTRAM_8LINES				BIT(1)
+#define   Y_UV_BYTE_SWAP			BIT(4)
+#define   A2R10G10B10_FORMAT_POS		8
+#define   A2R10G10B10_FORMAT_MASK		GENMASK(11, 8)
+#define DCSS_SCALER_BIT_DEPTH			0x0C
+#define   LUM_BIT_DEPTH_POS			0
+#define   LUM_BIT_DEPTH_MASK			GENMASK(1, 0)
+#define   CHR_BIT_DEPTH_POS			4
+#define   CHR_BIT_DEPTH_MASK			GENMASK(5, 4)
+#define DCSS_SCALER_SRC_FORMAT			0x10
+#define DCSS_SCALER_DST_FORMAT			0x14
+#define   FORMAT_MASK				GENMASK(1, 0)
+#define DCSS_SCALER_SRC_LUM_RES			0x18
+#define DCSS_SCALER_SRC_CHR_RES			0x1C
+#define DCSS_SCALER_DST_LUM_RES			0x20
+#define DCSS_SCALER_DST_CHR_RES			0x24
+#define   WIDTH_POS				0
+#define   WIDTH_MASK				GENMASK(11, 0)
+#define   HEIGHT_POS				16
+#define   HEIGHT_MASK				GENMASK(27, 16)
+#define DCSS_SCALER_V_LUM_START			0x48
+#define   V_START_MASK				GENMASK(15, 0)
+#define DCSS_SCALER_V_LUM_INC			0x4C
+#define   V_INC_MASK				GENMASK(15, 0)
+#define DCSS_SCALER_H_LUM_START			0x50
+#define   H_START_MASK				GENMASK(18, 0)
+#define DCSS_SCALER_H_LUM_INC			0x54
+#define   H_INC_MASK				GENMASK(15, 0)
+#define DCSS_SCALER_V_CHR_START			0x58
+#define DCSS_SCALER_V_CHR_INC			0x5C
+#define DCSS_SCALER_H_CHR_START			0x60
+#define DCSS_SCALER_H_CHR_INC			0x64
+#define DCSS_SCALER_COEF_VLUM			0x80
+#define DCSS_SCALER_COEF_HLUM			0x140
+#define DCSS_SCALER_COEF_VCHR			0x200
+#define DCSS_SCALER_COEF_HCHR			0x300
+
+struct dcss_scaler_ch {
+	void __iomem *base_reg;
+	u32 base_ofs;
+	struct dcss_scaler *scl;
+
+	u32 sdata_ctrl;
+	u32 scaler_ctrl;
+
+	bool scaler_ctrl_chgd;
+
+	u32 c_vstart;
+	u32 c_hstart;
+};
+
+struct dcss_scaler {
+	struct device *dev;
+
+	struct dcss_ctxld *ctxld;
+	u32 ctx_id;
+
+	struct dcss_scaler_ch ch[3];
+};
+
+/* scaler coefficients generator */
+#define PSC_FRAC_BITS 30
+#define PSC_FRAC_SCALE BIT(PSC_FRAC_BITS)
+#define PSC_BITS_FOR_PHASE 4
+#define PSC_NUM_PHASES 16
+#define PSC_STORED_PHASES (PSC_NUM_PHASES / 2 + 1)
+#define PSC_NUM_TAPS 7
+#define PSC_NUM_TAPS_RGBA 5
+#define PSC_COEFF_PRECISION 10
+#define PSC_PHASE_FRACTION_BITS 13
+#define PSC_PHASE_MASK (PSC_NUM_PHASES - 1)
+#define PSC_Q_FRACTION 19
+#define PSC_Q_ROUND_OFFSET (1 << (PSC_Q_FRACTION - 1))
+
+/**
+ * mult_q() - Performs fixed-point multiplication.
+ * @A: multiplier
+ * @B: multiplicand
+ */
+static int mult_q(int A, int B)
+{
+	int result;
+	s64 temp;
+
+	temp = (int64_t)A * (int64_t)B;
+	temp += PSC_Q_ROUND_OFFSET;
+	result = (int)(temp >> PSC_Q_FRACTION);
+	return result;
+}
+
+/**
+ * div_q() - Performs fixed-point division.
+ * @A: dividend
+ * @B: divisor
+ */
+static int div_q(int A, int B)
+{
+	int result;
+	s64 temp;
+
+	temp = (int64_t)A << PSC_Q_FRACTION;
+	if ((temp >= 0 && B >= 0) || (temp < 0 && B < 0))
+		temp += B / 2;
+	else
+		temp -= B / 2;
+
+	result = (int)(temp / B);
+	return result;
+}
+
+/**
+ * exp_approx_q() - Compute approximation to exp(x) function using Taylor
+ *		    series.
+ * @x: fixed-point argument of exp function
+ */
+static int exp_approx_q(int x)
+{
+	int sum = 1 << PSC_Q_FRACTION;
+	int term = 1 << PSC_Q_FRACTION;
+
+	term = mult_q(term, div_q(x, 1 << PSC_Q_FRACTION));
+	sum += term;
+	term = mult_q(term, div_q(x, 2 << PSC_Q_FRACTION));
+	sum += term;
+	term = mult_q(term, div_q(x, 3 << PSC_Q_FRACTION));
+	sum += term;
+	term = mult_q(term, div_q(x, 4 << PSC_Q_FRACTION));
+	sum += term;
+
+	return sum;
+}
+
+/**
+ * dcss_scaler_gaussian_filter() - Generate gaussian prototype filter.
+ * @fc_q: fixed-point cutoff frequency normalized to range [0, 1]
+ * @use_5_taps: indicates whether to use 5 taps or 7 taps
+ * @coef: output filter coefficients
+ */
+static void dcss_scaler_gaussian_filter(int fc_q, bool use_5_taps,
+					bool phase0_identity,
+					int coef[][PSC_NUM_TAPS])
+{
+	int sigma_q, g0_q, g1_q, g2_q;
+	int tap_cnt1, tap_cnt2, tap_idx, phase_cnt;
+	int mid;
+	int phase;
+	int i;
+	int taps;
+
+	if (use_5_taps)
+		for (phase = 0; phase < PSC_STORED_PHASES; phase++) {
+			coef[phase][0] = 0;
+			coef[phase][PSC_NUM_TAPS - 1] = 0;
+		}
+
+	/* seed coefficient scanner */
+	taps = use_5_taps ? PSC_NUM_TAPS_RGBA : PSC_NUM_TAPS;
+	mid = (PSC_NUM_PHASES * taps) / 2 - 1;
+	phase_cnt = (PSC_NUM_PHASES * (PSC_NUM_TAPS + 1)) / 2;
+	tap_cnt1 = (PSC_NUM_PHASES * PSC_NUM_TAPS) / 2;
+	tap_cnt2 = (PSC_NUM_PHASES * PSC_NUM_TAPS) / 2;
+
+	/* seed gaussian filter generator */
+	sigma_q = div_q(PSC_Q_ROUND_OFFSET, fc_q);
+	g0_q = 1 << PSC_Q_FRACTION;
+	g1_q = exp_approx_q(div_q(-PSC_Q_ROUND_OFFSET,
+				  mult_q(sigma_q, sigma_q)));
+	g2_q = mult_q(g1_q, g1_q);
+	coef[phase_cnt & PSC_PHASE_MASK][tap_cnt1 >> PSC_BITS_FOR_PHASE] = g0_q;
+
+	for (i = 0; i < mid; i++) {
+		phase_cnt++;
+		tap_cnt1--;
+		tap_cnt2++;
+
+		g0_q = mult_q(g0_q, g1_q);
+		g1_q = mult_q(g1_q, g2_q);
+
+		if ((phase_cnt & PSC_PHASE_MASK) <= 8) {
+			tap_idx = tap_cnt1 >> PSC_BITS_FOR_PHASE;
+			coef[phase_cnt & PSC_PHASE_MASK][tap_idx] = g0_q;
+		}
+		if (((-phase_cnt) & PSC_PHASE_MASK) <= 8) {
+			tap_idx = tap_cnt2 >> PSC_BITS_FOR_PHASE;
+			coef[(-phase_cnt) & PSC_PHASE_MASK][tap_idx] = g0_q;
+		}
+	}
+
+	phase_cnt++;
+	tap_cnt1--;
+	coef[phase_cnt & PSC_PHASE_MASK][tap_cnt1 >> PSC_BITS_FOR_PHASE] = 0;
+
+	/* override phase 0 with identity filter if specified */
+	if (phase0_identity)
+		for (i = 0; i < PSC_NUM_TAPS; i++)
+			coef[0][i] = i == (PSC_NUM_TAPS >> 1) ?
+						(1 << PSC_COEFF_PRECISION) : 0;
+
+	/* normalize coef */
+	for (phase = 0; phase < PSC_STORED_PHASES; phase++) {
+		int sum = 0;
+		s64 ll_temp;
+
+		for (i = 0; i < PSC_NUM_TAPS; i++)
+			sum += coef[phase][i];
+		for (i = 0; i < PSC_NUM_TAPS; i++) {
+			ll_temp = coef[phase][i];
+			ll_temp <<= PSC_COEFF_PRECISION;
+			ll_temp += sum >> 1;
+			ll_temp /= sum;
+			coef[phase][i] = (int)ll_temp;
+		}
+	}
+}
+
+/**
+ * dcss_scaler_filter_design() - Compute filter coefficients using
+ *				 Gaussian filter.
+ * @src_length: length of input
+ * @dst_length: length of output
+ * @use_5_taps: 0 for 7 taps per phase, 1 for 5 taps
+ * @coef: output coefficients
+ */
+static void dcss_scaler_filter_design(int src_length, int dst_length,
+				      bool use_5_taps, bool phase0_identity,
+				      int coef[][PSC_NUM_TAPS])
+{
+	int fc_q;
+
+	/* compute cutoff frequency */
+	if (dst_length >= src_length)
+		fc_q = div_q(1, PSC_NUM_PHASES);
+	else
+		fc_q = div_q(dst_length, src_length * PSC_NUM_PHASES);
+
+	/* compute gaussian filter coefficients */
+	dcss_scaler_gaussian_filter(fc_q, use_5_taps, phase0_identity, coef);
+}
+
+static void dcss_scaler_write(struct dcss_scaler_ch *ch, u32 val, u32 ofs)
+{
+	struct dcss_scaler *scl = ch->scl;
+
+	dcss_ctxld_write(scl->ctxld, scl->ctx_id, val, ch->base_ofs + ofs);
+}
+
+static int dcss_scaler_ch_init_all(struct dcss_scaler *scl,
+				   unsigned long scaler_base)
+{
+	struct dcss_scaler_ch *ch;
+	int i;
+
+	for (i = 0; i < 3; i++) {
+		ch = &scl->ch[i];
+
+		ch->base_ofs = scaler_base + i * 0x400;
+
+		ch->base_reg = ioremap(ch->base_ofs, SZ_4K);
+		if (!ch->base_reg) {
+			dev_err(scl->dev, "scaler: unable to remap ch base\n");
+			return -ENOMEM;
+		}
+
+		ch->scl = scl;
+	}
+
+	return 0;
+}
+
+int dcss_scaler_init(struct dcss_dev *dcss, unsigned long scaler_base)
+{
+	struct dcss_scaler *scaler;
+
+	scaler = kzalloc(sizeof(*scaler), GFP_KERNEL);
+	if (!scaler)
+		return -ENOMEM;
+
+	dcss->scaler = scaler;
+	scaler->dev = dcss->dev;
+	scaler->ctxld = dcss->ctxld;
+	scaler->ctx_id = CTX_SB_HP;
+
+	if (dcss_scaler_ch_init_all(scaler, scaler_base)) {
+		int i;
+
+		for (i = 0; i < 3; i++) {
+			if (scaler->ch[i].base_reg)
+				iounmap(scaler->ch[i].base_reg);
+		}
+
+		kfree(scaler);
+
+		return -ENOMEM;
+	}
+
+	return 0;
+}
+
+void dcss_scaler_exit(struct dcss_scaler *scl)
+{
+	int ch_no;
+
+	for (ch_no = 0; ch_no < 3; ch_no++) {
+		struct dcss_scaler_ch *ch = &scl->ch[ch_no];
+
+		dcss_writel(0, ch->base_reg + DCSS_SCALER_CTRL);
+
+		if (ch->base_reg)
+			iounmap(ch->base_reg);
+	}
+
+	kfree(scl);
+}
+
+void dcss_scaler_ch_enable(struct dcss_scaler *scl, int ch_num, bool en)
+{
+	struct dcss_scaler_ch *ch = &scl->ch[ch_num];
+	u32 scaler_ctrl;
+
+	scaler_ctrl = en ? SCALER_EN | REPEAT_EN : 0;
+
+	if (en)
+		dcss_scaler_write(ch, ch->sdata_ctrl, DCSS_SCALER_SDATA_CTRL);
+
+	if (ch->scaler_ctrl != scaler_ctrl)
+		ch->scaler_ctrl_chgd = true;
+
+	ch->scaler_ctrl = scaler_ctrl;
+}
+
+static void dcss_scaler_yuv_enable(struct dcss_scaler_ch *ch, bool en)
+{
+	ch->sdata_ctrl &= ~YUV_EN;
+	ch->sdata_ctrl |= en ? YUV_EN : 0;
+}
+
+static void dcss_scaler_rtr_8lines_enable(struct dcss_scaler_ch *ch, bool en)
+{
+	ch->sdata_ctrl &= ~RTRAM_8LINES;
+	ch->sdata_ctrl |= en ? RTRAM_8LINES : 0;
+}
+
+static void dcss_scaler_bit_depth_set(struct dcss_scaler_ch *ch, int depth)
+{
+	u32 val;
+
+	val = depth == 30 ? 2 : 0;
+
+	dcss_scaler_write(ch,
+			  ((val << CHR_BIT_DEPTH_POS) & CHR_BIT_DEPTH_MASK) |
+			  ((val << LUM_BIT_DEPTH_POS) & LUM_BIT_DEPTH_MASK),
+			  DCSS_SCALER_BIT_DEPTH);
+}
+
+enum buffer_format {
+	BUF_FMT_YUV420,
+	BUF_FMT_YUV422,
+	BUF_FMT_ARGB8888_YUV444,
+};
+
+enum chroma_location {
+	PSC_LOC_HORZ_0_VERT_1_OVER_4 = 0,
+	PSC_LOC_HORZ_1_OVER_4_VERT_1_OVER_4 = 1,
+	PSC_LOC_HORZ_0_VERT_0 = 2,
+	PSC_LOC_HORZ_1_OVER_4_VERT_0 = 3,
+	PSC_LOC_HORZ_0_VERT_1_OVER_2 = 4,
+	PSC_LOC_HORZ_1_OVER_4_VERT_1_OVER_2 = 5
+};
+
+static void dcss_scaler_format_set(struct dcss_scaler_ch *ch,
+				   enum buffer_format src_fmt,
+				   enum buffer_format dst_fmt)
+{
+	dcss_scaler_write(ch, src_fmt, DCSS_SCALER_SRC_FORMAT);
+	dcss_scaler_write(ch, dst_fmt, DCSS_SCALER_DST_FORMAT);
+}
+
+static void dcss_scaler_res_set(struct dcss_scaler_ch *ch,
+				int src_xres, int src_yres,
+				int dst_xres, int dst_yres,
+				u32 pix_format, enum buffer_format dst_format)
+{
+	u32 lsrc_xres, lsrc_yres, csrc_xres, csrc_yres;
+	u32 ldst_xres, ldst_yres, cdst_xres, cdst_yres;
+	bool src_is_444 = true;
+
+	lsrc_xres = src_xres;
+	csrc_xres = src_xres;
+	lsrc_yres = src_yres;
+	csrc_yres = src_yres;
+	ldst_xres = dst_xres;
+	cdst_xres = dst_xres;
+	ldst_yres = dst_yres;
+	cdst_yres = dst_yres;
+
+	if (pix_format == DRM_FORMAT_UYVY || pix_format == DRM_FORMAT_VYUY ||
+	    pix_format == DRM_FORMAT_YUYV || pix_format == DRM_FORMAT_YVYU) {
+		csrc_xres >>= 1;
+		src_is_444 = false;
+	} else if (pix_format == DRM_FORMAT_NV12 ||
+		   pix_format == DRM_FORMAT_NV21) {
+		csrc_xres >>= 1;
+		csrc_yres >>= 1;
+		src_is_444 = false;
+	}
+
+	if (dst_format == BUF_FMT_YUV422)
+		cdst_xres >>= 1;
+
+	/* for 4:4:4 to 4:2:2 conversion, source height should be 1 less */
+	if (src_is_444 && dst_format == BUF_FMT_YUV422) {
+		lsrc_yres--;
+		csrc_yres--;
+	}
+
+	dcss_scaler_write(ch, (((lsrc_yres - 1) << HEIGHT_POS) & HEIGHT_MASK) |
+			       (((lsrc_xres - 1) << WIDTH_POS) & WIDTH_MASK),
+			  DCSS_SCALER_SRC_LUM_RES);
+	dcss_scaler_write(ch, (((csrc_yres - 1) << HEIGHT_POS) & HEIGHT_MASK) |
+			       (((csrc_xres - 1) << WIDTH_POS) & WIDTH_MASK),
+			  DCSS_SCALER_SRC_CHR_RES);
+	dcss_scaler_write(ch, (((ldst_yres - 1) << HEIGHT_POS) & HEIGHT_MASK) |
+			       (((ldst_xres - 1) << WIDTH_POS) & WIDTH_MASK),
+			  DCSS_SCALER_DST_LUM_RES);
+	dcss_scaler_write(ch, (((cdst_yres - 1) << HEIGHT_POS) & HEIGHT_MASK) |
+			       (((cdst_xres - 1) << WIDTH_POS) & WIDTH_MASK),
+			  DCSS_SCALER_DST_CHR_RES);
+}
+
+#define downscale_fp(factor, fp_pos)		((factor) << (fp_pos))
+#define upscale_fp(factor, fp_pos)		((1 << (fp_pos)) / (factor))
+
+struct dcss_scaler_factors {
+	int downscale;
+	int upscale;
+};
+
+static const struct dcss_scaler_factors dcss_scaler_factors[] = {
+	{3, 8}, {5, 8}, {5, 8},
+};
+
+static void dcss_scaler_fractions_set(struct dcss_scaler_ch *ch,
+				      int src_xres, int src_yres,
+				      int dst_xres, int dst_yres,
+				      u32 src_format, u32 dst_format,
+				      enum chroma_location src_chroma_loc)
+{
+	int src_c_xres, src_c_yres, dst_c_xres, dst_c_yres;
+	u32 l_vinc, l_hinc, c_vinc, c_hinc;
+	u32 c_vstart, c_hstart;
+
+	src_c_xres = src_xres;
+	src_c_yres = src_yres;
+	dst_c_xres = dst_xres;
+	dst_c_yres = dst_yres;
+
+	c_vstart = 0;
+	c_hstart = 0;
+
+	/* adjustments for source chroma location */
+	if (src_format == BUF_FMT_YUV420) {
+		/* vertical input chroma position adjustment */
+		switch (src_chroma_loc) {
+		case PSC_LOC_HORZ_0_VERT_1_OVER_4:
+		case PSC_LOC_HORZ_1_OVER_4_VERT_1_OVER_4:
+			/*
+			 * move chroma up to first luma line
+			 * (1/4 chroma input line spacing)
+			 */
+			c_vstart -= (1 << (PSC_PHASE_FRACTION_BITS - 2));
+			break;
+		case PSC_LOC_HORZ_0_VERT_1_OVER_2:
+		case PSC_LOC_HORZ_1_OVER_4_VERT_1_OVER_2:
+			/*
+			 * move chroma up to first luma line
+			 * (1/2 chroma input line spacing)
+			 */
+			c_vstart -= (1 << (PSC_PHASE_FRACTION_BITS - 1));
+			break;
+		default:
+			break;
+		}
+		/* horizontal input chroma position adjustment */
+		switch (src_chroma_loc) {
+		case PSC_LOC_HORZ_1_OVER_4_VERT_1_OVER_4:
+		case PSC_LOC_HORZ_1_OVER_4_VERT_0:
+		case PSC_LOC_HORZ_1_OVER_4_VERT_1_OVER_2:
+			/* move chroma left 1/4 chroma input sample spacing */
+			c_hstart -= (1 << (PSC_PHASE_FRACTION_BITS - 2));
+			break;
+		default:
+			break;
+		}
+	}
+
+	/* adjustments to chroma resolution */
+	if (src_format == BUF_FMT_YUV420) {
+		src_c_xres >>= 1;
+		src_c_yres >>= 1;
+	} else if (src_format == BUF_FMT_YUV422) {
+		src_c_xres >>= 1;
+	}
+
+	if (dst_format == BUF_FMT_YUV422)
+		dst_c_xres >>= 1;
+
+	l_vinc = ((src_yres << 13) + (dst_yres >> 1)) / dst_yres;
+	c_vinc = ((src_c_yres << 13) + (dst_c_yres >> 1)) / dst_c_yres;
+	l_hinc = ((src_xres << 13) + (dst_xres >> 1)) / dst_xres;
+	c_hinc = ((src_c_xres << 13) + (dst_c_xres >> 1)) / dst_c_xres;
+
+	/* save chroma start phase */
+	ch->c_vstart = c_vstart;
+	ch->c_hstart = c_hstart;
+
+	dcss_scaler_write(ch, 0, DCSS_SCALER_V_LUM_START);
+	dcss_scaler_write(ch, l_vinc, DCSS_SCALER_V_LUM_INC);
+
+	dcss_scaler_write(ch, 0, DCSS_SCALER_H_LUM_START);
+	dcss_scaler_write(ch, l_hinc, DCSS_SCALER_H_LUM_INC);
+
+	dcss_scaler_write(ch, c_vstart, DCSS_SCALER_V_CHR_START);
+	dcss_scaler_write(ch, c_vinc, DCSS_SCALER_V_CHR_INC);
+
+	dcss_scaler_write(ch, c_hstart, DCSS_SCALER_H_CHR_START);
+	dcss_scaler_write(ch, c_hinc, DCSS_SCALER_H_CHR_INC);
+}
+
+int dcss_scaler_get_min_max_ratios(struct dcss_scaler *scl, int ch_num,
+				   int *min, int *max)
+{
+	*min = upscale_fp(dcss_scaler_factors[ch_num].upscale, 16);
+	*max = downscale_fp(dcss_scaler_factors[ch_num].downscale, 16);
+
+	return 0;
+}
+
+static void dcss_scaler_program_5_coef_set(struct dcss_scaler_ch *ch,
+					   int base_addr,
+					   int coef[][PSC_NUM_TAPS])
+{
+	int i, phase;
+
+	for (i = 0; i < PSC_STORED_PHASES; i++) {
+		dcss_scaler_write(ch, ((coef[i][1] & 0xfff) << 16 |
+				       (coef[i][2] & 0xfff) << 4  |
+				       (coef[i][3] & 0xf00) >> 8),
+				  base_addr + i * sizeof(u32));
+		dcss_scaler_write(ch, ((coef[i][3] & 0x0ff) << 20 |
+				       (coef[i][4] & 0xfff) << 8  |
+				       (coef[i][5] & 0xff0) >> 4),
+				  base_addr + 0x40 + i * sizeof(u32));
+		dcss_scaler_write(ch, ((coef[i][5] & 0x00f) << 24),
+				  base_addr + 0x80 + i * sizeof(u32));
+	}
+
+	/* reverse both phase and tap orderings */
+	for (phase = (PSC_NUM_PHASES >> 1) - 1;
+			i < PSC_NUM_PHASES; i++, phase--) {
+		dcss_scaler_write(ch, ((coef[phase][5] & 0xfff) << 16 |
+				       (coef[phase][4] & 0xfff) << 4  |
+				       (coef[phase][3] & 0xf00) >> 8),
+				  base_addr + i * sizeof(u32));
+		dcss_scaler_write(ch, ((coef[phase][3] & 0x0ff) << 20 |
+				       (coef[phase][2] & 0xfff) << 8  |
+				       (coef[phase][1] & 0xff0) >> 4),
+				  base_addr + 0x40 + i * sizeof(u32));
+		dcss_scaler_write(ch, ((coef[phase][1] & 0x00f) << 24),
+				  base_addr + 0x80 + i * sizeof(u32));
+	}
+}
+
+static void dcss_scaler_program_7_coef_set(struct dcss_scaler_ch *ch,
+					   int base_addr,
+					   int coef[][PSC_NUM_TAPS])
+{
+	int i, phase;
+
+	for (i = 0; i < PSC_STORED_PHASES; i++) {
+		dcss_scaler_write(ch, ((coef[i][0] & 0xfff) << 16 |
+				       (coef[i][1] & 0xfff) << 4  |
+				       (coef[i][2] & 0xf00) >> 8),
+				  base_addr + i * sizeof(u32));
+		dcss_scaler_write(ch, ((coef[i][2] & 0x0ff) << 20 |
+				       (coef[i][3] & 0xfff) << 8  |
+				       (coef[i][4] & 0xff0) >> 4),
+				  base_addr + 0x40 + i * sizeof(u32));
+		dcss_scaler_write(ch, ((coef[i][4] & 0x00f) << 24 |
+				       (coef[i][5] & 0xfff) << 12 |
+				       (coef[i][6] & 0xfff)),
+				  base_addr + 0x80 + i * sizeof(u32));
+	}
+
+	/* reverse both phase and tap orderings */
+	for (phase = (PSC_NUM_PHASES >> 1) - 1;
+			i < PSC_NUM_PHASES; i++, phase--) {
+		dcss_scaler_write(ch, ((coef[phase][6] & 0xfff) << 16 |
+				       (coef[phase][5] & 0xfff) << 4  |
+				       (coef[phase][4] & 0xf00) >> 8),
+				  base_addr + i * sizeof(u32));
+		dcss_scaler_write(ch, ((coef[phase][4] & 0x0ff) << 20 |
+				       (coef[phase][3] & 0xfff) << 8  |
+				       (coef[phase][2] & 0xff0) >> 4),
+				  base_addr + 0x40 + i * sizeof(u32));
+		dcss_scaler_write(ch, ((coef[phase][2] & 0x00f) << 24 |
+				       (coef[phase][1] & 0xfff) << 12 |
+				       (coef[phase][0] & 0xfff)),
+				  base_addr + 0x80 + i * sizeof(u32));
+	}
+}
+
+static void dcss_scaler_yuv_coef_set(struct dcss_scaler_ch *ch,
+				     enum buffer_format src_format,
+				     enum buffer_format dst_format,
+				     bool use_5_taps,
+				     int src_xres, int src_yres, int dst_xres,
+				     int dst_yres)
+{
+	int coef[PSC_STORED_PHASES][PSC_NUM_TAPS];
+	bool program_5_taps = use_5_taps ||
+			      (dst_format == BUF_FMT_YUV422 &&
+			       src_format == BUF_FMT_ARGB8888_YUV444);
+
+	/* horizontal luma */
+	dcss_scaler_filter_design(src_xres, dst_xres, false,
+				  src_xres == dst_xres, coef);
+	dcss_scaler_program_7_coef_set(ch, DCSS_SCALER_COEF_HLUM, coef);
+
+	/* vertical luma */
+	dcss_scaler_filter_design(src_yres, dst_yres, program_5_taps,
+				  src_yres == dst_yres, coef);
+
+	if (program_5_taps)
+		dcss_scaler_program_5_coef_set(ch, DCSS_SCALER_COEF_VLUM, coef);
+	else
+		dcss_scaler_program_7_coef_set(ch, DCSS_SCALER_COEF_VLUM, coef);
+
+	/* adjust chroma resolution */
+	if (src_format != BUF_FMT_ARGB8888_YUV444)
+		src_xres >>= 1;
+	if (src_format == BUF_FMT_YUV420)
+		src_yres >>= 1;
+	if (dst_format != BUF_FMT_ARGB8888_YUV444)
+		dst_xres >>= 1;
+	if (dst_format == BUF_FMT_YUV420) /* should not happen */
+		dst_yres >>= 1;
+
+	/* horizontal chroma */
+	dcss_scaler_filter_design(src_xres, dst_xres, false,
+				  (src_xres == dst_xres) && (ch->c_hstart == 0),
+				  coef);
+
+	dcss_scaler_program_7_coef_set(ch, DCSS_SCALER_COEF_HCHR, coef);
+
+	/* vertical chroma */
+	dcss_scaler_filter_design(src_yres, dst_yres, program_5_taps,
+				  (src_yres == dst_yres) && (ch->c_vstart == 0),
+				  coef);
+	if (program_5_taps)
+		dcss_scaler_program_5_coef_set(ch, DCSS_SCALER_COEF_VCHR, coef);
+	else
+		dcss_scaler_program_7_coef_set(ch, DCSS_SCALER_COEF_VCHR, coef);
+}
+
+static void dcss_scaler_rgb_coef_set(struct dcss_scaler_ch *ch,
+				     int src_xres, int src_yres, int dst_xres,
+				     int dst_yres)
+{
+	int coef[PSC_STORED_PHASES][PSC_NUM_TAPS];
+
+	/* horizontal RGB */
+	dcss_scaler_filter_design(src_xres, dst_xres, false,
+				  src_xres == dst_xres, coef);
+	dcss_scaler_program_7_coef_set(ch, DCSS_SCALER_COEF_HLUM, coef);
+
+	/* vertical RGB */
+	dcss_scaler_filter_design(src_yres, dst_yres, false,
+				  src_yres == dst_yres, coef);
+	dcss_scaler_program_7_coef_set(ch, DCSS_SCALER_COEF_VLUM, coef);
+}
+
+static void dcss_scaler_set_rgb10_order(struct dcss_scaler_ch *ch,
+					const struct drm_format_info *format)
+{
+	u32 a2r10g10b10_format;
+
+	if (format->is_yuv)
+		return;
+
+	ch->sdata_ctrl &= ~A2R10G10B10_FORMAT_MASK;
+
+	if (format->depth != 30)
+		return;
+
+	switch (format->format) {
+	case DRM_FORMAT_ARGB2101010:
+	case DRM_FORMAT_XRGB2101010:
+		a2r10g10b10_format = 0;
+		break;
+
+	case DRM_FORMAT_ABGR2101010:
+	case DRM_FORMAT_XBGR2101010:
+		a2r10g10b10_format = 5;
+		break;
+
+	case DRM_FORMAT_RGBA1010102:
+	case DRM_FORMAT_RGBX1010102:
+		a2r10g10b10_format = 6;
+		break;
+
+	case DRM_FORMAT_BGRA1010102:
+	case DRM_FORMAT_BGRX1010102:
+		a2r10g10b10_format = 11;
+		break;
+
+	default:
+		a2r10g10b10_format = 0;
+		break;
+	}
+
+	ch->sdata_ctrl |= a2r10g10b10_format << A2R10G10B10_FORMAT_POS;
+}
+
+void dcss_scaler_setup(struct dcss_scaler *scl, int ch_num,
+		       const struct drm_format_info *format,
+		       int src_xres, int src_yres, int dst_xres, int dst_yres,
+		       u32 vrefresh_hz)
+{
+	struct dcss_scaler_ch *ch = &scl->ch[ch_num];
+	unsigned int pixel_depth = 0;
+	bool rtr_8line_en = false;
+	bool use_5_taps = false;
+	enum buffer_format src_format = BUF_FMT_ARGB8888_YUV444;
+	enum buffer_format dst_format = BUF_FMT_ARGB8888_YUV444;
+	u32 pix_format = format->format;
+
+	if (format->is_yuv) {
+		dcss_scaler_yuv_enable(ch, true);
+
+		if (pix_format == DRM_FORMAT_NV12 ||
+		    pix_format == DRM_FORMAT_NV21) {
+			rtr_8line_en = true;
+			src_format = BUF_FMT_YUV420;
+		} else if (pix_format == DRM_FORMAT_UYVY ||
+			   pix_format == DRM_FORMAT_VYUY ||
+			   pix_format == DRM_FORMAT_YUYV ||
+			   pix_format == DRM_FORMAT_YVYU) {
+			src_format = BUF_FMT_YUV422;
+		}
+
+		use_5_taps = !rtr_8line_en;
+	} else {
+		dcss_scaler_yuv_enable(ch, false);
+
+		pixel_depth = format->depth;
+	}
+
+	dcss_scaler_fractions_set(ch, src_xres, src_yres, dst_xres,
+				  dst_yres, src_format, dst_format,
+				  PSC_LOC_HORZ_0_VERT_1_OVER_4);
+
+	if (format->is_yuv)
+		dcss_scaler_yuv_coef_set(ch, src_format, dst_format,
+					 use_5_taps, src_xres, src_yres,
+					 dst_xres, dst_yres);
+	else
+		dcss_scaler_rgb_coef_set(ch, src_xres, src_yres,
+					 dst_xres, dst_yres);
+
+	dcss_scaler_rtr_8lines_enable(ch, rtr_8line_en);
+	dcss_scaler_bit_depth_set(ch, pixel_depth);
+	dcss_scaler_set_rgb10_order(ch, format);
+	dcss_scaler_format_set(ch, src_format, dst_format);
+	dcss_scaler_res_set(ch, src_xres, src_yres, dst_xres, dst_yres,
+			    pix_format, dst_format);
+}
+
+/* This function will be called from interrupt context. */
+void dcss_scaler_write_sclctrl(struct dcss_scaler *scl)
+{
+	int chnum;
+
+	dcss_ctxld_assert_locked(scl->ctxld);
+
+	for (chnum = 0; chnum < 3; chnum++) {
+		struct dcss_scaler_ch *ch = &scl->ch[chnum];
+
+		if (ch->scaler_ctrl_chgd) {
+			dcss_ctxld_write_irqsafe(scl->ctxld, scl->ctx_id,
+						 ch->scaler_ctrl,
+						 ch->base_ofs +
+						 DCSS_SCALER_CTRL);
+			ch->scaler_ctrl_chgd = false;
+		}
+	}
+}
diff --git a/drivers/gpu/drm/imx/dcss/dcss-ss.c b/drivers/gpu/drm/imx/dcss/dcss-ss.c
new file mode 100644
index 000000000000..8ddf08da911b
--- /dev/null
+++ b/drivers/gpu/drm/imx/dcss/dcss-ss.c
@@ -0,0 +1,180 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Copyright 2019 NXP.
+ */
+
+#include <linux/device.h>
+#include <linux/slab.h>
+
+#include "dcss-dev.h"
+
+#define DCSS_SS_SYS_CTRL			0x00
+#define   RUN_EN				BIT(0)
+#define DCSS_SS_DISPLAY				0x10
+#define   LRC_X_POS				0
+#define   LRC_X_MASK				GENMASK(12, 0)
+#define   LRC_Y_POS				16
+#define   LRC_Y_MASK				GENMASK(28, 16)
+#define DCSS_SS_HSYNC				0x20
+#define DCSS_SS_VSYNC				0x30
+#define   SYNC_START_POS			0
+#define   SYNC_START_MASK			GENMASK(12, 0)
+#define   SYNC_END_POS				16
+#define   SYNC_END_MASK				GENMASK(28, 16)
+#define   SYNC_POL				BIT(31)
+#define DCSS_SS_DE_ULC				0x40
+#define   ULC_X_POS				0
+#define   ULC_X_MASK				GENMASK(12, 0)
+#define   ULC_Y_POS				16
+#define   ULC_Y_MASK				GENMASK(28, 16)
+#define   ULC_POL				BIT(31)
+#define DCSS_SS_DE_LRC				0x50
+#define DCSS_SS_MODE				0x60
+#define   PIPE_MODE_POS				0
+#define   PIPE_MODE_MASK			GENMASK(1, 0)
+#define DCSS_SS_COEFF				0x70
+#define   HORIZ_A_POS				0
+#define   HORIZ_A_MASK				GENMASK(3, 0)
+#define   HORIZ_B_POS				4
+#define   HORIZ_B_MASK				GENMASK(7, 4)
+#define   HORIZ_C_POS				8
+#define   HORIZ_C_MASK				GENMASK(11, 8)
+#define   HORIZ_H_NORM_POS			12
+#define   HORIZ_H_NORM_MASK			GENMASK(14, 12)
+#define   VERT_A_POS				16
+#define   VERT_A_MASK				GENMASK(19, 16)
+#define   VERT_B_POS				20
+#define   VERT_B_MASK				GENMASK(23, 20)
+#define   VERT_C_POS				24
+#define   VERT_C_MASK				GENMASK(27, 24)
+#define   VERT_H_NORM_POS			28
+#define   VERT_H_NORM_MASK			GENMASK(30, 28)
+#define DCSS_SS_CLIP_CB				0x80
+#define DCSS_SS_CLIP_CR				0x90
+#define   CLIP_MIN_POS				0
+#define   CLIP_MIN_MASK				GENMASK(9, 0)
+#define   CLIP_MAX_POS				0
+#define   CLIP_MAX_MASK				GENMASK(23, 16)
+#define DCSS_SS_INTER_MODE			0xA0
+#define   INT_EN				BIT(0)
+#define   VSYNC_SHIFT				BIT(1)
+
+struct dcss_ss {
+	struct device *dev;
+	void __iomem *base_reg;
+	u32 base_ofs;
+
+	struct dcss_ctxld *ctxld;
+	u32 ctx_id;
+
+	bool in_use;
+};
+
+static void dcss_ss_write(struct dcss_ss *ss, u32 val, u32 ofs)
+{
+	if (!ss->in_use)
+		dcss_writel(val, ss->base_reg + ofs);
+
+	dcss_ctxld_write(ss->ctxld, ss->ctx_id, val,
+			 ss->base_ofs + ofs);
+}
+
+int dcss_ss_init(struct dcss_dev *dcss, unsigned long ss_base)
+{
+	struct dcss_ss *ss;
+
+	ss = kzalloc(sizeof(*ss), GFP_KERNEL);
+	if (!ss)
+		return -ENOMEM;
+
+	dcss->ss = ss;
+	ss->dev = dcss->dev;
+	ss->ctxld = dcss->ctxld;
+
+	ss->base_reg = ioremap(ss_base, SZ_4K);
+	if (!ss->base_reg) {
+		dev_err(dcss->dev, "ss: unable to remap ss base\n");
+		kfree(ss);
+		return -ENOMEM;
+	}
+
+	ss->base_ofs = ss_base;
+	ss->ctx_id = CTX_SB_HP;
+
+	return 0;
+}
+
+void dcss_ss_exit(struct dcss_ss *ss)
+{
+	/* stop SS */
+	dcss_writel(0, ss->base_reg + DCSS_SS_SYS_CTRL);
+
+	if (ss->base_reg)
+		iounmap(ss->base_reg);
+
+	kfree(ss);
+}
+
+void dcss_ss_subsam_set(struct dcss_ss *ss)
+{
+	dcss_ss_write(ss, 0x41614161, DCSS_SS_COEFF);
+	dcss_ss_write(ss, 0, DCSS_SS_MODE);
+	dcss_ss_write(ss, 0x03ff0000, DCSS_SS_CLIP_CB);
+	dcss_ss_write(ss, 0x03ff0000, DCSS_SS_CLIP_CR);
+}
+
+void dcss_ss_sync_set(struct dcss_ss *ss, struct videomode *vm,
+		      bool phsync, bool pvsync)
+{
+	u16 lrc_x, lrc_y;
+	u16 hsync_start, hsync_end;
+	u16 vsync_start, vsync_end;
+	u16 de_ulc_x, de_ulc_y;
+	u16 de_lrc_x, de_lrc_y;
+
+	lrc_x = vm->hfront_porch + vm->hback_porch + vm->hsync_len +
+		vm->hactive - 1;
+	lrc_y = vm->vfront_porch + vm->vback_porch + vm->vsync_len +
+		vm->vactive - 1;
+
+	dcss_ss_write(ss, (lrc_y << LRC_Y_POS) | lrc_x, DCSS_SS_DISPLAY);
+
+	hsync_start = vm->hfront_porch + vm->hback_porch + vm->hsync_len +
+		      vm->hactive - 1;
+	hsync_end = vm->hsync_len - 1;
+
+	dcss_ss_write(ss, (phsync ? SYNC_POL : 0) |
+		      ((u32)hsync_end << SYNC_END_POS) | hsync_start,
+		      DCSS_SS_HSYNC);
+
+	vsync_start = vm->vfront_porch - 1;
+	vsync_end = vm->vfront_porch + vm->vsync_len - 1;
+
+	dcss_ss_write(ss, (pvsync ? SYNC_POL : 0) |
+		      ((u32)vsync_end << SYNC_END_POS) | vsync_start,
+		      DCSS_SS_VSYNC);
+
+	de_ulc_x = vm->hsync_len + vm->hback_porch - 1;
+	de_ulc_y = vm->vsync_len + vm->vfront_porch + vm->vback_porch;
+
+	dcss_ss_write(ss, SYNC_POL | ((u32)de_ulc_y << ULC_Y_POS) | de_ulc_x,
+		      DCSS_SS_DE_ULC);
+
+	de_lrc_x = vm->hsync_len + vm->hback_porch + vm->hactive - 1;
+	de_lrc_y = vm->vsync_len + vm->vfront_porch + vm->vback_porch +
+		   vm->vactive - 1;
+
+	dcss_ss_write(ss, (de_lrc_y << LRC_Y_POS) | de_lrc_x, DCSS_SS_DE_LRC);
+}
+
+void dcss_ss_enable(struct dcss_ss *ss)
+{
+	dcss_ss_write(ss, RUN_EN, DCSS_SS_SYS_CTRL);
+	ss->in_use = true;
+}
+
+void dcss_ss_shutoff(struct dcss_ss *ss)
+{
+	dcss_writel(0, ss->base_reg + DCSS_SS_SYS_CTRL);
+	ss->in_use = false;
+}
diff --git a/drivers/gpu/drm/ingenic/ingenic-drm-drv.c b/drivers/gpu/drm/ingenic/ingenic-drm-drv.c
index a3d1617d7c67..937d080f5d06 100644
--- a/drivers/gpu/drm/ingenic/ingenic-drm-drv.c
+++ b/drivers/gpu/drm/ingenic/ingenic-drm-drv.c
@@ -9,6 +9,8 @@
 #include <linux/component.h>
 #include <linux/clk.h>
 #include <linux/dma-mapping.h>
+#include <linux/dma-noncoherent.h>
+#include <linux/io.h>
 #include <linux/module.h>
 #include <linux/of_device.h>
 #include <linux/platform_device.h>
@@ -19,6 +21,7 @@
 #include <drm/drm_bridge.h>
 #include <drm/drm_crtc.h>
 #include <drm/drm_crtc_helper.h>
+#include <drm/drm_damage_helper.h>
 #include <drm/drm_drv.h>
 #include <drm/drm_gem_cma_helper.h>
 #include <drm/drm_fb_cma_helper.h>
@@ -76,6 +79,11 @@ static const u32 ingenic_drm_primary_formats[] = {
 	DRM_FORMAT_XRGB8888,
 };
 
+static bool ingenic_drm_cached_gem_buf;
+module_param_named(cached_gem_buffers, ingenic_drm_cached_gem_buf, bool, 0400);
+MODULE_PARM_DESC(cached_gem_buffers,
+		 "Enable fully cached GEM buffers [default=false]");
+
 static bool ingenic_drm_writeable_reg(struct device *dev, unsigned int reg)
 {
 	switch (reg) {
@@ -338,6 +346,8 @@ static int ingenic_drm_plane_atomic_check(struct drm_plane *plane,
 	     plane->state->fb->format->format != state->fb->format->format))
 		crtc_state->mode_changed = true;
 
+	drm_atomic_helper_check_plane_damage(state->state, state);
+
 	return 0;
 }
 
@@ -440,6 +450,38 @@ void ingenic_drm_plane_config(struct device *dev,
 	}
 }
 
+void ingenic_drm_sync_data(struct device *dev,
+			   struct drm_plane_state *old_state,
+			   struct drm_plane_state *state)
+{
+	const struct drm_format_info *finfo = state->fb->format;
+	struct ingenic_drm *priv = dev_get_drvdata(dev);
+	struct drm_atomic_helper_damage_iter iter;
+	unsigned int offset, i;
+	struct drm_rect clip;
+	dma_addr_t paddr;
+	void *addr;
+
+	if (!ingenic_drm_cached_gem_buf)
+		return;
+
+	drm_atomic_helper_damage_iter_init(&iter, old_state, state);
+
+	drm_atomic_for_each_plane_damage(&iter, &clip) {
+		for (i = 0; i < finfo->num_planes; i++) {
+			paddr = drm_fb_cma_get_gem_addr(state->fb, state, i);
+			addr = phys_to_virt(paddr);
+
+			/* Ignore x1/x2 values, invalidate complete lines */
+			offset = clip.y1 * state->fb->pitches[i];
+
+			dma_cache_sync(priv->dev, addr + offset,
+				       (clip.y2 - clip.y1) * state->fb->pitches[i],
+				       DMA_TO_DEVICE);
+		}
+	}
+}
+
 static void ingenic_drm_plane_atomic_update(struct drm_plane *plane,
 					    struct drm_plane_state *oldstate)
 {
@@ -450,6 +492,8 @@ static void ingenic_drm_plane_atomic_update(struct drm_plane *plane,
 	dma_addr_t addr;
 
 	if (state && state->fb) {
+		ingenic_drm_sync_data(priv->dev, oldstate, state);
+
 		addr = drm_fb_cma_get_gem_addr(state->fb, state, 0);
 		width = state->src_w >> 16;
 		height = state->src_h >> 16;
@@ -605,7 +649,69 @@ static void ingenic_drm_disable_vblank(struct drm_crtc *crtc)
 	regmap_update_bits(priv->map, JZ_REG_LCD_CTRL, JZ_LCD_CTRL_EOF_IRQ, 0);
 }
 
-DEFINE_DRM_GEM_CMA_FOPS(ingenic_drm_fops);
+static struct drm_framebuffer *
+ingenic_drm_gem_fb_create(struct drm_device *dev, struct drm_file *file,
+			  const struct drm_mode_fb_cmd2 *mode_cmd)
+{
+	if (ingenic_drm_cached_gem_buf)
+		return drm_gem_fb_create_with_dirty(dev, file, mode_cmd);
+
+	return drm_gem_fb_create(dev, file, mode_cmd);
+}
+
+static int ingenic_drm_gem_mmap(struct drm_gem_object *obj,
+				struct vm_area_struct *vma)
+{
+	struct drm_gem_cma_object *cma_obj = to_drm_gem_cma_obj(obj);
+	struct device *dev = cma_obj->base.dev->dev;
+	unsigned long attrs;
+	int ret;
+
+	if (ingenic_drm_cached_gem_buf)
+		attrs = DMA_ATTR_NON_CONSISTENT;
+	else
+		attrs = DMA_ATTR_WRITE_COMBINE;
+
+	/*
+	 * Clear the VM_PFNMAP flag that was set by drm_gem_mmap(), and set the
+	 * vm_pgoff (used as a fake buffer offset by DRM) to 0 as we want to map
+	 * the whole buffer.
+	 */
+	vma->vm_flags &= ~VM_PFNMAP;
+	vma->vm_pgoff = 0;
+	vma->vm_page_prot = vm_get_page_prot(vma->vm_flags);
+
+	ret = dma_mmap_attrs(dev, vma, cma_obj->vaddr, cma_obj->paddr,
+			     vma->vm_end - vma->vm_start, attrs);
+	if (ret)
+		drm_gem_vm_close(vma);
+
+	return ret;
+}
+
+static int ingenic_drm_gem_cma_mmap(struct file *filp,
+				    struct vm_area_struct *vma)
+{
+	int ret;
+
+	ret = drm_gem_mmap(filp, vma);
+	if (ret)
+		return ret;
+
+	return ingenic_drm_gem_mmap(vma->vm_private_data, vma);
+}
+
+static const struct file_operations ingenic_drm_fops = {
+	.owner		= THIS_MODULE,
+	.open		= drm_open,
+	.release	= drm_release,
+	.unlocked_ioctl	= drm_ioctl,
+	.compat_ioctl	= drm_compat_ioctl,
+	.poll		= drm_poll,
+	.read		= drm_read,
+	.llseek		= noop_llseek,
+	.mmap		= ingenic_drm_gem_cma_mmap,
+};
 
 static struct drm_driver ingenic_drm_driver_data = {
 	.driver_features	= DRIVER_MODESET | DRIVER_GEM | DRIVER_ATOMIC,
@@ -669,7 +775,7 @@ static const struct drm_encoder_helper_funcs ingenic_drm_encoder_helper_funcs =
 };
 
 static const struct drm_mode_config_funcs ingenic_drm_mode_config_funcs = {
-	.fb_create		= drm_gem_fb_create,
+	.fb_create		= ingenic_drm_gem_fb_create,
 	.output_poll_changed	= drm_fb_helper_output_poll_changed,
 	.atomic_check		= drm_atomic_helper_check,
 	.atomic_commit		= drm_atomic_helper_commit,
@@ -796,6 +902,8 @@ static int ingenic_drm_bind(struct device *dev, bool has_components)
 		return ret;
 	}
 
+	drm_plane_enable_fb_damage_clips(&priv->f1);
+
 	drm_crtc_helper_add(&priv->crtc, &ingenic_drm_crtc_helper_funcs);
 
 	ret = drm_crtc_init_with_planes(drm, &priv->crtc, &priv->f1,
@@ -821,6 +929,8 @@ static int ingenic_drm_bind(struct device *dev, bool has_components)
 			return ret;
 		}
 
+		drm_plane_enable_fb_damage_clips(&priv->f0);
+
 		if (IS_ENABLED(CONFIG_DRM_INGENIC_IPU) && has_components) {
 			ret = component_bind_all(dev, drm);
 			if (ret) {
diff --git a/drivers/gpu/drm/ingenic/ingenic-drm.h b/drivers/gpu/drm/ingenic/ingenic-drm.h
index 43f7d959cff7..df99f0f75d39 100644
--- a/drivers/gpu/drm/ingenic/ingenic-drm.h
+++ b/drivers/gpu/drm/ingenic/ingenic-drm.h
@@ -168,6 +168,10 @@ void ingenic_drm_plane_config(struct device *dev,
 			      struct drm_plane *plane, u32 fourcc);
 void ingenic_drm_plane_disable(struct device *dev, struct drm_plane *plane);
 
+void ingenic_drm_sync_data(struct device *dev,
+			   struct drm_plane_state *old_state,
+			   struct drm_plane_state *state);
+
 extern struct platform_driver *ingenic_ipu_driver_ptr;
 
 #endif /* DRIVERS_GPU_DRM_INGENIC_INGENIC_DRM_H */
diff --git a/drivers/gpu/drm/ingenic/ingenic-ipu.c b/drivers/gpu/drm/ingenic/ingenic-ipu.c
index fc8c6e970ee3..38c83e8cc6a5 100644
--- a/drivers/gpu/drm/ingenic/ingenic-ipu.c
+++ b/drivers/gpu/drm/ingenic/ingenic-ipu.c
@@ -20,6 +20,7 @@
 
 #include <drm/drm_atomic.h>
 #include <drm/drm_atomic_helper.h>
+#include <drm/drm_damage_helper.h>
 #include <drm/drm_drv.h>
 #include <drm/drm_fb_cma_helper.h>
 #include <drm/drm_fourcc.h>
@@ -316,6 +317,8 @@ static void ingenic_ipu_plane_atomic_update(struct drm_plane *plane,
 				JZ_IPU_CTRL_CHIP_EN | JZ_IPU_CTRL_LCDC_SEL);
 	}
 
+	ingenic_drm_sync_data(ipu->master, oldstate, state);
+
 	/* New addresses will be committed in vblank handler... */
 	ipu->addr_y = drm_fb_cma_get_gem_addr(state->fb, state, 0);
 	if (finfo->num_planes > 1)
@@ -534,7 +537,7 @@ static int ingenic_ipu_plane_atomic_check(struct drm_plane *plane,
 
 	if (!state->crtc ||
 	    !crtc_state->mode.hdisplay || !crtc_state->mode.vdisplay)
-		return 0;
+		goto out_check_damage;
 
 	/* Plane must be fully visible */
 	if (state->crtc_x < 0 || state->crtc_y < 0 ||
@@ -551,7 +554,7 @@ static int ingenic_ipu_plane_atomic_check(struct drm_plane *plane,
 		return -EINVAL;
 
 	if (!osd_changed(state, plane->state))
-		return 0;
+		goto out_check_damage;
 
 	crtc_state->mode_changed = true;
 
@@ -578,6 +581,9 @@ static int ingenic_ipu_plane_atomic_check(struct drm_plane *plane,
 	ipu->denom_w = denom_w;
 	ipu->denom_h = denom_h;
 
+out_check_damage:
+	drm_atomic_helper_check_plane_damage(state->state, state);
+
 	return 0;
 }
 
@@ -759,6 +765,8 @@ static int ingenic_ipu_bind(struct device *dev, struct device *master, void *d)
 		return err;
 	}
 
+	drm_plane_enable_fb_damage_clips(plane);
+
 	/*
 	 * Sharpness settings range is [0,32]
 	 * 0       : nearest-neighbor
diff --git a/drivers/gpu/drm/msm/msm_gem.c b/drivers/gpu/drm/msm/msm_gem.c
index 8c7ae812b813..e47958c3704a 100644
--- a/drivers/gpu/drm/msm/msm_gem.c
+++ b/drivers/gpu/drm/msm/msm_gem.c
@@ -123,7 +123,7 @@ static struct page **get_pages(struct drm_gem_object *obj)
 
 		msm_obj->pages = p;
 
-		msm_obj->sgt = drm_prime_pages_to_sg(p, npages);
+		msm_obj->sgt = drm_prime_pages_to_sg(obj->dev, p, npages);
 		if (IS_ERR(msm_obj->sgt)) {
 			void *ptr = ERR_CAST(msm_obj->sgt);
 
diff --git a/drivers/gpu/drm/msm/msm_gem_prime.c b/drivers/gpu/drm/msm/msm_gem_prime.c
index d7c8948427fe..515ef80816a0 100644
--- a/drivers/gpu/drm/msm/msm_gem_prime.c
+++ b/drivers/gpu/drm/msm/msm_gem_prime.c
@@ -19,7 +19,7 @@ struct sg_table *msm_gem_prime_get_sg_table(struct drm_gem_object *obj)
 	if (WARN_ON(!msm_obj->pages))  /* should have already pinned! */
 		return NULL;
 
-	return drm_prime_pages_to_sg(msm_obj->pages, npages);
+	return drm_prime_pages_to_sg(obj->dev, msm_obj->pages, npages);
 }
 
 void *msm_gem_prime_vmap(struct drm_gem_object *obj)
diff --git a/drivers/gpu/drm/mxsfb/mxsfb_drv.c b/drivers/gpu/drm/mxsfb/mxsfb_drv.c
index 8c549c3931af..35122aef037b 100644
--- a/drivers/gpu/drm/mxsfb/mxsfb_drv.c
+++ b/drivers/gpu/drm/mxsfb/mxsfb_drv.c
@@ -21,6 +21,7 @@
 #include <drm/drm_connector.h>
 #include <drm/drm_drv.h>
 #include <drm/drm_fb_helper.h>
+#include <drm/drm_fourcc.h>
 #include <drm/drm_gem_cma_helper.h>
 #include <drm/drm_gem_framebuffer_helper.h>
 #include <drm/drm_irq.h>
@@ -81,8 +82,26 @@ void mxsfb_disable_axi_clk(struct mxsfb_drm_private *mxsfb)
 		clk_disable_unprepare(mxsfb->clk_axi);
 }
 
+static struct drm_framebuffer *
+mxsfb_fb_create(struct drm_device *dev, struct drm_file *file_priv,
+		const struct drm_mode_fb_cmd2 *mode_cmd)
+{
+	const struct drm_format_info *info;
+
+	info = drm_get_format_info(dev, mode_cmd);
+	if (!info)
+		return ERR_PTR(-EINVAL);
+
+	if (mode_cmd->width * info->cpp[0] != mode_cmd->pitches[0]) {
+		dev_dbg(dev->dev, "Invalid pitch: fb width must match pitch\n");
+		return ERR_PTR(-EINVAL);
+	}
+
+	return drm_gem_fb_create(dev, file_priv, mode_cmd);
+}
+
 static const struct drm_mode_config_funcs mxsfb_mode_config_funcs = {
-	.fb_create		= drm_gem_fb_create,
+	.fb_create		= mxsfb_fb_create,
 	.atomic_check		= drm_atomic_helper_check,
 	.atomic_commit		= drm_atomic_helper_commit,
 };
diff --git a/drivers/gpu/drm/nouveau/dispnv04/crtc.c b/drivers/gpu/drm/nouveau/dispnv04/crtc.c
index 6416b6907aeb..f9e962fd94d0 100644
--- a/drivers/gpu/drm/nouveau/dispnv04/crtc.c
+++ b/drivers/gpu/drm/nouveau/dispnv04/crtc.c
@@ -615,7 +615,7 @@ nv_crtc_swap_fbs(struct drm_crtc *crtc, struct drm_framebuffer *old_fb)
 	struct nouveau_crtc *nv_crtc = nouveau_crtc(crtc);
 	int ret;
 
-	ret = nouveau_bo_pin(nvbo, TTM_PL_FLAG_VRAM, false);
+	ret = nouveau_bo_pin(nvbo, NOUVEAU_GEM_DOMAIN_VRAM, false);
 	if (ret == 0) {
 		if (disp->image[nv_crtc->index])
 			nouveau_bo_unpin(disp->image[nv_crtc->index]);
@@ -1172,7 +1172,7 @@ nv04_crtc_page_flip(struct drm_crtc *crtc, struct drm_framebuffer *fb,
 		return -ENOMEM;
 
 	if (new_bo != old_bo) {
-		ret = nouveau_bo_pin(new_bo, TTM_PL_FLAG_VRAM, true);
+		ret = nouveau_bo_pin(new_bo, NOUVEAU_GEM_DOMAIN_VRAM, true);
 		if (ret)
 			goto fail_free;
 	}
@@ -1336,10 +1336,11 @@ nv04_crtc_create(struct drm_device *dev, int crtc_num)
 	drm_mode_crtc_set_gamma_size(&nv_crtc->base, 256);
 
 	ret = nouveau_bo_new(&nouveau_drm(dev)->client, 64*64*4, 0x100,
-			     TTM_PL_FLAG_VRAM, 0, 0x0000, NULL, NULL,
+			     NOUVEAU_GEM_DOMAIN_VRAM, 0, 0x0000, NULL, NULL,
 			     &nv_crtc->cursor.nvbo);
 	if (!ret) {
-		ret = nouveau_bo_pin(nv_crtc->cursor.nvbo, TTM_PL_FLAG_VRAM, false);
+		ret = nouveau_bo_pin(nv_crtc->cursor.nvbo,
+				     NOUVEAU_GEM_DOMAIN_VRAM, false);
 		if (!ret) {
 			ret = nouveau_bo_map(nv_crtc->cursor.nvbo);
 			if (ret)
diff --git a/drivers/gpu/drm/nouveau/dispnv04/disp.c b/drivers/gpu/drm/nouveau/dispnv04/disp.c
index 3ee836dc5058..7739f46470d3 100644
--- a/drivers/gpu/drm/nouveau/dispnv04/disp.c
+++ b/drivers/gpu/drm/nouveau/dispnv04/disp.c
@@ -134,7 +134,7 @@ nv04_display_init(struct drm_device *dev, bool resume, bool runtime)
 		if (!fb || !fb->obj[0])
 			continue;
 		nvbo = nouveau_gem_object(fb->obj[0]);
-		ret = nouveau_bo_pin(nvbo, TTM_PL_FLAG_VRAM, true);
+		ret = nouveau_bo_pin(nvbo, NOUVEAU_GEM_DOMAIN_VRAM, true);
 		if (ret)
 			NV_ERROR(drm, "Could not pin framebuffer\n");
 	}
@@ -144,7 +144,8 @@ nv04_display_init(struct drm_device *dev, bool resume, bool runtime)
 		if (!nv_crtc->cursor.nvbo)
 			continue;
 
-		ret = nouveau_bo_pin(nv_crtc->cursor.nvbo, TTM_PL_FLAG_VRAM, true);
+		ret = nouveau_bo_pin(nv_crtc->cursor.nvbo,
+				     NOUVEAU_GEM_DOMAIN_VRAM, true);
 		if (!ret && nv_crtc->cursor.set_offset)
 			ret = nouveau_bo_map(nv_crtc->cursor.nvbo);
 		if (ret)
diff --git a/drivers/gpu/drm/nouveau/dispnv04/overlay.c b/drivers/gpu/drm/nouveau/dispnv04/overlay.c
index 193ba9498f3d..37e63e98cd08 100644
--- a/drivers/gpu/drm/nouveau/dispnv04/overlay.c
+++ b/drivers/gpu/drm/nouveau/dispnv04/overlay.c
@@ -142,7 +142,7 @@ nv10_update_plane(struct drm_plane *plane, struct drm_crtc *crtc,
 		return ret;
 
 	nvbo = nouveau_gem_object(fb->obj[0]);
-	ret = nouveau_bo_pin(nvbo, TTM_PL_FLAG_VRAM, false);
+	ret = nouveau_bo_pin(nvbo, NOUVEAU_GEM_DOMAIN_VRAM, false);
 	if (ret)
 		return ret;
 
@@ -387,7 +387,7 @@ nv04_update_plane(struct drm_plane *plane, struct drm_crtc *crtc,
 		return ret;
 
 	nvbo = nouveau_gem_object(fb->obj[0]);
-	ret = nouveau_bo_pin(nvbo, TTM_PL_FLAG_VRAM, false);
+	ret = nouveau_bo_pin(nvbo, NOUVEAU_GEM_DOMAIN_VRAM, false);
 	if (ret)
 		return ret;
 
diff --git a/drivers/gpu/drm/nouveau/dispnv50/disp.c b/drivers/gpu/drm/nouveau/dispnv50/disp.c
index 7799530e07c1..b111fe24a06b 100644
--- a/drivers/gpu/drm/nouveau/dispnv50/disp.c
+++ b/drivers/gpu/drm/nouveau/dispnv50/disp.c
@@ -2069,6 +2069,7 @@ nv50_disp_atomic_commit_tail(struct drm_atomic_state *state)
 	drm_atomic_helper_wait_for_fences(dev, state, false);
 	drm_atomic_helper_wait_for_dependencies(state);
 	drm_atomic_helper_update_legacy_modeset_state(dev, state);
+	drm_atomic_helper_calc_timestamping_constants(state);
 
 	if (atom->lock_core)
 		mutex_lock(&disp->mutex);
@@ -2622,10 +2623,11 @@ nv50_display_create(struct drm_device *dev)
 	dev->mode_config.normalize_zpos = true;
 
 	/* small shared memory area we use for notifiers and semaphores */
-	ret = nouveau_bo_new(&drm->client, 4096, 0x1000, TTM_PL_FLAG_VRAM,
+	ret = nouveau_bo_new(&drm->client, 4096, 0x1000,
+			     NOUVEAU_GEM_DOMAIN_VRAM,
 			     0, 0x0000, NULL, NULL, &disp->sync);
 	if (!ret) {
-		ret = nouveau_bo_pin(disp->sync, TTM_PL_FLAG_VRAM, true);
+		ret = nouveau_bo_pin(disp->sync, NOUVEAU_GEM_DOMAIN_VRAM, true);
 		if (!ret) {
 			ret = nouveau_bo_map(disp->sync);
 			if (ret)
diff --git a/drivers/gpu/drm/nouveau/dispnv50/wndw.c b/drivers/gpu/drm/nouveau/dispnv50/wndw.c
index 447ecc9fec42..0356474ad6f6 100644
--- a/drivers/gpu/drm/nouveau/dispnv50/wndw.c
+++ b/drivers/gpu/drm/nouveau/dispnv50/wndw.c
@@ -542,7 +542,7 @@ nv50_wndw_prepare_fb(struct drm_plane *plane, struct drm_plane_state *state)
 		return 0;
 
 	nvbo = nouveau_gem_object(fb->obj[0]);
-	ret = nouveau_bo_pin(nvbo, TTM_PL_FLAG_VRAM, true);
+	ret = nouveau_bo_pin(nvbo, NOUVEAU_GEM_DOMAIN_VRAM, true);
 	if (ret)
 		return ret;
 
diff --git a/drivers/gpu/drm/nouveau/nouveau_abi16.c b/drivers/gpu/drm/nouveau/nouveau_abi16.c
index 21537ca1dd39..9a5be6f32424 100644
--- a/drivers/gpu/drm/nouveau/nouveau_abi16.c
+++ b/drivers/gpu/drm/nouveau/nouveau_abi16.c
@@ -328,7 +328,8 @@ nouveau_abi16_ioctl_channel_alloc(ABI16_IOCTL_ARGS)
 	ret = nouveau_gem_new(cli, PAGE_SIZE, 0, NOUVEAU_GEM_DOMAIN_GART,
 			      0, 0, &chan->ntfy);
 	if (ret == 0)
-		ret = nouveau_bo_pin(chan->ntfy, TTM_PL_FLAG_TT, false);
+		ret = nouveau_bo_pin(chan->ntfy, NOUVEAU_GEM_DOMAIN_GART,
+				     false);
 	if (ret)
 		goto done;
 
diff --git a/drivers/gpu/drm/nouveau/nouveau_bo.c b/drivers/gpu/drm/nouveau/nouveau_bo.c
index 9140387f30dc..2ee75646ad6f 100644
--- a/drivers/gpu/drm/nouveau/nouveau_bo.c
+++ b/drivers/gpu/drm/nouveau/nouveau_bo.c
@@ -44,6 +44,9 @@
 #include <nvif/if500b.h>
 #include <nvif/if900b.h>
 
+static int nouveau_ttm_tt_bind(struct ttm_bo_device *bdev, struct ttm_tt *ttm,
+			       struct ttm_resource *reg);
+
 /*
  * NV10-NV40 tiling helpers
  */
@@ -137,6 +140,7 @@ nouveau_bo_del_ttm(struct ttm_buffer_object *bo)
 	struct nouveau_bo *nvbo = nouveau_bo(bo);
 
 	WARN_ON(nvbo->pin_refcnt > 0);
+	nouveau_bo_del_io_reserve_lru(bo);
 	nv10_bo_put_tile_region(dev, nvbo->tile, NULL);
 
 	/*
@@ -158,8 +162,7 @@ roundup_64(u64 x, u32 y)
 }
 
 static void
-nouveau_bo_fixup_align(struct nouveau_bo *nvbo, u32 flags,
-		       int *align, u64 *size)
+nouveau_bo_fixup_align(struct nouveau_bo *nvbo, int *align, u64 *size)
 {
 	struct nouveau_drm *drm = nouveau_bdev(nvbo->bo.bdev);
 	struct nvif_device *device = &drm->client.device;
@@ -192,7 +195,7 @@ nouveau_bo_fixup_align(struct nouveau_bo *nvbo, u32 flags,
 }
 
 struct nouveau_bo *
-nouveau_bo_alloc(struct nouveau_cli *cli, u64 *size, int *align, u32 flags,
+nouveau_bo_alloc(struct nouveau_cli *cli, u64 *size, int *align, u32 domain,
 		 u32 tile_mode, u32 tile_flags)
 {
 	struct nouveau_drm *drm = cli->drm;
@@ -218,7 +221,7 @@ nouveau_bo_alloc(struct nouveau_cli *cli, u64 *size, int *align, u32 flags,
 	 * mapping, but is what NOUVEAU_GEM_DOMAIN_COHERENT gets translated
 	 * into in nouveau_gem_new().
 	 */
-	if (flags & TTM_PL_FLAG_UNCACHED) {
+	if (domain & NOUVEAU_GEM_DOMAIN_COHERENT) {
 		/* Determine if we can get a cache-coherent map, forcing
 		 * uncached mapping if we can't.
 		 */
@@ -258,9 +261,9 @@ nouveau_bo_alloc(struct nouveau_cli *cli, u64 *size, int *align, u32 flags,
 		 * Skip page sizes that can't support needed domains.
 		 */
 		if (cli->device.info.family > NV_DEVICE_INFO_V0_CURIE &&
-		    (flags & TTM_PL_FLAG_VRAM) && !vmm->page[i].vram)
+		    (domain & NOUVEAU_GEM_DOMAIN_VRAM) && !vmm->page[i].vram)
 			continue;
-		if ((flags & TTM_PL_FLAG_TT) &&
+		if ((domain & NOUVEAU_GEM_DOMAIN_GART) &&
 		    (!vmm->page[i].host || vmm->page[i].shift > PAGE_SHIFT))
 			continue;
 
@@ -287,13 +290,13 @@ nouveau_bo_alloc(struct nouveau_cli *cli, u64 *size, int *align, u32 flags,
 	}
 	nvbo->page = vmm->page[pi].shift;
 
-	nouveau_bo_fixup_align(nvbo, flags, align, size);
+	nouveau_bo_fixup_align(nvbo, align, size);
 
 	return nvbo;
 }
 
 int
-nouveau_bo_init(struct nouveau_bo *nvbo, u64 size, int align, u32 flags,
+nouveau_bo_init(struct nouveau_bo *nvbo, u64 size, int align, u32 domain,
 		struct sg_table *sg, struct dma_resv *robj)
 {
 	int type = sg ? ttm_bo_type_sg : ttm_bo_type_device;
@@ -303,7 +306,8 @@ nouveau_bo_init(struct nouveau_bo *nvbo, u64 size, int align, u32 flags,
 	acc_size = ttm_bo_dma_acc_size(nvbo->bo.bdev, size, sizeof(*nvbo));
 
 	nvbo->bo.mem.num_pages = size >> PAGE_SHIFT;
-	nouveau_bo_placement_set(nvbo, flags, 0);
+	nouveau_bo_placement_set(nvbo, domain, 0);
+	INIT_LIST_HEAD(&nvbo->io_reserve_lru);
 
 	ret = ttm_bo_init(nvbo->bo.bdev, &nvbo->bo, size, type,
 			  &nvbo->placement, align >> PAGE_SHIFT, false,
@@ -318,19 +322,19 @@ nouveau_bo_init(struct nouveau_bo *nvbo, u64 size, int align, u32 flags,
 
 int
 nouveau_bo_new(struct nouveau_cli *cli, u64 size, int align,
-	       uint32_t flags, uint32_t tile_mode, uint32_t tile_flags,
+	       uint32_t domain, uint32_t tile_mode, uint32_t tile_flags,
 	       struct sg_table *sg, struct dma_resv *robj,
 	       struct nouveau_bo **pnvbo)
 {
 	struct nouveau_bo *nvbo;
 	int ret;
 
-	nvbo = nouveau_bo_alloc(cli, &size, &align, flags, tile_mode,
+	nvbo = nouveau_bo_alloc(cli, &size, &align, domain, tile_mode,
 				tile_flags);
 	if (IS_ERR(nvbo))
 		return PTR_ERR(nvbo);
 
-	ret = nouveau_bo_init(nvbo, size, align, flags, sg, robj);
+	ret = nouveau_bo_init(nvbo, size, align, domain, sg, robj);
 	if (ret)
 		return ret;
 
@@ -339,27 +343,49 @@ nouveau_bo_new(struct nouveau_cli *cli, u64 size, int align,
 }
 
 static void
-set_placement_list(struct ttm_place *pl, unsigned *n, uint32_t type, uint32_t flags)
+set_placement_list(struct nouveau_drm *drm, struct ttm_place *pl, unsigned *n,
+		   uint32_t domain, uint32_t flags)
 {
 	*n = 0;
 
-	if (type & TTM_PL_FLAG_VRAM)
-		pl[(*n)++].flags = TTM_PL_FLAG_VRAM | flags;
-	if (type & TTM_PL_FLAG_TT)
-		pl[(*n)++].flags = TTM_PL_FLAG_TT | flags;
-	if (type & TTM_PL_FLAG_SYSTEM)
-		pl[(*n)++].flags = TTM_PL_FLAG_SYSTEM | flags;
+	if (domain & NOUVEAU_GEM_DOMAIN_VRAM) {
+		struct nvif_mmu *mmu = &drm->client.mmu;
+		const u8 type = mmu->type[drm->ttm.type_vram].type;
+
+		pl[*n].mem_type = TTM_PL_VRAM;
+		pl[*n].flags = flags & ~TTM_PL_FLAG_CACHED;
+
+		/* Some BARs do not support being ioremapped WC */
+		if (drm->client.device.info.family >= NV_DEVICE_INFO_V0_TESLA &&
+		    type & NVIF_MEM_UNCACHED)
+			pl[*n].flags &= ~TTM_PL_FLAG_WC;
+
+		(*n)++;
+	}
+	if (domain & NOUVEAU_GEM_DOMAIN_GART) {
+		pl[*n].mem_type = TTM_PL_TT;
+		pl[*n].flags = flags;
+
+		if (drm->agp.bridge)
+			pl[*n].flags &= ~TTM_PL_FLAG_CACHED;
+
+		(*n)++;
+	}
+	if (domain & NOUVEAU_GEM_DOMAIN_CPU) {
+		pl[*n].mem_type = TTM_PL_SYSTEM;
+		pl[(*n)++].flags = flags;
+	}
 }
 
 static void
-set_placement_range(struct nouveau_bo *nvbo, uint32_t type)
+set_placement_range(struct nouveau_bo *nvbo, uint32_t domain)
 {
 	struct nouveau_drm *drm = nouveau_bdev(nvbo->bo.bdev);
 	u32 vram_pages = drm->client.device.info.ram_size >> PAGE_SHIFT;
 	unsigned i, fpfn, lpfn;
 
 	if (drm->client.device.info.family == NV_DEVICE_INFO_V0_CELSIUS &&
-	    nvbo->mode && (type & TTM_PL_FLAG_VRAM) &&
+	    nvbo->mode && (domain & NOUVEAU_GEM_DOMAIN_VRAM) &&
 	    nvbo->bo.mem.num_pages < vram_pages / 4) {
 		/*
 		 * Make sure that the color and depth buffers are handled
@@ -386,26 +412,28 @@ set_placement_range(struct nouveau_bo *nvbo, uint32_t type)
 }
 
 void
-nouveau_bo_placement_set(struct nouveau_bo *nvbo, uint32_t type, uint32_t busy)
+nouveau_bo_placement_set(struct nouveau_bo *nvbo, uint32_t domain,
+			 uint32_t busy)
 {
+	struct nouveau_drm *drm = nouveau_bdev(nvbo->bo.bdev);
 	struct ttm_placement *pl = &nvbo->placement;
 	uint32_t flags = (nvbo->force_coherent ? TTM_PL_FLAG_UNCACHED :
 						 TTM_PL_MASK_CACHING) |
 			 (nvbo->pin_refcnt ? TTM_PL_FLAG_NO_EVICT : 0);
 
 	pl->placement = nvbo->placements;
-	set_placement_list(nvbo->placements, &pl->num_placement,
-			   type, flags);
+	set_placement_list(drm, nvbo->placements, &pl->num_placement,
+			   domain, flags);
 
 	pl->busy_placement = nvbo->busy_placements;
-	set_placement_list(nvbo->busy_placements, &pl->num_busy_placement,
-			   type | busy, flags);
+	set_placement_list(drm, nvbo->busy_placements, &pl->num_busy_placement,
+			   domain | busy, flags);
 
-	set_placement_range(nvbo, type);
+	set_placement_range(nvbo, domain);
 }
 
 int
-nouveau_bo_pin(struct nouveau_bo *nvbo, uint32_t memtype, bool contig)
+nouveau_bo_pin(struct nouveau_bo *nvbo, uint32_t domain, bool contig)
 {
 	struct nouveau_drm *drm = nouveau_bdev(nvbo->bo.bdev);
 	struct ttm_buffer_object *bo = &nvbo->bo;
@@ -417,7 +445,7 @@ nouveau_bo_pin(struct nouveau_bo *nvbo, uint32_t memtype, bool contig)
 		return ret;
 
 	if (drm->client.device.info.family >= NV_DEVICE_INFO_V0_TESLA &&
-	    memtype == TTM_PL_FLAG_VRAM && contig) {
+	    domain == NOUVEAU_GEM_DOMAIN_VRAM && contig) {
 		if (!nvbo->contig) {
 			nvbo->contig = true;
 			force = true;
@@ -426,10 +454,22 @@ nouveau_bo_pin(struct nouveau_bo *nvbo, uint32_t memtype, bool contig)
 	}
 
 	if (nvbo->pin_refcnt) {
-		if (!(memtype & (1 << bo->mem.mem_type)) || evict) {
+		bool error = evict;
+
+		switch (bo->mem.mem_type) {
+		case TTM_PL_VRAM:
+			error |= !(domain & NOUVEAU_GEM_DOMAIN_VRAM);
+			break;
+		case TTM_PL_TT:
+			error |= !(domain & NOUVEAU_GEM_DOMAIN_GART);
+		default:
+			break;
+		}
+
+		if (error) {
 			NV_ERROR(drm, "bo %p pinned elsewhere: "
 				      "0x%08x vs 0x%08x\n", bo,
-				 1 << bo->mem.mem_type, memtype);
+				 bo->mem.mem_type, domain);
 			ret = -EBUSY;
 		}
 		nvbo->pin_refcnt++;
@@ -437,14 +477,14 @@ nouveau_bo_pin(struct nouveau_bo *nvbo, uint32_t memtype, bool contig)
 	}
 
 	if (evict) {
-		nouveau_bo_placement_set(nvbo, TTM_PL_FLAG_TT, 0);
+		nouveau_bo_placement_set(nvbo, NOUVEAU_GEM_DOMAIN_GART, 0);
 		ret = nouveau_bo_validate(nvbo, false, false);
 		if (ret)
 			goto out;
 	}
 
 	nvbo->pin_refcnt++;
-	nouveau_bo_placement_set(nvbo, memtype, 0);
+	nouveau_bo_placement_set(nvbo, domain, 0);
 
 	/* drop pin_refcnt temporarily, so we don't trip the assertion
 	 * in nouveau_bo_move() that makes sure we're not trying to
@@ -490,7 +530,16 @@ nouveau_bo_unpin(struct nouveau_bo *nvbo)
 	if (ref)
 		goto out;
 
-	nouveau_bo_placement_set(nvbo, bo->mem.placement, 0);
+	switch (bo->mem.mem_type) {
+	case TTM_PL_VRAM:
+		nouveau_bo_placement_set(nvbo, NOUVEAU_GEM_DOMAIN_VRAM, 0);
+		break;
+	case TTM_PL_TT:
+		nouveau_bo_placement_set(nvbo, NOUVEAU_GEM_DOMAIN_GART, 0);
+		break;
+	default:
+		break;
+	}
 
 	ret = nouveau_bo_validate(nvbo, false, false);
 	if (ret == 0) {
@@ -574,6 +623,26 @@ nouveau_bo_sync_for_cpu(struct nouveau_bo *nvbo)
 					PAGE_SIZE, DMA_FROM_DEVICE);
 }
 
+void nouveau_bo_add_io_reserve_lru(struct ttm_buffer_object *bo)
+{
+	struct nouveau_drm *drm = nouveau_bdev(bo->bdev);
+	struct nouveau_bo *nvbo = nouveau_bo(bo);
+
+	mutex_lock(&drm->ttm.io_reserve_mutex);
+	list_move_tail(&nvbo->io_reserve_lru, &drm->ttm.io_reserve_lru);
+	mutex_unlock(&drm->ttm.io_reserve_mutex);
+}
+
+void nouveau_bo_del_io_reserve_lru(struct ttm_buffer_object *bo)
+{
+	struct nouveau_drm *drm = nouveau_bdev(bo->bdev);
+	struct nouveau_bo *nvbo = nouveau_bo(bo);
+
+	mutex_lock(&drm->ttm.io_reserve_mutex);
+	list_del_init(&nvbo->io_reserve_lru);
+	mutex_unlock(&drm->ttm.io_reserve_mutex);
+}
+
 int
 nouveau_bo_validate(struct nouveau_bo *nvbo, bool interruptible,
 		    bool no_wait_gpu)
@@ -646,6 +715,36 @@ nouveau_ttm_tt_create(struct ttm_buffer_object *bo, uint32_t page_flags)
 	return nouveau_sgdma_create_ttm(bo, page_flags);
 }
 
+static int
+nouveau_ttm_tt_bind(struct ttm_bo_device *bdev, struct ttm_tt *ttm,
+		    struct ttm_resource *reg)
+{
+#if IS_ENABLED(CONFIG_AGP)
+	struct nouveau_drm *drm = nouveau_bdev(bdev);
+#endif
+	if (!reg)
+		return -EINVAL;
+#if IS_ENABLED(CONFIG_AGP)
+	if (drm->agp.bridge)
+		return ttm_agp_bind(ttm, reg);
+#endif
+	return nouveau_sgdma_bind(bdev, ttm, reg);
+}
+
+static void
+nouveau_ttm_tt_unbind(struct ttm_bo_device *bdev, struct ttm_tt *ttm)
+{
+#if IS_ENABLED(CONFIG_AGP)
+	struct nouveau_drm *drm = nouveau_bdev(bdev);
+
+	if (drm->agp.bridge) {
+		ttm_agp_unbind(ttm);
+		return;
+	}
+#endif
+	nouveau_sgdma_unbind(bdev, ttm);
+}
+
 static void
 nouveau_bo_evict_flags(struct ttm_buffer_object *bo, struct ttm_placement *pl)
 {
@@ -653,11 +752,11 @@ nouveau_bo_evict_flags(struct ttm_buffer_object *bo, struct ttm_placement *pl)
 
 	switch (bo->mem.mem_type) {
 	case TTM_PL_VRAM:
-		nouveau_bo_placement_set(nvbo, TTM_PL_FLAG_TT,
-					 TTM_PL_FLAG_SYSTEM);
+		nouveau_bo_placement_set(nvbo, NOUVEAU_GEM_DOMAIN_GART,
+					 NOUVEAU_GEM_DOMAIN_CPU);
 		break;
 	default:
-		nouveau_bo_placement_set(nvbo, TTM_PL_FLAG_SYSTEM, 0);
+		nouveau_bo_placement_set(nvbo, NOUVEAU_GEM_DOMAIN_CPU, 0);
 		break;
 	}
 
@@ -725,7 +824,7 @@ nouveau_bo_move_m2mf(struct ttm_buffer_object *bo, int evict, bool intr,
 			if (ret == 0) {
 				ret = ttm_bo_move_accel_cleanup(bo,
 								&fence->base,
-								evict,
+								evict, false,
 								new_reg);
 				nouveau_fence_unref(&fence);
 			}
@@ -811,7 +910,8 @@ nouveau_bo_move_flipd(struct ttm_buffer_object *bo, bool evict, bool intr,
 	struct ttm_place placement_memtype = {
 		.fpfn = 0,
 		.lpfn = 0,
-		.flags = TTM_PL_FLAG_TT | TTM_PL_MASK_CACHING
+		.mem_type = TTM_PL_TT,
+		.flags = TTM_PL_MASK_CACHING
 	};
 	struct ttm_placement placement;
 	struct ttm_resource tmp_reg;
@@ -826,7 +926,11 @@ nouveau_bo_move_flipd(struct ttm_buffer_object *bo, bool evict, bool intr,
 	if (ret)
 		return ret;
 
-	ret = ttm_tt_bind(bo->ttm, &tmp_reg, &ctx);
+	ret = ttm_tt_populate(bo->bdev, bo->ttm, &ctx);
+	if (ret)
+		goto out;
+
+	ret = nouveau_ttm_tt_bind(bo->bdev, bo->ttm, &tmp_reg);
 	if (ret)
 		goto out;
 
@@ -848,7 +952,8 @@ nouveau_bo_move_flips(struct ttm_buffer_object *bo, bool evict, bool intr,
 	struct ttm_place placement_memtype = {
 		.fpfn = 0,
 		.lpfn = 0,
-		.flags = TTM_PL_FLAG_TT | TTM_PL_MASK_CACHING
+		.mem_type = TTM_PL_TT,
+		.flags = TTM_PL_MASK_CACHING
 	};
 	struct ttm_placement placement;
 	struct ttm_resource tmp_reg;
@@ -888,6 +993,8 @@ nouveau_bo_move_ntfy(struct ttm_buffer_object *bo, bool evict,
 	if (bo->destroy != nouveau_bo_del_ttm)
 		return;
 
+	nouveau_bo_del_io_reserve_lru(bo);
+
 	if (mem && new_reg->mem_type != TTM_PL_SYSTEM &&
 	    mem->mem.page == nvbo->page) {
 		list_for_each_entry(vma, &nvbo->vma_list, head) {
@@ -969,9 +1076,7 @@ nouveau_bo_move(struct ttm_buffer_object *bo, bool evict,
 
 	/* Fake bo copy. */
 	if (old_reg->mem_type == TTM_PL_SYSTEM && !bo->ttm) {
-		BUG_ON(bo->mem.mm_node != NULL);
-		bo->mem = *new_reg;
-		new_reg->mm_node = NULL;
+		ttm_bo_move_null(bo, new_reg);
 		goto out;
 	}
 
@@ -1018,32 +1123,60 @@ nouveau_bo_verify_access(struct ttm_buffer_object *bo, struct file *filp)
 					  filp->private_data);
 }
 
+static void
+nouveau_ttm_io_mem_free_locked(struct nouveau_drm *drm,
+			       struct ttm_resource *reg)
+{
+	struct nouveau_mem *mem = nouveau_mem(reg);
+
+	if (drm->client.mem->oclass >= NVIF_CLASS_MEM_NV50) {
+		switch (reg->mem_type) {
+		case TTM_PL_TT:
+			if (mem->kind)
+				nvif_object_unmap_handle(&mem->mem.object);
+			break;
+		case TTM_PL_VRAM:
+			nvif_object_unmap_handle(&mem->mem.object);
+			break;
+		default:
+			break;
+		}
+	}
+}
+
 static int
 nouveau_ttm_io_mem_reserve(struct ttm_bo_device *bdev, struct ttm_resource *reg)
 {
 	struct nouveau_drm *drm = nouveau_bdev(bdev);
 	struct nvkm_device *device = nvxx_device(&drm->client.device);
 	struct nouveau_mem *mem = nouveau_mem(reg);
+	int ret;
 
+	mutex_lock(&drm->ttm.io_reserve_mutex);
+retry:
 	switch (reg->mem_type) {
 	case TTM_PL_SYSTEM:
 		/* System memory */
-		return 0;
+		ret = 0;
+		goto out;
 	case TTM_PL_TT:
 #if IS_ENABLED(CONFIG_AGP)
 		if (drm->agp.bridge) {
-			reg->bus.offset = reg->start << PAGE_SHIFT;
-			reg->bus.base = drm->agp.base;
+			reg->bus.offset = (reg->start << PAGE_SHIFT) +
+				drm->agp.base;
 			reg->bus.is_iomem = !drm->agp.cma;
 		}
 #endif
-		if (drm->client.mem->oclass < NVIF_CLASS_MEM_NV50 || !mem->kind)
+		if (drm->client.mem->oclass < NVIF_CLASS_MEM_NV50 ||
+		    !mem->kind) {
 			/* untiled */
+			ret = 0;
 			break;
+		}
 		fallthrough;	/* tiled memory */
 	case TTM_PL_VRAM:
-		reg->bus.offset = reg->start << PAGE_SHIFT;
-		reg->bus.base = device->func->resource_addr(device, 1);
+		reg->bus.offset = (reg->start << PAGE_SHIFT) +
+			device->func->resource_addr(device, 1);
 		reg->bus.is_iomem = true;
 		if (drm->client.mem->oclass >= NVIF_CLASS_MEM_NV50) {
 			union {
@@ -1052,7 +1185,6 @@ nouveau_ttm_io_mem_reserve(struct ttm_bo_device *bdev, struct ttm_resource *reg)
 			} args;
 			u64 handle, length;
 			u32 argc = 0;
-			int ret;
 
 			switch (mem->mem.object.oclass) {
 			case NVIF_CLASS_MEM_NV50:
@@ -1078,39 +1210,46 @@ nouveau_ttm_io_mem_reserve(struct ttm_bo_device *bdev, struct ttm_resource *reg)
 						     &handle, &length);
 			if (ret != 1) {
 				if (WARN_ON(ret == 0))
-					return -EINVAL;
-				return ret;
+					ret = -EINVAL;
+				goto out;
 			}
 
-			reg->bus.base = 0;
 			reg->bus.offset = handle;
+			ret = 0;
 		}
 		break;
 	default:
-		return -EINVAL;
+		ret = -EINVAL;
 	}
-	return 0;
+
+out:
+	if (ret == -ENOSPC) {
+		struct nouveau_bo *nvbo;
+
+		nvbo = list_first_entry_or_null(&drm->ttm.io_reserve_lru,
+						typeof(*nvbo),
+						io_reserve_lru);
+		if (nvbo) {
+			list_del_init(&nvbo->io_reserve_lru);
+			drm_vma_node_unmap(&nvbo->bo.base.vma_node,
+					   bdev->dev_mapping);
+			nouveau_ttm_io_mem_free_locked(drm, &nvbo->bo.mem);
+			goto retry;
+		}
+
+	}
+	mutex_unlock(&drm->ttm.io_reserve_mutex);
+	return ret;
 }
 
 static void
 nouveau_ttm_io_mem_free(struct ttm_bo_device *bdev, struct ttm_resource *reg)
 {
 	struct nouveau_drm *drm = nouveau_bdev(bdev);
-	struct nouveau_mem *mem = nouveau_mem(reg);
 
-	if (drm->client.mem->oclass >= NVIF_CLASS_MEM_NV50) {
-		switch (reg->mem_type) {
-		case TTM_PL_TT:
-			if (mem->kind)
-				nvif_object_unmap_handle(&mem->mem.object);
-			break;
-		case TTM_PL_VRAM:
-			nvif_object_unmap_handle(&mem->mem.object);
-			break;
-		default:
-			break;
-		}
-	}
+	mutex_lock(&drm->ttm.io_reserve_mutex);
+	nouveau_ttm_io_mem_free_locked(drm, reg);
+	mutex_unlock(&drm->ttm.io_reserve_mutex);
 }
 
 static int
@@ -1131,7 +1270,8 @@ nouveau_ttm_fault_reserve_notify(struct ttm_buffer_object *bo)
 			return 0;
 
 		if (bo->mem.mem_type == TTM_PL_SYSTEM) {
-			nouveau_bo_placement_set(nvbo, TTM_PL_TT, 0);
+			nouveau_bo_placement_set(nvbo, NOUVEAU_GEM_DOMAIN_GART,
+						 0);
 
 			ret = nouveau_bo_validate(nvbo, false, false);
 			if (ret)
@@ -1155,35 +1295,36 @@ nouveau_ttm_fault_reserve_notify(struct ttm_buffer_object *bo)
 		nvbo->busy_placements[i].lpfn = mappable;
 	}
 
-	nouveau_bo_placement_set(nvbo, TTM_PL_FLAG_VRAM, 0);
+	nouveau_bo_placement_set(nvbo, NOUVEAU_GEM_DOMAIN_VRAM, 0);
 	return nouveau_bo_validate(nvbo, false, false);
 }
 
 static int
-nouveau_ttm_tt_populate(struct ttm_tt *ttm, struct ttm_operation_ctx *ctx)
+nouveau_ttm_tt_populate(struct ttm_bo_device *bdev,
+			struct ttm_tt *ttm, struct ttm_operation_ctx *ctx)
 {
 	struct ttm_dma_tt *ttm_dma = (void *)ttm;
 	struct nouveau_drm *drm;
 	struct device *dev;
 	bool slave = !!(ttm->page_flags & TTM_PAGE_FLAG_SG);
 
-	if (ttm->state != tt_unpopulated)
+	if (ttm_tt_is_populated(ttm))
 		return 0;
 
 	if (slave && ttm->sg) {
 		/* make userspace faulting work */
 		drm_prime_sg_to_page_addr_arrays(ttm->sg, ttm->pages,
 						 ttm_dma->dma_address, ttm->num_pages);
-		ttm->state = tt_unbound;
+		ttm_tt_set_populated(ttm);
 		return 0;
 	}
 
-	drm = nouveau_bdev(ttm->bdev);
+	drm = nouveau_bdev(bdev);
 	dev = drm->dev->dev;
 
 #if IS_ENABLED(CONFIG_AGP)
 	if (drm->agp.bridge) {
-		return ttm_agp_tt_populate(ttm, ctx);
+		return ttm_pool_populate(ttm, ctx);
 	}
 #endif
 
@@ -1196,7 +1337,8 @@ nouveau_ttm_tt_populate(struct ttm_tt *ttm, struct ttm_operation_ctx *ctx)
 }
 
 static void
-nouveau_ttm_tt_unpopulate(struct ttm_tt *ttm)
+nouveau_ttm_tt_unpopulate(struct ttm_bo_device *bdev,
+			  struct ttm_tt *ttm)
 {
 	struct ttm_dma_tt *ttm_dma = (void *)ttm;
 	struct nouveau_drm *drm;
@@ -1206,12 +1348,12 @@ nouveau_ttm_tt_unpopulate(struct ttm_tt *ttm)
 	if (slave)
 		return;
 
-	drm = nouveau_bdev(ttm->bdev);
+	drm = nouveau_bdev(bdev);
 	dev = drm->dev->dev;
 
 #if IS_ENABLED(CONFIG_AGP)
 	if (drm->agp.bridge) {
-		ttm_agp_tt_unpopulate(ttm);
+		ttm_pool_unpopulate(ttm);
 		return;
 	}
 #endif
@@ -1226,6 +1368,22 @@ nouveau_ttm_tt_unpopulate(struct ttm_tt *ttm)
 	ttm_unmap_and_unpopulate_pages(dev, ttm_dma);
 }
 
+static void
+nouveau_ttm_tt_destroy(struct ttm_bo_device *bdev,
+		       struct ttm_tt *ttm)
+{
+#if IS_ENABLED(CONFIG_AGP)
+	struct nouveau_drm *drm = nouveau_bdev(bdev);
+	if (drm->agp.bridge) {
+		ttm_agp_unbind(ttm);
+		ttm_tt_destroy_common(bdev, ttm);
+		ttm_agp_destroy(ttm);
+		return;
+	}
+#endif
+	nouveau_sgdma_destroy(bdev, ttm);
+}
+
 void
 nouveau_bo_fence(struct nouveau_bo *nvbo, struct nouveau_fence *fence, bool exclusive)
 {
@@ -1241,6 +1399,9 @@ struct ttm_bo_driver nouveau_bo_driver = {
 	.ttm_tt_create = &nouveau_ttm_tt_create,
 	.ttm_tt_populate = &nouveau_ttm_tt_populate,
 	.ttm_tt_unpopulate = &nouveau_ttm_tt_unpopulate,
+	.ttm_tt_bind = &nouveau_ttm_tt_bind,
+	.ttm_tt_unbind = &nouveau_ttm_tt_unbind,
+	.ttm_tt_destroy = &nouveau_ttm_tt_destroy,
 	.eviction_valuable = ttm_bo_eviction_valuable,
 	.evict_flags = nouveau_bo_evict_flags,
 	.move_notify = nouveau_bo_move_ntfy,
diff --git a/drivers/gpu/drm/nouveau/nouveau_bo.h b/drivers/gpu/drm/nouveau/nouveau_bo.h
index aecb7481df0d..2a23c8207436 100644
--- a/drivers/gpu/drm/nouveau/nouveau_bo.h
+++ b/drivers/gpu/drm/nouveau/nouveau_bo.h
@@ -18,6 +18,7 @@ struct nouveau_bo {
 	bool force_coherent;
 	struct ttm_bo_kmap_obj kmap;
 	struct list_head head;
+	struct list_head io_reserve_lru;
 
 	/* protected by ttm_bo_reserve() */
 	struct drm_file *reserved_by;
@@ -76,10 +77,10 @@ extern struct ttm_bo_driver nouveau_bo_driver;
 
 void nouveau_bo_move_init(struct nouveau_drm *);
 struct nouveau_bo *nouveau_bo_alloc(struct nouveau_cli *, u64 *size, int *align,
-				    u32 flags, u32 tile_mode, u32 tile_flags);
-int  nouveau_bo_init(struct nouveau_bo *, u64 size, int align, u32 flags,
+				    u32 domain, u32 tile_mode, u32 tile_flags);
+int  nouveau_bo_init(struct nouveau_bo *, u64 size, int align, u32 domain,
 		     struct sg_table *sg, struct dma_resv *robj);
-int  nouveau_bo_new(struct nouveau_cli *, u64 size, int align, u32 flags,
+int  nouveau_bo_new(struct nouveau_cli *, u64 size, int align, u32 domain,
 		    u32 tile_mode, u32 tile_flags, struct sg_table *sg,
 		    struct dma_resv *robj,
 		    struct nouveau_bo **);
@@ -96,6 +97,8 @@ int  nouveau_bo_validate(struct nouveau_bo *, bool interruptible,
 			 bool no_wait_gpu);
 void nouveau_bo_sync_for_device(struct nouveau_bo *nvbo);
 void nouveau_bo_sync_for_cpu(struct nouveau_bo *nvbo);
+void nouveau_bo_add_io_reserve_lru(struct ttm_buffer_object *bo);
+void nouveau_bo_del_io_reserve_lru(struct ttm_buffer_object *bo);
 
 /* TODO: submit equivalent to TTM generic API upstream? */
 static inline void __iomem *
@@ -119,13 +122,13 @@ nouveau_bo_unmap_unpin_unref(struct nouveau_bo **pnvbo)
 }
 
 static inline int
-nouveau_bo_new_pin_map(struct nouveau_cli *cli, u64 size, int align, u32 flags,
+nouveau_bo_new_pin_map(struct nouveau_cli *cli, u64 size, int align, u32 domain,
 		       struct nouveau_bo **pnvbo)
 {
-	int ret = nouveau_bo_new(cli, size, align, flags,
+	int ret = nouveau_bo_new(cli, size, align, domain,
 				 0, 0, NULL, NULL, pnvbo);
 	if (ret == 0) {
-		ret = nouveau_bo_pin(*pnvbo, flags, true);
+		ret = nouveau_bo_pin(*pnvbo, domain, true);
 		if (ret == 0) {
 			ret = nouveau_bo_map(*pnvbo);
 			if (ret == 0)
diff --git a/drivers/gpu/drm/nouveau/nouveau_chan.c b/drivers/gpu/drm/nouveau/nouveau_chan.c
index b80e4ebf14a6..8f099601d2f2 100644
--- a/drivers/gpu/drm/nouveau/nouveau_chan.c
+++ b/drivers/gpu/drm/nouveau/nouveau_chan.c
@@ -163,9 +163,9 @@ nouveau_channel_prep(struct nouveau_drm *drm, struct nvif_device *device,
 	atomic_set(&chan->killed, 0);
 
 	/* allocate memory for dma push buffer */
-	target = TTM_PL_FLAG_TT | TTM_PL_FLAG_UNCACHED;
+	target = NOUVEAU_GEM_DOMAIN_GART | NOUVEAU_GEM_DOMAIN_COHERENT;
 	if (nouveau_vram_pushbuf)
-		target = TTM_PL_FLAG_VRAM;
+		target = NOUVEAU_GEM_DOMAIN_VRAM;
 
 	ret = nouveau_bo_new(cli, size, 0, target, 0, 0, NULL, NULL,
 			    &chan->push.buffer);
diff --git a/drivers/gpu/drm/nouveau/nouveau_dmem.c b/drivers/gpu/drm/nouveau/nouveau_dmem.c
index 4e8112fde3e6..af70ef8e5471 100644
--- a/drivers/gpu/drm/nouveau/nouveau_dmem.c
+++ b/drivers/gpu/drm/nouveau/nouveau_dmem.c
@@ -254,12 +254,12 @@ nouveau_dmem_chunk_alloc(struct nouveau_drm *drm, struct page **ppage)
 	chunk->pagemap.owner = drm->dev;
 
 	ret = nouveau_bo_new(&drm->client, DMEM_CHUNK_SIZE, 0,
-			     TTM_PL_FLAG_VRAM, 0, 0, NULL, NULL,
+			     NOUVEAU_GEM_DOMAIN_VRAM, 0, 0, NULL, NULL,
 			     &chunk->bo);
 	if (ret)
 		goto out_release;
 
-	ret = nouveau_bo_pin(chunk->bo, TTM_PL_FLAG_VRAM, false);
+	ret = nouveau_bo_pin(chunk->bo, NOUVEAU_GEM_DOMAIN_VRAM, false);
 	if (ret)
 		goto out_bo_free;
 
@@ -346,7 +346,7 @@ nouveau_dmem_resume(struct nouveau_drm *drm)
 
 	mutex_lock(&drm->dmem->mutex);
 	list_for_each_entry(chunk, &drm->dmem->chunks, list) {
-		ret = nouveau_bo_pin(chunk->bo, TTM_PL_FLAG_VRAM, false);
+		ret = nouveau_bo_pin(chunk->bo, NOUVEAU_GEM_DOMAIN_VRAM, false);
 		/* FIXME handle pin failure */
 		WARN_ON(ret);
 	}
diff --git a/drivers/gpu/drm/nouveau/nouveau_drv.h b/drivers/gpu/drm/nouveau/nouveau_drv.h
index 73ebf5fba2fc..b8025507a9e4 100644
--- a/drivers/gpu/drm/nouveau/nouveau_drv.h
+++ b/drivers/gpu/drm/nouveau/nouveau_drv.h
@@ -164,6 +164,8 @@ struct nouveau_drm {
 		int type_vram;
 		int type_host[2];
 		int type_ncoh[2];
+		struct mutex io_reserve_mutex;
+		struct list_head io_reserve_lru;
 	} ttm;
 
 	/* GEM interface support */
diff --git a/drivers/gpu/drm/nouveau/nouveau_fbcon.c b/drivers/gpu/drm/nouveau/nouveau_fbcon.c
index fad8030ec1f8..24ec5339efb4 100644
--- a/drivers/gpu/drm/nouveau/nouveau_fbcon.c
+++ b/drivers/gpu/drm/nouveau/nouveau_fbcon.c
@@ -341,7 +341,7 @@ nouveau_fbcon_create(struct drm_fb_helper *helper,
 	if (ret)
 		goto out_unref;
 
-	ret = nouveau_bo_pin(nvbo, TTM_PL_FLAG_VRAM, false);
+	ret = nouveau_bo_pin(nvbo, NOUVEAU_GEM_DOMAIN_VRAM, false);
 	if (ret) {
 		NV_ERROR(drm, "failed to pin fb: %d\n", ret);
 		goto out_unref;
@@ -378,8 +378,7 @@ nouveau_fbcon_create(struct drm_fb_helper *helper,
 			      FBINFO_HWACCEL_FILLRECT |
 			      FBINFO_HWACCEL_IMAGEBLIT;
 	info->fbops = &nouveau_fbcon_sw_ops;
-	info->fix.smem_start = nvbo->bo.mem.bus.base +
-			       nvbo->bo.mem.bus.offset;
+	info->fix.smem_start = nvbo->bo.mem.bus.offset;
 	info->fix.smem_len = nvbo->bo.mem.num_pages << PAGE_SHIFT;
 
 	info->screen_base = nvbo_kmap_obj_iovirtual(nvbo);
diff --git a/drivers/gpu/drm/nouveau/nouveau_gem.c b/drivers/gpu/drm/nouveau/nouveau_gem.c
index 81f111ad3f4f..89adadf4706b 100644
--- a/drivers/gpu/drm/nouveau/nouveau_gem.c
+++ b/drivers/gpu/drm/nouveau/nouveau_gem.c
@@ -176,20 +176,12 @@ nouveau_gem_new(struct nouveau_cli *cli, u64 size, int align, uint32_t domain,
 {
 	struct nouveau_drm *drm = cli->drm;
 	struct nouveau_bo *nvbo;
-	u32 flags = 0;
 	int ret;
 
-	if (domain & NOUVEAU_GEM_DOMAIN_VRAM)
-		flags |= TTM_PL_FLAG_VRAM;
-	if (domain & NOUVEAU_GEM_DOMAIN_GART)
-		flags |= TTM_PL_FLAG_TT;
-	if (!flags || domain & NOUVEAU_GEM_DOMAIN_CPU)
-		flags |= TTM_PL_FLAG_SYSTEM;
+	if (!(domain & (NOUVEAU_GEM_DOMAIN_VRAM | NOUVEAU_GEM_DOMAIN_GART)))
+		domain |= NOUVEAU_GEM_DOMAIN_CPU;
 
-	if (domain & NOUVEAU_GEM_DOMAIN_COHERENT)
-		flags |= TTM_PL_FLAG_UNCACHED;
-
-	nvbo = nouveau_bo_alloc(cli, &size, &align, flags, tile_mode,
+	nvbo = nouveau_bo_alloc(cli, &size, &align, domain, tile_mode,
 				tile_flags);
 	if (IS_ERR(nvbo))
 		return PTR_ERR(nvbo);
@@ -202,7 +194,7 @@ nouveau_gem_new(struct nouveau_cli *cli, u64 size, int align, uint32_t domain,
 		return ret;
 	}
 
-	ret = nouveau_bo_init(nvbo, size, align, flags, NULL, NULL);
+	ret = nouveau_bo_init(nvbo, size, align, domain, NULL, NULL);
 	if (ret) {
 		nouveau_bo_ref(NULL, &nvbo);
 		return ret;
@@ -296,32 +288,28 @@ nouveau_gem_set_domain(struct drm_gem_object *gem, uint32_t read_domains,
 	struct ttm_buffer_object *bo = &nvbo->bo;
 	uint32_t domains = valid_domains & nvbo->valid_domains &
 		(write_domains ? write_domains : read_domains);
-	uint32_t pref_flags = 0, valid_flags = 0;
+	uint32_t pref_domains = 0;;
 
 	if (!domains)
 		return -EINVAL;
 
-	if (valid_domains & NOUVEAU_GEM_DOMAIN_VRAM)
-		valid_flags |= TTM_PL_FLAG_VRAM;
-
-	if (valid_domains & NOUVEAU_GEM_DOMAIN_GART)
-		valid_flags |= TTM_PL_FLAG_TT;
+	valid_domains &= ~(NOUVEAU_GEM_DOMAIN_VRAM | NOUVEAU_GEM_DOMAIN_GART);
 
 	if ((domains & NOUVEAU_GEM_DOMAIN_VRAM) &&
 	    bo->mem.mem_type == TTM_PL_VRAM)
-		pref_flags |= TTM_PL_FLAG_VRAM;
+		pref_domains |= NOUVEAU_GEM_DOMAIN_VRAM;
 
 	else if ((domains & NOUVEAU_GEM_DOMAIN_GART) &&
 		 bo->mem.mem_type == TTM_PL_TT)
-		pref_flags |= TTM_PL_FLAG_TT;
+		pref_domains |= NOUVEAU_GEM_DOMAIN_GART;
 
 	else if (domains & NOUVEAU_GEM_DOMAIN_VRAM)
-		pref_flags |= TTM_PL_FLAG_VRAM;
+		pref_domains |= NOUVEAU_GEM_DOMAIN_VRAM;
 
 	else
-		pref_flags |= TTM_PL_FLAG_TT;
+		pref_domains |= NOUVEAU_GEM_DOMAIN_GART;
 
-	nouveau_bo_placement_set(nvbo, pref_flags, valid_flags);
+	nouveau_bo_placement_set(nvbo, pref_domains, valid_domains);
 
 	return 0;
 }
diff --git a/drivers/gpu/drm/nouveau/nouveau_prime.c b/drivers/gpu/drm/nouveau/nouveau_prime.c
index bae6a3eccee0..b2ecb91f8ddc 100644
--- a/drivers/gpu/drm/nouveau/nouveau_prime.c
+++ b/drivers/gpu/drm/nouveau/nouveau_prime.c
@@ -32,7 +32,7 @@ struct sg_table *nouveau_gem_prime_get_sg_table(struct drm_gem_object *obj)
 	struct nouveau_bo *nvbo = nouveau_gem_object(obj);
 	int npages = nvbo->bo.num_pages;
 
-	return drm_prime_pages_to_sg(nvbo->bo.ttm->pages, npages);
+	return drm_prime_pages_to_sg(obj->dev, nvbo->bo.ttm->pages, npages);
 }
 
 void *nouveau_gem_prime_vmap(struct drm_gem_object *obj)
@@ -64,14 +64,12 @@ struct drm_gem_object *nouveau_gem_prime_import_sg_table(struct drm_device *dev,
 	struct nouveau_bo *nvbo;
 	struct dma_resv *robj = attach->dmabuf->resv;
 	u64 size = attach->dmabuf->size;
-	u32 flags = 0;
 	int align = 0;
 	int ret;
 
-	flags = TTM_PL_FLAG_TT;
-
 	dma_resv_lock(robj, NULL);
-	nvbo = nouveau_bo_alloc(&drm->client, &size, &align, flags, 0, 0);
+	nvbo = nouveau_bo_alloc(&drm->client, &size, &align,
+				NOUVEAU_GEM_DOMAIN_GART, 0, 0);
 	if (IS_ERR(nvbo)) {
 		obj = ERR_CAST(nvbo);
 		goto unlock;
@@ -88,7 +86,8 @@ struct drm_gem_object *nouveau_gem_prime_import_sg_table(struct drm_device *dev,
 		goto unlock;
 	}
 
-	ret = nouveau_bo_init(nvbo, size, align, flags, sg, robj);
+	ret = nouveau_bo_init(nvbo, size, align, NOUVEAU_GEM_DOMAIN_GART,
+			      sg, robj);
 	if (ret) {
 		nouveau_bo_ref(NULL, &nvbo);
 		obj = ERR_PTR(ret);
@@ -108,7 +107,7 @@ int nouveau_gem_prime_pin(struct drm_gem_object *obj)
 	int ret;
 
 	/* pin buffer into GTT */
-	ret = nouveau_bo_pin(nvbo, TTM_PL_FLAG_TT, false);
+	ret = nouveau_bo_pin(nvbo, NOUVEAU_GEM_DOMAIN_GART, false);
 	if (ret)
 		return -EINVAL;
 
diff --git a/drivers/gpu/drm/nouveau/nouveau_sgdma.c b/drivers/gpu/drm/nouveau/nouveau_sgdma.c
index eef75c53a197..806d9ec310f5 100644
--- a/drivers/gpu/drm/nouveau/nouveau_sgdma.c
+++ b/drivers/gpu/drm/nouveau/nouveau_sgdma.c
@@ -14,87 +14,65 @@ struct nouveau_sgdma_be {
 	struct nouveau_mem *mem;
 };
 
-static void
-nouveau_sgdma_destroy(struct ttm_tt *ttm)
+void
+nouveau_sgdma_destroy(struct ttm_bo_device *bdev, struct ttm_tt *ttm)
 {
 	struct nouveau_sgdma_be *nvbe = (struct nouveau_sgdma_be *)ttm;
 
 	if (ttm) {
+		nouveau_sgdma_unbind(bdev, ttm);
+		ttm_tt_destroy_common(bdev, ttm);
 		ttm_dma_tt_fini(&nvbe->ttm);
 		kfree(nvbe);
 	}
 }
 
-static int
-nv04_sgdma_bind(struct ttm_tt *ttm, struct ttm_resource *reg)
+int
+nouveau_sgdma_bind(struct ttm_bo_device *bdev, struct ttm_tt *ttm, struct ttm_resource *reg)
 {
 	struct nouveau_sgdma_be *nvbe = (struct nouveau_sgdma_be *)ttm;
+	struct nouveau_drm *drm = nouveau_bdev(bdev);
 	struct nouveau_mem *mem = nouveau_mem(reg);
 	int ret;
 
+	if (nvbe->mem)
+		return 0;
+
 	ret = nouveau_mem_host(reg, &nvbe->ttm);
 	if (ret)
 		return ret;
 
-	ret = nouveau_mem_map(mem, &mem->cli->vmm.vmm, &mem->vma[0]);
-	if (ret) {
-		nouveau_mem_fini(mem);
-		return ret;
+	if (drm->client.device.info.family < NV_DEVICE_INFO_V0_TESLA) {
+		ret = nouveau_mem_map(mem, &mem->cli->vmm.vmm, &mem->vma[0]);
+		if (ret) {
+			nouveau_mem_fini(mem);
+			return ret;
+		}
 	}
 
 	nvbe->mem = mem;
 	return 0;
 }
 
-static void
-nv04_sgdma_unbind(struct ttm_tt *ttm)
+void
+nouveau_sgdma_unbind(struct ttm_bo_device *bdev, struct ttm_tt *ttm)
 {
 	struct nouveau_sgdma_be *nvbe = (struct nouveau_sgdma_be *)ttm;
-	nouveau_mem_fini(nvbe->mem);
-}
-
-static struct ttm_backend_func nv04_sgdma_backend = {
-	.bind			= nv04_sgdma_bind,
-	.unbind			= nv04_sgdma_unbind,
-	.destroy		= nouveau_sgdma_destroy
-};
-
-static int
-nv50_sgdma_bind(struct ttm_tt *ttm, struct ttm_resource *reg)
-{
-	struct nouveau_sgdma_be *nvbe = (struct nouveau_sgdma_be *)ttm;
-	struct nouveau_mem *mem = nouveau_mem(reg);
-	int ret;
-
-	ret = nouveau_mem_host(reg, &nvbe->ttm);
-	if (ret)
-		return ret;
-
-	nvbe->mem = mem;
-	return 0;
+	if (nvbe->mem) {
+		nouveau_mem_fini(nvbe->mem);
+		nvbe->mem = NULL;
+	}
 }
 
-static struct ttm_backend_func nv50_sgdma_backend = {
-	.bind			= nv50_sgdma_bind,
-	.unbind			= nv04_sgdma_unbind,
-	.destroy		= nouveau_sgdma_destroy
-};
-
 struct ttm_tt *
 nouveau_sgdma_create_ttm(struct ttm_buffer_object *bo, uint32_t page_flags)
 {
-	struct nouveau_drm *drm = nouveau_bdev(bo->bdev);
 	struct nouveau_sgdma_be *nvbe;
 
 	nvbe = kzalloc(sizeof(*nvbe), GFP_KERNEL);
 	if (!nvbe)
 		return NULL;
 
-	if (drm->client.device.info.family < NV_DEVICE_INFO_V0_TESLA)
-		nvbe->ttm.ttm.func = &nv04_sgdma_backend;
-	else
-		nvbe->ttm.ttm.func = &nv50_sgdma_backend;
-
 	if (ttm_dma_tt_init(&nvbe->ttm, bo, page_flags)) {
 		kfree(nvbe);
 		return NULL;
diff --git a/drivers/gpu/drm/nouveau/nouveau_ttm.c b/drivers/gpu/drm/nouveau/nouveau_ttm.c
index 53c6f8827322..427341753441 100644
--- a/drivers/gpu/drm/nouveau/nouveau_ttm.c
+++ b/drivers/gpu/drm/nouveau/nouveau_ttm.c
@@ -123,13 +123,51 @@ const struct ttm_resource_manager_func nv04_gart_manager = {
 	.free = nouveau_manager_del,
 };
 
+static vm_fault_t nouveau_ttm_fault(struct vm_fault *vmf)
+{
+	struct vm_area_struct *vma = vmf->vma;
+	struct ttm_buffer_object *bo = vma->vm_private_data;
+	pgprot_t prot;
+	vm_fault_t ret;
+
+	ret = ttm_bo_vm_reserve(bo, vmf);
+	if (ret)
+		return ret;
+
+	nouveau_bo_del_io_reserve_lru(bo);
+
+	prot = vm_get_page_prot(vma->vm_flags);
+	ret = ttm_bo_vm_fault_reserved(vmf, prot, TTM_BO_VM_NUM_PREFAULT, 1);
+	if (ret == VM_FAULT_RETRY && !(vmf->flags & FAULT_FLAG_RETRY_NOWAIT))
+		return ret;
+
+	nouveau_bo_add_io_reserve_lru(bo);
+
+	dma_resv_unlock(bo->base.resv);
+
+	return ret;
+}
+
+static struct vm_operations_struct nouveau_ttm_vm_ops = {
+	.fault = nouveau_ttm_fault,
+	.open = ttm_bo_vm_open,
+	.close = ttm_bo_vm_close,
+	.access = ttm_bo_vm_access
+};
+
 int
 nouveau_ttm_mmap(struct file *filp, struct vm_area_struct *vma)
 {
 	struct drm_file *file_priv = filp->private_data;
 	struct nouveau_drm *drm = nouveau_drm(file_priv->minor->dev);
+	int ret;
 
-	return ttm_bo_mmap(filp, vma, &drm->ttm.bdev);
+	ret = ttm_bo_mmap(filp, vma, &drm->ttm.bdev);
+	if (ret)
+		return ret;
+
+	vma->vm_ops = &nouveau_ttm_vm_ops;
+	return 0;
 }
 
 static int
@@ -156,24 +194,13 @@ nouveau_ttm_init_host(struct nouveau_drm *drm, u8 kind)
 static int
 nouveau_ttm_init_vram(struct nouveau_drm *drm)
 {
-	struct nvif_mmu *mmu = &drm->client.mmu;
 	if (drm->client.device.info.family >= NV_DEVICE_INFO_V0_TESLA) {
-		/* Some BARs do not support being ioremapped WC */
-		const u8 type = mmu->type[drm->ttm.type_vram].type;
 		struct ttm_resource_manager *man = kzalloc(sizeof(*man), GFP_KERNEL);
+
 		if (!man)
 			return -ENOMEM;
 
-		man->available_caching = TTM_PL_FLAG_UNCACHED | TTM_PL_FLAG_WC;
-		man->default_caching = TTM_PL_FLAG_WC;
-
-		if (type & NVIF_MEM_UNCACHED) {
-			man->available_caching = TTM_PL_FLAG_UNCACHED;
-			man->default_caching = TTM_PL_FLAG_UNCACHED;
-		}
-
 		man->func = &nouveau_vram_manager;
-		man->use_io_reserve_lru = true;
 
 		ttm_resource_manager_init(man,
 					  drm->gem.vram_available >> PAGE_SHIFT);
@@ -181,9 +208,7 @@ nouveau_ttm_init_vram(struct nouveau_drm *drm)
 		ttm_resource_manager_set_used(man, true);
 		return 0;
 	} else {
-		return ttm_range_man_init(&drm->ttm.bdev, TTM_PL_VRAM,
-					  TTM_PL_FLAG_UNCACHED | TTM_PL_FLAG_WC,
-					  TTM_PL_FLAG_WC, false,
+		return ttm_range_man_init(&drm->ttm.bdev, TTM_PL_VRAM, false,
 					  drm->gem.vram_available >> PAGE_SHIFT);
 	}
 }
@@ -208,25 +233,14 @@ nouveau_ttm_init_gtt(struct nouveau_drm *drm)
 {
 	struct ttm_resource_manager *man;
 	unsigned long size_pages = drm->gem.gart_available >> PAGE_SHIFT;
-	unsigned available_caching, default_caching;
 	const struct ttm_resource_manager_func *func = NULL;
-	if (drm->agp.bridge) {
-		available_caching = TTM_PL_FLAG_UNCACHED |
-			TTM_PL_FLAG_WC;
-		default_caching = TTM_PL_FLAG_WC;
-	} else {
-		available_caching = TTM_PL_MASK_CACHING;
-		default_caching = TTM_PL_FLAG_CACHED;
-	}
 
 	if (drm->client.device.info.family >= NV_DEVICE_INFO_V0_TESLA)
 		func = &nouveau_gart_manager;
 	else if (!drm->agp.bridge)
 		func = &nv04_gart_manager;
 	else
-		return ttm_range_man_init(&drm->ttm.bdev, TTM_PL_TT,
-					  available_caching, default_caching,
-					  true,
+		return ttm_range_man_init(&drm->ttm.bdev, TTM_PL_TT, true,
 					  size_pages);
 
 	man = kzalloc(sizeof(*man), GFP_KERNEL);
@@ -234,8 +248,6 @@ nouveau_ttm_init_gtt(struct nouveau_drm *drm)
 		return -ENOMEM;
 
 	man->func = func;
-	man->available_caching = available_caching;
-	man->default_caching = default_caching;
 	man->use_tt = true;
 	ttm_resource_manager_init(man, size_pages);
 	ttm_set_driver_manager(&drm->ttm.bdev, TTM_PL_TT, man);
@@ -339,6 +351,9 @@ nouveau_ttm_init(struct nouveau_drm *drm)
 		return ret;
 	}
 
+	mutex_init(&drm->ttm.io_reserve_mutex);
+	INIT_LIST_HEAD(&drm->ttm.io_reserve_lru);
+
 	NV_INFO(drm, "VRAM: %d MiB\n", (u32)(drm->gem.vram_available >> 20));
 	NV_INFO(drm, "GART: %d MiB\n", (u32)(drm->gem.gart_available >> 20));
 	return 0;
diff --git a/drivers/gpu/drm/nouveau/nouveau_ttm.h b/drivers/gpu/drm/nouveau/nouveau_ttm.h
index eaf25461cd91..69552049bb96 100644
--- a/drivers/gpu/drm/nouveau/nouveau_ttm.h
+++ b/drivers/gpu/drm/nouveau/nouveau_ttm.h
@@ -22,4 +22,7 @@ int  nouveau_ttm_mmap(struct file *, struct vm_area_struct *);
 int  nouveau_ttm_global_init(struct nouveau_drm *);
 void nouveau_ttm_global_release(struct nouveau_drm *);
 
+int nouveau_sgdma_bind(struct ttm_bo_device *bdev, struct ttm_tt *ttm, struct ttm_resource *reg);
+void nouveau_sgdma_unbind(struct ttm_bo_device *bdev, struct ttm_tt *ttm);
+void nouveau_sgdma_destroy(struct ttm_bo_device *bdev, struct ttm_tt *ttm);
 #endif
diff --git a/drivers/gpu/drm/nouveau/nv17_fence.c b/drivers/gpu/drm/nouveau/nv17_fence.c
index 6b697ee6bc0e..1253fdec712d 100644
--- a/drivers/gpu/drm/nouveau/nv17_fence.c
+++ b/drivers/gpu/drm/nouveau/nv17_fence.c
@@ -130,10 +130,11 @@ nv17_fence_create(struct nouveau_drm *drm)
 	priv->base.context_del = nv10_fence_context_del;
 	spin_lock_init(&priv->lock);
 
-	ret = nouveau_bo_new(&drm->client, 4096, 0x1000, TTM_PL_FLAG_VRAM,
+	ret = nouveau_bo_new(&drm->client, 4096, 0x1000,
+			     NOUVEAU_GEM_DOMAIN_VRAM,
 			     0, 0x0000, NULL, NULL, &priv->bo);
 	if (!ret) {
-		ret = nouveau_bo_pin(priv->bo, TTM_PL_FLAG_VRAM, false);
+		ret = nouveau_bo_pin(priv->bo, NOUVEAU_GEM_DOMAIN_VRAM, false);
 		if (!ret) {
 			ret = nouveau_bo_map(priv->bo);
 			if (ret)
diff --git a/drivers/gpu/drm/nouveau/nv50_fence.c b/drivers/gpu/drm/nouveau/nv50_fence.c
index 49b46f51073c..447238e3cbe7 100644
--- a/drivers/gpu/drm/nouveau/nv50_fence.c
+++ b/drivers/gpu/drm/nouveau/nv50_fence.c
@@ -81,10 +81,11 @@ nv50_fence_create(struct nouveau_drm *drm)
 	priv->base.context_del = nv10_fence_context_del;
 	spin_lock_init(&priv->lock);
 
-	ret = nouveau_bo_new(&drm->client, 4096, 0x1000, TTM_PL_FLAG_VRAM,
+	ret = nouveau_bo_new(&drm->client, 4096, 0x1000,
+			     NOUVEAU_GEM_DOMAIN_VRAM,
 			     0, 0x0000, NULL, NULL, &priv->bo);
 	if (!ret) {
-		ret = nouveau_bo_pin(priv->bo, TTM_PL_FLAG_VRAM, false);
+		ret = nouveau_bo_pin(priv->bo, NOUVEAU_GEM_DOMAIN_VRAM, false);
 		if (!ret) {
 			ret = nouveau_bo_map(priv->bo);
 			if (ret)
diff --git a/drivers/gpu/drm/nouveau/nv84_fence.c b/drivers/gpu/drm/nouveau/nv84_fence.c
index 7ed36b3a6b7d..7c9c928c3196 100644
--- a/drivers/gpu/drm/nouveau/nv84_fence.c
+++ b/drivers/gpu/drm/nouveau/nv84_fence.c
@@ -209,12 +209,13 @@ nv84_fence_create(struct nouveau_drm *drm)
 	mutex_init(&priv->mutex);
 
 	/* Use VRAM if there is any ; otherwise fallback to system memory */
-	domain = drm->client.device.info.ram_size != 0 ? TTM_PL_FLAG_VRAM :
-			 /*
-			  * fences created in sysmem must be non-cached or we
-			  * will lose CPU/GPU coherency!
-			  */
-			 TTM_PL_FLAG_TT | TTM_PL_FLAG_UNCACHED;
+	domain = drm->client.device.info.ram_size != 0 ?
+		NOUVEAU_GEM_DOMAIN_VRAM :
+		 /*
+		  * fences created in sysmem must be non-cached or we
+		  * will lose CPU/GPU coherency!
+		  */
+		NOUVEAU_GEM_DOMAIN_GART | NOUVEAU_GEM_DOMAIN_COHERENT;
 	ret = nouveau_bo_new(&drm->client, 16 * drm->chan.nr, 0,
 			     domain, 0, 0, NULL, NULL, &priv->bo);
 	if (ret == 0) {
diff --git a/drivers/gpu/drm/panel/Kconfig b/drivers/gpu/drm/panel/Kconfig
index 8d97d07c5871..b9dbedf8f15e 100644
--- a/drivers/gpu/drm/panel/Kconfig
+++ b/drivers/gpu/drm/panel/Kconfig
@@ -324,13 +324,30 @@ config DRM_PANEL_SAMSUNG_S6E63J0X03
 	select VIDEOMODE_HELPERS
 
 config DRM_PANEL_SAMSUNG_S6E63M0
-	tristate "Samsung S6E63M0 RGB/SPI panel"
+	tristate "Samsung S6E63M0 RGB panel"
 	depends on OF
-	depends on SPI
 	depends on BACKLIGHT_CLASS_DEVICE
 	help
 	  Say Y here if you want to enable support for Samsung S6E63M0
-	  AMOLED LCD panel.
+	  AMOLED LCD panel. This panel can be accessed using SPI or
+	  DSI.
+
+config DRM_PANEL_SAMSUNG_S6E63M0_SPI
+	tristate "Samsung S6E63M0 RGB SPI interface"
+	depends on SPI
+	depends on DRM_PANEL_SAMSUNG_S6E63M0
+	default DRM_PANEL_SAMSUNG_S6E63M0
+	help
+	  Say Y here if you want to be able to access the Samsung
+	  S6E63M0 panel using SPI.
+
+config DRM_PANEL_SAMSUNG_S6E63M0_DSI
+	tristate "Samsung S6E63M0 RGB DSI interface"
+	depends on DRM_MIPI_DSI
+	depends on DRM_PANEL_SAMSUNG_S6E63M0
+	help
+	  Say Y here if you want to be able to access the Samsung
+	  S6E63M0 panel using DSI.
 
 config DRM_PANEL_SAMSUNG_S6E88A0_AMS452EF01
 	tristate "Samsung AMS452EF01 panel with S6E88A0 DSI video mode controller"
diff --git a/drivers/gpu/drm/panel/Makefile b/drivers/gpu/drm/panel/Makefile
index 15a4e7752951..2ba560bca61d 100644
--- a/drivers/gpu/drm/panel/Makefile
+++ b/drivers/gpu/drm/panel/Makefile
@@ -34,6 +34,8 @@ obj-$(CONFIG_DRM_PANEL_SAMSUNG_S6D16D0) += panel-samsung-s6d16d0.o
 obj-$(CONFIG_DRM_PANEL_SAMSUNG_S6E3HA2) += panel-samsung-s6e3ha2.o
 obj-$(CONFIG_DRM_PANEL_SAMSUNG_S6E63J0X03) += panel-samsung-s6e63j0x03.o
 obj-$(CONFIG_DRM_PANEL_SAMSUNG_S6E63M0) += panel-samsung-s6e63m0.o
+obj-$(CONFIG_DRM_PANEL_SAMSUNG_S6E63M0_SPI) += panel-samsung-s6e63m0-spi.o
+obj-$(CONFIG_DRM_PANEL_SAMSUNG_S6E63M0_DSI) += panel-samsung-s6e63m0-dsi.o
 obj-$(CONFIG_DRM_PANEL_SAMSUNG_S6E88A0_AMS452EF01) += panel-samsung-s6e88a0-ams452ef01.o
 obj-$(CONFIG_DRM_PANEL_SAMSUNG_S6E8AA0) += panel-samsung-s6e8aa0.o
 obj-$(CONFIG_DRM_PANEL_SEIKO_43WVF1G) += panel-seiko-43wvf1g.o
diff --git a/drivers/gpu/drm/panel/panel-samsung-s6e63m0-dsi.c b/drivers/gpu/drm/panel/panel-samsung-s6e63m0-dsi.c
new file mode 100644
index 000000000000..eec74c10ddda
--- /dev/null
+++ b/drivers/gpu/drm/panel/panel-samsung-s6e63m0-dsi.c
@@ -0,0 +1,139 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * DSI interface to the Samsung S6E63M0 panel.
+ * (C) 2019 Linus Walleij
+ */
+
+#include <linux/module.h>
+#include <linux/delay.h>
+#include <linux/of_device.h>
+
+#include <drm/drm_mipi_dsi.h>
+#include <drm/drm_print.h>
+
+#include "panel-samsung-s6e63m0.h"
+
+#define MCS_GLOBAL_PARAM	0xb0
+#define S6E63M0_DSI_MAX_CHUNK	15 /* CMD + 15 bytes max */
+
+static int s6e63m0_dsi_dcs_read(struct device *dev, const u8 cmd, u8 *data)
+{
+	struct mipi_dsi_device *dsi = to_mipi_dsi_device(dev);
+	int ret;
+
+	ret = mipi_dsi_dcs_read(dsi, cmd, data, 1);
+	if (ret < 0) {
+		dev_err(dev, "could not read DCS CMD %02x\n", cmd);
+		return ret;
+	}
+
+	dev_info(dev, "DSI read CMD %02x = %02x\n", cmd, *data);
+
+	return 0;
+}
+
+static int s6e63m0_dsi_dcs_write(struct device *dev, const u8 *data, size_t len)
+{
+	struct mipi_dsi_device *dsi = to_mipi_dsi_device(dev);
+	const u8 *seqp = data;
+	u8 cmd;
+	u8 cmdwritten;
+	int remain;
+	int chunk;
+	int ret;
+
+	dev_info(dev, "DSI writing dcs seq: %*ph\n", (int)len, data);
+
+	/* Pick out and skip past the DCS command */
+	cmd = *seqp;
+	seqp++;
+	cmdwritten = 0;
+	remain = len - 1;
+	chunk = remain;
+
+	/* Send max S6E63M0_DSI_MAX_CHUNK bytes at a time */
+	if (chunk > S6E63M0_DSI_MAX_CHUNK)
+		chunk = S6E63M0_DSI_MAX_CHUNK;
+	ret = mipi_dsi_dcs_write(dsi, cmd, seqp, chunk);
+	if (ret < 0) {
+		dev_err(dev, "error sending DCS command seq cmd %02x\n", cmd);
+		return ret;
+	}
+	cmdwritten += chunk;
+	seqp += chunk;
+
+	while (cmdwritten < remain) {
+		chunk = remain - cmdwritten;
+		if (chunk > S6E63M0_DSI_MAX_CHUNK)
+			chunk = S6E63M0_DSI_MAX_CHUNK;
+		ret = mipi_dsi_dcs_write(dsi, MCS_GLOBAL_PARAM, &cmdwritten, 1);
+		if (ret < 0) {
+			dev_err(dev, "error sending CMD %02x global param %02x\n",
+				cmd, cmdwritten);
+			return ret;
+		}
+		ret = mipi_dsi_dcs_write(dsi, cmd, seqp, chunk);
+		if (ret < 0) {
+			dev_err(dev, "error sending CMD %02x chunk\n", cmd);
+			return ret;
+		}
+		cmdwritten += chunk;
+		seqp += chunk;
+	}
+	dev_info(dev, "sent command %02x %02x bytes\n", cmd, cmdwritten);
+
+	usleep_range(8000, 9000);
+
+	return 0;
+}
+
+static int s6e63m0_dsi_probe(struct mipi_dsi_device *dsi)
+{
+	struct device *dev = &dsi->dev;
+	int ret;
+
+	dsi->lanes = 2;
+	dsi->format = MIPI_DSI_FMT_RGB888;
+	dsi->hs_rate = 349440000;
+	dsi->lp_rate = 9600000;
+	dsi->mode_flags = MIPI_DSI_MODE_VIDEO |
+		MIPI_DSI_MODE_EOT_PACKET |
+		MIPI_DSI_MODE_VIDEO_BURST;
+
+	ret = s6e63m0_probe(dev, s6e63m0_dsi_dcs_read, s6e63m0_dsi_dcs_write,
+			    true);
+	if (ret)
+		return ret;
+
+	ret = mipi_dsi_attach(dsi);
+	if (ret < 0)
+		s6e63m0_remove(dev);
+
+	return ret;
+}
+
+static int s6e63m0_dsi_remove(struct mipi_dsi_device *dsi)
+{
+	mipi_dsi_detach(dsi);
+	return s6e63m0_remove(&dsi->dev);
+}
+
+static const struct of_device_id s6e63m0_dsi_of_match[] = {
+	{ .compatible = "samsung,s6e63m0" },
+	{ /* sentinel */ }
+};
+MODULE_DEVICE_TABLE(of, s6e63m0_dsi_of_match);
+
+static struct mipi_dsi_driver s6e63m0_dsi_driver = {
+	.probe			= s6e63m0_dsi_probe,
+	.remove			= s6e63m0_dsi_remove,
+	.driver			= {
+		.name		= "panel-samsung-s6e63m0",
+		.of_match_table = s6e63m0_dsi_of_match,
+	},
+};
+module_mipi_dsi_driver(s6e63m0_dsi_driver);
+
+MODULE_AUTHOR("Linus Walleij <linusw@kernel.org>");
+MODULE_DESCRIPTION("s6e63m0 LCD DSI Driver");
+MODULE_LICENSE("GPL v2");
diff --git a/drivers/gpu/drm/panel/panel-samsung-s6e63m0-spi.c b/drivers/gpu/drm/panel/panel-samsung-s6e63m0-spi.c
new file mode 100644
index 000000000000..d298d780220d
--- /dev/null
+++ b/drivers/gpu/drm/panel/panel-samsung-s6e63m0-spi.c
@@ -0,0 +1,101 @@
+// SPDX-License-Identifier: GPL-2.0
+
+#include <linux/device.h>
+#include <linux/module.h>
+#include <linux/spi/spi.h>
+#include <linux/delay.h>
+
+#include <drm/drm_print.h>
+
+#include "panel-samsung-s6e63m0.h"
+
+#define DATA_MASK	0x100
+
+static int s6e63m0_spi_dcs_read(struct device *dev, const u8 cmd, u8 *data)
+{
+	/*
+	 * FIXME: implement reading DCS commands over SPI so we can
+	 * properly identify which physical panel is connected.
+	 */
+	*data = 0;
+
+	return 0;
+}
+
+static int s6e63m0_spi_write_word(struct device *dev, u16 data)
+{
+	struct spi_device *spi = to_spi_device(dev);
+	struct spi_transfer xfer = {
+		.len	= 2,
+		.tx_buf = &data,
+	};
+	struct spi_message msg;
+
+	spi_message_init(&msg);
+	spi_message_add_tail(&xfer, &msg);
+
+	return spi_sync(spi, &msg);
+}
+
+static int s6e63m0_spi_dcs_write(struct device *dev, const u8 *data, size_t len)
+{
+	int ret = 0;
+
+	dev_dbg(dev, "SPI writing dcs seq: %*ph\n", (int)len, data);
+	ret = s6e63m0_spi_write_word(dev, *data);
+
+	while (!ret && --len) {
+		++data;
+		ret = s6e63m0_spi_write_word(dev, *data | DATA_MASK);
+	}
+
+	if (ret) {
+		dev_err(dev, "SPI error %d writing dcs seq: %*ph\n", ret,
+			(int)len, data);
+	}
+
+	usleep_range(300, 310);
+
+	return ret;
+}
+
+static int s6e63m0_spi_probe(struct spi_device *spi)
+{
+	struct device *dev = &spi->dev;
+	int ret;
+
+	spi->bits_per_word = 9;
+	spi->mode = SPI_MODE_3;
+	ret = spi_setup(spi);
+	if (ret < 0) {
+		dev_err(dev, "spi setup failed.\n");
+		return ret;
+	}
+	return s6e63m0_probe(dev, s6e63m0_spi_dcs_read, s6e63m0_spi_dcs_write,
+			     false);
+}
+
+static int s6e63m0_spi_remove(struct spi_device *spi)
+{
+	return s6e63m0_remove(&spi->dev);
+}
+
+static const struct of_device_id s6e63m0_spi_of_match[] = {
+	{ .compatible = "samsung,s6e63m0" },
+	{ /* sentinel */ }
+};
+MODULE_DEVICE_TABLE(of, s6e63m0_spi_of_match);
+
+static struct spi_driver s6e63m0_spi_driver = {
+	.probe			= s6e63m0_spi_probe,
+	.remove			= s6e63m0_spi_remove,
+	.driver			= {
+		.name		= "panel-samsung-s6e63m0",
+		.of_match_table = s6e63m0_spi_of_match,
+	},
+};
+module_spi_driver(s6e63m0_spi_driver);
+
+MODULE_AUTHOR("Paweł Chmiel <pawel.mikolaj.chmiel@gmail.com>");
+MODULE_DESCRIPTION("s6e63m0 LCD SPI Driver");
+MODULE_LICENSE("GPL v2");
diff --git a/drivers/gpu/drm/panel/panel-samsung-s6e63m0.c b/drivers/gpu/drm/panel/panel-samsung-s6e63m0.c
index 2cc772fdc456..3eee67e2d86a 100644
--- a/drivers/gpu/drm/panel/panel-samsung-s6e63m0.c
+++ b/drivers/gpu/drm/panel/panel-samsung-s6e63m0.c
@@ -16,25 +16,34 @@
 #include <linux/gpio/consumer.h>
 #include <linux/module.h>
 #include <linux/regulator/consumer.h>
-#include <linux/spi/spi.h>
 
 #include <video/mipi_display.h>
 
+#include "panel-samsung-s6e63m0.h"
+
 /* Manufacturer Command Set */
 #define MCS_ELVSS_ON                0xb1
 #define MCS_MIECTL1                0xc0
 #define MCS_BCMODE                              0xc1
+#define MCS_ERROR_CHECK		0xd5
+#define MCS_READ_ID1		0xda
+#define MCS_READ_ID2		0xdb
+#define MCS_READ_ID3		0xdc
+#define MCS_LEVEL_2_KEY		0xf0
+#define MCS_MTP_KEY		0xf1
 #define MCS_DISCTL   0xf2
 #define MCS_SRCCTL           0xf6
 #define MCS_IFCTL                       0xf7
 #define MCS_PANELCTL         0xF8
 #define MCS_PGAMMACTL                   0xfa
 
+#define S6E63M0_LCD_ID_VALUE_M2		0xA4
+#define S6E63M0_LCD_ID_VALUE_SM2	0xB4
+#define S6E63M0_LCD_ID_VALUE_SM2_1	0xB6
+
 #define NUM_GAMMA_LEVELS             11
 #define GAMMA_TABLE_COUNT           23
 
-#define DATA_MASK                                       0x100
-
 #define MAX_BRIGHTNESS              (NUM_GAMMA_LEVELS - 1)
 
 /* array of gamma tables for gamma value 2.2 */
@@ -87,8 +96,11 @@ static u8 const s6e63m0_gamma_22[NUM_GAMMA_LEVELS][GAMMA_TABLE_COUNT] = {
 
 struct s6e63m0 {
 	struct device *dev;
+	int (*dcs_read)(struct device *dev, const u8 cmd, u8 *val);
+	int (*dcs_write)(struct device *dev, const u8 *data, size_t len);
 	struct drm_panel panel;
 	struct backlight_device *bl_dev;
+	u8 lcd_type;
 
 	struct regulator_bulk_data supplies[2];
 	struct gpio_desc *reset_gpio;
@@ -134,42 +146,20 @@ static int s6e63m0_clear_error(struct s6e63m0 *ctx)
 	return ret;
 }
 
-static int s6e63m0_spi_write_word(struct s6e63m0 *ctx, u16 data)
+static void s6e63m0_dcs_read(struct s6e63m0 *ctx, const u8 cmd, u8 *data)
 {
-	struct spi_device *spi = to_spi_device(ctx->dev);
-	struct spi_transfer xfer = {
-		.len	= 2,
-		.tx_buf = &data,
-	};
-	struct spi_message msg;
-
-	spi_message_init(&msg);
-	spi_message_add_tail(&xfer, &msg);
+	if (ctx->error < 0)
+		return;
 
-	return spi_sync(spi, &msg);
+	ctx->error = ctx->dcs_read(ctx->dev, cmd, data);
 }
 
 static void s6e63m0_dcs_write(struct s6e63m0 *ctx, const u8 *data, size_t len)
 {
-	int ret = 0;
-
 	if (ctx->error < 0 || len == 0)
 		return;
 
-	dev_dbg(ctx->dev, "writing dcs seq: %*ph\n", (int)len, data);
-	ret = s6e63m0_spi_write_word(ctx, *data);
-
-	while (!ret && --len) {
-		++data;
-		ret = s6e63m0_spi_write_word(ctx, *data | DATA_MASK);
-	}
-
-	if (ret) {
-		dev_err(ctx->dev, "error %d writing dcs seq: %*ph\n", ret, (int)len, data);
-		ctx->error = ret;
-	}
-
-	usleep_range(300, 310);
+	ctx->error = ctx->dcs_write(ctx->dev, data, len);
 }
 
 #define s6e63m0_dcs_write_seq_static(ctx, seq ...) \
@@ -178,6 +168,43 @@ static void s6e63m0_dcs_write(struct s6e63m0 *ctx, const u8 *data, size_t len)
 		s6e63m0_dcs_write(ctx, d, ARRAY_SIZE(d)); \
 	})
 
+static int s6e63m0_check_lcd_type(struct s6e63m0 *ctx)
+{
+	u8 id1, id2, id3;
+	int ret;
+
+	s6e63m0_dcs_read(ctx, MCS_READ_ID1, &id1);
+	s6e63m0_dcs_read(ctx, MCS_READ_ID2, &id2);
+	s6e63m0_dcs_read(ctx, MCS_READ_ID3, &id3);
+
+	ret = s6e63m0_clear_error(ctx);
+	if (ret) {
+		dev_err(ctx->dev, "error checking LCD type (%d)\n", ret);
+		ctx->lcd_type = 0x00;
+		return ret;
+	}
+
+	dev_info(ctx->dev, "MTP ID: %02x %02x %02x\n", id1, id2, id3);
+
+	/* We attempt to detect what panel is mounted on the controller */
+	switch (id2) {
+	case S6E63M0_LCD_ID_VALUE_M2:
+		dev_info(ctx->dev, "detected LCD panel AMS397GE MIPI M2\n");
+		break;
+	case S6E63M0_LCD_ID_VALUE_SM2:
+	case S6E63M0_LCD_ID_VALUE_SM2_1:
+		dev_info(ctx->dev, "detected LCD panel AMS397GE MIPI SM2\n");
+		break;
+	default:
+		dev_info(ctx->dev, "unknown LCD panel type %02x\n", id2);
+		break;
+	}
+
+	ctx->lcd_type = id2;
+
+	return 0;
+}
+
 static void s6e63m0_init(struct s6e63m0 *ctx)
 {
 	s6e63m0_dcs_write_seq_static(ctx, MCS_PANELCTL,
@@ -249,8 +276,6 @@ static void s6e63m0_init(struct s6e63m0 *ctx)
 
 	s6e63m0_dcs_write_seq_static(ctx, MCS_ELVSS_ON,
 				     0x0b);
-
-	s6e63m0_dcs_write_seq_static(ctx, MIPI_DCS_EXIT_SLEEP_MODE);
 }
 
 static int s6e63m0_power_on(struct s6e63m0 *ctx)
@@ -263,6 +288,9 @@ static int s6e63m0_power_on(struct s6e63m0 *ctx)
 
 	msleep(25);
 
+	/* Be sure to send a reset pulse */
+	gpiod_set_value(ctx->reset_gpio, 1);
+	msleep(5);
 	gpiod_set_value(ctx->reset_gpio, 0);
 	msleep(120);
 
@@ -292,8 +320,10 @@ static int s6e63m0_disable(struct drm_panel *panel)
 
 	backlight_disable(ctx->bl_dev);
 
+	s6e63m0_dcs_write_seq_static(ctx, MIPI_DCS_SET_DISPLAY_OFF);
+	msleep(10);
 	s6e63m0_dcs_write_seq_static(ctx, MIPI_DCS_ENTER_SLEEP_MODE);
-	msleep(200);
+	msleep(120);
 
 	ctx->enabled = false;
 
@@ -331,6 +361,15 @@ static int s6e63m0_prepare(struct drm_panel *panel)
 	if (ret < 0)
 		return ret;
 
+	/* Magic to unlock level 2 control of the display */
+	s6e63m0_dcs_write_seq_static(ctx, MCS_LEVEL_2_KEY, 0x5a, 0x5a);
+	/* Magic to unlock MTP reading */
+	s6e63m0_dcs_write_seq_static(ctx, MCS_MTP_KEY, 0x5a, 0x5a);
+
+	ret = s6e63m0_check_lcd_type(ctx);
+	if (ret < 0)
+		return ret;
+
 	s6e63m0_init(ctx);
 
 	ret = s6e63m0_clear_error(ctx);
@@ -350,7 +389,15 @@ static int s6e63m0_enable(struct drm_panel *panel)
 	if (ctx->enabled)
 		return 0;
 
+	s6e63m0_dcs_write_seq_static(ctx, MIPI_DCS_EXIT_SLEEP_MODE);
+	msleep(120);
 	s6e63m0_dcs_write_seq_static(ctx, MIPI_DCS_SET_DISPLAY_ON);
+	msleep(10);
+
+	s6e63m0_dcs_write_seq_static(ctx, MCS_ERROR_CHECK,
+				     0xE7, 0x14, 0x60, 0x17, 0x0A, 0x49, 0xC3,
+				     0x8F, 0x19, 0x64, 0x91, 0x84, 0x76, 0x20,
+				     0x0F, 0x00);
 
 	backlight_enable(ctx->bl_dev);
 
@@ -429,9 +476,11 @@ static int s6e63m0_backlight_register(struct s6e63m0 *ctx)
 	return ret;
 }
 
-static int s6e63m0_probe(struct spi_device *spi)
+int s6e63m0_probe(struct device *dev,
+		  int (*dcs_read)(struct device *dev, const u8 cmd, u8 *val),
+		  int (*dcs_write)(struct device *dev, const u8 *data, size_t len),
+		  bool dsi_mode)
 {
-	struct device *dev = &spi->dev;
 	struct s6e63m0 *ctx;
 	int ret;
 
@@ -439,7 +488,9 @@ static int s6e63m0_probe(struct spi_device *spi)
 	if (!ctx)
 		return -ENOMEM;
 
-	spi_set_drvdata(spi, ctx);
+	ctx->dcs_read = dcs_read;
+	ctx->dcs_write = dcs_write;
+	dev_set_drvdata(dev, ctx);
 
 	ctx->dev = dev;
 	ctx->enabled = false;
@@ -460,15 +511,8 @@ static int s6e63m0_probe(struct spi_device *spi)
 		return PTR_ERR(ctx->reset_gpio);
 	}
 
-	spi->bits_per_word = 9;
-	spi->mode = SPI_MODE_3;
-	ret = spi_setup(spi);
-	if (ret < 0) {
-		dev_err(dev, "spi setup failed.\n");
-		return ret;
-	}
-
 	drm_panel_init(&ctx->panel, dev, &s6e63m0_drm_funcs,
+		       dsi_mode ? DRM_MODE_CONNECTOR_DSI :
 		       DRM_MODE_CONNECTOR_DPI);
 
 	ret = s6e63m0_backlight_register(ctx);
@@ -479,31 +523,17 @@ static int s6e63m0_probe(struct spi_device *spi)
 
 	return 0;
 }
+EXPORT_SYMBOL_GPL(s6e63m0_probe);
 
-static int s6e63m0_remove(struct spi_device *spi)
+int s6e63m0_remove(struct device *dev)
 {
-	struct s6e63m0 *ctx = spi_get_drvdata(spi);
+	struct s6e63m0 *ctx = dev_get_drvdata(dev);
 
 	drm_panel_remove(&ctx->panel);
 
 	return 0;
 }
-
-static const struct of_device_id s6e63m0_of_match[] = {
-	{ .compatible = "samsung,s6e63m0" },
-	{ /* sentinel */ }
-};
-MODULE_DEVICE_TABLE(of, s6e63m0_of_match);
-
-static struct spi_driver s6e63m0_driver = {
-	.probe			= s6e63m0_probe,
-	.remove			= s6e63m0_remove,
-	.driver			= {
-		.name		= "panel-samsung-s6e63m0",
-		.of_match_table = s6e63m0_of_match,
-	},
-};
-module_spi_driver(s6e63m0_driver);
+EXPORT_SYMBOL_GPL(s6e63m0_remove);
 
 MODULE_AUTHOR("Paweł Chmiel <pawel.mikolaj.chmiel@gmail.com>");
 MODULE_DESCRIPTION("s6e63m0 LCD Driver");
diff --git a/drivers/gpu/drm/panel/panel-samsung-s6e63m0.h b/drivers/gpu/drm/panel/panel-samsung-s6e63m0.h
new file mode 100644
index 000000000000..c669fec91763
--- /dev/null
+++ b/drivers/gpu/drm/panel/panel-samsung-s6e63m0.h
@@ -0,0 +1,13 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+
+#ifndef _PANEL_SAMSUNG_S6E63M0_H
+#define _PANEL_SAMSUNG_S6E63M0_H
+
+int s6e63m0_probe(struct device *dev,
+		  int (*dcs_read)(struct device *dev, const u8 cmd, u8 *val),
+		  int (*dcs_write)(struct device *dev, const u8 *data,
+				   size_t len),
+		  bool dsi_mode);
+int s6e63m0_remove(struct device *dev);
+
+#endif /* _PANEL_SAMSUNG_S6E63M0_H */
diff --git a/drivers/gpu/drm/panfrost/panfrost_device.h b/drivers/gpu/drm/panfrost/panfrost_device.h
index 953f7536a773..2e9cbd1c4a58 100644
--- a/drivers/gpu/drm/panfrost/panfrost_device.h
+++ b/drivers/gpu/drm/panfrost/panfrost_device.h
@@ -70,6 +70,9 @@ struct panfrost_compatible {
 	int num_pm_domains;
 	/* Only required if num_pm_domains > 1. */
 	const char * const *pm_domain_names;
+
+	/* Vendor implementation quirks callback */
+	void (*vendor_quirk)(struct panfrost_device *pfdev);
 };
 
 struct panfrost_device {
diff --git a/drivers/gpu/drm/panfrost/panfrost_drv.c b/drivers/gpu/drm/panfrost/panfrost_drv.c
index 36463c89e966..37d4cb7a5491 100644
--- a/drivers/gpu/drm/panfrost/panfrost_drv.c
+++ b/drivers/gpu/drm/panfrost/panfrost_drv.c
@@ -656,7 +656,18 @@ static const struct panfrost_compatible default_data = {
 	.pm_domain_names = NULL,
 };
 
+static const struct panfrost_compatible amlogic_data = {
+	.num_supplies = ARRAY_SIZE(default_supplies),
+	.supply_names = default_supplies,
+	.vendor_quirk = panfrost_gpu_amlogic_quirk,
+};
+
 static const struct of_device_id dt_match[] = {
+	/* Set first to probe before the generic compatibles */
+	{ .compatible = "amlogic,meson-gxm-mali",
+	  .data = &amlogic_data, },
+	{ .compatible = "amlogic,meson-g12a-mali",
+	  .data = &amlogic_data, },
 	{ .compatible = "arm,mali-t604", .data = &default_data, },
 	{ .compatible = "arm,mali-t624", .data = &default_data, },
 	{ .compatible = "arm,mali-t628", .data = &default_data, },
diff --git a/drivers/gpu/drm/panfrost/panfrost_gpu.c b/drivers/gpu/drm/panfrost/panfrost_gpu.c
index e0f190e43813..e1b2a3376624 100644
--- a/drivers/gpu/drm/panfrost/panfrost_gpu.c
+++ b/drivers/gpu/drm/panfrost/panfrost_gpu.c
@@ -76,6 +76,17 @@ int panfrost_gpu_soft_reset(struct panfrost_device *pfdev)
 	return 0;
 }
 
+void panfrost_gpu_amlogic_quirk(struct panfrost_device *pfdev)
+{
+	/*
+	 * The Amlogic integrated Mali-T820, Mali-G31 & Mali-G52 needs
+	 * these undocumented bits in GPU_PWR_OVERRIDE1 to be set in order
+	 * to operate correctly.
+	 */
+	gpu_write(pfdev, GPU_PWR_KEY, GPU_PWR_KEY_UNLOCK);
+	gpu_write(pfdev, GPU_PWR_OVERRIDE1, 0xfff | (0x20 << 16));
+}
+
 static void panfrost_gpu_init_quirks(struct panfrost_device *pfdev)
 {
 	u32 quirks = 0;
@@ -136,6 +147,10 @@ static void panfrost_gpu_init_quirks(struct panfrost_device *pfdev)
 
 	if (quirks)
 		gpu_write(pfdev, GPU_JM_CONFIG, quirks);
+
+	/* Here goes platform specific quirks */
+	if (pfdev->comp->vendor_quirk)
+		pfdev->comp->vendor_quirk(pfdev);
 }
 
 #define MAX_HW_REVS 6
@@ -305,6 +320,8 @@ void panfrost_gpu_power_on(struct panfrost_device *pfdev)
 	int ret;
 	u32 val;
 
+	panfrost_gpu_init_quirks(pfdev);
+
 	/* Just turn on everything for now */
 	gpu_write(pfdev, L2_PWRON_LO, pfdev->features.l2_present);
 	ret = readl_relaxed_poll_timeout(pfdev->iomem + L2_READY_LO,
@@ -344,6 +361,7 @@ int panfrost_gpu_init(struct panfrost_device *pfdev)
 
 	dma_set_mask_and_coherent(pfdev->dev,
 		DMA_BIT_MASK(FIELD_GET(0xff00, pfdev->features.mmu_features)));
+	dma_set_max_seg_size(pfdev->dev, UINT_MAX);
 
 	irq = platform_get_irq_byname(to_platform_device(pfdev->dev), "gpu");
 	if (irq <= 0)
@@ -356,7 +374,6 @@ int panfrost_gpu_init(struct panfrost_device *pfdev)
 		return err;
 	}
 
-	panfrost_gpu_init_quirks(pfdev);
 	panfrost_gpu_power_on(pfdev);
 
 	return 0;
diff --git a/drivers/gpu/drm/panfrost/panfrost_gpu.h b/drivers/gpu/drm/panfrost/panfrost_gpu.h
index 4112412087b2..468c51e7e46d 100644
--- a/drivers/gpu/drm/panfrost/panfrost_gpu.h
+++ b/drivers/gpu/drm/panfrost/panfrost_gpu.h
@@ -16,4 +16,6 @@ int panfrost_gpu_soft_reset(struct panfrost_device *pfdev);
 void panfrost_gpu_power_on(struct panfrost_device *pfdev);
 void panfrost_gpu_power_off(struct panfrost_device *pfdev);
 
+void panfrost_gpu_amlogic_quirk(struct panfrost_device *pfdev);
+
 #endif
diff --git a/drivers/gpu/drm/panfrost/panfrost_regs.h b/drivers/gpu/drm/panfrost/panfrost_regs.h
index ea38ac60581c..eddaa62ad8b0 100644
--- a/drivers/gpu/drm/panfrost/panfrost_regs.h
+++ b/drivers/gpu/drm/panfrost/panfrost_regs.h
@@ -51,6 +51,10 @@
 #define GPU_STATUS			0x34
 #define   GPU_STATUS_PRFCNT_ACTIVE	BIT(2)
 #define GPU_LATEST_FLUSH_ID		0x38
+#define GPU_PWR_KEY			0x50	/* (WO) Power manager key register */
+#define  GPU_PWR_KEY_UNLOCK		0x2968A819
+#define GPU_PWR_OVERRIDE0		0x54	/* (RW) Power manager override settings */
+#define GPU_PWR_OVERRIDE1		0x58	/* (RW) Power manager override settings */
 #define GPU_FAULT_STATUS		0x3C
 #define GPU_FAULT_ADDRESS_LO		0x40
 #define GPU_FAULT_ADDRESS_HI		0x44
diff --git a/drivers/gpu/drm/qxl/qxl_object.c b/drivers/gpu/drm/qxl/qxl_object.c
index f838b6d689aa..2bc364412e8b 100644
--- a/drivers/gpu/drm/qxl/qxl_object.c
+++ b/drivers/gpu/drm/qxl/qxl_object.c
@@ -64,16 +64,24 @@ void qxl_ttm_placement_from_domain(struct qxl_bo *qbo, u32 domain, bool pinned)
 
 	qbo->placement.placement = qbo->placements;
 	qbo->placement.busy_placement = qbo->placements;
-	if (domain == QXL_GEM_DOMAIN_VRAM)
-		qbo->placements[c++].flags = TTM_PL_FLAG_CACHED | TTM_PL_FLAG_VRAM | pflag;
+	if (domain == QXL_GEM_DOMAIN_VRAM) {
+		qbo->placements[c].mem_type = TTM_PL_VRAM;
+		qbo->placements[c++].flags = TTM_PL_FLAG_CACHED | pflag;
+	}
 	if (domain == QXL_GEM_DOMAIN_SURFACE) {
-		qbo->placements[c++].flags = TTM_PL_FLAG_CACHED | TTM_PL_FLAG_PRIV | pflag;
-		qbo->placements[c++].flags = TTM_PL_FLAG_CACHED | TTM_PL_FLAG_VRAM | pflag;
+		qbo->placements[c].mem_type = TTM_PL_PRIV;
+		qbo->placements[c++].flags = TTM_PL_FLAG_CACHED | pflag;
+		qbo->placements[c].mem_type = TTM_PL_VRAM;
+		qbo->placements[c++].flags = TTM_PL_FLAG_CACHED | pflag;
+	}
+	if (domain == QXL_GEM_DOMAIN_CPU) {
+		qbo->placements[c].mem_type = TTM_PL_SYSTEM;
+		qbo->placements[c++].flags = TTM_PL_MASK_CACHING | pflag;
+	}
+	if (!c) {
+		qbo->placements[c].mem_type = TTM_PL_SYSTEM;
+		qbo->placements[c++].flags = TTM_PL_MASK_CACHING;
 	}
-	if (domain == QXL_GEM_DOMAIN_CPU)
-		qbo->placements[c++].flags = TTM_PL_MASK_CACHING | TTM_PL_FLAG_SYSTEM | pflag;
-	if (!c)
-		qbo->placements[c++].flags = TTM_PL_MASK_CACHING | TTM_PL_FLAG_SYSTEM;
 	qbo->placement.num_placement = c;
 	qbo->placement.num_busy_placement = c;
 	for (i = 0; i < c; ++i) {
@@ -167,6 +175,7 @@ int qxl_bo_kmap(struct qxl_bo *bo, void **ptr)
 void *qxl_bo_kmap_atomic_page(struct qxl_device *qdev,
 			      struct qxl_bo *bo, int page_offset)
 {
+	unsigned long offset;
 	void *rptr;
 	int ret;
 	struct io_mapping *map;
@@ -178,9 +187,8 @@ void *qxl_bo_kmap_atomic_page(struct qxl_device *qdev,
 	else
 		goto fallback;
 
-	ret = qxl_ttm_io_mem_reserve(bo->tbo.bdev, &bo->tbo.mem);
-
-	return io_mapping_map_atomic_wc(map, bo->tbo.mem.bus.offset + page_offset);
+	offset = bo->tbo.mem.start << PAGE_SHIFT;
+	return io_mapping_map_atomic_wc(map, offset + page_offset);
 fallback:
 	if (bo->kptr) {
 		rptr = bo->kptr + (page_offset * PAGE_SIZE);
diff --git a/drivers/gpu/drm/qxl/qxl_ttm.c b/drivers/gpu/drm/qxl/qxl_ttm.c
index 7aae0a96f043..fd691fff8394 100644
--- a/drivers/gpu/drm/qxl/qxl_ttm.c
+++ b/drivers/gpu/drm/qxl/qxl_ttm.c
@@ -55,7 +55,8 @@ static void qxl_evict_flags(struct ttm_buffer_object *bo,
 	static const struct ttm_place placements = {
 		.fpfn = 0,
 		.lpfn = 0,
-		.flags = TTM_PL_MASK_CACHING | TTM_PL_FLAG_SYSTEM
+		.mem_type = TTM_PL_SYSTEM,
+		.flags = TTM_PL_MASK_CACHING
 	};
 
 	if (!qxl_ttm_bo_is_qxl_bo(bo)) {
@@ -81,13 +82,12 @@ int qxl_ttm_io_mem_reserve(struct ttm_bo_device *bdev,
 		return 0;
 	case TTM_PL_VRAM:
 		mem->bus.is_iomem = true;
-		mem->bus.base = qdev->vram_base;
-		mem->bus.offset = mem->start << PAGE_SHIFT;
+		mem->bus.offset = (mem->start << PAGE_SHIFT) + qdev->vram_base;
 		break;
 	case TTM_PL_PRIV:
 		mem->bus.is_iomem = true;
-		mem->bus.base = qdev->surfaceram_base;
-		mem->bus.offset = mem->start << PAGE_SHIFT;
+		mem->bus.offset = (mem->start << PAGE_SHIFT) +
+			qdev->surfaceram_base;
 		break;
 	default:
 		return -EINVAL;
@@ -104,7 +104,8 @@ struct qxl_ttm_tt {
 	u64				offset;
 };
 
-static int qxl_ttm_backend_bind(struct ttm_tt *ttm,
+static int qxl_ttm_backend_bind(struct ttm_bo_device *bdev,
+				struct ttm_tt *ttm,
 				struct ttm_resource *bo_mem)
 {
 	struct qxl_ttm_tt *gtt = (void *)ttm;
@@ -118,25 +119,22 @@ static int qxl_ttm_backend_bind(struct ttm_tt *ttm,
 	return -1;
 }
 
-static void qxl_ttm_backend_unbind(struct ttm_tt *ttm)
+static void qxl_ttm_backend_unbind(struct ttm_bo_device *bdev,
+				   struct ttm_tt *ttm)
 {
 	/* Not implemented */
 }
 
-static void qxl_ttm_backend_destroy(struct ttm_tt *ttm)
+static void qxl_ttm_backend_destroy(struct ttm_bo_device *bdev,
+				    struct ttm_tt *ttm)
 {
 	struct qxl_ttm_tt *gtt = (void *)ttm;
 
+	ttm_tt_destroy_common(bdev, ttm);
 	ttm_tt_fini(&gtt->ttm);
 	kfree(gtt);
 }
 
-static struct ttm_backend_func qxl_backend_func = {
-	.bind = &qxl_ttm_backend_bind,
-	.unbind = &qxl_ttm_backend_unbind,
-	.destroy = &qxl_ttm_backend_destroy,
-};
-
 static struct ttm_tt *qxl_ttm_tt_create(struct ttm_buffer_object *bo,
 					uint32_t page_flags)
 {
@@ -147,7 +145,6 @@ static struct ttm_tt *qxl_ttm_tt_create(struct ttm_buffer_object *bo,
 	gtt = kzalloc(sizeof(struct qxl_ttm_tt), GFP_KERNEL);
 	if (gtt == NULL)
 		return NULL;
-	gtt->ttm.func = &qxl_backend_func;
 	gtt->qdev = qdev;
 	if (ttm_tt_init(&gtt->ttm, bo, page_flags)) {
 		kfree(gtt);
@@ -156,16 +153,6 @@ static struct ttm_tt *qxl_ttm_tt_create(struct ttm_buffer_object *bo,
 	return &gtt->ttm;
 }
 
-static void qxl_move_null(struct ttm_buffer_object *bo,
-			     struct ttm_resource *new_mem)
-{
-	struct ttm_resource *old_mem = &bo->mem;
-
-	BUG_ON(old_mem->mm_node != NULL);
-	*old_mem = *new_mem;
-	new_mem->mm_node = NULL;
-}
-
 static int qxl_bo_move(struct ttm_buffer_object *bo, bool evict,
 		       struct ttm_operation_ctx *ctx,
 		       struct ttm_resource *new_mem)
@@ -178,7 +165,7 @@ static int qxl_bo_move(struct ttm_buffer_object *bo, bool evict,
 		return ret;
 
 	if (old_mem->mem_type == TTM_PL_SYSTEM && bo->ttm == NULL) {
-		qxl_move_null(bo, new_mem);
+		ttm_bo_move_null(bo, new_mem);
 		return 0;
 	}
 	return ttm_bo_move_memcpy(bo, ctx, new_mem);
@@ -202,6 +189,9 @@ static void qxl_bo_move_notify(struct ttm_buffer_object *bo,
 
 static struct ttm_bo_driver qxl_bo_driver = {
 	.ttm_tt_create = &qxl_ttm_tt_create,
+	.ttm_tt_bind = &qxl_ttm_backend_bind,
+	.ttm_tt_destroy = &qxl_ttm_backend_destroy,
+	.ttm_tt_unbind = &qxl_ttm_backend_unbind,
 	.eviction_valuable = ttm_bo_eviction_valuable,
 	.evict_flags = &qxl_evict_flags,
 	.move = &qxl_bo_move,
@@ -213,8 +203,7 @@ static int qxl_ttm_init_mem_type(struct qxl_device *qdev,
 				 unsigned int type,
 				 uint64_t size)
 {
-	return ttm_range_man_init(&qdev->mman.bdev, type, TTM_PL_MASK_CACHING,
-				  TTM_PL_FLAG_CACHED, false, size);
+	return ttm_range_man_init(&qdev->mman.bdev, type, false, size);
 }
 
 int qxl_ttm_init(struct qxl_device *qdev)
diff --git a/drivers/gpu/drm/radeon/radeon.h b/drivers/gpu/drm/radeon/radeon.h
index cc4f58d16589..a6d8de01194a 100644
--- a/drivers/gpu/drm/radeon/radeon.h
+++ b/drivers/gpu/drm/radeon/radeon.h
@@ -2815,10 +2815,12 @@ extern void radeon_legacy_set_clock_gating(struct radeon_device *rdev, int enabl
 extern void radeon_atom_set_clock_gating(struct radeon_device *rdev, int enable);
 extern void radeon_ttm_placement_from_domain(struct radeon_bo *rbo, u32 domain);
 extern bool radeon_ttm_bo_is_radeon_bo(struct ttm_buffer_object *bo);
-extern int radeon_ttm_tt_set_userptr(struct ttm_tt *ttm, uint64_t addr,
+extern int radeon_ttm_tt_set_userptr(struct radeon_device *rdev,
+				     struct ttm_tt *ttm, uint64_t addr,
 				     uint32_t flags);
-extern bool radeon_ttm_tt_has_userptr(struct ttm_tt *ttm);
-extern bool radeon_ttm_tt_is_readonly(struct ttm_tt *ttm);
+extern bool radeon_ttm_tt_has_userptr(struct radeon_device *rdev, struct ttm_tt *ttm);
+extern bool radeon_ttm_tt_is_readonly(struct radeon_device *rdev, struct ttm_tt *ttm);
+bool radeon_ttm_tt_is_bound(struct ttm_bo_device *bdev, struct ttm_tt *ttm);
 extern void radeon_vram_location(struct radeon_device *rdev, struct radeon_mc *mc, u64 base);
 extern void radeon_gtt_location(struct radeon_device *rdev, struct radeon_mc *mc);
 extern int radeon_resume_kms(struct drm_device *dev, bool resume, bool fbcon);
diff --git a/drivers/gpu/drm/radeon/radeon_cs.c b/drivers/gpu/drm/radeon/radeon_cs.c
index 33ae1b883268..21ce2f9502c0 100644
--- a/drivers/gpu/drm/radeon/radeon_cs.c
+++ b/drivers/gpu/drm/radeon/radeon_cs.c
@@ -160,7 +160,7 @@ static int radeon_cs_parser_relocs(struct radeon_cs_parser *p)
 			p->relocs[i].allowed_domains = domain;
 		}
 
-		if (radeon_ttm_tt_has_userptr(p->relocs[i].robj->tbo.ttm)) {
+		if (radeon_ttm_tt_has_userptr(p->rdev, p->relocs[i].robj->tbo.ttm)) {
 			uint32_t domain = p->relocs[i].preferred_domains;
 			if (!(domain & RADEON_GEM_DOMAIN_GTT)) {
 				DRM_ERROR("Only RADEON_GEM_DOMAIN_GTT is "
diff --git a/drivers/gpu/drm/radeon/radeon_gem.c b/drivers/gpu/drm/radeon/radeon_gem.c
index 7f5dfe04789e..e5c4271e64ed 100644
--- a/drivers/gpu/drm/radeon/radeon_gem.c
+++ b/drivers/gpu/drm/radeon/radeon_gem.c
@@ -331,7 +331,7 @@ int radeon_gem_userptr_ioctl(struct drm_device *dev, void *data,
 		goto handle_lockup;
 
 	bo = gem_to_radeon_bo(gobj);
-	r = radeon_ttm_tt_set_userptr(bo->tbo.ttm, args->addr, args->flags);
+	r = radeon_ttm_tt_set_userptr(rdev, bo->tbo.ttm, args->addr, args->flags);
 	if (r)
 		goto release_object;
 
@@ -420,7 +420,7 @@ int radeon_mode_dumb_mmap(struct drm_file *filp,
 		return -ENOENT;
 	}
 	robj = gem_to_radeon_bo(gobj);
-	if (radeon_ttm_tt_has_userptr(robj->tbo.ttm)) {
+	if (radeon_ttm_tt_has_userptr(robj->rdev, robj->tbo.ttm)) {
 		drm_gem_object_put(gobj);
 		return -EPERM;
 	}
@@ -721,7 +721,7 @@ int radeon_gem_op_ioctl(struct drm_device *dev, void *data,
 	robj = gem_to_radeon_bo(gobj);
 
 	r = -EPERM;
-	if (radeon_ttm_tt_has_userptr(robj->tbo.ttm))
+	if (radeon_ttm_tt_has_userptr(robj->rdev, robj->tbo.ttm))
 		goto out;
 
 	r = radeon_bo_reserve(robj, false);
diff --git a/drivers/gpu/drm/radeon/radeon_mn.c b/drivers/gpu/drm/radeon/radeon_mn.c
index f93829f08a4d..97b9b6dd6dd3 100644
--- a/drivers/gpu/drm/radeon/radeon_mn.c
+++ b/drivers/gpu/drm/radeon/radeon_mn.c
@@ -53,7 +53,7 @@ static bool radeon_mn_invalidate(struct mmu_interval_notifier *mn,
 	struct ttm_operation_ctx ctx = { false, false };
 	long r;
 
-	if (!bo->tbo.ttm || bo->tbo.ttm->state != tt_bound)
+	if (!bo->tbo.ttm || !radeon_ttm_tt_is_bound(bo->tbo.bdev, bo->tbo.ttm))
 		return true;
 
 	if (!mmu_notifier_range_blockable(range))
diff --git a/drivers/gpu/drm/radeon/radeon_object.c b/drivers/gpu/drm/radeon/radeon_object.c
index bb7582afd803..316e35d3f8a9 100644
--- a/drivers/gpu/drm/radeon/radeon_object.c
+++ b/drivers/gpu/drm/radeon/radeon_object.c
@@ -112,58 +112,58 @@ void radeon_ttm_placement_from_domain(struct radeon_bo *rbo, u32 domain)
 		    rbo->rdev->mc.visible_vram_size < rbo->rdev->mc.real_vram_size) {
 			rbo->placements[c].fpfn =
 				rbo->rdev->mc.visible_vram_size >> PAGE_SHIFT;
+			rbo->placements[c].mem_type = TTM_PL_VRAM;
 			rbo->placements[c++].flags = TTM_PL_FLAG_WC |
-						     TTM_PL_FLAG_UNCACHED |
-						     TTM_PL_FLAG_VRAM;
+						     TTM_PL_FLAG_UNCACHED;
 		}
 
 		rbo->placements[c].fpfn = 0;
+		rbo->placements[c].mem_type = TTM_PL_VRAM;
 		rbo->placements[c++].flags = TTM_PL_FLAG_WC |
-					     TTM_PL_FLAG_UNCACHED |
-					     TTM_PL_FLAG_VRAM;
+					     TTM_PL_FLAG_UNCACHED;
 	}
 
 	if (domain & RADEON_GEM_DOMAIN_GTT) {
 		if (rbo->flags & RADEON_GEM_GTT_UC) {
 			rbo->placements[c].fpfn = 0;
-			rbo->placements[c++].flags = TTM_PL_FLAG_UNCACHED |
-				TTM_PL_FLAG_TT;
+			rbo->placements[c].mem_type = TTM_PL_TT;
+			rbo->placements[c++].flags = TTM_PL_FLAG_UNCACHED;
 
 		} else if ((rbo->flags & RADEON_GEM_GTT_WC) ||
 			   (rbo->rdev->flags & RADEON_IS_AGP)) {
 			rbo->placements[c].fpfn = 0;
+			rbo->placements[c].mem_type = TTM_PL_TT;
 			rbo->placements[c++].flags = TTM_PL_FLAG_WC |
-				TTM_PL_FLAG_UNCACHED |
-				TTM_PL_FLAG_TT;
+				TTM_PL_FLAG_UNCACHED;
 		} else {
 			rbo->placements[c].fpfn = 0;
-			rbo->placements[c++].flags = TTM_PL_FLAG_CACHED |
-						     TTM_PL_FLAG_TT;
+			rbo->placements[c].mem_type = TTM_PL_TT;
+			rbo->placements[c++].flags = TTM_PL_FLAG_CACHED;
 		}
 	}
 
 	if (domain & RADEON_GEM_DOMAIN_CPU) {
 		if (rbo->flags & RADEON_GEM_GTT_UC) {
 			rbo->placements[c].fpfn = 0;
-			rbo->placements[c++].flags = TTM_PL_FLAG_UNCACHED |
-				TTM_PL_FLAG_SYSTEM;
+			rbo->placements[c].mem_type = TTM_PL_SYSTEM;
+			rbo->placements[c++].flags = TTM_PL_FLAG_UNCACHED;
 
 		} else if ((rbo->flags & RADEON_GEM_GTT_WC) ||
 		    rbo->rdev->flags & RADEON_IS_AGP) {
 			rbo->placements[c].fpfn = 0;
+			rbo->placements[c].mem_type = TTM_PL_SYSTEM;
 			rbo->placements[c++].flags = TTM_PL_FLAG_WC |
-				TTM_PL_FLAG_UNCACHED |
-				TTM_PL_FLAG_SYSTEM;
+				TTM_PL_FLAG_UNCACHED;
 		} else {
 			rbo->placements[c].fpfn = 0;
-			rbo->placements[c++].flags = TTM_PL_FLAG_CACHED |
-						     TTM_PL_FLAG_SYSTEM;
+			rbo->placements[c].mem_type = TTM_PL_SYSTEM;
+			rbo->placements[c++].flags = TTM_PL_FLAG_CACHED;
 		}
 	}
 	if (!c) {
 		rbo->placements[c].fpfn = 0;
-		rbo->placements[c++].flags = TTM_PL_MASK_CACHING |
-					     TTM_PL_FLAG_SYSTEM;
+		rbo->placements[c].mem_type = TTM_PL_SYSTEM;
+		rbo->placements[c++].flags = TTM_PL_MASK_CACHING;
 	}
 
 	rbo->placement.num_placement = c;
@@ -171,7 +171,7 @@ void radeon_ttm_placement_from_domain(struct radeon_bo *rbo, u32 domain)
 
 	for (i = 0; i < c; ++i) {
 		if ((rbo->flags & RADEON_GEM_CPU_ACCESS) &&
-		    (rbo->placements[i].flags & TTM_PL_FLAG_VRAM) &&
+		    (rbo->placements[i].mem_type == TTM_PL_VRAM) &&
 		    !rbo->placements[i].fpfn)
 			rbo->placements[i].lpfn =
 				rbo->rdev->mc.visible_vram_size >> PAGE_SHIFT;
@@ -331,7 +331,7 @@ int radeon_bo_pin_restricted(struct radeon_bo *bo, u32 domain, u64 max_offset,
 	struct ttm_operation_ctx ctx = { false, false };
 	int r, i;
 
-	if (radeon_ttm_tt_has_userptr(bo->tbo.ttm))
+	if (radeon_ttm_tt_has_userptr(bo->rdev, bo->tbo.ttm))
 		return -EPERM;
 
 	if (bo->pin_count) {
@@ -360,7 +360,7 @@ int radeon_bo_pin_restricted(struct radeon_bo *bo, u32 domain, u64 max_offset,
 	radeon_ttm_placement_from_domain(bo, domain);
 	for (i = 0; i < bo->placement.num_placement; i++) {
 		/* force to pin into visible video ram */
-		if ((bo->placements[i].flags & TTM_PL_FLAG_VRAM) &&
+		if ((bo->placements[i].mem_type == TTM_PL_VRAM) &&
 		    !(bo->flags & RADEON_GEM_NO_CPU_ACCESS) &&
 		    (!max_offset || max_offset > bo->rdev->mc.visible_vram_size))
 			bo->placements[i].lpfn =
@@ -824,7 +824,7 @@ int radeon_bo_fault_reserve_notify(struct ttm_buffer_object *bo)
 	lpfn =	rdev->mc.visible_vram_size >> PAGE_SHIFT;
 	for (i = 0; i < rbo->placement.num_placement; i++) {
 		/* Force into visible VRAM */
-		if ((rbo->placements[i].flags & TTM_PL_FLAG_VRAM) &&
+		if ((rbo->placements[i].mem_type == TTM_PL_VRAM) &&
 		    (!rbo->placements[i].lpfn || rbo->placements[i].lpfn > lpfn))
 			rbo->placements[i].lpfn = lpfn;
 	}
diff --git a/drivers/gpu/drm/radeon/radeon_prime.c b/drivers/gpu/drm/radeon/radeon_prime.c
index b906e8fbd5f3..b9de0e51c0be 100644
--- a/drivers/gpu/drm/radeon/radeon_prime.c
+++ b/drivers/gpu/drm/radeon/radeon_prime.c
@@ -36,7 +36,7 @@ struct sg_table *radeon_gem_prime_get_sg_table(struct drm_gem_object *obj)
 	struct radeon_bo *bo = gem_to_radeon_bo(obj);
 	int npages = bo->tbo.num_pages;
 
-	return drm_prime_pages_to_sg(bo->tbo.ttm->pages, npages);
+	return drm_prime_pages_to_sg(obj->dev, bo->tbo.ttm->pages, npages);
 }
 
 void *radeon_gem_prime_vmap(struct drm_gem_object *obj)
@@ -121,7 +121,7 @@ struct dma_buf *radeon_gem_prime_export(struct drm_gem_object *gobj,
 					int flags)
 {
 	struct radeon_bo *bo = gem_to_radeon_bo(gobj);
-	if (radeon_ttm_tt_has_userptr(bo->tbo.ttm))
+	if (radeon_ttm_tt_has_userptr(bo->rdev, bo->tbo.ttm))
 		return ERR_PTR(-EPERM);
 	return drm_gem_prime_export(gobj, flags);
 }
diff --git a/drivers/gpu/drm/radeon/radeon_ttm.c b/drivers/gpu/drm/radeon/radeon_ttm.c
index 74ad50c7491c..36150b7f31a9 100644
--- a/drivers/gpu/drm/radeon/radeon_ttm.c
+++ b/drivers/gpu/drm/radeon/radeon_ttm.c
@@ -56,6 +56,10 @@
 static int radeon_ttm_debugfs_init(struct radeon_device *rdev);
 static void radeon_ttm_debugfs_fini(struct radeon_device *rdev);
 
+static int radeon_ttm_tt_bind(struct ttm_bo_device *bdev,
+			      struct ttm_tt *ttm,
+			      struct ttm_resource *bo_mem);
+
 struct radeon_device *radeon_get_rdev(struct ttm_bo_device *bdev)
 {
 	struct radeon_mman *mman;
@@ -69,17 +73,13 @@ struct radeon_device *radeon_get_rdev(struct ttm_bo_device *bdev)
 static int radeon_ttm_init_vram(struct radeon_device *rdev)
 {
 	return ttm_range_man_init(&rdev->mman.bdev, TTM_PL_VRAM,
-				  TTM_PL_FLAG_UNCACHED | TTM_PL_FLAG_WC,
-				  TTM_PL_FLAG_WC, false,
-				  rdev->mc.real_vram_size >> PAGE_SHIFT);
+				  false, rdev->mc.real_vram_size >> PAGE_SHIFT);
 }
 
 static int radeon_ttm_init_gtt(struct radeon_device *rdev)
 {
 	return ttm_range_man_init(&rdev->mman.bdev, TTM_PL_TT,
-				  TTM_PL_MASK_CACHING,
-				  TTM_PL_FLAG_CACHED, true,
-				  rdev->mc.gtt_size >> PAGE_SHIFT);
+				  true, rdev->mc.gtt_size >> PAGE_SHIFT);
 }
 
 static void radeon_evict_flags(struct ttm_buffer_object *bo,
@@ -88,7 +88,8 @@ static void radeon_evict_flags(struct ttm_buffer_object *bo,
 	static const struct ttm_place placements = {
 		.fpfn = 0,
 		.lpfn = 0,
-		.flags = TTM_PL_MASK_CACHING | TTM_PL_FLAG_SYSTEM
+		.mem_type = TTM_PL_SYSTEM,
+		.flags = TTM_PL_MASK_CACHING
 	};
 
 	struct radeon_bo *rbo;
@@ -119,7 +120,7 @@ static void radeon_evict_flags(struct ttm_buffer_object *bo,
 							 RADEON_GEM_DOMAIN_GTT);
 			rbo->placement.num_busy_placement = 0;
 			for (i = 0; i < rbo->placement.num_placement; i++) {
-				if (rbo->placements[i].flags & TTM_PL_FLAG_VRAM) {
+				if (rbo->placements[i].mem_type == TTM_PL_VRAM) {
 					if (rbo->placements[i].fpfn < fpfn)
 						rbo->placements[i].fpfn = fpfn;
 				} else {
@@ -141,23 +142,14 @@ static void radeon_evict_flags(struct ttm_buffer_object *bo,
 static int radeon_verify_access(struct ttm_buffer_object *bo, struct file *filp)
 {
 	struct radeon_bo *rbo = container_of(bo, struct radeon_bo, tbo);
+	struct radeon_device *rdev = radeon_get_rdev(bo->bdev);
 
-	if (radeon_ttm_tt_has_userptr(bo->ttm))
+	if (radeon_ttm_tt_has_userptr(rdev, bo->ttm))
 		return -EPERM;
 	return drm_vma_node_verify_access(&rbo->tbo.base.vma_node,
 					  filp->private_data);
 }
 
-static void radeon_move_null(struct ttm_buffer_object *bo,
-			     struct ttm_resource *new_mem)
-{
-	struct ttm_resource *old_mem = &bo->mem;
-
-	BUG_ON(old_mem->mm_node != NULL);
-	*old_mem = *new_mem;
-	new_mem->mm_node = NULL;
-}
-
 static int radeon_move_blit(struct ttm_buffer_object *bo,
 			bool evict, bool no_wait_gpu,
 			struct ttm_resource *new_mem,
@@ -208,7 +200,7 @@ static int radeon_move_blit(struct ttm_buffer_object *bo,
 	if (IS_ERR(fence))
 		return PTR_ERR(fence);
 
-	r = ttm_bo_move_accel_cleanup(bo, &fence->base, evict, new_mem);
+	r = ttm_bo_move_accel_cleanup(bo, &fence->base, evict, false, new_mem);
 	radeon_fence_unref(&fence);
 	return r;
 }
@@ -233,7 +225,8 @@ static int radeon_move_vram_ram(struct ttm_buffer_object *bo,
 	placement.busy_placement = &placements;
 	placements.fpfn = 0;
 	placements.lpfn = 0;
-	placements.flags = TTM_PL_MASK_CACHING | TTM_PL_FLAG_TT;
+	placements.mem_type = TTM_PL_TT;
+	placements.flags = TTM_PL_MASK_CACHING;
 	r = ttm_bo_mem_space(bo, &placement, &tmp_mem, &ctx);
 	if (unlikely(r)) {
 		return r;
@@ -244,7 +237,12 @@ static int radeon_move_vram_ram(struct ttm_buffer_object *bo,
 		goto out_cleanup;
 	}
 
-	r = ttm_tt_bind(bo->ttm, &tmp_mem, &ctx);
+	r = ttm_tt_populate(bo->bdev, bo->ttm, &ctx);
+	if (unlikely(r)) {
+		goto out_cleanup;
+	}
+
+	r = radeon_ttm_tt_bind(bo->bdev, bo->ttm, &tmp_mem);
 	if (unlikely(r)) {
 		goto out_cleanup;
 	}
@@ -278,7 +276,8 @@ static int radeon_move_ram_vram(struct ttm_buffer_object *bo,
 	placement.busy_placement = &placements;
 	placements.fpfn = 0;
 	placements.lpfn = 0;
-	placements.flags = TTM_PL_MASK_CACHING | TTM_PL_FLAG_TT;
+	placements.mem_type = TTM_PL_TT;
+	placements.flags = TTM_PL_MASK_CACHING;
 	r = ttm_bo_mem_space(bo, &placement, &tmp_mem, &ctx);
 	if (unlikely(r)) {
 		return r;
@@ -316,7 +315,7 @@ static int radeon_bo_move(struct ttm_buffer_object *bo, bool evict,
 
 	rdev = radeon_get_rdev(bo->bdev);
 	if (old_mem->mem_type == TTM_PL_SYSTEM && bo->ttm == NULL) {
-		radeon_move_null(bo, new_mem);
+		ttm_bo_move_null(bo, new_mem);
 		return 0;
 	}
 	if ((old_mem->mem_type == TTM_PL_TT &&
@@ -324,7 +323,7 @@ static int radeon_bo_move(struct ttm_buffer_object *bo, bool evict,
 	    (old_mem->mem_type == TTM_PL_SYSTEM &&
 	     new_mem->mem_type == TTM_PL_TT)) {
 		/* bind is enough */
-		radeon_move_null(bo, new_mem);
+		ttm_bo_move_null(bo, new_mem);
 		return 0;
 	}
 	if (!rdev->ring[radeon_copy_ring_index(rdev)].ready ||
@@ -372,8 +371,8 @@ static int radeon_ttm_io_mem_reserve(struct ttm_bo_device *bdev, struct ttm_reso
 #if IS_ENABLED(CONFIG_AGP)
 		if (rdev->flags & RADEON_IS_AGP) {
 			/* RADEON_IS_AGP is set only if AGP is active */
-			mem->bus.offset = mem->start << PAGE_SHIFT;
-			mem->bus.base = rdev->mc.agp_base;
+			mem->bus.offset = (mem->start << PAGE_SHIFT) +
+				rdev->mc.agp_base;
 			mem->bus.is_iomem = !rdev->ddev->agp->cant_use_aperture;
 		}
 #endif
@@ -383,7 +382,7 @@ static int radeon_ttm_io_mem_reserve(struct ttm_bo_device *bdev, struct ttm_reso
 		/* check if it's visible */
 		if ((mem->bus.offset + bus_size) > rdev->mc.visible_vram_size)
 			return -EINVAL;
-		mem->bus.base = rdev->mc.aper_base;
+		mem->bus.offset += rdev->mc.aper_base;
 		mem->bus.is_iomem = true;
 #ifdef __alpha__
 		/*
@@ -392,12 +391,10 @@ static int radeon_ttm_io_mem_reserve(struct ttm_bo_device *bdev, struct ttm_reso
 		 */
 		if (mem->placement & TTM_PL_FLAG_WC)
 			mem->bus.addr =
-				ioremap_wc(mem->bus.base + mem->bus.offset,
-					   bus_size);
+				ioremap_wc(mem->bus.offset, bus_size);
 		else
 			mem->bus.addr =
-				ioremap(mem->bus.base + mem->bus.offset,
-					bus_size);
+				ioremap(mem->bus.offset, bus_size);
 		if (!mem->bus.addr)
 			return -ENOMEM;
 
@@ -407,7 +404,7 @@ static int radeon_ttm_io_mem_reserve(struct ttm_bo_device *bdev, struct ttm_reso
 		 * It then can be used to build PTEs for VRAM
 		 * access, as done in ttm_bo_vm_fault().
 		 */
-		mem->bus.base = (mem->bus.base & 0x0ffffffffUL) +
+		mem->bus.offset = (mem->bus.offset & 0x0ffffffffUL) +
 			rdev->ddev->hose->dense_mem_base;
 #endif
 		break;
@@ -427,12 +424,13 @@ struct radeon_ttm_tt {
 	uint64_t			userptr;
 	struct mm_struct		*usermm;
 	uint32_t			userflags;
+	bool bound;
 };
 
 /* prepare the sg table with the user pages */
-static int radeon_ttm_tt_pin_userptr(struct ttm_tt *ttm)
+static int radeon_ttm_tt_pin_userptr(struct ttm_bo_device *bdev, struct ttm_tt *ttm)
 {
-	struct radeon_device *rdev = radeon_get_rdev(ttm->bdev);
+	struct radeon_device *rdev = radeon_get_rdev(bdev);
 	struct radeon_ttm_tt *gtt = (void *)ttm;
 	unsigned pinned = 0;
 	int r;
@@ -491,9 +489,9 @@ release_pages:
 	return r;
 }
 
-static void radeon_ttm_tt_unpin_userptr(struct ttm_tt *ttm)
+static void radeon_ttm_tt_unpin_userptr(struct ttm_bo_device *bdev, struct ttm_tt *ttm)
 {
-	struct radeon_device *rdev = radeon_get_rdev(ttm->bdev);
+	struct radeon_device *rdev = radeon_get_rdev(bdev);
 	struct radeon_ttm_tt *gtt = (void *)ttm;
 	struct sg_page_iter sg_iter;
 
@@ -520,17 +518,28 @@ static void radeon_ttm_tt_unpin_userptr(struct ttm_tt *ttm)
 	sg_free_table(ttm->sg);
 }
 
-static int radeon_ttm_backend_bind(struct ttm_tt *ttm,
+static bool radeon_ttm_backend_is_bound(struct ttm_tt *ttm)
+{
+	struct radeon_ttm_tt *gtt = (void*)ttm;
+
+	return (gtt->bound);
+}
+
+static int radeon_ttm_backend_bind(struct ttm_bo_device *bdev,
+				   struct ttm_tt *ttm,
 				   struct ttm_resource *bo_mem)
 {
 	struct radeon_ttm_tt *gtt = (void*)ttm;
-	struct radeon_device *rdev = radeon_get_rdev(ttm->bdev);
+	struct radeon_device *rdev = radeon_get_rdev(bdev);
 	uint32_t flags = RADEON_GART_PAGE_VALID | RADEON_GART_PAGE_READ |
 		RADEON_GART_PAGE_WRITE;
 	int r;
 
+	if (gtt->bound)
+		return 0;
+
 	if (gtt->userptr) {
-		radeon_ttm_tt_pin_userptr(ttm);
+		radeon_ttm_tt_pin_userptr(bdev, ttm);
 		flags &= ~RADEON_GART_PAGE_WRITE;
 	}
 
@@ -548,34 +557,36 @@ static int radeon_ttm_backend_bind(struct ttm_tt *ttm,
 			  ttm->num_pages, (unsigned)gtt->offset);
 		return r;
 	}
+	gtt->bound = true;
 	return 0;
 }
 
-static void radeon_ttm_backend_unbind(struct ttm_tt *ttm)
+static void radeon_ttm_backend_unbind(struct ttm_bo_device *bdev, struct ttm_tt *ttm)
 {
 	struct radeon_ttm_tt *gtt = (void *)ttm;
-	struct radeon_device *rdev = radeon_get_rdev(ttm->bdev);
+	struct radeon_device *rdev = radeon_get_rdev(bdev);
+
+	if (!gtt->bound)
+		return;
 
 	radeon_gart_unbind(rdev, gtt->offset, ttm->num_pages);
 
 	if (gtt->userptr)
-		radeon_ttm_tt_unpin_userptr(ttm);
+		radeon_ttm_tt_unpin_userptr(bdev, ttm);
+	gtt->bound = false;
 }
 
-static void radeon_ttm_backend_destroy(struct ttm_tt *ttm)
+static void radeon_ttm_backend_destroy(struct ttm_bo_device *bdev, struct ttm_tt *ttm)
 {
 	struct radeon_ttm_tt *gtt = (void *)ttm;
 
+	radeon_ttm_backend_unbind(bdev, ttm);
+	ttm_tt_destroy_common(bdev, ttm);
+
 	ttm_dma_tt_fini(&gtt->ttm);
 	kfree(gtt);
 }
 
-static struct ttm_backend_func radeon_backend_func = {
-	.bind = &radeon_ttm_backend_bind,
-	.unbind = &radeon_ttm_backend_unbind,
-	.destroy = &radeon_ttm_backend_destroy,
-};
-
 static struct ttm_tt *radeon_ttm_tt_create(struct ttm_buffer_object *bo,
 					   uint32_t page_flags)
 {
@@ -594,7 +605,6 @@ static struct ttm_tt *radeon_ttm_tt_create(struct ttm_buffer_object *bo,
 	if (gtt == NULL) {
 		return NULL;
 	}
-	gtt->ttm.ttm.func = &radeon_backend_func;
 	if (ttm_dma_tt_init(&gtt->ttm, bo, page_flags)) {
 		kfree(gtt);
 		return NULL;
@@ -602,18 +612,25 @@ static struct ttm_tt *radeon_ttm_tt_create(struct ttm_buffer_object *bo,
 	return &gtt->ttm.ttm;
 }
 
-static struct radeon_ttm_tt *radeon_ttm_tt_to_gtt(struct ttm_tt *ttm)
+static struct radeon_ttm_tt *radeon_ttm_tt_to_gtt(struct radeon_device *rdev,
+						  struct ttm_tt *ttm)
 {
-	if (!ttm || ttm->func != &radeon_backend_func)
+#if IS_ENABLED(CONFIG_AGP)
+	if (rdev->flags & RADEON_IS_AGP)
 		return NULL;
-	return (struct radeon_ttm_tt *)ttm;
+#endif
+
+	if (!ttm)
+		return NULL;
+	return container_of(ttm, struct radeon_ttm_tt, ttm.ttm);
 }
 
-static int radeon_ttm_tt_populate(struct ttm_tt *ttm,
-			struct ttm_operation_ctx *ctx)
+static int radeon_ttm_tt_populate(struct ttm_bo_device *bdev,
+				  struct ttm_tt *ttm,
+				  struct ttm_operation_ctx *ctx)
 {
-	struct radeon_ttm_tt *gtt = radeon_ttm_tt_to_gtt(ttm);
-	struct radeon_device *rdev;
+	struct radeon_device *rdev = radeon_get_rdev(bdev);
+	struct radeon_ttm_tt *gtt = radeon_ttm_tt_to_gtt(rdev, ttm);
 	bool slave = !!(ttm->page_flags & TTM_PAGE_FLAG_SG);
 
 	if (gtt && gtt->userptr) {
@@ -622,21 +639,20 @@ static int radeon_ttm_tt_populate(struct ttm_tt *ttm,
 			return -ENOMEM;
 
 		ttm->page_flags |= TTM_PAGE_FLAG_SG;
-		ttm->state = tt_unbound;
+		ttm_tt_set_populated(ttm);
 		return 0;
 	}
 
 	if (slave && ttm->sg) {
 		drm_prime_sg_to_page_addr_arrays(ttm->sg, ttm->pages,
 						 gtt->ttm.dma_address, ttm->num_pages);
-		ttm->state = tt_unbound;
+		ttm_tt_set_populated(ttm);
 		return 0;
 	}
 
-	rdev = radeon_get_rdev(ttm->bdev);
 #if IS_ENABLED(CONFIG_AGP)
 	if (rdev->flags & RADEON_IS_AGP) {
-		return ttm_agp_tt_populate(ttm, ctx);
+		return ttm_pool_populate(ttm, ctx);
 	}
 #endif
 
@@ -649,10 +665,10 @@ static int radeon_ttm_tt_populate(struct ttm_tt *ttm,
 	return ttm_populate_and_map_pages(rdev->dev, &gtt->ttm, ctx);
 }
 
-static void radeon_ttm_tt_unpopulate(struct ttm_tt *ttm)
+static void radeon_ttm_tt_unpopulate(struct ttm_bo_device *bdev, struct ttm_tt *ttm)
 {
-	struct radeon_device *rdev;
-	struct radeon_ttm_tt *gtt = radeon_ttm_tt_to_gtt(ttm);
+	struct radeon_device *rdev = radeon_get_rdev(bdev);
+	struct radeon_ttm_tt *gtt = radeon_ttm_tt_to_gtt(rdev, ttm);
 	bool slave = !!(ttm->page_flags & TTM_PAGE_FLAG_SG);
 
 	if (gtt && gtt->userptr) {
@@ -664,10 +680,9 @@ static void radeon_ttm_tt_unpopulate(struct ttm_tt *ttm)
 	if (slave)
 		return;
 
-	rdev = radeon_get_rdev(ttm->bdev);
 #if IS_ENABLED(CONFIG_AGP)
 	if (rdev->flags & RADEON_IS_AGP) {
-		ttm_agp_tt_unpopulate(ttm);
+		ttm_pool_unpopulate(ttm);
 		return;
 	}
 #endif
@@ -682,10 +697,11 @@ static void radeon_ttm_tt_unpopulate(struct ttm_tt *ttm)
 	ttm_unmap_and_unpopulate_pages(rdev->dev, &gtt->ttm);
 }
 
-int radeon_ttm_tt_set_userptr(struct ttm_tt *ttm, uint64_t addr,
+int radeon_ttm_tt_set_userptr(struct radeon_device *rdev,
+			      struct ttm_tt *ttm, uint64_t addr,
 			      uint32_t flags)
 {
-	struct radeon_ttm_tt *gtt = radeon_ttm_tt_to_gtt(ttm);
+	struct radeon_ttm_tt *gtt = radeon_ttm_tt_to_gtt(rdev, ttm);
 
 	if (gtt == NULL)
 		return -EINVAL;
@@ -696,9 +712,69 @@ int radeon_ttm_tt_set_userptr(struct ttm_tt *ttm, uint64_t addr,
 	return 0;
 }
 
-bool radeon_ttm_tt_has_userptr(struct ttm_tt *ttm)
+bool radeon_ttm_tt_is_bound(struct ttm_bo_device *bdev,
+			    struct ttm_tt *ttm)
+{
+#if IS_ENABLED(CONFIG_AGP)
+	struct radeon_device *rdev = radeon_get_rdev(bdev);
+	if (rdev->flags & RADEON_IS_AGP)
+		return ttm_agp_is_bound(ttm);
+#endif
+	return radeon_ttm_backend_is_bound(ttm);
+}
+
+static int radeon_ttm_tt_bind(struct ttm_bo_device *bdev,
+			      struct ttm_tt *ttm,
+			      struct ttm_resource *bo_mem)
+{
+#if IS_ENABLED(CONFIG_AGP)
+	struct radeon_device *rdev = radeon_get_rdev(bdev);
+#endif
+
+	if (!bo_mem)
+		return -EINVAL;
+#if IS_ENABLED(CONFIG_AGP)
+	if (rdev->flags & RADEON_IS_AGP)
+		return ttm_agp_bind(ttm, bo_mem);
+#endif
+
+	return radeon_ttm_backend_bind(bdev, ttm, bo_mem);
+}
+
+static void radeon_ttm_tt_unbind(struct ttm_bo_device *bdev,
+				 struct ttm_tt *ttm)
+{
+#if IS_ENABLED(CONFIG_AGP)
+	struct radeon_device *rdev = radeon_get_rdev(bdev);
+
+	if (rdev->flags & RADEON_IS_AGP) {
+		ttm_agp_unbind(ttm);
+		return;
+	}
+#endif
+	radeon_ttm_backend_unbind(bdev, ttm);
+}
+
+static void radeon_ttm_tt_destroy(struct ttm_bo_device *bdev,
+				  struct ttm_tt *ttm)
+{
+#if IS_ENABLED(CONFIG_AGP)
+	struct radeon_device *rdev = radeon_get_rdev(bdev);
+
+	if (rdev->flags & RADEON_IS_AGP) {
+		ttm_agp_unbind(ttm);
+		ttm_tt_destroy_common(bdev, ttm);
+		ttm_agp_destroy(ttm);
+		return;
+	}
+#endif
+	radeon_ttm_backend_destroy(bdev, ttm);
+}
+
+bool radeon_ttm_tt_has_userptr(struct radeon_device *rdev,
+			       struct ttm_tt *ttm)
 {
-	struct radeon_ttm_tt *gtt = radeon_ttm_tt_to_gtt(ttm);
+	struct radeon_ttm_tt *gtt = radeon_ttm_tt_to_gtt(rdev, ttm);
 
 	if (gtt == NULL)
 		return false;
@@ -706,9 +782,10 @@ bool radeon_ttm_tt_has_userptr(struct ttm_tt *ttm)
 	return !!gtt->userptr;
 }
 
-bool radeon_ttm_tt_is_readonly(struct ttm_tt *ttm)
+bool radeon_ttm_tt_is_readonly(struct radeon_device *rdev,
+			       struct ttm_tt *ttm)
 {
-	struct radeon_ttm_tt *gtt = radeon_ttm_tt_to_gtt(ttm);
+	struct radeon_ttm_tt *gtt = radeon_ttm_tt_to_gtt(rdev, ttm);
 
 	if (gtt == NULL)
 		return false;
@@ -720,6 +797,9 @@ static struct ttm_bo_driver radeon_bo_driver = {
 	.ttm_tt_create = &radeon_ttm_tt_create,
 	.ttm_tt_populate = &radeon_ttm_tt_populate,
 	.ttm_tt_unpopulate = &radeon_ttm_tt_unpopulate,
+	.ttm_tt_bind = &radeon_ttm_tt_bind,
+	.ttm_tt_unbind = &radeon_ttm_tt_unbind,
+	.ttm_tt_destroy = &radeon_ttm_tt_destroy,
 	.eviction_valuable = ttm_bo_eviction_valuable,
 	.evict_flags = &radeon_evict_flags,
 	.move = &radeon_bo_move,
diff --git a/drivers/gpu/drm/radeon/radeon_vm.c b/drivers/gpu/drm/radeon/radeon_vm.c
index 71e2c3785ab9..ebad27c91a0d 100644
--- a/drivers/gpu/drm/radeon/radeon_vm.c
+++ b/drivers/gpu/drm/radeon/radeon_vm.c
@@ -942,7 +942,7 @@ int radeon_vm_bo_update(struct radeon_device *rdev,
 	bo_va->flags &= ~RADEON_VM_PAGE_VALID;
 	bo_va->flags &= ~RADEON_VM_PAGE_SYSTEM;
 	bo_va->flags &= ~RADEON_VM_PAGE_SNOOPED;
-	if (bo_va->bo && radeon_ttm_tt_is_readonly(bo_va->bo->tbo.ttm))
+	if (bo_va->bo && radeon_ttm_tt_is_readonly(rdev, bo_va->bo->tbo.ttm))
 		bo_va->flags &= ~RADEON_VM_PAGE_WRITEABLE;
 
 	if (mem) {
diff --git a/drivers/gpu/drm/rockchip/rockchip_drm_gem.c b/drivers/gpu/drm/rockchip/rockchip_drm_gem.c
index cb50f2ba2e46..62e5d0970525 100644
--- a/drivers/gpu/drm/rockchip/rockchip_drm_gem.c
+++ b/drivers/gpu/drm/rockchip/rockchip_drm_gem.c
@@ -85,7 +85,8 @@ static int rockchip_gem_get_pages(struct rockchip_gem_object *rk_obj)
 
 	rk_obj->num_pages = rk_obj->base.size >> PAGE_SHIFT;
 
-	rk_obj->sgt = drm_prime_pages_to_sg(rk_obj->pages, rk_obj->num_pages);
+	rk_obj->sgt = drm_prime_pages_to_sg(rk_obj->base.dev,
+					    rk_obj->pages, rk_obj->num_pages);
 	if (IS_ERR(rk_obj->sgt)) {
 		ret = PTR_ERR(rk_obj->sgt);
 		goto err_put_pages;
@@ -441,7 +442,7 @@ struct sg_table *rockchip_gem_prime_get_sg_table(struct drm_gem_object *obj)
 	int ret;
 
 	if (rk_obj->pages)
-		return drm_prime_pages_to_sg(rk_obj->pages, rk_obj->num_pages);
+		return drm_prime_pages_to_sg(obj->dev, rk_obj->pages, rk_obj->num_pages);
 
 	sgt = kzalloc(sizeof(*sgt), GFP_KERNEL);
 	if (!sgt)
diff --git a/drivers/gpu/drm/tegra/gem.c b/drivers/gpu/drm/tegra/gem.c
index 01d94befab11..a2bac20ff19d 100644
--- a/drivers/gpu/drm/tegra/gem.c
+++ b/drivers/gpu/drm/tegra/gem.c
@@ -282,7 +282,7 @@ static int tegra_bo_get_pages(struct drm_device *drm, struct tegra_bo *bo)
 
 	bo->num_pages = bo->gem.size >> PAGE_SHIFT;
 
-	bo->sgt = drm_prime_pages_to_sg(bo->pages, bo->num_pages);
+	bo->sgt = drm_prime_pages_to_sg(bo->gem.dev, bo->pages, bo->num_pages);
 	if (IS_ERR(bo->sgt)) {
 		err = PTR_ERR(bo->sgt);
 		goto put_pages;
diff --git a/drivers/gpu/drm/ttm/ttm_agp_backend.c b/drivers/gpu/drm/ttm/ttm_agp_backend.c
index 09fe80e215c5..a98fd795b752 100644
--- a/drivers/gpu/drm/ttm/ttm_agp_backend.c
+++ b/drivers/gpu/drm/ttm/ttm_agp_backend.c
@@ -48,7 +48,7 @@ struct ttm_agp_backend {
 	struct agp_bridge_data *bridge;
 };
 
-static int ttm_agp_bind(struct ttm_tt *ttm, struct ttm_resource *bo_mem)
+int ttm_agp_bind(struct ttm_tt *ttm, struct ttm_resource *bo_mem)
 {
 	struct ttm_agp_backend *agp_be = container_of(ttm, struct ttm_agp_backend, ttm);
 	struct page *dummy_read_page = ttm_bo_glob.dummy_read_page;
@@ -57,6 +57,9 @@ static int ttm_agp_bind(struct ttm_tt *ttm, struct ttm_resource *bo_mem)
 	int ret, cached = (bo_mem->placement & TTM_PL_FLAG_CACHED);
 	unsigned i;
 
+	if (agp_be->mem)
+		return 0;
+
 	mem = agp_allocate_memory(agp_be->bridge, ttm->num_pages, AGP_USER_MEMORY);
 	if (unlikely(mem == NULL))
 		return -ENOMEM;
@@ -81,8 +84,9 @@ static int ttm_agp_bind(struct ttm_tt *ttm, struct ttm_resource *bo_mem)
 
 	return ret;
 }
+EXPORT_SYMBOL(ttm_agp_bind);
 
-static void ttm_agp_unbind(struct ttm_tt *ttm)
+void ttm_agp_unbind(struct ttm_tt *ttm)
 {
 	struct ttm_agp_backend *agp_be = container_of(ttm, struct ttm_agp_backend, ttm);
 
@@ -95,8 +99,20 @@ static void ttm_agp_unbind(struct ttm_tt *ttm)
 		agp_be->mem = NULL;
 	}
 }
+EXPORT_SYMBOL(ttm_agp_unbind);
+
+bool ttm_agp_is_bound(struct ttm_tt *ttm)
+{
+	struct ttm_agp_backend *agp_be = container_of(ttm, struct ttm_agp_backend, ttm);
+
+	if (!ttm)
+		return false;
+
+	return (agp_be->mem != NULL);
+}
+EXPORT_SYMBOL(ttm_agp_is_bound);
 
-static void ttm_agp_destroy(struct ttm_tt *ttm)
+void ttm_agp_destroy(struct ttm_tt *ttm)
 {
 	struct ttm_agp_backend *agp_be = container_of(ttm, struct ttm_agp_backend, ttm);
 
@@ -105,12 +121,7 @@ static void ttm_agp_destroy(struct ttm_tt *ttm)
 	ttm_tt_fini(ttm);
 	kfree(agp_be);
 }
-
-static struct ttm_backend_func ttm_agp_func = {
-	.bind = ttm_agp_bind,
-	.unbind = ttm_agp_unbind,
-	.destroy = ttm_agp_destroy,
-};
+EXPORT_SYMBOL(ttm_agp_destroy);
 
 struct ttm_tt *ttm_agp_tt_create(struct ttm_buffer_object *bo,
 				 struct agp_bridge_data *bridge,
@@ -124,7 +135,6 @@ struct ttm_tt *ttm_agp_tt_create(struct ttm_buffer_object *bo,
 
 	agp_be->mem = NULL;
 	agp_be->bridge = bridge;
-	agp_be->ttm.func = &ttm_agp_func;
 
 	if (ttm_tt_init(&agp_be->ttm, bo, page_flags)) {
 		kfree(agp_be);
@@ -134,18 +144,3 @@ struct ttm_tt *ttm_agp_tt_create(struct ttm_buffer_object *bo,
 	return &agp_be->ttm;
 }
 EXPORT_SYMBOL(ttm_agp_tt_create);
-
-int ttm_agp_tt_populate(struct ttm_tt *ttm, struct ttm_operation_ctx *ctx)
-{
-	if (ttm->state != tt_unpopulated)
-		return 0;
-
-	return ttm_pool_populate(ttm, ctx);
-}
-EXPORT_SYMBOL(ttm_agp_tt_populate);
-
-void ttm_agp_tt_unpopulate(struct ttm_tt *ttm)
-{
-	ttm_pool_unpopulate(ttm);
-}
-EXPORT_SYMBOL(ttm_agp_tt_unpopulate);
diff --git a/drivers/gpu/drm/ttm/ttm_bo.c b/drivers/gpu/drm/ttm/ttm_bo.c
index e3931e515906..70b3bee27850 100644
--- a/drivers/gpu/drm/ttm/ttm_bo.c
+++ b/drivers/gpu/drm/ttm/ttm_bo.c
@@ -64,34 +64,18 @@ static void ttm_bo_default_destroy(struct ttm_buffer_object *bo)
 	kfree(bo);
 }
 
-static inline int ttm_mem_type_from_place(const struct ttm_place *place,
-					  uint32_t *mem_type)
-{
-	int pos;
-
-	pos = ffs(place->flags & TTM_PL_MASK_MEM);
-	if (unlikely(!pos))
-		return -EINVAL;
-
-	*mem_type = pos - 1;
-	return 0;
-}
-
 static void ttm_bo_mem_space_debug(struct ttm_buffer_object *bo,
 					struct ttm_placement *placement)
 {
 	struct drm_printer p = drm_debug_printer(TTM_PFX);
-	int i, ret, mem_type;
 	struct ttm_resource_manager *man;
+	int i, mem_type;
 
 	drm_printf(&p, "No space for %p (%lu pages, %luK, %luM)\n",
 		   bo, bo->mem.num_pages, bo->mem.size >> 10,
 		   bo->mem.size >> 20);
 	for (i = 0; i < placement->num_placement; i++) {
-		ret = ttm_mem_type_from_place(&placement->placement[i],
-						&mem_type);
-		if (ret)
-			return;
+		mem_type = placement->placement[i].mem_type;
 		drm_printf(&p, "  placement[%d]=0x%08X (%d)\n",
 			   i, placement->placement[i].flags, mem_type);
 		man = ttm_manager_type(bo->bdev, mem_type);
@@ -125,12 +109,6 @@ static struct kobj_type ttm_bo_glob_kobj_type  = {
 	.default_attrs = ttm_bo_global_attrs
 };
 
-
-static inline uint32_t ttm_bo_type_flags(unsigned type)
-{
-	return 1 << (type);
-}
-
 static void ttm_bo_add_mem_to_lru(struct ttm_buffer_object *bo,
 				  struct ttm_resource *mem)
 {
@@ -263,11 +241,7 @@ static int ttm_bo_handle_move_mem(struct ttm_buffer_object *bo,
 	struct ttm_resource_manager *new_man = ttm_manager_type(bdev, mem->mem_type);
 	int ret;
 
-	ret = ttm_mem_io_lock(old_man, true);
-	if (unlikely(ret != 0))
-		goto out_err;
-	ttm_bo_unmap_virtual_locked(bo);
-	ttm_mem_io_unlock(old_man);
+	ttm_bo_unmap_virtual(bo);
 
 	/*
 	 * Create and bind a ttm if required.
@@ -286,7 +260,11 @@ static int ttm_bo_handle_move_mem(struct ttm_buffer_object *bo,
 			goto out_err;
 
 		if (mem->mem_type != TTM_PL_SYSTEM) {
-			ret = ttm_tt_bind(bo->ttm, mem, ctx);
+			ret = ttm_tt_populate(bdev, bo->ttm, ctx);
+			if (ret)
+				goto out_err;
+
+			ret = ttm_bo_tt_bind(bo, mem);
 			if (ret)
 				goto out_err;
 		}
@@ -320,17 +298,13 @@ static int ttm_bo_handle_move_mem(struct ttm_buffer_object *bo,
 	}
 
 moved:
-	bo->evicted = false;
-
 	ctx->bytes_moved += bo->num_pages << PAGE_SHIFT;
 	return 0;
 
 out_err:
 	new_man = ttm_manager_type(bdev, bo->mem.mem_type);
-	if (!new_man->use_tt) {
-		ttm_tt_destroy(bo->ttm);
-		bo->ttm = NULL;
-	}
+	if (!new_man->use_tt)
+		ttm_bo_tt_destroy(bo);
 
 	return ret;
 }
@@ -348,8 +322,7 @@ static void ttm_bo_cleanup_memtype_use(struct ttm_buffer_object *bo)
 	if (bo->bdev->driver->move_notify)
 		bo->bdev->driver->move_notify(bo, false, NULL);
 
-	ttm_tt_destroy(bo->ttm);
-	bo->ttm = NULL;
+	ttm_bo_tt_destroy(bo);
 	ttm_resource_free(bo, &bo->mem);
 }
 
@@ -538,7 +511,6 @@ static void ttm_bo_release(struct kref *kref)
 	struct ttm_buffer_object *bo =
 	    container_of(kref, struct ttm_buffer_object, kref);
 	struct ttm_bo_device *bdev = bo->bdev;
-	struct ttm_resource_manager *man = ttm_manager_type(bdev, bo->mem.mem_type);
 	size_t acc_size = bo->acc_size;
 	int ret;
 
@@ -556,9 +528,7 @@ static void ttm_bo_release(struct kref *kref)
 			bo->bdev->driver->release_notify(bo);
 
 		drm_vma_offset_remove(bdev->vma_manager, &bo->base.vma_node);
-		ttm_mem_io_lock(man, false);
-		ttm_mem_io_free_vm(bo);
-		ttm_mem_io_unlock(man);
+		ttm_mem_io_free(bdev, &bo->mem);
 	}
 
 	if (!dma_resv_test_signaled_rcu(bo->base.resv, true) ||
@@ -648,8 +618,8 @@ static int ttm_bo_evict(struct ttm_buffer_object *bo,
 
 	evict_mem = bo->mem;
 	evict_mem.mm_node = NULL;
-	evict_mem.bus.io_reserved_vm = false;
-	evict_mem.bus.io_reserved_count = 0;
+	evict_mem.bus.offset = 0;
+	evict_mem.bus.addr = NULL;
 
 	ret = ttm_bo_mem_space(bo, &placement, &evict_mem, ctx);
 	if (ret) {
@@ -666,9 +636,7 @@ static int ttm_bo_evict(struct ttm_buffer_object *bo,
 		if (ret != -ERESTARTSYS)
 			pr_err("Buffer eviction failed\n");
 		ttm_resource_free(bo, &evict_mem);
-		goto out;
 	}
-	bo->evicted = true;
 out:
 	return ret;
 }
@@ -905,8 +873,6 @@ static uint32_t ttm_bo_select_caching(struct ttm_resource_manager *man,
 
 	if ((cur_placement & caching) != 0)
 		result |= (cur_placement & caching);
-	else if ((man->default_caching & caching) != 0)
-		result |= man->default_caching;
 	else if ((TTM_PL_FLAG_CACHED & caching) != 0)
 		result |= TTM_PL_FLAG_CACHED;
 	else if ((TTM_PL_FLAG_WC & caching) != 0)
@@ -917,25 +883,6 @@ static uint32_t ttm_bo_select_caching(struct ttm_resource_manager *man,
 	return result;
 }
 
-static bool ttm_bo_mt_compatible(struct ttm_resource_manager *man,
-				 uint32_t mem_type,
-				 const struct ttm_place *place,
-				 uint32_t *masked_placement)
-{
-	uint32_t cur_flags = ttm_bo_type_flags(mem_type);
-
-	if ((cur_flags & place->flags & TTM_PL_MASK_MEM) == 0)
-		return false;
-
-	if ((place->flags & man->available_caching) == 0)
-		return false;
-
-	cur_flags |= (place->flags & man->available_caching);
-
-	*masked_placement = cur_flags;
-	return true;
-}
-
 /**
  * ttm_bo_mem_placement - check if placement is compatible
  * @bo: BO to find memory for
@@ -953,30 +900,18 @@ static int ttm_bo_mem_placement(struct ttm_buffer_object *bo,
 				struct ttm_operation_ctx *ctx)
 {
 	struct ttm_bo_device *bdev = bo->bdev;
-	uint32_t mem_type = TTM_PL_SYSTEM;
 	struct ttm_resource_manager *man;
 	uint32_t cur_flags = 0;
-	int ret;
 
-	ret = ttm_mem_type_from_place(place, &mem_type);
-	if (ret)
-		return ret;
-
-	man = ttm_manager_type(bdev, mem_type);
+	man = ttm_manager_type(bdev, place->mem_type);
 	if (!man || !ttm_resource_manager_used(man))
 		return -EBUSY;
 
-	if (!ttm_bo_mt_compatible(man, mem_type, place, &cur_flags))
-		return -EBUSY;
-
-	cur_flags = ttm_bo_select_caching(man, bo->mem.placement, cur_flags);
-	/*
-	 * Use the access and other non-mapping-related flag bits from
-	 * the memory placement flags to the current flags
-	 */
-	ttm_flag_masked(&cur_flags, place->flags, ~TTM_PL_MASK_MEMTYPE);
+	cur_flags = ttm_bo_select_caching(man, bo->mem.placement,
+					  place->flags);
+	cur_flags |= place->flags & ~TTM_PL_MASK_CACHING;
 
-	mem->mem_type = mem_type;
+	mem->mem_type = place->mem_type;
 	mem->placement = cur_flags;
 
 	spin_lock(&ttm_bo_glob.lru_lock);
@@ -1013,10 +948,8 @@ int ttm_bo_mem_space(struct ttm_buffer_object *bo,
 		struct ttm_resource_manager *man;
 
 		ret = ttm_bo_mem_placement(bo, place, mem, ctx);
-		if (ret == -EBUSY)
-			continue;
 		if (ret)
-			goto error;
+			continue;
 
 		type_found = true;
 		ret = ttm_resource_alloc(bo, place, mem);
@@ -1041,10 +974,8 @@ int ttm_bo_mem_space(struct ttm_buffer_object *bo,
 		const struct ttm_place *place = &placement->busy_placement[i];
 
 		ret = ttm_bo_mem_placement(bo, place, mem, ctx);
-		if (ret == -EBUSY)
-			continue;
 		if (ret)
-			goto error;
+			continue;
 
 		type_found = true;
 		ret = ttm_bo_mem_force_space(bo, place, mem, ctx);
@@ -1082,8 +1013,8 @@ static int ttm_bo_move_buffer(struct ttm_buffer_object *bo,
 	mem.num_pages = bo->num_pages;
 	mem.size = mem.num_pages << PAGE_SHIFT;
 	mem.page_alignment = bo->mem.page_alignment;
-	mem.bus.io_reserved_vm = false;
-	mem.bus.io_reserved_count = 0;
+	mem.bus.offset = 0;
+	mem.bus.addr = NULL;
 	mem.mm_node = NULL;
 
 	/*
@@ -1115,7 +1046,7 @@ static bool ttm_bo_places_compat(const struct ttm_place *places,
 
 		*new_flags = heap->flags;
 		if ((*new_flags & mem->placement & TTM_PL_MASK_CACHING) &&
-		    (*new_flags & mem->placement & TTM_PL_MASK_MEM) &&
+		    (mem->mem_type == heap->mem_type) &&
 		    (!(*new_flags & TTM_PL_FLAG_CONTIGUOUS) ||
 		     (mem->placement & TTM_PL_FLAG_CONTIGUOUS)))
 			return true;
@@ -1170,12 +1101,8 @@ int ttm_bo_validate(struct ttm_buffer_object *bo,
 		if (ret)
 			return ret;
 	} else {
-		/*
-		 * Use the access and other non-mapping-related flag bits from
-		 * the compatible memory placement flags to the active flags
-		 */
-		ttm_flag_masked(&bo->mem.placement, new_flags,
-				~TTM_PL_MASK_MEMTYPE);
+		bo->mem.placement &= TTM_PL_MASK_CACHING;
+		bo->mem.placement |= new_flags & ~TTM_PL_MASK_CACHING;
 	}
 	/*
 	 * We might need to add a TTM.
@@ -1232,7 +1159,6 @@ int ttm_bo_init_reserved(struct ttm_bo_device *bdev,
 	INIT_LIST_HEAD(&bo->lru);
 	INIT_LIST_HEAD(&bo->ddestroy);
 	INIT_LIST_HEAD(&bo->swap);
-	INIT_LIST_HEAD(&bo->io_reserve_lru);
 	bo->bdev = bdev;
 	bo->type = type;
 	bo->num_pages = num_pages;
@@ -1241,10 +1167,10 @@ int ttm_bo_init_reserved(struct ttm_bo_device *bdev,
 	bo->mem.num_pages = bo->num_pages;
 	bo->mem.mm_node = NULL;
 	bo->mem.page_alignment = page_alignment;
-	bo->mem.bus.io_reserved_vm = false;
-	bo->mem.bus.io_reserved_count = 0;
+	bo->mem.bus.offset = 0;
+	bo->mem.bus.addr = NULL;
 	bo->moving = NULL;
-	bo->mem.placement = (TTM_PL_FLAG_SYSTEM | TTM_PL_FLAG_CACHED);
+	bo->mem.placement = TTM_PL_FLAG_CACHED;
 	bo->acc_size = acc_size;
 	bo->sg = sg;
 	if (resv) {
@@ -1325,9 +1251,9 @@ int ttm_bo_init(struct ttm_bo_device *bdev,
 }
 EXPORT_SYMBOL(ttm_bo_init);
 
-size_t ttm_bo_acc_size(struct ttm_bo_device *bdev,
-		       unsigned long bo_size,
-		       unsigned struct_size)
+static size_t ttm_bo_acc_size(struct ttm_bo_device *bdev,
+			      unsigned long bo_size,
+			      unsigned struct_size)
 {
 	unsigned npages = (PAGE_ALIGN(bo_size)) >> PAGE_SHIFT;
 	size_t size = 0;
@@ -1337,7 +1263,6 @@ size_t ttm_bo_acc_size(struct ttm_bo_device *bdev,
 	size += ttm_round_pot(sizeof(struct ttm_tt));
 	return size;
 }
-EXPORT_SYMBOL(ttm_bo_acc_size);
 
 size_t ttm_bo_dma_acc_size(struct ttm_bo_device *bdev,
 			   unsigned long bo_size,
@@ -1500,8 +1425,6 @@ static void ttm_bo_init_sysman(struct ttm_bo_device *bdev)
 	 * Other types need to be driver / IOCTL initialized.
 	 */
 	man->use_tt = true;
-	man->available_caching = TTM_PL_MASK_CACHING;
-	man->default_caching = TTM_PL_FLAG_CACHED;
 
 	ttm_resource_manager_init(man, 0);
 	ttm_set_driver_manager(bdev, TTM_PL_SYSTEM, man);
@@ -1545,25 +1468,13 @@ EXPORT_SYMBOL(ttm_bo_device_init);
  * buffer object vm functions.
  */
 
-void ttm_bo_unmap_virtual_locked(struct ttm_buffer_object *bo)
-{
-	struct ttm_bo_device *bdev = bo->bdev;
-
-	drm_vma_node_unmap(&bo->base.vma_node, bdev->dev_mapping);
-	ttm_mem_io_free_vm(bo);
-}
-
 void ttm_bo_unmap_virtual(struct ttm_buffer_object *bo)
 {
 	struct ttm_bo_device *bdev = bo->bdev;
-	struct ttm_resource_manager *man = ttm_manager_type(bdev, bo->mem.mem_type);
 
-	ttm_mem_io_lock(man, false);
-	ttm_bo_unmap_virtual_locked(bo);
-	ttm_mem_io_unlock(man);
+	drm_vma_node_unmap(&bo->base.vma_node, bdev->dev_mapping);
+	ttm_mem_io_free(bdev, &bo->mem);
 }
-
-
 EXPORT_SYMBOL(ttm_bo_unmap_virtual);
 
 int ttm_bo_wait(struct ttm_buffer_object *bo,
@@ -1647,7 +1558,7 @@ int ttm_bo_swapout(struct ttm_bo_global *glob, struct ttm_operation_ctx *ctx)
 
 		evict_mem = bo->mem;
 		evict_mem.mm_node = NULL;
-		evict_mem.placement = TTM_PL_FLAG_SYSTEM | TTM_PL_FLAG_CACHED;
+		evict_mem.placement = TTM_PL_FLAG_CACHED;
 		evict_mem.mem_type = TTM_PL_SYSTEM;
 
 		ret = ttm_bo_handle_move_mem(bo, &evict_mem, true, &ctx);
@@ -1673,7 +1584,7 @@ int ttm_bo_swapout(struct ttm_bo_global *glob, struct ttm_operation_ctx *ctx)
 	if (bo->bdev->driver->swap_notify)
 		bo->bdev->driver->swap_notify(bo);
 
-	ret = ttm_tt_swapout(bo->ttm, bo->persistent_swap_storage);
+	ret = ttm_tt_swapout(bo->bdev, bo->ttm, bo->persistent_swap_storage);
 out:
 
 	/**
@@ -1698,3 +1609,22 @@ void ttm_bo_swapout_all(void)
 	while (ttm_bo_swapout(&ttm_bo_glob, &ctx) == 0);
 }
 EXPORT_SYMBOL(ttm_bo_swapout_all);
+
+void ttm_bo_tt_destroy(struct ttm_buffer_object *bo)
+{
+	if (bo->ttm == NULL)
+		return;
+
+	ttm_tt_destroy(bo->bdev, bo->ttm);
+	bo->ttm = NULL;
+}
+
+int ttm_bo_tt_bind(struct ttm_buffer_object *bo, struct ttm_resource *mem)
+{
+	return bo->bdev->driver->ttm_tt_bind(bo->bdev, bo->ttm, mem);
+}
+
+void ttm_bo_tt_unbind(struct ttm_buffer_object *bo)
+{
+	bo->bdev->driver->ttm_tt_unbind(bo->bdev, bo->ttm);
+}
diff --git a/drivers/gpu/drm/ttm/ttm_bo_util.c b/drivers/gpu/drm/ttm/ttm_bo_util.c
index ee04716b2603..fb2a25f8408f 100644
--- a/drivers/gpu/drm/ttm/ttm_bo_util.c
+++ b/drivers/gpu/drm/ttm/ttm_bo_util.c
@@ -67,10 +67,8 @@ int ttm_bo_move_ttm(struct ttm_buffer_object *bo,
 			return ret;
 		}
 
-		ttm_tt_unbind(ttm);
+		ttm_bo_tt_unbind(bo);
 		ttm_bo_free_old_node(bo);
-		ttm_flag_masked(&old_mem->placement, TTM_PL_FLAG_SYSTEM,
-				TTM_PL_MASK_MEM);
 		old_mem->mem_type = TTM_PL_SYSTEM;
 	}
 
@@ -79,134 +77,56 @@ int ttm_bo_move_ttm(struct ttm_buffer_object *bo,
 		return ret;
 
 	if (new_mem->mem_type != TTM_PL_SYSTEM) {
-		ret = ttm_tt_bind(ttm, new_mem, ctx);
+
+		ret = ttm_tt_populate(bo->bdev, ttm, ctx);
 		if (unlikely(ret != 0))
 			return ret;
-	}
 
-	*old_mem = *new_mem;
-	new_mem->mm_node = NULL;
+		ret = ttm_bo_tt_bind(bo, new_mem);
+		if (unlikely(ret != 0))
+			return ret;
+	}
 
+	ttm_bo_assign_mem(bo, new_mem);
 	return 0;
 }
 EXPORT_SYMBOL(ttm_bo_move_ttm);
 
-int ttm_mem_io_lock(struct ttm_resource_manager *man, bool interruptible)
-{
-	if (likely(!man->use_io_reserve_lru))
-		return 0;
-
-	if (interruptible)
-		return mutex_lock_interruptible(&man->io_reserve_mutex);
-
-	mutex_lock(&man->io_reserve_mutex);
-	return 0;
-}
-
-void ttm_mem_io_unlock(struct ttm_resource_manager *man)
-{
-	if (likely(!man->use_io_reserve_lru))
-		return;
-
-	mutex_unlock(&man->io_reserve_mutex);
-}
-
-static int ttm_mem_io_evict(struct ttm_resource_manager *man)
-{
-	struct ttm_buffer_object *bo;
-
-	bo = list_first_entry_or_null(&man->io_reserve_lru,
-				      struct ttm_buffer_object,
-				      io_reserve_lru);
-	if (!bo)
-		return -ENOSPC;
-
-	list_del_init(&bo->io_reserve_lru);
-	ttm_bo_unmap_virtual_locked(bo);
-	return 0;
-}
-
 int ttm_mem_io_reserve(struct ttm_bo_device *bdev,
 		       struct ttm_resource *mem)
 {
-	struct ttm_resource_manager *man = ttm_manager_type(bdev, mem->mem_type);
-	int ret;
-
-	if (mem->bus.io_reserved_count++)
+	if (mem->bus.offset || mem->bus.addr)
 		return 0;
 
+	mem->bus.is_iomem = false;
 	if (!bdev->driver->io_mem_reserve)
 		return 0;
 
-	mem->bus.addr = NULL;
-	mem->bus.offset = 0;
-	mem->bus.base = 0;
-	mem->bus.is_iomem = false;
-retry:
-	ret = bdev->driver->io_mem_reserve(bdev, mem);
-	if (ret == -ENOSPC) {
-		ret = ttm_mem_io_evict(man);
-		if (ret == 0)
-			goto retry;
-	}
-	return ret;
+	return bdev->driver->io_mem_reserve(bdev, mem);
 }
 
 void ttm_mem_io_free(struct ttm_bo_device *bdev,
 		     struct ttm_resource *mem)
 {
-	if (--mem->bus.io_reserved_count)
+	if (!mem->bus.offset && !mem->bus.addr)
 		return;
 
-	if (!bdev->driver->io_mem_free)
-		return;
-
-	bdev->driver->io_mem_free(bdev, mem);
-}
-
-int ttm_mem_io_reserve_vm(struct ttm_buffer_object *bo)
-{
-	struct ttm_resource_manager *man = ttm_manager_type(bo->bdev, bo->mem.mem_type);
-	struct ttm_resource *mem = &bo->mem;
-	int ret;
-
-	if (mem->bus.io_reserved_vm)
-		return 0;
-
-	ret = ttm_mem_io_reserve(bo->bdev, mem);
-	if (unlikely(ret != 0))
-		return ret;
-	mem->bus.io_reserved_vm = true;
-	if (man->use_io_reserve_lru)
-		list_add_tail(&bo->io_reserve_lru,
-			      &man->io_reserve_lru);
-	return 0;
-}
+	if (bdev->driver->io_mem_free)
+		bdev->driver->io_mem_free(bdev, mem);
 
-void ttm_mem_io_free_vm(struct ttm_buffer_object *bo)
-{
-	struct ttm_resource *mem = &bo->mem;
-
-	if (!mem->bus.io_reserved_vm)
-		return;
-
-	mem->bus.io_reserved_vm = false;
-	list_del_init(&bo->io_reserve_lru);
-	ttm_mem_io_free(bo->bdev, mem);
+	mem->bus.offset = 0;
+	mem->bus.addr = NULL;
 }
 
 static int ttm_resource_ioremap(struct ttm_bo_device *bdev,
 			       struct ttm_resource *mem,
 			       void **virtual)
 {
-	struct ttm_resource_manager *man = ttm_manager_type(bdev, mem->mem_type);
 	int ret;
 	void *addr;
 
 	*virtual = NULL;
-	(void) ttm_mem_io_lock(man, false);
 	ret = ttm_mem_io_reserve(bdev, mem);
-	ttm_mem_io_unlock(man);
 	if (ret || !mem->bus.is_iomem)
 		return ret;
 
@@ -216,15 +136,11 @@ static int ttm_resource_ioremap(struct ttm_bo_device *bdev,
 		size_t bus_size = (size_t)mem->num_pages << PAGE_SHIFT;
 
 		if (mem->placement & TTM_PL_FLAG_WC)
-			addr = ioremap_wc(mem->bus.base + mem->bus.offset,
-					  bus_size);
+			addr = ioremap_wc(mem->bus.offset, bus_size);
 		else
-			addr = ioremap(mem->bus.base + mem->bus.offset,
-				       bus_size);
+			addr = ioremap(mem->bus.offset, bus_size);
 		if (!addr) {
-			(void) ttm_mem_io_lock(man, false);
 			ttm_mem_io_free(bdev, mem);
-			ttm_mem_io_unlock(man);
 			return -ENOMEM;
 		}
 	}
@@ -236,15 +152,9 @@ static void ttm_resource_iounmap(struct ttm_bo_device *bdev,
 				struct ttm_resource *mem,
 				void *virtual)
 {
-	struct ttm_resource_manager *man;
-
-	man = ttm_manager_type(bdev, mem->mem_type);
-
 	if (virtual && mem->bus.addr == NULL)
 		iounmap(virtual);
-	(void) ttm_mem_io_lock(man, false);
 	ttm_mem_io_free(bdev, mem);
-	ttm_mem_io_unlock(man);
 }
 
 static int ttm_copy_io_page(void *dst, void *src, unsigned long page)
@@ -342,7 +252,7 @@ int ttm_bo_move_memcpy(struct ttm_buffer_object *bo,
 	 * Don't move nonexistent data. Clear destination instead.
 	 */
 	if (old_iomap == NULL &&
-	    (ttm == NULL || (ttm->state == tt_unpopulated &&
+	    (ttm == NULL || (!ttm_tt_is_populated(ttm) &&
 			     !(ttm->page_flags & TTM_PAGE_FLAG_SWAPPED)))) {
 		memset_io(new_iomap, 0, new_mem->num_pages*PAGE_SIZE);
 		goto out2;
@@ -352,7 +262,7 @@ int ttm_bo_move_memcpy(struct ttm_buffer_object *bo,
 	 * TTM might be null for moves within the same region.
 	 */
 	if (ttm) {
-		ret = ttm_tt_populate(ttm, ctx);
+		ret = ttm_tt_populate(bdev, ttm, ctx);
 		if (ret)
 			goto out1;
 	}
@@ -387,13 +297,11 @@ int ttm_bo_move_memcpy(struct ttm_buffer_object *bo,
 	mb();
 out2:
 	old_copy = *old_mem;
-	*old_mem = *new_mem;
-	new_mem->mm_node = NULL;
 
-	if (!man->use_tt) {
-		ttm_tt_destroy(ttm);
-		bo->ttm = NULL;
-	}
+	ttm_bo_assign_mem(bo, new_mem);
+
+	if (!man->use_tt)
+		ttm_bo_tt_destroy(bo);
 
 out1:
 	ttm_resource_iounmap(bdev, old_mem, new_iomap);
@@ -458,7 +366,6 @@ static int ttm_buffer_object_transfer(struct ttm_buffer_object *bo,
 	INIT_LIST_HEAD(&fbo->base.ddestroy);
 	INIT_LIST_HEAD(&fbo->base.lru);
 	INIT_LIST_HEAD(&fbo->base.swap);
-	INIT_LIST_HEAD(&fbo->base.io_reserve_lru);
 	fbo->base.moving = NULL;
 	drm_vma_node_reset(&fbo->base.base.vma_node);
 
@@ -516,12 +423,10 @@ static int ttm_bo_ioremap(struct ttm_buffer_object *bo,
 	} else {
 		map->bo_kmap_type = ttm_bo_map_iomap;
 		if (mem->placement & TTM_PL_FLAG_WC)
-			map->virtual = ioremap_wc(bo->mem.bus.base +
-						  bo->mem.bus.offset + offset,
+			map->virtual = ioremap_wc(bo->mem.bus.offset + offset,
 						  size);
 		else
-			map->virtual = ioremap(bo->mem.bus.base +
-					       bo->mem.bus.offset + offset,
+			map->virtual = ioremap(bo->mem.bus.offset + offset,
 					       size);
 	}
 	return (!map->virtual) ? -ENOMEM : 0;
@@ -543,7 +448,7 @@ static int ttm_bo_kmap_ttm(struct ttm_buffer_object *bo,
 
 	BUG_ON(!ttm);
 
-	ret = ttm_tt_populate(ttm, &ctx);
+	ret = ttm_tt_populate(bo->bdev, ttm, &ctx);
 	if (ret)
 		return ret;
 
@@ -573,8 +478,6 @@ int ttm_bo_kmap(struct ttm_buffer_object *bo,
 		unsigned long start_page, unsigned long num_pages,
 		struct ttm_bo_kmap_obj *map)
 {
-	struct ttm_resource_manager *man =
-		ttm_manager_type(bo->bdev, bo->mem.mem_type);
 	unsigned long offset, size;
 	int ret;
 
@@ -585,9 +488,7 @@ int ttm_bo_kmap(struct ttm_buffer_object *bo,
 	if (start_page > bo->num_pages)
 		return -EINVAL;
 
-	(void) ttm_mem_io_lock(man, false);
 	ret = ttm_mem_io_reserve(bo->bdev, &bo->mem);
-	ttm_mem_io_unlock(man);
 	if (ret)
 		return ret;
 	if (!bo->mem.bus.is_iomem) {
@@ -602,10 +503,6 @@ EXPORT_SYMBOL(ttm_bo_kmap);
 
 void ttm_bo_kunmap(struct ttm_bo_kmap_obj *map)
 {
-	struct ttm_buffer_object *bo = map->bo;
-	struct ttm_resource_manager *man =
-		ttm_manager_type(bo->bdev, bo->mem.mem_type);
-
 	if (!map->virtual)
 		return;
 	switch (map->bo_kmap_type) {
@@ -623,167 +520,116 @@ void ttm_bo_kunmap(struct ttm_bo_kmap_obj *map)
 	default:
 		BUG();
 	}
-	(void) ttm_mem_io_lock(man, false);
 	ttm_mem_io_free(map->bo->bdev, &map->bo->mem);
-	ttm_mem_io_unlock(man);
 	map->virtual = NULL;
 	map->page = NULL;
 }
 EXPORT_SYMBOL(ttm_bo_kunmap);
 
-int ttm_bo_move_accel_cleanup(struct ttm_buffer_object *bo,
-			      struct dma_fence *fence,
-			      bool evict,
-			      struct ttm_resource *new_mem)
+static int ttm_bo_wait_free_node(struct ttm_buffer_object *bo,
+				 bool dst_use_tt)
 {
-	struct ttm_bo_device *bdev = bo->bdev;
-	struct ttm_resource_manager *man = ttm_manager_type(bdev, new_mem->mem_type);
-	struct ttm_resource *old_mem = &bo->mem;
 	int ret;
-	struct ttm_buffer_object *ghost_obj;
-
-	dma_resv_add_excl_fence(bo->base.resv, fence);
-	if (evict) {
-		ret = ttm_bo_wait(bo, false, false);
-		if (ret)
-			return ret;
-
-		if (!man->use_tt) {
-			ttm_tt_destroy(bo->ttm);
-			bo->ttm = NULL;
-		}
-		ttm_bo_free_old_node(bo);
-	} else {
-		/**
-		 * This should help pipeline ordinary buffer moves.
-		 *
-		 * Hang old buffer memory on a new buffer object,
-		 * and leave it to be released when the GPU
-		 * operation has completed.
-		 */
-
-		dma_fence_put(bo->moving);
-		bo->moving = dma_fence_get(fence);
-
-		ret = ttm_buffer_object_transfer(bo, &ghost_obj);
-		if (ret)
-			return ret;
-
-		dma_resv_add_excl_fence(&ghost_obj->base._resv, fence);
-
-		/**
-		 * If we're not moving to fixed memory, the TTM object
-		 * needs to stay alive. Otherwhise hang it on the ghost
-		 * bo to be unbound and destroyed.
-		 */
-
-		if (man->use_tt)
-			ghost_obj->ttm = NULL;
-		else
-			bo->ttm = NULL;
-
-		dma_resv_unlock(&ghost_obj->base._resv);
-		ttm_bo_put(ghost_obj);
-	}
-
-	*old_mem = *new_mem;
-	new_mem->mm_node = NULL;
+	ret = ttm_bo_wait(bo, false, false);
+	if (ret)
+		return ret;
 
+	if (!dst_use_tt)
+		ttm_bo_tt_destroy(bo);
+	ttm_bo_free_old_node(bo);
 	return 0;
 }
-EXPORT_SYMBOL(ttm_bo_move_accel_cleanup);
 
-int ttm_bo_pipeline_move(struct ttm_buffer_object *bo,
-			 struct dma_fence *fence, bool evict,
-			 struct ttm_resource *new_mem)
+static int ttm_bo_move_to_ghost(struct ttm_buffer_object *bo,
+				struct dma_fence *fence,
+				bool dst_use_tt)
 {
-	struct ttm_bo_device *bdev = bo->bdev;
-	struct ttm_resource *old_mem = &bo->mem;
-
-	struct ttm_resource_manager *from = ttm_manager_type(bdev, old_mem->mem_type);
-	struct ttm_resource_manager *to = ttm_manager_type(bdev, new_mem->mem_type);
-
+	struct ttm_buffer_object *ghost_obj;
 	int ret;
 
-	dma_resv_add_excl_fence(bo->base.resv, fence);
-
-	if (!evict) {
-		struct ttm_buffer_object *ghost_obj;
-
-		/**
-		 * This should help pipeline ordinary buffer moves.
-		 *
-		 * Hang old buffer memory on a new buffer object,
-		 * and leave it to be released when the GPU
-		 * operation has completed.
-		 */
+	/**
+	 * This should help pipeline ordinary buffer moves.
+	 *
+	 * Hang old buffer memory on a new buffer object,
+	 * and leave it to be released when the GPU
+	 * operation has completed.
+	 */
 
-		dma_fence_put(bo->moving);
-		bo->moving = dma_fence_get(fence);
+	dma_fence_put(bo->moving);
+	bo->moving = dma_fence_get(fence);
 
-		ret = ttm_buffer_object_transfer(bo, &ghost_obj);
-		if (ret)
-			return ret;
-
-		dma_resv_add_excl_fence(&ghost_obj->base._resv, fence);
+	ret = ttm_buffer_object_transfer(bo, &ghost_obj);
+	if (ret)
+		return ret;
 
-		/**
-		 * If we're not moving to fixed memory, the TTM object
-		 * needs to stay alive. Otherwhise hang it on the ghost
-		 * bo to be unbound and destroyed.
-		 */
+	dma_resv_add_excl_fence(&ghost_obj->base._resv, fence);
 
-		if (to->use_tt)
-			ghost_obj->ttm = NULL;
-		else
-			bo->ttm = NULL;
+	/**
+	 * If we're not moving to fixed memory, the TTM object
+	 * needs to stay alive. Otherwhise hang it on the ghost
+	 * bo to be unbound and destroyed.
+	 */
 
-		dma_resv_unlock(&ghost_obj->base._resv);
-		ttm_bo_put(ghost_obj);
+	if (dst_use_tt)
+		ghost_obj->ttm = NULL;
+	else
+		bo->ttm = NULL;
 
-	} else if (!from->use_tt) {
+	dma_resv_unlock(&ghost_obj->base._resv);
+	ttm_bo_put(ghost_obj);
+	return 0;
+}
 
-		/**
-		 * BO doesn't have a TTM we need to bind/unbind. Just remember
-		 * this eviction and free up the allocation
-		 */
+static void ttm_bo_move_pipeline_evict(struct ttm_buffer_object *bo,
+				       struct dma_fence *fence)
+{
+	struct ttm_bo_device *bdev = bo->bdev;
+	struct ttm_resource_manager *from = ttm_manager_type(bdev, bo->mem.mem_type);
 
-		spin_lock(&from->move_lock);
-		if (!from->move || dma_fence_is_later(fence, from->move)) {
-			dma_fence_put(from->move);
-			from->move = dma_fence_get(fence);
-		}
-		spin_unlock(&from->move_lock);
+	/**
+	 * BO doesn't have a TTM we need to bind/unbind. Just remember
+	 * this eviction and free up the allocation
+	 */
+	spin_lock(&from->move_lock);
+	if (!from->move || dma_fence_is_later(fence, from->move)) {
+		dma_fence_put(from->move);
+		from->move = dma_fence_get(fence);
+	}
+	spin_unlock(&from->move_lock);
 
-		ttm_bo_free_old_node(bo);
+	ttm_bo_free_old_node(bo);
 
-		dma_fence_put(bo->moving);
-		bo->moving = dma_fence_get(fence);
+	dma_fence_put(bo->moving);
+	bo->moving = dma_fence_get(fence);
+}
 
-	} else {
-		/**
-		 * Last resort, wait for the move to be completed.
-		 *
-		 * Should never happen in pratice.
-		 */
+int ttm_bo_move_accel_cleanup(struct ttm_buffer_object *bo,
+			      struct dma_fence *fence,
+			      bool evict,
+			      bool pipeline,
+			      struct ttm_resource *new_mem)
+{
+	struct ttm_bo_device *bdev = bo->bdev;
+	struct ttm_resource_manager *from = ttm_manager_type(bdev, bo->mem.mem_type);
+	struct ttm_resource_manager *man = ttm_manager_type(bdev, new_mem->mem_type);
+	int ret = 0;
 
-		ret = ttm_bo_wait(bo, false, false);
-		if (ret)
-			return ret;
+	dma_resv_add_excl_fence(bo->base.resv, fence);
+	if (!evict)
+		ret = ttm_bo_move_to_ghost(bo, fence, man->use_tt);
+	else if (!from->use_tt && pipeline)
+		ttm_bo_move_pipeline_evict(bo, fence);
+	else
+		ret = ttm_bo_wait_free_node(bo, man->use_tt);
 
-		if (!to->use_tt) {
-			ttm_tt_destroy(bo->ttm);
-			bo->ttm = NULL;
-		}
-		ttm_bo_free_old_node(bo);
-	}
+	if (ret)
+		return ret;
 
-	*old_mem = *new_mem;
-	new_mem->mm_node = NULL;
+	ttm_bo_assign_mem(bo, new_mem);
 
 	return 0;
 }
-EXPORT_SYMBOL(ttm_bo_pipeline_move);
+EXPORT_SYMBOL(ttm_bo_move_accel_cleanup);
 
 int ttm_bo_pipeline_gutting(struct ttm_buffer_object *bo)
 {
diff --git a/drivers/gpu/drm/ttm/ttm_bo_vm.c b/drivers/gpu/drm/ttm/ttm_bo_vm.c
index 01693e8f24b7..98a006fc30a5 100644
--- a/drivers/gpu/drm/ttm/ttm_bo_vm.c
+++ b/drivers/gpu/drm/ttm/ttm_bo_vm.c
@@ -101,8 +101,7 @@ static unsigned long ttm_bo_io_mem_pfn(struct ttm_buffer_object *bo,
 	if (bdev->driver->io_mem_pfn)
 		return bdev->driver->io_mem_pfn(bo, page_offset);
 
-	return ((bo->mem.bus.base + bo->mem.bus.offset) >> PAGE_SHIFT)
-		+ page_offset;
+	return (bo->mem.bus.offset >> PAGE_SHIFT) + page_offset;
 }
 
 /**
@@ -281,8 +280,6 @@ vm_fault_t ttm_bo_vm_fault_reserved(struct vm_fault *vmf,
 	pgoff_t i;
 	vm_fault_t ret = VM_FAULT_NOPAGE;
 	unsigned long address = vmf->address;
-	struct ttm_resource_manager *man =
-		ttm_manager_type(bdev, bo->mem.mem_type);
 
 	/*
 	 * Refuse to fault imported pages. This should be handled
@@ -321,24 +318,17 @@ vm_fault_t ttm_bo_vm_fault_reserved(struct vm_fault *vmf,
 	if (unlikely(ret != 0))
 		return ret;
 
-	err = ttm_mem_io_lock(man, true);
+	err = ttm_mem_io_reserve(bdev, &bo->mem);
 	if (unlikely(err != 0))
-		return VM_FAULT_NOPAGE;
-	err = ttm_mem_io_reserve_vm(bo);
-	if (unlikely(err != 0)) {
-		ret = VM_FAULT_SIGBUS;
-		goto out_io_unlock;
-	}
+		return VM_FAULT_SIGBUS;
 
 	page_offset = ((address - vma->vm_start) >> PAGE_SHIFT) +
 		vma->vm_pgoff - drm_vma_node_start(&bo->base.vma_node);
 	page_last = vma_pages(vma) + vma->vm_pgoff -
 		drm_vma_node_start(&bo->base.vma_node);
 
-	if (unlikely(page_offset >= bo->num_pages)) {
-		ret = VM_FAULT_SIGBUS;
-		goto out_io_unlock;
-	}
+	if (unlikely(page_offset >= bo->num_pages))
+		return VM_FAULT_SIGBUS;
 
 	prot = ttm_io_prot(bo->mem.placement, prot);
 	if (!bo->mem.bus.is_iomem) {
@@ -350,21 +340,17 @@ vm_fault_t ttm_bo_vm_fault_reserved(struct vm_fault *vmf,
 		};
 
 		ttm = bo->ttm;
-		if (ttm_tt_populate(bo->ttm, &ctx)) {
-			ret = VM_FAULT_OOM;
-			goto out_io_unlock;
-		}
+		if (ttm_tt_populate(bdev, bo->ttm, &ctx))
+			return VM_FAULT_OOM;
 	} else {
 		/* Iomem should not be marked encrypted */
 		prot = pgprot_decrypted(prot);
 	}
 
 	/* We don't prefault on huge faults. Yet. */
-	if (IS_ENABLED(CONFIG_TRANSPARENT_HUGEPAGE) && fault_page_size != 1) {
-		ret = ttm_bo_vm_insert_huge(vmf, bo, page_offset,
-					    fault_page_size, prot);
-		goto out_io_unlock;
-	}
+	if (IS_ENABLED(CONFIG_TRANSPARENT_HUGEPAGE) && fault_page_size != 1)
+		return ttm_bo_vm_insert_huge(vmf, bo, page_offset,
+					     fault_page_size, prot);
 
 	/*
 	 * Speculatively prefault a number of pages. Only error on
@@ -376,8 +362,7 @@ vm_fault_t ttm_bo_vm_fault_reserved(struct vm_fault *vmf,
 		} else {
 			page = ttm->pages[page_offset];
 			if (unlikely(!page && i == 0)) {
-				ret = VM_FAULT_OOM;
-				goto out_io_unlock;
+				return VM_FAULT_OOM;
 			} else if (unlikely(!page)) {
 				break;
 			}
@@ -404,7 +389,7 @@ vm_fault_t ttm_bo_vm_fault_reserved(struct vm_fault *vmf,
 		/* Never error on prefaulted PTEs */
 		if (unlikely((ret & VM_FAULT_ERROR))) {
 			if (i == 0)
-				goto out_io_unlock;
+				return VM_FAULT_NOPAGE;
 			else
 				break;
 		}
@@ -413,9 +398,6 @@ vm_fault_t ttm_bo_vm_fault_reserved(struct vm_fault *vmf,
 		if (unlikely(++page_offset >= page_last))
 			break;
 	}
-	ret = VM_FAULT_NOPAGE;
-out_io_unlock:
-	ttm_mem_io_unlock(man);
 	return ret;
 }
 EXPORT_SYMBOL(ttm_bo_vm_fault_reserved);
diff --git a/drivers/gpu/drm/ttm/ttm_memory.c b/drivers/gpu/drm/ttm/ttm_memory.c
index acd63b70d814..89d50f38c0f2 100644
--- a/drivers/gpu/drm/ttm/ttm_memory.c
+++ b/drivers/gpu/drm/ttm/ttm_memory.c
@@ -259,7 +259,7 @@ static bool ttm_zones_above_swap_target(struct ttm_mem_global *glob,
 	return false;
 }
 
-/**
+/*
  * At this point we only support a single shrink callback.
  * Extend this if needed, perhaps using a linked list of callbacks.
  * Note that this function is reentrant:
@@ -554,7 +554,6 @@ ttm_check_under_lowerlimit(struct ttm_mem_global *glob,
 
 	return false;
 }
-EXPORT_SYMBOL(ttm_check_under_lowerlimit);
 
 static int ttm_mem_global_reserve(struct ttm_mem_global *glob,
 				  struct ttm_mem_zone *single_zone,
@@ -682,9 +681,3 @@ size_t ttm_round_pot(size_t size)
 	return 0;
 }
 EXPORT_SYMBOL(ttm_round_pot);
-
-uint64_t ttm_get_kernel_zone_memory_size(struct ttm_mem_global *glob)
-{
-	return glob->zone_kernel->max_mem;
-}
-EXPORT_SYMBOL(ttm_get_kernel_zone_memory_size);
diff --git a/drivers/gpu/drm/ttm/ttm_page_alloc.c b/drivers/gpu/drm/ttm/ttm_page_alloc.c
index b40a4678c296..14660f723f71 100644
--- a/drivers/gpu/drm/ttm/ttm_page_alloc.c
+++ b/drivers/gpu/drm/ttm/ttm_page_alloc.c
@@ -1044,7 +1044,7 @@ ttm_pool_unpopulate_helper(struct ttm_tt *ttm, unsigned mem_count_update)
 put_pages:
 	ttm_put_pages(ttm->pages, ttm->num_pages, ttm->page_flags,
 		      ttm->caching_state);
-	ttm->state = tt_unpopulated;
+	ttm_tt_set_unpopulated(ttm);
 }
 
 int ttm_pool_populate(struct ttm_tt *ttm, struct ttm_operation_ctx *ctx)
@@ -1053,7 +1053,7 @@ int ttm_pool_populate(struct ttm_tt *ttm, struct ttm_operation_ctx *ctx)
 	unsigned i;
 	int ret;
 
-	if (ttm->state != tt_unpopulated)
+	if (ttm_tt_is_populated(ttm))
 		return 0;
 
 	if (ttm_check_under_lowerlimit(mem_glob, ttm->num_pages, ctx))
@@ -1083,7 +1083,7 @@ int ttm_pool_populate(struct ttm_tt *ttm, struct ttm_operation_ctx *ctx)
 		}
 	}
 
-	ttm->state = tt_unbound;
+	ttm_tt_set_populated(ttm);
 	return 0;
 }
 EXPORT_SYMBOL(ttm_pool_populate);
diff --git a/drivers/gpu/drm/ttm/ttm_page_alloc_dma.c b/drivers/gpu/drm/ttm/ttm_page_alloc_dma.c
index faefaaef7909..5e2df11685e7 100644
--- a/drivers/gpu/drm/ttm/ttm_page_alloc_dma.c
+++ b/drivers/gpu/drm/ttm/ttm_page_alloc_dma.c
@@ -894,7 +894,7 @@ int ttm_dma_populate(struct ttm_dma_tt *ttm_dma, struct device *dev,
 	unsigned i;
 	int ret;
 
-	if (ttm->state != tt_unpopulated)
+	if (ttm_tt_is_populated(ttm))
 		return 0;
 
 	if (ttm_check_under_lowerlimit(mem_glob, num_pages, ctx))
@@ -982,7 +982,7 @@ skip_huge:
 		}
 	}
 
-	ttm->state = tt_unbound;
+	ttm_tt_set_populated(ttm);
 	return 0;
 }
 EXPORT_SYMBOL_GPL(ttm_dma_populate);
@@ -1076,7 +1076,7 @@ void ttm_dma_unpopulate(struct ttm_dma_tt *ttm_dma, struct device *dev)
 	/* shrink pool if necessary (only on !is_cached pools)*/
 	if (npages)
 		ttm_dma_page_pool_free(pool, npages, false);
-	ttm->state = tt_unpopulated;
+	ttm_tt_set_unpopulated(ttm);
 }
 EXPORT_SYMBOL_GPL(ttm_dma_unpopulate);
 
diff --git a/drivers/gpu/drm/ttm/ttm_range_manager.c b/drivers/gpu/drm/ttm/ttm_range_manager.c
index 770c8988c139..1da0e277c511 100644
--- a/drivers/gpu/drm/ttm/ttm_range_manager.c
+++ b/drivers/gpu/drm/ttm/ttm_range_manager.c
@@ -113,10 +113,7 @@ static void ttm_range_man_free(struct ttm_resource_manager *man,
 static const struct ttm_resource_manager_func ttm_range_manager_func;
 
 int ttm_range_man_init(struct ttm_bo_device *bdev,
-		       unsigned type,
-		       uint32_t available_caching,
-		       uint32_t default_caching,
-		       bool use_tt,
+		       unsigned type, bool use_tt,
 		       unsigned long p_size)
 {
 	struct ttm_resource_manager *man;
@@ -127,8 +124,6 @@ int ttm_range_man_init(struct ttm_bo_device *bdev,
 		return -ENOMEM;
 
 	man = &rman->manager;
-	man->available_caching = available_caching;
-	man->default_caching = default_caching;
 	man->use_tt = use_tt;
 
 	man->func = &ttm_range_manager_func;
diff --git a/drivers/gpu/drm/ttm/ttm_resource.c b/drivers/gpu/drm/ttm/ttm_resource.c
index 33b642532e5c..b325b9264203 100644
--- a/drivers/gpu/drm/ttm/ttm_resource.c
+++ b/drivers/gpu/drm/ttm/ttm_resource.c
@@ -65,10 +65,7 @@ void ttm_resource_manager_init(struct ttm_resource_manager *man,
 {
 	unsigned i;
 
-	man->use_io_reserve_lru = false;
-	mutex_init(&man->io_reserve_mutex);
 	spin_lock_init(&man->move_lock);
-	INIT_LIST_HEAD(&man->io_reserve_lru);
 	man->size = p_size;
 
 	for (i = 0; i < TTM_MAX_BO_PRIORITY; ++i)
@@ -143,8 +140,6 @@ void ttm_resource_manager_debug(struct ttm_resource_manager *man,
 	drm_printf(p, "  use_type: %d\n", man->use_type);
 	drm_printf(p, "  use_tt: %d\n", man->use_tt);
 	drm_printf(p, "  size: %llu\n", man->size);
-	drm_printf(p, "  available_caching: 0x%08X\n", man->available_caching);
-	drm_printf(p, "  default_caching: 0x%08X\n", man->default_caching);
 	if (man->func && man->func->debug)
 		(*man->func->debug)(man, p);
 }
diff --git a/drivers/gpu/drm/ttm/ttm_tt.c b/drivers/gpu/drm/ttm/ttm_tt.c
index 1ccf1ef050d6..f43fa69a1e65 100644
--- a/drivers/gpu/drm/ttm/ttm_tt.c
+++ b/drivers/gpu/drm/ttm/ttm_tt.c
@@ -156,7 +156,7 @@ static int ttm_tt_set_caching(struct ttm_tt *ttm,
 	if (ttm->caching_state == c_state)
 		return 0;
 
-	if (ttm->state == tt_unpopulated) {
+	if (!ttm_tt_is_populated(ttm)) {
 		/* Change caching but don't populate */
 		ttm->caching_state = c_state;
 		return 0;
@@ -207,33 +207,31 @@ int ttm_tt_set_placement_caching(struct ttm_tt *ttm, uint32_t placement)
 }
 EXPORT_SYMBOL(ttm_tt_set_placement_caching);
 
-void ttm_tt_destroy(struct ttm_tt *ttm)
+void ttm_tt_destroy_common(struct ttm_bo_device *bdev, struct ttm_tt *ttm)
 {
-	if (ttm == NULL)
-		return;
-
-	ttm_tt_unbind(ttm);
-
-	if (ttm->state == tt_unbound)
-		ttm_tt_unpopulate(ttm);
+	ttm_tt_unpopulate(bdev, ttm);
 
 	if (!(ttm->page_flags & TTM_PAGE_FLAG_PERSISTENT_SWAP) &&
 	    ttm->swap_storage)
 		fput(ttm->swap_storage);
 
 	ttm->swap_storage = NULL;
-	ttm->func->destroy(ttm);
+}
+EXPORT_SYMBOL(ttm_tt_destroy_common);
+
+void ttm_tt_destroy(struct ttm_bo_device *bdev, struct ttm_tt *ttm)
+{
+	bdev->driver->ttm_tt_destroy(bdev, ttm);
 }
 
 static void ttm_tt_init_fields(struct ttm_tt *ttm,
 			       struct ttm_buffer_object *bo,
 			       uint32_t page_flags)
 {
-	ttm->bdev = bo->bdev;
 	ttm->num_pages = bo->num_pages;
 	ttm->caching_state = tt_cached;
 	ttm->page_flags = page_flags;
-	ttm->state = tt_unpopulated;
+	ttm_tt_set_unpopulated(ttm);
 	ttm->swap_storage = NULL;
 	ttm->sg = bo->sg;
 }
@@ -308,39 +306,6 @@ void ttm_dma_tt_fini(struct ttm_dma_tt *ttm_dma)
 }
 EXPORT_SYMBOL(ttm_dma_tt_fini);
 
-void ttm_tt_unbind(struct ttm_tt *ttm)
-{
-	if (ttm->state == tt_bound) {
-		ttm->func->unbind(ttm);
-		ttm->state = tt_unbound;
-	}
-}
-
-int ttm_tt_bind(struct ttm_tt *ttm, struct ttm_resource *bo_mem,
-		struct ttm_operation_ctx *ctx)
-{
-	int ret = 0;
-
-	if (!ttm)
-		return -EINVAL;
-
-	if (ttm->state == tt_bound)
-		return 0;
-
-	ret = ttm_tt_populate(ttm, ctx);
-	if (ret)
-		return ret;
-
-	ret = ttm->func->bind(ttm, bo_mem);
-	if (unlikely(ret != 0))
-		return ret;
-
-	ttm->state = tt_bound;
-
-	return 0;
-}
-EXPORT_SYMBOL(ttm_tt_bind);
-
 int ttm_tt_swapin(struct ttm_tt *ttm)
 {
 	struct address_space *swap_space;
@@ -383,7 +348,8 @@ out_err:
 	return ret;
 }
 
-int ttm_tt_swapout(struct ttm_tt *ttm, struct file *persistent_swap_storage)
+int ttm_tt_swapout(struct ttm_bo_device *bdev,
+		   struct ttm_tt *ttm, struct file *persistent_swap_storage)
 {
 	struct address_space *swap_space;
 	struct file *swap_storage;
@@ -392,7 +358,6 @@ int ttm_tt_swapout(struct ttm_tt *ttm, struct file *persistent_swap_storage)
 	int i;
 	int ret = -ENOMEM;
 
-	BUG_ON(ttm->state != tt_unbound && ttm->state != tt_unpopulated);
 	BUG_ON(ttm->caching_state != tt_cached);
 
 	if (!persistent_swap_storage) {
@@ -429,7 +394,7 @@ int ttm_tt_swapout(struct ttm_tt *ttm, struct file *persistent_swap_storage)
 		put_page(to_page);
 	}
 
-	ttm_tt_unpopulate(ttm);
+	ttm_tt_unpopulate(bdev, ttm);
 	ttm->swap_storage = swap_storage;
 	ttm->page_flags |= TTM_PAGE_FLAG_SWAPPED;
 	if (persistent_swap_storage)
@@ -443,7 +408,7 @@ out_err:
 	return ret;
 }
 
-static void ttm_tt_add_mapping(struct ttm_tt *ttm)
+static void ttm_tt_add_mapping(struct ttm_bo_device *bdev, struct ttm_tt *ttm)
 {
 	pgoff_t i;
 
@@ -451,24 +416,29 @@ static void ttm_tt_add_mapping(struct ttm_tt *ttm)
 		return;
 
 	for (i = 0; i < ttm->num_pages; ++i)
-		ttm->pages[i]->mapping = ttm->bdev->dev_mapping;
+		ttm->pages[i]->mapping = bdev->dev_mapping;
 }
 
-int ttm_tt_populate(struct ttm_tt *ttm, struct ttm_operation_ctx *ctx)
+int ttm_tt_populate(struct ttm_bo_device *bdev,
+		    struct ttm_tt *ttm, struct ttm_operation_ctx *ctx)
 {
 	int ret;
 
-	if (ttm->state != tt_unpopulated)
+	if (!ttm)
+		return -EINVAL;
+
+	if (ttm_tt_is_populated(ttm))
 		return 0;
 
-	if (ttm->bdev->driver->ttm_tt_populate)
-		ret = ttm->bdev->driver->ttm_tt_populate(ttm, ctx);
+	if (bdev->driver->ttm_tt_populate)
+		ret = bdev->driver->ttm_tt_populate(bdev, ttm, ctx);
 	else
 		ret = ttm_pool_populate(ttm, ctx);
 	if (!ret)
-		ttm_tt_add_mapping(ttm);
+		ttm_tt_add_mapping(bdev, ttm);
 	return ret;
 }
+EXPORT_SYMBOL(ttm_tt_populate);
 
 static void ttm_tt_clear_mapping(struct ttm_tt *ttm)
 {
@@ -484,14 +454,15 @@ static void ttm_tt_clear_mapping(struct ttm_tt *ttm)
 	}
 }
 
-void ttm_tt_unpopulate(struct ttm_tt *ttm)
+void ttm_tt_unpopulate(struct ttm_bo_device *bdev,
+		       struct ttm_tt *ttm)
 {
-	if (ttm->state == tt_unpopulated)
+	if (!ttm_tt_is_populated(ttm))
 		return;
 
 	ttm_tt_clear_mapping(ttm);
-	if (ttm->bdev->driver->ttm_tt_unpopulate)
-		ttm->bdev->driver->ttm_tt_unpopulate(ttm);
+	if (bdev->driver->ttm_tt_unpopulate)
+		bdev->driver->ttm_tt_unpopulate(bdev, ttm);
 	else
 		ttm_pool_unpopulate(ttm);
 }
diff --git a/drivers/gpu/drm/vboxvideo/vbox_mode.c b/drivers/gpu/drm/vboxvideo/vbox_mode.c
index d9a5af62af89..4fcc0a542b8a 100644
--- a/drivers/gpu/drm/vboxvideo/vbox_mode.c
+++ b/drivers/gpu/drm/vboxvideo/vbox_mode.c
@@ -397,11 +397,13 @@ static void vbox_cursor_atomic_update(struct drm_plane *plane,
 
 	vbox_crtc->cursor_enabled = true;
 
-	/* pinning is done in prepare/cleanup framebuffer */
-	src = drm_gem_vram_kmap(gbo, true, NULL);
+	src = drm_gem_vram_vmap(gbo);
 	if (IS_ERR(src)) {
+		/*
+		 * BUG: we should have pinned the BO in prepare_fb().
+		 */
 		mutex_unlock(&vbox->hw_mutex);
-		DRM_WARN("Could not kmap cursor bo, skipping update\n");
+		DRM_WARN("Could not map cursor bo, skipping update\n");
 		return;
 	}
 
@@ -414,7 +416,7 @@ static void vbox_cursor_atomic_update(struct drm_plane *plane,
 	data_size = width * height * 4 + mask_size;
 
 	copy_cursor_image(src, vbox->cursor_data, width, height, mask_size);
-	drm_gem_vram_kunmap(gbo);
+	drm_gem_vram_vunmap(gbo, src);
 
 	flags = VBOX_MOUSE_POINTER_VISIBLE | VBOX_MOUSE_POINTER_SHAPE |
 		VBOX_MOUSE_POINTER_ALPHA;
diff --git a/drivers/gpu/drm/vc4/Makefile b/drivers/gpu/drm/vc4/Makefile
index b303703bc7f3..d0163e18e9ca 100644
--- a/drivers/gpu/drm/vc4/Makefile
+++ b/drivers/gpu/drm/vc4/Makefile
@@ -12,6 +12,7 @@ vc4-y := \
 	vc4_kms.o \
 	vc4_gem.o \
 	vc4_hdmi.o \
+	vc4_hdmi_phy.o \
 	vc4_vec.o \
 	vc4_hvs.o \
 	vc4_irq.o \
diff --git a/drivers/gpu/drm/vc4/vc4_crtc.c b/drivers/gpu/drm/vc4/vc4_crtc.c
index 6d8fa6118fc1..a393f93390a2 100644
--- a/drivers/gpu/drm/vc4/vc4_crtc.c
+++ b/drivers/gpu/drm/vc4/vc4_crtc.c
@@ -65,6 +65,20 @@ static const struct debugfs_reg32 crtc_regs[] = {
 	VC4_REG32(PV_HACT_ACT),
 };
 
+static unsigned int
+vc4_crtc_get_cob_allocation(struct vc4_dev *vc4, unsigned int channel)
+{
+	u32 dispbase = HVS_READ(SCALER_DISPBASEX(channel));
+	/* Top/base are supposed to be 4-pixel aligned, but the
+	 * Raspberry Pi firmware fills the low bits (which are
+	 * presumably ignored).
+	 */
+	u32 top = VC4_GET_FIELD(dispbase, SCALER_DISPBASEX_TOP) & ~3;
+	u32 base = VC4_GET_FIELD(dispbase, SCALER_DISPBASEX_BASE) & ~3;
+
+	return top - base + 4;
+}
+
 static bool vc4_crtc_get_scanout_position(struct drm_crtc *crtc,
 					  bool in_vblank_irq,
 					  int *vpos, int *hpos,
@@ -74,6 +88,8 @@ static bool vc4_crtc_get_scanout_position(struct drm_crtc *crtc,
 	struct drm_device *dev = crtc->dev;
 	struct vc4_dev *vc4 = to_vc4_dev(dev);
 	struct vc4_crtc *vc4_crtc = to_vc4_crtc(crtc);
+	struct vc4_crtc_state *vc4_crtc_state = to_vc4_crtc_state(crtc->state);
+	unsigned int cob_size;
 	u32 val;
 	int fifo_lines;
 	int vblank_lines;
@@ -89,7 +105,7 @@ static bool vc4_crtc_get_scanout_position(struct drm_crtc *crtc,
 	 * Read vertical scanline which is currently composed for our
 	 * pixelvalve by the HVS, and also the scaler status.
 	 */
-	val = HVS_READ(SCALER_DISPSTATX(vc4_crtc->channel));
+	val = HVS_READ(SCALER_DISPSTATX(vc4_crtc_state->assigned_channel));
 
 	/* Get optional system timestamp after query. */
 	if (etime)
@@ -109,8 +125,9 @@ static bool vc4_crtc_get_scanout_position(struct drm_crtc *crtc,
 			*hpos += mode->crtc_htotal / 2;
 	}
 
+	cob_size = vc4_crtc_get_cob_allocation(vc4, vc4_crtc_state->assigned_channel);
 	/* This is the offset we need for translating hvs -> pv scanout pos. */
-	fifo_lines = vc4_crtc->cob_size / mode->crtc_hdisplay;
+	fifo_lines = cob_size / mode->crtc_hdisplay;
 
 	if (fifo_lines > 0)
 		ret = true;
@@ -189,10 +206,22 @@ void vc4_crtc_destroy(struct drm_crtc *crtc)
 	drm_crtc_cleanup(crtc);
 }
 
-static u32 vc4_get_fifo_full_level(u32 format)
+static u32 vc4_get_fifo_full_level(struct vc4_crtc *vc4_crtc, u32 format)
 {
-	static const u32 fifo_len_bytes = 64;
+	const struct vc4_crtc_data *crtc_data = vc4_crtc_to_vc4_crtc_data(vc4_crtc);
+	const struct vc4_pv_data *pv_data = vc4_crtc_to_vc4_pv_data(vc4_crtc);
+	u32 fifo_len_bytes = pv_data->fifo_depth;
 
+	/*
+	 * Pixels are pulled from the HVS if the number of bytes is
+	 * lower than the FIFO full level.
+	 *
+	 * The latency of the pixel fetch mechanism is 6 pixels, so we
+	 * need to convert those 6 pixels in bytes, depending on the
+	 * format, and then subtract that from the length of the FIFO
+	 * to make sure we never end up in a situation where the FIFO
+	 * is full.
+	 */
 	switch (format) {
 	case PV_CONTROL_FORMAT_DSIV_16:
 	case PV_CONTROL_FORMAT_DSIC_16:
@@ -202,10 +231,30 @@ static u32 vc4_get_fifo_full_level(u32 format)
 	case PV_CONTROL_FORMAT_24:
 	case PV_CONTROL_FORMAT_DSIV_24:
 	default:
+		/*
+		 * For some reason, the pixelvalve4 doesn't work with
+		 * the usual formula and will only work with 32.
+		 */
+		if (crtc_data->hvs_output == 5)
+			return 32;
+
 		return fifo_len_bytes - 3 * HVS_FIFO_LATENCY_PIX;
 	}
 }
 
+static u32 vc4_crtc_get_fifo_full_level_bits(struct vc4_crtc *vc4_crtc,
+					     u32 format)
+{
+	u32 level = vc4_get_fifo_full_level(vc4_crtc, format);
+	u32 ret = 0;
+
+	ret |= VC4_SET_FIELD((level >> 6),
+			     PV5_CONTROL_FIFO_LEVEL_HIGH);
+
+	return ret | VC4_SET_FIELD(level & 0x3f,
+				   PV_CONTROL_FIFO_LEVEL);
+}
+
 /*
  * Returns the encoder attached to the CRTC.
  *
@@ -230,11 +279,23 @@ static struct drm_encoder *vc4_get_crtc_encoder(struct drm_crtc *crtc)
 	return NULL;
 }
 
+static void vc4_crtc_pixelvalve_reset(struct drm_crtc *crtc)
+{
+	struct vc4_crtc *vc4_crtc = to_vc4_crtc(crtc);
+
+	/* The PV needs to be disabled before it can be flushed */
+	CRTC_WRITE(PV_CONTROL, CRTC_READ(PV_CONTROL) & ~PV_CONTROL_EN);
+	CRTC_WRITE(PV_CONTROL, CRTC_READ(PV_CONTROL) | PV_CONTROL_FIFO_CLR);
+}
+
 static void vc4_crtc_config_pv(struct drm_crtc *crtc)
 {
+	struct drm_device *dev = crtc->dev;
+	struct vc4_dev *vc4 = to_vc4_dev(dev);
 	struct drm_encoder *encoder = vc4_get_crtc_encoder(crtc);
 	struct vc4_encoder *vc4_encoder = to_vc4_encoder(encoder);
 	struct vc4_crtc *vc4_crtc = to_vc4_crtc(crtc);
+	const struct vc4_pv_data *pv_data = vc4_crtc_to_vc4_pv_data(vc4_crtc);
 	struct drm_crtc_state *state = crtc->state;
 	struct drm_display_mode *mode = &state->adjusted_mode;
 	bool interlace = mode->flags & DRM_MODE_FLAG_INTERLACE;
@@ -242,24 +303,29 @@ static void vc4_crtc_config_pv(struct drm_crtc *crtc)
 	bool is_dsi = (vc4_encoder->type == VC4_ENCODER_TYPE_DSI0 ||
 		       vc4_encoder->type == VC4_ENCODER_TYPE_DSI1);
 	u32 format = is_dsi ? PV_CONTROL_FORMAT_DSIV_24 : PV_CONTROL_FORMAT_24;
+	u8 ppc = pv_data->pixels_per_clock;
+	bool debug_dump_regs = false;
 
-	/* Reset the PV fifo. */
-	CRTC_WRITE(PV_CONTROL, 0);
-	CRTC_WRITE(PV_CONTROL, PV_CONTROL_FIFO_CLR | PV_CONTROL_EN);
-	CRTC_WRITE(PV_CONTROL, 0);
+	if (debug_dump_regs) {
+		struct drm_printer p = drm_info_printer(&vc4_crtc->pdev->dev);
+		dev_info(&vc4_crtc->pdev->dev, "CRTC %d regs before:\n",
+			 drm_crtc_index(crtc));
+		drm_print_regset32(&p, &vc4_crtc->regset);
+	}
+
+	vc4_crtc_pixelvalve_reset(crtc);
 
 	CRTC_WRITE(PV_HORZA,
-		   VC4_SET_FIELD((mode->htotal -
-				  mode->hsync_end) * pixel_rep,
+		   VC4_SET_FIELD((mode->htotal - mode->hsync_end) * pixel_rep / ppc,
 				 PV_HORZA_HBP) |
-		   VC4_SET_FIELD((mode->hsync_end -
-				  mode->hsync_start) * pixel_rep,
+		   VC4_SET_FIELD((mode->hsync_end - mode->hsync_start) * pixel_rep / ppc,
 				 PV_HORZA_HSYNC));
+
 	CRTC_WRITE(PV_HORZB,
-		   VC4_SET_FIELD((mode->hsync_start -
-				  mode->hdisplay) * pixel_rep,
+		   VC4_SET_FIELD((mode->hsync_start - mode->hdisplay) * pixel_rep / ppc,
 				 PV_HORZB_HFP) |
-		   VC4_SET_FIELD(mode->hdisplay * pixel_rep, PV_HORZB_HACTIVE));
+		   VC4_SET_FIELD(mode->hdisplay * pixel_rep / ppc,
+				 PV_HORZB_HACTIVE));
 
 	CRTC_WRITE(PV_VERTA,
 		   VC4_SET_FIELD(mode->crtc_vtotal - mode->crtc_vsync_end,
@@ -306,35 +372,20 @@ static void vc4_crtc_config_pv(struct drm_crtc *crtc)
 	if (is_dsi)
 		CRTC_WRITE(PV_HACT_ACT, mode->hdisplay * pixel_rep);
 
-	CRTC_WRITE(PV_CONTROL,
+	if (vc4->hvs->hvs5)
+		CRTC_WRITE(PV_MUX_CFG,
+			   VC4_SET_FIELD(PV_MUX_CFG_RGB_PIXEL_MUX_MODE_NO_SWAP,
+					 PV_MUX_CFG_RGB_PIXEL_MUX_MODE));
+
+	CRTC_WRITE(PV_CONTROL, PV_CONTROL_FIFO_CLR |
+		   vc4_crtc_get_fifo_full_level_bits(vc4_crtc, format) |
 		   VC4_SET_FIELD(format, PV_CONTROL_FORMAT) |
-		   VC4_SET_FIELD(vc4_get_fifo_full_level(format),
-				 PV_CONTROL_FIFO_LEVEL) |
 		   VC4_SET_FIELD(pixel_rep - 1, PV_CONTROL_PIXEL_REP) |
 		   PV_CONTROL_CLR_AT_START |
 		   PV_CONTROL_TRIGGER_UNDERFLOW |
 		   PV_CONTROL_WAIT_HSTART |
 		   VC4_SET_FIELD(vc4_encoder->clock_select,
-				 PV_CONTROL_CLK_SELECT) |
-		   PV_CONTROL_FIFO_CLR |
-		   PV_CONTROL_EN);
-}
-
-static void vc4_crtc_mode_set_nofb(struct drm_crtc *crtc)
-{
-	struct vc4_crtc *vc4_crtc = to_vc4_crtc(crtc);
-	bool debug_dump_regs = false;
-
-	if (debug_dump_regs) {
-		struct drm_printer p = drm_info_printer(&vc4_crtc->pdev->dev);
-		dev_info(&vc4_crtc->pdev->dev, "CRTC %d regs before:\n",
-			 drm_crtc_index(crtc));
-		drm_print_regset32(&p, &vc4_crtc->regset);
-	}
-
-	vc4_crtc_config_pv(crtc);
-
-	vc4_hvs_mode_set_nofb(crtc);
+				 PV_CONTROL_CLK_SELECT));
 
 	if (debug_dump_regs) {
 		struct drm_printer p = drm_info_printer(&vc4_crtc->pdev->dev);
@@ -352,24 +403,86 @@ static void require_hvs_enabled(struct drm_device *dev)
 		     SCALER_DISPCTRL_ENABLE);
 }
 
+static int vc4_crtc_disable(struct drm_crtc *crtc, unsigned int channel)
+{
+	struct drm_encoder *encoder = vc4_get_crtc_encoder(crtc);
+	struct vc4_encoder *vc4_encoder = to_vc4_encoder(encoder);
+	struct vc4_crtc *vc4_crtc = to_vc4_crtc(crtc);
+	struct drm_device *dev = crtc->dev;
+	int ret;
+
+	CRTC_WRITE(PV_V_CONTROL,
+		   CRTC_READ(PV_V_CONTROL) & ~PV_VCONTROL_VIDEN);
+	ret = wait_for(!(CRTC_READ(PV_V_CONTROL) & PV_VCONTROL_VIDEN), 1);
+	WARN_ONCE(ret, "Timeout waiting for !PV_VCONTROL_VIDEN\n");
+
+	/*
+	 * This delay is needed to avoid to get a pixel stuck in an
+	 * unflushable FIFO between the pixelvalve and the HDMI
+	 * controllers on the BCM2711.
+	 *
+	 * Timing is fairly sensitive here, so mdelay is the safest
+	 * approach.
+	 *
+	 * If it was to be reworked, the stuck pixel happens on a
+	 * BCM2711 when changing mode with a good probability, so a
+	 * script that changes mode on a regular basis should trigger
+	 * the bug after less than 10 attempts. It manifests itself with
+	 * every pixels being shifted by one to the right, and thus the
+	 * last pixel of a line actually being displayed as the first
+	 * pixel on the next line.
+	 */
+	mdelay(20);
+
+	if (vc4_encoder && vc4_encoder->post_crtc_disable)
+		vc4_encoder->post_crtc_disable(encoder);
+
+	vc4_crtc_pixelvalve_reset(crtc);
+	vc4_hvs_stop_channel(dev, channel);
+
+	if (vc4_encoder && vc4_encoder->post_crtc_powerdown)
+		vc4_encoder->post_crtc_powerdown(encoder);
+
+	return 0;
+}
+
+int vc4_crtc_disable_at_boot(struct drm_crtc *crtc)
+{
+	struct drm_device *drm = crtc->dev;
+	struct vc4_crtc *vc4_crtc = to_vc4_crtc(crtc);
+	int channel;
+
+	if (!(of_device_is_compatible(vc4_crtc->pdev->dev.of_node,
+				      "brcm,bcm2711-pixelvalve2") ||
+	      of_device_is_compatible(vc4_crtc->pdev->dev.of_node,
+				      "brcm,bcm2711-pixelvalve4")))
+		return 0;
+
+	if (!(CRTC_READ(PV_CONTROL) & PV_CONTROL_EN))
+		return 0;
+
+	if (!(CRTC_READ(PV_V_CONTROL) & PV_VCONTROL_VIDEN))
+		return 0;
+
+	channel = vc4_hvs_get_fifo_from_output(drm, vc4_crtc->data->hvs_output);
+	if (channel < 0)
+		return 0;
+
+	return vc4_crtc_disable(crtc, channel);
+}
+
 static void vc4_crtc_atomic_disable(struct drm_crtc *crtc,
 				    struct drm_crtc_state *old_state)
 {
+	struct vc4_crtc_state *old_vc4_state = to_vc4_crtc_state(old_state);
 	struct drm_device *dev = crtc->dev;
-	struct vc4_crtc *vc4_crtc = to_vc4_crtc(crtc);
-	int ret;
 
 	require_hvs_enabled(dev);
 
 	/* Disable vblank irq handling before crtc is disabled. */
 	drm_crtc_vblank_off(crtc);
 
-	CRTC_WRITE(PV_V_CONTROL,
-		   CRTC_READ(PV_V_CONTROL) & ~PV_VCONTROL_VIDEN);
-	ret = wait_for(!(CRTC_READ(PV_V_CONTROL) & PV_VCONTROL_VIDEN), 1);
-	WARN_ONCE(ret, "Timeout waiting for !PV_VCONTROL_VIDEN\n");
-
-	vc4_hvs_atomic_disable(crtc, old_state);
+	vc4_crtc_disable(crtc, old_vc4_state->assigned_channel);
 
 	/*
 	 * Make sure we issue a vblank event after disabling the CRTC if
@@ -390,6 +503,8 @@ static void vc4_crtc_atomic_enable(struct drm_crtc *crtc,
 {
 	struct drm_device *dev = crtc->dev;
 	struct vc4_crtc *vc4_crtc = to_vc4_crtc(crtc);
+	struct drm_encoder *encoder = vc4_get_crtc_encoder(crtc);
+	struct vc4_encoder *vc4_encoder = to_vc4_encoder(encoder);
 
 	require_hvs_enabled(dev);
 
@@ -400,11 +515,24 @@ static void vc4_crtc_atomic_enable(struct drm_crtc *crtc,
 
 	vc4_hvs_atomic_enable(crtc, old_state);
 
+	if (vc4_encoder->pre_crtc_configure)
+		vc4_encoder->pre_crtc_configure(encoder);
+
+	vc4_crtc_config_pv(crtc);
+
+	CRTC_WRITE(PV_CONTROL, CRTC_READ(PV_CONTROL) | PV_CONTROL_EN);
+
+	if (vc4_encoder->pre_crtc_enable)
+		vc4_encoder->pre_crtc_enable(encoder);
+
 	/* When feeding the transposer block the pixelvalve is unneeded and
 	 * should not be enabled.
 	 */
 	CRTC_WRITE(PV_V_CONTROL,
 		   CRTC_READ(PV_V_CONTROL) | PV_VCONTROL_VIDEN);
+
+	if (vc4_encoder->post_crtc_enable)
+		vc4_encoder->post_crtc_enable(encoder);
 }
 
 static enum drm_mode_status vc4_crtc_mode_valid(struct drm_crtc *crtc,
@@ -499,7 +627,7 @@ static void vc4_crtc_handle_page_flip(struct vc4_crtc *vc4_crtc)
 	struct drm_device *dev = crtc->dev;
 	struct vc4_dev *vc4 = to_vc4_dev(dev);
 	struct vc4_crtc_state *vc4_state = to_vc4_crtc_state(crtc->state);
-	u32 chan = vc4_crtc->channel;
+	u32 chan = vc4_state->assigned_channel;
 	unsigned long flags;
 
 	spin_lock_irqsave(&dev->event_lock, flags);
@@ -516,7 +644,7 @@ static void vc4_crtc_handle_page_flip(struct vc4_crtc *vc4_crtc)
 		 * the CRTC and encoder already reconfigured, leading to
 		 * underruns. This can be seen when reconfiguring the CRTC.
 		 */
-		vc4_hvs_unmask_underrun(dev, vc4_crtc->channel);
+		vc4_hvs_unmask_underrun(dev, chan);
 	}
 	spin_unlock_irqrestore(&dev->event_lock, flags);
 }
@@ -698,6 +826,7 @@ struct drm_crtc_state *vc4_crtc_duplicate_state(struct drm_crtc *crtc)
 	old_vc4_state = to_vc4_crtc_state(crtc->state);
 	vc4_state->feed_txp = old_vc4_state->feed_txp;
 	vc4_state->margins = old_vc4_state->margins;
+	vc4_state->assigned_channel = old_vc4_state->assigned_channel;
 
 	__drm_atomic_helper_crtc_duplicate_state(crtc, &vc4_state->base);
 	return &vc4_state->base;
@@ -747,7 +876,6 @@ static const struct drm_crtc_funcs vc4_crtc_funcs = {
 };
 
 static const struct drm_crtc_helper_funcs vc4_crtc_helper_funcs = {
-	.mode_set_nofb = vc4_crtc_mode_set_nofb,
 	.mode_valid = vc4_crtc_mode_valid,
 	.atomic_check = vc4_crtc_atomic_check,
 	.atomic_flush = vc4_hvs_atomic_flush,
@@ -758,9 +886,12 @@ static const struct drm_crtc_helper_funcs vc4_crtc_helper_funcs = {
 
 static const struct vc4_pv_data bcm2835_pv0_data = {
 	.base = {
-		.hvs_channel = 0,
+		.hvs_available_channels = BIT(0),
+		.hvs_output = 0,
 	},
 	.debugfs_name = "crtc0_regs",
+	.fifo_depth = 64,
+	.pixels_per_clock = 1,
 	.encoder_types = {
 		[PV_CONTROL_CLK_SELECT_DSI] = VC4_ENCODER_TYPE_DSI0,
 		[PV_CONTROL_CLK_SELECT_DPI_SMI_HDMI] = VC4_ENCODER_TYPE_DPI,
@@ -769,9 +900,12 @@ static const struct vc4_pv_data bcm2835_pv0_data = {
 
 static const struct vc4_pv_data bcm2835_pv1_data = {
 	.base = {
-		.hvs_channel = 2,
+		.hvs_available_channels = BIT(2),
+		.hvs_output = 2,
 	},
 	.debugfs_name = "crtc1_regs",
+	.fifo_depth = 64,
+	.pixels_per_clock = 1,
 	.encoder_types = {
 		[PV_CONTROL_CLK_SELECT_DSI] = VC4_ENCODER_TYPE_DSI1,
 		[PV_CONTROL_CLK_SELECT_DPI_SMI_HDMI] = VC4_ENCODER_TYPE_SMI,
@@ -780,19 +914,94 @@ static const struct vc4_pv_data bcm2835_pv1_data = {
 
 static const struct vc4_pv_data bcm2835_pv2_data = {
 	.base = {
-		.hvs_channel = 1,
+		.hvs_available_channels = BIT(1),
+		.hvs_output = 1,
 	},
 	.debugfs_name = "crtc2_regs",
+	.fifo_depth = 64,
+	.pixels_per_clock = 1,
 	.encoder_types = {
-		[PV_CONTROL_CLK_SELECT_DPI_SMI_HDMI] = VC4_ENCODER_TYPE_HDMI,
+		[PV_CONTROL_CLK_SELECT_DPI_SMI_HDMI] = VC4_ENCODER_TYPE_HDMI0,
 		[PV_CONTROL_CLK_SELECT_VEC] = VC4_ENCODER_TYPE_VEC,
 	},
 };
 
+static const struct vc4_pv_data bcm2711_pv0_data = {
+	.base = {
+		.hvs_available_channels = BIT(0),
+		.hvs_output = 0,
+	},
+	.debugfs_name = "crtc0_regs",
+	.fifo_depth = 64,
+	.pixels_per_clock = 1,
+	.encoder_types = {
+		[0] = VC4_ENCODER_TYPE_DSI0,
+		[1] = VC4_ENCODER_TYPE_DPI,
+	},
+};
+
+static const struct vc4_pv_data bcm2711_pv1_data = {
+	.base = {
+		.hvs_available_channels = BIT(0) | BIT(1) | BIT(2),
+		.hvs_output = 3,
+	},
+	.debugfs_name = "crtc1_regs",
+	.fifo_depth = 64,
+	.pixels_per_clock = 1,
+	.encoder_types = {
+		[0] = VC4_ENCODER_TYPE_DSI1,
+		[1] = VC4_ENCODER_TYPE_SMI,
+	},
+};
+
+static const struct vc4_pv_data bcm2711_pv2_data = {
+	.base = {
+		.hvs_available_channels = BIT(0) | BIT(1) | BIT(2),
+		.hvs_output = 4,
+	},
+	.debugfs_name = "crtc2_regs",
+	.fifo_depth = 256,
+	.pixels_per_clock = 2,
+	.encoder_types = {
+		[0] = VC4_ENCODER_TYPE_HDMI0,
+	},
+};
+
+static const struct vc4_pv_data bcm2711_pv3_data = {
+	.base = {
+		.hvs_available_channels = BIT(1),
+		.hvs_output = 1,
+	},
+	.debugfs_name = "crtc3_regs",
+	.fifo_depth = 64,
+	.pixels_per_clock = 1,
+	.encoder_types = {
+		[0] = VC4_ENCODER_TYPE_VEC,
+	},
+};
+
+static const struct vc4_pv_data bcm2711_pv4_data = {
+	.base = {
+		.hvs_available_channels = BIT(0) | BIT(1) | BIT(2),
+		.hvs_output = 5,
+	},
+	.debugfs_name = "crtc4_regs",
+	.fifo_depth = 64,
+	.pixels_per_clock = 2,
+	.encoder_types = {
+		[0] = VC4_ENCODER_TYPE_HDMI1,
+	},
+};
+
 static const struct of_device_id vc4_crtc_dt_match[] = {
 	{ .compatible = "brcm,bcm2835-pixelvalve0", .data = &bcm2835_pv0_data },
 	{ .compatible = "brcm,bcm2835-pixelvalve1", .data = &bcm2835_pv1_data },
 	{ .compatible = "brcm,bcm2835-pixelvalve2", .data = &bcm2835_pv2_data },
+	{ .compatible = "brcm,bcm2711-pixelvalve0", .data = &bcm2711_pv0_data },
+	{ .compatible = "brcm,bcm2711-pixelvalve1", .data = &bcm2711_pv1_data },
+	{ .compatible = "brcm,bcm2711-pixelvalve2", .data = &bcm2711_pv2_data },
+	{ .compatible = "brcm,bcm2711-pixelvalve3", .data = &bcm2711_pv3_data },
+	{ .compatible = "brcm,bcm2711-pixelvalve4", .data = &bcm2711_pv4_data },
 	{}
 };
 
@@ -819,26 +1028,11 @@ static void vc4_set_crtc_possible_masks(struct drm_device *drm,
 	}
 }
 
-static void
-vc4_crtc_get_cob_allocation(struct vc4_crtc *vc4_crtc)
-{
-	struct drm_device *drm = vc4_crtc->base.dev;
-	struct vc4_dev *vc4 = to_vc4_dev(drm);
-	u32 dispbase = HVS_READ(SCALER_DISPBASEX(vc4_crtc->channel));
-	/* Top/base are supposed to be 4-pixel aligned, but the
-	 * Raspberry Pi firmware fills the low bits (which are
-	 * presumably ignored).
-	 */
-	u32 top = VC4_GET_FIELD(dispbase, SCALER_DISPBASEX_TOP) & ~3;
-	u32 base = VC4_GET_FIELD(dispbase, SCALER_DISPBASEX_BASE) & ~3;
-
-	vc4_crtc->cob_size = top - base + 4;
-}
-
 int vc4_crtc_init(struct drm_device *drm, struct vc4_crtc *vc4_crtc,
 		  const struct drm_crtc_funcs *crtc_funcs,
 		  const struct drm_crtc_helper_funcs *crtc_helper_funcs)
 {
+	struct vc4_dev *vc4 = to_vc4_dev(drm);
 	struct drm_crtc *crtc = &vc4_crtc->base;
 	struct drm_plane *primary_plane;
 	unsigned int i;
@@ -858,15 +1052,17 @@ int vc4_crtc_init(struct drm_device *drm, struct vc4_crtc *vc4_crtc,
 	drm_crtc_init_with_planes(drm, crtc, primary_plane, NULL,
 				  crtc_funcs, NULL);
 	drm_crtc_helper_add(crtc, crtc_helper_funcs);
-	vc4_crtc->channel = vc4_crtc->data->hvs_channel;
-	drm_mode_crtc_set_gamma_size(crtc, ARRAY_SIZE(vc4_crtc->lut_r));
-	drm_crtc_enable_color_mgmt(crtc, 0, false, crtc->gamma_size);
 
-	/* We support CTM, but only for one CRTC at a time. It's therefore
-	 * implemented as private driver state in vc4_kms, not here.
-	 */
-	drm_crtc_enable_color_mgmt(crtc, 0, true, crtc->gamma_size);
-	vc4_crtc_get_cob_allocation(vc4_crtc);
+	if (!vc4->hvs->hvs5) {
+		drm_mode_crtc_set_gamma_size(crtc, ARRAY_SIZE(vc4_crtc->lut_r));
+
+		drm_crtc_enable_color_mgmt(crtc, 0, false, crtc->gamma_size);
+
+		/* We support CTM, but only for one CRTC at a time. It's therefore
+		 * implemented as private driver state in vc4_kms, not here.
+		 */
+		drm_crtc_enable_color_mgmt(crtc, 0, true, crtc->gamma_size);
+	}
 
 	for (i = 0; i < crtc->gamma_size; i++) {
 		vc4_crtc->lut_r[i] = i;
@@ -915,7 +1111,9 @@ static int vc4_crtc_bind(struct device *dev, struct device *master, void *data)
 	CRTC_WRITE(PV_INTEN, 0);
 	CRTC_WRITE(PV_INTSTAT, PV_INT_VFP_START);
 	ret = devm_request_irq(dev, platform_get_irq(pdev, 0),
-			       vc4_crtc_irq_handler, 0, "vc4 crtc", vc4_crtc);
+			       vc4_crtc_irq_handler,
+			       IRQF_SHARED,
+			       "vc4 crtc", vc4_crtc);
 	if (ret)
 		goto err_destroy_planes;
 
diff --git a/drivers/gpu/drm/vc4/vc4_drv.c b/drivers/gpu/drm/vc4/vc4_drv.c
index 38343d2fb4fb..f1a5fd5dab6f 100644
--- a/drivers/gpu/drm/vc4/vc4_drv.c
+++ b/drivers/gpu/drm/vc4/vc4_drv.c
@@ -252,6 +252,7 @@ static int vc4_drm_bind(struct device *dev)
 	struct drm_device *drm;
 	struct vc4_dev *vc4;
 	struct device_node *node;
+	struct drm_crtc *crtc;
 	int ret = 0;
 
 	dev->coherent_dma_mask = DMA_BIT_MASK(32);
@@ -298,6 +299,9 @@ static int vc4_drm_bind(struct device *dev)
 	if (ret < 0)
 		goto unbind_all;
 
+	drm_for_each_crtc(crtc, drm)
+		vc4_crtc_disable_at_boot(crtc);
+
 	ret = drm_dev_register(drm, 0);
 	if (ret < 0)
 		goto unbind_all;
@@ -368,6 +372,7 @@ static int vc4_platform_drm_remove(struct platform_device *pdev)
 }
 
 static const struct of_device_id vc4_of_match[] = {
+	{ .compatible = "brcm,bcm2711-vc5", },
 	{ .compatible = "brcm,bcm2835-vc4", },
 	{ .compatible = "brcm,cygnus-vc4", },
 	{},
diff --git a/drivers/gpu/drm/vc4/vc4_drv.h b/drivers/gpu/drm/vc4/vc4_drv.h
index fa19160c801f..8c8d96b6289f 100644
--- a/drivers/gpu/drm/vc4/vc4_drv.h
+++ b/drivers/gpu/drm/vc4/vc4_drv.h
@@ -73,7 +73,6 @@ struct vc4_perfmon {
 struct vc4_dev {
 	struct drm_device *dev;
 
-	struct vc4_hdmi *hdmi;
 	struct vc4_hvs *hvs;
 	struct vc4_v3d *v3d;
 	struct vc4_dpi *dpi;
@@ -201,6 +200,9 @@ struct vc4_dev {
 
 	int power_refcount;
 
+	/* Set to true when the load tracker is supported. */
+	bool load_tracker_available;
+
 	/* Set to true when the load tracker is active. */
 	bool load_tracker_enabled;
 
@@ -320,6 +322,8 @@ struct vc4_hvs {
 	void __iomem *regs;
 	u32 __iomem *dlist;
 
+	struct clk *core_clk;
+
 	/* Memory manager for CRTCs to allocate space in the display
 	 * list.  Units are dwords.
 	 */
@@ -329,7 +333,11 @@ struct vc4_hvs {
 	spinlock_t mm_lock;
 
 	struct drm_mm_node mitchell_netravali_filter;
+
 	struct debugfs_regset32 regset;
+
+	/* HVS version 5 flag, therefore requires updated dlist structures */
+	bool hvs5;
 };
 
 struct vc4_plane {
@@ -420,7 +428,8 @@ to_vc4_plane_state(struct drm_plane_state *state)
 
 enum vc4_encoder_type {
 	VC4_ENCODER_TYPE_NONE,
-	VC4_ENCODER_TYPE_HDMI,
+	VC4_ENCODER_TYPE_HDMI0,
+	VC4_ENCODER_TYPE_HDMI1,
 	VC4_ENCODER_TYPE_VEC,
 	VC4_ENCODER_TYPE_DSI0,
 	VC4_ENCODER_TYPE_DSI1,
@@ -432,6 +441,13 @@ struct vc4_encoder {
 	struct drm_encoder base;
 	enum vc4_encoder_type type;
 	u32 clock_select;
+
+	void (*pre_crtc_configure)(struct drm_encoder *encoder);
+	void (*pre_crtc_enable)(struct drm_encoder *encoder);
+	void (*post_crtc_enable)(struct drm_encoder *encoder);
+
+	void (*post_crtc_disable)(struct drm_encoder *encoder);
+	void (*post_crtc_powerdown)(struct drm_encoder *encoder);
 };
 
 static inline struct vc4_encoder *
@@ -441,13 +457,22 @@ to_vc4_encoder(struct drm_encoder *encoder)
 }
 
 struct vc4_crtc_data {
-	/* Which channel of the HVS this pixelvalve sources from. */
-	int hvs_channel;
+	/* Bitmask of channels (FIFOs) of the HVS that the output can source from */
+	unsigned int hvs_available_channels;
+
+	/* Which output of the HVS this pixelvalve sources from. */
+	int hvs_output;
 };
 
 struct vc4_pv_data {
 	struct vc4_crtc_data	base;
 
+	/* Depth of the PixelValve FIFO in bytes */
+	unsigned int fifo_depth;
+
+	/* Number of pixels output per clock period */
+	u8 pixels_per_clock;
+
 	enum vc4_encoder_type encoder_types[4];
 	const char *debugfs_name;
 
@@ -462,14 +487,9 @@ struct vc4_crtc {
 	/* Timestamp at start of vblank irq - unaffected by lock delays. */
 	ktime_t t_vblank;
 
-	/* Which HVS channel we're using for our CRTC. */
-	int channel;
-
 	u8 lut_r[256];
 	u8 lut_g[256];
 	u8 lut_b[256];
-	/* Size in pixels of the COB memory allocated to this CRTC. */
-	u32 cob_size;
 
 	struct drm_pending_vblank_event *event;
 
@@ -502,6 +522,7 @@ struct vc4_crtc_state {
 	struct drm_mm_node mm;
 	bool feed_txp;
 	bool txp_armed;
+	unsigned int assigned_channel;
 
 	struct {
 		unsigned int left;
@@ -794,6 +815,7 @@ void vc4_bo_remove_from_purgeable_pool(struct vc4_bo *bo);
 
 /* vc4_crtc.c */
 extern struct platform_driver vc4_crtc_driver;
+int vc4_crtc_disable_at_boot(struct drm_crtc *crtc);
 int vc4_crtc_init(struct drm_device *drm, struct vc4_crtc *vc4_crtc,
 		  const struct drm_crtc_funcs *crtc_funcs,
 		  const struct drm_crtc_helper_funcs *crtc_helper_funcs);
@@ -888,11 +910,12 @@ void vc4_irq_reset(struct drm_device *dev);
 
 /* vc4_hvs.c */
 extern struct platform_driver vc4_hvs_driver;
+void vc4_hvs_stop_channel(struct drm_device *dev, unsigned int output);
+int vc4_hvs_get_fifo_from_output(struct drm_device *dev, unsigned int output);
 int vc4_hvs_atomic_check(struct drm_crtc *crtc, struct drm_crtc_state *state);
 void vc4_hvs_atomic_enable(struct drm_crtc *crtc, struct drm_crtc_state *old_state);
 void vc4_hvs_atomic_disable(struct drm_crtc *crtc, struct drm_crtc_state *old_state);
 void vc4_hvs_atomic_flush(struct drm_crtc *crtc, struct drm_crtc_state *state);
-void vc4_hvs_mode_set_nofb(struct drm_crtc *crtc);
 void vc4_hvs_dump_state(struct drm_device *dev);
 void vc4_hvs_unmask_underrun(struct drm_device *dev, int channel);
 void vc4_hvs_mask_underrun(struct drm_device *dev, int channel);
diff --git a/drivers/gpu/drm/vc4/vc4_hdmi.c b/drivers/gpu/drm/vc4/vc4_hdmi.c
index 15a11cd4de25..03825596a308 100644
--- a/drivers/gpu/drm/vc4/vc4_hdmi.c
+++ b/drivers/gpu/drm/vc4/vc4_hdmi.c
@@ -43,177 +43,101 @@
 #include <linux/of_platform.h>
 #include <linux/pm_runtime.h>
 #include <linux/rational.h>
+#include <linux/reset.h>
 #include <sound/dmaengine_pcm.h>
 #include <sound/pcm_drm_eld.h>
 #include <sound/pcm_params.h>
 #include <sound/soc.h>
 #include "media/cec.h"
 #include "vc4_drv.h"
+#include "vc4_hdmi.h"
+#include "vc4_hdmi_regs.h"
 #include "vc4_regs.h"
 
-#define HSM_CLOCK_FREQ 163682864
-#define CEC_CLOCK_FREQ 40000
-#define CEC_CLOCK_DIV  (HSM_CLOCK_FREQ / CEC_CLOCK_FREQ)
-
-/* HDMI audio information */
-struct vc4_hdmi_audio {
-	struct snd_soc_card card;
-	struct snd_soc_dai_link link;
-	struct snd_soc_dai_link_component cpu;
-	struct snd_soc_dai_link_component codec;
-	struct snd_soc_dai_link_component platform;
-	int samplerate;
-	int channels;
-	struct snd_dmaengine_dai_dma_data dma_data;
-	struct snd_pcm_substream *substream;
-};
+#define VC5_HDMI_HORZA_HFP_SHIFT		16
+#define VC5_HDMI_HORZA_HFP_MASK			VC4_MASK(28, 16)
+#define VC5_HDMI_HORZA_VPOS			BIT(15)
+#define VC5_HDMI_HORZA_HPOS			BIT(14)
+#define VC5_HDMI_HORZA_HAP_SHIFT		0
+#define VC5_HDMI_HORZA_HAP_MASK			VC4_MASK(13, 0)
 
-/* General HDMI hardware state. */
-struct vc4_hdmi {
-	struct platform_device *pdev;
-
-	struct drm_encoder *encoder;
-	struct drm_connector *connector;
+#define VC5_HDMI_HORZB_HBP_SHIFT		16
+#define VC5_HDMI_HORZB_HBP_MASK			VC4_MASK(26, 16)
+#define VC5_HDMI_HORZB_HSP_SHIFT		0
+#define VC5_HDMI_HORZB_HSP_MASK			VC4_MASK(10, 0)
 
-	struct vc4_hdmi_audio audio;
+#define VC5_HDMI_VERTA_VSP_SHIFT		24
+#define VC5_HDMI_VERTA_VSP_MASK			VC4_MASK(28, 24)
+#define VC5_HDMI_VERTA_VFP_SHIFT		16
+#define VC5_HDMI_VERTA_VFP_MASK			VC4_MASK(22, 16)
+#define VC5_HDMI_VERTA_VAL_SHIFT		0
+#define VC5_HDMI_VERTA_VAL_MASK			VC4_MASK(12, 0)
 
-	struct i2c_adapter *ddc;
-	void __iomem *hdmicore_regs;
-	void __iomem *hd_regs;
-	int hpd_gpio;
-	bool hpd_active_low;
+#define VC5_HDMI_VERTB_VSPO_SHIFT		16
+#define VC5_HDMI_VERTB_VSPO_MASK		VC4_MASK(29, 16)
 
-	struct cec_adapter *cec_adap;
-	struct cec_msg cec_rx_msg;
-	bool cec_tx_ok;
-	bool cec_irq_was_rx;
+# define VC4_HD_M_SW_RST			BIT(2)
+# define VC4_HD_M_ENABLE			BIT(0)
 
-	struct clk *pixel_clock;
-	struct clk *hsm_clock;
+#define CEC_CLOCK_FREQ 40000
+#define VC4_HSM_MID_CLOCK 149985000
 
-	struct debugfs_regset32 hdmi_regset;
-	struct debugfs_regset32 hd_regset;
-};
+static int vc4_hdmi_debugfs_regs(struct seq_file *m, void *unused)
+{
+	struct drm_info_node *node = (struct drm_info_node *)m->private;
+	struct vc4_hdmi *vc4_hdmi = node->info_ent->data;
+	struct drm_printer p = drm_seq_file_printer(m);
 
-#define HDMI_READ(offset) readl(vc4->hdmi->hdmicore_regs + offset)
-#define HDMI_WRITE(offset, val) writel(val, vc4->hdmi->hdmicore_regs + offset)
-#define HD_READ(offset) readl(vc4->hdmi->hd_regs + offset)
-#define HD_WRITE(offset, val) writel(val, vc4->hdmi->hd_regs + offset)
+	drm_print_regset32(&p, &vc4_hdmi->hdmi_regset);
+	drm_print_regset32(&p, &vc4_hdmi->hd_regset);
 
-/* VC4 HDMI encoder KMS struct */
-struct vc4_hdmi_encoder {
-	struct vc4_encoder base;
-	bool hdmi_monitor;
-	bool limited_rgb_range;
-};
+	return 0;
+}
 
-static inline struct vc4_hdmi_encoder *
-to_vc4_hdmi_encoder(struct drm_encoder *encoder)
+static void vc4_hdmi_reset(struct vc4_hdmi *vc4_hdmi)
 {
-	return container_of(encoder, struct vc4_hdmi_encoder, base.base);
-}
+	HDMI_WRITE(HDMI_M_CTL, VC4_HD_M_SW_RST);
+	udelay(1);
+	HDMI_WRITE(HDMI_M_CTL, 0);
 
-/* VC4 HDMI connector KMS struct */
-struct vc4_hdmi_connector {
-	struct drm_connector base;
+	HDMI_WRITE(HDMI_M_CTL, VC4_HD_M_ENABLE);
 
-	/* Since the connector is attached to just the one encoder,
-	 * this is the reference to it so we can do the best_encoder()
-	 * hook.
-	 */
-	struct drm_encoder *encoder;
-};
+	HDMI_WRITE(HDMI_SW_RESET_CONTROL,
+		   VC4_HDMI_SW_RESET_HDMI |
+		   VC4_HDMI_SW_RESET_FORMAT_DETECT);
 
-static inline struct vc4_hdmi_connector *
-to_vc4_hdmi_connector(struct drm_connector *connector)
-{
-	return container_of(connector, struct vc4_hdmi_connector, base);
+	HDMI_WRITE(HDMI_SW_RESET_CONTROL, 0);
 }
 
-static const struct debugfs_reg32 hdmi_regs[] = {
-	VC4_REG32(VC4_HDMI_CORE_REV),
-	VC4_REG32(VC4_HDMI_SW_RESET_CONTROL),
-	VC4_REG32(VC4_HDMI_HOTPLUG_INT),
-	VC4_REG32(VC4_HDMI_HOTPLUG),
-	VC4_REG32(VC4_HDMI_MAI_CHANNEL_MAP),
-	VC4_REG32(VC4_HDMI_MAI_CONFIG),
-	VC4_REG32(VC4_HDMI_MAI_FORMAT),
-	VC4_REG32(VC4_HDMI_AUDIO_PACKET_CONFIG),
-	VC4_REG32(VC4_HDMI_RAM_PACKET_CONFIG),
-	VC4_REG32(VC4_HDMI_HORZA),
-	VC4_REG32(VC4_HDMI_HORZB),
-	VC4_REG32(VC4_HDMI_FIFO_CTL),
-	VC4_REG32(VC4_HDMI_SCHEDULER_CONTROL),
-	VC4_REG32(VC4_HDMI_VERTA0),
-	VC4_REG32(VC4_HDMI_VERTA1),
-	VC4_REG32(VC4_HDMI_VERTB0),
-	VC4_REG32(VC4_HDMI_VERTB1),
-	VC4_REG32(VC4_HDMI_TX_PHY_RESET_CTL),
-	VC4_REG32(VC4_HDMI_TX_PHY_CTL0),
-
-	VC4_REG32(VC4_HDMI_CEC_CNTRL_1),
-	VC4_REG32(VC4_HDMI_CEC_CNTRL_2),
-	VC4_REG32(VC4_HDMI_CEC_CNTRL_3),
-	VC4_REG32(VC4_HDMI_CEC_CNTRL_4),
-	VC4_REG32(VC4_HDMI_CEC_CNTRL_5),
-	VC4_REG32(VC4_HDMI_CPU_STATUS),
-	VC4_REG32(VC4_HDMI_CPU_MASK_STATUS),
-
-	VC4_REG32(VC4_HDMI_CEC_RX_DATA_1),
-	VC4_REG32(VC4_HDMI_CEC_RX_DATA_2),
-	VC4_REG32(VC4_HDMI_CEC_RX_DATA_3),
-	VC4_REG32(VC4_HDMI_CEC_RX_DATA_4),
-	VC4_REG32(VC4_HDMI_CEC_TX_DATA_1),
-	VC4_REG32(VC4_HDMI_CEC_TX_DATA_2),
-	VC4_REG32(VC4_HDMI_CEC_TX_DATA_3),
-	VC4_REG32(VC4_HDMI_CEC_TX_DATA_4),
-};
-
-static const struct debugfs_reg32 hd_regs[] = {
-	VC4_REG32(VC4_HD_M_CTL),
-	VC4_REG32(VC4_HD_MAI_CTL),
-	VC4_REG32(VC4_HD_MAI_THR),
-	VC4_REG32(VC4_HD_MAI_FMT),
-	VC4_REG32(VC4_HD_MAI_SMP),
-	VC4_REG32(VC4_HD_VID_CTL),
-	VC4_REG32(VC4_HD_CSC_CTL),
-	VC4_REG32(VC4_HD_FRAME_COUNT),
-};
-
-static int vc4_hdmi_debugfs_regs(struct seq_file *m, void *unused)
+static void vc5_hdmi_reset(struct vc4_hdmi *vc4_hdmi)
 {
-	struct drm_info_node *node = (struct drm_info_node *)m->private;
-	struct drm_device *dev = node->minor->dev;
-	struct vc4_dev *vc4 = to_vc4_dev(dev);
-	struct vc4_hdmi *hdmi = vc4->hdmi;
-	struct drm_printer p = drm_seq_file_printer(m);
+	reset_control_reset(vc4_hdmi->reset);
 
-	drm_print_regset32(&p, &hdmi->hdmi_regset);
-	drm_print_regset32(&p, &hdmi->hd_regset);
+	HDMI_WRITE(HDMI_DVP_CTL, 0);
 
-	return 0;
+	HDMI_WRITE(HDMI_CLOCK_STOP,
+		   HDMI_READ(HDMI_CLOCK_STOP) | VC4_DVP_HT_CLOCK_STOP_PIXEL);
 }
 
 static enum drm_connector_status
 vc4_hdmi_connector_detect(struct drm_connector *connector, bool force)
 {
-	struct drm_device *dev = connector->dev;
-	struct vc4_dev *vc4 = to_vc4_dev(dev);
+	struct vc4_hdmi *vc4_hdmi = connector_to_vc4_hdmi(connector);
 
-	if (vc4->hdmi->hpd_gpio) {
-		if (gpio_get_value_cansleep(vc4->hdmi->hpd_gpio) ^
-		    vc4->hdmi->hpd_active_low)
+	if (vc4_hdmi->hpd_gpio) {
+		if (gpio_get_value_cansleep(vc4_hdmi->hpd_gpio) ^
+		    vc4_hdmi->hpd_active_low)
 			return connector_status_connected;
-		cec_phys_addr_invalidate(vc4->hdmi->cec_adap);
+		cec_phys_addr_invalidate(vc4_hdmi->cec_adap);
 		return connector_status_disconnected;
 	}
 
-	if (drm_probe_ddc(vc4->hdmi->ddc))
+	if (drm_probe_ddc(vc4_hdmi->ddc))
 		return connector_status_connected;
 
-	if (HDMI_READ(VC4_HDMI_HOTPLUG) & VC4_HDMI_HOTPLUG_CONNECTED)
+	if (HDMI_READ(HDMI_HOTPLUG) & VC4_HDMI_HOTPLUG_CONNECTED)
 		return connector_status_connected;
-	cec_phys_addr_invalidate(vc4->hdmi->cec_adap);
+	cec_phys_addr_invalidate(vc4_hdmi->cec_adap);
 	return connector_status_disconnected;
 }
 
@@ -225,17 +149,13 @@ static void vc4_hdmi_connector_destroy(struct drm_connector *connector)
 
 static int vc4_hdmi_connector_get_modes(struct drm_connector *connector)
 {
-	struct vc4_hdmi_connector *vc4_connector =
-		to_vc4_hdmi_connector(connector);
-	struct drm_encoder *encoder = vc4_connector->encoder;
-	struct vc4_hdmi_encoder *vc4_encoder = to_vc4_hdmi_encoder(encoder);
-	struct drm_device *dev = connector->dev;
-	struct vc4_dev *vc4 = to_vc4_dev(dev);
+	struct vc4_hdmi *vc4_hdmi = connector_to_vc4_hdmi(connector);
+	struct vc4_hdmi_encoder *vc4_encoder = &vc4_hdmi->encoder;
 	int ret = 0;
 	struct edid *edid;
 
-	edid = drm_get_edid(connector, vc4->hdmi->ddc);
-	cec_s_phys_addr_from_edid(vc4->hdmi->cec_adap, edid);
+	edid = drm_get_edid(connector, vc4_hdmi->ddc);
+	cec_s_phys_addr_from_edid(vc4_hdmi->cec_adap, edid);
 	if (!edid)
 		return -ENODEV;
 
@@ -267,32 +187,23 @@ static const struct drm_connector_helper_funcs vc4_hdmi_connector_helper_funcs =
 	.get_modes = vc4_hdmi_connector_get_modes,
 };
 
-static struct drm_connector *vc4_hdmi_connector_init(struct drm_device *dev,
-						     struct drm_encoder *encoder,
-						     struct i2c_adapter *ddc)
+static int vc4_hdmi_connector_init(struct drm_device *dev,
+				   struct vc4_hdmi *vc4_hdmi)
 {
-	struct drm_connector *connector;
-	struct vc4_hdmi_connector *hdmi_connector;
+	struct drm_connector *connector = &vc4_hdmi->connector;
+	struct drm_encoder *encoder = &vc4_hdmi->encoder.base.base;
 	int ret;
 
-	hdmi_connector = devm_kzalloc(dev->dev, sizeof(*hdmi_connector),
-				      GFP_KERNEL);
-	if (!hdmi_connector)
-		return ERR_PTR(-ENOMEM);
-	connector = &hdmi_connector->base;
-
-	hdmi_connector->encoder = encoder;
-
 	drm_connector_init_with_ddc(dev, connector,
 				    &vc4_hdmi_connector_funcs,
 				    DRM_MODE_CONNECTOR_HDMIA,
-				    ddc);
+				    vc4_hdmi->ddc);
 	drm_connector_helper_add(connector, &vc4_hdmi_connector_helper_funcs);
 
 	/* Create and attach TV margin props to this connector. */
 	ret = drm_mode_create_tv_margin_properties(dev);
 	if (ret)
-		return ERR_PTR(ret);
+		return ret;
 
 	drm_connector_attach_tv_margin_properties(connector);
 
@@ -304,35 +215,37 @@ static struct drm_connector *vc4_hdmi_connector_init(struct drm_device *dev,
 
 	drm_connector_attach_encoder(connector, encoder);
 
-	return connector;
+	return 0;
 }
 
 static int vc4_hdmi_stop_packet(struct drm_encoder *encoder,
 				enum hdmi_infoframe_type type)
 {
-	struct drm_device *dev = encoder->dev;
-	struct vc4_dev *vc4 = to_vc4_dev(dev);
+	struct vc4_hdmi *vc4_hdmi = encoder_to_vc4_hdmi(encoder);
 	u32 packet_id = type - 0x80;
 
-	HDMI_WRITE(VC4_HDMI_RAM_PACKET_CONFIG,
-		   HDMI_READ(VC4_HDMI_RAM_PACKET_CONFIG) & ~BIT(packet_id));
+	HDMI_WRITE(HDMI_RAM_PACKET_CONFIG,
+		   HDMI_READ(HDMI_RAM_PACKET_CONFIG) & ~BIT(packet_id));
 
-	return wait_for(!(HDMI_READ(VC4_HDMI_RAM_PACKET_STATUS) &
+	return wait_for(!(HDMI_READ(HDMI_RAM_PACKET_STATUS) &
 			  BIT(packet_id)), 100);
 }
 
 static void vc4_hdmi_write_infoframe(struct drm_encoder *encoder,
 				     union hdmi_infoframe *frame)
 {
-	struct drm_device *dev = encoder->dev;
-	struct vc4_dev *vc4 = to_vc4_dev(dev);
+	struct vc4_hdmi *vc4_hdmi = encoder_to_vc4_hdmi(encoder);
 	u32 packet_id = frame->any.type - 0x80;
-	u32 packet_reg = VC4_HDMI_RAM_PACKET(packet_id);
+	const struct vc4_hdmi_register *ram_packet_start =
+		&vc4_hdmi->variant->registers[HDMI_RAM_PACKET_START];
+	u32 packet_reg = ram_packet_start->offset + VC4_HDMI_PACKET_STRIDE * packet_id;
+	void __iomem *base = __vc4_hdmi_get_field_base(vc4_hdmi,
+						       ram_packet_start->reg);
 	uint8_t buffer[VC4_HDMI_PACKET_STRIDE];
 	ssize_t len, i;
 	int ret;
 
-	WARN_ONCE(!(HDMI_READ(VC4_HDMI_RAM_PACKET_CONFIG) &
+	WARN_ONCE(!(HDMI_READ(HDMI_RAM_PACKET_CONFIG) &
 		    VC4_HDMI_RAM_PACKET_ENABLE),
 		  "Packet RAM has to be on to store the packet.");
 
@@ -347,23 +260,23 @@ static void vc4_hdmi_write_infoframe(struct drm_encoder *encoder,
 	}
 
 	for (i = 0; i < len; i += 7) {
-		HDMI_WRITE(packet_reg,
-			   buffer[i + 0] << 0 |
-			   buffer[i + 1] << 8 |
-			   buffer[i + 2] << 16);
+		writel(buffer[i + 0] << 0 |
+		       buffer[i + 1] << 8 |
+		       buffer[i + 2] << 16,
+		       base + packet_reg);
 		packet_reg += 4;
 
-		HDMI_WRITE(packet_reg,
-			   buffer[i + 3] << 0 |
-			   buffer[i + 4] << 8 |
-			   buffer[i + 5] << 16 |
-			   buffer[i + 6] << 24);
+		writel(buffer[i + 3] << 0 |
+		       buffer[i + 4] << 8 |
+		       buffer[i + 5] << 16 |
+		       buffer[i + 6] << 24,
+		       base + packet_reg);
 		packet_reg += 4;
 	}
 
-	HDMI_WRITE(VC4_HDMI_RAM_PACKET_CONFIG,
-		   HDMI_READ(VC4_HDMI_RAM_PACKET_CONFIG) | BIT(packet_id));
-	ret = wait_for((HDMI_READ(VC4_HDMI_RAM_PACKET_STATUS) &
+	HDMI_WRITE(HDMI_RAM_PACKET_CONFIG,
+		   HDMI_READ(HDMI_RAM_PACKET_CONFIG) | BIT(packet_id));
+	ret = wait_for((HDMI_READ(HDMI_RAM_PACKET_STATUS) &
 			BIT(packet_id)), 100);
 	if (ret)
 		DRM_ERROR("Failed to wait for infoframe to start: %d\n", ret);
@@ -371,24 +284,24 @@ static void vc4_hdmi_write_infoframe(struct drm_encoder *encoder,
 
 static void vc4_hdmi_set_avi_infoframe(struct drm_encoder *encoder)
 {
+	struct vc4_hdmi *vc4_hdmi = encoder_to_vc4_hdmi(encoder);
 	struct vc4_hdmi_encoder *vc4_encoder = to_vc4_hdmi_encoder(encoder);
-	struct vc4_dev *vc4 = encoder->dev->dev_private;
-	struct vc4_hdmi *hdmi = vc4->hdmi;
-	struct drm_connector_state *cstate = hdmi->connector->state;
+	struct drm_connector *connector = &vc4_hdmi->connector;
+	struct drm_connector_state *cstate = connector->state;
 	struct drm_crtc *crtc = encoder->crtc;
 	const struct drm_display_mode *mode = &crtc->state->adjusted_mode;
 	union hdmi_infoframe frame;
 	int ret;
 
 	ret = drm_hdmi_avi_infoframe_from_display_mode(&frame.avi,
-						       hdmi->connector, mode);
+						       connector, mode);
 	if (ret < 0) {
 		DRM_ERROR("couldn't fill AVI infoframe\n");
 		return;
 	}
 
 	drm_hdmi_avi_infoframe_quant_range(&frame.avi,
-					   hdmi->connector, mode,
+					   connector, mode,
 					   vc4_encoder->limited_rgb_range ?
 					   HDMI_QUANTIZATION_RANGE_LIMITED :
 					   HDMI_QUANTIZATION_RANGE_FULL);
@@ -416,9 +329,7 @@ static void vc4_hdmi_set_spd_infoframe(struct drm_encoder *encoder)
 
 static void vc4_hdmi_set_audio_infoframe(struct drm_encoder *encoder)
 {
-	struct drm_device *drm = encoder->dev;
-	struct vc4_dev *vc4 = drm->dev_private;
-	struct vc4_hdmi *hdmi = vc4->hdmi;
+	struct vc4_hdmi *vc4_hdmi = encoder_to_vc4_hdmi(encoder);
 	union hdmi_infoframe frame;
 	int ret;
 
@@ -427,45 +338,139 @@ static void vc4_hdmi_set_audio_infoframe(struct drm_encoder *encoder)
 	frame.audio.coding_type = HDMI_AUDIO_CODING_TYPE_STREAM;
 	frame.audio.sample_frequency = HDMI_AUDIO_SAMPLE_FREQUENCY_STREAM;
 	frame.audio.sample_size = HDMI_AUDIO_SAMPLE_SIZE_STREAM;
-	frame.audio.channels = hdmi->audio.channels;
+	frame.audio.channels = vc4_hdmi->audio.channels;
 
 	vc4_hdmi_write_infoframe(encoder, &frame);
 }
 
 static void vc4_hdmi_set_infoframes(struct drm_encoder *encoder)
 {
+	struct vc4_hdmi *vc4_hdmi = encoder_to_vc4_hdmi(encoder);
+
 	vc4_hdmi_set_avi_infoframe(encoder);
 	vc4_hdmi_set_spd_infoframe(encoder);
+	/*
+	 * If audio was streaming, then we need to reenabled the audio
+	 * infoframe here during encoder_enable.
+	 */
+	if (vc4_hdmi->audio.streaming)
+		vc4_hdmi_set_audio_infoframe(encoder);
 }
 
-static void vc4_hdmi_encoder_disable(struct drm_encoder *encoder)
+static void vc4_hdmi_encoder_post_crtc_disable(struct drm_encoder *encoder)
 {
-	struct drm_device *dev = encoder->dev;
-	struct vc4_dev *vc4 = to_vc4_dev(dev);
-	struct vc4_hdmi *hdmi = vc4->hdmi;
+	struct vc4_hdmi *vc4_hdmi = encoder_to_vc4_hdmi(encoder);
+
+	HDMI_WRITE(HDMI_RAM_PACKET_CONFIG, 0);
+
+	HDMI_WRITE(HDMI_VID_CTL, HDMI_READ(HDMI_VID_CTL) |
+		   VC4_HD_VID_CTL_CLRRGB | VC4_HD_VID_CTL_CLRSYNC);
+
+	HDMI_WRITE(HDMI_VID_CTL,
+		   HDMI_READ(HDMI_VID_CTL) | VC4_HD_VID_CTL_BLANKPIX);
+}
+
+static void vc4_hdmi_encoder_post_crtc_powerdown(struct drm_encoder *encoder)
+{
+	struct vc4_hdmi *vc4_hdmi = encoder_to_vc4_hdmi(encoder);
 	int ret;
 
-	HDMI_WRITE(VC4_HDMI_RAM_PACKET_CONFIG, 0);
+	if (vc4_hdmi->variant->phy_disable)
+		vc4_hdmi->variant->phy_disable(vc4_hdmi);
 
-	HDMI_WRITE(VC4_HDMI_TX_PHY_RESET_CTL, 0xf << 16);
-	HD_WRITE(VC4_HD_VID_CTL,
-		 HD_READ(VC4_HD_VID_CTL) & ~VC4_HD_VID_CTL_ENABLE);
+	HDMI_WRITE(HDMI_VID_CTL,
+		   HDMI_READ(HDMI_VID_CTL) & ~VC4_HD_VID_CTL_ENABLE);
 
-	clk_disable_unprepare(hdmi->pixel_clock);
+	clk_disable_unprepare(vc4_hdmi->pixel_bvb_clock);
+	clk_disable_unprepare(vc4_hdmi->hsm_clock);
+	clk_disable_unprepare(vc4_hdmi->pixel_clock);
 
-	ret = pm_runtime_put(&hdmi->pdev->dev);
+	ret = pm_runtime_put(&vc4_hdmi->pdev->dev);
 	if (ret < 0)
 		DRM_ERROR("Failed to release power domain: %d\n", ret);
 }
 
-static void vc4_hdmi_encoder_enable(struct drm_encoder *encoder)
+static void vc4_hdmi_encoder_disable(struct drm_encoder *encoder)
+{
+}
+
+static void vc4_hdmi_csc_setup(struct vc4_hdmi *vc4_hdmi, bool enable)
+{
+	u32 csc_ctl;
+
+	csc_ctl = VC4_SET_FIELD(VC4_HD_CSC_CTL_ORDER_BGR,
+				VC4_HD_CSC_CTL_ORDER);
+
+	if (enable) {
+		/* CEA VICs other than #1 requre limited range RGB
+		 * output unless overridden by an AVI infoframe.
+		 * Apply a colorspace conversion to squash 0-255 down
+		 * to 16-235.  The matrix here is:
+		 *
+		 * [ 0      0      0.8594 16]
+		 * [ 0      0.8594 0      16]
+		 * [ 0.8594 0      0      16]
+		 * [ 0      0      0       1]
+		 */
+		csc_ctl |= VC4_HD_CSC_CTL_ENABLE;
+		csc_ctl |= VC4_HD_CSC_CTL_RGB2YCC;
+		csc_ctl |= VC4_SET_FIELD(VC4_HD_CSC_CTL_MODE_CUSTOM,
+					 VC4_HD_CSC_CTL_MODE);
+
+		HDMI_WRITE(HDMI_CSC_12_11, (0x000 << 16) | 0x000);
+		HDMI_WRITE(HDMI_CSC_14_13, (0x100 << 16) | 0x6e0);
+		HDMI_WRITE(HDMI_CSC_22_21, (0x6e0 << 16) | 0x000);
+		HDMI_WRITE(HDMI_CSC_24_23, (0x100 << 16) | 0x000);
+		HDMI_WRITE(HDMI_CSC_32_31, (0x000 << 16) | 0x6e0);
+		HDMI_WRITE(HDMI_CSC_34_33, (0x100 << 16) | 0x000);
+	}
+
+	/* The RGB order applies even when CSC is disabled. */
+	HDMI_WRITE(HDMI_CSC_CTL, csc_ctl);
+}
+
+static void vc5_hdmi_csc_setup(struct vc4_hdmi *vc4_hdmi, bool enable)
+{
+	u32 csc_ctl;
+
+	csc_ctl = 0x07;	/* RGB_CONVERT_MODE = custom matrix, || USE_RGB_TO_YCBCR */
+
+	if (enable) {
+		/* CEA VICs other than #1 requre limited range RGB
+		 * output unless overridden by an AVI infoframe.
+		 * Apply a colorspace conversion to squash 0-255 down
+		 * to 16-235.  The matrix here is:
+		 *
+		 * [ 0.8594 0      0      16]
+		 * [ 0      0.8594 0      16]
+		 * [ 0      0      0.8594 16]
+		 * [ 0      0      0       1]
+		 * Matrix is signed 2p13 fixed point, with signed 9p6 offsets
+		 */
+		HDMI_WRITE(HDMI_CSC_12_11, (0x0000 << 16) | 0x1b80);
+		HDMI_WRITE(HDMI_CSC_14_13, (0x0400 << 16) | 0x0000);
+		HDMI_WRITE(HDMI_CSC_22_21, (0x1b80 << 16) | 0x0000);
+		HDMI_WRITE(HDMI_CSC_24_23, (0x0400 << 16) | 0x0000);
+		HDMI_WRITE(HDMI_CSC_32_31, (0x0000 << 16) | 0x0000);
+		HDMI_WRITE(HDMI_CSC_34_33, (0x0400 << 16) | 0x1b80);
+	} else {
+		/* Still use the matrix for full range, but make it unity.
+		 * Matrix is signed 2p13 fixed point, with signed 9p6 offsets
+		 */
+		HDMI_WRITE(HDMI_CSC_12_11, (0x0000 << 16) | 0x2000);
+		HDMI_WRITE(HDMI_CSC_14_13, (0x0000 << 16) | 0x0000);
+		HDMI_WRITE(HDMI_CSC_22_21, (0x2000 << 16) | 0x0000);
+		HDMI_WRITE(HDMI_CSC_24_23, (0x0000 << 16) | 0x0000);
+		HDMI_WRITE(HDMI_CSC_32_31, (0x0000 << 16) | 0x0000);
+		HDMI_WRITE(HDMI_CSC_34_33, (0x0000 << 16) | 0x2000);
+	}
+
+	HDMI_WRITE(HDMI_CSC_CTL, csc_ctl);
+}
+
+static void vc4_hdmi_set_timings(struct vc4_hdmi *vc4_hdmi,
+				 struct drm_display_mode *mode)
 {
-	struct drm_display_mode *mode = &encoder->crtc->state->adjusted_mode;
-	struct vc4_hdmi_encoder *vc4_encoder = to_vc4_hdmi_encoder(encoder);
-	struct drm_device *dev = encoder->dev;
-	struct vc4_dev *vc4 = to_vc4_dev(dev);
-	struct vc4_hdmi *hdmi = vc4->hdmi;
-	bool debug_dump_regs = false;
 	bool hsync_pos = mode->flags & DRM_MODE_FLAG_PHSYNC;
 	bool vsync_pos = mode->flags & DRM_MODE_FLAG_PVSYNC;
 	bool interlaced = mode->flags & DRM_MODE_FLAG_INTERLACE;
@@ -483,213 +488,285 @@ static void vc4_hdmi_encoder_enable(struct drm_encoder *encoder)
 					mode->crtc_vsync_end -
 					interlaced,
 					VC4_HDMI_VERTB_VBP));
-	u32 csc_ctl;
+
+	HDMI_WRITE(HDMI_HORZA,
+		   (vsync_pos ? VC4_HDMI_HORZA_VPOS : 0) |
+		   (hsync_pos ? VC4_HDMI_HORZA_HPOS : 0) |
+		   VC4_SET_FIELD(mode->hdisplay * pixel_rep,
+				 VC4_HDMI_HORZA_HAP));
+
+	HDMI_WRITE(HDMI_HORZB,
+		   VC4_SET_FIELD((mode->htotal -
+				  mode->hsync_end) * pixel_rep,
+				 VC4_HDMI_HORZB_HBP) |
+		   VC4_SET_FIELD((mode->hsync_end -
+				  mode->hsync_start) * pixel_rep,
+				 VC4_HDMI_HORZB_HSP) |
+		   VC4_SET_FIELD((mode->hsync_start -
+				  mode->hdisplay) * pixel_rep,
+				 VC4_HDMI_HORZB_HFP));
+
+	HDMI_WRITE(HDMI_VERTA0, verta);
+	HDMI_WRITE(HDMI_VERTA1, verta);
+
+	HDMI_WRITE(HDMI_VERTB0, vertb_even);
+	HDMI_WRITE(HDMI_VERTB1, vertb);
+}
+static void vc5_hdmi_set_timings(struct vc4_hdmi *vc4_hdmi,
+				 struct drm_display_mode *mode)
+{
+	bool hsync_pos = mode->flags & DRM_MODE_FLAG_PHSYNC;
+	bool vsync_pos = mode->flags & DRM_MODE_FLAG_PVSYNC;
+	bool interlaced = mode->flags & DRM_MODE_FLAG_INTERLACE;
+	u32 pixel_rep = (mode->flags & DRM_MODE_FLAG_DBLCLK) ? 2 : 1;
+	u32 verta = (VC4_SET_FIELD(mode->crtc_vsync_end - mode->crtc_vsync_start,
+				   VC5_HDMI_VERTA_VSP) |
+		     VC4_SET_FIELD(mode->crtc_vsync_start - mode->crtc_vdisplay,
+				   VC5_HDMI_VERTA_VFP) |
+		     VC4_SET_FIELD(mode->crtc_vdisplay, VC5_HDMI_VERTA_VAL));
+	u32 vertb = (VC4_SET_FIELD(0, VC5_HDMI_VERTB_VSPO) |
+		     VC4_SET_FIELD(mode->crtc_vtotal - mode->crtc_vsync_end,
+				   VC4_HDMI_VERTB_VBP));
+	u32 vertb_even = (VC4_SET_FIELD(0, VC5_HDMI_VERTB_VSPO) |
+			  VC4_SET_FIELD(mode->crtc_vtotal -
+					mode->crtc_vsync_end -
+					interlaced,
+					VC4_HDMI_VERTB_VBP));
+
+	HDMI_WRITE(HDMI_VEC_INTERFACE_XBAR, 0x354021);
+	HDMI_WRITE(HDMI_HORZA,
+		   (vsync_pos ? VC5_HDMI_HORZA_VPOS : 0) |
+		   (hsync_pos ? VC5_HDMI_HORZA_HPOS : 0) |
+		   VC4_SET_FIELD(mode->hdisplay * pixel_rep,
+				 VC5_HDMI_HORZA_HAP) |
+		   VC4_SET_FIELD((mode->hsync_start -
+				  mode->hdisplay) * pixel_rep,
+				 VC5_HDMI_HORZA_HFP));
+
+	HDMI_WRITE(HDMI_HORZB,
+		   VC4_SET_FIELD((mode->htotal -
+				  mode->hsync_end) * pixel_rep,
+				 VC5_HDMI_HORZB_HBP) |
+		   VC4_SET_FIELD((mode->hsync_end -
+				  mode->hsync_start) * pixel_rep,
+				 VC5_HDMI_HORZB_HSP));
+
+	HDMI_WRITE(HDMI_VERTA0, verta);
+	HDMI_WRITE(HDMI_VERTA1, verta);
+
+	HDMI_WRITE(HDMI_VERTB0, vertb_even);
+	HDMI_WRITE(HDMI_VERTB1, vertb);
+
+	HDMI_WRITE(HDMI_CLOCK_STOP, 0);
+}
+
+static void vc4_hdmi_recenter_fifo(struct vc4_hdmi *vc4_hdmi)
+{
+	u32 drift;
+	int ret;
+
+	drift = HDMI_READ(HDMI_FIFO_CTL);
+	drift &= VC4_HDMI_FIFO_VALID_WRITE_MASK;
+
+	HDMI_WRITE(HDMI_FIFO_CTL,
+		   drift & ~VC4_HDMI_FIFO_CTL_RECENTER);
+	HDMI_WRITE(HDMI_FIFO_CTL,
+		   drift | VC4_HDMI_FIFO_CTL_RECENTER);
+	usleep_range(1000, 1100);
+	HDMI_WRITE(HDMI_FIFO_CTL,
+		   drift & ~VC4_HDMI_FIFO_CTL_RECENTER);
+	HDMI_WRITE(HDMI_FIFO_CTL,
+		   drift | VC4_HDMI_FIFO_CTL_RECENTER);
+
+	ret = wait_for(HDMI_READ(HDMI_FIFO_CTL) &
+		       VC4_HDMI_FIFO_CTL_RECENTER_DONE, 1);
+	WARN_ONCE(ret, "Timeout waiting for "
+		  "VC4_HDMI_FIFO_CTL_RECENTER_DONE");
+}
+
+static void vc4_hdmi_encoder_pre_crtc_configure(struct drm_encoder *encoder)
+{
+	struct drm_display_mode *mode = &encoder->crtc->state->adjusted_mode;
+	struct vc4_hdmi *vc4_hdmi = encoder_to_vc4_hdmi(encoder);
+	unsigned long pixel_rate, hsm_rate;
 	int ret;
 
-	ret = pm_runtime_get_sync(&hdmi->pdev->dev);
+	ret = pm_runtime_get_sync(&vc4_hdmi->pdev->dev);
 	if (ret < 0) {
 		DRM_ERROR("Failed to retain power domain: %d\n", ret);
 		return;
 	}
 
-	ret = clk_set_rate(hdmi->pixel_clock,
-			   mode->clock * 1000 *
-			   ((mode->flags & DRM_MODE_FLAG_DBLCLK) ? 2 : 1));
+	pixel_rate = mode->clock * 1000 * ((mode->flags & DRM_MODE_FLAG_DBLCLK) ? 2 : 1);
+	ret = clk_set_rate(vc4_hdmi->pixel_clock, pixel_rate);
 	if (ret) {
 		DRM_ERROR("Failed to set pixel clock rate: %d\n", ret);
 		return;
 	}
 
-	ret = clk_prepare_enable(hdmi->pixel_clock);
+	ret = clk_prepare_enable(vc4_hdmi->pixel_clock);
 	if (ret) {
 		DRM_ERROR("Failed to turn on pixel clock: %d\n", ret);
 		return;
 	}
 
-	HDMI_WRITE(VC4_HDMI_SW_RESET_CONTROL,
-		   VC4_HDMI_SW_RESET_HDMI |
-		   VC4_HDMI_SW_RESET_FORMAT_DETECT);
-
-	HDMI_WRITE(VC4_HDMI_SW_RESET_CONTROL, 0);
-
-	/* PHY should be in reset, like
-	 * vc4_hdmi_encoder_disable() does.
+	/*
+	 * As stated in RPi's vc4 firmware "HDMI state machine (HSM) clock must
+	 * be faster than pixel clock, infinitesimally faster, tested in
+	 * simulation. Otherwise, exact value is unimportant for HDMI
+	 * operation." This conflicts with bcm2835's vc4 documentation, which
+	 * states HSM's clock has to be at least 108% of the pixel clock.
+	 *
+	 * Real life tests reveal that vc4's firmware statement holds up, and
+	 * users are able to use pixel clocks closer to HSM's, namely for
+	 * 1920x1200@60Hz. So it was decided to have leave a 1% margin between
+	 * both clocks. Which, for RPi0-3 implies a maximum pixel clock of
+	 * 162MHz.
+	 *
+	 * Additionally, the AXI clock needs to be at least 25% of
+	 * pixel clock, but HSM ends up being the limiting factor.
 	 */
-	HDMI_WRITE(VC4_HDMI_TX_PHY_RESET_CTL, 0xf << 16);
+	hsm_rate = max_t(unsigned long, 120000000, (pixel_rate / 100) * 101);
+	ret = clk_set_min_rate(vc4_hdmi->hsm_clock, hsm_rate);
+	if (ret) {
+		DRM_ERROR("Failed to set HSM clock rate: %d\n", ret);
+		return;
+	}
 
-	HDMI_WRITE(VC4_HDMI_TX_PHY_RESET_CTL, 0);
+	ret = clk_prepare_enable(vc4_hdmi->hsm_clock);
+	if (ret) {
+		DRM_ERROR("Failed to turn on HSM clock: %d\n", ret);
+		clk_disable_unprepare(vc4_hdmi->pixel_clock);
+		return;
+	}
 
-	if (debug_dump_regs) {
-		struct drm_printer p = drm_info_printer(&hdmi->pdev->dev);
+	/*
+	 * FIXME: When the pixel freq is 594MHz (4k60), this needs to be setup
+	 * at 300MHz.
+	 */
+	ret = clk_set_min_rate(vc4_hdmi->pixel_bvb_clock,
+			       (hsm_rate > VC4_HSM_MID_CLOCK ? 150000000 : 75000000));
+	if (ret) {
+		DRM_ERROR("Failed to set pixel bvb clock rate: %d\n", ret);
+		clk_disable_unprepare(vc4_hdmi->hsm_clock);
+		clk_disable_unprepare(vc4_hdmi->pixel_clock);
+		return;
+	}
 
-		dev_info(&hdmi->pdev->dev, "HDMI regs before:\n");
-		drm_print_regset32(&p, &hdmi->hdmi_regset);
-		drm_print_regset32(&p, &hdmi->hd_regset);
+	ret = clk_prepare_enable(vc4_hdmi->pixel_bvb_clock);
+	if (ret) {
+		DRM_ERROR("Failed to turn on pixel bvb clock: %d\n", ret);
+		clk_disable_unprepare(vc4_hdmi->hsm_clock);
+		clk_disable_unprepare(vc4_hdmi->pixel_clock);
+		return;
 	}
 
-	HD_WRITE(VC4_HD_VID_CTL, 0);
+	if (vc4_hdmi->variant->reset)
+		vc4_hdmi->variant->reset(vc4_hdmi);
 
-	HDMI_WRITE(VC4_HDMI_SCHEDULER_CONTROL,
-		   HDMI_READ(VC4_HDMI_SCHEDULER_CONTROL) |
+	if (vc4_hdmi->variant->phy_init)
+		vc4_hdmi->variant->phy_init(vc4_hdmi, mode);
+
+	HDMI_WRITE(HDMI_SCHEDULER_CONTROL,
+		   HDMI_READ(HDMI_SCHEDULER_CONTROL) |
 		   VC4_HDMI_SCHEDULER_CONTROL_MANUAL_FORMAT |
 		   VC4_HDMI_SCHEDULER_CONTROL_IGNORE_VSYNC_PREDICTS);
 
-	HDMI_WRITE(VC4_HDMI_HORZA,
-		   (vsync_pos ? VC4_HDMI_HORZA_VPOS : 0) |
-		   (hsync_pos ? VC4_HDMI_HORZA_HPOS : 0) |
-		   VC4_SET_FIELD(mode->hdisplay * pixel_rep,
-				 VC4_HDMI_HORZA_HAP));
-
-	HDMI_WRITE(VC4_HDMI_HORZB,
-		   VC4_SET_FIELD((mode->htotal -
-				  mode->hsync_end) * pixel_rep,
-				 VC4_HDMI_HORZB_HBP) |
-		   VC4_SET_FIELD((mode->hsync_end -
-				  mode->hsync_start) * pixel_rep,
-				 VC4_HDMI_HORZB_HSP) |
-		   VC4_SET_FIELD((mode->hsync_start -
-				  mode->hdisplay) * pixel_rep,
-				 VC4_HDMI_HORZB_HFP));
-
-	HDMI_WRITE(VC4_HDMI_VERTA0, verta);
-	HDMI_WRITE(VC4_HDMI_VERTA1, verta);
-
-	HDMI_WRITE(VC4_HDMI_VERTB0, vertb_even);
-	HDMI_WRITE(VC4_HDMI_VERTB1, vertb);
-
-	HD_WRITE(VC4_HD_VID_CTL,
-		 (vsync_pos ? 0 : VC4_HD_VID_CTL_VSYNC_LOW) |
-		 (hsync_pos ? 0 : VC4_HD_VID_CTL_HSYNC_LOW));
+	if (vc4_hdmi->variant->set_timings)
+		vc4_hdmi->variant->set_timings(vc4_hdmi, mode);
+}
 
-	csc_ctl = VC4_SET_FIELD(VC4_HD_CSC_CTL_ORDER_BGR,
-				VC4_HD_CSC_CTL_ORDER);
+static void vc4_hdmi_encoder_pre_crtc_enable(struct drm_encoder *encoder)
+{
+	struct drm_display_mode *mode = &encoder->crtc->state->adjusted_mode;
+	struct vc4_hdmi_encoder *vc4_encoder = to_vc4_hdmi_encoder(encoder);
+	struct vc4_hdmi *vc4_hdmi = encoder_to_vc4_hdmi(encoder);
 
 	if (vc4_encoder->hdmi_monitor &&
-	    drm_default_rgb_quant_range(mode) ==
-	    HDMI_QUANTIZATION_RANGE_LIMITED) {
-		/* CEA VICs other than #1 requre limited range RGB
-		 * output unless overridden by an AVI infoframe.
-		 * Apply a colorspace conversion to squash 0-255 down
-		 * to 16-235.  The matrix here is:
-		 *
-		 * [ 0      0      0.8594 16]
-		 * [ 0      0.8594 0      16]
-		 * [ 0.8594 0      0      16]
-		 * [ 0      0      0       1]
-		 */
-		csc_ctl |= VC4_HD_CSC_CTL_ENABLE;
-		csc_ctl |= VC4_HD_CSC_CTL_RGB2YCC;
-		csc_ctl |= VC4_SET_FIELD(VC4_HD_CSC_CTL_MODE_CUSTOM,
-					 VC4_HD_CSC_CTL_MODE);
+	    drm_default_rgb_quant_range(mode) == HDMI_QUANTIZATION_RANGE_LIMITED) {
+		if (vc4_hdmi->variant->csc_setup)
+			vc4_hdmi->variant->csc_setup(vc4_hdmi, true);
 
-		HD_WRITE(VC4_HD_CSC_12_11, (0x000 << 16) | 0x000);
-		HD_WRITE(VC4_HD_CSC_14_13, (0x100 << 16) | 0x6e0);
-		HD_WRITE(VC4_HD_CSC_22_21, (0x6e0 << 16) | 0x000);
-		HD_WRITE(VC4_HD_CSC_24_23, (0x100 << 16) | 0x000);
-		HD_WRITE(VC4_HD_CSC_32_31, (0x000 << 16) | 0x6e0);
-		HD_WRITE(VC4_HD_CSC_34_33, (0x100 << 16) | 0x000);
 		vc4_encoder->limited_rgb_range = true;
 	} else {
+		if (vc4_hdmi->variant->csc_setup)
+			vc4_hdmi->variant->csc_setup(vc4_hdmi, false);
+
 		vc4_encoder->limited_rgb_range = false;
 	}
 
-	/* The RGB order applies even when CSC is disabled. */
-	HD_WRITE(VC4_HD_CSC_CTL, csc_ctl);
-
-	HDMI_WRITE(VC4_HDMI_FIFO_CTL, VC4_HDMI_FIFO_CTL_MASTER_SLAVE_N);
+	HDMI_WRITE(HDMI_FIFO_CTL, VC4_HDMI_FIFO_CTL_MASTER_SLAVE_N);
+}
 
-	if (debug_dump_regs) {
-		struct drm_printer p = drm_info_printer(&hdmi->pdev->dev);
+static void vc4_hdmi_encoder_post_crtc_enable(struct drm_encoder *encoder)
+{
+	struct drm_display_mode *mode = &encoder->crtc->state->adjusted_mode;
+	struct vc4_hdmi *vc4_hdmi = encoder_to_vc4_hdmi(encoder);
+	struct vc4_hdmi_encoder *vc4_encoder = to_vc4_hdmi_encoder(encoder);
+	bool hsync_pos = mode->flags & DRM_MODE_FLAG_PHSYNC;
+	bool vsync_pos = mode->flags & DRM_MODE_FLAG_PVSYNC;
+	int ret;
 
-		dev_info(&hdmi->pdev->dev, "HDMI regs after:\n");
-		drm_print_regset32(&p, &hdmi->hdmi_regset);
-		drm_print_regset32(&p, &hdmi->hd_regset);
-	}
+	HDMI_WRITE(HDMI_VID_CTL,
+		   VC4_HD_VID_CTL_ENABLE |
+		   VC4_HD_VID_CTL_UNDERFLOW_ENABLE |
+		   VC4_HD_VID_CTL_FRAME_COUNTER_RESET |
+		   (vsync_pos ? 0 : VC4_HD_VID_CTL_VSYNC_LOW) |
+		   (hsync_pos ? 0 : VC4_HD_VID_CTL_HSYNC_LOW));
 
-	HD_WRITE(VC4_HD_VID_CTL,
-		 HD_READ(VC4_HD_VID_CTL) |
-		 VC4_HD_VID_CTL_ENABLE |
-		 VC4_HD_VID_CTL_UNDERFLOW_ENABLE |
-		 VC4_HD_VID_CTL_FRAME_COUNTER_RESET);
+	HDMI_WRITE(HDMI_VID_CTL,
+		   HDMI_READ(HDMI_VID_CTL) & ~VC4_HD_VID_CTL_BLANKPIX);
 
 	if (vc4_encoder->hdmi_monitor) {
-		HDMI_WRITE(VC4_HDMI_SCHEDULER_CONTROL,
-			   HDMI_READ(VC4_HDMI_SCHEDULER_CONTROL) |
+		HDMI_WRITE(HDMI_SCHEDULER_CONTROL,
+			   HDMI_READ(HDMI_SCHEDULER_CONTROL) |
 			   VC4_HDMI_SCHEDULER_CONTROL_MODE_HDMI);
 
-		ret = wait_for(HDMI_READ(VC4_HDMI_SCHEDULER_CONTROL) &
+		ret = wait_for(HDMI_READ(HDMI_SCHEDULER_CONTROL) &
 			       VC4_HDMI_SCHEDULER_CONTROL_HDMI_ACTIVE, 1000);
 		WARN_ONCE(ret, "Timeout waiting for "
 			  "VC4_HDMI_SCHEDULER_CONTROL_HDMI_ACTIVE\n");
 	} else {
-		HDMI_WRITE(VC4_HDMI_RAM_PACKET_CONFIG,
-			   HDMI_READ(VC4_HDMI_RAM_PACKET_CONFIG) &
+		HDMI_WRITE(HDMI_RAM_PACKET_CONFIG,
+			   HDMI_READ(HDMI_RAM_PACKET_CONFIG) &
 			   ~(VC4_HDMI_RAM_PACKET_ENABLE));
-		HDMI_WRITE(VC4_HDMI_SCHEDULER_CONTROL,
-			   HDMI_READ(VC4_HDMI_SCHEDULER_CONTROL) &
+		HDMI_WRITE(HDMI_SCHEDULER_CONTROL,
+			   HDMI_READ(HDMI_SCHEDULER_CONTROL) &
 			   ~VC4_HDMI_SCHEDULER_CONTROL_MODE_HDMI);
 
-		ret = wait_for(!(HDMI_READ(VC4_HDMI_SCHEDULER_CONTROL) &
+		ret = wait_for(!(HDMI_READ(HDMI_SCHEDULER_CONTROL) &
 				 VC4_HDMI_SCHEDULER_CONTROL_HDMI_ACTIVE), 1000);
 		WARN_ONCE(ret, "Timeout waiting for "
 			  "!VC4_HDMI_SCHEDULER_CONTROL_HDMI_ACTIVE\n");
 	}
 
 	if (vc4_encoder->hdmi_monitor) {
-		u32 drift;
-
-		WARN_ON(!(HDMI_READ(VC4_HDMI_SCHEDULER_CONTROL) &
+		WARN_ON(!(HDMI_READ(HDMI_SCHEDULER_CONTROL) &
 			  VC4_HDMI_SCHEDULER_CONTROL_HDMI_ACTIVE));
-		HDMI_WRITE(VC4_HDMI_SCHEDULER_CONTROL,
-			   HDMI_READ(VC4_HDMI_SCHEDULER_CONTROL) |
+		HDMI_WRITE(HDMI_SCHEDULER_CONTROL,
+			   HDMI_READ(HDMI_SCHEDULER_CONTROL) |
 			   VC4_HDMI_SCHEDULER_CONTROL_VERT_ALWAYS_KEEPOUT);
 
-		HDMI_WRITE(VC4_HDMI_RAM_PACKET_CONFIG,
+		HDMI_WRITE(HDMI_RAM_PACKET_CONFIG,
 			   VC4_HDMI_RAM_PACKET_ENABLE);
 
 		vc4_hdmi_set_infoframes(encoder);
-
-		drift = HDMI_READ(VC4_HDMI_FIFO_CTL);
-		drift &= VC4_HDMI_FIFO_VALID_WRITE_MASK;
-
-		HDMI_WRITE(VC4_HDMI_FIFO_CTL,
-			   drift & ~VC4_HDMI_FIFO_CTL_RECENTER);
-		HDMI_WRITE(VC4_HDMI_FIFO_CTL,
-			   drift | VC4_HDMI_FIFO_CTL_RECENTER);
-		usleep_range(1000, 1100);
-		HDMI_WRITE(VC4_HDMI_FIFO_CTL,
-			   drift & ~VC4_HDMI_FIFO_CTL_RECENTER);
-		HDMI_WRITE(VC4_HDMI_FIFO_CTL,
-			   drift | VC4_HDMI_FIFO_CTL_RECENTER);
-
-		ret = wait_for(HDMI_READ(VC4_HDMI_FIFO_CTL) &
-			       VC4_HDMI_FIFO_CTL_RECENTER_DONE, 1);
-		WARN_ONCE(ret, "Timeout waiting for "
-			  "VC4_HDMI_FIFO_CTL_RECENTER_DONE");
 	}
+
+	vc4_hdmi_recenter_fifo(vc4_hdmi);
+}
+
+static void vc4_hdmi_encoder_enable(struct drm_encoder *encoder)
+{
 }
 
 static enum drm_mode_status
-vc4_hdmi_encoder_mode_valid(struct drm_encoder *crtc,
+vc4_hdmi_encoder_mode_valid(struct drm_encoder *encoder,
 			    const struct drm_display_mode *mode)
 {
-	/*
-	 * As stated in RPi's vc4 firmware "HDMI state machine (HSM) clock must
-	 * be faster than pixel clock, infinitesimally faster, tested in
-	 * simulation. Otherwise, exact value is unimportant for HDMI
-	 * operation." This conflicts with bcm2835's vc4 documentation, which
-	 * states HSM's clock has to be at least 108% of the pixel clock.
-	 *
-	 * Real life tests reveal that vc4's firmware statement holds up, and
-	 * users are able to use pixel clocks closer to HSM's, namely for
-	 * 1920x1200@60Hz. So it was decided to have leave a 1% margin between
-	 * both clocks. Which, for RPi0-3 implies a maximum pixel clock of
-	 * 162MHz.
-	 *
-	 * Additionally, the AXI clock needs to be at least 25% of
-	 * pixel clock, but HSM ends up being the limiting factor.
-	 */
-	if (mode->clock > HSM_CLOCK_FREQ / (1000 * 101 / 100))
+	struct vc4_hdmi *vc4_hdmi = encoder_to_vc4_hdmi(encoder);
+
+	if ((mode->clock * 1000) > vc4_hdmi->variant->max_pixel_clock)
 		return MODE_CLOCK_HIGH;
 
 	return MODE_OK;
@@ -701,34 +778,54 @@ static const struct drm_encoder_helper_funcs vc4_hdmi_encoder_helper_funcs = {
 	.enable = vc4_hdmi_encoder_enable,
 };
 
+static u32 vc4_hdmi_channel_map(struct vc4_hdmi *vc4_hdmi, u32 channel_mask)
+{
+	int i;
+	u32 channel_map = 0;
+
+	for (i = 0; i < 8; i++) {
+		if (channel_mask & BIT(i))
+			channel_map |= i << (3 * i);
+	}
+	return channel_map;
+}
+
+static u32 vc5_hdmi_channel_map(struct vc4_hdmi *vc4_hdmi, u32 channel_mask)
+{
+	int i;
+	u32 channel_map = 0;
+
+	for (i = 0; i < 8; i++) {
+		if (channel_mask & BIT(i))
+			channel_map |= i << (4 * i);
+	}
+	return channel_map;
+}
+
 /* HDMI audio codec callbacks */
-static void vc4_hdmi_audio_set_mai_clock(struct vc4_hdmi *hdmi)
+static void vc4_hdmi_audio_set_mai_clock(struct vc4_hdmi *vc4_hdmi)
 {
-	struct drm_device *drm = hdmi->encoder->dev;
-	struct vc4_dev *vc4 = to_vc4_dev(drm);
-	u32 hsm_clock = clk_get_rate(hdmi->hsm_clock);
+	u32 hsm_clock = clk_get_rate(vc4_hdmi->audio_clock);
 	unsigned long n, m;
 
-	rational_best_approximation(hsm_clock, hdmi->audio.samplerate,
+	rational_best_approximation(hsm_clock, vc4_hdmi->audio.samplerate,
 				    VC4_HD_MAI_SMP_N_MASK >>
 				    VC4_HD_MAI_SMP_N_SHIFT,
 				    (VC4_HD_MAI_SMP_M_MASK >>
 				     VC4_HD_MAI_SMP_M_SHIFT) + 1,
 				    &n, &m);
 
-	HD_WRITE(VC4_HD_MAI_SMP,
-		 VC4_SET_FIELD(n, VC4_HD_MAI_SMP_N) |
-		 VC4_SET_FIELD(m - 1, VC4_HD_MAI_SMP_M));
+	HDMI_WRITE(HDMI_MAI_SMP,
+		   VC4_SET_FIELD(n, VC4_HD_MAI_SMP_N) |
+		   VC4_SET_FIELD(m - 1, VC4_HD_MAI_SMP_M));
 }
 
-static void vc4_hdmi_set_n_cts(struct vc4_hdmi *hdmi)
+static void vc4_hdmi_set_n_cts(struct vc4_hdmi *vc4_hdmi)
 {
-	struct drm_encoder *encoder = hdmi->encoder;
+	struct drm_encoder *encoder = &vc4_hdmi->encoder.base.base;
 	struct drm_crtc *crtc = encoder->crtc;
-	struct drm_device *drm = encoder->dev;
-	struct vc4_dev *vc4 = to_vc4_dev(drm);
 	const struct drm_display_mode *mode = &crtc->state->adjusted_mode;
-	u32 samplerate = hdmi->audio.samplerate;
+	u32 samplerate = vc4_hdmi->audio.samplerate;
 	u32 n, cts;
 	u64 tmp;
 
@@ -737,7 +834,7 @@ static void vc4_hdmi_set_n_cts(struct vc4_hdmi *hdmi)
 	do_div(tmp, 128 * samplerate);
 	cts = tmp;
 
-	HDMI_WRITE(VC4_HDMI_CRP_CFG,
+	HDMI_WRITE(HDMI_CRP_CFG,
 		   VC4_HDMI_CRP_CFG_EXTERNAL_CTS_EN |
 		   VC4_SET_FIELD(n, VC4_HDMI_CRP_CFG_N));
 
@@ -746,8 +843,8 @@ static void vc4_hdmi_set_n_cts(struct vc4_hdmi *hdmi)
 	 * providing a CTS_1 value.  The two CTS values are alternated
 	 * between based on the period fields
 	 */
-	HDMI_WRITE(VC4_HDMI_CTS_0, cts);
-	HDMI_WRITE(VC4_HDMI_CTS_1, cts);
+	HDMI_WRITE(HDMI_CTS_0, cts);
+	HDMI_WRITE(HDMI_CTS_1, cts);
 }
 
 static inline struct vc4_hdmi *dai_to_hdmi(struct snd_soc_dai *dai)
@@ -760,26 +857,25 @@ static inline struct vc4_hdmi *dai_to_hdmi(struct snd_soc_dai *dai)
 static int vc4_hdmi_audio_startup(struct snd_pcm_substream *substream,
 				  struct snd_soc_dai *dai)
 {
-	struct vc4_hdmi *hdmi = dai_to_hdmi(dai);
-	struct drm_encoder *encoder = hdmi->encoder;
-	struct vc4_dev *vc4 = to_vc4_dev(encoder->dev);
+	struct vc4_hdmi *vc4_hdmi = dai_to_hdmi(dai);
+	struct drm_encoder *encoder = &vc4_hdmi->encoder.base.base;
+	struct drm_connector *connector = &vc4_hdmi->connector;
 	int ret;
 
-	if (hdmi->audio.substream && hdmi->audio.substream != substream)
+	if (vc4_hdmi->audio.substream && vc4_hdmi->audio.substream != substream)
 		return -EINVAL;
 
-	hdmi->audio.substream = substream;
+	vc4_hdmi->audio.substream = substream;
 
 	/*
 	 * If the HDMI encoder hasn't probed, or the encoder is
 	 * currently in DVI mode, treat the codec dai as missing.
 	 */
-	if (!encoder->crtc || !(HDMI_READ(VC4_HDMI_RAM_PACKET_CONFIG) &
+	if (!encoder->crtc || !(HDMI_READ(HDMI_RAM_PACKET_CONFIG) &
 				VC4_HDMI_RAM_PACKET_ENABLE))
 		return -ENODEV;
 
-	ret = snd_pcm_hw_constraint_eld(substream->runtime,
-					hdmi->connector->eld);
+	ret = snd_pcm_hw_constraint_eld(substream->runtime, connector->eld);
 	if (ret)
 		return ret;
 
@@ -791,34 +887,33 @@ static int vc4_hdmi_audio_set_fmt(struct snd_soc_dai *dai, unsigned int fmt)
 	return 0;
 }
 
-static void vc4_hdmi_audio_reset(struct vc4_hdmi *hdmi)
+static void vc4_hdmi_audio_reset(struct vc4_hdmi *vc4_hdmi)
 {
-	struct drm_encoder *encoder = hdmi->encoder;
-	struct drm_device *drm = encoder->dev;
-	struct device *dev = &hdmi->pdev->dev;
-	struct vc4_dev *vc4 = to_vc4_dev(drm);
+	struct drm_encoder *encoder = &vc4_hdmi->encoder.base.base;
+	struct device *dev = &vc4_hdmi->pdev->dev;
 	int ret;
 
+	vc4_hdmi->audio.streaming = false;
 	ret = vc4_hdmi_stop_packet(encoder, HDMI_INFOFRAME_TYPE_AUDIO);
 	if (ret)
 		dev_err(dev, "Failed to stop audio infoframe: %d\n", ret);
 
-	HD_WRITE(VC4_HD_MAI_CTL, VC4_HD_MAI_CTL_RESET);
-	HD_WRITE(VC4_HD_MAI_CTL, VC4_HD_MAI_CTL_ERRORF);
-	HD_WRITE(VC4_HD_MAI_CTL, VC4_HD_MAI_CTL_FLUSH);
+	HDMI_WRITE(HDMI_MAI_CTL, VC4_HD_MAI_CTL_RESET);
+	HDMI_WRITE(HDMI_MAI_CTL, VC4_HD_MAI_CTL_ERRORF);
+	HDMI_WRITE(HDMI_MAI_CTL, VC4_HD_MAI_CTL_FLUSH);
 }
 
 static void vc4_hdmi_audio_shutdown(struct snd_pcm_substream *substream,
 				    struct snd_soc_dai *dai)
 {
-	struct vc4_hdmi *hdmi = dai_to_hdmi(dai);
+	struct vc4_hdmi *vc4_hdmi = dai_to_hdmi(dai);
 
-	if (substream != hdmi->audio.substream)
+	if (substream != vc4_hdmi->audio.substream)
 		return;
 
-	vc4_hdmi_audio_reset(hdmi);
+	vc4_hdmi_audio_reset(vc4_hdmi);
 
-	hdmi->audio.substream = NULL;
+	vc4_hdmi->audio.substream = NULL;
 }
 
 /* HDMI audio codec callbacks */
@@ -826,72 +921,65 @@ static int vc4_hdmi_audio_hw_params(struct snd_pcm_substream *substream,
 				    struct snd_pcm_hw_params *params,
 				    struct snd_soc_dai *dai)
 {
-	struct vc4_hdmi *hdmi = dai_to_hdmi(dai);
-	struct drm_encoder *encoder = hdmi->encoder;
-	struct drm_device *drm = encoder->dev;
-	struct device *dev = &hdmi->pdev->dev;
-	struct vc4_dev *vc4 = to_vc4_dev(drm);
+	struct vc4_hdmi *vc4_hdmi = dai_to_hdmi(dai);
+	struct device *dev = &vc4_hdmi->pdev->dev;
 	u32 audio_packet_config, channel_mask;
-	u32 channel_map, i;
+	u32 channel_map;
 
-	if (substream != hdmi->audio.substream)
+	if (substream != vc4_hdmi->audio.substream)
 		return -EINVAL;
 
 	dev_dbg(dev, "%s: %u Hz, %d bit, %d channels\n", __func__,
 		params_rate(params), params_width(params),
 		params_channels(params));
 
-	hdmi->audio.channels = params_channels(params);
-	hdmi->audio.samplerate = params_rate(params);
+	vc4_hdmi->audio.channels = params_channels(params);
+	vc4_hdmi->audio.samplerate = params_rate(params);
 
-	HD_WRITE(VC4_HD_MAI_CTL,
-		 VC4_HD_MAI_CTL_RESET |
-		 VC4_HD_MAI_CTL_FLUSH |
-		 VC4_HD_MAI_CTL_DLATE |
-		 VC4_HD_MAI_CTL_ERRORE |
-		 VC4_HD_MAI_CTL_ERRORF);
+	HDMI_WRITE(HDMI_MAI_CTL,
+		   VC4_HD_MAI_CTL_RESET |
+		   VC4_HD_MAI_CTL_FLUSH |
+		   VC4_HD_MAI_CTL_DLATE |
+		   VC4_HD_MAI_CTL_ERRORE |
+		   VC4_HD_MAI_CTL_ERRORF);
 
-	vc4_hdmi_audio_set_mai_clock(hdmi);
+	vc4_hdmi_audio_set_mai_clock(vc4_hdmi);
 
+	/* The B frame identifier should match the value used by alsa-lib (8) */
 	audio_packet_config =
 		VC4_HDMI_AUDIO_PACKET_ZERO_DATA_ON_SAMPLE_FLAT |
 		VC4_HDMI_AUDIO_PACKET_ZERO_DATA_ON_INACTIVE_CHANNELS |
-		VC4_SET_FIELD(0xf, VC4_HDMI_AUDIO_PACKET_B_FRAME_IDENTIFIER);
+		VC4_SET_FIELD(0x8, VC4_HDMI_AUDIO_PACKET_B_FRAME_IDENTIFIER);
 
-	channel_mask = GENMASK(hdmi->audio.channels - 1, 0);
+	channel_mask = GENMASK(vc4_hdmi->audio.channels - 1, 0);
 	audio_packet_config |= VC4_SET_FIELD(channel_mask,
 					     VC4_HDMI_AUDIO_PACKET_CEA_MASK);
 
 	/* Set the MAI threshold.  This logic mimics the firmware's. */
-	if (hdmi->audio.samplerate > 96000) {
-		HD_WRITE(VC4_HD_MAI_THR,
-			 VC4_SET_FIELD(0x12, VC4_HD_MAI_THR_DREQHIGH) |
-			 VC4_SET_FIELD(0x12, VC4_HD_MAI_THR_DREQLOW));
-	} else if (hdmi->audio.samplerate > 48000) {
-		HD_WRITE(VC4_HD_MAI_THR,
-			 VC4_SET_FIELD(0x14, VC4_HD_MAI_THR_DREQHIGH) |
-			 VC4_SET_FIELD(0x12, VC4_HD_MAI_THR_DREQLOW));
+	if (vc4_hdmi->audio.samplerate > 96000) {
+		HDMI_WRITE(HDMI_MAI_THR,
+			   VC4_SET_FIELD(0x12, VC4_HD_MAI_THR_DREQHIGH) |
+			   VC4_SET_FIELD(0x12, VC4_HD_MAI_THR_DREQLOW));
+	} else if (vc4_hdmi->audio.samplerate > 48000) {
+		HDMI_WRITE(HDMI_MAI_THR,
+			   VC4_SET_FIELD(0x14, VC4_HD_MAI_THR_DREQHIGH) |
+			   VC4_SET_FIELD(0x12, VC4_HD_MAI_THR_DREQLOW));
 	} else {
-		HD_WRITE(VC4_HD_MAI_THR,
-			 VC4_SET_FIELD(0x10, VC4_HD_MAI_THR_PANICHIGH) |
-			 VC4_SET_FIELD(0x10, VC4_HD_MAI_THR_PANICLOW) |
-			 VC4_SET_FIELD(0x10, VC4_HD_MAI_THR_DREQHIGH) |
-			 VC4_SET_FIELD(0x10, VC4_HD_MAI_THR_DREQLOW));
+		HDMI_WRITE(HDMI_MAI_THR,
+			   VC4_SET_FIELD(0x10, VC4_HD_MAI_THR_PANICHIGH) |
+			   VC4_SET_FIELD(0x10, VC4_HD_MAI_THR_PANICLOW) |
+			   VC4_SET_FIELD(0x10, VC4_HD_MAI_THR_DREQHIGH) |
+			   VC4_SET_FIELD(0x10, VC4_HD_MAI_THR_DREQLOW));
 	}
 
-	HDMI_WRITE(VC4_HDMI_MAI_CONFIG,
+	HDMI_WRITE(HDMI_MAI_CONFIG,
 		   VC4_HDMI_MAI_CONFIG_BIT_REVERSE |
 		   VC4_SET_FIELD(channel_mask, VC4_HDMI_MAI_CHANNEL_MASK));
 
-	channel_map = 0;
-	for (i = 0; i < 8; i++) {
-		if (channel_mask & BIT(i))
-			channel_map |= i << (3 * i);
-	}
-
-	HDMI_WRITE(VC4_HDMI_MAI_CHANNEL_MAP, channel_map);
-	HDMI_WRITE(VC4_HDMI_AUDIO_PACKET_CONFIG, audio_packet_config);
-	vc4_hdmi_set_n_cts(hdmi);
+	channel_map = vc4_hdmi->variant->channel_map(vc4_hdmi, channel_mask);
+	HDMI_WRITE(HDMI_MAI_CHANNEL_MAP, channel_map);
+	HDMI_WRITE(HDMI_AUDIO_PACKET_CONFIG, audio_packet_config);
+	vc4_hdmi_set_n_cts(vc4_hdmi);
 
 	return 0;
 }
@@ -899,30 +987,33 @@ static int vc4_hdmi_audio_hw_params(struct snd_pcm_substream *substream,
 static int vc4_hdmi_audio_trigger(struct snd_pcm_substream *substream, int cmd,
 				  struct snd_soc_dai *dai)
 {
-	struct vc4_hdmi *hdmi = dai_to_hdmi(dai);
-	struct drm_encoder *encoder = hdmi->encoder;
-	struct drm_device *drm = encoder->dev;
-	struct vc4_dev *vc4 = to_vc4_dev(drm);
+	struct vc4_hdmi *vc4_hdmi = dai_to_hdmi(dai);
+	struct drm_encoder *encoder = &vc4_hdmi->encoder.base.base;
 
 	switch (cmd) {
 	case SNDRV_PCM_TRIGGER_START:
 		vc4_hdmi_set_audio_infoframe(encoder);
-		HDMI_WRITE(VC4_HDMI_TX_PHY_CTL0,
-			   HDMI_READ(VC4_HDMI_TX_PHY_CTL0) &
-			   ~VC4_HDMI_TX_PHY_RNG_PWRDN);
-		HD_WRITE(VC4_HD_MAI_CTL,
-			 VC4_SET_FIELD(hdmi->audio.channels,
-				       VC4_HD_MAI_CTL_CHNUM) |
-			 VC4_HD_MAI_CTL_ENABLE);
+		vc4_hdmi->audio.streaming = true;
+
+		if (vc4_hdmi->variant->phy_rng_enable)
+			vc4_hdmi->variant->phy_rng_enable(vc4_hdmi);
+
+		HDMI_WRITE(HDMI_MAI_CTL,
+			   VC4_SET_FIELD(vc4_hdmi->audio.channels,
+					 VC4_HD_MAI_CTL_CHNUM) |
+			   VC4_HD_MAI_CTL_ENABLE);
 		break;
 	case SNDRV_PCM_TRIGGER_STOP:
-		HD_WRITE(VC4_HD_MAI_CTL,
-			 VC4_HD_MAI_CTL_DLATE |
-			 VC4_HD_MAI_CTL_ERRORE |
-			 VC4_HD_MAI_CTL_ERRORF);
-		HDMI_WRITE(VC4_HDMI_TX_PHY_CTL0,
-			   HDMI_READ(VC4_HDMI_TX_PHY_CTL0) |
-			   VC4_HDMI_TX_PHY_RNG_PWRDN);
+		HDMI_WRITE(HDMI_MAI_CTL,
+			   VC4_HD_MAI_CTL_DLATE |
+			   VC4_HD_MAI_CTL_ERRORE |
+			   VC4_HD_MAI_CTL_ERRORF);
+
+		if (vc4_hdmi->variant->phy_rng_disable)
+			vc4_hdmi->variant->phy_rng_disable(vc4_hdmi);
+
+		vc4_hdmi->audio.streaming = false;
+
 		break;
 	default:
 		break;
@@ -943,10 +1034,11 @@ static int vc4_hdmi_audio_eld_ctl_info(struct snd_kcontrol *kcontrol,
 				       struct snd_ctl_elem_info *uinfo)
 {
 	struct snd_soc_component *component = snd_kcontrol_chip(kcontrol);
-	struct vc4_hdmi *hdmi = snd_component_to_hdmi(component);
+	struct vc4_hdmi *vc4_hdmi = snd_component_to_hdmi(component);
+	struct drm_connector *connector = &vc4_hdmi->connector;
 
 	uinfo->type = SNDRV_CTL_ELEM_TYPE_BYTES;
-	uinfo->count = sizeof(hdmi->connector->eld);
+	uinfo->count = sizeof(connector->eld);
 
 	return 0;
 }
@@ -955,10 +1047,11 @@ static int vc4_hdmi_audio_eld_ctl_get(struct snd_kcontrol *kcontrol,
 				      struct snd_ctl_elem_value *ucontrol)
 {
 	struct snd_soc_component *component = snd_kcontrol_chip(kcontrol);
-	struct vc4_hdmi *hdmi = snd_component_to_hdmi(component);
+	struct vc4_hdmi *vc4_hdmi = snd_component_to_hdmi(component);
+	struct drm_connector *connector = &vc4_hdmi->connector;
 
-	memcpy(ucontrol->value.bytes.data, hdmi->connector->eld,
-	       sizeof(hdmi->connector->eld));
+	memcpy(ucontrol->value.bytes.data, connector->eld,
+	       sizeof(connector->eld));
 
 	return 0;
 }
@@ -1023,9 +1116,9 @@ static const struct snd_soc_component_driver vc4_hdmi_audio_cpu_dai_comp = {
 
 static int vc4_hdmi_audio_cpu_dai_probe(struct snd_soc_dai *dai)
 {
-	struct vc4_hdmi *hdmi = dai_to_hdmi(dai);
+	struct vc4_hdmi *vc4_hdmi = dai_to_hdmi(dai);
 
-	snd_soc_dai_init_dma_data(dai, &hdmi->audio.dma_data, NULL);
+	snd_soc_dai_init_dma_data(dai, &vc4_hdmi->audio.dma_data, NULL);
 
 	return 0;
 }
@@ -1051,12 +1144,15 @@ static const struct snd_dmaengine_pcm_config pcm_conf = {
 	.prepare_slave_config = snd_dmaengine_pcm_prepare_slave_config,
 };
 
-static int vc4_hdmi_audio_init(struct vc4_hdmi *hdmi)
+static int vc4_hdmi_audio_init(struct vc4_hdmi *vc4_hdmi)
 {
-	struct snd_soc_dai_link *dai_link = &hdmi->audio.link;
-	struct snd_soc_card *card = &hdmi->audio.card;
-	struct device *dev = &hdmi->pdev->dev;
+	const struct vc4_hdmi_register *mai_data =
+		&vc4_hdmi->variant->registers[HDMI_MAI_DATA];
+	struct snd_soc_dai_link *dai_link = &vc4_hdmi->audio.link;
+	struct snd_soc_card *card = &vc4_hdmi->audio.card;
+	struct device *dev = &vc4_hdmi->pdev->dev;
 	const __be32 *addr;
+	int index;
 	int ret;
 
 	if (!of_find_property(dev->of_node, "dmas", NULL)) {
@@ -1065,6 +1161,11 @@ static int vc4_hdmi_audio_init(struct vc4_hdmi *hdmi)
 		return 0;
 	}
 
+	if (mai_data->reg != VC4_HD) {
+		WARN_ONCE(true, "MAI isn't in the HD block\n");
+		return -EINVAL;
+	}
+
 	/*
 	 * Get the physical address of VC4_HD_MAI_DATA. We need to retrieve
 	 * the bus address specified in the DT, because the physical address
@@ -1072,10 +1173,16 @@ static int vc4_hdmi_audio_init(struct vc4_hdmi *hdmi)
 	 * for DMA transfers.
 	 * This VC/MMU should probably be exposed to avoid this kind of hacks.
 	 */
-	addr = of_get_address(dev->of_node, 1, NULL, NULL);
-	hdmi->audio.dma_data.addr = be32_to_cpup(addr) + VC4_HD_MAI_DATA;
-	hdmi->audio.dma_data.addr_width = DMA_SLAVE_BUSWIDTH_4_BYTES;
-	hdmi->audio.dma_data.maxburst = 2;
+	index = of_property_match_string(dev->of_node, "reg-names", "hd");
+	/* Before BCM2711, we don't have a named register range */
+	if (index < 0)
+		index = 1;
+
+	addr = of_get_address(dev->of_node, index, NULL, NULL);
+
+	vc4_hdmi->audio.dma_data.addr = be32_to_cpup(addr) + mai_data->offset;
+	vc4_hdmi->audio.dma_data.addr_width = DMA_SLAVE_BUSWIDTH_4_BYTES;
+	vc4_hdmi->audio.dma_data.maxburst = 2;
 
 	ret = devm_snd_dmaengine_pcm_register(dev, &pcm_conf, 0);
 	if (ret) {
@@ -1098,9 +1205,9 @@ static int vc4_hdmi_audio_init(struct vc4_hdmi *hdmi)
 		return ret;
 	}
 
-	dai_link->cpus		= &hdmi->audio.cpu;
-	dai_link->codecs	= &hdmi->audio.codec;
-	dai_link->platforms	= &hdmi->audio.platform;
+	dai_link->cpus		= &vc4_hdmi->audio.cpu;
+	dai_link->codecs	= &vc4_hdmi->audio.codec;
+	dai_link->platforms	= &vc4_hdmi->audio.platform;
 
 	dai_link->num_cpus	= 1;
 	dai_link->num_codecs	= 1;
@@ -1115,7 +1222,7 @@ static int vc4_hdmi_audio_init(struct vc4_hdmi *hdmi)
 
 	card->dai_link = dai_link;
 	card->num_links = 1;
-	card->name = "vc4-hdmi";
+	card->name = vc4_hdmi->variant->card_name;
 	card->dev = dev;
 
 	/*
@@ -1125,7 +1232,7 @@ static int vc4_hdmi_audio_init(struct vc4_hdmi *hdmi)
 	 * now stored in card->drvdata and should be retrieved with
 	 * snd_soc_card_get_drvdata() if needed.
 	 */
-	snd_soc_card_set_drvdata(card, hdmi);
+	snd_soc_card_set_drvdata(card, vc4_hdmi);
 	ret = devm_snd_soc_register_card(dev, card);
 	if (ret)
 		dev_err(dev, "Could not register sound card: %d\n", ret);
@@ -1137,35 +1244,35 @@ static int vc4_hdmi_audio_init(struct vc4_hdmi *hdmi)
 #ifdef CONFIG_DRM_VC4_HDMI_CEC
 static irqreturn_t vc4_cec_irq_handler_thread(int irq, void *priv)
 {
-	struct vc4_dev *vc4 = priv;
-	struct vc4_hdmi *hdmi = vc4->hdmi;
-
-	if (hdmi->cec_irq_was_rx) {
-		if (hdmi->cec_rx_msg.len)
-			cec_received_msg(hdmi->cec_adap, &hdmi->cec_rx_msg);
-	} else if (hdmi->cec_tx_ok) {
-		cec_transmit_done(hdmi->cec_adap, CEC_TX_STATUS_OK,
+	struct vc4_hdmi *vc4_hdmi = priv;
+
+	if (vc4_hdmi->cec_irq_was_rx) {
+		if (vc4_hdmi->cec_rx_msg.len)
+			cec_received_msg(vc4_hdmi->cec_adap,
+					 &vc4_hdmi->cec_rx_msg);
+	} else if (vc4_hdmi->cec_tx_ok) {
+		cec_transmit_done(vc4_hdmi->cec_adap, CEC_TX_STATUS_OK,
 				  0, 0, 0, 0);
 	} else {
 		/*
 		 * This CEC implementation makes 1 retry, so if we
 		 * get a NACK, then that means it made 2 attempts.
 		 */
-		cec_transmit_done(hdmi->cec_adap, CEC_TX_STATUS_NACK,
+		cec_transmit_done(vc4_hdmi->cec_adap, CEC_TX_STATUS_NACK,
 				  0, 2, 0, 0);
 	}
 	return IRQ_HANDLED;
 }
 
-static void vc4_cec_read_msg(struct vc4_dev *vc4, u32 cntrl1)
+static void vc4_cec_read_msg(struct vc4_hdmi *vc4_hdmi, u32 cntrl1)
 {
-	struct cec_msg *msg = &vc4->hdmi->cec_rx_msg;
+	struct cec_msg *msg = &vc4_hdmi->cec_rx_msg;
 	unsigned int i;
 
 	msg->len = 1 + ((cntrl1 & VC4_HDMI_CEC_REC_WRD_CNT_MASK) >>
 					VC4_HDMI_CEC_REC_WRD_CNT_SHIFT);
 	for (i = 0; i < msg->len; i += 4) {
-		u32 val = HDMI_READ(VC4_HDMI_CEC_RX_DATA_1 + i);
+		u32 val = HDMI_READ(HDMI_CEC_RX_DATA_1 + i);
 
 		msg->msg[i] = val & 0xff;
 		msg->msg[i + 1] = (val >> 8) & 0xff;
@@ -1176,38 +1283,37 @@ static void vc4_cec_read_msg(struct vc4_dev *vc4, u32 cntrl1)
 
 static irqreturn_t vc4_cec_irq_handler(int irq, void *priv)
 {
-	struct vc4_dev *vc4 = priv;
-	struct vc4_hdmi *hdmi = vc4->hdmi;
-	u32 stat = HDMI_READ(VC4_HDMI_CPU_STATUS);
+	struct vc4_hdmi *vc4_hdmi = priv;
+	u32 stat = HDMI_READ(HDMI_CEC_CPU_STATUS);
 	u32 cntrl1, cntrl5;
 
 	if (!(stat & VC4_HDMI_CPU_CEC))
 		return IRQ_NONE;
-	hdmi->cec_rx_msg.len = 0;
-	cntrl1 = HDMI_READ(VC4_HDMI_CEC_CNTRL_1);
-	cntrl5 = HDMI_READ(VC4_HDMI_CEC_CNTRL_5);
-	hdmi->cec_irq_was_rx = cntrl5 & VC4_HDMI_CEC_RX_CEC_INT;
-	if (hdmi->cec_irq_was_rx) {
-		vc4_cec_read_msg(vc4, cntrl1);
+	vc4_hdmi->cec_rx_msg.len = 0;
+	cntrl1 = HDMI_READ(HDMI_CEC_CNTRL_1);
+	cntrl5 = HDMI_READ(HDMI_CEC_CNTRL_5);
+	vc4_hdmi->cec_irq_was_rx = cntrl5 & VC4_HDMI_CEC_RX_CEC_INT;
+	if (vc4_hdmi->cec_irq_was_rx) {
+		vc4_cec_read_msg(vc4_hdmi, cntrl1);
 		cntrl1 |= VC4_HDMI_CEC_CLEAR_RECEIVE_OFF;
-		HDMI_WRITE(VC4_HDMI_CEC_CNTRL_1, cntrl1);
+		HDMI_WRITE(HDMI_CEC_CNTRL_1, cntrl1);
 		cntrl1 &= ~VC4_HDMI_CEC_CLEAR_RECEIVE_OFF;
 	} else {
-		hdmi->cec_tx_ok = cntrl1 & VC4_HDMI_CEC_TX_STATUS_GOOD;
+		vc4_hdmi->cec_tx_ok = cntrl1 & VC4_HDMI_CEC_TX_STATUS_GOOD;
 		cntrl1 &= ~VC4_HDMI_CEC_START_XMIT_BEGIN;
 	}
-	HDMI_WRITE(VC4_HDMI_CEC_CNTRL_1, cntrl1);
-	HDMI_WRITE(VC4_HDMI_CPU_CLEAR, VC4_HDMI_CPU_CEC);
+	HDMI_WRITE(HDMI_CEC_CNTRL_1, cntrl1);
+	HDMI_WRITE(HDMI_CEC_CPU_CLEAR, VC4_HDMI_CPU_CEC);
 
 	return IRQ_WAKE_THREAD;
 }
 
 static int vc4_hdmi_cec_adap_enable(struct cec_adapter *adap, bool enable)
 {
-	struct vc4_dev *vc4 = cec_get_drvdata(adap);
+	struct vc4_hdmi *vc4_hdmi = cec_get_drvdata(adap);
 	/* clock period in microseconds */
 	const u32 usecs = 1000000 / CEC_CLOCK_FREQ;
-	u32 val = HDMI_READ(VC4_HDMI_CEC_CNTRL_5);
+	u32 val = HDMI_READ(HDMI_CEC_CNTRL_5);
 
 	val &= ~(VC4_HDMI_CEC_TX_SW_RESET | VC4_HDMI_CEC_RX_SW_RESET |
 		 VC4_HDMI_CEC_CNT_TO_4700_US_MASK |
@@ -1216,30 +1322,30 @@ static int vc4_hdmi_cec_adap_enable(struct cec_adapter *adap, bool enable)
 	       ((4500 / usecs) << VC4_HDMI_CEC_CNT_TO_4500_US_SHIFT);
 
 	if (enable) {
-		HDMI_WRITE(VC4_HDMI_CEC_CNTRL_5, val |
+		HDMI_WRITE(HDMI_CEC_CNTRL_5, val |
 			   VC4_HDMI_CEC_TX_SW_RESET | VC4_HDMI_CEC_RX_SW_RESET);
-		HDMI_WRITE(VC4_HDMI_CEC_CNTRL_5, val);
-		HDMI_WRITE(VC4_HDMI_CEC_CNTRL_2,
-			 ((1500 / usecs) << VC4_HDMI_CEC_CNT_TO_1500_US_SHIFT) |
-			 ((1300 / usecs) << VC4_HDMI_CEC_CNT_TO_1300_US_SHIFT) |
-			 ((800 / usecs) << VC4_HDMI_CEC_CNT_TO_800_US_SHIFT) |
-			 ((600 / usecs) << VC4_HDMI_CEC_CNT_TO_600_US_SHIFT) |
-			 ((400 / usecs) << VC4_HDMI_CEC_CNT_TO_400_US_SHIFT));
-		HDMI_WRITE(VC4_HDMI_CEC_CNTRL_3,
-			 ((2750 / usecs) << VC4_HDMI_CEC_CNT_TO_2750_US_SHIFT) |
-			 ((2400 / usecs) << VC4_HDMI_CEC_CNT_TO_2400_US_SHIFT) |
-			 ((2050 / usecs) << VC4_HDMI_CEC_CNT_TO_2050_US_SHIFT) |
-			 ((1700 / usecs) << VC4_HDMI_CEC_CNT_TO_1700_US_SHIFT));
-		HDMI_WRITE(VC4_HDMI_CEC_CNTRL_4,
-			 ((4300 / usecs) << VC4_HDMI_CEC_CNT_TO_4300_US_SHIFT) |
-			 ((3900 / usecs) << VC4_HDMI_CEC_CNT_TO_3900_US_SHIFT) |
-			 ((3600 / usecs) << VC4_HDMI_CEC_CNT_TO_3600_US_SHIFT) |
-			 ((3500 / usecs) << VC4_HDMI_CEC_CNT_TO_3500_US_SHIFT));
-
-		HDMI_WRITE(VC4_HDMI_CPU_MASK_CLEAR, VC4_HDMI_CPU_CEC);
+		HDMI_WRITE(HDMI_CEC_CNTRL_5, val);
+		HDMI_WRITE(HDMI_CEC_CNTRL_2,
+			   ((1500 / usecs) << VC4_HDMI_CEC_CNT_TO_1500_US_SHIFT) |
+			   ((1300 / usecs) << VC4_HDMI_CEC_CNT_TO_1300_US_SHIFT) |
+			   ((800 / usecs) << VC4_HDMI_CEC_CNT_TO_800_US_SHIFT) |
+			   ((600 / usecs) << VC4_HDMI_CEC_CNT_TO_600_US_SHIFT) |
+			   ((400 / usecs) << VC4_HDMI_CEC_CNT_TO_400_US_SHIFT));
+		HDMI_WRITE(HDMI_CEC_CNTRL_3,
+			   ((2750 / usecs) << VC4_HDMI_CEC_CNT_TO_2750_US_SHIFT) |
+			   ((2400 / usecs) << VC4_HDMI_CEC_CNT_TO_2400_US_SHIFT) |
+			   ((2050 / usecs) << VC4_HDMI_CEC_CNT_TO_2050_US_SHIFT) |
+			   ((1700 / usecs) << VC4_HDMI_CEC_CNT_TO_1700_US_SHIFT));
+		HDMI_WRITE(HDMI_CEC_CNTRL_4,
+			   ((4300 / usecs) << VC4_HDMI_CEC_CNT_TO_4300_US_SHIFT) |
+			   ((3900 / usecs) << VC4_HDMI_CEC_CNT_TO_3900_US_SHIFT) |
+			   ((3600 / usecs) << VC4_HDMI_CEC_CNT_TO_3600_US_SHIFT) |
+			   ((3500 / usecs) << VC4_HDMI_CEC_CNT_TO_3500_US_SHIFT));
+
+		HDMI_WRITE(HDMI_CEC_CPU_MASK_CLEAR, VC4_HDMI_CPU_CEC);
 	} else {
-		HDMI_WRITE(VC4_HDMI_CPU_MASK_SET, VC4_HDMI_CPU_CEC);
-		HDMI_WRITE(VC4_HDMI_CEC_CNTRL_5, val |
+		HDMI_WRITE(HDMI_CEC_CPU_MASK_SET, VC4_HDMI_CPU_CEC);
+		HDMI_WRITE(HDMI_CEC_CNTRL_5, val |
 			   VC4_HDMI_CEC_TX_SW_RESET | VC4_HDMI_CEC_RX_SW_RESET);
 	}
 	return 0;
@@ -1247,10 +1353,10 @@ static int vc4_hdmi_cec_adap_enable(struct cec_adapter *adap, bool enable)
 
 static int vc4_hdmi_cec_adap_log_addr(struct cec_adapter *adap, u8 log_addr)
 {
-	struct vc4_dev *vc4 = cec_get_drvdata(adap);
+	struct vc4_hdmi *vc4_hdmi = cec_get_drvdata(adap);
 
-	HDMI_WRITE(VC4_HDMI_CEC_CNTRL_1,
-		   (HDMI_READ(VC4_HDMI_CEC_CNTRL_1) & ~VC4_HDMI_CEC_ADDR_MASK) |
+	HDMI_WRITE(HDMI_CEC_CNTRL_1,
+		   (HDMI_READ(HDMI_CEC_CNTRL_1) & ~VC4_HDMI_CEC_ADDR_MASK) |
 		   (log_addr & 0xf) << VC4_HDMI_CEC_ADDR_SHIFT);
 	return 0;
 }
@@ -1258,25 +1364,25 @@ static int vc4_hdmi_cec_adap_log_addr(struct cec_adapter *adap, u8 log_addr)
 static int vc4_hdmi_cec_adap_transmit(struct cec_adapter *adap, u8 attempts,
 				      u32 signal_free_time, struct cec_msg *msg)
 {
-	struct vc4_dev *vc4 = cec_get_drvdata(adap);
+	struct vc4_hdmi *vc4_hdmi = cec_get_drvdata(adap);
 	u32 val;
 	unsigned int i;
 
 	for (i = 0; i < msg->len; i += 4)
-		HDMI_WRITE(VC4_HDMI_CEC_TX_DATA_1 + i,
+		HDMI_WRITE(HDMI_CEC_TX_DATA_1 + i,
 			   (msg->msg[i]) |
 			   (msg->msg[i + 1] << 8) |
 			   (msg->msg[i + 2] << 16) |
 			   (msg->msg[i + 3] << 24));
 
-	val = HDMI_READ(VC4_HDMI_CEC_CNTRL_1);
+	val = HDMI_READ(HDMI_CEC_CNTRL_1);
 	val &= ~VC4_HDMI_CEC_START_XMIT_BEGIN;
-	HDMI_WRITE(VC4_HDMI_CEC_CNTRL_1, val);
+	HDMI_WRITE(HDMI_CEC_CNTRL_1, val);
 	val &= ~VC4_HDMI_CEC_MESSAGE_LENGTH_MASK;
 	val |= (msg->len - 1) << VC4_HDMI_CEC_MESSAGE_LENGTH_SHIFT;
 	val |= VC4_HDMI_CEC_START_XMIT_BEGIN;
 
-	HDMI_WRITE(VC4_HDMI_CEC_CNTRL_1, val);
+	HDMI_WRITE(HDMI_CEC_CNTRL_1, val);
 	return 0;
 }
 
@@ -1285,61 +1391,275 @@ static const struct cec_adap_ops vc4_hdmi_cec_adap_ops = {
 	.adap_log_addr = vc4_hdmi_cec_adap_log_addr,
 	.adap_transmit = vc4_hdmi_cec_adap_transmit,
 };
-#endif
 
-static int vc4_hdmi_bind(struct device *dev, struct device *master, void *data)
+static int vc4_hdmi_cec_init(struct vc4_hdmi *vc4_hdmi)
 {
-#ifdef CONFIG_DRM_VC4_HDMI_CEC
 	struct cec_connector_info conn_info;
-#endif
-	struct platform_device *pdev = to_platform_device(dev);
-	struct drm_device *drm = dev_get_drvdata(master);
-	struct vc4_dev *vc4 = drm->dev_private;
-	struct vc4_hdmi *hdmi;
-	struct vc4_hdmi_encoder *vc4_hdmi_encoder;
-	struct device_node *ddc_node;
+	struct platform_device *pdev = vc4_hdmi->pdev;
 	u32 value;
 	int ret;
 
-	hdmi = devm_kzalloc(dev, sizeof(*hdmi), GFP_KERNEL);
-	if (!hdmi)
+	if (!vc4_hdmi->variant->cec_available)
+		return 0;
+
+	vc4_hdmi->cec_adap = cec_allocate_adapter(&vc4_hdmi_cec_adap_ops,
+						  vc4_hdmi, "vc4",
+						  CEC_CAP_DEFAULTS |
+						  CEC_CAP_CONNECTOR_INFO, 1);
+	ret = PTR_ERR_OR_ZERO(vc4_hdmi->cec_adap);
+	if (ret < 0)
+		return ret;
+
+	cec_fill_conn_info_from_drm(&conn_info, &vc4_hdmi->connector);
+	cec_s_conn_info(vc4_hdmi->cec_adap, &conn_info);
+
+	HDMI_WRITE(HDMI_CEC_CPU_MASK_SET, 0xffffffff);
+	value = HDMI_READ(HDMI_CEC_CNTRL_1);
+	value &= ~VC4_HDMI_CEC_DIV_CLK_CNT_MASK;
+	/*
+	 * Set the logical address to Unregistered and set the clock
+	 * divider: the hsm_clock rate and this divider setting will
+	 * give a 40 kHz CEC clock.
+	 */
+	value |= VC4_HDMI_CEC_ADDR_MASK |
+		 (4091 << VC4_HDMI_CEC_DIV_CLK_CNT_SHIFT);
+	HDMI_WRITE(HDMI_CEC_CNTRL_1, value);
+	ret = devm_request_threaded_irq(&pdev->dev, platform_get_irq(pdev, 0),
+					vc4_cec_irq_handler,
+					vc4_cec_irq_handler_thread, 0,
+					"vc4 hdmi cec", vc4_hdmi);
+	if (ret)
+		goto err_delete_cec_adap;
+
+	ret = cec_register_adapter(vc4_hdmi->cec_adap, &pdev->dev);
+	if (ret < 0)
+		goto err_delete_cec_adap;
+
+	return 0;
+
+err_delete_cec_adap:
+	cec_delete_adapter(vc4_hdmi->cec_adap);
+
+	return ret;
+}
+
+static void vc4_hdmi_cec_exit(struct vc4_hdmi *vc4_hdmi)
+{
+	cec_unregister_adapter(vc4_hdmi->cec_adap);
+}
+#else
+static int vc4_hdmi_cec_init(struct vc4_hdmi *vc4_hdmi)
+{
+	return 0;
+}
+
+static void vc4_hdmi_cec_exit(struct vc4_hdmi *vc4_hdmi) {};
+
+#endif
+
+static int vc4_hdmi_build_regset(struct vc4_hdmi *vc4_hdmi,
+				 struct debugfs_regset32 *regset,
+				 enum vc4_hdmi_regs reg)
+{
+	const struct vc4_hdmi_variant *variant = vc4_hdmi->variant;
+	struct debugfs_reg32 *regs, *new_regs;
+	unsigned int count = 0;
+	unsigned int i;
+
+	regs = kcalloc(variant->num_registers, sizeof(*regs),
+		       GFP_KERNEL);
+	if (!regs)
 		return -ENOMEM;
 
-	vc4_hdmi_encoder = devm_kzalloc(dev, sizeof(*vc4_hdmi_encoder),
-					GFP_KERNEL);
-	if (!vc4_hdmi_encoder)
+	for (i = 0; i < variant->num_registers; i++) {
+		const struct vc4_hdmi_register *field =	&variant->registers[i];
+
+		if (field->reg != reg)
+			continue;
+
+		regs[count].name = field->name;
+		regs[count].offset = field->offset;
+		count++;
+	}
+
+	new_regs = krealloc(regs, count * sizeof(*regs), GFP_KERNEL);
+	if (!new_regs)
 		return -ENOMEM;
-	vc4_hdmi_encoder->base.type = VC4_ENCODER_TYPE_HDMI;
-	hdmi->encoder = &vc4_hdmi_encoder->base.base;
-
-	hdmi->pdev = pdev;
-	hdmi->hdmicore_regs = vc4_ioremap_regs(pdev, 0);
-	if (IS_ERR(hdmi->hdmicore_regs))
-		return PTR_ERR(hdmi->hdmicore_regs);
-
-	hdmi->hd_regs = vc4_ioremap_regs(pdev, 1);
-	if (IS_ERR(hdmi->hd_regs))
-		return PTR_ERR(hdmi->hd_regs);
-
-	hdmi->hdmi_regset.base = hdmi->hdmicore_regs;
-	hdmi->hdmi_regset.regs = hdmi_regs;
-	hdmi->hdmi_regset.nregs = ARRAY_SIZE(hdmi_regs);
-	hdmi->hd_regset.base = hdmi->hd_regs;
-	hdmi->hd_regset.regs = hd_regs;
-	hdmi->hd_regset.nregs = ARRAY_SIZE(hd_regs);
-
-	hdmi->pixel_clock = devm_clk_get(dev, "pixel");
-	if (IS_ERR(hdmi->pixel_clock)) {
-		ret = PTR_ERR(hdmi->pixel_clock);
+
+	regset->base = __vc4_hdmi_get_field_base(vc4_hdmi, reg);
+	regset->regs = new_regs;
+	regset->nregs = count;
+
+	return 0;
+}
+
+static int vc4_hdmi_init_resources(struct vc4_hdmi *vc4_hdmi)
+{
+	struct platform_device *pdev = vc4_hdmi->pdev;
+	struct device *dev = &pdev->dev;
+	int ret;
+
+	vc4_hdmi->hdmicore_regs = vc4_ioremap_regs(pdev, 0);
+	if (IS_ERR(vc4_hdmi->hdmicore_regs))
+		return PTR_ERR(vc4_hdmi->hdmicore_regs);
+
+	vc4_hdmi->hd_regs = vc4_ioremap_regs(pdev, 1);
+	if (IS_ERR(vc4_hdmi->hd_regs))
+		return PTR_ERR(vc4_hdmi->hd_regs);
+
+	ret = vc4_hdmi_build_regset(vc4_hdmi, &vc4_hdmi->hd_regset, VC4_HD);
+	if (ret)
+		return ret;
+
+	ret = vc4_hdmi_build_regset(vc4_hdmi, &vc4_hdmi->hdmi_regset, VC4_HDMI);
+	if (ret)
+		return ret;
+
+	vc4_hdmi->pixel_clock = devm_clk_get(dev, "pixel");
+	if (IS_ERR(vc4_hdmi->pixel_clock)) {
+		ret = PTR_ERR(vc4_hdmi->pixel_clock);
 		if (ret != -EPROBE_DEFER)
 			DRM_ERROR("Failed to get pixel clock\n");
 		return ret;
 	}
-	hdmi->hsm_clock = devm_clk_get(dev, "hdmi");
-	if (IS_ERR(hdmi->hsm_clock)) {
+
+	vc4_hdmi->hsm_clock = devm_clk_get(dev, "hdmi");
+	if (IS_ERR(vc4_hdmi->hsm_clock)) {
 		DRM_ERROR("Failed to get HDMI state machine clock\n");
-		return PTR_ERR(hdmi->hsm_clock);
+		return PTR_ERR(vc4_hdmi->hsm_clock);
 	}
+	vc4_hdmi->audio_clock = vc4_hdmi->hsm_clock;
+
+	return 0;
+}
+
+static int vc5_hdmi_init_resources(struct vc4_hdmi *vc4_hdmi)
+{
+	struct platform_device *pdev = vc4_hdmi->pdev;
+	struct device *dev = &pdev->dev;
+	struct resource *res;
+
+	res = platform_get_resource_byname(pdev, IORESOURCE_MEM, "hdmi");
+	if (!res)
+		return -ENODEV;
+
+	vc4_hdmi->hdmicore_regs = devm_ioremap(dev, res->start,
+					       resource_size(res));
+	if (!vc4_hdmi->hdmicore_regs)
+		return -ENOMEM;
+
+	res = platform_get_resource_byname(pdev, IORESOURCE_MEM, "hd");
+	if (!res)
+		return -ENODEV;
+
+	vc4_hdmi->hd_regs = devm_ioremap(dev, res->start, resource_size(res));
+	if (!vc4_hdmi->hd_regs)
+		return -ENOMEM;
+
+	res = platform_get_resource_byname(pdev, IORESOURCE_MEM, "cec");
+	if (!res)
+		return -ENODEV;
+
+	vc4_hdmi->cec_regs = devm_ioremap(dev, res->start, resource_size(res));
+	if (!vc4_hdmi->cec_regs)
+		return -ENOMEM;
+
+	res = platform_get_resource_byname(pdev, IORESOURCE_MEM, "csc");
+	if (!res)
+		return -ENODEV;
+
+	vc4_hdmi->csc_regs = devm_ioremap(dev, res->start, resource_size(res));
+	if (!vc4_hdmi->csc_regs)
+		return -ENOMEM;
+
+	res = platform_get_resource_byname(pdev, IORESOURCE_MEM, "dvp");
+	if (!res)
+		return -ENODEV;
+
+	vc4_hdmi->dvp_regs = devm_ioremap(dev, res->start, resource_size(res));
+	if (!vc4_hdmi->dvp_regs)
+		return -ENOMEM;
+
+	res = platform_get_resource_byname(pdev, IORESOURCE_MEM, "phy");
+	if (!res)
+		return -ENODEV;
+
+	vc4_hdmi->phy_regs = devm_ioremap(dev, res->start, resource_size(res));
+	if (!vc4_hdmi->phy_regs)
+		return -ENOMEM;
+
+	res = platform_get_resource_byname(pdev, IORESOURCE_MEM, "packet");
+	if (!res)
+		return -ENODEV;
+
+	vc4_hdmi->ram_regs = devm_ioremap(dev, res->start, resource_size(res));
+	if (!vc4_hdmi->ram_regs)
+		return -ENOMEM;
+
+	res = platform_get_resource_byname(pdev, IORESOURCE_MEM, "rm");
+	if (!res)
+		return -ENODEV;
+
+	vc4_hdmi->rm_regs = devm_ioremap(dev, res->start, resource_size(res));
+	if (!vc4_hdmi->rm_regs)
+		return -ENOMEM;
+
+	vc4_hdmi->hsm_clock = devm_clk_get(dev, "hdmi");
+	if (IS_ERR(vc4_hdmi->hsm_clock)) {
+		DRM_ERROR("Failed to get HDMI state machine clock\n");
+		return PTR_ERR(vc4_hdmi->hsm_clock);
+	}
+
+	vc4_hdmi->pixel_bvb_clock = devm_clk_get(dev, "bvb");
+	if (IS_ERR(vc4_hdmi->pixel_bvb_clock)) {
+		DRM_ERROR("Failed to get pixel bvb clock\n");
+		return PTR_ERR(vc4_hdmi->pixel_bvb_clock);
+	}
+
+	vc4_hdmi->audio_clock = devm_clk_get(dev, "audio");
+	if (IS_ERR(vc4_hdmi->audio_clock)) {
+		DRM_ERROR("Failed to get audio clock\n");
+		return PTR_ERR(vc4_hdmi->audio_clock);
+	}
+
+	vc4_hdmi->reset = devm_reset_control_get(dev, NULL);
+	if (IS_ERR(vc4_hdmi->reset)) {
+		DRM_ERROR("Failed to get HDMI reset line\n");
+		return PTR_ERR(vc4_hdmi->reset);
+	}
+
+	return 0;
+}
+
+static int vc4_hdmi_bind(struct device *dev, struct device *master, void *data)
+{
+	const struct vc4_hdmi_variant *variant = of_device_get_match_data(dev);
+	struct platform_device *pdev = to_platform_device(dev);
+	struct drm_device *drm = dev_get_drvdata(master);
+	struct vc4_hdmi *vc4_hdmi;
+	struct drm_encoder *encoder;
+	struct device_node *ddc_node;
+	u32 value;
+	int ret;
+
+	vc4_hdmi = devm_kzalloc(dev, sizeof(*vc4_hdmi), GFP_KERNEL);
+	if (!vc4_hdmi)
+		return -ENOMEM;
+
+	dev_set_drvdata(dev, vc4_hdmi);
+	encoder = &vc4_hdmi->encoder.base.base;
+	vc4_hdmi->encoder.base.type = variant->encoder_type;
+	vc4_hdmi->encoder.base.pre_crtc_configure = vc4_hdmi_encoder_pre_crtc_configure;
+	vc4_hdmi->encoder.base.pre_crtc_enable = vc4_hdmi_encoder_pre_crtc_enable;
+	vc4_hdmi->encoder.base.post_crtc_enable = vc4_hdmi_encoder_post_crtc_enable;
+	vc4_hdmi->encoder.base.post_crtc_disable = vc4_hdmi_encoder_post_crtc_disable;
+	vc4_hdmi->encoder.base.post_crtc_powerdown = vc4_hdmi_encoder_post_crtc_powerdown;
+	vc4_hdmi->pdev = pdev;
+	vc4_hdmi->variant = variant;
+
+	ret = variant->init_resources(vc4_hdmi);
+	if (ret)
+		return ret;
 
 	ddc_node = of_parse_phandle(dev->of_node, "ddc", 0);
 	if (!ddc_node) {
@@ -1347,123 +1667,62 @@ static int vc4_hdmi_bind(struct device *dev, struct device *master, void *data)
 		return -ENODEV;
 	}
 
-	hdmi->ddc = of_find_i2c_adapter_by_node(ddc_node);
+	vc4_hdmi->ddc = of_find_i2c_adapter_by_node(ddc_node);
 	of_node_put(ddc_node);
-	if (!hdmi->ddc) {
+	if (!vc4_hdmi->ddc) {
 		DRM_DEBUG("Failed to get ddc i2c adapter by node\n");
 		return -EPROBE_DEFER;
 	}
 
-	/* This is the rate that is set by the firmware.  The number
-	 * needs to be a bit higher than the pixel clock rate
-	 * (generally 148.5Mhz).
-	 */
-	ret = clk_set_rate(hdmi->hsm_clock, HSM_CLOCK_FREQ);
-	if (ret) {
-		DRM_ERROR("Failed to set HSM clock rate: %d\n", ret);
-		goto err_put_i2c;
-	}
-
-	ret = clk_prepare_enable(hdmi->hsm_clock);
-	if (ret) {
-		DRM_ERROR("Failed to turn on HDMI state machine clock: %d\n",
-			  ret);
-		goto err_put_i2c;
-	}
-
 	/* Only use the GPIO HPD pin if present in the DT, otherwise
 	 * we'll use the HDMI core's register.
 	 */
 	if (of_find_property(dev->of_node, "hpd-gpios", &value)) {
 		enum of_gpio_flags hpd_gpio_flags;
 
-		hdmi->hpd_gpio = of_get_named_gpio_flags(dev->of_node,
-							 "hpd-gpios", 0,
-							 &hpd_gpio_flags);
-		if (hdmi->hpd_gpio < 0) {
-			ret = hdmi->hpd_gpio;
+		vc4_hdmi->hpd_gpio = of_get_named_gpio_flags(dev->of_node,
+							     "hpd-gpios", 0,
+							     &hpd_gpio_flags);
+		if (vc4_hdmi->hpd_gpio < 0) {
+			ret = vc4_hdmi->hpd_gpio;
 			goto err_unprepare_hsm;
 		}
 
-		hdmi->hpd_active_low = hpd_gpio_flags & OF_GPIO_ACTIVE_LOW;
+		vc4_hdmi->hpd_active_low = hpd_gpio_flags & OF_GPIO_ACTIVE_LOW;
 	}
 
-	vc4->hdmi = hdmi;
-
-	/* HDMI core must be enabled. */
-	if (!(HD_READ(VC4_HD_M_CTL) & VC4_HD_M_ENABLE)) {
-		HD_WRITE(VC4_HD_M_CTL, VC4_HD_M_SW_RST);
-		udelay(1);
-		HD_WRITE(VC4_HD_M_CTL, 0);
-
-		HD_WRITE(VC4_HD_M_CTL, VC4_HD_M_ENABLE);
-	}
 	pm_runtime_enable(dev);
 
-	drm_simple_encoder_init(drm, hdmi->encoder, DRM_MODE_ENCODER_TMDS);
-	drm_encoder_helper_add(hdmi->encoder, &vc4_hdmi_encoder_helper_funcs);
+	drm_simple_encoder_init(drm, encoder, DRM_MODE_ENCODER_TMDS);
+	drm_encoder_helper_add(encoder, &vc4_hdmi_encoder_helper_funcs);
 
-	hdmi->connector =
-		vc4_hdmi_connector_init(drm, hdmi->encoder, hdmi->ddc);
-	if (IS_ERR(hdmi->connector)) {
-		ret = PTR_ERR(hdmi->connector);
+	ret = vc4_hdmi_connector_init(drm, vc4_hdmi);
+	if (ret)
 		goto err_destroy_encoder;
-	}
-#ifdef CONFIG_DRM_VC4_HDMI_CEC
-	hdmi->cec_adap = cec_allocate_adapter(&vc4_hdmi_cec_adap_ops,
-					      vc4, "vc4",
-					      CEC_CAP_DEFAULTS |
-					      CEC_CAP_CONNECTOR_INFO, 1);
-	ret = PTR_ERR_OR_ZERO(hdmi->cec_adap);
-	if (ret < 0)
-		goto err_destroy_conn;
 
-	cec_fill_conn_info_from_drm(&conn_info, hdmi->connector);
-	cec_s_conn_info(hdmi->cec_adap, &conn_info);
-
-	HDMI_WRITE(VC4_HDMI_CPU_MASK_SET, 0xffffffff);
-	value = HDMI_READ(VC4_HDMI_CEC_CNTRL_1);
-	value &= ~VC4_HDMI_CEC_DIV_CLK_CNT_MASK;
-	/*
-	 * Set the logical address to Unregistered and set the clock
-	 * divider: the hsm_clock rate and this divider setting will
-	 * give a 40 kHz CEC clock.
-	 */
-	value |= VC4_HDMI_CEC_ADDR_MASK |
-		 (4091 << VC4_HDMI_CEC_DIV_CLK_CNT_SHIFT);
-	HDMI_WRITE(VC4_HDMI_CEC_CNTRL_1, value);
-	ret = devm_request_threaded_irq(dev, platform_get_irq(pdev, 0),
-					vc4_cec_irq_handler,
-					vc4_cec_irq_handler_thread, 0,
-					"vc4 hdmi cec", vc4);
+	ret = vc4_hdmi_cec_init(vc4_hdmi);
 	if (ret)
-		goto err_delete_cec_adap;
-	ret = cec_register_adapter(hdmi->cec_adap, dev);
-	if (ret < 0)
-		goto err_delete_cec_adap;
-#endif
+		goto err_destroy_conn;
 
-	ret = vc4_hdmi_audio_init(hdmi);
+	ret = vc4_hdmi_audio_init(vc4_hdmi);
 	if (ret)
-		goto err_destroy_encoder;
+		goto err_free_cec;
 
-	vc4_debugfs_add_file(drm, "hdmi_regs", vc4_hdmi_debugfs_regs, hdmi);
+	vc4_debugfs_add_file(drm, variant->debugfs_name,
+			     vc4_hdmi_debugfs_regs,
+			     vc4_hdmi);
 
 	return 0;
 
-#ifdef CONFIG_DRM_VC4_HDMI_CEC
-err_delete_cec_adap:
-	cec_delete_adapter(hdmi->cec_adap);
+err_free_cec:
+	vc4_hdmi_cec_exit(vc4_hdmi);
 err_destroy_conn:
-	vc4_hdmi_connector_destroy(hdmi->connector);
-#endif
+	vc4_hdmi_connector_destroy(&vc4_hdmi->connector);
 err_destroy_encoder:
-	drm_encoder_cleanup(hdmi->encoder);
+	drm_encoder_cleanup(encoder);
 err_unprepare_hsm:
-	clk_disable_unprepare(hdmi->hsm_clock);
 	pm_runtime_disable(dev);
-err_put_i2c:
-	put_device(&hdmi->ddc->dev);
+	put_device(&vc4_hdmi->ddc->dev);
 
 	return ret;
 }
@@ -1471,20 +1730,39 @@ err_put_i2c:
 static void vc4_hdmi_unbind(struct device *dev, struct device *master,
 			    void *data)
 {
-	struct drm_device *drm = dev_get_drvdata(master);
-	struct vc4_dev *vc4 = drm->dev_private;
-	struct vc4_hdmi *hdmi = vc4->hdmi;
+	struct vc4_hdmi *vc4_hdmi;
 
-	cec_unregister_adapter(hdmi->cec_adap);
-	vc4_hdmi_connector_destroy(hdmi->connector);
-	drm_encoder_cleanup(hdmi->encoder);
+	/*
+	 * ASoC makes it a bit hard to retrieve a pointer to the
+	 * vc4_hdmi structure. Registering the card will overwrite our
+	 * device drvdata with a pointer to the snd_soc_card structure,
+	 * which can then be used to retrieve whatever drvdata we want
+	 * to associate.
+	 *
+	 * However, that doesn't fly in the case where we wouldn't
+	 * register an ASoC card (because of an old DT that is missing
+	 * the dmas properties for example), then the card isn't
+	 * registered and the device drvdata wouldn't be set.
+	 *
+	 * We can deal with both cases by making sure a snd_soc_card
+	 * pointer and a vc4_hdmi structure are pointing to the same
+	 * memory address, so we can treat them indistinctly without any
+	 * issue.
+	 */
+	BUILD_BUG_ON(offsetof(struct vc4_hdmi_audio, card) != 0);
+	BUILD_BUG_ON(offsetof(struct vc4_hdmi, audio) != 0);
+	vc4_hdmi = dev_get_drvdata(dev);
 
-	clk_disable_unprepare(hdmi->hsm_clock);
-	pm_runtime_disable(dev);
+	kfree(vc4_hdmi->hdmi_regset.regs);
+	kfree(vc4_hdmi->hd_regset.regs);
 
-	put_device(&hdmi->ddc->dev);
+	vc4_hdmi_cec_exit(vc4_hdmi);
+	vc4_hdmi_connector_destroy(&vc4_hdmi->connector);
+	drm_encoder_cleanup(&vc4_hdmi->encoder.base.base);
+
+	pm_runtime_disable(dev);
 
-	vc4->hdmi = NULL;
+	put_device(&vc4_hdmi->ddc->dev);
 }
 
 static const struct component_ops vc4_hdmi_ops = {
@@ -1503,8 +1781,80 @@ static int vc4_hdmi_dev_remove(struct platform_device *pdev)
 	return 0;
 }
 
+static const struct vc4_hdmi_variant bcm2835_variant = {
+	.encoder_type		= VC4_ENCODER_TYPE_HDMI0,
+	.debugfs_name		= "hdmi_regs",
+	.card_name		= "vc4-hdmi",
+	.max_pixel_clock	= 162000000,
+	.cec_available		= true,
+	.registers		= vc4_hdmi_fields,
+	.num_registers		= ARRAY_SIZE(vc4_hdmi_fields),
+
+	.init_resources		= vc4_hdmi_init_resources,
+	.csc_setup		= vc4_hdmi_csc_setup,
+	.reset			= vc4_hdmi_reset,
+	.set_timings		= vc4_hdmi_set_timings,
+	.phy_init		= vc4_hdmi_phy_init,
+	.phy_disable		= vc4_hdmi_phy_disable,
+	.phy_rng_enable		= vc4_hdmi_phy_rng_enable,
+	.phy_rng_disable	= vc4_hdmi_phy_rng_disable,
+	.channel_map		= vc4_hdmi_channel_map,
+};
+
+static const struct vc4_hdmi_variant bcm2711_hdmi0_variant = {
+	.encoder_type		= VC4_ENCODER_TYPE_HDMI0,
+	.debugfs_name		= "hdmi0_regs",
+	.card_name		= "vc4-hdmi-0",
+	.max_pixel_clock	= 297000000,
+	.registers		= vc5_hdmi_hdmi0_fields,
+	.num_registers		= ARRAY_SIZE(vc5_hdmi_hdmi0_fields),
+	.phy_lane_mapping	= {
+		PHY_LANE_0,
+		PHY_LANE_1,
+		PHY_LANE_2,
+		PHY_LANE_CK,
+	},
+
+	.init_resources		= vc5_hdmi_init_resources,
+	.csc_setup		= vc5_hdmi_csc_setup,
+	.reset			= vc5_hdmi_reset,
+	.set_timings		= vc5_hdmi_set_timings,
+	.phy_init		= vc5_hdmi_phy_init,
+	.phy_disable		= vc5_hdmi_phy_disable,
+	.phy_rng_enable		= vc5_hdmi_phy_rng_enable,
+	.phy_rng_disable	= vc5_hdmi_phy_rng_disable,
+	.channel_map		= vc5_hdmi_channel_map,
+};
+
+static const struct vc4_hdmi_variant bcm2711_hdmi1_variant = {
+	.encoder_type		= VC4_ENCODER_TYPE_HDMI1,
+	.debugfs_name		= "hdmi1_regs",
+	.card_name		= "vc4-hdmi-1",
+	.max_pixel_clock	= 297000000,
+	.registers		= vc5_hdmi_hdmi1_fields,
+	.num_registers		= ARRAY_SIZE(vc5_hdmi_hdmi1_fields),
+	.phy_lane_mapping	= {
+		PHY_LANE_1,
+		PHY_LANE_0,
+		PHY_LANE_CK,
+		PHY_LANE_2,
+	},
+
+	.init_resources		= vc5_hdmi_init_resources,
+	.csc_setup		= vc5_hdmi_csc_setup,
+	.reset			= vc5_hdmi_reset,
+	.set_timings		= vc5_hdmi_set_timings,
+	.phy_init		= vc5_hdmi_phy_init,
+	.phy_disable		= vc5_hdmi_phy_disable,
+	.phy_rng_enable		= vc5_hdmi_phy_rng_enable,
+	.phy_rng_disable	= vc5_hdmi_phy_rng_disable,
+	.channel_map		= vc5_hdmi_channel_map,
+};
+
 static const struct of_device_id vc4_hdmi_dt_match[] = {
-	{ .compatible = "brcm,bcm2835-hdmi" },
+	{ .compatible = "brcm,bcm2835-hdmi", .data = &bcm2835_variant },
+	{ .compatible = "brcm,bcm2711-hdmi0", .data = &bcm2711_hdmi0_variant },
+	{ .compatible = "brcm,bcm2711-hdmi1", .data = &bcm2711_hdmi1_variant },
 	{}
 };
 
diff --git a/drivers/gpu/drm/vc4/vc4_hdmi.h b/drivers/gpu/drm/vc4/vc4_hdmi.h
new file mode 100644
index 000000000000..63c6f8bddf1d
--- /dev/null
+++ b/drivers/gpu/drm/vc4/vc4_hdmi.h
@@ -0,0 +1,184 @@
+#ifndef _VC4_HDMI_H_
+#define _VC4_HDMI_H_
+
+#include <drm/drm_connector.h>
+#include <media/cec.h>
+#include <sound/dmaengine_pcm.h>
+#include <sound/soc.h>
+
+#include "vc4_drv.h"
+
+/* VC4 HDMI encoder KMS struct */
+struct vc4_hdmi_encoder {
+	struct vc4_encoder base;
+	bool hdmi_monitor;
+	bool limited_rgb_range;
+};
+
+static inline struct vc4_hdmi_encoder *
+to_vc4_hdmi_encoder(struct drm_encoder *encoder)
+{
+	return container_of(encoder, struct vc4_hdmi_encoder, base.base);
+}
+
+struct drm_display_mode;
+
+struct vc4_hdmi;
+struct vc4_hdmi_register;
+
+enum vc4_hdmi_phy_channel {
+	PHY_LANE_0 = 0,
+	PHY_LANE_1,
+	PHY_LANE_2,
+	PHY_LANE_CK,
+};
+
+struct vc4_hdmi_variant {
+	/* Encoder Type for that controller */
+	enum vc4_encoder_type encoder_type;
+
+	/* ALSA card name */
+	const char *card_name;
+
+	/* Filename to expose the registers in debugfs */
+	const char *debugfs_name;
+
+	/* Set to true when the CEC support is available */
+	bool cec_available;
+
+	/* Maximum pixel clock supported by the controller (in Hz) */
+	unsigned long long max_pixel_clock;
+
+	/* List of the registers available on that variant */
+	const struct vc4_hdmi_register *registers;
+
+	/* Number of registers on that variant */
+	unsigned int num_registers;
+
+	/* BCM2711 Only.
+	 * The variants don't map the lane in the same order in the
+	 * PHY, so this is an array mapping the HDMI channel (index)
+	 * to the PHY lane (value).
+	 */
+	enum vc4_hdmi_phy_channel phy_lane_mapping[4];
+
+	/* Callback to get the resources (memory region, interrupts,
+	 * clocks, etc) for that variant.
+	 */
+	int (*init_resources)(struct vc4_hdmi *vc4_hdmi);
+
+	/* Callback to reset the HDMI block */
+	void (*reset)(struct vc4_hdmi *vc4_hdmi);
+
+	/* Callback to enable / disable the CSC */
+	void (*csc_setup)(struct vc4_hdmi *vc4_hdmi, bool enable);
+
+	/* Callback to configure the video timings in the HDMI block */
+	void (*set_timings)(struct vc4_hdmi *vc4_hdmi,
+			    struct drm_display_mode *mode);
+
+	/* Callback to initialize the PHY according to the mode */
+	void (*phy_init)(struct vc4_hdmi *vc4_hdmi,
+			 struct drm_display_mode *mode);
+
+	/* Callback to disable the PHY */
+	void (*phy_disable)(struct vc4_hdmi *vc4_hdmi);
+
+	/* Callback to enable the RNG in the PHY */
+	void (*phy_rng_enable)(struct vc4_hdmi *vc4_hdmi);
+
+	/* Callback to disable the RNG in the PHY */
+	void (*phy_rng_disable)(struct vc4_hdmi *vc4_hdmi);
+
+	/* Callback to get channel map */
+	u32 (*channel_map)(struct vc4_hdmi *vc4_hdmi, u32 channel_mask);
+};
+
+/* HDMI audio information */
+struct vc4_hdmi_audio {
+	struct snd_soc_card card;
+	struct snd_soc_dai_link link;
+	struct snd_soc_dai_link_component cpu;
+	struct snd_soc_dai_link_component codec;
+	struct snd_soc_dai_link_component platform;
+	int samplerate;
+	int channels;
+	struct snd_dmaengine_dai_dma_data dma_data;
+	struct snd_pcm_substream *substream;
+
+	bool streaming;
+};
+
+/* General HDMI hardware state. */
+struct vc4_hdmi {
+	struct vc4_hdmi_audio audio;
+
+	struct platform_device *pdev;
+	const struct vc4_hdmi_variant *variant;
+
+	struct vc4_hdmi_encoder encoder;
+	struct drm_connector connector;
+
+	struct i2c_adapter *ddc;
+	void __iomem *hdmicore_regs;
+	void __iomem *hd_regs;
+
+	/* VC5 Only */
+	void __iomem *cec_regs;
+	/* VC5 Only */
+	void __iomem *csc_regs;
+	/* VC5 Only */
+	void __iomem *dvp_regs;
+	/* VC5 Only */
+	void __iomem *phy_regs;
+	/* VC5 Only */
+	void __iomem *ram_regs;
+	/* VC5 Only */
+	void __iomem *rm_regs;
+
+	int hpd_gpio;
+	bool hpd_active_low;
+
+	struct cec_adapter *cec_adap;
+	struct cec_msg cec_rx_msg;
+	bool cec_tx_ok;
+	bool cec_irq_was_rx;
+
+	struct clk *pixel_clock;
+	struct clk *hsm_clock;
+	struct clk *audio_clock;
+	struct clk *pixel_bvb_clock;
+
+	struct reset_control *reset;
+
+	struct debugfs_regset32 hdmi_regset;
+	struct debugfs_regset32 hd_regset;
+};
+
+static inline struct vc4_hdmi *
+connector_to_vc4_hdmi(struct drm_connector *connector)
+{
+	return container_of(connector, struct vc4_hdmi, connector);
+}
+
+static inline struct vc4_hdmi *
+encoder_to_vc4_hdmi(struct drm_encoder *encoder)
+{
+	struct vc4_hdmi_encoder *_encoder = to_vc4_hdmi_encoder(encoder);
+
+	return container_of(_encoder, struct vc4_hdmi, encoder);
+}
+
+void vc4_hdmi_phy_init(struct vc4_hdmi *vc4_hdmi,
+		       struct drm_display_mode *mode);
+void vc4_hdmi_phy_disable(struct vc4_hdmi *vc4_hdmi);
+void vc4_hdmi_phy_rng_enable(struct vc4_hdmi *vc4_hdmi);
+void vc4_hdmi_phy_rng_disable(struct vc4_hdmi *vc4_hdmi);
+
+void vc5_hdmi_phy_init(struct vc4_hdmi *vc4_hdmi,
+		       struct drm_display_mode *mode);
+void vc5_hdmi_phy_disable(struct vc4_hdmi *vc4_hdmi);
+void vc5_hdmi_phy_rng_enable(struct vc4_hdmi *vc4_hdmi);
+void vc5_hdmi_phy_rng_disable(struct vc4_hdmi *vc4_hdmi);
+
+#endif /* _VC4_HDMI_H_ */
diff --git a/drivers/gpu/drm/vc4/vc4_hdmi_phy.c b/drivers/gpu/drm/vc4/vc4_hdmi_phy.c
new file mode 100644
index 000000000000..057796b54c51
--- /dev/null
+++ b/drivers/gpu/drm/vc4/vc4_hdmi_phy.c
@@ -0,0 +1,521 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Copyright (C) 2015 Broadcom
+ * Copyright (c) 2014 The Linux Foundation. All rights reserved.
+ * Copyright (C) 2013 Red Hat
+ * Author: Rob Clark <robdclark@gmail.com>
+ */
+
+#include "vc4_hdmi.h"
+#include "vc4_regs.h"
+#include "vc4_hdmi_regs.h"
+
+#define VC4_HDMI_TX_PHY_RESET_CTL_PLL_RESETB	BIT(5)
+#define VC4_HDMI_TX_PHY_RESET_CTL_PLLDIV_RESETB	BIT(4)
+#define VC4_HDMI_TX_PHY_RESET_CTL_TX_CK_RESET	BIT(3)
+#define VC4_HDMI_TX_PHY_RESET_CTL_TX_2_RESET	BIT(2)
+#define VC4_HDMI_TX_PHY_RESET_CTL_TX_1_RESET	BIT(1)
+#define VC4_HDMI_TX_PHY_RESET_CTL_TX_0_RESET	BIT(0)
+
+#define VC4_HDMI_TX_PHY_POWERDOWN_CTL_RNDGEN_PWRDN	BIT(4)
+
+#define VC4_HDMI_TX_PHY_CTL_0_PREEMP_2_PREEMP_SHIFT	29
+#define VC4_HDMI_TX_PHY_CTL_0_PREEMP_2_PREEMP_MASK	VC4_MASK(31, 29)
+#define VC4_HDMI_TX_PHY_CTL_0_PREEMP_2_MAINDRV_SHIFT	24
+#define VC4_HDMI_TX_PHY_CTL_0_PREEMP_2_MAINDRV_MASK	VC4_MASK(28, 24)
+#define VC4_HDMI_TX_PHY_CTL_0_PREEMP_1_PREEMP_SHIFT	21
+#define VC4_HDMI_TX_PHY_CTL_0_PREEMP_1_PREEMP_MASK	VC4_MASK(23, 21)
+#define VC4_HDMI_TX_PHY_CTL_0_PREEMP_1_MAINDRV_SHIFT	16
+#define VC4_HDMI_TX_PHY_CTL_0_PREEMP_1_MAINDRV_MASK	VC4_MASK(20, 16)
+#define VC4_HDMI_TX_PHY_CTL_0_PREEMP_0_PREEMP_SHIFT	13
+#define VC4_HDMI_TX_PHY_CTL_0_PREEMP_0_PREEMP_MASK	VC4_MASK(15, 13)
+#define VC4_HDMI_TX_PHY_CTL_0_PREEMP_0_MAINDRV_SHIFT	8
+#define VC4_HDMI_TX_PHY_CTL_0_PREEMP_0_MAINDRV_MASK	VC4_MASK(12, 8)
+#define VC4_HDMI_TX_PHY_CTL_0_PREEMP_CK_PREEMP_SHIFT	5
+#define VC4_HDMI_TX_PHY_CTL_0_PREEMP_CK_PREEMP_MASK	VC4_MASK(7, 5)
+#define VC4_HDMI_TX_PHY_CTL_0_PREEMP_CK_MAINDRV_SHIFT	0
+#define VC4_HDMI_TX_PHY_CTL_0_PREEMP_CK_MAINDRV_MASK	VC4_MASK(4, 0)
+
+#define VC4_HDMI_TX_PHY_CTL_1_RES_SEL_DATA2_SHIFT	15
+#define VC4_HDMI_TX_PHY_CTL_1_RES_SEL_DATA2_MASK	VC4_MASK(19, 15)
+#define VC4_HDMI_TX_PHY_CTL_1_RES_SEL_DATA1_SHIFT	10
+#define VC4_HDMI_TX_PHY_CTL_1_RES_SEL_DATA1_MASK	VC4_MASK(14, 10)
+#define VC4_HDMI_TX_PHY_CTL_1_RES_SEL_DATA0_SHIFT	5
+#define VC4_HDMI_TX_PHY_CTL_1_RES_SEL_DATA0_MASK	VC4_MASK(9, 5)
+#define VC4_HDMI_TX_PHY_CTL_1_RES_SEL_CK_SHIFT		0
+#define VC4_HDMI_TX_PHY_CTL_1_RES_SEL_CK_MASK		VC4_MASK(4, 0)
+
+#define VC4_HDMI_TX_PHY_CTL_2_VCO_GAIN_SHIFT		16
+#define VC4_HDMI_TX_PHY_CTL_2_VCO_GAIN_MASK		VC4_MASK(19, 16)
+#define VC4_HDMI_TX_PHY_CTL_2_TERM_RES_SELDATA2_SHIFT	12
+#define VC4_HDMI_TX_PHY_CTL_2_TERM_RES_SELDATA2_MASK	VC4_MASK(15, 12)
+#define VC4_HDMI_TX_PHY_CTL_2_TERM_RES_SELDATA1_SHIFT	8
+#define VC4_HDMI_TX_PHY_CTL_2_TERM_RES_SELDATA1_MASK	VC4_MASK(11, 8)
+#define VC4_HDMI_TX_PHY_CTL_2_TERM_RES_SELDATA0_SHIFT	4
+#define VC4_HDMI_TX_PHY_CTL_2_TERM_RES_SELDATA0_MASK	VC4_MASK(7, 4)
+#define VC4_HDMI_TX_PHY_CTL_2_TERM_RES_SELCK_SHIFT	0
+#define VC4_HDMI_TX_PHY_CTL_2_TERM_RES_SELCK_MASK	VC4_MASK(3, 0)
+
+#define VC4_HDMI_TX_PHY_CTL_3_RP_SHIFT			17
+#define VC4_HDMI_TX_PHY_CTL_3_RP_MASK			VC4_MASK(19, 17)
+#define VC4_HDMI_TX_PHY_CTL_3_RZ_SHIFT			12
+#define VC4_HDMI_TX_PHY_CTL_3_RZ_MASK			VC4_MASK(16, 12)
+#define VC4_HDMI_TX_PHY_CTL_3_CP1_SHIFT			10
+#define VC4_HDMI_TX_PHY_CTL_3_CP1_MASK			VC4_MASK(11, 10)
+#define VC4_HDMI_TX_PHY_CTL_3_CP_SHIFT			8
+#define VC4_HDMI_TX_PHY_CTL_3_CP_MASK			VC4_MASK(9, 8)
+#define VC4_HDMI_TX_PHY_CTL_3_CZ_SHIFT			6
+#define VC4_HDMI_TX_PHY_CTL_3_CZ_MASK			VC4_MASK(7, 6)
+#define VC4_HDMI_TX_PHY_CTL_3_ICP_SHIFT			0
+#define VC4_HDMI_TX_PHY_CTL_3_ICP_MASK			VC4_MASK(5, 0)
+
+#define VC4_HDMI_TX_PHY_PLL_CTL_0_MASH11_MODE		BIT(13)
+#define VC4_HDMI_TX_PHY_PLL_CTL_0_VC_RANGE_EN		BIT(12)
+#define VC4_HDMI_TX_PHY_PLL_CTL_0_EMULATE_VC_LOW	BIT(11)
+#define VC4_HDMI_TX_PHY_PLL_CTL_0_EMULATE_VC_HIGH	BIT(10)
+#define VC4_HDMI_TX_PHY_PLL_CTL_0_VCO_SEL_SHIFT		9
+#define VC4_HDMI_TX_PHY_PLL_CTL_0_VCO_SEL_MASK		VC4_MASK(9, 9)
+#define VC4_HDMI_TX_PHY_PLL_CTL_0_VCO_FB_DIV2		BIT(8)
+#define VC4_HDMI_TX_PHY_PLL_CTL_0_VCO_POST_DIV2		BIT(7)
+#define VC4_HDMI_TX_PHY_PLL_CTL_0_VCO_CONT_EN		BIT(6)
+#define VC4_HDMI_TX_PHY_PLL_CTL_0_ENA_VCO_CLK		BIT(5)
+
+#define VC4_HDMI_TX_PHY_PLL_CTL_1_CPP_SHIFT			16
+#define VC4_HDMI_TX_PHY_PLL_CTL_1_CPP_MASK			VC4_MASK(27, 16)
+#define VC4_HDMI_TX_PHY_PLL_CTL_1_FREQ_DOUBLER_DELAY_SHIFT	14
+#define VC4_HDMI_TX_PHY_PLL_CTL_1_FREQ_DOUBLER_DELAY_MASK	VC4_MASK(15, 14)
+#define VC4_HDMI_TX_PHY_PLL_CTL_1_FREQ_DOUBLER_ENABLE		BIT(13)
+#define VC4_HDMI_TX_PHY_PLL_CTL_1_POST_RST_SEL_SHIFT		11
+#define VC4_HDMI_TX_PHY_PLL_CTL_1_POST_RST_SEL_MASK		VC4_MASK(12, 11)
+
+#define VC4_HDMI_TX_PHY_CLK_DIV_VCO_SHIFT		8
+#define VC4_HDMI_TX_PHY_CLK_DIV_VCO_MASK		VC4_MASK(15, 8)
+
+#define VC4_HDMI_TX_PHY_PLL_CFG_PDIV_SHIFT		0
+#define VC4_HDMI_TX_PHY_PLL_CFG_PDIV_MASK		VC4_MASK(3, 0)
+
+#define VC4_HDMI_TX_PHY_CHANNEL_SWAP_TXCK_OUT_SEL_MASK	VC4_MASK(13, 12)
+#define VC4_HDMI_TX_PHY_CHANNEL_SWAP_TXCK_OUT_SEL_SHIFT	12
+#define VC4_HDMI_TX_PHY_CHANNEL_SWAP_TX2_OUT_SEL_MASK	VC4_MASK(9, 8)
+#define VC4_HDMI_TX_PHY_CHANNEL_SWAP_TX2_OUT_SEL_SHIFT	8
+#define VC4_HDMI_TX_PHY_CHANNEL_SWAP_TX1_OUT_SEL_MASK	VC4_MASK(5, 4)
+#define VC4_HDMI_TX_PHY_CHANNEL_SWAP_TX1_OUT_SEL_SHIFT	4
+#define VC4_HDMI_TX_PHY_CHANNEL_SWAP_TX0_OUT_SEL_MASK	VC4_MASK(1, 0)
+#define VC4_HDMI_TX_PHY_CHANNEL_SWAP_TX0_OUT_SEL_SHIFT	0
+
+#define VC4_HDMI_TX_PHY_PLL_CALIBRATION_CONFIG_1_MIN_LIMIT_MASK		VC4_MASK(27, 0)
+#define VC4_HDMI_TX_PHY_PLL_CALIBRATION_CONFIG_1_MIN_LIMIT_SHIFT	0
+
+#define VC4_HDMI_TX_PHY_PLL_CALIBRATION_CONFIG_2_MAX_LIMIT_MASK		VC4_MASK(27, 0)
+#define VC4_HDMI_TX_PHY_PLL_CALIBRATION_CONFIG_2_MAX_LIMIT_SHIFT	0
+
+#define VC4_HDMI_TX_PHY_PLL_CALIBRATION_CONFIG_4_STABLE_THRESHOLD_MASK	VC4_MASK(31, 16)
+#define VC4_HDMI_TX_PHY_PLL_CALIBRATION_CONFIG_4_STABLE_THRESHOLD_SHIFT	16
+#define VC4_HDMI_TX_PHY_PLL_CALIBRATION_CONFIG_4_HOLD_THRESHOLD_MASK	VC4_MASK(15, 0)
+#define VC4_HDMI_TX_PHY_PLL_CALIBRATION_CONFIG_4_HOLD_THRESHOLD_SHIFT	0
+
+#define VC4_HDMI_RM_CONTROL_EN_FREEZE_COUNTERS		BIT(19)
+#define VC4_HDMI_RM_CONTROL_EN_LOAD_INTEGRATOR		BIT(17)
+#define VC4_HDMI_RM_CONTROL_FREE_RUN			BIT(4)
+
+#define VC4_HDMI_RM_OFFSET_ONLY				BIT(31)
+#define VC4_HDMI_RM_OFFSET_OFFSET_SHIFT			0
+#define VC4_HDMI_RM_OFFSET_OFFSET_MASK			VC4_MASK(30, 0)
+
+#define VC4_HDMI_RM_FORMAT_SHIFT_SHIFT			24
+#define VC4_HDMI_RM_FORMAT_SHIFT_MASK			VC4_MASK(25, 24)
+
+#define OSCILLATOR_FREQUENCY	54000000
+
+void vc4_hdmi_phy_init(struct vc4_hdmi *vc4_hdmi, struct drm_display_mode *mode)
+{
+	/* PHY should be in reset, like
+	 * vc4_hdmi_encoder_disable() does.
+	 */
+
+	HDMI_WRITE(HDMI_TX_PHY_RESET_CTL, 0xf << 16);
+	HDMI_WRITE(HDMI_TX_PHY_RESET_CTL, 0);
+}
+
+void vc4_hdmi_phy_disable(struct vc4_hdmi *vc4_hdmi)
+{
+	HDMI_WRITE(HDMI_TX_PHY_RESET_CTL, 0xf << 16);
+}
+
+void vc4_hdmi_phy_rng_enable(struct vc4_hdmi *vc4_hdmi)
+{
+	HDMI_WRITE(HDMI_TX_PHY_CTL_0,
+		   HDMI_READ(HDMI_TX_PHY_CTL_0) &
+		   ~VC4_HDMI_TX_PHY_RNG_PWRDN);
+}
+
+void vc4_hdmi_phy_rng_disable(struct vc4_hdmi *vc4_hdmi)
+{
+	HDMI_WRITE(HDMI_TX_PHY_CTL_0,
+		   HDMI_READ(HDMI_TX_PHY_CTL_0) |
+		   VC4_HDMI_TX_PHY_RNG_PWRDN);
+}
+
+static unsigned long long
+phy_get_vco_freq(unsigned long long clock, u8 *vco_sel, u8 *vco_div)
+{
+	unsigned long long vco_freq = clock;
+	unsigned int _vco_div = 0;
+	unsigned int _vco_sel = 0;
+
+	while (vco_freq < 3000000000ULL) {
+		_vco_div++;
+		vco_freq = clock * _vco_div * 10;
+	}
+
+	if (vco_freq > 4500000000ULL)
+		_vco_sel = 1;
+
+	*vco_sel = _vco_sel;
+	*vco_div = _vco_div;
+
+	return vco_freq;
+}
+
+static u8 phy_get_cp_current(unsigned long vco_freq)
+{
+	if (vco_freq < 3700000000ULL)
+		return 0x1c;
+
+	return 0x18;
+}
+
+static u32 phy_get_rm_offset(unsigned long long vco_freq)
+{
+	unsigned long long fref = OSCILLATOR_FREQUENCY;
+	u64 offset = 0;
+
+	/* RM offset is stored as 9.22 format */
+	offset = vco_freq * 2;
+	offset = offset << 22;
+	do_div(offset, fref);
+	offset >>= 2;
+
+	return offset;
+}
+
+static u8 phy_get_vco_gain(unsigned long long vco_freq)
+{
+	if (vco_freq < 3350000000ULL)
+		return 0xf;
+
+	if (vco_freq < 3700000000ULL)
+		return 0xc;
+
+	if (vco_freq < 4050000000ULL)
+		return 0x6;
+
+	if (vco_freq < 4800000000ULL)
+		return 0x5;
+
+	if (vco_freq < 5200000000ULL)
+		return 0x7;
+
+	return 0x2;
+}
+
+struct phy_lane_settings {
+	struct {
+		u8 preemphasis;
+		u8 main_driver;
+	} amplitude;
+
+	u8 res_sel_data;
+	u8 term_res_sel_data;
+};
+
+struct phy_settings {
+	unsigned long long min_rate;
+	unsigned long long max_rate;
+	struct phy_lane_settings channel[3];
+	struct phy_lane_settings clock;
+};
+
+static const struct phy_settings vc5_hdmi_phy_settings[] = {
+	{
+		0, 50000000,
+		{
+			{{0x0, 0x0A}, 0x12, 0x0},
+			{{0x0, 0x0A}, 0x12, 0x0},
+			{{0x0, 0x0A}, 0x12, 0x0}
+		},
+		{{0x0, 0x0A}, 0x18, 0x0},
+	},
+	{
+		50000001, 75000000,
+		{
+			{{0x0, 0x09}, 0x12, 0x0},
+			{{0x0, 0x09}, 0x12, 0x0},
+			{{0x0, 0x09}, 0x12, 0x0}
+		},
+		{{0x0, 0x0C}, 0x18, 0x3},
+	},
+	{
+		75000001,   165000000,
+		{
+			{{0x0, 0x09}, 0x12, 0x0},
+			{{0x0, 0x09}, 0x12, 0x0},
+			{{0x0, 0x09}, 0x12, 0x0}
+		},
+		{{0x0, 0x0C}, 0x18, 0x3},
+	},
+	{
+		165000001,  250000000,
+		{
+			{{0x0, 0x0F}, 0x12, 0x1},
+			{{0x0, 0x0F}, 0x12, 0x1},
+			{{0x0, 0x0F}, 0x12, 0x1}
+		},
+		{{0x0, 0x0C}, 0x18, 0x3},
+	},
+	{
+		250000001,  340000000,
+		{
+			{{0x2, 0x0D}, 0x12, 0x1},
+			{{0x2, 0x0D}, 0x12, 0x1},
+			{{0x2, 0x0D}, 0x12, 0x1}
+		},
+		{{0x0, 0x0C}, 0x18, 0xF},
+	},
+	{
+		340000001,  450000000,
+		{
+			{{0x0, 0x1B}, 0x12, 0xF},
+			{{0x0, 0x1B}, 0x12, 0xF},
+			{{0x0, 0x1B}, 0x12, 0xF}
+		},
+		{{0x0, 0x0A}, 0x12, 0xF},
+	},
+	{
+		450000001,  600000000,
+		{
+			{{0x0, 0x1C}, 0x12, 0xF},
+			{{0x0, 0x1C}, 0x12, 0xF},
+			{{0x0, 0x1C}, 0x12, 0xF}
+		},
+		{{0x0, 0x0B}, 0x13, 0xF},
+	},
+};
+
+static const struct phy_settings *phy_get_settings(unsigned long long tmds_rate)
+{
+	unsigned int count = ARRAY_SIZE(vc5_hdmi_phy_settings);
+	unsigned int i;
+
+	for (i = 0; i < count; i++) {
+		const struct phy_settings *s = &vc5_hdmi_phy_settings[i];
+
+		if (tmds_rate >= s->min_rate && tmds_rate <= s->max_rate)
+			return s;
+	}
+
+	/*
+	 * If the pixel clock exceeds our max setting, try the max
+	 * setting anyway.
+	 */
+	return &vc5_hdmi_phy_settings[count - 1];
+}
+
+static const struct phy_lane_settings *
+phy_get_channel_settings(enum vc4_hdmi_phy_channel chan,
+			 unsigned long long tmds_rate)
+{
+	const struct phy_settings *settings = phy_get_settings(tmds_rate);
+
+	if (chan == PHY_LANE_CK)
+		return &settings->clock;
+
+	return &settings->channel[chan];
+}
+
+static void vc5_hdmi_reset_phy(struct vc4_hdmi *vc4_hdmi)
+{
+	HDMI_WRITE(HDMI_TX_PHY_RESET_CTL, 0x0f);
+	HDMI_WRITE(HDMI_TX_PHY_POWERDOWN_CTL, BIT(10));
+}
+
+void vc5_hdmi_phy_init(struct vc4_hdmi *vc4_hdmi, struct drm_display_mode *mode)
+{
+	const struct phy_lane_settings *chan0_settings, *chan1_settings, *chan2_settings, *clock_settings;
+	const struct vc4_hdmi_variant *variant = vc4_hdmi->variant;
+	unsigned long long pixel_freq = mode->clock * 1000;
+	unsigned long long vco_freq;
+	unsigned char word_sel;
+	u8 vco_sel, vco_div;
+
+	vco_freq = phy_get_vco_freq(pixel_freq, &vco_sel, &vco_div);
+
+	vc5_hdmi_reset_phy(vc4_hdmi);
+
+	HDMI_WRITE(HDMI_TX_PHY_POWERDOWN_CTL,
+		   VC4_HDMI_TX_PHY_POWERDOWN_CTL_RNDGEN_PWRDN);
+
+	HDMI_WRITE(HDMI_TX_PHY_RESET_CTL,
+		   HDMI_READ(HDMI_TX_PHY_RESET_CTL) &
+		   ~VC4_HDMI_TX_PHY_RESET_CTL_TX_0_RESET &
+		   ~VC4_HDMI_TX_PHY_RESET_CTL_TX_1_RESET &
+		   ~VC4_HDMI_TX_PHY_RESET_CTL_TX_2_RESET &
+		   ~VC4_HDMI_TX_PHY_RESET_CTL_TX_CK_RESET);
+
+	HDMI_WRITE(HDMI_RM_CONTROL,
+		   HDMI_READ(HDMI_RM_CONTROL) |
+		   VC4_HDMI_RM_CONTROL_EN_FREEZE_COUNTERS |
+		   VC4_HDMI_RM_CONTROL_EN_LOAD_INTEGRATOR |
+		   VC4_HDMI_RM_CONTROL_FREE_RUN);
+
+	HDMI_WRITE(HDMI_TX_PHY_PLL_CALIBRATION_CONFIG_1,
+		   (HDMI_READ(HDMI_TX_PHY_PLL_CALIBRATION_CONFIG_1) &
+		    ~VC4_HDMI_TX_PHY_PLL_CALIBRATION_CONFIG_1_MIN_LIMIT_MASK) |
+		   VC4_SET_FIELD(0, VC4_HDMI_TX_PHY_PLL_CALIBRATION_CONFIG_1_MIN_LIMIT));
+
+	HDMI_WRITE(HDMI_TX_PHY_PLL_CALIBRATION_CONFIG_2,
+		   (HDMI_READ(HDMI_TX_PHY_PLL_CALIBRATION_CONFIG_2) &
+		    ~VC4_HDMI_TX_PHY_PLL_CALIBRATION_CONFIG_2_MAX_LIMIT_MASK) |
+		   VC4_SET_FIELD(0, VC4_HDMI_TX_PHY_PLL_CALIBRATION_CONFIG_2_MAX_LIMIT));
+
+	HDMI_WRITE(HDMI_RM_OFFSET,
+		   VC4_SET_FIELD(phy_get_rm_offset(vco_freq),
+				 VC4_HDMI_RM_OFFSET_OFFSET) |
+		   VC4_HDMI_RM_OFFSET_ONLY);
+
+	HDMI_WRITE(HDMI_TX_PHY_CLK_DIV,
+		   VC4_SET_FIELD(vco_div, VC4_HDMI_TX_PHY_CLK_DIV_VCO));
+
+	HDMI_WRITE(HDMI_TX_PHY_PLL_CALIBRATION_CONFIG_4,
+		   VC4_SET_FIELD(0xe147, VC4_HDMI_TX_PHY_PLL_CALIBRATION_CONFIG_4_HOLD_THRESHOLD) |
+		   VC4_SET_FIELD(0xe14, VC4_HDMI_TX_PHY_PLL_CALIBRATION_CONFIG_4_STABLE_THRESHOLD));
+
+	HDMI_WRITE(HDMI_TX_PHY_PLL_CTL_0,
+		   VC4_HDMI_TX_PHY_PLL_CTL_0_ENA_VCO_CLK |
+		   VC4_HDMI_TX_PHY_PLL_CTL_0_VCO_CONT_EN |
+		   VC4_HDMI_TX_PHY_PLL_CTL_0_MASH11_MODE |
+		   VC4_SET_FIELD(vco_sel, VC4_HDMI_TX_PHY_PLL_CTL_0_VCO_SEL));
+
+	HDMI_WRITE(HDMI_TX_PHY_PLL_CTL_1,
+		   HDMI_READ(HDMI_TX_PHY_PLL_CTL_1) |
+		   VC4_HDMI_TX_PHY_PLL_CTL_1_FREQ_DOUBLER_ENABLE |
+		   VC4_SET_FIELD(3, VC4_HDMI_TX_PHY_PLL_CTL_1_POST_RST_SEL) |
+		   VC4_SET_FIELD(1, VC4_HDMI_TX_PHY_PLL_CTL_1_FREQ_DOUBLER_DELAY) |
+		   VC4_SET_FIELD(0x8a, VC4_HDMI_TX_PHY_PLL_CTL_1_CPP));
+
+	HDMI_WRITE(HDMI_RM_FORMAT,
+		   HDMI_READ(HDMI_RM_FORMAT) |
+		   VC4_SET_FIELD(2, VC4_HDMI_RM_FORMAT_SHIFT));
+
+	HDMI_WRITE(HDMI_TX_PHY_PLL_CFG,
+		   HDMI_READ(HDMI_TX_PHY_PLL_CFG) |
+		   VC4_SET_FIELD(1, VC4_HDMI_TX_PHY_PLL_CFG_PDIV));
+
+	if (pixel_freq >= 340000000)
+		word_sel = 3;
+	else
+		word_sel = 0;
+	HDMI_WRITE(HDMI_TX_PHY_TMDS_CLK_WORD_SEL, word_sel);
+
+	HDMI_WRITE(HDMI_TX_PHY_CTL_3,
+		   VC4_SET_FIELD(phy_get_cp_current(vco_freq),
+				 VC4_HDMI_TX_PHY_CTL_3_ICP) |
+		   VC4_SET_FIELD(1, VC4_HDMI_TX_PHY_CTL_3_CP) |
+		   VC4_SET_FIELD(1, VC4_HDMI_TX_PHY_CTL_3_CP1) |
+		   VC4_SET_FIELD(3, VC4_HDMI_TX_PHY_CTL_3_CZ) |
+		   VC4_SET_FIELD(4, VC4_HDMI_TX_PHY_CTL_3_RP) |
+		   VC4_SET_FIELD(6, VC4_HDMI_TX_PHY_CTL_3_RZ));
+
+	chan0_settings =
+		phy_get_channel_settings(variant->phy_lane_mapping[PHY_LANE_0],
+					 pixel_freq);
+	chan1_settings =
+		phy_get_channel_settings(variant->phy_lane_mapping[PHY_LANE_1],
+					 pixel_freq);
+	chan2_settings =
+		phy_get_channel_settings(variant->phy_lane_mapping[PHY_LANE_2],
+					 pixel_freq);
+	clock_settings =
+		phy_get_channel_settings(variant->phy_lane_mapping[PHY_LANE_CK],
+					 pixel_freq);
+
+	HDMI_WRITE(HDMI_TX_PHY_CTL_0,
+		   VC4_SET_FIELD(chan0_settings->amplitude.preemphasis,
+				 VC4_HDMI_TX_PHY_CTL_0_PREEMP_0_PREEMP) |
+		   VC4_SET_FIELD(chan0_settings->amplitude.main_driver,
+				 VC4_HDMI_TX_PHY_CTL_0_PREEMP_0_MAINDRV) |
+		   VC4_SET_FIELD(chan1_settings->amplitude.preemphasis,
+				 VC4_HDMI_TX_PHY_CTL_0_PREEMP_1_PREEMP) |
+		   VC4_SET_FIELD(chan1_settings->amplitude.main_driver,
+				 VC4_HDMI_TX_PHY_CTL_0_PREEMP_1_MAINDRV) |
+		   VC4_SET_FIELD(chan2_settings->amplitude.preemphasis,
+				 VC4_HDMI_TX_PHY_CTL_0_PREEMP_2_PREEMP) |
+		   VC4_SET_FIELD(chan2_settings->amplitude.main_driver,
+				 VC4_HDMI_TX_PHY_CTL_0_PREEMP_2_MAINDRV) |
+		   VC4_SET_FIELD(clock_settings->amplitude.preemphasis,
+				 VC4_HDMI_TX_PHY_CTL_0_PREEMP_CK_PREEMP) |
+		   VC4_SET_FIELD(clock_settings->amplitude.main_driver,
+				 VC4_HDMI_TX_PHY_CTL_0_PREEMP_CK_MAINDRV));
+
+	HDMI_WRITE(HDMI_TX_PHY_CTL_1,
+		   HDMI_READ(HDMI_TX_PHY_CTL_1) |
+		   VC4_SET_FIELD(chan0_settings->res_sel_data,
+				 VC4_HDMI_TX_PHY_CTL_1_RES_SEL_DATA0) |
+		   VC4_SET_FIELD(chan1_settings->res_sel_data,
+				 VC4_HDMI_TX_PHY_CTL_1_RES_SEL_DATA1) |
+		   VC4_SET_FIELD(chan2_settings->res_sel_data,
+				 VC4_HDMI_TX_PHY_CTL_1_RES_SEL_DATA2) |
+		   VC4_SET_FIELD(clock_settings->res_sel_data,
+				 VC4_HDMI_TX_PHY_CTL_1_RES_SEL_CK));
+
+	HDMI_WRITE(HDMI_TX_PHY_CTL_2,
+		   VC4_SET_FIELD(chan0_settings->term_res_sel_data,
+				 VC4_HDMI_TX_PHY_CTL_2_TERM_RES_SELDATA0) |
+		   VC4_SET_FIELD(chan1_settings->term_res_sel_data,
+				 VC4_HDMI_TX_PHY_CTL_2_TERM_RES_SELDATA1) |
+		   VC4_SET_FIELD(chan2_settings->term_res_sel_data,
+				 VC4_HDMI_TX_PHY_CTL_2_TERM_RES_SELDATA2) |
+		   VC4_SET_FIELD(clock_settings->term_res_sel_data,
+				 VC4_HDMI_TX_PHY_CTL_2_TERM_RES_SELCK) |
+		   VC4_SET_FIELD(phy_get_vco_gain(vco_freq),
+				 VC4_HDMI_TX_PHY_CTL_2_VCO_GAIN));
+
+	HDMI_WRITE(HDMI_TX_PHY_CHANNEL_SWAP,
+		   VC4_SET_FIELD(variant->phy_lane_mapping[PHY_LANE_0],
+				 VC4_HDMI_TX_PHY_CHANNEL_SWAP_TX0_OUT_SEL) |
+		   VC4_SET_FIELD(variant->phy_lane_mapping[PHY_LANE_1],
+				 VC4_HDMI_TX_PHY_CHANNEL_SWAP_TX1_OUT_SEL) |
+		   VC4_SET_FIELD(variant->phy_lane_mapping[PHY_LANE_2],
+				 VC4_HDMI_TX_PHY_CHANNEL_SWAP_TX2_OUT_SEL) |
+		   VC4_SET_FIELD(variant->phy_lane_mapping[PHY_LANE_CK],
+				 VC4_HDMI_TX_PHY_CHANNEL_SWAP_TXCK_OUT_SEL));
+
+	HDMI_WRITE(HDMI_TX_PHY_RESET_CTL,
+		   HDMI_READ(HDMI_TX_PHY_RESET_CTL) &
+		   ~(VC4_HDMI_TX_PHY_RESET_CTL_PLL_RESETB |
+		     VC4_HDMI_TX_PHY_RESET_CTL_PLLDIV_RESETB));
+
+	HDMI_WRITE(HDMI_TX_PHY_RESET_CTL,
+		   HDMI_READ(HDMI_TX_PHY_RESET_CTL) |
+		   VC4_HDMI_TX_PHY_RESET_CTL_PLL_RESETB |
+		   VC4_HDMI_TX_PHY_RESET_CTL_PLLDIV_RESETB);
+}
+
+void vc5_hdmi_phy_disable(struct vc4_hdmi *vc4_hdmi)
+{
+	vc5_hdmi_reset_phy(vc4_hdmi);
+}
+
+void vc5_hdmi_phy_rng_enable(struct vc4_hdmi *vc4_hdmi)
+{
+	HDMI_WRITE(HDMI_TX_PHY_POWERDOWN_CTL,
+		   HDMI_READ(HDMI_TX_PHY_POWERDOWN_CTL) &
+		   ~VC4_HDMI_TX_PHY_POWERDOWN_CTL_RNDGEN_PWRDN);
+}
+
+void vc5_hdmi_phy_rng_disable(struct vc4_hdmi *vc4_hdmi)
+{
+	HDMI_WRITE(HDMI_TX_PHY_POWERDOWN_CTL,
+		   HDMI_READ(HDMI_TX_PHY_POWERDOWN_CTL) |
+		   VC4_HDMI_TX_PHY_POWERDOWN_CTL_RNDGEN_PWRDN);
+}
diff --git a/drivers/gpu/drm/vc4/vc4_hdmi_regs.h b/drivers/gpu/drm/vc4/vc4_hdmi_regs.h
new file mode 100644
index 000000000000..7c6b4818f245
--- /dev/null
+++ b/drivers/gpu/drm/vc4/vc4_hdmi_regs.h
@@ -0,0 +1,442 @@
+#ifndef _VC4_HDMI_REGS_H_
+#define _VC4_HDMI_REGS_H_
+
+#include "vc4_hdmi.h"
+
+#define VC4_HDMI_PACKET_STRIDE			0x24
+
+enum vc4_hdmi_regs {
+	VC4_INVALID = 0,
+	VC4_HDMI,
+	VC4_HD,
+	VC5_CEC,
+	VC5_CSC,
+	VC5_DVP,
+	VC5_PHY,
+	VC5_RAM,
+	VC5_RM,
+};
+
+enum vc4_hdmi_field {
+	HDMI_AUDIO_PACKET_CONFIG,
+	HDMI_CEC_CNTRL_1,
+	HDMI_CEC_CNTRL_2,
+	HDMI_CEC_CNTRL_3,
+	HDMI_CEC_CNTRL_4,
+	HDMI_CEC_CNTRL_5,
+	HDMI_CEC_CPU_CLEAR,
+	HDMI_CEC_CPU_MASK_CLEAR,
+	HDMI_CEC_CPU_MASK_SET,
+	HDMI_CEC_CPU_MASK_STATUS,
+	HDMI_CEC_CPU_STATUS,
+
+	/*
+	 * Transmit data, first byte is low byte of the 32-bit reg.
+	 * MSB of each byte transmitted first.
+	 */
+	HDMI_CEC_RX_DATA_1,
+	HDMI_CEC_RX_DATA_2,
+	HDMI_CEC_RX_DATA_3,
+	HDMI_CEC_RX_DATA_4,
+	HDMI_CEC_TX_DATA_1,
+	HDMI_CEC_TX_DATA_2,
+	HDMI_CEC_TX_DATA_3,
+	HDMI_CEC_TX_DATA_4,
+	HDMI_CLOCK_STOP,
+	HDMI_CORE_REV,
+	HDMI_CRP_CFG,
+	HDMI_CSC_12_11,
+	HDMI_CSC_14_13,
+	HDMI_CSC_22_21,
+	HDMI_CSC_24_23,
+	HDMI_CSC_32_31,
+	HDMI_CSC_34_33,
+	HDMI_CSC_CTL,
+
+	/*
+	 * 20-bit fields containing CTS values to be transmitted if
+	 * !EXTERNAL_CTS_EN
+	 */
+	HDMI_CTS_0,
+	HDMI_CTS_1,
+	HDMI_DVP_CTL,
+	HDMI_FIFO_CTL,
+	HDMI_FRAME_COUNT,
+	HDMI_HORZA,
+	HDMI_HORZB,
+	HDMI_HOTPLUG,
+	HDMI_HOTPLUG_INT,
+
+	/*
+	 * 3 bits per field, where each field maps from that
+	 * corresponding MAI bus channel to the given HDMI channel.
+	 */
+	HDMI_MAI_CHANNEL_MAP,
+	HDMI_MAI_CONFIG,
+	HDMI_MAI_CTL,
+
+	/*
+	 * Register for DMAing in audio data to be transported over
+	 * the MAI bus to the Falcon core.
+	 */
+	HDMI_MAI_DATA,
+
+	/* Format header to be placed on the MAI data. Unused. */
+	HDMI_MAI_FMT,
+
+	/* Last received format word on the MAI bus. */
+	HDMI_MAI_FORMAT,
+	HDMI_MAI_SMP,
+	HDMI_MAI_THR,
+	HDMI_M_CTL,
+	HDMI_RAM_PACKET_CONFIG,
+	HDMI_RAM_PACKET_START,
+	HDMI_RAM_PACKET_STATUS,
+	HDMI_RM_CONTROL,
+	HDMI_RM_FORMAT,
+	HDMI_RM_OFFSET,
+	HDMI_SCHEDULER_CONTROL,
+	HDMI_SW_RESET_CONTROL,
+	HDMI_TX_PHY_CHANNEL_SWAP,
+	HDMI_TX_PHY_CLK_DIV,
+	HDMI_TX_PHY_CTL_0,
+	HDMI_TX_PHY_CTL_1,
+	HDMI_TX_PHY_CTL_2,
+	HDMI_TX_PHY_CTL_3,
+	HDMI_TX_PHY_PLL_CALIBRATION_CONFIG_1,
+	HDMI_TX_PHY_PLL_CALIBRATION_CONFIG_2,
+	HDMI_TX_PHY_PLL_CALIBRATION_CONFIG_4,
+	HDMI_TX_PHY_PLL_CFG,
+	HDMI_TX_PHY_PLL_CTL_0,
+	HDMI_TX_PHY_PLL_CTL_1,
+	HDMI_TX_PHY_POWERDOWN_CTL,
+	HDMI_TX_PHY_RESET_CTL,
+	HDMI_TX_PHY_TMDS_CLK_WORD_SEL,
+	HDMI_VEC_INTERFACE_XBAR,
+	HDMI_VERTA0,
+	HDMI_VERTA1,
+	HDMI_VERTB0,
+	HDMI_VERTB1,
+	HDMI_VID_CTL,
+};
+
+struct vc4_hdmi_register {
+	char *name;
+	enum vc4_hdmi_regs reg;
+	unsigned int offset;
+};
+
+#define _VC4_REG(_base, _reg, _offset)	\
+	[_reg] = {				\
+		.name = #_reg,			\
+		.reg = _base,			\
+		.offset = _offset,		\
+	}
+
+#define VC4_HD_REG(reg, offset)		_VC4_REG(VC4_HD, reg, offset)
+#define VC4_HDMI_REG(reg, offset)	_VC4_REG(VC4_HDMI, reg, offset)
+#define VC5_CEC_REG(reg, offset)	_VC4_REG(VC5_CEC, reg, offset)
+#define VC5_CSC_REG(reg, offset)	_VC4_REG(VC5_CSC, reg, offset)
+#define VC5_DVP_REG(reg, offset)	_VC4_REG(VC5_DVP, reg, offset)
+#define VC5_PHY_REG(reg, offset)	_VC4_REG(VC5_PHY, reg, offset)
+#define VC5_RAM_REG(reg, offset)	_VC4_REG(VC5_RAM, reg, offset)
+#define VC5_RM_REG(reg, offset)		_VC4_REG(VC5_RM, reg, offset)
+
+static const struct vc4_hdmi_register vc4_hdmi_fields[] = {
+	VC4_HD_REG(HDMI_M_CTL, 0x000c),
+	VC4_HD_REG(HDMI_MAI_CTL, 0x0014),
+	VC4_HD_REG(HDMI_MAI_THR, 0x0018),
+	VC4_HD_REG(HDMI_MAI_FMT, 0x001c),
+	VC4_HD_REG(HDMI_MAI_DATA, 0x0020),
+	VC4_HD_REG(HDMI_MAI_SMP, 0x002c),
+	VC4_HD_REG(HDMI_VID_CTL, 0x0038),
+	VC4_HD_REG(HDMI_CSC_CTL, 0x0040),
+	VC4_HD_REG(HDMI_CSC_12_11, 0x0044),
+	VC4_HD_REG(HDMI_CSC_14_13, 0x0048),
+	VC4_HD_REG(HDMI_CSC_22_21, 0x004c),
+	VC4_HD_REG(HDMI_CSC_24_23, 0x0050),
+	VC4_HD_REG(HDMI_CSC_32_31, 0x0054),
+	VC4_HD_REG(HDMI_CSC_34_33, 0x0058),
+	VC4_HD_REG(HDMI_FRAME_COUNT, 0x0068),
+
+	VC4_HDMI_REG(HDMI_CORE_REV, 0x0000),
+	VC4_HDMI_REG(HDMI_SW_RESET_CONTROL, 0x0004),
+	VC4_HDMI_REG(HDMI_HOTPLUG_INT, 0x0008),
+	VC4_HDMI_REG(HDMI_HOTPLUG, 0x000c),
+	VC4_HDMI_REG(HDMI_FIFO_CTL, 0x005c),
+	VC4_HDMI_REG(HDMI_MAI_CHANNEL_MAP, 0x0090),
+	VC4_HDMI_REG(HDMI_MAI_CONFIG, 0x0094),
+	VC4_HDMI_REG(HDMI_MAI_FORMAT, 0x0098),
+	VC4_HDMI_REG(HDMI_AUDIO_PACKET_CONFIG, 0x009c),
+	VC4_HDMI_REG(HDMI_RAM_PACKET_CONFIG, 0x00a0),
+	VC4_HDMI_REG(HDMI_RAM_PACKET_STATUS, 0x00a4),
+	VC4_HDMI_REG(HDMI_CRP_CFG, 0x00a8),
+	VC4_HDMI_REG(HDMI_CTS_0, 0x00ac),
+	VC4_HDMI_REG(HDMI_CTS_1, 0x00b0),
+	VC4_HDMI_REG(HDMI_SCHEDULER_CONTROL, 0x00c0),
+	VC4_HDMI_REG(HDMI_HORZA, 0x00c4),
+	VC4_HDMI_REG(HDMI_HORZB, 0x00c8),
+	VC4_HDMI_REG(HDMI_VERTA0, 0x00cc),
+	VC4_HDMI_REG(HDMI_VERTB0, 0x00d0),
+	VC4_HDMI_REG(HDMI_VERTA1, 0x00d4),
+	VC4_HDMI_REG(HDMI_VERTB1, 0x00d8),
+	VC4_HDMI_REG(HDMI_CEC_CNTRL_1, 0x00e8),
+	VC4_HDMI_REG(HDMI_CEC_CNTRL_2, 0x00ec),
+	VC4_HDMI_REG(HDMI_CEC_CNTRL_3, 0x00f0),
+	VC4_HDMI_REG(HDMI_CEC_CNTRL_4, 0x00f4),
+	VC4_HDMI_REG(HDMI_CEC_CNTRL_5, 0x00f8),
+	VC4_HDMI_REG(HDMI_CEC_TX_DATA_1, 0x00fc),
+	VC4_HDMI_REG(HDMI_CEC_TX_DATA_2, 0x0100),
+	VC4_HDMI_REG(HDMI_CEC_TX_DATA_3, 0x0104),
+	VC4_HDMI_REG(HDMI_CEC_TX_DATA_4, 0x0108),
+	VC4_HDMI_REG(HDMI_CEC_RX_DATA_1, 0x010c),
+	VC4_HDMI_REG(HDMI_CEC_RX_DATA_2, 0x0110),
+	VC4_HDMI_REG(HDMI_CEC_RX_DATA_3, 0x0114),
+	VC4_HDMI_REG(HDMI_CEC_RX_DATA_4, 0x0118),
+	VC4_HDMI_REG(HDMI_TX_PHY_RESET_CTL, 0x02c0),
+	VC4_HDMI_REG(HDMI_TX_PHY_CTL_0, 0x02c4),
+	VC4_HDMI_REG(HDMI_CEC_CPU_STATUS, 0x0340),
+	VC4_HDMI_REG(HDMI_CEC_CPU_CLEAR, 0x0348),
+	VC4_HDMI_REG(HDMI_CEC_CPU_MASK_STATUS, 0x034c),
+	VC4_HDMI_REG(HDMI_CEC_CPU_MASK_SET, 0x034c),
+	VC4_HDMI_REG(HDMI_CEC_CPU_MASK_CLEAR, 0x0354),
+	VC4_HDMI_REG(HDMI_RAM_PACKET_START, 0x0400),
+};
+
+static const struct vc4_hdmi_register vc5_hdmi_hdmi0_fields[] = {
+	VC4_HD_REG(HDMI_DVP_CTL, 0x0000),
+	VC4_HD_REG(HDMI_MAI_CTL, 0x0010),
+	VC4_HD_REG(HDMI_MAI_THR, 0x0014),
+	VC4_HD_REG(HDMI_MAI_FMT, 0x0018),
+	VC4_HD_REG(HDMI_MAI_DATA, 0x001c),
+	VC4_HD_REG(HDMI_MAI_SMP, 0x0020),
+	VC4_HD_REG(HDMI_VID_CTL, 0x0044),
+	VC4_HD_REG(HDMI_FRAME_COUNT, 0x0060),
+
+	VC4_HDMI_REG(HDMI_FIFO_CTL, 0x074),
+	VC4_HDMI_REG(HDMI_AUDIO_PACKET_CONFIG, 0x0b8),
+	VC4_HDMI_REG(HDMI_RAM_PACKET_CONFIG, 0x0bc),
+	VC4_HDMI_REG(HDMI_RAM_PACKET_STATUS, 0x0c4),
+	VC4_HDMI_REG(HDMI_CRP_CFG, 0x0c8),
+	VC4_HDMI_REG(HDMI_CTS_0, 0x0cc),
+	VC4_HDMI_REG(HDMI_CTS_1, 0x0d0),
+	VC4_HDMI_REG(HDMI_SCHEDULER_CONTROL, 0x0e0),
+	VC4_HDMI_REG(HDMI_HORZA, 0x0e4),
+	VC4_HDMI_REG(HDMI_HORZB, 0x0e8),
+	VC4_HDMI_REG(HDMI_VERTA0, 0x0ec),
+	VC4_HDMI_REG(HDMI_VERTB0, 0x0f0),
+	VC4_HDMI_REG(HDMI_VERTA1, 0x0f4),
+	VC4_HDMI_REG(HDMI_VERTB1, 0x0f8),
+	VC4_HDMI_REG(HDMI_MAI_CHANNEL_MAP, 0x09c),
+	VC4_HDMI_REG(HDMI_MAI_CONFIG, 0x0a0),
+	VC4_HDMI_REG(HDMI_HOTPLUG, 0x1a8),
+
+	VC5_DVP_REG(HDMI_CLOCK_STOP, 0x0bc),
+	VC5_DVP_REG(HDMI_VEC_INTERFACE_XBAR, 0x0f0),
+
+	VC5_PHY_REG(HDMI_TX_PHY_RESET_CTL, 0x000),
+	VC5_PHY_REG(HDMI_TX_PHY_POWERDOWN_CTL, 0x004),
+	VC5_PHY_REG(HDMI_TX_PHY_CTL_0, 0x008),
+	VC5_PHY_REG(HDMI_TX_PHY_CTL_1, 0x00c),
+	VC5_PHY_REG(HDMI_TX_PHY_CTL_2, 0x010),
+	VC5_PHY_REG(HDMI_TX_PHY_CTL_3, 0x014),
+	VC5_PHY_REG(HDMI_TX_PHY_PLL_CTL_0, 0x01c),
+	VC5_PHY_REG(HDMI_TX_PHY_PLL_CTL_1, 0x020),
+	VC5_PHY_REG(HDMI_TX_PHY_CLK_DIV, 0x028),
+	VC5_PHY_REG(HDMI_TX_PHY_PLL_CFG, 0x034),
+	VC5_PHY_REG(HDMI_TX_PHY_TMDS_CLK_WORD_SEL, 0x044),
+	VC5_PHY_REG(HDMI_TX_PHY_CHANNEL_SWAP, 0x04c),
+	VC5_PHY_REG(HDMI_TX_PHY_PLL_CALIBRATION_CONFIG_1, 0x050),
+	VC5_PHY_REG(HDMI_TX_PHY_PLL_CALIBRATION_CONFIG_2, 0x054),
+	VC5_PHY_REG(HDMI_TX_PHY_PLL_CALIBRATION_CONFIG_4, 0x05c),
+
+	VC5_RM_REG(HDMI_RM_CONTROL, 0x000),
+	VC5_RM_REG(HDMI_RM_OFFSET, 0x018),
+	VC5_RM_REG(HDMI_RM_FORMAT, 0x01c),
+
+	VC5_RAM_REG(HDMI_RAM_PACKET_START, 0x000),
+
+	VC5_CEC_REG(HDMI_CEC_CNTRL_1, 0x010),
+	VC5_CEC_REG(HDMI_CEC_CNTRL_2, 0x014),
+	VC5_CEC_REG(HDMI_CEC_CNTRL_3, 0x018),
+	VC5_CEC_REG(HDMI_CEC_CNTRL_4, 0x01c),
+	VC5_CEC_REG(HDMI_CEC_CNTRL_5, 0x020),
+	VC5_CEC_REG(HDMI_CEC_TX_DATA_1, 0x028),
+	VC5_CEC_REG(HDMI_CEC_TX_DATA_2, 0x02c),
+	VC5_CEC_REG(HDMI_CEC_TX_DATA_3, 0x030),
+	VC5_CEC_REG(HDMI_CEC_TX_DATA_4, 0x034),
+	VC5_CEC_REG(HDMI_CEC_RX_DATA_1, 0x038),
+	VC5_CEC_REG(HDMI_CEC_RX_DATA_2, 0x03c),
+	VC5_CEC_REG(HDMI_CEC_RX_DATA_3, 0x040),
+	VC5_CEC_REG(HDMI_CEC_RX_DATA_4, 0x044),
+
+	VC5_CSC_REG(HDMI_CSC_CTL, 0x000),
+	VC5_CSC_REG(HDMI_CSC_12_11, 0x004),
+	VC5_CSC_REG(HDMI_CSC_14_13, 0x008),
+	VC5_CSC_REG(HDMI_CSC_22_21, 0x00c),
+	VC5_CSC_REG(HDMI_CSC_24_23, 0x010),
+	VC5_CSC_REG(HDMI_CSC_32_31, 0x014),
+	VC5_CSC_REG(HDMI_CSC_34_33, 0x018),
+};
+
+static const struct vc4_hdmi_register vc5_hdmi_hdmi1_fields[] = {
+	VC4_HD_REG(HDMI_DVP_CTL, 0x0000),
+	VC4_HD_REG(HDMI_MAI_CTL, 0x0030),
+	VC4_HD_REG(HDMI_MAI_THR, 0x0034),
+	VC4_HD_REG(HDMI_MAI_FMT, 0x0038),
+	VC4_HD_REG(HDMI_MAI_DATA, 0x003c),
+	VC4_HD_REG(HDMI_MAI_SMP, 0x0040),
+	VC4_HD_REG(HDMI_VID_CTL, 0x0048),
+	VC4_HD_REG(HDMI_FRAME_COUNT, 0x0064),
+
+	VC4_HDMI_REG(HDMI_FIFO_CTL, 0x074),
+	VC4_HDMI_REG(HDMI_AUDIO_PACKET_CONFIG, 0x0b8),
+	VC4_HDMI_REG(HDMI_RAM_PACKET_CONFIG, 0x0bc),
+	VC4_HDMI_REG(HDMI_RAM_PACKET_STATUS, 0x0c4),
+	VC4_HDMI_REG(HDMI_CRP_CFG, 0x0c8),
+	VC4_HDMI_REG(HDMI_CTS_0, 0x0cc),
+	VC4_HDMI_REG(HDMI_CTS_1, 0x0d0),
+	VC4_HDMI_REG(HDMI_SCHEDULER_CONTROL, 0x0e0),
+	VC4_HDMI_REG(HDMI_HORZA, 0x0e4),
+	VC4_HDMI_REG(HDMI_HORZB, 0x0e8),
+	VC4_HDMI_REG(HDMI_VERTA0, 0x0ec),
+	VC4_HDMI_REG(HDMI_VERTB0, 0x0f0),
+	VC4_HDMI_REG(HDMI_VERTA1, 0x0f4),
+	VC4_HDMI_REG(HDMI_VERTB1, 0x0f8),
+	VC4_HDMI_REG(HDMI_MAI_CHANNEL_MAP, 0x09c),
+	VC4_HDMI_REG(HDMI_MAI_CONFIG, 0x0a0),
+	VC4_HDMI_REG(HDMI_HOTPLUG, 0x1a8),
+
+	VC5_DVP_REG(HDMI_CLOCK_STOP, 0x0bc),
+	VC5_DVP_REG(HDMI_VEC_INTERFACE_XBAR, 0x0f0),
+
+	VC5_PHY_REG(HDMI_TX_PHY_RESET_CTL, 0x000),
+	VC5_PHY_REG(HDMI_TX_PHY_POWERDOWN_CTL, 0x004),
+	VC5_PHY_REG(HDMI_TX_PHY_CTL_0, 0x008),
+	VC5_PHY_REG(HDMI_TX_PHY_CTL_1, 0x00c),
+	VC5_PHY_REG(HDMI_TX_PHY_CTL_2, 0x010),
+	VC5_PHY_REG(HDMI_TX_PHY_CTL_3, 0x014),
+	VC5_PHY_REG(HDMI_TX_PHY_PLL_CTL_0, 0x01c),
+	VC5_PHY_REG(HDMI_TX_PHY_PLL_CTL_1, 0x020),
+	VC5_PHY_REG(HDMI_TX_PHY_CLK_DIV, 0x028),
+	VC5_PHY_REG(HDMI_TX_PHY_PLL_CFG, 0x034),
+	VC5_PHY_REG(HDMI_TX_PHY_CHANNEL_SWAP, 0x04c),
+	VC5_PHY_REG(HDMI_TX_PHY_TMDS_CLK_WORD_SEL, 0x044),
+	VC5_PHY_REG(HDMI_TX_PHY_PLL_CALIBRATION_CONFIG_1, 0x050),
+	VC5_PHY_REG(HDMI_TX_PHY_PLL_CALIBRATION_CONFIG_2, 0x054),
+	VC5_PHY_REG(HDMI_TX_PHY_PLL_CALIBRATION_CONFIG_4, 0x05c),
+
+	VC5_RM_REG(HDMI_RM_CONTROL, 0x000),
+	VC5_RM_REG(HDMI_RM_OFFSET, 0x018),
+	VC5_RM_REG(HDMI_RM_FORMAT, 0x01c),
+
+	VC5_RAM_REG(HDMI_RAM_PACKET_START, 0x000),
+
+	VC5_CEC_REG(HDMI_CEC_CNTRL_1, 0x010),
+	VC5_CEC_REG(HDMI_CEC_CNTRL_2, 0x014),
+	VC5_CEC_REG(HDMI_CEC_CNTRL_3, 0x018),
+	VC5_CEC_REG(HDMI_CEC_CNTRL_4, 0x01c),
+	VC5_CEC_REG(HDMI_CEC_CNTRL_5, 0x020),
+	VC5_CEC_REG(HDMI_CEC_TX_DATA_1, 0x028),
+	VC5_CEC_REG(HDMI_CEC_TX_DATA_2, 0x02c),
+	VC5_CEC_REG(HDMI_CEC_TX_DATA_3, 0x030),
+	VC5_CEC_REG(HDMI_CEC_TX_DATA_4, 0x034),
+	VC5_CEC_REG(HDMI_CEC_RX_DATA_1, 0x038),
+	VC5_CEC_REG(HDMI_CEC_RX_DATA_2, 0x03c),
+	VC5_CEC_REG(HDMI_CEC_RX_DATA_3, 0x040),
+	VC5_CEC_REG(HDMI_CEC_RX_DATA_4, 0x044),
+
+	VC5_CSC_REG(HDMI_CSC_CTL, 0x000),
+	VC5_CSC_REG(HDMI_CSC_12_11, 0x004),
+	VC5_CSC_REG(HDMI_CSC_14_13, 0x008),
+	VC5_CSC_REG(HDMI_CSC_22_21, 0x00c),
+	VC5_CSC_REG(HDMI_CSC_24_23, 0x010),
+	VC5_CSC_REG(HDMI_CSC_32_31, 0x014),
+	VC5_CSC_REG(HDMI_CSC_34_33, 0x018),
+};
+
+static inline
+void __iomem *__vc4_hdmi_get_field_base(struct vc4_hdmi *hdmi,
+					enum vc4_hdmi_regs reg)
+{
+	switch (reg) {
+	case VC4_HD:
+		return hdmi->hd_regs;
+
+	case VC4_HDMI:
+		return hdmi->hdmicore_regs;
+
+	case VC5_CSC:
+		return hdmi->csc_regs;
+
+	case VC5_CEC:
+		return hdmi->cec_regs;
+
+	case VC5_DVP:
+		return hdmi->dvp_regs;
+
+	case VC5_PHY:
+		return hdmi->phy_regs;
+
+	case VC5_RAM:
+		return hdmi->ram_regs;
+
+	case VC5_RM:
+		return hdmi->rm_regs;
+
+	default:
+		return NULL;
+	}
+
+	return NULL;
+}
+
+static inline u32 vc4_hdmi_read(struct vc4_hdmi *hdmi,
+				enum vc4_hdmi_field reg)
+{
+	const struct vc4_hdmi_register *field;
+	const struct vc4_hdmi_variant *variant = hdmi->variant;
+	void __iomem *base;
+
+	if (reg >= variant->num_registers) {
+		dev_warn(&hdmi->pdev->dev,
+			 "Invalid register ID %u\n", reg);
+		return 0;
+	}
+
+	field = &variant->registers[reg];
+	base = __vc4_hdmi_get_field_base(hdmi, field->reg);
+	if (!base) {
+		dev_warn(&hdmi->pdev->dev,
+			 "Unknown register ID %u\n", reg);
+		return 0;
+	}
+
+	return readl(base + field->offset);
+}
+#define HDMI_READ(reg)		vc4_hdmi_read(vc4_hdmi, reg)
+
+static inline void vc4_hdmi_write(struct vc4_hdmi *hdmi,
+				  enum vc4_hdmi_field reg,
+				  u32 value)
+{
+	const struct vc4_hdmi_register *field;
+	const struct vc4_hdmi_variant *variant = hdmi->variant;
+	void __iomem *base;
+
+	if (reg >= variant->num_registers) {
+		dev_warn(&hdmi->pdev->dev,
+			 "Invalid register ID %u\n", reg);
+		return;
+	}
+
+	field = &variant->registers[reg];
+	base = __vc4_hdmi_get_field_base(hdmi, field->reg);
+	if (!base)
+		return;
+
+	writel(value, base + field->offset);
+}
+#define HDMI_WRITE(reg, val)	vc4_hdmi_write(vc4_hdmi, reg, val)
+
+#endif /* _VC4_HDMI_REGS_H_ */
diff --git a/drivers/gpu/drm/vc4/vc4_hvs.c b/drivers/gpu/drm/vc4/vc4_hvs.c
index 2d2bf59c0503..4d0a833366ce 100644
--- a/drivers/gpu/drm/vc4/vc4_hvs.c
+++ b/drivers/gpu/drm/vc4/vc4_hvs.c
@@ -19,6 +19,8 @@
  * each CRTC.
  */
 
+#include <linux/bitfield.h>
+#include <linux/clk.h>
 #include <linux/component.h>
 #include <linux/platform_device.h>
 
@@ -160,6 +162,7 @@ static void vc4_hvs_lut_load(struct drm_crtc *crtc)
 	struct drm_device *dev = crtc->dev;
 	struct vc4_dev *vc4 = to_vc4_dev(dev);
 	struct vc4_crtc *vc4_crtc = to_vc4_crtc(crtc);
+	struct vc4_crtc_state *vc4_state = to_vc4_crtc_state(crtc->state);
 	u32 i;
 
 	/* The LUT memory is laid out with each HVS channel in order,
@@ -168,7 +171,7 @@ static void vc4_hvs_lut_load(struct drm_crtc *crtc)
 	 */
 	HVS_WRITE(SCALER_GAMADDR,
 		  SCALER_GAMADDR_AUTOINC |
-		  (vc4_crtc->channel * 3 * crtc->gamma_size));
+		  (vc4_state->assigned_channel * 3 * crtc->gamma_size));
 
 	for (i = 0; i < crtc->gamma_size; i++)
 		HVS_WRITE(SCALER_GAMDATA, vc4_crtc->lut_r[i]);
@@ -194,6 +197,135 @@ static void vc4_hvs_update_gamma_lut(struct drm_crtc *crtc)
 	vc4_hvs_lut_load(crtc);
 }
 
+int vc4_hvs_get_fifo_from_output(struct drm_device *dev, unsigned int output)
+{
+	struct vc4_dev *vc4 = to_vc4_dev(dev);
+	u32 reg;
+	int ret;
+
+	if (!vc4->hvs->hvs5)
+		return output;
+
+	switch (output) {
+	case 0:
+		return 0;
+
+	case 1:
+		return 1;
+
+	case 2:
+		reg = HVS_READ(SCALER_DISPECTRL);
+		ret = FIELD_GET(SCALER_DISPECTRL_DSP2_MUX_MASK, reg);
+		if (ret == 0)
+			return 2;
+
+		return 0;
+
+	case 3:
+		reg = HVS_READ(SCALER_DISPCTRL);
+		ret = FIELD_GET(SCALER_DISPCTRL_DSP3_MUX_MASK, reg);
+		if (ret == 3)
+			return -EPIPE;
+
+		return ret;
+
+	case 4:
+		reg = HVS_READ(SCALER_DISPEOLN);
+		ret = FIELD_GET(SCALER_DISPEOLN_DSP4_MUX_MASK, reg);
+		if (ret == 3)
+			return -EPIPE;
+
+		return ret;
+
+	case 5:
+		reg = HVS_READ(SCALER_DISPDITHER);
+		ret = FIELD_GET(SCALER_DISPDITHER_DSP5_MUX_MASK, reg);
+		if (ret == 3)
+			return -EPIPE;
+
+		return ret;
+
+	default:
+		return -EPIPE;
+	}
+}
+
+static int vc4_hvs_init_channel(struct vc4_dev *vc4, struct drm_crtc *crtc,
+				struct drm_display_mode *mode, bool oneshot)
+{
+	struct vc4_crtc_state *vc4_crtc_state = to_vc4_crtc_state(crtc->state);
+	unsigned int chan = vc4_crtc_state->assigned_channel;
+	bool interlace = mode->flags & DRM_MODE_FLAG_INTERLACE;
+	u32 dispbkgndx;
+	u32 dispctrl;
+
+	HVS_WRITE(SCALER_DISPCTRLX(chan), 0);
+	HVS_WRITE(SCALER_DISPCTRLX(chan), SCALER_DISPCTRLX_RESET);
+	HVS_WRITE(SCALER_DISPCTRLX(chan), 0);
+
+	/* Turn on the scaler, which will wait for vstart to start
+	 * compositing.
+	 * When feeding the transposer, we should operate in oneshot
+	 * mode.
+	 */
+	dispctrl = SCALER_DISPCTRLX_ENABLE;
+
+	if (!vc4->hvs->hvs5)
+		dispctrl |= VC4_SET_FIELD(mode->hdisplay,
+					  SCALER_DISPCTRLX_WIDTH) |
+			    VC4_SET_FIELD(mode->vdisplay,
+					  SCALER_DISPCTRLX_HEIGHT) |
+			    (oneshot ? SCALER_DISPCTRLX_ONESHOT : 0);
+	else
+		dispctrl |= VC4_SET_FIELD(mode->hdisplay,
+					  SCALER5_DISPCTRLX_WIDTH) |
+			    VC4_SET_FIELD(mode->vdisplay,
+					  SCALER5_DISPCTRLX_HEIGHT) |
+			    (oneshot ? SCALER5_DISPCTRLX_ONESHOT : 0);
+
+	HVS_WRITE(SCALER_DISPCTRLX(chan), dispctrl);
+
+	dispbkgndx = HVS_READ(SCALER_DISPBKGNDX(chan));
+	dispbkgndx &= ~SCALER_DISPBKGND_GAMMA;
+	dispbkgndx &= ~SCALER_DISPBKGND_INTERLACE;
+
+	HVS_WRITE(SCALER_DISPBKGNDX(chan), dispbkgndx |
+		  SCALER_DISPBKGND_AUTOHS |
+		  ((!vc4->hvs->hvs5) ? SCALER_DISPBKGND_GAMMA : 0) |
+		  (interlace ? SCALER_DISPBKGND_INTERLACE : 0));
+
+	/* Reload the LUT, since the SRAMs would have been disabled if
+	 * all CRTCs had SCALER_DISPBKGND_GAMMA unset at once.
+	 */
+	vc4_hvs_lut_load(crtc);
+
+	return 0;
+}
+
+void vc4_hvs_stop_channel(struct drm_device *dev, unsigned int chan)
+{
+	struct vc4_dev *vc4 = to_vc4_dev(dev);
+
+	if (HVS_READ(SCALER_DISPCTRLX(chan)) & SCALER_DISPCTRLX_ENABLE)
+		return;
+
+	HVS_WRITE(SCALER_DISPCTRLX(chan),
+		  HVS_READ(SCALER_DISPCTRLX(chan)) | SCALER_DISPCTRLX_RESET);
+	HVS_WRITE(SCALER_DISPCTRLX(chan),
+		  HVS_READ(SCALER_DISPCTRLX(chan)) & ~SCALER_DISPCTRLX_ENABLE);
+
+	/* Once we leave, the scaler should be disabled and its fifo empty. */
+	WARN_ON_ONCE(HVS_READ(SCALER_DISPCTRLX(chan)) & SCALER_DISPCTRLX_RESET);
+
+	WARN_ON_ONCE(VC4_GET_FIELD(HVS_READ(SCALER_DISPSTATX(chan)),
+				   SCALER_DISPSTATX_MODE) !=
+		     SCALER_DISPSTATX_MODE_DISABLED);
+
+	WARN_ON_ONCE((HVS_READ(SCALER_DISPSTATX(chan)) &
+		      (SCALER_DISPSTATX_FULL | SCALER_DISPSTATX_EMPTY)) !=
+		     SCALER_DISPSTATX_EMPTY);
+}
+
 int vc4_hvs_atomic_check(struct drm_crtc *crtc,
 			 struct drm_crtc_state *state)
 {
@@ -248,12 +380,12 @@ static void vc4_hvs_update_dlist(struct drm_crtc *crtc)
 			crtc->state->event = NULL;
 		}
 
-		HVS_WRITE(SCALER_DISPLISTX(vc4_crtc->channel),
+		HVS_WRITE(SCALER_DISPLISTX(vc4_state->assigned_channel),
 			  vc4_state->mm.start);
 
 		spin_unlock_irqrestore(&dev->event_lock, flags);
 	} else {
-		HVS_WRITE(SCALER_DISPLISTX(vc4_crtc->channel),
+		HVS_WRITE(SCALER_DISPLISTX(vc4_state->assigned_channel),
 			  vc4_state->mm.start);
 	}
 }
@@ -263,59 +395,22 @@ void vc4_hvs_atomic_enable(struct drm_crtc *crtc,
 {
 	struct drm_device *dev = crtc->dev;
 	struct vc4_dev *vc4 = to_vc4_dev(dev);
-	struct vc4_crtc *vc4_crtc = to_vc4_crtc(crtc);
 	struct vc4_crtc_state *vc4_state = to_vc4_crtc_state(crtc->state);
 	struct drm_display_mode *mode = &crtc->state->adjusted_mode;
 	bool oneshot = vc4_state->feed_txp;
-	u32 dispctrl;
 
 	vc4_hvs_update_dlist(crtc);
-
-	/* Turn on the scaler, which will wait for vstart to start
-	 * compositing.
-	 * When feeding the transposer, we should operate in oneshot
-	 * mode.
-	 */
-	dispctrl = SCALER_DISPCTRLX_ENABLE;
-	dispctrl |= VC4_SET_FIELD(mode->hdisplay,
-				  SCALER_DISPCTRLX_WIDTH) |
-		    VC4_SET_FIELD(mode->vdisplay,
-				  SCALER_DISPCTRLX_HEIGHT) |
-		    (oneshot ? SCALER_DISPCTRLX_ONESHOT : 0);
-
-	HVS_WRITE(SCALER_DISPCTRLX(vc4_crtc->channel), dispctrl);
+	vc4_hvs_init_channel(vc4, crtc, mode, oneshot);
 }
 
 void vc4_hvs_atomic_disable(struct drm_crtc *crtc,
 			    struct drm_crtc_state *old_state)
 {
 	struct drm_device *dev = crtc->dev;
-	struct vc4_dev *vc4 = to_vc4_dev(dev);
-	struct vc4_crtc *vc4_crtc = to_vc4_crtc(crtc);
-	u32 chan = vc4_crtc->channel;
+	struct vc4_crtc_state *vc4_state = to_vc4_crtc_state(old_state);
+	unsigned int chan = vc4_state->assigned_channel;
 
-	if (HVS_READ(SCALER_DISPCTRLX(chan)) &
-	    SCALER_DISPCTRLX_ENABLE) {
-		HVS_WRITE(SCALER_DISPCTRLX(chan),
-			  SCALER_DISPCTRLX_RESET);
-
-		/* While the docs say that reset is self-clearing, it
-		 * seems it doesn't actually.
-		 */
-		HVS_WRITE(SCALER_DISPCTRLX(chan), 0);
-	}
-
-	/* Once we leave, the scaler should be disabled and its fifo empty. */
-
-	WARN_ON_ONCE(HVS_READ(SCALER_DISPCTRLX(chan)) & SCALER_DISPCTRLX_RESET);
-
-	WARN_ON_ONCE(VC4_GET_FIELD(HVS_READ(SCALER_DISPSTATX(chan)),
-				   SCALER_DISPSTATX_MODE) !=
-		     SCALER_DISPSTATX_MODE_DISABLED);
-
-	WARN_ON_ONCE((HVS_READ(SCALER_DISPSTATX(chan)) &
-		      (SCALER_DISPSTATX_FULL | SCALER_DISPSTATX_EMPTY)) !=
-		     SCALER_DISPSTATX_EMPTY);
+	vc4_hvs_stop_channel(dev, chan);
 }
 
 void vc4_hvs_atomic_flush(struct drm_crtc *crtc,
@@ -323,7 +418,6 @@ void vc4_hvs_atomic_flush(struct drm_crtc *crtc,
 {
 	struct drm_device *dev = crtc->dev;
 	struct vc4_dev *vc4 = to_vc4_dev(dev);
-	struct vc4_crtc *vc4_crtc = to_vc4_crtc(crtc);
 	struct vc4_crtc_state *vc4_state = to_vc4_crtc_state(crtc->state);
 	struct drm_plane *plane;
 	struct vc4_plane_state *vc4_plane_state;
@@ -365,8 +459,8 @@ void vc4_hvs_atomic_flush(struct drm_crtc *crtc,
 		/* This sets a black background color fill, as is the case
 		 * with other DRM drivers.
 		 */
-		HVS_WRITE(SCALER_DISPBKGNDX(vc4_crtc->channel),
-			  HVS_READ(SCALER_DISPBKGNDX(vc4_crtc->channel)) |
+		HVS_WRITE(SCALER_DISPBKGNDX(vc4_state->assigned_channel),
+			  HVS_READ(SCALER_DISPBKGNDX(vc4_state->assigned_channel)) |
 			  SCALER_DISPBKGND_FILL);
 
 	/* Only update DISPLIST if the CRTC was already running and is not
@@ -380,7 +474,7 @@ void vc4_hvs_atomic_flush(struct drm_crtc *crtc,
 		vc4_hvs_update_dlist(crtc);
 
 	if (crtc->state->color_mgmt_changed) {
-		u32 dispbkgndx = HVS_READ(SCALER_DISPBKGNDX(vc4_crtc->channel));
+		u32 dispbkgndx = HVS_READ(SCALER_DISPBKGNDX(vc4_state->assigned_channel));
 
 		if (crtc->state->gamma_lut) {
 			vc4_hvs_update_gamma_lut(crtc);
@@ -392,7 +486,7 @@ void vc4_hvs_atomic_flush(struct drm_crtc *crtc,
 			 */
 			dispbkgndx &= ~SCALER_DISPBKGND_GAMMA;
 		}
-		HVS_WRITE(SCALER_DISPBKGNDX(vc4_crtc->channel), dispbkgndx);
+		HVS_WRITE(SCALER_DISPBKGNDX(vc4_state->assigned_channel), dispbkgndx);
 	}
 
 	if (debug_dump_regs) {
@@ -401,50 +495,6 @@ void vc4_hvs_atomic_flush(struct drm_crtc *crtc,
 	}
 }
 
-void vc4_hvs_mode_set_nofb(struct drm_crtc *crtc)
-{
-	struct drm_device *dev = crtc->dev;
-	struct vc4_dev *vc4 = to_vc4_dev(dev);
-	struct vc4_crtc *vc4_crtc = to_vc4_crtc(crtc);
-	struct vc4_crtc_state *vc4_state = to_vc4_crtc_state(crtc->state);
-	struct drm_display_mode *mode = &crtc->state->adjusted_mode;
-	bool interlace = mode->flags & DRM_MODE_FLAG_INTERLACE;
-
-	if (vc4_crtc->data->hvs_channel == 2) {
-		u32 dispctrl;
-		u32 dsp3_mux;
-
-		/*
-		 * SCALER_DISPCTRL_DSP3 = X, where X < 2 means 'connect DSP3 to
-		 * FIFO X'.
-		 * SCALER_DISPCTRL_DSP3 = 3 means 'disable DSP 3'.
-		 *
-		 * DSP3 is connected to FIFO2 unless the transposer is
-		 * enabled. In this case, FIFO 2 is directly accessed by the
-		 * TXP IP, and we need to disable the FIFO2 -> pixelvalve1
-		 * route.
-		 */
-		if (vc4_state->feed_txp)
-			dsp3_mux = VC4_SET_FIELD(3, SCALER_DISPCTRL_DSP3_MUX);
-		else
-			dsp3_mux = VC4_SET_FIELD(2, SCALER_DISPCTRL_DSP3_MUX);
-
-		dispctrl = HVS_READ(SCALER_DISPCTRL) &
-			   ~SCALER_DISPCTRL_DSP3_MUX_MASK;
-		HVS_WRITE(SCALER_DISPCTRL, dispctrl | dsp3_mux);
-	}
-
-	HVS_WRITE(SCALER_DISPBKGNDX(vc4_crtc->channel),
-		  SCALER_DISPBKGND_AUTOHS |
-		  SCALER_DISPBKGND_GAMMA |
-		  (interlace ? SCALER_DISPBKGND_INTERLACE : 0));
-
-	/* Reload the LUT, since the SRAMs would have been disabled if
-	 * all CRTCs had SCALER_DISPBKGND_GAMMA unset at once.
-	 */
-	vc4_hvs_lut_load(crtc);
-}
-
 void vc4_hvs_mask_underrun(struct drm_device *dev, int channel)
 {
 	struct vc4_dev *vc4 = to_vc4_dev(dev);
@@ -521,6 +571,9 @@ static int vc4_hvs_bind(struct device *dev, struct device *master, void *data)
 
 	hvs->pdev = pdev;
 
+	if (of_device_is_compatible(pdev->dev.of_node, "brcm,bcm2711-hvs"))
+		hvs->hvs5 = true;
+
 	hvs->regs = vc4_ioremap_regs(pdev, 0);
 	if (IS_ERR(hvs->regs))
 		return PTR_ERR(hvs->regs);
@@ -529,7 +582,24 @@ static int vc4_hvs_bind(struct device *dev, struct device *master, void *data)
 	hvs->regset.regs = hvs_regs;
 	hvs->regset.nregs = ARRAY_SIZE(hvs_regs);
 
-	hvs->dlist = hvs->regs + SCALER_DLIST_START;
+	if (hvs->hvs5) {
+		hvs->core_clk = devm_clk_get(&pdev->dev, NULL);
+		if (IS_ERR(hvs->core_clk)) {
+			dev_err(&pdev->dev, "Couldn't get core clock\n");
+			return PTR_ERR(hvs->core_clk);
+		}
+
+		ret = clk_prepare_enable(hvs->core_clk);
+		if (ret) {
+			dev_err(&pdev->dev, "Couldn't enable the core clock\n");
+			return ret;
+		}
+	}
+
+	if (!hvs->hvs5)
+		hvs->dlist = hvs->regs + SCALER_DLIST_START;
+	else
+		hvs->dlist = hvs->regs + SCALER5_DLIST_START;
 
 	spin_lock_init(&hvs->mm_lock);
 
@@ -547,7 +617,12 @@ static int vc4_hvs_bind(struct device *dev, struct device *master, void *data)
 	 * between planes when they don't overlap on the screen, but
 	 * for now we just allocate globally.
 	 */
-	drm_mm_init(&hvs->lbm_mm, 0, 96 * 1024);
+	if (!hvs->hvs5)
+		/* 96kB */
+		drm_mm_init(&hvs->lbm_mm, 0, 96 * 1024);
+	else
+		/* 70k words */
+		drm_mm_init(&hvs->lbm_mm, 0, 70 * 2 * 1024);
 
 	/* Upload filter kernels.  We only have the one for now, so we
 	 * keep it around for the lifetime of the driver.
@@ -605,6 +680,7 @@ static void vc4_hvs_unbind(struct device *dev, struct device *master,
 {
 	struct drm_device *drm = dev_get_drvdata(master);
 	struct vc4_dev *vc4 = drm->dev_private;
+	struct vc4_hvs *hvs = vc4->hvs;
 
 	if (drm_mm_node_allocated(&vc4->hvs->mitchell_netravali_filter))
 		drm_mm_remove_node(&vc4->hvs->mitchell_netravali_filter);
@@ -612,6 +688,8 @@ static void vc4_hvs_unbind(struct device *dev, struct device *master,
 	drm_mm_takedown(&vc4->hvs->dlist_mm);
 	drm_mm_takedown(&vc4->hvs->lbm_mm);
 
+	clk_disable_unprepare(hvs->core_clk);
+
 	vc4->hvs = NULL;
 }
 
@@ -632,6 +710,7 @@ static int vc4_hvs_dev_remove(struct platform_device *pdev)
 }
 
 static const struct of_device_id vc4_hvs_dt_match[] = {
+	{ .compatible = "brcm,bcm2711-hvs" },
 	{ .compatible = "brcm,bcm2835-hvs" },
 	{}
 };
diff --git a/drivers/gpu/drm/vc4/vc4_kms.c b/drivers/gpu/drm/vc4/vc4_kms.c
index 08318e69061b..af3ee3dcdab6 100644
--- a/drivers/gpu/drm/vc4/vc4_kms.c
+++ b/drivers/gpu/drm/vc4/vc4_kms.c
@@ -11,6 +11,8 @@
  * crtc, HDMI encoder).
  */
 
+#include <linux/clk.h>
+
 #include <drm/drm_atomic.h>
 #include <drm/drm_atomic_helper.h>
 #include <drm/drm_crtc.h>
@@ -144,22 +146,130 @@ vc4_ctm_commit(struct vc4_dev *vc4, struct drm_atomic_state *state)
 		  VC4_SET_FIELD(ctm_state->fifo, SCALER_OLEDOFFS_DISPFIFO));
 }
 
+static void vc4_hvs_pv_muxing_commit(struct vc4_dev *vc4,
+				     struct drm_atomic_state *state)
+{
+	struct drm_crtc_state *crtc_state;
+	struct drm_crtc *crtc;
+	unsigned int i;
+
+	for_each_new_crtc_in_state(state, crtc, crtc_state, i) {
+		struct vc4_crtc_state *vc4_state = to_vc4_crtc_state(crtc_state);
+		u32 dispctrl;
+		u32 dsp3_mux;
+
+		if (!crtc_state->active)
+			continue;
+
+		if (vc4_state->assigned_channel != 2)
+			continue;
+
+		/*
+		 * SCALER_DISPCTRL_DSP3 = X, where X < 2 means 'connect DSP3 to
+		 * FIFO X'.
+		 * SCALER_DISPCTRL_DSP3 = 3 means 'disable DSP 3'.
+		 *
+		 * DSP3 is connected to FIFO2 unless the transposer is
+		 * enabled. In this case, FIFO 2 is directly accessed by the
+		 * TXP IP, and we need to disable the FIFO2 -> pixelvalve1
+		 * route.
+		 */
+		if (vc4_state->feed_txp)
+			dsp3_mux = VC4_SET_FIELD(3, SCALER_DISPCTRL_DSP3_MUX);
+		else
+			dsp3_mux = VC4_SET_FIELD(2, SCALER_DISPCTRL_DSP3_MUX);
+
+		dispctrl = HVS_READ(SCALER_DISPCTRL) &
+			   ~SCALER_DISPCTRL_DSP3_MUX_MASK;
+		HVS_WRITE(SCALER_DISPCTRL, dispctrl | dsp3_mux);
+	}
+}
+
+static void vc5_hvs_pv_muxing_commit(struct vc4_dev *vc4,
+				     struct drm_atomic_state *state)
+{
+	struct drm_crtc_state *crtc_state;
+	struct drm_crtc *crtc;
+	unsigned char dsp2_mux = 0;
+	unsigned char dsp3_mux = 3;
+	unsigned char dsp4_mux = 3;
+	unsigned char dsp5_mux = 3;
+	unsigned int i;
+	u32 reg;
+
+	for_each_new_crtc_in_state(state, crtc, crtc_state, i) {
+		struct vc4_crtc_state *vc4_state = to_vc4_crtc_state(crtc_state);
+		struct vc4_crtc *vc4_crtc = to_vc4_crtc(crtc);
+
+		if (!crtc_state->active)
+			continue;
+
+		switch (vc4_crtc->data->hvs_output) {
+		case 2:
+			dsp2_mux = (vc4_state->assigned_channel == 2) ? 0 : 1;
+			break;
+
+		case 3:
+			dsp3_mux = vc4_state->assigned_channel;
+			break;
+
+		case 4:
+			dsp4_mux = vc4_state->assigned_channel;
+			break;
+
+		case 5:
+			dsp5_mux = vc4_state->assigned_channel;
+			break;
+
+		default:
+			break;
+		}
+	}
+
+	reg = HVS_READ(SCALER_DISPECTRL);
+	HVS_WRITE(SCALER_DISPECTRL,
+		  (reg & ~SCALER_DISPECTRL_DSP2_MUX_MASK) |
+		  VC4_SET_FIELD(dsp2_mux, SCALER_DISPECTRL_DSP2_MUX));
+
+	reg = HVS_READ(SCALER_DISPCTRL);
+	HVS_WRITE(SCALER_DISPCTRL,
+		  (reg & ~SCALER_DISPCTRL_DSP3_MUX_MASK) |
+		  VC4_SET_FIELD(dsp3_mux, SCALER_DISPCTRL_DSP3_MUX));
+
+	reg = HVS_READ(SCALER_DISPEOLN);
+	HVS_WRITE(SCALER_DISPEOLN,
+		  (reg & ~SCALER_DISPEOLN_DSP4_MUX_MASK) |
+		  VC4_SET_FIELD(dsp4_mux, SCALER_DISPEOLN_DSP4_MUX));
+
+	reg = HVS_READ(SCALER_DISPDITHER);
+	HVS_WRITE(SCALER_DISPDITHER,
+		  (reg & ~SCALER_DISPDITHER_DSP5_MUX_MASK) |
+		  VC4_SET_FIELD(dsp5_mux, SCALER_DISPDITHER_DSP5_MUX));
+}
+
 static void
 vc4_atomic_complete_commit(struct drm_atomic_state *state)
 {
 	struct drm_device *dev = state->dev;
 	struct vc4_dev *vc4 = to_vc4_dev(dev);
-	struct vc4_crtc *vc4_crtc;
+	struct vc4_hvs *hvs = vc4->hvs;
+	struct drm_crtc_state *new_crtc_state;
+	struct drm_crtc *crtc;
 	int i;
 
-	for (i = 0; i < dev->mode_config.num_crtc; i++) {
-		if (!state->crtcs[i].ptr || !state->crtcs[i].commit)
+	for_each_new_crtc_in_state(state, crtc, new_crtc_state, i) {
+		struct vc4_crtc_state *vc4_crtc_state;
+
+		if (!new_crtc_state->commit)
 			continue;
 
-		vc4_crtc = to_vc4_crtc(state->crtcs[i].ptr);
-		vc4_hvs_mask_underrun(dev, vc4_crtc->channel);
+		vc4_crtc_state = to_vc4_crtc_state(new_crtc_state);
+		vc4_hvs_mask_underrun(dev, vc4_crtc_state->assigned_channel);
 	}
 
+	if (vc4->hvs->hvs5)
+		clk_set_min_rate(hvs->core_clk, 500000000);
+
 	drm_atomic_helper_wait_for_fences(dev, state, false);
 
 	drm_atomic_helper_wait_for_dependencies(state);
@@ -168,6 +278,11 @@ vc4_atomic_complete_commit(struct drm_atomic_state *state)
 
 	vc4_ctm_commit(vc4, state);
 
+	if (vc4->hvs->hvs5)
+		vc5_hvs_pv_muxing_commit(vc4, state);
+	else
+		vc4_hvs_pv_muxing_commit(vc4, state);
+
 	drm_atomic_helper_commit_planes(dev, state, 0);
 
 	drm_atomic_helper_commit_modeset_enables(dev, state);
@@ -182,6 +297,9 @@ vc4_atomic_complete_commit(struct drm_atomic_state *state)
 
 	drm_atomic_helper_commit_cleanup_done(state);
 
+	if (vc4->hvs->hvs5)
+		clk_set_min_rate(hvs->core_clk, 0);
+
 	drm_atomic_state_put(state);
 
 	up(&vc4->async_modeset);
@@ -374,8 +492,11 @@ vc4_ctm_atomic_check(struct drm_device *dev, struct drm_atomic_state *state)
 
 		/* CTM is being enabled or the matrix changed. */
 		if (new_crtc_state->ctm) {
+			struct vc4_crtc_state *vc4_crtc_state =
+				to_vc4_crtc_state(new_crtc_state);
+
 			/* fifo is 1-based since 0 disables CTM. */
-			int fifo = to_vc4_crtc(crtc)->channel + 1;
+			int fifo = vc4_crtc_state->assigned_channel + 1;
 
 			/* Check userland isn't trying to turn on CTM for more
 			 * than one CRTC at a time.
@@ -415,6 +536,9 @@ static int vc4_load_tracker_atomic_check(struct drm_atomic_state *state)
 	struct drm_plane *plane;
 	int i;
 
+	if (!vc4->load_tracker_available)
+		return 0;
+
 	priv_state = drm_atomic_get_private_obj_state(state,
 						      &vc4->load_tracker);
 	if (IS_ERR(priv_state))
@@ -485,10 +609,77 @@ static const struct drm_private_state_funcs vc4_load_tracker_state_funcs = {
 	.atomic_destroy_state = vc4_load_tracker_destroy_state,
 };
 
+#define NUM_OUTPUTS  6
+#define NUM_CHANNELS 3
+
 static int
 vc4_atomic_check(struct drm_device *dev, struct drm_atomic_state *state)
 {
-	int ret;
+	unsigned long unassigned_channels = GENMASK(NUM_CHANNELS - 1, 0);
+	struct drm_crtc_state *crtc_state;
+	struct drm_crtc *crtc;
+	int i, ret;
+
+	/*
+	 * Since the HVS FIFOs are shared across all the pixelvalves and
+	 * the TXP (and thus all the CRTCs), we need to pull the current
+	 * state of all the enabled CRTCs so that an update to a single
+	 * CRTC still keeps the previous FIFOs enabled and assigned to
+	 * the same CRTCs, instead of evaluating only the CRTC being
+	 * modified.
+	 */
+	list_for_each_entry(crtc, &dev->mode_config.crtc_list, head) {
+		if (!crtc->state->enable)
+			continue;
+
+		crtc_state = drm_atomic_get_crtc_state(state, crtc);
+		if (IS_ERR(crtc_state))
+			return PTR_ERR(crtc_state);
+	}
+
+	for_each_new_crtc_in_state(state, crtc, crtc_state, i) {
+		struct vc4_crtc_state *vc4_crtc_state =
+			to_vc4_crtc_state(crtc_state);
+		struct vc4_crtc *vc4_crtc = to_vc4_crtc(crtc);
+		unsigned int matching_channels;
+
+		if (!crtc_state->active)
+			continue;
+
+		/*
+		 * The problem we have to solve here is that we have
+		 * up to 7 encoders, connected to up to 6 CRTCs.
+		 *
+		 * Those CRTCs, depending on the instance, can be
+		 * routed to 1, 2 or 3 HVS FIFOs, and we need to set
+		 * the change the muxing between FIFOs and outputs in
+		 * the HVS accordingly.
+		 *
+		 * It would be pretty hard to come up with an
+		 * algorithm that would generically solve
+		 * this. However, the current routing trees we support
+		 * allow us to simplify a bit the problem.
+		 *
+		 * Indeed, with the current supported layouts, if we
+		 * try to assign in the ascending crtc index order the
+		 * FIFOs, we can't fall into the situation where an
+		 * earlier CRTC that had multiple routes is assigned
+		 * one that was the only option for a later CRTC.
+		 *
+		 * If the layout changes and doesn't give us that in
+		 * the future, we will need to have something smarter,
+		 * but it works so far.
+		 */
+		matching_channels = unassigned_channels & vc4_crtc->data->hvs_available_channels;
+		if (matching_channels) {
+			unsigned int channel = ffs(matching_channels) - 1;
+
+			vc4_crtc_state->assigned_channel = channel;
+			unassigned_channels &= ~BIT(channel);
+		} else {
+			return -EINVAL;
+		}
+	}
 
 	ret = vc4_ctm_atomic_check(dev, state);
 	if (ret < 0)
@@ -512,12 +703,18 @@ int vc4_kms_load(struct drm_device *dev)
 	struct vc4_dev *vc4 = to_vc4_dev(dev);
 	struct vc4_ctm_state *ctm_state;
 	struct vc4_load_tracker_state *load_state;
+	bool is_vc5 = of_device_is_compatible(dev->dev->of_node,
+					      "brcm,bcm2711-vc5");
 	int ret;
 
-	/* Start with the load tracker enabled. Can be disabled through the
-	 * debugfs load_tracker file.
-	 */
-	vc4->load_tracker_enabled = true;
+	if (!is_vc5) {
+		vc4->load_tracker_available = true;
+
+		/* Start with the load tracker enabled. Can be
+		 * disabled through the debugfs load_tracker file.
+		 */
+		vc4->load_tracker_enabled = true;
+	}
 
 	sema_init(&vc4->async_modeset, 1);
 
@@ -531,8 +728,14 @@ int vc4_kms_load(struct drm_device *dev)
 		return ret;
 	}
 
-	dev->mode_config.max_width = 2048;
-	dev->mode_config.max_height = 2048;
+	if (is_vc5) {
+		dev->mode_config.max_width = 7680;
+		dev->mode_config.max_height = 7680;
+	} else {
+		dev->mode_config.max_width = 2048;
+		dev->mode_config.max_height = 2048;
+	}
+
 	dev->mode_config.funcs = &vc4_mode_funcs;
 	dev->mode_config.preferred_depth = 24;
 	dev->mode_config.async_page_flip = true;
@@ -547,14 +750,17 @@ int vc4_kms_load(struct drm_device *dev)
 	drm_atomic_private_obj_init(dev, &vc4->ctm_manager, &ctm_state->base,
 				    &vc4_ctm_state_funcs);
 
-	load_state = kzalloc(sizeof(*load_state), GFP_KERNEL);
-	if (!load_state) {
-		drm_atomic_private_obj_fini(&vc4->ctm_manager);
-		return -ENOMEM;
-	}
+	if (vc4->load_tracker_available) {
+		load_state = kzalloc(sizeof(*load_state), GFP_KERNEL);
+		if (!load_state) {
+			drm_atomic_private_obj_fini(&vc4->ctm_manager);
+			return -ENOMEM;
+		}
 
-	drm_atomic_private_obj_init(dev, &vc4->load_tracker, &load_state->base,
-				    &vc4_load_tracker_state_funcs);
+		drm_atomic_private_obj_init(dev, &vc4->load_tracker,
+					    &load_state->base,
+					    &vc4_load_tracker_state_funcs);
+	}
 
 	drm_mode_config_reset(dev);
 
diff --git a/drivers/gpu/drm/vc4/vc4_plane.c b/drivers/gpu/drm/vc4/vc4_plane.c
index d040d9f12c6d..6b39cc2ca18d 100644
--- a/drivers/gpu/drm/vc4/vc4_plane.c
+++ b/drivers/gpu/drm/vc4/vc4_plane.c
@@ -32,45 +32,60 @@ static const struct hvs_format {
 	u32 drm; /* DRM_FORMAT_* */
 	u32 hvs; /* HVS_FORMAT_* */
 	u32 pixel_order;
+	u32 pixel_order_hvs5;
 } hvs_formats[] = {
 	{
-		.drm = DRM_FORMAT_XRGB8888, .hvs = HVS_PIXEL_FORMAT_RGBA8888,
+		.drm = DRM_FORMAT_XRGB8888,
+		.hvs = HVS_PIXEL_FORMAT_RGBA8888,
 		.pixel_order = HVS_PIXEL_ORDER_ABGR,
+		.pixel_order_hvs5 = HVS_PIXEL_ORDER_ARGB,
 	},
 	{
-		.drm = DRM_FORMAT_ARGB8888, .hvs = HVS_PIXEL_FORMAT_RGBA8888,
+		.drm = DRM_FORMAT_ARGB8888,
+		.hvs = HVS_PIXEL_FORMAT_RGBA8888,
 		.pixel_order = HVS_PIXEL_ORDER_ABGR,
+		.pixel_order_hvs5 = HVS_PIXEL_ORDER_ARGB,
 	},
 	{
-		.drm = DRM_FORMAT_ABGR8888, .hvs = HVS_PIXEL_FORMAT_RGBA8888,
+		.drm = DRM_FORMAT_ABGR8888,
+		.hvs = HVS_PIXEL_FORMAT_RGBA8888,
 		.pixel_order = HVS_PIXEL_ORDER_ARGB,
+		.pixel_order_hvs5 = HVS_PIXEL_ORDER_ABGR,
 	},
 	{
-		.drm = DRM_FORMAT_XBGR8888, .hvs = HVS_PIXEL_FORMAT_RGBA8888,
+		.drm = DRM_FORMAT_XBGR8888,
+		.hvs = HVS_PIXEL_FORMAT_RGBA8888,
 		.pixel_order = HVS_PIXEL_ORDER_ARGB,
+		.pixel_order_hvs5 = HVS_PIXEL_ORDER_ABGR,
 	},
 	{
-		.drm = DRM_FORMAT_RGB565, .hvs = HVS_PIXEL_FORMAT_RGB565,
+		.drm = DRM_FORMAT_RGB565,
+		.hvs = HVS_PIXEL_FORMAT_RGB565,
 		.pixel_order = HVS_PIXEL_ORDER_XRGB,
 	},
 	{
-		.drm = DRM_FORMAT_BGR565, .hvs = HVS_PIXEL_FORMAT_RGB565,
+		.drm = DRM_FORMAT_BGR565,
+		.hvs = HVS_PIXEL_FORMAT_RGB565,
 		.pixel_order = HVS_PIXEL_ORDER_XBGR,
 	},
 	{
-		.drm = DRM_FORMAT_ARGB1555, .hvs = HVS_PIXEL_FORMAT_RGBA5551,
+		.drm = DRM_FORMAT_ARGB1555,
+		.hvs = HVS_PIXEL_FORMAT_RGBA5551,
 		.pixel_order = HVS_PIXEL_ORDER_ABGR,
 	},
 	{
-		.drm = DRM_FORMAT_XRGB1555, .hvs = HVS_PIXEL_FORMAT_RGBA5551,
+		.drm = DRM_FORMAT_XRGB1555,
+		.hvs = HVS_PIXEL_FORMAT_RGBA5551,
 		.pixel_order = HVS_PIXEL_ORDER_ABGR,
 	},
 	{
-		.drm = DRM_FORMAT_RGB888, .hvs = HVS_PIXEL_FORMAT_RGB888,
+		.drm = DRM_FORMAT_RGB888,
+		.hvs = HVS_PIXEL_FORMAT_RGB888,
 		.pixel_order = HVS_PIXEL_ORDER_XRGB,
 	},
 	{
-		.drm = DRM_FORMAT_BGR888, .hvs = HVS_PIXEL_FORMAT_RGB888,
+		.drm = DRM_FORMAT_BGR888,
+		.hvs = HVS_PIXEL_FORMAT_RGB888,
 		.pixel_order = HVS_PIXEL_ORDER_XBGR,
 	},
 	{
@@ -422,10 +437,7 @@ static void vc4_write_ppf(struct vc4_plane_state *vc4_state, u32 src, u32 dst)
 static u32 vc4_lbm_size(struct drm_plane_state *state)
 {
 	struct vc4_plane_state *vc4_state = to_vc4_plane_state(state);
-	/* This is the worst case number.  One of the two sizes will
-	 * be used depending on the scaling configuration.
-	 */
-	u32 pix_per_line = max(vc4_state->src_w[0], (u32)vc4_state->crtc_w);
+	u32 pix_per_line;
 	u32 lbm;
 
 	/* LBM is not needed when there's no vertical scaling. */
@@ -433,6 +445,18 @@ static u32 vc4_lbm_size(struct drm_plane_state *state)
 	    vc4_state->y_scaling[1] == VC4_SCALING_NONE)
 		return 0;
 
+	/*
+	 * This can be further optimized in the RGB/YUV444 case if the PPF
+	 * decimation factor is between 0.5 and 1.0 by using crtc_w.
+	 *
+	 * It's not an issue though, since in that case since src_w[0] is going
+	 * to be greater than or equal to crtc_w.
+	 */
+	if (vc4_state->x_scaling[0] == VC4_SCALING_TPZ)
+		pix_per_line = vc4_state->crtc_w;
+	else
+		pix_per_line = vc4_state->src_w[0];
+
 	if (!vc4_state->is_yuv) {
 		if (vc4_state->y_scaling[0] == VC4_SCALING_TPZ)
 			lbm = pix_per_line * 8;
@@ -492,6 +516,11 @@ static void vc4_plane_calc_load(struct drm_plane_state *state)
 	struct vc4_plane_state *vc4_state;
 	struct drm_crtc_state *crtc_state;
 	unsigned int vscale_factor;
+	struct vc4_dev *vc4;
+
+	vc4 = to_vc4_dev(state->plane->dev);
+	if (!vc4->load_tracker_available)
+		return;
 
 	vc4_state = to_vc4_plane_state(state);
 	crtc_state = drm_atomic_get_existing_crtc_state(state->state,
@@ -563,7 +592,9 @@ static int vc4_plane_allocate_lbm(struct drm_plane_state *state)
 		spin_lock_irqsave(&vc4->hvs->mm_lock, irqflags);
 		ret = drm_mm_insert_node_generic(&vc4->hvs->lbm_mm,
 						 &vc4_state->lbm,
-						 lbm_size, 32, 0, 0);
+						 lbm_size,
+						 vc4->hvs->hvs5 ? 64 : 32,
+						 0, 0);
 		spin_unlock_irqrestore(&vc4->hvs->mm_lock, irqflags);
 
 		if (ret)
@@ -776,35 +807,6 @@ static int vc4_plane_mode_set(struct drm_plane *plane,
 		return -EINVAL;
 	}
 
-	/* Control word */
-	vc4_dlist_write(vc4_state,
-			SCALER_CTL0_VALID |
-			(rotation & DRM_MODE_REFLECT_X ? SCALER_CTL0_HFLIP : 0) |
-			(rotation & DRM_MODE_REFLECT_Y ? SCALER_CTL0_VFLIP : 0) |
-			VC4_SET_FIELD(SCALER_CTL0_RGBA_EXPAND_ROUND, SCALER_CTL0_RGBA_EXPAND) |
-			(format->pixel_order << SCALER_CTL0_ORDER_SHIFT) |
-			(hvs_format << SCALER_CTL0_PIXEL_FORMAT_SHIFT) |
-			VC4_SET_FIELD(tiling, SCALER_CTL0_TILING) |
-			(vc4_state->is_unity ? SCALER_CTL0_UNITY : 0) |
-			VC4_SET_FIELD(scl0, SCALER_CTL0_SCL0) |
-			VC4_SET_FIELD(scl1, SCALER_CTL0_SCL1));
-
-	/* Position Word 0: Image Positions and Alpha Value */
-	vc4_state->pos0_offset = vc4_state->dlist_count;
-	vc4_dlist_write(vc4_state,
-			VC4_SET_FIELD(state->alpha >> 8, SCALER_POS0_FIXED_ALPHA) |
-			VC4_SET_FIELD(vc4_state->crtc_x, SCALER_POS0_START_X) |
-			VC4_SET_FIELD(vc4_state->crtc_y, SCALER_POS0_START_Y));
-
-	/* Position Word 1: Scaled Image Dimensions. */
-	if (!vc4_state->is_unity) {
-		vc4_dlist_write(vc4_state,
-				VC4_SET_FIELD(vc4_state->crtc_w,
-					      SCALER_POS1_SCL_WIDTH) |
-				VC4_SET_FIELD(vc4_state->crtc_h,
-					      SCALER_POS1_SCL_HEIGHT));
-	}
-
 	/* Don't waste cycles mixing with plane alpha if the set alpha
 	 * is opaque or there is no per-pixel alpha information.
 	 * In any case we use the alpha property value as the fixed alpha.
@@ -812,20 +814,120 @@ static int vc4_plane_mode_set(struct drm_plane *plane,
 	mix_plane_alpha = state->alpha != DRM_BLEND_ALPHA_OPAQUE &&
 			  fb->format->has_alpha;
 
-	/* Position Word 2: Source Image Size, Alpha */
-	vc4_state->pos2_offset = vc4_state->dlist_count;
-	vc4_dlist_write(vc4_state,
-			VC4_SET_FIELD(fb->format->has_alpha ?
-				      SCALER_POS2_ALPHA_MODE_PIPELINE :
-				      SCALER_POS2_ALPHA_MODE_FIXED,
-				      SCALER_POS2_ALPHA_MODE) |
-			(mix_plane_alpha ? SCALER_POS2_ALPHA_MIX : 0) |
-			(fb->format->has_alpha ? SCALER_POS2_ALPHA_PREMULT : 0) |
-			VC4_SET_FIELD(vc4_state->src_w[0], SCALER_POS2_WIDTH) |
-			VC4_SET_FIELD(vc4_state->src_h[0], SCALER_POS2_HEIGHT));
+	if (!vc4->hvs->hvs5) {
+	/* Control word */
+		vc4_dlist_write(vc4_state,
+				SCALER_CTL0_VALID |
+				(rotation & DRM_MODE_REFLECT_X ? SCALER_CTL0_HFLIP : 0) |
+				(rotation & DRM_MODE_REFLECT_Y ? SCALER_CTL0_VFLIP : 0) |
+				VC4_SET_FIELD(SCALER_CTL0_RGBA_EXPAND_ROUND, SCALER_CTL0_RGBA_EXPAND) |
+				(format->pixel_order << SCALER_CTL0_ORDER_SHIFT) |
+				(hvs_format << SCALER_CTL0_PIXEL_FORMAT_SHIFT) |
+				VC4_SET_FIELD(tiling, SCALER_CTL0_TILING) |
+				(vc4_state->is_unity ? SCALER_CTL0_UNITY : 0) |
+				VC4_SET_FIELD(scl0, SCALER_CTL0_SCL0) |
+				VC4_SET_FIELD(scl1, SCALER_CTL0_SCL1));
+
+		/* Position Word 0: Image Positions and Alpha Value */
+		vc4_state->pos0_offset = vc4_state->dlist_count;
+		vc4_dlist_write(vc4_state,
+				VC4_SET_FIELD(state->alpha >> 8, SCALER_POS0_FIXED_ALPHA) |
+				VC4_SET_FIELD(vc4_state->crtc_x, SCALER_POS0_START_X) |
+				VC4_SET_FIELD(vc4_state->crtc_y, SCALER_POS0_START_Y));
+
+		/* Position Word 1: Scaled Image Dimensions. */
+		if (!vc4_state->is_unity) {
+			vc4_dlist_write(vc4_state,
+					VC4_SET_FIELD(vc4_state->crtc_w,
+						      SCALER_POS1_SCL_WIDTH) |
+					VC4_SET_FIELD(vc4_state->crtc_h,
+						      SCALER_POS1_SCL_HEIGHT));
+		}
+
+		/* Position Word 2: Source Image Size, Alpha */
+		vc4_state->pos2_offset = vc4_state->dlist_count;
+		vc4_dlist_write(vc4_state,
+				VC4_SET_FIELD(fb->format->has_alpha ?
+					      SCALER_POS2_ALPHA_MODE_PIPELINE :
+					      SCALER_POS2_ALPHA_MODE_FIXED,
+					      SCALER_POS2_ALPHA_MODE) |
+				(mix_plane_alpha ? SCALER_POS2_ALPHA_MIX : 0) |
+				(fb->format->has_alpha ?
+						SCALER_POS2_ALPHA_PREMULT : 0) |
+				VC4_SET_FIELD(vc4_state->src_w[0],
+					      SCALER_POS2_WIDTH) |
+				VC4_SET_FIELD(vc4_state->src_h[0],
+					      SCALER_POS2_HEIGHT));
+
+		/* Position Word 3: Context.  Written by the HVS. */
+		vc4_dlist_write(vc4_state, 0xc0c0c0c0);
 
-	/* Position Word 3: Context.  Written by the HVS. */
-	vc4_dlist_write(vc4_state, 0xc0c0c0c0);
+	} else {
+		u32 hvs_pixel_order = format->pixel_order;
+
+		if (format->pixel_order_hvs5)
+			hvs_pixel_order = format->pixel_order_hvs5;
+
+		/* Control word */
+		vc4_dlist_write(vc4_state,
+				SCALER_CTL0_VALID |
+				(hvs_pixel_order << SCALER_CTL0_ORDER_SHIFT) |
+				(hvs_format << SCALER_CTL0_PIXEL_FORMAT_SHIFT) |
+				VC4_SET_FIELD(tiling, SCALER_CTL0_TILING) |
+				(vc4_state->is_unity ?
+						SCALER5_CTL0_UNITY : 0) |
+				VC4_SET_FIELD(scl0, SCALER_CTL0_SCL0) |
+				VC4_SET_FIELD(scl1, SCALER_CTL0_SCL1) |
+				SCALER5_CTL0_ALPHA_EXPAND |
+				SCALER5_CTL0_RGB_EXPAND);
+
+		/* Position Word 0: Image Positions and Alpha Value */
+		vc4_state->pos0_offset = vc4_state->dlist_count;
+		vc4_dlist_write(vc4_state,
+				(rotation & DRM_MODE_REFLECT_Y ?
+						SCALER5_POS0_VFLIP : 0) |
+				VC4_SET_FIELD(vc4_state->crtc_x,
+					      SCALER_POS0_START_X) |
+				(rotation & DRM_MODE_REFLECT_X ?
+					      SCALER5_POS0_HFLIP : 0) |
+				VC4_SET_FIELD(vc4_state->crtc_y,
+					      SCALER5_POS0_START_Y)
+			       );
+
+		/* Control Word 2 */
+		vc4_dlist_write(vc4_state,
+				VC4_SET_FIELD(state->alpha >> 4,
+					      SCALER5_CTL2_ALPHA) |
+				(fb->format->has_alpha ?
+					SCALER5_CTL2_ALPHA_PREMULT : 0) |
+				(mix_plane_alpha ?
+					SCALER5_CTL2_ALPHA_MIX : 0) |
+				VC4_SET_FIELD(fb->format->has_alpha ?
+				      SCALER5_CTL2_ALPHA_MODE_PIPELINE :
+				      SCALER5_CTL2_ALPHA_MODE_FIXED,
+				      SCALER5_CTL2_ALPHA_MODE)
+			       );
+
+		/* Position Word 1: Scaled Image Dimensions. */
+		if (!vc4_state->is_unity) {
+			vc4_dlist_write(vc4_state,
+					VC4_SET_FIELD(vc4_state->crtc_w,
+						      SCALER_POS1_SCL_WIDTH) |
+					VC4_SET_FIELD(vc4_state->crtc_h,
+						      SCALER_POS1_SCL_HEIGHT));
+		}
+
+		/* Position Word 2: Source Image Size */
+		vc4_state->pos2_offset = vc4_state->dlist_count;
+		vc4_dlist_write(vc4_state,
+				VC4_SET_FIELD(vc4_state->src_w[0],
+					      SCALER5_POS2_WIDTH) |
+				VC4_SET_FIELD(vc4_state->src_h[0],
+					      SCALER5_POS2_HEIGHT));
+
+		/* Position Word 3: Context.  Written by the HVS. */
+		vc4_dlist_write(vc4_state, 0xc0c0c0c0);
+	}
 
 
 	/* Pointer Word 0/1/2: RGB / Y / Cb / Cr Pointers
@@ -1203,6 +1305,10 @@ static bool vc4_format_mod_supported(struct drm_plane *plane,
 		default:
 			return false;
 		}
+	case DRM_FORMAT_RGBX1010102:
+	case DRM_FORMAT_BGRX1010102:
+	case DRM_FORMAT_RGBA1010102:
+	case DRM_FORMAT_BGRA1010102:
 	case DRM_FORMAT_YUV422:
 	case DRM_FORMAT_YVU422:
 	case DRM_FORMAT_YUV420:
@@ -1255,6 +1361,8 @@ struct drm_plane *vc4_plane_init(struct drm_device *dev,
 				       &vc4_plane_funcs,
 				       formats, ARRAY_SIZE(formats),
 				       modifiers, type, NULL);
+	if (ret)
+		return ERR_PTR(ret);
 
 	drm_plane_helper_add(plane, &vc4_plane_helper_funcs);
 
@@ -1283,7 +1391,7 @@ int vc4_plane_create_additional_planes(struct drm_device *drm)
 	 * modest number of planes to expose, that should hopefully
 	 * still cover any sane usecase.
 	 */
-	for (i = 0; i < 8; i++) {
+	for (i = 0; i < 16; i++) {
 		struct drm_plane *plane =
 			vc4_plane_init(drm, DRM_PLANE_TYPE_OVERLAY);
 
diff --git a/drivers/gpu/drm/vc4/vc4_regs.h b/drivers/gpu/drm/vc4/vc4_regs.h
index 324462cc9cd4..be2c32a519b3 100644
--- a/drivers/gpu/drm/vc4/vc4_regs.h
+++ b/drivers/gpu/drm/vc4/vc4_regs.h
@@ -129,6 +129,8 @@
 #define V3D_ERRSTAT  0x00f20
 
 #define PV_CONTROL				0x00
+# define PV5_CONTROL_FIFO_LEVEL_HIGH_MASK	VC4_MASK(26, 25)
+# define PV5_CONTROL_FIFO_LEVEL_HIGH_SHIFT	25
 # define PV_CONTROL_FORMAT_MASK			VC4_MASK(23, 21)
 # define PV_CONTROL_FORMAT_SHIFT		21
 # define PV_CONTROL_FORMAT_24			0
@@ -208,6 +210,11 @@
 
 #define PV_HACT_ACT				0x30
 
+#define PV_MUX_CFG				0x34
+# define PV_MUX_CFG_RGB_PIXEL_MUX_MODE_MASK	VC4_MASK(5, 2)
+# define PV_MUX_CFG_RGB_PIXEL_MUX_MODE_SHIFT	2
+# define PV_MUX_CFG_RGB_PIXEL_MUX_MODE_NO_SWAP	8
+
 #define SCALER_CHANNELS_COUNT			3
 
 #define SCALER_DISPCTRL                         0x00000000
@@ -286,9 +293,19 @@
 
 #define SCALER_DISPID                           0x00000008
 #define SCALER_DISPECTRL                        0x0000000c
+# define SCALER_DISPECTRL_DSP2_MUX_SHIFT	31
+# define SCALER_DISPECTRL_DSP2_MUX_MASK		VC4_MASK(31, 31)
+
 #define SCALER_DISPPROF                         0x00000010
+
 #define SCALER_DISPDITHER                       0x00000014
+# define SCALER_DISPDITHER_DSP5_MUX_SHIFT	30
+# define SCALER_DISPDITHER_DSP5_MUX_MASK	VC4_MASK(31, 30)
+
 #define SCALER_DISPEOLN                         0x00000018
+# define SCALER_DISPEOLN_DSP4_MUX_SHIFT		30
+# define SCALER_DISPEOLN_DSP4_MUX_MASK		VC4_MASK(31, 30)
+
 #define SCALER_DISPLIST0                        0x00000020
 #define SCALER_DISPLIST1                        0x00000024
 #define SCALER_DISPLIST2                        0x00000028
@@ -327,6 +344,20 @@
 # define SCALER_DISPCTRLX_HEIGHT_MASK		VC4_MASK(11, 0)
 # define SCALER_DISPCTRLX_HEIGHT_SHIFT		0
 
+# define SCALER5_DISPCTRLX_WIDTH_MASK		VC4_MASK(28, 16)
+# define SCALER5_DISPCTRLX_WIDTH_SHIFT		16
+/* Generates a single frame when VSTART is seen and stops at the last
+ * pixel read from the FIFO.
+ */
+# define SCALER5_DISPCTRLX_ONESHOT		BIT(15)
+/* Processes a single context in the dlist and then task switch,
+ * instead of an entire line.
+ */
+# define SCALER5_DISPCTRLX_ONECTX_MASK		VC4_MASK(14, 13)
+# define SCALER5_DISPCTRLX_ONECTX_SHIFT		13
+# define SCALER5_DISPCTRLX_HEIGHT_MASK		VC4_MASK(12, 0)
+# define SCALER5_DISPCTRLX_HEIGHT_SHIFT		0
+
 #define SCALER_DISPBKGND0                       0x00000044
 # define SCALER_DISPBKGND_AUTOHS		BIT(31)
 # define SCALER_DISPBKGND_INTERLACE		BIT(30)
@@ -460,32 +491,18 @@
 #define SCALER_DLIST_START                      0x00002000
 #define SCALER_DLIST_SIZE                       0x00004000
 
-#define VC4_HDMI_CORE_REV			0x000
+#define SCALER5_DLIST_START			0x00004000
 
-#define VC4_HDMI_SW_RESET_CONTROL		0x004
 # define VC4_HDMI_SW_RESET_FORMAT_DETECT	BIT(1)
 # define VC4_HDMI_SW_RESET_HDMI			BIT(0)
 
-#define VC4_HDMI_HOTPLUG_INT			0x008
-
-#define VC4_HDMI_HOTPLUG			0x00c
 # define VC4_HDMI_HOTPLUG_CONNECTED		BIT(0)
 
-/* 3 bits per field, where each field maps from that corresponding MAI
- * bus channel to the given HDMI channel.
- */
-#define VC4_HDMI_MAI_CHANNEL_MAP		0x090
-
-#define VC4_HDMI_MAI_CONFIG			0x094
 # define VC4_HDMI_MAI_CONFIG_FORMAT_REVERSE		BIT(27)
 # define VC4_HDMI_MAI_CONFIG_BIT_REVERSE		BIT(26)
 # define VC4_HDMI_MAI_CHANNEL_MASK_MASK			VC4_MASK(15, 0)
 # define VC4_HDMI_MAI_CHANNEL_MASK_SHIFT		0
 
-/* Last received format word on the MAI bus. */
-#define VC4_HDMI_MAI_FORMAT			0x098
-
-#define VC4_HDMI_AUDIO_PACKET_CONFIG		0x09c
 # define VC4_HDMI_AUDIO_PACKET_ZERO_DATA_ON_SAMPLE_FLAT		BIT(29)
 # define VC4_HDMI_AUDIO_PACKET_ZERO_DATA_ON_INACTIVE_CHANNELS	BIT(24)
 # define VC4_HDMI_AUDIO_PACKET_FORCE_SAMPLE_PRESENT		BIT(19)
@@ -499,12 +516,8 @@
 # define VC4_HDMI_AUDIO_PACKET_CEA_MASK_MASK			VC4_MASK(7, 0)
 # define VC4_HDMI_AUDIO_PACKET_CEA_MASK_SHIFT			0
 
-#define VC4_HDMI_RAM_PACKET_CONFIG		0x0a0
 # define VC4_HDMI_RAM_PACKET_ENABLE		BIT(16)
 
-#define VC4_HDMI_RAM_PACKET_STATUS		0x0a4
-
-#define VC4_HDMI_CRP_CFG			0x0a8
 /* When set, the CTS_PERIOD counts based on MAI bus sync pulse instead
  * of pixel clock.
  */
@@ -518,23 +531,12 @@
 # define VC4_HDMI_CRP_CFG_N_MASK		VC4_MASK(19, 0)
 # define VC4_HDMI_CRP_CFG_N_SHIFT		0
 
-/* 20-bit fields containing CTS values to be transmitted if !EXTERNAL_CTS_EN */
-#define VC4_HDMI_CTS_0				0x0ac
-#define VC4_HDMI_CTS_1				0x0b0
-/* 20-bit fields containing number of clocks to send CTS0/1 before
- * switching to the other one.
- */
-#define VC4_HDMI_CTS_PERIOD_0			0x0b4
-#define VC4_HDMI_CTS_PERIOD_1			0x0b8
-
-#define VC4_HDMI_HORZA				0x0c4
 # define VC4_HDMI_HORZA_VPOS			BIT(14)
 # define VC4_HDMI_HORZA_HPOS			BIT(13)
 /* Horizontal active pixels (hdisplay). */
 # define VC4_HDMI_HORZA_HAP_MASK		VC4_MASK(12, 0)
 # define VC4_HDMI_HORZA_HAP_SHIFT		0
 
-#define VC4_HDMI_HORZB				0x0c8
 /* Horizontal pack porch (htotal - hsync_end). */
 # define VC4_HDMI_HORZB_HBP_MASK		VC4_MASK(29, 20)
 # define VC4_HDMI_HORZB_HBP_SHIFT		20
@@ -545,7 +547,6 @@
 # define VC4_HDMI_HORZB_HFP_MASK		VC4_MASK(9, 0)
 # define VC4_HDMI_HORZB_HFP_SHIFT		0
 
-#define VC4_HDMI_FIFO_CTL			0x05c
 # define VC4_HDMI_FIFO_CTL_RECENTER_DONE	BIT(14)
 # define VC4_HDMI_FIFO_CTL_USE_EMPTY		BIT(13)
 # define VC4_HDMI_FIFO_CTL_ON_VB		BIT(7)
@@ -558,15 +559,12 @@
 # define VC4_HDMI_FIFO_CTL_MASTER_SLAVE_N	BIT(0)
 # define VC4_HDMI_FIFO_VALID_WRITE_MASK		0xefff
 
-#define VC4_HDMI_SCHEDULER_CONTROL		0x0c0
 # define VC4_HDMI_SCHEDULER_CONTROL_MANUAL_FORMAT BIT(15)
 # define VC4_HDMI_SCHEDULER_CONTROL_IGNORE_VSYNC_PREDICTS BIT(5)
 # define VC4_HDMI_SCHEDULER_CONTROL_VERT_ALWAYS_KEEPOUT	BIT(3)
 # define VC4_HDMI_SCHEDULER_CONTROL_HDMI_ACTIVE	BIT(1)
 # define VC4_HDMI_SCHEDULER_CONTROL_MODE_HDMI	BIT(0)
 
-#define VC4_HDMI_VERTA0				0x0cc
-#define VC4_HDMI_VERTA1				0x0d4
 /* Vertical sync pulse (vsync_end - vsync_start). */
 # define VC4_HDMI_VERTA_VSP_MASK		VC4_MASK(24, 20)
 # define VC4_HDMI_VERTA_VSP_SHIFT		20
@@ -577,8 +575,6 @@
 # define VC4_HDMI_VERTA_VAL_MASK		VC4_MASK(12, 0)
 # define VC4_HDMI_VERTA_VAL_SHIFT		0
 
-#define VC4_HDMI_VERTB0				0x0d0
-#define VC4_HDMI_VERTB1				0x0d8
 /* Vertical sync pulse offset (for interlaced) */
 # define VC4_HDMI_VERTB_VSPO_MASK		VC4_MASK(21, 9)
 # define VC4_HDMI_VERTB_VSPO_SHIFT		9
@@ -586,7 +582,6 @@
 # define VC4_HDMI_VERTB_VBP_MASK		VC4_MASK(8, 0)
 # define VC4_HDMI_VERTB_VBP_SHIFT		0
 
-#define VC4_HDMI_CEC_CNTRL_1			0x0e8
 /* Set when the transmission has ended. */
 # define VC4_HDMI_CEC_TX_EOM			BIT(31)
 /* If set, transmission was acked on the 1st or 2nd attempt (only one
@@ -627,7 +622,6 @@
 /* Set these fields to how many bit clock cycles get to that many
  * microseconds.
  */
-#define VC4_HDMI_CEC_CNTRL_2			0x0ec
 # define VC4_HDMI_CEC_CNT_TO_1500_US_MASK	VC4_MASK(30, 24)
 # define VC4_HDMI_CEC_CNT_TO_1500_US_SHIFT	24
 # define VC4_HDMI_CEC_CNT_TO_1300_US_MASK	VC4_MASK(23, 17)
@@ -639,7 +633,6 @@
 # define VC4_HDMI_CEC_CNT_TO_400_US_MASK	VC4_MASK(4, 0)
 # define VC4_HDMI_CEC_CNT_TO_400_US_SHIFT	0
 
-#define VC4_HDMI_CEC_CNTRL_3			0x0f0
 # define VC4_HDMI_CEC_CNT_TO_2750_US_MASK	VC4_MASK(31, 24)
 # define VC4_HDMI_CEC_CNT_TO_2750_US_SHIFT	24
 # define VC4_HDMI_CEC_CNT_TO_2400_US_MASK	VC4_MASK(23, 16)
@@ -649,7 +642,6 @@
 # define VC4_HDMI_CEC_CNT_TO_1700_US_MASK	VC4_MASK(7, 0)
 # define VC4_HDMI_CEC_CNT_TO_1700_US_SHIFT	0
 
-#define VC4_HDMI_CEC_CNTRL_4			0x0f4
 # define VC4_HDMI_CEC_CNT_TO_4300_US_MASK	VC4_MASK(31, 24)
 # define VC4_HDMI_CEC_CNT_TO_4300_US_SHIFT	24
 # define VC4_HDMI_CEC_CNT_TO_3900_US_MASK	VC4_MASK(23, 16)
@@ -659,7 +651,6 @@
 # define VC4_HDMI_CEC_CNT_TO_3500_US_MASK	VC4_MASK(7, 0)
 # define VC4_HDMI_CEC_CNT_TO_3500_US_SHIFT	0
 
-#define VC4_HDMI_CEC_CNTRL_5			0x0f8
 # define VC4_HDMI_CEC_TX_SW_RESET		BIT(27)
 # define VC4_HDMI_CEC_RX_SW_RESET		BIT(26)
 # define VC4_HDMI_CEC_PAD_SW_RESET		BIT(25)
@@ -672,39 +663,11 @@
 # define VC4_HDMI_CEC_CNT_TO_4500_US_MASK	VC4_MASK(7, 0)
 # define VC4_HDMI_CEC_CNT_TO_4500_US_SHIFT	0
 
-/* Transmit data, first byte is low byte of the 32-bit reg.  MSB of
- * each byte transmitted first.
- */
-#define VC4_HDMI_CEC_TX_DATA_1			0x0fc
-#define VC4_HDMI_CEC_TX_DATA_2			0x100
-#define VC4_HDMI_CEC_TX_DATA_3			0x104
-#define VC4_HDMI_CEC_TX_DATA_4			0x108
-#define VC4_HDMI_CEC_RX_DATA_1			0x10c
-#define VC4_HDMI_CEC_RX_DATA_2			0x110
-#define VC4_HDMI_CEC_RX_DATA_3			0x114
-#define VC4_HDMI_CEC_RX_DATA_4			0x118
-
-#define VC4_HDMI_TX_PHY_RESET_CTL		0x2c0
-
-#define VC4_HDMI_TX_PHY_CTL0			0x2c4
 # define VC4_HDMI_TX_PHY_RNG_PWRDN		BIT(25)
 
-/* Interrupt status bits */
-#define VC4_HDMI_CPU_STATUS			0x340
-#define VC4_HDMI_CPU_SET			0x344
-#define VC4_HDMI_CPU_CLEAR			0x348
 # define VC4_HDMI_CPU_CEC			BIT(6)
 # define VC4_HDMI_CPU_HOTPLUG			BIT(0)
 
-#define VC4_HDMI_CPU_MASK_STATUS		0x34c
-#define VC4_HDMI_CPU_MASK_SET			0x350
-#define VC4_HDMI_CPU_MASK_CLEAR			0x354
-
-#define VC4_HDMI_GCP(x)				(0x400 + ((x) * 0x4))
-#define VC4_HDMI_RAM_PACKET(x)			(0x400 + ((x) * 0x24))
-#define VC4_HDMI_PACKET_STRIDE			0x24
-
-#define VC4_HD_M_CTL				0x00c
 /* Debug: Current receive value on the CEC pad. */
 # define VC4_HD_CECRXD				BIT(9)
 /* Debug: Override CEC output to 0. */
@@ -714,7 +677,6 @@
 # define VC4_HD_M_SW_RST			BIT(2)
 # define VC4_HD_M_ENABLE			BIT(0)
 
-#define VC4_HD_MAI_CTL				0x014
 /* Set when audio stream is received at a slower rate than the
  * sampling period, so MAI fifo goes empty.  Write 1 to clear.
  */
@@ -739,7 +701,6 @@
 /* Single-shot reset bit.  Read value is undefined. */
 # define VC4_HD_MAI_CTL_RESET			BIT(0)
 
-#define VC4_HD_MAI_THR				0x018
 # define VC4_HD_MAI_THR_PANICHIGH_MASK		VC4_MASK(29, 24)
 # define VC4_HD_MAI_THR_PANICHIGH_SHIFT		24
 # define VC4_HD_MAI_THR_PANICLOW_MASK		VC4_MASK(21, 16)
@@ -749,31 +710,23 @@
 # define VC4_HD_MAI_THR_DREQLOW_MASK		VC4_MASK(5, 0)
 # define VC4_HD_MAI_THR_DREQLOW_SHIFT		0
 
-/* Format header to be placed on the MAI data. Unused. */
-#define VC4_HD_MAI_FMT				0x01c
-
-/* Register for DMAing in audio data to be transported over the MAI
- * bus to the Falcon core.
- */
-#define VC4_HD_MAI_DATA				0x020
-
 /* Divider from HDMI HSM clock to MAI serial clock.  Sampling period
  * converges to N / (M + 1) cycles.
  */
-#define VC4_HD_MAI_SMP				0x02c
 # define VC4_HD_MAI_SMP_N_MASK			VC4_MASK(31, 8)
 # define VC4_HD_MAI_SMP_N_SHIFT			8
 # define VC4_HD_MAI_SMP_M_MASK			VC4_MASK(7, 0)
 # define VC4_HD_MAI_SMP_M_SHIFT			0
 
-#define VC4_HD_VID_CTL				0x038
 # define VC4_HD_VID_CTL_ENABLE			BIT(31)
 # define VC4_HD_VID_CTL_UNDERFLOW_ENABLE	BIT(30)
 # define VC4_HD_VID_CTL_FRAME_COUNTER_RESET	BIT(29)
 # define VC4_HD_VID_CTL_VSYNC_LOW		BIT(28)
 # define VC4_HD_VID_CTL_HSYNC_LOW		BIT(27)
+# define VC4_HD_VID_CTL_CLRSYNC			BIT(24)
+# define VC4_HD_VID_CTL_CLRRGB			BIT(23)
+# define VC4_HD_VID_CTL_BLANKPIX		BIT(18)
 
-#define VC4_HD_CSC_CTL				0x040
 # define VC4_HD_CSC_CTL_ORDER_MASK		VC4_MASK(7, 5)
 # define VC4_HD_CSC_CTL_ORDER_SHIFT		5
 # define VC4_HD_CSC_CTL_ORDER_RGB		0
@@ -791,14 +744,7 @@
 # define VC4_HD_CSC_CTL_RGB2YCC			BIT(1)
 # define VC4_HD_CSC_CTL_ENABLE			BIT(0)
 
-#define VC4_HD_CSC_12_11			0x044
-#define VC4_HD_CSC_14_13			0x048
-#define VC4_HD_CSC_22_21			0x04c
-#define VC4_HD_CSC_24_23			0x050
-#define VC4_HD_CSC_32_31			0x054
-#define VC4_HD_CSC_34_33			0x058
-
-#define VC4_HD_FRAME_COUNT			0x068
+# define VC4_DVP_HT_CLOCK_STOP_PIXEL		BIT(1)
 
 /* HVS display list information. */
 #define HVS_BOOTLOADER_DLIST_END                32
@@ -825,6 +771,8 @@ enum hvs_pixel_format {
 	HVS_PIXEL_FORMAT_PALETTE = 13,
 	HVS_PIXEL_FORMAT_YUV444_RGB = 14,
 	HVS_PIXEL_FORMAT_AYUV444_RGB = 15,
+	HVS_PIXEL_FORMAT_RGBA1010102 = 16,
+	HVS_PIXEL_FORMAT_YCBCR_10BIT = 17,
 };
 
 /* Note: the LSB is the rightmost character shown.  Only valid for
@@ -879,6 +827,10 @@ enum hvs_pixel_format {
 #define SCALER_CTL0_RGBA_EXPAND_MSB		2
 #define SCALER_CTL0_RGBA_EXPAND_ROUND		3
 
+#define SCALER5_CTL0_ALPHA_EXPAND		BIT(12)
+
+#define SCALER5_CTL0_RGB_EXPAND			BIT(11)
+
 #define SCALER_CTL0_SCL1_MASK			VC4_MASK(10, 8)
 #define SCALER_CTL0_SCL1_SHIFT			8
 
@@ -896,10 +848,13 @@ enum hvs_pixel_format {
 
 /* Set to indicate no scaling. */
 #define SCALER_CTL0_UNITY			BIT(4)
+#define SCALER5_CTL0_UNITY			BIT(15)
 
 #define SCALER_CTL0_PIXEL_FORMAT_MASK		VC4_MASK(3, 0)
 #define SCALER_CTL0_PIXEL_FORMAT_SHIFT		0
 
+#define SCALER5_CTL0_PIXEL_FORMAT_MASK		VC4_MASK(4, 0)
+
 #define SCALER_POS0_FIXED_ALPHA_MASK		VC4_MASK(31, 24)
 #define SCALER_POS0_FIXED_ALPHA_SHIFT		24
 
@@ -909,12 +864,48 @@ enum hvs_pixel_format {
 #define SCALER_POS0_START_X_MASK		VC4_MASK(11, 0)
 #define SCALER_POS0_START_X_SHIFT		0
 
+#define SCALER5_POS0_START_Y_MASK		VC4_MASK(27, 16)
+#define SCALER5_POS0_START_Y_SHIFT		16
+
+#define SCALER5_POS0_START_X_MASK		VC4_MASK(13, 0)
+#define SCALER5_POS0_START_X_SHIFT		0
+
+#define SCALER5_POS0_VFLIP			BIT(31)
+#define SCALER5_POS0_HFLIP			BIT(15)
+
+#define SCALER5_CTL2_ALPHA_MODE_MASK		VC4_MASK(31, 30)
+#define SCALER5_CTL2_ALPHA_MODE_SHIFT		30
+#define SCALER5_CTL2_ALPHA_MODE_PIPELINE		0
+#define SCALER5_CTL2_ALPHA_MODE_FIXED		1
+#define SCALER5_CTL2_ALPHA_MODE_FIXED_NONZERO	2
+#define SCALER5_CTL2_ALPHA_MODE_FIXED_OVER_0x07	3
+
+#define SCALER5_CTL2_ALPHA_PREMULT		BIT(29)
+
+#define SCALER5_CTL2_ALPHA_MIX			BIT(28)
+
+#define SCALER5_CTL2_ALPHA_LOC			BIT(25)
+
+#define SCALER5_CTL2_MAP_SEL_MASK		VC4_MASK(18, 17)
+#define SCALER5_CTL2_MAP_SEL_SHIFT		17
+
+#define SCALER5_CTL2_GAMMA			BIT(16)
+
+#define SCALER5_CTL2_ALPHA_MASK			VC4_MASK(15, 4)
+#define SCALER5_CTL2_ALPHA_SHIFT		4
+
 #define SCALER_POS1_SCL_HEIGHT_MASK		VC4_MASK(27, 16)
 #define SCALER_POS1_SCL_HEIGHT_SHIFT		16
 
 #define SCALER_POS1_SCL_WIDTH_MASK		VC4_MASK(11, 0)
 #define SCALER_POS1_SCL_WIDTH_SHIFT		0
 
+#define SCALER5_POS1_SCL_HEIGHT_MASK		VC4_MASK(28, 16)
+#define SCALER5_POS1_SCL_HEIGHT_SHIFT		16
+
+#define SCALER5_POS1_SCL_WIDTH_MASK		VC4_MASK(12, 0)
+#define SCALER5_POS1_SCL_WIDTH_SHIFT		0
+
 #define SCALER_POS2_ALPHA_MODE_MASK		VC4_MASK(31, 30)
 #define SCALER_POS2_ALPHA_MODE_SHIFT		30
 #define SCALER_POS2_ALPHA_MODE_PIPELINE		0
@@ -930,6 +921,12 @@ enum hvs_pixel_format {
 #define SCALER_POS2_WIDTH_MASK			VC4_MASK(11, 0)
 #define SCALER_POS2_WIDTH_SHIFT			0
 
+#define SCALER5_POS2_HEIGHT_MASK		VC4_MASK(28, 16)
+#define SCALER5_POS2_HEIGHT_SHIFT		16
+
+#define SCALER5_POS2_WIDTH_MASK			VC4_MASK(12, 0)
+#define SCALER5_POS2_WIDTH_SHIFT		0
+
 /* Color Space Conversion words.  Some values are S2.8 signed
  * integers, except that the 2 integer bits map as {0x0: 0, 0x1: 1,
  * 0x2: 2, 0x3: -1}
diff --git a/drivers/gpu/drm/vc4/vc4_txp.c b/drivers/gpu/drm/vc4/vc4_txp.c
index a7c3af0005a0..849dcafbfff1 100644
--- a/drivers/gpu/drm/vc4/vc4_txp.c
+++ b/drivers/gpu/drm/vc4/vc4_txp.c
@@ -436,7 +436,6 @@ static const struct drm_crtc_helper_funcs vc4_txp_crtc_helper_funcs = {
 	.atomic_flush	= vc4_hvs_atomic_flush,
 	.atomic_enable	= vc4_txp_atomic_enable,
 	.atomic_disable	= vc4_txp_atomic_disable,
-	.mode_set_nofb	= vc4_hvs_mode_set_nofb,
 };
 
 static irqreturn_t vc4_txp_interrupt(int irq, void *data)
@@ -452,7 +451,8 @@ static irqreturn_t vc4_txp_interrupt(int irq, void *data)
 }
 
 static const struct vc4_crtc_data vc4_txp_crtc_data = {
-	.hvs_channel = 2,
+	.hvs_available_channels = BIT(2),
+	.hvs_output = 2,
 };
 
 static int vc4_txp_bind(struct device *dev, struct device *master, void *data)
diff --git a/drivers/gpu/drm/vgem/vgem_drv.c b/drivers/gpu/drm/vgem/vgem_drv.c
index 313339bbff90..cb884c890065 100644
--- a/drivers/gpu/drm/vgem/vgem_drv.c
+++ b/drivers/gpu/drm/vgem/vgem_drv.c
@@ -321,7 +321,7 @@ static struct sg_table *vgem_prime_get_sg_table(struct drm_gem_object *obj)
 {
 	struct drm_vgem_gem_object *bo = to_vgem_bo(obj);
 
-	return drm_prime_pages_to_sg(bo->pages, bo->base.size >> PAGE_SHIFT);
+	return drm_prime_pages_to_sg(obj->dev, bo->pages, bo->base.size >> PAGE_SHIFT);
 }
 
 static struct drm_gem_object* vgem_prime_import(struct drm_device *dev,
@@ -401,16 +401,8 @@ static int vgem_prime_mmap(struct drm_gem_object *obj,
 	return 0;
 }
 
-static void vgem_release(struct drm_device *dev)
-{
-	struct vgem_device *vgem = container_of(dev, typeof(*vgem), drm);
-
-	platform_device_unregister(vgem->platform);
-}
-
 static struct drm_driver vgem_driver = {
 	.driver_features		= DRIVER_GEM | DRIVER_RENDER,
-	.release			= vgem_release,
 	.open				= vgem_open,
 	.postclose			= vgem_postclose,
 	.gem_free_object_unlocked	= vgem_gem_free_object,
@@ -442,48 +434,49 @@ static struct drm_driver vgem_driver = {
 static int __init vgem_init(void)
 {
 	int ret;
+	struct platform_device *pdev;
 
-	vgem_device = kzalloc(sizeof(*vgem_device), GFP_KERNEL);
-	if (!vgem_device)
-		return -ENOMEM;
+	pdev = platform_device_register_simple("vgem", -1, NULL, 0);
+	if (IS_ERR(pdev))
+		return PTR_ERR(pdev);
 
-	vgem_device->platform =
-		platform_device_register_simple("vgem", -1, NULL, 0);
-	if (IS_ERR(vgem_device->platform)) {
-		ret = PTR_ERR(vgem_device->platform);
-		goto out_free;
+	if (!devres_open_group(&pdev->dev, NULL, GFP_KERNEL)) {
+		ret = -ENOMEM;
+		goto out_unregister;
 	}
 
-	dma_coerce_mask_and_coherent(&vgem_device->platform->dev,
+	dma_coerce_mask_and_coherent(&pdev->dev,
 				     DMA_BIT_MASK(64));
-	ret = drm_dev_init(&vgem_device->drm, &vgem_driver,
-			   &vgem_device->platform->dev);
-	if (ret)
-		goto out_unregister;
-	drmm_add_final_kfree(&vgem_device->drm, vgem_device);
+
+	vgem_device = devm_drm_dev_alloc(&pdev->dev, &vgem_driver,
+					 struct vgem_device, drm);
+	if (IS_ERR(vgem_device)) {
+		ret = PTR_ERR(vgem_device);
+		goto out_devres;
+	}
+	vgem_device->platform = pdev;
 
 	/* Final step: expose the device/driver to userspace */
 	ret = drm_dev_register(&vgem_device->drm, 0);
 	if (ret)
-		goto out_put;
+		goto out_devres;
 
 	return 0;
 
-out_put:
-	drm_dev_put(&vgem_device->drm);
-	platform_device_unregister(vgem_device->platform);
-	return ret;
+out_devres:
+	devres_release_group(&pdev->dev, NULL);
 out_unregister:
-	platform_device_unregister(vgem_device->platform);
-out_free:
-	kfree(vgem_device);
+	platform_device_unregister(pdev);
 	return ret;
 }
 
 static void __exit vgem_exit(void)
 {
+	struct platform_device *pdev = vgem_device->platform;
+
 	drm_dev_unregister(&vgem_device->drm);
-	drm_dev_put(&vgem_device->drm);
+	devres_release_group(&pdev->dev, NULL);
+	platform_device_unregister(pdev);
 }
 
 module_init(vgem_init);
diff --git a/drivers/gpu/drm/virtio/virtgpu_debugfs.c b/drivers/gpu/drm/virtio/virtgpu_debugfs.c
index 3221520f61f0..d5b0c543bd6d 100644
--- a/drivers/gpu/drm/virtio/virtgpu_debugfs.c
+++ b/drivers/gpu/drm/virtio/virtgpu_debugfs.c
@@ -48,6 +48,7 @@ static int virtio_gpu_features(struct seq_file *m, void *data)
 	virtio_add_bool(m, "virgl", vgdev->has_virgl_3d);
 	virtio_add_bool(m, "edid", vgdev->has_edid);
 	virtio_add_bool(m, "indirect", vgdev->has_indirect);
+	virtio_add_bool(m, "resource uuid", vgdev->has_resource_assign_uuid);
 	virtio_add_int(m, "cap sets", vgdev->num_capsets);
 	virtio_add_int(m, "scanouts", vgdev->num_scanouts);
 	return 0;
diff --git a/drivers/gpu/drm/virtio/virtgpu_display.c b/drivers/gpu/drm/virtio/virtgpu_display.c
index effea07abe62..f84b7e61311b 100644
--- a/drivers/gpu/drm/virtio/virtgpu_display.c
+++ b/drivers/gpu/drm/virtio/virtgpu_display.c
@@ -325,11 +325,14 @@ static const struct drm_mode_config_funcs virtio_gpu_mode_funcs = {
 	.atomic_commit = drm_atomic_helper_commit,
 };
 
-void virtio_gpu_modeset_init(struct virtio_gpu_device *vgdev)
+int virtio_gpu_modeset_init(struct virtio_gpu_device *vgdev)
 {
-	int i;
+	int i, ret;
+
+	ret = drmm_mode_config_init(vgdev->ddev);
+	if (ret)
+		return ret;
 
-	drm_mode_config_init(vgdev->ddev);
 	vgdev->ddev->mode_config.quirk_addfb_prefer_host_byte_order = true;
 	vgdev->ddev->mode_config.funcs = &virtio_gpu_mode_funcs;
 
@@ -343,6 +346,7 @@ void virtio_gpu_modeset_init(struct virtio_gpu_device *vgdev)
 		vgdev_output_init(vgdev, i);
 
 	drm_mode_config_reset(vgdev->ddev);
+	return 0;
 }
 
 void virtio_gpu_modeset_fini(struct virtio_gpu_device *vgdev)
@@ -351,5 +355,4 @@ void virtio_gpu_modeset_fini(struct virtio_gpu_device *vgdev)
 
 	for (i = 0 ; i < vgdev->num_scanouts; ++i)
 		kfree(vgdev->outputs[i].edid);
-	drm_mode_config_cleanup(vgdev->ddev);
 }
diff --git a/drivers/gpu/drm/virtio/virtgpu_drv.h b/drivers/gpu/drm/virtio/virtgpu_drv.h
index a52b7a39f286..55c34b4fc3e9 100644
--- a/drivers/gpu/drm/virtio/virtgpu_drv.h
+++ b/drivers/gpu/drm/virtio/virtgpu_drv.h
@@ -352,7 +352,7 @@ virtio_gpu_cmd_resource_assign_uuid(struct virtio_gpu_device *vgdev,
 				    struct virtio_gpu_object_array *objs);
 
 /* virtgpu_display.c */
-void virtio_gpu_modeset_init(struct virtio_gpu_device *vgdev);
+int virtio_gpu_modeset_init(struct virtio_gpu_device *vgdev);
 void virtio_gpu_modeset_fini(struct virtio_gpu_device *vgdev);
 
 /* virtgpu_plane.c */
diff --git a/drivers/gpu/drm/virtio/virtgpu_kms.c b/drivers/gpu/drm/virtio/virtgpu_kms.c
index 75d0dc2f6d28..eed57a931309 100644
--- a/drivers/gpu/drm/virtio/virtgpu_kms.c
+++ b/drivers/gpu/drm/virtio/virtgpu_kms.c
@@ -80,8 +80,10 @@ static void virtio_gpu_get_capsets(struct virtio_gpu_device *vgdev,
 					 vgdev->capsets[i].id > 0, 5 * HZ);
 		if (ret == 0) {
 			DRM_ERROR("timed out waiting for cap set %d\n", i);
+			spin_lock(&vgdev->display_info_lock);
 			kfree(vgdev->capsets);
 			vgdev->capsets = NULL;
+			spin_unlock(&vgdev->display_info_lock);
 			return;
 		}
 		DRM_INFO("cap set %d: id %d, max-version %d, max-size %d\n",
@@ -103,7 +105,7 @@ int virtio_gpu_init(struct drm_device *dev)
 	/* this will expand later */
 	struct virtqueue *vqs[2];
 	u32 num_scanouts, num_capsets;
-	int ret;
+	int ret = 0;
 
 	if (!virtio_has_feature(dev_to_virtio(dev->dev), VIRTIO_F_VERSION_1))
 		return -ENODEV;
@@ -184,7 +186,11 @@ int virtio_gpu_init(struct drm_device *dev)
 			num_capsets, &num_capsets);
 	DRM_INFO("number of cap sets: %d\n", num_capsets);
 
-	virtio_gpu_modeset_init(vgdev);
+	ret = virtio_gpu_modeset_init(vgdev);
+	if (ret) {
+		DRM_ERROR("modeset init failed\n");
+		goto err_scanouts;
+	}
 
 	virtio_device_ready(vgdev->vdev);
 
diff --git a/drivers/gpu/drm/virtio/virtgpu_vq.c b/drivers/gpu/drm/virtio/virtgpu_vq.c
index 651d1b0e8e8d..07945ca238e2 100644
--- a/drivers/gpu/drm/virtio/virtgpu_vq.c
+++ b/drivers/gpu/drm/virtio/virtgpu_vq.c
@@ -320,13 +320,13 @@ static struct sg_table *vmalloc_to_sgt(char *data, uint32_t size, int *sg_ents)
 	return sgt;
 }
 
-static void virtio_gpu_queue_ctrl_sgs(struct virtio_gpu_device *vgdev,
-				      struct virtio_gpu_vbuffer *vbuf,
-				      struct virtio_gpu_fence *fence,
-				      int elemcnt,
-				      struct scatterlist **sgs,
-				      int outcnt,
-				      int incnt)
+static int virtio_gpu_queue_ctrl_sgs(struct virtio_gpu_device *vgdev,
+				     struct virtio_gpu_vbuffer *vbuf,
+				     struct virtio_gpu_fence *fence,
+				     int elemcnt,
+				     struct scatterlist **sgs,
+				     int outcnt,
+				     int incnt)
 {
 	struct virtqueue *vq = vgdev->ctrlq.vq;
 	int ret, idx;
@@ -335,7 +335,7 @@ static void virtio_gpu_queue_ctrl_sgs(struct virtio_gpu_device *vgdev,
 		if (fence && vbuf->objs)
 			virtio_gpu_array_unlock_resv(vbuf->objs);
 		free_vbuf(vgdev, vbuf);
-		return;
+		return -1;
 	}
 
 	if (vgdev->has_indirect)
@@ -373,15 +373,16 @@ again:
 	spin_unlock(&vgdev->ctrlq.qlock);
 
 	drm_dev_exit(idx);
+	return 0;
 }
 
-static void virtio_gpu_queue_fenced_ctrl_buffer(struct virtio_gpu_device *vgdev,
-						struct virtio_gpu_vbuffer *vbuf,
-						struct virtio_gpu_fence *fence)
+static int virtio_gpu_queue_fenced_ctrl_buffer(struct virtio_gpu_device *vgdev,
+					       struct virtio_gpu_vbuffer *vbuf,
+					       struct virtio_gpu_fence *fence)
 {
 	struct scatterlist *sgs[3], vcmd, vout, vresp;
 	struct sg_table *sgt = NULL;
-	int elemcnt = 0, outcnt = 0, incnt = 0;
+	int elemcnt = 0, outcnt = 0, incnt = 0, ret;
 
 	/* set up vcmd */
 	sg_init_one(&vcmd, vbuf->buf, vbuf->size);
@@ -398,7 +399,7 @@ static void virtio_gpu_queue_fenced_ctrl_buffer(struct virtio_gpu_device *vgdev,
 			if (!sgt) {
 				if (fence && vbuf->objs)
 					virtio_gpu_array_unlock_resv(vbuf->objs);
-				return;
+				return -1;
 			}
 
 			elemcnt += sg_ents;
@@ -419,13 +420,14 @@ static void virtio_gpu_queue_fenced_ctrl_buffer(struct virtio_gpu_device *vgdev,
 		incnt++;
 	}
 
-	virtio_gpu_queue_ctrl_sgs(vgdev, vbuf, fence, elemcnt, sgs, outcnt,
-				  incnt);
+	ret = virtio_gpu_queue_ctrl_sgs(vgdev, vbuf, fence, elemcnt, sgs, outcnt,
+					incnt);
 
 	if (sgt) {
 		sg_free_table(sgt);
 		kfree(sgt);
 	}
+	return ret;
 }
 
 void virtio_gpu_notify(struct virtio_gpu_device *vgdev)
@@ -444,10 +446,10 @@ void virtio_gpu_notify(struct virtio_gpu_device *vgdev)
 		virtqueue_notify(vgdev->ctrlq.vq);
 }
 
-static void virtio_gpu_queue_ctrl_buffer(struct virtio_gpu_device *vgdev,
-					 struct virtio_gpu_vbuffer *vbuf)
+static int virtio_gpu_queue_ctrl_buffer(struct virtio_gpu_device *vgdev,
+					struct virtio_gpu_vbuffer *vbuf)
 {
-	virtio_gpu_queue_fenced_ctrl_buffer(vgdev, vbuf, NULL);
+	return virtio_gpu_queue_fenced_ctrl_buffer(vgdev, vbuf, NULL);
 }
 
 static void virtio_gpu_queue_cursor(struct virtio_gpu_device *vgdev,
@@ -534,6 +536,7 @@ void virtio_gpu_cmd_unref_resource(struct virtio_gpu_device *vgdev,
 {
 	struct virtio_gpu_resource_unref *cmd_p;
 	struct virtio_gpu_vbuffer *vbuf;
+	int ret;
 
 	cmd_p = virtio_gpu_alloc_cmd_cb(vgdev, &vbuf, sizeof(*cmd_p),
 					virtio_gpu_cmd_unref_cb);
@@ -543,7 +546,9 @@ void virtio_gpu_cmd_unref_resource(struct virtio_gpu_device *vgdev,
 	cmd_p->resource_id = cpu_to_le32(bo->hw_res_handle);
 
 	vbuf->resp_cb_data = bo;
-	virtio_gpu_queue_ctrl_buffer(vgdev, vbuf);
+	ret = virtio_gpu_queue_ctrl_buffer(vgdev, vbuf);
+	if (ret < 0)
+		virtio_gpu_cleanup_object(bo);
 }
 
 void virtio_gpu_cmd_set_scanout(struct virtio_gpu_device *vgdev,
@@ -683,9 +688,13 @@ static void virtio_gpu_cmd_get_capset_info_cb(struct virtio_gpu_device *vgdev,
 	int i = le32_to_cpu(cmd->capset_index);
 
 	spin_lock(&vgdev->display_info_lock);
-	vgdev->capsets[i].id = le32_to_cpu(resp->capset_id);
-	vgdev->capsets[i].max_version = le32_to_cpu(resp->capset_max_version);
-	vgdev->capsets[i].max_size = le32_to_cpu(resp->capset_max_size);
+	if (vgdev->capsets) {
+		vgdev->capsets[i].id = le32_to_cpu(resp->capset_id);
+		vgdev->capsets[i].max_version = le32_to_cpu(resp->capset_max_version);
+		vgdev->capsets[i].max_size = le32_to_cpu(resp->capset_max_size);
+	} else {
+		DRM_ERROR("invalid capset memory.");
+	}
 	spin_unlock(&vgdev->display_info_lock);
 	wake_up(&vgdev->resp_wq);
 }
diff --git a/drivers/gpu/drm/vkms/Makefile b/drivers/gpu/drm/vkms/Makefile
index 0b767d7efa24..333d3cead0e3 100644
--- a/drivers/gpu/drm/vkms/Makefile
+++ b/drivers/gpu/drm/vkms/Makefile
@@ -1,4 +1,11 @@
 # SPDX-License-Identifier: GPL-2.0-only
-vkms-y := vkms_drv.o vkms_plane.o vkms_output.o vkms_crtc.o vkms_gem.o vkms_composer.o
+vkms-y := \
+	vkms_drv.o \
+	vkms_plane.o \
+	vkms_output.o \
+	vkms_crtc.o \
+	vkms_gem.o \
+	vkms_composer.o \
+	vkms_writeback.o
 
 obj-$(CONFIG_DRM_VKMS) += vkms.o
diff --git a/drivers/gpu/drm/vkms/vkms_composer.c b/drivers/gpu/drm/vkms/vkms_composer.c
index eaecc5a6c5db..33c031f27c2c 100644
--- a/drivers/gpu/drm/vkms/vkms_composer.c
+++ b/drivers/gpu/drm/vkms/vkms_composer.c
@@ -9,31 +9,41 @@
 
 #include "vkms_drv.h"
 
+static u32 get_pixel_from_buffer(int x, int y, const u8 *buffer,
+				 const struct vkms_composer *composer)
+{
+	u32 pixel;
+	int src_offset = composer->offset + (y * composer->pitch)
+				      + (x * composer->cpp);
+
+	pixel = *(u32 *)&buffer[src_offset];
+
+	return pixel;
+}
+
 /**
  * compute_crc - Compute CRC value on output frame
  *
- * @vaddr_out: address to final framebuffer
+ * @vaddr: address to final framebuffer
  * @composer: framebuffer's metadata
  *
  * returns CRC value computed using crc32 on the visible portion of
  * the final framebuffer at vaddr_out
  */
-static uint32_t compute_crc(void *vaddr_out, struct vkms_composer *composer)
+static uint32_t compute_crc(const u8 *vaddr,
+			    const struct vkms_composer *composer)
 {
-	int i, j, src_offset;
+	int x, y;
+	u32 crc = 0, pixel = 0;
 	int x_src = composer->src.x1 >> 16;
 	int y_src = composer->src.y1 >> 16;
 	int h_src = drm_rect_height(&composer->src) >> 16;
 	int w_src = drm_rect_width(&composer->src) >> 16;
-	u32 crc = 0;
-
-	for (i = y_src; i < y_src + h_src; ++i) {
-		for (j = x_src; j < x_src + w_src; ++j) {
-			src_offset = composer->offset
-				     + (i * composer->pitch)
-				     + (j * composer->cpp);
-			crc = crc32_le(crc, vaddr_out + src_offset,
-				       sizeof(u32));
+
+	for (y = y_src; y < y_src + h_src; ++y) {
+		for (x = x_src; x < x_src + w_src; ++x) {
+			pixel = get_pixel_from_buffer(x, y, vaddr, composer);
+			crc = crc32_le(crc, (void *)&pixel, sizeof(u32));
 		}
 	}
 
@@ -131,35 +141,31 @@ static void compose_cursor(struct vkms_composer *cursor_composer,
 	      primary_composer, cursor_composer);
 }
 
-static uint32_t _vkms_get_crc(struct vkms_composer *primary_composer,
-			      struct vkms_composer *cursor_composer)
+static int compose_planes(void **vaddr_out,
+			  struct vkms_composer *primary_composer,
+			  struct vkms_composer *cursor_composer)
 {
 	struct drm_framebuffer *fb = &primary_composer->fb;
 	struct drm_gem_object *gem_obj = drm_gem_fb_get_obj(fb, 0);
 	struct vkms_gem_object *vkms_obj = drm_gem_to_vkms_gem(gem_obj);
-	void *vaddr_out = kzalloc(vkms_obj->gem.size, GFP_KERNEL);
-	u32 crc = 0;
 
-	if (!vaddr_out) {
-		DRM_ERROR("Failed to allocate memory for output frame.");
-		return 0;
+	if (!*vaddr_out) {
+		*vaddr_out = kzalloc(vkms_obj->gem.size, GFP_KERNEL);
+		if (!*vaddr_out) {
+			DRM_ERROR("Cannot allocate memory for output frame.");
+			return -ENOMEM;
+		}
 	}
 
-	if (WARN_ON(!vkms_obj->vaddr)) {
-		kfree(vaddr_out);
-		return crc;
-	}
+	if (WARN_ON(!vkms_obj->vaddr))
+		return -EINVAL;
 
-	memcpy(vaddr_out, vkms_obj->vaddr, vkms_obj->gem.size);
+	memcpy(*vaddr_out, vkms_obj->vaddr, vkms_obj->gem.size);
 
 	if (cursor_composer)
-		compose_cursor(cursor_composer, primary_composer, vaddr_out);
-
-	crc = compute_crc(vaddr_out, primary_composer);
-
-	kfree(vaddr_out);
+		compose_cursor(cursor_composer, primary_composer, *vaddr_out);
 
-	return crc;
+	return 0;
 }
 
 /**
@@ -180,14 +186,17 @@ void vkms_composer_worker(struct work_struct *work)
 	struct vkms_output *out = drm_crtc_to_vkms_output(crtc);
 	struct vkms_composer *primary_composer = NULL;
 	struct vkms_composer *cursor_composer = NULL;
+	bool crc_pending, wb_pending;
+	void *vaddr_out = NULL;
 	u32 crc32 = 0;
 	u64 frame_start, frame_end;
-	bool crc_pending;
+	int ret;
 
 	spin_lock_irq(&out->composer_lock);
 	frame_start = crtc_state->frame_start;
 	frame_end = crtc_state->frame_end;
 	crc_pending = crtc_state->crc_pending;
+	wb_pending = crtc_state->wb_pending;
 	crtc_state->frame_start = 0;
 	crtc_state->frame_end = 0;
 	crtc_state->crc_pending = false;
@@ -206,8 +215,29 @@ void vkms_composer_worker(struct work_struct *work)
 	if (crtc_state->num_active_planes == 2)
 		cursor_composer = crtc_state->active_planes[1]->composer;
 
-	if (primary_composer)
-		crc32 = _vkms_get_crc(primary_composer, cursor_composer);
+	if (!primary_composer)
+		return;
+
+	if (wb_pending)
+		vaddr_out = crtc_state->active_writeback;
+
+	ret = compose_planes(&vaddr_out, primary_composer, cursor_composer);
+	if (ret) {
+		if (ret == -EINVAL && !wb_pending)
+			kfree(vaddr_out);
+		return;
+	}
+
+	crc32 = compute_crc(vaddr_out, primary_composer);
+
+	if (wb_pending) {
+		drm_writeback_signal_completion(&out->wb_connector, 0);
+		spin_lock_irq(&out->composer_lock);
+		crtc_state->wb_pending = false;
+		spin_unlock_irq(&out->composer_lock);
+	} else {
+		kfree(vaddr_out);
+	}
 
 	/*
 	 * The worker can fall behind the vblank hrtimer, make sure we catch up.
@@ -256,7 +286,7 @@ int vkms_verify_crc_source(struct drm_crtc *crtc, const char *src_name,
 	return 0;
 }
 
-static void vkms_set_composer(struct vkms_output *out, bool enabled)
+void vkms_set_composer(struct vkms_output *out, bool enabled)
 {
 	bool old_enabled;
 
diff --git a/drivers/gpu/drm/vkms/vkms_drv.c b/drivers/gpu/drm/vkms/vkms_drv.c
index 83dd5567de8b..cb0b6230c22c 100644
--- a/drivers/gpu/drm/vkms/vkms_drv.c
+++ b/drivers/gpu/drm/vkms/vkms_drv.c
@@ -61,9 +61,6 @@ static void vkms_release(struct drm_device *dev)
 {
 	struct vkms_device *vkms = container_of(dev, struct vkms_device, drm);
 
-	platform_device_unregister(vkms->platform);
-	drm_atomic_helper_shutdown(&vkms->drm);
-	drm_mode_config_cleanup(&vkms->drm);
 	destroy_workqueue(vkms->output.composer_workq);
 }
 
@@ -144,30 +141,31 @@ static int vkms_modeset_init(struct vkms_device *vkmsdev)
 static int __init vkms_init(void)
 {
 	int ret;
+	struct platform_device *pdev;
 
-	vkms_device = kzalloc(sizeof(*vkms_device), GFP_KERNEL);
-	if (!vkms_device)
-		return -ENOMEM;
+	pdev = platform_device_register_simple(DRIVER_NAME, -1, NULL, 0);
+	if (IS_ERR(pdev))
+		return PTR_ERR(pdev);
 
-	vkms_device->platform =
-		platform_device_register_simple(DRIVER_NAME, -1, NULL, 0);
-	if (IS_ERR(vkms_device->platform)) {
-		ret = PTR_ERR(vkms_device->platform);
-		goto out_free;
+	if (!devres_open_group(&pdev->dev, NULL, GFP_KERNEL)) {
+		ret = -ENOMEM;
+		goto out_unregister;
 	}
 
-	ret = drm_dev_init(&vkms_device->drm, &vkms_driver,
-			   &vkms_device->platform->dev);
-	if (ret)
-		goto out_unregister;
-	drmm_add_final_kfree(&vkms_device->drm, vkms_device);
+	vkms_device = devm_drm_dev_alloc(&pdev->dev, &vkms_driver,
+					 struct vkms_device, drm);
+	if (IS_ERR(vkms_device)) {
+		ret = PTR_ERR(vkms_device);
+		goto out_devres;
+	}
+	vkms_device->platform = pdev;
 
 	ret = dma_coerce_mask_and_coherent(vkms_device->drm.dev,
 					   DMA_BIT_MASK(64));
 
 	if (ret) {
 		DRM_ERROR("Could not initialize DMA support\n");
-		goto out_put;
+		goto out_devres;
 	}
 
 	vkms_device->drm.irq_enabled = true;
@@ -175,39 +173,41 @@ static int __init vkms_init(void)
 	ret = drm_vblank_init(&vkms_device->drm, 1);
 	if (ret) {
 		DRM_ERROR("Failed to vblank\n");
-		goto out_put;
+		goto out_devres;
 	}
 
 	ret = vkms_modeset_init(vkms_device);
 	if (ret)
-		goto out_put;
+		goto out_devres;
 
 	ret = drm_dev_register(&vkms_device->drm, 0);
 	if (ret)
-		goto out_put;
+		goto out_devres;
 
 	return 0;
 
-out_put:
-	drm_dev_put(&vkms_device->drm);
-	platform_device_unregister(vkms_device->platform);
-	return ret;
+out_devres:
+	devres_release_group(&pdev->dev, NULL);
 out_unregister:
-	platform_device_unregister(vkms_device->platform);
-out_free:
-	kfree(vkms_device);
+	platform_device_unregister(pdev);
 	return ret;
 }
 
 static void __exit vkms_exit(void)
 {
+	struct platform_device *pdev;
+
 	if (!vkms_device) {
 		DRM_INFO("vkms_device is NULL.\n");
 		return;
 	}
 
+	pdev = vkms_device->platform;
+
 	drm_dev_unregister(&vkms_device->drm);
-	drm_dev_put(&vkms_device->drm);
+	drm_atomic_helper_shutdown(&vkms_device->drm);
+	devres_release_group(&pdev->dev, NULL);
+	platform_device_unregister(pdev);
 }
 
 module_init(vkms_init);
diff --git a/drivers/gpu/drm/vkms/vkms_drv.h b/drivers/gpu/drm/vkms/vkms_drv.h
index f4036bb0b9a8..380a8f27e156 100644
--- a/drivers/gpu/drm/vkms/vkms_drv.h
+++ b/drivers/gpu/drm/vkms/vkms_drv.h
@@ -8,6 +8,7 @@
 #include <drm/drm.h>
 #include <drm/drm_gem.h>
 #include <drm/drm_encoder.h>
+#include <drm/drm_writeback.h>
 
 #define XRES_MIN    20
 #define YRES_MIN    20
@@ -52,9 +53,11 @@ struct vkms_crtc_state {
 	int num_active_planes;
 	/* stack of active planes for crc computation, should be in z order */
 	struct vkms_plane_state **active_planes;
+	void *active_writeback;
 
-	/* below three are protected by vkms_output.composer_lock */
+	/* below four are protected by vkms_output.composer_lock */
 	bool crc_pending;
+	bool wb_pending;
 	u64 frame_start;
 	u64 frame_end;
 };
@@ -63,6 +66,7 @@ struct vkms_output {
 	struct drm_crtc crtc;
 	struct drm_encoder encoder;
 	struct drm_connector connector;
+	struct drm_writeback_connector wb_connector;
 	struct hrtimer vblank_hrtimer;
 	ktime_t period_ns;
 	struct drm_pending_vblank_event *event;
@@ -143,5 +147,9 @@ int vkms_verify_crc_source(struct drm_crtc *crtc, const char *source_name,
 
 /* Composer Support */
 void vkms_composer_worker(struct work_struct *work);
+void vkms_set_composer(struct vkms_output *out, bool enabled);
+
+/* Writeback */
+int vkms_enable_writeback_connector(struct vkms_device *vkmsdev);
 
 #endif /* _VKMS_DRV_H_ */
diff --git a/drivers/gpu/drm/vkms/vkms_output.c b/drivers/gpu/drm/vkms/vkms_output.c
index 85afb77e97f0..4a1848b0318f 100644
--- a/drivers/gpu/drm/vkms/vkms_output.c
+++ b/drivers/gpu/drm/vkms/vkms_output.c
@@ -80,6 +80,10 @@ int vkms_output_init(struct vkms_device *vkmsdev, int index)
 		goto err_attach;
 	}
 
+	ret = vkms_enable_writeback_connector(vkmsdev);
+	if (ret)
+		DRM_ERROR("Failed to init writeback connector\n");
+
 	drm_mode_config_reset(dev);
 
 	return 0;
diff --git a/drivers/gpu/drm/vkms/vkms_writeback.c b/drivers/gpu/drm/vkms/vkms_writeback.c
new file mode 100644
index 000000000000..094fa4aa061d
--- /dev/null
+++ b/drivers/gpu/drm/vkms/vkms_writeback.c
@@ -0,0 +1,142 @@
+// SPDX-License-Identifier: GPL-2.0+
+
+#include "vkms_drv.h"
+#include <drm/drm_fourcc.h>
+#include <drm/drm_writeback.h>
+#include <drm/drm_probe_helper.h>
+#include <drm/drm_atomic_helper.h>
+#include <drm/drm_gem_framebuffer_helper.h>
+
+static const u32 vkms_wb_formats[] = {
+	DRM_FORMAT_XRGB8888,
+};
+
+static const struct drm_connector_funcs vkms_wb_connector_funcs = {
+	.fill_modes = drm_helper_probe_single_connector_modes,
+	.destroy = drm_connector_cleanup,
+	.reset = drm_atomic_helper_connector_reset,
+	.atomic_duplicate_state = drm_atomic_helper_connector_duplicate_state,
+	.atomic_destroy_state = drm_atomic_helper_connector_destroy_state,
+};
+
+static int vkms_wb_encoder_atomic_check(struct drm_encoder *encoder,
+					struct drm_crtc_state *crtc_state,
+					struct drm_connector_state *conn_state)
+{
+	struct drm_framebuffer *fb;
+	const struct drm_display_mode *mode = &crtc_state->mode;
+
+	if (!conn_state->writeback_job || !conn_state->writeback_job->fb)
+		return 0;
+
+	fb = conn_state->writeback_job->fb;
+	if (fb->width != mode->hdisplay || fb->height != mode->vdisplay) {
+		DRM_DEBUG_KMS("Invalid framebuffer size %ux%u\n",
+			      fb->width, fb->height);
+		return -EINVAL;
+	}
+
+	if (fb->format->format != vkms_wb_formats[0]) {
+		struct drm_format_name_buf format_name;
+
+		DRM_DEBUG_KMS("Invalid pixel format %s\n",
+			      drm_get_format_name(fb->format->format,
+						  &format_name));
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static const struct drm_encoder_helper_funcs vkms_wb_encoder_helper_funcs = {
+	.atomic_check = vkms_wb_encoder_atomic_check,
+};
+
+static int vkms_wb_connector_get_modes(struct drm_connector *connector)
+{
+	struct drm_device *dev = connector->dev;
+
+	return drm_add_modes_noedid(connector, dev->mode_config.max_width,
+				    dev->mode_config.max_height);
+}
+
+static int vkms_wb_prepare_job(struct drm_writeback_connector *wb_connector,
+			       struct drm_writeback_job *job)
+{
+	struct vkms_gem_object *vkms_obj;
+	struct drm_gem_object *gem_obj;
+	int ret;
+
+	if (!job->fb)
+		return 0;
+
+	gem_obj = drm_gem_fb_get_obj(job->fb, 0);
+	ret = vkms_gem_vmap(gem_obj);
+	if (ret) {
+		DRM_ERROR("vmap failed: %d\n", ret);
+		return ret;
+	}
+
+	vkms_obj = drm_gem_to_vkms_gem(gem_obj);
+	job->priv = vkms_obj->vaddr;
+
+	return 0;
+}
+
+static void vkms_wb_cleanup_job(struct drm_writeback_connector *connector,
+				struct drm_writeback_job *job)
+{
+	struct drm_gem_object *gem_obj;
+	struct vkms_device *vkmsdev;
+
+	if (!job->fb)
+		return;
+
+	gem_obj = drm_gem_fb_get_obj(job->fb, 0);
+	vkms_gem_vunmap(gem_obj);
+
+	vkmsdev = drm_device_to_vkms_device(gem_obj->dev);
+	vkms_set_composer(&vkmsdev->output, false);
+}
+
+static void vkms_wb_atomic_commit(struct drm_connector *conn,
+				  struct drm_connector_state *state)
+{
+	struct vkms_device *vkmsdev = drm_device_to_vkms_device(conn->dev);
+	struct vkms_output *output = &vkmsdev->output;
+	struct drm_writeback_connector *wb_conn = &output->wb_connector;
+	struct drm_connector_state *conn_state = wb_conn->base.state;
+	struct vkms_crtc_state *crtc_state = output->composer_state;
+
+	if (!conn_state)
+		return;
+
+	vkms_set_composer(&vkmsdev->output, true);
+
+	spin_lock_irq(&output->composer_lock);
+	crtc_state->active_writeback = conn_state->writeback_job->priv;
+	crtc_state->wb_pending = true;
+	spin_unlock_irq(&output->composer_lock);
+	drm_writeback_queue_job(wb_conn, state);
+}
+
+static const struct drm_connector_helper_funcs vkms_wb_conn_helper_funcs = {
+	.get_modes = vkms_wb_connector_get_modes,
+	.prepare_writeback_job = vkms_wb_prepare_job,
+	.cleanup_writeback_job = vkms_wb_cleanup_job,
+	.atomic_commit = vkms_wb_atomic_commit,
+};
+
+int vkms_enable_writeback_connector(struct vkms_device *vkmsdev)
+{
+	struct drm_writeback_connector *wb = &vkmsdev->output.wb_connector;
+
+	vkmsdev->output.wb_connector.encoder.possible_crtcs = 1;
+	drm_connector_helper_add(&wb->base, &vkms_wb_conn_helper_funcs);
+
+	return drm_writeback_connector_init(&vkmsdev->drm, wb,
+					    &vkms_wb_connector_funcs,
+					    &vkms_wb_encoder_helper_funcs,
+					    vkms_wb_formats,
+					    ARRAY_SIZE(vkms_wb_formats));
+}
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_blit.c b/drivers/gpu/drm/vmwgfx/vmwgfx_blit.c
index 1629427d5734..e8d66182cd7b 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_blit.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_blit.c
@@ -464,14 +464,14 @@ int vmw_bo_cpu_blit(struct ttm_buffer_object *dst,
 	if (!(src->mem.placement & TTM_PL_FLAG_NO_EVICT))
 		dma_resv_assert_held(src->base.resv);
 
-	if (dst->ttm->state == tt_unpopulated) {
-		ret = dst->ttm->bdev->driver->ttm_tt_populate(dst->ttm, &ctx);
+	if (!ttm_tt_is_populated(dst->ttm)) {
+		ret = dst->bdev->driver->ttm_tt_populate(dst->bdev, dst->ttm, &ctx);
 		if (ret)
 			return ret;
 	}
 
-	if (src->ttm->state == tt_unpopulated) {
-		ret = src->ttm->bdev->driver->ttm_tt_populate(src->ttm, &ctx);
+	if (!ttm_tt_is_populated(src->ttm)) {
+		ret = src->bdev->driver->ttm_tt_populate(src->bdev, src->ttm, &ctx);
 		if (ret)
 			return ret;
 	}
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_bo.c b/drivers/gpu/drm/vmwgfx/vmwgfx_bo.c
index 3229451d0706..813f1b148094 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_bo.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_bo.c
@@ -354,10 +354,12 @@ void vmw_bo_pin_reserved(struct vmw_buffer_object *vbo, bool pin)
 
 	pl.fpfn = 0;
 	pl.lpfn = 0;
-	pl.flags = TTM_PL_FLAG_VRAM | VMW_PL_FLAG_GMR | VMW_PL_FLAG_MOB
-		| TTM_PL_FLAG_SYSTEM | TTM_PL_FLAG_CACHED;
+	pl.mem_type = bo->mem.mem_type;
+	pl.flags = bo->mem.placement;
 	if (pin)
 		pl.flags |= TTM_PL_FLAG_NO_EVICT;
+	else
+		pl.flags &= ~TTM_PL_FLAG_NO_EVICT;
 
 	memset(&placement, 0, sizeof(placement));
 	placement.num_placement = 1;
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c
index e60012886065..31e3e5c9f362 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c
@@ -626,9 +626,8 @@ static int vmw_vram_manager_init(struct vmw_private *dev_priv)
 #ifdef CONFIG_TRANSPARENT_HUGEPAGE
 	ret = vmw_thp_init(dev_priv);
 #else
-	ret = ttm_range_man_init(&dev_priv->bdev, TTM_PL_VRAM,
-				 TTM_PL_FLAG_CACHED, TTM_PL_FLAG_CACHED,
-				 false, dev_priv->vram_size >> PAGE_SHIFT);
+	ret = ttm_range_man_init(&dev_priv->bdev, TTM_PL_VRAM, false,
+				 dev_priv->vram_size >> PAGE_SHIFT);
 #endif
 	ttm_resource_manager_set_used(ttm_manager_type(&dev_priv->bdev, TTM_PL_VRAM), false);
 	return ret;
@@ -882,8 +881,6 @@ static int vmw_driver_load(struct drm_device *dev, unsigned long chipset)
 		DRM_ERROR("Failed initializing TTM buffer object driver.\n");
 		goto out_no_bdev;
 	}
-	ttm_manager_type(&dev_priv->bdev, TTM_PL_SYSTEM)->available_caching =
-		TTM_PL_FLAG_CACHED;
 
 	/*
 	 * Enable VRAM, but initially don't use it until SVGA is enabled and
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h
index 871ad738dadb..1523b51a7284 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h
@@ -82,9 +82,7 @@
 			VMWGFX_NUM_GB_SCREEN_TARGET)
 
 #define VMW_PL_GMR (TTM_PL_PRIV + 0)
-#define VMW_PL_FLAG_GMR (TTM_PL_FLAG_PRIV << 0)
 #define VMW_PL_MOB (TTM_PL_PRIV + 1)
-#define VMW_PL_FLAG_MOB (TTM_PL_FLAG_PRIV << 1)
 
 #define VMW_RES_CONTEXT ttm_driver_type0
 #define VMW_RES_SURFACE ttm_driver_type1
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_gmrid_manager.c b/drivers/gpu/drm/vmwgfx/vmwgfx_gmrid_manager.c
index bb76acb5b0fc..db64c3a90285 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_gmrid_manager.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_gmrid_manager.c
@@ -112,8 +112,6 @@ int vmw_gmrid_man_init(struct vmw_private *dev_priv, int type)
 	man = &gman->manager;
 
 	man->func = &vmw_gmrid_manager_func;
-	man->available_caching = TTM_PL_FLAG_CACHED;
-	man->default_caching = TTM_PL_FLAG_CACHED;
 	/* TODO: This is most likely not correct */
 	man->use_tt = true;
 	ttm_resource_manager_init(man, 0);
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_thp.c b/drivers/gpu/drm/vmwgfx/vmwgfx_thp.c
index 3c00a9e7cfcc..63fe7da4cbf4 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_thp.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_thp.c
@@ -123,25 +123,20 @@ static void vmw_thp_put_node(struct ttm_resource_manager *man,
 
 int vmw_thp_init(struct vmw_private *dev_priv)
 {
-	struct ttm_resource_manager *man;
 	struct vmw_thp_manager *rman;
 
 	rman = kzalloc(sizeof(*rman), GFP_KERNEL);
 	if (!rman)
 		return -ENOMEM;
 
-	man = &rman->manager;
-	man->available_caching = TTM_PL_FLAG_CACHED;
-	man->default_caching = TTM_PL_FLAG_CACHED;
-
-	ttm_resource_manager_init(man,
+	ttm_resource_manager_init(&rman->manager,
 				  dev_priv->vram_size >> PAGE_SHIFT);
 
-	drm_mm_init(&rman->mm, 0, man->size);
+	drm_mm_init(&rman->mm, 0, rman->manager.size);
 	spin_lock_init(&rman->lock);
 
 	ttm_set_driver_manager(&dev_priv->bdev, TTM_PL_VRAM, &rman->manager);
-	ttm_resource_manager_set_used(man, true);
+	ttm_resource_manager_set_used(&rman->manager, true);
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_ttm_buffer.c b/drivers/gpu/drm/vmwgfx/vmwgfx_ttm_buffer.c
index 13c31e2d7254..7f0310441da1 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_ttm_buffer.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_ttm_buffer.c
@@ -33,49 +33,57 @@
 static const struct ttm_place vram_placement_flags = {
 	.fpfn = 0,
 	.lpfn = 0,
-	.flags = TTM_PL_FLAG_VRAM | TTM_PL_FLAG_CACHED
+	.mem_type = TTM_PL_VRAM,
+	.flags = TTM_PL_FLAG_CACHED
 };
 
 static const struct ttm_place vram_ne_placement_flags = {
 	.fpfn = 0,
 	.lpfn = 0,
-	.flags = TTM_PL_FLAG_VRAM | TTM_PL_FLAG_CACHED | TTM_PL_FLAG_NO_EVICT
+	.mem_type = TTM_PL_VRAM,
+	.flags = TTM_PL_FLAG_CACHED | TTM_PL_FLAG_NO_EVICT
 };
 
 static const struct ttm_place sys_placement_flags = {
 	.fpfn = 0,
 	.lpfn = 0,
-	.flags = TTM_PL_FLAG_SYSTEM | TTM_PL_FLAG_CACHED
+	.mem_type = TTM_PL_SYSTEM,
+	.flags = TTM_PL_FLAG_CACHED
 };
 
 static const struct ttm_place sys_ne_placement_flags = {
 	.fpfn = 0,
 	.lpfn = 0,
-	.flags = TTM_PL_FLAG_SYSTEM | TTM_PL_FLAG_CACHED | TTM_PL_FLAG_NO_EVICT
+	.mem_type = TTM_PL_SYSTEM,
+	.flags = TTM_PL_FLAG_CACHED | TTM_PL_FLAG_NO_EVICT
 };
 
 static const struct ttm_place gmr_placement_flags = {
 	.fpfn = 0,
 	.lpfn = 0,
-	.flags = VMW_PL_FLAG_GMR | TTM_PL_FLAG_CACHED
+	.mem_type = VMW_PL_GMR,
+	.flags = TTM_PL_FLAG_CACHED
 };
 
 static const struct ttm_place gmr_ne_placement_flags = {
 	.fpfn = 0,
 	.lpfn = 0,
-	.flags = VMW_PL_FLAG_GMR | TTM_PL_FLAG_CACHED | TTM_PL_FLAG_NO_EVICT
+	.mem_type = VMW_PL_GMR,
+	.flags = TTM_PL_FLAG_CACHED | TTM_PL_FLAG_NO_EVICT
 };
 
 static const struct ttm_place mob_placement_flags = {
 	.fpfn = 0,
 	.lpfn = 0,
-	.flags = VMW_PL_FLAG_MOB | TTM_PL_FLAG_CACHED
+	.mem_type = VMW_PL_MOB,
+	.flags = TTM_PL_FLAG_CACHED
 };
 
 static const struct ttm_place mob_ne_placement_flags = {
 	.fpfn = 0,
 	.lpfn = 0,
-	.flags = VMW_PL_FLAG_MOB | TTM_PL_FLAG_CACHED | TTM_PL_FLAG_NO_EVICT
+	.mem_type = VMW_PL_MOB,
+	.flags = TTM_PL_FLAG_CACHED | TTM_PL_FLAG_NO_EVICT
 };
 
 struct ttm_placement vmw_vram_placement = {
@@ -89,11 +97,13 @@ static const struct ttm_place vram_gmr_placement_flags[] = {
 	{
 		.fpfn = 0,
 		.lpfn = 0,
-		.flags = TTM_PL_FLAG_VRAM | TTM_PL_FLAG_CACHED
+		.mem_type = TTM_PL_VRAM,
+		.flags = TTM_PL_FLAG_CACHED
 	}, {
 		.fpfn = 0,
 		.lpfn = 0,
-		.flags = VMW_PL_FLAG_GMR | TTM_PL_FLAG_CACHED
+		.mem_type = VMW_PL_GMR,
+		.flags = TTM_PL_FLAG_CACHED
 	}
 };
 
@@ -101,11 +111,13 @@ static const struct ttm_place gmr_vram_placement_flags[] = {
 	{
 		.fpfn = 0,
 		.lpfn = 0,
-		.flags = VMW_PL_FLAG_GMR | TTM_PL_FLAG_CACHED
+		.mem_type = VMW_PL_GMR,
+		.flags = TTM_PL_FLAG_CACHED
 	}, {
 		.fpfn = 0,
 		.lpfn = 0,
-		.flags = TTM_PL_FLAG_VRAM | TTM_PL_FLAG_CACHED
+		.mem_type = TTM_PL_VRAM,
+		.flags = TTM_PL_FLAG_CACHED
 	}
 };
 
@@ -120,12 +132,14 @@ static const struct ttm_place vram_gmr_ne_placement_flags[] = {
 	{
 		.fpfn = 0,
 		.lpfn = 0,
-		.flags = TTM_PL_FLAG_VRAM | TTM_PL_FLAG_CACHED |
+		.mem_type = TTM_PL_VRAM,
+		.flags = TTM_PL_FLAG_CACHED |
 			 TTM_PL_FLAG_NO_EVICT
 	}, {
 		.fpfn = 0,
 		.lpfn = 0,
-		.flags = VMW_PL_FLAG_GMR | TTM_PL_FLAG_CACHED |
+		.mem_type = VMW_PL_GMR,
+		.flags = TTM_PL_FLAG_CACHED |
 			 TTM_PL_FLAG_NO_EVICT
 	}
 };
@@ -169,19 +183,23 @@ static const struct ttm_place evictable_placement_flags[] = {
 	{
 		.fpfn = 0,
 		.lpfn = 0,
-		.flags = TTM_PL_FLAG_SYSTEM | TTM_PL_FLAG_CACHED
+		.mem_type = TTM_PL_SYSTEM,
+		.flags = TTM_PL_FLAG_CACHED
 	}, {
 		.fpfn = 0,
 		.lpfn = 0,
-		.flags = TTM_PL_FLAG_VRAM | TTM_PL_FLAG_CACHED
+		.mem_type = TTM_PL_VRAM,
+		.flags = TTM_PL_FLAG_CACHED
 	}, {
 		.fpfn = 0,
 		.lpfn = 0,
-		.flags = VMW_PL_FLAG_GMR | TTM_PL_FLAG_CACHED
+		.mem_type = VMW_PL_GMR,
+		.flags = TTM_PL_FLAG_CACHED
 	}, {
 		.fpfn = 0,
 		.lpfn = 0,
-		.flags = VMW_PL_FLAG_MOB | TTM_PL_FLAG_CACHED
+		.mem_type = VMW_PL_MOB,
+		.flags = TTM_PL_FLAG_CACHED
 	}
 };
 
@@ -189,15 +207,18 @@ static const struct ttm_place nonfixed_placement_flags[] = {
 	{
 		.fpfn = 0,
 		.lpfn = 0,
-		.flags = TTM_PL_FLAG_SYSTEM | TTM_PL_FLAG_CACHED
+		.mem_type = TTM_PL_SYSTEM,
+		.flags = TTM_PL_FLAG_CACHED
 	}, {
 		.fpfn = 0,
 		.lpfn = 0,
-		.flags = VMW_PL_FLAG_GMR | TTM_PL_FLAG_CACHED
+		.mem_type = VMW_PL_GMR,
+		.flags = TTM_PL_FLAG_CACHED
 	}, {
 		.fpfn = 0,
 		.lpfn = 0,
-		.flags = VMW_PL_FLAG_MOB | TTM_PL_FLAG_CACHED
+		.mem_type = VMW_PL_MOB,
+		.flags = TTM_PL_FLAG_CACHED
 	}
 };
 
@@ -246,6 +267,7 @@ struct vmw_ttm_tt {
 	struct vmw_sg_table vsgt;
 	uint64_t sg_alloc_size;
 	bool mapped;
+	bool bound;
 };
 
 const size_t vmw_tt_size = sizeof(struct vmw_ttm_tt);
@@ -530,11 +552,18 @@ const struct vmw_sg_table *vmw_bo_sg_table(struct ttm_buffer_object *bo)
 }
 
 
-static int vmw_ttm_bind(struct ttm_tt *ttm, struct ttm_resource *bo_mem)
+static int vmw_ttm_bind(struct ttm_bo_device *bdev,
+			struct ttm_tt *ttm, struct ttm_resource *bo_mem)
 {
 	struct vmw_ttm_tt *vmw_be =
 		container_of(ttm, struct vmw_ttm_tt, dma_ttm.ttm);
-	int ret;
+	int ret = 0;
+
+	if (!bo_mem)
+		return -EINVAL;
+
+	if (vmw_be->bound)
+		return 0;
 
 	ret = vmw_ttm_map_dma(vmw_be);
 	if (unlikely(ret != 0))
@@ -545,8 +574,9 @@ static int vmw_ttm_bind(struct ttm_tt *ttm, struct ttm_resource *bo_mem)
 
 	switch (bo_mem->mem_type) {
 	case VMW_PL_GMR:
-		return vmw_gmr_bind(vmw_be->dev_priv, &vmw_be->vsgt,
+		ret = vmw_gmr_bind(vmw_be->dev_priv, &vmw_be->vsgt,
 				    ttm->num_pages, vmw_be->gmr_id);
+		break;
 	case VMW_PL_MOB:
 		if (unlikely(vmw_be->mob == NULL)) {
 			vmw_be->mob =
@@ -555,20 +585,26 @@ static int vmw_ttm_bind(struct ttm_tt *ttm, struct ttm_resource *bo_mem)
 				return -ENOMEM;
 		}
 
-		return vmw_mob_bind(vmw_be->dev_priv, vmw_be->mob,
+		ret = vmw_mob_bind(vmw_be->dev_priv, vmw_be->mob,
 				    &vmw_be->vsgt, ttm->num_pages,
 				    vmw_be->gmr_id);
+		break;
 	default:
 		BUG();
 	}
-	return 0;
+	vmw_be->bound = true;
+	return ret;
 }
 
-static void vmw_ttm_unbind(struct ttm_tt *ttm)
+static void vmw_ttm_unbind(struct ttm_bo_device *bdev,
+			   struct ttm_tt *ttm)
 {
 	struct vmw_ttm_tt *vmw_be =
 		container_of(ttm, struct vmw_ttm_tt, dma_ttm.ttm);
 
+	if (!vmw_be->bound)
+		return;
+
 	switch (vmw_be->mem_type) {
 	case VMW_PL_GMR:
 		vmw_gmr_unbind(vmw_be->dev_priv, vmw_be->gmr_id);
@@ -582,14 +618,17 @@ static void vmw_ttm_unbind(struct ttm_tt *ttm)
 
 	if (vmw_be->dev_priv->map_mode == vmw_dma_map_bind)
 		vmw_ttm_unmap_dma(vmw_be);
+	vmw_be->bound = false;
 }
 
 
-static void vmw_ttm_destroy(struct ttm_tt *ttm)
+static void vmw_ttm_destroy(struct ttm_bo_device *bdev, struct ttm_tt *ttm)
 {
 	struct vmw_ttm_tt *vmw_be =
 		container_of(ttm, struct vmw_ttm_tt, dma_ttm.ttm);
 
+	vmw_ttm_unbind(bdev, ttm);
+	ttm_tt_destroy_common(bdev, ttm);
 	vmw_ttm_unmap_dma(vmw_be);
 	if (vmw_be->dev_priv->map_mode == vmw_dma_alloc_coherent)
 		ttm_dma_tt_fini(&vmw_be->dma_ttm);
@@ -603,7 +642,8 @@ static void vmw_ttm_destroy(struct ttm_tt *ttm)
 }
 
 
-static int vmw_ttm_populate(struct ttm_tt *ttm, struct ttm_operation_ctx *ctx)
+static int vmw_ttm_populate(struct ttm_bo_device *bdev,
+			    struct ttm_tt *ttm, struct ttm_operation_ctx *ctx)
 {
 	struct vmw_ttm_tt *vmw_tt =
 		container_of(ttm, struct vmw_ttm_tt, dma_ttm.ttm);
@@ -611,7 +651,7 @@ static int vmw_ttm_populate(struct ttm_tt *ttm, struct ttm_operation_ctx *ctx)
 	struct ttm_mem_global *glob = vmw_mem_glob(dev_priv);
 	int ret;
 
-	if (ttm->state != tt_unpopulated)
+	if (ttm_tt_is_populated(ttm))
 		return 0;
 
 	if (dev_priv->map_mode == vmw_dma_alloc_coherent) {
@@ -631,7 +671,8 @@ static int vmw_ttm_populate(struct ttm_tt *ttm, struct ttm_operation_ctx *ctx)
 	return ret;
 }
 
-static void vmw_ttm_unpopulate(struct ttm_tt *ttm)
+static void vmw_ttm_unpopulate(struct ttm_bo_device *bdev,
+			       struct ttm_tt *ttm)
 {
 	struct vmw_ttm_tt *vmw_tt = container_of(ttm, struct vmw_ttm_tt,
 						 dma_ttm.ttm);
@@ -655,12 +696,6 @@ static void vmw_ttm_unpopulate(struct ttm_tt *ttm)
 		ttm_pool_unpopulate(ttm);
 }
 
-static struct ttm_backend_func vmw_ttm_func = {
-	.bind = vmw_ttm_bind,
-	.unbind = vmw_ttm_unbind,
-	.destroy = vmw_ttm_destroy,
-};
-
 static struct ttm_tt *vmw_ttm_tt_create(struct ttm_buffer_object *bo,
 					uint32_t page_flags)
 {
@@ -671,7 +706,6 @@ static struct ttm_tt *vmw_ttm_tt_create(struct ttm_buffer_object *bo,
 	if (!vmw_be)
 		return NULL;
 
-	vmw_be->dma_ttm.ttm.func = &vmw_ttm_func;
 	vmw_be->dev_priv = container_of(bo->bdev, struct vmw_private, bdev);
 	vmw_be->mob = NULL;
 
@@ -712,8 +746,8 @@ static int vmw_ttm_io_mem_reserve(struct ttm_bo_device *bdev, struct ttm_resourc
 	case VMW_PL_MOB:
 		return 0;
 	case TTM_PL_VRAM:
-		mem->bus.offset = mem->start << PAGE_SHIFT;
-		mem->bus.base = dev_priv->vram_start;
+		mem->bus.offset = (mem->start << PAGE_SHIFT) +
+			dev_priv->vram_start;
 		mem->bus.is_iomem = true;
 		break;
 	default:
@@ -757,6 +791,9 @@ struct ttm_bo_driver vmw_bo_driver = {
 	.ttm_tt_create = &vmw_ttm_tt_create,
 	.ttm_tt_populate = &vmw_ttm_populate,
 	.ttm_tt_unpopulate = &vmw_ttm_unpopulate,
+	.ttm_tt_bind = &vmw_ttm_bind,
+	.ttm_tt_unbind = &vmw_ttm_unbind,
+	.ttm_tt_destroy = &vmw_ttm_destroy,
 	.eviction_valuable = ttm_bo_eviction_valuable,
 	.evict_flags = vmw_evict_flags,
 	.move = NULL,
@@ -787,7 +824,7 @@ int vmw_bo_create_and_populate(struct vmw_private *dev_priv,
 
 	ret = ttm_bo_reserve(bo, false, true, NULL);
 	BUG_ON(ret != 0);
-	ret = vmw_ttm_populate(bo->ttm, &ctx);
+	ret = vmw_ttm_populate(bo->bdev, bo->ttm, &ctx);
 	if (likely(ret == 0)) {
 		struct vmw_ttm_tt *vmw_tt =
 			container_of(bo->ttm, struct vmw_ttm_tt, dma_ttm.ttm);
diff --git a/drivers/gpu/drm/xen/xen_drm_front_gem.c b/drivers/gpu/drm/xen/xen_drm_front_gem.c
index a487384d5cb7..2f464ef2d53e 100644
--- a/drivers/gpu/drm/xen/xen_drm_front_gem.c
+++ b/drivers/gpu/drm/xen/xen_drm_front_gem.c
@@ -180,7 +180,8 @@ struct sg_table *xen_drm_front_gem_get_sg_table(struct drm_gem_object *gem_obj)
 	if (!xen_obj->pages)
 		return ERR_PTR(-ENOMEM);
 
-	return drm_prime_pages_to_sg(xen_obj->pages, xen_obj->num_pages);
+	return drm_prime_pages_to_sg(gem_obj->dev,
+				     xen_obj->pages, xen_obj->num_pages);
 }
 
 struct drm_gem_object *
diff --git a/drivers/gpu/drm/xlnx/zynqmp_disp.c b/drivers/gpu/drm/xlnx/zynqmp_disp.c
index a455cfc1bee5..98bd48f13fd1 100644
--- a/drivers/gpu/drm/xlnx/zynqmp_disp.c
+++ b/drivers/gpu/drm/xlnx/zynqmp_disp.c
@@ -242,12 +242,6 @@ static const u32 scaling_factors_565[] = {
 	ZYNQMP_DISP_AV_BUF_5BIT_SF,
 };
 
-static const u32 scaling_factors_666[] = {
-	ZYNQMP_DISP_AV_BUF_6BIT_SF,
-	ZYNQMP_DISP_AV_BUF_6BIT_SF,
-	ZYNQMP_DISP_AV_BUF_6BIT_SF,
-};
-
 static const u32 scaling_factors_888[] = {
 	ZYNQMP_DISP_AV_BUF_8BIT_SF,
 	ZYNQMP_DISP_AV_BUF_8BIT_SF,
diff --git a/drivers/gpu/drm/xlnx/zynqmp_dpsub.c b/drivers/gpu/drm/xlnx/zynqmp_dpsub.c
index 26328c76305b..8e69303aad3f 100644
--- a/drivers/gpu/drm/xlnx/zynqmp_dpsub.c
+++ b/drivers/gpu/drm/xlnx/zynqmp_dpsub.c
@@ -111,7 +111,7 @@ static int zynqmp_dpsub_drm_init(struct zynqmp_dpsub *dpsub)
 	/* Initialize mode config, vblank and the KMS poll helper. */
 	ret = drmm_mode_config_init(drm);
 	if (ret < 0)
-		goto err_dev_put;
+		return ret;
 
 	drm->mode_config.funcs = &zynqmp_dpsub_mode_config_funcs;
 	drm->mode_config.min_width = 0;
@@ -121,7 +121,7 @@ static int zynqmp_dpsub_drm_init(struct zynqmp_dpsub *dpsub)
 
 	ret = drm_vblank_init(drm, 1);
 	if (ret)
-		goto err_dev_put;
+		return ret;
 
 	drm->irq_enabled = 1;
 
@@ -154,8 +154,6 @@ static int zynqmp_dpsub_drm_init(struct zynqmp_dpsub *dpsub)
 
 err_poll_fini:
 	drm_kms_helper_poll_fini(drm);
-err_dev_put:
-	drm_dev_put(drm);
 	return ret;
 }
 
@@ -208,27 +206,16 @@ static int zynqmp_dpsub_probe(struct platform_device *pdev)
 	int ret;
 
 	/* Allocate private data. */
-	dpsub = kzalloc(sizeof(*dpsub), GFP_KERNEL);
-	if (!dpsub)
-		return -ENOMEM;
+	dpsub = devm_drm_dev_alloc(&pdev->dev, &zynqmp_dpsub_drm_driver,
+				   struct zynqmp_dpsub, drm);
+	if (IS_ERR(dpsub))
+		return PTR_ERR(dpsub);
 
 	dpsub->dev = &pdev->dev;
 	platform_set_drvdata(pdev, dpsub);
 
 	dma_set_mask(dpsub->dev, DMA_BIT_MASK(ZYNQMP_DISP_MAX_DMA_BIT));
 
-	/*
-	 * Initialize the DRM device early, as the DRM core mandates usage of
-	 * the managed memory helpers tied to the DRM device.
-	 */
-	ret = drm_dev_init(&dpsub->drm, &zynqmp_dpsub_drm_driver, &pdev->dev);
-	if (ret < 0) {
-		kfree(dpsub);
-		return ret;
-	}
-
-	drmm_add_final_kfree(&dpsub->drm, dpsub);
-
 	/* Try the reserved memory. Proceed if there's none. */
 	of_reserved_mem_device_init(&pdev->dev);
 
@@ -286,8 +273,6 @@ static int zynqmp_dpsub_remove(struct platform_device *pdev)
 	clk_disable_unprepare(dpsub->apb_clk);
 	of_reserved_mem_device_release(&pdev->dev);
 
-	drm_dev_put(drm);
-
 	return 0;
 }