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

Pull drm updates from Dave Airlie:
 "This is the main drm pull request for 3.10.

  Wierd bits:
   - OMAP drm changes required OMAP dss changes, in drivers/video, so I
     took them in here.
   - one more fbcon fix for font handover
   - VT switch avoidance in pm code
   - scatterlist helpers for gpu drivers - have acks from akpm

  Highlights:
   - qxl kms driver - driver for the spice qxl virtual GPU

  Nouveau:
   - fermi/kepler VRAM compression
   - GK110/nvf0 modesetting support.

  Tegra:
   - host1x core merged with 2D engine support

  i915:
   - vt switchless resume
   - more valleyview support
   - vblank fixes
   - modesetting pipe config rework

  radeon:
   - UVD engine support
   - SI chip tiling support
   - GPU registers initialisation from golden values.

  exynos:
   - device tree changes
   - fimc block support

  Otherwise:
   - bunches of fixes all over the place."

* 'drm-next' of git://people.freedesktop.org/~airlied/linux: (513 commits)
  qxl: update to new idr interfaces.
  drm/nouveau: fix build with nv50->nvc0
  drm/radeon: fix handling of v6 power tables
  drm/radeon: clarify family checks in pm table parsing
  drm/radeon: consolidate UVD clock programming
  drm/radeon: fix UPLL_REF_DIV_MASK definition
  radeon: add bo tracking debugfs
  drm/radeon: add new richland pci ids
  drm/radeon: add some new SI PCI ids
  drm/radeon: fix scratch reg handling for UVD fence
  drm/radeon: allocate SA bo in the requested domain
  drm/radeon: fix possible segfault when parsing pm tables
  drm/radeon: fix endian bugs in atom_allocate_fb_scratch()
  OMAPDSS: TFP410: return EPROBE_DEFER if the i2c adapter not found
  OMAPDSS: VENC: Add error handling for venc_probe_pdata
  OMAPDSS: HDMI: Add error handling for hdmi_probe_pdata
  OMAPDSS: RFBI: Add error handling for rfbi_probe_pdata
  OMAPDSS: DSI: Add error handling for dsi_probe_pdata
  OMAPDSS: SDI: Add error handling for sdi_probe_pdata
  OMAPDSS: DPI: Add error handling for dpi_probe_pdata
  ...

333 files changed, 26522 insertions, 5467 deletions

diff --git a/drivers/gpu/Makefile b/drivers/gpu/Makefile
index 30879df3daea..d8a22c2a579d 100644
--- a/drivers/gpu/Makefile
+++ b/drivers/gpu/Makefile
@@ -1 +1,2 @@
 obj-y			+= drm/ vga/
+obj-$(CONFIG_TEGRA_HOST1X)	+= host1x/
diff --git a/drivers/gpu/drm/Kconfig b/drivers/gpu/drm/Kconfig
index 1e82882da9de..b16c50ee769c 100644
--- a/drivers/gpu/drm/Kconfig
+++ b/drivers/gpu/drm/Kconfig
@@ -215,8 +215,8 @@ source "drivers/gpu/drm/cirrus/Kconfig"
 
 source "drivers/gpu/drm/shmobile/Kconfig"
 
-source "drivers/gpu/drm/tegra/Kconfig"
-
 source "drivers/gpu/drm/omapdrm/Kconfig"
 
 source "drivers/gpu/drm/tilcdc/Kconfig"
+
+source "drivers/gpu/drm/qxl/Kconfig"
diff --git a/drivers/gpu/drm/Makefile b/drivers/gpu/drm/Makefile
index 0d59b24f8d23..1c9f24396002 100644
--- a/drivers/gpu/drm/Makefile
+++ b/drivers/gpu/drm/Makefile
@@ -49,7 +49,7 @@ obj-$(CONFIG_DRM_GMA500) += gma500/
 obj-$(CONFIG_DRM_UDL) += udl/
 obj-$(CONFIG_DRM_AST) += ast/
 obj-$(CONFIG_DRM_SHMOBILE) +=shmobile/
-obj-$(CONFIG_DRM_TEGRA) += tegra/
 obj-$(CONFIG_DRM_OMAP)	+= omapdrm/
 obj-$(CONFIG_DRM_TILCDC)	+= tilcdc/
+obj-$(CONFIG_DRM_QXL) += qxl/
 obj-y			+= i2c/
diff --git a/drivers/gpu/drm/ast/ast_drv.h b/drivers/gpu/drm/ast/ast_drv.h
index 528429252f0f..02e52d543e4b 100644
--- a/drivers/gpu/drm/ast/ast_drv.h
+++ b/drivers/gpu/drm/ast/ast_drv.h
@@ -241,6 +241,8 @@ struct ast_fbdev {
 	void *sysram;
 	int size;
 	struct ttm_bo_kmap_obj mapping;
+	int x1, y1, x2, y2; /* dirty rect */
+	spinlock_t dirty_lock;
 };
 
 #define to_ast_crtc(x) container_of(x, struct ast_crtc, base)
diff --git a/drivers/gpu/drm/ast/ast_fb.c b/drivers/gpu/drm/ast/ast_fb.c
index 34931fe7d2c5..fbc0823cfa18 100644
--- a/drivers/gpu/drm/ast/ast_fb.c
+++ b/drivers/gpu/drm/ast/ast_fb.c
@@ -53,16 +53,52 @@ static void ast_dirty_update(struct ast_fbdev *afbdev,
 	int bpp = (afbdev->afb.base.bits_per_pixel + 7)/8;
 	int ret;
 	bool unmap = false;
+	bool store_for_later = false;
+	int x2, y2;
+	unsigned long flags;
 
 	obj = afbdev->afb.obj;
 	bo = gem_to_ast_bo(obj);
 
+	/*
+	 * try and reserve the BO, if we fail with busy
+	 * then the BO is being moved and we should
+	 * store up the damage until later.
+	 */
 	ret = ast_bo_reserve(bo, true);
 	if (ret) {
-		DRM_ERROR("failed to reserve fb bo\n");
+		if (ret != -EBUSY)
+			return;
+
+		store_for_later = true;
+	}
+
+	x2 = x + width - 1;
+	y2 = y + height - 1;
+	spin_lock_irqsave(&afbdev->dirty_lock, flags);
+
+	if (afbdev->y1 < y)
+		y = afbdev->y1;
+	if (afbdev->y2 > y2)
+		y2 = afbdev->y2;
+	if (afbdev->x1 < x)
+		x = afbdev->x1;
+	if (afbdev->x2 > x2)
+		x2 = afbdev->x2;
+
+	if (store_for_later) {
+		afbdev->x1 = x;
+		afbdev->x2 = x2;
+		afbdev->y1 = y;
+		afbdev->y2 = y2;
+		spin_unlock_irqrestore(&afbdev->dirty_lock, flags);
 		return;
 	}
 
+	afbdev->x1 = afbdev->y1 = INT_MAX;
+	afbdev->x2 = afbdev->y2 = 0;
+	spin_unlock_irqrestore(&afbdev->dirty_lock, flags);
+
 	if (!bo->kmap.virtual) {
 		ret = ttm_bo_kmap(&bo->bo, 0, bo->bo.num_pages, &bo->kmap);
 		if (ret) {
@@ -72,10 +108,10 @@ static void ast_dirty_update(struct ast_fbdev *afbdev,
 		}
 		unmap = true;
 	}
-	for (i = y; i < y + height; i++) {
+	for (i = y; i <= y2; i++) {
 		/* assume equal stride for now */
 		src_offset = dst_offset = i * afbdev->afb.base.pitches[0] + (x * bpp);
-		memcpy_toio(bo->kmap.virtual + src_offset, afbdev->sysram + src_offset, width * bpp);
+		memcpy_toio(bo->kmap.virtual + src_offset, afbdev->sysram + src_offset, (x2 - x + 1) * bpp);
 
 	}
 	if (unmap)
@@ -292,6 +328,7 @@ int ast_fbdev_init(struct drm_device *dev)
 
 	ast->fbdev = afbdev;
 	afbdev->helper.funcs = &ast_fb_helper_funcs;
+	spin_lock_init(&afbdev->dirty_lock);
 	ret = drm_fb_helper_init(dev, &afbdev->helper,
 				 1, 1);
 	if (ret) {
diff --git a/drivers/gpu/drm/ast/ast_ttm.c b/drivers/gpu/drm/ast/ast_ttm.c
index 3602731a6112..09da3393c527 100644
--- a/drivers/gpu/drm/ast/ast_ttm.c
+++ b/drivers/gpu/drm/ast/ast_ttm.c
@@ -316,7 +316,7 @@ int ast_bo_reserve(struct ast_bo *bo, bool no_wait)
 
 	ret = ttm_bo_reserve(&bo->bo, true, no_wait, false, 0);
 	if (ret) {
-		if (ret != -ERESTARTSYS)
+		if (ret != -ERESTARTSYS && ret != -EBUSY)
 			DRM_ERROR("reserve failed %p\n", bo);
 		return ret;
 	}
diff --git a/drivers/gpu/drm/cirrus/cirrus_drv.h b/drivers/gpu/drm/cirrus/cirrus_drv.h
index 6e0cc724e5a2..7ca059596887 100644
--- a/drivers/gpu/drm/cirrus/cirrus_drv.h
+++ b/drivers/gpu/drm/cirrus/cirrus_drv.h
@@ -154,6 +154,8 @@ struct cirrus_fbdev {
 	struct list_head fbdev_list;
 	void *sysram;
 	int size;
+	int x1, y1, x2, y2; /* dirty rect */
+	spinlock_t dirty_lock;
 };
 
 struct cirrus_bo {
diff --git a/drivers/gpu/drm/cirrus/cirrus_fbdev.c b/drivers/gpu/drm/cirrus/cirrus_fbdev.c
index e25afccaf85b..3541b567bbd8 100644
--- a/drivers/gpu/drm/cirrus/cirrus_fbdev.c
+++ b/drivers/gpu/drm/cirrus/cirrus_fbdev.c
@@ -27,16 +27,51 @@ static void cirrus_dirty_update(struct cirrus_fbdev *afbdev,
 	int bpp = (afbdev->gfb.base.bits_per_pixel + 7)/8;
 	int ret;
 	bool unmap = false;
+	bool store_for_later = false;
+	int x2, y2;
+	unsigned long flags;
 
 	obj = afbdev->gfb.obj;
 	bo = gem_to_cirrus_bo(obj);
 
+	/*
+	 * try and reserve the BO, if we fail with busy
+	 * then the BO is being moved and we should
+	 * store up the damage until later.
+	 */
 	ret = cirrus_bo_reserve(bo, true);
 	if (ret) {
-		DRM_ERROR("failed to reserve fb bo\n");
+		if (ret != -EBUSY)
+			return;
+		store_for_later = true;
+	}
+
+	x2 = x + width - 1;
+	y2 = y + height - 1;
+	spin_lock_irqsave(&afbdev->dirty_lock, flags);
+
+	if (afbdev->y1 < y)
+		y = afbdev->y1;
+	if (afbdev->y2 > y2)
+		y2 = afbdev->y2;
+	if (afbdev->x1 < x)
+		x = afbdev->x1;
+	if (afbdev->x2 > x2)
+		x2 = afbdev->x2;
+
+	if (store_for_later) {
+		afbdev->x1 = x;
+		afbdev->x2 = x2;
+		afbdev->y1 = y;
+		afbdev->y2 = y2;
+		spin_unlock_irqrestore(&afbdev->dirty_lock, flags);
 		return;
 	}
 
+	afbdev->x1 = afbdev->y1 = INT_MAX;
+	afbdev->x2 = afbdev->y2 = 0;
+	spin_unlock_irqrestore(&afbdev->dirty_lock, flags);
+
 	if (!bo->kmap.virtual) {
 		ret = ttm_bo_kmap(&bo->bo, 0, bo->bo.num_pages, &bo->kmap);
 		if (ret) {
@@ -268,6 +303,7 @@ int cirrus_fbdev_init(struct cirrus_device *cdev)
 
 	cdev->mode_info.gfbdev = gfbdev;
 	gfbdev->helper.funcs = &cirrus_fb_helper_funcs;
+	spin_lock_init(&gfbdev->dirty_lock);
 
 	ret = drm_fb_helper_init(cdev->dev, &gfbdev->helper,
 				 cdev->num_crtc, CIRRUSFB_CONN_LIMIT);
diff --git a/drivers/gpu/drm/cirrus/cirrus_ttm.c b/drivers/gpu/drm/cirrus/cirrus_ttm.c
index 1413a26e4905..2ed8cfc740c9 100644
--- a/drivers/gpu/drm/cirrus/cirrus_ttm.c
+++ b/drivers/gpu/drm/cirrus/cirrus_ttm.c
@@ -321,7 +321,7 @@ int cirrus_bo_reserve(struct cirrus_bo *bo, bool no_wait)
 
 	ret = ttm_bo_reserve(&bo->bo, true, no_wait, false, 0);
 	if (ret) {
-		if (ret != -ERESTARTSYS)
+		if (ret != -ERESTARTSYS && ret != -EBUSY)
 			DRM_ERROR("reserve failed %p\n", bo);
 		return ret;
 	}
diff --git a/drivers/gpu/drm/drm_cache.c b/drivers/gpu/drm/drm_cache.c
index a575cb2e6bdb..bb8f58012189 100644
--- a/drivers/gpu/drm/drm_cache.c
+++ b/drivers/gpu/drm/drm_cache.c
@@ -105,12 +105,11 @@ drm_clflush_sg(struct sg_table *st)
 {
 #if defined(CONFIG_X86)
 	if (cpu_has_clflush) {
-		struct scatterlist *sg;
-		int i;
+		struct sg_page_iter sg_iter;
 
 		mb();
-		for_each_sg(st->sgl, sg, st->nents, i)
-			drm_clflush_page(sg_page(sg));
+		for_each_sg_page(st->sgl, &sg_iter, st->nents, 0)
+			drm_clflush_page(sg_page_iter_page(&sg_iter));
 		mb();
 
 		return;
diff --git a/drivers/gpu/drm/drm_crtc.c b/drivers/gpu/drm/drm_crtc.c
index dd64a06dc5b4..3a8f7e6db295 100644
--- a/drivers/gpu/drm/drm_crtc.c
+++ b/drivers/gpu/drm/drm_crtc.c
@@ -178,9 +178,6 @@ static struct drm_prop_enum_list drm_dirty_info_enum_list[] = {
 	{ DRM_MODE_DIRTY_ANNOTATE, "Annotate" },
 };
 
-DRM_ENUM_NAME_FN(drm_get_dirty_info_name,
-		 drm_dirty_info_enum_list)
-
 struct drm_conn_prop_enum_list {
 	int type;
 	char *name;
@@ -412,7 +409,7 @@ struct drm_framebuffer *drm_framebuffer_lookup(struct drm_device *dev,
 	mutex_lock(&dev->mode_config.fb_lock);
 	fb = __drm_framebuffer_lookup(dev, id);
 	if (fb)
-		kref_get(&fb->refcount);
+		drm_framebuffer_reference(fb);
 	mutex_unlock(&dev->mode_config.fb_lock);
 
 	return fb;
@@ -706,7 +703,6 @@ int drm_connector_init(struct drm_device *dev,
 	connector->connector_type = connector_type;
 	connector->connector_type_id =
 		++drm_connector_enum_list[connector_type].count; /* TODO */
-	INIT_LIST_HEAD(&connector->user_modes);
 	INIT_LIST_HEAD(&connector->probed_modes);
 	INIT_LIST_HEAD(&connector->modes);
 	connector->edid_blob_ptr = NULL;
@@ -747,9 +743,6 @@ void drm_connector_cleanup(struct drm_connector *connector)
 	list_for_each_entry_safe(mode, t, &connector->modes, head)
 		drm_mode_remove(connector, mode);
 
-	list_for_each_entry_safe(mode, t, &connector->user_modes, head)
-		drm_mode_remove(connector, mode);
-
 	drm_mode_object_put(dev, &connector->base);
 	list_del(&connector->head);
 	dev->mode_config.num_connector--;
@@ -1120,45 +1113,7 @@ int drm_mode_create_dirty_info_property(struct drm_device *dev)
 }
 EXPORT_SYMBOL(drm_mode_create_dirty_info_property);
 
-/**
- * drm_mode_config_init - initialize DRM mode_configuration structure
- * @dev: DRM device
- *
- * Initialize @dev's mode_config structure, used for tracking the graphics
- * configuration of @dev.
- *
- * Since this initializes the modeset locks, no locking is possible. Which is no
- * problem, since this should happen single threaded at init time. It is the
- * driver's problem to ensure this guarantee.
- *
- */
-void drm_mode_config_init(struct drm_device *dev)
-{
-	mutex_init(&dev->mode_config.mutex);
-	mutex_init(&dev->mode_config.idr_mutex);
-	mutex_init(&dev->mode_config.fb_lock);
-	INIT_LIST_HEAD(&dev->mode_config.fb_list);
-	INIT_LIST_HEAD(&dev->mode_config.crtc_list);
-	INIT_LIST_HEAD(&dev->mode_config.connector_list);
-	INIT_LIST_HEAD(&dev->mode_config.encoder_list);
-	INIT_LIST_HEAD(&dev->mode_config.property_list);
-	INIT_LIST_HEAD(&dev->mode_config.property_blob_list);
-	INIT_LIST_HEAD(&dev->mode_config.plane_list);
-	idr_init(&dev->mode_config.crtc_idr);
-
-	drm_modeset_lock_all(dev);
-	drm_mode_create_standard_connector_properties(dev);
-	drm_modeset_unlock_all(dev);
-
-	/* Just to be sure */
-	dev->mode_config.num_fb = 0;
-	dev->mode_config.num_connector = 0;
-	dev->mode_config.num_crtc = 0;
-	dev->mode_config.num_encoder = 0;
-}
-EXPORT_SYMBOL(drm_mode_config_init);
-
-int drm_mode_group_init(struct drm_device *dev, struct drm_mode_group *group)
+static int drm_mode_group_init(struct drm_device *dev, struct drm_mode_group *group)
 {
 	uint32_t total_objects = 0;
 
@@ -1203,69 +1158,6 @@ int drm_mode_group_init_legacy_group(struct drm_device *dev,
 EXPORT_SYMBOL(drm_mode_group_init_legacy_group);
 
 /**
- * drm_mode_config_cleanup - free up DRM mode_config info
- * @dev: DRM device
- *
- * Free up all the connectors and CRTCs associated with this DRM device, then
- * free up the framebuffers and associated buffer objects.
- *
- * Note that since this /should/ happen single-threaded at driver/device
- * teardown time, no locking is required. It's the driver's job to ensure that
- * this guarantee actually holds true.
- *
- * FIXME: cleanup any dangling user buffer objects too
- */
-void drm_mode_config_cleanup(struct drm_device *dev)
-{
-	struct drm_connector *connector, *ot;
-	struct drm_crtc *crtc, *ct;
-	struct drm_encoder *encoder, *enct;
-	struct drm_framebuffer *fb, *fbt;
-	struct drm_property *property, *pt;
-	struct drm_plane *plane, *plt;
-
-	list_for_each_entry_safe(encoder, enct, &dev->mode_config.encoder_list,
-				 head) {
-		encoder->funcs->destroy(encoder);
-	}
-
-	list_for_each_entry_safe(connector, ot,
-				 &dev->mode_config.connector_list, head) {
-		connector->funcs->destroy(connector);
-	}
-
-	list_for_each_entry_safe(property, pt, &dev->mode_config.property_list,
-				 head) {
-		drm_property_destroy(dev, property);
-	}
-
-	/*
-	 * Single-threaded teardown context, so it's not required to grab the
-	 * fb_lock to protect against concurrent fb_list access. Contrary, it
-	 * would actually deadlock with the drm_framebuffer_cleanup function.
-	 *
-	 * Also, if there are any framebuffers left, that's a driver leak now,
-	 * so politely WARN about this.
-	 */
-	WARN_ON(!list_empty(&dev->mode_config.fb_list));
-	list_for_each_entry_safe(fb, fbt, &dev->mode_config.fb_list, head) {
-		drm_framebuffer_remove(fb);
-	}
-
-	list_for_each_entry_safe(plane, plt, &dev->mode_config.plane_list,
-				 head) {
-		plane->funcs->destroy(plane);
-	}
-
-	list_for_each_entry_safe(crtc, ct, &dev->mode_config.crtc_list, head) {
-		crtc->funcs->destroy(crtc);
-	}
-
-	idr_destroy(&dev->mode_config.crtc_idr);
-}
-EXPORT_SYMBOL(drm_mode_config_cleanup);
-
-/**
  * drm_crtc_convert_to_umode - convert a drm_display_mode into a modeinfo
  * @out: drm_mode_modeinfo struct to return to the user
  * @in: drm_display_mode to use
@@ -2717,192 +2609,6 @@ void drm_fb_release(struct drm_file *priv)
 	mutex_unlock(&priv->fbs_lock);
 }
 
-/**
- * drm_mode_attachmode - add a mode to the user mode list
- * @dev: DRM device
- * @connector: connector to add the mode to
- * @mode: mode to add
- *
- * Add @mode to @connector's user mode list.
- */
-static void drm_mode_attachmode(struct drm_device *dev,
-				struct drm_connector *connector,
-				struct drm_display_mode *mode)
-{
-	list_add_tail(&mode->head, &connector->user_modes);
-}
-
-int drm_mode_attachmode_crtc(struct drm_device *dev, struct drm_crtc *crtc,
-			     const struct drm_display_mode *mode)
-{
-	struct drm_connector *connector;
-	int ret = 0;
-	struct drm_display_mode *dup_mode, *next;
-	LIST_HEAD(list);
-
-	list_for_each_entry(connector, &dev->mode_config.connector_list, head) {
-		if (!connector->encoder)
-			continue;
-		if (connector->encoder->crtc == crtc) {
-			dup_mode = drm_mode_duplicate(dev, mode);
-			if (!dup_mode) {
-				ret = -ENOMEM;
-				goto out;
-			}
-			list_add_tail(&dup_mode->head, &list);
-		}
-	}
-
-	list_for_each_entry(connector, &dev->mode_config.connector_list, head) {
-		if (!connector->encoder)
-			continue;
-		if (connector->encoder->crtc == crtc)
-			list_move_tail(list.next, &connector->user_modes);
-	}
-
-	WARN_ON(!list_empty(&list));
-
- out:
-	list_for_each_entry_safe(dup_mode, next, &list, head)
-		drm_mode_destroy(dev, dup_mode);
-
-	return ret;
-}
-EXPORT_SYMBOL(drm_mode_attachmode_crtc);
-
-static int drm_mode_detachmode(struct drm_device *dev,
-			       struct drm_connector *connector,
-			       struct drm_display_mode *mode)
-{
-	int found = 0;
-	int ret = 0;
-	struct drm_display_mode *match_mode, *t;
-
-	list_for_each_entry_safe(match_mode, t, &connector->user_modes, head) {
-		if (drm_mode_equal(match_mode, mode)) {
-			list_del(&match_mode->head);
-			drm_mode_destroy(dev, match_mode);
-			found = 1;
-			break;
-		}
-	}
-
-	if (!found)
-		ret = -EINVAL;
-
-	return ret;
-}
-
-int drm_mode_detachmode_crtc(struct drm_device *dev, struct drm_display_mode *mode)
-{
-	struct drm_connector *connector;
-
-	list_for_each_entry(connector, &dev->mode_config.connector_list, head) {
-		drm_mode_detachmode(dev, connector, mode);
-	}
-	return 0;
-}
-EXPORT_SYMBOL(drm_mode_detachmode_crtc);
-
-/**
- * drm_fb_attachmode - Attach a user mode to an connector
- * @dev: drm device for the ioctl
- * @data: data pointer for the ioctl
- * @file_priv: drm file for the ioctl call
- *
- * This attaches a user specified mode to an connector.
- * Called by the user via ioctl.
- *
- * RETURNS:
- * Zero on success, errno on failure.
- */
-int drm_mode_attachmode_ioctl(struct drm_device *dev,
-			      void *data, struct drm_file *file_priv)
-{
-	struct drm_mode_mode_cmd *mode_cmd = data;
-	struct drm_connector *connector;
-	struct drm_display_mode *mode;
-	struct drm_mode_object *obj;
-	struct drm_mode_modeinfo *umode = &mode_cmd->mode;
-	int ret;
-
-	if (!drm_core_check_feature(dev, DRIVER_MODESET))
-		return -EINVAL;
-
-	drm_modeset_lock_all(dev);
-
-	obj = drm_mode_object_find(dev, mode_cmd->connector_id, DRM_MODE_OBJECT_CONNECTOR);
-	if (!obj) {
-		ret = -EINVAL;
-		goto out;
-	}
-	connector = obj_to_connector(obj);
-
-	mode = drm_mode_create(dev);
-	if (!mode) {
-		ret = -ENOMEM;
-		goto out;
-	}
-
-	ret = drm_crtc_convert_umode(mode, umode);
-	if (ret) {
-		DRM_DEBUG_KMS("Invalid mode\n");
-		drm_mode_destroy(dev, mode);
-		goto out;
-	}
-
-	drm_mode_attachmode(dev, connector, mode);
-out:
-	drm_modeset_unlock_all(dev);
-	return ret;
-}
-
-
-/**
- * drm_fb_detachmode - Detach a user specified mode from an connector
- * @dev: drm device for the ioctl
- * @data: data pointer for the ioctl
- * @file_priv: drm file for the ioctl call
- *
- * Called by the user via ioctl.
- *
- * RETURNS:
- * Zero on success, errno on failure.
- */
-int drm_mode_detachmode_ioctl(struct drm_device *dev,
-			      void *data, struct drm_file *file_priv)
-{
-	struct drm_mode_object *obj;
-	struct drm_mode_mode_cmd *mode_cmd = data;
-	struct drm_connector *connector;
-	struct drm_display_mode mode;
-	struct drm_mode_modeinfo *umode = &mode_cmd->mode;
-	int ret;
-
-	if (!drm_core_check_feature(dev, DRIVER_MODESET))
-		return -EINVAL;
-
-	drm_modeset_lock_all(dev);
-
-	obj = drm_mode_object_find(dev, mode_cmd->connector_id, DRM_MODE_OBJECT_CONNECTOR);
-	if (!obj) {
-		ret = -EINVAL;
-		goto out;
-	}
-	connector = obj_to_connector(obj);
-
-	ret = drm_crtc_convert_umode(&mode, umode);
-	if (ret) {
-		DRM_DEBUG_KMS("Invalid mode\n");
-		goto out;
-	}
-
-	ret = drm_mode_detachmode(dev, connector, &mode);
-out:
-	drm_modeset_unlock_all(dev);
-	return ret;
-}
-
 struct drm_property *drm_property_create(struct drm_device *dev, int flags,
 					 const char *name, int num_values)
 {
@@ -3739,6 +3445,12 @@ int drm_mode_page_flip_ioctl(struct drm_device *dev,
 		goto out;
 	}
 
+	if (crtc->fb->pixel_format != fb->pixel_format) {
+		DRM_DEBUG_KMS("Page flip is not allowed to change frame buffer format.\n");
+		ret = -EINVAL;
+		goto out;
+	}
+
 	if (page_flip->flags & DRM_MODE_PAGE_FLIP_EVENT) {
 		ret = -ENOMEM;
 		spin_lock_irqsave(&dev->event_lock, flags);
@@ -4064,3 +3776,110 @@ int drm_format_vert_chroma_subsampling(uint32_t format)
 	}
 }
 EXPORT_SYMBOL(drm_format_vert_chroma_subsampling);
+
+/**
+ * drm_mode_config_init - initialize DRM mode_configuration structure
+ * @dev: DRM device
+ *
+ * Initialize @dev's mode_config structure, used for tracking the graphics
+ * configuration of @dev.
+ *
+ * Since this initializes the modeset locks, no locking is possible. Which is no
+ * problem, since this should happen single threaded at init time. It is the
+ * driver's problem to ensure this guarantee.
+ *
+ */
+void drm_mode_config_init(struct drm_device *dev)
+{
+	mutex_init(&dev->mode_config.mutex);
+	mutex_init(&dev->mode_config.idr_mutex);
+	mutex_init(&dev->mode_config.fb_lock);
+	INIT_LIST_HEAD(&dev->mode_config.fb_list);
+	INIT_LIST_HEAD(&dev->mode_config.crtc_list);
+	INIT_LIST_HEAD(&dev->mode_config.connector_list);
+	INIT_LIST_HEAD(&dev->mode_config.encoder_list);
+	INIT_LIST_HEAD(&dev->mode_config.property_list);
+	INIT_LIST_HEAD(&dev->mode_config.property_blob_list);
+	INIT_LIST_HEAD(&dev->mode_config.plane_list);
+	idr_init(&dev->mode_config.crtc_idr);
+
+	drm_modeset_lock_all(dev);
+	drm_mode_create_standard_connector_properties(dev);
+	drm_modeset_unlock_all(dev);
+
+	/* Just to be sure */
+	dev->mode_config.num_fb = 0;
+	dev->mode_config.num_connector = 0;
+	dev->mode_config.num_crtc = 0;
+	dev->mode_config.num_encoder = 0;
+}
+EXPORT_SYMBOL(drm_mode_config_init);
+
+/**
+ * drm_mode_config_cleanup - free up DRM mode_config info
+ * @dev: DRM device
+ *
+ * Free up all the connectors and CRTCs associated with this DRM device, then
+ * free up the framebuffers and associated buffer objects.
+ *
+ * Note that since this /should/ happen single-threaded at driver/device
+ * teardown time, no locking is required. It's the driver's job to ensure that
+ * this guarantee actually holds true.
+ *
+ * FIXME: cleanup any dangling user buffer objects too
+ */
+void drm_mode_config_cleanup(struct drm_device *dev)
+{
+	struct drm_connector *connector, *ot;
+	struct drm_crtc *crtc, *ct;
+	struct drm_encoder *encoder, *enct;
+	struct drm_framebuffer *fb, *fbt;
+	struct drm_property *property, *pt;
+	struct drm_property_blob *blob, *bt;
+	struct drm_plane *plane, *plt;
+
+	list_for_each_entry_safe(encoder, enct, &dev->mode_config.encoder_list,
+				 head) {
+		encoder->funcs->destroy(encoder);
+	}
+
+	list_for_each_entry_safe(connector, ot,
+				 &dev->mode_config.connector_list, head) {
+		connector->funcs->destroy(connector);
+	}
+
+	list_for_each_entry_safe(property, pt, &dev->mode_config.property_list,
+				 head) {
+		drm_property_destroy(dev, property);
+	}
+
+	list_for_each_entry_safe(blob, bt, &dev->mode_config.property_blob_list,
+				 head) {
+		drm_property_destroy_blob(dev, blob);
+	}
+
+	/*
+	 * Single-threaded teardown context, so it's not required to grab the
+	 * fb_lock to protect against concurrent fb_list access. Contrary, it
+	 * would actually deadlock with the drm_framebuffer_cleanup function.
+	 *
+	 * Also, if there are any framebuffers left, that's a driver leak now,
+	 * so politely WARN about this.
+	 */
+	WARN_ON(!list_empty(&dev->mode_config.fb_list));
+	list_for_each_entry_safe(fb, fbt, &dev->mode_config.fb_list, head) {
+		drm_framebuffer_remove(fb);
+	}
+
+	list_for_each_entry_safe(plane, plt, &dev->mode_config.plane_list,
+				 head) {
+		plane->funcs->destroy(plane);
+	}
+
+	list_for_each_entry_safe(crtc, ct, &dev->mode_config.crtc_list, head) {
+		crtc->funcs->destroy(crtc);
+	}
+
+	idr_destroy(&dev->mode_config.crtc_idr);
+}
+EXPORT_SYMBOL(drm_mode_config_cleanup);
diff --git a/drivers/gpu/drm/drm_crtc_helper.c b/drivers/gpu/drm/drm_crtc_helper.c
index 7b2d378b2576..e974f9309b72 100644
--- a/drivers/gpu/drm/drm_crtc_helper.c
+++ b/drivers/gpu/drm/drm_crtc_helper.c
@@ -648,6 +648,9 @@ int drm_crtc_helper_set_config(struct drm_mode_set *set)
 		} else if (set->fb->bits_per_pixel !=
 			   set->crtc->fb->bits_per_pixel) {
 			mode_changed = true;
+		} else if (set->fb->pixel_format !=
+			   set->crtc->fb->pixel_format) {
+			mode_changed = true;
 		} else
 			fb_changed = true;
 	}
diff --git a/drivers/gpu/drm/drm_drv.c b/drivers/gpu/drm/drm_drv.c
index 25f91cd23e60..8d4f29075af5 100644
--- a/drivers/gpu/drm/drm_drv.c
+++ b/drivers/gpu/drm/drm_drv.c
@@ -60,7 +60,7 @@ static int drm_version(struct drm_device *dev, void *data,
 	[DRM_IOCTL_NR(ioctl)] = {.cmd = ioctl, .func = _func, .flags = _flags, .cmd_drv = 0}
 
 /** Ioctl table */
-static struct drm_ioctl_desc drm_ioctls[] = {
+static const struct drm_ioctl_desc drm_ioctls[] = {
 	DRM_IOCTL_DEF(DRM_IOCTL_VERSION, drm_version, DRM_UNLOCKED),
 	DRM_IOCTL_DEF(DRM_IOCTL_GET_UNIQUE, drm_getunique, 0),
 	DRM_IOCTL_DEF(DRM_IOCTL_GET_MAGIC, drm_getmagic, 0),
@@ -150,8 +150,8 @@ static struct drm_ioctl_desc drm_ioctls[] = {
 	DRM_IOCTL_DEF(DRM_IOCTL_MODE_SETGAMMA, drm_mode_gamma_set_ioctl, DRM_MASTER|DRM_UNLOCKED),
 	DRM_IOCTL_DEF(DRM_IOCTL_MODE_GETENCODER, drm_mode_getencoder, DRM_CONTROL_ALLOW|DRM_UNLOCKED),
 	DRM_IOCTL_DEF(DRM_IOCTL_MODE_GETCONNECTOR, drm_mode_getconnector, DRM_CONTROL_ALLOW|DRM_UNLOCKED),
-	DRM_IOCTL_DEF(DRM_IOCTL_MODE_ATTACHMODE, drm_mode_attachmode_ioctl, DRM_MASTER|DRM_CONTROL_ALLOW|DRM_UNLOCKED),
-	DRM_IOCTL_DEF(DRM_IOCTL_MODE_DETACHMODE, drm_mode_detachmode_ioctl, DRM_MASTER|DRM_CONTROL_ALLOW|DRM_UNLOCKED),
+	DRM_IOCTL_DEF(DRM_IOCTL_MODE_ATTACHMODE, drm_noop, DRM_MASTER|DRM_CONTROL_ALLOW|DRM_UNLOCKED),
+	DRM_IOCTL_DEF(DRM_IOCTL_MODE_DETACHMODE, drm_noop, DRM_MASTER|DRM_CONTROL_ALLOW|DRM_UNLOCKED),
 	DRM_IOCTL_DEF(DRM_IOCTL_MODE_GETPROPERTY, drm_mode_getproperty_ioctl, DRM_CONTROL_ALLOW|DRM_UNLOCKED),
 	DRM_IOCTL_DEF(DRM_IOCTL_MODE_SETPROPERTY, drm_mode_connector_property_set_ioctl, DRM_MASTER|DRM_CONTROL_ALLOW|DRM_UNLOCKED),
 	DRM_IOCTL_DEF(DRM_IOCTL_MODE_GETPROPBLOB, drm_mode_getblob_ioctl, DRM_CONTROL_ALLOW|DRM_UNLOCKED),
@@ -375,7 +375,7 @@ long drm_ioctl(struct file *filp,
 {
 	struct drm_file *file_priv = filp->private_data;
 	struct drm_device *dev;
-	struct drm_ioctl_desc *ioctl;
+	const struct drm_ioctl_desc *ioctl;
 	drm_ioctl_t *func;
 	unsigned int nr = DRM_IOCTL_NR(cmd);
 	int retcode = -EINVAL;
@@ -408,6 +408,7 @@ long drm_ioctl(struct file *filp,
 		usize = asize = _IOC_SIZE(cmd);
 		if (drv_size > asize)
 			asize = drv_size;
+		cmd = ioctl->cmd_drv;
 	}
 	else if ((nr >= DRM_COMMAND_END) || (nr < DRM_COMMAND_BASE)) {
 		ioctl = &drm_ioctls[nr];
diff --git a/drivers/gpu/drm/drm_edid.c b/drivers/gpu/drm/drm_edid.c
index e2acfdbf7d3c..9e62bbedb5ad 100644
--- a/drivers/gpu/drm/drm_edid.c
+++ b/drivers/gpu/drm/drm_edid.c
@@ -587,284 +587,348 @@ static const struct drm_display_mode edid_cea_modes[] = {
 	/* 1 - 640x480@60Hz */
 	{ DRM_MODE("640x480", DRM_MODE_TYPE_DRIVER, 25175, 640, 656,
 		   752, 800, 0, 480, 490, 492, 525, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) },
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC),
+	  .vrefresh = 60, },
 	/* 2 - 720x480@60Hz */
 	{ DRM_MODE("720x480", DRM_MODE_TYPE_DRIVER, 27000, 720, 736,
 		   798, 858, 0, 480, 489, 495, 525, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) },
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC),
+	  .vrefresh = 60, },
 	/* 3 - 720x480@60Hz */
 	{ DRM_MODE("720x480", DRM_MODE_TYPE_DRIVER, 27000, 720, 736,
 		   798, 858, 0, 480, 489, 495, 525, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) },
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC),
+	  .vrefresh = 60, },
 	/* 4 - 1280x720@60Hz */
 	{ DRM_MODE("1280x720", DRM_MODE_TYPE_DRIVER, 74250, 1280, 1390,
 		   1430, 1650, 0, 720, 725, 730, 750, 0,
-		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC),
+	  .vrefresh = 60, },
 	/* 5 - 1920x1080i@60Hz */
 	{ DRM_MODE("1920x1080i", DRM_MODE_TYPE_DRIVER, 74250, 1920, 2008,
 		   2052, 2200, 0, 1080, 1084, 1094, 1125, 0,
 		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC |
-			DRM_MODE_FLAG_INTERLACE) },
+			DRM_MODE_FLAG_INTERLACE),
+	  .vrefresh = 60, },
 	/* 6 - 1440x480i@60Hz */
 	{ DRM_MODE("1440x480i", DRM_MODE_TYPE_DRIVER, 27000, 1440, 1478,
 		   1602, 1716, 0, 480, 488, 494, 525, 0,
 		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC |
-			DRM_MODE_FLAG_INTERLACE | DRM_MODE_FLAG_DBLCLK) },
+			DRM_MODE_FLAG_INTERLACE | DRM_MODE_FLAG_DBLCLK),
+	  .vrefresh = 60, },
 	/* 7 - 1440x480i@60Hz */
 	{ DRM_MODE("1440x480i", DRM_MODE_TYPE_DRIVER, 27000, 1440, 1478,
 		   1602, 1716, 0, 480, 488, 494, 525, 0,
 		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC |
-			DRM_MODE_FLAG_INTERLACE | DRM_MODE_FLAG_DBLCLK) },
+			DRM_MODE_FLAG_INTERLACE | DRM_MODE_FLAG_DBLCLK),
+	  .vrefresh = 60, },
 	/* 8 - 1440x240@60Hz */
 	{ DRM_MODE("1440x240", DRM_MODE_TYPE_DRIVER, 27000, 1440, 1478,
 		   1602, 1716, 0, 240, 244, 247, 262, 0,
 		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC |
-			DRM_MODE_FLAG_DBLCLK) },
+			DRM_MODE_FLAG_DBLCLK),
+	  .vrefresh = 60, },
 	/* 9 - 1440x240@60Hz */
 	{ DRM_MODE("1440x240", DRM_MODE_TYPE_DRIVER, 27000, 1440, 1478,
 		   1602, 1716, 0, 240, 244, 247, 262, 0,
 		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC |
-			DRM_MODE_FLAG_DBLCLK) },
+			DRM_MODE_FLAG_DBLCLK),
+	  .vrefresh = 60, },
 	/* 10 - 2880x480i@60Hz */
 	{ DRM_MODE("2880x480i", DRM_MODE_TYPE_DRIVER, 54000, 2880, 2956,
 		   3204, 3432, 0, 480, 488, 494, 525, 0,
 		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC |
-			DRM_MODE_FLAG_INTERLACE) },
+			DRM_MODE_FLAG_INTERLACE),
+	  .vrefresh = 60, },
 	/* 11 - 2880x480i@60Hz */
 	{ DRM_MODE("2880x480i", DRM_MODE_TYPE_DRIVER, 54000, 2880, 2956,
 		   3204, 3432, 0, 480, 488, 494, 525, 0,
 		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC |
-			DRM_MODE_FLAG_INTERLACE) },
+			DRM_MODE_FLAG_INTERLACE),
+	  .vrefresh = 60, },
 	/* 12 - 2880x240@60Hz */
 	{ DRM_MODE("2880x240", DRM_MODE_TYPE_DRIVER, 54000, 2880, 2956,
 		   3204, 3432, 0, 240, 244, 247, 262, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) },
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC),
+	  .vrefresh = 60, },
 	/* 13 - 2880x240@60Hz */
 	{ DRM_MODE("2880x240", DRM_MODE_TYPE_DRIVER, 54000, 2880, 2956,
 		   3204, 3432, 0, 240, 244, 247, 262, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) },
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC),
+	  .vrefresh = 60, },
 	/* 14 - 1440x480@60Hz */
 	{ DRM_MODE("1440x480", DRM_MODE_TYPE_DRIVER, 54000, 1440, 1472,
 		   1596, 1716, 0, 480, 489, 495, 525, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) },
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC),
+	  .vrefresh = 60, },
 	/* 15 - 1440x480@60Hz */
 	{ DRM_MODE("1440x480", DRM_MODE_TYPE_DRIVER, 54000, 1440, 1472,
 		   1596, 1716, 0, 480, 489, 495, 525, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) },
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC),
+	  .vrefresh = 60, },
 	/* 16 - 1920x1080@60Hz */
 	{ DRM_MODE("1920x1080", DRM_MODE_TYPE_DRIVER, 148500, 1920, 2008,
 		   2052, 2200, 0, 1080, 1084, 1089, 1125, 0,
-		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC),
+	  .vrefresh = 60, },
 	/* 17 - 720x576@50Hz */
 	{ DRM_MODE("720x576", DRM_MODE_TYPE_DRIVER, 27000, 720, 732,
 		   796, 864, 0, 576, 581, 586, 625, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) },
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC),
+	  .vrefresh = 50, },
 	/* 18 - 720x576@50Hz */
 	{ DRM_MODE("720x576", DRM_MODE_TYPE_DRIVER, 27000, 720, 732,
 		   796, 864, 0, 576, 581, 586, 625, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) },
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC),
+	  .vrefresh = 50, },
 	/* 19 - 1280x720@50Hz */
 	{ DRM_MODE("1280x720", DRM_MODE_TYPE_DRIVER, 74250, 1280, 1720,
 		   1760, 1980, 0, 720, 725, 730, 750, 0,
-		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC),
+	  .vrefresh = 50, },
 	/* 20 - 1920x1080i@50Hz */
 	{ DRM_MODE("1920x1080i", DRM_MODE_TYPE_DRIVER, 74250, 1920, 2448,
 		   2492, 2640, 0, 1080, 1084, 1094, 1125, 0,
 		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC |
-			DRM_MODE_FLAG_INTERLACE) },
+			DRM_MODE_FLAG_INTERLACE),
+	  .vrefresh = 50, },
 	/* 21 - 1440x576i@50Hz */
 	{ DRM_MODE("1440x576i", DRM_MODE_TYPE_DRIVER, 27000, 1440, 1464,
 		   1590, 1728, 0, 576, 580, 586, 625, 0,
 		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC |
-			DRM_MODE_FLAG_INTERLACE | DRM_MODE_FLAG_DBLCLK) },
+			DRM_MODE_FLAG_INTERLACE | DRM_MODE_FLAG_DBLCLK),
+	  .vrefresh = 50, },
 	/* 22 - 1440x576i@50Hz */
 	{ DRM_MODE("1440x576i", DRM_MODE_TYPE_DRIVER, 27000, 1440, 1464,
 		   1590, 1728, 0, 576, 580, 586, 625, 0,
 		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC |
-			DRM_MODE_FLAG_INTERLACE | DRM_MODE_FLAG_DBLCLK) },
+			DRM_MODE_FLAG_INTERLACE | DRM_MODE_FLAG_DBLCLK),
+	  .vrefresh = 50, },
 	/* 23 - 1440x288@50Hz */
 	{ DRM_MODE("1440x288", DRM_MODE_TYPE_DRIVER, 27000, 1440, 1464,
 		   1590, 1728, 0, 288, 290, 293, 312, 0,
 		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC |
-			DRM_MODE_FLAG_DBLCLK) },
+			DRM_MODE_FLAG_DBLCLK),
+	  .vrefresh = 50, },
 	/* 24 - 1440x288@50Hz */
 	{ DRM_MODE("1440x288", DRM_MODE_TYPE_DRIVER, 27000, 1440, 1464,
 		   1590, 1728, 0, 288, 290, 293, 312, 0,
 		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC |
-			DRM_MODE_FLAG_DBLCLK) },
+			DRM_MODE_FLAG_DBLCLK),
+	  .vrefresh = 50, },
 	/* 25 - 2880x576i@50Hz */
 	{ DRM_MODE("2880x576i", DRM_MODE_TYPE_DRIVER, 54000, 2880, 2928,
 		   3180, 3456, 0, 576, 580, 586, 625, 0,
 		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC |
-			DRM_MODE_FLAG_INTERLACE) },
+			DRM_MODE_FLAG_INTERLACE),
+	  .vrefresh = 50, },
 	/* 26 - 2880x576i@50Hz */
 	{ DRM_MODE("2880x576i", DRM_MODE_TYPE_DRIVER, 54000, 2880, 2928,
 		   3180, 3456, 0, 576, 580, 586, 625, 0,
 		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC |
-			DRM_MODE_FLAG_INTERLACE) },
+			DRM_MODE_FLAG_INTERLACE),
+	  .vrefresh = 50, },
 	/* 27 - 2880x288@50Hz */
 	{ DRM_MODE("2880x288", DRM_MODE_TYPE_DRIVER, 54000, 2880, 2928,
 		   3180, 3456, 0, 288, 290, 293, 312, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) },
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC),
+	  .vrefresh = 50, },
 	/* 28 - 2880x288@50Hz */
 	{ DRM_MODE("2880x288", DRM_MODE_TYPE_DRIVER, 54000, 2880, 2928,
 		   3180, 3456, 0, 288, 290, 293, 312, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) },
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC),
+	  .vrefresh = 50, },
 	/* 29 - 1440x576@50Hz */
 	{ DRM_MODE("1440x576", DRM_MODE_TYPE_DRIVER, 54000, 1440, 1464,
 		   1592, 1728, 0, 576, 581, 586, 625, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) },
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC),
+	  .vrefresh = 50, },
 	/* 30 - 1440x576@50Hz */
 	{ DRM_MODE("1440x576", DRM_MODE_TYPE_DRIVER, 54000, 1440, 1464,
 		   1592, 1728, 0, 576, 581, 586, 625, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) },
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC),
+	  .vrefresh = 50, },
 	/* 31 - 1920x1080@50Hz */
 	{ DRM_MODE("1920x1080", DRM_MODE_TYPE_DRIVER, 148500, 1920, 2448,
 		   2492, 2640, 0, 1080, 1084, 1089, 1125, 0,
-		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC),
+	  .vrefresh = 50, },
 	/* 32 - 1920x1080@24Hz */
 	{ DRM_MODE("1920x1080", DRM_MODE_TYPE_DRIVER, 74250, 1920, 2558,
 		   2602, 2750, 0, 1080, 1084, 1089, 1125, 0,
-		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC),
+	  .vrefresh = 24, },
 	/* 33 - 1920x1080@25Hz */
 	{ DRM_MODE("1920x1080", DRM_MODE_TYPE_DRIVER, 74250, 1920, 2448,
 		   2492, 2640, 0, 1080, 1084, 1089, 1125, 0,
-		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC),
+	  .vrefresh = 25, },
 	/* 34 - 1920x1080@30Hz */
 	{ DRM_MODE("1920x1080", DRM_MODE_TYPE_DRIVER, 74250, 1920, 2008,
 		   2052, 2200, 0, 1080, 1084, 1089, 1125, 0,
-		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC),
+	  .vrefresh = 30, },
 	/* 35 - 2880x480@60Hz */
 	{ DRM_MODE("2880x480", DRM_MODE_TYPE_DRIVER, 108000, 2880, 2944,
 		   3192, 3432, 0, 480, 489, 495, 525, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) },
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC),
+	  .vrefresh = 60, },
 	/* 36 - 2880x480@60Hz */
 	{ DRM_MODE("2880x480", DRM_MODE_TYPE_DRIVER, 108000, 2880, 2944,
 		   3192, 3432, 0, 480, 489, 495, 525, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) },
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC),
+	  .vrefresh = 60, },
 	/* 37 - 2880x576@50Hz */
 	{ DRM_MODE("2880x576", DRM_MODE_TYPE_DRIVER, 108000, 2880, 2928,
 		   3184, 3456, 0, 576, 581, 586, 625, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) },
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC),
+	  .vrefresh = 50, },
 	/* 38 - 2880x576@50Hz */
 	{ DRM_MODE("2880x576", DRM_MODE_TYPE_DRIVER, 108000, 2880, 2928,
 		   3184, 3456, 0, 576, 581, 586, 625, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) },
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC),
+	  .vrefresh = 50, },
 	/* 39 - 1920x1080i@50Hz */
 	{ DRM_MODE("1920x1080i", DRM_MODE_TYPE_DRIVER, 72000, 1920, 1952,
 		   2120, 2304, 0, 1080, 1126, 1136, 1250, 0,
 		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_NVSYNC |
-			DRM_MODE_FLAG_INTERLACE) },
+			DRM_MODE_FLAG_INTERLACE),
+	  .vrefresh = 50, },
 	/* 40 - 1920x1080i@100Hz */
 	{ DRM_MODE("1920x1080i", DRM_MODE_TYPE_DRIVER, 148500, 1920, 2448,
 		   2492, 2640, 0, 1080, 1084, 1094, 1125, 0,
 		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC |
-			DRM_MODE_FLAG_INTERLACE) },
+			DRM_MODE_FLAG_INTERLACE),
+	  .vrefresh = 100, },
 	/* 41 - 1280x720@100Hz */
 	{ DRM_MODE("1280x720", DRM_MODE_TYPE_DRIVER, 148500, 1280, 1720,
 		   1760, 1980, 0, 720, 725, 730, 750, 0,
-		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC),
+	  .vrefresh = 100, },
 	/* 42 - 720x576@100Hz */
 	{ DRM_MODE("720x576", DRM_MODE_TYPE_DRIVER, 54000, 720, 732,
 		   796, 864, 0, 576, 581, 586, 625, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) },
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC),
+	  .vrefresh = 100, },
 	/* 43 - 720x576@100Hz */
 	{ DRM_MODE("720x576", DRM_MODE_TYPE_DRIVER, 54000, 720, 732,
 		   796, 864, 0, 576, 581, 586, 625, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) },
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC),
+	  .vrefresh = 100, },
 	/* 44 - 1440x576i@100Hz */
 	{ DRM_MODE("1440x576", DRM_MODE_TYPE_DRIVER, 54000, 1440, 1464,
 		   1590, 1728, 0, 576, 580, 586, 625, 0,
 		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC |
-			DRM_MODE_FLAG_DBLCLK) },
+			DRM_MODE_FLAG_DBLCLK),
+	  .vrefresh = 100, },
 	/* 45 - 1440x576i@100Hz */
 	{ DRM_MODE("1440x576", DRM_MODE_TYPE_DRIVER, 54000, 1440, 1464,
 		   1590, 1728, 0, 576, 580, 586, 625, 0,
 		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC |
-			DRM_MODE_FLAG_DBLCLK) },
+			DRM_MODE_FLAG_DBLCLK),
+	  .vrefresh = 100, },
 	/* 46 - 1920x1080i@120Hz */
 	{ DRM_MODE("1920x1080i", DRM_MODE_TYPE_DRIVER, 148500, 1920, 2008,
 		   2052, 2200, 0, 1080, 1084, 1094, 1125, 0,
 		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC |
-			DRM_MODE_FLAG_INTERLACE) },
+			DRM_MODE_FLAG_INTERLACE),
+	  .vrefresh = 120, },
 	/* 47 - 1280x720@120Hz */
 	{ DRM_MODE("1280x720", DRM_MODE_TYPE_DRIVER, 148500, 1280, 1390,
 		   1430, 1650, 0, 720, 725, 730, 750, 0,
-		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC),
+	  .vrefresh = 120, },
 	/* 48 - 720x480@120Hz */
 	{ DRM_MODE("720x480", DRM_MODE_TYPE_DRIVER, 54000, 720, 736,
 		   798, 858, 0, 480, 489, 495, 525, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) },
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC),
+	  .vrefresh = 120, },
 	/* 49 - 720x480@120Hz */
 	{ DRM_MODE("720x480", DRM_MODE_TYPE_DRIVER, 54000, 720, 736,
 		   798, 858, 0, 480, 489, 495, 525, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) },
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC),
+	  .vrefresh = 120, },
 	/* 50 - 1440x480i@120Hz */
 	{ DRM_MODE("1440x480i", DRM_MODE_TYPE_DRIVER, 54000, 1440, 1478,
 		   1602, 1716, 0, 480, 488, 494, 525, 0,
 		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC |
-			DRM_MODE_FLAG_INTERLACE | DRM_MODE_FLAG_DBLCLK) },
+			DRM_MODE_FLAG_INTERLACE | DRM_MODE_FLAG_DBLCLK),
+	  .vrefresh = 120, },
 	/* 51 - 1440x480i@120Hz */
 	{ DRM_MODE("1440x480i", DRM_MODE_TYPE_DRIVER, 54000, 1440, 1478,
 		   1602, 1716, 0, 480, 488, 494, 525, 0,
 		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC |
-			DRM_MODE_FLAG_INTERLACE | DRM_MODE_FLAG_DBLCLK) },
+			DRM_MODE_FLAG_INTERLACE | DRM_MODE_FLAG_DBLCLK),
+	  .vrefresh = 120, },
 	/* 52 - 720x576@200Hz */
 	{ DRM_MODE("720x576", DRM_MODE_TYPE_DRIVER, 108000, 720, 732,
 		   796, 864, 0, 576, 581, 586, 625, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) },
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC),
+	  .vrefresh = 200, },
 	/* 53 - 720x576@200Hz */
 	{ DRM_MODE("720x576", DRM_MODE_TYPE_DRIVER, 108000, 720, 732,
 		   796, 864, 0, 576, 581, 586, 625, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) },
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC),
+	  .vrefresh = 200, },
 	/* 54 - 1440x576i@200Hz */
 	{ DRM_MODE("1440x576i", DRM_MODE_TYPE_DRIVER, 108000, 1440, 1464,
 		   1590, 1728, 0, 576, 580, 586, 625, 0,
 		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC |
-			DRM_MODE_FLAG_INTERLACE | DRM_MODE_FLAG_DBLCLK) },
+			DRM_MODE_FLAG_INTERLACE | DRM_MODE_FLAG_DBLCLK),
+	  .vrefresh = 200, },
 	/* 55 - 1440x576i@200Hz */
 	{ DRM_MODE("1440x576i", DRM_MODE_TYPE_DRIVER, 108000, 1440, 1464,
 		   1590, 1728, 0, 576, 580, 586, 625, 0,
 		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC |
-			DRM_MODE_FLAG_INTERLACE | DRM_MODE_FLAG_DBLCLK) },
+			DRM_MODE_FLAG_INTERLACE | DRM_MODE_FLAG_DBLCLK),
+	  .vrefresh = 200, },
 	/* 56 - 720x480@240Hz */
 	{ DRM_MODE("720x480", DRM_MODE_TYPE_DRIVER, 108000, 720, 736,
 		   798, 858, 0, 480, 489, 495, 525, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) },
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC),
+	  .vrefresh = 240, },
 	/* 57 - 720x480@240Hz */
 	{ DRM_MODE("720x480", DRM_MODE_TYPE_DRIVER, 108000, 720, 736,
 		   798, 858, 0, 480, 489, 495, 525, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) },
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC),
+	  .vrefresh = 240, },
 	/* 58 - 1440x480i@240 */
 	{ DRM_MODE("1440x480i", DRM_MODE_TYPE_DRIVER, 108000, 1440, 1478,
 		   1602, 1716, 0, 480, 488, 494, 525, 0,
 		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC |
-			DRM_MODE_FLAG_INTERLACE | DRM_MODE_FLAG_DBLCLK) },
+			DRM_MODE_FLAG_INTERLACE | DRM_MODE_FLAG_DBLCLK),
+	  .vrefresh = 240, },
 	/* 59 - 1440x480i@240 */
 	{ DRM_MODE("1440x480i", DRM_MODE_TYPE_DRIVER, 108000, 1440, 1478,
 		   1602, 1716, 0, 480, 488, 494, 525, 0,
 		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC |
-			DRM_MODE_FLAG_INTERLACE | DRM_MODE_FLAG_DBLCLK) },
+			DRM_MODE_FLAG_INTERLACE | DRM_MODE_FLAG_DBLCLK),
+	  .vrefresh = 240, },
 	/* 60 - 1280x720@24Hz */
 	{ DRM_MODE("1280x720", DRM_MODE_TYPE_DRIVER, 59400, 1280, 3040,
 		   3080, 3300, 0, 720, 725, 730, 750, 0,
-		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC),
+	  .vrefresh = 24, },
 	/* 61 - 1280x720@25Hz */
 	{ DRM_MODE("1280x720", DRM_MODE_TYPE_DRIVER, 74250, 1280, 3700,
 		   3740, 3960, 0, 720, 725, 730, 750, 0,
-		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC),
+	  .vrefresh = 25, },
 	/* 62 - 1280x720@30Hz */
 	{ DRM_MODE("1280x720", DRM_MODE_TYPE_DRIVER, 74250, 1280, 3040,
 		   3080, 3300, 0, 720, 725, 730, 750, 0,
-		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC),
+	  .vrefresh = 30, },
 	/* 63 - 1920x1080@120Hz */
 	{ DRM_MODE("1920x1080", DRM_MODE_TYPE_DRIVER, 297000, 1920, 2008,
 		   2052, 2200, 0, 1080, 1084, 1089, 1125, 0,
-		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC),
+	 .vrefresh = 120, },
 	/* 64 - 1920x1080@100Hz */
 	{ DRM_MODE("1920x1080", DRM_MODE_TYPE_DRIVER, 297000, 1920, 2448,
 		   2492, 2640, 0, 1080, 1084, 1094, 1125, 0,
-		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC),
+	 .vrefresh = 100, },
 };
 
 /*** DDC fetch and block validation ***/
@@ -2266,13 +2330,34 @@ EXPORT_SYMBOL(drm_find_cea_extension);
  */
 u8 drm_match_cea_mode(const struct drm_display_mode *to_match)
 {
-	struct drm_display_mode *cea_mode;
 	u8 mode;
 
+	if (!to_match->clock)
+		return 0;
+
 	for (mode = 0; mode < ARRAY_SIZE(edid_cea_modes); mode++) {
-		cea_mode = (struct drm_display_mode *)&edid_cea_modes[mode];
+		const struct drm_display_mode *cea_mode = &edid_cea_modes[mode];
+		unsigned int clock1, clock2;
+
+		clock1 = clock2 = cea_mode->clock;
 
-		if (drm_mode_equal(to_match, cea_mode))
+		/* Check both 60Hz and 59.94Hz */
+		if (cea_mode->vrefresh % 6 == 0) {
+			/*
+			 * edid_cea_modes contains the 59.94Hz
+			 * variant for 240 and 480 line modes,
+			 * and the 60Hz variant otherwise.
+			 */
+			if (cea_mode->vdisplay == 240 ||
+			    cea_mode->vdisplay == 480)
+				clock1 = clock1 * 1001 / 1000;
+			else
+				clock2 = DIV_ROUND_UP(clock2 * 1000, 1001);
+		}
+
+		if ((KHZ2PICOS(to_match->clock) == KHZ2PICOS(clock1) ||
+		     KHZ2PICOS(to_match->clock) == KHZ2PICOS(clock2)) &&
+		    drm_mode_equal_no_clocks(to_match, cea_mode))
 			return mode + 1;
 	}
 	return 0;
@@ -2294,6 +2379,7 @@ do_cea_modes (struct drm_connector *connector, u8 *db, u8 len)
 			newmode = drm_mode_duplicate(dev,
 						     &edid_cea_modes[cea_mode]);
 			if (newmode) {
+				newmode->vrefresh = 0;
 				drm_mode_probed_add(connector, newmode);
 				modes++;
 			}
@@ -2511,6 +2597,65 @@ void drm_edid_to_eld(struct drm_connector *connector, struct edid *edid)
 EXPORT_SYMBOL(drm_edid_to_eld);
 
 /**
+ * drm_edid_to_sad - extracts SADs from EDID
+ * @edid: EDID to parse
+ * @sads: pointer that will be set to the extracted SADs
+ *
+ * Looks for CEA EDID block and extracts SADs (Short Audio Descriptors) from it.
+ * Note: returned pointer needs to be kfreed
+ *
+ * Return number of found SADs or negative number on error.
+ */
+int drm_edid_to_sad(struct edid *edid, struct cea_sad **sads)
+{
+	int count = 0;
+	int i, start, end, dbl;
+	u8 *cea;
+
+	cea = drm_find_cea_extension(edid);
+	if (!cea) {
+		DRM_DEBUG_KMS("SAD: no CEA Extension found\n");
+		return -ENOENT;
+	}
+
+	if (cea_revision(cea) < 3) {
+		DRM_DEBUG_KMS("SAD: wrong CEA revision\n");
+		return -ENOTSUPP;
+	}
+
+	if (cea_db_offsets(cea, &start, &end)) {
+		DRM_DEBUG_KMS("SAD: invalid data block offsets\n");
+		return -EPROTO;
+	}
+
+	for_each_cea_db(cea, i, start, end) {
+		u8 *db = &cea[i];
+
+		if (cea_db_tag(db) == AUDIO_BLOCK) {
+			int j;
+			dbl = cea_db_payload_len(db);
+
+			count = dbl / 3; /* SAD is 3B */
+			*sads = kcalloc(count, sizeof(**sads), GFP_KERNEL);
+			if (!*sads)
+				return -ENOMEM;
+			for (j = 0; j < count; j++) {
+				u8 *sad = &db[1 + j * 3];
+
+				(*sads)[j].format = (sad[0] & 0x78) >> 3;
+				(*sads)[j].channels = sad[0] & 0x7;
+				(*sads)[j].freq = sad[1] & 0x7F;
+				(*sads)[j].byte2 = sad[2];
+			}
+			break;
+		}
+	}
+
+	return count;
+}
+EXPORT_SYMBOL(drm_edid_to_sad);
+
+/**
  * drm_av_sync_delay - HDMI/DP sink audio-video sync delay in millisecond
  * @connector: connector associated with the HDMI/DP sink
  * @mode: the display mode
diff --git a/drivers/gpu/drm/drm_edid_load.c b/drivers/gpu/drm/drm_edid_load.c
index 38d3943f72de..fa445dd4dc00 100644
--- a/drivers/gpu/drm/drm_edid_load.c
+++ b/drivers/gpu/drm/drm_edid_load.c
@@ -31,10 +31,11 @@ module_param_string(edid_firmware, edid_firmware, sizeof(edid_firmware), 0644);
 MODULE_PARM_DESC(edid_firmware, "Do not probe monitor, use specified EDID blob "
 	"from built-in data or /lib/firmware instead. ");
 
-#define GENERIC_EDIDS 4
+#define GENERIC_EDIDS 5
 static char *generic_edid_name[GENERIC_EDIDS] = {
 	"edid/1024x768.bin",
 	"edid/1280x1024.bin",
+	"edid/1600x1200.bin",
 	"edid/1680x1050.bin",
 	"edid/1920x1080.bin",
 };
@@ -79,6 +80,24 @@ static u8 generic_edid[GENERIC_EDIDS][128] = {
 	{
 	0x00, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0x00,
 	0x31, 0xd8, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+	0x05, 0x16, 0x01, 0x03, 0x6d, 0x37, 0x29, 0x78,
+	0xea, 0x5e, 0xc0, 0xa4, 0x59, 0x4a, 0x98, 0x25,
+	0x20, 0x50, 0x54, 0x00, 0x00, 0x00, 0xa9, 0x40,
+	0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01,
+	0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x48, 0x3f,
+	0x40, 0x30, 0x62, 0xb0, 0x32, 0x40, 0x40, 0xc0,
+	0x13, 0x00, 0x2b, 0xa0, 0x21, 0x00, 0x00, 0x1e,
+	0x00, 0x00, 0x00, 0xff, 0x00, 0x4c, 0x69, 0x6e,
+	0x75, 0x78, 0x20, 0x23, 0x30, 0x0a, 0x20, 0x20,
+	0x20, 0x20, 0x00, 0x00, 0x00, 0xfd, 0x00, 0x3b,
+	0x3d, 0x4a, 0x4c, 0x11, 0x00, 0x0a, 0x20, 0x20,
+	0x20, 0x20, 0x20, 0x20, 0x00, 0x00, 0x00, 0xfc,
+	0x00, 0x4c, 0x69, 0x6e, 0x75, 0x78, 0x20, 0x55,
+	0x58, 0x47, 0x41, 0x0a, 0x20, 0x20, 0x00, 0x9d,
+	},
+	{
+	0x00, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0x00,
+	0x31, 0xd8, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
 	0x05, 0x16, 0x01, 0x03, 0x6d, 0x2b, 0x1b, 0x78,
 	0xea, 0x5e, 0xc0, 0xa4, 0x59, 0x4a, 0x98, 0x25,
 	0x20, 0x50, 0x54, 0x00, 0x00, 0x00, 0xb3, 0x00,
diff --git a/drivers/gpu/drm/drm_fb_helper.c b/drivers/gpu/drm/drm_fb_helper.c
index 892ff9f95975..b78cbe74dadf 100644
--- a/drivers/gpu/drm/drm_fb_helper.c
+++ b/drivers/gpu/drm/drm_fb_helper.c
@@ -1398,7 +1398,7 @@ static void drm_setup_crtcs(struct drm_fb_helper *fb_helper)
 	struct drm_mode_set *modeset;
 	bool *enabled;
 	int width, height;
-	int i, ret;
+	int i;
 
 	DRM_DEBUG_KMS("\n");
 
@@ -1419,16 +1419,23 @@ static void drm_setup_crtcs(struct drm_fb_helper *fb_helper)
 
 	drm_enable_connectors(fb_helper, enabled);
 
-	ret = drm_target_cloned(fb_helper, modes, enabled, width, height);
-	if (!ret) {
-		ret = drm_target_preferred(fb_helper, modes, enabled, width, height);
-		if (!ret)
+	if (!(fb_helper->funcs->initial_config &&
+	      fb_helper->funcs->initial_config(fb_helper, crtcs, modes,
+					       enabled, width, height))) {
+		memset(modes, 0, dev->mode_config.num_connector*sizeof(modes[0]));
+		memset(crtcs, 0, dev->mode_config.num_connector*sizeof(crtcs[0]));
+
+		if (!drm_target_cloned(fb_helper,
+				       modes, enabled, width, height) &&
+		    !drm_target_preferred(fb_helper,
+					  modes, enabled, width, height))
 			DRM_ERROR("Unable to find initial modes\n");
-	}
 
-	DRM_DEBUG_KMS("picking CRTCs for %dx%d config\n", width, height);
+		DRM_DEBUG_KMS("picking CRTCs for %dx%d config\n",
+			      width, height);
 
-	drm_pick_crtcs(fb_helper, crtcs, modes, 0, width, height);
+		drm_pick_crtcs(fb_helper, crtcs, modes, 0, width, height);
+	}
 
 	/* need to set the modesets up here for use later */
 	/* fill out the connector<->crtc mappings into the modesets */
diff --git a/drivers/gpu/drm/drm_gem.c b/drivers/gpu/drm/drm_gem.c
index af779ae19ebf..cf919e36e8ae 100644
--- a/drivers/gpu/drm/drm_gem.c
+++ b/drivers/gpu/drm/drm_gem.c
@@ -205,11 +205,11 @@ static void
 drm_gem_remove_prime_handles(struct drm_gem_object *obj, struct drm_file *filp)
 {
 	if (obj->import_attach) {
-		drm_prime_remove_imported_buf_handle(&filp->prime,
+		drm_prime_remove_buf_handle(&filp->prime,
 				obj->import_attach->dmabuf);
 	}
 	if (obj->export_dma_buf) {
-		drm_prime_remove_imported_buf_handle(&filp->prime,
+		drm_prime_remove_buf_handle(&filp->prime,
 				obj->export_dma_buf);
 	}
 }
diff --git a/drivers/gpu/drm/drm_modes.c b/drivers/gpu/drm/drm_modes.c
index f83f0719922e..faa79df02648 100644
--- a/drivers/gpu/drm/drm_modes.c
+++ b/drivers/gpu/drm/drm_modes.c
@@ -848,6 +848,26 @@ bool drm_mode_equal(const struct drm_display_mode *mode1, const struct drm_displ
 	} else if (mode1->clock != mode2->clock)
 		return false;
 
+	return drm_mode_equal_no_clocks(mode1, mode2);
+}
+EXPORT_SYMBOL(drm_mode_equal);
+
+/**
+ * drm_mode_equal_no_clocks - test modes for equality
+ * @mode1: first mode
+ * @mode2: second mode
+ *
+ * LOCKING:
+ * None.
+ *
+ * Check to see if @mode1 and @mode2 are equivalent, but
+ * don't check the pixel clocks.
+ *
+ * RETURNS:
+ * True if the modes are equal, false otherwise.
+ */
+bool drm_mode_equal_no_clocks(const struct drm_display_mode *mode1, const struct drm_display_mode *mode2)
+{
 	if (mode1->hdisplay == mode2->hdisplay &&
 	    mode1->hsync_start == mode2->hsync_start &&
 	    mode1->hsync_end == mode2->hsync_end &&
@@ -863,7 +883,7 @@ bool drm_mode_equal(const struct drm_display_mode *mode1, const struct drm_displ
 
 	return false;
 }
-EXPORT_SYMBOL(drm_mode_equal);
+EXPORT_SYMBOL(drm_mode_equal_no_clocks);
 
 /**
  * drm_mode_validate_size - make sure modes adhere to size constraints
diff --git a/drivers/gpu/drm/drm_pci.c b/drivers/gpu/drm/drm_pci.c
index bd719e936e13..14194b6ef644 100644
--- a/drivers/gpu/drm/drm_pci.c
+++ b/drivers/gpu/drm/drm_pci.c
@@ -152,7 +152,7 @@ static const char *drm_pci_get_name(struct drm_device *dev)
 	return pdriver->name;
 }
 
-int drm_pci_set_busid(struct drm_device *dev, struct drm_master *master)
+static int drm_pci_set_busid(struct drm_device *dev, struct drm_master *master)
 {
 	int len, ret;
 	struct pci_driver *pdriver = dev->driver->kdriver.pci;
@@ -194,9 +194,9 @@ err:
 	return ret;
 }
 
-int drm_pci_set_unique(struct drm_device *dev,
-		       struct drm_master *master,
-		       struct drm_unique *u)
+static int drm_pci_set_unique(struct drm_device *dev,
+			      struct drm_master *master,
+			      struct drm_unique *u)
 {
 	int domain, bus, slot, func, ret;
 	const char *bus_name;
@@ -266,7 +266,7 @@ static int drm_pci_irq_by_busid(struct drm_device *dev, struct drm_irq_busid *p)
 	return 0;
 }
 
-int drm_pci_agp_init(struct drm_device *dev)
+static int drm_pci_agp_init(struct drm_device *dev)
 {
 	if (drm_core_has_AGP(dev)) {
 		if (drm_pci_device_is_agp(dev))
diff --git a/drivers/gpu/drm/drm_prime.c b/drivers/gpu/drm/drm_prime.c
index 366910ddcfcb..dcde35231e25 100644
--- a/drivers/gpu/drm/drm_prime.c
+++ b/drivers/gpu/drm/drm_prime.c
@@ -62,6 +62,7 @@ struct drm_prime_member {
 	struct dma_buf *dma_buf;
 	uint32_t handle;
 };
+static int drm_prime_add_buf_handle(struct drm_prime_file_private *prime_fpriv, struct dma_buf *dma_buf, uint32_t handle);
 
 static struct sg_table *drm_gem_map_dma_buf(struct dma_buf_attachment *attach,
 		enum dma_data_direction dir)
@@ -200,7 +201,8 @@ int drm_gem_prime_handle_to_fd(struct drm_device *dev,
 {
 	struct drm_gem_object *obj;
 	void *buf;
-	int ret;
+	int ret = 0;
+	struct dma_buf *dmabuf;
 
 	obj = drm_gem_object_lookup(dev, file_priv, handle);
 	if (!obj)
@@ -209,43 +211,44 @@ int drm_gem_prime_handle_to_fd(struct drm_device *dev,
 	mutex_lock(&file_priv->prime.lock);
 	/* re-export the original imported object */
 	if (obj->import_attach) {
-		get_dma_buf(obj->import_attach->dmabuf);
-		*prime_fd = dma_buf_fd(obj->import_attach->dmabuf, flags);
-		drm_gem_object_unreference_unlocked(obj);
-		mutex_unlock(&file_priv->prime.lock);
-		return 0;
+		dmabuf = obj->import_attach->dmabuf;
+		goto out_have_obj;
 	}
 
 	if (obj->export_dma_buf) {
-		get_dma_buf(obj->export_dma_buf);
-		*prime_fd = dma_buf_fd(obj->export_dma_buf, flags);
-		drm_gem_object_unreference_unlocked(obj);
-	} else {
-		buf = dev->driver->gem_prime_export(dev, obj, flags);
-		if (IS_ERR(buf)) {
-			/* normally the created dma-buf takes ownership of the ref,
-			 * but if that fails then drop the ref
-			 */
-			drm_gem_object_unreference_unlocked(obj);
-			mutex_unlock(&file_priv->prime.lock);
-			return PTR_ERR(buf);
-		}
-		obj->export_dma_buf = buf;
-		*prime_fd = dma_buf_fd(buf, flags);
+		dmabuf = obj->export_dma_buf;
+		goto out_have_obj;
+	}
+
+	buf = dev->driver->gem_prime_export(dev, obj, flags);
+	if (IS_ERR(buf)) {
+		/* normally the created dma-buf takes ownership of the ref,
+		 * but if that fails then drop the ref
+		 */
+		ret = PTR_ERR(buf);
+		goto out;
 	}
+	obj->export_dma_buf = buf;
+
 	/* if we've exported this buffer the cheat and add it to the import list
 	 * so we get the correct handle back
 	 */
-	ret = drm_prime_add_imported_buf_handle(&file_priv->prime,
-			obj->export_dma_buf, handle);
-	if (ret) {
-		drm_gem_object_unreference_unlocked(obj);
-		mutex_unlock(&file_priv->prime.lock);
-		return ret;
-	}
+	ret = drm_prime_add_buf_handle(&file_priv->prime,
+				       obj->export_dma_buf, handle);
+	if (ret)
+		goto out;
 
+	*prime_fd = dma_buf_fd(buf, flags);
 	mutex_unlock(&file_priv->prime.lock);
 	return 0;
+
+out_have_obj:
+	get_dma_buf(dmabuf);
+	*prime_fd = dma_buf_fd(dmabuf, flags);
+out:
+	drm_gem_object_unreference_unlocked(obj);
+	mutex_unlock(&file_priv->prime.lock);
+	return ret;
 }
 EXPORT_SYMBOL(drm_gem_prime_handle_to_fd);
 
@@ -268,7 +271,6 @@ struct drm_gem_object *drm_gem_prime_import(struct drm_device *dev,
 			 * refcount on gem itself instead of f_count of dmabuf.
 			 */
 			drm_gem_object_reference(obj);
-			dma_buf_put(dma_buf);
 			return obj;
 		}
 	}
@@ -277,6 +279,8 @@ struct drm_gem_object *drm_gem_prime_import(struct drm_device *dev,
 	if (IS_ERR(attach))
 		return ERR_PTR(PTR_ERR(attach));
 
+	get_dma_buf(dma_buf);
+
 	sgt = dma_buf_map_attachment(attach, DMA_BIDIRECTIONAL);
 	if (IS_ERR_OR_NULL(sgt)) {
 		ret = PTR_ERR(sgt);
@@ -297,6 +301,8 @@ fail_unmap:
 	dma_buf_unmap_attachment(attach, sgt, DMA_BIDIRECTIONAL);
 fail_detach:
 	dma_buf_detach(dma_buf, attach);
+	dma_buf_put(dma_buf);
+
 	return ERR_PTR(ret);
 }
 EXPORT_SYMBOL(drm_gem_prime_import);
@@ -314,7 +320,7 @@ int drm_gem_prime_fd_to_handle(struct drm_device *dev,
 
 	mutex_lock(&file_priv->prime.lock);
 
-	ret = drm_prime_lookup_imported_buf_handle(&file_priv->prime,
+	ret = drm_prime_lookup_buf_handle(&file_priv->prime,
 			dma_buf, handle);
 	if (!ret) {
 		ret = 0;
@@ -333,12 +339,15 @@ int drm_gem_prime_fd_to_handle(struct drm_device *dev,
 	if (ret)
 		goto out_put;
 
-	ret = drm_prime_add_imported_buf_handle(&file_priv->prime,
+	ret = drm_prime_add_buf_handle(&file_priv->prime,
 			dma_buf, *handle);
 	if (ret)
 		goto fail;
 
 	mutex_unlock(&file_priv->prime.lock);
+
+	dma_buf_put(dma_buf);
+
 	return 0;
 
 fail:
@@ -401,21 +410,17 @@ int drm_prime_fd_to_handle_ioctl(struct drm_device *dev, void *data,
 struct sg_table *drm_prime_pages_to_sg(struct page **pages, int nr_pages)
 {
 	struct sg_table *sg = NULL;
-	struct scatterlist *iter;
-	int i;
 	int ret;
 
 	sg = kmalloc(sizeof(struct sg_table), GFP_KERNEL);
 	if (!sg)
 		goto out;
 
-	ret = sg_alloc_table(sg, nr_pages, GFP_KERNEL);
+	ret = sg_alloc_table_from_pages(sg, pages, nr_pages, 0,
+				nr_pages << PAGE_SHIFT, GFP_KERNEL);
 	if (ret)
 		goto out;
 
-	for_each_sg(sg->sgl, iter, nr_pages, i)
-		sg_set_page(iter, pages[i], PAGE_SIZE, 0);
-
 	return sg;
 out:
 	kfree(sg);
@@ -483,15 +488,12 @@ EXPORT_SYMBOL(drm_prime_init_file_private);
 
 void drm_prime_destroy_file_private(struct drm_prime_file_private *prime_fpriv)
 {
-	struct drm_prime_member *member, *safe;
-	list_for_each_entry_safe(member, safe, &prime_fpriv->head, entry) {
-		list_del(&member->entry);
-		kfree(member);
-	}
+	/* by now drm_gem_release should've made sure the list is empty */
+	WARN_ON(!list_empty(&prime_fpriv->head));
 }
 EXPORT_SYMBOL(drm_prime_destroy_file_private);
 
-int drm_prime_add_imported_buf_handle(struct drm_prime_file_private *prime_fpriv, struct dma_buf *dma_buf, uint32_t handle)
+static int drm_prime_add_buf_handle(struct drm_prime_file_private *prime_fpriv, struct dma_buf *dma_buf, uint32_t handle)
 {
 	struct drm_prime_member *member;
 
@@ -499,14 +501,14 @@ int drm_prime_add_imported_buf_handle(struct drm_prime_file_private *prime_fpriv
 	if (!member)
 		return -ENOMEM;
 
+	get_dma_buf(dma_buf);
 	member->dma_buf = dma_buf;
 	member->handle = handle;
 	list_add(&member->entry, &prime_fpriv->head);
 	return 0;
 }
-EXPORT_SYMBOL(drm_prime_add_imported_buf_handle);
 
-int drm_prime_lookup_imported_buf_handle(struct drm_prime_file_private *prime_fpriv, struct dma_buf *dma_buf, uint32_t *handle)
+int drm_prime_lookup_buf_handle(struct drm_prime_file_private *prime_fpriv, struct dma_buf *dma_buf, uint32_t *handle)
 {
 	struct drm_prime_member *member;
 
@@ -518,19 +520,20 @@ int drm_prime_lookup_imported_buf_handle(struct drm_prime_file_private *prime_fp
 	}
 	return -ENOENT;
 }
-EXPORT_SYMBOL(drm_prime_lookup_imported_buf_handle);
+EXPORT_SYMBOL(drm_prime_lookup_buf_handle);
 
-void drm_prime_remove_imported_buf_handle(struct drm_prime_file_private *prime_fpriv, struct dma_buf *dma_buf)
+void drm_prime_remove_buf_handle(struct drm_prime_file_private *prime_fpriv, struct dma_buf *dma_buf)
 {
 	struct drm_prime_member *member, *safe;
 
 	mutex_lock(&prime_fpriv->lock);
 	list_for_each_entry_safe(member, safe, &prime_fpriv->head, entry) {
 		if (member->dma_buf == dma_buf) {
+			dma_buf_put(dma_buf);
 			list_del(&member->entry);
 			kfree(member);
 		}
 	}
 	mutex_unlock(&prime_fpriv->lock);
 }
-EXPORT_SYMBOL(drm_prime_remove_imported_buf_handle);
+EXPORT_SYMBOL(drm_prime_remove_buf_handle);
diff --git a/drivers/gpu/drm/drm_vm.c b/drivers/gpu/drm/drm_vm.c
index db7bd292410b..1d4f7c9fe661 100644
--- a/drivers/gpu/drm/drm_vm.c
+++ b/drivers/gpu/drm/drm_vm.c
@@ -422,6 +422,7 @@ void drm_vm_open_locked(struct drm_device *dev,
 		list_add(&vma_entry->head, &dev->vmalist);
 	}
 }
+EXPORT_SYMBOL_GPL(drm_vm_open_locked);
 
 static void drm_vm_open(struct vm_area_struct *vma)
 {
diff --git a/drivers/gpu/drm/exynos/Kconfig b/drivers/gpu/drm/exynos/Kconfig
index 046bcda36abe..772c62a6e2ac 100644
--- a/drivers/gpu/drm/exynos/Kconfig
+++ b/drivers/gpu/drm/exynos/Kconfig
@@ -24,7 +24,9 @@ config DRM_EXYNOS_DMABUF
 
 config DRM_EXYNOS_FIMD
 	bool "Exynos DRM FIMD"
-	depends on DRM_EXYNOS && !FB_S3C && !ARCH_MULTIPLATFORM
+	depends on OF && DRM_EXYNOS && !FB_S3C && !ARCH_MULTIPLATFORM
+	select FB_MODE_HELPERS
+	select VIDEOMODE_HELPERS
 	help
 	  Choose this option if you want to use Exynos FIMD for DRM.
 
@@ -54,7 +56,7 @@ config DRM_EXYNOS_IPP
 
 config DRM_EXYNOS_FIMC
 	bool "Exynos DRM FIMC"
-	depends on DRM_EXYNOS_IPP
+	depends on DRM_EXYNOS_IPP && MFD_SYSCON && OF
 	help
 	  Choose this option if you want to use Exynos FIMC for DRM.
 
diff --git a/drivers/gpu/drm/exynos/exynos_drm_connector.c b/drivers/gpu/drm/exynos/exynos_drm_connector.c
index 4c5b6859c9ea..8bcc13ac9f73 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_connector.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_connector.c
@@ -124,7 +124,7 @@ static int exynos_drm_connector_get_modes(struct drm_connector *connector)
 		}
 
 		count = drm_add_edid_modes(connector, edid);
-		if (count < 0) {
+		if (!count) {
 			DRM_ERROR("Add edid modes failed %d\n", count);
 			goto out;
 		}
diff --git a/drivers/gpu/drm/exynos/exynos_drm_dmabuf.c b/drivers/gpu/drm/exynos/exynos_drm_dmabuf.c
index ba0a3aa78547..ff7f2a886a34 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_dmabuf.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_dmabuf.c
@@ -235,7 +235,6 @@ struct drm_gem_object *exynos_dmabuf_prime_import(struct drm_device *drm_dev,
 			 * refcount on gem itself instead of f_count of dmabuf.
 			 */
 			drm_gem_object_reference(obj);
-			dma_buf_put(dma_buf);
 			return obj;
 		}
 	}
@@ -244,6 +243,7 @@ struct drm_gem_object *exynos_dmabuf_prime_import(struct drm_device *drm_dev,
 	if (IS_ERR(attach))
 		return ERR_PTR(-EINVAL);
 
+	get_dma_buf(dma_buf);
 
 	sgt = dma_buf_map_attachment(attach, DMA_BIDIRECTIONAL);
 	if (IS_ERR_OR_NULL(sgt)) {
@@ -298,6 +298,8 @@ err_unmap_attach:
 	dma_buf_unmap_attachment(attach, sgt, DMA_BIDIRECTIONAL);
 err_buf_detach:
 	dma_buf_detach(dma_buf, attach);
+	dma_buf_put(dma_buf);
+
 	return ERR_PTR(ret);
 }
 
diff --git a/drivers/gpu/drm/exynos/exynos_drm_drv.c b/drivers/gpu/drm/exynos/exynos_drm_drv.c
index 3da5c2d214d8..ba6d995e4375 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_drv.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_drv.c
@@ -380,6 +380,10 @@ static int __init exynos_drm_init(void)
 	ret = platform_driver_register(&ipp_driver);
 	if (ret < 0)
 		goto out_ipp;
+
+	ret = exynos_platform_device_ipp_register();
+	if (ret < 0)
+		goto out_ipp_dev;
 #endif
 
 	ret = platform_driver_register(&exynos_drm_platform_driver);
@@ -388,7 +392,7 @@ static int __init exynos_drm_init(void)
 
 	exynos_drm_pdev = platform_device_register_simple("exynos-drm", -1,
 				NULL, 0);
-	if (IS_ERR_OR_NULL(exynos_drm_pdev)) {
+	if (IS_ERR(exynos_drm_pdev)) {
 		ret = PTR_ERR(exynos_drm_pdev);
 		goto out;
 	}
@@ -400,6 +404,8 @@ out:
 
 out_drm:
 #ifdef CONFIG_DRM_EXYNOS_IPP
+	exynos_platform_device_ipp_unregister();
+out_ipp_dev:
 	platform_driver_unregister(&ipp_driver);
 out_ipp:
 #endif
@@ -456,6 +462,7 @@ static void __exit exynos_drm_exit(void)
 	platform_driver_unregister(&exynos_drm_platform_driver);
 
 #ifdef CONFIG_DRM_EXYNOS_IPP
+	exynos_platform_device_ipp_unregister();
 	platform_driver_unregister(&ipp_driver);
 #endif
 
diff --git a/drivers/gpu/drm/exynos/exynos_drm_drv.h b/drivers/gpu/drm/exynos/exynos_drm_drv.h
index 4606fac7241a..680a7c1b9dea 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_drv.h
+++ b/drivers/gpu/drm/exynos/exynos_drm_drv.h
@@ -322,13 +322,23 @@ void exynos_drm_subdrv_close(struct drm_device *dev, struct drm_file *file);
  * this function registers exynos drm hdmi platform device. It ensures only one
  * instance of the device is created.
  */
-extern int exynos_platform_device_hdmi_register(void);
+int exynos_platform_device_hdmi_register(void);
 
 /*
  * this function unregisters exynos drm hdmi platform device if it exists.
  */
 void exynos_platform_device_hdmi_unregister(void);
 
+/*
+ * this function registers exynos drm ipp platform device.
+ */
+int exynos_platform_device_ipp_register(void);
+
+/*
+ * this function unregisters exynos drm ipp platform device if it exists.
+ */
+void exynos_platform_device_ipp_unregister(void);
+
 extern struct platform_driver fimd_driver;
 extern struct platform_driver hdmi_driver;
 extern struct platform_driver mixer_driver;
diff --git a/drivers/gpu/drm/exynos/exynos_drm_fimc.c b/drivers/gpu/drm/exynos/exynos_drm_fimc.c
index 411f69b76e84..773f583fa964 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_fimc.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_fimc.c
@@ -12,11 +12,12 @@
  *
  */
 #include <linux/kernel.h>
+#include <linux/mfd/syscon.h>
 #include <linux/module.h>
 #include <linux/platform_device.h>
+#include <linux/regmap.h>
 #include <linux/clk.h>
 #include <linux/pm_runtime.h>
-#include <plat/map-base.h>
 
 #include <drm/drmP.h>
 #include <drm/exynos_drm.h>
@@ -76,6 +77,27 @@ enum fimc_wb {
 	FIMC_WB_B,
 };
 
+enum {
+	FIMC_CLK_LCLK,
+	FIMC_CLK_GATE,
+	FIMC_CLK_WB_A,
+	FIMC_CLK_WB_B,
+	FIMC_CLK_MUX,
+	FIMC_CLK_PARENT,
+	FIMC_CLKS_MAX
+};
+
+static const char * const fimc_clock_names[] = {
+	[FIMC_CLK_LCLK]   = "sclk_fimc",
+	[FIMC_CLK_GATE]   = "fimc",
+	[FIMC_CLK_WB_A]   = "pxl_async0",
+	[FIMC_CLK_WB_B]   = "pxl_async1",
+	[FIMC_CLK_MUX]    = "mux",
+	[FIMC_CLK_PARENT] = "parent",
+};
+
+#define FIMC_DEFAULT_LCLK_FREQUENCY 133000000UL
+
 /*
  * A structure of scaler.
  *
@@ -119,28 +141,16 @@ struct fimc_capability {
 };
 
 /*
- * A structure of fimc driver data.
- *
- * @parent_clk: name of parent clock.
- */
-struct fimc_driverdata {
-	char	*parent_clk;
-};
-
-/*
  * A structure of fimc context.
  *
  * @ippdrv: prepare initialization using ippdrv.
  * @regs_res: register resources.
  * @regs: memory mapped io registers.
  * @lock: locking of operations.
- * @sclk_fimc_clk: fimc source clock.
- * @fimc_clk: fimc clock.
- * @wb_clk: writeback a clock.
- * @wb_b_clk: writeback b clock.
+ * @clocks: fimc clocks.
+ * @clk_frequency: LCLK clock frequency.
+ * @sysreg: handle to SYSREG block regmap.
  * @sc: scaler infomations.
- * @odr: ordering of YUV.
- * @ver: fimc version.
  * @pol: porarity of writeback.
  * @id: fimc id.
  * @irq: irq number.
@@ -151,12 +161,10 @@ struct fimc_context {
 	struct resource	*regs_res;
 	void __iomem	*regs;
 	struct mutex	lock;
-	struct clk	*sclk_fimc_clk;
-	struct clk	*fimc_clk;
-	struct clk	*wb_clk;
-	struct clk	*wb_b_clk;
+	struct clk	*clocks[FIMC_CLKS_MAX];
+	u32		clk_frequency;
+	struct regmap	*sysreg;
 	struct fimc_scaler	sc;
-	struct fimc_driverdata	*ddata;
 	struct exynos_drm_ipp_pol	pol;
 	int	id;
 	int	irq;
@@ -200,17 +208,13 @@ static void fimc_sw_reset(struct fimc_context *ctx)
 	fimc_write(0x0, EXYNOS_CIFCNTSEQ);
 }
 
-static void fimc_set_camblk_fimd0_wb(struct fimc_context *ctx)
+static int fimc_set_camblk_fimd0_wb(struct fimc_context *ctx)
 {
-	u32 camblk_cfg;
-
 	DRM_DEBUG_KMS("%s\n", __func__);
 
-	camblk_cfg = readl(SYSREG_CAMERA_BLK);
-	camblk_cfg &= ~(SYSREG_FIMD0WB_DEST_MASK);
-	camblk_cfg |= ctx->id << (SYSREG_FIMD0WB_DEST_SHIFT);
-
-	writel(camblk_cfg, SYSREG_CAMERA_BLK);
+	return regmap_update_bits(ctx->sysreg, SYSREG_CAMERA_BLK,
+				  SYSREG_FIMD0WB_DEST_MASK,
+				  ctx->id << SYSREG_FIMD0WB_DEST_SHIFT);
 }
 
 static void fimc_set_type_ctrl(struct fimc_context *ctx, enum fimc_wb wb)
@@ -1301,14 +1305,12 @@ static int fimc_clk_ctrl(struct fimc_context *ctx, bool enable)
 	DRM_DEBUG_KMS("%s:enable[%d]\n", __func__, enable);
 
 	if (enable) {
-		clk_enable(ctx->sclk_fimc_clk);
-		clk_enable(ctx->fimc_clk);
-		clk_enable(ctx->wb_clk);
+		clk_prepare_enable(ctx->clocks[FIMC_CLK_GATE]);
+		clk_prepare_enable(ctx->clocks[FIMC_CLK_WB_A]);
 		ctx->suspended = false;
 	} else {
-		clk_disable(ctx->sclk_fimc_clk);
-		clk_disable(ctx->fimc_clk);
-		clk_disable(ctx->wb_clk);
+		clk_disable_unprepare(ctx->clocks[FIMC_CLK_GATE]);
+		clk_disable_unprepare(ctx->clocks[FIMC_CLK_WB_A]);
 		ctx->suspended = true;
 	}
 
@@ -1613,7 +1615,11 @@ static int fimc_ippdrv_start(struct device *dev, enum drm_exynos_ipp_cmd cmd)
 		fimc_handle_lastend(ctx, true);
 
 		/* setup FIMD */
-		fimc_set_camblk_fimd0_wb(ctx);
+		ret = fimc_set_camblk_fimd0_wb(ctx);
+		if (ret < 0) {
+			dev_err(dev, "camblk setup failed.\n");
+			return ret;
+		}
 
 		set_wb.enable = 1;
 		set_wb.refresh = property->refresh_rate;
@@ -1713,76 +1719,118 @@ static void fimc_ippdrv_stop(struct device *dev, enum drm_exynos_ipp_cmd cmd)
 	fimc_write(cfg, EXYNOS_CIGCTRL);
 }
 
+static void fimc_put_clocks(struct fimc_context *ctx)
+{
+	int i;
+
+	for (i = 0; i < FIMC_CLKS_MAX; i++) {
+		if (IS_ERR(ctx->clocks[i]))
+			continue;
+		clk_put(ctx->clocks[i]);
+		ctx->clocks[i] = ERR_PTR(-EINVAL);
+	}
+}
+
+static int fimc_setup_clocks(struct fimc_context *ctx)
+{
+	struct device *fimc_dev = ctx->ippdrv.dev;
+	struct device *dev;
+	int ret, i;
+
+	for (i = 0; i < FIMC_CLKS_MAX; i++)
+		ctx->clocks[i] = ERR_PTR(-EINVAL);
+
+	for (i = 0; i < FIMC_CLKS_MAX; i++) {
+		if (i == FIMC_CLK_WB_A || i == FIMC_CLK_WB_B)
+			dev = fimc_dev->parent;
+		else
+			dev = fimc_dev;
+
+		ctx->clocks[i] = clk_get(dev, fimc_clock_names[i]);
+		if (IS_ERR(ctx->clocks[i])) {
+			if (i >= FIMC_CLK_MUX)
+				break;
+			ret = PTR_ERR(ctx->clocks[i]);
+			dev_err(fimc_dev, "failed to get clock: %s\n",
+						fimc_clock_names[i]);
+			goto e_clk_free;
+		}
+	}
+
+	/* Optional FIMC LCLK parent clock setting */
+	if (!IS_ERR(ctx->clocks[FIMC_CLK_PARENT])) {
+		ret = clk_set_parent(ctx->clocks[FIMC_CLK_MUX],
+				     ctx->clocks[FIMC_CLK_PARENT]);
+		if (ret < 0) {
+			dev_err(fimc_dev, "failed to set parent.\n");
+			goto e_clk_free;
+		}
+	}
+
+	ret = clk_set_rate(ctx->clocks[FIMC_CLK_LCLK], ctx->clk_frequency);
+	if (ret < 0)
+		goto e_clk_free;
+
+	ret = clk_prepare_enable(ctx->clocks[FIMC_CLK_LCLK]);
+	if (!ret)
+		return ret;
+e_clk_free:
+	fimc_put_clocks(ctx);
+	return ret;
+}
+
+static int fimc_parse_dt(struct fimc_context *ctx)
+{
+	struct device_node *node = ctx->ippdrv.dev->of_node;
+
+	/* Handle only devices that support the LCD Writeback data path */
+	if (!of_property_read_bool(node, "samsung,lcd-wb"))
+		return -ENODEV;
+
+	if (of_property_read_u32(node, "clock-frequency",
+					&ctx->clk_frequency))
+		ctx->clk_frequency = FIMC_DEFAULT_LCLK_FREQUENCY;
+
+	ctx->id = of_alias_get_id(node, "fimc");
+
+	if (ctx->id < 0) {
+		dev_err(ctx->ippdrv.dev, "failed to get node alias id.\n");
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
 static int fimc_probe(struct platform_device *pdev)
 {
 	struct device *dev = &pdev->dev;
 	struct fimc_context *ctx;
-	struct clk	*parent_clk;
 	struct resource *res;
 	struct exynos_drm_ippdrv *ippdrv;
-	struct exynos_drm_fimc_pdata *pdata;
-	struct fimc_driverdata *ddata;
 	int ret;
 
-	pdata = pdev->dev.platform_data;
-	if (!pdata) {
-		dev_err(dev, "no platform data specified.\n");
-		return -EINVAL;
+	if (!dev->of_node) {
+		dev_err(dev, "device tree node not found.\n");
+		return -ENODEV;
 	}
 
 	ctx = devm_kzalloc(dev, sizeof(*ctx), GFP_KERNEL);
 	if (!ctx)
 		return -ENOMEM;
 
-	ddata = (struct fimc_driverdata *)
-		platform_get_device_id(pdev)->driver_data;
-
-	/* clock control */
-	ctx->sclk_fimc_clk = devm_clk_get(dev, "sclk_fimc");
-	if (IS_ERR(ctx->sclk_fimc_clk)) {
-		dev_err(dev, "failed to get src fimc clock.\n");
-		return PTR_ERR(ctx->sclk_fimc_clk);
-	}
-	clk_enable(ctx->sclk_fimc_clk);
-
-	ctx->fimc_clk = devm_clk_get(dev, "fimc");
-	if (IS_ERR(ctx->fimc_clk)) {
-		dev_err(dev, "failed to get fimc clock.\n");
-		clk_disable(ctx->sclk_fimc_clk);
-		return PTR_ERR(ctx->fimc_clk);
-	}
-
-	ctx->wb_clk = devm_clk_get(dev, "pxl_async0");
-	if (IS_ERR(ctx->wb_clk)) {
-		dev_err(dev, "failed to get writeback a clock.\n");
-		clk_disable(ctx->sclk_fimc_clk);
-		return PTR_ERR(ctx->wb_clk);
-	}
-
-	ctx->wb_b_clk = devm_clk_get(dev, "pxl_async1");
-	if (IS_ERR(ctx->wb_b_clk)) {
-		dev_err(dev, "failed to get writeback b clock.\n");
-		clk_disable(ctx->sclk_fimc_clk);
-		return PTR_ERR(ctx->wb_b_clk);
-	}
+	ctx->ippdrv.dev = dev;
 
-	parent_clk = devm_clk_get(dev, ddata->parent_clk);
-
-	if (IS_ERR(parent_clk)) {
-		dev_err(dev, "failed to get parent clock.\n");
-		clk_disable(ctx->sclk_fimc_clk);
-		return PTR_ERR(parent_clk);
-	}
+	ret = fimc_parse_dt(ctx);
+	if (ret < 0)
+		return ret;
 
-	if (clk_set_parent(ctx->sclk_fimc_clk, parent_clk)) {
-		dev_err(dev, "failed to set parent.\n");
-		clk_disable(ctx->sclk_fimc_clk);
-		return -EINVAL;
+	ctx->sysreg = syscon_regmap_lookup_by_phandle(dev->of_node,
+						"samsung,sysreg");
+	if (IS_ERR(ctx->sysreg)) {
+		dev_err(dev, "syscon regmap lookup failed.\n");
+		return PTR_ERR(ctx->sysreg);
 	}
 
-	devm_clk_put(dev, parent_clk);
-	clk_set_rate(ctx->sclk_fimc_clk, pdata->clk_rate);
-
 	/* resource memory */
 	ctx->regs_res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
 	ctx->regs = devm_ioremap_resource(dev, ctx->regs_res);
@@ -1804,13 +1852,11 @@ static int fimc_probe(struct platform_device *pdev)
 		return ret;
 	}
 
-	/* context initailization */
-	ctx->id = pdev->id;
-	ctx->pol = pdata->pol;
-	ctx->ddata = ddata;
+	ret = fimc_setup_clocks(ctx);
+	if (ret < 0)
+		goto err_free_irq;
 
 	ippdrv = &ctx->ippdrv;
-	ippdrv->dev = dev;
 	ippdrv->ops[EXYNOS_DRM_OPS_SRC] = &fimc_src_ops;
 	ippdrv->ops[EXYNOS_DRM_OPS_DST] = &fimc_dst_ops;
 	ippdrv->check_property = fimc_ippdrv_check_property;
@@ -1820,7 +1866,7 @@ static int fimc_probe(struct platform_device *pdev)
 	ret = fimc_init_prop_list(ippdrv);
 	if (ret < 0) {
 		dev_err(dev, "failed to init property list.\n");
-		goto err_get_irq;
+		goto err_put_clk;
 	}
 
 	DRM_DEBUG_KMS("%s:id[%d]ippdrv[0x%x]\n", __func__, ctx->id,
@@ -1835,17 +1881,18 @@ static int fimc_probe(struct platform_device *pdev)
 	ret = exynos_drm_ippdrv_register(ippdrv);
 	if (ret < 0) {
 		dev_err(dev, "failed to register drm fimc device.\n");
-		goto err_ippdrv_register;
+		goto err_pm_dis;
 	}
 
 	dev_info(&pdev->dev, "drm fimc registered successfully.\n");
 
 	return 0;
 
-err_ippdrv_register:
-	devm_kfree(dev, ippdrv->prop_list);
+err_pm_dis:
 	pm_runtime_disable(dev);
-err_get_irq:
+err_put_clk:
+	fimc_put_clocks(ctx);
+err_free_irq:
 	free_irq(ctx->irq, ctx);
 
 	return ret;
@@ -1857,10 +1904,10 @@ static int fimc_remove(struct platform_device *pdev)
 	struct fimc_context *ctx = get_fimc_context(dev);
 	struct exynos_drm_ippdrv *ippdrv = &ctx->ippdrv;
 
-	devm_kfree(dev, ippdrv->prop_list);
 	exynos_drm_ippdrv_unregister(ippdrv);
 	mutex_destroy(&ctx->lock);
 
+	fimc_put_clocks(ctx);
 	pm_runtime_set_suspended(dev);
 	pm_runtime_disable(dev);
 
@@ -1915,36 +1962,22 @@ static int fimc_runtime_resume(struct device *dev)
 }
 #endif
 
-static struct fimc_driverdata exynos4210_fimc_data = {
-	.parent_clk = "mout_mpll",
-};
-
-static struct fimc_driverdata exynos4410_fimc_data = {
-	.parent_clk = "mout_mpll_user",
-};
-
-static struct platform_device_id fimc_driver_ids[] = {
-	{
-		.name		= "exynos4210-fimc",
-		.driver_data	= (unsigned long)&exynos4210_fimc_data,
-	}, {
-		.name		= "exynos4412-fimc",
-		.driver_data	= (unsigned long)&exynos4410_fimc_data,
-	},
-	{},
-};
-MODULE_DEVICE_TABLE(platform, fimc_driver_ids);
-
 static const struct dev_pm_ops fimc_pm_ops = {
 	SET_SYSTEM_SLEEP_PM_OPS(fimc_suspend, fimc_resume)
 	SET_RUNTIME_PM_OPS(fimc_runtime_suspend, fimc_runtime_resume, NULL)
 };
 
+static const struct of_device_id fimc_of_match[] = {
+	{ .compatible = "samsung,exynos4210-fimc" },
+	{ .compatible = "samsung,exynos4212-fimc" },
+	{ },
+};
+
 struct platform_driver fimc_driver = {
 	.probe		= fimc_probe,
 	.remove		= fimc_remove,
-	.id_table	= fimc_driver_ids,
 	.driver		= {
+		.of_match_table = fimc_of_match,
 		.name	= "exynos-drm-fimc",
 		.owner	= THIS_MODULE,
 		.pm	= &fimc_pm_ops,
diff --git a/drivers/gpu/drm/exynos/exynos_drm_fimd.c b/drivers/gpu/drm/exynos/exynos_drm_fimd.c
index 98cc14725ba9..746b282b343a 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_fimd.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_fimd.c
@@ -20,6 +20,7 @@
 #include <linux/of_device.h>
 #include <linux/pm_runtime.h>
 
+#include <video/of_display_timing.h>
 #include <video/samsung_fimd.h>
 #include <drm/exynos_drm.h>
 
@@ -800,18 +801,18 @@ static int fimd_clock(struct fimd_context *ctx, bool enable)
 	if (enable) {
 		int ret;
 
-		ret = clk_enable(ctx->bus_clk);
+		ret = clk_prepare_enable(ctx->bus_clk);
 		if (ret < 0)
 			return ret;
 
-		ret = clk_enable(ctx->lcd_clk);
+		ret = clk_prepare_enable(ctx->lcd_clk);
 		if  (ret < 0) {
-			clk_disable(ctx->bus_clk);
+			clk_disable_unprepare(ctx->bus_clk);
 			return ret;
 		}
 	} else {
-		clk_disable(ctx->lcd_clk);
-		clk_disable(ctx->bus_clk);
+		clk_disable_unprepare(ctx->lcd_clk);
+		clk_disable_unprepare(ctx->bus_clk);
 	}
 
 	return 0;
@@ -884,10 +885,25 @@ static int fimd_probe(struct platform_device *pdev)
 
 	DRM_DEBUG_KMS("%s\n", __FILE__);
 
-	pdata = pdev->dev.platform_data;
-	if (!pdata) {
-		dev_err(dev, "no platform data specified\n");
-		return -EINVAL;
+	if (pdev->dev.of_node) {
+		pdata = devm_kzalloc(dev, sizeof(*pdata), GFP_KERNEL);
+		if (!pdata) {
+			DRM_ERROR("memory allocation for pdata failed\n");
+			return -ENOMEM;
+		}
+
+		ret = of_get_fb_videomode(dev->of_node, &pdata->panel.timing,
+					OF_USE_NATIVE_MODE);
+		if (ret) {
+			DRM_ERROR("failed: of_get_fb_videomode() : %d\n", ret);
+			return ret;
+		}
+	} else {
+		pdata = pdev->dev.platform_data;
+		if (!pdata) {
+			DRM_ERROR("no platform data specified\n");
+			return -EINVAL;
+		}
 	}
 
 	panel = &pdata->panel;
@@ -918,7 +934,7 @@ static int fimd_probe(struct platform_device *pdev)
 	if (IS_ERR(ctx->regs))
 		return PTR_ERR(ctx->regs);
 
-	res = platform_get_resource(pdev, IORESOURCE_IRQ, 0);
+	res = platform_get_resource_byname(pdev, IORESOURCE_IRQ, "vsync");
 	if (!res) {
 		dev_err(dev, "irq request failed.\n");
 		return -ENXIO;
@@ -980,9 +996,6 @@ static int fimd_remove(struct platform_device *pdev)
 	if (ctx->suspended)
 		goto out;
 
-	clk_disable(ctx->lcd_clk);
-	clk_disable(ctx->bus_clk);
-
 	pm_runtime_set_suspended(dev);
 	pm_runtime_put_sync(dev);
 
diff --git a/drivers/gpu/drm/exynos/exynos_drm_gem.c b/drivers/gpu/drm/exynos/exynos_drm_gem.c
index 0e6fe000578c..cf4543ffa079 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_gem.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_gem.c
@@ -682,7 +682,8 @@ int exynos_drm_gem_dumb_create(struct drm_file *file_priv,
 	args->pitch = args->width * ((args->bpp + 7) / 8);
 	args->size = args->pitch * args->height;
 
-	exynos_gem_obj = exynos_drm_gem_create(dev, args->flags, args->size);
+	exynos_gem_obj = exynos_drm_gem_create(dev, EXYNOS_BO_CONTIG |
+						EXYNOS_BO_WC, args->size);
 	if (IS_ERR(exynos_gem_obj))
 		return PTR_ERR(exynos_gem_obj);
 
diff --git a/drivers/gpu/drm/exynos/exynos_drm_hdmi.c b/drivers/gpu/drm/exynos/exynos_drm_hdmi.c
index 7c27df03c9ff..ba2f0f1aa05f 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_hdmi.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_hdmi.c
@@ -51,21 +51,27 @@ struct drm_hdmi_context {
 
 int exynos_platform_device_hdmi_register(void)
 {
+	struct platform_device *pdev;
+
 	if (exynos_drm_hdmi_pdev)
 		return -EEXIST;
 
-	exynos_drm_hdmi_pdev = platform_device_register_simple(
+	pdev = platform_device_register_simple(
 			"exynos-drm-hdmi", -1, NULL, 0);
-	if (IS_ERR_OR_NULL(exynos_drm_hdmi_pdev))
-		return PTR_ERR(exynos_drm_hdmi_pdev);
+	if (IS_ERR(pdev))
+		return PTR_ERR(pdev);
+
+	exynos_drm_hdmi_pdev = pdev;
 
 	return 0;
 }
 
 void exynos_platform_device_hdmi_unregister(void)
 {
-	if (exynos_drm_hdmi_pdev)
+	if (exynos_drm_hdmi_pdev) {
 		platform_device_unregister(exynos_drm_hdmi_pdev);
+		exynos_drm_hdmi_pdev = NULL;
+	}
 }
 
 void exynos_hdmi_drv_attach(struct exynos_drm_hdmi_context *ctx)
@@ -205,13 +211,45 @@ static void drm_hdmi_mode_fixup(struct device *subdrv_dev,
 				const struct drm_display_mode *mode,
 				struct drm_display_mode *adjusted_mode)
 {
-	struct drm_hdmi_context *ctx = to_context(subdrv_dev);
+	struct drm_display_mode *m;
+	int mode_ok;
 
 	DRM_DEBUG_KMS("%s\n", __FILE__);
 
-	if (hdmi_ops && hdmi_ops->mode_fixup)
-		hdmi_ops->mode_fixup(ctx->hdmi_ctx->ctx, connector, mode,
-				     adjusted_mode);
+	drm_mode_set_crtcinfo(adjusted_mode, 0);
+
+	mode_ok = drm_hdmi_check_timing(subdrv_dev, adjusted_mode);
+
+	/* just return if user desired mode exists. */
+	if (mode_ok == 0)
+		return;
+
+	/*
+	 * otherwise, find the most suitable mode among modes and change it
+	 * to adjusted_mode.
+	 */
+	list_for_each_entry(m, &connector->modes, head) {
+		mode_ok = drm_hdmi_check_timing(subdrv_dev, m);
+
+		if (mode_ok == 0) {
+			struct drm_mode_object base;
+			struct list_head head;
+
+			DRM_INFO("desired mode doesn't exist so\n");
+			DRM_INFO("use the most suitable mode among modes.\n");
+
+			DRM_DEBUG_KMS("Adjusted Mode: [%d]x[%d] [%d]Hz\n",
+				m->hdisplay, m->vdisplay, m->vrefresh);
+
+			/* preserve display mode header while copying. */
+			head = adjusted_mode->head;
+			base = adjusted_mode->base;
+			memcpy(adjusted_mode, m, sizeof(*m));
+			adjusted_mode->head = head;
+			adjusted_mode->base = base;
+			break;
+		}
+	}
 }
 
 static void drm_hdmi_mode_set(struct device *subdrv_dev, void *mode)
diff --git a/drivers/gpu/drm/exynos/exynos_drm_hdmi.h b/drivers/gpu/drm/exynos/exynos_drm_hdmi.h
index b7faa3662307..6b709440df4c 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_hdmi.h
+++ b/drivers/gpu/drm/exynos/exynos_drm_hdmi.h
@@ -36,9 +36,6 @@ struct exynos_hdmi_ops {
 	int (*power_on)(void *ctx, int mode);
 
 	/* manager */
-	void (*mode_fixup)(void *ctx, struct drm_connector *connector,
-				const struct drm_display_mode *mode,
-				struct drm_display_mode *adjusted_mode);
 	void (*mode_set)(void *ctx, void *mode);
 	void (*get_max_resol)(void *ctx, unsigned int *width,
 				unsigned int *height);
diff --git a/drivers/gpu/drm/exynos/exynos_drm_ipp.c b/drivers/gpu/drm/exynos/exynos_drm_ipp.c
index 1adce07ecb5b..29d2ad314490 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_ipp.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_ipp.c
@@ -47,6 +47,9 @@
 #define get_ipp_context(dev)	platform_get_drvdata(to_platform_device(dev))
 #define ipp_is_m2m_cmd(c)	(c == IPP_CMD_M2M)
 
+/* platform device pointer for ipp device. */
+static struct platform_device *exynos_drm_ipp_pdev;
+
 /*
  * A structure of event.
  *
@@ -102,6 +105,30 @@ static LIST_HEAD(exynos_drm_ippdrv_list);
 static DEFINE_MUTEX(exynos_drm_ippdrv_lock);
 static BLOCKING_NOTIFIER_HEAD(exynos_drm_ippnb_list);
 
+int exynos_platform_device_ipp_register(void)
+{
+	struct platform_device *pdev;
+
+	if (exynos_drm_ipp_pdev)
+		return -EEXIST;
+
+	pdev = platform_device_register_simple("exynos-drm-ipp", -1, NULL, 0);
+	if (IS_ERR(pdev))
+		return PTR_ERR(pdev);
+
+	exynos_drm_ipp_pdev = pdev;
+
+	return 0;
+}
+
+void exynos_platform_device_ipp_unregister(void)
+{
+	if (exynos_drm_ipp_pdev) {
+		platform_device_unregister(exynos_drm_ipp_pdev);
+		exynos_drm_ipp_pdev = NULL;
+	}
+}
+
 int exynos_drm_ippdrv_register(struct exynos_drm_ippdrv *ippdrv)
 {
 	DRM_DEBUG_KMS("%s\n", __func__);
diff --git a/drivers/gpu/drm/exynos/exynos_drm_rotator.c b/drivers/gpu/drm/exynos/exynos_drm_rotator.c
index a40b9fb60240..947f09f15ad1 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_rotator.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_rotator.c
@@ -674,7 +674,7 @@ static int rotator_probe(struct platform_device *pdev)
 	}
 
 	rot->clock = devm_clk_get(dev, "rotator");
-	if (IS_ERR_OR_NULL(rot->clock)) {
+	if (IS_ERR(rot->clock)) {
 		dev_err(dev, "failed to get clock\n");
 		ret = PTR_ERR(rot->clock);
 		goto err_clk_get;
diff --git a/drivers/gpu/drm/exynos/exynos_hdmi.c b/drivers/gpu/drm/exynos/exynos_hdmi.c
index 2c5f266154ad..bbfc3840080c 100644
--- a/drivers/gpu/drm/exynos/exynos_hdmi.c
+++ b/drivers/gpu/drm/exynos/exynos_hdmi.c
@@ -108,7 +108,20 @@ struct hdmi_tg_regs {
 	u8 tg_3d[1];
 };
 
-struct hdmi_core_regs {
+struct hdmi_v13_core_regs {
+	u8 h_blank[2];
+	u8 v_blank[3];
+	u8 h_v_line[3];
+	u8 vsync_pol[1];
+	u8 int_pro_mode[1];
+	u8 v_blank_f[3];
+	u8 h_sync_gen[3];
+	u8 v_sync_gen1[3];
+	u8 v_sync_gen2[3];
+	u8 v_sync_gen3[3];
+};
+
+struct hdmi_v14_core_regs {
 	u8 h_blank[2];
 	u8 v2_blank[2];
 	u8 v1_blank[2];
@@ -147,11 +160,23 @@ struct hdmi_core_regs {
 	u8 vact_space_6[2];
 };
 
+struct hdmi_v13_conf {
+	struct hdmi_v13_core_regs core;
+	struct hdmi_tg_regs tg;
+};
+
 struct hdmi_v14_conf {
-	int pixel_clock;
-	struct hdmi_core_regs core;
+	struct hdmi_v14_core_regs core;
 	struct hdmi_tg_regs tg;
+};
+
+struct hdmi_conf_regs {
+	int pixel_clock;
 	int cea_video_id;
+	union {
+		struct hdmi_v13_conf v13_conf;
+		struct hdmi_v14_conf v14_conf;
+	} conf;
 };
 
 struct hdmi_context {
@@ -169,9 +194,8 @@ struct hdmi_context {
 	struct i2c_client		*ddc_port;
 	struct i2c_client		*hdmiphy_port;
 
-	/* current hdmiphy conf index */
-	int cur_conf;
-	struct hdmi_v14_conf		mode_conf;
+	/* current hdmiphy conf regs */
+	struct hdmi_conf_regs		mode_conf;
 
 	struct hdmi_resources		res;
 
@@ -180,292 +204,60 @@ struct hdmi_context {
 	enum hdmi_type			type;
 };
 
-/* HDMI Version 1.3 */
-static const u8 hdmiphy_v13_conf27[32] = {
-	0x01, 0x05, 0x00, 0xD8, 0x10, 0x1C, 0x30, 0x40,
-	0x6B, 0x10, 0x02, 0x51, 0xDF, 0xF2, 0x54, 0x87,
-	0x84, 0x00, 0x30, 0x38, 0x00, 0x08, 0x10, 0xE0,
-	0x22, 0x40, 0xE3, 0x26, 0x00, 0x00, 0x00, 0x00,
-};
-
-static const u8 hdmiphy_v13_conf27_027[32] = {
-	0x01, 0x05, 0x00, 0xD4, 0x10, 0x9C, 0x09, 0x64,
-	0x6B, 0x10, 0x02, 0x51, 0xDF, 0xF2, 0x54, 0x87,
-	0x84, 0x00, 0x30, 0x38, 0x00, 0x08, 0x10, 0xE0,
-	0x22, 0x40, 0xE3, 0x26, 0x00, 0x00, 0x00, 0x00,
-};
-
-static const u8 hdmiphy_v13_conf74_175[32] = {
-	0x01, 0x05, 0x00, 0xD8, 0x10, 0x9C, 0xef, 0x5B,
-	0x6D, 0x10, 0x01, 0x51, 0xef, 0xF3, 0x54, 0xb9,
-	0x84, 0x00, 0x30, 0x38, 0x00, 0x08, 0x10, 0xE0,
-	0x22, 0x40, 0xa5, 0x26, 0x01, 0x00, 0x00, 0x00,
-};
-
-static const u8 hdmiphy_v13_conf74_25[32] = {
-	0x01, 0x05, 0x00, 0xd8, 0x10, 0x9c, 0xf8, 0x40,
-	0x6a, 0x10, 0x01, 0x51, 0xff, 0xf1, 0x54, 0xba,
-	0x84, 0x00, 0x10, 0x38, 0x00, 0x08, 0x10, 0xe0,
-	0x22, 0x40, 0xa4, 0x26, 0x01, 0x00, 0x00, 0x00,
-};
-
-static const u8 hdmiphy_v13_conf148_5[32] = {
-	0x01, 0x05, 0x00, 0xD8, 0x10, 0x9C, 0xf8, 0x40,
-	0x6A, 0x18, 0x00, 0x51, 0xff, 0xF1, 0x54, 0xba,
-	0x84, 0x00, 0x10, 0x38, 0x00, 0x08, 0x10, 0xE0,
-	0x22, 0x40, 0xa4, 0x26, 0x02, 0x00, 0x00, 0x00,
-};
-
-struct hdmi_v13_tg_regs {
-	u8 cmd;
-	u8 h_fsz_l;
-	u8 h_fsz_h;
-	u8 hact_st_l;
-	u8 hact_st_h;
-	u8 hact_sz_l;
-	u8 hact_sz_h;
-	u8 v_fsz_l;
-	u8 v_fsz_h;
-	u8 vsync_l;
-	u8 vsync_h;
-	u8 vsync2_l;
-	u8 vsync2_h;
-	u8 vact_st_l;
-	u8 vact_st_h;
-	u8 vact_sz_l;
-	u8 vact_sz_h;
-	u8 field_chg_l;
-	u8 field_chg_h;
-	u8 vact_st2_l;
-	u8 vact_st2_h;
-	u8 vsync_top_hdmi_l;
-	u8 vsync_top_hdmi_h;
-	u8 vsync_bot_hdmi_l;
-	u8 vsync_bot_hdmi_h;
-	u8 field_top_hdmi_l;
-	u8 field_top_hdmi_h;
-	u8 field_bot_hdmi_l;
-	u8 field_bot_hdmi_h;
-};
-
-struct hdmi_v13_core_regs {
-	u8 h_blank[2];
-	u8 v_blank[3];
-	u8 h_v_line[3];
-	u8 vsync_pol[1];
-	u8 int_pro_mode[1];
-	u8 v_blank_f[3];
-	u8 h_sync_gen[3];
-	u8 v_sync_gen1[3];
-	u8 v_sync_gen2[3];
-	u8 v_sync_gen3[3];
-};
-
-struct hdmi_v13_preset_conf {
-	struct hdmi_v13_core_regs core;
-	struct hdmi_v13_tg_regs tg;
-};
-
-struct hdmi_v13_conf {
-	int width;
-	int height;
-	int vrefresh;
-	bool interlace;
-	int cea_video_id;
-	const u8 *hdmiphy_data;
-	const struct hdmi_v13_preset_conf *conf;
-};
-
-static const struct hdmi_v13_preset_conf hdmi_v13_conf_480p = {
-	.core = {
-		.h_blank = {0x8a, 0x00},
-		.v_blank = {0x0d, 0x6a, 0x01},
-		.h_v_line = {0x0d, 0xa2, 0x35},
-		.vsync_pol = {0x01},
-		.int_pro_mode = {0x00},
-		.v_blank_f = {0x00, 0x00, 0x00},
-		.h_sync_gen = {0x0e, 0x30, 0x11},
-		.v_sync_gen1 = {0x0f, 0x90, 0x00},
-		/* other don't care */
-	},
-	.tg = {
-		0x00, /* cmd */
-		0x5a, 0x03, /* h_fsz */
-		0x8a, 0x00, 0xd0, 0x02, /* hact */
-		0x0d, 0x02, /* v_fsz */
-		0x01, 0x00, 0x33, 0x02, /* vsync */
-		0x2d, 0x00, 0xe0, 0x01, /* vact */
-		0x33, 0x02, /* field_chg */
-		0x49, 0x02, /* vact_st2 */
-		0x01, 0x00, 0x33, 0x02, /* vsync top/bot */
-		0x01, 0x00, 0x33, 0x02, /* field top/bot */
-	},
-};
-
-static const struct hdmi_v13_preset_conf hdmi_v13_conf_720p60 = {
-	.core = {
-		.h_blank = {0x72, 0x01},
-		.v_blank = {0xee, 0xf2, 0x00},
-		.h_v_line = {0xee, 0x22, 0x67},
-		.vsync_pol = {0x00},
-		.int_pro_mode = {0x00},
-		.v_blank_f = {0x00, 0x00, 0x00}, /* don't care */
-		.h_sync_gen = {0x6c, 0x50, 0x02},
-		.v_sync_gen1 = {0x0a, 0x50, 0x00},
-		.v_sync_gen2 = {0x01, 0x10, 0x00},
-		.v_sync_gen3 = {0x01, 0x10, 0x00},
-		/* other don't care */
-	},
-	.tg = {
-		0x00, /* cmd */
-		0x72, 0x06, /* h_fsz */
-		0x71, 0x01, 0x01, 0x05, /* hact */
-		0xee, 0x02, /* v_fsz */
-		0x01, 0x00, 0x33, 0x02, /* vsync */
-		0x1e, 0x00, 0xd0, 0x02, /* vact */
-		0x33, 0x02, /* field_chg */
-		0x49, 0x02, /* vact_st2 */
-		0x01, 0x00, 0x01, 0x00, /* vsync top/bot */
-		0x01, 0x00, 0x33, 0x02, /* field top/bot */
-	},
-};
-
-static const struct hdmi_v13_preset_conf hdmi_v13_conf_1080i50 = {
-	.core = {
-		.h_blank = {0xd0, 0x02},
-		.v_blank = {0x32, 0xB2, 0x00},
-		.h_v_line = {0x65, 0x04, 0xa5},
-		.vsync_pol = {0x00},
-		.int_pro_mode = {0x01},
-		.v_blank_f = {0x49, 0x2A, 0x23},
-		.h_sync_gen = {0x0E, 0xEA, 0x08},
-		.v_sync_gen1 = {0x07, 0x20, 0x00},
-		.v_sync_gen2 = {0x39, 0x42, 0x23},
-		.v_sync_gen3 = {0x38, 0x87, 0x73},
-		/* other don't care */
-	},
-	.tg = {
-		0x00, /* cmd */
-		0x50, 0x0A, /* h_fsz */
-		0xCF, 0x02, 0x81, 0x07, /* hact */
-		0x65, 0x04, /* v_fsz */
-		0x01, 0x00, 0x33, 0x02, /* vsync */
-		0x16, 0x00, 0x1c, 0x02, /* vact */
-		0x33, 0x02, /* field_chg */
-		0x49, 0x02, /* vact_st2 */
-		0x01, 0x00, 0x33, 0x02, /* vsync top/bot */
-		0x01, 0x00, 0x33, 0x02, /* field top/bot */
-	},
+struct hdmiphy_config {
+	int pixel_clock;
+	u8 conf[32];
 };
 
-static const struct hdmi_v13_preset_conf hdmi_v13_conf_1080p50 = {
-	.core = {
-		.h_blank = {0xd0, 0x02},
-		.v_blank = {0x65, 0x6c, 0x01},
-		.h_v_line = {0x65, 0x04, 0xa5},
-		.vsync_pol = {0x00},
-		.int_pro_mode = {0x00},
-		.v_blank_f = {0x00, 0x00, 0x00}, /* don't care */
-		.h_sync_gen = {0x0e, 0xea, 0x08},
-		.v_sync_gen1 = {0x09, 0x40, 0x00},
-		.v_sync_gen2 = {0x01, 0x10, 0x00},
-		.v_sync_gen3 = {0x01, 0x10, 0x00},
-		/* other don't care */
-	},
-	.tg = {
-		0x00, /* cmd */
-		0x50, 0x0A, /* h_fsz */
-		0xCF, 0x02, 0x81, 0x07, /* hact */
-		0x65, 0x04, /* v_fsz */
-		0x01, 0x00, 0x33, 0x02, /* vsync */
-		0x2d, 0x00, 0x38, 0x04, /* vact */
-		0x33, 0x02, /* field_chg */
-		0x48, 0x02, /* vact_st2 */
-		0x01, 0x00, 0x01, 0x00, /* vsync top/bot */
-		0x01, 0x00, 0x33, 0x02, /* field top/bot */
+/* list of phy config settings */
+static const struct hdmiphy_config hdmiphy_v13_configs[] = {
+	{
+		.pixel_clock = 27000000,
+		.conf = {
+			0x01, 0x05, 0x00, 0xD8, 0x10, 0x1C, 0x30, 0x40,
+			0x6B, 0x10, 0x02, 0x51, 0xDF, 0xF2, 0x54, 0x87,
+			0x84, 0x00, 0x30, 0x38, 0x00, 0x08, 0x10, 0xE0,
+			0x22, 0x40, 0xE3, 0x26, 0x00, 0x00, 0x00, 0x00,
+		},
 	},
-};
-
-static const struct hdmi_v13_preset_conf hdmi_v13_conf_1080i60 = {
-	.core = {
-		.h_blank = {0x18, 0x01},
-		.v_blank = {0x32, 0xB2, 0x00},
-		.h_v_line = {0x65, 0x84, 0x89},
-		.vsync_pol = {0x00},
-		.int_pro_mode = {0x01},
-		.v_blank_f = {0x49, 0x2A, 0x23},
-		.h_sync_gen = {0x56, 0x08, 0x02},
-		.v_sync_gen1 = {0x07, 0x20, 0x00},
-		.v_sync_gen2 = {0x39, 0x42, 0x23},
-		.v_sync_gen3 = {0xa4, 0x44, 0x4a},
-		/* other don't care */
+	{
+		.pixel_clock = 27027000,
+		.conf = {
+			0x01, 0x05, 0x00, 0xD4, 0x10, 0x9C, 0x09, 0x64,
+			0x6B, 0x10, 0x02, 0x51, 0xDF, 0xF2, 0x54, 0x87,
+			0x84, 0x00, 0x30, 0x38, 0x00, 0x08, 0x10, 0xE0,
+			0x22, 0x40, 0xE3, 0x26, 0x00, 0x00, 0x00, 0x00,
+		},
 	},
-	.tg = {
-		0x00, /* cmd */
-		0x98, 0x08, /* h_fsz */
-		0x17, 0x01, 0x81, 0x07, /* hact */
-		0x65, 0x04, /* v_fsz */
-		0x01, 0x00, 0x33, 0x02, /* vsync */
-		0x16, 0x00, 0x1c, 0x02, /* vact */
-		0x33, 0x02, /* field_chg */
-		0x49, 0x02, /* vact_st2 */
-		0x01, 0x00, 0x33, 0x02, /* vsync top/bot */
-		0x01, 0x00, 0x33, 0x02, /* field top/bot */
+	{
+		.pixel_clock = 74176000,
+		.conf = {
+			0x01, 0x05, 0x00, 0xD8, 0x10, 0x9C, 0xef, 0x5B,
+			0x6D, 0x10, 0x01, 0x51, 0xef, 0xF3, 0x54, 0xb9,
+			0x84, 0x00, 0x30, 0x38, 0x00, 0x08, 0x10, 0xE0,
+			0x22, 0x40, 0xa5, 0x26, 0x01, 0x00, 0x00, 0x00,
+		},
 	},
-};
-
-static const struct hdmi_v13_preset_conf hdmi_v13_conf_1080p60 = {
-	.core = {
-		.h_blank = {0x18, 0x01},
-		.v_blank = {0x65, 0x6c, 0x01},
-		.h_v_line = {0x65, 0x84, 0x89},
-		.vsync_pol = {0x00},
-		.int_pro_mode = {0x00},
-		.v_blank_f = {0x00, 0x00, 0x00}, /* don't care */
-		.h_sync_gen = {0x56, 0x08, 0x02},
-		.v_sync_gen1 = {0x09, 0x40, 0x00},
-		.v_sync_gen2 = {0x01, 0x10, 0x00},
-		.v_sync_gen3 = {0x01, 0x10, 0x00},
-		/* other don't care */
+	{
+		.pixel_clock = 74250000,
+		.conf = {
+			0x01, 0x05, 0x00, 0xd8, 0x10, 0x9c, 0xf8, 0x40,
+			0x6a, 0x10, 0x01, 0x51, 0xff, 0xf1, 0x54, 0xba,
+			0x84, 0x00, 0x10, 0x38, 0x00, 0x08, 0x10, 0xe0,
+			0x22, 0x40, 0xa4, 0x26, 0x01, 0x00, 0x00, 0x00,
+		},
 	},
-	.tg = {
-		0x00, /* cmd */
-		0x98, 0x08, /* h_fsz */
-		0x17, 0x01, 0x81, 0x07, /* hact */
-		0x65, 0x04, /* v_fsz */
-		0x01, 0x00, 0x33, 0x02, /* vsync */
-		0x2d, 0x00, 0x38, 0x04, /* vact */
-		0x33, 0x02, /* field_chg */
-		0x48, 0x02, /* vact_st2 */
-		0x01, 0x00, 0x01, 0x00, /* vsync top/bot */
-		0x01, 0x00, 0x33, 0x02, /* field top/bot */
+	{
+		.pixel_clock = 148500000,
+		.conf = {
+			0x01, 0x05, 0x00, 0xD8, 0x10, 0x9C, 0xf8, 0x40,
+			0x6A, 0x18, 0x00, 0x51, 0xff, 0xF1, 0x54, 0xba,
+			0x84, 0x00, 0x10, 0x38, 0x00, 0x08, 0x10, 0xE0,
+			0x22, 0x40, 0xa4, 0x26, 0x02, 0x00, 0x00, 0x00,
+		},
 	},
 };
 
-static const struct hdmi_v13_conf hdmi_v13_confs[] = {
-	{ 1280, 720, 60, false, 4, hdmiphy_v13_conf74_25,
-			&hdmi_v13_conf_720p60 },
-	{ 1280, 720, 50, false, 19, hdmiphy_v13_conf74_25,
-			&hdmi_v13_conf_720p60 },
-	{ 720, 480, 60, false, 3, hdmiphy_v13_conf27_027,
-			&hdmi_v13_conf_480p },
-	{ 1920, 1080, 50, true, 20, hdmiphy_v13_conf74_25,
-			&hdmi_v13_conf_1080i50 },
-	{ 1920, 1080, 50, false, 31, hdmiphy_v13_conf148_5,
-			&hdmi_v13_conf_1080p50 },
-	{ 1920, 1080, 60, true, 5, hdmiphy_v13_conf74_25,
-			&hdmi_v13_conf_1080i60 },
-	{ 1920, 1080, 60, false, 16, hdmiphy_v13_conf148_5,
-			&hdmi_v13_conf_1080p60 },
-};
-
-/* HDMI Version 1.4 */
-struct hdmiphy_config {
-	int pixel_clock;
-	u8 conf[32];
-};
-
-/* list of all required phy config settings */
 static const struct hdmiphy_config hdmiphy_v14_configs[] = {
 	{
 		.pixel_clock = 25200000,
@@ -873,22 +665,6 @@ static void hdmi_regs_dump(struct hdmi_context *hdata, char *prefix)
 		hdmi_v14_regs_dump(hdata, prefix);
 }
 
-static int hdmi_v13_conf_index(struct drm_display_mode *mode)
-{
-	int i;
-
-	for (i = 0; i < ARRAY_SIZE(hdmi_v13_confs); ++i)
-		if (hdmi_v13_confs[i].width == mode->hdisplay &&
-				hdmi_v13_confs[i].height == mode->vdisplay &&
-				hdmi_v13_confs[i].vrefresh == mode->vrefresh &&
-				hdmi_v13_confs[i].interlace ==
-				((mode->flags & DRM_MODE_FLAG_INTERLACE) ?
-				 true : false))
-			return i;
-
-	return -EINVAL;
-}
-
 static u8 hdmi_chksum(struct hdmi_context *hdata,
 			u32 start, u8 len, u32 hdr_sum)
 {
@@ -943,11 +719,7 @@ static void hdmi_reg_infoframe(struct hdmi_context *hdata,
 		hdmi_reg_writeb(hdata, HDMI_AVI_BYTE(2), aspect_ratio |
 				AVI_SAME_AS_PIC_ASPECT_RATIO);
 
-		if (hdata->type == HDMI_TYPE13)
-			vic = hdmi_v13_confs[hdata->cur_conf].cea_video_id;
-		else
-			vic = hdata->mode_conf.cea_video_id;
-
+		vic = hdata->mode_conf.cea_video_id;
 		hdmi_reg_writeb(hdata, HDMI_AVI_BYTE(4), vic);
 
 		chksum = hdmi_chksum(hdata, HDMI_AVI_BYTE(1),
@@ -1000,63 +772,34 @@ static struct edid *hdmi_get_edid(void *ctx, struct drm_connector *connector)
 	return raw_edid;
 }
 
-static int hdmi_v13_check_timing(struct fb_videomode *check_timing)
+static int hdmi_find_phy_conf(struct hdmi_context *hdata, u32 pixel_clock)
 {
-	int i;
+	const struct hdmiphy_config *confs;
+	int count, i;
 
-	DRM_DEBUG_KMS("valid mode : xres=%d, yres=%d, refresh=%d, intl=%d\n",
-			check_timing->xres, check_timing->yres,
-			check_timing->refresh, (check_timing->vmode &
-			FB_VMODE_INTERLACED) ? true : false);
-
-	for (i = 0; i < ARRAY_SIZE(hdmi_v13_confs); ++i)
-		if (hdmi_v13_confs[i].width == check_timing->xres &&
-			hdmi_v13_confs[i].height == check_timing->yres &&
-			hdmi_v13_confs[i].vrefresh == check_timing->refresh &&
-			hdmi_v13_confs[i].interlace ==
-			((check_timing->vmode & FB_VMODE_INTERLACED) ?
-			 true : false))
-				return 0;
-
-	/* TODO */
-
-	return -EINVAL;
-}
+	DRM_DEBUG_KMS("[%d] %s\n", __LINE__, __func__);
 
-static int hdmi_v14_find_phy_conf(int pixel_clock)
-{
-	int i;
+	if (hdata->type == HDMI_TYPE13) {
+		confs = hdmiphy_v13_configs;
+		count = ARRAY_SIZE(hdmiphy_v13_configs);
+	} else if (hdata->type == HDMI_TYPE14) {
+		confs = hdmiphy_v14_configs;
+		count = ARRAY_SIZE(hdmiphy_v14_configs);
+	} else
+		return -EINVAL;
 
-	for (i = 0; i < ARRAY_SIZE(hdmiphy_v14_configs); i++) {
-		if (hdmiphy_v14_configs[i].pixel_clock == pixel_clock)
+	for (i = 0; i < count; i++)
+		if (confs[i].pixel_clock == pixel_clock)
 			return i;
-	}
 
 	DRM_DEBUG_KMS("Could not find phy config for %d\n", pixel_clock);
 	return -EINVAL;
 }
 
-static int hdmi_v14_check_timing(struct fb_videomode *check_timing)
-{
-	int i;
-
-	DRM_DEBUG_KMS("mode: xres=%d, yres=%d, refresh=%d, clock=%d, intl=%d\n",
-			check_timing->xres, check_timing->yres,
-			check_timing->refresh, check_timing->pixclock,
-			(check_timing->vmode & FB_VMODE_INTERLACED) ?
-			true : false);
-
-	for (i = 0; i < ARRAY_SIZE(hdmiphy_v14_configs); i++)
-		if (hdmiphy_v14_configs[i].pixel_clock ==
-			check_timing->pixclock)
-			return 0;
-
-	return -EINVAL;
-}
-
 static int hdmi_check_timing(void *ctx, struct fb_videomode *timing)
 {
 	struct hdmi_context *hdata = ctx;
+	int ret;
 
 	DRM_DEBUG_KMS("[%d] %s\n", __LINE__, __func__);
 
@@ -1064,10 +807,10 @@ static int hdmi_check_timing(void *ctx, struct fb_videomode *timing)
 			timing->yres, timing->refresh,
 			timing->vmode);
 
-	if (hdata->type == HDMI_TYPE13)
-		return hdmi_v13_check_timing(timing);
-	else
-		return hdmi_v14_check_timing(timing);
+	ret = hdmi_find_phy_conf(hdata, timing->pixclock);
+	if (ret < 0)
+		return ret;
+	return 0;
 }
 
 static void hdmi_set_acr(u32 freq, u8 *acr)
@@ -1301,10 +1044,9 @@ static void hdmi_conf_init(struct hdmi_context *hdata)
 
 static void hdmi_v13_timing_apply(struct hdmi_context *hdata)
 {
-	const struct hdmi_v13_preset_conf *conf =
-		hdmi_v13_confs[hdata->cur_conf].conf;
-	const struct hdmi_v13_core_regs *core = &conf->core;
-	const struct hdmi_v13_tg_regs *tg = &conf->tg;
+	const struct hdmi_tg_regs *tg = &hdata->mode_conf.conf.v13_conf.tg;
+	const struct hdmi_v13_core_regs *core =
+		&hdata->mode_conf.conf.v13_conf.core;
 	int tries;
 
 	/* setting core registers */
@@ -1334,34 +1076,34 @@ static void hdmi_v13_timing_apply(struct hdmi_context *hdata)
 	hdmi_reg_writeb(hdata, HDMI_V13_V_SYNC_GEN_3_1, core->v_sync_gen3[1]);
 	hdmi_reg_writeb(hdata, HDMI_V13_V_SYNC_GEN_3_2, core->v_sync_gen3[2]);
 	/* Timing generator registers */
-	hdmi_reg_writeb(hdata, HDMI_TG_H_FSZ_L, tg->h_fsz_l);
-	hdmi_reg_writeb(hdata, HDMI_TG_H_FSZ_H, tg->h_fsz_h);
-	hdmi_reg_writeb(hdata, HDMI_TG_HACT_ST_L, tg->hact_st_l);
-	hdmi_reg_writeb(hdata, HDMI_TG_HACT_ST_H, tg->hact_st_h);
-	hdmi_reg_writeb(hdata, HDMI_TG_HACT_SZ_L, tg->hact_sz_l);
-	hdmi_reg_writeb(hdata, HDMI_TG_HACT_SZ_H, tg->hact_sz_h);
-	hdmi_reg_writeb(hdata, HDMI_TG_V_FSZ_L, tg->v_fsz_l);
-	hdmi_reg_writeb(hdata, HDMI_TG_V_FSZ_H, tg->v_fsz_h);
-	hdmi_reg_writeb(hdata, HDMI_TG_VSYNC_L, tg->vsync_l);
-	hdmi_reg_writeb(hdata, HDMI_TG_VSYNC_H, tg->vsync_h);
-	hdmi_reg_writeb(hdata, HDMI_TG_VSYNC2_L, tg->vsync2_l);
-	hdmi_reg_writeb(hdata, HDMI_TG_VSYNC2_H, tg->vsync2_h);
-	hdmi_reg_writeb(hdata, HDMI_TG_VACT_ST_L, tg->vact_st_l);
-	hdmi_reg_writeb(hdata, HDMI_TG_VACT_ST_H, tg->vact_st_h);
-	hdmi_reg_writeb(hdata, HDMI_TG_VACT_SZ_L, tg->vact_sz_l);
-	hdmi_reg_writeb(hdata, HDMI_TG_VACT_SZ_H, tg->vact_sz_h);
-	hdmi_reg_writeb(hdata, HDMI_TG_FIELD_CHG_L, tg->field_chg_l);
-	hdmi_reg_writeb(hdata, HDMI_TG_FIELD_CHG_H, tg->field_chg_h);
-	hdmi_reg_writeb(hdata, HDMI_TG_VACT_ST2_L, tg->vact_st2_l);
-	hdmi_reg_writeb(hdata, HDMI_TG_VACT_ST2_H, tg->vact_st2_h);
-	hdmi_reg_writeb(hdata, HDMI_TG_VSYNC_TOP_HDMI_L, tg->vsync_top_hdmi_l);
-	hdmi_reg_writeb(hdata, HDMI_TG_VSYNC_TOP_HDMI_H, tg->vsync_top_hdmi_h);
-	hdmi_reg_writeb(hdata, HDMI_TG_VSYNC_BOT_HDMI_L, tg->vsync_bot_hdmi_l);
-	hdmi_reg_writeb(hdata, HDMI_TG_VSYNC_BOT_HDMI_H, tg->vsync_bot_hdmi_h);
-	hdmi_reg_writeb(hdata, HDMI_TG_FIELD_TOP_HDMI_L, tg->field_top_hdmi_l);
-	hdmi_reg_writeb(hdata, HDMI_TG_FIELD_TOP_HDMI_H, tg->field_top_hdmi_h);
-	hdmi_reg_writeb(hdata, HDMI_TG_FIELD_BOT_HDMI_L, tg->field_bot_hdmi_l);
-	hdmi_reg_writeb(hdata, HDMI_TG_FIELD_BOT_HDMI_H, tg->field_bot_hdmi_h);
+	hdmi_reg_writeb(hdata, HDMI_TG_H_FSZ_L, tg->h_fsz[0]);
+	hdmi_reg_writeb(hdata, HDMI_TG_H_FSZ_H, tg->h_fsz[1]);
+	hdmi_reg_writeb(hdata, HDMI_TG_HACT_ST_L, tg->hact_st[0]);
+	hdmi_reg_writeb(hdata, HDMI_TG_HACT_ST_H, tg->hact_st[1]);
+	hdmi_reg_writeb(hdata, HDMI_TG_HACT_SZ_L, tg->hact_sz[0]);
+	hdmi_reg_writeb(hdata, HDMI_TG_HACT_SZ_H, tg->hact_sz[1]);
+	hdmi_reg_writeb(hdata, HDMI_TG_V_FSZ_L, tg->v_fsz[0]);
+	hdmi_reg_writeb(hdata, HDMI_TG_V_FSZ_H, tg->v_fsz[1]);
+	hdmi_reg_writeb(hdata, HDMI_TG_VSYNC_L, tg->vsync[0]);
+	hdmi_reg_writeb(hdata, HDMI_TG_VSYNC_H, tg->vsync[1]);
+	hdmi_reg_writeb(hdata, HDMI_TG_VSYNC2_L, tg->vsync2[0]);
+	hdmi_reg_writeb(hdata, HDMI_TG_VSYNC2_H, tg->vsync2[1]);
+	hdmi_reg_writeb(hdata, HDMI_TG_VACT_ST_L, tg->vact_st[0]);
+	hdmi_reg_writeb(hdata, HDMI_TG_VACT_ST_H, tg->vact_st[1]);
+	hdmi_reg_writeb(hdata, HDMI_TG_VACT_SZ_L, tg->vact_sz[0]);
+	hdmi_reg_writeb(hdata, HDMI_TG_VACT_SZ_H, tg->vact_sz[1]);
+	hdmi_reg_writeb(hdata, HDMI_TG_FIELD_CHG_L, tg->field_chg[0]);
+	hdmi_reg_writeb(hdata, HDMI_TG_FIELD_CHG_H, tg->field_chg[1]);
+	hdmi_reg_writeb(hdata, HDMI_TG_VACT_ST2_L, tg->vact_st2[0]);
+	hdmi_reg_writeb(hdata, HDMI_TG_VACT_ST2_H, tg->vact_st2[1]);
+	hdmi_reg_writeb(hdata, HDMI_TG_VSYNC_TOP_HDMI_L, tg->vsync_top_hdmi[0]);
+	hdmi_reg_writeb(hdata, HDMI_TG_VSYNC_TOP_HDMI_H, tg->vsync_top_hdmi[1]);
+	hdmi_reg_writeb(hdata, HDMI_TG_VSYNC_BOT_HDMI_L, tg->vsync_bot_hdmi[0]);
+	hdmi_reg_writeb(hdata, HDMI_TG_VSYNC_BOT_HDMI_H, tg->vsync_bot_hdmi[1]);
+	hdmi_reg_writeb(hdata, HDMI_TG_FIELD_TOP_HDMI_L, tg->field_top_hdmi[0]);
+	hdmi_reg_writeb(hdata, HDMI_TG_FIELD_TOP_HDMI_H, tg->field_top_hdmi[1]);
+	hdmi_reg_writeb(hdata, HDMI_TG_FIELD_BOT_HDMI_L, tg->field_bot_hdmi[0]);
+	hdmi_reg_writeb(hdata, HDMI_TG_FIELD_BOT_HDMI_H, tg->field_bot_hdmi[1]);
 
 	/* waiting for HDMIPHY's PLL to get to steady state */
 	for (tries = 100; tries; --tries) {
@@ -1391,8 +1133,9 @@ static void hdmi_v13_timing_apply(struct hdmi_context *hdata)
 
 static void hdmi_v14_timing_apply(struct hdmi_context *hdata)
 {
-	struct hdmi_core_regs *core = &hdata->mode_conf.core;
-	struct hdmi_tg_regs *tg = &hdata->mode_conf.tg;
+	const struct hdmi_tg_regs *tg = &hdata->mode_conf.conf.v14_conf.tg;
+	const struct hdmi_v14_core_regs *core =
+		&hdata->mode_conf.conf.v14_conf.core;
 	int tries;
 
 	/* setting core registers */
@@ -1624,17 +1367,16 @@ static void hdmiphy_conf_apply(struct hdmi_context *hdata)
 	}
 
 	/* pixel clock */
-	if (hdata->type == HDMI_TYPE13) {
-		hdmiphy_data = hdmi_v13_confs[hdata->cur_conf].hdmiphy_data;
-	} else {
-		i = hdmi_v14_find_phy_conf(hdata->mode_conf.pixel_clock);
-		if (i < 0) {
-			DRM_ERROR("failed to find hdmiphy conf\n");
-			return;
-		}
+	i = hdmi_find_phy_conf(hdata, hdata->mode_conf.pixel_clock);
+	if (i < 0) {
+		DRM_ERROR("failed to find hdmiphy conf\n");
+		return;
+	}
 
+	if (hdata->type == HDMI_TYPE13)
+		hdmiphy_data = hdmiphy_v13_configs[i].conf;
+	else
 		hdmiphy_data = hdmiphy_v14_configs[i].conf;
-	}
 
 	memcpy(buffer, hdmiphy_data, 32);
 	ret = i2c_master_send(hdata->hdmiphy_port, buffer, 32);
@@ -1687,75 +1429,121 @@ static void hdmi_conf_apply(struct hdmi_context *hdata)
 	hdmi_regs_dump(hdata, "start");
 }
 
-static void hdmi_mode_fixup(void *ctx, struct drm_connector *connector,
-				const struct drm_display_mode *mode,
-				struct drm_display_mode *adjusted_mode)
+static void hdmi_set_reg(u8 *reg_pair, int num_bytes, u32 value)
 {
-	struct drm_display_mode *m;
-	struct hdmi_context *hdata = ctx;
-	int index;
+	int i;
+	BUG_ON(num_bytes > 4);
+	for (i = 0; i < num_bytes; i++)
+		reg_pair[i] = (value >> (8 * i)) & 0xff;
+}
 
-	DRM_DEBUG_KMS("[%d] %s\n", __LINE__, __func__);
+static void hdmi_v13_mode_set(struct hdmi_context *hdata,
+			struct drm_display_mode *m)
+{
+	struct hdmi_v13_core_regs *core = &hdata->mode_conf.conf.v13_conf.core;
+	struct hdmi_tg_regs *tg = &hdata->mode_conf.conf.v13_conf.tg;
+	unsigned int val;
 
-	drm_mode_set_crtcinfo(adjusted_mode, 0);
+	hdata->mode_conf.cea_video_id =
+		drm_match_cea_mode((struct drm_display_mode *)m);
+	hdata->mode_conf.pixel_clock = m->clock * 1000;
 
-	if (hdata->type == HDMI_TYPE13)
-		index = hdmi_v13_conf_index(adjusted_mode);
-	else
-		index = hdmi_v14_find_phy_conf(adjusted_mode->clock * 1000);
+	hdmi_set_reg(core->h_blank, 2, m->htotal - m->hdisplay);
+	hdmi_set_reg(core->h_v_line, 3, (m->htotal << 12) | m->vtotal);
 
-	/* just return if user desired mode exists. */
-	if (index >= 0)
-		return;
+	val = (m->flags & DRM_MODE_FLAG_NVSYNC) ? 1 : 0;
+	hdmi_set_reg(core->vsync_pol, 1, val);
+
+	val = (m->flags & DRM_MODE_FLAG_INTERLACE) ? 1 : 0;
+	hdmi_set_reg(core->int_pro_mode, 1, val);
+
+	val = (m->hsync_start - m->hdisplay - 2);
+	val |= ((m->hsync_end - m->hdisplay - 2) << 10);
+	val |= ((m->flags & DRM_MODE_FLAG_NHSYNC)  ? 1 : 0)<<20;
+	hdmi_set_reg(core->h_sync_gen, 3, val);
 
 	/*
-	 * otherwise, find the most suitable mode among modes and change it
-	 * to adjusted_mode.
+	 * Quirk requirement for exynos HDMI IP design,
+	 * 2 pixels less than the actual calculation for hsync_start
+	 * and end.
 	 */
-	list_for_each_entry(m, &connector->modes, head) {
-		if (hdata->type == HDMI_TYPE13)
-			index = hdmi_v13_conf_index(m);
-		else
-			index = hdmi_v14_find_phy_conf(m->clock * 1000);
-
-		if (index >= 0) {
-			struct drm_mode_object base;
-			struct list_head head;
-
-			DRM_INFO("desired mode doesn't exist so\n");
-			DRM_INFO("use the most suitable mode among modes.\n");
-
-			DRM_DEBUG_KMS("Adjusted Mode: [%d]x[%d] [%d]Hz\n",
-				m->hdisplay, m->vdisplay, m->vrefresh);
-
-			/* preserve display mode header while copying. */
-			head = adjusted_mode->head;
-			base = adjusted_mode->base;
-			memcpy(adjusted_mode, m, sizeof(*m));
-			adjusted_mode->head = head;
-			adjusted_mode->base = base;
-			break;
-		}
+
+	/* Following values & calculations differ for different type of modes */
+	if (m->flags & DRM_MODE_FLAG_INTERLACE) {
+		/* Interlaced Mode */
+		val = ((m->vsync_end - m->vdisplay) / 2);
+		val |= ((m->vsync_start - m->vdisplay) / 2) << 12;
+		hdmi_set_reg(core->v_sync_gen1, 3, val);
+
+		val = m->vtotal / 2;
+		val |= ((m->vtotal - m->vdisplay) / 2) << 11;
+		hdmi_set_reg(core->v_blank, 3, val);
+
+		val = (m->vtotal +
+			((m->vsync_end - m->vsync_start) * 4) + 5) / 2;
+		val |= m->vtotal << 11;
+		hdmi_set_reg(core->v_blank_f, 3, val);
+
+		val = ((m->vtotal / 2) + 7);
+		val |= ((m->vtotal / 2) + 2) << 12;
+		hdmi_set_reg(core->v_sync_gen2, 3, val);
+
+		val = ((m->htotal / 2) + (m->hsync_start - m->hdisplay));
+		val |= ((m->htotal / 2) +
+			(m->hsync_start - m->hdisplay)) << 12;
+		hdmi_set_reg(core->v_sync_gen3, 3, val);
+
+		hdmi_set_reg(tg->vact_st, 2, (m->vtotal - m->vdisplay) / 2);
+		hdmi_set_reg(tg->vact_sz, 2, m->vdisplay / 2);
+
+		hdmi_set_reg(tg->vact_st2, 2, 0x249);/* Reset value + 1*/
+	} else {
+		/* Progressive Mode */
+
+		val = m->vtotal;
+		val |= (m->vtotal - m->vdisplay) << 11;
+		hdmi_set_reg(core->v_blank, 3, val);
+
+		hdmi_set_reg(core->v_blank_f, 3, 0);
+
+		val = (m->vsync_end - m->vdisplay);
+		val |= ((m->vsync_start - m->vdisplay) << 12);
+		hdmi_set_reg(core->v_sync_gen1, 3, val);
+
+		hdmi_set_reg(core->v_sync_gen2, 3, 0x1001);/* Reset value  */
+		hdmi_set_reg(core->v_sync_gen3, 3, 0x1001);/* Reset value  */
+		hdmi_set_reg(tg->vact_st, 2, m->vtotal - m->vdisplay);
+		hdmi_set_reg(tg->vact_sz, 2, m->vdisplay);
+		hdmi_set_reg(tg->vact_st2, 2, 0x248); /* Reset value */
 	}
-}
 
-static void hdmi_set_reg(u8 *reg_pair, int num_bytes, u32 value)
-{
-	int i;
-	BUG_ON(num_bytes > 4);
-	for (i = 0; i < num_bytes; i++)
-		reg_pair[i] = (value >> (8 * i)) & 0xff;
+	/* Timing generator registers */
+	hdmi_set_reg(tg->cmd, 1, 0x0);
+	hdmi_set_reg(tg->h_fsz, 2, m->htotal);
+	hdmi_set_reg(tg->hact_st, 2, m->htotal - m->hdisplay);
+	hdmi_set_reg(tg->hact_sz, 2, m->hdisplay);
+	hdmi_set_reg(tg->v_fsz, 2, m->vtotal);
+	hdmi_set_reg(tg->vsync, 2, 0x1);
+	hdmi_set_reg(tg->vsync2, 2, 0x233); /* Reset value */
+	hdmi_set_reg(tg->field_chg, 2, 0x233); /* Reset value */
+	hdmi_set_reg(tg->vsync_top_hdmi, 2, 0x1); /* Reset value */
+	hdmi_set_reg(tg->vsync_bot_hdmi, 2, 0x233); /* Reset value */
+	hdmi_set_reg(tg->field_top_hdmi, 2, 0x1); /* Reset value */
+	hdmi_set_reg(tg->field_bot_hdmi, 2, 0x233); /* Reset value */
+	hdmi_set_reg(tg->tg_3d, 1, 0x0); /* Not used */
 }
 
 static void hdmi_v14_mode_set(struct hdmi_context *hdata,
 			struct drm_display_mode *m)
 {
-	struct hdmi_core_regs *core = &hdata->mode_conf.core;
-	struct hdmi_tg_regs *tg = &hdata->mode_conf.tg;
-
-	hdata->mode_conf.cea_video_id = drm_match_cea_mode(m);
+	struct hdmi_tg_regs *tg = &hdata->mode_conf.conf.v14_conf.tg;
+	struct hdmi_v14_core_regs *core =
+		&hdata->mode_conf.conf.v14_conf.core;
 
+	hdata->mode_conf.cea_video_id =
+		drm_match_cea_mode((struct drm_display_mode *)m);
 	hdata->mode_conf.pixel_clock = m->clock * 1000;
+
 	hdmi_set_reg(core->h_blank, 2, m->htotal - m->hdisplay);
 	hdmi_set_reg(core->v_line, 2, m->vtotal);
 	hdmi_set_reg(core->h_line, 2, m->htotal);
@@ -1852,25 +1640,22 @@ static void hdmi_v14_mode_set(struct hdmi_context *hdata,
 	hdmi_set_reg(tg->field_top_hdmi, 2, 0x1); /* Reset value */
 	hdmi_set_reg(tg->field_bot_hdmi, 2, 0x233); /* Reset value */
 	hdmi_set_reg(tg->tg_3d, 1, 0x0);
-
 }
 
 static void hdmi_mode_set(void *ctx, void *mode)
 {
 	struct hdmi_context *hdata = ctx;
-	int conf_idx;
+	struct drm_display_mode *m = mode;
 
-	DRM_DEBUG_KMS("[%d] %s\n", __LINE__, __func__);
+	DRM_DEBUG_KMS("[%s]: xres=%d, yres=%d, refresh=%d, intl=%s\n",
+		__func__, m->hdisplay, m->vdisplay,
+		m->vrefresh, (m->flags & DRM_MODE_FLAG_INTERLACE) ?
+		"INTERLACED" : "PROGERESSIVE");
 
-	if (hdata->type == HDMI_TYPE13) {
-		conf_idx = hdmi_v13_conf_index(mode);
-		if (conf_idx >= 0)
-			hdata->cur_conf = conf_idx;
-		else
-			DRM_DEBUG_KMS("not supported mode\n");
-	} else {
+	if (hdata->type == HDMI_TYPE13)
+		hdmi_v13_mode_set(hdata, mode);
+	else
 		hdmi_v14_mode_set(hdata, mode);
-	}
 }
 
 static void hdmi_get_max_resol(void *ctx, unsigned int *width,
@@ -1983,7 +1768,6 @@ static struct exynos_hdmi_ops hdmi_ops = {
 	.check_timing	= hdmi_check_timing,
 
 	/* manager */
-	.mode_fixup	= hdmi_mode_fixup,
 	.mode_set	= hdmi_mode_set,
 	.get_max_resol	= hdmi_get_max_resol,
 	.commit		= hdmi_commit,
@@ -2023,27 +1807,27 @@ static int hdmi_resources_init(struct hdmi_context *hdata)
 
 	/* get clocks, power */
 	res->hdmi = devm_clk_get(dev, "hdmi");
-	if (IS_ERR_OR_NULL(res->hdmi)) {
+	if (IS_ERR(res->hdmi)) {
 		DRM_ERROR("failed to get clock 'hdmi'\n");
 		goto fail;
 	}
 	res->sclk_hdmi = devm_clk_get(dev, "sclk_hdmi");
-	if (IS_ERR_OR_NULL(res->sclk_hdmi)) {
+	if (IS_ERR(res->sclk_hdmi)) {
 		DRM_ERROR("failed to get clock 'sclk_hdmi'\n");
 		goto fail;
 	}
 	res->sclk_pixel = devm_clk_get(dev, "sclk_pixel");
-	if (IS_ERR_OR_NULL(res->sclk_pixel)) {
+	if (IS_ERR(res->sclk_pixel)) {
 		DRM_ERROR("failed to get clock 'sclk_pixel'\n");
 		goto fail;
 	}
 	res->sclk_hdmiphy = devm_clk_get(dev, "sclk_hdmiphy");
-	if (IS_ERR_OR_NULL(res->sclk_hdmiphy)) {
+	if (IS_ERR(res->sclk_hdmiphy)) {
 		DRM_ERROR("failed to get clock 'sclk_hdmiphy'\n");
 		goto fail;
 	}
 	res->hdmiphy = devm_clk_get(dev, "hdmiphy");
-	if (IS_ERR_OR_NULL(res->hdmiphy)) {
+	if (IS_ERR(res->hdmiphy)) {
 		DRM_ERROR("failed to get clock 'hdmiphy'\n");
 		goto fail;
 	}
diff --git a/drivers/gpu/drm/exynos/exynos_mixer.c b/drivers/gpu/drm/exynos/exynos_mixer.c
index 2f4f72f07047..ec3e376b7e01 100644
--- a/drivers/gpu/drm/exynos/exynos_mixer.c
+++ b/drivers/gpu/drm/exynos/exynos_mixer.c
@@ -643,12 +643,14 @@ static void mixer_win_reset(struct mixer_context *ctx)
 	/* setting graphical layers */
 	val  = MXR_GRP_CFG_COLOR_KEY_DISABLE; /* no blank key */
 	val |= MXR_GRP_CFG_WIN_BLEND_EN;
-	val |= MXR_GRP_CFG_BLEND_PRE_MUL;
-	val |= MXR_GRP_CFG_PIXEL_BLEND_EN;
 	val |= MXR_GRP_CFG_ALPHA_VAL(0xff); /* non-transparent alpha */
 
-	/* the same configuration for both layers */
+	/* Don't blend layer 0 onto the mixer background */
 	mixer_reg_write(res, MXR_GRAPHIC_CFG(0), val);
+
+	/* Blend layer 1 into layer 0 */
+	val |= MXR_GRP_CFG_BLEND_PRE_MUL;
+	val |= MXR_GRP_CFG_PIXEL_BLEND_EN;
 	mixer_reg_write(res, MXR_GRAPHIC_CFG(1), val);
 
 	/* setting video layers */
@@ -820,7 +822,6 @@ static void mixer_win_disable(void *ctx, int win)
 
 static int mixer_check_timing(void *ctx, struct fb_videomode *timing)
 {
-	struct mixer_context *mixer_ctx = ctx;
 	u32 w, h;
 
 	w = timing->xres;
@@ -831,9 +832,6 @@ static int mixer_check_timing(void *ctx, struct fb_videomode *timing)
 		timing->refresh, (timing->vmode &
 		FB_VMODE_INTERLACED) ? true : false);
 
-	if (mixer_ctx->mxr_ver == MXR_VER_0_0_0_16)
-		return 0;
-
 	if ((w >= 464 && w <= 720 && h >= 261 && h <= 576) ||
 		(w >= 1024 && w <= 1280 && h >= 576 && h <= 720) ||
 		(w >= 1664 && w <= 1920 && h >= 936 && h <= 1080))
@@ -1047,13 +1045,13 @@ static int mixer_resources_init(struct exynos_drm_hdmi_context *ctx,
 	spin_lock_init(&mixer_res->reg_slock);
 
 	mixer_res->mixer = devm_clk_get(dev, "mixer");
-	if (IS_ERR_OR_NULL(mixer_res->mixer)) {
+	if (IS_ERR(mixer_res->mixer)) {
 		dev_err(dev, "failed to get clock 'mixer'\n");
 		return -ENODEV;
 	}
 
 	mixer_res->sclk_hdmi = devm_clk_get(dev, "sclk_hdmi");
-	if (IS_ERR_OR_NULL(mixer_res->sclk_hdmi)) {
+	if (IS_ERR(mixer_res->sclk_hdmi)) {
 		dev_err(dev, "failed to get clock 'sclk_hdmi'\n");
 		return -ENODEV;
 	}
@@ -1096,17 +1094,17 @@ static int vp_resources_init(struct exynos_drm_hdmi_context *ctx,
 	struct resource *res;
 
 	mixer_res->vp = devm_clk_get(dev, "vp");
-	if (IS_ERR_OR_NULL(mixer_res->vp)) {
+	if (IS_ERR(mixer_res->vp)) {
 		dev_err(dev, "failed to get clock 'vp'\n");
 		return -ENODEV;
 	}
 	mixer_res->sclk_mixer = devm_clk_get(dev, "sclk_mixer");
-	if (IS_ERR_OR_NULL(mixer_res->sclk_mixer)) {
+	if (IS_ERR(mixer_res->sclk_mixer)) {
 		dev_err(dev, "failed to get clock 'sclk_mixer'\n");
 		return -ENODEV;
 	}
 	mixer_res->sclk_dac = devm_clk_get(dev, "sclk_dac");
-	if (IS_ERR_OR_NULL(mixer_res->sclk_dac)) {
+	if (IS_ERR(mixer_res->sclk_dac)) {
 		dev_err(dev, "failed to get clock 'sclk_dac'\n");
 		return -ENODEV;
 	}
diff --git a/drivers/gpu/drm/exynos/regs-fimc.h b/drivers/gpu/drm/exynos/regs-fimc.h
index b4f9ca1fd851..30496134a3d0 100644
--- a/drivers/gpu/drm/exynos/regs-fimc.h
+++ b/drivers/gpu/drm/exynos/regs-fimc.h
@@ -661,9 +661,8 @@
 #define EXYNOS_CLKSRC_SCLK				(1 << 1)
 
 /* SYSREG for FIMC writeback */
-#define SYSREG_CAMERA_BLK			(S3C_VA_SYS + 0x0218)
-#define SYSREG_ISP_BLK				(S3C_VA_SYS + 0x020c)
-#define SYSREG_FIMD0WB_DEST_MASK	(0x3 << 23)
-#define SYSREG_FIMD0WB_DEST_SHIFT	23
+#define SYSREG_CAMERA_BLK			(0x0218)
+#define SYSREG_FIMD0WB_DEST_MASK		(0x3 << 23)
+#define SYSREG_FIMD0WB_DEST_SHIFT		23
 
 #endif /* EXYNOS_REGS_FIMC_H */
diff --git a/drivers/gpu/drm/gma500/Kconfig b/drivers/gpu/drm/gma500/Kconfig
index 1188f0fe7e4f..1f6e2dfaaeae 100644
--- a/drivers/gpu/drm/gma500/Kconfig
+++ b/drivers/gpu/drm/gma500/Kconfig
@@ -2,10 +2,15 @@ config DRM_GMA500
 	tristate "Intel GMA5/600 KMS Framebuffer"
 	depends on DRM && PCI && X86
 	select FB_CFB_COPYAREA
-        select FB_CFB_FILLRECT
-        select FB_CFB_IMAGEBLIT
-        select DRM_KMS_HELPER
-        select DRM_TTM
+	select FB_CFB_FILLRECT
+	select FB_CFB_IMAGEBLIT
+	select DRM_KMS_HELPER
+	select DRM_TTM
+	# GMA500 depends on ACPI_VIDEO when ACPI is enabled, just like i915
+	select ACPI_VIDEO if ACPI
+	select BACKLIGHT_CLASS_DEVICE if ACPI
+	select VIDEO_OUTPUT_CONTROL if ACPI
+	select INPUT if ACPI
 	help
 	  Say yes for an experimental 2D KMS framebuffer driver for the
 	  Intel GMA500 ('Poulsbo') and other Intel IMG based graphics
diff --git a/drivers/gpu/drm/gma500/cdv_intel_crt.c b/drivers/gpu/drm/gma500/cdv_intel_crt.c
index 8c175345d85c..7b8386fc3024 100644
--- a/drivers/gpu/drm/gma500/cdv_intel_crt.c
+++ b/drivers/gpu/drm/gma500/cdv_intel_crt.c
@@ -276,6 +276,7 @@ void cdv_intel_crt_init(struct drm_device *dev,
 		goto failed_connector;
 
 	connector = &psb_intel_connector->base;
+	connector->polled = DRM_CONNECTOR_POLL_HPD;
 	drm_connector_init(dev, connector,
 		&cdv_intel_crt_connector_funcs, DRM_MODE_CONNECTOR_VGA);
 
diff --git a/drivers/gpu/drm/gma500/cdv_intel_hdmi.c b/drivers/gpu/drm/gma500/cdv_intel_hdmi.c
index e223b500022e..464153d9d2df 100644
--- a/drivers/gpu/drm/gma500/cdv_intel_hdmi.c
+++ b/drivers/gpu/drm/gma500/cdv_intel_hdmi.c
@@ -319,6 +319,7 @@ void cdv_hdmi_init(struct drm_device *dev,
 		goto err_priv;
 
 	connector = &psb_intel_connector->base;
+	connector->polled = DRM_CONNECTOR_POLL_HPD;
 	encoder = &psb_intel_encoder->base;
 	drm_connector_init(dev, connector,
 			   &cdv_hdmi_connector_funcs,
diff --git a/drivers/gpu/drm/gma500/framebuffer.c b/drivers/gpu/drm/gma500/framebuffer.c
index 2590cac84257..1534e220097a 100644
--- a/drivers/gpu/drm/gma500/framebuffer.c
+++ b/drivers/gpu/drm/gma500/framebuffer.c
@@ -431,7 +431,7 @@ static int psbfb_create(struct psb_fbdev *fbdev,
 	fbdev->psb_fb_helper.fbdev = info;
 
 	drm_fb_helper_fill_fix(info, fb->pitches[0], fb->depth);
-	strcpy(info->fix.id, "psbfb");
+	strcpy(info->fix.id, "psbdrmfb");
 
 	info->flags = FBINFO_DEFAULT;
 	if (dev_priv->ops->accel_2d && pitch_lines > 8)	/* 2D engine */
@@ -772,8 +772,8 @@ void psb_modeset_init(struct drm_device *dev)
 	for (i = 0; i < dev_priv->num_pipe; i++)
 		psb_intel_crtc_init(dev, i, mode_dev);
 
-	dev->mode_config.max_width = 2048;
-	dev->mode_config.max_height = 2048;
+	dev->mode_config.max_width = 4096;
+	dev->mode_config.max_height = 4096;
 
 	psb_setup_outputs(dev);
 
diff --git a/drivers/gpu/drm/gma500/gtt.c b/drivers/gpu/drm/gma500/gtt.c
index 054e26e769ec..1f82183536a3 100644
--- a/drivers/gpu/drm/gma500/gtt.c
+++ b/drivers/gpu/drm/gma500/gtt.c
@@ -80,7 +80,8 @@ static u32 __iomem *psb_gtt_entry(struct drm_device *dev, struct gtt_range *r)
  *	the GTT. This is protected via the gtt mutex which the caller
  *	must hold.
  */
-static int psb_gtt_insert(struct drm_device *dev, struct gtt_range *r)
+static int psb_gtt_insert(struct drm_device *dev, struct gtt_range *r,
+			  int resume)
 {
 	u32 __iomem *gtt_slot;
 	u32 pte;
@@ -97,8 +98,10 @@ static int psb_gtt_insert(struct drm_device *dev, struct gtt_range *r)
 	gtt_slot = psb_gtt_entry(dev, r);
 	pages = r->pages;
 
-	/* Make sure changes are visible to the GPU */
-	set_pages_array_wc(pages, r->npage);
+	if (!resume) {
+		/* Make sure changes are visible to the GPU */
+		set_pages_array_wc(pages, r->npage);
+	}
 
 	/* Write our page entries into the GTT itself */
 	for (i = r->roll; i < r->npage; i++) {
@@ -269,7 +272,7 @@ int psb_gtt_pin(struct gtt_range *gt)
 		ret = psb_gtt_attach_pages(gt);
 		if (ret < 0)
 			goto out;
-		ret = psb_gtt_insert(dev, gt);
+		ret = psb_gtt_insert(dev, gt, 0);
 		if (ret < 0) {
 			psb_gtt_detach_pages(gt);
 			goto out;
@@ -421,9 +424,11 @@ int psb_gtt_init(struct drm_device *dev, int resume)
 	int ret = 0;
 	uint32_t pte;
 
-	mutex_init(&dev_priv->gtt_mutex);
+	if (!resume) {
+		mutex_init(&dev_priv->gtt_mutex);
+		psb_gtt_alloc(dev);
+	}
 
-	psb_gtt_alloc(dev);
 	pg = &dev_priv->gtt;
 
 	/* Enable the GTT */
@@ -505,7 +510,8 @@ int psb_gtt_init(struct drm_device *dev, int resume)
 	/*
 	 *	Map the GTT and the stolen memory area
 	 */
-	dev_priv->gtt_map = ioremap_nocache(pg->gtt_phys_start,
+	if (!resume)
+		dev_priv->gtt_map = ioremap_nocache(pg->gtt_phys_start,
 						gtt_pages << PAGE_SHIFT);
 	if (!dev_priv->gtt_map) {
 		dev_err(dev->dev, "Failure to map gtt.\n");
@@ -513,7 +519,9 @@ int psb_gtt_init(struct drm_device *dev, int resume)
 		goto out_err;
 	}
 
-	dev_priv->vram_addr = ioremap_wc(dev_priv->stolen_base, stolen_size);
+	if (!resume)
+		dev_priv->vram_addr = ioremap_wc(dev_priv->stolen_base,
+						 stolen_size);
 	if (!dev_priv->vram_addr) {
 		dev_err(dev->dev, "Failure to map stolen base.\n");
 		ret = -ENOMEM;
@@ -549,3 +557,31 @@ out_err:
 	psb_gtt_takedown(dev);
 	return ret;
 }
+
+int psb_gtt_restore(struct drm_device *dev)
+{
+	struct drm_psb_private *dev_priv = dev->dev_private;
+	struct resource *r = dev_priv->gtt_mem->child;
+	struct gtt_range *range;
+	unsigned int restored = 0, total = 0, size = 0;
+
+	/* On resume, the gtt_mutex is already initialized */
+	mutex_lock(&dev_priv->gtt_mutex);
+	psb_gtt_init(dev, 1);
+
+	while (r != NULL) {
+		range = container_of(r, struct gtt_range, resource);
+		if (range->pages) {
+			psb_gtt_insert(dev, range, 1);
+			size += range->resource.end - range->resource.start;
+			restored++;
+		}
+		r = r->sibling;
+		total++;
+	}
+	mutex_unlock(&dev_priv->gtt_mutex);
+	DRM_DEBUG_DRIVER("Restored %u of %u gtt ranges (%u KB)", restored,
+			 total, (size / 1024));
+
+	return 0;
+}
diff --git a/drivers/gpu/drm/gma500/gtt.h b/drivers/gpu/drm/gma500/gtt.h
index aa1742387f5a..6191d10acf33 100644
--- a/drivers/gpu/drm/gma500/gtt.h
+++ b/drivers/gpu/drm/gma500/gtt.h
@@ -60,5 +60,5 @@ extern int psb_gtt_pin(struct gtt_range *gt);
 extern void psb_gtt_unpin(struct gtt_range *gt);
 extern void psb_gtt_roll(struct drm_device *dev,
 					struct gtt_range *gt, int roll);
-
+extern int psb_gtt_restore(struct drm_device *dev);
 #endif
diff --git a/drivers/gpu/drm/gma500/intel_bios.c b/drivers/gpu/drm/gma500/intel_bios.c
index 403fffb03abd..d3497348c4d5 100644
--- a/drivers/gpu/drm/gma500/intel_bios.c
+++ b/drivers/gpu/drm/gma500/intel_bios.c
@@ -218,12 +218,11 @@ static void parse_backlight_data(struct drm_psb_private *dev_priv,
 	bl_start = find_section(bdb, BDB_LVDS_BACKLIGHT);
 	vbt_lvds_bl = (struct bdb_lvds_backlight *)(bl_start + 1) + p_type;
 
-	lvds_bl = kzalloc(sizeof(*vbt_lvds_bl), GFP_KERNEL);
+	lvds_bl = kmemdup(vbt_lvds_bl, sizeof(*vbt_lvds_bl), GFP_KERNEL);
 	if (!lvds_bl) {
 		dev_err(dev_priv->dev->dev, "out of memory for backlight data\n");
 		return;
 	}
-	memcpy(lvds_bl, vbt_lvds_bl, sizeof(*vbt_lvds_bl));
 	dev_priv->lvds_bl = lvds_bl;
 }
 
diff --git a/drivers/gpu/drm/gma500/intel_bios.h b/drivers/gpu/drm/gma500/intel_bios.h
index c6267c98c9e7..978ae4b25e82 100644
--- a/drivers/gpu/drm/gma500/intel_bios.h
+++ b/drivers/gpu/drm/gma500/intel_bios.h
@@ -19,8 +19,8 @@
  *
  */
 
-#ifndef _I830_BIOS_H_
-#define _I830_BIOS_H_
+#ifndef _INTEL_BIOS_H_
+#define _INTEL_BIOS_H_
 
 #include <drm/drmP.h>
 #include <drm/drm_dp_helper.h>
@@ -618,4 +618,4 @@ extern void psb_intel_destroy_bios(struct drm_device *dev);
 #define		PORT_IDPC	8
 #define		PORT_IDPD	9
 
-#endif /* _I830_BIOS_H_ */
+#endif /* _INTEL_BIOS_H_ */
diff --git a/drivers/gpu/drm/gma500/mdfld_dsi_output.c b/drivers/gpu/drm/gma500/mdfld_dsi_output.c
index 2d4ab48f07a2..3abf8315f57c 100644
--- a/drivers/gpu/drm/gma500/mdfld_dsi_output.c
+++ b/drivers/gpu/drm/gma500/mdfld_dsi_output.c
@@ -92,8 +92,8 @@ void mdfld_dsi_brightness_init(struct mdfld_dsi_config *dsi_config, int pipe)
 {
 	struct mdfld_dsi_pkg_sender *sender =
 				mdfld_dsi_get_pkg_sender(dsi_config);
-	struct drm_device *dev = sender->dev;
-	struct drm_psb_private *dev_priv = dev->dev_private;
+	struct drm_device *dev;
+	struct drm_psb_private *dev_priv;
 	u32 gen_ctrl_val;
 
 	if (!sender) {
@@ -101,6 +101,9 @@ void mdfld_dsi_brightness_init(struct mdfld_dsi_config *dsi_config, int pipe)
 		return;
 	}
 
+	dev = sender->dev;
+	dev_priv = dev->dev_private;
+
 	/* Set default display backlight value to 85% (0xd8)*/
 	mdfld_dsi_send_mcs_short(sender, write_display_brightness, 0xd8, 1,
 				true);
diff --git a/drivers/gpu/drm/gma500/power.c b/drivers/gpu/drm/gma500/power.c
index 889b854751da..b6b135fcd59c 100644
--- a/drivers/gpu/drm/gma500/power.c
+++ b/drivers/gpu/drm/gma500/power.c
@@ -110,6 +110,8 @@ static void gma_resume_display(struct pci_dev *pdev)
 	PSB_WVDC32(dev_priv->pge_ctl | _PSB_PGETBL_ENABLED, PSB_PGETBL_CTL);
 	pci_write_config_word(pdev, PSB_GMCH_CTRL,
 			dev_priv->gmch_ctrl | _PSB_GMCH_ENABLED);
+
+	psb_gtt_restore(dev); /* Rebuild our GTT mappings */
 	dev_priv->ops->restore_regs(dev);
 }
 
@@ -313,3 +315,18 @@ int psb_runtime_idle(struct device *dev)
 	else
 		return 1;
 }
+
+int gma_power_thaw(struct device *_dev)
+{
+	return gma_power_resume(_dev);
+}
+
+int gma_power_freeze(struct device *_dev)
+{
+	return gma_power_suspend(_dev);
+}
+
+int gma_power_restore(struct device *_dev)
+{
+	return gma_power_resume(_dev);
+}
diff --git a/drivers/gpu/drm/gma500/power.h b/drivers/gpu/drm/gma500/power.h
index 1969d2ecb328..56d8708bd41c 100644
--- a/drivers/gpu/drm/gma500/power.h
+++ b/drivers/gpu/drm/gma500/power.h
@@ -41,6 +41,9 @@ void gma_power_uninit(struct drm_device *dev);
  */
 int gma_power_suspend(struct device *dev);
 int gma_power_resume(struct device *dev);
+int gma_power_thaw(struct device *dev);
+int gma_power_freeze(struct device *dev);
+int gma_power_restore(struct device *_dev);
 
 /*
  * These are the functions the driver should use to wrap all hw access
diff --git a/drivers/gpu/drm/gma500/psb_drv.c b/drivers/gpu/drm/gma500/psb_drv.c
index 111e3df9c5de..bddea5807442 100644
--- a/drivers/gpu/drm/gma500/psb_drv.c
+++ b/drivers/gpu/drm/gma500/psb_drv.c
@@ -601,6 +601,9 @@ static void psb_remove(struct pci_dev *pdev)
 static const struct dev_pm_ops psb_pm_ops = {
 	.resume = gma_power_resume,
 	.suspend = gma_power_suspend,
+	.thaw = gma_power_thaw,
+	.freeze = gma_power_freeze,
+	.restore = gma_power_restore,
 	.runtime_suspend = psb_runtime_suspend,
 	.runtime_resume = psb_runtime_resume,
 	.runtime_idle = psb_runtime_idle,
diff --git a/drivers/gpu/drm/gma500/psb_drv.h b/drivers/gpu/drm/gma500/psb_drv.h
index a7fd6c48b793..6053b8abcd12 100644
--- a/drivers/gpu/drm/gma500/psb_drv.h
+++ b/drivers/gpu/drm/gma500/psb_drv.h
@@ -876,7 +876,6 @@ extern const struct psb_ops cdv_chip_ops;
 #define PSB_D_MSVDX   (1 << 9)
 #define PSB_D_TOPAZ   (1 << 10)
 
-extern int drm_psb_no_fb;
 extern int drm_idle_check_interval;
 
 /*
diff --git a/drivers/gpu/drm/gma500/psb_intel_display.c b/drivers/gpu/drm/gma500/psb_intel_display.c
index 9edb1902a096..6e8f42b61ff6 100644
--- a/drivers/gpu/drm/gma500/psb_intel_display.c
+++ b/drivers/gpu/drm/gma500/psb_intel_display.c
@@ -50,119 +50,41 @@ struct psb_intel_p2_t {
 	int p2_slow, p2_fast;
 };
 
-#define INTEL_P2_NUM		      2
-
 struct psb_intel_limit_t {
 	struct psb_intel_range_t dot, vco, n, m, m1, m2, p, p1;
 	struct psb_intel_p2_t p2;
 };
 
-#define I8XX_DOT_MIN		  25000
-#define I8XX_DOT_MAX		 350000
-#define I8XX_VCO_MIN		 930000
-#define I8XX_VCO_MAX		1400000
-#define I8XX_N_MIN		      3
-#define I8XX_N_MAX		     16
-#define I8XX_M_MIN		     96
-#define I8XX_M_MAX		    140
-#define I8XX_M1_MIN		     18
-#define I8XX_M1_MAX		     26
-#define I8XX_M2_MIN		      6
-#define I8XX_M2_MAX		     16
-#define I8XX_P_MIN		      4
-#define I8XX_P_MAX		    128
-#define I8XX_P1_MIN		      2
-#define I8XX_P1_MAX		     33
-#define I8XX_P1_LVDS_MIN	      1
-#define I8XX_P1_LVDS_MAX	      6
-#define I8XX_P2_SLOW		      4
-#define I8XX_P2_FAST		      2
-#define I8XX_P2_LVDS_SLOW	      14
-#define I8XX_P2_LVDS_FAST	      14	/* No fast option */
-#define I8XX_P2_SLOW_LIMIT	 165000
-
-#define I9XX_DOT_MIN		  20000
-#define I9XX_DOT_MAX		 400000
-#define I9XX_VCO_MIN		1400000
-#define I9XX_VCO_MAX		2800000
-#define I9XX_N_MIN		      1
-#define I9XX_N_MAX		      6
-#define I9XX_M_MIN		     70
-#define I9XX_M_MAX		    120
-#define I9XX_M1_MIN		      8
-#define I9XX_M1_MAX		     18
-#define I9XX_M2_MIN		      3
-#define I9XX_M2_MAX		      7
-#define I9XX_P_SDVO_DAC_MIN	      5
-#define I9XX_P_SDVO_DAC_MAX	     80
-#define I9XX_P_LVDS_MIN		      7
-#define I9XX_P_LVDS_MAX		     98
-#define I9XX_P1_MIN		      1
-#define I9XX_P1_MAX		      8
-#define I9XX_P2_SDVO_DAC_SLOW		     10
-#define I9XX_P2_SDVO_DAC_FAST		      5
-#define I9XX_P2_SDVO_DAC_SLOW_LIMIT	 200000
-#define I9XX_P2_LVDS_SLOW		     14
-#define I9XX_P2_LVDS_FAST		      7
-#define I9XX_P2_LVDS_SLOW_LIMIT		 112000
-
-#define INTEL_LIMIT_I8XX_DVO_DAC    0
-#define INTEL_LIMIT_I8XX_LVDS	    1
-#define INTEL_LIMIT_I9XX_SDVO_DAC   2
-#define INTEL_LIMIT_I9XX_LVDS	    3
+#define INTEL_LIMIT_I9XX_SDVO_DAC   0
+#define INTEL_LIMIT_I9XX_LVDS	    1
 
 static const struct psb_intel_limit_t psb_intel_limits[] = {
-	{			/* INTEL_LIMIT_I8XX_DVO_DAC */
-	 .dot = {.min = I8XX_DOT_MIN, .max = I8XX_DOT_MAX},
-	 .vco = {.min = I8XX_VCO_MIN, .max = I8XX_VCO_MAX},
-	 .n = {.min = I8XX_N_MIN, .max = I8XX_N_MAX},
-	 .m = {.min = I8XX_M_MIN, .max = I8XX_M_MAX},
-	 .m1 = {.min = I8XX_M1_MIN, .max = I8XX_M1_MAX},
-	 .m2 = {.min = I8XX_M2_MIN, .max = I8XX_M2_MAX},
-	 .p = {.min = I8XX_P_MIN, .max = I8XX_P_MAX},
-	 .p1 = {.min = I8XX_P1_MIN, .max = I8XX_P1_MAX},
-	 .p2 = {.dot_limit = I8XX_P2_SLOW_LIMIT,
-		.p2_slow = I8XX_P2_SLOW, .p2_fast = I8XX_P2_FAST},
-	 },
-	{			/* INTEL_LIMIT_I8XX_LVDS */
-	 .dot = {.min = I8XX_DOT_MIN, .max = I8XX_DOT_MAX},
-	 .vco = {.min = I8XX_VCO_MIN, .max = I8XX_VCO_MAX},
-	 .n = {.min = I8XX_N_MIN, .max = I8XX_N_MAX},
-	 .m = {.min = I8XX_M_MIN, .max = I8XX_M_MAX},
-	 .m1 = {.min = I8XX_M1_MIN, .max = I8XX_M1_MAX},
-	 .m2 = {.min = I8XX_M2_MIN, .max = I8XX_M2_MAX},
-	 .p = {.min = I8XX_P_MIN, .max = I8XX_P_MAX},
-	 .p1 = {.min = I8XX_P1_LVDS_MIN, .max = I8XX_P1_LVDS_MAX},
-	 .p2 = {.dot_limit = I8XX_P2_SLOW_LIMIT,
-		.p2_slow = I8XX_P2_LVDS_SLOW, .p2_fast = I8XX_P2_LVDS_FAST},
-	 },
 	{			/* INTEL_LIMIT_I9XX_SDVO_DAC */
-	 .dot = {.min = I9XX_DOT_MIN, .max = I9XX_DOT_MAX},
-	 .vco = {.min = I9XX_VCO_MIN, .max = I9XX_VCO_MAX},
-	 .n = {.min = I9XX_N_MIN, .max = I9XX_N_MAX},
-	 .m = {.min = I9XX_M_MIN, .max = I9XX_M_MAX},
-	 .m1 = {.min = I9XX_M1_MIN, .max = I9XX_M1_MAX},
-	 .m2 = {.min = I9XX_M2_MIN, .max = I9XX_M2_MAX},
-	 .p = {.min = I9XX_P_SDVO_DAC_MIN, .max = I9XX_P_SDVO_DAC_MAX},
-	 .p1 = {.min = I9XX_P1_MIN, .max = I9XX_P1_MAX},
-	 .p2 = {.dot_limit = I9XX_P2_SDVO_DAC_SLOW_LIMIT,
-		.p2_slow = I9XX_P2_SDVO_DAC_SLOW, .p2_fast =
-		I9XX_P2_SDVO_DAC_FAST},
+	 .dot = {.min = 20000, .max = 400000},
+	 .vco = {.min = 1400000, .max = 2800000},
+	 .n = {.min = 1, .max = 6},
+	 .m = {.min = 70, .max = 120},
+	 .m1 = {.min = 8, .max = 18},
+	 .m2 = {.min = 3, .max = 7},
+	 .p = {.min = 5, .max = 80},
+	 .p1 = {.min = 1, .max = 8},
+	 .p2 = {.dot_limit = 200000,
+		.p2_slow = 10, .p2_fast = 5},
 	 },
 	{			/* INTEL_LIMIT_I9XX_LVDS */
-	 .dot = {.min = I9XX_DOT_MIN, .max = I9XX_DOT_MAX},
-	 .vco = {.min = I9XX_VCO_MIN, .max = I9XX_VCO_MAX},
-	 .n = {.min = I9XX_N_MIN, .max = I9XX_N_MAX},
-	 .m = {.min = I9XX_M_MIN, .max = I9XX_M_MAX},
-	 .m1 = {.min = I9XX_M1_MIN, .max = I9XX_M1_MAX},
-	 .m2 = {.min = I9XX_M2_MIN, .max = I9XX_M2_MAX},
-	 .p = {.min = I9XX_P_LVDS_MIN, .max = I9XX_P_LVDS_MAX},
-	 .p1 = {.min = I9XX_P1_MIN, .max = I9XX_P1_MAX},
+	 .dot = {.min = 20000, .max = 400000},
+	 .vco = {.min = 1400000, .max = 2800000},
+	 .n = {.min = 1, .max = 6},
+	 .m = {.min = 70, .max = 120},
+	 .m1 = {.min = 8, .max = 18},
+	 .m2 = {.min = 3, .max = 7},
+	 .p = {.min = 7, .max = 98},
+	 .p1 = {.min = 1, .max = 8},
 	 /* The single-channel range is 25-112Mhz, and dual-channel
 	  * is 80-224Mhz.  Prefer single channel as much as possible.
 	  */
-	 .p2 = {.dot_limit = I9XX_P2_LVDS_SLOW_LIMIT,
-		.p2_slow = I9XX_P2_LVDS_SLOW, .p2_fast = I9XX_P2_LVDS_FAST},
+	 .p2 = {.dot_limit = 112000,
+		.p2_slow = 14, .p2_fast = 7},
 	 },
 };
 
@@ -177,9 +99,7 @@ static const struct psb_intel_limit_t *psb_intel_limit(struct drm_crtc *crtc)
 	return limit;
 }
 
-/** Derive the pixel clock for the given refclk and divisors for 8xx chips. */
-
-static void i8xx_clock(int refclk, struct psb_intel_clock_t *clock)
+static void psb_intel_clock(int refclk, struct psb_intel_clock_t *clock)
 {
 	clock->m = 5 * (clock->m1 + 2) + (clock->m2 + 2);
 	clock->p = clock->p1 * clock->p2;
@@ -187,22 +107,6 @@ static void i8xx_clock(int refclk, struct psb_intel_clock_t *clock)
 	clock->dot = clock->vco / clock->p;
 }
 
-/** Derive the pixel clock for the given refclk and divisors for 9xx chips. */
-
-static void i9xx_clock(int refclk, struct psb_intel_clock_t *clock)
-{
-	clock->m = 5 * (clock->m1 + 2) + (clock->m2 + 2);
-	clock->p = clock->p1 * clock->p2;
-	clock->vco = refclk * clock->m / (clock->n + 2);
-	clock->dot = clock->vco / clock->p;
-}
-
-static void psb_intel_clock(struct drm_device *dev, int refclk,
-			struct psb_intel_clock_t *clock)
-{
-	return i9xx_clock(refclk, clock);
-}
-
 /**
  * Returns whether any output on the specified pipe is of the specified type
  */
@@ -308,7 +212,7 @@ static bool psb_intel_find_best_PLL(struct drm_crtc *crtc, int target,
 				     clock.p1++) {
 					int this_err;
 
-					psb_intel_clock(dev, refclk, &clock);
+					psb_intel_clock(refclk, &clock);
 
 					if (!psb_intel_PLL_is_valid
 					    (crtc, &clock))
@@ -1068,7 +972,7 @@ static int psb_intel_crtc_cursor_move(struct drm_crtc *crtc, int x, int y)
 	return 0;
 }
 
-void psb_intel_crtc_gamma_set(struct drm_crtc *crtc, u16 *red,
+static void psb_intel_crtc_gamma_set(struct drm_crtc *crtc, u16 *red,
 			 u16 *green, u16 *blue, uint32_t type, uint32_t size)
 {
 	struct psb_intel_crtc *psb_intel_crtc = to_psb_intel_crtc(crtc);
@@ -1149,9 +1053,9 @@ static int psb_intel_crtc_clock_get(struct drm_device *dev,
 		if ((dpll & PLL_REF_INPUT_MASK) ==
 		    PLLB_REF_INPUT_SPREADSPECTRUMIN) {
 			/* XXX: might not be 66MHz */
-			i8xx_clock(66000, &clock);
+			psb_intel_clock(66000, &clock);
 		} else
-			i8xx_clock(48000, &clock);
+			psb_intel_clock(48000, &clock);
 	} else {
 		if (dpll & PLL_P1_DIVIDE_BY_TWO)
 			clock.p1 = 2;
@@ -1166,7 +1070,7 @@ static int psb_intel_crtc_clock_get(struct drm_device *dev,
 		else
 			clock.p2 = 2;
 
-		i8xx_clock(48000, &clock);
+		psb_intel_clock(48000, &clock);
 	}
 
 	/* XXX: It would be nice to validate the clocks, but we can't reuse
@@ -1225,7 +1129,7 @@ struct drm_display_mode *psb_intel_crtc_mode_get(struct drm_device *dev,
 	return mode;
 }
 
-void psb_intel_crtc_destroy(struct drm_crtc *crtc)
+static void psb_intel_crtc_destroy(struct drm_crtc *crtc)
 {
 	struct psb_intel_crtc *psb_intel_crtc = to_psb_intel_crtc(crtc);
 	struct gtt_range *gt;
diff --git a/drivers/gpu/drm/gma500/psb_intel_display.h b/drivers/gpu/drm/gma500/psb_intel_display.h
index 535b49a5e409..3724b971e91c 100644
--- a/drivers/gpu/drm/gma500/psb_intel_display.h
+++ b/drivers/gpu/drm/gma500/psb_intel_display.h
@@ -21,8 +21,5 @@
 #define _INTEL_DISPLAY_H_
 
 bool psb_intel_pipe_has_type(struct drm_crtc *crtc, int type);
-void psb_intel_crtc_gamma_set(struct drm_crtc *crtc, u16 *red,
-			 u16 *green, u16 *blue, uint32_t type, uint32_t size);
-void psb_intel_crtc_destroy(struct drm_crtc *crtc);
 
 #endif
diff --git a/drivers/gpu/drm/gma500/psb_intel_drv.h b/drivers/gpu/drm/gma500/psb_intel_drv.h
index 90f2d11e686b..4dcae421a58d 100644
--- a/drivers/gpu/drm/gma500/psb_intel_drv.h
+++ b/drivers/gpu/drm/gma500/psb_intel_drv.h
@@ -32,9 +32,6 @@
 /* maximum connectors per crtcs in the mode set */
 #define INTELFB_CONN_LIMIT 4
 
-#define INTEL_I2C_BUS_DVO 1
-#define INTEL_I2C_BUS_SDVO 2
-
 /* Intel Pipe Clone Bit */
 #define INTEL_HDMIB_CLONE_BIT 1
 #define INTEL_HDMIC_CLONE_BIT 2
@@ -68,11 +65,6 @@
 #define INTEL_OUTPUT_DISPLAYPORT 9
 #define INTEL_OUTPUT_EDP 10
 
-#define INTEL_DVO_CHIP_NONE 0
-#define INTEL_DVO_CHIP_LVDS 1
-#define INTEL_DVO_CHIP_TMDS 2
-#define INTEL_DVO_CHIP_TVOUT 4
-
 #define INTEL_MODE_PIXEL_MULTIPLIER_SHIFT (0x0)
 #define INTEL_MODE_PIXEL_MULTIPLIER_MASK (0xf << INTEL_MODE_PIXEL_MULTIPLIER_SHIFT)
 
diff --git a/drivers/gpu/drm/gma500/psb_intel_reg.h b/drivers/gpu/drm/gma500/psb_intel_reg.h
index d914719c4b60..0be30e4d146d 100644
--- a/drivers/gpu/drm/gma500/psb_intel_reg.h
+++ b/drivers/gpu/drm/gma500/psb_intel_reg.h
@@ -493,7 +493,6 @@
 #define PIPEACONF_DISABLE		0
 #define PIPEACONF_DOUBLE_WIDE		(1 << 30)
 #define PIPECONF_ACTIVE			(1 << 30)
-#define I965_PIPECONF_ACTIVE		(1 << 30)
 #define PIPECONF_DSIPLL_LOCK		(1 << 29)
 #define PIPEACONF_SINGLE_WIDE		0
 #define PIPEACONF_PIPE_UNLOCKED		0
diff --git a/drivers/gpu/drm/gma500/psb_intel_sdvo.c b/drivers/gpu/drm/gma500/psb_intel_sdvo.c
index a4cc777ab7a6..19e36603b23b 100644
--- a/drivers/gpu/drm/gma500/psb_intel_sdvo.c
+++ b/drivers/gpu/drm/gma500/psb_intel_sdvo.c
@@ -134,6 +134,9 @@ struct psb_intel_sdvo {
 
 	/* Input timings for adjusted_mode */
 	struct psb_intel_sdvo_dtd input_dtd;
+
+	/* Saved SDVO output states */
+	uint32_t saveSDVO; /* Can be SDVOB or SDVOC depending on sdvo_reg */
 };
 
 struct psb_intel_sdvo_connector {
@@ -1830,6 +1833,34 @@ done:
 #undef CHECK_PROPERTY
 }
 
+static void psb_intel_sdvo_save(struct drm_connector *connector)
+{
+	struct drm_device *dev = connector->dev;
+	struct psb_intel_encoder *psb_intel_encoder =
+					psb_intel_attached_encoder(connector);
+	struct psb_intel_sdvo *sdvo =
+				to_psb_intel_sdvo(&psb_intel_encoder->base);
+
+	sdvo->saveSDVO = REG_READ(sdvo->sdvo_reg);
+}
+
+static void psb_intel_sdvo_restore(struct drm_connector *connector)
+{
+	struct drm_device *dev = connector->dev;
+	struct drm_encoder *encoder =
+				&psb_intel_attached_encoder(connector)->base;
+	struct psb_intel_sdvo *sdvo = to_psb_intel_sdvo(encoder);
+	struct drm_crtc *crtc = encoder->crtc;
+
+	REG_WRITE(sdvo->sdvo_reg, sdvo->saveSDVO);
+
+	/* Force a full mode set on the crtc. We're supposed to have the
+	   mode_config lock already. */
+	if (connector->status == connector_status_connected)
+		drm_crtc_helper_set_mode(crtc, &crtc->mode, crtc->x, crtc->y,
+					 NULL);
+}
+
 static const struct drm_encoder_helper_funcs psb_intel_sdvo_helper_funcs = {
 	.dpms = psb_intel_sdvo_dpms,
 	.mode_fixup = psb_intel_sdvo_mode_fixup,
@@ -1840,6 +1871,8 @@ static const struct drm_encoder_helper_funcs psb_intel_sdvo_helper_funcs = {
 
 static const struct drm_connector_funcs psb_intel_sdvo_connector_funcs = {
 	.dpms = drm_helper_connector_dpms,
+	.save = psb_intel_sdvo_save,
+	.restore = psb_intel_sdvo_restore,
 	.detect = psb_intel_sdvo_detect,
 	.fill_modes = drm_helper_probe_single_connector_modes,
 	.set_property = psb_intel_sdvo_set_property,
diff --git a/drivers/gpu/drm/gma500/psb_irq.c b/drivers/gpu/drm/gma500/psb_irq.c
index 8652cdf3f03f..029eccf30137 100644
--- a/drivers/gpu/drm/gma500/psb_irq.c
+++ b/drivers/gpu/drm/gma500/psb_irq.c
@@ -211,7 +211,7 @@ irqreturn_t psb_irq_handler(DRM_IRQ_ARGS)
 
 	vdc_stat = PSB_RVDC32(PSB_INT_IDENTITY_R);
 
-	if (vdc_stat & _PSB_PIPE_EVENT_FLAG)
+	if (vdc_stat & (_PSB_PIPE_EVENT_FLAG|_PSB_IRQ_ASLE))
 		dsp_int = 1;
 
 	/* FIXME: Handle Medfield
diff --git a/drivers/gpu/drm/gma500/psb_irq.h b/drivers/gpu/drm/gma500/psb_irq.h
index 603045bee58a..debb7f190c06 100644
--- a/drivers/gpu/drm/gma500/psb_irq.h
+++ b/drivers/gpu/drm/gma500/psb_irq.h
@@ -21,8 +21,8 @@
  *
  **************************************************************************/
 
-#ifndef _SYSIRQ_H_
-#define _SYSIRQ_H_
+#ifndef _PSB_IRQ_H_
+#define _PSB_IRQ_H_
 
 #include <drm/drmP.h>
 
@@ -44,4 +44,4 @@ u32  psb_get_vblank_counter(struct drm_device *dev, int pipe);
 
 int mdfld_enable_te(struct drm_device *dev, int pipe);
 void mdfld_disable_te(struct drm_device *dev, int pipe);
-#endif /* _SYSIRQ_H_ */
+#endif /* _PSB_IRQ_H_ */
diff --git a/drivers/gpu/drm/i915/i915_debugfs.c b/drivers/gpu/drm/i915/i915_debugfs.c
index 7299ea45dd03..e913d325d5b8 100644
--- a/drivers/gpu/drm/i915/i915_debugfs.c
+++ b/drivers/gpu/drm/i915/i915_debugfs.c
@@ -772,6 +772,23 @@ static int i915_error_state(struct seq_file *m, void *unused)
 				}
 			}
 		}
+
+		obj = error->ring[i].ctx;
+		if (obj) {
+			seq_printf(m, "%s --- HW Context = 0x%08x\n",
+				   dev_priv->ring[i].name,
+				   obj->gtt_offset);
+			offset = 0;
+			for (elt = 0; elt < PAGE_SIZE/16; elt += 4) {
+				seq_printf(m, "[%04x] %08x %08x %08x %08x\n",
+					   offset,
+					   obj->pages[0][elt],
+					   obj->pages[0][elt+1],
+					   obj->pages[0][elt+2],
+					   obj->pages[0][elt+3]);
+					offset += 16;
+			}
+		}
 	}
 
 	if (error->overlay)
@@ -849,76 +866,42 @@ static const struct file_operations i915_error_state_fops = {
 	.release = i915_error_state_release,
 };
 
-static ssize_t
-i915_next_seqno_read(struct file *filp,
-		 char __user *ubuf,
-		 size_t max,
-		 loff_t *ppos)
+static int
+i915_next_seqno_get(void *data, u64 *val)
 {
-	struct drm_device *dev = filp->private_data;
+	struct drm_device *dev = data;
 	drm_i915_private_t *dev_priv = dev->dev_private;
-	char buf[80];
-	int len;
 	int ret;
 
 	ret = mutex_lock_interruptible(&dev->struct_mutex);
 	if (ret)
 		return ret;
 
-	len = snprintf(buf, sizeof(buf),
-		       "next_seqno :  0x%x\n",
-		       dev_priv->next_seqno);
-
+	*val = dev_priv->next_seqno;
 	mutex_unlock(&dev->struct_mutex);
 
-	if (len > sizeof(buf))
-		len = sizeof(buf);
-
-	return simple_read_from_buffer(ubuf, max, ppos, buf, len);
+	return 0;
 }
 
-static ssize_t
-i915_next_seqno_write(struct file *filp,
-		      const char __user *ubuf,
-		      size_t cnt,
-		      loff_t *ppos)
-{
-	struct drm_device *dev = filp->private_data;
-	char buf[20];
-	u32 val = 1;
+static int
+i915_next_seqno_set(void *data, u64 val)
+{
+	struct drm_device *dev = data;
 	int ret;
 
-	if (cnt > 0) {
-		if (cnt > sizeof(buf) - 1)
-			return -EINVAL;
-
-		if (copy_from_user(buf, ubuf, cnt))
-			return -EFAULT;
-		buf[cnt] = 0;
-
-		ret = kstrtouint(buf, 0, &val);
-		if (ret < 0)
-			return ret;
-	}
-
 	ret = mutex_lock_interruptible(&dev->struct_mutex);
 	if (ret)
 		return ret;
 
 	ret = i915_gem_set_seqno(dev, val);
-
 	mutex_unlock(&dev->struct_mutex);
 
-	return ret ?: cnt;
+	return ret;
 }
 
-static const struct file_operations i915_next_seqno_fops = {
-	.owner = THIS_MODULE,
-	.open = simple_open,
-	.read = i915_next_seqno_read,
-	.write = i915_next_seqno_write,
-	.llseek = default_llseek,
-};
+DEFINE_SIMPLE_ATTRIBUTE(i915_next_seqno_fops,
+			i915_next_seqno_get, i915_next_seqno_set,
+			"0x%llx\n");
 
 static int i915_rstdby_delays(struct seq_file *m, void *unused)
 {
@@ -1023,6 +1006,9 @@ static int i915_cur_delayinfo(struct seq_file *m, void *unused)
 		max_freq = rp_state_cap & 0xff;
 		seq_printf(m, "Max non-overclocked (RP0) frequency: %dMHz\n",
 			   max_freq * GT_FREQUENCY_MULTIPLIER);
+
+		seq_printf(m, "Max overclocked frequency: %dMHz\n",
+			   dev_priv->rps.hw_max * GT_FREQUENCY_MULTIPLIER);
 	} else {
 		seq_printf(m, "no P-state info available\n");
 	}
@@ -1371,7 +1357,7 @@ static int i915_ring_freq_table(struct seq_file *m, void *unused)
 	if (ret)
 		return ret;
 
-	seq_printf(m, "GPU freq (MHz)\tEffective CPU freq (MHz)\n");
+	seq_printf(m, "GPU freq (MHz)\tEffective CPU freq (MHz)\tEffective Ring freq (MHz)\n");
 
 	for (gpu_freq = dev_priv->rps.min_delay;
 	     gpu_freq <= dev_priv->rps.max_delay;
@@ -1380,7 +1366,10 @@ static int i915_ring_freq_table(struct seq_file *m, void *unused)
 		sandybridge_pcode_read(dev_priv,
 				       GEN6_PCODE_READ_MIN_FREQ_TABLE,
 				       &ia_freq);
-		seq_printf(m, "%d\t\t%d\n", gpu_freq * GT_FREQUENCY_MULTIPLIER, ia_freq * 100);
+		seq_printf(m, "%d\t\t%d\t\t\t\t%d\n",
+			   gpu_freq * GT_FREQUENCY_MULTIPLIER,
+			   ((ia_freq >> 0) & 0xff) * 100,
+			   ((ia_freq >> 8) & 0xff) * 100);
 	}
 
 	mutex_unlock(&dev_priv->rps.hw_lock);
@@ -1680,105 +1669,51 @@ static int i915_dpio_info(struct seq_file *m, void *data)
 	return 0;
 }
 
-static ssize_t
-i915_wedged_read(struct file *filp,
-		 char __user *ubuf,
-		 size_t max,
-		 loff_t *ppos)
+static int
+i915_wedged_get(void *data, u64 *val)
 {
-	struct drm_device *dev = filp->private_data;
+	struct drm_device *dev = data;
 	drm_i915_private_t *dev_priv = dev->dev_private;
-	char buf[80];
-	int len;
 
-	len = snprintf(buf, sizeof(buf),
-		       "wedged :  %d\n",
-		       atomic_read(&dev_priv->gpu_error.reset_counter));
+	*val = atomic_read(&dev_priv->gpu_error.reset_counter);
 
-	if (len > sizeof(buf))
-		len = sizeof(buf);
-
-	return simple_read_from_buffer(ubuf, max, ppos, buf, len);
+	return 0;
 }
 
-static ssize_t
-i915_wedged_write(struct file *filp,
-		  const char __user *ubuf,
-		  size_t cnt,
-		  loff_t *ppos)
+static int
+i915_wedged_set(void *data, u64 val)
 {
-	struct drm_device *dev = filp->private_data;
-	char buf[20];
-	int val = 1;
-
-	if (cnt > 0) {
-		if (cnt > sizeof(buf) - 1)
-			return -EINVAL;
+	struct drm_device *dev = data;
 
-		if (copy_from_user(buf, ubuf, cnt))
-			return -EFAULT;
-		buf[cnt] = 0;
-
-		val = simple_strtoul(buf, NULL, 0);
-	}
-
-	DRM_INFO("Manually setting wedged to %d\n", val);
+	DRM_INFO("Manually setting wedged to %llu\n", val);
 	i915_handle_error(dev, val);
 
-	return cnt;
+	return 0;
 }
 
-static const struct file_operations i915_wedged_fops = {
-	.owner = THIS_MODULE,
-	.open = simple_open,
-	.read = i915_wedged_read,
-	.write = i915_wedged_write,
-	.llseek = default_llseek,
-};
+DEFINE_SIMPLE_ATTRIBUTE(i915_wedged_fops,
+			i915_wedged_get, i915_wedged_set,
+			"%llu\n");
 
-static ssize_t
-i915_ring_stop_read(struct file *filp,
-		    char __user *ubuf,
-		    size_t max,
-		    loff_t *ppos)
+static int
+i915_ring_stop_get(void *data, u64 *val)
 {
-	struct drm_device *dev = filp->private_data;
+	struct drm_device *dev = data;
 	drm_i915_private_t *dev_priv = dev->dev_private;
-	char buf[20];
-	int len;
 
-	len = snprintf(buf, sizeof(buf),
-		       "0x%08x\n", dev_priv->gpu_error.stop_rings);
+	*val = dev_priv->gpu_error.stop_rings;
 
-	if (len > sizeof(buf))
-		len = sizeof(buf);
-
-	return simple_read_from_buffer(ubuf, max, ppos, buf, len);
+	return 0;
 }
 
-static ssize_t
-i915_ring_stop_write(struct file *filp,
-		     const char __user *ubuf,
-		     size_t cnt,
-		     loff_t *ppos)
+static int
+i915_ring_stop_set(void *data, u64 val)
 {
-	struct drm_device *dev = filp->private_data;
+	struct drm_device *dev = data;
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	char buf[20];
-	int val = 0, ret;
-
-	if (cnt > 0) {
-		if (cnt > sizeof(buf) - 1)
-			return -EINVAL;
-
-		if (copy_from_user(buf, ubuf, cnt))
-			return -EFAULT;
-		buf[cnt] = 0;
-
-		val = simple_strtoul(buf, NULL, 0);
-	}
+	int ret;
 
-	DRM_DEBUG_DRIVER("Stopping rings 0x%08x\n", val);
+	DRM_DEBUG_DRIVER("Stopping rings 0x%08llx\n", val);
 
 	ret = mutex_lock_interruptible(&dev->struct_mutex);
 	if (ret)
@@ -1787,16 +1722,12 @@ i915_ring_stop_write(struct file *filp,
 	dev_priv->gpu_error.stop_rings = val;
 	mutex_unlock(&dev->struct_mutex);
 
-	return cnt;
+	return 0;
 }
 
-static const struct file_operations i915_ring_stop_fops = {
-	.owner = THIS_MODULE,
-	.open = simple_open,
-	.read = i915_ring_stop_read,
-	.write = i915_ring_stop_write,
-	.llseek = default_llseek,
-};
+DEFINE_SIMPLE_ATTRIBUTE(i915_ring_stop_fops,
+			i915_ring_stop_get, i915_ring_stop_set,
+			"0x%08llx\n");
 
 #define DROP_UNBOUND 0x1
 #define DROP_BOUND 0x2
@@ -1806,46 +1737,23 @@ static const struct file_operations i915_ring_stop_fops = {
 		  DROP_BOUND | \
 		  DROP_RETIRE | \
 		  DROP_ACTIVE)
-static ssize_t
-i915_drop_caches_read(struct file *filp,
-		      char __user *ubuf,
-		      size_t max,
-		      loff_t *ppos)
+static int
+i915_drop_caches_get(void *data, u64 *val)
 {
-	char buf[20];
-	int len;
+	*val = DROP_ALL;
 
-	len = snprintf(buf, sizeof(buf), "0x%08x\n", DROP_ALL);
-	if (len > sizeof(buf))
-		len = sizeof(buf);
-
-	return simple_read_from_buffer(ubuf, max, ppos, buf, len);
+	return 0;
 }
 
-static ssize_t
-i915_drop_caches_write(struct file *filp,
-		       const char __user *ubuf,
-		       size_t cnt,
-		       loff_t *ppos)
+static int
+i915_drop_caches_set(void *data, u64 val)
 {
-	struct drm_device *dev = filp->private_data;
+	struct drm_device *dev = data;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct drm_i915_gem_object *obj, *next;
-	char buf[20];
-	int val = 0, ret;
-
-	if (cnt > 0) {
-		if (cnt > sizeof(buf) - 1)
-			return -EINVAL;
-
-		if (copy_from_user(buf, ubuf, cnt))
-			return -EFAULT;
-		buf[cnt] = 0;
-
-		val = simple_strtoul(buf, NULL, 0);
-	}
+	int ret;
 
-	DRM_DEBUG_DRIVER("Dropping caches: 0x%08x\n", val);
+	DRM_DEBUG_DRIVER("Dropping caches: 0x%08llx\n", val);
 
 	/* No need to check and wait for gpu resets, only libdrm auto-restarts
 	 * on ioctls on -EAGAIN. */
@@ -1883,27 +1791,19 @@ i915_drop_caches_write(struct file *filp,
 unlock:
 	mutex_unlock(&dev->struct_mutex);
 
-	return ret ?: cnt;
+	return ret;
 }
 
-static const struct file_operations i915_drop_caches_fops = {
-	.owner = THIS_MODULE,
-	.open = simple_open,
-	.read = i915_drop_caches_read,
-	.write = i915_drop_caches_write,
-	.llseek = default_llseek,
-};
+DEFINE_SIMPLE_ATTRIBUTE(i915_drop_caches_fops,
+			i915_drop_caches_get, i915_drop_caches_set,
+			"0x%08llx\n");
 
-static ssize_t
-i915_max_freq_read(struct file *filp,
-		   char __user *ubuf,
-		   size_t max,
-		   loff_t *ppos)
+static int
+i915_max_freq_get(void *data, u64 *val)
 {
-	struct drm_device *dev = filp->private_data;
+	struct drm_device *dev = data;
 	drm_i915_private_t *dev_priv = dev->dev_private;
-	char buf[80];
-	int len, ret;
+	int ret;
 
 	if (!(IS_GEN6(dev) || IS_GEN7(dev)))
 		return -ENODEV;
@@ -1912,42 +1812,23 @@ i915_max_freq_read(struct file *filp,
 	if (ret)
 		return ret;
 
-	len = snprintf(buf, sizeof(buf),
-		       "max freq: %d\n", dev_priv->rps.max_delay * GT_FREQUENCY_MULTIPLIER);
+	*val = dev_priv->rps.max_delay * GT_FREQUENCY_MULTIPLIER;
 	mutex_unlock(&dev_priv->rps.hw_lock);
 
-	if (len > sizeof(buf))
-		len = sizeof(buf);
-
-	return simple_read_from_buffer(ubuf, max, ppos, buf, len);
+	return 0;
 }
 
-static ssize_t
-i915_max_freq_write(struct file *filp,
-		  const char __user *ubuf,
-		  size_t cnt,
-		  loff_t *ppos)
+static int
+i915_max_freq_set(void *data, u64 val)
 {
-	struct drm_device *dev = filp->private_data;
+	struct drm_device *dev = data;
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	char buf[20];
-	int val = 1, ret;
+	int ret;
 
 	if (!(IS_GEN6(dev) || IS_GEN7(dev)))
 		return -ENODEV;
 
-	if (cnt > 0) {
-		if (cnt > sizeof(buf) - 1)
-			return -EINVAL;
-
-		if (copy_from_user(buf, ubuf, cnt))
-			return -EFAULT;
-		buf[cnt] = 0;
-
-		val = simple_strtoul(buf, NULL, 0);
-	}
-
-	DRM_DEBUG_DRIVER("Manually setting max freq to %d\n", val);
+	DRM_DEBUG_DRIVER("Manually setting max freq to %llu\n", val);
 
 	ret = mutex_lock_interruptible(&dev_priv->rps.hw_lock);
 	if (ret)
@@ -1956,30 +1837,24 @@ i915_max_freq_write(struct file *filp,
 	/*
 	 * Turbo will still be enabled, but won't go above the set value.
 	 */
-	dev_priv->rps.max_delay = val / GT_FREQUENCY_MULTIPLIER;
-
-	gen6_set_rps(dev, val / GT_FREQUENCY_MULTIPLIER);
+	do_div(val, GT_FREQUENCY_MULTIPLIER);
+	dev_priv->rps.max_delay = val;
+	gen6_set_rps(dev, val);
 	mutex_unlock(&dev_priv->rps.hw_lock);
 
-	return cnt;
+	return 0;
 }
 
-static const struct file_operations i915_max_freq_fops = {
-	.owner = THIS_MODULE,
-	.open = simple_open,
-	.read = i915_max_freq_read,
-	.write = i915_max_freq_write,
-	.llseek = default_llseek,
-};
+DEFINE_SIMPLE_ATTRIBUTE(i915_max_freq_fops,
+			i915_max_freq_get, i915_max_freq_set,
+			"%llu\n");
 
-static ssize_t
-i915_min_freq_read(struct file *filp, char __user *ubuf, size_t max,
-		   loff_t *ppos)
+static int
+i915_min_freq_get(void *data, u64 *val)
 {
-	struct drm_device *dev = filp->private_data;
+	struct drm_device *dev = data;
 	drm_i915_private_t *dev_priv = dev->dev_private;
-	char buf[80];
-	int len, ret;
+	int ret;
 
 	if (!(IS_GEN6(dev) || IS_GEN7(dev)))
 		return -ENODEV;
@@ -1988,40 +1863,23 @@ i915_min_freq_read(struct file *filp, char __user *ubuf, size_t max,
 	if (ret)
 		return ret;
 
-	len = snprintf(buf, sizeof(buf),
-		       "min freq: %d\n", dev_priv->rps.min_delay * GT_FREQUENCY_MULTIPLIER);
+	*val = dev_priv->rps.min_delay * GT_FREQUENCY_MULTIPLIER;
 	mutex_unlock(&dev_priv->rps.hw_lock);
 
-	if (len > sizeof(buf))
-		len = sizeof(buf);
-
-	return simple_read_from_buffer(ubuf, max, ppos, buf, len);
+	return 0;
 }
 
-static ssize_t
-i915_min_freq_write(struct file *filp, const char __user *ubuf, size_t cnt,
-		    loff_t *ppos)
+static int
+i915_min_freq_set(void *data, u64 val)
 {
-	struct drm_device *dev = filp->private_data;
+	struct drm_device *dev = data;
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	char buf[20];
-	int val = 1, ret;
+	int ret;
 
 	if (!(IS_GEN6(dev) || IS_GEN7(dev)))
 		return -ENODEV;
 
-	if (cnt > 0) {
-		if (cnt > sizeof(buf) - 1)
-			return -EINVAL;
-
-		if (copy_from_user(buf, ubuf, cnt))
-			return -EFAULT;
-		buf[cnt] = 0;
-
-		val = simple_strtoul(buf, NULL, 0);
-	}
-
-	DRM_DEBUG_DRIVER("Manually setting min freq to %d\n", val);
+	DRM_DEBUG_DRIVER("Manually setting min freq to %llu\n", val);
 
 	ret = mutex_lock_interruptible(&dev_priv->rps.hw_lock);
 	if (ret)
@@ -2030,33 +1888,25 @@ i915_min_freq_write(struct file *filp, const char __user *ubuf, size_t cnt,
 	/*
 	 * Turbo will still be enabled, but won't go below the set value.
 	 */
-	dev_priv->rps.min_delay = val / GT_FREQUENCY_MULTIPLIER;
-
-	gen6_set_rps(dev, val / GT_FREQUENCY_MULTIPLIER);
+	do_div(val, GT_FREQUENCY_MULTIPLIER);
+	dev_priv->rps.min_delay = val;
+	gen6_set_rps(dev, val);
 	mutex_unlock(&dev_priv->rps.hw_lock);
 
-	return cnt;
+	return 0;
 }
 
-static const struct file_operations i915_min_freq_fops = {
-	.owner = THIS_MODULE,
-	.open = simple_open,
-	.read = i915_min_freq_read,
-	.write = i915_min_freq_write,
-	.llseek = default_llseek,
-};
+DEFINE_SIMPLE_ATTRIBUTE(i915_min_freq_fops,
+			i915_min_freq_get, i915_min_freq_set,
+			"%llu\n");
 
-static ssize_t
-i915_cache_sharing_read(struct file *filp,
-		   char __user *ubuf,
-		   size_t max,
-		   loff_t *ppos)
+static int
+i915_cache_sharing_get(void *data, u64 *val)
 {
-	struct drm_device *dev = filp->private_data;
+	struct drm_device *dev = data;
 	drm_i915_private_t *dev_priv = dev->dev_private;
-	char buf[80];
 	u32 snpcr;
-	int len, ret;
+	int ret;
 
 	if (!(IS_GEN6(dev) || IS_GEN7(dev)))
 		return -ENODEV;
@@ -2068,46 +1918,25 @@ i915_cache_sharing_read(struct file *filp,
 	snpcr = I915_READ(GEN6_MBCUNIT_SNPCR);
 	mutex_unlock(&dev_priv->dev->struct_mutex);
 
-	len = snprintf(buf, sizeof(buf),
-		       "%d\n", (snpcr & GEN6_MBC_SNPCR_MASK) >>
-		       GEN6_MBC_SNPCR_SHIFT);
+	*val = (snpcr & GEN6_MBC_SNPCR_MASK) >> GEN6_MBC_SNPCR_SHIFT;
 
-	if (len > sizeof(buf))
-		len = sizeof(buf);
-
-	return simple_read_from_buffer(ubuf, max, ppos, buf, len);
+	return 0;
 }
 
-static ssize_t
-i915_cache_sharing_write(struct file *filp,
-		  const char __user *ubuf,
-		  size_t cnt,
-		  loff_t *ppos)
+static int
+i915_cache_sharing_set(void *data, u64 val)
 {
-	struct drm_device *dev = filp->private_data;
+	struct drm_device *dev = data;
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	char buf[20];
 	u32 snpcr;
-	int val = 1;
 
 	if (!(IS_GEN6(dev) || IS_GEN7(dev)))
 		return -ENODEV;
 
-	if (cnt > 0) {
-		if (cnt > sizeof(buf) - 1)
-			return -EINVAL;
-
-		if (copy_from_user(buf, ubuf, cnt))
-			return -EFAULT;
-		buf[cnt] = 0;
-
-		val = simple_strtoul(buf, NULL, 0);
-	}
-
-	if (val < 0 || val > 3)
+	if (val > 3)
 		return -EINVAL;
 
-	DRM_DEBUG_DRIVER("Manually setting uncore sharing to %d\n", val);
+	DRM_DEBUG_DRIVER("Manually setting uncore sharing to %llu\n", val);
 
 	/* Update the cache sharing policy here as well */
 	snpcr = I915_READ(GEN6_MBCUNIT_SNPCR);
@@ -2115,16 +1944,12 @@ i915_cache_sharing_write(struct file *filp,
 	snpcr |= (val << GEN6_MBC_SNPCR_SHIFT);
 	I915_WRITE(GEN6_MBCUNIT_SNPCR, snpcr);
 
-	return cnt;
+	return 0;
 }
 
-static const struct file_operations i915_cache_sharing_fops = {
-	.owner = THIS_MODULE,
-	.open = simple_open,
-	.read = i915_cache_sharing_read,
-	.write = i915_cache_sharing_write,
-	.llseek = default_llseek,
-};
+DEFINE_SIMPLE_ATTRIBUTE(i915_cache_sharing_fops,
+			i915_cache_sharing_get, i915_cache_sharing_set,
+			"%llu\n");
 
 /* As the drm_debugfs_init() routines are called before dev->dev_private is
  * allocated we need to hook into the minor for release. */
diff --git a/drivers/gpu/drm/i915/i915_dma.c b/drivers/gpu/drm/i915/i915_dma.c
index 4fa6beb14c77..3b315ba85a3e 100644
--- a/drivers/gpu/drm/i915/i915_dma.c
+++ b/drivers/gpu/drm/i915/i915_dma.c
@@ -1322,6 +1322,10 @@ static int i915_load_modeset_init(struct drm_device *dev)
 	/* Always safe in the mode setting case. */
 	/* FIXME: do pre/post-mode set stuff in core KMS code */
 	dev->vblank_disable_allowed = 1;
+	if (INTEL_INFO(dev)->num_pipes == 0) {
+		dev_priv->mm.suspended = 0;
+		return 0;
+	}
 
 	ret = intel_fbdev_init(dev);
 	if (ret)
@@ -1453,6 +1457,22 @@ static void i915_dump_device_info(struct drm_i915_private *dev_priv)
 }
 
 /**
+ * intel_early_sanitize_regs - clean up BIOS state
+ * @dev: DRM device
+ *
+ * This function must be called before we do any I915_READ or I915_WRITE. Its
+ * purpose is to clean up any state left by the BIOS that may affect us when
+ * reading and/or writing registers.
+ */
+static void intel_early_sanitize_regs(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+
+	if (IS_HASWELL(dev))
+		I915_WRITE_NOTRACE(FPGA_DBG, FPGA_DBG_RM_NOCLAIM);
+}
+
+/**
  * i915_driver_load - setup chip and create an initial config
  * @dev: DRM device
  * @flags: startup flags
@@ -1498,6 +1518,28 @@ int i915_driver_load(struct drm_device *dev, unsigned long flags)
 		goto free_priv;
 	}
 
+	mmio_bar = IS_GEN2(dev) ? 1 : 0;
+	/* Before gen4, the registers and the GTT are behind different BARs.
+	 * However, from gen4 onwards, the registers and the GTT are shared
+	 * in the same BAR, so we want to restrict this ioremap from
+	 * clobbering the GTT which we want ioremap_wc instead. Fortunately,
+	 * the register BAR remains the same size for all the earlier
+	 * generations up to Ironlake.
+	 */
+	if (info->gen < 5)
+		mmio_size = 512*1024;
+	else
+		mmio_size = 2*1024*1024;
+
+	dev_priv->regs = pci_iomap(dev->pdev, mmio_bar, mmio_size);
+	if (!dev_priv->regs) {
+		DRM_ERROR("failed to map registers\n");
+		ret = -EIO;
+		goto put_bridge;
+	}
+
+	intel_early_sanitize_regs(dev);
+
 	ret = i915_gem_gtt_init(dev);
 	if (ret)
 		goto put_bridge;
@@ -1522,26 +1564,6 @@ int i915_driver_load(struct drm_device *dev, unsigned long flags)
 	if (IS_BROADWATER(dev) || IS_CRESTLINE(dev))
 		dma_set_coherent_mask(&dev->pdev->dev, DMA_BIT_MASK(32));
 
-	mmio_bar = IS_GEN2(dev) ? 1 : 0;
-	/* Before gen4, the registers and the GTT are behind different BARs.
-	 * However, from gen4 onwards, the registers and the GTT are shared
-	 * in the same BAR, so we want to restrict this ioremap from
-	 * clobbering the GTT which we want ioremap_wc instead. Fortunately,
-	 * the register BAR remains the same size for all the earlier
-	 * generations up to Ironlake.
-	 */
-	if (info->gen < 5)
-		mmio_size = 512*1024;
-	else
-		mmio_size = 2*1024*1024;
-
-	dev_priv->regs = pci_iomap(dev->pdev, mmio_bar, mmio_size);
-	if (!dev_priv->regs) {
-		DRM_ERROR("failed to map registers\n");
-		ret = -EIO;
-		goto put_gmch;
-	}
-
 	aperture_size = dev_priv->gtt.mappable_end;
 
 	dev_priv->gtt.mappable =
@@ -1612,16 +1634,15 @@ int i915_driver_load(struct drm_device *dev, unsigned long flags)
 	mutex_init(&dev_priv->rps.hw_lock);
 	mutex_init(&dev_priv->modeset_restore_lock);
 
-	if (IS_IVYBRIDGE(dev) || IS_HASWELL(dev))
-		dev_priv->num_pipe = 3;
-	else if (IS_MOBILE(dev) || !IS_GEN2(dev))
-		dev_priv->num_pipe = 2;
-	else
-		dev_priv->num_pipe = 1;
+	dev_priv->num_plane = 1;
+	if (IS_VALLEYVIEW(dev))
+		dev_priv->num_plane = 2;
 
-	ret = drm_vblank_init(dev, dev_priv->num_pipe);
-	if (ret)
-		goto out_gem_unload;
+	if (INTEL_INFO(dev)->num_pipes) {
+		ret = drm_vblank_init(dev, INTEL_INFO(dev)->num_pipes);
+		if (ret)
+			goto out_gem_unload;
+	}
 
 	/* Start out suspended */
 	dev_priv->mm.suspended = 1;
@@ -1636,9 +1657,11 @@ int i915_driver_load(struct drm_device *dev, unsigned long flags)
 
 	i915_setup_sysfs(dev);
 
-	/* Must be done after probing outputs */
-	intel_opregion_init(dev);
-	acpi_video_register();
+	if (INTEL_INFO(dev)->num_pipes) {
+		/* Must be done after probing outputs */
+		intel_opregion_init(dev);
+		acpi_video_register();
+	}
 
 	if (IS_GEN5(dev))
 		intel_gpu_ips_init(dev_priv);
@@ -1663,10 +1686,9 @@ out_mtrrfree:
 		dev_priv->mm.gtt_mtrr = -1;
 	}
 	io_mapping_free(dev_priv->gtt.mappable);
+	dev_priv->gtt.gtt_remove(dev);
 out_rmmap:
 	pci_iounmap(dev->pdev, dev_priv->regs);
-put_gmch:
-	dev_priv->gtt.gtt_remove(dev);
 put_bridge:
 	pci_dev_put(dev_priv->bridge_dev);
 free_priv:
diff --git a/drivers/gpu/drm/i915/i915_drv.c b/drivers/gpu/drm/i915/i915_drv.c
index e9b57893db2b..9ebe895c17d6 100644
--- a/drivers/gpu/drm/i915/i915_drv.c
+++ b/drivers/gpu/drm/i915/i915_drv.c
@@ -121,9 +121,7 @@ MODULE_PARM_DESC(i915_enable_ppgtt,
 unsigned int i915_preliminary_hw_support __read_mostly = 0;
 module_param_named(preliminary_hw_support, i915_preliminary_hw_support, int, 0600);
 MODULE_PARM_DESC(preliminary_hw_support,
-		"Enable preliminary hardware support. "
-		"Enable Haswell and ValleyView Support. "
-		"(default: false)");
+		"Enable preliminary hardware support. (default: false)");
 
 int i915_disable_power_well __read_mostly = 0;
 module_param_named(disable_power_well, i915_disable_power_well, int, 0600);
@@ -142,75 +140,85 @@ extern int intel_agp_enabled;
 	.subdevice = PCI_ANY_ID,		\
 	.driver_data = (unsigned long) info }
 
+#define INTEL_QUANTA_VGA_DEVICE(info) {		\
+	.class = PCI_BASE_CLASS_DISPLAY << 16,	\
+	.class_mask = 0xff0000,			\
+	.vendor = 0x8086,			\
+	.device = 0x16a,			\
+	.subvendor = 0x152d,			\
+	.subdevice = 0x8990,			\
+	.driver_data = (unsigned long) info }
+
+
 static const struct intel_device_info intel_i830_info = {
-	.gen = 2, .is_mobile = 1, .cursor_needs_physical = 1,
+	.gen = 2, .is_mobile = 1, .cursor_needs_physical = 1, .num_pipes = 2,
 	.has_overlay = 1, .overlay_needs_physical = 1,
 };
 
 static const struct intel_device_info intel_845g_info = {
-	.gen = 2,
+	.gen = 2, .num_pipes = 1,
 	.has_overlay = 1, .overlay_needs_physical = 1,
 };
 
 static const struct intel_device_info intel_i85x_info = {
-	.gen = 2, .is_i85x = 1, .is_mobile = 1,
+	.gen = 2, .is_i85x = 1, .is_mobile = 1, .num_pipes = 2,
 	.cursor_needs_physical = 1,
 	.has_overlay = 1, .overlay_needs_physical = 1,
 };
 
 static const struct intel_device_info intel_i865g_info = {
-	.gen = 2,
+	.gen = 2, .num_pipes = 1,
 	.has_overlay = 1, .overlay_needs_physical = 1,
 };
 
 static const struct intel_device_info intel_i915g_info = {
-	.gen = 3, .is_i915g = 1, .cursor_needs_physical = 1,
+	.gen = 3, .is_i915g = 1, .cursor_needs_physical = 1, .num_pipes = 2,
 	.has_overlay = 1, .overlay_needs_physical = 1,
 };
 static const struct intel_device_info intel_i915gm_info = {
-	.gen = 3, .is_mobile = 1,
+	.gen = 3, .is_mobile = 1, .num_pipes = 2,
 	.cursor_needs_physical = 1,
 	.has_overlay = 1, .overlay_needs_physical = 1,
 	.supports_tv = 1,
 };
 static const struct intel_device_info intel_i945g_info = {
-	.gen = 3, .has_hotplug = 1, .cursor_needs_physical = 1,
+	.gen = 3, .has_hotplug = 1, .cursor_needs_physical = 1, .num_pipes = 2,
 	.has_overlay = 1, .overlay_needs_physical = 1,
 };
 static const struct intel_device_info intel_i945gm_info = {
-	.gen = 3, .is_i945gm = 1, .is_mobile = 1,
+	.gen = 3, .is_i945gm = 1, .is_mobile = 1, .num_pipes = 2,
 	.has_hotplug = 1, .cursor_needs_physical = 1,
 	.has_overlay = 1, .overlay_needs_physical = 1,
 	.supports_tv = 1,
 };
 
 static const struct intel_device_info intel_i965g_info = {
-	.gen = 4, .is_broadwater = 1,
+	.gen = 4, .is_broadwater = 1, .num_pipes = 2,
 	.has_hotplug = 1,
 	.has_overlay = 1,
 };
 
 static const struct intel_device_info intel_i965gm_info = {
-	.gen = 4, .is_crestline = 1,
+	.gen = 4, .is_crestline = 1, .num_pipes = 2,
 	.is_mobile = 1, .has_fbc = 1, .has_hotplug = 1,
 	.has_overlay = 1,
 	.supports_tv = 1,
 };
 
 static const struct intel_device_info intel_g33_info = {
-	.gen = 3, .is_g33 = 1,
+	.gen = 3, .is_g33 = 1, .num_pipes = 2,
 	.need_gfx_hws = 1, .has_hotplug = 1,
 	.has_overlay = 1,
 };
 
 static const struct intel_device_info intel_g45_info = {
-	.gen = 4, .is_g4x = 1, .need_gfx_hws = 1,
+	.gen = 4, .is_g4x = 1, .need_gfx_hws = 1, .num_pipes = 2,
 	.has_pipe_cxsr = 1, .has_hotplug = 1,
 	.has_bsd_ring = 1,
 };
 
 static const struct intel_device_info intel_gm45_info = {
-	.gen = 4, .is_g4x = 1,
+	.gen = 4, .is_g4x = 1, .num_pipes = 2,
 	.is_mobile = 1, .need_gfx_hws = 1, .has_fbc = 1,
 	.has_pipe_cxsr = 1, .has_hotplug = 1,
 	.supports_tv = 1,
@@ -218,26 +226,26 @@ static const struct intel_device_info intel_gm45_info = {
 };
 
 static const struct intel_device_info intel_pineview_info = {
-	.gen = 3, .is_g33 = 1, .is_pineview = 1, .is_mobile = 1,
+	.gen = 3, .is_g33 = 1, .is_pineview = 1, .is_mobile = 1, .num_pipes = 2,
 	.need_gfx_hws = 1, .has_hotplug = 1,
 	.has_overlay = 1,
 };
 
 static const struct intel_device_info intel_ironlake_d_info = {
-	.gen = 5,
+	.gen = 5, .num_pipes = 2,
 	.need_gfx_hws = 1, .has_hotplug = 1,
 	.has_bsd_ring = 1,
 };
 
 static const struct intel_device_info intel_ironlake_m_info = {
-	.gen = 5, .is_mobile = 1,
+	.gen = 5, .is_mobile = 1, .num_pipes = 2,
 	.need_gfx_hws = 1, .has_hotplug = 1,
 	.has_fbc = 1,
 	.has_bsd_ring = 1,
 };
 
 static const struct intel_device_info intel_sandybridge_d_info = {
-	.gen = 6,
+	.gen = 6, .num_pipes = 2,
 	.need_gfx_hws = 1, .has_hotplug = 1,
 	.has_bsd_ring = 1,
 	.has_blt_ring = 1,
@@ -246,7 +254,7 @@ static const struct intel_device_info intel_sandybridge_d_info = {
 };
 
 static const struct intel_device_info intel_sandybridge_m_info = {
-	.gen = 6, .is_mobile = 1,
+	.gen = 6, .is_mobile = 1, .num_pipes = 2,
 	.need_gfx_hws = 1, .has_hotplug = 1,
 	.has_fbc = 1,
 	.has_bsd_ring = 1,
@@ -255,61 +263,57 @@ static const struct intel_device_info intel_sandybridge_m_info = {
 	.has_force_wake = 1,
 };
 
+#define GEN7_FEATURES  \
+	.gen = 7, .num_pipes = 3, \
+	.need_gfx_hws = 1, .has_hotplug = 1, \
+	.has_bsd_ring = 1, \
+	.has_blt_ring = 1, \
+	.has_llc = 1, \
+	.has_force_wake = 1
+
 static const struct intel_device_info intel_ivybridge_d_info = {
-	.is_ivybridge = 1, .gen = 7,
-	.need_gfx_hws = 1, .has_hotplug = 1,
-	.has_bsd_ring = 1,
-	.has_blt_ring = 1,
-	.has_llc = 1,
-	.has_force_wake = 1,
+	GEN7_FEATURES,
+	.is_ivybridge = 1,
 };
 
 static const struct intel_device_info intel_ivybridge_m_info = {
-	.is_ivybridge = 1, .gen = 7, .is_mobile = 1,
-	.need_gfx_hws = 1, .has_hotplug = 1,
-	.has_fbc = 0,	/* FBC is not enabled on Ivybridge mobile yet */
-	.has_bsd_ring = 1,
-	.has_blt_ring = 1,
-	.has_llc = 1,
-	.has_force_wake = 1,
+	GEN7_FEATURES,
+	.is_ivybridge = 1,
+	.is_mobile = 1,
+};
+
+static const struct intel_device_info intel_ivybridge_q_info = {
+	GEN7_FEATURES,
+	.is_ivybridge = 1,
+	.num_pipes = 0, /* legal, last one wins */
 };
 
 static const struct intel_device_info intel_valleyview_m_info = {
-	.gen = 7, .is_mobile = 1,
-	.need_gfx_hws = 1, .has_hotplug = 1,
-	.has_fbc = 0,
-	.has_bsd_ring = 1,
-	.has_blt_ring = 1,
+	GEN7_FEATURES,
+	.is_mobile = 1,
+	.num_pipes = 2,
 	.is_valleyview = 1,
 	.display_mmio_offset = VLV_DISPLAY_BASE,
+	.has_llc = 0, /* legal, last one wins */
 };
 
 static const struct intel_device_info intel_valleyview_d_info = {
-	.gen = 7,
-	.need_gfx_hws = 1, .has_hotplug = 1,
-	.has_fbc = 0,
-	.has_bsd_ring = 1,
-	.has_blt_ring = 1,
+	GEN7_FEATURES,
+	.num_pipes = 2,
 	.is_valleyview = 1,
 	.display_mmio_offset = VLV_DISPLAY_BASE,
+	.has_llc = 0, /* legal, last one wins */
 };
 
 static const struct intel_device_info intel_haswell_d_info = {
-	.is_haswell = 1, .gen = 7,
-	.need_gfx_hws = 1, .has_hotplug = 1,
-	.has_bsd_ring = 1,
-	.has_blt_ring = 1,
-	.has_llc = 1,
-	.has_force_wake = 1,
+	GEN7_FEATURES,
+	.is_haswell = 1,
 };
 
 static const struct intel_device_info intel_haswell_m_info = {
-	.is_haswell = 1, .gen = 7, .is_mobile = 1,
-	.need_gfx_hws = 1, .has_hotplug = 1,
-	.has_bsd_ring = 1,
-	.has_blt_ring = 1,
-	.has_llc = 1,
-	.has_force_wake = 1,
+	GEN7_FEATURES,
+	.is_haswell = 1,
+	.is_mobile = 1,
 };
 
 static const struct pci_device_id pciidlist[] = {		/* aka */
@@ -356,6 +360,7 @@ static const struct pci_device_id pciidlist[] = {		/* aka */
 	INTEL_VGA_DEVICE(0x0152, &intel_ivybridge_d_info), /* GT1 desktop */
 	INTEL_VGA_DEVICE(0x0162, &intel_ivybridge_d_info), /* GT2 desktop */
 	INTEL_VGA_DEVICE(0x015a, &intel_ivybridge_d_info), /* GT1 server */
+	INTEL_QUANTA_VGA_DEVICE(&intel_ivybridge_q_info), /* Quanta transcode */
 	INTEL_VGA_DEVICE(0x016a, &intel_ivybridge_d_info), /* GT2 server */
 	INTEL_VGA_DEVICE(0x0402, &intel_haswell_d_info), /* GT1 desktop */
 	INTEL_VGA_DEVICE(0x0412, &intel_haswell_d_info), /* GT2 desktop */
@@ -394,6 +399,9 @@ static const struct pci_device_id pciidlist[] = {		/* aka */
 	INTEL_VGA_DEVICE(0x0D16, &intel_haswell_m_info), /* CRW GT2 mobile */
 	INTEL_VGA_DEVICE(0x0D26, &intel_haswell_m_info), /* CRW GT2 mobile */
 	INTEL_VGA_DEVICE(0x0f30, &intel_valleyview_m_info),
+	INTEL_VGA_DEVICE(0x0f31, &intel_valleyview_m_info),
+	INTEL_VGA_DEVICE(0x0f32, &intel_valleyview_m_info),
+	INTEL_VGA_DEVICE(0x0f33, &intel_valleyview_m_info),
 	INTEL_VGA_DEVICE(0x0157, &intel_valleyview_m_info),
 	INTEL_VGA_DEVICE(0x0155, &intel_valleyview_d_info),
 	{0, 0, 0}
@@ -408,6 +416,15 @@ void intel_detect_pch(struct drm_device *dev)
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct pci_dev *pch;
 
+	/* In all current cases, num_pipes is equivalent to the PCH_NOP setting
+	 * (which really amounts to a PCH but no South Display).
+	 */
+	if (INTEL_INFO(dev)->num_pipes == 0) {
+		dev_priv->pch_type = PCH_NOP;
+		dev_priv->num_pch_pll = 0;
+		return;
+	}
+
 	/*
 	 * The reason to probe ISA bridge instead of Dev31:Fun0 is to
 	 * make graphics device passthrough work easy for VMM, that only
@@ -442,11 +459,13 @@ void intel_detect_pch(struct drm_device *dev)
 				dev_priv->num_pch_pll = 0;
 				DRM_DEBUG_KMS("Found LynxPoint PCH\n");
 				WARN_ON(!IS_HASWELL(dev));
+				WARN_ON(IS_ULT(dev));
 			} else if (id == INTEL_PCH_LPT_LP_DEVICE_ID_TYPE) {
 				dev_priv->pch_type = PCH_LPT;
 				dev_priv->num_pch_pll = 0;
 				DRM_DEBUG_KMS("Found LynxPoint LP PCH\n");
 				WARN_ON(!IS_HASWELL(dev));
+				WARN_ON(!IS_ULT(dev));
 			}
 			BUG_ON(dev_priv->num_pch_pll > I915_NUM_PLLS);
 		}
@@ -474,6 +493,7 @@ bool i915_semaphore_is_enabled(struct drm_device *dev)
 static int i915_drm_freeze(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_crtc *crtc;
 
 	/* ignore lid events during suspend */
 	mutex_lock(&dev_priv->modeset_restore_lock);
@@ -497,10 +517,14 @@ static int i915_drm_freeze(struct drm_device *dev)
 
 		cancel_delayed_work_sync(&dev_priv->rps.delayed_resume_work);
 
-		intel_modeset_disable(dev);
-
 		drm_irq_uninstall(dev);
 		dev_priv->enable_hotplug_processing = false;
+		/*
+		 * Disable CRTCs directly since we want to preserve sw state
+		 * for _thaw.
+		 */
+		list_for_each_entry(crtc, &dev->mode_config.crtc_list, head)
+			dev_priv->display.crtc_disable(crtc);
 	}
 
 	i915_save_state(dev);
@@ -556,6 +580,24 @@ void intel_console_resume(struct work_struct *work)
 	console_unlock();
 }
 
+static void intel_resume_hotplug(struct drm_device *dev)
+{
+	struct drm_mode_config *mode_config = &dev->mode_config;
+	struct intel_encoder *encoder;
+
+	mutex_lock(&mode_config->mutex);
+	DRM_DEBUG_KMS("running encoder hotplug functions\n");
+
+	list_for_each_entry(encoder, &mode_config->encoder_list, base.head)
+		if (encoder->hot_plug)
+			encoder->hot_plug(encoder);
+
+	mutex_unlock(&mode_config->mutex);
+
+	/* Just fire off a uevent and let userspace tell us what to do */
+	drm_helper_hpd_irq_event(dev);
+}
+
 static int __i915_drm_thaw(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
@@ -578,7 +620,10 @@ static int __i915_drm_thaw(struct drm_device *dev)
 		drm_irq_install(dev);
 
 		intel_modeset_init_hw(dev);
-		intel_modeset_setup_hw_state(dev, false);
+
+		drm_modeset_lock_all(dev);
+		intel_modeset_setup_hw_state(dev, true);
+		drm_modeset_unlock_all(dev);
 
 		/*
 		 * ... but also need to make sure that hotplug processing
@@ -588,6 +633,8 @@ static int __i915_drm_thaw(struct drm_device *dev)
 		 * */
 		intel_hpd_init(dev);
 		dev_priv->enable_hotplug_processing = true;
+		/* Config may have changed between suspend and resume */
+		intel_resume_hotplug(dev);
 	}
 
 	intel_opregion_init(dev);
@@ -732,6 +779,7 @@ static int ironlake_do_reset(struct drm_device *dev)
 	int ret;
 
 	gdrst = I915_READ(MCHBAR_MIRROR_BASE + ILK_GDSR);
+	gdrst &= ~GRDOM_MASK;
 	I915_WRITE(MCHBAR_MIRROR_BASE + ILK_GDSR,
 		   gdrst | GRDOM_RENDER | GRDOM_RESET_ENABLE);
 	ret = wait_for(I915_READ(MCHBAR_MIRROR_BASE + ILK_GDSR) & 0x1, 500);
@@ -740,6 +788,7 @@ static int ironlake_do_reset(struct drm_device *dev)
 
 	/* We can't reset render&media without also resetting display ... */
 	gdrst = I915_READ(MCHBAR_MIRROR_BASE + ILK_GDSR);
+	gdrst &= ~GRDOM_MASK;
 	I915_WRITE(MCHBAR_MIRROR_BASE + ILK_GDSR,
 		   gdrst | GRDOM_MEDIA | GRDOM_RESET_ENABLE);
 	return wait_for(I915_READ(MCHBAR_MIRROR_BASE + ILK_GDSR) & 0x1, 500);
@@ -803,7 +852,7 @@ int intel_gpu_reset(struct drm_device *dev)
 
 	/* Also reset the gpu hangman. */
 	if (dev_priv->gpu_error.stop_rings) {
-		DRM_DEBUG("Simulated gpu hang, resetting stop_rings\n");
+		DRM_INFO("Simulated gpu hang, resetting stop_rings\n");
 		dev_priv->gpu_error.stop_rings = 0;
 		if (ret == -ENODEV) {
 			DRM_ERROR("Reset not implemented, but ignoring "
@@ -882,7 +931,11 @@ int i915_reset(struct drm_device *dev)
 			ring->init(ring);
 
 		i915_gem_context_init(dev);
-		i915_gem_init_ppgtt(dev);
+		if (dev_priv->mm.aliasing_ppgtt) {
+			ret = dev_priv->mm.aliasing_ppgtt->enable(dev);
+			if (ret)
+				i915_gem_cleanup_aliasing_ppgtt(dev);
+		}
 
 		/*
 		 * It would make sense to re-init all the other hw state, at
@@ -1147,6 +1200,27 @@ ilk_dummy_write(struct drm_i915_private *dev_priv)
 	I915_WRITE_NOTRACE(MI_MODE, 0);
 }
 
+static void
+hsw_unclaimed_reg_clear(struct drm_i915_private *dev_priv, u32 reg)
+{
+	if (IS_HASWELL(dev_priv->dev) &&
+	    (I915_READ_NOTRACE(FPGA_DBG) & FPGA_DBG_RM_NOCLAIM)) {
+		DRM_ERROR("Unknown unclaimed register before writing to %x\n",
+			  reg);
+		I915_WRITE_NOTRACE(FPGA_DBG, FPGA_DBG_RM_NOCLAIM);
+	}
+}
+
+static void
+hsw_unclaimed_reg_check(struct drm_i915_private *dev_priv, u32 reg)
+{
+	if (IS_HASWELL(dev_priv->dev) &&
+	    (I915_READ_NOTRACE(FPGA_DBG) & FPGA_DBG_RM_NOCLAIM)) {
+		DRM_ERROR("Unclaimed write to %x\n", reg);
+		I915_WRITE_NOTRACE(FPGA_DBG, FPGA_DBG_RM_NOCLAIM);
+	}
+}
+
 #define __i915_read(x, y) \
 u##x i915_read##x(struct drm_i915_private *dev_priv, u32 reg) { \
 	u##x val = 0; \
@@ -1183,18 +1257,12 @@ void i915_write##x(struct drm_i915_private *dev_priv, u32 reg, u##x val) { \
 	} \
 	if (IS_GEN5(dev_priv->dev)) \
 		ilk_dummy_write(dev_priv); \
-	if (IS_HASWELL(dev_priv->dev) && (I915_READ_NOTRACE(GEN7_ERR_INT) & ERR_INT_MMIO_UNCLAIMED)) { \
-		DRM_ERROR("Unknown unclaimed register before writing to %x\n", reg); \
-		I915_WRITE_NOTRACE(GEN7_ERR_INT, ERR_INT_MMIO_UNCLAIMED); \
-	} \
+	hsw_unclaimed_reg_clear(dev_priv, reg); \
 	write##y(val, dev_priv->regs + reg); \
 	if (unlikely(__fifo_ret)) { \
 		gen6_gt_check_fifodbg(dev_priv); \
 	} \
-	if (IS_HASWELL(dev_priv->dev) && (I915_READ_NOTRACE(GEN7_ERR_INT) & ERR_INT_MMIO_UNCLAIMED)) { \
-		DRM_ERROR("Unclaimed write to %x\n", reg); \
-		writel(ERR_INT_MMIO_UNCLAIMED, dev_priv->regs + GEN7_ERR_INT);	\
-	} \
+	hsw_unclaimed_reg_check(dev_priv, reg); \
 }
 __i915_write(8, b)
 __i915_write(16, w)
diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h
index 01769e2a9953..d5dcf7fe1ee9 100644
--- a/drivers/gpu/drm/i915/i915_drv.h
+++ b/drivers/gpu/drm/i915/i915_drv.h
@@ -86,6 +86,19 @@ enum port {
 };
 #define port_name(p) ((p) + 'A')
 
+enum hpd_pin {
+	HPD_NONE = 0,
+	HPD_PORT_A = HPD_NONE, /* PORT_A is internal */
+	HPD_TV = HPD_NONE,     /* TV is known to be unreliable */
+	HPD_CRT,
+	HPD_SDVO_B,
+	HPD_SDVO_C,
+	HPD_PORT_B,
+	HPD_PORT_C,
+	HPD_PORT_D,
+	HPD_NUM_PINS
+};
+
 #define I915_GEM_GPU_DOMAINS \
 	(I915_GEM_DOMAIN_RENDER | \
 	 I915_GEM_DOMAIN_SAMPLER | \
@@ -93,7 +106,7 @@ enum port {
 	 I915_GEM_DOMAIN_INSTRUCTION | \
 	 I915_GEM_DOMAIN_VERTEX)
 
-#define for_each_pipe(p) for ((p) = 0; (p) < dev_priv->num_pipe; (p)++)
+#define for_each_pipe(p) for ((p) = 0; (p) < INTEL_INFO(dev)->num_pipes; (p)++)
 
 #define for_each_encoder_on_crtc(dev, __crtc, intel_encoder) \
 	list_for_each_entry((intel_encoder), &(dev)->mode_config.encoder_list, base.head) \
@@ -182,9 +195,9 @@ struct drm_i915_master_private {
 	struct _drm_i915_sarea *sarea_priv;
 };
 #define I915_FENCE_REG_NONE -1
-#define I915_MAX_NUM_FENCES 16
-/* 16 fences + sign bit for FENCE_REG_NONE */
-#define I915_MAX_NUM_FENCE_BITS 5
+#define I915_MAX_NUM_FENCES 32
+/* 32 fences + sign bit for FENCE_REG_NONE */
+#define I915_MAX_NUM_FENCE_BITS 6
 
 struct drm_i915_fence_reg {
 	struct list_head lru_list;
@@ -243,7 +256,7 @@ struct drm_i915_error_state {
 			int page_count;
 			u32 gtt_offset;
 			u32 *pages[0];
-		} *ringbuffer, *batchbuffer;
+		} *ringbuffer, *batchbuffer, *ctx;
 		struct drm_i915_error_request {
 			long jiffies;
 			u32 seqno;
@@ -271,6 +284,9 @@ struct drm_i915_error_state {
 	struct intel_display_error_state *display;
 };
 
+struct intel_crtc_config;
+struct intel_crtc;
+
 struct drm_i915_display_funcs {
 	bool (*fbc_enabled)(struct drm_device *dev);
 	void (*enable_fbc)(struct drm_crtc *crtc, unsigned long interval);
@@ -283,9 +299,11 @@ struct drm_i915_display_funcs {
 	void (*update_linetime_wm)(struct drm_device *dev, int pipe,
 				 struct drm_display_mode *mode);
 	void (*modeset_global_resources)(struct drm_device *dev);
+	/* Returns the active state of the crtc, and if the crtc is active,
+	 * fills out the pipe-config with the hw state. */
+	bool (*get_pipe_config)(struct intel_crtc *,
+				struct intel_crtc_config *);
 	int (*crtc_mode_set)(struct drm_crtc *crtc,
-			     struct drm_display_mode *mode,
-			     struct drm_display_mode *adjusted_mode,
 			     int x, int y,
 			     struct drm_framebuffer *old_fb);
 	void (*crtc_enable)(struct drm_crtc *crtc);
@@ -341,6 +359,7 @@ struct drm_i915_gt_funcs {
 
 struct intel_device_info {
 	u32 display_mmio_offset;
+	u8 num_pipes:3;
 	u8 gen;
 	u8 is_mobile:1;
 	u8 is_i85x:1;
@@ -430,6 +449,7 @@ struct i915_hw_ppgtt {
 			       struct sg_table *st,
 			       unsigned int pg_start,
 			       enum i915_cache_level cache_level);
+	int (*enable)(struct drm_device *dev);
 	void (*cleanup)(struct i915_hw_ppgtt *ppgtt);
 };
 
@@ -460,6 +480,7 @@ enum intel_pch {
 	PCH_IBX,	/* Ibexpeak PCH */
 	PCH_CPT,	/* Cougarpoint PCH */
 	PCH_LPT,	/* Lynxpoint PCH */
+	PCH_NOP,
 };
 
 enum intel_sbi_destination {
@@ -647,6 +668,7 @@ struct intel_gen6_power_mgmt {
 	u8 cur_delay;
 	u8 min_delay;
 	u8 max_delay;
+	u8 hw_max;
 
 	struct delayed_work delayed_resume_work;
 
@@ -905,16 +927,24 @@ typedef struct drm_i915_private {
 	struct mutex dpio_lock;
 
 	/** Cached value of IMR to avoid reads in updating the bitfield */
-	u32 pipestat[2];
 	u32 irq_mask;
 	u32 gt_irq_mask;
 
-	u32 hotplug_supported_mask;
 	struct work_struct hotplug_work;
 	bool enable_hotplug_processing;
+	struct {
+		unsigned long hpd_last_jiffies;
+		int hpd_cnt;
+		enum {
+			HPD_ENABLED = 0,
+			HPD_DISABLED = 1,
+			HPD_MARK_DISABLED = 2
+		} hpd_mark;
+	} hpd_stats[HPD_NUM_PINS];
+	struct timer_list hotplug_reenable_timer;
 
-	int num_pipe;
 	int num_pch_pll;
+	int num_plane;
 
 	unsigned long cfb_size;
 	unsigned int cfb_fb;
@@ -928,9 +958,14 @@ typedef struct drm_i915_private {
 	struct intel_overlay *overlay;
 	unsigned int sprite_scaling_enabled;
 
+	/* backlight */
+	struct {
+		int level;
+		bool enabled;
+		struct backlight_device *device;
+	} backlight;
+
 	/* LVDS info */
-	int backlight_level;  /* restore backlight to this value */
-	bool backlight_enabled;
 	struct drm_display_mode *lfp_lvds_vbt_mode; /* if any */
 	struct drm_display_mode *sdvo_lvds_vbt_mode; /* if any */
 
@@ -941,6 +976,7 @@ typedef struct drm_i915_private {
 	unsigned int int_crt_support:1;
 	unsigned int lvds_use_ssc:1;
 	unsigned int display_clock_mode:1;
+	unsigned int fdi_rx_polarity_inverted:1;
 	int lvds_ssc_freq;
 	unsigned int bios_lvds_val; /* initial [PCH_]LVDS reg val in VBIOS */
 	struct {
@@ -1032,8 +1068,6 @@ typedef struct drm_i915_private {
 	 */
 	struct work_struct console_resume_work;
 
-	struct backlight_device *backlight;
-
 	struct drm_property *broadcast_rgb_property;
 	struct drm_property *force_audio_property;
 
@@ -1340,6 +1374,7 @@ struct drm_i915_file_private {
 #define HAS_PIPE_CONTROL(dev) (INTEL_INFO(dev)->gen >= 5)
 
 #define HAS_DDI(dev)		(IS_HASWELL(dev))
+#define HAS_POWER_WELL(dev)	(IS_HASWELL(dev))
 
 #define INTEL_PCH_DEVICE_ID_MASK		0xff00
 #define INTEL_PCH_IBX_DEVICE_ID_TYPE		0x3b00
@@ -1352,6 +1387,7 @@ struct drm_i915_file_private {
 #define HAS_PCH_LPT(dev) (INTEL_PCH_TYPE(dev) == PCH_LPT)
 #define HAS_PCH_CPT(dev) (INTEL_PCH_TYPE(dev) == PCH_CPT)
 #define HAS_PCH_IBX(dev) (INTEL_PCH_TYPE(dev) == PCH_IBX)
+#define HAS_PCH_NOP(dev) (INTEL_PCH_TYPE(dev) == PCH_NOP)
 #define HAS_PCH_SPLIT(dev) (INTEL_PCH_TYPE(dev) != PCH_NONE)
 
 #define HAS_FORCE_WAKE(dev) (INTEL_INFO(dev)->has_force_wake)
@@ -1529,17 +1565,12 @@ void i915_gem_lastclose(struct drm_device *dev);
 int __must_check i915_gem_object_get_pages(struct drm_i915_gem_object *obj);
 static inline struct page *i915_gem_object_get_page(struct drm_i915_gem_object *obj, int n)
 {
-	struct scatterlist *sg = obj->pages->sgl;
-	int nents = obj->pages->nents;
-	while (nents > SG_MAX_SINGLE_ALLOC) {
-		if (n < SG_MAX_SINGLE_ALLOC - 1)
-			break;
-
-		sg = sg_chain_ptr(sg + SG_MAX_SINGLE_ALLOC - 1);
-		n -= SG_MAX_SINGLE_ALLOC - 1;
-		nents -= SG_MAX_SINGLE_ALLOC - 1;
-	}
-	return sg_page(sg+n);
+	struct sg_page_iter sg_iter;
+
+	for_each_sg_page(obj->pages->sgl, &sg_iter, obj->pages->nents, n)
+		return sg_page_iter_page(&sg_iter);
+
+	return NULL;
 }
 static inline void i915_gem_object_pin_pages(struct drm_i915_gem_object *obj)
 {
@@ -1624,7 +1655,6 @@ int __must_check i915_gem_init(struct drm_device *dev);
 int __must_check i915_gem_init_hw(struct drm_device *dev);
 void i915_gem_l3_remap(struct drm_device *dev);
 void i915_gem_init_swizzling(struct drm_device *dev);
-void i915_gem_init_ppgtt(struct drm_device *dev);
 void i915_gem_cleanup_ringbuffer(struct drm_device *dev);
 int __must_check i915_gpu_idle(struct drm_device *dev);
 int __must_check i915_gem_idle(struct drm_device *dev);
@@ -1718,6 +1748,11 @@ void i915_gem_stolen_cleanup_compression(struct drm_device *dev);
 void i915_gem_cleanup_stolen(struct drm_device *dev);
 struct drm_i915_gem_object *
 i915_gem_object_create_stolen(struct drm_device *dev, u32 size);
+struct drm_i915_gem_object *
+i915_gem_object_create_stolen_for_preallocated(struct drm_device *dev,
+					       u32 stolen_offset,
+					       u32 gtt_offset,
+					       u32 size);
 void i915_gem_object_release_stolen(struct drm_i915_gem_object *obj);
 
 /* i915_gem_tiling.c */
@@ -1848,6 +1883,8 @@ int __gen6_gt_wait_for_fifo(struct drm_i915_private *dev_priv);
 
 int sandybridge_pcode_read(struct drm_i915_private *dev_priv, u8 mbox, u32 *val);
 int sandybridge_pcode_write(struct drm_i915_private *dev_priv, u8 mbox, u32 val);
+int valleyview_punit_read(struct drm_i915_private *dev_priv, u8 addr, u32 *val);
+int valleyview_punit_write(struct drm_i915_private *dev_priv, u8 addr, u32 val);
 
 #define __i915_read(x, y) \
 	u##x i915_read##x(struct drm_i915_private *dev_priv, u32 reg);
@@ -1901,4 +1938,9 @@ static inline uint32_t i915_vgacntrl_reg(struct drm_device *dev)
 		return VGACNTRL;
 }
 
+static inline void __user *to_user_ptr(u64 address)
+{
+	return (void __user *)(uintptr_t)address;
+}
+
 #endif
diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c
index 0e207e6e0df8..6be940effefd 100644
--- a/drivers/gpu/drm/i915/i915_gem.c
+++ b/drivers/gpu/drm/i915/i915_gem.c
@@ -411,10 +411,9 @@ i915_gem_shmem_pread(struct drm_device *dev,
 	int obj_do_bit17_swizzling, page_do_bit17_swizzling;
 	int prefaulted = 0;
 	int needs_clflush = 0;
-	struct scatterlist *sg;
-	int i;
+	struct sg_page_iter sg_iter;
 
-	user_data = (char __user *) (uintptr_t) args->data_ptr;
+	user_data = to_user_ptr(args->data_ptr);
 	remain = args->size;
 
 	obj_do_bit17_swizzling = i915_gem_object_needs_bit17_swizzle(obj);
@@ -441,11 +440,9 @@ i915_gem_shmem_pread(struct drm_device *dev,
 
 	offset = args->offset;
 
-	for_each_sg(obj->pages->sgl, sg, obj->pages->nents, i) {
-		struct page *page;
-
-		if (i < offset >> PAGE_SHIFT)
-			continue;
+	for_each_sg_page(obj->pages->sgl, &sg_iter, obj->pages->nents,
+			 offset >> PAGE_SHIFT) {
+		struct page *page = sg_page_iter_page(&sg_iter);
 
 		if (remain <= 0)
 			break;
@@ -460,7 +457,6 @@ i915_gem_shmem_pread(struct drm_device *dev,
 		if ((shmem_page_offset + page_length) > PAGE_SIZE)
 			page_length = PAGE_SIZE - shmem_page_offset;
 
-		page = sg_page(sg);
 		page_do_bit17_swizzling = obj_do_bit17_swizzling &&
 			(page_to_phys(page) & (1 << 17)) != 0;
 
@@ -522,7 +518,7 @@ i915_gem_pread_ioctl(struct drm_device *dev, void *data,
 		return 0;
 
 	if (!access_ok(VERIFY_WRITE,
-		       (char __user *)(uintptr_t)args->data_ptr,
+		       to_user_ptr(args->data_ptr),
 		       args->size))
 		return -EFAULT;
 
@@ -613,7 +609,7 @@ i915_gem_gtt_pwrite_fast(struct drm_device *dev,
 	if (ret)
 		goto out_unpin;
 
-	user_data = (char __user *) (uintptr_t) args->data_ptr;
+	user_data = to_user_ptr(args->data_ptr);
 	remain = args->size;
 
 	offset = obj->gtt_offset + args->offset;
@@ -732,10 +728,9 @@ i915_gem_shmem_pwrite(struct drm_device *dev,
 	int hit_slowpath = 0;
 	int needs_clflush_after = 0;
 	int needs_clflush_before = 0;
-	int i;
-	struct scatterlist *sg;
+	struct sg_page_iter sg_iter;
 
-	user_data = (char __user *) (uintptr_t) args->data_ptr;
+	user_data = to_user_ptr(args->data_ptr);
 	remain = args->size;
 
 	obj_do_bit17_swizzling = i915_gem_object_needs_bit17_swizzle(obj);
@@ -768,13 +763,11 @@ i915_gem_shmem_pwrite(struct drm_device *dev,
 	offset = args->offset;
 	obj->dirty = 1;
 
-	for_each_sg(obj->pages->sgl, sg, obj->pages->nents, i) {
-		struct page *page;
+	for_each_sg_page(obj->pages->sgl, &sg_iter, obj->pages->nents,
+			 offset >> PAGE_SHIFT) {
+		struct page *page = sg_page_iter_page(&sg_iter);
 		int partial_cacheline_write;
 
-		if (i < offset >> PAGE_SHIFT)
-			continue;
-
 		if (remain <= 0)
 			break;
 
@@ -796,7 +789,6 @@ i915_gem_shmem_pwrite(struct drm_device *dev,
 			((shmem_page_offset | page_length)
 				& (boot_cpu_data.x86_clflush_size - 1));
 
-		page = sg_page(sg);
 		page_do_bit17_swizzling = obj_do_bit17_swizzling &&
 			(page_to_phys(page) & (1 << 17)) != 0;
 
@@ -867,11 +859,11 @@ i915_gem_pwrite_ioctl(struct drm_device *dev, void *data,
 		return 0;
 
 	if (!access_ok(VERIFY_READ,
-		       (char __user *)(uintptr_t)args->data_ptr,
+		       to_user_ptr(args->data_ptr),
 		       args->size))
 		return -EFAULT;
 
-	ret = fault_in_multipages_readable((char __user *)(uintptr_t)args->data_ptr,
+	ret = fault_in_multipages_readable(to_user_ptr(args->data_ptr),
 					   args->size);
 	if (ret)
 		return -EFAULT;
@@ -1633,9 +1625,8 @@ i915_gem_object_is_purgeable(struct drm_i915_gem_object *obj)
 static void
 i915_gem_object_put_pages_gtt(struct drm_i915_gem_object *obj)
 {
-	int page_count = obj->base.size / PAGE_SIZE;
-	struct scatterlist *sg;
-	int ret, i;
+	struct sg_page_iter sg_iter;
+	int ret;
 
 	BUG_ON(obj->madv == __I915_MADV_PURGED);
 
@@ -1655,8 +1646,8 @@ i915_gem_object_put_pages_gtt(struct drm_i915_gem_object *obj)
 	if (obj->madv == I915_MADV_DONTNEED)
 		obj->dirty = 0;
 
-	for_each_sg(obj->pages->sgl, sg, page_count, i) {
-		struct page *page = sg_page(sg);
+	for_each_sg_page(obj->pages->sgl, &sg_iter, obj->pages->nents, 0) {
+		struct page *page = sg_page_iter_page(&sg_iter);
 
 		if (obj->dirty)
 			set_page_dirty(page);
@@ -1757,7 +1748,9 @@ i915_gem_object_get_pages_gtt(struct drm_i915_gem_object *obj)
 	struct address_space *mapping;
 	struct sg_table *st;
 	struct scatterlist *sg;
+	struct sg_page_iter sg_iter;
 	struct page *page;
+	unsigned long last_pfn = 0;	/* suppress gcc warning */
 	gfp_t gfp;
 
 	/* Assert that the object is not currently in any GPU domain. As it
@@ -1787,7 +1780,9 @@ i915_gem_object_get_pages_gtt(struct drm_i915_gem_object *obj)
 	gfp = mapping_gfp_mask(mapping);
 	gfp |= __GFP_NORETRY | __GFP_NOWARN | __GFP_NO_KSWAPD;
 	gfp &= ~(__GFP_IO | __GFP_WAIT);
-	for_each_sg(st->sgl, sg, page_count, i) {
+	sg = st->sgl;
+	st->nents = 0;
+	for (i = 0; i < page_count; i++) {
 		page = shmem_read_mapping_page_gfp(mapping, i, gfp);
 		if (IS_ERR(page)) {
 			i915_gem_purge(dev_priv, page_count);
@@ -1810,9 +1805,18 @@ i915_gem_object_get_pages_gtt(struct drm_i915_gem_object *obj)
 			gfp &= ~(__GFP_IO | __GFP_WAIT);
 		}
 
-		sg_set_page(sg, page, PAGE_SIZE, 0);
+		if (!i || page_to_pfn(page) != last_pfn + 1) {
+			if (i)
+				sg = sg_next(sg);
+			st->nents++;
+			sg_set_page(sg, page, PAGE_SIZE, 0);
+		} else {
+			sg->length += PAGE_SIZE;
+		}
+		last_pfn = page_to_pfn(page);
 	}
 
+	sg_mark_end(sg);
 	obj->pages = st;
 
 	if (i915_gem_object_needs_bit17_swizzle(obj))
@@ -1821,8 +1825,9 @@ i915_gem_object_get_pages_gtt(struct drm_i915_gem_object *obj)
 	return 0;
 
 err_pages:
-	for_each_sg(st->sgl, sg, i, page_count)
-		page_cache_release(sg_page(sg));
+	sg_mark_end(sg);
+	for_each_sg_page(st->sgl, &sg_iter, st->nents, 0)
+		page_cache_release(sg_page_iter_page(&sg_iter));
 	sg_free_table(st);
 	kfree(st);
 	return PTR_ERR(page);
@@ -2123,11 +2128,11 @@ static void i915_gem_reset_fences(struct drm_device *dev)
 	for (i = 0; i < dev_priv->num_fence_regs; i++) {
 		struct drm_i915_fence_reg *reg = &dev_priv->fence_regs[i];
 
-		i915_gem_write_fence(dev, i, NULL);
-
 		if (reg->obj)
 			i915_gem_object_fence_lost(reg->obj);
 
+		i915_gem_write_fence(dev, i, NULL);
+
 		reg->pin_count = 0;
 		reg->obj = NULL;
 		INIT_LIST_HEAD(&reg->lru_list);
@@ -2678,17 +2683,35 @@ static inline int fence_number(struct drm_i915_private *dev_priv,
 	return fence - dev_priv->fence_regs;
 }
 
+static void i915_gem_write_fence__ipi(void *data)
+{
+	wbinvd();
+}
+
 static void i915_gem_object_update_fence(struct drm_i915_gem_object *obj,
 					 struct drm_i915_fence_reg *fence,
 					 bool enable)
 {
-	struct drm_i915_private *dev_priv = obj->base.dev->dev_private;
-	int reg = fence_number(dev_priv, fence);
-
-	i915_gem_write_fence(obj->base.dev, reg, enable ? obj : NULL);
+	struct drm_device *dev = obj->base.dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	int fence_reg = fence_number(dev_priv, fence);
+
+	/* In order to fully serialize access to the fenced region and
+	 * the update to the fence register we need to take extreme
+	 * measures on SNB+. In theory, the write to the fence register
+	 * flushes all memory transactions before, and coupled with the
+	 * mb() placed around the register write we serialise all memory
+	 * operations with respect to the changes in the tiler. Yet, on
+	 * SNB+ we need to take a step further and emit an explicit wbinvd()
+	 * on each processor in order to manually flush all memory
+	 * transactions before updating the fence register.
+	 */
+	if (HAS_LLC(obj->base.dev))
+		on_each_cpu(i915_gem_write_fence__ipi, NULL, 1);
+	i915_gem_write_fence(dev, fence_reg, enable ? obj : NULL);
 
 	if (enable) {
-		obj->fence_reg = reg;
+		obj->fence_reg = fence_reg;
 		fence->obj = obj;
 		list_move_tail(&fence->lru_list, &dev_priv->mm.fence_list);
 	} else {
@@ -2717,6 +2740,7 @@ int
 i915_gem_object_put_fence(struct drm_i915_gem_object *obj)
 {
 	struct drm_i915_private *dev_priv = obj->base.dev->dev_private;
+	struct drm_i915_fence_reg *fence;
 	int ret;
 
 	ret = i915_gem_object_wait_fence(obj);
@@ -2726,10 +2750,10 @@ i915_gem_object_put_fence(struct drm_i915_gem_object *obj)
 	if (obj->fence_reg == I915_FENCE_REG_NONE)
 		return 0;
 
-	i915_gem_object_update_fence(obj,
-				     &dev_priv->fence_regs[obj->fence_reg],
-				     false);
+	fence = &dev_priv->fence_regs[obj->fence_reg];
+
 	i915_gem_object_fence_lost(obj);
+	i915_gem_object_update_fence(obj, fence, false);
 
 	return 0;
 }
@@ -3986,6 +4010,12 @@ i915_gem_init_hw(struct drm_device *dev)
 	if (IS_HASWELL(dev) && (I915_READ(0x120010) == 1))
 		I915_WRITE(0x9008, I915_READ(0x9008) | 0xf0000);
 
+	if (HAS_PCH_NOP(dev)) {
+		u32 temp = I915_READ(GEN7_MSG_CTL);
+		temp &= ~(WAIT_FOR_PCH_FLR_ACK | WAIT_FOR_PCH_RESET_ACK);
+		I915_WRITE(GEN7_MSG_CTL, temp);
+	}
+
 	i915_gem_l3_remap(dev);
 
 	i915_gem_init_swizzling(dev);
@@ -3999,7 +4029,13 @@ i915_gem_init_hw(struct drm_device *dev)
 	 * contexts before PPGTT.
 	 */
 	i915_gem_context_init(dev);
-	i915_gem_init_ppgtt(dev);
+	if (dev_priv->mm.aliasing_ppgtt) {
+		ret = dev_priv->mm.aliasing_ppgtt->enable(dev);
+		if (ret) {
+			i915_gem_cleanup_aliasing_ppgtt(dev);
+			DRM_INFO("PPGTT enable failed. This is not fatal, but unexpected\n");
+		}
+	}
 
 	return 0;
 }
@@ -4010,7 +4046,16 @@ int i915_gem_init(struct drm_device *dev)
 	int ret;
 
 	mutex_lock(&dev->struct_mutex);
+
+	if (IS_VALLEYVIEW(dev)) {
+		/* VLVA0 (potential hack), BIOS isn't actually waking us */
+		I915_WRITE(VLV_GTLC_WAKE_CTRL, 1);
+		if (wait_for((I915_READ(VLV_GTLC_PW_STATUS) & 1) == 1, 10))
+			DRM_DEBUG_DRIVER("allow wake ack timed out\n");
+	}
+
 	i915_gem_init_global_gtt(dev);
+
 	ret = i915_gem_init_hw(dev);
 	mutex_unlock(&dev->struct_mutex);
 	if (ret) {
@@ -4145,7 +4190,9 @@ i915_gem_load(struct drm_device *dev)
 	if (!drm_core_check_feature(dev, DRIVER_MODESET))
 		dev_priv->fence_reg_start = 3;
 
-	if (INTEL_INFO(dev)->gen >= 4 || IS_I945G(dev) || IS_I945GM(dev) || IS_G33(dev))
+	if (INTEL_INFO(dev)->gen >= 7 && !IS_VALLEYVIEW(dev))
+		dev_priv->num_fence_regs = 32;
+	else if (INTEL_INFO(dev)->gen >= 4 || IS_I945G(dev) || IS_I945GM(dev) || IS_G33(dev))
 		dev_priv->num_fence_regs = 16;
 	else
 		dev_priv->num_fence_regs = 8;
@@ -4327,7 +4374,7 @@ i915_gem_phys_pwrite(struct drm_device *dev,
 		     struct drm_file *file_priv)
 {
 	void *vaddr = obj->phys_obj->handle->vaddr + args->offset;
-	char __user *user_data = (char __user *) (uintptr_t) args->data_ptr;
+	char __user *user_data = to_user_ptr(args->data_ptr);
 
 	if (__copy_from_user_inatomic_nocache(vaddr, user_data, args->size)) {
 		unsigned long unwritten;
diff --git a/drivers/gpu/drm/i915/i915_gem_context.c b/drivers/gpu/drm/i915/i915_gem_context.c
index 94d873a6cffb..a1e8ecb6adf6 100644
--- a/drivers/gpu/drm/i915/i915_gem_context.c
+++ b/drivers/gpu/drm/i915/i915_gem_context.c
@@ -152,6 +152,13 @@ create_hw_context(struct drm_device *dev,
 		return ERR_PTR(-ENOMEM);
 	}
 
+	if (INTEL_INFO(dev)->gen >= 7) {
+		ret = i915_gem_object_set_cache_level(ctx->obj,
+						      I915_CACHE_LLC_MLC);
+		if (ret)
+			goto err_out;
+	}
+
 	/* The ring associated with the context object is handled by the normal
 	 * object tracking code. We give an initial ring value simple to pass an
 	 * assertion in the context switch code.
diff --git a/drivers/gpu/drm/i915/i915_gem_dmabuf.c b/drivers/gpu/drm/i915/i915_gem_dmabuf.c
index 6a5af6828624..dc53a527126b 100644
--- a/drivers/gpu/drm/i915/i915_gem_dmabuf.c
+++ b/drivers/gpu/drm/i915/i915_gem_dmabuf.c
@@ -62,7 +62,7 @@ static struct sg_table *i915_gem_map_dma_buf(struct dma_buf_attachment *attachme
 	src = obj->pages->sgl;
 	dst = st->sgl;
 	for (i = 0; i < obj->pages->nents; i++) {
-		sg_set_page(dst, sg_page(src), PAGE_SIZE, 0);
+		sg_set_page(dst, sg_page(src), src->length, 0);
 		dst = sg_next(dst);
 		src = sg_next(src);
 	}
@@ -105,7 +105,7 @@ static void *i915_gem_dmabuf_vmap(struct dma_buf *dma_buf)
 {
 	struct drm_i915_gem_object *obj = dma_buf->priv;
 	struct drm_device *dev = obj->base.dev;
-	struct scatterlist *sg;
+	struct sg_page_iter sg_iter;
 	struct page **pages;
 	int ret, i;
 
@@ -124,14 +124,15 @@ static void *i915_gem_dmabuf_vmap(struct dma_buf *dma_buf)
 
 	ret = -ENOMEM;
 
-	pages = drm_malloc_ab(obj->pages->nents, sizeof(struct page *));
+	pages = drm_malloc_ab(obj->base.size >> PAGE_SHIFT, sizeof(*pages));
 	if (pages == NULL)
 		goto error;
 
-	for_each_sg(obj->pages->sgl, sg, obj->pages->nents, i)
-		pages[i] = sg_page(sg);
+	i = 0;
+	for_each_sg_page(obj->pages->sgl, &sg_iter, obj->pages->nents, 0)
+		pages[i++] = sg_page_iter_page(&sg_iter);
 
-	obj->dma_buf_vmapping = vmap(pages, obj->pages->nents, 0, PAGE_KERNEL);
+	obj->dma_buf_vmapping = vmap(pages, i, 0, PAGE_KERNEL);
 	drm_free_large(pages);
 
 	if (!obj->dma_buf_vmapping)
@@ -271,7 +272,6 @@ struct drm_gem_object *i915_gem_prime_import(struct drm_device *dev,
 			 * refcount on gem itself instead of f_count of dmabuf.
 			 */
 			drm_gem_object_reference(&obj->base);
-			dma_buf_put(dma_buf);
 			return &obj->base;
 		}
 	}
@@ -281,6 +281,8 @@ struct drm_gem_object *i915_gem_prime_import(struct drm_device *dev,
 	if (IS_ERR(attach))
 		return ERR_CAST(attach);
 
+	get_dma_buf(dma_buf);
+
 	obj = i915_gem_object_alloc(dev);
 	if (obj == NULL) {
 		ret = -ENOMEM;
@@ -300,5 +302,7 @@ struct drm_gem_object *i915_gem_prime_import(struct drm_device *dev,
 
 fail_detach:
 	dma_buf_detach(dma_buf, attach);
+	dma_buf_put(dma_buf);
+
 	return ERR_PTR(ret);
 }
diff --git a/drivers/gpu/drm/i915/i915_gem_execbuffer.c b/drivers/gpu/drm/i915/i915_gem_execbuffer.c
index 9a48e1a2d417..117ce3813681 100644
--- a/drivers/gpu/drm/i915/i915_gem_execbuffer.c
+++ b/drivers/gpu/drm/i915/i915_gem_execbuffer.c
@@ -305,7 +305,7 @@ i915_gem_execbuffer_relocate_object(struct drm_i915_gem_object *obj,
 	struct drm_i915_gem_exec_object2 *entry = obj->exec_entry;
 	int remain, ret;
 
-	user_relocs = (void __user *)(uintptr_t)entry->relocs_ptr;
+	user_relocs = to_user_ptr(entry->relocs_ptr);
 
 	remain = entry->relocation_count;
 	while (remain) {
@@ -359,8 +359,7 @@ i915_gem_execbuffer_relocate_object_slow(struct drm_i915_gem_object *obj,
 }
 
 static int
-i915_gem_execbuffer_relocate(struct drm_device *dev,
-			     struct eb_objects *eb)
+i915_gem_execbuffer_relocate(struct eb_objects *eb)
 {
 	struct drm_i915_gem_object *obj;
 	int ret = 0;
@@ -475,7 +474,6 @@ i915_gem_execbuffer_unreserve_object(struct drm_i915_gem_object *obj)
 
 static int
 i915_gem_execbuffer_reserve(struct intel_ring_buffer *ring,
-			    struct drm_file *file,
 			    struct list_head *objects,
 			    bool *need_relocs)
 {
@@ -618,7 +616,7 @@ i915_gem_execbuffer_relocate_slow(struct drm_device *dev,
 		u64 invalid_offset = (u64)-1;
 		int j;
 
-		user_relocs = (void __user *)(uintptr_t)exec[i].relocs_ptr;
+		user_relocs = to_user_ptr(exec[i].relocs_ptr);
 
 		if (copy_from_user(reloc+total, user_relocs,
 				   exec[i].relocation_count * sizeof(*reloc))) {
@@ -663,7 +661,7 @@ i915_gem_execbuffer_relocate_slow(struct drm_device *dev,
 		goto err;
 
 	need_relocs = (args->flags & I915_EXEC_NO_RELOC) == 0;
-	ret = i915_gem_execbuffer_reserve(ring, file, &eb->objects, &need_relocs);
+	ret = i915_gem_execbuffer_reserve(ring, &eb->objects, &need_relocs);
 	if (ret)
 		goto err;
 
@@ -736,7 +734,7 @@ validate_exec_list(struct drm_i915_gem_exec_object2 *exec,
 	int relocs_max = INT_MAX / sizeof(struct drm_i915_gem_relocation_entry);
 
 	for (i = 0; i < count; i++) {
-		char __user *ptr = (char __user *)(uintptr_t)exec[i].relocs_ptr;
+		char __user *ptr = to_user_ptr(exec[i].relocs_ptr);
 		int length; /* limited by fault_in_pages_readable() */
 
 		if (exec[i].flags & __EXEC_OBJECT_UNKNOWN_FLAGS)
@@ -752,7 +750,11 @@ validate_exec_list(struct drm_i915_gem_exec_object2 *exec,
 
 		length = exec[i].relocation_count *
 			sizeof(struct drm_i915_gem_relocation_entry);
-		/* we may also need to update the presumed offsets */
+		/*
+		 * We must check that the entire relocation array is safe
+		 * to read, but since we may need to update the presumed
+		 * offsets during execution, check for full write access.
+		 */
 		if (!access_ok(VERIFY_WRITE, ptr, length))
 			return -EFAULT;
 
@@ -949,9 +951,8 @@ i915_gem_do_execbuffer(struct drm_device *dev, void *data,
 		}
 
 		if (copy_from_user(cliprects,
-				     (struct drm_clip_rect __user *)(uintptr_t)
-				     args->cliprects_ptr,
-				     sizeof(*cliprects)*args->num_cliprects)) {
+				   to_user_ptr(args->cliprects_ptr),
+				   sizeof(*cliprects)*args->num_cliprects)) {
 			ret = -EFAULT;
 			goto pre_mutex_err;
 		}
@@ -986,13 +987,13 @@ i915_gem_do_execbuffer(struct drm_device *dev, void *data,
 
 	/* Move the objects en-masse into the GTT, evicting if necessary. */
 	need_relocs = (args->flags & I915_EXEC_NO_RELOC) == 0;
-	ret = i915_gem_execbuffer_reserve(ring, file, &eb->objects, &need_relocs);
+	ret = i915_gem_execbuffer_reserve(ring, &eb->objects, &need_relocs);
 	if (ret)
 		goto err;
 
 	/* The objects are in their final locations, apply the relocations. */
 	if (need_relocs)
-		ret = i915_gem_execbuffer_relocate(dev, eb);
+		ret = i915_gem_execbuffer_relocate(eb);
 	if (ret) {
 		if (ret == -EFAULT) {
 			ret = i915_gem_execbuffer_relocate_slow(dev, args, file, ring,
@@ -1115,7 +1116,7 @@ i915_gem_execbuffer(struct drm_device *dev, void *data,
 		return -ENOMEM;
 	}
 	ret = copy_from_user(exec_list,
-			     (void __user *)(uintptr_t)args->buffers_ptr,
+			     to_user_ptr(args->buffers_ptr),
 			     sizeof(*exec_list) * args->buffer_count);
 	if (ret != 0) {
 		DRM_DEBUG("copy %d exec entries failed %d\n",
@@ -1154,7 +1155,7 @@ i915_gem_execbuffer(struct drm_device *dev, void *data,
 		for (i = 0; i < args->buffer_count; i++)
 			exec_list[i].offset = exec2_list[i].offset;
 		/* ... and back out to userspace */
-		ret = copy_to_user((void __user *)(uintptr_t)args->buffers_ptr,
+		ret = copy_to_user(to_user_ptr(args->buffers_ptr),
 				   exec_list,
 				   sizeof(*exec_list) * args->buffer_count);
 		if (ret) {
@@ -1195,8 +1196,7 @@ i915_gem_execbuffer2(struct drm_device *dev, void *data,
 		return -ENOMEM;
 	}
 	ret = copy_from_user(exec2_list,
-			     (struct drm_i915_relocation_entry __user *)
-			     (uintptr_t) args->buffers_ptr,
+			     to_user_ptr(args->buffers_ptr),
 			     sizeof(*exec2_list) * args->buffer_count);
 	if (ret != 0) {
 		DRM_DEBUG("copy %d exec entries failed %d\n",
@@ -1208,7 +1208,7 @@ i915_gem_execbuffer2(struct drm_device *dev, void *data,
 	ret = i915_gem_do_execbuffer(dev, data, file, args, exec2_list);
 	if (!ret) {
 		/* Copy the new buffer offsets back to the user's exec list. */
-		ret = copy_to_user((void __user *)(uintptr_t)args->buffers_ptr,
+		ret = copy_to_user(to_user_ptr(args->buffers_ptr),
 				   exec2_list,
 				   sizeof(*exec2_list) * args->buffer_count);
 		if (ret) {
diff --git a/drivers/gpu/drm/i915/i915_gem_gtt.c b/drivers/gpu/drm/i915/i915_gem_gtt.c
index 926a1e2dd234..dca614de71b6 100644
--- a/drivers/gpu/drm/i915/i915_gem_gtt.c
+++ b/drivers/gpu/drm/i915/i915_gem_gtt.c
@@ -28,7 +28,7 @@
 #include "i915_trace.h"
 #include "intel_drv.h"
 
-typedef uint32_t gtt_pte_t;
+typedef uint32_t gen6_gtt_pte_t;
 
 /* PPGTT stuff */
 #define GEN6_GTT_ADDR_ENCODE(addr)	((addr) | (((addr) >> 28) & 0xff0))
@@ -44,11 +44,11 @@ typedef uint32_t gtt_pte_t;
 #define GEN6_PTE_CACHE_LLC_MLC		(3 << 1)
 #define GEN6_PTE_ADDR_ENCODE(addr)	GEN6_GTT_ADDR_ENCODE(addr)
 
-static inline gtt_pte_t gen6_pte_encode(struct drm_device *dev,
-					dma_addr_t addr,
-					enum i915_cache_level level)
+static inline gen6_gtt_pte_t gen6_pte_encode(struct drm_device *dev,
+					     dma_addr_t addr,
+					     enum i915_cache_level level)
 {
-	gtt_pte_t pte = GEN6_PTE_VALID;
+	gen6_gtt_pte_t pte = GEN6_PTE_VALID;
 	pte |= GEN6_PTE_ADDR_ENCODE(addr);
 
 	switch (level) {
@@ -72,18 +72,85 @@ static inline gtt_pte_t gen6_pte_encode(struct drm_device *dev,
 		BUG();
 	}
 
-
 	return pte;
 }
 
+static int gen6_ppgtt_enable(struct drm_device *dev)
+{
+	drm_i915_private_t *dev_priv = dev->dev_private;
+	uint32_t pd_offset;
+	struct intel_ring_buffer *ring;
+	struct i915_hw_ppgtt *ppgtt = dev_priv->mm.aliasing_ppgtt;
+	gen6_gtt_pte_t __iomem *pd_addr;
+	uint32_t pd_entry;
+	int i;
+
+	pd_addr = (gen6_gtt_pte_t __iomem*)dev_priv->gtt.gsm +
+		ppgtt->pd_offset / sizeof(gen6_gtt_pte_t);
+	for (i = 0; i < ppgtt->num_pd_entries; i++) {
+		dma_addr_t pt_addr;
+
+		pt_addr = ppgtt->pt_dma_addr[i];
+		pd_entry = GEN6_PDE_ADDR_ENCODE(pt_addr);
+		pd_entry |= GEN6_PDE_VALID;
+
+		writel(pd_entry, pd_addr + i);
+	}
+	readl(pd_addr);
+
+	pd_offset = ppgtt->pd_offset;
+	pd_offset /= 64; /* in cachelines, */
+	pd_offset <<= 16;
+
+	if (INTEL_INFO(dev)->gen == 6) {
+		uint32_t ecochk, gab_ctl, ecobits;
+
+		ecobits = I915_READ(GAC_ECO_BITS);
+		I915_WRITE(GAC_ECO_BITS, ecobits | ECOBITS_SNB_BIT |
+					 ECOBITS_PPGTT_CACHE64B);
+
+		gab_ctl = I915_READ(GAB_CTL);
+		I915_WRITE(GAB_CTL, gab_ctl | GAB_CTL_CONT_AFTER_PAGEFAULT);
+
+		ecochk = I915_READ(GAM_ECOCHK);
+		I915_WRITE(GAM_ECOCHK, ecochk | ECOCHK_SNB_BIT |
+				       ECOCHK_PPGTT_CACHE64B);
+		I915_WRITE(GFX_MODE, _MASKED_BIT_ENABLE(GFX_PPGTT_ENABLE));
+	} else if (INTEL_INFO(dev)->gen >= 7) {
+		uint32_t ecochk, ecobits;
+
+		ecobits = I915_READ(GAC_ECO_BITS);
+		I915_WRITE(GAC_ECO_BITS, ecobits | ECOBITS_PPGTT_CACHE64B);
+
+		ecochk = I915_READ(GAM_ECOCHK);
+		if (IS_HASWELL(dev)) {
+			ecochk |= ECOCHK_PPGTT_WB_HSW;
+		} else {
+			ecochk |= ECOCHK_PPGTT_LLC_IVB;
+			ecochk &= ~ECOCHK_PPGTT_GFDT_IVB;
+		}
+		I915_WRITE(GAM_ECOCHK, ecochk);
+		/* GFX_MODE is per-ring on gen7+ */
+	}
+
+	for_each_ring(ring, dev_priv, i) {
+		if (INTEL_INFO(dev)->gen >= 7)
+			I915_WRITE(RING_MODE_GEN7(ring),
+				   _MASKED_BIT_ENABLE(GFX_PPGTT_ENABLE));
+
+		I915_WRITE(RING_PP_DIR_DCLV(ring), PP_DIR_DCLV_2G);
+		I915_WRITE(RING_PP_DIR_BASE(ring), pd_offset);
+	}
+	return 0;
+}
+
 /* PPGTT support for Sandybdrige/Gen6 and later */
 static void gen6_ppgtt_clear_range(struct i915_hw_ppgtt *ppgtt,
 				   unsigned first_entry,
 				   unsigned num_entries)
 {
-	gtt_pte_t *pt_vaddr;
-	gtt_pte_t scratch_pte;
-	unsigned act_pd = first_entry / I915_PPGTT_PT_ENTRIES;
+	gen6_gtt_pte_t *pt_vaddr, scratch_pte;
+	unsigned act_pt = first_entry / I915_PPGTT_PT_ENTRIES;
 	unsigned first_pte = first_entry % I915_PPGTT_PT_ENTRIES;
 	unsigned last_pte, i;
 
@@ -96,7 +163,7 @@ static void gen6_ppgtt_clear_range(struct i915_hw_ppgtt *ppgtt,
 		if (last_pte > I915_PPGTT_PT_ENTRIES)
 			last_pte = I915_PPGTT_PT_ENTRIES;
 
-		pt_vaddr = kmap_atomic(ppgtt->pt_pages[act_pd]);
+		pt_vaddr = kmap_atomic(ppgtt->pt_pages[act_pt]);
 
 		for (i = first_pte; i < last_pte; i++)
 			pt_vaddr[i] = scratch_pte;
@@ -105,7 +172,7 @@ static void gen6_ppgtt_clear_range(struct i915_hw_ppgtt *ppgtt,
 
 		num_entries -= last_pte - first_pte;
 		first_pte = 0;
-		act_pd++;
+		act_pt++;
 	}
 }
 
@@ -114,43 +181,27 @@ static void gen6_ppgtt_insert_entries(struct i915_hw_ppgtt *ppgtt,
 				      unsigned first_entry,
 				      enum i915_cache_level cache_level)
 {
-	gtt_pte_t *pt_vaddr;
-	unsigned act_pd = first_entry / I915_PPGTT_PT_ENTRIES;
-	unsigned first_pte = first_entry % I915_PPGTT_PT_ENTRIES;
-	unsigned i, j, m, segment_len;
-	dma_addr_t page_addr;
-	struct scatterlist *sg;
-
-	/* init sg walking */
-	sg = pages->sgl;
-	i = 0;
-	segment_len = sg_dma_len(sg) >> PAGE_SHIFT;
-	m = 0;
-
-	while (i < pages->nents) {
-		pt_vaddr = kmap_atomic(ppgtt->pt_pages[act_pd]);
-
-		for (j = first_pte; j < I915_PPGTT_PT_ENTRIES; j++) {
-			page_addr = sg_dma_address(sg) + (m << PAGE_SHIFT);
-			pt_vaddr[j] = gen6_pte_encode(ppgtt->dev, page_addr,
-						      cache_level);
-
-			/* grab the next page */
-			if (++m == segment_len) {
-				if (++i == pages->nents)
-					break;
-
-				sg = sg_next(sg);
-				segment_len = sg_dma_len(sg) >> PAGE_SHIFT;
-				m = 0;
-			}
-		}
+	gen6_gtt_pte_t *pt_vaddr;
+	unsigned act_pt = first_entry / I915_PPGTT_PT_ENTRIES;
+	unsigned act_pte = first_entry % I915_PPGTT_PT_ENTRIES;
+	struct sg_page_iter sg_iter;
 
-		kunmap_atomic(pt_vaddr);
+	pt_vaddr = kmap_atomic(ppgtt->pt_pages[act_pt]);
+	for_each_sg_page(pages->sgl, &sg_iter, pages->nents, 0) {
+		dma_addr_t page_addr;
 
-		first_pte = 0;
-		act_pd++;
+		page_addr = sg_page_iter_dma_address(&sg_iter);
+		pt_vaddr[act_pte] = gen6_pte_encode(ppgtt->dev, page_addr,
+						    cache_level);
+		if (++act_pte == I915_PPGTT_PT_ENTRIES) {
+			kunmap_atomic(pt_vaddr);
+			act_pt++;
+			pt_vaddr = kmap_atomic(ppgtt->pt_pages[act_pt]);
+			act_pte = 0;
+
+		}
 	}
+	kunmap_atomic(pt_vaddr);
 }
 
 static void gen6_ppgtt_cleanup(struct i915_hw_ppgtt *ppgtt)
@@ -182,10 +233,10 @@ static int gen6_ppgtt_init(struct i915_hw_ppgtt *ppgtt)
 	/* ppgtt PDEs reside in the global gtt pagetable, which has 512*1024
 	 * entries. For aliasing ppgtt support we just steal them at the end for
 	 * now. */
-	first_pd_entry_in_global_pt =
-		gtt_total_entries(dev_priv->gtt) - I915_PPGTT_PD_ENTRIES;
+       first_pd_entry_in_global_pt = gtt_total_entries(dev_priv->gtt);
 
 	ppgtt->num_pd_entries = I915_PPGTT_PD_ENTRIES;
+	ppgtt->enable = gen6_ppgtt_enable;
 	ppgtt->clear_range = gen6_ppgtt_clear_range;
 	ppgtt->insert_entries = gen6_ppgtt_insert_entries;
 	ppgtt->cleanup = gen6_ppgtt_cleanup;
@@ -219,12 +270,10 @@ static int gen6_ppgtt_init(struct i915_hw_ppgtt *ppgtt)
 		ppgtt->pt_dma_addr[i] = pt_addr;
 	}
 
-	ppgtt->scratch_page_dma_addr = dev_priv->gtt.scratch_page_dma;
-
 	ppgtt->clear_range(ppgtt, 0,
 			   ppgtt->num_pd_entries*I915_PPGTT_PT_ENTRIES);
 
-	ppgtt->pd_offset = (first_pd_entry_in_global_pt)*sizeof(gtt_pte_t);
+	ppgtt->pd_offset = first_pd_entry_in_global_pt * sizeof(gen6_gtt_pte_t);
 
 	return 0;
 
@@ -256,8 +305,13 @@ static int i915_gem_init_aliasing_ppgtt(struct drm_device *dev)
 		return -ENOMEM;
 
 	ppgtt->dev = dev;
+	ppgtt->scratch_page_dma_addr = dev_priv->gtt.scratch_page_dma;
+
+	if (INTEL_INFO(dev)->gen < 8)
+		ret = gen6_ppgtt_init(ppgtt);
+	else
+		BUG();
 
-	ret = gen6_ppgtt_init(ppgtt);
 	if (ret)
 		kfree(ppgtt);
 	else
@@ -275,6 +329,7 @@ void i915_gem_cleanup_aliasing_ppgtt(struct drm_device *dev)
 		return;
 
 	ppgtt->cleanup(ppgtt);
+	dev_priv->mm.aliasing_ppgtt = NULL;
 }
 
 void i915_ppgtt_bind_object(struct i915_hw_ppgtt *ppgtt,
@@ -294,64 +349,6 @@ void i915_ppgtt_unbind_object(struct i915_hw_ppgtt *ppgtt,
 			   obj->base.size >> PAGE_SHIFT);
 }
 
-void i915_gem_init_ppgtt(struct drm_device *dev)
-{
-	drm_i915_private_t *dev_priv = dev->dev_private;
-	uint32_t pd_offset;
-	struct intel_ring_buffer *ring;
-	struct i915_hw_ppgtt *ppgtt = dev_priv->mm.aliasing_ppgtt;
-	gtt_pte_t __iomem *pd_addr;
-	uint32_t pd_entry;
-	int i;
-
-	if (!dev_priv->mm.aliasing_ppgtt)
-		return;
-
-
-	pd_addr = (gtt_pte_t __iomem*)dev_priv->gtt.gsm + ppgtt->pd_offset/sizeof(gtt_pte_t);
-	for (i = 0; i < ppgtt->num_pd_entries; i++) {
-		dma_addr_t pt_addr;
-
-		pt_addr = ppgtt->pt_dma_addr[i];
-		pd_entry = GEN6_PDE_ADDR_ENCODE(pt_addr);
-		pd_entry |= GEN6_PDE_VALID;
-
-		writel(pd_entry, pd_addr + i);
-	}
-	readl(pd_addr);
-
-	pd_offset = ppgtt->pd_offset;
-	pd_offset /= 64; /* in cachelines, */
-	pd_offset <<= 16;
-
-	if (INTEL_INFO(dev)->gen == 6) {
-		uint32_t ecochk, gab_ctl, ecobits;
-
-		ecobits = I915_READ(GAC_ECO_BITS);
-		I915_WRITE(GAC_ECO_BITS, ecobits | ECOBITS_PPGTT_CACHE64B);
-
-		gab_ctl = I915_READ(GAB_CTL);
-		I915_WRITE(GAB_CTL, gab_ctl | GAB_CTL_CONT_AFTER_PAGEFAULT);
-
-		ecochk = I915_READ(GAM_ECOCHK);
-		I915_WRITE(GAM_ECOCHK, ecochk | ECOCHK_SNB_BIT |
-				       ECOCHK_PPGTT_CACHE64B);
-		I915_WRITE(GFX_MODE, _MASKED_BIT_ENABLE(GFX_PPGTT_ENABLE));
-	} else if (INTEL_INFO(dev)->gen >= 7) {
-		I915_WRITE(GAM_ECOCHK, ECOCHK_PPGTT_CACHE64B);
-		/* GFX_MODE is per-ring on gen7+ */
-	}
-
-	for_each_ring(ring, dev_priv, i) {
-		if (INTEL_INFO(dev)->gen >= 7)
-			I915_WRITE(RING_MODE_GEN7(ring),
-				   _MASKED_BIT_ENABLE(GFX_PPGTT_ENABLE));
-
-		I915_WRITE(RING_PP_DIR_DCLV(ring), PP_DIR_DCLV_2G);
-		I915_WRITE(RING_PP_DIR_BASE(ring), pd_offset);
-	}
-}
-
 extern int intel_iommu_gfx_mapped;
 /* Certain Gen5 chipsets require require idling the GPU before
  * unmapping anything from the GTT when VT-d is enabled.
@@ -432,21 +429,16 @@ static void gen6_ggtt_insert_entries(struct drm_device *dev,
 				     enum i915_cache_level level)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct scatterlist *sg = st->sgl;
-	gtt_pte_t __iomem *gtt_entries =
-		(gtt_pte_t __iomem *)dev_priv->gtt.gsm + first_entry;
-	int unused, i = 0;
-	unsigned int len, m = 0;
+	gen6_gtt_pte_t __iomem *gtt_entries =
+		(gen6_gtt_pte_t __iomem *)dev_priv->gtt.gsm + first_entry;
+	int i = 0;
+	struct sg_page_iter sg_iter;
 	dma_addr_t addr;
 
-	for_each_sg(st->sgl, sg, st->nents, unused) {
-		len = sg_dma_len(sg) >> PAGE_SHIFT;
-		for (m = 0; m < len; m++) {
-			addr = sg_dma_address(sg) + (m << PAGE_SHIFT);
-			iowrite32(gen6_pte_encode(dev, addr, level),
-				  &gtt_entries[i]);
-			i++;
-		}
+	for_each_sg_page(st->sgl, &sg_iter, st->nents, 0) {
+		addr = sg_page_iter_dma_address(&sg_iter);
+		iowrite32(gen6_pte_encode(dev, addr, level), &gtt_entries[i]);
+		i++;
 	}
 
 	/* XXX: This serves as a posting read to make sure that the PTE has
@@ -472,8 +464,8 @@ static void gen6_ggtt_clear_range(struct drm_device *dev,
 				  unsigned int num_entries)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	gtt_pte_t scratch_pte;
-	gtt_pte_t __iomem *gtt_base = (gtt_pte_t __iomem *) dev_priv->gtt.gsm + first_entry;
+	gen6_gtt_pte_t scratch_pte, __iomem *gtt_base =
+		(gen6_gtt_pte_t __iomem *) dev_priv->gtt.gsm + first_entry;
 	const int max_entries = gtt_total_entries(dev_priv->gtt) - first_entry;
 	int i;
 
@@ -647,9 +639,12 @@ void i915_gem_init_global_gtt(struct drm_device *dev)
 
 	if (intel_enable_ppgtt(dev) && HAS_ALIASING_PPGTT(dev)) {
 		int ret;
-		/* PPGTT pdes are stolen from global gtt ptes, so shrink the
-		 * aperture accordingly when using aliasing ppgtt. */
-		gtt_size -= I915_PPGTT_PD_ENTRIES*PAGE_SIZE;
+
+		if (INTEL_INFO(dev)->gen <= 7) {
+			/* PPGTT pdes are stolen from global gtt ptes, so shrink the
+			 * aperture accordingly when using aliasing ppgtt. */
+			gtt_size -= I915_PPGTT_PD_ENTRIES*PAGE_SIZE;
+		}
 
 		i915_gem_setup_global_gtt(dev, 0, mappable_size, gtt_size);
 
@@ -752,15 +747,17 @@ static int gen6_gmch_probe(struct drm_device *dev,
 	pci_read_config_word(dev->pdev, SNB_GMCH_CTRL, &snb_gmch_ctl);
 	gtt_size = gen6_get_total_gtt_size(snb_gmch_ctl);
 
-	if (IS_GEN7(dev))
+	if (IS_GEN7(dev) && !IS_VALLEYVIEW(dev))
 		*stolen = gen7_get_stolen_size(snb_gmch_ctl);
 	else
 		*stolen = gen6_get_stolen_size(snb_gmch_ctl);
 
-	*gtt_total = (gtt_size / sizeof(gtt_pte_t)) << PAGE_SHIFT;
+	*gtt_total = (gtt_size / sizeof(gen6_gtt_pte_t)) << PAGE_SHIFT;
+
+	/* For Modern GENs the PTEs and register space are split in the BAR */
+	gtt_bus_addr = pci_resource_start(dev->pdev, 0) +
+		(pci_resource_len(dev->pdev, 0) / 2);
 
-	/* For GEN6+ the PTEs for the ggtt live at 2MB + BAR0 */
-	gtt_bus_addr = pci_resource_start(dev->pdev, 0) + (2<<20);
 	dev_priv->gtt.gsm = ioremap_wc(gtt_bus_addr, gtt_size);
 	if (!dev_priv->gtt.gsm) {
 		DRM_ERROR("Failed to map the gtt page table\n");
@@ -817,7 +814,6 @@ int i915_gem_gtt_init(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct i915_gtt *gtt = &dev_priv->gtt;
-	unsigned long gtt_size;
 	int ret;
 
 	if (INTEL_INFO(dev)->gen <= 5) {
@@ -835,8 +831,6 @@ int i915_gem_gtt_init(struct drm_device *dev)
 	if (ret)
 		return ret;
 
-	gtt_size = (dev_priv->gtt.total >> PAGE_SHIFT) * sizeof(gtt_pte_t);
-
 	/* GMADR is the PCI mmio aperture into the global GTT. */
 	DRM_INFO("Memory usable by graphics device = %zdM\n",
 		 dev_priv->gtt.total >> 20);
diff --git a/drivers/gpu/drm/i915/i915_gem_stolen.c b/drivers/gpu/drm/i915/i915_gem_stolen.c
index 69d97cbac13c..130d1db27e28 100644
--- a/drivers/gpu/drm/i915/i915_gem_stolen.c
+++ b/drivers/gpu/drm/i915/i915_gem_stolen.c
@@ -312,6 +312,71 @@ i915_gem_object_create_stolen(struct drm_device *dev, u32 size)
 	return NULL;
 }
 
+struct drm_i915_gem_object *
+i915_gem_object_create_stolen_for_preallocated(struct drm_device *dev,
+					       u32 stolen_offset,
+					       u32 gtt_offset,
+					       u32 size)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_gem_object *obj;
+	struct drm_mm_node *stolen;
+
+	if (dev_priv->mm.stolen_base == 0)
+		return NULL;
+
+	DRM_DEBUG_KMS("creating preallocated stolen object: stolen_offset=%x, gtt_offset=%x, size=%x\n",
+			stolen_offset, gtt_offset, size);
+
+	/* KISS and expect everything to be page-aligned */
+	BUG_ON(stolen_offset & 4095);
+	BUG_ON(gtt_offset & 4095);
+	BUG_ON(size & 4095);
+
+	if (WARN_ON(size == 0))
+		return NULL;
+
+	stolen = drm_mm_create_block(&dev_priv->mm.stolen,
+				     stolen_offset, size,
+				     false);
+	if (stolen == NULL) {
+		DRM_DEBUG_KMS("failed to allocate stolen space\n");
+		return NULL;
+	}
+
+	obj = _i915_gem_object_create_stolen(dev, stolen);
+	if (obj == NULL) {
+		DRM_DEBUG_KMS("failed to allocate stolen object\n");
+		drm_mm_put_block(stolen);
+		return NULL;
+	}
+
+	/* To simplify the initialisation sequence between KMS and GTT,
+	 * we allow construction of the stolen object prior to
+	 * setting up the GTT space. The actual reservation will occur
+	 * later.
+	 */
+	if (drm_mm_initialized(&dev_priv->mm.gtt_space)) {
+		obj->gtt_space = drm_mm_create_block(&dev_priv->mm.gtt_space,
+						     gtt_offset, size,
+						     false);
+		if (obj->gtt_space == NULL) {
+			DRM_DEBUG_KMS("failed to allocate stolen GTT space\n");
+			drm_gem_object_unreference(&obj->base);
+			return NULL;
+		}
+	} else
+		obj->gtt_space = I915_GTT_RESERVED;
+
+	obj->gtt_offset = gtt_offset;
+	obj->has_global_gtt_mapping = 1;
+
+	list_add_tail(&obj->gtt_list, &dev_priv->mm.bound_list);
+	list_add_tail(&obj->mm_list, &dev_priv->mm.inactive_list);
+
+	return obj;
+}
+
 void
 i915_gem_object_release_stolen(struct drm_i915_gem_object *obj)
 {
diff --git a/drivers/gpu/drm/i915/i915_gem_tiling.c b/drivers/gpu/drm/i915/i915_gem_tiling.c
index abcba2f5a788..537545be69db 100644
--- a/drivers/gpu/drm/i915/i915_gem_tiling.c
+++ b/drivers/gpu/drm/i915/i915_gem_tiling.c
@@ -217,9 +217,12 @@ i915_tiling_ok(struct drm_device *dev, int stride, int size, int tiling_mode)
 		tile_width = 512;
 
 	/* check maximum stride & object size */
-	if (INTEL_INFO(dev)->gen >= 4) {
-		/* i965 stores the end address of the gtt mapping in the fence
-		 * reg, so dont bother to check the size */
+	/* i965+ stores the end address of the gtt mapping in the fence
+	 * reg, so dont bother to check the size */
+	if (INTEL_INFO(dev)->gen >= 7) {
+		if (stride / 128 > GEN7_FENCE_MAX_PITCH_VAL)
+			return false;
+	} else if (INTEL_INFO(dev)->gen >= 4) {
 		if (stride / 128 > I965_FENCE_MAX_PITCH_VAL)
 			return false;
 	} else {
@@ -235,6 +238,9 @@ i915_tiling_ok(struct drm_device *dev, int stride, int size, int tiling_mode)
 		}
 	}
 
+	if (stride < tile_width)
+		return false;
+
 	/* 965+ just needs multiples of tile width */
 	if (INTEL_INFO(dev)->gen >= 4) {
 		if (stride & (tile_width - 1))
@@ -243,9 +249,6 @@ i915_tiling_ok(struct drm_device *dev, int stride, int size, int tiling_mode)
 	}
 
 	/* Pre-965 needs power of two tile widths */
-	if (stride < tile_width)
-		return false;
-
 	if (stride & (stride - 1))
 		return false;
 
@@ -473,28 +476,29 @@ i915_gem_swizzle_page(struct page *page)
 void
 i915_gem_object_do_bit_17_swizzle(struct drm_i915_gem_object *obj)
 {
-	struct scatterlist *sg;
-	int page_count = obj->base.size >> PAGE_SHIFT;
+	struct sg_page_iter sg_iter;
 	int i;
 
 	if (obj->bit_17 == NULL)
 		return;
 
-	for_each_sg(obj->pages->sgl, sg, page_count, i) {
-		struct page *page = sg_page(sg);
+	i = 0;
+	for_each_sg_page(obj->pages->sgl, &sg_iter, obj->pages->nents, 0) {
+		struct page *page = sg_page_iter_page(&sg_iter);
 		char new_bit_17 = page_to_phys(page) >> 17;
 		if ((new_bit_17 & 0x1) !=
 		    (test_bit(i, obj->bit_17) != 0)) {
 			i915_gem_swizzle_page(page);
 			set_page_dirty(page);
 		}
+		i++;
 	}
 }
 
 void
 i915_gem_object_save_bit_17_swizzle(struct drm_i915_gem_object *obj)
 {
-	struct scatterlist *sg;
+	struct sg_page_iter sg_iter;
 	int page_count = obj->base.size >> PAGE_SHIFT;
 	int i;
 
@@ -508,11 +512,12 @@ i915_gem_object_save_bit_17_swizzle(struct drm_i915_gem_object *obj)
 		}
 	}
 
-	for_each_sg(obj->pages->sgl, sg, page_count, i) {
-		struct page *page = sg_page(sg);
-		if (page_to_phys(page) & (1 << 17))
+	i = 0;
+	for_each_sg_page(obj->pages->sgl, &sg_iter, obj->pages->nents, 0) {
+		if (page_to_phys(sg_page_iter_page(&sg_iter)) & (1 << 17))
 			__set_bit(i, obj->bit_17);
 		else
 			__clear_bit(i, obj->bit_17);
+		i++;
 	}
 }
diff --git a/drivers/gpu/drm/i915/i915_irq.c b/drivers/gpu/drm/i915/i915_irq.c
index 3c7bb0410b51..0aa2ef0d2ae0 100644
--- a/drivers/gpu/drm/i915/i915_irq.c
+++ b/drivers/gpu/drm/i915/i915_irq.c
@@ -36,6 +36,61 @@
 #include "i915_trace.h"
 #include "intel_drv.h"
 
+static const u32 hpd_ibx[] = {
+	[HPD_CRT] = SDE_CRT_HOTPLUG,
+	[HPD_SDVO_B] = SDE_SDVOB_HOTPLUG,
+	[HPD_PORT_B] = SDE_PORTB_HOTPLUG,
+	[HPD_PORT_C] = SDE_PORTC_HOTPLUG,
+	[HPD_PORT_D] = SDE_PORTD_HOTPLUG
+};
+
+static const u32 hpd_cpt[] = {
+	[HPD_CRT] = SDE_CRT_HOTPLUG_CPT,
+	[HPD_SDVO_B] = SDE_SDVOB_HOTPLUG_CPT,
+	[HPD_PORT_B] = SDE_PORTB_HOTPLUG_CPT,
+	[HPD_PORT_C] = SDE_PORTC_HOTPLUG_CPT,
+	[HPD_PORT_D] = SDE_PORTD_HOTPLUG_CPT
+};
+
+static const u32 hpd_mask_i915[] = {
+	[HPD_CRT] = CRT_HOTPLUG_INT_EN,
+	[HPD_SDVO_B] = SDVOB_HOTPLUG_INT_EN,
+	[HPD_SDVO_C] = SDVOC_HOTPLUG_INT_EN,
+	[HPD_PORT_B] = PORTB_HOTPLUG_INT_EN,
+	[HPD_PORT_C] = PORTC_HOTPLUG_INT_EN,
+	[HPD_PORT_D] = PORTD_HOTPLUG_INT_EN
+};
+
+static const u32 hpd_status_gen4[] = {
+	[HPD_CRT] = CRT_HOTPLUG_INT_STATUS,
+	[HPD_SDVO_B] = SDVOB_HOTPLUG_INT_STATUS_G4X,
+	[HPD_SDVO_C] = SDVOC_HOTPLUG_INT_STATUS_G4X,
+	[HPD_PORT_B] = PORTB_HOTPLUG_INT_STATUS,
+	[HPD_PORT_C] = PORTC_HOTPLUG_INT_STATUS,
+	[HPD_PORT_D] = PORTD_HOTPLUG_INT_STATUS
+};
+
+static const u32 hpd_status_i965[] = {
+	 [HPD_CRT] = CRT_HOTPLUG_INT_STATUS,
+	 [HPD_SDVO_B] = SDVOB_HOTPLUG_INT_STATUS_I965,
+	 [HPD_SDVO_C] = SDVOC_HOTPLUG_INT_STATUS_I965,
+	 [HPD_PORT_B] = PORTB_HOTPLUG_INT_STATUS,
+	 [HPD_PORT_C] = PORTC_HOTPLUG_INT_STATUS,
+	 [HPD_PORT_D] = PORTD_HOTPLUG_INT_STATUS
+};
+
+static const u32 hpd_status_i915[] = { /* i915 and valleyview are the same */
+	[HPD_CRT] = CRT_HOTPLUG_INT_STATUS,
+	[HPD_SDVO_B] = SDVOB_HOTPLUG_INT_STATUS_I915,
+	[HPD_SDVO_C] = SDVOC_HOTPLUG_INT_STATUS_I915,
+	[HPD_PORT_B] = PORTB_HOTPLUG_INT_STATUS,
+	[HPD_PORT_C] = PORTC_HOTPLUG_INT_STATUS,
+	[HPD_PORT_D] = PORTD_HOTPLUG_INT_STATUS
+};
+
+static void ibx_hpd_irq_setup(struct drm_device *dev);
+static void i915_hpd_irq_setup(struct drm_device *dev);
+
 /* For display hotplug interrupt */
 static void
 ironlake_enable_display_irq(drm_i915_private_t *dev_priv, u32 mask)
@@ -47,7 +102,7 @@ ironlake_enable_display_irq(drm_i915_private_t *dev_priv, u32 mask)
 	}
 }
 
-static inline void
+static void
 ironlake_disable_display_irq(drm_i915_private_t *dev_priv, u32 mask)
 {
 	if ((dev_priv->irq_mask & mask) != mask) {
@@ -60,26 +115,30 @@ ironlake_disable_display_irq(drm_i915_private_t *dev_priv, u32 mask)
 void
 i915_enable_pipestat(drm_i915_private_t *dev_priv, int pipe, u32 mask)
 {
-	if ((dev_priv->pipestat[pipe] & mask) != mask) {
-		u32 reg = PIPESTAT(pipe);
+	u32 reg = PIPESTAT(pipe);
+	u32 pipestat = I915_READ(reg) & 0x7fff0000;
 
-		dev_priv->pipestat[pipe] |= mask;
-		/* Enable the interrupt, clear any pending status */
-		I915_WRITE(reg, dev_priv->pipestat[pipe] | (mask >> 16));
-		POSTING_READ(reg);
-	}
+	if ((pipestat & mask) == mask)
+		return;
+
+	/* Enable the interrupt, clear any pending status */
+	pipestat |= mask | (mask >> 16);
+	I915_WRITE(reg, pipestat);
+	POSTING_READ(reg);
 }
 
 void
 i915_disable_pipestat(drm_i915_private_t *dev_priv, int pipe, u32 mask)
 {
-	if ((dev_priv->pipestat[pipe] & mask) != 0) {
-		u32 reg = PIPESTAT(pipe);
+	u32 reg = PIPESTAT(pipe);
+	u32 pipestat = I915_READ(reg) & 0x7fff0000;
 
-		dev_priv->pipestat[pipe] &= ~mask;
-		I915_WRITE(reg, dev_priv->pipestat[pipe]);
-		POSTING_READ(reg);
-	}
+	if ((pipestat & mask) == 0)
+		return;
+
+	pipestat &= ~mask;
+	I915_WRITE(reg, pipestat);
+	POSTING_READ(reg);
 }
 
 /**
@@ -250,10 +309,9 @@ static int i915_get_vblank_timestamp(struct drm_device *dev, int pipe,
 			      struct timeval *vblank_time,
 			      unsigned flags)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct drm_crtc *crtc;
 
-	if (pipe < 0 || pipe >= dev_priv->num_pipe) {
+	if (pipe < 0 || pipe >= INTEL_INFO(dev)->num_pipes) {
 		DRM_ERROR("Invalid crtc %d\n", pipe);
 		return -EINVAL;
 	}
@@ -279,13 +337,19 @@ static int i915_get_vblank_timestamp(struct drm_device *dev, int pipe,
 /*
  * Handle hotplug events outside the interrupt handler proper.
  */
+#define I915_REENABLE_HOTPLUG_DELAY (2*60*1000)
+
 static void i915_hotplug_work_func(struct work_struct *work)
 {
 	drm_i915_private_t *dev_priv = container_of(work, drm_i915_private_t,
 						    hotplug_work);
 	struct drm_device *dev = dev_priv->dev;
 	struct drm_mode_config *mode_config = &dev->mode_config;
-	struct intel_encoder *encoder;
+	struct intel_connector *intel_connector;
+	struct intel_encoder *intel_encoder;
+	struct drm_connector *connector;
+	unsigned long irqflags;
+	bool hpd_disabled = false;
 
 	/* HPD irq before everything is fully set up. */
 	if (!dev_priv->enable_hotplug_processing)
@@ -294,9 +358,36 @@ static void i915_hotplug_work_func(struct work_struct *work)
 	mutex_lock(&mode_config->mutex);
 	DRM_DEBUG_KMS("running encoder hotplug functions\n");
 
-	list_for_each_entry(encoder, &mode_config->encoder_list, base.head)
-		if (encoder->hot_plug)
-			encoder->hot_plug(encoder);
+	spin_lock_irqsave(&dev_priv->irq_lock, irqflags);
+	list_for_each_entry(connector, &mode_config->connector_list, head) {
+		intel_connector = to_intel_connector(connector);
+		intel_encoder = intel_connector->encoder;
+		if (intel_encoder->hpd_pin > HPD_NONE &&
+		    dev_priv->hpd_stats[intel_encoder->hpd_pin].hpd_mark == HPD_MARK_DISABLED &&
+		    connector->polled == DRM_CONNECTOR_POLL_HPD) {
+			DRM_INFO("HPD interrupt storm detected on connector %s: "
+				 "switching from hotplug detection to polling\n",
+				drm_get_connector_name(connector));
+			dev_priv->hpd_stats[intel_encoder->hpd_pin].hpd_mark = HPD_DISABLED;
+			connector->polled = DRM_CONNECTOR_POLL_CONNECT
+				| DRM_CONNECTOR_POLL_DISCONNECT;
+			hpd_disabled = true;
+		}
+	}
+	 /* if there were no outputs to poll, poll was disabled,
+	  * therefore make sure it's enabled when disabling HPD on
+	  * some connectors */
+	if (hpd_disabled) {
+		drm_kms_helper_poll_enable(dev);
+		mod_timer(&dev_priv->hotplug_reenable_timer,
+			  jiffies + msecs_to_jiffies(I915_REENABLE_HOTPLUG_DELAY));
+	}
+
+	spin_unlock_irqrestore(&dev_priv->irq_lock, irqflags);
+
+	list_for_each_entry(intel_encoder, &mode_config->encoder_list, base.head)
+		if (intel_encoder->hot_plug)
+			intel_encoder->hot_plug(intel_encoder);
 
 	mutex_unlock(&mode_config->mutex);
 
@@ -525,6 +616,45 @@ static void gen6_queue_rps_work(struct drm_i915_private *dev_priv,
 	queue_work(dev_priv->wq, &dev_priv->rps.work);
 }
 
+#define HPD_STORM_DETECT_PERIOD 1000
+#define HPD_STORM_THRESHOLD 5
+
+static inline bool hotplug_irq_storm_detect(struct drm_device *dev,
+					    u32 hotplug_trigger,
+					    const u32 *hpd)
+{
+	drm_i915_private_t *dev_priv = dev->dev_private;
+	unsigned long irqflags;
+	int i;
+	bool ret = false;
+
+	spin_lock_irqsave(&dev_priv->irq_lock, irqflags);
+
+	for (i = 1; i < HPD_NUM_PINS; i++) {
+
+		if (!(hpd[i] & hotplug_trigger) ||
+		    dev_priv->hpd_stats[i].hpd_mark != HPD_ENABLED)
+			continue;
+
+		if (!time_in_range(jiffies, dev_priv->hpd_stats[i].hpd_last_jiffies,
+				   dev_priv->hpd_stats[i].hpd_last_jiffies
+				   + msecs_to_jiffies(HPD_STORM_DETECT_PERIOD))) {
+			dev_priv->hpd_stats[i].hpd_last_jiffies = jiffies;
+			dev_priv->hpd_stats[i].hpd_cnt = 0;
+		} else if (dev_priv->hpd_stats[i].hpd_cnt > HPD_STORM_THRESHOLD) {
+			dev_priv->hpd_stats[i].hpd_mark = HPD_MARK_DISABLED;
+			DRM_DEBUG_KMS("HPD interrupt storm detected on PIN %d\n", i);
+			ret = true;
+		} else {
+			dev_priv->hpd_stats[i].hpd_cnt++;
+		}
+	}
+
+	spin_unlock_irqrestore(&dev_priv->irq_lock, irqflags);
+
+	return ret;
+}
+
 static void gmbus_irq_handler(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = (drm_i915_private_t *) dev->dev_private;
@@ -593,13 +723,16 @@ static irqreturn_t valleyview_irq_handler(int irq, void *arg)
 		/* Consume port.  Then clear IIR or we'll miss events */
 		if (iir & I915_DISPLAY_PORT_INTERRUPT) {
 			u32 hotplug_status = I915_READ(PORT_HOTPLUG_STAT);
+			u32 hotplug_trigger = hotplug_status & HOTPLUG_INT_STATUS_I915;
 
 			DRM_DEBUG_DRIVER("hotplug event received, stat 0x%08x\n",
 					 hotplug_status);
-			if (hotplug_status & dev_priv->hotplug_supported_mask)
+			if (hotplug_trigger) {
+				if (hotplug_irq_storm_detect(dev, hotplug_trigger, hpd_status_i915))
+					i915_hpd_irq_setup(dev);
 				queue_work(dev_priv->wq,
 					   &dev_priv->hotplug_work);
-
+			}
 			I915_WRITE(PORT_HOTPLUG_STAT, hotplug_status);
 			I915_READ(PORT_HOTPLUG_STAT);
 		}
@@ -623,10 +756,13 @@ static void ibx_irq_handler(struct drm_device *dev, u32 pch_iir)
 {
 	drm_i915_private_t *dev_priv = (drm_i915_private_t *) dev->dev_private;
 	int pipe;
+	u32 hotplug_trigger = pch_iir & SDE_HOTPLUG_MASK;
 
-	if (pch_iir & SDE_HOTPLUG_MASK)
+	if (hotplug_trigger) {
+		if (hotplug_irq_storm_detect(dev, hotplug_trigger, hpd_ibx))
+			ibx_hpd_irq_setup(dev);
 		queue_work(dev_priv->wq, &dev_priv->hotplug_work);
-
+	}
 	if (pch_iir & SDE_AUDIO_POWER_MASK)
 		DRM_DEBUG_DRIVER("PCH audio power change on port %d\n",
 				 (pch_iir & SDE_AUDIO_POWER_MASK) >>
@@ -669,10 +805,13 @@ static void cpt_irq_handler(struct drm_device *dev, u32 pch_iir)
 {
 	drm_i915_private_t *dev_priv = (drm_i915_private_t *) dev->dev_private;
 	int pipe;
+	u32 hotplug_trigger = pch_iir & SDE_HOTPLUG_MASK_CPT;
 
-	if (pch_iir & SDE_HOTPLUG_MASK_CPT)
+	if (hotplug_trigger) {
+		if (hotplug_irq_storm_detect(dev, hotplug_trigger, hpd_cpt))
+			ibx_hpd_irq_setup(dev);
 		queue_work(dev_priv->wq, &dev_priv->hotplug_work);
-
+	}
 	if (pch_iir & SDE_AUDIO_POWER_MASK_CPT)
 		DRM_DEBUG_DRIVER("PCH audio power change on port %d\n",
 				 (pch_iir & SDE_AUDIO_POWER_MASK_CPT) >>
@@ -701,7 +840,7 @@ static irqreturn_t ivybridge_irq_handler(int irq, void *arg)
 {
 	struct drm_device *dev = (struct drm_device *) arg;
 	drm_i915_private_t *dev_priv = (drm_i915_private_t *) dev->dev_private;
-	u32 de_iir, gt_iir, de_ier, pm_iir, sde_ier;
+	u32 de_iir, gt_iir, de_ier, pm_iir, sde_ier = 0;
 	irqreturn_t ret = IRQ_NONE;
 	int i;
 
@@ -716,9 +855,11 @@ static irqreturn_t ivybridge_irq_handler(int irq, void *arg)
 	 * able to process them after we restore SDEIER (as soon as we restore
 	 * it, we'll get an interrupt if SDEIIR still has something to process
 	 * due to its back queue). */
-	sde_ier = I915_READ(SDEIER);
-	I915_WRITE(SDEIER, 0);
-	POSTING_READ(SDEIER);
+	if (!HAS_PCH_NOP(dev)) {
+		sde_ier = I915_READ(SDEIER);
+		I915_WRITE(SDEIER, 0);
+		POSTING_READ(SDEIER);
+	}
 
 	gt_iir = I915_READ(GTIIR);
 	if (gt_iir) {
@@ -745,7 +886,7 @@ static irqreturn_t ivybridge_irq_handler(int irq, void *arg)
 		}
 
 		/* check event from PCH */
-		if (de_iir & DE_PCH_EVENT_IVB) {
+		if (!HAS_PCH_NOP(dev) && (de_iir & DE_PCH_EVENT_IVB)) {
 			u32 pch_iir = I915_READ(SDEIIR);
 
 			cpt_irq_handler(dev, pch_iir);
@@ -768,8 +909,10 @@ static irqreturn_t ivybridge_irq_handler(int irq, void *arg)
 
 	I915_WRITE(DEIER, de_ier);
 	POSTING_READ(DEIER);
-	I915_WRITE(SDEIER, sde_ier);
-	POSTING_READ(SDEIER);
+	if (!HAS_PCH_NOP(dev)) {
+		I915_WRITE(SDEIER, sde_ier);
+		POSTING_READ(SDEIER);
+	}
 
 	return ret;
 }
@@ -937,6 +1080,8 @@ static void i915_error_work_func(struct work_struct *work)
 		for_each_ring(ring, dev_priv, i)
 			wake_up_all(&ring->irq_queue);
 
+		intel_display_handle_reset(dev);
+
 		wake_up_all(&dev_priv->gpu_error.reset_queue);
 	}
 }
@@ -972,24 +1117,23 @@ static void i915_get_extra_instdone(struct drm_device *dev,
 
 #ifdef CONFIG_DEBUG_FS
 static struct drm_i915_error_object *
-i915_error_object_create(struct drm_i915_private *dev_priv,
-			 struct drm_i915_gem_object *src)
+i915_error_object_create_sized(struct drm_i915_private *dev_priv,
+			       struct drm_i915_gem_object *src,
+			       const int num_pages)
 {
 	struct drm_i915_error_object *dst;
-	int i, count;
+	int i;
 	u32 reloc_offset;
 
 	if (src == NULL || src->pages == NULL)
 		return NULL;
 
-	count = src->base.size / PAGE_SIZE;
-
-	dst = kmalloc(sizeof(*dst) + count * sizeof(u32 *), GFP_ATOMIC);
+	dst = kmalloc(sizeof(*dst) + num_pages * sizeof(u32 *), GFP_ATOMIC);
 	if (dst == NULL)
 		return NULL;
 
 	reloc_offset = src->gtt_offset;
-	for (i = 0; i < count; i++) {
+	for (i = 0; i < num_pages; i++) {
 		unsigned long flags;
 		void *d;
 
@@ -1039,7 +1183,7 @@ i915_error_object_create(struct drm_i915_private *dev_priv,
 
 		reloc_offset += PAGE_SIZE;
 	}
-	dst->page_count = count;
+	dst->page_count = num_pages;
 	dst->gtt_offset = src->gtt_offset;
 
 	return dst;
@@ -1050,6 +1194,9 @@ unwind:
 	kfree(dst);
 	return NULL;
 }
+#define i915_error_object_create(dev_priv, src) \
+	i915_error_object_create_sized((dev_priv), (src), \
+				       (src)->base.size>>PAGE_SHIFT)
 
 static void
 i915_error_object_free(struct drm_i915_error_object *obj)
@@ -1148,7 +1295,7 @@ static void i915_gem_record_fences(struct drm_device *dev,
 	switch (INTEL_INFO(dev)->gen) {
 	case 7:
 	case 6:
-		for (i = 0; i < 16; i++)
+		for (i = 0; i < dev_priv->num_fence_regs; i++)
 			error->fence[i] = I915_READ64(FENCE_REG_SANDYBRIDGE_0 + (i * 8));
 		break;
 	case 5:
@@ -1256,6 +1403,26 @@ static void i915_record_ring_state(struct drm_device *dev,
 	error->cpu_ring_tail[ring->id] = ring->tail;
 }
 
+
+static void i915_gem_record_active_context(struct intel_ring_buffer *ring,
+					   struct drm_i915_error_state *error,
+					   struct drm_i915_error_ring *ering)
+{
+	struct drm_i915_private *dev_priv = ring->dev->dev_private;
+	struct drm_i915_gem_object *obj;
+
+	/* Currently render ring is the only HW context user */
+	if (ring->id != RCS || !error->ccid)
+		return;
+
+	list_for_each_entry(obj, &dev_priv->mm.bound_list, gtt_list) {
+		if ((error->ccid & PAGE_MASK) == obj->gtt_offset) {
+			ering->ctx = i915_error_object_create_sized(dev_priv,
+								    obj, 1);
+		}
+	}
+}
+
 static void i915_gem_record_rings(struct drm_device *dev,
 				  struct drm_i915_error_state *error)
 {
@@ -1273,6 +1440,9 @@ static void i915_gem_record_rings(struct drm_device *dev,
 		error->ring[i].ringbuffer =
 			i915_error_object_create(dev_priv, ring->obj);
 
+
+		i915_gem_record_active_context(ring, error, &error->ring[i]);
+
 		count = 0;
 		list_for_each_entry(request, &ring->request_list, list)
 			count++;
@@ -1328,14 +1498,15 @@ static void i915_capture_error_state(struct drm_device *dev)
 		return;
 	}
 
-	DRM_INFO("capturing error event; look for more information in"
+	DRM_INFO("capturing error event; look for more information in "
 		 "/sys/kernel/debug/dri/%d/i915_error_state\n",
 		 dev->primary->index);
 
 	kref_init(&error->ref);
 	error->eir = I915_READ(EIR);
 	error->pgtbl_er = I915_READ(PGTBL_ER);
-	error->ccid = I915_READ(CCID);
+	if (HAS_HW_CONTEXTS(dev))
+		error->ccid = I915_READ(CCID);
 
 	if (HAS_PCH_SPLIT(dev))
 		error->ier = I915_READ(DEIER) | I915_READ(GTIER);
@@ -1356,8 +1527,9 @@ static void i915_capture_error_state(struct drm_device *dev)
 	else if (INTEL_INFO(dev)->gen == 6)
 		error->forcewake = I915_READ(FORCEWAKE);
 
-	for_each_pipe(pipe)
-		error->pipestat[pipe] = I915_READ(PIPESTAT(pipe));
+	if (!HAS_PCH_SPLIT(dev))
+		for_each_pipe(pipe)
+			error->pipestat[pipe] = I915_READ(PIPESTAT(pipe));
 
 	if (INTEL_INFO(dev)->gen >= 6) {
 		error->error = I915_READ(ERROR_GEN6);
@@ -1567,7 +1739,7 @@ void i915_handle_error(struct drm_device *dev, bool wedged)
 	queue_work(dev_priv->wq, &dev_priv->gpu_error.work);
 }
 
-static void i915_pageflip_stall_check(struct drm_device *dev, int pipe)
+static void __always_unused i915_pageflip_stall_check(struct drm_device *dev, int pipe)
 {
 	drm_i915_private_t *dev_priv = dev->dev_private;
 	struct drm_crtc *crtc = dev_priv->pipe_to_crtc_mapping[pipe];
@@ -1777,6 +1949,37 @@ static bool i915_hangcheck_ring_idle(struct intel_ring_buffer *ring, bool *err)
 	return false;
 }
 
+static bool semaphore_passed(struct intel_ring_buffer *ring)
+{
+	struct drm_i915_private *dev_priv = ring->dev->dev_private;
+	u32 acthd = intel_ring_get_active_head(ring) & HEAD_ADDR;
+	struct intel_ring_buffer *signaller;
+	u32 cmd, ipehr, acthd_min;
+
+	ipehr = I915_READ(RING_IPEHR(ring->mmio_base));
+	if ((ipehr & ~(0x3 << 16)) !=
+	    (MI_SEMAPHORE_MBOX | MI_SEMAPHORE_COMPARE | MI_SEMAPHORE_REGISTER))
+		return false;
+
+	/* ACTHD is likely pointing to the dword after the actual command,
+	 * so scan backwards until we find the MBOX.
+	 */
+	acthd_min = max((int)acthd - 3 * 4, 0);
+	do {
+		cmd = ioread32(ring->virtual_start + acthd);
+		if (cmd == ipehr)
+			break;
+
+		acthd -= 4;
+		if (acthd < acthd_min)
+			return false;
+	} while (1);
+
+	signaller = &dev_priv->ring[(ring->id + (((ipehr >> 17) & 1) + 1)) % 3];
+	return i915_seqno_passed(signaller->get_seqno(signaller, false),
+				 ioread32(ring->virtual_start+acthd+4)+1);
+}
+
 static bool kick_ring(struct intel_ring_buffer *ring)
 {
 	struct drm_device *dev = ring->dev;
@@ -1788,6 +1991,15 @@ static bool kick_ring(struct intel_ring_buffer *ring)
 		I915_WRITE_CTL(ring, tmp);
 		return true;
 	}
+
+	if (INTEL_INFO(dev)->gen >= 6 &&
+	    tmp & RING_WAIT_SEMAPHORE &&
+	    semaphore_passed(ring)) {
+		DRM_ERROR("Kicking stuck semaphore on %s\n",
+			  ring->name);
+		I915_WRITE_CTL(ring, tmp);
+		return true;
+	}
 	return false;
 }
 
@@ -1901,9 +2113,18 @@ static void ironlake_irq_preinstall(struct drm_device *dev)
 	I915_WRITE(GTIER, 0x0);
 	POSTING_READ(GTIER);
 
+	if (HAS_PCH_NOP(dev))
+		return;
+
 	/* south display irq */
 	I915_WRITE(SDEIMR, 0xffffffff);
-	I915_WRITE(SDEIER, 0x0);
+	/*
+	 * SDEIER is also touched by the interrupt handler to work around missed
+	 * PCH interrupts. Hence we can't update it after the interrupt handler
+	 * is enabled - instead we unconditionally enable all PCH interrupt
+	 * sources here, but then only unmask them as needed with SDEIMR.
+	 */
+	I915_WRITE(SDEIER, 0xffffffff);
 	POSTING_READ(SDEIER);
 }
 
@@ -1939,18 +2160,34 @@ static void valleyview_irq_preinstall(struct drm_device *dev)
 	POSTING_READ(VLV_IER);
 }
 
-/*
- * Enable digital hotplug on the PCH, and configure the DP short pulse
- * duration to 2ms (which is the minimum in the Display Port spec)
- *
- * This register is the same on all known PCH chips.
- */
-
-static void ibx_enable_hotplug(struct drm_device *dev)
+static void ibx_hpd_irq_setup(struct drm_device *dev)
 {
 	drm_i915_private_t *dev_priv = (drm_i915_private_t *) dev->dev_private;
-	u32	hotplug;
+	struct drm_mode_config *mode_config = &dev->mode_config;
+	struct intel_encoder *intel_encoder;
+	u32 mask = ~I915_READ(SDEIMR);
+	u32 hotplug;
+
+	if (HAS_PCH_IBX(dev)) {
+		mask &= ~SDE_HOTPLUG_MASK;
+		list_for_each_entry(intel_encoder, &mode_config->encoder_list, base.head)
+			if (dev_priv->hpd_stats[intel_encoder->hpd_pin].hpd_mark == HPD_ENABLED)
+				mask |= hpd_ibx[intel_encoder->hpd_pin];
+	} else {
+		mask &= ~SDE_HOTPLUG_MASK_CPT;
+		list_for_each_entry(intel_encoder, &mode_config->encoder_list, base.head)
+			if (dev_priv->hpd_stats[intel_encoder->hpd_pin].hpd_mark == HPD_ENABLED)
+				mask |= hpd_cpt[intel_encoder->hpd_pin];
+	}
 
+	I915_WRITE(SDEIMR, ~mask);
+
+	/*
+	 * Enable digital hotplug on the PCH, and configure the DP short pulse
+	 * duration to 2ms (which is the minimum in the Display Port spec)
+	 *
+	 * This register is the same on all known PCH chips.
+	 */
 	hotplug = I915_READ(PCH_PORT_HOTPLUG);
 	hotplug &= ~(PORTD_PULSE_DURATION_MASK|PORTC_PULSE_DURATION_MASK|PORTB_PULSE_DURATION_MASK);
 	hotplug |= PORTD_HOTPLUG_ENABLE | PORTD_PULSE_DURATION_2ms;
@@ -1965,20 +2202,15 @@ static void ibx_irq_postinstall(struct drm_device *dev)
 	u32 mask;
 
 	if (HAS_PCH_IBX(dev))
-		mask = SDE_HOTPLUG_MASK |
-		       SDE_GMBUS |
-		       SDE_AUX_MASK;
+		mask = SDE_GMBUS | SDE_AUX_MASK;
 	else
-		mask = SDE_HOTPLUG_MASK_CPT |
-		       SDE_GMBUS_CPT |
-		       SDE_AUX_MASK_CPT;
+		mask = SDE_GMBUS_CPT | SDE_AUX_MASK_CPT;
+
+	if (HAS_PCH_NOP(dev))
+		return;
 
 	I915_WRITE(SDEIIR, I915_READ(SDEIIR));
 	I915_WRITE(SDEIMR, ~mask);
-	I915_WRITE(SDEIER, mask);
-	POSTING_READ(SDEIER);
-
-	ibx_enable_hotplug(dev);
 }
 
 static int ironlake_irq_postinstall(struct drm_device *dev)
@@ -2089,9 +2321,6 @@ static int valleyview_irq_postinstall(struct drm_device *dev)
 		I915_DISPLAY_PIPE_A_VBLANK_INTERRUPT |
 		I915_DISPLAY_PIPE_B_VBLANK_INTERRUPT;
 
-	dev_priv->pipestat[0] = 0;
-	dev_priv->pipestat[1] = 0;
-
 	/* Hack for broken MSIs on VLV */
 	pci_write_config_dword(dev_priv->dev->pdev, 0x94, 0xfee00000);
 	pci_read_config_word(dev->pdev, 0x98, &msid);
@@ -2135,30 +2364,6 @@ static int valleyview_irq_postinstall(struct drm_device *dev)
 	return 0;
 }
 
-static void valleyview_hpd_irq_setup(struct drm_device *dev)
-{
-	drm_i915_private_t *dev_priv = (drm_i915_private_t *) dev->dev_private;
-	u32 hotplug_en = I915_READ(PORT_HOTPLUG_EN);
-
-	/* Note HDMI and DP share bits */
-	if (dev_priv->hotplug_supported_mask & PORTB_HOTPLUG_INT_STATUS)
-		hotplug_en |= PORTB_HOTPLUG_INT_EN;
-	if (dev_priv->hotplug_supported_mask & PORTC_HOTPLUG_INT_STATUS)
-		hotplug_en |= PORTC_HOTPLUG_INT_EN;
-	if (dev_priv->hotplug_supported_mask & PORTD_HOTPLUG_INT_STATUS)
-		hotplug_en |= PORTD_HOTPLUG_INT_EN;
-	if (dev_priv->hotplug_supported_mask & SDVOC_HOTPLUG_INT_STATUS_I915)
-		hotplug_en |= SDVOC_HOTPLUG_INT_EN;
-	if (dev_priv->hotplug_supported_mask & SDVOB_HOTPLUG_INT_STATUS_I915)
-		hotplug_en |= SDVOB_HOTPLUG_INT_EN;
-	if (dev_priv->hotplug_supported_mask & CRT_HOTPLUG_INT_STATUS) {
-		hotplug_en |= CRT_HOTPLUG_INT_EN;
-		hotplug_en |= CRT_HOTPLUG_VOLTAGE_COMPARE_50;
-	}
-
-	I915_WRITE(PORT_HOTPLUG_EN, hotplug_en);
-}
-
 static void valleyview_irq_uninstall(struct drm_device *dev)
 {
 	drm_i915_private_t *dev_priv = (drm_i915_private_t *) dev->dev_private;
@@ -2167,6 +2372,8 @@ static void valleyview_irq_uninstall(struct drm_device *dev)
 	if (!dev_priv)
 		return;
 
+	del_timer_sync(&dev_priv->hotplug_reenable_timer);
+
 	for_each_pipe(pipe)
 		I915_WRITE(PIPESTAT(pipe), 0xffff);
 
@@ -2188,6 +2395,8 @@ static void ironlake_irq_uninstall(struct drm_device *dev)
 	if (!dev_priv)
 		return;
 
+	del_timer_sync(&dev_priv->hotplug_reenable_timer);
+
 	I915_WRITE(HWSTAM, 0xffffffff);
 
 	I915_WRITE(DEIMR, 0xffffffff);
@@ -2198,6 +2407,9 @@ static void ironlake_irq_uninstall(struct drm_device *dev)
 	I915_WRITE(GTIER, 0x0);
 	I915_WRITE(GTIIR, I915_READ(GTIIR));
 
+	if (HAS_PCH_NOP(dev))
+		return;
+
 	I915_WRITE(SDEIMR, 0xffffffff);
 	I915_WRITE(SDEIER, 0x0);
 	I915_WRITE(SDEIIR, I915_READ(SDEIIR));
@@ -2221,9 +2433,6 @@ static int i8xx_irq_postinstall(struct drm_device *dev)
 {
 	drm_i915_private_t *dev_priv = (drm_i915_private_t *) dev->dev_private;
 
-	dev_priv->pipestat[0] = 0;
-	dev_priv->pipestat[1] = 0;
-
 	I915_WRITE16(EMR,
 		     ~(I915_ERROR_PAGE_TABLE | I915_ERROR_MEMORY_REFRESH));
 
@@ -2246,6 +2455,37 @@ static int i8xx_irq_postinstall(struct drm_device *dev)
 	return 0;
 }
 
+/*
+ * Returns true when a page flip has completed.
+ */
+static bool i8xx_handle_vblank(struct drm_device *dev,
+			       int pipe, u16 iir)
+{
+	drm_i915_private_t *dev_priv = dev->dev_private;
+	u16 flip_pending = DISPLAY_PLANE_FLIP_PENDING(pipe);
+
+	if (!drm_handle_vblank(dev, pipe))
+		return false;
+
+	if ((iir & flip_pending) == 0)
+		return false;
+
+	intel_prepare_page_flip(dev, pipe);
+
+	/* We detect FlipDone by looking for the change in PendingFlip from '1'
+	 * to '0' on the following vblank, i.e. IIR has the Pendingflip
+	 * asserted following the MI_DISPLAY_FLIP, but ISR is deasserted, hence
+	 * the flip is completed (no longer pending). Since this doesn't raise
+	 * an interrupt per se, we watch for the change at vblank.
+	 */
+	if (I915_READ16(ISR) & flip_pending)
+		return false;
+
+	intel_finish_page_flip(dev, pipe);
+
+	return true;
+}
+
 static irqreturn_t i8xx_irq_handler(int irq, void *arg)
 {
 	struct drm_device *dev = (struct drm_device *) arg;
@@ -2301,22 +2541,12 @@ static irqreturn_t i8xx_irq_handler(int irq, void *arg)
 			notify_ring(dev, &dev_priv->ring[RCS]);
 
 		if (pipe_stats[0] & PIPE_VBLANK_INTERRUPT_STATUS &&
-		    drm_handle_vblank(dev, 0)) {
-			if (iir & I915_DISPLAY_PLANE_A_FLIP_PENDING_INTERRUPT) {
-				intel_prepare_page_flip(dev, 0);
-				intel_finish_page_flip(dev, 0);
-				flip_mask &= ~I915_DISPLAY_PLANE_A_FLIP_PENDING_INTERRUPT;
-			}
-		}
+		    i8xx_handle_vblank(dev, 0, iir))
+			flip_mask &= ~DISPLAY_PLANE_FLIP_PENDING(0);
 
 		if (pipe_stats[1] & PIPE_VBLANK_INTERRUPT_STATUS &&
-		    drm_handle_vblank(dev, 1)) {
-			if (iir & I915_DISPLAY_PLANE_B_FLIP_PENDING_INTERRUPT) {
-				intel_prepare_page_flip(dev, 1);
-				intel_finish_page_flip(dev, 1);
-				flip_mask &= ~I915_DISPLAY_PLANE_B_FLIP_PENDING_INTERRUPT;
-			}
-		}
+		    i8xx_handle_vblank(dev, 1, iir))
+			flip_mask &= ~DISPLAY_PLANE_FLIP_PENDING(1);
 
 		iir = new_iir;
 	}
@@ -2364,9 +2594,6 @@ static int i915_irq_postinstall(struct drm_device *dev)
 	drm_i915_private_t *dev_priv = (drm_i915_private_t *) dev->dev_private;
 	u32 enable_mask;
 
-	dev_priv->pipestat[0] = 0;
-	dev_priv->pipestat[1] = 0;
-
 	I915_WRITE(EMR, ~(I915_ERROR_PAGE_TABLE | I915_ERROR_MEMORY_REFRESH));
 
 	/* Unmask the interrupts that we always want on. */
@@ -2404,33 +2631,35 @@ static int i915_irq_postinstall(struct drm_device *dev)
 	return 0;
 }
 
-static void i915_hpd_irq_setup(struct drm_device *dev)
+/*
+ * Returns true when a page flip has completed.
+ */
+static bool i915_handle_vblank(struct drm_device *dev,
+			       int plane, int pipe, u32 iir)
 {
-	drm_i915_private_t *dev_priv = (drm_i915_private_t *) dev->dev_private;
-	u32 hotplug_en;
+	drm_i915_private_t *dev_priv = dev->dev_private;
+	u32 flip_pending = DISPLAY_PLANE_FLIP_PENDING(plane);
 
-	if (I915_HAS_HOTPLUG(dev)) {
-		hotplug_en = I915_READ(PORT_HOTPLUG_EN);
+	if (!drm_handle_vblank(dev, pipe))
+		return false;
 
-		if (dev_priv->hotplug_supported_mask & PORTB_HOTPLUG_INT_STATUS)
-			hotplug_en |= PORTB_HOTPLUG_INT_EN;
-		if (dev_priv->hotplug_supported_mask & PORTC_HOTPLUG_INT_STATUS)
-			hotplug_en |= PORTC_HOTPLUG_INT_EN;
-		if (dev_priv->hotplug_supported_mask & PORTD_HOTPLUG_INT_STATUS)
-			hotplug_en |= PORTD_HOTPLUG_INT_EN;
-		if (dev_priv->hotplug_supported_mask & SDVOC_HOTPLUG_INT_STATUS_I915)
-			hotplug_en |= SDVOC_HOTPLUG_INT_EN;
-		if (dev_priv->hotplug_supported_mask & SDVOB_HOTPLUG_INT_STATUS_I915)
-			hotplug_en |= SDVOB_HOTPLUG_INT_EN;
-		if (dev_priv->hotplug_supported_mask & CRT_HOTPLUG_INT_STATUS) {
-			hotplug_en |= CRT_HOTPLUG_INT_EN;
-			hotplug_en |= CRT_HOTPLUG_VOLTAGE_COMPARE_50;
-		}
+	if ((iir & flip_pending) == 0)
+		return false;
 
-		/* Ignore TV since it's buggy */
+	intel_prepare_page_flip(dev, plane);
 
-		I915_WRITE(PORT_HOTPLUG_EN, hotplug_en);
-	}
+	/* We detect FlipDone by looking for the change in PendingFlip from '1'
+	 * to '0' on the following vblank, i.e. IIR has the Pendingflip
+	 * asserted following the MI_DISPLAY_FLIP, but ISR is deasserted, hence
+	 * the flip is completed (no longer pending). Since this doesn't raise
+	 * an interrupt per se, we watch for the change at vblank.
+	 */
+	if (I915_READ(ISR) & flip_pending)
+		return false;
+
+	intel_finish_page_flip(dev, pipe);
+
+	return true;
 }
 
 static irqreturn_t i915_irq_handler(int irq, void *arg)
@@ -2442,10 +2671,6 @@ static irqreturn_t i915_irq_handler(int irq, void *arg)
 	u32 flip_mask =
 		I915_DISPLAY_PLANE_A_FLIP_PENDING_INTERRUPT |
 		I915_DISPLAY_PLANE_B_FLIP_PENDING_INTERRUPT;
-	u32 flip[2] = {
-		I915_DISPLAY_PLANE_A_FLIP_PENDING_INTERRUPT,
-		I915_DISPLAY_PLANE_B_FLIP_PENDING_INTERRUPT
-	};
 	int pipe, ret = IRQ_NONE;
 
 	atomic_inc(&dev_priv->irq_received);
@@ -2486,13 +2711,16 @@ static irqreturn_t i915_irq_handler(int irq, void *arg)
 		if ((I915_HAS_HOTPLUG(dev)) &&
 		    (iir & I915_DISPLAY_PORT_INTERRUPT)) {
 			u32 hotplug_status = I915_READ(PORT_HOTPLUG_STAT);
+			u32 hotplug_trigger = hotplug_status & HOTPLUG_INT_STATUS_I915;
 
 			DRM_DEBUG_DRIVER("hotplug event received, stat 0x%08x\n",
 				  hotplug_status);
-			if (hotplug_status & dev_priv->hotplug_supported_mask)
+			if (hotplug_trigger) {
+				if (hotplug_irq_storm_detect(dev, hotplug_trigger, hpd_status_i915))
+					i915_hpd_irq_setup(dev);
 				queue_work(dev_priv->wq,
 					   &dev_priv->hotplug_work);
-
+			}
 			I915_WRITE(PORT_HOTPLUG_STAT, hotplug_status);
 			POSTING_READ(PORT_HOTPLUG_STAT);
 		}
@@ -2507,14 +2735,10 @@ static irqreturn_t i915_irq_handler(int irq, void *arg)
 			int plane = pipe;
 			if (IS_MOBILE(dev))
 				plane = !plane;
+
 			if (pipe_stats[pipe] & PIPE_VBLANK_INTERRUPT_STATUS &&
-			    drm_handle_vblank(dev, pipe)) {
-				if (iir & flip[plane]) {
-					intel_prepare_page_flip(dev, plane);
-					intel_finish_page_flip(dev, pipe);
-					flip_mask &= ~flip[plane];
-				}
-			}
+			    i915_handle_vblank(dev, plane, pipe, iir))
+				flip_mask &= ~DISPLAY_PLANE_FLIP_PENDING(plane);
 
 			if (pipe_stats[pipe] & PIPE_LEGACY_BLC_EVENT_STATUS)
 				blc_event = true;
@@ -2552,6 +2776,8 @@ static void i915_irq_uninstall(struct drm_device * dev)
 	drm_i915_private_t *dev_priv = (drm_i915_private_t *) dev->dev_private;
 	int pipe;
 
+	del_timer_sync(&dev_priv->hotplug_reenable_timer);
+
 	if (I915_HAS_HOTPLUG(dev)) {
 		I915_WRITE(PORT_HOTPLUG_EN, 0);
 		I915_WRITE(PORT_HOTPLUG_STAT, I915_READ(PORT_HOTPLUG_STAT));
@@ -2603,13 +2829,13 @@ static int i965_irq_postinstall(struct drm_device *dev)
 			       I915_RENDER_COMMAND_PARSER_ERROR_INTERRUPT);
 
 	enable_mask = ~dev_priv->irq_mask;
+	enable_mask &= ~(I915_DISPLAY_PLANE_A_FLIP_PENDING_INTERRUPT |
+			 I915_DISPLAY_PLANE_B_FLIP_PENDING_INTERRUPT);
 	enable_mask |= I915_USER_INTERRUPT;
 
 	if (IS_G4X(dev))
 		enable_mask |= I915_BSD_USER_INTERRUPT;
 
-	dev_priv->pipestat[0] = 0;
-	dev_priv->pipestat[1] = 0;
 	i915_enable_pipestat(dev_priv, 0, PIPE_GMBUS_EVENT_ENABLE);
 
 	/*
@@ -2639,45 +2865,33 @@ static int i965_irq_postinstall(struct drm_device *dev)
 	return 0;
 }
 
-static void i965_hpd_irq_setup(struct drm_device *dev)
+static void i915_hpd_irq_setup(struct drm_device *dev)
 {
 	drm_i915_private_t *dev_priv = (drm_i915_private_t *) dev->dev_private;
+	struct drm_mode_config *mode_config = &dev->mode_config;
+	struct intel_encoder *intel_encoder;
 	u32 hotplug_en;
 
-	/* Note HDMI and DP share hotplug bits */
-	hotplug_en = 0;
-	if (dev_priv->hotplug_supported_mask & PORTB_HOTPLUG_INT_STATUS)
-		hotplug_en |= PORTB_HOTPLUG_INT_EN;
-	if (dev_priv->hotplug_supported_mask & PORTC_HOTPLUG_INT_STATUS)
-		hotplug_en |= PORTC_HOTPLUG_INT_EN;
-	if (dev_priv->hotplug_supported_mask & PORTD_HOTPLUG_INT_STATUS)
-		hotplug_en |= PORTD_HOTPLUG_INT_EN;
-	if (IS_G4X(dev)) {
-		if (dev_priv->hotplug_supported_mask & SDVOC_HOTPLUG_INT_STATUS_G4X)
-			hotplug_en |= SDVOC_HOTPLUG_INT_EN;
-		if (dev_priv->hotplug_supported_mask & SDVOB_HOTPLUG_INT_STATUS_G4X)
-			hotplug_en |= SDVOB_HOTPLUG_INT_EN;
-	} else {
-		if (dev_priv->hotplug_supported_mask & SDVOC_HOTPLUG_INT_STATUS_I965)
-			hotplug_en |= SDVOC_HOTPLUG_INT_EN;
-		if (dev_priv->hotplug_supported_mask & SDVOB_HOTPLUG_INT_STATUS_I965)
-			hotplug_en |= SDVOB_HOTPLUG_INT_EN;
-	}
-	if (dev_priv->hotplug_supported_mask & CRT_HOTPLUG_INT_STATUS) {
-		hotplug_en |= CRT_HOTPLUG_INT_EN;
-
+	if (I915_HAS_HOTPLUG(dev)) {
+		hotplug_en = I915_READ(PORT_HOTPLUG_EN);
+		hotplug_en &= ~HOTPLUG_INT_EN_MASK;
+		/* Note HDMI and DP share hotplug bits */
+		/* enable bits are the same for all generations */
+		list_for_each_entry(intel_encoder, &mode_config->encoder_list, base.head)
+			if (dev_priv->hpd_stats[intel_encoder->hpd_pin].hpd_mark == HPD_ENABLED)
+				hotplug_en |= hpd_mask_i915[intel_encoder->hpd_pin];
 		/* Programming the CRT detection parameters tends
 		   to generate a spurious hotplug event about three
 		   seconds later.  So just do it once.
-		   */
+		*/
 		if (IS_G4X(dev))
 			hotplug_en |= CRT_HOTPLUG_ACTIVATION_PERIOD_64;
+		hotplug_en &= ~CRT_HOTPLUG_VOLTAGE_COMPARE_MASK;
 		hotplug_en |= CRT_HOTPLUG_VOLTAGE_COMPARE_50;
-	}
-
-	/* Ignore TV since it's buggy */
 
-	I915_WRITE(PORT_HOTPLUG_EN, hotplug_en);
+		/* Ignore TV since it's buggy */
+		I915_WRITE(PORT_HOTPLUG_EN, hotplug_en);
+	}
 }
 
 static irqreturn_t i965_irq_handler(int irq, void *arg)
@@ -2689,6 +2903,9 @@ static irqreturn_t i965_irq_handler(int irq, void *arg)
 	unsigned long irqflags;
 	int irq_received;
 	int ret = IRQ_NONE, pipe;
+	u32 flip_mask =
+		I915_DISPLAY_PLANE_A_FLIP_PENDING_INTERRUPT |
+		I915_DISPLAY_PLANE_B_FLIP_PENDING_INTERRUPT;
 
 	atomic_inc(&dev_priv->irq_received);
 
@@ -2697,7 +2914,7 @@ static irqreturn_t i965_irq_handler(int irq, void *arg)
 	for (;;) {
 		bool blc_event = false;
 
-		irq_received = iir != 0;
+		irq_received = (iir & ~flip_mask) != 0;
 
 		/* Can't rely on pipestat interrupt bit in iir as it might
 		 * have been cleared after the pipestat interrupt was received.
@@ -2733,18 +2950,24 @@ static irqreturn_t i965_irq_handler(int irq, void *arg)
 		/* Consume port.  Then clear IIR or we'll miss events */
 		if (iir & I915_DISPLAY_PORT_INTERRUPT) {
 			u32 hotplug_status = I915_READ(PORT_HOTPLUG_STAT);
+			u32 hotplug_trigger = hotplug_status & (IS_G4X(dev) ?
+								  HOTPLUG_INT_STATUS_G4X :
+								  HOTPLUG_INT_STATUS_I965);
 
 			DRM_DEBUG_DRIVER("hotplug event received, stat 0x%08x\n",
 				  hotplug_status);
-			if (hotplug_status & dev_priv->hotplug_supported_mask)
+			if (hotplug_trigger) {
+				if (hotplug_irq_storm_detect(dev, hotplug_trigger,
+							    IS_G4X(dev) ? hpd_status_gen4 : hpd_status_i965))
+					i915_hpd_irq_setup(dev);
 				queue_work(dev_priv->wq,
 					   &dev_priv->hotplug_work);
-
+			}
 			I915_WRITE(PORT_HOTPLUG_STAT, hotplug_status);
 			I915_READ(PORT_HOTPLUG_STAT);
 		}
 
-		I915_WRITE(IIR, iir);
+		I915_WRITE(IIR, iir & ~flip_mask);
 		new_iir = I915_READ(IIR); /* Flush posted writes */
 
 		if (iir & I915_USER_INTERRUPT)
@@ -2752,18 +2975,10 @@ static irqreturn_t i965_irq_handler(int irq, void *arg)
 		if (iir & I915_BSD_USER_INTERRUPT)
 			notify_ring(dev, &dev_priv->ring[VCS]);
 
-		if (iir & I915_DISPLAY_PLANE_A_FLIP_PENDING_INTERRUPT)
-			intel_prepare_page_flip(dev, 0);
-
-		if (iir & I915_DISPLAY_PLANE_B_FLIP_PENDING_INTERRUPT)
-			intel_prepare_page_flip(dev, 1);
-
 		for_each_pipe(pipe) {
 			if (pipe_stats[pipe] & PIPE_START_VBLANK_INTERRUPT_STATUS &&
-			    drm_handle_vblank(dev, pipe)) {
-				i915_pageflip_stall_check(dev, pipe);
-				intel_finish_page_flip(dev, pipe);
-			}
+			    i915_handle_vblank(dev, pipe, pipe, iir))
+				flip_mask &= ~DISPLAY_PLANE_FLIP_PENDING(pipe);
 
 			if (pipe_stats[pipe] & PIPE_LEGACY_BLC_EVENT_STATUS)
 				blc_event = true;
@@ -2807,6 +3022,8 @@ static void i965_irq_uninstall(struct drm_device * dev)
 	if (!dev_priv)
 		return;
 
+	del_timer_sync(&dev_priv->hotplug_reenable_timer);
+
 	I915_WRITE(PORT_HOTPLUG_EN, 0);
 	I915_WRITE(PORT_HOTPLUG_STAT, I915_READ(PORT_HOTPLUG_STAT));
 
@@ -2822,6 +3039,41 @@ static void i965_irq_uninstall(struct drm_device * dev)
 	I915_WRITE(IIR, I915_READ(IIR));
 }
 
+static void i915_reenable_hotplug_timer_func(unsigned long data)
+{
+	drm_i915_private_t *dev_priv = (drm_i915_private_t *)data;
+	struct drm_device *dev = dev_priv->dev;
+	struct drm_mode_config *mode_config = &dev->mode_config;
+	unsigned long irqflags;
+	int i;
+
+	spin_lock_irqsave(&dev_priv->irq_lock, irqflags);
+	for (i = (HPD_NONE + 1); i < HPD_NUM_PINS; i++) {
+		struct drm_connector *connector;
+
+		if (dev_priv->hpd_stats[i].hpd_mark != HPD_DISABLED)
+			continue;
+
+		dev_priv->hpd_stats[i].hpd_mark = HPD_ENABLED;
+
+		list_for_each_entry(connector, &mode_config->connector_list, head) {
+			struct intel_connector *intel_connector = to_intel_connector(connector);
+
+			if (intel_connector->encoder->hpd_pin == i) {
+				if (connector->polled != intel_connector->polled)
+					DRM_DEBUG_DRIVER("Reenabling HPD on connector %s\n",
+							 drm_get_connector_name(connector));
+				connector->polled = intel_connector->polled;
+				if (!connector->polled)
+					connector->polled = DRM_CONNECTOR_POLL_HPD;
+			}
+		}
+	}
+	if (dev_priv->display.hpd_irq_setup)
+		dev_priv->display.hpd_irq_setup(dev);
+	spin_unlock_irqrestore(&dev_priv->irq_lock, irqflags);
+}
+
 void intel_irq_init(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
@@ -2834,6 +3086,8 @@ void intel_irq_init(struct drm_device *dev)
 	setup_timer(&dev_priv->gpu_error.hangcheck_timer,
 		    i915_hangcheck_elapsed,
 		    (unsigned long) dev);
+	setup_timer(&dev_priv->hotplug_reenable_timer, i915_reenable_hotplug_timer_func,
+		    (unsigned long) dev_priv);
 
 	pm_qos_add_request(&dev_priv->pm_qos, PM_QOS_CPU_DMA_LATENCY, PM_QOS_DEFAULT_VALUE);
 
@@ -2857,7 +3111,7 @@ void intel_irq_init(struct drm_device *dev)
 		dev->driver->irq_uninstall = valleyview_irq_uninstall;
 		dev->driver->enable_vblank = valleyview_enable_vblank;
 		dev->driver->disable_vblank = valleyview_disable_vblank;
-		dev_priv->display.hpd_irq_setup = valleyview_hpd_irq_setup;
+		dev_priv->display.hpd_irq_setup = i915_hpd_irq_setup;
 	} else if (IS_IVYBRIDGE(dev) || IS_HASWELL(dev)) {
 		/* Share pre & uninstall handlers with ILK/SNB */
 		dev->driver->irq_handler = ivybridge_irq_handler;
@@ -2866,6 +3120,7 @@ void intel_irq_init(struct drm_device *dev)
 		dev->driver->irq_uninstall = ironlake_irq_uninstall;
 		dev->driver->enable_vblank = ivybridge_enable_vblank;
 		dev->driver->disable_vblank = ivybridge_disable_vblank;
+		dev_priv->display.hpd_irq_setup = ibx_hpd_irq_setup;
 	} else if (HAS_PCH_SPLIT(dev)) {
 		dev->driver->irq_handler = ironlake_irq_handler;
 		dev->driver->irq_preinstall = ironlake_irq_preinstall;
@@ -2873,6 +3128,7 @@ void intel_irq_init(struct drm_device *dev)
 		dev->driver->irq_uninstall = ironlake_irq_uninstall;
 		dev->driver->enable_vblank = ironlake_enable_vblank;
 		dev->driver->disable_vblank = ironlake_disable_vblank;
+		dev_priv->display.hpd_irq_setup = ibx_hpd_irq_setup;
 	} else {
 		if (INTEL_INFO(dev)->gen == 2) {
 			dev->driver->irq_preinstall = i8xx_irq_preinstall;
@@ -2890,7 +3146,7 @@ void intel_irq_init(struct drm_device *dev)
 			dev->driver->irq_postinstall = i965_irq_postinstall;
 			dev->driver->irq_uninstall = i965_irq_uninstall;
 			dev->driver->irq_handler = i965_irq_handler;
-			dev_priv->display.hpd_irq_setup = i965_hpd_irq_setup;
+			dev_priv->display.hpd_irq_setup = i915_hpd_irq_setup;
 		}
 		dev->driver->enable_vblank = i915_enable_vblank;
 		dev->driver->disable_vblank = i915_disable_vblank;
@@ -2900,7 +3156,20 @@ void intel_irq_init(struct drm_device *dev)
 void intel_hpd_init(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_mode_config *mode_config = &dev->mode_config;
+	struct drm_connector *connector;
+	int i;
 
+	for (i = 1; i < HPD_NUM_PINS; i++) {
+		dev_priv->hpd_stats[i].hpd_cnt = 0;
+		dev_priv->hpd_stats[i].hpd_mark = HPD_ENABLED;
+	}
+	list_for_each_entry(connector, &mode_config->connector_list, head) {
+		struct intel_connector *intel_connector = to_intel_connector(connector);
+		connector->polled = intel_connector->polled;
+		if (!connector->polled && I915_HAS_HOTPLUG(dev) && intel_connector->encoder->hpd_pin > HPD_NONE)
+			connector->polled = DRM_CONNECTOR_POLL_HPD;
+	}
 	if (dev_priv->display.hpd_irq_setup)
 		dev_priv->display.hpd_irq_setup(dev);
 }
diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h
index 848992f67d56..83f9c26e1adb 100644
--- a/drivers/gpu/drm/i915/i915_reg.h
+++ b/drivers/gpu/drm/i915/i915_reg.h
@@ -91,6 +91,7 @@
 #define  GRDOM_FULL	(0<<2)
 #define  GRDOM_RENDER	(1<<2)
 #define  GRDOM_MEDIA	(3<<2)
+#define  GRDOM_MASK	(3<<2)
 #define  GRDOM_RESET_ENABLE (1<<0)
 
 #define GEN6_MBCUNIT_SNPCR	0x900c /* for LLC config */
@@ -121,10 +122,17 @@
 
 #define GAM_ECOCHK			0x4090
 #define   ECOCHK_SNB_BIT		(1<<10)
+#define   HSW_ECOCHK_ARB_PRIO_SOL	(1<<6)
 #define   ECOCHK_PPGTT_CACHE64B		(0x3<<3)
 #define   ECOCHK_PPGTT_CACHE4B		(0x0<<3)
+#define   ECOCHK_PPGTT_GFDT_IVB		(0x1<<4)
+#define   ECOCHK_PPGTT_LLC_IVB		(0x1<<3)
+#define   ECOCHK_PPGTT_UC_HSW		(0x1<<3)
+#define   ECOCHK_PPGTT_WT_HSW		(0x2<<3)
+#define   ECOCHK_PPGTT_WB_HSW		(0x3<<3)
 
 #define GAC_ECO_BITS			0x14090
+#define   ECOBITS_SNB_BIT		(1<<13)
 #define   ECOBITS_PPGTT_CACHE64B	(3<<8)
 #define   ECOBITS_PPGTT_CACHE4B		(0<<8)
 
@@ -422,6 +430,7 @@
 
 #define FENCE_REG_SANDYBRIDGE_0		0x100000
 #define   SANDYBRIDGE_FENCE_PITCH_SHIFT	32
+#define   GEN7_FENCE_MAX_PITCH_VAL	0x0800
 
 /* control register for cpu gtt access */
 #define TILECTL				0x101000
@@ -522,6 +531,9 @@
 #define GEN7_ERR_INT	0x44040
 #define   ERR_INT_MMIO_UNCLAIMED (1<<13)
 
+#define FPGA_DBG		0x42300
+#define   FPGA_DBG_RM_NOCLAIM	(1<<31)
+
 #define DERRMR		0x44050
 
 /* GM45+ chicken bits -- debug workaround bits that may be required
@@ -591,6 +603,7 @@
 #define   I915_USER_INTERRUPT				(1<<1)
 #define   I915_ASLE_INTERRUPT				(1<<0)
 #define   I915_BSD_USER_INTERRUPT                      (1<<25)
+#define   DISPLAY_PLANE_FLIP_PENDING(plane) (1<<(11-(plane))) /* A and B only */
 #define EIR		0x020b0
 #define EMR		0x020b4
 #define ESR		0x020b8
@@ -1197,6 +1210,9 @@
 
 #define MCHBAR_MIRROR_BASE_SNB	0x140000
 
+/* Memory controller frequency in MCHBAR for Haswell (possible SNB+) */
+#define DCLK 0x5e04
+
 /** 915-945 and GM965 MCH register controlling DRAM channel access */
 #define DCC			0x10200
 #define DCC_ADDRESSING_MODE_SINGLE_CHANNEL		(0 << 0)
@@ -1637,6 +1653,12 @@
 #define   SDVOC_HOTPLUG_INT_EN			(1 << 25)
 #define   TV_HOTPLUG_INT_EN			(1 << 18)
 #define   CRT_HOTPLUG_INT_EN			(1 << 9)
+#define HOTPLUG_INT_EN_MASK			(PORTB_HOTPLUG_INT_EN | \
+						 PORTC_HOTPLUG_INT_EN | \
+						 PORTD_HOTPLUG_INT_EN | \
+						 SDVOC_HOTPLUG_INT_EN | \
+						 SDVOB_HOTPLUG_INT_EN | \
+						 CRT_HOTPLUG_INT_EN)
 #define   CRT_HOTPLUG_FORCE_DETECT		(1 << 3)
 #define CRT_HOTPLUG_ACTIVATION_PERIOD_32	(0 << 8)
 /* must use period 64 on GM45 according to docs */
@@ -1675,43 +1697,84 @@
 #define   SDVOB_HOTPLUG_INT_STATUS_I965		(3 << 2)
 #define   SDVOC_HOTPLUG_INT_STATUS_I915		(1 << 7)
 #define   SDVOB_HOTPLUG_INT_STATUS_I915		(1 << 6)
-
-/* SDVO port control */
-#define SDVOB			0x61140
-#define SDVOC			0x61160
-#define   SDVO_ENABLE		(1 << 31)
-#define   SDVO_PIPE_B_SELECT	(1 << 30)
-#define   SDVO_STALL_SELECT	(1 << 29)
-#define   SDVO_INTERRUPT_ENABLE	(1 << 26)
+#define   HOTPLUG_INT_STATUS_G4X		(CRT_HOTPLUG_INT_STATUS | \
+						 SDVOB_HOTPLUG_INT_STATUS_G4X | \
+						 SDVOC_HOTPLUG_INT_STATUS_G4X | \
+						 PORTB_HOTPLUG_INT_STATUS | \
+						 PORTC_HOTPLUG_INT_STATUS | \
+						 PORTD_HOTPLUG_INT_STATUS)
+
+#define HOTPLUG_INT_STATUS_I965			(CRT_HOTPLUG_INT_STATUS | \
+						 SDVOB_HOTPLUG_INT_STATUS_I965 | \
+						 SDVOC_HOTPLUG_INT_STATUS_I965 | \
+						 PORTB_HOTPLUG_INT_STATUS | \
+						 PORTC_HOTPLUG_INT_STATUS | \
+						 PORTD_HOTPLUG_INT_STATUS)
+
+#define HOTPLUG_INT_STATUS_I915			(CRT_HOTPLUG_INT_STATUS | \
+						 SDVOB_HOTPLUG_INT_STATUS_I915 | \
+						 SDVOC_HOTPLUG_INT_STATUS_I915 | \
+						 PORTB_HOTPLUG_INT_STATUS | \
+						 PORTC_HOTPLUG_INT_STATUS | \
+						 PORTD_HOTPLUG_INT_STATUS)
+
+/* SDVO and HDMI port control.
+ * The same register may be used for SDVO or HDMI */
+#define GEN3_SDVOB	0x61140
+#define GEN3_SDVOC	0x61160
+#define GEN4_HDMIB	GEN3_SDVOB
+#define GEN4_HDMIC	GEN3_SDVOC
+#define PCH_SDVOB	0xe1140
+#define PCH_HDMIB	PCH_SDVOB
+#define PCH_HDMIC	0xe1150
+#define PCH_HDMID	0xe1160
+
+/* Gen 3 SDVO bits: */
+#define   SDVO_ENABLE				(1 << 31)
+#define   SDVO_PIPE_SEL(pipe)			((pipe) << 30)
+#define   SDVO_PIPE_SEL_MASK			(1 << 30)
+#define   SDVO_PIPE_B_SELECT			(1 << 30)
+#define   SDVO_STALL_SELECT			(1 << 29)
+#define   SDVO_INTERRUPT_ENABLE			(1 << 26)
 /**
  * 915G/GM SDVO pixel multiplier.
- *
  * Programmed value is multiplier - 1, up to 5x.
- *
  * \sa DPLL_MD_UDI_MULTIPLIER_MASK
  */
-#define   SDVO_PORT_MULTIPLY_MASK	(7 << 23)
+#define   SDVO_PORT_MULTIPLY_MASK		(7 << 23)
 #define   SDVO_PORT_MULTIPLY_SHIFT		23
-#define   SDVO_PHASE_SELECT_MASK	(15 << 19)
-#define   SDVO_PHASE_SELECT_DEFAULT	(6 << 19)
-#define   SDVO_CLOCK_OUTPUT_INVERT	(1 << 18)
-#define   SDVOC_GANG_MODE		(1 << 16)
-#define   SDVO_ENCODING_SDVO		(0x0 << 10)
-#define   SDVO_ENCODING_HDMI		(0x2 << 10)
-/** Requird for HDMI operation */
-#define   SDVO_NULL_PACKETS_DURING_VSYNC (1 << 9)
-#define   SDVO_COLOR_RANGE_16_235	(1 << 8)
-#define   SDVO_BORDER_ENABLE		(1 << 7)
-#define   SDVO_AUDIO_ENABLE		(1 << 6)
-/** New with 965, default is to be set */
-#define   SDVO_VSYNC_ACTIVE_HIGH	(1 << 4)
-/** New with 965, default is to be set */
-#define   SDVO_HSYNC_ACTIVE_HIGH	(1 << 3)
-#define   SDVOB_PCIE_CONCURRENCY	(1 << 3)
-#define   SDVO_DETECTED			(1 << 2)
+#define   SDVO_PHASE_SELECT_MASK		(15 << 19)
+#define   SDVO_PHASE_SELECT_DEFAULT		(6 << 19)
+#define   SDVO_CLOCK_OUTPUT_INVERT		(1 << 18)
+#define   SDVOC_GANG_MODE			(1 << 16) /* Port C only */
+#define   SDVO_BORDER_ENABLE			(1 << 7) /* SDVO only */
+#define   SDVOB_PCIE_CONCURRENCY		(1 << 3) /* Port B only */
+#define   SDVO_DETECTED				(1 << 2)
 /* Bits to be preserved when writing */
-#define   SDVOB_PRESERVE_MASK ((1 << 17) | (1 << 16) | (1 << 14) | (1 << 26))
-#define   SDVOC_PRESERVE_MASK ((1 << 17) | (1 << 26))
+#define   SDVOB_PRESERVE_MASK ((1 << 17) | (1 << 16) | (1 << 14) | \
+			       SDVO_INTERRUPT_ENABLE)
+#define   SDVOC_PRESERVE_MASK ((1 << 17) | SDVO_INTERRUPT_ENABLE)
+
+/* Gen 4 SDVO/HDMI bits: */
+#define   SDVO_COLOR_FORMAT_8bpc		(0 << 26)
+#define   SDVO_ENCODING_SDVO			(0 << 10)
+#define   SDVO_ENCODING_HDMI			(2 << 10)
+#define   HDMI_MODE_SELECT_HDMI			(1 << 9) /* HDMI only */
+#define   HDMI_MODE_SELECT_DVI			(0 << 9) /* HDMI only */
+#define   HDMI_COLOR_RANGE_16_235		(1 << 8) /* HDMI only */
+#define   SDVO_AUDIO_ENABLE			(1 << 6)
+/* VSYNC/HSYNC bits new with 965, default is to be set */
+#define   SDVO_VSYNC_ACTIVE_HIGH		(1 << 4)
+#define   SDVO_HSYNC_ACTIVE_HIGH		(1 << 3)
+
+/* Gen 5 (IBX) SDVO/HDMI bits: */
+#define   HDMI_COLOR_FORMAT_12bpc		(3 << 26) /* HDMI only */
+#define   SDVOB_HOTPLUG_ENABLE			(1 << 23) /* SDVO only */
+
+/* Gen 6 (CPT) SDVO/HDMI bits: */
+#define   SDVO_PIPE_SEL_CPT(pipe)		((pipe) << 29)
+#define   SDVO_PIPE_SEL_MASK_CPT		(3 << 29)
+
 
 /* DVO port control */
 #define DVOA			0x61120
@@ -1898,7 +1961,7 @@
 #define PFIT_AUTO_RATIOS (dev_priv->info->display_mmio_offset + 0x61238)
 
 /* Backlight control */
-#define BLC_PWM_CTL2		0x61250 /* 965+ only */
+#define BLC_PWM_CTL2	(dev_priv->info->display_mmio_offset + 0x61250) /* 965+ only */
 #define   BLM_PWM_ENABLE		(1 << 31)
 #define   BLM_COMBINATION_MODE		(1 << 30) /* gen4 only */
 #define   BLM_PIPE_SELECT		(1 << 29)
@@ -1917,7 +1980,7 @@
 #define   BLM_PHASE_IN_COUNT_MASK	(0xff << 8)
 #define   BLM_PHASE_IN_INCR_SHIFT	(0)
 #define   BLM_PHASE_IN_INCR_MASK	(0xff << 0)
-#define BLC_PWM_CTL		0x61254
+#define BLC_PWM_CTL	(dev_priv->info->display_mmio_offset + 0x61254)
 /*
  * This is the most significant 15 bits of the number of backlight cycles in a
  * complete cycle of the modulated backlight control.
@@ -1939,7 +2002,7 @@
 #define   BACKLIGHT_DUTY_CYCLE_MASK_PNV		(0xfffe)
 #define   BLM_POLARITY_PNV			(1 << 0) /* pnv only */
 
-#define BLC_HIST_CTL		0x61260
+#define BLC_HIST_CTL	(dev_priv->info->display_mmio_offset + 0x61260)
 
 /* New registers for PCH-split platforms. Safe where new bits show up, the
  * register layout machtes with gen4 BLC_PWM_CTL[12]. */
@@ -2589,14 +2652,14 @@
 #define _PIPEB_GMCH_DATA_M			0x71050
 
 /* Transfer unit size for display port - 1, default is 0x3f (for TU size 64) */
-#define   PIPE_GMCH_DATA_M_TU_SIZE_MASK		(0x3f << 25)
-#define   PIPE_GMCH_DATA_M_TU_SIZE_SHIFT	25
+#define  TU_SIZE(x)             (((x)-1) << 25) /* default size 64 */
+#define  TU_SIZE_MASK           (0x3f << 25)
 
-#define   PIPE_GMCH_DATA_M_MASK			(0xffffff)
+#define  DATA_LINK_M_N_MASK	(0xffffff)
+#define  DATA_LINK_N_MAX	(0x800000)
 
 #define _PIPEA_GMCH_DATA_N			0x70054
 #define _PIPEB_GMCH_DATA_N			0x71054
-#define   PIPE_GMCH_DATA_N_MASK			(0xffffff)
 
 /*
  * Computing Link M and N values for the Display Port link
@@ -2611,11 +2674,9 @@
 
 #define _PIPEA_DP_LINK_M				0x70060
 #define _PIPEB_DP_LINK_M				0x71060
-#define   PIPEA_DP_LINK_M_MASK			(0xffffff)
 
 #define _PIPEA_DP_LINK_N				0x70064
 #define _PIPEB_DP_LINK_N				0x71064
-#define   PIPEA_DP_LINK_N_MASK			(0xffffff)
 
 #define PIPE_GMCH_DATA_M(pipe) _PIPE(pipe, _PIPEA_GMCH_DATA_M, _PIPEB_GMCH_DATA_M)
 #define PIPE_GMCH_DATA_N(pipe) _PIPE(pipe, _PIPEA_GMCH_DATA_N, _PIPEB_GMCH_DATA_N)
@@ -2776,6 +2837,8 @@
 #define   DSPFW_HPLL_CURSOR_SHIFT	16
 #define   DSPFW_HPLL_CURSOR_MASK	(0x3f<<16)
 #define   DSPFW_HPLL_SR_MASK		(0x1ff)
+#define DSPFW4			(dev_priv->info->display_mmio_offset + 0x70070)
+#define DSPFW7			(dev_priv->info->display_mmio_offset + 0x7007c)
 
 /* drain latency register values*/
 #define DRAIN_LATENCY_PRECISION_32	32
@@ -3233,6 +3296,63 @@
 #define SPRGAMC(pipe) _PIPE(pipe, _SPRA_GAMC, _SPRB_GAMC)
 #define SPRSURFLIVE(pipe) _PIPE(pipe, _SPRA_SURFLIVE, _SPRB_SURFLIVE)
 
+#define _SPACNTR		0x72180
+#define   SP_ENABLE			(1<<31)
+#define   SP_GEAMMA_ENABLE		(1<<30)
+#define   SP_PIXFORMAT_MASK		(0xf<<26)
+#define   SP_FORMAT_YUV422		(0<<26)
+#define   SP_FORMAT_BGR565		(5<<26)
+#define   SP_FORMAT_BGRX8888		(6<<26)
+#define   SP_FORMAT_BGRA8888		(7<<26)
+#define   SP_FORMAT_RGBX1010102		(8<<26)
+#define   SP_FORMAT_RGBA1010102		(9<<26)
+#define   SP_FORMAT_RGBX8888		(0xe<<26)
+#define   SP_FORMAT_RGBA8888		(0xf<<26)
+#define   SP_SOURCE_KEY			(1<<22)
+#define   SP_YUV_BYTE_ORDER_MASK	(3<<16)
+#define   SP_YUV_ORDER_YUYV		(0<<16)
+#define   SP_YUV_ORDER_UYVY		(1<<16)
+#define   SP_YUV_ORDER_YVYU		(2<<16)
+#define   SP_YUV_ORDER_VYUY		(3<<16)
+#define   SP_TILED			(1<<10)
+#define _SPALINOFF		0x72184
+#define _SPASTRIDE		0x72188
+#define _SPAPOS			0x7218c
+#define _SPASIZE		0x72190
+#define _SPAKEYMINVAL		0x72194
+#define _SPAKEYMSK		0x72198
+#define _SPASURF		0x7219c
+#define _SPAKEYMAXVAL		0x721a0
+#define _SPATILEOFF		0x721a4
+#define _SPACONSTALPHA		0x721a8
+#define _SPAGAMC		0x721f4
+
+#define _SPBCNTR		0x72280
+#define _SPBLINOFF		0x72284
+#define _SPBSTRIDE		0x72288
+#define _SPBPOS			0x7228c
+#define _SPBSIZE		0x72290
+#define _SPBKEYMINVAL		0x72294
+#define _SPBKEYMSK		0x72298
+#define _SPBSURF		0x7229c
+#define _SPBKEYMAXVAL		0x722a0
+#define _SPBTILEOFF		0x722a4
+#define _SPBCONSTALPHA		0x722a8
+#define _SPBGAMC		0x722f4
+
+#define SPCNTR(pipe, plane) _PIPE(pipe * 2 + plane, _SPACNTR, _SPBCNTR)
+#define SPLINOFF(pipe, plane) _PIPE(pipe * 2 + plane, _SPALINOFF, _SPBLINOFF)
+#define SPSTRIDE(pipe, plane) _PIPE(pipe * 2 + plane, _SPASTRIDE, _SPBSTRIDE)
+#define SPPOS(pipe, plane) _PIPE(pipe * 2 + plane, _SPAPOS, _SPBPOS)
+#define SPSIZE(pipe, plane) _PIPE(pipe * 2 + plane, _SPASIZE, _SPBSIZE)
+#define SPKEYMINVAL(pipe, plane) _PIPE(pipe * 2 + plane, _SPAKEYMINVAL, _SPBKEYMINVAL)
+#define SPKEYMSK(pipe, plane) _PIPE(pipe * 2 + plane, _SPAKEYMSK, _SPBKEYMSK)
+#define SPSURF(pipe, plane) _PIPE(pipe * 2 + plane, _SPASURF, _SPBSURF)
+#define SPKEYMAXVAL(pipe, plane) _PIPE(pipe * 2 + plane, _SPAKEYMAXVAL, _SPBKEYMAXVAL)
+#define SPTILEOFF(pipe, plane) _PIPE(pipe * 2 + plane, _SPATILEOFF, _SPBTILEOFF)
+#define SPCONSTALPHA(pipe, plane) _PIPE(pipe * 2 + plane, _SPACONSTALPHA, _SPBCONSTALPHA)
+#define SPGAMC(pipe, plane) _PIPE(pipe * 2 + plane, _SPAGAMC, _SPBGAMC)
+
 /* VBIOS regs */
 #define VGACNTRL		0x71400
 # define VGA_DISP_DISABLE			(1 << 31)
@@ -3282,8 +3402,6 @@
 
 
 #define _PIPEA_DATA_M1           (dev_priv->info->display_mmio_offset + 0x60030)
-#define  TU_SIZE(x)             (((x)-1) << 25) /* default size 64 */
-#define  TU_SIZE_MASK           0x7e000000
 #define  PIPE_DATA_M1_OFFSET    0
 #define _PIPEA_DATA_N1           (dev_priv->info->display_mmio_offset + 0x60034)
 #define  PIPE_DATA_N1_OFFSET    0
@@ -3456,6 +3574,9 @@
 #define DISP_ARB_CTL	0x45000
 #define  DISP_TILE_SURFACE_SWIZZLING	(1<<13)
 #define  DISP_FBC_WM_DIS		(1<<15)
+#define GEN7_MSG_CTL	0x45010
+#define  WAIT_FOR_PCH_RESET_ACK		(1<<1)
+#define  WAIT_FOR_PCH_FLR_ACK		(1<<0)
 
 /* GEN7 chicken */
 #define GEN7_COMMON_SLICE_CHICKEN1		0x7010
@@ -3508,7 +3629,11 @@
 #define SDE_PORTC_HOTPLUG       (1 << 9)
 #define SDE_PORTB_HOTPLUG       (1 << 8)
 #define SDE_SDVOB_HOTPLUG       (1 << 6)
-#define SDE_HOTPLUG_MASK	(0xf << 8)
+#define SDE_HOTPLUG_MASK        (SDE_CRT_HOTPLUG | \
+				 SDE_SDVOB_HOTPLUG |	\
+				 SDE_PORTB_HOTPLUG |	\
+				 SDE_PORTC_HOTPLUG |	\
+				 SDE_PORTD_HOTPLUG)
 #define SDE_TRANSB_CRC_DONE	(1 << 5)
 #define SDE_TRANSB_CRC_ERR	(1 << 4)
 #define SDE_TRANSB_FIFO_UNDER	(1 << 3)
@@ -3531,7 +3656,9 @@
 #define SDE_PORTC_HOTPLUG_CPT	(1 << 22)
 #define SDE_PORTB_HOTPLUG_CPT	(1 << 21)
 #define SDE_CRT_HOTPLUG_CPT	(1 << 19)
+#define SDE_SDVOB_HOTPLUG_CPT	(1 << 18)
 #define SDE_HOTPLUG_MASK_CPT	(SDE_CRT_HOTPLUG_CPT |		\
+				 SDE_SDVOB_HOTPLUG_CPT |	\
 				 SDE_PORTD_HOTPLUG_CPT |	\
 				 SDE_PORTC_HOTPLUG_CPT |	\
 				 SDE_PORTB_HOTPLUG_CPT)
@@ -3754,14 +3881,16 @@
 #define HSW_VIDEO_DIP_VSC_ECC_B		0x61344
 #define HSW_VIDEO_DIP_GCP_B		0x61210
 
-#define HSW_TVIDEO_DIP_CTL(pipe) \
-	 _PIPE(pipe, HSW_VIDEO_DIP_CTL_A, HSW_VIDEO_DIP_CTL_B)
-#define HSW_TVIDEO_DIP_AVI_DATA(pipe) \
-	 _PIPE(pipe, HSW_VIDEO_DIP_AVI_DATA_A, HSW_VIDEO_DIP_AVI_DATA_B)
-#define HSW_TVIDEO_DIP_SPD_DATA(pipe) \
-	 _PIPE(pipe, HSW_VIDEO_DIP_SPD_DATA_A, HSW_VIDEO_DIP_SPD_DATA_B)
-#define HSW_TVIDEO_DIP_GCP(pipe) \
-	_PIPE(pipe, HSW_VIDEO_DIP_GCP_A, HSW_VIDEO_DIP_GCP_B)
+#define HSW_TVIDEO_DIP_CTL(trans) \
+	 _TRANSCODER(trans, HSW_VIDEO_DIP_CTL_A, HSW_VIDEO_DIP_CTL_B)
+#define HSW_TVIDEO_DIP_AVI_DATA(trans) \
+	 _TRANSCODER(trans, HSW_VIDEO_DIP_AVI_DATA_A, HSW_VIDEO_DIP_AVI_DATA_B)
+#define HSW_TVIDEO_DIP_SPD_DATA(trans) \
+	 _TRANSCODER(trans, HSW_VIDEO_DIP_SPD_DATA_A, HSW_VIDEO_DIP_SPD_DATA_B)
+#define HSW_TVIDEO_DIP_GCP(trans) \
+	_TRANSCODER(trans, HSW_VIDEO_DIP_GCP_A, HSW_VIDEO_DIP_GCP_B)
+#define HSW_TVIDEO_DIP_VSC_DATA(trans) \
+	 _TRANSCODER(trans, HSW_VIDEO_DIP_VSC_DATA_A, HSW_VIDEO_DIP_VSC_DATA_B)
 
 #define _TRANS_HTOTAL_B          0xe1000
 #define _TRANS_HBLANK_B          0xe1004
@@ -3826,8 +3955,11 @@
 #define _TRANSA_CHICKEN2	 0xf0064
 #define _TRANSB_CHICKEN2	 0xf1064
 #define TRANS_CHICKEN2(pipe) _PIPE(pipe, _TRANSA_CHICKEN2, _TRANSB_CHICKEN2)
-#define  TRANS_CHICKEN2_TIMING_OVERRIDE		(1<<31)
-
+#define  TRANS_CHICKEN2_TIMING_OVERRIDE			(1<<31)
+#define  TRANS_CHICKEN2_FDI_POLARITY_REVERSED		(1<<29)
+#define  TRANS_CHICKEN2_FRAME_START_DELAY_MASK		(3<<27)
+#define  TRANS_CHICKEN2_DISABLE_DEEP_COLOR_COUNTER	(1<<26)
+#define  TRANS_CHICKEN2_DISABLE_DEEP_COLOR_MODESWITCH	(1<<25)
 
 #define SOUTH_CHICKEN1		0xc2000
 #define  FDIA_PHASE_SYNC_SHIFT_OVR	19
@@ -3976,34 +4108,6 @@
 #define FDI_PLL_CTL_1           0xfe000
 #define FDI_PLL_CTL_2           0xfe004
 
-/* or SDVOB */
-#define HDMIB   0xe1140
-#define  PORT_ENABLE    (1 << 31)
-#define  TRANSCODER(pipe)       ((pipe) << 30)
-#define  TRANSCODER_CPT(pipe)   ((pipe) << 29)
-#define  TRANSCODER_MASK        (1 << 30)
-#define  TRANSCODER_MASK_CPT    (3 << 29)
-#define  COLOR_FORMAT_8bpc      (0)
-#define  COLOR_FORMAT_12bpc     (3 << 26)
-#define  SDVOB_HOTPLUG_ENABLE   (1 << 23)
-#define  SDVO_ENCODING          (0)
-#define  TMDS_ENCODING          (2 << 10)
-#define  NULL_PACKET_VSYNC_ENABLE       (1 << 9)
-/* CPT */
-#define  HDMI_MODE_SELECT	(1 << 9)
-#define  DVI_MODE_SELECT	(0)
-#define  SDVOB_BORDER_ENABLE    (1 << 7)
-#define  AUDIO_ENABLE           (1 << 6)
-#define  VSYNC_ACTIVE_HIGH      (1 << 4)
-#define  HSYNC_ACTIVE_HIGH      (1 << 3)
-#define  PORT_DETECTED          (1 << 2)
-
-/* PCH SDVOB multiplex with HDMIB */
-#define PCH_SDVOB	HDMIB
-
-#define HDMIC   0xe1150
-#define HDMID   0xe1160
-
 #define PCH_LVDS	0xe1180
 #define  LVDS_DETECTED	(1 << 1)
 
@@ -4020,6 +4124,15 @@
 #define PIPEB_PP_OFF_DELAYS     (VLV_DISPLAY_BASE + 0x6130c)
 #define PIPEB_PP_DIVISOR        (VLV_DISPLAY_BASE + 0x61310)
 
+#define VLV_PIPE_PP_STATUS(pipe) _PIPE(pipe, PIPEA_PP_STATUS, PIPEB_PP_STATUS)
+#define VLV_PIPE_PP_CONTROL(pipe) _PIPE(pipe, PIPEA_PP_CONTROL, PIPEB_PP_CONTROL)
+#define VLV_PIPE_PP_ON_DELAYS(pipe) \
+		_PIPE(pipe, PIPEA_PP_ON_DELAYS, PIPEB_PP_ON_DELAYS)
+#define VLV_PIPE_PP_OFF_DELAYS(pipe) \
+		_PIPE(pipe, PIPEA_PP_OFF_DELAYS, PIPEB_PP_OFF_DELAYS)
+#define VLV_PIPE_PP_DIVISOR(pipe) \
+		_PIPE(pipe, PIPEA_PP_DIVISOR, PIPEB_PP_DIVISOR)
+
 #define PCH_PP_STATUS		0xc7200
 #define PCH_PP_CONTROL		0xc7204
 #define  PANEL_UNLOCK_REGS	(0xabcd << 16)
@@ -4149,8 +4262,12 @@
 #define  FORCEWAKE				0xA18C
 #define  FORCEWAKE_VLV				0x1300b0
 #define  FORCEWAKE_ACK_VLV			0x1300b4
+#define  FORCEWAKE_MEDIA_VLV			0x1300b8
+#define  FORCEWAKE_ACK_MEDIA_VLV		0x1300bc
 #define  FORCEWAKE_ACK_HSW			0x130044
 #define  FORCEWAKE_ACK				0x130090
+#define  VLV_GTLC_WAKE_CTRL			0x130090
+#define  VLV_GTLC_PW_STATUS			0x130094
 #define  FORCEWAKE_MT				0xa188 /* multi-threaded */
 #define   FORCEWAKE_KERNEL			0x1
 #define   FORCEWAKE_USER			0x2
@@ -4184,6 +4301,7 @@
 #define GEN6_RPNSWREQ				0xA008
 #define   GEN6_TURBO_DISABLE			(1<<31)
 #define   GEN6_FREQUENCY(x)			((x)<<25)
+#define   HSW_FREQUENCY(x)			((x)<<24)
 #define   GEN6_OFFSET(x)			((x)<<19)
 #define   GEN6_AGGRESSIVE_TURBO			(0<<15)
 #define GEN6_RC_VIDEO_FREQ			0xA00C
@@ -4274,6 +4392,21 @@
 #define   GEN6_DECODE_RC6_VID(vids)		(((vids) * 5) + 245)
 #define GEN6_PCODE_DATA				0x138128
 #define   GEN6_PCODE_FREQ_IA_RATIO_SHIFT	8
+#define   GEN6_PCODE_FREQ_RING_RATIO_SHIFT	16
+
+#define VLV_IOSF_DOORBELL_REQ			0x182100
+#define   IOSF_DEVFN_SHIFT			24
+#define   IOSF_OPCODE_SHIFT			16
+#define   IOSF_PORT_SHIFT			8
+#define   IOSF_BYTE_ENABLES_SHIFT		4
+#define   IOSF_BAR_SHIFT			1
+#define   IOSF_SB_BUSY				(1<<0)
+#define   IOSF_PORT_PUNIT			0x4
+#define VLV_IOSF_DATA				0x182104
+#define VLV_IOSF_ADDR				0x182108
+
+#define PUNIT_OPCODE_REG_READ			6
+#define PUNIT_OPCODE_REG_WRITE			7
 
 #define GEN6_GT_CORE_STATUS		0x138060
 #define   GEN6_CORE_CPD_STATE_MASK	(7<<4)
diff --git a/drivers/gpu/drm/i915/i915_suspend.c b/drivers/gpu/drm/i915/i915_suspend.c
index 2135f21ea458..41f0fdecfbdc 100644
--- a/drivers/gpu/drm/i915/i915_suspend.c
+++ b/drivers/gpu/drm/i915/i915_suspend.c
@@ -209,7 +209,8 @@ static void i915_save_display(struct drm_device *dev)
 		dev_priv->regfile.saveBLC_PWM_CTL2 = I915_READ(BLC_PWM_PCH_CTL2);
 		dev_priv->regfile.saveBLC_CPU_PWM_CTL = I915_READ(BLC_PWM_CPU_CTL);
 		dev_priv->regfile.saveBLC_CPU_PWM_CTL2 = I915_READ(BLC_PWM_CPU_CTL2);
-		dev_priv->regfile.saveLVDS = I915_READ(PCH_LVDS);
+		if (HAS_PCH_IBX(dev) || HAS_PCH_CPT(dev))
+			dev_priv->regfile.saveLVDS = I915_READ(PCH_LVDS);
 	} else {
 		dev_priv->regfile.savePP_CONTROL = I915_READ(PP_CONTROL);
 		dev_priv->regfile.savePFIT_PGM_RATIOS = I915_READ(PFIT_PGM_RATIOS);
@@ -255,6 +256,7 @@ static void i915_save_display(struct drm_device *dev)
 static void i915_restore_display(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
+	u32 mask = 0xffffffff;
 
 	/* Display arbitration */
 	if (INTEL_INFO(dev)->gen <= 4)
@@ -267,10 +269,13 @@ static void i915_restore_display(struct drm_device *dev)
 	if (INTEL_INFO(dev)->gen >= 4 && !HAS_PCH_SPLIT(dev))
 		I915_WRITE(BLC_PWM_CTL2, dev_priv->regfile.saveBLC_PWM_CTL2);
 
-	if (HAS_PCH_SPLIT(dev)) {
-		I915_WRITE(PCH_LVDS, dev_priv->regfile.saveLVDS);
-	} else if (IS_MOBILE(dev) && !IS_I830(dev))
-		I915_WRITE(LVDS, dev_priv->regfile.saveLVDS);
+	if (drm_core_check_feature(dev, DRIVER_MODESET))
+		mask = ~LVDS_PORT_EN;
+
+	if (HAS_PCH_IBX(dev) || HAS_PCH_CPT(dev))
+		I915_WRITE(PCH_LVDS, dev_priv->regfile.saveLVDS & mask);
+	else if (INTEL_INFO(dev)->gen <= 4 && IS_MOBILE(dev) && !IS_I830(dev))
+		I915_WRITE(LVDS, dev_priv->regfile.saveLVDS & mask);
 
 	if (!IS_I830(dev) && !IS_845G(dev) && !HAS_PCH_SPLIT(dev))
 		I915_WRITE(PFIT_CONTROL, dev_priv->regfile.savePFIT_CONTROL);
diff --git a/drivers/gpu/drm/i915/i915_sysfs.c b/drivers/gpu/drm/i915/i915_sysfs.c
index 9462081b1e60..d5e1890678f9 100644
--- a/drivers/gpu/drm/i915/i915_sysfs.c
+++ b/drivers/gpu/drm/i915/i915_sysfs.c
@@ -49,7 +49,7 @@ static ssize_t
 show_rc6_mask(struct device *kdev, struct device_attribute *attr, char *buf)
 {
 	struct drm_minor *dminor = container_of(kdev, struct drm_minor, kdev);
-	return snprintf(buf, PAGE_SIZE, "%x", intel_enable_rc6(dminor->dev));
+	return snprintf(buf, PAGE_SIZE, "%x\n", intel_enable_rc6(dminor->dev));
 }
 
 static ssize_t
@@ -57,7 +57,7 @@ show_rc6_ms(struct device *kdev, struct device_attribute *attr, char *buf)
 {
 	struct drm_minor *dminor = container_of(kdev, struct drm_minor, kdev);
 	u32 rc6_residency = calc_residency(dminor->dev, GEN6_GT_GFX_RC6);
-	return snprintf(buf, PAGE_SIZE, "%u", rc6_residency);
+	return snprintf(buf, PAGE_SIZE, "%u\n", rc6_residency);
 }
 
 static ssize_t
@@ -65,7 +65,7 @@ show_rc6p_ms(struct device *kdev, struct device_attribute *attr, char *buf)
 {
 	struct drm_minor *dminor = container_of(kdev, struct drm_minor, kdev);
 	u32 rc6p_residency = calc_residency(dminor->dev, GEN6_GT_GFX_RC6p);
-	return snprintf(buf, PAGE_SIZE, "%u", rc6p_residency);
+	return snprintf(buf, PAGE_SIZE, "%u\n", rc6p_residency);
 }
 
 static ssize_t
@@ -73,7 +73,7 @@ show_rc6pp_ms(struct device *kdev, struct device_attribute *attr, char *buf)
 {
 	struct drm_minor *dminor = container_of(kdev, struct drm_minor, kdev);
 	u32 rc6pp_residency = calc_residency(dminor->dev, GEN6_GT_GFX_RC6pp);
-	return snprintf(buf, PAGE_SIZE, "%u", rc6pp_residency);
+	return snprintf(buf, PAGE_SIZE, "%u\n", rc6pp_residency);
 }
 
 static DEVICE_ATTR(rc6_enable, S_IRUGO, show_rc6_mask, NULL);
@@ -215,7 +215,7 @@ static ssize_t gt_cur_freq_mhz_show(struct device *kdev,
 	ret = dev_priv->rps.cur_delay * GT_FREQUENCY_MULTIPLIER;
 	mutex_unlock(&dev_priv->rps.hw_lock);
 
-	return snprintf(buf, PAGE_SIZE, "%d", ret);
+	return snprintf(buf, PAGE_SIZE, "%d\n", ret);
 }
 
 static ssize_t gt_max_freq_mhz_show(struct device *kdev, struct device_attribute *attr, char *buf)
@@ -229,7 +229,7 @@ static ssize_t gt_max_freq_mhz_show(struct device *kdev, struct device_attribute
 	ret = dev_priv->rps.max_delay * GT_FREQUENCY_MULTIPLIER;
 	mutex_unlock(&dev_priv->rps.hw_lock);
 
-	return snprintf(buf, PAGE_SIZE, "%d", ret);
+	return snprintf(buf, PAGE_SIZE, "%d\n", ret);
 }
 
 static ssize_t gt_max_freq_mhz_store(struct device *kdev,
@@ -239,7 +239,7 @@ static ssize_t gt_max_freq_mhz_store(struct device *kdev,
 	struct drm_minor *minor = container_of(kdev, struct drm_minor, kdev);
 	struct drm_device *dev = minor->dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	u32 val, rp_state_cap, hw_max, hw_min;
+	u32 val, rp_state_cap, hw_max, hw_min, non_oc_max;
 	ssize_t ret;
 
 	ret = kstrtou32(buf, 0, &val);
@@ -251,7 +251,8 @@ static ssize_t gt_max_freq_mhz_store(struct device *kdev,
 	mutex_lock(&dev_priv->rps.hw_lock);
 
 	rp_state_cap = I915_READ(GEN6_RP_STATE_CAP);
-	hw_max = (rp_state_cap & 0xff);
+	hw_max = dev_priv->rps.hw_max;
+	non_oc_max = (rp_state_cap & 0xff);
 	hw_min = ((rp_state_cap & 0xff0000) >> 16);
 
 	if (val < hw_min || val > hw_max || val < dev_priv->rps.min_delay) {
@@ -259,6 +260,10 @@ static ssize_t gt_max_freq_mhz_store(struct device *kdev,
 		return -EINVAL;
 	}
 
+	if (val > non_oc_max)
+		DRM_DEBUG("User requested overclocking to %d\n",
+			  val * GT_FREQUENCY_MULTIPLIER);
+
 	if (dev_priv->rps.cur_delay > val)
 		gen6_set_rps(dev_priv->dev, val);
 
@@ -280,7 +285,7 @@ static ssize_t gt_min_freq_mhz_show(struct device *kdev, struct device_attribute
 	ret = dev_priv->rps.min_delay * GT_FREQUENCY_MULTIPLIER;
 	mutex_unlock(&dev_priv->rps.hw_lock);
 
-	return snprintf(buf, PAGE_SIZE, "%d", ret);
+	return snprintf(buf, PAGE_SIZE, "%d\n", ret);
 }
 
 static ssize_t gt_min_freq_mhz_store(struct device *kdev,
@@ -302,7 +307,7 @@ static ssize_t gt_min_freq_mhz_store(struct device *kdev,
 	mutex_lock(&dev_priv->rps.hw_lock);
 
 	rp_state_cap = I915_READ(GEN6_RP_STATE_CAP);
-	hw_max = (rp_state_cap & 0xff);
+	hw_max = dev_priv->rps.hw_max;
 	hw_min = ((rp_state_cap & 0xff0000) >> 16);
 
 	if (val < hw_min || val > hw_max || val > dev_priv->rps.max_delay) {
@@ -355,7 +360,7 @@ static ssize_t gt_rp_mhz_show(struct device *kdev, struct device_attribute *attr
 	} else {
 		BUG();
 	}
-	return snprintf(buf, PAGE_SIZE, "%d", val);
+	return snprintf(buf, PAGE_SIZE, "%d\n", val);
 }
 
 static const struct attribute *gen6_attrs[] = {
diff --git a/drivers/gpu/drm/i915/intel_bios.c b/drivers/gpu/drm/i915/intel_bios.c
index 55ffba1f5818..95070b2124c6 100644
--- a/drivers/gpu/drm/i915/intel_bios.c
+++ b/drivers/gpu/drm/i915/intel_bios.c
@@ -351,12 +351,14 @@ parse_general_features(struct drm_i915_private *dev_priv,
 		dev_priv->lvds_ssc_freq =
 			intel_bios_ssc_frequency(dev, general->ssc_freq);
 		dev_priv->display_clock_mode = general->display_clock_mode;
-		DRM_DEBUG_KMS("BDB_GENERAL_FEATURES int_tv_support %d int_crt_support %d lvds_use_ssc %d lvds_ssc_freq %d display_clock_mode %d\n",
+		dev_priv->fdi_rx_polarity_inverted = general->fdi_rx_polarity_inverted;
+		DRM_DEBUG_KMS("BDB_GENERAL_FEATURES int_tv_support %d int_crt_support %d lvds_use_ssc %d lvds_ssc_freq %d display_clock_mode %d fdi_rx_polarity_inverted %d\n",
 			      dev_priv->int_tv_support,
 			      dev_priv->int_crt_support,
 			      dev_priv->lvds_use_ssc,
 			      dev_priv->lvds_ssc_freq,
-			      dev_priv->display_clock_mode);
+			      dev_priv->display_clock_mode,
+			      dev_priv->fdi_rx_polarity_inverted);
 	}
 }
 
@@ -692,6 +694,9 @@ intel_parse_bios(struct drm_device *dev)
 	struct bdb_header *bdb = NULL;
 	u8 __iomem *bios = NULL;
 
+	if (HAS_PCH_NOP(dev))
+		return -ENODEV;
+
 	init_vbt_defaults(dev_priv);
 
 	/* XXX Should this validation be moved to intel_opregion.c? */
diff --git a/drivers/gpu/drm/i915/intel_bios.h b/drivers/gpu/drm/i915/intel_bios.h
index 36e57f934373..e088d6f0956a 100644
--- a/drivers/gpu/drm/i915/intel_bios.h
+++ b/drivers/gpu/drm/i915/intel_bios.h
@@ -127,7 +127,9 @@ struct bdb_general_features {
         /* bits 3 */
 	u8 disable_smooth_vision:1;
 	u8 single_dvi:1;
-	u8 rsvd9:6; /* finish byte */
+	u8 rsvd9:1;
+	u8 fdi_rx_polarity_inverted:1;
+	u8 rsvd10:4; /* finish byte */
 
         /* bits 4 */
 	u8 legacy_monitor_detect;
diff --git a/drivers/gpu/drm/i915/intel_crt.c b/drivers/gpu/drm/i915/intel_crt.c
index 1ce45a0a2d3e..58b4a53715cd 100644
--- a/drivers/gpu/drm/i915/intel_crt.c
+++ b/drivers/gpu/drm/i915/intel_crt.c
@@ -199,10 +199,14 @@ static int intel_crt_mode_valid(struct drm_connector *connector,
 	return MODE_OK;
 }
 
-static bool intel_crt_mode_fixup(struct drm_encoder *encoder,
-				 const struct drm_display_mode *mode,
-				 struct drm_display_mode *adjusted_mode)
+static bool intel_crt_compute_config(struct intel_encoder *encoder,
+				     struct intel_crtc_config *pipe_config)
 {
+	struct drm_device *dev = encoder->base.dev;
+
+	if (HAS_PCH_SPLIT(dev))
+		pipe_config->has_pch_encoder = true;
+
 	return true;
 }
 
@@ -676,7 +680,6 @@ static void intel_crt_reset(struct drm_connector *connector)
  */
 
 static const struct drm_encoder_helper_funcs crt_encoder_funcs = {
-	.mode_fixup = intel_crt_mode_fixup,
 	.mode_set = intel_crt_mode_set,
 };
 
@@ -768,8 +771,11 @@ void intel_crt_init(struct drm_device *dev)
 	else
 		crt->adpa_reg = ADPA;
 
+	crt->base.compute_config = intel_crt_compute_config;
 	crt->base.disable = intel_disable_crt;
 	crt->base.enable = intel_enable_crt;
+	if (I915_HAS_HOTPLUG(dev))
+		crt->base.hpd_pin = HPD_CRT;
 	if (HAS_DDI(dev))
 		crt->base.get_hw_state = intel_ddi_get_hw_state;
 	else
@@ -781,18 +787,14 @@ void intel_crt_init(struct drm_device *dev)
 
 	drm_sysfs_connector_add(connector);
 
-	if (I915_HAS_HOTPLUG(dev))
-		connector->polled = DRM_CONNECTOR_POLL_HPD;
-	else
-		connector->polled = DRM_CONNECTOR_POLL_CONNECT;
+	if (!I915_HAS_HOTPLUG(dev))
+		intel_connector->polled = DRM_CONNECTOR_POLL_CONNECT;
 
 	/*
 	 * Configure the automatic hotplug detection stuff
 	 */
 	crt->force_hotplug_required = 0;
 
-	dev_priv->hotplug_supported_mask |= CRT_HOTPLUG_INT_STATUS;
-
 	/*
 	 * TODO: find a proper way to discover whether we need to set the the
 	 * polarity and link reversal bits or not, instead of relying on the
diff --git a/drivers/gpu/drm/i915/intel_ddi.c b/drivers/gpu/drm/i915/intel_ddi.c
index 8d0bac3c35d7..26a0a570f92e 100644
--- a/drivers/gpu/drm/i915/intel_ddi.c
+++ b/drivers/gpu/drm/i915/intel_ddi.c
@@ -898,6 +898,9 @@ bool intel_ddi_pll_mode_set(struct drm_crtc *crtc, int clock)
 			plls->spll_refcount++;
 			reg = SPLL_CTL;
 			intel_crtc->ddi_pll_sel = PORT_CLK_SEL_SPLL;
+		} else {
+			DRM_ERROR("SPLL already in use\n");
+			return false;
 		}
 
 		WARN(I915_READ(reg) & SPLL_PLL_ENABLE,
@@ -921,14 +924,14 @@ void intel_ddi_set_pipe_settings(struct drm_crtc *crtc)
 	struct drm_i915_private *dev_priv = crtc->dev->dev_private;
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 	struct intel_encoder *intel_encoder = intel_ddi_get_crtc_encoder(crtc);
-	enum transcoder cpu_transcoder = intel_crtc->cpu_transcoder;
+	enum transcoder cpu_transcoder = intel_crtc->config.cpu_transcoder;
 	int type = intel_encoder->type;
 	uint32_t temp;
 
 	if (type == INTEL_OUTPUT_DISPLAYPORT || type == INTEL_OUTPUT_EDP) {
 
 		temp = TRANS_MSA_SYNC_CLK;
-		switch (intel_crtc->bpp) {
+		switch (intel_crtc->config.pipe_bpp) {
 		case 18:
 			temp |= TRANS_MSA_6_BPC;
 			break;
@@ -942,22 +945,20 @@ void intel_ddi_set_pipe_settings(struct drm_crtc *crtc)
 			temp |= TRANS_MSA_12_BPC;
 			break;
 		default:
-			temp |= TRANS_MSA_8_BPC;
-			WARN(1, "%d bpp unsupported by DDI function\n",
-			     intel_crtc->bpp);
+			BUG();
 		}
 		I915_WRITE(TRANS_MSA_MISC(cpu_transcoder), temp);
 	}
 }
 
-void intel_ddi_enable_pipe_func(struct drm_crtc *crtc)
+void intel_ddi_enable_transcoder_func(struct drm_crtc *crtc)
 {
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 	struct intel_encoder *intel_encoder = intel_ddi_get_crtc_encoder(crtc);
 	struct drm_encoder *encoder = &intel_encoder->base;
 	struct drm_i915_private *dev_priv = crtc->dev->dev_private;
 	enum pipe pipe = intel_crtc->pipe;
-	enum transcoder cpu_transcoder = intel_crtc->cpu_transcoder;
+	enum transcoder cpu_transcoder = intel_crtc->config.cpu_transcoder;
 	enum port port = intel_ddi_get_encoder_port(intel_encoder);
 	int type = intel_encoder->type;
 	uint32_t temp;
@@ -966,7 +967,7 @@ void intel_ddi_enable_pipe_func(struct drm_crtc *crtc)
 	temp = TRANS_DDI_FUNC_ENABLE;
 	temp |= TRANS_DDI_SELECT_PORT(port);
 
-	switch (intel_crtc->bpp) {
+	switch (intel_crtc->config.pipe_bpp) {
 	case 18:
 		temp |= TRANS_DDI_BPC_6;
 		break;
@@ -980,8 +981,7 @@ void intel_ddi_enable_pipe_func(struct drm_crtc *crtc)
 		temp |= TRANS_DDI_BPC_12;
 		break;
 	default:
-		WARN(1, "%d bpp unsupported by transcoder DDI function\n",
-		     intel_crtc->bpp);
+		BUG();
 	}
 
 	if (crtc->mode.flags & DRM_MODE_FLAG_PVSYNC)
@@ -1150,14 +1150,14 @@ bool intel_ddi_get_hw_state(struct intel_encoder *encoder,
 
 	DRM_DEBUG_KMS("No pipe for ddi port %i found\n", port);
 
-	return true;
+	return false;
 }
 
 static uint32_t intel_ddi_get_crtc_pll(struct drm_i915_private *dev_priv,
 				       enum pipe pipe)
 {
 	uint32_t temp, ret;
-	enum port port;
+	enum port port = I915_MAX_PORTS;
 	enum transcoder cpu_transcoder = intel_pipe_to_cpu_transcoder(dev_priv,
 								      pipe);
 	int i;
@@ -1173,10 +1173,16 @@ static uint32_t intel_ddi_get_crtc_pll(struct drm_i915_private *dev_priv,
 				port = i;
 	}
 
-	ret = I915_READ(PORT_CLK_SEL(port));
-
-	DRM_DEBUG_KMS("Pipe %c connected to port %c using clock 0x%08x\n",
-		      pipe_name(pipe), port_name(port), ret);
+	if (port == I915_MAX_PORTS) {
+		WARN(1, "Pipe %c enabled on an unknown port\n",
+		     pipe_name(pipe));
+		ret = PORT_CLK_SEL_NONE;
+	} else {
+		ret = I915_READ(PORT_CLK_SEL(port));
+		DRM_DEBUG_KMS("Pipe %c connected to port %c using clock "
+			      "0x%08x\n", pipe_name(pipe), port_name(port),
+			      ret);
+	}
 
 	return ret;
 }
@@ -1217,7 +1223,7 @@ void intel_ddi_enable_pipe_clock(struct intel_crtc *intel_crtc)
 	struct drm_i915_private *dev_priv = crtc->dev->dev_private;
 	struct intel_encoder *intel_encoder = intel_ddi_get_crtc_encoder(crtc);
 	enum port port = intel_ddi_get_encoder_port(intel_encoder);
-	enum transcoder cpu_transcoder = intel_crtc->cpu_transcoder;
+	enum transcoder cpu_transcoder = intel_crtc->config.cpu_transcoder;
 
 	if (cpu_transcoder != TRANSCODER_EDP)
 		I915_WRITE(TRANS_CLK_SEL(cpu_transcoder),
@@ -1227,7 +1233,7 @@ void intel_ddi_enable_pipe_clock(struct intel_crtc *intel_crtc)
 void intel_ddi_disable_pipe_clock(struct intel_crtc *intel_crtc)
 {
 	struct drm_i915_private *dev_priv = intel_crtc->base.dev->dev_private;
-	enum transcoder cpu_transcoder = intel_crtc->cpu_transcoder;
+	enum transcoder cpu_transcoder = intel_crtc->config.cpu_transcoder;
 
 	if (cpu_transcoder != TRANSCODER_EDP)
 		I915_WRITE(TRANS_CLK_SEL(cpu_transcoder),
@@ -1341,15 +1347,15 @@ static void intel_disable_ddi(struct intel_encoder *intel_encoder)
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	uint32_t tmp;
 
+	tmp = I915_READ(HSW_AUD_PIN_ELD_CP_VLD);
+	tmp &= ~((AUDIO_OUTPUT_ENABLE_A | AUDIO_ELD_VALID_A) << (pipe * 4));
+	I915_WRITE(HSW_AUD_PIN_ELD_CP_VLD, tmp);
+
 	if (type == INTEL_OUTPUT_EDP) {
 		struct intel_dp *intel_dp = enc_to_intel_dp(encoder);
 
 		ironlake_edp_backlight_off(intel_dp);
 	}
-
-	tmp = I915_READ(HSW_AUD_PIN_ELD_CP_VLD);
-	tmp &= ~((AUDIO_OUTPUT_ENABLE_A | AUDIO_ELD_VALID_A) << (pipe * 4));
-	I915_WRITE(HSW_AUD_PIN_ELD_CP_VLD, tmp);
 }
 
 int intel_ddi_get_cdclk_freq(struct drm_i915_private *dev_priv)
@@ -1467,19 +1473,17 @@ static void intel_ddi_destroy(struct drm_encoder *encoder)
 	intel_dp_encoder_destroy(encoder);
 }
 
-static bool intel_ddi_mode_fixup(struct drm_encoder *encoder,
-				 const struct drm_display_mode *mode,
-				 struct drm_display_mode *adjusted_mode)
+static bool intel_ddi_compute_config(struct intel_encoder *encoder,
+				     struct intel_crtc_config *pipe_config)
 {
-	struct intel_encoder *intel_encoder = to_intel_encoder(encoder);
-	int type = intel_encoder->type;
+	int type = encoder->type;
 
-	WARN(type == INTEL_OUTPUT_UNKNOWN, "mode_fixup() on unknown output!\n");
+	WARN(type == INTEL_OUTPUT_UNKNOWN, "compute_config() on unknown output!\n");
 
 	if (type == INTEL_OUTPUT_HDMI)
-		return intel_hdmi_mode_fixup(encoder, mode, adjusted_mode);
+		return intel_hdmi_compute_config(encoder, pipe_config);
 	else
-		return intel_dp_mode_fixup(encoder, mode, adjusted_mode);
+		return intel_dp_compute_config(encoder, pipe_config);
 }
 
 static const struct drm_encoder_funcs intel_ddi_funcs = {
@@ -1487,7 +1491,6 @@ static const struct drm_encoder_funcs intel_ddi_funcs = {
 };
 
 static const struct drm_encoder_helper_funcs intel_ddi_helper_funcs = {
-	.mode_fixup = intel_ddi_mode_fixup,
 	.mode_set = intel_ddi_mode_set,
 };
 
@@ -1527,6 +1530,7 @@ void intel_ddi_init(struct drm_device *dev, enum port port)
 			 DRM_MODE_ENCODER_TMDS);
 	drm_encoder_helper_add(encoder, &intel_ddi_helper_funcs);
 
+	intel_encoder->compute_config = intel_ddi_compute_config;
 	intel_encoder->enable = intel_enable_ddi;
 	intel_encoder->pre_enable = intel_ddi_pre_enable;
 	intel_encoder->disable = intel_disable_ddi;
@@ -1537,9 +1541,7 @@ void intel_ddi_init(struct drm_device *dev, enum port port)
 	intel_dig_port->port_reversal = I915_READ(DDI_BUF_CTL(port)) &
 					DDI_BUF_PORT_REVERSAL;
 	if (hdmi_connector)
-		intel_dig_port->hdmi.sdvox_reg = DDI_BUF_CTL(port);
-	else
-		intel_dig_port->hdmi.sdvox_reg = 0;
+		intel_dig_port->hdmi.hdmi_reg = DDI_BUF_CTL(port);
 	intel_dig_port->dp.output_reg = DDI_BUF_CTL(port);
 
 	intel_encoder->type = INTEL_OUTPUT_UNKNOWN;
diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
index b20d50192fcc..efe829919755 100644
--- a/drivers/gpu/drm/i915/intel_display.c
+++ b/drivers/gpu/drm/i915/intel_display.c
@@ -71,8 +71,24 @@ typedef struct intel_limit intel_limit_t;
 struct intel_limit {
 	intel_range_t   dot, vco, n, m, m1, m2, p, p1;
 	intel_p2_t	    p2;
-	bool (* find_pll)(const intel_limit_t *, struct drm_crtc *,
-			int, int, intel_clock_t *, intel_clock_t *);
+	/**
+	 * find_pll() - Find the best values for the PLL
+	 * @limit: limits for the PLL
+	 * @crtc: current CRTC
+	 * @target: target frequency in kHz
+	 * @refclk: reference clock frequency in kHz
+	 * @match_clock: if provided, @best_clock P divider must
+	 *               match the P divider from @match_clock
+	 *               used for LVDS downclocking
+	 * @best_clock: best PLL values found
+	 *
+	 * Returns true on success, false on failure.
+	 */
+	bool (*find_pll)(const intel_limit_t *limit,
+			 struct drm_crtc *crtc,
+			 int target, int refclk,
+			 intel_clock_t *match_clock,
+			 intel_clock_t *best_clock);
 };
 
 /* FDI */
@@ -471,7 +487,6 @@ static const intel_limit_t *intel_ironlake_limit(struct drm_crtc *crtc,
 
 	if (intel_pipe_has_type(crtc, INTEL_OUTPUT_LVDS)) {
 		if (intel_is_dual_link_lvds(dev)) {
-			/* LVDS dual channel */
 			if (refclk == 100000)
 				limit = &intel_limits_ironlake_dual_lvds_100m;
 			else
@@ -498,10 +513,8 @@ static const intel_limit_t *intel_g4x_limit(struct drm_crtc *crtc)
 
 	if (intel_pipe_has_type(crtc, INTEL_OUTPUT_LVDS)) {
 		if (intel_is_dual_link_lvds(dev))
-			/* LVDS with dual channel */
 			limit = &intel_limits_g4x_dual_channel_lvds;
 		else
-			/* LVDS with dual channel */
 			limit = &intel_limits_g4x_single_channel_lvds;
 	} else if (intel_pipe_has_type(crtc, INTEL_OUTPUT_HDMI) ||
 		   intel_pipe_has_type(crtc, INTEL_OUTPUT_ANALOG)) {
@@ -879,7 +892,7 @@ enum transcoder intel_pipe_to_cpu_transcoder(struct drm_i915_private *dev_priv,
 	struct drm_crtc *crtc = dev_priv->pipe_to_crtc_mapping[pipe];
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 
-	return intel_crtc->cpu_transcoder;
+	return intel_crtc->config.cpu_transcoder;
 }
 
 static void ironlake_wait_for_vblank(struct drm_device *dev, int pipe)
@@ -1214,8 +1227,8 @@ void assert_pipe(struct drm_i915_private *dev_priv,
 	if (pipe == PIPE_A && dev_priv->quirks & QUIRK_PIPEA_FORCE)
 		state = true;
 
-	if (IS_HASWELL(dev_priv->dev) && cpu_transcoder != TRANSCODER_EDP &&
-	    !(I915_READ(HSW_PWR_WELL_DRIVER) & HSW_PWR_WELL_ENABLE)) {
+	if (!intel_using_power_well(dev_priv->dev) &&
+	    cpu_transcoder != TRANSCODER_EDP) {
 		cur_state = false;
 	} else {
 		reg = PIPECONF(cpu_transcoder);
@@ -1254,7 +1267,7 @@ static void assert_planes_disabled(struct drm_i915_private *dev_priv,
 	int cur_pipe;
 
 	/* Planes are fixed to pipes on ILK+ */
-	if (HAS_PCH_SPLIT(dev_priv->dev)) {
+	if (HAS_PCH_SPLIT(dev_priv->dev) || IS_VALLEYVIEW(dev_priv->dev)) {
 		reg = DSPCNTR(pipe);
 		val = I915_READ(reg);
 		WARN((val & DISPLAY_PLANE_ENABLE),
@@ -1275,6 +1288,25 @@ static void assert_planes_disabled(struct drm_i915_private *dev_priv,
 	}
 }
 
+static void assert_sprites_disabled(struct drm_i915_private *dev_priv,
+				    enum pipe pipe)
+{
+	int reg, i;
+	u32 val;
+
+	if (!IS_VALLEYVIEW(dev_priv->dev))
+		return;
+
+	/* Need to check both planes against the pipe */
+	for (i = 0; i < dev_priv->num_plane; i++) {
+		reg = SPCNTR(pipe, i);
+		val = I915_READ(reg);
+		WARN((val & SP_ENABLE),
+		     "sprite %d assertion failure, should be off on pipe %c but is still active\n",
+		     pipe * 2 + i, pipe_name(pipe));
+	}
+}
+
 static void assert_pch_refclk_enabled(struct drm_i915_private *dev_priv)
 {
 	u32 val;
@@ -1327,14 +1359,14 @@ static bool dp_pipe_enabled(struct drm_i915_private *dev_priv,
 static bool hdmi_pipe_enabled(struct drm_i915_private *dev_priv,
 			      enum pipe pipe, u32 val)
 {
-	if ((val & PORT_ENABLE) == 0)
+	if ((val & SDVO_ENABLE) == 0)
 		return false;
 
 	if (HAS_PCH_CPT(dev_priv->dev)) {
-		if ((val & PORT_TRANS_SEL_MASK) != PORT_TRANS_SEL_CPT(pipe))
+		if ((val & SDVO_PIPE_SEL_MASK_CPT) != SDVO_PIPE_SEL_CPT(pipe))
 			return false;
 	} else {
-		if ((val & TRANSCODER_MASK) != TRANSCODER(pipe))
+		if ((val & SDVO_PIPE_SEL_MASK) != SDVO_PIPE_SEL(pipe))
 			return false;
 	}
 	return true;
@@ -1392,7 +1424,7 @@ static void assert_pch_hdmi_disabled(struct drm_i915_private *dev_priv,
 	     "PCH HDMI (0x%08x) enabled on transcoder %c, should be disabled\n",
 	     reg, pipe_name(pipe));
 
-	WARN(HAS_PCH_IBX(dev_priv->dev) && (val & PORT_ENABLE) == 0
+	WARN(HAS_PCH_IBX(dev_priv->dev) && (val & SDVO_ENABLE) == 0
 	     && (val & SDVO_PIPE_B_SELECT),
 	     "IBX PCH hdmi port still using transcoder B\n");
 }
@@ -1419,9 +1451,9 @@ static void assert_pch_ports_disabled(struct drm_i915_private *dev_priv,
 	     "PCH LVDS enabled on transcoder %c, should be disabled\n",
 	     pipe_name(pipe));
 
-	assert_pch_hdmi_disabled(dev_priv, pipe, HDMIB);
-	assert_pch_hdmi_disabled(dev_priv, pipe, HDMIC);
-	assert_pch_hdmi_disabled(dev_priv, pipe, HDMID);
+	assert_pch_hdmi_disabled(dev_priv, pipe, PCH_HDMIB);
+	assert_pch_hdmi_disabled(dev_priv, pipe, PCH_HDMIC);
+	assert_pch_hdmi_disabled(dev_priv, pipe, PCH_HDMID);
 }
 
 /**
@@ -1859,6 +1891,7 @@ static void intel_disable_pipe(struct drm_i915_private *dev_priv,
 	 * or we might hang the display.
 	 */
 	assert_planes_disabled(dev_priv, pipe);
+	assert_sprites_disabled(dev_priv, pipe);
 
 	/* Don't disable pipe A or pipe A PLLs if needed */
 	if (pipe == PIPE_A && (dev_priv->quirks & QUIRK_PIPEA_FORCE))
@@ -1937,6 +1970,15 @@ static void intel_disable_plane(struct drm_i915_private *dev_priv,
 	intel_wait_for_vblank(dev_priv->dev, pipe);
 }
 
+static bool need_vtd_wa(struct drm_device *dev)
+{
+#ifdef CONFIG_INTEL_IOMMU
+	if (INTEL_INFO(dev)->gen >= 6 && intel_iommu_gfx_mapped)
+		return true;
+#endif
+	return false;
+}
+
 int
 intel_pin_and_fence_fb_obj(struct drm_device *dev,
 			   struct drm_i915_gem_object *obj,
@@ -1960,13 +2002,23 @@ intel_pin_and_fence_fb_obj(struct drm_device *dev,
 		alignment = 0;
 		break;
 	case I915_TILING_Y:
-		/* FIXME: Is this true? */
-		DRM_ERROR("Y tiled not allowed for scan out buffers\n");
+		/* Despite that we check this in framebuffer_init userspace can
+		 * screw us over and change the tiling after the fact. Only
+		 * pinned buffers can't change their tiling. */
+		DRM_DEBUG_DRIVER("Y tiled not allowed for scan out buffers\n");
 		return -EINVAL;
 	default:
 		BUG();
 	}
 
+	/* Note that the w/a also requires 64 PTE of padding following the
+	 * bo. We currently fill all unused PTE with the shadow page and so
+	 * we should always have valid PTE following the scanout preventing
+	 * the VT-d warning.
+	 */
+	if (need_vtd_wa(dev) && alignment < 256 * 1024)
+		alignment = 256 * 1024;
+
 	dev_priv->mm.interruptible = false;
 	ret = i915_gem_object_pin_to_display_plane(obj, alignment, pipelined);
 	if (ret)
@@ -2083,8 +2135,7 @@ static int i9xx_update_plane(struct drm_crtc *crtc, struct drm_framebuffer *fb,
 		dspcntr |= DISPPLANE_RGBX101010;
 		break;
 	default:
-		DRM_ERROR("Unknown pixel format 0x%08x\n", fb->pixel_format);
-		return -EINVAL;
+		BUG();
 	}
 
 	if (INTEL_INFO(dev)->gen >= 4) {
@@ -2177,8 +2228,7 @@ static int ironlake_update_plane(struct drm_crtc *crtc,
 		dspcntr |= DISPPLANE_RGBX101010;
 		break;
 	default:
-		DRM_ERROR("Unknown pixel format 0x%08x\n", fb->pixel_format);
-		return -EINVAL;
+		BUG();
 	}
 
 	if (obj->tiling_mode != I915_TILING_NONE)
@@ -2229,6 +2279,44 @@ intel_pipe_set_base_atomic(struct drm_crtc *crtc, struct drm_framebuffer *fb,
 	return dev_priv->display.update_plane(crtc, fb, x, y);
 }
 
+void intel_display_handle_reset(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_crtc *crtc;
+
+	/*
+	 * Flips in the rings have been nuked by the reset,
+	 * so complete all pending flips so that user space
+	 * will get its events and not get stuck.
+	 *
+	 * Also update the base address of all primary
+	 * planes to the the last fb to make sure we're
+	 * showing the correct fb after a reset.
+	 *
+	 * Need to make two loops over the crtcs so that we
+	 * don't try to grab a crtc mutex before the
+	 * pending_flip_queue really got woken up.
+	 */
+
+	list_for_each_entry(crtc, &dev->mode_config.crtc_list, head) {
+		struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
+		enum plane plane = intel_crtc->plane;
+
+		intel_prepare_page_flip(dev, plane);
+		intel_finish_page_flip_plane(dev, plane);
+	}
+
+	list_for_each_entry(crtc, &dev->mode_config.crtc_list, head) {
+		struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
+
+		mutex_lock(&crtc->mutex);
+		if (intel_crtc->active)
+			dev_priv->display.update_plane(crtc, crtc->fb,
+						       crtc->x, crtc->y);
+		mutex_unlock(&crtc->mutex);
+	}
+}
+
 static int
 intel_finish_fb(struct drm_framebuffer *old_fb)
 {
@@ -2295,10 +2383,10 @@ intel_pipe_set_base(struct drm_crtc *crtc, int x, int y,
 		return 0;
 	}
 
-	if(intel_crtc->plane > dev_priv->num_pipe) {
+	if (intel_crtc->plane > INTEL_INFO(dev)->num_pipes) {
 		DRM_ERROR("no plane for crtc: plane %d, num_pipes %d\n",
 				intel_crtc->plane,
-				dev_priv->num_pipe);
+				INTEL_INFO(dev)->num_pipes);
 		return -EINVAL;
 	}
 
@@ -2312,9 +2400,6 @@ intel_pipe_set_base(struct drm_crtc *crtc, int x, int y,
 		return ret;
 	}
 
-	if (crtc->fb)
-		intel_finish_fb(crtc->fb);
-
 	ret = dev_priv->display.update_plane(crtc, fb, x, y);
 	if (ret) {
 		intel_unpin_fb_obj(to_intel_framebuffer(fb)->obj);
@@ -2912,32 +2997,6 @@ static void intel_crtc_wait_for_pending_flips(struct drm_crtc *crtc)
 	mutex_unlock(&dev->struct_mutex);
 }
 
-static bool ironlake_crtc_driving_pch(struct drm_crtc *crtc)
-{
-	struct drm_device *dev = crtc->dev;
-	struct intel_encoder *intel_encoder;
-
-	/*
-	 * If there's a non-PCH eDP on this crtc, it must be DP_A, and that
-	 * must be driven by its own crtc; no sharing is possible.
-	 */
-	for_each_encoder_on_crtc(dev, crtc, intel_encoder) {
-		switch (intel_encoder->type) {
-		case INTEL_OUTPUT_EDP:
-			if (!intel_encoder_is_pch_edp(&intel_encoder->base))
-				return false;
-			continue;
-		}
-	}
-
-	return true;
-}
-
-static bool haswell_crtc_driving_pch(struct drm_crtc *crtc)
-{
-	return intel_pipe_has_type(crtc, INTEL_OUTPUT_ANALOG);
-}
-
 /* Program iCLKIP clock to the desired frequency */
 static void lpt_program_iclkip(struct drm_crtc *crtc)
 {
@@ -3144,7 +3203,7 @@ static void lpt_pch_enable(struct drm_crtc *crtc)
 	struct drm_device *dev = crtc->dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
-	enum transcoder cpu_transcoder = intel_crtc->cpu_transcoder;
+	enum transcoder cpu_transcoder = intel_crtc->config.cpu_transcoder;
 
 	assert_transcoder_disabled(dev_priv, TRANSCODER_A);
 
@@ -3273,7 +3332,6 @@ static void ironlake_crtc_enable(struct drm_crtc *crtc)
 	int pipe = intel_crtc->pipe;
 	int plane = intel_crtc->plane;
 	u32 temp;
-	bool is_pch_port;
 
 	WARN_ON(!crtc->enabled);
 
@@ -3289,9 +3347,8 @@ static void ironlake_crtc_enable(struct drm_crtc *crtc)
 			I915_WRITE(PCH_LVDS, temp | LVDS_PORT_EN);
 	}
 
-	is_pch_port = ironlake_crtc_driving_pch(crtc);
 
-	if (is_pch_port) {
+	if (intel_crtc->config.has_pch_encoder) {
 		/* Note: FDI PLL enabling _must_ be done before we enable the
 		 * cpu pipes, hence this is separate from all the other fdi/pch
 		 * enabling. */
@@ -3328,10 +3385,11 @@ static void ironlake_crtc_enable(struct drm_crtc *crtc)
 	 */
 	intel_crtc_load_lut(crtc);
 
-	intel_enable_pipe(dev_priv, pipe, is_pch_port);
+	intel_enable_pipe(dev_priv, pipe,
+			  intel_crtc->config.has_pch_encoder);
 	intel_enable_plane(dev_priv, plane, pipe);
 
-	if (is_pch_port)
+	if (intel_crtc->config.has_pch_encoder)
 		ironlake_pch_enable(crtc);
 
 	mutex_lock(&dev->struct_mutex);
@@ -3365,7 +3423,6 @@ static void haswell_crtc_enable(struct drm_crtc *crtc)
 	struct intel_encoder *encoder;
 	int pipe = intel_crtc->pipe;
 	int plane = intel_crtc->plane;
-	bool is_pch_port;
 
 	WARN_ON(!crtc->enabled);
 
@@ -3375,9 +3432,7 @@ static void haswell_crtc_enable(struct drm_crtc *crtc)
 	intel_crtc->active = true;
 	intel_update_watermarks(dev);
 
-	is_pch_port = haswell_crtc_driving_pch(crtc);
-
-	if (is_pch_port)
+	if (intel_crtc->config.has_pch_encoder)
 		dev_priv->display.fdi_link_train(crtc);
 
 	for_each_encoder_on_crtc(dev, crtc, encoder)
@@ -3406,12 +3461,13 @@ static void haswell_crtc_enable(struct drm_crtc *crtc)
 	intel_crtc_load_lut(crtc);
 
 	intel_ddi_set_pipe_settings(crtc);
-	intel_ddi_enable_pipe_func(crtc);
+	intel_ddi_enable_transcoder_func(crtc);
 
-	intel_enable_pipe(dev_priv, pipe, is_pch_port);
+	intel_enable_pipe(dev_priv, pipe,
+			  intel_crtc->config.has_pch_encoder);
 	intel_enable_plane(dev_priv, plane, pipe);
 
-	if (is_pch_port)
+	if (intel_crtc->config.has_pch_encoder)
 		lpt_pch_enable(crtc);
 
 	mutex_lock(&dev->struct_mutex);
@@ -3522,14 +3578,11 @@ static void haswell_crtc_disable(struct drm_crtc *crtc)
 	struct intel_encoder *encoder;
 	int pipe = intel_crtc->pipe;
 	int plane = intel_crtc->plane;
-	enum transcoder cpu_transcoder = intel_crtc->cpu_transcoder;
-	bool is_pch_port;
+	enum transcoder cpu_transcoder = intel_crtc->config.cpu_transcoder;
 
 	if (!intel_crtc->active)
 		return;
 
-	is_pch_port = haswell_crtc_driving_pch(crtc);
-
 	for_each_encoder_on_crtc(dev, crtc, encoder)
 		encoder->disable(encoder);
 
@@ -3546,9 +3599,13 @@ static void haswell_crtc_disable(struct drm_crtc *crtc)
 
 	intel_ddi_disable_transcoder_func(dev_priv, cpu_transcoder);
 
-	/* Disable PF */
-	I915_WRITE(PF_CTL(pipe), 0);
-	I915_WRITE(PF_WIN_SZ(pipe), 0);
+	/* XXX: Once we have proper panel fitter state tracking implemented with
+	 * hardware state read/check support we should switch to only disable
+	 * the panel fitter when we know it's used. */
+	if (intel_using_power_well(dev)) {
+		I915_WRITE(PF_CTL(pipe), 0);
+		I915_WRITE(PF_WIN_SZ(pipe), 0);
+	}
 
 	intel_ddi_disable_pipe_clock(intel_crtc);
 
@@ -3556,7 +3613,7 @@ static void haswell_crtc_disable(struct drm_crtc *crtc)
 		if (encoder->post_disable)
 			encoder->post_disable(encoder);
 
-	if (is_pch_port) {
+	if (intel_crtc->config.has_pch_encoder) {
 		lpt_disable_pch_transcoder(dev_priv);
 		intel_ddi_fdi_disable(crtc);
 	}
@@ -3581,7 +3638,7 @@ static void haswell_crtc_off(struct drm_crtc *crtc)
 
 	/* Stop saying we're using TRANSCODER_EDP because some other CRTC might
 	 * start using it. */
-	intel_crtc->cpu_transcoder = (enum transcoder) intel_crtc->pipe;
+	intel_crtc->config.cpu_transcoder = (enum transcoder) intel_crtc->pipe;
 
 	intel_ddi_put_crtc_pll(crtc);
 }
@@ -3667,6 +3724,26 @@ static void i9xx_crtc_enable(struct drm_crtc *crtc)
 		encoder->enable(encoder);
 }
 
+static void i9xx_pfit_disable(struct intel_crtc *crtc)
+{
+	struct drm_device *dev = crtc->base.dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	enum pipe pipe;
+	uint32_t pctl = I915_READ(PFIT_CONTROL);
+
+	assert_pipe_disabled(dev_priv, crtc->pipe);
+
+	if (INTEL_INFO(dev)->gen >= 4)
+		pipe = (pctl & PFIT_PIPE_MASK) >> PFIT_PIPE_SHIFT;
+	else
+		pipe = PIPE_B;
+
+	if (pipe == crtc->pipe) {
+		DRM_DEBUG_DRIVER("disabling pfit, current: 0x%08x\n", pctl);
+		I915_WRITE(PFIT_CONTROL, 0);
+	}
+}
+
 static void i9xx_crtc_disable(struct drm_crtc *crtc)
 {
 	struct drm_device *dev = crtc->dev;
@@ -3675,8 +3752,6 @@ static void i9xx_crtc_disable(struct drm_crtc *crtc)
 	struct intel_encoder *encoder;
 	int pipe = intel_crtc->pipe;
 	int plane = intel_crtc->plane;
-	u32 pctl;
-
 
 	if (!intel_crtc->active)
 		return;
@@ -3696,11 +3771,7 @@ static void i9xx_crtc_disable(struct drm_crtc *crtc)
 	intel_disable_plane(dev_priv, plane, pipe);
 	intel_disable_pipe(dev_priv, pipe);
 
-	/* Disable pannel fitter if it is on this pipe. */
-	pctl = I915_READ(PFIT_CONTROL);
-	if ((pctl & PFIT_ENABLE) &&
-	    ((pctl & PFIT_PIPE_MASK) >> PFIT_PIPE_SHIFT) == pipe)
-		I915_WRITE(PFIT_CONTROL, 0);
+	i9xx_pfit_disable(intel_crtc);
 
 	intel_disable_pll(dev_priv, pipe);
 
@@ -3906,22 +3977,23 @@ bool intel_connector_get_hw_state(struct intel_connector *connector)
 	return encoder->get_hw_state(encoder, &pipe);
 }
 
-static bool intel_crtc_mode_fixup(struct drm_crtc *crtc,
-				  const struct drm_display_mode *mode,
-				  struct drm_display_mode *adjusted_mode)
+static bool intel_crtc_compute_config(struct drm_crtc *crtc,
+				      struct intel_crtc_config *pipe_config)
 {
 	struct drm_device *dev = crtc->dev;
+	struct drm_display_mode *adjusted_mode = &pipe_config->adjusted_mode;
 
 	if (HAS_PCH_SPLIT(dev)) {
 		/* FDI link clock is fixed at 2.7G */
-		if (mode->clock * 3 > IRONLAKE_FDI_FREQ * 4)
+		if (pipe_config->requested_mode.clock * 3
+		    > IRONLAKE_FDI_FREQ * 4)
 			return false;
 	}
 
 	/* All interlaced capable intel hw wants timings in frames. Note though
 	 * that intel_lvds_mode_fixup does some funny tricks with the crtc
 	 * timings, so we need to be careful not to clobber these.*/
-	if (!(adjusted_mode->private_flags & INTEL_MODE_CRTC_TIMINGS_SET))
+	if (!pipe_config->timings_set)
 		drm_mode_set_crtcinfo(adjusted_mode, 0);
 
 	/* WaPruneModeWithIncorrectHsyncOffset: Cantiga+ cannot handle modes
@@ -3931,6 +4003,14 @@ static bool intel_crtc_mode_fixup(struct drm_crtc *crtc,
 		adjusted_mode->hsync_start == adjusted_mode->hdisplay)
 		return false;
 
+	if ((IS_G4X(dev) || IS_VALLEYVIEW(dev)) && pipe_config->pipe_bpp > 10*3) {
+		pipe_config->pipe_bpp = 10*3; /* 12bpc is gen5+ */
+	} else if (INTEL_INFO(dev)->gen <= 4 && pipe_config->pipe_bpp > 8*3) {
+		/* only a 8bpc pipe, with 6bpc dither through the panel fitter
+		 * for lvds. */
+		pipe_config->pipe_bpp = 8*3;
+	}
+
 	return true;
 }
 
@@ -4004,26 +4084,36 @@ static int i830_get_display_clock_speed(struct drm_device *dev)
 }
 
 static void
-intel_reduce_ratio(uint32_t *num, uint32_t *den)
+intel_reduce_m_n_ratio(uint32_t *num, uint32_t *den)
 {
-	while (*num > 0xffffff || *den > 0xffffff) {
+	while (*num > DATA_LINK_M_N_MASK ||
+	       *den > DATA_LINK_M_N_MASK) {
 		*num >>= 1;
 		*den >>= 1;
 	}
 }
 
+static void compute_m_n(unsigned int m, unsigned int n,
+			uint32_t *ret_m, uint32_t *ret_n)
+{
+	*ret_n = min_t(unsigned int, roundup_pow_of_two(n), DATA_LINK_N_MAX);
+	*ret_m = div_u64((uint64_t) m * *ret_n, n);
+	intel_reduce_m_n_ratio(ret_m, ret_n);
+}
+
 void
 intel_link_compute_m_n(int bits_per_pixel, int nlanes,
 		       int pixel_clock, int link_clock,
 		       struct intel_link_m_n *m_n)
 {
 	m_n->tu = 64;
-	m_n->gmch_m = bits_per_pixel * pixel_clock;
-	m_n->gmch_n = link_clock * nlanes * 8;
-	intel_reduce_ratio(&m_n->gmch_m, &m_n->gmch_n);
-	m_n->link_m = pixel_clock;
-	m_n->link_n = link_clock;
-	intel_reduce_ratio(&m_n->link_m, &m_n->link_n);
+
+	compute_m_n(bits_per_pixel * pixel_clock,
+		    link_clock * nlanes * 8,
+		    &m_n->gmch_m, &m_n->gmch_n);
+
+	compute_m_n(pixel_clock, link_clock,
+		    &m_n->link_m, &m_n->link_n);
 }
 
 static inline bool intel_panel_use_ssc(struct drm_i915_private *dev_priv)
@@ -4034,142 +4124,6 @@ static inline bool intel_panel_use_ssc(struct drm_i915_private *dev_priv)
 		&& !(dev_priv->quirks & QUIRK_LVDS_SSC_DISABLE);
 }
 
-/**
- * intel_choose_pipe_bpp_dither - figure out what color depth the pipe should send
- * @crtc: CRTC structure
- * @mode: requested mode
- *
- * A pipe may be connected to one or more outputs.  Based on the depth of the
- * attached framebuffer, choose a good color depth to use on the pipe.
- *
- * If possible, match the pipe depth to the fb depth.  In some cases, this
- * isn't ideal, because the connected output supports a lesser or restricted
- * set of depths.  Resolve that here:
- *    LVDS typically supports only 6bpc, so clamp down in that case
- *    HDMI supports only 8bpc or 12bpc, so clamp to 8bpc with dither for 10bpc
- *    Displays may support a restricted set as well, check EDID and clamp as
- *      appropriate.
- *    DP may want to dither down to 6bpc to fit larger modes
- *
- * RETURNS:
- * Dithering requirement (i.e. false if display bpc and pipe bpc match,
- * true if they don't match).
- */
-static bool intel_choose_pipe_bpp_dither(struct drm_crtc *crtc,
-					 struct drm_framebuffer *fb,
-					 unsigned int *pipe_bpp,
-					 struct drm_display_mode *mode)
-{
-	struct drm_device *dev = crtc->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct drm_connector *connector;
-	struct intel_encoder *intel_encoder;
-	unsigned int display_bpc = UINT_MAX, bpc;
-
-	/* Walk the encoders & connectors on this crtc, get min bpc */
-	for_each_encoder_on_crtc(dev, crtc, intel_encoder) {
-
-		if (intel_encoder->type == INTEL_OUTPUT_LVDS) {
-			unsigned int lvds_bpc;
-
-			if ((I915_READ(PCH_LVDS) & LVDS_A3_POWER_MASK) ==
-			    LVDS_A3_POWER_UP)
-				lvds_bpc = 8;
-			else
-				lvds_bpc = 6;
-
-			if (lvds_bpc < display_bpc) {
-				DRM_DEBUG_KMS("clamping display bpc (was %d) to LVDS (%d)\n", display_bpc, lvds_bpc);
-				display_bpc = lvds_bpc;
-			}
-			continue;
-		}
-
-		/* Not one of the known troublemakers, check the EDID */
-		list_for_each_entry(connector, &dev->mode_config.connector_list,
-				    head) {
-			if (connector->encoder != &intel_encoder->base)
-				continue;
-
-			/* Don't use an invalid EDID bpc value */
-			if (connector->display_info.bpc &&
-			    connector->display_info.bpc < display_bpc) {
-				DRM_DEBUG_KMS("clamping display bpc (was %d) to EDID reported max of %d\n", display_bpc, connector->display_info.bpc);
-				display_bpc = connector->display_info.bpc;
-			}
-		}
-
-		if (intel_encoder->type == INTEL_OUTPUT_EDP) {
-			/* Use VBT settings if we have an eDP panel */
-			unsigned int edp_bpc = dev_priv->edp.bpp / 3;
-
-			if (edp_bpc && edp_bpc < display_bpc) {
-				DRM_DEBUG_KMS("clamping display bpc (was %d) to eDP (%d)\n", display_bpc, edp_bpc);
-				display_bpc = edp_bpc;
-			}
-			continue;
-		}
-
-		/*
-		 * HDMI is either 12 or 8, so if the display lets 10bpc sneak
-		 * through, clamp it down.  (Note: >12bpc will be caught below.)
-		 */
-		if (intel_encoder->type == INTEL_OUTPUT_HDMI) {
-			if (display_bpc > 8 && display_bpc < 12) {
-				DRM_DEBUG_KMS("forcing bpc to 12 for HDMI\n");
-				display_bpc = 12;
-			} else {
-				DRM_DEBUG_KMS("forcing bpc to 8 for HDMI\n");
-				display_bpc = 8;
-			}
-		}
-	}
-
-	if (mode->private_flags & INTEL_MODE_DP_FORCE_6BPC) {
-		DRM_DEBUG_KMS("Dithering DP to 6bpc\n");
-		display_bpc = 6;
-	}
-
-	/*
-	 * We could just drive the pipe at the highest bpc all the time and
-	 * enable dithering as needed, but that costs bandwidth.  So choose
-	 * the minimum value that expresses the full color range of the fb but
-	 * also stays within the max display bpc discovered above.
-	 */
-
-	switch (fb->depth) {
-	case 8:
-		bpc = 8; /* since we go through a colormap */
-		break;
-	case 15:
-	case 16:
-		bpc = 6; /* min is 18bpp */
-		break;
-	case 24:
-		bpc = 8;
-		break;
-	case 30:
-		bpc = 10;
-		break;
-	case 48:
-		bpc = 12;
-		break;
-	default:
-		DRM_DEBUG("unsupported depth, assuming 24 bits\n");
-		bpc = min((unsigned int)8, display_bpc);
-		break;
-	}
-
-	display_bpc = min(display_bpc, bpc);
-
-	DRM_DEBUG_KMS("setting pipe bpc to %d (max display bpc %d)\n",
-		      bpc, display_bpc);
-
-	*pipe_bpp = display_bpc * 3;
-
-	return display_bpc != bpc;
-}
-
 static int vlv_get_refclk(struct drm_crtc *crtc)
 {
 	struct drm_device *dev = crtc->dev;
@@ -4214,37 +4168,38 @@ static int i9xx_get_refclk(struct drm_crtc *crtc, int num_connectors)
 	return refclk;
 }
 
-static void i9xx_adjust_sdvo_tv_clock(struct drm_display_mode *adjusted_mode,
-				      intel_clock_t *clock)
+static void i9xx_adjust_sdvo_tv_clock(struct intel_crtc *crtc)
 {
+	unsigned dotclock = crtc->config.adjusted_mode.clock;
+	struct dpll *clock = &crtc->config.dpll;
+
 	/* SDVO TV has fixed PLL values depend on its clock range,
 	   this mirrors vbios setting. */
-	if (adjusted_mode->clock >= 100000
-	    && adjusted_mode->clock < 140500) {
+	if (dotclock >= 100000 && dotclock < 140500) {
 		clock->p1 = 2;
 		clock->p2 = 10;
 		clock->n = 3;
 		clock->m1 = 16;
 		clock->m2 = 8;
-	} else if (adjusted_mode->clock >= 140500
-		   && adjusted_mode->clock <= 200000) {
+	} else if (dotclock >= 140500 && dotclock <= 200000) {
 		clock->p1 = 1;
 		clock->p2 = 10;
 		clock->n = 6;
 		clock->m1 = 12;
 		clock->m2 = 8;
 	}
+
+	crtc->config.clock_set = true;
 }
 
-static void i9xx_update_pll_dividers(struct drm_crtc *crtc,
-				     intel_clock_t *clock,
+static void i9xx_update_pll_dividers(struct intel_crtc *crtc,
 				     intel_clock_t *reduced_clock)
 {
-	struct drm_device *dev = crtc->dev;
+	struct drm_device *dev = crtc->base.dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
-	int pipe = intel_crtc->pipe;
+	int pipe = crtc->pipe;
 	u32 fp, fp2 = 0;
+	struct dpll *clock = &crtc->config.dpll;
 
 	if (IS_PINEVIEW(dev)) {
 		fp = (1 << clock->n) << 16 | clock->m1 << 8 | clock->m2;
@@ -4260,26 +4215,29 @@ static void i9xx_update_pll_dividers(struct drm_crtc *crtc,
 
 	I915_WRITE(FP0(pipe), fp);
 
-	intel_crtc->lowfreq_avail = false;
-	if (intel_pipe_has_type(crtc, INTEL_OUTPUT_LVDS) &&
+	crtc->lowfreq_avail = false;
+	if (intel_pipe_has_type(&crtc->base, INTEL_OUTPUT_LVDS) &&
 	    reduced_clock && i915_powersave) {
 		I915_WRITE(FP1(pipe), fp2);
-		intel_crtc->lowfreq_avail = true;
+		crtc->lowfreq_avail = true;
 	} else {
 		I915_WRITE(FP1(pipe), fp);
 	}
 }
 
-static void vlv_update_pll(struct drm_crtc *crtc,
-			   struct drm_display_mode *mode,
-			   struct drm_display_mode *adjusted_mode,
-			   intel_clock_t *clock, intel_clock_t *reduced_clock,
-			   int num_connectors)
+static void intel_dp_set_m_n(struct intel_crtc *crtc)
 {
-	struct drm_device *dev = crtc->dev;
+	if (crtc->config.has_pch_encoder)
+		intel_pch_transcoder_set_m_n(crtc, &crtc->config.dp_m_n);
+	else
+		intel_cpu_transcoder_set_m_n(crtc, &crtc->config.dp_m_n);
+}
+
+static void vlv_update_pll(struct intel_crtc *crtc)
+{
+	struct drm_device *dev = crtc->base.dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
-	int pipe = intel_crtc->pipe;
+	int pipe = crtc->pipe;
 	u32 dpll, mdiv, pdiv;
 	u32 bestn, bestm1, bestm2, bestp1, bestp2;
 	bool is_sdvo;
@@ -4287,8 +4245,8 @@ static void vlv_update_pll(struct drm_crtc *crtc,
 
 	mutex_lock(&dev_priv->dpio_lock);
 
-	is_sdvo = intel_pipe_has_type(crtc, INTEL_OUTPUT_SDVO) ||
-		intel_pipe_has_type(crtc, INTEL_OUTPUT_HDMI);
+	is_sdvo = intel_pipe_has_type(&crtc->base, INTEL_OUTPUT_SDVO) ||
+		intel_pipe_has_type(&crtc->base, INTEL_OUTPUT_HDMI);
 
 	dpll = DPLL_VGA_MODE_DIS;
 	dpll |= DPLL_EXT_BUFFER_ENABLE_VLV;
@@ -4298,11 +4256,11 @@ static void vlv_update_pll(struct drm_crtc *crtc,
 	I915_WRITE(DPLL(pipe), dpll);
 	POSTING_READ(DPLL(pipe));
 
-	bestn = clock->n;
-	bestm1 = clock->m1;
-	bestm2 = clock->m2;
-	bestp1 = clock->p1;
-	bestp2 = clock->p2;
+	bestn = crtc->config.dpll.n;
+	bestm1 = crtc->config.dpll.m1;
+	bestm2 = crtc->config.dpll.m2;
+	bestp1 = crtc->config.dpll.p1;
+	bestp2 = crtc->config.dpll.p2;
 
 	/*
 	 * In Valleyview PLL and program lane counter registers are exposed
@@ -4334,8 +4292,8 @@ static void vlv_update_pll(struct drm_crtc *crtc,
 
 	intel_dpio_write(dev_priv, DPIO_FASTCLK_DISABLE, 0x620);
 
-	if (intel_pipe_has_type(crtc, INTEL_OUTPUT_DISPLAYPORT))
-		intel_dp_set_m_n(crtc, mode, adjusted_mode);
+	if (crtc->config.has_dp_encoder)
+		intel_dp_set_m_n(crtc);
 
 	I915_WRITE(DPLL(pipe), dpll);
 
@@ -4345,26 +4303,25 @@ static void vlv_update_pll(struct drm_crtc *crtc,
 
 	temp = 0;
 	if (is_sdvo) {
-		temp = intel_mode_get_pixel_multiplier(adjusted_mode);
-		if (temp > 1)
-			temp = (temp - 1) << DPLL_MD_UDI_MULTIPLIER_SHIFT;
-		else
-			temp = 0;
+		temp = 0;
+		if (crtc->config.pixel_multiplier > 1) {
+			temp = (crtc->config.pixel_multiplier - 1)
+				<< DPLL_MD_UDI_MULTIPLIER_SHIFT;
+		}
 	}
 	I915_WRITE(DPLL_MD(pipe), temp);
 	POSTING_READ(DPLL_MD(pipe));
 
 	/* Now program lane control registers */
-	if(intel_pipe_has_type(crtc, INTEL_OUTPUT_DISPLAYPORT)
-			|| intel_pipe_has_type(crtc, INTEL_OUTPUT_HDMI))
-	{
+	if(intel_pipe_has_type(&crtc->base, INTEL_OUTPUT_DISPLAYPORT)
+	   || intel_pipe_has_type(&crtc->base, INTEL_OUTPUT_HDMI)) {
 		temp = 0x1000C4;
 		if(pipe == 1)
 			temp |= (1 << 21);
 		intel_dpio_write(dev_priv, DPIO_DATA_CHANNEL1, temp);
 	}
-	if(intel_pipe_has_type(crtc,INTEL_OUTPUT_EDP))
-	{
+
+	if(intel_pipe_has_type(&crtc->base, INTEL_OUTPUT_EDP)) {
 		temp = 0x1000C4;
 		if(pipe == 1)
 			temp |= (1 << 21);
@@ -4374,40 +4331,39 @@ static void vlv_update_pll(struct drm_crtc *crtc,
 	mutex_unlock(&dev_priv->dpio_lock);
 }
 
-static void i9xx_update_pll(struct drm_crtc *crtc,
-			    struct drm_display_mode *mode,
-			    struct drm_display_mode *adjusted_mode,
-			    intel_clock_t *clock, intel_clock_t *reduced_clock,
+static void i9xx_update_pll(struct intel_crtc *crtc,
+			    intel_clock_t *reduced_clock,
 			    int num_connectors)
 {
-	struct drm_device *dev = crtc->dev;
+	struct drm_device *dev = crtc->base.dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 	struct intel_encoder *encoder;
-	int pipe = intel_crtc->pipe;
+	int pipe = crtc->pipe;
 	u32 dpll;
 	bool is_sdvo;
+	struct dpll *clock = &crtc->config.dpll;
 
-	i9xx_update_pll_dividers(crtc, clock, reduced_clock);
+	i9xx_update_pll_dividers(crtc, reduced_clock);
 
-	is_sdvo = intel_pipe_has_type(crtc, INTEL_OUTPUT_SDVO) ||
-		intel_pipe_has_type(crtc, INTEL_OUTPUT_HDMI);
+	is_sdvo = intel_pipe_has_type(&crtc->base, INTEL_OUTPUT_SDVO) ||
+		intel_pipe_has_type(&crtc->base, INTEL_OUTPUT_HDMI);
 
 	dpll = DPLL_VGA_MODE_DIS;
 
-	if (intel_pipe_has_type(crtc, INTEL_OUTPUT_LVDS))
+	if (intel_pipe_has_type(&crtc->base, INTEL_OUTPUT_LVDS))
 		dpll |= DPLLB_MODE_LVDS;
 	else
 		dpll |= DPLLB_MODE_DAC_SERIAL;
+
 	if (is_sdvo) {
-		int pixel_multiplier = intel_mode_get_pixel_multiplier(adjusted_mode);
-		if (pixel_multiplier > 1) {
-			if (IS_I945G(dev) || IS_I945GM(dev) || IS_G33(dev))
-				dpll |= (pixel_multiplier - 1) << SDVO_MULTIPLIER_SHIFT_HIRES;
+		if ((crtc->config.pixel_multiplier > 1) &&
+		    (IS_I945G(dev) || IS_I945GM(dev) || IS_G33(dev))) {
+			dpll |= (crtc->config.pixel_multiplier - 1)
+				<< SDVO_MULTIPLIER_SHIFT_HIRES;
 		}
 		dpll |= DPLL_DVO_HIGH_SPEED;
 	}
-	if (intel_pipe_has_type(crtc, INTEL_OUTPUT_DISPLAYPORT))
+	if (intel_pipe_has_type(&crtc->base, INTEL_OUTPUT_DISPLAYPORT))
 		dpll |= DPLL_DVO_HIGH_SPEED;
 
 	/* compute bitmask from p1 value */
@@ -4435,13 +4391,13 @@ static void i9xx_update_pll(struct drm_crtc *crtc,
 	if (INTEL_INFO(dev)->gen >= 4)
 		dpll |= (6 << PLL_LOAD_PULSE_PHASE_SHIFT);
 
-	if (is_sdvo && intel_pipe_has_type(crtc, INTEL_OUTPUT_TVOUT))
+	if (is_sdvo && intel_pipe_has_type(&crtc->base, INTEL_OUTPUT_TVOUT))
 		dpll |= PLL_REF_INPUT_TVCLKINBC;
-	else if (intel_pipe_has_type(crtc, INTEL_OUTPUT_TVOUT))
+	else if (intel_pipe_has_type(&crtc->base, INTEL_OUTPUT_TVOUT))
 		/* XXX: just matching BIOS for now */
 		/*	dpll |= PLL_REF_INPUT_TVCLKINBC; */
 		dpll |= 3;
-	else if (intel_pipe_has_type(crtc, INTEL_OUTPUT_LVDS) &&
+	else if (intel_pipe_has_type(&crtc->base, INTEL_OUTPUT_LVDS) &&
 		 intel_panel_use_ssc(dev_priv) && num_connectors < 2)
 		dpll |= PLLB_REF_INPUT_SPREADSPECTRUMIN;
 	else
@@ -4452,12 +4408,12 @@ static void i9xx_update_pll(struct drm_crtc *crtc,
 	POSTING_READ(DPLL(pipe));
 	udelay(150);
 
-	for_each_encoder_on_crtc(dev, crtc, encoder)
+	for_each_encoder_on_crtc(dev, &crtc->base, encoder)
 		if (encoder->pre_pll_enable)
 			encoder->pre_pll_enable(encoder);
 
-	if (intel_pipe_has_type(crtc, INTEL_OUTPUT_DISPLAYPORT))
-		intel_dp_set_m_n(crtc, mode, adjusted_mode);
+	if (crtc->config.has_dp_encoder)
+		intel_dp_set_m_n(crtc);
 
 	I915_WRITE(DPLL(pipe), dpll);
 
@@ -4468,11 +4424,11 @@ static void i9xx_update_pll(struct drm_crtc *crtc,
 	if (INTEL_INFO(dev)->gen >= 4) {
 		u32 temp = 0;
 		if (is_sdvo) {
-			temp = intel_mode_get_pixel_multiplier(adjusted_mode);
-			if (temp > 1)
-				temp = (temp - 1) << DPLL_MD_UDI_MULTIPLIER_SHIFT;
-			else
-				temp = 0;
+			temp = 0;
+			if (crtc->config.pixel_multiplier > 1) {
+				temp = (crtc->config.pixel_multiplier - 1)
+					<< DPLL_MD_UDI_MULTIPLIER_SHIFT;
+			}
 		}
 		I915_WRITE(DPLL_MD(pipe), temp);
 	} else {
@@ -4485,23 +4441,23 @@ static void i9xx_update_pll(struct drm_crtc *crtc,
 	}
 }
 
-static void i8xx_update_pll(struct drm_crtc *crtc,
+static void i8xx_update_pll(struct intel_crtc *crtc,
 			    struct drm_display_mode *adjusted_mode,
-			    intel_clock_t *clock, intel_clock_t *reduced_clock,
+			    intel_clock_t *reduced_clock,
 			    int num_connectors)
 {
-	struct drm_device *dev = crtc->dev;
+	struct drm_device *dev = crtc->base.dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 	struct intel_encoder *encoder;
-	int pipe = intel_crtc->pipe;
+	int pipe = crtc->pipe;
 	u32 dpll;
+	struct dpll *clock = &crtc->config.dpll;
 
-	i9xx_update_pll_dividers(crtc, clock, reduced_clock);
+	i9xx_update_pll_dividers(crtc, reduced_clock);
 
 	dpll = DPLL_VGA_MODE_DIS;
 
-	if (intel_pipe_has_type(crtc, INTEL_OUTPUT_LVDS)) {
+	if (intel_pipe_has_type(&crtc->base, INTEL_OUTPUT_LVDS)) {
 		dpll |= (1 << (clock->p1 - 1)) << DPLL_FPA01_P1_POST_DIV_SHIFT;
 	} else {
 		if (clock->p1 == 2)
@@ -4512,11 +4468,7 @@ static void i8xx_update_pll(struct drm_crtc *crtc,
 			dpll |= PLL_P2_DIVIDE_BY_4;
 	}
 
-	if (intel_pipe_has_type(crtc, INTEL_OUTPUT_TVOUT))
-		/* XXX: just matching BIOS for now */
-		/*	dpll |= PLL_REF_INPUT_TVCLKINBC; */
-		dpll |= 3;
-	else if (intel_pipe_has_type(crtc, INTEL_OUTPUT_LVDS) &&
+	if (intel_pipe_has_type(&crtc->base, INTEL_OUTPUT_LVDS) &&
 		 intel_panel_use_ssc(dev_priv) && num_connectors < 2)
 		dpll |= PLLB_REF_INPUT_SPREADSPECTRUMIN;
 	else
@@ -4527,7 +4479,7 @@ static void i8xx_update_pll(struct drm_crtc *crtc,
 	POSTING_READ(DPLL(pipe));
 	udelay(150);
 
-	for_each_encoder_on_crtc(dev, crtc, encoder)
+	for_each_encoder_on_crtc(dev, &crtc->base, encoder)
 		if (encoder->pre_pll_enable)
 			encoder->pre_pll_enable(encoder);
 
@@ -4552,7 +4504,7 @@ static void intel_set_pipe_timings(struct intel_crtc *intel_crtc,
 	struct drm_device *dev = intel_crtc->base.dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	enum pipe pipe = intel_crtc->pipe;
-	enum transcoder cpu_transcoder = intel_crtc->cpu_transcoder;
+	enum transcoder cpu_transcoder = intel_crtc->config.cpu_transcoder;
 	uint32_t vsyncshift;
 
 	if (!IS_GEN2(dev) && adjusted_mode->flags & DRM_MODE_FLAG_INTERLACE) {
@@ -4603,22 +4555,92 @@ static void intel_set_pipe_timings(struct intel_crtc *intel_crtc,
 		   ((mode->hdisplay - 1) << 16) | (mode->vdisplay - 1));
 }
 
+static void i9xx_set_pipeconf(struct intel_crtc *intel_crtc)
+{
+	struct drm_device *dev = intel_crtc->base.dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	uint32_t pipeconf;
+
+	pipeconf = I915_READ(PIPECONF(intel_crtc->pipe));
+
+	if (intel_crtc->pipe == 0 && INTEL_INFO(dev)->gen < 4) {
+		/* Enable pixel doubling when the dot clock is > 90% of the (display)
+		 * core speed.
+		 *
+		 * XXX: No double-wide on 915GM pipe B. Is that the only reason for the
+		 * pipe == 0 check?
+		 */
+		if (intel_crtc->config.requested_mode.clock >
+		    dev_priv->display.get_display_clock_speed(dev) * 9 / 10)
+			pipeconf |= PIPECONF_DOUBLE_WIDE;
+		else
+			pipeconf &= ~PIPECONF_DOUBLE_WIDE;
+	}
+
+	/* default to 8bpc */
+	pipeconf &= ~(PIPECONF_BPC_MASK | PIPECONF_DITHER_EN);
+	if (intel_crtc->config.has_dp_encoder) {
+		if (intel_crtc->config.dither) {
+			pipeconf |= PIPECONF_6BPC |
+				    PIPECONF_DITHER_EN |
+				    PIPECONF_DITHER_TYPE_SP;
+		}
+	}
+
+	if (IS_VALLEYVIEW(dev) && intel_pipe_has_type(&intel_crtc->base,
+						      INTEL_OUTPUT_EDP)) {
+		if (intel_crtc->config.dither) {
+			pipeconf |= PIPECONF_6BPC |
+					PIPECONF_ENABLE |
+					I965_PIPECONF_ACTIVE;
+		}
+	}
+
+	if (HAS_PIPE_CXSR(dev)) {
+		if (intel_crtc->lowfreq_avail) {
+			DRM_DEBUG_KMS("enabling CxSR downclocking\n");
+			pipeconf |= PIPECONF_CXSR_DOWNCLOCK;
+		} else {
+			DRM_DEBUG_KMS("disabling CxSR downclocking\n");
+			pipeconf &= ~PIPECONF_CXSR_DOWNCLOCK;
+		}
+	}
+
+	pipeconf &= ~PIPECONF_INTERLACE_MASK;
+	if (!IS_GEN2(dev) &&
+	    intel_crtc->config.adjusted_mode.flags & DRM_MODE_FLAG_INTERLACE)
+		pipeconf |= PIPECONF_INTERLACE_W_FIELD_INDICATION;
+	else
+		pipeconf |= PIPECONF_PROGRESSIVE;
+
+	if (IS_VALLEYVIEW(dev)) {
+		if (intel_crtc->config.limited_color_range)
+			pipeconf |= PIPECONF_COLOR_RANGE_SELECT;
+		else
+			pipeconf &= ~PIPECONF_COLOR_RANGE_SELECT;
+	}
+
+	I915_WRITE(PIPECONF(intel_crtc->pipe), pipeconf);
+	POSTING_READ(PIPECONF(intel_crtc->pipe));
+}
+
 static int i9xx_crtc_mode_set(struct drm_crtc *crtc,
-			      struct drm_display_mode *mode,
-			      struct drm_display_mode *adjusted_mode,
 			      int x, int y,
 			      struct drm_framebuffer *fb)
 {
 	struct drm_device *dev = crtc->dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
+	struct drm_display_mode *adjusted_mode =
+		&intel_crtc->config.adjusted_mode;
+	struct drm_display_mode *mode = &intel_crtc->config.requested_mode;
 	int pipe = intel_crtc->pipe;
 	int plane = intel_crtc->plane;
 	int refclk, num_connectors = 0;
 	intel_clock_t clock, reduced_clock;
-	u32 dspcntr, pipeconf;
+	u32 dspcntr;
 	bool ok, has_reduced_clock = false, is_sdvo = false;
-	bool is_lvds = false, is_tv = false, is_dp = false;
+	bool is_lvds = false, is_tv = false;
 	struct intel_encoder *encoder;
 	const intel_limit_t *limit;
 	int ret;
@@ -4637,9 +4659,6 @@ static int i9xx_crtc_mode_set(struct drm_crtc *crtc,
 		case INTEL_OUTPUT_TVOUT:
 			is_tv = true;
 			break;
-		case INTEL_OUTPUT_DISPLAYPORT:
-			is_dp = true;
-			break;
 		}
 
 		num_connectors++;
@@ -4676,86 +4695,42 @@ static int i9xx_crtc_mode_set(struct drm_crtc *crtc,
 						    &clock,
 						    &reduced_clock);
 	}
+	/* Compat-code for transition, will disappear. */
+	if (!intel_crtc->config.clock_set) {
+		intel_crtc->config.dpll.n = clock.n;
+		intel_crtc->config.dpll.m1 = clock.m1;
+		intel_crtc->config.dpll.m2 = clock.m2;
+		intel_crtc->config.dpll.p1 = clock.p1;
+		intel_crtc->config.dpll.p2 = clock.p2;
+	}
 
 	if (is_sdvo && is_tv)
-		i9xx_adjust_sdvo_tv_clock(adjusted_mode, &clock);
+		i9xx_adjust_sdvo_tv_clock(intel_crtc);
 
 	if (IS_GEN2(dev))
-		i8xx_update_pll(crtc, adjusted_mode, &clock,
+		i8xx_update_pll(intel_crtc, adjusted_mode,
 				has_reduced_clock ? &reduced_clock : NULL,
 				num_connectors);
 	else if (IS_VALLEYVIEW(dev))
-		vlv_update_pll(crtc, mode, adjusted_mode, &clock,
-				has_reduced_clock ? &reduced_clock : NULL,
-				num_connectors);
+		vlv_update_pll(intel_crtc);
 	else
-		i9xx_update_pll(crtc, mode, adjusted_mode, &clock,
+		i9xx_update_pll(intel_crtc,
 				has_reduced_clock ? &reduced_clock : NULL,
 				num_connectors);
 
-	/* setup pipeconf */
-	pipeconf = I915_READ(PIPECONF(pipe));
-
 	/* Set up the display plane register */
 	dspcntr = DISPPLANE_GAMMA_ENABLE;
 
-	if (pipe == 0)
-		dspcntr &= ~DISPPLANE_SEL_PIPE_MASK;
-	else
-		dspcntr |= DISPPLANE_SEL_PIPE_B;
-
-	if (pipe == 0 && INTEL_INFO(dev)->gen < 4) {
-		/* Enable pixel doubling when the dot clock is > 90% of the (display)
-		 * core speed.
-		 *
-		 * XXX: No double-wide on 915GM pipe B. Is that the only reason for the
-		 * pipe == 0 check?
-		 */
-		if (mode->clock >
-		    dev_priv->display.get_display_clock_speed(dev) * 9 / 10)
-			pipeconf |= PIPECONF_DOUBLE_WIDE;
+	if (!IS_VALLEYVIEW(dev)) {
+		if (pipe == 0)
+			dspcntr &= ~DISPPLANE_SEL_PIPE_MASK;
 		else
-			pipeconf &= ~PIPECONF_DOUBLE_WIDE;
-	}
-
-	/* default to 8bpc */
-	pipeconf &= ~(PIPECONF_BPC_MASK | PIPECONF_DITHER_EN);
-	if (is_dp) {
-		if (adjusted_mode->private_flags & INTEL_MODE_DP_FORCE_6BPC) {
-			pipeconf |= PIPECONF_6BPC |
-				    PIPECONF_DITHER_EN |
-				    PIPECONF_DITHER_TYPE_SP;
-		}
-	}
-
-	if (IS_VALLEYVIEW(dev) && intel_pipe_has_type(crtc, INTEL_OUTPUT_EDP)) {
-		if (adjusted_mode->private_flags & INTEL_MODE_DP_FORCE_6BPC) {
-			pipeconf |= PIPECONF_6BPC |
-					PIPECONF_ENABLE |
-					I965_PIPECONF_ACTIVE;
-		}
+			dspcntr |= DISPPLANE_SEL_PIPE_B;
 	}
 
 	DRM_DEBUG_KMS("Mode for pipe %c:\n", pipe == 0 ? 'A' : 'B');
 	drm_mode_debug_printmodeline(mode);
 
-	if (HAS_PIPE_CXSR(dev)) {
-		if (intel_crtc->lowfreq_avail) {
-			DRM_DEBUG_KMS("enabling CxSR downclocking\n");
-			pipeconf |= PIPECONF_CXSR_DOWNCLOCK;
-		} else {
-			DRM_DEBUG_KMS("disabling CxSR downclocking\n");
-			pipeconf &= ~PIPECONF_CXSR_DOWNCLOCK;
-		}
-	}
-
-	pipeconf &= ~PIPECONF_INTERLACE_MASK;
-	if (!IS_GEN2(dev) &&
-	    adjusted_mode->flags & DRM_MODE_FLAG_INTERLACE)
-		pipeconf |= PIPECONF_INTERLACE_W_FIELD_INDICATION;
-	else
-		pipeconf |= PIPECONF_PROGRESSIVE;
-
 	intel_set_pipe_timings(intel_crtc, mode, adjusted_mode);
 
 	/* pipesrc and dspsize control the size that is scaled from,
@@ -4766,8 +4741,8 @@ static int i9xx_crtc_mode_set(struct drm_crtc *crtc,
 		   (mode->hdisplay - 1));
 	I915_WRITE(DSPPOS(plane), 0);
 
-	I915_WRITE(PIPECONF(pipe), pipeconf);
-	POSTING_READ(PIPECONF(pipe));
+	i9xx_set_pipeconf(intel_crtc);
+
 	intel_enable_pipe(dev_priv, pipe, false);
 
 	intel_wait_for_vblank(dev, pipe);
@@ -4782,12 +4757,26 @@ static int i9xx_crtc_mode_set(struct drm_crtc *crtc,
 	return ret;
 }
 
+static bool i9xx_get_pipe_config(struct intel_crtc *crtc,
+				 struct intel_crtc_config *pipe_config)
+{
+	struct drm_device *dev = crtc->base.dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	uint32_t tmp;
+
+	tmp = I915_READ(PIPECONF(crtc->pipe));
+	if (!(tmp & PIPECONF_ENABLE))
+		return false;
+
+	return true;
+}
+
 static void ironlake_init_pch_refclk(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct drm_mode_config *mode_config = &dev->mode_config;
 	struct intel_encoder *encoder;
-	u32 temp;
+	u32 val, final;
 	bool has_lvds = false;
 	bool has_cpu_edp = false;
 	bool has_pch_edp = false;
@@ -4830,70 +4819,109 @@ static void ironlake_init_pch_refclk(struct drm_device *dev)
 	 * PCH B stepping, previous chipset stepping should be
 	 * ignoring this setting.
 	 */
-	temp = I915_READ(PCH_DREF_CONTROL);
+	val = I915_READ(PCH_DREF_CONTROL);
+
+	/* As we must carefully and slowly disable/enable each source in turn,
+	 * compute the final state we want first and check if we need to
+	 * make any changes at all.
+	 */
+	final = val;
+	final &= ~DREF_NONSPREAD_SOURCE_MASK;
+	if (has_ck505)
+		final |= DREF_NONSPREAD_CK505_ENABLE;
+	else
+		final |= DREF_NONSPREAD_SOURCE_ENABLE;
+
+	final &= ~DREF_SSC_SOURCE_MASK;
+	final &= ~DREF_CPU_SOURCE_OUTPUT_MASK;
+	final &= ~DREF_SSC1_ENABLE;
+
+	if (has_panel) {
+		final |= DREF_SSC_SOURCE_ENABLE;
+
+		if (intel_panel_use_ssc(dev_priv) && can_ssc)
+			final |= DREF_SSC1_ENABLE;
+
+		if (has_cpu_edp) {
+			if (intel_panel_use_ssc(dev_priv) && can_ssc)
+				final |= DREF_CPU_SOURCE_OUTPUT_DOWNSPREAD;
+			else
+				final |= DREF_CPU_SOURCE_OUTPUT_NONSPREAD;
+		} else
+			final |= DREF_CPU_SOURCE_OUTPUT_DISABLE;
+	} else {
+		final |= DREF_SSC_SOURCE_DISABLE;
+		final |= DREF_CPU_SOURCE_OUTPUT_DISABLE;
+	}
+
+	if (final == val)
+		return;
+
 	/* Always enable nonspread source */
-	temp &= ~DREF_NONSPREAD_SOURCE_MASK;
+	val &= ~DREF_NONSPREAD_SOURCE_MASK;
 
 	if (has_ck505)
-		temp |= DREF_NONSPREAD_CK505_ENABLE;
+		val |= DREF_NONSPREAD_CK505_ENABLE;
 	else
-		temp |= DREF_NONSPREAD_SOURCE_ENABLE;
+		val |= DREF_NONSPREAD_SOURCE_ENABLE;
 
 	if (has_panel) {
-		temp &= ~DREF_SSC_SOURCE_MASK;
-		temp |= DREF_SSC_SOURCE_ENABLE;
+		val &= ~DREF_SSC_SOURCE_MASK;
+		val |= DREF_SSC_SOURCE_ENABLE;
 
 		/* SSC must be turned on before enabling the CPU output  */
 		if (intel_panel_use_ssc(dev_priv) && can_ssc) {
 			DRM_DEBUG_KMS("Using SSC on panel\n");
-			temp |= DREF_SSC1_ENABLE;
+			val |= DREF_SSC1_ENABLE;
 		} else
-			temp &= ~DREF_SSC1_ENABLE;
+			val &= ~DREF_SSC1_ENABLE;
 
 		/* Get SSC going before enabling the outputs */
-		I915_WRITE(PCH_DREF_CONTROL, temp);
+		I915_WRITE(PCH_DREF_CONTROL, val);
 		POSTING_READ(PCH_DREF_CONTROL);
 		udelay(200);
 
-		temp &= ~DREF_CPU_SOURCE_OUTPUT_MASK;
+		val &= ~DREF_CPU_SOURCE_OUTPUT_MASK;
 
 		/* Enable CPU source on CPU attached eDP */
 		if (has_cpu_edp) {
 			if (intel_panel_use_ssc(dev_priv) && can_ssc) {
 				DRM_DEBUG_KMS("Using SSC on eDP\n");
-				temp |= DREF_CPU_SOURCE_OUTPUT_DOWNSPREAD;
+				val |= DREF_CPU_SOURCE_OUTPUT_DOWNSPREAD;
 			}
 			else
-				temp |= DREF_CPU_SOURCE_OUTPUT_NONSPREAD;
+				val |= DREF_CPU_SOURCE_OUTPUT_NONSPREAD;
 		} else
-			temp |= DREF_CPU_SOURCE_OUTPUT_DISABLE;
+			val |= DREF_CPU_SOURCE_OUTPUT_DISABLE;
 
-		I915_WRITE(PCH_DREF_CONTROL, temp);
+		I915_WRITE(PCH_DREF_CONTROL, val);
 		POSTING_READ(PCH_DREF_CONTROL);
 		udelay(200);
 	} else {
 		DRM_DEBUG_KMS("Disabling SSC entirely\n");
 
-		temp &= ~DREF_CPU_SOURCE_OUTPUT_MASK;
+		val &= ~DREF_CPU_SOURCE_OUTPUT_MASK;
 
 		/* Turn off CPU output */
-		temp |= DREF_CPU_SOURCE_OUTPUT_DISABLE;
+		val |= DREF_CPU_SOURCE_OUTPUT_DISABLE;
 
-		I915_WRITE(PCH_DREF_CONTROL, temp);
+		I915_WRITE(PCH_DREF_CONTROL, val);
 		POSTING_READ(PCH_DREF_CONTROL);
 		udelay(200);
 
 		/* Turn off the SSC source */
-		temp &= ~DREF_SSC_SOURCE_MASK;
-		temp |= DREF_SSC_SOURCE_DISABLE;
+		val &= ~DREF_SSC_SOURCE_MASK;
+		val |= DREF_SSC_SOURCE_DISABLE;
 
 		/* Turn off SSC1 */
-		temp &= ~ DREF_SSC1_ENABLE;
+		val &= ~DREF_SSC1_ENABLE;
 
-		I915_WRITE(PCH_DREF_CONTROL, temp);
+		I915_WRITE(PCH_DREF_CONTROL, val);
 		POSTING_READ(PCH_DREF_CONTROL);
 		udelay(200);
 	}
+
+	BUG_ON(val != final);
 }
 
 /* Sequence to enable CLKOUT_DP for FDI usage and configure PCH FDI I/O. */
@@ -4958,13 +4986,6 @@ static void lpt_init_pch_refclk(struct drm_device *dev)
 	tmp |= (0x12 << 24);
 	intel_sbi_write(dev_priv, 0x8008, tmp, SBI_MPHY);
 
-	if (!is_sdv) {
-		tmp = intel_sbi_read(dev_priv, 0x808C, SBI_MPHY);
-		tmp &= ~(0x3 << 6);
-		tmp |= (1 << 6) | (1 << 0);
-		intel_sbi_write(dev_priv, 0x808C, tmp, SBI_MPHY);
-	}
-
 	if (is_sdv) {
 		tmp = intel_sbi_read(dev_priv, 0x800C, SBI_MPHY);
 		tmp |= 0x7FFF;
@@ -5118,7 +5139,7 @@ static void ironlake_set_pipeconf(struct drm_crtc *crtc,
 	val = I915_READ(PIPECONF(pipe));
 
 	val &= ~PIPECONF_BPC_MASK;
-	switch (intel_crtc->bpp) {
+	switch (intel_crtc->config.pipe_bpp) {
 	case 18:
 		val |= PIPECONF_6BPC;
 		break;
@@ -5146,7 +5167,7 @@ static void ironlake_set_pipeconf(struct drm_crtc *crtc,
 	else
 		val |= PIPECONF_PROGRESSIVE;
 
-	if (adjusted_mode->private_flags & INTEL_MODE_LIMITED_COLOR_RANGE)
+	if (intel_crtc->config.limited_color_range)
 		val |= PIPECONF_COLOR_RANGE_SELECT;
 	else
 		val &= ~PIPECONF_COLOR_RANGE_SELECT;
@@ -5162,8 +5183,7 @@ static void ironlake_set_pipeconf(struct drm_crtc *crtc,
  * is supported, but eventually this should handle various
  * RGB<->YCbCr scenarios as well.
  */
-static void intel_set_pipe_csc(struct drm_crtc *crtc,
-			       const struct drm_display_mode *adjusted_mode)
+static void intel_set_pipe_csc(struct drm_crtc *crtc)
 {
 	struct drm_device *dev = crtc->dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
@@ -5178,7 +5198,7 @@ static void intel_set_pipe_csc(struct drm_crtc *crtc,
 	 * consideration.
 	 */
 
-	if (adjusted_mode->private_flags & INTEL_MODE_LIMITED_COLOR_RANGE)
+	if (intel_crtc->config.limited_color_range)
 		coeff = ((235 - 16) * (1 << 12) / 255) & 0xff8; /* 0.xxx... */
 
 	/*
@@ -5202,7 +5222,7 @@ static void intel_set_pipe_csc(struct drm_crtc *crtc,
 	if (INTEL_INFO(dev)->gen > 6) {
 		uint16_t postoff = 0;
 
-		if (adjusted_mode->private_flags & INTEL_MODE_LIMITED_COLOR_RANGE)
+		if (intel_crtc->config.limited_color_range)
 			postoff = (16 * (1 << 13) / 255) & 0x1fff;
 
 		I915_WRITE(PIPE_CSC_POSTOFF_HI(pipe), postoff);
@@ -5213,7 +5233,7 @@ static void intel_set_pipe_csc(struct drm_crtc *crtc,
 	} else {
 		uint32_t mode = CSC_MODE_YUV_TO_RGB;
 
-		if (adjusted_mode->private_flags & INTEL_MODE_LIMITED_COLOR_RANGE)
+		if (intel_crtc->config.limited_color_range)
 			mode |= CSC_BLACK_SCREEN_OFFSET;
 
 		I915_WRITE(PIPE_CSC_MODE(pipe), mode);
@@ -5226,7 +5246,7 @@ static void haswell_set_pipeconf(struct drm_crtc *crtc,
 {
 	struct drm_i915_private *dev_priv = crtc->dev->dev_private;
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
-	enum transcoder cpu_transcoder = intel_crtc->cpu_transcoder;
+	enum transcoder cpu_transcoder = intel_crtc->config.cpu_transcoder;
 	uint32_t val;
 
 	val = I915_READ(PIPECONF(cpu_transcoder));
@@ -5303,7 +5323,7 @@ static bool ironlake_compute_clocks(struct drm_crtc *crtc,
 	}
 
 	if (is_sdvo && is_tv)
-		i9xx_adjust_sdvo_tv_clock(adjusted_mode, clock);
+		i9xx_adjust_sdvo_tv_clock(to_intel_crtc(crtc));
 
 	return true;
 }
@@ -5344,7 +5364,7 @@ static bool ironlake_check_fdi_lanes(struct intel_crtc *intel_crtc)
 		return false;
 	}
 
-	if (dev_priv->num_pipe == 2)
+	if (INTEL_INFO(dev)->num_pipes == 2)
 		return true;
 
 	switch (intel_crtc->pipe) {
@@ -5401,87 +5421,87 @@ int ironlake_get_lanes_required(int target_clock, int link_bw, int bpp)
 	return bps / (link_bw * 8) + 1;
 }
 
-static void ironlake_set_m_n(struct drm_crtc *crtc,
-			     struct drm_display_mode *mode,
-			     struct drm_display_mode *adjusted_mode)
+void intel_pch_transcoder_set_m_n(struct intel_crtc *crtc,
+				  struct intel_link_m_n *m_n)
 {
-	struct drm_device *dev = crtc->dev;
+	struct drm_device *dev = crtc->base.dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
-	enum transcoder cpu_transcoder = intel_crtc->cpu_transcoder;
-	struct intel_encoder *intel_encoder, *edp_encoder = NULL;
-	struct intel_link_m_n m_n = {0};
-	int target_clock, pixel_multiplier, lane, link_bw;
-	bool is_dp = false, is_cpu_edp = false;
+	int pipe = crtc->pipe;
 
-	for_each_encoder_on_crtc(dev, crtc, intel_encoder) {
-		switch (intel_encoder->type) {
-		case INTEL_OUTPUT_DISPLAYPORT:
-			is_dp = true;
-			break;
-		case INTEL_OUTPUT_EDP:
-			is_dp = true;
-			if (!intel_encoder_is_pch_edp(&intel_encoder->base))
-				is_cpu_edp = true;
-			edp_encoder = intel_encoder;
-			break;
-		}
-	}
+	I915_WRITE(TRANSDATA_M1(pipe), TU_SIZE(m_n->tu) | m_n->gmch_m);
+	I915_WRITE(TRANSDATA_N1(pipe), m_n->gmch_n);
+	I915_WRITE(TRANSDPLINK_M1(pipe), m_n->link_m);
+	I915_WRITE(TRANSDPLINK_N1(pipe), m_n->link_n);
+}
+
+void intel_cpu_transcoder_set_m_n(struct intel_crtc *crtc,
+				  struct intel_link_m_n *m_n)
+{
+	struct drm_device *dev = crtc->base.dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	int pipe = crtc->pipe;
+	enum transcoder transcoder = crtc->config.cpu_transcoder;
 
-	/* FDI link */
-	pixel_multiplier = intel_mode_get_pixel_multiplier(adjusted_mode);
-	lane = 0;
-	/* CPU eDP doesn't require FDI link, so just set DP M/N
-	   according to current link config */
-	if (is_cpu_edp) {
-		intel_edp_link_config(edp_encoder, &lane, &link_bw);
+	if (INTEL_INFO(dev)->gen >= 5) {
+		I915_WRITE(PIPE_DATA_M1(transcoder), TU_SIZE(m_n->tu) | m_n->gmch_m);
+		I915_WRITE(PIPE_DATA_N1(transcoder), m_n->gmch_n);
+		I915_WRITE(PIPE_LINK_M1(transcoder), m_n->link_m);
+		I915_WRITE(PIPE_LINK_N1(transcoder), m_n->link_n);
 	} else {
-		/* FDI is a binary signal running at ~2.7GHz, encoding
-		 * each output octet as 10 bits. The actual frequency
-		 * is stored as a divider into a 100MHz clock, and the
-		 * mode pixel clock is stored in units of 1KHz.
-		 * Hence the bw of each lane in terms of the mode signal
-		 * is:
-		 */
-		link_bw = intel_fdi_link_freq(dev) * MHz(100)/KHz(1)/10;
+		I915_WRITE(PIPE_GMCH_DATA_M(pipe), TU_SIZE(m_n->tu) | m_n->gmch_m);
+		I915_WRITE(PIPE_GMCH_DATA_N(pipe), m_n->gmch_n);
+		I915_WRITE(PIPE_DP_LINK_M(pipe), m_n->link_m);
+		I915_WRITE(PIPE_DP_LINK_N(pipe), m_n->link_n);
 	}
+}
+
+static void ironlake_fdi_set_m_n(struct drm_crtc *crtc)
+{
+	struct drm_device *dev = crtc->dev;
+	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
+	struct drm_display_mode *adjusted_mode =
+		&intel_crtc->config.adjusted_mode;
+	struct intel_link_m_n m_n = {0};
+	int target_clock, lane, link_bw;
+
+	/* FDI is a binary signal running at ~2.7GHz, encoding
+	 * each output octet as 10 bits. The actual frequency
+	 * is stored as a divider into a 100MHz clock, and the
+	 * mode pixel clock is stored in units of 1KHz.
+	 * Hence the bw of each lane in terms of the mode signal
+	 * is:
+	 */
+	link_bw = intel_fdi_link_freq(dev) * MHz(100)/KHz(1)/10;
 
-	/* [e]DP over FDI requires target mode clock instead of link clock. */
-	if (edp_encoder)
-		target_clock = intel_edp_target_clock(edp_encoder, mode);
-	else if (is_dp)
-		target_clock = mode->clock;
+	if (intel_crtc->config.pixel_target_clock)
+		target_clock = intel_crtc->config.pixel_target_clock;
 	else
 		target_clock = adjusted_mode->clock;
 
-	if (!lane)
-		lane = ironlake_get_lanes_required(target_clock, link_bw,
-						   intel_crtc->bpp);
+	lane = ironlake_get_lanes_required(target_clock, link_bw,
+					   intel_crtc->config.pipe_bpp);
 
 	intel_crtc->fdi_lanes = lane;
 
-	if (pixel_multiplier > 1)
-		link_bw *= pixel_multiplier;
-	intel_link_compute_m_n(intel_crtc->bpp, lane, target_clock, link_bw, &m_n);
+	if (intel_crtc->config.pixel_multiplier > 1)
+		link_bw *= intel_crtc->config.pixel_multiplier;
+	intel_link_compute_m_n(intel_crtc->config.pipe_bpp, lane, target_clock,
+			       link_bw, &m_n);
 
-	I915_WRITE(PIPE_DATA_M1(cpu_transcoder), TU_SIZE(m_n.tu) | m_n.gmch_m);
-	I915_WRITE(PIPE_DATA_N1(cpu_transcoder), m_n.gmch_n);
-	I915_WRITE(PIPE_LINK_M1(cpu_transcoder), m_n.link_m);
-	I915_WRITE(PIPE_LINK_N1(cpu_transcoder), m_n.link_n);
+	intel_cpu_transcoder_set_m_n(intel_crtc, &m_n);
 }
 
 static uint32_t ironlake_compute_dpll(struct intel_crtc *intel_crtc,
-				      struct drm_display_mode *adjusted_mode,
-				      intel_clock_t *clock, u32 fp)
+				      intel_clock_t *clock, u32 *fp,
+				      intel_clock_t *reduced_clock, u32 *fp2)
 {
 	struct drm_crtc *crtc = &intel_crtc->base;
 	struct drm_device *dev = crtc->dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_encoder *intel_encoder;
 	uint32_t dpll;
-	int factor, pixel_multiplier, num_connectors = 0;
+	int factor, num_connectors = 0;
 	bool is_lvds = false, is_sdvo = false, is_tv = false;
-	bool is_dp = false, is_cpu_edp = false;
 
 	for_each_encoder_on_crtc(dev, crtc, intel_encoder) {
 		switch (intel_encoder->type) {
@@ -5497,14 +5517,6 @@ static uint32_t ironlake_compute_dpll(struct intel_crtc *intel_crtc,
 		case INTEL_OUTPUT_TVOUT:
 			is_tv = true;
 			break;
-		case INTEL_OUTPUT_DISPLAYPORT:
-			is_dp = true;
-			break;
-		case INTEL_OUTPUT_EDP:
-			is_dp = true;
-			if (!intel_encoder_is_pch_edp(&intel_encoder->base))
-				is_cpu_edp = true;
-			break;
 		}
 
 		num_connectors++;
@@ -5515,13 +5527,16 @@ static uint32_t ironlake_compute_dpll(struct intel_crtc *intel_crtc,
 	if (is_lvds) {
 		if ((intel_panel_use_ssc(dev_priv) &&
 		     dev_priv->lvds_ssc_freq == 100) ||
-		    intel_is_dual_link_lvds(dev))
+		    (HAS_PCH_IBX(dev) && intel_is_dual_link_lvds(dev)))
 			factor = 25;
 	} else if (is_sdvo && is_tv)
 		factor = 20;
 
 	if (clock->m < factor * clock->n)
-		fp |= FP_CB_TUNE;
+		*fp |= FP_CB_TUNE;
+
+	if (fp2 && (reduced_clock->m < factor * reduced_clock->n))
+		*fp2 |= FP_CB_TUNE;
 
 	dpll = 0;
 
@@ -5530,13 +5545,14 @@ static uint32_t ironlake_compute_dpll(struct intel_crtc *intel_crtc,
 	else
 		dpll |= DPLLB_MODE_DAC_SERIAL;
 	if (is_sdvo) {
-		pixel_multiplier = intel_mode_get_pixel_multiplier(adjusted_mode);
-		if (pixel_multiplier > 1) {
-			dpll |= (pixel_multiplier - 1) << PLL_REF_SDVO_HDMI_MULTIPLIER_SHIFT;
+		if (intel_crtc->config.pixel_multiplier > 1) {
+			dpll |= (intel_crtc->config.pixel_multiplier - 1)
+				<< PLL_REF_SDVO_HDMI_MULTIPLIER_SHIFT;
 		}
 		dpll |= DPLL_DVO_HIGH_SPEED;
 	}
-	if (is_dp && !is_cpu_edp)
+	if (intel_crtc->config.has_dp_encoder &&
+	    intel_crtc->config.has_pch_encoder)
 		dpll |= DPLL_DVO_HIGH_SPEED;
 
 	/* compute bitmask from p1 value */
@@ -5574,21 +5590,22 @@ static uint32_t ironlake_compute_dpll(struct intel_crtc *intel_crtc,
 }
 
 static int ironlake_crtc_mode_set(struct drm_crtc *crtc,
-				  struct drm_display_mode *mode,
-				  struct drm_display_mode *adjusted_mode,
 				  int x, int y,
 				  struct drm_framebuffer *fb)
 {
 	struct drm_device *dev = crtc->dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
+	struct drm_display_mode *adjusted_mode =
+		&intel_crtc->config.adjusted_mode;
+	struct drm_display_mode *mode = &intel_crtc->config.requested_mode;
 	int pipe = intel_crtc->pipe;
 	int plane = intel_crtc->plane;
 	int num_connectors = 0;
 	intel_clock_t clock, reduced_clock;
 	u32 dpll, fp = 0, fp2 = 0;
 	bool ok, has_reduced_clock = false;
-	bool is_lvds = false, is_dp = false, is_cpu_edp = false;
+	bool is_lvds = false;
 	struct intel_encoder *encoder;
 	int ret;
 	bool dither, fdi_config_ok;
@@ -5598,14 +5615,6 @@ static int ironlake_crtc_mode_set(struct drm_crtc *crtc,
 		case INTEL_OUTPUT_LVDS:
 			is_lvds = true;
 			break;
-		case INTEL_OUTPUT_DISPLAYPORT:
-			is_dp = true;
-			break;
-		case INTEL_OUTPUT_EDP:
-			is_dp = true;
-			if (!intel_encoder_is_pch_edp(&encoder->base))
-				is_cpu_edp = true;
-			break;
 		}
 
 		num_connectors++;
@@ -5614,19 +5623,28 @@ static int ironlake_crtc_mode_set(struct drm_crtc *crtc,
 	WARN(!(HAS_PCH_IBX(dev) || HAS_PCH_CPT(dev)),
 	     "Unexpected PCH type %d\n", INTEL_PCH_TYPE(dev));
 
+	intel_crtc->config.cpu_transcoder = pipe;
+
 	ok = ironlake_compute_clocks(crtc, adjusted_mode, &clock,
 				     &has_reduced_clock, &reduced_clock);
 	if (!ok) {
 		DRM_ERROR("Couldn't find PLL settings for mode!\n");
 		return -EINVAL;
 	}
+	/* Compat-code for transition, will disappear. */
+	if (!intel_crtc->config.clock_set) {
+		intel_crtc->config.dpll.n = clock.n;
+		intel_crtc->config.dpll.m1 = clock.m1;
+		intel_crtc->config.dpll.m2 = clock.m2;
+		intel_crtc->config.dpll.p1 = clock.p1;
+		intel_crtc->config.dpll.p2 = clock.p2;
+	}
 
 	/* Ensure that the cursor is valid for the new mode before changing... */
 	intel_crtc_update_cursor(crtc, true);
 
 	/* determine panel color depth */
-	dither = intel_choose_pipe_bpp_dither(crtc, fb, &intel_crtc->bpp,
-					      adjusted_mode);
+	dither = intel_crtc->config.dither;
 	if (is_lvds && dev_priv->lvds_dither)
 		dither = true;
 
@@ -5635,13 +5653,14 @@ static int ironlake_crtc_mode_set(struct drm_crtc *crtc,
 		fp2 = reduced_clock.n << 16 | reduced_clock.m1 << 8 |
 			reduced_clock.m2;
 
-	dpll = ironlake_compute_dpll(intel_crtc, adjusted_mode, &clock, fp);
+	dpll = ironlake_compute_dpll(intel_crtc, &clock, &fp, &reduced_clock,
+				     has_reduced_clock ? &fp2 : NULL);
 
 	DRM_DEBUG_KMS("Mode for pipe %d:\n", pipe);
 	drm_mode_debug_printmodeline(mode);
 
 	/* CPU eDP is the only output that doesn't need a PCH PLL of its own. */
-	if (!is_cpu_edp) {
+	if (intel_crtc->config.has_pch_encoder) {
 		struct intel_pch_pll *pll;
 
 		pll = intel_get_pch_pll(intel_crtc, dpll, fp);
@@ -5653,8 +5672,8 @@ static int ironlake_crtc_mode_set(struct drm_crtc *crtc,
 	} else
 		intel_put_pch_pll(intel_crtc);
 
-	if (is_dp && !is_cpu_edp)
-		intel_dp_set_m_n(crtc, mode, adjusted_mode);
+	if (intel_crtc->config.has_dp_encoder)
+		intel_dp_set_m_n(intel_crtc);
 
 	for_each_encoder_on_crtc(dev, crtc, encoder)
 		if (encoder->pre_pll_enable)
@@ -5689,7 +5708,9 @@ static int ironlake_crtc_mode_set(struct drm_crtc *crtc,
 
 	/* Note, this also computes intel_crtc->fdi_lanes which is used below in
 	 * ironlake_check_fdi_lanes. */
-	ironlake_set_m_n(crtc, mode, adjusted_mode);
+	intel_crtc->fdi_lanes = 0;
+	if (intel_crtc->config.has_pch_encoder)
+		ironlake_fdi_set_m_n(crtc);
 
 	fdi_config_ok = ironlake_check_fdi_lanes(intel_crtc);
 
@@ -5710,6 +5731,23 @@ static int ironlake_crtc_mode_set(struct drm_crtc *crtc,
 	return fdi_config_ok ? ret : -EINVAL;
 }
 
+static bool ironlake_get_pipe_config(struct intel_crtc *crtc,
+				     struct intel_crtc_config *pipe_config)
+{
+	struct drm_device *dev = crtc->base.dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	uint32_t tmp;
+
+	tmp = I915_READ(PIPECONF(crtc->pipe));
+	if (!(tmp & PIPECONF_ENABLE))
+		return false;
+
+	if (I915_READ(TRANSCONF(crtc->pipe)) & TRANS_ENABLE)
+		pipe_config->has_pch_encoder = true;
+
+	return true;
+}
+
 static void haswell_modeset_global_resources(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
@@ -5740,29 +5778,26 @@ static void haswell_modeset_global_resources(struct drm_device *dev)
 }
 
 static int haswell_crtc_mode_set(struct drm_crtc *crtc,
-				 struct drm_display_mode *mode,
-				 struct drm_display_mode *adjusted_mode,
 				 int x, int y,
 				 struct drm_framebuffer *fb)
 {
 	struct drm_device *dev = crtc->dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
+	struct drm_display_mode *adjusted_mode =
+		&intel_crtc->config.adjusted_mode;
+	struct drm_display_mode *mode = &intel_crtc->config.requested_mode;
 	int pipe = intel_crtc->pipe;
 	int plane = intel_crtc->plane;
 	int num_connectors = 0;
-	bool is_dp = false, is_cpu_edp = false;
+	bool is_cpu_edp = false;
 	struct intel_encoder *encoder;
 	int ret;
 	bool dither;
 
 	for_each_encoder_on_crtc(dev, crtc, encoder) {
 		switch (encoder->type) {
-		case INTEL_OUTPUT_DISPLAYPORT:
-			is_dp = true;
-			break;
 		case INTEL_OUTPUT_EDP:
-			is_dp = true;
 			if (!intel_encoder_is_pch_edp(&encoder->base))
 				is_cpu_edp = true;
 			break;
@@ -5772,9 +5807,9 @@ static int haswell_crtc_mode_set(struct drm_crtc *crtc,
 	}
 
 	if (is_cpu_edp)
-		intel_crtc->cpu_transcoder = TRANSCODER_EDP;
+		intel_crtc->config.cpu_transcoder = TRANSCODER_EDP;
 	else
-		intel_crtc->cpu_transcoder = pipe;
+		intel_crtc->config.cpu_transcoder = pipe;
 
 	/* We are not sure yet this won't happen. */
 	WARN(!HAS_PCH_LPT(dev), "Unexpected PCH type %d\n",
@@ -5783,7 +5818,7 @@ static int haswell_crtc_mode_set(struct drm_crtc *crtc,
 	WARN(num_connectors != 1, "%d connectors attached to pipe %c\n",
 	     num_connectors, pipe_name(pipe));
 
-	WARN_ON(I915_READ(PIPECONF(intel_crtc->cpu_transcoder)) &
+	WARN_ON(I915_READ(PIPECONF(intel_crtc->config.cpu_transcoder)) &
 		(PIPECONF_ENABLE | I965_PIPECONF_ACTIVE));
 
 	WARN_ON(I915_READ(DSPCNTR(plane)) & DISPLAY_PLANE_ENABLE);
@@ -5795,25 +5830,24 @@ static int haswell_crtc_mode_set(struct drm_crtc *crtc,
 	intel_crtc_update_cursor(crtc, true);
 
 	/* determine panel color depth */
-	dither = intel_choose_pipe_bpp_dither(crtc, fb, &intel_crtc->bpp,
-					      adjusted_mode);
+	dither = intel_crtc->config.dither;
 
 	DRM_DEBUG_KMS("Mode for pipe %d:\n", pipe);
 	drm_mode_debug_printmodeline(mode);
 
-	if (is_dp && !is_cpu_edp)
-		intel_dp_set_m_n(crtc, mode, adjusted_mode);
+	if (intel_crtc->config.has_dp_encoder)
+		intel_dp_set_m_n(intel_crtc);
 
 	intel_crtc->lowfreq_avail = false;
 
 	intel_set_pipe_timings(intel_crtc, mode, adjusted_mode);
 
-	if (!is_dp || is_cpu_edp)
-		ironlake_set_m_n(crtc, mode, adjusted_mode);
+	if (intel_crtc->config.has_pch_encoder)
+		ironlake_fdi_set_m_n(crtc);
 
 	haswell_set_pipeconf(crtc, adjusted_mode, dither);
 
-	intel_set_pipe_csc(crtc, adjusted_mode);
+	intel_set_pipe_csc(crtc);
 
 	/* Set up the display plane register */
 	I915_WRITE(DSPCNTR(plane), DISPPLANE_GAMMA_ENABLE | DISPPLANE_PIPE_CSC_ENABLE);
@@ -5828,9 +5862,32 @@ static int haswell_crtc_mode_set(struct drm_crtc *crtc,
 	return ret;
 }
 
+static bool haswell_get_pipe_config(struct intel_crtc *crtc,
+				    struct intel_crtc_config *pipe_config)
+{
+	struct drm_device *dev = crtc->base.dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	uint32_t tmp;
+
+	tmp = I915_READ(PIPECONF(crtc->config.cpu_transcoder));
+	if (!(tmp & PIPECONF_ENABLE))
+		return false;
+
+	/*
+	 * aswell has only FDI/PCH transcoder A. It is which is connected to
+	 * DDI E. So just check whether this pipe is wired to DDI E and whether
+	 * the PCH transcoder is on.
+	 */
+	tmp = I915_READ(TRANS_DDI_FUNC_CTL(crtc->pipe));
+	if ((tmp & TRANS_DDI_PORT_MASK) == TRANS_DDI_SELECT_PORT(PORT_E) &&
+	    I915_READ(TRANSCONF(PIPE_A)) & TRANS_ENABLE)
+		pipe_config->has_pch_encoder = true;
+
+
+	return true;
+}
+
 static int intel_crtc_mode_set(struct drm_crtc *crtc,
-			       struct drm_display_mode *mode,
-			       struct drm_display_mode *adjusted_mode,
 			       int x, int y,
 			       struct drm_framebuffer *fb)
 {
@@ -5839,13 +5896,16 @@ static int intel_crtc_mode_set(struct drm_crtc *crtc,
 	struct drm_encoder_helper_funcs *encoder_funcs;
 	struct intel_encoder *encoder;
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
+	struct drm_display_mode *adjusted_mode =
+		&intel_crtc->config.adjusted_mode;
+	struct drm_display_mode *mode = &intel_crtc->config.requested_mode;
 	int pipe = intel_crtc->pipe;
 	int ret;
 
 	drm_vblank_pre_modeset(dev, pipe);
 
-	ret = dev_priv->display.crtc_mode_set(crtc, mode, adjusted_mode,
-					      x, y, fb);
+	ret = dev_priv->display.crtc_mode_set(crtc, x, y, fb);
+
 	drm_vblank_post_modeset(dev, pipe);
 
 	if (ret != 0)
@@ -5856,8 +5916,12 @@ static int intel_crtc_mode_set(struct drm_crtc *crtc,
 			encoder->base.base.id,
 			drm_get_encoder_name(&encoder->base),
 			mode->base.id, mode->name);
-		encoder_funcs = encoder->base.helper_private;
-		encoder_funcs->mode_set(&encoder->base, mode, adjusted_mode);
+		if (encoder->mode_set) {
+			encoder->mode_set(encoder);
+		} else {
+			encoder_funcs = encoder->base.helper_private;
+			encoder_funcs->mode_set(&encoder->base, mode, adjusted_mode);
+		}
 	}
 
 	return 0;
@@ -6325,13 +6389,24 @@ static int intel_crtc_cursor_set(struct drm_crtc *crtc,
 	/* we only need to pin inside GTT if cursor is non-phy */
 	mutex_lock(&dev->struct_mutex);
 	if (!dev_priv->info->cursor_needs_physical) {
+		unsigned alignment;
+
 		if (obj->tiling_mode) {
 			DRM_ERROR("cursor cannot be tiled\n");
 			ret = -EINVAL;
 			goto fail_locked;
 		}
 
-		ret = i915_gem_object_pin_to_display_plane(obj, 0, NULL);
+		/* Note that the w/a also requires 2 PTE of padding following
+		 * the bo. We currently fill all unused PTE with the shadow
+		 * page and so we should always have valid PTE following the
+		 * cursor preventing the VT-d warning.
+		 */
+		alignment = 0;
+		if (need_vtd_wa(dev))
+			alignment = 64*1024;
+
+		ret = i915_gem_object_pin_to_display_plane(obj, alignment, NULL);
 		if (ret) {
 			DRM_ERROR("failed to move cursor bo into the GTT\n");
 			goto fail_locked;
@@ -6436,20 +6511,6 @@ static void intel_crtc_gamma_set(struct drm_crtc *crtc, u16 *red, u16 *green,
 	intel_crtc_load_lut(crtc);
 }
 
-/**
- * Get a pipe with a simple mode set on it for doing load-based monitor
- * detection.
- *
- * It will be up to the load-detect code to adjust the pipe as appropriate for
- * its requirements.  The pipe will be connected to no other encoders.
- *
- * Currently this code will only succeed if there is a pipe with no encoders
- * configured for it.  In the future, it could choose to temporarily disable
- * some outputs to free up a pipe for its use.
- *
- * \return crtc, or NULL if no pipes are available.
- */
-
 /* VESA 640x480x72Hz mode to set on the pipe */
 static struct drm_display_mode load_detect_mode = {
 	DRM_MODE("640x480", DRM_MODE_TYPE_DEFAULT, 31500, 640, 664,
@@ -6776,7 +6837,7 @@ struct drm_display_mode *intel_crtc_mode_get(struct drm_device *dev,
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
-	enum transcoder cpu_transcoder = intel_crtc->cpu_transcoder;
+	enum transcoder cpu_transcoder = intel_crtc->config.cpu_transcoder;
 	struct drm_display_mode *mode;
 	int htot = I915_READ(HTOTAL(cpu_transcoder));
 	int hsync = I915_READ(HSYNC(cpu_transcoder));
@@ -6954,7 +7015,6 @@ static void do_intel_finish_page_flip(struct drm_device *dev,
 	drm_i915_private_t *dev_priv = dev->dev_private;
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 	struct intel_unpin_work *work;
-	struct drm_i915_gem_object *obj;
 	unsigned long flags;
 
 	/* Ignore early vblank irqs */
@@ -6984,8 +7044,6 @@ static void do_intel_finish_page_flip(struct drm_device *dev,
 
 	spin_unlock_irqrestore(&dev->event_lock, flags);
 
-	obj = work->old_fb_obj;
-
 	wake_up_all(&dev_priv->pending_flip_queue);
 
 	queue_work(dev_priv->wq, &work->work);
@@ -7473,19 +7531,93 @@ static void intel_modeset_commit_output_state(struct drm_device *dev)
 	}
 }
 
-static struct drm_display_mode *
-intel_modeset_adjusted_mode(struct drm_crtc *crtc,
-			    struct drm_display_mode *mode)
+static int
+pipe_config_set_bpp(struct drm_crtc *crtc,
+		    struct drm_framebuffer *fb,
+		    struct intel_crtc_config *pipe_config)
+{
+	struct drm_device *dev = crtc->dev;
+	struct drm_connector *connector;
+	int bpp;
+
+	switch (fb->pixel_format) {
+	case DRM_FORMAT_C8:
+		bpp = 8*3; /* since we go through a colormap */
+		break;
+	case DRM_FORMAT_XRGB1555:
+	case DRM_FORMAT_ARGB1555:
+		/* checked in intel_framebuffer_init already */
+		if (WARN_ON(INTEL_INFO(dev)->gen > 3))
+			return -EINVAL;
+	case DRM_FORMAT_RGB565:
+		bpp = 6*3; /* min is 18bpp */
+		break;
+	case DRM_FORMAT_XBGR8888:
+	case DRM_FORMAT_ABGR8888:
+		/* checked in intel_framebuffer_init already */
+		if (WARN_ON(INTEL_INFO(dev)->gen < 4))
+			return -EINVAL;
+	case DRM_FORMAT_XRGB8888:
+	case DRM_FORMAT_ARGB8888:
+		bpp = 8*3;
+		break;
+	case DRM_FORMAT_XRGB2101010:
+	case DRM_FORMAT_ARGB2101010:
+	case DRM_FORMAT_XBGR2101010:
+	case DRM_FORMAT_ABGR2101010:
+		/* checked in intel_framebuffer_init already */
+		if (WARN_ON(INTEL_INFO(dev)->gen < 4))
+			return -EINVAL;
+		bpp = 10*3;
+		break;
+	/* TODO: gen4+ supports 16 bpc floating point, too. */
+	default:
+		DRM_DEBUG_KMS("unsupported depth\n");
+		return -EINVAL;
+	}
+
+	pipe_config->pipe_bpp = bpp;
+
+	/* Clamp display bpp to EDID value */
+	list_for_each_entry(connector, &dev->mode_config.connector_list,
+			    head) {
+		if (connector->encoder && connector->encoder->crtc != crtc)
+			continue;
+
+		/* Don't use an invalid EDID bpc value */
+		if (connector->display_info.bpc &&
+		    connector->display_info.bpc * 3 < bpp) {
+			DRM_DEBUG_KMS("clamping display bpp (was %d) to EDID reported max of %d\n",
+				      bpp, connector->display_info.bpc*3);
+			pipe_config->pipe_bpp = connector->display_info.bpc*3;
+		}
+	}
+
+	return bpp;
+}
+
+static struct intel_crtc_config *
+intel_modeset_pipe_config(struct drm_crtc *crtc,
+			  struct drm_framebuffer *fb,
+			  struct drm_display_mode *mode)
 {
 	struct drm_device *dev = crtc->dev;
-	struct drm_display_mode *adjusted_mode;
 	struct drm_encoder_helper_funcs *encoder_funcs;
 	struct intel_encoder *encoder;
+	struct intel_crtc_config *pipe_config;
+	int plane_bpp;
 
-	adjusted_mode = drm_mode_duplicate(dev, mode);
-	if (!adjusted_mode)
+	pipe_config = kzalloc(sizeof(*pipe_config), GFP_KERNEL);
+	if (!pipe_config)
 		return ERR_PTR(-ENOMEM);
 
+	drm_mode_copy(&pipe_config->adjusted_mode, mode);
+	drm_mode_copy(&pipe_config->requested_mode, mode);
+
+	plane_bpp = pipe_config_set_bpp(crtc, fb, pipe_config);
+	if (plane_bpp < 0)
+		goto fail;
+
 	/* Pass our mode to the connectors and the CRTC to give them a chance to
 	 * adjust it according to limitations or connector properties, and also
 	 * a chance to reject the mode entirely.
@@ -7495,23 +7627,38 @@ intel_modeset_adjusted_mode(struct drm_crtc *crtc,
 
 		if (&encoder->new_crtc->base != crtc)
 			continue;
+
+		if (encoder->compute_config) {
+			if (!(encoder->compute_config(encoder, pipe_config))) {
+				DRM_DEBUG_KMS("Encoder config failure\n");
+				goto fail;
+			}
+
+			continue;
+		}
+
 		encoder_funcs = encoder->base.helper_private;
-		if (!(encoder_funcs->mode_fixup(&encoder->base, mode,
-						adjusted_mode))) {
+		if (!(encoder_funcs->mode_fixup(&encoder->base,
+						&pipe_config->requested_mode,
+						&pipe_config->adjusted_mode))) {
 			DRM_DEBUG_KMS("Encoder fixup failed\n");
 			goto fail;
 		}
 	}
 
-	if (!(intel_crtc_mode_fixup(crtc, mode, adjusted_mode))) {
+	if (!(intel_crtc_compute_config(crtc, pipe_config))) {
 		DRM_DEBUG_KMS("CRTC fixup failed\n");
 		goto fail;
 	}
 	DRM_DEBUG_KMS("[CRTC:%d]\n", crtc->base.id);
 
-	return adjusted_mode;
+	pipe_config->dither = pipe_config->pipe_bpp != plane_bpp;
+	DRM_DEBUG_KMS("plane bpp: %i, pipe bpp: %i, dithering: %i\n",
+		      plane_bpp, pipe_config->pipe_bpp, pipe_config->dither);
+
+	return pipe_config;
 fail:
-	drm_mode_destroy(dev, adjusted_mode);
+	kfree(pipe_config);
 	return ERR_PTR(-EINVAL);
 }
 
@@ -7589,22 +7736,25 @@ intel_modeset_affected_pipes(struct drm_crtc *crtc, unsigned *modeset_pipes,
 	if (crtc->enabled)
 		*prepare_pipes |= 1 << intel_crtc->pipe;
 
-	/* We only support modeset on one single crtc, hence we need to do that
-	 * only for the passed in crtc iff we change anything else than just
-	 * disable crtcs.
-	 *
-	 * This is actually not true, to be fully compatible with the old crtc
-	 * helper we automatically disable _any_ output (i.e. doesn't need to be
-	 * connected to the crtc we're modesetting on) if it's disconnected.
-	 * Which is a rather nutty api (since changed the output configuration
-	 * without userspace's explicit request can lead to confusion), but
-	 * alas. Hence we currently need to modeset on all pipes we prepare. */
+	/*
+	 * For simplicity do a full modeset on any pipe where the output routing
+	 * changed. We could be more clever, but that would require us to be
+	 * more careful with calling the relevant encoder->mode_set functions.
+	 */
 	if (*prepare_pipes)
 		*modeset_pipes = *prepare_pipes;
 
 	/* ... and mask these out. */
 	*modeset_pipes &= ~(*disable_pipes);
 	*prepare_pipes &= ~(*disable_pipes);
+
+	/*
+	 * HACK: We don't (yet) fully support global modesets. intel_set_config
+	 * obies this rule, but the modeset restore mode of
+	 * intel_modeset_setup_hw_state does not.
+	 */
+	*modeset_pipes &= 1 << intel_crtc->pipe;
+	*prepare_pipes &= 1 << intel_crtc->pipe;
 }
 
 static bool intel_crtc_in_use(struct drm_crtc *crtc)
@@ -7673,12 +7823,29 @@ intel_modeset_update_state(struct drm_device *dev, unsigned prepare_pipes)
 			    base.head) \
 		if (mask & (1 <<(intel_crtc)->pipe)) \
 
+static bool
+intel_pipe_config_compare(struct intel_crtc_config *current_config,
+			  struct intel_crtc_config *pipe_config)
+{
+	if (current_config->has_pch_encoder != pipe_config->has_pch_encoder) {
+		DRM_ERROR("mismatch in has_pch_encoder "
+			  "(expected %i, found %i)\n",
+			  current_config->has_pch_encoder,
+			  pipe_config->has_pch_encoder);
+		return false;
+	}
+
+	return true;
+}
+
 void
 intel_modeset_check_state(struct drm_device *dev)
 {
+	drm_i915_private_t *dev_priv = dev->dev_private;
 	struct intel_crtc *crtc;
 	struct intel_encoder *encoder;
 	struct intel_connector *connector;
+	struct intel_crtc_config pipe_config;
 
 	list_for_each_entry(connector, &dev->mode_config.connector_list,
 			    base.head) {
@@ -7767,17 +7934,27 @@ intel_modeset_check_state(struct drm_device *dev)
 		     "crtc's computed enabled state doesn't match tracked enabled state "
 		     "(expected %i, found %i)\n", enabled, crtc->base.enabled);
 
-		assert_pipe(dev->dev_private, crtc->pipe, crtc->active);
+		memset(&pipe_config, 0, sizeof(pipe_config));
+		active = dev_priv->display.get_pipe_config(crtc,
+							   &pipe_config);
+		WARN(crtc->active != active,
+		     "crtc active state doesn't match with hw state "
+		     "(expected %i, found %i)\n", crtc->active, active);
+
+		WARN(active &&
+		     !intel_pipe_config_compare(&crtc->config, &pipe_config),
+		     "pipe state doesn't match!\n");
 	}
 }
 
-int intel_set_mode(struct drm_crtc *crtc,
-		   struct drm_display_mode *mode,
-		   int x, int y, struct drm_framebuffer *fb)
+static int __intel_set_mode(struct drm_crtc *crtc,
+			    struct drm_display_mode *mode,
+			    int x, int y, struct drm_framebuffer *fb)
 {
 	struct drm_device *dev = crtc->dev;
 	drm_i915_private_t *dev_priv = dev->dev_private;
-	struct drm_display_mode *adjusted_mode, *saved_mode, *saved_hwmode;
+	struct drm_display_mode *saved_mode, *saved_hwmode;
+	struct intel_crtc_config *pipe_config = NULL;
 	struct intel_crtc *intel_crtc;
 	unsigned disable_pipes, prepare_pipes, modeset_pipes;
 	int ret = 0;
@@ -7790,12 +7967,6 @@ int intel_set_mode(struct drm_crtc *crtc,
 	intel_modeset_affected_pipes(crtc, &modeset_pipes,
 				     &prepare_pipes, &disable_pipes);
 
-	DRM_DEBUG_KMS("set mode pipe masks: modeset: %x, prepare: %x, disable: %x\n",
-		      modeset_pipes, prepare_pipes, disable_pipes);
-
-	for_each_intel_crtc_masked(dev, disable_pipes, intel_crtc)
-		intel_crtc_disable(&intel_crtc->base);
-
 	*saved_hwmode = crtc->hwmode;
 	*saved_mode = crtc->mode;
 
@@ -7804,15 +7975,22 @@ int intel_set_mode(struct drm_crtc *crtc,
 	 * Hence simply check whether any bit is set in modeset_pipes in all the
 	 * pieces of code that are not yet converted to deal with mutliple crtcs
 	 * changing their mode at the same time. */
-	adjusted_mode = NULL;
 	if (modeset_pipes) {
-		adjusted_mode = intel_modeset_adjusted_mode(crtc, mode);
-		if (IS_ERR(adjusted_mode)) {
-			ret = PTR_ERR(adjusted_mode);
+		pipe_config = intel_modeset_pipe_config(crtc, fb, mode);
+		if (IS_ERR(pipe_config)) {
+			ret = PTR_ERR(pipe_config);
+			pipe_config = NULL;
+
 			goto out;
 		}
 	}
 
+	DRM_DEBUG_KMS("set mode pipe masks: modeset: %x, prepare: %x, disable: %x\n",
+		      modeset_pipes, prepare_pipes, disable_pipes);
+
+	for_each_intel_crtc_masked(dev, disable_pipes, intel_crtc)
+		intel_crtc_disable(&intel_crtc->base);
+
 	for_each_intel_crtc_masked(dev, prepare_pipes, intel_crtc) {
 		if (intel_crtc->base.enabled)
 			dev_priv->display.crtc_disable(&intel_crtc->base);
@@ -7821,8 +7999,14 @@ int intel_set_mode(struct drm_crtc *crtc,
 	/* crtc->mode is already used by the ->mode_set callbacks, hence we need
 	 * to set it here already despite that we pass it down the callchain.
 	 */
-	if (modeset_pipes)
+	if (modeset_pipes) {
+		enum transcoder tmp = to_intel_crtc(crtc)->config.cpu_transcoder;
 		crtc->mode = *mode;
+		/* mode_set/enable/disable functions rely on a correct pipe
+		 * config. */
+		to_intel_crtc(crtc)->config = *pipe_config;
+		to_intel_crtc(crtc)->config.cpu_transcoder = tmp;
+	}
 
 	/* Only after disabling all output pipelines that will be changed can we
 	 * update the the output configuration. */
@@ -7836,7 +8020,6 @@ int intel_set_mode(struct drm_crtc *crtc,
 	 */
 	for_each_intel_crtc_masked(dev, modeset_pipes, intel_crtc) {
 		ret = intel_crtc_mode_set(&intel_crtc->base,
-					  mode, adjusted_mode,
 					  x, y, fb);
 		if (ret)
 			goto done;
@@ -7848,7 +8031,7 @@ int intel_set_mode(struct drm_crtc *crtc,
 
 	if (modeset_pipes) {
 		/* Store real post-adjustment hardware mode. */
-		crtc->hwmode = *adjusted_mode;
+		crtc->hwmode = pipe_config->adjusted_mode;
 
 		/* Calculate and store various constants which
 		 * are later needed by vblank and swap-completion
@@ -7859,19 +8042,31 @@ int intel_set_mode(struct drm_crtc *crtc,
 
 	/* FIXME: add subpixel order */
 done:
-	drm_mode_destroy(dev, adjusted_mode);
 	if (ret && crtc->enabled) {
 		crtc->hwmode = *saved_hwmode;
 		crtc->mode = *saved_mode;
-	} else {
-		intel_modeset_check_state(dev);
 	}
 
 out:
+	kfree(pipe_config);
 	kfree(saved_mode);
 	return ret;
 }
 
+int intel_set_mode(struct drm_crtc *crtc,
+		     struct drm_display_mode *mode,
+		     int x, int y, struct drm_framebuffer *fb)
+{
+	int ret;
+
+	ret = __intel_set_mode(crtc, mode, x, y, fb);
+
+	if (ret == 0)
+		intel_modeset_check_state(crtc->dev);
+
+	return ret;
+}
+
 void intel_crtc_restore_mode(struct drm_crtc *crtc)
 {
 	intel_set_mode(crtc, &crtc->mode, crtc->x, crtc->y, crtc->fb);
@@ -7959,10 +8154,8 @@ intel_set_config_compute_mode_changes(struct drm_mode_set *set,
 			config->mode_changed = true;
 		} else if (set->fb == NULL) {
 			config->mode_changed = true;
-		} else if (set->fb->depth != set->crtc->fb->depth) {
-			config->mode_changed = true;
-		} else if (set->fb->bits_per_pixel !=
-			   set->crtc->fb->bits_per_pixel) {
+		} else if (set->fb->pixel_format !=
+			   set->crtc->fb->pixel_format) {
 			config->mode_changed = true;
 		} else
 			config->fb_changed = true;
@@ -8145,6 +8338,8 @@ static int intel_crtc_set_config(struct drm_mode_set *set)
 			goto fail;
 		}
 	} else if (config->fb_changed) {
+		intel_crtc_wait_for_pending_flips(set->crtc);
+
 		ret = intel_pipe_set_base(set->crtc,
 					  set->x, set->y, set->fb);
 	}
@@ -8221,7 +8416,7 @@ static void intel_crtc_init(struct drm_device *dev, int pipe)
 	/* Swap pipes & planes for FBC on pre-965 */
 	intel_crtc->pipe = pipe;
 	intel_crtc->plane = pipe;
-	intel_crtc->cpu_transcoder = pipe;
+	intel_crtc->config.cpu_transcoder = pipe;
 	if (IS_MOBILE(dev) && IS_GEN3(dev)) {
 		DRM_DEBUG_KMS("swapping pipes & planes for FBC\n");
 		intel_crtc->plane = !pipe;
@@ -8232,8 +8427,6 @@ static void intel_crtc_init(struct drm_device *dev, int pipe)
 	dev_priv->plane_to_crtc_mapping[intel_crtc->plane] = &intel_crtc->base;
 	dev_priv->pipe_to_crtc_mapping[intel_crtc->pipe] = &intel_crtc->base;
 
-	intel_crtc->bpp = 24; /* default for pre-Ironlake */
-
 	drm_crtc_helper_add(&intel_crtc->base, &intel_helper_funcs);
 }
 
@@ -8314,7 +8507,7 @@ static void intel_setup_outputs(struct drm_device *dev)
 		I915_WRITE(PFIT_CONTROL, 0);
 	}
 
-	if (!(HAS_DDI(dev) && (I915_READ(DDI_BUF_CTL(PORT_A)) & DDI_A_4_LANES)))
+	if (!IS_ULT(dev))
 		intel_crt_init(dev);
 
 	if (HAS_DDI(dev)) {
@@ -8343,20 +8536,20 @@ static void intel_setup_outputs(struct drm_device *dev)
 		if (has_edp_a(dev))
 			intel_dp_init(dev, DP_A, PORT_A);
 
-		if (I915_READ(HDMIB) & PORT_DETECTED) {
+		if (I915_READ(PCH_HDMIB) & SDVO_DETECTED) {
 			/* PCH SDVOB multiplex with HDMIB */
 			found = intel_sdvo_init(dev, PCH_SDVOB, true);
 			if (!found)
-				intel_hdmi_init(dev, HDMIB, PORT_B);
+				intel_hdmi_init(dev, PCH_HDMIB, PORT_B);
 			if (!found && (I915_READ(PCH_DP_B) & DP_DETECTED))
 				intel_dp_init(dev, PCH_DP_B, PORT_B);
 		}
 
-		if (I915_READ(HDMIC) & PORT_DETECTED)
-			intel_hdmi_init(dev, HDMIC, PORT_C);
+		if (I915_READ(PCH_HDMIC) & SDVO_DETECTED)
+			intel_hdmi_init(dev, PCH_HDMIC, PORT_C);
 
-		if (!dpd_is_edp && I915_READ(HDMID) & PORT_DETECTED)
-			intel_hdmi_init(dev, HDMID, PORT_D);
+		if (!dpd_is_edp && I915_READ(PCH_HDMID) & SDVO_DETECTED)
+			intel_hdmi_init(dev, PCH_HDMID, PORT_D);
 
 		if (I915_READ(PCH_DP_C) & DP_DETECTED)
 			intel_dp_init(dev, PCH_DP_C, PORT_C);
@@ -8368,24 +8561,21 @@ static void intel_setup_outputs(struct drm_device *dev)
 		if (I915_READ(VLV_DISPLAY_BASE + DP_C) & DP_DETECTED)
 			intel_dp_init(dev, VLV_DISPLAY_BASE + DP_C, PORT_C);
 
-		if (I915_READ(VLV_DISPLAY_BASE + SDVOB) & PORT_DETECTED) {
-			intel_hdmi_init(dev, VLV_DISPLAY_BASE + SDVOB, PORT_B);
+		if (I915_READ(VLV_DISPLAY_BASE + GEN4_HDMIB) & SDVO_DETECTED) {
+			intel_hdmi_init(dev, VLV_DISPLAY_BASE + GEN4_HDMIB,
+					PORT_B);
 			if (I915_READ(VLV_DISPLAY_BASE + DP_B) & DP_DETECTED)
 				intel_dp_init(dev, VLV_DISPLAY_BASE + DP_B, PORT_B);
 		}
-
-		if (I915_READ(VLV_DISPLAY_BASE + SDVOC) & PORT_DETECTED)
-			intel_hdmi_init(dev, VLV_DISPLAY_BASE + SDVOC, PORT_C);
-
 	} else if (SUPPORTS_DIGITAL_OUTPUTS(dev)) {
 		bool found = false;
 
-		if (I915_READ(SDVOB) & SDVO_DETECTED) {
+		if (I915_READ(GEN3_SDVOB) & SDVO_DETECTED) {
 			DRM_DEBUG_KMS("probing SDVOB\n");
-			found = intel_sdvo_init(dev, SDVOB, true);
+			found = intel_sdvo_init(dev, GEN3_SDVOB, true);
 			if (!found && SUPPORTS_INTEGRATED_HDMI(dev)) {
 				DRM_DEBUG_KMS("probing HDMI on SDVOB\n");
-				intel_hdmi_init(dev, SDVOB, PORT_B);
+				intel_hdmi_init(dev, GEN4_HDMIB, PORT_B);
 			}
 
 			if (!found && SUPPORTS_INTEGRATED_DP(dev)) {
@@ -8396,16 +8586,16 @@ static void intel_setup_outputs(struct drm_device *dev)
 
 		/* Before G4X SDVOC doesn't have its own detect register */
 
-		if (I915_READ(SDVOB) & SDVO_DETECTED) {
+		if (I915_READ(GEN3_SDVOB) & SDVO_DETECTED) {
 			DRM_DEBUG_KMS("probing SDVOC\n");
-			found = intel_sdvo_init(dev, SDVOC, false);
+			found = intel_sdvo_init(dev, GEN3_SDVOC, false);
 		}
 
-		if (!found && (I915_READ(SDVOC) & SDVO_DETECTED)) {
+		if (!found && (I915_READ(GEN3_SDVOC) & SDVO_DETECTED)) {
 
 			if (SUPPORTS_INTEGRATED_HDMI(dev)) {
 				DRM_DEBUG_KMS("probing HDMI on SDVOC\n");
-				intel_hdmi_init(dev, SDVOC, PORT_C);
+				intel_hdmi_init(dev, GEN4_HDMIC, PORT_C);
 			}
 			if (SUPPORTS_INTEGRATED_DP(dev)) {
 				DRM_DEBUG_KMS("probing DP_C\n");
@@ -8572,20 +8762,22 @@ static void intel_init_display(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 
-	/* We always want a DPMS function */
 	if (HAS_DDI(dev)) {
+		dev_priv->display.get_pipe_config = haswell_get_pipe_config;
 		dev_priv->display.crtc_mode_set = haswell_crtc_mode_set;
 		dev_priv->display.crtc_enable = haswell_crtc_enable;
 		dev_priv->display.crtc_disable = haswell_crtc_disable;
 		dev_priv->display.off = haswell_crtc_off;
 		dev_priv->display.update_plane = ironlake_update_plane;
 	} else if (HAS_PCH_SPLIT(dev)) {
+		dev_priv->display.get_pipe_config = ironlake_get_pipe_config;
 		dev_priv->display.crtc_mode_set = ironlake_crtc_mode_set;
 		dev_priv->display.crtc_enable = ironlake_crtc_enable;
 		dev_priv->display.crtc_disable = ironlake_crtc_disable;
 		dev_priv->display.off = ironlake_crtc_off;
 		dev_priv->display.update_plane = ironlake_update_plane;
 	} else {
+		dev_priv->display.get_pipe_config = i9xx_get_pipe_config;
 		dev_priv->display.crtc_mode_set = i9xx_crtc_mode_set;
 		dev_priv->display.crtc_enable = i9xx_crtc_enable;
 		dev_priv->display.crtc_disable = i9xx_crtc_disable;
@@ -8828,7 +9020,7 @@ void intel_modeset_init_hw(struct drm_device *dev)
 void intel_modeset_init(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	int i, ret;
+	int i, j, ret;
 
 	drm_mode_config_init(dev);
 
@@ -8844,6 +9036,9 @@ void intel_modeset_init(struct drm_device *dev)
 
 	intel_init_pm(dev);
 
+	if (INTEL_INFO(dev)->num_pipes == 0)
+		return;
+
 	intel_init_display(dev);
 
 	if (IS_GEN2(dev)) {
@@ -8859,13 +9054,17 @@ void intel_modeset_init(struct drm_device *dev)
 	dev->mode_config.fb_base = dev_priv->gtt.mappable_base;
 
 	DRM_DEBUG_KMS("%d display pipe%s available.\n",
-		      dev_priv->num_pipe, dev_priv->num_pipe > 1 ? "s" : "");
+		      INTEL_INFO(dev)->num_pipes,
+		      INTEL_INFO(dev)->num_pipes > 1 ? "s" : "");
 
-	for (i = 0; i < dev_priv->num_pipe; i++) {
+	for (i = 0; i < INTEL_INFO(dev)->num_pipes; i++) {
 		intel_crtc_init(dev, i);
-		ret = intel_plane_init(dev, i);
-		if (ret)
-			DRM_DEBUG_KMS("plane %d init failed: %d\n", i, ret);
+		for (j = 0; j < dev_priv->num_plane; j++) {
+			ret = intel_plane_init(dev, i, j);
+			if (ret)
+				DRM_DEBUG_KMS("pipe %d plane %d init failed: %d\n",
+					      i, j, ret);
+		}
 	}
 
 	intel_cpu_pll_init(dev);
@@ -8918,10 +9117,11 @@ static void intel_enable_pipe_a(struct drm_device *dev)
 static bool
 intel_check_plane_mapping(struct intel_crtc *crtc)
 {
-	struct drm_i915_private *dev_priv = crtc->base.dev->dev_private;
+	struct drm_device *dev = crtc->base.dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
 	u32 reg, val;
 
-	if (dev_priv->num_pipe == 1)
+	if (INTEL_INFO(dev)->num_pipes == 1)
 		return true;
 
 	reg = DSPCNTR(!crtc->plane);
@@ -8941,7 +9141,7 @@ static void intel_sanitize_crtc(struct intel_crtc *crtc)
 	u32 reg;
 
 	/* Clear any frame start delays used for debugging left by the BIOS */
-	reg = PIPECONF(crtc->cpu_transcoder);
+	reg = PIPECONF(crtc->config.cpu_transcoder);
 	I915_WRITE(reg, I915_READ(reg) & ~PIPECONF_FRAME_START_DELAY_MASK);
 
 	/* We need to sanitize the plane -> pipe mapping first because this will
@@ -9077,6 +9277,7 @@ void intel_modeset_setup_hw_state(struct drm_device *dev,
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	enum pipe pipe;
 	u32 tmp;
+	struct drm_plane *plane;
 	struct intel_crtc *crtc;
 	struct intel_encoder *encoder;
 	struct intel_connector *connector;
@@ -9096,24 +9297,32 @@ void intel_modeset_setup_hw_state(struct drm_device *dev,
 			case TRANS_DDI_EDP_INPUT_C_ONOFF:
 				pipe = PIPE_C;
 				break;
+			default:
+				/* A bogus value has been programmed, disable
+				 * the transcoder */
+				WARN(1, "Bogus eDP source %08x\n", tmp);
+				intel_ddi_disable_transcoder_func(dev_priv,
+						TRANSCODER_EDP);
+				goto setup_pipes;
 			}
 
 			crtc = to_intel_crtc(dev_priv->pipe_to_crtc_mapping[pipe]);
-			crtc->cpu_transcoder = TRANSCODER_EDP;
+			crtc->config.cpu_transcoder = TRANSCODER_EDP;
 
 			DRM_DEBUG_KMS("Pipe %c using transcoder EDP\n",
 				      pipe_name(pipe));
 		}
 	}
 
-	for_each_pipe(pipe) {
-		crtc = to_intel_crtc(dev_priv->pipe_to_crtc_mapping[pipe]);
+setup_pipes:
+	list_for_each_entry(crtc, &dev->mode_config.crtc_list,
+			    base.head) {
+		enum transcoder tmp = crtc->config.cpu_transcoder;
+		memset(&crtc->config, 0, sizeof(crtc->config));
+		crtc->config.cpu_transcoder = tmp;
 
-		tmp = I915_READ(PIPECONF(crtc->cpu_transcoder));
-		if (tmp & PIPECONF_ENABLE)
-			crtc->active = true;
-		else
-			crtc->active = false;
+		crtc->active = dev_priv->display.get_pipe_config(crtc,
+								 &crtc->config);
 
 		crtc->base.enabled = crtc->active;
 
@@ -9172,9 +9381,19 @@ void intel_modeset_setup_hw_state(struct drm_device *dev,
 	}
 
 	if (force_restore) {
+		/*
+		 * We need to use raw interfaces for restoring state to avoid
+		 * checking (bogus) intermediate states.
+		 */
 		for_each_pipe(pipe) {
-			intel_crtc_restore_mode(dev_priv->pipe_to_crtc_mapping[pipe]);
+			struct drm_crtc *crtc =
+				dev_priv->pipe_to_crtc_mapping[pipe];
+
+			__intel_set_mode(crtc, &crtc->mode, crtc->x, crtc->y,
+					 crtc->fb);
 		}
+		list_for_each_entry(plane, &dev->mode_config.plane_list, head)
+			intel_plane_restore(plane);
 
 		i915_redisable_vga(dev);
 	} else {
@@ -9236,6 +9455,9 @@ void intel_modeset_cleanup(struct drm_device *dev)
 	/* flush any delayed tasks or pending work */
 	flush_scheduled_work();
 
+	/* destroy backlight, if any, before the connectors */
+	intel_panel_destroy_backlight(dev);
+
 	drm_mode_config_cleanup(dev);
 
 	intel_cleanup_overlay(dev);
@@ -9323,15 +9545,24 @@ intel_display_capture_error_state(struct drm_device *dev)
 	for_each_pipe(i) {
 		cpu_transcoder = intel_pipe_to_cpu_transcoder(dev_priv, i);
 
-		error->cursor[i].control = I915_READ(CURCNTR(i));
-		error->cursor[i].position = I915_READ(CURPOS(i));
-		error->cursor[i].base = I915_READ(CURBASE(i));
+		if (INTEL_INFO(dev)->gen <= 6 || IS_VALLEYVIEW(dev)) {
+			error->cursor[i].control = I915_READ(CURCNTR(i));
+			error->cursor[i].position = I915_READ(CURPOS(i));
+			error->cursor[i].base = I915_READ(CURBASE(i));
+		} else {
+			error->cursor[i].control = I915_READ(CURCNTR_IVB(i));
+			error->cursor[i].position = I915_READ(CURPOS_IVB(i));
+			error->cursor[i].base = I915_READ(CURBASE_IVB(i));
+		}
 
 		error->plane[i].control = I915_READ(DSPCNTR(i));
 		error->plane[i].stride = I915_READ(DSPSTRIDE(i));
-		error->plane[i].size = I915_READ(DSPSIZE(i));
-		error->plane[i].pos = I915_READ(DSPPOS(i));
-		error->plane[i].addr = I915_READ(DSPADDR(i));
+		if (INTEL_INFO(dev)->gen <= 3) {
+			error->plane[i].size = I915_READ(DSPSIZE(i));
+			error->plane[i].pos = I915_READ(DSPPOS(i));
+		}
+		if (INTEL_INFO(dev)->gen <= 7 && !IS_HASWELL(dev))
+			error->plane[i].addr = I915_READ(DSPADDR(i));
 		if (INTEL_INFO(dev)->gen >= 4) {
 			error->plane[i].surface = I915_READ(DSPSURF(i));
 			error->plane[i].tile_offset = I915_READ(DSPTILEOFF(i));
@@ -9355,10 +9586,9 @@ intel_display_print_error_state(struct seq_file *m,
 				struct drm_device *dev,
 				struct intel_display_error_state *error)
 {
-	drm_i915_private_t *dev_priv = dev->dev_private;
 	int i;
 
-	seq_printf(m, "Num Pipes: %d\n", dev_priv->num_pipe);
+	seq_printf(m, "Num Pipes: %d\n", INTEL_INFO(dev)->num_pipes);
 	for_each_pipe(i) {
 		seq_printf(m, "Pipe [%d]:\n", i);
 		seq_printf(m, "  CONF: %08x\n", error->pipe[i].conf);
@@ -9373,9 +9603,12 @@ intel_display_print_error_state(struct seq_file *m,
 		seq_printf(m, "Plane [%d]:\n", i);
 		seq_printf(m, "  CNTR: %08x\n", error->plane[i].control);
 		seq_printf(m, "  STRIDE: %08x\n", error->plane[i].stride);
-		seq_printf(m, "  SIZE: %08x\n", error->plane[i].size);
-		seq_printf(m, "  POS: %08x\n", error->plane[i].pos);
-		seq_printf(m, "  ADDR: %08x\n", error->plane[i].addr);
+		if (INTEL_INFO(dev)->gen <= 3) {
+			seq_printf(m, "  SIZE: %08x\n", error->plane[i].size);
+			seq_printf(m, "  POS: %08x\n", error->plane[i].pos);
+		}
+		if (INTEL_INFO(dev)->gen <= 7 && !IS_HASWELL(dev))
+			seq_printf(m, "  ADDR: %08x\n", error->plane[i].addr);
 		if (INTEL_INFO(dev)->gen >= 4) {
 			seq_printf(m, "  SURF: %08x\n", error->plane[i].surface);
 			seq_printf(m, "  TILEOFF: %08x\n", error->plane[i].tile_offset);
diff --git a/drivers/gpu/drm/i915/intel_dp.c b/drivers/gpu/drm/i915/intel_dp.c
index c3f5bd8a5077..fb2fbc1e08b9 100644
--- a/drivers/gpu/drm/i915/intel_dp.c
+++ b/drivers/gpu/drm/i915/intel_dp.c
@@ -109,29 +109,6 @@ bool intel_encoder_is_pch_edp(struct drm_encoder *encoder)
 
 static void intel_dp_link_down(struct intel_dp *intel_dp);
 
-void
-intel_edp_link_config(struct intel_encoder *intel_encoder,
-		       int *lane_num, int *link_bw)
-{
-	struct intel_dp *intel_dp = enc_to_intel_dp(&intel_encoder->base);
-
-	*lane_num = intel_dp->lane_count;
-	*link_bw = drm_dp_bw_code_to_link_rate(intel_dp->link_bw);
-}
-
-int
-intel_edp_target_clock(struct intel_encoder *intel_encoder,
-		       struct drm_display_mode *mode)
-{
-	struct intel_dp *intel_dp = enc_to_intel_dp(&intel_encoder->base);
-	struct intel_connector *intel_connector = intel_dp->attached_connector;
-
-	if (intel_connector->panel.fixed_mode)
-		return intel_connector->panel.fixed_mode->clock;
-	else
-		return mode->clock;
-}
-
 static int
 intel_dp_max_link_bw(struct intel_dp *intel_dp)
 {
@@ -177,34 +154,6 @@ intel_dp_max_data_rate(int max_link_clock, int max_lanes)
 	return (max_link_clock * max_lanes * 8) / 10;
 }
 
-static bool
-intel_dp_adjust_dithering(struct intel_dp *intel_dp,
-			  struct drm_display_mode *mode,
-			  bool adjust_mode)
-{
-	int max_link_clock =
-		drm_dp_bw_code_to_link_rate(intel_dp_max_link_bw(intel_dp));
-	int max_lanes = drm_dp_max_lane_count(intel_dp->dpcd);
-	int max_rate, mode_rate;
-
-	mode_rate = intel_dp_link_required(mode->clock, 24);
-	max_rate = intel_dp_max_data_rate(max_link_clock, max_lanes);
-
-	if (mode_rate > max_rate) {
-		mode_rate = intel_dp_link_required(mode->clock, 18);
-		if (mode_rate > max_rate)
-			return false;
-
-		if (adjust_mode)
-			mode->private_flags
-				|= INTEL_MODE_DP_FORCE_6BPC;
-
-		return true;
-	}
-
-	return true;
-}
-
 static int
 intel_dp_mode_valid(struct drm_connector *connector,
 		    struct drm_display_mode *mode)
@@ -212,6 +161,8 @@ intel_dp_mode_valid(struct drm_connector *connector,
 	struct intel_dp *intel_dp = intel_attached_dp(connector);
 	struct intel_connector *intel_connector = to_intel_connector(connector);
 	struct drm_display_mode *fixed_mode = intel_connector->panel.fixed_mode;
+	int target_clock = mode->clock;
+	int max_rate, mode_rate, max_lanes, max_link_clock;
 
 	if (is_edp(intel_dp) && fixed_mode) {
 		if (mode->hdisplay > fixed_mode->hdisplay)
@@ -219,9 +170,17 @@ intel_dp_mode_valid(struct drm_connector *connector,
 
 		if (mode->vdisplay > fixed_mode->vdisplay)
 			return MODE_PANEL;
+
+		target_clock = fixed_mode->clock;
 	}
 
-	if (!intel_dp_adjust_dithering(intel_dp, mode, false))
+	max_link_clock = drm_dp_bw_code_to_link_rate(intel_dp_max_link_bw(intel_dp));
+	max_lanes = drm_dp_max_lane_count(intel_dp->dpcd);
+
+	max_rate = intel_dp_max_data_rate(max_link_clock, max_lanes);
+	mode_rate = intel_dp_link_required(target_clock, 18);
+
+	if (mode_rate > max_rate)
 		return MODE_CLOCK_HIGH;
 
 	if (mode->clock < 10000)
@@ -294,16 +253,20 @@ static bool ironlake_edp_have_panel_power(struct intel_dp *intel_dp)
 {
 	struct drm_device *dev = intel_dp_to_dev(intel_dp);
 	struct drm_i915_private *dev_priv = dev->dev_private;
+	u32 pp_stat_reg;
 
-	return (I915_READ(PCH_PP_STATUS) & PP_ON) != 0;
+	pp_stat_reg = IS_VALLEYVIEW(dev) ? PIPEA_PP_STATUS : PCH_PP_STATUS;
+	return (I915_READ(pp_stat_reg) & PP_ON) != 0;
 }
 
 static bool ironlake_edp_have_panel_vdd(struct intel_dp *intel_dp)
 {
 	struct drm_device *dev = intel_dp_to_dev(intel_dp);
 	struct drm_i915_private *dev_priv = dev->dev_private;
+	u32 pp_ctrl_reg;
 
-	return (I915_READ(PCH_PP_CONTROL) & EDP_FORCE_VDD) != 0;
+	pp_ctrl_reg = IS_VALLEYVIEW(dev) ? PIPEA_PP_CONTROL : PCH_PP_CONTROL;
+	return (I915_READ(pp_ctrl_reg) & EDP_FORCE_VDD) != 0;
 }
 
 static void
@@ -311,14 +274,19 @@ intel_dp_check_edp(struct intel_dp *intel_dp)
 {
 	struct drm_device *dev = intel_dp_to_dev(intel_dp);
 	struct drm_i915_private *dev_priv = dev->dev_private;
+	u32 pp_stat_reg, pp_ctrl_reg;
 
 	if (!is_edp(intel_dp))
 		return;
+
+	pp_stat_reg = IS_VALLEYVIEW(dev) ? PIPEA_PP_STATUS : PCH_PP_STATUS;
+	pp_ctrl_reg = IS_VALLEYVIEW(dev) ? PIPEA_PP_CONTROL : PCH_PP_CONTROL;
+
 	if (!ironlake_edp_have_panel_power(intel_dp) && !ironlake_edp_have_panel_vdd(intel_dp)) {
 		WARN(1, "eDP powered off while attempting aux channel communication.\n");
 		DRM_DEBUG_KMS("Status 0x%08x Control 0x%08x\n",
-			      I915_READ(PCH_PP_STATUS),
-			      I915_READ(PCH_PP_CONTROL));
+				I915_READ(pp_stat_reg),
+				I915_READ(pp_ctrl_reg));
 	}
 }
 
@@ -328,29 +296,10 @@ intel_dp_aux_wait_done(struct intel_dp *intel_dp, bool has_aux_irq)
 	struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp);
 	struct drm_device *dev = intel_dig_port->base.base.dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	uint32_t ch_ctl = intel_dp->output_reg + 0x10;
+	uint32_t ch_ctl = intel_dp->aux_ch_ctl_reg;
 	uint32_t status;
 	bool done;
 
-	if (IS_HASWELL(dev)) {
-		switch (intel_dig_port->port) {
-		case PORT_A:
-			ch_ctl = DPA_AUX_CH_CTL;
-			break;
-		case PORT_B:
-			ch_ctl = PCH_DPB_AUX_CH_CTL;
-			break;
-		case PORT_C:
-			ch_ctl = PCH_DPC_AUX_CH_CTL;
-			break;
-		case PORT_D:
-			ch_ctl = PCH_DPD_AUX_CH_CTL;
-			break;
-		default:
-			BUG();
-		}
-	}
-
 #define C (((status = I915_READ_NOTRACE(ch_ctl)) & DP_AUX_CH_CTL_SEND_BUSY) == 0)
 	if (has_aux_irq)
 		done = wait_event_timeout(dev_priv->gmbus_wait_queue, C,
@@ -370,11 +319,10 @@ intel_dp_aux_ch(struct intel_dp *intel_dp,
 		uint8_t *send, int send_bytes,
 		uint8_t *recv, int recv_size)
 {
-	uint32_t output_reg = intel_dp->output_reg;
 	struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp);
 	struct drm_device *dev = intel_dig_port->base.base.dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	uint32_t ch_ctl = output_reg + 0x10;
+	uint32_t ch_ctl = intel_dp->aux_ch_ctl_reg;
 	uint32_t ch_data = ch_ctl + 4;
 	int i, ret, recv_bytes;
 	uint32_t status;
@@ -388,29 +336,6 @@ intel_dp_aux_ch(struct intel_dp *intel_dp,
 	 */
 	pm_qos_update_request(&dev_priv->pm_qos, 0);
 
-	if (IS_HASWELL(dev)) {
-		switch (intel_dig_port->port) {
-		case PORT_A:
-			ch_ctl = DPA_AUX_CH_CTL;
-			ch_data = DPA_AUX_CH_DATA1;
-			break;
-		case PORT_B:
-			ch_ctl = PCH_DPB_AUX_CH_CTL;
-			ch_data = PCH_DPB_AUX_CH_DATA1;
-			break;
-		case PORT_C:
-			ch_ctl = PCH_DPC_AUX_CH_CTL;
-			ch_data = PCH_DPC_AUX_CH_DATA1;
-			break;
-		case PORT_D:
-			ch_ctl = PCH_DPD_AUX_CH_CTL;
-			ch_data = PCH_DPD_AUX_CH_DATA1;
-			break;
-		default:
-			BUG();
-		}
-	}
-
 	intel_dp_check_edp(intel_dp);
 	/* The clock divider is based off the hrawclk,
 	 * and would like to run at 2MHz. So, take the
@@ -428,10 +353,14 @@ intel_dp_aux_ch(struct intel_dp *intel_dp,
 			aux_clock_divider = 200; /* SNB & IVB eDP input clock at 400Mhz */
 		else
 			aux_clock_divider = 225; /* eDP input clock at 450Mhz */
-	} else if (HAS_PCH_SPLIT(dev))
+	} else if (dev_priv->pch_id == INTEL_PCH_LPT_DEVICE_ID_TYPE) {
+		/* Workaround for non-ULT HSW */
+		aux_clock_divider = 74;
+	} else if (HAS_PCH_SPLIT(dev)) {
 		aux_clock_divider = DIV_ROUND_UP(intel_pch_rawclk(dev), 2);
-	else
+	} else {
 		aux_clock_divider = intel_hrawclk(dev) / 2;
+	}
 
 	if (IS_GEN6(dev))
 		precharge = 3;
@@ -732,18 +661,26 @@ intel_dp_i2c_init(struct intel_dp *intel_dp,
 }
 
 bool
-intel_dp_mode_fixup(struct drm_encoder *encoder,
-		    const struct drm_display_mode *mode,
-		    struct drm_display_mode *adjusted_mode)
+intel_dp_compute_config(struct intel_encoder *encoder,
+			struct intel_crtc_config *pipe_config)
 {
-	struct drm_device *dev = encoder->dev;
-	struct intel_dp *intel_dp = enc_to_intel_dp(encoder);
+	struct drm_device *dev = encoder->base.dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_display_mode *adjusted_mode = &pipe_config->adjusted_mode;
+	struct drm_display_mode *mode = &pipe_config->requested_mode;
+	struct intel_dp *intel_dp = enc_to_intel_dp(&encoder->base);
 	struct intel_connector *intel_connector = intel_dp->attached_connector;
 	int lane_count, clock;
 	int max_lane_count = drm_dp_max_lane_count(intel_dp->dpcd);
 	int max_clock = intel_dp_max_link_bw(intel_dp) == DP_LINK_BW_2_7 ? 1 : 0;
 	int bpp, mode_rate;
 	static int bws[2] = { DP_LINK_BW_1_62, DP_LINK_BW_2_7 };
+	int target_clock, link_avail, link_clock;
+
+	if (HAS_PCH_SPLIT(dev) && !HAS_DDI(dev) && !is_cpu_edp(intel_dp))
+		pipe_config->has_pch_encoder = true;
+
+	pipe_config->has_dp_encoder = true;
 
 	if (is_edp(intel_dp) && intel_connector->panel.fixed_mode) {
 		intel_fixed_panel_mode(intel_connector->panel.fixed_mode,
@@ -752,6 +689,8 @@ intel_dp_mode_fixup(struct drm_encoder *encoder,
 					intel_connector->panel.fitting_mode,
 					mode, adjusted_mode);
 	}
+	/* We need to take the panel's fixed mode into account. */
+	target_clock = adjusted_mode->clock;
 
 	if (adjusted_mode->flags & DRM_MODE_FLAG_DBLCLK)
 		return false;
@@ -760,11 +699,28 @@ intel_dp_mode_fixup(struct drm_encoder *encoder,
 		      "max bw %02x pixel clock %iKHz\n",
 		      max_lane_count, bws[max_clock], adjusted_mode->clock);
 
-	if (!intel_dp_adjust_dithering(intel_dp, adjusted_mode, true))
-		return false;
+	/* Walk through all bpp values. Luckily they're all nicely spaced with 2
+	 * bpc in between. */
+	bpp = min_t(int, 8*3, pipe_config->pipe_bpp);
+	for (; bpp >= 6*3; bpp -= 2*3) {
+		mode_rate = intel_dp_link_required(target_clock, bpp);
+
+		for (clock = 0; clock <= max_clock; clock++) {
+			for (lane_count = 1; lane_count <= max_lane_count; lane_count <<= 1) {
+				link_clock = drm_dp_bw_code_to_link_rate(bws[clock]);
+				link_avail = intel_dp_max_data_rate(link_clock,
+								    lane_count);
+
+				if (mode_rate <= link_avail) {
+					goto found;
+				}
+			}
+		}
+	}
 
-	bpp = adjusted_mode->private_flags & INTEL_MODE_DP_FORCE_6BPC ? 18 : 24;
+	return false;
 
+found:
 	if (intel_dp->color_range_auto) {
 		/*
 		 * See:
@@ -778,104 +734,38 @@ intel_dp_mode_fixup(struct drm_encoder *encoder,
 	}
 
 	if (intel_dp->color_range)
-		adjusted_mode->private_flags |= INTEL_MODE_LIMITED_COLOR_RANGE;
-
-	mode_rate = intel_dp_link_required(adjusted_mode->clock, bpp);
-
-	for (clock = 0; clock <= max_clock; clock++) {
-		for (lane_count = 1; lane_count <= max_lane_count; lane_count <<= 1) {
-			int link_bw_clock =
-				drm_dp_bw_code_to_link_rate(bws[clock]);
-			int link_avail = intel_dp_max_data_rate(link_bw_clock,
-								lane_count);
-
-			if (mode_rate <= link_avail) {
-				intel_dp->link_bw = bws[clock];
-				intel_dp->lane_count = lane_count;
-				adjusted_mode->clock = link_bw_clock;
-				DRM_DEBUG_KMS("DP link bw %02x lane "
-						"count %d clock %d bpp %d\n",
-				       intel_dp->link_bw, intel_dp->lane_count,
-				       adjusted_mode->clock, bpp);
-				DRM_DEBUG_KMS("DP link bw required %i available %i\n",
-					      mode_rate, link_avail);
-				return true;
-			}
-		}
-	}
-
-	return false;
-}
+		pipe_config->limited_color_range = true;
 
-void
-intel_dp_set_m_n(struct drm_crtc *crtc, struct drm_display_mode *mode,
-		 struct drm_display_mode *adjusted_mode)
-{
-	struct drm_device *dev = crtc->dev;
-	struct intel_encoder *intel_encoder;
-	struct intel_dp *intel_dp;
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
-	int lane_count = 4;
-	struct intel_link_m_n m_n;
-	int pipe = intel_crtc->pipe;
-	enum transcoder cpu_transcoder = intel_crtc->cpu_transcoder;
-	int target_clock;
+	intel_dp->link_bw = bws[clock];
+	intel_dp->lane_count = lane_count;
+	adjusted_mode->clock = drm_dp_bw_code_to_link_rate(intel_dp->link_bw);
+	pipe_config->pixel_target_clock = target_clock;
 
-	/*
-	 * Find the lane count in the intel_encoder private
-	 */
-	for_each_encoder_on_crtc(dev, crtc, intel_encoder) {
-		intel_dp = enc_to_intel_dp(&intel_encoder->base);
+	DRM_DEBUG_KMS("DP link bw %02x lane count %d clock %d bpp %d\n",
+		      intel_dp->link_bw, intel_dp->lane_count,
+		      adjusted_mode->clock, bpp);
+	DRM_DEBUG_KMS("DP link bw required %i available %i\n",
+		      mode_rate, link_avail);
 
-		if (intel_encoder->type == INTEL_OUTPUT_DISPLAYPORT ||
-		    intel_encoder->type == INTEL_OUTPUT_EDP)
-		{
-			lane_count = intel_dp->lane_count;
-			break;
-		}
-	}
-
-	target_clock = mode->clock;
-	for_each_encoder_on_crtc(dev, crtc, intel_encoder) {
-		if (intel_encoder->type == INTEL_OUTPUT_EDP) {
-			target_clock = intel_edp_target_clock(intel_encoder,
-							      mode);
-			break;
-		}
-	}
+	intel_link_compute_m_n(bpp, lane_count,
+			       target_clock, adjusted_mode->clock,
+			       &pipe_config->dp_m_n);
 
 	/*
-	 * Compute the GMCH and Link ratios. The '3' here is
-	 * the number of bytes_per_pixel post-LUT, which we always
-	 * set up for 8-bits of R/G/B, or 3 bytes total.
+	 * XXX: We have a strange regression where using the vbt edp bpp value
+	 * for the link bw computation results in black screens, the panel only
+	 * works when we do the computation at the usual 24bpp (but still
+	 * requires us to use 18bpp). Until that's fully debugged, stay
+	 * bug-for-bug compatible with the old code.
 	 */
-	intel_link_compute_m_n(intel_crtc->bpp, lane_count,
-			       target_clock, adjusted_mode->clock, &m_n);
-
-	if (IS_HASWELL(dev)) {
-		I915_WRITE(PIPE_DATA_M1(cpu_transcoder),
-			   TU_SIZE(m_n.tu) | m_n.gmch_m);
-		I915_WRITE(PIPE_DATA_N1(cpu_transcoder), m_n.gmch_n);
-		I915_WRITE(PIPE_LINK_M1(cpu_transcoder), m_n.link_m);
-		I915_WRITE(PIPE_LINK_N1(cpu_transcoder), m_n.link_n);
-	} else if (HAS_PCH_SPLIT(dev)) {
-		I915_WRITE(TRANSDATA_M1(pipe), TU_SIZE(m_n.tu) | m_n.gmch_m);
-		I915_WRITE(TRANSDATA_N1(pipe), m_n.gmch_n);
-		I915_WRITE(TRANSDPLINK_M1(pipe), m_n.link_m);
-		I915_WRITE(TRANSDPLINK_N1(pipe), m_n.link_n);
-	} else if (IS_VALLEYVIEW(dev)) {
-		I915_WRITE(PIPE_DATA_M1(pipe), TU_SIZE(m_n.tu) | m_n.gmch_m);
-		I915_WRITE(PIPE_DATA_N1(pipe), m_n.gmch_n);
-		I915_WRITE(PIPE_LINK_M1(pipe), m_n.link_m);
-		I915_WRITE(PIPE_LINK_N1(pipe), m_n.link_n);
-	} else {
-		I915_WRITE(PIPE_GMCH_DATA_M(pipe),
-			   TU_SIZE(m_n.tu) | m_n.gmch_m);
-		I915_WRITE(PIPE_GMCH_DATA_N(pipe), m_n.gmch_n);
-		I915_WRITE(PIPE_DP_LINK_M(pipe), m_n.link_m);
-		I915_WRITE(PIPE_DP_LINK_N(pipe), m_n.link_n);
+	if (is_edp(intel_dp) && dev_priv->edp.bpp) {
+		DRM_DEBUG_KMS("clamping display bpc (was %d) to eDP (%d)\n",
+			      bpp, dev_priv->edp.bpp);
+		bpp = min_t(int, bpp, dev_priv->edp.bpp);
 	}
+	pipe_config->pipe_bpp = bpp;
+
+	return true;
 }
 
 void intel_dp_init_link_config(struct intel_dp *intel_dp)
@@ -994,7 +884,7 @@ intel_dp_mode_set(struct drm_encoder *encoder, struct drm_display_mode *mode,
 		else
 			intel_dp->DP |= DP_PLL_FREQ_270MHZ;
 	} else if (!HAS_PCH_CPT(dev) || is_cpu_edp(intel_dp)) {
-		if (!HAS_PCH_SPLIT(dev))
+		if (!HAS_PCH_SPLIT(dev) && !IS_VALLEYVIEW(dev))
 			intel_dp->DP |= intel_dp->color_range;
 
 		if (adjusted_mode->flags & DRM_MODE_FLAG_PHSYNC)
@@ -1009,7 +899,7 @@ intel_dp_mode_set(struct drm_encoder *encoder, struct drm_display_mode *mode,
 		if (intel_crtc->pipe == 1)
 			intel_dp->DP |= DP_PIPEB_SELECT;
 
-		if (is_cpu_edp(intel_dp)) {
+		if (is_cpu_edp(intel_dp) && !IS_VALLEYVIEW(dev)) {
 			/* don't miss out required setting for eDP */
 			if (adjusted_mode->clock < 200000)
 				intel_dp->DP |= DP_PLL_FREQ_160MHZ;
@@ -1020,7 +910,7 @@ intel_dp_mode_set(struct drm_encoder *encoder, struct drm_display_mode *mode,
 		intel_dp->DP |= DP_LINK_TRAIN_OFF_CPT;
 	}
 
-	if (is_cpu_edp(intel_dp))
+	if (is_cpu_edp(intel_dp) && !IS_VALLEYVIEW(dev))
 		ironlake_set_pll_edp(crtc, adjusted_mode->clock);
 }
 
@@ -1039,16 +929,20 @@ static void ironlake_wait_panel_status(struct intel_dp *intel_dp,
 {
 	struct drm_device *dev = intel_dp_to_dev(intel_dp);
 	struct drm_i915_private *dev_priv = dev->dev_private;
+	u32 pp_stat_reg, pp_ctrl_reg;
+
+	pp_stat_reg = IS_VALLEYVIEW(dev) ? PIPEA_PP_STATUS : PCH_PP_STATUS;
+	pp_ctrl_reg = IS_VALLEYVIEW(dev) ? PIPEA_PP_CONTROL : PCH_PP_CONTROL;
 
 	DRM_DEBUG_KMS("mask %08x value %08x status %08x control %08x\n",
-		      mask, value,
-		      I915_READ(PCH_PP_STATUS),
-		      I915_READ(PCH_PP_CONTROL));
+			mask, value,
+			I915_READ(pp_stat_reg),
+			I915_READ(pp_ctrl_reg));
 
-	if (_wait_for((I915_READ(PCH_PP_STATUS) & mask) == value, 5000, 10)) {
+	if (_wait_for((I915_READ(pp_stat_reg) & mask) == value, 5000, 10)) {
 		DRM_ERROR("Panel status timeout: status %08x control %08x\n",
-			  I915_READ(PCH_PP_STATUS),
-			  I915_READ(PCH_PP_CONTROL));
+				I915_READ(pp_stat_reg),
+				I915_READ(pp_ctrl_reg));
 	}
 }
 
@@ -1075,9 +969,15 @@ static void ironlake_wait_panel_power_cycle(struct intel_dp *intel_dp)
  * is locked
  */
 
-static  u32 ironlake_get_pp_control(struct drm_i915_private *dev_priv)
+static  u32 ironlake_get_pp_control(struct intel_dp *intel_dp)
 {
-	u32	control = I915_READ(PCH_PP_CONTROL);
+	struct drm_device *dev = intel_dp_to_dev(intel_dp);
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	u32 control;
+	u32 pp_ctrl_reg;
+
+	pp_ctrl_reg = IS_VALLEYVIEW(dev) ? PIPEA_PP_CONTROL : PCH_PP_CONTROL;
+	control = I915_READ(pp_ctrl_reg);
 
 	control &= ~PANEL_UNLOCK_MASK;
 	control |= PANEL_UNLOCK_REGS;
@@ -1089,6 +989,7 @@ void ironlake_edp_panel_vdd_on(struct intel_dp *intel_dp)
 	struct drm_device *dev = intel_dp_to_dev(intel_dp);
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	u32 pp;
+	u32 pp_stat_reg, pp_ctrl_reg;
 
 	if (!is_edp(intel_dp))
 		return;
@@ -1107,13 +1008,16 @@ void ironlake_edp_panel_vdd_on(struct intel_dp *intel_dp)
 	if (!ironlake_edp_have_panel_power(intel_dp))
 		ironlake_wait_panel_power_cycle(intel_dp);
 
-	pp = ironlake_get_pp_control(dev_priv);
+	pp = ironlake_get_pp_control(intel_dp);
 	pp |= EDP_FORCE_VDD;
-	I915_WRITE(PCH_PP_CONTROL, pp);
-	POSTING_READ(PCH_PP_CONTROL);
-	DRM_DEBUG_KMS("PCH_PP_STATUS: 0x%08x PCH_PP_CONTROL: 0x%08x\n",
-		      I915_READ(PCH_PP_STATUS), I915_READ(PCH_PP_CONTROL));
 
+	pp_stat_reg = IS_VALLEYVIEW(dev) ? PIPEA_PP_STATUS : PCH_PP_STATUS;
+	pp_ctrl_reg = IS_VALLEYVIEW(dev) ? PIPEA_PP_CONTROL : PCH_PP_CONTROL;
+
+	I915_WRITE(pp_ctrl_reg, pp);
+	POSTING_READ(pp_ctrl_reg);
+	DRM_DEBUG_KMS("PP_STATUS: 0x%08x PP_CONTROL: 0x%08x\n",
+			I915_READ(pp_stat_reg), I915_READ(pp_ctrl_reg));
 	/*
 	 * If the panel wasn't on, delay before accessing aux channel
 	 */
@@ -1128,19 +1032,23 @@ static void ironlake_panel_vdd_off_sync(struct intel_dp *intel_dp)
 	struct drm_device *dev = intel_dp_to_dev(intel_dp);
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	u32 pp;
+	u32 pp_stat_reg, pp_ctrl_reg;
 
 	WARN_ON(!mutex_is_locked(&dev->mode_config.mutex));
 
 	if (!intel_dp->want_panel_vdd && ironlake_edp_have_panel_vdd(intel_dp)) {
-		pp = ironlake_get_pp_control(dev_priv);
+		pp = ironlake_get_pp_control(intel_dp);
 		pp &= ~EDP_FORCE_VDD;
-		I915_WRITE(PCH_PP_CONTROL, pp);
-		POSTING_READ(PCH_PP_CONTROL);
 
-		/* Make sure sequencer is idle before allowing subsequent activity */
-		DRM_DEBUG_KMS("PCH_PP_STATUS: 0x%08x PCH_PP_CONTROL: 0x%08x\n",
-			      I915_READ(PCH_PP_STATUS), I915_READ(PCH_PP_CONTROL));
+		pp_stat_reg = IS_VALLEYVIEW(dev) ? PIPEA_PP_STATUS : PCH_PP_STATUS;
+		pp_ctrl_reg = IS_VALLEYVIEW(dev) ? PIPEA_PP_CONTROL : PCH_PP_CONTROL;
+
+		I915_WRITE(pp_ctrl_reg, pp);
+		POSTING_READ(pp_ctrl_reg);
 
+		/* Make sure sequencer is idle before allowing subsequent activity */
+		DRM_DEBUG_KMS("PP_STATUS: 0x%08x PP_CONTROL: 0x%08x\n",
+		I915_READ(pp_stat_reg), I915_READ(pp_ctrl_reg));
 		msleep(intel_dp->panel_power_down_delay);
 	}
 }
@@ -1184,6 +1092,7 @@ void ironlake_edp_panel_on(struct intel_dp *intel_dp)
 	struct drm_device *dev = intel_dp_to_dev(intel_dp);
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	u32 pp;
+	u32 pp_ctrl_reg;
 
 	if (!is_edp(intel_dp))
 		return;
@@ -1197,7 +1106,7 @@ void ironlake_edp_panel_on(struct intel_dp *intel_dp)
 
 	ironlake_wait_panel_power_cycle(intel_dp);
 
-	pp = ironlake_get_pp_control(dev_priv);
+	pp = ironlake_get_pp_control(intel_dp);
 	if (IS_GEN5(dev)) {
 		/* ILK workaround: disable reset around power sequence */
 		pp &= ~PANEL_POWER_RESET;
@@ -1209,8 +1118,10 @@ void ironlake_edp_panel_on(struct intel_dp *intel_dp)
 	if (!IS_GEN5(dev))
 		pp |= PANEL_POWER_RESET;
 
-	I915_WRITE(PCH_PP_CONTROL, pp);
-	POSTING_READ(PCH_PP_CONTROL);
+	pp_ctrl_reg = IS_VALLEYVIEW(dev) ? PIPEA_PP_CONTROL : PCH_PP_CONTROL;
+
+	I915_WRITE(pp_ctrl_reg, pp);
+	POSTING_READ(pp_ctrl_reg);
 
 	ironlake_wait_panel_on(intel_dp);
 
@@ -1226,6 +1137,7 @@ void ironlake_edp_panel_off(struct intel_dp *intel_dp)
 	struct drm_device *dev = intel_dp_to_dev(intel_dp);
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	u32 pp;
+	u32 pp_ctrl_reg;
 
 	if (!is_edp(intel_dp))
 		return;
@@ -1234,12 +1146,15 @@ void ironlake_edp_panel_off(struct intel_dp *intel_dp)
 
 	WARN(!intel_dp->want_panel_vdd, "Need VDD to turn off panel\n");
 
-	pp = ironlake_get_pp_control(dev_priv);
+	pp = ironlake_get_pp_control(intel_dp);
 	/* We need to switch off panel power _and_ force vdd, for otherwise some
 	 * panels get very unhappy and cease to work. */
 	pp &= ~(POWER_TARGET_ON | EDP_FORCE_VDD | PANEL_POWER_RESET | EDP_BLC_ENABLE);
-	I915_WRITE(PCH_PP_CONTROL, pp);
-	POSTING_READ(PCH_PP_CONTROL);
+
+	pp_ctrl_reg = IS_VALLEYVIEW(dev) ? PIPEA_PP_CONTROL : PCH_PP_CONTROL;
+
+	I915_WRITE(pp_ctrl_reg, pp);
+	POSTING_READ(pp_ctrl_reg);
 
 	intel_dp->want_panel_vdd = false;
 
@@ -1253,6 +1168,7 @@ void ironlake_edp_backlight_on(struct intel_dp *intel_dp)
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	int pipe = to_intel_crtc(intel_dig_port->base.base.crtc)->pipe;
 	u32 pp;
+	u32 pp_ctrl_reg;
 
 	if (!is_edp(intel_dp))
 		return;
@@ -1265,10 +1181,13 @@ void ironlake_edp_backlight_on(struct intel_dp *intel_dp)
 	 * allowing it to appear.
 	 */
 	msleep(intel_dp->backlight_on_delay);
-	pp = ironlake_get_pp_control(dev_priv);
+	pp = ironlake_get_pp_control(intel_dp);
 	pp |= EDP_BLC_ENABLE;
-	I915_WRITE(PCH_PP_CONTROL, pp);
-	POSTING_READ(PCH_PP_CONTROL);
+
+	pp_ctrl_reg = IS_VALLEYVIEW(dev) ? PIPEA_PP_CONTROL : PCH_PP_CONTROL;
+
+	I915_WRITE(pp_ctrl_reg, pp);
+	POSTING_READ(pp_ctrl_reg);
 
 	intel_panel_enable_backlight(dev, pipe);
 }
@@ -1278,6 +1197,7 @@ void ironlake_edp_backlight_off(struct intel_dp *intel_dp)
 	struct drm_device *dev = intel_dp_to_dev(intel_dp);
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	u32 pp;
+	u32 pp_ctrl_reg;
 
 	if (!is_edp(intel_dp))
 		return;
@@ -1285,10 +1205,13 @@ void ironlake_edp_backlight_off(struct intel_dp *intel_dp)
 	intel_panel_disable_backlight(dev);
 
 	DRM_DEBUG_KMS("\n");
-	pp = ironlake_get_pp_control(dev_priv);
+	pp = ironlake_get_pp_control(intel_dp);
 	pp &= ~EDP_BLC_ENABLE;
-	I915_WRITE(PCH_PP_CONTROL, pp);
-	POSTING_READ(PCH_PP_CONTROL);
+
+	pp_ctrl_reg = IS_VALLEYVIEW(dev) ? PIPEA_PP_CONTROL : PCH_PP_CONTROL;
+
+	I915_WRITE(pp_ctrl_reg, pp);
+	POSTING_READ(pp_ctrl_reg);
 	msleep(intel_dp->backlight_off_delay);
 }
 
@@ -1384,7 +1307,7 @@ static bool intel_dp_get_hw_state(struct intel_encoder *encoder,
 	if (!(tmp & DP_PORT_EN))
 		return false;
 
-	if (is_cpu_edp(intel_dp) && IS_GEN7(dev)) {
+	if (is_cpu_edp(intel_dp) && IS_GEN7(dev) && !IS_VALLEYVIEW(dev)) {
 		*pipe = PORT_TO_PIPE_CPT(tmp);
 	} else if (!HAS_PCH_CPT(dev) || is_cpu_edp(intel_dp)) {
 		*pipe = PORT_TO_PIPE(tmp);
@@ -1441,10 +1364,12 @@ static void intel_disable_dp(struct intel_encoder *encoder)
 static void intel_post_disable_dp(struct intel_encoder *encoder)
 {
 	struct intel_dp *intel_dp = enc_to_intel_dp(&encoder->base);
+	struct drm_device *dev = encoder->base.dev;
 
 	if (is_cpu_edp(intel_dp)) {
 		intel_dp_link_down(intel_dp);
-		ironlake_edp_pll_off(intel_dp);
+		if (!IS_VALLEYVIEW(dev))
+			ironlake_edp_pll_off(intel_dp);
 	}
 }
 
@@ -1470,8 +1395,9 @@ static void intel_enable_dp(struct intel_encoder *encoder)
 static void intel_pre_enable_dp(struct intel_encoder *encoder)
 {
 	struct intel_dp *intel_dp = enc_to_intel_dp(&encoder->base);
+	struct drm_device *dev = encoder->base.dev;
 
-	if (is_cpu_edp(intel_dp))
+	if (is_cpu_edp(intel_dp) && !IS_VALLEYVIEW(dev))
 		ironlake_edp_pll_on(intel_dp);
 }
 
@@ -1548,7 +1474,7 @@ intel_dp_pre_emphasis_max(struct intel_dp *intel_dp, uint8_t voltage_swing)
 {
 	struct drm_device *dev = intel_dp_to_dev(intel_dp);
 
-	if (IS_HASWELL(dev)) {
+	if (HAS_DDI(dev)) {
 		switch (voltage_swing & DP_TRAIN_VOLTAGE_SWING_MASK) {
 		case DP_TRAIN_VOLTAGE_SWING_400:
 			return DP_TRAIN_PRE_EMPHASIS_9_5;
@@ -1756,7 +1682,7 @@ intel_dp_set_signal_levels(struct intel_dp *intel_dp, uint32_t *DP)
 	uint32_t signal_levels, mask;
 	uint8_t train_set = intel_dp->train_set[0];
 
-	if (IS_HASWELL(dev)) {
+	if (HAS_DDI(dev)) {
 		signal_levels = intel_hsw_signal_levels(train_set);
 		mask = DDI_BUF_EMP_MASK;
 	} else if (IS_GEN7(dev) && is_cpu_edp(intel_dp) && !IS_VALLEYVIEW(dev)) {
@@ -1787,7 +1713,7 @@ intel_dp_set_link_train(struct intel_dp *intel_dp,
 	int ret;
 	uint32_t temp;
 
-	if (IS_HASWELL(dev)) {
+	if (HAS_DDI(dev)) {
 		temp = I915_READ(DP_TP_CTL(port));
 
 		if (dp_train_pat & DP_LINK_SCRAMBLING_DISABLE)
@@ -2311,6 +2237,16 @@ g4x_dp_detect(struct intel_dp *intel_dp)
 	struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp);
 	uint32_t bit;
 
+	/* Can't disconnect eDP, but you can close the lid... */
+	if (is_edp(intel_dp)) {
+		enum drm_connector_status status;
+
+		status = intel_panel_detect(dev);
+		if (status == connector_status_unknown)
+			status = connector_status_connected;
+		return status;
+	}
+
 	switch (intel_dig_port->port) {
 	case PORT_B:
 		bit = PORTB_HOTPLUG_LIVE_STATUS;
@@ -2492,6 +2428,9 @@ intel_dp_set_property(struct drm_connector *connector,
 	}
 
 	if (property == dev_priv->broadcast_rgb_property) {
+		bool old_auto = intel_dp->color_range_auto;
+		uint32_t old_range = intel_dp->color_range;
+
 		switch (val) {
 		case INTEL_BROADCAST_RGB_AUTO:
 			intel_dp->color_range_auto = true;
@@ -2507,6 +2446,11 @@ intel_dp_set_property(struct drm_connector *connector,
 		default:
 			return -EINVAL;
 		}
+
+		if (old_auto == intel_dp->color_range_auto &&
+		    old_range == intel_dp->color_range)
+			return 0;
+
 		goto done;
 	}
 
@@ -2538,17 +2482,14 @@ done:
 static void
 intel_dp_destroy(struct drm_connector *connector)
 {
-	struct drm_device *dev = connector->dev;
 	struct intel_dp *intel_dp = intel_attached_dp(connector);
 	struct intel_connector *intel_connector = to_intel_connector(connector);
 
 	if (!IS_ERR_OR_NULL(intel_connector->edid))
 		kfree(intel_connector->edid);
 
-	if (is_edp(intel_dp)) {
-		intel_panel_destroy_backlight(dev);
+	if (is_edp(intel_dp))
 		intel_panel_fini(&intel_connector->panel);
-	}
 
 	drm_sysfs_connector_remove(connector);
 	drm_connector_cleanup(connector);
@@ -2573,7 +2514,6 @@ void intel_dp_encoder_destroy(struct drm_encoder *encoder)
 }
 
 static const struct drm_encoder_helper_funcs intel_dp_helper_funcs = {
-	.mode_fixup = intel_dp_mode_fixup,
 	.mode_set = intel_dp_mode_set,
 };
 
@@ -2669,15 +2609,28 @@ intel_dp_init_panel_power_sequencer(struct drm_device *dev,
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct edp_power_seq cur, vbt, spec, final;
 	u32 pp_on, pp_off, pp_div, pp;
+	int pp_control_reg, pp_on_reg, pp_off_reg, pp_div_reg;
+
+	if (HAS_PCH_SPLIT(dev)) {
+		pp_control_reg = PCH_PP_CONTROL;
+		pp_on_reg = PCH_PP_ON_DELAYS;
+		pp_off_reg = PCH_PP_OFF_DELAYS;
+		pp_div_reg = PCH_PP_DIVISOR;
+	} else {
+		pp_control_reg = PIPEA_PP_CONTROL;
+		pp_on_reg = PIPEA_PP_ON_DELAYS;
+		pp_off_reg = PIPEA_PP_OFF_DELAYS;
+		pp_div_reg = PIPEA_PP_DIVISOR;
+	}
 
 	/* Workaround: Need to write PP_CONTROL with the unlock key as
 	 * the very first thing. */
-	pp = ironlake_get_pp_control(dev_priv);
-	I915_WRITE(PCH_PP_CONTROL, pp);
+	pp = ironlake_get_pp_control(intel_dp);
+	I915_WRITE(pp_control_reg, pp);
 
-	pp_on = I915_READ(PCH_PP_ON_DELAYS);
-	pp_off = I915_READ(PCH_PP_OFF_DELAYS);
-	pp_div = I915_READ(PCH_PP_DIVISOR);
+	pp_on = I915_READ(pp_on_reg);
+	pp_off = I915_READ(pp_off_reg);
+	pp_div = I915_READ(pp_div_reg);
 
 	/* Pull timing values out of registers */
 	cur.t1_t3 = (pp_on & PANEL_POWER_UP_DELAY_MASK) >>
@@ -2752,7 +2705,22 @@ intel_dp_init_panel_power_sequencer_registers(struct drm_device *dev,
 					      struct edp_power_seq *seq)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	u32 pp_on, pp_off, pp_div;
+	u32 pp_on, pp_off, pp_div, port_sel = 0;
+	int div = HAS_PCH_SPLIT(dev) ? intel_pch_rawclk(dev) : intel_hrawclk(dev);
+	int pp_on_reg, pp_off_reg, pp_div_reg;
+
+	if (HAS_PCH_SPLIT(dev)) {
+		pp_on_reg = PCH_PP_ON_DELAYS;
+		pp_off_reg = PCH_PP_OFF_DELAYS;
+		pp_div_reg = PCH_PP_DIVISOR;
+	} else {
+		pp_on_reg = PIPEA_PP_ON_DELAYS;
+		pp_off_reg = PIPEA_PP_OFF_DELAYS;
+		pp_div_reg = PIPEA_PP_DIVISOR;
+	}
+
+	if (IS_VALLEYVIEW(dev))
+		port_sel = I915_READ(pp_on_reg) & 0xc0000000;
 
 	/* And finally store the new values in the power sequencer. */
 	pp_on = (seq->t1_t3 << PANEL_POWER_UP_DELAY_SHIFT) |
@@ -2761,8 +2729,7 @@ intel_dp_init_panel_power_sequencer_registers(struct drm_device *dev,
 		 (seq->t10 << PANEL_POWER_DOWN_DELAY_SHIFT);
 	/* Compute the divisor for the pp clock, simply match the Bspec
 	 * formula. */
-	pp_div = ((100 * intel_pch_rawclk(dev))/2 - 1)
-			<< PP_REFERENCE_DIVIDER_SHIFT;
+	pp_div = ((100 * div)/2 - 1) << PP_REFERENCE_DIVIDER_SHIFT;
 	pp_div |= (DIV_ROUND_UP(seq->t11_t12, 1000)
 			<< PANEL_POWER_CYCLE_DELAY_SHIFT);
 
@@ -2770,19 +2737,21 @@ intel_dp_init_panel_power_sequencer_registers(struct drm_device *dev,
 	 * power sequencer any more. */
 	if (HAS_PCH_IBX(dev) || HAS_PCH_CPT(dev)) {
 		if (is_cpu_edp(intel_dp))
-			pp_on |= PANEL_POWER_PORT_DP_A;
+			port_sel = PANEL_POWER_PORT_DP_A;
 		else
-			pp_on |= PANEL_POWER_PORT_DP_D;
+			port_sel = PANEL_POWER_PORT_DP_D;
 	}
 
-	I915_WRITE(PCH_PP_ON_DELAYS, pp_on);
-	I915_WRITE(PCH_PP_OFF_DELAYS, pp_off);
-	I915_WRITE(PCH_PP_DIVISOR, pp_div);
+	pp_on |= port_sel;
+
+	I915_WRITE(pp_on_reg, pp_on);
+	I915_WRITE(pp_off_reg, pp_off);
+	I915_WRITE(pp_div_reg, pp_div);
 
 	DRM_DEBUG_KMS("panel power sequencer register settings: PP_ON %#x, PP_OFF %#x, PP_DIV %#x\n",
-		      I915_READ(PCH_PP_ON_DELAYS),
-		      I915_READ(PCH_PP_OFF_DELAYS),
-		      I915_READ(PCH_PP_DIVISOR));
+		      I915_READ(pp_on_reg),
+		      I915_READ(pp_off_reg),
+		      I915_READ(pp_div_reg));
 }
 
 void
@@ -2829,7 +2798,6 @@ intel_dp_init_connector(struct intel_digital_port *intel_dig_port,
 	drm_connector_init(dev, connector, &intel_dp_connector_funcs, type);
 	drm_connector_helper_add(connector, &intel_dp_connector_helper_funcs);
 
-	connector->polled = DRM_CONNECTOR_POLL_HPD;
 	connector->interlace_allowed = true;
 	connector->doublescan_allowed = 0;
 
@@ -2844,27 +2812,46 @@ intel_dp_init_connector(struct intel_digital_port *intel_dig_port,
 	else
 		intel_connector->get_hw_state = intel_connector_get_hw_state;
 
+	intel_dp->aux_ch_ctl_reg = intel_dp->output_reg + 0x10;
+	if (HAS_DDI(dev)) {
+		switch (intel_dig_port->port) {
+		case PORT_A:
+			intel_dp->aux_ch_ctl_reg = DPA_AUX_CH_CTL;
+			break;
+		case PORT_B:
+			intel_dp->aux_ch_ctl_reg = PCH_DPB_AUX_CH_CTL;
+			break;
+		case PORT_C:
+			intel_dp->aux_ch_ctl_reg = PCH_DPC_AUX_CH_CTL;
+			break;
+		case PORT_D:
+			intel_dp->aux_ch_ctl_reg = PCH_DPD_AUX_CH_CTL;
+			break;
+		default:
+			BUG();
+		}
+	}
 
 	/* Set up the DDC bus. */
 	switch (port) {
 	case PORT_A:
+		intel_encoder->hpd_pin = HPD_PORT_A;
 		name = "DPDDC-A";
 		break;
 	case PORT_B:
-		dev_priv->hotplug_supported_mask |= PORTB_HOTPLUG_INT_STATUS;
+		intel_encoder->hpd_pin = HPD_PORT_B;
 		name = "DPDDC-B";
 		break;
 	case PORT_C:
-		dev_priv->hotplug_supported_mask |= PORTC_HOTPLUG_INT_STATUS;
+		intel_encoder->hpd_pin = HPD_PORT_C;
 		name = "DPDDC-C";
 		break;
 	case PORT_D:
-		dev_priv->hotplug_supported_mask |= PORTD_HOTPLUG_INT_STATUS;
+		intel_encoder->hpd_pin = HPD_PORT_D;
 		name = "DPDDC-D";
 		break;
 	default:
-		WARN(1, "Invalid port %c\n", port_name(port));
-		break;
+		BUG();
 	}
 
 	if (is_edp(intel_dp))
@@ -2974,6 +2961,7 @@ intel_dp_init(struct drm_device *dev, int output_reg, enum port port)
 			 DRM_MODE_ENCODER_TMDS);
 	drm_encoder_helper_add(&intel_encoder->base, &intel_dp_helper_funcs);
 
+	intel_encoder->compute_config = intel_dp_compute_config;
 	intel_encoder->enable = intel_enable_dp;
 	intel_encoder->pre_enable = intel_pre_enable_dp;
 	intel_encoder->disable = intel_disable_dp;
diff --git a/drivers/gpu/drm/i915/intel_drv.h b/drivers/gpu/drm/i915/intel_drv.h
index 07ebac6fe8ca..b5b6d19e6dd3 100644
--- a/drivers/gpu/drm/i915/intel_drv.h
+++ b/drivers/gpu/drm/i915/intel_drv.h
@@ -33,12 +33,21 @@
 #include <drm/drm_fb_helper.h>
 #include <drm/drm_dp_helper.h>
 
+/**
+ * _wait_for - magic (register) wait macro
+ *
+ * Does the right thing for modeset paths when run under kdgb or similar atomic
+ * contexts. Note that it's important that we check the condition again after
+ * having timed out, since the timeout could be due to preemption or similar and
+ * we've never had a chance to check the condition before the timeout.
+ */
 #define _wait_for(COND, MS, W) ({ \
-	unsigned long timeout__ = jiffies + msecs_to_jiffies(MS);	\
+	unsigned long timeout__ = jiffies + msecs_to_jiffies(MS) + 1;	\
 	int ret__ = 0;							\
 	while (!(COND)) {						\
 		if (time_after(jiffies, timeout__)) {			\
-			ret__ = -ETIMEDOUT;				\
+			if (!(COND))					\
+				ret__ = -ETIMEDOUT;			\
 			break;						\
 		}							\
 		if (W && drm_can_sleep())  {				\
@@ -50,21 +59,10 @@
 	ret__;								\
 })
 
-#define wait_for_atomic_us(COND, US) ({ \
-	unsigned long timeout__ = jiffies + usecs_to_jiffies(US);	\
-	int ret__ = 0;							\
-	while (!(COND)) {						\
-		if (time_after(jiffies, timeout__)) {			\
-			ret__ = -ETIMEDOUT;				\
-			break;						\
-		}							\
-		cpu_relax();						\
-	}								\
-	ret__;								\
-})
-
 #define wait_for(COND, MS) _wait_for(COND, MS, 1)
 #define wait_for_atomic(COND, MS) _wait_for(COND, MS, 0)
+#define wait_for_atomic_us(COND, US) _wait_for((COND), \
+					       DIV_ROUND_UP((US), 1000), 0)
 
 #define KHz(x) (1000*x)
 #define MHz(x) KHz(1000*x)
@@ -101,34 +99,6 @@
 #define INTEL_DVO_CHIP_TMDS 2
 #define INTEL_DVO_CHIP_TVOUT 4
 
-/* drm_display_mode->private_flags */
-#define INTEL_MODE_PIXEL_MULTIPLIER_SHIFT (0x0)
-#define INTEL_MODE_PIXEL_MULTIPLIER_MASK (0xf << INTEL_MODE_PIXEL_MULTIPLIER_SHIFT)
-#define INTEL_MODE_DP_FORCE_6BPC (0x10)
-/* This flag must be set by the encoder's mode_fixup if it changes the crtc
- * timings in the mode to prevent the crtc fixup from overwriting them.
- * Currently only lvds needs that. */
-#define INTEL_MODE_CRTC_TIMINGS_SET (0x20)
-/*
- * Set when limited 16-235 (as opposed to full 0-255) RGB color range is
- * to be used.
- */
-#define INTEL_MODE_LIMITED_COLOR_RANGE (0x40)
-
-static inline void
-intel_mode_set_pixel_multiplier(struct drm_display_mode *mode,
-				int multiplier)
-{
-	mode->clock *= multiplier;
-	mode->private_flags |= multiplier;
-}
-
-static inline int
-intel_mode_get_pixel_multiplier(const struct drm_display_mode *mode)
-{
-	return (mode->private_flags & INTEL_MODE_PIXEL_MULTIPLIER_MASK) >> INTEL_MODE_PIXEL_MULTIPLIER_SHIFT;
-}
-
 struct intel_framebuffer {
 	struct drm_framebuffer base;
 	struct drm_i915_gem_object *obj;
@@ -158,9 +128,12 @@ struct intel_encoder {
 	bool cloneable;
 	bool connectors_active;
 	void (*hot_plug)(struct intel_encoder *);
+	bool (*compute_config)(struct intel_encoder *,
+			       struct intel_crtc_config *);
 	void (*pre_pll_enable)(struct intel_encoder *);
 	void (*pre_enable)(struct intel_encoder *);
 	void (*enable)(struct intel_encoder *);
+	void (*mode_set)(struct intel_encoder *intel_encoder);
 	void (*disable)(struct intel_encoder *);
 	void (*post_disable)(struct intel_encoder *);
 	/* Read out the current hw state of this connector, returning true if
@@ -168,6 +141,7 @@ struct intel_encoder {
 	 * it is connected to in the pipe parameter. */
 	bool (*get_hw_state)(struct intel_encoder *, enum pipe *pipe);
 	int crtc_mask;
+	enum hpd_pin hpd_pin;
 };
 
 struct intel_panel {
@@ -197,13 +171,65 @@ struct intel_connector {
 
 	/* Cached EDID for eDP and LVDS. May hold ERR_PTR for invalid EDID. */
 	struct edid *edid;
+
+	/* since POLL and HPD connectors may use the same HPD line keep the native
+	   state of connector->polled in case hotplug storm detection changes it */
+	u8 polled;
+};
+
+struct intel_crtc_config {
+	struct drm_display_mode requested_mode;
+	struct drm_display_mode adjusted_mode;
+	/* This flag must be set by the encoder's compute_config callback if it
+	 * changes the crtc timings in the mode to prevent the crtc fixup from
+	 * overwriting them.  Currently only lvds needs that. */
+	bool timings_set;
+	/* Whether to set up the PCH/FDI. Note that we never allow sharing
+	 * between pch encoders and cpu encoders. */
+	bool has_pch_encoder;
+
+	/* CPU Transcoder for the pipe. Currently this can only differ from the
+	 * pipe on Haswell (where we have a special eDP transcoder). */
+	enum transcoder cpu_transcoder;
+
+	/*
+	 * Use reduced/limited/broadcast rbg range, compressing from the full
+	 * range fed into the crtcs.
+	 */
+	bool limited_color_range;
+
+	/* DP has a bunch of special case unfortunately, so mark the pipe
+	 * accordingly. */
+	bool has_dp_encoder;
+	bool dither;
+
+	/* Controls for the clock computation, to override various stages. */
+	bool clock_set;
+
+	/* Settings for the intel dpll used on pretty much everything but
+	 * haswell. */
+	struct dpll {
+		unsigned n;
+		unsigned m1, m2;
+		unsigned p1, p2;
+	} dpll;
+
+	int pipe_bpp;
+	struct intel_link_m_n dp_m_n;
+	/**
+	 * This is currently used by DP and HDMI encoders since those can have a
+	 * target pixel clock != the port link clock (which is currently stored
+	 * in adjusted_mode->clock).
+	 */
+	int pixel_target_clock;
+	/* Used by SDVO (and if we ever fix it, HDMI). */
+	unsigned pixel_multiplier;
 };
 
 struct intel_crtc {
 	struct drm_crtc base;
 	enum pipe pipe;
 	enum plane plane;
-	enum transcoder cpu_transcoder;
 	u8 lut_r[256], lut_g[256], lut_b[256];
 	/*
 	 * Whether the crtc and the connected output pipeline is active. Implies
@@ -230,7 +256,8 @@ struct intel_crtc {
 	int16_t cursor_x, cursor_y;
 	int16_t cursor_width, cursor_height;
 	bool cursor_visible;
-	unsigned int bpp;
+
+	struct intel_crtc_config config;
 
 	/* We can share PLLs across outputs if the timings match */
 	struct intel_pch_pll *pch_pll;
@@ -242,11 +269,16 @@ struct intel_crtc {
 
 struct intel_plane {
 	struct drm_plane base;
+	int plane;
 	enum pipe pipe;
 	struct drm_i915_gem_object *obj;
 	bool can_scale;
 	int max_downscale;
 	u32 lut_r[1024], lut_g[1024], lut_b[1024];
+	int crtc_x, crtc_y;
+	unsigned int crtc_w, crtc_h;
+	uint32_t src_x, src_y;
+	uint32_t src_w, src_h;
 	void (*update_plane)(struct drm_plane *plane,
 			     struct drm_framebuffer *fb,
 			     struct drm_i915_gem_object *obj,
@@ -347,7 +379,7 @@ struct dip_infoframe {
 } __attribute__((packed));
 
 struct intel_hdmi {
-	u32 sdvox_reg;
+	u32 hdmi_reg;
 	int ddc_bus;
 	uint32_t color_range;
 	bool color_range_auto;
@@ -366,6 +398,7 @@ struct intel_hdmi {
 
 struct intel_dp {
 	uint32_t output_reg;
+	uint32_t aux_ch_ctl_reg;
 	uint32_t DP;
 	uint8_t  link_configuration[DP_LINK_CONFIGURATION_SIZE];
 	bool has_audio;
@@ -443,13 +476,12 @@ extern void intel_attach_broadcast_rgb_property(struct drm_connector *connector)
 
 extern void intel_crt_init(struct drm_device *dev);
 extern void intel_hdmi_init(struct drm_device *dev,
-			    int sdvox_reg, enum port port);
+			    int hdmi_reg, enum port port);
 extern void intel_hdmi_init_connector(struct intel_digital_port *intel_dig_port,
 				      struct intel_connector *intel_connector);
 extern struct intel_hdmi *enc_to_intel_hdmi(struct drm_encoder *encoder);
-extern bool intel_hdmi_mode_fixup(struct drm_encoder *encoder,
-				  const struct drm_display_mode *mode,
-				  struct drm_display_mode *adjusted_mode);
+extern bool intel_hdmi_compute_config(struct intel_encoder *encoder,
+				      struct intel_crtc_config *pipe_config);
 extern void intel_dip_infoframe_csum(struct dip_infoframe *avi_if);
 extern bool intel_sdvo_init(struct drm_device *dev, uint32_t sdvo_reg,
 			    bool is_sdvob);
@@ -464,18 +496,14 @@ extern void intel_dp_init(struct drm_device *dev, int output_reg,
 			  enum port port);
 extern void intel_dp_init_connector(struct intel_digital_port *intel_dig_port,
 				    struct intel_connector *intel_connector);
-void
-intel_dp_set_m_n(struct drm_crtc *crtc, struct drm_display_mode *mode,
-		 struct drm_display_mode *adjusted_mode);
 extern void intel_dp_init_link_config(struct intel_dp *intel_dp);
 extern void intel_dp_start_link_train(struct intel_dp *intel_dp);
 extern void intel_dp_complete_link_train(struct intel_dp *intel_dp);
 extern void intel_dp_sink_dpms(struct intel_dp *intel_dp, int mode);
 extern void intel_dp_encoder_destroy(struct drm_encoder *encoder);
 extern void intel_dp_check_link_status(struct intel_dp *intel_dp);
-extern bool intel_dp_mode_fixup(struct drm_encoder *encoder,
-				const struct drm_display_mode *mode,
-				struct drm_display_mode *adjusted_mode);
+extern bool intel_dp_compute_config(struct intel_encoder *encoder,
+				    struct intel_crtc_config *pipe_config);
 extern bool intel_dpd_is_edp(struct drm_device *dev);
 extern void ironlake_edp_backlight_on(struct intel_dp *intel_dp);
 extern void ironlake_edp_backlight_off(struct intel_dp *intel_dp);
@@ -483,11 +511,8 @@ extern void ironlake_edp_panel_on(struct intel_dp *intel_dp);
 extern void ironlake_edp_panel_off(struct intel_dp *intel_dp);
 extern void ironlake_edp_panel_vdd_on(struct intel_dp *intel_dp);
 extern void ironlake_edp_panel_vdd_off(struct intel_dp *intel_dp, bool sync);
-extern void intel_edp_link_config(struct intel_encoder *, int *, int *);
-extern int intel_edp_target_clock(struct intel_encoder *,
-				  struct drm_display_mode *mode);
 extern bool intel_encoder_is_pch_edp(struct drm_encoder *encoder);
-extern int intel_plane_init(struct drm_device *dev, enum pipe pipe);
+extern int intel_plane_init(struct drm_device *dev, enum pipe pipe, int plane);
 extern void intel_flush_display_plane(struct drm_i915_private *dev_priv,
 				      enum plane plane);
 
@@ -531,6 +556,7 @@ extern bool intel_encoder_check_is_cloned(struct intel_encoder *encoder);
 extern void intel_connector_dpms(struct drm_connector *, int mode);
 extern bool intel_connector_get_hw_state(struct intel_connector *connector);
 extern void intel_modeset_check_state(struct drm_device *dev);
+extern void intel_plane_restore(struct drm_plane *plane);
 
 
 static inline struct intel_encoder *intel_attached_encoder(struct drm_connector *connector)
@@ -636,6 +662,10 @@ extern void intel_init_clock_gating(struct drm_device *dev);
 extern void intel_write_eld(struct drm_encoder *encoder,
 			    struct drm_display_mode *mode);
 extern void intel_cpt_verify_modeset(struct drm_device *dev, int pipe);
+extern void intel_cpu_transcoder_set_m_n(struct intel_crtc *crtc,
+					 struct intel_link_m_n *m_n);
+extern void intel_pch_transcoder_set_m_n(struct intel_crtc *crtc,
+					 struct intel_link_m_n *m_n);
 extern void intel_prepare_ddi(struct drm_device *dev);
 extern void hsw_fdi_link_train(struct drm_crtc *crtc);
 extern void intel_ddi_init(struct drm_device *dev, enum port port);
@@ -670,6 +700,7 @@ extern void intel_update_fbc(struct drm_device *dev);
 extern void intel_gpu_ips_init(struct drm_i915_private *dev_priv);
 extern void intel_gpu_ips_teardown(void);
 
+extern bool intel_using_power_well(struct drm_device *dev);
 extern void intel_init_power_well(struct drm_device *dev);
 extern void intel_set_power_well(struct drm_device *dev, bool enable);
 extern void intel_enable_gt_powersave(struct drm_device *dev);
@@ -681,7 +712,7 @@ extern bool intel_ddi_get_hw_state(struct intel_encoder *encoder,
 				   enum pipe *pipe);
 extern int intel_ddi_get_cdclk_freq(struct drm_i915_private *dev_priv);
 extern void intel_ddi_pll_init(struct drm_device *dev);
-extern void intel_ddi_enable_pipe_func(struct drm_crtc *crtc);
+extern void intel_ddi_enable_transcoder_func(struct drm_crtc *crtc);
 extern void intel_ddi_disable_transcoder_func(struct drm_i915_private *dev_priv,
 					      enum transcoder cpu_transcoder);
 extern void intel_ddi_enable_pipe_clock(struct intel_crtc *intel_crtc);
@@ -695,4 +726,6 @@ extern bool
 intel_ddi_connector_get_hw_state(struct intel_connector *intel_connector);
 extern void intel_ddi_fdi_disable(struct drm_crtc *crtc);
 
+extern void intel_display_handle_reset(struct drm_device *dev);
+
 #endif /* __INTEL_DRV_H__ */
diff --git a/drivers/gpu/drm/i915/intel_dvo.c b/drivers/gpu/drm/i915/intel_dvo.c
index 00e70dbe82da..cc70b16d5d42 100644
--- a/drivers/gpu/drm/i915/intel_dvo.c
+++ b/drivers/gpu/drm/i915/intel_dvo.c
@@ -448,6 +448,7 @@ void intel_dvo_init(struct drm_device *dev)
 		const struct intel_dvo_device *dvo = &intel_dvo_devices[i];
 		struct i2c_adapter *i2c;
 		int gpio;
+		bool dvoinit;
 
 		/* Allow the I2C driver info to specify the GPIO to be used in
 		 * special cases, but otherwise default to what's defined
@@ -467,7 +468,17 @@ void intel_dvo_init(struct drm_device *dev)
 		i2c = intel_gmbus_get_adapter(dev_priv, gpio);
 
 		intel_dvo->dev = *dvo;
-		if (!dvo->dev_ops->init(&intel_dvo->dev, i2c))
+
+		/* GMBUS NAK handling seems to be unstable, hence let the
+		 * transmitter detection run in bit banging mode for now.
+		 */
+		intel_gmbus_force_bit(i2c, true);
+
+		dvoinit = dvo->dev_ops->init(&intel_dvo->dev, i2c);
+
+		intel_gmbus_force_bit(i2c, false);
+
+		if (!dvoinit)
 			continue;
 
 		intel_encoder->type = INTEL_OUTPUT_DVO;
diff --git a/drivers/gpu/drm/i915/intel_fb.c b/drivers/gpu/drm/i915/intel_fb.c
index 981bdce3634e..0e19e575a1b4 100644
--- a/drivers/gpu/drm/i915/intel_fb.c
+++ b/drivers/gpu/drm/i915/intel_fb.c
@@ -150,7 +150,8 @@ static int intelfb_create(struct drm_fb_helper *helper,
 	}
 	info->screen_size = size;
 
-//	memset(info->screen_base, 0, size);
+	/* This driver doesn't need a VT switch to restore the mode on resume */
+	info->skip_vt_switch = true;
 
 	drm_fb_helper_fill_fix(info, fb->pitches[0], fb->depth);
 	drm_fb_helper_fill_var(info, &ifbdev->helper, sizes->fb_width, sizes->fb_height);
@@ -227,7 +228,7 @@ int intel_fbdev_init(struct drm_device *dev)
 	ifbdev->helper.funcs = &intel_fb_helper_funcs;
 
 	ret = drm_fb_helper_init(dev, &ifbdev->helper,
-				 dev_priv->num_pipe,
+				 INTEL_INFO(dev)->num_pipes,
 				 INTELFB_CONN_LIMIT);
 	if (ret) {
 		kfree(ifbdev);
@@ -282,6 +283,9 @@ void intel_fb_restore_mode(struct drm_device *dev)
 	struct drm_mode_config *config = &dev->mode_config;
 	struct drm_plane *plane;
 
+	if (INTEL_INFO(dev)->num_pipes == 0)
+		return;
+
 	drm_modeset_lock_all(dev);
 
 	ret = drm_fb_helper_restore_fbdev_mode(&dev_priv->fbdev->helper);
diff --git a/drivers/gpu/drm/i915/intel_hdmi.c b/drivers/gpu/drm/i915/intel_hdmi.c
index fa8ec4a26041..a9057930f2b2 100644
--- a/drivers/gpu/drm/i915/intel_hdmi.c
+++ b/drivers/gpu/drm/i915/intel_hdmi.c
@@ -50,7 +50,7 @@ assert_hdmi_port_disabled(struct intel_hdmi *intel_hdmi)
 
 	enabled_bits = HAS_DDI(dev) ? DDI_BUF_CTL_ENABLE : SDVO_ENABLE;
 
-	WARN(I915_READ(intel_hdmi->sdvox_reg) & enabled_bits,
+	WARN(I915_READ(intel_hdmi->hdmi_reg) & enabled_bits,
 	     "HDMI port enabled, expecting disabled\n");
 }
 
@@ -120,13 +120,14 @@ static u32 hsw_infoframe_enable(struct dip_infoframe *frame)
 	}
 }
 
-static u32 hsw_infoframe_data_reg(struct dip_infoframe *frame, enum pipe pipe)
+static u32 hsw_infoframe_data_reg(struct dip_infoframe *frame,
+				  enum transcoder cpu_transcoder)
 {
 	switch (frame->type) {
 	case DIP_TYPE_AVI:
-		return HSW_TVIDEO_DIP_AVI_DATA(pipe);
+		return HSW_TVIDEO_DIP_AVI_DATA(cpu_transcoder);
 	case DIP_TYPE_SPD:
-		return HSW_TVIDEO_DIP_SPD_DATA(pipe);
+		return HSW_TVIDEO_DIP_SPD_DATA(cpu_transcoder);
 	default:
 		DRM_DEBUG_DRIVER("unknown info frame type %d\n", frame->type);
 		return 0;
@@ -293,8 +294,8 @@ static void hsw_write_infoframe(struct drm_encoder *encoder,
 	struct drm_device *dev = encoder->dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_crtc *intel_crtc = to_intel_crtc(encoder->crtc);
-	u32 ctl_reg = HSW_TVIDEO_DIP_CTL(intel_crtc->pipe);
-	u32 data_reg = hsw_infoframe_data_reg(frame, intel_crtc->pipe);
+	u32 ctl_reg = HSW_TVIDEO_DIP_CTL(intel_crtc->config.cpu_transcoder);
+	u32 data_reg = hsw_infoframe_data_reg(frame, intel_crtc->config.cpu_transcoder);
 	unsigned int i, len = DIP_HEADER_SIZE + frame->len;
 	u32 val = I915_READ(ctl_reg);
 
@@ -332,6 +333,7 @@ static void intel_hdmi_set_avi_infoframe(struct drm_encoder *encoder,
 					 struct drm_display_mode *adjusted_mode)
 {
 	struct intel_hdmi *intel_hdmi = enc_to_intel_hdmi(encoder);
+	struct intel_crtc *intel_crtc = to_intel_crtc(encoder->crtc);
 	struct dip_infoframe avi_if = {
 		.type = DIP_TYPE_AVI,
 		.ver = DIP_VERSION_AVI,
@@ -342,7 +344,7 @@ static void intel_hdmi_set_avi_infoframe(struct drm_encoder *encoder,
 		avi_if.body.avi.YQ_CN_PR |= DIP_AVI_PR_2;
 
 	if (intel_hdmi->rgb_quant_range_selectable) {
-		if (adjusted_mode->private_flags & INTEL_MODE_LIMITED_COLOR_RANGE)
+		if (intel_crtc->config.limited_color_range)
 			avi_if.body.avi.ITC_EC_Q_SC |= DIP_AVI_RGB_QUANT_RANGE_LIMITED;
 		else
 			avi_if.body.avi.ITC_EC_Q_SC |= DIP_AVI_RGB_QUANT_RANGE_FULL;
@@ -568,7 +570,7 @@ static void hsw_set_infoframes(struct drm_encoder *encoder,
 	struct drm_i915_private *dev_priv = encoder->dev->dev_private;
 	struct intel_crtc *intel_crtc = to_intel_crtc(encoder->crtc);
 	struct intel_hdmi *intel_hdmi = enc_to_intel_hdmi(encoder);
-	u32 reg = HSW_TVIDEO_DIP_CTL(intel_crtc->pipe);
+	u32 reg = HSW_TVIDEO_DIP_CTL(intel_crtc->config.cpu_transcoder);
 	u32 val = I915_READ(reg);
 
 	assert_hdmi_port_disabled(intel_hdmi);
@@ -597,40 +599,40 @@ static void intel_hdmi_mode_set(struct drm_encoder *encoder,
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_crtc *intel_crtc = to_intel_crtc(encoder->crtc);
 	struct intel_hdmi *intel_hdmi = enc_to_intel_hdmi(encoder);
-	u32 sdvox;
+	u32 hdmi_val;
 
-	sdvox = SDVO_ENCODING_HDMI;
-	if (!HAS_PCH_SPLIT(dev))
-		sdvox |= intel_hdmi->color_range;
+	hdmi_val = SDVO_ENCODING_HDMI;
+	if (!HAS_PCH_SPLIT(dev) && !IS_VALLEYVIEW(dev))
+		hdmi_val |= intel_hdmi->color_range;
 	if (adjusted_mode->flags & DRM_MODE_FLAG_PVSYNC)
-		sdvox |= SDVO_VSYNC_ACTIVE_HIGH;
+		hdmi_val |= SDVO_VSYNC_ACTIVE_HIGH;
 	if (adjusted_mode->flags & DRM_MODE_FLAG_PHSYNC)
-		sdvox |= SDVO_HSYNC_ACTIVE_HIGH;
+		hdmi_val |= SDVO_HSYNC_ACTIVE_HIGH;
 
-	if (intel_crtc->bpp > 24)
-		sdvox |= COLOR_FORMAT_12bpc;
+	if (intel_crtc->config.pipe_bpp > 24)
+		hdmi_val |= HDMI_COLOR_FORMAT_12bpc;
 	else
-		sdvox |= COLOR_FORMAT_8bpc;
+		hdmi_val |= SDVO_COLOR_FORMAT_8bpc;
 
 	/* Required on CPT */
 	if (intel_hdmi->has_hdmi_sink && HAS_PCH_CPT(dev))
-		sdvox |= HDMI_MODE_SELECT;
+		hdmi_val |= HDMI_MODE_SELECT_HDMI;
 
 	if (intel_hdmi->has_audio) {
 		DRM_DEBUG_DRIVER("Enabling HDMI audio on pipe %c\n",
 				 pipe_name(intel_crtc->pipe));
-		sdvox |= SDVO_AUDIO_ENABLE;
-		sdvox |= SDVO_NULL_PACKETS_DURING_VSYNC;
+		hdmi_val |= SDVO_AUDIO_ENABLE;
+		hdmi_val |= HDMI_MODE_SELECT_HDMI;
 		intel_write_eld(encoder, adjusted_mode);
 	}
 
 	if (HAS_PCH_CPT(dev))
-		sdvox |= PORT_TRANS_SEL_CPT(intel_crtc->pipe);
-	else if (intel_crtc->pipe == PIPE_B)
-		sdvox |= SDVO_PIPE_B_SELECT;
+		hdmi_val |= SDVO_PIPE_SEL_CPT(intel_crtc->pipe);
+	else
+		hdmi_val |= SDVO_PIPE_SEL(intel_crtc->pipe);
 
-	I915_WRITE(intel_hdmi->sdvox_reg, sdvox);
-	POSTING_READ(intel_hdmi->sdvox_reg);
+	I915_WRITE(intel_hdmi->hdmi_reg, hdmi_val);
+	POSTING_READ(intel_hdmi->hdmi_reg);
 
 	intel_hdmi->set_infoframes(encoder, adjusted_mode);
 }
@@ -643,7 +645,7 @@ static bool intel_hdmi_get_hw_state(struct intel_encoder *encoder,
 	struct intel_hdmi *intel_hdmi = enc_to_intel_hdmi(&encoder->base);
 	u32 tmp;
 
-	tmp = I915_READ(intel_hdmi->sdvox_reg);
+	tmp = I915_READ(intel_hdmi->hdmi_reg);
 
 	if (!(tmp & SDVO_ENABLE))
 		return false;
@@ -660,6 +662,7 @@ static void intel_enable_hdmi(struct intel_encoder *encoder)
 {
 	struct drm_device *dev = encoder->base.dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct intel_crtc *intel_crtc = to_intel_crtc(encoder->base.crtc);
 	struct intel_hdmi *intel_hdmi = enc_to_intel_hdmi(&encoder->base);
 	u32 temp;
 	u32 enable_bits = SDVO_ENABLE;
@@ -667,38 +670,32 @@ static void intel_enable_hdmi(struct intel_encoder *encoder)
 	if (intel_hdmi->has_audio)
 		enable_bits |= SDVO_AUDIO_ENABLE;
 
-	temp = I915_READ(intel_hdmi->sdvox_reg);
+	temp = I915_READ(intel_hdmi->hdmi_reg);
 
 	/* HW workaround for IBX, we need to move the port to transcoder A
-	 * before disabling it. */
-	if (HAS_PCH_IBX(dev)) {
-		struct drm_crtc *crtc = encoder->base.crtc;
-		int pipe = crtc ? to_intel_crtc(crtc)->pipe : -1;
-
-		/* Restore the transcoder select bit. */
-		if (pipe == PIPE_B)
-			enable_bits |= SDVO_PIPE_B_SELECT;
-	}
+	 * before disabling it, so restore the transcoder select bit here. */
+	if (HAS_PCH_IBX(dev))
+		enable_bits |= SDVO_PIPE_SEL(intel_crtc->pipe);
 
 	/* HW workaround, need to toggle enable bit off and on for 12bpc, but
 	 * we do this anyway which shows more stable in testing.
 	 */
 	if (HAS_PCH_SPLIT(dev)) {
-		I915_WRITE(intel_hdmi->sdvox_reg, temp & ~SDVO_ENABLE);
-		POSTING_READ(intel_hdmi->sdvox_reg);
+		I915_WRITE(intel_hdmi->hdmi_reg, temp & ~SDVO_ENABLE);
+		POSTING_READ(intel_hdmi->hdmi_reg);
 	}
 
 	temp |= enable_bits;
 
-	I915_WRITE(intel_hdmi->sdvox_reg, temp);
-	POSTING_READ(intel_hdmi->sdvox_reg);
+	I915_WRITE(intel_hdmi->hdmi_reg, temp);
+	POSTING_READ(intel_hdmi->hdmi_reg);
 
 	/* HW workaround, need to write this twice for issue that may result
 	 * in first write getting masked.
 	 */
 	if (HAS_PCH_SPLIT(dev)) {
-		I915_WRITE(intel_hdmi->sdvox_reg, temp);
-		POSTING_READ(intel_hdmi->sdvox_reg);
+		I915_WRITE(intel_hdmi->hdmi_reg, temp);
+		POSTING_READ(intel_hdmi->hdmi_reg);
 	}
 }
 
@@ -710,7 +707,7 @@ static void intel_disable_hdmi(struct intel_encoder *encoder)
 	u32 temp;
 	u32 enable_bits = SDVO_ENABLE | SDVO_AUDIO_ENABLE;
 
-	temp = I915_READ(intel_hdmi->sdvox_reg);
+	temp = I915_READ(intel_hdmi->hdmi_reg);
 
 	/* HW workaround for IBX, we need to move the port to transcoder A
 	 * before disabling it. */
@@ -720,12 +717,12 @@ static void intel_disable_hdmi(struct intel_encoder *encoder)
 
 		if (temp & SDVO_PIPE_B_SELECT) {
 			temp &= ~SDVO_PIPE_B_SELECT;
-			I915_WRITE(intel_hdmi->sdvox_reg, temp);
-			POSTING_READ(intel_hdmi->sdvox_reg);
+			I915_WRITE(intel_hdmi->hdmi_reg, temp);
+			POSTING_READ(intel_hdmi->hdmi_reg);
 
 			/* Again we need to write this twice. */
-			I915_WRITE(intel_hdmi->sdvox_reg, temp);
-			POSTING_READ(intel_hdmi->sdvox_reg);
+			I915_WRITE(intel_hdmi->hdmi_reg, temp);
+			POSTING_READ(intel_hdmi->hdmi_reg);
 
 			/* Transcoder selection bits only update
 			 * effectively on vblank. */
@@ -740,21 +737,21 @@ static void intel_disable_hdmi(struct intel_encoder *encoder)
 	 * we do this anyway which shows more stable in testing.
 	 */
 	if (HAS_PCH_SPLIT(dev)) {
-		I915_WRITE(intel_hdmi->sdvox_reg, temp & ~SDVO_ENABLE);
-		POSTING_READ(intel_hdmi->sdvox_reg);
+		I915_WRITE(intel_hdmi->hdmi_reg, temp & ~SDVO_ENABLE);
+		POSTING_READ(intel_hdmi->hdmi_reg);
 	}
 
 	temp &= ~enable_bits;
 
-	I915_WRITE(intel_hdmi->sdvox_reg, temp);
-	POSTING_READ(intel_hdmi->sdvox_reg);
+	I915_WRITE(intel_hdmi->hdmi_reg, temp);
+	POSTING_READ(intel_hdmi->hdmi_reg);
 
 	/* HW workaround, need to write this twice for issue that may result
 	 * in first write getting masked.
 	 */
 	if (HAS_PCH_SPLIT(dev)) {
-		I915_WRITE(intel_hdmi->sdvox_reg, temp);
-		POSTING_READ(intel_hdmi->sdvox_reg);
+		I915_WRITE(intel_hdmi->hdmi_reg, temp);
+		POSTING_READ(intel_hdmi->hdmi_reg);
 	}
 }
 
@@ -772,23 +769,40 @@ static int intel_hdmi_mode_valid(struct drm_connector *connector,
 	return MODE_OK;
 }
 
-bool intel_hdmi_mode_fixup(struct drm_encoder *encoder,
-			   const struct drm_display_mode *mode,
-			   struct drm_display_mode *adjusted_mode)
+bool intel_hdmi_compute_config(struct intel_encoder *encoder,
+			       struct intel_crtc_config *pipe_config)
 {
-	struct intel_hdmi *intel_hdmi = enc_to_intel_hdmi(encoder);
+	struct intel_hdmi *intel_hdmi = enc_to_intel_hdmi(&encoder->base);
+	struct drm_device *dev = encoder->base.dev;
+	struct drm_display_mode *adjusted_mode = &pipe_config->adjusted_mode;
 
 	if (intel_hdmi->color_range_auto) {
 		/* See CEA-861-E - 5.1 Default Encoding Parameters */
 		if (intel_hdmi->has_hdmi_sink &&
 		    drm_match_cea_mode(adjusted_mode) > 1)
-			intel_hdmi->color_range = SDVO_COLOR_RANGE_16_235;
+			intel_hdmi->color_range = HDMI_COLOR_RANGE_16_235;
 		else
 			intel_hdmi->color_range = 0;
 	}
 
 	if (intel_hdmi->color_range)
-		adjusted_mode->private_flags |= INTEL_MODE_LIMITED_COLOR_RANGE;
+		pipe_config->limited_color_range = true;
+
+	if (HAS_PCH_SPLIT(dev) && !HAS_DDI(dev))
+		pipe_config->has_pch_encoder = true;
+
+	/*
+	 * HDMI is either 12 or 8, so if the display lets 10bpc sneak
+	 * through, clamp it down. Note that g4x/vlv don't support 12bpc hdmi
+	 * outputs.
+	 */
+	if (pipe_config->pipe_bpp > 8*3 && HAS_PCH_SPLIT(dev)) {
+		DRM_DEBUG_KMS("forcing bpc to 12 for HDMI\n");
+		pipe_config->pipe_bpp = 12*3;
+	} else {
+		DRM_DEBUG_KMS("forcing bpc to 8 for HDMI\n");
+		pipe_config->pipe_bpp = 8*3;
+	}
 
 	return true;
 }
@@ -906,6 +920,9 @@ intel_hdmi_set_property(struct drm_connector *connector,
 	}
 
 	if (property == dev_priv->broadcast_rgb_property) {
+		bool old_auto = intel_hdmi->color_range_auto;
+		uint32_t old_range = intel_hdmi->color_range;
+
 		switch (val) {
 		case INTEL_BROADCAST_RGB_AUTO:
 			intel_hdmi->color_range_auto = true;
@@ -916,11 +933,16 @@ intel_hdmi_set_property(struct drm_connector *connector,
 			break;
 		case INTEL_BROADCAST_RGB_LIMITED:
 			intel_hdmi->color_range_auto = false;
-			intel_hdmi->color_range = SDVO_COLOR_RANGE_16_235;
+			intel_hdmi->color_range = HDMI_COLOR_RANGE_16_235;
 			break;
 		default:
 			return -EINVAL;
 		}
+
+		if (old_auto == intel_hdmi->color_range_auto &&
+		    old_range == intel_hdmi->color_range)
+			return 0;
+
 		goto done;
 	}
 
@@ -941,7 +963,6 @@ static void intel_hdmi_destroy(struct drm_connector *connector)
 }
 
 static const struct drm_encoder_helper_funcs intel_hdmi_helper_funcs = {
-	.mode_fixup = intel_hdmi_mode_fixup,
 	.mode_set = intel_hdmi_mode_set,
 };
 
@@ -985,36 +1006,36 @@ void intel_hdmi_init_connector(struct intel_digital_port *intel_dig_port,
 			   DRM_MODE_CONNECTOR_HDMIA);
 	drm_connector_helper_add(connector, &intel_hdmi_connector_helper_funcs);
 
-	connector->polled = DRM_CONNECTOR_POLL_HPD;
 	connector->interlace_allowed = 1;
 	connector->doublescan_allowed = 0;
 
 	switch (port) {
 	case PORT_B:
 		intel_hdmi->ddc_bus = GMBUS_PORT_DPB;
-		dev_priv->hotplug_supported_mask |= PORTB_HOTPLUG_INT_STATUS;
+		intel_encoder->hpd_pin = HPD_PORT_B;
 		break;
 	case PORT_C:
 		intel_hdmi->ddc_bus = GMBUS_PORT_DPC;
-		dev_priv->hotplug_supported_mask |= PORTC_HOTPLUG_INT_STATUS;
+		intel_encoder->hpd_pin = HPD_PORT_C;
 		break;
 	case PORT_D:
 		intel_hdmi->ddc_bus = GMBUS_PORT_DPD;
-		dev_priv->hotplug_supported_mask |= PORTD_HOTPLUG_INT_STATUS;
+		intel_encoder->hpd_pin = HPD_PORT_D;
 		break;
 	case PORT_A:
+		intel_encoder->hpd_pin = HPD_PORT_A;
 		/* Internal port only for eDP. */
 	default:
 		BUG();
 	}
 
-	if (!HAS_PCH_SPLIT(dev)) {
-		intel_hdmi->write_infoframe = g4x_write_infoframe;
-		intel_hdmi->set_infoframes = g4x_set_infoframes;
-	} else if (IS_VALLEYVIEW(dev)) {
+	if (IS_VALLEYVIEW(dev)) {
 		intel_hdmi->write_infoframe = vlv_write_infoframe;
 		intel_hdmi->set_infoframes = vlv_set_infoframes;
-	} else if (IS_HASWELL(dev)) {
+	} else if (!HAS_PCH_SPLIT(dev)) {
+		intel_hdmi->write_infoframe = g4x_write_infoframe;
+		intel_hdmi->set_infoframes = g4x_set_infoframes;
+	} else if (HAS_DDI(dev)) {
 		intel_hdmi->write_infoframe = hsw_write_infoframe;
 		intel_hdmi->set_infoframes = hsw_set_infoframes;
 	} else if (HAS_PCH_IBX(dev)) {
@@ -1045,7 +1066,7 @@ void intel_hdmi_init_connector(struct intel_digital_port *intel_dig_port,
 	}
 }
 
-void intel_hdmi_init(struct drm_device *dev, int sdvox_reg, enum port port)
+void intel_hdmi_init(struct drm_device *dev, int hdmi_reg, enum port port)
 {
 	struct intel_digital_port *intel_dig_port;
 	struct intel_encoder *intel_encoder;
@@ -1069,6 +1090,7 @@ void intel_hdmi_init(struct drm_device *dev, int sdvox_reg, enum port port)
 			 DRM_MODE_ENCODER_TMDS);
 	drm_encoder_helper_add(&intel_encoder->base, &intel_hdmi_helper_funcs);
 
+	intel_encoder->compute_config = intel_hdmi_compute_config;
 	intel_encoder->enable = intel_enable_hdmi;
 	intel_encoder->disable = intel_disable_hdmi;
 	intel_encoder->get_hw_state = intel_hdmi_get_hw_state;
@@ -1078,7 +1100,7 @@ void intel_hdmi_init(struct drm_device *dev, int sdvox_reg, enum port port)
 	intel_encoder->cloneable = false;
 
 	intel_dig_port->port = port;
-	intel_dig_port->hdmi.sdvox_reg = sdvox_reg;
+	intel_dig_port->hdmi.hdmi_reg = hdmi_reg;
 	intel_dig_port->dp.output_reg = 0;
 
 	intel_hdmi_init_connector(intel_dig_port, intel_connector);
diff --git a/drivers/gpu/drm/i915/intel_i2c.c b/drivers/gpu/drm/i915/intel_i2c.c
index ef4744e1bf0b..5d245031e391 100644
--- a/drivers/gpu/drm/i915/intel_i2c.c
+++ b/drivers/gpu/drm/i915/intel_i2c.c
@@ -522,7 +522,9 @@ int intel_setup_gmbus(struct drm_device *dev)
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	int ret, i;
 
-	if (HAS_PCH_SPLIT(dev))
+	if (HAS_PCH_NOP(dev))
+		return 0;
+	else if (HAS_PCH_SPLIT(dev))
 		dev_priv->gpio_mmio_base = PCH_GPIOA - GPIOA;
 	else if (IS_VALLEYVIEW(dev))
 		dev_priv->gpio_mmio_base = VLV_DISPLAY_BASE;
diff --git a/drivers/gpu/drm/i915/intel_lvds.c b/drivers/gpu/drm/i915/intel_lvds.c
index 3d1d97488cc9..f36f1baabd5a 100644
--- a/drivers/gpu/drm/i915/intel_lvds.c
+++ b/drivers/gpu/drm/i915/intel_lvds.c
@@ -261,8 +261,6 @@ centre_horizontally(struct drm_display_mode *mode,
 
 	mode->crtc_hsync_start = mode->crtc_hblank_start + sync_pos;
 	mode->crtc_hsync_end = mode->crtc_hsync_start + sync_width;
-
-	mode->private_flags |= INTEL_MODE_CRTC_TIMINGS_SET;
 }
 
 static void
@@ -284,8 +282,6 @@ centre_vertically(struct drm_display_mode *mode,
 
 	mode->crtc_vsync_start = mode->crtc_vblank_start + sync_pos;
 	mode->crtc_vsync_end = mode->crtc_vsync_start + sync_width;
-
-	mode->private_flags |= INTEL_MODE_CRTC_TIMINGS_SET;
 }
 
 static inline u32 panel_fitter_scaling(u32 source, u32 target)
@@ -301,17 +297,20 @@ static inline u32 panel_fitter_scaling(u32 source, u32 target)
 	return (FACTOR * ratio + FACTOR/2) / FACTOR;
 }
 
-static bool intel_lvds_mode_fixup(struct drm_encoder *encoder,
-				  const struct drm_display_mode *mode,
-				  struct drm_display_mode *adjusted_mode)
+static bool intel_lvds_compute_config(struct intel_encoder *intel_encoder,
+				      struct intel_crtc_config *pipe_config)
 {
-	struct drm_device *dev = encoder->dev;
+	struct drm_device *dev = intel_encoder->base.dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct intel_lvds_encoder *lvds_encoder = to_lvds_encoder(encoder);
+	struct intel_lvds_encoder *lvds_encoder =
+		to_lvds_encoder(&intel_encoder->base);
 	struct intel_connector *intel_connector =
 		&lvds_encoder->attached_connector->base;
+	struct drm_display_mode *adjusted_mode = &pipe_config->adjusted_mode;
+	struct drm_display_mode *mode = &pipe_config->requested_mode;
 	struct intel_crtc *intel_crtc = lvds_encoder->base.new_crtc;
 	u32 pfit_control = 0, pfit_pgm_ratios = 0, border = 0;
+	unsigned int lvds_bpp;
 	int pipe;
 
 	/* Should never happen!! */
@@ -323,6 +322,17 @@ static bool intel_lvds_mode_fixup(struct drm_encoder *encoder,
 	if (intel_encoder_check_is_cloned(&lvds_encoder->base))
 		return false;
 
+	if ((I915_READ(lvds_encoder->reg) & LVDS_A3_POWER_MASK) ==
+	    LVDS_A3_POWER_UP)
+		lvds_bpp = 8*3;
+	else
+		lvds_bpp = 6*3;
+
+	if (lvds_bpp != pipe_config->pipe_bpp) {
+		DRM_DEBUG_KMS("forcing display bpp (was %d) to LVDS (%d)\n",
+			      pipe_config->pipe_bpp, lvds_bpp);
+		pipe_config->pipe_bpp = lvds_bpp;
+	}
 	/*
 	 * We have timings from the BIOS for the panel, put them in
 	 * to the adjusted mode.  The CRTC will be set up for this mode,
@@ -333,6 +343,8 @@ static bool intel_lvds_mode_fixup(struct drm_encoder *encoder,
 			       adjusted_mode);
 
 	if (HAS_PCH_SPLIT(dev)) {
+		pipe_config->has_pch_encoder = true;
+
 		intel_pch_panel_fitting(dev,
 					intel_connector->panel.fitting_mode,
 					mode, adjusted_mode);
@@ -359,6 +371,7 @@ static bool intel_lvds_mode_fixup(struct drm_encoder *encoder,
 		I915_WRITE(BCLRPAT(pipe), 0);
 
 	drm_mode_set_crtcinfo(adjusted_mode, 0);
+	pipe_config->timings_set = true;
 
 	switch (intel_connector->panel.fitting_mode) {
 	case DRM_MODE_SCALE_CENTER:
@@ -618,7 +631,6 @@ static void intel_lvds_destroy(struct drm_connector *connector)
 	if (!IS_ERR_OR_NULL(lvds_connector->base.edid))
 		kfree(lvds_connector->base.edid);
 
-	intel_panel_destroy_backlight(connector->dev);
 	intel_panel_fini(&lvds_connector->base.panel);
 
 	drm_sysfs_connector_remove(connector);
@@ -661,7 +673,6 @@ static int intel_lvds_set_property(struct drm_connector *connector,
 }
 
 static const struct drm_encoder_helper_funcs intel_lvds_helper_funcs = {
-	.mode_fixup = intel_lvds_mode_fixup,
 	.mode_set = intel_lvds_mode_set,
 };
 
@@ -850,6 +861,14 @@ static const struct dmi_system_id intel_no_lvds[] = {
 			DMI_MATCH(DMI_PRODUCT_NAME, "X7SPA-H"),
 		},
 	},
+	{
+		.callback = intel_no_lvds_dmi_callback,
+		.ident = "Fujitsu Esprimo Q900",
+		.matches = {
+			DMI_MATCH(DMI_SYS_VENDOR, "FUJITSU"),
+			DMI_MATCH(DMI_PRODUCT_NAME, "ESPRIMO Q900"),
+		},
+	},
 
 	{ }	/* terminating entry */
 };
@@ -1019,12 +1038,15 @@ static bool intel_lvds_supported(struct drm_device *dev)
 {
 	/* With the introduction of the PCH we gained a dedicated
 	 * LVDS presence pin, use it. */
-	if (HAS_PCH_SPLIT(dev))
+	if (HAS_PCH_IBX(dev) || HAS_PCH_CPT(dev))
 		return true;
 
 	/* Otherwise LVDS was only attached to mobile products,
 	 * except for the inglorious 830gm */
-	return IS_MOBILE(dev) && !IS_I830(dev);
+	if (INTEL_INFO(dev)->gen <= 4 && IS_MOBILE(dev) && !IS_I830(dev))
+		return true;
+
+	return false;
 }
 
 /**
@@ -1102,6 +1124,7 @@ bool intel_lvds_init(struct drm_device *dev)
 	intel_encoder->enable = intel_enable_lvds;
 	intel_encoder->pre_enable = intel_pre_enable_lvds;
 	intel_encoder->pre_pll_enable = intel_pre_pll_enable_lvds;
+	intel_encoder->compute_config = intel_lvds_compute_config;
 	intel_encoder->disable = intel_disable_lvds;
 	intel_encoder->get_hw_state = intel_lvds_get_hw_state;
 	intel_connector->get_hw_state = intel_connector_get_hw_state;
diff --git a/drivers/gpu/drm/i915/intel_panel.c b/drivers/gpu/drm/i915/intel_panel.c
index bee8cb6108a7..eb5e6e95f3c7 100644
--- a/drivers/gpu/drm/i915/intel_panel.c
+++ b/drivers/gpu/drm/i915/intel_panel.c
@@ -286,8 +286,11 @@ void intel_panel_set_backlight(struct drm_device *dev, u32 level)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 
-	dev_priv->backlight_level = level;
-	if (dev_priv->backlight_enabled)
+	dev_priv->backlight.level = level;
+	if (dev_priv->backlight.device)
+		dev_priv->backlight.device->props.brightness = level;
+
+	if (dev_priv->backlight.enabled)
 		intel_panel_actually_set_backlight(dev, level);
 }
 
@@ -295,7 +298,7 @@ void intel_panel_disable_backlight(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 
-	dev_priv->backlight_enabled = false;
+	dev_priv->backlight.enabled = false;
 	intel_panel_actually_set_backlight(dev, 0);
 
 	if (INTEL_INFO(dev)->gen >= 4) {
@@ -318,8 +321,12 @@ void intel_panel_enable_backlight(struct drm_device *dev,
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 
-	if (dev_priv->backlight_level == 0)
-		dev_priv->backlight_level = intel_panel_get_max_backlight(dev);
+	if (dev_priv->backlight.level == 0) {
+		dev_priv->backlight.level = intel_panel_get_max_backlight(dev);
+		if (dev_priv->backlight.device)
+			dev_priv->backlight.device->props.brightness =
+				dev_priv->backlight.level;
+	}
 
 	if (INTEL_INFO(dev)->gen >= 4) {
 		uint32_t reg, tmp;
@@ -335,7 +342,7 @@ void intel_panel_enable_backlight(struct drm_device *dev,
 		if (tmp & BLM_PWM_ENABLE)
 			goto set_level;
 
-		if (dev_priv->num_pipe == 3)
+		if (INTEL_INFO(dev)->num_pipes == 3)
 			tmp &= ~BLM_PIPE_SELECT_IVB;
 		else
 			tmp &= ~BLM_PIPE_SELECT;
@@ -360,16 +367,16 @@ set_level:
 	 * BLC_PWM_CPU_CTL may be cleared to zero automatically when these
 	 * registers are set.
 	 */
-	dev_priv->backlight_enabled = true;
-	intel_panel_actually_set_backlight(dev, dev_priv->backlight_level);
+	dev_priv->backlight.enabled = true;
+	intel_panel_actually_set_backlight(dev, dev_priv->backlight.level);
 }
 
 static void intel_panel_init_backlight(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 
-	dev_priv->backlight_level = intel_panel_get_backlight(dev);
-	dev_priv->backlight_enabled = dev_priv->backlight_level != 0;
+	dev_priv->backlight.level = intel_panel_get_backlight(dev);
+	dev_priv->backlight.enabled = dev_priv->backlight.level != 0;
 }
 
 enum drm_connector_status
@@ -405,8 +412,7 @@ static int intel_panel_update_status(struct backlight_device *bd)
 static int intel_panel_get_brightness(struct backlight_device *bd)
 {
 	struct drm_device *dev = bl_get_data(bd);
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	return dev_priv->backlight_level;
+	return intel_panel_get_backlight(dev);
 }
 
 static const struct backlight_ops intel_panel_bl_ops = {
@@ -422,33 +428,38 @@ int intel_panel_setup_backlight(struct drm_connector *connector)
 
 	intel_panel_init_backlight(dev);
 
+	if (WARN_ON(dev_priv->backlight.device))
+		return -ENODEV;
+
 	memset(&props, 0, sizeof(props));
 	props.type = BACKLIGHT_RAW;
+	props.brightness = dev_priv->backlight.level;
 	props.max_brightness = _intel_panel_get_max_backlight(dev);
 	if (props.max_brightness == 0) {
 		DRM_DEBUG_DRIVER("Failed to get maximum backlight value\n");
 		return -ENODEV;
 	}
-	dev_priv->backlight =
+	dev_priv->backlight.device =
 		backlight_device_register("intel_backlight",
 					  &connector->kdev, dev,
 					  &intel_panel_bl_ops, &props);
 
-	if (IS_ERR(dev_priv->backlight)) {
+	if (IS_ERR(dev_priv->backlight.device)) {
 		DRM_ERROR("Failed to register backlight: %ld\n",
-			  PTR_ERR(dev_priv->backlight));
-		dev_priv->backlight = NULL;
+			  PTR_ERR(dev_priv->backlight.device));
+		dev_priv->backlight.device = NULL;
 		return -ENODEV;
 	}
-	dev_priv->backlight->props.brightness = intel_panel_get_backlight(dev);
 	return 0;
 }
 
 void intel_panel_destroy_backlight(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	if (dev_priv->backlight)
-		backlight_device_unregister(dev_priv->backlight);
+	if (dev_priv->backlight.device) {
+		backlight_device_unregister(dev_priv->backlight.device);
+		dev_priv->backlight.device = NULL;
+	}
 }
 #else
 int intel_panel_setup_backlight(struct drm_connector *connector)
diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c
index adca00783e61..de3b0dc5658b 100644
--- a/drivers/gpu/drm/i915/intel_pm.c
+++ b/drivers/gpu/drm/i915/intel_pm.c
@@ -2460,10 +2460,14 @@ void gen6_set_rps(struct drm_device *dev, u8 val)
 	if (val == dev_priv->rps.cur_delay)
 		return;
 
-	I915_WRITE(GEN6_RPNSWREQ,
-		   GEN6_FREQUENCY(val) |
-		   GEN6_OFFSET(0) |
-		   GEN6_AGGRESSIVE_TURBO);
+	if (IS_HASWELL(dev))
+		I915_WRITE(GEN6_RPNSWREQ,
+			   HSW_FREQUENCY(val));
+	else
+		I915_WRITE(GEN6_RPNSWREQ,
+			   GEN6_FREQUENCY(val) |
+			   GEN6_OFFSET(0) |
+			   GEN6_AGGRESSIVE_TURBO);
 
 	/* Make sure we continue to get interrupts
 	 * until we hit the minimum or maximum frequencies.
@@ -2554,8 +2558,8 @@ static void gen6_enable_rps(struct drm_device *dev)
 	rp_state_cap = I915_READ(GEN6_RP_STATE_CAP);
 	gt_perf_status = I915_READ(GEN6_GT_PERF_STATUS);
 
-	/* In units of 100MHz */
-	dev_priv->rps.max_delay = rp_state_cap & 0xff;
+	/* In units of 50MHz */
+	dev_priv->rps.hw_max = dev_priv->rps.max_delay = rp_state_cap & 0xff;
 	dev_priv->rps.min_delay = (rp_state_cap & 0xff0000) >> 16;
 	dev_priv->rps.cur_delay = 0;
 
@@ -2601,12 +2605,19 @@ static void gen6_enable_rps(struct drm_device *dev)
 		   GEN6_RC_CTL_EI_MODE(1) |
 		   GEN6_RC_CTL_HW_ENABLE);
 
-	I915_WRITE(GEN6_RPNSWREQ,
-		   GEN6_FREQUENCY(10) |
-		   GEN6_OFFSET(0) |
-		   GEN6_AGGRESSIVE_TURBO);
-	I915_WRITE(GEN6_RC_VIDEO_FREQ,
-		   GEN6_FREQUENCY(12));
+	if (IS_HASWELL(dev)) {
+		I915_WRITE(GEN6_RPNSWREQ,
+			   HSW_FREQUENCY(10));
+		I915_WRITE(GEN6_RC_VIDEO_FREQ,
+			   HSW_FREQUENCY(12));
+	} else {
+		I915_WRITE(GEN6_RPNSWREQ,
+			   GEN6_FREQUENCY(10) |
+			   GEN6_OFFSET(0) |
+			   GEN6_AGGRESSIVE_TURBO);
+		I915_WRITE(GEN6_RC_VIDEO_FREQ,
+			   GEN6_FREQUENCY(12));
+	}
 
 	I915_WRITE(GEN6_RP_DOWN_TIMEOUT, 1000000);
 	I915_WRITE(GEN6_RP_INTERRUPT_LIMITS,
@@ -2631,9 +2642,11 @@ static void gen6_enable_rps(struct drm_device *dev)
 	if (!ret) {
 		pcu_mbox = 0;
 		ret = sandybridge_pcode_read(dev_priv, GEN6_READ_OC_PARAMS, &pcu_mbox);
-		if (ret && pcu_mbox & (1<<31)) { /* OC supported */
-			dev_priv->rps.max_delay = pcu_mbox & 0xff;
-			DRM_DEBUG_DRIVER("overclocking supported, adjusting frequency max to %dMHz\n", pcu_mbox * 50);
+		if (!ret && (pcu_mbox & (1<<31))) { /* OC supported */
+			DRM_DEBUG_DRIVER("Overclocking supported. Max: %dMHz, Overclock max: %dMHz\n",
+					 (dev_priv->rps.max_delay & 0xff) * 50,
+					 (pcu_mbox & 0xff) * 50);
+			dev_priv->rps.hw_max = pcu_mbox & 0xff;
 		}
 	} else {
 		DRM_DEBUG_DRIVER("Failed to set the min frequency\n");
@@ -2671,8 +2684,8 @@ static void gen6_update_ring_freq(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	int min_freq = 15;
-	int gpu_freq;
-	unsigned int ia_freq, max_ia_freq;
+	unsigned int gpu_freq;
+	unsigned int max_ia_freq, min_ring_freq;
 	int scaling_factor = 180;
 
 	WARN_ON(!mutex_is_locked(&dev_priv->rps.hw_lock));
@@ -2688,6 +2701,10 @@ static void gen6_update_ring_freq(struct drm_device *dev)
 	/* Convert from kHz to MHz */
 	max_ia_freq /= 1000;
 
+	min_ring_freq = I915_READ(MCHBAR_MIRROR_BASE_SNB + DCLK);
+	/* convert DDR frequency from units of 133.3MHz to bandwidth */
+	min_ring_freq = (2 * 4 * min_ring_freq + 2) / 3;
+
 	/*
 	 * For each potential GPU frequency, load a ring frequency we'd like
 	 * to use for memory access.  We do this by specifying the IA frequency
@@ -2696,21 +2713,32 @@ static void gen6_update_ring_freq(struct drm_device *dev)
 	for (gpu_freq = dev_priv->rps.max_delay; gpu_freq >= dev_priv->rps.min_delay;
 	     gpu_freq--) {
 		int diff = dev_priv->rps.max_delay - gpu_freq;
-
-		/*
-		 * For GPU frequencies less than 750MHz, just use the lowest
-		 * ring freq.
-		 */
-		if (gpu_freq < min_freq)
-			ia_freq = 800;
-		else
-			ia_freq = max_ia_freq - ((diff * scaling_factor) / 2);
-		ia_freq = DIV_ROUND_CLOSEST(ia_freq, 100);
-		ia_freq <<= GEN6_PCODE_FREQ_IA_RATIO_SHIFT;
+		unsigned int ia_freq = 0, ring_freq = 0;
+
+		if (IS_HASWELL(dev)) {
+			ring_freq = (gpu_freq * 5 + 3) / 4;
+			ring_freq = max(min_ring_freq, ring_freq);
+			/* leave ia_freq as the default, chosen by cpufreq */
+		} else {
+			/* On older processors, there is no separate ring
+			 * clock domain, so in order to boost the bandwidth
+			 * of the ring, we need to upclock the CPU (ia_freq).
+			 *
+			 * For GPU frequencies less than 750MHz,
+			 * just use the lowest ring freq.
+			 */
+			if (gpu_freq < min_freq)
+				ia_freq = 800;
+			else
+				ia_freq = max_ia_freq - ((diff * scaling_factor) / 2);
+			ia_freq = DIV_ROUND_CLOSEST(ia_freq, 100);
+		}
 
 		sandybridge_pcode_write(dev_priv,
 					GEN6_PCODE_WRITE_MIN_FREQ_TABLE,
-					ia_freq | gpu_freq);
+					ia_freq << GEN6_PCODE_FREQ_IA_RATIO_SHIFT |
+					ring_freq << GEN6_PCODE_FREQ_RING_RATIO_SHIFT |
+					gpu_freq);
 	}
 }
 
@@ -2821,7 +2849,7 @@ static void ironlake_enable_rc6(struct drm_device *dev)
 	ret = intel_ring_idle(ring);
 	dev_priv->mm.interruptible = was_interruptible;
 	if (ret) {
-		DRM_ERROR("failed to enable ironlake power power savings\n");
+		DRM_ERROR("failed to enable ironlake power savings\n");
 		ironlake_teardown_rc6(dev);
 		return;
 	}
@@ -3562,6 +3590,7 @@ static void cpt_init_clock_gating(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	int pipe;
+	uint32_t val;
 
 	/*
 	 * On Ibex Peak and Cougar Point, we need to disable clock
@@ -3574,8 +3603,17 @@ static void cpt_init_clock_gating(struct drm_device *dev)
 	/* The below fixes the weird display corruption, a few pixels shifted
 	 * downward, on (only) LVDS of some HP laptops with IVY.
 	 */
-	for_each_pipe(pipe)
-		I915_WRITE(TRANS_CHICKEN2(pipe), TRANS_CHICKEN2_TIMING_OVERRIDE);
+	for_each_pipe(pipe) {
+		val = I915_READ(TRANS_CHICKEN2(pipe));
+		val |= TRANS_CHICKEN2_TIMING_OVERRIDE;
+		val &= ~TRANS_CHICKEN2_FDI_POLARITY_REVERSED;
+		if (dev_priv->fdi_rx_polarity_inverted)
+			val |= TRANS_CHICKEN2_FDI_POLARITY_REVERSED;
+		val &= ~TRANS_CHICKEN2_FRAME_START_DELAY_MASK;
+		val &= ~TRANS_CHICKEN2_DISABLE_DEEP_COLOR_COUNTER;
+		val &= ~TRANS_CHICKEN2_DISABLE_DEEP_COLOR_MODESWITCH;
+		I915_WRITE(TRANS_CHICKEN2(pipe), val);
+	}
 	/* WADP0ClockGatingDisable */
 	for_each_pipe(pipe) {
 		I915_WRITE(TRANS_CHICKEN1(pipe),
@@ -3768,6 +3806,9 @@ static void haswell_init_clock_gating(struct drm_device *dev)
 	I915_WRITE(GEN6_MBCTL, I915_READ(GEN6_MBCTL) |
 		   GEN6_MBCTL_ENABLE_BOOT_FETCH);
 
+	/* WaSwitchSolVfFArbitrationPriority */
+	I915_WRITE(GAM_ECOCHK, I915_READ(GAM_ECOCHK) | HSW_ECOCHK_ARB_PRIO_SOL);
+
 	/* XXX: This is a workaround for early silicon revisions and should be
 	 * removed later.
 	 */
@@ -3874,7 +3915,8 @@ static void ivybridge_init_clock_gating(struct drm_device *dev)
 	snpcr |= GEN6_MBC_SNPCR_MED;
 	I915_WRITE(GEN6_MBCUNIT_SNPCR, snpcr);
 
-	cpt_init_clock_gating(dev);
+	if (!HAS_PCH_NOP(dev))
+		cpt_init_clock_gating(dev);
 
 	gen6_check_mch_setup(dev);
 }
@@ -3899,8 +3941,10 @@ static void valleyview_init_clock_gating(struct drm_device *dev)
 		   CHICKEN3_DGMG_REQ_OUT_FIX_DISABLE |
 		   CHICKEN3_DGMG_DONE_FIX_DISABLE);
 
+	/* WaDisablePSDDualDispatchEnable */
 	I915_WRITE(GEN7_HALF_SLICE_CHICKEN1,
-		   _MASKED_BIT_ENABLE(GEN7_PSD_SINGLE_PORT_DISPATCH_ENABLE));
+		   _MASKED_BIT_ENABLE(GEN7_MAX_PS_THREAD_DEP |
+				      GEN7_PSD_SINGLE_PORT_DISPATCH_ENABLE));
 
 	/* Apply the WaDisableRHWOOptimizationForRenderHang workaround. */
 	I915_WRITE(GEN7_COMMON_SLICE_CHICKEN1,
@@ -3968,24 +4012,20 @@ static void valleyview_init_clock_gating(struct drm_device *dev)
 		   _MASKED_BIT_ENABLE(PIXEL_SUBSPAN_COLLECT_OPT_DISABLE));
 
 	/*
-	 * On ValleyView, the GUnit needs to signal the GT
-	 * when flip and other events complete.  So enable
-	 * all the GUnit->GT interrupts here
-	 */
-	I915_WRITE(VLV_DPFLIPSTAT, PIPEB_LINE_COMPARE_INT_EN |
-		   PIPEB_HLINE_INT_EN | PIPEB_VBLANK_INT_EN |
-		   SPRITED_FLIPDONE_INT_EN | SPRITEC_FLIPDONE_INT_EN |
-		   PLANEB_FLIPDONE_INT_EN | PIPEA_LINE_COMPARE_INT_EN |
-		   PIPEA_HLINE_INT_EN | PIPEA_VBLANK_INT_EN |
-		   SPRITEB_FLIPDONE_INT_EN | SPRITEA_FLIPDONE_INT_EN |
-		   PLANEA_FLIPDONE_INT_EN);
-
-	/*
 	 * WaDisableVLVClockGating_VBIIssue
 	 * Disable clock gating on th GCFG unit to prevent a delay
 	 * in the reporting of vblank events.
 	 */
-	I915_WRITE(VLV_GUNIT_CLOCK_GATE, GCFG_DIS);
+	I915_WRITE(VLV_GUNIT_CLOCK_GATE, 0xffffffff);
+
+	/* Conservative clock gating settings for now */
+	I915_WRITE(0x9400, 0xffffffff);
+	I915_WRITE(0x9404, 0xffffffff);
+	I915_WRITE(0x9408, 0xffffffff);
+	I915_WRITE(0x940c, 0xffffffff);
+	I915_WRITE(0x9410, 0xffffffff);
+	I915_WRITE(0x9414, 0xffffffff);
+	I915_WRITE(0x9418, 0xffffffff);
 }
 
 static void g4x_init_clock_gating(struct drm_device *dev)
@@ -4070,13 +4110,29 @@ void intel_init_clock_gating(struct drm_device *dev)
 	dev_priv->display.init_clock_gating(dev);
 }
 
+/**
+ * We should only use the power well if we explicitly asked the hardware to
+ * enable it, so check if it's enabled and also check if we've requested it to
+ * be enabled.
+ */
+bool intel_using_power_well(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+
+	if (IS_HASWELL(dev))
+		return I915_READ(HSW_PWR_WELL_DRIVER) ==
+		       (HSW_PWR_WELL_ENABLE | HSW_PWR_WELL_STATE);
+	else
+		return true;
+}
+
 void intel_set_power_well(struct drm_device *dev, bool enable)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	bool is_enabled, enable_requested;
 	uint32_t tmp;
 
-	if (!IS_HASWELL(dev))
+	if (!HAS_POWER_WELL(dev))
 		return;
 
 	if (!i915_disable_power_well && !enable)
@@ -4114,7 +4170,7 @@ void intel_init_power_well(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 
-	if (!IS_HASWELL(dev))
+	if (!HAS_POWER_WELL(dev))
 		return;
 
 	/* For now, we need the power well to be always enabled. */
@@ -4176,7 +4232,6 @@ void intel_init_pm(struct drm_device *dev)
 			}
 			dev_priv->display.init_clock_gating = gen6_init_clock_gating;
 		} else if (IS_IVYBRIDGE(dev)) {
-			/* FIXME: detect B0+ stepping and use auto training */
 			if (SNB_READ_WM0_LATENCY()) {
 				dev_priv->display.update_wm = ivybridge_update_wm;
 				dev_priv->display.update_sprite_wm = sandybridge_update_sprite_wm;
@@ -4274,21 +4329,14 @@ static void __gen6_gt_force_wake_reset(struct drm_i915_private *dev_priv)
 
 static void __gen6_gt_force_wake_get(struct drm_i915_private *dev_priv)
 {
-	u32 forcewake_ack;
-
-	if (IS_HASWELL(dev_priv->dev))
-		forcewake_ack = FORCEWAKE_ACK_HSW;
-	else
-		forcewake_ack = FORCEWAKE_ACK;
-
-	if (wait_for_atomic((I915_READ_NOTRACE(forcewake_ack) & 1) == 0,
+	if (wait_for_atomic((I915_READ_NOTRACE(FORCEWAKE_ACK) & 1) == 0,
 			    FORCEWAKE_ACK_TIMEOUT_MS))
 		DRM_ERROR("Timed out waiting for forcewake old ack to clear.\n");
 
-	I915_WRITE_NOTRACE(FORCEWAKE, FORCEWAKE_KERNEL);
+	I915_WRITE_NOTRACE(FORCEWAKE, 1);
 	POSTING_READ(ECOBUS); /* something from same cacheline, but !FORCEWAKE */
 
-	if (wait_for_atomic((I915_READ_NOTRACE(forcewake_ack) & 1),
+	if (wait_for_atomic((I915_READ_NOTRACE(FORCEWAKE_ACK) & 1),
 			    FORCEWAKE_ACK_TIMEOUT_MS))
 		DRM_ERROR("Timed out waiting for forcewake to ack request.\n");
 
@@ -4311,7 +4359,7 @@ static void __gen6_gt_force_wake_mt_get(struct drm_i915_private *dev_priv)
 	else
 		forcewake_ack = FORCEWAKE_MT_ACK;
 
-	if (wait_for_atomic((I915_READ_NOTRACE(forcewake_ack) & 1) == 0,
+	if (wait_for_atomic((I915_READ_NOTRACE(forcewake_ack) & FORCEWAKE_KERNEL) == 0,
 			    FORCEWAKE_ACK_TIMEOUT_MS))
 		DRM_ERROR("Timed out waiting for forcewake old ack to clear.\n");
 
@@ -4319,7 +4367,7 @@ static void __gen6_gt_force_wake_mt_get(struct drm_i915_private *dev_priv)
 	/* something from same cacheline, but !FORCEWAKE_MT */
 	POSTING_READ(ECOBUS);
 
-	if (wait_for_atomic((I915_READ_NOTRACE(forcewake_ack) & 1),
+	if (wait_for_atomic((I915_READ_NOTRACE(forcewake_ack) & FORCEWAKE_KERNEL),
 			    FORCEWAKE_ACK_TIMEOUT_MS))
 		DRM_ERROR("Timed out waiting for forcewake to ack request.\n");
 
@@ -4409,15 +4457,22 @@ static void vlv_force_wake_reset(struct drm_i915_private *dev_priv)
 
 static void vlv_force_wake_get(struct drm_i915_private *dev_priv)
 {
-	if (wait_for_atomic((I915_READ_NOTRACE(FORCEWAKE_ACK_VLV) & 1) == 0,
+	if (wait_for_atomic((I915_READ_NOTRACE(FORCEWAKE_ACK_VLV) & FORCEWAKE_KERNEL) == 0,
 			    FORCEWAKE_ACK_TIMEOUT_MS))
 		DRM_ERROR("Timed out waiting for forcewake old ack to clear.\n");
 
 	I915_WRITE_NOTRACE(FORCEWAKE_VLV, _MASKED_BIT_ENABLE(FORCEWAKE_KERNEL));
+	I915_WRITE_NOTRACE(FORCEWAKE_MEDIA_VLV,
+			   _MASKED_BIT_ENABLE(FORCEWAKE_KERNEL));
 
-	if (wait_for_atomic((I915_READ_NOTRACE(FORCEWAKE_ACK_VLV) & 1),
+	if (wait_for_atomic((I915_READ_NOTRACE(FORCEWAKE_ACK_VLV) & FORCEWAKE_KERNEL),
 			    FORCEWAKE_ACK_TIMEOUT_MS))
-		DRM_ERROR("Timed out waiting for forcewake to ack request.\n");
+		DRM_ERROR("Timed out waiting for GT to ack forcewake request.\n");
+
+	if (wait_for_atomic((I915_READ_NOTRACE(FORCEWAKE_ACK_MEDIA_VLV) &
+			     FORCEWAKE_KERNEL),
+			    FORCEWAKE_ACK_TIMEOUT_MS))
+		DRM_ERROR("Timed out waiting for media to ack forcewake request.\n");
 
 	__gen6_gt_wait_for_thread_c0(dev_priv);
 }
@@ -4425,8 +4480,9 @@ static void vlv_force_wake_get(struct drm_i915_private *dev_priv)
 static void vlv_force_wake_put(struct drm_i915_private *dev_priv)
 {
 	I915_WRITE_NOTRACE(FORCEWAKE_VLV, _MASKED_BIT_DISABLE(FORCEWAKE_KERNEL));
-	/* something from same cacheline, but !FORCEWAKE_VLV */
-	POSTING_READ(FORCEWAKE_ACK_VLV);
+	I915_WRITE_NOTRACE(FORCEWAKE_MEDIA_VLV,
+			   _MASKED_BIT_DISABLE(FORCEWAKE_KERNEL));
+	/* The below doubles as a POSTING_READ */
 	gen6_gt_check_fifodbg(dev_priv);
 }
 
@@ -4511,3 +4567,56 @@ int sandybridge_pcode_write(struct drm_i915_private *dev_priv, u8 mbox, u32 val)
 
 	return 0;
 }
+
+static int vlv_punit_rw(struct drm_i915_private *dev_priv, u8 opcode,
+			u8 addr, u32 *val)
+{
+	u32 cmd, devfn, port, be, bar;
+
+	bar = 0;
+	be = 0xf;
+	port = IOSF_PORT_PUNIT;
+	devfn = PCI_DEVFN(2, 0);
+
+	cmd = (devfn << IOSF_DEVFN_SHIFT) | (opcode << IOSF_OPCODE_SHIFT) |
+		(port << IOSF_PORT_SHIFT) | (be << IOSF_BYTE_ENABLES_SHIFT) |
+		(bar << IOSF_BAR_SHIFT);
+
+	WARN_ON(!mutex_is_locked(&dev_priv->rps.hw_lock));
+
+	if (I915_READ(VLV_IOSF_DOORBELL_REQ) & IOSF_SB_BUSY) {
+		DRM_DEBUG_DRIVER("warning: pcode (%s) mailbox access failed\n",
+				 opcode == PUNIT_OPCODE_REG_READ ?
+				 "read" : "write");
+		return -EAGAIN;
+	}
+
+	I915_WRITE(VLV_IOSF_ADDR, addr);
+	if (opcode == PUNIT_OPCODE_REG_WRITE)
+		I915_WRITE(VLV_IOSF_DATA, *val);
+	I915_WRITE(VLV_IOSF_DOORBELL_REQ, cmd);
+
+	if (wait_for((I915_READ(VLV_IOSF_DOORBELL_REQ) & IOSF_SB_BUSY) == 0,
+		     500)) {
+		DRM_ERROR("timeout waiting for pcode %s (%d) to finish\n",
+			  opcode == PUNIT_OPCODE_REG_READ ? "read" : "write",
+			  addr);
+		return -ETIMEDOUT;
+	}
+
+	if (opcode == PUNIT_OPCODE_REG_READ)
+		*val = I915_READ(VLV_IOSF_DATA);
+	I915_WRITE(VLV_IOSF_DATA, 0);
+
+	return 0;
+}
+
+int valleyview_punit_read(struct drm_i915_private *dev_priv, u8 addr, u32 *val)
+{
+	return vlv_punit_rw(dev_priv, PUNIT_OPCODE_REG_READ, addr, val);
+}
+
+int valleyview_punit_write(struct drm_i915_private *dev_priv, u8 addr, u32 val)
+{
+	return vlv_punit_rw(dev_priv, PUNIT_OPCODE_REG_WRITE, addr, &val);
+}
diff --git a/drivers/gpu/drm/i915/intel_sdvo.c b/drivers/gpu/drm/i915/intel_sdvo.c
index 78413ec623c9..d15428404b9a 100644
--- a/drivers/gpu/drm/i915/intel_sdvo.c
+++ b/drivers/gpu/drm/i915/intel_sdvo.c
@@ -246,11 +246,11 @@ static void intel_sdvo_write_sdvox(struct intel_sdvo *intel_sdvo, u32 val)
 		return;
 	}
 
-	if (intel_sdvo->sdvo_reg == SDVOB) {
-		cval = I915_READ(SDVOC);
-	} else {
-		bval = I915_READ(SDVOB);
-	}
+	if (intel_sdvo->sdvo_reg == GEN3_SDVOB)
+		cval = I915_READ(GEN3_SDVOC);
+	else
+		bval = I915_READ(GEN3_SDVOB);
+
 	/*
 	 * Write the registers twice for luck. Sometimes,
 	 * writing them only once doesn't appear to 'stick'.
@@ -258,10 +258,10 @@ static void intel_sdvo_write_sdvox(struct intel_sdvo *intel_sdvo, u32 val)
 	 */
 	for (i = 0; i < 2; i++)
 	{
-		I915_WRITE(SDVOB, bval);
-		I915_READ(SDVOB);
-		I915_WRITE(SDVOC, cval);
-		I915_READ(SDVOC);
+		I915_WRITE(GEN3_SDVOB, bval);
+		I915_READ(GEN3_SDVOB);
+		I915_WRITE(GEN3_SDVOC, cval);
+		I915_READ(GEN3_SDVOC);
 	}
 }
 
@@ -788,7 +788,6 @@ static void intel_sdvo_get_dtd_from_mode(struct intel_sdvo_dtd *dtd,
 	v_sync_offset = mode->vsync_start - mode->vdisplay;
 
 	mode_clock = mode->clock;
-	mode_clock /= intel_mode_get_pixel_multiplier(mode) ?: 1;
 	mode_clock /= 10;
 	dtd->part1.clock = mode_clock;
 
@@ -957,14 +956,17 @@ static bool intel_sdvo_set_avi_infoframe(struct intel_sdvo *intel_sdvo,
 		.len = DIP_LEN_AVI,
 	};
 	uint8_t sdvo_data[4 + sizeof(avi_if.body.avi)];
+	struct intel_crtc *intel_crtc = to_intel_crtc(intel_sdvo->base.base.crtc);
 
 	if (intel_sdvo->rgb_quant_range_selectable) {
-		if (adjusted_mode->private_flags & INTEL_MODE_LIMITED_COLOR_RANGE)
+		if (intel_crtc->config.limited_color_range)
 			avi_if.body.avi.ITC_EC_Q_SC |= DIP_AVI_RGB_QUANT_RANGE_LIMITED;
 		else
 			avi_if.body.avi.ITC_EC_Q_SC |= DIP_AVI_RGB_QUANT_RANGE_FULL;
 	}
 
+	avi_if.body.avi.VIC = drm_match_cea_mode(adjusted_mode);
+
 	intel_dip_infoframe_csum(&avi_if);
 
 	/* sdvo spec says that the ecc is handled by the hw, and it looks like
@@ -1039,12 +1041,18 @@ intel_sdvo_get_preferred_input_mode(struct intel_sdvo *intel_sdvo,
 	return true;
 }
 
-static bool intel_sdvo_mode_fixup(struct drm_encoder *encoder,
-				  const struct drm_display_mode *mode,
-				  struct drm_display_mode *adjusted_mode)
+static bool intel_sdvo_compute_config(struct intel_encoder *encoder,
+				      struct intel_crtc_config *pipe_config)
 {
-	struct intel_sdvo *intel_sdvo = to_intel_sdvo(encoder);
-	int multiplier;
+	struct intel_sdvo *intel_sdvo = to_intel_sdvo(&encoder->base);
+	struct drm_display_mode *adjusted_mode = &pipe_config->adjusted_mode;
+	struct drm_display_mode *mode = &pipe_config->requested_mode;
+
+	DRM_DEBUG_KMS("forcing bpc to 8 for SDVO\n");
+	pipe_config->pipe_bpp = 8*3;
+
+	if (HAS_PCH_SPLIT(encoder->base.dev))
+		pipe_config->has_pch_encoder = true;
 
 	/* We need to construct preferred input timings based on our
 	 * output timings.  To do that, we have to set the output
@@ -1071,37 +1079,40 @@ static bool intel_sdvo_mode_fixup(struct drm_encoder *encoder,
 	/* Make the CRTC code factor in the SDVO pixel multiplier.  The
 	 * SDVO device will factor out the multiplier during mode_set.
 	 */
-	multiplier = intel_sdvo_get_pixel_multiplier(adjusted_mode);
-	intel_mode_set_pixel_multiplier(adjusted_mode, multiplier);
+	pipe_config->pixel_multiplier =
+		intel_sdvo_get_pixel_multiplier(adjusted_mode);
+	adjusted_mode->clock *= pipe_config->pixel_multiplier;
 
 	if (intel_sdvo->color_range_auto) {
 		/* See CEA-861-E - 5.1 Default Encoding Parameters */
+		/* FIXME: This bit is only valid when using TMDS encoding and 8
+		 * bit per color mode. */
 		if (intel_sdvo->has_hdmi_monitor &&
 		    drm_match_cea_mode(adjusted_mode) > 1)
-			intel_sdvo->color_range = SDVO_COLOR_RANGE_16_235;
+			intel_sdvo->color_range = HDMI_COLOR_RANGE_16_235;
 		else
 			intel_sdvo->color_range = 0;
 	}
 
 	if (intel_sdvo->color_range)
-		adjusted_mode->private_flags |= INTEL_MODE_LIMITED_COLOR_RANGE;
+		pipe_config->limited_color_range = true;
 
 	return true;
 }
 
-static void intel_sdvo_mode_set(struct drm_encoder *encoder,
-				struct drm_display_mode *mode,
-				struct drm_display_mode *adjusted_mode)
+static void intel_sdvo_mode_set(struct intel_encoder *intel_encoder)
 {
-	struct drm_device *dev = encoder->dev;
+	struct drm_device *dev = intel_encoder->base.dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct drm_crtc *crtc = encoder->crtc;
+	struct drm_crtc *crtc = intel_encoder->base.crtc;
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
-	struct intel_sdvo *intel_sdvo = to_intel_sdvo(encoder);
+	struct drm_display_mode *adjusted_mode =
+		&intel_crtc->config.adjusted_mode;
+	struct drm_display_mode *mode = &intel_crtc->config.requested_mode;
+	struct intel_sdvo *intel_sdvo = to_intel_sdvo(&intel_encoder->base);
 	u32 sdvox;
 	struct intel_sdvo_in_out_map in_out;
 	struct intel_sdvo_dtd input_dtd, output_dtd;
-	int pixel_multiplier = intel_mode_get_pixel_multiplier(adjusted_mode);
 	int rate;
 
 	if (!mode)
@@ -1161,7 +1172,7 @@ static void intel_sdvo_mode_set(struct drm_encoder *encoder,
 		DRM_INFO("Setting input timings on %s failed\n",
 			 SDVO_NAME(intel_sdvo));
 
-	switch (pixel_multiplier) {
+	switch (intel_crtc->config.pixel_multiplier) {
 	default:
 	case 1: rate = SDVO_CLOCK_RATE_MULT_1X; break;
 	case 2: rate = SDVO_CLOCK_RATE_MULT_2X; break;
@@ -1182,10 +1193,10 @@ static void intel_sdvo_mode_set(struct drm_encoder *encoder,
 	} else {
 		sdvox = I915_READ(intel_sdvo->sdvo_reg);
 		switch (intel_sdvo->sdvo_reg) {
-		case SDVOB:
+		case GEN3_SDVOB:
 			sdvox &= SDVOB_PRESERVE_MASK;
 			break;
-		case SDVOC:
+		case GEN3_SDVOC:
 			sdvox &= SDVOC_PRESERVE_MASK;
 			break;
 		}
@@ -1193,9 +1204,9 @@ static void intel_sdvo_mode_set(struct drm_encoder *encoder,
 	}
 
 	if (INTEL_PCH_TYPE(dev) >= PCH_CPT)
-		sdvox |= TRANSCODER_CPT(intel_crtc->pipe);
+		sdvox |= SDVO_PIPE_SEL_CPT(intel_crtc->pipe);
 	else
-		sdvox |= TRANSCODER(intel_crtc->pipe);
+		sdvox |= SDVO_PIPE_SEL(intel_crtc->pipe);
 
 	if (intel_sdvo->has_hdmi_audio)
 		sdvox |= SDVO_AUDIO_ENABLE;
@@ -1205,7 +1216,8 @@ static void intel_sdvo_mode_set(struct drm_encoder *encoder,
 	} else if (IS_I945G(dev) || IS_I945GM(dev) || IS_G33(dev)) {
 		/* done in crtc_mode_set as it lives inside the dpll register */
 	} else {
-		sdvox |= (pixel_multiplier - 1) << SDVO_PORT_MULTIPLY_SHIFT;
+		sdvox |= (intel_crtc->config.pixel_multiplier - 1)
+			<< SDVO_PORT_MULTIPLY_SHIFT;
 	}
 
 	if (input_dtd.part2.sdvo_flags & SDVO_NEED_TO_STALL &&
@@ -1235,11 +1247,13 @@ static bool intel_sdvo_get_hw_state(struct intel_encoder *encoder,
 	struct drm_device *dev = encoder->base.dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_sdvo *intel_sdvo = to_intel_sdvo(&encoder->base);
+	u16 active_outputs;
 	u32 tmp;
 
 	tmp = I915_READ(intel_sdvo->sdvo_reg);
+	intel_sdvo_get_active_outputs(intel_sdvo, &active_outputs);
 
-	if (!(tmp & SDVO_ENABLE))
+	if (!(tmp & SDVO_ENABLE) && (active_outputs == 0))
 		return false;
 
 	if (HAS_PCH_CPT(dev))
@@ -1305,15 +1319,9 @@ static void intel_enable_sdvo(struct intel_encoder *encoder)
 	temp = I915_READ(intel_sdvo->sdvo_reg);
 	if ((temp & SDVO_ENABLE) == 0) {
 		/* HW workaround for IBX, we need to move the port
-		 * to transcoder A before disabling it. */
-		if (HAS_PCH_IBX(dev)) {
-			struct drm_crtc *crtc = encoder->base.crtc;
-			int pipe = crtc ? to_intel_crtc(crtc)->pipe : -1;
-
-			/* Restore the transcoder select bit. */
-			if (pipe == PIPE_B)
-				temp |= SDVO_PIPE_B_SELECT;
-		}
+		 * to transcoder A before disabling it, so restore it here. */
+		if (HAS_PCH_IBX(dev))
+			temp |= SDVO_PIPE_SEL(intel_crtc->pipe);
 
 		intel_sdvo_write_sdvox(intel_sdvo, temp | SDVO_ENABLE);
 	}
@@ -1922,6 +1930,9 @@ intel_sdvo_set_property(struct drm_connector *connector,
 	}
 
 	if (property == dev_priv->broadcast_rgb_property) {
+		bool old_auto = intel_sdvo->color_range_auto;
+		uint32_t old_range = intel_sdvo->color_range;
+
 		switch (val) {
 		case INTEL_BROADCAST_RGB_AUTO:
 			intel_sdvo->color_range_auto = true;
@@ -1932,11 +1943,18 @@ intel_sdvo_set_property(struct drm_connector *connector,
 			break;
 		case INTEL_BROADCAST_RGB_LIMITED:
 			intel_sdvo->color_range_auto = false;
-			intel_sdvo->color_range = SDVO_COLOR_RANGE_16_235;
+			/* FIXME: this bit is only valid when using TMDS
+			 * encoding and 8 bit per color mode. */
+			intel_sdvo->color_range = HDMI_COLOR_RANGE_16_235;
 			break;
 		default:
 			return -EINVAL;
 		}
+
+		if (old_auto == intel_sdvo->color_range_auto &&
+		    old_range == intel_sdvo->color_range)
+			return 0;
+
 		goto done;
 	}
 
@@ -2040,11 +2058,6 @@ done:
 #undef CHECK_PROPERTY
 }
 
-static const struct drm_encoder_helper_funcs intel_sdvo_helper_funcs = {
-	.mode_fixup = intel_sdvo_mode_fixup,
-	.mode_set = intel_sdvo_mode_set,
-};
-
 static const struct drm_connector_funcs intel_sdvo_connector_funcs = {
 	.dpms = intel_sdvo_dpms,
 	.detect = intel_sdvo_detect,
@@ -2269,7 +2282,6 @@ intel_sdvo_dvi_init(struct intel_sdvo *intel_sdvo, int device)
 	connector = &intel_connector->base;
 	if (intel_sdvo_get_hotplug_support(intel_sdvo) &
 		intel_sdvo_connector->output_flag) {
-		connector->polled = DRM_CONNECTOR_POLL_HPD;
 		intel_sdvo->hotplug_active |= intel_sdvo_connector->output_flag;
 		/* Some SDVO devices have one-shot hotplug interrupts.
 		 * Ensure that they get re-enabled when an interrupt happens.
@@ -2277,7 +2289,7 @@ intel_sdvo_dvi_init(struct intel_sdvo *intel_sdvo, int device)
 		intel_encoder->hot_plug = intel_sdvo_enable_hotplug;
 		intel_sdvo_enable_hotplug(intel_encoder);
 	} else {
-		connector->polled = DRM_CONNECTOR_POLL_CONNECT | DRM_CONNECTOR_POLL_DISCONNECT;
+		intel_connector->polled = DRM_CONNECTOR_POLL_CONNECT | DRM_CONNECTOR_POLL_DISCONNECT;
 	}
 	encoder->encoder_type = DRM_MODE_ENCODER_TMDS;
 	connector->connector_type = DRM_MODE_CONNECTOR_DVID;
@@ -2346,7 +2358,7 @@ intel_sdvo_analog_init(struct intel_sdvo *intel_sdvo, int device)
 
 	intel_connector = &intel_sdvo_connector->base;
 	connector = &intel_connector->base;
-	connector->polled = DRM_CONNECTOR_POLL_CONNECT;
+	intel_connector->polled = DRM_CONNECTOR_POLL_CONNECT;
 	encoder->encoder_type = DRM_MODE_ENCODER_DAC;
 	connector->connector_type = DRM_MODE_CONNECTOR_VGA;
 
@@ -2739,7 +2751,6 @@ bool intel_sdvo_init(struct drm_device *dev, uint32_t sdvo_reg, bool is_sdvob)
 	struct intel_sdvo *intel_sdvo;
 	u32 hotplug_mask;
 	int i;
-
 	intel_sdvo = kzalloc(sizeof(struct intel_sdvo), GFP_KERNEL);
 	if (!intel_sdvo)
 		return false;
@@ -2779,9 +2790,15 @@ bool intel_sdvo_init(struct drm_device *dev, uint32_t sdvo_reg, bool is_sdvob)
 			SDVOB_HOTPLUG_INT_STATUS_I915 : SDVOC_HOTPLUG_INT_STATUS_I915;
 	}
 
-	drm_encoder_helper_add(&intel_encoder->base, &intel_sdvo_helper_funcs);
+	/* Only enable the hotplug irq if we need it, to work around noisy
+	 * hotplug lines.
+	 */
+	if (intel_sdvo->hotplug_active)
+		intel_encoder->hpd_pin = HPD_SDVO_B ? HPD_SDVO_B : HPD_SDVO_C;
 
+	intel_encoder->compute_config = intel_sdvo_compute_config;
 	intel_encoder->disable = intel_disable_sdvo;
+	intel_encoder->mode_set = intel_sdvo_mode_set;
 	intel_encoder->enable = intel_enable_sdvo;
 	intel_encoder->get_hw_state = intel_sdvo_get_hw_state;
 
@@ -2807,12 +2824,6 @@ bool intel_sdvo_init(struct drm_device *dev, uint32_t sdvo_reg, bool is_sdvob)
 	 */
 	intel_sdvo->base.cloneable = false;
 
-	/* Only enable the hotplug irq if we need it, to work around noisy
-	 * hotplug lines.
-	 */
-	if (intel_sdvo->hotplug_active)
-		dev_priv->hotplug_supported_mask |= hotplug_mask;
-
 	intel_sdvo_select_ddc_bus(dev_priv, intel_sdvo, sdvo_reg);
 
 	/* Set the input timing to the screen. Assume always input 0. */
diff --git a/drivers/gpu/drm/i915/intel_sprite.c b/drivers/gpu/drm/i915/intel_sprite.c
index 1b6eb76beb7c..c7d25c5dd4e6 100644
--- a/drivers/gpu/drm/i915/intel_sprite.c
+++ b/drivers/gpu/drm/i915/intel_sprite.c
@@ -37,6 +37,174 @@
 #include "i915_drv.h"
 
 static void
+vlv_update_plane(struct drm_plane *dplane, struct drm_framebuffer *fb,
+		 struct drm_i915_gem_object *obj, int crtc_x, int crtc_y,
+		 unsigned int crtc_w, unsigned int crtc_h,
+		 uint32_t x, uint32_t y,
+		 uint32_t src_w, uint32_t src_h)
+{
+	struct drm_device *dev = dplane->dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct intel_plane *intel_plane = to_intel_plane(dplane);
+	int pipe = intel_plane->pipe;
+	int plane = intel_plane->plane;
+	u32 sprctl;
+	unsigned long sprsurf_offset, linear_offset;
+	int pixel_size = drm_format_plane_cpp(fb->pixel_format, 0);
+
+	sprctl = I915_READ(SPCNTR(pipe, plane));
+
+	/* Mask out pixel format bits in case we change it */
+	sprctl &= ~SP_PIXFORMAT_MASK;
+	sprctl &= ~SP_YUV_BYTE_ORDER_MASK;
+	sprctl &= ~SP_TILED;
+
+	switch (fb->pixel_format) {
+	case DRM_FORMAT_YUYV:
+		sprctl |= SP_FORMAT_YUV422 | SP_YUV_ORDER_YUYV;
+		break;
+	case DRM_FORMAT_YVYU:
+		sprctl |= SP_FORMAT_YUV422 | SP_YUV_ORDER_YVYU;
+		break;
+	case DRM_FORMAT_UYVY:
+		sprctl |= SP_FORMAT_YUV422 | SP_YUV_ORDER_UYVY;
+		break;
+	case DRM_FORMAT_VYUY:
+		sprctl |= SP_FORMAT_YUV422 | SP_YUV_ORDER_VYUY;
+		break;
+	case DRM_FORMAT_RGB565:
+		sprctl |= SP_FORMAT_BGR565;
+		break;
+	case DRM_FORMAT_XRGB8888:
+		sprctl |= SP_FORMAT_BGRX8888;
+		break;
+	case DRM_FORMAT_ARGB8888:
+		sprctl |= SP_FORMAT_BGRA8888;
+		break;
+	case DRM_FORMAT_XBGR2101010:
+		sprctl |= SP_FORMAT_RGBX1010102;
+		break;
+	case DRM_FORMAT_ABGR2101010:
+		sprctl |= SP_FORMAT_RGBA1010102;
+		break;
+	case DRM_FORMAT_XBGR8888:
+		sprctl |= SP_FORMAT_RGBX8888;
+		break;
+	case DRM_FORMAT_ABGR8888:
+		sprctl |= SP_FORMAT_RGBA8888;
+		break;
+	default:
+		/*
+		 * If we get here one of the upper layers failed to filter
+		 * out the unsupported plane formats
+		 */
+		BUG();
+		break;
+	}
+
+	if (obj->tiling_mode != I915_TILING_NONE)
+		sprctl |= SP_TILED;
+
+	sprctl |= SP_ENABLE;
+
+	/* Sizes are 0 based */
+	src_w--;
+	src_h--;
+	crtc_w--;
+	crtc_h--;
+
+	intel_update_sprite_watermarks(dev, pipe, crtc_w, pixel_size);
+
+	I915_WRITE(SPSTRIDE(pipe, plane), fb->pitches[0]);
+	I915_WRITE(SPPOS(pipe, plane), (crtc_y << 16) | crtc_x);
+
+	linear_offset = y * fb->pitches[0] + x * pixel_size;
+	sprsurf_offset = intel_gen4_compute_page_offset(&x, &y,
+							obj->tiling_mode,
+							pixel_size,
+							fb->pitches[0]);
+	linear_offset -= sprsurf_offset;
+
+	if (obj->tiling_mode != I915_TILING_NONE)
+		I915_WRITE(SPTILEOFF(pipe, plane), (y << 16) | x);
+	else
+		I915_WRITE(SPLINOFF(pipe, plane), linear_offset);
+
+	I915_WRITE(SPSIZE(pipe, plane), (crtc_h << 16) | crtc_w);
+	I915_WRITE(SPCNTR(pipe, plane), sprctl);
+	I915_MODIFY_DISPBASE(SPSURF(pipe, plane), obj->gtt_offset +
+			     sprsurf_offset);
+	POSTING_READ(SPSURF(pipe, plane));
+}
+
+static void
+vlv_disable_plane(struct drm_plane *dplane)
+{
+	struct drm_device *dev = dplane->dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct intel_plane *intel_plane = to_intel_plane(dplane);
+	int pipe = intel_plane->pipe;
+	int plane = intel_plane->plane;
+
+	I915_WRITE(SPCNTR(pipe, plane), I915_READ(SPCNTR(pipe, plane)) &
+		   ~SP_ENABLE);
+	/* Activate double buffered register update */
+	I915_MODIFY_DISPBASE(SPSURF(pipe, plane), 0);
+	POSTING_READ(SPSURF(pipe, plane));
+}
+
+static int
+vlv_update_colorkey(struct drm_plane *dplane,
+		    struct drm_intel_sprite_colorkey *key)
+{
+	struct drm_device *dev = dplane->dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct intel_plane *intel_plane = to_intel_plane(dplane);
+	int pipe = intel_plane->pipe;
+	int plane = intel_plane->plane;
+	u32 sprctl;
+
+	if (key->flags & I915_SET_COLORKEY_DESTINATION)
+		return -EINVAL;
+
+	I915_WRITE(SPKEYMINVAL(pipe, plane), key->min_value);
+	I915_WRITE(SPKEYMAXVAL(pipe, plane), key->max_value);
+	I915_WRITE(SPKEYMSK(pipe, plane), key->channel_mask);
+
+	sprctl = I915_READ(SPCNTR(pipe, plane));
+	sprctl &= ~SP_SOURCE_KEY;
+	if (key->flags & I915_SET_COLORKEY_SOURCE)
+		sprctl |= SP_SOURCE_KEY;
+	I915_WRITE(SPCNTR(pipe, plane), sprctl);
+
+	POSTING_READ(SPKEYMSK(pipe, plane));
+
+	return 0;
+}
+
+static void
+vlv_get_colorkey(struct drm_plane *dplane,
+		 struct drm_intel_sprite_colorkey *key)
+{
+	struct drm_device *dev = dplane->dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct intel_plane *intel_plane = to_intel_plane(dplane);
+	int pipe = intel_plane->pipe;
+	int plane = intel_plane->plane;
+	u32 sprctl;
+
+	key->min_value = I915_READ(SPKEYMINVAL(pipe, plane));
+	key->max_value = I915_READ(SPKEYMAXVAL(pipe, plane));
+	key->channel_mask = I915_READ(SPKEYMSK(pipe, plane));
+
+	sprctl = I915_READ(SPCNTR(pipe, plane));
+	if (sprctl & SP_SOURCE_KEY)
+		key->flags = I915_SET_COLORKEY_SOURCE;
+	else
+		key->flags = I915_SET_COLORKEY_NONE;
+}
+
+static void
 ivb_update_plane(struct drm_plane *plane, struct drm_framebuffer *fb,
 		 struct drm_i915_gem_object *obj, int crtc_x, int crtc_y,
 		 unsigned int crtc_w, unsigned int crtc_h,
@@ -441,6 +609,15 @@ intel_update_plane(struct drm_plane *plane, struct drm_crtc *crtc,
 
 	old_obj = intel_plane->obj;
 
+	intel_plane->crtc_x = crtc_x;
+	intel_plane->crtc_y = crtc_y;
+	intel_plane->crtc_w = crtc_w;
+	intel_plane->crtc_h = crtc_h;
+	intel_plane->src_x = src_x;
+	intel_plane->src_y = src_y;
+	intel_plane->src_w = src_w;
+	intel_plane->src_h = src_h;
+
 	src_w = src_w >> 16;
 	src_h = src_h >> 16;
 
@@ -513,6 +690,11 @@ intel_update_plane(struct drm_plane *plane, struct drm_crtc *crtc,
 
 	mutex_lock(&dev->struct_mutex);
 
+	/* Note that this will apply the VT-d workaround for scanouts,
+	 * which is more restrictive than required for sprites. (The
+	 * primary plane requires 256KiB alignment with 64 PTE padding,
+	 * the sprite planes only require 128KiB alignment and 32 PTE padding.
+	 */
 	ret = intel_pin_and_fence_fb_obj(dev, obj, NULL);
 	if (ret)
 		goto out_unlock;
@@ -568,6 +750,8 @@ intel_disable_plane(struct drm_plane *plane)
 	if (!intel_plane->obj)
 		goto out;
 
+	intel_wait_for_vblank(dev, intel_plane->pipe);
+
 	mutex_lock(&dev->struct_mutex);
 	intel_unpin_fb_obj(intel_plane->obj);
 	intel_plane->obj = NULL;
@@ -647,6 +831,20 @@ out_unlock:
 	return ret;
 }
 
+void intel_plane_restore(struct drm_plane *plane)
+{
+	struct intel_plane *intel_plane = to_intel_plane(plane);
+
+	if (!plane->crtc || !plane->fb)
+		return;
+
+	intel_update_plane(plane, plane->crtc, plane->fb,
+			   intel_plane->crtc_x, intel_plane->crtc_y,
+			   intel_plane->crtc_w, intel_plane->crtc_h,
+			   intel_plane->src_x, intel_plane->src_y,
+			   intel_plane->src_w, intel_plane->src_h);
+}
+
 static const struct drm_plane_funcs intel_plane_funcs = {
 	.update_plane = intel_update_plane,
 	.disable_plane = intel_disable_plane,
@@ -670,8 +868,22 @@ static uint32_t snb_plane_formats[] = {
 	DRM_FORMAT_VYUY,
 };
 
+static uint32_t vlv_plane_formats[] = {
+	DRM_FORMAT_RGB565,
+	DRM_FORMAT_ABGR8888,
+	DRM_FORMAT_ARGB8888,
+	DRM_FORMAT_XBGR8888,
+	DRM_FORMAT_XRGB8888,
+	DRM_FORMAT_XBGR2101010,
+	DRM_FORMAT_ABGR2101010,
+	DRM_FORMAT_YUYV,
+	DRM_FORMAT_YVYU,
+	DRM_FORMAT_UYVY,
+	DRM_FORMAT_VYUY,
+};
+
 int
-intel_plane_init(struct drm_device *dev, enum pipe pipe)
+intel_plane_init(struct drm_device *dev, enum pipe pipe, int plane)
 {
 	struct intel_plane *intel_plane;
 	unsigned long possible_crtcs;
@@ -710,14 +922,26 @@ intel_plane_init(struct drm_device *dev, enum pipe pipe)
 			intel_plane->can_scale = false;
 		else
 			intel_plane->can_scale = true;
-		intel_plane->max_downscale = 2;
-		intel_plane->update_plane = ivb_update_plane;
-		intel_plane->disable_plane = ivb_disable_plane;
-		intel_plane->update_colorkey = ivb_update_colorkey;
-		intel_plane->get_colorkey = ivb_get_colorkey;
-
-		plane_formats = snb_plane_formats;
-		num_plane_formats = ARRAY_SIZE(snb_plane_formats);
+
+		if (IS_VALLEYVIEW(dev)) {
+			intel_plane->max_downscale = 1;
+			intel_plane->update_plane = vlv_update_plane;
+			intel_plane->disable_plane = vlv_disable_plane;
+			intel_plane->update_colorkey = vlv_update_colorkey;
+			intel_plane->get_colorkey = vlv_get_colorkey;
+
+			plane_formats = vlv_plane_formats;
+			num_plane_formats = ARRAY_SIZE(vlv_plane_formats);
+		} else {
+			intel_plane->max_downscale = 2;
+			intel_plane->update_plane = ivb_update_plane;
+			intel_plane->disable_plane = ivb_disable_plane;
+			intel_plane->update_colorkey = ivb_update_colorkey;
+			intel_plane->get_colorkey = ivb_get_colorkey;
+
+			plane_formats = snb_plane_formats;
+			num_plane_formats = ARRAY_SIZE(snb_plane_formats);
+		}
 		break;
 
 	default:
@@ -726,6 +950,7 @@ intel_plane_init(struct drm_device *dev, enum pipe pipe)
 	}
 
 	intel_plane->pipe = pipe;
+	intel_plane->plane = plane;
 	possible_crtcs = (1 << pipe);
 	ret = drm_plane_init(dev, &intel_plane->base, possible_crtcs,
 			     &intel_plane_funcs,
diff --git a/drivers/gpu/drm/i915/intel_tv.c b/drivers/gpu/drm/i915/intel_tv.c
index d808421c1c80..b945bc54207a 100644
--- a/drivers/gpu/drm/i915/intel_tv.c
+++ b/drivers/gpu/drm/i915/intel_tv.c
@@ -905,11 +905,10 @@ intel_tv_mode_valid(struct drm_connector *connector,
 
 
 static bool
-intel_tv_mode_fixup(struct drm_encoder *encoder,
-		    const struct drm_display_mode *mode,
-		    struct drm_display_mode *adjusted_mode)
+intel_tv_compute_config(struct intel_encoder *encoder,
+			struct intel_crtc_config *pipe_config)
 {
-	struct intel_tv *intel_tv = enc_to_intel_tv(encoder);
+	struct intel_tv *intel_tv = enc_to_intel_tv(&encoder->base);
 	const struct tv_mode *tv_mode = intel_tv_mode_find(intel_tv);
 
 	if (!tv_mode)
@@ -918,7 +917,10 @@ intel_tv_mode_fixup(struct drm_encoder *encoder,
 	if (intel_encoder_check_is_cloned(&intel_tv->base))
 		return false;
 
-	adjusted_mode->clock = tv_mode->clock;
+	pipe_config->adjusted_mode.clock = tv_mode->clock;
+	DRM_DEBUG_KMS("forcing bpc to 8 for TV\n");
+	pipe_config->pipe_bpp = 8*3;
+
 	return true;
 }
 
@@ -1485,7 +1487,6 @@ out:
 }
 
 static const struct drm_encoder_helper_funcs intel_tv_helper_funcs = {
-	.mode_fixup = intel_tv_mode_fixup,
 	.mode_set = intel_tv_mode_set,
 };
 
@@ -1612,7 +1613,7 @@ intel_tv_init(struct drm_device *dev)
 	 *
 	 * More recent chipsets favour HDMI rather than integrated S-Video.
 	 */
-	connector->polled = DRM_CONNECTOR_POLL_CONNECT;
+	intel_connector->polled = DRM_CONNECTOR_POLL_CONNECT;
 
 	drm_connector_init(dev, connector, &intel_tv_connector_funcs,
 			   DRM_MODE_CONNECTOR_SVIDEO);
@@ -1620,6 +1621,7 @@ intel_tv_init(struct drm_device *dev)
 	drm_encoder_init(dev, &intel_encoder->base, &intel_tv_enc_funcs,
 			 DRM_MODE_ENCODER_TVDAC);
 
+	intel_encoder->compute_config = intel_tv_compute_config;
 	intel_encoder->enable = intel_enable_tv;
 	intel_encoder->disable = intel_disable_tv;
 	intel_encoder->get_hw_state = intel_tv_get_hw_state;
diff --git a/drivers/gpu/drm/mgag200/mgag200_drv.h b/drivers/gpu/drm/mgag200/mgag200_drv.h
index 4d932c46725d..bf29b2f4d68d 100644
--- a/drivers/gpu/drm/mgag200/mgag200_drv.h
+++ b/drivers/gpu/drm/mgag200/mgag200_drv.h
@@ -115,6 +115,8 @@ struct mga_fbdev {
 	void *sysram;
 	int size;
 	struct ttm_bo_kmap_obj mapping;
+	int x1, y1, x2, y2; /* dirty rect */
+	spinlock_t dirty_lock;
 };
 
 struct mga_crtc {
@@ -215,7 +217,7 @@ mgag200_bo(struct ttm_buffer_object *bo)
 {
 	return container_of(bo, struct mgag200_bo, bo);
 }
-				/* mga_crtc.c */
+				/* mgag200_crtc.c */
 void mga_crtc_fb_gamma_set(struct drm_crtc *crtc, u16 red, u16 green,
 			     u16 blue, int regno);
 void mga_crtc_fb_gamma_get(struct drm_crtc *crtc, u16 *red, u16 *green,
@@ -225,7 +227,7 @@ void mga_crtc_fb_gamma_get(struct drm_crtc *crtc, u16 *red, u16 *green,
 int mgag200_modeset_init(struct mga_device *mdev);
 void mgag200_modeset_fini(struct mga_device *mdev);
 
-				/* mga_fbdev.c */
+				/* mgag200_fb.c */
 int mgag200_fbdev_init(struct mga_device *mdev);
 void mgag200_fbdev_fini(struct mga_device *mdev);
 
@@ -254,7 +256,7 @@ mgag200_dumb_mmap_offset(struct drm_file *file,
 			 struct drm_device *dev,
 			 uint32_t handle,
 			 uint64_t *offset);
-				/* mga_i2c.c */
+				/* mgag200_i2c.c */
 struct mga_i2c_chan *mgag200_i2c_create(struct drm_device *dev);
 void mgag200_i2c_destroy(struct mga_i2c_chan *i2c);
 
diff --git a/drivers/gpu/drm/mgag200/mgag200_fb.c b/drivers/gpu/drm/mgag200/mgag200_fb.c
index d2253f639481..5da824ce9ba1 100644
--- a/drivers/gpu/drm/mgag200/mgag200_fb.c
+++ b/drivers/gpu/drm/mgag200/mgag200_fb.c
@@ -29,16 +29,52 @@ static void mga_dirty_update(struct mga_fbdev *mfbdev,
 	int bpp = (mfbdev->mfb.base.bits_per_pixel + 7)/8;
 	int ret;
 	bool unmap = false;
+	bool store_for_later = false;
+	int x2, y2;
+	unsigned long flags;
 
 	obj = mfbdev->mfb.obj;
 	bo = gem_to_mga_bo(obj);
 
+	/*
+	 * try and reserve the BO, if we fail with busy
+	 * then the BO is being moved and we should
+	 * store up the damage until later.
+	 */
 	ret = mgag200_bo_reserve(bo, true);
 	if (ret) {
-		DRM_ERROR("failed to reserve fb bo\n");
+		if (ret != -EBUSY)
+			return;
+
+		store_for_later = true;
+	}
+
+	x2 = x + width - 1;
+	y2 = y + height - 1;
+	spin_lock_irqsave(&mfbdev->dirty_lock, flags);
+
+	if (mfbdev->y1 < y)
+		y = mfbdev->y1;
+	if (mfbdev->y2 > y2)
+		y2 = mfbdev->y2;
+	if (mfbdev->x1 < x)
+		x = mfbdev->x1;
+	if (mfbdev->x2 > x2)
+		x2 = mfbdev->x2;
+
+	if (store_for_later) {
+		mfbdev->x1 = x;
+		mfbdev->x2 = x2;
+		mfbdev->y1 = y;
+		mfbdev->y2 = y2;
+		spin_unlock_irqrestore(&mfbdev->dirty_lock, flags);
 		return;
 	}
 
+	mfbdev->x1 = mfbdev->y1 = INT_MAX;
+	mfbdev->x2 = mfbdev->y2 = 0;
+	spin_unlock_irqrestore(&mfbdev->dirty_lock, flags);
+
 	if (!bo->kmap.virtual) {
 		ret = ttm_bo_kmap(&bo->bo, 0, bo->bo.num_pages, &bo->kmap);
 		if (ret) {
@@ -48,10 +84,10 @@ static void mga_dirty_update(struct mga_fbdev *mfbdev,
 		}
 		unmap = true;
 	}
-	for (i = y; i < y + height; i++) {
+	for (i = y; i <= y2; i++) {
 		/* assume equal stride for now */
 		src_offset = dst_offset = i * mfbdev->mfb.base.pitches[0] + (x * bpp);
-		memcpy_toio(bo->kmap.virtual + src_offset, mfbdev->sysram + src_offset, width * bpp);
+		memcpy_toio(bo->kmap.virtual + src_offset, mfbdev->sysram + src_offset, (x2 - x + 1) * bpp);
 
 	}
 	if (unmap)
@@ -105,12 +141,9 @@ static int mgag200fb_create_object(struct mga_fbdev *afbdev,
 				   struct drm_gem_object **gobj_p)
 {
 	struct drm_device *dev = afbdev->helper.dev;
-	u32 bpp, depth;
 	u32 size;
 	struct drm_gem_object *gobj;
-
 	int ret = 0;
-	drm_fb_get_bpp_depth(mode_cmd->pixel_format, &depth, &bpp);
 
 	size = mode_cmd->pitches[0] * mode_cmd->height;
 	ret = mgag200_gem_create(dev, size, true, &gobj);
@@ -249,19 +282,19 @@ int mgag200_fbdev_init(struct mga_device *mdev)
 	struct mga_fbdev *mfbdev;
 	int ret;
 
-	mfbdev = kzalloc(sizeof(struct mga_fbdev), GFP_KERNEL);
+	mfbdev = devm_kzalloc(mdev->dev->dev, sizeof(struct mga_fbdev), GFP_KERNEL);
 	if (!mfbdev)
 		return -ENOMEM;
 
 	mdev->mfbdev = mfbdev;
 	mfbdev->helper.funcs = &mga_fb_helper_funcs;
+	spin_lock_init(&mfbdev->dirty_lock);
 
 	ret = drm_fb_helper_init(mdev->dev, &mfbdev->helper,
 				 mdev->num_crtc, MGAG200FB_CONN_LIMIT);
-	if (ret) {
-		kfree(mfbdev);
+	if (ret)
 		return ret;
-	}
+
 	drm_fb_helper_single_add_all_connectors(&mfbdev->helper);
 
 	/* disable all the possible outputs/crtcs before entering KMS mode */
@@ -278,6 +311,4 @@ void mgag200_fbdev_fini(struct mga_device *mdev)
 		return;
 
 	mga_fbdev_destroy(mdev->dev, mdev->mfbdev);
-	kfree(mdev->mfbdev);
-	mdev->mfbdev = NULL;
 }
diff --git a/drivers/gpu/drm/mgag200/mgag200_main.c b/drivers/gpu/drm/mgag200/mgag200_main.c
index 64297c72464f..99059237da38 100644
--- a/drivers/gpu/drm/mgag200/mgag200_main.c
+++ b/drivers/gpu/drm/mgag200/mgag200_main.c
@@ -76,15 +76,6 @@ static const struct drm_mode_config_funcs mga_mode_funcs = {
 	.fb_create = mgag200_user_framebuffer_create,
 };
 
-/* Unmap the framebuffer from the core and release the memory */
-static void mga_vram_fini(struct mga_device *mdev)
-{
-	pci_iounmap(mdev->dev->pdev, mdev->rmmio);
-	mdev->rmmio = NULL;
-	if (mdev->mc.vram_base)
-		release_mem_region(mdev->mc.vram_base, mdev->mc.vram_window);
-}
-
 static int mga_probe_vram(struct mga_device *mdev, void __iomem *mem)
 {
 	int offset;
@@ -140,7 +131,7 @@ static int mga_vram_init(struct mga_device *mdev)
 	remove_conflicting_framebuffers(aper, "mgafb", true);
 	kfree(aper);
 
-	if (!request_mem_region(mdev->mc.vram_base, mdev->mc.vram_window,
+	if (!devm_request_mem_region(mdev->dev->dev, mdev->mc.vram_base, mdev->mc.vram_window,
 				"mgadrmfb_vram")) {
 		DRM_ERROR("can't reserve VRAM\n");
 		return -ENXIO;
@@ -173,13 +164,13 @@ static int mgag200_device_init(struct drm_device *dev,
 	mdev->rmmio_base = pci_resource_start(mdev->dev->pdev, 1);
 	mdev->rmmio_size = pci_resource_len(mdev->dev->pdev, 1);
 
-	if (!request_mem_region(mdev->rmmio_base, mdev->rmmio_size,
+	if (!devm_request_mem_region(mdev->dev->dev, mdev->rmmio_base, mdev->rmmio_size,
 				"mgadrmfb_mmio")) {
 		DRM_ERROR("can't reserve mmio registers\n");
 		return -ENOMEM;
 	}
 
-	mdev->rmmio = pci_iomap(dev->pdev, 1, 0);
+	mdev->rmmio = pcim_iomap(dev->pdev, 1, 0);
 	if (mdev->rmmio == NULL)
 		return -ENOMEM;
 
@@ -188,10 +179,8 @@ static int mgag200_device_init(struct drm_device *dev,
 		mdev->reg_1e24 = RREG32(0x1e24);
 
 	ret = mga_vram_init(mdev);
-	if (ret) {
-		release_mem_region(mdev->rmmio_base, mdev->rmmio_size);
+	if (ret)
 		return ret;
-	}
 
 	mdev->bpp_shifts[0] = 0;
 	mdev->bpp_shifts[1] = 1;
@@ -200,12 +189,6 @@ static int mgag200_device_init(struct drm_device *dev,
 	return 0;
 }
 
-void mgag200_device_fini(struct mga_device *mdev)
-{
-	release_mem_region(mdev->rmmio_base, mdev->rmmio_size);
-	mga_vram_fini(mdev);
-}
-
 /*
  * Functions here will be called by the core once it's bound the driver to
  * a PCI device
@@ -217,7 +200,7 @@ int mgag200_driver_load(struct drm_device *dev, unsigned long flags)
 	struct mga_device *mdev;
 	int r;
 
-	mdev = kzalloc(sizeof(struct mga_device), GFP_KERNEL);
+	mdev = devm_kzalloc(dev->dev, sizeof(struct mga_device), GFP_KERNEL);
 	if (mdev == NULL)
 		return -ENOMEM;
 	dev->dev_private = (void *)mdev;
@@ -234,8 +217,6 @@ int mgag200_driver_load(struct drm_device *dev, unsigned long flags)
 
 	drm_mode_config_init(dev);
 	dev->mode_config.funcs = (void *)&mga_mode_funcs;
-	dev->mode_config.min_width = 0;
-	dev->mode_config.min_height = 0;
 	dev->mode_config.preferred_depth = 24;
 	dev->mode_config.prefer_shadow = 1;
 
@@ -258,8 +239,6 @@ int mgag200_driver_unload(struct drm_device *dev)
 	mgag200_fbdev_fini(mdev);
 	drm_mode_config_cleanup(dev);
 	mgag200_mm_fini(mdev);
-	mgag200_device_fini(mdev);
-	kfree(mdev);
 	dev->dev_private = NULL;
 	return 0;
 }
diff --git a/drivers/gpu/drm/mgag200/mgag200_mode.c b/drivers/gpu/drm/mgag200/mgag200_mode.c
index 78d8e919509f..f9889658329b 100644
--- a/drivers/gpu/drm/mgag200/mgag200_mode.c
+++ b/drivers/gpu/drm/mgag200/mgag200_mode.c
@@ -1254,9 +1254,8 @@ static const struct drm_crtc_helper_funcs mga_helper_funcs = {
 };
 
 /* CRTC setup */
-static void mga_crtc_init(struct drm_device *dev)
+static void mga_crtc_init(struct mga_device *mdev)
 {
-	struct mga_device *mdev = dev->dev_private;
 	struct mga_crtc *mga_crtc;
 	int i;
 
@@ -1267,7 +1266,7 @@ static void mga_crtc_init(struct drm_device *dev)
 	if (mga_crtc == NULL)
 		return;
 
-	drm_crtc_init(dev, &mga_crtc->base, &mga_crtc_funcs);
+	drm_crtc_init(mdev->dev, &mga_crtc->base, &mga_crtc_funcs);
 
 	drm_mode_crtc_set_gamma_size(&mga_crtc->base, MGAG200_LUT_SIZE);
 	mdev->mode_info.crtc = mga_crtc;
@@ -1522,7 +1521,7 @@ int mgag200_modeset_init(struct mga_device *mdev)
 
 	mdev->dev->mode_config.fb_base = mdev->mc.vram_base;
 
-	mga_crtc_init(mdev->dev);
+	mga_crtc_init(mdev);
 
 	encoder = mga_encoder_init(mdev->dev);
 	if (!encoder) {
diff --git a/drivers/gpu/drm/mgag200/mgag200_ttm.c b/drivers/gpu/drm/mgag200/mgag200_ttm.c
index 8fc9d9201945..401c9891d3a8 100644
--- a/drivers/gpu/drm/mgag200/mgag200_ttm.c
+++ b/drivers/gpu/drm/mgag200/mgag200_ttm.c
@@ -315,8 +315,8 @@ int mgag200_bo_reserve(struct mgag200_bo *bo, bool no_wait)
 
 	ret = ttm_bo_reserve(&bo->bo, true, no_wait, false, 0);
 	if (ret) {
-		if (ret != -ERESTARTSYS)
-			DRM_ERROR("reserve failed %p\n", bo);
+		if (ret != -ERESTARTSYS && ret != -EBUSY)
+			DRM_ERROR("reserve failed %p %d\n", bo, ret);
 		return ret;
 	}
 	return 0;
diff --git a/drivers/gpu/drm/nouveau/Makefile b/drivers/gpu/drm/nouveau/Makefile
index 90f9140eeefd..998e8b4444f3 100644
--- a/drivers/gpu/drm/nouveau/Makefile
+++ b/drivers/gpu/drm/nouveau/Makefile
@@ -53,15 +53,6 @@ nouveau-y += core/subdev/clock/nva3.o
 nouveau-y += core/subdev/clock/nvc0.o
 nouveau-y += core/subdev/clock/pllnv04.o
 nouveau-y += core/subdev/clock/pllnva3.o
-nouveau-y += core/subdev/device/base.o
-nouveau-y += core/subdev/device/nv04.o
-nouveau-y += core/subdev/device/nv10.o
-nouveau-y += core/subdev/device/nv20.o
-nouveau-y += core/subdev/device/nv30.o
-nouveau-y += core/subdev/device/nv40.o
-nouveau-y += core/subdev/device/nv50.o
-nouveau-y += core/subdev/device/nvc0.o
-nouveau-y += core/subdev/device/nve0.o
 nouveau-y += core/subdev/devinit/base.o
 nouveau-y += core/subdev/devinit/nv04.o
 nouveau-y += core/subdev/devinit/nv05.o
@@ -126,6 +117,7 @@ nouveau-y += core/subdev/therm/ic.o
 nouveau-y += core/subdev/therm/temp.o
 nouveau-y += core/subdev/therm/nv40.o
 nouveau-y += core/subdev/therm/nv50.o
+nouveau-y += core/subdev/therm/nv84.o
 nouveau-y += core/subdev/therm/nva3.o
 nouveau-y += core/subdev/therm/nvd0.o
 nouveau-y += core/subdev/timer/base.o
@@ -150,6 +142,15 @@ nouveau-y += core/engine/copy/nvc0.o
 nouveau-y += core/engine/copy/nve0.o
 nouveau-y += core/engine/crypt/nv84.o
 nouveau-y += core/engine/crypt/nv98.o
+nouveau-y += core/engine/device/base.o
+nouveau-y += core/engine/device/nv04.o
+nouveau-y += core/engine/device/nv10.o
+nouveau-y += core/engine/device/nv20.o
+nouveau-y += core/engine/device/nv30.o
+nouveau-y += core/engine/device/nv40.o
+nouveau-y += core/engine/device/nv50.o
+nouveau-y += core/engine/device/nvc0.o
+nouveau-y += core/engine/device/nve0.o
 nouveau-y += core/engine/disp/base.o
 nouveau-y += core/engine/disp/nv04.o
 nouveau-y += core/engine/disp/nv50.o
@@ -159,6 +160,7 @@ nouveau-y += core/engine/disp/nva0.o
 nouveau-y += core/engine/disp/nva3.o
 nouveau-y += core/engine/disp/nvd0.o
 nouveau-y += core/engine/disp/nve0.o
+nouveau-y += core/engine/disp/nvf0.o
 nouveau-y += core/engine/disp/dacnv50.o
 nouveau-y += core/engine/disp/dport.o
 nouveau-y += core/engine/disp/hdanva3.o
@@ -212,7 +214,7 @@ nouveau-y += core/engine/vp/nve0.o
 
 # drm/core
 nouveau-y += nouveau_drm.o nouveau_chan.o nouveau_dma.o nouveau_fence.o
-nouveau-y += nouveau_irq.o nouveau_vga.o nouveau_agp.o
+nouveau-y += nouveau_vga.o nouveau_agp.o
 nouveau-y += nouveau_ttm.o nouveau_sgdma.o nouveau_bo.o nouveau_gem.o
 nouveau-y += nouveau_prime.o nouveau_abi16.o
 nouveau-y += nv04_fence.o nv10_fence.o nv17_fence.o
@@ -224,9 +226,7 @@ nouveau-y += nouveau_connector.o nouveau_dp.o
 nouveau-y += nv04_fbcon.o nv50_fbcon.o nvc0_fbcon.o
 
 # drm/kms/nv04:nv50
-nouveau-y += nouveau_hw.o nouveau_calc.o
-nouveau-y += nv04_dac.o nv04_dfp.o nv04_tv.o nv17_tv.o nv17_tv_modes.o
-nouveau-y += nv04_crtc.o nv04_display.o nv04_cursor.o
+include $(src)/dispnv04/Makefile
 
 # drm/kms/nv50-
 nouveau-y += nv50_display.o
diff --git a/drivers/gpu/drm/nouveau/core/core/client.c b/drivers/gpu/drm/nouveau/core/core/client.c
index 295c22165eac..9079c0ac58e6 100644
--- a/drivers/gpu/drm/nouveau/core/core/client.c
+++ b/drivers/gpu/drm/nouveau/core/core/client.c
@@ -27,7 +27,7 @@
 #include <core/handle.h>
 #include <core/option.h>
 
-#include <subdev/device.h>
+#include <engine/device.h>
 
 static void
 nouveau_client_dtor(struct nouveau_object *object)
@@ -58,8 +58,9 @@ nouveau_client_create_(const char *name, u64 devname, const char *cfg,
 		return -ENODEV;
 
 	ret = nouveau_namedb_create_(NULL, NULL, &nouveau_client_oclass,
-				     NV_CLIENT_CLASS, nouveau_device_sclass,
-				     0, length, pobject);
+				     NV_CLIENT_CLASS, NULL,
+				     (1ULL << NVDEV_ENGINE_DEVICE),
+				     length, pobject);
 	client = *pobject;
 	if (ret)
 		return ret;
diff --git a/drivers/gpu/drm/nouveau/core/core/engine.c b/drivers/gpu/drm/nouveau/core/core/engine.c
index 09b3bd502fd0..c8bed4a26833 100644
--- a/drivers/gpu/drm/nouveau/core/core/engine.c
+++ b/drivers/gpu/drm/nouveau/core/core/engine.c
@@ -33,7 +33,6 @@ nouveau_engine_create_(struct nouveau_object *parent,
 		       const char *iname, const char *fname,
 		       int length, void **pobject)
 {
-	struct nouveau_device *device = nv_device(parent);
 	struct nouveau_engine *engine;
 	int ret;
 
@@ -43,7 +42,8 @@ nouveau_engine_create_(struct nouveau_object *parent,
 	if (ret)
 		return ret;
 
-	if (!nouveau_boolopt(device->cfgopt, iname, enable)) {
+	if ( parent &&
+	    !nouveau_boolopt(nv_device(parent)->cfgopt, iname, enable)) {
 		if (!enable)
 			nv_warn(engine, "disabled, %s=1 to enable\n", iname);
 		return -ENODEV;
diff --git a/drivers/gpu/drm/nouveau/core/core/event.c b/drivers/gpu/drm/nouveau/core/core/event.c
index 6d01e0f0fc8a..7eb81c1b6fab 100644
--- a/drivers/gpu/drm/nouveau/core/core/event.c
+++ b/drivers/gpu/drm/nouveau/core/core/event.c
@@ -27,8 +27,10 @@ static void
 nouveau_event_put_locked(struct nouveau_event *event, int index,
 			 struct nouveau_eventh *handler)
 {
-	if (!--event->index[index].refs)
-		event->disable(event, index);
+	if (!--event->index[index].refs) {
+		if (event->disable)
+			event->disable(event, index);
+	}
 	list_del(&handler->head);
 }
 
@@ -53,8 +55,10 @@ nouveau_event_get(struct nouveau_event *event, int index,
 	spin_lock_irqsave(&event->lock, flags);
 	if (index < event->index_nr) {
 		list_add(&handler->head, &event->index[index].list);
-		if (!event->index[index].refs++)
-			event->enable(event, index);
+		if (!event->index[index].refs++) {
+			if (event->enable)
+				event->enable(event, index);
+		}
 	}
 	spin_unlock_irqrestore(&event->lock, flags);
 }
diff --git a/drivers/gpu/drm/nouveau/core/core/object.c b/drivers/gpu/drm/nouveau/core/core/object.c
index 3b2e7b6304d3..7f48e288215f 100644
--- a/drivers/gpu/drm/nouveau/core/core/object.c
+++ b/drivers/gpu/drm/nouveau/core/core/object.c
@@ -136,26 +136,30 @@ nouveau_object_ctor(struct nouveau_object *parent,
 		    struct nouveau_object **pobject)
 {
 	struct nouveau_ofuncs *ofuncs = oclass->ofuncs;
+	struct nouveau_object *object = NULL;
 	int ret;
 
-	*pobject = NULL;
-
-	ret = ofuncs->ctor(parent, engine, oclass, data, size, pobject);
+	ret = ofuncs->ctor(parent, engine, oclass, data, size, &object);
+	*pobject = object;
 	if (ret < 0) {
 		if (ret != -ENODEV) {
 			nv_error(parent, "failed to create 0x%08x, %d\n",
 				 oclass->handle, ret);
 		}
 
-		if (*pobject) {
-			ofuncs->dtor(*pobject);
+		if (object) {
+			ofuncs->dtor(object);
 			*pobject = NULL;
 		}
 
 		return ret;
 	}
 
-	nv_debug(*pobject, "created\n");
+	if (ret == 0) {
+		nv_debug(object, "created\n");
+		atomic_set(&object->refcount, 1);
+	}
+
 	return 0;
 }
 
@@ -327,6 +331,7 @@ nouveau_object_inc(struct nouveau_object *object)
 	}
 
 	ret = nv_ofuncs(object)->init(object);
+	atomic_set(&object->usecount, 1);
 	if (ret) {
 		nv_error(object, "init failed, %d\n", ret);
 		goto fail_self;
@@ -357,6 +362,7 @@ nouveau_object_decf(struct nouveau_object *object)
 	nv_trace(object, "stopping...\n");
 
 	ret = nv_ofuncs(object)->fini(object, false);
+	atomic_set(&object->usecount, 0);
 	if (ret)
 		nv_warn(object, "failed fini, %d\n", ret);
 
@@ -381,6 +387,7 @@ nouveau_object_decs(struct nouveau_object *object)
 	nv_trace(object, "suspending...\n");
 
 	ret = nv_ofuncs(object)->fini(object, true);
+	atomic_set(&object->usecount, 0);
 	if (ret) {
 		nv_error(object, "failed suspend, %d\n", ret);
 		return ret;
diff --git a/drivers/gpu/drm/nouveau/core/core/parent.c b/drivers/gpu/drm/nouveau/core/core/parent.c
index db7c54943102..313380ce632d 100644
--- a/drivers/gpu/drm/nouveau/core/core/parent.c
+++ b/drivers/gpu/drm/nouveau/core/core/parent.c
@@ -24,6 +24,7 @@
 
 #include <core/object.h>
 #include <core/parent.h>
+#include <core/client.h>
 
 int
 nouveau_parent_sclass(struct nouveau_object *parent, u16 handle,
@@ -50,7 +51,12 @@ nouveau_parent_sclass(struct nouveau_object *parent, u16 handle,
 	while (mask) {
 		int i = ffsll(mask) - 1;
 
-		if ((engine = nouveau_engine(parent, i))) {
+		if (nv_iclass(parent, NV_CLIENT_CLASS))
+			engine = nv_engine(nv_client(parent)->device);
+		else
+			engine = nouveau_engine(parent, i);
+
+		if (engine) {
 			oclass = engine->sclass;
 			while (oclass->ofuncs) {
 				if ((oclass->handle & 0xffff) == handle) {
diff --git a/drivers/gpu/drm/nouveau/core/subdev/device/base.c b/drivers/gpu/drm/nouveau/core/engine/device/base.c
index 3937ced5c753..4c72571655ad 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/device/base.c
+++ b/drivers/gpu/drm/nouveau/core/engine/device/base.c
@@ -29,7 +29,7 @@
 
 #include <core/class.h>
 
-#include <subdev/device.h>
+#include <engine/device.h>
 
 static DEFINE_MUTEX(nv_devices_mutex);
 static LIST_HEAD(nv_devices);
@@ -55,7 +55,6 @@ nouveau_device_find(u64 name)
 struct nouveau_devobj {
 	struct nouveau_parent base;
 	struct nouveau_object *subdev[NVDEV_SUBDEV_NR];
-	bool created;
 };
 
 static const u64 disable_map[] = {
@@ -173,7 +172,8 @@ nouveau_devobj_ctor(struct nouveau_object *parent,
 			case 0xa0: device->card_type = NV_50; break;
 			case 0xc0: device->card_type = NV_C0; break;
 			case 0xd0: device->card_type = NV_D0; break;
-			case 0xe0: device->card_type = NV_E0; break;
+			case 0xe0:
+			case 0xf0: device->card_type = NV_E0; break;
 			default:
 				break;
 			}
@@ -238,26 +238,24 @@ nouveau_devobj_ctor(struct nouveau_object *parent,
 	}
 
 	/* ensure requested subsystems are available for use */
-	for (i = 0, c = 0; i < NVDEV_SUBDEV_NR; i++) {
+	for (i = 1, c = 1; i < NVDEV_SUBDEV_NR; i++) {
 		if (!(oclass = device->oclass[i]) || (disable & (1ULL << i)))
 			continue;
 
-		if (!device->subdev[i]) {
-			ret = nouveau_object_ctor(nv_object(device), NULL,
-						  oclass, NULL, i,
-						  &devobj->subdev[i]);
-			if (ret == -ENODEV)
-				continue;
-			if (ret)
-				return ret;
-
-			if (nv_iclass(devobj->subdev[i], NV_ENGINE_CLASS))
-				nouveau_subdev_reset(devobj->subdev[i]);
-		} else {
+		if (device->subdev[i]) {
 			nouveau_object_ref(device->subdev[i],
 					  &devobj->subdev[i]);
+			continue;
 		}
 
+		ret = nouveau_object_ctor(nv_object(device), NULL,
+					  oclass, NULL, i,
+					  &devobj->subdev[i]);
+		if (ret == -ENODEV)
+			continue;
+		if (ret)
+			return ret;
+
 		/* note: can't init *any* subdevs until devinit has been run
 		 * due to not knowing exactly what the vbios init tables will
 		 * mess with.  devinit also can't be run until all of its
@@ -273,6 +271,10 @@ nouveau_devobj_ctor(struct nouveau_object *parent,
 				ret = nouveau_object_inc(subdev);
 				if (ret)
 					return ret;
+				atomic_dec(&nv_object(device)->usecount);
+			} else
+			if (subdev) {
+				nouveau_subdev_reset(subdev);
 			}
 		}
 	}
@@ -292,74 +294,6 @@ nouveau_devobj_dtor(struct nouveau_object *object)
 	nouveau_parent_destroy(&devobj->base);
 }
 
-static int
-nouveau_devobj_init(struct nouveau_object *object)
-{
-	struct nouveau_devobj *devobj = (void *)object;
-	struct nouveau_object *subdev;
-	int ret, i;
-
-	ret = nouveau_parent_init(&devobj->base);
-	if (ret)
-		return ret;
-
-	for (i = 0; devobj->created && i < NVDEV_SUBDEV_NR; i++) {
-		if ((subdev = devobj->subdev[i])) {
-			if (!nv_iclass(subdev, NV_ENGINE_CLASS)) {
-				ret = nouveau_object_inc(subdev);
-				if (ret)
-					goto fail;
-			}
-		}
-	}
-
-	devobj->created = true;
-	return 0;
-
-fail:
-	for (--i; i >= 0; i--) {
-		if ((subdev = devobj->subdev[i])) {
-			if (!nv_iclass(subdev, NV_ENGINE_CLASS))
-				nouveau_object_dec(subdev, false);
-		}
-	}
-
-	return ret;
-}
-
-static int
-nouveau_devobj_fini(struct nouveau_object *object, bool suspend)
-{
-	struct nouveau_devobj *devobj = (void *)object;
-	struct nouveau_object *subdev;
-	int ret, i;
-
-	for (i = NVDEV_SUBDEV_NR - 1; i >= 0; i--) {
-		if ((subdev = devobj->subdev[i])) {
-			if (!nv_iclass(subdev, NV_ENGINE_CLASS)) {
-				ret = nouveau_object_dec(subdev, suspend);
-				if (ret && suspend)
-					goto fail;
-			}
-		}
-	}
-
-	ret = nouveau_parent_fini(&devobj->base, suspend);
-fail:
-	for (; ret && suspend && i < NVDEV_SUBDEV_NR; i++) {
-		if ((subdev = devobj->subdev[i])) {
-			if (!nv_iclass(subdev, NV_ENGINE_CLASS)) {
-				ret = nouveau_object_inc(subdev);
-				if (ret) {
-					/* XXX */
-				}
-			}
-		}
-	}
-
-	return ret;
-}
-
 static u8
 nouveau_devobj_rd08(struct nouveau_object *object, u64 addr)
 {
@@ -400,8 +334,8 @@ static struct nouveau_ofuncs
 nouveau_devobj_ofuncs = {
 	.ctor = nouveau_devobj_ctor,
 	.dtor = nouveau_devobj_dtor,
-	.init = nouveau_devobj_init,
-	.fini = nouveau_devobj_fini,
+	.init = _nouveau_parent_init,
+	.fini = _nouveau_parent_fini,
 	.rd08 = nouveau_devobj_rd08,
 	.rd16 = nouveau_devobj_rd16,
 	.rd32 = nouveau_devobj_rd32,
@@ -413,12 +347,76 @@ nouveau_devobj_ofuncs = {
 /******************************************************************************
  * nouveau_device: engine functions
  *****************************************************************************/
-struct nouveau_oclass
+static struct nouveau_oclass
 nouveau_device_sclass[] = {
 	{ 0x0080, &nouveau_devobj_ofuncs },
 	{}
 };
 
+static int
+nouveau_device_fini(struct nouveau_object *object, bool suspend)
+{
+	struct nouveau_device *device = (void *)object;
+	struct nouveau_object *subdev;
+	int ret, i;
+
+	for (i = NVDEV_SUBDEV_NR - 1; i >= 0; i--) {
+		if ((subdev = device->subdev[i])) {
+			if (!nv_iclass(subdev, NV_ENGINE_CLASS)) {
+				ret = nouveau_object_dec(subdev, suspend);
+				if (ret && suspend)
+					goto fail;
+			}
+		}
+	}
+
+	ret = 0;
+fail:
+	for (; ret && i < NVDEV_SUBDEV_NR; i++) {
+		if ((subdev = device->subdev[i])) {
+			if (!nv_iclass(subdev, NV_ENGINE_CLASS)) {
+				ret = nouveau_object_inc(subdev);
+				if (ret) {
+					/* XXX */
+				}
+			}
+		}
+	}
+
+	return ret;
+}
+
+static int
+nouveau_device_init(struct nouveau_object *object)
+{
+	struct nouveau_device *device = (void *)object;
+	struct nouveau_object *subdev;
+	int ret, i;
+
+	for (i = 0; i < NVDEV_SUBDEV_NR; i++) {
+		if ((subdev = device->subdev[i])) {
+			if (!nv_iclass(subdev, NV_ENGINE_CLASS)) {
+				ret = nouveau_object_inc(subdev);
+				if (ret)
+					goto fail;
+			} else {
+				nouveau_subdev_reset(subdev);
+			}
+		}
+	}
+
+	ret = 0;
+fail:
+	for (--i; ret && i >= 0; i--) {
+		if ((subdev = device->subdev[i])) {
+			if (!nv_iclass(subdev, NV_ENGINE_CLASS))
+				nouveau_object_dec(subdev, false);
+		}
+	}
+
+	return ret;
+}
+
 static void
 nouveau_device_dtor(struct nouveau_object *object)
 {
@@ -428,17 +426,19 @@ nouveau_device_dtor(struct nouveau_object *object)
 	list_del(&device->head);
 	mutex_unlock(&nv_devices_mutex);
 
-	if (device->base.mmio)
-		iounmap(device->base.mmio);
+	if (nv_subdev(device)->mmio)
+		iounmap(nv_subdev(device)->mmio);
 
-	nouveau_subdev_destroy(&device->base);
+	nouveau_engine_destroy(&device->base);
 }
 
 static struct nouveau_oclass
 nouveau_device_oclass = {
-	.handle = NV_SUBDEV(DEVICE, 0x00),
+	.handle = NV_ENGINE(DEVICE, 0x00),
 	.ofuncs = &(struct nouveau_ofuncs) {
 		.dtor = nouveau_device_dtor,
+		.init = nouveau_device_init,
+		.fini = nouveau_device_fini,
 	},
 };
 
@@ -456,13 +456,12 @@ nouveau_device_create_(struct pci_dev *pdev, u64 name, const char *sname,
 			goto done;
 	}
 
-	ret = nouveau_subdev_create_(NULL, NULL, &nouveau_device_oclass, 0,
+	ret = nouveau_engine_create_(NULL, NULL, &nouveau_device_oclass, true,
 				     "DEVICE", "device", length, pobject);
 	device = *pobject;
 	if (ret)
 		goto done;
 
-	atomic_set(&nv_object(device)->usecount, 2);
 	device->pdev = pdev;
 	device->handle = name;
 	device->cfgopt = cfg;
@@ -470,6 +469,7 @@ nouveau_device_create_(struct pci_dev *pdev, u64 name, const char *sname,
 	device->name = sname;
 
 	nv_subdev(device)->debug = nouveau_dbgopt(device->dbgopt, "DEVICE");
+	nv_engine(device)->sclass = nouveau_device_sclass;
 	list_add(&device->head, &nv_devices);
 done:
 	mutex_unlock(&nv_devices_mutex);
diff --git a/drivers/gpu/drm/nouveau/core/subdev/device/nv04.c b/drivers/gpu/drm/nouveau/core/engine/device/nv04.c
index 473c5c03d3c9..a0284cf09c0f 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/device/nv04.c
+++ b/drivers/gpu/drm/nouveau/core/engine/device/nv04.c
@@ -22,7 +22,6 @@
  * Authors: Ben Skeggs
  */
 
-#include <subdev/device.h>
 #include <subdev/bios.h>
 #include <subdev/bus.h>
 #include <subdev/i2c.h>
@@ -34,6 +33,7 @@
 #include <subdev/instmem.h>
 #include <subdev/vm.h>
 
+#include <engine/device.h>
 #include <engine/dmaobj.h>
 #include <engine/fifo.h>
 #include <engine/software.h>
diff --git a/drivers/gpu/drm/nouveau/core/subdev/device/nv10.c b/drivers/gpu/drm/nouveau/core/engine/device/nv10.c
index d0774f5bebe1..1b7809a095c3 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/device/nv10.c
+++ b/drivers/gpu/drm/nouveau/core/engine/device/nv10.c
@@ -22,7 +22,6 @@
  * Authors: Ben Skeggs
  */
 
-#include <subdev/device.h>
 #include <subdev/bios.h>
 #include <subdev/bus.h>
 #include <subdev/gpio.h>
@@ -35,6 +34,7 @@
 #include <subdev/instmem.h>
 #include <subdev/vm.h>
 
+#include <engine/device.h>
 #include <engine/dmaobj.h>
 #include <engine/fifo.h>
 #include <engine/software.h>
diff --git a/drivers/gpu/drm/nouveau/core/subdev/device/nv20.c b/drivers/gpu/drm/nouveau/core/engine/device/nv20.c
index ab920e0dc45b..12a4005fa619 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/device/nv20.c
+++ b/drivers/gpu/drm/nouveau/core/engine/device/nv20.c
@@ -22,7 +22,6 @@
  * Authors: Ben Skeggs
  */
 
-#include <subdev/device.h>
 #include <subdev/bios.h>
 #include <subdev/bus.h>
 #include <subdev/gpio.h>
@@ -36,6 +35,7 @@
 #include <subdev/instmem.h>
 #include <subdev/vm.h>
 
+#include <engine/device.h>
 #include <engine/dmaobj.h>
 #include <engine/fifo.h>
 #include <engine/software.h>
diff --git a/drivers/gpu/drm/nouveau/core/subdev/device/nv30.c b/drivers/gpu/drm/nouveau/core/engine/device/nv30.c
index 5f2110261b04..cef0f1ea4c21 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/device/nv30.c
+++ b/drivers/gpu/drm/nouveau/core/engine/device/nv30.c
@@ -22,7 +22,6 @@
  * Authors: Ben Skeggs
  */
 
-#include <subdev/device.h>
 #include <subdev/bios.h>
 #include <subdev/bus.h>
 #include <subdev/gpio.h>
@@ -35,6 +34,7 @@
 #include <subdev/instmem.h>
 #include <subdev/vm.h>
 
+#include <engine/device.h>
 #include <engine/dmaobj.h>
 #include <engine/fifo.h>
 #include <engine/software.h>
diff --git a/drivers/gpu/drm/nouveau/core/subdev/device/nv40.c b/drivers/gpu/drm/nouveau/core/engine/device/nv40.c
index f3d55efe9ac9..1719cb0ee595 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/device/nv40.c
+++ b/drivers/gpu/drm/nouveau/core/engine/device/nv40.c
@@ -22,7 +22,6 @@
  * Authors: Ben Skeggs
  */
 
-#include <subdev/device.h>
 #include <subdev/bios.h>
 #include <subdev/bus.h>
 #include <subdev/vm.h>
@@ -37,6 +36,7 @@
 #include <subdev/instmem.h>
 #include <subdev/vm.h>
 
+#include <engine/device.h>
 #include <engine/dmaobj.h>
 #include <engine/fifo.h>
 #include <engine/software.h>
diff --git a/drivers/gpu/drm/nouveau/core/subdev/device/nv50.c b/drivers/gpu/drm/nouveau/core/engine/device/nv50.c
index 5ed2fa51ddc2..5e8c3de75593 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/device/nv50.c
+++ b/drivers/gpu/drm/nouveau/core/engine/device/nv50.c
@@ -22,7 +22,6 @@
  * Authors: Ben Skeggs
  */
 
-#include <subdev/device.h>
 #include <subdev/bios.h>
 #include <subdev/bus.h>
 #include <subdev/gpio.h>
@@ -38,6 +37,7 @@
 #include <subdev/vm.h>
 #include <subdev/bar.h>
 
+#include <engine/device.h>
 #include <engine/dmaobj.h>
 #include <engine/fifo.h>
 #include <engine/software.h>
@@ -83,7 +83,7 @@ nv50_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_GPIO   ] = &nv50_gpio_oclass;
 		device->oclass[NVDEV_SUBDEV_I2C    ] = &nv50_i2c_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nv50_clock_oclass;
-		device->oclass[NVDEV_SUBDEV_THERM  ] = &nv50_therm_oclass;
+		device->oclass[NVDEV_SUBDEV_THERM  ] = &nv84_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv50_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nv50_mc_oclass;
@@ -109,7 +109,7 @@ nv50_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_GPIO   ] = &nv50_gpio_oclass;
 		device->oclass[NVDEV_SUBDEV_I2C    ] = &nv50_i2c_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nv50_clock_oclass;
-		device->oclass[NVDEV_SUBDEV_THERM  ] = &nv50_therm_oclass;
+		device->oclass[NVDEV_SUBDEV_THERM  ] = &nv84_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv50_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nv50_mc_oclass;
@@ -135,7 +135,7 @@ nv50_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_GPIO   ] = &nv50_gpio_oclass;
 		device->oclass[NVDEV_SUBDEV_I2C    ] = &nv50_i2c_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nv50_clock_oclass;
-		device->oclass[NVDEV_SUBDEV_THERM  ] = &nv50_therm_oclass;
+		device->oclass[NVDEV_SUBDEV_THERM  ] = &nv84_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv50_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nv50_mc_oclass;
@@ -161,7 +161,7 @@ nv50_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_GPIO   ] = &nv50_gpio_oclass;
 		device->oclass[NVDEV_SUBDEV_I2C    ] = &nv94_i2c_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nv50_clock_oclass;
-		device->oclass[NVDEV_SUBDEV_THERM  ] = &nv50_therm_oclass;
+		device->oclass[NVDEV_SUBDEV_THERM  ] = &nv84_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv50_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nv50_mc_oclass;
@@ -187,7 +187,7 @@ nv50_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_GPIO   ] = &nv50_gpio_oclass;
 		device->oclass[NVDEV_SUBDEV_I2C    ] = &nv94_i2c_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nv50_clock_oclass;
-		device->oclass[NVDEV_SUBDEV_THERM  ] = &nv50_therm_oclass;
+		device->oclass[NVDEV_SUBDEV_THERM  ] = &nv84_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv50_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nv50_mc_oclass;
@@ -213,7 +213,7 @@ nv50_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_GPIO   ] = &nv50_gpio_oclass;
 		device->oclass[NVDEV_SUBDEV_I2C    ] = &nv94_i2c_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nv50_clock_oclass;
-		device->oclass[NVDEV_SUBDEV_THERM  ] = &nv50_therm_oclass;
+		device->oclass[NVDEV_SUBDEV_THERM  ] = &nv84_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv50_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nv98_mc_oclass;
@@ -239,7 +239,7 @@ nv50_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_GPIO   ] = &nv50_gpio_oclass;
 		device->oclass[NVDEV_SUBDEV_I2C    ] = &nv50_i2c_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nv50_clock_oclass;
-		device->oclass[NVDEV_SUBDEV_THERM  ] = &nv50_therm_oclass;
+		device->oclass[NVDEV_SUBDEV_THERM  ] = &nv84_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv50_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nv98_mc_oclass;
@@ -265,7 +265,7 @@ nv50_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_GPIO   ] = &nv50_gpio_oclass;
 		device->oclass[NVDEV_SUBDEV_I2C    ] = &nv94_i2c_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nv50_clock_oclass;
-		device->oclass[NVDEV_SUBDEV_THERM  ] = &nv50_therm_oclass;
+		device->oclass[NVDEV_SUBDEV_THERM  ] = &nv84_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv50_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nv98_mc_oclass;
@@ -291,7 +291,7 @@ nv50_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_GPIO   ] = &nv50_gpio_oclass;
 		device->oclass[NVDEV_SUBDEV_I2C    ] = &nv94_i2c_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nv50_clock_oclass;
-		device->oclass[NVDEV_SUBDEV_THERM  ] = &nv50_therm_oclass;
+		device->oclass[NVDEV_SUBDEV_THERM  ] = &nv84_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv50_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nv98_mc_oclass;
diff --git a/drivers/gpu/drm/nouveau/core/subdev/device/nvc0.c b/drivers/gpu/drm/nouveau/core/engine/device/nvc0.c
index 4393eb4d6564..955af122c3a6 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/device/nvc0.c
+++ b/drivers/gpu/drm/nouveau/core/engine/device/nvc0.c
@@ -22,7 +22,6 @@
  * Authors: Ben Skeggs
  */
 
-#include <subdev/device.h>
 #include <subdev/bios.h>
 #include <subdev/bus.h>
 #include <subdev/gpio.h>
@@ -40,6 +39,7 @@
 #include <subdev/vm.h>
 #include <subdev/bar.h>
 
+#include <engine/device.h>
 #include <engine/dmaobj.h>
 #include <engine/fifo.h>
 #include <engine/software.h>
@@ -285,6 +285,34 @@ nvc0_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_ENGINE_COPY0  ] = &nvc0_copy0_oclass;
 		device->oclass[NVDEV_ENGINE_DISP   ] = &nvd0_disp_oclass;
 		break;
+	case 0xd7:
+		device->cname = "GF117";
+		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
+		device->oclass[NVDEV_SUBDEV_GPIO   ] = &nvd0_gpio_oclass;
+		device->oclass[NVDEV_SUBDEV_I2C    ] = &nvd0_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nvc0_clock_oclass;
+		device->oclass[NVDEV_SUBDEV_THERM  ] = &nvd0_therm_oclass;
+		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
+		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv50_devinit_oclass;
+		device->oclass[NVDEV_SUBDEV_MC     ] = &nvc0_mc_oclass;
+		device->oclass[NVDEV_SUBDEV_BUS    ] = &nvc0_bus_oclass;
+		device->oclass[NVDEV_SUBDEV_TIMER  ] = &nv04_timer_oclass;
+		device->oclass[NVDEV_SUBDEV_FB     ] = &nvc0_fb_oclass;
+		device->oclass[NVDEV_SUBDEV_LTCG   ] = &nvc0_ltcg_oclass;
+		device->oclass[NVDEV_SUBDEV_IBUS   ] = &nvc0_ibus_oclass;
+		device->oclass[NVDEV_SUBDEV_INSTMEM] = &nv50_instmem_oclass;
+		device->oclass[NVDEV_SUBDEV_VM     ] = &nvc0_vmmgr_oclass;
+		device->oclass[NVDEV_SUBDEV_BAR    ] = &nvc0_bar_oclass;
+		device->oclass[NVDEV_ENGINE_DMAOBJ ] = &nvd0_dmaeng_oclass;
+		device->oclass[NVDEV_ENGINE_FIFO   ] = &nvc0_fifo_oclass;
+		device->oclass[NVDEV_ENGINE_SW     ] = &nvc0_software_oclass;
+		device->oclass[NVDEV_ENGINE_GR     ] = &nvc0_graph_oclass;
+		device->oclass[NVDEV_ENGINE_VP     ] = &nvc0_vp_oclass;
+		device->oclass[NVDEV_ENGINE_BSP    ] = &nvc0_bsp_oclass;
+		device->oclass[NVDEV_ENGINE_PPP    ] = &nvc0_ppp_oclass;
+		device->oclass[NVDEV_ENGINE_COPY0  ] = &nvc0_copy0_oclass;
+		device->oclass[NVDEV_ENGINE_DISP   ] = &nvd0_disp_oclass;
+		break;
 	default:
 		nv_fatal(device, "unknown Fermi chipset\n");
 		return -EINVAL;
diff --git a/drivers/gpu/drm/nouveau/core/subdev/device/nve0.c b/drivers/gpu/drm/nouveau/core/engine/device/nve0.c
index 5c12391619fd..a354e409cdff 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/device/nve0.c
+++ b/drivers/gpu/drm/nouveau/core/engine/device/nve0.c
@@ -22,7 +22,6 @@
  * Authors: Ben Skeggs
  */
 
-#include <subdev/device.h>
 #include <subdev/bios.h>
 #include <subdev/bus.h>
 #include <subdev/gpio.h>
@@ -40,6 +39,7 @@
 #include <subdev/vm.h>
 #include <subdev/bar.h>
 
+#include <engine/device.h>
 #include <engine/dmaobj.h>
 #include <engine/fifo.h>
 #include <engine/software.h>
@@ -141,6 +141,40 @@ nve0_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_ENGINE_VP     ] = &nve0_vp_oclass;
 		device->oclass[NVDEV_ENGINE_PPP    ] = &nvc0_ppp_oclass;
 		break;
+	case 0xf0:
+		device->cname = "GK110";
+		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
+		device->oclass[NVDEV_SUBDEV_GPIO   ] = &nve0_gpio_oclass;
+		device->oclass[NVDEV_SUBDEV_I2C    ] = &nvd0_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nvc0_clock_oclass;
+		device->oclass[NVDEV_SUBDEV_THERM  ] = &nvd0_therm_oclass;
+		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
+		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv50_devinit_oclass;
+		device->oclass[NVDEV_SUBDEV_MC     ] = &nvc0_mc_oclass;
+		device->oclass[NVDEV_SUBDEV_BUS    ] = &nvc0_bus_oclass;
+		device->oclass[NVDEV_SUBDEV_TIMER  ] = &nv04_timer_oclass;
+		device->oclass[NVDEV_SUBDEV_FB     ] = &nvc0_fb_oclass;
+		device->oclass[NVDEV_SUBDEV_LTCG   ] = &nvc0_ltcg_oclass;
+		device->oclass[NVDEV_SUBDEV_IBUS   ] = &nve0_ibus_oclass;
+		device->oclass[NVDEV_SUBDEV_INSTMEM] = &nv50_instmem_oclass;
+		device->oclass[NVDEV_SUBDEV_VM     ] = &nvc0_vmmgr_oclass;
+		device->oclass[NVDEV_SUBDEV_BAR    ] = &nvc0_bar_oclass;
+		device->oclass[NVDEV_ENGINE_DMAOBJ ] = &nvd0_dmaeng_oclass;
+#if 0
+		device->oclass[NVDEV_ENGINE_FIFO   ] = &nve0_fifo_oclass;
+		device->oclass[NVDEV_ENGINE_SW     ] = &nvc0_software_oclass;
+		device->oclass[NVDEV_ENGINE_GR     ] = &nve0_graph_oclass;
+#endif
+		device->oclass[NVDEV_ENGINE_DISP   ] = &nvf0_disp_oclass;
+#if 0
+		device->oclass[NVDEV_ENGINE_COPY0  ] = &nve0_copy0_oclass;
+		device->oclass[NVDEV_ENGINE_COPY1  ] = &nve0_copy1_oclass;
+		device->oclass[NVDEV_ENGINE_COPY2  ] = &nve0_copy2_oclass;
+		device->oclass[NVDEV_ENGINE_BSP    ] = &nve0_bsp_oclass;
+		device->oclass[NVDEV_ENGINE_VP     ] = &nve0_vp_oclass;
+		device->oclass[NVDEV_ENGINE_PPP    ] = &nvc0_ppp_oclass;
+#endif
+		break;
 	default:
 		nv_fatal(device, "unknown Kepler chipset\n");
 		return -EINVAL;
diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/dport.c b/drivers/gpu/drm/nouveau/core/engine/disp/dport.c
index fa27b02ff829..31cc8fe8e7f0 100644
--- a/drivers/gpu/drm/nouveau/core/engine/disp/dport.c
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/dport.c
@@ -191,7 +191,7 @@ dp_link_train_cr(struct dp_state *dp)
 static int
 dp_link_train_eq(struct dp_state *dp)
 {
-	bool eq_done, cr_done = true;
+	bool eq_done = false, cr_done = true;
 	int tries = 0, i;
 
 	dp_set_training_pattern(dp, 2);
diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/nv50.c b/drivers/gpu/drm/nouveau/core/engine/disp/nv50.c
index 02e369f80449..6a38402fa56c 100644
--- a/drivers/gpu/drm/nouveau/core/engine/disp/nv50.c
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/nv50.c
@@ -572,7 +572,8 @@ nv50_disp_base_ctor(struct nouveau_object *parent,
 	priv->base.vblank->priv = priv;
 	priv->base.vblank->enable = nv50_disp_base_vblank_enable;
 	priv->base.vblank->disable = nv50_disp_base_vblank_disable;
-	return nouveau_ramht_new(parent, parent, 0x1000, 0, &base->ramht);
+	return nouveau_ramht_new(nv_object(base), nv_object(base), 0x1000, 0,
+				&base->ramht);
 }
 
 static void
@@ -719,7 +720,7 @@ nv50_disp_data_ctor(struct nouveau_object *parent,
 	if (nv_mclass(parent) != NV_DEVICE_CLASS) {
 		atomic_inc(&parent->refcount);
 		*pobject = parent;
-		return 0;
+		return 1;
 	}
 
 	/* allocate display hardware to client */
diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/nvd0.c b/drivers/gpu/drm/nouveau/core/engine/disp/nvd0.c
index 788dd34ccb54..019eacd8a68f 100644
--- a/drivers/gpu/drm/nouveau/core/engine/disp/nvd0.c
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/nvd0.c
@@ -473,7 +473,8 @@ nvd0_disp_base_ctor(struct nouveau_object *parent,
 	priv->base.vblank->enable = nvd0_disp_base_vblank_enable;
 	priv->base.vblank->disable = nvd0_disp_base_vblank_disable;
 
-	return nouveau_ramht_new(parent, parent, 0x1000, 0, &base->ramht);
+	return nouveau_ramht_new(nv_object(base), nv_object(base), 0x1000, 0,
+				&base->ramht);
 }
 
 static void
diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/nvf0.c b/drivers/gpu/drm/nouveau/core/engine/disp/nvf0.c
new file mode 100644
index 000000000000..a488c36e40f9
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/nvf0.c
@@ -0,0 +1,89 @@
+/*
+ * Copyright 2012 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs
+ */
+
+#include <engine/software.h>
+#include <engine/disp.h>
+
+#include <core/class.h>
+
+#include "nv50.h"
+
+static struct nouveau_oclass
+nvf0_disp_sclass[] = {
+	{ NVF0_DISP_MAST_CLASS, &nvd0_disp_mast_ofuncs },
+	{ NVF0_DISP_SYNC_CLASS, &nvd0_disp_sync_ofuncs },
+	{ NVF0_DISP_OVLY_CLASS, &nvd0_disp_ovly_ofuncs },
+	{ NVF0_DISP_OIMM_CLASS, &nvd0_disp_oimm_ofuncs },
+	{ NVF0_DISP_CURS_CLASS, &nvd0_disp_curs_ofuncs },
+	{}
+};
+
+static struct nouveau_oclass
+nvf0_disp_base_oclass[] = {
+	{ NVF0_DISP_CLASS, &nvd0_disp_base_ofuncs, nva3_disp_base_omthds },
+	{}
+};
+
+static int
+nvf0_disp_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
+	       struct nouveau_oclass *oclass, void *data, u32 size,
+	       struct nouveau_object **pobject)
+{
+	struct nv50_disp_priv *priv;
+	int heads = nv_rd32(parent, 0x022448);
+	int ret;
+
+	ret = nouveau_disp_create(parent, engine, oclass, heads,
+				  "PDISP", "display", &priv);
+	*pobject = nv_object(priv);
+	if (ret)
+		return ret;
+
+	nv_engine(priv)->sclass = nvf0_disp_base_oclass;
+	nv_engine(priv)->cclass = &nv50_disp_cclass;
+	nv_subdev(priv)->intr = nvd0_disp_intr;
+	INIT_WORK(&priv->supervisor, nvd0_disp_intr_supervisor);
+	priv->sclass = nvf0_disp_sclass;
+	priv->head.nr = heads;
+	priv->dac.nr = 3;
+	priv->sor.nr = 4;
+	priv->dac.power = nv50_dac_power;
+	priv->dac.sense = nv50_dac_sense;
+	priv->sor.power = nv50_sor_power;
+	priv->sor.hda_eld = nvd0_hda_eld;
+	priv->sor.hdmi = nvd0_hdmi_ctrl;
+	priv->sor.dp = &nvd0_sor_dp_func;
+	return 0;
+}
+
+struct nouveau_oclass
+nvf0_disp_oclass = {
+	.handle = NV_ENGINE(DISP, 0x92),
+	.ofuncs = &(struct nouveau_ofuncs) {
+		.ctor = nvf0_disp_ctor,
+		.dtor = _nouveau_disp_dtor,
+		.init = _nouveau_disp_init,
+		.fini = _nouveau_disp_fini,
+	},
+};
diff --git a/drivers/gpu/drm/nouveau/core/engine/dmaobj/nvd0.c b/drivers/gpu/drm/nouveau/core/engine/dmaobj/nvd0.c
index d1528752980c..944e73ac485c 100644
--- a/drivers/gpu/drm/nouveau/core/engine/dmaobj/nvd0.c
+++ b/drivers/gpu/drm/nouveau/core/engine/dmaobj/nvd0.c
@@ -50,6 +50,9 @@ nvd0_dmaobj_bind(struct nouveau_dmaeng *dmaeng,
 		case NVE0_DISP_MAST_CLASS:
 		case NVE0_DISP_SYNC_CLASS:
 		case NVE0_DISP_OVLY_CLASS:
+		case NVF0_DISP_MAST_CLASS:
+		case NVF0_DISP_SYNC_CLASS:
+		case NVF0_DISP_OVLY_CLASS:
 			break;
 		default:
 			return -EINVAL;
diff --git a/drivers/gpu/drm/nouveau/core/engine/fifo/base.c b/drivers/gpu/drm/nouveau/core/engine/fifo/base.c
index 7341ebe131fa..d3ec436d9cb5 100644
--- a/drivers/gpu/drm/nouveau/core/engine/fifo/base.c
+++ b/drivers/gpu/drm/nouveau/core/engine/fifo/base.c
@@ -91,6 +91,8 @@ nouveau_fifo_channel_create_(struct nouveau_object *parent,
 	if (!chan->user)
 		return -EFAULT;
 
+	nouveau_event_trigger(priv->cevent, 0);
+
 	chan->size = size;
 	return 0;
 }
@@ -167,6 +169,7 @@ nouveau_fifo_destroy(struct nouveau_fifo *priv)
 {
 	kfree(priv->channel);
 	nouveau_event_destroy(&priv->uevent);
+	nouveau_event_destroy(&priv->cevent);
 	nouveau_engine_destroy(&priv->base);
 }
 
@@ -191,6 +194,10 @@ nouveau_fifo_create_(struct nouveau_object *parent,
 	if (!priv->channel)
 		return -ENOMEM;
 
+	ret = nouveau_event_create(1, &priv->cevent);
+	if (ret)
+		return ret;
+
 	ret = nouveau_event_create(1, &priv->uevent);
 	if (ret)
 		return ret;
diff --git a/drivers/gpu/drm/nouveau/core/engine/fifo/nv50.c b/drivers/gpu/drm/nouveau/core/engine/fifo/nv50.c
index 840af6172788..ddaeb5572903 100644
--- a/drivers/gpu/drm/nouveau/core/engine/fifo/nv50.c
+++ b/drivers/gpu/drm/nouveau/core/engine/fifo/nv50.c
@@ -210,7 +210,8 @@ nv50_fifo_chan_ctor_dma(struct nouveau_object *parent,
 	nv_parent(chan)->object_attach = nv50_fifo_object_attach;
 	nv_parent(chan)->object_detach = nv50_fifo_object_detach;
 
-	ret = nouveau_ramht_new(parent, parent, 0x8000, 16, &chan->ramht);
+	ret = nouveau_ramht_new(nv_object(chan), nv_object(chan), 0x8000, 16,
+				&chan->ramht);
 	if (ret)
 		return ret;
 
@@ -263,7 +264,8 @@ nv50_fifo_chan_ctor_ind(struct nouveau_object *parent,
 	nv_parent(chan)->object_attach = nv50_fifo_object_attach;
 	nv_parent(chan)->object_detach = nv50_fifo_object_detach;
 
-	ret = nouveau_ramht_new(parent, parent, 0x8000, 16, &chan->ramht);
+	ret = nouveau_ramht_new(nv_object(chan), nv_object(chan), 0x8000, 16,
+			       &chan->ramht);
 	if (ret)
 		return ret;
 
@@ -373,17 +375,17 @@ nv50_fifo_context_ctor(struct nouveau_object *parent,
 	if (ret)
 		return ret;
 
-	ret = nouveau_gpuobj_new(parent, nv_object(base), 0x0200, 0x1000,
-				 NVOBJ_FLAG_ZERO_ALLOC, &base->ramfc);
+	ret = nouveau_gpuobj_new(nv_object(base), nv_object(base), 0x0200,
+				 0x1000, NVOBJ_FLAG_ZERO_ALLOC, &base->ramfc);
 	if (ret)
 		return ret;
 
-	ret = nouveau_gpuobj_new(parent, nv_object(base), 0x1200, 0,
+	ret = nouveau_gpuobj_new(nv_object(base), nv_object(base), 0x1200, 0,
 				 NVOBJ_FLAG_ZERO_ALLOC, &base->eng);
 	if (ret)
 		return ret;
 
-	ret = nouveau_gpuobj_new(parent, nv_object(base), 0x4000, 0, 0,
+	ret = nouveau_gpuobj_new(nv_object(base), nv_object(base), 0x4000, 0, 0,
 				&base->pgd);
 	if (ret)
 		return ret;
@@ -437,12 +439,12 @@ nv50_fifo_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	if (ret)
 		return ret;
 
-	ret = nouveau_gpuobj_new(parent, NULL, 128 * 4, 0x1000, 0,
+	ret = nouveau_gpuobj_new(nv_object(priv), NULL, 128 * 4, 0x1000, 0,
 				&priv->playlist[0]);
 	if (ret)
 		return ret;
 
-	ret = nouveau_gpuobj_new(parent, NULL, 128 * 4, 0x1000, 0,
+	ret = nouveau_gpuobj_new(nv_object(priv), NULL, 128 * 4, 0x1000, 0,
 				&priv->playlist[1]);
 	if (ret)
 		return ret;
diff --git a/drivers/gpu/drm/nouveau/core/engine/fifo/nv84.c b/drivers/gpu/drm/nouveau/core/engine/fifo/nv84.c
index 094000e87871..35b94bd18808 100644
--- a/drivers/gpu/drm/nouveau/core/engine/fifo/nv84.c
+++ b/drivers/gpu/drm/nouveau/core/engine/fifo/nv84.c
@@ -180,7 +180,8 @@ nv84_fifo_chan_ctor_dma(struct nouveau_object *parent,
 	if (ret)
 		return ret;
 
-	ret = nouveau_ramht_new(parent, parent, 0x8000, 16, &chan->ramht);
+	ret = nouveau_ramht_new(nv_object(chan), nv_object(chan), 0x8000, 16,
+			       &chan->ramht);
 	if (ret)
 		return ret;
 
@@ -242,7 +243,8 @@ nv84_fifo_chan_ctor_ind(struct nouveau_object *parent,
 	if (ret)
 		return ret;
 
-	ret = nouveau_ramht_new(parent, parent, 0x8000, 16, &chan->ramht);
+	ret = nouveau_ramht_new(nv_object(chan), nv_object(chan), 0x8000, 16,
+			       &chan->ramht);
 	if (ret)
 		return ret;
 
@@ -336,12 +338,12 @@ nv84_fifo_context_ctor(struct nouveau_object *parent,
 	if (ret)
 		return ret;
 
-	ret = nouveau_gpuobj_new(parent, nv_object(base), 0x0200, 0,
+	ret = nouveau_gpuobj_new(nv_object(base), nv_object(base), 0x0200, 0,
 				 NVOBJ_FLAG_ZERO_ALLOC, &base->eng);
 	if (ret)
 		return ret;
 
-	ret = nouveau_gpuobj_new(parent, nv_object(base), 0x4000, 0,
+	ret = nouveau_gpuobj_new(nv_object(base), nv_object(base), 0x4000, 0,
 				 0, &base->pgd);
 	if (ret)
 		return ret;
@@ -350,13 +352,13 @@ nv84_fifo_context_ctor(struct nouveau_object *parent,
 	if (ret)
 		return ret;
 
-	ret = nouveau_gpuobj_new(parent, nv_object(base), 0x1000, 0x400,
-				 NVOBJ_FLAG_ZERO_ALLOC, &base->cache);
+	ret = nouveau_gpuobj_new(nv_object(base), nv_object(base), 0x1000,
+				 0x400, NVOBJ_FLAG_ZERO_ALLOC, &base->cache);
 	if (ret)
 		return ret;
 
-	ret = nouveau_gpuobj_new(parent, nv_object(base), 0x0100, 0x100,
-				 NVOBJ_FLAG_ZERO_ALLOC, &base->ramfc);
+	ret = nouveau_gpuobj_new(nv_object(base), nv_object(base), 0x0100,
+				 0x100, NVOBJ_FLAG_ZERO_ALLOC, &base->ramfc);
 	if (ret)
 		return ret;
 
@@ -407,12 +409,12 @@ nv84_fifo_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	if (ret)
 		return ret;
 
-	ret = nouveau_gpuobj_new(parent, NULL, 128 * 4, 0x1000, 0,
+	ret = nouveau_gpuobj_new(nv_object(priv), NULL, 128 * 4, 0x1000, 0,
 				&priv->playlist[0]);
 	if (ret)
 		return ret;
 
-	ret = nouveau_gpuobj_new(parent, NULL, 128 * 4, 0x1000, 0,
+	ret = nouveau_gpuobj_new(nv_object(priv), NULL, 128 * 4, 0x1000, 0,
 				&priv->playlist[1]);
 	if (ret)
 		return ret;
diff --git a/drivers/gpu/drm/nouveau/core/engine/fifo/nvc0.c b/drivers/gpu/drm/nouveau/core/engine/fifo/nvc0.c
index 4f226afb5591..4d4a6b905370 100644
--- a/drivers/gpu/drm/nouveau/core/engine/fifo/nvc0.c
+++ b/drivers/gpu/drm/nouveau/core/engine/fifo/nvc0.c
@@ -292,7 +292,8 @@ nvc0_fifo_context_ctor(struct nouveau_object *parent,
 	if (ret)
 		return ret;
 
-	ret = nouveau_gpuobj_new(parent, NULL, 0x10000, 0x1000, 0, &base->pgd);
+	ret = nouveau_gpuobj_new(nv_object(base), NULL, 0x10000, 0x1000, 0,
+				&base->pgd);
 	if (ret)
 		return ret;
 
@@ -623,17 +624,17 @@ nvc0_fifo_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	if (ret)
 		return ret;
 
-	ret = nouveau_gpuobj_new(parent, NULL, 0x1000, 0x1000, 0,
+	ret = nouveau_gpuobj_new(nv_object(priv), NULL, 0x1000, 0x1000, 0,
 				&priv->playlist[0]);
 	if (ret)
 		return ret;
 
-	ret = nouveau_gpuobj_new(parent, NULL, 0x1000, 0x1000, 0,
+	ret = nouveau_gpuobj_new(nv_object(priv), NULL, 0x1000, 0x1000, 0,
 				&priv->playlist[1]);
 	if (ret)
 		return ret;
 
-	ret = nouveau_gpuobj_new(parent, NULL, 128 * 0x1000, 0x1000, 0,
+	ret = nouveau_gpuobj_new(nv_object(priv), NULL, 128 * 0x1000, 0x1000, 0,
 				&priv->user.mem);
 	if (ret)
 		return ret;
diff --git a/drivers/gpu/drm/nouveau/core/engine/fifo/nve0.c b/drivers/gpu/drm/nouveau/core/engine/fifo/nve0.c
index 4419e40d88e9..9151919fb831 100644
--- a/drivers/gpu/drm/nouveau/core/engine/fifo/nve0.c
+++ b/drivers/gpu/drm/nouveau/core/engine/fifo/nve0.c
@@ -96,7 +96,7 @@ nve0_fifo_playlist_update(struct nve0_fifo_priv *priv, u32 engine)
 
 	cur = engn->playlist[engn->cur_playlist];
 	if (unlikely(cur == NULL)) {
-		int ret = nouveau_gpuobj_new(nv_object(priv)->parent, NULL,
+		int ret = nouveau_gpuobj_new(nv_object(priv), NULL,
 					     0x8000, 0x1000, 0, &cur);
 		if (ret) {
 			nv_error(priv, "playlist alloc failed\n");
@@ -333,7 +333,8 @@ nve0_fifo_context_ctor(struct nouveau_object *parent,
 	if (ret)
 		return ret;
 
-	ret = nouveau_gpuobj_new(parent, NULL, 0x10000, 0x1000, 0, &base->pgd);
+	ret = nouveau_gpuobj_new(nv_object(base), NULL, 0x10000, 0x1000, 0,
+				&base->pgd);
 	if (ret)
 		return ret;
 
@@ -595,7 +596,7 @@ nve0_fifo_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	if (ret)
 		return ret;
 
-	ret = nouveau_gpuobj_new(parent, NULL, 4096 * 0x200, 0x1000,
+	ret = nouveau_gpuobj_new(nv_object(priv), NULL, 4096 * 0x200, 0x1000,
 				 NVOBJ_FLAG_ZERO_ALLOC, &priv->user.mem);
 	if (ret)
 		return ret;
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/ctxnvc0.c b/drivers/gpu/drm/nouveau/core/engine/graph/ctxnvc0.c
index 0b7951a85943..4cc6269d4077 100644
--- a/drivers/gpu/drm/nouveau/core/engine/graph/ctxnvc0.c
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/ctxnvc0.c
@@ -36,7 +36,6 @@ int
 nvc0_grctx_init(struct nvc0_graph_priv *priv, struct nvc0_grctx *info)
 {
 	struct nouveau_bar *bar = nouveau_bar(priv);
-	struct nouveau_object *parent = nv_object(priv);
 	struct nouveau_gpuobj *chan;
 	u32 size = (0x80000 + priv->size + 4095) & ~4095;
 	int ret, i;
@@ -44,7 +43,7 @@ nvc0_grctx_init(struct nvc0_graph_priv *priv, struct nvc0_grctx *info)
 	/* allocate memory to for a "channel", which we'll use to generate
 	 * the default context values
 	 */
-	ret = nouveau_gpuobj_new(parent, NULL, size, 0x1000,
+	ret = nouveau_gpuobj_new(nv_object(priv), NULL, size, 0x1000,
 				 NVOBJ_FLAG_ZERO_ALLOC, &info->chan);
 	chan = info->chan;
 	if (ret) {
@@ -1399,7 +1398,7 @@ nvc0_grctx_generate_90c0(struct nvc0_graph_priv *priv)
 {
 	int i;
 
-	for (i = 0; nv_device(priv)->chipset == 0xd9 && i < 4; i++) {
+	for (i = 0; nv_device(priv)->chipset >= 0xd0 && i < 4; i++) {
 		nv_mthd(priv, 0x90c0, 0x2700 + (i * 0x40), 0x00000000);
 		nv_mthd(priv, 0x90c0, 0x2720 + (i * 0x40), 0x00000000);
 		nv_mthd(priv, 0x90c0, 0x2704 + (i * 0x40), 0x00000000);
@@ -1415,7 +1414,7 @@ nvc0_grctx_generate_90c0(struct nvc0_graph_priv *priv)
 	nv_mthd(priv, 0x90c0, 0x27ac, 0x00000000);
 	nv_mthd(priv, 0x90c0, 0x27cc, 0x00000000);
 	nv_mthd(priv, 0x90c0, 0x27ec, 0x00000000);
-	for (i = 0; nv_device(priv)->chipset == 0xd9 && i < 4; i++) {
+	for (i = 0; nv_device(priv)->chipset >= 0xd0 && i < 4; i++) {
 		nv_mthd(priv, 0x90c0, 0x2710 + (i * 0x40), 0x00014000);
 		nv_mthd(priv, 0x90c0, 0x2730 + (i * 0x40), 0x00014000);
 		nv_mthd(priv, 0x90c0, 0x2714 + (i * 0x40), 0x00000040);
@@ -1615,7 +1614,7 @@ static void
 nvc0_grctx_generate_shaders(struct nvc0_graph_priv *priv)
 {
 
-	if (nv_device(priv)->chipset == 0xd9) {
+	if (nv_device(priv)->chipset >= 0xd0) {
 		nv_wr32(priv, 0x405800, 0x0f8000bf);
 		nv_wr32(priv, 0x405830, 0x02180218);
 		nv_wr32(priv, 0x405834, 0x08000000);
@@ -1658,10 +1657,10 @@ nvc0_grctx_generate_unk64xx(struct nvc0_graph_priv *priv)
 	nv_wr32(priv, 0x4064ac, 0x00003fff);
 	nv_wr32(priv, 0x4064b4, 0x00000000);
 	nv_wr32(priv, 0x4064b8, 0x00000000);
-	if (nv_device(priv)->chipset == 0xd9)
+	if (nv_device(priv)->chipset >= 0xd0)
 		nv_wr32(priv, 0x4064bc, 0x00000000);
 	if (nv_device(priv)->chipset == 0xc1 ||
-	    nv_device(priv)->chipset == 0xd9) {
+	    nv_device(priv)->chipset >= 0xd0) {
 		nv_wr32(priv, 0x4064c0, 0x80140078);
 		nv_wr32(priv, 0x4064c4, 0x0086ffff);
 	}
@@ -1701,7 +1700,7 @@ nvc0_grctx_generate_rop(struct nvc0_graph_priv *priv)
 	/* ROPC_BROADCAST */
 	nv_wr32(priv, 0x408800, 0x02802a3c);
 	nv_wr32(priv, 0x408804, 0x00000040);
-	if (chipset == 0xd9) {
+	if (chipset >= 0xd0) {
 		nv_wr32(priv, 0x408808, 0x1043e005);
 		nv_wr32(priv, 0x408900, 0x3080b801);
 		nv_wr32(priv, 0x408904, 0x1043e005);
@@ -1735,7 +1734,7 @@ nvc0_grctx_generate_gpc(struct nvc0_graph_priv *priv)
 	nv_wr32(priv, 0x418408, 0x00000000);
 	nv_wr32(priv, 0x41840c, 0x00001008);
 	nv_wr32(priv, 0x418410, 0x0fff0fff);
-	nv_wr32(priv, 0x418414, chipset != 0xd9 ? 0x00200fff : 0x02200fff);
+	nv_wr32(priv, 0x418414, chipset < 0xd0 ? 0x00200fff : 0x02200fff);
 	nv_wr32(priv, 0x418450, 0x00000000);
 	nv_wr32(priv, 0x418454, 0x00000000);
 	nv_wr32(priv, 0x418458, 0x00000000);
@@ -1750,14 +1749,14 @@ nvc0_grctx_generate_gpc(struct nvc0_graph_priv *priv)
 	nv_wr32(priv, 0x418700, 0x00000002);
 	nv_wr32(priv, 0x418704, 0x00000080);
 	nv_wr32(priv, 0x418708, 0x00000000);
-	nv_wr32(priv, 0x41870c, chipset != 0xd9 ? 0x07c80000 : 0x00000000);
+	nv_wr32(priv, 0x41870c, chipset < 0xd0 ? 0x07c80000 : 0x00000000);
 	nv_wr32(priv, 0x418710, 0x00000000);
-	nv_wr32(priv, 0x418800, chipset != 0xd9 ? 0x0006860a : 0x7006860a);
+	nv_wr32(priv, 0x418800, chipset < 0xd0 ? 0x0006860a : 0x7006860a);
 	nv_wr32(priv, 0x418808, 0x00000000);
 	nv_wr32(priv, 0x41880c, 0x00000000);
 	nv_wr32(priv, 0x418810, 0x00000000);
 	nv_wr32(priv, 0x418828, 0x00008442);
-	if (chipset == 0xc1 || chipset == 0xd9)
+	if (chipset == 0xc1 || chipset >= 0xd0)
 		nv_wr32(priv, 0x418830, 0x10000001);
 	else
 		nv_wr32(priv, 0x418830, 0x00000001);
@@ -1768,7 +1767,7 @@ nvc0_grctx_generate_gpc(struct nvc0_graph_priv *priv)
 	nv_wr32(priv, 0x4188f0, 0x00000000);
 	nv_wr32(priv, 0x4188f4, 0x00000000);
 	nv_wr32(priv, 0x4188f8, 0x00000000);
-	if (chipset == 0xd9)
+	if (chipset >= 0xd0)
 		nv_wr32(priv, 0x4188fc, 0x20100008);
 	else if (chipset == 0xc1)
 		nv_wr32(priv, 0x4188fc, 0x00100018);
@@ -1787,7 +1786,7 @@ nvc0_grctx_generate_gpc(struct nvc0_graph_priv *priv)
 		nv_wr32(priv, 0x418a14 + (i * 0x20), 0x00000000);
 		nv_wr32(priv, 0x418a18 + (i * 0x20), 0x00000000);
 	}
-	nv_wr32(priv, 0x418b00, chipset != 0xd9 ? 0x00000000 : 0x00000006);
+	nv_wr32(priv, 0x418b00, chipset < 0xd0 ? 0x00000000 : 0x00000006);
 	nv_wr32(priv, 0x418b08, 0x0a418820);
 	nv_wr32(priv, 0x418b0c, 0x062080e6);
 	nv_wr32(priv, 0x418b10, 0x020398a4);
@@ -1804,7 +1803,7 @@ nvc0_grctx_generate_gpc(struct nvc0_graph_priv *priv)
 	nv_wr32(priv, 0x418c24, 0x00000000);
 	nv_wr32(priv, 0x418c28, 0x00000000);
 	nv_wr32(priv, 0x418c2c, 0x00000000);
-	if (chipset == 0xc1 || chipset == 0xd9)
+	if (chipset == 0xc1 || chipset >= 0xd0)
 		nv_wr32(priv, 0x418c6c, 0x00000001);
 	nv_wr32(priv, 0x418c80, 0x20200004);
 	nv_wr32(priv, 0x418c8c, 0x00000001);
@@ -1823,7 +1822,7 @@ nvc0_grctx_generate_tp(struct nvc0_graph_priv *priv)
 	nv_wr32(priv, 0x419818, 0x00000000);
 	nv_wr32(priv, 0x41983c, 0x00038bc7);
 	nv_wr32(priv, 0x419848, 0x00000000);
-	if (chipset == 0xc1 || chipset == 0xd9)
+	if (chipset == 0xc1 || chipset >= 0xd0)
 		nv_wr32(priv, 0x419864, 0x00000129);
 	else
 		nv_wr32(priv, 0x419864, 0x0000012a);
@@ -1836,7 +1835,7 @@ nvc0_grctx_generate_tp(struct nvc0_graph_priv *priv)
 	nv_wr32(priv, 0x419a14, 0x00000200);
 	nv_wr32(priv, 0x419a1c, 0x00000000);
 	nv_wr32(priv, 0x419a20, 0x00000800);
-	if (chipset == 0xd9)
+	if (chipset >= 0xd0)
 		nv_wr32(priv, 0x00419ac4, 0x0017f440);
 	else if (chipset != 0xc0 && chipset != 0xc8)
 		nv_wr32(priv, 0x00419ac4, 0x0007f440);
@@ -1847,16 +1846,16 @@ nvc0_grctx_generate_tp(struct nvc0_graph_priv *priv)
 	nv_wr32(priv, 0x419b10, 0x0a418820);
 	nv_wr32(priv, 0x419b14, 0x000000e6);
 	nv_wr32(priv, 0x419bd0, 0x00900103);
-	if (chipset == 0xc1 || chipset == 0xd9)
+	if (chipset == 0xc1 || chipset >= 0xd0)
 		nv_wr32(priv, 0x419be0, 0x00400001);
 	else
 		nv_wr32(priv, 0x419be0, 0x00000001);
 	nv_wr32(priv, 0x419be4, 0x00000000);
-	nv_wr32(priv, 0x419c00, chipset != 0xd9 ? 0x00000002 : 0x0000000a);
+	nv_wr32(priv, 0x419c00, chipset < 0xd0 ? 0x00000002 : 0x0000000a);
 	nv_wr32(priv, 0x419c04, 0x00000006);
 	nv_wr32(priv, 0x419c08, 0x00000002);
 	nv_wr32(priv, 0x419c20, 0x00000000);
-	if (nv_device(priv)->chipset == 0xd9) {
+	if (nv_device(priv)->chipset >= 0xd0) {
 		nv_wr32(priv, 0x419c24, 0x00084210);
 		nv_wr32(priv, 0x419c28, 0x3cf3cf3c);
 		nv_wr32(priv, 0x419cb0, 0x00020048);
@@ -1868,12 +1867,12 @@ nvc0_grctx_generate_tp(struct nvc0_graph_priv *priv)
 	}
 	nv_wr32(priv, 0x419ce8, 0x00000000);
 	nv_wr32(priv, 0x419cf4, 0x00000183);
-	if (chipset == 0xc1 || chipset == 0xd9)
+	if (chipset == 0xc1 || chipset >= 0xd0)
 		nv_wr32(priv, 0x419d20, 0x12180000);
 	else
 		nv_wr32(priv, 0x419d20, 0x02180000);
 	nv_wr32(priv, 0x419d24, 0x00001fff);
-	if (chipset == 0xc1 || chipset == 0xd9)
+	if (chipset == 0xc1 || chipset >= 0xd0)
 		nv_wr32(priv, 0x419d44, 0x02180218);
 	nv_wr32(priv, 0x419e04, 0x00000000);
 	nv_wr32(priv, 0x419e08, 0x00000000);
@@ -2210,7 +2209,7 @@ nvc0_grctx_generate(struct nvc0_graph_priv *priv)
 	nv_icmd(priv, 0x00000215, 0x00000040);
 	nv_icmd(priv, 0x00000216, 0x00000040);
 	nv_icmd(priv, 0x00000217, 0x00000040);
-	if (nv_device(priv)->chipset == 0xd9) {
+	if (nv_device(priv)->chipset >= 0xd0) {
 		for (i = 0x0400; i <= 0x0417; i++)
 			nv_icmd(priv, i, 0x00000040);
 	}
@@ -2222,7 +2221,7 @@ nvc0_grctx_generate(struct nvc0_graph_priv *priv)
 	nv_icmd(priv, 0x0000021d, 0x0000c080);
 	nv_icmd(priv, 0x0000021e, 0x0000c080);
 	nv_icmd(priv, 0x0000021f, 0x0000c080);
-	if (nv_device(priv)->chipset == 0xd9) {
+	if (nv_device(priv)->chipset >= 0xd0) {
 		for (i = 0x0440; i <= 0x0457; i++)
 			nv_icmd(priv, i, 0x0000c080);
 	}
@@ -2789,7 +2788,7 @@ nvc0_grctx_generate(struct nvc0_graph_priv *priv)
 	nv_icmd(priv, 0x00000585, 0x0000003f);
 	nv_icmd(priv, 0x00000576, 0x00000003);
 	if (nv_device(priv)->chipset == 0xc1 ||
-	    nv_device(priv)->chipset == 0xd9)
+	    nv_device(priv)->chipset >= 0xd0)
 		nv_icmd(priv, 0x0000057b, 0x00000059);
 	nv_icmd(priv, 0x00000586, 0x00000040);
 	nv_icmd(priv, 0x00000582, 0x00000080);
@@ -2891,7 +2890,7 @@ nvc0_grctx_generate(struct nvc0_graph_priv *priv)
 	nv_icmd(priv, 0x00000957, 0x00000003);
 	nv_icmd(priv, 0x0000095e, 0x20164010);
 	nv_icmd(priv, 0x0000095f, 0x00000020);
-	if (nv_device(priv)->chipset == 0xd9)
+	if (nv_device(priv)->chipset >= 0xd0)
 		nv_icmd(priv, 0x0000097d, 0x00000020);
 	nv_icmd(priv, 0x00000683, 0x00000006);
 	nv_icmd(priv, 0x00000685, 0x003fffff);
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/ctxnve0.c b/drivers/gpu/drm/nouveau/core/engine/graph/ctxnve0.c
index 6d8c63931ee6..ae27dae3fe38 100644
--- a/drivers/gpu/drm/nouveau/core/engine/graph/ctxnve0.c
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/ctxnve0.c
@@ -2772,10 +2772,15 @@ nve0_grctx_generate(struct nvc0_graph_priv *priv)
 	for (i = 0; i < 8; i++)
 		nv_wr32(priv, 0x4064d0 + (i * 0x04), 0x00000000);
 
-	nv_wr32(priv, 0x405b00, 0x201);
-	nv_wr32(priv, 0x408850, 0x2);
-	nv_wr32(priv, 0x408958, 0x2);
-	nv_wr32(priv, 0x419f78, 0xa);
+	nv_wr32(priv, 0x405b00, (priv->tpc_total << 8) | priv->gpc_nr);
+	if (priv->gpc_nr == 1) {
+		nv_mask(priv, 0x408850, 0x0000000f, priv->tpc_nr[0]);
+		nv_mask(priv, 0x408958, 0x0000000f, priv->tpc_nr[0]);
+	} else {
+		nv_mask(priv, 0x408850, 0x0000000f, priv->gpc_nr);
+		nv_mask(priv, 0x408958, 0x0000000f, priv->gpc_nr);
+	}
+	nv_mask(priv, 0x419f78, 0x00000001, 0x00000000);
 
 	nve0_grctx_generate_icmd(priv);
 	nve0_grctx_generate_a097(priv);
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/fuc/gpcnvc0.fuc b/drivers/gpu/drm/nouveau/core/engine/graph/fuc/gpcnvc0.fuc
index b86cc60dcd56..f7055af0f2a6 100644
--- a/drivers/gpu/drm/nouveau/core/engine/graph/fuc/gpcnvc0.fuc
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/fuc/gpcnvc0.fuc
@@ -87,6 +87,11 @@ chipsets:
 .b16 #nvd9_gpc_mmio_tail
 .b16 #nvd9_tpc_mmio_head
 .b16 #nvd9_tpc_mmio_tail
+.b8  0xd7 0 0 0
+.b16 #nvd9_gpc_mmio_head
+.b16 #nvd9_gpc_mmio_tail
+.b16 #nvd9_tpc_mmio_head
+.b16 #nvd9_tpc_mmio_tail
 .b8  0 0 0 0
 
 // GPC mmio lists
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/fuc/hubnvc0.fuc b/drivers/gpu/drm/nouveau/core/engine/graph/fuc/hubnvc0.fuc
index 0bcfa4d447e5..7fbdebb2bafb 100644
--- a/drivers/gpu/drm/nouveau/core/engine/graph/fuc/hubnvc0.fuc
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/fuc/hubnvc0.fuc
@@ -62,6 +62,9 @@ chipsets:
 .b8  0xd9 0 0 0
 .b16 #nvd9_hub_mmio_head
 .b16 #nvd9_hub_mmio_tail
+.b8  0xd7 0 0 0
+.b16 #nvd9_hub_mmio_head
+.b16 #nvd9_hub_mmio_tail
 .b8  0 0 0 0
 
 nvc0_hub_mmio_head:
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/nv20.c b/drivers/gpu/drm/nouveau/core/engine/graph/nv20.c
index 0607b9801748..b24559315903 100644
--- a/drivers/gpu/drm/nouveau/core/engine/graph/nv20.c
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/nv20.c
@@ -254,7 +254,7 @@ nv20_graph_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	if (ret)
 		return ret;
 
-	ret = nouveau_gpuobj_new(parent, NULL, 32 * 4, 16,
+	ret = nouveau_gpuobj_new(nv_object(priv), NULL, 32 * 4, 16,
 				 NVOBJ_FLAG_ZERO_ALLOC, &priv->ctxtab);
 	if (ret)
 		return ret;
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/nv25.c b/drivers/gpu/drm/nouveau/core/engine/graph/nv25.c
index b2b650dd8b28..7a80d005a974 100644
--- a/drivers/gpu/drm/nouveau/core/engine/graph/nv25.c
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/nv25.c
@@ -142,7 +142,7 @@ nv25_graph_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	if (ret)
 		return ret;
 
-	ret = nouveau_gpuobj_new(parent, NULL, 32 * 4, 16,
+	ret = nouveau_gpuobj_new(nv_object(priv), NULL, 32 * 4, 16,
 				 NVOBJ_FLAG_ZERO_ALLOC, &priv->ctxtab);
 	if (ret)
 		return ret;
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/nv2a.c b/drivers/gpu/drm/nouveau/core/engine/graph/nv2a.c
index 700462fa0ae0..3e1f32ee43d4 100644
--- a/drivers/gpu/drm/nouveau/core/engine/graph/nv2a.c
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/nv2a.c
@@ -109,7 +109,7 @@ nv2a_graph_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	if (ret)
 		return ret;
 
-	ret = nouveau_gpuobj_new(parent, NULL, 32 * 4, 16,
+	ret = nouveau_gpuobj_new(nv_object(priv), NULL, 32 * 4, 16,
 				 NVOBJ_FLAG_ZERO_ALLOC, &priv->ctxtab);
 	if (ret)
 		return ret;
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/nv30.c b/drivers/gpu/drm/nouveau/core/engine/graph/nv30.c
index cedadaa92d3f..e451db32e92a 100644
--- a/drivers/gpu/drm/nouveau/core/engine/graph/nv30.c
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/nv30.c
@@ -143,7 +143,7 @@ nv30_graph_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	if (ret)
 		return ret;
 
-	ret = nouveau_gpuobj_new(parent, NULL, 32 * 4, 16,
+	ret = nouveau_gpuobj_new(nv_object(priv), NULL, 32 * 4, 16,
 				 NVOBJ_FLAG_ZERO_ALLOC, &priv->ctxtab);
 	if (ret)
 		return ret;
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/nv34.c b/drivers/gpu/drm/nouveau/core/engine/graph/nv34.c
index 273f6320027b..9385ac7b44a4 100644
--- a/drivers/gpu/drm/nouveau/core/engine/graph/nv34.c
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/nv34.c
@@ -143,7 +143,7 @@ nv34_graph_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	if (ret)
 		return ret;
 
-	ret = nouveau_gpuobj_new(parent, NULL, 32 * 4, 16,
+	ret = nouveau_gpuobj_new(nv_object(priv), NULL, 32 * 4, 16,
 				 NVOBJ_FLAG_ZERO_ALLOC, &priv->ctxtab);
 	if (ret)
 		return ret;
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/nv35.c b/drivers/gpu/drm/nouveau/core/engine/graph/nv35.c
index f40ee2116ee1..9ce84b73f86a 100644
--- a/drivers/gpu/drm/nouveau/core/engine/graph/nv35.c
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/nv35.c
@@ -141,7 +141,7 @@ nv35_graph_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	if (ret)
 		return ret;
 
-	ret = nouveau_gpuobj_new(parent, NULL, 32 * 4, 16,
+	ret = nouveau_gpuobj_new(nv_object(priv), NULL, 32 * 4, 16,
 				 NVOBJ_FLAG_ZERO_ALLOC, &priv->ctxtab);
 	if (ret)
 		return ret;
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/nv40.c b/drivers/gpu/drm/nouveau/core/engine/graph/nv40.c
index 17049d5c723d..193a5de1b482 100644
--- a/drivers/gpu/drm/nouveau/core/engine/graph/nv40.c
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/nv40.c
@@ -46,6 +46,14 @@ struct nv40_graph_chan {
 	struct nouveau_graph_chan base;
 };
 
+static u64
+nv40_graph_units(struct nouveau_graph *graph)
+{
+	struct nv40_graph_priv *priv = (void *)graph;
+
+	return nv_rd32(priv, 0x1540);
+}
+
 /*******************************************************************************
  * Graphics object classes
  ******************************************************************************/
@@ -359,6 +367,8 @@ nv40_graph_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	else
 		nv_engine(priv)->sclass = nv40_graph_sclass;
 	nv_engine(priv)->tile_prog = nv40_graph_tile_prog;
+
+	priv->base.units = nv40_graph_units;
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/nv50.c b/drivers/gpu/drm/nouveau/core/engine/graph/nv50.c
index f2b1a7a124f2..1ac36110ca19 100644
--- a/drivers/gpu/drm/nouveau/core/engine/graph/nv50.c
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/nv50.c
@@ -48,6 +48,14 @@ struct nv50_graph_chan {
 	struct nouveau_graph_chan base;
 };
 
+static u64
+nv50_graph_units(struct nouveau_graph *graph)
+{
+	struct nv50_graph_priv *priv = (void *)graph;
+
+	return nv_rd32(priv, 0x1540);
+}
+
 /*******************************************************************************
  * Graphics object classes
  ******************************************************************************/
@@ -819,6 +827,8 @@ nv50_graph_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	nv_subdev(priv)->intr = nv50_graph_intr;
 	nv_engine(priv)->cclass = &nv50_graph_cclass;
 
+	priv->base.units = nv50_graph_units;
+
 	switch (nv_device(priv)->chipset) {
 	case 0x50:
 		nv_engine(priv)->sclass = nv50_graph_sclass;
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/nvc0.c b/drivers/gpu/drm/nouveau/core/engine/graph/nvc0.c
index 0de0dd724aff..f9b9d82c287f 100644
--- a/drivers/gpu/drm/nouveau/core/engine/graph/nvc0.c
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/nvc0.c
@@ -60,6 +60,19 @@ nvc8_graph_sclass[] = {
 	{}
 };
 
+u64
+nvc0_graph_units(struct nouveau_graph *graph)
+{
+	struct nvc0_graph_priv *priv = (void *)graph;
+	u64 cfg;
+
+	cfg  = (u32)priv->gpc_nr;
+	cfg |= (u32)priv->tpc_total << 8;
+	cfg |= (u64)priv->rop_nr << 32;
+
+	return cfg;
+}
+
 /*******************************************************************************
  * PGRAPH context
  ******************************************************************************/
@@ -89,7 +102,8 @@ nvc0_graph_context_ctor(struct nouveau_object *parent,
 	 * fuc to modify some per-context register settings on first load
 	 * of the context.
 	 */
-	ret = nouveau_gpuobj_new(parent, NULL, 0x1000, 0x100, 0, &chan->mmio);
+	ret = nouveau_gpuobj_new(nv_object(chan), NULL, 0x1000, 0x100, 0,
+				&chan->mmio);
 	if (ret)
 		return ret;
 
@@ -101,8 +115,8 @@ nvc0_graph_context_ctor(struct nouveau_object *parent,
 
 	/* allocate buffers referenced by mmio list */
 	for (i = 0; data->size && i < ARRAY_SIZE(priv->mmio_data); i++) {
-		ret = nouveau_gpuobj_new(parent, NULL, data->size, data->align,
-					 0, &chan->data[i].mem);
+		ret = nouveau_gpuobj_new(nv_object(chan), NULL, data->size,
+					 data->align, 0, &chan->data[i].mem);
 		if (ret)
 			return ret;
 
@@ -518,9 +532,10 @@ nvc0_graph_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 {
 	struct nouveau_device *device = nv_device(parent);
 	struct nvc0_graph_priv *priv;
+	bool enable = device->chipset != 0xd7;
 	int ret, i;
 
-	ret = nouveau_graph_create(parent, engine, oclass, true, &priv);
+	ret = nouveau_graph_create(parent, engine, oclass, enable, &priv);
 	*pobject = nv_object(priv);
 	if (ret)
 		return ret;
@@ -529,6 +544,8 @@ nvc0_graph_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	nv_subdev(priv)->intr = nvc0_graph_intr;
 	nv_engine(priv)->cclass = &nvc0_graph_cclass;
 
+	priv->base.units = nvc0_graph_units;
+
 	if (nouveau_boolopt(device->cfgopt, "NvGrUseFW", false)) {
 		nv_info(priv, "using external firmware\n");
 		if (nvc0_graph_ctor_fw(priv, "fuc409c", &priv->fuc409c) ||
@@ -551,11 +568,13 @@ nvc0_graph_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 		break;
 	}
 
-	ret = nouveau_gpuobj_new(parent, NULL, 0x1000, 256, 0, &priv->unk4188b4);
+	ret = nouveau_gpuobj_new(nv_object(priv), NULL, 0x1000, 256, 0,
+				&priv->unk4188b4);
 	if (ret)
 		return ret;
 
-	ret = nouveau_gpuobj_new(parent, NULL, 0x1000, 256, 0, &priv->unk4188b8);
+	ret = nouveau_gpuobj_new(nv_object(priv), NULL, 0x1000, 256, 0,
+				&priv->unk4188b8);
 	if (ret)
 		return ret;
 
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/nvc0.h b/drivers/gpu/drm/nouveau/core/engine/graph/nvc0.h
index a1e78de46456..c870dad0f670 100644
--- a/drivers/gpu/drm/nouveau/core/engine/graph/nvc0.h
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/nvc0.h
@@ -118,6 +118,7 @@ nvc0_graph_class(void *obj)
 		return 0x9197;
 	case 0xc8:
 	case 0xd9:
+	case 0xd7:
 		return 0x9297;
 	case 0xe4:
 	case 0xe7:
@@ -169,4 +170,6 @@ int  nvc0_graph_context_ctor(struct nouveau_object *, struct nouveau_object *,
 			     struct nouveau_object **);
 void nvc0_graph_context_dtor(struct nouveau_object *);
 
+u64 nvc0_graph_units(struct nouveau_graph *);
+
 #endif
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/nve0.c b/drivers/gpu/drm/nouveau/core/engine/graph/nve0.c
index 4857f913efdd..678c16f63055 100644
--- a/drivers/gpu/drm/nouveau/core/engine/graph/nve0.c
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/nve0.c
@@ -77,11 +77,207 @@ nve0_graph_ctxctl_isr(struct nvc0_graph_priv *priv)
 	nv_wr32(priv, 0x409c20, ustat);
 }
 
+static const struct nouveau_enum nve0_mp_warp_error[] = {
+	{ 0x00, "NO_ERROR" },
+	{ 0x01, "STACK_MISMATCH" },
+	{ 0x05, "MISALIGNED_PC" },
+	{ 0x08, "MISALIGNED_GPR" },
+	{ 0x09, "INVALID_OPCODE" },
+	{ 0x0d, "GPR_OUT_OF_BOUNDS" },
+	{ 0x0e, "MEM_OUT_OF_BOUNDS" },
+	{ 0x0f, "UNALIGNED_MEM_ACCESS" },
+	{ 0x11, "INVALID_PARAM" },
+	{}
+};
+
+static const struct nouveau_enum nve0_mp_global_error[] = {
+	{ 2, "MULTIPLE_WARP_ERRORS" },
+	{ 3, "OUT_OF_STACK_SPACE" },
+	{}
+};
+
+static const struct nouveau_enum nve0_gpc_rop_error[] = {
+	{ 1, "RT_PITCH_OVERRUN" },
+	{ 4, "RT_WIDTH_OVERRUN" },
+	{ 5, "RT_HEIGHT_OVERRUN" },
+	{ 7, "ZETA_STORAGE_TYPE_MISMATCH" },
+	{ 8, "RT_STORAGE_TYPE_MISMATCH" },
+	{ 10, "RT_LINEAR_MISMATCH" },
+	{}
+};
+
+static const struct nouveau_enum nve0_sked_error[] = {
+	{ 7, "CONSTANT_BUFFER_SIZE" },
+	{ 9, "LOCAL_MEMORY_SIZE_POS" },
+	{ 10, "LOCAL_MEMORY_SIZE_NEG" },
+	{ 11, "WARP_CSTACK_SIZE" },
+	{ 12, "TOTAL_TEMP_SIZE" },
+	{ 13, "REGISTER_COUNT" },
+	{ 18, "TOTAL_THREADS" },
+	{ 20, "PROGRAM_OFFSET" },
+	{ 21, "SHARED_MEMORY_SIZE" },
+	{ 25, "SHARED_CONFIG_TOO_SMALL" },
+	{ 26, "TOTAL_REGISTER_COUNT" },
+	{}
+};
+
+static void
+nve0_graph_mp_trap(struct nvc0_graph_priv *priv, int gpc, int tp)
+{
+	int i;
+	u32 werr = nv_rd32(priv, TPC_UNIT(gpc, tp, 0x648));
+	u32 gerr = nv_rd32(priv, TPC_UNIT(gpc, tp, 0x650));
+
+	nv_error(priv, "GPC%i/TP%i/MP trap:", gpc, tp);
+
+	for (i = 0; i <= 31; ++i) {
+		if (!(gerr & (1 << i)))
+			continue;
+		pr_cont(" ");
+		nouveau_enum_print(nve0_mp_global_error, i);
+	}
+	if (werr) {
+		pr_cont(" ");
+		nouveau_enum_print(nve0_mp_warp_error, werr & 0xffff);
+	}
+	pr_cont("\n");
+
+	/* disable MP trap to avoid spam */
+	nv_mask(priv, TPC_UNIT(gpc, tp, 0x50c), 0x2, 0x0);
+
+	/* TODO: figure out how to resume after an MP trap */
+}
+
+static void
+nve0_graph_tp_trap(struct nvc0_graph_priv *priv, int gpc, int tp)
+{
+	u32 stat = nv_rd32(priv, TPC_UNIT(gpc, tp, 0x508));
+
+	if (stat & 0x1) {
+		u32 trap = nv_rd32(priv, TPC_UNIT(gpc, tp, 0x224));
+		nv_error(priv, "GPC%i/TP%i/TEX trap: %08x\n",
+			 gpc, tp, trap);
+
+		nv_wr32(priv, TPC_UNIT(gpc, tp, 0x224), 0xc0000000);
+		stat &= ~0x1;
+	}
+
+	if (stat & 0x2) {
+		nve0_graph_mp_trap(priv, gpc, tp);
+		stat &= ~0x2;
+	}
+
+	if (stat & 0x4) {
+		u32 trap = nv_rd32(priv, TPC_UNIT(gpc, tp, 0x084));
+		nv_error(priv, "GPC%i/TP%i/POLY trap: %08x\n",
+			 gpc, tp, trap);
+
+		nv_wr32(priv, TPC_UNIT(gpc, tp, 0x084), 0xc0000000);
+		stat &= ~0x4;
+	}
+
+	if (stat & 0x8) {
+		u32 trap = nv_rd32(priv, TPC_UNIT(gpc, tp, 0x48c));
+		nv_error(priv, "GPC%i/TP%i/L1C trap: %08x\n",
+			 gpc, tp, trap);
+
+		nv_wr32(priv, TPC_UNIT(gpc, tp, 0x48c), 0xc0000000);
+		stat &= ~0x8;
+	}
+
+	if (stat) {
+		nv_error(priv, "GPC%i/TP%i: unknown stat %08x\n",
+			 gpc, tp, stat);
+	}
+}
+
+static void
+nve0_graph_gpc_trap(struct nvc0_graph_priv *priv)
+{
+	const u32 mask = nv_rd32(priv, 0x400118);
+	int gpc;
+
+	for (gpc = 0; gpc < 4; ++gpc) {
+		u32 stat;
+		int tp;
+
+		if (!(mask & (1 << gpc)))
+			continue;
+		stat = nv_rd32(priv, GPC_UNIT(gpc, 0x2c90));
+
+		if (stat & 0x0001) {
+			u32 trap[4];
+			int i;
+
+			trap[0] = nv_rd32(priv, GPC_UNIT(gpc, 0x0420));
+			trap[1] = nv_rd32(priv, GPC_UNIT(gpc, 0x0434));
+			trap[2] = nv_rd32(priv, GPC_UNIT(gpc, 0x0438));
+			trap[3] = nv_rd32(priv, GPC_UNIT(gpc, 0x043c));
+
+			nv_error(priv, "GPC%i/PROP trap:", gpc);
+			for (i = 0; i <= 29; ++i) {
+				if (!(trap[0] & (1 << i)))
+					continue;
+				pr_cont(" ");
+				nouveau_enum_print(nve0_gpc_rop_error, i);
+			}
+			pr_cont("\n");
+
+			nv_error(priv, "x = %u, y = %u, "
+				 "format = %x, storage type = %x\n",
+				 trap[1] & 0xffff,
+				 trap[1] >> 16,
+				 (trap[2] >> 8) & 0x3f,
+				 trap[3] & 0xff);
+
+			nv_wr32(priv, GPC_UNIT(gpc, 0x0420), 0xc0000000);
+			stat &= ~0x0001;
+		}
+
+		if (stat & 0x0002) {
+			u32 trap = nv_rd32(priv, GPC_UNIT(gpc, 0x0900));
+			nv_error(priv, "GPC%i/ZCULL trap: %08x\n", gpc,
+				 trap);
+			nv_wr32(priv, GPC_UNIT(gpc, 0x0900), 0xc0000000);
+			stat &= ~0x0002;
+		}
+
+		if (stat & 0x0004) {
+			u32 trap = nv_rd32(priv, GPC_UNIT(gpc, 0x1028));
+			nv_error(priv, "GPC%i/CCACHE trap: %08x\n", gpc,
+				 trap);
+			nv_wr32(priv, GPC_UNIT(gpc, 0x1028), 0xc0000000);
+			stat &= ~0x0004;
+		}
+
+		if (stat & 0x0008) {
+			u32 trap = nv_rd32(priv, GPC_UNIT(gpc, 0x0824));
+			nv_error(priv, "GPC%i/ESETUP trap %08x\n", gpc,
+				 trap);
+			nv_wr32(priv, GPC_UNIT(gpc, 0x0824), 0xc0000000);
+			stat &= ~0x0008;
+		}
+
+		for (tp = 0; tp < 8; ++tp) {
+			if (stat & (1 << (16 + tp)))
+				nve0_graph_tp_trap(priv, gpc, tp);
+		}
+		stat &= ~0xff0000;
+
+		if (stat) {
+			nv_error(priv, "GPC%i: unknown stat %08x\n",
+				 gpc, stat);
+		}
+	}
+}
+
+
 static void
 nve0_graph_trap_isr(struct nvc0_graph_priv *priv, int chid, u64 inst,
 		struct nouveau_object *engctx)
 {
 	u32 trap = nv_rd32(priv, 0x400108);
+	int i;
 	int rop;
 
 	if (trap & 0x00000001) {
@@ -102,6 +298,32 @@ nve0_graph_trap_isr(struct nvc0_graph_priv *priv, int chid, u64 inst,
 		trap &= ~0x00000010;
 	}
 
+	if (trap & 0x00000100) {
+		u32 stat = nv_rd32(priv, 0x407020);
+		nv_error(priv, "SKED ch %d [0x%010llx %s]:",
+			 chid, inst, nouveau_client_name(engctx));
+
+		for (i = 0; i <= 29; ++i) {
+			if (!(stat & (1 << i)))
+				continue;
+			pr_cont(" ");
+			nouveau_enum_print(nve0_sked_error, i);
+		}
+		pr_cont("\n");
+
+		if (stat & 0x3fffffff)
+			nv_wr32(priv, 0x407020, 0x40000000);
+		nv_wr32(priv, 0x400108, 0x00000100);
+		trap &= ~0x00000100;
+	}
+
+	if (trap & 0x01000000) {
+		nv_error(priv, "GPC ch %d [0x%010llx %s]:\n",
+			 chid, inst, nouveau_client_name(engctx));
+		nve0_graph_gpc_trap(priv);
+		trap &= ~0x01000000;
+	}
+
 	if (trap & 0x02000000) {
 		for (rop = 0; rop < priv->rop_nr; rop++) {
 			u32 statz = nv_rd32(priv, ROP_UNIT(rop, 0x070));
@@ -217,6 +439,8 @@ nve0_graph_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	nv_engine(priv)->cclass = &nve0_graph_cclass;
 	nv_engine(priv)->sclass = nve0_graph_sclass;
 
+	priv->base.units = nvc0_graph_units;
+
 	if (nouveau_boolopt(device->cfgopt, "NvGrUseFW", false)) {
 		nv_info(priv, "using external firmware\n");
 		if (nvc0_graph_ctor_fw(priv, "fuc409c", &priv->fuc409c) ||
@@ -227,11 +451,13 @@ nve0_graph_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 		priv->firmware = true;
 	}
 
-	ret = nouveau_gpuobj_new(parent, NULL, 0x1000, 256, 0, &priv->unk4188b4);
+	ret = nouveau_gpuobj_new(nv_object(priv), NULL, 0x1000, 256, 0,
+				&priv->unk4188b4);
 	if (ret)
 		return ret;
 
-	ret = nouveau_gpuobj_new(parent, NULL, 0x1000, 256, 0, &priv->unk4188b8);
+	ret = nouveau_gpuobj_new(nv_object(priv), NULL, 0x1000, 256, 0,
+				&priv->unk4188b8);
 	if (ret)
 		return ret;
 
diff --git a/drivers/gpu/drm/nouveau/core/engine/software/nvc0.c b/drivers/gpu/drm/nouveau/core/engine/software/nvc0.c
index a523eaad47e3..d698e710ddd4 100644
--- a/drivers/gpu/drm/nouveau/core/engine/software/nvc0.c
+++ b/drivers/gpu/drm/nouveau/core/engine/software/nvc0.c
@@ -94,6 +94,32 @@ nvc0_software_mthd_flip(struct nouveau_object *object, u32 mthd,
 	return -EINVAL;
 }
 
+static int
+nvc0_software_mthd_mp_control(struct nouveau_object *object, u32 mthd,
+                              void *args, u32 size)
+{
+	struct nvc0_software_chan *chan = (void *)nv_engctx(object->parent);
+	struct nvc0_software_priv *priv = (void *)nv_object(chan)->engine;
+	u32 data = *(u32 *)args;
+
+	switch (mthd) {
+	case 0x600:
+		nv_wr32(priv, 0x419e00, data); /* MP.PM_UNK000 */
+		break;
+	case 0x644:
+		if (data & ~0x1ffffe)
+			return -EINVAL;
+		nv_wr32(priv, 0x419e44, data); /* MP.TRAP_WARP_ERROR_EN */
+		break;
+	case 0x6ac:
+		nv_wr32(priv, 0x419eac, data); /* MP.PM_UNK0AC */
+		break;
+	default:
+		return -EINVAL;
+	}
+	return 0;
+}
+
 static struct nouveau_omthds
 nvc0_software_omthds[] = {
 	{ 0x0400, 0x0400, nvc0_software_mthd_vblsem_offset },
@@ -101,6 +127,9 @@ nvc0_software_omthds[] = {
 	{ 0x0408, 0x0408, nvc0_software_mthd_vblsem_value },
 	{ 0x040c, 0x040c, nvc0_software_mthd_vblsem_release },
 	{ 0x0500, 0x0500, nvc0_software_mthd_flip },
+	{ 0x0600, 0x0600, nvc0_software_mthd_mp_control },
+	{ 0x0644, 0x0644, nvc0_software_mthd_mp_control },
+	{ 0x06ac, 0x06ac, nvc0_software_mthd_mp_control },
 	{}
 };
 
diff --git a/drivers/gpu/drm/nouveau/core/include/core/class.h b/drivers/gpu/drm/nouveau/core/include/core/class.h
index 92d3ab11d962..0a393f7f055f 100644
--- a/drivers/gpu/drm/nouveau/core/include/core/class.h
+++ b/drivers/gpu/drm/nouveau/core/include/core/class.h
@@ -169,6 +169,7 @@ struct nv04_display_class {
  * 8570: NVA3_DISP
  * 9070: NVD0_DISP
  * 9170: NVE0_DISP
+ * 9270: NVF0_DISP
  */
 
 #define NV50_DISP_CLASS                                              0x00005070
@@ -178,6 +179,7 @@ struct nv04_display_class {
 #define NVA3_DISP_CLASS                                              0x00008570
 #define NVD0_DISP_CLASS                                              0x00009070
 #define NVE0_DISP_CLASS                                              0x00009170
+#define NVF0_DISP_CLASS                                              0x00009270
 
 #define NV50_DISP_SOR_MTHD                                           0x00010000
 #define NV50_DISP_SOR_MTHD_TYPE                                      0x0000f000
@@ -246,6 +248,7 @@ struct nv50_display_class {
  * 857a: NVA3_DISP_CURS
  * 907a: NVD0_DISP_CURS
  * 917a: NVE0_DISP_CURS
+ * 927a: NVF0_DISP_CURS
  */
 
 #define NV50_DISP_CURS_CLASS                                         0x0000507a
@@ -255,6 +258,7 @@ struct nv50_display_class {
 #define NVA3_DISP_CURS_CLASS                                         0x0000857a
 #define NVD0_DISP_CURS_CLASS                                         0x0000907a
 #define NVE0_DISP_CURS_CLASS                                         0x0000917a
+#define NVF0_DISP_CURS_CLASS                                         0x0000927a
 
 struct nv50_display_curs_class {
 	u32 head;
@@ -267,6 +271,7 @@ struct nv50_display_curs_class {
  * 857b: NVA3_DISP_OIMM
  * 907b: NVD0_DISP_OIMM
  * 917b: NVE0_DISP_OIMM
+ * 927b: NVE0_DISP_OIMM
  */
 
 #define NV50_DISP_OIMM_CLASS                                         0x0000507b
@@ -276,6 +281,7 @@ struct nv50_display_curs_class {
 #define NVA3_DISP_OIMM_CLASS                                         0x0000857b
 #define NVD0_DISP_OIMM_CLASS                                         0x0000907b
 #define NVE0_DISP_OIMM_CLASS                                         0x0000917b
+#define NVF0_DISP_OIMM_CLASS                                         0x0000927b
 
 struct nv50_display_oimm_class {
 	u32 head;
@@ -288,6 +294,7 @@ struct nv50_display_oimm_class {
  * 857c: NVA3_DISP_SYNC
  * 907c: NVD0_DISP_SYNC
  * 917c: NVE0_DISP_SYNC
+ * 927c: NVF0_DISP_SYNC
  */
 
 #define NV50_DISP_SYNC_CLASS                                         0x0000507c
@@ -297,6 +304,7 @@ struct nv50_display_oimm_class {
 #define NVA3_DISP_SYNC_CLASS                                         0x0000857c
 #define NVD0_DISP_SYNC_CLASS                                         0x0000907c
 #define NVE0_DISP_SYNC_CLASS                                         0x0000917c
+#define NVF0_DISP_SYNC_CLASS                                         0x0000927c
 
 struct nv50_display_sync_class {
 	u32 pushbuf;
@@ -310,6 +318,7 @@ struct nv50_display_sync_class {
  * 857d: NVA3_DISP_MAST
  * 907d: NVD0_DISP_MAST
  * 917d: NVE0_DISP_MAST
+ * 927d: NVF0_DISP_MAST
  */
 
 #define NV50_DISP_MAST_CLASS                                         0x0000507d
@@ -319,6 +328,7 @@ struct nv50_display_sync_class {
 #define NVA3_DISP_MAST_CLASS                                         0x0000857d
 #define NVD0_DISP_MAST_CLASS                                         0x0000907d
 #define NVE0_DISP_MAST_CLASS                                         0x0000917d
+#define NVF0_DISP_MAST_CLASS                                         0x0000927d
 
 struct nv50_display_mast_class {
 	u32 pushbuf;
@@ -331,6 +341,7 @@ struct nv50_display_mast_class {
  * 857e: NVA3_DISP_OVLY
  * 907e: NVD0_DISP_OVLY
  * 917e: NVE0_DISP_OVLY
+ * 927e: NVF0_DISP_OVLY
  */
 
 #define NV50_DISP_OVLY_CLASS                                         0x0000507e
@@ -340,6 +351,7 @@ struct nv50_display_mast_class {
 #define NVA3_DISP_OVLY_CLASS                                         0x0000857e
 #define NVD0_DISP_OVLY_CLASS                                         0x0000907e
 #define NVE0_DISP_OVLY_CLASS                                         0x0000917e
+#define NVF0_DISP_OVLY_CLASS                                         0x0000927e
 
 struct nv50_display_ovly_class {
 	u32 pushbuf;
diff --git a/drivers/gpu/drm/nouveau/core/include/core/device.h b/drivers/gpu/drm/nouveau/core/include/core/device.h
index d351a4e5819c..05840f3eee98 100644
--- a/drivers/gpu/drm/nouveau/core/include/core/device.h
+++ b/drivers/gpu/drm/nouveau/core/include/core/device.h
@@ -6,7 +6,7 @@
 #include <core/engine.h>
 
 enum nv_subdev_type {
-	NVDEV_SUBDEV_DEVICE,
+	NVDEV_ENGINE_DEVICE,
 	NVDEV_SUBDEV_VBIOS,
 
 	/* All subdevs from DEVINIT to DEVINIT_LAST will be created before
@@ -57,7 +57,7 @@ enum nv_subdev_type {
 };
 
 struct nouveau_device {
-	struct nouveau_subdev base;
+	struct nouveau_engine base;
 	struct list_head head;
 
 	struct pci_dev *pdev;
@@ -99,7 +99,7 @@ nv_device(void *obj)
 
 #if CONFIG_NOUVEAU_DEBUG >= NV_DBG_PARANOIA
 	if (unlikely(!nv_iclass(device, NV_SUBDEV_CLASS) ||
-		     (nv_hclass(device) & 0xff) != NVDEV_SUBDEV_DEVICE)) {
+		     (nv_hclass(device) & 0xff) != NVDEV_ENGINE_DEVICE)) {
 		nv_assert("BAD CAST -> NvDevice, 0x%08x 0x%08x",
 			  nv_hclass(object), nv_hclass(device));
 	}
diff --git a/drivers/gpu/drm/nouveau/core/include/core/parent.h b/drivers/gpu/drm/nouveau/core/include/core/parent.h
index 31cd852c96df..9f5ea900ff00 100644
--- a/drivers/gpu/drm/nouveau/core/include/core/parent.h
+++ b/drivers/gpu/drm/nouveau/core/include/core/parent.h
@@ -51,8 +51,8 @@ int  nouveau_parent_create_(struct nouveau_object *, struct nouveau_object *,
 void nouveau_parent_destroy(struct nouveau_parent *);
 
 void _nouveau_parent_dtor(struct nouveau_object *);
-#define _nouveau_parent_init _nouveau_object_init
-#define _nouveau_parent_fini _nouveau_object_fini
+#define _nouveau_parent_init nouveau_object_init
+#define _nouveau_parent_fini nouveau_object_fini
 
 int nouveau_parent_sclass(struct nouveau_object *, u16 handle,
 			  struct nouveau_object **pengine,
diff --git a/drivers/gpu/drm/nouveau/core/include/subdev/device.h b/drivers/gpu/drm/nouveau/core/include/engine/device.h
index c9e4c4afa50e..b3dd2c4c2f1e 100644
--- a/drivers/gpu/drm/nouveau/core/include/subdev/device.h
+++ b/drivers/gpu/drm/nouveau/core/include/engine/device.h
@@ -18,7 +18,6 @@ int nv50_identify(struct nouveau_device *);
 int nvc0_identify(struct nouveau_device *);
 int nve0_identify(struct nouveau_device *);
 
-extern struct nouveau_oclass nouveau_device_sclass[];
 struct nouveau_device *nouveau_device_find(u64 name);
 
 #endif
diff --git a/drivers/gpu/drm/nouveau/core/include/engine/disp.h b/drivers/gpu/drm/nouveau/core/include/engine/disp.h
index 28da6772c095..4b21fabfbddb 100644
--- a/drivers/gpu/drm/nouveau/core/include/engine/disp.h
+++ b/drivers/gpu/drm/nouveau/core/include/engine/disp.h
@@ -44,5 +44,6 @@ extern struct nouveau_oclass nv94_disp_oclass;
 extern struct nouveau_oclass nva3_disp_oclass;
 extern struct nouveau_oclass nvd0_disp_oclass;
 extern struct nouveau_oclass nve0_disp_oclass;
+extern struct nouveau_oclass nvf0_disp_oclass;
 
 #endif
diff --git a/drivers/gpu/drm/nouveau/core/include/engine/fifo.h b/drivers/gpu/drm/nouveau/core/include/engine/fifo.h
index b46c197709f3..633c2f806482 100644
--- a/drivers/gpu/drm/nouveau/core/include/engine/fifo.h
+++ b/drivers/gpu/drm/nouveau/core/include/engine/fifo.h
@@ -65,7 +65,8 @@ struct nouveau_fifo_base {
 struct nouveau_fifo {
 	struct nouveau_engine base;
 
-	struct nouveau_event *uevent;
+	struct nouveau_event *cevent; /* channel creation event */
+	struct nouveau_event *uevent; /* async user trigger */
 
 	struct nouveau_object **channel;
 	spinlock_t lock;
diff --git a/drivers/gpu/drm/nouveau/core/include/engine/graph.h b/drivers/gpu/drm/nouveau/core/include/engine/graph.h
index 6943b40d0817..5d392439f2ac 100644
--- a/drivers/gpu/drm/nouveau/core/include/engine/graph.h
+++ b/drivers/gpu/drm/nouveau/core/include/engine/graph.h
@@ -26,6 +26,10 @@ struct nouveau_graph_chan {
 
 struct nouveau_graph {
 	struct nouveau_engine base;
+
+	/* Returns chipset-specific counts of units packed into an u64.
+	 */
+	u64 (*units)(struct nouveau_graph *);
 };
 
 static inline struct nouveau_graph *
diff --git a/drivers/gpu/drm/nouveau/core/include/subdev/ltcg.h b/drivers/gpu/drm/nouveau/core/include/subdev/ltcg.h
index f351f63bc654..a1985ed3d58d 100644
--- a/drivers/gpu/drm/nouveau/core/include/subdev/ltcg.h
+++ b/drivers/gpu/drm/nouveau/core/include/subdev/ltcg.h
@@ -4,8 +4,15 @@
 #include <core/subdev.h>
 #include <core/device.h>
 
+struct nouveau_mm_node;
+
 struct nouveau_ltcg {
 	struct nouveau_subdev base;
+
+	int  (*tags_alloc)(struct nouveau_ltcg *, u32 count,
+	                   struct nouveau_mm_node **);
+	void (*tags_free)(struct nouveau_ltcg *, struct nouveau_mm_node **);
+	void (*tags_clear)(struct nouveau_ltcg *, u32 first, u32 count);
 };
 
 static inline struct nouveau_ltcg *
diff --git a/drivers/gpu/drm/nouveau/core/include/subdev/mc.h b/drivers/gpu/drm/nouveau/core/include/subdev/mc.h
index fded97cea500..d5502267c30f 100644
--- a/drivers/gpu/drm/nouveau/core/include/subdev/mc.h
+++ b/drivers/gpu/drm/nouveau/core/include/subdev/mc.h
@@ -21,18 +21,22 @@ nouveau_mc(void *obj)
 }
 
 #define nouveau_mc_create(p,e,o,d)                                             \
-	nouveau_subdev_create_((p), (e), (o), 0, "PMC", "master",              \
-			       sizeof(**d), (void **)d)
-#define nouveau_mc_destroy(p)                                                  \
-	nouveau_subdev_destroy(&(p)->base)
-#define nouveau_mc_init(p)                                                     \
-	nouveau_subdev_init(&(p)->base)
-#define nouveau_mc_fini(p,s)                                                   \
-	nouveau_subdev_fini(&(p)->base, (s))
-
-#define _nouveau_mc_dtor _nouveau_subdev_dtor
-#define _nouveau_mc_init _nouveau_subdev_init
-#define _nouveau_mc_fini _nouveau_subdev_fini
+	nouveau_mc_create_((p), (e), (o), sizeof(**d), (void **)d)
+#define nouveau_mc_destroy(p) ({                                               \
+	struct nouveau_mc *pmc = (p); _nouveau_mc_dtor(nv_object(pmc));        \
+})
+#define nouveau_mc_init(p) ({                                                  \
+	struct nouveau_mc *pmc = (p); _nouveau_mc_init(nv_object(pmc));        \
+})
+#define nouveau_mc_fini(p,s) ({                                                \
+	struct nouveau_mc *pmc = (p); _nouveau_mc_fini(nv_object(pmc), (s));   \
+})
+
+int  nouveau_mc_create_(struct nouveau_object *, struct nouveau_object *,
+			struct nouveau_oclass *, int, void **);
+void _nouveau_mc_dtor(struct nouveau_object *);
+int  _nouveau_mc_init(struct nouveau_object *);
+int  _nouveau_mc_fini(struct nouveau_object *, bool);
 
 extern struct nouveau_oclass nv04_mc_oclass;
 extern struct nouveau_oclass nv44_mc_oclass;
@@ -40,8 +44,6 @@ extern struct nouveau_oclass nv50_mc_oclass;
 extern struct nouveau_oclass nv98_mc_oclass;
 extern struct nouveau_oclass nvc0_mc_oclass;
 
-void nouveau_mc_intr(struct nouveau_subdev *);
-
 extern const struct nouveau_mc_intr nv04_mc_intr[];
 int nv04_mc_init(struct nouveau_object *);
 int nv50_mc_init(struct nouveau_object *);
diff --git a/drivers/gpu/drm/nouveau/core/include/subdev/therm.h b/drivers/gpu/drm/nouveau/core/include/subdev/therm.h
index 0b20fc0d19c1..c075998d82e6 100644
--- a/drivers/gpu/drm/nouveau/core/include/subdev/therm.h
+++ b/drivers/gpu/drm/nouveau/core/include/subdev/therm.h
@@ -73,6 +73,7 @@ int  _nouveau_therm_fini(struct nouveau_object *, bool);
 
 extern struct nouveau_oclass nv40_therm_oclass;
 extern struct nouveau_oclass nv50_therm_oclass;
+extern struct nouveau_oclass nv84_therm_oclass;
 extern struct nouveau_oclass nva3_therm_oclass;
 extern struct nouveau_oclass nvd0_therm_oclass;
 
diff --git a/drivers/gpu/drm/nouveau/core/os.h b/drivers/gpu/drm/nouveau/core/os.h
index eb496033b55c..3bd9be2ab37f 100644
--- a/drivers/gpu/drm/nouveau/core/os.h
+++ b/drivers/gpu/drm/nouveau/core/os.h
@@ -17,6 +17,7 @@
 #include <linux/acpi.h>
 #include <linux/dmi.h>
 #include <linux/reboot.h>
+#include <linux/interrupt.h>
 
 #include <asm/unaligned.h>
 
diff --git a/drivers/gpu/drm/nouveau/core/subdev/bar/nv50.c b/drivers/gpu/drm/nouveau/core/subdev/bar/nv50.c
index c3acf5b70d9e..649f1ced1fe0 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/bar/nv50.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/bar/nv50.c
@@ -122,18 +122,20 @@ nv50_bar_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	if (ret)
 		return ret;
 
-	ret = nouveau_gpuobj_new(parent, NULL, 0x20000, 0, NVOBJ_FLAG_HEAP,
-				&priv->mem);
+	ret = nouveau_gpuobj_new(nv_object(priv), NULL, 0x20000, 0,
+				 NVOBJ_FLAG_HEAP, &priv->mem);
 	heap = nv_object(priv->mem);
 	if (ret)
 		return ret;
 
-	ret = nouveau_gpuobj_new(parent, heap, (device->chipset == 0x50) ?
-				 0x1400 : 0x0200, 0, 0, &priv->pad);
+	ret = nouveau_gpuobj_new(nv_object(priv), heap,
+				(device->chipset == 0x50) ? 0x1400 : 0x0200,
+				 0, 0, &priv->pad);
 	if (ret)
 		return ret;
 
-	ret = nouveau_gpuobj_new(parent, heap, 0x4000, 0, 0, &priv->pgd);
+	ret = nouveau_gpuobj_new(nv_object(priv), heap, 0x4000, 0,
+				 0, &priv->pgd);
 	if (ret)
 		return ret;
 
@@ -145,9 +147,9 @@ nv50_bar_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	if (ret)
 		return ret;
 
-	ret = nouveau_gpuobj_new(parent, heap, ((limit-- - start) >> 12) * 8,
-				 0x1000, NVOBJ_FLAG_ZERO_ALLOC,
-				 &vm->pgt[0].obj[0]);
+	ret = nouveau_gpuobj_new(nv_object(priv), heap,
+				 ((limit-- - start) >> 12) * 8, 0x1000,
+				 NVOBJ_FLAG_ZERO_ALLOC, &vm->pgt[0].obj[0]);
 	vm->pgt[0].refcount[0] = 1;
 	if (ret)
 		return ret;
@@ -157,7 +159,7 @@ nv50_bar_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	if (ret)
 		return ret;
 
-	ret = nouveau_gpuobj_new(parent, heap, 24, 16, 0, &priv->bar3);
+	ret = nouveau_gpuobj_new(nv_object(priv), heap, 24, 16, 0, &priv->bar3);
 	if (ret)
 		return ret;
 
@@ -182,7 +184,7 @@ nv50_bar_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	if (ret)
 		return ret;
 
-	ret = nouveau_gpuobj_new(parent, heap, 24, 16, 0, &priv->bar1);
+	ret = nouveau_gpuobj_new(nv_object(priv), heap, 24, 16, 0, &priv->bar1);
 	if (ret)
 		return ret;
 
diff --git a/drivers/gpu/drm/nouveau/core/subdev/bar/nvc0.c b/drivers/gpu/drm/nouveau/core/subdev/bar/nvc0.c
index 77a6fb725d3f..f8a44956dec1 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/bar/nvc0.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/bar/nvc0.c
@@ -101,12 +101,14 @@ nvc0_bar_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 		return ret;
 
 	/* BAR3 */
-	ret = nouveau_gpuobj_new(parent, NULL, 0x1000, 0, 0, &priv->bar[0].mem);
+	ret = nouveau_gpuobj_new(nv_object(priv), NULL, 0x1000, 0, 0,
+				&priv->bar[0].mem);
 	mem = priv->bar[0].mem;
 	if (ret)
 		return ret;
 
-	ret = nouveau_gpuobj_new(parent, NULL, 0x8000, 0, 0, &priv->bar[0].pgd);
+	ret = nouveau_gpuobj_new(nv_object(priv), NULL, 0x8000, 0, 0,
+				&priv->bar[0].pgd);
 	if (ret)
 		return ret;
 
@@ -114,7 +116,7 @@ nvc0_bar_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	if (ret)
 		return ret;
 
-	ret = nouveau_gpuobj_new(parent, NULL,
+	ret = nouveau_gpuobj_new(nv_object(priv), NULL,
 				 (pci_resource_len(pdev, 3) >> 12) * 8,
 				 0x1000, NVOBJ_FLAG_ZERO_ALLOC,
 				 &vm->pgt[0].obj[0]);
@@ -133,12 +135,14 @@ nvc0_bar_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	nv_wo32(mem, 0x020c, upper_32_bits(pci_resource_len(pdev, 3) - 1));
 
 	/* BAR1 */
-	ret = nouveau_gpuobj_new(parent, NULL, 0x1000, 0, 0, &priv->bar[1].mem);
+	ret = nouveau_gpuobj_new(nv_object(priv), NULL, 0x1000, 0, 0,
+				&priv->bar[1].mem);
 	mem = priv->bar[1].mem;
 	if (ret)
 		return ret;
 
-	ret = nouveau_gpuobj_new(parent, NULL, 0x8000, 0, 0, &priv->bar[1].pgd);
+	ret = nouveau_gpuobj_new(nv_object(priv), NULL, 0x8000, 0, 0,
+				&priv->bar[1].pgd);
 	if (ret)
 		return ret;
 
diff --git a/drivers/gpu/drm/nouveau/core/subdev/bios/init.c b/drivers/gpu/drm/nouveau/core/subdev/bios/init.c
index 9c41b58d57e2..c300b5e7b670 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/bios/init.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/bios/init.c
@@ -64,27 +64,33 @@ init_exec_force(struct nvbios_init *init, bool exec)
 static inline int
 init_or(struct nvbios_init *init)
 {
-	if (init->outp)
-		return ffs(init->outp->or) - 1;
-	error("script needs OR!!\n");
+	if (init_exec(init)) {
+		if (init->outp)
+			return ffs(init->outp->or) - 1;
+		error("script needs OR!!\n");
+	}
 	return 0;
 }
 
 static inline int
 init_link(struct nvbios_init *init)
 {
-	if (init->outp)
-		return !(init->outp->sorconf.link & 1);
-	error("script needs OR link\n");
+	if (init_exec(init)) {
+		if (init->outp)
+			return !(init->outp->sorconf.link & 1);
+		error("script needs OR link\n");
+	}
 	return 0;
 }
 
 static inline int
 init_crtc(struct nvbios_init *init)
 {
-	if (init->crtc >= 0)
-		return init->crtc;
-	error("script needs crtc\n");
+	if (init_exec(init)) {
+		if (init->crtc >= 0)
+			return init->crtc;
+		error("script needs crtc\n");
+	}
 	return 0;
 }
 
@@ -92,16 +98,21 @@ static u8
 init_conn(struct nvbios_init *init)
 {
 	struct nouveau_bios *bios = init->bios;
+	u8  ver, len;
+	u16 conn;
 
-	if (init->outp) {
-		u8  ver, len;
-		u16 conn = dcb_conn(bios, init->outp->connector, &ver, &len);
-		if (conn)
-			return nv_ro08(bios, conn);
+	if (init_exec(init)) {
+		if (init->outp) {
+			conn = init->outp->connector;
+			conn = dcb_conn(bios, conn, &ver, &len);
+			if (conn)
+				return nv_ro08(bios, conn);
+		}
+
+		error("script needs connector type\n");
 	}
 
-	error("script needs connector type\n");
-	return 0x00;
+	return 0xff;
 }
 
 static inline u32
@@ -227,7 +238,8 @@ init_i2c(struct nvbios_init *init, int index)
 	} else
 	if (index < 0) {
 		if (!init->outp) {
-			error("script needs output for i2c\n");
+			if (init_exec(init))
+				error("script needs output for i2c\n");
 			return NULL;
 		}
 
@@ -544,7 +556,8 @@ init_tmds_reg(struct nvbios_init *init, u8 tmds)
 			return 0x6808b0 + dacoffset;
 		}
 
-		error("tmds opcodes need dcb\n");
+		if (init_exec(init))
+			error("tmds opcodes need dcb\n");
 	} else {
 		if (tmds < ARRAY_SIZE(pramdac_table))
 			return pramdac_table[tmds];
@@ -792,7 +805,8 @@ init_dp_condition(struct nvbios_init *init)
 			break;
 		}
 
-		warn("script needs dp output table data\n");
+		if (init_exec(init))
+			warn("script needs dp output table data\n");
 		break;
 	case 5:
 		if (!(init_rdauxr(init, 0x0d) & 1))
@@ -816,7 +830,7 @@ init_io_mask_or(struct nvbios_init *init)
 	u8    or = init_or(init);
 	u8  data;
 
-	trace("IO_MASK_OR\t0x03d4[0x%02x] &= ~(1 << 0x%02x)", index, or);
+	trace("IO_MASK_OR\t0x03d4[0x%02x] &= ~(1 << 0x%02x)\n", index, or);
 	init->offset += 2;
 
 	data = init_rdvgai(init, 0x03d4, index);
@@ -835,7 +849,7 @@ init_io_or(struct nvbios_init *init)
 	u8    or = init_or(init);
 	u8  data;
 
-	trace("IO_OR\t0x03d4[0x%02x] |= (1 << 0x%02x)", index, or);
+	trace("IO_OR\t0x03d4[0x%02x] |= (1 << 0x%02x)\n", index, or);
 	init->offset += 2;
 
 	data = init_rdvgai(init, 0x03d4, index);
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fb/nvc0.c b/drivers/gpu/drm/nouveau/core/subdev/fb/nvc0.c
index 7606ed15b6fa..86ad59203c8b 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/fb/nvc0.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/fb/nvc0.c
@@ -23,6 +23,7 @@
  */
 
 #include <subdev/fb.h>
+#include <subdev/ltcg.h>
 #include <subdev/bios.h>
 
 struct nvc0_fb_priv {
@@ -31,34 +32,14 @@ struct nvc0_fb_priv {
 	dma_addr_t r100c10;
 };
 
-/* 0 = unsupported
- * 1 = non-compressed
- * 3 = compressed
- */
-static const u8 types[256] = {
-	1, 1, 3, 3, 3, 3, 0, 3, 3, 3, 3, 0, 0, 0, 0, 0,
-	0, 1, 0, 0, 0, 0, 0, 3, 3, 3, 3, 0, 0, 0, 0, 0,
-	0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0,
-	0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 3, 3, 3, 3, 3,
-	3, 3, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0,
-	0, 3, 3, 3, 3, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
-	0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
-	0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0,
-	0, 0, 0, 0, 0, 0, 3, 3, 3, 3, 0, 1, 1, 1, 1, 0,
-	0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
-	0, 0, 0, 3, 3, 3, 3, 1, 1, 1, 1, 0, 0, 0, 0, 0,
-	0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3,
-	3, 3, 3, 1, 0, 0, 0, 0, 0, 0, 0, 0, 3, 3, 3, 3,
-	3, 3, 0, 0, 0, 0, 0, 0, 3, 0, 0, 3, 0, 3, 0, 3,
-	3, 0, 3, 3, 3, 3, 3, 0, 0, 3, 0, 3, 0, 3, 3, 0,
-	3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 0, 1, 1, 0
-};
+extern const u8 nvc0_pte_storage_type_map[256];
+
 
 static bool
 nvc0_fb_memtype_valid(struct nouveau_fb *pfb, u32 tile_flags)
 {
 	u8 memtype = (tile_flags & 0x0000ff00) >> 8;
-	return likely((types[memtype] == 1));
+	return likely((nvc0_pte_storage_type_map[memtype] != 0xff));
 }
 
 static int
@@ -130,6 +111,7 @@ nvc0_fb_vram_new(struct nouveau_fb *pfb, u64 size, u32 align, u32 ncmin,
 	int type = (memtype & 0x0ff);
 	int back = (memtype & 0x800);
 	int ret;
+	const bool comp = nvc0_pte_storage_type_map[type] != type;
 
 	size  >>= 12;
 	align >>= 12;
@@ -142,10 +124,22 @@ nvc0_fb_vram_new(struct nouveau_fb *pfb, u64 size, u32 align, u32 ncmin,
 		return -ENOMEM;
 
 	INIT_LIST_HEAD(&mem->regions);
-	mem->memtype = type;
 	mem->size = size;
 
 	mutex_lock(&pfb->base.mutex);
+	if (comp) {
+		struct nouveau_ltcg *ltcg = nouveau_ltcg(pfb->base.base.parent);
+
+		/* compression only works with lpages */
+		if (align == (1 << (17 - 12))) {
+			int n = size >> 5;
+			ltcg->tags_alloc(ltcg, n, &mem->tag);
+		}
+		if (unlikely(!mem->tag))
+			type = nvc0_pte_storage_type_map[type];
+	}
+	mem->memtype = type;
+
 	do {
 		if (back)
 			ret = nouveau_mm_tail(mm, 1, size, ncmin, align, &r);
@@ -168,6 +162,17 @@ nvc0_fb_vram_new(struct nouveau_fb *pfb, u64 size, u32 align, u32 ncmin,
 	return 0;
 }
 
+static void
+nvc0_fb_vram_del(struct nouveau_fb *pfb, struct nouveau_mem **pmem)
+{
+	struct nouveau_ltcg *ltcg = nouveau_ltcg(pfb->base.base.parent);
+
+	if ((*pmem)->tag)
+		ltcg->tags_free(ltcg, &(*pmem)->tag);
+
+	nv50_fb_vram_del(pfb, pmem);
+}
+
 static int
 nvc0_fb_init(struct nouveau_object *object)
 {
@@ -178,7 +183,8 @@ nvc0_fb_init(struct nouveau_object *object)
 	if (ret)
 		return ret;
 
-	nv_wr32(priv, 0x100c10, priv->r100c10 >> 8);
+	if (priv->r100c10_page)
+		nv_wr32(priv, 0x100c10, priv->r100c10 >> 8);
 	return 0;
 }
 
@@ -214,16 +220,16 @@ nvc0_fb_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	priv->base.memtype_valid = nvc0_fb_memtype_valid;
 	priv->base.ram.init = nvc0_fb_vram_init;
 	priv->base.ram.get = nvc0_fb_vram_new;
-	priv->base.ram.put = nv50_fb_vram_del;
+	priv->base.ram.put = nvc0_fb_vram_del;
 
 	priv->r100c10_page = alloc_page(GFP_KERNEL | __GFP_ZERO);
-	if (!priv->r100c10_page)
-		return -ENOMEM;
-
-	priv->r100c10 = pci_map_page(device->pdev, priv->r100c10_page, 0,
-				     PAGE_SIZE, PCI_DMA_BIDIRECTIONAL);
-	if (pci_dma_mapping_error(device->pdev, priv->r100c10))
-		return -EFAULT;
+	if (priv->r100c10_page) {
+		priv->r100c10 = pci_map_page(device->pdev, priv->r100c10_page,
+					     0, PAGE_SIZE,
+					     PCI_DMA_BIDIRECTIONAL);
+		if (pci_dma_mapping_error(device->pdev, priv->r100c10))
+			return -EFAULT;
+	}
 
 	return nouveau_fb_preinit(&priv->base);
 }
diff --git a/drivers/gpu/drm/nouveau/core/subdev/i2c/base.c b/drivers/gpu/drm/nouveau/core/subdev/i2c/base.c
index 2e98e8a3f1aa..8ae2625415e1 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/i2c/base.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/i2c/base.c
@@ -140,12 +140,8 @@ nouveau_i2c_port_create_(struct nouveau_object *parent,
 	}
 
 	/* drop port's i2c subdev refcount, i2c handles this itself */
-	if (ret == 0) {
+	if (ret == 0)
 		list_add_tail(&port->head, &i2c->ports);
-		atomic_dec(&parent->refcount);
-		atomic_dec(&engine->refcount);
-	}
-
 	return ret;
 }
 
diff --git a/drivers/gpu/drm/nouveau/core/subdev/instmem/nv04.c b/drivers/gpu/drm/nouveau/core/subdev/instmem/nv04.c
index f5bbd3834116..795393d7b2f5 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/instmem/nv04.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/instmem/nv04.c
@@ -93,7 +93,6 @@ nv04_instmem_alloc(struct nouveau_instmem *imem, struct nouveau_object *parent,
 		   u32 size, u32 align, struct nouveau_object **pobject)
 {
 	struct nouveau_object *engine = nv_object(imem);
-	struct nv04_instmem_priv *priv = (void *)(imem);
 	int ret;
 
 	ret = nouveau_object_ctor(parent, engine, &nv04_instobj_oclass,
@@ -101,14 +100,6 @@ nv04_instmem_alloc(struct nouveau_instmem *imem, struct nouveau_object *parent,
 	if (ret)
 		return ret;
 
-	/* INSTMEM itself creates objects to reserve (and preserve across
-	 * suspend/resume) various fixed data locations, each one of these
-	 * takes a reference on INSTMEM itself, causing it to never be
-	 * freed.  We drop all the self-references here to avoid this.
-	 */
-	if (unlikely(!priv->created))
-		atomic_dec(&engine->refcount);
-
 	return 0;
 }
 
@@ -134,27 +125,28 @@ nv04_instmem_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 		return ret;
 
 	/* 0x00000-0x10000: reserve for probable vbios image */
-	ret = nouveau_gpuobj_new(parent, NULL, 0x10000, 0, 0, &priv->vbios);
+	ret = nouveau_gpuobj_new(nv_object(priv), NULL, 0x10000, 0, 0,
+				&priv->vbios);
 	if (ret)
 		return ret;
 
 	/* 0x10000-0x18000: reserve for RAMHT */
-	ret = nouveau_ramht_new(parent, NULL, 0x08000, 0, &priv->ramht);
+	ret = nouveau_ramht_new(nv_object(priv), NULL, 0x08000, 0, &priv->ramht);
 	if (ret)
 		return ret;
 
 	/* 0x18000-0x18800: reserve for RAMFC (enough for 32 nv30 channels) */
-	ret = nouveau_gpuobj_new(parent, NULL, 0x00800, 0,
+	ret = nouveau_gpuobj_new(nv_object(priv), NULL, 0x00800, 0,
 				 NVOBJ_FLAG_ZERO_ALLOC, &priv->ramfc);
 	if (ret)
 		return ret;
 
 	/* 0x18800-0x18a00: reserve for RAMRO */
-	ret = nouveau_gpuobj_new(parent, NULL, 0x00200, 0, 0, &priv->ramro);
+	ret = nouveau_gpuobj_new(nv_object(priv), NULL, 0x00200, 0, 0,
+				&priv->ramro);
 	if (ret)
 		return ret;
 
-	priv->created = true;
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/nouveau/core/subdev/instmem/nv04.h b/drivers/gpu/drm/nouveau/core/subdev/instmem/nv04.h
index 7983d8d9b358..b15b61310236 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/instmem/nv04.h
+++ b/drivers/gpu/drm/nouveau/core/subdev/instmem/nv04.h
@@ -9,7 +9,6 @@
 
 struct nv04_instmem_priv {
 	struct nouveau_instmem base;
-	bool created;
 
 	void __iomem *iomem;
 	struct nouveau_mm heap;
diff --git a/drivers/gpu/drm/nouveau/core/subdev/instmem/nv40.c b/drivers/gpu/drm/nouveau/core/subdev/instmem/nv40.c
index da64253201ef..716bf41bc3c1 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/instmem/nv40.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/instmem/nv40.c
@@ -82,31 +82,33 @@ nv40_instmem_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 		return ret;
 
 	/* 0x00000-0x10000: reserve for probable vbios image */
-	ret = nouveau_gpuobj_new(parent, NULL, 0x10000, 0, 0, &priv->vbios);
+	ret = nouveau_gpuobj_new(nv_object(priv), NULL, 0x10000, 0, 0,
+				&priv->vbios);
 	if (ret)
 		return ret;
 
 	/* 0x10000-0x18000: reserve for RAMHT */
-	ret = nouveau_ramht_new(parent, NULL, 0x08000, 0, &priv->ramht);
+	ret = nouveau_ramht_new(nv_object(priv), NULL, 0x08000, 0,
+			       &priv->ramht);
 	if (ret)
 		return ret;
 
 	/* 0x18000-0x18200: reserve for RAMRO
 	 * 0x18200-0x20000: padding
 	 */
-	ret = nouveau_gpuobj_new(parent, NULL, 0x08000, 0, 0, &priv->ramro);
+	ret = nouveau_gpuobj_new(nv_object(priv), NULL, 0x08000, 0, 0,
+				&priv->ramro);
 	if (ret)
 		return ret;
 
 	/* 0x20000-0x21000: reserve for RAMFC
 	 * 0x21000-0x40000: padding and some unknown crap
 	 */
-	ret = nouveau_gpuobj_new(parent, NULL, 0x20000, 0,
+	ret = nouveau_gpuobj_new(nv_object(priv), NULL, 0x20000, 0,
 				 NVOBJ_FLAG_ZERO_ALLOC, &priv->ramfc);
 	if (ret)
 		return ret;
 
-	priv->created = true;
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/nouveau/core/subdev/ltcg/nvc0.c b/drivers/gpu/drm/nouveau/core/subdev/ltcg/nvc0.c
index 078a2b9d6bd6..e4940fb166e8 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/ltcg/nvc0.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/ltcg/nvc0.c
@@ -23,10 +23,17 @@
  */
 
 #include <subdev/ltcg.h>
+#include <subdev/fb.h>
+#include <subdev/timer.h>
 
 struct nvc0_ltcg_priv {
 	struct nouveau_ltcg base;
+	u32 part_nr;
+	u32 part_mask;
 	u32 subp_nr;
+	struct nouveau_mm tags;
+	u32 num_tags;
+	struct nouveau_mm_node *tag_ram;
 };
 
 static void
@@ -62,11 +69,104 @@ nvc0_ltcg_intr(struct nouveau_subdev *subdev)
 }
 
 static int
+nvc0_ltcg_tags_alloc(struct nouveau_ltcg *ltcg, u32 n,
+		     struct nouveau_mm_node **pnode)
+{
+	struct nvc0_ltcg_priv *priv = (struct nvc0_ltcg_priv *)ltcg;
+	int ret;
+
+	ret = nouveau_mm_head(&priv->tags, 1, n, n, 1, pnode);
+	if (ret)
+		*pnode = NULL;
+
+	return ret;
+}
+
+static void
+nvc0_ltcg_tags_free(struct nouveau_ltcg *ltcg, struct nouveau_mm_node **pnode)
+{
+	struct nvc0_ltcg_priv *priv = (struct nvc0_ltcg_priv *)ltcg;
+
+	nouveau_mm_free(&priv->tags, pnode);
+}
+
+static void
+nvc0_ltcg_tags_clear(struct nouveau_ltcg *ltcg, u32 first, u32 count)
+{
+	struct nvc0_ltcg_priv *priv = (struct nvc0_ltcg_priv *)ltcg;
+	u32 last = first + count - 1;
+	int p, i;
+
+	BUG_ON((first > last) || (last >= priv->num_tags));
+
+	nv_wr32(priv, 0x17e8cc, first);
+	nv_wr32(priv, 0x17e8d0, last);
+	nv_wr32(priv, 0x17e8c8, 0x4); /* trigger clear */
+
+	/* wait until it's finished with clearing */
+	for (p = 0; p < priv->part_nr; ++p) {
+		if (!(priv->part_mask & (1 << p)))
+			continue;
+		for (i = 0; i < priv->subp_nr; ++i)
+			nv_wait(priv, 0x1410c8 + p * 0x2000 + i * 0x400, ~0, 0);
+	}
+}
+
+/* TODO: Figure out tag memory details and drop the over-cautious allocation.
+ */
+static int
+nvc0_ltcg_init_tag_ram(struct nouveau_fb *pfb, struct nvc0_ltcg_priv *priv)
+{
+	u32 tag_size, tag_margin, tag_align;
+	int ret;
+
+	nv_wr32(priv, 0x17e8d8, priv->part_nr);
+
+	/* tags for 1/4 of VRAM should be enough (8192/4 per GiB of VRAM) */
+	priv->num_tags = (pfb->ram.size >> 17) / 4;
+	if (priv->num_tags > (1 << 17))
+		priv->num_tags = 1 << 17; /* we have 17 bits in PTE */
+	priv->num_tags = (priv->num_tags + 63) & ~63; /* round up to 64 */
+
+	tag_align = priv->part_nr * 0x800;
+	tag_margin = (tag_align < 0x6000) ? 0x6000 : tag_align;
+
+	/* 4 part 4 sub: 0x2000 bytes for 56 tags */
+	/* 3 part 4 sub: 0x6000 bytes for 168 tags */
+	/*
+	 * About 147 bytes per tag. Let's be safe and allocate x2, which makes
+	 * 0x4980 bytes for 64 tags, and round up to 0x6000 bytes for 64 tags.
+	 *
+	 * For 4 GiB of memory we'll have 8192 tags which makes 3 MiB, < 0.1 %.
+	 */
+	tag_size  = (priv->num_tags / 64) * 0x6000 + tag_margin;
+	tag_size += tag_align;
+	tag_size  = (tag_size + 0xfff) >> 12; /* round up */
+
+	ret = nouveau_mm_tail(&pfb->vram, 0, tag_size, tag_size, 1,
+	                      &priv->tag_ram);
+	if (ret) {
+		priv->num_tags = 0;
+	} else {
+		u64 tag_base = (priv->tag_ram->offset << 12) + tag_margin;
+
+		tag_base += tag_align - 1;
+		ret = do_div(tag_base, tag_align);
+
+		nv_wr32(priv, 0x17e8d4, tag_base);
+	}
+	ret = nouveau_mm_init(&priv->tags, 0, priv->num_tags, 1);
+
+	return ret;
+}
+
+static int
 nvc0_ltcg_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	       struct nouveau_oclass *oclass, void *data, u32 size,
 	       struct nouveau_object **pobject)
 {
 	struct nvc0_ltcg_priv *priv;
+	struct nouveau_fb *pfb = nouveau_fb(parent);
 	int ret;
 
 	ret = nouveau_ltcg_create(parent, engine, oclass, &priv);
@@ -74,19 +174,44 @@ nvc0_ltcg_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	if (ret)
 		return ret;
 
-	priv->subp_nr = nv_rd32(priv, 0x17e8dc) >> 24;
+	priv->part_nr = nv_rd32(priv, 0x022438);
+	priv->part_mask = nv_rd32(priv, 0x022554);
+
+	priv->subp_nr = nv_rd32(priv, 0x17e8dc) >> 28;
+
 	nv_mask(priv, 0x17e820, 0x00100000, 0x00000000); /* INTR_EN &= ~0x10 */
 
+	ret = nvc0_ltcg_init_tag_ram(pfb, priv);
+	if (ret)
+		return ret;
+
+	priv->base.tags_alloc = nvc0_ltcg_tags_alloc;
+	priv->base.tags_free  = nvc0_ltcg_tags_free;
+	priv->base.tags_clear = nvc0_ltcg_tags_clear;
+
 	nv_subdev(priv)->intr = nvc0_ltcg_intr;
 	return 0;
 }
 
+static void
+nvc0_ltcg_dtor(struct nouveau_object *object)
+{
+	struct nouveau_ltcg *ltcg = (struct nouveau_ltcg *)object;
+	struct nvc0_ltcg_priv *priv = (struct nvc0_ltcg_priv *)ltcg;
+	struct nouveau_fb *pfb = nouveau_fb(ltcg->base.base.parent);
+
+	nouveau_mm_fini(&priv->tags);
+	nouveau_mm_free(&pfb->vram, &priv->tag_ram);
+
+	nouveau_ltcg_destroy(ltcg);
+}
+
 struct nouveau_oclass
 nvc0_ltcg_oclass = {
 	.handle = NV_SUBDEV(LTCG, 0xc0),
 	.ofuncs = &(struct nouveau_ofuncs) {
 		.ctor = nvc0_ltcg_ctor,
-		.dtor = _nouveau_ltcg_dtor,
+		.dtor = nvc0_ltcg_dtor,
 		.init = _nouveau_ltcg_init,
 		.fini = _nouveau_ltcg_fini,
 	},
diff --git a/drivers/gpu/drm/nouveau/core/subdev/mc/base.c b/drivers/gpu/drm/nouveau/core/subdev/mc/base.c
index 8379aafa6e1b..1c0330b8c9a4 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/mc/base.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/mc/base.c
@@ -24,10 +24,10 @@
 
 #include <subdev/mc.h>
 
-void
-nouveau_mc_intr(struct nouveau_subdev *subdev)
+static irqreturn_t
+nouveau_mc_intr(int irq, void *arg)
 {
-	struct nouveau_mc *pmc = nouveau_mc(subdev);
+	struct nouveau_mc *pmc = arg;
 	const struct nouveau_mc_intr *map = pmc->intr_map;
 	struct nouveau_subdev *unit;
 	u32 stat, intr;
@@ -35,7 +35,7 @@ nouveau_mc_intr(struct nouveau_subdev *subdev)
 	intr = stat = nv_rd32(pmc, 0x000100);
 	while (stat && map->stat) {
 		if (stat & map->stat) {
-			unit = nouveau_subdev(subdev, map->unit);
+			unit = nouveau_subdev(pmc, map->unit);
 			if (unit && unit->intr)
 				unit->intr(unit);
 			intr &= ~map->stat;
@@ -46,4 +46,56 @@ nouveau_mc_intr(struct nouveau_subdev *subdev)
 	if (intr) {
 		nv_error(pmc, "unknown intr 0x%08x\n", stat);
 	}
+
+	return stat ? IRQ_HANDLED : IRQ_NONE;
+}
+
+int
+_nouveau_mc_fini(struct nouveau_object *object, bool suspend)
+{
+	struct nouveau_mc *pmc = (void *)object;
+	nv_wr32(pmc, 0x000140, 0x00000000);
+	return nouveau_subdev_fini(&pmc->base, suspend);
+}
+
+int
+_nouveau_mc_init(struct nouveau_object *object)
+{
+	struct nouveau_mc *pmc = (void *)object;
+	int ret = nouveau_subdev_init(&pmc->base);
+	if (ret)
+		return ret;
+	nv_wr32(pmc, 0x000140, 0x00000001);
+	return 0;
+}
+
+void
+_nouveau_mc_dtor(struct nouveau_object *object)
+{
+	struct nouveau_device *device = nv_device(object);
+	struct nouveau_mc *pmc = (void *)object;
+	free_irq(device->pdev->irq, pmc);
+	nouveau_subdev_destroy(&pmc->base);
+}
+
+int
+nouveau_mc_create_(struct nouveau_object *parent, struct nouveau_object *engine,
+		   struct nouveau_oclass *oclass, int length, void **pobject)
+{
+	struct nouveau_device *device = nv_device(parent);
+	struct nouveau_mc *pmc;
+	int ret;
+
+	ret = nouveau_subdev_create_(parent, engine, oclass, 0, "PMC",
+				     "master", length, pobject);
+	pmc = *pobject;
+	if (ret)
+		return ret;
+
+	ret = request_irq(device->pdev->irq, nouveau_mc_intr,
+			  IRQF_SHARED, "nouveau", pmc);
+	if (ret < 0)
+		return ret;
+
+	return 0;
 }
diff --git a/drivers/gpu/drm/nouveau/core/subdev/mc/nv04.c b/drivers/gpu/drm/nouveau/core/subdev/mc/nv04.c
index 89da8fa7ea0f..8c769715227b 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/mc/nv04.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/mc/nv04.c
@@ -55,7 +55,6 @@ nv04_mc_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	if (ret)
 		return ret;
 
-	nv_subdev(priv)->intr = nouveau_mc_intr;
 	priv->base.intr_map = nv04_mc_intr;
 	return 0;
 }
diff --git a/drivers/gpu/drm/nouveau/core/subdev/mc/nv44.c b/drivers/gpu/drm/nouveau/core/subdev/mc/nv44.c
index 397d868359ad..51919371810f 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/mc/nv44.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/mc/nv44.c
@@ -41,7 +41,6 @@ nv44_mc_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	if (ret)
 		return ret;
 
-	nv_subdev(priv)->intr = nouveau_mc_intr;
 	priv->base.intr_map = nv04_mc_intr;
 	return 0;
 }
diff --git a/drivers/gpu/drm/nouveau/core/subdev/mc/nv50.c b/drivers/gpu/drm/nouveau/core/subdev/mc/nv50.c
index 5965add6daee..d796924f9930 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/mc/nv50.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/mc/nv50.c
@@ -57,7 +57,6 @@ nv50_mc_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	if (ret)
 		return ret;
 
-	nv_subdev(priv)->intr = nouveau_mc_intr;
 	priv->base.intr_map = nv50_mc_intr;
 	return 0;
 }
diff --git a/drivers/gpu/drm/nouveau/core/subdev/mc/nv98.c b/drivers/gpu/drm/nouveau/core/subdev/mc/nv98.c
index 3a80b29dce0f..e82fd21b5041 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/mc/nv98.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/mc/nv98.c
@@ -59,7 +59,6 @@ nv98_mc_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	if (ret)
 		return ret;
 
-	nv_subdev(priv)->intr = nouveau_mc_intr;
 	priv->base.intr_map = nv98_mc_intr;
 	return 0;
 }
diff --git a/drivers/gpu/drm/nouveau/core/subdev/mc/nvc0.c b/drivers/gpu/drm/nouveau/core/subdev/mc/nvc0.c
index 42bbf72023a8..737bd4b682e1 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/mc/nvc0.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/mc/nvc0.c
@@ -61,7 +61,6 @@ nvc0_mc_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	if (ret)
 		return ret;
 
-	nv_subdev(priv)->intr = nouveau_mc_intr;
 	priv->base.intr_map = nvc0_mc_intr;
 	return 0;
 }
diff --git a/drivers/gpu/drm/nouveau/core/subdev/therm/nv40.c b/drivers/gpu/drm/nouveau/core/subdev/therm/nv40.c
index a70d1b7e397b..002e51b3af93 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/therm/nv40.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/therm/nv40.c
@@ -165,7 +165,7 @@ nv40_fan_pwm_set(struct nouveau_therm *therm, int line, u32 divs, u32 duty)
 	return 0;
 }
 
-static void
+void
 nv40_therm_intr(struct nouveau_subdev *subdev)
 {
 	struct nouveau_therm *therm = nouveau_therm(subdev);
diff --git a/drivers/gpu/drm/nouveau/core/subdev/therm/nv50.c b/drivers/gpu/drm/nouveau/core/subdev/therm/nv50.c
index 86632cbd65ce..8cf7597a2182 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/therm/nv50.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/therm/nv50.c
@@ -118,145 +118,36 @@ nv50_fan_pwm_clock(struct nouveau_therm *therm)
 	return pwm_clock;
 }
 
-int
-nv50_temp_get(struct nouveau_therm *therm)
-{
-	return nv_rd32(therm, 0x20400);
-}
-
-static void
-nv50_therm_program_alarms(struct nouveau_therm *therm)
-{
-	struct nouveau_therm_priv *priv = (void *)therm;
-	struct nvbios_therm_sensor *sensor = &priv->bios_sensor;
-	unsigned long flags;
-
-	spin_lock_irqsave(&priv->sensor.alarm_program_lock, flags);
-
-	/* enable RISING and FALLING IRQs for shutdown, THRS 0, 1, 2 and 4 */
-	nv_wr32(therm, 0x20000, 0x000003ff);
-
-	/* shutdown: The computer should be shutdown when reached */
-	nv_wr32(therm, 0x20484, sensor->thrs_shutdown.hysteresis);
-	nv_wr32(therm, 0x20480, sensor->thrs_shutdown.temp);
-
-	/* THRS_1 : fan boost*/
-	nv_wr32(therm, 0x204c4, sensor->thrs_fan_boost.temp);
-
-	/* THRS_2 : critical */
-	nv_wr32(therm, 0x204c0, sensor->thrs_critical.temp);
-
-	/* THRS_4 : down clock */
-	nv_wr32(therm, 0x20414, sensor->thrs_down_clock.temp);
-	spin_unlock_irqrestore(&priv->sensor.alarm_program_lock, flags);
-
-	nv_info(therm,
-		"Programmed thresholds [ %d(%d), %d(%d), %d(%d), %d(%d) ]\n",
-		sensor->thrs_fan_boost.temp, sensor->thrs_fan_boost.hysteresis,
-		sensor->thrs_down_clock.temp,
-		sensor->thrs_down_clock.hysteresis,
-		sensor->thrs_critical.temp, sensor->thrs_critical.hysteresis,
-		sensor->thrs_shutdown.temp, sensor->thrs_shutdown.hysteresis);
-
-}
-
-/* must be called with alarm_program_lock taken ! */
 static void
-nv50_therm_threshold_hyst_emulation(struct nouveau_therm *therm,
-				   uint32_t thrs_reg, u8 status_bit,
-				   const struct nvbios_therm_threshold *thrs,
-				   enum nouveau_therm_thrs thrs_name)
+nv50_sensor_setup(struct nouveau_therm *therm)
 {
-	enum nouveau_therm_thrs_direction direction;
-	enum nouveau_therm_thrs_state prev_state, new_state;
-	int temp, cur;
-
-	prev_state = nouveau_therm_sensor_get_threshold_state(therm, thrs_name);
-	temp = nv_rd32(therm, thrs_reg);
-
-	/* program the next threshold */
-	if (temp == thrs->temp) {
-		nv_wr32(therm, thrs_reg, thrs->temp - thrs->hysteresis);
-		new_state = NOUVEAU_THERM_THRS_HIGHER;
-	} else {
-		nv_wr32(therm, thrs_reg, thrs->temp);
-		new_state = NOUVEAU_THERM_THRS_LOWER;
-	}
-
-	/* fix the state (in case someone reprogrammed the alarms) */
-	cur = therm->temp_get(therm);
-	if (new_state == NOUVEAU_THERM_THRS_LOWER && cur > thrs->temp)
-		new_state = NOUVEAU_THERM_THRS_HIGHER;
-	else if (new_state == NOUVEAU_THERM_THRS_HIGHER &&
-		cur < thrs->temp - thrs->hysteresis)
-		new_state = NOUVEAU_THERM_THRS_LOWER;
-	nouveau_therm_sensor_set_threshold_state(therm, thrs_name, new_state);
-
-	/* find the direction */
-	if (prev_state < new_state)
-		direction = NOUVEAU_THERM_THRS_RISING;
-	else if (prev_state > new_state)
-		direction = NOUVEAU_THERM_THRS_FALLING;
-	else
-		return;
-
-	/* advertise a change in direction */
-	nouveau_therm_sensor_event(therm, thrs_name, direction);
+	nv_mask(therm, 0x20010, 0x40000000, 0x0);
+	mdelay(20); /* wait for the temperature to stabilize */
 }
 
-static void
-nv50_therm_intr(struct nouveau_subdev *subdev)
+static int
+nv50_temp_get(struct nouveau_therm *therm)
 {
-	struct nouveau_therm *therm = nouveau_therm(subdev);
 	struct nouveau_therm_priv *priv = (void *)therm;
 	struct nvbios_therm_sensor *sensor = &priv->bios_sensor;
-	unsigned long flags;
-	uint32_t intr;
-
-	spin_lock_irqsave(&priv->sensor.alarm_program_lock, flags);
-
-	intr = nv_rd32(therm, 0x20100);
-
-	/* THRS_4: downclock */
-	if (intr & 0x002) {
-		nv50_therm_threshold_hyst_emulation(therm, 0x20414, 24,
-						  &sensor->thrs_down_clock,
-						  NOUVEAU_THERM_THRS_DOWNCLOCK);
-		intr &= ~0x002;
-	}
+	int core_temp;
 
-	/* shutdown */
-	if (intr & 0x004) {
-		nv50_therm_threshold_hyst_emulation(therm, 0x20480, 20,
-						   &sensor->thrs_shutdown,
-						   NOUVEAU_THERM_THRS_SHUTDOWN);
-		intr &= ~0x004;
-	}
-
-	/* THRS_1 : fan boost */
-	if (intr & 0x008) {
-		nv50_therm_threshold_hyst_emulation(therm, 0x204c4, 21,
-						   &sensor->thrs_fan_boost,
-						   NOUVEAU_THERM_THRS_FANBOOST);
-		intr &= ~0x008;
-	}
+	core_temp = nv_rd32(therm, 0x20014) & 0x3fff;
 
-	/* THRS_2 : critical */
-	if (intr & 0x010) {
-		nv50_therm_threshold_hyst_emulation(therm, 0x204c0, 22,
-						   &sensor->thrs_critical,
-						   NOUVEAU_THERM_THRS_CRITICAL);
-		intr &= ~0x010;
-	}
+	/* if the slope or the offset is unset, do no use the sensor */
+	if (!sensor->slope_div || !sensor->slope_mult ||
+	    !sensor->offset_num || !sensor->offset_den)
+	    return -ENODEV;
 
-	if (intr)
-		nv_error(therm, "unhandled intr 0x%08x\n", intr);
+	core_temp = core_temp * sensor->slope_mult / sensor->slope_div;
+	core_temp = core_temp + sensor->offset_num / sensor->offset_den;
+	core_temp = core_temp + sensor->offset_constant - 8;
 
-	/* ACK everything */
-	nv_wr32(therm, 0x20100, 0xffffffff);
-	nv_wr32(therm, 0x1100, 0x10000); /* PBUS */
+	/* reserve negative temperatures for errors */
+	if (core_temp < 0)
+		core_temp = 0;
 
-	spin_unlock_irqrestore(&priv->sensor.alarm_program_lock, flags);
+	return core_temp;
 }
 
 static int
@@ -278,33 +169,29 @@ nv50_therm_ctor(struct nouveau_object *parent,
 	priv->base.base.pwm_set = nv50_fan_pwm_set;
 	priv->base.base.pwm_clock = nv50_fan_pwm_clock;
 	priv->base.base.temp_get = nv50_temp_get;
-	priv->base.sensor.program_alarms = nv50_therm_program_alarms;
-	nv_subdev(priv)->intr = nv50_therm_intr;
-
-	/* init the thresholds */
-	nouveau_therm_sensor_set_threshold_state(&priv->base.base,
-						 NOUVEAU_THERM_THRS_SHUTDOWN,
-						 NOUVEAU_THERM_THRS_LOWER);
-	nouveau_therm_sensor_set_threshold_state(&priv->base.base,
-						 NOUVEAU_THERM_THRS_FANBOOST,
-						 NOUVEAU_THERM_THRS_LOWER);
-	nouveau_therm_sensor_set_threshold_state(&priv->base.base,
-						 NOUVEAU_THERM_THRS_CRITICAL,
-						 NOUVEAU_THERM_THRS_LOWER);
-	nouveau_therm_sensor_set_threshold_state(&priv->base.base,
-						 NOUVEAU_THERM_THRS_DOWNCLOCK,
-						 NOUVEAU_THERM_THRS_LOWER);
+	priv->base.sensor.program_alarms = nouveau_therm_program_alarms_polling;
+	nv_subdev(priv)->intr = nv40_therm_intr;
 
 	return nouveau_therm_preinit(&priv->base.base);
 }
 
+static int
+nv50_therm_init(struct nouveau_object *object)
+{
+	struct nouveau_therm *therm = (void *)object;
+
+	nv50_sensor_setup(therm);
+
+	return _nouveau_therm_init(object);
+}
+
 struct nouveau_oclass
 nv50_therm_oclass = {
 	.handle = NV_SUBDEV(THERM, 0x50),
 	.ofuncs = &(struct nouveau_ofuncs) {
 		.ctor = nv50_therm_ctor,
 		.dtor = _nouveau_therm_dtor,
-		.init = _nouveau_therm_init,
+		.init = nv50_therm_init,
 		.fini = _nouveau_therm_fini,
 	},
 };
diff --git a/drivers/gpu/drm/nouveau/core/subdev/therm/nv84.c b/drivers/gpu/drm/nouveau/core/subdev/therm/nv84.c
new file mode 100644
index 000000000000..42ba633ccff7
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/subdev/therm/nv84.c
@@ -0,0 +1,221 @@
+/*
+ * Copyright 2012 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs
+ * 	    Martin Peres
+ */
+
+#include "priv.h"
+
+struct nv84_therm_priv {
+	struct nouveau_therm_priv base;
+};
+
+int
+nv84_temp_get(struct nouveau_therm *therm)
+{
+	return nv_rd32(therm, 0x20400);
+}
+
+static void
+nv84_therm_program_alarms(struct nouveau_therm *therm)
+{
+	struct nouveau_therm_priv *priv = (void *)therm;
+	struct nvbios_therm_sensor *sensor = &priv->bios_sensor;
+	unsigned long flags;
+
+	spin_lock_irqsave(&priv->sensor.alarm_program_lock, flags);
+
+	/* enable RISING and FALLING IRQs for shutdown, THRS 0, 1, 2 and 4 */
+	nv_wr32(therm, 0x20000, 0x000003ff);
+
+	/* shutdown: The computer should be shutdown when reached */
+	nv_wr32(therm, 0x20484, sensor->thrs_shutdown.hysteresis);
+	nv_wr32(therm, 0x20480, sensor->thrs_shutdown.temp);
+
+	/* THRS_1 : fan boost*/
+	nv_wr32(therm, 0x204c4, sensor->thrs_fan_boost.temp);
+
+	/* THRS_2 : critical */
+	nv_wr32(therm, 0x204c0, sensor->thrs_critical.temp);
+
+	/* THRS_4 : down clock */
+	nv_wr32(therm, 0x20414, sensor->thrs_down_clock.temp);
+	spin_unlock_irqrestore(&priv->sensor.alarm_program_lock, flags);
+
+	nv_debug(therm,
+		 "Programmed thresholds [ %d(%d), %d(%d), %d(%d), %d(%d) ]\n",
+		 sensor->thrs_fan_boost.temp, sensor->thrs_fan_boost.hysteresis,
+		 sensor->thrs_down_clock.temp,
+		 sensor->thrs_down_clock.hysteresis,
+		 sensor->thrs_critical.temp, sensor->thrs_critical.hysteresis,
+		 sensor->thrs_shutdown.temp, sensor->thrs_shutdown.hysteresis);
+
+}
+
+/* must be called with alarm_program_lock taken ! */
+static void
+nv84_therm_threshold_hyst_emulation(struct nouveau_therm *therm,
+				   uint32_t thrs_reg, u8 status_bit,
+				   const struct nvbios_therm_threshold *thrs,
+				   enum nouveau_therm_thrs thrs_name)
+{
+	enum nouveau_therm_thrs_direction direction;
+	enum nouveau_therm_thrs_state prev_state, new_state;
+	int temp, cur;
+
+	prev_state = nouveau_therm_sensor_get_threshold_state(therm, thrs_name);
+	temp = nv_rd32(therm, thrs_reg);
+
+	/* program the next threshold */
+	if (temp == thrs->temp) {
+		nv_wr32(therm, thrs_reg, thrs->temp - thrs->hysteresis);
+		new_state = NOUVEAU_THERM_THRS_HIGHER;
+	} else {
+		nv_wr32(therm, thrs_reg, thrs->temp);
+		new_state = NOUVEAU_THERM_THRS_LOWER;
+	}
+
+	/* fix the state (in case someone reprogrammed the alarms) */
+	cur = therm->temp_get(therm);
+	if (new_state == NOUVEAU_THERM_THRS_LOWER && cur > thrs->temp)
+		new_state = NOUVEAU_THERM_THRS_HIGHER;
+	else if (new_state == NOUVEAU_THERM_THRS_HIGHER &&
+		cur < thrs->temp - thrs->hysteresis)
+		new_state = NOUVEAU_THERM_THRS_LOWER;
+	nouveau_therm_sensor_set_threshold_state(therm, thrs_name, new_state);
+
+	/* find the direction */
+	if (prev_state < new_state)
+		direction = NOUVEAU_THERM_THRS_RISING;
+	else if (prev_state > new_state)
+		direction = NOUVEAU_THERM_THRS_FALLING;
+	else
+		return;
+
+	/* advertise a change in direction */
+	nouveau_therm_sensor_event(therm, thrs_name, direction);
+}
+
+static void
+nv84_therm_intr(struct nouveau_subdev *subdev)
+{
+	struct nouveau_therm *therm = nouveau_therm(subdev);
+	struct nouveau_therm_priv *priv = (void *)therm;
+	struct nvbios_therm_sensor *sensor = &priv->bios_sensor;
+	unsigned long flags;
+	uint32_t intr;
+
+	spin_lock_irqsave(&priv->sensor.alarm_program_lock, flags);
+
+	intr = nv_rd32(therm, 0x20100);
+
+	/* THRS_4: downclock */
+	if (intr & 0x002) {
+		nv84_therm_threshold_hyst_emulation(therm, 0x20414, 24,
+						  &sensor->thrs_down_clock,
+						  NOUVEAU_THERM_THRS_DOWNCLOCK);
+		intr &= ~0x002;
+	}
+
+	/* shutdown */
+	if (intr & 0x004) {
+		nv84_therm_threshold_hyst_emulation(therm, 0x20480, 20,
+						   &sensor->thrs_shutdown,
+						   NOUVEAU_THERM_THRS_SHUTDOWN);
+		intr &= ~0x004;
+	}
+
+	/* THRS_1 : fan boost */
+	if (intr & 0x008) {
+		nv84_therm_threshold_hyst_emulation(therm, 0x204c4, 21,
+						   &sensor->thrs_fan_boost,
+						   NOUVEAU_THERM_THRS_FANBOOST);
+		intr &= ~0x008;
+	}
+
+	/* THRS_2 : critical */
+	if (intr & 0x010) {
+		nv84_therm_threshold_hyst_emulation(therm, 0x204c0, 22,
+						   &sensor->thrs_critical,
+						   NOUVEAU_THERM_THRS_CRITICAL);
+		intr &= ~0x010;
+	}
+
+	if (intr)
+		nv_error(therm, "unhandled intr 0x%08x\n", intr);
+
+	/* ACK everything */
+	nv_wr32(therm, 0x20100, 0xffffffff);
+	nv_wr32(therm, 0x1100, 0x10000); /* PBUS */
+
+	spin_unlock_irqrestore(&priv->sensor.alarm_program_lock, flags);
+}
+
+static int
+nv84_therm_ctor(struct nouveau_object *parent,
+		struct nouveau_object *engine,
+		struct nouveau_oclass *oclass, void *data, u32 size,
+		struct nouveau_object **pobject)
+{
+	struct nv84_therm_priv *priv;
+	int ret;
+
+	ret = nouveau_therm_create(parent, engine, oclass, &priv);
+	*pobject = nv_object(priv);
+	if (ret)
+		return ret;
+
+	priv->base.base.pwm_ctrl = nv50_fan_pwm_ctrl;
+	priv->base.base.pwm_get = nv50_fan_pwm_get;
+	priv->base.base.pwm_set = nv50_fan_pwm_set;
+	priv->base.base.pwm_clock = nv50_fan_pwm_clock;
+	priv->base.base.temp_get = nv84_temp_get;
+	priv->base.sensor.program_alarms = nv84_therm_program_alarms;
+	nv_subdev(priv)->intr = nv84_therm_intr;
+
+	/* init the thresholds */
+	nouveau_therm_sensor_set_threshold_state(&priv->base.base,
+						 NOUVEAU_THERM_THRS_SHUTDOWN,
+						 NOUVEAU_THERM_THRS_LOWER);
+	nouveau_therm_sensor_set_threshold_state(&priv->base.base,
+						 NOUVEAU_THERM_THRS_FANBOOST,
+						 NOUVEAU_THERM_THRS_LOWER);
+	nouveau_therm_sensor_set_threshold_state(&priv->base.base,
+						 NOUVEAU_THERM_THRS_CRITICAL,
+						 NOUVEAU_THERM_THRS_LOWER);
+	nouveau_therm_sensor_set_threshold_state(&priv->base.base,
+						 NOUVEAU_THERM_THRS_DOWNCLOCK,
+						 NOUVEAU_THERM_THRS_LOWER);
+
+	return nouveau_therm_preinit(&priv->base.base);
+}
+
+struct nouveau_oclass
+nv84_therm_oclass = {
+	.handle = NV_SUBDEV(THERM, 0x84),
+	.ofuncs = &(struct nouveau_ofuncs) {
+		.ctor = nv84_therm_ctor,
+		.dtor = _nouveau_therm_dtor,
+		.init = _nouveau_therm_init,
+		.fini = _nouveau_therm_fini,
+	},
+};
diff --git a/drivers/gpu/drm/nouveau/core/subdev/therm/nva3.c b/drivers/gpu/drm/nouveau/core/subdev/therm/nva3.c
index 2dcc5437116a..d11a7c400813 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/therm/nva3.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/therm/nva3.c
@@ -81,7 +81,7 @@ nva3_therm_ctor(struct nouveau_object *parent,
 	priv->base.base.pwm_get = nv50_fan_pwm_get;
 	priv->base.base.pwm_set = nv50_fan_pwm_set;
 	priv->base.base.pwm_clock = nv50_fan_pwm_clock;
-	priv->base.base.temp_get = nv50_temp_get;
+	priv->base.base.temp_get = nv84_temp_get;
 	priv->base.base.fan_sense = nva3_therm_fan_sense;
 	priv->base.sensor.program_alarms = nouveau_therm_program_alarms_polling;
 	return nouveau_therm_preinit(&priv->base.base);
diff --git a/drivers/gpu/drm/nouveau/core/subdev/therm/nvd0.c b/drivers/gpu/drm/nouveau/core/subdev/therm/nvd0.c
index d7d30ee8332e..54c28bdc4204 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/therm/nvd0.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/therm/nvd0.c
@@ -135,7 +135,7 @@ nvd0_therm_ctor(struct nouveau_object *parent,
 	priv->base.base.pwm_get = nvd0_fan_pwm_get;
 	priv->base.base.pwm_set = nvd0_fan_pwm_set;
 	priv->base.base.pwm_clock = nvd0_fan_pwm_clock;
-	priv->base.base.temp_get = nv50_temp_get;
+	priv->base.base.temp_get = nv84_temp_get;
 	priv->base.base.fan_sense = nva3_therm_fan_sense;
 	priv->base.sensor.program_alarms = nouveau_therm_program_alarms_polling;
 	return nouveau_therm_preinit(&priv->base.base);
diff --git a/drivers/gpu/drm/nouveau/core/subdev/therm/priv.h b/drivers/gpu/drm/nouveau/core/subdev/therm/priv.h
index 438d9824b774..15ca64e481f1 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/therm/priv.h
+++ b/drivers/gpu/drm/nouveau/core/subdev/therm/priv.h
@@ -134,11 +134,12 @@ void nouveau_therm_sensor_event(struct nouveau_therm *therm,
 			        enum nouveau_therm_thrs_direction dir);
 void nouveau_therm_program_alarms_polling(struct nouveau_therm *therm);
 
+void nv40_therm_intr(struct nouveau_subdev *);
 int nv50_fan_pwm_ctrl(struct nouveau_therm *, int, bool);
 int nv50_fan_pwm_get(struct nouveau_therm *, int, u32 *, u32 *);
 int nv50_fan_pwm_set(struct nouveau_therm *, int, u32, u32);
 int nv50_fan_pwm_clock(struct nouveau_therm *);
-int nv50_temp_get(struct nouveau_therm *therm);
+int nv84_temp_get(struct nouveau_therm *therm);
 
 int nva3_therm_fan_sense(struct nouveau_therm *);
 
diff --git a/drivers/gpu/drm/nouveau/core/subdev/therm/temp.c b/drivers/gpu/drm/nouveau/core/subdev/therm/temp.c
index 470f6a47b656..dde746c78c8a 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/therm/temp.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/therm/temp.c
@@ -205,13 +205,13 @@ nouveau_therm_program_alarms_polling(struct nouveau_therm *therm)
 	struct nouveau_therm_priv *priv = (void *)therm;
 	struct nvbios_therm_sensor *sensor = &priv->bios_sensor;
 
-	nv_info(therm,
-		"programmed thresholds [ %d(%d), %d(%d), %d(%d), %d(%d) ]\n",
-		sensor->thrs_fan_boost.temp, sensor->thrs_fan_boost.hysteresis,
-		sensor->thrs_down_clock.temp,
-		sensor->thrs_down_clock.hysteresis,
-		sensor->thrs_critical.temp, sensor->thrs_critical.hysteresis,
-		sensor->thrs_shutdown.temp, sensor->thrs_shutdown.hysteresis);
+	nv_debug(therm,
+		 "programmed thresholds [ %d(%d), %d(%d), %d(%d), %d(%d) ]\n",
+		 sensor->thrs_fan_boost.temp, sensor->thrs_fan_boost.hysteresis,
+		 sensor->thrs_down_clock.temp,
+		 sensor->thrs_down_clock.hysteresis,
+		 sensor->thrs_critical.temp, sensor->thrs_critical.hysteresis,
+		 sensor->thrs_shutdown.temp, sensor->thrs_shutdown.hysteresis);
 
 	alarm_timer_callback(&priv->sensor.therm_poll_alarm);
 }
diff --git a/drivers/gpu/drm/nouveau/core/subdev/timer/nv04.c b/drivers/gpu/drm/nouveau/core/subdev/timer/nv04.c
index 8e1bae4f12e8..9469b8275675 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/timer/nv04.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/timer/nv04.c
@@ -96,11 +96,16 @@ nv04_timer_alarm(struct nouveau_timer *ptimer, u64 time,
 
 	/* append new alarm to list, in soonest-alarm-first order */
 	spin_lock_irqsave(&priv->lock, flags);
-	list_for_each_entry(list, &priv->alarms, head) {
-		if (list->timestamp > alarm->timestamp)
-			break;
+	if (!time) {
+		if (!list_empty(&alarm->head))
+			list_del(&alarm->head);
+	} else {
+		list_for_each_entry(list, &priv->alarms, head) {
+			if (list->timestamp > alarm->timestamp)
+				break;
+		}
+		list_add_tail(&alarm->head, &list->head);
 	}
-	list_add_tail(&alarm->head, &list->head);
 	spin_unlock_irqrestore(&priv->lock, flags);
 
 	/* process pending alarms */
diff --git a/drivers/gpu/drm/nouveau/core/subdev/vm/nv04.c b/drivers/gpu/drm/nouveau/core/subdev/vm/nv04.c
index 6adbbc9cc361..ed45437167f2 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/vm/nv04.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/vm/nv04.c
@@ -110,7 +110,7 @@ nv04_vmmgr_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	if (ret)
 		return ret;
 
-	ret = nouveau_gpuobj_new(parent, NULL,
+	ret = nouveau_gpuobj_new(nv_object(priv), NULL,
 				 (NV04_PDMA_SIZE / NV04_PDMA_PAGE) * 4 +
 				 8, 16, NVOBJ_FLAG_ZERO_ALLOC,
 				 &priv->vm->pgt[0].obj[0]);
diff --git a/drivers/gpu/drm/nouveau/core/subdev/vm/nv41.c b/drivers/gpu/drm/nouveau/core/subdev/vm/nv41.c
index 9474cfca6e4c..064c76262876 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/vm/nv41.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/vm/nv41.c
@@ -119,7 +119,7 @@ nv41_vmmgr_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	if (ret)
 		return ret;
 
-	ret = nouveau_gpuobj_new(parent, NULL,
+	ret = nouveau_gpuobj_new(nv_object(priv), NULL,
 				(NV41_GART_SIZE / NV41_GART_PAGE) * 4,
 				 16, NVOBJ_FLAG_ZERO_ALLOC,
 				 &priv->vm->pgt[0].obj[0]);
diff --git a/drivers/gpu/drm/nouveau/core/subdev/vm/nv44.c b/drivers/gpu/drm/nouveau/core/subdev/vm/nv44.c
index aa8131436e3d..fae1f67d5948 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/vm/nv44.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/vm/nv44.c
@@ -196,7 +196,7 @@ nv44_vmmgr_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	if (ret)
 		return ret;
 
-	ret = nouveau_gpuobj_new(parent, NULL,
+	ret = nouveau_gpuobj_new(nv_object(priv), NULL,
 				(NV44_GART_SIZE / NV44_GART_PAGE) * 4,
 				 512 * 1024, NVOBJ_FLAG_ZERO_ALLOC,
 				 &priv->vm->pgt[0].obj[0]);
diff --git a/drivers/gpu/drm/nouveau/core/subdev/vm/nvc0.c b/drivers/gpu/drm/nouveau/core/subdev/vm/nvc0.c
index 30c61e6c2017..4c3b0a23b9d6 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/vm/nvc0.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/vm/nvc0.c
@@ -28,12 +28,54 @@
 #include <subdev/timer.h>
 #include <subdev/fb.h>
 #include <subdev/vm.h>
+#include <subdev/ltcg.h>
 
 struct nvc0_vmmgr_priv {
 	struct nouveau_vmmgr base;
 	spinlock_t lock;
 };
 
+
+/* Map from compressed to corresponding uncompressed storage type.
+ * The value 0xff represents an invalid storage type.
+ */
+const u8 nvc0_pte_storage_type_map[256] =
+{
+	0x00, 0x01, 0x01, 0x01, 0x01, 0x01, 0xff, 0x01, /* 0x00 */
+	0x01, 0x01, 0x01, 0xff, 0xff, 0xff, 0xff, 0xff,
+	0xff, 0x11, 0xff, 0xff, 0xff, 0xff, 0xff, 0x11, /* 0x10 */
+	0x11, 0x11, 0x11, 0xff, 0xff, 0xff, 0xff, 0xff,
+	0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0x26, 0x27, /* 0x20 */
+	0x28, 0x29, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff,
+	0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, /* 0x30 */
+	0xff, 0xff, 0x26, 0x27, 0x28, 0x29, 0x26, 0x27,
+	0x28, 0x29, 0xff, 0xff, 0xff, 0xff, 0x46, 0xff, /* 0x40 */
+	0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff,
+	0xff, 0x46, 0x46, 0x46, 0x46, 0xff, 0xff, 0xff, /* 0x50 */
+	0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff,
+	0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, /* 0x60 */
+	0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff,
+	0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, /* 0x70 */
+	0xff, 0xff, 0xff, 0x7b, 0xff, 0xff, 0xff, 0xff,
+	0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0x7b, 0x7b, /* 0x80 */
+	0x7b, 0x7b, 0xff, 0x8b, 0x8c, 0x8d, 0x8e, 0xff,
+	0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, /* 0x90 */
+	0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff,
+	0xff, 0xff, 0xff, 0x8b, 0x8c, 0x8d, 0x8e, 0xa7, /* 0xa0 */
+	0xa8, 0xa9, 0xaa, 0xff, 0xff, 0xff, 0xff, 0xff,
+	0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, /* 0xb0 */
+	0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xa7,
+	0xa8, 0xa9, 0xaa, 0xc3, 0xff, 0xff, 0xff, 0xff, /* 0xc0 */
+	0xff, 0xff, 0xff, 0xff, 0xfe, 0xfe, 0xc3, 0xc3,
+	0xc3, 0xc3, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, /* 0xd0 */
+	0xfe, 0xff, 0xff, 0xfe, 0xff, 0xfe, 0xff, 0xfe,
+	0xfe, 0xff, 0xfe, 0xfe, 0xfe, 0xfe, 0xfe, 0xff, /* 0xe0 */
+	0xff, 0xfe, 0xff, 0xfe, 0xff, 0xfe, 0xfe, 0xff,
+	0xfe, 0xfe, 0xfe, 0xfe, 0xfe, 0xfe, 0xfe, 0xfe, /* 0xf0 */
+	0xfe, 0xfe, 0xfe, 0xfe, 0xff, 0xfd, 0xfe, 0xff
+};
+
+
 static void
 nvc0_vm_map_pgt(struct nouveau_gpuobj *pgd, u32 index,
 		struct nouveau_gpuobj *pgt[2])
@@ -68,10 +110,20 @@ static void
 nvc0_vm_map(struct nouveau_vma *vma, struct nouveau_gpuobj *pgt,
 	    struct nouveau_mem *mem, u32 pte, u32 cnt, u64 phys, u64 delta)
 {
-	u32 next = 1 << (vma->node->type - 8);
+	u64 next = 1 << (vma->node->type - 8);
 
 	phys  = nvc0_vm_addr(vma, phys, mem->memtype, 0);
 	pte <<= 3;
+
+	if (mem->tag) {
+		struct nouveau_ltcg *ltcg =
+			nouveau_ltcg(vma->vm->vmm->base.base.parent);
+		u32 tag = mem->tag->offset + (delta >> 17);
+		phys |= (u64)tag << (32 + 12);
+		next |= (u64)1   << (32 + 12);
+		ltcg->tags_clear(ltcg, tag, cnt);
+	}
+
 	while (cnt--) {
 		nv_wo32(pgt, pte + 0, lower_32_bits(phys));
 		nv_wo32(pgt, pte + 4, upper_32_bits(phys));
@@ -85,10 +137,12 @@ nvc0_vm_map_sg(struct nouveau_vma *vma, struct nouveau_gpuobj *pgt,
 	       struct nouveau_mem *mem, u32 pte, u32 cnt, dma_addr_t *list)
 {
 	u32 target = (vma->access & NV_MEM_ACCESS_NOSNOOP) ? 7 : 5;
+	/* compressed storage types are invalid for system memory */
+	u32 memtype = nvc0_pte_storage_type_map[mem->memtype & 0xff];
 
 	pte <<= 3;
 	while (cnt--) {
-		u64 phys = nvc0_vm_addr(vma, *list++, mem->memtype, target);
+		u64 phys = nvc0_vm_addr(vma, *list++, memtype, target);
 		nv_wo32(pgt, pte + 0, lower_32_bits(phys));
 		nv_wo32(pgt, pte + 4, upper_32_bits(phys));
 		pte += 8;
diff --git a/drivers/gpu/drm/nouveau/dispnv04/Makefile b/drivers/gpu/drm/nouveau/dispnv04/Makefile
new file mode 100644
index 000000000000..ea3f5b8a0f95
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/dispnv04/Makefile
@@ -0,0 +1,10 @@
+nouveau-y += dispnv04/arb.o
+nouveau-y += dispnv04/crtc.o
+nouveau-y += dispnv04/cursor.o
+nouveau-y += dispnv04/dac.o
+nouveau-y += dispnv04/dfp.o
+nouveau-y += dispnv04/disp.o
+nouveau-y += dispnv04/hw.o
+nouveau-y += dispnv04/tvmodesnv17.o
+nouveau-y += dispnv04/tvnv04.o
+nouveau-y += dispnv04/tvnv17.o
diff --git a/drivers/gpu/drm/nouveau/nouveau_calc.c b/drivers/gpu/drm/nouveau/dispnv04/arb.c
index 6da576445b3d..2e70462883e8 100644
--- a/drivers/gpu/drm/nouveau/nouveau_calc.c
+++ b/drivers/gpu/drm/nouveau/dispnv04/arb.c
@@ -25,7 +25,7 @@
 
 #include "nouveau_drm.h"
 #include "nouveau_reg.h"
-#include "nouveau_hw.h"
+#include "hw.h"
 
 /****************************************************************************\
 *                                                                            *
diff --git a/drivers/gpu/drm/nouveau/nv04_crtc.c b/drivers/gpu/drm/nouveau/dispnv04/crtc.c
index 6578cd28c556..0782bd2f1e04 100644
--- a/drivers/gpu/drm/nouveau/nv04_crtc.c
+++ b/drivers/gpu/drm/nouveau/dispnv04/crtc.c
@@ -33,10 +33,10 @@
 #include "nouveau_encoder.h"
 #include "nouveau_connector.h"
 #include "nouveau_crtc.h"
-#include "nouveau_hw.h"
+#include "hw.h"
 #include "nvreg.h"
 #include "nouveau_fbcon.h"
-#include "nv04_display.h"
+#include "disp.h"
 
 #include <subdev/bios/pll.h>
 #include <subdev/clock.h>
@@ -1070,4 +1070,3 @@ nv04_crtc_create(struct drm_device *dev, int crtc_num)
 
 	return 0;
 }
-
diff --git a/drivers/gpu/drm/nouveau/nv04_cursor.c b/drivers/gpu/drm/nouveau/dispnv04/cursor.c
index fe86f0de348f..a810303169de 100644
--- a/drivers/gpu/drm/nouveau/nv04_cursor.c
+++ b/drivers/gpu/drm/nouveau/dispnv04/cursor.c
@@ -3,7 +3,7 @@
 #include "nouveau_drm.h"
 #include "nouveau_reg.h"
 #include "nouveau_crtc.h"
-#include "nouveau_hw.h"
+#include "hw.h"
 
 static void
 nv04_cursor_show(struct nouveau_crtc *nv_crtc, bool update)
@@ -68,4 +68,3 @@ nv04_cursor_init(struct nouveau_crtc *crtc)
 	crtc->cursor.show = nv04_cursor_show;
 	return 0;
 }
-
diff --git a/drivers/gpu/drm/nouveau/nv04_dac.c b/drivers/gpu/drm/nouveau/dispnv04/dac.c
index 64f7020fb605..434b920f6bd4 100644
--- a/drivers/gpu/drm/nouveau/nv04_dac.c
+++ b/drivers/gpu/drm/nouveau/dispnv04/dac.c
@@ -31,7 +31,7 @@
 #include "nouveau_encoder.h"
 #include "nouveau_connector.h"
 #include "nouveau_crtc.h"
-#include "nouveau_hw.h"
+#include "hw.h"
 #include "nvreg.h"
 
 #include <subdev/bios/gpio.h>
diff --git a/drivers/gpu/drm/nouveau/nv04_dfp.c b/drivers/gpu/drm/nouveau/dispnv04/dfp.c
index 7e24cdf1cb39..93dd23ff0093 100644
--- a/drivers/gpu/drm/nouveau/nv04_dfp.c
+++ b/drivers/gpu/drm/nouveau/dispnv04/dfp.c
@@ -32,7 +32,7 @@
 #include "nouveau_encoder.h"
 #include "nouveau_connector.h"
 #include "nouveau_crtc.h"
-#include "nouveau_hw.h"
+#include "hw.h"
 #include "nvreg.h"
 
 #include <drm/i2c/sil164.h>
diff --git a/drivers/gpu/drm/nouveau/nv04_display.c b/drivers/gpu/drm/nouveau/dispnv04/disp.c
index ad48444c385c..4908d3fd0486 100644
--- a/drivers/gpu/drm/nouveau/nv04_display.c
+++ b/drivers/gpu/drm/nouveau/dispnv04/disp.c
@@ -30,7 +30,7 @@
 
 #include "nouveau_drm.h"
 #include "nouveau_reg.h"
-#include "nouveau_hw.h"
+#include "hw.h"
 #include "nouveau_encoder.h"
 #include "nouveau_connector.h"
 
diff --git a/drivers/gpu/drm/nouveau/nv04_display.h b/drivers/gpu/drm/nouveau/dispnv04/disp.h
index a0a031dad13f..a0a031dad13f 100644
--- a/drivers/gpu/drm/nouveau/nv04_display.h
+++ b/drivers/gpu/drm/nouveau/dispnv04/disp.h
diff --git a/drivers/gpu/drm/nouveau/nouveau_hw.c b/drivers/gpu/drm/nouveau/dispnv04/hw.c
index 617a06ffdb46..973056b86207 100644
--- a/drivers/gpu/drm/nouveau/nouveau_hw.c
+++ b/drivers/gpu/drm/nouveau/dispnv04/hw.c
@@ -24,7 +24,7 @@
 
 #include <drm/drmP.h>
 #include "nouveau_drm.h"
-#include "nouveau_hw.h"
+#include "hw.h"
 
 #include <subdev/bios/pll.h>
 #include <subdev/clock.h>
diff --git a/drivers/gpu/drm/nouveau/nouveau_hw.h b/drivers/gpu/drm/nouveau/dispnv04/hw.h
index 7dff1021fab4..eeb70d912d99 100644
--- a/drivers/gpu/drm/nouveau/nouveau_hw.h
+++ b/drivers/gpu/drm/nouveau/dispnv04/hw.h
@@ -24,7 +24,8 @@
 #define __NOUVEAU_HW_H__
 
 #include <drm/drmP.h>
-#include "nv04_display.h"
+#include "disp.h"
+#include "nvreg.h"
 
 #include <subdev/bios/pll.h>
 
diff --git a/drivers/gpu/drm/nouveau/nvreg.h b/drivers/gpu/drm/nouveau/dispnv04/nvreg.h
index bbfb1a68fb11..bbfb1a68fb11 100644
--- a/drivers/gpu/drm/nouveau/nvreg.h
+++ b/drivers/gpu/drm/nouveau/dispnv04/nvreg.h
diff --git a/drivers/gpu/drm/nouveau/nv17_tv_modes.c b/drivers/gpu/drm/nouveau/dispnv04/tvmodesnv17.c
index 1cdfe2a5875d..08c6f5e50610 100644
--- a/drivers/gpu/drm/nouveau/nv17_tv_modes.c
+++ b/drivers/gpu/drm/nouveau/dispnv04/tvmodesnv17.c
@@ -29,8 +29,8 @@
 #include "nouveau_drm.h"
 #include "nouveau_encoder.h"
 #include "nouveau_crtc.h"
-#include "nouveau_hw.h"
-#include "nv17_tv.h"
+#include "hw.h"
+#include "tvnv17.h"
 
 char *nv17_tv_norm_names[NUM_TV_NORMS] = {
 	[TV_NORM_PAL] = "PAL",
diff --git a/drivers/gpu/drm/nouveau/nv04_tv.c b/drivers/gpu/drm/nouveau/dispnv04/tvnv04.c
index 4a69ccdef9b4..bf13db4e8631 100644
--- a/drivers/gpu/drm/nouveau/nv04_tv.c
+++ b/drivers/gpu/drm/nouveau/dispnv04/tvnv04.c
@@ -30,7 +30,7 @@
 #include "nouveau_encoder.h"
 #include "nouveau_connector.h"
 #include "nouveau_crtc.h"
-#include "nouveau_hw.h"
+#include "hw.h"
 #include <drm/drm_crtc_helper.h>
 
 #include <drm/i2c/ch7006.h>
diff --git a/drivers/gpu/drm/nouveau/nv17_tv.c b/drivers/gpu/drm/nouveau/dispnv04/tvnv17.c
index 977e42be2050..acef48f4a4ea 100644
--- a/drivers/gpu/drm/nouveau/nv17_tv.c
+++ b/drivers/gpu/drm/nouveau/dispnv04/tvnv17.c
@@ -31,8 +31,8 @@
 #include "nouveau_encoder.h"
 #include "nouveau_connector.h"
 #include "nouveau_crtc.h"
-#include "nouveau_hw.h"
-#include "nv17_tv.h"
+#include "hw.h"
+#include "tvnv17.h"
 
 #include <core/device.h>
 
diff --git a/drivers/gpu/drm/nouveau/nv17_tv.h b/drivers/gpu/drm/nouveau/dispnv04/tvnv17.h
index 7b331543a41b..7b331543a41b 100644
--- a/drivers/gpu/drm/nouveau/nv17_tv.h
+++ b/drivers/gpu/drm/nouveau/dispnv04/tvnv17.h
diff --git a/drivers/gpu/drm/nouveau/nouveau_abi16.c b/drivers/gpu/drm/nouveau/nouveau_abi16.c
index 5eb3e0da7c6e..1c4c6c9161ac 100644
--- a/drivers/gpu/drm/nouveau/nouveau_abi16.c
+++ b/drivers/gpu/drm/nouveau/nouveau_abi16.c
@@ -30,6 +30,7 @@
 #include <subdev/fb.h>
 #include <subdev/timer.h>
 #include <subdev/instmem.h>
+#include <engine/graph.h>
 
 #include "nouveau_drm.h"
 #include "nouveau_dma.h"
@@ -168,6 +169,7 @@ nouveau_abi16_ioctl_getparam(ABI16_IOCTL_ARGS)
 	struct nouveau_drm *drm = nouveau_drm(dev);
 	struct nouveau_device *device = nv_device(drm->device);
 	struct nouveau_timer *ptimer = nouveau_timer(device);
+	struct nouveau_graph *graph = (void *)nouveau_engine(device, NVDEV_ENGINE_GR);
 	struct drm_nouveau_getparam *getparam = data;
 
 	switch (getparam->param) {
@@ -208,14 +210,8 @@ nouveau_abi16_ioctl_getparam(ABI16_IOCTL_ARGS)
 		getparam->value = 1;
 		break;
 	case NOUVEAU_GETPARAM_GRAPH_UNITS:
-		/* NV40 and NV50 versions are quite different, but register
-		 * address is the same. User is supposed to know the card
-		 * family anyway... */
-		if (device->chipset >= 0x40) {
-			getparam->value = nv_rd32(device, 0x001540);
-			break;
-		}
-		/* FALLTHRU */
+		getparam->value = graph->units ? graph->units(graph) : 0;
+		break;
 	default:
 		nv_debug(device, "unknown parameter %lld\n", getparam->param);
 		return -EINVAL;
diff --git a/drivers/gpu/drm/nouveau/nouveau_backlight.c b/drivers/gpu/drm/nouveau/nouveau_backlight.c
index 5d940302d2aa..2ffad2176b7f 100644
--- a/drivers/gpu/drm/nouveau/nouveau_backlight.c
+++ b/drivers/gpu/drm/nouveau/nouveau_backlight.c
@@ -239,6 +239,9 @@ nouveau_backlight_init(struct drm_device *dev)
 		case NV_40:
 			return nv40_backlight_init(connector);
 		case NV_50:
+		case NV_C0:
+		case NV_D0:
+		case NV_E0:
 			return nv50_backlight_init(connector);
 		default:
 			break;
diff --git a/drivers/gpu/drm/nouveau/nouveau_bios.c b/drivers/gpu/drm/nouveau/nouveau_bios.c
index 50a6dd02f7c5..6aa2137e093a 100644
--- a/drivers/gpu/drm/nouveau/nouveau_bios.c
+++ b/drivers/gpu/drm/nouveau/nouveau_bios.c
@@ -28,7 +28,7 @@
 
 #include "nouveau_drm.h"
 #include "nouveau_reg.h"
-#include "nouveau_hw.h"
+#include "dispnv04/hw.h"
 #include "nouveau_encoder.h"
 
 #include <linux/io-mapping.h>
diff --git a/drivers/gpu/drm/nouveau/nouveau_bios.h b/drivers/gpu/drm/nouveau/nouveau_bios.h
index 7ccd28f11adf..0067586eb015 100644
--- a/drivers/gpu/drm/nouveau/nouveau_bios.h
+++ b/drivers/gpu/drm/nouveau/nouveau_bios.h
@@ -24,8 +24,6 @@
 #ifndef __NOUVEAU_DISPBIOS_H__
 #define __NOUVEAU_DISPBIOS_H__
 
-#include "nvreg.h"
-
 #define DCB_MAX_NUM_ENTRIES 16
 #define DCB_MAX_NUM_I2C_ENTRIES 16
 #define DCB_MAX_NUM_GPIO_ENTRIES 32
diff --git a/drivers/gpu/drm/nouveau/nouveau_connector.c b/drivers/gpu/drm/nouveau/nouveau_connector.c
index 4dd7ae2ac6c6..4da776f344d7 100644
--- a/drivers/gpu/drm/nouveau/nouveau_connector.c
+++ b/drivers/gpu/drm/nouveau/nouveau_connector.c
@@ -32,7 +32,7 @@
 
 #include "nouveau_reg.h"
 #include "nouveau_drm.h"
-#include "nouveau_hw.h"
+#include "dispnv04/hw.h"
 #include "nouveau_acpi.h"
 
 #include "nouveau_display.h"
diff --git a/drivers/gpu/drm/nouveau/nouveau_display.c b/drivers/gpu/drm/nouveau/nouveau_display.c
index 4610c3a29bbe..7bf22d4a3d96 100644
--- a/drivers/gpu/drm/nouveau/nouveau_display.c
+++ b/drivers/gpu/drm/nouveau/nouveau_display.c
@@ -28,7 +28,7 @@
 #include <drm/drm_crtc_helper.h>
 
 #include "nouveau_fbcon.h"
-#include "nouveau_hw.h"
+#include "dispnv04/hw.h"
 #include "nouveau_crtc.h"
 #include "nouveau_dma.h"
 #include "nouveau_gem.h"
diff --git a/drivers/gpu/drm/nouveau/nouveau_drm.c b/drivers/gpu/drm/nouveau/nouveau_drm.c
index c95decf543e9..46c152ff0a80 100644
--- a/drivers/gpu/drm/nouveau/nouveau_drm.c
+++ b/drivers/gpu/drm/nouveau/nouveau_drm.c
@@ -31,13 +31,13 @@
 #include <core/gpuobj.h>
 #include <core/class.h>
 
-#include <subdev/device.h>
-#include <subdev/vm.h>
-
+#include <engine/device.h>
 #include <engine/disp.h>
+#include <engine/fifo.h>
+
+#include <subdev/vm.h>
 
 #include "nouveau_drm.h"
-#include "nouveau_irq.h"
 #include "nouveau_dma.h"
 #include "nouveau_ttm.h"
 #include "nouveau_gem.h"
@@ -165,7 +165,7 @@ nouveau_accel_init(struct nouveau_drm *drm)
 	u32 arg0, arg1;
 	int ret;
 
-	if (nouveau_noaccel)
+	if (nouveau_noaccel || !nouveau_fifo(device) /*XXX*/)
 		return;
 
 	/* initialise synchronisation routines */
@@ -365,10 +365,6 @@ nouveau_drm_load(struct drm_device *dev, unsigned long flags)
 	if (ret)
 		goto fail_bios;
 
-	ret = nouveau_irq_init(dev);
-	if (ret)
-		goto fail_irq;
-
 	ret = nouveau_display_create(dev);
 	if (ret)
 		goto fail_dispctor;
@@ -388,8 +384,6 @@ nouveau_drm_load(struct drm_device *dev, unsigned long flags)
 fail_dispinit:
 	nouveau_display_destroy(dev);
 fail_dispctor:
-	nouveau_irq_fini(dev);
-fail_irq:
 	nouveau_bios_takedown(dev);
 fail_bios:
 	nouveau_ttm_fini(drm);
@@ -415,7 +409,6 @@ nouveau_drm_unload(struct drm_device *dev)
 		nouveau_display_fini(dev);
 	nouveau_display_destroy(dev);
 
-	nouveau_irq_fini(dev);
 	nouveau_bios_takedown(dev);
 
 	nouveau_ttm_fini(drm);
@@ -533,7 +526,6 @@ nouveau_do_resume(struct drm_device *dev)
 		nouveau_fence(drm)->resume(drm);
 
 	nouveau_run_vbios_init(dev);
-	nouveau_irq_postinstall(dev);
 	nouveau_pm_resume(dev);
 
 	if (dev->mode_config.num_crtc) {
@@ -669,8 +661,7 @@ static struct drm_driver
 driver = {
 	.driver_features =
 		DRIVER_USE_AGP | DRIVER_PCI_DMA | DRIVER_SG |
-		DRIVER_HAVE_IRQ | DRIVER_IRQ_SHARED | DRIVER_GEM |
-		DRIVER_MODESET | DRIVER_PRIME,
+		DRIVER_GEM | DRIVER_MODESET | DRIVER_PRIME,
 
 	.load = nouveau_drm_load,
 	.unload = nouveau_drm_unload,
@@ -684,11 +675,6 @@ driver = {
 	.debugfs_cleanup = nouveau_debugfs_takedown,
 #endif
 
-	.irq_preinstall = nouveau_irq_preinstall,
-	.irq_postinstall = nouveau_irq_postinstall,
-	.irq_uninstall = nouveau_irq_uninstall,
-	.irq_handler = nouveau_irq_handler,
-
 	.get_vblank_counter = drm_vblank_count,
 	.enable_vblank = nouveau_drm_vblank_enable,
 	.disable_vblank = nouveau_drm_vblank_disable,
diff --git a/drivers/gpu/drm/nouveau/nouveau_drm.h b/drivers/gpu/drm/nouveau/nouveau_drm.h
index 9c39bafbef2c..f2b30f89dee0 100644
--- a/drivers/gpu/drm/nouveau/nouveau_drm.h
+++ b/drivers/gpu/drm/nouveau/nouveau_drm.h
@@ -10,7 +10,18 @@
 
 #define DRIVER_MAJOR		1
 #define DRIVER_MINOR		1
-#define DRIVER_PATCHLEVEL	0
+#define DRIVER_PATCHLEVEL	1
+
+/*
+ * 1.1.1:
+ * 	- added support for tiled system memory buffer objects
+ *      - added support for NOUVEAU_GETPARAM_GRAPH_UNITS on [nvc0,nve0].
+ *      - added support for compressed memory storage types on [nvc0,nve0].
+ *      - added support for software methods 0x600,0x644,0x6ac on nvc0
+ *        to control registers on the MPs to enable performance counters,
+ *        and to control the warp error enable mask (OpenGL requires out of
+ *        bounds access to local memory to be silently ignored / return 0).
+ */
 
 #include <core/client.h>
 #include <core/event.h>
diff --git a/drivers/gpu/drm/nouveau/nouveau_encoder.h b/drivers/gpu/drm/nouveau/nouveau_encoder.h
index e24341229d5e..24660c0f713d 100644
--- a/drivers/gpu/drm/nouveau/nouveau_encoder.h
+++ b/drivers/gpu/drm/nouveau/nouveau_encoder.h
@@ -30,7 +30,7 @@
 #include <subdev/bios/dcb.h>
 
 #include <drm/drm_encoder_slave.h>
-#include "nv04_display.h"
+#include "dispnv04/disp.h"
 
 #define NV_DPMS_CLEARED 0x80
 
diff --git a/drivers/gpu/drm/nouveau/nouveau_irq.c b/drivers/gpu/drm/nouveau/nouveau_irq.c
deleted file mode 100644
index 1303680affd3..000000000000
--- a/drivers/gpu/drm/nouveau/nouveau_irq.c
+++ /dev/null
@@ -1,76 +0,0 @@
-/*
- * Copyright 2012 Red Hat Inc.
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * the rights to use, copy, modify, merge, publish, distribute, sublicense,
- * and/or sell copies of the Software, and to permit persons to whom the
- * Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in
- * all copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
- * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
- * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
- * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
- * OTHER DEALINGS IN THE SOFTWARE.
- *
- * Authors: Ben Skeggs
- */
-
-#include <subdev/mc.h>
-
-#include "nouveau_drm.h"
-#include "nouveau_irq.h"
-#include "nv50_display.h"
-
-void
-nouveau_irq_preinstall(struct drm_device *dev)
-{
-	nv_wr32(nouveau_dev(dev), 0x000140, 0x00000000);
-}
-
-int
-nouveau_irq_postinstall(struct drm_device *dev)
-{
-	nv_wr32(nouveau_dev(dev), 0x000140, 0x00000001);
-	return 0;
-}
-
-void
-nouveau_irq_uninstall(struct drm_device *dev)
-{
-	nv_wr32(nouveau_dev(dev), 0x000140, 0x00000000);
-}
-
-irqreturn_t
-nouveau_irq_handler(DRM_IRQ_ARGS)
-{
-	struct drm_device *dev = arg;
-	struct nouveau_device *device = nouveau_dev(dev);
-	struct nouveau_mc *pmc = nouveau_mc(device);
-	u32 stat;
-
-	stat = nv_rd32(device, 0x000100);
-	if (stat == 0 || stat == ~0)
-		return IRQ_NONE;
-
-	nv_subdev(pmc)->intr(nv_subdev(pmc));
-	return IRQ_HANDLED;
-}
-
-int
-nouveau_irq_init(struct drm_device *dev)
-{
-	return drm_irq_install(dev);
-}
-
-void
-nouveau_irq_fini(struct drm_device *dev)
-{
-	drm_irq_uninstall(dev);
-}
diff --git a/drivers/gpu/drm/nouveau/nouveau_irq.h b/drivers/gpu/drm/nouveau/nouveau_irq.h
deleted file mode 100644
index 06714ad857bb..000000000000
--- a/drivers/gpu/drm/nouveau/nouveau_irq.h
+++ /dev/null
@@ -1,11 +0,0 @@
-#ifndef __NOUVEAU_IRQ_H__
-#define __NOUVEAU_IRQ_H__
-
-extern int         nouveau_irq_init(struct drm_device *);
-extern void        nouveau_irq_fini(struct drm_device *);
-extern irqreturn_t nouveau_irq_handler(DRM_IRQ_ARGS);
-extern void        nouveau_irq_preinstall(struct drm_device *);
-extern int         nouveau_irq_postinstall(struct drm_device *);
-extern void        nouveau_irq_uninstall(struct drm_device *);
-
-#endif
diff --git a/drivers/gpu/drm/nouveau/nouveau_ttm.c b/drivers/gpu/drm/nouveau/nouveau_ttm.c
index 9be9cb58e19b..f19a15a3bc03 100644
--- a/drivers/gpu/drm/nouveau/nouveau_ttm.c
+++ b/drivers/gpu/drm/nouveau/nouveau_ttm.c
@@ -35,14 +35,16 @@
 static int
 nouveau_vram_manager_init(struct ttm_mem_type_manager *man, unsigned long psize)
 {
-	/* nothing to do */
+	struct nouveau_drm *drm = nouveau_bdev(man->bdev);
+	struct nouveau_fb *pfb = nouveau_fb(drm->device);
+	man->priv = pfb;
 	return 0;
 }
 
 static int
 nouveau_vram_manager_fini(struct ttm_mem_type_manager *man)
 {
-	/* nothing to do */
+	man->priv = NULL;
 	return 0;
 }
 
@@ -104,7 +106,8 @@ nouveau_vram_manager_new(struct ttm_mem_type_manager *man,
 static void
 nouveau_vram_manager_debug(struct ttm_mem_type_manager *man, const char *prefix)
 {
-	struct nouveau_mm *mm = man->priv;
+	struct nouveau_fb *pfb = man->priv;
+	struct nouveau_mm *mm = &pfb->vram;
 	struct nouveau_mm_node *r;
 	u32 total = 0, free = 0;
 
@@ -161,6 +164,8 @@ nouveau_gart_manager_new(struct ttm_mem_type_manager *man,
 			 struct ttm_placement *placement,
 			 struct ttm_mem_reg *mem)
 {
+	struct nouveau_drm *drm = nouveau_bdev(bo->bdev);
+	struct nouveau_bo *nvbo = nouveau_bo(bo);
 	struct nouveau_mem *node;
 
 	if (unlikely((mem->num_pages << PAGE_SHIFT) >= 512 * 1024 * 1024))
@@ -171,6 +176,20 @@ nouveau_gart_manager_new(struct ttm_mem_type_manager *man,
 		return -ENOMEM;
 	node->page_shift = 12;
 
+	switch (nv_device(drm->device)->card_type) {
+	case NV_50:
+		if (nv_device(drm->device)->chipset != 0x50)
+			node->memtype = (nvbo->tile_flags & 0x7f00) >> 8;
+		break;
+	case NV_C0:
+	case NV_D0:
+	case NV_E0:
+		node->memtype = (nvbo->tile_flags & 0xff00) >> 8;
+		break;
+	default:
+		break;
+	}
+
 	mem->mm_node = node;
 	mem->start   = 0;
 	return 0;
diff --git a/drivers/gpu/drm/nouveau/nv04_pm.c b/drivers/gpu/drm/nouveau/nv04_pm.c
index 2a0cc9d0614a..27afc0ea28b0 100644
--- a/drivers/gpu/drm/nouveau/nv04_pm.c
+++ b/drivers/gpu/drm/nouveau/nv04_pm.c
@@ -25,7 +25,7 @@
 #include <drm/drmP.h>
 #include "nouveau_drm.h"
 #include "nouveau_reg.h"
-#include "nouveau_hw.h"
+#include "dispnv04/hw.h"
 #include "nouveau_pm.h"
 
 #include <subdev/bios/pll.h>
diff --git a/drivers/gpu/drm/nouveau/nv40_pm.c b/drivers/gpu/drm/nouveau/nv40_pm.c
index 3382064c7f33..3af5bcd0b203 100644
--- a/drivers/gpu/drm/nouveau/nv40_pm.c
+++ b/drivers/gpu/drm/nouveau/nv40_pm.c
@@ -26,7 +26,7 @@
 #include "nouveau_drm.h"
 #include "nouveau_bios.h"
 #include "nouveau_pm.h"
-#include "nouveau_hw.h"
+#include "dispnv04/hw.h"
 
 #include <subdev/bios/pll.h>
 #include <subdev/clock.h>
diff --git a/drivers/gpu/drm/nouveau/nv50_display.c b/drivers/gpu/drm/nouveau/nv50_display.c
index 1ddc03e51bf4..ebf0a683305e 100644
--- a/drivers/gpu/drm/nouveau/nv50_display.c
+++ b/drivers/gpu/drm/nouveau/nv50_display.c
@@ -2174,6 +2174,7 @@ int
 nv50_display_create(struct drm_device *dev)
 {
 	static const u16 oclass[] = {
+		NVF0_DISP_CLASS,
 		NVE0_DISP_CLASS,
 		NVD0_DISP_CLASS,
 		NVA3_DISP_CLASS,
diff --git a/drivers/gpu/drm/nouveau/nv50_pm.c b/drivers/gpu/drm/nouveau/nv50_pm.c
index 8bd5d2781baf..69620e39c90c 100644
--- a/drivers/gpu/drm/nouveau/nv50_pm.c
+++ b/drivers/gpu/drm/nouveau/nv50_pm.c
@@ -25,7 +25,7 @@
 #include <drm/drmP.h>
 #include "nouveau_drm.h"
 #include "nouveau_bios.h"
-#include "nouveau_hw.h"
+#include "dispnv04/hw.h"
 #include "nouveau_pm.h"
 #include "nouveau_hwsq.h"
 
diff --git a/drivers/gpu/drm/omapdrm/omap_connector.c b/drivers/gpu/drm/omapdrm/omap_connector.c
index c451c41a7a7d..912759daf562 100644
--- a/drivers/gpu/drm/omapdrm/omap_connector.c
+++ b/drivers/gpu/drm/omapdrm/omap_connector.c
@@ -110,6 +110,11 @@ static enum drm_connector_status omap_connector_detect(
 			ret = connector_status_connected;
 		else
 			ret = connector_status_disconnected;
+	} else if (dssdev->type == OMAP_DISPLAY_TYPE_DPI ||
+			dssdev->type == OMAP_DISPLAY_TYPE_DBI ||
+			dssdev->type == OMAP_DISPLAY_TYPE_SDI ||
+			dssdev->type == OMAP_DISPLAY_TYPE_DSI) {
+		ret = connector_status_connected;
 	} else {
 		ret = connector_status_unknown;
 	}
@@ -189,12 +194,30 @@ static int omap_connector_mode_valid(struct drm_connector *connector,
 	struct omap_video_timings timings = {0};
 	struct drm_device *dev = connector->dev;
 	struct drm_display_mode *new_mode;
-	int ret = MODE_BAD;
+	int r, ret = MODE_BAD;
 
 	copy_timings_drm_to_omap(&timings, mode);
 	mode->vrefresh = drm_mode_vrefresh(mode);
 
-	if (!dssdrv->check_timings(dssdev, &timings)) {
+	/*
+	 * if the panel driver doesn't have a check_timings, it's most likely
+	 * a fixed resolution panel, check if the timings match with the
+	 * panel's timings
+	 */
+	if (dssdrv->check_timings) {
+		r = dssdrv->check_timings(dssdev, &timings);
+	} else {
+		struct omap_video_timings t = {0};
+
+		dssdrv->get_timings(dssdev, &t);
+
+		if (memcmp(&timings, &t, sizeof(struct omap_video_timings)))
+			r = -EINVAL;
+		else
+			r = 0;
+	}
+
+	if (!r) {
 		/* check if vrefresh is still valid */
 		new_mode = drm_mode_duplicate(dev, mode);
 		new_mode->clock = timings.pixel_clock;
diff --git a/drivers/gpu/drm/omapdrm/omap_crtc.c b/drivers/gpu/drm/omapdrm/omap_crtc.c
index bec66a490b8f..79b200aee18a 100644
--- a/drivers/gpu/drm/omapdrm/omap_crtc.c
+++ b/drivers/gpu/drm/omapdrm/omap_crtc.c
@@ -74,6 +74,13 @@ struct omap_crtc {
 	struct work_struct page_flip_work;
 };
 
+uint32_t pipe2vbl(struct drm_crtc *crtc)
+{
+	struct omap_crtc *omap_crtc = to_omap_crtc(crtc);
+
+	return dispc_mgr_get_vsync_irq(omap_crtc->channel);
+}
+
 /*
  * Manager-ops, callbacks from output when they need to configure
  * the upstream part of the video pipe.
@@ -613,7 +620,13 @@ struct drm_crtc *omap_crtc_init(struct drm_device *dev,
 	omap_crtc->apply.pre_apply  = omap_crtc_pre_apply;
 	omap_crtc->apply.post_apply = omap_crtc_post_apply;
 
-	omap_crtc->apply_irq.irqmask = pipe2vbl(id);
+	omap_crtc->channel = channel;
+	omap_crtc->plane = plane;
+	omap_crtc->plane->crtc = crtc;
+	omap_crtc->name = channel_names[channel];
+	omap_crtc->pipe = id;
+
+	omap_crtc->apply_irq.irqmask = pipe2vbl(crtc);
 	omap_crtc->apply_irq.irq = omap_crtc_apply_irq;
 
 	omap_crtc->error_irq.irqmask =
@@ -621,12 +634,6 @@ struct drm_crtc *omap_crtc_init(struct drm_device *dev,
 	omap_crtc->error_irq.irq = omap_crtc_error_irq;
 	omap_irq_register(dev, &omap_crtc->error_irq);
 
-	omap_crtc->channel = channel;
-	omap_crtc->plane = plane;
-	omap_crtc->plane->crtc = crtc;
-	omap_crtc->name = channel_names[channel];
-	omap_crtc->pipe = id;
-
 	/* temporary: */
 	omap_crtc->mgr.id = channel;
 
diff --git a/drivers/gpu/drm/omapdrm/omap_drv.c b/drivers/gpu/drm/omapdrm/omap_drv.c
index 079c54c6f94c..9c53c25e5201 100644
--- a/drivers/gpu/drm/omapdrm/omap_drv.c
+++ b/drivers/gpu/drm/omapdrm/omap_drv.c
@@ -74,54 +74,53 @@ static int get_connector_type(struct omap_dss_device *dssdev)
 	}
 }
 
+static bool channel_used(struct drm_device *dev, enum omap_channel channel)
+{
+	struct omap_drm_private *priv = dev->dev_private;
+	int i;
+
+	for (i = 0; i < priv->num_crtcs; i++) {
+		struct drm_crtc *crtc = priv->crtcs[i];
+
+		if (omap_crtc_channel(crtc) == channel)
+			return true;
+	}
+
+	return false;
+}
+
 static int omap_modeset_init(struct drm_device *dev)
 {
 	struct omap_drm_private *priv = dev->dev_private;
 	struct omap_dss_device *dssdev = NULL;
 	int num_ovls = dss_feat_get_num_ovls();
-	int id;
+	int num_mgrs = dss_feat_get_num_mgrs();
+	int num_crtcs;
+	int i, id = 0;
 
 	drm_mode_config_init(dev);
 
 	omap_drm_irq_install(dev);
 
 	/*
-	 * Create private planes and CRTCs for the last NUM_CRTCs overlay
-	 * plus manager:
+	 * We usually don't want to create a CRTC for each manager, at least
+	 * not until we have a way to expose private planes to userspace.
+	 * Otherwise there would not be enough video pipes left for drm planes.
+	 * We use the num_crtc argument to limit the number of crtcs we create.
 	 */
-	for (id = 0; id < min(num_crtc, num_ovls); id++) {
-		struct drm_plane *plane;
-		struct drm_crtc *crtc;
-
-		plane = omap_plane_init(dev, id, true);
-		crtc = omap_crtc_init(dev, plane, pipe2chan(id), id);
+	num_crtcs = min3(num_crtc, num_mgrs, num_ovls);
 
-		BUG_ON(priv->num_crtcs >= ARRAY_SIZE(priv->crtcs));
-		priv->crtcs[id] = crtc;
-		priv->num_crtcs++;
-
-		priv->planes[id] = plane;
-		priv->num_planes++;
-	}
-
-	/*
-	 * Create normal planes for the remaining overlays:
-	 */
-	for (; id < num_ovls; id++) {
-		struct drm_plane *plane = omap_plane_init(dev, id, false);
-
-		BUG_ON(priv->num_planes >= ARRAY_SIZE(priv->planes));
-		priv->planes[priv->num_planes++] = plane;
-	}
+	dssdev = NULL;
 
 	for_each_dss_dev(dssdev) {
 		struct drm_connector *connector;
 		struct drm_encoder *encoder;
+		enum omap_channel channel;
 
 		if (!dssdev->driver) {
 			dev_warn(dev->dev, "%s has no driver.. skipping it\n",
 					dssdev->name);
-			return 0;
+			continue;
 		}
 
 		if (!(dssdev->driver->get_timings ||
@@ -129,7 +128,7 @@ static int omap_modeset_init(struct drm_device *dev)
 			dev_warn(dev->dev, "%s driver does not support "
 				"get_timings or read_edid.. skipping it!\n",
 				dssdev->name);
-			return 0;
+			continue;
 		}
 
 		encoder = omap_encoder_init(dev, dssdev);
@@ -157,16 +156,118 @@ static int omap_modeset_init(struct drm_device *dev)
 
 		drm_mode_connector_attach_encoder(connector, encoder);
 
+		/*
+		 * if we have reached the limit of the crtcs we are allowed to
+		 * create, let's not try to look for a crtc for this
+		 * panel/encoder and onwards, we will, of course, populate the
+		 * the possible_crtcs field for all the encoders with the final
+		 * set of crtcs we create
+		 */
+		if (id == num_crtcs)
+			continue;
+
+		/*
+		 * get the recommended DISPC channel for this encoder. For now,
+		 * we only try to get create a crtc out of the recommended, the
+		 * other possible channels to which the encoder can connect are
+		 * not considered.
+		 */
+		channel = dssdev->output->dispc_channel;
+
+		/*
+		 * if this channel hasn't already been taken by a previously
+		 * allocated crtc, we create a new crtc for it
+		 */
+		if (!channel_used(dev, channel)) {
+			struct drm_plane *plane;
+			struct drm_crtc *crtc;
+
+			plane = omap_plane_init(dev, id, true);
+			crtc = omap_crtc_init(dev, plane, channel, id);
+
+			BUG_ON(priv->num_crtcs >= ARRAY_SIZE(priv->crtcs));
+			priv->crtcs[id] = crtc;
+			priv->num_crtcs++;
+
+			priv->planes[id] = plane;
+			priv->num_planes++;
+
+			id++;
+		}
+	}
+
+	/*
+	 * we have allocated crtcs according to the need of the panels/encoders,
+	 * adding more crtcs here if needed
+	 */
+	for (; id < num_crtcs; id++) {
+
+		/* find a free manager for this crtc */
+		for (i = 0; i < num_mgrs; i++) {
+			if (!channel_used(dev, i)) {
+				struct drm_plane *plane;
+				struct drm_crtc *crtc;
+
+				plane = omap_plane_init(dev, id, true);
+				crtc = omap_crtc_init(dev, plane, i, id);
+
+				BUG_ON(priv->num_crtcs >=
+					ARRAY_SIZE(priv->crtcs));
+
+				priv->crtcs[id] = crtc;
+				priv->num_crtcs++;
+
+				priv->planes[id] = plane;
+				priv->num_planes++;
+
+				break;
+			} else {
+				continue;
+			}
+		}
+
+		if (i == num_mgrs) {
+			/* this shouldn't really happen */
+			dev_err(dev->dev, "no managers left for crtc\n");
+			return -ENOMEM;
+		}
+	}
+
+	/*
+	 * Create normal planes for the remaining overlays:
+	 */
+	for (; id < num_ovls; id++) {
+		struct drm_plane *plane = omap_plane_init(dev, id, false);
+
+		BUG_ON(priv->num_planes >= ARRAY_SIZE(priv->planes));
+		priv->planes[priv->num_planes++] = plane;
+	}
+
+	for (i = 0; i < priv->num_encoders; i++) {
+		struct drm_encoder *encoder = priv->encoders[i];
+		struct omap_dss_device *dssdev =
+					omap_encoder_get_dssdev(encoder);
+
 		/* figure out which crtc's we can connect the encoder to: */
 		encoder->possible_crtcs = 0;
 		for (id = 0; id < priv->num_crtcs; id++) {
-			enum omap_dss_output_id supported_outputs =
-					dss_feat_get_supported_outputs(pipe2chan(id));
+			struct drm_crtc *crtc = priv->crtcs[id];
+			enum omap_channel crtc_channel;
+			enum omap_dss_output_id supported_outputs;
+
+			crtc_channel = omap_crtc_channel(crtc);
+			supported_outputs =
+				dss_feat_get_supported_outputs(crtc_channel);
+
 			if (supported_outputs & dssdev->output->id)
 				encoder->possible_crtcs |= (1 << id);
 		}
 	}
 
+	DBG("registered %d planes, %d crtcs, %d encoders and %d connectors\n",
+		priv->num_planes, priv->num_crtcs, priv->num_encoders,
+		priv->num_connectors);
+
 	dev->mode_config.min_width = 32;
 	dev->mode_config.min_height = 32;
 
@@ -303,7 +404,7 @@ static int ioctl_gem_info(struct drm_device *dev, void *data,
 	return ret;
 }
 
-struct drm_ioctl_desc ioctls[DRM_COMMAND_END - DRM_COMMAND_BASE] = {
+static struct drm_ioctl_desc ioctls[DRM_COMMAND_END - DRM_COMMAND_BASE] = {
 	DRM_IOCTL_DEF_DRV(OMAP_GET_PARAM, ioctl_get_param, DRM_UNLOCKED|DRM_AUTH),
 	DRM_IOCTL_DEF_DRV(OMAP_SET_PARAM, ioctl_set_param, DRM_UNLOCKED|DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
 	DRM_IOCTL_DEF_DRV(OMAP_GEM_NEW, ioctl_gem_new, DRM_UNLOCKED|DRM_AUTH),
@@ -567,7 +668,7 @@ static const struct dev_pm_ops omapdrm_pm_ops = {
 };
 #endif
 
-struct platform_driver pdev = {
+static struct platform_driver pdev = {
 		.driver = {
 			.name = DRIVER_NAME,
 			.owner = THIS_MODULE,
diff --git a/drivers/gpu/drm/omapdrm/omap_drv.h b/drivers/gpu/drm/omapdrm/omap_drv.h
index d4f997bb4ac0..215a20dd340c 100644
--- a/drivers/gpu/drm/omapdrm/omap_drv.h
+++ b/drivers/gpu/drm/omapdrm/omap_drv.h
@@ -139,8 +139,8 @@ void omap_gem_describe_objects(struct list_head *list, struct seq_file *m);
 int omap_gem_resume(struct device *dev);
 #endif
 
-int omap_irq_enable_vblank(struct drm_device *dev, int crtc);
-void omap_irq_disable_vblank(struct drm_device *dev, int crtc);
+int omap_irq_enable_vblank(struct drm_device *dev, int crtc_id);
+void omap_irq_disable_vblank(struct drm_device *dev, int crtc_id);
 irqreturn_t omap_irq_handler(DRM_IRQ_ARGS);
 void omap_irq_preinstall(struct drm_device *dev);
 int omap_irq_postinstall(struct drm_device *dev);
@@ -271,39 +271,9 @@ static inline int align_pitch(int pitch, int width, int bpp)
 	return ALIGN(pitch, 8 * bytespp);
 }
 
-static inline enum omap_channel pipe2chan(int pipe)
-{
-	int num_mgrs = dss_feat_get_num_mgrs();
-
-	/*
-	 * We usually don't want to create a CRTC for each manager,
-	 * at least not until we have a way to expose private planes
-	 * to userspace.  Otherwise there would not be enough video
-	 * pipes left for drm planes.  The higher #'d managers tend
-	 * to have more features so start in reverse order.
-	 */
-	return num_mgrs - pipe - 1;
-}
-
 /* map crtc to vblank mask */
-static inline uint32_t pipe2vbl(int crtc)
-{
-	enum omap_channel channel = pipe2chan(crtc);
-	return dispc_mgr_get_vsync_irq(channel);
-}
-
-static inline int crtc2pipe(struct drm_device *dev, struct drm_crtc *crtc)
-{
-	struct omap_drm_private *priv = dev->dev_private;
-	int i;
-
-	for (i = 0; i < ARRAY_SIZE(priv->crtcs); i++)
-		if (priv->crtcs[i] == crtc)
-			return i;
-
-	BUG();  /* bogus CRTC ptr */
-	return -1;
-}
+uint32_t pipe2vbl(struct drm_crtc *crtc);
+struct omap_dss_device *omap_encoder_get_dssdev(struct drm_encoder *encoder);
 
 /* should these be made into common util helpers?
  */
diff --git a/drivers/gpu/drm/omapdrm/omap_encoder.c b/drivers/gpu/drm/omapdrm/omap_encoder.c
index 21d126d0317e..c29451ba65da 100644
--- a/drivers/gpu/drm/omapdrm/omap_encoder.c
+++ b/drivers/gpu/drm/omapdrm/omap_encoder.c
@@ -41,6 +41,13 @@ struct omap_encoder {
 	struct omap_dss_device *dssdev;
 };
 
+struct omap_dss_device *omap_encoder_get_dssdev(struct drm_encoder *encoder)
+{
+	struct omap_encoder *omap_encoder = to_omap_encoder(encoder);
+
+	return omap_encoder->dssdev;
+}
+
 static void omap_encoder_destroy(struct drm_encoder *encoder)
 {
 	struct omap_encoder *omap_encoder = to_omap_encoder(encoder);
@@ -128,13 +135,26 @@ int omap_encoder_update(struct drm_encoder *encoder,
 
 	dssdev->output->manager = mgr;
 
-	ret = dssdrv->check_timings(dssdev, timings);
+	if (dssdrv->check_timings) {
+		ret = dssdrv->check_timings(dssdev, timings);
+	} else {
+		struct omap_video_timings t = {0};
+
+		dssdrv->get_timings(dssdev, &t);
+
+		if (memcmp(timings, &t, sizeof(struct omap_video_timings)))
+			ret = -EINVAL;
+		else
+			ret = 0;
+	}
+
 	if (ret) {
 		dev_err(dev->dev, "could not set timings: %d\n", ret);
 		return ret;
 	}
 
-	dssdrv->set_timings(dssdev, timings);
+	if (dssdrv->set_timings)
+		dssdrv->set_timings(dssdev, timings);
 
 	return 0;
 }
diff --git a/drivers/gpu/drm/omapdrm/omap_gem_dmabuf.c b/drivers/gpu/drm/omapdrm/omap_gem_dmabuf.c
index ac74d1bc67bf..be7cd97a0db0 100644
--- a/drivers/gpu/drm/omapdrm/omap_gem_dmabuf.c
+++ b/drivers/gpu/drm/omapdrm/omap_gem_dmabuf.c
@@ -178,7 +178,7 @@ out_unlock:
 	return omap_gem_mmap_obj(obj, vma);
 }
 
-struct dma_buf_ops omap_dmabuf_ops = {
+static struct dma_buf_ops omap_dmabuf_ops = {
 		.map_dma_buf = omap_gem_map_dma_buf,
 		.unmap_dma_buf = omap_gem_unmap_dma_buf,
 		.release = omap_gem_dmabuf_release,
@@ -212,7 +212,6 @@ struct drm_gem_object *omap_gem_prime_import(struct drm_device *dev,
 			 * refcount on gem itself instead of f_count of dmabuf.
 			 */
 			drm_gem_object_reference(obj);
-			dma_buf_put(buffer);
 			return obj;
 		}
 	}
diff --git a/drivers/gpu/drm/omapdrm/omap_irq.c b/drivers/gpu/drm/omapdrm/omap_irq.c
index e01303ee00c3..9263db117ff8 100644
--- a/drivers/gpu/drm/omapdrm/omap_irq.c
+++ b/drivers/gpu/drm/omapdrm/omap_irq.c
@@ -130,12 +130,13 @@ int omap_irq_wait(struct drm_device *dev, struct omap_irq_wait *wait,
  * Zero on success, appropriate errno if the given @crtc's vblank
  * interrupt cannot be enabled.
  */
-int omap_irq_enable_vblank(struct drm_device *dev, int crtc)
+int omap_irq_enable_vblank(struct drm_device *dev, int crtc_id)
 {
 	struct omap_drm_private *priv = dev->dev_private;
+	struct drm_crtc *crtc = priv->crtcs[crtc_id];
 	unsigned long flags;
 
-	DBG("dev=%p, crtc=%d", dev, crtc);
+	DBG("dev=%p, crtc=%d", dev, crtc_id);
 
 	dispc_runtime_get();
 	spin_lock_irqsave(&list_lock, flags);
@@ -156,12 +157,13 @@ int omap_irq_enable_vblank(struct drm_device *dev, int crtc)
  * a hardware vblank counter, this routine should be a no-op, since
  * interrupts will have to stay on to keep the count accurate.
  */
-void omap_irq_disable_vblank(struct drm_device *dev, int crtc)
+void omap_irq_disable_vblank(struct drm_device *dev, int crtc_id)
 {
 	struct omap_drm_private *priv = dev->dev_private;
+	struct drm_crtc *crtc = priv->crtcs[crtc_id];
 	unsigned long flags;
 
-	DBG("dev=%p, crtc=%d", dev, crtc);
+	DBG("dev=%p, crtc=%d", dev, crtc_id);
 
 	dispc_runtime_get();
 	spin_lock_irqsave(&list_lock, flags);
@@ -186,9 +188,12 @@ irqreturn_t omap_irq_handler(DRM_IRQ_ARGS)
 
 	VERB("irqs: %08x", irqstatus);
 
-	for (id = 0; id < priv->num_crtcs; id++)
-		if (irqstatus & pipe2vbl(id))
+	for (id = 0; id < priv->num_crtcs; id++) {
+		struct drm_crtc *crtc = priv->crtcs[id];
+
+		if (irqstatus & pipe2vbl(crtc))
 			drm_handle_vblank(dev, id);
+	}
 
 	spin_lock_irqsave(&list_lock, flags);
 	list_for_each_entry_safe(handler, n, &priv->irq_list, node) {
diff --git a/drivers/gpu/drm/omapdrm/omap_plane.c b/drivers/gpu/drm/omapdrm/omap_plane.c
index 2882cda6ea19..8d225d7ff4e3 100644
--- a/drivers/gpu/drm/omapdrm/omap_plane.c
+++ b/drivers/gpu/drm/omapdrm/omap_plane.c
@@ -247,6 +247,12 @@ static int omap_plane_update(struct drm_plane *plane,
 {
 	struct omap_plane *omap_plane = to_omap_plane(plane);
 	omap_plane->enabled = true;
+
+	if (plane->fb)
+		drm_framebuffer_unreference(plane->fb);
+
+	drm_framebuffer_reference(fb);
+
 	return omap_plane_mode_set(plane, crtc, fb,
 			crtc_x, crtc_y, crtc_w, crtc_h,
 			src_x, src_y, src_w, src_h,
diff --git a/drivers/gpu/drm/qxl/Kconfig b/drivers/gpu/drm/qxl/Kconfig
new file mode 100644
index 000000000000..2f1a57e11140
--- /dev/null
+++ b/drivers/gpu/drm/qxl/Kconfig
@@ -0,0 +1,10 @@
+config DRM_QXL
+	tristate "QXL virtual GPU"
+	depends on DRM && PCI
+	select FB_SYS_FILLRECT
+	select FB_SYS_COPYAREA
+	select FB_SYS_IMAGEBLIT
+        select DRM_KMS_HELPER
+        select DRM_TTM
+	help
+		QXL virtual GPU for Spice virtualization desktop integration. Do not enable this driver unless your distro ships a corresponding X.org QXL driver that can handle kernel modesetting.
diff --git a/drivers/gpu/drm/qxl/Makefile b/drivers/gpu/drm/qxl/Makefile
new file mode 100644
index 000000000000..ea046ba691d2
--- /dev/null
+++ b/drivers/gpu/drm/qxl/Makefile
@@ -0,0 +1,9 @@
+#
+# Makefile for the drm device driver.  This driver provides support for the
+# Direct Rendering Infrastructure (DRI) in XFree86 4.1.0 and higher.
+
+ccflags-y := -Iinclude/drm
+
+qxl-y := qxl_drv.o qxl_kms.o qxl_display.o qxl_ttm.o qxl_fb.o qxl_object.o qxl_gem.o qxl_cmd.o qxl_image.o qxl_draw.o qxl_debugfs.o qxl_irq.o qxl_dumb.o qxl_ioctl.o qxl_fence.o qxl_release.o
+
+obj-$(CONFIG_DRM_QXL)+= qxl.o
diff --git a/drivers/gpu/drm/qxl/qxl_cmd.c b/drivers/gpu/drm/qxl/qxl_cmd.c
new file mode 100644
index 000000000000..08b0823c93d5
--- /dev/null
+++ b/drivers/gpu/drm/qxl/qxl_cmd.c
@@ -0,0 +1,685 @@
+/*
+ * Copyright 2013 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Dave Airlie
+ *          Alon Levy
+ */
+
+/* QXL cmd/ring handling */
+
+#include "qxl_drv.h"
+#include "qxl_object.h"
+
+static int qxl_reap_surface_id(struct qxl_device *qdev, int max_to_reap);
+
+struct ring {
+	struct qxl_ring_header      header;
+	uint8_t                     elements[0];
+};
+
+struct qxl_ring {
+	struct ring	       *ring;
+	int			element_size;
+	int			n_elements;
+	int			prod_notify;
+	wait_queue_head_t      *push_event;
+	spinlock_t             lock;
+};
+
+void qxl_ring_free(struct qxl_ring *ring)
+{
+	kfree(ring);
+}
+
+struct qxl_ring *
+qxl_ring_create(struct qxl_ring_header *header,
+		int element_size,
+		int n_elements,
+		int prod_notify,
+		bool set_prod_notify,
+		wait_queue_head_t *push_event)
+{
+	struct qxl_ring *ring;
+
+	ring = kmalloc(sizeof(*ring), GFP_KERNEL);
+	if (!ring)
+		return NULL;
+
+	ring->ring = (struct ring *)header;
+	ring->element_size = element_size;
+	ring->n_elements = n_elements;
+	ring->prod_notify = prod_notify;
+	ring->push_event = push_event;
+	if (set_prod_notify)
+		header->notify_on_prod = ring->n_elements;
+	spin_lock_init(&ring->lock);
+	return ring;
+}
+
+static int qxl_check_header(struct qxl_ring *ring)
+{
+	int ret;
+	struct qxl_ring_header *header = &(ring->ring->header);
+	unsigned long flags;
+	spin_lock_irqsave(&ring->lock, flags);
+	ret = header->prod - header->cons < header->num_items;
+	if (ret == 0)
+		header->notify_on_cons = header->cons + 1;
+	spin_unlock_irqrestore(&ring->lock, flags);
+	return ret;
+}
+
+static int qxl_check_idle(struct qxl_ring *ring)
+{
+	int ret;
+	struct qxl_ring_header *header = &(ring->ring->header);
+	unsigned long flags;
+	spin_lock_irqsave(&ring->lock, flags);
+	ret = header->prod == header->cons;
+	spin_unlock_irqrestore(&ring->lock, flags);
+	return ret;
+}
+
+int qxl_ring_push(struct qxl_ring *ring,
+		  const void *new_elt, bool interruptible)
+{
+	struct qxl_ring_header *header = &(ring->ring->header);
+	uint8_t *elt;
+	int idx, ret;
+	unsigned long flags;
+	spin_lock_irqsave(&ring->lock, flags);
+	if (header->prod - header->cons == header->num_items) {
+		header->notify_on_cons = header->cons + 1;
+		mb();
+		spin_unlock_irqrestore(&ring->lock, flags);
+		if (!drm_can_sleep()) {
+			while (!qxl_check_header(ring))
+				udelay(1);
+		} else {
+			if (interruptible) {
+				ret = wait_event_interruptible(*ring->push_event,
+							       qxl_check_header(ring));
+				if (ret)
+					return ret;
+			} else {
+				wait_event(*ring->push_event,
+					   qxl_check_header(ring));
+			}
+
+		}
+		spin_lock_irqsave(&ring->lock, flags);
+	}
+
+	idx = header->prod & (ring->n_elements - 1);
+	elt = ring->ring->elements + idx * ring->element_size;
+
+	memcpy((void *)elt, new_elt, ring->element_size);
+
+	header->prod++;
+
+	mb();
+
+	if (header->prod == header->notify_on_prod)
+		outb(0, ring->prod_notify);
+
+	spin_unlock_irqrestore(&ring->lock, flags);
+	return 0;
+}
+
+static bool qxl_ring_pop(struct qxl_ring *ring,
+			 void *element)
+{
+	volatile struct qxl_ring_header *header = &(ring->ring->header);
+	volatile uint8_t *ring_elt;
+	int idx;
+	unsigned long flags;
+	spin_lock_irqsave(&ring->lock, flags);
+	if (header->cons == header->prod) {
+		header->notify_on_prod = header->cons + 1;
+		spin_unlock_irqrestore(&ring->lock, flags);
+		return false;
+	}
+
+	idx = header->cons & (ring->n_elements - 1);
+	ring_elt = ring->ring->elements + idx * ring->element_size;
+
+	memcpy(element, (void *)ring_elt, ring->element_size);
+
+	header->cons++;
+
+	spin_unlock_irqrestore(&ring->lock, flags);
+	return true;
+}
+
+int
+qxl_push_command_ring_release(struct qxl_device *qdev, struct qxl_release *release,
+			      uint32_t type, bool interruptible)
+{
+	struct qxl_command cmd;
+
+	cmd.type = type;
+	cmd.data = qxl_bo_physical_address(qdev, release->bos[0], release->release_offset);
+
+	return qxl_ring_push(qdev->command_ring, &cmd, interruptible);
+}
+
+int
+qxl_push_cursor_ring_release(struct qxl_device *qdev, struct qxl_release *release,
+			     uint32_t type, bool interruptible)
+{
+	struct qxl_command cmd;
+
+	cmd.type = type;
+	cmd.data = qxl_bo_physical_address(qdev, release->bos[0], release->release_offset);
+
+	return qxl_ring_push(qdev->cursor_ring, &cmd, interruptible);
+}
+
+bool qxl_queue_garbage_collect(struct qxl_device *qdev, bool flush)
+{
+	if (!qxl_check_idle(qdev->release_ring)) {
+		queue_work(qdev->gc_queue, &qdev->gc_work);
+		if (flush)
+			flush_work(&qdev->gc_work);
+		return true;
+	}
+	return false;
+}
+
+int qxl_garbage_collect(struct qxl_device *qdev)
+{
+	struct qxl_release *release;
+	uint64_t id, next_id;
+	int i = 0;
+	int ret;
+	union qxl_release_info *info;
+
+	while (qxl_ring_pop(qdev->release_ring, &id)) {
+		QXL_INFO(qdev, "popped %lld\n", id);
+		while (id) {
+			release = qxl_release_from_id_locked(qdev, id);
+			if (release == NULL)
+				break;
+
+			ret = qxl_release_reserve(qdev, release, false);
+			if (ret) {
+				qxl_io_log(qdev, "failed to reserve release on garbage collect %lld\n", id);
+				DRM_ERROR("failed to reserve release %lld\n", id);
+			}
+
+			info = qxl_release_map(qdev, release);
+			next_id = info->next;
+			qxl_release_unmap(qdev, release, info);
+
+			qxl_release_unreserve(qdev, release);
+			QXL_INFO(qdev, "popped %lld, next %lld\n", id,
+				next_id);
+
+			switch (release->type) {
+			case QXL_RELEASE_DRAWABLE:
+			case QXL_RELEASE_SURFACE_CMD:
+			case QXL_RELEASE_CURSOR_CMD:
+				break;
+			default:
+				DRM_ERROR("unexpected release type\n");
+				break;
+			}
+			id = next_id;
+
+			qxl_release_free(qdev, release);
+			++i;
+		}
+	}
+
+	QXL_INFO(qdev, "%s: %lld\n", __func__, i);
+
+	return i;
+}
+
+int qxl_alloc_bo_reserved(struct qxl_device *qdev, unsigned long size,
+			  struct qxl_bo **_bo)
+{
+	struct qxl_bo *bo;
+	int ret;
+
+	ret = qxl_bo_create(qdev, size, false /* not kernel - device */,
+			    QXL_GEM_DOMAIN_VRAM, NULL, &bo);
+	if (ret) {
+		DRM_ERROR("failed to allocate VRAM BO\n");
+		return ret;
+	}
+	ret = qxl_bo_reserve(bo, false);
+	if (unlikely(ret != 0))
+		goto out_unref;
+
+	*_bo = bo;
+	return 0;
+out_unref:
+	qxl_bo_unref(&bo);
+	return 0;
+}
+
+static int wait_for_io_cmd_user(struct qxl_device *qdev, uint8_t val, long port)
+{
+	int irq_num;
+	long addr = qdev->io_base + port;
+	int ret;
+
+	mutex_lock(&qdev->async_io_mutex);
+	irq_num = atomic_read(&qdev->irq_received_io_cmd);
+
+
+	if (qdev->last_sent_io_cmd > irq_num) {
+		ret = wait_event_interruptible(qdev->io_cmd_event,
+					       atomic_read(&qdev->irq_received_io_cmd) > irq_num);
+		if (ret)
+			goto out;
+		irq_num = atomic_read(&qdev->irq_received_io_cmd);
+	}
+	outb(val, addr);
+	qdev->last_sent_io_cmd = irq_num + 1;
+	ret = wait_event_interruptible(qdev->io_cmd_event,
+				       atomic_read(&qdev->irq_received_io_cmd) > irq_num);
+out:
+	mutex_unlock(&qdev->async_io_mutex);
+	return ret;
+}
+
+static void wait_for_io_cmd(struct qxl_device *qdev, uint8_t val, long port)
+{
+	int ret;
+
+restart:
+	ret = wait_for_io_cmd_user(qdev, val, port);
+	if (ret == -ERESTARTSYS)
+		goto restart;
+}
+
+int qxl_io_update_area(struct qxl_device *qdev, struct qxl_bo *surf,
+			const struct qxl_rect *area)
+{
+	int surface_id;
+	uint32_t surface_width, surface_height;
+	int ret;
+
+	if (!surf->hw_surf_alloc)
+		DRM_ERROR("got io update area with no hw surface\n");
+
+	if (surf->is_primary)
+		surface_id = 0;
+	else
+		surface_id = surf->surface_id;
+	surface_width = surf->surf.width;
+	surface_height = surf->surf.height;
+
+	if (area->left < 0 || area->top < 0 ||
+	    area->right > surface_width || area->bottom > surface_height) {
+		qxl_io_log(qdev, "%s: not doing area update for "
+			   "%d, (%d,%d,%d,%d) (%d,%d)\n", __func__, surface_id, area->left,
+			   area->top, area->right, area->bottom, surface_width, surface_height);
+		return -EINVAL;
+	}
+	mutex_lock(&qdev->update_area_mutex);
+	qdev->ram_header->update_area = *area;
+	qdev->ram_header->update_surface = surface_id;
+	ret = wait_for_io_cmd_user(qdev, 0, QXL_IO_UPDATE_AREA_ASYNC);
+	mutex_unlock(&qdev->update_area_mutex);
+	return ret;
+}
+
+void qxl_io_notify_oom(struct qxl_device *qdev)
+{
+	outb(0, qdev->io_base + QXL_IO_NOTIFY_OOM);
+}
+
+void qxl_io_flush_release(struct qxl_device *qdev)
+{
+	outb(0, qdev->io_base + QXL_IO_FLUSH_RELEASE);
+}
+
+void qxl_io_flush_surfaces(struct qxl_device *qdev)
+{
+	wait_for_io_cmd(qdev, 0, QXL_IO_FLUSH_SURFACES_ASYNC);
+}
+
+
+void qxl_io_destroy_primary(struct qxl_device *qdev)
+{
+	wait_for_io_cmd(qdev, 0, QXL_IO_DESTROY_PRIMARY_ASYNC);
+}
+
+void qxl_io_create_primary(struct qxl_device *qdev, unsigned width,
+			   unsigned height, unsigned offset, struct qxl_bo *bo)
+{
+	struct qxl_surface_create *create;
+
+	QXL_INFO(qdev, "%s: qdev %p, ram_header %p\n", __func__, qdev,
+		 qdev->ram_header);
+	create = &qdev->ram_header->create_surface;
+	create->format = bo->surf.format;
+	create->width = width;
+	create->height = height;
+	create->stride = bo->surf.stride;
+	create->mem = qxl_bo_physical_address(qdev, bo, offset);
+
+	QXL_INFO(qdev, "%s: mem = %llx, from %p\n", __func__, create->mem,
+		 bo->kptr);
+
+	create->flags = QXL_SURF_FLAG_KEEP_DATA;
+	create->type = QXL_SURF_TYPE_PRIMARY;
+
+	wait_for_io_cmd(qdev, 0, QXL_IO_CREATE_PRIMARY_ASYNC);
+}
+
+void qxl_io_memslot_add(struct qxl_device *qdev, uint8_t id)
+{
+	QXL_INFO(qdev, "qxl_memslot_add %d\n", id);
+	wait_for_io_cmd(qdev, id, QXL_IO_MEMSLOT_ADD_ASYNC);
+}
+
+void qxl_io_log(struct qxl_device *qdev, const char *fmt, ...)
+{
+	va_list args;
+
+	va_start(args, fmt);
+	vsnprintf(qdev->ram_header->log_buf, QXL_LOG_BUF_SIZE, fmt, args);
+	va_end(args);
+	/*
+	 * DO not do a DRM output here - this will call printk, which will
+	 * call back into qxl for rendering (qxl_fb)
+	 */
+	outb(0, qdev->io_base + QXL_IO_LOG);
+}
+
+void qxl_io_reset(struct qxl_device *qdev)
+{
+	outb(0, qdev->io_base + QXL_IO_RESET);
+}
+
+void qxl_io_monitors_config(struct qxl_device *qdev)
+{
+	qxl_io_log(qdev, "%s: %d [%dx%d+%d+%d]\n", __func__,
+		   qdev->monitors_config ?
+		   qdev->monitors_config->count : -1,
+		   qdev->monitors_config && qdev->monitors_config->count ?
+		   qdev->monitors_config->heads[0].width : -1,
+		   qdev->monitors_config && qdev->monitors_config->count ?
+		   qdev->monitors_config->heads[0].height : -1,
+		   qdev->monitors_config && qdev->monitors_config->count ?
+		   qdev->monitors_config->heads[0].x : -1,
+		   qdev->monitors_config && qdev->monitors_config->count ?
+		   qdev->monitors_config->heads[0].y : -1
+		   );
+
+	wait_for_io_cmd(qdev, 0, QXL_IO_MONITORS_CONFIG_ASYNC);
+}
+
+int qxl_surface_id_alloc(struct qxl_device *qdev,
+		      struct qxl_bo *surf)
+{
+	uint32_t handle;
+	int idr_ret;
+	int count = 0;
+again:
+	idr_preload(GFP_ATOMIC);
+	spin_lock(&qdev->surf_id_idr_lock);
+	idr_ret = idr_alloc(&qdev->surf_id_idr, NULL, 1, 0, GFP_NOWAIT);
+	spin_unlock(&qdev->surf_id_idr_lock);
+	idr_preload_end();
+	if (idr_ret < 0)
+		return idr_ret;
+	handle = idr_ret;
+
+	if (handle >= qdev->rom->n_surfaces) {
+		count++;
+		spin_lock(&qdev->surf_id_idr_lock);
+		idr_remove(&qdev->surf_id_idr, handle);
+		spin_unlock(&qdev->surf_id_idr_lock);
+		qxl_reap_surface_id(qdev, 2);
+		goto again;
+	}
+	surf->surface_id = handle;
+
+	spin_lock(&qdev->surf_id_idr_lock);
+	qdev->last_alloced_surf_id = handle;
+	spin_unlock(&qdev->surf_id_idr_lock);
+	return 0;
+}
+
+void qxl_surface_id_dealloc(struct qxl_device *qdev,
+			    uint32_t surface_id)
+{
+	spin_lock(&qdev->surf_id_idr_lock);
+	idr_remove(&qdev->surf_id_idr, surface_id);
+	spin_unlock(&qdev->surf_id_idr_lock);
+}
+
+int qxl_hw_surface_alloc(struct qxl_device *qdev,
+			 struct qxl_bo *surf,
+			 struct ttm_mem_reg *new_mem)
+{
+	struct qxl_surface_cmd *cmd;
+	struct qxl_release *release;
+	int ret;
+
+	if (surf->hw_surf_alloc)
+		return 0;
+
+	ret = qxl_alloc_surface_release_reserved(qdev, QXL_SURFACE_CMD_CREATE,
+						 NULL,
+						 &release);
+	if (ret)
+		return ret;
+
+	cmd = (struct qxl_surface_cmd *)qxl_release_map(qdev, release);
+	cmd->type = QXL_SURFACE_CMD_CREATE;
+	cmd->u.surface_create.format = surf->surf.format;
+	cmd->u.surface_create.width = surf->surf.width;
+	cmd->u.surface_create.height = surf->surf.height;
+	cmd->u.surface_create.stride = surf->surf.stride;
+	if (new_mem) {
+		int slot_id = surf->type == QXL_GEM_DOMAIN_VRAM ? qdev->main_mem_slot : qdev->surfaces_mem_slot;
+		struct qxl_memslot *slot = &(qdev->mem_slots[slot_id]);
+
+		/* TODO - need to hold one of the locks to read tbo.offset */
+		cmd->u.surface_create.data = slot->high_bits;
+
+		cmd->u.surface_create.data |= (new_mem->start << PAGE_SHIFT) + surf->tbo.bdev->man[new_mem->mem_type].gpu_offset;
+	} else
+		cmd->u.surface_create.data = qxl_bo_physical_address(qdev, surf, 0);
+	cmd->surface_id = surf->surface_id;
+	qxl_release_unmap(qdev, release, &cmd->release_info);
+
+	surf->surf_create = release;
+
+	/* no need to add a release to the fence for this bo,
+	   since it is only released when we ask to destroy the surface
+	   and it would never signal otherwise */
+	qxl_fence_releaseable(qdev, release);
+
+	qxl_push_command_ring_release(qdev, release, QXL_CMD_SURFACE, false);
+
+	qxl_release_unreserve(qdev, release);
+
+	surf->hw_surf_alloc = true;
+	spin_lock(&qdev->surf_id_idr_lock);
+	idr_replace(&qdev->surf_id_idr, surf, surf->surface_id);
+	spin_unlock(&qdev->surf_id_idr_lock);
+	return 0;
+}
+
+int qxl_hw_surface_dealloc(struct qxl_device *qdev,
+			   struct qxl_bo *surf)
+{
+	struct qxl_surface_cmd *cmd;
+	struct qxl_release *release;
+	int ret;
+	int id;
+
+	if (!surf->hw_surf_alloc)
+		return 0;
+
+	ret = qxl_alloc_surface_release_reserved(qdev, QXL_SURFACE_CMD_DESTROY,
+						 surf->surf_create,
+						 &release);
+	if (ret)
+		return ret;
+
+	surf->surf_create = NULL;
+	/* remove the surface from the idr, but not the surface id yet */
+	spin_lock(&qdev->surf_id_idr_lock);
+	idr_replace(&qdev->surf_id_idr, NULL, surf->surface_id);
+	spin_unlock(&qdev->surf_id_idr_lock);
+	surf->hw_surf_alloc = false;
+
+	id = surf->surface_id;
+	surf->surface_id = 0;
+
+	release->surface_release_id = id;
+	cmd = (struct qxl_surface_cmd *)qxl_release_map(qdev, release);
+	cmd->type = QXL_SURFACE_CMD_DESTROY;
+	cmd->surface_id = id;
+	qxl_release_unmap(qdev, release, &cmd->release_info);
+
+	qxl_fence_releaseable(qdev, release);
+
+	qxl_push_command_ring_release(qdev, release, QXL_CMD_SURFACE, false);
+
+	qxl_release_unreserve(qdev, release);
+
+
+	return 0;
+}
+
+int qxl_update_surface(struct qxl_device *qdev, struct qxl_bo *surf)
+{
+	struct qxl_rect rect;
+	int ret;
+
+	/* if we are evicting, we need to make sure the surface is up
+	   to date */
+	rect.left = 0;
+	rect.right = surf->surf.width;
+	rect.top = 0;
+	rect.bottom = surf->surf.height;
+retry:
+	ret = qxl_io_update_area(qdev, surf, &rect);
+	if (ret == -ERESTARTSYS)
+		goto retry;
+	return ret;
+}
+
+static void qxl_surface_evict_locked(struct qxl_device *qdev, struct qxl_bo *surf, bool do_update_area)
+{
+	/* no need to update area if we are just freeing the surface normally */
+	if (do_update_area)
+		qxl_update_surface(qdev, surf);
+
+	/* nuke the surface id at the hw */
+	qxl_hw_surface_dealloc(qdev, surf);
+}
+
+void qxl_surface_evict(struct qxl_device *qdev, struct qxl_bo *surf, bool do_update_area)
+{
+	mutex_lock(&qdev->surf_evict_mutex);
+	qxl_surface_evict_locked(qdev, surf, do_update_area);
+	mutex_unlock(&qdev->surf_evict_mutex);
+}
+
+static int qxl_reap_surf(struct qxl_device *qdev, struct qxl_bo *surf, bool stall)
+{
+	int ret;
+
+	ret = qxl_bo_reserve(surf, false);
+	if (ret == -EBUSY)
+		return -EBUSY;
+
+	if (surf->fence.num_active_releases > 0 && stall == false) {
+		qxl_bo_unreserve(surf);
+		return -EBUSY;
+	}
+
+	if (stall)
+		mutex_unlock(&qdev->surf_evict_mutex);
+
+	spin_lock(&surf->tbo.bdev->fence_lock);
+	ret = ttm_bo_wait(&surf->tbo, true, true, !stall);
+	spin_unlock(&surf->tbo.bdev->fence_lock);
+
+	if (stall)
+		mutex_lock(&qdev->surf_evict_mutex);
+	if (ret == -EBUSY) {
+		qxl_bo_unreserve(surf);
+		return -EBUSY;
+	}
+
+	qxl_surface_evict_locked(qdev, surf, true);
+	qxl_bo_unreserve(surf);
+	return 0;
+}
+
+static int qxl_reap_surface_id(struct qxl_device *qdev, int max_to_reap)
+{
+	int num_reaped = 0;
+	int i, ret;
+	bool stall = false;
+	int start = 0;
+
+	mutex_lock(&qdev->surf_evict_mutex);
+again:
+
+	spin_lock(&qdev->surf_id_idr_lock);
+	start = qdev->last_alloced_surf_id + 1;
+	spin_unlock(&qdev->surf_id_idr_lock);
+
+	for (i = start; i < start + qdev->rom->n_surfaces; i++) {
+		void *objptr;
+		int surfid = i % qdev->rom->n_surfaces;
+
+		/* this avoids the case where the objects is in the
+		   idr but has been evicted half way - its makes
+		   the idr lookup atomic with the eviction */
+		spin_lock(&qdev->surf_id_idr_lock);
+		objptr = idr_find(&qdev->surf_id_idr, surfid);
+		spin_unlock(&qdev->surf_id_idr_lock);
+
+		if (!objptr)
+			continue;
+
+		ret = qxl_reap_surf(qdev, objptr, stall);
+		if (ret == 0)
+			num_reaped++;
+		if (num_reaped >= max_to_reap)
+			break;
+	}
+	if (num_reaped == 0 && stall == false) {
+		stall = true;
+		goto again;
+	}
+
+	mutex_unlock(&qdev->surf_evict_mutex);
+	if (num_reaped) {
+		usleep_range(500, 1000);
+		qxl_queue_garbage_collect(qdev, true);
+	}
+
+	return 0;
+}
diff --git a/drivers/gpu/drm/qxl/qxl_debugfs.c b/drivers/gpu/drm/qxl/qxl_debugfs.c
new file mode 100644
index 000000000000..c3c2bbdc6674
--- /dev/null
+++ b/drivers/gpu/drm/qxl/qxl_debugfs.c
@@ -0,0 +1,141 @@
+/*
+ * Copyright (C) 2009 Red Hat <bskeggs@redhat.com>
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining
+ * a copy of this software and associated documentation files (the
+ * "Software"), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sublicense, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the
+ * next paragraph) shall be included in all copies or substantial
+ * portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
+ * IN NO EVENT SHALL THE COPYRIGHT OWNER(S) AND/OR ITS SUPPLIERS BE
+ * LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION
+ * OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION
+ * WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+/*
+ * Authors:
+ *  Alon Levy <alevy@redhat.com>
+ */
+
+#include <linux/debugfs.h>
+
+#include "drmP.h"
+#include "qxl_drv.h"
+#include "qxl_object.h"
+
+
+#if defined(CONFIG_DEBUG_FS)
+static int
+qxl_debugfs_irq_received(struct seq_file *m, void *data)
+{
+	struct drm_info_node *node = (struct drm_info_node *) m->private;
+	struct qxl_device *qdev = node->minor->dev->dev_private;
+
+	seq_printf(m, "%d\n", atomic_read(&qdev->irq_received));
+	seq_printf(m, "%d\n", atomic_read(&qdev->irq_received_display));
+	seq_printf(m, "%d\n", atomic_read(&qdev->irq_received_cursor));
+	seq_printf(m, "%d\n", atomic_read(&qdev->irq_received_io_cmd));
+	seq_printf(m, "%d\n", qdev->irq_received_error);
+	return 0;
+}
+
+static int
+qxl_debugfs_buffers_info(struct seq_file *m, void *data)
+{
+	struct drm_info_node *node = (struct drm_info_node *) m->private;
+	struct qxl_device *qdev = node->minor->dev->dev_private;
+	struct qxl_bo *bo;
+
+	list_for_each_entry(bo, &qdev->gem.objects, list) {
+		seq_printf(m, "size %ld, pc %d, sync obj %p, num releases %d\n",
+			   (unsigned long)bo->gem_base.size, bo->pin_count,
+			   bo->tbo.sync_obj, bo->fence.num_active_releases);
+	}
+	return 0;
+}
+
+static struct drm_info_list qxl_debugfs_list[] = {
+	{ "irq_received", qxl_debugfs_irq_received, 0, NULL },
+	{ "qxl_buffers", qxl_debugfs_buffers_info, 0, NULL },
+};
+#define QXL_DEBUGFS_ENTRIES ARRAY_SIZE(qxl_debugfs_list)
+#endif
+
+int
+qxl_debugfs_init(struct drm_minor *minor)
+{
+#if defined(CONFIG_DEBUG_FS)
+	drm_debugfs_create_files(qxl_debugfs_list, QXL_DEBUGFS_ENTRIES,
+				 minor->debugfs_root, minor);
+#endif
+	return 0;
+}
+
+void
+qxl_debugfs_takedown(struct drm_minor *minor)
+{
+#if defined(CONFIG_DEBUG_FS)
+	drm_debugfs_remove_files(qxl_debugfs_list, QXL_DEBUGFS_ENTRIES,
+				 minor);
+#endif
+}
+
+int qxl_debugfs_add_files(struct qxl_device *qdev,
+			  struct drm_info_list *files,
+			  unsigned nfiles)
+{
+	unsigned i;
+
+	for (i = 0; i < qdev->debugfs_count; i++) {
+		if (qdev->debugfs[i].files == files) {
+			/* Already registered */
+			return 0;
+		}
+	}
+
+	i = qdev->debugfs_count + 1;
+	if (i > QXL_DEBUGFS_MAX_COMPONENTS) {
+		DRM_ERROR("Reached maximum number of debugfs components.\n");
+		DRM_ERROR("Report so we increase QXL_DEBUGFS_MAX_COMPONENTS.\n");
+		return -EINVAL;
+	}
+	qdev->debugfs[qdev->debugfs_count].files = files;
+	qdev->debugfs[qdev->debugfs_count].num_files = nfiles;
+	qdev->debugfs_count = i;
+#if defined(CONFIG_DEBUG_FS)
+	drm_debugfs_create_files(files, nfiles,
+				 qdev->ddev->control->debugfs_root,
+				 qdev->ddev->control);
+	drm_debugfs_create_files(files, nfiles,
+				 qdev->ddev->primary->debugfs_root,
+				 qdev->ddev->primary);
+#endif
+	return 0;
+}
+
+void qxl_debugfs_remove_files(struct qxl_device *qdev)
+{
+#if defined(CONFIG_DEBUG_FS)
+	unsigned i;
+
+	for (i = 0; i < qdev->debugfs_count; i++) {
+		drm_debugfs_remove_files(qdev->debugfs[i].files,
+					 qdev->debugfs[i].num_files,
+					 qdev->ddev->control);
+		drm_debugfs_remove_files(qdev->debugfs[i].files,
+					 qdev->debugfs[i].num_files,
+					 qdev->ddev->primary);
+	}
+#endif
+}
diff --git a/drivers/gpu/drm/qxl/qxl_dev.h b/drivers/gpu/drm/qxl/qxl_dev.h
new file mode 100644
index 000000000000..94c5aec71920
--- /dev/null
+++ b/drivers/gpu/drm/qxl/qxl_dev.h
@@ -0,0 +1,879 @@
+/*
+   Copyright (C) 2009 Red Hat, Inc.
+
+   Redistribution and use in source and binary forms, with or without
+   modification, are permitted provided that the following conditions are
+   met:
+
+       * Redistributions of source code must retain the above copyright
+	 notice, this list of conditions and the following disclaimer.
+       * Redistributions in binary form must reproduce the above copyright
+	 notice, this list of conditions and the following disclaimer in
+	 the documentation and/or other materials provided with the
+	 distribution.
+       * Neither the name of the copyright holder nor the names of its
+	 contributors may be used to endorse or promote products derived
+	 from this software without specific prior written permission.
+
+   THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDER AND CONTRIBUTORS "AS
+   IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED
+   TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A
+   PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
+   HOLDER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+   SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
+   LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+   DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+   THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+   (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
+   OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+*/
+
+
+#ifndef H_QXL_DEV
+#define H_QXL_DEV
+
+#include <linux/types.h>
+
+/*
+ * from spice-protocol
+ * Release 0.10.0
+ */
+
+/* enums.h */
+
+enum SpiceImageType {
+	SPICE_IMAGE_TYPE_BITMAP,
+	SPICE_IMAGE_TYPE_QUIC,
+	SPICE_IMAGE_TYPE_RESERVED,
+	SPICE_IMAGE_TYPE_LZ_PLT = 100,
+	SPICE_IMAGE_TYPE_LZ_RGB,
+	SPICE_IMAGE_TYPE_GLZ_RGB,
+	SPICE_IMAGE_TYPE_FROM_CACHE,
+	SPICE_IMAGE_TYPE_SURFACE,
+	SPICE_IMAGE_TYPE_JPEG,
+	SPICE_IMAGE_TYPE_FROM_CACHE_LOSSLESS,
+	SPICE_IMAGE_TYPE_ZLIB_GLZ_RGB,
+	SPICE_IMAGE_TYPE_JPEG_ALPHA,
+
+	SPICE_IMAGE_TYPE_ENUM_END
+};
+
+enum SpiceBitmapFmt {
+	SPICE_BITMAP_FMT_INVALID,
+	SPICE_BITMAP_FMT_1BIT_LE,
+	SPICE_BITMAP_FMT_1BIT_BE,
+	SPICE_BITMAP_FMT_4BIT_LE,
+	SPICE_BITMAP_FMT_4BIT_BE,
+	SPICE_BITMAP_FMT_8BIT,
+	SPICE_BITMAP_FMT_16BIT,
+	SPICE_BITMAP_FMT_24BIT,
+	SPICE_BITMAP_FMT_32BIT,
+	SPICE_BITMAP_FMT_RGBA,
+
+	SPICE_BITMAP_FMT_ENUM_END
+};
+
+enum SpiceSurfaceFmt {
+	SPICE_SURFACE_FMT_INVALID,
+	SPICE_SURFACE_FMT_1_A,
+	SPICE_SURFACE_FMT_8_A = 8,
+	SPICE_SURFACE_FMT_16_555 = 16,
+	SPICE_SURFACE_FMT_32_xRGB = 32,
+	SPICE_SURFACE_FMT_16_565 = 80,
+	SPICE_SURFACE_FMT_32_ARGB = 96,
+
+	SPICE_SURFACE_FMT_ENUM_END
+};
+
+enum SpiceClipType {
+	SPICE_CLIP_TYPE_NONE,
+	SPICE_CLIP_TYPE_RECTS,
+
+	SPICE_CLIP_TYPE_ENUM_END
+};
+
+enum SpiceRopd {
+	SPICE_ROPD_INVERS_SRC = (1 << 0),
+	SPICE_ROPD_INVERS_BRUSH = (1 << 1),
+	SPICE_ROPD_INVERS_DEST = (1 << 2),
+	SPICE_ROPD_OP_PUT = (1 << 3),
+	SPICE_ROPD_OP_OR = (1 << 4),
+	SPICE_ROPD_OP_AND = (1 << 5),
+	SPICE_ROPD_OP_XOR = (1 << 6),
+	SPICE_ROPD_OP_BLACKNESS = (1 << 7),
+	SPICE_ROPD_OP_WHITENESS = (1 << 8),
+	SPICE_ROPD_OP_INVERS = (1 << 9),
+	SPICE_ROPD_INVERS_RES = (1 << 10),
+
+	SPICE_ROPD_MASK = 0x7ff
+};
+
+enum SpiceBrushType {
+	SPICE_BRUSH_TYPE_NONE,
+	SPICE_BRUSH_TYPE_SOLID,
+	SPICE_BRUSH_TYPE_PATTERN,
+
+	SPICE_BRUSH_TYPE_ENUM_END
+};
+
+enum SpiceCursorType {
+	SPICE_CURSOR_TYPE_ALPHA,
+	SPICE_CURSOR_TYPE_MONO,
+	SPICE_CURSOR_TYPE_COLOR4,
+	SPICE_CURSOR_TYPE_COLOR8,
+	SPICE_CURSOR_TYPE_COLOR16,
+	SPICE_CURSOR_TYPE_COLOR24,
+	SPICE_CURSOR_TYPE_COLOR32,
+
+	SPICE_CURSOR_TYPE_ENUM_END
+};
+
+/* qxl_dev.h */
+
+#pragma pack(push, 1)
+
+#define REDHAT_PCI_VENDOR_ID 0x1b36
+
+/* 0x100-0x11f reserved for spice, 0x1ff used for unstable work */
+#define QXL_DEVICE_ID_STABLE 0x0100
+
+enum {
+	QXL_REVISION_STABLE_V04 = 0x01,
+	QXL_REVISION_STABLE_V06 = 0x02,
+	QXL_REVISION_STABLE_V10 = 0x03,
+	QXL_REVISION_STABLE_V12 = 0x04,
+};
+
+#define QXL_DEVICE_ID_DEVEL 0x01ff
+#define QXL_REVISION_DEVEL 0x01
+
+#define QXL_ROM_MAGIC (*(uint32_t *)"QXRO")
+#define QXL_RAM_MAGIC (*(uint32_t *)"QXRA")
+
+enum {
+	QXL_RAM_RANGE_INDEX,
+	QXL_VRAM_RANGE_INDEX,
+	QXL_ROM_RANGE_INDEX,
+	QXL_IO_RANGE_INDEX,
+
+	QXL_PCI_RANGES
+};
+
+/* qxl-1 compat: append only */
+enum {
+	QXL_IO_NOTIFY_CMD,
+	QXL_IO_NOTIFY_CURSOR,
+	QXL_IO_UPDATE_AREA,
+	QXL_IO_UPDATE_IRQ,
+	QXL_IO_NOTIFY_OOM,
+	QXL_IO_RESET,
+	QXL_IO_SET_MODE,                  /* qxl-1 */
+	QXL_IO_LOG,
+	/* appended for qxl-2 */
+	QXL_IO_MEMSLOT_ADD,
+	QXL_IO_MEMSLOT_DEL,
+	QXL_IO_DETACH_PRIMARY,
+	QXL_IO_ATTACH_PRIMARY,
+	QXL_IO_CREATE_PRIMARY,
+	QXL_IO_DESTROY_PRIMARY,
+	QXL_IO_DESTROY_SURFACE_WAIT,
+	QXL_IO_DESTROY_ALL_SURFACES,
+	/* appended for qxl-3 */
+	QXL_IO_UPDATE_AREA_ASYNC,
+	QXL_IO_MEMSLOT_ADD_ASYNC,
+	QXL_IO_CREATE_PRIMARY_ASYNC,
+	QXL_IO_DESTROY_PRIMARY_ASYNC,
+	QXL_IO_DESTROY_SURFACE_ASYNC,
+	QXL_IO_DESTROY_ALL_SURFACES_ASYNC,
+	QXL_IO_FLUSH_SURFACES_ASYNC,
+	QXL_IO_FLUSH_RELEASE,
+	/* appended for qxl-4 */
+	QXL_IO_MONITORS_CONFIG_ASYNC,
+
+	QXL_IO_RANGE_SIZE
+};
+
+typedef uint64_t QXLPHYSICAL;
+typedef int32_t QXLFIXED; /* fixed 28.4 */
+
+struct qxl_point_fix {
+	QXLFIXED x;
+	QXLFIXED y;
+};
+
+struct qxl_point {
+	int32_t x;
+	int32_t y;
+};
+
+struct qxl_point_1_6 {
+	int16_t x;
+	int16_t y;
+};
+
+struct qxl_rect {
+	int32_t top;
+	int32_t left;
+	int32_t bottom;
+	int32_t right;
+};
+
+struct qxl_urect {
+	uint32_t top;
+	uint32_t left;
+	uint32_t bottom;
+	uint32_t right;
+};
+
+/* qxl-1 compat: append only */
+struct qxl_rom {
+	uint32_t magic;
+	uint32_t id;
+	uint32_t update_id;
+	uint32_t compression_level;
+	uint32_t log_level;
+	uint32_t mode;			  /* qxl-1 */
+	uint32_t modes_offset;
+	uint32_t num_io_pages;
+	uint32_t pages_offset;		  /* qxl-1 */
+	uint32_t draw_area_offset;	  /* qxl-1 */
+	uint32_t surface0_area_size;	  /* qxl-1 name: draw_area_size */
+	uint32_t ram_header_offset;
+	uint32_t mm_clock;
+	/* appended for qxl-2 */
+	uint32_t n_surfaces;
+	uint64_t flags;
+	uint8_t slots_start;
+	uint8_t slots_end;
+	uint8_t slot_gen_bits;
+	uint8_t slot_id_bits;
+	uint8_t slot_generation;
+	/* appended for qxl-4 */
+	uint8_t client_present;
+	uint8_t client_capabilities[58];
+	uint32_t client_monitors_config_crc;
+	struct {
+		uint16_t count;
+	uint16_t padding;
+		struct qxl_urect heads[64];
+	} client_monitors_config;
+};
+
+/* qxl-1 compat: fixed */
+struct qxl_mode {
+	uint32_t id;
+	uint32_t x_res;
+	uint32_t y_res;
+	uint32_t bits;
+	uint32_t stride;
+	uint32_t x_mili;
+	uint32_t y_mili;
+	uint32_t orientation;
+};
+
+/* qxl-1 compat: fixed */
+struct qxl_modes {
+	uint32_t n_modes;
+	struct qxl_mode modes[0];
+};
+
+/* qxl-1 compat: append only */
+enum qxl_cmd_type {
+	QXL_CMD_NOP,
+	QXL_CMD_DRAW,
+	QXL_CMD_UPDATE,
+	QXL_CMD_CURSOR,
+	QXL_CMD_MESSAGE,
+	QXL_CMD_SURFACE,
+};
+
+/* qxl-1 compat: fixed */
+struct qxl_command {
+	QXLPHYSICAL data;
+	uint32_t type;
+	uint32_t padding;
+};
+
+#define QXL_COMMAND_FLAG_COMPAT		(1<<0)
+#define QXL_COMMAND_FLAG_COMPAT_16BPP	(2<<0)
+
+struct qxl_command_ext {
+	struct qxl_command cmd;
+	uint32_t group_id;
+	uint32_t flags;
+};
+
+struct qxl_mem_slot {
+	uint64_t mem_start;
+	uint64_t mem_end;
+};
+
+#define QXL_SURF_TYPE_PRIMARY	   0
+
+#define QXL_SURF_FLAG_KEEP_DATA	   (1 << 0)
+
+struct qxl_surface_create {
+	uint32_t width;
+	uint32_t height;
+	int32_t stride;
+	uint32_t format;
+	uint32_t position;
+	uint32_t mouse_mode;
+	uint32_t flags;
+	uint32_t type;
+	QXLPHYSICAL mem;
+};
+
+#define QXL_COMMAND_RING_SIZE 32
+#define QXL_CURSOR_RING_SIZE 32
+#define QXL_RELEASE_RING_SIZE 8
+
+#define QXL_LOG_BUF_SIZE 4096
+
+#define QXL_INTERRUPT_DISPLAY (1 << 0)
+#define QXL_INTERRUPT_CURSOR (1 << 1)
+#define QXL_INTERRUPT_IO_CMD (1 << 2)
+#define QXL_INTERRUPT_ERROR  (1 << 3)
+#define QXL_INTERRUPT_CLIENT (1 << 4)
+#define QXL_INTERRUPT_CLIENT_MONITORS_CONFIG  (1 << 5)
+
+struct qxl_ring_header {
+	uint32_t num_items;
+	uint32_t prod;
+	uint32_t notify_on_prod;
+	uint32_t cons;
+	uint32_t notify_on_cons;
+};
+
+/* qxl-1 compat: append only */
+struct qxl_ram_header {
+	uint32_t magic;
+	uint32_t int_pending;
+	uint32_t int_mask;
+	uint8_t log_buf[QXL_LOG_BUF_SIZE];
+	struct qxl_ring_header  cmd_ring_hdr;
+	struct qxl_command	cmd_ring[QXL_COMMAND_RING_SIZE];
+	struct qxl_ring_header  cursor_ring_hdr;
+	struct qxl_command	cursor_ring[QXL_CURSOR_RING_SIZE];
+	struct qxl_ring_header  release_ring_hdr;
+	uint64_t		release_ring[QXL_RELEASE_RING_SIZE];
+	struct qxl_rect update_area;
+	/* appended for qxl-2 */
+	uint32_t update_surface;
+	struct qxl_mem_slot mem_slot;
+	struct qxl_surface_create create_surface;
+	uint64_t flags;
+
+	/* appended for qxl-4 */
+
+	/* used by QXL_IO_MONITORS_CONFIG_ASYNC */
+	QXLPHYSICAL monitors_config;
+	uint8_t guest_capabilities[64];
+};
+
+union qxl_release_info {
+	uint64_t id;	  /* in  */
+	uint64_t next;	  /* out */
+};
+
+struct qxl_release_info_ext {
+	union qxl_release_info *info;
+	uint32_t group_id;
+};
+
+struct qxl_data_chunk {
+	uint32_t data_size;
+	QXLPHYSICAL prev_chunk;
+	QXLPHYSICAL next_chunk;
+	uint8_t data[0];
+};
+
+struct qxl_message {
+	union qxl_release_info release_info;
+	uint8_t data[0];
+};
+
+struct qxl_compat_update_cmd {
+	union qxl_release_info release_info;
+	struct qxl_rect area;
+	uint32_t update_id;
+};
+
+struct qxl_update_cmd {
+	union qxl_release_info release_info;
+	struct qxl_rect area;
+	uint32_t update_id;
+	uint32_t surface_id;
+};
+
+struct qxl_cursor_header {
+	uint64_t unique;
+	uint16_t type;
+	uint16_t width;
+	uint16_t height;
+	uint16_t hot_spot_x;
+	uint16_t hot_spot_y;
+};
+
+struct qxl_cursor {
+	struct qxl_cursor_header header;
+	uint32_t data_size;
+	struct qxl_data_chunk chunk;
+};
+
+enum {
+	QXL_CURSOR_SET,
+	QXL_CURSOR_MOVE,
+	QXL_CURSOR_HIDE,
+	QXL_CURSOR_TRAIL,
+};
+
+#define QXL_CURSOR_DEVICE_DATA_SIZE 128
+
+struct qxl_cursor_cmd {
+	union qxl_release_info release_info;
+	uint8_t type;
+	union {
+		struct {
+			struct qxl_point_1_6 position;
+			uint8_t visible;
+			QXLPHYSICAL shape;
+		} set;
+		struct {
+			uint16_t length;
+			uint16_t frequency;
+		} trail;
+		struct qxl_point_1_6 position;
+	} u;
+	/* todo: dynamic size from rom */
+	uint8_t device_data[QXL_CURSOR_DEVICE_DATA_SIZE];
+};
+
+enum {
+	QXL_DRAW_NOP,
+	QXL_DRAW_FILL,
+	QXL_DRAW_OPAQUE,
+	QXL_DRAW_COPY,
+	QXL_COPY_BITS,
+	QXL_DRAW_BLEND,
+	QXL_DRAW_BLACKNESS,
+	QXL_DRAW_WHITENESS,
+	QXL_DRAW_INVERS,
+	QXL_DRAW_ROP3,
+	QXL_DRAW_STROKE,
+	QXL_DRAW_TEXT,
+	QXL_DRAW_TRANSPARENT,
+	QXL_DRAW_ALPHA_BLEND,
+	QXL_DRAW_COMPOSITE
+};
+
+struct qxl_raster_glyph {
+	struct qxl_point render_pos;
+	struct qxl_point glyph_origin;
+	uint16_t width;
+	uint16_t height;
+	uint8_t data[0];
+};
+
+struct qxl_string {
+	uint32_t data_size;
+	uint16_t length;
+	uint16_t flags;
+	struct qxl_data_chunk chunk;
+};
+
+struct qxl_copy_bits {
+	struct qxl_point src_pos;
+};
+
+enum qxl_effect_type {
+	QXL_EFFECT_BLEND = 0,
+	QXL_EFFECT_OPAQUE = 1,
+	QXL_EFFECT_REVERT_ON_DUP = 2,
+	QXL_EFFECT_BLACKNESS_ON_DUP = 3,
+	QXL_EFFECT_WHITENESS_ON_DUP = 4,
+	QXL_EFFECT_NOP_ON_DUP = 5,
+	QXL_EFFECT_NOP = 6,
+	QXL_EFFECT_OPAQUE_BRUSH = 7
+};
+
+struct qxl_pattern {
+	QXLPHYSICAL pat;
+	struct qxl_point pos;
+};
+
+struct qxl_brush {
+	uint32_t type;
+	union {
+		uint32_t color;
+		struct qxl_pattern pattern;
+	} u;
+};
+
+struct qxl_q_mask {
+	uint8_t flags;
+	struct qxl_point pos;
+	QXLPHYSICAL bitmap;
+};
+
+struct qxl_fill {
+	struct qxl_brush brush;
+	uint16_t rop_descriptor;
+	struct qxl_q_mask mask;
+};
+
+struct qxl_opaque {
+	QXLPHYSICAL src_bitmap;
+	struct qxl_rect src_area;
+	struct qxl_brush brush;
+	uint16_t rop_descriptor;
+	uint8_t scale_mode;
+	struct qxl_q_mask mask;
+};
+
+struct qxl_copy {
+	QXLPHYSICAL src_bitmap;
+	struct qxl_rect src_area;
+	uint16_t rop_descriptor;
+	uint8_t scale_mode;
+	struct qxl_q_mask mask;
+};
+
+struct qxl_transparent {
+	QXLPHYSICAL src_bitmap;
+	struct qxl_rect src_area;
+	uint32_t src_color;
+	uint32_t true_color;
+};
+
+struct qxl_alpha_blend {
+	uint16_t alpha_flags;
+	uint8_t alpha;
+	QXLPHYSICAL src_bitmap;
+	struct qxl_rect src_area;
+};
+
+struct qxl_compat_alpha_blend {
+	uint8_t alpha;
+	QXLPHYSICAL src_bitmap;
+	struct qxl_rect src_area;
+};
+
+struct qxl_rop_3 {
+	QXLPHYSICAL src_bitmap;
+	struct qxl_rect src_area;
+	struct qxl_brush brush;
+	uint8_t rop3;
+	uint8_t scale_mode;
+	struct qxl_q_mask mask;
+};
+
+struct qxl_line_attr {
+	uint8_t flags;
+	uint8_t join_style;
+	uint8_t end_style;
+	uint8_t style_nseg;
+	QXLFIXED width;
+	QXLFIXED miter_limit;
+	QXLPHYSICAL style;
+};
+
+struct qxl_stroke {
+	QXLPHYSICAL path;
+	struct qxl_line_attr attr;
+	struct qxl_brush brush;
+	uint16_t fore_mode;
+	uint16_t back_mode;
+};
+
+struct qxl_text {
+	QXLPHYSICAL str;
+	struct qxl_rect back_area;
+	struct qxl_brush fore_brush;
+	struct qxl_brush back_brush;
+	uint16_t fore_mode;
+	uint16_t back_mode;
+};
+
+struct qxl_mask {
+	struct qxl_q_mask mask;
+};
+
+struct qxl_clip {
+	uint32_t type;
+	QXLPHYSICAL data;
+};
+
+enum qxl_operator {
+	QXL_OP_CLEAR			 = 0x00,
+	QXL_OP_SOURCE			 = 0x01,
+	QXL_OP_DST			 = 0x02,
+	QXL_OP_OVER			 = 0x03,
+	QXL_OP_OVER_REVERSE		 = 0x04,
+	QXL_OP_IN			 = 0x05,
+	QXL_OP_IN_REVERSE		 = 0x06,
+	QXL_OP_OUT			 = 0x07,
+	QXL_OP_OUT_REVERSE		 = 0x08,
+	QXL_OP_ATOP			 = 0x09,
+	QXL_OP_ATOP_REVERSE		 = 0x0a,
+	QXL_OP_XOR			 = 0x0b,
+	QXL_OP_ADD			 = 0x0c,
+	QXL_OP_SATURATE			 = 0x0d,
+	/* Note the jump here from 0x0d to 0x30 */
+	QXL_OP_MULTIPLY			 = 0x30,
+	QXL_OP_SCREEN			 = 0x31,
+	QXL_OP_OVERLAY			 = 0x32,
+	QXL_OP_DARKEN			 = 0x33,
+	QXL_OP_LIGHTEN			 = 0x34,
+	QXL_OP_COLOR_DODGE		 = 0x35,
+	QXL_OP_COLOR_BURN		 = 0x36,
+	QXL_OP_HARD_LIGHT		 = 0x37,
+	QXL_OP_SOFT_LIGHT		 = 0x38,
+	QXL_OP_DIFFERENCE		 = 0x39,
+	QXL_OP_EXCLUSION		 = 0x3a,
+	QXL_OP_HSL_HUE			 = 0x3b,
+	QXL_OP_HSL_SATURATION		 = 0x3c,
+	QXL_OP_HSL_COLOR		 = 0x3d,
+	QXL_OP_HSL_LUMINOSITY		 = 0x3e
+};
+
+struct qxl_transform {
+	uint32_t	t00;
+	uint32_t	t01;
+	uint32_t	t02;
+	uint32_t	t10;
+	uint32_t	t11;
+	uint32_t	t12;
+};
+
+/* The flags field has the following bit fields:
+ *
+ *     operator:		[  0 -  7 ]
+ *     src_filter:		[  8 - 10 ]
+ *     mask_filter:		[ 11 - 13 ]
+ *     src_repeat:		[ 14 - 15 ]
+ *     mask_repeat:		[ 16 - 17 ]
+ *     component_alpha:		[ 18 - 18 ]
+ *     reserved:		[ 19 - 31 ]
+ *
+ * The repeat and filter values are those of pixman:
+ *		REPEAT_NONE =		0
+ *              REPEAT_NORMAL =		1
+ *		REPEAT_PAD =		2
+ *		REPEAT_REFLECT =	3
+ *
+ * The filter values are:
+ *		FILTER_NEAREST =	0
+ *		FILTER_BILINEAR	=	1
+ */
+struct qxl_composite {
+	uint32_t		flags;
+
+	QXLPHYSICAL			src;
+	QXLPHYSICAL			src_transform;	/* May be NULL */
+	QXLPHYSICAL			mask;		/* May be NULL */
+	QXLPHYSICAL			mask_transform;	/* May be NULL */
+	struct qxl_point_1_6	src_origin;
+	struct qxl_point_1_6	mask_origin;
+};
+
+struct qxl_compat_drawable {
+	union qxl_release_info release_info;
+	uint8_t effect;
+	uint8_t type;
+	uint16_t bitmap_offset;
+	struct qxl_rect bitmap_area;
+	struct qxl_rect bbox;
+	struct qxl_clip clip;
+	uint32_t mm_time;
+	union {
+		struct qxl_fill fill;
+		struct qxl_opaque opaque;
+		struct qxl_copy copy;
+		struct qxl_transparent transparent;
+		struct qxl_compat_alpha_blend alpha_blend;
+		struct qxl_copy_bits copy_bits;
+		struct qxl_copy blend;
+		struct qxl_rop_3 rop3;
+		struct qxl_stroke stroke;
+		struct qxl_text text;
+		struct qxl_mask blackness;
+		struct qxl_mask invers;
+		struct qxl_mask whiteness;
+	} u;
+};
+
+struct qxl_drawable {
+	union qxl_release_info release_info;
+	uint32_t surface_id;
+	uint8_t effect;
+	uint8_t type;
+	uint8_t self_bitmap;
+	struct qxl_rect self_bitmap_area;
+	struct qxl_rect bbox;
+	struct qxl_clip clip;
+	uint32_t mm_time;
+	int32_t surfaces_dest[3];
+	struct qxl_rect surfaces_rects[3];
+	union {
+		struct qxl_fill fill;
+		struct qxl_opaque opaque;
+		struct qxl_copy copy;
+		struct qxl_transparent transparent;
+		struct qxl_alpha_blend alpha_blend;
+		struct qxl_copy_bits copy_bits;
+		struct qxl_copy blend;
+		struct qxl_rop_3 rop3;
+		struct qxl_stroke stroke;
+		struct qxl_text text;
+		struct qxl_mask blackness;
+		struct qxl_mask invers;
+		struct qxl_mask whiteness;
+		struct qxl_composite composite;
+	} u;
+};
+
+enum qxl_surface_cmd_type {
+	QXL_SURFACE_CMD_CREATE,
+	QXL_SURFACE_CMD_DESTROY,
+};
+
+struct qxl_surface {
+	uint32_t format;
+	uint32_t width;
+	uint32_t height;
+	int32_t stride;
+	QXLPHYSICAL data;
+};
+
+struct qxl_surface_cmd {
+	union qxl_release_info release_info;
+	uint32_t surface_id;
+	uint8_t type;
+	uint32_t flags;
+	union {
+		struct qxl_surface surface_create;
+	} u;
+};
+
+struct qxl_clip_rects {
+	uint32_t num_rects;
+	struct qxl_data_chunk chunk;
+};
+
+enum {
+	QXL_PATH_BEGIN = (1 << 0),
+	QXL_PATH_END = (1 << 1),
+	QXL_PATH_CLOSE = (1 << 3),
+	QXL_PATH_BEZIER = (1 << 4),
+};
+
+struct qxl_path_seg {
+	uint32_t flags;
+	uint32_t count;
+	struct qxl_point_fix points[0];
+};
+
+struct qxl_path {
+	uint32_t data_size;
+	struct qxl_data_chunk chunk;
+};
+
+enum {
+	QXL_IMAGE_GROUP_DRIVER,
+	QXL_IMAGE_GROUP_DEVICE,
+	QXL_IMAGE_GROUP_RED,
+	QXL_IMAGE_GROUP_DRIVER_DONT_CACHE,
+};
+
+struct qxl_image_id {
+	uint32_t group;
+	uint32_t unique;
+};
+
+union qxl_image_id_union {
+	struct qxl_image_id id;
+	uint64_t value;
+};
+
+enum qxl_image_flags {
+	QXL_IMAGE_CACHE = (1 << 0),
+	QXL_IMAGE_HIGH_BITS_SET = (1 << 1),
+};
+
+enum qxl_bitmap_flags {
+	QXL_BITMAP_DIRECT = (1 << 0),
+	QXL_BITMAP_UNSTABLE = (1 << 1),
+	QXL_BITMAP_TOP_DOWN = (1 << 2), /* == SPICE_BITMAP_FLAGS_TOP_DOWN */
+};
+
+#define QXL_SET_IMAGE_ID(image, _group, _unique) {              \
+	(image)->descriptor.id = (((uint64_t)_unique) << 32) | _group;	\
+}
+
+struct qxl_image_descriptor {
+	uint64_t id;
+	uint8_t type;
+	uint8_t flags;
+	uint32_t width;
+	uint32_t height;
+};
+
+struct qxl_palette {
+	uint64_t unique;
+	uint16_t num_ents;
+	uint32_t ents[0];
+};
+
+struct qxl_bitmap {
+	uint8_t format;
+	uint8_t flags;
+	uint32_t x;
+	uint32_t y;
+	uint32_t stride;
+	QXLPHYSICAL palette;
+	QXLPHYSICAL data; /* data[0] ? */
+};
+
+struct qxl_surface_id {
+	uint32_t surface_id;
+};
+
+struct qxl_encoder_data {
+	uint32_t data_size;
+	uint8_t data[0];
+};
+
+struct qxl_image {
+	struct qxl_image_descriptor descriptor;
+	union { /* variable length */
+		struct qxl_bitmap bitmap;
+		struct qxl_encoder_data quic;
+		struct qxl_surface_id surface_image;
+	} u;
+};
+
+/* A QXLHead is a single monitor output backed by a QXLSurface.
+ * x and y offsets are unsigned since they are used in relation to
+ * the given surface, not the same as the x, y coordinates in the guest
+ * screen reference frame. */
+struct qxl_head {
+	uint32_t id;
+	uint32_t surface_id;
+	uint32_t width;
+	uint32_t height;
+	uint32_t x;
+	uint32_t y;
+	uint32_t flags;
+};
+
+struct qxl_monitors_config {
+	uint16_t count;
+	uint16_t max_allowed; /* If it is 0 no fixed limit is given by the
+				 driver */
+	struct qxl_head heads[0];
+};
+
+#pragma pack(pop)
+
+#endif /* _H_QXL_DEV */
diff --git a/drivers/gpu/drm/qxl/qxl_display.c b/drivers/gpu/drm/qxl/qxl_display.c
new file mode 100644
index 000000000000..fcfd4436ceed
--- /dev/null
+++ b/drivers/gpu/drm/qxl/qxl_display.c
@@ -0,0 +1,982 @@
+/*
+ * Copyright 2013 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Dave Airlie
+ *          Alon Levy
+ */
+
+
+#include "linux/crc32.h"
+
+#include "qxl_drv.h"
+#include "qxl_object.h"
+#include "drm_crtc_helper.h"
+
+static void qxl_crtc_set_to_mode(struct qxl_device *qdev,
+				 struct drm_connector *connector,
+				 struct qxl_head *head)
+{
+	struct drm_device *dev = connector->dev;
+	struct drm_display_mode *mode, *t;
+	int width = head->width;
+	int height = head->height;
+
+	if (width < 320 || height < 240) {
+		qxl_io_log(qdev, "%s: bad head: %dx%d", width, height);
+		width = 1024;
+		height = 768;
+	}
+	if (width * height * 4 > 16*1024*1024) {
+		width = 1024;
+		height = 768;
+	}
+	/* TODO: go over regular modes and removed preferred? */
+	list_for_each_entry_safe(mode, t, &connector->probed_modes, head)
+		drm_mode_remove(connector, mode);
+	mode = drm_cvt_mode(dev, width, height, 60, false, false, false);
+	mode->type |= DRM_MODE_TYPE_PREFERRED;
+	mode->status = MODE_OK;
+	drm_mode_probed_add(connector, mode);
+	qxl_io_log(qdev, "%s: %d x %d\n", __func__, width, height);
+}
+
+void qxl_crtc_set_from_monitors_config(struct qxl_device *qdev)
+{
+	struct drm_connector *connector;
+	int i;
+	struct drm_device *dev = qdev->ddev;
+
+	i = 0;
+	qxl_io_log(qdev, "%s: %d, %d\n", __func__,
+		   dev->mode_config.num_connector,
+		   qdev->monitors_config->count);
+	list_for_each_entry(connector, &dev->mode_config.connector_list, head) {
+		if (i > qdev->monitors_config->count) {
+			/* crtc will be reported as disabled */
+			continue;
+		}
+		qxl_crtc_set_to_mode(qdev, connector,
+				     &qdev->monitors_config->heads[i]);
+		++i;
+	}
+}
+
+void qxl_alloc_client_monitors_config(struct qxl_device *qdev, unsigned count)
+{
+	if (qdev->client_monitors_config &&
+	    count > qdev->client_monitors_config->count) {
+		kfree(qdev->client_monitors_config);
+		qdev->client_monitors_config = NULL;
+	}
+	if (!qdev->client_monitors_config) {
+		qdev->client_monitors_config = kzalloc(
+				sizeof(struct qxl_monitors_config) +
+				sizeof(struct qxl_head) * count, GFP_KERNEL);
+		if (!qdev->client_monitors_config) {
+			qxl_io_log(qdev,
+				   "%s: allocation failure for %u heads\n",
+				   __func__, count);
+			return;
+		}
+	}
+	qdev->client_monitors_config->count = count;
+}
+
+static int qxl_display_copy_rom_client_monitors_config(struct qxl_device *qdev)
+{
+	int i;
+	int num_monitors;
+	uint32_t crc;
+
+	BUG_ON(!qdev->monitors_config);
+	num_monitors = qdev->rom->client_monitors_config.count;
+	crc = crc32(0, (const uint8_t *)&qdev->rom->client_monitors_config,
+		  sizeof(qdev->rom->client_monitors_config));
+	if (crc != qdev->rom->client_monitors_config_crc) {
+		qxl_io_log(qdev, "crc mismatch: have %X (%d) != %X\n", crc,
+			   sizeof(qdev->rom->client_monitors_config),
+			   qdev->rom->client_monitors_config_crc);
+		return 1;
+	}
+	if (num_monitors > qdev->monitors_config->max_allowed) {
+		DRM_INFO("client monitors list will be truncated: %d < %d\n",
+			 qdev->monitors_config->max_allowed, num_monitors);
+		num_monitors = qdev->monitors_config->max_allowed;
+	} else {
+		num_monitors = qdev->rom->client_monitors_config.count;
+	}
+	qxl_alloc_client_monitors_config(qdev, num_monitors);
+	/* we copy max from the client but it isn't used */
+	qdev->client_monitors_config->max_allowed =
+				qdev->monitors_config->max_allowed;
+	for (i = 0 ; i < qdev->client_monitors_config->count ; ++i) {
+		struct qxl_urect *c_rect =
+			&qdev->rom->client_monitors_config.heads[i];
+		struct qxl_head *client_head =
+			&qdev->client_monitors_config->heads[i];
+		struct qxl_head *head = &qdev->monitors_config->heads[i];
+		client_head->x = head->x = c_rect->left;
+		client_head->y = head->y = c_rect->top;
+		client_head->width = head->width =
+						c_rect->right - c_rect->left;
+		client_head->height = head->height =
+						c_rect->bottom - c_rect->top;
+		client_head->surface_id = head->surface_id = 0;
+		client_head->id = head->id = i;
+		client_head->flags = head->flags = 0;
+		QXL_DEBUG(qdev, "read %dx%d+%d+%d\n", head->width, head->height,
+			  head->x, head->y);
+	}
+	return 0;
+}
+
+void qxl_display_read_client_monitors_config(struct qxl_device *qdev)
+{
+
+	while (qxl_display_copy_rom_client_monitors_config(qdev)) {
+		qxl_io_log(qdev, "failed crc check for client_monitors_config,"
+				 " retrying\n");
+	}
+	qxl_crtc_set_from_monitors_config(qdev);
+	/* fire off a uevent and let userspace tell us what to do */
+	qxl_io_log(qdev, "calling drm_sysfs_hotplug_event\n");
+	drm_sysfs_hotplug_event(qdev->ddev);
+}
+
+static int qxl_add_monitors_config_modes(struct drm_connector *connector)
+{
+	struct drm_device *dev = connector->dev;
+	struct qxl_device *qdev = dev->dev_private;
+	struct qxl_output *output = drm_connector_to_qxl_output(connector);
+	int h = output->index;
+	struct drm_display_mode *mode = NULL;
+	struct qxl_head *head;
+
+	if (!qdev->monitors_config)
+		return 0;
+	head = &qdev->monitors_config->heads[h];
+
+	mode = drm_cvt_mode(dev, head->width, head->height, 60, false, false,
+			    false);
+	mode->type |= DRM_MODE_TYPE_PREFERRED;
+	drm_mode_probed_add(connector, mode);
+	return 1;
+}
+
+static int qxl_add_common_modes(struct drm_connector *connector)
+{
+	struct drm_device *dev = connector->dev;
+	struct drm_display_mode *mode = NULL;
+	int i;
+	struct mode_size {
+		int w;
+		int h;
+	} common_modes[] = {
+		{ 640,  480},
+		{ 720,  480},
+		{ 800,  600},
+		{ 848,  480},
+		{1024,  768},
+		{1152,  768},
+		{1280,  720},
+		{1280,  800},
+		{1280,  854},
+		{1280,  960},
+		{1280, 1024},
+		{1440,  900},
+		{1400, 1050},
+		{1680, 1050},
+		{1600, 1200},
+		{1920, 1080},
+		{1920, 1200}
+	};
+
+	for (i = 0; i < ARRAY_SIZE(common_modes); i++) {
+		if (common_modes[i].w < 320 || common_modes[i].h < 200)
+			continue;
+
+		mode = drm_cvt_mode(dev, common_modes[i].w, common_modes[i].h,
+				    60, false, false, false);
+		if (common_modes[i].w == 1024 && common_modes[i].h == 768)
+			mode->type |= DRM_MODE_TYPE_PREFERRED;
+		drm_mode_probed_add(connector, mode);
+	}
+	return i - 1;
+}
+
+static void qxl_crtc_gamma_set(struct drm_crtc *crtc, u16 *red, u16 *green,
+			       u16 *blue, uint32_t start, uint32_t size)
+{
+	/* TODO */
+}
+
+static void qxl_crtc_destroy(struct drm_crtc *crtc)
+{
+	struct qxl_crtc *qxl_crtc = to_qxl_crtc(crtc);
+
+	drm_crtc_cleanup(crtc);
+	kfree(qxl_crtc);
+}
+
+static void
+qxl_hide_cursor(struct qxl_device *qdev)
+{
+	struct qxl_release *release;
+	struct qxl_cursor_cmd *cmd;
+	int ret;
+
+	ret = qxl_alloc_release_reserved(qdev, sizeof(*cmd), QXL_RELEASE_CURSOR_CMD,
+					 &release, NULL);
+
+	cmd = (struct qxl_cursor_cmd *)qxl_release_map(qdev, release);
+	cmd->type = QXL_CURSOR_HIDE;
+	qxl_release_unmap(qdev, release, &cmd->release_info);
+
+	qxl_fence_releaseable(qdev, release);
+	qxl_push_cursor_ring_release(qdev, release, QXL_CMD_CURSOR, false);
+	qxl_release_unreserve(qdev, release);
+}
+
+static int qxl_crtc_cursor_set(struct drm_crtc *crtc,
+			       struct drm_file *file_priv,
+			       uint32_t handle,
+			       uint32_t width,
+			       uint32_t height)
+{
+	struct drm_device *dev = crtc->dev;
+	struct qxl_device *qdev = dev->dev_private;
+	struct qxl_crtc *qcrtc = to_qxl_crtc(crtc);
+	struct drm_gem_object *obj;
+	struct qxl_cursor *cursor;
+	struct qxl_cursor_cmd *cmd;
+	struct qxl_bo *cursor_bo, *user_bo;
+	struct qxl_release *release;
+	void *user_ptr;
+
+	int size = 64*64*4;
+	int ret = 0;
+	if (!handle) {
+		qxl_hide_cursor(qdev);
+		return 0;
+	}
+
+	obj = drm_gem_object_lookup(crtc->dev, file_priv, handle);
+	if (!obj) {
+		DRM_ERROR("cannot find cursor object\n");
+		return -ENOENT;
+	}
+
+	user_bo = gem_to_qxl_bo(obj);
+
+	ret = qxl_bo_reserve(user_bo, false);
+	if (ret)
+		goto out_unref;
+
+	ret = qxl_bo_pin(user_bo, QXL_GEM_DOMAIN_CPU, NULL);
+	if (ret)
+		goto out_unreserve;
+
+	ret = qxl_bo_kmap(user_bo, &user_ptr);
+	if (ret)
+		goto out_unpin;
+
+	ret = qxl_alloc_release_reserved(qdev, sizeof(*cmd),
+					 QXL_RELEASE_CURSOR_CMD,
+					 &release, NULL);
+	if (ret)
+		goto out_kunmap;
+	ret = qxl_alloc_bo_reserved(qdev, sizeof(struct qxl_cursor) + size,
+				    &cursor_bo);
+	if (ret)
+		goto out_free_release;
+	ret = qxl_bo_kmap(cursor_bo, (void **)&cursor);
+	if (ret)
+		goto out_free_bo;
+
+	cursor->header.unique = 0;
+	cursor->header.type = SPICE_CURSOR_TYPE_ALPHA;
+	cursor->header.width = 64;
+	cursor->header.height = 64;
+	cursor->header.hot_spot_x = 0;
+	cursor->header.hot_spot_y = 0;
+	cursor->data_size = size;
+	cursor->chunk.next_chunk = 0;
+	cursor->chunk.prev_chunk = 0;
+	cursor->chunk.data_size = size;
+
+	memcpy(cursor->chunk.data, user_ptr, size);
+
+	qxl_bo_kunmap(cursor_bo);
+
+	/* finish with the userspace bo */
+	qxl_bo_kunmap(user_bo);
+	qxl_bo_unpin(user_bo);
+	qxl_bo_unreserve(user_bo);
+	drm_gem_object_unreference_unlocked(obj);
+
+	cmd = (struct qxl_cursor_cmd *)qxl_release_map(qdev, release);
+	cmd->type = QXL_CURSOR_SET;
+	cmd->u.set.position.x = qcrtc->cur_x;
+	cmd->u.set.position.y = qcrtc->cur_y;
+
+	cmd->u.set.shape = qxl_bo_physical_address(qdev, cursor_bo, 0);
+	qxl_release_add_res(qdev, release, cursor_bo);
+
+	cmd->u.set.visible = 1;
+	qxl_release_unmap(qdev, release, &cmd->release_info);
+
+	qxl_fence_releaseable(qdev, release);
+	qxl_push_cursor_ring_release(qdev, release, QXL_CMD_CURSOR, false);
+	qxl_release_unreserve(qdev, release);
+
+	qxl_bo_unreserve(cursor_bo);
+	qxl_bo_unref(&cursor_bo);
+
+	return ret;
+out_free_bo:
+	qxl_bo_unref(&cursor_bo);
+out_free_release:
+	qxl_release_unreserve(qdev, release);
+	qxl_release_free(qdev, release);
+out_kunmap:
+	qxl_bo_kunmap(user_bo);
+out_unpin:
+	qxl_bo_unpin(user_bo);
+out_unreserve:
+	qxl_bo_unreserve(user_bo);
+out_unref:
+	drm_gem_object_unreference_unlocked(obj);
+	return ret;
+}
+
+static int qxl_crtc_cursor_move(struct drm_crtc *crtc,
+				int x, int y)
+{
+	struct drm_device *dev = crtc->dev;
+	struct qxl_device *qdev = dev->dev_private;
+	struct qxl_crtc *qcrtc = to_qxl_crtc(crtc);
+	struct qxl_release *release;
+	struct qxl_cursor_cmd *cmd;
+	int ret;
+
+	ret = qxl_alloc_release_reserved(qdev, sizeof(*cmd), QXL_RELEASE_CURSOR_CMD,
+				   &release, NULL);
+
+	qcrtc->cur_x = x;
+	qcrtc->cur_y = y;
+
+	cmd = (struct qxl_cursor_cmd *)qxl_release_map(qdev, release);
+	cmd->type = QXL_CURSOR_MOVE;
+	cmd->u.position.x = qcrtc->cur_x;
+	cmd->u.position.y = qcrtc->cur_y;
+	qxl_release_unmap(qdev, release, &cmd->release_info);
+
+	qxl_fence_releaseable(qdev, release);
+	qxl_push_cursor_ring_release(qdev, release, QXL_CMD_CURSOR, false);
+	qxl_release_unreserve(qdev, release);
+	return 0;
+}
+
+
+static const struct drm_crtc_funcs qxl_crtc_funcs = {
+	.cursor_set = qxl_crtc_cursor_set,
+	.cursor_move = qxl_crtc_cursor_move,
+	.gamma_set = qxl_crtc_gamma_set,
+	.set_config = drm_crtc_helper_set_config,
+	.destroy = qxl_crtc_destroy,
+};
+
+static void qxl_user_framebuffer_destroy(struct drm_framebuffer *fb)
+{
+	struct qxl_framebuffer *qxl_fb = to_qxl_framebuffer(fb);
+
+	if (qxl_fb->obj)
+		drm_gem_object_unreference_unlocked(qxl_fb->obj);
+	drm_framebuffer_cleanup(fb);
+	kfree(qxl_fb);
+}
+
+static int qxl_framebuffer_surface_dirty(struct drm_framebuffer *fb,
+					 struct drm_file *file_priv,
+					 unsigned flags, unsigned color,
+					 struct drm_clip_rect *clips,
+					 unsigned num_clips)
+{
+	/* TODO: vmwgfx where this was cribbed from had locking. Why? */
+	struct qxl_framebuffer *qxl_fb = to_qxl_framebuffer(fb);
+	struct qxl_device *qdev = qxl_fb->base.dev->dev_private;
+	struct drm_clip_rect norect;
+	struct qxl_bo *qobj;
+	int inc = 1;
+
+	qobj = gem_to_qxl_bo(qxl_fb->obj);
+	if (qxl_fb != qdev->active_user_framebuffer) {
+		DRM_INFO("%s: qxl_fb 0x%p != qdev->active_user_framebuffer 0x%p\n",
+			__func__, qxl_fb, qdev->active_user_framebuffer);
+	}
+	if (!num_clips) {
+		num_clips = 1;
+		clips = &norect;
+		norect.x1 = norect.y1 = 0;
+		norect.x2 = fb->width;
+		norect.y2 = fb->height;
+	} else if (flags & DRM_MODE_FB_DIRTY_ANNOTATE_COPY) {
+		num_clips /= 2;
+		inc = 2; /* skip source rects */
+	}
+
+	qxl_draw_dirty_fb(qdev, qxl_fb, qobj, flags, color,
+			  clips, num_clips, inc);
+	return 0;
+}
+
+static const struct drm_framebuffer_funcs qxl_fb_funcs = {
+	.destroy = qxl_user_framebuffer_destroy,
+	.dirty = qxl_framebuffer_surface_dirty,
+/*	TODO?
+ *	.create_handle = qxl_user_framebuffer_create_handle, */
+};
+
+int
+qxl_framebuffer_init(struct drm_device *dev,
+		     struct qxl_framebuffer *qfb,
+		     struct drm_mode_fb_cmd2 *mode_cmd,
+		     struct drm_gem_object *obj)
+{
+	int ret;
+
+	qfb->obj = obj;
+	ret = drm_framebuffer_init(dev, &qfb->base, &qxl_fb_funcs);
+	if (ret) {
+		qfb->obj = NULL;
+		return ret;
+	}
+	drm_helper_mode_fill_fb_struct(&qfb->base, mode_cmd);
+	return 0;
+}
+
+static void qxl_crtc_dpms(struct drm_crtc *crtc, int mode)
+{
+}
+
+static bool qxl_crtc_mode_fixup(struct drm_crtc *crtc,
+				  const struct drm_display_mode *mode,
+				  struct drm_display_mode *adjusted_mode)
+{
+	struct drm_device *dev = crtc->dev;
+	struct qxl_device *qdev = dev->dev_private;
+
+	qxl_io_log(qdev, "%s: (%d,%d) => (%d,%d)\n",
+		   __func__,
+		   mode->hdisplay, mode->vdisplay,
+		   adjusted_mode->hdisplay,
+		   adjusted_mode->vdisplay);
+	return true;
+}
+
+void
+qxl_send_monitors_config(struct qxl_device *qdev)
+{
+	int i;
+
+	BUG_ON(!qdev->ram_header->monitors_config);
+
+	if (qdev->monitors_config->count == 0) {
+		qxl_io_log(qdev, "%s: 0 monitors??\n", __func__);
+		return;
+	}
+	for (i = 0 ; i < qdev->monitors_config->count ; ++i) {
+		struct qxl_head *head = &qdev->monitors_config->heads[i];
+
+		if (head->y > 8192 || head->y < head->x ||
+		    head->width > 8192 || head->height > 8192) {
+			DRM_ERROR("head %d wrong: %dx%d+%d+%d\n",
+				  i, head->width, head->height,
+				  head->x, head->y);
+			return;
+		}
+	}
+	qxl_io_monitors_config(qdev);
+}
+
+static void qxl_monitors_config_set_single(struct qxl_device *qdev,
+					   unsigned x, unsigned y,
+					   unsigned width, unsigned height)
+{
+	DRM_DEBUG("%dx%d+%d+%d\n", width, height, x, y);
+	qdev->monitors_config->count = 1;
+	qdev->monitors_config->heads[0].x = x;
+	qdev->monitors_config->heads[0].y = y;
+	qdev->monitors_config->heads[0].width = width;
+	qdev->monitors_config->heads[0].height = height;
+}
+
+static int qxl_crtc_mode_set(struct drm_crtc *crtc,
+			       struct drm_display_mode *mode,
+			       struct drm_display_mode *adjusted_mode,
+			       int x, int y,
+			       struct drm_framebuffer *old_fb)
+{
+	struct drm_device *dev = crtc->dev;
+	struct qxl_device *qdev = dev->dev_private;
+	struct qxl_mode *m = (void *)mode->private;
+	struct qxl_framebuffer *qfb;
+	struct qxl_bo *bo, *old_bo = NULL;
+	uint32_t width, height, base_offset;
+	bool recreate_primary = false;
+	int ret;
+
+	if (!crtc->fb) {
+		DRM_DEBUG_KMS("No FB bound\n");
+		return 0;
+	}
+
+	if (old_fb) {
+		qfb = to_qxl_framebuffer(old_fb);
+		old_bo = gem_to_qxl_bo(qfb->obj);
+	}
+	qfb = to_qxl_framebuffer(crtc->fb);
+	bo = gem_to_qxl_bo(qfb->obj);
+	if (!m)
+		/* and do we care? */
+		DRM_DEBUG("%dx%d: not a native mode\n", x, y);
+	else
+		DRM_DEBUG("%dx%d: qxl id %d\n",
+			  mode->hdisplay, mode->vdisplay, m->id);
+	DRM_DEBUG("+%d+%d (%d,%d) => (%d,%d)\n",
+		  x, y,
+		  mode->hdisplay, mode->vdisplay,
+		  adjusted_mode->hdisplay,
+		  adjusted_mode->vdisplay);
+
+	recreate_primary = true;
+
+	width = mode->hdisplay;
+	height = mode->vdisplay;
+	base_offset = 0;
+
+	ret = qxl_bo_reserve(bo, false);
+	if (ret != 0)
+		return ret;
+	ret = qxl_bo_pin(bo, bo->type, NULL);
+	if (ret != 0) {
+		qxl_bo_unreserve(bo);
+		return -EINVAL;
+	}
+	qxl_bo_unreserve(bo);
+	if (recreate_primary) {
+		qxl_io_destroy_primary(qdev);
+		qxl_io_log(qdev,
+			   "recreate primary: %dx%d (was %dx%d,%d,%d)\n",
+			   width, height, bo->surf.width,
+			   bo->surf.height, bo->surf.stride, bo->surf.format);
+		qxl_io_create_primary(qdev, width, height, base_offset, bo);
+		bo->is_primary = true;
+	}
+
+	if (old_bo && old_bo != bo) {
+		old_bo->is_primary = false;
+		ret = qxl_bo_reserve(old_bo, false);
+		qxl_bo_unpin(old_bo);
+		qxl_bo_unreserve(old_bo);
+	}
+
+	if (qdev->monitors_config->count == 0) {
+		qxl_monitors_config_set_single(qdev, x, y,
+					       mode->hdisplay,
+					       mode->vdisplay);
+	}
+	qdev->mode_set = true;
+	return 0;
+}
+
+static void qxl_crtc_prepare(struct drm_crtc *crtc)
+{
+	DRM_DEBUG("current: %dx%d+%d+%d (%d).\n",
+		  crtc->mode.hdisplay, crtc->mode.vdisplay,
+		  crtc->x, crtc->y, crtc->enabled);
+}
+
+static void qxl_crtc_commit(struct drm_crtc *crtc)
+{
+	DRM_DEBUG("\n");
+}
+
+static void qxl_crtc_load_lut(struct drm_crtc *crtc)
+{
+	DRM_DEBUG("\n");
+}
+
+static const struct drm_crtc_helper_funcs qxl_crtc_helper_funcs = {
+	.dpms = qxl_crtc_dpms,
+	.mode_fixup = qxl_crtc_mode_fixup,
+	.mode_set = qxl_crtc_mode_set,
+	.prepare = qxl_crtc_prepare,
+	.commit = qxl_crtc_commit,
+	.load_lut = qxl_crtc_load_lut,
+};
+
+static int qdev_crtc_init(struct drm_device *dev, int num_crtc)
+{
+	struct qxl_crtc *qxl_crtc;
+
+	qxl_crtc = kzalloc(sizeof(struct qxl_crtc), GFP_KERNEL);
+	if (!qxl_crtc)
+		return -ENOMEM;
+
+	drm_crtc_init(dev, &qxl_crtc->base, &qxl_crtc_funcs);
+
+	drm_mode_crtc_set_gamma_size(&qxl_crtc->base, 256);
+	drm_crtc_helper_add(&qxl_crtc->base, &qxl_crtc_helper_funcs);
+	return 0;
+}
+
+static void qxl_enc_dpms(struct drm_encoder *encoder, int mode)
+{
+	DRM_DEBUG("\n");
+}
+
+static bool qxl_enc_mode_fixup(struct drm_encoder *encoder,
+			       const struct drm_display_mode *mode,
+			       struct drm_display_mode *adjusted_mode)
+{
+	DRM_DEBUG("\n");
+	return true;
+}
+
+static void qxl_enc_prepare(struct drm_encoder *encoder)
+{
+	DRM_DEBUG("\n");
+}
+
+static void qxl_write_monitors_config_for_encoder(struct qxl_device *qdev,
+		struct drm_encoder *encoder)
+{
+	int i;
+	struct qxl_head *head;
+	struct drm_display_mode *mode;
+
+	BUG_ON(!encoder);
+	/* TODO: ugly, do better */
+	for (i = 0 ; (encoder->possible_crtcs != (1 << i)) && i < 32; ++i)
+		;
+	if (encoder->possible_crtcs != (1 << i)) {
+		DRM_ERROR("encoder has wrong possible_crtcs: %x\n",
+			  encoder->possible_crtcs);
+		return;
+	}
+	if (!qdev->monitors_config ||
+	    qdev->monitors_config->max_allowed <= i) {
+		DRM_ERROR(
+		"head number too large or missing monitors config: %p, %d",
+		qdev->monitors_config,
+		qdev->monitors_config ?
+			qdev->monitors_config->max_allowed : -1);
+		return;
+	}
+	if (!encoder->crtc) {
+		DRM_ERROR("missing crtc on encoder %p\n", encoder);
+		return;
+	}
+	if (i != 0)
+		DRM_DEBUG("missing for multiple monitors: no head holes\n");
+	head = &qdev->monitors_config->heads[i];
+	head->id = i;
+	head->surface_id = 0;
+	if (encoder->crtc->enabled) {
+		mode = &encoder->crtc->mode;
+		head->width = mode->hdisplay;
+		head->height = mode->vdisplay;
+		head->x = encoder->crtc->x;
+		head->y = encoder->crtc->y;
+		if (qdev->monitors_config->count < i + 1)
+			qdev->monitors_config->count = i + 1;
+	} else {
+		head->width = 0;
+		head->height = 0;
+		head->x = 0;
+		head->y = 0;
+	}
+	DRM_DEBUG("setting head %d to +%d+%d %dx%d\n",
+		  i, head->x, head->y, head->width, head->height);
+	head->flags = 0;
+	/* TODO - somewhere else to call this for multiple monitors
+	 * (config_commit?) */
+	qxl_send_monitors_config(qdev);
+}
+
+static void qxl_enc_commit(struct drm_encoder *encoder)
+{
+	struct qxl_device *qdev = encoder->dev->dev_private;
+
+	qxl_write_monitors_config_for_encoder(qdev, encoder);
+	DRM_DEBUG("\n");
+}
+
+static void qxl_enc_mode_set(struct drm_encoder *encoder,
+				struct drm_display_mode *mode,
+				struct drm_display_mode *adjusted_mode)
+{
+	DRM_DEBUG("\n");
+}
+
+static int qxl_conn_get_modes(struct drm_connector *connector)
+{
+	int ret = 0;
+	struct qxl_device *qdev = connector->dev->dev_private;
+
+	DRM_DEBUG_KMS("monitors_config=%p\n", qdev->monitors_config);
+	/* TODO: what should we do here? only show the configured modes for the
+	 * device, or allow the full list, or both? */
+	if (qdev->monitors_config && qdev->monitors_config->count) {
+		ret = qxl_add_monitors_config_modes(connector);
+		if (ret < 0)
+			return ret;
+	}
+	ret += qxl_add_common_modes(connector);
+	return ret;
+}
+
+static int qxl_conn_mode_valid(struct drm_connector *connector,
+			       struct drm_display_mode *mode)
+{
+	/* TODO: is this called for user defined modes? (xrandr --add-mode)
+	 * TODO: check that the mode fits in the framebuffer */
+	DRM_DEBUG("%s: %dx%d status=%d\n", mode->name, mode->hdisplay,
+		  mode->vdisplay, mode->status);
+	return MODE_OK;
+}
+
+static struct drm_encoder *qxl_best_encoder(struct drm_connector *connector)
+{
+	struct qxl_output *qxl_output =
+		drm_connector_to_qxl_output(connector);
+
+	DRM_DEBUG("\n");
+	return &qxl_output->enc;
+}
+
+
+static const struct drm_encoder_helper_funcs qxl_enc_helper_funcs = {
+	.dpms = qxl_enc_dpms,
+	.mode_fixup = qxl_enc_mode_fixup,
+	.prepare = qxl_enc_prepare,
+	.mode_set = qxl_enc_mode_set,
+	.commit = qxl_enc_commit,
+};
+
+static const struct drm_connector_helper_funcs qxl_connector_helper_funcs = {
+	.get_modes = qxl_conn_get_modes,
+	.mode_valid = qxl_conn_mode_valid,
+	.best_encoder = qxl_best_encoder,
+};
+
+static void qxl_conn_save(struct drm_connector *connector)
+{
+	DRM_DEBUG("\n");
+}
+
+static void qxl_conn_restore(struct drm_connector *connector)
+{
+	DRM_DEBUG("\n");
+}
+
+static enum drm_connector_status qxl_conn_detect(
+			struct drm_connector *connector,
+			bool force)
+{
+	struct qxl_output *output =
+		drm_connector_to_qxl_output(connector);
+	struct drm_device *ddev = connector->dev;
+	struct qxl_device *qdev = ddev->dev_private;
+	int connected;
+
+	/* The first monitor is always connected */
+	connected = (output->index == 0) ||
+		    (qdev->monitors_config &&
+		     qdev->monitors_config->count > output->index);
+
+	DRM_DEBUG("\n");
+	return connected ? connector_status_connected
+			 : connector_status_disconnected;
+}
+
+static int qxl_conn_set_property(struct drm_connector *connector,
+				   struct drm_property *property,
+				   uint64_t value)
+{
+	DRM_DEBUG("\n");
+	return 0;
+}
+
+static void qxl_conn_destroy(struct drm_connector *connector)
+{
+	struct qxl_output *qxl_output =
+		drm_connector_to_qxl_output(connector);
+
+	drm_sysfs_connector_remove(connector);
+	drm_connector_cleanup(connector);
+	kfree(qxl_output);
+}
+
+static const struct drm_connector_funcs qxl_connector_funcs = {
+	.dpms = drm_helper_connector_dpms,
+	.save = qxl_conn_save,
+	.restore = qxl_conn_restore,
+	.detect = qxl_conn_detect,
+	.fill_modes = drm_helper_probe_single_connector_modes,
+	.set_property = qxl_conn_set_property,
+	.destroy = qxl_conn_destroy,
+};
+
+static void qxl_enc_destroy(struct drm_encoder *encoder)
+{
+	drm_encoder_cleanup(encoder);
+}
+
+static const struct drm_encoder_funcs qxl_enc_funcs = {
+	.destroy = qxl_enc_destroy,
+};
+
+static int qdev_output_init(struct drm_device *dev, int num_output)
+{
+	struct qxl_output *qxl_output;
+	struct drm_connector *connector;
+	struct drm_encoder *encoder;
+
+	qxl_output = kzalloc(sizeof(struct qxl_output), GFP_KERNEL);
+	if (!qxl_output)
+		return -ENOMEM;
+
+	qxl_output->index = num_output;
+
+	connector = &qxl_output->base;
+	encoder = &qxl_output->enc;
+	drm_connector_init(dev, &qxl_output->base,
+			   &qxl_connector_funcs, DRM_MODE_CONNECTOR_VIRTUAL);
+
+	drm_encoder_init(dev, &qxl_output->enc, &qxl_enc_funcs,
+			 DRM_MODE_ENCODER_VIRTUAL);
+
+	encoder->possible_crtcs = 1 << num_output;
+	drm_mode_connector_attach_encoder(&qxl_output->base,
+					  &qxl_output->enc);
+	drm_encoder_helper_add(encoder, &qxl_enc_helper_funcs);
+	drm_connector_helper_add(connector, &qxl_connector_helper_funcs);
+
+	drm_sysfs_connector_add(connector);
+	return 0;
+}
+
+static struct drm_framebuffer *
+qxl_user_framebuffer_create(struct drm_device *dev,
+			    struct drm_file *file_priv,
+			    struct drm_mode_fb_cmd2 *mode_cmd)
+{
+	struct drm_gem_object *obj;
+	struct qxl_framebuffer *qxl_fb;
+	struct qxl_device *qdev = dev->dev_private;
+	int ret;
+
+	obj = drm_gem_object_lookup(dev, file_priv, mode_cmd->handles[0]);
+
+	qxl_fb = kzalloc(sizeof(*qxl_fb), GFP_KERNEL);
+	if (qxl_fb == NULL)
+		return NULL;
+
+	ret = qxl_framebuffer_init(dev, qxl_fb, mode_cmd, obj);
+	if (ret) {
+		kfree(qxl_fb);
+		drm_gem_object_unreference_unlocked(obj);
+		return NULL;
+	}
+
+	if (qdev->active_user_framebuffer) {
+		DRM_INFO("%s: active_user_framebuffer %p -> %p\n",
+			 __func__,
+			 qdev->active_user_framebuffer, qxl_fb);
+	}
+	qdev->active_user_framebuffer = qxl_fb;
+
+	return &qxl_fb->base;
+}
+
+static const struct drm_mode_config_funcs qxl_mode_funcs = {
+	.fb_create = qxl_user_framebuffer_create,
+};
+
+int qxl_modeset_init(struct qxl_device *qdev)
+{
+	int i;
+	int ret;
+	struct drm_gem_object *gobj;
+	int max_allowed = QXL_NUM_OUTPUTS;
+	int monitors_config_size = sizeof(struct qxl_monitors_config) +
+				   max_allowed * sizeof(struct qxl_head);
+
+	drm_mode_config_init(qdev->ddev);
+	ret = qxl_gem_object_create(qdev, monitors_config_size, 0,
+				    QXL_GEM_DOMAIN_VRAM,
+				    false, false, NULL, &gobj);
+	if (ret) {
+		DRM_ERROR("%s: failed to create gem ret=%d\n", __func__, ret);
+		return -ENOMEM;
+	}
+	qdev->monitors_config_bo = gem_to_qxl_bo(gobj);
+	qxl_bo_kmap(qdev->monitors_config_bo, NULL);
+	qdev->monitors_config = qdev->monitors_config_bo->kptr;
+	qdev->ram_header->monitors_config =
+		qxl_bo_physical_address(qdev, qdev->monitors_config_bo, 0);
+
+	memset(qdev->monitors_config, 0, monitors_config_size);
+	qdev->monitors_config->max_allowed = max_allowed;
+
+	qdev->ddev->mode_config.funcs = (void *)&qxl_mode_funcs;
+
+	/* modes will be validated against the framebuffer size */
+	qdev->ddev->mode_config.min_width = 320;
+	qdev->ddev->mode_config.min_height = 200;
+	qdev->ddev->mode_config.max_width = 8192;
+	qdev->ddev->mode_config.max_height = 8192;
+
+	qdev->ddev->mode_config.fb_base = qdev->vram_base;
+	for (i = 0 ; i < QXL_NUM_OUTPUTS; ++i) {
+		qdev_crtc_init(qdev->ddev, i);
+		qdev_output_init(qdev->ddev, i);
+	}
+
+	qdev->mode_info.mode_config_initialized = true;
+
+	/* primary surface must be created by this point, to allow
+	 * issuing command queue commands and having them read by
+	 * spice server. */
+	qxl_fbdev_init(qdev);
+	return 0;
+}
+
+void qxl_modeset_fini(struct qxl_device *qdev)
+{
+	qxl_fbdev_fini(qdev);
+	if (qdev->mode_info.mode_config_initialized) {
+		drm_mode_config_cleanup(qdev->ddev);
+		qdev->mode_info.mode_config_initialized = false;
+	}
+}
diff --git a/drivers/gpu/drm/qxl/qxl_draw.c b/drivers/gpu/drm/qxl/qxl_draw.c
new file mode 100644
index 000000000000..3c8c3dbf9378
--- /dev/null
+++ b/drivers/gpu/drm/qxl/qxl_draw.c
@@ -0,0 +1,390 @@
+/*
+ * Copyright 2011 Red Hat, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * on the rights to use, copy, modify, merge, publish, distribute, sub
+ * license, and/or sell copies of the Software, and to permit persons to whom
+ * the Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER
+ * IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ */
+
+#include "qxl_drv.h"
+#include "qxl_object.h"
+
+/* returns a pointer to the already allocated qxl_rect array inside
+ * the qxl_clip_rects. This is *not* the same as the memory allocated
+ * on the device, it is offset to qxl_clip_rects.chunk.data */
+static struct qxl_rect *drawable_set_clipping(struct qxl_device *qdev,
+					      struct qxl_drawable *drawable,
+					      unsigned num_clips,
+					      struct qxl_bo **clips_bo,
+					      struct qxl_release *release)
+{
+	struct qxl_clip_rects *dev_clips;
+	int ret;
+	int size = sizeof(*dev_clips) + sizeof(struct qxl_rect) * num_clips;
+	ret = qxl_alloc_bo_reserved(qdev, size, clips_bo);
+	if (ret)
+		return NULL;
+
+	ret = qxl_bo_kmap(*clips_bo, (void **)&dev_clips);
+	if (ret) {
+		qxl_bo_unref(clips_bo);
+		return NULL;
+	}
+	dev_clips->num_rects = num_clips;
+	dev_clips->chunk.next_chunk = 0;
+	dev_clips->chunk.prev_chunk = 0;
+	dev_clips->chunk.data_size = sizeof(struct qxl_rect) * num_clips;
+	return (struct qxl_rect *)dev_clips->chunk.data;
+}
+
+static int
+make_drawable(struct qxl_device *qdev, int surface, uint8_t type,
+	      const struct qxl_rect *rect,
+	      struct qxl_release **release)
+{
+	struct qxl_drawable *drawable;
+	int i, ret;
+
+	ret = qxl_alloc_release_reserved(qdev, sizeof(*drawable),
+					 QXL_RELEASE_DRAWABLE, release,
+					 NULL);
+	if (ret)
+		return ret;
+
+	drawable = (struct qxl_drawable *)qxl_release_map(qdev, *release);
+	drawable->type = type;
+
+	drawable->surface_id = surface;		/* Only primary for now */
+	drawable->effect = QXL_EFFECT_OPAQUE;
+	drawable->self_bitmap = 0;
+	drawable->self_bitmap_area.top = 0;
+	drawable->self_bitmap_area.left = 0;
+	drawable->self_bitmap_area.bottom = 0;
+	drawable->self_bitmap_area.right = 0;
+	/* FIXME: add clipping */
+	drawable->clip.type = SPICE_CLIP_TYPE_NONE;
+
+	/*
+	 * surfaces_dest[i] should apparently be filled out with the
+	 * surfaces that we depend on, and surface_rects should be
+	 * filled with the rectangles of those surfaces that we
+	 * are going to use.
+	 */
+	for (i = 0; i < 3; ++i)
+		drawable->surfaces_dest[i] = -1;
+
+	if (rect)
+		drawable->bbox = *rect;
+
+	drawable->mm_time = qdev->rom->mm_clock;
+	qxl_release_unmap(qdev, *release, &drawable->release_info);
+	return 0;
+}
+
+static int qxl_palette_create_1bit(struct qxl_bo **palette_bo,
+				   const struct qxl_fb_image *qxl_fb_image)
+{
+	struct qxl_device *qdev = qxl_fb_image->qdev;
+	const struct fb_image *fb_image = &qxl_fb_image->fb_image;
+	uint32_t visual = qxl_fb_image->visual;
+	const uint32_t *pseudo_palette = qxl_fb_image->pseudo_palette;
+	struct qxl_palette *pal;
+	int ret;
+	uint32_t fgcolor, bgcolor;
+	static uint64_t unique; /* we make no attempt to actually set this
+				 * correctly globaly, since that would require
+				 * tracking all of our palettes. */
+
+	ret = qxl_alloc_bo_reserved(qdev,
+				    sizeof(struct qxl_palette) + sizeof(uint32_t) * 2,
+				    palette_bo);
+
+	ret = qxl_bo_kmap(*palette_bo, (void **)&pal);
+	pal->num_ents = 2;
+	pal->unique = unique++;
+	if (visual == FB_VISUAL_TRUECOLOR || visual == FB_VISUAL_DIRECTCOLOR) {
+		/* NB: this is the only used branch currently. */
+		fgcolor = pseudo_palette[fb_image->fg_color];
+		bgcolor = pseudo_palette[fb_image->bg_color];
+	} else {
+		fgcolor = fb_image->fg_color;
+		bgcolor = fb_image->bg_color;
+	}
+	pal->ents[0] = bgcolor;
+	pal->ents[1] = fgcolor;
+	qxl_bo_kunmap(*palette_bo);
+	return 0;
+}
+
+void qxl_draw_opaque_fb(const struct qxl_fb_image *qxl_fb_image,
+			int stride /* filled in if 0 */)
+{
+	struct qxl_device *qdev = qxl_fb_image->qdev;
+	struct qxl_drawable *drawable;
+	struct qxl_rect rect;
+	const struct fb_image *fb_image = &qxl_fb_image->fb_image;
+	int x = fb_image->dx;
+	int y = fb_image->dy;
+	int width = fb_image->width;
+	int height = fb_image->height;
+	const char *src = fb_image->data;
+	int depth = fb_image->depth;
+	struct qxl_release *release;
+	struct qxl_bo *image_bo;
+	struct qxl_image *image;
+	int ret;
+
+	if (stride == 0)
+		stride = depth * width / 8;
+
+	rect.left = x;
+	rect.right = x + width;
+	rect.top = y;
+	rect.bottom = y + height;
+
+	ret = make_drawable(qdev, 0, QXL_DRAW_COPY, &rect, &release);
+	if (ret)
+		return;
+
+	ret = qxl_image_create(qdev, release, &image_bo,
+			       (const uint8_t *)src, 0, 0,
+			       width, height, depth, stride);
+	if (ret) {
+		qxl_release_unreserve(qdev, release);
+		qxl_release_free(qdev, release);
+		return;
+	}
+
+	if (depth == 1) {
+		struct qxl_bo *palette_bo;
+		void *ptr;
+		ret = qxl_palette_create_1bit(&palette_bo, qxl_fb_image);
+		qxl_release_add_res(qdev, release, palette_bo);
+
+		ptr = qxl_bo_kmap_atomic_page(qdev, image_bo, 0);
+		image = ptr;
+		image->u.bitmap.palette =
+			qxl_bo_physical_address(qdev, palette_bo, 0);
+		qxl_bo_kunmap_atomic_page(qdev, image_bo, ptr);
+		qxl_bo_unreserve(palette_bo);
+		qxl_bo_unref(&palette_bo);
+	}
+
+	drawable = (struct qxl_drawable *)qxl_release_map(qdev, release);
+
+	drawable->u.copy.src_area.top = 0;
+	drawable->u.copy.src_area.bottom = height;
+	drawable->u.copy.src_area.left = 0;
+	drawable->u.copy.src_area.right = width;
+
+	drawable->u.copy.rop_descriptor = SPICE_ROPD_OP_PUT;
+	drawable->u.copy.scale_mode = 0;
+	drawable->u.copy.mask.flags = 0;
+	drawable->u.copy.mask.pos.x = 0;
+	drawable->u.copy.mask.pos.y = 0;
+	drawable->u.copy.mask.bitmap = 0;
+
+	drawable->u.copy.src_bitmap =
+		qxl_bo_physical_address(qdev, image_bo, 0);
+	qxl_release_unmap(qdev, release, &drawable->release_info);
+
+	qxl_release_add_res(qdev, release, image_bo);
+	qxl_bo_unreserve(image_bo);
+	qxl_bo_unref(&image_bo);
+
+	qxl_fence_releaseable(qdev, release);
+	qxl_push_command_ring_release(qdev, release, QXL_CMD_DRAW, false);
+	qxl_release_unreserve(qdev, release);
+}
+
+/* push a draw command using the given clipping rectangles as
+ * the sources from the shadow framebuffer.
+ *
+ * Right now implementing with a single draw and a clip list. Clip
+ * lists are known to be a problem performance wise, this can be solved
+ * by treating them differently in the server.
+ */
+void qxl_draw_dirty_fb(struct qxl_device *qdev,
+		       struct qxl_framebuffer *qxl_fb,
+		       struct qxl_bo *bo,
+		       unsigned flags, unsigned color,
+		       struct drm_clip_rect *clips,
+		       unsigned num_clips, int inc)
+{
+	/*
+	 * TODO: if flags & DRM_MODE_FB_DIRTY_ANNOTATE_FILL then we should
+	 * send a fill command instead, much cheaper.
+	 *
+	 * See include/drm/drm_mode.h
+	 */
+	struct drm_clip_rect *clips_ptr;
+	int i;
+	int left, right, top, bottom;
+	int width, height;
+	struct qxl_drawable *drawable;
+	struct qxl_rect drawable_rect;
+	struct qxl_rect *rects;
+	int stride = qxl_fb->base.pitches[0];
+	/* depth is not actually interesting, we don't mask with it */
+	int depth = qxl_fb->base.bits_per_pixel;
+	uint8_t *surface_base;
+	struct qxl_release *release;
+	struct qxl_bo *image_bo;
+	struct qxl_bo *clips_bo;
+	int ret;
+
+	left = clips->x1;
+	right = clips->x2;
+	top = clips->y1;
+	bottom = clips->y2;
+
+	/* skip the first clip rect */
+	for (i = 1, clips_ptr = clips + inc;
+	     i < num_clips; i++, clips_ptr += inc) {
+		left = min_t(int, left, (int)clips_ptr->x1);
+		right = max_t(int, right, (int)clips_ptr->x2);
+		top = min_t(int, top, (int)clips_ptr->y1);
+		bottom = max_t(int, bottom, (int)clips_ptr->y2);
+	}
+
+	width = right - left;
+	height = bottom - top;
+	drawable_rect.left = left;
+	drawable_rect.right = right;
+	drawable_rect.top = top;
+	drawable_rect.bottom = bottom;
+	ret = make_drawable(qdev, 0, QXL_DRAW_COPY, &drawable_rect,
+			    &release);
+	if (ret)
+		return;
+
+	ret = qxl_bo_kmap(bo, (void **)&surface_base);
+	if (ret)
+		goto out_unref;
+
+	ret = qxl_image_create(qdev, release, &image_bo, surface_base,
+			       left, top, width, height, depth, stride);
+	qxl_bo_kunmap(bo);
+	if (ret)
+		goto out_unref;
+
+	rects = drawable_set_clipping(qdev, drawable, num_clips, &clips_bo, release);
+	if (!rects) {
+		qxl_bo_unref(&image_bo);
+		goto out_unref;
+	}
+	drawable = (struct qxl_drawable *)qxl_release_map(qdev, release);
+
+	drawable->clip.type = SPICE_CLIP_TYPE_RECTS;
+	drawable->clip.data = qxl_bo_physical_address(qdev,
+						      clips_bo, 0);
+	qxl_release_add_res(qdev, release, clips_bo);
+
+	drawable->u.copy.src_area.top = 0;
+	drawable->u.copy.src_area.bottom = height;
+	drawable->u.copy.src_area.left = 0;
+	drawable->u.copy.src_area.right = width;
+
+	drawable->u.copy.rop_descriptor = SPICE_ROPD_OP_PUT;
+	drawable->u.copy.scale_mode = 0;
+	drawable->u.copy.mask.flags = 0;
+	drawable->u.copy.mask.pos.x = 0;
+	drawable->u.copy.mask.pos.y = 0;
+	drawable->u.copy.mask.bitmap = 0;
+
+	drawable->u.copy.src_bitmap = qxl_bo_physical_address(qdev, image_bo, 0);
+	qxl_release_unmap(qdev, release, &drawable->release_info);
+	qxl_release_add_res(qdev, release, image_bo);
+	qxl_bo_unreserve(image_bo);
+	qxl_bo_unref(&image_bo);
+	clips_ptr = clips;
+	for (i = 0; i < num_clips; i++, clips_ptr += inc) {
+		rects[i].left   = clips_ptr->x1;
+		rects[i].right  = clips_ptr->x2;
+		rects[i].top    = clips_ptr->y1;
+		rects[i].bottom = clips_ptr->y2;
+	}
+	qxl_bo_kunmap(clips_bo);
+	qxl_bo_unreserve(clips_bo);
+	qxl_bo_unref(&clips_bo);
+
+	qxl_fence_releaseable(qdev, release);
+	qxl_push_command_ring_release(qdev, release, QXL_CMD_DRAW, false);
+	qxl_release_unreserve(qdev, release);
+	return;
+
+out_unref:
+	qxl_release_unreserve(qdev, release);
+	qxl_release_free(qdev, release);
+}
+
+void qxl_draw_copyarea(struct qxl_device *qdev,
+		       u32 width, u32 height,
+		       u32 sx, u32 sy,
+		       u32 dx, u32 dy)
+{
+	struct qxl_drawable *drawable;
+	struct qxl_rect rect;
+	struct qxl_release *release;
+	int ret;
+
+	rect.left = dx;
+	rect.top = dy;
+	rect.right = dx + width;
+	rect.bottom = dy + height;
+	ret = make_drawable(qdev, 0, QXL_COPY_BITS, &rect, &release);
+	if (ret)
+		return;
+
+	drawable = (struct qxl_drawable *)qxl_release_map(qdev, release);
+	drawable->u.copy_bits.src_pos.x = sx;
+	drawable->u.copy_bits.src_pos.y = sy;
+
+	qxl_release_unmap(qdev, release, &drawable->release_info);
+	qxl_fence_releaseable(qdev, release);
+	qxl_push_command_ring_release(qdev, release, QXL_CMD_DRAW, false);
+	qxl_release_unreserve(qdev, release);
+}
+
+void qxl_draw_fill(struct qxl_draw_fill *qxl_draw_fill_rec)
+{
+	struct qxl_device *qdev = qxl_draw_fill_rec->qdev;
+	struct qxl_rect rect = qxl_draw_fill_rec->rect;
+	uint32_t color = qxl_draw_fill_rec->color;
+	uint16_t rop = qxl_draw_fill_rec->rop;
+	struct qxl_drawable *drawable;
+	struct qxl_release *release;
+	int ret;
+
+	ret = make_drawable(qdev, 0, QXL_DRAW_FILL, &rect, &release);
+	if (ret)
+		return;
+
+	drawable = (struct qxl_drawable *)qxl_release_map(qdev, release);
+	drawable->u.fill.brush.type = SPICE_BRUSH_TYPE_SOLID;
+	drawable->u.fill.brush.u.color = color;
+	drawable->u.fill.rop_descriptor = rop;
+	drawable->u.fill.mask.flags = 0;
+	drawable->u.fill.mask.pos.x = 0;
+	drawable->u.fill.mask.pos.y = 0;
+	drawable->u.fill.mask.bitmap = 0;
+
+	qxl_release_unmap(qdev, release, &drawable->release_info);
+	qxl_fence_releaseable(qdev, release);
+	qxl_push_command_ring_release(qdev, release, QXL_CMD_DRAW, false);
+	qxl_release_unreserve(qdev, release);
+}
diff --git a/drivers/gpu/drm/qxl/qxl_drv.c b/drivers/gpu/drm/qxl/qxl_drv.c
new file mode 100644
index 000000000000..aa291d8a98a2
--- /dev/null
+++ b/drivers/gpu/drm/qxl/qxl_drv.c
@@ -0,0 +1,145 @@
+/* vim: set ts=8 sw=8 tw=78 ai noexpandtab */
+/* qxl_drv.c -- QXL driver -*- linux-c -*-
+ *
+ * Copyright 2011 Red Hat, Inc.
+ * All Rights Reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * VA LINUX SYSTEMS AND/OR ITS SUPPLIERS BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors:
+ *    Dave Airlie <airlie@redhat.com>
+ *    Alon Levy <alevy@redhat.com>
+ */
+
+#include <linux/module.h>
+#include <linux/console.h>
+
+#include "drmP.h"
+#include "drm/drm.h"
+
+#include "qxl_drv.h"
+
+extern int qxl_max_ioctls;
+static DEFINE_PCI_DEVICE_TABLE(pciidlist) = {
+	{ 0x1b36, 0x100, PCI_ANY_ID, PCI_ANY_ID, PCI_CLASS_DISPLAY_VGA << 8,
+	  0xffff00, 0 },
+	{ 0x1b36, 0x100, PCI_ANY_ID, PCI_ANY_ID, PCI_CLASS_DISPLAY_OTHER << 8,
+	  0xffff00, 0 },
+	{ 0, 0, 0 },
+};
+MODULE_DEVICE_TABLE(pci, pciidlist);
+
+static int qxl_modeset = -1;
+
+MODULE_PARM_DESC(modeset, "Disable/Enable modesetting");
+module_param_named(modeset, qxl_modeset, int, 0400);
+
+static struct drm_driver qxl_driver;
+static struct pci_driver qxl_pci_driver;
+
+static int
+qxl_pci_probe(struct pci_dev *pdev, const struct pci_device_id *ent)
+{
+	if (pdev->revision < 4) {
+		DRM_ERROR("qxl too old, doesn't support client_monitors_config,"
+			  " use xf86-video-qxl in user mode");
+		return -EINVAL; /* TODO: ENODEV ? */
+	}
+	return drm_get_pci_dev(pdev, ent, &qxl_driver);
+}
+
+static void
+qxl_pci_remove(struct pci_dev *pdev)
+{
+	struct drm_device *dev = pci_get_drvdata(pdev);
+
+	drm_put_dev(dev);
+}
+
+static struct pci_driver qxl_pci_driver = {
+	 .name = DRIVER_NAME,
+	 .id_table = pciidlist,
+	 .probe = qxl_pci_probe,
+	 .remove = qxl_pci_remove,
+};
+
+static const struct file_operations qxl_fops = {
+	.owner = THIS_MODULE,
+	.open = drm_open,
+	.release = drm_release,
+	.unlocked_ioctl = drm_ioctl,
+	.poll = drm_poll,
+	.fasync = drm_fasync,
+	.mmap = qxl_mmap,
+};
+
+static struct drm_driver qxl_driver = {
+	.driver_features = DRIVER_GEM | DRIVER_MODESET |
+			   DRIVER_HAVE_IRQ | DRIVER_IRQ_SHARED,
+	.dev_priv_size = 0,
+	.load = qxl_driver_load,
+	.unload = qxl_driver_unload,
+
+	.dumb_create = qxl_mode_dumb_create,
+	.dumb_map_offset = qxl_mode_dumb_mmap,
+	.dumb_destroy = qxl_mode_dumb_destroy,
+#if defined(CONFIG_DEBUG_FS)
+	.debugfs_init = qxl_debugfs_init,
+	.debugfs_cleanup = qxl_debugfs_takedown,
+#endif
+	.gem_init_object = qxl_gem_object_init,
+	.gem_free_object = qxl_gem_object_free,
+	.gem_open_object = qxl_gem_object_open,
+	.gem_close_object = qxl_gem_object_close,
+	.fops = &qxl_fops,
+	.ioctls = qxl_ioctls,
+	.irq_handler = qxl_irq_handler,
+	.name = DRIVER_NAME,
+	.desc = DRIVER_DESC,
+	.date = DRIVER_DATE,
+	.major = 0,
+	.minor = 1,
+	.patchlevel = 0,
+};
+
+static int __init qxl_init(void)
+{
+#ifdef CONFIG_VGA_CONSOLE
+	if (vgacon_text_force() && qxl_modeset == -1)
+		return -EINVAL;
+#endif
+
+	if (qxl_modeset == 0)
+		return -EINVAL;
+	qxl_driver.num_ioctls = qxl_max_ioctls;
+	return drm_pci_init(&qxl_driver, &qxl_pci_driver);
+}
+
+static void __exit qxl_exit(void)
+{
+	drm_pci_exit(&qxl_driver, &qxl_pci_driver);
+}
+
+module_init(qxl_init);
+module_exit(qxl_exit);
+
+MODULE_AUTHOR(DRIVER_AUTHOR);
+MODULE_DESCRIPTION(DRIVER_DESC);
+MODULE_LICENSE("GPL and additional rights");
diff --git a/drivers/gpu/drm/qxl/qxl_drv.h b/drivers/gpu/drm/qxl/qxl_drv.h
new file mode 100644
index 000000000000..52b582c211da
--- /dev/null
+++ b/drivers/gpu/drm/qxl/qxl_drv.h
@@ -0,0 +1,566 @@
+/*
+ * Copyright 2013 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Dave Airlie
+ *          Alon Levy
+ */
+
+
+#ifndef QXL_DRV_H
+#define QXL_DRV_H
+
+/*
+ * Definitions taken from spice-protocol, plus kernel driver specific bits.
+ */
+
+#include <linux/workqueue.h>
+#include <linux/firmware.h>
+#include <linux/platform_device.h>
+
+#include "drmP.h"
+#include "drm_crtc.h"
+#include <ttm/ttm_bo_api.h>
+#include <ttm/ttm_bo_driver.h>
+#include <ttm/ttm_placement.h>
+#include <ttm/ttm_module.h>
+
+#include <drm/qxl_drm.h>
+#include "qxl_dev.h"
+
+#define DRIVER_AUTHOR		"Dave Airlie"
+
+#define DRIVER_NAME		"qxl"
+#define DRIVER_DESC		"RH QXL"
+#define DRIVER_DATE		"20120117"
+
+#define DRIVER_MAJOR 0
+#define DRIVER_MINOR 1
+#define DRIVER_PATCHLEVEL 0
+
+#define QXL_NUM_OUTPUTS 1
+
+#define QXL_DEBUGFS_MAX_COMPONENTS		32
+
+extern int qxl_log_level;
+
+enum {
+	QXL_INFO_LEVEL = 1,
+	QXL_DEBUG_LEVEL = 2,
+};
+
+#define QXL_INFO(qdev, fmt, ...) do { \
+		if (qxl_log_level >= QXL_INFO_LEVEL) {	\
+			qxl_io_log(qdev, fmt, __VA_ARGS__); \
+		}	\
+	} while (0)
+#define QXL_DEBUG(qdev, fmt, ...) do { \
+		if (qxl_log_level >= QXL_DEBUG_LEVEL) {	\
+			qxl_io_log(qdev, fmt, __VA_ARGS__); \
+		}	\
+	} while (0)
+#define QXL_INFO_ONCE(qdev, fmt, ...) do { \
+		static int done;		\
+		if (!done) {			\
+			done = 1;			\
+			QXL_INFO(qdev, fmt, __VA_ARGS__);	\
+		}						\
+	} while (0)
+
+#define DRM_FILE_OFFSET 0x100000000ULL
+#define DRM_FILE_PAGE_OFFSET (DRM_FILE_OFFSET >> PAGE_SHIFT)
+
+#define QXL_INTERRUPT_MASK (\
+	QXL_INTERRUPT_DISPLAY |\
+	QXL_INTERRUPT_CURSOR |\
+	QXL_INTERRUPT_IO_CMD |\
+	QXL_INTERRUPT_CLIENT_MONITORS_CONFIG)
+
+struct qxl_fence {
+	struct qxl_device *qdev;
+	uint32_t num_active_releases;
+	uint32_t *release_ids;
+	struct radix_tree_root tree;
+};
+
+struct qxl_bo {
+	/* Protected by gem.mutex */
+	struct list_head		list;
+	/* Protected by tbo.reserved */
+	u32				placements[3];
+	struct ttm_placement		placement;
+	struct ttm_buffer_object	tbo;
+	struct ttm_bo_kmap_obj		kmap;
+	unsigned			pin_count;
+	void				*kptr;
+	int                             type;
+	/* Constant after initialization */
+	struct drm_gem_object		gem_base;
+	bool is_primary; /* is this now a primary surface */
+	bool hw_surf_alloc;
+	struct qxl_surface surf;
+	uint32_t surface_id;
+	struct qxl_fence fence; /* per bo fence  - list of releases */
+	struct qxl_release *surf_create;
+	atomic_t reserve_count;
+};
+#define gem_to_qxl_bo(gobj) container_of((gobj), struct qxl_bo, gem_base)
+
+struct qxl_gem {
+	struct mutex		mutex;
+	struct list_head	objects;
+};
+
+struct qxl_bo_list {
+	struct list_head lhead;
+	struct qxl_bo *bo;
+};
+
+struct qxl_reloc_list {
+	struct list_head bos;
+};
+
+struct qxl_crtc {
+	struct drm_crtc base;
+	int cur_x;
+	int cur_y;
+};
+
+struct qxl_output {
+	int index;
+	struct drm_connector base;
+	struct drm_encoder enc;
+};
+
+struct qxl_framebuffer {
+	struct drm_framebuffer base;
+	struct drm_gem_object *obj;
+};
+
+#define to_qxl_crtc(x) container_of(x, struct qxl_crtc, base)
+#define drm_connector_to_qxl_output(x) container_of(x, struct qxl_output, base)
+#define drm_encoder_to_qxl_output(x) container_of(x, struct qxl_output, base)
+#define to_qxl_framebuffer(x) container_of(x, struct qxl_framebuffer, base)
+
+struct qxl_mman {
+	struct ttm_bo_global_ref        bo_global_ref;
+	struct drm_global_reference	mem_global_ref;
+	bool				mem_global_referenced;
+	struct ttm_bo_device		bdev;
+};
+
+struct qxl_mode_info {
+	int num_modes;
+	struct qxl_mode *modes;
+	bool mode_config_initialized;
+
+	/* pointer to fbdev info structure */
+	struct qxl_fbdev *qfbdev;
+};
+
+
+struct qxl_memslot {
+	uint8_t		generation;
+	uint64_t	start_phys_addr;
+	uint64_t	end_phys_addr;
+	uint64_t	high_bits;
+};
+
+enum {
+	QXL_RELEASE_DRAWABLE,
+	QXL_RELEASE_SURFACE_CMD,
+	QXL_RELEASE_CURSOR_CMD,
+};
+
+/* drm_ prefix to differentiate from qxl_release_info in
+ * spice-protocol/qxl_dev.h */
+#define QXL_MAX_RES 96
+struct qxl_release {
+	int id;
+	int type;
+	int bo_count;
+	uint32_t release_offset;
+	uint32_t surface_release_id;
+	struct qxl_bo *bos[QXL_MAX_RES];
+};
+
+struct qxl_fb_image {
+	struct qxl_device *qdev;
+	uint32_t pseudo_palette[16];
+	struct fb_image fb_image;
+	uint32_t visual;
+};
+
+struct qxl_draw_fill {
+	struct qxl_device *qdev;
+	struct qxl_rect rect;
+	uint32_t color;
+	uint16_t rop;
+};
+
+/*
+ * Debugfs
+ */
+struct qxl_debugfs {
+	struct drm_info_list	*files;
+	unsigned		num_files;
+};
+
+int qxl_debugfs_add_files(struct qxl_device *rdev,
+			     struct drm_info_list *files,
+			     unsigned nfiles);
+int qxl_debugfs_fence_init(struct qxl_device *rdev);
+void qxl_debugfs_remove_files(struct qxl_device *qdev);
+
+struct qxl_device;
+
+struct qxl_device {
+	struct device			*dev;
+	struct drm_device		*ddev;
+	struct pci_dev			*pdev;
+	unsigned long flags;
+
+	resource_size_t vram_base, vram_size;
+	resource_size_t surfaceram_base, surfaceram_size;
+	resource_size_t rom_base, rom_size;
+	struct qxl_rom *rom;
+
+	struct qxl_mode *modes;
+	struct qxl_bo *monitors_config_bo;
+	struct qxl_monitors_config *monitors_config;
+
+	/* last received client_monitors_config */
+	struct qxl_monitors_config *client_monitors_config;
+
+	int io_base;
+	void *ram;
+	struct qxl_mman		mman;
+	struct qxl_gem		gem;
+	struct qxl_mode_info mode_info;
+
+	/*
+	 * last created framebuffer with fb_create
+	 * only used by debugfs dumbppm
+	 */
+	struct qxl_framebuffer *active_user_framebuffer;
+
+	struct fb_info			*fbdev_info;
+	struct qxl_framebuffer	*fbdev_qfb;
+	void *ram_physical;
+
+	struct qxl_ring *release_ring;
+	struct qxl_ring *command_ring;
+	struct qxl_ring *cursor_ring;
+
+	struct qxl_ram_header *ram_header;
+	bool mode_set;
+
+	bool primary_created;
+
+	struct qxl_memslot	*mem_slots;
+	uint8_t		n_mem_slots;
+
+	uint8_t		main_mem_slot;
+	uint8_t		surfaces_mem_slot;
+	uint8_t		slot_id_bits;
+	uint8_t		slot_gen_bits;
+	uint64_t	va_slot_mask;
+
+	struct idr	release_idr;
+	spinlock_t release_idr_lock;
+	struct mutex	async_io_mutex;
+	unsigned int last_sent_io_cmd;
+
+	/* interrupt handling */
+	atomic_t irq_received;
+	atomic_t irq_received_display;
+	atomic_t irq_received_cursor;
+	atomic_t irq_received_io_cmd;
+	unsigned irq_received_error;
+	wait_queue_head_t display_event;
+	wait_queue_head_t cursor_event;
+	wait_queue_head_t io_cmd_event;
+	struct work_struct client_monitors_config_work;
+
+	/* debugfs */
+	struct qxl_debugfs	debugfs[QXL_DEBUGFS_MAX_COMPONENTS];
+	unsigned		debugfs_count;
+
+	struct mutex		update_area_mutex;
+
+	struct idr	surf_id_idr;
+	spinlock_t surf_id_idr_lock;
+	int last_alloced_surf_id;
+
+	struct mutex surf_evict_mutex;
+	struct io_mapping *vram_mapping;
+	struct io_mapping *surface_mapping;
+
+	/* */
+	struct mutex release_mutex;
+	struct qxl_bo *current_release_bo[3];
+	int current_release_bo_offset[3];
+
+	struct workqueue_struct *gc_queue;
+	struct work_struct gc_work;
+
+};
+
+/* forward declaration for QXL_INFO_IO */
+void qxl_io_log(struct qxl_device *qdev, const char *fmt, ...);
+
+extern struct drm_ioctl_desc qxl_ioctls[];
+extern int qxl_max_ioctl;
+
+int qxl_driver_load(struct drm_device *dev, unsigned long flags);
+int qxl_driver_unload(struct drm_device *dev);
+
+int qxl_modeset_init(struct qxl_device *qdev);
+void qxl_modeset_fini(struct qxl_device *qdev);
+
+int qxl_bo_init(struct qxl_device *qdev);
+void qxl_bo_fini(struct qxl_device *qdev);
+
+struct qxl_ring *qxl_ring_create(struct qxl_ring_header *header,
+				 int element_size,
+				 int n_elements,
+				 int prod_notify,
+				 bool set_prod_notify,
+				 wait_queue_head_t *push_event);
+void qxl_ring_free(struct qxl_ring *ring);
+
+static inline void *
+qxl_fb_virtual_address(struct qxl_device *qdev, unsigned long physical)
+{
+	QXL_INFO(qdev, "not implemented (%lu)\n", physical);
+	return 0;
+}
+
+static inline uint64_t
+qxl_bo_physical_address(struct qxl_device *qdev, struct qxl_bo *bo,
+			unsigned long offset)
+{
+	int slot_id = bo->type == QXL_GEM_DOMAIN_VRAM ? qdev->main_mem_slot : qdev->surfaces_mem_slot;
+	struct qxl_memslot *slot = &(qdev->mem_slots[slot_id]);
+
+	/* TODO - need to hold one of the locks to read tbo.offset */
+	return slot->high_bits | (bo->tbo.offset + offset);
+}
+
+/* qxl_fb.c */
+#define QXLFB_CONN_LIMIT 1
+
+int qxl_fbdev_init(struct qxl_device *qdev);
+void qxl_fbdev_fini(struct qxl_device *qdev);
+int qxl_get_handle_for_primary_fb(struct qxl_device *qdev,
+				  struct drm_file *file_priv,
+				  uint32_t *handle);
+
+/* qxl_display.c */
+int
+qxl_framebuffer_init(struct drm_device *dev,
+		     struct qxl_framebuffer *rfb,
+		     struct drm_mode_fb_cmd2 *mode_cmd,
+		     struct drm_gem_object *obj);
+void qxl_display_read_client_monitors_config(struct qxl_device *qdev);
+void qxl_send_monitors_config(struct qxl_device *qdev);
+
+/* used by qxl_debugfs only */
+void qxl_crtc_set_from_monitors_config(struct qxl_device *qdev);
+void qxl_alloc_client_monitors_config(struct qxl_device *qdev, unsigned count);
+
+/* qxl_gem.c */
+int qxl_gem_init(struct qxl_device *qdev);
+void qxl_gem_fini(struct qxl_device *qdev);
+int qxl_gem_object_create(struct qxl_device *qdev, int size,
+			  int alignment, int initial_domain,
+			  bool discardable, bool kernel,
+			  struct qxl_surface *surf,
+			  struct drm_gem_object **obj);
+int qxl_gem_object_pin(struct drm_gem_object *obj, uint32_t pin_domain,
+			  uint64_t *gpu_addr);
+void qxl_gem_object_unpin(struct drm_gem_object *obj);
+int qxl_gem_object_create_with_handle(struct qxl_device *qdev,
+				      struct drm_file *file_priv,
+				      u32 domain,
+				      size_t size,
+				      struct qxl_surface *surf,
+				      struct qxl_bo **qobj,
+				      uint32_t *handle);
+int qxl_gem_object_init(struct drm_gem_object *obj);
+void qxl_gem_object_free(struct drm_gem_object *gobj);
+int qxl_gem_object_open(struct drm_gem_object *obj, struct drm_file *file_priv);
+void qxl_gem_object_close(struct drm_gem_object *obj,
+			  struct drm_file *file_priv);
+void qxl_bo_force_delete(struct qxl_device *qdev);
+int qxl_bo_kmap(struct qxl_bo *bo, void **ptr);
+
+/* qxl_dumb.c */
+int qxl_mode_dumb_create(struct drm_file *file_priv,
+			 struct drm_device *dev,
+			 struct drm_mode_create_dumb *args);
+int qxl_mode_dumb_destroy(struct drm_file *file_priv,
+			  struct drm_device *dev,
+			  uint32_t handle);
+int qxl_mode_dumb_mmap(struct drm_file *filp,
+		       struct drm_device *dev,
+		       uint32_t handle, uint64_t *offset_p);
+
+
+/* qxl ttm */
+int qxl_ttm_init(struct qxl_device *qdev);
+void qxl_ttm_fini(struct qxl_device *qdev);
+int qxl_mmap(struct file *filp, struct vm_area_struct *vma);
+
+/* qxl image */
+
+int qxl_image_create(struct qxl_device *qdev,
+		     struct qxl_release *release,
+		     struct qxl_bo **image_bo,
+		     const uint8_t *data,
+		     int x, int y, int width, int height,
+		     int depth, int stride);
+void qxl_update_screen(struct qxl_device *qxl);
+
+/* qxl io operations (qxl_cmd.c) */
+
+void qxl_io_create_primary(struct qxl_device *qdev,
+			   unsigned width, unsigned height, unsigned offset,
+			   struct qxl_bo *bo);
+void qxl_io_destroy_primary(struct qxl_device *qdev);
+void qxl_io_memslot_add(struct qxl_device *qdev, uint8_t id);
+void qxl_io_notify_oom(struct qxl_device *qdev);
+
+int qxl_io_update_area(struct qxl_device *qdev, struct qxl_bo *surf,
+		       const struct qxl_rect *area);
+
+void qxl_io_reset(struct qxl_device *qdev);
+void qxl_io_monitors_config(struct qxl_device *qdev);
+int qxl_ring_push(struct qxl_ring *ring, const void *new_elt, bool interruptible);
+void qxl_io_flush_release(struct qxl_device *qdev);
+void qxl_io_flush_surfaces(struct qxl_device *qdev);
+
+int qxl_release_reserve(struct qxl_device *qdev,
+			struct qxl_release *release, bool no_wait);
+void qxl_release_unreserve(struct qxl_device *qdev,
+			   struct qxl_release *release);
+union qxl_release_info *qxl_release_map(struct qxl_device *qdev,
+					struct qxl_release *release);
+void qxl_release_unmap(struct qxl_device *qdev,
+		       struct qxl_release *release,
+		       union qxl_release_info *info);
+/*
+ * qxl_bo_add_resource.
+ *
+ */
+void qxl_bo_add_resource(struct qxl_bo *main_bo, struct qxl_bo *resource);
+
+int qxl_alloc_surface_release_reserved(struct qxl_device *qdev,
+				       enum qxl_surface_cmd_type surface_cmd_type,
+				       struct qxl_release *create_rel,
+				       struct qxl_release **release);
+int qxl_alloc_release_reserved(struct qxl_device *qdev, unsigned long size,
+			       int type, struct qxl_release **release,
+			       struct qxl_bo **rbo);
+int qxl_fence_releaseable(struct qxl_device *qdev,
+			  struct qxl_release *release);
+int
+qxl_push_command_ring_release(struct qxl_device *qdev, struct qxl_release *release,
+			      uint32_t type, bool interruptible);
+int
+qxl_push_cursor_ring_release(struct qxl_device *qdev, struct qxl_release *release,
+			     uint32_t type, bool interruptible);
+int qxl_alloc_bo_reserved(struct qxl_device *qdev, unsigned long size,
+			  struct qxl_bo **_bo);
+/* qxl drawing commands */
+
+void qxl_draw_opaque_fb(const struct qxl_fb_image *qxl_fb_image,
+			int stride /* filled in if 0 */);
+
+void qxl_draw_dirty_fb(struct qxl_device *qdev,
+		       struct qxl_framebuffer *qxl_fb,
+		       struct qxl_bo *bo,
+		       unsigned flags, unsigned color,
+		       struct drm_clip_rect *clips,
+		       unsigned num_clips, int inc);
+
+void qxl_draw_fill(struct qxl_draw_fill *qxl_draw_fill_rec);
+
+void qxl_draw_copyarea(struct qxl_device *qdev,
+		       u32 width, u32 height,
+		       u32 sx, u32 sy,
+		       u32 dx, u32 dy);
+
+uint64_t
+qxl_release_alloc(struct qxl_device *qdev, int type,
+		  struct qxl_release **ret);
+
+void qxl_release_free(struct qxl_device *qdev,
+		      struct qxl_release *release);
+void qxl_release_add_res(struct qxl_device *qdev,
+			 struct qxl_release *release,
+			 struct qxl_bo *bo);
+/* used by qxl_debugfs_release */
+struct qxl_release *qxl_release_from_id_locked(struct qxl_device *qdev,
+						   uint64_t id);
+
+bool qxl_queue_garbage_collect(struct qxl_device *qdev, bool flush);
+int qxl_garbage_collect(struct qxl_device *qdev);
+
+/* debugfs */
+
+int qxl_debugfs_init(struct drm_minor *minor);
+void qxl_debugfs_takedown(struct drm_minor *minor);
+
+/* qxl_irq.c */
+int qxl_irq_init(struct qxl_device *qdev);
+irqreturn_t qxl_irq_handler(DRM_IRQ_ARGS);
+
+/* qxl_fb.c */
+int qxl_fb_init(struct qxl_device *qdev);
+
+int qxl_debugfs_add_files(struct qxl_device *qdev,
+			  struct drm_info_list *files,
+			  unsigned nfiles);
+
+int qxl_surface_id_alloc(struct qxl_device *qdev,
+			 struct qxl_bo *surf);
+void qxl_surface_id_dealloc(struct qxl_device *qdev,
+			    uint32_t surface_id);
+int qxl_hw_surface_alloc(struct qxl_device *qdev,
+			 struct qxl_bo *surf,
+			 struct ttm_mem_reg *mem);
+int qxl_hw_surface_dealloc(struct qxl_device *qdev,
+			   struct qxl_bo *surf);
+
+int qxl_bo_check_id(struct qxl_device *qdev, struct qxl_bo *bo);
+
+struct qxl_drv_surface *
+qxl_surface_lookup(struct drm_device *dev, int surface_id);
+void qxl_surface_evict(struct qxl_device *qdev, struct qxl_bo *surf, bool freeing);
+int qxl_update_surface(struct qxl_device *qdev, struct qxl_bo *surf);
+
+/* qxl_fence.c */
+int qxl_fence_add_release(struct qxl_fence *qfence, uint32_t rel_id);
+int qxl_fence_remove_release(struct qxl_fence *qfence, uint32_t rel_id);
+int qxl_fence_init(struct qxl_device *qdev, struct qxl_fence *qfence);
+void qxl_fence_fini(struct qxl_fence *qfence);
+
+#endif
diff --git a/drivers/gpu/drm/qxl/qxl_dumb.c b/drivers/gpu/drm/qxl/qxl_dumb.c
new file mode 100644
index 000000000000..847c4ee798f7
--- /dev/null
+++ b/drivers/gpu/drm/qxl/qxl_dumb.c
@@ -0,0 +1,93 @@
+/*
+ * Copyright 2013 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Dave Airlie
+ *          Alon Levy
+ */
+
+#include "qxl_drv.h"
+#include "qxl_object.h"
+
+/* dumb ioctls implementation */
+
+int qxl_mode_dumb_create(struct drm_file *file_priv,
+			    struct drm_device *dev,
+			    struct drm_mode_create_dumb *args)
+{
+	struct qxl_device *qdev = dev->dev_private;
+	struct qxl_bo *qobj;
+	uint32_t handle;
+	int r;
+	struct qxl_surface surf;
+	uint32_t pitch, format;
+	pitch = args->width * ((args->bpp + 1) / 8);
+	args->size = pitch * args->height;
+	args->size = ALIGN(args->size, PAGE_SIZE);
+
+	switch (args->bpp) {
+	case 16:
+		format = SPICE_SURFACE_FMT_16_565;
+		break;
+	case 32:
+		format = SPICE_SURFACE_FMT_32_xRGB;
+		break;
+	default:
+		return -EINVAL;
+	}
+	  
+	surf.width = args->width;
+	surf.height = args->height;
+	surf.stride = pitch;
+	surf.format = format;
+	r = qxl_gem_object_create_with_handle(qdev, file_priv,
+					      QXL_GEM_DOMAIN_VRAM,
+					      args->size, &surf, &qobj,
+					      &handle);
+	if (r)
+		return r;
+	args->pitch = pitch;
+	args->handle = handle;
+	return 0;
+}
+
+int qxl_mode_dumb_destroy(struct drm_file *file_priv,
+			     struct drm_device *dev,
+			     uint32_t handle)
+{
+	return drm_gem_handle_delete(file_priv, handle);
+}
+
+int qxl_mode_dumb_mmap(struct drm_file *file_priv,
+		       struct drm_device *dev,
+		       uint32_t handle, uint64_t *offset_p)
+{
+	struct drm_gem_object *gobj;
+	struct qxl_bo *qobj;
+
+	BUG_ON(!offset_p);
+	gobj = drm_gem_object_lookup(dev, file_priv, handle);
+	if (gobj == NULL)
+		return -ENOENT;
+	qobj = gem_to_qxl_bo(gobj);
+	*offset_p = qxl_bo_mmap_offset(qobj);
+	drm_gem_object_unreference_unlocked(gobj);
+	return 0;
+}
diff --git a/drivers/gpu/drm/qxl/qxl_fb.c b/drivers/gpu/drm/qxl/qxl_fb.c
new file mode 100644
index 000000000000..b3c51275df5c
--- /dev/null
+++ b/drivers/gpu/drm/qxl/qxl_fb.c
@@ -0,0 +1,567 @@
+/*
+ * Copyright © 2013 Red Hat
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ *
+ * Authors:
+ *     David Airlie
+ */
+#include <linux/module.h>
+#include <linux/fb.h>
+
+#include "drmP.h"
+#include "drm/drm.h"
+#include "drm/drm_crtc.h"
+#include "drm/drm_crtc_helper.h"
+#include "qxl_drv.h"
+
+#include "qxl_object.h"
+#include "drm_fb_helper.h"
+
+#define QXL_DIRTY_DELAY (HZ / 30)
+
+struct qxl_fbdev {
+	struct drm_fb_helper helper;
+	struct qxl_framebuffer	qfb;
+	struct list_head	fbdev_list;
+	struct qxl_device	*qdev;
+
+	void *shadow;
+	int size;
+
+	/* dirty memory logging */
+	struct {
+		spinlock_t lock;
+		bool active;
+		unsigned x1;
+		unsigned y1;
+		unsigned x2;
+		unsigned y2;
+	} dirty;
+};
+
+static void qxl_fb_image_init(struct qxl_fb_image *qxl_fb_image,
+			      struct qxl_device *qdev, struct fb_info *info,
+			      const struct fb_image *image)
+{
+	qxl_fb_image->qdev = qdev;
+	if (info) {
+		qxl_fb_image->visual = info->fix.visual;
+		if (qxl_fb_image->visual == FB_VISUAL_TRUECOLOR ||
+		    qxl_fb_image->visual == FB_VISUAL_DIRECTCOLOR)
+			memcpy(&qxl_fb_image->pseudo_palette,
+			       info->pseudo_palette,
+			       sizeof(qxl_fb_image->pseudo_palette));
+	} else {
+		 /* fallback */
+		if (image->depth == 1)
+			qxl_fb_image->visual = FB_VISUAL_MONO10;
+		else
+			qxl_fb_image->visual = FB_VISUAL_DIRECTCOLOR;
+	}
+	if (image) {
+		memcpy(&qxl_fb_image->fb_image, image,
+		       sizeof(qxl_fb_image->fb_image));
+	}
+}
+
+static void qxl_fb_dirty_flush(struct fb_info *info)
+{
+	struct qxl_fbdev *qfbdev = info->par;
+	struct qxl_device *qdev = qfbdev->qdev;
+	struct qxl_fb_image qxl_fb_image;
+	struct fb_image *image = &qxl_fb_image.fb_image;
+	u32 x1, x2, y1, y2;
+
+	/* TODO: hard coding 32 bpp */
+	int stride = qfbdev->qfb.base.pitches[0] * 4;
+
+	x1 = qfbdev->dirty.x1;
+	x2 = qfbdev->dirty.x2;
+	y1 = qfbdev->dirty.y1;
+	y2 = qfbdev->dirty.y2;
+	/*
+	 * we are using a shadow draw buffer, at qdev->surface0_shadow
+	 */
+	qxl_io_log(qdev, "dirty x[%d, %d], y[%d, %d]", x1, x2, y1, y2);
+	image->dx = x1;
+	image->dy = y1;
+	image->width = x2 - x1;
+	image->height = y2 - y1;
+	image->fg_color = 0xffffffff; /* unused, just to avoid uninitialized
+					 warnings */
+	image->bg_color = 0;
+	image->depth = 32;	     /* TODO: take from somewhere? */
+	image->cmap.start = 0;
+	image->cmap.len = 0;
+	image->cmap.red = NULL;
+	image->cmap.green = NULL;
+	image->cmap.blue = NULL;
+	image->cmap.transp = NULL;
+	image->data = qfbdev->shadow + (x1 * 4) + (stride * y1);
+
+	qxl_fb_image_init(&qxl_fb_image, qdev, info, NULL);
+	qxl_draw_opaque_fb(&qxl_fb_image, stride);
+	qfbdev->dirty.x1 = 0;
+	qfbdev->dirty.x2 = 0;
+	qfbdev->dirty.y1 = 0;
+	qfbdev->dirty.y2 = 0;
+}
+
+static void qxl_deferred_io(struct fb_info *info,
+			    struct list_head *pagelist)
+{
+	struct qxl_fbdev *qfbdev = info->par;
+	unsigned long start, end, min, max;
+	struct page *page;
+	int y1, y2;
+
+	min = ULONG_MAX;
+	max = 0;
+	list_for_each_entry(page, pagelist, lru) {
+		start = page->index << PAGE_SHIFT;
+		end = start + PAGE_SIZE - 1;
+		min = min(min, start);
+		max = max(max, end);
+	}
+
+	if (min < max) {
+		y1 = min / info->fix.line_length;
+		y2 = (max / info->fix.line_length) + 1;
+
+		/* TODO: add spin lock? */
+		/* spin_lock_irqsave(&qfbdev->dirty.lock, flags); */
+		qfbdev->dirty.x1 = 0;
+		qfbdev->dirty.y1 = y1;
+		qfbdev->dirty.x2 = info->var.xres;
+		qfbdev->dirty.y2 = y2;
+		/* spin_unlock_irqrestore(&qfbdev->dirty.lock, flags); */
+	}
+
+	qxl_fb_dirty_flush(info);
+};
+
+
+static struct fb_deferred_io qxl_defio = {
+	.delay		= QXL_DIRTY_DELAY,
+	.deferred_io	= qxl_deferred_io,
+};
+
+static void qxl_fb_fillrect(struct fb_info *info,
+			    const struct fb_fillrect *fb_rect)
+{
+	struct qxl_fbdev *qfbdev = info->par;
+	struct qxl_device *qdev = qfbdev->qdev;
+	struct qxl_rect rect;
+	uint32_t color;
+	int x = fb_rect->dx;
+	int y = fb_rect->dy;
+	int width = fb_rect->width;
+	int height = fb_rect->height;
+	uint16_t rop;
+	struct qxl_draw_fill qxl_draw_fill_rec;
+
+	if (info->fix.visual == FB_VISUAL_TRUECOLOR ||
+	    info->fix.visual == FB_VISUAL_DIRECTCOLOR)
+		color = ((u32 *) (info->pseudo_palette))[fb_rect->color];
+	else
+		color = fb_rect->color;
+	rect.left = x;
+	rect.right = x + width;
+	rect.top = y;
+	rect.bottom = y + height;
+	switch (fb_rect->rop) {
+	case ROP_XOR:
+		rop = SPICE_ROPD_OP_XOR;
+		break;
+	case ROP_COPY:
+		rop = SPICE_ROPD_OP_PUT;
+		break;
+	default:
+		pr_err("qxl_fb_fillrect(): unknown rop, "
+		       "defaulting to SPICE_ROPD_OP_PUT\n");
+		rop = SPICE_ROPD_OP_PUT;
+	}
+	qxl_draw_fill_rec.qdev = qdev;
+	qxl_draw_fill_rec.rect = rect;
+	qxl_draw_fill_rec.color = color;
+	qxl_draw_fill_rec.rop = rop;
+	if (!drm_can_sleep()) {
+		qxl_io_log(qdev,
+			"%s: TODO use RCU, mysterious locks with spin_lock\n",
+			__func__);
+		return;
+	}
+	qxl_draw_fill(&qxl_draw_fill_rec);
+}
+
+static void qxl_fb_copyarea(struct fb_info *info,
+			    const struct fb_copyarea *region)
+{
+	struct qxl_fbdev *qfbdev = info->par;
+
+	qxl_draw_copyarea(qfbdev->qdev,
+			  region->width, region->height,
+			  region->sx, region->sy,
+			  region->dx, region->dy);
+}
+
+static void qxl_fb_imageblit_safe(struct qxl_fb_image *qxl_fb_image)
+{
+	qxl_draw_opaque_fb(qxl_fb_image, 0);
+}
+
+static void qxl_fb_imageblit(struct fb_info *info,
+			     const struct fb_image *image)
+{
+	struct qxl_fbdev *qfbdev = info->par;
+	struct qxl_device *qdev = qfbdev->qdev;
+	struct qxl_fb_image qxl_fb_image;
+
+	if (!drm_can_sleep()) {
+		/* we cannot do any ttm_bo allocation since that will fail on
+		 * ioremap_wc..__get_vm_area_node, so queue the work item
+		 * instead This can happen from printk inside an interrupt
+		 * context, i.e.: smp_apic_timer_interrupt..check_cpu_stall */
+		qxl_io_log(qdev,
+			"%s: TODO use RCU, mysterious locks with spin_lock\n",
+			   __func__);
+		return;
+	}
+
+	/* ensure proper order of rendering operations - TODO: must do this
+	 * for everything. */
+	qxl_fb_image_init(&qxl_fb_image, qfbdev->qdev, info, image);
+	qxl_fb_imageblit_safe(&qxl_fb_image);
+}
+
+int qxl_fb_init(struct qxl_device *qdev)
+{
+	return 0;
+}
+
+static struct fb_ops qxlfb_ops = {
+	.owner = THIS_MODULE,
+	.fb_check_var = drm_fb_helper_check_var,
+	.fb_set_par = drm_fb_helper_set_par, /* TODO: copy vmwgfx */
+	.fb_fillrect = qxl_fb_fillrect,
+	.fb_copyarea = qxl_fb_copyarea,
+	.fb_imageblit = qxl_fb_imageblit,
+	.fb_pan_display = drm_fb_helper_pan_display,
+	.fb_blank = drm_fb_helper_blank,
+	.fb_setcmap = drm_fb_helper_setcmap,
+	.fb_debug_enter = drm_fb_helper_debug_enter,
+	.fb_debug_leave = drm_fb_helper_debug_leave,
+};
+
+static void qxlfb_destroy_pinned_object(struct drm_gem_object *gobj)
+{
+	struct qxl_bo *qbo = gem_to_qxl_bo(gobj);
+	int ret;
+
+	ret = qxl_bo_reserve(qbo, false);
+	if (likely(ret == 0)) {
+		qxl_bo_kunmap(qbo);
+		qxl_bo_unpin(qbo);
+		qxl_bo_unreserve(qbo);
+	}
+	drm_gem_object_unreference_unlocked(gobj);
+}
+
+int qxl_get_handle_for_primary_fb(struct qxl_device *qdev,
+				  struct drm_file *file_priv,
+				  uint32_t *handle)
+{
+	int r;
+	struct drm_gem_object *gobj = qdev->fbdev_qfb->obj;
+
+	BUG_ON(!gobj);
+	/* drm_get_handle_create adds a reference - good */
+	r = drm_gem_handle_create(file_priv, gobj, handle);
+	if (r)
+		return r;
+	return 0;
+}
+
+static int qxlfb_create_pinned_object(struct qxl_fbdev *qfbdev,
+				      struct drm_mode_fb_cmd2 *mode_cmd,
+				      struct drm_gem_object **gobj_p)
+{
+	struct qxl_device *qdev = qfbdev->qdev;
+	struct drm_gem_object *gobj = NULL;
+	struct qxl_bo *qbo = NULL;
+	int ret;
+	int aligned_size, size;
+	int height = mode_cmd->height;
+	int bpp;
+	int depth;
+
+	drm_fb_get_bpp_depth(mode_cmd->pixel_format, &bpp, &depth);
+
+	size = mode_cmd->pitches[0] * height;
+	aligned_size = ALIGN(size, PAGE_SIZE);
+	/* TODO: unallocate and reallocate surface0 for real. Hack to just
+	 * have a large enough surface0 for 1024x768 Xorg 32bpp mode */
+	ret = qxl_gem_object_create(qdev, aligned_size, 0,
+				    QXL_GEM_DOMAIN_SURFACE,
+				    false, /* is discardable */
+				    false, /* is kernel (false means device) */
+				    NULL,
+				    &gobj);
+	if (ret) {
+		pr_err("failed to allocate framebuffer (%d)\n",
+		       aligned_size);
+		return -ENOMEM;
+	}
+	qbo = gem_to_qxl_bo(gobj);
+
+	qbo->surf.width = mode_cmd->width;
+	qbo->surf.height = mode_cmd->height;
+	qbo->surf.stride = mode_cmd->pitches[0];
+	qbo->surf.format = SPICE_SURFACE_FMT_32_xRGB;
+	ret = qxl_bo_reserve(qbo, false);
+	if (unlikely(ret != 0))
+		goto out_unref;
+	ret = qxl_bo_pin(qbo, QXL_GEM_DOMAIN_SURFACE, NULL);
+	if (ret) {
+		qxl_bo_unreserve(qbo);
+		goto out_unref;
+	}
+	ret = qxl_bo_kmap(qbo, NULL);
+	qxl_bo_unreserve(qbo); /* unreserve, will be mmaped */
+	if (ret)
+		goto out_unref;
+
+	*gobj_p = gobj;
+	return 0;
+out_unref:
+	qxlfb_destroy_pinned_object(gobj);
+	*gobj_p = NULL;
+	return ret;
+}
+
+static int qxlfb_create(struct qxl_fbdev *qfbdev,
+			struct drm_fb_helper_surface_size *sizes)
+{
+	struct qxl_device *qdev = qfbdev->qdev;
+	struct fb_info *info;
+	struct drm_framebuffer *fb = NULL;
+	struct drm_mode_fb_cmd2 mode_cmd;
+	struct drm_gem_object *gobj = NULL;
+	struct qxl_bo *qbo = NULL;
+	struct device *device = &qdev->pdev->dev;
+	int ret;
+	int size;
+	int bpp = sizes->surface_bpp;
+	int depth = sizes->surface_depth;
+	void *shadow;
+
+	mode_cmd.width = sizes->surface_width;
+	mode_cmd.height = sizes->surface_height;
+
+	mode_cmd.pitches[0] = ALIGN(mode_cmd.width * ((bpp + 1) / 8), 64);
+	mode_cmd.pixel_format = drm_mode_legacy_fb_format(bpp, depth);
+
+	ret = qxlfb_create_pinned_object(qfbdev, &mode_cmd, &gobj);
+	qbo = gem_to_qxl_bo(gobj);
+	QXL_INFO(qdev, "%s: %dx%d %d\n", __func__, mode_cmd.width,
+		 mode_cmd.height, mode_cmd.pitches[0]);
+
+	shadow = vmalloc(mode_cmd.pitches[0] * mode_cmd.height);
+	/* TODO: what's the usual response to memory allocation errors? */
+	BUG_ON(!shadow);
+	QXL_INFO(qdev,
+	"surface0 at gpu offset %lld, mmap_offset %lld (virt %p, shadow %p)\n",
+		 qxl_bo_gpu_offset(qbo),
+		 qxl_bo_mmap_offset(qbo),
+		 qbo->kptr,
+		 shadow);
+	size = mode_cmd.pitches[0] * mode_cmd.height;
+
+	info = framebuffer_alloc(0, device);
+	if (info == NULL) {
+		ret = -ENOMEM;
+		goto out_unref;
+	}
+
+	info->par = qfbdev;
+
+	qxl_framebuffer_init(qdev->ddev, &qfbdev->qfb, &mode_cmd, gobj);
+
+	fb = &qfbdev->qfb.base;
+
+	/* setup helper with fb data */
+	qfbdev->helper.fb = fb;
+	qfbdev->helper.fbdev = info;
+	qfbdev->shadow = shadow;
+	strcpy(info->fix.id, "qxldrmfb");
+
+	drm_fb_helper_fill_fix(info, fb->pitches[0], fb->depth);
+
+	info->flags = FBINFO_DEFAULT | FBINFO_HWACCEL_COPYAREA | FBINFO_HWACCEL_FILLRECT;
+	info->fbops = &qxlfb_ops;
+
+	/*
+	 * TODO: using gobj->size in various places in this function. Not sure
+	 * what the difference between the different sizes is.
+	 */
+	info->fix.smem_start = qdev->vram_base; /* TODO - correct? */
+	info->fix.smem_len = gobj->size;
+	info->screen_base = qfbdev->shadow;
+	info->screen_size = gobj->size;
+
+	drm_fb_helper_fill_var(info, &qfbdev->helper, sizes->fb_width,
+			       sizes->fb_height);
+
+	/* setup aperture base/size for vesafb takeover */
+	info->apertures = alloc_apertures(1);
+	if (!info->apertures) {
+		ret = -ENOMEM;
+		goto out_unref;
+	}
+	info->apertures->ranges[0].base = qdev->ddev->mode_config.fb_base;
+	info->apertures->ranges[0].size = qdev->vram_size;
+
+	info->fix.mmio_start = 0;
+	info->fix.mmio_len = 0;
+
+	if (info->screen_base == NULL) {
+		ret = -ENOSPC;
+		goto out_unref;
+	}
+
+	ret = fb_alloc_cmap(&info->cmap, 256, 0);
+	if (ret) {
+		ret = -ENOMEM;
+		goto out_unref;
+	}
+
+	info->fbdefio = &qxl_defio;
+	fb_deferred_io_init(info);
+
+	qdev->fbdev_info = info;
+	qdev->fbdev_qfb = &qfbdev->qfb;
+	DRM_INFO("fb mappable at 0x%lX, size %lu\n",  info->fix.smem_start, (unsigned long)info->screen_size);
+	DRM_INFO("fb: depth %d, pitch %d, width %d, height %d\n", fb->depth, fb->pitches[0], fb->width, fb->height);
+	return 0;
+
+out_unref:
+	if (qbo) {
+		ret = qxl_bo_reserve(qbo, false);
+		if (likely(ret == 0)) {
+			qxl_bo_kunmap(qbo);
+			qxl_bo_unpin(qbo);
+			qxl_bo_unreserve(qbo);
+		}
+	}
+	if (fb && ret) {
+		drm_gem_object_unreference(gobj);
+		drm_framebuffer_cleanup(fb);
+		kfree(fb);
+	}
+	drm_gem_object_unreference(gobj);
+	return ret;
+}
+
+static int qxl_fb_find_or_create_single(
+		struct drm_fb_helper *helper,
+		struct drm_fb_helper_surface_size *sizes)
+{
+	struct qxl_fbdev *qfbdev = (struct qxl_fbdev *)helper;
+	int new_fb = 0;
+	int ret;
+
+	if (!helper->fb) {
+		ret = qxlfb_create(qfbdev, sizes);
+		if (ret)
+			return ret;
+		new_fb = 1;
+	}
+	return new_fb;
+}
+
+static int qxl_fbdev_destroy(struct drm_device *dev, struct qxl_fbdev *qfbdev)
+{
+	struct fb_info *info;
+	struct qxl_framebuffer *qfb = &qfbdev->qfb;
+
+	if (qfbdev->helper.fbdev) {
+		info = qfbdev->helper.fbdev;
+
+		unregister_framebuffer(info);
+		framebuffer_release(info);
+	}
+	if (qfb->obj) {
+		qxlfb_destroy_pinned_object(qfb->obj);
+		qfb->obj = NULL;
+	}
+	drm_fb_helper_fini(&qfbdev->helper);
+	vfree(qfbdev->shadow);
+	drm_framebuffer_cleanup(&qfb->base);
+
+	return 0;
+}
+
+static struct drm_fb_helper_funcs qxl_fb_helper_funcs = {
+	/* TODO
+	.gamma_set = qxl_crtc_fb_gamma_set,
+	.gamma_get = qxl_crtc_fb_gamma_get,
+	*/
+	.fb_probe = qxl_fb_find_or_create_single,
+};
+
+int qxl_fbdev_init(struct qxl_device *qdev)
+{
+	struct qxl_fbdev *qfbdev;
+	int bpp_sel = 32; /* TODO: parameter from somewhere? */
+	int ret;
+
+	qfbdev = kzalloc(sizeof(struct qxl_fbdev), GFP_KERNEL);
+	if (!qfbdev)
+		return -ENOMEM;
+
+	qfbdev->qdev = qdev;
+	qdev->mode_info.qfbdev = qfbdev;
+	qfbdev->helper.funcs = &qxl_fb_helper_funcs;
+
+	ret = drm_fb_helper_init(qdev->ddev, &qfbdev->helper,
+				 1 /* num_crtc - QXL supports just 1 */,
+				 QXLFB_CONN_LIMIT);
+	if (ret) {
+		kfree(qfbdev);
+		return ret;
+	}
+
+	drm_fb_helper_single_add_all_connectors(&qfbdev->helper);
+	drm_fb_helper_initial_config(&qfbdev->helper, bpp_sel);
+	return 0;
+}
+
+void qxl_fbdev_fini(struct qxl_device *qdev)
+{
+	if (!qdev->mode_info.qfbdev)
+		return;
+
+	qxl_fbdev_destroy(qdev->ddev, qdev->mode_info.qfbdev);
+	kfree(qdev->mode_info.qfbdev);
+	qdev->mode_info.qfbdev = NULL;
+}
+
+
diff --git a/drivers/gpu/drm/qxl/qxl_fence.c b/drivers/gpu/drm/qxl/qxl_fence.c
new file mode 100644
index 000000000000..63c6715ad385
--- /dev/null
+++ b/drivers/gpu/drm/qxl/qxl_fence.c
@@ -0,0 +1,97 @@
+/*
+ * Copyright 2013 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Dave Airlie
+ *          Alon Levy
+ */
+
+
+#include "qxl_drv.h"
+
+/* QXL fencing-
+
+   When we submit operations to the GPU we pass a release reference to the GPU
+   with them, the release reference is then added to the release ring when
+   the GPU is finished with that particular operation and has removed it from
+   its tree.
+
+   So we have can have multiple outstanding non linear fences per object.
+
+   From a TTM POV we only care if the object has any outstanding releases on
+   it.
+
+   we wait until all outstanding releases are processeed.
+
+   sync object is just a list of release ids that represent that fence on
+   that buffer.
+
+   we just add new releases onto the sync object attached to the object.
+
+   This currently uses a radix tree to store the list of release ids.
+
+   For some reason every so often qxl hw fails to release, things go wrong.
+*/
+
+
+int qxl_fence_add_release(struct qxl_fence *qfence, uint32_t rel_id)
+{
+	struct qxl_bo *bo = container_of(qfence, struct qxl_bo, fence);
+
+	spin_lock(&bo->tbo.bdev->fence_lock);
+	radix_tree_insert(&qfence->tree, rel_id, qfence);
+	qfence->num_active_releases++;
+	spin_unlock(&bo->tbo.bdev->fence_lock);
+	return 0;
+}
+
+int qxl_fence_remove_release(struct qxl_fence *qfence, uint32_t rel_id)
+{
+	void *ret;
+	int retval = 0;
+	struct qxl_bo *bo = container_of(qfence, struct qxl_bo, fence);
+
+	spin_lock(&bo->tbo.bdev->fence_lock);
+
+	ret = radix_tree_delete(&qfence->tree, rel_id);
+	if (ret == qfence)
+		qfence->num_active_releases--;
+	else {
+		DRM_DEBUG("didn't find fence in radix tree for %d\n", rel_id);
+		retval = -ENOENT;
+	}
+	spin_unlock(&bo->tbo.bdev->fence_lock);
+	return retval;
+}
+
+
+int qxl_fence_init(struct qxl_device *qdev, struct qxl_fence *qfence)
+{
+	qfence->qdev = qdev;
+	qfence->num_active_releases = 0;
+	INIT_RADIX_TREE(&qfence->tree, GFP_ATOMIC);
+	return 0;
+}
+
+void qxl_fence_fini(struct qxl_fence *qfence)
+{
+	kfree(qfence->release_ids);
+	qfence->num_active_releases = 0;
+}
diff --git a/drivers/gpu/drm/qxl/qxl_gem.c b/drivers/gpu/drm/qxl/qxl_gem.c
new file mode 100644
index 000000000000..a235693aabba
--- /dev/null
+++ b/drivers/gpu/drm/qxl/qxl_gem.c
@@ -0,0 +1,149 @@
+/*
+ * Copyright 2013 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Dave Airlie
+ *          Alon Levy
+ */
+
+#include "drmP.h"
+#include "drm/drm.h"
+#include "qxl_drv.h"
+#include "qxl_object.h"
+
+int qxl_gem_object_init(struct drm_gem_object *obj)
+{
+	/* we do nothings here */
+	return 0;
+}
+
+void qxl_gem_object_free(struct drm_gem_object *gobj)
+{
+	struct qxl_bo *qobj = gem_to_qxl_bo(gobj);
+
+	if (qobj)
+		qxl_bo_unref(&qobj);
+}
+
+int qxl_gem_object_create(struct qxl_device *qdev, int size,
+			  int alignment, int initial_domain,
+			  bool discardable, bool kernel,
+			  struct qxl_surface *surf,
+			  struct drm_gem_object **obj)
+{
+	struct qxl_bo *qbo;
+	int r;
+
+	*obj = NULL;
+	/* At least align on page size */
+	if (alignment < PAGE_SIZE)
+		alignment = PAGE_SIZE;
+	r = qxl_bo_create(qdev, size, kernel, initial_domain, surf, &qbo);
+	if (r) {
+		if (r != -ERESTARTSYS)
+			DRM_ERROR(
+			"Failed to allocate GEM object (%d, %d, %u, %d)\n",
+				  size, initial_domain, alignment, r);
+		return r;
+	}
+	*obj = &qbo->gem_base;
+
+	mutex_lock(&qdev->gem.mutex);
+	list_add_tail(&qbo->list, &qdev->gem.objects);
+	mutex_unlock(&qdev->gem.mutex);
+
+	return 0;
+}
+
+int qxl_gem_object_create_with_handle(struct qxl_device *qdev,
+				      struct drm_file *file_priv,
+				      u32 domain,
+				      size_t size,
+				      struct qxl_surface *surf,
+				      struct qxl_bo **qobj,
+				      uint32_t *handle)
+{
+	struct drm_gem_object *gobj;
+	int r;
+
+	BUG_ON(!qobj);
+	BUG_ON(!handle);
+
+	r = qxl_gem_object_create(qdev, size, 0,
+				  domain,
+				  false, false, surf,
+				  &gobj);
+	if (r)
+		return -ENOMEM;
+	r = drm_gem_handle_create(file_priv, gobj, handle);
+	if (r)
+		return r;
+	/* drop reference from allocate - handle holds it now */
+	*qobj = gem_to_qxl_bo(gobj);
+	drm_gem_object_unreference_unlocked(gobj);
+	return 0;
+}
+
+int qxl_gem_object_pin(struct drm_gem_object *obj, uint32_t pin_domain,
+			  uint64_t *gpu_addr)
+{
+	struct qxl_bo *qobj = obj->driver_private;
+	int r;
+
+	r = qxl_bo_reserve(qobj, false);
+	if (unlikely(r != 0))
+		return r;
+	r = qxl_bo_pin(qobj, pin_domain, gpu_addr);
+	qxl_bo_unreserve(qobj);
+	return r;
+}
+
+void qxl_gem_object_unpin(struct drm_gem_object *obj)
+{
+	struct qxl_bo *qobj = obj->driver_private;
+	int r;
+
+	r = qxl_bo_reserve(qobj, false);
+	if (likely(r == 0)) {
+		qxl_bo_unpin(qobj);
+		qxl_bo_unreserve(qobj);
+	}
+}
+
+int qxl_gem_object_open(struct drm_gem_object *obj, struct drm_file *file_priv)
+{
+	return 0;
+}
+
+void qxl_gem_object_close(struct drm_gem_object *obj,
+			  struct drm_file *file_priv)
+{
+}
+
+int qxl_gem_init(struct qxl_device *qdev)
+{
+	INIT_LIST_HEAD(&qdev->gem.objects);
+	return 0;
+}
+
+void qxl_gem_fini(struct qxl_device *qdev)
+{
+	qxl_bo_force_delete(qdev);
+}
diff --git a/drivers/gpu/drm/qxl/qxl_image.c b/drivers/gpu/drm/qxl/qxl_image.c
new file mode 100644
index 000000000000..cf856206996b
--- /dev/null
+++ b/drivers/gpu/drm/qxl/qxl_image.c
@@ -0,0 +1,176 @@
+/*
+ * Copyright 2013 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Dave Airlie
+ *          Alon Levy
+ */
+
+#include <linux/gfp.h>
+#include <linux/slab.h>
+
+#include "qxl_drv.h"
+#include "qxl_object.h"
+
+static int
+qxl_image_create_helper(struct qxl_device *qdev,
+			struct qxl_release *release,
+			struct qxl_bo **image_bo,
+			const uint8_t *data,
+			int width, int height,
+			int depth, unsigned int hash,
+			int stride)
+{
+	struct qxl_image *image;
+	struct qxl_data_chunk *chunk;
+	int i;
+	int chunk_stride;
+	int linesize = width * depth / 8;
+	struct qxl_bo *chunk_bo;
+	int ret;
+	void *ptr;
+	/* Chunk */
+	/* FIXME: Check integer overflow */
+	/* TODO: variable number of chunks */
+	chunk_stride = stride; /* TODO: should use linesize, but it renders
+				  wrong (check the bitmaps are sent correctly
+				  first) */
+	ret = qxl_alloc_bo_reserved(qdev, sizeof(*chunk) + height * chunk_stride,
+				    &chunk_bo);
+	
+	ptr = qxl_bo_kmap_atomic_page(qdev, chunk_bo, 0);
+	chunk = ptr;
+	chunk->data_size = height * chunk_stride;
+	chunk->prev_chunk = 0;
+	chunk->next_chunk = 0;
+	qxl_bo_kunmap_atomic_page(qdev, chunk_bo, ptr);
+
+	{
+		void *k_data, *i_data;
+		int remain;
+		int page;
+		int size;
+		if (stride == linesize && chunk_stride == stride) {
+			remain = linesize * height;
+			page = 0;
+			i_data = (void *)data;
+
+			while (remain > 0) {
+				ptr = qxl_bo_kmap_atomic_page(qdev, chunk_bo, page << PAGE_SHIFT);
+
+				if (page == 0) {
+					chunk = ptr;
+					k_data = chunk->data;
+					size = PAGE_SIZE - offsetof(struct qxl_data_chunk, data);
+				} else {
+					k_data = ptr;
+					size = PAGE_SIZE;
+				}
+				size = min(size, remain);
+
+				memcpy(k_data, i_data, size);
+
+				qxl_bo_kunmap_atomic_page(qdev, chunk_bo, ptr);
+				i_data += size;
+				remain -= size;
+				page++;
+			}
+		} else {
+			unsigned page_base, page_offset, out_offset;
+			for (i = 0 ; i < height ; ++i) {
+				i_data = (void *)data + i * stride;
+				remain = linesize;
+				out_offset = offsetof(struct qxl_data_chunk, data) + i * chunk_stride;
+
+				while (remain > 0) {
+					page_base = out_offset & PAGE_MASK;
+					page_offset = offset_in_page(out_offset);
+					
+					size = min((int)(PAGE_SIZE - page_offset), remain);
+
+					ptr = qxl_bo_kmap_atomic_page(qdev, chunk_bo, page_base);
+					k_data = ptr + page_offset;
+					memcpy(k_data, i_data, size);
+					qxl_bo_kunmap_atomic_page(qdev, chunk_bo, ptr);
+					remain -= size;
+					i_data += size;
+					out_offset += size;
+				}
+			}
+		}
+	}
+
+
+	qxl_bo_kunmap(chunk_bo);
+
+	/* Image */
+	ret = qxl_alloc_bo_reserved(qdev, sizeof(*image), image_bo);
+
+	ptr = qxl_bo_kmap_atomic_page(qdev, *image_bo, 0);
+	image = ptr;
+
+	image->descriptor.id = 0;
+	image->descriptor.type = SPICE_IMAGE_TYPE_BITMAP;
+
+	image->descriptor.flags = 0;
+	image->descriptor.width = width;
+	image->descriptor.height = height;
+
+	switch (depth) {
+	case 1:
+		/* TODO: BE? check by arch? */
+		image->u.bitmap.format = SPICE_BITMAP_FMT_1BIT_BE;
+		break;
+	case 24:
+		image->u.bitmap.format = SPICE_BITMAP_FMT_24BIT;
+		break;
+	case 32:
+		image->u.bitmap.format = SPICE_BITMAP_FMT_32BIT;
+		break;
+	default:
+		DRM_ERROR("unsupported image bit depth\n");
+		return -EINVAL; /* TODO: cleanup */
+	}
+	image->u.bitmap.flags = QXL_BITMAP_TOP_DOWN;
+	image->u.bitmap.x = width;
+	image->u.bitmap.y = height;
+	image->u.bitmap.stride = chunk_stride;
+	image->u.bitmap.palette = 0;
+	image->u.bitmap.data = qxl_bo_physical_address(qdev, chunk_bo, 0);
+	qxl_release_add_res(qdev, release, chunk_bo);
+	qxl_bo_unreserve(chunk_bo);
+	qxl_bo_unref(&chunk_bo);
+
+	qxl_bo_kunmap_atomic_page(qdev, *image_bo, ptr);
+
+	return 0;
+}
+
+int qxl_image_create(struct qxl_device *qdev,
+		     struct qxl_release *release,
+		     struct qxl_bo **image_bo,
+		     const uint8_t *data,
+		     int x, int y, int width, int height,
+		     int depth, int stride)
+{
+	data += y * stride + x * (depth / 8);
+	return qxl_image_create_helper(qdev, release, image_bo, data,
+				       width, height, depth, 0, stride);
+}
diff --git a/drivers/gpu/drm/qxl/qxl_ioctl.c b/drivers/gpu/drm/qxl/qxl_ioctl.c
new file mode 100644
index 000000000000..04b64f9cbfdb
--- /dev/null
+++ b/drivers/gpu/drm/qxl/qxl_ioctl.c
@@ -0,0 +1,411 @@
+/*
+ * Copyright 2013 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Dave Airlie
+ *          Alon Levy
+ */
+
+#include "qxl_drv.h"
+#include "qxl_object.h"
+
+/*
+ * TODO: allocating a new gem(in qxl_bo) for each request.
+ * This is wasteful since bo's are page aligned.
+ */
+static int qxl_alloc_ioctl(struct drm_device *dev, void *data,
+			   struct drm_file *file_priv)
+{
+	struct qxl_device *qdev = dev->dev_private;
+	struct drm_qxl_alloc *qxl_alloc = data;
+	int ret;
+	struct qxl_bo *qobj;
+	uint32_t handle;
+	u32 domain = QXL_GEM_DOMAIN_VRAM;
+
+	if (qxl_alloc->size == 0) {
+		DRM_ERROR("invalid size %d\n", qxl_alloc->size);
+		return -EINVAL;
+	}
+	ret = qxl_gem_object_create_with_handle(qdev, file_priv,
+						domain,
+						qxl_alloc->size,
+						NULL,
+						&qobj, &handle);
+	if (ret) {
+		DRM_ERROR("%s: failed to create gem ret=%d\n",
+			  __func__, ret);
+		return -ENOMEM;
+	}
+	qxl_alloc->handle = handle;
+	return 0;
+}
+
+static int qxl_map_ioctl(struct drm_device *dev, void *data,
+			 struct drm_file *file_priv)
+{
+	struct qxl_device *qdev = dev->dev_private;
+	struct drm_qxl_map *qxl_map = data;
+
+	return qxl_mode_dumb_mmap(file_priv, qdev->ddev, qxl_map->handle,
+				  &qxl_map->offset);
+}
+
+/*
+ * dst must be validated, i.e. whole bo on vram/surfacesram (right now all bo's
+ * are on vram).
+ * *(dst + dst_off) = qxl_bo_physical_address(src, src_off)
+ */
+static void
+apply_reloc(struct qxl_device *qdev, struct qxl_bo *dst, uint64_t dst_off,
+	    struct qxl_bo *src, uint64_t src_off)
+{
+	void *reloc_page;
+
+	reloc_page = qxl_bo_kmap_atomic_page(qdev, dst, dst_off & PAGE_MASK);
+	*(uint64_t *)(reloc_page + (dst_off & ~PAGE_MASK)) = qxl_bo_physical_address(qdev,
+								     src, src_off);
+	qxl_bo_kunmap_atomic_page(qdev, dst, reloc_page);
+}
+
+static void
+apply_surf_reloc(struct qxl_device *qdev, struct qxl_bo *dst, uint64_t dst_off,
+		 struct qxl_bo *src)
+{
+	uint32_t id = 0;
+	void *reloc_page;
+
+	if (src && !src->is_primary)
+		id = src->surface_id;
+
+	reloc_page = qxl_bo_kmap_atomic_page(qdev, dst, dst_off & PAGE_MASK);
+	*(uint32_t *)(reloc_page + (dst_off & ~PAGE_MASK)) = id;
+	qxl_bo_kunmap_atomic_page(qdev, dst, reloc_page);
+}
+
+/* return holding the reference to this object */
+static struct qxl_bo *qxlhw_handle_to_bo(struct qxl_device *qdev,
+					 struct drm_file *file_priv, uint64_t handle,
+					 struct qxl_reloc_list *reloc_list)
+{
+	struct drm_gem_object *gobj;
+	struct qxl_bo *qobj;
+	int ret;
+
+	gobj = drm_gem_object_lookup(qdev->ddev, file_priv, handle);
+	if (!gobj) {
+		DRM_ERROR("bad bo handle %lld\n", handle);
+		return NULL;
+	}
+	qobj = gem_to_qxl_bo(gobj);
+
+	ret = qxl_bo_list_add(reloc_list, qobj);
+	if (ret)
+		return NULL;
+
+	return qobj;
+}
+
+/*
+ * Usage of execbuffer:
+ * Relocations need to take into account the full QXLDrawable size.
+ * However, the command as passed from user space must *not* contain the initial
+ * QXLReleaseInfo struct (first XXX bytes)
+ */
+static int qxl_execbuffer_ioctl(struct drm_device *dev, void *data,
+				struct drm_file *file_priv)
+{
+	struct qxl_device *qdev = dev->dev_private;
+	struct drm_qxl_execbuffer *execbuffer = data;
+	struct drm_qxl_command user_cmd;
+	int cmd_num;
+	struct qxl_bo *reloc_src_bo;
+	struct qxl_bo *reloc_dst_bo;
+	struct drm_qxl_reloc reloc;
+	void *fb_cmd;
+	int i, ret;
+	struct qxl_reloc_list reloc_list;
+	int unwritten;
+	uint32_t reloc_dst_offset;
+	INIT_LIST_HEAD(&reloc_list.bos);
+
+	for (cmd_num = 0; cmd_num < execbuffer->commands_num; ++cmd_num) {
+		struct qxl_release *release;
+		struct qxl_bo *cmd_bo;
+		int release_type;
+		struct drm_qxl_command *commands =
+			(struct drm_qxl_command *)execbuffer->commands;
+
+		if (DRM_COPY_FROM_USER(&user_cmd, &commands[cmd_num],
+				       sizeof(user_cmd)))
+			return -EFAULT;
+		switch (user_cmd.type) {
+		case QXL_CMD_DRAW:
+			release_type = QXL_RELEASE_DRAWABLE;
+			break;
+		case QXL_CMD_SURFACE:
+		case QXL_CMD_CURSOR:
+		default:
+			DRM_DEBUG("Only draw commands in execbuffers\n");
+			return -EINVAL;
+			break;
+		}
+
+		if (user_cmd.command_size > PAGE_SIZE - sizeof(union qxl_release_info))
+			return -EINVAL;
+
+		ret = qxl_alloc_release_reserved(qdev,
+						 sizeof(union qxl_release_info) +
+						 user_cmd.command_size,
+						 release_type,
+						 &release,
+						 &cmd_bo);
+		if (ret)
+			return ret;
+
+		/* TODO copy slow path code from i915 */
+		fb_cmd = qxl_bo_kmap_atomic_page(qdev, cmd_bo, (release->release_offset & PAGE_SIZE));
+		unwritten = __copy_from_user_inatomic_nocache(fb_cmd + sizeof(union qxl_release_info) + (release->release_offset & ~PAGE_SIZE), (void *)(unsigned long)user_cmd.command, user_cmd.command_size);
+		qxl_bo_kunmap_atomic_page(qdev, cmd_bo, fb_cmd);
+		if (unwritten) {
+			DRM_ERROR("got unwritten %d\n", unwritten);
+			qxl_release_unreserve(qdev, release);
+			qxl_release_free(qdev, release);
+			return -EFAULT;
+		}
+
+		for (i = 0 ; i < user_cmd.relocs_num; ++i) {
+			if (DRM_COPY_FROM_USER(&reloc,
+					       &((struct drm_qxl_reloc *)user_cmd.relocs)[i],
+					       sizeof(reloc))) {
+				qxl_bo_list_unreserve(&reloc_list, true);
+				qxl_release_unreserve(qdev, release);
+				qxl_release_free(qdev, release);
+				return -EFAULT;
+			}
+
+			/* add the bos to the list of bos to validate -
+			   need to validate first then process relocs? */
+			if (reloc.dst_handle) {
+				reloc_dst_bo = qxlhw_handle_to_bo(qdev, file_priv,
+								  reloc.dst_handle, &reloc_list);
+				if (!reloc_dst_bo) {
+					qxl_bo_list_unreserve(&reloc_list, true);
+					qxl_release_unreserve(qdev, release);
+					qxl_release_free(qdev, release);
+					return -EINVAL;
+				}
+				reloc_dst_offset = 0;
+			} else {
+				reloc_dst_bo = cmd_bo;
+				reloc_dst_offset = release->release_offset;
+			}
+
+			/* reserve and validate the reloc dst bo */
+			if (reloc.reloc_type == QXL_RELOC_TYPE_BO || reloc.src_handle > 0) {
+				reloc_src_bo =
+					qxlhw_handle_to_bo(qdev, file_priv,
+							   reloc.src_handle, &reloc_list);
+				if (!reloc_src_bo) {
+					if (reloc_dst_bo != cmd_bo)
+						drm_gem_object_unreference_unlocked(&reloc_dst_bo->gem_base);
+					qxl_bo_list_unreserve(&reloc_list, true);
+					qxl_release_unreserve(qdev, release);
+					qxl_release_free(qdev, release);
+					return -EINVAL;
+				}
+			} else
+				reloc_src_bo = NULL;
+			if (reloc.reloc_type == QXL_RELOC_TYPE_BO) {
+				apply_reloc(qdev, reloc_dst_bo, reloc_dst_offset + reloc.dst_offset,
+					    reloc_src_bo, reloc.src_offset);
+			} else if (reloc.reloc_type == QXL_RELOC_TYPE_SURF) {
+				apply_surf_reloc(qdev, reloc_dst_bo, reloc_dst_offset + reloc.dst_offset, reloc_src_bo);
+			} else {
+				DRM_ERROR("unknown reloc type %d\n", reloc.reloc_type);
+				return -EINVAL;
+			}
+
+			if (reloc_src_bo && reloc_src_bo != cmd_bo) {
+				qxl_release_add_res(qdev, release, reloc_src_bo);
+				drm_gem_object_unreference_unlocked(&reloc_src_bo->gem_base);
+			}
+
+			if (reloc_dst_bo != cmd_bo)
+				drm_gem_object_unreference_unlocked(&reloc_dst_bo->gem_base);
+		}
+		qxl_fence_releaseable(qdev, release);
+
+		ret = qxl_push_command_ring_release(qdev, release, user_cmd.type, true);
+		if (ret == -ERESTARTSYS) {
+			qxl_release_unreserve(qdev, release);
+			qxl_release_free(qdev, release);
+			qxl_bo_list_unreserve(&reloc_list, true);
+			return ret;
+		}
+		qxl_release_unreserve(qdev, release);
+	}
+	qxl_bo_list_unreserve(&reloc_list, 0);
+	return 0;
+}
+
+static int qxl_update_area_ioctl(struct drm_device *dev, void *data,
+				 struct drm_file *file)
+{
+	struct qxl_device *qdev = dev->dev_private;
+	struct drm_qxl_update_area *update_area = data;
+	struct qxl_rect area = {.left = update_area->left,
+				.top = update_area->top,
+				.right = update_area->right,
+				.bottom = update_area->bottom};
+	int ret;
+	struct drm_gem_object *gobj = NULL;
+	struct qxl_bo *qobj = NULL;
+
+	if (update_area->left >= update_area->right ||
+	    update_area->top >= update_area->bottom)
+		return -EINVAL;
+
+	gobj = drm_gem_object_lookup(dev, file, update_area->handle);
+	if (gobj == NULL)
+		return -ENOENT;
+
+	qobj = gem_to_qxl_bo(gobj);
+
+	ret = qxl_bo_reserve(qobj, false);
+	if (ret)
+		goto out;
+
+	if (!qobj->pin_count) {
+		ret = ttm_bo_validate(&qobj->tbo, &qobj->placement,
+				      true, false);
+		if (unlikely(ret))
+			goto out;
+	}
+
+	ret = qxl_bo_check_id(qdev, qobj);
+	if (ret)
+		goto out2;
+	if (!qobj->surface_id)
+		DRM_ERROR("got update area for surface with no id %d\n", update_area->handle);
+	ret = qxl_io_update_area(qdev, qobj, &area);
+
+out2:
+	qxl_bo_unreserve(qobj);
+
+out:
+	drm_gem_object_unreference_unlocked(gobj);
+	return ret;
+}
+
+static int qxl_getparam_ioctl(struct drm_device *dev, void *data,
+		       struct drm_file *file_priv)
+{
+	struct qxl_device *qdev = dev->dev_private;
+	struct drm_qxl_getparam *param = data;
+
+	switch (param->param) {
+	case QXL_PARAM_NUM_SURFACES:
+		param->value = qdev->rom->n_surfaces;
+		break;
+	case QXL_PARAM_MAX_RELOCS:
+		param->value = QXL_MAX_RES;
+		break;
+	default:
+		return -EINVAL;
+	}
+	return 0;
+}
+
+static int qxl_clientcap_ioctl(struct drm_device *dev, void *data,
+				  struct drm_file *file_priv)
+{
+	struct qxl_device *qdev = dev->dev_private;
+	struct drm_qxl_clientcap *param = data;
+	int byte, idx;
+
+	byte = param->index / 8;
+	idx = param->index % 8;
+
+	if (qdev->pdev->revision < 4)
+		return -ENOSYS;
+
+	if (byte >= 58)
+		return -ENOSYS;
+
+	if (qdev->rom->client_capabilities[byte] & (1 << idx))
+		return 0;
+	return -ENOSYS;
+}
+
+static int qxl_alloc_surf_ioctl(struct drm_device *dev, void *data,
+				struct drm_file *file)
+{
+	struct qxl_device *qdev = dev->dev_private;
+	struct drm_qxl_alloc_surf *param = data;
+	struct qxl_bo *qobj;
+	int handle;
+	int ret;
+	int size, actual_stride;
+	struct qxl_surface surf;
+
+	/* work out size allocate bo with handle */
+	actual_stride = param->stride < 0 ? -param->stride : param->stride;
+	size = actual_stride * param->height + actual_stride;
+
+	surf.format = param->format;
+	surf.width = param->width;
+	surf.height = param->height;
+	surf.stride = param->stride;
+	surf.data = 0;
+
+	ret = qxl_gem_object_create_with_handle(qdev, file,
+						QXL_GEM_DOMAIN_SURFACE,
+						size,
+						&surf,
+						&qobj, &handle);
+	if (ret) {
+		DRM_ERROR("%s: failed to create gem ret=%d\n",
+			  __func__, ret);
+		return -ENOMEM;
+	} else
+		param->handle = handle;
+	return ret;
+}
+
+struct drm_ioctl_desc qxl_ioctls[] = {
+	DRM_IOCTL_DEF_DRV(QXL_ALLOC, qxl_alloc_ioctl, DRM_AUTH|DRM_UNLOCKED),
+
+	DRM_IOCTL_DEF_DRV(QXL_MAP, qxl_map_ioctl, DRM_AUTH|DRM_UNLOCKED),
+
+	DRM_IOCTL_DEF_DRV(QXL_EXECBUFFER, qxl_execbuffer_ioctl,
+							DRM_AUTH|DRM_UNLOCKED),
+	DRM_IOCTL_DEF_DRV(QXL_UPDATE_AREA, qxl_update_area_ioctl,
+							DRM_AUTH|DRM_UNLOCKED),
+	DRM_IOCTL_DEF_DRV(QXL_GETPARAM, qxl_getparam_ioctl,
+							DRM_AUTH|DRM_UNLOCKED),
+	DRM_IOCTL_DEF_DRV(QXL_CLIENTCAP, qxl_clientcap_ioctl,
+							DRM_AUTH|DRM_UNLOCKED),
+
+	DRM_IOCTL_DEF_DRV(QXL_ALLOC_SURF, qxl_alloc_surf_ioctl,
+			  DRM_AUTH|DRM_UNLOCKED),
+};
+
+int qxl_max_ioctls = DRM_ARRAY_SIZE(qxl_ioctls);
diff --git a/drivers/gpu/drm/qxl/qxl_irq.c b/drivers/gpu/drm/qxl/qxl_irq.c
new file mode 100644
index 000000000000..21393dc4700a
--- /dev/null
+++ b/drivers/gpu/drm/qxl/qxl_irq.c
@@ -0,0 +1,97 @@
+/*
+ * Copyright 2013 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Dave Airlie
+ *          Alon Levy
+ */
+
+#include "qxl_drv.h"
+
+irqreturn_t qxl_irq_handler(DRM_IRQ_ARGS)
+{
+	struct drm_device *dev = (struct drm_device *) arg;
+	struct qxl_device *qdev = (struct qxl_device *)dev->dev_private;
+	uint32_t pending;
+
+	pending = xchg(&qdev->ram_header->int_pending, 0);
+
+	atomic_inc(&qdev->irq_received);
+
+	if (pending & QXL_INTERRUPT_DISPLAY) {
+		atomic_inc(&qdev->irq_received_display);
+		wake_up_all(&qdev->display_event);
+		qxl_queue_garbage_collect(qdev, false);
+	}
+	if (pending & QXL_INTERRUPT_CURSOR) {
+		atomic_inc(&qdev->irq_received_cursor);
+		wake_up_all(&qdev->cursor_event);
+	}
+	if (pending & QXL_INTERRUPT_IO_CMD) {
+		atomic_inc(&qdev->irq_received_io_cmd);
+		wake_up_all(&qdev->io_cmd_event);
+	}
+	if (pending & QXL_INTERRUPT_ERROR) {
+		/* TODO: log it, reset device (only way to exit this condition)
+		 * (do it a certain number of times, afterwards admit defeat,
+		 * to avoid endless loops).
+		 */
+		qdev->irq_received_error++;
+		qxl_io_log(qdev, "%s: driver is in bug mode.\n", __func__);
+	}
+	if (pending & QXL_INTERRUPT_CLIENT_MONITORS_CONFIG) {
+		qxl_io_log(qdev, "QXL_INTERRUPT_CLIENT_MONITORS_CONFIG\n");
+		schedule_work(&qdev->client_monitors_config_work);
+	}
+	qdev->ram_header->int_mask = QXL_INTERRUPT_MASK;
+	outb(0, qdev->io_base + QXL_IO_UPDATE_IRQ);
+	return IRQ_HANDLED;
+}
+
+static void qxl_client_monitors_config_work_func(struct work_struct *work)
+{
+	struct qxl_device *qdev = container_of(work, struct qxl_device,
+					       client_monitors_config_work);
+
+	qxl_display_read_client_monitors_config(qdev);
+}
+
+int qxl_irq_init(struct qxl_device *qdev)
+{
+	int ret;
+
+	init_waitqueue_head(&qdev->display_event);
+	init_waitqueue_head(&qdev->cursor_event);
+	init_waitqueue_head(&qdev->io_cmd_event);
+	INIT_WORK(&qdev->client_monitors_config_work,
+		  qxl_client_monitors_config_work_func);
+	atomic_set(&qdev->irq_received, 0);
+	atomic_set(&qdev->irq_received_display, 0);
+	atomic_set(&qdev->irq_received_cursor, 0);
+	atomic_set(&qdev->irq_received_io_cmd, 0);
+	qdev->irq_received_error = 0;
+	ret = drm_irq_install(qdev->ddev);
+	qdev->ram_header->int_mask = QXL_INTERRUPT_MASK;
+	if (unlikely(ret != 0)) {
+		DRM_ERROR("Failed installing irq: %d\n", ret);
+		return 1;
+	}
+	return 0;
+}
diff --git a/drivers/gpu/drm/qxl/qxl_kms.c b/drivers/gpu/drm/qxl/qxl_kms.c
new file mode 100644
index 000000000000..85127ed24cfd
--- /dev/null
+++ b/drivers/gpu/drm/qxl/qxl_kms.c
@@ -0,0 +1,302 @@
+/*
+ * Copyright 2013 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Dave Airlie
+ *          Alon Levy
+ */
+
+#include "qxl_drv.h"
+#include "qxl_object.h"
+
+#include <linux/io-mapping.h>
+
+int qxl_log_level;
+
+static void qxl_dump_mode(struct qxl_device *qdev, void *p)
+{
+	struct qxl_mode *m = p;
+	DRM_DEBUG_KMS("%d: %dx%d %d bits, stride %d, %dmm x %dmm, orientation %d\n",
+		      m->id, m->x_res, m->y_res, m->bits, m->stride, m->x_mili,
+		      m->y_mili, m->orientation);
+}
+
+static bool qxl_check_device(struct qxl_device *qdev)
+{
+	struct qxl_rom *rom = qdev->rom;
+	int mode_offset;
+	int i;
+
+	if (rom->magic != 0x4f525851) {
+		DRM_ERROR("bad rom signature %x\n", rom->magic);
+		return false;
+	}
+
+	DRM_INFO("Device Version %d.%d\n", rom->id, rom->update_id);
+	DRM_INFO("Compression level %d log level %d\n", rom->compression_level,
+		 rom->log_level);
+	DRM_INFO("Currently using mode #%d, list at 0x%x\n",
+		 rom->mode, rom->modes_offset);
+	DRM_INFO("%d io pages at offset 0x%x\n",
+		 rom->num_io_pages, rom->pages_offset);
+	DRM_INFO("%d byte draw area at offset 0x%x\n",
+		 rom->surface0_area_size, rom->draw_area_offset);
+
+	qdev->vram_size = rom->surface0_area_size;
+	DRM_INFO("RAM header offset: 0x%x\n", rom->ram_header_offset);
+
+	mode_offset = rom->modes_offset / 4;
+	qdev->mode_info.num_modes = ((u32 *)rom)[mode_offset];
+	DRM_INFO("rom modes offset 0x%x for %d modes\n", rom->modes_offset,
+		 qdev->mode_info.num_modes);
+	qdev->mode_info.modes = (void *)((uint32_t *)rom + mode_offset + 1);
+	for (i = 0; i < qdev->mode_info.num_modes; i++)
+		qxl_dump_mode(qdev, qdev->mode_info.modes + i);
+	return true;
+}
+
+static uint8_t setup_slot(struct qxl_device *qdev, uint8_t slot_index_offset,
+	unsigned long start_phys_addr, unsigned long end_phys_addr)
+{
+	uint64_t high_bits;
+	struct qxl_memslot *slot;
+	uint8_t slot_index;
+	struct qxl_ram_header *ram_header = qdev->ram_header;
+
+	slot_index = qdev->rom->slots_start + slot_index_offset;
+	slot = &qdev->mem_slots[slot_index];
+	slot->start_phys_addr = start_phys_addr;
+	slot->end_phys_addr = end_phys_addr;
+	ram_header->mem_slot.mem_start = slot->start_phys_addr;
+	ram_header->mem_slot.mem_end = slot->end_phys_addr;
+	qxl_io_memslot_add(qdev, slot_index);
+	slot->generation = qdev->rom->slot_generation;
+	high_bits = slot_index << qdev->slot_gen_bits;
+	high_bits |= slot->generation;
+	high_bits <<= (64 - (qdev->slot_gen_bits + qdev->slot_id_bits));
+	slot->high_bits = high_bits;
+	return slot_index;
+}
+
+static void qxl_gc_work(struct work_struct *work)
+{
+	struct qxl_device *qdev = container_of(work, struct qxl_device, gc_work);
+	qxl_garbage_collect(qdev);
+}
+
+int qxl_device_init(struct qxl_device *qdev,
+		    struct drm_device *ddev,
+		    struct pci_dev *pdev,
+		    unsigned long flags)
+{
+	int r;
+
+	qdev->dev = &pdev->dev;
+	qdev->ddev = ddev;
+	qdev->pdev = pdev;
+	qdev->flags = flags;
+
+	mutex_init(&qdev->gem.mutex);
+	mutex_init(&qdev->update_area_mutex);
+	mutex_init(&qdev->release_mutex);
+	mutex_init(&qdev->surf_evict_mutex);
+	INIT_LIST_HEAD(&qdev->gem.objects);
+
+	qdev->rom_base = pci_resource_start(pdev, 2);
+	qdev->rom_size = pci_resource_len(pdev, 2);
+	qdev->vram_base = pci_resource_start(pdev, 0);
+	qdev->surfaceram_base = pci_resource_start(pdev, 1);
+	qdev->surfaceram_size = pci_resource_len(pdev, 1);
+	qdev->io_base = pci_resource_start(pdev, 3);
+
+	qdev->vram_mapping = io_mapping_create_wc(qdev->vram_base, pci_resource_len(pdev, 0));
+	qdev->surface_mapping = io_mapping_create_wc(qdev->surfaceram_base, qdev->surfaceram_size);
+	DRM_DEBUG_KMS("qxl: vram %p-%p(%dM %dk), surface %p-%p(%dM %dk)\n",
+		 (void *)qdev->vram_base, (void *)pci_resource_end(pdev, 0),
+		 (int)pci_resource_len(pdev, 0) / 1024 / 1024,
+		 (int)pci_resource_len(pdev, 0) / 1024,
+		 (void *)qdev->surfaceram_base,
+		 (void *)pci_resource_end(pdev, 1),
+		 (int)qdev->surfaceram_size / 1024 / 1024,
+		 (int)qdev->surfaceram_size / 1024);
+
+	qdev->rom = ioremap(qdev->rom_base, qdev->rom_size);
+	if (!qdev->rom) {
+		pr_err("Unable to ioremap ROM\n");
+		return -ENOMEM;
+	}
+
+	qxl_check_device(qdev);
+
+	r = qxl_bo_init(qdev);
+	if (r) {
+		DRM_ERROR("bo init failed %d\n", r);
+		return r;
+	}
+
+	qdev->ram_header = ioremap(qdev->vram_base +
+				   qdev->rom->ram_header_offset,
+				   sizeof(*qdev->ram_header));
+
+	qdev->command_ring = qxl_ring_create(&(qdev->ram_header->cmd_ring_hdr),
+					     sizeof(struct qxl_command),
+					     QXL_COMMAND_RING_SIZE,
+					     qdev->io_base + QXL_IO_NOTIFY_CMD,
+					     false,
+					     &qdev->display_event);
+
+	qdev->cursor_ring = qxl_ring_create(
+				&(qdev->ram_header->cursor_ring_hdr),
+				sizeof(struct qxl_command),
+				QXL_CURSOR_RING_SIZE,
+				qdev->io_base + QXL_IO_NOTIFY_CMD,
+				false,
+				&qdev->cursor_event);
+
+	qdev->release_ring = qxl_ring_create(
+				&(qdev->ram_header->release_ring_hdr),
+				sizeof(uint64_t),
+				QXL_RELEASE_RING_SIZE, 0, true,
+				NULL);
+
+	/* TODO - slot initialization should happen on reset. where is our
+	 * reset handler? */
+	qdev->n_mem_slots = qdev->rom->slots_end;
+	qdev->slot_gen_bits = qdev->rom->slot_gen_bits;
+	qdev->slot_id_bits = qdev->rom->slot_id_bits;
+	qdev->va_slot_mask =
+		(~(uint64_t)0) >> (qdev->slot_id_bits + qdev->slot_gen_bits);
+
+	qdev->mem_slots =
+		kmalloc(qdev->n_mem_slots * sizeof(struct qxl_memslot),
+			GFP_KERNEL);
+
+	idr_init(&qdev->release_idr);
+	spin_lock_init(&qdev->release_idr_lock);
+
+	idr_init(&qdev->surf_id_idr);
+	spin_lock_init(&qdev->surf_id_idr_lock);
+
+	mutex_init(&qdev->async_io_mutex);
+
+	/* reset the device into a known state - no memslots, no primary
+	 * created, no surfaces. */
+	qxl_io_reset(qdev);
+
+	/* must initialize irq before first async io - slot creation */
+	r = qxl_irq_init(qdev);
+	if (r)
+		return r;
+
+	/*
+	 * Note that virtual is surface0. We rely on the single ioremap done
+	 * before.
+	 */
+	qdev->main_mem_slot = setup_slot(qdev, 0,
+		(unsigned long)qdev->vram_base,
+		(unsigned long)qdev->vram_base + qdev->rom->ram_header_offset);
+	qdev->surfaces_mem_slot = setup_slot(qdev, 1,
+		(unsigned long)qdev->surfaceram_base,
+		(unsigned long)qdev->surfaceram_base + qdev->surfaceram_size);
+	DRM_INFO("main mem slot %d [%lx,%x)\n",
+		qdev->main_mem_slot,
+		(unsigned long)qdev->vram_base, qdev->rom->ram_header_offset);
+
+
+	qdev->gc_queue = create_singlethread_workqueue("qxl_gc");
+	INIT_WORK(&qdev->gc_work, qxl_gc_work);
+
+	r = qxl_fb_init(qdev);
+	if (r)
+		return r;
+
+	return 0;
+}
+
+static void qxl_device_fini(struct qxl_device *qdev)
+{
+	if (qdev->current_release_bo[0])
+		qxl_bo_unref(&qdev->current_release_bo[0]);
+	if (qdev->current_release_bo[1])
+		qxl_bo_unref(&qdev->current_release_bo[1]);
+	flush_workqueue(qdev->gc_queue);
+	destroy_workqueue(qdev->gc_queue);
+	qdev->gc_queue = NULL;
+
+	qxl_ring_free(qdev->command_ring);
+	qxl_ring_free(qdev->cursor_ring);
+	qxl_ring_free(qdev->release_ring);
+	qxl_bo_fini(qdev);
+	io_mapping_free(qdev->surface_mapping);
+	io_mapping_free(qdev->vram_mapping);
+	iounmap(qdev->ram_header);
+	iounmap(qdev->rom);
+	qdev->rom = NULL;
+	qdev->mode_info.modes = NULL;
+	qdev->mode_info.num_modes = 0;
+	qxl_debugfs_remove_files(qdev);
+}
+
+int qxl_driver_unload(struct drm_device *dev)
+{
+	struct qxl_device *qdev = dev->dev_private;
+
+	if (qdev == NULL)
+		return 0;
+	qxl_modeset_fini(qdev);
+	qxl_device_fini(qdev);
+
+	kfree(qdev);
+	dev->dev_private = NULL;
+	return 0;
+}
+
+int qxl_driver_load(struct drm_device *dev, unsigned long flags)
+{
+	struct qxl_device *qdev;
+	int r;
+
+	/* require kms */
+	if (!drm_core_check_feature(dev, DRIVER_MODESET))
+		return -ENODEV;
+
+	qdev = kzalloc(sizeof(struct qxl_device), GFP_KERNEL);
+	if (qdev == NULL)
+		return -ENOMEM;
+
+	dev->dev_private = qdev;
+
+	r = qxl_device_init(qdev, dev, dev->pdev, flags);
+	if (r)
+		goto out;
+
+	r = qxl_modeset_init(qdev);
+	if (r) {
+		qxl_driver_unload(dev);
+		goto out;
+	}
+
+	return 0;
+out:
+	kfree(qdev);
+	return r;
+}
+
+
diff --git a/drivers/gpu/drm/qxl/qxl_object.c b/drivers/gpu/drm/qxl/qxl_object.c
new file mode 100644
index 000000000000..d9b12e7bc6e1
--- /dev/null
+++ b/drivers/gpu/drm/qxl/qxl_object.c
@@ -0,0 +1,365 @@
+/*
+ * Copyright 2013 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Dave Airlie
+ *          Alon Levy
+ */
+
+#include "qxl_drv.h"
+#include "qxl_object.h"
+
+#include <linux/io-mapping.h>
+static void qxl_ttm_bo_destroy(struct ttm_buffer_object *tbo)
+{
+	struct qxl_bo *bo;
+	struct qxl_device *qdev;
+
+	bo = container_of(tbo, struct qxl_bo, tbo);
+	qdev = (struct qxl_device *)bo->gem_base.dev->dev_private;
+
+	qxl_surface_evict(qdev, bo, false);
+	qxl_fence_fini(&bo->fence);
+	mutex_lock(&qdev->gem.mutex);
+	list_del_init(&bo->list);
+	mutex_unlock(&qdev->gem.mutex);
+	drm_gem_object_release(&bo->gem_base);
+	kfree(bo);
+}
+
+bool qxl_ttm_bo_is_qxl_bo(struct ttm_buffer_object *bo)
+{
+	if (bo->destroy == &qxl_ttm_bo_destroy)
+		return true;
+	return false;
+}
+
+void qxl_ttm_placement_from_domain(struct qxl_bo *qbo, u32 domain)
+{
+	u32 c = 0;
+
+	qbo->placement.fpfn = 0;
+	qbo->placement.lpfn = 0;
+	qbo->placement.placement = qbo->placements;
+	qbo->placement.busy_placement = qbo->placements;
+	if (domain == QXL_GEM_DOMAIN_VRAM)
+		qbo->placements[c++] = TTM_PL_FLAG_CACHED | TTM_PL_FLAG_VRAM;
+	if (domain == QXL_GEM_DOMAIN_SURFACE)
+		qbo->placements[c++] = TTM_PL_FLAG_CACHED | TTM_PL_FLAG_PRIV0;
+	if (domain == QXL_GEM_DOMAIN_CPU)
+		qbo->placements[c++] = TTM_PL_MASK_CACHING | TTM_PL_FLAG_SYSTEM;
+	if (!c)
+		qbo->placements[c++] = TTM_PL_MASK_CACHING | TTM_PL_FLAG_SYSTEM;
+	qbo->placement.num_placement = c;
+	qbo->placement.num_busy_placement = c;
+}
+
+
+int qxl_bo_create(struct qxl_device *qdev,
+		  unsigned long size, bool kernel, u32 domain,
+		  struct qxl_surface *surf,
+		  struct qxl_bo **bo_ptr)
+{
+	struct qxl_bo *bo;
+	enum ttm_bo_type type;
+	int r;
+
+	if (unlikely(qdev->mman.bdev.dev_mapping == NULL))
+		qdev->mman.bdev.dev_mapping = qdev->ddev->dev_mapping;
+	if (kernel)
+		type = ttm_bo_type_kernel;
+	else
+		type = ttm_bo_type_device;
+	*bo_ptr = NULL;
+	bo = kzalloc(sizeof(struct qxl_bo), GFP_KERNEL);
+	if (bo == NULL)
+		return -ENOMEM;
+	size = roundup(size, PAGE_SIZE);
+	r = drm_gem_object_init(qdev->ddev, &bo->gem_base, size);
+	if (unlikely(r)) {
+		kfree(bo);
+		return r;
+	}
+	bo->gem_base.driver_private = NULL;
+	bo->type = domain;
+	bo->pin_count = 0;
+	bo->surface_id = 0;
+	qxl_fence_init(qdev, &bo->fence);
+	INIT_LIST_HEAD(&bo->list);
+	atomic_set(&bo->reserve_count, 0);
+	if (surf)
+		bo->surf = *surf;
+
+	qxl_ttm_placement_from_domain(bo, domain);
+
+	r = ttm_bo_init(&qdev->mman.bdev, &bo->tbo, size, type,
+			&bo->placement, 0, !kernel, NULL, size,
+			NULL, &qxl_ttm_bo_destroy);
+	if (unlikely(r != 0)) {
+		if (r != -ERESTARTSYS)
+			dev_err(qdev->dev,
+				"object_init failed for (%lu, 0x%08X)\n",
+				size, domain);
+		return r;
+	}
+	*bo_ptr = bo;
+	return 0;
+}
+
+int qxl_bo_kmap(struct qxl_bo *bo, void **ptr)
+{
+	bool is_iomem;
+	int r;
+
+	if (bo->kptr) {
+		if (ptr)
+			*ptr = bo->kptr;
+		return 0;
+	}
+	r = ttm_bo_kmap(&bo->tbo, 0, bo->tbo.num_pages, &bo->kmap);
+	if (r)
+		return r;
+	bo->kptr = ttm_kmap_obj_virtual(&bo->kmap, &is_iomem);
+	if (ptr)
+		*ptr = bo->kptr;
+	return 0;
+}
+
+void *qxl_bo_kmap_atomic_page(struct qxl_device *qdev,
+			      struct qxl_bo *bo, int page_offset)
+{
+	struct ttm_mem_type_manager *man = &bo->tbo.bdev->man[bo->tbo.mem.mem_type];
+	void *rptr;
+	int ret;
+	struct io_mapping *map;
+
+	if (bo->tbo.mem.mem_type == TTM_PL_VRAM)
+		map = qdev->vram_mapping;
+	else if (bo->tbo.mem.mem_type == TTM_PL_PRIV0)
+		map = qdev->surface_mapping;
+	else
+		goto fallback;
+
+	(void) ttm_mem_io_lock(man, false);
+	ret = ttm_mem_io_reserve(bo->tbo.bdev, &bo->tbo.mem);
+	ttm_mem_io_unlock(man);
+
+	return io_mapping_map_atomic_wc(map, bo->tbo.mem.bus.offset + page_offset);
+fallback:
+	if (bo->kptr) {
+		rptr = bo->kptr + (page_offset * PAGE_SIZE);
+		return rptr;
+	}
+
+	ret = qxl_bo_kmap(bo, &rptr);
+	if (ret)
+		return NULL;
+
+	rptr += page_offset * PAGE_SIZE;
+	return rptr;
+}
+
+void qxl_bo_kunmap(struct qxl_bo *bo)
+{
+	if (bo->kptr == NULL)
+		return;
+	bo->kptr = NULL;
+	ttm_bo_kunmap(&bo->kmap);
+}
+
+void qxl_bo_kunmap_atomic_page(struct qxl_device *qdev,
+			       struct qxl_bo *bo, void *pmap)
+{
+	struct ttm_mem_type_manager *man = &bo->tbo.bdev->man[bo->tbo.mem.mem_type];
+	struct io_mapping *map;
+
+	if (bo->tbo.mem.mem_type == TTM_PL_VRAM)
+		map = qdev->vram_mapping;
+	else if (bo->tbo.mem.mem_type == TTM_PL_PRIV0)
+		map = qdev->surface_mapping;
+	else
+		goto fallback;
+
+	io_mapping_unmap_atomic(pmap);
+
+	(void) ttm_mem_io_lock(man, false);
+	ttm_mem_io_free(bo->tbo.bdev, &bo->tbo.mem);
+	ttm_mem_io_unlock(man);
+	return ;
+ fallback:
+	qxl_bo_kunmap(bo);
+}
+
+void qxl_bo_unref(struct qxl_bo **bo)
+{
+	struct ttm_buffer_object *tbo;
+
+	if ((*bo) == NULL)
+		return;
+	tbo = &((*bo)->tbo);
+	ttm_bo_unref(&tbo);
+	if (tbo == NULL)
+		*bo = NULL;
+}
+
+struct qxl_bo *qxl_bo_ref(struct qxl_bo *bo)
+{
+	ttm_bo_reference(&bo->tbo);
+	return bo;
+}
+
+int qxl_bo_pin(struct qxl_bo *bo, u32 domain, u64 *gpu_addr)
+{
+	struct qxl_device *qdev = (struct qxl_device *)bo->gem_base.dev->dev_private;
+	int r, i;
+
+	if (bo->pin_count) {
+		bo->pin_count++;
+		if (gpu_addr)
+			*gpu_addr = qxl_bo_gpu_offset(bo);
+		return 0;
+	}
+	qxl_ttm_placement_from_domain(bo, domain);
+	for (i = 0; i < bo->placement.num_placement; i++)
+		bo->placements[i] |= TTM_PL_FLAG_NO_EVICT;
+	r = ttm_bo_validate(&bo->tbo, &bo->placement, false, false);
+	if (likely(r == 0)) {
+		bo->pin_count = 1;
+		if (gpu_addr != NULL)
+			*gpu_addr = qxl_bo_gpu_offset(bo);
+	}
+	if (unlikely(r != 0))
+		dev_err(qdev->dev, "%p pin failed\n", bo);
+	return r;
+}
+
+int qxl_bo_unpin(struct qxl_bo *bo)
+{
+	struct qxl_device *qdev = (struct qxl_device *)bo->gem_base.dev->dev_private;
+	int r, i;
+
+	if (!bo->pin_count) {
+		dev_warn(qdev->dev, "%p unpin not necessary\n", bo);
+		return 0;
+	}
+	bo->pin_count--;
+	if (bo->pin_count)
+		return 0;
+	for (i = 0; i < bo->placement.num_placement; i++)
+		bo->placements[i] &= ~TTM_PL_FLAG_NO_EVICT;
+	r = ttm_bo_validate(&bo->tbo, &bo->placement, false, false);
+	if (unlikely(r != 0))
+		dev_err(qdev->dev, "%p validate failed for unpin\n", bo);
+	return r;
+}
+
+void qxl_bo_force_delete(struct qxl_device *qdev)
+{
+	struct qxl_bo *bo, *n;
+
+	if (list_empty(&qdev->gem.objects))
+		return;
+	dev_err(qdev->dev, "Userspace still has active objects !\n");
+	list_for_each_entry_safe(bo, n, &qdev->gem.objects, list) {
+		mutex_lock(&qdev->ddev->struct_mutex);
+		dev_err(qdev->dev, "%p %p %lu %lu force free\n",
+			&bo->gem_base, bo, (unsigned long)bo->gem_base.size,
+			*((unsigned long *)&bo->gem_base.refcount));
+		mutex_lock(&qdev->gem.mutex);
+		list_del_init(&bo->list);
+		mutex_unlock(&qdev->gem.mutex);
+		/* this should unref the ttm bo */
+		drm_gem_object_unreference(&bo->gem_base);
+		mutex_unlock(&qdev->ddev->struct_mutex);
+	}
+}
+
+int qxl_bo_init(struct qxl_device *qdev)
+{
+	return qxl_ttm_init(qdev);
+}
+
+void qxl_bo_fini(struct qxl_device *qdev)
+{
+	qxl_ttm_fini(qdev);
+}
+
+int qxl_bo_check_id(struct qxl_device *qdev, struct qxl_bo *bo)
+{
+	int ret;
+	if (bo->type == QXL_GEM_DOMAIN_SURFACE && bo->surface_id == 0) {
+		/* allocate a surface id for this surface now */
+		ret = qxl_surface_id_alloc(qdev, bo);
+		if (ret)
+			return ret;
+
+		ret = qxl_hw_surface_alloc(qdev, bo, NULL);
+		if (ret)
+			return ret;
+	}
+	return 0;
+}
+
+void qxl_bo_list_unreserve(struct qxl_reloc_list *reloc_list, bool failed)
+{
+	struct qxl_bo_list *entry, *sf;
+
+	list_for_each_entry_safe(entry, sf, &reloc_list->bos, lhead) {
+		qxl_bo_unreserve(entry->bo);
+		list_del(&entry->lhead);
+		kfree(entry);
+	}
+}
+
+int qxl_bo_list_add(struct qxl_reloc_list *reloc_list, struct qxl_bo *bo)
+{
+	struct qxl_bo_list *entry;
+	int ret;
+
+	list_for_each_entry(entry, &reloc_list->bos, lhead) {
+		if (entry->bo == bo)
+			return 0;
+	}
+
+	entry = kmalloc(sizeof(struct qxl_bo_list), GFP_KERNEL);
+	if (!entry)
+		return -ENOMEM;
+
+	entry->bo = bo;
+	list_add(&entry->lhead, &reloc_list->bos);
+
+	ret = qxl_bo_reserve(bo, false);
+	if (ret)
+		return ret;
+
+	if (!bo->pin_count) {
+		qxl_ttm_placement_from_domain(bo, bo->type);
+		ret = ttm_bo_validate(&bo->tbo, &bo->placement,
+				      true, false);
+		if (ret)
+			return ret;
+	}
+
+	/* allocate a surface for reserved + validated buffers */
+	ret = qxl_bo_check_id(bo->gem_base.dev->dev_private, bo);
+	if (ret)
+		return ret;
+	return 0;
+}
diff --git a/drivers/gpu/drm/qxl/qxl_object.h b/drivers/gpu/drm/qxl/qxl_object.h
new file mode 100644
index 000000000000..b4fd89fbd8b7
--- /dev/null
+++ b/drivers/gpu/drm/qxl/qxl_object.h
@@ -0,0 +1,112 @@
+/*
+ * Copyright 2013 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Dave Airlie
+ *          Alon Levy
+ */
+#ifndef QXL_OBJECT_H
+#define QXL_OBJECT_H
+
+#include "qxl_drv.h"
+
+static inline int qxl_bo_reserve(struct qxl_bo *bo, bool no_wait)
+{
+	int r;
+
+	r = ttm_bo_reserve(&bo->tbo, true, no_wait, false, 0);
+	if (unlikely(r != 0)) {
+		if (r != -ERESTARTSYS) {
+			struct qxl_device *qdev = (struct qxl_device *)bo->gem_base.dev->dev_private;
+			dev_err(qdev->dev, "%p reserve failed\n", bo);
+		}
+		return r;
+	}
+	return 0;
+}
+
+static inline void qxl_bo_unreserve(struct qxl_bo *bo)
+{
+	ttm_bo_unreserve(&bo->tbo);
+}
+
+static inline u64 qxl_bo_gpu_offset(struct qxl_bo *bo)
+{
+	return bo->tbo.offset;
+}
+
+static inline unsigned long qxl_bo_size(struct qxl_bo *bo)
+{
+	return bo->tbo.num_pages << PAGE_SHIFT;
+}
+
+static inline bool qxl_bo_is_reserved(struct qxl_bo *bo)
+{
+	return !!atomic_read(&bo->tbo.reserved);
+}
+
+static inline u64 qxl_bo_mmap_offset(struct qxl_bo *bo)
+{
+	return bo->tbo.addr_space_offset;
+}
+
+static inline int qxl_bo_wait(struct qxl_bo *bo, u32 *mem_type,
+			      bool no_wait)
+{
+	int r;
+
+	r = ttm_bo_reserve(&bo->tbo, true, no_wait, false, 0);
+	if (unlikely(r != 0)) {
+		if (r != -ERESTARTSYS) {
+			struct qxl_device *qdev = (struct qxl_device *)bo->gem_base.dev->dev_private;
+			dev_err(qdev->dev, "%p reserve failed for wait\n",
+				bo);
+		}
+		return r;
+	}
+	spin_lock(&bo->tbo.bdev->fence_lock);
+	if (mem_type)
+		*mem_type = bo->tbo.mem.mem_type;
+	if (bo->tbo.sync_obj)
+		r = ttm_bo_wait(&bo->tbo, true, true, no_wait);
+	spin_unlock(&bo->tbo.bdev->fence_lock);
+	ttm_bo_unreserve(&bo->tbo);
+	return r;
+}
+
+extern int qxl_bo_create(struct qxl_device *qdev,
+			 unsigned long size,
+			 bool kernel, u32 domain,
+			 struct qxl_surface *surf,
+			 struct qxl_bo **bo_ptr);
+extern int qxl_bo_kmap(struct qxl_bo *bo, void **ptr);
+extern void qxl_bo_kunmap(struct qxl_bo *bo);
+void *qxl_bo_kmap_atomic_page(struct qxl_device *qdev, struct qxl_bo *bo, int page_offset);
+void qxl_bo_kunmap_atomic_page(struct qxl_device *qdev, struct qxl_bo *bo, void *map);
+extern struct qxl_bo *qxl_bo_ref(struct qxl_bo *bo);
+extern void qxl_bo_unref(struct qxl_bo **bo);
+extern int qxl_bo_pin(struct qxl_bo *bo, u32 domain, u64 *gpu_addr);
+extern int qxl_bo_unpin(struct qxl_bo *bo);
+extern void qxl_ttm_placement_from_domain(struct qxl_bo *qbo, u32 domain);
+extern bool qxl_ttm_bo_is_qxl_bo(struct ttm_buffer_object *bo);
+
+extern int qxl_bo_list_add(struct qxl_reloc_list *reloc_list, struct qxl_bo *bo);
+extern void qxl_bo_list_unreserve(struct qxl_reloc_list *reloc_list, bool failed);
+#endif
diff --git a/drivers/gpu/drm/qxl/qxl_release.c b/drivers/gpu/drm/qxl/qxl_release.c
new file mode 100644
index 000000000000..b443d6751d5f
--- /dev/null
+++ b/drivers/gpu/drm/qxl/qxl_release.c
@@ -0,0 +1,304 @@
+/*
+ * Copyright 2011 Red Hat, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * on the rights to use, copy, modify, merge, publish, distribute, sub
+ * license, and/or sell copies of the Software, and to permit persons to whom
+ * the Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER
+ * IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ */
+#include "qxl_drv.h"
+#include "qxl_object.h"
+
+/*
+ * drawable cmd cache - allocate a bunch of VRAM pages, suballocate
+ * into 256 byte chunks for now - gives 16 cmds per page.
+ *
+ * use an ida to index into the chunks?
+ */
+/* manage releaseables */
+/* stack them 16 high for now -drawable object is 191 */
+#define RELEASE_SIZE 256
+#define RELEASES_PER_BO (4096 / RELEASE_SIZE)
+/* put an alloc/dealloc surface cmd into one bo and round up to 128 */
+#define SURFACE_RELEASE_SIZE 128
+#define SURFACE_RELEASES_PER_BO (4096 / SURFACE_RELEASE_SIZE)
+
+static const int release_size_per_bo[] = { RELEASE_SIZE, SURFACE_RELEASE_SIZE, RELEASE_SIZE };
+static const int releases_per_bo[] = { RELEASES_PER_BO, SURFACE_RELEASES_PER_BO, RELEASES_PER_BO };
+uint64_t
+qxl_release_alloc(struct qxl_device *qdev, int type,
+		  struct qxl_release **ret)
+{
+	struct qxl_release *release;
+	int handle;
+	size_t size = sizeof(*release);
+	int idr_ret;
+
+	release = kmalloc(size, GFP_KERNEL);
+	if (!release) {
+		DRM_ERROR("Out of memory\n");
+		return 0;
+	}
+	release->type = type;
+	release->bo_count = 0;
+	release->release_offset = 0;
+	release->surface_release_id = 0;
+
+	idr_preload(GFP_KERNEL);
+	spin_lock(&qdev->release_idr_lock);
+	idr_ret = idr_alloc(&qdev->release_idr, release, 1, 0, GFP_NOWAIT);
+	spin_unlock(&qdev->release_idr_lock);
+	idr_preload_end();
+	handle = idr_ret;
+	if (idr_ret < 0)
+		goto release_fail;
+	*ret = release;
+	QXL_INFO(qdev, "allocated release %lld\n", handle);
+	release->id = handle;
+release_fail:
+
+	return handle;
+}
+
+void
+qxl_release_free(struct qxl_device *qdev,
+		 struct qxl_release *release)
+{
+	int i;
+
+	QXL_INFO(qdev, "release %d, type %d, %d bos\n", release->id,
+		 release->type, release->bo_count);
+
+	if (release->surface_release_id)
+		qxl_surface_id_dealloc(qdev, release->surface_release_id);
+
+	for (i = 0 ; i < release->bo_count; ++i) {
+		QXL_INFO(qdev, "release %llx\n",
+			release->bos[i]->tbo.addr_space_offset
+						- DRM_FILE_OFFSET);
+		qxl_fence_remove_release(&release->bos[i]->fence, release->id);
+		qxl_bo_unref(&release->bos[i]);
+	}
+	spin_lock(&qdev->release_idr_lock);
+	idr_remove(&qdev->release_idr, release->id);
+	spin_unlock(&qdev->release_idr_lock);
+	kfree(release);
+}
+
+void
+qxl_release_add_res(struct qxl_device *qdev, struct qxl_release *release,
+		    struct qxl_bo *bo)
+{
+	int i;
+	for (i = 0; i < release->bo_count; i++)
+		if (release->bos[i] == bo)
+			return;
+
+	if (release->bo_count >= QXL_MAX_RES) {
+		DRM_ERROR("exceeded max resource on a qxl_release item\n");
+		return;
+	}
+	release->bos[release->bo_count++] = qxl_bo_ref(bo);
+}
+
+static int qxl_release_bo_alloc(struct qxl_device *qdev,
+				struct qxl_bo **bo)
+{
+	int ret;
+	ret = qxl_bo_create(qdev, PAGE_SIZE, false, QXL_GEM_DOMAIN_VRAM, NULL,
+			    bo);
+	return ret;
+}
+
+int qxl_release_reserve(struct qxl_device *qdev,
+			struct qxl_release *release, bool no_wait)
+{
+	int ret;
+	if (atomic_inc_return(&release->bos[0]->reserve_count) == 1) {
+		ret = qxl_bo_reserve(release->bos[0], no_wait);
+		if (ret)
+			return ret;
+	}
+	return 0;
+}
+
+void qxl_release_unreserve(struct qxl_device *qdev,
+			  struct qxl_release *release)
+{
+	if (atomic_dec_and_test(&release->bos[0]->reserve_count))
+		qxl_bo_unreserve(release->bos[0]);
+}
+
+int qxl_alloc_surface_release_reserved(struct qxl_device *qdev,
+				       enum qxl_surface_cmd_type surface_cmd_type,
+				       struct qxl_release *create_rel,
+				       struct qxl_release **release)
+{
+	int ret;
+
+	if (surface_cmd_type == QXL_SURFACE_CMD_DESTROY && create_rel) {
+		int idr_ret;
+		struct qxl_bo *bo;
+		union qxl_release_info *info;
+
+		/* stash the release after the create command */
+		idr_ret = qxl_release_alloc(qdev, QXL_RELEASE_SURFACE_CMD, release);
+		bo = qxl_bo_ref(create_rel->bos[0]);
+
+		(*release)->release_offset = create_rel->release_offset + 64;
+
+		qxl_release_add_res(qdev, *release, bo);
+
+		ret = qxl_release_reserve(qdev, *release, false);
+		if (ret) {
+			DRM_ERROR("release reserve failed\n");
+			goto out_unref;
+		}
+		info = qxl_release_map(qdev, *release);
+		info->id = idr_ret;
+		qxl_release_unmap(qdev, *release, info);
+
+
+out_unref:
+		qxl_bo_unref(&bo);
+		return ret;
+	}
+
+	return qxl_alloc_release_reserved(qdev, sizeof(struct qxl_surface_cmd),
+					 QXL_RELEASE_SURFACE_CMD, release, NULL);
+}
+
+int qxl_alloc_release_reserved(struct qxl_device *qdev, unsigned long size,
+				       int type, struct qxl_release **release,
+				       struct qxl_bo **rbo)
+{
+	struct qxl_bo *bo;
+	int idr_ret;
+	int ret;
+	union qxl_release_info *info;
+	int cur_idx;
+
+	if (type == QXL_RELEASE_DRAWABLE)
+		cur_idx = 0;
+	else if (type == QXL_RELEASE_SURFACE_CMD)
+		cur_idx = 1;
+	else if (type == QXL_RELEASE_CURSOR_CMD)
+		cur_idx = 2;
+	else {
+		DRM_ERROR("got illegal type: %d\n", type);
+		return -EINVAL;
+	}
+
+	idr_ret = qxl_release_alloc(qdev, type, release);
+
+	mutex_lock(&qdev->release_mutex);
+	if (qdev->current_release_bo_offset[cur_idx] + 1 >= releases_per_bo[cur_idx]) {
+		qxl_bo_unref(&qdev->current_release_bo[cur_idx]);
+		qdev->current_release_bo_offset[cur_idx] = 0;
+		qdev->current_release_bo[cur_idx] = NULL;
+	}
+	if (!qdev->current_release_bo[cur_idx]) {
+		ret = qxl_release_bo_alloc(qdev, &qdev->current_release_bo[cur_idx]);
+		if (ret) {
+			mutex_unlock(&qdev->release_mutex);
+			return ret;
+		}
+
+		/* pin releases bo's they are too messy to evict */
+		ret = qxl_bo_reserve(qdev->current_release_bo[cur_idx], false);
+		qxl_bo_pin(qdev->current_release_bo[cur_idx], QXL_GEM_DOMAIN_VRAM, NULL);
+		qxl_bo_unreserve(qdev->current_release_bo[cur_idx]);
+	}
+
+	bo = qxl_bo_ref(qdev->current_release_bo[cur_idx]);
+
+	(*release)->release_offset = qdev->current_release_bo_offset[cur_idx] * release_size_per_bo[cur_idx];
+	qdev->current_release_bo_offset[cur_idx]++;
+
+	if (rbo)
+		*rbo = bo;
+
+	qxl_release_add_res(qdev, *release, bo);
+
+	ret = qxl_release_reserve(qdev, *release, false);
+	mutex_unlock(&qdev->release_mutex);
+	if (ret)
+		goto out_unref;
+
+	info = qxl_release_map(qdev, *release);
+	info->id = idr_ret;
+	qxl_release_unmap(qdev, *release, info);
+
+out_unref:
+	qxl_bo_unref(&bo);
+	return ret;
+}
+
+int qxl_fence_releaseable(struct qxl_device *qdev,
+			  struct qxl_release *release)
+{
+	int i, ret;
+	for (i = 0; i < release->bo_count; i++) {
+		if (!release->bos[i]->tbo.sync_obj)
+			release->bos[i]->tbo.sync_obj = &release->bos[i]->fence;
+		ret = qxl_fence_add_release(&release->bos[i]->fence, release->id);
+		if (ret)
+			return ret;
+	}
+	return 0;
+}
+
+struct qxl_release *qxl_release_from_id_locked(struct qxl_device *qdev,
+						   uint64_t id)
+{
+	struct qxl_release *release;
+
+	spin_lock(&qdev->release_idr_lock);
+	release = idr_find(&qdev->release_idr, id);
+	spin_unlock(&qdev->release_idr_lock);
+	if (!release) {
+		DRM_ERROR("failed to find id in release_idr\n");
+		return NULL;
+	}
+	if (release->bo_count < 1) {
+		DRM_ERROR("read a released resource with 0 bos\n");
+		return NULL;
+	}
+	return release;
+}
+
+union qxl_release_info *qxl_release_map(struct qxl_device *qdev,
+					struct qxl_release *release)
+{
+	void *ptr;
+	union qxl_release_info *info;
+	struct qxl_bo *bo = release->bos[0];
+
+	ptr = qxl_bo_kmap_atomic_page(qdev, bo, release->release_offset & PAGE_SIZE);
+	info = ptr + (release->release_offset & ~PAGE_SIZE);
+	return info;
+}
+
+void qxl_release_unmap(struct qxl_device *qdev,
+		       struct qxl_release *release,
+		       union qxl_release_info *info)
+{
+	struct qxl_bo *bo = release->bos[0];
+	void *ptr;
+
+	ptr = ((void *)info) - (release->release_offset & ~PAGE_SIZE);
+	qxl_bo_kunmap_atomic_page(qdev, bo, ptr);
+}
diff --git a/drivers/gpu/drm/qxl/qxl_ttm.c b/drivers/gpu/drm/qxl/qxl_ttm.c
new file mode 100644
index 000000000000..489cb8cece4d
--- /dev/null
+++ b/drivers/gpu/drm/qxl/qxl_ttm.c
@@ -0,0 +1,581 @@
+/*
+ * Copyright 2013 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Dave Airlie
+ *          Alon Levy
+ */
+
+#include <ttm/ttm_bo_api.h>
+#include <ttm/ttm_bo_driver.h>
+#include <ttm/ttm_placement.h>
+#include <ttm/ttm_page_alloc.h>
+#include <ttm/ttm_module.h>
+#include <drm/drmP.h>
+#include <drm/drm.h>
+#include <drm/qxl_drm.h>
+#include "qxl_drv.h"
+#include "qxl_object.h"
+
+#include <linux/delay.h>
+static int qxl_ttm_debugfs_init(struct qxl_device *qdev);
+
+static struct qxl_device *qxl_get_qdev(struct ttm_bo_device *bdev)
+{
+	struct qxl_mman *mman;
+	struct qxl_device *qdev;
+
+	mman = container_of(bdev, struct qxl_mman, bdev);
+	qdev = container_of(mman, struct qxl_device, mman);
+	return qdev;
+}
+
+static int qxl_ttm_mem_global_init(struct drm_global_reference *ref)
+{
+	return ttm_mem_global_init(ref->object);
+}
+
+static void qxl_ttm_mem_global_release(struct drm_global_reference *ref)
+{
+	ttm_mem_global_release(ref->object);
+}
+
+static int qxl_ttm_global_init(struct qxl_device *qdev)
+{
+	struct drm_global_reference *global_ref;
+	int r;
+
+	qdev->mman.mem_global_referenced = false;
+	global_ref = &qdev->mman.mem_global_ref;
+	global_ref->global_type = DRM_GLOBAL_TTM_MEM;
+	global_ref->size = sizeof(struct ttm_mem_global);
+	global_ref->init = &qxl_ttm_mem_global_init;
+	global_ref->release = &qxl_ttm_mem_global_release;
+
+	r = drm_global_item_ref(global_ref);
+	if (r != 0) {
+		DRM_ERROR("Failed setting up TTM memory accounting "
+			  "subsystem.\n");
+		return r;
+	}
+
+	qdev->mman.bo_global_ref.mem_glob =
+		qdev->mman.mem_global_ref.object;
+	global_ref = &qdev->mman.bo_global_ref.ref;
+	global_ref->global_type = DRM_GLOBAL_TTM_BO;
+	global_ref->size = sizeof(struct ttm_bo_global);
+	global_ref->init = &ttm_bo_global_init;
+	global_ref->release = &ttm_bo_global_release;
+	r = drm_global_item_ref(global_ref);
+	if (r != 0) {
+		DRM_ERROR("Failed setting up TTM BO subsystem.\n");
+		drm_global_item_unref(&qdev->mman.mem_global_ref);
+		return r;
+	}
+
+	qdev->mman.mem_global_referenced = true;
+	return 0;
+}
+
+static void qxl_ttm_global_fini(struct qxl_device *qdev)
+{
+	if (qdev->mman.mem_global_referenced) {
+		drm_global_item_unref(&qdev->mman.bo_global_ref.ref);
+		drm_global_item_unref(&qdev->mman.mem_global_ref);
+		qdev->mman.mem_global_referenced = false;
+	}
+}
+
+static struct vm_operations_struct qxl_ttm_vm_ops;
+static const struct vm_operations_struct *ttm_vm_ops;
+
+static int qxl_ttm_fault(struct vm_area_struct *vma, struct vm_fault *vmf)
+{
+	struct ttm_buffer_object *bo;
+	struct qxl_device *qdev;
+	int r;
+
+	bo = (struct ttm_buffer_object *)vma->vm_private_data;
+	if (bo == NULL)
+		return VM_FAULT_NOPAGE;
+	qdev = qxl_get_qdev(bo->bdev);
+	r = ttm_vm_ops->fault(vma, vmf);
+	return r;
+}
+
+int qxl_mmap(struct file *filp, struct vm_area_struct *vma)
+{
+	struct drm_file *file_priv;
+	struct qxl_device *qdev;
+	int r;
+
+	if (unlikely(vma->vm_pgoff < DRM_FILE_PAGE_OFFSET)) {
+		pr_info("%s: vma->vm_pgoff (%ld) < DRM_FILE_PAGE_OFFSET\n",
+			__func__, vma->vm_pgoff);
+		return drm_mmap(filp, vma);
+	}
+
+	file_priv = filp->private_data;
+	qdev = file_priv->minor->dev->dev_private;
+	if (qdev == NULL) {
+		DRM_ERROR(
+		 "filp->private_data->minor->dev->dev_private == NULL\n");
+		return -EINVAL;
+	}
+	QXL_INFO(qdev, "%s: filp->private_data = 0x%p, vma->vm_pgoff = %lx\n",
+		 __func__, filp->private_data, vma->vm_pgoff);
+
+	r = ttm_bo_mmap(filp, vma, &qdev->mman.bdev);
+	if (unlikely(r != 0))
+		return r;
+	if (unlikely(ttm_vm_ops == NULL)) {
+		ttm_vm_ops = vma->vm_ops;
+		qxl_ttm_vm_ops = *ttm_vm_ops;
+		qxl_ttm_vm_ops.fault = &qxl_ttm_fault;
+	}
+	vma->vm_ops = &qxl_ttm_vm_ops;
+	return 0;
+}
+
+static int qxl_invalidate_caches(struct ttm_bo_device *bdev, uint32_t flags)
+{
+	return 0;
+}
+
+static int qxl_init_mem_type(struct ttm_bo_device *bdev, uint32_t type,
+			     struct ttm_mem_type_manager *man)
+{
+	struct qxl_device *qdev;
+
+	qdev = qxl_get_qdev(bdev);
+
+	switch (type) {
+	case TTM_PL_SYSTEM:
+		/* System memory */
+		man->flags = TTM_MEMTYPE_FLAG_MAPPABLE;
+		man->available_caching = TTM_PL_MASK_CACHING;
+		man->default_caching = TTM_PL_FLAG_CACHED;
+		break;
+	case TTM_PL_VRAM:
+	case TTM_PL_PRIV0:
+		/* "On-card" video ram */
+		man->func = &ttm_bo_manager_func;
+		man->gpu_offset = 0;
+		man->flags = TTM_MEMTYPE_FLAG_FIXED |
+			     TTM_MEMTYPE_FLAG_MAPPABLE;
+		man->available_caching = TTM_PL_MASK_CACHING;
+		man->default_caching = TTM_PL_FLAG_CACHED;
+		break;
+	default:
+		DRM_ERROR("Unsupported memory type %u\n", (unsigned)type);
+		return -EINVAL;
+	}
+	return 0;
+}
+
+static void qxl_evict_flags(struct ttm_buffer_object *bo,
+				struct ttm_placement *placement)
+{
+	struct qxl_bo *qbo;
+	static u32 placements = TTM_PL_MASK_CACHING | TTM_PL_FLAG_SYSTEM;
+
+	if (!qxl_ttm_bo_is_qxl_bo(bo)) {
+		placement->fpfn = 0;
+		placement->lpfn = 0;
+		placement->placement = &placements;
+		placement->busy_placement = &placements;
+		placement->num_placement = 1;
+		placement->num_busy_placement = 1;
+		return;
+	}
+	qbo = container_of(bo, struct qxl_bo, tbo);
+	qxl_ttm_placement_from_domain(qbo, QXL_GEM_DOMAIN_CPU);
+	*placement = qbo->placement;
+}
+
+static int qxl_verify_access(struct ttm_buffer_object *bo, struct file *filp)
+{
+	return 0;
+}
+
+static int qxl_ttm_io_mem_reserve(struct ttm_bo_device *bdev,
+				  struct ttm_mem_reg *mem)
+{
+	struct ttm_mem_type_manager *man = &bdev->man[mem->mem_type];
+	struct qxl_device *qdev = qxl_get_qdev(bdev);
+
+	mem->bus.addr = NULL;
+	mem->bus.offset = 0;
+	mem->bus.size = mem->num_pages << PAGE_SHIFT;
+	mem->bus.base = 0;
+	mem->bus.is_iomem = false;
+	if (!(man->flags & TTM_MEMTYPE_FLAG_MAPPABLE))
+		return -EINVAL;
+	switch (mem->mem_type) {
+	case TTM_PL_SYSTEM:
+		/* system memory */
+		return 0;
+	case TTM_PL_VRAM:
+		mem->bus.is_iomem = true;
+		mem->bus.base = qdev->vram_base;
+		mem->bus.offset = mem->start << PAGE_SHIFT;
+		break;
+	case TTM_PL_PRIV0:
+		mem->bus.is_iomem = true;
+		mem->bus.base = qdev->surfaceram_base;
+		mem->bus.offset = mem->start << PAGE_SHIFT;
+		break;
+	default:
+		return -EINVAL;
+	}
+	return 0;
+}
+
+static void qxl_ttm_io_mem_free(struct ttm_bo_device *bdev,
+				struct ttm_mem_reg *mem)
+{
+}
+
+/*
+ * TTM backend functions.
+ */
+struct qxl_ttm_tt {
+	struct ttm_dma_tt		ttm;
+	struct qxl_device		*qdev;
+	u64				offset;
+};
+
+static int qxl_ttm_backend_bind(struct ttm_tt *ttm,
+				struct ttm_mem_reg *bo_mem)
+{
+	struct qxl_ttm_tt *gtt = (void *)ttm;
+
+	gtt->offset = (unsigned long)(bo_mem->start << PAGE_SHIFT);
+	if (!ttm->num_pages) {
+		WARN(1, "nothing to bind %lu pages for mreg %p back %p!\n",
+		     ttm->num_pages, bo_mem, ttm);
+	}
+	/* Not implemented */
+	return -1;
+}
+
+static int qxl_ttm_backend_unbind(struct ttm_tt *ttm)
+{
+	/* Not implemented */
+	return -1;
+}
+
+static void qxl_ttm_backend_destroy(struct ttm_tt *ttm)
+{
+	struct qxl_ttm_tt *gtt = (void *)ttm;
+
+	ttm_dma_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 int qxl_ttm_tt_populate(struct ttm_tt *ttm)
+{
+	int r;
+
+	if (ttm->state != tt_unpopulated)
+		return 0;
+
+	r = ttm_pool_populate(ttm);
+	if (r)
+		return r;
+
+	return 0;
+}
+
+static void qxl_ttm_tt_unpopulate(struct ttm_tt *ttm)
+{
+	ttm_pool_unpopulate(ttm);
+}
+
+static struct ttm_tt *qxl_ttm_tt_create(struct ttm_bo_device *bdev,
+					unsigned long size, uint32_t page_flags,
+					struct page *dummy_read_page)
+{
+	struct qxl_device *qdev;
+	struct qxl_ttm_tt *gtt;
+
+	qdev = qxl_get_qdev(bdev);
+	gtt = kzalloc(sizeof(struct qxl_ttm_tt), GFP_KERNEL);
+	if (gtt == NULL)
+		return NULL;
+	gtt->ttm.ttm.func = &qxl_backend_func;
+	gtt->qdev = qdev;
+	if (ttm_dma_tt_init(&gtt->ttm, bdev, size, page_flags,
+			    dummy_read_page)) {
+		kfree(gtt);
+		return NULL;
+	}
+	return &gtt->ttm.ttm;
+}
+
+static void qxl_move_null(struct ttm_buffer_object *bo,
+			     struct ttm_mem_reg *new_mem)
+{
+	struct ttm_mem_reg *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, bool interruptible,
+		       bool no_wait_gpu,
+		       struct ttm_mem_reg *new_mem)
+{
+	struct ttm_mem_reg *old_mem = &bo->mem;
+	if (old_mem->mem_type == TTM_PL_SYSTEM && bo->ttm == NULL) {
+		qxl_move_null(bo, new_mem);
+		return 0;
+	}
+	return ttm_bo_move_memcpy(bo, evict, no_wait_gpu, new_mem);
+}
+
+
+static int qxl_sync_obj_wait(void *sync_obj,
+			     bool lazy, bool interruptible)
+{
+	struct qxl_fence *qfence = (struct qxl_fence *)sync_obj;
+	int count = 0, sc = 0;
+	struct qxl_bo *bo = container_of(qfence, struct qxl_bo, fence);
+
+	if (qfence->num_active_releases == 0)
+		return 0;
+
+retry:
+	if (sc == 0) {
+		if (bo->type == QXL_GEM_DOMAIN_SURFACE)
+			qxl_update_surface(qfence->qdev, bo);
+	} else if (sc >= 1) {
+		qxl_io_notify_oom(qfence->qdev);
+	}
+
+	sc++;
+
+	for (count = 0; count < 10; count++) {
+		bool ret;
+		ret = qxl_queue_garbage_collect(qfence->qdev, true);
+		if (ret == false)
+			break;
+
+		if (qfence->num_active_releases == 0)
+			return 0;
+	}
+
+	if (qfence->num_active_releases) {
+		bool have_drawable_releases = false;
+		void **slot;
+		struct radix_tree_iter iter;
+		int release_id;
+
+		radix_tree_for_each_slot(slot, &qfence->tree, &iter, 0) {
+			struct qxl_release *release;
+
+			release_id = iter.index;
+			release = qxl_release_from_id_locked(qfence->qdev, release_id);
+			if (release == NULL)
+				continue;
+
+			if (release->type == QXL_RELEASE_DRAWABLE)
+				have_drawable_releases = true;
+		}
+
+		qxl_queue_garbage_collect(qfence->qdev, true);
+
+		if (have_drawable_releases || sc < 4) {
+			if (sc > 2)
+				/* back off */
+				usleep_range(500, 1000);
+			if (have_drawable_releases && sc > 300) {
+				WARN(1, "sync obj %d still has outstanding releases %d %d %d %ld %d\n", sc, bo->surface_id, bo->is_primary, bo->pin_count, (unsigned long)bo->gem_base.size, qfence->num_active_releases);
+				return -EBUSY;
+			}
+			goto retry;
+		}
+	}
+	return 0;
+}
+
+static int qxl_sync_obj_flush(void *sync_obj)
+{
+	return 0;
+}
+
+static void qxl_sync_obj_unref(void **sync_obj)
+{
+}
+
+static void *qxl_sync_obj_ref(void *sync_obj)
+{
+	return sync_obj;
+}
+
+static bool qxl_sync_obj_signaled(void *sync_obj)
+{
+	struct qxl_fence *qfence = (struct qxl_fence *)sync_obj;
+	return (qfence->num_active_releases == 0);
+}
+
+static void qxl_bo_move_notify(struct ttm_buffer_object *bo,
+			       struct ttm_mem_reg *new_mem)
+{
+	struct qxl_bo *qbo;
+	struct qxl_device *qdev;
+
+	if (!qxl_ttm_bo_is_qxl_bo(bo))
+		return;
+	qbo = container_of(bo, struct qxl_bo, tbo);
+	qdev = qbo->gem_base.dev->dev_private;
+
+	if (bo->mem.mem_type == TTM_PL_PRIV0 && qbo->surface_id)
+		qxl_surface_evict(qdev, qbo, new_mem ? true : false);
+}
+
+static struct ttm_bo_driver qxl_bo_driver = {
+	.ttm_tt_create = &qxl_ttm_tt_create,
+	.ttm_tt_populate = &qxl_ttm_tt_populate,
+	.ttm_tt_unpopulate = &qxl_ttm_tt_unpopulate,
+	.invalidate_caches = &qxl_invalidate_caches,
+	.init_mem_type = &qxl_init_mem_type,
+	.evict_flags = &qxl_evict_flags,
+	.move = &qxl_bo_move,
+	.verify_access = &qxl_verify_access,
+	.io_mem_reserve = &qxl_ttm_io_mem_reserve,
+	.io_mem_free = &qxl_ttm_io_mem_free,
+	.sync_obj_signaled = &qxl_sync_obj_signaled,
+	.sync_obj_wait = &qxl_sync_obj_wait,
+	.sync_obj_flush = &qxl_sync_obj_flush,
+	.sync_obj_unref = &qxl_sync_obj_unref,
+	.sync_obj_ref = &qxl_sync_obj_ref,
+	.move_notify = &qxl_bo_move_notify,
+};
+
+
+
+int qxl_ttm_init(struct qxl_device *qdev)
+{
+	int r;
+	int num_io_pages; /* != rom->num_io_pages, we include surface0 */
+
+	r = qxl_ttm_global_init(qdev);
+	if (r)
+		return r;
+	/* No others user of address space so set it to 0 */
+	r = ttm_bo_device_init(&qdev->mman.bdev,
+			       qdev->mman.bo_global_ref.ref.object,
+			       &qxl_bo_driver, DRM_FILE_PAGE_OFFSET, 0);
+	if (r) {
+		DRM_ERROR("failed initializing buffer object driver(%d).\n", r);
+		return r;
+	}
+	/* NOTE: this includes the framebuffer (aka surface 0) */
+	num_io_pages = qdev->rom->ram_header_offset / PAGE_SIZE;
+	r = ttm_bo_init_mm(&qdev->mman.bdev, TTM_PL_VRAM,
+			   num_io_pages);
+	if (r) {
+		DRM_ERROR("Failed initializing VRAM heap.\n");
+		return r;
+	}
+	r = ttm_bo_init_mm(&qdev->mman.bdev, TTM_PL_PRIV0,
+			   qdev->surfaceram_size / PAGE_SIZE);
+	if (r) {
+		DRM_ERROR("Failed initializing Surfaces heap.\n");
+		return r;
+	}
+	DRM_INFO("qxl: %uM of VRAM memory size\n",
+		 (unsigned)qdev->vram_size / (1024 * 1024));
+	DRM_INFO("qxl: %luM of IO pages memory ready (VRAM domain)\n",
+		 ((unsigned)num_io_pages * PAGE_SIZE) / (1024 * 1024));
+	if (unlikely(qdev->mman.bdev.dev_mapping == NULL))
+		qdev->mman.bdev.dev_mapping = qdev->ddev->dev_mapping;
+	r = qxl_ttm_debugfs_init(qdev);
+	if (r) {
+		DRM_ERROR("Failed to init debugfs\n");
+		return r;
+	}
+	return 0;
+}
+
+void qxl_ttm_fini(struct qxl_device *qdev)
+{
+	ttm_bo_clean_mm(&qdev->mman.bdev, TTM_PL_VRAM);
+	ttm_bo_clean_mm(&qdev->mman.bdev, TTM_PL_PRIV0);
+	ttm_bo_device_release(&qdev->mman.bdev);
+	qxl_ttm_global_fini(qdev);
+	DRM_INFO("qxl: ttm finalized\n");
+}
+
+
+#define QXL_DEBUGFS_MEM_TYPES 2
+
+#if defined(CONFIG_DEBUG_FS)
+static int qxl_mm_dump_table(struct seq_file *m, void *data)
+{
+	struct drm_info_node *node = (struct drm_info_node *)m->private;
+	struct drm_mm *mm = (struct drm_mm *)node->info_ent->data;
+	struct drm_device *dev = node->minor->dev;
+	struct qxl_device *rdev = dev->dev_private;
+	int ret;
+	struct ttm_bo_global *glob = rdev->mman.bdev.glob;
+
+	spin_lock(&glob->lru_lock);
+	ret = drm_mm_dump_table(m, mm);
+	spin_unlock(&glob->lru_lock);
+	return ret;
+}
+#endif
+
+static int qxl_ttm_debugfs_init(struct qxl_device *qdev)
+{
+#if defined(CONFIG_DEBUG_FS)
+	static struct drm_info_list qxl_mem_types_list[QXL_DEBUGFS_MEM_TYPES];
+	static char qxl_mem_types_names[QXL_DEBUGFS_MEM_TYPES][32];
+	unsigned i;
+
+	for (i = 0; i < QXL_DEBUGFS_MEM_TYPES; i++) {
+		if (i == 0)
+			sprintf(qxl_mem_types_names[i], "qxl_mem_mm");
+		else
+			sprintf(qxl_mem_types_names[i], "qxl_surf_mm");
+		qxl_mem_types_list[i].name = qxl_mem_types_names[i];
+		qxl_mem_types_list[i].show = &qxl_mm_dump_table;
+		qxl_mem_types_list[i].driver_features = 0;
+		if (i == 0)
+			qxl_mem_types_list[i].data = qdev->mman.bdev.man[TTM_PL_VRAM].priv;
+		else
+			qxl_mem_types_list[i].data = qdev->mman.bdev.man[TTM_PL_PRIV0].priv;
+
+	}
+	return qxl_debugfs_add_files(qdev, qxl_mem_types_list, i);
+#else
+	return 0;
+#endif
+}
diff --git a/drivers/gpu/drm/radeon/Makefile b/drivers/gpu/drm/radeon/Makefile
index bf172522ea68..86c5e3611892 100644
--- a/drivers/gpu/drm/radeon/Makefile
+++ b/drivers/gpu/drm/radeon/Makefile
@@ -76,7 +76,7 @@ radeon-y += radeon_device.o radeon_asic.o radeon_kms.o \
 	evergreen.o evergreen_cs.o evergreen_blit_shaders.o evergreen_blit_kms.o \
 	evergreen_hdmi.o radeon_trace_points.o ni.o cayman_blit_shaders.o \
 	atombios_encoders.o radeon_semaphore.o radeon_sa.o atombios_i2c.o si.o \
-	si_blit_shaders.o radeon_prime.o
+	si_blit_shaders.o radeon_prime.o radeon_uvd.o
 
 radeon-$(CONFIG_COMPAT) += radeon_ioc32.o
 radeon-$(CONFIG_VGA_SWITCHEROO) += radeon_atpx_handler.o
diff --git a/drivers/gpu/drm/radeon/atom.c b/drivers/gpu/drm/radeon/atom.c
index 46a9c3772850..fb441a790f3d 100644
--- a/drivers/gpu/drm/radeon/atom.c
+++ b/drivers/gpu/drm/radeon/atom.c
@@ -1394,10 +1394,10 @@ int atom_allocate_fb_scratch(struct atom_context *ctx)
 		firmware_usage = (struct _ATOM_VRAM_USAGE_BY_FIRMWARE *)(ctx->bios + data_offset);
 
 		DRM_DEBUG("atom firmware requested %08x %dkb\n",
-			  firmware_usage->asFirmwareVramReserveInfo[0].ulStartAddrUsedByFirmware,
-			  firmware_usage->asFirmwareVramReserveInfo[0].usFirmwareUseInKb);
+			  le32_to_cpu(firmware_usage->asFirmwareVramReserveInfo[0].ulStartAddrUsedByFirmware),
+			  le16_to_cpu(firmware_usage->asFirmwareVramReserveInfo[0].usFirmwareUseInKb));
 
-		usage_bytes = firmware_usage->asFirmwareVramReserveInfo[0].usFirmwareUseInKb * 1024;
+		usage_bytes = le16_to_cpu(firmware_usage->asFirmwareVramReserveInfo[0].usFirmwareUseInKb) * 1024;
 	}
 	ctx->scratch_size_bytes = 0;
 	if (usage_bytes == 0)
diff --git a/drivers/gpu/drm/radeon/atombios.h b/drivers/gpu/drm/radeon/atombios.h
index 4b04ba3828e8..0ee573743de9 100644
--- a/drivers/gpu/drm/radeon/atombios.h
+++ b/drivers/gpu/drm/radeon/atombios.h
@@ -458,6 +458,7 @@ typedef struct _COMPUTE_MEMORY_ENGINE_PLL_PARAMETERS_V3
   union
   {
     ATOM_COMPUTE_CLOCK_FREQ  ulClock;         //Input Parameter
+    ULONG ulClockParams;                      //ULONG access for BE
     ATOM_S_MPLL_FB_DIVIDER   ulFbDiv;         //Output Parameter
   };
   UCHAR   ucRefDiv;                           //Output Parameter      
@@ -490,6 +491,7 @@ typedef struct _COMPUTE_MEMORY_ENGINE_PLL_PARAMETERS_V5
   union
   {
     ATOM_COMPUTE_CLOCK_FREQ  ulClock;         //Input Parameter
+    ULONG ulClockParams;                      //ULONG access for BE
     ATOM_S_MPLL_FB_DIVIDER   ulFbDiv;         //Output Parameter
   };
   UCHAR   ucRefDiv;                           //Output Parameter      
diff --git a/drivers/gpu/drm/radeon/atombios_crtc.c b/drivers/gpu/drm/radeon/atombios_crtc.c
index 21a892c6ab9c..6d6fdb3ba0d0 100644
--- a/drivers/gpu/drm/radeon/atombios_crtc.c
+++ b/drivers/gpu/drm/radeon/atombios_crtc.c
@@ -557,6 +557,9 @@ static u32 atombios_adjust_pll(struct drm_crtc *crtc,
 		/* use frac fb div on APUs */
 		if (ASIC_IS_DCE41(rdev) || ASIC_IS_DCE61(rdev))
 			radeon_crtc->pll_flags |= RADEON_PLL_USE_FRAC_FB_DIV;
+		/* use frac fb div on RS780/RS880 */
+		if ((rdev->family == CHIP_RS780) || (rdev->family == CHIP_RS880))
+			radeon_crtc->pll_flags |= RADEON_PLL_USE_FRAC_FB_DIV;
 		if (ASIC_IS_DCE32(rdev) && mode->clock > 165000)
 			radeon_crtc->pll_flags |= RADEON_PLL_USE_FRAC_FB_DIV;
 	} else {
diff --git a/drivers/gpu/drm/radeon/atombios_encoders.c b/drivers/gpu/drm/radeon/atombios_encoders.c
index 4552d4aff317..44a7da66e081 100644
--- a/drivers/gpu/drm/radeon/atombios_encoders.c
+++ b/drivers/gpu/drm/radeon/atombios_encoders.c
@@ -2150,13 +2150,10 @@ radeon_atom_encoder_mode_set(struct drm_encoder *encoder,
 	atombios_apply_encoder_quirks(encoder, adjusted_mode);
 
 	if (atombios_get_encoder_mode(encoder) == ATOM_ENCODER_MODE_HDMI) {
-		r600_hdmi_enable(encoder);
-		if (ASIC_IS_DCE6(rdev))
-			; /* TODO (use pointers instead of if-s?) */
-		else if (ASIC_IS_DCE4(rdev))
-			evergreen_hdmi_setmode(encoder, adjusted_mode);
-		else
-			r600_hdmi_setmode(encoder, adjusted_mode);
+		if (rdev->asic->display.hdmi_enable)
+			radeon_hdmi_enable(rdev, encoder, true);
+		if (rdev->asic->display.hdmi_setmode)
+			radeon_hdmi_setmode(rdev, encoder, adjusted_mode);
 	}
 }
 
@@ -2413,8 +2410,10 @@ static void radeon_atom_encoder_disable(struct drm_encoder *encoder)
 
 disable_done:
 	if (radeon_encoder_is_digital(encoder)) {
-		if (atombios_get_encoder_mode(encoder) == ATOM_ENCODER_MODE_HDMI)
-			r600_hdmi_disable(encoder);
+		if (atombios_get_encoder_mode(encoder) == ATOM_ENCODER_MODE_HDMI) {
+			if (rdev->asic->display.hdmi_enable)
+				radeon_hdmi_enable(rdev, encoder, false);
+		}
 		dig = radeon_encoder->enc_priv;
 		dig->dig_encoder = -1;
 	}
diff --git a/drivers/gpu/drm/radeon/evergreen.c b/drivers/gpu/drm/radeon/evergreen.c
index 305a657bf215..105bafb6c29d 100644
--- a/drivers/gpu/drm/radeon/evergreen.c
+++ b/drivers/gpu/drm/radeon/evergreen.c
@@ -53,6 +53,864 @@ void evergreen_pcie_gen2_enable(struct radeon_device *rdev);
 extern void cayman_cp_int_cntl_setup(struct radeon_device *rdev,
 				     int ring, u32 cp_int_cntl);
 
+static const u32 evergreen_golden_registers[] =
+{
+	0x3f90, 0xffff0000, 0xff000000,
+	0x9148, 0xffff0000, 0xff000000,
+	0x3f94, 0xffff0000, 0xff000000,
+	0x914c, 0xffff0000, 0xff000000,
+	0x9b7c, 0xffffffff, 0x00000000,
+	0x8a14, 0xffffffff, 0x00000007,
+	0x8b10, 0xffffffff, 0x00000000,
+	0x960c, 0xffffffff, 0x54763210,
+	0x88c4, 0xffffffff, 0x000000c2,
+	0x88d4, 0xffffffff, 0x00000010,
+	0x8974, 0xffffffff, 0x00000000,
+	0xc78, 0x00000080, 0x00000080,
+	0x5eb4, 0xffffffff, 0x00000002,
+	0x5e78, 0xffffffff, 0x001000f0,
+	0x6104, 0x01000300, 0x00000000,
+	0x5bc0, 0x00300000, 0x00000000,
+	0x7030, 0xffffffff, 0x00000011,
+	0x7c30, 0xffffffff, 0x00000011,
+	0x10830, 0xffffffff, 0x00000011,
+	0x11430, 0xffffffff, 0x00000011,
+	0x12030, 0xffffffff, 0x00000011,
+	0x12c30, 0xffffffff, 0x00000011,
+	0xd02c, 0xffffffff, 0x08421000,
+	0x240c, 0xffffffff, 0x00000380,
+	0x8b24, 0xffffffff, 0x00ff0fff,
+	0x28a4c, 0x06000000, 0x06000000,
+	0x10c, 0x00000001, 0x00000001,
+	0x8d00, 0xffffffff, 0x100e4848,
+	0x8d04, 0xffffffff, 0x00164745,
+	0x8c00, 0xffffffff, 0xe4000003,
+	0x8c04, 0xffffffff, 0x40600060,
+	0x8c08, 0xffffffff, 0x001c001c,
+	0x8cf0, 0xffffffff, 0x08e00620,
+	0x8c20, 0xffffffff, 0x00800080,
+	0x8c24, 0xffffffff, 0x00800080,
+	0x8c18, 0xffffffff, 0x20202078,
+	0x8c1c, 0xffffffff, 0x00001010,
+	0x28350, 0xffffffff, 0x00000000,
+	0xa008, 0xffffffff, 0x00010000,
+	0x5cc, 0xffffffff, 0x00000001,
+	0x9508, 0xffffffff, 0x00000002,
+	0x913c, 0x0000000f, 0x0000000a
+};
+
+static const u32 evergreen_golden_registers2[] =
+{
+	0x2f4c, 0xffffffff, 0x00000000,
+	0x54f4, 0xffffffff, 0x00000000,
+	0x54f0, 0xffffffff, 0x00000000,
+	0x5498, 0xffffffff, 0x00000000,
+	0x549c, 0xffffffff, 0x00000000,
+	0x5494, 0xffffffff, 0x00000000,
+	0x53cc, 0xffffffff, 0x00000000,
+	0x53c8, 0xffffffff, 0x00000000,
+	0x53c4, 0xffffffff, 0x00000000,
+	0x53c0, 0xffffffff, 0x00000000,
+	0x53bc, 0xffffffff, 0x00000000,
+	0x53b8, 0xffffffff, 0x00000000,
+	0x53b4, 0xffffffff, 0x00000000,
+	0x53b0, 0xffffffff, 0x00000000
+};
+
+static const u32 cypress_mgcg_init[] =
+{
+	0x802c, 0xffffffff, 0xc0000000,
+	0x5448, 0xffffffff, 0x00000100,
+	0x55e4, 0xffffffff, 0x00000100,
+	0x160c, 0xffffffff, 0x00000100,
+	0x5644, 0xffffffff, 0x00000100,
+	0xc164, 0xffffffff, 0x00000100,
+	0x8a18, 0xffffffff, 0x00000100,
+	0x897c, 0xffffffff, 0x06000100,
+	0x8b28, 0xffffffff, 0x00000100,
+	0x9144, 0xffffffff, 0x00000100,
+	0x9a60, 0xffffffff, 0x00000100,
+	0x9868, 0xffffffff, 0x00000100,
+	0x8d58, 0xffffffff, 0x00000100,
+	0x9510, 0xffffffff, 0x00000100,
+	0x949c, 0xffffffff, 0x00000100,
+	0x9654, 0xffffffff, 0x00000100,
+	0x9030, 0xffffffff, 0x00000100,
+	0x9034, 0xffffffff, 0x00000100,
+	0x9038, 0xffffffff, 0x00000100,
+	0x903c, 0xffffffff, 0x00000100,
+	0x9040, 0xffffffff, 0x00000100,
+	0xa200, 0xffffffff, 0x00000100,
+	0xa204, 0xffffffff, 0x00000100,
+	0xa208, 0xffffffff, 0x00000100,
+	0xa20c, 0xffffffff, 0x00000100,
+	0x971c, 0xffffffff, 0x00000100,
+	0x977c, 0xffffffff, 0x00000100,
+	0x3f80, 0xffffffff, 0x00000100,
+	0xa210, 0xffffffff, 0x00000100,
+	0xa214, 0xffffffff, 0x00000100,
+	0x4d8, 0xffffffff, 0x00000100,
+	0x9784, 0xffffffff, 0x00000100,
+	0x9698, 0xffffffff, 0x00000100,
+	0x4d4, 0xffffffff, 0x00000200,
+	0x30cc, 0xffffffff, 0x00000100,
+	0xd0c0, 0xffffffff, 0xff000100,
+	0x802c, 0xffffffff, 0x40000000,
+	0x915c, 0xffffffff, 0x00010000,
+	0x9160, 0xffffffff, 0x00030002,
+	0x9178, 0xffffffff, 0x00070000,
+	0x917c, 0xffffffff, 0x00030002,
+	0x9180, 0xffffffff, 0x00050004,
+	0x918c, 0xffffffff, 0x00010006,
+	0x9190, 0xffffffff, 0x00090008,
+	0x9194, 0xffffffff, 0x00070000,
+	0x9198, 0xffffffff, 0x00030002,
+	0x919c, 0xffffffff, 0x00050004,
+	0x91a8, 0xffffffff, 0x00010006,
+	0x91ac, 0xffffffff, 0x00090008,
+	0x91b0, 0xffffffff, 0x00070000,
+	0x91b4, 0xffffffff, 0x00030002,
+	0x91b8, 0xffffffff, 0x00050004,
+	0x91c4, 0xffffffff, 0x00010006,
+	0x91c8, 0xffffffff, 0x00090008,
+	0x91cc, 0xffffffff, 0x00070000,
+	0x91d0, 0xffffffff, 0x00030002,
+	0x91d4, 0xffffffff, 0x00050004,
+	0x91e0, 0xffffffff, 0x00010006,
+	0x91e4, 0xffffffff, 0x00090008,
+	0x91e8, 0xffffffff, 0x00000000,
+	0x91ec, 0xffffffff, 0x00070000,
+	0x91f0, 0xffffffff, 0x00030002,
+	0x91f4, 0xffffffff, 0x00050004,
+	0x9200, 0xffffffff, 0x00010006,
+	0x9204, 0xffffffff, 0x00090008,
+	0x9208, 0xffffffff, 0x00070000,
+	0x920c, 0xffffffff, 0x00030002,
+	0x9210, 0xffffffff, 0x00050004,
+	0x921c, 0xffffffff, 0x00010006,
+	0x9220, 0xffffffff, 0x00090008,
+	0x9224, 0xffffffff, 0x00070000,
+	0x9228, 0xffffffff, 0x00030002,
+	0x922c, 0xffffffff, 0x00050004,
+	0x9238, 0xffffffff, 0x00010006,
+	0x923c, 0xffffffff, 0x00090008,
+	0x9240, 0xffffffff, 0x00070000,
+	0x9244, 0xffffffff, 0x00030002,
+	0x9248, 0xffffffff, 0x00050004,
+	0x9254, 0xffffffff, 0x00010006,
+	0x9258, 0xffffffff, 0x00090008,
+	0x925c, 0xffffffff, 0x00070000,
+	0x9260, 0xffffffff, 0x00030002,
+	0x9264, 0xffffffff, 0x00050004,
+	0x9270, 0xffffffff, 0x00010006,
+	0x9274, 0xffffffff, 0x00090008,
+	0x9278, 0xffffffff, 0x00070000,
+	0x927c, 0xffffffff, 0x00030002,
+	0x9280, 0xffffffff, 0x00050004,
+	0x928c, 0xffffffff, 0x00010006,
+	0x9290, 0xffffffff, 0x00090008,
+	0x9294, 0xffffffff, 0x00000000,
+	0x929c, 0xffffffff, 0x00000001,
+	0x802c, 0xffffffff, 0x40010000,
+	0x915c, 0xffffffff, 0x00010000,
+	0x9160, 0xffffffff, 0x00030002,
+	0x9178, 0xffffffff, 0x00070000,
+	0x917c, 0xffffffff, 0x00030002,
+	0x9180, 0xffffffff, 0x00050004,
+	0x918c, 0xffffffff, 0x00010006,
+	0x9190, 0xffffffff, 0x00090008,
+	0x9194, 0xffffffff, 0x00070000,
+	0x9198, 0xffffffff, 0x00030002,
+	0x919c, 0xffffffff, 0x00050004,
+	0x91a8, 0xffffffff, 0x00010006,
+	0x91ac, 0xffffffff, 0x00090008,
+	0x91b0, 0xffffffff, 0x00070000,
+	0x91b4, 0xffffffff, 0x00030002,
+	0x91b8, 0xffffffff, 0x00050004,
+	0x91c4, 0xffffffff, 0x00010006,
+	0x91c8, 0xffffffff, 0x00090008,
+	0x91cc, 0xffffffff, 0x00070000,
+	0x91d0, 0xffffffff, 0x00030002,
+	0x91d4, 0xffffffff, 0x00050004,
+	0x91e0, 0xffffffff, 0x00010006,
+	0x91e4, 0xffffffff, 0x00090008,
+	0x91e8, 0xffffffff, 0x00000000,
+	0x91ec, 0xffffffff, 0x00070000,
+	0x91f0, 0xffffffff, 0x00030002,
+	0x91f4, 0xffffffff, 0x00050004,
+	0x9200, 0xffffffff, 0x00010006,
+	0x9204, 0xffffffff, 0x00090008,
+	0x9208, 0xffffffff, 0x00070000,
+	0x920c, 0xffffffff, 0x00030002,
+	0x9210, 0xffffffff, 0x00050004,
+	0x921c, 0xffffffff, 0x00010006,
+	0x9220, 0xffffffff, 0x00090008,
+	0x9224, 0xffffffff, 0x00070000,
+	0x9228, 0xffffffff, 0x00030002,
+	0x922c, 0xffffffff, 0x00050004,
+	0x9238, 0xffffffff, 0x00010006,
+	0x923c, 0xffffffff, 0x00090008,
+	0x9240, 0xffffffff, 0x00070000,
+	0x9244, 0xffffffff, 0x00030002,
+	0x9248, 0xffffffff, 0x00050004,
+	0x9254, 0xffffffff, 0x00010006,
+	0x9258, 0xffffffff, 0x00090008,
+	0x925c, 0xffffffff, 0x00070000,
+	0x9260, 0xffffffff, 0x00030002,
+	0x9264, 0xffffffff, 0x00050004,
+	0x9270, 0xffffffff, 0x00010006,
+	0x9274, 0xffffffff, 0x00090008,
+	0x9278, 0xffffffff, 0x00070000,
+	0x927c, 0xffffffff, 0x00030002,
+	0x9280, 0xffffffff, 0x00050004,
+	0x928c, 0xffffffff, 0x00010006,
+	0x9290, 0xffffffff, 0x00090008,
+	0x9294, 0xffffffff, 0x00000000,
+	0x929c, 0xffffffff, 0x00000001,
+	0x802c, 0xffffffff, 0xc0000000
+};
+
+static const u32 redwood_mgcg_init[] =
+{
+	0x802c, 0xffffffff, 0xc0000000,
+	0x5448, 0xffffffff, 0x00000100,
+	0x55e4, 0xffffffff, 0x00000100,
+	0x160c, 0xffffffff, 0x00000100,
+	0x5644, 0xffffffff, 0x00000100,
+	0xc164, 0xffffffff, 0x00000100,
+	0x8a18, 0xffffffff, 0x00000100,
+	0x897c, 0xffffffff, 0x06000100,
+	0x8b28, 0xffffffff, 0x00000100,
+	0x9144, 0xffffffff, 0x00000100,
+	0x9a60, 0xffffffff, 0x00000100,
+	0x9868, 0xffffffff, 0x00000100,
+	0x8d58, 0xffffffff, 0x00000100,
+	0x9510, 0xffffffff, 0x00000100,
+	0x949c, 0xffffffff, 0x00000100,
+	0x9654, 0xffffffff, 0x00000100,
+	0x9030, 0xffffffff, 0x00000100,
+	0x9034, 0xffffffff, 0x00000100,
+	0x9038, 0xffffffff, 0x00000100,
+	0x903c, 0xffffffff, 0x00000100,
+	0x9040, 0xffffffff, 0x00000100,
+	0xa200, 0xffffffff, 0x00000100,
+	0xa204, 0xffffffff, 0x00000100,
+	0xa208, 0xffffffff, 0x00000100,
+	0xa20c, 0xffffffff, 0x00000100,
+	0x971c, 0xffffffff, 0x00000100,
+	0x977c, 0xffffffff, 0x00000100,
+	0x3f80, 0xffffffff, 0x00000100,
+	0xa210, 0xffffffff, 0x00000100,
+	0xa214, 0xffffffff, 0x00000100,
+	0x4d8, 0xffffffff, 0x00000100,
+	0x9784, 0xffffffff, 0x00000100,
+	0x9698, 0xffffffff, 0x00000100,
+	0x4d4, 0xffffffff, 0x00000200,
+	0x30cc, 0xffffffff, 0x00000100,
+	0xd0c0, 0xffffffff, 0xff000100,
+	0x802c, 0xffffffff, 0x40000000,
+	0x915c, 0xffffffff, 0x00010000,
+	0x9160, 0xffffffff, 0x00030002,
+	0x9178, 0xffffffff, 0x00070000,
+	0x917c, 0xffffffff, 0x00030002,
+	0x9180, 0xffffffff, 0x00050004,
+	0x918c, 0xffffffff, 0x00010006,
+	0x9190, 0xffffffff, 0x00090008,
+	0x9194, 0xffffffff, 0x00070000,
+	0x9198, 0xffffffff, 0x00030002,
+	0x919c, 0xffffffff, 0x00050004,
+	0x91a8, 0xffffffff, 0x00010006,
+	0x91ac, 0xffffffff, 0x00090008,
+	0x91b0, 0xffffffff, 0x00070000,
+	0x91b4, 0xffffffff, 0x00030002,
+	0x91b8, 0xffffffff, 0x00050004,
+	0x91c4, 0xffffffff, 0x00010006,
+	0x91c8, 0xffffffff, 0x00090008,
+	0x91cc, 0xffffffff, 0x00070000,
+	0x91d0, 0xffffffff, 0x00030002,
+	0x91d4, 0xffffffff, 0x00050004,
+	0x91e0, 0xffffffff, 0x00010006,
+	0x91e4, 0xffffffff, 0x00090008,
+	0x91e8, 0xffffffff, 0x00000000,
+	0x91ec, 0xffffffff, 0x00070000,
+	0x91f0, 0xffffffff, 0x00030002,
+	0x91f4, 0xffffffff, 0x00050004,
+	0x9200, 0xffffffff, 0x00010006,
+	0x9204, 0xffffffff, 0x00090008,
+	0x9294, 0xffffffff, 0x00000000,
+	0x929c, 0xffffffff, 0x00000001,
+	0x802c, 0xffffffff, 0xc0000000
+};
+
+static const u32 cedar_golden_registers[] =
+{
+	0x3f90, 0xffff0000, 0xff000000,
+	0x9148, 0xffff0000, 0xff000000,
+	0x3f94, 0xffff0000, 0xff000000,
+	0x914c, 0xffff0000, 0xff000000,
+	0x9b7c, 0xffffffff, 0x00000000,
+	0x8a14, 0xffffffff, 0x00000007,
+	0x8b10, 0xffffffff, 0x00000000,
+	0x960c, 0xffffffff, 0x54763210,
+	0x88c4, 0xffffffff, 0x000000c2,
+	0x88d4, 0xffffffff, 0x00000000,
+	0x8974, 0xffffffff, 0x00000000,
+	0xc78, 0x00000080, 0x00000080,
+	0x5eb4, 0xffffffff, 0x00000002,
+	0x5e78, 0xffffffff, 0x001000f0,
+	0x6104, 0x01000300, 0x00000000,
+	0x5bc0, 0x00300000, 0x00000000,
+	0x7030, 0xffffffff, 0x00000011,
+	0x7c30, 0xffffffff, 0x00000011,
+	0x10830, 0xffffffff, 0x00000011,
+	0x11430, 0xffffffff, 0x00000011,
+	0xd02c, 0xffffffff, 0x08421000,
+	0x240c, 0xffffffff, 0x00000380,
+	0x8b24, 0xffffffff, 0x00ff0fff,
+	0x28a4c, 0x06000000, 0x06000000,
+	0x10c, 0x00000001, 0x00000001,
+	0x8d00, 0xffffffff, 0x100e4848,
+	0x8d04, 0xffffffff, 0x00164745,
+	0x8c00, 0xffffffff, 0xe4000003,
+	0x8c04, 0xffffffff, 0x40600060,
+	0x8c08, 0xffffffff, 0x001c001c,
+	0x8cf0, 0xffffffff, 0x08e00410,
+	0x8c20, 0xffffffff, 0x00800080,
+	0x8c24, 0xffffffff, 0x00800080,
+	0x8c18, 0xffffffff, 0x20202078,
+	0x8c1c, 0xffffffff, 0x00001010,
+	0x28350, 0xffffffff, 0x00000000,
+	0xa008, 0xffffffff, 0x00010000,
+	0x5cc, 0xffffffff, 0x00000001,
+	0x9508, 0xffffffff, 0x00000002
+};
+
+static const u32 cedar_mgcg_init[] =
+{
+	0x802c, 0xffffffff, 0xc0000000,
+	0x5448, 0xffffffff, 0x00000100,
+	0x55e4, 0xffffffff, 0x00000100,
+	0x160c, 0xffffffff, 0x00000100,
+	0x5644, 0xffffffff, 0x00000100,
+	0xc164, 0xffffffff, 0x00000100,
+	0x8a18, 0xffffffff, 0x00000100,
+	0x897c, 0xffffffff, 0x06000100,
+	0x8b28, 0xffffffff, 0x00000100,
+	0x9144, 0xffffffff, 0x00000100,
+	0x9a60, 0xffffffff, 0x00000100,
+	0x9868, 0xffffffff, 0x00000100,
+	0x8d58, 0xffffffff, 0x00000100,
+	0x9510, 0xffffffff, 0x00000100,
+	0x949c, 0xffffffff, 0x00000100,
+	0x9654, 0xffffffff, 0x00000100,
+	0x9030, 0xffffffff, 0x00000100,
+	0x9034, 0xffffffff, 0x00000100,
+	0x9038, 0xffffffff, 0x00000100,
+	0x903c, 0xffffffff, 0x00000100,
+	0x9040, 0xffffffff, 0x00000100,
+	0xa200, 0xffffffff, 0x00000100,
+	0xa204, 0xffffffff, 0x00000100,
+	0xa208, 0xffffffff, 0x00000100,
+	0xa20c, 0xffffffff, 0x00000100,
+	0x971c, 0xffffffff, 0x00000100,
+	0x977c, 0xffffffff, 0x00000100,
+	0x3f80, 0xffffffff, 0x00000100,
+	0xa210, 0xffffffff, 0x00000100,
+	0xa214, 0xffffffff, 0x00000100,
+	0x4d8, 0xffffffff, 0x00000100,
+	0x9784, 0xffffffff, 0x00000100,
+	0x9698, 0xffffffff, 0x00000100,
+	0x4d4, 0xffffffff, 0x00000200,
+	0x30cc, 0xffffffff, 0x00000100,
+	0xd0c0, 0xffffffff, 0xff000100,
+	0x802c, 0xffffffff, 0x40000000,
+	0x915c, 0xffffffff, 0x00010000,
+	0x9178, 0xffffffff, 0x00050000,
+	0x917c, 0xffffffff, 0x00030002,
+	0x918c, 0xffffffff, 0x00010004,
+	0x9190, 0xffffffff, 0x00070006,
+	0x9194, 0xffffffff, 0x00050000,
+	0x9198, 0xffffffff, 0x00030002,
+	0x91a8, 0xffffffff, 0x00010004,
+	0x91ac, 0xffffffff, 0x00070006,
+	0x91e8, 0xffffffff, 0x00000000,
+	0x9294, 0xffffffff, 0x00000000,
+	0x929c, 0xffffffff, 0x00000001,
+	0x802c, 0xffffffff, 0xc0000000
+};
+
+static const u32 juniper_mgcg_init[] =
+{
+	0x802c, 0xffffffff, 0xc0000000,
+	0x5448, 0xffffffff, 0x00000100,
+	0x55e4, 0xffffffff, 0x00000100,
+	0x160c, 0xffffffff, 0x00000100,
+	0x5644, 0xffffffff, 0x00000100,
+	0xc164, 0xffffffff, 0x00000100,
+	0x8a18, 0xffffffff, 0x00000100,
+	0x897c, 0xffffffff, 0x06000100,
+	0x8b28, 0xffffffff, 0x00000100,
+	0x9144, 0xffffffff, 0x00000100,
+	0x9a60, 0xffffffff, 0x00000100,
+	0x9868, 0xffffffff, 0x00000100,
+	0x8d58, 0xffffffff, 0x00000100,
+	0x9510, 0xffffffff, 0x00000100,
+	0x949c, 0xffffffff, 0x00000100,
+	0x9654, 0xffffffff, 0x00000100,
+	0x9030, 0xffffffff, 0x00000100,
+	0x9034, 0xffffffff, 0x00000100,
+	0x9038, 0xffffffff, 0x00000100,
+	0x903c, 0xffffffff, 0x00000100,
+	0x9040, 0xffffffff, 0x00000100,
+	0xa200, 0xffffffff, 0x00000100,
+	0xa204, 0xffffffff, 0x00000100,
+	0xa208, 0xffffffff, 0x00000100,
+	0xa20c, 0xffffffff, 0x00000100,
+	0x971c, 0xffffffff, 0x00000100,
+	0xd0c0, 0xffffffff, 0xff000100,
+	0x802c, 0xffffffff, 0x40000000,
+	0x915c, 0xffffffff, 0x00010000,
+	0x9160, 0xffffffff, 0x00030002,
+	0x9178, 0xffffffff, 0x00070000,
+	0x917c, 0xffffffff, 0x00030002,
+	0x9180, 0xffffffff, 0x00050004,
+	0x918c, 0xffffffff, 0x00010006,
+	0x9190, 0xffffffff, 0x00090008,
+	0x9194, 0xffffffff, 0x00070000,
+	0x9198, 0xffffffff, 0x00030002,
+	0x919c, 0xffffffff, 0x00050004,
+	0x91a8, 0xffffffff, 0x00010006,
+	0x91ac, 0xffffffff, 0x00090008,
+	0x91b0, 0xffffffff, 0x00070000,
+	0x91b4, 0xffffffff, 0x00030002,
+	0x91b8, 0xffffffff, 0x00050004,
+	0x91c4, 0xffffffff, 0x00010006,
+	0x91c8, 0xffffffff, 0x00090008,
+	0x91cc, 0xffffffff, 0x00070000,
+	0x91d0, 0xffffffff, 0x00030002,
+	0x91d4, 0xffffffff, 0x00050004,
+	0x91e0, 0xffffffff, 0x00010006,
+	0x91e4, 0xffffffff, 0x00090008,
+	0x91e8, 0xffffffff, 0x00000000,
+	0x91ec, 0xffffffff, 0x00070000,
+	0x91f0, 0xffffffff, 0x00030002,
+	0x91f4, 0xffffffff, 0x00050004,
+	0x9200, 0xffffffff, 0x00010006,
+	0x9204, 0xffffffff, 0x00090008,
+	0x9208, 0xffffffff, 0x00070000,
+	0x920c, 0xffffffff, 0x00030002,
+	0x9210, 0xffffffff, 0x00050004,
+	0x921c, 0xffffffff, 0x00010006,
+	0x9220, 0xffffffff, 0x00090008,
+	0x9224, 0xffffffff, 0x00070000,
+	0x9228, 0xffffffff, 0x00030002,
+	0x922c, 0xffffffff, 0x00050004,
+	0x9238, 0xffffffff, 0x00010006,
+	0x923c, 0xffffffff, 0x00090008,
+	0x9240, 0xffffffff, 0x00070000,
+	0x9244, 0xffffffff, 0x00030002,
+	0x9248, 0xffffffff, 0x00050004,
+	0x9254, 0xffffffff, 0x00010006,
+	0x9258, 0xffffffff, 0x00090008,
+	0x925c, 0xffffffff, 0x00070000,
+	0x9260, 0xffffffff, 0x00030002,
+	0x9264, 0xffffffff, 0x00050004,
+	0x9270, 0xffffffff, 0x00010006,
+	0x9274, 0xffffffff, 0x00090008,
+	0x9278, 0xffffffff, 0x00070000,
+	0x927c, 0xffffffff, 0x00030002,
+	0x9280, 0xffffffff, 0x00050004,
+	0x928c, 0xffffffff, 0x00010006,
+	0x9290, 0xffffffff, 0x00090008,
+	0x9294, 0xffffffff, 0x00000000,
+	0x929c, 0xffffffff, 0x00000001,
+	0x802c, 0xffffffff, 0xc0000000,
+	0x977c, 0xffffffff, 0x00000100,
+	0x3f80, 0xffffffff, 0x00000100,
+	0xa210, 0xffffffff, 0x00000100,
+	0xa214, 0xffffffff, 0x00000100,
+	0x4d8, 0xffffffff, 0x00000100,
+	0x9784, 0xffffffff, 0x00000100,
+	0x9698, 0xffffffff, 0x00000100,
+	0x4d4, 0xffffffff, 0x00000200,
+	0x30cc, 0xffffffff, 0x00000100,
+	0x802c, 0xffffffff, 0xc0000000
+};
+
+static const u32 supersumo_golden_registers[] =
+{
+	0x5eb4, 0xffffffff, 0x00000002,
+	0x5cc, 0xffffffff, 0x00000001,
+	0x7030, 0xffffffff, 0x00000011,
+	0x7c30, 0xffffffff, 0x00000011,
+	0x6104, 0x01000300, 0x00000000,
+	0x5bc0, 0x00300000, 0x00000000,
+	0x8c04, 0xffffffff, 0x40600060,
+	0x8c08, 0xffffffff, 0x001c001c,
+	0x8c20, 0xffffffff, 0x00800080,
+	0x8c24, 0xffffffff, 0x00800080,
+	0x8c18, 0xffffffff, 0x20202078,
+	0x8c1c, 0xffffffff, 0x00001010,
+	0x918c, 0xffffffff, 0x00010006,
+	0x91a8, 0xffffffff, 0x00010006,
+	0x91c4, 0xffffffff, 0x00010006,
+	0x91e0, 0xffffffff, 0x00010006,
+	0x9200, 0xffffffff, 0x00010006,
+	0x9150, 0xffffffff, 0x6e944040,
+	0x917c, 0xffffffff, 0x00030002,
+	0x9180, 0xffffffff, 0x00050004,
+	0x9198, 0xffffffff, 0x00030002,
+	0x919c, 0xffffffff, 0x00050004,
+	0x91b4, 0xffffffff, 0x00030002,
+	0x91b8, 0xffffffff, 0x00050004,
+	0x91d0, 0xffffffff, 0x00030002,
+	0x91d4, 0xffffffff, 0x00050004,
+	0x91f0, 0xffffffff, 0x00030002,
+	0x91f4, 0xffffffff, 0x00050004,
+	0x915c, 0xffffffff, 0x00010000,
+	0x9160, 0xffffffff, 0x00030002,
+	0x3f90, 0xffff0000, 0xff000000,
+	0x9178, 0xffffffff, 0x00070000,
+	0x9194, 0xffffffff, 0x00070000,
+	0x91b0, 0xffffffff, 0x00070000,
+	0x91cc, 0xffffffff, 0x00070000,
+	0x91ec, 0xffffffff, 0x00070000,
+	0x9148, 0xffff0000, 0xff000000,
+	0x9190, 0xffffffff, 0x00090008,
+	0x91ac, 0xffffffff, 0x00090008,
+	0x91c8, 0xffffffff, 0x00090008,
+	0x91e4, 0xffffffff, 0x00090008,
+	0x9204, 0xffffffff, 0x00090008,
+	0x3f94, 0xffff0000, 0xff000000,
+	0x914c, 0xffff0000, 0xff000000,
+	0x929c, 0xffffffff, 0x00000001,
+	0x8a18, 0xffffffff, 0x00000100,
+	0x8b28, 0xffffffff, 0x00000100,
+	0x9144, 0xffffffff, 0x00000100,
+	0x5644, 0xffffffff, 0x00000100,
+	0x9b7c, 0xffffffff, 0x00000000,
+	0x8030, 0xffffffff, 0x0000100a,
+	0x8a14, 0xffffffff, 0x00000007,
+	0x8b24, 0xffffffff, 0x00ff0fff,
+	0x8b10, 0xffffffff, 0x00000000,
+	0x28a4c, 0x06000000, 0x06000000,
+	0x4d8, 0xffffffff, 0x00000100,
+	0x913c, 0xffff000f, 0x0100000a,
+	0x960c, 0xffffffff, 0x54763210,
+	0x88c4, 0xffffffff, 0x000000c2,
+	0x88d4, 0xffffffff, 0x00000010,
+	0x8974, 0xffffffff, 0x00000000,
+	0xc78, 0x00000080, 0x00000080,
+	0x5e78, 0xffffffff, 0x001000f0,
+	0xd02c, 0xffffffff, 0x08421000,
+	0xa008, 0xffffffff, 0x00010000,
+	0x8d00, 0xffffffff, 0x100e4848,
+	0x8d04, 0xffffffff, 0x00164745,
+	0x8c00, 0xffffffff, 0xe4000003,
+	0x8cf0, 0x1fffffff, 0x08e00620,
+	0x28350, 0xffffffff, 0x00000000,
+	0x9508, 0xffffffff, 0x00000002
+};
+
+static const u32 sumo_golden_registers[] =
+{
+	0x900c, 0x00ffffff, 0x0017071f,
+	0x8c18, 0xffffffff, 0x10101060,
+	0x8c1c, 0xffffffff, 0x00001010,
+	0x8c30, 0x0000000f, 0x00000005,
+	0x9688, 0x0000000f, 0x00000007
+};
+
+static const u32 wrestler_golden_registers[] =
+{
+	0x5eb4, 0xffffffff, 0x00000002,
+	0x5cc, 0xffffffff, 0x00000001,
+	0x7030, 0xffffffff, 0x00000011,
+	0x7c30, 0xffffffff, 0x00000011,
+	0x6104, 0x01000300, 0x00000000,
+	0x5bc0, 0x00300000, 0x00000000,
+	0x918c, 0xffffffff, 0x00010006,
+	0x91a8, 0xffffffff, 0x00010006,
+	0x9150, 0xffffffff, 0x6e944040,
+	0x917c, 0xffffffff, 0x00030002,
+	0x9198, 0xffffffff, 0x00030002,
+	0x915c, 0xffffffff, 0x00010000,
+	0x3f90, 0xffff0000, 0xff000000,
+	0x9178, 0xffffffff, 0x00070000,
+	0x9194, 0xffffffff, 0x00070000,
+	0x9148, 0xffff0000, 0xff000000,
+	0x9190, 0xffffffff, 0x00090008,
+	0x91ac, 0xffffffff, 0x00090008,
+	0x3f94, 0xffff0000, 0xff000000,
+	0x914c, 0xffff0000, 0xff000000,
+	0x929c, 0xffffffff, 0x00000001,
+	0x8a18, 0xffffffff, 0x00000100,
+	0x8b28, 0xffffffff, 0x00000100,
+	0x9144, 0xffffffff, 0x00000100,
+	0x9b7c, 0xffffffff, 0x00000000,
+	0x8030, 0xffffffff, 0x0000100a,
+	0x8a14, 0xffffffff, 0x00000001,
+	0x8b24, 0xffffffff, 0x00ff0fff,
+	0x8b10, 0xffffffff, 0x00000000,
+	0x28a4c, 0x06000000, 0x06000000,
+	0x4d8, 0xffffffff, 0x00000100,
+	0x913c, 0xffff000f, 0x0100000a,
+	0x960c, 0xffffffff, 0x54763210,
+	0x88c4, 0xffffffff, 0x000000c2,
+	0x88d4, 0xffffffff, 0x00000010,
+	0x8974, 0xffffffff, 0x00000000,
+	0xc78, 0x00000080, 0x00000080,
+	0x5e78, 0xffffffff, 0x001000f0,
+	0xd02c, 0xffffffff, 0x08421000,
+	0xa008, 0xffffffff, 0x00010000,
+	0x8d00, 0xffffffff, 0x100e4848,
+	0x8d04, 0xffffffff, 0x00164745,
+	0x8c00, 0xffffffff, 0xe4000003,
+	0x8cf0, 0x1fffffff, 0x08e00410,
+	0x28350, 0xffffffff, 0x00000000,
+	0x9508, 0xffffffff, 0x00000002,
+	0x900c, 0xffffffff, 0x0017071f,
+	0x8c18, 0xffffffff, 0x10101060,
+	0x8c1c, 0xffffffff, 0x00001010
+};
+
+static const u32 barts_golden_registers[] =
+{
+	0x5eb4, 0xffffffff, 0x00000002,
+	0x5e78, 0x8f311ff1, 0x001000f0,
+	0x3f90, 0xffff0000, 0xff000000,
+	0x9148, 0xffff0000, 0xff000000,
+	0x3f94, 0xffff0000, 0xff000000,
+	0x914c, 0xffff0000, 0xff000000,
+	0xc78, 0x00000080, 0x00000080,
+	0xbd4, 0x70073777, 0x00010001,
+	0xd02c, 0xbfffff1f, 0x08421000,
+	0xd0b8, 0x03773777, 0x02011003,
+	0x5bc0, 0x00200000, 0x50100000,
+	0x98f8, 0x33773777, 0x02011003,
+	0x98fc, 0xffffffff, 0x76543210,
+	0x7030, 0x31000311, 0x00000011,
+	0x2f48, 0x00000007, 0x02011003,
+	0x6b28, 0x00000010, 0x00000012,
+	0x7728, 0x00000010, 0x00000012,
+	0x10328, 0x00000010, 0x00000012,
+	0x10f28, 0x00000010, 0x00000012,
+	0x11b28, 0x00000010, 0x00000012,
+	0x12728, 0x00000010, 0x00000012,
+	0x240c, 0x000007ff, 0x00000380,
+	0x8a14, 0xf000001f, 0x00000007,
+	0x8b24, 0x3fff3fff, 0x00ff0fff,
+	0x8b10, 0x0000ff0f, 0x00000000,
+	0x28a4c, 0x07ffffff, 0x06000000,
+	0x10c, 0x00000001, 0x00010003,
+	0xa02c, 0xffffffff, 0x0000009b,
+	0x913c, 0x0000000f, 0x0100000a,
+	0x8d00, 0xffff7f7f, 0x100e4848,
+	0x8d04, 0x00ffffff, 0x00164745,
+	0x8c00, 0xfffc0003, 0xe4000003,
+	0x8c04, 0xf8ff00ff, 0x40600060,
+	0x8c08, 0x00ff00ff, 0x001c001c,
+	0x8cf0, 0x1fff1fff, 0x08e00620,
+	0x8c20, 0x0fff0fff, 0x00800080,
+	0x8c24, 0x0fff0fff, 0x00800080,
+	0x8c18, 0xffffffff, 0x20202078,
+	0x8c1c, 0x0000ffff, 0x00001010,
+	0x28350, 0x00000f01, 0x00000000,
+	0x9508, 0x3700001f, 0x00000002,
+	0x960c, 0xffffffff, 0x54763210,
+	0x88c4, 0x001f3ae3, 0x000000c2,
+	0x88d4, 0x0000001f, 0x00000010,
+	0x8974, 0xffffffff, 0x00000000
+};
+
+static const u32 turks_golden_registers[] =
+{
+	0x5eb4, 0xffffffff, 0x00000002,
+	0x5e78, 0x8f311ff1, 0x001000f0,
+	0x8c8, 0x00003000, 0x00001070,
+	0x8cc, 0x000fffff, 0x00040035,
+	0x3f90, 0xffff0000, 0xfff00000,
+	0x9148, 0xffff0000, 0xfff00000,
+	0x3f94, 0xffff0000, 0xfff00000,
+	0x914c, 0xffff0000, 0xfff00000,
+	0xc78, 0x00000080, 0x00000080,
+	0xbd4, 0x00073007, 0x00010002,
+	0xd02c, 0xbfffff1f, 0x08421000,
+	0xd0b8, 0x03773777, 0x02010002,
+	0x5bc0, 0x00200000, 0x50100000,
+	0x98f8, 0x33773777, 0x00010002,
+	0x98fc, 0xffffffff, 0x33221100,
+	0x7030, 0x31000311, 0x00000011,
+	0x2f48, 0x33773777, 0x00010002,
+	0x6b28, 0x00000010, 0x00000012,
+	0x7728, 0x00000010, 0x00000012,
+	0x10328, 0x00000010, 0x00000012,
+	0x10f28, 0x00000010, 0x00000012,
+	0x11b28, 0x00000010, 0x00000012,
+	0x12728, 0x00000010, 0x00000012,
+	0x240c, 0x000007ff, 0x00000380,
+	0x8a14, 0xf000001f, 0x00000007,
+	0x8b24, 0x3fff3fff, 0x00ff0fff,
+	0x8b10, 0x0000ff0f, 0x00000000,
+	0x28a4c, 0x07ffffff, 0x06000000,
+	0x10c, 0x00000001, 0x00010003,
+	0xa02c, 0xffffffff, 0x0000009b,
+	0x913c, 0x0000000f, 0x0100000a,
+	0x8d00, 0xffff7f7f, 0x100e4848,
+	0x8d04, 0x00ffffff, 0x00164745,
+	0x8c00, 0xfffc0003, 0xe4000003,
+	0x8c04, 0xf8ff00ff, 0x40600060,
+	0x8c08, 0x00ff00ff, 0x001c001c,
+	0x8cf0, 0x1fff1fff, 0x08e00410,
+	0x8c20, 0x0fff0fff, 0x00800080,
+	0x8c24, 0x0fff0fff, 0x00800080,
+	0x8c18, 0xffffffff, 0x20202078,
+	0x8c1c, 0x0000ffff, 0x00001010,
+	0x28350, 0x00000f01, 0x00000000,
+	0x9508, 0x3700001f, 0x00000002,
+	0x960c, 0xffffffff, 0x54763210,
+	0x88c4, 0x001f3ae3, 0x000000c2,
+	0x88d4, 0x0000001f, 0x00000010,
+	0x8974, 0xffffffff, 0x00000000
+};
+
+static const u32 caicos_golden_registers[] =
+{
+	0x5eb4, 0xffffffff, 0x00000002,
+	0x5e78, 0x8f311ff1, 0x001000f0,
+	0x8c8, 0x00003420, 0x00001450,
+	0x8cc, 0x000fffff, 0x00040035,
+	0x3f90, 0xffff0000, 0xfffc0000,
+	0x9148, 0xffff0000, 0xfffc0000,
+	0x3f94, 0xffff0000, 0xfffc0000,
+	0x914c, 0xffff0000, 0xfffc0000,
+	0xc78, 0x00000080, 0x00000080,
+	0xbd4, 0x00073007, 0x00010001,
+	0xd02c, 0xbfffff1f, 0x08421000,
+	0xd0b8, 0x03773777, 0x02010001,
+	0x5bc0, 0x00200000, 0x50100000,
+	0x98f8, 0x33773777, 0x02010001,
+	0x98fc, 0xffffffff, 0x33221100,
+	0x7030, 0x31000311, 0x00000011,
+	0x2f48, 0x33773777, 0x02010001,
+	0x6b28, 0x00000010, 0x00000012,
+	0x7728, 0x00000010, 0x00000012,
+	0x10328, 0x00000010, 0x00000012,
+	0x10f28, 0x00000010, 0x00000012,
+	0x11b28, 0x00000010, 0x00000012,
+	0x12728, 0x00000010, 0x00000012,
+	0x240c, 0x000007ff, 0x00000380,
+	0x8a14, 0xf000001f, 0x00000001,
+	0x8b24, 0x3fff3fff, 0x00ff0fff,
+	0x8b10, 0x0000ff0f, 0x00000000,
+	0x28a4c, 0x07ffffff, 0x06000000,
+	0x10c, 0x00000001, 0x00010003,
+	0xa02c, 0xffffffff, 0x0000009b,
+	0x913c, 0x0000000f, 0x0100000a,
+	0x8d00, 0xffff7f7f, 0x100e4848,
+	0x8d04, 0x00ffffff, 0x00164745,
+	0x8c00, 0xfffc0003, 0xe4000003,
+	0x8c04, 0xf8ff00ff, 0x40600060,
+	0x8c08, 0x00ff00ff, 0x001c001c,
+	0x8cf0, 0x1fff1fff, 0x08e00410,
+	0x8c20, 0x0fff0fff, 0x00800080,
+	0x8c24, 0x0fff0fff, 0x00800080,
+	0x8c18, 0xffffffff, 0x20202078,
+	0x8c1c, 0x0000ffff, 0x00001010,
+	0x28350, 0x00000f01, 0x00000000,
+	0x9508, 0x3700001f, 0x00000002,
+	0x960c, 0xffffffff, 0x54763210,
+	0x88c4, 0x001f3ae3, 0x000000c2,
+	0x88d4, 0x0000001f, 0x00000010,
+	0x8974, 0xffffffff, 0x00000000
+};
+
+static void evergreen_init_golden_registers(struct radeon_device *rdev)
+{
+	switch (rdev->family) {
+	case CHIP_CYPRESS:
+	case CHIP_HEMLOCK:
+		radeon_program_register_sequence(rdev,
+						 evergreen_golden_registers,
+						 (const u32)ARRAY_SIZE(evergreen_golden_registers));
+		radeon_program_register_sequence(rdev,
+						 evergreen_golden_registers2,
+						 (const u32)ARRAY_SIZE(evergreen_golden_registers2));
+		radeon_program_register_sequence(rdev,
+						 cypress_mgcg_init,
+						 (const u32)ARRAY_SIZE(cypress_mgcg_init));
+		break;
+	case CHIP_JUNIPER:
+		radeon_program_register_sequence(rdev,
+						 evergreen_golden_registers,
+						 (const u32)ARRAY_SIZE(evergreen_golden_registers));
+		radeon_program_register_sequence(rdev,
+						 evergreen_golden_registers2,
+						 (const u32)ARRAY_SIZE(evergreen_golden_registers2));
+		radeon_program_register_sequence(rdev,
+						 juniper_mgcg_init,
+						 (const u32)ARRAY_SIZE(juniper_mgcg_init));
+		break;
+	case CHIP_REDWOOD:
+		radeon_program_register_sequence(rdev,
+						 evergreen_golden_registers,
+						 (const u32)ARRAY_SIZE(evergreen_golden_registers));
+		radeon_program_register_sequence(rdev,
+						 evergreen_golden_registers2,
+						 (const u32)ARRAY_SIZE(evergreen_golden_registers2));
+		radeon_program_register_sequence(rdev,
+						 redwood_mgcg_init,
+						 (const u32)ARRAY_SIZE(redwood_mgcg_init));
+		break;
+	case CHIP_CEDAR:
+		radeon_program_register_sequence(rdev,
+						 cedar_golden_registers,
+						 (const u32)ARRAY_SIZE(cedar_golden_registers));
+		radeon_program_register_sequence(rdev,
+						 evergreen_golden_registers2,
+						 (const u32)ARRAY_SIZE(evergreen_golden_registers2));
+		radeon_program_register_sequence(rdev,
+						 cedar_mgcg_init,
+						 (const u32)ARRAY_SIZE(cedar_mgcg_init));
+		break;
+	case CHIP_PALM:
+		radeon_program_register_sequence(rdev,
+						 wrestler_golden_registers,
+						 (const u32)ARRAY_SIZE(wrestler_golden_registers));
+		break;
+	case CHIP_SUMO:
+		radeon_program_register_sequence(rdev,
+						 supersumo_golden_registers,
+						 (const u32)ARRAY_SIZE(supersumo_golden_registers));
+		break;
+	case CHIP_SUMO2:
+		radeon_program_register_sequence(rdev,
+						 supersumo_golden_registers,
+						 (const u32)ARRAY_SIZE(supersumo_golden_registers));
+		radeon_program_register_sequence(rdev,
+						 sumo_golden_registers,
+						 (const u32)ARRAY_SIZE(sumo_golden_registers));
+		break;
+	case CHIP_BARTS:
+		radeon_program_register_sequence(rdev,
+						 barts_golden_registers,
+						 (const u32)ARRAY_SIZE(barts_golden_registers));
+		break;
+	case CHIP_TURKS:
+		radeon_program_register_sequence(rdev,
+						 turks_golden_registers,
+						 (const u32)ARRAY_SIZE(turks_golden_registers));
+		break;
+	case CHIP_CAICOS:
+		radeon_program_register_sequence(rdev,
+						 caicos_golden_registers,
+						 (const u32)ARRAY_SIZE(caicos_golden_registers));
+		break;
+	default:
+		break;
+	}
+}
+
 void evergreen_tiling_fields(unsigned tiling_flags, unsigned *bankw,
 			     unsigned *bankh, unsigned *mtaspect,
 			     unsigned *tile_split)
@@ -84,6 +942,142 @@ void evergreen_tiling_fields(unsigned tiling_flags, unsigned *bankw,
 	}
 }
 
+static int sumo_set_uvd_clock(struct radeon_device *rdev, u32 clock,
+			      u32 cntl_reg, u32 status_reg)
+{
+	int r, i;
+	struct atom_clock_dividers dividers;
+
+        r = radeon_atom_get_clock_dividers(rdev, COMPUTE_ENGINE_PLL_PARAM,
+					   clock, false, &dividers);
+	if (r)
+		return r;
+
+	WREG32_P(cntl_reg, dividers.post_div, ~(DCLK_DIR_CNTL_EN|DCLK_DIVIDER_MASK));
+
+	for (i = 0; i < 100; i++) {
+		if (RREG32(status_reg) & DCLK_STATUS)
+			break;
+		mdelay(10);
+	}
+	if (i == 100)
+		return -ETIMEDOUT;
+
+	return 0;
+}
+
+int sumo_set_uvd_clocks(struct radeon_device *rdev, u32 vclk, u32 dclk)
+{
+	int r = 0;
+	u32 cg_scratch = RREG32(CG_SCRATCH1);
+
+	r = sumo_set_uvd_clock(rdev, vclk, CG_VCLK_CNTL, CG_VCLK_STATUS);
+	if (r)
+		goto done;
+	cg_scratch &= 0xffff0000;
+	cg_scratch |= vclk / 100; /* Mhz */
+
+	r = sumo_set_uvd_clock(rdev, dclk, CG_DCLK_CNTL, CG_DCLK_STATUS);
+	if (r)
+		goto done;
+	cg_scratch &= 0x0000ffff;
+	cg_scratch |= (dclk / 100) << 16; /* Mhz */
+
+done:
+	WREG32(CG_SCRATCH1, cg_scratch);
+
+	return r;
+}
+
+int evergreen_set_uvd_clocks(struct radeon_device *rdev, u32 vclk, u32 dclk)
+{
+	/* start off with something large */
+	unsigned fb_div = 0, vclk_div = 0, dclk_div = 0;
+	int r;
+
+	/* bypass vclk and dclk with bclk */
+	WREG32_P(CG_UPLL_FUNC_CNTL_2,
+		VCLK_SRC_SEL(1) | DCLK_SRC_SEL(1),
+		~(VCLK_SRC_SEL_MASK | DCLK_SRC_SEL_MASK));
+
+	/* put PLL in bypass mode */
+	WREG32_P(CG_UPLL_FUNC_CNTL, UPLL_BYPASS_EN_MASK, ~UPLL_BYPASS_EN_MASK);
+
+	if (!vclk || !dclk) {
+		/* keep the Bypass mode, put PLL to sleep */
+		WREG32_P(CG_UPLL_FUNC_CNTL, UPLL_SLEEP_MASK, ~UPLL_SLEEP_MASK);
+		return 0;
+	}
+
+	r = radeon_uvd_calc_upll_dividers(rdev, vclk, dclk, 125000, 250000,
+					  16384, 0x03FFFFFF, 0, 128, 5,
+					  &fb_div, &vclk_div, &dclk_div);
+	if (r)
+		return r;
+
+	/* set VCO_MODE to 1 */
+	WREG32_P(CG_UPLL_FUNC_CNTL, UPLL_VCO_MODE_MASK, ~UPLL_VCO_MODE_MASK);
+
+	/* toggle UPLL_SLEEP to 1 then back to 0 */
+	WREG32_P(CG_UPLL_FUNC_CNTL, UPLL_SLEEP_MASK, ~UPLL_SLEEP_MASK);
+	WREG32_P(CG_UPLL_FUNC_CNTL, 0, ~UPLL_SLEEP_MASK);
+
+	/* deassert UPLL_RESET */
+	WREG32_P(CG_UPLL_FUNC_CNTL, 0, ~UPLL_RESET_MASK);
+
+	mdelay(1);
+
+	r = radeon_uvd_send_upll_ctlreq(rdev, CG_UPLL_FUNC_CNTL);
+	if (r)
+		return r;
+
+	/* assert UPLL_RESET again */
+	WREG32_P(CG_UPLL_FUNC_CNTL, UPLL_RESET_MASK, ~UPLL_RESET_MASK);
+
+	/* disable spread spectrum. */
+	WREG32_P(CG_UPLL_SPREAD_SPECTRUM, 0, ~SSEN_MASK);
+
+	/* set feedback divider */
+	WREG32_P(CG_UPLL_FUNC_CNTL_3, UPLL_FB_DIV(fb_div), ~UPLL_FB_DIV_MASK);
+
+	/* set ref divider to 0 */
+	WREG32_P(CG_UPLL_FUNC_CNTL, 0, ~UPLL_REF_DIV_MASK);
+
+	if (fb_div < 307200)
+		WREG32_P(CG_UPLL_FUNC_CNTL_4, 0, ~UPLL_SPARE_ISPARE9);
+	else
+		WREG32_P(CG_UPLL_FUNC_CNTL_4, UPLL_SPARE_ISPARE9, ~UPLL_SPARE_ISPARE9);
+
+	/* set PDIV_A and PDIV_B */
+	WREG32_P(CG_UPLL_FUNC_CNTL_2,
+		UPLL_PDIV_A(vclk_div) | UPLL_PDIV_B(dclk_div),
+		~(UPLL_PDIV_A_MASK | UPLL_PDIV_B_MASK));
+
+	/* give the PLL some time to settle */
+	mdelay(15);
+
+	/* deassert PLL_RESET */
+	WREG32_P(CG_UPLL_FUNC_CNTL, 0, ~UPLL_RESET_MASK);
+
+	mdelay(15);
+
+	/* switch from bypass mode to normal mode */
+	WREG32_P(CG_UPLL_FUNC_CNTL, 0, ~UPLL_BYPASS_EN_MASK);
+
+	r = radeon_uvd_send_upll_ctlreq(rdev, CG_UPLL_FUNC_CNTL);
+	if (r)
+		return r;
+
+	/* switch VCLK and DCLK selection */
+	WREG32_P(CG_UPLL_FUNC_CNTL_2,
+		VCLK_SRC_SEL(2) | DCLK_SRC_SEL(2),
+		~(VCLK_SRC_SEL_MASK | DCLK_SRC_SEL_MASK));
+
+	mdelay(100);
+
+	return 0;
+}
+
 void evergreen_fix_pci_max_read_req_size(struct radeon_device *rdev)
 {
 	u16 ctl, v;
@@ -105,6 +1099,27 @@ void evergreen_fix_pci_max_read_req_size(struct radeon_device *rdev)
 	}
 }
 
+static bool dce4_is_in_vblank(struct radeon_device *rdev, int crtc)
+{
+	if (RREG32(EVERGREEN_CRTC_STATUS + crtc_offsets[crtc]) & EVERGREEN_CRTC_V_BLANK)
+		return true;
+	else
+		return false;
+}
+
+static bool dce4_is_counter_moving(struct radeon_device *rdev, int crtc)
+{
+	u32 pos1, pos2;
+
+	pos1 = RREG32(EVERGREEN_CRTC_STATUS_POSITION + crtc_offsets[crtc]);
+	pos2 = RREG32(EVERGREEN_CRTC_STATUS_POSITION + crtc_offsets[crtc]);
+
+	if (pos1 != pos2)
+		return true;
+	else
+		return false;
+}
+
 /**
  * dce4_wait_for_vblank - vblank wait asic callback.
  *
@@ -115,21 +1130,28 @@ void evergreen_fix_pci_max_read_req_size(struct radeon_device *rdev)
  */
 void dce4_wait_for_vblank(struct radeon_device *rdev, int crtc)
 {
-	int i;
+	unsigned i = 0;
 
 	if (crtc >= rdev->num_crtc)
 		return;
 
-	if (RREG32(EVERGREEN_CRTC_CONTROL + crtc_offsets[crtc]) & EVERGREEN_CRTC_MASTER_EN) {
-		for (i = 0; i < rdev->usec_timeout; i++) {
-			if (!(RREG32(EVERGREEN_CRTC_STATUS + crtc_offsets[crtc]) & EVERGREEN_CRTC_V_BLANK))
+	if (!(RREG32(EVERGREEN_CRTC_CONTROL + crtc_offsets[crtc]) & EVERGREEN_CRTC_MASTER_EN))
+		return;
+
+	/* depending on when we hit vblank, we may be close to active; if so,
+	 * wait for another frame.
+	 */
+	while (dce4_is_in_vblank(rdev, crtc)) {
+		if (i++ % 100 == 0) {
+			if (!dce4_is_counter_moving(rdev, crtc))
 				break;
-			udelay(1);
 		}
-		for (i = 0; i < rdev->usec_timeout; i++) {
-			if (RREG32(EVERGREEN_CRTC_STATUS + crtc_offsets[crtc]) & EVERGREEN_CRTC_V_BLANK)
+	}
+
+	while (!dce4_is_in_vblank(rdev, crtc)) {
+		if (i++ % 100 == 0) {
+			if (!dce4_is_counter_moving(rdev, crtc))
 				break;
-			udelay(1);
 		}
 	}
 }
@@ -608,6 +1630,16 @@ void evergreen_hpd_init(struct radeon_device *rdev)
 
 	list_for_each_entry(connector, &dev->mode_config.connector_list, head) {
 		struct radeon_connector *radeon_connector = to_radeon_connector(connector);
+
+		if (connector->connector_type == DRM_MODE_CONNECTOR_eDP ||
+		    connector->connector_type == DRM_MODE_CONNECTOR_LVDS) {
+			/* don't try to enable hpd on eDP or LVDS avoid breaking the
+			 * aux dp channel on imac and help (but not completely fix)
+			 * https://bugzilla.redhat.com/show_bug.cgi?id=726143
+			 * also avoid interrupt storms during dpms.
+			 */
+			continue;
+		}
 		switch (radeon_connector->hpd.hpd) {
 		case RADEON_HPD_1:
 			WREG32(DC_HPD1_CONTROL, tmp);
@@ -1325,17 +2357,16 @@ void evergreen_mc_stop(struct radeon_device *rdev, struct evergreen_mc_save *sav
 				tmp = RREG32(EVERGREEN_CRTC_BLANK_CONTROL + crtc_offsets[i]);
 				if (!(tmp & EVERGREEN_CRTC_BLANK_DATA_EN)) {
 					radeon_wait_for_vblank(rdev, i);
-					tmp |= EVERGREEN_CRTC_BLANK_DATA_EN;
 					WREG32(EVERGREEN_CRTC_UPDATE_LOCK + crtc_offsets[i], 1);
+					tmp |= EVERGREEN_CRTC_BLANK_DATA_EN;
 					WREG32(EVERGREEN_CRTC_BLANK_CONTROL + crtc_offsets[i], tmp);
-					WREG32(EVERGREEN_CRTC_UPDATE_LOCK + crtc_offsets[i], 0);
 				}
 			} else {
 				tmp = RREG32(EVERGREEN_CRTC_CONTROL + crtc_offsets[i]);
 				if (!(tmp & EVERGREEN_CRTC_DISP_READ_REQUEST_DISABLE)) {
 					radeon_wait_for_vblank(rdev, i);
-					tmp |= EVERGREEN_CRTC_DISP_READ_REQUEST_DISABLE;
 					WREG32(EVERGREEN_CRTC_UPDATE_LOCK + crtc_offsets[i], 1);
+					tmp |= EVERGREEN_CRTC_DISP_READ_REQUEST_DISABLE;
 					WREG32(EVERGREEN_CRTC_CONTROL + crtc_offsets[i], tmp);
 					WREG32(EVERGREEN_CRTC_UPDATE_LOCK + crtc_offsets[i], 0);
 				}
@@ -1347,6 +2378,15 @@ void evergreen_mc_stop(struct radeon_device *rdev, struct evergreen_mc_save *sav
 					break;
 				udelay(1);
 			}
+
+			/* XXX this is a hack to avoid strange behavior with EFI on certain systems */
+			WREG32(EVERGREEN_CRTC_UPDATE_LOCK + crtc_offsets[i], 1);
+			tmp = RREG32(EVERGREEN_CRTC_CONTROL + crtc_offsets[i]);
+			tmp &= ~EVERGREEN_CRTC_MASTER_EN;
+			WREG32(EVERGREEN_CRTC_CONTROL + crtc_offsets[i], tmp);
+			WREG32(EVERGREEN_CRTC_UPDATE_LOCK + crtc_offsets[i], 0);
+			save->crtc_enabled[i] = false;
+			/* ***** */
 		} else {
 			save->crtc_enabled[i] = false;
 		}
@@ -1364,6 +2404,22 @@ void evergreen_mc_stop(struct radeon_device *rdev, struct evergreen_mc_save *sav
 	}
 	/* wait for the MC to settle */
 	udelay(100);
+
+	/* lock double buffered regs */
+	for (i = 0; i < rdev->num_crtc; i++) {
+		if (save->crtc_enabled[i]) {
+			tmp = RREG32(EVERGREEN_GRPH_UPDATE + crtc_offsets[i]);
+			if (!(tmp & EVERGREEN_GRPH_UPDATE_LOCK)) {
+				tmp |= EVERGREEN_GRPH_UPDATE_LOCK;
+				WREG32(EVERGREEN_GRPH_UPDATE + crtc_offsets[i], tmp);
+			}
+			tmp = RREG32(EVERGREEN_MASTER_UPDATE_LOCK + crtc_offsets[i]);
+			if (!(tmp & 1)) {
+				tmp |= 1;
+				WREG32(EVERGREEN_MASTER_UPDATE_LOCK + crtc_offsets[i], tmp);
+			}
+		}
+	}
 }
 
 void evergreen_mc_resume(struct radeon_device *rdev, struct evergreen_mc_save *save)
@@ -1385,6 +2441,33 @@ void evergreen_mc_resume(struct radeon_device *rdev, struct evergreen_mc_save *s
 	WREG32(EVERGREEN_VGA_MEMORY_BASE_ADDRESS_HIGH, upper_32_bits(rdev->mc.vram_start));
 	WREG32(EVERGREEN_VGA_MEMORY_BASE_ADDRESS, (u32)rdev->mc.vram_start);
 
+	/* unlock regs and wait for update */
+	for (i = 0; i < rdev->num_crtc; i++) {
+		if (save->crtc_enabled[i]) {
+			tmp = RREG32(EVERGREEN_MASTER_UPDATE_MODE + crtc_offsets[i]);
+			if ((tmp & 0x3) != 0) {
+				tmp &= ~0x3;
+				WREG32(EVERGREEN_MASTER_UPDATE_MODE + crtc_offsets[i], tmp);
+			}
+			tmp = RREG32(EVERGREEN_GRPH_UPDATE + crtc_offsets[i]);
+			if (tmp & EVERGREEN_GRPH_UPDATE_LOCK) {
+				tmp &= ~EVERGREEN_GRPH_UPDATE_LOCK;
+				WREG32(EVERGREEN_GRPH_UPDATE + crtc_offsets[i], tmp);
+			}
+			tmp = RREG32(EVERGREEN_MASTER_UPDATE_LOCK + crtc_offsets[i]);
+			if (tmp & 1) {
+				tmp &= ~1;
+				WREG32(EVERGREEN_MASTER_UPDATE_LOCK + crtc_offsets[i], tmp);
+			}
+			for (j = 0; j < rdev->usec_timeout; j++) {
+				tmp = RREG32(EVERGREEN_GRPH_UPDATE + crtc_offsets[i]);
+				if ((tmp & EVERGREEN_GRPH_SURFACE_UPDATE_PENDING) == 0)
+					break;
+				udelay(1);
+			}
+		}
+	}
+
 	/* unblackout the MC */
 	tmp = RREG32(MC_SHARED_BLACKOUT_CNTL);
 	tmp &= ~BLACKOUT_MODE_MASK;
@@ -2050,6 +3133,14 @@ static void evergreen_gpu_init(struct radeon_device *rdev)
 	}
 	/* enabled rb are just the one not disabled :) */
 	disabled_rb_mask = tmp;
+	tmp = 0;
+	for (i = 0; i < rdev->config.evergreen.max_backends; i++)
+		tmp |= (1 << i);
+	/* if all the backends are disabled, fix it up here */
+	if ((disabled_rb_mask & tmp) == tmp) {
+		for (i = 0; i < rdev->config.evergreen.max_backends; i++)
+			disabled_rb_mask &= ~(1 << i);
+	}
 
 	WREG32(GRBM_GFX_INDEX, INSTANCE_BROADCAST_WRITES | SE_BROADCAST_WRITES);
 	WREG32(RLC_GFX_INDEX, INSTANCE_BROADCAST_WRITES | SE_BROADCAST_WRITES);
@@ -2058,6 +3149,9 @@ static void evergreen_gpu_init(struct radeon_device *rdev)
 	WREG32(DMIF_ADDR_CONFIG, gb_addr_config);
 	WREG32(HDP_ADDR_CONFIG, gb_addr_config);
 	WREG32(DMA_TILING_CONFIG, gb_addr_config);
+	WREG32(UVD_UDEC_ADDR_CONFIG, gb_addr_config);
+	WREG32(UVD_UDEC_DB_ADDR_CONFIG, gb_addr_config);
+	WREG32(UVD_UDEC_DBW_ADDR_CONFIG, gb_addr_config);
 
 	if ((rdev->config.evergreen.max_backends == 1) &&
 	    (rdev->flags & RADEON_IS_IGP)) {
@@ -3360,6 +4454,9 @@ restart_ih:
 				DRM_ERROR("Unhandled interrupt: %d %d\n", src_id, src_data);
 				break;
 			}
+		case 124: /* UVD */
+			DRM_DEBUG("IH: UVD int: 0x%08x\n", src_data);
+			radeon_fence_process(rdev, R600_RING_TYPE_UVD_INDEX);
 			break;
 		case 146:
 		case 147:
@@ -3571,7 +4668,7 @@ int evergreen_copy_dma(struct radeon_device *rdev,
 
 static int evergreen_startup(struct radeon_device *rdev)
 {
-	struct radeon_ring *ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX];
+	struct radeon_ring *ring;
 	int r;
 
 	/* enable pcie gen2 link */
@@ -3638,6 +4735,17 @@ static int evergreen_startup(struct radeon_device *rdev)
 		return r;
 	}
 
+	r = rv770_uvd_resume(rdev);
+	if (!r) {
+		r = radeon_fence_driver_start_ring(rdev,
+						   R600_RING_TYPE_UVD_INDEX);
+		if (r)
+			dev_err(rdev->dev, "UVD fences init error (%d).\n", r);
+	}
+
+	if (r)
+		rdev->ring[R600_RING_TYPE_UVD_INDEX].ring_size = 0;
+
 	/* Enable IRQ */
 	r = r600_irq_init(rdev);
 	if (r) {
@@ -3647,6 +4755,7 @@ static int evergreen_startup(struct radeon_device *rdev)
 	}
 	evergreen_irq_set(rdev);
 
+	ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX];
 	r = radeon_ring_init(rdev, ring, ring->ring_size, RADEON_WB_CP_RPTR_OFFSET,
 			     R600_CP_RB_RPTR, R600_CP_RB_WPTR,
 			     0, 0xfffff, RADEON_CP_PACKET2);
@@ -3670,6 +4779,19 @@ static int evergreen_startup(struct radeon_device *rdev)
 	if (r)
 		return r;
 
+	ring = &rdev->ring[R600_RING_TYPE_UVD_INDEX];
+	if (ring->ring_size) {
+		r = radeon_ring_init(rdev, ring, ring->ring_size,
+				     R600_WB_UVD_RPTR_OFFSET,
+				     UVD_RBC_RB_RPTR, UVD_RBC_RB_WPTR,
+				     0, 0xfffff, RADEON_CP_PACKET2);
+		if (!r)
+			r = r600_uvd_init(rdev);
+
+		if (r)
+			DRM_ERROR("radeon: error initializing UVD (%d).\n", r);
+	}
+
 	r = radeon_ib_pool_init(rdev);
 	if (r) {
 		dev_err(rdev->dev, "IB initialization failed (%d).\n", r);
@@ -3701,6 +4823,9 @@ int evergreen_resume(struct radeon_device *rdev)
 	/* post card */
 	atom_asic_init(rdev->mode_info.atom_context);
 
+	/* init golden registers */
+	evergreen_init_golden_registers(rdev);
+
 	rdev->accel_working = true;
 	r = evergreen_startup(rdev);
 	if (r) {
@@ -3716,8 +4841,10 @@ int evergreen_resume(struct radeon_device *rdev)
 int evergreen_suspend(struct radeon_device *rdev)
 {
 	r600_audio_fini(rdev);
+	radeon_uvd_suspend(rdev);
 	r700_cp_stop(rdev);
 	r600_dma_stop(rdev);
+	r600_uvd_rbc_stop(rdev);
 	evergreen_irq_suspend(rdev);
 	radeon_wb_disable(rdev);
 	evergreen_pcie_gart_disable(rdev);
@@ -3762,6 +4889,8 @@ int evergreen_init(struct radeon_device *rdev)
 		DRM_INFO("GPU not posted. posting now...\n");
 		atom_asic_init(rdev->mode_info.atom_context);
 	}
+	/* init golden registers */
+	evergreen_init_golden_registers(rdev);
 	/* Initialize scratch registers */
 	r600_scratch_init(rdev);
 	/* Initialize surface registers */
@@ -3797,6 +4926,13 @@ int evergreen_init(struct radeon_device *rdev)
 	rdev->ring[R600_RING_TYPE_DMA_INDEX].ring_obj = NULL;
 	r600_ring_init(rdev, &rdev->ring[R600_RING_TYPE_DMA_INDEX], 64 * 1024);
 
+	r = radeon_uvd_init(rdev);
+	if (!r) {
+		rdev->ring[R600_RING_TYPE_UVD_INDEX].ring_obj = NULL;
+		r600_ring_init(rdev, &rdev->ring[R600_RING_TYPE_UVD_INDEX],
+			       4096);
+	}
+
 	rdev->ih.ring_obj = NULL;
 	r600_ih_ring_init(rdev, 64 * 1024);
 
@@ -3843,6 +4979,7 @@ void evergreen_fini(struct radeon_device *rdev)
 	radeon_ib_pool_fini(rdev);
 	radeon_irq_kms_fini(rdev);
 	evergreen_pcie_gart_fini(rdev);
+	radeon_uvd_fini(rdev);
 	r600_vram_scratch_fini(rdev);
 	radeon_gem_fini(rdev);
 	radeon_fence_driver_fini(rdev);
@@ -3878,7 +5015,7 @@ void evergreen_pcie_gen2_enable(struct radeon_device *rdev)
 	if (!(mask & DRM_PCIE_SPEED_50))
 		return;
 
-	speed_cntl = RREG32_PCIE_P(PCIE_LC_SPEED_CNTL);
+	speed_cntl = RREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL);
 	if (speed_cntl & LC_CURRENT_DATA_RATE) {
 		DRM_INFO("PCIE gen 2 link speeds already enabled\n");
 		return;
@@ -3889,33 +5026,33 @@ void evergreen_pcie_gen2_enable(struct radeon_device *rdev)
 	if ((speed_cntl & LC_OTHER_SIDE_EVER_SENT_GEN2) ||
 	    (speed_cntl & LC_OTHER_SIDE_SUPPORTS_GEN2)) {
 
-		link_width_cntl = RREG32_PCIE_P(PCIE_LC_LINK_WIDTH_CNTL);
+		link_width_cntl = RREG32_PCIE_PORT(PCIE_LC_LINK_WIDTH_CNTL);
 		link_width_cntl &= ~LC_UPCONFIGURE_DIS;
-		WREG32_PCIE_P(PCIE_LC_LINK_WIDTH_CNTL, link_width_cntl);
+		WREG32_PCIE_PORT(PCIE_LC_LINK_WIDTH_CNTL, link_width_cntl);
 
-		speed_cntl = RREG32_PCIE_P(PCIE_LC_SPEED_CNTL);
+		speed_cntl = RREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL);
 		speed_cntl &= ~LC_TARGET_LINK_SPEED_OVERRIDE_EN;
-		WREG32_PCIE_P(PCIE_LC_SPEED_CNTL, speed_cntl);
+		WREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL, speed_cntl);
 
-		speed_cntl = RREG32_PCIE_P(PCIE_LC_SPEED_CNTL);
+		speed_cntl = RREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL);
 		speed_cntl |= LC_CLR_FAILED_SPD_CHANGE_CNT;
-		WREG32_PCIE_P(PCIE_LC_SPEED_CNTL, speed_cntl);
+		WREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL, speed_cntl);
 
-		speed_cntl = RREG32_PCIE_P(PCIE_LC_SPEED_CNTL);
+		speed_cntl = RREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL);
 		speed_cntl &= ~LC_CLR_FAILED_SPD_CHANGE_CNT;
-		WREG32_PCIE_P(PCIE_LC_SPEED_CNTL, speed_cntl);
+		WREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL, speed_cntl);
 
-		speed_cntl = RREG32_PCIE_P(PCIE_LC_SPEED_CNTL);
+		speed_cntl = RREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL);
 		speed_cntl |= LC_GEN2_EN_STRAP;
-		WREG32_PCIE_P(PCIE_LC_SPEED_CNTL, speed_cntl);
+		WREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL, speed_cntl);
 
 	} else {
-		link_width_cntl = RREG32_PCIE_P(PCIE_LC_LINK_WIDTH_CNTL);
+		link_width_cntl = RREG32_PCIE_PORT(PCIE_LC_LINK_WIDTH_CNTL);
 		/* XXX: only disable it if gen1 bridge vendor == 0x111d or 0x1106 */
 		if (1)
 			link_width_cntl |= LC_UPCONFIGURE_DIS;
 		else
 			link_width_cntl &= ~LC_UPCONFIGURE_DIS;
-		WREG32_PCIE_P(PCIE_LC_LINK_WIDTH_CNTL, link_width_cntl);
+		WREG32_PCIE_PORT(PCIE_LC_LINK_WIDTH_CNTL, link_width_cntl);
 	}
 }
diff --git a/drivers/gpu/drm/radeon/evergreen_hdmi.c b/drivers/gpu/drm/radeon/evergreen_hdmi.c
index 4fdecc2b4040..b4ab8ceb1654 100644
--- a/drivers/gpu/drm/radeon/evergreen_hdmi.c
+++ b/drivers/gpu/drm/radeon/evergreen_hdmi.c
@@ -54,6 +54,68 @@ static void evergreen_hdmi_update_ACR(struct drm_encoder *encoder, uint32_t cloc
 	WREG32(HDMI_ACR_48_1 + offset, acr.n_48khz);
 }
 
+static void evergreen_hdmi_write_sad_regs(struct drm_encoder *encoder)
+{
+	struct radeon_device *rdev = encoder->dev->dev_private;
+	struct drm_connector *connector;
+	struct radeon_connector *radeon_connector = NULL;
+	struct cea_sad *sads;
+	int i, sad_count;
+
+	static const u16 eld_reg_to_type[][2] = {
+		{ AZ_F0_CODEC_PIN0_CONTROL_AUDIO_DESCRIPTOR0, HDMI_AUDIO_CODING_TYPE_PCM },
+		{ AZ_F0_CODEC_PIN0_CONTROL_AUDIO_DESCRIPTOR1, HDMI_AUDIO_CODING_TYPE_AC3 },
+		{ AZ_F0_CODEC_PIN0_CONTROL_AUDIO_DESCRIPTOR2, HDMI_AUDIO_CODING_TYPE_MPEG1 },
+		{ AZ_F0_CODEC_PIN0_CONTROL_AUDIO_DESCRIPTOR3, HDMI_AUDIO_CODING_TYPE_MP3 },
+		{ AZ_F0_CODEC_PIN0_CONTROL_AUDIO_DESCRIPTOR4, HDMI_AUDIO_CODING_TYPE_MPEG2 },
+		{ AZ_F0_CODEC_PIN0_CONTROL_AUDIO_DESCRIPTOR5, HDMI_AUDIO_CODING_TYPE_AAC_LC },
+		{ AZ_F0_CODEC_PIN0_CONTROL_AUDIO_DESCRIPTOR6, HDMI_AUDIO_CODING_TYPE_DTS },
+		{ AZ_F0_CODEC_PIN0_CONTROL_AUDIO_DESCRIPTOR7, HDMI_AUDIO_CODING_TYPE_ATRAC },
+		{ AZ_F0_CODEC_PIN0_CONTROL_AUDIO_DESCRIPTOR9, HDMI_AUDIO_CODING_TYPE_EAC3 },
+		{ AZ_F0_CODEC_PIN0_CONTROL_AUDIO_DESCRIPTOR10, HDMI_AUDIO_CODING_TYPE_DTS_HD },
+		{ AZ_F0_CODEC_PIN0_CONTROL_AUDIO_DESCRIPTOR11, HDMI_AUDIO_CODING_TYPE_MLP },
+		{ AZ_F0_CODEC_PIN0_CONTROL_AUDIO_DESCRIPTOR13, HDMI_AUDIO_CODING_TYPE_WMA_PRO },
+	};
+
+	list_for_each_entry(connector, &encoder->dev->mode_config.connector_list, head) {
+		if (connector->encoder == encoder)
+			radeon_connector = to_radeon_connector(connector);
+	}
+
+	if (!radeon_connector) {
+		DRM_ERROR("Couldn't find encoder's connector\n");
+		return;
+	}
+
+	sad_count = drm_edid_to_sad(radeon_connector->edid, &sads);
+	if (sad_count < 0) {
+		DRM_ERROR("Couldn't read SADs: %d\n", sad_count);
+		return;
+	}
+	BUG_ON(!sads);
+
+	for (i = 0; i < ARRAY_SIZE(eld_reg_to_type); i++) {
+		u32 value = 0;
+		int j;
+
+		for (j = 0; j < sad_count; j++) {
+			struct cea_sad *sad = &sads[j];
+
+			if (sad->format == eld_reg_to_type[i][1]) {
+				value = MAX_CHANNELS(sad->channels) |
+					DESCRIPTOR_BYTE_2(sad->byte2) |
+					SUPPORTED_FREQUENCIES(sad->freq);
+				if (sad->format == HDMI_AUDIO_CODING_TYPE_PCM)
+					value |= SUPPORTED_FREQUENCIES_STEREO(sad->freq);
+				break;
+			}
+		}
+		WREG32(eld_reg_to_type[i][0], value);
+	}
+
+	kfree(sads);
+}
+
 /*
  * build a HDMI Video Info Frame
  */
@@ -85,6 +147,30 @@ static void evergreen_hdmi_update_avi_infoframe(struct drm_encoder *encoder,
 		frame[0xC] | (frame[0xD] << 8));
 }
 
+static void evergreen_audio_set_dto(struct drm_encoder *encoder, u32 clock)
+{
+	struct drm_device *dev = encoder->dev;
+	struct radeon_device *rdev = dev->dev_private;
+	struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder);
+	struct radeon_encoder_atom_dig *dig = radeon_encoder->enc_priv;
+	struct radeon_crtc *radeon_crtc = to_radeon_crtc(encoder->crtc);
+	u32 base_rate = 48000;
+
+	if (!dig || !dig->afmt)
+		return;
+
+	/* XXX: properly calculate this */
+	/* XXX two dtos; generally use dto0 for hdmi */
+	/* Express [24MHz / target pixel clock] as an exact rational
+	 * number (coefficient of two integer numbers.  DCCG_AUDIO_DTOx_PHASE
+	 * is the numerator, DCCG_AUDIO_DTOx_MODULE is the denominator
+	 */
+	WREG32(DCCG_AUDIO_DTO0_PHASE, (base_rate*50) & 0xffffff);
+	WREG32(DCCG_AUDIO_DTO0_MODULE, (clock*100) & 0xffffff);
+	WREG32(DCCG_AUDIO_DTO_SOURCE, DCCG_AUDIO_DTO0_SOURCE_SEL(radeon_crtc->crtc_id));
+}
+
+
 /*
  * update the info frames with the data from the current display mode
  */
@@ -104,33 +190,19 @@ void evergreen_hdmi_setmode(struct drm_encoder *encoder, struct drm_display_mode
 		return;
 	offset = dig->afmt->offset;
 
-	r600_audio_set_clock(encoder, mode->clock);
+	evergreen_audio_set_dto(encoder, mode->clock);
 
 	WREG32(HDMI_VBI_PACKET_CONTROL + offset,
 	       HDMI_NULL_SEND); /* send null packets when required */
 
 	WREG32(AFMT_AUDIO_CRC_CONTROL + offset, 0x1000);
 
-	WREG32(HDMI_AUDIO_PACKET_CONTROL + offset,
-	       HDMI_AUDIO_DELAY_EN(1) | /* set the default audio delay */
-	       HDMI_AUDIO_PACKETS_PER_LINE(3)); /* should be suffient for all audio modes and small enough for all hblanks */
-
-	WREG32(AFMT_AUDIO_PACKET_CONTROL + offset,
-	       AFMT_AUDIO_SAMPLE_SEND | /* send audio packets */
-	       AFMT_60958_CS_UPDATE); /* allow 60958 channel status fields to be updated */
-
-	WREG32(HDMI_ACR_PACKET_CONTROL + offset,
-	       HDMI_ACR_AUTO_SEND | /* allow hw to sent ACR packets when required */
-	       HDMI_ACR_SOURCE); /* select SW CTS value */
-
 	WREG32(HDMI_VBI_PACKET_CONTROL + offset,
 	       HDMI_NULL_SEND | /* send null packets when required */
 	       HDMI_GC_SEND | /* send general control packets */
 	       HDMI_GC_CONT); /* send general control packets every frame */
 
 	WREG32(HDMI_INFOFRAME_CONTROL0 + offset,
-	       HDMI_AVI_INFO_SEND | /* enable AVI info frames */
-	       HDMI_AVI_INFO_CONT | /* send AVI info frames every frame/field */
 	       HDMI_AUDIO_INFO_SEND | /* enable audio info frames (frames won't be set until audio is enabled) */
 	       HDMI_AUDIO_INFO_CONT); /* required for audio info values to be updated */
 
@@ -138,11 +210,47 @@ void evergreen_hdmi_setmode(struct drm_encoder *encoder, struct drm_display_mode
 	       AFMT_AUDIO_INFO_UPDATE); /* required for audio info values to be updated */
 
 	WREG32(HDMI_INFOFRAME_CONTROL1 + offset,
-	       HDMI_AVI_INFO_LINE(2) | /* anything other than 0 */
 	       HDMI_AUDIO_INFO_LINE(2)); /* anything other than 0 */
 
 	WREG32(HDMI_GC + offset, 0); /* unset HDMI_GC_AVMUTE */
 
+	WREG32(HDMI_AUDIO_PACKET_CONTROL + offset,
+	       HDMI_AUDIO_DELAY_EN(1) | /* set the default audio delay */
+	       HDMI_AUDIO_PACKETS_PER_LINE(3)); /* should be suffient for all audio modes and small enough for all hblanks */
+
+	WREG32(AFMT_AUDIO_PACKET_CONTROL + offset,
+	       AFMT_60958_CS_UPDATE); /* allow 60958 channel status fields to be updated */
+
+	/* fglrx clears sth in AFMT_AUDIO_PACKET_CONTROL2 here */
+
+	WREG32(HDMI_ACR_PACKET_CONTROL + offset,
+	       HDMI_ACR_AUTO_SEND | /* allow hw to sent ACR packets when required */
+	       HDMI_ACR_SOURCE); /* select SW CTS value */
+
+	evergreen_hdmi_update_ACR(encoder, mode->clock);
+
+	WREG32(AFMT_60958_0 + offset,
+	       AFMT_60958_CS_CHANNEL_NUMBER_L(1));
+
+	WREG32(AFMT_60958_1 + offset,
+	       AFMT_60958_CS_CHANNEL_NUMBER_R(2));
+
+	WREG32(AFMT_60958_2 + offset,
+	       AFMT_60958_CS_CHANNEL_NUMBER_2(3) |
+	       AFMT_60958_CS_CHANNEL_NUMBER_3(4) |
+	       AFMT_60958_CS_CHANNEL_NUMBER_4(5) |
+	       AFMT_60958_CS_CHANNEL_NUMBER_5(6) |
+	       AFMT_60958_CS_CHANNEL_NUMBER_6(7) |
+	       AFMT_60958_CS_CHANNEL_NUMBER_7(8));
+
+	/* fglrx sets 0x0001005f | (x & 0x00fc0000) in 0x5f78 here */
+
+	WREG32(AFMT_AUDIO_PACKET_CONTROL2 + offset,
+	       AFMT_AUDIO_CHANNEL_ENABLE(0xff));
+
+	/* fglrx sets 0x40 in 0x5f80 here */
+	evergreen_hdmi_write_sad_regs(encoder);
+
 	err = drm_hdmi_avi_infoframe_from_display_mode(&frame, mode);
 	if (err < 0) {
 		DRM_ERROR("failed to setup AVI infoframe: %zd\n", err);
@@ -156,7 +264,17 @@ void evergreen_hdmi_setmode(struct drm_encoder *encoder, struct drm_display_mode
 	}
 
 	evergreen_hdmi_update_avi_infoframe(encoder, buffer, sizeof(buffer));
-	evergreen_hdmi_update_ACR(encoder, mode->clock);
+
+	WREG32_OR(HDMI_INFOFRAME_CONTROL0 + offset,
+		  HDMI_AVI_INFO_SEND | /* enable AVI info frames */
+		  HDMI_AVI_INFO_CONT); /* required for audio info values to be updated */
+
+	WREG32_P(HDMI_INFOFRAME_CONTROL1 + offset,
+		 HDMI_AVI_INFO_LINE(2), /* anything other than 0 */
+		 ~HDMI_AVI_INFO_LINE_MASK);
+
+	WREG32_OR(AFMT_AUDIO_PACKET_CONTROL + offset,
+		  AFMT_AUDIO_SAMPLE_SEND); /* send audio packets */
 
 	/* it's unknown what these bits do excatly, but it's indeed quite useful for debugging */
 	WREG32(AFMT_RAMP_CONTROL0 + offset, 0x00FFFFFF);
@@ -164,3 +282,20 @@ void evergreen_hdmi_setmode(struct drm_encoder *encoder, struct drm_display_mode
 	WREG32(AFMT_RAMP_CONTROL2 + offset, 0x00000001);
 	WREG32(AFMT_RAMP_CONTROL3 + offset, 0x00000001);
 }
+
+void evergreen_hdmi_enable(struct drm_encoder *encoder, bool enable)
+{
+	struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder);
+	struct radeon_encoder_atom_dig *dig = radeon_encoder->enc_priv;
+
+	/* Silent, r600_hdmi_enable will raise WARN for us */
+	if (enable && dig->afmt->enabled)
+		return;
+	if (!enable && !dig->afmt->enabled)
+		return;
+
+	dig->afmt->enabled = enable;
+
+	DRM_DEBUG("%sabling HDMI interface @ 0x%04X for encoder 0x%x\n",
+		  enable ? "En" : "Dis", dig->afmt->offset, radeon_encoder->encoder_id);
+}
diff --git a/drivers/gpu/drm/radeon/evergreen_reg.h b/drivers/gpu/drm/radeon/evergreen_reg.h
index f585be16e2d5..881aba23c477 100644
--- a/drivers/gpu/drm/radeon/evergreen_reg.h
+++ b/drivers/gpu/drm/radeon/evergreen_reg.h
@@ -226,6 +226,8 @@
 #define EVERGREEN_CRTC_STATUS_HV_COUNT                  0x6ea0
 #define EVERGREEN_MASTER_UPDATE_MODE                    0x6ef8
 #define EVERGREEN_CRTC_UPDATE_LOCK                      0x6ed4
+#define EVERGREEN_MASTER_UPDATE_LOCK                    0x6ef4
+#define EVERGREEN_MASTER_UPDATE_MODE                    0x6ef8
 
 #define EVERGREEN_DC_GPIO_HPD_MASK                      0x64b0
 #define EVERGREEN_DC_GPIO_HPD_A                         0x64b4
diff --git a/drivers/gpu/drm/radeon/evergreend.h b/drivers/gpu/drm/radeon/evergreend.h
index 982d25ad9af3..75c05631146d 100644
--- a/drivers/gpu/drm/radeon/evergreend.h
+++ b/drivers/gpu/drm/radeon/evergreend.h
@@ -53,6 +53,43 @@
 #define RCU_IND_INDEX           			0x100
 #define RCU_IND_DATA            			0x104
 
+/* discrete uvd clocks */
+#define CG_UPLL_FUNC_CNTL				0x718
+#	define UPLL_RESET_MASK				0x00000001
+#	define UPLL_SLEEP_MASK				0x00000002
+#	define UPLL_BYPASS_EN_MASK			0x00000004
+#	define UPLL_CTLREQ_MASK				0x00000008
+#	define UPLL_REF_DIV_MASK			0x003F0000
+#	define UPLL_VCO_MODE_MASK			0x00000200
+#	define UPLL_CTLACK_MASK				0x40000000
+#	define UPLL_CTLACK2_MASK			0x80000000
+#define CG_UPLL_FUNC_CNTL_2				0x71c
+#	define UPLL_PDIV_A(x)				((x) << 0)
+#	define UPLL_PDIV_A_MASK				0x0000007F
+#	define UPLL_PDIV_B(x)				((x) << 8)
+#	define UPLL_PDIV_B_MASK				0x00007F00
+#	define VCLK_SRC_SEL(x)				((x) << 20)
+#	define VCLK_SRC_SEL_MASK			0x01F00000
+#	define DCLK_SRC_SEL(x)				((x) << 25)
+#	define DCLK_SRC_SEL_MASK			0x3E000000
+#define CG_UPLL_FUNC_CNTL_3				0x720
+#	define UPLL_FB_DIV(x)				((x) << 0)
+#	define UPLL_FB_DIV_MASK				0x01FFFFFF
+#define CG_UPLL_FUNC_CNTL_4				0x854
+#	define UPLL_SPARE_ISPARE9			0x00020000
+#define CG_UPLL_SPREAD_SPECTRUM				0x79c
+#	define SSEN_MASK				0x00000001
+
+/* fusion uvd clocks */
+#define CG_DCLK_CNTL                                    0x610
+#       define DCLK_DIVIDER_MASK                        0x7f
+#       define DCLK_DIR_CNTL_EN                         (1 << 8)
+#define CG_DCLK_STATUS                                  0x614
+#       define DCLK_STATUS                              (1 << 0)
+#define CG_VCLK_CNTL                                    0x618
+#define CG_VCLK_STATUS                                  0x61c
+#define	CG_SCRATCH1					0x820
+
 #define GRBM_GFX_INDEX          			0x802C
 #define		INSTANCE_INDEX(x)			((x) << 0)
 #define		SE_INDEX(x)     			((x) << 16)
@@ -197,6 +234,7 @@
 #       define HDMI_MPEG_INFO_CONT           (1 << 9)
 #define HDMI_INFOFRAME_CONTROL1              0x7048
 #       define HDMI_AVI_INFO_LINE(x)         (((x) & 0x3f) << 0)
+#       define HDMI_AVI_INFO_LINE_MASK       (0x3f << 0)
 #       define HDMI_AUDIO_INFO_LINE(x)       (((x) & 0x3f) << 8)
 #       define HDMI_MPEG_INFO_LINE(x)        (((x) & 0x3f) << 16)
 #define HDMI_GENERIC_PACKET_CONTROL          0x704c
@@ -992,6 +1030,16 @@
 #       define TARGET_LINK_SPEED_MASK                     (0xf << 0)
 #       define SELECTABLE_DEEMPHASIS                      (1 << 6)
 
+
+/*
+ * UVD
+ */
+#define UVD_UDEC_ADDR_CONFIG				0xef4c
+#define UVD_UDEC_DB_ADDR_CONFIG				0xef50
+#define UVD_UDEC_DBW_ADDR_CONFIG			0xef54
+#define UVD_RBC_RB_RPTR					0xf690
+#define UVD_RBC_RB_WPTR					0xf694
+
 /*
  * PM4
  */
diff --git a/drivers/gpu/drm/radeon/ni.c b/drivers/gpu/drm/radeon/ni.c
index 27769e724b6d..7969c0c8ec20 100644
--- a/drivers/gpu/drm/radeon/ni.c
+++ b/drivers/gpu/drm/radeon/ni.c
@@ -78,6 +78,282 @@ MODULE_FIRMWARE("radeon/ARUBA_pfp.bin");
 MODULE_FIRMWARE("radeon/ARUBA_me.bin");
 MODULE_FIRMWARE("radeon/ARUBA_rlc.bin");
 
+
+static const u32 cayman_golden_registers2[] =
+{
+	0x3e5c, 0xffffffff, 0x00000000,
+	0x3e48, 0xffffffff, 0x00000000,
+	0x3e4c, 0xffffffff, 0x00000000,
+	0x3e64, 0xffffffff, 0x00000000,
+	0x3e50, 0xffffffff, 0x00000000,
+	0x3e60, 0xffffffff, 0x00000000
+};
+
+static const u32 cayman_golden_registers[] =
+{
+	0x5eb4, 0xffffffff, 0x00000002,
+	0x5e78, 0x8f311ff1, 0x001000f0,
+	0x3f90, 0xffff0000, 0xff000000,
+	0x9148, 0xffff0000, 0xff000000,
+	0x3f94, 0xffff0000, 0xff000000,
+	0x914c, 0xffff0000, 0xff000000,
+	0xc78, 0x00000080, 0x00000080,
+	0xbd4, 0x70073777, 0x00011003,
+	0xd02c, 0xbfffff1f, 0x08421000,
+	0xd0b8, 0x73773777, 0x02011003,
+	0x5bc0, 0x00200000, 0x50100000,
+	0x98f8, 0x33773777, 0x02011003,
+	0x98fc, 0xffffffff, 0x76541032,
+	0x7030, 0x31000311, 0x00000011,
+	0x2f48, 0x33773777, 0x42010001,
+	0x6b28, 0x00000010, 0x00000012,
+	0x7728, 0x00000010, 0x00000012,
+	0x10328, 0x00000010, 0x00000012,
+	0x10f28, 0x00000010, 0x00000012,
+	0x11b28, 0x00000010, 0x00000012,
+	0x12728, 0x00000010, 0x00000012,
+	0x240c, 0x000007ff, 0x00000000,
+	0x8a14, 0xf000001f, 0x00000007,
+	0x8b24, 0x3fff3fff, 0x00ff0fff,
+	0x8b10, 0x0000ff0f, 0x00000000,
+	0x28a4c, 0x07ffffff, 0x06000000,
+	0x10c, 0x00000001, 0x00010003,
+	0xa02c, 0xffffffff, 0x0000009b,
+	0x913c, 0x0000010f, 0x01000100,
+	0x8c04, 0xf8ff00ff, 0x40600060,
+	0x28350, 0x00000f01, 0x00000000,
+	0x9508, 0x3700001f, 0x00000002,
+	0x960c, 0xffffffff, 0x54763210,
+	0x88c4, 0x001f3ae3, 0x00000082,
+	0x88d0, 0xffffffff, 0x0f40df40,
+	0x88d4, 0x0000001f, 0x00000010,
+	0x8974, 0xffffffff, 0x00000000
+};
+
+static const u32 dvst_golden_registers2[] =
+{
+	0x8f8, 0xffffffff, 0,
+	0x8fc, 0x00380000, 0,
+	0x8f8, 0xffffffff, 1,
+	0x8fc, 0x0e000000, 0
+};
+
+static const u32 dvst_golden_registers[] =
+{
+	0x690, 0x3fff3fff, 0x20c00033,
+	0x918c, 0x0fff0fff, 0x00010006,
+	0x91a8, 0x0fff0fff, 0x00010006,
+	0x9150, 0xffffdfff, 0x6e944040,
+	0x917c, 0x0fff0fff, 0x00030002,
+	0x9198, 0x0fff0fff, 0x00030002,
+	0x915c, 0x0fff0fff, 0x00010000,
+	0x3f90, 0xffff0001, 0xff000000,
+	0x9178, 0x0fff0fff, 0x00070000,
+	0x9194, 0x0fff0fff, 0x00070000,
+	0x9148, 0xffff0001, 0xff000000,
+	0x9190, 0x0fff0fff, 0x00090008,
+	0x91ac, 0x0fff0fff, 0x00090008,
+	0x3f94, 0xffff0000, 0xff000000,
+	0x914c, 0xffff0000, 0xff000000,
+	0x929c, 0x00000fff, 0x00000001,
+	0x55e4, 0xff607fff, 0xfc000100,
+	0x8a18, 0xff000fff, 0x00000100,
+	0x8b28, 0xff000fff, 0x00000100,
+	0x9144, 0xfffc0fff, 0x00000100,
+	0x6ed8, 0x00010101, 0x00010000,
+	0x9830, 0xffffffff, 0x00000000,
+	0x9834, 0xf00fffff, 0x00000400,
+	0x9838, 0xfffffffe, 0x00000000,
+	0xd0c0, 0xff000fff, 0x00000100,
+	0xd02c, 0xbfffff1f, 0x08421000,
+	0xd0b8, 0x73773777, 0x12010001,
+	0x5bb0, 0x000000f0, 0x00000070,
+	0x98f8, 0x73773777, 0x12010001,
+	0x98fc, 0xffffffff, 0x00000010,
+	0x9b7c, 0x00ff0000, 0x00fc0000,
+	0x8030, 0x00001f0f, 0x0000100a,
+	0x2f48, 0x73773777, 0x12010001,
+	0x2408, 0x00030000, 0x000c007f,
+	0x8a14, 0xf000003f, 0x00000007,
+	0x8b24, 0x3fff3fff, 0x00ff0fff,
+	0x8b10, 0x0000ff0f, 0x00000000,
+	0x28a4c, 0x07ffffff, 0x06000000,
+	0x4d8, 0x00000fff, 0x00000100,
+	0xa008, 0xffffffff, 0x00010000,
+	0x913c, 0xffff03ff, 0x01000100,
+	0x8c00, 0x000000ff, 0x00000003,
+	0x8c04, 0xf8ff00ff, 0x40600060,
+	0x8cf0, 0x1fff1fff, 0x08e00410,
+	0x28350, 0x00000f01, 0x00000000,
+	0x9508, 0xf700071f, 0x00000002,
+	0x960c, 0xffffffff, 0x54763210,
+	0x20ef8, 0x01ff01ff, 0x00000002,
+	0x20e98, 0xfffffbff, 0x00200000,
+	0x2015c, 0xffffffff, 0x00000f40,
+	0x88c4, 0x001f3ae3, 0x00000082,
+	0x8978, 0x3fffffff, 0x04050140,
+	0x88d4, 0x0000001f, 0x00000010,
+	0x8974, 0xffffffff, 0x00000000
+};
+
+static const u32 scrapper_golden_registers[] =
+{
+	0x690, 0x3fff3fff, 0x20c00033,
+	0x918c, 0x0fff0fff, 0x00010006,
+	0x918c, 0x0fff0fff, 0x00010006,
+	0x91a8, 0x0fff0fff, 0x00010006,
+	0x91a8, 0x0fff0fff, 0x00010006,
+	0x9150, 0xffffdfff, 0x6e944040,
+	0x9150, 0xffffdfff, 0x6e944040,
+	0x917c, 0x0fff0fff, 0x00030002,
+	0x917c, 0x0fff0fff, 0x00030002,
+	0x9198, 0x0fff0fff, 0x00030002,
+	0x9198, 0x0fff0fff, 0x00030002,
+	0x915c, 0x0fff0fff, 0x00010000,
+	0x915c, 0x0fff0fff, 0x00010000,
+	0x3f90, 0xffff0001, 0xff000000,
+	0x3f90, 0xffff0001, 0xff000000,
+	0x9178, 0x0fff0fff, 0x00070000,
+	0x9178, 0x0fff0fff, 0x00070000,
+	0x9194, 0x0fff0fff, 0x00070000,
+	0x9194, 0x0fff0fff, 0x00070000,
+	0x9148, 0xffff0001, 0xff000000,
+	0x9148, 0xffff0001, 0xff000000,
+	0x9190, 0x0fff0fff, 0x00090008,
+	0x9190, 0x0fff0fff, 0x00090008,
+	0x91ac, 0x0fff0fff, 0x00090008,
+	0x91ac, 0x0fff0fff, 0x00090008,
+	0x3f94, 0xffff0000, 0xff000000,
+	0x3f94, 0xffff0000, 0xff000000,
+	0x914c, 0xffff0000, 0xff000000,
+	0x914c, 0xffff0000, 0xff000000,
+	0x929c, 0x00000fff, 0x00000001,
+	0x929c, 0x00000fff, 0x00000001,
+	0x55e4, 0xff607fff, 0xfc000100,
+	0x8a18, 0xff000fff, 0x00000100,
+	0x8a18, 0xff000fff, 0x00000100,
+	0x8b28, 0xff000fff, 0x00000100,
+	0x8b28, 0xff000fff, 0x00000100,
+	0x9144, 0xfffc0fff, 0x00000100,
+	0x9144, 0xfffc0fff, 0x00000100,
+	0x6ed8, 0x00010101, 0x00010000,
+	0x9830, 0xffffffff, 0x00000000,
+	0x9830, 0xffffffff, 0x00000000,
+	0x9834, 0xf00fffff, 0x00000400,
+	0x9834, 0xf00fffff, 0x00000400,
+	0x9838, 0xfffffffe, 0x00000000,
+	0x9838, 0xfffffffe, 0x00000000,
+	0xd0c0, 0xff000fff, 0x00000100,
+	0xd02c, 0xbfffff1f, 0x08421000,
+	0xd02c, 0xbfffff1f, 0x08421000,
+	0xd0b8, 0x73773777, 0x12010001,
+	0xd0b8, 0x73773777, 0x12010001,
+	0x5bb0, 0x000000f0, 0x00000070,
+	0x98f8, 0x73773777, 0x12010001,
+	0x98f8, 0x73773777, 0x12010001,
+	0x98fc, 0xffffffff, 0x00000010,
+	0x98fc, 0xffffffff, 0x00000010,
+	0x9b7c, 0x00ff0000, 0x00fc0000,
+	0x9b7c, 0x00ff0000, 0x00fc0000,
+	0x8030, 0x00001f0f, 0x0000100a,
+	0x8030, 0x00001f0f, 0x0000100a,
+	0x2f48, 0x73773777, 0x12010001,
+	0x2f48, 0x73773777, 0x12010001,
+	0x2408, 0x00030000, 0x000c007f,
+	0x8a14, 0xf000003f, 0x00000007,
+	0x8a14, 0xf000003f, 0x00000007,
+	0x8b24, 0x3fff3fff, 0x00ff0fff,
+	0x8b24, 0x3fff3fff, 0x00ff0fff,
+	0x8b10, 0x0000ff0f, 0x00000000,
+	0x8b10, 0x0000ff0f, 0x00000000,
+	0x28a4c, 0x07ffffff, 0x06000000,
+	0x28a4c, 0x07ffffff, 0x06000000,
+	0x4d8, 0x00000fff, 0x00000100,
+	0x4d8, 0x00000fff, 0x00000100,
+	0xa008, 0xffffffff, 0x00010000,
+	0xa008, 0xffffffff, 0x00010000,
+	0x913c, 0xffff03ff, 0x01000100,
+	0x913c, 0xffff03ff, 0x01000100,
+	0x90e8, 0x001fffff, 0x010400c0,
+	0x8c00, 0x000000ff, 0x00000003,
+	0x8c00, 0x000000ff, 0x00000003,
+	0x8c04, 0xf8ff00ff, 0x40600060,
+	0x8c04, 0xf8ff00ff, 0x40600060,
+	0x8c30, 0x0000000f, 0x00040005,
+	0x8cf0, 0x1fff1fff, 0x08e00410,
+	0x8cf0, 0x1fff1fff, 0x08e00410,
+	0x900c, 0x00ffffff, 0x0017071f,
+	0x28350, 0x00000f01, 0x00000000,
+	0x28350, 0x00000f01, 0x00000000,
+	0x9508, 0xf700071f, 0x00000002,
+	0x9508, 0xf700071f, 0x00000002,
+	0x9688, 0x00300000, 0x0017000f,
+	0x960c, 0xffffffff, 0x54763210,
+	0x960c, 0xffffffff, 0x54763210,
+	0x20ef8, 0x01ff01ff, 0x00000002,
+	0x20e98, 0xfffffbff, 0x00200000,
+	0x2015c, 0xffffffff, 0x00000f40,
+	0x88c4, 0x001f3ae3, 0x00000082,
+	0x88c4, 0x001f3ae3, 0x00000082,
+	0x8978, 0x3fffffff, 0x04050140,
+	0x8978, 0x3fffffff, 0x04050140,
+	0x88d4, 0x0000001f, 0x00000010,
+	0x88d4, 0x0000001f, 0x00000010,
+	0x8974, 0xffffffff, 0x00000000,
+	0x8974, 0xffffffff, 0x00000000
+};
+
+static void ni_init_golden_registers(struct radeon_device *rdev)
+{
+	switch (rdev->family) {
+	case CHIP_CAYMAN:
+		radeon_program_register_sequence(rdev,
+						 cayman_golden_registers,
+						 (const u32)ARRAY_SIZE(cayman_golden_registers));
+		radeon_program_register_sequence(rdev,
+						 cayman_golden_registers2,
+						 (const u32)ARRAY_SIZE(cayman_golden_registers2));
+		break;
+	case CHIP_ARUBA:
+		if ((rdev->pdev->device == 0x9900) ||
+		    (rdev->pdev->device == 0x9901) ||
+		    (rdev->pdev->device == 0x9903) ||
+		    (rdev->pdev->device == 0x9904) ||
+		    (rdev->pdev->device == 0x9905) ||
+		    (rdev->pdev->device == 0x9906) ||
+		    (rdev->pdev->device == 0x9907) ||
+		    (rdev->pdev->device == 0x9908) ||
+		    (rdev->pdev->device == 0x9909) ||
+		    (rdev->pdev->device == 0x990A) ||
+		    (rdev->pdev->device == 0x990B) ||
+		    (rdev->pdev->device == 0x990C) ||
+		    (rdev->pdev->device == 0x990D) ||
+		    (rdev->pdev->device == 0x990E) ||
+		    (rdev->pdev->device == 0x990F) ||
+		    (rdev->pdev->device == 0x9910) ||
+		    (rdev->pdev->device == 0x9913) ||
+		    (rdev->pdev->device == 0x9917) ||
+		    (rdev->pdev->device == 0x9918)) {
+			radeon_program_register_sequence(rdev,
+							 dvst_golden_registers,
+							 (const u32)ARRAY_SIZE(dvst_golden_registers));
+			radeon_program_register_sequence(rdev,
+							 dvst_golden_registers2,
+							 (const u32)ARRAY_SIZE(dvst_golden_registers2));
+		} else {
+			radeon_program_register_sequence(rdev,
+							 scrapper_golden_registers,
+							 (const u32)ARRAY_SIZE(scrapper_golden_registers));
+			radeon_program_register_sequence(rdev,
+							 dvst_golden_registers2,
+							 (const u32)ARRAY_SIZE(dvst_golden_registers2));
+		}
+		break;
+	default:
+		break;
+	}
+}
+
 #define BTC_IO_MC_REGS_SIZE 29
 
 static const u32 barts_io_mc_regs[BTC_IO_MC_REGS_SIZE][2] = {
@@ -473,7 +749,8 @@ static void cayman_gpu_init(struct radeon_device *rdev)
 		    (rdev->pdev->device == 0x990F) ||
 		    (rdev->pdev->device == 0x9910) ||
 		    (rdev->pdev->device == 0x9917) ||
-		    (rdev->pdev->device == 0x9999)) {
+		    (rdev->pdev->device == 0x9999) ||
+		    (rdev->pdev->device == 0x999C)) {
 			rdev->config.cayman.max_simds_per_se = 6;
 			rdev->config.cayman.max_backends_per_se = 2;
 		} else if ((rdev->pdev->device == 0x9903) ||
@@ -482,7 +759,8 @@ static void cayman_gpu_init(struct radeon_device *rdev)
 			   (rdev->pdev->device == 0x990D) ||
 			   (rdev->pdev->device == 0x990E) ||
 			   (rdev->pdev->device == 0x9913) ||
-			   (rdev->pdev->device == 0x9918)) {
+			   (rdev->pdev->device == 0x9918) ||
+			   (rdev->pdev->device == 0x999D)) {
 			rdev->config.cayman.max_simds_per_se = 4;
 			rdev->config.cayman.max_backends_per_se = 2;
 		} else if ((rdev->pdev->device == 0x9919) ||
@@ -615,15 +893,28 @@ static void cayman_gpu_init(struct radeon_device *rdev)
 	}
 	/* enabled rb are just the one not disabled :) */
 	disabled_rb_mask = tmp;
+	tmp = 0;
+	for (i = 0; i < (rdev->config.cayman.max_backends_per_se * rdev->config.cayman.max_shader_engines); i++)
+		tmp |= (1 << i);
+	/* if all the backends are disabled, fix it up here */
+	if ((disabled_rb_mask & tmp) == tmp) {
+		for (i = 0; i < (rdev->config.cayman.max_backends_per_se * rdev->config.cayman.max_shader_engines); i++)
+			disabled_rb_mask &= ~(1 << i);
+	}
 
 	WREG32(GRBM_GFX_INDEX, INSTANCE_BROADCAST_WRITES | SE_BROADCAST_WRITES);
 	WREG32(RLC_GFX_INDEX, INSTANCE_BROADCAST_WRITES | SE_BROADCAST_WRITES);
 
 	WREG32(GB_ADDR_CONFIG, gb_addr_config);
 	WREG32(DMIF_ADDR_CONFIG, gb_addr_config);
+	if (ASIC_IS_DCE6(rdev))
+		WREG32(DMIF_ADDR_CALC, gb_addr_config);
 	WREG32(HDP_ADDR_CONFIG, gb_addr_config);
 	WREG32(DMA_TILING_CONFIG + DMA0_REGISTER_OFFSET, gb_addr_config);
 	WREG32(DMA_TILING_CONFIG + DMA1_REGISTER_OFFSET, gb_addr_config);
+	WREG32(UVD_UDEC_ADDR_CONFIG, gb_addr_config);
+	WREG32(UVD_UDEC_DB_ADDR_CONFIG, gb_addr_config);
+	WREG32(UVD_UDEC_DBW_ADDR_CONFIG, gb_addr_config);
 
 	if ((rdev->config.cayman.max_backends_per_se == 1) &&
 	    (rdev->flags & RADEON_IS_IGP)) {
@@ -931,6 +1222,23 @@ void cayman_ring_ib_execute(struct radeon_device *rdev, struct radeon_ib *ib)
 	radeon_ring_write(ring, 10); /* poll interval */
 }
 
+void cayman_uvd_semaphore_emit(struct radeon_device *rdev,
+			       struct radeon_ring *ring,
+			       struct radeon_semaphore *semaphore,
+			       bool emit_wait)
+{
+	uint64_t addr = semaphore->gpu_addr;
+
+	radeon_ring_write(ring, PACKET0(UVD_SEMA_ADDR_LOW, 0));
+	radeon_ring_write(ring, (addr >> 3) & 0x000FFFFF);
+
+	radeon_ring_write(ring, PACKET0(UVD_SEMA_ADDR_HIGH, 0));
+	radeon_ring_write(ring, (addr >> 23) & 0x000FFFFF);
+
+	radeon_ring_write(ring, PACKET0(UVD_SEMA_CMD, 0));
+	radeon_ring_write(ring, 0x80 | (emit_wait ? 1 : 0));
+}
+
 static void cayman_cp_enable(struct radeon_device *rdev, bool enable)
 {
 	if (enable)
@@ -1682,6 +1990,16 @@ static int cayman_startup(struct radeon_device *rdev)
 		return r;
 	}
 
+	r = rv770_uvd_resume(rdev);
+	if (!r) {
+		r = radeon_fence_driver_start_ring(rdev,
+						   R600_RING_TYPE_UVD_INDEX);
+		if (r)
+			dev_err(rdev->dev, "UVD fences init error (%d).\n", r);
+	}
+	if (r)
+		rdev->ring[R600_RING_TYPE_UVD_INDEX].ring_size = 0;
+
 	r = radeon_fence_driver_start_ring(rdev, CAYMAN_RING_TYPE_CP1_INDEX);
 	if (r) {
 		dev_err(rdev->dev, "failed initializing CP fences (%d).\n", r);
@@ -1748,6 +2066,18 @@ static int cayman_startup(struct radeon_device *rdev)
 	if (r)
 		return r;
 
+	ring = &rdev->ring[R600_RING_TYPE_UVD_INDEX];
+	if (ring->ring_size) {
+		r = radeon_ring_init(rdev, ring, ring->ring_size,
+				     R600_WB_UVD_RPTR_OFFSET,
+				     UVD_RBC_RB_RPTR, UVD_RBC_RB_WPTR,
+				     0, 0xfffff, RADEON_CP_PACKET2);
+		if (!r)
+			r = r600_uvd_init(rdev);
+		if (r)
+			DRM_ERROR("radeon: failed initializing UVD (%d).\n", r);
+	}
+
 	r = radeon_ib_pool_init(rdev);
 	if (r) {
 		dev_err(rdev->dev, "IB initialization failed (%d).\n", r);
@@ -1778,6 +2108,9 @@ int cayman_resume(struct radeon_device *rdev)
 	/* post card */
 	atom_asic_init(rdev->mode_info.atom_context);
 
+	/* init golden registers */
+	ni_init_golden_registers(rdev);
+
 	rdev->accel_working = true;
 	r = cayman_startup(rdev);
 	if (r) {
@@ -1794,6 +2127,8 @@ int cayman_suspend(struct radeon_device *rdev)
 	radeon_vm_manager_fini(rdev);
 	cayman_cp_enable(rdev, false);
 	cayman_dma_stop(rdev);
+	r600_uvd_rbc_stop(rdev);
+	radeon_uvd_suspend(rdev);
 	evergreen_irq_suspend(rdev);
 	radeon_wb_disable(rdev);
 	cayman_pcie_gart_disable(rdev);
@@ -1834,6 +2169,8 @@ int cayman_init(struct radeon_device *rdev)
 		DRM_INFO("GPU not posted. posting now...\n");
 		atom_asic_init(rdev->mode_info.atom_context);
 	}
+	/* init golden registers */
+	ni_init_golden_registers(rdev);
 	/* Initialize scratch registers */
 	r600_scratch_init(rdev);
 	/* Initialize surface registers */
@@ -1868,6 +2205,13 @@ int cayman_init(struct radeon_device *rdev)
 	ring->ring_obj = NULL;
 	r600_ring_init(rdev, ring, 64 * 1024);
 
+	r = radeon_uvd_init(rdev);
+	if (!r) {
+		ring = &rdev->ring[R600_RING_TYPE_UVD_INDEX];
+		ring->ring_obj = NULL;
+		r600_ring_init(rdev, ring, 4096);
+	}
+
 	rdev->ih.ring_obj = NULL;
 	r600_ih_ring_init(rdev, 64 * 1024);
 
@@ -1919,6 +2263,7 @@ void cayman_fini(struct radeon_device *rdev)
 	radeon_vm_manager_fini(rdev);
 	radeon_ib_pool_fini(rdev);
 	radeon_irq_kms_fini(rdev);
+	radeon_uvd_fini(rdev);
 	cayman_pcie_gart_fini(rdev);
 	r600_vram_scratch_fini(rdev);
 	radeon_gem_fini(rdev);
@@ -2017,28 +2362,57 @@ void cayman_vm_set_page(struct radeon_device *rdev,
 			}
 		}
 	} else {
-		while (count) {
-			ndw = count * 2;
-			if (ndw > 0xFFFFE)
-				ndw = 0xFFFFE;
+		if ((flags & RADEON_VM_PAGE_SYSTEM) ||
+		    (count == 1)) {
+			while (count) {
+				ndw = count * 2;
+				if (ndw > 0xFFFFE)
+					ndw = 0xFFFFE;
+
+				/* for non-physically contiguous pages (system) */
+				ib->ptr[ib->length_dw++] = DMA_PACKET(DMA_PACKET_WRITE, 0, 0, ndw);
+				ib->ptr[ib->length_dw++] = pe;
+				ib->ptr[ib->length_dw++] = upper_32_bits(pe) & 0xff;
+				for (; ndw > 0; ndw -= 2, --count, pe += 8) {
+					if (flags & RADEON_VM_PAGE_SYSTEM) {
+						value = radeon_vm_map_gart(rdev, addr);
+						value &= 0xFFFFFFFFFFFFF000ULL;
+					} else if (flags & RADEON_VM_PAGE_VALID) {
+						value = addr;
+					} else {
+						value = 0;
+					}
+					addr += incr;
+					value |= r600_flags;
+					ib->ptr[ib->length_dw++] = value;
+					ib->ptr[ib->length_dw++] = upper_32_bits(value);
+				}
+			}
+			while (ib->length_dw & 0x7)
+				ib->ptr[ib->length_dw++] = DMA_PACKET(DMA_PACKET_NOP, 0, 0, 0);
+		} else {
+			while (count) {
+				ndw = count * 2;
+				if (ndw > 0xFFFFE)
+					ndw = 0xFFFFE;
 
-			/* for non-physically contiguous pages (system) */
-			ib->ptr[ib->length_dw++] = DMA_PACKET(DMA_PACKET_WRITE, 0, 0, ndw);
-			ib->ptr[ib->length_dw++] = pe;
-			ib->ptr[ib->length_dw++] = upper_32_bits(pe) & 0xff;
-			for (; ndw > 0; ndw -= 2, --count, pe += 8) {
-				if (flags & RADEON_VM_PAGE_SYSTEM) {
-					value = radeon_vm_map_gart(rdev, addr);
-					value &= 0xFFFFFFFFFFFFF000ULL;
-				} else if (flags & RADEON_VM_PAGE_VALID) {
+				if (flags & RADEON_VM_PAGE_VALID)
 					value = addr;
-				} else {
+				else
 					value = 0;
-				}
-				addr += incr;
-				value |= r600_flags;
-				ib->ptr[ib->length_dw++] = value;
+				/* for physically contiguous pages (vram) */
+				ib->ptr[ib->length_dw++] = DMA_PTE_PDE_PACKET(ndw);
+				ib->ptr[ib->length_dw++] = pe; /* dst addr */
+				ib->ptr[ib->length_dw++] = upper_32_bits(pe) & 0xff;
+				ib->ptr[ib->length_dw++] = r600_flags; /* mask */
+				ib->ptr[ib->length_dw++] = 0;
+				ib->ptr[ib->length_dw++] = value; /* value */
 				ib->ptr[ib->length_dw++] = upper_32_bits(value);
+				ib->ptr[ib->length_dw++] = incr; /* increment size */
+				ib->ptr[ib->length_dw++] = 0;
+				pe += ndw * 4;
+				addr += (ndw / 2) * incr;
+				count -= ndw / 2;
 			}
 		}
 		while (ib->length_dw & 0x7)
diff --git a/drivers/gpu/drm/radeon/nid.h b/drivers/gpu/drm/radeon/nid.h
index 079dee202a9e..e226faf16fea 100644
--- a/drivers/gpu/drm/radeon/nid.h
+++ b/drivers/gpu/drm/radeon/nid.h
@@ -45,6 +45,10 @@
 #define ARUBA_GB_ADDR_CONFIG_GOLDEN        0x12010001
 
 #define DMIF_ADDR_CONFIG  				0xBD4
+
+/* DCE6 only */
+#define DMIF_ADDR_CALC  				0xC00
+
 #define	SRBM_GFX_CNTL				        0x0E44
 #define		RINGID(x)					(((x) & 0x3) << 0)
 #define		VMID(x)						(((x) & 0x7) << 0)
@@ -486,6 +490,18 @@
 #       define CACHE_FLUSH_AND_INV_EVENT                        (0x16 << 0)
 
 /*
+ * UVD
+ */
+#define UVD_SEMA_ADDR_LOW				0xEF00
+#define UVD_SEMA_ADDR_HIGH				0xEF04
+#define UVD_SEMA_CMD					0xEF08
+#define UVD_UDEC_ADDR_CONFIG				0xEF4C
+#define UVD_UDEC_DB_ADDR_CONFIG				0xEF50
+#define UVD_UDEC_DBW_ADDR_CONFIG			0xEF54
+#define UVD_RBC_RB_RPTR					0xF690
+#define UVD_RBC_RB_WPTR					0xF694
+
+/*
  * PM4
  */
 #define PACKET0(reg, n)	((RADEON_PACKET_TYPE0 << 30) |			\
@@ -668,6 +684,11 @@
 					 (((vmid) & 0xF) << 20) |	\
 					 (((n) & 0xFFFFF) << 0))
 
+#define DMA_PTE_PDE_PACKET(n)		((2 << 28) |			\
+					 (1 << 26) |			\
+					 (1 << 21) |			\
+					 (((n) & 0xFFFFF) << 0))
+
 /* async DMA Packet types */
 #define	DMA_PACKET_WRITE				  0x2
 #define	DMA_PACKET_COPY					  0x3
diff --git a/drivers/gpu/drm/radeon/r100.c b/drivers/gpu/drm/radeon/r100.c
index 9db58530be37..4973bff37fec 100644
--- a/drivers/gpu/drm/radeon/r100.c
+++ b/drivers/gpu/drm/radeon/r100.c
@@ -69,6 +69,38 @@ MODULE_FIRMWARE(FIRMWARE_R520);
  * and others in some cases.
  */
 
+static bool r100_is_in_vblank(struct radeon_device *rdev, int crtc)
+{
+	if (crtc == 0) {
+		if (RREG32(RADEON_CRTC_STATUS) & RADEON_CRTC_VBLANK_CUR)
+			return true;
+		else
+			return false;
+	} else {
+		if (RREG32(RADEON_CRTC2_STATUS) & RADEON_CRTC2_VBLANK_CUR)
+			return true;
+		else
+			return false;
+	}
+}
+
+static bool r100_is_counter_moving(struct radeon_device *rdev, int crtc)
+{
+	u32 vline1, vline2;
+
+	if (crtc == 0) {
+		vline1 = (RREG32(RADEON_CRTC_VLINE_CRNT_VLINE) >> 16) & RADEON_CRTC_V_TOTAL;
+		vline2 = (RREG32(RADEON_CRTC_VLINE_CRNT_VLINE) >> 16) & RADEON_CRTC_V_TOTAL;
+	} else {
+		vline1 = (RREG32(RADEON_CRTC2_VLINE_CRNT_VLINE) >> 16) & RADEON_CRTC_V_TOTAL;
+		vline2 = (RREG32(RADEON_CRTC2_VLINE_CRNT_VLINE) >> 16) & RADEON_CRTC_V_TOTAL;
+	}
+	if (vline1 != vline2)
+		return true;
+	else
+		return false;
+}
+
 /**
  * r100_wait_for_vblank - vblank wait asic callback.
  *
@@ -79,36 +111,33 @@ MODULE_FIRMWARE(FIRMWARE_R520);
  */
 void r100_wait_for_vblank(struct radeon_device *rdev, int crtc)
 {
-	int i;
+	unsigned i = 0;
 
 	if (crtc >= rdev->num_crtc)
 		return;
 
 	if (crtc == 0) {
-		if (RREG32(RADEON_CRTC_GEN_CNTL) & RADEON_CRTC_EN) {
-			for (i = 0; i < rdev->usec_timeout; i++) {
-				if (!(RREG32(RADEON_CRTC_STATUS) & RADEON_CRTC_VBLANK_CUR))
-					break;
-				udelay(1);
-			}
-			for (i = 0; i < rdev->usec_timeout; i++) {
-				if (RREG32(RADEON_CRTC_STATUS) & RADEON_CRTC_VBLANK_CUR)
-					break;
-				udelay(1);
-			}
-		}
+		if (!(RREG32(RADEON_CRTC_GEN_CNTL) & RADEON_CRTC_EN))
+			return;
 	} else {
-		if (RREG32(RADEON_CRTC2_GEN_CNTL) & RADEON_CRTC2_EN) {
-			for (i = 0; i < rdev->usec_timeout; i++) {
-				if (!(RREG32(RADEON_CRTC2_STATUS) & RADEON_CRTC2_VBLANK_CUR))
-					break;
-				udelay(1);
-			}
-			for (i = 0; i < rdev->usec_timeout; i++) {
-				if (RREG32(RADEON_CRTC2_STATUS) & RADEON_CRTC2_VBLANK_CUR)
-					break;
-				udelay(1);
-			}
+		if (!(RREG32(RADEON_CRTC2_GEN_CNTL) & RADEON_CRTC2_EN))
+			return;
+	}
+
+	/* depending on when we hit vblank, we may be close to active; if so,
+	 * wait for another frame.
+	 */
+	while (r100_is_in_vblank(rdev, crtc)) {
+		if (i++ % 100 == 0) {
+			if (!r100_is_counter_moving(rdev, crtc))
+				break;
+		}
+	}
+
+	while (!r100_is_in_vblank(rdev, crtc)) {
+		if (i++ % 100 == 0) {
+			if (!r100_is_counter_moving(rdev, crtc))
+				break;
 		}
 	}
 }
diff --git a/drivers/gpu/drm/radeon/r500_reg.h b/drivers/gpu/drm/radeon/r500_reg.h
index c0dc8d3ba0bb..1dd0d32993d5 100644
--- a/drivers/gpu/drm/radeon/r500_reg.h
+++ b/drivers/gpu/drm/radeon/r500_reg.h
@@ -358,7 +358,9 @@
 #define AVIVO_D1CRTC_STATUS_HV_COUNT                            0x60ac
 #define AVIVO_D1CRTC_STEREO_CONTROL                             0x60c4
 
+#define AVIVO_D1MODE_MASTER_UPDATE_LOCK                         0x60e0
 #define AVIVO_D1MODE_MASTER_UPDATE_MODE                         0x60e4
+#define AVIVO_D1CRTC_UPDATE_LOCK                                0x60e8
 
 /* master controls */
 #define AVIVO_DC_CRTC_MASTER_EN                                 0x60f8
diff --git a/drivers/gpu/drm/radeon/r600.c b/drivers/gpu/drm/radeon/r600.c
index 0740db3fcd22..1a08008c978b 100644
--- a/drivers/gpu/drm/radeon/r600.c
+++ b/drivers/gpu/drm/radeon/r600.c
@@ -1145,7 +1145,7 @@ static void r600_vram_gtt_location(struct radeon_device *rdev, struct radeon_mc
 	}
 	if (rdev->flags & RADEON_IS_AGP) {
 		size_bf = mc->gtt_start;
-		size_af = 0xFFFFFFFF - mc->gtt_end;
+		size_af = mc->mc_mask - mc->gtt_end;
 		if (size_bf > size_af) {
 			if (mc->mc_vram_size > size_bf) {
 				dev_warn(rdev->dev, "limiting VRAM\n");
@@ -2552,6 +2552,193 @@ void r600_dma_fini(struct radeon_device *rdev)
 }
 
 /*
+ * UVD
+ */
+int r600_uvd_rbc_start(struct radeon_device *rdev)
+{
+	struct radeon_ring *ring = &rdev->ring[R600_RING_TYPE_UVD_INDEX];
+	uint64_t rptr_addr;
+	uint32_t rb_bufsz, tmp;
+	int r;
+
+	rptr_addr = rdev->wb.gpu_addr + R600_WB_UVD_RPTR_OFFSET;
+
+	if (upper_32_bits(rptr_addr) != upper_32_bits(ring->gpu_addr)) {
+		DRM_ERROR("UVD ring and rptr not in the same 4GB segment!\n");
+		return -EINVAL;
+	}
+
+	/* force RBC into idle state */
+	WREG32(UVD_RBC_RB_CNTL, 0x11010101);
+
+	/* Set the write pointer delay */
+	WREG32(UVD_RBC_RB_WPTR_CNTL, 0);
+
+	/* set the wb address */
+	WREG32(UVD_RBC_RB_RPTR_ADDR, rptr_addr >> 2);
+
+	/* programm the 4GB memory segment for rptr and ring buffer */
+	WREG32(UVD_LMI_EXT40_ADDR, upper_32_bits(rptr_addr) |
+				   (0x7 << 16) | (0x1 << 31));
+
+	/* Initialize the ring buffer's read and write pointers */
+	WREG32(UVD_RBC_RB_RPTR, 0x0);
+
+	ring->wptr = ring->rptr = RREG32(UVD_RBC_RB_RPTR);
+	WREG32(UVD_RBC_RB_WPTR, ring->wptr);
+
+	/* set the ring address */
+	WREG32(UVD_RBC_RB_BASE, ring->gpu_addr);
+
+	/* Set ring buffer size */
+	rb_bufsz = drm_order(ring->ring_size);
+	rb_bufsz = (0x1 << 8) | rb_bufsz;
+	WREG32(UVD_RBC_RB_CNTL, rb_bufsz);
+
+	ring->ready = true;
+	r = radeon_ring_test(rdev, R600_RING_TYPE_UVD_INDEX, ring);
+	if (r) {
+		ring->ready = false;
+		return r;
+	}
+
+	r = radeon_ring_lock(rdev, ring, 10);
+	if (r) {
+		DRM_ERROR("radeon: ring failed to lock UVD ring (%d).\n", r);
+		return r;
+	}
+
+	tmp = PACKET0(UVD_SEMA_WAIT_FAULT_TIMEOUT_CNTL, 0);
+	radeon_ring_write(ring, tmp);
+	radeon_ring_write(ring, 0xFFFFF);
+
+	tmp = PACKET0(UVD_SEMA_WAIT_INCOMPLETE_TIMEOUT_CNTL, 0);
+	radeon_ring_write(ring, tmp);
+	radeon_ring_write(ring, 0xFFFFF);
+
+	tmp = PACKET0(UVD_SEMA_SIGNAL_INCOMPLETE_TIMEOUT_CNTL, 0);
+	radeon_ring_write(ring, tmp);
+	radeon_ring_write(ring, 0xFFFFF);
+
+	/* Clear timeout status bits */
+	radeon_ring_write(ring, PACKET0(UVD_SEMA_TIMEOUT_STATUS, 0));
+	radeon_ring_write(ring, 0x8);
+
+	radeon_ring_write(ring, PACKET0(UVD_SEMA_CNTL, 0));
+	radeon_ring_write(ring, 3);
+
+	radeon_ring_unlock_commit(rdev, ring);
+
+	return 0;
+}
+
+void r600_uvd_rbc_stop(struct radeon_device *rdev)
+{
+	struct radeon_ring *ring = &rdev->ring[R600_RING_TYPE_UVD_INDEX];
+
+	/* force RBC into idle state */
+	WREG32(UVD_RBC_RB_CNTL, 0x11010101);
+	ring->ready = false;
+}
+
+int r600_uvd_init(struct radeon_device *rdev)
+{
+	int i, j, r;
+
+	/* raise clocks while booting up the VCPU */
+	radeon_set_uvd_clocks(rdev, 53300, 40000);
+
+	/* disable clock gating */
+	WREG32(UVD_CGC_GATE, 0);
+
+	/* disable interupt */
+	WREG32_P(UVD_MASTINT_EN, 0, ~(1 << 1));
+
+	/* put LMI, VCPU, RBC etc... into reset */
+	WREG32(UVD_SOFT_RESET, LMI_SOFT_RESET | VCPU_SOFT_RESET |
+	       LBSI_SOFT_RESET | RBC_SOFT_RESET | CSM_SOFT_RESET |
+	       CXW_SOFT_RESET | TAP_SOFT_RESET | LMI_UMC_SOFT_RESET);
+	mdelay(5);
+
+	/* take UVD block out of reset */
+	WREG32_P(SRBM_SOFT_RESET, 0, ~SOFT_RESET_UVD);
+	mdelay(5);
+
+	/* initialize UVD memory controller */
+	WREG32(UVD_LMI_CTRL, 0x40 | (1 << 8) | (1 << 13) |
+			     (1 << 21) | (1 << 9) | (1 << 20));
+
+	/* disable byte swapping */
+	WREG32(UVD_LMI_SWAP_CNTL, 0);
+	WREG32(UVD_MP_SWAP_CNTL, 0);
+
+	WREG32(UVD_MPC_SET_MUXA0, 0x40c2040);
+	WREG32(UVD_MPC_SET_MUXA1, 0x0);
+	WREG32(UVD_MPC_SET_MUXB0, 0x40c2040);
+	WREG32(UVD_MPC_SET_MUXB1, 0x0);
+	WREG32(UVD_MPC_SET_ALU, 0);
+	WREG32(UVD_MPC_SET_MUX, 0x88);
+
+	/* Stall UMC */
+	WREG32_P(UVD_LMI_CTRL2, 1 << 8, ~(1 << 8));
+	WREG32_P(UVD_RB_ARB_CTRL, 1 << 3, ~(1 << 3));
+
+	/* take all subblocks out of reset, except VCPU */
+	WREG32(UVD_SOFT_RESET, VCPU_SOFT_RESET);
+	mdelay(5);
+
+	/* enable VCPU clock */
+	WREG32(UVD_VCPU_CNTL,  1 << 9);
+
+	/* enable UMC */
+	WREG32_P(UVD_LMI_CTRL2, 0, ~(1 << 8));
+
+	/* boot up the VCPU */
+	WREG32(UVD_SOFT_RESET, 0);
+	mdelay(10);
+
+	WREG32_P(UVD_RB_ARB_CTRL, 0, ~(1 << 3));
+
+	for (i = 0; i < 10; ++i) {
+		uint32_t status;
+		for (j = 0; j < 100; ++j) {
+			status = RREG32(UVD_STATUS);
+			if (status & 2)
+				break;
+			mdelay(10);
+		}
+		r = 0;
+		if (status & 2)
+			break;
+
+		DRM_ERROR("UVD not responding, trying to reset the VCPU!!!\n");
+		WREG32_P(UVD_SOFT_RESET, VCPU_SOFT_RESET, ~VCPU_SOFT_RESET);
+		mdelay(10);
+		WREG32_P(UVD_SOFT_RESET, 0, ~VCPU_SOFT_RESET);
+		mdelay(10);
+		r = -1;
+	}
+
+	if (r) {
+		DRM_ERROR("UVD not responding, giving up!!!\n");
+		radeon_set_uvd_clocks(rdev, 0, 0);
+		return r;
+	}
+
+	/* enable interupt */
+	WREG32_P(UVD_MASTINT_EN, 3<<1, ~(3 << 1));
+
+	r = r600_uvd_rbc_start(rdev);
+	if (!r)
+		DRM_INFO("UVD initialized successfully.\n");
+
+	/* lower clocks again */
+	radeon_set_uvd_clocks(rdev, 0, 0);
+
+	return r;
+}
+
+/*
  * GPU scratch registers helpers function.
  */
 void r600_scratch_init(struct radeon_device *rdev)
@@ -2660,6 +2847,40 @@ int r600_dma_ring_test(struct radeon_device *rdev,
 	return r;
 }
 
+int r600_uvd_ring_test(struct radeon_device *rdev, struct radeon_ring *ring)
+{
+	uint32_t tmp = 0;
+	unsigned i;
+	int r;
+
+	WREG32(UVD_CONTEXT_ID, 0xCAFEDEAD);
+	r = radeon_ring_lock(rdev, ring, 3);
+	if (r) {
+		DRM_ERROR("radeon: cp failed to lock ring %d (%d).\n",
+			  ring->idx, r);
+		return r;
+	}
+	radeon_ring_write(ring, PACKET0(UVD_CONTEXT_ID, 0));
+	radeon_ring_write(ring, 0xDEADBEEF);
+	radeon_ring_unlock_commit(rdev, ring);
+	for (i = 0; i < rdev->usec_timeout; i++) {
+		tmp = RREG32(UVD_CONTEXT_ID);
+		if (tmp == 0xDEADBEEF)
+			break;
+		DRM_UDELAY(1);
+	}
+
+	if (i < rdev->usec_timeout) {
+		DRM_INFO("ring test on %d succeeded in %d usecs\n",
+			 ring->idx, i);
+	} else {
+		DRM_ERROR("radeon: ring %d test failed (0x%08X)\n",
+			  ring->idx, tmp);
+		r = -EINVAL;
+	}
+	return r;
+}
+
 /*
  * CP fences/semaphores
  */
@@ -2711,6 +2932,30 @@ void r600_fence_ring_emit(struct radeon_device *rdev,
 	}
 }
 
+void r600_uvd_fence_emit(struct radeon_device *rdev,
+			 struct radeon_fence *fence)
+{
+	struct radeon_ring *ring = &rdev->ring[fence->ring];
+	uint32_t addr = rdev->fence_drv[fence->ring].gpu_addr;
+
+	radeon_ring_write(ring, PACKET0(UVD_CONTEXT_ID, 0));
+	radeon_ring_write(ring, fence->seq);
+	radeon_ring_write(ring, PACKET0(UVD_GPCOM_VCPU_DATA0, 0));
+	radeon_ring_write(ring, addr & 0xffffffff);
+	radeon_ring_write(ring, PACKET0(UVD_GPCOM_VCPU_DATA1, 0));
+	radeon_ring_write(ring, upper_32_bits(addr) & 0xff);
+	radeon_ring_write(ring, PACKET0(UVD_GPCOM_VCPU_CMD, 0));
+	radeon_ring_write(ring, 0);
+
+	radeon_ring_write(ring, PACKET0(UVD_GPCOM_VCPU_DATA0, 0));
+	radeon_ring_write(ring, 0);
+	radeon_ring_write(ring, PACKET0(UVD_GPCOM_VCPU_DATA1, 0));
+	radeon_ring_write(ring, 0);
+	radeon_ring_write(ring, PACKET0(UVD_GPCOM_VCPU_CMD, 0));
+	radeon_ring_write(ring, 2);
+	return;
+}
+
 void r600_semaphore_ring_emit(struct radeon_device *rdev,
 			      struct radeon_ring *ring,
 			      struct radeon_semaphore *semaphore,
@@ -2780,6 +3025,23 @@ void r600_dma_semaphore_ring_emit(struct radeon_device *rdev,
 	radeon_ring_write(ring, upper_32_bits(addr) & 0xff);
 }
 
+void r600_uvd_semaphore_emit(struct radeon_device *rdev,
+			     struct radeon_ring *ring,
+			     struct radeon_semaphore *semaphore,
+			     bool emit_wait)
+{
+	uint64_t addr = semaphore->gpu_addr;
+
+	radeon_ring_write(ring, PACKET0(UVD_SEMA_ADDR_LOW, 0));
+	radeon_ring_write(ring, (addr >> 3) & 0x000FFFFF);
+
+	radeon_ring_write(ring, PACKET0(UVD_SEMA_ADDR_HIGH, 0));
+	radeon_ring_write(ring, (addr >> 23) & 0x000FFFFF);
+
+	radeon_ring_write(ring, PACKET0(UVD_SEMA_CMD, 0));
+	radeon_ring_write(ring, emit_wait ? 1 : 0);
+}
+
 int r600_copy_blit(struct radeon_device *rdev,
 		   uint64_t src_offset,
 		   uint64_t dst_offset,
@@ -3183,6 +3445,16 @@ void r600_ring_ib_execute(struct radeon_device *rdev, struct radeon_ib *ib)
 	radeon_ring_write(ring, ib->length_dw);
 }
 
+void r600_uvd_ib_execute(struct radeon_device *rdev, struct radeon_ib *ib)
+{
+	struct radeon_ring *ring = &rdev->ring[ib->ring];
+
+	radeon_ring_write(ring, PACKET0(UVD_RBC_IB_BASE, 0));
+	radeon_ring_write(ring, ib->gpu_addr);
+	radeon_ring_write(ring, PACKET0(UVD_RBC_IB_SIZE, 0));
+	radeon_ring_write(ring, ib->length_dw);
+}
+
 int r600_ib_test(struct radeon_device *rdev, struct radeon_ring *ring)
 {
 	struct radeon_ib ib;
@@ -3300,6 +3572,41 @@ int r600_dma_ib_test(struct radeon_device *rdev, struct radeon_ring *ring)
 	return r;
 }
 
+int r600_uvd_ib_test(struct radeon_device *rdev, struct radeon_ring *ring)
+{
+	struct radeon_fence *fence = NULL;
+	int r;
+
+	r = radeon_set_uvd_clocks(rdev, 53300, 40000);
+	if (r) {
+		DRM_ERROR("radeon: failed to raise UVD clocks (%d).\n", r);
+		return r;
+	}
+
+	r = radeon_uvd_get_create_msg(rdev, ring->idx, 1, NULL);
+	if (r) {
+		DRM_ERROR("radeon: failed to get create msg (%d).\n", r);
+		goto error;
+	}
+
+	r = radeon_uvd_get_destroy_msg(rdev, ring->idx, 1, &fence);
+	if (r) {
+		DRM_ERROR("radeon: failed to get destroy ib (%d).\n", r);
+		goto error;
+	}
+
+	r = radeon_fence_wait(fence, false);
+	if (r) {
+		DRM_ERROR("radeon: fence wait failed (%d).\n", r);
+		goto error;
+	}
+	DRM_INFO("ib test on ring %d succeeded\n",  ring->idx);
+error:
+	radeon_fence_unref(&fence);
+	radeon_set_uvd_clocks(rdev, 0, 0);
+	return r;
+}
+
 /**
  * r600_dma_ring_ib_execute - Schedule an IB on the DMA engine
  *
@@ -4232,7 +4539,7 @@ void r600_ioctl_wait_idle(struct radeon_device *rdev, struct radeon_bo *bo)
 
 void r600_set_pcie_lanes(struct radeon_device *rdev, int lanes)
 {
-	u32 link_width_cntl, mask, target_reg;
+	u32 link_width_cntl, mask;
 
 	if (rdev->flags & RADEON_IS_IGP)
 		return;
@@ -4244,7 +4551,7 @@ void r600_set_pcie_lanes(struct radeon_device *rdev, int lanes)
 	if (ASIC_IS_X2(rdev))
 		return;
 
-	/* FIXME wait for idle */
+	radeon_gui_idle(rdev);
 
 	switch (lanes) {
 	case 0:
@@ -4263,53 +4570,24 @@ void r600_set_pcie_lanes(struct radeon_device *rdev, int lanes)
 		mask = RADEON_PCIE_LC_LINK_WIDTH_X8;
 		break;
 	case 12:
+		/* not actually supported */
 		mask = RADEON_PCIE_LC_LINK_WIDTH_X12;
 		break;
 	case 16:
-	default:
 		mask = RADEON_PCIE_LC_LINK_WIDTH_X16;
 		break;
-	}
-
-	link_width_cntl = RREG32_PCIE_P(RADEON_PCIE_LC_LINK_WIDTH_CNTL);
-
-	if ((link_width_cntl & RADEON_PCIE_LC_LINK_WIDTH_RD_MASK) ==
-	    (mask << RADEON_PCIE_LC_LINK_WIDTH_RD_SHIFT))
-		return;
-
-	if (link_width_cntl & R600_PCIE_LC_UPCONFIGURE_DIS)
+	default:
+		DRM_ERROR("invalid pcie lane request: %d\n", lanes);
 		return;
+	}
 
-	link_width_cntl &= ~(RADEON_PCIE_LC_LINK_WIDTH_MASK |
-			     RADEON_PCIE_LC_RECONFIG_NOW |
-			     R600_PCIE_LC_RENEGOTIATE_EN |
-			     R600_PCIE_LC_RECONFIG_ARC_MISSING_ESCAPE);
-	link_width_cntl |= mask;
-
-	WREG32_PCIE_P(RADEON_PCIE_LC_LINK_WIDTH_CNTL, link_width_cntl);
-
-        /* some northbridges can renegotiate the link rather than requiring                                  
-         * a complete re-config.                                                                             
-         * e.g., AMD 780/790 northbridges (pci ids: 0x5956, 0x5957, 0x5958, etc.)                            
-         */
-        if (link_width_cntl & R600_PCIE_LC_RENEGOTIATION_SUPPORT)
-		link_width_cntl |= R600_PCIE_LC_RENEGOTIATE_EN | R600_PCIE_LC_UPCONFIGURE_SUPPORT;
-        else
-		link_width_cntl |= R600_PCIE_LC_RECONFIG_ARC_MISSING_ESCAPE;
-
-	WREG32_PCIE_P(RADEON_PCIE_LC_LINK_WIDTH_CNTL, (link_width_cntl |
-						       RADEON_PCIE_LC_RECONFIG_NOW));
-
-        if (rdev->family >= CHIP_RV770)
-		target_reg = R700_TARGET_AND_CURRENT_PROFILE_INDEX;
-        else
-		target_reg = R600_TARGET_AND_CURRENT_PROFILE_INDEX;
-
-        /* wait for lane set to complete */
-        link_width_cntl = RREG32(target_reg);
-        while (link_width_cntl == 0xffffffff)
-		link_width_cntl = RREG32(target_reg);
+	link_width_cntl = RREG32_PCIE_PORT(RADEON_PCIE_LC_LINK_WIDTH_CNTL);
+	link_width_cntl &= ~RADEON_PCIE_LC_LINK_WIDTH_MASK;
+	link_width_cntl |= mask << RADEON_PCIE_LC_LINK_WIDTH_SHIFT;
+	link_width_cntl |= (RADEON_PCIE_LC_RECONFIG_NOW |
+			    R600_PCIE_LC_RECONFIG_ARC_MISSING_ESCAPE);
 
+	WREG32_PCIE_PORT(RADEON_PCIE_LC_LINK_WIDTH_CNTL, link_width_cntl);
 }
 
 int r600_get_pcie_lanes(struct radeon_device *rdev)
@@ -4326,13 +4604,11 @@ int r600_get_pcie_lanes(struct radeon_device *rdev)
 	if (ASIC_IS_X2(rdev))
 		return 0;
 
-	/* FIXME wait for idle */
+	radeon_gui_idle(rdev);
 
-	link_width_cntl = RREG32_PCIE_P(RADEON_PCIE_LC_LINK_WIDTH_CNTL);
+	link_width_cntl = RREG32_PCIE_PORT(RADEON_PCIE_LC_LINK_WIDTH_CNTL);
 
 	switch ((link_width_cntl & RADEON_PCIE_LC_LINK_WIDTH_RD_MASK) >> RADEON_PCIE_LC_LINK_WIDTH_RD_SHIFT) {
-	case RADEON_PCIE_LC_LINK_WIDTH_X0:
-		return 0;
 	case RADEON_PCIE_LC_LINK_WIDTH_X1:
 		return 1;
 	case RADEON_PCIE_LC_LINK_WIDTH_X2:
@@ -4341,6 +4617,10 @@ int r600_get_pcie_lanes(struct radeon_device *rdev)
 		return 4;
 	case RADEON_PCIE_LC_LINK_WIDTH_X8:
 		return 8;
+	case RADEON_PCIE_LC_LINK_WIDTH_X12:
+		/* not actually supported */
+		return 12;
+	case RADEON_PCIE_LC_LINK_WIDTH_X0:
 	case RADEON_PCIE_LC_LINK_WIDTH_X16:
 	default:
 		return 16;
@@ -4378,7 +4658,7 @@ static void r600_pcie_gen2_enable(struct radeon_device *rdev)
 	if (!(mask & DRM_PCIE_SPEED_50))
 		return;
 
-	speed_cntl = RREG32_PCIE_P(PCIE_LC_SPEED_CNTL);
+	speed_cntl = RREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL);
 	if (speed_cntl & LC_CURRENT_DATA_RATE) {
 		DRM_INFO("PCIE gen 2 link speeds already enabled\n");
 		return;
@@ -4391,23 +4671,23 @@ static void r600_pcie_gen2_enable(struct radeon_device *rdev)
 	    (rdev->family == CHIP_RV620) ||
 	    (rdev->family == CHIP_RV635)) {
 		/* advertise upconfig capability */
-		link_width_cntl = RREG32_PCIE_P(PCIE_LC_LINK_WIDTH_CNTL);
+		link_width_cntl = RREG32_PCIE_PORT(PCIE_LC_LINK_WIDTH_CNTL);
 		link_width_cntl &= ~LC_UPCONFIGURE_DIS;
-		WREG32_PCIE_P(PCIE_LC_LINK_WIDTH_CNTL, link_width_cntl);
-		link_width_cntl = RREG32_PCIE_P(PCIE_LC_LINK_WIDTH_CNTL);
+		WREG32_PCIE_PORT(PCIE_LC_LINK_WIDTH_CNTL, link_width_cntl);
+		link_width_cntl = RREG32_PCIE_PORT(PCIE_LC_LINK_WIDTH_CNTL);
 		if (link_width_cntl & LC_RENEGOTIATION_SUPPORT) {
 			lanes = (link_width_cntl & LC_LINK_WIDTH_RD_MASK) >> LC_LINK_WIDTH_RD_SHIFT;
 			link_width_cntl &= ~(LC_LINK_WIDTH_MASK |
 					     LC_RECONFIG_ARC_MISSING_ESCAPE);
 			link_width_cntl |= lanes | LC_RECONFIG_NOW | LC_RENEGOTIATE_EN;
-			WREG32_PCIE_P(PCIE_LC_LINK_WIDTH_CNTL, link_width_cntl);
+			WREG32_PCIE_PORT(PCIE_LC_LINK_WIDTH_CNTL, link_width_cntl);
 		} else {
 			link_width_cntl |= LC_UPCONFIGURE_DIS;
-			WREG32_PCIE_P(PCIE_LC_LINK_WIDTH_CNTL, link_width_cntl);
+			WREG32_PCIE_PORT(PCIE_LC_LINK_WIDTH_CNTL, link_width_cntl);
 		}
 	}
 
-	speed_cntl = RREG32_PCIE_P(PCIE_LC_SPEED_CNTL);
+	speed_cntl = RREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL);
 	if ((speed_cntl & LC_OTHER_SIDE_EVER_SENT_GEN2) &&
 	    (speed_cntl & LC_OTHER_SIDE_SUPPORTS_GEN2)) {
 
@@ -4428,7 +4708,7 @@ static void r600_pcie_gen2_enable(struct radeon_device *rdev)
 		speed_cntl &= ~LC_VOLTAGE_TIMER_SEL_MASK;
 		speed_cntl &= ~LC_FORCE_DIS_HW_SPEED_CHANGE;
 		speed_cntl |= LC_FORCE_EN_HW_SPEED_CHANGE;
-		WREG32_PCIE_P(PCIE_LC_SPEED_CNTL, speed_cntl);
+		WREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL, speed_cntl);
 
 		tmp = RREG32(0x541c);
 		WREG32(0x541c, tmp | 0x8);
@@ -4442,27 +4722,27 @@ static void r600_pcie_gen2_enable(struct radeon_device *rdev)
 		if ((rdev->family == CHIP_RV670) ||
 		    (rdev->family == CHIP_RV620) ||
 		    (rdev->family == CHIP_RV635)) {
-			training_cntl = RREG32_PCIE_P(PCIE_LC_TRAINING_CNTL);
+			training_cntl = RREG32_PCIE_PORT(PCIE_LC_TRAINING_CNTL);
 			training_cntl &= ~LC_POINT_7_PLUS_EN;
-			WREG32_PCIE_P(PCIE_LC_TRAINING_CNTL, training_cntl);
+			WREG32_PCIE_PORT(PCIE_LC_TRAINING_CNTL, training_cntl);
 		} else {
-			speed_cntl = RREG32_PCIE_P(PCIE_LC_SPEED_CNTL);
+			speed_cntl = RREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL);
 			speed_cntl &= ~LC_TARGET_LINK_SPEED_OVERRIDE_EN;
-			WREG32_PCIE_P(PCIE_LC_SPEED_CNTL, speed_cntl);
+			WREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL, speed_cntl);
 		}
 
-		speed_cntl = RREG32_PCIE_P(PCIE_LC_SPEED_CNTL);
+		speed_cntl = RREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL);
 		speed_cntl |= LC_GEN2_EN_STRAP;
-		WREG32_PCIE_P(PCIE_LC_SPEED_CNTL, speed_cntl);
+		WREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL, speed_cntl);
 
 	} else {
-		link_width_cntl = RREG32_PCIE_P(PCIE_LC_LINK_WIDTH_CNTL);
+		link_width_cntl = RREG32_PCIE_PORT(PCIE_LC_LINK_WIDTH_CNTL);
 		/* XXX: only disable it if gen1 bridge vendor == 0x111d or 0x1106 */
 		if (1)
 			link_width_cntl |= LC_UPCONFIGURE_DIS;
 		else
 			link_width_cntl &= ~LC_UPCONFIGURE_DIS;
-		WREG32_PCIE_P(PCIE_LC_LINK_WIDTH_CNTL, link_width_cntl);
+		WREG32_PCIE_PORT(PCIE_LC_LINK_WIDTH_CNTL, link_width_cntl);
 	}
 }
 
diff --git a/drivers/gpu/drm/radeon/r600_audio.c b/drivers/gpu/drm/radeon/r600_audio.c
index cb03fe22b0ab..c92eb86a8e55 100644
--- a/drivers/gpu/drm/radeon/r600_audio.c
+++ b/drivers/gpu/drm/radeon/r600_audio.c
@@ -57,10 +57,7 @@ static bool radeon_dig_encoder(struct drm_encoder *encoder)
  */
 static int r600_audio_chipset_supported(struct radeon_device *rdev)
 {
-	return (rdev->family >= CHIP_R600 && !ASIC_IS_DCE6(rdev))
-		|| rdev->family == CHIP_RS600
-		|| rdev->family == CHIP_RS690
-		|| rdev->family == CHIP_RS740;
+	return ASIC_IS_DCE2(rdev) && !ASIC_IS_DCE6(rdev);
 }
 
 struct r600_audio r600_audio_status(struct radeon_device *rdev)
@@ -184,65 +181,6 @@ int r600_audio_init(struct radeon_device *rdev)
 }
 
 /*
- * atach the audio codec to the clock source of the encoder
- */
-void r600_audio_set_clock(struct drm_encoder *encoder, int clock)
-{
-	struct drm_device *dev = encoder->dev;
-	struct radeon_device *rdev = dev->dev_private;
-	struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder);
-	struct radeon_encoder_atom_dig *dig = radeon_encoder->enc_priv;
-	struct radeon_crtc *radeon_crtc = to_radeon_crtc(encoder->crtc);
-	int base_rate = 48000;
-
-	switch (radeon_encoder->encoder_id) {
-	case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_TMDS1:
-	case ENCODER_OBJECT_ID_INTERNAL_LVTM1:
-		WREG32_P(R600_AUDIO_TIMING, 0, ~0x301);
-		break;
-	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY:
-	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY1:
-	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY2:
-	case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_LVTMA:
-		WREG32_P(R600_AUDIO_TIMING, 0x100, ~0x301);
-		break;
-	default:
-		dev_err(rdev->dev, "Unsupported encoder type 0x%02X\n",
-			  radeon_encoder->encoder_id);
-		return;
-	}
-
-	if (ASIC_IS_DCE4(rdev)) {
-		/* TODO: other PLLs? */
-		WREG32(EVERGREEN_AUDIO_PLL1_MUL, base_rate * 10);
-		WREG32(EVERGREEN_AUDIO_PLL1_DIV, clock * 10);
-		WREG32(EVERGREEN_AUDIO_PLL1_UNK, 0x00000071);
-
-		/* Select DTO source */
-		WREG32(0x5ac, radeon_crtc->crtc_id);
-	} else {
-		switch (dig->dig_encoder) {
-		case 0:
-			WREG32(R600_AUDIO_PLL1_MUL, base_rate * 50);
-			WREG32(R600_AUDIO_PLL1_DIV, clock * 100);
-			WREG32(R600_AUDIO_CLK_SRCSEL, 0);
-			break;
-
-		case 1:
-			WREG32(R600_AUDIO_PLL2_MUL, base_rate * 50);
-			WREG32(R600_AUDIO_PLL2_DIV, clock * 100);
-			WREG32(R600_AUDIO_CLK_SRCSEL, 1);
-			break;
-		default:
-			dev_err(rdev->dev,
-				"Unsupported DIG on encoder 0x%02X\n",
-				radeon_encoder->encoder_id);
-			return;
-		}
-	}
-}
-
-/*
  * release the audio timer
  * TODO: How to do this correctly on SMP systems?
  */
diff --git a/drivers/gpu/drm/radeon/r600_hdmi.c b/drivers/gpu/drm/radeon/r600_hdmi.c
index 21ecc0e12dc4..47f180a79352 100644
--- a/drivers/gpu/drm/radeon/r600_hdmi.c
+++ b/drivers/gpu/drm/radeon/r600_hdmi.c
@@ -226,6 +226,39 @@ static void r600_hdmi_audio_workaround(struct drm_encoder *encoder)
 		 value, ~HDMI0_AUDIO_TEST_EN);
 }
 
+void r600_audio_set_dto(struct drm_encoder *encoder, u32 clock)
+{
+	struct drm_device *dev = encoder->dev;
+	struct radeon_device *rdev = dev->dev_private;
+	struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder);
+	struct radeon_encoder_atom_dig *dig = radeon_encoder->enc_priv;
+	u32 base_rate = 48000;
+
+	if (!dig || !dig->afmt)
+		return;
+
+	/* there are two DTOs selected by DCCG_AUDIO_DTO_SELECT.
+	 * doesn't matter which one you use.  Just use the first one.
+	 */
+	/* XXX: properly calculate this */
+	/* XXX two dtos; generally use dto0 for hdmi */
+	/* Express [24MHz / target pixel clock] as an exact rational
+	 * number (coefficient of two integer numbers.  DCCG_AUDIO_DTOx_PHASE
+	 * is the numerator, DCCG_AUDIO_DTOx_MODULE is the denominator
+	 */
+	if (ASIC_IS_DCE3(rdev)) {
+		/* according to the reg specs, this should DCE3.2 only, but in
+		 * practice it seems to cover DCE3.0 as well.
+		 */
+		WREG32(DCCG_AUDIO_DTO0_PHASE, base_rate * 50);
+		WREG32(DCCG_AUDIO_DTO0_MODULE, clock * 100);
+		WREG32(DCCG_AUDIO_DTO_SELECT, 0); /* select DTO0 */
+	} else {
+		/* according to the reg specs, this should be DCE2.0 and DCE3.0 */
+		WREG32(AUDIO_DTO, AUDIO_DTO_PHASE(base_rate * 50) |
+		       AUDIO_DTO_MODULE(clock * 100));
+	}
+}
 
 /*
  * update the info frames with the data from the current display mode
@@ -246,7 +279,7 @@ void r600_hdmi_setmode(struct drm_encoder *encoder, struct drm_display_mode *mod
 		return;
 	offset = dig->afmt->offset;
 
-	r600_audio_set_clock(encoder, mode->clock);
+	r600_audio_set_dto(encoder, mode->clock);
 
 	WREG32(HDMI0_VBI_PACKET_CONTROL + offset,
 	       HDMI0_NULL_SEND); /* send null packets when required */
@@ -415,114 +448,73 @@ void r600_hdmi_update_audio_settings(struct drm_encoder *encoder)
 /*
  * enable the HDMI engine
  */
-void r600_hdmi_enable(struct drm_encoder *encoder)
+void r600_hdmi_enable(struct drm_encoder *encoder, bool enable)
 {
 	struct drm_device *dev = encoder->dev;
 	struct radeon_device *rdev = dev->dev_private;
 	struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder);
 	struct radeon_encoder_atom_dig *dig = radeon_encoder->enc_priv;
-	uint32_t offset;
-	u32 hdmi;
-
-	if (ASIC_IS_DCE6(rdev))
-		return;
+	u32 hdmi = HDMI0_ERROR_ACK;
 
 	/* Silent, r600_hdmi_enable will raise WARN for us */
-	if (dig->afmt->enabled)
+	if (enable && dig->afmt->enabled)
+		return;
+	if (!enable && !dig->afmt->enabled)
 		return;
-	offset = dig->afmt->offset;
 
 	/* Older chipsets require setting HDMI and routing manually */
-	if (rdev->family >= CHIP_R600 && !ASIC_IS_DCE3(rdev)) {
-		hdmi = HDMI0_ERROR_ACK | HDMI0_ENABLE;
+	if (!ASIC_IS_DCE3(rdev)) {
+		if (enable)
+			hdmi |= HDMI0_ENABLE;
 		switch (radeon_encoder->encoder_id) {
 		case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_TMDS1:
-			WREG32_P(AVIVO_TMDSA_CNTL, AVIVO_TMDSA_CNTL_HDMI_EN,
-				 ~AVIVO_TMDSA_CNTL_HDMI_EN);
-			hdmi |= HDMI0_STREAM(HDMI0_STREAM_TMDSA);
+			if (enable) {
+				WREG32_OR(AVIVO_TMDSA_CNTL, AVIVO_TMDSA_CNTL_HDMI_EN);
+				hdmi |= HDMI0_STREAM(HDMI0_STREAM_TMDSA);
+			} else {
+				WREG32_AND(AVIVO_TMDSA_CNTL, ~AVIVO_TMDSA_CNTL_HDMI_EN);
+			}
 			break;
 		case ENCODER_OBJECT_ID_INTERNAL_LVTM1:
-			WREG32_P(AVIVO_LVTMA_CNTL, AVIVO_LVTMA_CNTL_HDMI_EN,
-				 ~AVIVO_LVTMA_CNTL_HDMI_EN);
-			hdmi |= HDMI0_STREAM(HDMI0_STREAM_LVTMA);
+			if (enable) {
+				WREG32_OR(AVIVO_LVTMA_CNTL, AVIVO_LVTMA_CNTL_HDMI_EN);
+				hdmi |= HDMI0_STREAM(HDMI0_STREAM_LVTMA);
+			} else {
+				WREG32_AND(AVIVO_LVTMA_CNTL, ~AVIVO_LVTMA_CNTL_HDMI_EN);
+			}
 			break;
 		case ENCODER_OBJECT_ID_INTERNAL_DDI:
-			WREG32_P(DDIA_CNTL, DDIA_HDMI_EN, ~DDIA_HDMI_EN);
-			hdmi |= HDMI0_STREAM(HDMI0_STREAM_DDIA);
+			if (enable) {
+				WREG32_OR(DDIA_CNTL, DDIA_HDMI_EN);
+				hdmi |= HDMI0_STREAM(HDMI0_STREAM_DDIA);
+			} else {
+				WREG32_AND(DDIA_CNTL, ~DDIA_HDMI_EN);
+			}
 			break;
 		case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_DVO1:
-			hdmi |= HDMI0_STREAM(HDMI0_STREAM_DVOA);
+			if (enable)
+				hdmi |= HDMI0_STREAM(HDMI0_STREAM_DVOA);
 			break;
 		default:
 			dev_err(rdev->dev, "Invalid encoder for HDMI: 0x%X\n",
 				radeon_encoder->encoder_id);
 			break;
 		}
-		WREG32(HDMI0_CONTROL + offset, hdmi);
+		WREG32(HDMI0_CONTROL + dig->afmt->offset, hdmi);
 	}
 
 	if (rdev->irq.installed) {
 		/* if irq is available use it */
-		radeon_irq_kms_enable_afmt(rdev, dig->afmt->id);
+		/* XXX: shouldn't need this on any asics.  Double check DCE2/3 */
+		if (enable)
+			radeon_irq_kms_enable_afmt(rdev, dig->afmt->id);
+		else
+			radeon_irq_kms_disable_afmt(rdev, dig->afmt->id);
 	}
 
-	dig->afmt->enabled = true;
+	dig->afmt->enabled = enable;
 
-	DRM_DEBUG("Enabling HDMI interface @ 0x%04X for encoder 0x%x\n",
-		  offset, radeon_encoder->encoder_id);
+	DRM_DEBUG("%sabling HDMI interface @ 0x%04X for encoder 0x%x\n",
+		  enable ? "En" : "Dis", dig->afmt->offset, radeon_encoder->encoder_id);
 }
 
-/*
- * disable the HDMI engine
- */
-void r600_hdmi_disable(struct drm_encoder *encoder)
-{
-	struct drm_device *dev = encoder->dev;
-	struct radeon_device *rdev = dev->dev_private;
-	struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder);
-	struct radeon_encoder_atom_dig *dig = radeon_encoder->enc_priv;
-	uint32_t offset;
-
-	if (ASIC_IS_DCE6(rdev))
-		return;
-
-	/* Called for ATOM_ENCODER_MODE_HDMI only */
-	if (!dig || !dig->afmt) {
-		return;
-	}
-	if (!dig->afmt->enabled)
-		return;
-	offset = dig->afmt->offset;
-
-	DRM_DEBUG("Disabling HDMI interface @ 0x%04X for encoder 0x%x\n",
-		  offset, radeon_encoder->encoder_id);
-
-	/* disable irq */
-	radeon_irq_kms_disable_afmt(rdev, dig->afmt->id);
-
-	/* Older chipsets not handled by AtomBIOS */
-	if (rdev->family >= CHIP_R600 && !ASIC_IS_DCE3(rdev)) {
-		switch (radeon_encoder->encoder_id) {
-		case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_TMDS1:
-			WREG32_P(AVIVO_TMDSA_CNTL, 0,
-				 ~AVIVO_TMDSA_CNTL_HDMI_EN);
-			break;
-		case ENCODER_OBJECT_ID_INTERNAL_LVTM1:
-			WREG32_P(AVIVO_LVTMA_CNTL, 0,
-				 ~AVIVO_LVTMA_CNTL_HDMI_EN);
-			break;
-		case ENCODER_OBJECT_ID_INTERNAL_DDI:
-			WREG32_P(DDIA_CNTL, 0, ~DDIA_HDMI_EN);
-			break;
-		case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_DVO1:
-			break;
-		default:
-			dev_err(rdev->dev, "Invalid encoder for HDMI: 0x%X\n",
-				radeon_encoder->encoder_id);
-			break;
-		}
-		WREG32(HDMI0_CONTROL + offset, HDMI0_ERROR_ACK);
-	}
-
-	dig->afmt->enabled = false;
-}
diff --git a/drivers/gpu/drm/radeon/r600d.h b/drivers/gpu/drm/radeon/r600d.h
index a42ba11a3bed..acb146c06973 100644
--- a/drivers/gpu/drm/radeon/r600d.h
+++ b/drivers/gpu/drm/radeon/r600d.h
@@ -691,6 +691,7 @@
 #define SRBM_SOFT_RESET                                   0xe60
 #       define SOFT_RESET_DMA                             (1 << 12)
 #       define SOFT_RESET_RLC                             (1 << 13)
+#       define SOFT_RESET_UVD                             (1 << 18)
 #       define RV770_SOFT_RESET_DMA                       (1 << 20)
 
 #define CP_INT_CNTL                                       0xc124
@@ -909,7 +910,12 @@
 #       define TARGET_LINK_SPEED_MASK                     (0xf << 0)
 #       define SELECTABLE_DEEMPHASIS                      (1 << 6)
 
-/* Audio clocks */
+/* Audio clocks DCE 2.0/3.0 */
+#define AUDIO_DTO                         0x7340
+#       define AUDIO_DTO_PHASE(x)         (((x) & 0xffff) << 0)
+#       define AUDIO_DTO_MODULE(x)        (((x) & 0xffff) << 16)
+
+/* Audio clocks DCE 3.2 */
 #define DCCG_AUDIO_DTO0_PHASE             0x0514
 #define DCCG_AUDIO_DTO0_MODULE            0x0518
 #define DCCG_AUDIO_DTO0_LOAD              0x051c
@@ -1143,6 +1149,70 @@
 #       define AFMT_AZ_AUDIO_ENABLE_CHG_ACK  (1 << 30)
 
 /*
+ * UVD
+ */
+#define UVD_SEMA_ADDR_LOW				0xef00
+#define UVD_SEMA_ADDR_HIGH				0xef04
+#define UVD_SEMA_CMD					0xef08
+
+#define UVD_GPCOM_VCPU_CMD				0xef0c
+#define UVD_GPCOM_VCPU_DATA0				0xef10
+#define UVD_GPCOM_VCPU_DATA1				0xef14
+#define UVD_ENGINE_CNTL					0xef18
+
+#define UVD_SEMA_CNTL					0xf400
+#define UVD_RB_ARB_CTRL					0xf480
+
+#define UVD_LMI_EXT40_ADDR				0xf498
+#define UVD_CGC_GATE					0xf4a8
+#define UVD_LMI_CTRL2					0xf4f4
+#define UVD_MASTINT_EN					0xf500
+#define UVD_LMI_ADDR_EXT				0xf594
+#define UVD_LMI_CTRL					0xf598
+#define UVD_LMI_SWAP_CNTL				0xf5b4
+#define UVD_MP_SWAP_CNTL				0xf5bC
+#define UVD_MPC_CNTL					0xf5dC
+#define UVD_MPC_SET_MUXA0				0xf5e4
+#define UVD_MPC_SET_MUXA1				0xf5e8
+#define UVD_MPC_SET_MUXB0				0xf5eC
+#define UVD_MPC_SET_MUXB1				0xf5f0
+#define UVD_MPC_SET_MUX					0xf5f4
+#define UVD_MPC_SET_ALU					0xf5f8
+
+#define UVD_VCPU_CNTL					0xf660
+#define UVD_SOFT_RESET					0xf680
+#define		RBC_SOFT_RESET					(1<<0)
+#define		LBSI_SOFT_RESET					(1<<1)
+#define		LMI_SOFT_RESET					(1<<2)
+#define		VCPU_SOFT_RESET					(1<<3)
+#define		CSM_SOFT_RESET					(1<<5)
+#define		CXW_SOFT_RESET					(1<<6)
+#define		TAP_SOFT_RESET					(1<<7)
+#define		LMI_UMC_SOFT_RESET				(1<<13)
+#define UVD_RBC_IB_BASE					0xf684
+#define UVD_RBC_IB_SIZE					0xf688
+#define UVD_RBC_RB_BASE					0xf68c
+#define UVD_RBC_RB_RPTR					0xf690
+#define UVD_RBC_RB_WPTR					0xf694
+#define UVD_RBC_RB_WPTR_CNTL				0xf698
+
+#define UVD_STATUS					0xf6bc
+
+#define UVD_SEMA_TIMEOUT_STATUS				0xf6c0
+#define UVD_SEMA_WAIT_INCOMPLETE_TIMEOUT_CNTL		0xf6c4
+#define UVD_SEMA_WAIT_FAULT_TIMEOUT_CNTL		0xf6c8
+#define UVD_SEMA_SIGNAL_INCOMPLETE_TIMEOUT_CNTL		0xf6cc
+
+#define UVD_RBC_RB_CNTL					0xf6a4
+#define UVD_RBC_RB_RPTR_ADDR				0xf6a8
+
+#define UVD_CONTEXT_ID					0xf6f4
+
+#	define UPLL_CTLREQ_MASK				0x00000008
+#	define UPLL_CTLACK_MASK				0x40000000
+#	define UPLL_CTLACK2_MASK			0x80000000
+
+/*
  * PM4
  */
 #define PACKET0(reg, n)	((RADEON_PACKET_TYPE0 << 30) |			\
diff --git a/drivers/gpu/drm/radeon/radeon.h b/drivers/gpu/drm/radeon/radeon.h
index 8263af3fd832..1442ce765d48 100644
--- a/drivers/gpu/drm/radeon/radeon.h
+++ b/drivers/gpu/drm/radeon/radeon.h
@@ -95,6 +95,7 @@ extern int radeon_hw_i2c;
 extern int radeon_pcie_gen2;
 extern int radeon_msi;
 extern int radeon_lockup_timeout;
+extern int radeon_fastfb;
 
 /*
  * Copy from radeon_drv.h so we don't have to include both and have conflicting
@@ -109,24 +110,27 @@ extern int radeon_lockup_timeout;
 #define RADEON_BIOS_NUM_SCRATCH			8
 
 /* max number of rings */
-#define RADEON_NUM_RINGS			5
+#define RADEON_NUM_RINGS			6
 
 /* fence seq are set to this number when signaled */
 #define RADEON_FENCE_SIGNALED_SEQ		0LL
 
 /* internal ring indices */
 /* r1xx+ has gfx CP ring */
-#define RADEON_RING_TYPE_GFX_INDEX		0
+#define RADEON_RING_TYPE_GFX_INDEX	0
 
 /* cayman has 2 compute CP rings */
-#define CAYMAN_RING_TYPE_CP1_INDEX		1
-#define CAYMAN_RING_TYPE_CP2_INDEX		2
+#define CAYMAN_RING_TYPE_CP1_INDEX	1
+#define CAYMAN_RING_TYPE_CP2_INDEX	2
 
 /* R600+ has an async dma ring */
 #define R600_RING_TYPE_DMA_INDEX		3
 /* cayman add a second async dma ring */
 #define CAYMAN_RING_TYPE_DMA1_INDEX		4
 
+/* R600+ */
+#define R600_RING_TYPE_UVD_INDEX	5
+
 /* hardcode those limit for now */
 #define RADEON_VA_IB_OFFSET			(1 << 20)
 #define RADEON_VA_RESERVED_SIZE			(8 << 20)
@@ -202,6 +206,11 @@ void radeon_pm_suspend(struct radeon_device *rdev);
 void radeon_pm_resume(struct radeon_device *rdev);
 void radeon_combios_get_power_modes(struct radeon_device *rdev);
 void radeon_atombios_get_power_modes(struct radeon_device *rdev);
+int radeon_atom_get_clock_dividers(struct radeon_device *rdev,
+				   u8 clock_type,
+				   u32 clock,
+				   bool strobe_mode,
+				   struct atom_clock_dividers *dividers);
 void radeon_atom_set_voltage(struct radeon_device *rdev, u16 voltage_level, u8 voltage_type);
 void rs690_pm_info(struct radeon_device *rdev);
 extern int rv6xx_get_temp(struct radeon_device *rdev);
@@ -349,7 +358,8 @@ struct radeon_bo {
 	struct radeon_device		*rdev;
 	struct drm_gem_object		gem_base;
 
-	struct ttm_bo_kmap_obj dma_buf_vmap;
+	struct ttm_bo_kmap_obj		dma_buf_vmap;
+	pid_t				pid;
 };
 #define gem_to_radeon_bo(gobj) container_of((gobj), struct radeon_bo, gem_base)
 
@@ -357,11 +367,14 @@ struct radeon_bo_list {
 	struct ttm_validate_buffer tv;
 	struct radeon_bo	*bo;
 	uint64_t		gpu_offset;
-	unsigned		rdomain;
-	unsigned		wdomain;
+	bool			written;
+	unsigned		domain;
+	unsigned		alt_domain;
 	u32			tiling_flags;
 };
 
+int radeon_gem_debugfs_init(struct radeon_device *rdev);
+
 /* sub-allocation manager, it has to be protected by another lock.
  * By conception this is an helper for other part of the driver
  * like the indirect buffer or semaphore, which both have their
@@ -517,6 +530,7 @@ struct radeon_mc {
 	bool			vram_is_ddr;
 	bool			igp_sideport_enabled;
 	u64                     gtt_base_align;
+	u64                     mc_mask;
 };
 
 bool radeon_combios_sideport_present(struct radeon_device *rdev);
@@ -918,6 +932,7 @@ struct radeon_wb {
 #define R600_WB_DMA_RPTR_OFFSET   1792
 #define R600_WB_IH_WPTR_OFFSET   2048
 #define CAYMAN_WB_DMA1_RPTR_OFFSET   2304
+#define R600_WB_UVD_RPTR_OFFSET  2560
 #define R600_WB_EVENT_OFFSET     3072
 
 /**
@@ -1118,6 +1133,46 @@ struct radeon_pm {
 int radeon_pm_get_type_index(struct radeon_device *rdev,
 			     enum radeon_pm_state_type ps_type,
 			     int instance);
+/*
+ * UVD
+ */
+#define RADEON_MAX_UVD_HANDLES	10
+#define RADEON_UVD_STACK_SIZE	(1024*1024)
+#define RADEON_UVD_HEAP_SIZE	(1024*1024)
+
+struct radeon_uvd {
+	struct radeon_bo	*vcpu_bo;
+	void			*cpu_addr;
+	uint64_t		gpu_addr;
+	atomic_t		handles[RADEON_MAX_UVD_HANDLES];
+	struct drm_file		*filp[RADEON_MAX_UVD_HANDLES];
+	struct delayed_work	idle_work;
+};
+
+int radeon_uvd_init(struct radeon_device *rdev);
+void radeon_uvd_fini(struct radeon_device *rdev);
+int radeon_uvd_suspend(struct radeon_device *rdev);
+int radeon_uvd_resume(struct radeon_device *rdev);
+int radeon_uvd_get_create_msg(struct radeon_device *rdev, int ring,
+			      uint32_t handle, struct radeon_fence **fence);
+int radeon_uvd_get_destroy_msg(struct radeon_device *rdev, int ring,
+			       uint32_t handle, struct radeon_fence **fence);
+void radeon_uvd_force_into_uvd_segment(struct radeon_bo *rbo);
+void radeon_uvd_free_handles(struct radeon_device *rdev,
+			     struct drm_file *filp);
+int radeon_uvd_cs_parse(struct radeon_cs_parser *parser);
+void radeon_uvd_note_usage(struct radeon_device *rdev);
+int radeon_uvd_calc_upll_dividers(struct radeon_device *rdev,
+				  unsigned vclk, unsigned dclk,
+				  unsigned vco_min, unsigned vco_max,
+				  unsigned fb_factor, unsigned fb_mask,
+				  unsigned pd_min, unsigned pd_max,
+				  unsigned pd_even,
+				  unsigned *optimal_fb_div,
+				  unsigned *optimal_vclk_div,
+				  unsigned *optimal_dclk_div);
+int radeon_uvd_send_upll_ctlreq(struct radeon_device *rdev,
+                                unsigned cg_upll_func_cntl);
 
 struct r600_audio {
 	int			channels;
@@ -1229,6 +1284,9 @@ struct radeon_asic {
 		void (*set_backlight_level)(struct radeon_encoder *radeon_encoder, u8 level);
 		/* get backlight level */
 		u8 (*get_backlight_level)(struct radeon_encoder *radeon_encoder);
+		/* audio callbacks */
+		void (*hdmi_enable)(struct drm_encoder *encoder, bool enable);
+		void (*hdmi_setmode)(struct drm_encoder *encoder, struct drm_display_mode *mode);
 	} display;
 	/* copy functions for bo handling */
 	struct {
@@ -1281,6 +1339,7 @@ struct radeon_asic {
 		int (*get_pcie_lanes)(struct radeon_device *rdev);
 		void (*set_pcie_lanes)(struct radeon_device *rdev, int lanes);
 		void (*set_clock_gating)(struct radeon_device *rdev, int enable);
+		int (*set_uvd_clocks)(struct radeon_device *rdev, u32 vclk, u32 dclk);
 	} pm;
 	/* pageflipping */
 	struct {
@@ -1443,6 +1502,7 @@ struct si_asic {
 	unsigned multi_gpu_tile_size;
 
 	unsigned tile_config;
+	uint32_t tile_mode_array[32];
 };
 
 union radeon_asic_config {
@@ -1608,6 +1668,7 @@ struct radeon_device {
 	struct radeon_asic		*asic;
 	struct radeon_gem		gem;
 	struct radeon_pm		pm;
+	struct radeon_uvd		uvd;
 	uint32_t			bios_scratch[RADEON_BIOS_NUM_SCRATCH];
 	struct radeon_wb		wb;
 	struct radeon_dummy_page	dummy_page;
@@ -1615,12 +1676,14 @@ struct radeon_device {
 	bool				suspend;
 	bool				need_dma32;
 	bool				accel_working;
+	bool				fastfb_working; /* IGP feature*/
 	struct radeon_surface_reg surface_regs[RADEON_GEM_MAX_SURFACES];
 	const struct firmware *me_fw;	/* all family ME firmware */
 	const struct firmware *pfp_fw;	/* r6/700 PFP firmware */
 	const struct firmware *rlc_fw;	/* r6/700 RLC firmware */
 	const struct firmware *mc_fw;	/* NI MC firmware */
 	const struct firmware *ce_fw;	/* SI CE firmware */
+	const struct firmware *uvd_fw;	/* UVD firmware */
 	struct r600_blit r600_blit;
 	struct r600_vram_scratch vram_scratch;
 	int msi_enabled; /* msi enabled */
@@ -1688,8 +1751,8 @@ void r100_io_wreg(struct radeon_device *rdev, u32 reg, u32 v);
 #define WREG32_MC(reg, v) rdev->mc_wreg(rdev, (reg), (v))
 #define RREG32_PCIE(reg) rv370_pcie_rreg(rdev, (reg))
 #define WREG32_PCIE(reg, v) rv370_pcie_wreg(rdev, (reg), (v))
-#define RREG32_PCIE_P(reg) rdev->pciep_rreg(rdev, (reg))
-#define WREG32_PCIE_P(reg, v) rdev->pciep_wreg(rdev, (reg), (v))
+#define RREG32_PCIE_PORT(reg) rdev->pciep_rreg(rdev, (reg))
+#define WREG32_PCIE_PORT(reg, v) rdev->pciep_wreg(rdev, (reg), (v))
 #define WREG32_P(reg, val, mask)				\
 	do {							\
 		uint32_t tmp_ = RREG32(reg);			\
@@ -1697,6 +1760,8 @@ void r100_io_wreg(struct radeon_device *rdev, u32 reg, u32 v);
 		tmp_ |= ((val) & ~(mask));			\
 		WREG32(reg, tmp_);				\
 	} while (0)
+#define WREG32_AND(reg, and) WREG32_P(reg, 0, and)
+#define WREG32_OR(reg, or) WREG32_P(reg, or, ~or)
 #define WREG32_PLL_P(reg, val, mask)				\
 	do {							\
 		uint32_t tmp_ = RREG32_PLL(reg);		\
@@ -1830,6 +1895,8 @@ void radeon_ring_write(struct radeon_ring *ring, uint32_t v);
 #define radeon_get_vblank_counter(rdev, crtc) (rdev)->asic->display.get_vblank_counter((rdev), (crtc))
 #define radeon_set_backlight_level(rdev, e, l) (rdev)->asic->display.set_backlight_level((e), (l))
 #define radeon_get_backlight_level(rdev, e) (rdev)->asic->display.get_backlight_level((e))
+#define radeon_hdmi_enable(rdev, e, b) (rdev)->asic->display.hdmi_enable((e), (b))
+#define radeon_hdmi_setmode(rdev, e, m) (rdev)->asic->display.hdmi_setmode((e), (m))
 #define radeon_fence_ring_emit(rdev, r, fence) (rdev)->asic->ring[(r)].emit_fence((rdev), (fence))
 #define radeon_semaphore_ring_emit(rdev, r, cp, semaphore, emit_wait) (rdev)->asic->ring[(r)].emit_semaphore((rdev), (cp), (semaphore), (emit_wait))
 #define radeon_copy_blit(rdev, s, d, np, f) (rdev)->asic->copy.blit((rdev), (s), (d), (np), (f))
@@ -1845,6 +1912,7 @@ void radeon_ring_write(struct radeon_ring *ring, uint32_t v);
 #define radeon_get_pcie_lanes(rdev) (rdev)->asic->pm.get_pcie_lanes((rdev))
 #define radeon_set_pcie_lanes(rdev, l) (rdev)->asic->pm.set_pcie_lanes((rdev), (l))
 #define radeon_set_clock_gating(rdev, e) (rdev)->asic->pm.set_clock_gating((rdev), (e))
+#define radeon_set_uvd_clocks(rdev, v, d) (rdev)->asic->pm.set_uvd_clocks((rdev), (v), (d))
 #define radeon_set_surface_reg(rdev, r, f, p, o, s) ((rdev)->asic->surface.set_reg((rdev), (r), (f), (p), (o), (s)))
 #define radeon_clear_surface_reg(rdev, r) ((rdev)->asic->surface.clear_reg((rdev), (r)))
 #define radeon_bandwidth_update(rdev) (rdev)->asic->display.bandwidth_update((rdev))
@@ -1892,6 +1960,9 @@ extern void radeon_gtt_location(struct radeon_device *rdev, struct radeon_mc *mc
 extern int radeon_resume_kms(struct drm_device *dev);
 extern int radeon_suspend_kms(struct drm_device *dev, pm_message_t state);
 extern void radeon_ttm_set_active_vram_size(struct radeon_device *rdev, u64 size);
+extern void radeon_program_register_sequence(struct radeon_device *rdev,
+					     const u32 *registers,
+					     const u32 array_size);
 
 /*
  * vm
@@ -1964,9 +2035,6 @@ struct radeon_hdmi_acr {
 
 extern struct radeon_hdmi_acr r600_hdmi_acr(uint32_t clock);
 
-extern void r600_hdmi_enable(struct drm_encoder *encoder);
-extern void r600_hdmi_disable(struct drm_encoder *encoder);
-extern void r600_hdmi_setmode(struct drm_encoder *encoder, struct drm_display_mode *mode);
 extern u32 r6xx_remap_render_backend(struct radeon_device *rdev,
 				     u32 tiling_pipe_num,
 				     u32 max_rb_num,
@@ -1977,8 +2045,6 @@ extern u32 r6xx_remap_render_backend(struct radeon_device *rdev,
  * evergreen functions used by radeon_encoder.c
  */
 
-extern void evergreen_hdmi_setmode(struct drm_encoder *encoder, struct drm_display_mode *mode);
-
 extern int ni_init_microcode(struct radeon_device *rdev);
 extern int ni_mc_load_microcode(struct radeon_device *rdev);
 
diff --git a/drivers/gpu/drm/radeon/radeon_asic.c b/drivers/gpu/drm/radeon/radeon_asic.c
index aba0a893ea98..6417132c50cf 100644
--- a/drivers/gpu/drm/radeon/radeon_asic.c
+++ b/drivers/gpu/drm/radeon/radeon_asic.c
@@ -656,6 +656,8 @@ static struct radeon_asic rs600_asic = {
 		.wait_for_vblank = &avivo_wait_for_vblank,
 		.set_backlight_level = &atombios_set_backlight_level,
 		.get_backlight_level = &atombios_get_backlight_level,
+		.hdmi_enable = &r600_hdmi_enable,
+		.hdmi_setmode = &r600_hdmi_setmode,
 	},
 	.copy = {
 		.blit = &r100_copy_blit,
@@ -732,6 +734,8 @@ static struct radeon_asic rs690_asic = {
 		.wait_for_vblank = &avivo_wait_for_vblank,
 		.set_backlight_level = &atombios_set_backlight_level,
 		.get_backlight_level = &atombios_get_backlight_level,
+		.hdmi_enable = &r600_hdmi_enable,
+		.hdmi_setmode = &r600_hdmi_setmode,
 	},
 	.copy = {
 		.blit = &r100_copy_blit,
@@ -970,6 +974,8 @@ static struct radeon_asic r600_asic = {
 		.wait_for_vblank = &avivo_wait_for_vblank,
 		.set_backlight_level = &atombios_set_backlight_level,
 		.get_backlight_level = &atombios_get_backlight_level,
+		.hdmi_enable = &r600_hdmi_enable,
+		.hdmi_setmode = &r600_hdmi_setmode,
 	},
 	.copy = {
 		.blit = &r600_copy_blit,
@@ -1056,6 +1062,8 @@ static struct radeon_asic rs780_asic = {
 		.wait_for_vblank = &avivo_wait_for_vblank,
 		.set_backlight_level = &atombios_set_backlight_level,
 		.get_backlight_level = &atombios_get_backlight_level,
+		.hdmi_enable = &r600_hdmi_enable,
+		.hdmi_setmode = &r600_hdmi_setmode,
 	},
 	.copy = {
 		.blit = &r600_copy_blit,
@@ -1130,6 +1138,15 @@ static struct radeon_asic rv770_asic = {
 			.ring_test = &r600_dma_ring_test,
 			.ib_test = &r600_dma_ib_test,
 			.is_lockup = &r600_dma_is_lockup,
+		},
+		[R600_RING_TYPE_UVD_INDEX] = {
+			.ib_execute = &r600_uvd_ib_execute,
+			.emit_fence = &r600_uvd_fence_emit,
+			.emit_semaphore = &r600_uvd_semaphore_emit,
+			.cs_parse = &radeon_uvd_cs_parse,
+			.ring_test = &r600_uvd_ring_test,
+			.ib_test = &r600_uvd_ib_test,
+			.is_lockup = &radeon_ring_test_lockup,
 		}
 	},
 	.irq = {
@@ -1142,6 +1159,8 @@ static struct radeon_asic rv770_asic = {
 		.wait_for_vblank = &avivo_wait_for_vblank,
 		.set_backlight_level = &atombios_set_backlight_level,
 		.get_backlight_level = &atombios_get_backlight_level,
+		.hdmi_enable = &r600_hdmi_enable,
+		.hdmi_setmode = &r600_hdmi_setmode,
 	},
 	.copy = {
 		.blit = &r600_copy_blit,
@@ -1174,6 +1193,7 @@ static struct radeon_asic rv770_asic = {
 		.get_pcie_lanes = &r600_get_pcie_lanes,
 		.set_pcie_lanes = &r600_set_pcie_lanes,
 		.set_clock_gating = &radeon_atom_set_clock_gating,
+		.set_uvd_clocks = &rv770_set_uvd_clocks,
 	},
 	.pflip = {
 		.pre_page_flip = &rs600_pre_page_flip,
@@ -1216,6 +1236,15 @@ static struct radeon_asic evergreen_asic = {
 			.ring_test = &r600_dma_ring_test,
 			.ib_test = &r600_dma_ib_test,
 			.is_lockup = &evergreen_dma_is_lockup,
+		},
+		[R600_RING_TYPE_UVD_INDEX] = {
+			.ib_execute = &r600_uvd_ib_execute,
+			.emit_fence = &r600_uvd_fence_emit,
+			.emit_semaphore = &r600_uvd_semaphore_emit,
+			.cs_parse = &radeon_uvd_cs_parse,
+			.ring_test = &r600_uvd_ring_test,
+			.ib_test = &r600_uvd_ib_test,
+			.is_lockup = &radeon_ring_test_lockup,
 		}
 	},
 	.irq = {
@@ -1228,6 +1257,8 @@ static struct radeon_asic evergreen_asic = {
 		.wait_for_vblank = &dce4_wait_for_vblank,
 		.set_backlight_level = &atombios_set_backlight_level,
 		.get_backlight_level = &atombios_get_backlight_level,
+		.hdmi_enable = &evergreen_hdmi_enable,
+		.hdmi_setmode = &evergreen_hdmi_setmode,
 	},
 	.copy = {
 		.blit = &r600_copy_blit,
@@ -1260,6 +1291,7 @@ static struct radeon_asic evergreen_asic = {
 		.get_pcie_lanes = &r600_get_pcie_lanes,
 		.set_pcie_lanes = &r600_set_pcie_lanes,
 		.set_clock_gating = NULL,
+		.set_uvd_clocks = &evergreen_set_uvd_clocks,
 	},
 	.pflip = {
 		.pre_page_flip = &evergreen_pre_page_flip,
@@ -1302,6 +1334,15 @@ static struct radeon_asic sumo_asic = {
 			.ring_test = &r600_dma_ring_test,
 			.ib_test = &r600_dma_ib_test,
 			.is_lockup = &evergreen_dma_is_lockup,
+		},
+		[R600_RING_TYPE_UVD_INDEX] = {
+			.ib_execute = &r600_uvd_ib_execute,
+			.emit_fence = &r600_uvd_fence_emit,
+			.emit_semaphore = &r600_uvd_semaphore_emit,
+			.cs_parse = &radeon_uvd_cs_parse,
+			.ring_test = &r600_uvd_ring_test,
+			.ib_test = &r600_uvd_ib_test,
+			.is_lockup = &radeon_ring_test_lockup,
 		}
 	},
 	.irq = {
@@ -1314,6 +1355,8 @@ static struct radeon_asic sumo_asic = {
 		.wait_for_vblank = &dce4_wait_for_vblank,
 		.set_backlight_level = &atombios_set_backlight_level,
 		.get_backlight_level = &atombios_get_backlight_level,
+		.hdmi_enable = &evergreen_hdmi_enable,
+		.hdmi_setmode = &evergreen_hdmi_setmode,
 	},
 	.copy = {
 		.blit = &r600_copy_blit,
@@ -1346,6 +1389,7 @@ static struct radeon_asic sumo_asic = {
 		.get_pcie_lanes = NULL,
 		.set_pcie_lanes = NULL,
 		.set_clock_gating = NULL,
+		.set_uvd_clocks = &sumo_set_uvd_clocks,
 	},
 	.pflip = {
 		.pre_page_flip = &evergreen_pre_page_flip,
@@ -1388,6 +1432,15 @@ static struct radeon_asic btc_asic = {
 			.ring_test = &r600_dma_ring_test,
 			.ib_test = &r600_dma_ib_test,
 			.is_lockup = &evergreen_dma_is_lockup,
+		},
+		[R600_RING_TYPE_UVD_INDEX] = {
+			.ib_execute = &r600_uvd_ib_execute,
+			.emit_fence = &r600_uvd_fence_emit,
+			.emit_semaphore = &r600_uvd_semaphore_emit,
+			.cs_parse = &radeon_uvd_cs_parse,
+			.ring_test = &r600_uvd_ring_test,
+			.ib_test = &r600_uvd_ib_test,
+			.is_lockup = &radeon_ring_test_lockup,
 		}
 	},
 	.irq = {
@@ -1400,6 +1453,8 @@ static struct radeon_asic btc_asic = {
 		.wait_for_vblank = &dce4_wait_for_vblank,
 		.set_backlight_level = &atombios_set_backlight_level,
 		.get_backlight_level = &atombios_get_backlight_level,
+		.hdmi_enable = &evergreen_hdmi_enable,
+		.hdmi_setmode = &evergreen_hdmi_setmode,
 	},
 	.copy = {
 		.blit = &r600_copy_blit,
@@ -1429,9 +1484,10 @@ static struct radeon_asic btc_asic = {
 		.set_engine_clock = &radeon_atom_set_engine_clock,
 		.get_memory_clock = &radeon_atom_get_memory_clock,
 		.set_memory_clock = &radeon_atom_set_memory_clock,
-		.get_pcie_lanes = NULL,
-		.set_pcie_lanes = NULL,
+		.get_pcie_lanes = &r600_get_pcie_lanes,
+		.set_pcie_lanes = &r600_set_pcie_lanes,
 		.set_clock_gating = NULL,
+		.set_uvd_clocks = &evergreen_set_uvd_clocks,
 	},
 	.pflip = {
 		.pre_page_flip = &evergreen_pre_page_flip,
@@ -1517,6 +1573,15 @@ static struct radeon_asic cayman_asic = {
 			.ib_test = &r600_dma_ib_test,
 			.is_lockup = &cayman_dma_is_lockup,
 			.vm_flush = &cayman_dma_vm_flush,
+		},
+		[R600_RING_TYPE_UVD_INDEX] = {
+			.ib_execute = &r600_uvd_ib_execute,
+			.emit_fence = &r600_uvd_fence_emit,
+			.emit_semaphore = &cayman_uvd_semaphore_emit,
+			.cs_parse = &radeon_uvd_cs_parse,
+			.ring_test = &r600_uvd_ring_test,
+			.ib_test = &r600_uvd_ib_test,
+			.is_lockup = &radeon_ring_test_lockup,
 		}
 	},
 	.irq = {
@@ -1529,6 +1594,8 @@ static struct radeon_asic cayman_asic = {
 		.wait_for_vblank = &dce4_wait_for_vblank,
 		.set_backlight_level = &atombios_set_backlight_level,
 		.get_backlight_level = &atombios_get_backlight_level,
+		.hdmi_enable = &evergreen_hdmi_enable,
+		.hdmi_setmode = &evergreen_hdmi_setmode,
 	},
 	.copy = {
 		.blit = &r600_copy_blit,
@@ -1558,9 +1625,10 @@ static struct radeon_asic cayman_asic = {
 		.set_engine_clock = &radeon_atom_set_engine_clock,
 		.get_memory_clock = &radeon_atom_get_memory_clock,
 		.set_memory_clock = &radeon_atom_set_memory_clock,
-		.get_pcie_lanes = NULL,
-		.set_pcie_lanes = NULL,
+		.get_pcie_lanes = &r600_get_pcie_lanes,
+		.set_pcie_lanes = &r600_set_pcie_lanes,
 		.set_clock_gating = NULL,
+		.set_uvd_clocks = &evergreen_set_uvd_clocks,
 	},
 	.pflip = {
 		.pre_page_flip = &evergreen_pre_page_flip,
@@ -1646,6 +1714,15 @@ static struct radeon_asic trinity_asic = {
 			.ib_test = &r600_dma_ib_test,
 			.is_lockup = &cayman_dma_is_lockup,
 			.vm_flush = &cayman_dma_vm_flush,
+		},
+		[R600_RING_TYPE_UVD_INDEX] = {
+			.ib_execute = &r600_uvd_ib_execute,
+			.emit_fence = &r600_uvd_fence_emit,
+			.emit_semaphore = &cayman_uvd_semaphore_emit,
+			.cs_parse = &radeon_uvd_cs_parse,
+			.ring_test = &r600_uvd_ring_test,
+			.ib_test = &r600_uvd_ib_test,
+			.is_lockup = &radeon_ring_test_lockup,
 		}
 	},
 	.irq = {
@@ -1690,6 +1767,7 @@ static struct radeon_asic trinity_asic = {
 		.get_pcie_lanes = NULL,
 		.set_pcie_lanes = NULL,
 		.set_clock_gating = NULL,
+		.set_uvd_clocks = &sumo_set_uvd_clocks,
 	},
 	.pflip = {
 		.pre_page_flip = &evergreen_pre_page_flip,
@@ -1775,6 +1853,15 @@ static struct radeon_asic si_asic = {
 			.ib_test = &r600_dma_ib_test,
 			.is_lockup = &si_dma_is_lockup,
 			.vm_flush = &si_dma_vm_flush,
+		},
+		[R600_RING_TYPE_UVD_INDEX] = {
+			.ib_execute = &r600_uvd_ib_execute,
+			.emit_fence = &r600_uvd_fence_emit,
+			.emit_semaphore = &cayman_uvd_semaphore_emit,
+			.cs_parse = &radeon_uvd_cs_parse,
+			.ring_test = &r600_uvd_ring_test,
+			.ib_test = &r600_uvd_ib_test,
+			.is_lockup = &radeon_ring_test_lockup,
 		}
 	},
 	.irq = {
@@ -1816,9 +1903,10 @@ static struct radeon_asic si_asic = {
 		.set_engine_clock = &radeon_atom_set_engine_clock,
 		.get_memory_clock = &radeon_atom_get_memory_clock,
 		.set_memory_clock = &radeon_atom_set_memory_clock,
-		.get_pcie_lanes = NULL,
-		.set_pcie_lanes = NULL,
+		.get_pcie_lanes = &r600_get_pcie_lanes,
+		.set_pcie_lanes = &r600_set_pcie_lanes,
 		.set_clock_gating = NULL,
+		.set_uvd_clocks = &si_set_uvd_clocks,
 	},
 	.pflip = {
 		.pre_page_flip = &evergreen_pre_page_flip,
diff --git a/drivers/gpu/drm/radeon/radeon_asic.h b/drivers/gpu/drm/radeon/radeon_asic.h
index 3535f73ad3e2..2c87365d345f 100644
--- a/drivers/gpu/drm/radeon/radeon_asic.h
+++ b/drivers/gpu/drm/radeon/radeon_asic.h
@@ -330,6 +330,7 @@ int r600_dma_ib_test(struct radeon_device *rdev, struct radeon_ring *ring);
 void r600_ring_ib_execute(struct radeon_device *rdev, struct radeon_ib *ib);
 int r600_ring_test(struct radeon_device *rdev, struct radeon_ring *cp);
 int r600_dma_ring_test(struct radeon_device *rdev, struct radeon_ring *cp);
+int r600_uvd_ring_test(struct radeon_device *rdev, struct radeon_ring *ring);
 int r600_copy_blit(struct radeon_device *rdev,
 		   uint64_t src_offset, uint64_t dst_offset,
 		   unsigned num_gpu_pages, struct radeon_fence **fence);
@@ -373,11 +374,12 @@ void r600_disable_interrupts(struct radeon_device *rdev);
 void r600_rlc_stop(struct radeon_device *rdev);
 /* r600 audio */
 int r600_audio_init(struct radeon_device *rdev);
-void r600_audio_set_clock(struct drm_encoder *encoder, int clock);
 struct r600_audio r600_audio_status(struct radeon_device *rdev);
 void r600_audio_fini(struct radeon_device *rdev);
 int r600_hdmi_buffer_status_changed(struct drm_encoder *encoder);
 void r600_hdmi_update_audio_settings(struct drm_encoder *encoder);
+void r600_hdmi_enable(struct drm_encoder *encoder, bool enable);
+void r600_hdmi_setmode(struct drm_encoder *encoder, struct drm_display_mode *mode);
 /* r600 blit */
 int r600_blit_prepare_copy(struct radeon_device *rdev, unsigned num_gpu_pages,
 			   struct radeon_fence **fence, struct radeon_sa_bo **vb,
@@ -392,6 +394,19 @@ int r600_mc_wait_for_idle(struct radeon_device *rdev);
 u32 r600_get_xclk(struct radeon_device *rdev);
 uint64_t r600_get_gpu_clock_counter(struct radeon_device *rdev);
 
+/* uvd */
+int r600_uvd_init(struct radeon_device *rdev);
+int r600_uvd_rbc_start(struct radeon_device *rdev);
+void r600_uvd_rbc_stop(struct radeon_device *rdev);
+int r600_uvd_ib_test(struct radeon_device *rdev, struct radeon_ring *ring);
+void r600_uvd_fence_emit(struct radeon_device *rdev,
+			 struct radeon_fence *fence);
+void r600_uvd_semaphore_emit(struct radeon_device *rdev,
+			     struct radeon_ring *ring,
+			     struct radeon_semaphore *semaphore,
+			     bool emit_wait);
+void r600_uvd_ib_execute(struct radeon_device *rdev, struct radeon_ib *ib);
+
 /*
  * rv770,rv730,rv710,rv740
  */
@@ -409,6 +424,8 @@ int rv770_copy_dma(struct radeon_device *rdev,
 		  unsigned num_gpu_pages,
 		   struct radeon_fence **fence);
 u32 rv770_get_xclk(struct radeon_device *rdev);
+int rv770_uvd_resume(struct radeon_device *rdev);
+int rv770_set_uvd_clocks(struct radeon_device *rdev, u32 vclk, u32 dclk);
 
 /*
  * evergreen
@@ -444,6 +461,8 @@ extern void evergreen_pm_prepare(struct radeon_device *rdev);
 extern void evergreen_pm_finish(struct radeon_device *rdev);
 extern void sumo_pm_init_profile(struct radeon_device *rdev);
 extern void btc_pm_init_profile(struct radeon_device *rdev);
+int sumo_set_uvd_clocks(struct radeon_device *rdev, u32 vclk, u32 dclk);
+int evergreen_set_uvd_clocks(struct radeon_device *rdev, u32 vclk, u32 dclk);
 extern void evergreen_pre_page_flip(struct radeon_device *rdev, int crtc);
 extern u32 evergreen_page_flip(struct radeon_device *rdev, int crtc, u64 crtc_base);
 extern void evergreen_post_page_flip(struct radeon_device *rdev, int crtc);
@@ -459,12 +478,18 @@ int evergreen_copy_dma(struct radeon_device *rdev,
 		       uint64_t src_offset, uint64_t dst_offset,
 		       unsigned num_gpu_pages,
 		       struct radeon_fence **fence);
+void evergreen_hdmi_enable(struct drm_encoder *encoder, bool enable);
+void evergreen_hdmi_setmode(struct drm_encoder *encoder, struct drm_display_mode *mode);
 
 /*
  * cayman
  */
 void cayman_fence_ring_emit(struct radeon_device *rdev,
 			    struct radeon_fence *fence);
+void cayman_uvd_semaphore_emit(struct radeon_device *rdev,
+			       struct radeon_ring *ring,
+			       struct radeon_semaphore *semaphore,
+			       bool emit_wait);
 void cayman_pcie_gart_tlb_flush(struct radeon_device *rdev);
 int cayman_init(struct radeon_device *rdev);
 void cayman_fini(struct radeon_device *rdev);
@@ -524,5 +549,6 @@ int si_copy_dma(struct radeon_device *rdev,
 void si_dma_vm_flush(struct radeon_device *rdev, int ridx, struct radeon_vm *vm);
 u32 si_get_xclk(struct radeon_device *rdev);
 uint64_t si_get_gpu_clock_counter(struct radeon_device *rdev);
+int si_set_uvd_clocks(struct radeon_device *rdev, u32 vclk, u32 dclk);
 
 #endif
diff --git a/drivers/gpu/drm/radeon/radeon_atombios.c b/drivers/gpu/drm/radeon/radeon_atombios.c
index f22eb5713528..dea6f63c9724 100644
--- a/drivers/gpu/drm/radeon/radeon_atombios.c
+++ b/drivers/gpu/drm/radeon/radeon_atombios.c
@@ -2028,6 +2028,8 @@ static int radeon_atombios_parse_power_table_1_3(struct radeon_device *rdev)
 	num_modes = power_info->info.ucNumOfPowerModeEntries;
 	if (num_modes > ATOM_MAX_NUMBEROF_POWER_BLOCK)
 		num_modes = ATOM_MAX_NUMBEROF_POWER_BLOCK;
+	if (num_modes == 0)
+		return state_index;
 	rdev->pm.power_state = kzalloc(sizeof(struct radeon_power_state) * num_modes, GFP_KERNEL);
 	if (!rdev->pm.power_state)
 		return state_index;
@@ -2307,7 +2309,7 @@ static void radeon_atombios_parse_pplib_non_clock_info(struct radeon_device *rde
 		rdev->pm.default_power_state_index = state_index;
 		rdev->pm.power_state[state_index].default_clock_mode =
 			&rdev->pm.power_state[state_index].clock_info[mode_index - 1];
-		if (ASIC_IS_DCE5(rdev) && !(rdev->flags & RADEON_IS_IGP)) {
+		if ((rdev->family >= CHIP_BARTS) && !(rdev->flags & RADEON_IS_IGP)) {
 			/* NI chips post without MC ucode, so default clocks are strobe mode only */
 			rdev->pm.default_sclk = rdev->pm.power_state[state_index].clock_info[0].sclk;
 			rdev->pm.default_mclk = rdev->pm.power_state[state_index].clock_info[0].mclk;
@@ -2345,7 +2347,7 @@ static bool radeon_atombios_parse_pplib_clock_info(struct radeon_device *rdev,
 			sclk |= clock_info->rs780.ucLowEngineClockHigh << 16;
 			rdev->pm.power_state[state_index].clock_info[mode_index].sclk = sclk;
 		}
-	} else if (ASIC_IS_DCE6(rdev)) {
+	} else if (rdev->family >= CHIP_TAHITI) {
 		sclk = le16_to_cpu(clock_info->si.usEngineClockLow);
 		sclk |= clock_info->si.ucEngineClockHigh << 16;
 		mclk = le16_to_cpu(clock_info->si.usMemoryClockLow);
@@ -2358,7 +2360,7 @@ static bool radeon_atombios_parse_pplib_clock_info(struct radeon_device *rdev,
 			le16_to_cpu(clock_info->si.usVDDC);
 		rdev->pm.power_state[state_index].clock_info[mode_index].voltage.vddci =
 			le16_to_cpu(clock_info->si.usVDDCI);
-	} else if (ASIC_IS_DCE4(rdev)) {
+	} else if (rdev->family >= CHIP_CEDAR) {
 		sclk = le16_to_cpu(clock_info->evergreen.usEngineClockLow);
 		sclk |= clock_info->evergreen.ucEngineClockHigh << 16;
 		mclk = le16_to_cpu(clock_info->evergreen.usMemoryClockLow);
@@ -2432,6 +2434,8 @@ static int radeon_atombios_parse_power_table_4_5(struct radeon_device *rdev)
 	power_info = (union power_info *)(mode_info->atom_context->bios + data_offset);
 
 	radeon_atombios_add_pplib_thermal_controller(rdev, &power_info->pplib.sThermalController);
+	if (power_info->pplib.ucNumStates == 0)
+		return state_index;
 	rdev->pm.power_state = kzalloc(sizeof(struct radeon_power_state) *
 				       power_info->pplib.ucNumStates, GFP_KERNEL);
 	if (!rdev->pm.power_state)
@@ -2514,6 +2518,7 @@ static int radeon_atombios_parse_power_table_6(struct radeon_device *rdev)
 	int index = GetIndexIntoMasterTable(DATA, PowerPlayInfo);
         u16 data_offset;
 	u8 frev, crev;
+	u8 *power_state_offset;
 
 	if (!atom_parse_data_header(mode_info->atom_context, index, NULL,
 				   &frev, &crev, &data_offset))
@@ -2530,15 +2535,17 @@ static int radeon_atombios_parse_power_table_6(struct radeon_device *rdev)
 	non_clock_info_array = (struct _NonClockInfoArray *)
 		(mode_info->atom_context->bios + data_offset +
 		 le16_to_cpu(power_info->pplib.usNonClockInfoArrayOffset));
+	if (state_array->ucNumEntries == 0)
+		return state_index;
 	rdev->pm.power_state = kzalloc(sizeof(struct radeon_power_state) *
 				       state_array->ucNumEntries, GFP_KERNEL);
 	if (!rdev->pm.power_state)
 		return state_index;
+	power_state_offset = (u8 *)state_array->states;
 	for (i = 0; i < state_array->ucNumEntries; i++) {
 		mode_index = 0;
-		power_state = (union pplib_power_state *)&state_array->states[i];
-		/* XXX this might be an inagua bug... */
-		non_clock_array_index = i; /* power_state->v2.nonClockInfoIndex */
+		power_state = (union pplib_power_state *)power_state_offset;
+		non_clock_array_index = power_state->v2.nonClockInfoIndex;
 		non_clock_info = (struct _ATOM_PPLIB_NONCLOCK_INFO *)
 			&non_clock_info_array->nonClockInfo[non_clock_array_index];
 		rdev->pm.power_state[i].clock_info = kzalloc(sizeof(struct radeon_pm_clock_info) *
@@ -2550,9 +2557,6 @@ static int radeon_atombios_parse_power_table_6(struct radeon_device *rdev)
 		if (power_state->v2.ucNumDPMLevels) {
 			for (j = 0; j < power_state->v2.ucNumDPMLevels; j++) {
 				clock_array_index = power_state->v2.clockInfoIndex[j];
-				/* XXX this might be an inagua bug... */
-				if (clock_array_index >= clock_info_array->ucNumEntries)
-					continue;
 				clock_info = (union pplib_clock_info *)
 					&clock_info_array->clockInfo[clock_array_index * clock_info_array->ucEntrySize];
 				valid = radeon_atombios_parse_pplib_clock_info(rdev,
@@ -2574,6 +2578,7 @@ static int radeon_atombios_parse_power_table_6(struct radeon_device *rdev)
 								   non_clock_info);
 			state_index++;
 		}
+		power_state_offset += 2 + power_state->v2.ucNumDPMLevels;
 	}
 	/* if multiple clock modes, mark the lowest as no display */
 	for (i = 0; i < state_index; i++) {
@@ -2620,7 +2625,9 @@ void radeon_atombios_get_power_modes(struct radeon_device *rdev)
 		default:
 			break;
 		}
-	} else {
+	}
+
+	if (state_index == 0) {
 		rdev->pm.power_state = kzalloc(sizeof(struct radeon_power_state), GFP_KERNEL);
 		if (rdev->pm.power_state) {
 			rdev->pm.power_state[0].clock_info =
@@ -2654,6 +2661,111 @@ void radeon_atombios_get_power_modes(struct radeon_device *rdev)
 		rdev->pm.current_vddc = 0;
 }
 
+union get_clock_dividers {
+	struct _COMPUTE_MEMORY_ENGINE_PLL_PARAMETERS v1;
+	struct _COMPUTE_MEMORY_ENGINE_PLL_PARAMETERS_V2 v2;
+	struct _COMPUTE_MEMORY_ENGINE_PLL_PARAMETERS_V3 v3;
+	struct _COMPUTE_MEMORY_ENGINE_PLL_PARAMETERS_V4 v4;
+	struct _COMPUTE_MEMORY_ENGINE_PLL_PARAMETERS_V5 v5;
+};
+
+int radeon_atom_get_clock_dividers(struct radeon_device *rdev,
+				   u8 clock_type,
+				   u32 clock,
+				   bool strobe_mode,
+				   struct atom_clock_dividers *dividers)
+{
+	union get_clock_dividers args;
+	int index = GetIndexIntoMasterTable(COMMAND, ComputeMemoryEnginePLL);
+	u8 frev, crev;
+
+	memset(&args, 0, sizeof(args));
+	memset(dividers, 0, sizeof(struct atom_clock_dividers));
+
+	if (!atom_parse_cmd_header(rdev->mode_info.atom_context, index, &frev, &crev))
+		return -EINVAL;
+
+	switch (crev) {
+	case 1:
+		/* r4xx, r5xx */
+		args.v1.ucAction = clock_type;
+		args.v1.ulClock = cpu_to_le32(clock);	/* 10 khz */
+
+		atom_execute_table(rdev->mode_info.atom_context, index, (uint32_t *)&args);
+
+		dividers->post_div = args.v1.ucPostDiv;
+		dividers->fb_div = args.v1.ucFbDiv;
+		dividers->enable_post_div = true;
+		break;
+	case 2:
+	case 3:
+		/* r6xx, r7xx, evergreen, ni */
+		if (rdev->family <= CHIP_RV770) {
+			args.v2.ucAction = clock_type;
+			args.v2.ulClock = cpu_to_le32(clock);	/* 10 khz */
+
+			atom_execute_table(rdev->mode_info.atom_context, index, (uint32_t *)&args);
+
+			dividers->post_div = args.v2.ucPostDiv;
+			dividers->fb_div = le16_to_cpu(args.v2.usFbDiv);
+			dividers->ref_div = args.v2.ucAction;
+			if (rdev->family == CHIP_RV770) {
+				dividers->enable_post_div = (le32_to_cpu(args.v2.ulClock) & (1 << 24)) ?
+					true : false;
+				dividers->vco_mode = (le32_to_cpu(args.v2.ulClock) & (1 << 25)) ? 1 : 0;
+			} else
+				dividers->enable_post_div = (dividers->fb_div & 1) ? true : false;
+		} else {
+			if (clock_type == COMPUTE_ENGINE_PLL_PARAM) {
+				args.v3.ulClockParams = cpu_to_le32((clock_type << 24) | clock);
+
+				atom_execute_table(rdev->mode_info.atom_context, index, (uint32_t *)&args);
+
+				dividers->post_div = args.v3.ucPostDiv;
+				dividers->enable_post_div = (args.v3.ucCntlFlag &
+							     ATOM_PLL_CNTL_FLAG_PLL_POST_DIV_EN) ? true : false;
+				dividers->enable_dithen = (args.v3.ucCntlFlag &
+							   ATOM_PLL_CNTL_FLAG_FRACTION_DISABLE) ? false : true;
+				dividers->fb_div = le16_to_cpu(args.v3.ulFbDiv.usFbDiv);
+				dividers->frac_fb_div = le16_to_cpu(args.v3.ulFbDiv.usFbDivFrac);
+				dividers->ref_div = args.v3.ucRefDiv;
+				dividers->vco_mode = (args.v3.ucCntlFlag &
+						      ATOM_PLL_CNTL_FLAG_MPLL_VCO_MODE) ? 1 : 0;
+			} else {
+				args.v5.ulClockParams = cpu_to_le32((clock_type << 24) | clock);
+				if (strobe_mode)
+					args.v5.ucInputFlag = ATOM_PLL_INPUT_FLAG_PLL_STROBE_MODE_EN;
+
+				atom_execute_table(rdev->mode_info.atom_context, index, (uint32_t *)&args);
+
+				dividers->post_div = args.v5.ucPostDiv;
+				dividers->enable_post_div = (args.v5.ucCntlFlag &
+							     ATOM_PLL_CNTL_FLAG_PLL_POST_DIV_EN) ? true : false;
+				dividers->enable_dithen = (args.v5.ucCntlFlag &
+							   ATOM_PLL_CNTL_FLAG_FRACTION_DISABLE) ? false : true;
+				dividers->whole_fb_div = le16_to_cpu(args.v5.ulFbDiv.usFbDiv);
+				dividers->frac_fb_div = le16_to_cpu(args.v5.ulFbDiv.usFbDivFrac);
+				dividers->ref_div = args.v5.ucRefDiv;
+				dividers->vco_mode = (args.v5.ucCntlFlag &
+						      ATOM_PLL_CNTL_FLAG_MPLL_VCO_MODE) ? 1 : 0;
+			}
+		}
+		break;
+	case 4:
+		/* fusion */
+		args.v4.ulClock = cpu_to_le32(clock);	/* 10 khz */
+
+		atom_execute_table(rdev->mode_info.atom_context, index, (uint32_t *)&args);
+
+		dividers->post_div = args.v4.ucPostDiv;
+		dividers->real_clock = le32_to_cpu(args.v4.ulClock);
+		break;
+	default:
+		return -EINVAL;
+	}
+	return 0;
+}
+
 void radeon_atom_set_clock_gating(struct radeon_device *rdev, int enable)
 {
 	DYNAMIC_CLOCK_GATING_PS_ALLOCATION args;
diff --git a/drivers/gpu/drm/radeon/radeon_cs.c b/drivers/gpu/drm/radeon/radeon_cs.c
index 70d38241b083..7e265a58141f 100644
--- a/drivers/gpu/drm/radeon/radeon_cs.c
+++ b/drivers/gpu/drm/radeon/radeon_cs.c
@@ -63,30 +63,50 @@ static int radeon_cs_parser_relocs(struct radeon_cs_parser *p)
 				break;
 			}
 		}
-		if (!duplicate) {
-			p->relocs[i].gobj = drm_gem_object_lookup(ddev,
-								  p->filp,
-								  r->handle);
-			if (p->relocs[i].gobj == NULL) {
-				DRM_ERROR("gem object lookup failed 0x%x\n",
-					  r->handle);
-				return -ENOENT;
-			}
-			p->relocs_ptr[i] = &p->relocs[i];
-			p->relocs[i].robj = gem_to_radeon_bo(p->relocs[i].gobj);
-			p->relocs[i].lobj.bo = p->relocs[i].robj;
-			p->relocs[i].lobj.wdomain = r->write_domain;
-			p->relocs[i].lobj.rdomain = r->read_domains;
-			p->relocs[i].lobj.tv.bo = &p->relocs[i].robj->tbo;
-			p->relocs[i].handle = r->handle;
-			p->relocs[i].flags = r->flags;
-			radeon_bo_list_add_object(&p->relocs[i].lobj,
-						  &p->validated);
-
-		} else
+		if (duplicate) {
 			p->relocs[i].handle = 0;
+			continue;
+		}
+
+		p->relocs[i].gobj = drm_gem_object_lookup(ddev, p->filp,
+							  r->handle);
+		if (p->relocs[i].gobj == NULL) {
+			DRM_ERROR("gem object lookup failed 0x%x\n",
+				  r->handle);
+			return -ENOENT;
+		}
+		p->relocs_ptr[i] = &p->relocs[i];
+		p->relocs[i].robj = gem_to_radeon_bo(p->relocs[i].gobj);
+		p->relocs[i].lobj.bo = p->relocs[i].robj;
+		p->relocs[i].lobj.written = !!r->write_domain;
+
+		/* the first reloc of an UVD job is the
+		   msg and that must be in VRAM */
+		if (p->ring == R600_RING_TYPE_UVD_INDEX && i == 0) {
+			/* TODO: is this still needed for NI+ ? */
+			p->relocs[i].lobj.domain =
+				RADEON_GEM_DOMAIN_VRAM;
+
+			p->relocs[i].lobj.alt_domain =
+				RADEON_GEM_DOMAIN_VRAM;
+
+		} else {
+			uint32_t domain = r->write_domain ?
+				r->write_domain : r->read_domains;
+
+			p->relocs[i].lobj.domain = domain;
+			if (domain == RADEON_GEM_DOMAIN_VRAM)
+				domain |= RADEON_GEM_DOMAIN_GTT;
+			p->relocs[i].lobj.alt_domain = domain;
+		}
+
+		p->relocs[i].lobj.tv.bo = &p->relocs[i].robj->tbo;
+		p->relocs[i].handle = r->handle;
+
+		radeon_bo_list_add_object(&p->relocs[i].lobj,
+					  &p->validated);
 	}
-	return radeon_bo_list_validate(&p->validated);
+	return radeon_bo_list_validate(&p->validated, p->ring);
 }
 
 static int radeon_cs_get_ring(struct radeon_cs_parser *p, u32 ring, s32 priority)
@@ -121,6 +141,9 @@ static int radeon_cs_get_ring(struct radeon_cs_parser *p, u32 ring, s32 priority
 			return -EINVAL;
 		}
 		break;
+	case RADEON_CS_RING_UVD:
+		p->ring = R600_RING_TYPE_UVD_INDEX;
+		break;
 	}
 	return 0;
 }
@@ -241,15 +264,15 @@ int radeon_cs_parser_init(struct radeon_cs_parser *p, void *data)
 			return -EINVAL;
 		}
 
-		/* we only support VM on SI+ */
-		if ((p->rdev->family >= CHIP_TAHITI) &&
-		    ((p->cs_flags & RADEON_CS_USE_VM) == 0)) {
-			DRM_ERROR("VM required on SI+!\n");
+		if (radeon_cs_get_ring(p, ring, priority))
 			return -EINVAL;
-		}
 
-		if (radeon_cs_get_ring(p, ring, priority))
+		/* we only support VM on some SI+ rings */
+		if ((p->rdev->asic->ring[p->ring].cs_parse == NULL) &&
+		   ((p->cs_flags & RADEON_CS_USE_VM) == 0)) {
+			DRM_ERROR("Ring %d requires VM!\n", p->ring);
 			return -EINVAL;
+		}
 	}
 
 	/* deal with non-vm */
@@ -526,6 +549,10 @@ int radeon_cs_ioctl(struct drm_device *dev, void *data, struct drm_file *filp)
 		r = radeon_cs_handle_lockup(rdev, r);
 		return r;
 	}
+
+	if (parser.ring == R600_RING_TYPE_UVD_INDEX)
+		radeon_uvd_note_usage(rdev);
+
 	r = radeon_cs_ib_chunk(rdev, &parser);
 	if (r) {
 		goto out;
diff --git a/drivers/gpu/drm/radeon/radeon_device.c b/drivers/gpu/drm/radeon/radeon_device.c
index 44b8034a400d..a8f608903989 100644
--- a/drivers/gpu/drm/radeon/radeon_device.c
+++ b/drivers/gpu/drm/radeon/radeon_device.c
@@ -98,6 +98,42 @@ static const char radeon_family_name[][16] = {
 };
 
 /**
+ * radeon_program_register_sequence - program an array of registers.
+ *
+ * @rdev: radeon_device pointer
+ * @registers: pointer to the register array
+ * @array_size: size of the register array
+ *
+ * Programs an array or registers with and and or masks.
+ * This is a helper for setting golden registers.
+ */
+void radeon_program_register_sequence(struct radeon_device *rdev,
+				      const u32 *registers,
+				      const u32 array_size)
+{
+	u32 tmp, reg, and_mask, or_mask;
+	int i;
+
+	if (array_size % 3)
+		return;
+
+	for (i = 0; i < array_size; i +=3) {
+		reg = registers[i + 0];
+		and_mask = registers[i + 1];
+		or_mask = registers[i + 2];
+
+		if (and_mask == 0xffffffff) {
+			tmp = or_mask;
+		} else {
+			tmp = RREG32(reg);
+			tmp &= ~and_mask;
+			tmp |= or_mask;
+		}
+		WREG32(reg, tmp);
+	}
+}
+
+/**
  * radeon_surface_init - Clear GPU surface registers.
  *
  * @rdev: radeon_device pointer
@@ -359,7 +395,7 @@ void radeon_vram_location(struct radeon_device *rdev, struct radeon_mc *mc, u64
 	uint64_t limit = (uint64_t)radeon_vram_limit << 20;
 
 	mc->vram_start = base;
-	if (mc->mc_vram_size > (0xFFFFFFFF - base + 1)) {
+	if (mc->mc_vram_size > (rdev->mc.mc_mask - base + 1)) {
 		dev_warn(rdev->dev, "limiting VRAM to PCI aperture size\n");
 		mc->real_vram_size = mc->aper_size;
 		mc->mc_vram_size = mc->aper_size;
@@ -394,7 +430,7 @@ void radeon_gtt_location(struct radeon_device *rdev, struct radeon_mc *mc)
 {
 	u64 size_af, size_bf;
 
-	size_af = ((0xFFFFFFFF - mc->vram_end) + mc->gtt_base_align) & ~mc->gtt_base_align;
+	size_af = ((rdev->mc.mc_mask - mc->vram_end) + mc->gtt_base_align) & ~mc->gtt_base_align;
 	size_bf = mc->vram_start & ~mc->gtt_base_align;
 	if (size_bf > size_af) {
 		if (mc->gtt_size > size_bf) {
@@ -1068,6 +1104,17 @@ int radeon_device_init(struct radeon_device *rdev,
 		radeon_agp_disable(rdev);
 	}
 
+	/* Set the internal MC address mask
+	 * This is the max address of the GPU's
+	 * internal address space.
+	 */
+	if (rdev->family >= CHIP_CAYMAN)
+		rdev->mc.mc_mask = 0xffffffffffULL; /* 40 bit MC */
+	else if (rdev->family >= CHIP_CEDAR)
+		rdev->mc.mc_mask = 0xfffffffffULL; /* 36 bit MC */
+	else
+		rdev->mc.mc_mask = 0xffffffffULL; /* 32 bit MC */
+
 	/* set DMA mask + need_dma32 flags.
 	 * PCIE - can handle 40-bits.
 	 * IGP - can handle 40-bits
@@ -1131,6 +1178,11 @@ int radeon_device_init(struct radeon_device *rdev,
 	if (r)
 		DRM_ERROR("ib ring test failed (%d).\n", r);
 
+	r = radeon_gem_debugfs_init(rdev);
+	if (r) {
+		DRM_ERROR("registering gem debugfs failed (%d).\n", r);
+	}
+
 	if (rdev->flags & RADEON_IS_AGP && !rdev->accel_working) {
 		/* Acceleration not working on AGP card try again
 		 * with fallback to PCI or PCIE GART
diff --git a/drivers/gpu/drm/radeon/radeon_drv.c b/drivers/gpu/drm/radeon/radeon_drv.c
index 66a7f0fd9620..d33f484ace48 100644
--- a/drivers/gpu/drm/radeon/radeon_drv.c
+++ b/drivers/gpu/drm/radeon/radeon_drv.c
@@ -71,9 +71,12 @@
  *   2.28.0 - r600-eg: Add MEM_WRITE packet support
  *   2.29.0 - R500 FP16 color clear registers
  *   2.30.0 - fix for FMASK texturing
+ *   2.31.0 - Add fastfb support for rs690
+ *   2.32.0 - new info request for rings working
+ *   2.33.0 - Add SI tiling mode array query
  */
 #define KMS_DRIVER_MAJOR	2
-#define KMS_DRIVER_MINOR	30
+#define KMS_DRIVER_MINOR	33
 #define KMS_DRIVER_PATCHLEVEL	0
 int radeon_driver_load_kms(struct drm_device *dev, unsigned long flags);
 int radeon_driver_unload_kms(struct drm_device *dev);
@@ -160,6 +163,7 @@ int radeon_hw_i2c = 0;
 int radeon_pcie_gen2 = -1;
 int radeon_msi = -1;
 int radeon_lockup_timeout = 10000;
+int radeon_fastfb = 0;
 
 MODULE_PARM_DESC(no_wb, "Disable AGP writeback for scratch registers");
 module_param_named(no_wb, radeon_no_wb, int, 0444);
@@ -212,6 +216,9 @@ module_param_named(msi, radeon_msi, int, 0444);
 MODULE_PARM_DESC(lockup_timeout, "GPU lockup timeout in ms (defaul 10000 = 10 seconds, 0 = disable)");
 module_param_named(lockup_timeout, radeon_lockup_timeout, int, 0444);
 
+MODULE_PARM_DESC(fastfb, "Direct FB access for IGP chips (0 = disable, 1 = enable)");
+module_param_named(fastfb, radeon_fastfb, int, 0444);
+
 static struct pci_device_id pciidlist[] = {
 	radeon_PCI_IDS
 };
diff --git a/drivers/gpu/drm/radeon/radeon_fence.c b/drivers/gpu/drm/radeon/radeon_fence.c
index 34356252567a..5b937dfe6f65 100644
--- a/drivers/gpu/drm/radeon/radeon_fence.c
+++ b/drivers/gpu/drm/radeon/radeon_fence.c
@@ -31,9 +31,9 @@
 #include <linux/seq_file.h>
 #include <linux/atomic.h>
 #include <linux/wait.h>
-#include <linux/list.h>
 #include <linux/kref.h>
 #include <linux/slab.h>
+#include <linux/firmware.h>
 #include <drm/drmP.h>
 #include "radeon_reg.h"
 #include "radeon.h"
@@ -768,7 +768,19 @@ int radeon_fence_driver_start_ring(struct radeon_device *rdev, int ring)
 	radeon_scratch_free(rdev, rdev->fence_drv[ring].scratch_reg);
 	if (rdev->wb.use_event || !radeon_ring_supports_scratch_reg(rdev, &rdev->ring[ring])) {
 		rdev->fence_drv[ring].scratch_reg = 0;
-		index = R600_WB_EVENT_OFFSET + ring * 4;
+		if (ring != R600_RING_TYPE_UVD_INDEX) {
+			index = R600_WB_EVENT_OFFSET + ring * 4;
+			rdev->fence_drv[ring].cpu_addr = &rdev->wb.wb[index/4];
+			rdev->fence_drv[ring].gpu_addr = rdev->wb.gpu_addr +
+							 index;
+
+		} else {
+			/* put fence directly behind firmware */
+			index = ALIGN(rdev->uvd_fw->size, 8);
+			rdev->fence_drv[ring].cpu_addr = rdev->uvd.cpu_addr + index;
+			rdev->fence_drv[ring].gpu_addr = rdev->uvd.gpu_addr + index;
+		}
+
 	} else {
 		r = radeon_scratch_get(rdev, &rdev->fence_drv[ring].scratch_reg);
 		if (r) {
@@ -778,9 +790,9 @@ int radeon_fence_driver_start_ring(struct radeon_device *rdev, int ring)
 		index = RADEON_WB_SCRATCH_OFFSET +
 			rdev->fence_drv[ring].scratch_reg -
 			rdev->scratch.reg_base;
+		rdev->fence_drv[ring].cpu_addr = &rdev->wb.wb[index/4];
+		rdev->fence_drv[ring].gpu_addr = rdev->wb.gpu_addr + index;
 	}
-	rdev->fence_drv[ring].cpu_addr = &rdev->wb.wb[index/4];
-	rdev->fence_drv[ring].gpu_addr = rdev->wb.gpu_addr + index;
 	radeon_fence_write(rdev, atomic64_read(&rdev->fence_drv[ring].last_seq), ring);
 	rdev->fence_drv[ring].initialized = true;
 	dev_info(rdev->dev, "fence driver on ring %d use gpu addr 0x%016llx and cpu addr 0x%p\n",
diff --git a/drivers/gpu/drm/radeon/radeon_gem.c b/drivers/gpu/drm/radeon/radeon_gem.c
index fe5c1f6b7957..aa796031ab65 100644
--- a/drivers/gpu/drm/radeon/radeon_gem.c
+++ b/drivers/gpu/drm/radeon/radeon_gem.c
@@ -84,6 +84,7 @@ retry:
 		return r;
 	}
 	*obj = &robj->gem_base;
+	robj->pid = task_pid_nr(current);
 
 	mutex_lock(&rdev->gem.mutex);
 	list_add_tail(&robj->list, &rdev->gem.objects);
@@ -575,3 +576,52 @@ int radeon_mode_dumb_destroy(struct drm_file *file_priv,
 {
 	return drm_gem_handle_delete(file_priv, handle);
 }
+
+#if defined(CONFIG_DEBUG_FS)
+static int radeon_debugfs_gem_info(struct seq_file *m, void *data)
+{
+	struct drm_info_node *node = (struct drm_info_node *)m->private;
+	struct drm_device *dev = node->minor->dev;
+	struct radeon_device *rdev = dev->dev_private;
+	struct radeon_bo *rbo;
+	unsigned i = 0;
+
+	mutex_lock(&rdev->gem.mutex);
+	list_for_each_entry(rbo, &rdev->gem.objects, list) {
+		unsigned domain;
+		const char *placement;
+
+		domain = radeon_mem_type_to_domain(rbo->tbo.mem.mem_type);
+		switch (domain) {
+		case RADEON_GEM_DOMAIN_VRAM:
+			placement = "VRAM";
+			break;
+		case RADEON_GEM_DOMAIN_GTT:
+			placement = " GTT";
+			break;
+		case RADEON_GEM_DOMAIN_CPU:
+		default:
+			placement = " CPU";
+			break;
+		}
+		seq_printf(m, "bo[0x%08x] %8ldkB %8ldMB %s pid %8ld\n",
+			   i, radeon_bo_size(rbo) >> 10, radeon_bo_size(rbo) >> 20,
+			   placement, (unsigned long)rbo->pid);
+		i++;
+	}
+	mutex_unlock(&rdev->gem.mutex);
+	return 0;
+}
+
+static struct drm_info_list radeon_debugfs_gem_list[] = {
+	{"radeon_gem_info", &radeon_debugfs_gem_info, 0, NULL},
+};
+#endif
+
+int radeon_gem_debugfs_init(struct radeon_device *rdev)
+{
+#if defined(CONFIG_DEBUG_FS)
+	return radeon_debugfs_add_files(rdev, radeon_debugfs_gem_list, 1);
+#endif
+	return 0;
+}
diff --git a/drivers/gpu/drm/radeon/radeon_kms.c b/drivers/gpu/drm/radeon/radeon_kms.c
index c75cb2c6ba71..4f2d4f4c1dab 100644
--- a/drivers/gpu/drm/radeon/radeon_kms.c
+++ b/drivers/gpu/drm/radeon/radeon_kms.c
@@ -50,9 +50,13 @@ int radeon_driver_unload_kms(struct drm_device *dev)
 
 	if (rdev == NULL)
 		return 0;
+	if (rdev->rmmio == NULL)
+		goto done_free;
 	radeon_acpi_fini(rdev);
 	radeon_modeset_fini(rdev);
 	radeon_device_fini(rdev);
+
+done_free:
 	kfree(rdev);
 	dev->dev_private = NULL;
 	return 0;
@@ -176,80 +180,65 @@ int radeon_info_ioctl(struct drm_device *dev, void *data, struct drm_file *filp)
 	struct radeon_device *rdev = dev->dev_private;
 	struct drm_radeon_info *info = data;
 	struct radeon_mode_info *minfo = &rdev->mode_info;
-	uint32_t value, *value_ptr;
-	uint64_t value64, *value_ptr64;
+	uint32_t *value, value_tmp, *value_ptr, value_size;
+	uint64_t value64;
 	struct drm_crtc *crtc;
 	int i, found;
 
-	/* TIMESTAMP is a 64-bit value, needs special handling. */
-	if (info->request == RADEON_INFO_TIMESTAMP) {
-		if (rdev->family >= CHIP_R600) {
-			value_ptr64 = (uint64_t*)((unsigned long)info->value);
-			value64 = radeon_get_gpu_clock_counter(rdev);
-
-			if (DRM_COPY_TO_USER(value_ptr64, &value64, sizeof(value64))) {
-				DRM_ERROR("copy_to_user %s:%u\n", __func__, __LINE__);
-				return -EFAULT;
-			}
-			return 0;
-		} else {
-			DRM_DEBUG_KMS("timestamp is r6xx+ only!\n");
-			return -EINVAL;
-		}
-	}
-
 	value_ptr = (uint32_t *)((unsigned long)info->value);
-	if (DRM_COPY_FROM_USER(&value, value_ptr, sizeof(value))) {
-		DRM_ERROR("copy_from_user %s:%u\n", __func__, __LINE__);
-		return -EFAULT;
-	}
+	value = &value_tmp;
+	value_size = sizeof(uint32_t);
 
 	switch (info->request) {
 	case RADEON_INFO_DEVICE_ID:
-		value = dev->pci_device;
+		*value = dev->pci_device;
 		break;
 	case RADEON_INFO_NUM_GB_PIPES:
-		value = rdev->num_gb_pipes;
+		*value = rdev->num_gb_pipes;
 		break;
 	case RADEON_INFO_NUM_Z_PIPES:
-		value = rdev->num_z_pipes;
+		*value = rdev->num_z_pipes;
 		break;
 	case RADEON_INFO_ACCEL_WORKING:
 		/* xf86-video-ati 6.13.0 relies on this being false for evergreen */
 		if ((rdev->family >= CHIP_CEDAR) && (rdev->family <= CHIP_HEMLOCK))
-			value = false;
+			*value = false;
 		else
-			value = rdev->accel_working;
+			*value = rdev->accel_working;
 		break;
 	case RADEON_INFO_CRTC_FROM_ID:
+		if (DRM_COPY_FROM_USER(value, value_ptr, sizeof(uint32_t))) {
+			DRM_ERROR("copy_from_user %s:%u\n", __func__, __LINE__);
+			return -EFAULT;
+		}
 		for (i = 0, found = 0; i < rdev->num_crtc; i++) {
 			crtc = (struct drm_crtc *)minfo->crtcs[i];
-			if (crtc && crtc->base.id == value) {
+			if (crtc && crtc->base.id == *value) {
 				struct radeon_crtc *radeon_crtc = to_radeon_crtc(crtc);
-				value = radeon_crtc->crtc_id;
+				*value = radeon_crtc->crtc_id;
 				found = 1;
 				break;
 			}
 		}
 		if (!found) {
-			DRM_DEBUG_KMS("unknown crtc id %d\n", value);
+			DRM_DEBUG_KMS("unknown crtc id %d\n", *value);
 			return -EINVAL;
 		}
 		break;
 	case RADEON_INFO_ACCEL_WORKING2:
-		value = rdev->accel_working;
+		*value = rdev->accel_working;
 		break;
 	case RADEON_INFO_TILING_CONFIG:
 		if (rdev->family >= CHIP_TAHITI)
-			value = rdev->config.si.tile_config;
+			*value = rdev->config.si.tile_config;
 		else if (rdev->family >= CHIP_CAYMAN)
-			value = rdev->config.cayman.tile_config;
+			*value = rdev->config.cayman.tile_config;
 		else if (rdev->family >= CHIP_CEDAR)
-			value = rdev->config.evergreen.tile_config;
+			*value = rdev->config.evergreen.tile_config;
 		else if (rdev->family >= CHIP_RV770)
-			value = rdev->config.rv770.tile_config;
+			*value = rdev->config.rv770.tile_config;
 		else if (rdev->family >= CHIP_R600)
-			value = rdev->config.r600.tile_config;
+			*value = rdev->config.r600.tile_config;
 		else {
 			DRM_DEBUG_KMS("tiling config is r6xx+ only!\n");
 			return -EINVAL;
@@ -262,73 +251,81 @@ int radeon_info_ioctl(struct drm_device *dev, void *data, struct drm_file *filp)
 		 *
 		 * When returning, the value is 1 if filp owns hyper-z access,
 		 * 0 otherwise. */
-		if (value >= 2) {
-			DRM_DEBUG_KMS("WANT_HYPERZ: invalid value %d\n", value);
+		if (DRM_COPY_FROM_USER(value, value_ptr, sizeof(uint32_t))) {
+			DRM_ERROR("copy_from_user %s:%u\n", __func__, __LINE__);
+			return -EFAULT;
+		}
+		if (*value >= 2) {
+			DRM_DEBUG_KMS("WANT_HYPERZ: invalid value %d\n", *value);
 			return -EINVAL;
 		}
-		radeon_set_filp_rights(dev, &rdev->hyperz_filp, filp, &value);
+		radeon_set_filp_rights(dev, &rdev->hyperz_filp, filp, value);
 		break;
 	case RADEON_INFO_WANT_CMASK:
 		/* The same logic as Hyper-Z. */
-		if (value >= 2) {
-			DRM_DEBUG_KMS("WANT_CMASK: invalid value %d\n", value);
+		if (DRM_COPY_FROM_USER(value, value_ptr, sizeof(uint32_t))) {
+			DRM_ERROR("copy_from_user %s:%u\n", __func__, __LINE__);
+			return -EFAULT;
+		}
+		if (*value >= 2) {
+			DRM_DEBUG_KMS("WANT_CMASK: invalid value %d\n", *value);
 			return -EINVAL;
 		}
-		radeon_set_filp_rights(dev, &rdev->cmask_filp, filp, &value);
+		radeon_set_filp_rights(dev, &rdev->cmask_filp, filp, value);
 		break;
 	case RADEON_INFO_CLOCK_CRYSTAL_FREQ:
 		/* return clock value in KHz */
 		if (rdev->asic->get_xclk)
-			value = radeon_get_xclk(rdev) * 10;
+			*value = radeon_get_xclk(rdev) * 10;
 		else
-			value = rdev->clock.spll.reference_freq * 10;
+			*value = rdev->clock.spll.reference_freq * 10;
 		break;
 	case RADEON_INFO_NUM_BACKENDS:
 		if (rdev->family >= CHIP_TAHITI)
-			value = rdev->config.si.max_backends_per_se *
+			*value = rdev->config.si.max_backends_per_se *
 				rdev->config.si.max_shader_engines;
 		else if (rdev->family >= CHIP_CAYMAN)
-			value = rdev->config.cayman.max_backends_per_se *
+			*value = rdev->config.cayman.max_backends_per_se *
 				rdev->config.cayman.max_shader_engines;
 		else if (rdev->family >= CHIP_CEDAR)
-			value = rdev->config.evergreen.max_backends;
+			*value = rdev->config.evergreen.max_backends;
 		else if (rdev->family >= CHIP_RV770)
-			value = rdev->config.rv770.max_backends;
+			*value = rdev->config.rv770.max_backends;
 		else if (rdev->family >= CHIP_R600)
-			value = rdev->config.r600.max_backends;
+			*value = rdev->config.r600.max_backends;
 		else {
 			return -EINVAL;
 		}
 		break;
 	case RADEON_INFO_NUM_TILE_PIPES:
 		if (rdev->family >= CHIP_TAHITI)
-			value = rdev->config.si.max_tile_pipes;
+			*value = rdev->config.si.max_tile_pipes;
 		else if (rdev->family >= CHIP_CAYMAN)
-			value = rdev->config.cayman.max_tile_pipes;
+			*value = rdev->config.cayman.max_tile_pipes;
 		else if (rdev->family >= CHIP_CEDAR)
-			value = rdev->config.evergreen.max_tile_pipes;
+			*value = rdev->config.evergreen.max_tile_pipes;
 		else if (rdev->family >= CHIP_RV770)
-			value = rdev->config.rv770.max_tile_pipes;
+			*value = rdev->config.rv770.max_tile_pipes;
 		else if (rdev->family >= CHIP_R600)
-			value = rdev->config.r600.max_tile_pipes;
+			*value = rdev->config.r600.max_tile_pipes;
 		else {
 			return -EINVAL;
 		}
 		break;
 	case RADEON_INFO_FUSION_GART_WORKING:
-		value = 1;
+		*value = 1;
 		break;
 	case RADEON_INFO_BACKEND_MAP:
 		if (rdev->family >= CHIP_TAHITI)
-			value = rdev->config.si.backend_map;
+			*value = rdev->config.si.backend_map;
 		else if (rdev->family >= CHIP_CAYMAN)
-			value = rdev->config.cayman.backend_map;
+			*value = rdev->config.cayman.backend_map;
 		else if (rdev->family >= CHIP_CEDAR)
-			value = rdev->config.evergreen.backend_map;
+			*value = rdev->config.evergreen.backend_map;
 		else if (rdev->family >= CHIP_RV770)
-			value = rdev->config.rv770.backend_map;
+			*value = rdev->config.rv770.backend_map;
 		else if (rdev->family >= CHIP_R600)
-			value = rdev->config.r600.backend_map;
+			*value = rdev->config.r600.backend_map;
 		else {
 			return -EINVAL;
 		}
@@ -337,50 +334,91 @@ int radeon_info_ioctl(struct drm_device *dev, void *data, struct drm_file *filp)
 		/* this is where we report if vm is supported or not */
 		if (rdev->family < CHIP_CAYMAN)
 			return -EINVAL;
-		value = RADEON_VA_RESERVED_SIZE;
+		*value = RADEON_VA_RESERVED_SIZE;
 		break;
 	case RADEON_INFO_IB_VM_MAX_SIZE:
 		/* this is where we report if vm is supported or not */
 		if (rdev->family < CHIP_CAYMAN)
 			return -EINVAL;
-		value = RADEON_IB_VM_MAX_SIZE;
+		*value = RADEON_IB_VM_MAX_SIZE;
 		break;
 	case RADEON_INFO_MAX_PIPES:
 		if (rdev->family >= CHIP_TAHITI)
-			value = rdev->config.si.max_cu_per_sh;
+			*value = rdev->config.si.max_cu_per_sh;
 		else if (rdev->family >= CHIP_CAYMAN)
-			value = rdev->config.cayman.max_pipes_per_simd;
+			*value = rdev->config.cayman.max_pipes_per_simd;
 		else if (rdev->family >= CHIP_CEDAR)
-			value = rdev->config.evergreen.max_pipes;
+			*value = rdev->config.evergreen.max_pipes;
 		else if (rdev->family >= CHIP_RV770)
-			value = rdev->config.rv770.max_pipes;
+			*value = rdev->config.rv770.max_pipes;
 		else if (rdev->family >= CHIP_R600)
-			value = rdev->config.r600.max_pipes;
+			*value = rdev->config.r600.max_pipes;
 		else {
 			return -EINVAL;
 		}
 		break;
+	case RADEON_INFO_TIMESTAMP:
+		if (rdev->family < CHIP_R600) {
+			DRM_DEBUG_KMS("timestamp is r6xx+ only!\n");
+			return -EINVAL;
+		}
+		value = (uint32_t*)&value64;
+		value_size = sizeof(uint64_t);
+		value64 = radeon_get_gpu_clock_counter(rdev);
+		break;
 	case RADEON_INFO_MAX_SE:
 		if (rdev->family >= CHIP_TAHITI)
-			value = rdev->config.si.max_shader_engines;
+			*value = rdev->config.si.max_shader_engines;
 		else if (rdev->family >= CHIP_CAYMAN)
-			value = rdev->config.cayman.max_shader_engines;
+			*value = rdev->config.cayman.max_shader_engines;
 		else if (rdev->family >= CHIP_CEDAR)
-			value = rdev->config.evergreen.num_ses;
+			*value = rdev->config.evergreen.num_ses;
 		else
-			value = 1;
+			*value = 1;
 		break;
 	case RADEON_INFO_MAX_SH_PER_SE:
 		if (rdev->family >= CHIP_TAHITI)
-			value = rdev->config.si.max_sh_per_se;
+			*value = rdev->config.si.max_sh_per_se;
 		else
 			return -EINVAL;
 		break;
+	case RADEON_INFO_FASTFB_WORKING:
+		*value = rdev->fastfb_working;
+		break;
+	case RADEON_INFO_RING_WORKING:
+		if (DRM_COPY_FROM_USER(value, value_ptr, sizeof(uint32_t))) {
+			DRM_ERROR("copy_from_user %s:%u\n", __func__, __LINE__);
+			return -EFAULT;
+		}
+		switch (*value) {
+		case RADEON_CS_RING_GFX:
+		case RADEON_CS_RING_COMPUTE:
+			*value = rdev->ring[RADEON_RING_TYPE_GFX_INDEX].ready;
+			break;
+		case RADEON_CS_RING_DMA:
+			*value = rdev->ring[R600_RING_TYPE_DMA_INDEX].ready;
+			*value |= rdev->ring[CAYMAN_RING_TYPE_DMA1_INDEX].ready;
+			break;
+		case RADEON_CS_RING_UVD:
+			*value = rdev->ring[R600_RING_TYPE_UVD_INDEX].ready;
+			break;
+		default:
+			return -EINVAL;
+		}
+		break;
+	case RADEON_INFO_SI_TILE_MODE_ARRAY:
+		if (rdev->family < CHIP_TAHITI) {
+			DRM_DEBUG_KMS("tile mode array is si only!\n");
+			return -EINVAL;
+		}
+		value = rdev->config.si.tile_mode_array;
+		value_size = sizeof(uint32_t)*32;
+		break;
 	default:
 		DRM_DEBUG_KMS("Invalid request %d\n", info->request);
 		return -EINVAL;
 	}
-	if (DRM_COPY_TO_USER(value_ptr, &value, sizeof(uint32_t))) {
+	if (DRM_COPY_TO_USER(value_ptr, (char*)value, value_size)) {
 		DRM_ERROR("copy_to_user %s:%u\n", __func__, __LINE__);
 		return -EFAULT;
 	}
@@ -513,6 +551,7 @@ void radeon_driver_preclose_kms(struct drm_device *dev,
 		rdev->hyperz_filp = NULL;
 	if (rdev->cmask_filp == file_priv)
 		rdev->cmask_filp = NULL;
+	radeon_uvd_free_handles(rdev, file_priv);
 }
 
 /*
diff --git a/drivers/gpu/drm/radeon/radeon_mode.h b/drivers/gpu/drm/radeon/radeon_mode.h
index 4003f5a68c09..44e579e75fd0 100644
--- a/drivers/gpu/drm/radeon/radeon_mode.h
+++ b/drivers/gpu/drm/radeon/radeon_mode.h
@@ -492,6 +492,29 @@ struct radeon_framebuffer {
 #define ENCODER_MODE_IS_DP(em) (((em) == ATOM_ENCODER_MODE_DP) || \
 				((em) == ATOM_ENCODER_MODE_DP_MST))
 
+struct atom_clock_dividers {
+	u32 post_div;
+	union {
+		struct {
+#ifdef __BIG_ENDIAN
+			u32 reserved : 6;
+			u32 whole_fb_div : 12;
+			u32 frac_fb_div : 14;
+#else
+			u32 frac_fb_div : 14;
+			u32 whole_fb_div : 12;
+			u32 reserved : 6;
+#endif
+		};
+		u32 fb_div;
+	};
+	u32 ref_div;
+	bool enable_post_div;
+	bool enable_dithen;
+	u32 vco_mode;
+	u32 real_clock;
+};
+
 extern enum radeon_tv_std
 radeon_combios_get_tv_info(struct radeon_device *rdev);
 extern enum radeon_tv_std
diff --git a/drivers/gpu/drm/radeon/radeon_object.c b/drivers/gpu/drm/radeon/radeon_object.c
index d3aface2d12d..1424ccde2377 100644
--- a/drivers/gpu/drm/radeon/radeon_object.c
+++ b/drivers/gpu/drm/radeon/radeon_object.c
@@ -321,8 +321,10 @@ void radeon_bo_force_delete(struct radeon_device *rdev)
 int radeon_bo_init(struct radeon_device *rdev)
 {
 	/* Add an MTRR for the VRAM */
-	rdev->mc.vram_mtrr = mtrr_add(rdev->mc.aper_base, rdev->mc.aper_size,
+	if (!rdev->fastfb_working) {
+		rdev->mc.vram_mtrr = mtrr_add(rdev->mc.aper_base, rdev->mc.aper_size,
 			MTRR_TYPE_WRCOMB, 1);
+	}
 	DRM_INFO("Detected VRAM RAM=%lluM, BAR=%lluM\n",
 		rdev->mc.mc_vram_size >> 20,
 		(unsigned long long)rdev->mc.aper_size >> 20);
@@ -339,14 +341,14 @@ void radeon_bo_fini(struct radeon_device *rdev)
 void radeon_bo_list_add_object(struct radeon_bo_list *lobj,
 				struct list_head *head)
 {
-	if (lobj->wdomain) {
+	if (lobj->written) {
 		list_add(&lobj->tv.head, head);
 	} else {
 		list_add_tail(&lobj->tv.head, head);
 	}
 }
 
-int radeon_bo_list_validate(struct list_head *head)
+int radeon_bo_list_validate(struct list_head *head, int ring)
 {
 	struct radeon_bo_list *lobj;
 	struct radeon_bo *bo;
@@ -360,15 +362,17 @@ int radeon_bo_list_validate(struct list_head *head)
 	list_for_each_entry(lobj, head, tv.head) {
 		bo = lobj->bo;
 		if (!bo->pin_count) {
-			domain = lobj->wdomain ? lobj->wdomain : lobj->rdomain;
+			domain = lobj->domain;
 			
 		retry:
 			radeon_ttm_placement_from_domain(bo, domain);
+			if (ring == R600_RING_TYPE_UVD_INDEX)
+				radeon_uvd_force_into_uvd_segment(bo);
 			r = ttm_bo_validate(&bo->tbo, &bo->placement,
 						true, false);
 			if (unlikely(r)) {
-				if (r != -ERESTARTSYS && domain == RADEON_GEM_DOMAIN_VRAM) {
-					domain |= RADEON_GEM_DOMAIN_GTT;
+				if (r != -ERESTARTSYS && domain != lobj->alt_domain) {
+					domain = lobj->alt_domain;
 					goto retry;
 				}
 				return r;
diff --git a/drivers/gpu/drm/radeon/radeon_object.h b/drivers/gpu/drm/radeon/radeon_object.h
index 5fc86b03043b..e2cb80a96b51 100644
--- a/drivers/gpu/drm/radeon/radeon_object.h
+++ b/drivers/gpu/drm/radeon/radeon_object.h
@@ -128,7 +128,7 @@ extern int radeon_bo_init(struct radeon_device *rdev);
 extern void radeon_bo_fini(struct radeon_device *rdev);
 extern void radeon_bo_list_add_object(struct radeon_bo_list *lobj,
 				struct list_head *head);
-extern int radeon_bo_list_validate(struct list_head *head);
+extern int radeon_bo_list_validate(struct list_head *head, int ring);
 extern int radeon_bo_fbdev_mmap(struct radeon_bo *bo,
 				struct vm_area_struct *vma);
 extern int radeon_bo_set_tiling_flags(struct radeon_bo *bo,
diff --git a/drivers/gpu/drm/radeon/radeon_pm.c b/drivers/gpu/drm/radeon/radeon_pm.c
index 338fd6a74e87..788c64cb4b47 100644
--- a/drivers/gpu/drm/radeon/radeon_pm.c
+++ b/drivers/gpu/drm/radeon/radeon_pm.c
@@ -843,7 +843,11 @@ static int radeon_debugfs_pm_info(struct seq_file *m, void *data)
 	struct radeon_device *rdev = dev->dev_private;
 
 	seq_printf(m, "default engine clock: %u0 kHz\n", rdev->pm.default_sclk);
-	seq_printf(m, "current engine clock: %u0 kHz\n", radeon_get_engine_clock(rdev));
+	/* radeon_get_engine_clock is not reliable on APUs so just print the current clock */
+	if ((rdev->family >= CHIP_PALM) && (rdev->flags & RADEON_IS_IGP))
+		seq_printf(m, "current engine clock: %u0 kHz\n", rdev->pm.current_sclk);
+	else
+		seq_printf(m, "current engine clock: %u0 kHz\n", radeon_get_engine_clock(rdev));
 	seq_printf(m, "default memory clock: %u0 kHz\n", rdev->pm.default_mclk);
 	if (rdev->asic->pm.get_memory_clock)
 		seq_printf(m, "current memory clock: %u0 kHz\n", radeon_get_memory_clock(rdev));
diff --git a/drivers/gpu/drm/radeon/radeon_ring.c b/drivers/gpu/drm/radeon/radeon_ring.c
index 8d58e268ff6d..e17faa7cf732 100644
--- a/drivers/gpu/drm/radeon/radeon_ring.c
+++ b/drivers/gpu/drm/radeon/radeon_ring.c
@@ -180,7 +180,8 @@ int radeon_ib_schedule(struct radeon_device *rdev, struct radeon_ib *ib,
 		radeon_semaphore_free(rdev, &ib->semaphore, NULL);
 	}
 	/* if we can't remember our last VM flush then flush now! */
-	if (ib->vm && !ib->vm->last_flush) {
+	/* XXX figure out why we have to flush for every IB */
+	if (ib->vm /*&& !ib->vm->last_flush*/) {
 		radeon_ring_vm_flush(rdev, ib->ring, ib->vm);
 	}
 	if (const_ib) {
@@ -368,7 +369,7 @@ void radeon_ring_free_size(struct radeon_device *rdev, struct radeon_ring *ring)
 {
 	u32 rptr;
 
-	if (rdev->wb.enabled)
+	if (rdev->wb.enabled && ring != &rdev->ring[R600_RING_TYPE_UVD_INDEX])
 		rptr = le32_to_cpu(rdev->wb.wb[ring->rptr_offs/4]);
 	else
 		rptr = RREG32(ring->rptr_reg);
@@ -821,18 +822,20 @@ static int radeon_debugfs_ring_info(struct seq_file *m, void *data)
 	return 0;
 }
 
-static int radeon_ring_type_gfx_index = RADEON_RING_TYPE_GFX_INDEX;
-static int cayman_ring_type_cp1_index = CAYMAN_RING_TYPE_CP1_INDEX;
-static int cayman_ring_type_cp2_index = CAYMAN_RING_TYPE_CP2_INDEX;
-static int radeon_ring_type_dma1_index = R600_RING_TYPE_DMA_INDEX;
-static int radeon_ring_type_dma2_index = CAYMAN_RING_TYPE_DMA1_INDEX;
+static int radeon_gfx_index = RADEON_RING_TYPE_GFX_INDEX;
+static int cayman_cp1_index = CAYMAN_RING_TYPE_CP1_INDEX;
+static int cayman_cp2_index = CAYMAN_RING_TYPE_CP2_INDEX;
+static int radeon_dma1_index = R600_RING_TYPE_DMA_INDEX;
+static int radeon_dma2_index = CAYMAN_RING_TYPE_DMA1_INDEX;
+static int r600_uvd_index = R600_RING_TYPE_UVD_INDEX;
 
 static struct drm_info_list radeon_debugfs_ring_info_list[] = {
-	{"radeon_ring_gfx", radeon_debugfs_ring_info, 0, &radeon_ring_type_gfx_index},
-	{"radeon_ring_cp1", radeon_debugfs_ring_info, 0, &cayman_ring_type_cp1_index},
-	{"radeon_ring_cp2", radeon_debugfs_ring_info, 0, &cayman_ring_type_cp2_index},
-	{"radeon_ring_dma1", radeon_debugfs_ring_info, 0, &radeon_ring_type_dma1_index},
-	{"radeon_ring_dma2", radeon_debugfs_ring_info, 0, &radeon_ring_type_dma2_index},
+	{"radeon_ring_gfx", radeon_debugfs_ring_info, 0, &radeon_gfx_index},
+	{"radeon_ring_cp1", radeon_debugfs_ring_info, 0, &cayman_cp1_index},
+	{"radeon_ring_cp2", radeon_debugfs_ring_info, 0, &cayman_cp2_index},
+	{"radeon_ring_dma1", radeon_debugfs_ring_info, 0, &radeon_dma1_index},
+	{"radeon_ring_dma2", radeon_debugfs_ring_info, 0, &radeon_dma2_index},
+	{"radeon_ring_uvd", radeon_debugfs_ring_info, 0, &r600_uvd_index},
 };
 
 static int radeon_debugfs_sa_info(struct seq_file *m, void *data)
diff --git a/drivers/gpu/drm/radeon/radeon_sa.c b/drivers/gpu/drm/radeon/radeon_sa.c
index cb800995d4f9..0abe5a9431bb 100644
--- a/drivers/gpu/drm/radeon/radeon_sa.c
+++ b/drivers/gpu/drm/radeon/radeon_sa.c
@@ -64,7 +64,7 @@ int radeon_sa_bo_manager_init(struct radeon_device *rdev,
 	}
 
 	r = radeon_bo_create(rdev, size, RADEON_GPU_PAGE_SIZE, true,
-			     RADEON_GEM_DOMAIN_CPU, NULL, &sa_manager->bo);
+			     domain, NULL, &sa_manager->bo);
 	if (r) {
 		dev_err(rdev->dev, "(%d) failed to allocate bo for manager\n", r);
 		return r;
diff --git a/drivers/gpu/drm/radeon/radeon_test.c b/drivers/gpu/drm/radeon/radeon_test.c
index fda09c9ea689..bbed4af8d0bc 100644
--- a/drivers/gpu/drm/radeon/radeon_test.c
+++ b/drivers/gpu/drm/radeon/radeon_test.c
@@ -252,6 +252,36 @@ void radeon_test_moves(struct radeon_device *rdev)
 		radeon_do_test_moves(rdev, RADEON_TEST_COPY_BLIT);
 }
 
+static int radeon_test_create_and_emit_fence(struct radeon_device *rdev,
+					     struct radeon_ring *ring,
+					     struct radeon_fence **fence)
+{
+	int r;
+
+	if (ring->idx == R600_RING_TYPE_UVD_INDEX) {
+		r = radeon_uvd_get_create_msg(rdev, ring->idx, 1, NULL);
+		if (r) {
+			DRM_ERROR("Failed to get dummy create msg\n");
+			return r;
+		}
+
+		r = radeon_uvd_get_destroy_msg(rdev, ring->idx, 1, fence);
+		if (r) {
+			DRM_ERROR("Failed to get dummy destroy msg\n");
+			return r;
+		}
+	} else {
+		r = radeon_ring_lock(rdev, ring, 64);
+		if (r) {
+			DRM_ERROR("Failed to lock ring A %d\n", ring->idx);
+			return r;
+		}
+		radeon_fence_emit(rdev, fence, ring->idx);
+		radeon_ring_unlock_commit(rdev, ring);
+	}
+	return 0;
+}
+
 void radeon_test_ring_sync(struct radeon_device *rdev,
 			   struct radeon_ring *ringA,
 			   struct radeon_ring *ringB)
@@ -272,21 +302,24 @@ void radeon_test_ring_sync(struct radeon_device *rdev,
 		goto out_cleanup;
 	}
 	radeon_semaphore_emit_wait(rdev, ringA->idx, semaphore);
-	r = radeon_fence_emit(rdev, &fence1, ringA->idx);
-	if (r) {
-		DRM_ERROR("Failed to emit fence 1\n");
-		radeon_ring_unlock_undo(rdev, ringA);
+	radeon_ring_unlock_commit(rdev, ringA);
+
+	r = radeon_test_create_and_emit_fence(rdev, ringA, &fence1);
+	if (r)
 		goto out_cleanup;
-	}
-	radeon_semaphore_emit_wait(rdev, ringA->idx, semaphore);
-	r = radeon_fence_emit(rdev, &fence2, ringA->idx);
+
+	r = radeon_ring_lock(rdev, ringA, 64);
 	if (r) {
-		DRM_ERROR("Failed to emit fence 2\n");
-		radeon_ring_unlock_undo(rdev, ringA);
+		DRM_ERROR("Failed to lock ring A %d\n", ringA->idx);
 		goto out_cleanup;
 	}
+	radeon_semaphore_emit_wait(rdev, ringA->idx, semaphore);
 	radeon_ring_unlock_commit(rdev, ringA);
 
+	r = radeon_test_create_and_emit_fence(rdev, ringA, &fence2);
+	if (r)
+		goto out_cleanup;
+
 	mdelay(1000);
 
 	if (radeon_fence_signaled(fence1)) {
@@ -364,27 +397,22 @@ static void radeon_test_ring_sync2(struct radeon_device *rdev,
 		goto out_cleanup;
 	}
 	radeon_semaphore_emit_wait(rdev, ringA->idx, semaphore);
-	r = radeon_fence_emit(rdev, &fenceA, ringA->idx);
-	if (r) {
-		DRM_ERROR("Failed to emit sync fence 1\n");
-		radeon_ring_unlock_undo(rdev, ringA);
-		goto out_cleanup;
-	}
 	radeon_ring_unlock_commit(rdev, ringA);
 
+	r = radeon_test_create_and_emit_fence(rdev, ringA, &fenceA);
+	if (r)
+		goto out_cleanup;
+
 	r = radeon_ring_lock(rdev, ringB, 64);
 	if (r) {
 		DRM_ERROR("Failed to lock ring B %d\n", ringB->idx);
 		goto out_cleanup;
 	}
 	radeon_semaphore_emit_wait(rdev, ringB->idx, semaphore);
-	r = radeon_fence_emit(rdev, &fenceB, ringB->idx);
-	if (r) {
-		DRM_ERROR("Failed to create sync fence 2\n");
-		radeon_ring_unlock_undo(rdev, ringB);
-		goto out_cleanup;
-	}
 	radeon_ring_unlock_commit(rdev, ringB);
+	r = radeon_test_create_and_emit_fence(rdev, ringB, &fenceB);
+	if (r)
+		goto out_cleanup;
 
 	mdelay(1000);
 
@@ -393,7 +421,7 @@ static void radeon_test_ring_sync2(struct radeon_device *rdev,
 		goto out_cleanup;
 	}
 	if (radeon_fence_signaled(fenceB)) {
-		DRM_ERROR("Fence A signaled without waiting for semaphore.\n");
+		DRM_ERROR("Fence B signaled without waiting for semaphore.\n");
 		goto out_cleanup;
 	}
 
diff --git a/drivers/gpu/drm/radeon/radeon_uvd.c b/drivers/gpu/drm/radeon/radeon_uvd.c
new file mode 100644
index 000000000000..906e5c0ca3b9
--- /dev/null
+++ b/drivers/gpu/drm/radeon/radeon_uvd.c
@@ -0,0 +1,831 @@
+/*
+ * Copyright 2011 Advanced Micro Devices, Inc.
+ * All Rights Reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the
+ * "Software"), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sub license, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDERS, AUTHORS AND/OR ITS SUPPLIERS BE LIABLE FOR ANY CLAIM,
+ * DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR
+ * OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE
+ * USE OR OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * The above copyright notice and this permission notice (including the
+ * next paragraph) shall be included in all copies or substantial portions
+ * of the Software.
+ *
+ */
+/*
+ * Authors:
+ *    Christian König <deathsimple@vodafone.de>
+ */
+
+#include <linux/firmware.h>
+#include <linux/module.h>
+#include <drm/drmP.h>
+#include <drm/drm.h>
+
+#include "radeon.h"
+#include "r600d.h"
+
+/* 1 second timeout */
+#define UVD_IDLE_TIMEOUT_MS	1000
+
+/* Firmware Names */
+#define FIRMWARE_RV710		"radeon/RV710_uvd.bin"
+#define FIRMWARE_CYPRESS	"radeon/CYPRESS_uvd.bin"
+#define FIRMWARE_SUMO		"radeon/SUMO_uvd.bin"
+#define FIRMWARE_TAHITI		"radeon/TAHITI_uvd.bin"
+
+MODULE_FIRMWARE(FIRMWARE_RV710);
+MODULE_FIRMWARE(FIRMWARE_CYPRESS);
+MODULE_FIRMWARE(FIRMWARE_SUMO);
+MODULE_FIRMWARE(FIRMWARE_TAHITI);
+
+static void radeon_uvd_idle_work_handler(struct work_struct *work);
+
+int radeon_uvd_init(struct radeon_device *rdev)
+{
+	struct platform_device *pdev;
+	unsigned long bo_size;
+	const char *fw_name;
+	int i, r;
+
+	INIT_DELAYED_WORK(&rdev->uvd.idle_work, radeon_uvd_idle_work_handler);
+
+	pdev = platform_device_register_simple("radeon_uvd", 0, NULL, 0);
+	r = IS_ERR(pdev);
+	if (r) {
+		dev_err(rdev->dev, "radeon_uvd: Failed to register firmware\n");
+		return -EINVAL;
+	}
+
+	switch (rdev->family) {
+	case CHIP_RV710:
+	case CHIP_RV730:
+	case CHIP_RV740:
+		fw_name = FIRMWARE_RV710;
+		break;
+
+	case CHIP_CYPRESS:
+	case CHIP_HEMLOCK:
+	case CHIP_JUNIPER:
+	case CHIP_REDWOOD:
+	case CHIP_CEDAR:
+		fw_name = FIRMWARE_CYPRESS;
+		break;
+
+	case CHIP_SUMO:
+	case CHIP_SUMO2:
+	case CHIP_PALM:
+	case CHIP_CAYMAN:
+	case CHIP_BARTS:
+	case CHIP_TURKS:
+	case CHIP_CAICOS:
+		fw_name = FIRMWARE_SUMO;
+		break;
+
+	case CHIP_TAHITI:
+	case CHIP_VERDE:
+	case CHIP_PITCAIRN:
+	case CHIP_ARUBA:
+		fw_name = FIRMWARE_TAHITI;
+		break;
+
+	default:
+		return -EINVAL;
+	}
+
+	r = request_firmware(&rdev->uvd_fw, fw_name, &pdev->dev);
+	if (r) {
+		dev_err(rdev->dev, "radeon_uvd: Can't load firmware \"%s\"\n",
+			fw_name);
+		platform_device_unregister(pdev);
+		return r;
+	}
+
+	platform_device_unregister(pdev);
+
+	bo_size = RADEON_GPU_PAGE_ALIGN(rdev->uvd_fw->size + 8) +
+		  RADEON_UVD_STACK_SIZE + RADEON_UVD_HEAP_SIZE;
+	r = radeon_bo_create(rdev, bo_size, PAGE_SIZE, true,
+			     RADEON_GEM_DOMAIN_VRAM, NULL, &rdev->uvd.vcpu_bo);
+	if (r) {
+		dev_err(rdev->dev, "(%d) failed to allocate UVD bo\n", r);
+		return r;
+	}
+
+	r = radeon_uvd_resume(rdev);
+	if (r)
+		return r;
+
+	memset(rdev->uvd.cpu_addr, 0, bo_size);
+	memcpy(rdev->uvd.cpu_addr, rdev->uvd_fw->data, rdev->uvd_fw->size);
+
+	r = radeon_uvd_suspend(rdev);
+	if (r)
+		return r;
+
+	for (i = 0; i < RADEON_MAX_UVD_HANDLES; ++i) {
+		atomic_set(&rdev->uvd.handles[i], 0);
+		rdev->uvd.filp[i] = NULL;
+	}
+
+	return 0;
+}
+
+void radeon_uvd_fini(struct radeon_device *rdev)
+{
+	radeon_uvd_suspend(rdev);
+	radeon_bo_unref(&rdev->uvd.vcpu_bo);
+}
+
+int radeon_uvd_suspend(struct radeon_device *rdev)
+{
+	int r;
+
+	if (rdev->uvd.vcpu_bo == NULL)
+		return 0;
+
+	r = radeon_bo_reserve(rdev->uvd.vcpu_bo, false);
+	if (!r) {
+		radeon_bo_kunmap(rdev->uvd.vcpu_bo);
+		radeon_bo_unpin(rdev->uvd.vcpu_bo);
+		radeon_bo_unreserve(rdev->uvd.vcpu_bo);
+	}
+	return r;
+}
+
+int radeon_uvd_resume(struct radeon_device *rdev)
+{
+	int r;
+
+	if (rdev->uvd.vcpu_bo == NULL)
+		return -EINVAL;
+
+	r = radeon_bo_reserve(rdev->uvd.vcpu_bo, false);
+	if (r) {
+		radeon_bo_unref(&rdev->uvd.vcpu_bo);
+		dev_err(rdev->dev, "(%d) failed to reserve UVD bo\n", r);
+		return r;
+	}
+
+	r = radeon_bo_pin(rdev->uvd.vcpu_bo, RADEON_GEM_DOMAIN_VRAM,
+			  &rdev->uvd.gpu_addr);
+	if (r) {
+		radeon_bo_unreserve(rdev->uvd.vcpu_bo);
+		radeon_bo_unref(&rdev->uvd.vcpu_bo);
+		dev_err(rdev->dev, "(%d) UVD bo pin failed\n", r);
+		return r;
+	}
+
+	r = radeon_bo_kmap(rdev->uvd.vcpu_bo, &rdev->uvd.cpu_addr);
+	if (r) {
+		dev_err(rdev->dev, "(%d) UVD map failed\n", r);
+		return r;
+	}
+
+	radeon_bo_unreserve(rdev->uvd.vcpu_bo);
+
+	return 0;
+}
+
+void radeon_uvd_force_into_uvd_segment(struct radeon_bo *rbo)
+{
+	rbo->placement.fpfn = 0 >> PAGE_SHIFT;
+	rbo->placement.lpfn = (256 * 1024 * 1024) >> PAGE_SHIFT;
+}
+
+void radeon_uvd_free_handles(struct radeon_device *rdev, struct drm_file *filp)
+{
+	int i, r;
+	for (i = 0; i < RADEON_MAX_UVD_HANDLES; ++i) {
+		if (rdev->uvd.filp[i] == filp) {
+			uint32_t handle = atomic_read(&rdev->uvd.handles[i]);
+			struct radeon_fence *fence;
+
+			r = radeon_uvd_get_destroy_msg(rdev,
+				R600_RING_TYPE_UVD_INDEX, handle, &fence);
+			if (r) {
+				DRM_ERROR("Error destroying UVD (%d)!\n", r);
+				continue;
+			}
+
+			radeon_fence_wait(fence, false);
+			radeon_fence_unref(&fence);
+
+			rdev->uvd.filp[i] = NULL;
+			atomic_set(&rdev->uvd.handles[i], 0);
+		}
+	}
+}
+
+static int radeon_uvd_cs_msg_decode(uint32_t *msg, unsigned buf_sizes[])
+{
+	unsigned stream_type = msg[4];
+	unsigned width = msg[6];
+	unsigned height = msg[7];
+	unsigned dpb_size = msg[9];
+	unsigned pitch = msg[28];
+
+	unsigned width_in_mb = width / 16;
+	unsigned height_in_mb = ALIGN(height / 16, 2);
+
+	unsigned image_size, tmp, min_dpb_size;
+
+	image_size = width * height;
+	image_size += image_size / 2;
+	image_size = ALIGN(image_size, 1024);
+
+	switch (stream_type) {
+	case 0: /* H264 */
+
+		/* reference picture buffer */
+		min_dpb_size = image_size * 17;
+
+		/* macroblock context buffer */
+		min_dpb_size += width_in_mb * height_in_mb * 17 * 192;
+
+		/* IT surface buffer */
+		min_dpb_size += width_in_mb * height_in_mb * 32;
+		break;
+
+	case 1: /* VC1 */
+
+		/* reference picture buffer */
+		min_dpb_size = image_size * 3;
+
+		/* CONTEXT_BUFFER */
+		min_dpb_size += width_in_mb * height_in_mb * 128;
+
+		/* IT surface buffer */
+		min_dpb_size += width_in_mb * 64;
+
+		/* DB surface buffer */
+		min_dpb_size += width_in_mb * 128;
+
+		/* BP */
+		tmp = max(width_in_mb, height_in_mb);
+		min_dpb_size += ALIGN(tmp * 7 * 16, 64);
+		break;
+
+	case 3: /* MPEG2 */
+
+		/* reference picture buffer */
+		min_dpb_size = image_size * 3;
+		break;
+
+	case 4: /* MPEG4 */
+
+		/* reference picture buffer */
+		min_dpb_size = image_size * 3;
+
+		/* CM */
+		min_dpb_size += width_in_mb * height_in_mb * 64;
+
+		/* IT surface buffer */
+		min_dpb_size += ALIGN(width_in_mb * height_in_mb * 32, 64);
+		break;
+
+	default:
+		DRM_ERROR("UVD codec not handled %d!\n", stream_type);
+		return -EINVAL;
+	}
+
+	if (width > pitch) {
+		DRM_ERROR("Invalid UVD decoding target pitch!\n");
+		return -EINVAL;
+	}
+
+	if (dpb_size < min_dpb_size) {
+		DRM_ERROR("Invalid dpb_size in UVD message (%d / %d)!\n",
+			  dpb_size, min_dpb_size);
+		return -EINVAL;
+	}
+
+	buf_sizes[0x1] = dpb_size;
+	buf_sizes[0x2] = image_size;
+	return 0;
+}
+
+static int radeon_uvd_cs_msg(struct radeon_cs_parser *p, struct radeon_bo *bo,
+			     unsigned offset, unsigned buf_sizes[])
+{
+	int32_t *msg, msg_type, handle;
+	void *ptr;
+
+	int i, r;
+
+	if (offset & 0x3F) {
+		DRM_ERROR("UVD messages must be 64 byte aligned!\n");
+		return -EINVAL;
+	}
+
+	r = radeon_bo_kmap(bo, &ptr);
+	if (r)
+		return r;
+
+	msg = ptr + offset;
+
+	msg_type = msg[1];
+	handle = msg[2];
+
+	if (handle == 0) {
+		DRM_ERROR("Invalid UVD handle!\n");
+		return -EINVAL;
+	}
+
+	if (msg_type == 1) {
+		/* it's a decode msg, calc buffer sizes */
+		r = radeon_uvd_cs_msg_decode(msg, buf_sizes);
+		radeon_bo_kunmap(bo);
+		if (r)
+			return r;
+
+	} else if (msg_type == 2) {
+		/* it's a destroy msg, free the handle */
+		for (i = 0; i < RADEON_MAX_UVD_HANDLES; ++i)
+			atomic_cmpxchg(&p->rdev->uvd.handles[i], handle, 0);
+		radeon_bo_kunmap(bo);
+		return 0;
+	} else {
+		/* it's a create msg, no special handling needed */
+		radeon_bo_kunmap(bo);
+	}
+
+	/* create or decode, validate the handle */
+	for (i = 0; i < RADEON_MAX_UVD_HANDLES; ++i) {
+		if (atomic_read(&p->rdev->uvd.handles[i]) == handle)
+			return 0;
+	}
+
+	/* handle not found try to alloc a new one */
+	for (i = 0; i < RADEON_MAX_UVD_HANDLES; ++i) {
+		if (!atomic_cmpxchg(&p->rdev->uvd.handles[i], 0, handle)) {
+			p->rdev->uvd.filp[i] = p->filp;
+			return 0;
+		}
+	}
+
+	DRM_ERROR("No more free UVD handles!\n");
+	return -EINVAL;
+}
+
+static int radeon_uvd_cs_reloc(struct radeon_cs_parser *p,
+			       int data0, int data1,
+			       unsigned buf_sizes[])
+{
+	struct radeon_cs_chunk *relocs_chunk;
+	struct radeon_cs_reloc *reloc;
+	unsigned idx, cmd, offset;
+	uint64_t start, end;
+	int r;
+
+	relocs_chunk = &p->chunks[p->chunk_relocs_idx];
+	offset = radeon_get_ib_value(p, data0);
+	idx = radeon_get_ib_value(p, data1);
+	if (idx >= relocs_chunk->length_dw) {
+		DRM_ERROR("Relocs at %d after relocations chunk end %d !\n",
+			  idx, relocs_chunk->length_dw);
+		return -EINVAL;
+	}
+
+	reloc = p->relocs_ptr[(idx / 4)];
+	start = reloc->lobj.gpu_offset;
+	end = start + radeon_bo_size(reloc->robj);
+	start += offset;
+
+	p->ib.ptr[data0] = start & 0xFFFFFFFF;
+	p->ib.ptr[data1] = start >> 32;
+
+	cmd = radeon_get_ib_value(p, p->idx) >> 1;
+
+	if (cmd < 0x4) {
+		if ((end - start) < buf_sizes[cmd]) {
+			DRM_ERROR("buffer to small (%d / %d)!\n",
+				  (unsigned)(end - start), buf_sizes[cmd]);
+			return -EINVAL;
+		}
+
+	} else if (cmd != 0x100) {
+		DRM_ERROR("invalid UVD command %X!\n", cmd);
+		return -EINVAL;
+	}
+
+	if ((start >> 28) != (end >> 28)) {
+		DRM_ERROR("reloc %LX-%LX crossing 256MB boundary!\n",
+			  start, end);
+		return -EINVAL;
+	}
+
+	/* TODO: is this still necessary on NI+ ? */
+	if ((cmd == 0 || cmd == 0x3) &&
+	    (start >> 28) != (p->rdev->uvd.gpu_addr >> 28)) {
+		DRM_ERROR("msg/fb buffer %LX-%LX out of 256MB segment!\n",
+			  start, end);
+		return -EINVAL;
+	}
+
+	if (cmd == 0) {
+		r = radeon_uvd_cs_msg(p, reloc->robj, offset, buf_sizes);
+		if (r)
+			return r;
+	}
+
+	return 0;
+}
+
+static int radeon_uvd_cs_reg(struct radeon_cs_parser *p,
+			     struct radeon_cs_packet *pkt,
+			     int *data0, int *data1,
+			     unsigned buf_sizes[])
+{
+	int i, r;
+
+	p->idx++;
+	for (i = 0; i <= pkt->count; ++i) {
+		switch (pkt->reg + i*4) {
+		case UVD_GPCOM_VCPU_DATA0:
+			*data0 = p->idx;
+			break;
+		case UVD_GPCOM_VCPU_DATA1:
+			*data1 = p->idx;
+			break;
+		case UVD_GPCOM_VCPU_CMD:
+			r = radeon_uvd_cs_reloc(p, *data0, *data1, buf_sizes);
+			if (r)
+				return r;
+			break;
+		case UVD_ENGINE_CNTL:
+			break;
+		default:
+			DRM_ERROR("Invalid reg 0x%X!\n",
+				  pkt->reg + i*4);
+			return -EINVAL;
+		}
+		p->idx++;
+	}
+	return 0;
+}
+
+int radeon_uvd_cs_parse(struct radeon_cs_parser *p)
+{
+	struct radeon_cs_packet pkt;
+	int r, data0 = 0, data1 = 0;
+
+	/* minimum buffer sizes */
+	unsigned buf_sizes[] = {
+		[0x00000000]	=	2048,
+		[0x00000001]	=	32 * 1024 * 1024,
+		[0x00000002]	=	2048 * 1152 * 3,
+		[0x00000003]	=	2048,
+	};
+
+	if (p->chunks[p->chunk_ib_idx].length_dw % 16) {
+		DRM_ERROR("UVD IB length (%d) not 16 dwords aligned!\n",
+			  p->chunks[p->chunk_ib_idx].length_dw);
+		return -EINVAL;
+	}
+
+	if (p->chunk_relocs_idx == -1) {
+		DRM_ERROR("No relocation chunk !\n");
+		return -EINVAL;
+	}
+
+
+	do {
+		r = radeon_cs_packet_parse(p, &pkt, p->idx);
+		if (r)
+			return r;
+		switch (pkt.type) {
+		case RADEON_PACKET_TYPE0:
+			r = radeon_uvd_cs_reg(p, &pkt, &data0,
+					      &data1, buf_sizes);
+			if (r)
+				return r;
+			break;
+		case RADEON_PACKET_TYPE2:
+			p->idx += pkt.count + 2;
+			break;
+		default:
+			DRM_ERROR("Unknown packet type %d !\n", pkt.type);
+			return -EINVAL;
+		}
+	} while (p->idx < p->chunks[p->chunk_ib_idx].length_dw);
+	return 0;
+}
+
+static int radeon_uvd_send_msg(struct radeon_device *rdev,
+			       int ring, struct radeon_bo *bo,
+			       struct radeon_fence **fence)
+{
+	struct ttm_validate_buffer tv;
+	struct list_head head;
+	struct radeon_ib ib;
+	uint64_t addr;
+	int i, r;
+
+	memset(&tv, 0, sizeof(tv));
+	tv.bo = &bo->tbo;
+
+	INIT_LIST_HEAD(&head);
+	list_add(&tv.head, &head);
+
+	r = ttm_eu_reserve_buffers(&head);
+	if (r)
+		return r;
+
+	radeon_ttm_placement_from_domain(bo, RADEON_GEM_DOMAIN_VRAM);
+	radeon_uvd_force_into_uvd_segment(bo);
+
+	r = ttm_bo_validate(&bo->tbo, &bo->placement, true, false);
+	if (r) {
+		ttm_eu_backoff_reservation(&head);
+		return r;
+	}
+
+	r = radeon_ib_get(rdev, ring, &ib, NULL, 16);
+	if (r) {
+		ttm_eu_backoff_reservation(&head);
+		return r;
+	}
+
+	addr = radeon_bo_gpu_offset(bo);
+	ib.ptr[0] = PACKET0(UVD_GPCOM_VCPU_DATA0, 0);
+	ib.ptr[1] = addr;
+	ib.ptr[2] = PACKET0(UVD_GPCOM_VCPU_DATA1, 0);
+	ib.ptr[3] = addr >> 32;
+	ib.ptr[4] = PACKET0(UVD_GPCOM_VCPU_CMD, 0);
+	ib.ptr[5] = 0;
+	for (i = 6; i < 16; ++i)
+		ib.ptr[i] = PACKET2(0);
+	ib.length_dw = 16;
+
+	r = radeon_ib_schedule(rdev, &ib, NULL);
+	if (r) {
+		ttm_eu_backoff_reservation(&head);
+		return r;
+	}
+	ttm_eu_fence_buffer_objects(&head, ib.fence);
+
+	if (fence)
+		*fence = radeon_fence_ref(ib.fence);
+
+	radeon_ib_free(rdev, &ib);
+	radeon_bo_unref(&bo);
+	return 0;
+}
+
+/* multiple fence commands without any stream commands in between can
+   crash the vcpu so just try to emmit a dummy create/destroy msg to
+   avoid this */
+int radeon_uvd_get_create_msg(struct radeon_device *rdev, int ring,
+			      uint32_t handle, struct radeon_fence **fence)
+{
+	struct radeon_bo *bo;
+	uint32_t *msg;
+	int r, i;
+
+	r = radeon_bo_create(rdev, 1024, PAGE_SIZE, true,
+			     RADEON_GEM_DOMAIN_VRAM, NULL, &bo);
+	if (r)
+		return r;
+
+	r = radeon_bo_reserve(bo, false);
+	if (r) {
+		radeon_bo_unref(&bo);
+		return r;
+	}
+
+	r = radeon_bo_kmap(bo, (void **)&msg);
+	if (r) {
+		radeon_bo_unreserve(bo);
+		radeon_bo_unref(&bo);
+		return r;
+	}
+
+	/* stitch together an UVD create msg */
+	msg[0] = 0x00000de4;
+	msg[1] = 0x00000000;
+	msg[2] = handle;
+	msg[3] = 0x00000000;
+	msg[4] = 0x00000000;
+	msg[5] = 0x00000000;
+	msg[6] = 0x00000000;
+	msg[7] = 0x00000780;
+	msg[8] = 0x00000440;
+	msg[9] = 0x00000000;
+	msg[10] = 0x01b37000;
+	for (i = 11; i < 1024; ++i)
+		msg[i] = 0x0;
+
+	radeon_bo_kunmap(bo);
+	radeon_bo_unreserve(bo);
+
+	return radeon_uvd_send_msg(rdev, ring, bo, fence);
+}
+
+int radeon_uvd_get_destroy_msg(struct radeon_device *rdev, int ring,
+			       uint32_t handle, struct radeon_fence **fence)
+{
+	struct radeon_bo *bo;
+	uint32_t *msg;
+	int r, i;
+
+	r = radeon_bo_create(rdev, 1024, PAGE_SIZE, true,
+			     RADEON_GEM_DOMAIN_VRAM, NULL, &bo);
+	if (r)
+		return r;
+
+	r = radeon_bo_reserve(bo, false);
+	if (r) {
+		radeon_bo_unref(&bo);
+		return r;
+	}
+
+	r = radeon_bo_kmap(bo, (void **)&msg);
+	if (r) {
+		radeon_bo_unreserve(bo);
+		radeon_bo_unref(&bo);
+		return r;
+	}
+
+	/* stitch together an UVD destroy msg */
+	msg[0] = 0x00000de4;
+	msg[1] = 0x00000002;
+	msg[2] = handle;
+	msg[3] = 0x00000000;
+	for (i = 4; i < 1024; ++i)
+		msg[i] = 0x0;
+
+	radeon_bo_kunmap(bo);
+	radeon_bo_unreserve(bo);
+
+	return radeon_uvd_send_msg(rdev, ring, bo, fence);
+}
+
+static void radeon_uvd_idle_work_handler(struct work_struct *work)
+{
+	struct radeon_device *rdev =
+		container_of(work, struct radeon_device, uvd.idle_work.work);
+
+	if (radeon_fence_count_emitted(rdev, R600_RING_TYPE_UVD_INDEX) == 0)
+		radeon_set_uvd_clocks(rdev, 0, 0);
+	else
+		schedule_delayed_work(&rdev->uvd.idle_work,
+				      msecs_to_jiffies(UVD_IDLE_TIMEOUT_MS));
+}
+
+void radeon_uvd_note_usage(struct radeon_device *rdev)
+{
+	bool set_clocks = !cancel_delayed_work_sync(&rdev->uvd.idle_work);
+	set_clocks &= schedule_delayed_work(&rdev->uvd.idle_work,
+					    msecs_to_jiffies(UVD_IDLE_TIMEOUT_MS));
+	if (set_clocks)
+		radeon_set_uvd_clocks(rdev, 53300, 40000);
+}
+
+static unsigned radeon_uvd_calc_upll_post_div(unsigned vco_freq,
+					      unsigned target_freq,
+					      unsigned pd_min,
+					      unsigned pd_even)
+{
+	unsigned post_div = vco_freq / target_freq;
+
+	/* adjust to post divider minimum value */
+	if (post_div < pd_min)
+		post_div = pd_min;
+
+	/* we alway need a frequency less than or equal the target */
+	if ((vco_freq / post_div) > target_freq)
+		post_div += 1;
+
+	/* post dividers above a certain value must be even */
+	if (post_div > pd_even && post_div % 2)
+		post_div += 1;
+
+	return post_div;
+}
+
+/**
+ * radeon_uvd_calc_upll_dividers - calc UPLL clock dividers
+ *
+ * @rdev: radeon_device pointer
+ * @vclk: wanted VCLK
+ * @dclk: wanted DCLK
+ * @vco_min: minimum VCO frequency
+ * @vco_max: maximum VCO frequency
+ * @fb_factor: factor to multiply vco freq with
+ * @fb_mask: limit and bitmask for feedback divider
+ * @pd_min: post divider minimum
+ * @pd_max: post divider maximum
+ * @pd_even: post divider must be even above this value
+ * @optimal_fb_div: resulting feedback divider
+ * @optimal_vclk_div: resulting vclk post divider
+ * @optimal_dclk_div: resulting dclk post divider
+ *
+ * Calculate dividers for UVDs UPLL (R6xx-SI, except APUs).
+ * Returns zero on success -EINVAL on error.
+ */
+int radeon_uvd_calc_upll_dividers(struct radeon_device *rdev,
+				  unsigned vclk, unsigned dclk,
+				  unsigned vco_min, unsigned vco_max,
+				  unsigned fb_factor, unsigned fb_mask,
+				  unsigned pd_min, unsigned pd_max,
+				  unsigned pd_even,
+				  unsigned *optimal_fb_div,
+				  unsigned *optimal_vclk_div,
+				  unsigned *optimal_dclk_div)
+{
+	unsigned vco_freq, ref_freq = rdev->clock.spll.reference_freq;
+
+	/* start off with something large */
+	unsigned optimal_score = ~0;
+
+	/* loop through vco from low to high */
+	vco_min = max(max(vco_min, vclk), dclk);
+	for (vco_freq = vco_min; vco_freq <= vco_max; vco_freq += 100) {
+
+		uint64_t fb_div = (uint64_t)vco_freq * fb_factor;
+		unsigned vclk_div, dclk_div, score;
+
+		do_div(fb_div, ref_freq);
+
+		/* fb div out of range ? */
+		if (fb_div > fb_mask)
+			break; /* it can oly get worse */
+
+		fb_div &= fb_mask;
+
+		/* calc vclk divider with current vco freq */
+		vclk_div = radeon_uvd_calc_upll_post_div(vco_freq, vclk,
+							 pd_min, pd_even);
+		if (vclk_div > pd_max)
+			break; /* vco is too big, it has to stop */
+
+		/* calc dclk divider with current vco freq */
+		dclk_div = radeon_uvd_calc_upll_post_div(vco_freq, dclk,
+							 pd_min, pd_even);
+		if (vclk_div > pd_max)
+			break; /* vco is too big, it has to stop */
+
+		/* calc score with current vco freq */
+		score = vclk - (vco_freq / vclk_div) + dclk - (vco_freq / dclk_div);
+
+		/* determine if this vco setting is better than current optimal settings */
+		if (score < optimal_score) {
+			*optimal_fb_div = fb_div;
+			*optimal_vclk_div = vclk_div;
+			*optimal_dclk_div = dclk_div;
+			optimal_score = score;
+			if (optimal_score == 0)
+				break; /* it can't get better than this */
+		}
+	}
+
+	/* did we found a valid setup ? */
+	if (optimal_score == ~0)
+		return -EINVAL;
+
+	return 0;
+}
+
+int radeon_uvd_send_upll_ctlreq(struct radeon_device *rdev,
+				unsigned cg_upll_func_cntl)
+{
+	unsigned i;
+
+	/* make sure UPLL_CTLREQ is deasserted */
+	WREG32_P(cg_upll_func_cntl, 0, ~UPLL_CTLREQ_MASK);
+
+	mdelay(10);
+
+	/* assert UPLL_CTLREQ */
+	WREG32_P(cg_upll_func_cntl, UPLL_CTLREQ_MASK, ~UPLL_CTLREQ_MASK);
+
+	/* wait for CTLACK and CTLACK2 to get asserted */
+	for (i = 0; i < 100; ++i) {
+		uint32_t mask = UPLL_CTLACK_MASK | UPLL_CTLACK2_MASK;
+		if ((RREG32(cg_upll_func_cntl) & mask) == mask)
+			break;
+		mdelay(10);
+	}
+
+	/* deassert UPLL_CTLREQ */
+	WREG32_P(cg_upll_func_cntl, 0, ~UPLL_CTLREQ_MASK);
+
+	if (i == 100) {
+		DRM_ERROR("Timeout setting UVD clocks!\n");
+		return -ETIMEDOUT;
+	}
+
+	return 0;
+}
diff --git a/drivers/gpu/drm/radeon/rs600.c b/drivers/gpu/drm/radeon/rs600.c
index 5a0fc74c2ba6..46fa1b07c560 100644
--- a/drivers/gpu/drm/radeon/rs600.c
+++ b/drivers/gpu/drm/radeon/rs600.c
@@ -52,23 +52,59 @@ static const u32 crtc_offsets[2] =
 	AVIVO_D2CRTC_H_TOTAL - AVIVO_D1CRTC_H_TOTAL
 };
 
+static bool avivo_is_in_vblank(struct radeon_device *rdev, int crtc)
+{
+	if (RREG32(AVIVO_D1CRTC_STATUS + crtc_offsets[crtc]) & AVIVO_D1CRTC_V_BLANK)
+		return true;
+	else
+		return false;
+}
+
+static bool avivo_is_counter_moving(struct radeon_device *rdev, int crtc)
+{
+	u32 pos1, pos2;
+
+	pos1 = RREG32(AVIVO_D1CRTC_STATUS_POSITION + crtc_offsets[crtc]);
+	pos2 = RREG32(AVIVO_D1CRTC_STATUS_POSITION + crtc_offsets[crtc]);
+
+	if (pos1 != pos2)
+		return true;
+	else
+		return false;
+}
+
+/**
+ * avivo_wait_for_vblank - vblank wait asic callback.
+ *
+ * @rdev: radeon_device pointer
+ * @crtc: crtc to wait for vblank on
+ *
+ * Wait for vblank on the requested crtc (r5xx-r7xx).
+ */
 void avivo_wait_for_vblank(struct radeon_device *rdev, int crtc)
 {
-	int i;
+	unsigned i = 0;
 
 	if (crtc >= rdev->num_crtc)
 		return;
 
-	if (RREG32(AVIVO_D1CRTC_CONTROL + crtc_offsets[crtc]) & AVIVO_CRTC_EN) {
-		for (i = 0; i < rdev->usec_timeout; i++) {
-			if (!(RREG32(AVIVO_D1CRTC_STATUS + crtc_offsets[crtc]) & AVIVO_D1CRTC_V_BLANK))
+	if (!(RREG32(AVIVO_D1CRTC_CONTROL + crtc_offsets[crtc]) & AVIVO_CRTC_EN))
+		return;
+
+	/* depending on when we hit vblank, we may be close to active; if so,
+	 * wait for another frame.
+	 */
+	while (avivo_is_in_vblank(rdev, crtc)) {
+		if (i++ % 100 == 0) {
+			if (!avivo_is_counter_moving(rdev, crtc))
 				break;
-			udelay(1);
 		}
-		for (i = 0; i < rdev->usec_timeout; i++) {
-			if (RREG32(AVIVO_D1CRTC_STATUS + crtc_offsets[crtc]) & AVIVO_D1CRTC_V_BLANK)
+	}
+
+	while (!avivo_is_in_vblank(rdev, crtc)) {
+		if (i++ % 100 == 0) {
+			if (!avivo_is_counter_moving(rdev, crtc))
 				break;
-			udelay(1);
 		}
 	}
 }
diff --git a/drivers/gpu/drm/radeon/rs690.c b/drivers/gpu/drm/radeon/rs690.c
index 5706d2ac75ab..ab4c86cfd552 100644
--- a/drivers/gpu/drm/radeon/rs690.c
+++ b/drivers/gpu/drm/radeon/rs690.c
@@ -148,6 +148,8 @@ void rs690_pm_info(struct radeon_device *rdev)
 static void rs690_mc_init(struct radeon_device *rdev)
 {
 	u64 base;
+	uint32_t h_addr, l_addr;
+	unsigned long long k8_addr;
 
 	rs400_gart_adjust_size(rdev);
 	rdev->mc.vram_is_ddr = true;
@@ -160,6 +162,27 @@ static void rs690_mc_init(struct radeon_device *rdev)
 	base = RREG32_MC(R_000100_MCCFG_FB_LOCATION);
 	base = G_000100_MC_FB_START(base) << 16;
 	rdev->mc.igp_sideport_enabled = radeon_atombios_sideport_present(rdev);
+
+	/* Use K8 direct mapping for fast fb access. */ 
+	rdev->fastfb_working = false;
+	h_addr = G_00005F_K8_ADDR_EXT(RREG32_MC(R_00005F_MC_MISC_UMA_CNTL));
+	l_addr = RREG32_MC(R_00001E_K8_FB_LOCATION);
+	k8_addr = ((unsigned long long)h_addr) << 32 | l_addr;
+#if defined(CONFIG_X86_32) && !defined(CONFIG_X86_PAE)
+	if (k8_addr + rdev->mc.visible_vram_size < 0x100000000ULL)	
+#endif
+	{
+		/* FastFB shall be used with UMA memory. Here it is simply disabled when sideport 
+		 * memory is present.
+		 */
+		if (rdev->mc.igp_sideport_enabled == false && radeon_fastfb == 1) {
+			DRM_INFO("Direct mapping: aper base at 0x%llx, replaced by direct mapping base 0x%llx.\n", 
+					(unsigned long long)rdev->mc.aper_base, k8_addr);
+			rdev->mc.aper_base = (resource_size_t)k8_addr;
+			rdev->fastfb_working = true;
+		}
+	}  
+
 	rs690_pm_info(rdev);
 	radeon_vram_location(rdev, &rdev->mc, base);
 	rdev->mc.gtt_base_align = rdev->mc.gtt_size - 1;
diff --git a/drivers/gpu/drm/radeon/rs690d.h b/drivers/gpu/drm/radeon/rs690d.h
index 36e6398a98ae..8af3ccf20cc0 100644
--- a/drivers/gpu/drm/radeon/rs690d.h
+++ b/drivers/gpu/drm/radeon/rs690d.h
@@ -29,6 +29,9 @@
 #define __RS690D_H__
 
 /* Registers */
+#define R_00001E_K8_FB_LOCATION                      0x00001E
+#define R_00005F_MC_MISC_UMA_CNTL                    0x00005F
+#define   G_00005F_K8_ADDR_EXT(x)                      (((x) >> 0) & 0xFF)
 #define R_000078_MC_INDEX                            0x000078
 #define   S_000078_MC_IND_ADDR(x)                      (((x) & 0x1FF) << 0)
 #define   G_000078_MC_IND_ADDR(x)                      (((x) >> 0) & 0x1FF)
diff --git a/drivers/gpu/drm/radeon/rv515.c b/drivers/gpu/drm/radeon/rv515.c
index 435ed3551364..ffcba730c57c 100644
--- a/drivers/gpu/drm/radeon/rv515.c
+++ b/drivers/gpu/drm/radeon/rv515.c
@@ -303,8 +303,10 @@ void rv515_mc_stop(struct radeon_device *rdev, struct rv515_mc_save *save)
 			tmp = RREG32(AVIVO_D1CRTC_CONTROL + crtc_offsets[i]);
 			if (!(tmp & AVIVO_CRTC_DISP_READ_REQUEST_DISABLE)) {
 				radeon_wait_for_vblank(rdev, i);
+				WREG32(AVIVO_D1CRTC_UPDATE_LOCK + crtc_offsets[i], 1);
 				tmp |= AVIVO_CRTC_DISP_READ_REQUEST_DISABLE;
 				WREG32(AVIVO_D1CRTC_CONTROL + crtc_offsets[i], tmp);
+				WREG32(AVIVO_D1CRTC_UPDATE_LOCK + crtc_offsets[i], 0);
 			}
 			/* wait for the next frame */
 			frame_count = radeon_get_vblank_counter(rdev, i);
@@ -313,6 +315,15 @@ void rv515_mc_stop(struct radeon_device *rdev, struct rv515_mc_save *save)
 					break;
 				udelay(1);
 			}
+
+			/* XXX this is a hack to avoid strange behavior with EFI on certain systems */
+			WREG32(AVIVO_D1CRTC_UPDATE_LOCK + crtc_offsets[i], 1);
+			tmp = RREG32(AVIVO_D1CRTC_CONTROL + crtc_offsets[i]);
+			tmp &= ~AVIVO_CRTC_EN;
+			WREG32(AVIVO_D1CRTC_CONTROL + crtc_offsets[i], tmp);
+			WREG32(AVIVO_D1CRTC_UPDATE_LOCK + crtc_offsets[i], 0);
+			save->crtc_enabled[i] = false;
+			/* ***** */
 		} else {
 			save->crtc_enabled[i] = false;
 		}
@@ -338,6 +349,22 @@ void rv515_mc_stop(struct radeon_device *rdev, struct rv515_mc_save *save)
 	}
 	/* wait for the MC to settle */
 	udelay(100);
+
+	/* lock double buffered regs */
+	for (i = 0; i < rdev->num_crtc; i++) {
+		if (save->crtc_enabled[i]) {
+			tmp = RREG32(AVIVO_D1GRPH_UPDATE + crtc_offsets[i]);
+			if (!(tmp & AVIVO_D1GRPH_UPDATE_LOCK)) {
+				tmp |= AVIVO_D1GRPH_UPDATE_LOCK;
+				WREG32(AVIVO_D1GRPH_UPDATE + crtc_offsets[i], tmp);
+			}
+			tmp = RREG32(AVIVO_D1MODE_MASTER_UPDATE_LOCK + crtc_offsets[i]);
+			if (!(tmp & 1)) {
+				tmp |= 1;
+				WREG32(AVIVO_D1MODE_MASTER_UPDATE_LOCK + crtc_offsets[i], tmp);
+			}
+		}
+	}
 }
 
 void rv515_mc_resume(struct radeon_device *rdev, struct rv515_mc_save *save)
@@ -348,7 +375,7 @@ void rv515_mc_resume(struct radeon_device *rdev, struct rv515_mc_save *save)
 	/* update crtc base addresses */
 	for (i = 0; i < rdev->num_crtc; i++) {
 		if (rdev->family >= CHIP_RV770) {
-			if (i == 1) {
+			if (i == 0) {
 				WREG32(R700_D1GRPH_PRIMARY_SURFACE_ADDRESS_HIGH,
 				       upper_32_bits(rdev->mc.vram_start));
 				WREG32(R700_D1GRPH_SECONDARY_SURFACE_ADDRESS_HIGH,
@@ -367,6 +394,33 @@ void rv515_mc_resume(struct radeon_device *rdev, struct rv515_mc_save *save)
 	}
 	WREG32(R_000310_VGA_MEMORY_BASE_ADDRESS, (u32)rdev->mc.vram_start);
 
+	/* unlock regs and wait for update */
+	for (i = 0; i < rdev->num_crtc; i++) {
+		if (save->crtc_enabled[i]) {
+			tmp = RREG32(AVIVO_D1MODE_MASTER_UPDATE_MODE + crtc_offsets[i]);
+			if ((tmp & 0x3) != 0) {
+				tmp &= ~0x3;
+				WREG32(AVIVO_D1MODE_MASTER_UPDATE_MODE + crtc_offsets[i], tmp);
+			}
+			tmp = RREG32(AVIVO_D1GRPH_UPDATE + crtc_offsets[i]);
+			if (tmp & AVIVO_D1GRPH_UPDATE_LOCK) {
+				tmp &= ~AVIVO_D1GRPH_UPDATE_LOCK;
+				WREG32(AVIVO_D1GRPH_UPDATE + crtc_offsets[i], tmp);
+			}
+			tmp = RREG32(AVIVO_D1MODE_MASTER_UPDATE_LOCK + crtc_offsets[i]);
+			if (tmp & 1) {
+				tmp &= ~1;
+				WREG32(AVIVO_D1MODE_MASTER_UPDATE_LOCK + crtc_offsets[i], tmp);
+			}
+			for (j = 0; j < rdev->usec_timeout; j++) {
+				tmp = RREG32(AVIVO_D1GRPH_UPDATE + crtc_offsets[i]);
+				if ((tmp & AVIVO_D1GRPH_SURFACE_UPDATE_PENDING) == 0)
+					break;
+				udelay(1);
+			}
+		}
+	}
+
 	if (rdev->family >= CHIP_R600) {
 		/* unblackout the MC */
 		if (rdev->family >= CHIP_RV770)
diff --git a/drivers/gpu/drm/radeon/rv770.c b/drivers/gpu/drm/radeon/rv770.c
index d63fe1d0f53f..83f612a9500b 100644
--- a/drivers/gpu/drm/radeon/rv770.c
+++ b/drivers/gpu/drm/radeon/rv770.c
@@ -42,6 +42,739 @@
 static void rv770_gpu_init(struct radeon_device *rdev);
 void rv770_fini(struct radeon_device *rdev);
 static void rv770_pcie_gen2_enable(struct radeon_device *rdev);
+int evergreen_set_uvd_clocks(struct radeon_device *rdev, u32 vclk, u32 dclk);
+
+int rv770_set_uvd_clocks(struct radeon_device *rdev, u32 vclk, u32 dclk)
+{
+	unsigned fb_div = 0, vclk_div = 0, dclk_div = 0;
+	int r;
+
+	/* RV740 uses evergreen uvd clk programming */
+	if (rdev->family == CHIP_RV740)
+		return evergreen_set_uvd_clocks(rdev, vclk, dclk);
+
+	/* bypass vclk and dclk with bclk */
+	WREG32_P(CG_UPLL_FUNC_CNTL_2,
+		 VCLK_SRC_SEL(1) | DCLK_SRC_SEL(1),
+		 ~(VCLK_SRC_SEL_MASK | DCLK_SRC_SEL_MASK));
+
+	if (!vclk || !dclk) {
+		/* keep the Bypass mode, put PLL to sleep */
+		WREG32_P(CG_UPLL_FUNC_CNTL, UPLL_SLEEP_MASK, ~UPLL_SLEEP_MASK);
+		return 0;
+	}
+
+	r = radeon_uvd_calc_upll_dividers(rdev, vclk, dclk, 50000, 160000,
+					  43663, 0x03FFFFFE, 1, 30, ~0,
+					  &fb_div, &vclk_div, &dclk_div);
+	if (r)
+		return r;
+
+	fb_div |= 1;
+	vclk_div -= 1;
+	dclk_div -= 1;
+
+	/* set UPLL_FB_DIV to 0x50000 */
+	WREG32_P(CG_UPLL_FUNC_CNTL_3, UPLL_FB_DIV(0x50000), ~UPLL_FB_DIV_MASK);
+
+	/* deassert UPLL_RESET and UPLL_SLEEP */
+	WREG32_P(CG_UPLL_FUNC_CNTL, 0, ~(UPLL_RESET_MASK | UPLL_SLEEP_MASK));
+
+	/* assert BYPASS EN and FB_DIV[0] <- ??? why? */
+	WREG32_P(CG_UPLL_FUNC_CNTL, UPLL_BYPASS_EN_MASK, ~UPLL_BYPASS_EN_MASK);
+	WREG32_P(CG_UPLL_FUNC_CNTL_3, UPLL_FB_DIV(1), ~UPLL_FB_DIV(1));
+
+	r = radeon_uvd_send_upll_ctlreq(rdev, CG_UPLL_FUNC_CNTL);
+	if (r)
+		return r;
+
+	/* assert PLL_RESET */
+	WREG32_P(CG_UPLL_FUNC_CNTL, UPLL_RESET_MASK, ~UPLL_RESET_MASK);
+
+	/* set the required FB_DIV, REF_DIV, Post divder values */
+	WREG32_P(CG_UPLL_FUNC_CNTL, UPLL_REF_DIV(1), ~UPLL_REF_DIV_MASK);
+	WREG32_P(CG_UPLL_FUNC_CNTL_2,
+		 UPLL_SW_HILEN(vclk_div >> 1) |
+		 UPLL_SW_LOLEN((vclk_div >> 1) + (vclk_div & 1)) |
+		 UPLL_SW_HILEN2(dclk_div >> 1) |
+		 UPLL_SW_LOLEN2((dclk_div >> 1) + (dclk_div & 1)),
+		 ~UPLL_SW_MASK);
+
+	WREG32_P(CG_UPLL_FUNC_CNTL_3, UPLL_FB_DIV(fb_div),
+		 ~UPLL_FB_DIV_MASK);
+
+	/* give the PLL some time to settle */
+	mdelay(15);
+
+	/* deassert PLL_RESET */
+	WREG32_P(CG_UPLL_FUNC_CNTL, 0, ~UPLL_RESET_MASK);
+
+	mdelay(15);
+
+	/* deassert BYPASS EN and FB_DIV[0] <- ??? why? */
+	WREG32_P(CG_UPLL_FUNC_CNTL, 0, ~UPLL_BYPASS_EN_MASK);
+	WREG32_P(CG_UPLL_FUNC_CNTL_3, 0, ~UPLL_FB_DIV(1));
+
+	r = radeon_uvd_send_upll_ctlreq(rdev, CG_UPLL_FUNC_CNTL);
+	if (r)
+		return r;
+
+	/* switch VCLK and DCLK selection */
+	WREG32_P(CG_UPLL_FUNC_CNTL_2,
+		 VCLK_SRC_SEL(2) | DCLK_SRC_SEL(2),
+		 ~(VCLK_SRC_SEL_MASK | DCLK_SRC_SEL_MASK));
+
+	mdelay(100);
+
+	return 0;
+}
+
+static const u32 r7xx_golden_registers[] =
+{
+	0x8d00, 0xffffffff, 0x0e0e0074,
+	0x8d04, 0xffffffff, 0x013a2b34,
+	0x9508, 0xffffffff, 0x00000002,
+	0x8b20, 0xffffffff, 0,
+	0x88c4, 0xffffffff, 0x000000c2,
+	0x28350, 0xffffffff, 0,
+	0x9058, 0xffffffff, 0x0fffc40f,
+	0x240c, 0xffffffff, 0x00000380,
+	0x733c, 0xffffffff, 0x00000002,
+	0x2650, 0x00040000, 0,
+	0x20bc, 0x00040000, 0,
+	0x7300, 0xffffffff, 0x001000f0
+};
+
+static const u32 r7xx_golden_dyn_gpr_registers[] =
+{
+	0x8db0, 0xffffffff, 0x98989898,
+	0x8db4, 0xffffffff, 0x98989898,
+	0x8db8, 0xffffffff, 0x98989898,
+	0x8dbc, 0xffffffff, 0x98989898,
+	0x8dc0, 0xffffffff, 0x98989898,
+	0x8dc4, 0xffffffff, 0x98989898,
+	0x8dc8, 0xffffffff, 0x98989898,
+	0x8dcc, 0xffffffff, 0x98989898,
+	0x88c4, 0xffffffff, 0x00000082
+};
+
+static const u32 rv770_golden_registers[] =
+{
+	0x562c, 0xffffffff, 0,
+	0x3f90, 0xffffffff, 0,
+	0x9148, 0xffffffff, 0,
+	0x3f94, 0xffffffff, 0,
+	0x914c, 0xffffffff, 0,
+	0x9698, 0x18000000, 0x18000000
+};
+
+static const u32 rv770ce_golden_registers[] =
+{
+	0x562c, 0xffffffff, 0,
+	0x3f90, 0xffffffff, 0x00cc0000,
+	0x9148, 0xffffffff, 0x00cc0000,
+	0x3f94, 0xffffffff, 0x00cc0000,
+	0x914c, 0xffffffff, 0x00cc0000,
+	0x9b7c, 0xffffffff, 0x00fa0000,
+	0x3f8c, 0xffffffff, 0x00fa0000,
+	0x9698, 0x18000000, 0x18000000
+};
+
+static const u32 rv770_mgcg_init[] =
+{
+	0x8bcc, 0xffffffff, 0x130300f9,
+	0x5448, 0xffffffff, 0x100,
+	0x55e4, 0xffffffff, 0x100,
+	0x160c, 0xffffffff, 0x100,
+	0x5644, 0xffffffff, 0x100,
+	0xc164, 0xffffffff, 0x100,
+	0x8a18, 0xffffffff, 0x100,
+	0x897c, 0xffffffff, 0x8000100,
+	0x8b28, 0xffffffff, 0x3c000100,
+	0x9144, 0xffffffff, 0x100,
+	0x9a1c, 0xffffffff, 0x10000,
+	0x9a50, 0xffffffff, 0x100,
+	0x9a1c, 0xffffffff, 0x10001,
+	0x9a50, 0xffffffff, 0x100,
+	0x9a1c, 0xffffffff, 0x10002,
+	0x9a50, 0xffffffff, 0x100,
+	0x9a1c, 0xffffffff, 0x10003,
+	0x9a50, 0xffffffff, 0x100,
+	0x9a1c, 0xffffffff, 0x0,
+	0x9870, 0xffffffff, 0x100,
+	0x8d58, 0xffffffff, 0x100,
+	0x9500, 0xffffffff, 0x0,
+	0x9510, 0xffffffff, 0x100,
+	0x9500, 0xffffffff, 0x1,
+	0x9510, 0xffffffff, 0x100,
+	0x9500, 0xffffffff, 0x2,
+	0x9510, 0xffffffff, 0x100,
+	0x9500, 0xffffffff, 0x3,
+	0x9510, 0xffffffff, 0x100,
+	0x9500, 0xffffffff, 0x4,
+	0x9510, 0xffffffff, 0x100,
+	0x9500, 0xffffffff, 0x5,
+	0x9510, 0xffffffff, 0x100,
+	0x9500, 0xffffffff, 0x6,
+	0x9510, 0xffffffff, 0x100,
+	0x9500, 0xffffffff, 0x7,
+	0x9510, 0xffffffff, 0x100,
+	0x9500, 0xffffffff, 0x8,
+	0x9510, 0xffffffff, 0x100,
+	0x9500, 0xffffffff, 0x9,
+	0x9510, 0xffffffff, 0x100,
+	0x9500, 0xffffffff, 0x8000,
+	0x9490, 0xffffffff, 0x0,
+	0x949c, 0xffffffff, 0x100,
+	0x9490, 0xffffffff, 0x1,
+	0x949c, 0xffffffff, 0x100,
+	0x9490, 0xffffffff, 0x2,
+	0x949c, 0xffffffff, 0x100,
+	0x9490, 0xffffffff, 0x3,
+	0x949c, 0xffffffff, 0x100,
+	0x9490, 0xffffffff, 0x4,
+	0x949c, 0xffffffff, 0x100,
+	0x9490, 0xffffffff, 0x5,
+	0x949c, 0xffffffff, 0x100,
+	0x9490, 0xffffffff, 0x6,
+	0x949c, 0xffffffff, 0x100,
+	0x9490, 0xffffffff, 0x7,
+	0x949c, 0xffffffff, 0x100,
+	0x9490, 0xffffffff, 0x8,
+	0x949c, 0xffffffff, 0x100,
+	0x9490, 0xffffffff, 0x9,
+	0x949c, 0xffffffff, 0x100,
+	0x9490, 0xffffffff, 0x8000,
+	0x9604, 0xffffffff, 0x0,
+	0x9654, 0xffffffff, 0x100,
+	0x9604, 0xffffffff, 0x1,
+	0x9654, 0xffffffff, 0x100,
+	0x9604, 0xffffffff, 0x2,
+	0x9654, 0xffffffff, 0x100,
+	0x9604, 0xffffffff, 0x3,
+	0x9654, 0xffffffff, 0x100,
+	0x9604, 0xffffffff, 0x4,
+	0x9654, 0xffffffff, 0x100,
+	0x9604, 0xffffffff, 0x5,
+	0x9654, 0xffffffff, 0x100,
+	0x9604, 0xffffffff, 0x6,
+	0x9654, 0xffffffff, 0x100,
+	0x9604, 0xffffffff, 0x7,
+	0x9654, 0xffffffff, 0x100,
+	0x9604, 0xffffffff, 0x8,
+	0x9654, 0xffffffff, 0x100,
+	0x9604, 0xffffffff, 0x9,
+	0x9654, 0xffffffff, 0x100,
+	0x9604, 0xffffffff, 0x80000000,
+	0x9030, 0xffffffff, 0x100,
+	0x9034, 0xffffffff, 0x100,
+	0x9038, 0xffffffff, 0x100,
+	0x903c, 0xffffffff, 0x100,
+	0x9040, 0xffffffff, 0x100,
+	0xa200, 0xffffffff, 0x100,
+	0xa204, 0xffffffff, 0x100,
+	0xa208, 0xffffffff, 0x100,
+	0xa20c, 0xffffffff, 0x100,
+	0x971c, 0xffffffff, 0x100,
+	0x915c, 0xffffffff, 0x00020001,
+	0x9160, 0xffffffff, 0x00040003,
+	0x916c, 0xffffffff, 0x00060005,
+	0x9170, 0xffffffff, 0x00080007,
+	0x9174, 0xffffffff, 0x000a0009,
+	0x9178, 0xffffffff, 0x000c000b,
+	0x917c, 0xffffffff, 0x000e000d,
+	0x9180, 0xffffffff, 0x0010000f,
+	0x918c, 0xffffffff, 0x00120011,
+	0x9190, 0xffffffff, 0x00140013,
+	0x9194, 0xffffffff, 0x00020001,
+	0x9198, 0xffffffff, 0x00040003,
+	0x919c, 0xffffffff, 0x00060005,
+	0x91a8, 0xffffffff, 0x00080007,
+	0x91ac, 0xffffffff, 0x000a0009,
+	0x91b0, 0xffffffff, 0x000c000b,
+	0x91b4, 0xffffffff, 0x000e000d,
+	0x91b8, 0xffffffff, 0x0010000f,
+	0x91c4, 0xffffffff, 0x00120011,
+	0x91c8, 0xffffffff, 0x00140013,
+	0x91cc, 0xffffffff, 0x00020001,
+	0x91d0, 0xffffffff, 0x00040003,
+	0x91d4, 0xffffffff, 0x00060005,
+	0x91e0, 0xffffffff, 0x00080007,
+	0x91e4, 0xffffffff, 0x000a0009,
+	0x91e8, 0xffffffff, 0x000c000b,
+	0x91ec, 0xffffffff, 0x00020001,
+	0x91f0, 0xffffffff, 0x00040003,
+	0x91f4, 0xffffffff, 0x00060005,
+	0x9200, 0xffffffff, 0x00080007,
+	0x9204, 0xffffffff, 0x000a0009,
+	0x9208, 0xffffffff, 0x000c000b,
+	0x920c, 0xffffffff, 0x000e000d,
+	0x9210, 0xffffffff, 0x0010000f,
+	0x921c, 0xffffffff, 0x00120011,
+	0x9220, 0xffffffff, 0x00140013,
+	0x9224, 0xffffffff, 0x00020001,
+	0x9228, 0xffffffff, 0x00040003,
+	0x922c, 0xffffffff, 0x00060005,
+	0x9238, 0xffffffff, 0x00080007,
+	0x923c, 0xffffffff, 0x000a0009,
+	0x9240, 0xffffffff, 0x000c000b,
+	0x9244, 0xffffffff, 0x000e000d,
+	0x9248, 0xffffffff, 0x0010000f,
+	0x9254, 0xffffffff, 0x00120011,
+	0x9258, 0xffffffff, 0x00140013,
+	0x925c, 0xffffffff, 0x00020001,
+	0x9260, 0xffffffff, 0x00040003,
+	0x9264, 0xffffffff, 0x00060005,
+	0x9270, 0xffffffff, 0x00080007,
+	0x9274, 0xffffffff, 0x000a0009,
+	0x9278, 0xffffffff, 0x000c000b,
+	0x927c, 0xffffffff, 0x000e000d,
+	0x9280, 0xffffffff, 0x0010000f,
+	0x928c, 0xffffffff, 0x00120011,
+	0x9290, 0xffffffff, 0x00140013,
+	0x9294, 0xffffffff, 0x00020001,
+	0x929c, 0xffffffff, 0x00040003,
+	0x92a0, 0xffffffff, 0x00060005,
+	0x92a4, 0xffffffff, 0x00080007
+};
+
+static const u32 rv710_golden_registers[] =
+{
+	0x3f90, 0x00ff0000, 0x00fc0000,
+	0x9148, 0x00ff0000, 0x00fc0000,
+	0x3f94, 0x00ff0000, 0x00fc0000,
+	0x914c, 0x00ff0000, 0x00fc0000,
+	0xb4c, 0x00000020, 0x00000020,
+	0xa180, 0xffffffff, 0x00003f3f
+};
+
+static const u32 rv710_mgcg_init[] =
+{
+	0x8bcc, 0xffffffff, 0x13030040,
+	0x5448, 0xffffffff, 0x100,
+	0x55e4, 0xffffffff, 0x100,
+	0x160c, 0xffffffff, 0x100,
+	0x5644, 0xffffffff, 0x100,
+	0xc164, 0xffffffff, 0x100,
+	0x8a18, 0xffffffff, 0x100,
+	0x897c, 0xffffffff, 0x8000100,
+	0x8b28, 0xffffffff, 0x3c000100,
+	0x9144, 0xffffffff, 0x100,
+	0x9a1c, 0xffffffff, 0x10000,
+	0x9a50, 0xffffffff, 0x100,
+	0x9a1c, 0xffffffff, 0x0,
+	0x9870, 0xffffffff, 0x100,
+	0x8d58, 0xffffffff, 0x100,
+	0x9500, 0xffffffff, 0x0,
+	0x9510, 0xffffffff, 0x100,
+	0x9500, 0xffffffff, 0x1,
+	0x9510, 0xffffffff, 0x100,
+	0x9500, 0xffffffff, 0x8000,
+	0x9490, 0xffffffff, 0x0,
+	0x949c, 0xffffffff, 0x100,
+	0x9490, 0xffffffff, 0x1,
+	0x949c, 0xffffffff, 0x100,
+	0x9490, 0xffffffff, 0x8000,
+	0x9604, 0xffffffff, 0x0,
+	0x9654, 0xffffffff, 0x100,
+	0x9604, 0xffffffff, 0x1,
+	0x9654, 0xffffffff, 0x100,
+	0x9604, 0xffffffff, 0x80000000,
+	0x9030, 0xffffffff, 0x100,
+	0x9034, 0xffffffff, 0x100,
+	0x9038, 0xffffffff, 0x100,
+	0x903c, 0xffffffff, 0x100,
+	0x9040, 0xffffffff, 0x100,
+	0xa200, 0xffffffff, 0x100,
+	0xa204, 0xffffffff, 0x100,
+	0xa208, 0xffffffff, 0x100,
+	0xa20c, 0xffffffff, 0x100,
+	0x971c, 0xffffffff, 0x100,
+	0x915c, 0xffffffff, 0x00020001,
+	0x9174, 0xffffffff, 0x00000003,
+	0x9178, 0xffffffff, 0x00050001,
+	0x917c, 0xffffffff, 0x00030002,
+	0x918c, 0xffffffff, 0x00000004,
+	0x9190, 0xffffffff, 0x00070006,
+	0x9194, 0xffffffff, 0x00050001,
+	0x9198, 0xffffffff, 0x00030002,
+	0x91a8, 0xffffffff, 0x00000004,
+	0x91ac, 0xffffffff, 0x00070006,
+	0x91e8, 0xffffffff, 0x00000001,
+	0x9294, 0xffffffff, 0x00000001,
+	0x929c, 0xffffffff, 0x00000002,
+	0x92a0, 0xffffffff, 0x00040003,
+	0x9150, 0xffffffff, 0x4d940000
+};
+
+static const u32 rv730_golden_registers[] =
+{
+	0x3f90, 0x00ff0000, 0x00f00000,
+	0x9148, 0x00ff0000, 0x00f00000,
+	0x3f94, 0x00ff0000, 0x00f00000,
+	0x914c, 0x00ff0000, 0x00f00000,
+	0x900c, 0xffffffff, 0x003b033f,
+	0xb4c, 0x00000020, 0x00000020,
+	0xa180, 0xffffffff, 0x00003f3f
+};
+
+static const u32 rv730_mgcg_init[] =
+{
+	0x8bcc, 0xffffffff, 0x130300f9,
+	0x5448, 0xffffffff, 0x100,
+	0x55e4, 0xffffffff, 0x100,
+	0x160c, 0xffffffff, 0x100,
+	0x5644, 0xffffffff, 0x100,
+	0xc164, 0xffffffff, 0x100,
+	0x8a18, 0xffffffff, 0x100,
+	0x897c, 0xffffffff, 0x8000100,
+	0x8b28, 0xffffffff, 0x3c000100,
+	0x9144, 0xffffffff, 0x100,
+	0x9a1c, 0xffffffff, 0x10000,
+	0x9a50, 0xffffffff, 0x100,
+	0x9a1c, 0xffffffff, 0x10001,
+	0x9a50, 0xffffffff, 0x100,
+	0x9a1c, 0xffffffff, 0x0,
+	0x9870, 0xffffffff, 0x100,
+	0x8d58, 0xffffffff, 0x100,
+	0x9500, 0xffffffff, 0x0,
+	0x9510, 0xffffffff, 0x100,
+	0x9500, 0xffffffff, 0x1,
+	0x9510, 0xffffffff, 0x100,
+	0x9500, 0xffffffff, 0x2,
+	0x9510, 0xffffffff, 0x100,
+	0x9500, 0xffffffff, 0x3,
+	0x9510, 0xffffffff, 0x100,
+	0x9500, 0xffffffff, 0x4,
+	0x9510, 0xffffffff, 0x100,
+	0x9500, 0xffffffff, 0x5,
+	0x9510, 0xffffffff, 0x100,
+	0x9500, 0xffffffff, 0x6,
+	0x9510, 0xffffffff, 0x100,
+	0x9500, 0xffffffff, 0x7,
+	0x9510, 0xffffffff, 0x100,
+	0x9500, 0xffffffff, 0x8000,
+	0x9490, 0xffffffff, 0x0,
+	0x949c, 0xffffffff, 0x100,
+	0x9490, 0xffffffff, 0x1,
+	0x949c, 0xffffffff, 0x100,
+	0x9490, 0xffffffff, 0x2,
+	0x949c, 0xffffffff, 0x100,
+	0x9490, 0xffffffff, 0x3,
+	0x949c, 0xffffffff, 0x100,
+	0x9490, 0xffffffff, 0x4,
+	0x949c, 0xffffffff, 0x100,
+	0x9490, 0xffffffff, 0x5,
+	0x949c, 0xffffffff, 0x100,
+	0x9490, 0xffffffff, 0x6,
+	0x949c, 0xffffffff, 0x100,
+	0x9490, 0xffffffff, 0x7,
+	0x949c, 0xffffffff, 0x100,
+	0x9490, 0xffffffff, 0x8000,
+	0x9604, 0xffffffff, 0x0,
+	0x9654, 0xffffffff, 0x100,
+	0x9604, 0xffffffff, 0x1,
+	0x9654, 0xffffffff, 0x100,
+	0x9604, 0xffffffff, 0x2,
+	0x9654, 0xffffffff, 0x100,
+	0x9604, 0xffffffff, 0x3,
+	0x9654, 0xffffffff, 0x100,
+	0x9604, 0xffffffff, 0x4,
+	0x9654, 0xffffffff, 0x100,
+	0x9604, 0xffffffff, 0x5,
+	0x9654, 0xffffffff, 0x100,
+	0x9604, 0xffffffff, 0x6,
+	0x9654, 0xffffffff, 0x100,
+	0x9604, 0xffffffff, 0x7,
+	0x9654, 0xffffffff, 0x100,
+	0x9604, 0xffffffff, 0x80000000,
+	0x9030, 0xffffffff, 0x100,
+	0x9034, 0xffffffff, 0x100,
+	0x9038, 0xffffffff, 0x100,
+	0x903c, 0xffffffff, 0x100,
+	0x9040, 0xffffffff, 0x100,
+	0xa200, 0xffffffff, 0x100,
+	0xa204, 0xffffffff, 0x100,
+	0xa208, 0xffffffff, 0x100,
+	0xa20c, 0xffffffff, 0x100,
+	0x971c, 0xffffffff, 0x100,
+	0x915c, 0xffffffff, 0x00020001,
+	0x916c, 0xffffffff, 0x00040003,
+	0x9170, 0xffffffff, 0x00000005,
+	0x9178, 0xffffffff, 0x00050001,
+	0x917c, 0xffffffff, 0x00030002,
+	0x918c, 0xffffffff, 0x00000004,
+	0x9190, 0xffffffff, 0x00070006,
+	0x9194, 0xffffffff, 0x00050001,
+	0x9198, 0xffffffff, 0x00030002,
+	0x91a8, 0xffffffff, 0x00000004,
+	0x91ac, 0xffffffff, 0x00070006,
+	0x91b0, 0xffffffff, 0x00050001,
+	0x91b4, 0xffffffff, 0x00030002,
+	0x91c4, 0xffffffff, 0x00000004,
+	0x91c8, 0xffffffff, 0x00070006,
+	0x91cc, 0xffffffff, 0x00050001,
+	0x91d0, 0xffffffff, 0x00030002,
+	0x91e0, 0xffffffff, 0x00000004,
+	0x91e4, 0xffffffff, 0x00070006,
+	0x91e8, 0xffffffff, 0x00000001,
+	0x91ec, 0xffffffff, 0x00050001,
+	0x91f0, 0xffffffff, 0x00030002,
+	0x9200, 0xffffffff, 0x00000004,
+	0x9204, 0xffffffff, 0x00070006,
+	0x9208, 0xffffffff, 0x00050001,
+	0x920c, 0xffffffff, 0x00030002,
+	0x921c, 0xffffffff, 0x00000004,
+	0x9220, 0xffffffff, 0x00070006,
+	0x9224, 0xffffffff, 0x00050001,
+	0x9228, 0xffffffff, 0x00030002,
+	0x9238, 0xffffffff, 0x00000004,
+	0x923c, 0xffffffff, 0x00070006,
+	0x9240, 0xffffffff, 0x00050001,
+	0x9244, 0xffffffff, 0x00030002,
+	0x9254, 0xffffffff, 0x00000004,
+	0x9258, 0xffffffff, 0x00070006,
+	0x9294, 0xffffffff, 0x00000001,
+	0x929c, 0xffffffff, 0x00000002,
+	0x92a0, 0xffffffff, 0x00040003,
+	0x92a4, 0xffffffff, 0x00000005
+};
+
+static const u32 rv740_golden_registers[] =
+{
+	0x88c4, 0xffffffff, 0x00000082,
+	0x28a50, 0xfffffffc, 0x00000004,
+	0x2650, 0x00040000, 0,
+	0x20bc, 0x00040000, 0,
+	0x733c, 0xffffffff, 0x00000002,
+	0x7300, 0xffffffff, 0x001000f0,
+	0x3f90, 0x00ff0000, 0,
+	0x9148, 0x00ff0000, 0,
+	0x3f94, 0x00ff0000, 0,
+	0x914c, 0x00ff0000, 0,
+	0x240c, 0xffffffff, 0x00000380,
+	0x8a14, 0x00000007, 0x00000007,
+	0x8b24, 0xffffffff, 0x00ff0fff,
+	0x28a4c, 0xffffffff, 0x00004000,
+	0xa180, 0xffffffff, 0x00003f3f,
+	0x8d00, 0xffffffff, 0x0e0e003a,
+	0x8d04, 0xffffffff, 0x013a0e2a,
+	0x8c00, 0xffffffff, 0xe400000f,
+	0x8db0, 0xffffffff, 0x98989898,
+	0x8db4, 0xffffffff, 0x98989898,
+	0x8db8, 0xffffffff, 0x98989898,
+	0x8dbc, 0xffffffff, 0x98989898,
+	0x8dc0, 0xffffffff, 0x98989898,
+	0x8dc4, 0xffffffff, 0x98989898,
+	0x8dc8, 0xffffffff, 0x98989898,
+	0x8dcc, 0xffffffff, 0x98989898,
+	0x9058, 0xffffffff, 0x0fffc40f,
+	0x900c, 0xffffffff, 0x003b033f,
+	0x28350, 0xffffffff, 0,
+	0x8cf0, 0x1fffffff, 0x08e00420,
+	0x9508, 0xffffffff, 0x00000002,
+	0x88c4, 0xffffffff, 0x000000c2,
+	0x9698, 0x18000000, 0x18000000
+};
+
+static const u32 rv740_mgcg_init[] =
+{
+	0x8bcc, 0xffffffff, 0x13030100,
+	0x5448, 0xffffffff, 0x100,
+	0x55e4, 0xffffffff, 0x100,
+	0x160c, 0xffffffff, 0x100,
+	0x5644, 0xffffffff, 0x100,
+	0xc164, 0xffffffff, 0x100,
+	0x8a18, 0xffffffff, 0x100,
+	0x897c, 0xffffffff, 0x100,
+	0x8b28, 0xffffffff, 0x100,
+	0x9144, 0xffffffff, 0x100,
+	0x9a1c, 0xffffffff, 0x10000,
+	0x9a50, 0xffffffff, 0x100,
+	0x9a1c, 0xffffffff, 0x10001,
+	0x9a50, 0xffffffff, 0x100,
+	0x9a1c, 0xffffffff, 0x10002,
+	0x9a50, 0xffffffff, 0x100,
+	0x9a1c, 0xffffffff, 0x10003,
+	0x9a50, 0xffffffff, 0x100,
+	0x9a1c, 0xffffffff, 0x0,
+	0x9870, 0xffffffff, 0x100,
+	0x8d58, 0xffffffff, 0x100,
+	0x9500, 0xffffffff, 0x0,
+	0x9510, 0xffffffff, 0x100,
+	0x9500, 0xffffffff, 0x1,
+	0x9510, 0xffffffff, 0x100,
+	0x9500, 0xffffffff, 0x2,
+	0x9510, 0xffffffff, 0x100,
+	0x9500, 0xffffffff, 0x3,
+	0x9510, 0xffffffff, 0x100,
+	0x9500, 0xffffffff, 0x4,
+	0x9510, 0xffffffff, 0x100,
+	0x9500, 0xffffffff, 0x5,
+	0x9510, 0xffffffff, 0x100,
+	0x9500, 0xffffffff, 0x6,
+	0x9510, 0xffffffff, 0x100,
+	0x9500, 0xffffffff, 0x7,
+	0x9510, 0xffffffff, 0x100,
+	0x9500, 0xffffffff, 0x8000,
+	0x9490, 0xffffffff, 0x0,
+	0x949c, 0xffffffff, 0x100,
+	0x9490, 0xffffffff, 0x1,
+	0x949c, 0xffffffff, 0x100,
+	0x9490, 0xffffffff, 0x2,
+	0x949c, 0xffffffff, 0x100,
+	0x9490, 0xffffffff, 0x3,
+	0x949c, 0xffffffff, 0x100,
+	0x9490, 0xffffffff, 0x4,
+	0x949c, 0xffffffff, 0x100,
+	0x9490, 0xffffffff, 0x5,
+	0x949c, 0xffffffff, 0x100,
+	0x9490, 0xffffffff, 0x6,
+	0x949c, 0xffffffff, 0x100,
+	0x9490, 0xffffffff, 0x7,
+	0x949c, 0xffffffff, 0x100,
+	0x9490, 0xffffffff, 0x8000,
+	0x9604, 0xffffffff, 0x0,
+	0x9654, 0xffffffff, 0x100,
+	0x9604, 0xffffffff, 0x1,
+	0x9654, 0xffffffff, 0x100,
+	0x9604, 0xffffffff, 0x2,
+	0x9654, 0xffffffff, 0x100,
+	0x9604, 0xffffffff, 0x3,
+	0x9654, 0xffffffff, 0x100,
+	0x9604, 0xffffffff, 0x4,
+	0x9654, 0xffffffff, 0x100,
+	0x9604, 0xffffffff, 0x5,
+	0x9654, 0xffffffff, 0x100,
+	0x9604, 0xffffffff, 0x6,
+	0x9654, 0xffffffff, 0x100,
+	0x9604, 0xffffffff, 0x7,
+	0x9654, 0xffffffff, 0x100,
+	0x9604, 0xffffffff, 0x80000000,
+	0x9030, 0xffffffff, 0x100,
+	0x9034, 0xffffffff, 0x100,
+	0x9038, 0xffffffff, 0x100,
+	0x903c, 0xffffffff, 0x100,
+	0x9040, 0xffffffff, 0x100,
+	0xa200, 0xffffffff, 0x100,
+	0xa204, 0xffffffff, 0x100,
+	0xa208, 0xffffffff, 0x100,
+	0xa20c, 0xffffffff, 0x100,
+	0x971c, 0xffffffff, 0x100,
+	0x915c, 0xffffffff, 0x00020001,
+	0x9160, 0xffffffff, 0x00040003,
+	0x916c, 0xffffffff, 0x00060005,
+	0x9170, 0xffffffff, 0x00080007,
+	0x9174, 0xffffffff, 0x000a0009,
+	0x9178, 0xffffffff, 0x000c000b,
+	0x917c, 0xffffffff, 0x000e000d,
+	0x9180, 0xffffffff, 0x0010000f,
+	0x918c, 0xffffffff, 0x00120011,
+	0x9190, 0xffffffff, 0x00140013,
+	0x9194, 0xffffffff, 0x00020001,
+	0x9198, 0xffffffff, 0x00040003,
+	0x919c, 0xffffffff, 0x00060005,
+	0x91a8, 0xffffffff, 0x00080007,
+	0x91ac, 0xffffffff, 0x000a0009,
+	0x91b0, 0xffffffff, 0x000c000b,
+	0x91b4, 0xffffffff, 0x000e000d,
+	0x91b8, 0xffffffff, 0x0010000f,
+	0x91c4, 0xffffffff, 0x00120011,
+	0x91c8, 0xffffffff, 0x00140013,
+	0x91cc, 0xffffffff, 0x00020001,
+	0x91d0, 0xffffffff, 0x00040003,
+	0x91d4, 0xffffffff, 0x00060005,
+	0x91e0, 0xffffffff, 0x00080007,
+	0x91e4, 0xffffffff, 0x000a0009,
+	0x91e8, 0xffffffff, 0x000c000b,
+	0x91ec, 0xffffffff, 0x00020001,
+	0x91f0, 0xffffffff, 0x00040003,
+	0x91f4, 0xffffffff, 0x00060005,
+	0x9200, 0xffffffff, 0x00080007,
+	0x9204, 0xffffffff, 0x000a0009,
+	0x9208, 0xffffffff, 0x000c000b,
+	0x920c, 0xffffffff, 0x000e000d,
+	0x9210, 0xffffffff, 0x0010000f,
+	0x921c, 0xffffffff, 0x00120011,
+	0x9220, 0xffffffff, 0x00140013,
+	0x9224, 0xffffffff, 0x00020001,
+	0x9228, 0xffffffff, 0x00040003,
+	0x922c, 0xffffffff, 0x00060005,
+	0x9238, 0xffffffff, 0x00080007,
+	0x923c, 0xffffffff, 0x000a0009,
+	0x9240, 0xffffffff, 0x000c000b,
+	0x9244, 0xffffffff, 0x000e000d,
+	0x9248, 0xffffffff, 0x0010000f,
+	0x9254, 0xffffffff, 0x00120011,
+	0x9258, 0xffffffff, 0x00140013,
+	0x9294, 0xffffffff, 0x00020001,
+	0x929c, 0xffffffff, 0x00040003,
+	0x92a0, 0xffffffff, 0x00060005,
+	0x92a4, 0xffffffff, 0x00080007
+};
+
+static void rv770_init_golden_registers(struct radeon_device *rdev)
+{
+	switch (rdev->family) {
+	case CHIP_RV770:
+		radeon_program_register_sequence(rdev,
+						 r7xx_golden_registers,
+						 (const u32)ARRAY_SIZE(r7xx_golden_registers));
+		radeon_program_register_sequence(rdev,
+						 r7xx_golden_dyn_gpr_registers,
+						 (const u32)ARRAY_SIZE(r7xx_golden_dyn_gpr_registers));
+		if (rdev->pdev->device == 0x994e)
+			radeon_program_register_sequence(rdev,
+							 rv770ce_golden_registers,
+							 (const u32)ARRAY_SIZE(rv770ce_golden_registers));
+		else
+			radeon_program_register_sequence(rdev,
+							 rv770_golden_registers,
+							 (const u32)ARRAY_SIZE(rv770_golden_registers));
+		radeon_program_register_sequence(rdev,
+						 rv770_mgcg_init,
+						 (const u32)ARRAY_SIZE(rv770_mgcg_init));
+		break;
+	case CHIP_RV730:
+		radeon_program_register_sequence(rdev,
+						 r7xx_golden_registers,
+						 (const u32)ARRAY_SIZE(r7xx_golden_registers));
+		radeon_program_register_sequence(rdev,
+						 r7xx_golden_dyn_gpr_registers,
+						 (const u32)ARRAY_SIZE(r7xx_golden_dyn_gpr_registers));
+		radeon_program_register_sequence(rdev,
+						 rv730_golden_registers,
+						 (const u32)ARRAY_SIZE(rv770_golden_registers));
+		radeon_program_register_sequence(rdev,
+						 rv730_mgcg_init,
+						 (const u32)ARRAY_SIZE(rv770_mgcg_init));
+		break;
+	case CHIP_RV710:
+		radeon_program_register_sequence(rdev,
+						 r7xx_golden_registers,
+						 (const u32)ARRAY_SIZE(r7xx_golden_registers));
+		radeon_program_register_sequence(rdev,
+						 r7xx_golden_dyn_gpr_registers,
+						 (const u32)ARRAY_SIZE(r7xx_golden_dyn_gpr_registers));
+		radeon_program_register_sequence(rdev,
+						 rv710_golden_registers,
+						 (const u32)ARRAY_SIZE(rv770_golden_registers));
+		radeon_program_register_sequence(rdev,
+						 rv710_mgcg_init,
+						 (const u32)ARRAY_SIZE(rv770_mgcg_init));
+		break;
+	case CHIP_RV740:
+		radeon_program_register_sequence(rdev,
+						 rv740_golden_registers,
+						 (const u32)ARRAY_SIZE(rv770_golden_registers));
+		radeon_program_register_sequence(rdev,
+						 rv740_mgcg_init,
+						 (const u32)ARRAY_SIZE(rv770_mgcg_init));
+		break;
+	default:
+		break;
+	}
+}
 
 #define PCIE_BUS_CLK                10000
 #define TCLK                        (PCIE_BUS_CLK / 10)
@@ -68,6 +801,105 @@ u32 rv770_get_xclk(struct radeon_device *rdev)
 	return reference_clock;
 }
 
+int rv770_uvd_resume(struct radeon_device *rdev)
+{
+	uint64_t addr;
+	uint32_t chip_id, size;
+	int r;
+
+	r = radeon_uvd_resume(rdev);
+	if (r)
+		return r;
+
+	/* programm the VCPU memory controller bits 0-27 */
+	addr = rdev->uvd.gpu_addr >> 3;
+	size = RADEON_GPU_PAGE_ALIGN(rdev->uvd_fw->size + 4) >> 3;
+	WREG32(UVD_VCPU_CACHE_OFFSET0, addr);
+	WREG32(UVD_VCPU_CACHE_SIZE0, size);
+
+	addr += size;
+	size = RADEON_UVD_STACK_SIZE >> 3;
+	WREG32(UVD_VCPU_CACHE_OFFSET1, addr);
+	WREG32(UVD_VCPU_CACHE_SIZE1, size);
+
+	addr += size;
+	size = RADEON_UVD_HEAP_SIZE >> 3;
+	WREG32(UVD_VCPU_CACHE_OFFSET2, addr);
+	WREG32(UVD_VCPU_CACHE_SIZE2, size);
+
+	/* bits 28-31 */
+	addr = (rdev->uvd.gpu_addr >> 28) & 0xF;
+	WREG32(UVD_LMI_ADDR_EXT, (addr << 12) | (addr << 0));
+
+	/* bits 32-39 */
+	addr = (rdev->uvd.gpu_addr >> 32) & 0xFF;
+	WREG32(UVD_LMI_EXT40_ADDR, addr | (0x9 << 16) | (0x1 << 31));
+
+	/* tell firmware which hardware it is running on */
+	switch (rdev->family) {
+	default:
+		return -EINVAL;
+	case CHIP_RV710:
+		chip_id = 0x01000005;
+		break;
+	case CHIP_RV730:
+		chip_id = 0x01000006;
+		break;
+	case CHIP_RV740:
+		chip_id = 0x01000007;
+		break;
+	case CHIP_CYPRESS:
+	case CHIP_HEMLOCK:
+		chip_id = 0x01000008;
+		break;
+	case CHIP_JUNIPER:
+		chip_id = 0x01000009;
+		break;
+	case CHIP_REDWOOD:
+		chip_id = 0x0100000a;
+		break;
+	case CHIP_CEDAR:
+		chip_id = 0x0100000b;
+		break;
+	case CHIP_SUMO:
+		chip_id = 0x0100000c;
+		break;
+	case CHIP_SUMO2:
+		chip_id = 0x0100000d;
+		break;
+	case CHIP_PALM:
+		chip_id = 0x0100000e;
+		break;
+	case CHIP_CAYMAN:
+		chip_id = 0x0100000f;
+		break;
+	case CHIP_BARTS:
+		chip_id = 0x01000010;
+		break;
+	case CHIP_TURKS:
+		chip_id = 0x01000011;
+		break;
+	case CHIP_CAICOS:
+		chip_id = 0x01000012;
+		break;
+	case CHIP_TAHITI:
+		chip_id = 0x01000014;
+		break;
+	case CHIP_VERDE:
+		chip_id = 0x01000015;
+		break;
+	case CHIP_PITCAIRN:
+		chip_id = 0x01000016;
+		break;
+	case CHIP_ARUBA:
+		chip_id = 0x01000017;
+		break;
+	}
+	WREG32(UVD_VCPU_CHIP_ID, chip_id);
+
+	return 0;
+}
+
 u32 rv770_page_flip(struct radeon_device *rdev, int crtc_id, u64 crtc_base)
 {
 	struct radeon_crtc *radeon_crtc = rdev->mode_info.crtcs[crtc_id];
@@ -611,6 +1443,11 @@ static void rv770_gpu_init(struct radeon_device *rdev)
 	WREG32(HDP_TILING_CONFIG, (gb_tiling_config & 0xffff));
 	WREG32(DMA_TILING_CONFIG, (gb_tiling_config & 0xffff));
 	WREG32(DMA_TILING_CONFIG2, (gb_tiling_config & 0xffff));
+	if (rdev->family == CHIP_RV730) {
+		WREG32(UVD_UDEC_DB_TILING_CONFIG, (gb_tiling_config & 0xffff));
+		WREG32(UVD_UDEC_DBW_TILING_CONFIG, (gb_tiling_config & 0xffff));
+		WREG32(UVD_UDEC_TILING_CONFIG, (gb_tiling_config & 0xffff));
+	}
 
 	WREG32(CGTS_SYS_TCC_DISABLE, 0);
 	WREG32(CGTS_TCC_DISABLE, 0);
@@ -840,7 +1677,7 @@ void r700_vram_gtt_location(struct radeon_device *rdev, struct radeon_mc *mc)
 	}
 	if (rdev->flags & RADEON_IS_AGP) {
 		size_bf = mc->gtt_start;
-		size_af = 0xFFFFFFFF - mc->gtt_end;
+		size_af = mc->mc_mask - mc->gtt_end;
 		if (size_bf > size_af) {
 			if (mc->mc_vram_size > size_bf) {
 				dev_warn(rdev->dev, "limiting VRAM\n");
@@ -1040,6 +1877,17 @@ static int rv770_startup(struct radeon_device *rdev)
 		return r;
 	}
 
+	r = rv770_uvd_resume(rdev);
+	if (!r) {
+		r = radeon_fence_driver_start_ring(rdev,
+						   R600_RING_TYPE_UVD_INDEX);
+		if (r)
+			dev_err(rdev->dev, "UVD fences init error (%d).\n", r);
+	}
+
+	if (r)
+		rdev->ring[R600_RING_TYPE_UVD_INDEX].ring_size = 0;
+
 	/* Enable IRQ */
 	r = r600_irq_init(rdev);
 	if (r) {
@@ -1074,6 +1922,19 @@ static int rv770_startup(struct radeon_device *rdev)
 	if (r)
 		return r;
 
+	ring = &rdev->ring[R600_RING_TYPE_UVD_INDEX];
+	if (ring->ring_size) {
+		r = radeon_ring_init(rdev, ring, ring->ring_size,
+				     R600_WB_UVD_RPTR_OFFSET,
+				     UVD_RBC_RB_RPTR, UVD_RBC_RB_WPTR,
+				     0, 0xfffff, RADEON_CP_PACKET2);
+		if (!r)
+			r = r600_uvd_init(rdev);
+
+		if (r)
+			DRM_ERROR("radeon: failed initializing UVD (%d).\n", r);
+	}
+
 	r = radeon_ib_pool_init(rdev);
 	if (r) {
 		dev_err(rdev->dev, "IB initialization failed (%d).\n", r);
@@ -1100,6 +1961,9 @@ int rv770_resume(struct radeon_device *rdev)
 	/* post card */
 	atom_asic_init(rdev->mode_info.atom_context);
 
+	/* init golden registers */
+	rv770_init_golden_registers(rdev);
+
 	rdev->accel_working = true;
 	r = rv770_startup(rdev);
 	if (r) {
@@ -1115,6 +1979,7 @@ int rv770_resume(struct radeon_device *rdev)
 int rv770_suspend(struct radeon_device *rdev)
 {
 	r600_audio_fini(rdev);
+	radeon_uvd_suspend(rdev);
 	r700_cp_stop(rdev);
 	r600_dma_stop(rdev);
 	r600_irq_suspend(rdev);
@@ -1156,6 +2021,8 @@ int rv770_init(struct radeon_device *rdev)
 		DRM_INFO("GPU not posted. posting now...\n");
 		atom_asic_init(rdev->mode_info.atom_context);
 	}
+	/* init golden registers */
+	rv770_init_golden_registers(rdev);
 	/* Initialize scratch registers */
 	r600_scratch_init(rdev);
 	/* Initialize surface registers */
@@ -1190,6 +2057,13 @@ int rv770_init(struct radeon_device *rdev)
 	rdev->ring[R600_RING_TYPE_DMA_INDEX].ring_obj = NULL;
 	r600_ring_init(rdev, &rdev->ring[R600_RING_TYPE_DMA_INDEX], 64 * 1024);
 
+	r = radeon_uvd_init(rdev);
+	if (!r) {
+		rdev->ring[R600_RING_TYPE_UVD_INDEX].ring_obj = NULL;
+		r600_ring_init(rdev, &rdev->ring[R600_RING_TYPE_UVD_INDEX],
+			       4096);
+	}
+
 	rdev->ih.ring_obj = NULL;
 	r600_ih_ring_init(rdev, 64 * 1024);
 
@@ -1224,6 +2098,7 @@ void rv770_fini(struct radeon_device *rdev)
 	radeon_ib_pool_fini(rdev);
 	radeon_irq_kms_fini(rdev);
 	rv770_pcie_gart_fini(rdev);
+	radeon_uvd_fini(rdev);
 	r600_vram_scratch_fini(rdev);
 	radeon_gem_fini(rdev);
 	radeon_fence_driver_fini(rdev);
@@ -1264,23 +2139,23 @@ static void rv770_pcie_gen2_enable(struct radeon_device *rdev)
 	DRM_INFO("enabling PCIE gen 2 link speeds, disable with radeon.pcie_gen2=0\n");
 
 	/* advertise upconfig capability */
-	link_width_cntl = RREG32_PCIE_P(PCIE_LC_LINK_WIDTH_CNTL);
+	link_width_cntl = RREG32_PCIE_PORT(PCIE_LC_LINK_WIDTH_CNTL);
 	link_width_cntl &= ~LC_UPCONFIGURE_DIS;
-	WREG32_PCIE_P(PCIE_LC_LINK_WIDTH_CNTL, link_width_cntl);
-	link_width_cntl = RREG32_PCIE_P(PCIE_LC_LINK_WIDTH_CNTL);
+	WREG32_PCIE_PORT(PCIE_LC_LINK_WIDTH_CNTL, link_width_cntl);
+	link_width_cntl = RREG32_PCIE_PORT(PCIE_LC_LINK_WIDTH_CNTL);
 	if (link_width_cntl & LC_RENEGOTIATION_SUPPORT) {
 		lanes = (link_width_cntl & LC_LINK_WIDTH_RD_MASK) >> LC_LINK_WIDTH_RD_SHIFT;
 		link_width_cntl &= ~(LC_LINK_WIDTH_MASK |
 				     LC_RECONFIG_ARC_MISSING_ESCAPE);
 		link_width_cntl |= lanes | LC_RECONFIG_NOW |
 			LC_RENEGOTIATE_EN | LC_UPCONFIGURE_SUPPORT;
-		WREG32_PCIE_P(PCIE_LC_LINK_WIDTH_CNTL, link_width_cntl);
+		WREG32_PCIE_PORT(PCIE_LC_LINK_WIDTH_CNTL, link_width_cntl);
 	} else {
 		link_width_cntl |= LC_UPCONFIGURE_DIS;
-		WREG32_PCIE_P(PCIE_LC_LINK_WIDTH_CNTL, link_width_cntl);
+		WREG32_PCIE_PORT(PCIE_LC_LINK_WIDTH_CNTL, link_width_cntl);
 	}
 
-	speed_cntl = RREG32_PCIE_P(PCIE_LC_SPEED_CNTL);
+	speed_cntl = RREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL);
 	if ((speed_cntl & LC_OTHER_SIDE_EVER_SENT_GEN2) &&
 	    (speed_cntl & LC_OTHER_SIDE_SUPPORTS_GEN2)) {
 
@@ -1293,29 +2168,29 @@ static void rv770_pcie_gen2_enable(struct radeon_device *rdev)
 		WREG16(0x4088, link_cntl2);
 		WREG32(MM_CFGREGS_CNTL, 0);
 
-		speed_cntl = RREG32_PCIE_P(PCIE_LC_SPEED_CNTL);
+		speed_cntl = RREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL);
 		speed_cntl &= ~LC_TARGET_LINK_SPEED_OVERRIDE_EN;
-		WREG32_PCIE_P(PCIE_LC_SPEED_CNTL, speed_cntl);
+		WREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL, speed_cntl);
 
-		speed_cntl = RREG32_PCIE_P(PCIE_LC_SPEED_CNTL);
+		speed_cntl = RREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL);
 		speed_cntl |= LC_CLR_FAILED_SPD_CHANGE_CNT;
-		WREG32_PCIE_P(PCIE_LC_SPEED_CNTL, speed_cntl);
+		WREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL, speed_cntl);
 
-		speed_cntl = RREG32_PCIE_P(PCIE_LC_SPEED_CNTL);
+		speed_cntl = RREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL);
 		speed_cntl &= ~LC_CLR_FAILED_SPD_CHANGE_CNT;
-		WREG32_PCIE_P(PCIE_LC_SPEED_CNTL, speed_cntl);
+		WREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL, speed_cntl);
 
-		speed_cntl = RREG32_PCIE_P(PCIE_LC_SPEED_CNTL);
+		speed_cntl = RREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL);
 		speed_cntl |= LC_GEN2_EN_STRAP;
-		WREG32_PCIE_P(PCIE_LC_SPEED_CNTL, speed_cntl);
+		WREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL, speed_cntl);
 
 	} else {
-		link_width_cntl = RREG32_PCIE_P(PCIE_LC_LINK_WIDTH_CNTL);
+		link_width_cntl = RREG32_PCIE_PORT(PCIE_LC_LINK_WIDTH_CNTL);
 		/* XXX: only disable it if gen1 bridge vendor == 0x111d or 0x1106 */
 		if (1)
 			link_width_cntl |= LC_UPCONFIGURE_DIS;
 		else
 			link_width_cntl &= ~LC_UPCONFIGURE_DIS;
-		WREG32_PCIE_P(PCIE_LC_LINK_WIDTH_CNTL, link_width_cntl);
+		WREG32_PCIE_PORT(PCIE_LC_LINK_WIDTH_CNTL, link_width_cntl);
 	}
 }
diff --git a/drivers/gpu/drm/radeon/rv770d.h b/drivers/gpu/drm/radeon/rv770d.h
index c55f950a4af7..85b16266f748 100644
--- a/drivers/gpu/drm/radeon/rv770d.h
+++ b/drivers/gpu/drm/radeon/rv770d.h
@@ -38,6 +38,30 @@
 #define R7XX_MAX_PIPES             8
 #define R7XX_MAX_PIPES_MASK        0xff
 
+/* discrete uvd clocks */
+#define CG_UPLL_FUNC_CNTL				0x718
+#	define UPLL_RESET_MASK				0x00000001
+#	define UPLL_SLEEP_MASK				0x00000002
+#	define UPLL_BYPASS_EN_MASK			0x00000004
+#	define UPLL_CTLREQ_MASK				0x00000008
+#	define UPLL_REF_DIV(x)				((x) << 16)
+#	define UPLL_REF_DIV_MASK			0x003F0000
+#	define UPLL_CTLACK_MASK				0x40000000
+#	define UPLL_CTLACK2_MASK			0x80000000
+#define CG_UPLL_FUNC_CNTL_2				0x71c
+#	define UPLL_SW_HILEN(x)				((x) << 0)
+#	define UPLL_SW_LOLEN(x)				((x) << 4)
+#	define UPLL_SW_HILEN2(x)			((x) << 8)
+#	define UPLL_SW_LOLEN2(x)			((x) << 12)
+#	define UPLL_SW_MASK				0x0000FFFF
+#	define VCLK_SRC_SEL(x)				((x) << 20)
+#	define VCLK_SRC_SEL_MASK			0x01F00000
+#	define DCLK_SRC_SEL(x)				((x) << 25)
+#	define DCLK_SRC_SEL_MASK			0x3E000000
+#define CG_UPLL_FUNC_CNTL_3				0x720
+#	define UPLL_FB_DIV(x)				((x) << 0)
+#	define UPLL_FB_DIV_MASK				0x01FFFFFF
+
 /* Registers */
 #define	CB_COLOR0_BASE					0x28040
 #define	CB_COLOR1_BASE					0x28044
@@ -112,6 +136,11 @@
 #define DMA_TILING_CONFIG                               0x3ec8
 #define DMA_TILING_CONFIG2                              0xd0b8
 
+/* RV730 only */
+#define UVD_UDEC_TILING_CONFIG                          0xef40
+#define UVD_UDEC_DB_TILING_CONFIG                       0xef44
+#define UVD_UDEC_DBW_TILING_CONFIG                      0xef48
+
 #define	GC_USER_SHADER_PIPE_CONFIG			0x8954
 #define		INACTIVE_QD_PIPES(x)				((x) << 8)
 #define		INACTIVE_QD_PIPES_MASK				0x0000FF00
@@ -671,4 +700,18 @@
 #       define TARGET_LINK_SPEED_MASK                     (0xf << 0)
 #       define SELECTABLE_DEEMPHASIS                      (1 << 6)
 
+/* UVD */
+#define UVD_LMI_EXT40_ADDR				0xf498
+#define UVD_VCPU_CHIP_ID				0xf4d4
+#define UVD_VCPU_CACHE_OFFSET0				0xf4d8
+#define UVD_VCPU_CACHE_SIZE0				0xf4dc
+#define UVD_VCPU_CACHE_OFFSET1				0xf4e0
+#define UVD_VCPU_CACHE_SIZE1				0xf4e4
+#define UVD_VCPU_CACHE_OFFSET2				0xf4e8
+#define UVD_VCPU_CACHE_SIZE2				0xf4ec
+#define UVD_LMI_ADDR_EXT				0xf594
+
+#define UVD_RBC_RB_RPTR					0xf690
+#define UVD_RBC_RB_WPTR					0xf694
+
 #endif
diff --git a/drivers/gpu/drm/radeon/si.c b/drivers/gpu/drm/radeon/si.c
index bafbe3216952..f0b6c2f87c4d 100644
--- a/drivers/gpu/drm/radeon/si.c
+++ b/drivers/gpu/drm/radeon/si.c
@@ -70,6 +70,794 @@ extern u32 evergreen_get_number_of_dram_channels(struct radeon_device *rdev);
 extern void evergreen_print_gpu_status_regs(struct radeon_device *rdev);
 extern bool evergreen_is_display_hung(struct radeon_device *rdev);
 
+static const u32 tahiti_golden_rlc_registers[] =
+{
+	0xc424, 0xffffffff, 0x00601005,
+	0xc47c, 0xffffffff, 0x10104040,
+	0xc488, 0xffffffff, 0x0100000a,
+	0xc314, 0xffffffff, 0x00000800,
+	0xc30c, 0xffffffff, 0x800000f4,
+	0xf4a8, 0xffffffff, 0x00000000
+};
+
+static const u32 tahiti_golden_registers[] =
+{
+	0x9a10, 0x00010000, 0x00018208,
+	0x9830, 0xffffffff, 0x00000000,
+	0x9834, 0xf00fffff, 0x00000400,
+	0x9838, 0x0002021c, 0x00020200,
+	0xc78, 0x00000080, 0x00000000,
+	0xd030, 0x000300c0, 0x00800040,
+	0xd830, 0x000300c0, 0x00800040,
+	0x5bb0, 0x000000f0, 0x00000070,
+	0x5bc0, 0x00200000, 0x50100000,
+	0x7030, 0x31000311, 0x00000011,
+	0x277c, 0x00000003, 0x000007ff,
+	0x240c, 0x000007ff, 0x00000000,
+	0x8a14, 0xf000001f, 0x00000007,
+	0x8b24, 0xffffffff, 0x00ffffff,
+	0x8b10, 0x0000ff0f, 0x00000000,
+	0x28a4c, 0x07ffffff, 0x4e000000,
+	0x28350, 0x3f3f3fff, 0x2a00126a,
+	0x30, 0x000000ff, 0x0040,
+	0x34, 0x00000040, 0x00004040,
+	0x9100, 0x07ffffff, 0x03000000,
+	0x8e88, 0x01ff1f3f, 0x00000000,
+	0x8e84, 0x01ff1f3f, 0x00000000,
+	0x9060, 0x0000007f, 0x00000020,
+	0x9508, 0x00010000, 0x00010000,
+	0xac14, 0x00000200, 0x000002fb,
+	0xac10, 0xffffffff, 0x0000543b,
+	0xac0c, 0xffffffff, 0xa9210876,
+	0x88d0, 0xffffffff, 0x000fff40,
+	0x88d4, 0x0000001f, 0x00000010,
+	0x1410, 0x20000000, 0x20fffed8,
+	0x15c0, 0x000c0fc0, 0x000c0400
+};
+
+static const u32 tahiti_golden_registers2[] =
+{
+	0xc64, 0x00000001, 0x00000001
+};
+
+static const u32 pitcairn_golden_rlc_registers[] =
+{
+	0xc424, 0xffffffff, 0x00601004,
+	0xc47c, 0xffffffff, 0x10102020,
+	0xc488, 0xffffffff, 0x01000020,
+	0xc314, 0xffffffff, 0x00000800,
+	0xc30c, 0xffffffff, 0x800000a4
+};
+
+static const u32 pitcairn_golden_registers[] =
+{
+	0x9a10, 0x00010000, 0x00018208,
+	0x9830, 0xffffffff, 0x00000000,
+	0x9834, 0xf00fffff, 0x00000400,
+	0x9838, 0x0002021c, 0x00020200,
+	0xc78, 0x00000080, 0x00000000,
+	0xd030, 0x000300c0, 0x00800040,
+	0xd830, 0x000300c0, 0x00800040,
+	0x5bb0, 0x000000f0, 0x00000070,
+	0x5bc0, 0x00200000, 0x50100000,
+	0x7030, 0x31000311, 0x00000011,
+	0x2ae4, 0x00073ffe, 0x000022a2,
+	0x240c, 0x000007ff, 0x00000000,
+	0x8a14, 0xf000001f, 0x00000007,
+	0x8b24, 0xffffffff, 0x00ffffff,
+	0x8b10, 0x0000ff0f, 0x00000000,
+	0x28a4c, 0x07ffffff, 0x4e000000,
+	0x28350, 0x3f3f3fff, 0x2a00126a,
+	0x30, 0x000000ff, 0x0040,
+	0x34, 0x00000040, 0x00004040,
+	0x9100, 0x07ffffff, 0x03000000,
+	0x9060, 0x0000007f, 0x00000020,
+	0x9508, 0x00010000, 0x00010000,
+	0xac14, 0x000003ff, 0x000000f7,
+	0xac10, 0xffffffff, 0x00000000,
+	0xac0c, 0xffffffff, 0x32761054,
+	0x88d4, 0x0000001f, 0x00000010,
+	0x15c0, 0x000c0fc0, 0x000c0400
+};
+
+static const u32 verde_golden_rlc_registers[] =
+{
+	0xc424, 0xffffffff, 0x033f1005,
+	0xc47c, 0xffffffff, 0x10808020,
+	0xc488, 0xffffffff, 0x00800008,
+	0xc314, 0xffffffff, 0x00001000,
+	0xc30c, 0xffffffff, 0x80010014
+};
+
+static const u32 verde_golden_registers[] =
+{
+	0x9a10, 0x00010000, 0x00018208,
+	0x9830, 0xffffffff, 0x00000000,
+	0x9834, 0xf00fffff, 0x00000400,
+	0x9838, 0x0002021c, 0x00020200,
+	0xc78, 0x00000080, 0x00000000,
+	0xd030, 0x000300c0, 0x00800040,
+	0xd030, 0x000300c0, 0x00800040,
+	0xd830, 0x000300c0, 0x00800040,
+	0xd830, 0x000300c0, 0x00800040,
+	0x5bb0, 0x000000f0, 0x00000070,
+	0x5bc0, 0x00200000, 0x50100000,
+	0x7030, 0x31000311, 0x00000011,
+	0x2ae4, 0x00073ffe, 0x000022a2,
+	0x2ae4, 0x00073ffe, 0x000022a2,
+	0x2ae4, 0x00073ffe, 0x000022a2,
+	0x240c, 0x000007ff, 0x00000000,
+	0x240c, 0x000007ff, 0x00000000,
+	0x240c, 0x000007ff, 0x00000000,
+	0x8a14, 0xf000001f, 0x00000007,
+	0x8a14, 0xf000001f, 0x00000007,
+	0x8a14, 0xf000001f, 0x00000007,
+	0x8b24, 0xffffffff, 0x00ffffff,
+	0x8b10, 0x0000ff0f, 0x00000000,
+	0x28a4c, 0x07ffffff, 0x4e000000,
+	0x28350, 0x3f3f3fff, 0x0000124a,
+	0x28350, 0x3f3f3fff, 0x0000124a,
+	0x28350, 0x3f3f3fff, 0x0000124a,
+	0x30, 0x000000ff, 0x0040,
+	0x34, 0x00000040, 0x00004040,
+	0x9100, 0x07ffffff, 0x03000000,
+	0x9100, 0x07ffffff, 0x03000000,
+	0x8e88, 0x01ff1f3f, 0x00000000,
+	0x8e88, 0x01ff1f3f, 0x00000000,
+	0x8e88, 0x01ff1f3f, 0x00000000,
+	0x8e84, 0x01ff1f3f, 0x00000000,
+	0x8e84, 0x01ff1f3f, 0x00000000,
+	0x8e84, 0x01ff1f3f, 0x00000000,
+	0x9060, 0x0000007f, 0x00000020,
+	0x9508, 0x00010000, 0x00010000,
+	0xac14, 0x000003ff, 0x00000003,
+	0xac14, 0x000003ff, 0x00000003,
+	0xac14, 0x000003ff, 0x00000003,
+	0xac10, 0xffffffff, 0x00000000,
+	0xac10, 0xffffffff, 0x00000000,
+	0xac10, 0xffffffff, 0x00000000,
+	0xac0c, 0xffffffff, 0x00001032,
+	0xac0c, 0xffffffff, 0x00001032,
+	0xac0c, 0xffffffff, 0x00001032,
+	0x88d4, 0x0000001f, 0x00000010,
+	0x88d4, 0x0000001f, 0x00000010,
+	0x88d4, 0x0000001f, 0x00000010,
+	0x15c0, 0x000c0fc0, 0x000c0400
+};
+
+static const u32 oland_golden_rlc_registers[] =
+{
+	0xc424, 0xffffffff, 0x00601005,
+	0xc47c, 0xffffffff, 0x10104040,
+	0xc488, 0xffffffff, 0x0100000a,
+	0xc314, 0xffffffff, 0x00000800,
+	0xc30c, 0xffffffff, 0x800000f4
+};
+
+static const u32 oland_golden_registers[] =
+{
+	0x9a10, 0x00010000, 0x00018208,
+	0x9830, 0xffffffff, 0x00000000,
+	0x9834, 0xf00fffff, 0x00000400,
+	0x9838, 0x0002021c, 0x00020200,
+	0xc78, 0x00000080, 0x00000000,
+	0xd030, 0x000300c0, 0x00800040,
+	0xd830, 0x000300c0, 0x00800040,
+	0x5bb0, 0x000000f0, 0x00000070,
+	0x5bc0, 0x00200000, 0x50100000,
+	0x7030, 0x31000311, 0x00000011,
+	0x2ae4, 0x00073ffe, 0x000022a2,
+	0x240c, 0x000007ff, 0x00000000,
+	0x8a14, 0xf000001f, 0x00000007,
+	0x8b24, 0xffffffff, 0x00ffffff,
+	0x8b10, 0x0000ff0f, 0x00000000,
+	0x28a4c, 0x07ffffff, 0x4e000000,
+	0x28350, 0x3f3f3fff, 0x00000082,
+	0x30, 0x000000ff, 0x0040,
+	0x34, 0x00000040, 0x00004040,
+	0x9100, 0x07ffffff, 0x03000000,
+	0x9060, 0x0000007f, 0x00000020,
+	0x9508, 0x00010000, 0x00010000,
+	0xac14, 0x000003ff, 0x000000f3,
+	0xac10, 0xffffffff, 0x00000000,
+	0xac0c, 0xffffffff, 0x00003210,
+	0x88d4, 0x0000001f, 0x00000010,
+	0x15c0, 0x000c0fc0, 0x000c0400
+};
+
+static const u32 tahiti_mgcg_cgcg_init[] =
+{
+	0xc400, 0xffffffff, 0xfffffffc,
+	0x802c, 0xffffffff, 0xe0000000,
+	0x9a60, 0xffffffff, 0x00000100,
+	0x92a4, 0xffffffff, 0x00000100,
+	0xc164, 0xffffffff, 0x00000100,
+	0x9774, 0xffffffff, 0x00000100,
+	0x8984, 0xffffffff, 0x06000100,
+	0x8a18, 0xffffffff, 0x00000100,
+	0x92a0, 0xffffffff, 0x00000100,
+	0xc380, 0xffffffff, 0x00000100,
+	0x8b28, 0xffffffff, 0x00000100,
+	0x9144, 0xffffffff, 0x00000100,
+	0x8d88, 0xffffffff, 0x00000100,
+	0x8d8c, 0xffffffff, 0x00000100,
+	0x9030, 0xffffffff, 0x00000100,
+	0x9034, 0xffffffff, 0x00000100,
+	0x9038, 0xffffffff, 0x00000100,
+	0x903c, 0xffffffff, 0x00000100,
+	0xad80, 0xffffffff, 0x00000100,
+	0xac54, 0xffffffff, 0x00000100,
+	0x897c, 0xffffffff, 0x06000100,
+	0x9868, 0xffffffff, 0x00000100,
+	0x9510, 0xffffffff, 0x00000100,
+	0xaf04, 0xffffffff, 0x00000100,
+	0xae04, 0xffffffff, 0x00000100,
+	0x949c, 0xffffffff, 0x00000100,
+	0x802c, 0xffffffff, 0xe0000000,
+	0x9160, 0xffffffff, 0x00010000,
+	0x9164, 0xffffffff, 0x00030002,
+	0x9168, 0xffffffff, 0x00040007,
+	0x916c, 0xffffffff, 0x00060005,
+	0x9170, 0xffffffff, 0x00090008,
+	0x9174, 0xffffffff, 0x00020001,
+	0x9178, 0xffffffff, 0x00040003,
+	0x917c, 0xffffffff, 0x00000007,
+	0x9180, 0xffffffff, 0x00060005,
+	0x9184, 0xffffffff, 0x00090008,
+	0x9188, 0xffffffff, 0x00030002,
+	0x918c, 0xffffffff, 0x00050004,
+	0x9190, 0xffffffff, 0x00000008,
+	0x9194, 0xffffffff, 0x00070006,
+	0x9198, 0xffffffff, 0x000a0009,
+	0x919c, 0xffffffff, 0x00040003,
+	0x91a0, 0xffffffff, 0x00060005,
+	0x91a4, 0xffffffff, 0x00000009,
+	0x91a8, 0xffffffff, 0x00080007,
+	0x91ac, 0xffffffff, 0x000b000a,
+	0x91b0, 0xffffffff, 0x00050004,
+	0x91b4, 0xffffffff, 0x00070006,
+	0x91b8, 0xffffffff, 0x0008000b,
+	0x91bc, 0xffffffff, 0x000a0009,
+	0x91c0, 0xffffffff, 0x000d000c,
+	0x91c4, 0xffffffff, 0x00060005,
+	0x91c8, 0xffffffff, 0x00080007,
+	0x91cc, 0xffffffff, 0x0000000b,
+	0x91d0, 0xffffffff, 0x000a0009,
+	0x91d4, 0xffffffff, 0x000d000c,
+	0x91d8, 0xffffffff, 0x00070006,
+	0x91dc, 0xffffffff, 0x00090008,
+	0x91e0, 0xffffffff, 0x0000000c,
+	0x91e4, 0xffffffff, 0x000b000a,
+	0x91e8, 0xffffffff, 0x000e000d,
+	0x91ec, 0xffffffff, 0x00080007,
+	0x91f0, 0xffffffff, 0x000a0009,
+	0x91f4, 0xffffffff, 0x0000000d,
+	0x91f8, 0xffffffff, 0x000c000b,
+	0x91fc, 0xffffffff, 0x000f000e,
+	0x9200, 0xffffffff, 0x00090008,
+	0x9204, 0xffffffff, 0x000b000a,
+	0x9208, 0xffffffff, 0x000c000f,
+	0x920c, 0xffffffff, 0x000e000d,
+	0x9210, 0xffffffff, 0x00110010,
+	0x9214, 0xffffffff, 0x000a0009,
+	0x9218, 0xffffffff, 0x000c000b,
+	0x921c, 0xffffffff, 0x0000000f,
+	0x9220, 0xffffffff, 0x000e000d,
+	0x9224, 0xffffffff, 0x00110010,
+	0x9228, 0xffffffff, 0x000b000a,
+	0x922c, 0xffffffff, 0x000d000c,
+	0x9230, 0xffffffff, 0x00000010,
+	0x9234, 0xffffffff, 0x000f000e,
+	0x9238, 0xffffffff, 0x00120011,
+	0x923c, 0xffffffff, 0x000c000b,
+	0x9240, 0xffffffff, 0x000e000d,
+	0x9244, 0xffffffff, 0x00000011,
+	0x9248, 0xffffffff, 0x0010000f,
+	0x924c, 0xffffffff, 0x00130012,
+	0x9250, 0xffffffff, 0x000d000c,
+	0x9254, 0xffffffff, 0x000f000e,
+	0x9258, 0xffffffff, 0x00100013,
+	0x925c, 0xffffffff, 0x00120011,
+	0x9260, 0xffffffff, 0x00150014,
+	0x9264, 0xffffffff, 0x000e000d,
+	0x9268, 0xffffffff, 0x0010000f,
+	0x926c, 0xffffffff, 0x00000013,
+	0x9270, 0xffffffff, 0x00120011,
+	0x9274, 0xffffffff, 0x00150014,
+	0x9278, 0xffffffff, 0x000f000e,
+	0x927c, 0xffffffff, 0x00110010,
+	0x9280, 0xffffffff, 0x00000014,
+	0x9284, 0xffffffff, 0x00130012,
+	0x9288, 0xffffffff, 0x00160015,
+	0x928c, 0xffffffff, 0x0010000f,
+	0x9290, 0xffffffff, 0x00120011,
+	0x9294, 0xffffffff, 0x00000015,
+	0x9298, 0xffffffff, 0x00140013,
+	0x929c, 0xffffffff, 0x00170016,
+	0x9150, 0xffffffff, 0x96940200,
+	0x8708, 0xffffffff, 0x00900100,
+	0xc478, 0xffffffff, 0x00000080,
+	0xc404, 0xffffffff, 0x0020003f,
+	0x30, 0xffffffff, 0x0000001c,
+	0x34, 0x000f0000, 0x000f0000,
+	0x160c, 0xffffffff, 0x00000100,
+	0x1024, 0xffffffff, 0x00000100,
+	0x102c, 0x00000101, 0x00000000,
+	0x20a8, 0xffffffff, 0x00000104,
+	0x264c, 0x000c0000, 0x000c0000,
+	0x2648, 0x000c0000, 0x000c0000,
+	0x55e4, 0xff000fff, 0x00000100,
+	0x55e8, 0x00000001, 0x00000001,
+	0x2f50, 0x00000001, 0x00000001,
+	0x30cc, 0xc0000fff, 0x00000104,
+	0xc1e4, 0x00000001, 0x00000001,
+	0xd0c0, 0xfffffff0, 0x00000100,
+	0xd8c0, 0xfffffff0, 0x00000100
+};
+
+static const u32 pitcairn_mgcg_cgcg_init[] =
+{
+	0xc400, 0xffffffff, 0xfffffffc,
+	0x802c, 0xffffffff, 0xe0000000,
+	0x9a60, 0xffffffff, 0x00000100,
+	0x92a4, 0xffffffff, 0x00000100,
+	0xc164, 0xffffffff, 0x00000100,
+	0x9774, 0xffffffff, 0x00000100,
+	0x8984, 0xffffffff, 0x06000100,
+	0x8a18, 0xffffffff, 0x00000100,
+	0x92a0, 0xffffffff, 0x00000100,
+	0xc380, 0xffffffff, 0x00000100,
+	0x8b28, 0xffffffff, 0x00000100,
+	0x9144, 0xffffffff, 0x00000100,
+	0x8d88, 0xffffffff, 0x00000100,
+	0x8d8c, 0xffffffff, 0x00000100,
+	0x9030, 0xffffffff, 0x00000100,
+	0x9034, 0xffffffff, 0x00000100,
+	0x9038, 0xffffffff, 0x00000100,
+	0x903c, 0xffffffff, 0x00000100,
+	0xad80, 0xffffffff, 0x00000100,
+	0xac54, 0xffffffff, 0x00000100,
+	0x897c, 0xffffffff, 0x06000100,
+	0x9868, 0xffffffff, 0x00000100,
+	0x9510, 0xffffffff, 0x00000100,
+	0xaf04, 0xffffffff, 0x00000100,
+	0xae04, 0xffffffff, 0x00000100,
+	0x949c, 0xffffffff, 0x00000100,
+	0x802c, 0xffffffff, 0xe0000000,
+	0x9160, 0xffffffff, 0x00010000,
+	0x9164, 0xffffffff, 0x00030002,
+	0x9168, 0xffffffff, 0x00040007,
+	0x916c, 0xffffffff, 0x00060005,
+	0x9170, 0xffffffff, 0x00090008,
+	0x9174, 0xffffffff, 0x00020001,
+	0x9178, 0xffffffff, 0x00040003,
+	0x917c, 0xffffffff, 0x00000007,
+	0x9180, 0xffffffff, 0x00060005,
+	0x9184, 0xffffffff, 0x00090008,
+	0x9188, 0xffffffff, 0x00030002,
+	0x918c, 0xffffffff, 0x00050004,
+	0x9190, 0xffffffff, 0x00000008,
+	0x9194, 0xffffffff, 0x00070006,
+	0x9198, 0xffffffff, 0x000a0009,
+	0x919c, 0xffffffff, 0x00040003,
+	0x91a0, 0xffffffff, 0x00060005,
+	0x91a4, 0xffffffff, 0x00000009,
+	0x91a8, 0xffffffff, 0x00080007,
+	0x91ac, 0xffffffff, 0x000b000a,
+	0x91b0, 0xffffffff, 0x00050004,
+	0x91b4, 0xffffffff, 0x00070006,
+	0x91b8, 0xffffffff, 0x0008000b,
+	0x91bc, 0xffffffff, 0x000a0009,
+	0x91c0, 0xffffffff, 0x000d000c,
+	0x9200, 0xffffffff, 0x00090008,
+	0x9204, 0xffffffff, 0x000b000a,
+	0x9208, 0xffffffff, 0x000c000f,
+	0x920c, 0xffffffff, 0x000e000d,
+	0x9210, 0xffffffff, 0x00110010,
+	0x9214, 0xffffffff, 0x000a0009,
+	0x9218, 0xffffffff, 0x000c000b,
+	0x921c, 0xffffffff, 0x0000000f,
+	0x9220, 0xffffffff, 0x000e000d,
+	0x9224, 0xffffffff, 0x00110010,
+	0x9228, 0xffffffff, 0x000b000a,
+	0x922c, 0xffffffff, 0x000d000c,
+	0x9230, 0xffffffff, 0x00000010,
+	0x9234, 0xffffffff, 0x000f000e,
+	0x9238, 0xffffffff, 0x00120011,
+	0x923c, 0xffffffff, 0x000c000b,
+	0x9240, 0xffffffff, 0x000e000d,
+	0x9244, 0xffffffff, 0x00000011,
+	0x9248, 0xffffffff, 0x0010000f,
+	0x924c, 0xffffffff, 0x00130012,
+	0x9250, 0xffffffff, 0x000d000c,
+	0x9254, 0xffffffff, 0x000f000e,
+	0x9258, 0xffffffff, 0x00100013,
+	0x925c, 0xffffffff, 0x00120011,
+	0x9260, 0xffffffff, 0x00150014,
+	0x9150, 0xffffffff, 0x96940200,
+	0x8708, 0xffffffff, 0x00900100,
+	0xc478, 0xffffffff, 0x00000080,
+	0xc404, 0xffffffff, 0x0020003f,
+	0x30, 0xffffffff, 0x0000001c,
+	0x34, 0x000f0000, 0x000f0000,
+	0x160c, 0xffffffff, 0x00000100,
+	0x1024, 0xffffffff, 0x00000100,
+	0x102c, 0x00000101, 0x00000000,
+	0x20a8, 0xffffffff, 0x00000104,
+	0x55e4, 0xff000fff, 0x00000100,
+	0x55e8, 0x00000001, 0x00000001,
+	0x2f50, 0x00000001, 0x00000001,
+	0x30cc, 0xc0000fff, 0x00000104,
+	0xc1e4, 0x00000001, 0x00000001,
+	0xd0c0, 0xfffffff0, 0x00000100,
+	0xd8c0, 0xfffffff0, 0x00000100
+};
+
+static const u32 verde_mgcg_cgcg_init[] =
+{
+	0xc400, 0xffffffff, 0xfffffffc,
+	0x802c, 0xffffffff, 0xe0000000,
+	0x9a60, 0xffffffff, 0x00000100,
+	0x92a4, 0xffffffff, 0x00000100,
+	0xc164, 0xffffffff, 0x00000100,
+	0x9774, 0xffffffff, 0x00000100,
+	0x8984, 0xffffffff, 0x06000100,
+	0x8a18, 0xffffffff, 0x00000100,
+	0x92a0, 0xffffffff, 0x00000100,
+	0xc380, 0xffffffff, 0x00000100,
+	0x8b28, 0xffffffff, 0x00000100,
+	0x9144, 0xffffffff, 0x00000100,
+	0x8d88, 0xffffffff, 0x00000100,
+	0x8d8c, 0xffffffff, 0x00000100,
+	0x9030, 0xffffffff, 0x00000100,
+	0x9034, 0xffffffff, 0x00000100,
+	0x9038, 0xffffffff, 0x00000100,
+	0x903c, 0xffffffff, 0x00000100,
+	0xad80, 0xffffffff, 0x00000100,
+	0xac54, 0xffffffff, 0x00000100,
+	0x897c, 0xffffffff, 0x06000100,
+	0x9868, 0xffffffff, 0x00000100,
+	0x9510, 0xffffffff, 0x00000100,
+	0xaf04, 0xffffffff, 0x00000100,
+	0xae04, 0xffffffff, 0x00000100,
+	0x949c, 0xffffffff, 0x00000100,
+	0x802c, 0xffffffff, 0xe0000000,
+	0x9160, 0xffffffff, 0x00010000,
+	0x9164, 0xffffffff, 0x00030002,
+	0x9168, 0xffffffff, 0x00040007,
+	0x916c, 0xffffffff, 0x00060005,
+	0x9170, 0xffffffff, 0x00090008,
+	0x9174, 0xffffffff, 0x00020001,
+	0x9178, 0xffffffff, 0x00040003,
+	0x917c, 0xffffffff, 0x00000007,
+	0x9180, 0xffffffff, 0x00060005,
+	0x9184, 0xffffffff, 0x00090008,
+	0x9188, 0xffffffff, 0x00030002,
+	0x918c, 0xffffffff, 0x00050004,
+	0x9190, 0xffffffff, 0x00000008,
+	0x9194, 0xffffffff, 0x00070006,
+	0x9198, 0xffffffff, 0x000a0009,
+	0x919c, 0xffffffff, 0x00040003,
+	0x91a0, 0xffffffff, 0x00060005,
+	0x91a4, 0xffffffff, 0x00000009,
+	0x91a8, 0xffffffff, 0x00080007,
+	0x91ac, 0xffffffff, 0x000b000a,
+	0x91b0, 0xffffffff, 0x00050004,
+	0x91b4, 0xffffffff, 0x00070006,
+	0x91b8, 0xffffffff, 0x0008000b,
+	0x91bc, 0xffffffff, 0x000a0009,
+	0x91c0, 0xffffffff, 0x000d000c,
+	0x9200, 0xffffffff, 0x00090008,
+	0x9204, 0xffffffff, 0x000b000a,
+	0x9208, 0xffffffff, 0x000c000f,
+	0x920c, 0xffffffff, 0x000e000d,
+	0x9210, 0xffffffff, 0x00110010,
+	0x9214, 0xffffffff, 0x000a0009,
+	0x9218, 0xffffffff, 0x000c000b,
+	0x921c, 0xffffffff, 0x0000000f,
+	0x9220, 0xffffffff, 0x000e000d,
+	0x9224, 0xffffffff, 0x00110010,
+	0x9228, 0xffffffff, 0x000b000a,
+	0x922c, 0xffffffff, 0x000d000c,
+	0x9230, 0xffffffff, 0x00000010,
+	0x9234, 0xffffffff, 0x000f000e,
+	0x9238, 0xffffffff, 0x00120011,
+	0x923c, 0xffffffff, 0x000c000b,
+	0x9240, 0xffffffff, 0x000e000d,
+	0x9244, 0xffffffff, 0x00000011,
+	0x9248, 0xffffffff, 0x0010000f,
+	0x924c, 0xffffffff, 0x00130012,
+	0x9250, 0xffffffff, 0x000d000c,
+	0x9254, 0xffffffff, 0x000f000e,
+	0x9258, 0xffffffff, 0x00100013,
+	0x925c, 0xffffffff, 0x00120011,
+	0x9260, 0xffffffff, 0x00150014,
+	0x9150, 0xffffffff, 0x96940200,
+	0x8708, 0xffffffff, 0x00900100,
+	0xc478, 0xffffffff, 0x00000080,
+	0xc404, 0xffffffff, 0x0020003f,
+	0x30, 0xffffffff, 0x0000001c,
+	0x34, 0x000f0000, 0x000f0000,
+	0x160c, 0xffffffff, 0x00000100,
+	0x1024, 0xffffffff, 0x00000100,
+	0x102c, 0x00000101, 0x00000000,
+	0x20a8, 0xffffffff, 0x00000104,
+	0x264c, 0x000c0000, 0x000c0000,
+	0x2648, 0x000c0000, 0x000c0000,
+	0x55e4, 0xff000fff, 0x00000100,
+	0x55e8, 0x00000001, 0x00000001,
+	0x2f50, 0x00000001, 0x00000001,
+	0x30cc, 0xc0000fff, 0x00000104,
+	0xc1e4, 0x00000001, 0x00000001,
+	0xd0c0, 0xfffffff0, 0x00000100,
+	0xd8c0, 0xfffffff0, 0x00000100
+};
+
+static const u32 oland_mgcg_cgcg_init[] =
+{
+	0xc400, 0xffffffff, 0xfffffffc,
+	0x802c, 0xffffffff, 0xe0000000,
+	0x9a60, 0xffffffff, 0x00000100,
+	0x92a4, 0xffffffff, 0x00000100,
+	0xc164, 0xffffffff, 0x00000100,
+	0x9774, 0xffffffff, 0x00000100,
+	0x8984, 0xffffffff, 0x06000100,
+	0x8a18, 0xffffffff, 0x00000100,
+	0x92a0, 0xffffffff, 0x00000100,
+	0xc380, 0xffffffff, 0x00000100,
+	0x8b28, 0xffffffff, 0x00000100,
+	0x9144, 0xffffffff, 0x00000100,
+	0x8d88, 0xffffffff, 0x00000100,
+	0x8d8c, 0xffffffff, 0x00000100,
+	0x9030, 0xffffffff, 0x00000100,
+	0x9034, 0xffffffff, 0x00000100,
+	0x9038, 0xffffffff, 0x00000100,
+	0x903c, 0xffffffff, 0x00000100,
+	0xad80, 0xffffffff, 0x00000100,
+	0xac54, 0xffffffff, 0x00000100,
+	0x897c, 0xffffffff, 0x06000100,
+	0x9868, 0xffffffff, 0x00000100,
+	0x9510, 0xffffffff, 0x00000100,
+	0xaf04, 0xffffffff, 0x00000100,
+	0xae04, 0xffffffff, 0x00000100,
+	0x949c, 0xffffffff, 0x00000100,
+	0x802c, 0xffffffff, 0xe0000000,
+	0x9160, 0xffffffff, 0x00010000,
+	0x9164, 0xffffffff, 0x00030002,
+	0x9168, 0xffffffff, 0x00040007,
+	0x916c, 0xffffffff, 0x00060005,
+	0x9170, 0xffffffff, 0x00090008,
+	0x9174, 0xffffffff, 0x00020001,
+	0x9178, 0xffffffff, 0x00040003,
+	0x917c, 0xffffffff, 0x00000007,
+	0x9180, 0xffffffff, 0x00060005,
+	0x9184, 0xffffffff, 0x00090008,
+	0x9188, 0xffffffff, 0x00030002,
+	0x918c, 0xffffffff, 0x00050004,
+	0x9190, 0xffffffff, 0x00000008,
+	0x9194, 0xffffffff, 0x00070006,
+	0x9198, 0xffffffff, 0x000a0009,
+	0x919c, 0xffffffff, 0x00040003,
+	0x91a0, 0xffffffff, 0x00060005,
+	0x91a4, 0xffffffff, 0x00000009,
+	0x91a8, 0xffffffff, 0x00080007,
+	0x91ac, 0xffffffff, 0x000b000a,
+	0x91b0, 0xffffffff, 0x00050004,
+	0x91b4, 0xffffffff, 0x00070006,
+	0x91b8, 0xffffffff, 0x0008000b,
+	0x91bc, 0xffffffff, 0x000a0009,
+	0x91c0, 0xffffffff, 0x000d000c,
+	0x91c4, 0xffffffff, 0x00060005,
+	0x91c8, 0xffffffff, 0x00080007,
+	0x91cc, 0xffffffff, 0x0000000b,
+	0x91d0, 0xffffffff, 0x000a0009,
+	0x91d4, 0xffffffff, 0x000d000c,
+	0x9150, 0xffffffff, 0x96940200,
+	0x8708, 0xffffffff, 0x00900100,
+	0xc478, 0xffffffff, 0x00000080,
+	0xc404, 0xffffffff, 0x0020003f,
+	0x30, 0xffffffff, 0x0000001c,
+	0x34, 0x000f0000, 0x000f0000,
+	0x160c, 0xffffffff, 0x00000100,
+	0x1024, 0xffffffff, 0x00000100,
+	0x102c, 0x00000101, 0x00000000,
+	0x20a8, 0xffffffff, 0x00000104,
+	0x264c, 0x000c0000, 0x000c0000,
+	0x2648, 0x000c0000, 0x000c0000,
+	0x55e4, 0xff000fff, 0x00000100,
+	0x55e8, 0x00000001, 0x00000001,
+	0x2f50, 0x00000001, 0x00000001,
+	0x30cc, 0xc0000fff, 0x00000104,
+	0xc1e4, 0x00000001, 0x00000001,
+	0xd0c0, 0xfffffff0, 0x00000100,
+	0xd8c0, 0xfffffff0, 0x00000100
+};
+
+static u32 verde_pg_init[] =
+{
+	0x353c, 0xffffffff, 0x40000,
+	0x3538, 0xffffffff, 0x200010ff,
+	0x353c, 0xffffffff, 0x0,
+	0x353c, 0xffffffff, 0x0,
+	0x353c, 0xffffffff, 0x0,
+	0x353c, 0xffffffff, 0x0,
+	0x353c, 0xffffffff, 0x0,
+	0x353c, 0xffffffff, 0x7007,
+	0x3538, 0xffffffff, 0x300010ff,
+	0x353c, 0xffffffff, 0x0,
+	0x353c, 0xffffffff, 0x0,
+	0x353c, 0xffffffff, 0x0,
+	0x353c, 0xffffffff, 0x0,
+	0x353c, 0xffffffff, 0x0,
+	0x353c, 0xffffffff, 0x400000,
+	0x3538, 0xffffffff, 0x100010ff,
+	0x353c, 0xffffffff, 0x0,
+	0x353c, 0xffffffff, 0x0,
+	0x353c, 0xffffffff, 0x0,
+	0x353c, 0xffffffff, 0x0,
+	0x353c, 0xffffffff, 0x0,
+	0x353c, 0xffffffff, 0x120200,
+	0x3538, 0xffffffff, 0x500010ff,
+	0x353c, 0xffffffff, 0x0,
+	0x353c, 0xffffffff, 0x0,
+	0x353c, 0xffffffff, 0x0,
+	0x353c, 0xffffffff, 0x0,
+	0x353c, 0xffffffff, 0x0,
+	0x353c, 0xffffffff, 0x1e1e16,
+	0x3538, 0xffffffff, 0x600010ff,
+	0x353c, 0xffffffff, 0x0,
+	0x353c, 0xffffffff, 0x0,
+	0x353c, 0xffffffff, 0x0,
+	0x353c, 0xffffffff, 0x0,
+	0x353c, 0xffffffff, 0x0,
+	0x353c, 0xffffffff, 0x171f1e,
+	0x3538, 0xffffffff, 0x700010ff,
+	0x353c, 0xffffffff, 0x0,
+	0x353c, 0xffffffff, 0x0,
+	0x353c, 0xffffffff, 0x0,
+	0x353c, 0xffffffff, 0x0,
+	0x353c, 0xffffffff, 0x0,
+	0x353c, 0xffffffff, 0x0,
+	0x3538, 0xffffffff, 0x9ff,
+	0x3500, 0xffffffff, 0x0,
+	0x3504, 0xffffffff, 0x10000800,
+	0x3504, 0xffffffff, 0xf,
+	0x3504, 0xffffffff, 0xf,
+	0x3500, 0xffffffff, 0x4,
+	0x3504, 0xffffffff, 0x1000051e,
+	0x3504, 0xffffffff, 0xffff,
+	0x3504, 0xffffffff, 0xffff,
+	0x3500, 0xffffffff, 0x8,
+	0x3504, 0xffffffff, 0x80500,
+	0x3500, 0xffffffff, 0x12,
+	0x3504, 0xffffffff, 0x9050c,
+	0x3500, 0xffffffff, 0x1d,
+	0x3504, 0xffffffff, 0xb052c,
+	0x3500, 0xffffffff, 0x2a,
+	0x3504, 0xffffffff, 0x1053e,
+	0x3500, 0xffffffff, 0x2d,
+	0x3504, 0xffffffff, 0x10546,
+	0x3500, 0xffffffff, 0x30,
+	0x3504, 0xffffffff, 0xa054e,
+	0x3500, 0xffffffff, 0x3c,
+	0x3504, 0xffffffff, 0x1055f,
+	0x3500, 0xffffffff, 0x3f,
+	0x3504, 0xffffffff, 0x10567,
+	0x3500, 0xffffffff, 0x42,
+	0x3504, 0xffffffff, 0x1056f,
+	0x3500, 0xffffffff, 0x45,
+	0x3504, 0xffffffff, 0x10572,
+	0x3500, 0xffffffff, 0x48,
+	0x3504, 0xffffffff, 0x20575,
+	0x3500, 0xffffffff, 0x4c,
+	0x3504, 0xffffffff, 0x190801,
+	0x3500, 0xffffffff, 0x67,
+	0x3504, 0xffffffff, 0x1082a,
+	0x3500, 0xffffffff, 0x6a,
+	0x3504, 0xffffffff, 0x1b082d,
+	0x3500, 0xffffffff, 0x87,
+	0x3504, 0xffffffff, 0x310851,
+	0x3500, 0xffffffff, 0xba,
+	0x3504, 0xffffffff, 0x891,
+	0x3500, 0xffffffff, 0xbc,
+	0x3504, 0xffffffff, 0x893,
+	0x3500, 0xffffffff, 0xbe,
+	0x3504, 0xffffffff, 0x20895,
+	0x3500, 0xffffffff, 0xc2,
+	0x3504, 0xffffffff, 0x20899,
+	0x3500, 0xffffffff, 0xc6,
+	0x3504, 0xffffffff, 0x2089d,
+	0x3500, 0xffffffff, 0xca,
+	0x3504, 0xffffffff, 0x8a1,
+	0x3500, 0xffffffff, 0xcc,
+	0x3504, 0xffffffff, 0x8a3,
+	0x3500, 0xffffffff, 0xce,
+	0x3504, 0xffffffff, 0x308a5,
+	0x3500, 0xffffffff, 0xd3,
+	0x3504, 0xffffffff, 0x6d08cd,
+	0x3500, 0xffffffff, 0x142,
+	0x3504, 0xffffffff, 0x2000095a,
+	0x3504, 0xffffffff, 0x1,
+	0x3500, 0xffffffff, 0x144,
+	0x3504, 0xffffffff, 0x301f095b,
+	0x3500, 0xffffffff, 0x165,
+	0x3504, 0xffffffff, 0xc094d,
+	0x3500, 0xffffffff, 0x173,
+	0x3504, 0xffffffff, 0xf096d,
+	0x3500, 0xffffffff, 0x184,
+	0x3504, 0xffffffff, 0x15097f,
+	0x3500, 0xffffffff, 0x19b,
+	0x3504, 0xffffffff, 0xc0998,
+	0x3500, 0xffffffff, 0x1a9,
+	0x3504, 0xffffffff, 0x409a7,
+	0x3500, 0xffffffff, 0x1af,
+	0x3504, 0xffffffff, 0xcdc,
+	0x3500, 0xffffffff, 0x1b1,
+	0x3504, 0xffffffff, 0x800,
+	0x3508, 0xffffffff, 0x6c9b2000,
+	0x3510, 0xfc00, 0x2000,
+	0x3544, 0xffffffff, 0xfc0,
+	0x28d4, 0x00000100, 0x100
+};
+
+static void si_init_golden_registers(struct radeon_device *rdev)
+{
+	switch (rdev->family) {
+	case CHIP_TAHITI:
+		radeon_program_register_sequence(rdev,
+						 tahiti_golden_registers,
+						 (const u32)ARRAY_SIZE(tahiti_golden_registers));
+		radeon_program_register_sequence(rdev,
+						 tahiti_golden_rlc_registers,
+						 (const u32)ARRAY_SIZE(tahiti_golden_rlc_registers));
+		radeon_program_register_sequence(rdev,
+						 tahiti_mgcg_cgcg_init,
+						 (const u32)ARRAY_SIZE(tahiti_mgcg_cgcg_init));
+		radeon_program_register_sequence(rdev,
+						 tahiti_golden_registers2,
+						 (const u32)ARRAY_SIZE(tahiti_golden_registers2));
+		break;
+	case CHIP_PITCAIRN:
+		radeon_program_register_sequence(rdev,
+						 pitcairn_golden_registers,
+						 (const u32)ARRAY_SIZE(pitcairn_golden_registers));
+		radeon_program_register_sequence(rdev,
+						 pitcairn_golden_rlc_registers,
+						 (const u32)ARRAY_SIZE(pitcairn_golden_rlc_registers));
+		radeon_program_register_sequence(rdev,
+						 pitcairn_mgcg_cgcg_init,
+						 (const u32)ARRAY_SIZE(pitcairn_mgcg_cgcg_init));
+		break;
+	case CHIP_VERDE:
+		radeon_program_register_sequence(rdev,
+						 verde_golden_registers,
+						 (const u32)ARRAY_SIZE(verde_golden_registers));
+		radeon_program_register_sequence(rdev,
+						 verde_golden_rlc_registers,
+						 (const u32)ARRAY_SIZE(verde_golden_rlc_registers));
+		radeon_program_register_sequence(rdev,
+						 verde_mgcg_cgcg_init,
+						 (const u32)ARRAY_SIZE(verde_mgcg_cgcg_init));
+		radeon_program_register_sequence(rdev,
+						 verde_pg_init,
+						 (const u32)ARRAY_SIZE(verde_pg_init));
+		break;
+	case CHIP_OLAND:
+		radeon_program_register_sequence(rdev,
+						 oland_golden_registers,
+						 (const u32)ARRAY_SIZE(oland_golden_registers));
+		radeon_program_register_sequence(rdev,
+						 oland_golden_rlc_registers,
+						 (const u32)ARRAY_SIZE(oland_golden_rlc_registers));
+		radeon_program_register_sequence(rdev,
+						 oland_mgcg_cgcg_init,
+						 (const u32)ARRAY_SIZE(oland_mgcg_cgcg_init));
+		break;
+	default:
+		break;
+	}
+}
+
 #define PCIE_BUS_CLK                10000
 #define TCLK                        (PCIE_BUS_CLK / 10)
 
@@ -1211,6 +1999,7 @@ static void si_tiling_mode_table_init(struct radeon_device *rdev)
 				gb_tile_moden = 0;
 				break;
 			}
+			rdev->config.si.tile_mode_array[reg_offset] = gb_tile_moden;
 			WREG32(GB_TILE_MODE0 + (reg_offset * 4), gb_tile_moden);
 		}
 	} else if ((rdev->family == CHIP_VERDE) ||
@@ -1451,6 +2240,7 @@ static void si_tiling_mode_table_init(struct radeon_device *rdev)
 				gb_tile_moden = 0;
 				break;
 			}
+			rdev->config.si.tile_mode_array[reg_offset] = gb_tile_moden;
 			WREG32(GB_TILE_MODE0 + (reg_offset * 4), gb_tile_moden);
 		}
 	} else
@@ -1463,7 +2253,7 @@ static void si_select_se_sh(struct radeon_device *rdev,
 	u32 data = INSTANCE_BROADCAST_WRITES;
 
 	if ((se_num == 0xffffffff) && (sh_num == 0xffffffff))
-		data = SH_BROADCAST_WRITES | SE_BROADCAST_WRITES;
+		data |= SH_BROADCAST_WRITES | SE_BROADCAST_WRITES;
 	else if (se_num == 0xffffffff)
 		data |= SE_BROADCAST_WRITES | SH_INDEX(sh_num);
 	else if (sh_num == 0xffffffff)
@@ -1765,9 +2555,13 @@ static void si_gpu_init(struct radeon_device *rdev)
 
 	WREG32(GB_ADDR_CONFIG, gb_addr_config);
 	WREG32(DMIF_ADDR_CONFIG, gb_addr_config);
+	WREG32(DMIF_ADDR_CALC, gb_addr_config);
 	WREG32(HDP_ADDR_CONFIG, gb_addr_config);
 	WREG32(DMA_TILING_CONFIG + DMA0_REGISTER_OFFSET, gb_addr_config);
 	WREG32(DMA_TILING_CONFIG + DMA1_REGISTER_OFFSET, gb_addr_config);
+	WREG32(UVD_UDEC_ADDR_CONFIG, gb_addr_config);
+	WREG32(UVD_UDEC_DB_ADDR_CONFIG, gb_addr_config);
+	WREG32(UVD_UDEC_DBW_ADDR_CONFIG, gb_addr_config);
 
 	si_tiling_mode_table_init(rdev);
 
@@ -2538,46 +3332,6 @@ static void si_mc_program(struct radeon_device *rdev)
 	rv515_vga_render_disable(rdev);
 }
 
-/* SI MC address space is 40 bits */
-static void si_vram_location(struct radeon_device *rdev,
-			     struct radeon_mc *mc, u64 base)
-{
-	mc->vram_start = base;
-	if (mc->mc_vram_size > (0xFFFFFFFFFFULL - base + 1)) {
-		dev_warn(rdev->dev, "limiting VRAM to PCI aperture size\n");
-		mc->real_vram_size = mc->aper_size;
-		mc->mc_vram_size = mc->aper_size;
-	}
-	mc->vram_end = mc->vram_start + mc->mc_vram_size - 1;
-	dev_info(rdev->dev, "VRAM: %lluM 0x%016llX - 0x%016llX (%lluM used)\n",
-			mc->mc_vram_size >> 20, mc->vram_start,
-			mc->vram_end, mc->real_vram_size >> 20);
-}
-
-static void si_gtt_location(struct radeon_device *rdev, struct radeon_mc *mc)
-{
-	u64 size_af, size_bf;
-
-	size_af = ((0xFFFFFFFFFFULL - mc->vram_end) + mc->gtt_base_align) & ~mc->gtt_base_align;
-	size_bf = mc->vram_start & ~mc->gtt_base_align;
-	if (size_bf > size_af) {
-		if (mc->gtt_size > size_bf) {
-			dev_warn(rdev->dev, "limiting GTT\n");
-			mc->gtt_size = size_bf;
-		}
-		mc->gtt_start = (mc->vram_start & ~mc->gtt_base_align) - mc->gtt_size;
-	} else {
-		if (mc->gtt_size > size_af) {
-			dev_warn(rdev->dev, "limiting GTT\n");
-			mc->gtt_size = size_af;
-		}
-		mc->gtt_start = (mc->vram_end + 1 + mc->gtt_base_align) & ~mc->gtt_base_align;
-	}
-	mc->gtt_end = mc->gtt_start + mc->gtt_size - 1;
-	dev_info(rdev->dev, "GTT: %lluM 0x%016llX - 0x%016llX\n",
-			mc->gtt_size >> 20, mc->gtt_start, mc->gtt_end);
-}
-
 static void si_vram_gtt_location(struct radeon_device *rdev,
 				 struct radeon_mc *mc)
 {
@@ -2587,9 +3341,9 @@ static void si_vram_gtt_location(struct radeon_device *rdev,
 		mc->real_vram_size = 0xFFC0000000ULL;
 		mc->mc_vram_size = 0xFFC0000000ULL;
 	}
-	si_vram_location(rdev, &rdev->mc, 0);
+	radeon_vram_location(rdev, &rdev->mc, 0);
 	rdev->mc.gtt_base_align = 0;
-	si_gtt_location(rdev, mc);
+	radeon_gtt_location(rdev, mc);
 }
 
 static int si_mc_init(struct radeon_device *rdev)
@@ -4322,14 +5076,6 @@ static int si_startup(struct radeon_device *rdev)
 		return r;
 	si_gpu_init(rdev);
 
-#if 0
-	r = evergreen_blit_init(rdev);
-	if (r) {
-		r600_blit_fini(rdev);
-		rdev->asic->copy = NULL;
-		dev_warn(rdev->dev, "failed blitter (%d) falling back to memcpy\n", r);
-	}
-#endif
 	/* allocate rlc buffers */
 	r = si_rlc_init(rdev);
 	if (r) {
@@ -4372,6 +5118,16 @@ static int si_startup(struct radeon_device *rdev)
 		return r;
 	}
 
+	r = rv770_uvd_resume(rdev);
+	if (!r) {
+		r = radeon_fence_driver_start_ring(rdev,
+						   R600_RING_TYPE_UVD_INDEX);
+		if (r)
+			dev_err(rdev->dev, "UVD fences init error (%d).\n", r);
+	}
+	if (r)
+		rdev->ring[R600_RING_TYPE_UVD_INDEX].ring_size = 0;
+
 	/* Enable IRQ */
 	r = si_irq_init(rdev);
 	if (r) {
@@ -4429,6 +5185,18 @@ static int si_startup(struct radeon_device *rdev)
 	if (r)
 		return r;
 
+	ring = &rdev->ring[R600_RING_TYPE_UVD_INDEX];
+	if (ring->ring_size) {
+		r = radeon_ring_init(rdev, ring, ring->ring_size,
+				     R600_WB_UVD_RPTR_OFFSET,
+				     UVD_RBC_RB_RPTR, UVD_RBC_RB_WPTR,
+				     0, 0xfffff, RADEON_CP_PACKET2);
+		if (!r)
+			r = r600_uvd_init(rdev);
+		if (r)
+			DRM_ERROR("radeon: failed initializing UVD (%d).\n", r);
+	}
+
 	r = radeon_ib_pool_init(rdev);
 	if (r) {
 		dev_err(rdev->dev, "IB initialization failed (%d).\n", r);
@@ -4455,6 +5223,9 @@ int si_resume(struct radeon_device *rdev)
 	/* post card */
 	atom_asic_init(rdev->mode_info.atom_context);
 
+	/* init golden registers */
+	si_init_golden_registers(rdev);
+
 	rdev->accel_working = true;
 	r = si_startup(rdev);
 	if (r) {
@@ -4472,6 +5243,8 @@ int si_suspend(struct radeon_device *rdev)
 	radeon_vm_manager_fini(rdev);
 	si_cp_enable(rdev, false);
 	cayman_dma_stop(rdev);
+	r600_uvd_rbc_stop(rdev);
+	radeon_uvd_suspend(rdev);
 	si_irq_suspend(rdev);
 	radeon_wb_disable(rdev);
 	si_pcie_gart_disable(rdev);
@@ -4512,6 +5285,8 @@ int si_init(struct radeon_device *rdev)
 		DRM_INFO("GPU not posted. posting now...\n");
 		atom_asic_init(rdev->mode_info.atom_context);
 	}
+	/* init golden registers */
+	si_init_golden_registers(rdev);
 	/* Initialize scratch registers */
 	si_scratch_init(rdev);
 	/* Initialize surface registers */
@@ -4557,6 +5332,13 @@ int si_init(struct radeon_device *rdev)
 	ring->ring_obj = NULL;
 	r600_ring_init(rdev, ring, 64 * 1024);
 
+	r = radeon_uvd_init(rdev);
+	if (!r) {
+		ring = &rdev->ring[R600_RING_TYPE_UVD_INDEX];
+		ring->ring_obj = NULL;
+		r600_ring_init(rdev, ring, 4096);
+	}
+
 	rdev->ih.ring_obj = NULL;
 	r600_ih_ring_init(rdev, 64 * 1024);
 
@@ -4594,9 +5376,6 @@ int si_init(struct radeon_device *rdev)
 
 void si_fini(struct radeon_device *rdev)
 {
-#if 0
-	r600_blit_fini(rdev);
-#endif
 	si_cp_fini(rdev);
 	cayman_dma_fini(rdev);
 	si_irq_fini(rdev);
@@ -4605,6 +5384,7 @@ void si_fini(struct radeon_device *rdev)
 	radeon_vm_manager_fini(rdev);
 	radeon_ib_pool_fini(rdev);
 	radeon_irq_kms_fini(rdev);
+	radeon_uvd_fini(rdev);
 	si_pcie_gart_fini(rdev);
 	r600_vram_scratch_fini(rdev);
 	radeon_gem_fini(rdev);
@@ -4634,3 +5414,94 @@ uint64_t si_get_gpu_clock_counter(struct radeon_device *rdev)
 	mutex_unlock(&rdev->gpu_clock_mutex);
 	return clock;
 }
+
+int si_set_uvd_clocks(struct radeon_device *rdev, u32 vclk, u32 dclk)
+{
+	unsigned fb_div = 0, vclk_div = 0, dclk_div = 0;
+	int r;
+
+	/* bypass vclk and dclk with bclk */
+	WREG32_P(CG_UPLL_FUNC_CNTL_2,
+		VCLK_SRC_SEL(1) | DCLK_SRC_SEL(1),
+		~(VCLK_SRC_SEL_MASK | DCLK_SRC_SEL_MASK));
+
+	/* put PLL in bypass mode */
+	WREG32_P(CG_UPLL_FUNC_CNTL, UPLL_BYPASS_EN_MASK, ~UPLL_BYPASS_EN_MASK);
+
+	if (!vclk || !dclk) {
+		/* keep the Bypass mode, put PLL to sleep */
+		WREG32_P(CG_UPLL_FUNC_CNTL, UPLL_SLEEP_MASK, ~UPLL_SLEEP_MASK);
+		return 0;
+	}
+
+	r = radeon_uvd_calc_upll_dividers(rdev, vclk, dclk, 125000, 250000,
+					  16384, 0x03FFFFFF, 0, 128, 5,
+					  &fb_div, &vclk_div, &dclk_div);
+	if (r)
+		return r;
+
+	/* set RESET_ANTI_MUX to 0 */
+	WREG32_P(CG_UPLL_FUNC_CNTL_5, 0, ~RESET_ANTI_MUX_MASK);
+
+	/* set VCO_MODE to 1 */
+	WREG32_P(CG_UPLL_FUNC_CNTL, UPLL_VCO_MODE_MASK, ~UPLL_VCO_MODE_MASK);
+
+	/* toggle UPLL_SLEEP to 1 then back to 0 */
+	WREG32_P(CG_UPLL_FUNC_CNTL, UPLL_SLEEP_MASK, ~UPLL_SLEEP_MASK);
+	WREG32_P(CG_UPLL_FUNC_CNTL, 0, ~UPLL_SLEEP_MASK);
+
+	/* deassert UPLL_RESET */
+	WREG32_P(CG_UPLL_FUNC_CNTL, 0, ~UPLL_RESET_MASK);
+
+	mdelay(1);
+
+	r = radeon_uvd_send_upll_ctlreq(rdev, CG_UPLL_FUNC_CNTL);
+	if (r)
+		return r;
+
+	/* assert UPLL_RESET again */
+	WREG32_P(CG_UPLL_FUNC_CNTL, UPLL_RESET_MASK, ~UPLL_RESET_MASK);
+
+	/* disable spread spectrum. */
+	WREG32_P(CG_UPLL_SPREAD_SPECTRUM, 0, ~SSEN_MASK);
+
+	/* set feedback divider */
+	WREG32_P(CG_UPLL_FUNC_CNTL_3, UPLL_FB_DIV(fb_div), ~UPLL_FB_DIV_MASK);
+
+	/* set ref divider to 0 */
+	WREG32_P(CG_UPLL_FUNC_CNTL, 0, ~UPLL_REF_DIV_MASK);
+
+	if (fb_div < 307200)
+		WREG32_P(CG_UPLL_FUNC_CNTL_4, 0, ~UPLL_SPARE_ISPARE9);
+	else
+		WREG32_P(CG_UPLL_FUNC_CNTL_4, UPLL_SPARE_ISPARE9, ~UPLL_SPARE_ISPARE9);
+
+	/* set PDIV_A and PDIV_B */
+	WREG32_P(CG_UPLL_FUNC_CNTL_2,
+		UPLL_PDIV_A(vclk_div) | UPLL_PDIV_B(dclk_div),
+		~(UPLL_PDIV_A_MASK | UPLL_PDIV_B_MASK));
+
+	/* give the PLL some time to settle */
+	mdelay(15);
+
+	/* deassert PLL_RESET */
+	WREG32_P(CG_UPLL_FUNC_CNTL, 0, ~UPLL_RESET_MASK);
+
+	mdelay(15);
+
+	/* switch from bypass mode to normal mode */
+	WREG32_P(CG_UPLL_FUNC_CNTL, 0, ~UPLL_BYPASS_EN_MASK);
+
+	r = radeon_uvd_send_upll_ctlreq(rdev, CG_UPLL_FUNC_CNTL);
+	if (r)
+		return r;
+
+	/* switch VCLK and DCLK selection */
+	WREG32_P(CG_UPLL_FUNC_CNTL_2,
+		VCLK_SRC_SEL(2) | DCLK_SRC_SEL(2),
+		~(VCLK_SRC_SEL_MASK | DCLK_SRC_SEL_MASK));
+
+	mdelay(100);
+
+	return 0;
+}
diff --git a/drivers/gpu/drm/radeon/sid.h b/drivers/gpu/drm/radeon/sid.h
index 23fc08fc8e7f..222877ba6cf5 100644
--- a/drivers/gpu/drm/radeon/sid.h
+++ b/drivers/gpu/drm/radeon/sid.h
@@ -29,6 +29,35 @@
 #define TAHITI_GB_ADDR_CONFIG_GOLDEN        0x12011003
 #define VERDE_GB_ADDR_CONFIG_GOLDEN         0x12010002
 
+/* discrete uvd clocks */
+#define	CG_UPLL_FUNC_CNTL				0x634
+#	define UPLL_RESET_MASK				0x00000001
+#	define UPLL_SLEEP_MASK				0x00000002
+#	define UPLL_BYPASS_EN_MASK			0x00000004
+#	define UPLL_CTLREQ_MASK				0x00000008
+#	define UPLL_VCO_MODE_MASK			0x00000600
+#	define UPLL_REF_DIV_MASK			0x003F0000
+#	define UPLL_CTLACK_MASK				0x40000000
+#	define UPLL_CTLACK2_MASK			0x80000000
+#define	CG_UPLL_FUNC_CNTL_2				0x638
+#	define UPLL_PDIV_A(x)				((x) << 0)
+#	define UPLL_PDIV_A_MASK				0x0000007F
+#	define UPLL_PDIV_B(x)				((x) << 8)
+#	define UPLL_PDIV_B_MASK				0x00007F00
+#	define VCLK_SRC_SEL(x)				((x) << 20)
+#	define VCLK_SRC_SEL_MASK			0x01F00000
+#	define DCLK_SRC_SEL(x)				((x) << 25)
+#	define DCLK_SRC_SEL_MASK			0x3E000000
+#define	CG_UPLL_FUNC_CNTL_3				0x63C
+#	define UPLL_FB_DIV(x)				((x) << 0)
+#	define UPLL_FB_DIV_MASK				0x01FFFFFF
+#define	CG_UPLL_FUNC_CNTL_4                             0x644
+#	define UPLL_SPARE_ISPARE9			0x00020000
+#define	CG_UPLL_FUNC_CNTL_5				0x648
+#	define RESET_ANTI_MUX_MASK			0x00000200
+#define	CG_UPLL_SPREAD_SPECTRUM				0x650
+#	define SSEN_MASK				0x00000001
+
 #define	CG_MULT_THERMAL_STATUS					0x714
 #define		ASIC_MAX_TEMP(x)				((x) << 0)
 #define		ASIC_MAX_TEMP_MASK				0x000001ff
@@ -65,6 +94,8 @@
 
 #define DMIF_ADDR_CONFIG  				0xBD4
 
+#define DMIF_ADDR_CALC  				0xC00
+
 #define	SRBM_STATUS				        0xE50
 #define		GRBM_RQ_PENDING 			(1 << 5)
 #define		VMC_BUSY 				(1 << 8)
@@ -798,6 +829,15 @@
 #       define THREAD_TRACE_FINISH                      (55 << 0)
 
 /*
+ * UVD
+ */
+#define UVD_UDEC_ADDR_CONFIG				0xEF4C
+#define UVD_UDEC_DB_ADDR_CONFIG				0xEF50
+#define UVD_UDEC_DBW_ADDR_CONFIG			0xEF54
+#define UVD_RBC_RB_RPTR					0xF690
+#define UVD_RBC_RB_WPTR					0xF694
+
+/*
  * PM4
  */
 #define PACKET0(reg, n)	((RADEON_PACKET_TYPE0 << 30) |			\
diff --git a/drivers/gpu/drm/shmobile/shmob_drm_crtc.c b/drivers/gpu/drm/shmobile/shmob_drm_crtc.c
index d917a411ca85..7dff49ed66e7 100644
--- a/drivers/gpu/drm/shmobile/shmob_drm_crtc.c
+++ b/drivers/gpu/drm/shmobile/shmob_drm_crtc.c
@@ -494,10 +494,10 @@ static int shmob_drm_crtc_page_flip(struct drm_crtc *crtc,
 
 	if (event) {
 		event->pipe = 0;
+		drm_vblank_get(dev, 0);
 		spin_lock_irqsave(&dev->event_lock, flags);
 		scrtc->event = event;
 		spin_unlock_irqrestore(&dev->event_lock, flags);
-		drm_vblank_get(dev, 0);
 	}
 
 	return 0;
diff --git a/drivers/gpu/drm/tegra/Makefile b/drivers/gpu/drm/tegra/Makefile
deleted file mode 100644
index 80f73d1315d0..000000000000
--- a/drivers/gpu/drm/tegra/Makefile
+++ /dev/null
@@ -1,7 +0,0 @@
-ccflags-y := -Iinclude/drm
-ccflags-$(CONFIG_DRM_TEGRA_DEBUG) += -DDEBUG
-
-tegra-drm-y := drm.o fb.o dc.o host1x.o
-tegra-drm-y += output.o rgb.o hdmi.o
-
-obj-$(CONFIG_DRM_TEGRA) += tegra-drm.o
diff --git a/drivers/gpu/drm/tegra/drm.c b/drivers/gpu/drm/tegra/drm.c
deleted file mode 100644
index 9d452df5bcad..000000000000
--- a/drivers/gpu/drm/tegra/drm.c
+++ /dev/null
@@ -1,217 +0,0 @@
-/*
- * Copyright (C) 2012 Avionic Design GmbH
- * Copyright (C) 2012 NVIDIA CORPORATION.  All rights reserved.
- *
- * This program is free software; you can redistribute it and/or modify
- * it under the terms of the GNU General Public License version 2 as
- * published by the Free Software Foundation.
- */
-
-#include <linux/module.h>
-#include <linux/of_address.h>
-#include <linux/of_platform.h>
-
-#include <linux/dma-mapping.h>
-#include <asm/dma-iommu.h>
-
-#include "drm.h"
-
-#define DRIVER_NAME "tegra"
-#define DRIVER_DESC "NVIDIA Tegra graphics"
-#define DRIVER_DATE "20120330"
-#define DRIVER_MAJOR 0
-#define DRIVER_MINOR 0
-#define DRIVER_PATCHLEVEL 0
-
-static int tegra_drm_load(struct drm_device *drm, unsigned long flags)
-{
-	struct device *dev = drm->dev;
-	struct host1x *host1x;
-	int err;
-
-	host1x = dev_get_drvdata(dev);
-	drm->dev_private = host1x;
-	host1x->drm = drm;
-
-	drm_mode_config_init(drm);
-
-	err = host1x_drm_init(host1x, drm);
-	if (err < 0)
-		return err;
-
-	err = drm_vblank_init(drm, drm->mode_config.num_crtc);
-	if (err < 0)
-		return err;
-
-	err = tegra_drm_fb_init(drm);
-	if (err < 0)
-		return err;
-
-	drm_kms_helper_poll_init(drm);
-
-	return 0;
-}
-
-static int tegra_drm_unload(struct drm_device *drm)
-{
-	drm_kms_helper_poll_fini(drm);
-	tegra_drm_fb_exit(drm);
-
-	drm_mode_config_cleanup(drm);
-
-	return 0;
-}
-
-static int tegra_drm_open(struct drm_device *drm, struct drm_file *filp)
-{
-	return 0;
-}
-
-static void tegra_drm_lastclose(struct drm_device *drm)
-{
-	struct host1x *host1x = drm->dev_private;
-
-	drm_fbdev_cma_restore_mode(host1x->fbdev);
-}
-
-static struct drm_ioctl_desc tegra_drm_ioctls[] = {
-};
-
-static const struct file_operations tegra_drm_fops = {
-	.owner = THIS_MODULE,
-	.open = drm_open,
-	.release = drm_release,
-	.unlocked_ioctl = drm_ioctl,
-	.mmap = drm_gem_cma_mmap,
-	.poll = drm_poll,
-	.fasync = drm_fasync,
-	.read = drm_read,
-#ifdef CONFIG_COMPAT
-	.compat_ioctl = drm_compat_ioctl,
-#endif
-	.llseek = noop_llseek,
-};
-
-static struct drm_crtc *tegra_crtc_from_pipe(struct drm_device *drm, int pipe)
-{
-	struct drm_crtc *crtc;
-
-	list_for_each_entry(crtc, &drm->mode_config.crtc_list, head) {
-		struct tegra_dc *dc = to_tegra_dc(crtc);
-
-		if (dc->pipe == pipe)
-			return crtc;
-	}
-
-	return NULL;
-}
-
-static u32 tegra_drm_get_vblank_counter(struct drm_device *dev, int crtc)
-{
-	/* TODO: implement real hardware counter using syncpoints */
-	return drm_vblank_count(dev, crtc);
-}
-
-static int tegra_drm_enable_vblank(struct drm_device *drm, int pipe)
-{
-	struct drm_crtc *crtc = tegra_crtc_from_pipe(drm, pipe);
-	struct tegra_dc *dc = to_tegra_dc(crtc);
-
-	if (!crtc)
-		return -ENODEV;
-
-	tegra_dc_enable_vblank(dc);
-
-	return 0;
-}
-
-static void tegra_drm_disable_vblank(struct drm_device *drm, int pipe)
-{
-	struct drm_crtc *crtc = tegra_crtc_from_pipe(drm, pipe);
-	struct tegra_dc *dc = to_tegra_dc(crtc);
-
-	if (crtc)
-		tegra_dc_disable_vblank(dc);
-}
-
-static void tegra_drm_preclose(struct drm_device *drm, struct drm_file *file)
-{
-	struct drm_crtc *crtc;
-
-	list_for_each_entry(crtc, &drm->mode_config.crtc_list, head)
-		tegra_dc_cancel_page_flip(crtc, file);
-}
-
-#ifdef CONFIG_DEBUG_FS
-static int tegra_debugfs_framebuffers(struct seq_file *s, void *data)
-{
-	struct drm_info_node *node = (struct drm_info_node *)s->private;
-	struct drm_device *drm = node->minor->dev;
-	struct drm_framebuffer *fb;
-
-	mutex_lock(&drm->mode_config.fb_lock);
-
-	list_for_each_entry(fb, &drm->mode_config.fb_list, head) {
-		seq_printf(s, "%3d: user size: %d x %d, depth %d, %d bpp, refcount %d\n",
-			   fb->base.id, fb->width, fb->height, fb->depth,
-			   fb->bits_per_pixel,
-			   atomic_read(&fb->refcount.refcount));
-	}
-
-	mutex_unlock(&drm->mode_config.fb_lock);
-
-	return 0;
-}
-
-static struct drm_info_list tegra_debugfs_list[] = {
-	{ "framebuffers", tegra_debugfs_framebuffers, 0 },
-};
-
-static int tegra_debugfs_init(struct drm_minor *minor)
-{
-	return drm_debugfs_create_files(tegra_debugfs_list,
-					ARRAY_SIZE(tegra_debugfs_list),
-					minor->debugfs_root, minor);
-}
-
-static void tegra_debugfs_cleanup(struct drm_minor *minor)
-{
-	drm_debugfs_remove_files(tegra_debugfs_list,
-				 ARRAY_SIZE(tegra_debugfs_list), minor);
-}
-#endif
-
-struct drm_driver tegra_drm_driver = {
-	.driver_features = DRIVER_BUS_PLATFORM | DRIVER_MODESET | DRIVER_GEM,
-	.load = tegra_drm_load,
-	.unload = tegra_drm_unload,
-	.open = tegra_drm_open,
-	.preclose = tegra_drm_preclose,
-	.lastclose = tegra_drm_lastclose,
-
-	.get_vblank_counter = tegra_drm_get_vblank_counter,
-	.enable_vblank = tegra_drm_enable_vblank,
-	.disable_vblank = tegra_drm_disable_vblank,
-
-#if defined(CONFIG_DEBUG_FS)
-	.debugfs_init = tegra_debugfs_init,
-	.debugfs_cleanup = tegra_debugfs_cleanup,
-#endif
-
-	.gem_free_object = drm_gem_cma_free_object,
-	.gem_vm_ops = &drm_gem_cma_vm_ops,
-	.dumb_create = drm_gem_cma_dumb_create,
-	.dumb_map_offset = drm_gem_cma_dumb_map_offset,
-	.dumb_destroy = drm_gem_cma_dumb_destroy,
-
-	.ioctls = tegra_drm_ioctls,
-	.num_ioctls = ARRAY_SIZE(tegra_drm_ioctls),
-	.fops = &tegra_drm_fops,
-
-	.name = DRIVER_NAME,
-	.desc = DRIVER_DESC,
-	.date = DRIVER_DATE,
-	.major = DRIVER_MAJOR,
-	.minor = DRIVER_MINOR,
-	.patchlevel = DRIVER_PATCHLEVEL,
-};
diff --git a/drivers/gpu/drm/tegra/fb.c b/drivers/gpu/drm/tegra/fb.c
deleted file mode 100644
index 03914953cb1c..000000000000
--- a/drivers/gpu/drm/tegra/fb.c
+++ /dev/null
@@ -1,52 +0,0 @@
-/*
- * Copyright (C) 2012 Avionic Design GmbH
- * Copyright (C) 2012 NVIDIA CORPORATION.  All rights reserved.
- *
- * This program is free software; you can redistribute it and/or modify
- * it under the terms of the GNU General Public License version 2 as
- * published by the Free Software Foundation.
- */
-
-#include "drm.h"
-
-static void tegra_drm_fb_output_poll_changed(struct drm_device *drm)
-{
-	struct host1x *host1x = drm->dev_private;
-
-	drm_fbdev_cma_hotplug_event(host1x->fbdev);
-}
-
-static const struct drm_mode_config_funcs tegra_drm_mode_funcs = {
-	.fb_create = drm_fb_cma_create,
-	.output_poll_changed = tegra_drm_fb_output_poll_changed,
-};
-
-int tegra_drm_fb_init(struct drm_device *drm)
-{
-	struct host1x *host1x = drm->dev_private;
-	struct drm_fbdev_cma *fbdev;
-
-	drm->mode_config.min_width = 0;
-	drm->mode_config.min_height = 0;
-
-	drm->mode_config.max_width = 4096;
-	drm->mode_config.max_height = 4096;
-
-	drm->mode_config.funcs = &tegra_drm_mode_funcs;
-
-	fbdev = drm_fbdev_cma_init(drm, 32, drm->mode_config.num_crtc,
-				   drm->mode_config.num_connector);
-	if (IS_ERR(fbdev))
-		return PTR_ERR(fbdev);
-
-	host1x->fbdev = fbdev;
-
-	return 0;
-}
-
-void tegra_drm_fb_exit(struct drm_device *drm)
-{
-	struct host1x *host1x = drm->dev_private;
-
-	drm_fbdev_cma_fini(host1x->fbdev);
-}
diff --git a/drivers/gpu/drm/tegra/host1x.c b/drivers/gpu/drm/tegra/host1x.c
deleted file mode 100644
index 92e25a7e00ea..000000000000
--- a/drivers/gpu/drm/tegra/host1x.c
+++ /dev/null
@@ -1,327 +0,0 @@
-/*
- * Copyright (C) 2012 Avionic Design GmbH
- * Copyright (C) 2012 NVIDIA CORPORATION.  All rights reserved.
- *
- * This program is free software; you can redistribute it and/or modify
- * it under the terms of the GNU General Public License version 2 as
- * published by the Free Software Foundation.
- */
-
-#include <linux/clk.h>
-#include <linux/err.h>
-#include <linux/module.h>
-#include <linux/of.h>
-#include <linux/platform_device.h>
-
-#include "drm.h"
-
-struct host1x_drm_client {
-	struct host1x_client *client;
-	struct device_node *np;
-	struct list_head list;
-};
-
-static int host1x_add_drm_client(struct host1x *host1x, struct device_node *np)
-{
-	struct host1x_drm_client *client;
-
-	client = kzalloc(sizeof(*client), GFP_KERNEL);
-	if (!client)
-		return -ENOMEM;
-
-	INIT_LIST_HEAD(&client->list);
-	client->np = of_node_get(np);
-
-	list_add_tail(&client->list, &host1x->drm_clients);
-
-	return 0;
-}
-
-static int host1x_activate_drm_client(struct host1x *host1x,
-				      struct host1x_drm_client *drm,
-				      struct host1x_client *client)
-{
-	mutex_lock(&host1x->drm_clients_lock);
-	list_del_init(&drm->list);
-	list_add_tail(&drm->list, &host1x->drm_active);
-	drm->client = client;
-	mutex_unlock(&host1x->drm_clients_lock);
-
-	return 0;
-}
-
-static int host1x_remove_drm_client(struct host1x *host1x,
-				    struct host1x_drm_client *client)
-{
-	mutex_lock(&host1x->drm_clients_lock);
-	list_del_init(&client->list);
-	mutex_unlock(&host1x->drm_clients_lock);
-
-	of_node_put(client->np);
-	kfree(client);
-
-	return 0;
-}
-
-static int host1x_parse_dt(struct host1x *host1x)
-{
-	static const char * const compat[] = {
-		"nvidia,tegra20-dc",
-		"nvidia,tegra20-hdmi",
-		"nvidia,tegra30-dc",
-		"nvidia,tegra30-hdmi",
-	};
-	unsigned int i;
-	int err;
-
-	for (i = 0; i < ARRAY_SIZE(compat); i++) {
-		struct device_node *np;
-
-		for_each_child_of_node(host1x->dev->of_node, np) {
-			if (of_device_is_compatible(np, compat[i]) &&
-			    of_device_is_available(np)) {
-				err = host1x_add_drm_client(host1x, np);
-				if (err < 0)
-					return err;
-			}
-		}
-	}
-
-	return 0;
-}
-
-static int tegra_host1x_probe(struct platform_device *pdev)
-{
-	struct host1x *host1x;
-	struct resource *regs;
-	int err;
-
-	host1x = devm_kzalloc(&pdev->dev, sizeof(*host1x), GFP_KERNEL);
-	if (!host1x)
-		return -ENOMEM;
-
-	mutex_init(&host1x->drm_clients_lock);
-	INIT_LIST_HEAD(&host1x->drm_clients);
-	INIT_LIST_HEAD(&host1x->drm_active);
-	mutex_init(&host1x->clients_lock);
-	INIT_LIST_HEAD(&host1x->clients);
-	host1x->dev = &pdev->dev;
-
-	err = host1x_parse_dt(host1x);
-	if (err < 0) {
-		dev_err(&pdev->dev, "failed to parse DT: %d\n", err);
-		return err;
-	}
-
-	host1x->clk = devm_clk_get(&pdev->dev, NULL);
-	if (IS_ERR(host1x->clk))
-		return PTR_ERR(host1x->clk);
-
-	err = clk_prepare_enable(host1x->clk);
-	if (err < 0)
-		return err;
-
-	regs = platform_get_resource(pdev, IORESOURCE_MEM, 0);
-	if (!regs) {
-		err = -ENXIO;
-		goto err;
-	}
-
-	err = platform_get_irq(pdev, 0);
-	if (err < 0)
-		goto err;
-
-	host1x->syncpt = err;
-
-	err = platform_get_irq(pdev, 1);
-	if (err < 0)
-		goto err;
-
-	host1x->irq = err;
-
-	host1x->regs = devm_ioremap_resource(&pdev->dev, regs);
-	if (IS_ERR(host1x->regs)) {
-		err = PTR_ERR(host1x->regs);
-		goto err;
-	}
-
-	platform_set_drvdata(pdev, host1x);
-
-	return 0;
-
-err:
-	clk_disable_unprepare(host1x->clk);
-	return err;
-}
-
-static int tegra_host1x_remove(struct platform_device *pdev)
-{
-	struct host1x *host1x = platform_get_drvdata(pdev);
-
-	clk_disable_unprepare(host1x->clk);
-
-	return 0;
-}
-
-int host1x_drm_init(struct host1x *host1x, struct drm_device *drm)
-{
-	struct host1x_client *client;
-
-	mutex_lock(&host1x->clients_lock);
-
-	list_for_each_entry(client, &host1x->clients, list) {
-		if (client->ops && client->ops->drm_init) {
-			int err = client->ops->drm_init(client, drm);
-			if (err < 0) {
-				dev_err(host1x->dev,
-					"DRM setup failed for %s: %d\n",
-					dev_name(client->dev), err);
-				return err;
-			}
-		}
-	}
-
-	mutex_unlock(&host1x->clients_lock);
-
-	return 0;
-}
-
-int host1x_drm_exit(struct host1x *host1x)
-{
-	struct platform_device *pdev = to_platform_device(host1x->dev);
-	struct host1x_client *client;
-
-	if (!host1x->drm)
-		return 0;
-
-	mutex_lock(&host1x->clients_lock);
-
-	list_for_each_entry_reverse(client, &host1x->clients, list) {
-		if (client->ops && client->ops->drm_exit) {
-			int err = client->ops->drm_exit(client);
-			if (err < 0) {
-				dev_err(host1x->dev,
-					"DRM cleanup failed for %s: %d\n",
-					dev_name(client->dev), err);
-				return err;
-			}
-		}
-	}
-
-	mutex_unlock(&host1x->clients_lock);
-
-	drm_platform_exit(&tegra_drm_driver, pdev);
-	host1x->drm = NULL;
-
-	return 0;
-}
-
-int host1x_register_client(struct host1x *host1x, struct host1x_client *client)
-{
-	struct host1x_drm_client *drm, *tmp;
-	int err;
-
-	mutex_lock(&host1x->clients_lock);
-	list_add_tail(&client->list, &host1x->clients);
-	mutex_unlock(&host1x->clients_lock);
-
-	list_for_each_entry_safe(drm, tmp, &host1x->drm_clients, list)
-		if (drm->np == client->dev->of_node)
-			host1x_activate_drm_client(host1x, drm, client);
-
-	if (list_empty(&host1x->drm_clients)) {
-		struct platform_device *pdev = to_platform_device(host1x->dev);
-
-		err = drm_platform_init(&tegra_drm_driver, pdev);
-		if (err < 0) {
-			dev_err(host1x->dev, "drm_platform_init(): %d\n", err);
-			return err;
-		}
-	}
-
-	client->host1x = host1x;
-
-	return 0;
-}
-
-int host1x_unregister_client(struct host1x *host1x,
-			     struct host1x_client *client)
-{
-	struct host1x_drm_client *drm, *tmp;
-	int err;
-
-	list_for_each_entry_safe(drm, tmp, &host1x->drm_active, list) {
-		if (drm->client == client) {
-			err = host1x_drm_exit(host1x);
-			if (err < 0) {
-				dev_err(host1x->dev, "host1x_drm_exit(): %d\n",
-					err);
-				return err;
-			}
-
-			host1x_remove_drm_client(host1x, drm);
-			break;
-		}
-	}
-
-	mutex_lock(&host1x->clients_lock);
-	list_del_init(&client->list);
-	mutex_unlock(&host1x->clients_lock);
-
-	return 0;
-}
-
-static struct of_device_id tegra_host1x_of_match[] = {
-	{ .compatible = "nvidia,tegra30-host1x", },
-	{ .compatible = "nvidia,tegra20-host1x", },
-	{ },
-};
-MODULE_DEVICE_TABLE(of, tegra_host1x_of_match);
-
-struct platform_driver tegra_host1x_driver = {
-	.driver = {
-		.name = "tegra-host1x",
-		.owner = THIS_MODULE,
-		.of_match_table = tegra_host1x_of_match,
-	},
-	.probe = tegra_host1x_probe,
-	.remove = tegra_host1x_remove,
-};
-
-static int __init tegra_host1x_init(void)
-{
-	int err;
-
-	err = platform_driver_register(&tegra_host1x_driver);
-	if (err < 0)
-		return err;
-
-	err = platform_driver_register(&tegra_dc_driver);
-	if (err < 0)
-		goto unregister_host1x;
-
-	err = platform_driver_register(&tegra_hdmi_driver);
-	if (err < 0)
-		goto unregister_dc;
-
-	return 0;
-
-unregister_dc:
-	platform_driver_unregister(&tegra_dc_driver);
-unregister_host1x:
-	platform_driver_unregister(&tegra_host1x_driver);
-	return err;
-}
-module_init(tegra_host1x_init);
-
-static void __exit tegra_host1x_exit(void)
-{
-	platform_driver_unregister(&tegra_hdmi_driver);
-	platform_driver_unregister(&tegra_dc_driver);
-	platform_driver_unregister(&tegra_host1x_driver);
-}
-module_exit(tegra_host1x_exit);
-
-MODULE_AUTHOR("Thierry Reding <thierry.reding@avionic-design.de>");
-MODULE_DESCRIPTION("NVIDIA Tegra DRM driver");
-MODULE_LICENSE("GPL");
diff --git a/drivers/gpu/drm/tilcdc/Makefile b/drivers/gpu/drm/tilcdc/Makefile
index deda656b10e7..7d2eefe94bf7 100644
--- a/drivers/gpu/drm/tilcdc/Makefile
+++ b/drivers/gpu/drm/tilcdc/Makefile
@@ -1,4 +1,7 @@
-ccflags-y := -Iinclude/drm -Werror
+ccflags-y := -Iinclude/drm
+ifeq (, $(findstring -W,$(EXTRA_CFLAGS)))
+	ccflags-y += -Werror
+endif
 
 tilcdc-y := \
 	tilcdc_crtc.o \
diff --git a/drivers/gpu/drm/tilcdc/tilcdc_drv.c b/drivers/gpu/drm/tilcdc/tilcdc_drv.c
index c5b592dc1970..2b5461bcd9fb 100644
--- a/drivers/gpu/drm/tilcdc/tilcdc_drv.c
+++ b/drivers/gpu/drm/tilcdc/tilcdc_drv.c
@@ -75,7 +75,7 @@ static int modeset_init(struct drm_device *dev)
 		mod->funcs->modeset_init(mod, dev);
 	}
 
-	if ((priv->num_encoders = 0) || (priv->num_connectors == 0)) {
+	if ((priv->num_encoders == 0) || (priv->num_connectors == 0)) {
 		/* oh nos! */
 		dev_err(dev->dev, "no encoders/connectors found\n");
 		return -ENXIO;
@@ -299,11 +299,10 @@ static int tilcdc_irq_postinstall(struct drm_device *dev)
 	struct tilcdc_drm_private *priv = dev->dev_private;
 
 	/* enable FIFO underflow irq: */
-	if (priv->rev == 1) {
+	if (priv->rev == 1)
 		tilcdc_set(dev, LCDC_RASTER_CTRL_REG, LCDC_V1_UNDERFLOW_INT_ENA);
-	} else {
+	else
 		tilcdc_set(dev, LCDC_INT_ENABLE_SET_REG, LCDC_V2_UNDERFLOW_INT_ENA);
-	}
 
 	return 0;
 }
@@ -363,7 +362,7 @@ static const struct {
 	uint8_t  rev;
 	uint8_t  save;
 	uint32_t reg;
-} registers[] = 	{
+} registers[] =		{
 #define REG(rev, save, reg) { #reg, rev, save, reg }
 		/* exists in revision 1: */
 		REG(1, false, LCDC_PID_REG),
diff --git a/drivers/gpu/drm/tilcdc/tilcdc_panel.c b/drivers/gpu/drm/tilcdc/tilcdc_panel.c
index 90ee49786372..09176654fddb 100644
--- a/drivers/gpu/drm/tilcdc/tilcdc_panel.c
+++ b/drivers/gpu/drm/tilcdc/tilcdc_panel.c
@@ -305,7 +305,7 @@ static const struct tilcdc_module_ops panel_module_ops = {
  */
 
 /* maybe move this somewhere common if it is needed by other outputs? */
-static struct tilcdc_panel_info * of_get_panel_info(struct device_node *np)
+static struct tilcdc_panel_info *of_get_panel_info(struct device_node *np)
 {
 	struct device_node *info_np;
 	struct tilcdc_panel_info *info;
@@ -413,7 +413,6 @@ static struct of_device_id panel_of_match[] = {
 		{ .compatible = "ti,tilcdc,panel", },
 		{ },
 };
-MODULE_DEVICE_TABLE(of, panel_of_match);
 
 struct platform_driver panel_driver = {
 	.probe = panel_probe,
diff --git a/drivers/gpu/drm/tilcdc/tilcdc_slave.c b/drivers/gpu/drm/tilcdc/tilcdc_slave.c
index 568dc1c08e6c..db1d2fc9dfb5 100644
--- a/drivers/gpu/drm/tilcdc/tilcdc_slave.c
+++ b/drivers/gpu/drm/tilcdc/tilcdc_slave.c
@@ -353,7 +353,6 @@ static struct of_device_id slave_of_match[] = {
 		{ .compatible = "ti,tilcdc,slave", },
 		{ },
 };
-MODULE_DEVICE_TABLE(of, slave_of_match);
 
 struct platform_driver slave_driver = {
 	.probe = slave_probe,
diff --git a/drivers/gpu/drm/tilcdc/tilcdc_tfp410.c b/drivers/gpu/drm/tilcdc/tilcdc_tfp410.c
index 58d487ba2414..a36788fbcd98 100644
--- a/drivers/gpu/drm/tilcdc/tilcdc_tfp410.c
+++ b/drivers/gpu/drm/tilcdc/tilcdc_tfp410.c
@@ -396,7 +396,6 @@ static struct of_device_id tfp410_of_match[] = {
 		{ .compatible = "ti,tilcdc,tfp410", },
 		{ },
 };
-MODULE_DEVICE_TABLE(of, tfp410_of_match);
 
 struct platform_driver tfp410_driver = {
 	.probe = tfp410_probe,
diff --git a/drivers/gpu/drm/ttm/ttm_bo_util.c b/drivers/gpu/drm/ttm/ttm_bo_util.c
index 8be35c809c7b..af894584dd90 100644
--- a/drivers/gpu/drm/ttm/ttm_bo_util.c
+++ b/drivers/gpu/drm/ttm/ttm_bo_util.c
@@ -86,6 +86,7 @@ int ttm_mem_io_lock(struct ttm_mem_type_manager *man, bool interruptible)
 	mutex_lock(&man->io_reserve_mutex);
 	return 0;
 }
+EXPORT_SYMBOL(ttm_mem_io_lock);
 
 void ttm_mem_io_unlock(struct ttm_mem_type_manager *man)
 {
@@ -94,6 +95,7 @@ void ttm_mem_io_unlock(struct ttm_mem_type_manager *man)
 
 	mutex_unlock(&man->io_reserve_mutex);
 }
+EXPORT_SYMBOL(ttm_mem_io_unlock);
 
 static int ttm_mem_io_evict(struct ttm_mem_type_manager *man)
 {
@@ -111,8 +113,9 @@ static int ttm_mem_io_evict(struct ttm_mem_type_manager *man)
 	return 0;
 }
 
-static int ttm_mem_io_reserve(struct ttm_bo_device *bdev,
-			      struct ttm_mem_reg *mem)
+
+int ttm_mem_io_reserve(struct ttm_bo_device *bdev,
+		       struct ttm_mem_reg *mem)
 {
 	struct ttm_mem_type_manager *man = &bdev->man[mem->mem_type];
 	int ret = 0;
@@ -134,9 +137,10 @@ retry:
 	}
 	return ret;
 }
+EXPORT_SYMBOL(ttm_mem_io_reserve);
 
-static void ttm_mem_io_free(struct ttm_bo_device *bdev,
-			    struct ttm_mem_reg *mem)
+void ttm_mem_io_free(struct ttm_bo_device *bdev,
+		     struct ttm_mem_reg *mem)
 {
 	struct ttm_mem_type_manager *man = &bdev->man[mem->mem_type];
 
@@ -149,6 +153,7 @@ static void ttm_mem_io_free(struct ttm_bo_device *bdev,
 		bdev->driver->io_mem_free(bdev, mem);
 
 }
+EXPORT_SYMBOL(ttm_mem_io_free);
 
 int ttm_mem_io_reserve_vm(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 74705f329d99..3df9f16b041c 100644
--- a/drivers/gpu/drm/ttm/ttm_bo_vm.c
+++ b/drivers/gpu/drm/ttm/ttm_bo_vm.c
@@ -147,7 +147,7 @@ static int ttm_bo_vm_fault(struct vm_area_struct *vma, struct vm_fault *vmf)
 
 	page_offset = ((address - vma->vm_start) >> PAGE_SHIFT) +
 	    bo->vm_node->start - vma->vm_pgoff;
-	page_last = ((vma->vm_end - vma->vm_start) >> PAGE_SHIFT) +
+	page_last = vma_pages(vma) +
 	    bo->vm_node->start - vma->vm_pgoff;
 
 	if (unlikely(page_offset >= bo->num_pages)) {
@@ -258,7 +258,7 @@ int ttm_bo_mmap(struct file *filp, struct vm_area_struct *vma,
 
 	read_lock(&bdev->vm_lock);
 	bo = ttm_bo_vm_lookup_rb(bdev, vma->vm_pgoff,
-				 (vma->vm_end - vma->vm_start) >> PAGE_SHIFT);
+				 vma_pages(vma));
 	if (likely(bo != NULL) && !kref_get_unless_zero(&bo->kref))
 		bo = NULL;
 	read_unlock(&bdev->vm_lock);
diff --git a/drivers/gpu/drm/udl/udl_fb.c b/drivers/gpu/drm/udl/udl_fb.c
index 9f4be3d4a02e..dc0c065f8d39 100644
--- a/drivers/gpu/drm/udl/udl_fb.c
+++ b/drivers/gpu/drm/udl/udl_fb.c
@@ -482,7 +482,7 @@ static int udlfb_create(struct drm_fb_helper *helper,
 	struct udl_fbdev *ufbdev = (struct udl_fbdev *)helper;
 	struct drm_device *dev = ufbdev->helper.dev;
 	struct fb_info *info;
-	struct device *device = &dev->usbdev->dev;
+	struct device *device = dev->dev;
 	struct drm_framebuffer *fb;
 	struct drm_mode_fb_cmd2 mode_cmd;
 	struct udl_gem_object *obj;
diff --git a/drivers/gpu/drm/udl/udl_gem.c b/drivers/gpu/drm/udl/udl_gem.c
index 3816270ba49b..ef034fa3e6f5 100644
--- a/drivers/gpu/drm/udl/udl_gem.c
+++ b/drivers/gpu/drm/udl/udl_gem.c
@@ -303,6 +303,8 @@ struct drm_gem_object *udl_gem_prime_import(struct drm_device *dev,
 	if (IS_ERR(attach))
 		return ERR_CAST(attach);
 
+	get_dma_buf(dma_buf);
+
 	sg = dma_buf_map_attachment(attach, DMA_BIDIRECTIONAL);
 	if (IS_ERR(sg)) {
 		ret = PTR_ERR(sg);
@@ -322,5 +324,7 @@ fail_unmap:
 	dma_buf_unmap_attachment(attach, sg, DMA_BIDIRECTIONAL);
 fail_detach:
 	dma_buf_detach(dma_buf, attach);
+	dma_buf_put(dma_buf);
+
 	return ERR_PTR(ret);
 }
diff --git a/drivers/gpu/host1x/Kconfig b/drivers/gpu/host1x/Kconfig
new file mode 100644
index 000000000000..ccfd42b23606
--- /dev/null
+++ b/drivers/gpu/host1x/Kconfig
@@ -0,0 +1,24 @@
+config TEGRA_HOST1X
+	tristate "NVIDIA Tegra host1x driver"
+	depends on ARCH_TEGRA || ARCH_MULTIPLATFORM
+	help
+	  Driver for the NVIDIA Tegra host1x hardware.
+
+	  The Tegra host1x module is the DMA engine for register access to
+	  Tegra's graphics- and multimedia-related modules. The modules served
+	  by host1x are referred to as clients. host1x includes some other
+	  functionality, such as synchronization.
+
+if TEGRA_HOST1X
+
+config TEGRA_HOST1X_FIREWALL
+	bool "Enable HOST1X security firewall"
+	default y
+	help
+	  Say yes if kernel should protect command streams from tampering.
+
+	  If unsure, choose Y.
+
+source "drivers/gpu/host1x/drm/Kconfig"
+
+endif
diff --git a/drivers/gpu/host1x/Makefile b/drivers/gpu/host1x/Makefile
new file mode 100644
index 000000000000..3b037b6e0298
--- /dev/null
+++ b/drivers/gpu/host1x/Makefile
@@ -0,0 +1,20 @@
+ccflags-y = -Idrivers/gpu/host1x
+
+host1x-y = \
+	syncpt.o \
+	dev.o \
+	intr.o \
+	cdma.o \
+	channel.o \
+	job.o \
+	debug.o \
+	hw/host1x01.o
+
+ccflags-y += -Iinclude/drm
+ccflags-$(CONFIG_DRM_TEGRA_DEBUG) += -DDEBUG
+
+host1x-$(CONFIG_DRM_TEGRA) += drm/drm.o drm/fb.o drm/dc.o
+host1x-$(CONFIG_DRM_TEGRA) += drm/output.o drm/rgb.o drm/hdmi.o
+host1x-$(CONFIG_DRM_TEGRA) += drm/gem.o
+host1x-$(CONFIG_DRM_TEGRA) += drm/gr2d.o
+obj-$(CONFIG_TEGRA_HOST1X) += host1x.o
diff --git a/drivers/gpu/host1x/cdma.c b/drivers/gpu/host1x/cdma.c
new file mode 100644
index 000000000000..de72172d3b5f
--- /dev/null
+++ b/drivers/gpu/host1x/cdma.c
@@ -0,0 +1,491 @@
+/*
+ * Tegra host1x Command DMA
+ *
+ * Copyright (c) 2010-2013, NVIDIA Corporation.
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms and conditions of the GNU General Public License,
+ * version 2, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program.  If not, see <http://www.gnu.org/licenses/>.
+ */
+
+
+#include <asm/cacheflush.h>
+#include <linux/device.h>
+#include <linux/dma-mapping.h>
+#include <linux/interrupt.h>
+#include <linux/kernel.h>
+#include <linux/kfifo.h>
+#include <linux/slab.h>
+#include <trace/events/host1x.h>
+
+#include "cdma.h"
+#include "channel.h"
+#include "dev.h"
+#include "debug.h"
+#include "host1x_bo.h"
+#include "job.h"
+
+/*
+ * push_buffer
+ *
+ * The push buffer is a circular array of words to be fetched by command DMA.
+ * Note that it works slightly differently to the sync queue; fence == pos
+ * means that the push buffer is full, not empty.
+ */
+
+#define HOST1X_PUSHBUFFER_SLOTS	512
+
+/*
+ * Clean up push buffer resources
+ */
+static void host1x_pushbuffer_destroy(struct push_buffer *pb)
+{
+	struct host1x_cdma *cdma = pb_to_cdma(pb);
+	struct host1x *host1x = cdma_to_host1x(cdma);
+
+	if (pb->phys != 0)
+		dma_free_writecombine(host1x->dev, pb->size_bytes + 4,
+				      pb->mapped, pb->phys);
+
+	pb->mapped = NULL;
+	pb->phys = 0;
+}
+
+/*
+ * Init push buffer resources
+ */
+static int host1x_pushbuffer_init(struct push_buffer *pb)
+{
+	struct host1x_cdma *cdma = pb_to_cdma(pb);
+	struct host1x *host1x = cdma_to_host1x(cdma);
+
+	pb->mapped = NULL;
+	pb->phys = 0;
+	pb->size_bytes = HOST1X_PUSHBUFFER_SLOTS * 8;
+
+	/* initialize buffer pointers */
+	pb->fence = pb->size_bytes - 8;
+	pb->pos = 0;
+
+	/* allocate and map pushbuffer memory */
+	pb->mapped = dma_alloc_writecombine(host1x->dev, pb->size_bytes + 4,
+					    &pb->phys, GFP_KERNEL);
+	if (!pb->mapped)
+		goto fail;
+
+	host1x_hw_pushbuffer_init(host1x, pb);
+
+	return 0;
+
+fail:
+	host1x_pushbuffer_destroy(pb);
+	return -ENOMEM;
+}
+
+/*
+ * Push two words to the push buffer
+ * Caller must ensure push buffer is not full
+ */
+static void host1x_pushbuffer_push(struct push_buffer *pb, u32 op1, u32 op2)
+{
+	u32 pos = pb->pos;
+	u32 *p = (u32 *)((u32)pb->mapped + pos);
+	WARN_ON(pos == pb->fence);
+	*(p++) = op1;
+	*(p++) = op2;
+	pb->pos = (pos + 8) & (pb->size_bytes - 1);
+}
+
+/*
+ * Pop a number of two word slots from the push buffer
+ * Caller must ensure push buffer is not empty
+ */
+static void host1x_pushbuffer_pop(struct push_buffer *pb, unsigned int slots)
+{
+	/* Advance the next write position */
+	pb->fence = (pb->fence + slots * 8) & (pb->size_bytes - 1);
+}
+
+/*
+ * Return the number of two word slots free in the push buffer
+ */
+static u32 host1x_pushbuffer_space(struct push_buffer *pb)
+{
+	return ((pb->fence - pb->pos) & (pb->size_bytes - 1)) / 8;
+}
+
+/*
+ * Sleep (if necessary) until the requested event happens
+ *   - CDMA_EVENT_SYNC_QUEUE_EMPTY : sync queue is completely empty.
+ *     - Returns 1
+ *   - CDMA_EVENT_PUSH_BUFFER_SPACE : there is space in the push buffer
+ *     - Return the amount of space (> 0)
+ * Must be called with the cdma lock held.
+ */
+unsigned int host1x_cdma_wait_locked(struct host1x_cdma *cdma,
+				     enum cdma_event event)
+{
+	for (;;) {
+		unsigned int space;
+
+		if (event == CDMA_EVENT_SYNC_QUEUE_EMPTY)
+			space = list_empty(&cdma->sync_queue) ? 1 : 0;
+		else if (event == CDMA_EVENT_PUSH_BUFFER_SPACE) {
+			struct push_buffer *pb = &cdma->push_buffer;
+			space = host1x_pushbuffer_space(pb);
+		} else {
+			WARN_ON(1);
+			return -EINVAL;
+		}
+
+		if (space)
+			return space;
+
+		trace_host1x_wait_cdma(dev_name(cdma_to_channel(cdma)->dev),
+				       event);
+
+		/* If somebody has managed to already start waiting, yield */
+		if (cdma->event != CDMA_EVENT_NONE) {
+			mutex_unlock(&cdma->lock);
+			schedule();
+			mutex_lock(&cdma->lock);
+			continue;
+		}
+		cdma->event = event;
+
+		mutex_unlock(&cdma->lock);
+		down(&cdma->sem);
+		mutex_lock(&cdma->lock);
+	}
+	return 0;
+}
+
+/*
+ * Start timer that tracks the time spent by the job.
+ * Must be called with the cdma lock held.
+ */
+static void cdma_start_timer_locked(struct host1x_cdma *cdma,
+				    struct host1x_job *job)
+{
+	struct host1x *host = cdma_to_host1x(cdma);
+
+	if (cdma->timeout.client) {
+		/* timer already started */
+		return;
+	}
+
+	cdma->timeout.client = job->client;
+	cdma->timeout.syncpt = host1x_syncpt_get(host, job->syncpt_id);
+	cdma->timeout.syncpt_val = job->syncpt_end;
+	cdma->timeout.start_ktime = ktime_get();
+
+	schedule_delayed_work(&cdma->timeout.wq,
+			      msecs_to_jiffies(job->timeout));
+}
+
+/*
+ * Stop timer when a buffer submission completes.
+ * Must be called with the cdma lock held.
+ */
+static void stop_cdma_timer_locked(struct host1x_cdma *cdma)
+{
+	cancel_delayed_work(&cdma->timeout.wq);
+	cdma->timeout.client = 0;
+}
+
+/*
+ * For all sync queue entries that have already finished according to the
+ * current sync point registers:
+ *  - unpin & unref their mems
+ *  - pop their push buffer slots
+ *  - remove them from the sync queue
+ * This is normally called from the host code's worker thread, but can be
+ * called manually if necessary.
+ * Must be called with the cdma lock held.
+ */
+static void update_cdma_locked(struct host1x_cdma *cdma)
+{
+	bool signal = false;
+	struct host1x *host1x = cdma_to_host1x(cdma);
+	struct host1x_job *job, *n;
+
+	/* If CDMA is stopped, queue is cleared and we can return */
+	if (!cdma->running)
+		return;
+
+	/*
+	 * Walk the sync queue, reading the sync point registers as necessary,
+	 * to consume as many sync queue entries as possible without blocking
+	 */
+	list_for_each_entry_safe(job, n, &cdma->sync_queue, list) {
+		struct host1x_syncpt *sp =
+			host1x_syncpt_get(host1x, job->syncpt_id);
+
+		/* Check whether this syncpt has completed, and bail if not */
+		if (!host1x_syncpt_is_expired(sp, job->syncpt_end)) {
+			/* Start timer on next pending syncpt */
+			if (job->timeout)
+				cdma_start_timer_locked(cdma, job);
+			break;
+		}
+
+		/* Cancel timeout, when a buffer completes */
+		if (cdma->timeout.client)
+			stop_cdma_timer_locked(cdma);
+
+		/* Unpin the memory */
+		host1x_job_unpin(job);
+
+		/* Pop push buffer slots */
+		if (job->num_slots) {
+			struct push_buffer *pb = &cdma->push_buffer;
+			host1x_pushbuffer_pop(pb, job->num_slots);
+			if (cdma->event == CDMA_EVENT_PUSH_BUFFER_SPACE)
+				signal = true;
+		}
+
+		list_del(&job->list);
+		host1x_job_put(job);
+	}
+
+	if (cdma->event == CDMA_EVENT_SYNC_QUEUE_EMPTY &&
+	    list_empty(&cdma->sync_queue))
+		signal = true;
+
+	if (signal) {
+		cdma->event = CDMA_EVENT_NONE;
+		up(&cdma->sem);
+	}
+}
+
+void host1x_cdma_update_sync_queue(struct host1x_cdma *cdma,
+				   struct device *dev)
+{
+	u32 restart_addr;
+	u32 syncpt_incrs;
+	struct host1x_job *job = NULL;
+	u32 syncpt_val;
+	struct host1x *host1x = cdma_to_host1x(cdma);
+
+	syncpt_val = host1x_syncpt_load(cdma->timeout.syncpt);
+
+	dev_dbg(dev, "%s: starting cleanup (thresh %d)\n",
+		__func__, syncpt_val);
+
+	/*
+	 * Move the sync_queue read pointer to the first entry that hasn't
+	 * completed based on the current HW syncpt value. It's likely there
+	 * won't be any (i.e. we're still at the head), but covers the case
+	 * where a syncpt incr happens just prior/during the teardown.
+	 */
+
+	dev_dbg(dev, "%s: skip completed buffers still in sync_queue\n",
+		__func__);
+
+	list_for_each_entry(job, &cdma->sync_queue, list) {
+		if (syncpt_val < job->syncpt_end)
+			break;
+
+		host1x_job_dump(dev, job);
+	}
+
+	/*
+	 * Walk the sync_queue, first incrementing with the CPU syncpts that
+	 * are partially executed (the first buffer) or fully skipped while
+	 * still in the current context (slots are also NOP-ed).
+	 *
+	 * At the point contexts are interleaved, syncpt increments must be
+	 * done inline with the pushbuffer from a GATHER buffer to maintain
+	 * the order (slots are modified to be a GATHER of syncpt incrs).
+	 *
+	 * Note: save in restart_addr the location where the timed out buffer
+	 * started in the PB, so we can start the refetch from there (with the
+	 * modified NOP-ed PB slots). This lets things appear to have completed
+	 * properly for this buffer and resources are freed.
+	 */
+
+	dev_dbg(dev, "%s: perform CPU incr on pending same ctx buffers\n",
+		__func__);
+
+	if (!list_empty(&cdma->sync_queue))
+		restart_addr = job->first_get;
+	else
+		restart_addr = cdma->last_pos;
+
+	/* do CPU increments as long as this context continues */
+	list_for_each_entry_from(job, &cdma->sync_queue, list) {
+		/* different context, gets us out of this loop */
+		if (job->client != cdma->timeout.client)
+			break;
+
+		/* won't need a timeout when replayed */
+		job->timeout = 0;
+
+		syncpt_incrs = job->syncpt_end - syncpt_val;
+		dev_dbg(dev, "%s: CPU incr (%d)\n", __func__, syncpt_incrs);
+
+		host1x_job_dump(dev, job);
+
+		/* safe to use CPU to incr syncpts */
+		host1x_hw_cdma_timeout_cpu_incr(host1x, cdma, job->first_get,
+						syncpt_incrs, job->syncpt_end,
+						job->num_slots);
+
+		syncpt_val += syncpt_incrs;
+	}
+
+	/* The following sumbits from the same client may be dependent on the
+	 * failed submit and therefore they may fail. Force a small timeout
+	 * to make the queue cleanup faster */
+
+	list_for_each_entry_from(job, &cdma->sync_queue, list)
+		if (job->client == cdma->timeout.client)
+			job->timeout = min_t(unsigned int, job->timeout, 500);
+
+	dev_dbg(dev, "%s: finished sync_queue modification\n", __func__);
+
+	/* roll back DMAGET and start up channel again */
+	host1x_hw_cdma_resume(host1x, cdma, restart_addr);
+}
+
+/*
+ * Create a cdma
+ */
+int host1x_cdma_init(struct host1x_cdma *cdma)
+{
+	int err;
+
+	mutex_init(&cdma->lock);
+	sema_init(&cdma->sem, 0);
+
+	INIT_LIST_HEAD(&cdma->sync_queue);
+
+	cdma->event = CDMA_EVENT_NONE;
+	cdma->running = false;
+	cdma->torndown = false;
+
+	err = host1x_pushbuffer_init(&cdma->push_buffer);
+	if (err)
+		return err;
+	return 0;
+}
+
+/*
+ * Destroy a cdma
+ */
+int host1x_cdma_deinit(struct host1x_cdma *cdma)
+{
+	struct push_buffer *pb = &cdma->push_buffer;
+	struct host1x *host1x = cdma_to_host1x(cdma);
+
+	if (cdma->running) {
+		pr_warn("%s: CDMA still running\n", __func__);
+		return -EBUSY;
+	}
+
+	host1x_pushbuffer_destroy(pb);
+	host1x_hw_cdma_timeout_destroy(host1x, cdma);
+
+	return 0;
+}
+
+/*
+ * Begin a cdma submit
+ */
+int host1x_cdma_begin(struct host1x_cdma *cdma, struct host1x_job *job)
+{
+	struct host1x *host1x = cdma_to_host1x(cdma);
+
+	mutex_lock(&cdma->lock);
+
+	if (job->timeout) {
+		/* init state on first submit with timeout value */
+		if (!cdma->timeout.initialized) {
+			int err;
+			err = host1x_hw_cdma_timeout_init(host1x, cdma,
+							  job->syncpt_id);
+			if (err) {
+				mutex_unlock(&cdma->lock);
+				return err;
+			}
+		}
+	}
+	if (!cdma->running)
+		host1x_hw_cdma_start(host1x, cdma);
+
+	cdma->slots_free = 0;
+	cdma->slots_used = 0;
+	cdma->first_get = cdma->push_buffer.pos;
+
+	trace_host1x_cdma_begin(dev_name(job->channel->dev));
+	return 0;
+}
+
+/*
+ * Push two words into a push buffer slot
+ * Blocks as necessary if the push buffer is full.
+ */
+void host1x_cdma_push(struct host1x_cdma *cdma, u32 op1, u32 op2)
+{
+	struct host1x *host1x = cdma_to_host1x(cdma);
+	struct push_buffer *pb = &cdma->push_buffer;
+	u32 slots_free = cdma->slots_free;
+
+	if (host1x_debug_trace_cmdbuf)
+		trace_host1x_cdma_push(dev_name(cdma_to_channel(cdma)->dev),
+				       op1, op2);
+
+	if (slots_free == 0) {
+		host1x_hw_cdma_flush(host1x, cdma);
+		slots_free = host1x_cdma_wait_locked(cdma,
+						CDMA_EVENT_PUSH_BUFFER_SPACE);
+	}
+	cdma->slots_free = slots_free - 1;
+	cdma->slots_used++;
+	host1x_pushbuffer_push(pb, op1, op2);
+}
+
+/*
+ * End a cdma submit
+ * Kick off DMA, add job to the sync queue, and a number of slots to be freed
+ * from the pushbuffer. The handles for a submit must all be pinned at the same
+ * time, but they can be unpinned in smaller chunks.
+ */
+void host1x_cdma_end(struct host1x_cdma *cdma,
+		     struct host1x_job *job)
+{
+	struct host1x *host1x = cdma_to_host1x(cdma);
+	bool idle = list_empty(&cdma->sync_queue);
+
+	host1x_hw_cdma_flush(host1x, cdma);
+
+	job->first_get = cdma->first_get;
+	job->num_slots = cdma->slots_used;
+	host1x_job_get(job);
+	list_add_tail(&job->list, &cdma->sync_queue);
+
+	/* start timer on idle -> active transitions */
+	if (job->timeout && idle)
+		cdma_start_timer_locked(cdma, job);
+
+	trace_host1x_cdma_end(dev_name(job->channel->dev));
+	mutex_unlock(&cdma->lock);
+}
+
+/*
+ * Update cdma state according to current sync point values
+ */
+void host1x_cdma_update(struct host1x_cdma *cdma)
+{
+	mutex_lock(&cdma->lock);
+	update_cdma_locked(cdma);
+	mutex_unlock(&cdma->lock);
+}
diff --git a/drivers/gpu/host1x/cdma.h b/drivers/gpu/host1x/cdma.h
new file mode 100644
index 000000000000..313c4b784348
--- /dev/null
+++ b/drivers/gpu/host1x/cdma.h
@@ -0,0 +1,100 @@
+/*
+ * Tegra host1x Command DMA
+ *
+ * Copyright (c) 2010-2013, NVIDIA Corporation.
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms and conditions of the GNU General Public License,
+ * version 2, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program.  If not, see <http://www.gnu.org/licenses/>.
+ */
+
+#ifndef __HOST1X_CDMA_H
+#define __HOST1X_CDMA_H
+
+#include <linux/sched.h>
+#include <linux/semaphore.h>
+#include <linux/list.h>
+
+struct host1x_syncpt;
+struct host1x_userctx_timeout;
+struct host1x_job;
+
+/*
+ * cdma
+ *
+ * This is in charge of a host command DMA channel.
+ * Sends ops to a push buffer, and takes responsibility for unpinning
+ * (& possibly freeing) of memory after those ops have completed.
+ * Producer:
+ *	begin
+ *		push - send ops to the push buffer
+ *	end - start command DMA and enqueue handles to be unpinned
+ * Consumer:
+ *	update - call to update sync queue and push buffer, unpin memory
+ */
+
+struct push_buffer {
+	u32 *mapped;			/* mapped pushbuffer memory */
+	dma_addr_t phys;		/* physical address of pushbuffer */
+	u32 fence;			/* index we've written */
+	u32 pos;			/* index to write to */
+	u32 size_bytes;
+};
+
+struct buffer_timeout {
+	struct delayed_work wq;		/* work queue */
+	bool initialized;		/* timer one-time setup flag */
+	struct host1x_syncpt *syncpt;	/* buffer completion syncpt */
+	u32 syncpt_val;			/* syncpt value when completed */
+	ktime_t start_ktime;		/* starting time */
+	/* context timeout information */
+	int client;
+};
+
+enum cdma_event {
+	CDMA_EVENT_NONE,		/* not waiting for any event */
+	CDMA_EVENT_SYNC_QUEUE_EMPTY,	/* wait for empty sync queue */
+	CDMA_EVENT_PUSH_BUFFER_SPACE	/* wait for space in push buffer */
+};
+
+struct host1x_cdma {
+	struct mutex lock;		/* controls access to shared state */
+	struct semaphore sem;		/* signalled when event occurs */
+	enum cdma_event event;		/* event that sem is waiting for */
+	unsigned int slots_used;	/* pb slots used in current submit */
+	unsigned int slots_free;	/* pb slots free in current submit */
+	unsigned int first_get;		/* DMAGET value, where submit begins */
+	unsigned int last_pos;		/* last value written to DMAPUT */
+	struct push_buffer push_buffer;	/* channel's push buffer */
+	struct list_head sync_queue;	/* job queue */
+	struct buffer_timeout timeout;	/* channel's timeout state/wq */
+	bool running;
+	bool torndown;
+};
+
+#define cdma_to_channel(cdma) container_of(cdma, struct host1x_channel, cdma)
+#define cdma_to_host1x(cdma) dev_get_drvdata(cdma_to_channel(cdma)->dev->parent)
+#define pb_to_cdma(pb) container_of(pb, struct host1x_cdma, push_buffer)
+
+int host1x_cdma_init(struct host1x_cdma *cdma);
+int host1x_cdma_deinit(struct host1x_cdma *cdma);
+void host1x_cdma_stop(struct host1x_cdma *cdma);
+int host1x_cdma_begin(struct host1x_cdma *cdma, struct host1x_job *job);
+void host1x_cdma_push(struct host1x_cdma *cdma, u32 op1, u32 op2);
+void host1x_cdma_end(struct host1x_cdma *cdma, struct host1x_job *job);
+void host1x_cdma_update(struct host1x_cdma *cdma);
+void host1x_cdma_peek(struct host1x_cdma *cdma, u32 dmaget, int slot,
+		      u32 *out);
+unsigned int host1x_cdma_wait_locked(struct host1x_cdma *cdma,
+				     enum cdma_event event);
+void host1x_cdma_update_sync_queue(struct host1x_cdma *cdma,
+				   struct device *dev);
+#endif
diff --git a/drivers/gpu/host1x/channel.c b/drivers/gpu/host1x/channel.c
new file mode 100644
index 000000000000..83ea51b9f0fc
--- /dev/null
+++ b/drivers/gpu/host1x/channel.c
@@ -0,0 +1,126 @@
+/*
+ * Tegra host1x Channel
+ *
+ * Copyright (c) 2010-2013, NVIDIA Corporation.
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms and conditions of the GNU General Public License,
+ * version 2, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program.  If not, see <http://www.gnu.org/licenses/>.
+ */
+
+#include <linux/slab.h>
+#include <linux/module.h>
+
+#include "channel.h"
+#include "dev.h"
+#include "job.h"
+
+/* Constructor for the host1x device list */
+int host1x_channel_list_init(struct host1x *host)
+{
+	INIT_LIST_HEAD(&host->chlist.list);
+	mutex_init(&host->chlist_mutex);
+
+	if (host->info->nb_channels > BITS_PER_LONG) {
+		WARN(1, "host1x hardware has more channels than supported by the driver\n");
+		return -ENOSYS;
+	}
+
+	return 0;
+}
+
+int host1x_job_submit(struct host1x_job *job)
+{
+	struct host1x *host = dev_get_drvdata(job->channel->dev->parent);
+
+	return host1x_hw_channel_submit(host, job);
+}
+
+struct host1x_channel *host1x_channel_get(struct host1x_channel *channel)
+{
+	int err = 0;
+
+	mutex_lock(&channel->reflock);
+
+	if (channel->refcount == 0)
+		err = host1x_cdma_init(&channel->cdma);
+
+	if (!err)
+		channel->refcount++;
+
+	mutex_unlock(&channel->reflock);
+
+	return err ? NULL : channel;
+}
+
+void host1x_channel_put(struct host1x_channel *channel)
+{
+	mutex_lock(&channel->reflock);
+
+	if (channel->refcount == 1) {
+		struct host1x *host = dev_get_drvdata(channel->dev->parent);
+
+		host1x_hw_cdma_stop(host, &channel->cdma);
+		host1x_cdma_deinit(&channel->cdma);
+	}
+
+	channel->refcount--;
+
+	mutex_unlock(&channel->reflock);
+}
+
+struct host1x_channel *host1x_channel_request(struct device *dev)
+{
+	struct host1x *host = dev_get_drvdata(dev->parent);
+	int max_channels = host->info->nb_channels;
+	struct host1x_channel *channel = NULL;
+	int index, err;
+
+	mutex_lock(&host->chlist_mutex);
+
+	index = find_first_zero_bit(&host->allocated_channels, max_channels);
+	if (index >= max_channels)
+		goto fail;
+
+	channel = kzalloc(sizeof(*channel), GFP_KERNEL);
+	if (!channel)
+		goto fail;
+
+	err = host1x_hw_channel_init(host, channel, index);
+	if (err < 0)
+		goto fail;
+
+	/* Link device to host1x_channel */
+	channel->dev = dev;
+
+	/* Add to channel list */
+	list_add_tail(&channel->list, &host->chlist.list);
+
+	host->allocated_channels |= BIT(index);
+
+	mutex_unlock(&host->chlist_mutex);
+	return channel;
+
+fail:
+	dev_err(dev, "failed to init channel\n");
+	kfree(channel);
+	mutex_unlock(&host->chlist_mutex);
+	return NULL;
+}
+
+void host1x_channel_free(struct host1x_channel *channel)
+{
+	struct host1x *host = dev_get_drvdata(channel->dev->parent);
+
+	host->allocated_channels &= ~BIT(channel->id);
+	list_del(&channel->list);
+	kfree(channel);
+}
diff --git a/drivers/gpu/host1x/channel.h b/drivers/gpu/host1x/channel.h
new file mode 100644
index 000000000000..48723b8eea42
--- /dev/null
+++ b/drivers/gpu/host1x/channel.h
@@ -0,0 +1,52 @@
+/*
+ * Tegra host1x Channel
+ *
+ * Copyright (c) 2010-2013, NVIDIA Corporation.
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms and conditions of the GNU General Public License,
+ * version 2, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program.  If not, see <http://www.gnu.org/licenses/>.
+ */
+
+#ifndef __HOST1X_CHANNEL_H
+#define __HOST1X_CHANNEL_H
+
+#include <linux/io.h>
+
+#include "cdma.h"
+
+struct host1x;
+
+struct host1x_channel {
+	struct list_head list;
+
+	unsigned int refcount;
+	unsigned int id;
+	struct mutex reflock;
+	struct mutex submitlock;
+	void __iomem *regs;
+	struct device *dev;
+	struct host1x_cdma cdma;
+};
+
+/* channel list operations */
+int host1x_channel_list_init(struct host1x *host);
+
+struct host1x_channel *host1x_channel_request(struct device *dev);
+void host1x_channel_free(struct host1x_channel *channel);
+struct host1x_channel *host1x_channel_get(struct host1x_channel *channel);
+void host1x_channel_put(struct host1x_channel *channel);
+int host1x_job_submit(struct host1x_job *job);
+
+#define host1x_for_each_channel(host, channel)				\
+	list_for_each_entry(channel, &host->chlist.list, list)
+
+#endif
diff --git a/drivers/gpu/host1x/debug.c b/drivers/gpu/host1x/debug.c
new file mode 100644
index 000000000000..3ec7d77de24d
--- /dev/null
+++ b/drivers/gpu/host1x/debug.c
@@ -0,0 +1,210 @@
+/*
+ * Copyright (C) 2010 Google, Inc.
+ * Author: Erik Gilling <konkers@android.com>
+ *
+ * Copyright (C) 2011-2013 NVIDIA Corporation
+ *
+ * This software is licensed under the terms of the GNU General Public
+ * License version 2, as published by the Free Software Foundation, and
+ * may be copied, distributed, and modified under those terms.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ */
+
+#include <linux/debugfs.h>
+#include <linux/seq_file.h>
+#include <linux/uaccess.h>
+
+#include <linux/io.h>
+
+#include "dev.h"
+#include "debug.h"
+#include "channel.h"
+
+unsigned int host1x_debug_trace_cmdbuf;
+
+static pid_t host1x_debug_force_timeout_pid;
+static u32 host1x_debug_force_timeout_val;
+static u32 host1x_debug_force_timeout_channel;
+
+void host1x_debug_output(struct output *o, const char *fmt, ...)
+{
+	va_list args;
+	int len;
+
+	va_start(args, fmt);
+	len = vsnprintf(o->buf, sizeof(o->buf), fmt, args);
+	va_end(args);
+	o->fn(o->ctx, o->buf, len);
+}
+
+static int show_channels(struct host1x_channel *ch, void *data, bool show_fifo)
+{
+	struct host1x *m = dev_get_drvdata(ch->dev->parent);
+	struct output *o = data;
+
+	mutex_lock(&ch->reflock);
+	if (ch->refcount) {
+		mutex_lock(&ch->cdma.lock);
+		if (show_fifo)
+			host1x_hw_show_channel_fifo(m, ch, o);
+		host1x_hw_show_channel_cdma(m, ch, o);
+		mutex_unlock(&ch->cdma.lock);
+	}
+	mutex_unlock(&ch->reflock);
+
+	return 0;
+}
+
+static void show_syncpts(struct host1x *m, struct output *o)
+{
+	int i;
+	host1x_debug_output(o, "---- syncpts ----\n");
+	for (i = 0; i < host1x_syncpt_nb_pts(m); i++) {
+		u32 max = host1x_syncpt_read_max(m->syncpt + i);
+		u32 min = host1x_syncpt_load(m->syncpt + i);
+		if (!min && !max)
+			continue;
+		host1x_debug_output(o, "id %d (%s) min %d max %d\n",
+				    i, m->syncpt[i].name, min, max);
+	}
+
+	for (i = 0; i < host1x_syncpt_nb_bases(m); i++) {
+		u32 base_val;
+		base_val = host1x_syncpt_load_wait_base(m->syncpt + i);
+		if (base_val)
+			host1x_debug_output(o, "waitbase id %d val %d\n", i,
+					    base_val);
+	}
+
+	host1x_debug_output(o, "\n");
+}
+
+static void show_all(struct host1x *m, struct output *o)
+{
+	struct host1x_channel *ch;
+
+	host1x_hw_show_mlocks(m, o);
+	show_syncpts(m, o);
+	host1x_debug_output(o, "---- channels ----\n");
+
+	host1x_for_each_channel(m, ch)
+		show_channels(ch, o, true);
+}
+
+#ifdef CONFIG_DEBUG_FS
+static void show_all_no_fifo(struct host1x *host1x, struct output *o)
+{
+	struct host1x_channel *ch;
+
+	host1x_hw_show_mlocks(host1x, o);
+	show_syncpts(host1x, o);
+	host1x_debug_output(o, "---- channels ----\n");
+
+	host1x_for_each_channel(host1x, ch)
+		show_channels(ch, o, false);
+}
+
+static int host1x_debug_show_all(struct seq_file *s, void *unused)
+{
+	struct output o = {
+		.fn = write_to_seqfile,
+		.ctx = s
+	};
+	show_all(s->private, &o);
+	return 0;
+}
+
+static int host1x_debug_show(struct seq_file *s, void *unused)
+{
+	struct output o = {
+		.fn = write_to_seqfile,
+		.ctx = s
+	};
+	show_all_no_fifo(s->private, &o);
+	return 0;
+}
+
+static int host1x_debug_open_all(struct inode *inode, struct file *file)
+{
+	return single_open(file, host1x_debug_show_all, inode->i_private);
+}
+
+static const struct file_operations host1x_debug_all_fops = {
+	.open		= host1x_debug_open_all,
+	.read		= seq_read,
+	.llseek		= seq_lseek,
+	.release	= single_release,
+};
+
+static int host1x_debug_open(struct inode *inode, struct file *file)
+{
+	return single_open(file, host1x_debug_show, inode->i_private);
+}
+
+static const struct file_operations host1x_debug_fops = {
+	.open		= host1x_debug_open,
+	.read		= seq_read,
+	.llseek		= seq_lseek,
+	.release	= single_release,
+};
+
+void host1x_debug_init(struct host1x *host1x)
+{
+	struct dentry *de = debugfs_create_dir("tegra-host1x", NULL);
+
+	if (!de)
+		return;
+
+	/* Store the created entry */
+	host1x->debugfs = de;
+
+	debugfs_create_file("status", S_IRUGO, de, host1x, &host1x_debug_fops);
+	debugfs_create_file("status_all", S_IRUGO, de, host1x,
+			    &host1x_debug_all_fops);
+
+	debugfs_create_u32("trace_cmdbuf", S_IRUGO|S_IWUSR, de,
+			   &host1x_debug_trace_cmdbuf);
+
+	host1x_hw_debug_init(host1x, de);
+
+	debugfs_create_u32("force_timeout_pid", S_IRUGO|S_IWUSR, de,
+			   &host1x_debug_force_timeout_pid);
+	debugfs_create_u32("force_timeout_val", S_IRUGO|S_IWUSR, de,
+			   &host1x_debug_force_timeout_val);
+	debugfs_create_u32("force_timeout_channel", S_IRUGO|S_IWUSR, de,
+			   &host1x_debug_force_timeout_channel);
+}
+
+void host1x_debug_deinit(struct host1x *host1x)
+{
+	debugfs_remove_recursive(host1x->debugfs);
+}
+#else
+void host1x_debug_init(struct host1x *host1x)
+{
+}
+void host1x_debug_deinit(struct host1x *host1x)
+{
+}
+#endif
+
+void host1x_debug_dump(struct host1x *host1x)
+{
+	struct output o = {
+		.fn = write_to_printk
+	};
+	show_all(host1x, &o);
+}
+
+void host1x_debug_dump_syncpts(struct host1x *host1x)
+{
+	struct output o = {
+		.fn = write_to_printk
+	};
+	show_syncpts(host1x, &o);
+}
diff --git a/drivers/gpu/host1x/debug.h b/drivers/gpu/host1x/debug.h
new file mode 100644
index 000000000000..4595b2e0799f
--- /dev/null
+++ b/drivers/gpu/host1x/debug.h
@@ -0,0 +1,51 @@
+/*
+ * Tegra host1x Debug
+ *
+ * Copyright (c) 2011-2013 NVIDIA Corporation.
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms and conditions of the GNU General Public License,
+ * version 2, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program.  If not, see <http://www.gnu.org/licenses/>.
+ */
+#ifndef __HOST1X_DEBUG_H
+#define __HOST1X_DEBUG_H
+
+#include <linux/debugfs.h>
+#include <linux/seq_file.h>
+
+struct host1x;
+
+struct output {
+	void (*fn)(void *ctx, const char *str, size_t len);
+	void *ctx;
+	char buf[256];
+};
+
+static inline void write_to_seqfile(void *ctx, const char *str, size_t len)
+{
+	seq_write((struct seq_file *)ctx, str, len);
+}
+
+static inline void write_to_printk(void *ctx, const char *str, size_t len)
+{
+	pr_info("%s", str);
+}
+
+void __printf(2, 3) host1x_debug_output(struct output *o, const char *fmt, ...);
+
+extern unsigned int host1x_debug_trace_cmdbuf;
+
+void host1x_debug_init(struct host1x *host1x);
+void host1x_debug_deinit(struct host1x *host1x);
+void host1x_debug_dump(struct host1x *host1x);
+void host1x_debug_dump_syncpts(struct host1x *host1x);
+
+#endif
diff --git a/drivers/gpu/host1x/dev.c b/drivers/gpu/host1x/dev.c
new file mode 100644
index 000000000000..28e28a23d444
--- /dev/null
+++ b/drivers/gpu/host1x/dev.c
@@ -0,0 +1,246 @@
+/*
+ * Tegra host1x driver
+ *
+ * Copyright (c) 2010-2013, NVIDIA Corporation.
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms and conditions of the GNU General Public License,
+ * version 2, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program.  If not, see <http://www.gnu.org/licenses/>.
+ */
+
+#include <linux/module.h>
+#include <linux/list.h>
+#include <linux/slab.h>
+#include <linux/of.h>
+#include <linux/of_device.h>
+#include <linux/clk.h>
+#include <linux/io.h>
+
+#define CREATE_TRACE_POINTS
+#include <trace/events/host1x.h>
+
+#include "dev.h"
+#include "intr.h"
+#include "channel.h"
+#include "debug.h"
+#include "hw/host1x01.h"
+#include "host1x_client.h"
+
+void host1x_set_drm_data(struct device *dev, void *data)
+{
+	struct host1x *host1x = dev_get_drvdata(dev);
+	host1x->drm_data = data;
+}
+
+void *host1x_get_drm_data(struct device *dev)
+{
+	struct host1x *host1x = dev_get_drvdata(dev);
+	return host1x->drm_data;
+}
+
+void host1x_sync_writel(struct host1x *host1x, u32 v, u32 r)
+{
+	void __iomem *sync_regs = host1x->regs + host1x->info->sync_offset;
+
+	writel(v, sync_regs + r);
+}
+
+u32 host1x_sync_readl(struct host1x *host1x, u32 r)
+{
+	void __iomem *sync_regs = host1x->regs + host1x->info->sync_offset;
+
+	return readl(sync_regs + r);
+}
+
+void host1x_ch_writel(struct host1x_channel *ch, u32 v, u32 r)
+{
+	writel(v, ch->regs + r);
+}
+
+u32 host1x_ch_readl(struct host1x_channel *ch, u32 r)
+{
+	return readl(ch->regs + r);
+}
+
+static const struct host1x_info host1x01_info = {
+	.nb_channels	= 8,
+	.nb_pts		= 32,
+	.nb_mlocks	= 16,
+	.nb_bases	= 8,
+	.init		= host1x01_init,
+	.sync_offset	= 0x3000,
+};
+
+static struct of_device_id host1x_of_match[] = {
+	{ .compatible = "nvidia,tegra30-host1x", .data = &host1x01_info, },
+	{ .compatible = "nvidia,tegra20-host1x", .data = &host1x01_info, },
+	{ },
+};
+MODULE_DEVICE_TABLE(of, host1x_of_match);
+
+static int host1x_probe(struct platform_device *pdev)
+{
+	const struct of_device_id *id;
+	struct host1x *host;
+	struct resource *regs;
+	int syncpt_irq;
+	int err;
+
+	id = of_match_device(host1x_of_match, &pdev->dev);
+	if (!id)
+		return -EINVAL;
+
+	regs = platform_get_resource(pdev, IORESOURCE_MEM, 0);
+	if (!regs) {
+		dev_err(&pdev->dev, "failed to get registers\n");
+		return -ENXIO;
+	}
+
+	syncpt_irq = platform_get_irq(pdev, 0);
+	if (syncpt_irq < 0) {
+		dev_err(&pdev->dev, "failed to get IRQ\n");
+		return -ENXIO;
+	}
+
+	host = devm_kzalloc(&pdev->dev, sizeof(*host), GFP_KERNEL);
+	if (!host)
+		return -ENOMEM;
+
+	host->dev = &pdev->dev;
+	host->info = id->data;
+
+	/* set common host1x device data */
+	platform_set_drvdata(pdev, host);
+
+	host->regs = devm_ioremap_resource(&pdev->dev, regs);
+	if (IS_ERR(host->regs))
+		return PTR_ERR(host->regs);
+
+	if (host->info->init) {
+		err = host->info->init(host);
+		if (err)
+			return err;
+	}
+
+	host->clk = devm_clk_get(&pdev->dev, NULL);
+	if (IS_ERR(host->clk)) {
+		dev_err(&pdev->dev, "failed to get clock\n");
+		err = PTR_ERR(host->clk);
+		return err;
+	}
+
+	err = host1x_channel_list_init(host);
+	if (err) {
+		dev_err(&pdev->dev, "failed to initialize channel list\n");
+		return err;
+	}
+
+	err = clk_prepare_enable(host->clk);
+	if (err < 0) {
+		dev_err(&pdev->dev, "failed to enable clock\n");
+		return err;
+	}
+
+	err = host1x_syncpt_init(host);
+	if (err) {
+		dev_err(&pdev->dev, "failed to initialize syncpts\n");
+		return err;
+	}
+
+	err = host1x_intr_init(host, syncpt_irq);
+	if (err) {
+		dev_err(&pdev->dev, "failed to initialize interrupts\n");
+		goto fail_deinit_syncpt;
+	}
+
+	host1x_debug_init(host);
+
+	host1x_drm_alloc(pdev);
+
+	return 0;
+
+fail_deinit_syncpt:
+	host1x_syncpt_deinit(host);
+	return err;
+}
+
+static int __exit host1x_remove(struct platform_device *pdev)
+{
+	struct host1x *host = platform_get_drvdata(pdev);
+
+	host1x_intr_deinit(host);
+	host1x_syncpt_deinit(host);
+	clk_disable_unprepare(host->clk);
+
+	return 0;
+}
+
+static struct platform_driver tegra_host1x_driver = {
+	.probe = host1x_probe,
+	.remove = __exit_p(host1x_remove),
+	.driver = {
+		.owner = THIS_MODULE,
+		.name = "tegra-host1x",
+		.of_match_table = host1x_of_match,
+	},
+};
+
+static int __init tegra_host1x_init(void)
+{
+	int err;
+
+	err = platform_driver_register(&tegra_host1x_driver);
+	if (err < 0)
+		return err;
+
+#ifdef CONFIG_DRM_TEGRA
+	err = platform_driver_register(&tegra_dc_driver);
+	if (err < 0)
+		goto unregister_host1x;
+
+	err = platform_driver_register(&tegra_hdmi_driver);
+	if (err < 0)
+		goto unregister_dc;
+
+	err = platform_driver_register(&tegra_gr2d_driver);
+	if (err < 0)
+		goto unregister_hdmi;
+#endif
+
+	return 0;
+
+#ifdef CONFIG_DRM_TEGRA
+unregister_hdmi:
+	platform_driver_unregister(&tegra_hdmi_driver);
+unregister_dc:
+	platform_driver_unregister(&tegra_dc_driver);
+unregister_host1x:
+	platform_driver_unregister(&tegra_host1x_driver);
+	return err;
+#endif
+}
+module_init(tegra_host1x_init);
+
+static void __exit tegra_host1x_exit(void)
+{
+#ifdef CONFIG_DRM_TEGRA
+	platform_driver_unregister(&tegra_gr2d_driver);
+	platform_driver_unregister(&tegra_hdmi_driver);
+	platform_driver_unregister(&tegra_dc_driver);
+#endif
+	platform_driver_unregister(&tegra_host1x_driver);
+}
+module_exit(tegra_host1x_exit);
+
+MODULE_AUTHOR("Thierry Reding <thierry.reding@avionic-design.de>");
+MODULE_AUTHOR("Terje Bergstrom <tbergstrom@nvidia.com>");
+MODULE_DESCRIPTION("Host1x driver for Tegra products");
+MODULE_LICENSE("GPL");
diff --git a/drivers/gpu/host1x/dev.h b/drivers/gpu/host1x/dev.h
new file mode 100644
index 000000000000..a1607d6e135b
--- /dev/null
+++ b/drivers/gpu/host1x/dev.h
@@ -0,0 +1,308 @@
+/*
+ * Copyright (c) 2012-2013, NVIDIA Corporation.
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms and conditions of the GNU General Public License,
+ * version 2, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program.  If not, see <http://www.gnu.org/licenses/>.
+ */
+
+#ifndef HOST1X_DEV_H
+#define HOST1X_DEV_H
+
+#include <linux/platform_device.h>
+#include <linux/device.h>
+
+#include "channel.h"
+#include "syncpt.h"
+#include "intr.h"
+#include "cdma.h"
+#include "job.h"
+
+struct host1x_syncpt;
+struct host1x_channel;
+struct host1x_cdma;
+struct host1x_job;
+struct push_buffer;
+struct output;
+struct dentry;
+
+struct host1x_channel_ops {
+	int (*init)(struct host1x_channel *channel, struct host1x *host,
+		    unsigned int id);
+	int (*submit)(struct host1x_job *job);
+};
+
+struct host1x_cdma_ops {
+	void (*start)(struct host1x_cdma *cdma);
+	void (*stop)(struct host1x_cdma *cdma);
+	void (*flush)(struct  host1x_cdma *cdma);
+	int (*timeout_init)(struct host1x_cdma *cdma, u32 syncpt_id);
+	void (*timeout_destroy)(struct host1x_cdma *cdma);
+	void (*freeze)(struct host1x_cdma *cdma);
+	void (*resume)(struct host1x_cdma *cdma, u32 getptr);
+	void (*timeout_cpu_incr)(struct host1x_cdma *cdma, u32 getptr,
+				 u32 syncpt_incrs, u32 syncval, u32 nr_slots);
+};
+
+struct host1x_pushbuffer_ops {
+	void (*init)(struct push_buffer *pb);
+};
+
+struct host1x_debug_ops {
+	void (*debug_init)(struct dentry *de);
+	void (*show_channel_cdma)(struct host1x *host,
+				  struct host1x_channel *ch,
+				  struct output *o);
+	void (*show_channel_fifo)(struct host1x *host,
+				  struct host1x_channel *ch,
+				  struct output *o);
+	void (*show_mlocks)(struct host1x *host, struct output *output);
+
+};
+
+struct host1x_syncpt_ops {
+	void (*restore)(struct host1x_syncpt *syncpt);
+	void (*restore_wait_base)(struct host1x_syncpt *syncpt);
+	void (*load_wait_base)(struct host1x_syncpt *syncpt);
+	u32 (*load)(struct host1x_syncpt *syncpt);
+	void (*cpu_incr)(struct host1x_syncpt *syncpt);
+	int (*patch_wait)(struct host1x_syncpt *syncpt, void *patch_addr);
+};
+
+struct host1x_intr_ops {
+	int (*init_host_sync)(struct host1x *host, u32 cpm,
+		void (*syncpt_thresh_work)(struct work_struct *work));
+	void (*set_syncpt_threshold)(
+		struct host1x *host, u32 id, u32 thresh);
+	void (*enable_syncpt_intr)(struct host1x *host, u32 id);
+	void (*disable_syncpt_intr)(struct host1x *host, u32 id);
+	void (*disable_all_syncpt_intrs)(struct host1x *host);
+	int (*free_syncpt_irq)(struct host1x *host);
+};
+
+struct host1x_info {
+	int	nb_channels;		/* host1x: num channels supported */
+	int	nb_pts;			/* host1x: num syncpoints supported */
+	int	nb_bases;		/* host1x: num syncpoints supported */
+	int	nb_mlocks;		/* host1x: number of mlocks */
+	int	(*init)(struct host1x *); /* initialize per SoC ops */
+	int	sync_offset;
+};
+
+struct host1x {
+	const struct host1x_info *info;
+
+	void __iomem *regs;
+	struct host1x_syncpt *syncpt;
+	struct device *dev;
+	struct clk *clk;
+
+	struct mutex intr_mutex;
+	struct workqueue_struct *intr_wq;
+	int intr_syncpt_irq;
+
+	const struct host1x_syncpt_ops *syncpt_op;
+	const struct host1x_intr_ops *intr_op;
+	const struct host1x_channel_ops *channel_op;
+	const struct host1x_cdma_ops *cdma_op;
+	const struct host1x_pushbuffer_ops *cdma_pb_op;
+	const struct host1x_debug_ops *debug_op;
+
+	struct host1x_syncpt *nop_sp;
+
+	struct mutex chlist_mutex;
+	struct host1x_channel chlist;
+	unsigned long allocated_channels;
+	unsigned int num_allocated_channels;
+
+	struct dentry *debugfs;
+
+	void *drm_data;
+};
+
+void host1x_sync_writel(struct host1x *host1x, u32 r, u32 v);
+u32 host1x_sync_readl(struct host1x *host1x, u32 r);
+void host1x_ch_writel(struct host1x_channel *ch, u32 r, u32 v);
+u32 host1x_ch_readl(struct host1x_channel *ch, u32 r);
+
+static inline void host1x_hw_syncpt_restore(struct host1x *host,
+					    struct host1x_syncpt *sp)
+{
+	host->syncpt_op->restore(sp);
+}
+
+static inline void host1x_hw_syncpt_restore_wait_base(struct host1x *host,
+						      struct host1x_syncpt *sp)
+{
+	host->syncpt_op->restore_wait_base(sp);
+}
+
+static inline void host1x_hw_syncpt_load_wait_base(struct host1x *host,
+						   struct host1x_syncpt *sp)
+{
+	host->syncpt_op->load_wait_base(sp);
+}
+
+static inline u32 host1x_hw_syncpt_load(struct host1x *host,
+					struct host1x_syncpt *sp)
+{
+	return host->syncpt_op->load(sp);
+}
+
+static inline void host1x_hw_syncpt_cpu_incr(struct host1x *host,
+					     struct host1x_syncpt *sp)
+{
+	host->syncpt_op->cpu_incr(sp);
+}
+
+static inline int host1x_hw_syncpt_patch_wait(struct host1x *host,
+					      struct host1x_syncpt *sp,
+					      void *patch_addr)
+{
+	return host->syncpt_op->patch_wait(sp, patch_addr);
+}
+
+static inline int host1x_hw_intr_init_host_sync(struct host1x *host, u32 cpm,
+			void (*syncpt_thresh_work)(struct work_struct *))
+{
+	return host->intr_op->init_host_sync(host, cpm, syncpt_thresh_work);
+}
+
+static inline void host1x_hw_intr_set_syncpt_threshold(struct host1x *host,
+						       u32 id, u32 thresh)
+{
+	host->intr_op->set_syncpt_threshold(host, id, thresh);
+}
+
+static inline void host1x_hw_intr_enable_syncpt_intr(struct host1x *host,
+						     u32 id)
+{
+	host->intr_op->enable_syncpt_intr(host, id);
+}
+
+static inline void host1x_hw_intr_disable_syncpt_intr(struct host1x *host,
+						      u32 id)
+{
+	host->intr_op->disable_syncpt_intr(host, id);
+}
+
+static inline void host1x_hw_intr_disable_all_syncpt_intrs(struct host1x *host)
+{
+	host->intr_op->disable_all_syncpt_intrs(host);
+}
+
+static inline int host1x_hw_intr_free_syncpt_irq(struct host1x *host)
+{
+	return host->intr_op->free_syncpt_irq(host);
+}
+
+static inline int host1x_hw_channel_init(struct host1x *host,
+					 struct host1x_channel *channel,
+					 int chid)
+{
+	return host->channel_op->init(channel, host, chid);
+}
+
+static inline int host1x_hw_channel_submit(struct host1x *host,
+					   struct host1x_job *job)
+{
+	return host->channel_op->submit(job);
+}
+
+static inline void host1x_hw_cdma_start(struct host1x *host,
+					struct host1x_cdma *cdma)
+{
+	host->cdma_op->start(cdma);
+}
+
+static inline void host1x_hw_cdma_stop(struct host1x *host,
+				       struct host1x_cdma *cdma)
+{
+	host->cdma_op->stop(cdma);
+}
+
+static inline void host1x_hw_cdma_flush(struct host1x *host,
+					struct host1x_cdma *cdma)
+{
+	host->cdma_op->flush(cdma);
+}
+
+static inline int host1x_hw_cdma_timeout_init(struct host1x *host,
+					      struct host1x_cdma *cdma,
+					      u32 syncpt_id)
+{
+	return host->cdma_op->timeout_init(cdma, syncpt_id);
+}
+
+static inline void host1x_hw_cdma_timeout_destroy(struct host1x *host,
+						  struct host1x_cdma *cdma)
+{
+	host->cdma_op->timeout_destroy(cdma);
+}
+
+static inline void host1x_hw_cdma_freeze(struct host1x *host,
+					 struct host1x_cdma *cdma)
+{
+	host->cdma_op->freeze(cdma);
+}
+
+static inline void host1x_hw_cdma_resume(struct host1x *host,
+					 struct host1x_cdma *cdma, u32 getptr)
+{
+	host->cdma_op->resume(cdma, getptr);
+}
+
+static inline void host1x_hw_cdma_timeout_cpu_incr(struct host1x *host,
+						   struct host1x_cdma *cdma,
+						   u32 getptr,
+						   u32 syncpt_incrs,
+						   u32 syncval, u32 nr_slots)
+{
+	host->cdma_op->timeout_cpu_incr(cdma, getptr, syncpt_incrs, syncval,
+					nr_slots);
+}
+
+static inline void host1x_hw_pushbuffer_init(struct host1x *host,
+					     struct push_buffer *pb)
+{
+	host->cdma_pb_op->init(pb);
+}
+
+static inline void host1x_hw_debug_init(struct host1x *host, struct dentry *de)
+{
+	if (host->debug_op && host->debug_op->debug_init)
+		host->debug_op->debug_init(de);
+}
+
+static inline void host1x_hw_show_channel_cdma(struct host1x *host,
+					       struct host1x_channel *channel,
+					       struct output *o)
+{
+	host->debug_op->show_channel_cdma(host, channel, o);
+}
+
+static inline void host1x_hw_show_channel_fifo(struct host1x *host,
+					       struct host1x_channel *channel,
+					       struct output *o)
+{
+	host->debug_op->show_channel_fifo(host, channel, o);
+}
+
+static inline void host1x_hw_show_mlocks(struct host1x *host, struct output *o)
+{
+	host->debug_op->show_mlocks(host, o);
+}
+
+extern struct platform_driver tegra_hdmi_driver;
+extern struct platform_driver tegra_dc_driver;
+extern struct platform_driver tegra_gr2d_driver;
+
+#endif
diff --git a/drivers/gpu/drm/tegra/Kconfig b/drivers/gpu/host1x/drm/Kconfig
index be1daf7344d3..69853a4de40a 100644
--- a/drivers/gpu/drm/tegra/Kconfig
+++ b/drivers/gpu/host1x/drm/Kconfig
@@ -1,12 +1,10 @@
 config DRM_TEGRA
-	tristate "NVIDIA Tegra DRM"
-	depends on DRM && OF && ARCH_TEGRA
+	bool "NVIDIA Tegra DRM"
+	depends on DRM
 	select DRM_KMS_HELPER
-	select DRM_GEM_CMA_HELPER
-	select DRM_KMS_CMA_HELPER
-	select FB_CFB_FILLRECT
-	select FB_CFB_COPYAREA
-	select FB_CFB_IMAGEBLIT
+	select FB_SYS_FILLRECT
+	select FB_SYS_COPYAREA
+	select FB_SYS_IMAGEBLIT
 	help
 	  Choose this option if you have an NVIDIA Tegra SoC.
 
@@ -15,6 +13,14 @@ config DRM_TEGRA
 
 if DRM_TEGRA
 
+config DRM_TEGRA_STAGING
+	bool "Enable HOST1X interface"
+	depends on STAGING
+	help
+	  Say yes if HOST1X should be available for userspace DRM users.
+
+	  If unsure, choose N.
+
 config DRM_TEGRA_DEBUG
 	bool "NVIDIA Tegra DRM debug support"
 	help
diff --git a/drivers/gpu/drm/tegra/dc.c b/drivers/gpu/host1x/drm/dc.c
index de94707b9dbe..1e2060324f02 100644
--- a/drivers/gpu/drm/tegra/dc.c
+++ b/drivers/gpu/host1x/drm/dc.c
@@ -14,8 +14,10 @@
 #include <linux/platform_device.h>
 #include <linux/clk/tegra.h>
 
-#include "drm.h"
+#include "host1x_client.h"
 #include "dc.h"
+#include "drm.h"
+#include "gem.h"
 
 struct tegra_plane {
 	struct drm_plane base;
@@ -51,9 +53,9 @@ static int tegra_plane_update(struct drm_plane *plane, struct drm_crtc *crtc,
 	window.bits_per_pixel = fb->bits_per_pixel;
 
 	for (i = 0; i < drm_format_num_planes(fb->pixel_format); i++) {
-		struct drm_gem_cma_object *gem = drm_fb_cma_get_gem_obj(fb, i);
+		struct tegra_bo *bo = tegra_fb_get_plane(fb, i);
 
-		window.base[i] = gem->paddr + fb->offsets[i];
+		window.base[i] = bo->paddr + fb->offsets[i];
 
 		/*
 		 * Tegra doesn't support different strides for U and V planes
@@ -103,7 +105,9 @@ static const struct drm_plane_funcs tegra_plane_funcs = {
 };
 
 static const uint32_t plane_formats[] = {
+	DRM_FORMAT_XBGR8888,
 	DRM_FORMAT_XRGB8888,
+	DRM_FORMAT_RGB565,
 	DRM_FORMAT_UYVY,
 	DRM_FORMAT_YUV420,
 	DRM_FORMAT_YUV422,
@@ -136,7 +140,7 @@ static int tegra_dc_add_planes(struct drm_device *drm, struct tegra_dc *dc)
 static int tegra_dc_set_base(struct tegra_dc *dc, int x, int y,
 			     struct drm_framebuffer *fb)
 {
-	struct drm_gem_cma_object *gem = drm_fb_cma_get_gem_obj(fb, 0);
+	struct tegra_bo *bo = tegra_fb_get_plane(fb, 0);
 	unsigned long value;
 
 	tegra_dc_writel(dc, WINDOW_A_SELECT, DC_CMD_DISPLAY_WINDOW_HEADER);
@@ -144,7 +148,7 @@ static int tegra_dc_set_base(struct tegra_dc *dc, int x, int y,
 	value = fb->offsets[0] + y * fb->pitches[0] +
 		x * fb->bits_per_pixel / 8;
 
-	tegra_dc_writel(dc, gem->paddr + value, DC_WINBUF_START_ADDR);
+	tegra_dc_writel(dc, bo->paddr + value, DC_WINBUF_START_ADDR);
 	tegra_dc_writel(dc, fb->pitches[0], DC_WIN_LINE_STRIDE);
 
 	value = GENERAL_UPDATE | WIN_A_UPDATE;
@@ -186,20 +190,20 @@ static void tegra_dc_finish_page_flip(struct tegra_dc *dc)
 {
 	struct drm_device *drm = dc->base.dev;
 	struct drm_crtc *crtc = &dc->base;
-	struct drm_gem_cma_object *gem;
 	unsigned long flags, base;
+	struct tegra_bo *bo;
 
 	if (!dc->event)
 		return;
 
-	gem = drm_fb_cma_get_gem_obj(crtc->fb, 0);
+	bo = tegra_fb_get_plane(crtc->fb, 0);
 
 	/* check if new start address has been latched */
 	tegra_dc_writel(dc, READ_MUX, DC_CMD_STATE_ACCESS);
 	base = tegra_dc_readl(dc, DC_WINBUF_START_ADDR);
 	tegra_dc_writel(dc, 0, DC_CMD_STATE_ACCESS);
 
-	if (base == gem->paddr + crtc->fb->offsets[0]) {
+	if (base == bo->paddr + crtc->fb->offsets[0]) {
 		spin_lock_irqsave(&drm->event_lock, flags);
 		drm_send_vblank_event(drm, dc->pipe, dc->event);
 		drm_vblank_put(drm, dc->pipe);
@@ -541,6 +545,9 @@ int tegra_dc_setup_window(struct tegra_dc *dc, unsigned int index,
 unsigned int tegra_dc_format(uint32_t format)
 {
 	switch (format) {
+	case DRM_FORMAT_XBGR8888:
+		return WIN_COLOR_DEPTH_R8G8B8A8;
+
 	case DRM_FORMAT_XRGB8888:
 		return WIN_COLOR_DEPTH_B8G8R8A8;
 
@@ -569,7 +576,7 @@ static int tegra_crtc_mode_set(struct drm_crtc *crtc,
 			       struct drm_display_mode *adjusted,
 			       int x, int y, struct drm_framebuffer *old_fb)
 {
-	struct drm_gem_cma_object *gem = drm_fb_cma_get_gem_obj(crtc->fb, 0);
+	struct tegra_bo *bo = tegra_fb_get_plane(crtc->fb, 0);
 	struct tegra_dc *dc = to_tegra_dc(crtc);
 	struct tegra_dc_window window;
 	unsigned long div, value;
@@ -616,7 +623,7 @@ static int tegra_crtc_mode_set(struct drm_crtc *crtc,
 	window.format = tegra_dc_format(crtc->fb->pixel_format);
 	window.bits_per_pixel = crtc->fb->bits_per_pixel;
 	window.stride[0] = crtc->fb->pitches[0];
-	window.base[0] = gem->paddr;
+	window.base[0] = bo->paddr;
 
 	err = tegra_dc_setup_window(dc, 0, &window);
 	if (err < 0)
@@ -1097,7 +1104,7 @@ static const struct host1x_client_ops dc_client_ops = {
 
 static int tegra_dc_probe(struct platform_device *pdev)
 {
-	struct host1x *host1x = dev_get_drvdata(pdev->dev.parent);
+	struct host1x_drm *host1x = host1x_get_drm_data(pdev->dev.parent);
 	struct resource *regs;
 	struct tegra_dc *dc;
 	int err;
@@ -1160,7 +1167,7 @@ static int tegra_dc_probe(struct platform_device *pdev)
 
 static int tegra_dc_remove(struct platform_device *pdev)
 {
-	struct host1x *host1x = dev_get_drvdata(pdev->dev.parent);
+	struct host1x_drm *host1x = host1x_get_drm_data(pdev->dev.parent);
 	struct tegra_dc *dc = platform_get_drvdata(pdev);
 	int err;
 
diff --git a/drivers/gpu/drm/tegra/dc.h b/drivers/gpu/host1x/drm/dc.h
index 79eaec9aac77..79eaec9aac77 100644
--- a/drivers/gpu/drm/tegra/dc.h
+++ b/drivers/gpu/host1x/drm/dc.h
diff --git a/drivers/gpu/host1x/drm/drm.c b/drivers/gpu/host1x/drm/drm.c
new file mode 100644
index 000000000000..2b561c9118c6
--- /dev/null
+++ b/drivers/gpu/host1x/drm/drm.c
@@ -0,0 +1,640 @@
+/*
+ * Copyright (C) 2012 Avionic Design GmbH
+ * Copyright (C) 2012-2013 NVIDIA CORPORATION.  All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+
+#include <linux/module.h>
+#include <linux/of_address.h>
+#include <linux/of_platform.h>
+
+#include <linux/dma-mapping.h>
+#include <asm/dma-iommu.h>
+
+#include <drm/drm.h>
+#include <drm/drmP.h>
+
+#include "host1x_client.h"
+#include "dev.h"
+#include "drm.h"
+#include "gem.h"
+#include "syncpt.h"
+
+#define DRIVER_NAME "tegra"
+#define DRIVER_DESC "NVIDIA Tegra graphics"
+#define DRIVER_DATE "20120330"
+#define DRIVER_MAJOR 0
+#define DRIVER_MINOR 0
+#define DRIVER_PATCHLEVEL 0
+
+struct host1x_drm_client {
+	struct host1x_client *client;
+	struct device_node *np;
+	struct list_head list;
+};
+
+static int host1x_add_drm_client(struct host1x_drm *host1x,
+				 struct device_node *np)
+{
+	struct host1x_drm_client *client;
+
+	client = kzalloc(sizeof(*client), GFP_KERNEL);
+	if (!client)
+		return -ENOMEM;
+
+	INIT_LIST_HEAD(&client->list);
+	client->np = of_node_get(np);
+
+	list_add_tail(&client->list, &host1x->drm_clients);
+
+	return 0;
+}
+
+static int host1x_activate_drm_client(struct host1x_drm *host1x,
+				      struct host1x_drm_client *drm,
+				      struct host1x_client *client)
+{
+	mutex_lock(&host1x->drm_clients_lock);
+	list_del_init(&drm->list);
+	list_add_tail(&drm->list, &host1x->drm_active);
+	drm->client = client;
+	mutex_unlock(&host1x->drm_clients_lock);
+
+	return 0;
+}
+
+static int host1x_remove_drm_client(struct host1x_drm *host1x,
+				    struct host1x_drm_client *client)
+{
+	mutex_lock(&host1x->drm_clients_lock);
+	list_del_init(&client->list);
+	mutex_unlock(&host1x->drm_clients_lock);
+
+	of_node_put(client->np);
+	kfree(client);
+
+	return 0;
+}
+
+static int host1x_parse_dt(struct host1x_drm *host1x)
+{
+	static const char * const compat[] = {
+		"nvidia,tegra20-dc",
+		"nvidia,tegra20-hdmi",
+		"nvidia,tegra20-gr2d",
+		"nvidia,tegra30-dc",
+		"nvidia,tegra30-hdmi",
+		"nvidia,tegra30-gr2d",
+	};
+	unsigned int i;
+	int err;
+
+	for (i = 0; i < ARRAY_SIZE(compat); i++) {
+		struct device_node *np;
+
+		for_each_child_of_node(host1x->dev->of_node, np) {
+			if (of_device_is_compatible(np, compat[i]) &&
+			    of_device_is_available(np)) {
+				err = host1x_add_drm_client(host1x, np);
+				if (err < 0)
+					return err;
+			}
+		}
+	}
+
+	return 0;
+}
+
+int host1x_drm_alloc(struct platform_device *pdev)
+{
+	struct host1x_drm *host1x;
+	int err;
+
+	host1x = devm_kzalloc(&pdev->dev, sizeof(*host1x), GFP_KERNEL);
+	if (!host1x)
+		return -ENOMEM;
+
+	mutex_init(&host1x->drm_clients_lock);
+	INIT_LIST_HEAD(&host1x->drm_clients);
+	INIT_LIST_HEAD(&host1x->drm_active);
+	mutex_init(&host1x->clients_lock);
+	INIT_LIST_HEAD(&host1x->clients);
+	host1x->dev = &pdev->dev;
+
+	err = host1x_parse_dt(host1x);
+	if (err < 0) {
+		dev_err(&pdev->dev, "failed to parse DT: %d\n", err);
+		return err;
+	}
+
+	host1x_set_drm_data(&pdev->dev, host1x);
+
+	return 0;
+}
+
+int host1x_drm_init(struct host1x_drm *host1x, struct drm_device *drm)
+{
+	struct host1x_client *client;
+
+	mutex_lock(&host1x->clients_lock);
+
+	list_for_each_entry(client, &host1x->clients, list) {
+		if (client->ops && client->ops->drm_init) {
+			int err = client->ops->drm_init(client, drm);
+			if (err < 0) {
+				dev_err(host1x->dev,
+					"DRM setup failed for %s: %d\n",
+					dev_name(client->dev), err);
+				return err;
+			}
+		}
+	}
+
+	mutex_unlock(&host1x->clients_lock);
+
+	return 0;
+}
+
+int host1x_drm_exit(struct host1x_drm *host1x)
+{
+	struct platform_device *pdev = to_platform_device(host1x->dev);
+	struct host1x_client *client;
+
+	if (!host1x->drm)
+		return 0;
+
+	mutex_lock(&host1x->clients_lock);
+
+	list_for_each_entry_reverse(client, &host1x->clients, list) {
+		if (client->ops && client->ops->drm_exit) {
+			int err = client->ops->drm_exit(client);
+			if (err < 0) {
+				dev_err(host1x->dev,
+					"DRM cleanup failed for %s: %d\n",
+					dev_name(client->dev), err);
+				return err;
+			}
+		}
+	}
+
+	mutex_unlock(&host1x->clients_lock);
+
+	drm_platform_exit(&tegra_drm_driver, pdev);
+	host1x->drm = NULL;
+
+	return 0;
+}
+
+int host1x_register_client(struct host1x_drm *host1x,
+			   struct host1x_client *client)
+{
+	struct host1x_drm_client *drm, *tmp;
+	int err;
+
+	mutex_lock(&host1x->clients_lock);
+	list_add_tail(&client->list, &host1x->clients);
+	mutex_unlock(&host1x->clients_lock);
+
+	list_for_each_entry_safe(drm, tmp, &host1x->drm_clients, list)
+		if (drm->np == client->dev->of_node)
+			host1x_activate_drm_client(host1x, drm, client);
+
+	if (list_empty(&host1x->drm_clients)) {
+		struct platform_device *pdev = to_platform_device(host1x->dev);
+
+		err = drm_platform_init(&tegra_drm_driver, pdev);
+		if (err < 0) {
+			dev_err(host1x->dev, "drm_platform_init(): %d\n", err);
+			return err;
+		}
+	}
+
+	return 0;
+}
+
+int host1x_unregister_client(struct host1x_drm *host1x,
+			     struct host1x_client *client)
+{
+	struct host1x_drm_client *drm, *tmp;
+	int err;
+
+	list_for_each_entry_safe(drm, tmp, &host1x->drm_active, list) {
+		if (drm->client == client) {
+			err = host1x_drm_exit(host1x);
+			if (err < 0) {
+				dev_err(host1x->dev, "host1x_drm_exit(): %d\n",
+					err);
+				return err;
+			}
+
+			host1x_remove_drm_client(host1x, drm);
+			break;
+		}
+	}
+
+	mutex_lock(&host1x->clients_lock);
+	list_del_init(&client->list);
+	mutex_unlock(&host1x->clients_lock);
+
+	return 0;
+}
+
+static int tegra_drm_load(struct drm_device *drm, unsigned long flags)
+{
+	struct host1x_drm *host1x;
+	int err;
+
+	host1x = host1x_get_drm_data(drm->dev);
+	drm->dev_private = host1x;
+	host1x->drm = drm;
+
+	drm_mode_config_init(drm);
+
+	err = host1x_drm_init(host1x, drm);
+	if (err < 0)
+		return err;
+
+	err = drm_vblank_init(drm, drm->mode_config.num_crtc);
+	if (err < 0)
+		return err;
+
+	err = tegra_drm_fb_init(drm);
+	if (err < 0)
+		return err;
+
+	drm_kms_helper_poll_init(drm);
+
+	return 0;
+}
+
+static int tegra_drm_unload(struct drm_device *drm)
+{
+	drm_kms_helper_poll_fini(drm);
+	tegra_drm_fb_exit(drm);
+
+	drm_mode_config_cleanup(drm);
+
+	return 0;
+}
+
+static int tegra_drm_open(struct drm_device *drm, struct drm_file *filp)
+{
+	struct host1x_drm_file *fpriv;
+
+	fpriv = kzalloc(sizeof(*fpriv), GFP_KERNEL);
+	if (!fpriv)
+		return -ENOMEM;
+
+	INIT_LIST_HEAD(&fpriv->contexts);
+	filp->driver_priv = fpriv;
+
+	return 0;
+}
+
+static void host1x_drm_context_free(struct host1x_drm_context *context)
+{
+	context->client->ops->close_channel(context);
+	kfree(context);
+}
+
+static void tegra_drm_lastclose(struct drm_device *drm)
+{
+	struct host1x_drm *host1x = drm->dev_private;
+
+	tegra_fbdev_restore_mode(host1x->fbdev);
+}
+
+#ifdef CONFIG_DRM_TEGRA_STAGING
+static bool host1x_drm_file_owns_context(struct host1x_drm_file *file,
+					 struct host1x_drm_context *context)
+{
+	struct host1x_drm_context *ctx;
+
+	list_for_each_entry(ctx, &file->contexts, list)
+		if (ctx == context)
+			return true;
+
+	return false;
+}
+
+static int tegra_gem_create(struct drm_device *drm, void *data,
+			    struct drm_file *file)
+{
+	struct drm_tegra_gem_create *args = data;
+	struct tegra_bo *bo;
+
+	bo = tegra_bo_create_with_handle(file, drm, args->size,
+					 &args->handle);
+	if (IS_ERR(bo))
+		return PTR_ERR(bo);
+
+	return 0;
+}
+
+static int tegra_gem_mmap(struct drm_device *drm, void *data,
+			  struct drm_file *file)
+{
+	struct drm_tegra_gem_mmap *args = data;
+	struct drm_gem_object *gem;
+	struct tegra_bo *bo;
+
+	gem = drm_gem_object_lookup(drm, file, args->handle);
+	if (!gem)
+		return -EINVAL;
+
+	bo = to_tegra_bo(gem);
+
+	args->offset = tegra_bo_get_mmap_offset(bo);
+
+	drm_gem_object_unreference(gem);
+
+	return 0;
+}
+
+static int tegra_syncpt_read(struct drm_device *drm, void *data,
+			     struct drm_file *file)
+{
+	struct drm_tegra_syncpt_read *args = data;
+	struct host1x *host = dev_get_drvdata(drm->dev);
+	struct host1x_syncpt *sp = host1x_syncpt_get(host, args->id);
+
+	if (!sp)
+		return -EINVAL;
+
+	args->value = host1x_syncpt_read_min(sp);
+	return 0;
+}
+
+static int tegra_syncpt_incr(struct drm_device *drm, void *data,
+			     struct drm_file *file)
+{
+	struct drm_tegra_syncpt_incr *args = data;
+	struct host1x *host = dev_get_drvdata(drm->dev);
+	struct host1x_syncpt *sp = host1x_syncpt_get(host, args->id);
+
+	if (!sp)
+		return -EINVAL;
+
+	host1x_syncpt_incr(sp);
+	return 0;
+}
+
+static int tegra_syncpt_wait(struct drm_device *drm, void *data,
+			     struct drm_file *file)
+{
+	struct drm_tegra_syncpt_wait *args = data;
+	struct host1x *host = dev_get_drvdata(drm->dev);
+	struct host1x_syncpt *sp = host1x_syncpt_get(host, args->id);
+
+	if (!sp)
+		return -EINVAL;
+
+	return host1x_syncpt_wait(sp, args->thresh, args->timeout,
+				  &args->value);
+}
+
+static int tegra_open_channel(struct drm_device *drm, void *data,
+			      struct drm_file *file)
+{
+	struct drm_tegra_open_channel *args = data;
+	struct host1x_client *client;
+	struct host1x_drm_context *context;
+	struct host1x_drm_file *fpriv = file->driver_priv;
+	struct host1x_drm *host1x = drm->dev_private;
+	int err = -ENODEV;
+
+	context = kzalloc(sizeof(*context), GFP_KERNEL);
+	if (!context)
+		return -ENOMEM;
+
+	list_for_each_entry(client, &host1x->clients, list)
+		if (client->class == args->client) {
+			err = client->ops->open_channel(client, context);
+			if (err)
+				break;
+
+			context->client = client;
+			list_add(&context->list, &fpriv->contexts);
+			args->context = (uintptr_t)context;
+			return 0;
+		}
+
+	kfree(context);
+	return err;
+}
+
+static int tegra_close_channel(struct drm_device *drm, void *data,
+			       struct drm_file *file)
+{
+	struct drm_tegra_close_channel *args = data;
+	struct host1x_drm_file *fpriv = file->driver_priv;
+	struct host1x_drm_context *context =
+		(struct host1x_drm_context *)(uintptr_t)args->context;
+
+	if (!host1x_drm_file_owns_context(fpriv, context))
+		return -EINVAL;
+
+	list_del(&context->list);
+	host1x_drm_context_free(context);
+
+	return 0;
+}
+
+static int tegra_get_syncpt(struct drm_device *drm, void *data,
+			    struct drm_file *file)
+{
+	struct drm_tegra_get_syncpt *args = data;
+	struct host1x_drm_file *fpriv = file->driver_priv;
+	struct host1x_drm_context *context =
+		(struct host1x_drm_context *)(uintptr_t)args->context;
+	struct host1x_syncpt *syncpt;
+
+	if (!host1x_drm_file_owns_context(fpriv, context))
+		return -ENODEV;
+
+	if (args->index >= context->client->num_syncpts)
+		return -EINVAL;
+
+	syncpt = context->client->syncpts[args->index];
+	args->id = host1x_syncpt_id(syncpt);
+
+	return 0;
+}
+
+static int tegra_submit(struct drm_device *drm, void *data,
+			struct drm_file *file)
+{
+	struct drm_tegra_submit *args = data;
+	struct host1x_drm_file *fpriv = file->driver_priv;
+	struct host1x_drm_context *context =
+		(struct host1x_drm_context *)(uintptr_t)args->context;
+
+	if (!host1x_drm_file_owns_context(fpriv, context))
+		return -ENODEV;
+
+	return context->client->ops->submit(context, args, drm, file);
+}
+#endif
+
+static struct drm_ioctl_desc tegra_drm_ioctls[] = {
+#ifdef CONFIG_DRM_TEGRA_STAGING
+	DRM_IOCTL_DEF_DRV(TEGRA_GEM_CREATE, tegra_gem_create, DRM_UNLOCKED | DRM_AUTH),
+	DRM_IOCTL_DEF_DRV(TEGRA_GEM_MMAP, tegra_gem_mmap, DRM_UNLOCKED),
+	DRM_IOCTL_DEF_DRV(TEGRA_SYNCPT_READ, tegra_syncpt_read, DRM_UNLOCKED),
+	DRM_IOCTL_DEF_DRV(TEGRA_SYNCPT_INCR, tegra_syncpt_incr, DRM_UNLOCKED),
+	DRM_IOCTL_DEF_DRV(TEGRA_SYNCPT_WAIT, tegra_syncpt_wait, DRM_UNLOCKED),
+	DRM_IOCTL_DEF_DRV(TEGRA_OPEN_CHANNEL, tegra_open_channel, DRM_UNLOCKED),
+	DRM_IOCTL_DEF_DRV(TEGRA_CLOSE_CHANNEL, tegra_close_channel, DRM_UNLOCKED),
+	DRM_IOCTL_DEF_DRV(TEGRA_GET_SYNCPT, tegra_get_syncpt, DRM_UNLOCKED),
+	DRM_IOCTL_DEF_DRV(TEGRA_SUBMIT, tegra_submit, DRM_UNLOCKED),
+#endif
+};
+
+static const struct file_operations tegra_drm_fops = {
+	.owner = THIS_MODULE,
+	.open = drm_open,
+	.release = drm_release,
+	.unlocked_ioctl = drm_ioctl,
+	.mmap = tegra_drm_mmap,
+	.poll = drm_poll,
+	.fasync = drm_fasync,
+	.read = drm_read,
+#ifdef CONFIG_COMPAT
+	.compat_ioctl = drm_compat_ioctl,
+#endif
+	.llseek = noop_llseek,
+};
+
+static struct drm_crtc *tegra_crtc_from_pipe(struct drm_device *drm, int pipe)
+{
+	struct drm_crtc *crtc;
+
+	list_for_each_entry(crtc, &drm->mode_config.crtc_list, head) {
+		struct tegra_dc *dc = to_tegra_dc(crtc);
+
+		if (dc->pipe == pipe)
+			return crtc;
+	}
+
+	return NULL;
+}
+
+static u32 tegra_drm_get_vblank_counter(struct drm_device *dev, int crtc)
+{
+	/* TODO: implement real hardware counter using syncpoints */
+	return drm_vblank_count(dev, crtc);
+}
+
+static int tegra_drm_enable_vblank(struct drm_device *drm, int pipe)
+{
+	struct drm_crtc *crtc = tegra_crtc_from_pipe(drm, pipe);
+	struct tegra_dc *dc = to_tegra_dc(crtc);
+
+	if (!crtc)
+		return -ENODEV;
+
+	tegra_dc_enable_vblank(dc);
+
+	return 0;
+}
+
+static void tegra_drm_disable_vblank(struct drm_device *drm, int pipe)
+{
+	struct drm_crtc *crtc = tegra_crtc_from_pipe(drm, pipe);
+	struct tegra_dc *dc = to_tegra_dc(crtc);
+
+	if (crtc)
+		tegra_dc_disable_vblank(dc);
+}
+
+static void tegra_drm_preclose(struct drm_device *drm, struct drm_file *file)
+{
+	struct host1x_drm_file *fpriv = file->driver_priv;
+	struct host1x_drm_context *context, *tmp;
+	struct drm_crtc *crtc;
+
+	list_for_each_entry(crtc, &drm->mode_config.crtc_list, head)
+		tegra_dc_cancel_page_flip(crtc, file);
+
+	list_for_each_entry_safe(context, tmp, &fpriv->contexts, list)
+		host1x_drm_context_free(context);
+
+	kfree(fpriv);
+}
+
+#ifdef CONFIG_DEBUG_FS
+static int tegra_debugfs_framebuffers(struct seq_file *s, void *data)
+{
+	struct drm_info_node *node = (struct drm_info_node *)s->private;
+	struct drm_device *drm = node->minor->dev;
+	struct drm_framebuffer *fb;
+
+	mutex_lock(&drm->mode_config.fb_lock);
+
+	list_for_each_entry(fb, &drm->mode_config.fb_list, head) {
+		seq_printf(s, "%3d: user size: %d x %d, depth %d, %d bpp, refcount %d\n",
+			   fb->base.id, fb->width, fb->height, fb->depth,
+			   fb->bits_per_pixel,
+			   atomic_read(&fb->refcount.refcount));
+	}
+
+	mutex_unlock(&drm->mode_config.fb_lock);
+
+	return 0;
+}
+
+static struct drm_info_list tegra_debugfs_list[] = {
+	{ "framebuffers", tegra_debugfs_framebuffers, 0 },
+};
+
+static int tegra_debugfs_init(struct drm_minor *minor)
+{
+	return drm_debugfs_create_files(tegra_debugfs_list,
+					ARRAY_SIZE(tegra_debugfs_list),
+					minor->debugfs_root, minor);
+}
+
+static void tegra_debugfs_cleanup(struct drm_minor *minor)
+{
+	drm_debugfs_remove_files(tegra_debugfs_list,
+				 ARRAY_SIZE(tegra_debugfs_list), minor);
+}
+#endif
+
+struct drm_driver tegra_drm_driver = {
+	.driver_features = DRIVER_BUS_PLATFORM | DRIVER_MODESET | DRIVER_GEM,
+	.load = tegra_drm_load,
+	.unload = tegra_drm_unload,
+	.open = tegra_drm_open,
+	.preclose = tegra_drm_preclose,
+	.lastclose = tegra_drm_lastclose,
+
+	.get_vblank_counter = tegra_drm_get_vblank_counter,
+	.enable_vblank = tegra_drm_enable_vblank,
+	.disable_vblank = tegra_drm_disable_vblank,
+
+#if defined(CONFIG_DEBUG_FS)
+	.debugfs_init = tegra_debugfs_init,
+	.debugfs_cleanup = tegra_debugfs_cleanup,
+#endif
+
+	.gem_free_object = tegra_bo_free_object,
+	.gem_vm_ops = &tegra_bo_vm_ops,
+	.dumb_create = tegra_bo_dumb_create,
+	.dumb_map_offset = tegra_bo_dumb_map_offset,
+	.dumb_destroy = tegra_bo_dumb_destroy,
+
+	.ioctls = tegra_drm_ioctls,
+	.num_ioctls = ARRAY_SIZE(tegra_drm_ioctls),
+	.fops = &tegra_drm_fops,
+
+	.name = DRIVER_NAME,
+	.desc = DRIVER_DESC,
+	.date = DRIVER_DATE,
+	.major = DRIVER_MAJOR,
+	.minor = DRIVER_MINOR,
+	.patchlevel = DRIVER_PATCHLEVEL,
+};
diff --git a/drivers/gpu/drm/tegra/drm.h b/drivers/gpu/host1x/drm/drm.h
index 6dd75a2600eb..02ce020f2575 100644
--- a/drivers/gpu/drm/tegra/drm.h
+++ b/drivers/gpu/host1x/drm/drm.h
@@ -1,24 +1,36 @@
 /*
  * Copyright (C) 2012 Avionic Design GmbH
- * Copyright (C) 2012 NVIDIA CORPORATION.  All rights reserved.
+ * Copyright (C) 2012-2013 NVIDIA CORPORATION.  All rights reserved.
  *
  * This program is free software; you can redistribute it and/or modify
  * it under the terms of the GNU General Public License version 2 as
  * published by the Free Software Foundation.
  */
 
-#ifndef TEGRA_DRM_H
-#define TEGRA_DRM_H 1
+#ifndef HOST1X_DRM_H
+#define HOST1X_DRM_H 1
 
 #include <drm/drmP.h>
 #include <drm/drm_crtc_helper.h>
 #include <drm/drm_edid.h>
 #include <drm/drm_fb_helper.h>
-#include <drm/drm_gem_cma_helper.h>
-#include <drm/drm_fb_cma_helper.h>
 #include <drm/drm_fixed.h>
+#include <uapi/drm/tegra_drm.h>
 
-struct host1x {
+#include "host1x.h"
+
+struct tegra_fb {
+	struct drm_framebuffer base;
+	struct tegra_bo **planes;
+	unsigned int num_planes;
+};
+
+struct tegra_fbdev {
+	struct drm_fb_helper base;
+	struct tegra_fb *fb;
+};
+
+struct host1x_drm {
 	struct drm_device *drm;
 	struct device *dev;
 	void __iomem *regs;
@@ -33,31 +45,53 @@ struct host1x {
 	struct mutex clients_lock;
 	struct list_head clients;
 
-	struct drm_fbdev_cma *fbdev;
+	struct tegra_fbdev *fbdev;
 };
 
 struct host1x_client;
 
+struct host1x_drm_context {
+	struct host1x_client *client;
+	struct host1x_channel *channel;
+	struct list_head list;
+};
+
 struct host1x_client_ops {
 	int (*drm_init)(struct host1x_client *client, struct drm_device *drm);
 	int (*drm_exit)(struct host1x_client *client);
+	int (*open_channel)(struct host1x_client *client,
+			    struct host1x_drm_context *context);
+	void (*close_channel)(struct host1x_drm_context *context);
+	int (*submit)(struct host1x_drm_context *context,
+		      struct drm_tegra_submit *args, struct drm_device *drm,
+		      struct drm_file *file);
+};
+
+struct host1x_drm_file {
+	struct list_head contexts;
 };
 
 struct host1x_client {
-	struct host1x *host1x;
+	struct host1x_drm *host1x;
 	struct device *dev;
 
 	const struct host1x_client_ops *ops;
 
+	enum host1x_class class;
+	struct host1x_channel *channel;
+
+	struct host1x_syncpt **syncpts;
+	unsigned int num_syncpts;
+
 	struct list_head list;
 };
 
-extern int host1x_drm_init(struct host1x *host1x, struct drm_device *drm);
-extern int host1x_drm_exit(struct host1x *host1x);
+extern int host1x_drm_init(struct host1x_drm *host1x, struct drm_device *drm);
+extern int host1x_drm_exit(struct host1x_drm *host1x);
 
-extern int host1x_register_client(struct host1x *host1x,
+extern int host1x_register_client(struct host1x_drm *host1x,
 				  struct host1x_client *client);
-extern int host1x_unregister_client(struct host1x *host1x,
+extern int host1x_unregister_client(struct host1x_drm *host1x,
 				    struct host1x_client *client);
 
 struct tegra_output;
@@ -66,7 +100,7 @@ struct tegra_dc {
 	struct host1x_client client;
 	spinlock_t lock;
 
-	struct host1x *host1x;
+	struct host1x_drm *host1x;
 	struct device *dev;
 
 	struct drm_crtc base;
@@ -226,12 +260,12 @@ extern int tegra_output_init(struct drm_device *drm, struct tegra_output *output
 extern int tegra_output_exit(struct tegra_output *output);
 
 /* from fb.c */
+struct tegra_bo *tegra_fb_get_plane(struct drm_framebuffer *framebuffer,
+				    unsigned int index);
 extern int tegra_drm_fb_init(struct drm_device *drm);
 extern void tegra_drm_fb_exit(struct drm_device *drm);
+extern void tegra_fbdev_restore_mode(struct tegra_fbdev *fbdev);
 
-extern struct platform_driver tegra_host1x_driver;
-extern struct platform_driver tegra_hdmi_driver;
-extern struct platform_driver tegra_dc_driver;
 extern struct drm_driver tegra_drm_driver;
 
-#endif /* TEGRA_DRM_H */
+#endif /* HOST1X_DRM_H */
diff --git a/drivers/gpu/host1x/drm/fb.c b/drivers/gpu/host1x/drm/fb.c
new file mode 100644
index 000000000000..979a3e32b78b
--- /dev/null
+++ b/drivers/gpu/host1x/drm/fb.c
@@ -0,0 +1,374 @@
+/*
+ * Copyright (C) 2012-2013 Avionic Design GmbH
+ * Copyright (C) 2012 NVIDIA CORPORATION.  All rights reserved.
+ *
+ * Based on the KMS/FB CMA helpers
+ *   Copyright (C) 2012 Analog Device Inc.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+
+#include <linux/module.h>
+
+#include "drm.h"
+#include "gem.h"
+
+static inline struct tegra_fb *to_tegra_fb(struct drm_framebuffer *fb)
+{
+	return container_of(fb, struct tegra_fb, base);
+}
+
+static inline struct tegra_fbdev *to_tegra_fbdev(struct drm_fb_helper *helper)
+{
+	return container_of(helper, struct tegra_fbdev, base);
+}
+
+struct tegra_bo *tegra_fb_get_plane(struct drm_framebuffer *framebuffer,
+				    unsigned int index)
+{
+	struct tegra_fb *fb = to_tegra_fb(framebuffer);
+
+	if (index >= drm_format_num_planes(framebuffer->pixel_format))
+		return NULL;
+
+	return fb->planes[index];
+}
+
+static void tegra_fb_destroy(struct drm_framebuffer *framebuffer)
+{
+	struct tegra_fb *fb = to_tegra_fb(framebuffer);
+	unsigned int i;
+
+	for (i = 0; i < fb->num_planes; i++) {
+		struct tegra_bo *bo = fb->planes[i];
+
+		if (bo)
+			drm_gem_object_unreference_unlocked(&bo->gem);
+	}
+
+	drm_framebuffer_cleanup(framebuffer);
+	kfree(fb->planes);
+	kfree(fb);
+}
+
+static int tegra_fb_create_handle(struct drm_framebuffer *framebuffer,
+				  struct drm_file *file, unsigned int *handle)
+{
+	struct tegra_fb *fb = to_tegra_fb(framebuffer);
+
+	return drm_gem_handle_create(file, &fb->planes[0]->gem, handle);
+}
+
+static struct drm_framebuffer_funcs tegra_fb_funcs = {
+	.destroy = tegra_fb_destroy,
+	.create_handle = tegra_fb_create_handle,
+};
+
+static struct tegra_fb *tegra_fb_alloc(struct drm_device *drm,
+				       struct drm_mode_fb_cmd2 *mode_cmd,
+				       struct tegra_bo **planes,
+				       unsigned int num_planes)
+{
+	struct tegra_fb *fb;
+	unsigned int i;
+	int err;
+
+	fb = kzalloc(sizeof(*fb), GFP_KERNEL);
+	if (!fb)
+		return ERR_PTR(-ENOMEM);
+
+	fb->planes = kzalloc(num_planes * sizeof(*planes), GFP_KERNEL);
+	if (!fb->planes)
+		return ERR_PTR(-ENOMEM);
+
+	fb->num_planes = num_planes;
+
+	drm_helper_mode_fill_fb_struct(&fb->base, mode_cmd);
+
+	for (i = 0; i < fb->num_planes; i++)
+		fb->planes[i] = planes[i];
+
+	err = drm_framebuffer_init(drm, &fb->base, &tegra_fb_funcs);
+	if (err < 0) {
+		dev_err(drm->dev, "failed to initialize framebuffer: %d\n",
+			err);
+		kfree(fb->planes);
+		kfree(fb);
+		return ERR_PTR(err);
+	}
+
+	return fb;
+}
+
+static struct drm_framebuffer *tegra_fb_create(struct drm_device *drm,
+					       struct drm_file *file,
+					       struct drm_mode_fb_cmd2 *cmd)
+{
+	unsigned int hsub, vsub, i;
+	struct tegra_bo *planes[4];
+	struct drm_gem_object *gem;
+	struct tegra_fb *fb;
+	int err;
+
+	hsub = drm_format_horz_chroma_subsampling(cmd->pixel_format);
+	vsub = drm_format_vert_chroma_subsampling(cmd->pixel_format);
+
+	for (i = 0; i < drm_format_num_planes(cmd->pixel_format); i++) {
+		unsigned int width = cmd->width / (i ? hsub : 1);
+		unsigned int height = cmd->height / (i ? vsub : 1);
+		unsigned int size, bpp;
+
+		gem = drm_gem_object_lookup(drm, file, cmd->handles[i]);
+		if (!gem) {
+			err = -ENXIO;
+			goto unreference;
+		}
+
+		bpp = drm_format_plane_cpp(cmd->pixel_format, i);
+
+		size = (height - 1) * cmd->pitches[i] +
+		       width * bpp + cmd->offsets[i];
+
+		if (gem->size < size) {
+			err = -EINVAL;
+			goto unreference;
+		}
+
+		planes[i] = to_tegra_bo(gem);
+	}
+
+	fb = tegra_fb_alloc(drm, cmd, planes, i);
+	if (IS_ERR(fb)) {
+		err = PTR_ERR(fb);
+		goto unreference;
+	}
+
+	return &fb->base;
+
+unreference:
+	while (i--)
+		drm_gem_object_unreference_unlocked(&planes[i]->gem);
+
+	return ERR_PTR(err);
+}
+
+static struct fb_ops tegra_fb_ops = {
+	.owner = THIS_MODULE,
+	.fb_fillrect = sys_fillrect,
+	.fb_copyarea = sys_copyarea,
+	.fb_imageblit = sys_imageblit,
+	.fb_check_var = drm_fb_helper_check_var,
+	.fb_set_par = drm_fb_helper_set_par,
+	.fb_blank = drm_fb_helper_blank,
+	.fb_pan_display = drm_fb_helper_pan_display,
+	.fb_setcmap = drm_fb_helper_setcmap,
+};
+
+static int tegra_fbdev_probe(struct drm_fb_helper *helper,
+			     struct drm_fb_helper_surface_size *sizes)
+{
+	struct tegra_fbdev *fbdev = to_tegra_fbdev(helper);
+	struct drm_device *drm = helper->dev;
+	struct drm_mode_fb_cmd2 cmd = { 0 };
+	unsigned int bytes_per_pixel;
+	struct drm_framebuffer *fb;
+	unsigned long offset;
+	struct fb_info *info;
+	struct tegra_bo *bo;
+	size_t size;
+	int err;
+
+	bytes_per_pixel = DIV_ROUND_UP(sizes->surface_bpp, 8);
+
+	cmd.width = sizes->surface_width;
+	cmd.height = sizes->surface_height;
+	cmd.pitches[0] = sizes->surface_width * bytes_per_pixel;
+	cmd.pixel_format = drm_mode_legacy_fb_format(sizes->surface_bpp,
+						     sizes->surface_depth);
+
+	size = cmd.pitches[0] * cmd.height;
+
+	bo = tegra_bo_create(drm, size);
+	if (IS_ERR(bo))
+		return PTR_ERR(bo);
+
+	info = framebuffer_alloc(0, drm->dev);
+	if (!info) {
+		dev_err(drm->dev, "failed to allocate framebuffer info\n");
+		tegra_bo_free_object(&bo->gem);
+		return -ENOMEM;
+	}
+
+	fbdev->fb = tegra_fb_alloc(drm, &cmd, &bo, 1);
+	if (IS_ERR(fbdev->fb)) {
+		dev_err(drm->dev, "failed to allocate DRM framebuffer\n");
+		err = PTR_ERR(fbdev->fb);
+		goto release;
+	}
+
+	fb = &fbdev->fb->base;
+	helper->fb = fb;
+	helper->fbdev = info;
+
+	info->par = helper;
+	info->flags = FBINFO_FLAG_DEFAULT;
+	info->fbops = &tegra_fb_ops;
+
+	err = fb_alloc_cmap(&info->cmap, 256, 0);
+	if (err < 0) {
+		dev_err(drm->dev, "failed to allocate color map: %d\n", err);
+		goto destroy;
+	}
+
+	drm_fb_helper_fill_fix(info, fb->pitches[0], fb->depth);
+	drm_fb_helper_fill_var(info, helper, fb->width, fb->height);
+
+	offset = info->var.xoffset * bytes_per_pixel +
+		 info->var.yoffset * fb->pitches[0];
+
+	drm->mode_config.fb_base = (resource_size_t)bo->paddr;
+	info->screen_base = bo->vaddr + offset;
+	info->screen_size = size;
+	info->fix.smem_start = (unsigned long)(bo->paddr + offset);
+	info->fix.smem_len = size;
+
+	return 0;
+
+destroy:
+	drm_framebuffer_unregister_private(fb);
+	tegra_fb_destroy(fb);
+release:
+	framebuffer_release(info);
+	return err;
+}
+
+static struct drm_fb_helper_funcs tegra_fb_helper_funcs = {
+	.fb_probe = tegra_fbdev_probe,
+};
+
+static struct tegra_fbdev *tegra_fbdev_create(struct drm_device *drm,
+					      unsigned int preferred_bpp,
+					      unsigned int num_crtc,
+					      unsigned int max_connectors)
+{
+	struct drm_fb_helper *helper;
+	struct tegra_fbdev *fbdev;
+	int err;
+
+	fbdev = kzalloc(sizeof(*fbdev), GFP_KERNEL);
+	if (!fbdev) {
+		dev_err(drm->dev, "failed to allocate DRM fbdev\n");
+		return ERR_PTR(-ENOMEM);
+	}
+
+	fbdev->base.funcs = &tegra_fb_helper_funcs;
+	helper = &fbdev->base;
+
+	err = drm_fb_helper_init(drm, &fbdev->base, num_crtc, max_connectors);
+	if (err < 0) {
+		dev_err(drm->dev, "failed to initialize DRM FB helper\n");
+		goto free;
+	}
+
+	err = drm_fb_helper_single_add_all_connectors(&fbdev->base);
+	if (err < 0) {
+		dev_err(drm->dev, "failed to add connectors\n");
+		goto fini;
+	}
+
+	drm_helper_disable_unused_functions(drm);
+
+	err = drm_fb_helper_initial_config(&fbdev->base, preferred_bpp);
+	if (err < 0) {
+		dev_err(drm->dev, "failed to set initial configuration\n");
+		goto fini;
+	}
+
+	return fbdev;
+
+fini:
+	drm_fb_helper_fini(&fbdev->base);
+free:
+	kfree(fbdev);
+	return ERR_PTR(err);
+}
+
+static void tegra_fbdev_free(struct tegra_fbdev *fbdev)
+{
+	struct fb_info *info = fbdev->base.fbdev;
+
+	if (info) {
+		int err;
+
+		err = unregister_framebuffer(info);
+		if (err < 0)
+			DRM_DEBUG_KMS("failed to unregister framebuffer\n");
+
+		if (info->cmap.len)
+			fb_dealloc_cmap(&info->cmap);
+
+		framebuffer_release(info);
+	}
+
+	if (fbdev->fb) {
+		drm_framebuffer_unregister_private(&fbdev->fb->base);
+		tegra_fb_destroy(&fbdev->fb->base);
+	}
+
+	drm_fb_helper_fini(&fbdev->base);
+	kfree(fbdev);
+}
+
+static void tegra_fb_output_poll_changed(struct drm_device *drm)
+{
+	struct host1x_drm *host1x = drm->dev_private;
+
+	if (host1x->fbdev)
+		drm_fb_helper_hotplug_event(&host1x->fbdev->base);
+}
+
+static const struct drm_mode_config_funcs tegra_drm_mode_funcs = {
+	.fb_create = tegra_fb_create,
+	.output_poll_changed = tegra_fb_output_poll_changed,
+};
+
+int tegra_drm_fb_init(struct drm_device *drm)
+{
+	struct host1x_drm *host1x = drm->dev_private;
+	struct tegra_fbdev *fbdev;
+
+	drm->mode_config.min_width = 0;
+	drm->mode_config.min_height = 0;
+
+	drm->mode_config.max_width = 4096;
+	drm->mode_config.max_height = 4096;
+
+	drm->mode_config.funcs = &tegra_drm_mode_funcs;
+
+	fbdev = tegra_fbdev_create(drm, 32, drm->mode_config.num_crtc,
+				   drm->mode_config.num_connector);
+	if (IS_ERR(fbdev))
+		return PTR_ERR(fbdev);
+
+	host1x->fbdev = fbdev;
+
+	return 0;
+}
+
+void tegra_drm_fb_exit(struct drm_device *drm)
+{
+	struct host1x_drm *host1x = drm->dev_private;
+
+	tegra_fbdev_free(host1x->fbdev);
+}
+
+void tegra_fbdev_restore_mode(struct tegra_fbdev *fbdev)
+{
+	if (fbdev) {
+		drm_modeset_lock_all(fbdev->base.dev);
+		drm_fb_helper_restore_fbdev_mode(&fbdev->base);
+		drm_modeset_unlock_all(fbdev->base.dev);
+	}
+}
diff --git a/drivers/gpu/host1x/drm/gem.c b/drivers/gpu/host1x/drm/gem.c
new file mode 100644
index 000000000000..c5e9a9b494c2
--- /dev/null
+++ b/drivers/gpu/host1x/drm/gem.c
@@ -0,0 +1,270 @@
+/*
+ * NVIDIA Tegra DRM GEM helper functions
+ *
+ * Copyright (C) 2012 Sascha Hauer, Pengutronix
+ * Copyright (C) 2013 NVIDIA CORPORATION, All rights reserved.
+ *
+ * Based on the GEM/CMA helpers
+ *
+ * Copyright (c) 2011 Samsung Electronics Co., Ltd.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License
+ * as published by the Free Software Foundation; either version 2
+ * of the License, or (at your option) any later version.
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ */
+
+#include <linux/mm.h>
+#include <linux/slab.h>
+#include <linux/mutex.h>
+#include <linux/export.h>
+#include <linux/dma-mapping.h>
+
+#include <drm/drmP.h>
+#include <drm/drm.h>
+
+#include "gem.h"
+
+static inline struct tegra_bo *host1x_to_drm_bo(struct host1x_bo *bo)
+{
+	return container_of(bo, struct tegra_bo, base);
+}
+
+static void tegra_bo_put(struct host1x_bo *bo)
+{
+	struct tegra_bo *obj = host1x_to_drm_bo(bo);
+	struct drm_device *drm = obj->gem.dev;
+
+	mutex_lock(&drm->struct_mutex);
+	drm_gem_object_unreference(&obj->gem);
+	mutex_unlock(&drm->struct_mutex);
+}
+
+static dma_addr_t tegra_bo_pin(struct host1x_bo *bo, struct sg_table **sgt)
+{
+	struct tegra_bo *obj = host1x_to_drm_bo(bo);
+
+	return obj->paddr;
+}
+
+static void tegra_bo_unpin(struct host1x_bo *bo, struct sg_table *sgt)
+{
+}
+
+static void *tegra_bo_mmap(struct host1x_bo *bo)
+{
+	struct tegra_bo *obj = host1x_to_drm_bo(bo);
+
+	return obj->vaddr;
+}
+
+static void tegra_bo_munmap(struct host1x_bo *bo, void *addr)
+{
+}
+
+static void *tegra_bo_kmap(struct host1x_bo *bo, unsigned int page)
+{
+	struct tegra_bo *obj = host1x_to_drm_bo(bo);
+
+	return obj->vaddr + page * PAGE_SIZE;
+}
+
+static void tegra_bo_kunmap(struct host1x_bo *bo, unsigned int page,
+			    void *addr)
+{
+}
+
+static struct host1x_bo *tegra_bo_get(struct host1x_bo *bo)
+{
+	struct tegra_bo *obj = host1x_to_drm_bo(bo);
+	struct drm_device *drm = obj->gem.dev;
+
+	mutex_lock(&drm->struct_mutex);
+	drm_gem_object_reference(&obj->gem);
+	mutex_unlock(&drm->struct_mutex);
+
+	return bo;
+}
+
+const struct host1x_bo_ops tegra_bo_ops = {
+	.get = tegra_bo_get,
+	.put = tegra_bo_put,
+	.pin = tegra_bo_pin,
+	.unpin = tegra_bo_unpin,
+	.mmap = tegra_bo_mmap,
+	.munmap = tegra_bo_munmap,
+	.kmap = tegra_bo_kmap,
+	.kunmap = tegra_bo_kunmap,
+};
+
+static void tegra_bo_destroy(struct drm_device *drm, struct tegra_bo *bo)
+{
+	dma_free_writecombine(drm->dev, bo->gem.size, bo->vaddr, bo->paddr);
+}
+
+unsigned int tegra_bo_get_mmap_offset(struct tegra_bo *bo)
+{
+	return (unsigned int)bo->gem.map_list.hash.key << PAGE_SHIFT;
+}
+
+struct tegra_bo *tegra_bo_create(struct drm_device *drm, unsigned int size)
+{
+	struct tegra_bo *bo;
+	int err;
+
+	bo = kzalloc(sizeof(*bo), GFP_KERNEL);
+	if (!bo)
+		return ERR_PTR(-ENOMEM);
+
+	host1x_bo_init(&bo->base, &tegra_bo_ops);
+	size = round_up(size, PAGE_SIZE);
+
+	bo->vaddr = dma_alloc_writecombine(drm->dev, size, &bo->paddr,
+					   GFP_KERNEL | __GFP_NOWARN);
+	if (!bo->vaddr) {
+		dev_err(drm->dev, "failed to allocate buffer with size %u\n",
+			size);
+		err = -ENOMEM;
+		goto err_dma;
+	}
+
+	err = drm_gem_object_init(drm, &bo->gem, size);
+	if (err)
+		goto err_init;
+
+	err = drm_gem_create_mmap_offset(&bo->gem);
+	if (err)
+		goto err_mmap;
+
+	return bo;
+
+err_mmap:
+	drm_gem_object_release(&bo->gem);
+err_init:
+	tegra_bo_destroy(drm, bo);
+err_dma:
+	kfree(bo);
+
+	return ERR_PTR(err);
+
+}
+
+struct tegra_bo *tegra_bo_create_with_handle(struct drm_file *file,
+					    struct drm_device *drm,
+					    unsigned int size,
+					    unsigned int *handle)
+{
+	struct tegra_bo *bo;
+	int ret;
+
+	bo = tegra_bo_create(drm, size);
+	if (IS_ERR(bo))
+		return bo;
+
+	ret = drm_gem_handle_create(file, &bo->gem, handle);
+	if (ret)
+		goto err;
+
+	drm_gem_object_unreference_unlocked(&bo->gem);
+
+	return bo;
+
+err:
+	tegra_bo_free_object(&bo->gem);
+	return ERR_PTR(ret);
+}
+
+void tegra_bo_free_object(struct drm_gem_object *gem)
+{
+	struct tegra_bo *bo = to_tegra_bo(gem);
+
+	if (gem->map_list.map)
+		drm_gem_free_mmap_offset(gem);
+
+	drm_gem_object_release(gem);
+	tegra_bo_destroy(gem->dev, bo);
+
+	kfree(bo);
+}
+
+int tegra_bo_dumb_create(struct drm_file *file, struct drm_device *drm,
+			 struct drm_mode_create_dumb *args)
+{
+	int min_pitch = DIV_ROUND_UP(args->width * args->bpp, 8);
+	struct tegra_bo *bo;
+
+	if (args->pitch < min_pitch)
+		args->pitch = min_pitch;
+
+	if (args->size < args->pitch * args->height)
+		args->size = args->pitch * args->height;
+
+	bo = tegra_bo_create_with_handle(file, drm, args->size,
+					    &args->handle);
+	if (IS_ERR(bo))
+		return PTR_ERR(bo);
+
+	return 0;
+}
+
+int tegra_bo_dumb_map_offset(struct drm_file *file, struct drm_device *drm,
+			     uint32_t handle, uint64_t *offset)
+{
+	struct drm_gem_object *gem;
+	struct tegra_bo *bo;
+
+	mutex_lock(&drm->struct_mutex);
+
+	gem = drm_gem_object_lookup(drm, file, handle);
+	if (!gem) {
+		dev_err(drm->dev, "failed to lookup GEM object\n");
+		mutex_unlock(&drm->struct_mutex);
+		return -EINVAL;
+	}
+
+	bo = to_tegra_bo(gem);
+
+	*offset = tegra_bo_get_mmap_offset(bo);
+
+	drm_gem_object_unreference(gem);
+
+	mutex_unlock(&drm->struct_mutex);
+
+	return 0;
+}
+
+const struct vm_operations_struct tegra_bo_vm_ops = {
+	.open = drm_gem_vm_open,
+	.close = drm_gem_vm_close,
+};
+
+int tegra_drm_mmap(struct file *file, struct vm_area_struct *vma)
+{
+	struct drm_gem_object *gem;
+	struct tegra_bo *bo;
+	int ret;
+
+	ret = drm_gem_mmap(file, vma);
+	if (ret)
+		return ret;
+
+	gem = vma->vm_private_data;
+	bo = to_tegra_bo(gem);
+
+	ret = remap_pfn_range(vma, vma->vm_start, bo->paddr >> PAGE_SHIFT,
+			      vma->vm_end - vma->vm_start, vma->vm_page_prot);
+	if (ret)
+		drm_gem_vm_close(vma);
+
+	return ret;
+}
+
+int tegra_bo_dumb_destroy(struct drm_file *file, struct drm_device *drm,
+			  unsigned int handle)
+{
+	return drm_gem_handle_delete(file, handle);
+}
diff --git a/drivers/gpu/host1x/drm/gem.h b/drivers/gpu/host1x/drm/gem.h
new file mode 100644
index 000000000000..34de2b486eb7
--- /dev/null
+++ b/drivers/gpu/host1x/drm/gem.h
@@ -0,0 +1,59 @@
+/*
+ * Tegra host1x GEM implementation
+ *
+ * Copyright (c) 2012-2013, NVIDIA Corporation.
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms and conditions of the GNU General Public License,
+ * version 2, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program.  If not, see <http://www.gnu.org/licenses/>.
+ */
+
+#ifndef __HOST1X_GEM_H
+#define __HOST1X_GEM_H
+
+#include <drm/drm.h>
+#include <drm/drmP.h>
+
+#include "host1x_bo.h"
+
+struct tegra_bo {
+	struct drm_gem_object gem;
+	struct host1x_bo base;
+	dma_addr_t paddr;
+	void *vaddr;
+};
+
+static inline struct tegra_bo *to_tegra_bo(struct drm_gem_object *gem)
+{
+	return container_of(gem, struct tegra_bo, gem);
+}
+
+extern const struct host1x_bo_ops tegra_bo_ops;
+
+struct tegra_bo *tegra_bo_create(struct drm_device *drm, unsigned int size);
+struct tegra_bo *tegra_bo_create_with_handle(struct drm_file *file,
+					    struct drm_device *drm,
+					    unsigned int size,
+					    unsigned int *handle);
+void tegra_bo_free_object(struct drm_gem_object *gem);
+unsigned int tegra_bo_get_mmap_offset(struct tegra_bo *bo);
+int tegra_bo_dumb_create(struct drm_file *file, struct drm_device *drm,
+			 struct drm_mode_create_dumb *args);
+int tegra_bo_dumb_map_offset(struct drm_file *file, struct drm_device *drm,
+			     uint32_t handle, uint64_t *offset);
+int tegra_bo_dumb_destroy(struct drm_file *file, struct drm_device *drm,
+			  unsigned int handle);
+
+int tegra_drm_mmap(struct file *file, struct vm_area_struct *vma);
+
+extern const struct vm_operations_struct tegra_bo_vm_ops;
+
+#endif
diff --git a/drivers/gpu/host1x/drm/gr2d.c b/drivers/gpu/host1x/drm/gr2d.c
new file mode 100644
index 000000000000..6a45ae090ee7
--- /dev/null
+++ b/drivers/gpu/host1x/drm/gr2d.c
@@ -0,0 +1,339 @@
+/*
+ * drivers/video/tegra/host/gr2d/gr2d.c
+ *
+ * Tegra Graphics 2D
+ *
+ * Copyright (c) 2012-2013, NVIDIA Corporation.
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms and conditions of the GNU General Public License,
+ * version 2, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program.  If not, see <http://www.gnu.org/licenses/>.
+ */
+
+#include <linux/export.h>
+#include <linux/of.h>
+#include <linux/of_device.h>
+#include <linux/clk.h>
+
+#include "channel.h"
+#include "drm.h"
+#include "gem.h"
+#include "job.h"
+#include "host1x.h"
+#include "host1x_bo.h"
+#include "host1x_client.h"
+#include "syncpt.h"
+
+struct gr2d {
+	struct host1x_client client;
+	struct clk *clk;
+	struct host1x_channel *channel;
+	unsigned long *addr_regs;
+};
+
+static inline struct gr2d *to_gr2d(struct host1x_client *client)
+{
+	return container_of(client, struct gr2d, client);
+}
+
+static int gr2d_is_addr_reg(struct device *dev, u32 class, u32 reg);
+
+static int gr2d_client_init(struct host1x_client *client,
+			    struct drm_device *drm)
+{
+	return 0;
+}
+
+static int gr2d_client_exit(struct host1x_client *client)
+{
+	return 0;
+}
+
+static int gr2d_open_channel(struct host1x_client *client,
+			     struct host1x_drm_context *context)
+{
+	struct gr2d *gr2d = to_gr2d(client);
+
+	context->channel = host1x_channel_get(gr2d->channel);
+
+	if (!context->channel)
+		return -ENOMEM;
+
+	return 0;
+}
+
+static void gr2d_close_channel(struct host1x_drm_context *context)
+{
+	host1x_channel_put(context->channel);
+}
+
+static struct host1x_bo *host1x_bo_lookup(struct drm_device *drm,
+					  struct drm_file *file,
+					  u32 handle)
+{
+	struct drm_gem_object *gem;
+	struct tegra_bo *bo;
+
+	gem = drm_gem_object_lookup(drm, file, handle);
+	if (!gem)
+		return 0;
+
+	mutex_lock(&drm->struct_mutex);
+	drm_gem_object_unreference(gem);
+	mutex_unlock(&drm->struct_mutex);
+
+	bo = to_tegra_bo(gem);
+	return &bo->base;
+}
+
+static int gr2d_submit(struct host1x_drm_context *context,
+		       struct drm_tegra_submit *args, struct drm_device *drm,
+		       struct drm_file *file)
+{
+	struct host1x_job *job;
+	unsigned int num_cmdbufs = args->num_cmdbufs;
+	unsigned int num_relocs = args->num_relocs;
+	unsigned int num_waitchks = args->num_waitchks;
+	struct drm_tegra_cmdbuf __user *cmdbufs =
+		(void * __user)(uintptr_t)args->cmdbufs;
+	struct drm_tegra_reloc __user *relocs =
+		(void * __user)(uintptr_t)args->relocs;
+	struct drm_tegra_waitchk __user *waitchks =
+		(void * __user)(uintptr_t)args->waitchks;
+	struct drm_tegra_syncpt syncpt;
+	int err;
+
+	/* We don't yet support other than one syncpt_incr struct per submit */
+	if (args->num_syncpts != 1)
+		return -EINVAL;
+
+	job = host1x_job_alloc(context->channel, args->num_cmdbufs,
+			       args->num_relocs, args->num_waitchks);
+	if (!job)
+		return -ENOMEM;
+
+	job->num_relocs = args->num_relocs;
+	job->num_waitchk = args->num_waitchks;
+	job->client = (u32)args->context;
+	job->class = context->client->class;
+	job->serialize = true;
+
+	while (num_cmdbufs) {
+		struct drm_tegra_cmdbuf cmdbuf;
+		struct host1x_bo *bo;
+
+		err = copy_from_user(&cmdbuf, cmdbufs, sizeof(cmdbuf));
+		if (err)
+			goto fail;
+
+		bo = host1x_bo_lookup(drm, file, cmdbuf.handle);
+		if (!bo)
+			goto fail;
+
+		host1x_job_add_gather(job, bo, cmdbuf.words, cmdbuf.offset);
+		num_cmdbufs--;
+		cmdbufs++;
+	}
+
+	err = copy_from_user(job->relocarray, relocs,
+			     sizeof(*relocs) * num_relocs);
+	if (err)
+		goto fail;
+
+	while (num_relocs--) {
+		struct host1x_reloc *reloc = &job->relocarray[num_relocs];
+		struct host1x_bo *cmdbuf, *target;
+
+		cmdbuf = host1x_bo_lookup(drm, file, (u32)reloc->cmdbuf);
+		target = host1x_bo_lookup(drm, file, (u32)reloc->target);
+
+		reloc->cmdbuf = cmdbuf;
+		reloc->target = target;
+
+		if (!reloc->target || !reloc->cmdbuf)
+			goto fail;
+	}
+
+	err = copy_from_user(job->waitchk, waitchks,
+			     sizeof(*waitchks) * num_waitchks);
+	if (err)
+		goto fail;
+
+	err = copy_from_user(&syncpt, (void * __user)(uintptr_t)args->syncpts,
+			     sizeof(syncpt));
+	if (err)
+		goto fail;
+
+	job->syncpt_id = syncpt.id;
+	job->syncpt_incrs = syncpt.incrs;
+	job->timeout = 10000;
+	job->is_addr_reg = gr2d_is_addr_reg;
+
+	if (args->timeout && args->timeout < 10000)
+		job->timeout = args->timeout;
+
+	err = host1x_job_pin(job, context->client->dev);
+	if (err)
+		goto fail;
+
+	err = host1x_job_submit(job);
+	if (err)
+		goto fail_submit;
+
+	args->fence = job->syncpt_end;
+
+	host1x_job_put(job);
+	return 0;
+
+fail_submit:
+	host1x_job_unpin(job);
+fail:
+	host1x_job_put(job);
+	return err;
+}
+
+static struct host1x_client_ops gr2d_client_ops = {
+	.drm_init = gr2d_client_init,
+	.drm_exit = gr2d_client_exit,
+	.open_channel = gr2d_open_channel,
+	.close_channel = gr2d_close_channel,
+	.submit = gr2d_submit,
+};
+
+static void gr2d_init_addr_reg_map(struct device *dev, struct gr2d *gr2d)
+{
+	const u32 gr2d_addr_regs[] = {0x1a, 0x1b, 0x26, 0x2b, 0x2c, 0x2d, 0x31,
+				      0x32, 0x48, 0x49, 0x4a, 0x4b, 0x4c};
+	unsigned long *bitmap;
+	int i;
+
+	bitmap = devm_kzalloc(dev, DIV_ROUND_UP(256, BITS_PER_BYTE),
+			      GFP_KERNEL);
+
+	for (i = 0; i < ARRAY_SIZE(gr2d_addr_regs); ++i) {
+		u32 reg = gr2d_addr_regs[i];
+		bitmap[BIT_WORD(reg)] |= BIT_MASK(reg);
+	}
+
+	gr2d->addr_regs = bitmap;
+}
+
+static int gr2d_is_addr_reg(struct device *dev, u32 class, u32 reg)
+{
+	struct gr2d *gr2d = dev_get_drvdata(dev);
+
+	switch (class) {
+	case HOST1X_CLASS_HOST1X:
+		return reg == 0x2b;
+	case HOST1X_CLASS_GR2D:
+	case HOST1X_CLASS_GR2D_SB:
+		reg &= 0xff;
+		if (gr2d->addr_regs[BIT_WORD(reg)] & BIT_MASK(reg))
+			return 1;
+	default:
+		return 0;
+	}
+}
+
+static const struct of_device_id gr2d_match[] = {
+	{ .compatible = "nvidia,tegra30-gr2d" },
+	{ .compatible = "nvidia,tegra20-gr2d" },
+	{ },
+};
+
+static int gr2d_probe(struct platform_device *pdev)
+{
+	struct device *dev = &pdev->dev;
+	struct host1x_drm *host1x = host1x_get_drm_data(dev->parent);
+	int err;
+	struct gr2d *gr2d = NULL;
+	struct host1x_syncpt **syncpts;
+
+	gr2d = devm_kzalloc(dev, sizeof(*gr2d), GFP_KERNEL);
+	if (!gr2d)
+		return -ENOMEM;
+
+	syncpts = devm_kzalloc(dev, sizeof(*syncpts), GFP_KERNEL);
+	if (!syncpts)
+		return -ENOMEM;
+
+	gr2d->clk = devm_clk_get(dev, NULL);
+	if (IS_ERR(gr2d->clk)) {
+		dev_err(dev, "cannot get clock\n");
+		return PTR_ERR(gr2d->clk);
+	}
+
+	err = clk_prepare_enable(gr2d->clk);
+	if (err) {
+		dev_err(dev, "cannot turn on clock\n");
+		return err;
+	}
+
+	gr2d->channel = host1x_channel_request(dev);
+	if (!gr2d->channel)
+		return -ENOMEM;
+
+	*syncpts = host1x_syncpt_request(dev, 0);
+	if (!(*syncpts)) {
+		host1x_channel_free(gr2d->channel);
+		return -ENOMEM;
+	}
+
+	gr2d->client.ops = &gr2d_client_ops;
+	gr2d->client.dev = dev;
+	gr2d->client.class = HOST1X_CLASS_GR2D;
+	gr2d->client.syncpts = syncpts;
+	gr2d->client.num_syncpts = 1;
+
+	err = host1x_register_client(host1x, &gr2d->client);
+	if (err < 0) {
+		dev_err(dev, "failed to register host1x client: %d\n", err);
+		return err;
+	}
+
+	gr2d_init_addr_reg_map(dev, gr2d);
+
+	platform_set_drvdata(pdev, gr2d);
+
+	return 0;
+}
+
+static int __exit gr2d_remove(struct platform_device *pdev)
+{
+	struct host1x_drm *host1x = host1x_get_drm_data(pdev->dev.parent);
+	struct gr2d *gr2d = platform_get_drvdata(pdev);
+	unsigned int i;
+	int err;
+
+	err = host1x_unregister_client(host1x, &gr2d->client);
+	if (err < 0) {
+		dev_err(&pdev->dev, "failed to unregister client: %d\n", err);
+		return err;
+	}
+
+	for (i = 0; i < gr2d->client.num_syncpts; i++)
+		host1x_syncpt_free(gr2d->client.syncpts[i]);
+
+	host1x_channel_free(gr2d->channel);
+	clk_disable_unprepare(gr2d->clk);
+
+	return 0;
+}
+
+struct platform_driver tegra_gr2d_driver = {
+	.probe = gr2d_probe,
+	.remove = __exit_p(gr2d_remove),
+	.driver = {
+		.owner = THIS_MODULE,
+		.name = "gr2d",
+		.of_match_table = gr2d_match,
+	}
+};
diff --git a/drivers/gpu/drm/tegra/hdmi.c b/drivers/gpu/host1x/drm/hdmi.c
index bb747f6cd1a4..01097da09f7f 100644
--- a/drivers/gpu/drm/tegra/hdmi.c
+++ b/drivers/gpu/host1x/drm/hdmi.c
@@ -22,6 +22,7 @@
 #include "hdmi.h"
 #include "drm.h"
 #include "dc.h"
+#include "host1x_client.h"
 
 struct tegra_hdmi {
 	struct host1x_client client;
@@ -1189,7 +1190,7 @@ static const struct host1x_client_ops hdmi_client_ops = {
 
 static int tegra_hdmi_probe(struct platform_device *pdev)
 {
-	struct host1x *host1x = dev_get_drvdata(pdev->dev.parent);
+	struct host1x_drm *host1x = host1x_get_drm_data(pdev->dev.parent);
 	struct tegra_hdmi *hdmi;
 	struct resource *regs;
 	int err;
@@ -1278,7 +1279,7 @@ static int tegra_hdmi_probe(struct platform_device *pdev)
 
 static int tegra_hdmi_remove(struct platform_device *pdev)
 {
-	struct host1x *host1x = dev_get_drvdata(pdev->dev.parent);
+	struct host1x_drm *host1x = host1x_get_drm_data(pdev->dev.parent);
 	struct tegra_hdmi *hdmi = platform_get_drvdata(pdev);
 	int err;
 
diff --git a/drivers/gpu/drm/tegra/hdmi.h b/drivers/gpu/host1x/drm/hdmi.h
index 52ac36e08ccb..52ac36e08ccb 100644
--- a/drivers/gpu/drm/tegra/hdmi.h
+++ b/drivers/gpu/host1x/drm/hdmi.h
diff --git a/drivers/gpu/drm/tegra/output.c b/drivers/gpu/host1x/drm/output.c
index 8140fc6c34d8..8140fc6c34d8 100644
--- a/drivers/gpu/drm/tegra/output.c
+++ b/drivers/gpu/host1x/drm/output.c
diff --git a/drivers/gpu/drm/tegra/rgb.c b/drivers/gpu/host1x/drm/rgb.c
index ed4416f20260..ed4416f20260 100644
--- a/drivers/gpu/drm/tegra/rgb.c
+++ b/drivers/gpu/host1x/drm/rgb.c
diff --git a/drivers/gpu/host1x/host1x.h b/drivers/gpu/host1x/host1x.h
new file mode 100644
index 000000000000..a2bc1e65e972
--- /dev/null
+++ b/drivers/gpu/host1x/host1x.h
@@ -0,0 +1,30 @@
+/*
+ * Tegra host1x driver
+ *
+ * Copyright (c) 2009-2013, NVIDIA Corporation. All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along
+ * with this program; if not, write to the Free Software Foundation, Inc.,
+ * 51 Franklin Street, Fifth Floor, Boston, MA  02110-1301, USA.
+ */
+
+#ifndef __LINUX_HOST1X_H
+#define __LINUX_HOST1X_H
+
+enum host1x_class {
+	HOST1X_CLASS_HOST1X	= 0x1,
+	HOST1X_CLASS_GR2D	= 0x51,
+	HOST1X_CLASS_GR2D_SB    = 0x52
+};
+
+#endif
diff --git a/drivers/gpu/host1x/host1x_bo.h b/drivers/gpu/host1x/host1x_bo.h
new file mode 100644
index 000000000000..4c1f10bd773d
--- /dev/null
+++ b/drivers/gpu/host1x/host1x_bo.h
@@ -0,0 +1,87 @@
+/*
+ * Tegra host1x Memory Management Abstraction header
+ *
+ * Copyright (c) 2012-2013, NVIDIA Corporation.
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms and conditions of the GNU General Public License,
+ * version 2, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program.  If not, see <http://www.gnu.org/licenses/>.
+ */
+
+#ifndef _HOST1X_BO_H
+#define _HOST1X_BO_H
+
+struct host1x_bo;
+
+struct host1x_bo_ops {
+	struct host1x_bo *(*get)(struct host1x_bo *bo);
+	void (*put)(struct host1x_bo *bo);
+	dma_addr_t (*pin)(struct host1x_bo *bo, struct sg_table **sgt);
+	void (*unpin)(struct host1x_bo *bo, struct sg_table *sgt);
+	void *(*mmap)(struct host1x_bo *bo);
+	void (*munmap)(struct host1x_bo *bo, void *addr);
+	void *(*kmap)(struct host1x_bo *bo, unsigned int pagenum);
+	void (*kunmap)(struct host1x_bo *bo, unsigned int pagenum, void *addr);
+};
+
+struct host1x_bo {
+	const struct host1x_bo_ops *ops;
+};
+
+static inline void host1x_bo_init(struct host1x_bo *bo,
+				  const struct host1x_bo_ops *ops)
+{
+	bo->ops = ops;
+}
+
+static inline struct host1x_bo *host1x_bo_get(struct host1x_bo *bo)
+{
+	return bo->ops->get(bo);
+}
+
+static inline void host1x_bo_put(struct host1x_bo *bo)
+{
+	bo->ops->put(bo);
+}
+
+static inline dma_addr_t host1x_bo_pin(struct host1x_bo *bo,
+				       struct sg_table **sgt)
+{
+	return bo->ops->pin(bo, sgt);
+}
+
+static inline void host1x_bo_unpin(struct host1x_bo *bo, struct sg_table *sgt)
+{
+	bo->ops->unpin(bo, sgt);
+}
+
+static inline void *host1x_bo_mmap(struct host1x_bo *bo)
+{
+	return bo->ops->mmap(bo);
+}
+
+static inline void host1x_bo_munmap(struct host1x_bo *bo, void *addr)
+{
+	bo->ops->munmap(bo, addr);
+}
+
+static inline void *host1x_bo_kmap(struct host1x_bo *bo, unsigned int pagenum)
+{
+	return bo->ops->kmap(bo, pagenum);
+}
+
+static inline void host1x_bo_kunmap(struct host1x_bo *bo,
+				    unsigned int pagenum, void *addr)
+{
+	bo->ops->kunmap(bo, pagenum, addr);
+}
+
+#endif
diff --git a/drivers/gpu/host1x/host1x_client.h b/drivers/gpu/host1x/host1x_client.h
new file mode 100644
index 000000000000..9b85f10f4a44
--- /dev/null
+++ b/drivers/gpu/host1x/host1x_client.h
@@ -0,0 +1,35 @@
+/*
+ * Copyright (c) 2013, NVIDIA Corporation.
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms and conditions of the GNU General Public License,
+ * version 2, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program.  If not, see <http://www.gnu.org/licenses/>.
+ */
+
+#ifndef HOST1X_CLIENT_H
+#define HOST1X_CLIENT_H
+
+struct device;
+struct platform_device;
+
+#ifdef CONFIG_DRM_TEGRA
+int host1x_drm_alloc(struct platform_device *pdev);
+#else
+static inline int host1x_drm_alloc(struct platform_device *pdev)
+{
+	return 0;
+}
+#endif
+
+void host1x_set_drm_data(struct device *dev, void *data);
+void *host1x_get_drm_data(struct device *dev);
+
+#endif
diff --git a/drivers/gpu/host1x/hw/Makefile b/drivers/gpu/host1x/hw/Makefile
new file mode 100644
index 000000000000..9b50863a2236
--- /dev/null
+++ b/drivers/gpu/host1x/hw/Makefile
@@ -0,0 +1,6 @@
+ccflags-y = -Idrivers/gpu/host1x
+
+host1x-hw-objs  = \
+	host1x01.o
+
+obj-$(CONFIG_TEGRA_HOST1X) += host1x-hw.o
diff --git a/drivers/gpu/host1x/hw/cdma_hw.c b/drivers/gpu/host1x/hw/cdma_hw.c
new file mode 100644
index 000000000000..590b69d91dab
--- /dev/null
+++ b/drivers/gpu/host1x/hw/cdma_hw.c
@@ -0,0 +1,326 @@
+/*
+ * Tegra host1x Command DMA
+ *
+ * Copyright (c) 2010-2013, NVIDIA Corporation.
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms and conditions of the GNU General Public License,
+ * version 2, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program.  If not, see <http://www.gnu.org/licenses/>.
+ */
+
+#include <linux/slab.h>
+#include <linux/scatterlist.h>
+#include <linux/dma-mapping.h>
+
+#include "cdma.h"
+#include "channel.h"
+#include "dev.h"
+#include "debug.h"
+
+/*
+ * Put the restart at the end of pushbuffer memor
+ */
+static void push_buffer_init(struct push_buffer *pb)
+{
+	*(pb->mapped + (pb->size_bytes >> 2)) = host1x_opcode_restart(0);
+}
+
+/*
+ * Increment timedout buffer's syncpt via CPU.
+ */
+static void cdma_timeout_cpu_incr(struct host1x_cdma *cdma, u32 getptr,
+				u32 syncpt_incrs, u32 syncval, u32 nr_slots)
+{
+	struct host1x *host1x = cdma_to_host1x(cdma);
+	struct push_buffer *pb = &cdma->push_buffer;
+	u32 i;
+
+	for (i = 0; i < syncpt_incrs; i++)
+		host1x_syncpt_cpu_incr(cdma->timeout.syncpt);
+
+	/* after CPU incr, ensure shadow is up to date */
+	host1x_syncpt_load(cdma->timeout.syncpt);
+
+	/* NOP all the PB slots */
+	while (nr_slots--) {
+		u32 *p = (u32 *)((u32)pb->mapped + getptr);
+		*(p++) = HOST1X_OPCODE_NOP;
+		*(p++) = HOST1X_OPCODE_NOP;
+		dev_dbg(host1x->dev, "%s: NOP at 0x%x\n", __func__,
+			pb->phys + getptr);
+		getptr = (getptr + 8) & (pb->size_bytes - 1);
+	}
+	wmb();
+}
+
+/*
+ * Start channel DMA
+ */
+static void cdma_start(struct host1x_cdma *cdma)
+{
+	struct host1x_channel *ch = cdma_to_channel(cdma);
+
+	if (cdma->running)
+		return;
+
+	cdma->last_pos = cdma->push_buffer.pos;
+
+	host1x_ch_writel(ch, HOST1X_CHANNEL_DMACTRL_DMASTOP,
+			 HOST1X_CHANNEL_DMACTRL);
+
+	/* set base, put and end pointer */
+	host1x_ch_writel(ch, cdma->push_buffer.phys, HOST1X_CHANNEL_DMASTART);
+	host1x_ch_writel(ch, cdma->push_buffer.pos, HOST1X_CHANNEL_DMAPUT);
+	host1x_ch_writel(ch, cdma->push_buffer.phys +
+			 cdma->push_buffer.size_bytes + 4,
+			 HOST1X_CHANNEL_DMAEND);
+
+	/* reset GET */
+	host1x_ch_writel(ch, HOST1X_CHANNEL_DMACTRL_DMASTOP |
+			 HOST1X_CHANNEL_DMACTRL_DMAGETRST |
+			 HOST1X_CHANNEL_DMACTRL_DMAINITGET,
+			 HOST1X_CHANNEL_DMACTRL);
+
+	/* start the command DMA */
+	host1x_ch_writel(ch, 0, HOST1X_CHANNEL_DMACTRL);
+
+	cdma->running = true;
+}
+
+/*
+ * Similar to cdma_start(), but rather than starting from an idle
+ * state (where DMA GET is set to DMA PUT), on a timeout we restore
+ * DMA GET from an explicit value (so DMA may again be pending).
+ */
+static void cdma_timeout_restart(struct host1x_cdma *cdma, u32 getptr)
+{
+	struct host1x *host1x = cdma_to_host1x(cdma);
+	struct host1x_channel *ch = cdma_to_channel(cdma);
+
+	if (cdma->running)
+		return;
+
+	cdma->last_pos = cdma->push_buffer.pos;
+
+	host1x_ch_writel(ch, HOST1X_CHANNEL_DMACTRL_DMASTOP,
+			 HOST1X_CHANNEL_DMACTRL);
+
+	/* set base, end pointer (all of memory) */
+	host1x_ch_writel(ch, cdma->push_buffer.phys, HOST1X_CHANNEL_DMASTART);
+	host1x_ch_writel(ch, cdma->push_buffer.phys +
+			 cdma->push_buffer.size_bytes,
+			 HOST1X_CHANNEL_DMAEND);
+
+	/* set GET, by loading the value in PUT (then reset GET) */
+	host1x_ch_writel(ch, getptr, HOST1X_CHANNEL_DMAPUT);
+	host1x_ch_writel(ch, HOST1X_CHANNEL_DMACTRL_DMASTOP |
+			 HOST1X_CHANNEL_DMACTRL_DMAGETRST |
+			 HOST1X_CHANNEL_DMACTRL_DMAINITGET,
+			 HOST1X_CHANNEL_DMACTRL);
+
+	dev_dbg(host1x->dev,
+		"%s: DMA GET 0x%x, PUT HW 0x%x / shadow 0x%x\n", __func__,
+		host1x_ch_readl(ch, HOST1X_CHANNEL_DMAGET),
+		host1x_ch_readl(ch, HOST1X_CHANNEL_DMAPUT),
+		cdma->last_pos);
+
+	/* deassert GET reset and set PUT */
+	host1x_ch_writel(ch, HOST1X_CHANNEL_DMACTRL_DMASTOP,
+			 HOST1X_CHANNEL_DMACTRL);
+	host1x_ch_writel(ch, cdma->push_buffer.pos, HOST1X_CHANNEL_DMAPUT);
+
+	/* start the command DMA */
+	host1x_ch_writel(ch, 0, HOST1X_CHANNEL_DMACTRL);
+
+	cdma->running = true;
+}
+
+/*
+ * Kick channel DMA into action by writing its PUT offset (if it has changed)
+ */
+static void cdma_flush(struct host1x_cdma *cdma)
+{
+	struct host1x_channel *ch = cdma_to_channel(cdma);
+
+	if (cdma->push_buffer.pos != cdma->last_pos) {
+		host1x_ch_writel(ch, cdma->push_buffer.pos,
+				 HOST1X_CHANNEL_DMAPUT);
+		cdma->last_pos = cdma->push_buffer.pos;
+	}
+}
+
+static void cdma_stop(struct host1x_cdma *cdma)
+{
+	struct host1x_channel *ch = cdma_to_channel(cdma);
+
+	mutex_lock(&cdma->lock);
+	if (cdma->running) {
+		host1x_cdma_wait_locked(cdma, CDMA_EVENT_SYNC_QUEUE_EMPTY);
+		host1x_ch_writel(ch, HOST1X_CHANNEL_DMACTRL_DMASTOP,
+				 HOST1X_CHANNEL_DMACTRL);
+		cdma->running = false;
+	}
+	mutex_unlock(&cdma->lock);
+}
+
+/*
+ * Stops both channel's command processor and CDMA immediately.
+ * Also, tears down the channel and resets corresponding module.
+ */
+static void cdma_freeze(struct host1x_cdma *cdma)
+{
+	struct host1x *host = cdma_to_host1x(cdma);
+	struct host1x_channel *ch = cdma_to_channel(cdma);
+	u32 cmdproc_stop;
+
+	if (cdma->torndown && !cdma->running) {
+		dev_warn(host->dev, "Already torn down\n");
+		return;
+	}
+
+	dev_dbg(host->dev, "freezing channel (id %d)\n", ch->id);
+
+	cmdproc_stop = host1x_sync_readl(host, HOST1X_SYNC_CMDPROC_STOP);
+	cmdproc_stop |= BIT(ch->id);
+	host1x_sync_writel(host, cmdproc_stop, HOST1X_SYNC_CMDPROC_STOP);
+
+	dev_dbg(host->dev, "%s: DMA GET 0x%x, PUT HW 0x%x / shadow 0x%x\n",
+		__func__, host1x_ch_readl(ch, HOST1X_CHANNEL_DMAGET),
+		host1x_ch_readl(ch, HOST1X_CHANNEL_DMAPUT),
+		cdma->last_pos);
+
+	host1x_ch_writel(ch, HOST1X_CHANNEL_DMACTRL_DMASTOP,
+			 HOST1X_CHANNEL_DMACTRL);
+
+	host1x_sync_writel(host, BIT(ch->id), HOST1X_SYNC_CH_TEARDOWN);
+
+	cdma->running = false;
+	cdma->torndown = true;
+}
+
+static void cdma_resume(struct host1x_cdma *cdma, u32 getptr)
+{
+	struct host1x *host1x = cdma_to_host1x(cdma);
+	struct host1x_channel *ch = cdma_to_channel(cdma);
+	u32 cmdproc_stop;
+
+	dev_dbg(host1x->dev,
+		"resuming channel (id %d, DMAGET restart = 0x%x)\n",
+		ch->id, getptr);
+
+	cmdproc_stop = host1x_sync_readl(host1x, HOST1X_SYNC_CMDPROC_STOP);
+	cmdproc_stop &= ~(BIT(ch->id));
+	host1x_sync_writel(host1x, cmdproc_stop, HOST1X_SYNC_CMDPROC_STOP);
+
+	cdma->torndown = false;
+	cdma_timeout_restart(cdma, getptr);
+}
+
+/*
+ * If this timeout fires, it indicates the current sync_queue entry has
+ * exceeded its TTL and the userctx should be timed out and remaining
+ * submits already issued cleaned up (future submits return an error).
+ */
+static void cdma_timeout_handler(struct work_struct *work)
+{
+	struct host1x_cdma *cdma;
+	struct host1x *host1x;
+	struct host1x_channel *ch;
+
+	u32 syncpt_val;
+
+	u32 prev_cmdproc, cmdproc_stop;
+
+	cdma = container_of(to_delayed_work(work), struct host1x_cdma,
+			    timeout.wq);
+	host1x = cdma_to_host1x(cdma);
+	ch = cdma_to_channel(cdma);
+
+	host1x_debug_dump(cdma_to_host1x(cdma));
+
+	mutex_lock(&cdma->lock);
+
+	if (!cdma->timeout.client) {
+		dev_dbg(host1x->dev,
+			"cdma_timeout: expired, but has no clientid\n");
+		mutex_unlock(&cdma->lock);
+		return;
+	}
+
+	/* stop processing to get a clean snapshot */
+	prev_cmdproc = host1x_sync_readl(host1x, HOST1X_SYNC_CMDPROC_STOP);
+	cmdproc_stop = prev_cmdproc | BIT(ch->id);
+	host1x_sync_writel(host1x, cmdproc_stop, HOST1X_SYNC_CMDPROC_STOP);
+
+	dev_dbg(host1x->dev, "cdma_timeout: cmdproc was 0x%x is 0x%x\n",
+		prev_cmdproc, cmdproc_stop);
+
+	syncpt_val = host1x_syncpt_load(cdma->timeout.syncpt);
+
+	/* has buffer actually completed? */
+	if ((s32)(syncpt_val - cdma->timeout.syncpt_val) >= 0) {
+		dev_dbg(host1x->dev,
+			"cdma_timeout: expired, but buffer had completed\n");
+		/* restore */
+		cmdproc_stop = prev_cmdproc & ~(BIT(ch->id));
+		host1x_sync_writel(host1x, cmdproc_stop,
+				   HOST1X_SYNC_CMDPROC_STOP);
+		mutex_unlock(&cdma->lock);
+		return;
+	}
+
+	dev_warn(host1x->dev, "%s: timeout: %d (%s), HW thresh %d, done %d\n",
+		__func__, cdma->timeout.syncpt->id, cdma->timeout.syncpt->name,
+		syncpt_val, cdma->timeout.syncpt_val);
+
+	/* stop HW, resetting channel/module */
+	host1x_hw_cdma_freeze(host1x, cdma);
+
+	host1x_cdma_update_sync_queue(cdma, ch->dev);
+	mutex_unlock(&cdma->lock);
+}
+
+/*
+ * Init timeout resources
+ */
+static int cdma_timeout_init(struct host1x_cdma *cdma, u32 syncpt_id)
+{
+	INIT_DELAYED_WORK(&cdma->timeout.wq, cdma_timeout_handler);
+	cdma->timeout.initialized = true;
+
+	return 0;
+}
+
+/*
+ * Clean up timeout resources
+ */
+static void cdma_timeout_destroy(struct host1x_cdma *cdma)
+{
+	if (cdma->timeout.initialized)
+		cancel_delayed_work(&cdma->timeout.wq);
+	cdma->timeout.initialized = false;
+}
+
+static const struct host1x_cdma_ops host1x_cdma_ops = {
+	.start = cdma_start,
+	.stop = cdma_stop,
+	.flush = cdma_flush,
+
+	.timeout_init = cdma_timeout_init,
+	.timeout_destroy = cdma_timeout_destroy,
+	.freeze = cdma_freeze,
+	.resume = cdma_resume,
+	.timeout_cpu_incr = cdma_timeout_cpu_incr,
+};
+
+static const struct host1x_pushbuffer_ops host1x_pushbuffer_ops = {
+	.init = push_buffer_init,
+};
diff --git a/drivers/gpu/host1x/hw/channel_hw.c b/drivers/gpu/host1x/hw/channel_hw.c
new file mode 100644
index 000000000000..ee199623e365
--- /dev/null
+++ b/drivers/gpu/host1x/hw/channel_hw.c
@@ -0,0 +1,168 @@
+/*
+ * Tegra host1x Channel
+ *
+ * Copyright (c) 2010-2013, NVIDIA Corporation.
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms and conditions of the GNU General Public License,
+ * version 2, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program.  If not, see <http://www.gnu.org/licenses/>.
+ */
+
+#include <linux/slab.h>
+#include <trace/events/host1x.h>
+
+#include "host1x.h"
+#include "host1x_bo.h"
+#include "channel.h"
+#include "dev.h"
+#include "intr.h"
+#include "job.h"
+
+#define HOST1X_CHANNEL_SIZE 16384
+#define TRACE_MAX_LENGTH 128U
+
+static void trace_write_gather(struct host1x_cdma *cdma, struct host1x_bo *bo,
+			       u32 offset, u32 words)
+{
+	void *mem = NULL;
+
+	if (host1x_debug_trace_cmdbuf)
+		mem = host1x_bo_mmap(bo);
+
+	if (mem) {
+		u32 i;
+		/*
+		 * Write in batches of 128 as there seems to be a limit
+		 * of how much you can output to ftrace at once.
+		 */
+		for (i = 0; i < words; i += TRACE_MAX_LENGTH) {
+			trace_host1x_cdma_push_gather(
+				dev_name(cdma_to_channel(cdma)->dev),
+				(u32)bo, min(words - i, TRACE_MAX_LENGTH),
+				offset + i * sizeof(u32), mem);
+		}
+		host1x_bo_munmap(bo, mem);
+	}
+}
+
+static void submit_gathers(struct host1x_job *job)
+{
+	struct host1x_cdma *cdma = &job->channel->cdma;
+	unsigned int i;
+
+	for (i = 0; i < job->num_gathers; i++) {
+		struct host1x_job_gather *g = &job->gathers[i];
+		u32 op1 = host1x_opcode_gather(g->words);
+		u32 op2 = g->base + g->offset;
+		trace_write_gather(cdma, g->bo, g->offset, op1 & 0xffff);
+		host1x_cdma_push(cdma, op1, op2);
+	}
+}
+
+static int channel_submit(struct host1x_job *job)
+{
+	struct host1x_channel *ch = job->channel;
+	struct host1x_syncpt *sp;
+	u32 user_syncpt_incrs = job->syncpt_incrs;
+	u32 prev_max = 0;
+	u32 syncval;
+	int err;
+	struct host1x_waitlist *completed_waiter = NULL;
+	struct host1x *host = dev_get_drvdata(ch->dev->parent);
+
+	sp = host->syncpt + job->syncpt_id;
+	trace_host1x_channel_submit(dev_name(ch->dev),
+				    job->num_gathers, job->num_relocs,
+				    job->num_waitchk, job->syncpt_id,
+				    job->syncpt_incrs);
+
+	/* before error checks, return current max */
+	prev_max = job->syncpt_end = host1x_syncpt_read_max(sp);
+
+	/* get submit lock */
+	err = mutex_lock_interruptible(&ch->submitlock);
+	if (err)
+		goto error;
+
+	completed_waiter = kzalloc(sizeof(*completed_waiter), GFP_KERNEL);
+	if (!completed_waiter) {
+		mutex_unlock(&ch->submitlock);
+		err = -ENOMEM;
+		goto error;
+	}
+
+	/* begin a CDMA submit */
+	err = host1x_cdma_begin(&ch->cdma, job);
+	if (err) {
+		mutex_unlock(&ch->submitlock);
+		goto error;
+	}
+
+	if (job->serialize) {
+		/*
+		 * Force serialization by inserting a host wait for the
+		 * previous job to finish before this one can commence.
+		 */
+		host1x_cdma_push(&ch->cdma,
+				 host1x_opcode_setclass(HOST1X_CLASS_HOST1X,
+					host1x_uclass_wait_syncpt_r(), 1),
+				 host1x_class_host_wait_syncpt(job->syncpt_id,
+					host1x_syncpt_read_max(sp)));
+	}
+
+	syncval = host1x_syncpt_incr_max(sp, user_syncpt_incrs);
+
+	job->syncpt_end = syncval;
+
+	/* add a setclass for modules that require it */
+	if (job->class)
+		host1x_cdma_push(&ch->cdma,
+				 host1x_opcode_setclass(job->class, 0, 0),
+				 HOST1X_OPCODE_NOP);
+
+	submit_gathers(job);
+
+	/* end CDMA submit & stash pinned hMems into sync queue */
+	host1x_cdma_end(&ch->cdma, job);
+
+	trace_host1x_channel_submitted(dev_name(ch->dev), prev_max, syncval);
+
+	/* schedule a submit complete interrupt */
+	err = host1x_intr_add_action(host, job->syncpt_id, syncval,
+				     HOST1X_INTR_ACTION_SUBMIT_COMPLETE, ch,
+				     completed_waiter, NULL);
+	completed_waiter = NULL;
+	WARN(err, "Failed to set submit complete interrupt");
+
+	mutex_unlock(&ch->submitlock);
+
+	return 0;
+
+error:
+	kfree(completed_waiter);
+	return err;
+}
+
+static int host1x_channel_init(struct host1x_channel *ch, struct host1x *dev,
+			       unsigned int index)
+{
+	ch->id = index;
+	mutex_init(&ch->reflock);
+	mutex_init(&ch->submitlock);
+
+	ch->regs = dev->regs + index * HOST1X_CHANNEL_SIZE;
+	return 0;
+}
+
+static const struct host1x_channel_ops host1x_channel_ops = {
+	.init = host1x_channel_init,
+	.submit = channel_submit,
+};
diff --git a/drivers/gpu/host1x/hw/debug_hw.c b/drivers/gpu/host1x/hw/debug_hw.c
new file mode 100644
index 000000000000..334c038052f5
--- /dev/null
+++ b/drivers/gpu/host1x/hw/debug_hw.c
@@ -0,0 +1,322 @@
+/*
+ * Copyright (C) 2010 Google, Inc.
+ * Author: Erik Gilling <konkers@android.com>
+ *
+ * Copyright (C) 2011-2013 NVIDIA Corporation
+ *
+ * This software is licensed under the terms of the GNU General Public
+ * License version 2, as published by the Free Software Foundation, and
+ * may be copied, distributed, and modified under those terms.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ */
+
+#include <linux/debugfs.h>
+#include <linux/seq_file.h>
+#include <linux/mm.h>
+#include <linux/scatterlist.h>
+
+#include <linux/io.h>
+
+#include "dev.h"
+#include "debug.h"
+#include "cdma.h"
+#include "channel.h"
+#include "host1x_bo.h"
+
+#define HOST1X_DEBUG_MAX_PAGE_OFFSET 102400
+
+enum {
+	HOST1X_OPCODE_SETCLASS	= 0x00,
+	HOST1X_OPCODE_INCR	= 0x01,
+	HOST1X_OPCODE_NONINCR	= 0x02,
+	HOST1X_OPCODE_MASK	= 0x03,
+	HOST1X_OPCODE_IMM	= 0x04,
+	HOST1X_OPCODE_RESTART	= 0x05,
+	HOST1X_OPCODE_GATHER	= 0x06,
+	HOST1X_OPCODE_EXTEND	= 0x0e,
+};
+
+enum {
+	HOST1X_OPCODE_EXTEND_ACQUIRE_MLOCK	= 0x00,
+	HOST1X_OPCODE_EXTEND_RELEASE_MLOCK	= 0x01,
+};
+
+static unsigned int show_channel_command(struct output *o, u32 val)
+{
+	unsigned mask;
+	unsigned subop;
+
+	switch (val >> 28) {
+	case HOST1X_OPCODE_SETCLASS:
+		mask = val & 0x3f;
+		if (mask) {
+			host1x_debug_output(o, "SETCL(class=%03x, offset=%03x, mask=%02x, [",
+					    val >> 6 & 0x3ff,
+					    val >> 16 & 0xfff, mask);
+			return hweight8(mask);
+		} else {
+			host1x_debug_output(o, "SETCL(class=%03x)\n",
+					    val >> 6 & 0x3ff);
+			return 0;
+		}
+
+	case HOST1X_OPCODE_INCR:
+		host1x_debug_output(o, "INCR(offset=%03x, [",
+				    val >> 16 & 0xfff);
+		return val & 0xffff;
+
+	case HOST1X_OPCODE_NONINCR:
+		host1x_debug_output(o, "NONINCR(offset=%03x, [",
+				    val >> 16 & 0xfff);
+		return val & 0xffff;
+
+	case HOST1X_OPCODE_MASK:
+		mask = val & 0xffff;
+		host1x_debug_output(o, "MASK(offset=%03x, mask=%03x, [",
+				    val >> 16 & 0xfff, mask);
+		return hweight16(mask);
+
+	case HOST1X_OPCODE_IMM:
+		host1x_debug_output(o, "IMM(offset=%03x, data=%03x)\n",
+				    val >> 16 & 0xfff, val & 0xffff);
+		return 0;
+
+	case HOST1X_OPCODE_RESTART:
+		host1x_debug_output(o, "RESTART(offset=%08x)\n", val << 4);
+		return 0;
+
+	case HOST1X_OPCODE_GATHER:
+		host1x_debug_output(o, "GATHER(offset=%03x, insert=%d, type=%d, count=%04x, addr=[",
+				    val >> 16 & 0xfff, val >> 15 & 0x1,
+				    val >> 14 & 0x1, val & 0x3fff);
+		return 1;
+
+	case HOST1X_OPCODE_EXTEND:
+		subop = val >> 24 & 0xf;
+		if (subop == HOST1X_OPCODE_EXTEND_ACQUIRE_MLOCK)
+			host1x_debug_output(o, "ACQUIRE_MLOCK(index=%d)\n",
+					    val & 0xff);
+		else if (subop == HOST1X_OPCODE_EXTEND_RELEASE_MLOCK)
+			host1x_debug_output(o, "RELEASE_MLOCK(index=%d)\n",
+					    val & 0xff);
+		else
+			host1x_debug_output(o, "EXTEND_UNKNOWN(%08x)\n", val);
+		return 0;
+
+	default:
+		return 0;
+	}
+}
+
+static void show_gather(struct output *o, phys_addr_t phys_addr,
+			unsigned int words, struct host1x_cdma *cdma,
+			phys_addr_t pin_addr, u32 *map_addr)
+{
+	/* Map dmaget cursor to corresponding mem handle */
+	u32 offset = phys_addr - pin_addr;
+	unsigned int data_count = 0, i;
+
+	/*
+	 * Sometimes we're given different hardware address to the same
+	 * page - in these cases the offset will get an invalid number and
+	 * we just have to bail out.
+	 */
+	if (offset > HOST1X_DEBUG_MAX_PAGE_OFFSET) {
+		host1x_debug_output(o, "[address mismatch]\n");
+		return;
+	}
+
+	for (i = 0; i < words; i++) {
+		u32 addr = phys_addr + i * 4;
+		u32 val = *(map_addr + offset / 4 + i);
+
+		if (!data_count) {
+			host1x_debug_output(o, "%08x: %08x:", addr, val);
+			data_count = show_channel_command(o, val);
+		} else {
+			host1x_debug_output(o, "%08x%s", val,
+					    data_count > 0 ? ", " : "])\n");
+			data_count--;
+		}
+	}
+}
+
+static void show_channel_gathers(struct output *o, struct host1x_cdma *cdma)
+{
+	struct host1x_job *job;
+
+	list_for_each_entry(job, &cdma->sync_queue, list) {
+		int i;
+		host1x_debug_output(o, "\n%p: JOB, syncpt_id=%d, syncpt_val=%d, first_get=%08x, timeout=%d num_slots=%d, num_handles=%d\n",
+				    job, job->syncpt_id, job->syncpt_end,
+				    job->first_get, job->timeout,
+				    job->num_slots, job->num_unpins);
+
+		for (i = 0; i < job->num_gathers; i++) {
+			struct host1x_job_gather *g = &job->gathers[i];
+			u32 *mapped;
+
+			if (job->gather_copy_mapped)
+				mapped = (u32 *)job->gather_copy_mapped;
+			else
+				mapped = host1x_bo_mmap(g->bo);
+
+			if (!mapped) {
+				host1x_debug_output(o, "[could not mmap]\n");
+				continue;
+			}
+
+			host1x_debug_output(o, "    GATHER at %08x+%04x, %d words\n",
+					    g->base, g->offset, g->words);
+
+			show_gather(o, g->base + g->offset, g->words, cdma,
+				    g->base, mapped);
+
+			if (!job->gather_copy_mapped)
+				host1x_bo_munmap(g->bo, mapped);
+		}
+	}
+}
+
+static void host1x_debug_show_channel_cdma(struct host1x *host,
+					   struct host1x_channel *ch,
+					   struct output *o)
+{
+	struct host1x_cdma *cdma = &ch->cdma;
+	u32 dmaput, dmaget, dmactrl;
+	u32 cbstat, cbread;
+	u32 val, base, baseval;
+
+	dmaput = host1x_ch_readl(ch, HOST1X_CHANNEL_DMAPUT);
+	dmaget = host1x_ch_readl(ch, HOST1X_CHANNEL_DMAGET);
+	dmactrl = host1x_ch_readl(ch, HOST1X_CHANNEL_DMACTRL);
+	cbread = host1x_sync_readl(host, HOST1X_SYNC_CBREAD(ch->id));
+	cbstat = host1x_sync_readl(host, HOST1X_SYNC_CBSTAT(ch->id));
+
+	host1x_debug_output(o, "%d-%s: ", ch->id, dev_name(ch->dev));
+
+	if (HOST1X_CHANNEL_DMACTRL_DMASTOP_V(dmactrl) ||
+	    !ch->cdma.push_buffer.mapped) {
+		host1x_debug_output(o, "inactive\n\n");
+		return;
+	}
+
+	if (HOST1X_SYNC_CBSTAT_CBCLASS_V(cbstat) == HOST1X_CLASS_HOST1X &&
+	    HOST1X_SYNC_CBSTAT_CBOFFSET_V(cbstat) ==
+	    HOST1X_UCLASS_WAIT_SYNCPT)
+		host1x_debug_output(o, "waiting on syncpt %d val %d\n",
+				    cbread >> 24, cbread & 0xffffff);
+	else if (HOST1X_SYNC_CBSTAT_CBCLASS_V(cbstat) ==
+	   HOST1X_CLASS_HOST1X &&
+	   HOST1X_SYNC_CBSTAT_CBOFFSET_V(cbstat) ==
+	   HOST1X_UCLASS_WAIT_SYNCPT_BASE) {
+
+		base = (cbread >> 16) & 0xff;
+		baseval =
+			host1x_sync_readl(host, HOST1X_SYNC_SYNCPT_BASE(base));
+		val = cbread & 0xffff;
+		host1x_debug_output(o, "waiting on syncpt %d val %d (base %d = %d; offset = %d)\n",
+				    cbread >> 24, baseval + val, base,
+				    baseval, val);
+	} else
+		host1x_debug_output(o, "active class %02x, offset %04x, val %08x\n",
+				    HOST1X_SYNC_CBSTAT_CBCLASS_V(cbstat),
+				    HOST1X_SYNC_CBSTAT_CBOFFSET_V(cbstat),
+				    cbread);
+
+	host1x_debug_output(o, "DMAPUT %08x, DMAGET %08x, DMACTL %08x\n",
+			    dmaput, dmaget, dmactrl);
+	host1x_debug_output(o, "CBREAD %08x, CBSTAT %08x\n", cbread, cbstat);
+
+	show_channel_gathers(o, cdma);
+	host1x_debug_output(o, "\n");
+}
+
+static void host1x_debug_show_channel_fifo(struct host1x *host,
+					   struct host1x_channel *ch,
+					   struct output *o)
+{
+	u32 val, rd_ptr, wr_ptr, start, end;
+	unsigned int data_count = 0;
+
+	host1x_debug_output(o, "%d: fifo:\n", ch->id);
+
+	val = host1x_ch_readl(ch, HOST1X_CHANNEL_FIFOSTAT);
+	host1x_debug_output(o, "FIFOSTAT %08x\n", val);
+	if (HOST1X_CHANNEL_FIFOSTAT_CFEMPTY_V(val)) {
+		host1x_debug_output(o, "[empty]\n");
+		return;
+	}
+
+	host1x_sync_writel(host, 0x0, HOST1X_SYNC_CFPEEK_CTRL);
+	host1x_sync_writel(host, HOST1X_SYNC_CFPEEK_CTRL_ENA_F(1) |
+			   HOST1X_SYNC_CFPEEK_CTRL_CHANNR_F(ch->id),
+			   HOST1X_SYNC_CFPEEK_CTRL);
+
+	val = host1x_sync_readl(host, HOST1X_SYNC_CFPEEK_PTRS);
+	rd_ptr = HOST1X_SYNC_CFPEEK_PTRS_CF_RD_PTR_V(val);
+	wr_ptr = HOST1X_SYNC_CFPEEK_PTRS_CF_WR_PTR_V(val);
+
+	val = host1x_sync_readl(host, HOST1X_SYNC_CF_SETUP(ch->id));
+	start = HOST1X_SYNC_CF_SETUP_BASE_V(val);
+	end = HOST1X_SYNC_CF_SETUP_LIMIT_V(val);
+
+	do {
+		host1x_sync_writel(host, 0x0, HOST1X_SYNC_CFPEEK_CTRL);
+		host1x_sync_writel(host, HOST1X_SYNC_CFPEEK_CTRL_ENA_F(1) |
+				   HOST1X_SYNC_CFPEEK_CTRL_CHANNR_F(ch->id) |
+				   HOST1X_SYNC_CFPEEK_CTRL_ADDR_F(rd_ptr),
+				   HOST1X_SYNC_CFPEEK_CTRL);
+		val = host1x_sync_readl(host, HOST1X_SYNC_CFPEEK_READ);
+
+		if (!data_count) {
+			host1x_debug_output(o, "%08x:", val);
+			data_count = show_channel_command(o, val);
+		} else {
+			host1x_debug_output(o, "%08x%s", val,
+					    data_count > 0 ? ", " : "])\n");
+			data_count--;
+		}
+
+		if (rd_ptr == end)
+			rd_ptr = start;
+		else
+			rd_ptr++;
+	} while (rd_ptr != wr_ptr);
+
+	if (data_count)
+		host1x_debug_output(o, ", ...])\n");
+	host1x_debug_output(o, "\n");
+
+	host1x_sync_writel(host, 0x0, HOST1X_SYNC_CFPEEK_CTRL);
+}
+
+static void host1x_debug_show_mlocks(struct host1x *host, struct output *o)
+{
+	int i;
+
+	host1x_debug_output(o, "---- mlocks ----\n");
+	for (i = 0; i < host1x_syncpt_nb_mlocks(host); i++) {
+		u32 owner =
+			host1x_sync_readl(host, HOST1X_SYNC_MLOCK_OWNER(i));
+		if (HOST1X_SYNC_MLOCK_OWNER_CH_OWNS_V(owner))
+			host1x_debug_output(o, "%d: locked by channel %d\n",
+				i, HOST1X_SYNC_MLOCK_OWNER_CHID_F(owner));
+		else if (HOST1X_SYNC_MLOCK_OWNER_CPU_OWNS_V(owner))
+			host1x_debug_output(o, "%d: locked by cpu\n", i);
+		else
+			host1x_debug_output(o, "%d: unlocked\n", i);
+	}
+	host1x_debug_output(o, "\n");
+}
+
+static const struct host1x_debug_ops host1x_debug_ops = {
+	.show_channel_cdma = host1x_debug_show_channel_cdma,
+	.show_channel_fifo = host1x_debug_show_channel_fifo,
+	.show_mlocks = host1x_debug_show_mlocks,
+};
diff --git a/drivers/gpu/host1x/hw/host1x01.c b/drivers/gpu/host1x/hw/host1x01.c
new file mode 100644
index 000000000000..a14e91cd1e58
--- /dev/null
+++ b/drivers/gpu/host1x/hw/host1x01.c
@@ -0,0 +1,42 @@
+/*
+ * Host1x init for T20 and T30 Architecture Chips
+ *
+ * Copyright (c) 2011-2013, NVIDIA Corporation.
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms and conditions of the GNU General Public License,
+ * version 2, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program.  If not, see <http://www.gnu.org/licenses/>.
+ */
+
+/* include hw specification */
+#include "hw/host1x01.h"
+#include "hw/host1x01_hardware.h"
+
+/* include code */
+#include "hw/cdma_hw.c"
+#include "hw/channel_hw.c"
+#include "hw/debug_hw.c"
+#include "hw/intr_hw.c"
+#include "hw/syncpt_hw.c"
+
+#include "dev.h"
+
+int host1x01_init(struct host1x *host)
+{
+	host->channel_op = &host1x_channel_ops;
+	host->cdma_op = &host1x_cdma_ops;
+	host->cdma_pb_op = &host1x_pushbuffer_ops;
+	host->syncpt_op = &host1x_syncpt_ops;
+	host->intr_op = &host1x_intr_ops;
+	host->debug_op = &host1x_debug_ops;
+
+	return 0;
+}
diff --git a/drivers/gpu/host1x/hw/host1x01.h b/drivers/gpu/host1x/hw/host1x01.h
new file mode 100644
index 000000000000..2706b6743250
--- /dev/null
+++ b/drivers/gpu/host1x/hw/host1x01.h
@@ -0,0 +1,25 @@
+/*
+ * Host1x init for T20 and T30 Architecture Chips
+ *
+ * Copyright (c) 2011-2013, NVIDIA Corporation.
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms and conditions of the GNU General Public License,
+ * version 2, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program.  If not, see <http://www.gnu.org/licenses/>.
+ */
+#ifndef HOST1X_HOST1X01_H
+#define HOST1X_HOST1X01_H
+
+struct host1x;
+
+int host1x01_init(struct host1x *host);
+
+#endif /* HOST1X_HOST1X01_H_ */
diff --git a/drivers/gpu/host1x/hw/host1x01_hardware.h b/drivers/gpu/host1x/hw/host1x01_hardware.h
new file mode 100644
index 000000000000..5f0fb866efa8
--- /dev/null
+++ b/drivers/gpu/host1x/hw/host1x01_hardware.h
@@ -0,0 +1,143 @@
+/*
+ * Tegra host1x Register Offsets for Tegra20 and Tegra30
+ *
+ * Copyright (c) 2010-2013 NVIDIA Corporation.
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms and conditions of the GNU General Public License,
+ * version 2, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program.  If not, see <http://www.gnu.org/licenses/>.
+ */
+
+#ifndef __HOST1X_HOST1X01_HARDWARE_H
+#define __HOST1X_HOST1X01_HARDWARE_H
+
+#include <linux/types.h>
+#include <linux/bitops.h>
+
+#include "hw_host1x01_channel.h"
+#include "hw_host1x01_sync.h"
+#include "hw_host1x01_uclass.h"
+
+static inline u32 host1x_class_host_wait_syncpt(
+	unsigned indx, unsigned threshold)
+{
+	return host1x_uclass_wait_syncpt_indx_f(indx)
+		| host1x_uclass_wait_syncpt_thresh_f(threshold);
+}
+
+static inline u32 host1x_class_host_load_syncpt_base(
+	unsigned indx, unsigned threshold)
+{
+	return host1x_uclass_load_syncpt_base_base_indx_f(indx)
+		| host1x_uclass_load_syncpt_base_value_f(threshold);
+}
+
+static inline u32 host1x_class_host_wait_syncpt_base(
+	unsigned indx, unsigned base_indx, unsigned offset)
+{
+	return host1x_uclass_wait_syncpt_base_indx_f(indx)
+		| host1x_uclass_wait_syncpt_base_base_indx_f(base_indx)
+		| host1x_uclass_wait_syncpt_base_offset_f(offset);
+}
+
+static inline u32 host1x_class_host_incr_syncpt_base(
+	unsigned base_indx, unsigned offset)
+{
+	return host1x_uclass_incr_syncpt_base_base_indx_f(base_indx)
+		| host1x_uclass_incr_syncpt_base_offset_f(offset);
+}
+
+static inline u32 host1x_class_host_incr_syncpt(
+	unsigned cond, unsigned indx)
+{
+	return host1x_uclass_incr_syncpt_cond_f(cond)
+		| host1x_uclass_incr_syncpt_indx_f(indx);
+}
+
+static inline u32 host1x_class_host_indoff_reg_write(
+	unsigned mod_id, unsigned offset, bool auto_inc)
+{
+	u32 v = host1x_uclass_indoff_indbe_f(0xf)
+		| host1x_uclass_indoff_indmodid_f(mod_id)
+		| host1x_uclass_indoff_indroffset_f(offset);
+	if (auto_inc)
+		v |= host1x_uclass_indoff_autoinc_f(1);
+	return v;
+}
+
+static inline u32 host1x_class_host_indoff_reg_read(
+	unsigned mod_id, unsigned offset, bool auto_inc)
+{
+	u32 v = host1x_uclass_indoff_indmodid_f(mod_id)
+		| host1x_uclass_indoff_indroffset_f(offset)
+		| host1x_uclass_indoff_rwn_read_v();
+	if (auto_inc)
+		v |= host1x_uclass_indoff_autoinc_f(1);
+	return v;
+}
+
+
+/* cdma opcodes */
+static inline u32 host1x_opcode_setclass(
+	unsigned class_id, unsigned offset, unsigned mask)
+{
+	return (0 << 28) | (offset << 16) | (class_id << 6) | mask;
+}
+
+static inline u32 host1x_opcode_incr(unsigned offset, unsigned count)
+{
+	return (1 << 28) | (offset << 16) | count;
+}
+
+static inline u32 host1x_opcode_nonincr(unsigned offset, unsigned count)
+{
+	return (2 << 28) | (offset << 16) | count;
+}
+
+static inline u32 host1x_opcode_mask(unsigned offset, unsigned mask)
+{
+	return (3 << 28) | (offset << 16) | mask;
+}
+
+static inline u32 host1x_opcode_imm(unsigned offset, unsigned value)
+{
+	return (4 << 28) | (offset << 16) | value;
+}
+
+static inline u32 host1x_opcode_imm_incr_syncpt(unsigned cond, unsigned indx)
+{
+	return host1x_opcode_imm(host1x_uclass_incr_syncpt_r(),
+		host1x_class_host_incr_syncpt(cond, indx));
+}
+
+static inline u32 host1x_opcode_restart(unsigned address)
+{
+	return (5 << 28) | (address >> 4);
+}
+
+static inline u32 host1x_opcode_gather(unsigned count)
+{
+	return (6 << 28) | count;
+}
+
+static inline u32 host1x_opcode_gather_nonincr(unsigned offset,	unsigned count)
+{
+	return (6 << 28) | (offset << 16) | BIT(15) | count;
+}
+
+static inline u32 host1x_opcode_gather_incr(unsigned offset, unsigned count)
+{
+	return (6 << 28) | (offset << 16) | BIT(15) | BIT(14) | count;
+}
+
+#define HOST1X_OPCODE_NOP host1x_opcode_nonincr(0, 0)
+
+#endif
diff --git a/drivers/gpu/host1x/hw/hw_host1x01_channel.h b/drivers/gpu/host1x/hw/hw_host1x01_channel.h
new file mode 100644
index 000000000000..b4bc7ca4e051
--- /dev/null
+++ b/drivers/gpu/host1x/hw/hw_host1x01_channel.h
@@ -0,0 +1,120 @@
+/*
+ * Copyright (c) 2012-2013, NVIDIA Corporation.
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms and conditions of the GNU General Public License,
+ * version 2, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program.  If not, see <http://www.gnu.org/licenses/>.
+ *
+ */
+
+ /*
+  * Function naming determines intended use:
+  *
+  *     <x>_r(void) : Returns the offset for register <x>.
+  *
+  *     <x>_w(void) : Returns the word offset for word (4 byte) element <x>.
+  *
+  *     <x>_<y>_s(void) : Returns size of field <y> of register <x> in bits.
+  *
+  *     <x>_<y>_f(u32 v) : Returns a value based on 'v' which has been shifted
+  *         and masked to place it at field <y> of register <x>.  This value
+  *         can be |'d with others to produce a full register value for
+  *         register <x>.
+  *
+  *     <x>_<y>_m(void) : Returns a mask for field <y> of register <x>.  This
+  *         value can be ~'d and then &'d to clear the value of field <y> for
+  *         register <x>.
+  *
+  *     <x>_<y>_<z>_f(void) : Returns the constant value <z> after being shifted
+  *         to place it at field <y> of register <x>.  This value can be |'d
+  *         with others to produce a full register value for <x>.
+  *
+  *     <x>_<y>_v(u32 r) : Returns the value of field <y> from a full register
+  *         <x> value 'r' after being shifted to place its LSB at bit 0.
+  *         This value is suitable for direct comparison with other unshifted
+  *         values appropriate for use in field <y> of register <x>.
+  *
+  *     <x>_<y>_<z>_v(void) : Returns the constant value for <z> defined for
+  *         field <y> of register <x>.  This value is suitable for direct
+  *         comparison with unshifted values appropriate for use in field <y>
+  *         of register <x>.
+  */
+
+#ifndef __hw_host1x_channel_host1x_h__
+#define __hw_host1x_channel_host1x_h__
+
+static inline u32 host1x_channel_fifostat_r(void)
+{
+	return 0x0;
+}
+#define HOST1X_CHANNEL_FIFOSTAT \
+	host1x_channel_fifostat_r()
+static inline u32 host1x_channel_fifostat_cfempty_v(u32 r)
+{
+	return (r >> 10) & 0x1;
+}
+#define HOST1X_CHANNEL_FIFOSTAT_CFEMPTY_V(r) \
+	host1x_channel_fifostat_cfempty_v(r)
+static inline u32 host1x_channel_dmastart_r(void)
+{
+	return 0x14;
+}
+#define HOST1X_CHANNEL_DMASTART \
+	host1x_channel_dmastart_r()
+static inline u32 host1x_channel_dmaput_r(void)
+{
+	return 0x18;
+}
+#define HOST1X_CHANNEL_DMAPUT \
+	host1x_channel_dmaput_r()
+static inline u32 host1x_channel_dmaget_r(void)
+{
+	return 0x1c;
+}
+#define HOST1X_CHANNEL_DMAGET \
+	host1x_channel_dmaget_r()
+static inline u32 host1x_channel_dmaend_r(void)
+{
+	return 0x20;
+}
+#define HOST1X_CHANNEL_DMAEND \
+	host1x_channel_dmaend_r()
+static inline u32 host1x_channel_dmactrl_r(void)
+{
+	return 0x24;
+}
+#define HOST1X_CHANNEL_DMACTRL \
+	host1x_channel_dmactrl_r()
+static inline u32 host1x_channel_dmactrl_dmastop(void)
+{
+	return 1 << 0;
+}
+#define HOST1X_CHANNEL_DMACTRL_DMASTOP \
+	host1x_channel_dmactrl_dmastop()
+static inline u32 host1x_channel_dmactrl_dmastop_v(u32 r)
+{
+	return (r >> 0) & 0x1;
+}
+#define HOST1X_CHANNEL_DMACTRL_DMASTOP_V(r) \
+	host1x_channel_dmactrl_dmastop_v(r)
+static inline u32 host1x_channel_dmactrl_dmagetrst(void)
+{
+	return 1 << 1;
+}
+#define HOST1X_CHANNEL_DMACTRL_DMAGETRST \
+	host1x_channel_dmactrl_dmagetrst()
+static inline u32 host1x_channel_dmactrl_dmainitget(void)
+{
+	return 1 << 2;
+}
+#define HOST1X_CHANNEL_DMACTRL_DMAINITGET \
+	host1x_channel_dmactrl_dmainitget()
+#endif
diff --git a/drivers/gpu/host1x/hw/hw_host1x01_sync.h b/drivers/gpu/host1x/hw/hw_host1x01_sync.h
new file mode 100644
index 000000000000..ac704e579977
--- /dev/null
+++ b/drivers/gpu/host1x/hw/hw_host1x01_sync.h
@@ -0,0 +1,243 @@
+/*
+ * Copyright (c) 2012-2013, NVIDIA Corporation.
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms and conditions of the GNU General Public License,
+ * version 2, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program.  If not, see <http://www.gnu.org/licenses/>.
+ *
+ */
+
+ /*
+  * Function naming determines intended use:
+  *
+  *     <x>_r(void) : Returns the offset for register <x>.
+  *
+  *     <x>_w(void) : Returns the word offset for word (4 byte) element <x>.
+  *
+  *     <x>_<y>_s(void) : Returns size of field <y> of register <x> in bits.
+  *
+  *     <x>_<y>_f(u32 v) : Returns a value based on 'v' which has been shifted
+  *         and masked to place it at field <y> of register <x>.  This value
+  *         can be |'d with others to produce a full register value for
+  *         register <x>.
+  *
+  *     <x>_<y>_m(void) : Returns a mask for field <y> of register <x>.  This
+  *         value can be ~'d and then &'d to clear the value of field <y> for
+  *         register <x>.
+  *
+  *     <x>_<y>_<z>_f(void) : Returns the constant value <z> after being shifted
+  *         to place it at field <y> of register <x>.  This value can be |'d
+  *         with others to produce a full register value for <x>.
+  *
+  *     <x>_<y>_v(u32 r) : Returns the value of field <y> from a full register
+  *         <x> value 'r' after being shifted to place its LSB at bit 0.
+  *         This value is suitable for direct comparison with other unshifted
+  *         values appropriate for use in field <y> of register <x>.
+  *
+  *     <x>_<y>_<z>_v(void) : Returns the constant value for <z> defined for
+  *         field <y> of register <x>.  This value is suitable for direct
+  *         comparison with unshifted values appropriate for use in field <y>
+  *         of register <x>.
+  */
+
+#ifndef __hw_host1x01_sync_h__
+#define __hw_host1x01_sync_h__
+
+#define REGISTER_STRIDE	4
+
+static inline u32 host1x_sync_syncpt_r(unsigned int id)
+{
+	return 0x400 + id * REGISTER_STRIDE;
+}
+#define HOST1X_SYNC_SYNCPT(id) \
+	host1x_sync_syncpt_r(id)
+static inline u32 host1x_sync_syncpt_thresh_cpu0_int_status_r(unsigned int id)
+{
+	return 0x40 + id * REGISTER_STRIDE;
+}
+#define HOST1X_SYNC_SYNCPT_THRESH_CPU0_INT_STATUS(id) \
+	host1x_sync_syncpt_thresh_cpu0_int_status_r(id)
+static inline u32 host1x_sync_syncpt_thresh_int_disable_r(unsigned int id)
+{
+	return 0x60 + id * REGISTER_STRIDE;
+}
+#define HOST1X_SYNC_SYNCPT_THRESH_INT_DISABLE(id) \
+	host1x_sync_syncpt_thresh_int_disable_r(id)
+static inline u32 host1x_sync_syncpt_thresh_int_enable_cpu0_r(unsigned int id)
+{
+	return 0x68 + id * REGISTER_STRIDE;
+}
+#define HOST1X_SYNC_SYNCPT_THRESH_INT_ENABLE_CPU0(id) \
+	host1x_sync_syncpt_thresh_int_enable_cpu0_r(id)
+static inline u32 host1x_sync_cf_setup_r(unsigned int channel)
+{
+	return 0x80 + channel * REGISTER_STRIDE;
+}
+#define HOST1X_SYNC_CF_SETUP(channel) \
+	host1x_sync_cf_setup_r(channel)
+static inline u32 host1x_sync_cf_setup_base_v(u32 r)
+{
+	return (r >> 0) & 0x1ff;
+}
+#define HOST1X_SYNC_CF_SETUP_BASE_V(r) \
+	host1x_sync_cf_setup_base_v(r)
+static inline u32 host1x_sync_cf_setup_limit_v(u32 r)
+{
+	return (r >> 16) & 0x1ff;
+}
+#define HOST1X_SYNC_CF_SETUP_LIMIT_V(r) \
+	host1x_sync_cf_setup_limit_v(r)
+static inline u32 host1x_sync_cmdproc_stop_r(void)
+{
+	return 0xac;
+}
+#define HOST1X_SYNC_CMDPROC_STOP \
+	host1x_sync_cmdproc_stop_r()
+static inline u32 host1x_sync_ch_teardown_r(void)
+{
+	return 0xb0;
+}
+#define HOST1X_SYNC_CH_TEARDOWN \
+	host1x_sync_ch_teardown_r()
+static inline u32 host1x_sync_usec_clk_r(void)
+{
+	return 0x1a4;
+}
+#define HOST1X_SYNC_USEC_CLK \
+	host1x_sync_usec_clk_r()
+static inline u32 host1x_sync_ctxsw_timeout_cfg_r(void)
+{
+	return 0x1a8;
+}
+#define HOST1X_SYNC_CTXSW_TIMEOUT_CFG \
+	host1x_sync_ctxsw_timeout_cfg_r()
+static inline u32 host1x_sync_ip_busy_timeout_r(void)
+{
+	return 0x1bc;
+}
+#define HOST1X_SYNC_IP_BUSY_TIMEOUT \
+	host1x_sync_ip_busy_timeout_r()
+static inline u32 host1x_sync_mlock_owner_r(unsigned int id)
+{
+	return 0x340 + id * REGISTER_STRIDE;
+}
+#define HOST1X_SYNC_MLOCK_OWNER(id) \
+	host1x_sync_mlock_owner_r(id)
+static inline u32 host1x_sync_mlock_owner_chid_f(u32 v)
+{
+	return (v & 0xf) << 8;
+}
+#define HOST1X_SYNC_MLOCK_OWNER_CHID_F(v) \
+	host1x_sync_mlock_owner_chid_f(v)
+static inline u32 host1x_sync_mlock_owner_cpu_owns_v(u32 r)
+{
+	return (r >> 1) & 0x1;
+}
+#define HOST1X_SYNC_MLOCK_OWNER_CPU_OWNS_V(r) \
+	host1x_sync_mlock_owner_cpu_owns_v(r)
+static inline u32 host1x_sync_mlock_owner_ch_owns_v(u32 r)
+{
+	return (r >> 0) & 0x1;
+}
+#define HOST1X_SYNC_MLOCK_OWNER_CH_OWNS_V(r) \
+	host1x_sync_mlock_owner_ch_owns_v(r)
+static inline u32 host1x_sync_syncpt_int_thresh_r(unsigned int id)
+{
+	return 0x500 + id * REGISTER_STRIDE;
+}
+#define HOST1X_SYNC_SYNCPT_INT_THRESH(id) \
+	host1x_sync_syncpt_int_thresh_r(id)
+static inline u32 host1x_sync_syncpt_base_r(unsigned int id)
+{
+	return 0x600 + id * REGISTER_STRIDE;
+}
+#define HOST1X_SYNC_SYNCPT_BASE(id) \
+	host1x_sync_syncpt_base_r(id)
+static inline u32 host1x_sync_syncpt_cpu_incr_r(unsigned int id)
+{
+	return 0x700 + id * REGISTER_STRIDE;
+}
+#define HOST1X_SYNC_SYNCPT_CPU_INCR(id) \
+	host1x_sync_syncpt_cpu_incr_r(id)
+static inline u32 host1x_sync_cbread_r(unsigned int channel)
+{
+	return 0x720 + channel * REGISTER_STRIDE;
+}
+#define HOST1X_SYNC_CBREAD(channel) \
+	host1x_sync_cbread_r(channel)
+static inline u32 host1x_sync_cfpeek_ctrl_r(void)
+{
+	return 0x74c;
+}
+#define HOST1X_SYNC_CFPEEK_CTRL \
+	host1x_sync_cfpeek_ctrl_r()
+static inline u32 host1x_sync_cfpeek_ctrl_addr_f(u32 v)
+{
+	return (v & 0x1ff) << 0;
+}
+#define HOST1X_SYNC_CFPEEK_CTRL_ADDR_F(v) \
+	host1x_sync_cfpeek_ctrl_addr_f(v)
+static inline u32 host1x_sync_cfpeek_ctrl_channr_f(u32 v)
+{
+	return (v & 0x7) << 16;
+}
+#define HOST1X_SYNC_CFPEEK_CTRL_CHANNR_F(v) \
+	host1x_sync_cfpeek_ctrl_channr_f(v)
+static inline u32 host1x_sync_cfpeek_ctrl_ena_f(u32 v)
+{
+	return (v & 0x1) << 31;
+}
+#define HOST1X_SYNC_CFPEEK_CTRL_ENA_F(v) \
+	host1x_sync_cfpeek_ctrl_ena_f(v)
+static inline u32 host1x_sync_cfpeek_read_r(void)
+{
+	return 0x750;
+}
+#define HOST1X_SYNC_CFPEEK_READ \
+	host1x_sync_cfpeek_read_r()
+static inline u32 host1x_sync_cfpeek_ptrs_r(void)
+{
+	return 0x754;
+}
+#define HOST1X_SYNC_CFPEEK_PTRS \
+	host1x_sync_cfpeek_ptrs_r()
+static inline u32 host1x_sync_cfpeek_ptrs_cf_rd_ptr_v(u32 r)
+{
+	return (r >> 0) & 0x1ff;
+}
+#define HOST1X_SYNC_CFPEEK_PTRS_CF_RD_PTR_V(r) \
+	host1x_sync_cfpeek_ptrs_cf_rd_ptr_v(r)
+static inline u32 host1x_sync_cfpeek_ptrs_cf_wr_ptr_v(u32 r)
+{
+	return (r >> 16) & 0x1ff;
+}
+#define HOST1X_SYNC_CFPEEK_PTRS_CF_WR_PTR_V(r) \
+	host1x_sync_cfpeek_ptrs_cf_wr_ptr_v(r)
+static inline u32 host1x_sync_cbstat_r(unsigned int channel)
+{
+	return 0x758 + channel * REGISTER_STRIDE;
+}
+#define HOST1X_SYNC_CBSTAT(channel) \
+	host1x_sync_cbstat_r(channel)
+static inline u32 host1x_sync_cbstat_cboffset_v(u32 r)
+{
+	return (r >> 0) & 0xffff;
+}
+#define HOST1X_SYNC_CBSTAT_CBOFFSET_V(r) \
+	host1x_sync_cbstat_cboffset_v(r)
+static inline u32 host1x_sync_cbstat_cbclass_v(u32 r)
+{
+	return (r >> 16) & 0x3ff;
+}
+#define HOST1X_SYNC_CBSTAT_CBCLASS_V(r) \
+	host1x_sync_cbstat_cbclass_v(r)
+
+#endif /* __hw_host1x01_sync_h__ */
diff --git a/drivers/gpu/host1x/hw/hw_host1x01_uclass.h b/drivers/gpu/host1x/hw/hw_host1x01_uclass.h
new file mode 100644
index 000000000000..42f3ce19ca32
--- /dev/null
+++ b/drivers/gpu/host1x/hw/hw_host1x01_uclass.h
@@ -0,0 +1,174 @@
+/*
+ * Copyright (c) 2012-2013, NVIDIA Corporation.
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms and conditions of the GNU General Public License,
+ * version 2, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program.  If not, see <http://www.gnu.org/licenses/>.
+ *
+ */
+
+ /*
+  * Function naming determines intended use:
+  *
+  *     <x>_r(void) : Returns the offset for register <x>.
+  *
+  *     <x>_w(void) : Returns the word offset for word (4 byte) element <x>.
+  *
+  *     <x>_<y>_s(void) : Returns size of field <y> of register <x> in bits.
+  *
+  *     <x>_<y>_f(u32 v) : Returns a value based on 'v' which has been shifted
+  *         and masked to place it at field <y> of register <x>.  This value
+  *         can be |'d with others to produce a full register value for
+  *         register <x>.
+  *
+  *     <x>_<y>_m(void) : Returns a mask for field <y> of register <x>.  This
+  *         value can be ~'d and then &'d to clear the value of field <y> for
+  *         register <x>.
+  *
+  *     <x>_<y>_<z>_f(void) : Returns the constant value <z> after being shifted
+  *         to place it at field <y> of register <x>.  This value can be |'d
+  *         with others to produce a full register value for <x>.
+  *
+  *     <x>_<y>_v(u32 r) : Returns the value of field <y> from a full register
+  *         <x> value 'r' after being shifted to place its LSB at bit 0.
+  *         This value is suitable for direct comparison with other unshifted
+  *         values appropriate for use in field <y> of register <x>.
+  *
+  *     <x>_<y>_<z>_v(void) : Returns the constant value for <z> defined for
+  *         field <y> of register <x>.  This value is suitable for direct
+  *         comparison with unshifted values appropriate for use in field <y>
+  *         of register <x>.
+  */
+
+#ifndef __hw_host1x_uclass_host1x_h__
+#define __hw_host1x_uclass_host1x_h__
+
+static inline u32 host1x_uclass_incr_syncpt_r(void)
+{
+	return 0x0;
+}
+#define HOST1X_UCLASS_INCR_SYNCPT \
+	host1x_uclass_incr_syncpt_r()
+static inline u32 host1x_uclass_incr_syncpt_cond_f(u32 v)
+{
+	return (v & 0xff) << 8;
+}
+#define HOST1X_UCLASS_INCR_SYNCPT_COND_F(v) \
+	host1x_uclass_incr_syncpt_cond_f(v)
+static inline u32 host1x_uclass_incr_syncpt_indx_f(u32 v)
+{
+	return (v & 0xff) << 0;
+}
+#define HOST1X_UCLASS_INCR_SYNCPT_INDX_F(v) \
+	host1x_uclass_incr_syncpt_indx_f(v)
+static inline u32 host1x_uclass_wait_syncpt_r(void)
+{
+	return 0x8;
+}
+#define HOST1X_UCLASS_WAIT_SYNCPT \
+	host1x_uclass_wait_syncpt_r()
+static inline u32 host1x_uclass_wait_syncpt_indx_f(u32 v)
+{
+	return (v & 0xff) << 24;
+}
+#define HOST1X_UCLASS_WAIT_SYNCPT_INDX_F(v) \
+	host1x_uclass_wait_syncpt_indx_f(v)
+static inline u32 host1x_uclass_wait_syncpt_thresh_f(u32 v)
+{
+	return (v & 0xffffff) << 0;
+}
+#define HOST1X_UCLASS_WAIT_SYNCPT_THRESH_F(v) \
+	host1x_uclass_wait_syncpt_thresh_f(v)
+static inline u32 host1x_uclass_wait_syncpt_base_r(void)
+{
+	return 0x9;
+}
+#define HOST1X_UCLASS_WAIT_SYNCPT_BASE \
+	host1x_uclass_wait_syncpt_base_r()
+static inline u32 host1x_uclass_wait_syncpt_base_indx_f(u32 v)
+{
+	return (v & 0xff) << 24;
+}
+#define HOST1X_UCLASS_WAIT_SYNCPT_BASE_INDX_F(v) \
+	host1x_uclass_wait_syncpt_base_indx_f(v)
+static inline u32 host1x_uclass_wait_syncpt_base_base_indx_f(u32 v)
+{
+	return (v & 0xff) << 16;
+}
+#define HOST1X_UCLASS_WAIT_SYNCPT_BASE_BASE_INDX_F(v) \
+	host1x_uclass_wait_syncpt_base_base_indx_f(v)
+static inline u32 host1x_uclass_wait_syncpt_base_offset_f(u32 v)
+{
+	return (v & 0xffff) << 0;
+}
+#define HOST1X_UCLASS_WAIT_SYNCPT_BASE_OFFSET_F(v) \
+	host1x_uclass_wait_syncpt_base_offset_f(v)
+static inline u32 host1x_uclass_load_syncpt_base_base_indx_f(u32 v)
+{
+	return (v & 0xff) << 24;
+}
+#define HOST1X_UCLASS_LOAD_SYNCPT_BASE_BASE_INDX_F(v) \
+	host1x_uclass_load_syncpt_base_base_indx_f(v)
+static inline u32 host1x_uclass_load_syncpt_base_value_f(u32 v)
+{
+	return (v & 0xffffff) << 0;
+}
+#define HOST1X_UCLASS_LOAD_SYNCPT_BASE_VALUE_F(v) \
+	host1x_uclass_load_syncpt_base_value_f(v)
+static inline u32 host1x_uclass_incr_syncpt_base_base_indx_f(u32 v)
+{
+	return (v & 0xff) << 24;
+}
+#define HOST1X_UCLASS_INCR_SYNCPT_BASE_BASE_INDX_F(v) \
+	host1x_uclass_incr_syncpt_base_base_indx_f(v)
+static inline u32 host1x_uclass_incr_syncpt_base_offset_f(u32 v)
+{
+	return (v & 0xffffff) << 0;
+}
+#define HOST1X_UCLASS_INCR_SYNCPT_BASE_OFFSET_F(v) \
+	host1x_uclass_incr_syncpt_base_offset_f(v)
+static inline u32 host1x_uclass_indoff_r(void)
+{
+	return 0x2d;
+}
+#define HOST1X_UCLASS_INDOFF \
+	host1x_uclass_indoff_r()
+static inline u32 host1x_uclass_indoff_indbe_f(u32 v)
+{
+	return (v & 0xf) << 28;
+}
+#define HOST1X_UCLASS_INDOFF_INDBE_F(v) \
+	host1x_uclass_indoff_indbe_f(v)
+static inline u32 host1x_uclass_indoff_autoinc_f(u32 v)
+{
+	return (v & 0x1) << 27;
+}
+#define HOST1X_UCLASS_INDOFF_AUTOINC_F(v) \
+	host1x_uclass_indoff_autoinc_f(v)
+static inline u32 host1x_uclass_indoff_indmodid_f(u32 v)
+{
+	return (v & 0xff) << 18;
+}
+#define HOST1X_UCLASS_INDOFF_INDMODID_F(v) \
+	host1x_uclass_indoff_indmodid_f(v)
+static inline u32 host1x_uclass_indoff_indroffset_f(u32 v)
+{
+	return (v & 0xffff) << 2;
+}
+#define HOST1X_UCLASS_INDOFF_INDROFFSET_F(v) \
+	host1x_uclass_indoff_indroffset_f(v)
+static inline u32 host1x_uclass_indoff_rwn_read_v(void)
+{
+	return 1;
+}
+#define HOST1X_UCLASS_INDOFF_INDROFFSET_F(v) \
+	host1x_uclass_indoff_indroffset_f(v)
+#endif
diff --git a/drivers/gpu/host1x/hw/intr_hw.c b/drivers/gpu/host1x/hw/intr_hw.c
new file mode 100644
index 000000000000..b592eef1efcb
--- /dev/null
+++ b/drivers/gpu/host1x/hw/intr_hw.c
@@ -0,0 +1,143 @@
+/*
+ * Tegra host1x Interrupt Management
+ *
+ * Copyright (C) 2010 Google, Inc.
+ * Copyright (c) 2010-2013, NVIDIA Corporation.
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms and conditions of the GNU General Public License,
+ * version 2, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program.  If not, see <http://www.gnu.org/licenses/>.
+ */
+
+#include <linux/interrupt.h>
+#include <linux/irq.h>
+#include <linux/io.h>
+#include <asm/mach/irq.h>
+
+#include "intr.h"
+#include "dev.h"
+
+/*
+ * Sync point threshold interrupt service function
+ * Handles sync point threshold triggers, in interrupt context
+ */
+static void host1x_intr_syncpt_handle(struct host1x_syncpt *syncpt)
+{
+	unsigned int id = syncpt->id;
+	struct host1x *host = syncpt->host;
+
+	host1x_sync_writel(host, BIT_MASK(id),
+		HOST1X_SYNC_SYNCPT_THRESH_INT_DISABLE(BIT_WORD(id)));
+	host1x_sync_writel(host, BIT_MASK(id),
+		HOST1X_SYNC_SYNCPT_THRESH_CPU0_INT_STATUS(BIT_WORD(id)));
+
+	queue_work(host->intr_wq, &syncpt->intr.work);
+}
+
+static irqreturn_t syncpt_thresh_isr(int irq, void *dev_id)
+{
+	struct host1x *host = dev_id;
+	unsigned long reg;
+	int i, id;
+
+	for (i = 0; i <= BIT_WORD(host->info->nb_pts); i++) {
+		reg = host1x_sync_readl(host,
+			HOST1X_SYNC_SYNCPT_THRESH_CPU0_INT_STATUS(i));
+		for_each_set_bit(id, &reg, BITS_PER_LONG) {
+			struct host1x_syncpt *syncpt =
+				host->syncpt + (i * BITS_PER_LONG + id);
+			host1x_intr_syncpt_handle(syncpt);
+		}
+	}
+
+	return IRQ_HANDLED;
+}
+
+static void _host1x_intr_disable_all_syncpt_intrs(struct host1x *host)
+{
+	u32 i;
+
+	for (i = 0; i <= BIT_WORD(host->info->nb_pts); ++i) {
+		host1x_sync_writel(host, 0xffffffffu,
+			HOST1X_SYNC_SYNCPT_THRESH_INT_DISABLE(i));
+		host1x_sync_writel(host, 0xffffffffu,
+			HOST1X_SYNC_SYNCPT_THRESH_CPU0_INT_STATUS(i));
+	}
+}
+
+static int _host1x_intr_init_host_sync(struct host1x *host, u32 cpm,
+	void (*syncpt_thresh_work)(struct work_struct *))
+{
+	int i, err;
+
+	host1x_hw_intr_disable_all_syncpt_intrs(host);
+
+	for (i = 0; i < host->info->nb_pts; i++)
+		INIT_WORK(&host->syncpt[i].intr.work, syncpt_thresh_work);
+
+	err = devm_request_irq(host->dev, host->intr_syncpt_irq,
+			       syncpt_thresh_isr, IRQF_SHARED,
+			       "host1x_syncpt", host);
+	if (IS_ERR_VALUE(err)) {
+		WARN_ON(1);
+		return err;
+	}
+
+	/* disable the ip_busy_timeout. this prevents write drops */
+	host1x_sync_writel(host, 0, HOST1X_SYNC_IP_BUSY_TIMEOUT);
+
+	/*
+	 * increase the auto-ack timout to the maximum value. 2d will hang
+	 * otherwise on Tegra2.
+	 */
+	host1x_sync_writel(host, 0xff, HOST1X_SYNC_CTXSW_TIMEOUT_CFG);
+
+	/* update host clocks per usec */
+	host1x_sync_writel(host, cpm, HOST1X_SYNC_USEC_CLK);
+
+	return 0;
+}
+
+static void _host1x_intr_set_syncpt_threshold(struct host1x *host,
+	u32 id, u32 thresh)
+{
+	host1x_sync_writel(host, thresh, HOST1X_SYNC_SYNCPT_INT_THRESH(id));
+}
+
+static void _host1x_intr_enable_syncpt_intr(struct host1x *host, u32 id)
+{
+	host1x_sync_writel(host, BIT_MASK(id),
+		HOST1X_SYNC_SYNCPT_THRESH_INT_ENABLE_CPU0(BIT_WORD(id)));
+}
+
+static void _host1x_intr_disable_syncpt_intr(struct host1x *host, u32 id)
+{
+	host1x_sync_writel(host, BIT_MASK(id),
+		HOST1X_SYNC_SYNCPT_THRESH_INT_DISABLE(BIT_WORD(id)));
+	host1x_sync_writel(host, BIT_MASK(id),
+		HOST1X_SYNC_SYNCPT_THRESH_CPU0_INT_STATUS(BIT_WORD(id)));
+}
+
+static int _host1x_free_syncpt_irq(struct host1x *host)
+{
+	devm_free_irq(host->dev, host->intr_syncpt_irq, host);
+	flush_workqueue(host->intr_wq);
+	return 0;
+}
+
+static const struct host1x_intr_ops host1x_intr_ops = {
+	.init_host_sync = _host1x_intr_init_host_sync,
+	.set_syncpt_threshold = _host1x_intr_set_syncpt_threshold,
+	.enable_syncpt_intr = _host1x_intr_enable_syncpt_intr,
+	.disable_syncpt_intr = _host1x_intr_disable_syncpt_intr,
+	.disable_all_syncpt_intrs = _host1x_intr_disable_all_syncpt_intrs,
+	.free_syncpt_irq = _host1x_free_syncpt_irq,
+};
diff --git a/drivers/gpu/host1x/hw/syncpt_hw.c b/drivers/gpu/host1x/hw/syncpt_hw.c
new file mode 100644
index 000000000000..61174990102a
--- /dev/null
+++ b/drivers/gpu/host1x/hw/syncpt_hw.c
@@ -0,0 +1,114 @@
+/*
+ * Tegra host1x Syncpoints
+ *
+ * Copyright (c) 2010-2013, NVIDIA Corporation.
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms and conditions of the GNU General Public License,
+ * version 2, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program.  If not, see <http://www.gnu.org/licenses/>.
+ */
+
+#include <linux/io.h>
+
+#include "dev.h"
+#include "syncpt.h"
+
+/*
+ * Write the current syncpoint value back to hw.
+ */
+static void syncpt_restore(struct host1x_syncpt *sp)
+{
+	struct host1x *host = sp->host;
+	int min = host1x_syncpt_read_min(sp);
+	host1x_sync_writel(host, min, HOST1X_SYNC_SYNCPT(sp->id));
+}
+
+/*
+ * Write the current waitbase value back to hw.
+ */
+static void syncpt_restore_wait_base(struct host1x_syncpt *sp)
+{
+	struct host1x *host = sp->host;
+	host1x_sync_writel(host, sp->base_val,
+			   HOST1X_SYNC_SYNCPT_BASE(sp->id));
+}
+
+/*
+ * Read waitbase value from hw.
+ */
+static void syncpt_read_wait_base(struct host1x_syncpt *sp)
+{
+	struct host1x *host = sp->host;
+	sp->base_val =
+		host1x_sync_readl(host, HOST1X_SYNC_SYNCPT_BASE(sp->id));
+}
+
+/*
+ * Updates the last value read from hardware.
+ */
+static u32 syncpt_load(struct host1x_syncpt *sp)
+{
+	struct host1x *host = sp->host;
+	u32 old, live;
+
+	/* Loop in case there's a race writing to min_val */
+	do {
+		old = host1x_syncpt_read_min(sp);
+		live = host1x_sync_readl(host, HOST1X_SYNC_SYNCPT(sp->id));
+	} while ((u32)atomic_cmpxchg(&sp->min_val, old, live) != old);
+
+	if (!host1x_syncpt_check_max(sp, live))
+		dev_err(host->dev, "%s failed: id=%u, min=%d, max=%d\n",
+			__func__, sp->id, host1x_syncpt_read_min(sp),
+			host1x_syncpt_read_max(sp));
+
+	return live;
+}
+
+/*
+ * Write a cpu syncpoint increment to the hardware, without touching
+ * the cache.
+ */
+static void syncpt_cpu_incr(struct host1x_syncpt *sp)
+{
+	struct host1x *host = sp->host;
+	u32 reg_offset = sp->id / 32;
+
+	if (!host1x_syncpt_client_managed(sp) &&
+	    host1x_syncpt_idle(sp)) {
+		dev_err(host->dev, "Trying to increment syncpoint id %d beyond max\n",
+			sp->id);
+		host1x_debug_dump(sp->host);
+		return;
+	}
+	host1x_sync_writel(host, BIT_MASK(sp->id),
+			   HOST1X_SYNC_SYNCPT_CPU_INCR(reg_offset));
+	wmb();
+}
+
+/* remove a wait pointed to by patch_addr */
+static int syncpt_patch_wait(struct host1x_syncpt *sp, void *patch_addr)
+{
+	u32 override = host1x_class_host_wait_syncpt(
+		HOST1X_SYNCPT_RESERVED, 0);
+
+	*((u32 *)patch_addr) = override;
+	return 0;
+}
+
+static const struct host1x_syncpt_ops host1x_syncpt_ops = {
+	.restore = syncpt_restore,
+	.restore_wait_base = syncpt_restore_wait_base,
+	.load_wait_base = syncpt_read_wait_base,
+	.load = syncpt_load,
+	.cpu_incr = syncpt_cpu_incr,
+	.patch_wait = syncpt_patch_wait,
+};
diff --git a/drivers/gpu/host1x/intr.c b/drivers/gpu/host1x/intr.c
new file mode 100644
index 000000000000..2491bf82e30c
--- /dev/null
+++ b/drivers/gpu/host1x/intr.c
@@ -0,0 +1,354 @@
+/*
+ * Tegra host1x Interrupt Management
+ *
+ * Copyright (c) 2010-2013, NVIDIA Corporation.
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms and conditions of the GNU General Public License,
+ * version 2, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program.  If not, see <http://www.gnu.org/licenses/>.
+ */
+
+#include <linux/clk.h>
+#include <linux/interrupt.h>
+#include <linux/slab.h>
+#include <linux/irq.h>
+
+#include <trace/events/host1x.h>
+#include "channel.h"
+#include "dev.h"
+#include "intr.h"
+
+/* Wait list management */
+
+enum waitlist_state {
+	WLS_PENDING,
+	WLS_REMOVED,
+	WLS_CANCELLED,
+	WLS_HANDLED
+};
+
+static void waiter_release(struct kref *kref)
+{
+	kfree(container_of(kref, struct host1x_waitlist, refcount));
+}
+
+/*
+ * add a waiter to a waiter queue, sorted by threshold
+ * returns true if it was added at the head of the queue
+ */
+static bool add_waiter_to_queue(struct host1x_waitlist *waiter,
+				struct list_head *queue)
+{
+	struct host1x_waitlist *pos;
+	u32 thresh = waiter->thresh;
+
+	list_for_each_entry_reverse(pos, queue, list)
+		if ((s32)(pos->thresh - thresh) <= 0) {
+			list_add(&waiter->list, &pos->list);
+			return false;
+		}
+
+	list_add(&waiter->list, queue);
+	return true;
+}
+
+/*
+ * run through a waiter queue for a single sync point ID
+ * and gather all completed waiters into lists by actions
+ */
+static void remove_completed_waiters(struct list_head *head, u32 sync,
+			struct list_head completed[HOST1X_INTR_ACTION_COUNT])
+{
+	struct list_head *dest;
+	struct host1x_waitlist *waiter, *next, *prev;
+
+	list_for_each_entry_safe(waiter, next, head, list) {
+		if ((s32)(waiter->thresh - sync) > 0)
+			break;
+
+		dest = completed + waiter->action;
+
+		/* consolidate submit cleanups */
+		if (waiter->action == HOST1X_INTR_ACTION_SUBMIT_COMPLETE &&
+		    !list_empty(dest)) {
+			prev = list_entry(dest->prev,
+					  struct host1x_waitlist, list);
+			if (prev->data == waiter->data) {
+				prev->count++;
+				dest = NULL;
+			}
+		}
+
+		/* PENDING->REMOVED or CANCELLED->HANDLED */
+		if (atomic_inc_return(&waiter->state) == WLS_HANDLED || !dest) {
+			list_del(&waiter->list);
+			kref_put(&waiter->refcount, waiter_release);
+		} else
+			list_move_tail(&waiter->list, dest);
+	}
+}
+
+static void reset_threshold_interrupt(struct host1x *host,
+				      struct list_head *head,
+				      unsigned int id)
+{
+	u32 thresh =
+		list_first_entry(head, struct host1x_waitlist, list)->thresh;
+
+	host1x_hw_intr_set_syncpt_threshold(host, id, thresh);
+	host1x_hw_intr_enable_syncpt_intr(host, id);
+}
+
+static void action_submit_complete(struct host1x_waitlist *waiter)
+{
+	struct host1x_channel *channel = waiter->data;
+
+	host1x_cdma_update(&channel->cdma);
+
+	/*  Add nr_completed to trace */
+	trace_host1x_channel_submit_complete(dev_name(channel->dev),
+					     waiter->count, waiter->thresh);
+
+}
+
+static void action_wakeup(struct host1x_waitlist *waiter)
+{
+	wait_queue_head_t *wq = waiter->data;
+	wake_up(wq);
+}
+
+static void action_wakeup_interruptible(struct host1x_waitlist *waiter)
+{
+	wait_queue_head_t *wq = waiter->data;
+	wake_up_interruptible(wq);
+}
+
+typedef void (*action_handler)(struct host1x_waitlist *waiter);
+
+static action_handler action_handlers[HOST1X_INTR_ACTION_COUNT] = {
+	action_submit_complete,
+	action_wakeup,
+	action_wakeup_interruptible,
+};
+
+static void run_handlers(struct list_head completed[HOST1X_INTR_ACTION_COUNT])
+{
+	struct list_head *head = completed;
+	int i;
+
+	for (i = 0; i < HOST1X_INTR_ACTION_COUNT; ++i, ++head) {
+		action_handler handler = action_handlers[i];
+		struct host1x_waitlist *waiter, *next;
+
+		list_for_each_entry_safe(waiter, next, head, list) {
+			list_del(&waiter->list);
+			handler(waiter);
+			WARN_ON(atomic_xchg(&waiter->state, WLS_HANDLED) !=
+				WLS_REMOVED);
+			kref_put(&waiter->refcount, waiter_release);
+		}
+	}
+}
+
+/*
+ * Remove & handle all waiters that have completed for the given syncpt
+ */
+static int process_wait_list(struct host1x *host,
+			     struct host1x_syncpt *syncpt,
+			     u32 threshold)
+{
+	struct list_head completed[HOST1X_INTR_ACTION_COUNT];
+	unsigned int i;
+	int empty;
+
+	for (i = 0; i < HOST1X_INTR_ACTION_COUNT; ++i)
+		INIT_LIST_HEAD(completed + i);
+
+	spin_lock(&syncpt->intr.lock);
+
+	remove_completed_waiters(&syncpt->intr.wait_head, threshold,
+				 completed);
+
+	empty = list_empty(&syncpt->intr.wait_head);
+	if (empty)
+		host1x_hw_intr_disable_syncpt_intr(host, syncpt->id);
+	else
+		reset_threshold_interrupt(host, &syncpt->intr.wait_head,
+					  syncpt->id);
+
+	spin_unlock(&syncpt->intr.lock);
+
+	run_handlers(completed);
+
+	return empty;
+}
+
+/*
+ * Sync point threshold interrupt service thread function
+ * Handles sync point threshold triggers, in thread context
+ */
+
+static void syncpt_thresh_work(struct work_struct *work)
+{
+	struct host1x_syncpt_intr *syncpt_intr =
+		container_of(work, struct host1x_syncpt_intr, work);
+	struct host1x_syncpt *syncpt =
+		container_of(syncpt_intr, struct host1x_syncpt, intr);
+	unsigned int id = syncpt->id;
+	struct host1x *host = syncpt->host;
+
+	(void)process_wait_list(host, syncpt,
+				host1x_syncpt_load(host->syncpt + id));
+}
+
+int host1x_intr_add_action(struct host1x *host, u32 id, u32 thresh,
+			   enum host1x_intr_action action, void *data,
+			   struct host1x_waitlist *waiter, void **ref)
+{
+	struct host1x_syncpt *syncpt;
+	int queue_was_empty;
+
+	if (waiter == NULL) {
+		pr_warn("%s: NULL waiter\n", __func__);
+		return -EINVAL;
+	}
+
+	/* initialize a new waiter */
+	INIT_LIST_HEAD(&waiter->list);
+	kref_init(&waiter->refcount);
+	if (ref)
+		kref_get(&waiter->refcount);
+	waiter->thresh = thresh;
+	waiter->action = action;
+	atomic_set(&waiter->state, WLS_PENDING);
+	waiter->data = data;
+	waiter->count = 1;
+
+	syncpt = host->syncpt + id;
+
+	spin_lock(&syncpt->intr.lock);
+
+	queue_was_empty = list_empty(&syncpt->intr.wait_head);
+
+	if (add_waiter_to_queue(waiter, &syncpt->intr.wait_head)) {
+		/* added at head of list - new threshold value */
+		host1x_hw_intr_set_syncpt_threshold(host, id, thresh);
+
+		/* added as first waiter - enable interrupt */
+		if (queue_was_empty)
+			host1x_hw_intr_enable_syncpt_intr(host, id);
+	}
+
+	spin_unlock(&syncpt->intr.lock);
+
+	if (ref)
+		*ref = waiter;
+	return 0;
+}
+
+void host1x_intr_put_ref(struct host1x *host, u32 id, void *ref)
+{
+	struct host1x_waitlist *waiter = ref;
+	struct host1x_syncpt *syncpt;
+
+	while (atomic_cmpxchg(&waiter->state, WLS_PENDING, WLS_CANCELLED) ==
+	       WLS_REMOVED)
+		schedule();
+
+	syncpt = host->syncpt + id;
+	(void)process_wait_list(host, syncpt,
+				host1x_syncpt_load(host->syncpt + id));
+
+	kref_put(&waiter->refcount, waiter_release);
+}
+
+int host1x_intr_init(struct host1x *host, unsigned int irq_sync)
+{
+	unsigned int id;
+	u32 nb_pts = host1x_syncpt_nb_pts(host);
+
+	mutex_init(&host->intr_mutex);
+	host->intr_syncpt_irq = irq_sync;
+	host->intr_wq = create_workqueue("host_syncpt");
+	if (!host->intr_wq)
+		return -ENOMEM;
+
+	for (id = 0; id < nb_pts; ++id) {
+		struct host1x_syncpt *syncpt = host->syncpt + id;
+
+		spin_lock_init(&syncpt->intr.lock);
+		INIT_LIST_HEAD(&syncpt->intr.wait_head);
+		snprintf(syncpt->intr.thresh_irq_name,
+			 sizeof(syncpt->intr.thresh_irq_name),
+			 "host1x_sp_%02d", id);
+	}
+
+	host1x_intr_start(host);
+
+	return 0;
+}
+
+void host1x_intr_deinit(struct host1x *host)
+{
+	host1x_intr_stop(host);
+	destroy_workqueue(host->intr_wq);
+}
+
+void host1x_intr_start(struct host1x *host)
+{
+	u32 hz = clk_get_rate(host->clk);
+	int err;
+
+	mutex_lock(&host->intr_mutex);
+	err = host1x_hw_intr_init_host_sync(host, DIV_ROUND_UP(hz, 1000000),
+					    syncpt_thresh_work);
+	if (err) {
+		mutex_unlock(&host->intr_mutex);
+		return;
+	}
+	mutex_unlock(&host->intr_mutex);
+}
+
+void host1x_intr_stop(struct host1x *host)
+{
+	unsigned int id;
+	struct host1x_syncpt *syncpt = host->syncpt;
+	u32 nb_pts = host1x_syncpt_nb_pts(host);
+
+	mutex_lock(&host->intr_mutex);
+
+	host1x_hw_intr_disable_all_syncpt_intrs(host);
+
+	for (id = 0; id < nb_pts; ++id) {
+		struct host1x_waitlist *waiter, *next;
+
+		list_for_each_entry_safe(waiter, next,
+			&syncpt[id].intr.wait_head, list) {
+			if (atomic_cmpxchg(&waiter->state,
+			    WLS_CANCELLED, WLS_HANDLED) == WLS_CANCELLED) {
+				list_del(&waiter->list);
+				kref_put(&waiter->refcount, waiter_release);
+			}
+		}
+
+		if (!list_empty(&syncpt[id].intr.wait_head)) {
+			/* output diagnostics */
+			mutex_unlock(&host->intr_mutex);
+			pr_warn("%s cannot stop syncpt intr id=%d\n",
+				__func__, id);
+			return;
+		}
+	}
+
+	host1x_hw_intr_free_syncpt_irq(host);
+
+	mutex_unlock(&host->intr_mutex);
+}
diff --git a/drivers/gpu/host1x/intr.h b/drivers/gpu/host1x/intr.h
new file mode 100644
index 000000000000..2b8adf016a05
--- /dev/null
+++ b/drivers/gpu/host1x/intr.h
@@ -0,0 +1,102 @@
+/*
+ * Tegra host1x Interrupt Management
+ *
+ * Copyright (c) 2010-2013, NVIDIA Corporation.
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms and conditions of the GNU General Public License,
+ * version 2, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program.  If not, see <http://www.gnu.org/licenses/>.
+ */
+
+#ifndef __HOST1X_INTR_H
+#define __HOST1X_INTR_H
+
+#include <linux/interrupt.h>
+#include <linux/workqueue.h>
+
+struct host1x;
+
+enum host1x_intr_action {
+	/*
+	 * Perform cleanup after a submit has completed.
+	 * 'data' points to a channel
+	 */
+	HOST1X_INTR_ACTION_SUBMIT_COMPLETE = 0,
+
+	/*
+	 * Wake up a  task.
+	 * 'data' points to a wait_queue_head_t
+	 */
+	HOST1X_INTR_ACTION_WAKEUP,
+
+	/*
+	 * Wake up a interruptible task.
+	 * 'data' points to a wait_queue_head_t
+	 */
+	HOST1X_INTR_ACTION_WAKEUP_INTERRUPTIBLE,
+
+	HOST1X_INTR_ACTION_COUNT
+};
+
+struct host1x_syncpt_intr {
+	spinlock_t lock;
+	struct list_head wait_head;
+	char thresh_irq_name[12];
+	struct work_struct work;
+};
+
+struct host1x_waitlist {
+	struct list_head list;
+	struct kref refcount;
+	u32 thresh;
+	enum host1x_intr_action action;
+	atomic_t state;
+	void *data;
+	int count;
+};
+
+/*
+ * Schedule an action to be taken when a sync point reaches the given threshold.
+ *
+ * @id the sync point
+ * @thresh the threshold
+ * @action the action to take
+ * @data a pointer to extra data depending on action, see above
+ * @waiter waiter structure - assumes ownership
+ * @ref must be passed if cancellation is possible, else NULL
+ *
+ * This is a non-blocking api.
+ */
+int host1x_intr_add_action(struct host1x *host, u32 id, u32 thresh,
+	enum host1x_intr_action action, void *data,
+	struct host1x_waitlist *waiter, void **ref);
+
+/*
+ * Unreference an action submitted to host1x_intr_add_action().
+ * You must call this if you passed non-NULL as ref.
+ * @ref the ref returned from host1x_intr_add_action()
+ */
+void host1x_intr_put_ref(struct host1x *host, u32 id, void *ref);
+
+/* Initialize host1x sync point interrupt */
+int host1x_intr_init(struct host1x *host, unsigned int irq_sync);
+
+/* Deinitialize host1x sync point interrupt */
+void host1x_intr_deinit(struct host1x *host);
+
+/* Enable host1x sync point interrupt */
+void host1x_intr_start(struct host1x *host);
+
+/* Disable host1x sync point interrupt */
+void host1x_intr_stop(struct host1x *host);
+
+irqreturn_t host1x_syncpt_thresh_fn(void *dev_id);
+#endif
diff --git a/drivers/gpu/host1x/job.c b/drivers/gpu/host1x/job.c
new file mode 100644
index 000000000000..f665d679031c
--- /dev/null
+++ b/drivers/gpu/host1x/job.c
@@ -0,0 +1,603 @@
+/*
+ * Tegra host1x Job
+ *
+ * Copyright (c) 2010-2013, NVIDIA Corporation.
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms and conditions of the GNU General Public License,
+ * version 2, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program.  If not, see <http://www.gnu.org/licenses/>.
+ */
+
+#include <linux/dma-mapping.h>
+#include <linux/err.h>
+#include <linux/kref.h>
+#include <linux/module.h>
+#include <linux/scatterlist.h>
+#include <linux/slab.h>
+#include <linux/vmalloc.h>
+#include <trace/events/host1x.h>
+
+#include "channel.h"
+#include "dev.h"
+#include "host1x_bo.h"
+#include "job.h"
+#include "syncpt.h"
+
+struct host1x_job *host1x_job_alloc(struct host1x_channel *ch,
+				    u32 num_cmdbufs, u32 num_relocs,
+				    u32 num_waitchks)
+{
+	struct host1x_job *job = NULL;
+	unsigned int num_unpins = num_cmdbufs + num_relocs;
+	u64 total;
+	void *mem;
+
+	/* Check that we're not going to overflow */
+	total = sizeof(struct host1x_job) +
+		num_relocs * sizeof(struct host1x_reloc) +
+		num_unpins * sizeof(struct host1x_job_unpin_data) +
+		num_waitchks * sizeof(struct host1x_waitchk) +
+		num_cmdbufs * sizeof(struct host1x_job_gather) +
+		num_unpins * sizeof(dma_addr_t) +
+		num_unpins * sizeof(u32 *);
+	if (total > ULONG_MAX)
+		return NULL;
+
+	mem = job = kzalloc(total, GFP_KERNEL);
+	if (!job)
+		return NULL;
+
+	kref_init(&job->ref);
+	job->channel = ch;
+
+	/* Redistribute memory to the structs  */
+	mem += sizeof(struct host1x_job);
+	job->relocarray = num_relocs ? mem : NULL;
+	mem += num_relocs * sizeof(struct host1x_reloc);
+	job->unpins = num_unpins ? mem : NULL;
+	mem += num_unpins * sizeof(struct host1x_job_unpin_data);
+	job->waitchk = num_waitchks ? mem : NULL;
+	mem += num_waitchks * sizeof(struct host1x_waitchk);
+	job->gathers = num_cmdbufs ? mem : NULL;
+	mem += num_cmdbufs * sizeof(struct host1x_job_gather);
+	job->addr_phys = num_unpins ? mem : NULL;
+
+	job->reloc_addr_phys = job->addr_phys;
+	job->gather_addr_phys = &job->addr_phys[num_relocs];
+
+	return job;
+}
+
+struct host1x_job *host1x_job_get(struct host1x_job *job)
+{
+	kref_get(&job->ref);
+	return job;
+}
+
+static void job_free(struct kref *ref)
+{
+	struct host1x_job *job = container_of(ref, struct host1x_job, ref);
+
+	kfree(job);
+}
+
+void host1x_job_put(struct host1x_job *job)
+{
+	kref_put(&job->ref, job_free);
+}
+
+void host1x_job_add_gather(struct host1x_job *job, struct host1x_bo *bo,
+			   u32 words, u32 offset)
+{
+	struct host1x_job_gather *cur_gather = &job->gathers[job->num_gathers];
+
+	cur_gather->words = words;
+	cur_gather->bo = bo;
+	cur_gather->offset = offset;
+	job->num_gathers++;
+}
+
+/*
+ * NULL an already satisfied WAIT_SYNCPT host method, by patching its
+ * args in the command stream. The method data is changed to reference
+ * a reserved (never given out or incr) HOST1X_SYNCPT_RESERVED syncpt
+ * with a matching threshold value of 0, so is guaranteed to be popped
+ * by the host HW.
+ */
+static void host1x_syncpt_patch_offset(struct host1x_syncpt *sp,
+				       struct host1x_bo *h, u32 offset)
+{
+	void *patch_addr = NULL;
+
+	/* patch the wait */
+	patch_addr = host1x_bo_kmap(h, offset >> PAGE_SHIFT);
+	if (patch_addr) {
+		host1x_syncpt_patch_wait(sp,
+					 patch_addr + (offset & ~PAGE_MASK));
+		host1x_bo_kunmap(h, offset >> PAGE_SHIFT, patch_addr);
+	} else
+		pr_err("Could not map cmdbuf for wait check\n");
+}
+
+/*
+ * Check driver supplied waitchk structs for syncpt thresholds
+ * that have already been satisfied and NULL the comparison (to
+ * avoid a wrap condition in the HW).
+ */
+static int do_waitchks(struct host1x_job *job, struct host1x *host,
+		       struct host1x_bo *patch)
+{
+	int i;
+
+	/* compare syncpt vs wait threshold */
+	for (i = 0; i < job->num_waitchk; i++) {
+		struct host1x_waitchk *wait = &job->waitchk[i];
+		struct host1x_syncpt *sp =
+			host1x_syncpt_get(host, wait->syncpt_id);
+
+		/* validate syncpt id */
+		if (wait->syncpt_id > host1x_syncpt_nb_pts(host))
+			continue;
+
+		/* skip all other gathers */
+		if (patch != wait->bo)
+			continue;
+
+		trace_host1x_syncpt_wait_check(wait->bo, wait->offset,
+					       wait->syncpt_id, wait->thresh,
+					       host1x_syncpt_read_min(sp));
+
+		if (host1x_syncpt_is_expired(sp, wait->thresh)) {
+			dev_dbg(host->dev,
+				"drop WAIT id %d (%s) thresh 0x%x, min 0x%x\n",
+				wait->syncpt_id, sp->name, wait->thresh,
+				host1x_syncpt_read_min(sp));
+
+			host1x_syncpt_patch_offset(sp, patch, wait->offset);
+		}
+
+		wait->bo = NULL;
+	}
+
+	return 0;
+}
+
+static unsigned int pin_job(struct host1x_job *job)
+{
+	unsigned int i;
+
+	job->num_unpins = 0;
+
+	for (i = 0; i < job->num_relocs; i++) {
+		struct host1x_reloc *reloc = &job->relocarray[i];
+		struct sg_table *sgt;
+		dma_addr_t phys_addr;
+
+		reloc->target = host1x_bo_get(reloc->target);
+		if (!reloc->target)
+			goto unpin;
+
+		phys_addr = host1x_bo_pin(reloc->target, &sgt);
+		if (!phys_addr)
+			goto unpin;
+
+		job->addr_phys[job->num_unpins] = phys_addr;
+		job->unpins[job->num_unpins].bo = reloc->target;
+		job->unpins[job->num_unpins].sgt = sgt;
+		job->num_unpins++;
+	}
+
+	for (i = 0; i < job->num_gathers; i++) {
+		struct host1x_job_gather *g = &job->gathers[i];
+		struct sg_table *sgt;
+		dma_addr_t phys_addr;
+
+		g->bo = host1x_bo_get(g->bo);
+		if (!g->bo)
+			goto unpin;
+
+		phys_addr = host1x_bo_pin(g->bo, &sgt);
+		if (!phys_addr)
+			goto unpin;
+
+		job->addr_phys[job->num_unpins] = phys_addr;
+		job->unpins[job->num_unpins].bo = g->bo;
+		job->unpins[job->num_unpins].sgt = sgt;
+		job->num_unpins++;
+	}
+
+	return job->num_unpins;
+
+unpin:
+	host1x_job_unpin(job);
+	return 0;
+}
+
+static unsigned int do_relocs(struct host1x_job *job, struct host1x_bo *cmdbuf)
+{
+	int i = 0;
+	u32 last_page = ~0;
+	void *cmdbuf_page_addr = NULL;
+
+	/* pin & patch the relocs for one gather */
+	while (i < job->num_relocs) {
+		struct host1x_reloc *reloc = &job->relocarray[i];
+		u32 reloc_addr = (job->reloc_addr_phys[i] +
+			reloc->target_offset) >> reloc->shift;
+		u32 *target;
+
+		/* skip all other gathers */
+		if (!(reloc->cmdbuf && cmdbuf == reloc->cmdbuf)) {
+			i++;
+			continue;
+		}
+
+		if (last_page != reloc->cmdbuf_offset >> PAGE_SHIFT) {
+			if (cmdbuf_page_addr)
+				host1x_bo_kunmap(cmdbuf, last_page,
+						 cmdbuf_page_addr);
+
+			cmdbuf_page_addr = host1x_bo_kmap(cmdbuf,
+					reloc->cmdbuf_offset >> PAGE_SHIFT);
+			last_page = reloc->cmdbuf_offset >> PAGE_SHIFT;
+
+			if (unlikely(!cmdbuf_page_addr)) {
+				pr_err("Could not map cmdbuf for relocation\n");
+				return -ENOMEM;
+			}
+		}
+
+		target = cmdbuf_page_addr + (reloc->cmdbuf_offset & ~PAGE_MASK);
+		*target = reloc_addr;
+
+		/* mark this gather as handled */
+		reloc->cmdbuf = 0;
+	}
+
+	if (cmdbuf_page_addr)
+		host1x_bo_kunmap(cmdbuf, last_page, cmdbuf_page_addr);
+
+	return 0;
+}
+
+static int check_reloc(struct host1x_reloc *reloc, struct host1x_bo *cmdbuf,
+		       unsigned int offset)
+{
+	offset *= sizeof(u32);
+
+	if (reloc->cmdbuf != cmdbuf || reloc->cmdbuf_offset != offset)
+		return -EINVAL;
+
+	return 0;
+}
+
+struct host1x_firewall {
+	struct host1x_job *job;
+	struct device *dev;
+
+	unsigned int num_relocs;
+	struct host1x_reloc *reloc;
+
+	struct host1x_bo *cmdbuf_id;
+	unsigned int offset;
+
+	u32 words;
+	u32 class;
+	u32 reg;
+	u32 mask;
+	u32 count;
+};
+
+static int check_mask(struct host1x_firewall *fw)
+{
+	u32 mask = fw->mask;
+	u32 reg = fw->reg;
+
+	while (mask) {
+		if (fw->words == 0)
+			return -EINVAL;
+
+		if (mask & 1) {
+			if (fw->job->is_addr_reg(fw->dev, fw->class, reg)) {
+				bool bad_reloc = check_reloc(fw->reloc,
+							     fw->cmdbuf_id,
+							     fw->offset);
+				if (!fw->num_relocs || bad_reloc)
+					return -EINVAL;
+				fw->reloc++;
+				fw->num_relocs--;
+			}
+			fw->words--;
+			fw->offset++;
+		}
+		mask >>= 1;
+		reg++;
+	}
+
+	return 0;
+}
+
+static int check_incr(struct host1x_firewall *fw)
+{
+	u32 count = fw->count;
+	u32 reg = fw->reg;
+
+	while (fw) {
+		if (fw->words == 0)
+			return -EINVAL;
+
+		if (fw->job->is_addr_reg(fw->dev, fw->class, reg)) {
+			bool bad_reloc = check_reloc(fw->reloc, fw->cmdbuf_id,
+						     fw->offset);
+			if (!fw->num_relocs || bad_reloc)
+				return -EINVAL;
+			fw->reloc++;
+			fw->num_relocs--;
+		}
+		reg++;
+		fw->words--;
+		fw->offset++;
+		count--;
+	}
+
+	return 0;
+}
+
+static int check_nonincr(struct host1x_firewall *fw)
+{
+	int is_addr_reg = fw->job->is_addr_reg(fw->dev, fw->class, fw->reg);
+	u32 count = fw->count;
+
+	while (count) {
+		if (fw->words == 0)
+			return -EINVAL;
+
+		if (is_addr_reg) {
+			bool bad_reloc = check_reloc(fw->reloc, fw->cmdbuf_id,
+						     fw->offset);
+			if (!fw->num_relocs || bad_reloc)
+				return -EINVAL;
+			fw->reloc++;
+			fw->num_relocs--;
+		}
+		fw->words--;
+		fw->offset++;
+		count--;
+	}
+
+	return 0;
+}
+
+static int validate(struct host1x_job *job, struct device *dev,
+		    struct host1x_job_gather *g)
+{
+	u32 *cmdbuf_base;
+	int err = 0;
+	struct host1x_firewall fw;
+
+	fw.job = job;
+	fw.dev = dev;
+	fw.reloc = job->relocarray;
+	fw.num_relocs = job->num_relocs;
+	fw.cmdbuf_id = g->bo;
+
+	fw.offset = 0;
+	fw.class = 0;
+
+	if (!job->is_addr_reg)
+		return 0;
+
+	cmdbuf_base = host1x_bo_mmap(g->bo);
+	if (!cmdbuf_base)
+		return -ENOMEM;
+
+	fw.words = g->words;
+	while (fw.words && !err) {
+		u32 word = cmdbuf_base[fw.offset];
+		u32 opcode = (word & 0xf0000000) >> 28;
+
+		fw.mask = 0;
+		fw.reg = 0;
+		fw.count = 0;
+		fw.words--;
+		fw.offset++;
+
+		switch (opcode) {
+		case 0:
+			fw.class = word >> 6 & 0x3ff;
+			fw.mask = word & 0x3f;
+			fw.reg = word >> 16 & 0xfff;
+			err = check_mask(&fw);
+			if (err)
+				goto out;
+			break;
+		case 1:
+			fw.reg = word >> 16 & 0xfff;
+			fw.count = word & 0xffff;
+			err = check_incr(&fw);
+			if (err)
+				goto out;
+			break;
+
+		case 2:
+			fw.reg = word >> 16 & 0xfff;
+			fw.count = word & 0xffff;
+			err = check_nonincr(&fw);
+			if (err)
+				goto out;
+			break;
+
+		case 3:
+			fw.mask = word & 0xffff;
+			fw.reg = word >> 16 & 0xfff;
+			err = check_mask(&fw);
+			if (err)
+				goto out;
+			break;
+		case 4:
+		case 5:
+		case 14:
+			break;
+		default:
+			err = -EINVAL;
+			break;
+		}
+	}
+
+	/* No relocs should remain at this point */
+	if (fw.num_relocs)
+		err = -EINVAL;
+
+out:
+	host1x_bo_munmap(g->bo, cmdbuf_base);
+
+	return err;
+}
+
+static inline int copy_gathers(struct host1x_job *job, struct device *dev)
+{
+	size_t size = 0;
+	size_t offset = 0;
+	int i;
+
+	for (i = 0; i < job->num_gathers; i++) {
+		struct host1x_job_gather *g = &job->gathers[i];
+		size += g->words * sizeof(u32);
+	}
+
+	job->gather_copy_mapped = dma_alloc_writecombine(dev, size,
+							 &job->gather_copy,
+							 GFP_KERNEL);
+	if (!job->gather_copy_mapped) {
+		int err = PTR_ERR(job->gather_copy_mapped);
+		job->gather_copy_mapped = NULL;
+		return err;
+	}
+
+	job->gather_copy_size = size;
+
+	for (i = 0; i < job->num_gathers; i++) {
+		struct host1x_job_gather *g = &job->gathers[i];
+		void *gather;
+
+		gather = host1x_bo_mmap(g->bo);
+		memcpy(job->gather_copy_mapped + offset, gather + g->offset,
+		       g->words * sizeof(u32));
+		host1x_bo_munmap(g->bo, gather);
+
+		g->base = job->gather_copy;
+		g->offset = offset;
+		g->bo = NULL;
+
+		offset += g->words * sizeof(u32);
+	}
+
+	return 0;
+}
+
+int host1x_job_pin(struct host1x_job *job, struct device *dev)
+{
+	int err;
+	unsigned int i, j;
+	struct host1x *host = dev_get_drvdata(dev->parent);
+	DECLARE_BITMAP(waitchk_mask, host1x_syncpt_nb_pts(host));
+
+	bitmap_zero(waitchk_mask, host1x_syncpt_nb_pts(host));
+	for (i = 0; i < job->num_waitchk; i++) {
+		u32 syncpt_id = job->waitchk[i].syncpt_id;
+		if (syncpt_id < host1x_syncpt_nb_pts(host))
+			set_bit(syncpt_id, waitchk_mask);
+	}
+
+	/* get current syncpt values for waitchk */
+	for_each_set_bit(i, waitchk_mask, host1x_syncpt_nb_pts(host))
+		host1x_syncpt_load(host->syncpt + i);
+
+	/* pin memory */
+	err = pin_job(job);
+	if (!err)
+		goto out;
+
+	/* patch gathers */
+	for (i = 0; i < job->num_gathers; i++) {
+		struct host1x_job_gather *g = &job->gathers[i];
+
+		/* process each gather mem only once */
+		if (g->handled)
+			continue;
+
+		g->base = job->gather_addr_phys[i];
+
+		for (j = 0; j < job->num_gathers; j++)
+			if (job->gathers[j].bo == g->bo)
+				job->gathers[j].handled = true;
+
+		err = 0;
+
+		if (IS_ENABLED(CONFIG_TEGRA_HOST1X_FIREWALL))
+			err = validate(job, dev, g);
+
+		if (err)
+			dev_err(dev, "Job invalid (err=%d)\n", err);
+
+		if (!err)
+			err = do_relocs(job, g->bo);
+
+		if (!err)
+			err = do_waitchks(job, host, g->bo);
+
+		if (err)
+			break;
+	}
+
+	if (IS_ENABLED(CONFIG_TEGRA_HOST1X_FIREWALL) && !err) {
+		err = copy_gathers(job, dev);
+		if (err) {
+			host1x_job_unpin(job);
+			return err;
+		}
+	}
+
+out:
+	wmb();
+
+	return err;
+}
+
+void host1x_job_unpin(struct host1x_job *job)
+{
+	unsigned int i;
+
+	for (i = 0; i < job->num_unpins; i++) {
+		struct host1x_job_unpin_data *unpin = &job->unpins[i];
+		host1x_bo_unpin(unpin->bo, unpin->sgt);
+		host1x_bo_put(unpin->bo);
+	}
+	job->num_unpins = 0;
+
+	if (job->gather_copy_size)
+		dma_free_writecombine(job->channel->dev, job->gather_copy_size,
+				      job->gather_copy_mapped,
+				      job->gather_copy);
+}
+
+/*
+ * Debug routine used to dump job entries
+ */
+void host1x_job_dump(struct device *dev, struct host1x_job *job)
+{
+	dev_dbg(dev, "    SYNCPT_ID   %d\n", job->syncpt_id);
+	dev_dbg(dev, "    SYNCPT_VAL  %d\n", job->syncpt_end);
+	dev_dbg(dev, "    FIRST_GET   0x%x\n", job->first_get);
+	dev_dbg(dev, "    TIMEOUT     %d\n", job->timeout);
+	dev_dbg(dev, "    NUM_SLOTS   %d\n", job->num_slots);
+	dev_dbg(dev, "    NUM_HANDLES %d\n", job->num_unpins);
+}
diff --git a/drivers/gpu/host1x/job.h b/drivers/gpu/host1x/job.h
new file mode 100644
index 000000000000..fba45f20458e
--- /dev/null
+++ b/drivers/gpu/host1x/job.h
@@ -0,0 +1,162 @@
+/*
+ * Tegra host1x Job
+ *
+ * Copyright (c) 2011-2013, NVIDIA Corporation.
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms and conditions of the GNU General Public License,
+ * version 2, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program.  If not, see <http://www.gnu.org/licenses/>.
+ */
+
+#ifndef __HOST1X_JOB_H
+#define __HOST1X_JOB_H
+
+struct host1x_job_gather {
+	u32 words;
+	dma_addr_t base;
+	struct host1x_bo *bo;
+	int offset;
+	bool handled;
+};
+
+struct host1x_cmdbuf {
+	u32 handle;
+	u32 offset;
+	u32 words;
+	u32 pad;
+};
+
+struct host1x_reloc {
+	struct host1x_bo *cmdbuf;
+	u32 cmdbuf_offset;
+	struct host1x_bo *target;
+	u32 target_offset;
+	u32 shift;
+	u32 pad;
+};
+
+struct host1x_waitchk {
+	struct host1x_bo *bo;
+	u32 offset;
+	u32 syncpt_id;
+	u32 thresh;
+};
+
+struct host1x_job_unpin_data {
+	struct host1x_bo *bo;
+	struct sg_table *sgt;
+};
+
+/*
+ * Each submit is tracked as a host1x_job.
+ */
+struct host1x_job {
+	/* When refcount goes to zero, job can be freed */
+	struct kref ref;
+
+	/* List entry */
+	struct list_head list;
+
+	/* Channel where job is submitted to */
+	struct host1x_channel *channel;
+
+	u32 client;
+
+	/* Gathers and their memory */
+	struct host1x_job_gather *gathers;
+	unsigned int num_gathers;
+
+	/* Wait checks to be processed at submit time */
+	struct host1x_waitchk *waitchk;
+	unsigned int num_waitchk;
+	u32 waitchk_mask;
+
+	/* Array of handles to be pinned & unpinned */
+	struct host1x_reloc *relocarray;
+	unsigned int num_relocs;
+	struct host1x_job_unpin_data *unpins;
+	unsigned int num_unpins;
+
+	dma_addr_t *addr_phys;
+	dma_addr_t *gather_addr_phys;
+	dma_addr_t *reloc_addr_phys;
+
+	/* Sync point id, number of increments and end related to the submit */
+	u32 syncpt_id;
+	u32 syncpt_incrs;
+	u32 syncpt_end;
+
+	/* Maximum time to wait for this job */
+	unsigned int timeout;
+
+	/* Index and number of slots used in the push buffer */
+	unsigned int first_get;
+	unsigned int num_slots;
+
+	/* Copy of gathers */
+	size_t gather_copy_size;
+	dma_addr_t gather_copy;
+	u8 *gather_copy_mapped;
+
+	/* Check if register is marked as an address reg */
+	int (*is_addr_reg)(struct device *dev, u32 reg, u32 class);
+
+	/* Request a SETCLASS to this class */
+	u32 class;
+
+	/* Add a channel wait for previous ops to complete */
+	bool serialize;
+};
+/*
+ * Allocate memory for a job. Just enough memory will be allocated to
+ * accomodate the submit.
+ */
+struct host1x_job *host1x_job_alloc(struct host1x_channel *ch,
+				    u32 num_cmdbufs, u32 num_relocs,
+				    u32 num_waitchks);
+
+/*
+ * Add a gather to a job.
+ */
+void host1x_job_add_gather(struct host1x_job *job, struct host1x_bo *mem_id,
+			   u32 words, u32 offset);
+
+/*
+ * Increment reference going to host1x_job.
+ */
+struct host1x_job *host1x_job_get(struct host1x_job *job);
+
+/*
+ * Decrement reference job, free if goes to zero.
+ */
+void host1x_job_put(struct host1x_job *job);
+
+/*
+ * Pin memory related to job. This handles relocation of addresses to the
+ * host1x address space. Handles both the gather memory and any other memory
+ * referred to from the gather buffers.
+ *
+ * Handles also patching out host waits that would wait for an expired sync
+ * point value.
+ */
+int host1x_job_pin(struct host1x_job *job, struct device *dev);
+
+/*
+ * Unpin memory related to job.
+ */
+void host1x_job_unpin(struct host1x_job *job);
+
+/*
+ * Dump contents of job to debug output.
+ */
+void host1x_job_dump(struct device *dev, struct host1x_job *job);
+
+#endif
diff --git a/drivers/gpu/host1x/syncpt.c b/drivers/gpu/host1x/syncpt.c
new file mode 100644
index 000000000000..4b493453e805
--- /dev/null
+++ b/drivers/gpu/host1x/syncpt.c
@@ -0,0 +1,387 @@
+/*
+ * Tegra host1x Syncpoints
+ *
+ * Copyright (c) 2010-2013, NVIDIA Corporation.
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms and conditions of the GNU General Public License,
+ * version 2, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program.  If not, see <http://www.gnu.org/licenses/>.
+ */
+
+#include <linux/module.h>
+#include <linux/device.h>
+#include <linux/slab.h>
+
+#include <trace/events/host1x.h>
+
+#include "syncpt.h"
+#include "dev.h"
+#include "intr.h"
+#include "debug.h"
+
+#define SYNCPT_CHECK_PERIOD (2 * HZ)
+#define MAX_STUCK_CHECK_COUNT 15
+
+static struct host1x_syncpt *_host1x_syncpt_alloc(struct host1x *host,
+						  struct device *dev,
+						  int client_managed)
+{
+	int i;
+	struct host1x_syncpt *sp = host->syncpt;
+	char *name;
+
+	for (i = 0; i < host->info->nb_pts && sp->name; i++, sp++)
+		;
+	if (sp->dev)
+		return NULL;
+
+	name = kasprintf(GFP_KERNEL, "%02d-%s", sp->id,
+			dev ? dev_name(dev) : NULL);
+	if (!name)
+		return NULL;
+
+	sp->dev = dev;
+	sp->name = name;
+	sp->client_managed = client_managed;
+
+	return sp;
+}
+
+u32 host1x_syncpt_id(struct host1x_syncpt *sp)
+{
+	return sp->id;
+}
+
+/*
+ * Updates the value sent to hardware.
+ */
+u32 host1x_syncpt_incr_max(struct host1x_syncpt *sp, u32 incrs)
+{
+	return (u32)atomic_add_return(incrs, &sp->max_val);
+}
+
+ /*
+ * Write cached syncpoint and waitbase values to hardware.
+ */
+void host1x_syncpt_restore(struct host1x *host)
+{
+	struct host1x_syncpt *sp_base = host->syncpt;
+	u32 i;
+
+	for (i = 0; i < host1x_syncpt_nb_pts(host); i++)
+		host1x_hw_syncpt_restore(host, sp_base + i);
+	for (i = 0; i < host1x_syncpt_nb_bases(host); i++)
+		host1x_hw_syncpt_restore_wait_base(host, sp_base + i);
+	wmb();
+}
+
+/*
+ * Update the cached syncpoint and waitbase values by reading them
+ * from the registers.
+  */
+void host1x_syncpt_save(struct host1x *host)
+{
+	struct host1x_syncpt *sp_base = host->syncpt;
+	u32 i;
+
+	for (i = 0; i < host1x_syncpt_nb_pts(host); i++) {
+		if (host1x_syncpt_client_managed(sp_base + i))
+			host1x_hw_syncpt_load(host, sp_base + i);
+		else
+			WARN_ON(!host1x_syncpt_idle(sp_base + i));
+	}
+
+	for (i = 0; i < host1x_syncpt_nb_bases(host); i++)
+		host1x_hw_syncpt_load_wait_base(host, sp_base + i);
+}
+
+/*
+ * Updates the cached syncpoint value by reading a new value from the hardware
+ * register
+ */
+u32 host1x_syncpt_load(struct host1x_syncpt *sp)
+{
+	u32 val;
+	val = host1x_hw_syncpt_load(sp->host, sp);
+	trace_host1x_syncpt_load_min(sp->id, val);
+
+	return val;
+}
+
+/*
+ * Get the current syncpoint base
+ */
+u32 host1x_syncpt_load_wait_base(struct host1x_syncpt *sp)
+{
+	u32 val;
+	host1x_hw_syncpt_load_wait_base(sp->host, sp);
+	val = sp->base_val;
+	return val;
+}
+
+/*
+ * Write a cpu syncpoint increment to the hardware, without touching
+ * the cache. Caller is responsible for host being powered.
+ */
+void host1x_syncpt_cpu_incr(struct host1x_syncpt *sp)
+{
+	host1x_hw_syncpt_cpu_incr(sp->host, sp);
+}
+
+/*
+ * Increment syncpoint value from cpu, updating cache
+ */
+void host1x_syncpt_incr(struct host1x_syncpt *sp)
+{
+	if (host1x_syncpt_client_managed(sp))
+		host1x_syncpt_incr_max(sp, 1);
+	host1x_syncpt_cpu_incr(sp);
+}
+
+/*
+ * Updated sync point form hardware, and returns true if syncpoint is expired,
+ * false if we may need to wait
+ */
+static bool syncpt_load_min_is_expired(struct host1x_syncpt *sp, u32 thresh)
+{
+	host1x_hw_syncpt_load(sp->host, sp);
+	return host1x_syncpt_is_expired(sp, thresh);
+}
+
+/*
+ * Main entrypoint for syncpoint value waits.
+ */
+int host1x_syncpt_wait(struct host1x_syncpt *sp, u32 thresh, long timeout,
+			u32 *value)
+{
+	DECLARE_WAIT_QUEUE_HEAD_ONSTACK(wq);
+	void *ref;
+	struct host1x_waitlist *waiter;
+	int err = 0, check_count = 0;
+	u32 val;
+
+	if (value)
+		*value = 0;
+
+	/* first check cache */
+	if (host1x_syncpt_is_expired(sp, thresh)) {
+		if (value)
+			*value = host1x_syncpt_load(sp);
+		return 0;
+	}
+
+	/* try to read from register */
+	val = host1x_hw_syncpt_load(sp->host, sp);
+	if (host1x_syncpt_is_expired(sp, thresh)) {
+		if (value)
+			*value = val;
+		goto done;
+	}
+
+	if (!timeout) {
+		err = -EAGAIN;
+		goto done;
+	}
+
+	/* allocate a waiter */
+	waiter = kzalloc(sizeof(*waiter), GFP_KERNEL);
+	if (!waiter) {
+		err = -ENOMEM;
+		goto done;
+	}
+
+	/* schedule a wakeup when the syncpoint value is reached */
+	err = host1x_intr_add_action(sp->host, sp->id, thresh,
+				     HOST1X_INTR_ACTION_WAKEUP_INTERRUPTIBLE,
+				     &wq, waiter, &ref);
+	if (err)
+		goto done;
+
+	err = -EAGAIN;
+	/* Caller-specified timeout may be impractically low */
+	if (timeout < 0)
+		timeout = LONG_MAX;
+
+	/* wait for the syncpoint, or timeout, or signal */
+	while (timeout) {
+		long check = min_t(long, SYNCPT_CHECK_PERIOD, timeout);
+		int remain = wait_event_interruptible_timeout(wq,
+				syncpt_load_min_is_expired(sp, thresh),
+				check);
+		if (remain > 0 || host1x_syncpt_is_expired(sp, thresh)) {
+			if (value)
+				*value = host1x_syncpt_load(sp);
+			err = 0;
+			break;
+		}
+		if (remain < 0) {
+			err = remain;
+			break;
+		}
+		timeout -= check;
+		if (timeout && check_count <= MAX_STUCK_CHECK_COUNT) {
+			dev_warn(sp->host->dev,
+				"%s: syncpoint id %d (%s) stuck waiting %d, timeout=%ld\n",
+				 current->comm, sp->id, sp->name,
+				 thresh, timeout);
+
+			host1x_debug_dump_syncpts(sp->host);
+			if (check_count == MAX_STUCK_CHECK_COUNT)
+				host1x_debug_dump(sp->host);
+			check_count++;
+		}
+	}
+	host1x_intr_put_ref(sp->host, sp->id, ref);
+
+done:
+	return err;
+}
+EXPORT_SYMBOL(host1x_syncpt_wait);
+
+/*
+ * Returns true if syncpoint is expired, false if we may need to wait
+ */
+bool host1x_syncpt_is_expired(struct host1x_syncpt *sp, u32 thresh)
+{
+	u32 current_val;
+	u32 future_val;
+	smp_rmb();
+	current_val = (u32)atomic_read(&sp->min_val);
+	future_val = (u32)atomic_read(&sp->max_val);
+
+	/* Note the use of unsigned arithmetic here (mod 1<<32).
+	 *
+	 * c = current_val = min_val	= the current value of the syncpoint.
+	 * t = thresh			= the value we are checking
+	 * f = future_val  = max_val	= the value c will reach when all
+	 *				  outstanding increments have completed.
+	 *
+	 * Note that c always chases f until it reaches f.
+	 *
+	 * Dtf = (f - t)
+	 * Dtc = (c - t)
+	 *
+	 *  Consider all cases:
+	 *
+	 *	A) .....c..t..f.....	Dtf < Dtc	need to wait
+	 *	B) .....c.....f..t..	Dtf > Dtc	expired
+	 *	C) ..t..c.....f.....	Dtf > Dtc	expired	   (Dct very large)
+	 *
+	 *  Any case where f==c: always expired (for any t).	Dtf == Dcf
+	 *  Any case where t==c: always expired (for any f).	Dtf >= Dtc (because Dtc==0)
+	 *  Any case where t==f!=c: always wait.		Dtf <  Dtc (because Dtf==0,
+	 *							Dtc!=0)
+	 *
+	 *  Other cases:
+	 *
+	 *	A) .....t..f..c.....	Dtf < Dtc	need to wait
+	 *	A) .....f..c..t.....	Dtf < Dtc	need to wait
+	 *	A) .....f..t..c.....	Dtf > Dtc	expired
+	 *
+	 *   So:
+	 *	   Dtf >= Dtc implies EXPIRED	(return true)
+	 *	   Dtf <  Dtc implies WAIT	(return false)
+	 *
+	 * Note: If t is expired then we *cannot* wait on it. We would wait
+	 * forever (hang the system).
+	 *
+	 * Note: do NOT get clever and remove the -thresh from both sides. It
+	 * is NOT the same.
+	 *
+	 * If future valueis zero, we have a client managed sync point. In that
+	 * case we do a direct comparison.
+	 */
+	if (!host1x_syncpt_client_managed(sp))
+		return future_val - thresh >= current_val - thresh;
+	else
+		return (s32)(current_val - thresh) >= 0;
+}
+
+/* remove a wait pointed to by patch_addr */
+int host1x_syncpt_patch_wait(struct host1x_syncpt *sp, void *patch_addr)
+{
+	return host1x_hw_syncpt_patch_wait(sp->host, sp, patch_addr);
+}
+
+int host1x_syncpt_init(struct host1x *host)
+{
+	struct host1x_syncpt *syncpt;
+	int i;
+
+	syncpt = devm_kzalloc(host->dev, sizeof(*syncpt) * host->info->nb_pts,
+		GFP_KERNEL);
+	if (!syncpt)
+		return -ENOMEM;
+
+	for (i = 0; i < host->info->nb_pts; ++i) {
+		syncpt[i].id = i;
+		syncpt[i].host = host;
+	}
+
+	host->syncpt = syncpt;
+
+	host1x_syncpt_restore(host);
+
+	/* Allocate sync point to use for clearing waits for expired fences */
+	host->nop_sp = _host1x_syncpt_alloc(host, NULL, 0);
+	if (!host->nop_sp)
+		return -ENOMEM;
+
+	return 0;
+}
+
+struct host1x_syncpt *host1x_syncpt_request(struct device *dev,
+					    int client_managed)
+{
+	struct host1x *host = dev_get_drvdata(dev->parent);
+	return _host1x_syncpt_alloc(host, dev, client_managed);
+}
+
+void host1x_syncpt_free(struct host1x_syncpt *sp)
+{
+	if (!sp)
+		return;
+
+	kfree(sp->name);
+	sp->dev = NULL;
+	sp->name = NULL;
+	sp->client_managed = 0;
+}
+
+void host1x_syncpt_deinit(struct host1x *host)
+{
+	int i;
+	struct host1x_syncpt *sp = host->syncpt;
+	for (i = 0; i < host->info->nb_pts; i++, sp++)
+		kfree(sp->name);
+}
+
+int host1x_syncpt_nb_pts(struct host1x *host)
+{
+	return host->info->nb_pts;
+}
+
+int host1x_syncpt_nb_bases(struct host1x *host)
+{
+	return host->info->nb_bases;
+}
+
+int host1x_syncpt_nb_mlocks(struct host1x *host)
+{
+	return host->info->nb_mlocks;
+}
+
+struct host1x_syncpt *host1x_syncpt_get(struct host1x *host, u32 id)
+{
+	if (host->info->nb_pts < id)
+		return NULL;
+	return host->syncpt + id;
+}
diff --git a/drivers/gpu/host1x/syncpt.h b/drivers/gpu/host1x/syncpt.h
new file mode 100644
index 000000000000..c99806130f2e
--- /dev/null
+++ b/drivers/gpu/host1x/syncpt.h
@@ -0,0 +1,165 @@
+/*
+ * Tegra host1x Syncpoints
+ *
+ * Copyright (c) 2010-2013, NVIDIA Corporation.
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms and conditions of the GNU General Public License,
+ * version 2, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program.  If not, see <http://www.gnu.org/licenses/>.
+ */
+
+#ifndef __HOST1X_SYNCPT_H
+#define __HOST1X_SYNCPT_H
+
+#include <linux/atomic.h>
+#include <linux/kernel.h>
+#include <linux/sched.h>
+
+#include "intr.h"
+
+struct host1x;
+
+/* Reserved for replacing an expired wait with a NOP */
+#define HOST1X_SYNCPT_RESERVED			0
+
+struct host1x_syncpt {
+	int id;
+	atomic_t min_val;
+	atomic_t max_val;
+	u32 base_val;
+	const char *name;
+	int client_managed;
+	struct host1x *host;
+	struct device *dev;
+
+	/* interrupt data */
+	struct host1x_syncpt_intr intr;
+};
+
+/* Initialize sync point array  */
+int host1x_syncpt_init(struct host1x *host);
+
+/*  Free sync point array */
+void host1x_syncpt_deinit(struct host1x *host);
+
+/*
+ * Read max. It indicates how many operations there are in queue, either in
+ * channel or in a software thread.
+ * */
+static inline u32 host1x_syncpt_read_max(struct host1x_syncpt *sp)
+{
+	smp_rmb();
+	return (u32)atomic_read(&sp->max_val);
+}
+
+/*
+ * Read min, which is a shadow of the current sync point value in hardware.
+ */
+static inline u32 host1x_syncpt_read_min(struct host1x_syncpt *sp)
+{
+	smp_rmb();
+	return (u32)atomic_read(&sp->min_val);
+}
+
+/* Return number of sync point supported. */
+int host1x_syncpt_nb_pts(struct host1x *host);
+
+/* Return number of wait bases supported. */
+int host1x_syncpt_nb_bases(struct host1x *host);
+
+/* Return number of mlocks supported. */
+int host1x_syncpt_nb_mlocks(struct host1x *host);
+
+/*
+ * Check sync point sanity. If max is larger than min, there have too many
+ * sync point increments.
+ *
+ * Client managed sync point are not tracked.
+ * */
+static inline bool host1x_syncpt_check_max(struct host1x_syncpt *sp, u32 real)
+{
+	u32 max;
+	if (sp->client_managed)
+		return true;
+	max = host1x_syncpt_read_max(sp);
+	return (s32)(max - real) >= 0;
+}
+
+/* Return true if sync point is client managed. */
+static inline int host1x_syncpt_client_managed(struct host1x_syncpt *sp)
+{
+	return sp->client_managed;
+}
+
+/*
+ * Returns true if syncpoint min == max, which means that there are no
+ * outstanding operations.
+ */
+static inline bool host1x_syncpt_idle(struct host1x_syncpt *sp)
+{
+	int min, max;
+	smp_rmb();
+	min = atomic_read(&sp->min_val);
+	max = atomic_read(&sp->max_val);
+	return (min == max);
+}
+
+/* Return pointer to struct denoting sync point id. */
+struct host1x_syncpt *host1x_syncpt_get(struct host1x *host, u32 id);
+
+/* Request incrementing a sync point. */
+void host1x_syncpt_cpu_incr(struct host1x_syncpt *sp);
+
+/* Load current value from hardware to the shadow register. */
+u32 host1x_syncpt_load(struct host1x_syncpt *sp);
+
+/* Check if the given syncpoint value has already passed */
+bool host1x_syncpt_is_expired(struct host1x_syncpt *sp, u32 thresh);
+
+/* Save host1x sync point state into shadow registers. */
+void host1x_syncpt_save(struct host1x *host);
+
+/* Reset host1x sync point state from shadow registers. */
+void host1x_syncpt_restore(struct host1x *host);
+
+/* Read current wait base value into shadow register and return it. */
+u32 host1x_syncpt_load_wait_base(struct host1x_syncpt *sp);
+
+/* Increment sync point and its max. */
+void host1x_syncpt_incr(struct host1x_syncpt *sp);
+
+/* Indicate future operations by incrementing the sync point max. */
+u32 host1x_syncpt_incr_max(struct host1x_syncpt *sp, u32 incrs);
+
+/* Wait until sync point reaches a threshold value, or a timeout. */
+int host1x_syncpt_wait(struct host1x_syncpt *sp, u32 thresh,
+			long timeout, u32 *value);
+
+/* Check if sync point id is valid. */
+static inline int host1x_syncpt_is_valid(struct host1x_syncpt *sp)
+{
+	return sp->id < host1x_syncpt_nb_pts(sp->host);
+}
+
+/* Patch a wait by replacing it with a wait for syncpt 0 value 0 */
+int host1x_syncpt_patch_wait(struct host1x_syncpt *sp, void *patch_addr);
+
+/* Return id of the sync point */
+u32 host1x_syncpt_id(struct host1x_syncpt *sp);
+
+/* Allocate a sync point for a device. */
+struct host1x_syncpt *host1x_syncpt_request(struct device *dev,
+		int client_managed);
+
+/* Free a sync point. */
+void host1x_syncpt_free(struct host1x_syncpt *sp);
+
+#endif