Merge branch 'master' into for-next

Merge with latest Linus' tree, as I have incoming patches
that fix code that is newer than current HEAD of for-next.

Conflicts:
	drivers/net/ethernet/realtek/r8169.c

276 files changed, 30687 insertions, 7017 deletions

diff --git a/drivers/gpu/drm/Kconfig b/drivers/gpu/drm/Kconfig
index 2418429a9836..e354bc0b052a 100644
--- a/drivers/gpu/drm/Kconfig
+++ b/drivers/gpu/drm/Kconfig
@@ -9,6 +9,7 @@ menuconfig DRM
 	depends on (AGP || AGP=n) && !EMULATED_CMPXCHG && MMU
 	select I2C
 	select I2C_ALGOBIT
+	select DMA_SHARED_BUFFER
 	help
 	  Kernel-level support for the Direct Rendering Infrastructure (DRI)
 	  introduced in XFree86 4.0. If you say Y here, you need to select
@@ -18,6 +19,11 @@ menuconfig DRM
 	  details.  You should also select and configure AGP
 	  (/dev/agpgart) support if it is available for your platform.
 
+config DRM_USB
+	tristate
+	depends on DRM
+	select USB
+
 config DRM_KMS_HELPER
 	tristate
 	depends on DRM
@@ -27,6 +33,18 @@ config DRM_KMS_HELPER
 	help
 	  FB and CRTC helpers for KMS drivers.
 
+config DRM_LOAD_EDID_FIRMWARE
+	bool "Allow to specify an EDID data set instead of probing for it"
+	depends on DRM_KMS_HELPER
+	help
+	  Say Y here, if you want to use EDID data to be loaded from the
+	  /lib/firmware directory or one of the provided built-in
+	  data sets. This may be necessary, if the graphics adapter or
+	  monitor are unable to provide appropriate EDID data. Since this
+	  feature is provided as a workaround for broken hardware, the
+	  default case is N. Details and instructions how to build your own
+	  EDID data are given in Documentation/EDID/HOWTO.txt.
+
 config DRM_TTM
 	tristate
 	depends on DRM
@@ -71,6 +89,8 @@ config DRM_RADEON
 
 source "drivers/gpu/drm/radeon/Kconfig"
 
+source "drivers/gpu/drm/nouveau/Kconfig"
+
 config DRM_I810
 	tristate "Intel I810"
 	# !PREEMPT because of missing ioctl locking
@@ -165,3 +185,4 @@ source "drivers/gpu/drm/vmwgfx/Kconfig"
 
 source "drivers/gpu/drm/gma500/Kconfig"
 
+source "drivers/gpu/drm/udl/Kconfig"
diff --git a/drivers/gpu/drm/Makefile b/drivers/gpu/drm/Makefile
index 0cde1b80fdb1..c20da5bda355 100644
--- a/drivers/gpu/drm/Makefile
+++ b/drivers/gpu/drm/Makefile
@@ -12,17 +12,21 @@ drm-y       :=	drm_auth.o drm_buffer.o drm_bufs.o drm_cache.o \
 		drm_platform.o drm_sysfs.o drm_hashtab.o drm_mm.o \
 		drm_crtc.o drm_modes.o drm_edid.o \
 		drm_info.o drm_debugfs.o drm_encoder_slave.o \
-		drm_trace_points.o drm_global.o drm_usb.o
+		drm_trace_points.o drm_global.o drm_prime.o
 
 drm-$(CONFIG_COMPAT) += drm_ioc32.o
 
+drm-usb-y   := drm_usb.o
+
 drm_kms_helper-y := drm_fb_helper.o drm_crtc_helper.o drm_dp_i2c_helper.o
+drm_kms_helper-$(CONFIG_DRM_LOAD_EDID_FIRMWARE) += drm_edid_load.o
 
 obj-$(CONFIG_DRM_KMS_HELPER) += drm_kms_helper.o
 
 CFLAGS_drm_trace_points.o := -I$(src)
 
 obj-$(CONFIG_DRM)	+= drm.o
+obj-$(CONFIG_DRM_USB)   += drm_usb.o
 obj-$(CONFIG_DRM_TTM)	+= ttm/
 obj-$(CONFIG_DRM_TDFX)	+= tdfx/
 obj-$(CONFIG_DRM_R128)	+= r128/
@@ -37,4 +41,5 @@ obj-$(CONFIG_DRM_VIA)	+=via/
 obj-$(CONFIG_DRM_NOUVEAU) +=nouveau/
 obj-$(CONFIG_DRM_EXYNOS) +=exynos/
 obj-$(CONFIG_DRM_GMA500) += gma500/
+obj-$(CONFIG_DRM_UDL) += udl/
 obj-y			+= i2c/
diff --git a/drivers/gpu/drm/drm_cache.c b/drivers/gpu/drm/drm_cache.c
index 592865381c6e..4b8653b932f9 100644
--- a/drivers/gpu/drm/drm_cache.c
+++ b/drivers/gpu/drm/drm_cache.c
@@ -41,10 +41,10 @@ drm_clflush_page(struct page *page)
 	if (unlikely(page == NULL))
 		return;
 
-	page_virtual = kmap_atomic(page, KM_USER0);
+	page_virtual = kmap_atomic(page);
 	for (i = 0; i < PAGE_SIZE; i += boot_cpu_data.x86_clflush_size)
 		clflush(page_virtual + i);
-	kunmap_atomic(page_virtual, KM_USER0);
+	kunmap_atomic(page_virtual);
 }
 
 static void drm_cache_flush_clflush(struct page *pages[],
@@ -87,10 +87,10 @@ drm_clflush_pages(struct page *pages[], unsigned long num_pages)
 		if (unlikely(page == NULL))
 			continue;
 
-		page_virtual = kmap_atomic(page, KM_USER0);
+		page_virtual = kmap_atomic(page);
 		flush_dcache_range((unsigned long)page_virtual,
 				   (unsigned long)page_virtual + PAGE_SIZE);
-		kunmap_atomic(page_virtual, KM_USER0);
+		kunmap_atomic(page_virtual);
 	}
 #else
 	printk(KERN_ERR "Architecture has no drm_cache.c support\n");
diff --git a/drivers/gpu/drm/drm_crtc.c b/drivers/gpu/drm/drm_crtc.c
index 5e818a808ace..d3aaeb6ae236 100644
--- a/drivers/gpu/drm/drm_crtc.c
+++ b/drivers/gpu/drm/drm_crtc.c
@@ -38,11 +38,6 @@
 #include "drm_edid.h"
 #include "drm_fourcc.h"
 
-struct drm_prop_enum_list {
-	int type;
-	char *name;
-};
-
 /* Avoid boilerplate.  I'm tired of typing. */
 #define DRM_ENUM_NAME_FN(fnname, list)				\
 	char *fnname(int val)					\
@@ -298,9 +293,8 @@ int drm_framebuffer_init(struct drm_device *dev, struct drm_framebuffer *fb,
 	int ret;
 
 	ret = drm_mode_object_get(dev, &fb->base, DRM_MODE_OBJECT_FB);
-	if (ret) {
+	if (ret)
 		return ret;
-	}
 
 	fb->dev = dev;
 	fb->funcs = funcs;
@@ -370,19 +364,31 @@ EXPORT_SYMBOL(drm_framebuffer_cleanup);
  * Caller must hold mode config lock.
  *
  * Inits a new object created as base part of an driver crtc object.
+ *
+ * RETURNS:
+ * Zero on success, error code on failure.
  */
-void drm_crtc_init(struct drm_device *dev, struct drm_crtc *crtc,
+int drm_crtc_init(struct drm_device *dev, struct drm_crtc *crtc,
 		   const struct drm_crtc_funcs *funcs)
 {
+	int ret;
+
 	crtc->dev = dev;
 	crtc->funcs = funcs;
 
 	mutex_lock(&dev->mode_config.mutex);
-	drm_mode_object_get(dev, &crtc->base, DRM_MODE_OBJECT_CRTC);
+
+	ret = drm_mode_object_get(dev, &crtc->base, DRM_MODE_OBJECT_CRTC);
+	if (ret)
+		goto out;
 
 	list_add_tail(&crtc->head, &dev->mode_config.crtc_list);
 	dev->mode_config.num_crtc++;
+
+ out:
 	mutex_unlock(&dev->mode_config.mutex);
+
+	return ret;
 }
 EXPORT_SYMBOL(drm_crtc_init);
 
@@ -442,7 +448,7 @@ void drm_mode_remove(struct drm_connector *connector,
 		     struct drm_display_mode *mode)
 {
 	list_del(&mode->head);
-	kfree(mode);
+	drm_mode_destroy(connector->dev, mode);
 }
 EXPORT_SYMBOL(drm_mode_remove);
 
@@ -454,21 +460,29 @@ EXPORT_SYMBOL(drm_mode_remove);
  * @name: user visible name of the connector
  *
  * LOCKING:
- * Caller must hold @dev's mode_config lock.
+ * Takes mode config lock.
  *
  * Initialises a preallocated connector. Connectors should be
  * subclassed as part of driver connector objects.
+ *
+ * RETURNS:
+ * Zero on success, error code on failure.
  */
-void drm_connector_init(struct drm_device *dev,
-		     struct drm_connector *connector,
-		     const struct drm_connector_funcs *funcs,
-		     int connector_type)
+int drm_connector_init(struct drm_device *dev,
+		       struct drm_connector *connector,
+		       const struct drm_connector_funcs *funcs,
+		       int connector_type)
 {
+	int ret;
+
 	mutex_lock(&dev->mode_config.mutex);
 
+	ret = drm_mode_object_get(dev, &connector->base, DRM_MODE_OBJECT_CONNECTOR);
+	if (ret)
+		goto out;
+
 	connector->dev = dev;
 	connector->funcs = funcs;
-	drm_mode_object_get(dev, &connector->base, DRM_MODE_OBJECT_CONNECTOR);
 	connector->connector_type = connector_type;
 	connector->connector_type_id =
 		++drm_connector_enum_list[connector_type].count; /* TODO */
@@ -488,7 +502,10 @@ void drm_connector_init(struct drm_device *dev,
 	drm_connector_attach_property(connector,
 				      dev->mode_config.dpms_property, 0);
 
+ out:
 	mutex_unlock(&dev->mode_config.mutex);
+
+	return ret;
 }
 EXPORT_SYMBOL(drm_connector_init);
 
@@ -497,7 +514,7 @@ EXPORT_SYMBOL(drm_connector_init);
  * @connector: connector to cleanup
  *
  * LOCKING:
- * Caller must hold @dev's mode_config lock.
+ * Takes mode config lock.
  *
  * Cleans up the connector but doesn't free the object.
  */
@@ -523,23 +540,41 @@ void drm_connector_cleanup(struct drm_connector *connector)
 }
 EXPORT_SYMBOL(drm_connector_cleanup);
 
-void drm_encoder_init(struct drm_device *dev,
+void drm_connector_unplug_all(struct drm_device *dev)
+{
+	struct drm_connector *connector;
+
+	/* taking the mode config mutex ends up in a clash with sysfs */
+	list_for_each_entry(connector, &dev->mode_config.connector_list, head)
+		drm_sysfs_connector_remove(connector);
+
+}
+EXPORT_SYMBOL(drm_connector_unplug_all);
+
+int drm_encoder_init(struct drm_device *dev,
 		      struct drm_encoder *encoder,
 		      const struct drm_encoder_funcs *funcs,
 		      int encoder_type)
 {
+	int ret;
+
 	mutex_lock(&dev->mode_config.mutex);
 
-	encoder->dev = dev;
+	ret = drm_mode_object_get(dev, &encoder->base, DRM_MODE_OBJECT_ENCODER);
+	if (ret)
+		goto out;
 
-	drm_mode_object_get(dev, &encoder->base, DRM_MODE_OBJECT_ENCODER);
+	encoder->dev = dev;
 	encoder->encoder_type = encoder_type;
 	encoder->funcs = funcs;
 
 	list_add_tail(&encoder->head, &dev->mode_config.encoder_list);
 	dev->mode_config.num_encoder++;
 
+ out:
 	mutex_unlock(&dev->mode_config.mutex);
+
+	return ret;
 }
 EXPORT_SYMBOL(drm_encoder_init);
 
@@ -560,18 +595,23 @@ int drm_plane_init(struct drm_device *dev, struct drm_plane *plane,
 		   const uint32_t *formats, uint32_t format_count,
 		   bool priv)
 {
+	int ret;
+
 	mutex_lock(&dev->mode_config.mutex);
 
+	ret = drm_mode_object_get(dev, &plane->base, DRM_MODE_OBJECT_PLANE);
+	if (ret)
+		goto out;
+
 	plane->dev = dev;
-	drm_mode_object_get(dev, &plane->base, DRM_MODE_OBJECT_PLANE);
 	plane->funcs = funcs;
 	plane->format_types = kmalloc(sizeof(uint32_t) * format_count,
 				      GFP_KERNEL);
 	if (!plane->format_types) {
 		DRM_DEBUG_KMS("out of memory when allocating plane\n");
 		drm_mode_object_put(dev, &plane->base);
-		mutex_unlock(&dev->mode_config.mutex);
-		return -ENOMEM;
+		ret = -ENOMEM;
+		goto out;
 	}
 
 	memcpy(plane->format_types, formats, format_count * sizeof(uint32_t));
@@ -589,9 +629,10 @@ int drm_plane_init(struct drm_device *dev, struct drm_plane *plane,
 		INIT_LIST_HEAD(&plane->head);
 	}
 
+ out:
 	mutex_unlock(&dev->mode_config.mutex);
 
-	return 0;
+	return ret;
 }
 EXPORT_SYMBOL(drm_plane_init);
 
@@ -631,7 +672,11 @@ struct drm_display_mode *drm_mode_create(struct drm_device *dev)
 	if (!nmode)
 		return NULL;
 
-	drm_mode_object_get(dev, &nmode->base, DRM_MODE_OBJECT_MODE);
+	if (drm_mode_object_get(dev, &nmode->base, DRM_MODE_OBJECT_MODE)) {
+		kfree(nmode);
+		return NULL;
+	}
+
 	return nmode;
 }
 EXPORT_SYMBOL(drm_mode_create);
@@ -648,6 +693,9 @@ EXPORT_SYMBOL(drm_mode_create);
  */
 void drm_mode_destroy(struct drm_device *dev, struct drm_display_mode *mode)
 {
+	if (!mode)
+		return;
+
 	drm_mode_object_put(dev, &mode->base);
 
 	kfree(mode);
@@ -658,7 +706,6 @@ static int drm_mode_create_standard_connector_properties(struct drm_device *dev)
 {
 	struct drm_property *edid;
 	struct drm_property *dpms;
-	int i;
 
 	/*
 	 * Standard properties (apply to all connectors)
@@ -668,11 +715,9 @@ static int drm_mode_create_standard_connector_properties(struct drm_device *dev)
 				   "EDID", 0);
 	dev->mode_config.edid_property = edid;
 
-	dpms = drm_property_create(dev, DRM_MODE_PROP_ENUM,
-				   "DPMS", ARRAY_SIZE(drm_dpms_enum_list));
-	for (i = 0; i < ARRAY_SIZE(drm_dpms_enum_list); i++)
-		drm_property_add_enum(dpms, i, drm_dpms_enum_list[i].type,
-				      drm_dpms_enum_list[i].name);
+	dpms = drm_property_create_enum(dev, 0,
+				   "DPMS", drm_dpms_enum_list,
+				   ARRAY_SIZE(drm_dpms_enum_list));
 	dev->mode_config.dpms_property = dpms;
 
 	return 0;
@@ -688,30 +733,21 @@ int drm_mode_create_dvi_i_properties(struct drm_device *dev)
 {
 	struct drm_property *dvi_i_selector;
 	struct drm_property *dvi_i_subconnector;
-	int i;
 
 	if (dev->mode_config.dvi_i_select_subconnector_property)
 		return 0;
 
 	dvi_i_selector =
-		drm_property_create(dev, DRM_MODE_PROP_ENUM,
+		drm_property_create_enum(dev, 0,
 				    "select subconnector",
+				    drm_dvi_i_select_enum_list,
 				    ARRAY_SIZE(drm_dvi_i_select_enum_list));
-	for (i = 0; i < ARRAY_SIZE(drm_dvi_i_select_enum_list); i++)
-		drm_property_add_enum(dvi_i_selector, i,
-				      drm_dvi_i_select_enum_list[i].type,
-				      drm_dvi_i_select_enum_list[i].name);
 	dev->mode_config.dvi_i_select_subconnector_property = dvi_i_selector;
 
-	dvi_i_subconnector =
-		drm_property_create(dev, DRM_MODE_PROP_ENUM |
-				    DRM_MODE_PROP_IMMUTABLE,
+	dvi_i_subconnector = drm_property_create_enum(dev, DRM_MODE_PROP_IMMUTABLE,
 				    "subconnector",
+				    drm_dvi_i_subconnector_enum_list,
 				    ARRAY_SIZE(drm_dvi_i_subconnector_enum_list));
-	for (i = 0; i < ARRAY_SIZE(drm_dvi_i_subconnector_enum_list); i++)
-		drm_property_add_enum(dvi_i_subconnector, i,
-				      drm_dvi_i_subconnector_enum_list[i].type,
-				      drm_dvi_i_subconnector_enum_list[i].name);
 	dev->mode_config.dvi_i_subconnector_property = dvi_i_subconnector;
 
 	return 0;
@@ -742,51 +778,33 @@ int drm_mode_create_tv_properties(struct drm_device *dev, int num_modes,
 	/*
 	 * Basic connector properties
 	 */
-	tv_selector = drm_property_create(dev, DRM_MODE_PROP_ENUM,
+	tv_selector = drm_property_create_enum(dev, 0,
 					  "select subconnector",
+					  drm_tv_select_enum_list,
 					  ARRAY_SIZE(drm_tv_select_enum_list));
-	for (i = 0; i < ARRAY_SIZE(drm_tv_select_enum_list); i++)
-		drm_property_add_enum(tv_selector, i,
-				      drm_tv_select_enum_list[i].type,
-				      drm_tv_select_enum_list[i].name);
 	dev->mode_config.tv_select_subconnector_property = tv_selector;
 
 	tv_subconnector =
-		drm_property_create(dev, DRM_MODE_PROP_ENUM |
-				    DRM_MODE_PROP_IMMUTABLE, "subconnector",
+		drm_property_create_enum(dev, DRM_MODE_PROP_IMMUTABLE,
+				    "subconnector",
+				    drm_tv_subconnector_enum_list,
 				    ARRAY_SIZE(drm_tv_subconnector_enum_list));
-	for (i = 0; i < ARRAY_SIZE(drm_tv_subconnector_enum_list); i++)
-		drm_property_add_enum(tv_subconnector, i,
-				      drm_tv_subconnector_enum_list[i].type,
-				      drm_tv_subconnector_enum_list[i].name);
 	dev->mode_config.tv_subconnector_property = tv_subconnector;
 
 	/*
 	 * Other, TV specific properties: margins & TV modes.
 	 */
 	dev->mode_config.tv_left_margin_property =
-		drm_property_create(dev, DRM_MODE_PROP_RANGE,
-				    "left margin", 2);
-	dev->mode_config.tv_left_margin_property->values[0] = 0;
-	dev->mode_config.tv_left_margin_property->values[1] = 100;
+		drm_property_create_range(dev, 0, "left margin", 0, 100);
 
 	dev->mode_config.tv_right_margin_property =
-		drm_property_create(dev, DRM_MODE_PROP_RANGE,
-				    "right margin", 2);
-	dev->mode_config.tv_right_margin_property->values[0] = 0;
-	dev->mode_config.tv_right_margin_property->values[1] = 100;
+		drm_property_create_range(dev, 0, "right margin", 0, 100);
 
 	dev->mode_config.tv_top_margin_property =
-		drm_property_create(dev, DRM_MODE_PROP_RANGE,
-				    "top margin", 2);
-	dev->mode_config.tv_top_margin_property->values[0] = 0;
-	dev->mode_config.tv_top_margin_property->values[1] = 100;
+		drm_property_create_range(dev, 0, "top margin", 0, 100);
 
 	dev->mode_config.tv_bottom_margin_property =
-		drm_property_create(dev, DRM_MODE_PROP_RANGE,
-				    "bottom margin", 2);
-	dev->mode_config.tv_bottom_margin_property->values[0] = 0;
-	dev->mode_config.tv_bottom_margin_property->values[1] = 100;
+		drm_property_create_range(dev, 0, "bottom margin", 0, 100);
 
 	dev->mode_config.tv_mode_property =
 		drm_property_create(dev, DRM_MODE_PROP_ENUM,
@@ -796,40 +814,22 @@ int drm_mode_create_tv_properties(struct drm_device *dev, int num_modes,
 				      i, modes[i]);
 
 	dev->mode_config.tv_brightness_property =
-		drm_property_create(dev, DRM_MODE_PROP_RANGE,
-				    "brightness", 2);
-	dev->mode_config.tv_brightness_property->values[0] = 0;
-	dev->mode_config.tv_brightness_property->values[1] = 100;
+		drm_property_create_range(dev, 0, "brightness", 0, 100);
 
 	dev->mode_config.tv_contrast_property =
-		drm_property_create(dev, DRM_MODE_PROP_RANGE,
-				    "contrast", 2);
-	dev->mode_config.tv_contrast_property->values[0] = 0;
-	dev->mode_config.tv_contrast_property->values[1] = 100;
+		drm_property_create_range(dev, 0, "contrast", 0, 100);
 
 	dev->mode_config.tv_flicker_reduction_property =
-		drm_property_create(dev, DRM_MODE_PROP_RANGE,
-				    "flicker reduction", 2);
-	dev->mode_config.tv_flicker_reduction_property->values[0] = 0;
-	dev->mode_config.tv_flicker_reduction_property->values[1] = 100;
+		drm_property_create_range(dev, 0, "flicker reduction", 0, 100);
 
 	dev->mode_config.tv_overscan_property =
-		drm_property_create(dev, DRM_MODE_PROP_RANGE,
-				    "overscan", 2);
-	dev->mode_config.tv_overscan_property->values[0] = 0;
-	dev->mode_config.tv_overscan_property->values[1] = 100;
+		drm_property_create_range(dev, 0, "overscan", 0, 100);
 
 	dev->mode_config.tv_saturation_property =
-		drm_property_create(dev, DRM_MODE_PROP_RANGE,
-				    "saturation", 2);
-	dev->mode_config.tv_saturation_property->values[0] = 0;
-	dev->mode_config.tv_saturation_property->values[1] = 100;
+		drm_property_create_range(dev, 0, "saturation", 0, 100);
 
 	dev->mode_config.tv_hue_property =
-		drm_property_create(dev, DRM_MODE_PROP_RANGE,
-				    "hue", 2);
-	dev->mode_config.tv_hue_property->values[0] = 0;
-	dev->mode_config.tv_hue_property->values[1] = 100;
+		drm_property_create_range(dev, 0, "hue", 0, 100);
 
 	return 0;
 }
@@ -845,18 +845,14 @@ EXPORT_SYMBOL(drm_mode_create_tv_properties);
 int drm_mode_create_scaling_mode_property(struct drm_device *dev)
 {
 	struct drm_property *scaling_mode;
-	int i;
 
 	if (dev->mode_config.scaling_mode_property)
 		return 0;
 
 	scaling_mode =
-		drm_property_create(dev, DRM_MODE_PROP_ENUM, "scaling mode",
+		drm_property_create_enum(dev, 0, "scaling mode",
+				drm_scaling_mode_enum_list,
 				    ARRAY_SIZE(drm_scaling_mode_enum_list));
-	for (i = 0; i < ARRAY_SIZE(drm_scaling_mode_enum_list); i++)
-		drm_property_add_enum(scaling_mode, i,
-				      drm_scaling_mode_enum_list[i].type,
-				      drm_scaling_mode_enum_list[i].name);
 
 	dev->mode_config.scaling_mode_property = scaling_mode;
 
@@ -874,18 +870,14 @@ EXPORT_SYMBOL(drm_mode_create_scaling_mode_property);
 int drm_mode_create_dithering_property(struct drm_device *dev)
 {
 	struct drm_property *dithering_mode;
-	int i;
 
 	if (dev->mode_config.dithering_mode_property)
 		return 0;
 
 	dithering_mode =
-		drm_property_create(dev, DRM_MODE_PROP_ENUM, "dithering",
+		drm_property_create_enum(dev, 0, "dithering",
+				drm_dithering_mode_enum_list,
 				    ARRAY_SIZE(drm_dithering_mode_enum_list));
-	for (i = 0; i < ARRAY_SIZE(drm_dithering_mode_enum_list); i++)
-		drm_property_add_enum(dithering_mode, i,
-				      drm_dithering_mode_enum_list[i].type,
-				      drm_dithering_mode_enum_list[i].name);
 	dev->mode_config.dithering_mode_property = dithering_mode;
 
 	return 0;
@@ -902,20 +894,15 @@ EXPORT_SYMBOL(drm_mode_create_dithering_property);
 int drm_mode_create_dirty_info_property(struct drm_device *dev)
 {
 	struct drm_property *dirty_info;
-	int i;
 
 	if (dev->mode_config.dirty_info_property)
 		return 0;
 
 	dirty_info =
-		drm_property_create(dev, DRM_MODE_PROP_ENUM |
-				    DRM_MODE_PROP_IMMUTABLE,
+		drm_property_create_enum(dev, DRM_MODE_PROP_IMMUTABLE,
 				    "dirty",
+				    drm_dirty_info_enum_list,
 				    ARRAY_SIZE(drm_dirty_info_enum_list));
-	for (i = 0; i < ARRAY_SIZE(drm_dirty_info_enum_list); i++)
-		drm_property_add_enum(dirty_info, i,
-				      drm_dirty_info_enum_list[i].type,
-				      drm_dirty_info_enum_list[i].name);
 	dev->mode_config.dirty_info_property = dirty_info;
 
 	return 0;
@@ -999,6 +986,7 @@ int drm_mode_group_init_legacy_group(struct drm_device *dev,
 
 	return 0;
 }
+EXPORT_SYMBOL(drm_mode_group_init_legacy_group);
 
 /**
  * drm_mode_config_cleanup - free up DRM mode_config info
@@ -1048,6 +1036,9 @@ void drm_mode_config_cleanup(struct drm_device *dev)
 				 head) {
 		plane->funcs->destroy(plane);
 	}
+
+	idr_remove_all(&dev->mode_config.crtc_idr);
+	idr_destroy(&dev->mode_config.crtc_idr);
 }
 EXPORT_SYMBOL(drm_mode_config_cleanup);
 
@@ -1062,9 +1053,16 @@ EXPORT_SYMBOL(drm_mode_config_cleanup);
  * Convert a drm_display_mode into a drm_mode_modeinfo structure to return to
  * the user.
  */
-void drm_crtc_convert_to_umode(struct drm_mode_modeinfo *out,
-			       struct drm_display_mode *in)
+static void drm_crtc_convert_to_umode(struct drm_mode_modeinfo *out,
+				      const struct drm_display_mode *in)
 {
+	WARN(in->hdisplay > USHRT_MAX || in->hsync_start > USHRT_MAX ||
+	     in->hsync_end > USHRT_MAX || in->htotal > USHRT_MAX ||
+	     in->hskew > USHRT_MAX || in->vdisplay > USHRT_MAX ||
+	     in->vsync_start > USHRT_MAX || in->vsync_end > USHRT_MAX ||
+	     in->vtotal > USHRT_MAX || in->vscan > USHRT_MAX,
+	     "timing values too large for mode info\n");
+
 	out->clock = in->clock;
 	out->hdisplay = in->hdisplay;
 	out->hsync_start = in->hsync_start;
@@ -1093,10 +1091,16 @@ void drm_crtc_convert_to_umode(struct drm_mode_modeinfo *out,
  *
  * Convert a drm_mode_modeinfo into a drm_display_mode structure to return to
  * the caller.
+ *
+ * RETURNS:
+ * Zero on success, errno on failure.
  */
-void drm_crtc_convert_umode(struct drm_display_mode *out,
-			    struct drm_mode_modeinfo *in)
+static int drm_crtc_convert_umode(struct drm_display_mode *out,
+				  const struct drm_mode_modeinfo *in)
 {
+	if (in->clock > INT_MAX || in->vrefresh > INT_MAX)
+		return -ERANGE;
+
 	out->clock = in->clock;
 	out->hdisplay = in->hdisplay;
 	out->hsync_start = in->hsync_start;
@@ -1113,6 +1117,8 @@ void drm_crtc_convert_umode(struct drm_display_mode *out,
 	out->type = in->type;
 	strncpy(out->name, in->name, DRM_DISPLAY_MODE_LEN);
 	out->name[DRM_DISPLAY_MODE_LEN-1] = 0;
+
+	return 0;
 }
 
 /**
@@ -1311,7 +1317,7 @@ out:
  * @arg: arg from ioctl
  *
  * LOCKING:
- * Caller? (FIXME)
+ * Takes mode config lock.
  *
  * Construct a CRTC configuration structure to return to the user.
  *
@@ -1371,7 +1377,7 @@ out:
  * @arg: arg from ioctl
  *
  * LOCKING:
- * Caller? (FIXME)
+ * Takes mode config lock.
  *
  * Construct a connector configuration structure to return to the user.
  *
@@ -1553,6 +1559,9 @@ out:
  * @data: ioctl data
  * @file_priv: DRM file info
  *
+ * LOCKING:
+ * Takes mode config lock.
+ *
  * Return an plane count and set of IDs.
  */
 int drm_mode_getplane_res(struct drm_device *dev, void *data,
@@ -1599,6 +1608,9 @@ out:
  * @data: ioctl data
  * @file_priv: DRM file info
  *
+ * LOCKING:
+ * Takes mode config lock.
+ *
  * Return plane info, including formats supported, gamma size, any
  * current fb, etc.
  */
@@ -1664,6 +1676,9 @@ out:
  * @data: ioctl data*
  * @file_prive: DRM file info
  *
+ * LOCKING:
+ * Takes mode config lock.
+ *
  * Set plane info, including placement, fb, scaling, and other factors.
  * Or pass a NULL fb to disable.
  */
@@ -1794,7 +1809,7 @@ out:
  * @arg: arg from ioctl
  *
  * LOCKING:
- * Caller? (FIXME)
+ * Takes mode config lock.
  *
  * Build a new CRTC configuration based on user request.
  *
@@ -1809,7 +1824,7 @@ int drm_mode_setcrtc(struct drm_device *dev, void *data,
 	struct drm_mode_config *config = &dev->mode_config;
 	struct drm_mode_crtc *crtc_req = data;
 	struct drm_mode_object *obj;
-	struct drm_crtc *crtc, *crtcfb;
+	struct drm_crtc *crtc;
 	struct drm_connector **connector_set = NULL, *connector;
 	struct drm_framebuffer *fb = NULL;
 	struct drm_display_mode *mode = NULL;
@@ -1821,6 +1836,10 @@ int drm_mode_setcrtc(struct drm_device *dev, void *data,
 	if (!drm_core_check_feature(dev, DRIVER_MODESET))
 		return -EINVAL;
 
+	/* For some reason crtc x/y offsets are signed internally. */
+	if (crtc_req->x > INT_MAX || crtc_req->y > INT_MAX)
+		return -ERANGE;
+
 	mutex_lock(&dev->mode_config.mutex);
 	obj = drm_mode_object_find(dev, crtc_req->crtc_id,
 				   DRM_MODE_OBJECT_CRTC);
@@ -1836,14 +1855,12 @@ int drm_mode_setcrtc(struct drm_device *dev, void *data,
 		/* If we have a mode we need a framebuffer. */
 		/* If we pass -1, set the mode with the currently bound fb */
 		if (crtc_req->fb_id == -1) {
-			list_for_each_entry(crtcfb,
-					    &dev->mode_config.crtc_list, head) {
-				if (crtcfb == crtc) {
-					DRM_DEBUG_KMS("Using current fb for "
-							"setmode\n");
-					fb = crtc->fb;
-				}
+			if (!crtc->fb) {
+				DRM_DEBUG_KMS("CRTC doesn't have current FB\n");
+				ret = -EINVAL;
+				goto out;
 			}
+			fb = crtc->fb;
 		} else {
 			obj = drm_mode_object_find(dev, crtc_req->fb_id,
 						   DRM_MODE_OBJECT_FB);
@@ -1857,8 +1874,30 @@ int drm_mode_setcrtc(struct drm_device *dev, void *data,
 		}
 
 		mode = drm_mode_create(dev);
-		drm_crtc_convert_umode(mode, &crtc_req->mode);
+		if (!mode) {
+			ret = -ENOMEM;
+			goto out;
+		}
+
+		ret = drm_crtc_convert_umode(mode, &crtc_req->mode);
+		if (ret) {
+			DRM_DEBUG_KMS("Invalid mode\n");
+			goto out;
+		}
+
 		drm_mode_set_crtcinfo(mode, CRTC_INTERLACE_HALVE_V);
+
+		if (mode->hdisplay > fb->width ||
+		    mode->vdisplay > fb->height ||
+		    crtc_req->x > fb->width - mode->hdisplay ||
+		    crtc_req->y > fb->height - mode->vdisplay) {
+			DRM_DEBUG_KMS("Invalid CRTC viewport %ux%u+%u+%u for fb size %ux%u.\n",
+				      mode->hdisplay, mode->vdisplay,
+				      crtc_req->x, crtc_req->y,
+				      fb->width, fb->height);
+			ret = -ENOSPC;
+			goto out;
+		}
 	}
 
 	if (crtc_req->count_connectors == 0 && mode) {
@@ -1926,6 +1965,7 @@ int drm_mode_setcrtc(struct drm_device *dev, void *data,
 
 out:
 	kfree(connector_set);
+	drm_mode_destroy(dev, mode);
 	mutex_unlock(&dev->mode_config.mutex);
 	return ret;
 }
@@ -2275,7 +2315,7 @@ out:
  * @arg: arg from ioctl
  *
  * LOCKING:
- * Caller? (FIXME)
+ * Takes mode config lock.
  *
  * Lookup the FB given its ID and return info about it.
  *
@@ -2424,38 +2464,48 @@ void drm_fb_release(struct drm_file *priv)
  *
  * Add @mode to @connector's user mode list.
  */
-static int drm_mode_attachmode(struct drm_device *dev,
-			       struct drm_connector *connector,
-			       struct drm_display_mode *mode)
+static void drm_mode_attachmode(struct drm_device *dev,
+				struct drm_connector *connector,
+				struct drm_display_mode *mode)
 {
-	int ret = 0;
-
 	list_add_tail(&mode->head, &connector->user_modes);
-	return ret;
 }
 
 int drm_mode_attachmode_crtc(struct drm_device *dev, struct drm_crtc *crtc,
-			     struct drm_display_mode *mode)
+			     const struct drm_display_mode *mode)
 {
 	struct drm_connector *connector;
 	int ret = 0;
-	struct drm_display_mode *dup_mode;
-	int need_dup = 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)
-			break;
+			continue;
 		if (connector->encoder->crtc == crtc) {
-			if (need_dup)
-				dup_mode = drm_mode_duplicate(dev, mode);
-			else
-				dup_mode = mode;
-			ret = drm_mode_attachmode(dev, connector, dup_mode);
-			if (ret)
-				return ret;
-			need_dup = 1;
+			dup_mode = drm_mode_duplicate(dev, mode);
+			if (!dup_mode) {
+				ret = -ENOMEM;
+				goto out;
+			}
+			list_add_tail(&dup_mode->head, &list);
 		}
 	}
-	return 0;
+
+	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);
 
@@ -2534,9 +2584,14 @@ int drm_mode_attachmode_ioctl(struct drm_device *dev,
 		goto out;
 	}
 
-	drm_crtc_convert_umode(mode, umode);
+	ret = drm_crtc_convert_umode(mode, umode);
+	if (ret) {
+		DRM_DEBUG_KMS("Invalid mode\n");
+		drm_mode_destroy(dev, mode);
+		goto out;
+	}
 
-	ret = drm_mode_attachmode(dev, connector, mode);
+	drm_mode_attachmode(dev, connector, mode);
 out:
 	mutex_unlock(&dev->mode_config.mutex);
 	return ret;
@@ -2577,7 +2632,12 @@ int drm_mode_detachmode_ioctl(struct drm_device *dev,
 	}
 	connector = obj_to_connector(obj);
 
-	drm_crtc_convert_umode(&mode, umode);
+	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:
 	mutex_unlock(&dev->mode_config.mutex);
@@ -2588,6 +2648,7 @@ struct drm_property *drm_property_create(struct drm_device *dev, int flags,
 					 const char *name, int num_values)
 {
 	struct drm_property *property = NULL;
+	int ret;
 
 	property = kzalloc(sizeof(struct drm_property), GFP_KERNEL);
 	if (!property)
@@ -2599,7 +2660,10 @@ struct drm_property *drm_property_create(struct drm_device *dev, int flags,
 			goto fail;
 	}
 
-	drm_mode_object_get(dev, &property->base, DRM_MODE_OBJECT_PROPERTY);
+	ret = drm_mode_object_get(dev, &property->base, DRM_MODE_OBJECT_PROPERTY);
+	if (ret)
+		goto fail;
+
 	property->flags = flags;
 	property->num_values = num_values;
 	INIT_LIST_HEAD(&property->enum_blob_list);
@@ -2612,11 +2676,59 @@ struct drm_property *drm_property_create(struct drm_device *dev, int flags,
 	list_add_tail(&property->head, &dev->mode_config.property_list);
 	return property;
 fail:
+	kfree(property->values);
 	kfree(property);
 	return NULL;
 }
 EXPORT_SYMBOL(drm_property_create);
 
+struct drm_property *drm_property_create_enum(struct drm_device *dev, int flags,
+					 const char *name,
+					 const struct drm_prop_enum_list *props,
+					 int num_values)
+{
+	struct drm_property *property;
+	int i, ret;
+
+	flags |= DRM_MODE_PROP_ENUM;
+
+	property = drm_property_create(dev, flags, name, num_values);
+	if (!property)
+		return NULL;
+
+	for (i = 0; i < num_values; i++) {
+		ret = drm_property_add_enum(property, i,
+				      props[i].type,
+				      props[i].name);
+		if (ret) {
+			drm_property_destroy(dev, property);
+			return NULL;
+		}
+	}
+
+	return property;
+}
+EXPORT_SYMBOL(drm_property_create_enum);
+
+struct drm_property *drm_property_create_range(struct drm_device *dev, int flags,
+					 const char *name,
+					 uint64_t min, uint64_t max)
+{
+	struct drm_property *property;
+
+	flags |= DRM_MODE_PROP_RANGE;
+
+	property = drm_property_create(dev, flags, name, 2);
+	if (!property)
+		return NULL;
+
+	property->values[0] = min;
+	property->values[1] = max;
+
+	return property;
+}
+EXPORT_SYMBOL(drm_property_create_range);
+
 int drm_property_add_enum(struct drm_property *property, int index,
 			  uint64_t value, const char *name)
 {
@@ -2828,6 +2940,7 @@ static struct drm_property_blob *drm_property_create_blob(struct drm_device *dev
 							  void *data)
 {
 	struct drm_property_blob *blob;
+	int ret;
 
 	if (!length || !data)
 		return NULL;
@@ -2836,13 +2949,16 @@ static struct drm_property_blob *drm_property_create_blob(struct drm_device *dev
 	if (!blob)
 		return NULL;
 
-	blob->data = (void *)((char *)blob + sizeof(struct drm_property_blob));
+	ret = drm_mode_object_get(dev, &blob->base, DRM_MODE_OBJECT_BLOB);
+	if (ret) {
+		kfree(blob);
+		return NULL;
+	}
+
 	blob->length = length;
 
 	memcpy(blob->data, data, length);
 
-	drm_mode_object_get(dev, &blob->base, DRM_MODE_OBJECT_BLOB);
-
 	list_add_tail(&blob->head, &dev->mode_config.property_blob_list);
 	return blob;
 }
@@ -3021,7 +3137,7 @@ void drm_mode_connector_detach_encoder(struct drm_connector *connector,
 }
 EXPORT_SYMBOL(drm_mode_connector_detach_encoder);
 
-bool drm_mode_crtc_set_gamma_size(struct drm_crtc *crtc,
+int drm_mode_crtc_set_gamma_size(struct drm_crtc *crtc,
 				  int gamma_size)
 {
 	crtc->gamma_size = gamma_size;
@@ -3029,10 +3145,10 @@ bool drm_mode_crtc_set_gamma_size(struct drm_crtc *crtc,
 	crtc->gamma_store = kzalloc(gamma_size * sizeof(uint16_t) * 3, GFP_KERNEL);
 	if (!crtc->gamma_store) {
 		crtc->gamma_size = 0;
-		return false;
+		return -ENOMEM;
 	}
 
-	return true;
+	return 0;
 }
 EXPORT_SYMBOL(drm_mode_crtc_set_gamma_size);
 
@@ -3178,6 +3294,18 @@ int drm_mode_page_flip_ioctl(struct drm_device *dev,
 		goto out;
 	fb = obj_to_fb(obj);
 
+	if (crtc->mode.hdisplay > fb->width ||
+	    crtc->mode.vdisplay > fb->height ||
+	    crtc->x > fb->width - crtc->mode.hdisplay ||
+	    crtc->y > fb->height - crtc->mode.vdisplay) {
+		DRM_DEBUG_KMS("Invalid fb size %ux%u for CRTC viewport %ux%u+%d+%d.\n",
+			      fb->width, fb->height,
+			      crtc->mode.hdisplay, crtc->mode.vdisplay,
+			      crtc->x, crtc->y);
+		ret = -ENOSPC;
+		goto out;
+	}
+
 	if (page_flip->flags & DRM_MODE_PAGE_FLIP_EVENT) {
 		ret = -ENOMEM;
 		spin_lock_irqsave(&dev->event_lock, flags);
diff --git a/drivers/gpu/drm/drm_crtc_helper.c b/drivers/gpu/drm/drm_crtc_helper.c
index 84a4a809793f..81118893264c 100644
--- a/drivers/gpu/drm/drm_crtc_helper.c
+++ b/drivers/gpu/drm/drm_crtc_helper.c
@@ -37,6 +37,7 @@
 #include "drm_fourcc.h"
 #include "drm_crtc_helper.h"
 #include "drm_fb_helper.h"
+#include "drm_edid.h"
 
 static bool drm_kms_helper_poll = true;
 module_param_named(poll, drm_kms_helper_poll, bool, 0600);
@@ -44,12 +45,12 @@ module_param_named(poll, drm_kms_helper_poll, bool, 0600);
 static void drm_mode_validate_flag(struct drm_connector *connector,
 				   int flags)
 {
-	struct drm_display_mode *mode, *t;
+	struct drm_display_mode *mode;
 
 	if (flags == (DRM_MODE_FLAG_DBLSCAN | DRM_MODE_FLAG_INTERLACE))
 		return;
 
-	list_for_each_entry_safe(mode, t, &connector->modes, head) {
+	list_for_each_entry(mode, &connector->modes, head) {
 		if ((mode->flags & DRM_MODE_FLAG_INTERLACE) &&
 				!(flags & DRM_MODE_FLAG_INTERLACE))
 			mode->status = MODE_NO_INTERLACE;
@@ -87,7 +88,7 @@ int drm_helper_probe_single_connector_modes(struct drm_connector *connector,
 					    uint32_t maxX, uint32_t maxY)
 {
 	struct drm_device *dev = connector->dev;
-	struct drm_display_mode *mode, *t;
+	struct drm_display_mode *mode;
 	struct drm_connector_helper_funcs *connector_funcs =
 		connector->helper_private;
 	int count = 0;
@@ -96,7 +97,7 @@ int drm_helper_probe_single_connector_modes(struct drm_connector *connector,
 	DRM_DEBUG_KMS("[CONNECTOR:%d:%s]\n", connector->base.id,
 			drm_get_connector_name(connector));
 	/* set all modes to the unverified state */
-	list_for_each_entry_safe(mode, t, &connector->modes, head)
+	list_for_each_entry(mode, &connector->modes, head)
 		mode->status = MODE_UNVERIFIED;
 
 	if (connector->force) {
@@ -118,7 +119,12 @@ int drm_helper_probe_single_connector_modes(struct drm_connector *connector,
 		goto prune;
 	}
 
-	count = (*connector_funcs->get_modes)(connector);
+#ifdef CONFIG_DRM_LOAD_EDID_FIRMWARE
+	count = drm_load_edid_firmware(connector);
+	if (count == 0)
+#endif
+		count = (*connector_funcs->get_modes)(connector);
+
 	if (count == 0 && connector->status == connector_status_connected)
 		count = drm_add_modes_noedid(connector, 1024, 768);
 	if (count == 0)
@@ -136,7 +142,7 @@ int drm_helper_probe_single_connector_modes(struct drm_connector *connector,
 		mode_flags |= DRM_MODE_FLAG_DBLSCAN;
 	drm_mode_validate_flag(connector, mode_flags);
 
-	list_for_each_entry_safe(mode, t, &connector->modes, head) {
+	list_for_each_entry(mode, &connector->modes, head) {
 		if (mode->status == MODE_OK)
 			mode->status = connector_funcs->mode_valid(connector,
 								   mode);
@@ -152,7 +158,7 @@ prune:
 
 	DRM_DEBUG_KMS("[CONNECTOR:%d:%s] probed modes :\n", connector->base.id,
 			drm_get_connector_name(connector));
-	list_for_each_entry_safe(mode, t, &connector->modes, head) {
+	list_for_each_entry(mode, &connector->modes, head) {
 		mode->vrefresh = drm_mode_vrefresh(mode);
 
 		drm_mode_set_crtcinfo(mode, CRTC_INTERLACE_HALVE_V);
@@ -352,6 +358,8 @@ bool drm_crtc_helper_set_mode(struct drm_crtc *crtc,
 		return true;
 
 	adjusted_mode = drm_mode_duplicate(dev, mode);
+	if (!adjusted_mode)
+		return false;
 
 	saved_hwmode = crtc->hwmode;
 	saved_mode = crtc->mode;
diff --git a/drivers/gpu/drm/drm_drv.c b/drivers/gpu/drm/drm_drv.c
index ebf7d3f68fc4..6116e3b75393 100644
--- a/drivers/gpu/drm/drm_drv.c
+++ b/drivers/gpu/drm/drm_drv.c
@@ -135,23 +135,27 @@ static struct drm_ioctl_desc drm_ioctls[] = {
 	DRM_IOCTL_DEF(DRM_IOCTL_GEM_FLINK, drm_gem_flink_ioctl, DRM_AUTH|DRM_UNLOCKED),
 	DRM_IOCTL_DEF(DRM_IOCTL_GEM_OPEN, drm_gem_open_ioctl, DRM_AUTH|DRM_UNLOCKED),
 
-	DRM_IOCTL_DEF(DRM_IOCTL_MODE_GETRESOURCES, drm_mode_getresources, DRM_MASTER|DRM_CONTROL_ALLOW|DRM_UNLOCKED),
+	DRM_IOCTL_DEF(DRM_IOCTL_MODE_GETRESOURCES, drm_mode_getresources, DRM_CONTROL_ALLOW|DRM_UNLOCKED),
+
+	DRM_IOCTL_DEF(DRM_IOCTL_PRIME_HANDLE_TO_FD, drm_prime_handle_to_fd_ioctl, DRM_AUTH|DRM_UNLOCKED),
+	DRM_IOCTL_DEF(DRM_IOCTL_PRIME_FD_TO_HANDLE, drm_prime_fd_to_handle_ioctl, DRM_AUTH|DRM_UNLOCKED),
+
 	DRM_IOCTL_DEF(DRM_IOCTL_MODE_GETPLANERESOURCES, drm_mode_getplane_res, DRM_MASTER|DRM_CONTROL_ALLOW|DRM_UNLOCKED),
-	DRM_IOCTL_DEF(DRM_IOCTL_MODE_GETCRTC, drm_mode_getcrtc, DRM_MASTER|DRM_CONTROL_ALLOW|DRM_UNLOCKED),
+	DRM_IOCTL_DEF(DRM_IOCTL_MODE_GETCRTC, drm_mode_getcrtc, DRM_CONTROL_ALLOW|DRM_UNLOCKED),
 	DRM_IOCTL_DEF(DRM_IOCTL_MODE_SETCRTC, drm_mode_setcrtc, DRM_MASTER|DRM_CONTROL_ALLOW|DRM_UNLOCKED),
 	DRM_IOCTL_DEF(DRM_IOCTL_MODE_GETPLANE, drm_mode_getplane, DRM_MASTER|DRM_CONTROL_ALLOW|DRM_UNLOCKED),
 	DRM_IOCTL_DEF(DRM_IOCTL_MODE_SETPLANE, drm_mode_setplane, DRM_MASTER|DRM_CONTROL_ALLOW|DRM_UNLOCKED),
 	DRM_IOCTL_DEF(DRM_IOCTL_MODE_CURSOR, drm_mode_cursor_ioctl, DRM_MASTER|DRM_CONTROL_ALLOW|DRM_UNLOCKED),
-	DRM_IOCTL_DEF(DRM_IOCTL_MODE_GETGAMMA, drm_mode_gamma_get_ioctl, DRM_MASTER|DRM_UNLOCKED),
+	DRM_IOCTL_DEF(DRM_IOCTL_MODE_GETGAMMA, drm_mode_gamma_get_ioctl, DRM_UNLOCKED),
 	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_MASTER|DRM_CONTROL_ALLOW|DRM_UNLOCKED),
-	DRM_IOCTL_DEF(DRM_IOCTL_MODE_GETCONNECTOR, drm_mode_getconnector, DRM_MASTER|DRM_CONTROL_ALLOW|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_GETPROPERTY, drm_mode_getproperty_ioctl, DRM_MASTER | 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_MASTER|DRM_CONTROL_ALLOW|DRM_UNLOCKED),
-	DRM_IOCTL_DEF(DRM_IOCTL_MODE_GETFB, drm_mode_getfb, DRM_MASTER|DRM_CONTROL_ALLOW|DRM_UNLOCKED),
+	DRM_IOCTL_DEF(DRM_IOCTL_MODE_GETFB, drm_mode_getfb, DRM_CONTROL_ALLOW|DRM_UNLOCKED),
 	DRM_IOCTL_DEF(DRM_IOCTL_MODE_ADDFB, drm_mode_addfb, DRM_MASTER|DRM_CONTROL_ALLOW|DRM_UNLOCKED),
 	DRM_IOCTL_DEF(DRM_IOCTL_MODE_ADDFB2, drm_mode_addfb2, DRM_MASTER|DRM_CONTROL_ALLOW|DRM_UNLOCKED),
 	DRM_IOCTL_DEF(DRM_IOCTL_MODE_RMFB, drm_mode_rmfb, DRM_MASTER|DRM_CONTROL_ALLOW|DRM_UNLOCKED),
@@ -390,6 +394,10 @@ long drm_ioctl(struct file *filp,
 	unsigned int usize, asize;
 
 	dev = file_priv->minor->dev;
+
+	if (drm_device_is_unplugged(dev))
+		return -ENODEV;
+
 	atomic_inc(&dev->ioctl_count);
 	atomic_inc(&dev->counts[_DRM_STAT_IOCTLS]);
 	++file_priv->ioctl_count;
diff --git a/drivers/gpu/drm/drm_edid.c b/drivers/gpu/drm/drm_edid.c
index ece03fc2d386..5a18b0df8285 100644
--- a/drivers/gpu/drm/drm_edid.c
+++ b/drivers/gpu/drm/drm_edid.c
@@ -149,8 +149,7 @@ EXPORT_SYMBOL(drm_edid_header_is_valid);
  * Sanity check the EDID block (base or extension).  Return 0 if the block
  * doesn't check out, or 1 if it's valid.
  */
-static bool
-drm_edid_block_valid(u8 *raw_edid)
+bool drm_edid_block_valid(u8 *raw_edid)
 {
 	int i;
 	u8 csum = 0;
@@ -203,6 +202,7 @@ bad:
 	}
 	return 0;
 }
+EXPORT_SYMBOL(drm_edid_block_valid);
 
 /**
  * drm_edid_is_valid - sanity check EDID data
@@ -226,7 +226,6 @@ bool drm_edid_is_valid(struct edid *edid)
 }
 EXPORT_SYMBOL(drm_edid_is_valid);
 
-#define DDC_ADDR 0x50
 #define DDC_SEGMENT_ADDR 0x30
 /**
  * Get EDID information via I2C.
@@ -266,6 +265,11 @@ drm_do_probe_ddc_edid(struct i2c_adapter *adapter, unsigned char *buf,
 			}
 		};
 		ret = i2c_transfer(adapter, msgs, 2);
+		if (ret == -ENXIO) {
+			DRM_DEBUG_KMS("drm: skipping non-existent adapter %s\n",
+					adapter->name);
+			break;
+		}
 	} while (ret != 2 && --retries);
 
 	return ret == 2 ? 0 : -1;
@@ -745,7 +749,7 @@ drm_mode_std(struct drm_connector *connector, struct edid *edid,
 		 */
 		mode = drm_gtf_mode(dev, hsize, vsize, vrefresh_rate, 0, 0);
 		if (drm_mode_hsync(mode) > drm_gtf2_hbreak(edid)) {
-			kfree(mode);
+			drm_mode_destroy(dev, mode);
 			mode = drm_gtf_mode_complex(dev, hsize, vsize,
 						    vrefresh_rate, 0, 0,
 						    drm_gtf2_m(edid),
diff --git a/drivers/gpu/drm/drm_edid_load.c b/drivers/gpu/drm/drm_edid_load.c
new file mode 100644
index 000000000000..da9acba2dd6c
--- /dev/null
+++ b/drivers/gpu/drm/drm_edid_load.c
@@ -0,0 +1,250 @@
+/*
+   drm_edid_load.c: use a built-in EDID data set or load it via the firmware
+		    interface
+
+   Copyright (C) 2012 Carsten Emde <C.Emde@osadl.org>
+
+   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 St, Fifth Floor, Boston, MA  02110-1301  USA.
+*/
+
+#include <linux/module.h>
+#include <linux/firmware.h>
+#include "drmP.h"
+#include "drm_crtc.h"
+#include "drm_crtc_helper.h"
+#include "drm_edid.h"
+
+static char edid_firmware[PATH_MAX];
+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
+static char *generic_edid_name[GENERIC_EDIDS] = {
+	"edid/1024x768.bin",
+	"edid/1280x1024.bin",
+	"edid/1680x1050.bin",
+	"edid/1920x1080.bin",
+};
+
+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, 0x23, 0x1a, 0x78,
+	0xea, 0x5e, 0xc0, 0xa4, 0x59, 0x4a, 0x98, 0x25,
+	0x20, 0x50, 0x54, 0x00, 0x08, 0x00, 0x61, 0x40,
+	0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01,
+	0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x64, 0x19,
+	0x00, 0x40, 0x41, 0x00, 0x26, 0x30, 0x08, 0x90,
+	0x36, 0x00, 0x63, 0x0a, 0x11, 0x00, 0x00, 0x18,
+	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, 0x2f, 0x31, 0x07, 0x00, 0x0a, 0x20, 0x20,
+	0x20, 0x20, 0x20, 0x20, 0x00, 0x00, 0x00, 0xfc,
+	0x00, 0x4c, 0x69, 0x6e, 0x75, 0x78, 0x20, 0x58,
+	0x47, 0x41, 0x0a, 0x20, 0x20, 0x20, 0x00, 0x55,
+	},
+	{
+	0x00, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0x00,
+	0x31, 0xd8, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+	0x05, 0x16, 0x01, 0x03, 0x6d, 0x2c, 0x23, 0x78,
+	0xea, 0x5e, 0xc0, 0xa4, 0x59, 0x4a, 0x98, 0x25,
+	0x20, 0x50, 0x54, 0x00, 0x00, 0x00, 0x81, 0x80,
+	0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01,
+	0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x30, 0x2a,
+	0x00, 0x98, 0x51, 0x00, 0x2a, 0x40, 0x30, 0x70,
+	0x13, 0x00, 0xbc, 0x63, 0x11, 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, 0x3e, 0x40, 0x0b, 0x00, 0x0a, 0x20, 0x20,
+	0x20, 0x20, 0x20, 0x20, 0x00, 0x00, 0x00, 0xfc,
+	0x00, 0x4c, 0x69, 0x6e, 0x75, 0x78, 0x20, 0x53,
+	0x58, 0x47, 0x41, 0x0a, 0x20, 0x20, 0x00, 0xa0,
+	},
+	{
+	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,
+	0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01,
+	0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x21, 0x39,
+	0x90, 0x30, 0x62, 0x1a, 0x27, 0x40, 0x68, 0xb0,
+	0x36, 0x00, 0xb5, 0x11, 0x11, 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, 0x40, 0x42, 0x0f, 0x00, 0x0a, 0x20, 0x20,
+	0x20, 0x20, 0x20, 0x20, 0x00, 0x00, 0x00, 0xfc,
+	0x00, 0x4c, 0x69, 0x6e, 0x75, 0x78, 0x20, 0x57,
+	0x53, 0x58, 0x47, 0x41, 0x0a, 0x20, 0x00, 0x26,
+	},
+	{
+	0x00, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0x00,
+	0x31, 0xd8, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+	0x05, 0x16, 0x01, 0x03, 0x6d, 0x32, 0x1c, 0x78,
+	0xea, 0x5e, 0xc0, 0xa4, 0x59, 0x4a, 0x98, 0x25,
+	0x20, 0x50, 0x54, 0x00, 0x00, 0x00, 0xd1, 0xc0,
+	0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01,
+	0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x02, 0x3a,
+	0x80, 0x18, 0x71, 0x38, 0x2d, 0x40, 0x58, 0x2c,
+	0x45, 0x00, 0xf4, 0x19, 0x11, 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, 0x42, 0x44, 0x0f, 0x00, 0x0a, 0x20, 0x20,
+	0x20, 0x20, 0x20, 0x20, 0x00, 0x00, 0x00, 0xfc,
+	0x00, 0x4c, 0x69, 0x6e, 0x75, 0x78, 0x20, 0x46,
+	0x48, 0x44, 0x0a, 0x20, 0x20, 0x20, 0x00, 0x05,
+	},
+};
+
+static int edid_load(struct drm_connector *connector, char *name,
+		     char *connector_name)
+{
+	const struct firmware *fw;
+	struct platform_device *pdev;
+	u8 *fwdata = NULL, *edid;
+	int fwsize, expected;
+	int builtin = 0, err = 0;
+	int i, valid_extensions = 0;
+
+	pdev = platform_device_register_simple(connector_name, -1, NULL, 0);
+	if (IS_ERR(pdev)) {
+		DRM_ERROR("Failed to register EDID firmware platform device "
+		    "for connector \"%s\"\n", connector_name);
+		err = -EINVAL;
+		goto out;
+	}
+
+	err = request_firmware(&fw, name, &pdev->dev);
+	platform_device_unregister(pdev);
+
+	if (err) {
+		i = 0;
+		while (i < GENERIC_EDIDS && strcmp(name, generic_edid_name[i]))
+			i++;
+		if (i < GENERIC_EDIDS) {
+			err = 0;
+			builtin = 1;
+			fwdata = generic_edid[i];
+			fwsize = sizeof(generic_edid[i]);
+		}
+	}
+
+	if (err) {
+		DRM_ERROR("Requesting EDID firmware \"%s\" failed (err=%d)\n",
+		    name, err);
+		goto out;
+	}
+
+	if (fwdata == NULL) {
+		fwdata = (u8 *) fw->data;
+		fwsize = fw->size;
+	}
+
+	expected = (fwdata[0x7e] + 1) * EDID_LENGTH;
+	if (expected != fwsize) {
+		DRM_ERROR("Size of EDID firmware \"%s\" is invalid "
+		    "(expected %d, got %d)\n", name, expected, (int) fwsize);
+		err = -EINVAL;
+		goto relfw_out;
+	}
+
+	edid = kmalloc(fwsize, GFP_KERNEL);
+	if (edid == NULL) {
+		err = -ENOMEM;
+		goto relfw_out;
+	}
+	memcpy(edid, fwdata, fwsize);
+
+	if (!drm_edid_block_valid(edid)) {
+		DRM_ERROR("Base block of EDID firmware \"%s\" is invalid ",
+		    name);
+		kfree(edid);
+		err = -EINVAL;
+		goto relfw_out;
+	}
+
+	for (i = 1; i <= edid[0x7e]; i++) {
+		if (i != valid_extensions + 1)
+			memcpy(edid + (valid_extensions + 1) * EDID_LENGTH,
+			    edid + i * EDID_LENGTH, EDID_LENGTH);
+		if (drm_edid_block_valid(edid + i * EDID_LENGTH))
+			valid_extensions++;
+	}
+
+	if (valid_extensions != edid[0x7e]) {
+		edid[EDID_LENGTH-1] += edid[0x7e] - valid_extensions;
+		DRM_INFO("Found %d valid extensions instead of %d in EDID data "
+		    "\"%s\" for connector \"%s\"\n", valid_extensions,
+		    edid[0x7e], name, connector_name);
+		edid[0x7e] = valid_extensions;
+		edid = krealloc(edid, (valid_extensions + 1) * EDID_LENGTH,
+		    GFP_KERNEL);
+		if (edid == NULL) {
+			err = -ENOMEM;
+			goto relfw_out;
+		}
+	}
+
+	connector->display_info.raw_edid = edid;
+	DRM_INFO("Got %s EDID base block and %d extension%s from "
+	    "\"%s\" for connector \"%s\"\n", builtin ? "built-in" :
+	    "external", valid_extensions, valid_extensions == 1 ? "" : "s",
+	    name, connector_name);
+
+relfw_out:
+	release_firmware(fw);
+
+out:
+	return err;
+}
+
+int drm_load_edid_firmware(struct drm_connector *connector)
+{
+	char *connector_name = drm_get_connector_name(connector);
+	char *edidname = edid_firmware, *last, *colon;
+	int ret = 0;
+
+	if (*edidname == '\0')
+		return ret;
+
+	colon = strchr(edidname, ':');
+	if (colon != NULL) {
+		if (strncmp(connector_name, edidname, colon - edidname))
+			return ret;
+		edidname = colon + 1;
+		if (*edidname == '\0')
+			return ret;
+	}
+
+	last = edidname + strlen(edidname) - 1;
+	if (*last == '\n')
+		*last = '\0';
+
+	ret = edid_load(connector, edidname, connector_name);
+	if (ret)
+		return 0;
+
+	drm_mode_connector_update_edid_property(connector,
+	    (struct edid *) connector->display_info.raw_edid);
+
+	return drm_add_edid_modes(connector, (struct edid *)
+	    connector->display_info.raw_edid);
+}
diff --git a/drivers/gpu/drm/drm_fb_helper.c b/drivers/gpu/drm/drm_fb_helper.c
index aada26f63dec..a0d6e894d97c 100644
--- a/drivers/gpu/drm/drm_fb_helper.c
+++ b/drivers/gpu/drm/drm_fb_helper.c
@@ -306,91 +306,31 @@ static struct sysrq_key_op sysrq_drm_fb_helper_restore_op = {
 static struct sysrq_key_op sysrq_drm_fb_helper_restore_op = { };
 #endif
 
-static void drm_fb_helper_on(struct fb_info *info)
+static void drm_fb_helper_dpms(struct fb_info *info, int dpms_mode)
 {
 	struct drm_fb_helper *fb_helper = info->par;
 	struct drm_device *dev = fb_helper->dev;
 	struct drm_crtc *crtc;
-	struct drm_crtc_helper_funcs *crtc_funcs;
-	struct drm_connector *connector;
-	struct drm_encoder *encoder;
-	int i, j;
-
-	/*
-	 * For each CRTC in this fb, turn the crtc on then,
-	 * find all associated encoders and turn them on.
-	 */
-	mutex_lock(&dev->mode_config.mutex);
-	for (i = 0; i < fb_helper->crtc_count; i++) {
-		crtc = fb_helper->crtc_info[i].mode_set.crtc;
-		crtc_funcs = crtc->helper_private;
-
-		if (!crtc->enabled)
-			continue;
-
-		crtc_funcs->dpms(crtc, DRM_MODE_DPMS_ON);
-
-		/* Walk the connectors & encoders on this fb turning them on */
-		for (j = 0; j < fb_helper->connector_count; j++) {
-			connector = fb_helper->connector_info[j]->connector;
-			connector->dpms = DRM_MODE_DPMS_ON;
-			drm_connector_property_set_value(connector,
-							 dev->mode_config.dpms_property,
-							 DRM_MODE_DPMS_ON);
-		}
-		/* Found a CRTC on this fb, now find encoders */
-		list_for_each_entry(encoder, &dev->mode_config.encoder_list, head) {
-			if (encoder->crtc == crtc) {
-				struct drm_encoder_helper_funcs *encoder_funcs;
-
-				encoder_funcs = encoder->helper_private;
-				encoder_funcs->dpms(encoder, DRM_MODE_DPMS_ON);
-			}
-		}
-	}
-	mutex_unlock(&dev->mode_config.mutex);
-}
-
-static void drm_fb_helper_off(struct fb_info *info, int dpms_mode)
-{
-	struct drm_fb_helper *fb_helper = info->par;
-	struct drm_device *dev = fb_helper->dev;
-	struct drm_crtc *crtc;
-	struct drm_crtc_helper_funcs *crtc_funcs;
 	struct drm_connector *connector;
-	struct drm_encoder *encoder;
 	int i, j;
 
 	/*
-	 * For each CRTC in this fb, find all associated encoders
-	 * and turn them off, then turn off the CRTC.
+	 * For each CRTC in this fb, turn the connectors on/off.
 	 */
 	mutex_lock(&dev->mode_config.mutex);
 	for (i = 0; i < fb_helper->crtc_count; i++) {
 		crtc = fb_helper->crtc_info[i].mode_set.crtc;
-		crtc_funcs = crtc->helper_private;
 
 		if (!crtc->enabled)
 			continue;
 
-		/* Walk the connectors on this fb and mark them off */
+		/* Walk the connectors & encoders on this fb turning them on/off */
 		for (j = 0; j < fb_helper->connector_count; j++) {
 			connector = fb_helper->connector_info[j]->connector;
-			connector->dpms = dpms_mode;
+			drm_helper_connector_dpms(connector, dpms_mode);
 			drm_connector_property_set_value(connector,
-							 dev->mode_config.dpms_property,
-							 dpms_mode);
-		}
-		/* Found a CRTC on this fb, now find encoders */
-		list_for_each_entry(encoder, &dev->mode_config.encoder_list, head) {
-			if (encoder->crtc == crtc) {
-				struct drm_encoder_helper_funcs *encoder_funcs;
-
-				encoder_funcs = encoder->helper_private;
-				encoder_funcs->dpms(encoder, dpms_mode);
-			}
+				dev->mode_config.dpms_property, dpms_mode);
 		}
-		crtc_funcs->dpms(crtc, DRM_MODE_DPMS_OFF);
 	}
 	mutex_unlock(&dev->mode_config.mutex);
 }
@@ -400,23 +340,23 @@ int drm_fb_helper_blank(int blank, struct fb_info *info)
 	switch (blank) {
 	/* Display: On; HSync: On, VSync: On */
 	case FB_BLANK_UNBLANK:
-		drm_fb_helper_on(info);
+		drm_fb_helper_dpms(info, DRM_MODE_DPMS_ON);
 		break;
 	/* Display: Off; HSync: On, VSync: On */
 	case FB_BLANK_NORMAL:
-		drm_fb_helper_off(info, DRM_MODE_DPMS_STANDBY);
+		drm_fb_helper_dpms(info, DRM_MODE_DPMS_STANDBY);
 		break;
 	/* Display: Off; HSync: Off, VSync: On */
 	case FB_BLANK_HSYNC_SUSPEND:
-		drm_fb_helper_off(info, DRM_MODE_DPMS_STANDBY);
+		drm_fb_helper_dpms(info, DRM_MODE_DPMS_STANDBY);
 		break;
 	/* Display: Off; HSync: On, VSync: Off */
 	case FB_BLANK_VSYNC_SUSPEND:
-		drm_fb_helper_off(info, DRM_MODE_DPMS_SUSPEND);
+		drm_fb_helper_dpms(info, DRM_MODE_DPMS_SUSPEND);
 		break;
 	/* Display: Off; HSync: Off, VSync: Off */
 	case FB_BLANK_POWERDOWN:
-		drm_fb_helper_off(info, DRM_MODE_DPMS_OFF);
+		drm_fb_helper_dpms(info, DRM_MODE_DPMS_OFF);
 		break;
 	}
 	return 0;
@@ -430,8 +370,11 @@ static void drm_fb_helper_crtc_free(struct drm_fb_helper *helper)
 	for (i = 0; i < helper->connector_count; i++)
 		kfree(helper->connector_info[i]);
 	kfree(helper->connector_info);
-	for (i = 0; i < helper->crtc_count; i++)
+	for (i = 0; i < helper->crtc_count; i++) {
 		kfree(helper->crtc_info[i].mode_set.connectors);
+		if (helper->crtc_info[i].mode_set.mode)
+			drm_mode_destroy(helper->dev, helper->crtc_info[i].mode_set.mode);
+	}
 	kfree(helper->crtc_info);
 }
 
@@ -474,11 +417,10 @@ int drm_fb_helper_init(struct drm_device *dev,
 
 	i = 0;
 	list_for_each_entry(crtc, &dev->mode_config.crtc_list, head) {
-		fb_helper->crtc_info[i].crtc_id = crtc->base.id;
 		fb_helper->crtc_info[i].mode_set.crtc = crtc;
 		i++;
 	}
-	fb_helper->conn_limit = max_conn_count;
+
 	return 0;
 out_free:
 	drm_fb_helper_crtc_free(fb_helper);
@@ -617,9 +559,13 @@ int drm_fb_helper_check_var(struct fb_var_screeninfo *var,
 		return -EINVAL;
 
 	/* Need to resize the fb object !!! */
-	if (var->bits_per_pixel > fb->bits_per_pixel || var->xres > fb->width || var->yres > fb->height) {
+	if (var->bits_per_pixel > fb->bits_per_pixel ||
+	    var->xres > fb->width || var->yres > fb->height ||
+	    var->xres_virtual > fb->width || var->yres_virtual > fb->height) {
 		DRM_DEBUG("fb userspace requested width/height/bpp is greater than current fb "
-			  "object %dx%d-%d > %dx%d-%d\n", var->xres, var->yres, var->bits_per_pixel,
+			  "request %dx%d-%d (virtual %dx%d) > %dx%d-%d\n",
+			  var->xres, var->yres, var->bits_per_pixel,
+			  var->xres_virtual, var->yres_virtual,
 			  fb->width, fb->height, fb->bits_per_pixel);
 		return -EINVAL;
 	}
diff --git a/drivers/gpu/drm/drm_fops.c b/drivers/gpu/drm/drm_fops.c
index 6263b0147598..cdfbf27b2b3c 100644
--- a/drivers/gpu/drm/drm_fops.c
+++ b/drivers/gpu/drm/drm_fops.c
@@ -133,6 +133,9 @@ int drm_open(struct inode *inode, struct file *filp)
 	if (!(dev = minor->dev))
 		return -ENODEV;
 
+	if (drm_device_is_unplugged(dev))
+		return -ENODEV;
+
 	retcode = drm_open_helper(inode, filp, dev);
 	if (!retcode) {
 		atomic_inc(&dev->counts[_DRM_STAT_OPENS]);
@@ -181,6 +184,9 @@ int drm_stub_open(struct inode *inode, struct file *filp)
 	if (!(dev = minor->dev))
 		goto out;
 
+	if (drm_device_is_unplugged(dev))
+		goto out;
+
 	old_fops = filp->f_op;
 	filp->f_op = fops_get(dev->driver->fops);
 	if (filp->f_op == NULL) {
@@ -265,6 +271,9 @@ static int drm_open_helper(struct inode *inode, struct file *filp,
 	if (dev->driver->driver_features & DRIVER_GEM)
 		drm_gem_open(dev, priv);
 
+	if (drm_core_check_feature(dev, DRIVER_PRIME))
+		drm_prime_init_file_private(&priv->prime);
+
 	if (dev->driver->open) {
 		ret = dev->driver->open(dev, priv);
 		if (ret < 0)
@@ -565,6 +574,10 @@ int drm_release(struct inode *inode, struct file *filp)
 
 	if (dev->driver->postclose)
 		dev->driver->postclose(dev, file_priv);
+
+	if (drm_core_check_feature(dev, DRIVER_PRIME))
+		drm_prime_destroy_file_private(&file_priv->prime);
+
 	kfree(file_priv);
 
 	/* ========================================================
@@ -579,6 +592,8 @@ int drm_release(struct inode *inode, struct file *filp)
 			retcode = -EBUSY;
 		} else
 			retcode = drm_lastclose(dev);
+		if (drm_device_is_unplugged(dev))
+			drm_put_dev(dev);
 	}
 	mutex_unlock(&drm_global_mutex);
 
diff --git a/drivers/gpu/drm/drm_gem.c b/drivers/gpu/drm/drm_gem.c
index f8625e290728..83114b5e3cee 100644
--- a/drivers/gpu/drm/drm_gem.c
+++ b/drivers/gpu/drm/drm_gem.c
@@ -35,6 +35,7 @@
 #include <linux/mman.h>
 #include <linux/pagemap.h>
 #include <linux/shmem_fs.h>
+#include <linux/dma-buf.h>
 #include "drmP.h"
 
 /** @file drm_gem.c
@@ -232,6 +233,10 @@ drm_gem_handle_delete(struct drm_file *filp, u32 handle)
 	idr_remove(&filp->object_idr, handle);
 	spin_unlock(&filp->table_lock);
 
+	if (obj->import_attach)
+		drm_prime_remove_imported_buf_handle(&filp->prime,
+				obj->import_attach->dmabuf);
+
 	if (dev->driver->gem_close_object)
 		dev->driver->gem_close_object(obj, filp);
 	drm_gem_object_handle_unreference_unlocked(obj);
@@ -527,6 +532,10 @@ drm_gem_object_release_handle(int id, void *ptr, void *data)
 	struct drm_gem_object *obj = ptr;
 	struct drm_device *dev = obj->dev;
 
+	if (obj->import_attach)
+		drm_prime_remove_imported_buf_handle(&file_priv->prime,
+				obj->import_attach->dmabuf);
+
 	if (dev->driver->gem_close_object)
 		dev->driver->gem_close_object(obj, file_priv);
 
@@ -661,6 +670,9 @@ int drm_gem_mmap(struct file *filp, struct vm_area_struct *vma)
 	struct drm_hash_item *hash;
 	int ret = 0;
 
+	if (drm_device_is_unplugged(dev))
+		return -ENODEV;
+
 	mutex_lock(&dev->struct_mutex);
 
 	if (drm_ht_find_item(&mm->offset_hash, vma->vm_pgoff, &hash)) {
@@ -700,7 +712,6 @@ int drm_gem_mmap(struct file *filp, struct vm_area_struct *vma)
 	 */
 	drm_gem_object_reference(obj);
 
-	vma->vm_file = filp;	/* Needed for drm_vm_open() */
 	drm_vm_open_locked(vma);
 
 out_unlock:
diff --git a/drivers/gpu/drm/drm_ioc32.c b/drivers/gpu/drm/drm_ioc32.c
index ddd70db45f76..637fcc3766c7 100644
--- a/drivers/gpu/drm/drm_ioc32.c
+++ b/drivers/gpu/drm/drm_ioc32.c
@@ -315,7 +315,8 @@ static int compat_drm_getclient(struct file *file, unsigned int cmd,
 	if (err)
 		return err;
 
-	if (__get_user(c32.auth, &client->auth)
+	if (__get_user(c32.idx, &client->idx)
+	    || __get_user(c32.auth, &client->auth)
 	    || __get_user(c32.pid, &client->pid)
 	    || __get_user(c32.uid, &client->uid)
 	    || __get_user(c32.magic, &client->magic)
diff --git a/drivers/gpu/drm/drm_ioctl.c b/drivers/gpu/drm/drm_ioctl.c
index 956fd38d7c9e..cf85155da2a0 100644
--- a/drivers/gpu/drm/drm_ioctl.c
+++ b/drivers/gpu/drm/drm_ioctl.c
@@ -37,6 +37,7 @@
 #include "drm_core.h"
 
 #include "linux/pci.h"
+#include "linux/export.h"
 
 /**
  * Get the bus id.
@@ -276,6 +277,12 @@ int drm_getcap(struct drm_device *dev, void *data, struct drm_file *file_priv)
 	case DRM_CAP_VBLANK_HIGH_CRTC:
 		req->value = 1;
 		break;
+	case DRM_CAP_DUMB_PREFERRED_DEPTH:
+		req->value = dev->mode_config.preferred_depth;
+		break;
+	case DRM_CAP_DUMB_PREFER_SHADOW:
+		req->value = dev->mode_config.prefer_shadow;
+		break;
 	default:
 		return -EINVAL;
 	}
@@ -346,3 +353,4 @@ int drm_noop(struct drm_device *dev, void *data,
 	DRM_DEBUG("\n");
 	return 0;
 }
+EXPORT_SYMBOL(drm_noop);
diff --git a/drivers/gpu/drm/drm_irq.c b/drivers/gpu/drm/drm_irq.c
index 44a5d0ad8b7c..c869436e238a 100644
--- a/drivers/gpu/drm/drm_irq.c
+++ b/drivers/gpu/drm/drm_irq.c
@@ -305,7 +305,7 @@ static void drm_irq_vgaarb_nokms(void *cookie, bool state)
  * \param dev DRM device.
  *
  * Initializes the IRQ related data. Installs the handler, calling the driver
- * \c drm_driver_irq_preinstall() and \c drm_driver_irq_postinstall() functions
+ * \c irq_preinstall() and \c irq_postinstall() functions
  * before and after the installation.
  */
 int drm_irq_install(struct drm_device *dev)
@@ -385,7 +385,7 @@ EXPORT_SYMBOL(drm_irq_install);
  *
  * \param dev DRM device.
  *
- * Calls the driver's \c drm_driver_irq_uninstall() function, and stops the irq.
+ * Calls the driver's \c irq_uninstall() function, and stops the irq.
  */
 int drm_irq_uninstall(struct drm_device *dev)
 {
diff --git a/drivers/gpu/drm/drm_memory.c b/drivers/gpu/drm/drm_memory.c
index c8b6b66d428d..c86a0f1a435c 100644
--- a/drivers/gpu/drm/drm_memory.c
+++ b/drivers/gpu/drm/drm_memory.c
@@ -37,25 +37,6 @@
 #include <linux/export.h>
 #include "drmP.h"
 
-/**
- * Called when "/proc/dri/%dev%/mem" is read.
- *
- * \param buf output buffer.
- * \param start start of output data.
- * \param offset requested start offset.
- * \param len requested number of bytes.
- * \param eof whether there is no more data to return.
- * \param data private data.
- * \return number of written bytes.
- *
- * No-op.
- */
-int drm_mem_info(char *buf, char **start, off_t offset,
-		 int len, int *eof, void *data)
-{
-	return 0;
-}
-
 #if __OS_HAS_AGP
 static void *agp_remap(unsigned long offset, unsigned long size,
 		       struct drm_device * dev)
diff --git a/drivers/gpu/drm/drm_modes.c b/drivers/gpu/drm/drm_modes.c
index fb8e46b4e8bc..b7adb4a967fd 100644
--- a/drivers/gpu/drm/drm_modes.c
+++ b/drivers/gpu/drm/drm_modes.c
@@ -686,8 +686,6 @@ void drm_mode_set_crtcinfo(struct drm_display_mode *p, int adjust_flags)
 			p->crtc_vsync_end /= 2;
 			p->crtc_vtotal /= 2;
 		}
-
-		p->crtc_vtotal |= 1;
 	}
 
 	if (p->flags & DRM_MODE_FLAG_DBLSCAN) {
@@ -716,6 +714,27 @@ EXPORT_SYMBOL(drm_mode_set_crtcinfo);
 
 
 /**
+ * drm_mode_copy - copy the mode
+ * @dst: mode to overwrite
+ * @src: mode to copy
+ *
+ * LOCKING:
+ * None.
+ *
+ * Copy an existing mode into another mode, preserving the object id
+ * of the destination mode.
+ */
+void drm_mode_copy(struct drm_display_mode *dst, const struct drm_display_mode *src)
+{
+	int id = dst->base.id;
+
+	*dst = *src;
+	dst->base.id = id;
+	INIT_LIST_HEAD(&dst->head);
+}
+EXPORT_SYMBOL(drm_mode_copy);
+
+/**
  * drm_mode_duplicate - allocate and duplicate an existing mode
  * @m: mode to duplicate
  *
@@ -729,16 +748,13 @@ struct drm_display_mode *drm_mode_duplicate(struct drm_device *dev,
 					    const struct drm_display_mode *mode)
 {
 	struct drm_display_mode *nmode;
-	int new_id;
 
 	nmode = drm_mode_create(dev);
 	if (!nmode)
 		return NULL;
 
-	new_id = nmode->base.id;
-	*nmode = *mode;
-	nmode->base.id = new_id;
-	INIT_LIST_HEAD(&nmode->head);
+	drm_mode_copy(nmode, mode);
+
 	return nmode;
 }
 EXPORT_SYMBOL(drm_mode_duplicate);
diff --git a/drivers/gpu/drm/drm_pci.c b/drivers/gpu/drm/drm_pci.c
index d4d10b7880cf..13f3d936472f 100644
--- a/drivers/gpu/drm/drm_pci.c
+++ b/drivers/gpu/drm/drm_pci.c
@@ -324,8 +324,6 @@ int drm_get_pci_dev(struct pci_dev *pdev, const struct pci_device_id *ent,
 	if (ret)
 		goto err_g1;
 
-	pci_set_master(pdev);
-
 	dev->pdev = pdev;
 	dev->dev = &pdev->dev;
 
diff --git a/drivers/gpu/drm/drm_platform.c b/drivers/gpu/drm/drm_platform.c
index ae9db5e2b27c..82431dcae37b 100644
--- a/drivers/gpu/drm/drm_platform.c
+++ b/drivers/gpu/drm/drm_platform.c
@@ -122,7 +122,7 @@ static const char *drm_platform_get_name(struct drm_device *dev)
 
 static int drm_platform_set_busid(struct drm_device *dev, struct drm_master *master)
 {
-	int len, ret;
+	int len, ret, id;
 
 	master->unique_len = 13 + strlen(dev->platformdev->name);
 	master->unique_size = master->unique_len;
@@ -131,8 +131,16 @@ static int drm_platform_set_busid(struct drm_device *dev, struct drm_master *mas
 	if (master->unique == NULL)
 		return -ENOMEM;
 
+	id = dev->platformdev->id;
+
+	/* if only a single instance of the platform device, id will be
+	 * set to -1.. use 0 instead to avoid a funny looking bus-id:
+	 */
+	if (id == -1)
+		id = 0;
+
 	len = snprintf(master->unique, master->unique_len,
-			"platform:%s:%02d", dev->platformdev->name, dev->platformdev->id);
+			"platform:%s:%02d", dev->platformdev->name, id);
 
 	if (len > master->unique_len) {
 		DRM_ERROR("Unique buffer overflowed\n");
diff --git a/drivers/gpu/drm/drm_prime.c b/drivers/gpu/drm/drm_prime.c
new file mode 100644
index 000000000000..1bdf2b54eaf6
--- /dev/null
+++ b/drivers/gpu/drm/drm_prime.c
@@ -0,0 +1,304 @@
+/*
+ * Copyright © 2012 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:
+ *      Dave Airlie <airlied@redhat.com>
+ *      Rob Clark <rob.clark@linaro.org>
+ *
+ */
+
+#include <linux/export.h>
+#include <linux/dma-buf.h>
+#include "drmP.h"
+
+/*
+ * DMA-BUF/GEM Object references and lifetime overview:
+ *
+ * On the export the dma_buf holds a reference to the exporting GEM
+ * object. It takes this reference in handle_to_fd_ioctl, when it
+ * first calls .prime_export and stores the exporting GEM object in
+ * the dma_buf priv. This reference is released when the dma_buf
+ * object goes away in the driver .release function.
+ *
+ * On the import the importing GEM object holds a reference to the
+ * dma_buf (which in turn holds a ref to the exporting GEM object).
+ * It takes that reference in the fd_to_handle ioctl.
+ * It calls dma_buf_get, creates an attachment to it and stores the
+ * attachment in the GEM object. When this attachment is destroyed
+ * when the imported object is destroyed, we remove the attachment
+ * and drop the reference to the dma_buf.
+ *
+ * Thus the chain of references always flows in one direction
+ * (avoiding loops): importing_gem -> dmabuf -> exporting_gem
+ *
+ * Self-importing: if userspace is using PRIME as a replacement for flink
+ * then it will get a fd->handle request for a GEM object that it created.
+ * Drivers should detect this situation and return back the gem object
+ * from the dma-buf private.
+ */
+
+struct drm_prime_member {
+	struct list_head entry;
+	struct dma_buf *dma_buf;
+	uint32_t handle;
+};
+
+int drm_gem_prime_handle_to_fd(struct drm_device *dev,
+		struct drm_file *file_priv, uint32_t handle, uint32_t flags,
+		int *prime_fd)
+{
+	struct drm_gem_object *obj;
+	void *buf;
+
+	obj = drm_gem_object_lookup(dev, file_priv, handle);
+	if (!obj)
+		return -ENOENT;
+
+	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;
+	}
+
+	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);
+	}
+	mutex_unlock(&file_priv->prime.lock);
+	return 0;
+}
+EXPORT_SYMBOL(drm_gem_prime_handle_to_fd);
+
+int drm_gem_prime_fd_to_handle(struct drm_device *dev,
+		struct drm_file *file_priv, int prime_fd, uint32_t *handle)
+{
+	struct dma_buf *dma_buf;
+	struct drm_gem_object *obj;
+	int ret;
+
+	dma_buf = dma_buf_get(prime_fd);
+	if (IS_ERR(dma_buf))
+		return PTR_ERR(dma_buf);
+
+	mutex_lock(&file_priv->prime.lock);
+
+	ret = drm_prime_lookup_imported_buf_handle(&file_priv->prime,
+			dma_buf, handle);
+	if (!ret) {
+		ret = 0;
+		goto out_put;
+	}
+
+	/* never seen this one, need to import */
+	obj = dev->driver->gem_prime_import(dev, dma_buf);
+	if (IS_ERR(obj)) {
+		ret = PTR_ERR(obj);
+		goto out_put;
+	}
+
+	ret = drm_gem_handle_create(file_priv, obj, handle);
+	drm_gem_object_unreference_unlocked(obj);
+	if (ret)
+		goto out_put;
+
+	ret = drm_prime_add_imported_buf_handle(&file_priv->prime,
+			dma_buf, *handle);
+	if (ret)
+		goto fail;
+
+	mutex_unlock(&file_priv->prime.lock);
+	return 0;
+
+fail:
+	/* hmm, if driver attached, we are relying on the free-object path
+	 * to detach.. which seems ok..
+	 */
+	drm_gem_object_handle_unreference_unlocked(obj);
+out_put:
+	dma_buf_put(dma_buf);
+	mutex_unlock(&file_priv->prime.lock);
+	return ret;
+}
+EXPORT_SYMBOL(drm_gem_prime_fd_to_handle);
+
+int drm_prime_handle_to_fd_ioctl(struct drm_device *dev, void *data,
+				 struct drm_file *file_priv)
+{
+	struct drm_prime_handle *args = data;
+	uint32_t flags;
+
+	if (!drm_core_check_feature(dev, DRIVER_PRIME))
+		return -EINVAL;
+
+	if (!dev->driver->prime_handle_to_fd)
+		return -ENOSYS;
+
+	/* check flags are valid */
+	if (args->flags & ~DRM_CLOEXEC)
+		return -EINVAL;
+
+	/* we only want to pass DRM_CLOEXEC which is == O_CLOEXEC */
+	flags = args->flags & DRM_CLOEXEC;
+
+	return dev->driver->prime_handle_to_fd(dev, file_priv,
+			args->handle, flags, &args->fd);
+}
+
+int drm_prime_fd_to_handle_ioctl(struct drm_device *dev, void *data,
+				 struct drm_file *file_priv)
+{
+	struct drm_prime_handle *args = data;
+
+	if (!drm_core_check_feature(dev, DRIVER_PRIME))
+		return -EINVAL;
+
+	if (!dev->driver->prime_fd_to_handle)
+		return -ENOSYS;
+
+	return dev->driver->prime_fd_to_handle(dev, file_priv,
+			args->fd, &args->handle);
+}
+
+/*
+ * drm_prime_pages_to_sg
+ *
+ * this helper creates an sg table object from a set of pages
+ * the driver is responsible for mapping the pages into the
+ * importers address space
+ */
+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);
+	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);
+	return NULL;
+}
+EXPORT_SYMBOL(drm_prime_pages_to_sg);
+
+/* helper function to cleanup a GEM/prime object */
+void drm_prime_gem_destroy(struct drm_gem_object *obj, struct sg_table *sg)
+{
+	struct dma_buf_attachment *attach;
+	struct dma_buf *dma_buf;
+	attach = obj->import_attach;
+	if (sg)
+		dma_buf_unmap_attachment(attach, sg, DMA_BIDIRECTIONAL);
+	dma_buf = attach->dmabuf;
+	dma_buf_detach(attach->dmabuf, attach);
+	/* remove the reference */
+	dma_buf_put(dma_buf);
+}
+EXPORT_SYMBOL(drm_prime_gem_destroy);
+
+void drm_prime_init_file_private(struct drm_prime_file_private *prime_fpriv)
+{
+	INIT_LIST_HEAD(&prime_fpriv->head);
+	mutex_init(&prime_fpriv->lock);
+}
+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);
+	}
+}
+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)
+{
+	struct drm_prime_member *member;
+
+	member = kmalloc(sizeof(*member), GFP_KERNEL);
+	if (!member)
+		return -ENOMEM;
+
+	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)
+{
+	struct drm_prime_member *member;
+
+	list_for_each_entry(member, &prime_fpriv->head, entry) {
+		if (member->dma_buf == dma_buf) {
+			*handle = member->handle;
+			return 0;
+		}
+	}
+	return -ENOENT;
+}
+EXPORT_SYMBOL(drm_prime_lookup_imported_buf_handle);
+
+void drm_prime_remove_imported_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) {
+			list_del(&member->entry);
+			kfree(member);
+		}
+	}
+	mutex_unlock(&prime_fpriv->lock);
+}
+EXPORT_SYMBOL(drm_prime_remove_imported_buf_handle);
diff --git a/drivers/gpu/drm/drm_stub.c b/drivers/gpu/drm/drm_stub.c
index 6d7b083c5b77..aa454f80e109 100644
--- a/drivers/gpu/drm/drm_stub.c
+++ b/drivers/gpu/drm/drm_stub.c
@@ -319,6 +319,7 @@ int drm_fill_in_dev(struct drm_device *dev,
 	drm_lastclose(dev);
 	return retcode;
 }
+EXPORT_SYMBOL(drm_fill_in_dev);
 
 
 /**
@@ -397,6 +398,7 @@ err_idr:
 	*minor = NULL;
 	return ret;
 }
+EXPORT_SYMBOL(drm_get_minor);
 
 /**
  * Put a secondary minor number.
@@ -428,6 +430,12 @@ int drm_put_minor(struct drm_minor **minor_p)
 	*minor_p = NULL;
 	return 0;
 }
+EXPORT_SYMBOL(drm_put_minor);
+
+static void drm_unplug_minor(struct drm_minor *minor)
+{
+	drm_sysfs_device_remove(minor);
+}
 
 /**
  * Called via drm_exit() at module unload time or when pci device is
@@ -492,3 +500,21 @@ void drm_put_dev(struct drm_device *dev)
 	kfree(dev);
 }
 EXPORT_SYMBOL(drm_put_dev);
+
+void drm_unplug_dev(struct drm_device *dev)
+{
+	/* for a USB device */
+	if (drm_core_check_feature(dev, DRIVER_MODESET))
+		drm_unplug_minor(dev->control);
+	drm_unplug_minor(dev->primary);
+
+	mutex_lock(&drm_global_mutex);
+
+	drm_device_set_unplugged(dev);
+
+	if (dev->open_count == 0) {
+		drm_put_dev(dev);
+	}
+	mutex_unlock(&drm_global_mutex);
+}
+EXPORT_SYMBOL(drm_unplug_dev);
diff --git a/drivers/gpu/drm/drm_sysfs.c b/drivers/gpu/drm/drm_sysfs.c
index 62c3675045ac..5a7bd51fc3d8 100644
--- a/drivers/gpu/drm/drm_sysfs.c
+++ b/drivers/gpu/drm/drm_sysfs.c
@@ -454,6 +454,8 @@ void drm_sysfs_connector_remove(struct drm_connector *connector)
 {
 	int i;
 
+	if (!connector->kdev.parent)
+		return;
 	DRM_DEBUG("removing \"%s\" from sysfs\n",
 		  drm_get_connector_name(connector));
 
@@ -461,6 +463,7 @@ void drm_sysfs_connector_remove(struct drm_connector *connector)
 		device_remove_file(&connector->kdev, &connector_attrs[i]);
 	sysfs_remove_bin_file(&connector->kdev.kobj, &edid_attr);
 	device_unregister(&connector->kdev);
+	connector->kdev.parent = NULL;
 }
 EXPORT_SYMBOL(drm_sysfs_connector_remove);
 
@@ -533,7 +536,9 @@ err_out:
  */
 void drm_sysfs_device_remove(struct drm_minor *minor)
 {
-	device_unregister(&minor->kdev);
+	if (minor->kdev.parent)
+		device_unregister(&minor->kdev);
+	minor->kdev.parent = NULL;
 }
 
 
diff --git a/drivers/gpu/drm/drm_usb.c b/drivers/gpu/drm/drm_usb.c
index 445003f4dc93..c8c83dad2ce1 100644
--- a/drivers/gpu/drm/drm_usb.c
+++ b/drivers/gpu/drm/drm_usb.c
@@ -2,7 +2,6 @@
 #include <linux/usb.h>
 #include <linux/export.h>
 
-#ifdef CONFIG_USB
 int drm_get_usb_dev(struct usb_interface *interface,
 		    const struct usb_device_id *id,
 		    struct drm_driver *driver)
@@ -115,4 +114,3 @@ void drm_usb_exit(struct drm_driver *driver,
 	usb_deregister(udriver);
 }
 EXPORT_SYMBOL(drm_usb_exit);
-#endif
diff --git a/drivers/gpu/drm/drm_vm.c b/drivers/gpu/drm/drm_vm.c
index 8c03eaf41448..149561818349 100644
--- a/drivers/gpu/drm/drm_vm.c
+++ b/drivers/gpu/drm/drm_vm.c
@@ -519,7 +519,6 @@ static int drm_mmap_dma(struct file *filp, struct vm_area_struct *vma)
 	vma->vm_flags |= VM_RESERVED;	/* Don't swap */
 	vma->vm_flags |= VM_DONTEXPAND;
 
-	vma->vm_file = filp;	/* Needed for drm_vm_open() */
 	drm_vm_open_locked(vma);
 	return 0;
 }
@@ -671,7 +670,6 @@ int drm_mmap_locked(struct file *filp, struct vm_area_struct *vma)
 	vma->vm_flags |= VM_RESERVED;	/* Don't swap */
 	vma->vm_flags |= VM_DONTEXPAND;
 
-	vma->vm_file = filp;	/* Needed for drm_vm_open() */
 	drm_vm_open_locked(vma);
 	return 0;
 }
@@ -682,6 +680,9 @@ int drm_mmap(struct file *filp, struct vm_area_struct *vma)
 	struct drm_device *dev = priv->minor->dev;
 	int ret;
 
+	if (drm_device_is_unplugged(dev))
+		return -ENODEV;
+
 	mutex_lock(&dev->struct_mutex);
 	ret = drm_mmap_locked(filp, vma);
 	mutex_unlock(&dev->struct_mutex);
diff --git a/drivers/gpu/drm/exynos/Kconfig b/drivers/gpu/drm/exynos/Kconfig
index b9e5266c341b..3343ac437fe5 100644
--- a/drivers/gpu/drm/exynos/Kconfig
+++ b/drivers/gpu/drm/exynos/Kconfig
@@ -1,7 +1,6 @@
 config DRM_EXYNOS
 	tristate "DRM Support for Samsung SoC EXYNOS Series"
 	depends on DRM && PLAT_SAMSUNG
-	default	n
 	select DRM_KMS_HELPER
 	select FB_CFB_FILLRECT
 	select FB_CFB_COPYAREA
@@ -12,16 +11,19 @@ config DRM_EXYNOS
 	  If M is selected the module will be called exynosdrm.
 
 config DRM_EXYNOS_FIMD
-	tristate "Exynos DRM FIMD"
+	bool "Exynos DRM FIMD"
 	depends on DRM_EXYNOS && !FB_S3C
-	default n
 	help
 	  Choose this option if you want to use Exynos FIMD for DRM.
-	  If M is selected, the module will be called exynos_drm_fimd
 
 config DRM_EXYNOS_HDMI
-	tristate "Exynos DRM HDMI"
+	bool "Exynos DRM HDMI"
 	depends on DRM_EXYNOS && !VIDEO_SAMSUNG_S5P_TV
 	help
 	  Choose this option if you want to use Exynos HDMI for DRM.
-	  If M is selected, the module will be called exynos_drm_hdmi
+
+config DRM_EXYNOS_VIDI
+	bool "Exynos DRM Virtual Display"
+	depends on DRM_EXYNOS
+	help
+	  Choose this option if you want to use Exynos VIDI for DRM.
diff --git a/drivers/gpu/drm/exynos/Makefile b/drivers/gpu/drm/exynos/Makefile
index 395e69c9a96e..9e0bff8badf9 100644
--- a/drivers/gpu/drm/exynos/Makefile
+++ b/drivers/gpu/drm/exynos/Makefile
@@ -8,7 +8,10 @@ exynosdrm-y := exynos_drm_drv.o exynos_drm_encoder.o exynos_drm_connector.o \
 		exynos_drm_buf.o exynos_drm_gem.o exynos_drm_core.o \
 		exynos_drm_plane.o
 
-obj-$(CONFIG_DRM_EXYNOS) += exynosdrm.o
-obj-$(CONFIG_DRM_EXYNOS_FIMD) += exynos_drm_fimd.o
-obj-$(CONFIG_DRM_EXYNOS_HDMI) += exynos_hdmi.o exynos_mixer.o exynos_ddc.o \
-				 exynos_hdmiphy.o exynos_drm_hdmi.o
+exynosdrm-$(CONFIG_DRM_EXYNOS_FIMD)	+= exynos_drm_fimd.o
+exynosdrm-$(CONFIG_DRM_EXYNOS_HDMI)	+= exynos_hdmi.o exynos_mixer.o \
+					   exynos_ddc.o exynos_hdmiphy.o \
+					   exynos_drm_hdmi.o
+exynosdrm-$(CONFIG_DRM_EXYNOS_VIDI)	+= exynos_drm_vidi.o
+
+obj-$(CONFIG_DRM_EXYNOS)		+= exynosdrm.o
diff --git a/drivers/gpu/drm/exynos/exynos_ddc.c b/drivers/gpu/drm/exynos/exynos_ddc.c
index 84b614fe26fd..7e1051d07f1f 100644
--- a/drivers/gpu/drm/exynos/exynos_ddc.c
+++ b/drivers/gpu/drm/exynos/exynos_ddc.c
@@ -55,4 +55,3 @@ struct i2c_driver ddc_driver = {
 	.remove		= __devexit_p(s5p_ddc_remove),
 	.command		= NULL,
 };
-EXPORT_SYMBOL(ddc_driver);
diff --git a/drivers/gpu/drm/exynos/exynos_drm_buf.c b/drivers/gpu/drm/exynos/exynos_drm_buf.c
index 3cf785c58186..4a3a5f72ed4a 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_buf.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_buf.c
@@ -25,45 +25,161 @@
 
 #include "drmP.h"
 #include "drm.h"
+#include "exynos_drm.h"
 
 #include "exynos_drm_drv.h"
 #include "exynos_drm_gem.h"
 #include "exynos_drm_buf.h"
 
 static int lowlevel_buffer_allocate(struct drm_device *dev,
-		struct exynos_drm_gem_buf *buffer)
+		unsigned int flags, struct exynos_drm_gem_buf *buf)
 {
+	dma_addr_t start_addr, end_addr;
+	unsigned int npages, page_size, i = 0;
+	struct scatterlist *sgl;
+	int ret = 0;
+
 	DRM_DEBUG_KMS("%s\n", __FILE__);
 
-	buffer->kvaddr = dma_alloc_writecombine(dev->dev, buffer->size,
-			&buffer->dma_addr, GFP_KERNEL);
-	if (!buffer->kvaddr) {
-		DRM_ERROR("failed to allocate buffer.\n");
+	if (flags & EXYNOS_BO_NONCONTIG) {
+		DRM_DEBUG_KMS("not support allocation type.\n");
+		return -EINVAL;
+	}
+
+	if (buf->dma_addr) {
+		DRM_DEBUG_KMS("already allocated.\n");
+		return 0;
+	}
+
+	if (buf->size >= SZ_1M) {
+		npages = (buf->size >> SECTION_SHIFT) + 1;
+		page_size = SECTION_SIZE;
+	} else if (buf->size >= SZ_64K) {
+		npages = (buf->size >> 16) + 1;
+		page_size = SZ_64K;
+	} else {
+		npages = (buf->size >> PAGE_SHIFT) + 1;
+		page_size = PAGE_SIZE;
+	}
+
+	buf->sgt = kzalloc(sizeof(struct sg_table), GFP_KERNEL);
+	if (!buf->sgt) {
+		DRM_ERROR("failed to allocate sg table.\n");
 		return -ENOMEM;
 	}
 
-	DRM_DEBUG_KMS("vaddr(0x%lx), dma_addr(0x%lx), size(0x%lx)\n",
-			(unsigned long)buffer->kvaddr,
-			(unsigned long)buffer->dma_addr,
-			buffer->size);
+	ret = sg_alloc_table(buf->sgt, npages, GFP_KERNEL);
+	if (ret < 0) {
+		DRM_ERROR("failed to initialize sg table.\n");
+		kfree(buf->sgt);
+		buf->sgt = NULL;
+		return -ENOMEM;
+	}
 
-	return 0;
+		buf->kvaddr = dma_alloc_writecombine(dev->dev, buf->size,
+				&buf->dma_addr, GFP_KERNEL);
+		if (!buf->kvaddr) {
+			DRM_ERROR("failed to allocate buffer.\n");
+			ret = -ENOMEM;
+			goto err1;
+		}
+
+		start_addr = buf->dma_addr;
+		end_addr = buf->dma_addr + buf->size;
+
+		buf->pages = kzalloc(sizeof(struct page) * npages, GFP_KERNEL);
+		if (!buf->pages) {
+			DRM_ERROR("failed to allocate pages.\n");
+			ret = -ENOMEM;
+			goto err2;
+		}
+
+	start_addr = buf->dma_addr;
+	end_addr = buf->dma_addr + buf->size;
+
+	buf->pages = kzalloc(sizeof(struct page) * npages, GFP_KERNEL);
+	if (!buf->pages) {
+		DRM_ERROR("failed to allocate pages.\n");
+		ret = -ENOMEM;
+		goto err2;
+	}
+
+	sgl = buf->sgt->sgl;
+
+	while (i < npages) {
+		buf->pages[i] = phys_to_page(start_addr);
+		sg_set_page(sgl, buf->pages[i], page_size, 0);
+		sg_dma_address(sgl) = start_addr;
+		start_addr += page_size;
+		if (end_addr - start_addr < page_size)
+			break;
+		sgl = sg_next(sgl);
+		i++;
+	}
+
+	buf->pages[i] = phys_to_page(start_addr);
+
+	sgl = sg_next(sgl);
+	sg_set_page(sgl, buf->pages[i+1], end_addr - start_addr, 0);
+
+	DRM_DEBUG_KMS("vaddr(0x%lx), dma_addr(0x%lx), size(0x%lx)\n",
+			(unsigned long)buf->kvaddr,
+			(unsigned long)buf->dma_addr,
+			buf->size);
+
+	return ret;
+err2:
+	dma_free_writecombine(dev->dev, buf->size, buf->kvaddr,
+			(dma_addr_t)buf->dma_addr);
+	buf->dma_addr = (dma_addr_t)NULL;
+err1:
+	sg_free_table(buf->sgt);
+	kfree(buf->sgt);
+	buf->sgt = NULL;
+
+	return ret;
 }
 
 static void lowlevel_buffer_deallocate(struct drm_device *dev,
-		struct exynos_drm_gem_buf *buffer)
+		unsigned int flags, struct exynos_drm_gem_buf *buf)
 {
 	DRM_DEBUG_KMS("%s.\n", __FILE__);
 
-	if (buffer->dma_addr && buffer->size)
-		dma_free_writecombine(dev->dev, buffer->size, buffer->kvaddr,
-				(dma_addr_t)buffer->dma_addr);
-	else
-		DRM_DEBUG_KMS("buffer data are invalid.\n");
+	/*
+	 * release only physically continuous memory and
+	 * non-continuous memory would be released by exynos
+	 * gem framework.
+	 */
+	if (flags & EXYNOS_BO_NONCONTIG) {
+		DRM_DEBUG_KMS("not support allocation type.\n");
+		return;
+	}
+
+	if (!buf->dma_addr) {
+		DRM_DEBUG_KMS("dma_addr is invalid.\n");
+		return;
+	}
+
+	DRM_DEBUG_KMS("vaddr(0x%lx), dma_addr(0x%lx), size(0x%lx)\n",
+			(unsigned long)buf->kvaddr,
+			(unsigned long)buf->dma_addr,
+			buf->size);
+
+	sg_free_table(buf->sgt);
+
+	kfree(buf->sgt);
+	buf->sgt = NULL;
+
+	kfree(buf->pages);
+	buf->pages = NULL;
+
+	dma_free_writecombine(dev->dev, buf->size, buf->kvaddr,
+				(dma_addr_t)buf->dma_addr);
+	buf->dma_addr = (dma_addr_t)NULL;
 }
 
-struct exynos_drm_gem_buf *exynos_drm_buf_create(struct drm_device *dev,
-		unsigned int size)
+struct exynos_drm_gem_buf *exynos_drm_init_buf(struct drm_device *dev,
+						unsigned int size)
 {
 	struct exynos_drm_gem_buf *buffer;
 
@@ -77,21 +193,11 @@ struct exynos_drm_gem_buf *exynos_drm_buf_create(struct drm_device *dev,
 	}
 
 	buffer->size = size;
-
-	/*
-	 * allocate memory region with size and set the memory information
-	 * to vaddr and dma_addr of a buffer object.
-	 */
-	if (lowlevel_buffer_allocate(dev, buffer) < 0) {
-		kfree(buffer);
-		return NULL;
-	}
-
 	return buffer;
 }
 
-void exynos_drm_buf_destroy(struct drm_device *dev,
-		struct exynos_drm_gem_buf *buffer)
+void exynos_drm_fini_buf(struct drm_device *dev,
+				struct exynos_drm_gem_buf *buffer)
 {
 	DRM_DEBUG_KMS("%s.\n", __FILE__);
 
@@ -100,12 +206,27 @@ void exynos_drm_buf_destroy(struct drm_device *dev,
 		return;
 	}
 
-	lowlevel_buffer_deallocate(dev, buffer);
-
 	kfree(buffer);
 	buffer = NULL;
 }
 
-MODULE_AUTHOR("Inki Dae <inki.dae@samsung.com>");
-MODULE_DESCRIPTION("Samsung SoC DRM Buffer Management Module");
-MODULE_LICENSE("GPL");
+int exynos_drm_alloc_buf(struct drm_device *dev,
+		struct exynos_drm_gem_buf *buf, unsigned int flags)
+{
+
+	/*
+	 * allocate memory region and set the memory information
+	 * to vaddr and dma_addr of a buffer object.
+	 */
+	if (lowlevel_buffer_allocate(dev, flags, buf) < 0)
+		return -ENOMEM;
+
+	return 0;
+}
+
+void exynos_drm_free_buf(struct drm_device *dev,
+		unsigned int flags, struct exynos_drm_gem_buf *buffer)
+{
+
+	lowlevel_buffer_deallocate(dev, flags, buffer);
+}
diff --git a/drivers/gpu/drm/exynos/exynos_drm_buf.h b/drivers/gpu/drm/exynos/exynos_drm_buf.h
index c913f2bad760..3388e4eb4ba2 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_buf.h
+++ b/drivers/gpu/drm/exynos/exynos_drm_buf.h
@@ -26,12 +26,22 @@
 #ifndef _EXYNOS_DRM_BUF_H_
 #define _EXYNOS_DRM_BUF_H_
 
-/* allocate physical memory. */
-struct exynos_drm_gem_buf *exynos_drm_buf_create(struct drm_device *dev,
-		unsigned int size);
+/* create and initialize buffer object. */
+struct exynos_drm_gem_buf *exynos_drm_init_buf(struct drm_device *dev,
+						unsigned int size);
 
-/* remove allocated physical memory. */
-void exynos_drm_buf_destroy(struct drm_device *dev,
-		struct exynos_drm_gem_buf *buffer);
+/* destroy buffer object. */
+void exynos_drm_fini_buf(struct drm_device *dev,
+				struct exynos_drm_gem_buf *buffer);
+
+/* allocate physical memory region and setup sgt and pages. */
+int exynos_drm_alloc_buf(struct drm_device *dev,
+				struct exynos_drm_gem_buf *buf,
+				unsigned int flags);
+
+/* release physical memory region, sgt and pages. */
+void exynos_drm_free_buf(struct drm_device *dev,
+				unsigned int flags,
+				struct exynos_drm_gem_buf *buffer);
 
 #endif
diff --git a/drivers/gpu/drm/exynos/exynos_drm_connector.c b/drivers/gpu/drm/exynos/exynos_drm_connector.c
index d620b0784257..bf791fa0e50d 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_connector.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_connector.c
@@ -28,6 +28,7 @@
 #include "drmP.h"
 #include "drm_crtc_helper.h"
 
+#include <drm/exynos_drm.h>
 #include "exynos_drm_drv.h"
 #include "exynos_drm_encoder.h"
 
@@ -44,22 +45,25 @@ struct exynos_drm_connector {
 /* convert exynos_video_timings to drm_display_mode */
 static inline void
 convert_to_display_mode(struct drm_display_mode *mode,
-			struct fb_videomode *timing)
+			struct exynos_drm_panel_info *panel)
 {
+	struct fb_videomode *timing = &panel->timing;
 	DRM_DEBUG_KMS("%s\n", __FILE__);
 
 	mode->clock = timing->pixclock / 1000;
 	mode->vrefresh = timing->refresh;
 
 	mode->hdisplay = timing->xres;
-	mode->hsync_start = mode->hdisplay + timing->left_margin;
+	mode->hsync_start = mode->hdisplay + timing->right_margin;
 	mode->hsync_end = mode->hsync_start + timing->hsync_len;
-	mode->htotal = mode->hsync_end + timing->right_margin;
+	mode->htotal = mode->hsync_end + timing->left_margin;
 
 	mode->vdisplay = timing->yres;
-	mode->vsync_start = mode->vdisplay + timing->upper_margin;
+	mode->vsync_start = mode->vdisplay + timing->lower_margin;
 	mode->vsync_end = mode->vsync_start + timing->vsync_len;
-	mode->vtotal = mode->vsync_end + timing->lower_margin;
+	mode->vtotal = mode->vsync_end + timing->upper_margin;
+	mode->width_mm = panel->width_mm;
+	mode->height_mm = panel->height_mm;
 
 	if (timing->vmode & FB_VMODE_INTERLACED)
 		mode->flags |= DRM_MODE_FLAG_INTERLACE;
@@ -81,14 +85,14 @@ convert_to_video_timing(struct fb_videomode *timing,
 	timing->refresh = drm_mode_vrefresh(mode);
 
 	timing->xres = mode->hdisplay;
-	timing->left_margin = mode->hsync_start - mode->hdisplay;
+	timing->right_margin = mode->hsync_start - mode->hdisplay;
 	timing->hsync_len = mode->hsync_end - mode->hsync_start;
-	timing->right_margin = mode->htotal - mode->hsync_end;
+	timing->left_margin = mode->htotal - mode->hsync_end;
 
 	timing->yres = mode->vdisplay;
-	timing->upper_margin = mode->vsync_start - mode->vdisplay;
+	timing->lower_margin = mode->vsync_start - mode->vdisplay;
 	timing->vsync_len = mode->vsync_end - mode->vsync_start;
-	timing->lower_margin = mode->vtotal - mode->vsync_end;
+	timing->upper_margin = mode->vtotal - mode->vsync_end;
 
 	if (mode->flags & DRM_MODE_FLAG_INTERLACE)
 		timing->vmode = FB_VMODE_INTERLACED;
@@ -148,16 +152,18 @@ static int exynos_drm_connector_get_modes(struct drm_connector *connector)
 		connector->display_info.raw_edid = edid;
 	} else {
 		struct drm_display_mode *mode = drm_mode_create(connector->dev);
-		struct fb_videomode *timing;
+		struct exynos_drm_panel_info *panel;
 
-		if (display_ops->get_timing)
-			timing = display_ops->get_timing(manager->dev);
+		if (display_ops->get_panel)
+			panel = display_ops->get_panel(manager->dev);
 		else {
 			drm_mode_destroy(connector->dev, mode);
 			return 0;
 		}
 
-		convert_to_display_mode(mode, timing);
+		convert_to_display_mode(mode, panel);
+		connector->display_info.width_mm = mode->width_mm;
+		connector->display_info.height_mm = mode->height_mm;
 
 		mode->type = DRM_MODE_TYPE_DRIVER | DRM_MODE_TYPE_PREFERRED;
 		drm_mode_set_name(mode);
@@ -219,6 +225,29 @@ static struct drm_connector_helper_funcs exynos_connector_helper_funcs = {
 	.best_encoder	= exynos_drm_best_encoder,
 };
 
+static int exynos_drm_connector_fill_modes(struct drm_connector *connector,
+				unsigned int max_width, unsigned int max_height)
+{
+	struct exynos_drm_connector *exynos_connector =
+					to_exynos_connector(connector);
+	struct exynos_drm_manager *manager = exynos_connector->manager;
+	struct exynos_drm_manager_ops *ops = manager->ops;
+	unsigned int width, height;
+
+	width = max_width;
+	height = max_height;
+
+	/*
+	 * if specific driver want to find desired_mode using maxmum
+	 * resolution then get max width and height from that driver.
+	 */
+	if (ops && ops->get_max_resol)
+		ops->get_max_resol(manager->dev, &width, &height);
+
+	return drm_helper_probe_single_connector_modes(connector, width,
+							height);
+}
+
 /* get detection status of display device. */
 static enum drm_connector_status
 exynos_drm_connector_detect(struct drm_connector *connector, bool force)
@@ -256,7 +285,7 @@ static void exynos_drm_connector_destroy(struct drm_connector *connector)
 
 static struct drm_connector_funcs exynos_connector_funcs = {
 	.dpms		= drm_helper_connector_dpms,
-	.fill_modes	= drm_helper_probe_single_connector_modes,
+	.fill_modes	= exynos_drm_connector_fill_modes,
 	.detect		= exynos_drm_connector_detect,
 	.destroy	= exynos_drm_connector_destroy,
 };
@@ -286,6 +315,10 @@ struct drm_connector *exynos_drm_connector_create(struct drm_device *dev,
 		connector->interlace_allowed = true;
 		connector->polled = DRM_CONNECTOR_POLL_HPD;
 		break;
+	case EXYNOS_DISPLAY_TYPE_VIDI:
+		type = DRM_MODE_CONNECTOR_VIRTUAL;
+		connector->polled = DRM_CONNECTOR_POLL_HPD;
+		break;
 	default:
 		type = DRM_MODE_CONNECTOR_Unknown;
 		break;
@@ -319,9 +352,3 @@ err_connector:
 	kfree(exynos_connector);
 	return NULL;
 }
-
-MODULE_AUTHOR("Inki Dae <inki.dae@samsung.com>");
-MODULE_AUTHOR("Joonyoung Shim <jy0922.shim@samsung.com>");
-MODULE_AUTHOR("Seung-Woo Kim <sw0312.kim@samsung.com>");
-MODULE_DESCRIPTION("Samsung SoC DRM Connector Driver");
-MODULE_LICENSE("GPL");
diff --git a/drivers/gpu/drm/exynos/exynos_drm_core.c b/drivers/gpu/drm/exynos/exynos_drm_core.c
index 661a03571d0c..411832e8e17a 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_core.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_core.c
@@ -32,7 +32,6 @@
 #include "exynos_drm_connector.h"
 #include "exynos_drm_fbdev.h"
 
-static DEFINE_MUTEX(exynos_drm_mutex);
 static LIST_HEAD(exynos_drm_subdrv_list);
 static struct drm_device *drm_dev;
 
@@ -60,6 +59,9 @@ static int exynos_drm_subdrv_probe(struct drm_device *dev,
 			return ret;
 	}
 
+	if (subdrv->is_local)
+		return 0;
+
 	/* create and initialize a encoder for this sub driver. */
 	encoder = exynos_drm_encoder_create(dev, &subdrv->manager,
 			(1 << MAX_CRTC) - 1);
@@ -116,13 +118,10 @@ int exynos_drm_device_register(struct drm_device *dev)
 	if (!dev)
 		return -EINVAL;
 
-	if (drm_dev) {
-		DRM_ERROR("Already drm device were registered\n");
-		return -EBUSY;
-	}
+	drm_dev = dev;
 
-	mutex_lock(&exynos_drm_mutex);
 	list_for_each_entry_safe(subdrv, n, &exynos_drm_subdrv_list, list) {
+		subdrv->drm_dev = dev;
 		err = exynos_drm_subdrv_probe(dev, subdrv);
 		if (err) {
 			DRM_DEBUG("exynos drm subdrv probe failed.\n");
@@ -130,9 +129,6 @@ int exynos_drm_device_register(struct drm_device *dev)
 		}
 	}
 
-	drm_dev = dev;
-	mutex_unlock(&exynos_drm_mutex);
-
 	return 0;
 }
 EXPORT_SYMBOL_GPL(exynos_drm_device_register);
@@ -143,83 +139,28 @@ int exynos_drm_device_unregister(struct drm_device *dev)
 
 	DRM_DEBUG_DRIVER("%s\n", __FILE__);
 
-	if (!dev || dev != drm_dev) {
+	if (!dev) {
 		WARN(1, "Unexpected drm device unregister!\n");
 		return -EINVAL;
 	}
 
-	mutex_lock(&exynos_drm_mutex);
 	list_for_each_entry(subdrv, &exynos_drm_subdrv_list, list)
 		exynos_drm_subdrv_remove(dev, subdrv);
 
 	drm_dev = NULL;
-	mutex_unlock(&exynos_drm_mutex);
 
 	return 0;
 }
 EXPORT_SYMBOL_GPL(exynos_drm_device_unregister);
 
-static int exynos_drm_mode_group_reinit(struct drm_device *dev)
-{
-	struct drm_mode_group *group = &dev->primary->mode_group;
-	uint32_t *id_list = group->id_list;
-	int ret;
-
-	DRM_DEBUG_DRIVER("%s\n", __FILE__);
-
-	ret = drm_mode_group_init_legacy_group(dev, group);
-	if (ret < 0)
-		return ret;
-
-	kfree(id_list);
-	return 0;
-}
-
 int exynos_drm_subdrv_register(struct exynos_drm_subdrv *subdrv)
 {
-	int err;
-
 	DRM_DEBUG_DRIVER("%s\n", __FILE__);
 
 	if (!subdrv)
 		return -EINVAL;
 
-	mutex_lock(&exynos_drm_mutex);
-	if (drm_dev) {
-		err = exynos_drm_subdrv_probe(drm_dev, subdrv);
-		if (err) {
-			DRM_ERROR("failed to probe exynos drm subdrv\n");
-			mutex_unlock(&exynos_drm_mutex);
-			return err;
-		}
-
-		/*
-		 * if any specific driver such as fimd or hdmi driver called
-		 * exynos_drm_subdrv_register() later than drm_load(),
-		 * the fb helper should be re-initialized and re-configured.
-		 */
-		err = exynos_drm_fbdev_reinit(drm_dev);
-		if (err) {
-			DRM_ERROR("failed to reinitialize exynos drm fbdev\n");
-			exynos_drm_subdrv_remove(drm_dev, subdrv);
-			mutex_unlock(&exynos_drm_mutex);
-			return err;
-		}
-
-		err = exynos_drm_mode_group_reinit(drm_dev);
-		if (err) {
-			DRM_ERROR("failed to reinitialize mode group\n");
-			exynos_drm_fbdev_fini(drm_dev);
-			exynos_drm_subdrv_remove(drm_dev, subdrv);
-			mutex_unlock(&exynos_drm_mutex);
-			return err;
-		}
-	}
-
-	subdrv->drm_dev = drm_dev;
-
 	list_add_tail(&subdrv->list, &exynos_drm_subdrv_list);
-	mutex_unlock(&exynos_drm_mutex);
 
 	return 0;
 }
@@ -227,46 +168,48 @@ EXPORT_SYMBOL_GPL(exynos_drm_subdrv_register);
 
 int exynos_drm_subdrv_unregister(struct exynos_drm_subdrv *subdrv)
 {
-	int ret = -EFAULT;
-
 	DRM_DEBUG_DRIVER("%s\n", __FILE__);
 
-	if (!subdrv) {
-		DRM_DEBUG("Unexpected exynos drm subdrv unregister!\n");
-		return ret;
-	}
+	if (!subdrv)
+		return -EINVAL;
 
-	mutex_lock(&exynos_drm_mutex);
-	if (drm_dev) {
-		exynos_drm_subdrv_remove(drm_dev, subdrv);
-		list_del(&subdrv->list);
+	list_del(&subdrv->list);
 
-		/*
-		 * fb helper should be updated once a sub driver is released
-		 * to re-configure crtc and connector and also to re-setup
-		 * drm framebuffer.
-		 */
-		ret = exynos_drm_fbdev_reinit(drm_dev);
-		if (ret < 0) {
-			DRM_ERROR("failed fb helper reinit.\n");
-			goto fail;
-		}
+	return 0;
+}
+EXPORT_SYMBOL_GPL(exynos_drm_subdrv_unregister);
 
-		ret = exynos_drm_mode_group_reinit(drm_dev);
-		if (ret < 0) {
-			DRM_ERROR("failed drm mode group reinit.\n");
-			goto fail;
+int exynos_drm_subdrv_open(struct drm_device *dev, struct drm_file *file)
+{
+	struct exynos_drm_subdrv *subdrv;
+	int ret;
+
+	list_for_each_entry(subdrv, &exynos_drm_subdrv_list, list) {
+		if (subdrv->open) {
+			ret = subdrv->open(dev, subdrv->manager.dev, file);
+			if (ret)
+				goto err;
 		}
 	}
 
-fail:
-	mutex_unlock(&exynos_drm_mutex);
+	return 0;
+
+err:
+	list_for_each_entry_reverse(subdrv, &subdrv->list, list) {
+		if (subdrv->close)
+			subdrv->close(dev, subdrv->manager.dev, file);
+	}
 	return ret;
 }
-EXPORT_SYMBOL_GPL(exynos_drm_subdrv_unregister);
+EXPORT_SYMBOL_GPL(exynos_drm_subdrv_open);
+
+void exynos_drm_subdrv_close(struct drm_device *dev, struct drm_file *file)
+{
+	struct exynos_drm_subdrv *subdrv;
 
-MODULE_AUTHOR("Inki Dae <inki.dae@samsung.com>");
-MODULE_AUTHOR("Joonyoung Shim <jy0922.shim@samsung.com>");
-MODULE_AUTHOR("Seung-Woo Kim <sw0312.kim@samsung.com>");
-MODULE_DESCRIPTION("Samsung SoC DRM Core Driver");
-MODULE_LICENSE("GPL");
+	list_for_each_entry(subdrv, &exynos_drm_subdrv_list, list) {
+		if (subdrv->close)
+			subdrv->close(dev, subdrv->manager.dev, file);
+	}
+}
+EXPORT_SYMBOL_GPL(exynos_drm_subdrv_close);
diff --git a/drivers/gpu/drm/exynos/exynos_drm_crtc.c b/drivers/gpu/drm/exynos/exynos_drm_crtc.c
index e3861ac49295..3486ffed0bf0 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_crtc.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_crtc.c
@@ -249,7 +249,11 @@ exynos_drm_crtc_mode_set(struct drm_crtc *crtc, struct drm_display_mode *mode,
 {
 	DRM_DEBUG_KMS("%s\n", __FILE__);
 
-	mode = adjusted_mode;
+	/*
+	 * copy the mode data adjusted by mode_fixup() into crtc->mode
+	 * so that hardware can be seet to proper mode.
+	 */
+	memcpy(&crtc->mode, adjusted_mode, sizeof(*adjusted_mode));
 
 	return exynos_drm_crtc_update(crtc);
 }
@@ -307,9 +311,6 @@ static int exynos_drm_crtc_page_flip(struct drm_crtc *crtc,
 		 */
 		event->pipe = exynos_crtc->pipe;
 
-		list_add_tail(&event->base.link,
-				&dev_priv->pageflip_event_list);
-
 		ret = drm_vblank_get(dev, exynos_crtc->pipe);
 		if (ret) {
 			DRM_DEBUG("failed to acquire vblank counter\n");
@@ -318,6 +319,9 @@ static int exynos_drm_crtc_page_flip(struct drm_crtc *crtc,
 			goto out;
 		}
 
+		list_add_tail(&event->base.link,
+				&dev_priv->pageflip_event_list);
+
 		crtc->fb = fb;
 		ret = exynos_drm_crtc_update(crtc);
 		if (ret) {
@@ -426,9 +430,3 @@ void exynos_drm_crtc_disable_vblank(struct drm_device *dev, int crtc)
 	exynos_drm_fn_encoder(private->crtc[crtc], &crtc,
 			exynos_drm_disable_vblank);
 }
-
-MODULE_AUTHOR("Inki Dae <inki.dae@samsung.com>");
-MODULE_AUTHOR("Joonyoung Shim <jy0922.shim@samsung.com>");
-MODULE_AUTHOR("Seung-Woo Kim <sw0312.kim@samsung.com>");
-MODULE_DESCRIPTION("Samsung SoC DRM CRTC Driver");
-MODULE_LICENSE("GPL");
diff --git a/drivers/gpu/drm/exynos/exynos_drm_drv.c b/drivers/gpu/drm/exynos/exynos_drm_drv.c
index 35889ca255e9..a6819b5f8428 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_drv.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_drv.c
@@ -33,10 +33,12 @@
 
 #include "exynos_drm_drv.h"
 #include "exynos_drm_crtc.h"
+#include "exynos_drm_encoder.h"
 #include "exynos_drm_fbdev.h"
 #include "exynos_drm_fb.h"
 #include "exynos_drm_gem.h"
 #include "exynos_drm_plane.h"
+#include "exynos_drm_vidi.h"
 
 #define DRIVER_NAME	"exynos"
 #define DRIVER_DESC	"Samsung SoC DRM"
@@ -99,6 +101,9 @@ static int exynos_drm_load(struct drm_device *dev, unsigned long flags)
 	if (ret)
 		goto err_vblank;
 
+	/* setup possible_clones. */
+	exynos_drm_encoder_setup(dev);
+
 	/*
 	 * create and configure fb helper and also exynos specific
 	 * fbdev object.
@@ -140,17 +145,45 @@ static int exynos_drm_unload(struct drm_device *dev)
 	return 0;
 }
 
+static int exynos_drm_open(struct drm_device *dev, struct drm_file *file)
+{
+	DRM_DEBUG_DRIVER("%s\n", __FILE__);
+
+	return exynos_drm_subdrv_open(dev, file);
+}
+
 static void exynos_drm_preclose(struct drm_device *dev,
-					struct drm_file *file_priv)
+					struct drm_file *file)
 {
-	struct exynos_drm_private *dev_priv = dev->dev_private;
+	struct exynos_drm_private *private = dev->dev_private;
+	struct drm_pending_vblank_event *e, *t;
+	unsigned long flags;
 
-	/*
-	 * drm framework frees all events at release time,
-	 * so private event list should be cleared.
-	 */
-	if (!list_empty(&dev_priv->pageflip_event_list))
-		INIT_LIST_HEAD(&dev_priv->pageflip_event_list);
+	DRM_DEBUG_DRIVER("%s\n", __FILE__);
+
+	/* release events of current file */
+	spin_lock_irqsave(&dev->event_lock, flags);
+	list_for_each_entry_safe(e, t, &private->pageflip_event_list,
+			base.link) {
+		if (e->base.file_priv == file) {
+			list_del(&e->base.link);
+			e->base.destroy(&e->base);
+		}
+	}
+	spin_unlock_irqrestore(&dev->event_lock, flags);
+
+	exynos_drm_subdrv_close(dev, file);
+}
+
+static void exynos_drm_postclose(struct drm_device *dev, struct drm_file *file)
+{
+	DRM_DEBUG_DRIVER("%s\n", __FILE__);
+
+	if (!file->driver_priv)
+		return;
+
+	kfree(file->driver_priv);
+	file->driver_priv = NULL;
 }
 
 static void exynos_drm_lastclose(struct drm_device *dev)
@@ -176,6 +209,8 @@ static struct drm_ioctl_desc exynos_ioctls[] = {
 			exynos_drm_gem_mmap_ioctl, DRM_UNLOCKED | DRM_AUTH),
 	DRM_IOCTL_DEF_DRV(EXYNOS_PLANE_SET_ZPOS, exynos_plane_set_zpos_ioctl,
 			DRM_UNLOCKED | DRM_AUTH),
+	DRM_IOCTL_DEF_DRV(EXYNOS_VIDI_CONNECTION,
+			vidi_connection_ioctl, DRM_UNLOCKED | DRM_AUTH),
 };
 
 static const struct file_operations exynos_drm_driver_fops = {
@@ -193,8 +228,10 @@ static struct drm_driver exynos_drm_driver = {
 				  DRIVER_MODESET | DRIVER_GEM,
 	.load			= exynos_drm_load,
 	.unload			= exynos_drm_unload,
+	.open			= exynos_drm_open,
 	.preclose		= exynos_drm_preclose,
 	.lastclose		= exynos_drm_lastclose,
+	.postclose		= exynos_drm_postclose,
 	.get_vblank_counter	= drm_vblank_count,
 	.enable_vblank		= exynos_drm_crtc_enable_vblank,
 	.disable_vblank		= exynos_drm_crtc_disable_vblank,
@@ -236,15 +273,66 @@ static struct platform_driver exynos_drm_platform_driver = {
 	.remove		= __devexit_p(exynos_drm_platform_remove),
 	.driver		= {
 		.owner	= THIS_MODULE,
-		.name	= DRIVER_NAME,
+		.name	= "exynos-drm",
 	},
 };
 
 static int __init exynos_drm_init(void)
 {
+	int ret;
+
 	DRM_DEBUG_DRIVER("%s\n", __FILE__);
 
-	return platform_driver_register(&exynos_drm_platform_driver);
+#ifdef CONFIG_DRM_EXYNOS_FIMD
+	ret = platform_driver_register(&fimd_driver);
+	if (ret < 0)
+		goto out_fimd;
+#endif
+
+#ifdef CONFIG_DRM_EXYNOS_HDMI
+	ret = platform_driver_register(&hdmi_driver);
+	if (ret < 0)
+		goto out_hdmi;
+	ret = platform_driver_register(&mixer_driver);
+	if (ret < 0)
+		goto out_mixer;
+	ret = platform_driver_register(&exynos_drm_common_hdmi_driver);
+	if (ret < 0)
+		goto out_common_hdmi;
+#endif
+
+#ifdef CONFIG_DRM_EXYNOS_VIDI
+	ret = platform_driver_register(&vidi_driver);
+	if (ret < 0)
+		goto out_vidi;
+#endif
+
+	ret = platform_driver_register(&exynos_drm_platform_driver);
+	if (ret < 0)
+		goto out;
+
+	return 0;
+
+out:
+#ifdef CONFIG_DRM_EXYNOS_VIDI
+out_vidi:
+	platform_driver_unregister(&vidi_driver);
+#endif
+
+#ifdef CONFIG_DRM_EXYNOS_HDMI
+	platform_driver_unregister(&exynos_drm_common_hdmi_driver);
+out_common_hdmi:
+	platform_driver_unregister(&mixer_driver);
+out_mixer:
+	platform_driver_unregister(&hdmi_driver);
+out_hdmi:
+#endif
+
+#ifdef CONFIG_DRM_EXYNOS_FIMD
+	platform_driver_unregister(&fimd_driver);
+out_fimd:
+#endif
+	return ret;
 }
 
 static void __exit exynos_drm_exit(void)
@@ -252,6 +340,20 @@ static void __exit exynos_drm_exit(void)
 	DRM_DEBUG_DRIVER("%s\n", __FILE__);
 
 	platform_driver_unregister(&exynos_drm_platform_driver);
+
+#ifdef CONFIG_DRM_EXYNOS_HDMI
+	platform_driver_unregister(&exynos_drm_common_hdmi_driver);
+	platform_driver_unregister(&mixer_driver);
+	platform_driver_unregister(&hdmi_driver);
+#endif
+
+#ifdef CONFIG_DRM_EXYNOS_VIDI
+	platform_driver_unregister(&vidi_driver);
+#endif
+
+#ifdef CONFIG_DRM_EXYNOS_FIMD
+	platform_driver_unregister(&fimd_driver);
+#endif
 }
 
 module_init(exynos_drm_init);
diff --git a/drivers/gpu/drm/exynos/exynos_drm_drv.h b/drivers/gpu/drm/exynos/exynos_drm_drv.h
index e685e1e33055..fbd0a232c93d 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_drv.h
+++ b/drivers/gpu/drm/exynos/exynos_drm_drv.h
@@ -32,9 +32,9 @@
 #include <linux/module.h>
 #include "drm.h"
 
-#define MAX_CRTC	2
+#define MAX_CRTC	3
 #define MAX_PLANE	5
-#define MAX_FB_BUFFER	3
+#define MAX_FB_BUFFER	4
 #define DEFAULT_ZPOS	-1
 
 struct drm_device;
@@ -50,6 +50,8 @@ enum exynos_drm_output_type {
 	EXYNOS_DISPLAY_TYPE_LCD,
 	/* HDMI Interface. */
 	EXYNOS_DISPLAY_TYPE_HDMI,
+	/* Virtual Display Interface. */
+	EXYNOS_DISPLAY_TYPE_VIDI,
 };
 
 /*
@@ -136,7 +138,7 @@ struct exynos_drm_overlay {
  * @type: one of EXYNOS_DISPLAY_TYPE_LCD and HDMI.
  * @is_connected: check for that display is connected or not.
  * @get_edid: get edid modes from display driver.
- * @get_timing: get timing object from display driver.
+ * @get_panel: get panel object from display driver.
  * @check_timing: check if timing is valid or not.
  * @power_on: display device on or off.
  */
@@ -145,7 +147,7 @@ struct exynos_drm_display_ops {
 	bool (*is_connected)(struct device *dev);
 	int (*get_edid)(struct device *dev, struct drm_connector *connector,
 				u8 *edid, int len);
-	void *(*get_timing)(struct device *dev);
+	void *(*get_panel)(struct device *dev);
 	int (*check_timing)(struct device *dev, void *timing);
 	int (*power_on)(struct device *dev, int mode);
 };
@@ -155,8 +157,10 @@ struct exynos_drm_display_ops {
  *
  * @dpms: control device power.
  * @apply: set timing, vblank and overlay data to registers.
+ * @mode_fixup: fix mode data comparing to hw specific display mode.
  * @mode_set: convert drm_display_mode to hw specific display mode and
  *	      would be called by encoder->mode_set().
+ * @get_max_resol: get maximum resolution to specific hardware.
  * @commit: set current hw specific display mode to hw.
  * @enable_vblank: specific driver callback for enabling vblank interrupt.
  * @disable_vblank: specific driver callback for disabling vblank interrupt.
@@ -164,7 +168,13 @@ struct exynos_drm_display_ops {
 struct exynos_drm_manager_ops {
 	void (*dpms)(struct device *subdrv_dev, int mode);
 	void (*apply)(struct device *subdrv_dev);
+	void (*mode_fixup)(struct device *subdrv_dev,
+				struct drm_connector *connector,
+				struct drm_display_mode *mode,
+				struct drm_display_mode *adjusted_mode);
 	void (*mode_set)(struct device *subdrv_dev, void *mode);
+	void (*get_max_resol)(struct device *subdrv_dev, unsigned int *width,
+				unsigned int *height);
 	void (*commit)(struct device *subdrv_dev);
 	int (*enable_vblank)(struct device *subdrv_dev);
 	void (*disable_vblank)(struct device *subdrv_dev);
@@ -217,10 +227,13 @@ struct exynos_drm_private {
  * @list: sub driver has its own list object to register to exynos drm driver.
  * @drm_dev: pointer to drm_device and this pointer would be set
  *	when sub driver calls exynos_drm_subdrv_register().
+ * @is_local: appear encoder and connector disrelated device.
  * @probe: this callback would be called by exynos drm driver after
  *	subdrv is registered to it.
  * @remove: this callback is used to release resources created
  *	by probe callback.
+ * @open: this would be called with drm device file open.
+ * @close: this would be called with drm device file close.
  * @manager: subdrv has its own manager to control a hardware appropriately
  *	and we can access a hardware drawing on this manager.
  * @encoder: encoder object owned by this sub driver.
@@ -229,9 +242,14 @@ struct exynos_drm_private {
 struct exynos_drm_subdrv {
 	struct list_head list;
 	struct drm_device *drm_dev;
+	bool is_local;
 
 	int (*probe)(struct drm_device *drm_dev, struct device *dev);
 	void (*remove)(struct drm_device *dev);
+	int (*open)(struct drm_device *drm_dev, struct device *dev,
+			struct drm_file *file);
+	void (*close)(struct drm_device *drm_dev, struct device *dev,
+			struct drm_file *file);
 
 	struct exynos_drm_manager manager;
 	struct drm_encoder *encoder;
@@ -254,15 +272,19 @@ int exynos_drm_device_unregister(struct drm_device *dev);
  * this function would be called by sub drivers such as display controller
  * or hdmi driver to register this sub driver object to exynos drm driver
  * and when a sub driver is registered to exynos drm driver a probe callback
- * of the sub driver is called and creates its own encoder and connector
- * and then fb helper and drm mode group would be re-initialized.
+ * of the sub driver is called and creates its own encoder and connector.
  */
 int exynos_drm_subdrv_register(struct exynos_drm_subdrv *drm_subdrv);
 
-/*
- * this function removes subdrv list from exynos drm driver and fb helper
- * and drm mode group would be re-initialized.
- */
+/* this function removes subdrv list from exynos drm driver */
 int exynos_drm_subdrv_unregister(struct exynos_drm_subdrv *drm_subdrv);
 
+int exynos_drm_subdrv_open(struct drm_device *dev, struct drm_file *file);
+void exynos_drm_subdrv_close(struct drm_device *dev, struct drm_file *file);
+
+extern struct platform_driver fimd_driver;
+extern struct platform_driver hdmi_driver;
+extern struct platform_driver mixer_driver;
+extern struct platform_driver exynos_drm_common_hdmi_driver;
+extern struct platform_driver vidi_driver;
 #endif
diff --git a/drivers/gpu/drm/exynos/exynos_drm_encoder.c b/drivers/gpu/drm/exynos/exynos_drm_encoder.c
index 86b93dde219a..6e9ac7bd1dcf 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_encoder.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_encoder.c
@@ -111,9 +111,19 @@ exynos_drm_encoder_mode_fixup(struct drm_encoder *encoder,
 			       struct drm_display_mode *mode,
 			       struct drm_display_mode *adjusted_mode)
 {
+	struct drm_device *dev = encoder->dev;
+	struct drm_connector *connector;
+	struct exynos_drm_manager *manager = exynos_drm_get_manager(encoder);
+	struct exynos_drm_manager_ops *manager_ops = manager->ops;
+
 	DRM_DEBUG_KMS("%s\n", __FILE__);
 
-	/* drm framework doesn't check NULL. */
+	list_for_each_entry(connector, &dev->mode_config.connector_list, head) {
+		if (connector->encoder == encoder)
+			if (manager_ops && manager_ops->mode_fixup)
+				manager_ops->mode_fixup(manager->dev, connector,
+							mode, adjusted_mode);
+	}
 
 	return true;
 }
@@ -132,12 +142,11 @@ static void exynos_drm_encoder_mode_set(struct drm_encoder *encoder,
 
 	DRM_DEBUG_KMS("%s\n", __FILE__);
 
-	mode = adjusted_mode;
-
 	list_for_each_entry(connector, &dev->mode_config.connector_list, head) {
 		if (connector->encoder == encoder) {
 			if (manager_ops && manager_ops->mode_set)
-				manager_ops->mode_set(manager->dev, mode);
+				manager_ops->mode_set(manager->dev,
+							adjusted_mode);
 
 			if (overlay_ops && overlay_ops->mode_set)
 				overlay_ops->mode_set(manager->dev, overlay);
@@ -195,6 +204,41 @@ static struct drm_encoder_funcs exynos_encoder_funcs = {
 	.destroy = exynos_drm_encoder_destroy,
 };
 
+static unsigned int exynos_drm_encoder_clones(struct drm_encoder *encoder)
+{
+	struct drm_encoder *clone;
+	struct drm_device *dev = encoder->dev;
+	struct exynos_drm_encoder *exynos_encoder = to_exynos_encoder(encoder);
+	struct exynos_drm_display_ops *display_ops =
+				exynos_encoder->manager->display_ops;
+	unsigned int clone_mask = 0;
+	int cnt = 0;
+
+	list_for_each_entry(clone, &dev->mode_config.encoder_list, head) {
+		switch (display_ops->type) {
+		case EXYNOS_DISPLAY_TYPE_LCD:
+		case EXYNOS_DISPLAY_TYPE_HDMI:
+		case EXYNOS_DISPLAY_TYPE_VIDI:
+			clone_mask |= (1 << (cnt++));
+			break;
+		default:
+			continue;
+		}
+	}
+
+	return clone_mask;
+}
+
+void exynos_drm_encoder_setup(struct drm_device *dev)
+{
+	struct drm_encoder *encoder;
+
+	DRM_DEBUG_KMS("%s\n", __FILE__);
+
+	list_for_each_entry(encoder, &dev->mode_config.encoder_list, head)
+		encoder->possible_clones = exynos_drm_encoder_clones(encoder);
+}
+
 struct drm_encoder *
 exynos_drm_encoder_create(struct drm_device *dev,
 			   struct exynos_drm_manager *manager,
@@ -399,9 +443,3 @@ void exynos_drm_encoder_crtc_disable(struct drm_encoder *encoder, void *data)
 	if (overlay_ops && overlay_ops->disable)
 		overlay_ops->disable(manager->dev, zpos);
 }
-
-MODULE_AUTHOR("Inki Dae <inki.dae@samsung.com>");
-MODULE_AUTHOR("Joonyoung Shim <jy0922.shim@samsung.com>");
-MODULE_AUTHOR("Seung-Woo Kim <sw0312.kim@samsung.com>");
-MODULE_DESCRIPTION("Samsung SoC DRM Encoder Driver");
-MODULE_LICENSE("GPL");
diff --git a/drivers/gpu/drm/exynos/exynos_drm_encoder.h b/drivers/gpu/drm/exynos/exynos_drm_encoder.h
index 97b087a51cb6..eb7d2316847e 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_encoder.h
+++ b/drivers/gpu/drm/exynos/exynos_drm_encoder.h
@@ -30,6 +30,7 @@
 
 struct exynos_drm_manager;
 
+void exynos_drm_encoder_setup(struct drm_device *dev);
 struct drm_encoder *exynos_drm_encoder_create(struct drm_device *dev,
 					       struct exynos_drm_manager *mgr,
 					       unsigned int possible_crtcs);
diff --git a/drivers/gpu/drm/exynos/exynos_drm_fb.c b/drivers/gpu/drm/exynos/exynos_drm_fb.c
index 3733fe6723d3..c38c8f468fa3 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_fb.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_fb.c
@@ -211,9 +211,3 @@ void exynos_drm_mode_config_init(struct drm_device *dev)
 
 	dev->mode_config.funcs = &exynos_drm_mode_config_funcs;
 }
-
-MODULE_AUTHOR("Inki Dae <inki.dae@samsung.com>");
-MODULE_AUTHOR("Joonyoung Shim <jy0922.shim@samsung.com>");
-MODULE_AUTHOR("Seung-Woo Kim <sw0312.kim@samsung.com>");
-MODULE_DESCRIPTION("Samsung SoC DRM FB Driver");
-MODULE_LICENSE("GPL");
diff --git a/drivers/gpu/drm/exynos/exynos_drm_fbdev.c b/drivers/gpu/drm/exynos/exynos_drm_fbdev.c
index d7ae29d2f3d6..d5586cc75163 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_fbdev.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_fbdev.c
@@ -46,39 +46,13 @@ struct exynos_drm_fbdev {
 	struct exynos_drm_gem_obj	*exynos_gem_obj;
 };
 
-static int exynos_drm_fbdev_set_par(struct fb_info *info)
-{
-	struct fb_var_screeninfo *var = &info->var;
-
-	switch (var->bits_per_pixel) {
-	case 32:
-	case 24:
-	case 18:
-	case 16:
-	case 12:
-		info->fix.visual = FB_VISUAL_TRUECOLOR;
-		break;
-	case 1:
-		info->fix.visual = FB_VISUAL_MONO01;
-		break;
-	default:
-		info->fix.visual = FB_VISUAL_PSEUDOCOLOR;
-		break;
-	}
-
-	info->fix.line_length = (var->xres_virtual * var->bits_per_pixel) / 8;
-
-	return drm_fb_helper_set_par(info);
-}
-
-
 static struct fb_ops exynos_drm_fb_ops = {
 	.owner		= THIS_MODULE,
 	.fb_fillrect	= cfb_fillrect,
 	.fb_copyarea	= cfb_copyarea,
 	.fb_imageblit	= cfb_imageblit,
 	.fb_check_var	= drm_fb_helper_check_var,
-	.fb_set_par	= exynos_drm_fbdev_set_par,
+	.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,
@@ -151,7 +125,9 @@ static int exynos_drm_fbdev_create(struct drm_fb_helper *helper,
 	}
 
 	size = mode_cmd.pitches[0] * mode_cmd.height;
-	exynos_gem_obj = exynos_drm_gem_create(dev, size);
+
+	/* 0 means to allocate physically continuous memory */
+	exynos_gem_obj = exynos_drm_gem_create(dev, 0, size);
 	if (IS_ERR(exynos_gem_obj)) {
 		ret = PTR_ERR(exynos_gem_obj);
 		goto out;
@@ -195,66 +171,6 @@ out:
 	return ret;
 }
 
-static bool
-exynos_drm_fbdev_is_samefb(struct drm_framebuffer *fb,
-			    struct drm_fb_helper_surface_size *sizes)
-{
-	if (fb->width != sizes->surface_width)
-		return false;
-	if (fb->height != sizes->surface_height)
-		return false;
-	if (fb->bits_per_pixel != sizes->surface_bpp)
-		return false;
-	if (fb->depth != sizes->surface_depth)
-		return false;
-
-	return true;
-}
-
-static int exynos_drm_fbdev_recreate(struct drm_fb_helper *helper,
-				      struct drm_fb_helper_surface_size *sizes)
-{
-	struct drm_device *dev = helper->dev;
-	struct exynos_drm_fbdev *exynos_fbdev = to_exynos_fbdev(helper);
-	struct exynos_drm_gem_obj *exynos_gem_obj;
-	struct drm_framebuffer *fb = helper->fb;
-	struct drm_mode_fb_cmd2 mode_cmd = { 0 };
-	unsigned long size;
-
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
-	if (exynos_drm_fbdev_is_samefb(fb, sizes))
-		return 0;
-
-	mode_cmd.width = sizes->surface_width;
-	mode_cmd.height = sizes->surface_height;
-	mode_cmd.pitches[0] = sizes->surface_width * (sizes->surface_bpp >> 3);
-	mode_cmd.pixel_format = drm_mode_legacy_fb_format(sizes->surface_bpp,
-							  sizes->surface_depth);
-
-	if (exynos_fbdev->exynos_gem_obj)
-		exynos_drm_gem_destroy(exynos_fbdev->exynos_gem_obj);
-
-	if (fb->funcs->destroy)
-		fb->funcs->destroy(fb);
-
-	size = mode_cmd.pitches[0] * mode_cmd.height;
-	exynos_gem_obj = exynos_drm_gem_create(dev, size);
-	if (IS_ERR(exynos_gem_obj))
-		return PTR_ERR(exynos_gem_obj);
-
-	exynos_fbdev->exynos_gem_obj = exynos_gem_obj;
-
-	helper->fb = exynos_drm_framebuffer_init(dev, &mode_cmd,
-			&exynos_gem_obj->base);
-	if (IS_ERR_OR_NULL(helper->fb)) {
-		DRM_ERROR("failed to create drm framebuffer.\n");
-		return PTR_ERR(helper->fb);
-	}
-
-	return exynos_drm_fbdev_update(helper, helper->fb);
-}
-
 static int exynos_drm_fbdev_probe(struct drm_fb_helper *helper,
 				   struct drm_fb_helper_surface_size *sizes)
 {
@@ -262,6 +178,10 @@ static int exynos_drm_fbdev_probe(struct drm_fb_helper *helper,
 
 	DRM_DEBUG_KMS("%s\n", __FILE__);
 
+	/*
+	 * with !helper->fb, it means that this funcion is called first time
+	 * and after that, the helper->fb would be used as clone mode.
+	 */
 	if (!helper->fb) {
 		ret = exynos_drm_fbdev_create(helper, sizes);
 		if (ret < 0) {
@@ -274,12 +194,6 @@ static int exynos_drm_fbdev_probe(struct drm_fb_helper *helper,
 		 * because register_framebuffer() should be called.
 		 */
 		ret = 1;
-	} else {
-		ret = exynos_drm_fbdev_recreate(helper, sizes);
-		if (ret < 0) {
-			DRM_ERROR("failed to reconfigure fbdev\n");
-			return ret;
-		}
 	}
 
 	return ret;
@@ -402,89 +316,3 @@ void exynos_drm_fbdev_restore_mode(struct drm_device *dev)
 
 	drm_fb_helper_restore_fbdev_mode(private->fb_helper);
 }
-
-int exynos_drm_fbdev_reinit(struct drm_device *dev)
-{
-	struct exynos_drm_private *private = dev->dev_private;
-	struct drm_fb_helper *fb_helper;
-	int ret;
-
-	if (!private)
-		return -EINVAL;
-
-	/*
-	 * if all sub drivers were unloaded then num_connector is 0
-	 * so at this time, the framebuffers also should be destroyed.
-	 */
-	if (!dev->mode_config.num_connector) {
-		exynos_drm_fbdev_fini(dev);
-		return 0;
-	}
-
-	fb_helper = private->fb_helper;
-
-	if (fb_helper) {
-		struct list_head temp_list;
-
-		INIT_LIST_HEAD(&temp_list);
-
-		/*
-		 * fb_helper is reintialized but kernel fb is reused
-		 * so kernel_fb_list need to be backuped and restored
-		 */
-		if (!list_empty(&fb_helper->kernel_fb_list))
-			list_replace_init(&fb_helper->kernel_fb_list,
-					&temp_list);
-
-		drm_fb_helper_fini(fb_helper);
-
-		ret = drm_fb_helper_init(dev, fb_helper,
-				dev->mode_config.num_crtc, MAX_CONNECTOR);
-		if (ret < 0) {
-			DRM_ERROR("failed to initialize drm fb helper\n");
-			return ret;
-		}
-
-		if (!list_empty(&temp_list))
-			list_replace(&temp_list, &fb_helper->kernel_fb_list);
-
-		ret = drm_fb_helper_single_add_all_connectors(fb_helper);
-		if (ret < 0) {
-			DRM_ERROR("failed to add fb helper to connectors\n");
-			goto err;
-		}
-
-		ret = drm_fb_helper_initial_config(fb_helper, PREFERRED_BPP);
-		if (ret < 0) {
-			DRM_ERROR("failed to set up hw configuration.\n");
-			goto err;
-		}
-	} else {
-		/*
-		 * if drm_load() failed whem drm load() was called prior
-		 * to specific drivers, fb_helper must be NULL and so
-		 * this fuction should be called again to re-initialize and
-		 * re-configure the fb helper. it means that this function
-		 * has been called by the specific drivers.
-		 */
-		ret = exynos_drm_fbdev_init(dev);
-	}
-
-	return ret;
-
-err:
-	/*
-	 * if drm_load() failed when drm load() was called prior
-	 * to specific drivers, the fb_helper must be NULL and so check it.
-	 */
-	if (fb_helper)
-		drm_fb_helper_fini(fb_helper);
-
-	return ret;
-}
-
-MODULE_AUTHOR("Inki Dae <inki.dae@samsung.com>");
-MODULE_AUTHOR("Joonyoung Shim <jy0922.shim@samsung.com>");
-MODULE_AUTHOR("Seung-Woo Kim <sw0312.kim@samsung.com>");
-MODULE_DESCRIPTION("Samsung SoC DRM FBDEV Driver");
-MODULE_LICENSE("GPL");
diff --git a/drivers/gpu/drm/exynos/exynos_drm_fimd.c b/drivers/gpu/drm/exynos/exynos_drm_fimd.c
index b6a737d196ae..ecb6db229700 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_fimd.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_fimd.c
@@ -89,7 +89,7 @@ struct fimd_context {
 	bool				suspended;
 	struct mutex			lock;
 
-	struct fb_videomode		*timing;
+	struct exynos_drm_panel_info *panel;
 };
 
 static bool fimd_display_is_connected(struct device *dev)
@@ -101,13 +101,13 @@ static bool fimd_display_is_connected(struct device *dev)
 	return true;
 }
 
-static void *fimd_get_timing(struct device *dev)
+static void *fimd_get_panel(struct device *dev)
 {
 	struct fimd_context *ctx = get_fimd_context(dev);
 
 	DRM_DEBUG_KMS("%s\n", __FILE__);
 
-	return ctx->timing;
+	return ctx->panel;
 }
 
 static int fimd_check_timing(struct device *dev, void *timing)
@@ -131,7 +131,7 @@ static int fimd_display_power_on(struct device *dev, int mode)
 static struct exynos_drm_display_ops fimd_display_ops = {
 	.type = EXYNOS_DISPLAY_TYPE_LCD,
 	.is_connected = fimd_display_is_connected,
-	.get_timing = fimd_get_timing,
+	.get_panel = fimd_get_panel,
 	.check_timing = fimd_check_timing,
 	.power_on = fimd_display_power_on,
 };
@@ -193,7 +193,8 @@ static void fimd_apply(struct device *subdrv_dev)
 static void fimd_commit(struct device *dev)
 {
 	struct fimd_context *ctx = get_fimd_context(dev);
-	struct fb_videomode *timing = ctx->timing;
+	struct exynos_drm_panel_info *panel = ctx->panel;
+	struct fb_videomode *timing = &panel->timing;
 	u32 val;
 
 	if (ctx->suspended)
@@ -604,7 +605,12 @@ static void fimd_finish_pageflip(struct drm_device *drm_dev, int crtc)
 	}
 
 	if (is_checked) {
-		drm_vblank_put(drm_dev, crtc);
+		/*
+		 * call drm_vblank_put only in case that drm_vblank_get was
+		 * called.
+		 */
+		if (atomic_read(&drm_dev->vblank_refcount[crtc]) > 0)
+			drm_vblank_put(drm_dev, crtc);
 
 		/*
 		 * don't off vblank if vblank_disable_allowed is 1,
@@ -781,7 +787,7 @@ static int __devinit fimd_probe(struct platform_device *pdev)
 	struct fimd_context *ctx;
 	struct exynos_drm_subdrv *subdrv;
 	struct exynos_drm_fimd_pdata *pdata;
-	struct fb_videomode *timing;
+	struct exynos_drm_panel_info *panel;
 	struct resource *res;
 	int win;
 	int ret = -EINVAL;
@@ -794,9 +800,9 @@ static int __devinit fimd_probe(struct platform_device *pdev)
 		return -EINVAL;
 	}
 
-	timing = &pdata->timing;
-	if (!timing) {
-		dev_err(dev, "timing is null.\n");
+	panel = &pdata->panel;
+	if (!panel) {
+		dev_err(dev, "panel is null.\n");
 		return -EINVAL;
 	}
 
@@ -811,8 +817,6 @@ static int __devinit fimd_probe(struct platform_device *pdev)
 		goto err_clk_get;
 	}
 
-	clk_enable(ctx->bus_clk);
-
 	ctx->lcd_clk = clk_get(dev, "sclk_fimd");
 	if (IS_ERR(ctx->lcd_clk)) {
 		dev_err(dev, "failed to get lcd clock\n");
@@ -820,8 +824,6 @@ static int __devinit fimd_probe(struct platform_device *pdev)
 		goto err_bus_clk;
 	}
 
-	clk_enable(ctx->lcd_clk);
-
 	res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
 	if (!res) {
 		dev_err(dev, "failed to find registers\n");
@@ -858,16 +860,10 @@ static int __devinit fimd_probe(struct platform_device *pdev)
 		goto err_req_irq;
 	}
 
-	ctx->clkdiv = fimd_calc_clkdiv(ctx, timing);
 	ctx->vidcon0 = pdata->vidcon0;
 	ctx->vidcon1 = pdata->vidcon1;
 	ctx->default_win = pdata->default_win;
-	ctx->timing = timing;
-
-	timing->pixclock = clk_get_rate(ctx->lcd_clk) / ctx->clkdiv;
-
-	DRM_DEBUG_KMS("pixel clock = %d, clkdiv = %d\n",
-			timing->pixclock, ctx->clkdiv);
+	ctx->panel = panel;
 
 	subdrv = &ctx->subdrv;
 
@@ -883,10 +879,15 @@ static int __devinit fimd_probe(struct platform_device *pdev)
 
 	platform_set_drvdata(pdev, ctx);
 
-	pm_runtime_set_active(dev);
 	pm_runtime_enable(dev);
 	pm_runtime_get_sync(dev);
 
+	ctx->clkdiv = fimd_calc_clkdiv(ctx, &panel->timing);
+	panel->timing.pixclock = clk_get_rate(ctx->lcd_clk) / ctx->clkdiv;
+
+	DRM_DEBUG_KMS("pixel clock = %d, clkdiv = %d\n",
+			panel->timing.pixclock, ctx->clkdiv);
+
 	for (win = 0; win < WINDOWS_NR; win++)
 		fimd_clear_win(ctx, win);
 
@@ -1006,7 +1007,7 @@ static const struct dev_pm_ops fimd_pm_ops = {
 	SET_RUNTIME_PM_OPS(fimd_runtime_suspend, fimd_runtime_resume, NULL)
 };
 
-static struct platform_driver fimd_driver = {
+struct platform_driver fimd_driver = {
 	.probe		= fimd_probe,
 	.remove		= __devexit_p(fimd_remove),
 	.driver		= {
@@ -1015,21 +1016,3 @@ static struct platform_driver fimd_driver = {
 		.pm	= &fimd_pm_ops,
 	},
 };
-
-static int __init fimd_init(void)
-{
-	return platform_driver_register(&fimd_driver);
-}
-
-static void __exit fimd_exit(void)
-{
-	platform_driver_unregister(&fimd_driver);
-}
-
-module_init(fimd_init);
-module_exit(fimd_exit);
-
-MODULE_AUTHOR("Joonyoung Shim <jy0922.shim@samsung.com>");
-MODULE_AUTHOR("Inki Dae <inki.dae@samsung.com>");
-MODULE_DESCRIPTION("Samsung DRM FIMD Driver");
-MODULE_LICENSE("GPL");
diff --git a/drivers/gpu/drm/exynos/exynos_drm_gem.c b/drivers/gpu/drm/exynos/exynos_drm_gem.c
index 025abb3e3b67..fa1aa94a3d8e 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_gem.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_gem.c
@@ -26,6 +26,7 @@
 #include "drmP.h"
 #include "drm.h"
 
+#include <linux/shmem_fs.h>
 #include <drm/exynos_drm.h>
 
 #include "exynos_drm_drv.h"
@@ -55,6 +56,178 @@ static unsigned int convert_to_vm_err_msg(int msg)
 	return out_msg;
 }
 
+static unsigned int mask_gem_flags(unsigned int flags)
+{
+	return flags &= EXYNOS_BO_NONCONTIG;
+}
+
+static struct page **exynos_gem_get_pages(struct drm_gem_object *obj,
+						gfp_t gfpmask)
+{
+	struct inode *inode;
+	struct address_space *mapping;
+	struct page *p, **pages;
+	int i, npages;
+
+	/* This is the shared memory object that backs the GEM resource */
+	inode = obj->filp->f_path.dentry->d_inode;
+	mapping = inode->i_mapping;
+
+	npages = obj->size >> PAGE_SHIFT;
+
+	pages = drm_malloc_ab(npages, sizeof(struct page *));
+	if (pages == NULL)
+		return ERR_PTR(-ENOMEM);
+
+	gfpmask |= mapping_gfp_mask(mapping);
+
+	for (i = 0; i < npages; i++) {
+		p = shmem_read_mapping_page_gfp(mapping, i, gfpmask);
+		if (IS_ERR(p))
+			goto fail;
+		pages[i] = p;
+	}
+
+	return pages;
+
+fail:
+	while (i--)
+		page_cache_release(pages[i]);
+
+	drm_free_large(pages);
+	return ERR_PTR(PTR_ERR(p));
+}
+
+static void exynos_gem_put_pages(struct drm_gem_object *obj,
+					struct page **pages,
+					bool dirty, bool accessed)
+{
+	int i, npages;
+
+	npages = obj->size >> PAGE_SHIFT;
+
+	for (i = 0; i < npages; i++) {
+		if (dirty)
+			set_page_dirty(pages[i]);
+
+		if (accessed)
+			mark_page_accessed(pages[i]);
+
+		/* Undo the reference we took when populating the table */
+		page_cache_release(pages[i]);
+	}
+
+	drm_free_large(pages);
+}
+
+static int exynos_drm_gem_map_pages(struct drm_gem_object *obj,
+					struct vm_area_struct *vma,
+					unsigned long f_vaddr,
+					pgoff_t page_offset)
+{
+	struct exynos_drm_gem_obj *exynos_gem_obj = to_exynos_gem_obj(obj);
+	struct exynos_drm_gem_buf *buf = exynos_gem_obj->buffer;
+	unsigned long pfn;
+
+	if (exynos_gem_obj->flags & EXYNOS_BO_NONCONTIG) {
+		unsigned long usize = buf->size;
+
+		if (!buf->pages)
+			return -EINTR;
+
+		while (usize > 0) {
+			pfn = page_to_pfn(buf->pages[page_offset++]);
+			vm_insert_mixed(vma, f_vaddr, pfn);
+			f_vaddr += PAGE_SIZE;
+			usize -= PAGE_SIZE;
+		}
+
+		return 0;
+	}
+
+	pfn = (buf->dma_addr >> PAGE_SHIFT) + page_offset;
+
+	return vm_insert_mixed(vma, f_vaddr, pfn);
+}
+
+static int exynos_drm_gem_get_pages(struct drm_gem_object *obj)
+{
+	struct exynos_drm_gem_obj *exynos_gem_obj = to_exynos_gem_obj(obj);
+	struct exynos_drm_gem_buf *buf = exynos_gem_obj->buffer;
+	struct scatterlist *sgl;
+	struct page **pages;
+	unsigned int npages, i = 0;
+	int ret;
+
+	if (buf->pages) {
+		DRM_DEBUG_KMS("already allocated.\n");
+		return -EINVAL;
+	}
+
+	pages = exynos_gem_get_pages(obj, GFP_KERNEL);
+	if (IS_ERR(pages)) {
+		DRM_ERROR("failed to get pages.\n");
+		return PTR_ERR(pages);
+	}
+
+	npages = obj->size >> PAGE_SHIFT;
+
+	buf->sgt = kzalloc(sizeof(struct sg_table), GFP_KERNEL);
+	if (!buf->sgt) {
+		DRM_ERROR("failed to allocate sg table.\n");
+		ret = -ENOMEM;
+		goto err;
+	}
+
+	ret = sg_alloc_table(buf->sgt, npages, GFP_KERNEL);
+	if (ret < 0) {
+		DRM_ERROR("failed to initialize sg table.\n");
+		ret = -EFAULT;
+		goto err1;
+	}
+
+	sgl = buf->sgt->sgl;
+
+	/* set all pages to sg list. */
+	while (i < npages) {
+		sg_set_page(sgl, pages[i], PAGE_SIZE, 0);
+		sg_dma_address(sgl) = page_to_phys(pages[i]);
+		i++;
+		sgl = sg_next(sgl);
+	}
+
+	/* add some codes for UNCACHED type here. TODO */
+
+	buf->pages = pages;
+	return ret;
+err1:
+	kfree(buf->sgt);
+	buf->sgt = NULL;
+err:
+	exynos_gem_put_pages(obj, pages, true, false);
+	return ret;
+
+}
+
+static void exynos_drm_gem_put_pages(struct drm_gem_object *obj)
+{
+	struct exynos_drm_gem_obj *exynos_gem_obj = to_exynos_gem_obj(obj);
+	struct exynos_drm_gem_buf *buf = exynos_gem_obj->buffer;
+
+	/*
+	 * if buffer typs is EXYNOS_BO_NONCONTIG then release all pages
+	 * allocated at gem fault handler.
+	 */
+	sg_free_table(buf->sgt);
+	kfree(buf->sgt);
+	buf->sgt = NULL;
+
+	exynos_gem_put_pages(obj, buf->pages, true, false);
+	buf->pages = NULL;
+
+	/* add some codes for UNCACHED type here. TODO */
+}
+
 static int exynos_drm_gem_handle_create(struct drm_gem_object *obj,
 					struct drm_file *file_priv,
 					unsigned int *handle)
@@ -90,7 +263,15 @@ void exynos_drm_gem_destroy(struct exynos_drm_gem_obj *exynos_gem_obj)
 
 	DRM_DEBUG_KMS("handle count = %d\n", atomic_read(&obj->handle_count));
 
-	exynos_drm_buf_destroy(obj->dev, exynos_gem_obj->buffer);
+	if ((exynos_gem_obj->flags & EXYNOS_BO_NONCONTIG) &&
+			exynos_gem_obj->buffer->pages)
+		exynos_drm_gem_put_pages(obj);
+	else
+		exynos_drm_free_buf(obj->dev, exynos_gem_obj->flags,
+					exynos_gem_obj->buffer);
+
+	exynos_drm_fini_buf(obj->dev, exynos_gem_obj->buffer);
+	exynos_gem_obj->buffer = NULL;
 
 	if (obj->map_list.map)
 		drm_gem_free_mmap_offset(obj);
@@ -99,6 +280,7 @@ void exynos_drm_gem_destroy(struct exynos_drm_gem_obj *exynos_gem_obj)
 	drm_gem_object_release(obj);
 
 	kfree(exynos_gem_obj);
+	exynos_gem_obj = NULL;
 }
 
 static struct exynos_drm_gem_obj *exynos_drm_gem_init(struct drm_device *dev,
@@ -114,6 +296,7 @@ static struct exynos_drm_gem_obj *exynos_drm_gem_init(struct drm_device *dev,
 		return NULL;
 	}
 
+	exynos_gem_obj->size = size;
 	obj = &exynos_gem_obj->base;
 
 	ret = drm_gem_object_init(dev, obj, size);
@@ -129,27 +312,55 @@ static struct exynos_drm_gem_obj *exynos_drm_gem_init(struct drm_device *dev,
 }
 
 struct exynos_drm_gem_obj *exynos_drm_gem_create(struct drm_device *dev,
-						 unsigned long size)
+						unsigned int flags,
+						unsigned long size)
 {
-	struct exynos_drm_gem_buf *buffer;
 	struct exynos_drm_gem_obj *exynos_gem_obj;
+	struct exynos_drm_gem_buf *buf;
+	int ret;
 
 	size = roundup(size, PAGE_SIZE);
 	DRM_DEBUG_KMS("%s: size = 0x%lx\n", __FILE__, size);
 
-	buffer = exynos_drm_buf_create(dev, size);
-	if (!buffer)
+	flags = mask_gem_flags(flags);
+
+	buf = exynos_drm_init_buf(dev, size);
+	if (!buf)
 		return ERR_PTR(-ENOMEM);
 
 	exynos_gem_obj = exynos_drm_gem_init(dev, size);
 	if (!exynos_gem_obj) {
-		exynos_drm_buf_destroy(dev, buffer);
-		return ERR_PTR(-ENOMEM);
+		ret = -ENOMEM;
+		goto err;
 	}
 
-	exynos_gem_obj->buffer = buffer;
+	exynos_gem_obj->buffer = buf;
+
+	/* set memory type and cache attribute from user side. */
+	exynos_gem_obj->flags = flags;
+
+	/*
+	 * allocate all pages as desired size if user wants to allocate
+	 * physically non-continuous memory.
+	 */
+	if (flags & EXYNOS_BO_NONCONTIG) {
+		ret = exynos_drm_gem_get_pages(&exynos_gem_obj->base);
+		if (ret < 0) {
+			drm_gem_object_release(&exynos_gem_obj->base);
+			goto err;
+		}
+	} else {
+		ret = exynos_drm_alloc_buf(dev, buf, flags);
+		if (ret < 0) {
+			drm_gem_object_release(&exynos_gem_obj->base);
+			goto err;
+		}
+	}
 
 	return exynos_gem_obj;
+err:
+	exynos_drm_fini_buf(dev, buf);
+	return ERR_PTR(ret);
 }
 
 int exynos_drm_gem_create_ioctl(struct drm_device *dev, void *data,
@@ -161,7 +372,7 @@ int exynos_drm_gem_create_ioctl(struct drm_device *dev, void *data,
 
 	DRM_DEBUG_KMS("%s\n", __FILE__);
 
-	exynos_gem_obj = exynos_drm_gem_create(dev, args->size);
+	exynos_gem_obj = exynos_drm_gem_create(dev, args->flags, args->size);
 	if (IS_ERR(exynos_gem_obj))
 		return PTR_ERR(exynos_gem_obj);
 
@@ -175,6 +386,64 @@ int exynos_drm_gem_create_ioctl(struct drm_device *dev, void *data,
 	return 0;
 }
 
+void *exynos_drm_gem_get_dma_addr(struct drm_device *dev,
+					unsigned int gem_handle,
+					struct drm_file *file_priv)
+{
+	struct exynos_drm_gem_obj *exynos_gem_obj;
+	struct drm_gem_object *obj;
+
+	obj = drm_gem_object_lookup(dev, file_priv, gem_handle);
+	if (!obj) {
+		DRM_ERROR("failed to lookup gem object.\n");
+		return ERR_PTR(-EINVAL);
+	}
+
+	exynos_gem_obj = to_exynos_gem_obj(obj);
+
+	if (exynos_gem_obj->flags & EXYNOS_BO_NONCONTIG) {
+		DRM_DEBUG_KMS("not support NONCONTIG type.\n");
+		drm_gem_object_unreference_unlocked(obj);
+
+		/* TODO */
+		return ERR_PTR(-EINVAL);
+	}
+
+	return &exynos_gem_obj->buffer->dma_addr;
+}
+
+void exynos_drm_gem_put_dma_addr(struct drm_device *dev,
+					unsigned int gem_handle,
+					struct drm_file *file_priv)
+{
+	struct exynos_drm_gem_obj *exynos_gem_obj;
+	struct drm_gem_object *obj;
+
+	obj = drm_gem_object_lookup(dev, file_priv, gem_handle);
+	if (!obj) {
+		DRM_ERROR("failed to lookup gem object.\n");
+		return;
+	}
+
+	exynos_gem_obj = to_exynos_gem_obj(obj);
+
+	if (exynos_gem_obj->flags & EXYNOS_BO_NONCONTIG) {
+		DRM_DEBUG_KMS("not support NONCONTIG type.\n");
+		drm_gem_object_unreference_unlocked(obj);
+
+		/* TODO */
+		return;
+	}
+
+	drm_gem_object_unreference_unlocked(obj);
+
+	/*
+	 * decrease obj->refcount one more time because we has already
+	 * increased it at exynos_drm_gem_get_dma_addr().
+	 */
+	drm_gem_object_unreference_unlocked(obj);
+}
+
 int exynos_drm_gem_map_offset_ioctl(struct drm_device *dev, void *data,
 				    struct drm_file *file_priv)
 {
@@ -200,7 +469,8 @@ static int exynos_drm_gem_mmap_buffer(struct file *filp,
 	struct drm_gem_object *obj = filp->private_data;
 	struct exynos_drm_gem_obj *exynos_gem_obj = to_exynos_gem_obj(obj);
 	struct exynos_drm_gem_buf *buffer;
-	unsigned long pfn, vm_size;
+	unsigned long pfn, vm_size, usize, uaddr = vma->vm_start;
+	int ret;
 
 	DRM_DEBUG_KMS("%s\n", __FILE__);
 
@@ -208,9 +478,9 @@ static int exynos_drm_gem_mmap_buffer(struct file *filp,
 
 	/* in case of direct mapping, always having non-cachable attribute */
 	vma->vm_page_prot = pgprot_noncached(vma->vm_page_prot);
-	vma->vm_file = filp;
 
-	vm_size = vma->vm_end - vma->vm_start;
+	vm_size = usize = vma->vm_end - vma->vm_start;
+
 	/*
 	 * a buffer contains information to physically continuous memory
 	 * allocated by user request or at framebuffer creation.
@@ -221,18 +491,37 @@ static int exynos_drm_gem_mmap_buffer(struct file *filp,
 	if (vm_size > buffer->size)
 		return -EINVAL;
 
-	/*
-	 * get page frame number to physical memory to be mapped
-	 * to user space.
-	 */
-	pfn = ((unsigned long)exynos_gem_obj->buffer->dma_addr) >> PAGE_SHIFT;
-
-	DRM_DEBUG_KMS("pfn = 0x%lx\n", pfn);
-
-	if (remap_pfn_range(vma, vma->vm_start, pfn, vm_size,
-				vma->vm_page_prot)) {
-		DRM_ERROR("failed to remap pfn range.\n");
-		return -EAGAIN;
+	if (exynos_gem_obj->flags & EXYNOS_BO_NONCONTIG) {
+		int i = 0;
+
+		if (!buffer->pages)
+			return -EINVAL;
+
+		do {
+			ret = vm_insert_page(vma, uaddr, buffer->pages[i++]);
+			if (ret) {
+				DRM_ERROR("failed to remap user space.\n");
+				return ret;
+			}
+
+			uaddr += PAGE_SIZE;
+			usize -= PAGE_SIZE;
+		} while (usize > 0);
+	} else {
+		/*
+		 * get page frame number to physical memory to be mapped
+		 * to user space.
+		 */
+		pfn = ((unsigned long)exynos_gem_obj->buffer->dma_addr) >>
+								PAGE_SHIFT;
+
+		DRM_DEBUG_KMS("pfn = 0x%lx\n", pfn);
+
+		if (remap_pfn_range(vma, vma->vm_start, pfn, vm_size,
+					vma->vm_page_prot)) {
+			DRM_ERROR("failed to remap pfn range.\n");
+			return -EAGAIN;
+		}
 	}
 
 	return 0;
@@ -312,9 +601,9 @@ int exynos_drm_gem_dumb_create(struct drm_file *file_priv,
 	 */
 
 	args->pitch = args->width * args->bpp >> 3;
-	args->size = args->pitch * args->height;
+	args->size = PAGE_ALIGN(args->pitch * args->height);
 
-	exynos_gem_obj = exynos_drm_gem_create(dev, args->size);
+	exynos_gem_obj = exynos_drm_gem_create(dev, args->flags, args->size);
 	if (IS_ERR(exynos_gem_obj))
 		return PTR_ERR(exynos_gem_obj);
 
@@ -398,20 +687,31 @@ int exynos_drm_gem_fault(struct vm_area_struct *vma, struct vm_fault *vmf)
 	struct drm_gem_object *obj = vma->vm_private_data;
 	struct exynos_drm_gem_obj *exynos_gem_obj = to_exynos_gem_obj(obj);
 	struct drm_device *dev = obj->dev;
-	unsigned long pfn;
+	unsigned long f_vaddr;
 	pgoff_t page_offset;
 	int ret;
 
 	page_offset = ((unsigned long)vmf->virtual_address -
 			vma->vm_start) >> PAGE_SHIFT;
+	f_vaddr = (unsigned long)vmf->virtual_address;
 
 	mutex_lock(&dev->struct_mutex);
 
-	pfn = (((unsigned long)exynos_gem_obj->buffer->dma_addr) >>
-			PAGE_SHIFT) + page_offset;
+	/*
+	 * allocate all pages as desired size if user wants to allocate
+	 * physically non-continuous memory.
+	 */
+	if (exynos_gem_obj->flags & EXYNOS_BO_NONCONTIG) {
+		ret = exynos_drm_gem_get_pages(obj);
+		if (ret < 0)
+			goto err;
+	}
 
-	ret = vm_insert_mixed(vma, (unsigned long)vmf->virtual_address, pfn);
+	ret = exynos_drm_gem_map_pages(obj, vma, f_vaddr, page_offset);
+	if (ret < 0)
+		DRM_ERROR("failed to map pages.\n");
 
+err:
 	mutex_unlock(&dev->struct_mutex);
 
 	return convert_to_vm_err_msg(ret);
@@ -435,7 +735,3 @@ int exynos_drm_gem_mmap(struct file *filp, struct vm_area_struct *vma)
 
 	return ret;
 }
-
-MODULE_AUTHOR("Inki Dae <inki.dae@samsung.com>");
-MODULE_DESCRIPTION("Samsung SoC DRM GEM Module");
-MODULE_LICENSE("GPL");
diff --git a/drivers/gpu/drm/exynos/exynos_drm_gem.h b/drivers/gpu/drm/exynos/exynos_drm_gem.h
index 67cdc9168708..e40fbad8b705 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_gem.h
+++ b/drivers/gpu/drm/exynos/exynos_drm_gem.h
@@ -36,11 +36,15 @@
  * @dma_addr: bus address(accessed by dma) to allocated memory region.
  *	- this address could be physical address without IOMMU and
  *	device address with IOMMU.
+ * @sgt: sg table to transfer page data.
+ * @pages: contain all pages to allocated memory region.
  * @size: size of allocated memory region.
  */
 struct exynos_drm_gem_buf {
 	void __iomem		*kvaddr;
 	dma_addr_t		dma_addr;
+	struct sg_table		*sgt;
+	struct page		**pages;
 	unsigned long		size;
 };
 
@@ -55,6 +59,8 @@ struct exynos_drm_gem_buf {
  *	by user request or at framebuffer creation.
  *	continuous memory region allocated by user request
  *	or at framebuffer creation.
+ * @size: total memory size to physically non-continuous memory region.
+ * @flags: indicate memory type to allocated buffer and cache attruibute.
  *
  * P.S. this object would be transfered to user as kms_bo.handle so
  *	user can access the buffer through kms_bo.handle.
@@ -62,6 +68,8 @@ struct exynos_drm_gem_buf {
 struct exynos_drm_gem_obj {
 	struct drm_gem_object		base;
 	struct exynos_drm_gem_buf	*buffer;
+	unsigned long			size;
+	unsigned int			flags;
 };
 
 /* destroy a buffer with gem object */
@@ -69,7 +77,8 @@ void exynos_drm_gem_destroy(struct exynos_drm_gem_obj *exynos_gem_obj);
 
 /* create a new buffer with gem object */
 struct exynos_drm_gem_obj *exynos_drm_gem_create(struct drm_device *dev,
-						 unsigned long size);
+						unsigned int flags,
+						unsigned long size);
 
 /*
  * request gem object creation and buffer allocation as the size
@@ -79,6 +88,24 @@ struct exynos_drm_gem_obj *exynos_drm_gem_create(struct drm_device *dev,
 int exynos_drm_gem_create_ioctl(struct drm_device *dev, void *data,
 				struct drm_file *file_priv);
 
+/*
+ * get dma address from gem handle and this function could be used for
+ * other drivers such as 2d/3d acceleration drivers.
+ * with this function call, gem object reference count would be increased.
+ */
+void *exynos_drm_gem_get_dma_addr(struct drm_device *dev,
+					unsigned int gem_handle,
+					struct drm_file *file_priv);
+
+/*
+ * put dma address from gem handle and this function could be used for
+ * other drivers such as 2d/3d acceleration drivers.
+ * with this function call, gem object reference count would be decreased.
+ */
+void exynos_drm_gem_put_dma_addr(struct drm_device *dev,
+					unsigned int gem_handle,
+					struct drm_file *file_priv);
+
 /* get buffer offset to map to user space. */
 int exynos_drm_gem_map_offset_ioctl(struct drm_device *dev, void *data,
 				    struct drm_file *file_priv);
diff --git a/drivers/gpu/drm/exynos/exynos_drm_hdmi.c b/drivers/gpu/drm/exynos/exynos_drm_hdmi.c
index ed8a319ed84b..14eb26b0ba1c 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_hdmi.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_hdmi.c
@@ -38,7 +38,6 @@ struct drm_hdmi_context {
 	struct exynos_drm_subdrv	subdrv;
 	struct exynos_drm_hdmi_context	*hdmi_ctx;
 	struct exynos_drm_hdmi_context	*mixer_ctx;
-	struct work_struct		work;
 };
 
 void exynos_drm_display_ops_register(struct exynos_hdmi_display_ops
@@ -49,7 +48,6 @@ void exynos_drm_display_ops_register(struct exynos_hdmi_display_ops
 	if (display_ops)
 		hdmi_display_ops = display_ops;
 }
-EXPORT_SYMBOL(exynos_drm_display_ops_register);
 
 void exynos_drm_manager_ops_register(struct exynos_hdmi_manager_ops
 					*manager_ops)
@@ -59,7 +57,6 @@ void exynos_drm_manager_ops_register(struct exynos_hdmi_manager_ops
 	if (manager_ops)
 		hdmi_manager_ops = manager_ops;
 }
-EXPORT_SYMBOL(exynos_drm_manager_ops_register);
 
 void exynos_drm_overlay_ops_register(struct exynos_hdmi_overlay_ops
 					*overlay_ops)
@@ -69,7 +66,6 @@ void exynos_drm_overlay_ops_register(struct exynos_hdmi_overlay_ops
 	if (overlay_ops)
 		hdmi_overlay_ops = overlay_ops;
 }
-EXPORT_SYMBOL(exynos_drm_overlay_ops_register);
 
 static bool drm_hdmi_is_connected(struct device *dev)
 {
@@ -155,6 +151,20 @@ static void drm_hdmi_disable_vblank(struct device *subdrv_dev)
 		return hdmi_overlay_ops->disable_vblank(ctx->mixer_ctx->ctx);
 }
 
+static void drm_hdmi_mode_fixup(struct device *subdrv_dev,
+				struct drm_connector *connector,
+				struct drm_display_mode *mode,
+				struct drm_display_mode *adjusted_mode)
+{
+	struct drm_hdmi_context *ctx = to_context(subdrv_dev);
+
+	DRM_DEBUG_KMS("%s\n", __FILE__);
+
+	if (hdmi_manager_ops && hdmi_manager_ops->mode_fixup)
+		hdmi_manager_ops->mode_fixup(ctx->hdmi_ctx->ctx, connector,
+						mode, adjusted_mode);
+}
+
 static void drm_hdmi_mode_set(struct device *subdrv_dev, void *mode)
 {
 	struct drm_hdmi_context *ctx = to_context(subdrv_dev);
@@ -165,6 +175,18 @@ static void drm_hdmi_mode_set(struct device *subdrv_dev, void *mode)
 		hdmi_manager_ops->mode_set(ctx->hdmi_ctx->ctx, mode);
 }
 
+static void drm_hdmi_get_max_resol(struct device *subdrv_dev,
+				unsigned int *width, unsigned int *height)
+{
+	struct drm_hdmi_context *ctx = to_context(subdrv_dev);
+
+	DRM_DEBUG_KMS("%s\n", __FILE__);
+
+	if (hdmi_manager_ops && hdmi_manager_ops->get_max_resol)
+		hdmi_manager_ops->get_max_resol(ctx->hdmi_ctx->ctx, width,
+							height);
+}
+
 static void drm_hdmi_commit(struct device *subdrv_dev)
 {
 	struct drm_hdmi_context *ctx = to_context(subdrv_dev);
@@ -200,7 +222,9 @@ static struct exynos_drm_manager_ops drm_hdmi_manager_ops = {
 	.dpms = drm_hdmi_dpms,
 	.enable_vblank = drm_hdmi_enable_vblank,
 	.disable_vblank = drm_hdmi_disable_vblank,
+	.mode_fixup = drm_hdmi_mode_fixup,
 	.mode_set = drm_hdmi_mode_set,
+	.get_max_resol = drm_hdmi_get_max_resol,
 	.commit = drm_hdmi_commit,
 };
 
@@ -249,7 +273,6 @@ static int hdmi_subdrv_probe(struct drm_device *drm_dev,
 	struct drm_hdmi_context *ctx;
 	struct platform_device *pdev = to_platform_device(dev);
 	struct exynos_drm_common_hdmi_pd *pd;
-	int ret;
 
 	DRM_DEBUG_KMS("%s\n", __FILE__);
 
@@ -270,26 +293,13 @@ static int hdmi_subdrv_probe(struct drm_device *drm_dev,
 		return -EFAULT;
 	}
 
-	ret = platform_driver_register(&hdmi_driver);
-	if (ret) {
-		DRM_DEBUG_KMS("failed to register hdmi driver.\n");
-		return ret;
-	}
-
-	ret = platform_driver_register(&mixer_driver);
-	if (ret) {
-		DRM_DEBUG_KMS("failed to register mixer driver.\n");
-		goto err_hdmidrv;
-	}
-
 	ctx = get_ctx_from_subdrv(subdrv);
 
 	ctx->hdmi_ctx = (struct exynos_drm_hdmi_context *)
 				to_context(pd->hdmi_dev);
 	if (!ctx->hdmi_ctx) {
 		DRM_DEBUG_KMS("hdmi context is null.\n");
-		ret = -EFAULT;
-		goto err_mixerdrv;
+		return -EFAULT;
 	}
 
 	ctx->hdmi_ctx->drm_dev = drm_dev;
@@ -298,42 +308,12 @@ static int hdmi_subdrv_probe(struct drm_device *drm_dev,
 				to_context(pd->mixer_dev);
 	if (!ctx->mixer_ctx) {
 		DRM_DEBUG_KMS("mixer context is null.\n");
-		ret = -EFAULT;
-		goto err_mixerdrv;
+		return -EFAULT;
 	}
 
 	ctx->mixer_ctx->drm_dev = drm_dev;
 
 	return 0;
-
-err_mixerdrv:
-	platform_driver_unregister(&mixer_driver);
-err_hdmidrv:
-	platform_driver_unregister(&hdmi_driver);
-	return ret;
-}
-
-static void hdmi_subdrv_remove(struct drm_device *drm_dev)
-{
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
-	platform_driver_unregister(&hdmi_driver);
-	platform_driver_unregister(&mixer_driver);
-}
-
-static void exynos_drm_hdmi_late_probe(struct work_struct *work)
-{
-	struct drm_hdmi_context *ctx = container_of(work,
-				struct drm_hdmi_context, work);
-
-	/*
-	 * this function calls subdrv->probe() so this must be called
-	 * after probe context.
-	 *
-	 * PS. subdrv->probe() will call platform_driver_register() to probe
-	 * hdmi and mixer driver.
-	 */
-	exynos_drm_subdrv_register(&ctx->subdrv);
 }
 
 static int __devinit exynos_drm_hdmi_probe(struct platform_device *pdev)
@@ -353,7 +333,6 @@ static int __devinit exynos_drm_hdmi_probe(struct platform_device *pdev)
 	subdrv = &ctx->subdrv;
 
 	subdrv->probe = hdmi_subdrv_probe;
-	subdrv->remove = hdmi_subdrv_remove;
 	subdrv->manager.pipe = -1;
 	subdrv->manager.ops = &drm_hdmi_manager_ops;
 	subdrv->manager.overlay_ops = &drm_hdmi_overlay_ops;
@@ -362,9 +341,7 @@ static int __devinit exynos_drm_hdmi_probe(struct platform_device *pdev)
 
 	platform_set_drvdata(pdev, subdrv);
 
-	INIT_WORK(&ctx->work, exynos_drm_hdmi_late_probe);
-
-	schedule_work(&ctx->work);
+	exynos_drm_subdrv_register(subdrv);
 
 	return 0;
 }
@@ -400,7 +377,7 @@ static int __devexit exynos_drm_hdmi_remove(struct platform_device *pdev)
 	return 0;
 }
 
-static struct platform_driver exynos_drm_common_hdmi_driver = {
+struct platform_driver exynos_drm_common_hdmi_driver = {
 	.probe		= exynos_drm_hdmi_probe,
 	.remove		= __devexit_p(exynos_drm_hdmi_remove),
 	.driver		= {
@@ -409,31 +386,3 @@ static struct platform_driver exynos_drm_common_hdmi_driver = {
 		.pm = &hdmi_pm_ops,
 	},
 };
-
-static int __init exynos_drm_hdmi_init(void)
-{
-	int ret;
-
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
-	ret = platform_driver_register(&exynos_drm_common_hdmi_driver);
-	if (ret) {
-		DRM_DEBUG_KMS("failed to register hdmi common driver.\n");
-		return ret;
-	}
-
-	return ret;
-}
-
-static void __exit exynos_drm_hdmi_exit(void)
-{
-	platform_driver_unregister(&exynos_drm_common_hdmi_driver);
-}
-
-module_init(exynos_drm_hdmi_init);
-module_exit(exynos_drm_hdmi_exit);
-
-MODULE_AUTHOR("Inki Dae <inki.dae@samsung.com>");
-MODULE_AUTHOR("Seung-Woo Kim, <sw0312.kim@samsung.com>");
-MODULE_DESCRIPTION("Samsung SoC DRM HDMI Driver");
-MODULE_LICENSE("GPL");
diff --git a/drivers/gpu/drm/exynos/exynos_drm_hdmi.h b/drivers/gpu/drm/exynos/exynos_drm_hdmi.h
index 3c29f790ee45..44497cfb6c74 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_hdmi.h
+++ b/drivers/gpu/drm/exynos/exynos_drm_hdmi.h
@@ -47,7 +47,12 @@ struct exynos_hdmi_display_ops {
 };
 
 struct exynos_hdmi_manager_ops {
+	void (*mode_fixup)(void *ctx, struct drm_connector *connector,
+				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);
 	void (*commit)(void *ctx);
 	void (*disable)(void *ctx);
 };
diff --git a/drivers/gpu/drm/exynos/exynos_drm_plane.c b/drivers/gpu/drm/exynos/exynos_drm_plane.c
index bdcf770aa22e..c277a3a445f5 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_plane.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_plane.c
@@ -22,6 +22,10 @@ struct exynos_plane {
 	bool				enabled;
 };
 
+static const uint32_t formats[] = {
+	DRM_FORMAT_XRGB8888,
+};
+
 static int
 exynos_update_plane(struct drm_plane *plane, struct drm_crtc *crtc,
 		     struct drm_framebuffer *fb, int crtc_x, int crtc_y,
@@ -115,9 +119,9 @@ int exynos_plane_init(struct drm_device *dev, unsigned int nr)
 
 	exynos_plane->overlay.zpos = DEFAULT_ZPOS;
 
-	/* TODO: format */
 	return drm_plane_init(dev, &exynos_plane->base, possible_crtcs,
-			      &exynos_plane_funcs, NULL, 0, false);
+			      &exynos_plane_funcs, formats, ARRAY_SIZE(formats),
+			      false);
 }
 
 int exynos_plane_set_zpos_ioctl(struct drm_device *dev, void *data,
diff --git a/drivers/gpu/drm/exynos/exynos_drm_vidi.c b/drivers/gpu/drm/exynos/exynos_drm_vidi.c
new file mode 100644
index 000000000000..8e1339f9fe1f
--- /dev/null
+++ b/drivers/gpu/drm/exynos/exynos_drm_vidi.c
@@ -0,0 +1,676 @@
+/* exynos_drm_vidi.c
+ *
+ * Copyright (C) 2012 Samsung Electronics Co.Ltd
+ * Authors:
+ *	Inki Dae <inki.dae@samsung.com>
+ *
+ * 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.
+ *
+ */
+#include "drmP.h"
+
+#include <linux/kernel.h>
+#include <linux/module.h>
+#include <linux/platform_device.h>
+
+#include <drm/exynos_drm.h>
+
+#include "drm_edid.h"
+#include "drm_crtc_helper.h"
+
+#include "exynos_drm_drv.h"
+#include "exynos_drm_crtc.h"
+#include "exynos_drm_encoder.h"
+
+/* vidi has totally three virtual windows. */
+#define WINDOWS_NR		3
+
+#define get_vidi_context(dev)	platform_get_drvdata(to_platform_device(dev))
+
+struct vidi_win_data {
+	unsigned int		offset_x;
+	unsigned int		offset_y;
+	unsigned int		ovl_width;
+	unsigned int		ovl_height;
+	unsigned int		fb_width;
+	unsigned int		fb_height;
+	unsigned int		bpp;
+	dma_addr_t		dma_addr;
+	void __iomem		*vaddr;
+	unsigned int		buf_offsize;
+	unsigned int		line_size;	/* bytes */
+	bool			enabled;
+};
+
+struct vidi_context {
+	struct exynos_drm_subdrv	subdrv;
+	struct drm_crtc			*crtc;
+	struct vidi_win_data		win_data[WINDOWS_NR];
+	struct edid			*raw_edid;
+	unsigned int			clkdiv;
+	unsigned int			default_win;
+	unsigned long			irq_flags;
+	unsigned int			connected;
+	bool				vblank_on;
+	bool				suspended;
+	struct work_struct		work;
+	struct mutex			lock;
+};
+
+static const char fake_edid_info[] = {
+	0x00, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0x00, 0x4c, 0x2d, 0x05, 0x05,
+	0x00, 0x00, 0x00, 0x00, 0x30, 0x12, 0x01, 0x03, 0x80, 0x10, 0x09, 0x78,
+	0x0a, 0xee, 0x91, 0xa3, 0x54, 0x4c, 0x99, 0x26, 0x0f, 0x50, 0x54, 0xbd,
+	0xee, 0x00, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01,
+	0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x66, 0x21, 0x50, 0xb0, 0x51, 0x00,
+	0x1b, 0x30, 0x40, 0x70, 0x36, 0x00, 0xa0, 0x5a, 0x00, 0x00, 0x00, 0x1e,
+	0x01, 0x1d, 0x00, 0x72, 0x51, 0xd0, 0x1e, 0x20, 0x6e, 0x28, 0x55, 0x00,
+	0xa0, 0x5a, 0x00, 0x00, 0x00, 0x1e, 0x00, 0x00, 0x00, 0xfd, 0x00, 0x18,
+	0x4b, 0x1a, 0x44, 0x17, 0x00, 0x0a, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20,
+	0x00, 0x00, 0x00, 0xfc, 0x00, 0x53, 0x41, 0x4d, 0x53, 0x55, 0x4e, 0x47,
+	0x0a, 0x20, 0x20, 0x20, 0x20, 0x20, 0x01, 0xbc, 0x02, 0x03, 0x1e, 0xf1,
+	0x46, 0x84, 0x05, 0x03, 0x10, 0x20, 0x22, 0x23, 0x09, 0x07, 0x07, 0x83,
+	0x01, 0x00, 0x00, 0xe2, 0x00, 0x0f, 0x67, 0x03, 0x0c, 0x00, 0x10, 0x00,
+	0xb8, 0x2d, 0x01, 0x1d, 0x80, 0x18, 0x71, 0x1c, 0x16, 0x20, 0x58, 0x2c,
+	0x25, 0x00, 0xa0, 0x5a, 0x00, 0x00, 0x00, 0x9e, 0x8c, 0x0a, 0xd0, 0x8a,
+	0x20, 0xe0, 0x2d, 0x10, 0x10, 0x3e, 0x96, 0x00, 0xa0, 0x5a, 0x00, 0x00,
+	0x00, 0x18, 0x02, 0x3a, 0x80, 0x18, 0x71, 0x38, 0x2d, 0x40, 0x58, 0x2c,
+	0x45, 0x00, 0xa0, 0x5a, 0x00, 0x00, 0x00, 0x1e, 0x00, 0x00, 0x00, 0x00,
+	0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+	0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+	0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+	0x00, 0x00, 0x00, 0x06
+};
+
+static void vidi_fake_vblank_handler(struct work_struct *work);
+
+static bool vidi_display_is_connected(struct device *dev)
+{
+	struct vidi_context *ctx = get_vidi_context(dev);
+
+	DRM_DEBUG_KMS("%s\n", __FILE__);
+
+	/*
+	 * connection request would come from user side
+	 * to do hotplug through specific ioctl.
+	 */
+	return ctx->connected ? true : false;
+}
+
+static int vidi_get_edid(struct device *dev, struct drm_connector *connector,
+				u8 *edid, int len)
+{
+	struct vidi_context *ctx = get_vidi_context(dev);
+	struct edid *raw_edid;
+
+	DRM_DEBUG_KMS("%s\n", __FILE__);
+
+	/*
+	 * the edid data comes from user side and it would be set
+	 * to ctx->raw_edid through specific ioctl.
+	 */
+	if (!ctx->raw_edid) {
+		DRM_DEBUG_KMS("raw_edid is null.\n");
+		return -EFAULT;
+	}
+
+	raw_edid = kzalloc(len, GFP_KERNEL);
+	if (!raw_edid) {
+		DRM_DEBUG_KMS("failed to allocate raw_edid.\n");
+		return -ENOMEM;
+	}
+
+	memcpy(raw_edid, ctx->raw_edid, min((1 + ctx->raw_edid->extensions)
+						* EDID_LENGTH, len));
+
+	/* attach the edid data to connector. */
+	connector->display_info.raw_edid = (char *)raw_edid;
+
+	memcpy(edid, ctx->raw_edid, min((1 + ctx->raw_edid->extensions)
+					* EDID_LENGTH, len));
+
+	return 0;
+}
+
+static void *vidi_get_panel(struct device *dev)
+{
+	DRM_DEBUG_KMS("%s\n", __FILE__);
+
+	/* TODO. */
+
+	return NULL;
+}
+
+static int vidi_check_timing(struct device *dev, void *timing)
+{
+	DRM_DEBUG_KMS("%s\n", __FILE__);
+
+	/* TODO. */
+
+	return 0;
+}
+
+static int vidi_display_power_on(struct device *dev, int mode)
+{
+	DRM_DEBUG_KMS("%s\n", __FILE__);
+
+	/* TODO */
+
+	return 0;
+}
+
+static struct exynos_drm_display_ops vidi_display_ops = {
+	.type = EXYNOS_DISPLAY_TYPE_VIDI,
+	.is_connected = vidi_display_is_connected,
+	.get_edid = vidi_get_edid,
+	.get_panel = vidi_get_panel,
+	.check_timing = vidi_check_timing,
+	.power_on = vidi_display_power_on,
+};
+
+static void vidi_dpms(struct device *subdrv_dev, int mode)
+{
+	struct vidi_context *ctx = get_vidi_context(subdrv_dev);
+
+	DRM_DEBUG_KMS("%s, %d\n", __FILE__, mode);
+
+	mutex_lock(&ctx->lock);
+
+	switch (mode) {
+	case DRM_MODE_DPMS_ON:
+		/* TODO. */
+		break;
+	case DRM_MODE_DPMS_STANDBY:
+	case DRM_MODE_DPMS_SUSPEND:
+	case DRM_MODE_DPMS_OFF:
+		/* TODO. */
+		break;
+	default:
+		DRM_DEBUG_KMS("unspecified mode %d\n", mode);
+		break;
+	}
+
+	mutex_unlock(&ctx->lock);
+}
+
+static void vidi_apply(struct device *subdrv_dev)
+{
+	struct vidi_context *ctx = get_vidi_context(subdrv_dev);
+	struct exynos_drm_manager *mgr = &ctx->subdrv.manager;
+	struct exynos_drm_manager_ops *mgr_ops = mgr->ops;
+	struct exynos_drm_overlay_ops *ovl_ops = mgr->overlay_ops;
+	struct vidi_win_data *win_data;
+	int i;
+
+	DRM_DEBUG_KMS("%s\n", __FILE__);
+
+	for (i = 0; i < WINDOWS_NR; i++) {
+		win_data = &ctx->win_data[i];
+		if (win_data->enabled && (ovl_ops && ovl_ops->commit))
+			ovl_ops->commit(subdrv_dev, i);
+	}
+
+	if (mgr_ops && mgr_ops->commit)
+		mgr_ops->commit(subdrv_dev);
+}
+
+static void vidi_commit(struct device *dev)
+{
+	struct vidi_context *ctx = get_vidi_context(dev);
+
+	DRM_DEBUG_KMS("%s\n", __FILE__);
+
+	if (ctx->suspended)
+		return;
+}
+
+static int vidi_enable_vblank(struct device *dev)
+{
+	struct vidi_context *ctx = get_vidi_context(dev);
+
+	DRM_DEBUG_KMS("%s\n", __FILE__);
+
+	if (ctx->suspended)
+		return -EPERM;
+
+	if (!test_and_set_bit(0, &ctx->irq_flags))
+		ctx->vblank_on = true;
+
+	return 0;
+}
+
+static void vidi_disable_vblank(struct device *dev)
+{
+	struct vidi_context *ctx = get_vidi_context(dev);
+
+	DRM_DEBUG_KMS("%s\n", __FILE__);
+
+	if (ctx->suspended)
+		return;
+
+	if (test_and_clear_bit(0, &ctx->irq_flags))
+		ctx->vblank_on = false;
+}
+
+static struct exynos_drm_manager_ops vidi_manager_ops = {
+	.dpms = vidi_dpms,
+	.apply = vidi_apply,
+	.commit = vidi_commit,
+	.enable_vblank = vidi_enable_vblank,
+	.disable_vblank = vidi_disable_vblank,
+};
+
+static void vidi_win_mode_set(struct device *dev,
+			      struct exynos_drm_overlay *overlay)
+{
+	struct vidi_context *ctx = get_vidi_context(dev);
+	struct vidi_win_data *win_data;
+	int win;
+	unsigned long offset;
+
+	DRM_DEBUG_KMS("%s\n", __FILE__);
+
+	if (!overlay) {
+		dev_err(dev, "overlay is NULL\n");
+		return;
+	}
+
+	win = overlay->zpos;
+	if (win == DEFAULT_ZPOS)
+		win = ctx->default_win;
+
+	if (win < 0 || win > WINDOWS_NR)
+		return;
+
+	offset = overlay->fb_x * (overlay->bpp >> 3);
+	offset += overlay->fb_y * overlay->pitch;
+
+	DRM_DEBUG_KMS("offset = 0x%lx, pitch = %x\n", offset, overlay->pitch);
+
+	win_data = &ctx->win_data[win];
+
+	win_data->offset_x = overlay->crtc_x;
+	win_data->offset_y = overlay->crtc_y;
+	win_data->ovl_width = overlay->crtc_width;
+	win_data->ovl_height = overlay->crtc_height;
+	win_data->fb_width = overlay->fb_width;
+	win_data->fb_height = overlay->fb_height;
+	win_data->dma_addr = overlay->dma_addr[0] + offset;
+	win_data->vaddr = overlay->vaddr[0] + offset;
+	win_data->bpp = overlay->bpp;
+	win_data->buf_offsize = (overlay->fb_width - overlay->crtc_width) *
+				(overlay->bpp >> 3);
+	win_data->line_size = overlay->crtc_width * (overlay->bpp >> 3);
+
+	/*
+	 * some parts of win_data should be transferred to user side
+	 * through specific ioctl.
+	 */
+
+	DRM_DEBUG_KMS("offset_x = %d, offset_y = %d\n",
+			win_data->offset_x, win_data->offset_y);
+	DRM_DEBUG_KMS("ovl_width = %d, ovl_height = %d\n",
+			win_data->ovl_width, win_data->ovl_height);
+	DRM_DEBUG_KMS("paddr = 0x%lx, vaddr = 0x%lx\n",
+			(unsigned long)win_data->dma_addr,
+			(unsigned long)win_data->vaddr);
+	DRM_DEBUG_KMS("fb_width = %d, crtc_width = %d\n",
+			overlay->fb_width, overlay->crtc_width);
+}
+
+static void vidi_win_commit(struct device *dev, int zpos)
+{
+	struct vidi_context *ctx = get_vidi_context(dev);
+	struct vidi_win_data *win_data;
+	int win = zpos;
+
+	DRM_DEBUG_KMS("%s\n", __FILE__);
+
+	if (ctx->suspended)
+		return;
+
+	if (win == DEFAULT_ZPOS)
+		win = ctx->default_win;
+
+	if (win < 0 || win > WINDOWS_NR)
+		return;
+
+	win_data = &ctx->win_data[win];
+
+	win_data->enabled = true;
+
+	DRM_DEBUG_KMS("dma_addr = 0x%x\n", win_data->dma_addr);
+
+	if (ctx->vblank_on)
+		schedule_work(&ctx->work);
+}
+
+static void vidi_win_disable(struct device *dev, int zpos)
+{
+	struct vidi_context *ctx = get_vidi_context(dev);
+	struct vidi_win_data *win_data;
+	int win = zpos;
+
+	DRM_DEBUG_KMS("%s\n", __FILE__);
+
+	if (win == DEFAULT_ZPOS)
+		win = ctx->default_win;
+
+	if (win < 0 || win > WINDOWS_NR)
+		return;
+
+	win_data = &ctx->win_data[win];
+	win_data->enabled = false;
+
+	/* TODO. */
+}
+
+static struct exynos_drm_overlay_ops vidi_overlay_ops = {
+	.mode_set = vidi_win_mode_set,
+	.commit = vidi_win_commit,
+	.disable = vidi_win_disable,
+};
+
+static void vidi_finish_pageflip(struct drm_device *drm_dev, int crtc)
+{
+	struct exynos_drm_private *dev_priv = drm_dev->dev_private;
+	struct drm_pending_vblank_event *e, *t;
+	struct timeval now;
+	unsigned long flags;
+	bool is_checked = false;
+
+	spin_lock_irqsave(&drm_dev->event_lock, flags);
+
+	list_for_each_entry_safe(e, t, &dev_priv->pageflip_event_list,
+			base.link) {
+		/* if event's pipe isn't same as crtc then ignore it. */
+		if (crtc != e->pipe)
+			continue;
+
+		is_checked = true;
+
+		do_gettimeofday(&now);
+		e->event.sequence = 0;
+		e->event.tv_sec = now.tv_sec;
+		e->event.tv_usec = now.tv_usec;
+
+		list_move_tail(&e->base.link, &e->base.file_priv->event_list);
+		wake_up_interruptible(&e->base.file_priv->event_wait);
+	}
+
+	if (is_checked) {
+		/*
+		 * call drm_vblank_put only in case that drm_vblank_get was
+		 * called.
+		 */
+		if (atomic_read(&drm_dev->vblank_refcount[crtc]) > 0)
+			drm_vblank_put(drm_dev, crtc);
+
+		/*
+		 * don't off vblank if vblank_disable_allowed is 1,
+		 * because vblank would be off by timer handler.
+		 */
+		if (!drm_dev->vblank_disable_allowed)
+			drm_vblank_off(drm_dev, crtc);
+	}
+
+	spin_unlock_irqrestore(&drm_dev->event_lock, flags);
+}
+
+static void vidi_fake_vblank_handler(struct work_struct *work)
+{
+	struct vidi_context *ctx = container_of(work, struct vidi_context,
+					work);
+	struct exynos_drm_subdrv *subdrv = &ctx->subdrv;
+	struct exynos_drm_manager *manager = &subdrv->manager;
+
+	if (manager->pipe < 0)
+		return;
+
+	/* refresh rate is about 50Hz. */
+	usleep_range(16000, 20000);
+
+	drm_handle_vblank(subdrv->drm_dev, manager->pipe);
+	vidi_finish_pageflip(subdrv->drm_dev, manager->pipe);
+}
+
+static int vidi_subdrv_probe(struct drm_device *drm_dev, struct device *dev)
+{
+	DRM_DEBUG_KMS("%s\n", __FILE__);
+
+	/*
+	 * enable drm irq mode.
+	 * - with irq_enabled = 1, we can use the vblank feature.
+	 *
+	 * P.S. note that we wouldn't use drm irq handler but
+	 *	just specific driver own one instead because
+	 *	drm framework supports only one irq handler.
+	 */
+	drm_dev->irq_enabled = 1;
+
+	/*
+	 * with vblank_disable_allowed = 1, vblank interrupt will be disabled
+	 * by drm timer once a current process gives up ownership of
+	 * vblank event.(after drm_vblank_put function is called)
+	 */
+	drm_dev->vblank_disable_allowed = 1;
+
+	return 0;
+}
+
+static void vidi_subdrv_remove(struct drm_device *drm_dev)
+{
+	DRM_DEBUG_KMS("%s\n", __FILE__);
+
+	/* TODO. */
+}
+
+static int vidi_power_on(struct vidi_context *ctx, bool enable)
+{
+	struct exynos_drm_subdrv *subdrv = &ctx->subdrv;
+	struct device *dev = subdrv->manager.dev;
+
+	DRM_DEBUG_KMS("%s\n", __FILE__);
+
+	if (enable != false && enable != true)
+		return -EINVAL;
+
+	if (enable) {
+		ctx->suspended = false;
+
+		/* if vblank was enabled status, enable it again. */
+		if (test_and_clear_bit(0, &ctx->irq_flags))
+			vidi_enable_vblank(dev);
+
+		vidi_apply(dev);
+	} else {
+		ctx->suspended = true;
+	}
+
+	return 0;
+}
+
+static int vidi_show_connection(struct device *dev,
+				struct device_attribute *attr, char *buf)
+{
+	int rc;
+	struct vidi_context *ctx = get_vidi_context(dev);
+
+	mutex_lock(&ctx->lock);
+
+	rc = sprintf(buf, "%d\n", ctx->connected);
+
+	mutex_unlock(&ctx->lock);
+
+	return rc;
+}
+
+static int vidi_store_connection(struct device *dev,
+				struct device_attribute *attr,
+				const char *buf, size_t len)
+{
+	struct vidi_context *ctx = get_vidi_context(dev);
+	int ret;
+
+	DRM_DEBUG_KMS("%s\n", __FILE__);
+
+	ret = kstrtoint(buf, 0, &ctx->connected);
+	if (ret)
+		return ret;
+
+	if (ctx->connected > 1)
+		return -EINVAL;
+
+	DRM_DEBUG_KMS("requested connection.\n");
+
+	drm_helper_hpd_irq_event(ctx->subdrv.drm_dev);
+
+	return len;
+}
+
+static DEVICE_ATTR(connection, 0644, vidi_show_connection,
+			vidi_store_connection);
+
+int vidi_connection_ioctl(struct drm_device *drm_dev, void *data,
+				struct drm_file *file_priv)
+{
+	struct vidi_context *ctx = NULL;
+	struct drm_encoder *encoder;
+	struct exynos_drm_manager *manager;
+	struct exynos_drm_display_ops *display_ops;
+	struct drm_exynos_vidi_connection *vidi = data;
+
+	DRM_DEBUG_KMS("%s\n", __FILE__);
+
+	if (!vidi) {
+		DRM_DEBUG_KMS("user data for vidi is null.\n");
+		return -EINVAL;
+	}
+
+	if (!vidi->edid) {
+		DRM_DEBUG_KMS("edid data is null.\n");
+		return -EINVAL;
+	}
+
+	if (vidi->connection > 1) {
+		DRM_DEBUG_KMS("connection should be 0 or 1.\n");
+		return -EINVAL;
+	}
+
+	list_for_each_entry(encoder, &drm_dev->mode_config.encoder_list,
+								head) {
+		manager = exynos_drm_get_manager(encoder);
+		display_ops = manager->display_ops;
+
+		if (display_ops->type == EXYNOS_DISPLAY_TYPE_VIDI) {
+			ctx = get_vidi_context(manager->dev);
+			break;
+		}
+	}
+
+	if (!ctx) {
+		DRM_DEBUG_KMS("not found virtual device type encoder.\n");
+		return -EINVAL;
+	}
+
+	if (ctx->connected == vidi->connection) {
+		DRM_DEBUG_KMS("same connection request.\n");
+		return -EINVAL;
+	}
+
+	if (vidi->connection)
+		ctx->raw_edid = (struct edid *)vidi->edid;
+
+	ctx->connected = vidi->connection;
+	drm_helper_hpd_irq_event(ctx->subdrv.drm_dev);
+
+	return 0;
+}
+
+static int __devinit vidi_probe(struct platform_device *pdev)
+{
+	struct device *dev = &pdev->dev;
+	struct vidi_context *ctx;
+	struct exynos_drm_subdrv *subdrv;
+	int ret;
+
+	DRM_DEBUG_KMS("%s\n", __FILE__);
+
+	ctx = kzalloc(sizeof(*ctx), GFP_KERNEL);
+	if (!ctx)
+		return -ENOMEM;
+
+	ctx->default_win = 0;
+
+	INIT_WORK(&ctx->work, vidi_fake_vblank_handler);
+
+	/* for test */
+	ctx->raw_edid = (struct edid *)fake_edid_info;
+
+	subdrv = &ctx->subdrv;
+	subdrv->probe = vidi_subdrv_probe;
+	subdrv->remove = vidi_subdrv_remove;
+	subdrv->manager.pipe = -1;
+	subdrv->manager.ops = &vidi_manager_ops;
+	subdrv->manager.overlay_ops = &vidi_overlay_ops;
+	subdrv->manager.display_ops = &vidi_display_ops;
+	subdrv->manager.dev = dev;
+
+	mutex_init(&ctx->lock);
+
+	platform_set_drvdata(pdev, ctx);
+
+	ret = device_create_file(&pdev->dev, &dev_attr_connection);
+	if (ret < 0)
+		DRM_INFO("failed to create connection sysfs.\n");
+
+	exynos_drm_subdrv_register(subdrv);
+
+	return 0;
+}
+
+static int __devexit vidi_remove(struct platform_device *pdev)
+{
+	struct vidi_context *ctx = platform_get_drvdata(pdev);
+
+	DRM_DEBUG_KMS("%s\n", __FILE__);
+
+	exynos_drm_subdrv_unregister(&ctx->subdrv);
+
+	kfree(ctx);
+
+	return 0;
+}
+
+#ifdef CONFIG_PM_SLEEP
+static int vidi_suspend(struct device *dev)
+{
+	struct vidi_context *ctx = get_vidi_context(dev);
+
+	return vidi_power_on(ctx, false);
+}
+
+static int vidi_resume(struct device *dev)
+{
+	struct vidi_context *ctx = get_vidi_context(dev);
+
+	return vidi_power_on(ctx, true);
+}
+#endif
+
+static const struct dev_pm_ops vidi_pm_ops = {
+	SET_SYSTEM_SLEEP_PM_OPS(vidi_suspend, vidi_resume)
+};
+
+struct platform_driver vidi_driver = {
+	.probe		= vidi_probe,
+	.remove		= __devexit_p(vidi_remove),
+	.driver		= {
+		.name	= "exynos-drm-vidi",
+		.owner	= THIS_MODULE,
+		.pm	= &vidi_pm_ops,
+	},
+};
diff --git a/drivers/gpu/drm/exynos/exynos_drm_vidi.h b/drivers/gpu/drm/exynos/exynos_drm_vidi.h
new file mode 100644
index 000000000000..a4babe4e65d7
--- /dev/null
+++ b/drivers/gpu/drm/exynos/exynos_drm_vidi.h
@@ -0,0 +1,36 @@
+/* exynos_drm_vidi.h
+ *
+ * Copyright (c) 2012 Samsung Electronics Co., Ltd.
+ * Author: Inki Dae <inki.dae@samsung.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
+ * 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.
+ */
+
+#ifndef _EXYNOS_DRM_VIDI_H_
+#define _EXYNOS_DRM_VIDI_H_
+
+#ifdef CONFIG_DRM_EXYNOS_VIDI
+int vidi_connection_ioctl(struct drm_device *drm_dev, void *data,
+				struct drm_file *file_priv);
+#else
+#define vidi_connection_ioctl	NULL
+#endif
+
+#endif
diff --git a/drivers/gpu/drm/exynos/exynos_hdmi.c b/drivers/gpu/drm/exynos/exynos_hdmi.c
index 3429d3fd93f3..575a8cbd3533 100644
--- a/drivers/gpu/drm/exynos/exynos_hdmi.c
+++ b/drivers/gpu/drm/exynos/exynos_hdmi.c
@@ -41,44 +41,83 @@
 #include "exynos_hdmi.h"
 
 #define HDMI_OVERLAY_NUMBER	3
+#define MAX_WIDTH		1920
+#define MAX_HEIGHT		1080
 #define get_hdmi_context(dev)	platform_get_drvdata(to_platform_device(dev))
 
-static const u8 hdmiphy_conf27[32] = {
+struct hdmi_resources {
+	struct clk			*hdmi;
+	struct clk			*sclk_hdmi;
+	struct clk			*sclk_pixel;
+	struct clk			*sclk_hdmiphy;
+	struct clk			*hdmiphy;
+	struct regulator_bulk_data	*regul_bulk;
+	int				regul_count;
+};
+
+struct hdmi_context {
+	struct device			*dev;
+	struct drm_device		*drm_dev;
+	struct fb_videomode		*default_timing;
+	unsigned int			is_v13:1;
+	unsigned int			default_win;
+	unsigned int			default_bpp;
+	bool				hpd_handle;
+	bool				enabled;
+
+	struct resource			*regs_res;
+	void __iomem			*regs;
+	unsigned int			irq;
+	struct workqueue_struct		*wq;
+	struct work_struct		hotplug_work;
+
+	struct i2c_client		*ddc_port;
+	struct i2c_client		*hdmiphy_port;
+
+	/* current hdmiphy conf index */
+	int cur_conf;
+
+	struct hdmi_resources		res;
+	void				*parent_ctx;
+};
+
+/* 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_conf27_027[32] = {
+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_conf74_175[32] = {
+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_conf74_25[32] = {
+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_conf148_5[32] = {
+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_tg_regs {
+struct hdmi_v13_tg_regs {
 	u8 cmd;
 	u8 h_fsz_l;
 	u8 h_fsz_h;
@@ -110,7 +149,7 @@ struct hdmi_tg_regs {
 	u8 field_bot_hdmi_h;
 };
 
-struct hdmi_core_regs {
+struct hdmi_v13_core_regs {
 	u8 h_blank[2];
 	u8 v_blank[3];
 	u8 h_v_line[3];
@@ -123,12 +162,21 @@ struct hdmi_core_regs {
 	u8 v_sync_gen3[3];
 };
 
-struct hdmi_preset_conf {
-	struct hdmi_core_regs core;
-	struct hdmi_tg_regs tg;
+struct hdmi_v13_preset_conf {
+	struct hdmi_v13_core_regs core;
+	struct hdmi_v13_tg_regs tg;
 };
 
-static const struct hdmi_preset_conf hdmi_conf_480p = {
+struct hdmi_v13_conf {
+	int width;
+	int height;
+	int vrefresh;
+	bool interlace;
+	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},
@@ -154,7 +202,7 @@ static const struct hdmi_preset_conf hdmi_conf_480p = {
 	},
 };
 
-static const struct hdmi_preset_conf hdmi_conf_720p60 = {
+static const struct hdmi_v13_preset_conf hdmi_v13_conf_720p60 = {
 	.core = {
 		.h_blank = {0x72, 0x01},
 		.v_blank = {0xee, 0xf2, 0x00},
@@ -182,7 +230,7 @@ static const struct hdmi_preset_conf hdmi_conf_720p60 = {
 	},
 };
 
-static const struct hdmi_preset_conf hdmi_conf_1080i50 = {
+static const struct hdmi_v13_preset_conf hdmi_v13_conf_1080i50 = {
 	.core = {
 		.h_blank = {0xd0, 0x02},
 		.v_blank = {0x32, 0xB2, 0x00},
@@ -210,7 +258,7 @@ static const struct hdmi_preset_conf hdmi_conf_1080i50 = {
 	},
 };
 
-static const struct hdmi_preset_conf hdmi_conf_1080p50 = {
+static const struct hdmi_v13_preset_conf hdmi_v13_conf_1080p50 = {
 	.core = {
 		.h_blank = {0xd0, 0x02},
 		.v_blank = {0x65, 0x6c, 0x01},
@@ -238,7 +286,7 @@ static const struct hdmi_preset_conf hdmi_conf_1080p50 = {
 	},
 };
 
-static const struct hdmi_preset_conf hdmi_conf_1080i60 = {
+static const struct hdmi_v13_preset_conf hdmi_v13_conf_1080i60 = {
 	.core = {
 		.h_blank = {0x18, 0x01},
 		.v_blank = {0x32, 0xB2, 0x00},
@@ -266,7 +314,7 @@ static const struct hdmi_preset_conf hdmi_conf_1080i60 = {
 	},
 };
 
-static const struct hdmi_preset_conf hdmi_conf_1080p60 = {
+static const struct hdmi_v13_preset_conf hdmi_v13_conf_1080p60 = {
 	.core = {
 		.h_blank = {0x18, 0x01},
 		.v_blank = {0x65, 0x6c, 0x01},
@@ -294,13 +342,530 @@ static const struct hdmi_preset_conf hdmi_conf_1080p60 = {
 	},
 };
 
+static const struct hdmi_v13_conf hdmi_v13_confs[] = {
+	{ 1280, 720, 60, false, hdmiphy_v13_conf74_25, &hdmi_v13_conf_720p60 },
+	{ 1280, 720, 50, false, hdmiphy_v13_conf74_25, &hdmi_v13_conf_720p60 },
+	{ 720, 480, 60, false, hdmiphy_v13_conf27_027, &hdmi_v13_conf_480p },
+	{ 1920, 1080, 50, true, hdmiphy_v13_conf74_25, &hdmi_v13_conf_1080i50 },
+	{ 1920, 1080, 50, false, hdmiphy_v13_conf148_5,
+				 &hdmi_v13_conf_1080p50 },
+	{ 1920, 1080, 60, true, hdmiphy_v13_conf74_25, &hdmi_v13_conf_1080i60 },
+	{ 1920, 1080, 60, false, hdmiphy_v13_conf148_5,
+				 &hdmi_v13_conf_1080p60 },
+};
+
+/* HDMI Version 1.4 */
+static const u8 hdmiphy_conf27_027[32] = {
+	0x01, 0xd1, 0x2d, 0x72, 0x40, 0x64, 0x12, 0x08,
+	0x43, 0xa0, 0x0e, 0xd9, 0x45, 0xa0, 0xac, 0x80,
+	0x08, 0x80, 0x11, 0x04, 0x02, 0x22, 0x44, 0x86,
+	0x54, 0xe3, 0x24, 0x00, 0x00, 0x00, 0x01, 0x00,
+};
+
+static const u8 hdmiphy_conf74_25[32] = {
+	0x01, 0xd1, 0x1f, 0x10, 0x40, 0x40, 0xf8, 0x08,
+	0x81, 0xa0, 0xba, 0xd8, 0x45, 0xa0, 0xac, 0x80,
+	0x3c, 0x80, 0x11, 0x04, 0x02, 0x22, 0x44, 0x86,
+	0x54, 0xa5, 0x24, 0x01, 0x00, 0x00, 0x01, 0x00,
+};
+
+static const u8 hdmiphy_conf148_5[32] = {
+	0x01, 0xd1, 0x1f, 0x00, 0x40, 0x40, 0xf8, 0x08,
+	0x81, 0xa0, 0xba, 0xd8, 0x45, 0xa0, 0xac, 0x80,
+	0x3c, 0x80, 0x11, 0x04, 0x02, 0x22, 0x44, 0x86,
+	0x54, 0x4b, 0x25, 0x03, 0x00, 0x00, 0x01, 0x00,
+};
+
+struct hdmi_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 vact_st3_l;
+	u8 vact_st3_h;
+	u8 vact_st4_l;
+	u8 vact_st4_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;
+	u8 tg_3d;
+};
+
+struct hdmi_core_regs {
+	u8 h_blank[2];
+	u8 v2_blank[2];
+	u8 v1_blank[2];
+	u8 v_line[2];
+	u8 h_line[2];
+	u8 hsync_pol[1];
+	u8 vsync_pol[1];
+	u8 int_pro_mode[1];
+	u8 v_blank_f0[2];
+	u8 v_blank_f1[2];
+	u8 h_sync_start[2];
+	u8 h_sync_end[2];
+	u8 v_sync_line_bef_2[2];
+	u8 v_sync_line_bef_1[2];
+	u8 v_sync_line_aft_2[2];
+	u8 v_sync_line_aft_1[2];
+	u8 v_sync_line_aft_pxl_2[2];
+	u8 v_sync_line_aft_pxl_1[2];
+	u8 v_blank_f2[2]; /* for 3D mode */
+	u8 v_blank_f3[2]; /* for 3D mode */
+	u8 v_blank_f4[2]; /* for 3D mode */
+	u8 v_blank_f5[2]; /* for 3D mode */
+	u8 v_sync_line_aft_3[2];
+	u8 v_sync_line_aft_4[2];
+	u8 v_sync_line_aft_5[2];
+	u8 v_sync_line_aft_6[2];
+	u8 v_sync_line_aft_pxl_3[2];
+	u8 v_sync_line_aft_pxl_4[2];
+	u8 v_sync_line_aft_pxl_5[2];
+	u8 v_sync_line_aft_pxl_6[2];
+	u8 vact_space_1[2];
+	u8 vact_space_2[2];
+	u8 vact_space_3[2];
+	u8 vact_space_4[2];
+	u8 vact_space_5[2];
+	u8 vact_space_6[2];
+};
+
+struct hdmi_preset_conf {
+	struct hdmi_core_regs core;
+	struct hdmi_tg_regs tg;
+};
+
+struct hdmi_conf {
+	int width;
+	int height;
+	int vrefresh;
+	bool interlace;
+	const u8 *hdmiphy_data;
+	const struct hdmi_preset_conf *conf;
+};
+
+static const struct hdmi_preset_conf hdmi_conf_480p60 = {
+	.core = {
+		.h_blank = {0x8a, 0x00},
+		.v2_blank = {0x0d, 0x02},
+		.v1_blank = {0x2d, 0x00},
+		.v_line = {0x0d, 0x02},
+		.h_line = {0x5a, 0x03},
+		.hsync_pol = {0x01},
+		.vsync_pol = {0x01},
+		.int_pro_mode = {0x00},
+		.v_blank_f0 = {0xff, 0xff},
+		.v_blank_f1 = {0xff, 0xff},
+		.h_sync_start = {0x0e, 0x00},
+		.h_sync_end = {0x4c, 0x00},
+		.v_sync_line_bef_2 = {0x0f, 0x00},
+		.v_sync_line_bef_1 = {0x09, 0x00},
+		.v_sync_line_aft_2 = {0xff, 0xff},
+		.v_sync_line_aft_1 = {0xff, 0xff},
+		.v_sync_line_aft_pxl_2 = {0xff, 0xff},
+		.v_sync_line_aft_pxl_1 = {0xff, 0xff},
+		.v_blank_f2 = {0xff, 0xff},
+		.v_blank_f3 = {0xff, 0xff},
+		.v_blank_f4 = {0xff, 0xff},
+		.v_blank_f5 = {0xff, 0xff},
+		.v_sync_line_aft_3 = {0xff, 0xff},
+		.v_sync_line_aft_4 = {0xff, 0xff},
+		.v_sync_line_aft_5 = {0xff, 0xff},
+		.v_sync_line_aft_6 = {0xff, 0xff},
+		.v_sync_line_aft_pxl_3 = {0xff, 0xff},
+		.v_sync_line_aft_pxl_4 = {0xff, 0xff},
+		.v_sync_line_aft_pxl_5 = {0xff, 0xff},
+		.v_sync_line_aft_pxl_6 = {0xff, 0xff},
+		.vact_space_1 = {0xff, 0xff},
+		.vact_space_2 = {0xff, 0xff},
+		.vact_space_3 = {0xff, 0xff},
+		.vact_space_4 = {0xff, 0xff},
+		.vact_space_5 = {0xff, 0xff},
+		.vact_space_6 = {0xff, 0xff},
+		/* 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 */
+		0x48, 0x02, /* vact_st2 */
+		0x00, 0x00, /* vact_st3 */
+		0x00, 0x00, /* vact_st4 */
+		0x01, 0x00, 0x01, 0x00, /* vsync top/bot */
+		0x01, 0x00, 0x33, 0x02, /* field top/bot */
+		0x00, /* 3d FP */
+	},
+};
+
+static const struct hdmi_preset_conf hdmi_conf_720p50 = {
+	.core = {
+		.h_blank = {0xbc, 0x02},
+		.v2_blank = {0xee, 0x02},
+		.v1_blank = {0x1e, 0x00},
+		.v_line = {0xee, 0x02},
+		.h_line = {0xbc, 0x07},
+		.hsync_pol = {0x00},
+		.vsync_pol = {0x00},
+		.int_pro_mode = {0x00},
+		.v_blank_f0 = {0xff, 0xff},
+		.v_blank_f1 = {0xff, 0xff},
+		.h_sync_start = {0xb6, 0x01},
+		.h_sync_end = {0xde, 0x01},
+		.v_sync_line_bef_2 = {0x0a, 0x00},
+		.v_sync_line_bef_1 = {0x05, 0x00},
+		.v_sync_line_aft_2 = {0xff, 0xff},
+		.v_sync_line_aft_1 = {0xff, 0xff},
+		.v_sync_line_aft_pxl_2 = {0xff, 0xff},
+		.v_sync_line_aft_pxl_1 = {0xff, 0xff},
+		.v_blank_f2 = {0xff, 0xff},
+		.v_blank_f3 = {0xff, 0xff},
+		.v_blank_f4 = {0xff, 0xff},
+		.v_blank_f5 = {0xff, 0xff},
+		.v_sync_line_aft_3 = {0xff, 0xff},
+		.v_sync_line_aft_4 = {0xff, 0xff},
+		.v_sync_line_aft_5 = {0xff, 0xff},
+		.v_sync_line_aft_6 = {0xff, 0xff},
+		.v_sync_line_aft_pxl_3 = {0xff, 0xff},
+		.v_sync_line_aft_pxl_4 = {0xff, 0xff},
+		.v_sync_line_aft_pxl_5 = {0xff, 0xff},
+		.v_sync_line_aft_pxl_6 = {0xff, 0xff},
+		.vact_space_1 = {0xff, 0xff},
+		.vact_space_2 = {0xff, 0xff},
+		.vact_space_3 = {0xff, 0xff},
+		.vact_space_4 = {0xff, 0xff},
+		.vact_space_5 = {0xff, 0xff},
+		.vact_space_6 = {0xff, 0xff},
+		/* other don't care */
+	},
+	.tg = {
+		0x00, /* cmd */
+		0xbc, 0x07, /* h_fsz */
+		0xbc, 0x02, 0x00, 0x05, /* hact */
+		0xee, 0x02, /* v_fsz */
+		0x01, 0x00, 0x33, 0x02, /* vsync */
+		0x1e, 0x00, 0xd0, 0x02, /* vact */
+		0x33, 0x02, /* field_chg */
+		0x48, 0x02, /* vact_st2 */
+		0x00, 0x00, /* vact_st3 */
+		0x00, 0x00, /* vact_st4 */
+		0x01, 0x00, 0x01, 0x00, /* vsync top/bot */
+		0x01, 0x00, 0x33, 0x02, /* field top/bot */
+		0x00, /* 3d FP */
+	},
+};
+
+static const struct hdmi_preset_conf hdmi_conf_720p60 = {
+	.core = {
+		.h_blank = {0x72, 0x01},
+		.v2_blank = {0xee, 0x02},
+		.v1_blank = {0x1e, 0x00},
+		.v_line = {0xee, 0x02},
+		.h_line = {0x72, 0x06},
+		.hsync_pol = {0x00},
+		.vsync_pol = {0x00},
+		.int_pro_mode = {0x00},
+		.v_blank_f0 = {0xff, 0xff},
+		.v_blank_f1 = {0xff, 0xff},
+		.h_sync_start = {0x6c, 0x00},
+		.h_sync_end = {0x94, 0x00},
+		.v_sync_line_bef_2 = {0x0a, 0x00},
+		.v_sync_line_bef_1 = {0x05, 0x00},
+		.v_sync_line_aft_2 = {0xff, 0xff},
+		.v_sync_line_aft_1 = {0xff, 0xff},
+		.v_sync_line_aft_pxl_2 = {0xff, 0xff},
+		.v_sync_line_aft_pxl_1 = {0xff, 0xff},
+		.v_blank_f2 = {0xff, 0xff},
+		.v_blank_f3 = {0xff, 0xff},
+		.v_blank_f4 = {0xff, 0xff},
+		.v_blank_f5 = {0xff, 0xff},
+		.v_sync_line_aft_3 = {0xff, 0xff},
+		.v_sync_line_aft_4 = {0xff, 0xff},
+		.v_sync_line_aft_5 = {0xff, 0xff},
+		.v_sync_line_aft_6 = {0xff, 0xff},
+		.v_sync_line_aft_pxl_3 = {0xff, 0xff},
+		.v_sync_line_aft_pxl_4 = {0xff, 0xff},
+		.v_sync_line_aft_pxl_5 = {0xff, 0xff},
+		.v_sync_line_aft_pxl_6 = {0xff, 0xff},
+		.vact_space_1 = {0xff, 0xff},
+		.vact_space_2 = {0xff, 0xff},
+		.vact_space_3 = {0xff, 0xff},
+		.vact_space_4 = {0xff, 0xff},
+		.vact_space_5 = {0xff, 0xff},
+		.vact_space_6 = {0xff, 0xff},
+		/* other don't care */
+	},
+	.tg = {
+		0x00, /* cmd */
+		0x72, 0x06, /* h_fsz */
+		0x72, 0x01, 0x00, 0x05, /* hact */
+		0xee, 0x02, /* v_fsz */
+		0x01, 0x00, 0x33, 0x02, /* vsync */
+		0x1e, 0x00, 0xd0, 0x02, /* vact */
+		0x33, 0x02, /* field_chg */
+		0x48, 0x02, /* vact_st2 */
+		0x00, 0x00, /* vact_st3 */
+		0x00, 0x00, /* vact_st4 */
+		0x01, 0x00, 0x01, 0x00, /* vsync top/bot */
+		0x01, 0x00, 0x33, 0x02, /* field top/bot */
+		0x00, /* 3d FP */
+	},
+};
+
+static const struct hdmi_preset_conf hdmi_conf_1080i50 = {
+	.core = {
+		.h_blank = {0xd0, 0x02},
+		.v2_blank = {0x32, 0x02},
+		.v1_blank = {0x16, 0x00},
+		.v_line = {0x65, 0x04},
+		.h_line = {0x50, 0x0a},
+		.hsync_pol = {0x00},
+		.vsync_pol = {0x00},
+		.int_pro_mode = {0x01},
+		.v_blank_f0 = {0x49, 0x02},
+		.v_blank_f1 = {0x65, 0x04},
+		.h_sync_start = {0x0e, 0x02},
+		.h_sync_end = {0x3a, 0x02},
+		.v_sync_line_bef_2 = {0x07, 0x00},
+		.v_sync_line_bef_1 = {0x02, 0x00},
+		.v_sync_line_aft_2 = {0x39, 0x02},
+		.v_sync_line_aft_1 = {0x34, 0x02},
+		.v_sync_line_aft_pxl_2 = {0x38, 0x07},
+		.v_sync_line_aft_pxl_1 = {0x38, 0x07},
+		.v_blank_f2 = {0xff, 0xff},
+		.v_blank_f3 = {0xff, 0xff},
+		.v_blank_f4 = {0xff, 0xff},
+		.v_blank_f5 = {0xff, 0xff},
+		.v_sync_line_aft_3 = {0xff, 0xff},
+		.v_sync_line_aft_4 = {0xff, 0xff},
+		.v_sync_line_aft_5 = {0xff, 0xff},
+		.v_sync_line_aft_6 = {0xff, 0xff},
+		.v_sync_line_aft_pxl_3 = {0xff, 0xff},
+		.v_sync_line_aft_pxl_4 = {0xff, 0xff},
+		.v_sync_line_aft_pxl_5 = {0xff, 0xff},
+		.v_sync_line_aft_pxl_6 = {0xff, 0xff},
+		.vact_space_1 = {0xff, 0xff},
+		.vact_space_2 = {0xff, 0xff},
+		.vact_space_3 = {0xff, 0xff},
+		.vact_space_4 = {0xff, 0xff},
+		.vact_space_5 = {0xff, 0xff},
+		.vact_space_6 = {0xff, 0xff},
+		/* other don't care */
+	},
+	.tg = {
+		0x00, /* cmd */
+		0x50, 0x0a, /* h_fsz */
+		0xd0, 0x02, 0x80, 0x07, /* hact */
+		0x65, 0x04, /* v_fsz */
+		0x01, 0x00, 0x33, 0x02, /* vsync */
+		0x16, 0x00, 0x1c, 0x02, /* vact */
+		0x33, 0x02, /* field_chg */
+		0x49, 0x02, /* vact_st2 */
+		0x00, 0x00, /* vact_st3 */
+		0x00, 0x00, /* vact_st4 */
+		0x01, 0x00, 0x33, 0x02, /* vsync top/bot */
+		0x01, 0x00, 0x33, 0x02, /* field top/bot */
+		0x00, /* 3d FP */
+	},
+};
+
+static const struct hdmi_preset_conf hdmi_conf_1080i60 = {
+	.core = {
+		.h_blank = {0x18, 0x01},
+		.v2_blank = {0x32, 0x02},
+		.v1_blank = {0x16, 0x00},
+		.v_line = {0x65, 0x04},
+		.h_line = {0x98, 0x08},
+		.hsync_pol = {0x00},
+		.vsync_pol = {0x00},
+		.int_pro_mode = {0x01},
+		.v_blank_f0 = {0x49, 0x02},
+		.v_blank_f1 = {0x65, 0x04},
+		.h_sync_start = {0x56, 0x00},
+		.h_sync_end = {0x82, 0x00},
+		.v_sync_line_bef_2 = {0x07, 0x00},
+		.v_sync_line_bef_1 = {0x02, 0x00},
+		.v_sync_line_aft_2 = {0x39, 0x02},
+		.v_sync_line_aft_1 = {0x34, 0x02},
+		.v_sync_line_aft_pxl_2 = {0xa4, 0x04},
+		.v_sync_line_aft_pxl_1 = {0xa4, 0x04},
+		.v_blank_f2 = {0xff, 0xff},
+		.v_blank_f3 = {0xff, 0xff},
+		.v_blank_f4 = {0xff, 0xff},
+		.v_blank_f5 = {0xff, 0xff},
+		.v_sync_line_aft_3 = {0xff, 0xff},
+		.v_sync_line_aft_4 = {0xff, 0xff},
+		.v_sync_line_aft_5 = {0xff, 0xff},
+		.v_sync_line_aft_6 = {0xff, 0xff},
+		.v_sync_line_aft_pxl_3 = {0xff, 0xff},
+		.v_sync_line_aft_pxl_4 = {0xff, 0xff},
+		.v_sync_line_aft_pxl_5 = {0xff, 0xff},
+		.v_sync_line_aft_pxl_6 = {0xff, 0xff},
+		.vact_space_1 = {0xff, 0xff},
+		.vact_space_2 = {0xff, 0xff},
+		.vact_space_3 = {0xff, 0xff},
+		.vact_space_4 = {0xff, 0xff},
+		.vact_space_5 = {0xff, 0xff},
+		.vact_space_6 = {0xff, 0xff},
+		/* other don't care */
+	},
+	.tg = {
+		0x00, /* cmd */
+		0x98, 0x08, /* h_fsz */
+		0x18, 0x01, 0x80, 0x07, /* hact */
+		0x65, 0x04, /* v_fsz */
+		0x01, 0x00, 0x33, 0x02, /* vsync */
+		0x16, 0x00, 0x1c, 0x02, /* vact */
+		0x33, 0x02, /* field_chg */
+		0x49, 0x02, /* vact_st2 */
+		0x00, 0x00, /* vact_st3 */
+		0x00, 0x00, /* vact_st4 */
+		0x01, 0x00, 0x33, 0x02, /* vsync top/bot */
+		0x01, 0x00, 0x33, 0x02, /* field top/bot */
+		0x00, /* 3d FP */
+	},
+};
+
+static const struct hdmi_preset_conf hdmi_conf_1080p50 = {
+	.core = {
+		.h_blank = {0xd0, 0x02},
+		.v2_blank = {0x65, 0x04},
+		.v1_blank = {0x2d, 0x00},
+		.v_line = {0x65, 0x04},
+		.h_line = {0x50, 0x0a},
+		.hsync_pol = {0x00},
+		.vsync_pol = {0x00},
+		.int_pro_mode = {0x00},
+		.v_blank_f0 = {0xff, 0xff},
+		.v_blank_f1 = {0xff, 0xff},
+		.h_sync_start = {0x0e, 0x02},
+		.h_sync_end = {0x3a, 0x02},
+		.v_sync_line_bef_2 = {0x09, 0x00},
+		.v_sync_line_bef_1 = {0x04, 0x00},
+		.v_sync_line_aft_2 = {0xff, 0xff},
+		.v_sync_line_aft_1 = {0xff, 0xff},
+		.v_sync_line_aft_pxl_2 = {0xff, 0xff},
+		.v_sync_line_aft_pxl_1 = {0xff, 0xff},
+		.v_blank_f2 = {0xff, 0xff},
+		.v_blank_f3 = {0xff, 0xff},
+		.v_blank_f4 = {0xff, 0xff},
+		.v_blank_f5 = {0xff, 0xff},
+		.v_sync_line_aft_3 = {0xff, 0xff},
+		.v_sync_line_aft_4 = {0xff, 0xff},
+		.v_sync_line_aft_5 = {0xff, 0xff},
+		.v_sync_line_aft_6 = {0xff, 0xff},
+		.v_sync_line_aft_pxl_3 = {0xff, 0xff},
+		.v_sync_line_aft_pxl_4 = {0xff, 0xff},
+		.v_sync_line_aft_pxl_5 = {0xff, 0xff},
+		.v_sync_line_aft_pxl_6 = {0xff, 0xff},
+		.vact_space_1 = {0xff, 0xff},
+		.vact_space_2 = {0xff, 0xff},
+		.vact_space_3 = {0xff, 0xff},
+		.vact_space_4 = {0xff, 0xff},
+		.vact_space_5 = {0xff, 0xff},
+		.vact_space_6 = {0xff, 0xff},
+		/* other don't care */
+	},
+	.tg = {
+		0x00, /* cmd */
+		0x50, 0x0a, /* h_fsz */
+		0xd0, 0x02, 0x80, 0x07, /* hact */
+		0x65, 0x04, /* v_fsz */
+		0x01, 0x00, 0x33, 0x02, /* vsync */
+		0x2d, 0x00, 0x38, 0x04, /* vact */
+		0x33, 0x02, /* field_chg */
+		0x48, 0x02, /* vact_st2 */
+		0x00, 0x00, /* vact_st3 */
+		0x00, 0x00, /* vact_st4 */
+		0x01, 0x00, 0x01, 0x00, /* vsync top/bot */
+		0x01, 0x00, 0x33, 0x02, /* field top/bot */
+		0x00, /* 3d FP */
+	},
+};
+
+static const struct hdmi_preset_conf hdmi_conf_1080p60 = {
+	.core = {
+		.h_blank = {0x18, 0x01},
+		.v2_blank = {0x65, 0x04},
+		.v1_blank = {0x2d, 0x00},
+		.v_line = {0x65, 0x04},
+		.h_line = {0x98, 0x08},
+		.hsync_pol = {0x00},
+		.vsync_pol = {0x00},
+		.int_pro_mode = {0x00},
+		.v_blank_f0 = {0xff, 0xff},
+		.v_blank_f1 = {0xff, 0xff},
+		.h_sync_start = {0x56, 0x00},
+		.h_sync_end = {0x82, 0x00},
+		.v_sync_line_bef_2 = {0x09, 0x00},
+		.v_sync_line_bef_1 = {0x04, 0x00},
+		.v_sync_line_aft_2 = {0xff, 0xff},
+		.v_sync_line_aft_1 = {0xff, 0xff},
+		.v_sync_line_aft_pxl_2 = {0xff, 0xff},
+		.v_sync_line_aft_pxl_1 = {0xff, 0xff},
+		.v_blank_f2 = {0xff, 0xff},
+		.v_blank_f3 = {0xff, 0xff},
+		.v_blank_f4 = {0xff, 0xff},
+		.v_blank_f5 = {0xff, 0xff},
+		.v_sync_line_aft_3 = {0xff, 0xff},
+		.v_sync_line_aft_4 = {0xff, 0xff},
+		.v_sync_line_aft_5 = {0xff, 0xff},
+		.v_sync_line_aft_6 = {0xff, 0xff},
+		.v_sync_line_aft_pxl_3 = {0xff, 0xff},
+		.v_sync_line_aft_pxl_4 = {0xff, 0xff},
+		.v_sync_line_aft_pxl_5 = {0xff, 0xff},
+		.v_sync_line_aft_pxl_6 = {0xff, 0xff},
+		/* other don't care */
+	},
+	.tg = {
+		0x00, /* cmd */
+		0x98, 0x08, /* h_fsz */
+		0x18, 0x01, 0x80, 0x07, /* hact */
+		0x65, 0x04, /* v_fsz */
+		0x01, 0x00, 0x33, 0x02, /* vsync */
+		0x2d, 0x00, 0x38, 0x04, /* vact */
+		0x33, 0x02, /* field_chg */
+		0x48, 0x02, /* vact_st2 */
+		0x00, 0x00, /* vact_st3 */
+		0x00, 0x00, /* vact_st4 */
+		0x01, 0x00, 0x01, 0x00, /* vsync top/bot */
+		0x01, 0x00, 0x33, 0x02, /* field top/bot */
+		0x00, /* 3d FP */
+	},
+};
+
 static const struct hdmi_conf hdmi_confs[] = {
+	{ 720, 480, 60, false, hdmiphy_conf27_027, &hdmi_conf_480p60 },
+	{ 1280, 720, 50, false, hdmiphy_conf74_25, &hdmi_conf_720p50 },
 	{ 1280, 720, 60, false, hdmiphy_conf74_25, &hdmi_conf_720p60 },
-	{ 1280, 720, 50, false, hdmiphy_conf74_25, &hdmi_conf_720p60 },
-	{ 720, 480, 60, false, hdmiphy_conf27_027, &hdmi_conf_480p },
 	{ 1920, 1080, 50, true, hdmiphy_conf74_25, &hdmi_conf_1080i50 },
-	{ 1920, 1080, 50, false, hdmiphy_conf148_5, &hdmi_conf_1080p50 },
 	{ 1920, 1080, 60, true, hdmiphy_conf74_25, &hdmi_conf_1080i60 },
+	{ 1920, 1080, 50, false, hdmiphy_conf148_5, &hdmi_conf_1080p50 },
 	{ 1920, 1080, 60, false, hdmiphy_conf148_5, &hdmi_conf_1080p60 },
 };
 
@@ -324,7 +889,7 @@ static inline void hdmi_reg_writemask(struct hdmi_context *hdata,
 	writel(value, hdata->regs + reg_id);
 }
 
-static void hdmi_regs_dump(struct hdmi_context *hdata, char *prefix)
+static void hdmi_v13_regs_dump(struct hdmi_context *hdata, char *prefix)
 {
 #define DUMPREG(reg_id) \
 	DRM_DEBUG_KMS("%s:" #reg_id " = %08x\n", prefix, \
@@ -333,6 +898,101 @@ static void hdmi_regs_dump(struct hdmi_context *hdata, char *prefix)
 	DUMPREG(HDMI_INTC_FLAG);
 	DUMPREG(HDMI_INTC_CON);
 	DUMPREG(HDMI_HPD_STATUS);
+	DUMPREG(HDMI_V13_PHY_RSTOUT);
+	DUMPREG(HDMI_V13_PHY_VPLL);
+	DUMPREG(HDMI_V13_PHY_CMU);
+	DUMPREG(HDMI_V13_CORE_RSTOUT);
+
+	DRM_DEBUG_KMS("%s: ---- CORE REGISTERS ----\n", prefix);
+	DUMPREG(HDMI_CON_0);
+	DUMPREG(HDMI_CON_1);
+	DUMPREG(HDMI_CON_2);
+	DUMPREG(HDMI_SYS_STATUS);
+	DUMPREG(HDMI_V13_PHY_STATUS);
+	DUMPREG(HDMI_STATUS_EN);
+	DUMPREG(HDMI_HPD);
+	DUMPREG(HDMI_MODE_SEL);
+	DUMPREG(HDMI_V13_HPD_GEN);
+	DUMPREG(HDMI_V13_DC_CONTROL);
+	DUMPREG(HDMI_V13_VIDEO_PATTERN_GEN);
+
+	DRM_DEBUG_KMS("%s: ---- CORE SYNC REGISTERS ----\n", prefix);
+	DUMPREG(HDMI_H_BLANK_0);
+	DUMPREG(HDMI_H_BLANK_1);
+	DUMPREG(HDMI_V13_V_BLANK_0);
+	DUMPREG(HDMI_V13_V_BLANK_1);
+	DUMPREG(HDMI_V13_V_BLANK_2);
+	DUMPREG(HDMI_V13_H_V_LINE_0);
+	DUMPREG(HDMI_V13_H_V_LINE_1);
+	DUMPREG(HDMI_V13_H_V_LINE_2);
+	DUMPREG(HDMI_VSYNC_POL);
+	DUMPREG(HDMI_INT_PRO_MODE);
+	DUMPREG(HDMI_V13_V_BLANK_F_0);
+	DUMPREG(HDMI_V13_V_BLANK_F_1);
+	DUMPREG(HDMI_V13_V_BLANK_F_2);
+	DUMPREG(HDMI_V13_H_SYNC_GEN_0);
+	DUMPREG(HDMI_V13_H_SYNC_GEN_1);
+	DUMPREG(HDMI_V13_H_SYNC_GEN_2);
+	DUMPREG(HDMI_V13_V_SYNC_GEN_1_0);
+	DUMPREG(HDMI_V13_V_SYNC_GEN_1_1);
+	DUMPREG(HDMI_V13_V_SYNC_GEN_1_2);
+	DUMPREG(HDMI_V13_V_SYNC_GEN_2_0);
+	DUMPREG(HDMI_V13_V_SYNC_GEN_2_1);
+	DUMPREG(HDMI_V13_V_SYNC_GEN_2_2);
+	DUMPREG(HDMI_V13_V_SYNC_GEN_3_0);
+	DUMPREG(HDMI_V13_V_SYNC_GEN_3_1);
+	DUMPREG(HDMI_V13_V_SYNC_GEN_3_2);
+
+	DRM_DEBUG_KMS("%s: ---- TG REGISTERS ----\n", prefix);
+	DUMPREG(HDMI_TG_CMD);
+	DUMPREG(HDMI_TG_H_FSZ_L);
+	DUMPREG(HDMI_TG_H_FSZ_H);
+	DUMPREG(HDMI_TG_HACT_ST_L);
+	DUMPREG(HDMI_TG_HACT_ST_H);
+	DUMPREG(HDMI_TG_HACT_SZ_L);
+	DUMPREG(HDMI_TG_HACT_SZ_H);
+	DUMPREG(HDMI_TG_V_FSZ_L);
+	DUMPREG(HDMI_TG_V_FSZ_H);
+	DUMPREG(HDMI_TG_VSYNC_L);
+	DUMPREG(HDMI_TG_VSYNC_H);
+	DUMPREG(HDMI_TG_VSYNC2_L);
+	DUMPREG(HDMI_TG_VSYNC2_H);
+	DUMPREG(HDMI_TG_VACT_ST_L);
+	DUMPREG(HDMI_TG_VACT_ST_H);
+	DUMPREG(HDMI_TG_VACT_SZ_L);
+	DUMPREG(HDMI_TG_VACT_SZ_H);
+	DUMPREG(HDMI_TG_FIELD_CHG_L);
+	DUMPREG(HDMI_TG_FIELD_CHG_H);
+	DUMPREG(HDMI_TG_VACT_ST2_L);
+	DUMPREG(HDMI_TG_VACT_ST2_H);
+	DUMPREG(HDMI_TG_VSYNC_TOP_HDMI_L);
+	DUMPREG(HDMI_TG_VSYNC_TOP_HDMI_H);
+	DUMPREG(HDMI_TG_VSYNC_BOT_HDMI_L);
+	DUMPREG(HDMI_TG_VSYNC_BOT_HDMI_H);
+	DUMPREG(HDMI_TG_FIELD_TOP_HDMI_L);
+	DUMPREG(HDMI_TG_FIELD_TOP_HDMI_H);
+	DUMPREG(HDMI_TG_FIELD_BOT_HDMI_L);
+	DUMPREG(HDMI_TG_FIELD_BOT_HDMI_H);
+#undef DUMPREG
+}
+
+static void hdmi_v14_regs_dump(struct hdmi_context *hdata, char *prefix)
+{
+	int i;
+
+#define DUMPREG(reg_id) \
+	DRM_DEBUG_KMS("%s:" #reg_id " = %08x\n", prefix, \
+	readl(hdata->regs + reg_id))
+
+	DRM_DEBUG_KMS("%s: ---- CONTROL REGISTERS ----\n", prefix);
+	DUMPREG(HDMI_INTC_CON);
+	DUMPREG(HDMI_INTC_FLAG);
+	DUMPREG(HDMI_HPD_STATUS);
+	DUMPREG(HDMI_INTC_CON_1);
+	DUMPREG(HDMI_INTC_FLAG_1);
+	DUMPREG(HDMI_PHY_STATUS_0);
+	DUMPREG(HDMI_PHY_STATUS_PLL);
+	DUMPREG(HDMI_PHY_CON_0);
 	DUMPREG(HDMI_PHY_RSTOUT);
 	DUMPREG(HDMI_PHY_VPLL);
 	DUMPREG(HDMI_PHY_CMU);
@@ -343,40 +1003,93 @@ static void hdmi_regs_dump(struct hdmi_context *hdata, char *prefix)
 	DUMPREG(HDMI_CON_1);
 	DUMPREG(HDMI_CON_2);
 	DUMPREG(HDMI_SYS_STATUS);
-	DUMPREG(HDMI_PHY_STATUS);
+	DUMPREG(HDMI_PHY_STATUS_0);
 	DUMPREG(HDMI_STATUS_EN);
 	DUMPREG(HDMI_HPD);
 	DUMPREG(HDMI_MODE_SEL);
-	DUMPREG(HDMI_HPD_GEN);
+	DUMPREG(HDMI_ENC_EN);
 	DUMPREG(HDMI_DC_CONTROL);
 	DUMPREG(HDMI_VIDEO_PATTERN_GEN);
 
 	DRM_DEBUG_KMS("%s: ---- CORE SYNC REGISTERS ----\n", prefix);
 	DUMPREG(HDMI_H_BLANK_0);
 	DUMPREG(HDMI_H_BLANK_1);
-	DUMPREG(HDMI_V_BLANK_0);
-	DUMPREG(HDMI_V_BLANK_1);
-	DUMPREG(HDMI_V_BLANK_2);
-	DUMPREG(HDMI_H_V_LINE_0);
-	DUMPREG(HDMI_H_V_LINE_1);
-	DUMPREG(HDMI_H_V_LINE_2);
+	DUMPREG(HDMI_V2_BLANK_0);
+	DUMPREG(HDMI_V2_BLANK_1);
+	DUMPREG(HDMI_V1_BLANK_0);
+	DUMPREG(HDMI_V1_BLANK_1);
+	DUMPREG(HDMI_V_LINE_0);
+	DUMPREG(HDMI_V_LINE_1);
+	DUMPREG(HDMI_H_LINE_0);
+	DUMPREG(HDMI_H_LINE_1);
+	DUMPREG(HDMI_HSYNC_POL);
+
 	DUMPREG(HDMI_VSYNC_POL);
 	DUMPREG(HDMI_INT_PRO_MODE);
-	DUMPREG(HDMI_V_BLANK_F_0);
-	DUMPREG(HDMI_V_BLANK_F_1);
-	DUMPREG(HDMI_V_BLANK_F_2);
-	DUMPREG(HDMI_H_SYNC_GEN_0);
-	DUMPREG(HDMI_H_SYNC_GEN_1);
-	DUMPREG(HDMI_H_SYNC_GEN_2);
-	DUMPREG(HDMI_V_SYNC_GEN_1_0);
-	DUMPREG(HDMI_V_SYNC_GEN_1_1);
-	DUMPREG(HDMI_V_SYNC_GEN_1_2);
-	DUMPREG(HDMI_V_SYNC_GEN_2_0);
-	DUMPREG(HDMI_V_SYNC_GEN_2_1);
-	DUMPREG(HDMI_V_SYNC_GEN_2_2);
-	DUMPREG(HDMI_V_SYNC_GEN_3_0);
-	DUMPREG(HDMI_V_SYNC_GEN_3_1);
-	DUMPREG(HDMI_V_SYNC_GEN_3_2);
+	DUMPREG(HDMI_V_BLANK_F0_0);
+	DUMPREG(HDMI_V_BLANK_F0_1);
+	DUMPREG(HDMI_V_BLANK_F1_0);
+	DUMPREG(HDMI_V_BLANK_F1_1);
+
+	DUMPREG(HDMI_H_SYNC_START_0);
+	DUMPREG(HDMI_H_SYNC_START_1);
+	DUMPREG(HDMI_H_SYNC_END_0);
+	DUMPREG(HDMI_H_SYNC_END_1);
+
+	DUMPREG(HDMI_V_SYNC_LINE_BEF_2_0);
+	DUMPREG(HDMI_V_SYNC_LINE_BEF_2_1);
+	DUMPREG(HDMI_V_SYNC_LINE_BEF_1_0);
+	DUMPREG(HDMI_V_SYNC_LINE_BEF_1_1);
+
+	DUMPREG(HDMI_V_SYNC_LINE_AFT_2_0);
+	DUMPREG(HDMI_V_SYNC_LINE_AFT_2_1);
+	DUMPREG(HDMI_V_SYNC_LINE_AFT_1_0);
+	DUMPREG(HDMI_V_SYNC_LINE_AFT_1_1);
+
+	DUMPREG(HDMI_V_SYNC_LINE_AFT_PXL_2_0);
+	DUMPREG(HDMI_V_SYNC_LINE_AFT_PXL_2_1);
+	DUMPREG(HDMI_V_SYNC_LINE_AFT_PXL_1_0);
+	DUMPREG(HDMI_V_SYNC_LINE_AFT_PXL_1_1);
+
+	DUMPREG(HDMI_V_BLANK_F2_0);
+	DUMPREG(HDMI_V_BLANK_F2_1);
+	DUMPREG(HDMI_V_BLANK_F3_0);
+	DUMPREG(HDMI_V_BLANK_F3_1);
+	DUMPREG(HDMI_V_BLANK_F4_0);
+	DUMPREG(HDMI_V_BLANK_F4_1);
+	DUMPREG(HDMI_V_BLANK_F5_0);
+	DUMPREG(HDMI_V_BLANK_F5_1);
+
+	DUMPREG(HDMI_V_SYNC_LINE_AFT_3_0);
+	DUMPREG(HDMI_V_SYNC_LINE_AFT_3_1);
+	DUMPREG(HDMI_V_SYNC_LINE_AFT_4_0);
+	DUMPREG(HDMI_V_SYNC_LINE_AFT_4_1);
+	DUMPREG(HDMI_V_SYNC_LINE_AFT_5_0);
+	DUMPREG(HDMI_V_SYNC_LINE_AFT_5_1);
+	DUMPREG(HDMI_V_SYNC_LINE_AFT_6_0);
+	DUMPREG(HDMI_V_SYNC_LINE_AFT_6_1);
+
+	DUMPREG(HDMI_V_SYNC_LINE_AFT_PXL_3_0);
+	DUMPREG(HDMI_V_SYNC_LINE_AFT_PXL_3_1);
+	DUMPREG(HDMI_V_SYNC_LINE_AFT_PXL_4_0);
+	DUMPREG(HDMI_V_SYNC_LINE_AFT_PXL_4_1);
+	DUMPREG(HDMI_V_SYNC_LINE_AFT_PXL_5_0);
+	DUMPREG(HDMI_V_SYNC_LINE_AFT_PXL_5_1);
+	DUMPREG(HDMI_V_SYNC_LINE_AFT_PXL_6_0);
+	DUMPREG(HDMI_V_SYNC_LINE_AFT_PXL_6_1);
+
+	DUMPREG(HDMI_VACT_SPACE_1_0);
+	DUMPREG(HDMI_VACT_SPACE_1_1);
+	DUMPREG(HDMI_VACT_SPACE_2_0);
+	DUMPREG(HDMI_VACT_SPACE_2_1);
+	DUMPREG(HDMI_VACT_SPACE_3_0);
+	DUMPREG(HDMI_VACT_SPACE_3_1);
+	DUMPREG(HDMI_VACT_SPACE_4_0);
+	DUMPREG(HDMI_VACT_SPACE_4_1);
+	DUMPREG(HDMI_VACT_SPACE_5_0);
+	DUMPREG(HDMI_VACT_SPACE_5_1);
+	DUMPREG(HDMI_VACT_SPACE_6_0);
+	DUMPREG(HDMI_VACT_SPACE_6_1);
 
 	DRM_DEBUG_KMS("%s: ---- TG REGISTERS ----\n", prefix);
 	DUMPREG(HDMI_TG_CMD);
@@ -400,6 +1113,10 @@ static void hdmi_regs_dump(struct hdmi_context *hdata, char *prefix)
 	DUMPREG(HDMI_TG_FIELD_CHG_H);
 	DUMPREG(HDMI_TG_VACT_ST2_L);
 	DUMPREG(HDMI_TG_VACT_ST2_H);
+	DUMPREG(HDMI_TG_VACT_ST3_L);
+	DUMPREG(HDMI_TG_VACT_ST3_H);
+	DUMPREG(HDMI_TG_VACT_ST4_L);
+	DUMPREG(HDMI_TG_VACT_ST4_H);
 	DUMPREG(HDMI_TG_VSYNC_TOP_HDMI_L);
 	DUMPREG(HDMI_TG_VSYNC_TOP_HDMI_H);
 	DUMPREG(HDMI_TG_VSYNC_BOT_HDMI_L);
@@ -408,10 +1125,49 @@ static void hdmi_regs_dump(struct hdmi_context *hdata, char *prefix)
 	DUMPREG(HDMI_TG_FIELD_TOP_HDMI_H);
 	DUMPREG(HDMI_TG_FIELD_BOT_HDMI_L);
 	DUMPREG(HDMI_TG_FIELD_BOT_HDMI_H);
+	DUMPREG(HDMI_TG_3D);
+
+	DRM_DEBUG_KMS("%s: ---- PACKET REGISTERS ----\n", prefix);
+	DUMPREG(HDMI_AVI_CON);
+	DUMPREG(HDMI_AVI_HEADER0);
+	DUMPREG(HDMI_AVI_HEADER1);
+	DUMPREG(HDMI_AVI_HEADER2);
+	DUMPREG(HDMI_AVI_CHECK_SUM);
+	DUMPREG(HDMI_VSI_CON);
+	DUMPREG(HDMI_VSI_HEADER0);
+	DUMPREG(HDMI_VSI_HEADER1);
+	DUMPREG(HDMI_VSI_HEADER2);
+	for (i = 0; i < 7; ++i)
+		DUMPREG(HDMI_VSI_DATA(i));
+
 #undef DUMPREG
 }
 
-static int hdmi_conf_index(struct drm_display_mode *mode)
+static void hdmi_regs_dump(struct hdmi_context *hdata, char *prefix)
+{
+	if (hdata->is_v13)
+		hdmi_v13_regs_dump(hdata, prefix);
+	else
+		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 int hdmi_v14_conf_index(struct drm_display_mode *mode)
 {
 	int i;
 
@@ -424,7 +1180,16 @@ static int hdmi_conf_index(struct drm_display_mode *mode)
 				 true : false))
 			return i;
 
-	return -1;
+	return -EINVAL;
+}
+
+static int hdmi_conf_index(struct hdmi_context *hdata,
+			   struct drm_display_mode *mode)
+{
+	if (hdata->is_v13)
+		return hdmi_v13_conf_index(mode);
+
+	return hdmi_v14_conf_index(mode);
 }
 
 static bool hdmi_is_connected(void *ctx)
@@ -462,29 +1227,69 @@ static int hdmi_get_edid(void *ctx, struct drm_connector *connector,
 	return 0;
 }
 
-static int hdmi_check_timing(void *ctx, void *timing)
+static int hdmi_v13_check_timing(struct fb_videomode *check_timing)
 {
-	struct fb_videomode *check_timing = timing;
 	int i;
 
-	DRM_DEBUG_KMS("[%d] %s\n", __LINE__, __func__);
+	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);
 
-	DRM_DEBUG_KMS("[%d]x[%d] [%d]Hz [%x]\n", check_timing->xres,
-			check_timing->yres, check_timing->refresh,
-			check_timing->vmode);
+	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;
 
-	for (i = 0; i < ARRAY_SIZE(hdmi_confs); ++i)
+	/* TODO */
+
+	return -EINVAL;
+}
+
+static int hdmi_v14_check_timing(struct fb_videomode *check_timing)
+{
+	int 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_confs); i++)
 		if (hdmi_confs[i].width == check_timing->xres &&
 			hdmi_confs[i].height == check_timing->yres &&
 			hdmi_confs[i].vrefresh == check_timing->refresh &&
 			hdmi_confs[i].interlace ==
 			((check_timing->vmode & FB_VMODE_INTERLACED) ?
 			 true : false))
-			return 0;
+				return 0;
+
+	/* TODO */
 
 	return -EINVAL;
 }
 
+static int hdmi_check_timing(void *ctx, void *timing)
+{
+	struct hdmi_context *hdata = (struct hdmi_context *)ctx;
+	struct fb_videomode *check_timing = timing;
+
+	DRM_DEBUG_KMS("[%d] %s\n", __LINE__, __func__);
+
+	DRM_DEBUG_KMS("[%d]x[%d] [%d]Hz [%x]\n", check_timing->xres,
+			check_timing->yres, check_timing->refresh,
+			check_timing->vmode);
+
+	if (hdata->is_v13)
+		return hdmi_v13_check_timing(check_timing);
+	else
+		return hdmi_v14_check_timing(check_timing);
+}
+
 static int hdmi_display_power_on(void *ctx, int mode)
 {
 	DRM_DEBUG_KMS("[%d] %s\n", __LINE__, __func__);
@@ -514,15 +1319,185 @@ static struct exynos_hdmi_display_ops display_ops = {
 	.power_on	= hdmi_display_power_on,
 };
 
+static void hdmi_set_acr(u32 freq, u8 *acr)
+{
+	u32 n, cts;
+
+	switch (freq) {
+	case 32000:
+		n = 4096;
+		cts = 27000;
+		break;
+	case 44100:
+		n = 6272;
+		cts = 30000;
+		break;
+	case 88200:
+		n = 12544;
+		cts = 30000;
+		break;
+	case 176400:
+		n = 25088;
+		cts = 30000;
+		break;
+	case 48000:
+		n = 6144;
+		cts = 27000;
+		break;
+	case 96000:
+		n = 12288;
+		cts = 27000;
+		break;
+	case 192000:
+		n = 24576;
+		cts = 27000;
+		break;
+	default:
+		n = 0;
+		cts = 0;
+		break;
+	}
+
+	acr[1] = cts >> 16;
+	acr[2] = cts >> 8 & 0xff;
+	acr[3] = cts & 0xff;
+
+	acr[4] = n >> 16;
+	acr[5] = n >> 8 & 0xff;
+	acr[6] = n & 0xff;
+}
+
+static void hdmi_reg_acr(struct hdmi_context *hdata, u8 *acr)
+{
+	hdmi_reg_writeb(hdata, HDMI_ACR_N0, acr[6]);
+	hdmi_reg_writeb(hdata, HDMI_ACR_N1, acr[5]);
+	hdmi_reg_writeb(hdata, HDMI_ACR_N2, acr[4]);
+	hdmi_reg_writeb(hdata, HDMI_ACR_MCTS0, acr[3]);
+	hdmi_reg_writeb(hdata, HDMI_ACR_MCTS1, acr[2]);
+	hdmi_reg_writeb(hdata, HDMI_ACR_MCTS2, acr[1]);
+	hdmi_reg_writeb(hdata, HDMI_ACR_CTS0, acr[3]);
+	hdmi_reg_writeb(hdata, HDMI_ACR_CTS1, acr[2]);
+	hdmi_reg_writeb(hdata, HDMI_ACR_CTS2, acr[1]);
+
+	if (hdata->is_v13)
+		hdmi_reg_writeb(hdata, HDMI_V13_ACR_CON, 4);
+	else
+		hdmi_reg_writeb(hdata, HDMI_ACR_CON, 4);
+}
+
+static void hdmi_audio_init(struct hdmi_context *hdata)
+{
+	u32 sample_rate, bits_per_sample, frame_size_code;
+	u32 data_num, bit_ch, sample_frq;
+	u32 val;
+	u8 acr[7];
+
+	sample_rate = 44100;
+	bits_per_sample = 16;
+	frame_size_code = 0;
+
+	switch (bits_per_sample) {
+	case 20:
+		data_num = 2;
+		bit_ch  = 1;
+		break;
+	case 24:
+		data_num = 3;
+		bit_ch  = 1;
+		break;
+	default:
+		data_num = 1;
+		bit_ch  = 0;
+		break;
+	}
+
+	hdmi_set_acr(sample_rate, acr);
+	hdmi_reg_acr(hdata, acr);
+
+	hdmi_reg_writeb(hdata, HDMI_I2S_MUX_CON, HDMI_I2S_IN_DISABLE
+				| HDMI_I2S_AUD_I2S | HDMI_I2S_CUV_I2S_ENABLE
+				| HDMI_I2S_MUX_ENABLE);
+
+	hdmi_reg_writeb(hdata, HDMI_I2S_MUX_CH, HDMI_I2S_CH0_EN
+			| HDMI_I2S_CH1_EN | HDMI_I2S_CH2_EN);
+
+	hdmi_reg_writeb(hdata, HDMI_I2S_MUX_CUV, HDMI_I2S_CUV_RL_EN);
+
+	sample_frq = (sample_rate == 44100) ? 0 :
+			(sample_rate == 48000) ? 2 :
+			(sample_rate == 32000) ? 3 :
+			(sample_rate == 96000) ? 0xa : 0x0;
+
+	hdmi_reg_writeb(hdata, HDMI_I2S_CLK_CON, HDMI_I2S_CLK_DIS);
+	hdmi_reg_writeb(hdata, HDMI_I2S_CLK_CON, HDMI_I2S_CLK_EN);
+
+	val = hdmi_reg_read(hdata, HDMI_I2S_DSD_CON) | 0x01;
+	hdmi_reg_writeb(hdata, HDMI_I2S_DSD_CON, val);
+
+	/* Configuration I2S input ports. Configure I2S_PIN_SEL_0~4 */
+	hdmi_reg_writeb(hdata, HDMI_I2S_PIN_SEL_0, HDMI_I2S_SEL_SCLK(5)
+			| HDMI_I2S_SEL_LRCK(6));
+	hdmi_reg_writeb(hdata, HDMI_I2S_PIN_SEL_1, HDMI_I2S_SEL_SDATA1(1)
+			| HDMI_I2S_SEL_SDATA2(4));
+	hdmi_reg_writeb(hdata, HDMI_I2S_PIN_SEL_2, HDMI_I2S_SEL_SDATA3(1)
+			| HDMI_I2S_SEL_SDATA2(2));
+	hdmi_reg_writeb(hdata, HDMI_I2S_PIN_SEL_3, HDMI_I2S_SEL_DSD(0));
+
+	/* I2S_CON_1 & 2 */
+	hdmi_reg_writeb(hdata, HDMI_I2S_CON_1, HDMI_I2S_SCLK_FALLING_EDGE
+			| HDMI_I2S_L_CH_LOW_POL);
+	hdmi_reg_writeb(hdata, HDMI_I2S_CON_2, HDMI_I2S_MSB_FIRST_MODE
+			| HDMI_I2S_SET_BIT_CH(bit_ch)
+			| HDMI_I2S_SET_SDATA_BIT(data_num)
+			| HDMI_I2S_BASIC_FORMAT);
+
+	/* Configure register related to CUV information */
+	hdmi_reg_writeb(hdata, HDMI_I2S_CH_ST_0, HDMI_I2S_CH_STATUS_MODE_0
+			| HDMI_I2S_2AUD_CH_WITHOUT_PREEMPH
+			| HDMI_I2S_COPYRIGHT
+			| HDMI_I2S_LINEAR_PCM
+			| HDMI_I2S_CONSUMER_FORMAT);
+	hdmi_reg_writeb(hdata, HDMI_I2S_CH_ST_1, HDMI_I2S_CD_PLAYER);
+	hdmi_reg_writeb(hdata, HDMI_I2S_CH_ST_2, HDMI_I2S_SET_SOURCE_NUM(0));
+	hdmi_reg_writeb(hdata, HDMI_I2S_CH_ST_3, HDMI_I2S_CLK_ACCUR_LEVEL_2
+			| HDMI_I2S_SET_SMP_FREQ(sample_frq));
+	hdmi_reg_writeb(hdata, HDMI_I2S_CH_ST_4,
+			HDMI_I2S_ORG_SMP_FREQ_44_1
+			| HDMI_I2S_WORD_LEN_MAX24_24BITS
+			| HDMI_I2S_WORD_LEN_MAX_24BITS);
+
+	hdmi_reg_writeb(hdata, HDMI_I2S_CH_ST_CON, HDMI_I2S_CH_STATUS_RELOAD);
+}
+
+static void hdmi_audio_control(struct hdmi_context *hdata, bool onoff)
+{
+	u32 mod;
+
+	mod = hdmi_reg_read(hdata, HDMI_MODE_SEL);
+	if (mod & HDMI_DVI_MODE_EN)
+		return;
+
+	hdmi_reg_writeb(hdata, HDMI_AUI_CON, onoff ? 2 : 0);
+	hdmi_reg_writemask(hdata, HDMI_CON_0, onoff ?
+			HDMI_ASP_EN : HDMI_ASP_DIS, HDMI_ASP_MASK);
+}
+
 static void hdmi_conf_reset(struct hdmi_context *hdata)
 {
+	u32 reg;
+
 	/* disable hpd handle for drm */
 	hdata->hpd_handle = false;
 
+	if (hdata->is_v13)
+		reg = HDMI_V13_CORE_RSTOUT;
+	else
+		reg = HDMI_CORE_RSTOUT;
+
 	/* resetting HDMI core */
-	hdmi_reg_writemask(hdata, HDMI_CORE_RSTOUT,  0, HDMI_CORE_SW_RSTOUT);
+	hdmi_reg_writemask(hdata, reg,  0, HDMI_CORE_SW_RSTOUT);
 	mdelay(10);
-	hdmi_reg_writemask(hdata, HDMI_CORE_RSTOUT, ~0, HDMI_CORE_SW_RSTOUT);
+	hdmi_reg_writemask(hdata, reg, ~0, HDMI_CORE_SW_RSTOUT);
 	mdelay(10);
 
 	/* enable hpd handle for drm */
@@ -546,27 +1521,126 @@ static void hdmi_conf_init(struct hdmi_context *hdata)
 		HDMI_MODE_HDMI_EN, HDMI_MODE_MASK);
 	/* disable bluescreen */
 	hdmi_reg_writemask(hdata, HDMI_CON_0, 0, HDMI_BLUE_SCR_EN);
-	/* choose bluescreen (fecal) color */
-	hdmi_reg_writeb(hdata, HDMI_BLUE_SCREEN_0, 0x12);
-	hdmi_reg_writeb(hdata, HDMI_BLUE_SCREEN_1, 0x34);
-	hdmi_reg_writeb(hdata, HDMI_BLUE_SCREEN_2, 0x56);
-	/* enable AVI packet every vsync, fixes purple line problem */
-	hdmi_reg_writeb(hdata, HDMI_AVI_CON, 0x02);
-	/* force RGB, look to CEA-861-D, table 7 for more detail */
-	hdmi_reg_writeb(hdata, HDMI_AVI_BYTE(0), 0 << 5);
-	hdmi_reg_writemask(hdata, HDMI_CON_1, 0x10 << 5, 0x11 << 5);
-
-	hdmi_reg_writeb(hdata, HDMI_SPD_CON, 0x02);
-	hdmi_reg_writeb(hdata, HDMI_AUI_CON, 0x02);
-	hdmi_reg_writeb(hdata, HDMI_ACR_CON, 0x04);
+
+	if (hdata->is_v13) {
+		/* choose bluescreen (fecal) color */
+		hdmi_reg_writeb(hdata, HDMI_V13_BLUE_SCREEN_0, 0x12);
+		hdmi_reg_writeb(hdata, HDMI_V13_BLUE_SCREEN_1, 0x34);
+		hdmi_reg_writeb(hdata, HDMI_V13_BLUE_SCREEN_2, 0x56);
+
+		/* enable AVI packet every vsync, fixes purple line problem */
+		hdmi_reg_writeb(hdata, HDMI_V13_AVI_CON, 0x02);
+		/* force RGB, look to CEA-861-D, table 7 for more detail */
+		hdmi_reg_writeb(hdata, HDMI_V13_AVI_BYTE(0), 0 << 5);
+		hdmi_reg_writemask(hdata, HDMI_CON_1, 0x10 << 5, 0x11 << 5);
+
+		hdmi_reg_writeb(hdata, HDMI_V13_SPD_CON, 0x02);
+		hdmi_reg_writeb(hdata, HDMI_V13_AUI_CON, 0x02);
+		hdmi_reg_writeb(hdata, HDMI_V13_ACR_CON, 0x04);
+	} else {
+		/* enable AVI packet every vsync, fixes purple line problem */
+		hdmi_reg_writeb(hdata, HDMI_AVI_CON, 0x02);
+		hdmi_reg_writeb(hdata, HDMI_AVI_BYTE(1), 2 << 5);
+		hdmi_reg_writemask(hdata, HDMI_CON_1, 2, 3 << 5);
+	}
 
 	/* enable hpd handle for drm */
 	hdata->hpd_handle = true;
 }
 
-static void hdmi_timing_apply(struct hdmi_context *hdata,
-				 const struct hdmi_preset_conf *conf)
+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;
+	int tries;
+
+	/* setting core registers */
+	hdmi_reg_writeb(hdata, HDMI_H_BLANK_0, core->h_blank[0]);
+	hdmi_reg_writeb(hdata, HDMI_H_BLANK_1, core->h_blank[1]);
+	hdmi_reg_writeb(hdata, HDMI_V13_V_BLANK_0, core->v_blank[0]);
+	hdmi_reg_writeb(hdata, HDMI_V13_V_BLANK_1, core->v_blank[1]);
+	hdmi_reg_writeb(hdata, HDMI_V13_V_BLANK_2, core->v_blank[2]);
+	hdmi_reg_writeb(hdata, HDMI_V13_H_V_LINE_0, core->h_v_line[0]);
+	hdmi_reg_writeb(hdata, HDMI_V13_H_V_LINE_1, core->h_v_line[1]);
+	hdmi_reg_writeb(hdata, HDMI_V13_H_V_LINE_2, core->h_v_line[2]);
+	hdmi_reg_writeb(hdata, HDMI_VSYNC_POL, core->vsync_pol[0]);
+	hdmi_reg_writeb(hdata, HDMI_INT_PRO_MODE, core->int_pro_mode[0]);
+	hdmi_reg_writeb(hdata, HDMI_V13_V_BLANK_F_0, core->v_blank_f[0]);
+	hdmi_reg_writeb(hdata, HDMI_V13_V_BLANK_F_1, core->v_blank_f[1]);
+	hdmi_reg_writeb(hdata, HDMI_V13_V_BLANK_F_2, core->v_blank_f[2]);
+	hdmi_reg_writeb(hdata, HDMI_V13_H_SYNC_GEN_0, core->h_sync_gen[0]);
+	hdmi_reg_writeb(hdata, HDMI_V13_H_SYNC_GEN_1, core->h_sync_gen[1]);
+	hdmi_reg_writeb(hdata, HDMI_V13_H_SYNC_GEN_2, core->h_sync_gen[2]);
+	hdmi_reg_writeb(hdata, HDMI_V13_V_SYNC_GEN_1_0, core->v_sync_gen1[0]);
+	hdmi_reg_writeb(hdata, HDMI_V13_V_SYNC_GEN_1_1, core->v_sync_gen1[1]);
+	hdmi_reg_writeb(hdata, HDMI_V13_V_SYNC_GEN_1_2, core->v_sync_gen1[2]);
+	hdmi_reg_writeb(hdata, HDMI_V13_V_SYNC_GEN_2_0, core->v_sync_gen2[0]);
+	hdmi_reg_writeb(hdata, HDMI_V13_V_SYNC_GEN_2_1, core->v_sync_gen2[1]);
+	hdmi_reg_writeb(hdata, HDMI_V13_V_SYNC_GEN_2_2, core->v_sync_gen2[2]);
+	hdmi_reg_writeb(hdata, HDMI_V13_V_SYNC_GEN_3_0, core->v_sync_gen3[0]);
+	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);
+
+	/* waiting for HDMIPHY's PLL to get to steady state */
+	for (tries = 100; tries; --tries) {
+		u32 val = hdmi_reg_read(hdata, HDMI_V13_PHY_STATUS);
+		if (val & HDMI_PHY_STATUS_READY)
+			break;
+		mdelay(1);
+	}
+	/* steady state not achieved */
+	if (tries == 0) {
+		DRM_ERROR("hdmiphy's pll could not reach steady state.\n");
+		hdmi_regs_dump(hdata, "timing apply");
+	}
+
+	clk_disable(hdata->res.sclk_hdmi);
+	clk_set_parent(hdata->res.sclk_hdmi, hdata->res.sclk_hdmiphy);
+	clk_enable(hdata->res.sclk_hdmi);
+
+	/* enable HDMI and timing generator */
+	hdmi_reg_writemask(hdata, HDMI_CON_0, ~0, HDMI_EN);
+	if (core->int_pro_mode[0])
+		hdmi_reg_writemask(hdata, HDMI_TG_CMD, ~0, HDMI_TG_EN |
+				HDMI_FIELD_EN);
+	else
+		hdmi_reg_writemask(hdata, HDMI_TG_CMD, ~0, HDMI_TG_EN);
+}
+
+static void hdmi_v14_timing_apply(struct hdmi_context *hdata)
+{
+	const struct hdmi_preset_conf *conf = hdmi_confs[hdata->cur_conf].conf;
 	const struct hdmi_core_regs *core = &conf->core;
 	const struct hdmi_tg_regs *tg = &conf->tg;
 	int tries;
@@ -574,29 +1648,102 @@ static void hdmi_timing_apply(struct hdmi_context *hdata,
 	/* setting core registers */
 	hdmi_reg_writeb(hdata, HDMI_H_BLANK_0, core->h_blank[0]);
 	hdmi_reg_writeb(hdata, HDMI_H_BLANK_1, core->h_blank[1]);
-	hdmi_reg_writeb(hdata, HDMI_V_BLANK_0, core->v_blank[0]);
-	hdmi_reg_writeb(hdata, HDMI_V_BLANK_1, core->v_blank[1]);
-	hdmi_reg_writeb(hdata, HDMI_V_BLANK_2, core->v_blank[2]);
-	hdmi_reg_writeb(hdata, HDMI_H_V_LINE_0, core->h_v_line[0]);
-	hdmi_reg_writeb(hdata, HDMI_H_V_LINE_1, core->h_v_line[1]);
-	hdmi_reg_writeb(hdata, HDMI_H_V_LINE_2, core->h_v_line[2]);
+	hdmi_reg_writeb(hdata, HDMI_V2_BLANK_0, core->v2_blank[0]);
+	hdmi_reg_writeb(hdata, HDMI_V2_BLANK_1, core->v2_blank[1]);
+	hdmi_reg_writeb(hdata, HDMI_V1_BLANK_0, core->v1_blank[0]);
+	hdmi_reg_writeb(hdata, HDMI_V1_BLANK_1, core->v1_blank[1]);
+	hdmi_reg_writeb(hdata, HDMI_V_LINE_0, core->v_line[0]);
+	hdmi_reg_writeb(hdata, HDMI_V_LINE_1, core->v_line[1]);
+	hdmi_reg_writeb(hdata, HDMI_H_LINE_0, core->h_line[0]);
+	hdmi_reg_writeb(hdata, HDMI_H_LINE_1, core->h_line[1]);
+	hdmi_reg_writeb(hdata, HDMI_HSYNC_POL, core->hsync_pol[0]);
 	hdmi_reg_writeb(hdata, HDMI_VSYNC_POL, core->vsync_pol[0]);
 	hdmi_reg_writeb(hdata, HDMI_INT_PRO_MODE, core->int_pro_mode[0]);
-	hdmi_reg_writeb(hdata, HDMI_V_BLANK_F_0, core->v_blank_f[0]);
-	hdmi_reg_writeb(hdata, HDMI_V_BLANK_F_1, core->v_blank_f[1]);
-	hdmi_reg_writeb(hdata, HDMI_V_BLANK_F_2, core->v_blank_f[2]);
-	hdmi_reg_writeb(hdata, HDMI_H_SYNC_GEN_0, core->h_sync_gen[0]);
-	hdmi_reg_writeb(hdata, HDMI_H_SYNC_GEN_1, core->h_sync_gen[1]);
-	hdmi_reg_writeb(hdata, HDMI_H_SYNC_GEN_2, core->h_sync_gen[2]);
-	hdmi_reg_writeb(hdata, HDMI_V_SYNC_GEN_1_0, core->v_sync_gen1[0]);
-	hdmi_reg_writeb(hdata, HDMI_V_SYNC_GEN_1_1, core->v_sync_gen1[1]);
-	hdmi_reg_writeb(hdata, HDMI_V_SYNC_GEN_1_2, core->v_sync_gen1[2]);
-	hdmi_reg_writeb(hdata, HDMI_V_SYNC_GEN_2_0, core->v_sync_gen2[0]);
-	hdmi_reg_writeb(hdata, HDMI_V_SYNC_GEN_2_1, core->v_sync_gen2[1]);
-	hdmi_reg_writeb(hdata, HDMI_V_SYNC_GEN_2_2, core->v_sync_gen2[2]);
-	hdmi_reg_writeb(hdata, HDMI_V_SYNC_GEN_3_0, core->v_sync_gen3[0]);
-	hdmi_reg_writeb(hdata, HDMI_V_SYNC_GEN_3_1, core->v_sync_gen3[1]);
-	hdmi_reg_writeb(hdata, HDMI_V_SYNC_GEN_3_2, core->v_sync_gen3[2]);
+	hdmi_reg_writeb(hdata, HDMI_V_BLANK_F0_0, core->v_blank_f0[0]);
+	hdmi_reg_writeb(hdata, HDMI_V_BLANK_F0_1, core->v_blank_f0[1]);
+	hdmi_reg_writeb(hdata, HDMI_V_BLANK_F1_0, core->v_blank_f1[0]);
+	hdmi_reg_writeb(hdata, HDMI_V_BLANK_F1_1, core->v_blank_f1[1]);
+	hdmi_reg_writeb(hdata, HDMI_H_SYNC_START_0, core->h_sync_start[0]);
+	hdmi_reg_writeb(hdata, HDMI_H_SYNC_START_1, core->h_sync_start[1]);
+	hdmi_reg_writeb(hdata, HDMI_H_SYNC_END_0, core->h_sync_end[0]);
+	hdmi_reg_writeb(hdata, HDMI_H_SYNC_END_1, core->h_sync_end[1]);
+	hdmi_reg_writeb(hdata, HDMI_V_SYNC_LINE_BEF_2_0,
+			core->v_sync_line_bef_2[0]);
+	hdmi_reg_writeb(hdata, HDMI_V_SYNC_LINE_BEF_2_1,
+			core->v_sync_line_bef_2[1]);
+	hdmi_reg_writeb(hdata, HDMI_V_SYNC_LINE_BEF_1_0,
+			core->v_sync_line_bef_1[0]);
+	hdmi_reg_writeb(hdata, HDMI_V_SYNC_LINE_BEF_1_1,
+			core->v_sync_line_bef_1[1]);
+	hdmi_reg_writeb(hdata, HDMI_V_SYNC_LINE_AFT_2_0,
+			core->v_sync_line_aft_2[0]);
+	hdmi_reg_writeb(hdata, HDMI_V_SYNC_LINE_AFT_2_1,
+			core->v_sync_line_aft_2[1]);
+	hdmi_reg_writeb(hdata, HDMI_V_SYNC_LINE_AFT_1_0,
+			core->v_sync_line_aft_1[0]);
+	hdmi_reg_writeb(hdata, HDMI_V_SYNC_LINE_AFT_1_1,
+			core->v_sync_line_aft_1[1]);
+	hdmi_reg_writeb(hdata, HDMI_V_SYNC_LINE_AFT_PXL_2_0,
+			core->v_sync_line_aft_pxl_2[0]);
+	hdmi_reg_writeb(hdata, HDMI_V_SYNC_LINE_AFT_PXL_2_1,
+			core->v_sync_line_aft_pxl_2[1]);
+	hdmi_reg_writeb(hdata, HDMI_V_SYNC_LINE_AFT_PXL_1_0,
+			core->v_sync_line_aft_pxl_1[0]);
+	hdmi_reg_writeb(hdata, HDMI_V_SYNC_LINE_AFT_PXL_1_1,
+			core->v_sync_line_aft_pxl_1[1]);
+	hdmi_reg_writeb(hdata, HDMI_V_BLANK_F2_0, core->v_blank_f2[0]);
+	hdmi_reg_writeb(hdata, HDMI_V_BLANK_F2_1, core->v_blank_f2[1]);
+	hdmi_reg_writeb(hdata, HDMI_V_BLANK_F3_0, core->v_blank_f3[0]);
+	hdmi_reg_writeb(hdata, HDMI_V_BLANK_F3_1, core->v_blank_f3[1]);
+	hdmi_reg_writeb(hdata, HDMI_V_BLANK_F4_0, core->v_blank_f4[0]);
+	hdmi_reg_writeb(hdata, HDMI_V_BLANK_F4_1, core->v_blank_f4[1]);
+	hdmi_reg_writeb(hdata, HDMI_V_BLANK_F5_0, core->v_blank_f5[0]);
+	hdmi_reg_writeb(hdata, HDMI_V_BLANK_F5_1, core->v_blank_f5[1]);
+	hdmi_reg_writeb(hdata, HDMI_V_SYNC_LINE_AFT_3_0,
+			core->v_sync_line_aft_3[0]);
+	hdmi_reg_writeb(hdata, HDMI_V_SYNC_LINE_AFT_3_1,
+			core->v_sync_line_aft_3[1]);
+	hdmi_reg_writeb(hdata, HDMI_V_SYNC_LINE_AFT_4_0,
+			core->v_sync_line_aft_4[0]);
+	hdmi_reg_writeb(hdata, HDMI_V_SYNC_LINE_AFT_4_1,
+			core->v_sync_line_aft_4[1]);
+	hdmi_reg_writeb(hdata, HDMI_V_SYNC_LINE_AFT_5_0,
+			core->v_sync_line_aft_5[0]);
+	hdmi_reg_writeb(hdata, HDMI_V_SYNC_LINE_AFT_5_1,
+			core->v_sync_line_aft_5[1]);
+	hdmi_reg_writeb(hdata, HDMI_V_SYNC_LINE_AFT_6_0,
+			core->v_sync_line_aft_6[0]);
+	hdmi_reg_writeb(hdata, HDMI_V_SYNC_LINE_AFT_6_1,
+			core->v_sync_line_aft_6[1]);
+	hdmi_reg_writeb(hdata, HDMI_V_SYNC_LINE_AFT_PXL_3_0,
+			core->v_sync_line_aft_pxl_3[0]);
+	hdmi_reg_writeb(hdata, HDMI_V_SYNC_LINE_AFT_PXL_3_1,
+			core->v_sync_line_aft_pxl_3[1]);
+	hdmi_reg_writeb(hdata, HDMI_V_SYNC_LINE_AFT_PXL_4_0,
+			core->v_sync_line_aft_pxl_4[0]);
+	hdmi_reg_writeb(hdata, HDMI_V_SYNC_LINE_AFT_PXL_4_1,
+			core->v_sync_line_aft_pxl_4[1]);
+	hdmi_reg_writeb(hdata, HDMI_V_SYNC_LINE_AFT_PXL_5_0,
+			core->v_sync_line_aft_pxl_5[0]);
+	hdmi_reg_writeb(hdata, HDMI_V_SYNC_LINE_AFT_PXL_5_1,
+			core->v_sync_line_aft_pxl_5[1]);
+	hdmi_reg_writeb(hdata, HDMI_V_SYNC_LINE_AFT_PXL_6_0,
+			core->v_sync_line_aft_pxl_6[0]);
+	hdmi_reg_writeb(hdata, HDMI_V_SYNC_LINE_AFT_PXL_6_1,
+			core->v_sync_line_aft_pxl_6[1]);
+	hdmi_reg_writeb(hdata, HDMI_VACT_SPACE_1_0, core->vact_space_1[0]);
+	hdmi_reg_writeb(hdata, HDMI_VACT_SPACE_1_1, core->vact_space_1[1]);
+	hdmi_reg_writeb(hdata, HDMI_VACT_SPACE_2_0, core->vact_space_2[0]);
+	hdmi_reg_writeb(hdata, HDMI_VACT_SPACE_2_1, core->vact_space_2[1]);
+	hdmi_reg_writeb(hdata, HDMI_VACT_SPACE_3_0, core->vact_space_3[0]);
+	hdmi_reg_writeb(hdata, HDMI_VACT_SPACE_3_1, core->vact_space_3[1]);
+	hdmi_reg_writeb(hdata, HDMI_VACT_SPACE_4_0, core->vact_space_4[0]);
+	hdmi_reg_writeb(hdata, HDMI_VACT_SPACE_4_1, core->vact_space_4[1]);
+	hdmi_reg_writeb(hdata, HDMI_VACT_SPACE_5_0, core->vact_space_5[0]);
+	hdmi_reg_writeb(hdata, HDMI_VACT_SPACE_5_1, core->vact_space_5[1]);
+	hdmi_reg_writeb(hdata, HDMI_VACT_SPACE_6_0, core->vact_space_6[0]);
+	hdmi_reg_writeb(hdata, HDMI_VACT_SPACE_6_1, core->vact_space_6[1]);
+
 	/* 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);
@@ -618,6 +1765,10 @@ static void hdmi_timing_apply(struct hdmi_context *hdata,
 	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_VACT_ST3_L, tg->vact_st3_l);
+	hdmi_reg_writeb(hdata, HDMI_TG_VACT_ST3_H, tg->vact_st3_h);
+	hdmi_reg_writeb(hdata, HDMI_TG_VACT_ST4_L, tg->vact_st4_l);
+	hdmi_reg_writeb(hdata, HDMI_TG_VACT_ST4_H, tg->vact_st4_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);
@@ -626,10 +1777,11 @@ static void hdmi_timing_apply(struct hdmi_context *hdata,
 	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_3D, tg->tg_3d);
 
 	/* waiting for HDMIPHY's PLL to get to steady state */
 	for (tries = 100; tries; --tries) {
-		u32 val = hdmi_reg_read(hdata, HDMI_PHY_STATUS);
+		u32 val = hdmi_reg_read(hdata, HDMI_PHY_STATUS_0);
 		if (val & HDMI_PHY_STATUS_READY)
 			break;
 		mdelay(1);
@@ -653,9 +1805,18 @@ static void hdmi_timing_apply(struct hdmi_context *hdata,
 		hdmi_reg_writemask(hdata, HDMI_TG_CMD, ~0, HDMI_TG_EN);
 }
 
+static void hdmi_timing_apply(struct hdmi_context *hdata)
+{
+	if (hdata->is_v13)
+		hdmi_v13_timing_apply(hdata);
+	else
+		hdmi_v14_timing_apply(hdata);
+}
+
 static void hdmiphy_conf_reset(struct hdmi_context *hdata)
 {
 	u8 buffer[2];
+	u32 reg;
 
 	clk_disable(hdata->res.sclk_hdmi);
 	clk_set_parent(hdata->res.sclk_hdmi, hdata->res.sclk_pixel);
@@ -668,15 +1829,21 @@ static void hdmiphy_conf_reset(struct hdmi_context *hdata)
 	if (hdata->hdmiphy_port)
 		i2c_master_send(hdata->hdmiphy_port, buffer, 2);
 
+	if (hdata->is_v13)
+		reg = HDMI_V13_PHY_RSTOUT;
+	else
+		reg = HDMI_PHY_RSTOUT;
+
 	/* reset hdmiphy */
-	hdmi_reg_writemask(hdata, HDMI_PHY_RSTOUT, ~0, HDMI_PHY_SW_RSTOUT);
+	hdmi_reg_writemask(hdata, reg, ~0, HDMI_PHY_SW_RSTOUT);
 	mdelay(10);
-	hdmi_reg_writemask(hdata, HDMI_PHY_RSTOUT,  0, HDMI_PHY_SW_RSTOUT);
+	hdmi_reg_writemask(hdata, reg,  0, HDMI_PHY_SW_RSTOUT);
 	mdelay(10);
 }
 
 static void hdmiphy_conf_apply(struct hdmi_context *hdata)
 {
+	const u8 *hdmiphy_data;
 	u8 buffer[32];
 	u8 operation[2];
 	u8 read_buffer[32] = {0, };
@@ -689,7 +1856,12 @@ static void hdmiphy_conf_apply(struct hdmi_context *hdata)
 	}
 
 	/* pixel clock */
-	memcpy(buffer, hdmi_confs[hdata->cur_conf].hdmiphy_data, 32);
+	if (hdata->is_v13)
+		hdmiphy_data = hdmi_v13_confs[hdata->cur_conf].hdmiphy_data;
+	else
+		hdmiphy_data = hdmi_confs[hdata->cur_conf].hdmiphy_data;
+
+	memcpy(buffer, hdmiphy_data, 32);
 	ret = i2c_master_send(hdata->hdmiphy_port, buffer, 32);
 	if (ret != 32) {
 		DRM_ERROR("failed to configure HDMIPHY via I2C\n");
@@ -721,9 +1893,6 @@ static void hdmiphy_conf_apply(struct hdmi_context *hdata)
 
 static void hdmi_conf_apply(struct hdmi_context *hdata)
 {
-	const struct hdmi_preset_conf *conf =
-		  hdmi_confs[hdata->cur_conf].conf;
-
 	DRM_DEBUG_KMS("[%d] %s\n", __LINE__, __func__);
 
 	hdmiphy_conf_reset(hdata);
@@ -731,13 +1900,55 @@ static void hdmi_conf_apply(struct hdmi_context *hdata)
 
 	hdmi_conf_reset(hdata);
 	hdmi_conf_init(hdata);
+	hdmi_audio_init(hdata);
 
 	/* setting core registers */
-	hdmi_timing_apply(hdata, conf);
+	hdmi_timing_apply(hdata);
+	hdmi_audio_control(hdata, true);
 
 	hdmi_regs_dump(hdata, "start");
 }
 
+static void hdmi_mode_fixup(void *ctx, struct drm_connector *connector,
+				struct drm_display_mode *mode,
+				struct drm_display_mode *adjusted_mode)
+{
+	struct drm_display_mode *m;
+	struct hdmi_context *hdata = (struct hdmi_context *)ctx;
+	int index;
+
+	DRM_DEBUG_KMS("[%d] %s\n", __LINE__, __func__);
+
+	drm_mode_set_crtcinfo(adjusted_mode, 0);
+
+	if (hdata->is_v13)
+		index = hdmi_v13_conf_index(adjusted_mode);
+	else
+		index = hdmi_v14_conf_index(adjusted_mode);
+
+	/* just return if user desired mode exists. */
+	if (index >= 0)
+		return;
+
+	/*
+	 * otherwise, find the most suitable mode among modes and change it
+	 * to adjusted_mode.
+	 */
+	list_for_each_entry(m, &connector->modes, head) {
+		if (hdata->is_v13)
+			index = hdmi_v13_conf_index(m);
+		else
+			index = hdmi_v14_conf_index(m);
+
+		if (index >= 0) {
+			DRM_INFO("desired mode doesn't exist so\n");
+			DRM_INFO("use the most suitable mode among modes.\n");
+			memcpy(adjusted_mode, m, sizeof(*m));
+			break;
+		}
+	}
+}
+
 static void hdmi_mode_set(void *ctx, void *mode)
 {
 	struct hdmi_context *hdata = (struct hdmi_context *)ctx;
@@ -745,13 +1956,22 @@ static void hdmi_mode_set(void *ctx, void *mode)
 
 	DRM_DEBUG_KMS("[%d] %s\n", __LINE__, __func__);
 
-	conf_idx = hdmi_conf_index(mode);
-	if (conf_idx >= 0 && conf_idx < ARRAY_SIZE(hdmi_confs))
+	conf_idx = hdmi_conf_index(hdata, mode);
+	if (conf_idx >= 0)
 		hdata->cur_conf = conf_idx;
 	else
 		DRM_DEBUG_KMS("not supported mode\n");
 }
 
+static void hdmi_get_max_resol(void *ctx, unsigned int *width,
+					unsigned int *height)
+{
+	DRM_DEBUG_KMS("[%d] %s\n", __LINE__, __func__);
+
+	*width = MAX_WIDTH;
+	*height = MAX_HEIGHT;
+}
+
 static void hdmi_commit(void *ctx)
 {
 	struct hdmi_context *hdata = (struct hdmi_context *)ctx;
@@ -770,13 +1990,16 @@ static void hdmi_disable(void *ctx)
 	DRM_DEBUG_KMS("[%d] %s\n", __LINE__, __func__);
 
 	if (hdata->enabled) {
+		hdmi_audio_control(hdata, false);
 		hdmiphy_conf_reset(hdata);
 		hdmi_conf_reset(hdata);
 	}
 }
 
 static struct exynos_hdmi_manager_ops manager_ops = {
+	.mode_fixup	= hdmi_mode_fixup,
 	.mode_set	= hdmi_mode_set,
+	.get_max_resol	= hdmi_get_max_resol,
 	.commit		= hdmi_commit,
 	.disable	= hdmi_disable,
 };
@@ -926,7 +2149,7 @@ static void hdmi_resource_poweron(struct hdmi_context *hdata)
 	hdmiphy_conf_reset(hdata);
 	hdmi_conf_reset(hdata);
 	hdmi_conf_init(hdata);
-
+	hdmi_audio_init(hdata);
 }
 
 static void hdmi_resource_poweroff(struct hdmi_context *hdata)
@@ -978,14 +2201,12 @@ void hdmi_attach_ddc_client(struct i2c_client *ddc)
 	if (ddc)
 		hdmi_ddc = ddc;
 }
-EXPORT_SYMBOL(hdmi_attach_ddc_client);
 
 void hdmi_attach_hdmiphy_client(struct i2c_client *hdmiphy)
 {
 	if (hdmiphy)
 		hdmi_hdmiphy = hdmiphy;
 }
-EXPORT_SYMBOL(hdmi_attach_hdmiphy_client);
 
 static int __devinit hdmi_probe(struct platform_device *pdev)
 {
@@ -1022,6 +2243,7 @@ static int __devinit hdmi_probe(struct platform_device *pdev)
 
 	platform_set_drvdata(pdev, drm_hdmi_ctx);
 
+	hdata->is_v13 = pdata->is_v13;
 	hdata->default_win = pdata->default_win;
 	hdata->default_timing = &pdata->timing;
 	hdata->default_bpp = pdata->bpp;
@@ -1167,10 +2389,3 @@ struct platform_driver hdmi_driver = {
 		.pm = &hdmi_pm_ops,
 	},
 };
-EXPORT_SYMBOL(hdmi_driver);
-
-MODULE_AUTHOR("Seung-Woo Kim, <sw0312.kim@samsung.com>");
-MODULE_AUTHOR("Inki Dae <inki.dae@samsung.com>");
-MODULE_AUTHOR("Joonyoung Shim <jy0922.shim@samsung.com>");
-MODULE_DESCRIPTION("Samsung DRM HDMI core Driver");
-MODULE_LICENSE("GPL");
diff --git a/drivers/gpu/drm/exynos/exynos_hdmi.h b/drivers/gpu/drm/exynos/exynos_hdmi.h
index 31d6cf84c1aa..1c3b6d8f1fe7 100644
--- a/drivers/gpu/drm/exynos/exynos_hdmi.h
+++ b/drivers/gpu/drm/exynos/exynos_hdmi.h
@@ -28,56 +28,6 @@
 #ifndef _EXYNOS_HDMI_H_
 #define _EXYNOS_HDMI_H_
 
-struct hdmi_conf {
-	int width;
-	int height;
-	int vrefresh;
-	bool interlace;
-	const u8 *hdmiphy_data;
-	const struct hdmi_preset_conf *conf;
-};
-
-struct hdmi_resources {
-	struct clk *hdmi;
-	struct clk *sclk_hdmi;
-	struct clk *sclk_pixel;
-	struct clk *sclk_hdmiphy;
-	struct clk *hdmiphy;
-	struct regulator_bulk_data *regul_bulk;
-	int regul_count;
-};
-
-struct hdmi_context {
-	struct device			*dev;
-	struct drm_device		*drm_dev;
-	struct fb_videomode		*default_timing;
-	unsigned int			default_win;
-	unsigned int			default_bpp;
-	bool				hpd_handle;
-	bool				enabled;
-
-	struct resource			*regs_res;
-	/** base address of HDMI registers */
-	void __iomem *regs;
-	/** HDMI hotplug interrupt */
-	unsigned int irq;
-	/** workqueue for delayed work */
-	struct workqueue_struct *wq;
-	/** hotplug handling work */
-	struct work_struct hotplug_work;
-
-	struct i2c_client *ddc_port;
-	struct i2c_client *hdmiphy_port;
-
-	/** current hdmiphy conf index */
-	int cur_conf;
-	/** other resources */
-	struct hdmi_resources res;
-
-	void *parent_ctx;
-};
-
-
 void hdmi_attach_ddc_client(struct i2c_client *ddc);
 void hdmi_attach_hdmiphy_client(struct i2c_client *hdmiphy);
 
diff --git a/drivers/gpu/drm/exynos/exynos_mixer.c b/drivers/gpu/drm/exynos/exynos_mixer.c
index ac24cff39775..4d5f41e19527 100644
--- a/drivers/gpu/drm/exynos/exynos_mixer.c
+++ b/drivers/gpu/drm/exynos/exynos_mixer.c
@@ -36,11 +36,57 @@
 
 #include "exynos_drm_drv.h"
 #include "exynos_drm_hdmi.h"
-#include "exynos_hdmi.h"
-#include "exynos_mixer.h"
+
+#define HDMI_OVERLAY_NUMBER	3
 
 #define get_mixer_context(dev)	platform_get_drvdata(to_platform_device(dev))
 
+struct hdmi_win_data {
+	dma_addr_t		dma_addr;
+	void __iomem		*vaddr;
+	dma_addr_t		chroma_dma_addr;
+	void __iomem		*chroma_vaddr;
+	uint32_t		pixel_format;
+	unsigned int		bpp;
+	unsigned int		crtc_x;
+	unsigned int		crtc_y;
+	unsigned int		crtc_width;
+	unsigned int		crtc_height;
+	unsigned int		fb_x;
+	unsigned int		fb_y;
+	unsigned int		fb_width;
+	unsigned int		fb_height;
+	unsigned int		mode_width;
+	unsigned int		mode_height;
+	unsigned int		scan_flags;
+};
+
+struct mixer_resources {
+	struct device		*dev;
+	int			irq;
+	void __iomem		*mixer_regs;
+	void __iomem		*vp_regs;
+	spinlock_t		reg_slock;
+	struct clk		*mixer;
+	struct clk		*vp;
+	struct clk		*sclk_mixer;
+	struct clk		*sclk_hdmi;
+	struct clk		*sclk_dac;
+};
+
+struct mixer_context {
+	struct fb_videomode	*default_timing;
+	unsigned int		default_win;
+	unsigned int		default_bpp;
+	unsigned int		irq;
+	int			pipe;
+	bool			interlace;
+	bool			vp_enabled;
+
+	struct mixer_resources	mixer_res;
+	struct hdmi_win_data	win_data[HDMI_OVERLAY_NUMBER];
+};
+
 static const u8 filter_y_horiz_tap8[] = {
 	0,	-1,	-1,	-1,	-1,	-1,	-1,	-1,
 	-1,	-1,	-1,	-1,	-1,	0,	0,	0,
@@ -712,7 +758,12 @@ static void mixer_finish_pageflip(struct drm_device *drm_dev, int crtc)
 	}
 
 	if (is_checked)
-		drm_vblank_put(drm_dev, crtc);
+		/*
+		 * call drm_vblank_put only in case that drm_vblank_get was
+		 * called.
+		 */
+		if (atomic_read(&drm_dev->vblank_refcount[crtc]) > 0)
+			drm_vblank_put(drm_dev, crtc);
 
 	spin_unlock_irqrestore(&drm_dev->event_lock, flags);
 }
@@ -779,15 +830,15 @@ static void mixer_win_reset(struct mixer_context *ctx)
 	mixer_reg_writemask(res, MXR_STATUS, MXR_STATUS_16_BURST,
 		MXR_STATUS_BURST_MASK);
 
-	/* setting default layer priority: layer1 > video > layer0
+	/* setting default layer priority: layer1 > layer0 > video
 	 * because typical usage scenario would be
+	 * layer1 - OSD
 	 * layer0 - framebuffer
 	 * video - video overlay
-	 * layer1 - OSD
 	 */
-	val  = MXR_LAYER_CFG_GRP0_VAL(1);
-	val |= MXR_LAYER_CFG_VP_VAL(2);
-	val |= MXR_LAYER_CFG_GRP1_VAL(3);
+	val = MXR_LAYER_CFG_GRP1_VAL(3);
+	val |= MXR_LAYER_CFG_GRP0_VAL(2);
+	val |= MXR_LAYER_CFG_VP_VAL(1);
 	mixer_reg_write(res, MXR_LAYER_CFG, val);
 
 	/* setting background color */
@@ -1044,7 +1095,7 @@ static int mixer_remove(struct platform_device *pdev)
 					platform_get_drvdata(pdev);
 	struct mixer_context *ctx = (struct mixer_context *)drm_hdmi_ctx->ctx;
 
-	dev_info(dev, "remove sucessful\n");
+	dev_info(dev, "remove successful\n");
 
 	mixer_resource_poweroff(ctx);
 	mixer_resources_cleanup(ctx);
@@ -1061,10 +1112,3 @@ struct platform_driver mixer_driver = {
 	.probe = mixer_probe,
 	.remove = __devexit_p(mixer_remove),
 };
-EXPORT_SYMBOL(mixer_driver);
-
-MODULE_AUTHOR("Seung-Woo Kim, <sw0312.kim@samsung.com>");
-MODULE_AUTHOR("Inki Dae <inki.dae@samsung.com>");
-MODULE_AUTHOR("Joonyoung Shim <jy0922.shim@samsung.com>");
-MODULE_DESCRIPTION("Samsung DRM HDMI mixer Driver");
-MODULE_LICENSE("GPL");
diff --git a/drivers/gpu/drm/exynos/exynos_mixer.h b/drivers/gpu/drm/exynos/exynos_mixer.h
deleted file mode 100644
index cebacfefc077..000000000000
--- a/drivers/gpu/drm/exynos/exynos_mixer.h
+++ /dev/null
@@ -1,92 +0,0 @@
-/*
- *
- * Copyright (c) 2011 Samsung Electronics Co., Ltd.
- * Authors:
- *	Seung-Woo Kim <sw0312.kim@samsung.com>
- *	Inki Dae <inki.dae@samsung.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
- * 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.
- */
-
-#ifndef _EXYNOS_MIXER_H_
-#define _EXYNOS_MIXER_H_
-
-#define HDMI_OVERLAY_NUMBER	3
-
-struct hdmi_win_data {
-	dma_addr_t		dma_addr;
-	void __iomem		*vaddr;
-	dma_addr_t		chroma_dma_addr;
-	void __iomem		*chroma_vaddr;
-	uint32_t		pixel_format;
-	unsigned int		bpp;
-	unsigned int		crtc_x;
-	unsigned int		crtc_y;
-	unsigned int		crtc_width;
-	unsigned int		crtc_height;
-	unsigned int		fb_x;
-	unsigned int		fb_y;
-	unsigned int		fb_width;
-	unsigned int		fb_height;
-	unsigned int		mode_width;
-	unsigned int		mode_height;
-	unsigned int		scan_flags;
-};
-
-struct mixer_resources {
-	struct device *dev;
-	/** interrupt index */
-	int irq;
-	/** pointer to Mixer registers */
-	void __iomem *mixer_regs;
-	/** pointer to Video Processor registers */
-	void __iomem *vp_regs;
-	/** spinlock for protection of registers */
-	spinlock_t reg_slock;
-	/** other resources */
-	struct clk *mixer;
-	struct clk *vp;
-	struct clk *sclk_mixer;
-	struct clk *sclk_hdmi;
-	struct clk *sclk_dac;
-};
-
-struct mixer_context {
-	unsigned int			default_win;
-	struct fb_videomode		*default_timing;
-	unsigned int			default_bpp;
-
-	/** mixer interrupt */
-	unsigned int irq;
-	/** current crtc pipe for vblank */
-	int pipe;
-	/** interlace scan mode */
-	bool interlace;
-	/** vp enabled status */
-	bool vp_enabled;
-
-	/** mixer and vp resources */
-	struct mixer_resources mixer_res;
-
-	/** overlay window data */
-	struct hdmi_win_data		win_data[HDMI_OVERLAY_NUMBER];
-};
-
-#endif
diff --git a/drivers/gpu/drm/exynos/regs-hdmi.h b/drivers/gpu/drm/exynos/regs-hdmi.h
index 72e6b52be740..3c04bea842ce 100644
--- a/drivers/gpu/drm/exynos/regs-hdmi.h
+++ b/drivers/gpu/drm/exynos/regs-hdmi.h
@@ -19,64 +19,67 @@
  * Register part
 */
 
+/* HDMI Version 1.3 & Common */
 #define HDMI_CTRL_BASE(x)		((x) + 0x00000000)
 #define HDMI_CORE_BASE(x)		((x) + 0x00010000)
+#define HDMI_I2S_BASE(x)		((x) + 0x00040000)
 #define HDMI_TG_BASE(x)			((x) + 0x00050000)
 
 /* Control registers */
 #define HDMI_INTC_CON			HDMI_CTRL_BASE(0x0000)
 #define HDMI_INTC_FLAG			HDMI_CTRL_BASE(0x0004)
 #define HDMI_HPD_STATUS			HDMI_CTRL_BASE(0x000C)
-#define HDMI_PHY_RSTOUT			HDMI_CTRL_BASE(0x0014)
-#define HDMI_PHY_VPLL			HDMI_CTRL_BASE(0x0018)
-#define HDMI_PHY_CMU			HDMI_CTRL_BASE(0x001C)
-#define HDMI_CORE_RSTOUT		HDMI_CTRL_BASE(0x0020)
+#define HDMI_V13_PHY_RSTOUT		HDMI_CTRL_BASE(0x0014)
+#define HDMI_V13_PHY_VPLL		HDMI_CTRL_BASE(0x0018)
+#define HDMI_V13_PHY_CMU		HDMI_CTRL_BASE(0x001C)
+#define HDMI_V13_CORE_RSTOUT		HDMI_CTRL_BASE(0x0020)
 
 /* Core registers */
 #define HDMI_CON_0			HDMI_CORE_BASE(0x0000)
 #define HDMI_CON_1			HDMI_CORE_BASE(0x0004)
 #define HDMI_CON_2			HDMI_CORE_BASE(0x0008)
 #define HDMI_SYS_STATUS			HDMI_CORE_BASE(0x0010)
-#define HDMI_PHY_STATUS			HDMI_CORE_BASE(0x0014)
+#define HDMI_V13_PHY_STATUS		HDMI_CORE_BASE(0x0014)
 #define HDMI_STATUS_EN			HDMI_CORE_BASE(0x0020)
 #define HDMI_HPD			HDMI_CORE_BASE(0x0030)
 #define HDMI_MODE_SEL			HDMI_CORE_BASE(0x0040)
-#define HDMI_BLUE_SCREEN_0		HDMI_CORE_BASE(0x0050)
-#define HDMI_BLUE_SCREEN_1		HDMI_CORE_BASE(0x0054)
-#define HDMI_BLUE_SCREEN_2		HDMI_CORE_BASE(0x0058)
+#define HDMI_ENC_EN			HDMI_CORE_BASE(0x0044)
+#define HDMI_V13_BLUE_SCREEN_0		HDMI_CORE_BASE(0x0050)
+#define HDMI_V13_BLUE_SCREEN_1		HDMI_CORE_BASE(0x0054)
+#define HDMI_V13_BLUE_SCREEN_2		HDMI_CORE_BASE(0x0058)
 #define HDMI_H_BLANK_0			HDMI_CORE_BASE(0x00A0)
 #define HDMI_H_BLANK_1			HDMI_CORE_BASE(0x00A4)
-#define HDMI_V_BLANK_0			HDMI_CORE_BASE(0x00B0)
-#define HDMI_V_BLANK_1			HDMI_CORE_BASE(0x00B4)
-#define HDMI_V_BLANK_2			HDMI_CORE_BASE(0x00B8)
-#define HDMI_H_V_LINE_0			HDMI_CORE_BASE(0x00C0)
-#define HDMI_H_V_LINE_1			HDMI_CORE_BASE(0x00C4)
-#define HDMI_H_V_LINE_2			HDMI_CORE_BASE(0x00C8)
+#define HDMI_V13_V_BLANK_0		HDMI_CORE_BASE(0x00B0)
+#define HDMI_V13_V_BLANK_1		HDMI_CORE_BASE(0x00B4)
+#define HDMI_V13_V_BLANK_2		HDMI_CORE_BASE(0x00B8)
+#define HDMI_V13_H_V_LINE_0		HDMI_CORE_BASE(0x00C0)
+#define HDMI_V13_H_V_LINE_1		HDMI_CORE_BASE(0x00C4)
+#define HDMI_V13_H_V_LINE_2		HDMI_CORE_BASE(0x00C8)
 #define HDMI_VSYNC_POL			HDMI_CORE_BASE(0x00E4)
 #define HDMI_INT_PRO_MODE		HDMI_CORE_BASE(0x00E8)
-#define HDMI_V_BLANK_F_0		HDMI_CORE_BASE(0x0110)
-#define HDMI_V_BLANK_F_1		HDMI_CORE_BASE(0x0114)
-#define HDMI_V_BLANK_F_2		HDMI_CORE_BASE(0x0118)
-#define HDMI_H_SYNC_GEN_0		HDMI_CORE_BASE(0x0120)
-#define HDMI_H_SYNC_GEN_1		HDMI_CORE_BASE(0x0124)
-#define HDMI_H_SYNC_GEN_2		HDMI_CORE_BASE(0x0128)
-#define HDMI_V_SYNC_GEN_1_0		HDMI_CORE_BASE(0x0130)
-#define HDMI_V_SYNC_GEN_1_1		HDMI_CORE_BASE(0x0134)
-#define HDMI_V_SYNC_GEN_1_2		HDMI_CORE_BASE(0x0138)
-#define HDMI_V_SYNC_GEN_2_0		HDMI_CORE_BASE(0x0140)
-#define HDMI_V_SYNC_GEN_2_1		HDMI_CORE_BASE(0x0144)
-#define HDMI_V_SYNC_GEN_2_2		HDMI_CORE_BASE(0x0148)
-#define HDMI_V_SYNC_GEN_3_0		HDMI_CORE_BASE(0x0150)
-#define HDMI_V_SYNC_GEN_3_1		HDMI_CORE_BASE(0x0154)
-#define HDMI_V_SYNC_GEN_3_2		HDMI_CORE_BASE(0x0158)
-#define HDMI_ACR_CON			HDMI_CORE_BASE(0x0180)
-#define HDMI_AVI_CON			HDMI_CORE_BASE(0x0300)
-#define HDMI_AVI_BYTE(n)		HDMI_CORE_BASE(0x0320 + 4 * (n))
-#define HDMI_DC_CONTROL			HDMI_CORE_BASE(0x05C0)
-#define HDMI_VIDEO_PATTERN_GEN		HDMI_CORE_BASE(0x05C4)
-#define HDMI_HPD_GEN			HDMI_CORE_BASE(0x05C8)
-#define HDMI_AUI_CON			HDMI_CORE_BASE(0x0360)
-#define HDMI_SPD_CON			HDMI_CORE_BASE(0x0400)
+#define HDMI_V13_V_BLANK_F_0		HDMI_CORE_BASE(0x0110)
+#define HDMI_V13_V_BLANK_F_1		HDMI_CORE_BASE(0x0114)
+#define HDMI_V13_V_BLANK_F_2		HDMI_CORE_BASE(0x0118)
+#define HDMI_V13_H_SYNC_GEN_0		HDMI_CORE_BASE(0x0120)
+#define HDMI_V13_H_SYNC_GEN_1		HDMI_CORE_BASE(0x0124)
+#define HDMI_V13_H_SYNC_GEN_2		HDMI_CORE_BASE(0x0128)
+#define HDMI_V13_V_SYNC_GEN_1_0		HDMI_CORE_BASE(0x0130)
+#define HDMI_V13_V_SYNC_GEN_1_1		HDMI_CORE_BASE(0x0134)
+#define HDMI_V13_V_SYNC_GEN_1_2		HDMI_CORE_BASE(0x0138)
+#define HDMI_V13_V_SYNC_GEN_2_0		HDMI_CORE_BASE(0x0140)
+#define HDMI_V13_V_SYNC_GEN_2_1		HDMI_CORE_BASE(0x0144)
+#define HDMI_V13_V_SYNC_GEN_2_2		HDMI_CORE_BASE(0x0148)
+#define HDMI_V13_V_SYNC_GEN_3_0		HDMI_CORE_BASE(0x0150)
+#define HDMI_V13_V_SYNC_GEN_3_1		HDMI_CORE_BASE(0x0154)
+#define HDMI_V13_V_SYNC_GEN_3_2		HDMI_CORE_BASE(0x0158)
+#define HDMI_V13_ACR_CON		HDMI_CORE_BASE(0x0180)
+#define HDMI_V13_AVI_CON		HDMI_CORE_BASE(0x0300)
+#define HDMI_V13_AVI_BYTE(n)		HDMI_CORE_BASE(0x0320 + 4 * (n))
+#define HDMI_V13_DC_CONTROL		HDMI_CORE_BASE(0x05C0)
+#define HDMI_V13_VIDEO_PATTERN_GEN	HDMI_CORE_BASE(0x05C4)
+#define HDMI_V13_HPD_GEN		HDMI_CORE_BASE(0x05C8)
+#define HDMI_V13_AUI_CON		HDMI_CORE_BASE(0x0360)
+#define HDMI_V13_SPD_CON		HDMI_CORE_BASE(0x0400)
 
 /* Timing generator registers */
 #define HDMI_TG_CMD			HDMI_TG_BASE(0x0000)
@@ -130,6 +133,9 @@
 
 /* HDMI_CON_0 */
 #define HDMI_BLUE_SCR_EN		(1 << 5)
+#define HDMI_ASP_EN			(1 << 2)
+#define HDMI_ASP_DIS			(0 << 2)
+#define HDMI_ASP_MASK			(1 << 2)
 #define HDMI_EN				(1 << 0)
 
 /* HDMI_PHY_STATUS */
@@ -138,10 +144,418 @@
 /* HDMI_MODE_SEL */
 #define HDMI_MODE_HDMI_EN		(1 << 1)
 #define HDMI_MODE_DVI_EN		(1 << 0)
+#define HDMI_DVI_MODE_EN		(1)
+#define HDMI_DVI_MODE_DIS		(0)
 #define HDMI_MODE_MASK			(3 << 0)
 
 /* HDMI_TG_CMD */
 #define HDMI_TG_EN			(1 << 0)
 #define HDMI_FIELD_EN			(1 << 1)
 
+
+/* HDMI Version 1.4 */
+/* Control registers */
+/* #define HDMI_INTC_CON		HDMI_CTRL_BASE(0x0000) */
+/* #define HDMI_INTC_FLAG		HDMI_CTRL_BASE(0x0004) */
+#define HDMI_HDCP_KEY_LOAD		HDMI_CTRL_BASE(0x0008)
+/* #define HDMI_HPD_STATUS		HDMI_CTRL_BASE(0x000C) */
+#define HDMI_INTC_CON_1			HDMI_CTRL_BASE(0x0010)
+#define HDMI_INTC_FLAG_1		HDMI_CTRL_BASE(0x0014)
+#define HDMI_PHY_STATUS_0		HDMI_CTRL_BASE(0x0020)
+#define HDMI_PHY_STATUS_CMU		HDMI_CTRL_BASE(0x0024)
+#define HDMI_PHY_STATUS_PLL		HDMI_CTRL_BASE(0x0028)
+#define HDMI_PHY_CON_0			HDMI_CTRL_BASE(0x0030)
+#define HDMI_HPD_CTRL			HDMI_CTRL_BASE(0x0040)
+#define HDMI_HPD_ST			HDMI_CTRL_BASE(0x0044)
+#define HDMI_HPD_TH_X			HDMI_CTRL_BASE(0x0050)
+#define HDMI_AUDIO_CLKSEL		HDMI_CTRL_BASE(0x0070)
+#define HDMI_PHY_RSTOUT			HDMI_CTRL_BASE(0x0074)
+#define HDMI_PHY_VPLL			HDMI_CTRL_BASE(0x0078)
+#define HDMI_PHY_CMU			HDMI_CTRL_BASE(0x007C)
+#define HDMI_CORE_RSTOUT		HDMI_CTRL_BASE(0x0080)
+
+/* Video related registers */
+#define HDMI_YMAX			HDMI_CORE_BASE(0x0060)
+#define HDMI_YMIN			HDMI_CORE_BASE(0x0064)
+#define HDMI_CMAX			HDMI_CORE_BASE(0x0068)
+#define HDMI_CMIN			HDMI_CORE_BASE(0x006C)
+
+#define HDMI_V2_BLANK_0			HDMI_CORE_BASE(0x00B0)
+#define HDMI_V2_BLANK_1			HDMI_CORE_BASE(0x00B4)
+#define HDMI_V1_BLANK_0			HDMI_CORE_BASE(0x00B8)
+#define HDMI_V1_BLANK_1			HDMI_CORE_BASE(0x00BC)
+
+#define HDMI_V_LINE_0			HDMI_CORE_BASE(0x00C0)
+#define HDMI_V_LINE_1			HDMI_CORE_BASE(0x00C4)
+#define HDMI_H_LINE_0			HDMI_CORE_BASE(0x00C8)
+#define HDMI_H_LINE_1			HDMI_CORE_BASE(0x00CC)
+
+#define HDMI_HSYNC_POL			HDMI_CORE_BASE(0x00E0)
+
+#define HDMI_V_BLANK_F0_0		HDMI_CORE_BASE(0x0110)
+#define HDMI_V_BLANK_F0_1		HDMI_CORE_BASE(0x0114)
+#define HDMI_V_BLANK_F1_0		HDMI_CORE_BASE(0x0118)
+#define HDMI_V_BLANK_F1_1		HDMI_CORE_BASE(0x011C)
+
+#define HDMI_H_SYNC_START_0		HDMI_CORE_BASE(0x0120)
+#define HDMI_H_SYNC_START_1		HDMI_CORE_BASE(0x0124)
+#define HDMI_H_SYNC_END_0		HDMI_CORE_BASE(0x0128)
+#define HDMI_H_SYNC_END_1		HDMI_CORE_BASE(0x012C)
+
+#define HDMI_V_SYNC_LINE_BEF_2_0	HDMI_CORE_BASE(0x0130)
+#define HDMI_V_SYNC_LINE_BEF_2_1	HDMI_CORE_BASE(0x0134)
+#define HDMI_V_SYNC_LINE_BEF_1_0	HDMI_CORE_BASE(0x0138)
+#define HDMI_V_SYNC_LINE_BEF_1_1	HDMI_CORE_BASE(0x013C)
+
+#define HDMI_V_SYNC_LINE_AFT_2_0	HDMI_CORE_BASE(0x0140)
+#define HDMI_V_SYNC_LINE_AFT_2_1	HDMI_CORE_BASE(0x0144)
+#define HDMI_V_SYNC_LINE_AFT_1_0	HDMI_CORE_BASE(0x0148)
+#define HDMI_V_SYNC_LINE_AFT_1_1	HDMI_CORE_BASE(0x014C)
+
+#define HDMI_V_SYNC_LINE_AFT_PXL_2_0	HDMI_CORE_BASE(0x0150)
+#define HDMI_V_SYNC_LINE_AFT_PXL_2_1	HDMI_CORE_BASE(0x0154)
+#define HDMI_V_SYNC_LINE_AFT_PXL_1_0	HDMI_CORE_BASE(0x0158)
+#define HDMI_V_SYNC_LINE_AFT_PXL_1_1	HDMI_CORE_BASE(0x015C)
+
+#define HDMI_V_BLANK_F2_0		HDMI_CORE_BASE(0x0160)
+#define HDMI_V_BLANK_F2_1		HDMI_CORE_BASE(0x0164)
+#define HDMI_V_BLANK_F3_0		HDMI_CORE_BASE(0x0168)
+#define HDMI_V_BLANK_F3_1		HDMI_CORE_BASE(0x016C)
+#define HDMI_V_BLANK_F4_0		HDMI_CORE_BASE(0x0170)
+#define HDMI_V_BLANK_F4_1		HDMI_CORE_BASE(0x0174)
+#define HDMI_V_BLANK_F5_0		HDMI_CORE_BASE(0x0178)
+#define HDMI_V_BLANK_F5_1		HDMI_CORE_BASE(0x017C)
+
+#define HDMI_V_SYNC_LINE_AFT_3_0	HDMI_CORE_BASE(0x0180)
+#define HDMI_V_SYNC_LINE_AFT_3_1	HDMI_CORE_BASE(0x0184)
+#define HDMI_V_SYNC_LINE_AFT_4_0	HDMI_CORE_BASE(0x0188)
+#define HDMI_V_SYNC_LINE_AFT_4_1	HDMI_CORE_BASE(0x018C)
+#define HDMI_V_SYNC_LINE_AFT_5_0	HDMI_CORE_BASE(0x0190)
+#define HDMI_V_SYNC_LINE_AFT_5_1	HDMI_CORE_BASE(0x0194)
+#define HDMI_V_SYNC_LINE_AFT_6_0	HDMI_CORE_BASE(0x0198)
+#define HDMI_V_SYNC_LINE_AFT_6_1	HDMI_CORE_BASE(0x019C)
+
+#define HDMI_V_SYNC_LINE_AFT_PXL_3_0	HDMI_CORE_BASE(0x01A0)
+#define HDMI_V_SYNC_LINE_AFT_PXL_3_1	HDMI_CORE_BASE(0x01A4)
+#define HDMI_V_SYNC_LINE_AFT_PXL_4_0	HDMI_CORE_BASE(0x01A8)
+#define HDMI_V_SYNC_LINE_AFT_PXL_4_1	HDMI_CORE_BASE(0x01AC)
+#define HDMI_V_SYNC_LINE_AFT_PXL_5_0	HDMI_CORE_BASE(0x01B0)
+#define HDMI_V_SYNC_LINE_AFT_PXL_5_1	HDMI_CORE_BASE(0x01B4)
+#define HDMI_V_SYNC_LINE_AFT_PXL_6_0	HDMI_CORE_BASE(0x01B8)
+#define HDMI_V_SYNC_LINE_AFT_PXL_6_1	HDMI_CORE_BASE(0x01BC)
+
+#define HDMI_VACT_SPACE_1_0		HDMI_CORE_BASE(0x01C0)
+#define HDMI_VACT_SPACE_1_1		HDMI_CORE_BASE(0x01C4)
+#define HDMI_VACT_SPACE_2_0		HDMI_CORE_BASE(0x01C8)
+#define HDMI_VACT_SPACE_2_1		HDMI_CORE_BASE(0x01CC)
+#define HDMI_VACT_SPACE_3_0		HDMI_CORE_BASE(0x01D0)
+#define HDMI_VACT_SPACE_3_1		HDMI_CORE_BASE(0x01D4)
+#define HDMI_VACT_SPACE_4_0		HDMI_CORE_BASE(0x01D8)
+#define HDMI_VACT_SPACE_4_1		HDMI_CORE_BASE(0x01DC)
+#define HDMI_VACT_SPACE_5_0		HDMI_CORE_BASE(0x01E0)
+#define HDMI_VACT_SPACE_5_1		HDMI_CORE_BASE(0x01E4)
+#define HDMI_VACT_SPACE_6_0		HDMI_CORE_BASE(0x01E8)
+#define HDMI_VACT_SPACE_6_1		HDMI_CORE_BASE(0x01EC)
+
+#define HDMI_GCP_CON			HDMI_CORE_BASE(0x0200)
+#define HDMI_GCP_BYTE1			HDMI_CORE_BASE(0x0210)
+#define HDMI_GCP_BYTE2			HDMI_CORE_BASE(0x0214)
+#define HDMI_GCP_BYTE3			HDMI_CORE_BASE(0x0218)
+
+/* Audio related registers */
+#define HDMI_ASP_CON			HDMI_CORE_BASE(0x0300)
+#define HDMI_ASP_SP_FLAT		HDMI_CORE_BASE(0x0304)
+#define HDMI_ASP_CHCFG0			HDMI_CORE_BASE(0x0310)
+#define HDMI_ASP_CHCFG1			HDMI_CORE_BASE(0x0314)
+#define HDMI_ASP_CHCFG2			HDMI_CORE_BASE(0x0318)
+#define HDMI_ASP_CHCFG3			HDMI_CORE_BASE(0x031C)
+
+#define HDMI_ACR_CON			HDMI_CORE_BASE(0x0400)
+#define HDMI_ACR_MCTS0			HDMI_CORE_BASE(0x0410)
+#define HDMI_ACR_MCTS1			HDMI_CORE_BASE(0x0414)
+#define HDMI_ACR_MCTS2			HDMI_CORE_BASE(0x0418)
+#define HDMI_ACR_CTS0			HDMI_CORE_BASE(0x0420)
+#define HDMI_ACR_CTS1			HDMI_CORE_BASE(0x0424)
+#define HDMI_ACR_CTS2			HDMI_CORE_BASE(0x0428)
+#define HDMI_ACR_N0			HDMI_CORE_BASE(0x0430)
+#define HDMI_ACR_N1			HDMI_CORE_BASE(0x0434)
+#define HDMI_ACR_N2			HDMI_CORE_BASE(0x0438)
+
+/* Packet related registers */
+#define HDMI_ACP_CON			HDMI_CORE_BASE(0x0500)
+#define HDMI_ACP_TYPE			HDMI_CORE_BASE(0x0514)
+#define HDMI_ACP_DATA(n)		HDMI_CORE_BASE(0x0520 + 4 * (n))
+
+#define HDMI_ISRC_CON			HDMI_CORE_BASE(0x0600)
+#define HDMI_ISRC1_HEADER1		HDMI_CORE_BASE(0x0614)
+#define HDMI_ISRC1_DATA(n)		HDMI_CORE_BASE(0x0620 + 4 * (n))
+#define HDMI_ISRC2_DATA(n)		HDMI_CORE_BASE(0x06A0 + 4 * (n))
+
+#define HDMI_AVI_CON			HDMI_CORE_BASE(0x0700)
+#define HDMI_AVI_HEADER0		HDMI_CORE_BASE(0x0710)
+#define HDMI_AVI_HEADER1		HDMI_CORE_BASE(0x0714)
+#define HDMI_AVI_HEADER2		HDMI_CORE_BASE(0x0718)
+#define HDMI_AVI_CHECK_SUM		HDMI_CORE_BASE(0x071C)
+#define HDMI_AVI_BYTE(n)		HDMI_CORE_BASE(0x0720 + 4 * (n))
+
+#define HDMI_AUI_CON			HDMI_CORE_BASE(0x0800)
+#define HDMI_AUI_HEADER0		HDMI_CORE_BASE(0x0810)
+#define HDMI_AUI_HEADER1		HDMI_CORE_BASE(0x0814)
+#define HDMI_AUI_HEADER2		HDMI_CORE_BASE(0x0818)
+#define HDMI_AUI_CHECK_SUM		HDMI_CORE_BASE(0x081C)
+#define HDMI_AUI_BYTE(n)		HDMI_CORE_BASE(0x0820 + 4 * (n))
+
+#define HDMI_MPG_CON			HDMI_CORE_BASE(0x0900)
+#define HDMI_MPG_CHECK_SUM		HDMI_CORE_BASE(0x091C)
+#define HDMI_MPG_DATA(n)		HDMI_CORE_BASE(0x0920 + 4 * (n))
+
+#define HDMI_SPD_CON			HDMI_CORE_BASE(0x0A00)
+#define HDMI_SPD_HEADER0		HDMI_CORE_BASE(0x0A10)
+#define HDMI_SPD_HEADER1		HDMI_CORE_BASE(0x0A14)
+#define HDMI_SPD_HEADER2		HDMI_CORE_BASE(0x0A18)
+#define HDMI_SPD_DATA(n)		HDMI_CORE_BASE(0x0A20 + 4 * (n))
+
+#define HDMI_GAMUT_CON			HDMI_CORE_BASE(0x0B00)
+#define HDMI_GAMUT_HEADER0		HDMI_CORE_BASE(0x0B10)
+#define HDMI_GAMUT_HEADER1		HDMI_CORE_BASE(0x0B14)
+#define HDMI_GAMUT_HEADER2		HDMI_CORE_BASE(0x0B18)
+#define HDMI_GAMUT_METADATA(n)		HDMI_CORE_BASE(0x0B20 + 4 * (n))
+
+#define HDMI_VSI_CON			HDMI_CORE_BASE(0x0C00)
+#define HDMI_VSI_HEADER0		HDMI_CORE_BASE(0x0C10)
+#define HDMI_VSI_HEADER1		HDMI_CORE_BASE(0x0C14)
+#define HDMI_VSI_HEADER2		HDMI_CORE_BASE(0x0C18)
+#define HDMI_VSI_DATA(n)		HDMI_CORE_BASE(0x0C20 + 4 * (n))
+
+#define HDMI_DC_CONTROL			HDMI_CORE_BASE(0x0D00)
+#define HDMI_VIDEO_PATTERN_GEN		HDMI_CORE_BASE(0x0D04)
+
+#define HDMI_AN_SEED_SEL		HDMI_CORE_BASE(0x0E48)
+#define HDMI_AN_SEED_0			HDMI_CORE_BASE(0x0E58)
+#define HDMI_AN_SEED_1			HDMI_CORE_BASE(0x0E5C)
+#define HDMI_AN_SEED_2			HDMI_CORE_BASE(0x0E60)
+#define HDMI_AN_SEED_3			HDMI_CORE_BASE(0x0E64)
+
+/* HDCP related registers */
+#define HDMI_HDCP_SHA1(n)		HDMI_CORE_BASE(0x7000 + 4 * (n))
+#define HDMI_HDCP_KSV_LIST(n)		HDMI_CORE_BASE(0x7050 + 4 * (n))
+
+#define HDMI_HDCP_KSV_LIST_CON		HDMI_CORE_BASE(0x7064)
+#define HDMI_HDCP_SHA_RESULT		HDMI_CORE_BASE(0x7070)
+#define HDMI_HDCP_CTRL1			HDMI_CORE_BASE(0x7080)
+#define HDMI_HDCP_CTRL2			HDMI_CORE_BASE(0x7084)
+#define HDMI_HDCP_CHECK_RESULT		HDMI_CORE_BASE(0x7090)
+#define HDMI_HDCP_BKSV(n)		HDMI_CORE_BASE(0x70A0 + 4 * (n))
+#define HDMI_HDCP_AKSV(n)		HDMI_CORE_BASE(0x70C0 + 4 * (n))
+#define HDMI_HDCP_AN(n)			HDMI_CORE_BASE(0x70E0 + 4 * (n))
+
+#define HDMI_HDCP_BCAPS			HDMI_CORE_BASE(0x7100)
+#define HDMI_HDCP_BSTATUS_0		HDMI_CORE_BASE(0x7110)
+#define HDMI_HDCP_BSTATUS_1		HDMI_CORE_BASE(0x7114)
+#define HDMI_HDCP_RI_0			HDMI_CORE_BASE(0x7140)
+#define HDMI_HDCP_RI_1			HDMI_CORE_BASE(0x7144)
+#define HDMI_HDCP_I2C_INT		HDMI_CORE_BASE(0x7180)
+#define HDMI_HDCP_AN_INT		HDMI_CORE_BASE(0x7190)
+#define HDMI_HDCP_WDT_INT		HDMI_CORE_BASE(0x71A0)
+#define HDMI_HDCP_RI_INT		HDMI_CORE_BASE(0x71B0)
+#define HDMI_HDCP_RI_COMPARE_0		HDMI_CORE_BASE(0x71D0)
+#define HDMI_HDCP_RI_COMPARE_1		HDMI_CORE_BASE(0x71D4)
+#define HDMI_HDCP_FRAME_COUNT		HDMI_CORE_BASE(0x71E0)
+
+#define HDMI_RGB_ROUND_EN		HDMI_CORE_BASE(0xD500)
+#define HDMI_VACT_SPACE_R_0		HDMI_CORE_BASE(0xD504)
+#define HDMI_VACT_SPACE_R_1		HDMI_CORE_BASE(0xD508)
+#define HDMI_VACT_SPACE_G_0		HDMI_CORE_BASE(0xD50C)
+#define HDMI_VACT_SPACE_G_1		HDMI_CORE_BASE(0xD510)
+#define HDMI_VACT_SPACE_B_0		HDMI_CORE_BASE(0xD514)
+#define HDMI_VACT_SPACE_B_1		HDMI_CORE_BASE(0xD518)
+
+#define HDMI_BLUE_SCREEN_B_0		HDMI_CORE_BASE(0xD520)
+#define HDMI_BLUE_SCREEN_B_1		HDMI_CORE_BASE(0xD524)
+#define HDMI_BLUE_SCREEN_G_0		HDMI_CORE_BASE(0xD528)
+#define HDMI_BLUE_SCREEN_G_1		HDMI_CORE_BASE(0xD52C)
+#define HDMI_BLUE_SCREEN_R_0		HDMI_CORE_BASE(0xD530)
+#define HDMI_BLUE_SCREEN_R_1		HDMI_CORE_BASE(0xD534)
+
+/* HDMI I2S register */
+#define HDMI_I2S_CLK_CON		HDMI_I2S_BASE(0x000)
+#define HDMI_I2S_CON_1			HDMI_I2S_BASE(0x004)
+#define HDMI_I2S_CON_2			HDMI_I2S_BASE(0x008)
+#define HDMI_I2S_PIN_SEL_0		HDMI_I2S_BASE(0x00c)
+#define HDMI_I2S_PIN_SEL_1		HDMI_I2S_BASE(0x010)
+#define HDMI_I2S_PIN_SEL_2		HDMI_I2S_BASE(0x014)
+#define HDMI_I2S_PIN_SEL_3		HDMI_I2S_BASE(0x018)
+#define HDMI_I2S_DSD_CON		HDMI_I2S_BASE(0x01c)
+#define HDMI_I2S_MUX_CON		HDMI_I2S_BASE(0x020)
+#define HDMI_I2S_CH_ST_CON		HDMI_I2S_BASE(0x024)
+#define HDMI_I2S_CH_ST_0		HDMI_I2S_BASE(0x028)
+#define HDMI_I2S_CH_ST_1		HDMI_I2S_BASE(0x02c)
+#define HDMI_I2S_CH_ST_2		HDMI_I2S_BASE(0x030)
+#define HDMI_I2S_CH_ST_3		HDMI_I2S_BASE(0x034)
+#define HDMI_I2S_CH_ST_4		HDMI_I2S_BASE(0x038)
+#define HDMI_I2S_CH_ST_SH_0		HDMI_I2S_BASE(0x03c)
+#define HDMI_I2S_CH_ST_SH_1		HDMI_I2S_BASE(0x040)
+#define HDMI_I2S_CH_ST_SH_2		HDMI_I2S_BASE(0x044)
+#define HDMI_I2S_CH_ST_SH_3		HDMI_I2S_BASE(0x048)
+#define HDMI_I2S_CH_ST_SH_4		HDMI_I2S_BASE(0x04c)
+#define HDMI_I2S_MUX_CH			HDMI_I2S_BASE(0x054)
+#define HDMI_I2S_MUX_CUV		HDMI_I2S_BASE(0x058)
+
+/* I2S bit definition */
+
+/* I2S_CLK_CON */
+#define HDMI_I2S_CLK_DIS		(0)
+#define HDMI_I2S_CLK_EN			(1)
+
+/* I2S_CON_1 */
+#define HDMI_I2S_SCLK_FALLING_EDGE	(0 << 1)
+#define HDMI_I2S_SCLK_RISING_EDGE	(1 << 1)
+#define HDMI_I2S_L_CH_LOW_POL		(0)
+#define HDMI_I2S_L_CH_HIGH_POL		(1)
+
+/* I2S_CON_2 */
+#define HDMI_I2S_MSB_FIRST_MODE		(0 << 6)
+#define HDMI_I2S_LSB_FIRST_MODE		(1 << 6)
+#define HDMI_I2S_BIT_CH_32FS		(0 << 4)
+#define HDMI_I2S_BIT_CH_48FS		(1 << 4)
+#define HDMI_I2S_BIT_CH_RESERVED	(2 << 4)
+#define HDMI_I2S_SDATA_16BIT		(1 << 2)
+#define HDMI_I2S_SDATA_20BIT		(2 << 2)
+#define HDMI_I2S_SDATA_24BIT		(3 << 2)
+#define HDMI_I2S_BASIC_FORMAT		(0)
+#define HDMI_I2S_L_JUST_FORMAT		(2)
+#define HDMI_I2S_R_JUST_FORMAT		(3)
+#define HDMI_I2S_CON_2_CLR		(~(0xFF))
+#define HDMI_I2S_SET_BIT_CH(x)		(((x) & 0x7) << 4)
+#define HDMI_I2S_SET_SDATA_BIT(x)	(((x) & 0x7) << 2)
+
+/* I2S_PIN_SEL_0 */
+#define HDMI_I2S_SEL_SCLK(x)		(((x) & 0x7) << 4)
+#define HDMI_I2S_SEL_LRCK(x)		((x) & 0x7)
+
+/* I2S_PIN_SEL_1 */
+#define HDMI_I2S_SEL_SDATA1(x)		(((x) & 0x7) << 4)
+#define HDMI_I2S_SEL_SDATA2(x)		((x) & 0x7)
+
+/* I2S_PIN_SEL_2 */
+#define HDMI_I2S_SEL_SDATA3(x)		(((x) & 0x7) << 4)
+#define HDMI_I2S_SEL_SDATA2(x)		((x) & 0x7)
+
+/* I2S_PIN_SEL_3 */
+#define HDMI_I2S_SEL_DSD(x)		((x) & 0x7)
+
+/* I2S_DSD_CON */
+#define HDMI_I2S_DSD_CLK_RI_EDGE	(1 << 1)
+#define HDMI_I2S_DSD_CLK_FA_EDGE	(0 << 1)
+#define HDMI_I2S_DSD_ENABLE		(1)
+#define HDMI_I2S_DSD_DISABLE		(0)
+
+/* I2S_MUX_CON */
+#define HDMI_I2S_NOISE_FILTER_ZERO	(0 << 5)
+#define HDMI_I2S_NOISE_FILTER_2_STAGE	(1 << 5)
+#define HDMI_I2S_NOISE_FILTER_3_STAGE	(2 << 5)
+#define HDMI_I2S_NOISE_FILTER_4_STAGE	(3 << 5)
+#define HDMI_I2S_NOISE_FILTER_5_STAGE	(4 << 5)
+#define HDMI_I2S_IN_DISABLE		(1 << 4)
+#define HDMI_I2S_IN_ENABLE		(0 << 4)
+#define HDMI_I2S_AUD_SPDIF		(0 << 2)
+#define HDMI_I2S_AUD_I2S		(1 << 2)
+#define HDMI_I2S_AUD_DSD		(2 << 2)
+#define HDMI_I2S_CUV_SPDIF_ENABLE	(0 << 1)
+#define HDMI_I2S_CUV_I2S_ENABLE		(1 << 1)
+#define HDMI_I2S_MUX_DISABLE		(0)
+#define HDMI_I2S_MUX_ENABLE		(1)
+#define HDMI_I2S_MUX_CON_CLR		(~(0xFF))
+
+/* I2S_CH_ST_CON */
+#define HDMI_I2S_CH_STATUS_RELOAD	(1)
+#define HDMI_I2S_CH_ST_CON_CLR		(~(1))
+
+/* I2S_CH_ST_0 / I2S_CH_ST_SH_0 */
+#define HDMI_I2S_CH_STATUS_MODE_0	(0 << 6)
+#define HDMI_I2S_2AUD_CH_WITHOUT_PREEMPH	(0 << 3)
+#define HDMI_I2S_2AUD_CH_WITH_PREEMPH	(1 << 3)
+#define HDMI_I2S_DEFAULT_EMPHASIS	(0 << 3)
+#define HDMI_I2S_COPYRIGHT		(0 << 2)
+#define HDMI_I2S_NO_COPYRIGHT		(1 << 2)
+#define HDMI_I2S_LINEAR_PCM		(0 << 1)
+#define HDMI_I2S_NO_LINEAR_PCM		(1 << 1)
+#define HDMI_I2S_CONSUMER_FORMAT	(0)
+#define HDMI_I2S_PROF_FORMAT		(1)
+#define HDMI_I2S_CH_ST_0_CLR		(~(0xFF))
+
+/* I2S_CH_ST_1 / I2S_CH_ST_SH_1 */
+#define HDMI_I2S_CD_PLAYER		(0x00)
+#define HDMI_I2S_DAT_PLAYER		(0x03)
+#define HDMI_I2S_DCC_PLAYER		(0x43)
+#define HDMI_I2S_MINI_DISC_PLAYER	(0x49)
+
+/* I2S_CH_ST_2 / I2S_CH_ST_SH_2 */
+#define HDMI_I2S_CHANNEL_NUM_MASK	(0xF << 4)
+#define HDMI_I2S_SOURCE_NUM_MASK	(0xF)
+#define HDMI_I2S_SET_CHANNEL_NUM(x)	(((x) & (0xF)) << 4)
+#define HDMI_I2S_SET_SOURCE_NUM(x)	((x) & (0xF))
+
+/* I2S_CH_ST_3 / I2S_CH_ST_SH_3 */
+#define HDMI_I2S_CLK_ACCUR_LEVEL_1	(1 << 4)
+#define HDMI_I2S_CLK_ACCUR_LEVEL_2	(0 << 4)
+#define HDMI_I2S_CLK_ACCUR_LEVEL_3	(2 << 4)
+#define HDMI_I2S_SMP_FREQ_44_1		(0x0)
+#define HDMI_I2S_SMP_FREQ_48		(0x2)
+#define HDMI_I2S_SMP_FREQ_32		(0x3)
+#define HDMI_I2S_SMP_FREQ_96		(0xA)
+#define HDMI_I2S_SET_SMP_FREQ(x)	((x) & (0xF))
+
+/* I2S_CH_ST_4 / I2S_CH_ST_SH_4 */
+#define HDMI_I2S_ORG_SMP_FREQ_44_1	(0xF << 4)
+#define HDMI_I2S_ORG_SMP_FREQ_88_2	(0x7 << 4)
+#define HDMI_I2S_ORG_SMP_FREQ_22_05	(0xB << 4)
+#define HDMI_I2S_ORG_SMP_FREQ_176_4	(0x3 << 4)
+#define HDMI_I2S_WORD_LEN_NOT_DEFINE	(0x0 << 1)
+#define HDMI_I2S_WORD_LEN_MAX24_20BITS	(0x1 << 1)
+#define HDMI_I2S_WORD_LEN_MAX24_22BITS	(0x2 << 1)
+#define HDMI_I2S_WORD_LEN_MAX24_23BITS	(0x4 << 1)
+#define HDMI_I2S_WORD_LEN_MAX24_24BITS	(0x5 << 1)
+#define HDMI_I2S_WORD_LEN_MAX24_21BITS	(0x6 << 1)
+#define HDMI_I2S_WORD_LEN_MAX20_16BITS	(0x1 << 1)
+#define HDMI_I2S_WORD_LEN_MAX20_18BITS	(0x2 << 1)
+#define HDMI_I2S_WORD_LEN_MAX20_19BITS	(0x4 << 1)
+#define HDMI_I2S_WORD_LEN_MAX20_20BITS	(0x5 << 1)
+#define HDMI_I2S_WORD_LEN_MAX20_17BITS	(0x6 << 1)
+#define HDMI_I2S_WORD_LEN_MAX_24BITS	(1)
+#define HDMI_I2S_WORD_LEN_MAX_20BITS	(0)
+
+/* I2S_MUX_CH */
+#define HDMI_I2S_CH3_R_EN		(1 << 7)
+#define HDMI_I2S_CH3_L_EN		(1 << 6)
+#define HDMI_I2S_CH3_EN			(3 << 6)
+#define HDMI_I2S_CH2_R_EN		(1 << 5)
+#define HDMI_I2S_CH2_L_EN		(1 << 4)
+#define HDMI_I2S_CH2_EN			(3 << 4)
+#define HDMI_I2S_CH1_R_EN		(1 << 3)
+#define HDMI_I2S_CH1_L_EN		(1 << 2)
+#define HDMI_I2S_CH1_EN			(3 << 2)
+#define HDMI_I2S_CH0_R_EN		(1 << 1)
+#define HDMI_I2S_CH0_L_EN		(1)
+#define HDMI_I2S_CH0_EN			(3)
+#define HDMI_I2S_CH_ALL_EN		(0xFF)
+#define HDMI_I2S_MUX_CH_CLR		(~HDMI_I2S_CH_ALL_EN)
+
+/* I2S_MUX_CUV */
+#define HDMI_I2S_CUV_R_EN		(1 << 1)
+#define HDMI_I2S_CUV_L_EN		(1)
+#define HDMI_I2S_CUV_RL_EN		(0x03)
+
+/* I2S_CUV_L_R */
+#define HDMI_I2S_CUV_R_DATA_MASK	(0x7 << 4)
+#define HDMI_I2S_CUV_L_DATA_MASK	(0x7)
+
+/* Timing generator registers */
+/* TG configure/status registers */
+#define HDMI_TG_VACT_ST3_L		HDMI_TG_BASE(0x0068)
+#define HDMI_TG_VACT_ST3_H		HDMI_TG_BASE(0x006c)
+#define HDMI_TG_VACT_ST4_L		HDMI_TG_BASE(0x0070)
+#define HDMI_TG_VACT_ST4_H		HDMI_TG_BASE(0x0074)
+#define HDMI_TG_3D			HDMI_TG_BASE(0x00F0)
+
 #endif /* SAMSUNG_REGS_HDMI_H */
diff --git a/drivers/gpu/drm/gma500/Kconfig b/drivers/gpu/drm/gma500/Kconfig
index 754e14bdc801..42e665c7e90a 100644
--- a/drivers/gpu/drm/gma500/Kconfig
+++ b/drivers/gpu/drm/gma500/Kconfig
@@ -16,8 +16,7 @@ config DRM_GMA600
 	depends on DRM_GMA500
 	help
 	  Say yes to include support for GMA600 (Intel Moorestown/Oaktrail)
-	  platforms with LVDS ports. HDMI and MIPI are not currently
-	  supported.
+	  platforms with LVDS ports. MIPI is not currently supported.
 
 config DRM_GMA3600
 	bool "Intel GMA3600/3650 support (Experimental)"
@@ -25,3 +24,10 @@ config DRM_GMA3600
 	help
 	  Say yes to include basic support for Intel GMA3600/3650 (Intel
 	  Cedar Trail) platforms.
+
+config DRM_MEDFIELD
+	bool "Intel Medfield support (Experimental)"
+	depends on DRM_GMA500 && X86_INTEL_MID
+	help
+	  Say yes to include support for the Intel Medfield platform.
+
diff --git a/drivers/gpu/drm/gma500/Makefile b/drivers/gpu/drm/gma500/Makefile
index 81c103be5e21..1583982917ce 100644
--- a/drivers/gpu/drm/gma500/Makefile
+++ b/drivers/gpu/drm/gma500/Makefile
@@ -37,4 +37,14 @@ gma500_gfx-$(CONFIG_DRM_GMA600) += oaktrail_device.o \
 	  oaktrail_hdmi.o \
 	  oaktrail_hdmi_i2c.o
 
+gma500_gfx-$(CONFIG_DRM_MEDFIELD) += mdfld_device.o \
+	  mdfld_output.o \
+	  mdfld_intel_display.o \
+	  mdfld_dsi_output.o \
+	  mdfld_dsi_dpi.o \
+	  mdfld_dsi_pkg_sender.o \
+	  mdfld_tpo_vid.o \
+	  mdfld_tmd_vid.o \
+	  tc35876x-dsi-lvds.o
+
 obj-$(CONFIG_DRM_GMA500) += gma500_gfx.o
diff --git a/drivers/gpu/drm/gma500/cdv_device.c b/drivers/gpu/drm/gma500/cdv_device.c
index 4a5b099c3bc5..a54cc738926a 100644
--- a/drivers/gpu/drm/gma500/cdv_device.c
+++ b/drivers/gpu/drm/gma500/cdv_device.c
@@ -202,13 +202,12 @@ static inline void CDV_MSG_WRITE32(uint port, uint offset, u32 value)
 	pci_dev_put(pci_root);
 }
 
-#define PSB_APM_CMD			0x0
-#define PSB_APM_STS			0x04
 #define PSB_PM_SSC			0x20
 #define PSB_PM_SSS			0x30
-#define PSB_PWRGT_GFX_MASK		0x3
-#define CDV_PWRGT_DISPLAY_CNTR		0x000fc00c
-#define CDV_PWRGT_DISPLAY_STS		0x000fc00c
+#define PSB_PWRGT_GFX_ON		0x02
+#define PSB_PWRGT_GFX_OFF		0x01
+#define PSB_PWRGT_GFX_D0		0x00
+#define PSB_PWRGT_GFX_D3		0x03
 
 static void cdv_init_pm(struct drm_device *dev)
 {
@@ -221,26 +220,22 @@ static void cdv_init_pm(struct drm_device *dev)
 	dev_priv->ospm_base = CDV_MSG_READ32(PSB_PUNIT_PORT,
 							PSB_OSPMBA) & 0xFFFF;
 
-	/* Force power on for now */
+	/* Power status */
 	pwr_cnt = inl(dev_priv->apm_base + PSB_APM_CMD);
-	pwr_cnt &= ~PSB_PWRGT_GFX_MASK;
 
+	/* Enable the GPU */
+	pwr_cnt &= ~PSB_PWRGT_GFX_MASK;
+	pwr_cnt |= PSB_PWRGT_GFX_ON;
 	outl(pwr_cnt, dev_priv->apm_base + PSB_APM_CMD);
+
+	/* Wait for the GPU power */
 	for (i = 0; i < 5; i++) {
 		u32 pwr_sts = inl(dev_priv->apm_base + PSB_APM_STS);
 		if ((pwr_sts & PSB_PWRGT_GFX_MASK) == 0)
-			break;
-		udelay(10);
-	}
-	pwr_cnt = inl(dev_priv->ospm_base + PSB_PM_SSC);
-	pwr_cnt &= ~CDV_PWRGT_DISPLAY_CNTR;
-	outl(pwr_cnt, dev_priv->ospm_base + PSB_PM_SSC);
-	for (i = 0; i < 5; i++) {
-		u32 pwr_sts = inl(dev_priv->ospm_base + PSB_PM_SSS);
-		if ((pwr_sts & CDV_PWRGT_DISPLAY_STS) == 0)
-			break;
+			return;
 		udelay(10);
 	}
+	dev_err(dev->dev, "GPU: power management timed out.\n");
 }
 
 /**
@@ -249,11 +244,50 @@ static void cdv_init_pm(struct drm_device *dev)
  *
  *	Save the state we need in order to be able to restore the interface
  *	upon resume from suspend
- *
- *	FIXME: review
  */
 static int cdv_save_display_registers(struct drm_device *dev)
 {
+	struct drm_psb_private *dev_priv = dev->dev_private;
+	struct psb_save_area *regs = &dev_priv->regs;
+	struct drm_connector *connector;
+
+	dev_info(dev->dev, "Saving GPU registers.\n");
+
+	pci_read_config_byte(dev->pdev, 0xF4, &regs->cdv.saveLBB);
+
+	regs->cdv.saveDSPCLK_GATE_D = REG_READ(DSPCLK_GATE_D);
+	regs->cdv.saveRAMCLK_GATE_D = REG_READ(RAMCLK_GATE_D);
+
+	regs->cdv.saveDSPARB = REG_READ(DSPARB);
+	regs->cdv.saveDSPFW[0] = REG_READ(DSPFW1);
+	regs->cdv.saveDSPFW[1] = REG_READ(DSPFW2);
+	regs->cdv.saveDSPFW[2] = REG_READ(DSPFW3);
+	regs->cdv.saveDSPFW[3] = REG_READ(DSPFW4);
+	regs->cdv.saveDSPFW[4] = REG_READ(DSPFW5);
+	regs->cdv.saveDSPFW[5] = REG_READ(DSPFW6);
+
+	regs->cdv.saveADPA = REG_READ(ADPA);
+
+	regs->cdv.savePP_CONTROL = REG_READ(PP_CONTROL);
+	regs->cdv.savePFIT_PGM_RATIOS = REG_READ(PFIT_PGM_RATIOS);
+	regs->saveBLC_PWM_CTL = REG_READ(BLC_PWM_CTL);
+	regs->saveBLC_PWM_CTL2 = REG_READ(BLC_PWM_CTL2);
+	regs->cdv.saveLVDS = REG_READ(LVDS);
+
+	regs->cdv.savePFIT_CONTROL = REG_READ(PFIT_CONTROL);
+
+	regs->cdv.savePP_ON_DELAYS = REG_READ(PP_ON_DELAYS);
+	regs->cdv.savePP_OFF_DELAYS = REG_READ(PP_OFF_DELAYS);
+	regs->cdv.savePP_CYCLE = REG_READ(PP_CYCLE);
+
+	regs->cdv.saveVGACNTRL = REG_READ(VGACNTRL);
+
+	regs->cdv.saveIER = REG_READ(PSB_INT_ENABLE_R);
+	regs->cdv.saveIMR = REG_READ(PSB_INT_MASK_R);
+
+	list_for_each_entry(connector, &dev->mode_config.connector_list, head)
+		connector->funcs->dpms(connector, DRM_MODE_DPMS_OFF);
+
 	return 0;
 }
 
@@ -267,16 +301,113 @@ static int cdv_save_display_registers(struct drm_device *dev)
  */
 static int cdv_restore_display_registers(struct drm_device *dev)
 {
+	struct drm_psb_private *dev_priv = dev->dev_private;
+	struct psb_save_area *regs = &dev_priv->regs;
+	struct drm_connector *connector;
+	u32 temp;
+
+	pci_write_config_byte(dev->pdev, 0xF4, regs->cdv.saveLBB);
+
+	REG_WRITE(DSPCLK_GATE_D, regs->cdv.saveDSPCLK_GATE_D);
+	REG_WRITE(RAMCLK_GATE_D, regs->cdv.saveRAMCLK_GATE_D);
+
+	/* BIOS does below anyway */
+	REG_WRITE(DPIO_CFG, 0);
+	REG_WRITE(DPIO_CFG, DPIO_MODE_SELECT_0 | DPIO_CMN_RESET_N);
+
+	temp = REG_READ(DPLL_A);
+	if ((temp & DPLL_SYNCLOCK_ENABLE) == 0) {
+		REG_WRITE(DPLL_A, temp | DPLL_SYNCLOCK_ENABLE);
+		REG_READ(DPLL_A);
+	}
+
+	temp = REG_READ(DPLL_B);
+	if ((temp & DPLL_SYNCLOCK_ENABLE) == 0) {
+		REG_WRITE(DPLL_B, temp | DPLL_SYNCLOCK_ENABLE);
+		REG_READ(DPLL_B);
+	}
+
+	udelay(500);
+
+	REG_WRITE(DSPFW1, regs->cdv.saveDSPFW[0]);
+	REG_WRITE(DSPFW2, regs->cdv.saveDSPFW[1]);
+	REG_WRITE(DSPFW3, regs->cdv.saveDSPFW[2]);
+	REG_WRITE(DSPFW4, regs->cdv.saveDSPFW[3]);
+	REG_WRITE(DSPFW5, regs->cdv.saveDSPFW[4]);
+	REG_WRITE(DSPFW6, regs->cdv.saveDSPFW[5]);
+
+	REG_WRITE(DSPARB, regs->cdv.saveDSPARB);
+	REG_WRITE(ADPA, regs->cdv.saveADPA);
+
+	REG_WRITE(BLC_PWM_CTL2, regs->saveBLC_PWM_CTL2);
+	REG_WRITE(LVDS, regs->cdv.saveLVDS);
+	REG_WRITE(PFIT_CONTROL, regs->cdv.savePFIT_CONTROL);
+	REG_WRITE(PFIT_PGM_RATIOS, regs->cdv.savePFIT_PGM_RATIOS);
+	REG_WRITE(BLC_PWM_CTL, regs->saveBLC_PWM_CTL);
+	REG_WRITE(PP_ON_DELAYS, regs->cdv.savePP_ON_DELAYS);
+	REG_WRITE(PP_OFF_DELAYS, regs->cdv.savePP_OFF_DELAYS);
+	REG_WRITE(PP_CYCLE, regs->cdv.savePP_CYCLE);
+	REG_WRITE(PP_CONTROL, regs->cdv.savePP_CONTROL);
+
+	REG_WRITE(VGACNTRL, regs->cdv.saveVGACNTRL);
+
+	REG_WRITE(PSB_INT_ENABLE_R, regs->cdv.saveIER);
+	REG_WRITE(PSB_INT_MASK_R, regs->cdv.saveIMR);
+
+	/* Fix arbitration bug */
+	CDV_MSG_WRITE32(3, 0x30, 0x08027108);
+
+	drm_mode_config_reset(dev);
+
+	list_for_each_entry(connector, &dev->mode_config.connector_list, head)
+		connector->funcs->dpms(connector, DRM_MODE_DPMS_ON);
+
+	/* Resume the modeset for every activated CRTC */
+	drm_helper_resume_force_mode(dev);
 	return 0;
 }
 
 static int cdv_power_down(struct drm_device *dev)
 {
+	struct drm_psb_private *dev_priv = dev->dev_private;
+	u32 pwr_cnt, pwr_mask, pwr_sts;
+	int tries = 5;
+
+	pwr_cnt = inl(dev_priv->apm_base + PSB_APM_CMD);
+	pwr_cnt &= ~PSB_PWRGT_GFX_MASK;
+	pwr_cnt |= PSB_PWRGT_GFX_OFF;
+	pwr_mask = PSB_PWRGT_GFX_MASK;
+
+	outl(pwr_cnt, dev_priv->apm_base + PSB_APM_CMD);
+
+	while (tries--) {
+		pwr_sts = inl(dev_priv->apm_base + PSB_APM_STS);
+		if ((pwr_sts & pwr_mask) == PSB_PWRGT_GFX_D3)
+			return 0;
+		udelay(10);
+	}
 	return 0;
 }
 
 static int cdv_power_up(struct drm_device *dev)
 {
+	struct drm_psb_private *dev_priv = dev->dev_private;
+	u32 pwr_cnt, pwr_mask, pwr_sts;
+	int tries = 5;
+
+	pwr_cnt = inl(dev_priv->apm_base + PSB_APM_CMD);
+	pwr_cnt &= ~PSB_PWRGT_GFX_MASK;
+	pwr_cnt |= PSB_PWRGT_GFX_ON;
+	pwr_mask = PSB_PWRGT_GFX_MASK;
+
+	outl(pwr_cnt, dev_priv->apm_base + PSB_APM_CMD);
+
+	while (tries--) {
+		pwr_sts = inl(dev_priv->apm_base + PSB_APM_STS);
+		if ((pwr_sts & pwr_mask) == PSB_PWRGT_GFX_D0)
+			return 0;
+		udelay(10);
+	}
 	return 0;
 }
 
@@ -321,6 +452,8 @@ static int cdv_chip_setup(struct drm_device *dev)
 	cdv_get_core_freq(dev);
 	gma_intel_opregion_init(dev);
 	psb_intel_init_bios(dev);
+	REG_WRITE(PORT_HOTPLUG_EN, 0);
+	REG_WRITE(PORT_HOTPLUG_STAT, REG_READ(PORT_HOTPLUG_STAT));
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/gma500/cdv_device.h b/drivers/gpu/drm/gma500/cdv_device.h
index 2a88b7beb551..9561e17621b3 100644
--- a/drivers/gpu/drm/gma500/cdv_device.h
+++ b/drivers/gpu/drm/gma500/cdv_device.h
@@ -26,7 +26,7 @@ extern void cdv_hdmi_init(struct drm_device *dev, struct psb_intel_mode_device *
 extern struct drm_display_mode *cdv_intel_crtc_mode_get(struct drm_device *dev,
 					     struct drm_crtc *crtc);
 
-extern inline void cdv_intel_wait_for_vblank(struct drm_device *dev)
+static inline void cdv_intel_wait_for_vblank(struct drm_device *dev)
 {
 	/* Wait for 20ms, i.e. one cycle at 50hz. */
         /* FIXME: msleep ?? */
diff --git a/drivers/gpu/drm/gma500/cdv_intel_crt.c b/drivers/gpu/drm/gma500/cdv_intel_crt.c
index c100f3e9c920..a71a6cd95bdd 100644
--- a/drivers/gpu/drm/gma500/cdv_intel_crt.c
+++ b/drivers/gpu/drm/gma500/cdv_intel_crt.c
@@ -32,6 +32,7 @@
 #include "psb_intel_drv.h"
 #include "psb_intel_reg.h"
 #include "power.h"
+#include "cdv_device.h"
 #include <linux/pm_runtime.h>
 
 
diff --git a/drivers/gpu/drm/gma500/cdv_intel_display.c b/drivers/gpu/drm/gma500/cdv_intel_display.c
index 18d11525095e..be8455919b33 100644
--- a/drivers/gpu/drm/gma500/cdv_intel_display.c
+++ b/drivers/gpu/drm/gma500/cdv_intel_display.c
@@ -344,7 +344,7 @@ cdv_dpll_set_clock_cdv(struct drm_device *dev, struct drm_crtc *crtc,
 /*
  * Returns whether any encoder on the specified pipe is of the specified type
  */
-bool cdv_intel_pipe_has_type(struct drm_crtc *crtc, int type)
+static bool cdv_intel_pipe_has_type(struct drm_crtc *crtc, int type)
 {
 	struct drm_device *dev = crtc->dev;
 	struct drm_mode_config *mode_config = &dev->mode_config;
@@ -476,7 +476,7 @@ static bool cdv_intel_find_best_PLL(struct drm_crtc *crtc, int target,
 	return err != target;
 }
 
-int cdv_intel_pipe_set_base(struct drm_crtc *crtc,
+static int cdv_intel_pipe_set_base(struct drm_crtc *crtc,
 			    int x, int y, struct drm_framebuffer *old_fb)
 {
 	struct drm_device *dev = crtc->dev;
@@ -569,7 +569,6 @@ static void cdv_intel_crtc_dpms(struct drm_crtc *crtc, int mode)
 	int dspbase_reg = (pipe == 0) ? DSPABASE : DSPBBASE;
 	int pipeconf_reg = (pipe == 0) ? PIPEACONF : PIPEBCONF;
 	u32 temp;
-	bool enabled;
 
 	/* XXX: When our outputs are all unaware of DPMS modes other than off
 	 * and on, we should map those modes to DRM_MODE_DPMS_OFF in the CRTC.
@@ -663,7 +662,6 @@ static void cdv_intel_crtc_dpms(struct drm_crtc *crtc, int mode)
 		udelay(150);
 		break;
 	}
-	enabled = crtc->enabled && mode != DRM_MODE_DPMS_OFF;
 	/*Set FIFO Watermarks*/
 	REG_WRITE(DSPARB, 0x3F3E);
 }
@@ -680,22 +678,6 @@ static void cdv_intel_crtc_commit(struct drm_crtc *crtc)
 	crtc_funcs->dpms(crtc, DRM_MODE_DPMS_ON);
 }
 
-void cdv_intel_encoder_prepare(struct drm_encoder *encoder)
-{
-	struct drm_encoder_helper_funcs *encoder_funcs =
-	    encoder->helper_private;
-	/* lvds has its own version of prepare see cdv_intel_lvds_prepare */
-	encoder_funcs->dpms(encoder, DRM_MODE_DPMS_OFF);
-}
-
-void cdv_intel_encoder_commit(struct drm_encoder *encoder)
-{
-	struct drm_encoder_helper_funcs *encoder_funcs =
-	    encoder->helper_private;
-	/* lvds has its own version of commit see cdv_intel_lvds_commit */
-	encoder_funcs->dpms(encoder, DRM_MODE_DPMS_ON);
-}
-
 static bool cdv_intel_crtc_mode_fixup(struct drm_crtc *crtc,
 				  struct drm_display_mode *mode,
 				  struct drm_display_mode *adjusted_mode)
@@ -745,7 +727,7 @@ static int cdv_intel_crtc_mode_set(struct drm_crtc *crtc,
 	int refclk;
 	struct cdv_intel_clock_t clock;
 	u32 dpll = 0, dspcntr, pipeconf;
-	bool ok, is_sdvo = false, is_dvo = false;
+	bool ok;
 	bool is_crt = false, is_lvds = false, is_tv = false;
 	bool is_hdmi = false;
 	struct drm_mode_config *mode_config = &dev->mode_config;
@@ -763,12 +745,6 @@ static int cdv_intel_crtc_mode_set(struct drm_crtc *crtc,
 		case INTEL_OUTPUT_LVDS:
 			is_lvds = true;
 			break;
-		case INTEL_OUTPUT_SDVO:
-			is_sdvo = true;
-			break;
-		case INTEL_OUTPUT_DVO:
-			is_dvo = true;
-			break;
 		case INTEL_OUTPUT_TVOUT:
 			is_tv = true;
 			break;
@@ -928,7 +904,7 @@ static int cdv_intel_crtc_mode_set(struct drm_crtc *crtc,
 }
 
 /** Loads the palette/gamma unit for the CRTC with the prepared values */
-void cdv_intel_crtc_load_lut(struct drm_crtc *crtc)
+static void cdv_intel_crtc_load_lut(struct drm_crtc *crtc)
 {
 	struct drm_device *dev = crtc->dev;
 	struct drm_psb_private *dev_priv =
@@ -968,7 +944,7 @@ void cdv_intel_crtc_load_lut(struct drm_crtc *crtc)
 		gma_power_end(dev);
 	} else {
 		for (i = 0; i < 256; i++) {
-			dev_priv->save_palette_a[i] =
+			dev_priv->regs.psb.save_palette_a[i] =
 				  ((psb_intel_crtc->lut_r[i] +
 				  psb_intel_crtc->lut_adj[i]) << 16) |
 				  ((psb_intel_crtc->lut_g[i] +
@@ -1338,18 +1314,20 @@ static int cdv_intel_crtc_clock_get(struct drm_device *dev,
 		gma_power_end(dev);
 	} else {
 		dpll = (pipe == 0) ?
-			dev_priv->saveDPLL_A : dev_priv->saveDPLL_B;
+			dev_priv->regs.psb.saveDPLL_A :
+			dev_priv->regs.psb.saveDPLL_B;
 
 		if ((dpll & DISPLAY_RATE_SELECT_FPA1) == 0)
 			fp = (pipe == 0) ?
-				dev_priv->saveFPA0 :
-				dev_priv->saveFPB0;
+				dev_priv->regs.psb.saveFPA0 :
+				dev_priv->regs.psb.saveFPB0;
 		else
 			fp = (pipe == 0) ?
-				dev_priv->saveFPA1 :
-				dev_priv->saveFPB1;
+				dev_priv->regs.psb.saveFPA1 :
+				dev_priv->regs.psb.saveFPB1;
 
-		is_lvds = (pipe == 1) && (dev_priv->saveLVDS & LVDS_PORT_EN);
+		is_lvds = (pipe == 1) &&
+				(dev_priv->regs.psb.saveLVDS & LVDS_PORT_EN);
 	}
 
 	clock.m1 = (fp & FP_M1_DIV_MASK) >> FP_M1_DIV_SHIFT;
@@ -1419,13 +1397,17 @@ struct drm_display_mode *cdv_intel_crtc_mode_get(struct drm_device *dev,
 		gma_power_end(dev);
 	} else {
 		htot = (pipe == 0) ?
-			dev_priv->saveHTOTAL_A : dev_priv->saveHTOTAL_B;
+			dev_priv->regs.psb.saveHTOTAL_A :
+			dev_priv->regs.psb.saveHTOTAL_B;
 		hsync = (pipe == 0) ?
-			dev_priv->saveHSYNC_A : dev_priv->saveHSYNC_B;
+			dev_priv->regs.psb.saveHSYNC_A :
+			dev_priv->regs.psb.saveHSYNC_B;
 		vtot = (pipe == 0) ?
-			dev_priv->saveVTOTAL_A : dev_priv->saveVTOTAL_B;
+			dev_priv->regs.psb.saveVTOTAL_A :
+			dev_priv->regs.psb.saveVTOTAL_B;
 		vsync = (pipe == 0) ?
-			dev_priv->saveVSYNC_A : dev_priv->saveVSYNC_B;
+			dev_priv->regs.psb.saveVSYNC_A :
+			dev_priv->regs.psb.saveVSYNC_B;
 	}
 
 	mode = kzalloc(sizeof(*mode), GFP_KERNEL);
@@ -1475,34 +1457,3 @@ const struct drm_crtc_funcs cdv_intel_crtc_funcs = {
 	.set_config = cdv_crtc_set_config,
 	.destroy = cdv_intel_crtc_destroy,
 };
-
-/*
- * Set the default value of cursor control and base register
- * to zero. This is a workaround for h/w defect on oaktrail
- */
-void cdv_intel_cursor_init(struct drm_device *dev, int pipe)
-{
-	uint32_t control;
-	uint32_t base;
-
-	switch (pipe) {
-	case 0:
-		control = CURACNTR;
-		base = CURABASE;
-		break;
-	case 1:
-		control = CURBCNTR;
-		base = CURBBASE;
-		break;
-	case 2:
-		control = CURCCNTR;
-		base = CURCBASE;
-		break;
-	default:
-		return;
-	}
-
-	REG_WRITE(control, 0);
-	REG_WRITE(base, 0);
-}
-
diff --git a/drivers/gpu/drm/gma500/cdv_intel_hdmi.c b/drivers/gpu/drm/gma500/cdv_intel_hdmi.c
index de25560e629d..8d5269555005 100644
--- a/drivers/gpu/drm/gma500/cdv_intel_hdmi.c
+++ b/drivers/gpu/drm/gma500/cdv_intel_hdmi.c
@@ -34,6 +34,7 @@
 #include "psb_intel_drv.h"
 #include "psb_drv.h"
 #include "psb_intel_reg.h"
+#include "cdv_device.h"
 #include <linux/pm_runtime.h>
 
 /* hdmi control bits */
diff --git a/drivers/gpu/drm/gma500/cdv_intel_lvds.c b/drivers/gpu/drm/gma500/cdv_intel_lvds.c
index 50e744be9852..8359c1a3f45f 100644
--- a/drivers/gpu/drm/gma500/cdv_intel_lvds.c
+++ b/drivers/gpu/drm/gma500/cdv_intel_lvds.c
@@ -78,13 +78,14 @@ static u32 cdv_intel_lvds_get_max_backlight(struct drm_device *dev)
 
 		gma_power_end(dev);
 	} else
-		retval = ((dev_priv->saveBLC_PWM_CTL &
+		retval = ((dev_priv->regs.saveBLC_PWM_CTL &
 			  BACKLIGHT_MODULATION_FREQ_MASK) >>
 			  BACKLIGHT_MODULATION_FREQ_SHIFT) * 2;
 
 	return retval;
 }
 
+#if 0
 /*
  * Set LVDS backlight level by I2C command
  */
@@ -165,6 +166,7 @@ void cdv_intel_lvds_set_brightness(struct drm_device *dev, int level)
 	else
 		cdv_lvds_pwm_set_brightness(dev, level);
 }
+#endif
 
 /**
  * Sets the backlight level.
@@ -184,9 +186,9 @@ static void cdv_intel_lvds_set_backlight(struct drm_device *dev, int level)
 				(level << BACKLIGHT_DUTY_CYCLE_SHIFT)));
 		gma_power_end(dev);
 	} else {
-		blc_pwm_ctl = dev_priv->saveBLC_PWM_CTL &
+		blc_pwm_ctl = dev_priv->regs.saveBLC_PWM_CTL &
 				~BACKLIGHT_DUTY_CYCLE_MASK;
-		dev_priv->saveBLC_PWM_CTL = (blc_pwm_ctl |
+		dev_priv->regs.saveBLC_PWM_CTL = (blc_pwm_ctl |
 					(level << BACKLIGHT_DUTY_CYCLE_SHIFT));
 	}
 }
@@ -242,7 +244,7 @@ static void cdv_intel_lvds_restore(struct drm_connector *connector)
 {
 }
 
-int cdv_intel_lvds_mode_valid(struct drm_connector *connector,
+static int cdv_intel_lvds_mode_valid(struct drm_connector *connector,
 			      struct drm_display_mode *mode)
 {
 	struct drm_device *dev = connector->dev;
@@ -267,7 +269,7 @@ int cdv_intel_lvds_mode_valid(struct drm_connector *connector,
 	return MODE_OK;
 }
 
-bool cdv_intel_lvds_mode_fixup(struct drm_encoder *encoder,
+static bool cdv_intel_lvds_mode_fixup(struct drm_encoder *encoder,
 				  struct drm_display_mode *mode,
 				  struct drm_display_mode *adjusted_mode)
 {
@@ -436,7 +438,7 @@ static int cdv_intel_lvds_get_modes(struct drm_connector *connector)
  * Unregister the DDC bus for this connector then free the driver private
  * structure.
  */
-void cdv_intel_lvds_destroy(struct drm_connector *connector)
+static void cdv_intel_lvds_destroy(struct drm_connector *connector)
 {
 	struct psb_intel_encoder *psb_intel_encoder =
 					psb_intel_attached_encoder(connector);
@@ -448,7 +450,7 @@ void cdv_intel_lvds_destroy(struct drm_connector *connector)
 	kfree(connector);
 }
 
-int cdv_intel_lvds_set_property(struct drm_connector *connector,
+static int cdv_intel_lvds_set_property(struct drm_connector *connector,
 				       struct drm_property *property,
 				       uint64_t value)
 {
diff --git a/drivers/gpu/drm/gma500/framebuffer.c b/drivers/gpu/drm/gma500/framebuffer.c
index 830dfdd6bf15..8ea202f1ba50 100644
--- a/drivers/gpu/drm/gma500/framebuffer.c
+++ b/drivers/gpu/drm/gma500/framebuffer.c
@@ -111,39 +111,6 @@ static int psbfb_pan(struct fb_var_screeninfo *var, struct fb_info *info)
         return 0;
 }
 
-void psbfb_suspend(struct drm_device *dev)
-{
-	struct drm_framebuffer *fb;
-
-	console_lock();
-	mutex_lock(&dev->mode_config.mutex);
-	list_for_each_entry(fb, &dev->mode_config.fb_list, head) {
-		struct psb_framebuffer *psbfb = to_psb_fb(fb);
-		struct fb_info *info = psbfb->fbdev;
-		fb_set_suspend(info, 1);
-		drm_fb_helper_blank(FB_BLANK_POWERDOWN, info);
-	}
-	mutex_unlock(&dev->mode_config.mutex);
-	console_unlock();
-}
-
-void psbfb_resume(struct drm_device *dev)
-{
-	struct drm_framebuffer *fb;
-
-	console_lock();
-	mutex_lock(&dev->mode_config.mutex);
-	list_for_each_entry(fb, &dev->mode_config.fb_list, head) {
-		struct psb_framebuffer *psbfb = to_psb_fb(fb);
-		struct fb_info *info = psbfb->fbdev;
-		fb_set_suspend(info, 0);
-		drm_fb_helper_blank(FB_BLANK_UNBLANK, info);
-	}
-	mutex_unlock(&dev->mode_config.mutex);
-	console_unlock();
-	drm_helper_disable_unused_functions(dev);
-}
-
 static int psbfb_vm_fault(struct vm_area_struct *vma, struct vm_fault *vmf)
 {
 	struct psb_framebuffer *psbfb = vma->vm_private_data;
@@ -158,7 +125,7 @@ static int psbfb_vm_fault(struct vm_area_struct *vma, struct vm_fault *vmf)
 	unsigned long phys_addr = (unsigned long)dev_priv->stolen_base;
 
 	page_num = (vma->vm_end - vma->vm_start) >> PAGE_SHIFT;
-	address = (unsigned long)vmf->virtual_address;
+	address = (unsigned long)vmf->virtual_address - (vmf->pgoff << PAGE_SHIFT);
 
 	vma->vm_page_prot = pgprot_noncached(vma->vm_page_prot);
 
@@ -247,7 +214,6 @@ static struct fb_ops psbfb_roll_ops = {
 	.fb_imageblit = cfb_imageblit,
 	.fb_pan_display = psbfb_pan,
 	.fb_mmap = psbfb_mmap,
-	.fb_sync = psbfb_sync,
 	.fb_ioctl = psbfb_ioctl,
 };
 
@@ -391,6 +357,7 @@ static int psbfb_create(struct psb_fbdev *fbdev,
 	mode_cmd.width = sizes->surface_width;
 	mode_cmd.height = sizes->surface_height;
 	bpp = sizes->surface_bpp;
+	depth = sizes->surface_depth;
 
 	/* No 24bit packed */
 	if (bpp == 24)
@@ -403,7 +370,6 @@ static int psbfb_create(struct psb_fbdev *fbdev,
 		 * is ok with some fonts
 		 */
         	mode_cmd.pitches[0] =  ALIGN(mode_cmd.width * ((bpp + 7) / 8), 4096 >> pitch_lines);
-        	depth = sizes->surface_depth;
 
         	size = mode_cmd.pitches[0] * mode_cmd.height;
         	size = ALIGN(size, PAGE_SIZE);
@@ -463,6 +429,7 @@ static int psbfb_create(struct psb_fbdev *fbdev,
 	fbdev->psb_fb_helper.fb = fb;
 	fbdev->psb_fb_helper.fbdev = info;
 
+	drm_fb_helper_fill_fix(info, fb->pitches[0], fb->depth);
 	strcpy(info->fix.id, "psbfb");
 
 	info->flags = FBINFO_DEFAULT;
@@ -500,18 +467,13 @@ static int psbfb_create(struct psb_fbdev *fbdev,
 		info->apertures->ranges[0].size = dev_priv->gtt.stolen_size;
 	}
 
-	drm_fb_helper_fill_fix(info, fb->pitches[0], fb->depth);
 	drm_fb_helper_fill_var(info, &fbdev->psb_fb_helper,
 				sizes->fb_width, sizes->fb_height);
 
 	info->fix.mmio_start = pci_resource_start(dev->pdev, 0);
 	info->fix.mmio_len = pci_resource_len(dev->pdev, 0);
 
-	info->pixmap.size = 64 * 1024;
-	info->pixmap.buf_align = 8;
-	info->pixmap.access_align = 32;
-	info->pixmap.flags = FB_PIXMAP_SYSTEM;
-	info->pixmap.scan_align = 1;
+	/* Use default scratch pixmap (info->pixmap.flags = FB_PIXMAP_SYSTEM) */
 
 	dev_info(dev->dev, "allocated %dx%d fb\n",
 					psbfb->base.width, psbfb->base.height);
@@ -560,11 +522,21 @@ static struct drm_framebuffer *psb_user_framebuffer_create
 static void psbfb_gamma_set(struct drm_crtc *crtc, u16 red, u16 green,
 							u16 blue, int regno)
 {
+	struct psb_intel_crtc *intel_crtc = to_psb_intel_crtc(crtc);
+
+	intel_crtc->lut_r[regno] = red >> 8;
+	intel_crtc->lut_g[regno] = green >> 8;
+	intel_crtc->lut_b[regno] = blue >> 8;
 }
 
 static void psbfb_gamma_get(struct drm_crtc *crtc, u16 *red,
 					u16 *green, u16 *blue, int regno)
 {
+	struct psb_intel_crtc *intel_crtc = to_psb_intel_crtc(crtc);
+
+	*red = intel_crtc->lut_r[regno] << 8;
+	*green = intel_crtc->lut_g[regno] << 8;
+	*blue = intel_crtc->lut_b[regno] << 8;
 }
 
 static int psbfb_probe(struct drm_fb_helper *helper,
@@ -589,7 +561,7 @@ struct drm_fb_helper_funcs psb_fb_helper_funcs = {
 	.fb_probe = psbfb_probe,
 };
 
-int psb_fbdev_destroy(struct drm_device *dev, struct psb_fbdev *fbdev)
+static int psb_fbdev_destroy(struct drm_device *dev, struct psb_fbdev *fbdev)
 {
 	struct fb_info *info;
 	struct psb_framebuffer *psbfb = &fbdev->pfb;
@@ -631,7 +603,7 @@ int psb_fbdev_init(struct drm_device *dev)
 	return 0;
 }
 
-void psb_fbdev_fini(struct drm_device *dev)
+static void psb_fbdev_fini(struct drm_device *dev)
 {
 	struct drm_psb_private *dev_priv = dev->dev_private;
 
@@ -725,10 +697,7 @@ static int psb_create_backlight_property(struct drm_device *dev)
 	if (dev_priv->backlight_property)
 		return 0;
 
-	backlight = drm_property_create(dev, DRM_MODE_PROP_RANGE,
-							"backlight", 2);
-	backlight->values[0] = 0;
-	backlight->values[1] = 100;
+	backlight = drm_property_create_range(dev, 0, "backlight", 0, 100);
 
 	dev_priv->backlight_property = backlight;
 
diff --git a/drivers/gpu/drm/gma500/gem_glue.c b/drivers/gpu/drm/gma500/gem_glue.c
index daac12120653..3c17634f6061 100644
--- a/drivers/gpu/drm/gma500/gem_glue.c
+++ b/drivers/gpu/drm/gma500/gem_glue.c
@@ -19,6 +19,7 @@
 
 #include <drm/drmP.h>
 #include <drm/drm.h>
+#include "gem_glue.h"
 
 void drm_gem_object_release_wrap(struct drm_gem_object *obj)
 {
diff --git a/drivers/gpu/drm/gma500/gtt.c b/drivers/gpu/drm/gma500/gtt.c
index 5d5330f667f1..c6465b40090f 100644
--- a/drivers/gpu/drm/gma500/gtt.c
+++ b/drivers/gpu/drm/gma500/gtt.c
@@ -57,7 +57,7 @@ static inline uint32_t psb_gtt_mask_pte(uint32_t pfn, int type)
  *	Given a gtt_range object return the GTT offset of the page table
  *	entries for this gtt_range
  */
-u32 *psb_gtt_entry(struct drm_device *dev, struct gtt_range *r)
+static u32 *psb_gtt_entry(struct drm_device *dev, struct gtt_range *r)
 {
 	struct drm_psb_private *dev_priv = dev->dev_private;
 	unsigned long offset;
@@ -378,7 +378,7 @@ void psb_gtt_free_range(struct drm_device *dev, struct gtt_range *gt)
 	kfree(gt);
 }
 
-void psb_gtt_alloc(struct drm_device *dev)
+static void psb_gtt_alloc(struct drm_device *dev)
 {
 	struct drm_psb_private *dev_priv = dev->dev_private;
 	init_rwsem(&dev_priv->gtt.sem);
@@ -446,10 +446,9 @@ int psb_gtt_init(struct drm_device *dev, int resume)
 	pg->gtt_start = pci_resource_start(dev->pdev, PSB_GTT_RESOURCE);
 	gtt_pages = pci_resource_len(dev->pdev, PSB_GTT_RESOURCE)
 								>> PAGE_SHIFT;
-	/* Some CDV firmware doesn't report this currently. In which case the
-	   system has 64 gtt pages */
+	/* CDV doesn't report this. In which case the system has 64 gtt pages */
 	if (pg->gtt_start == 0 || gtt_pages == 0) {
-		dev_err(dev->dev, "GTT PCI BAR not initialized.\n");
+		dev_dbg(dev->dev, "GTT PCI BAR not initialized.\n");
 		gtt_pages = 64;
 		pg->gtt_start = dev_priv->pge_ctl;
 	}
@@ -461,10 +460,10 @@ int psb_gtt_init(struct drm_device *dev, int resume)
 
 	if (pg->gatt_pages == 0 || pg->gatt_start == 0) {
 		static struct resource fudge;	/* Preferably peppermint */
-		/* This can occur on CDV SDV systems. Fudge it in this case.
+		/* This can occur on CDV systems. Fudge it in this case.
 		   We really don't care what imaginary space is being allocated
 		   at this point */
-		dev_err(dev->dev, "GATT PCI BAR not initialized.\n");
+		dev_dbg(dev->dev, "GATT PCI BAR not initialized.\n");
 		pg->gatt_start = 0x40000000;
 		pg->gatt_pages = (128 * 1024 * 1024) >> PAGE_SHIFT;
 		/* This is a little confusing but in fact the GTT is providing
diff --git a/drivers/gpu/drm/gma500/intel_gmbus.c b/drivers/gpu/drm/gma500/intel_gmbus.c
index 147584ac8d02..9db90527bf0f 100644
--- a/drivers/gpu/drm/gma500/intel_gmbus.c
+++ b/drivers/gpu/drm/gma500/intel_gmbus.c
@@ -395,7 +395,7 @@ int gma_intel_setup_gmbus(struct drm_device *dev)
 	struct drm_psb_private *dev_priv = dev->dev_private;
 	int ret, i;
 
-	dev_priv->gmbus = kcalloc(sizeof(struct intel_gmbus), GMBUS_NUM_PORTS,
+	dev_priv->gmbus = kcalloc(GMBUS_NUM_PORTS, sizeof(struct intel_gmbus),
 				  GFP_KERNEL);
 	if (dev_priv->gmbus == NULL)
 		return -ENOMEM;
diff --git a/drivers/gpu/drm/gma500/mdfld_device.c b/drivers/gpu/drm/gma500/mdfld_device.c
new file mode 100644
index 000000000000..af656787db0f
--- /dev/null
+++ b/drivers/gpu/drm/gma500/mdfld_device.c
@@ -0,0 +1,691 @@
+/**************************************************************************
+ * Copyright (c) 2011, Intel Corporation.
+ * All Rights Reserved.
+ *
+ * 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, write to the Free Software Foundation, Inc.,
+ * 51 Franklin St - Fifth Floor, Boston, MA 02110-1301 USA.
+ *
+ **************************************************************************/
+
+#include "psb_drv.h"
+#include "mid_bios.h"
+#include "mdfld_output.h"
+#include "mdfld_dsi_output.h"
+#include "tc35876x-dsi-lvds.h"
+
+#include <asm/intel_scu_ipc.h>
+
+#ifdef CONFIG_BACKLIGHT_CLASS_DEVICE
+
+#define MRST_BLC_MAX_PWM_REG_FREQ	    0xFFFF
+#define BLC_PWM_PRECISION_FACTOR 100	/* 10000000 */
+#define BLC_PWM_FREQ_CALC_CONSTANT 32
+#define MHz 1000000
+#define BRIGHTNESS_MIN_LEVEL 1
+#define BRIGHTNESS_MAX_LEVEL 100
+#define BRIGHTNESS_MASK	0xFF
+#define BLC_POLARITY_NORMAL 0
+#define BLC_POLARITY_INVERSE 1
+#define BLC_ADJUSTMENT_MAX 100
+
+#define MDFLD_BLC_PWM_PRECISION_FACTOR    10
+#define MDFLD_BLC_MAX_PWM_REG_FREQ        0xFFFE
+#define MDFLD_BLC_MIN_PWM_REG_FREQ        0x2
+
+#define MDFLD_BACKLIGHT_PWM_POLARITY_BIT_CLEAR (0xFFFE)
+#define MDFLD_BACKLIGHT_PWM_CTL_SHIFT	(16)
+
+static struct backlight_device *mdfld_backlight_device;
+
+int mdfld_set_brightness(struct backlight_device *bd)
+{
+	struct drm_device *dev =
+		(struct drm_device *)bl_get_data(mdfld_backlight_device);
+	struct drm_psb_private *dev_priv = dev->dev_private;
+	int level = bd->props.brightness;
+
+	DRM_DEBUG_DRIVER("backlight level set to %d\n", level);
+
+	/* Perform value bounds checking */
+	if (level < BRIGHTNESS_MIN_LEVEL)
+		level = BRIGHTNESS_MIN_LEVEL;
+
+	if (gma_power_begin(dev, false)) {
+		u32 adjusted_level = 0;
+
+		/*
+		 * Adjust the backlight level with the percent in
+		 * dev_priv->blc_adj2
+		 */
+		adjusted_level = level * dev_priv->blc_adj2;
+		adjusted_level = adjusted_level / BLC_ADJUSTMENT_MAX;
+		dev_priv->brightness_adjusted = adjusted_level;
+
+		if (mdfld_get_panel_type(dev, 0) == TC35876X) {
+			if (dev_priv->dpi_panel_on[0] ||
+					dev_priv->dpi_panel_on[2])
+				tc35876x_brightness_control(dev,
+						dev_priv->brightness_adjusted);
+		} else {
+			if (dev_priv->dpi_panel_on[0])
+				mdfld_dsi_brightness_control(dev, 0,
+						dev_priv->brightness_adjusted);
+		}
+
+		if (dev_priv->dpi_panel_on[2])
+			mdfld_dsi_brightness_control(dev, 2,
+					dev_priv->brightness_adjusted);
+		gma_power_end(dev);
+	}
+
+	/* cache the brightness for later use */
+	dev_priv->brightness = level;
+	return 0;
+}
+
+static int mdfld_get_brightness(struct backlight_device *bd)
+{
+	struct drm_device *dev =
+		(struct drm_device *)bl_get_data(mdfld_backlight_device);
+	struct drm_psb_private *dev_priv = dev->dev_private;
+
+	DRM_DEBUG_DRIVER("brightness = 0x%x \n", dev_priv->brightness);
+
+	/* return locally cached var instead of HW read (due to DPST etc.) */
+	return dev_priv->brightness;
+}
+
+static const struct backlight_ops mdfld_ops = {
+	.get_brightness = mdfld_get_brightness,
+	.update_status  = mdfld_set_brightness,
+};
+
+static int device_backlight_init(struct drm_device *dev)
+{
+	struct drm_psb_private *dev_priv = (struct drm_psb_private *)
+		dev->dev_private;
+
+	dev_priv->blc_adj1 = BLC_ADJUSTMENT_MAX;
+	dev_priv->blc_adj2 = BLC_ADJUSTMENT_MAX;
+
+	return 0;
+}
+
+static int mdfld_backlight_init(struct drm_device *dev)
+{
+	struct backlight_properties props;
+	int ret = 0;
+
+	memset(&props, 0, sizeof(struct backlight_properties));
+	props.max_brightness = BRIGHTNESS_MAX_LEVEL;
+	props.type = BACKLIGHT_PLATFORM;
+	mdfld_backlight_device = backlight_device_register("mdfld-bl",
+				NULL, (void *)dev, &mdfld_ops, &props);
+
+	if (IS_ERR(mdfld_backlight_device))
+		return PTR_ERR(mdfld_backlight_device);
+
+	ret = device_backlight_init(dev);
+	if (ret)
+		return ret;
+
+	mdfld_backlight_device->props.brightness = BRIGHTNESS_MAX_LEVEL;
+	mdfld_backlight_device->props.max_brightness = BRIGHTNESS_MAX_LEVEL;
+	backlight_update_status(mdfld_backlight_device);
+	return 0;
+}
+#endif
+
+struct backlight_device *mdfld_get_backlight_device(void)
+{
+#ifdef CONFIG_BACKLIGHT_CLASS_DEVICE
+	return mdfld_backlight_device;
+#else
+	return NULL;
+#endif
+}
+
+/*
+ * mdfld_save_display_registers
+ *
+ * Description: We are going to suspend so save current display
+ * register state.
+ *
+ * Notes: FIXME_JLIU7 need to add the support for DPI MIPI & HDMI audio
+ */
+static int mdfld_save_display_registers(struct drm_device *dev, int pipe)
+{
+	struct drm_psb_private *dev_priv = dev->dev_private;
+	struct medfield_state *regs = &dev_priv->regs.mdfld;
+	int i;
+
+	/* register */
+	u32 dpll_reg = MRST_DPLL_A;
+	u32 fp_reg = MRST_FPA0;
+	u32 pipeconf_reg = PIPEACONF;
+	u32 htot_reg = HTOTAL_A;
+	u32 hblank_reg = HBLANK_A;
+	u32 hsync_reg = HSYNC_A;
+	u32 vtot_reg = VTOTAL_A;
+	u32 vblank_reg = VBLANK_A;
+	u32 vsync_reg = VSYNC_A;
+	u32 pipesrc_reg = PIPEASRC;
+	u32 dspstride_reg = DSPASTRIDE;
+	u32 dsplinoff_reg = DSPALINOFF;
+	u32 dsptileoff_reg = DSPATILEOFF;
+	u32 dspsize_reg = DSPASIZE;
+	u32 dsppos_reg = DSPAPOS;
+	u32 dspsurf_reg = DSPASURF;
+	u32 mipi_reg = MIPI;
+	u32 dspcntr_reg = DSPACNTR;
+	u32 dspstatus_reg = PIPEASTAT;
+	u32 palette_reg = PALETTE_A;
+
+	/* pointer to values */
+	u32 *dpll_val = &regs->saveDPLL_A;
+	u32 *fp_val = &regs->saveFPA0;
+	u32 *pipeconf_val = &regs->savePIPEACONF;
+	u32 *htot_val = &regs->saveHTOTAL_A;
+	u32 *hblank_val = &regs->saveHBLANK_A;
+	u32 *hsync_val = &regs->saveHSYNC_A;
+	u32 *vtot_val = &regs->saveVTOTAL_A;
+	u32 *vblank_val = &regs->saveVBLANK_A;
+	u32 *vsync_val = &regs->saveVSYNC_A;
+	u32 *pipesrc_val = &regs->savePIPEASRC;
+	u32 *dspstride_val = &regs->saveDSPASTRIDE;
+	u32 *dsplinoff_val = &regs->saveDSPALINOFF;
+	u32 *dsptileoff_val = &regs->saveDSPATILEOFF;
+	u32 *dspsize_val = &regs->saveDSPASIZE;
+	u32 *dsppos_val = &regs->saveDSPAPOS;
+	u32 *dspsurf_val = &regs->saveDSPASURF;
+	u32 *mipi_val = &regs->saveMIPI;
+	u32 *dspcntr_val = &regs->saveDSPACNTR;
+	u32 *dspstatus_val = &regs->saveDSPASTATUS;
+	u32 *palette_val = regs->save_palette_a;
+
+	switch (pipe) {
+	case 0:
+		break;
+	case 1:
+		/* regester */
+		dpll_reg = MDFLD_DPLL_B;
+		fp_reg = MDFLD_DPLL_DIV0;
+		pipeconf_reg = PIPEBCONF;
+		htot_reg = HTOTAL_B;
+		hblank_reg = HBLANK_B;
+		hsync_reg = HSYNC_B;
+		vtot_reg = VTOTAL_B;
+		vblank_reg = VBLANK_B;
+		vsync_reg = VSYNC_B;
+		pipesrc_reg = PIPEBSRC;
+		dspstride_reg = DSPBSTRIDE;
+		dsplinoff_reg = DSPBLINOFF;
+		dsptileoff_reg = DSPBTILEOFF;
+		dspsize_reg = DSPBSIZE;
+		dsppos_reg = DSPBPOS;
+		dspsurf_reg = DSPBSURF;
+		dspcntr_reg = DSPBCNTR;
+		dspstatus_reg = PIPEBSTAT;
+		palette_reg = PALETTE_B;
+
+		/* values */
+		dpll_val = &regs->saveDPLL_B;
+		fp_val = &regs->saveFPB0;
+		pipeconf_val = &regs->savePIPEBCONF;
+		htot_val = &regs->saveHTOTAL_B;
+		hblank_val = &regs->saveHBLANK_B;
+		hsync_val = &regs->saveHSYNC_B;
+		vtot_val = &regs->saveVTOTAL_B;
+		vblank_val = &regs->saveVBLANK_B;
+		vsync_val = &regs->saveVSYNC_B;
+		pipesrc_val = &regs->savePIPEBSRC;
+		dspstride_val = &regs->saveDSPBSTRIDE;
+		dsplinoff_val = &regs->saveDSPBLINOFF;
+		dsptileoff_val = &regs->saveDSPBTILEOFF;
+		dspsize_val = &regs->saveDSPBSIZE;
+		dsppos_val = &regs->saveDSPBPOS;
+		dspsurf_val = &regs->saveDSPBSURF;
+		dspcntr_val = &regs->saveDSPBCNTR;
+		dspstatus_val = &regs->saveDSPBSTATUS;
+		palette_val = regs->save_palette_b;
+		break;
+	case 2:
+		/* register */
+		pipeconf_reg = PIPECCONF;
+		htot_reg = HTOTAL_C;
+		hblank_reg = HBLANK_C;
+		hsync_reg = HSYNC_C;
+		vtot_reg = VTOTAL_C;
+		vblank_reg = VBLANK_C;
+		vsync_reg = VSYNC_C;
+		pipesrc_reg = PIPECSRC;
+		dspstride_reg = DSPCSTRIDE;
+		dsplinoff_reg = DSPCLINOFF;
+		dsptileoff_reg = DSPCTILEOFF;
+		dspsize_reg = DSPCSIZE;
+		dsppos_reg = DSPCPOS;
+		dspsurf_reg = DSPCSURF;
+		mipi_reg = MIPI_C;
+		dspcntr_reg = DSPCCNTR;
+		dspstatus_reg = PIPECSTAT;
+		palette_reg = PALETTE_C;
+
+		/* pointer to values */
+		pipeconf_val = &regs->savePIPECCONF;
+		htot_val = &regs->saveHTOTAL_C;
+		hblank_val = &regs->saveHBLANK_C;
+		hsync_val = &regs->saveHSYNC_C;
+		vtot_val = &regs->saveVTOTAL_C;
+		vblank_val = &regs->saveVBLANK_C;
+		vsync_val = &regs->saveVSYNC_C;
+		pipesrc_val = &regs->savePIPECSRC;
+		dspstride_val = &regs->saveDSPCSTRIDE;
+		dsplinoff_val = &regs->saveDSPCLINOFF;
+		dsptileoff_val = &regs->saveDSPCTILEOFF;
+		dspsize_val = &regs->saveDSPCSIZE;
+		dsppos_val = &regs->saveDSPCPOS;
+		dspsurf_val = &regs->saveDSPCSURF;
+		mipi_val = &regs->saveMIPI_C;
+		dspcntr_val = &regs->saveDSPCCNTR;
+		dspstatus_val = &regs->saveDSPCSTATUS;
+		palette_val = regs->save_palette_c;
+		break;
+	default:
+		DRM_ERROR("%s, invalid pipe number.\n", __func__);
+		return -EINVAL;
+	}
+
+	/* Pipe & plane A info */
+	*dpll_val = PSB_RVDC32(dpll_reg);
+	*fp_val = PSB_RVDC32(fp_reg);
+	*pipeconf_val = PSB_RVDC32(pipeconf_reg);
+	*htot_val = PSB_RVDC32(htot_reg);
+	*hblank_val = PSB_RVDC32(hblank_reg);
+	*hsync_val = PSB_RVDC32(hsync_reg);
+	*vtot_val = PSB_RVDC32(vtot_reg);
+	*vblank_val = PSB_RVDC32(vblank_reg);
+	*vsync_val = PSB_RVDC32(vsync_reg);
+	*pipesrc_val = PSB_RVDC32(pipesrc_reg);
+	*dspstride_val = PSB_RVDC32(dspstride_reg);
+	*dsplinoff_val = PSB_RVDC32(dsplinoff_reg);
+	*dsptileoff_val = PSB_RVDC32(dsptileoff_reg);
+	*dspsize_val = PSB_RVDC32(dspsize_reg);
+	*dsppos_val = PSB_RVDC32(dsppos_reg);
+	*dspsurf_val = PSB_RVDC32(dspsurf_reg);
+	*dspcntr_val = PSB_RVDC32(dspcntr_reg);
+	*dspstatus_val = PSB_RVDC32(dspstatus_reg);
+
+	/*save palette (gamma) */
+	for (i = 0; i < 256; i++)
+		palette_val[i] = PSB_RVDC32(palette_reg + (i << 2));
+
+	if (pipe == 1) {
+		regs->savePFIT_CONTROL = PSB_RVDC32(PFIT_CONTROL);
+		regs->savePFIT_PGM_RATIOS = PSB_RVDC32(PFIT_PGM_RATIOS);
+
+		regs->saveHDMIPHYMISCCTL = PSB_RVDC32(HDMIPHYMISCCTL);
+		regs->saveHDMIB_CONTROL = PSB_RVDC32(HDMIB_CONTROL);
+		return 0;
+	}
+
+	*mipi_val = PSB_RVDC32(mipi_reg);
+	return 0;
+}
+
+/*
+ * mdfld_restore_display_registers
+ *
+ * Description: We are going to resume so restore display register state.
+ *
+ * Notes: FIXME_JLIU7 need to add the support for DPI MIPI & HDMI audio
+ */
+static int mdfld_restore_display_registers(struct drm_device *dev, int pipe)
+{
+	/* To get  panel out of ULPS mode. */
+	u32 temp = 0;
+	u32 device_ready_reg = DEVICE_READY_REG;
+	struct drm_psb_private *dev_priv = dev->dev_private;
+	struct mdfld_dsi_config *dsi_config = NULL;
+	struct medfield_state *regs = &dev_priv->regs.mdfld;
+	u32 i = 0;
+	u32 dpll = 0;
+	u32 timeout = 0;
+
+	/* regester */
+	u32 dpll_reg = MRST_DPLL_A;
+	u32 fp_reg = MRST_FPA0;
+	u32 pipeconf_reg = PIPEACONF;
+	u32 htot_reg = HTOTAL_A;
+	u32 hblank_reg = HBLANK_A;
+	u32 hsync_reg = HSYNC_A;
+	u32 vtot_reg = VTOTAL_A;
+	u32 vblank_reg = VBLANK_A;
+	u32 vsync_reg = VSYNC_A;
+	u32 pipesrc_reg = PIPEASRC;
+	u32 dspstride_reg = DSPASTRIDE;
+	u32 dsplinoff_reg = DSPALINOFF;
+	u32 dsptileoff_reg = DSPATILEOFF;
+	u32 dspsize_reg = DSPASIZE;
+	u32 dsppos_reg = DSPAPOS;
+	u32 dspsurf_reg = DSPASURF;
+	u32 dspstatus_reg = PIPEASTAT;
+	u32 mipi_reg = MIPI;
+	u32 dspcntr_reg = DSPACNTR;
+	u32 palette_reg = PALETTE_A;
+
+	/* values */
+	u32 dpll_val = regs->saveDPLL_A & ~DPLL_VCO_ENABLE;
+	u32 fp_val = regs->saveFPA0;
+	u32 pipeconf_val = regs->savePIPEACONF;
+	u32 htot_val = regs->saveHTOTAL_A;
+	u32 hblank_val = regs->saveHBLANK_A;
+	u32 hsync_val = regs->saveHSYNC_A;
+	u32 vtot_val = regs->saveVTOTAL_A;
+	u32 vblank_val = regs->saveVBLANK_A;
+	u32 vsync_val = regs->saveVSYNC_A;
+	u32 pipesrc_val = regs->savePIPEASRC;
+	u32 dspstride_val = regs->saveDSPASTRIDE;
+	u32 dsplinoff_val = regs->saveDSPALINOFF;
+	u32 dsptileoff_val = regs->saveDSPATILEOFF;
+	u32 dspsize_val = regs->saveDSPASIZE;
+	u32 dsppos_val = regs->saveDSPAPOS;
+	u32 dspsurf_val = regs->saveDSPASURF;
+	u32 dspstatus_val = regs->saveDSPASTATUS;
+	u32 mipi_val = regs->saveMIPI;
+	u32 dspcntr_val = regs->saveDSPACNTR;
+	u32 *palette_val = regs->save_palette_a;
+
+	switch (pipe) {
+	case 0:
+		dsi_config = dev_priv->dsi_configs[0];
+		break;
+	case 1:
+		/* regester */
+		dpll_reg = MDFLD_DPLL_B;
+		fp_reg = MDFLD_DPLL_DIV0;
+		pipeconf_reg = PIPEBCONF;
+		htot_reg = HTOTAL_B;
+		hblank_reg = HBLANK_B;
+		hsync_reg = HSYNC_B;
+		vtot_reg = VTOTAL_B;
+		vblank_reg = VBLANK_B;
+		vsync_reg = VSYNC_B;
+		pipesrc_reg = PIPEBSRC;
+		dspstride_reg = DSPBSTRIDE;
+		dsplinoff_reg = DSPBLINOFF;
+		dsptileoff_reg = DSPBTILEOFF;
+		dspsize_reg = DSPBSIZE;
+		dsppos_reg = DSPBPOS;
+		dspsurf_reg = DSPBSURF;
+		dspcntr_reg = DSPBCNTR;
+		dspstatus_reg = PIPEBSTAT;
+		palette_reg = PALETTE_B;
+
+		/* values */
+		dpll_val = regs->saveDPLL_B & ~DPLL_VCO_ENABLE;
+		fp_val = regs->saveFPB0;
+		pipeconf_val = regs->savePIPEBCONF;
+		htot_val = regs->saveHTOTAL_B;
+		hblank_val = regs->saveHBLANK_B;
+		hsync_val = regs->saveHSYNC_B;
+		vtot_val = regs->saveVTOTAL_B;
+		vblank_val = regs->saveVBLANK_B;
+		vsync_val = regs->saveVSYNC_B;
+		pipesrc_val = regs->savePIPEBSRC;
+		dspstride_val = regs->saveDSPBSTRIDE;
+		dsplinoff_val = regs->saveDSPBLINOFF;
+		dsptileoff_val = regs->saveDSPBTILEOFF;
+		dspsize_val = regs->saveDSPBSIZE;
+		dsppos_val = regs->saveDSPBPOS;
+		dspsurf_val = regs->saveDSPBSURF;
+		dspcntr_val = regs->saveDSPBCNTR;
+		dspstatus_val = regs->saveDSPBSTATUS;
+		palette_val = regs->save_palette_b;
+		break;
+	case 2:
+		/* regester */
+		pipeconf_reg = PIPECCONF;
+		htot_reg = HTOTAL_C;
+		hblank_reg = HBLANK_C;
+		hsync_reg = HSYNC_C;
+		vtot_reg = VTOTAL_C;
+		vblank_reg = VBLANK_C;
+		vsync_reg = VSYNC_C;
+		pipesrc_reg = PIPECSRC;
+		dspstride_reg = DSPCSTRIDE;
+		dsplinoff_reg = DSPCLINOFF;
+		dsptileoff_reg = DSPCTILEOFF;
+		dspsize_reg = DSPCSIZE;
+		dsppos_reg = DSPCPOS;
+		dspsurf_reg = DSPCSURF;
+		mipi_reg = MIPI_C;
+		dspcntr_reg = DSPCCNTR;
+		dspstatus_reg = PIPECSTAT;
+		palette_reg = PALETTE_C;
+
+		/* values */
+		pipeconf_val = regs->savePIPECCONF;
+		htot_val = regs->saveHTOTAL_C;
+		hblank_val = regs->saveHBLANK_C;
+		hsync_val = regs->saveHSYNC_C;
+		vtot_val = regs->saveVTOTAL_C;
+		vblank_val = regs->saveVBLANK_C;
+		vsync_val = regs->saveVSYNC_C;
+		pipesrc_val = regs->savePIPECSRC;
+		dspstride_val = regs->saveDSPCSTRIDE;
+		dsplinoff_val = regs->saveDSPCLINOFF;
+		dsptileoff_val = regs->saveDSPCTILEOFF;
+		dspsize_val = regs->saveDSPCSIZE;
+		dsppos_val = regs->saveDSPCPOS;
+		dspsurf_val = regs->saveDSPCSURF;
+		mipi_val = regs->saveMIPI_C;
+		dspcntr_val = regs->saveDSPCCNTR;
+		dspstatus_val = regs->saveDSPCSTATUS;
+		palette_val = regs->save_palette_c;
+
+		dsi_config = dev_priv->dsi_configs[1];
+		break;
+	default:
+		DRM_ERROR("%s, invalid pipe number.\n", __func__);
+		return -EINVAL;
+	}
+
+	/*make sure VGA plane is off. it initializes to on after reset!*/
+	PSB_WVDC32(0x80000000, VGACNTRL);
+
+	if (pipe == 1) {
+		PSB_WVDC32(dpll_val & ~DPLL_VCO_ENABLE, dpll_reg);
+		PSB_RVDC32(dpll_reg);
+
+		PSB_WVDC32(fp_val, fp_reg);
+	} else {
+
+		dpll = PSB_RVDC32(dpll_reg);
+
+		if (!(dpll & DPLL_VCO_ENABLE)) {
+
+			/* When ungating power of DPLL, needs to wait 0.5us
+			   before enable the VCO */
+			if (dpll & MDFLD_PWR_GATE_EN) {
+				dpll &= ~MDFLD_PWR_GATE_EN;
+				PSB_WVDC32(dpll, dpll_reg);
+				/* FIXME_MDFLD PO - change 500 to 1 after PO */
+				udelay(500);
+			}
+
+			PSB_WVDC32(fp_val, fp_reg);
+			PSB_WVDC32(dpll_val, dpll_reg);
+			/* FIXME_MDFLD PO - change 500 to 1 after PO */
+			udelay(500);
+
+			dpll_val |= DPLL_VCO_ENABLE;
+			PSB_WVDC32(dpll_val, dpll_reg);
+			PSB_RVDC32(dpll_reg);
+
+			/* wait for DSI PLL to lock */
+			while (timeout < 20000 &&
+			  !(PSB_RVDC32(pipeconf_reg) & PIPECONF_DSIPLL_LOCK)) {
+				udelay(150);
+				timeout++;
+			}
+
+			if (timeout == 20000) {
+				DRM_ERROR("%s, can't lock DSIPLL.\n",
+								__func__);
+				return -EINVAL;
+			}
+		}
+	}
+	/* Restore mode */
+	PSB_WVDC32(htot_val, htot_reg);
+	PSB_WVDC32(hblank_val, hblank_reg);
+	PSB_WVDC32(hsync_val, hsync_reg);
+	PSB_WVDC32(vtot_val, vtot_reg);
+	PSB_WVDC32(vblank_val, vblank_reg);
+	PSB_WVDC32(vsync_val, vsync_reg);
+	PSB_WVDC32(pipesrc_val, pipesrc_reg);
+	PSB_WVDC32(dspstatus_val, dspstatus_reg);
+
+	/*set up the plane*/
+	PSB_WVDC32(dspstride_val, dspstride_reg);
+	PSB_WVDC32(dsplinoff_val, dsplinoff_reg);
+	PSB_WVDC32(dsptileoff_val, dsptileoff_reg);
+	PSB_WVDC32(dspsize_val, dspsize_reg);
+	PSB_WVDC32(dsppos_val, dsppos_reg);
+	PSB_WVDC32(dspsurf_val, dspsurf_reg);
+
+	if (pipe == 1) {
+		/* restore palette (gamma) */
+		/*DRM_UDELAY(50000); */
+		for (i = 0; i < 256; i++)
+			PSB_WVDC32(palette_val[i], palette_reg + (i << 2));
+
+		PSB_WVDC32(regs->savePFIT_CONTROL, PFIT_CONTROL);
+		PSB_WVDC32(regs->savePFIT_PGM_RATIOS, PFIT_PGM_RATIOS);
+
+		/*TODO: resume HDMI port */
+
+		/*TODO: resume pipe*/
+
+		/*enable the plane*/
+		PSB_WVDC32(dspcntr_val & ~DISPLAY_PLANE_ENABLE, dspcntr_reg);
+
+		return 0;
+	}
+
+	/*set up pipe related registers*/
+	PSB_WVDC32(mipi_val, mipi_reg);
+
+	/*setup MIPI adapter + MIPI IP registers*/
+	if (dsi_config)
+		mdfld_dsi_controller_init(dsi_config, pipe);
+
+	if (in_atomic() || in_interrupt())
+		mdelay(20);
+	else
+		msleep(20);
+
+	/*enable the plane*/
+	PSB_WVDC32(dspcntr_val, dspcntr_reg);
+
+	if (in_atomic() || in_interrupt())
+		mdelay(20);
+	else
+		msleep(20);
+
+	/* LP Hold Release */
+	temp = REG_READ(mipi_reg);
+	temp |= LP_OUTPUT_HOLD_RELEASE;
+	REG_WRITE(mipi_reg, temp);
+	mdelay(1);
+
+
+	/* Set DSI host to exit from Utra Low Power State */
+	temp = REG_READ(device_ready_reg);
+	temp &= ~ULPS_MASK;
+	temp |= 0x3;
+	temp |= EXIT_ULPS_DEV_READY;
+	REG_WRITE(device_ready_reg, temp);
+	mdelay(1);
+
+	temp = REG_READ(device_ready_reg);
+	temp &= ~ULPS_MASK;
+	temp |= EXITING_ULPS;
+	REG_WRITE(device_ready_reg, temp);
+	mdelay(1);
+
+	/*enable the pipe*/
+	PSB_WVDC32(pipeconf_val, pipeconf_reg);
+
+	/* restore palette (gamma) */
+	/*DRM_UDELAY(50000); */
+	for (i = 0; i < 256; i++)
+		PSB_WVDC32(palette_val[i], palette_reg + (i << 2));
+
+	return 0;
+}
+
+static int mdfld_save_registers(struct drm_device *dev)
+{
+	/* mdfld_save_cursor_overlay_registers(dev); */
+	mdfld_save_display_registers(dev, 0);
+	mdfld_save_display_registers(dev, 2);
+	mdfld_disable_crtc(dev, 0);
+	mdfld_disable_crtc(dev, 2);
+
+	return 0;
+}
+
+static int mdfld_restore_registers(struct drm_device *dev)
+{
+	mdfld_restore_display_registers(dev, 2);
+	mdfld_restore_display_registers(dev, 0);
+	/* mdfld_restore_cursor_overlay_registers(dev); */
+
+	return 0;
+}
+
+static int mdfld_power_down(struct drm_device *dev)
+{
+	/* FIXME */
+	return 0;
+}
+
+static int mdfld_power_up(struct drm_device *dev)
+{
+	/* FIXME */
+	return 0;
+}
+
+const struct psb_ops mdfld_chip_ops = {
+	.name = "mdfld",
+	.accel_2d = 0,
+	.pipes = 3,
+	.crtcs = 3,
+	.sgx_offset = MRST_SGX_OFFSET,
+
+	.chip_setup = mid_chip_setup,
+	.crtc_helper = &mdfld_helper_funcs,
+	.crtc_funcs = &psb_intel_crtc_funcs,
+
+	.output_init = mdfld_output_init,
+
+#ifdef CONFIG_BACKLIGHT_CLASS_DEVICE
+	.backlight_init = mdfld_backlight_init,
+#endif
+
+	.save_regs = mdfld_save_registers,
+	.restore_regs = mdfld_restore_registers,
+	.power_down = mdfld_power_down,
+	.power_up = mdfld_power_up,
+};
diff --git a/drivers/gpu/drm/gma500/mdfld_dsi_dpi.c b/drivers/gpu/drm/gma500/mdfld_dsi_dpi.c
new file mode 100644
index 000000000000..d52358b744a0
--- /dev/null
+++ b/drivers/gpu/drm/gma500/mdfld_dsi_dpi.c
@@ -0,0 +1,1017 @@
+/*
+ * Copyright © 2010 Intel Corporation
+ *
+ * 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:
+ * jim liu <jim.liu@intel.com>
+ * Jackie Li<yaodong.li@intel.com>
+ */
+
+#include "mdfld_dsi_dpi.h"
+#include "mdfld_output.h"
+#include "mdfld_dsi_pkg_sender.h"
+#include "psb_drv.h"
+#include "tc35876x-dsi-lvds.h"
+
+static void mdfld_dsi_dpi_shut_down(struct mdfld_dsi_dpi_output *output,
+								int pipe);
+
+static void mdfld_wait_for_HS_DATA_FIFO(struct drm_device *dev, u32 pipe)
+{
+	u32 gen_fifo_stat_reg = MIPI_GEN_FIFO_STAT_REG(pipe);
+	int timeout = 0;
+
+	udelay(500);
+
+	/* This will time out after approximately 2+ seconds */
+	while ((timeout < 20000) &&
+		(REG_READ(gen_fifo_stat_reg) & DSI_FIFO_GEN_HS_DATA_FULL)) {
+		udelay(100);
+		timeout++;
+	}
+
+	if (timeout == 20000)
+		DRM_INFO("MIPI: HS Data FIFO was never cleared!\n");
+}
+
+static void mdfld_wait_for_HS_CTRL_FIFO(struct drm_device *dev, u32 pipe)
+{
+	u32 gen_fifo_stat_reg = MIPI_GEN_FIFO_STAT_REG(pipe);
+	int timeout = 0;
+
+	udelay(500);
+
+	/* This will time out after approximately 2+ seconds */
+	while ((timeout < 20000) && (REG_READ(gen_fifo_stat_reg)
+					& DSI_FIFO_GEN_HS_CTRL_FULL)) {
+		udelay(100);
+		timeout++;
+	}
+	if (timeout == 20000)
+		DRM_INFO("MIPI: HS CMD FIFO was never cleared!\n");
+}
+
+static void mdfld_wait_for_DPI_CTRL_FIFO(struct drm_device *dev, u32 pipe)
+{
+	u32 gen_fifo_stat_reg = MIPI_GEN_FIFO_STAT_REG(pipe);
+	int timeout = 0;
+
+	udelay(500);
+
+	/* This will time out after approximately 2+ seconds */
+	while ((timeout < 20000) && ((REG_READ(gen_fifo_stat_reg) &
+					DPI_FIFO_EMPTY) != DPI_FIFO_EMPTY)) {
+		udelay(100);
+		timeout++;
+	}
+
+	if (timeout == 20000)
+		DRM_ERROR("MIPI: DPI FIFO was never cleared\n");
+}
+
+static void mdfld_wait_for_SPL_PKG_SENT(struct drm_device *dev, u32 pipe)
+{
+	u32 intr_stat_reg = MIPI_INTR_STAT_REG(pipe);
+	int timeout = 0;
+
+	udelay(500);
+
+	/* This will time out after approximately 2+ seconds */
+	while ((timeout < 20000) && (!(REG_READ(intr_stat_reg)
+					& DSI_INTR_STATE_SPL_PKG_SENT))) {
+		udelay(100);
+		timeout++;
+	}
+
+	if (timeout == 20000)
+                DRM_ERROR("MIPI: SPL_PKT_SENT_INTERRUPT was not sent successfully!\n");
+}
+
+/* For TC35876X */
+
+static void dsi_set_device_ready_state(struct drm_device *dev, int state,
+				int pipe)
+{
+	REG_FLD_MOD(MIPI_DEVICE_READY_REG(pipe), !!state, 0, 0);
+}
+
+static void dsi_set_pipe_plane_enable_state(struct drm_device *dev,
+							int state, int pipe)
+{
+	struct drm_psb_private *dev_priv = dev->dev_private;
+	u32 pipeconf_reg = PIPEACONF;
+	u32 dspcntr_reg = DSPACNTR;
+
+	u32 dspcntr = dev_priv->dspcntr[pipe];
+	u32 mipi = MIPI_PORT_EN | PASS_FROM_SPHY_TO_AFE | SEL_FLOPPED_HSTX;
+
+	if (pipe) {
+		pipeconf_reg = PIPECCONF;
+		dspcntr_reg = DSPCCNTR;
+	} else
+		mipi &= (~0x03);
+
+	if (state) {
+		/*Set up pipe */
+		REG_WRITE(pipeconf_reg, BIT(31));
+
+		if (REG_BIT_WAIT(pipeconf_reg, 1, 30))
+			dev_err(&dev->pdev->dev, "%s: Pipe enable timeout\n",
+				__func__);
+
+		/*Set up display plane */
+		REG_WRITE(dspcntr_reg, dspcntr);
+	} else {
+		u32 dspbase_reg = pipe ? MDFLD_DSPCBASE : MRST_DSPABASE;
+
+		/* Put DSI lanes to ULPS to disable pipe */
+		REG_FLD_MOD(MIPI_DEVICE_READY_REG(pipe), 2, 2, 1);
+		REG_READ(MIPI_DEVICE_READY_REG(pipe)); /* posted write? */
+
+		/* LP Hold */
+		REG_FLD_MOD(MIPI_PORT_CONTROL(pipe), 0, 16, 16);
+		REG_READ(MIPI_PORT_CONTROL(pipe)); /* posted write? */
+
+		/* Disable display plane */
+		REG_FLD_MOD(dspcntr_reg, 0, 31, 31);
+
+		/* Flush the plane changes ??? posted write? */
+		REG_WRITE(dspbase_reg, REG_READ(dspbase_reg));
+		REG_READ(dspbase_reg);
+
+		/* Disable PIPE */
+		REG_FLD_MOD(pipeconf_reg, 0, 31, 31);
+
+		if (REG_BIT_WAIT(pipeconf_reg, 0, 30))
+			dev_err(&dev->pdev->dev, "%s: Pipe disable timeout\n",
+				__func__);
+
+		if (REG_BIT_WAIT(MIPI_GEN_FIFO_STAT_REG(pipe), 1, 28))
+			dev_err(&dev->pdev->dev, "%s: FIFO not empty\n",
+				__func__);
+	}
+}
+
+static void mdfld_dsi_configure_down(struct mdfld_dsi_encoder *dsi_encoder,
+								int pipe)
+{
+	struct mdfld_dsi_dpi_output *dpi_output =
+				MDFLD_DSI_DPI_OUTPUT(dsi_encoder);
+	struct mdfld_dsi_config *dsi_config =
+				mdfld_dsi_encoder_get_config(dsi_encoder);
+	struct drm_device *dev = dsi_config->dev;
+	struct drm_psb_private *dev_priv = dev->dev_private;
+
+	if (!dev_priv->dpi_panel_on[pipe]) {
+		dev_err(dev->dev, "DPI panel is already off\n");
+		return;
+	}
+	tc35876x_toshiba_bridge_panel_off(dev);
+	tc35876x_set_bridge_reset_state(dev, 1);
+	dsi_set_pipe_plane_enable_state(dev, 0, pipe);
+	mdfld_dsi_dpi_shut_down(dpi_output, pipe);
+	dsi_set_device_ready_state(dev, 0, pipe);
+}
+
+static void mdfld_dsi_configure_up(struct mdfld_dsi_encoder *dsi_encoder,
+								int pipe)
+{
+	struct mdfld_dsi_dpi_output *dpi_output =
+				MDFLD_DSI_DPI_OUTPUT(dsi_encoder);
+	struct mdfld_dsi_config *dsi_config =
+				mdfld_dsi_encoder_get_config(dsi_encoder);
+	struct drm_device *dev = dsi_config->dev;
+	struct drm_psb_private *dev_priv = dev->dev_private;
+
+	if (dev_priv->dpi_panel_on[pipe]) {
+		dev_err(dev->dev, "DPI panel is already on\n");
+		return;
+	}
+
+	/* For resume path sequence */
+	mdfld_dsi_dpi_shut_down(dpi_output, pipe);
+	dsi_set_device_ready_state(dev, 0, pipe);
+
+	dsi_set_device_ready_state(dev, 1, pipe);
+	tc35876x_set_bridge_reset_state(dev, 0);
+	tc35876x_configure_lvds_bridge(dev);
+	mdfld_dsi_dpi_turn_on(dpi_output, pipe);  /* Send turn on command */
+	dsi_set_pipe_plane_enable_state(dev, 1, pipe);
+}
+/* End for TC35876X */
+
+/* ************************************************************************* *\
+ * FUNCTION: mdfld_dsi_tpo_ic_init
+ *
+ * DESCRIPTION:  This function is called only by mrst_dsi_mode_set and
+ *               restore_display_registers.  since this function does not
+ *               acquire the mutex, it is important that the calling function
+ *               does!
+\* ************************************************************************* */
+static void mdfld_dsi_tpo_ic_init(struct mdfld_dsi_config *dsi_config, u32 pipe)
+{
+	struct drm_device *dev = dsi_config->dev;
+	u32 dcsChannelNumber = dsi_config->channel_num;
+	u32 gen_data_reg = MIPI_HS_GEN_DATA_REG(pipe);
+	u32 gen_ctrl_reg = MIPI_HS_GEN_CTRL_REG(pipe);
+	u32 gen_ctrl_val = GEN_LONG_WRITE;
+
+	DRM_INFO("Enter mrst init TPO MIPI display.\n");
+
+	gen_ctrl_val |= dcsChannelNumber << DCS_CHANNEL_NUMBER_POS;
+
+	/* Flip page order */
+	mdfld_wait_for_HS_DATA_FIFO(dev, pipe);
+	REG_WRITE(gen_data_reg, 0x00008036);
+	mdfld_wait_for_HS_CTRL_FIFO(dev, pipe);
+	REG_WRITE(gen_ctrl_reg, gen_ctrl_val | (0x02 << WORD_COUNTS_POS));
+
+	/* 0xF0 */
+	mdfld_wait_for_HS_DATA_FIFO(dev, pipe);
+	REG_WRITE(gen_data_reg, 0x005a5af0);
+	mdfld_wait_for_HS_CTRL_FIFO(dev, pipe);
+	REG_WRITE(gen_ctrl_reg, gen_ctrl_val | (0x03 << WORD_COUNTS_POS));
+
+	/* Write protection key */
+	mdfld_wait_for_HS_DATA_FIFO(dev, pipe);
+	REG_WRITE(gen_data_reg, 0x005a5af1);
+	mdfld_wait_for_HS_CTRL_FIFO(dev, pipe);
+	REG_WRITE(gen_ctrl_reg, gen_ctrl_val | (0x03 << WORD_COUNTS_POS));
+
+	/* 0xFC */
+	mdfld_wait_for_HS_DATA_FIFO(dev, pipe);
+	REG_WRITE(gen_data_reg, 0x005a5afc);
+	mdfld_wait_for_HS_CTRL_FIFO(dev, pipe);
+	REG_WRITE(gen_ctrl_reg, gen_ctrl_val | (0x03 << WORD_COUNTS_POS));
+
+	/* 0xB7 */
+	mdfld_wait_for_HS_DATA_FIFO(dev, pipe);
+	REG_WRITE(gen_data_reg, 0x770000b7);
+	mdfld_wait_for_HS_DATA_FIFO(dev, pipe);
+	REG_WRITE(gen_data_reg, 0x00000044);
+	mdfld_wait_for_HS_CTRL_FIFO(dev, pipe);
+	REG_WRITE(gen_ctrl_reg, gen_ctrl_val | (0x05 << WORD_COUNTS_POS));
+
+	/* 0xB6 */
+	mdfld_wait_for_HS_DATA_FIFO(dev, pipe);
+	REG_WRITE(gen_data_reg, 0x000a0ab6);
+	mdfld_wait_for_HS_CTRL_FIFO(dev, pipe);
+	REG_WRITE(gen_ctrl_reg, gen_ctrl_val | (0x03 << WORD_COUNTS_POS));
+
+	/* 0xF2 */
+	mdfld_wait_for_HS_DATA_FIFO(dev, pipe);
+	REG_WRITE(gen_data_reg, 0x081010f2);
+	mdfld_wait_for_HS_DATA_FIFO(dev, pipe);
+	REG_WRITE(gen_data_reg, 0x4a070708);
+	mdfld_wait_for_HS_DATA_FIFO(dev, pipe);
+	REG_WRITE(gen_data_reg, 0x000000c5);
+	mdfld_wait_for_HS_CTRL_FIFO(dev, pipe);
+	REG_WRITE(gen_ctrl_reg, gen_ctrl_val | (0x09 << WORD_COUNTS_POS));
+
+	/* 0xF8 */
+	mdfld_wait_for_HS_DATA_FIFO(dev, pipe);
+	REG_WRITE(gen_data_reg, 0x024003f8);
+	mdfld_wait_for_HS_DATA_FIFO(dev, pipe);
+	REG_WRITE(gen_data_reg, 0x01030a04);
+	mdfld_wait_for_HS_DATA_FIFO(dev, pipe);
+	REG_WRITE(gen_data_reg, 0x0e020220);
+	mdfld_wait_for_HS_DATA_FIFO(dev, pipe);
+	REG_WRITE(gen_data_reg, 0x00000004);
+	mdfld_wait_for_HS_CTRL_FIFO(dev, pipe);
+	REG_WRITE(gen_ctrl_reg, gen_ctrl_val | (0x0d << WORD_COUNTS_POS));
+
+	/* 0xE2 */
+	mdfld_wait_for_HS_DATA_FIFO(dev, pipe);
+	REG_WRITE(gen_data_reg, 0x398fc3e2);
+	mdfld_wait_for_HS_DATA_FIFO(dev, pipe);
+	REG_WRITE(gen_data_reg, 0x0000916f);
+	mdfld_wait_for_HS_CTRL_FIFO(dev, pipe);
+	REG_WRITE(gen_ctrl_reg, gen_ctrl_val | (0x06 << WORD_COUNTS_POS));
+
+	/* 0xB0 */
+	mdfld_wait_for_HS_DATA_FIFO(dev, pipe);
+	REG_WRITE(gen_data_reg, 0x000000b0);
+	mdfld_wait_for_HS_CTRL_FIFO(dev, pipe);
+	REG_WRITE(gen_ctrl_reg, gen_ctrl_val | (0x02 << WORD_COUNTS_POS));
+
+	/* 0xF4 */
+	mdfld_wait_for_HS_DATA_FIFO(dev, pipe);
+	REG_WRITE(gen_data_reg, 0x240242f4);
+	mdfld_wait_for_HS_DATA_FIFO(dev, pipe);
+	REG_WRITE(gen_data_reg, 0x78ee2002);
+	mdfld_wait_for_HS_DATA_FIFO(dev, pipe);
+	REG_WRITE(gen_data_reg, 0x2a071050);
+	mdfld_wait_for_HS_DATA_FIFO(dev, pipe);
+	REG_WRITE(gen_data_reg, 0x507fee10);
+	mdfld_wait_for_HS_DATA_FIFO(dev, pipe);
+	REG_WRITE(gen_data_reg, 0x10300710);
+	mdfld_wait_for_HS_CTRL_FIFO(dev, pipe);
+	REG_WRITE(gen_ctrl_reg, gen_ctrl_val | (0x14 << WORD_COUNTS_POS));
+
+	/* 0xBA */
+	mdfld_wait_for_HS_DATA_FIFO(dev, pipe);
+	REG_WRITE(gen_data_reg, 0x19fe07ba);
+	mdfld_wait_for_HS_DATA_FIFO(dev, pipe);
+	REG_WRITE(gen_data_reg, 0x101c0a31);
+	mdfld_wait_for_HS_DATA_FIFO(dev, pipe);
+	REG_WRITE(gen_data_reg, 0x00000010);
+	mdfld_wait_for_HS_CTRL_FIFO(dev, pipe);
+	REG_WRITE(gen_ctrl_reg, gen_ctrl_val | (0x09 << WORD_COUNTS_POS));
+
+	/* 0xBB */
+	mdfld_wait_for_HS_DATA_FIFO(dev, pipe);
+	REG_WRITE(gen_data_reg, 0x28ff07bb);
+	mdfld_wait_for_HS_DATA_FIFO(dev, pipe);
+	REG_WRITE(gen_data_reg, 0x24280a31);
+	mdfld_wait_for_HS_DATA_FIFO(dev, pipe);
+	REG_WRITE(gen_data_reg, 0x00000034);
+	mdfld_wait_for_HS_CTRL_FIFO(dev, pipe);
+	REG_WRITE(gen_ctrl_reg, gen_ctrl_val | (0x09 << WORD_COUNTS_POS));
+
+	/* 0xFB */
+	mdfld_wait_for_HS_DATA_FIFO(dev, pipe);
+	REG_WRITE(gen_data_reg, 0x535d05fb);
+	mdfld_wait_for_HS_DATA_FIFO(dev, pipe);
+	REG_WRITE(gen_data_reg, 0x1b1a2130);
+	mdfld_wait_for_HS_DATA_FIFO(dev, pipe);
+	REG_WRITE(gen_data_reg, 0x221e180e);
+	mdfld_wait_for_HS_DATA_FIFO(dev, pipe);
+	REG_WRITE(gen_data_reg, 0x131d2120);
+	mdfld_wait_for_HS_DATA_FIFO(dev, pipe);
+	REG_WRITE(gen_data_reg, 0x535d0508);
+	mdfld_wait_for_HS_DATA_FIFO(dev, pipe);
+	REG_WRITE(gen_data_reg, 0x1c1a2131);
+	mdfld_wait_for_HS_DATA_FIFO(dev, pipe);
+	REG_WRITE(gen_data_reg, 0x231f160d);
+	mdfld_wait_for_HS_DATA_FIFO(dev, pipe);
+	REG_WRITE(gen_data_reg, 0x111b2220);
+	mdfld_wait_for_HS_DATA_FIFO(dev, pipe);
+	REG_WRITE(gen_data_reg, 0x535c2008);
+	mdfld_wait_for_HS_DATA_FIFO(dev, pipe);
+	REG_WRITE(gen_data_reg, 0x1f1d2433);
+	mdfld_wait_for_HS_DATA_FIFO(dev, pipe);
+	REG_WRITE(gen_data_reg, 0x2c251a10);
+	mdfld_wait_for_HS_DATA_FIFO(dev, pipe);
+	REG_WRITE(gen_data_reg, 0x2c34372d);
+	mdfld_wait_for_HS_DATA_FIFO(dev, pipe);
+	REG_WRITE(gen_data_reg, 0x00000023);
+	mdfld_wait_for_HS_CTRL_FIFO(dev, pipe);
+	REG_WRITE(gen_ctrl_reg, gen_ctrl_val | (0x31 << WORD_COUNTS_POS));
+
+	/* 0xFA */
+	mdfld_wait_for_HS_DATA_FIFO(dev, pipe);
+	REG_WRITE(gen_data_reg, 0x525c0bfa);
+	mdfld_wait_for_HS_DATA_FIFO(dev, pipe);
+	REG_WRITE(gen_data_reg, 0x1c1c232f);
+	mdfld_wait_for_HS_DATA_FIFO(dev, pipe);
+	REG_WRITE(gen_data_reg, 0x2623190e);
+	mdfld_wait_for_HS_DATA_FIFO(dev, pipe);
+	REG_WRITE(gen_data_reg, 0x18212625);
+	mdfld_wait_for_HS_DATA_FIFO(dev, pipe);
+	REG_WRITE(gen_data_reg, 0x545d0d0e);
+	mdfld_wait_for_HS_DATA_FIFO(dev, pipe);
+	REG_WRITE(gen_data_reg, 0x1e1d2333);
+	mdfld_wait_for_HS_DATA_FIFO(dev, pipe);
+	REG_WRITE(gen_data_reg, 0x26231a10);
+	mdfld_wait_for_HS_DATA_FIFO(dev, pipe);
+	REG_WRITE(gen_data_reg, 0x1a222725);
+	mdfld_wait_for_HS_DATA_FIFO(dev, pipe);
+	REG_WRITE(gen_data_reg, 0x545d280f);
+	mdfld_wait_for_HS_DATA_FIFO(dev, pipe);
+	REG_WRITE(gen_data_reg, 0x21202635);
+	mdfld_wait_for_HS_DATA_FIFO(dev, pipe);
+	REG_WRITE(gen_data_reg, 0x31292013);
+	mdfld_wait_for_HS_DATA_FIFO(dev, pipe);
+	REG_WRITE(gen_data_reg, 0x31393d33);
+	mdfld_wait_for_HS_DATA_FIFO(dev, pipe);
+	REG_WRITE(gen_data_reg, 0x00000029);
+	mdfld_wait_for_HS_CTRL_FIFO(dev, pipe);
+	REG_WRITE(gen_ctrl_reg, gen_ctrl_val | (0x31 << WORD_COUNTS_POS));
+
+	/* Set DM */
+	mdfld_wait_for_HS_DATA_FIFO(dev, pipe);
+	REG_WRITE(gen_data_reg, 0x000100f7);
+	mdfld_wait_for_HS_CTRL_FIFO(dev, pipe);
+	REG_WRITE(gen_ctrl_reg, gen_ctrl_val | (0x03 << WORD_COUNTS_POS));
+}
+
+static u16 mdfld_dsi_dpi_to_byte_clock_count(int pixel_clock_count,
+						int num_lane, int bpp)
+{
+	return (u16)((pixel_clock_count * bpp) / (num_lane * 8));
+}
+
+/*
+ * Calculate the dpi time basing on a given drm mode @mode
+ * return 0 on success.
+ * FIXME: I was using proposed mode value for calculation, may need to
+ * use crtc mode values later
+ */
+int mdfld_dsi_dpi_timing_calculation(struct drm_display_mode *mode,
+				struct mdfld_dsi_dpi_timing *dpi_timing,
+				int num_lane, int bpp)
+{
+	int pclk_hsync, pclk_hfp, pclk_hbp, pclk_hactive;
+	int pclk_vsync, pclk_vfp, pclk_vbp;
+
+	pclk_hactive = mode->hdisplay;
+	pclk_hfp = mode->hsync_start - mode->hdisplay;
+	pclk_hsync = mode->hsync_end - mode->hsync_start;
+	pclk_hbp = mode->htotal - mode->hsync_end;
+
+	pclk_vfp = mode->vsync_start - mode->vdisplay;
+	pclk_vsync = mode->vsync_end - mode->vsync_start;
+	pclk_vbp = mode->vtotal - mode->vsync_end;
+
+	/*
+	 * byte clock counts were calculated by following formula
+	 * bclock_count = pclk_count * bpp / num_lane / 8
+	 */
+	dpi_timing->hsync_count = mdfld_dsi_dpi_to_byte_clock_count(
+						pclk_hsync, num_lane, bpp);
+	dpi_timing->hbp_count = mdfld_dsi_dpi_to_byte_clock_count(
+						pclk_hbp, num_lane, bpp);
+	dpi_timing->hfp_count = mdfld_dsi_dpi_to_byte_clock_count(
+						pclk_hfp, num_lane, bpp);
+	dpi_timing->hactive_count = mdfld_dsi_dpi_to_byte_clock_count(
+						pclk_hactive, num_lane, bpp);
+	dpi_timing->vsync_count = mdfld_dsi_dpi_to_byte_clock_count(
+						pclk_vsync, num_lane, bpp);
+	dpi_timing->vbp_count = mdfld_dsi_dpi_to_byte_clock_count(
+						pclk_vbp, num_lane, bpp);
+	dpi_timing->vfp_count = mdfld_dsi_dpi_to_byte_clock_count(
+						pclk_vfp, num_lane, bpp);
+
+	return 0;
+}
+
+void mdfld_dsi_dpi_controller_init(struct mdfld_dsi_config *dsi_config,
+								int pipe)
+{
+	struct drm_device *dev = dsi_config->dev;
+	int lane_count = dsi_config->lane_count;
+	struct mdfld_dsi_dpi_timing dpi_timing;
+	struct drm_display_mode *mode = dsi_config->mode;
+	u32 val;
+
+	/*un-ready device*/
+	REG_FLD_MOD(MIPI_DEVICE_READY_REG(pipe), 0, 0, 0);
+
+	/*init dsi adapter before kicking off*/
+	REG_WRITE(MIPI_CTRL_REG(pipe), 0x00000018);
+
+	/*enable all interrupts*/
+	REG_WRITE(MIPI_INTR_EN_REG(pipe), 0xffffffff);
+
+	/*set up func_prg*/
+	val = lane_count;
+	val |= dsi_config->channel_num << DSI_DPI_VIRT_CHANNEL_OFFSET;
+
+	switch (dsi_config->bpp) {
+	case 16:
+		val |= DSI_DPI_COLOR_FORMAT_RGB565;
+		break;
+	case 18:
+		val |= DSI_DPI_COLOR_FORMAT_RGB666;
+		break;
+	case 24:
+		val |= DSI_DPI_COLOR_FORMAT_RGB888;
+		break;
+	default:
+		DRM_ERROR("unsupported color format, bpp = %d\n",
+							dsi_config->bpp);
+	}
+	REG_WRITE(MIPI_DSI_FUNC_PRG_REG(pipe), val);
+
+	REG_WRITE(MIPI_HS_TX_TIMEOUT_REG(pipe),
+			(mode->vtotal * mode->htotal * dsi_config->bpp /
+				(8 * lane_count)) & DSI_HS_TX_TIMEOUT_MASK);
+	REG_WRITE(MIPI_LP_RX_TIMEOUT_REG(pipe),
+				0xffff & DSI_LP_RX_TIMEOUT_MASK);
+
+	/*max value: 20 clock cycles of txclkesc*/
+	REG_WRITE(MIPI_TURN_AROUND_TIMEOUT_REG(pipe),
+				0x14 & DSI_TURN_AROUND_TIMEOUT_MASK);
+
+	/*min 21 txclkesc, max: ffffh*/
+	REG_WRITE(MIPI_DEVICE_RESET_TIMER_REG(pipe),
+				0xffff & DSI_RESET_TIMER_MASK);
+
+	REG_WRITE(MIPI_DPI_RESOLUTION_REG(pipe),
+				mode->vdisplay << 16 | mode->hdisplay);
+
+	/*set DPI timing registers*/
+	mdfld_dsi_dpi_timing_calculation(mode, &dpi_timing,
+				dsi_config->lane_count, dsi_config->bpp);
+
+	REG_WRITE(MIPI_HSYNC_COUNT_REG(pipe),
+			dpi_timing.hsync_count & DSI_DPI_TIMING_MASK);
+	REG_WRITE(MIPI_HBP_COUNT_REG(pipe),
+			dpi_timing.hbp_count & DSI_DPI_TIMING_MASK);
+	REG_WRITE(MIPI_HFP_COUNT_REG(pipe),
+			dpi_timing.hfp_count & DSI_DPI_TIMING_MASK);
+	REG_WRITE(MIPI_HACTIVE_COUNT_REG(pipe),
+			dpi_timing.hactive_count & DSI_DPI_TIMING_MASK);
+	REG_WRITE(MIPI_VSYNC_COUNT_REG(pipe),
+			dpi_timing.vsync_count & DSI_DPI_TIMING_MASK);
+	REG_WRITE(MIPI_VBP_COUNT_REG(pipe),
+			dpi_timing.vbp_count & DSI_DPI_TIMING_MASK);
+	REG_WRITE(MIPI_VFP_COUNT_REG(pipe),
+			dpi_timing.vfp_count & DSI_DPI_TIMING_MASK);
+
+	REG_WRITE(MIPI_HIGH_LOW_SWITCH_COUNT_REG(pipe), 0x46);
+
+	/*min: 7d0 max: 4e20*/
+	REG_WRITE(MIPI_INIT_COUNT_REG(pipe), 0x000007d0);
+
+	/*set up video mode*/
+	val = dsi_config->video_mode | DSI_DPI_COMPLETE_LAST_LINE;
+	REG_WRITE(MIPI_VIDEO_MODE_FORMAT_REG(pipe), val);
+
+	REG_WRITE(MIPI_EOT_DISABLE_REG(pipe), 0x00000000);
+
+	REG_WRITE(MIPI_LP_BYTECLK_REG(pipe), 0x00000004);
+
+	/*TODO: figure out how to setup these registers*/
+	if (mdfld_get_panel_type(dev, pipe) == TC35876X)
+		REG_WRITE(MIPI_DPHY_PARAM_REG(pipe), 0x2A0c6008);
+	else
+		REG_WRITE(MIPI_DPHY_PARAM_REG(pipe), 0x150c3408);
+
+	REG_WRITE(MIPI_CLK_LANE_SWITCH_TIME_CNT_REG(pipe), (0xa << 16) | 0x14);
+
+	if (mdfld_get_panel_type(dev, pipe) == TC35876X)
+		tc35876x_set_bridge_reset_state(dev, 0);  /*Pull High Reset */
+
+	/*set device ready*/
+	REG_FLD_MOD(MIPI_DEVICE_READY_REG(pipe), 1, 0, 0);
+}
+
+void mdfld_dsi_dpi_turn_on(struct mdfld_dsi_dpi_output *output, int pipe)
+{
+	struct drm_device *dev = output->dev;
+
+	/* clear special packet sent bit */
+	if (REG_READ(MIPI_INTR_STAT_REG(pipe)) & DSI_INTR_STATE_SPL_PKG_SENT)
+		REG_WRITE(MIPI_INTR_STAT_REG(pipe),
+					DSI_INTR_STATE_SPL_PKG_SENT);
+
+	/*send turn on package*/
+	REG_WRITE(MIPI_DPI_CONTROL_REG(pipe), DSI_DPI_CTRL_HS_TURN_ON);
+
+	/*wait for SPL_PKG_SENT interrupt*/
+	mdfld_wait_for_SPL_PKG_SENT(dev, pipe);
+
+	if (REG_READ(MIPI_INTR_STAT_REG(pipe)) & DSI_INTR_STATE_SPL_PKG_SENT)
+		REG_WRITE(MIPI_INTR_STAT_REG(pipe),
+					DSI_INTR_STATE_SPL_PKG_SENT);
+
+	output->panel_on = 1;
+
+	/* FIXME the following is disabled to WA the X slow start issue
+	   for TMD panel
+	if (pipe == 2)
+		dev_priv->dpi_panel_on2 = true;
+	else if (pipe == 0)
+		dev_priv->dpi_panel_on = true; */
+}
+
+static void mdfld_dsi_dpi_shut_down(struct mdfld_dsi_dpi_output *output,
+								int pipe)
+{
+	struct drm_device *dev = output->dev;
+
+	/*if output is on, or mode setting didn't happen, ignore this*/
+	if ((!output->panel_on) || output->first_boot) {
+		output->first_boot = 0;
+		return;
+	}
+
+	/* Wait for dpi fifo to empty */
+	mdfld_wait_for_DPI_CTRL_FIFO(dev, pipe);
+
+	/* Clear the special packet interrupt bit if set */
+	if (REG_READ(MIPI_INTR_STAT_REG(pipe)) & DSI_INTR_STATE_SPL_PKG_SENT)
+		REG_WRITE(MIPI_INTR_STAT_REG(pipe),
+					DSI_INTR_STATE_SPL_PKG_SENT);
+
+	if (REG_READ(MIPI_DPI_CONTROL_REG(pipe)) == DSI_DPI_CTRL_HS_SHUTDOWN)
+		goto shutdown_out;
+
+	REG_WRITE(MIPI_DPI_CONTROL_REG(pipe), DSI_DPI_CTRL_HS_SHUTDOWN);
+
+shutdown_out:
+	output->panel_on = 0;
+	output->first_boot = 0;
+
+	/* FIXME the following is disabled to WA the X slow start issue
+	   for TMD panel
+	if (pipe == 2)
+		dev_priv->dpi_panel_on2 = false;
+	else if (pipe == 0)
+		dev_priv->dpi_panel_on = false;	 */
+}
+
+static void mdfld_dsi_dpi_set_power(struct drm_encoder *encoder, bool on)
+{
+	struct mdfld_dsi_encoder *dsi_encoder = mdfld_dsi_encoder(encoder);
+	struct mdfld_dsi_dpi_output *dpi_output =
+				MDFLD_DSI_DPI_OUTPUT(dsi_encoder);
+	struct mdfld_dsi_config *dsi_config =
+				mdfld_dsi_encoder_get_config(dsi_encoder);
+	int pipe = mdfld_dsi_encoder_get_pipe(dsi_encoder);
+	struct drm_device *dev = dsi_config->dev;
+	struct drm_psb_private *dev_priv = dev->dev_private;
+
+	/*start up display island if it was shutdown*/
+	if (!gma_power_begin(dev, true))
+		return;
+
+	if (on) {
+		if (mdfld_get_panel_type(dev, pipe) == TMD_VID)
+			mdfld_dsi_dpi_turn_on(dpi_output, pipe);
+		else if (mdfld_get_panel_type(dev, pipe) == TC35876X)
+			mdfld_dsi_configure_up(dsi_encoder, pipe);
+		else {
+			/*enable mipi port*/
+			REG_WRITE(MIPI_PORT_CONTROL(pipe),
+				REG_READ(MIPI_PORT_CONTROL(pipe)) | BIT(31));
+			REG_READ(MIPI_PORT_CONTROL(pipe));
+
+			mdfld_dsi_dpi_turn_on(dpi_output, pipe);
+			mdfld_dsi_tpo_ic_init(dsi_config, pipe);
+		}
+		dev_priv->dpi_panel_on[pipe] = true;
+	} else {
+		if (mdfld_get_panel_type(dev, pipe) == TMD_VID)
+			mdfld_dsi_dpi_shut_down(dpi_output, pipe);
+		else if (mdfld_get_panel_type(dev, pipe) == TC35876X)
+			mdfld_dsi_configure_down(dsi_encoder, pipe);
+		else {
+			mdfld_dsi_dpi_shut_down(dpi_output, pipe);
+
+			/*disable mipi port*/
+			REG_WRITE(MIPI_PORT_CONTROL(pipe),
+				REG_READ(MIPI_PORT_CONTROL(pipe)) & ~BIT(31));
+			REG_READ(MIPI_PORT_CONTROL(pipe));
+		}
+		dev_priv->dpi_panel_on[pipe] = false;
+	}
+	gma_power_end(dev);
+}
+
+void mdfld_dsi_dpi_dpms(struct drm_encoder *encoder, int mode)
+{
+	mdfld_dsi_dpi_set_power(encoder, mode == DRM_MODE_DPMS_ON);
+}
+
+bool mdfld_dsi_dpi_mode_fixup(struct drm_encoder *encoder,
+				     struct drm_display_mode *mode,
+				     struct drm_display_mode *adjusted_mode)
+{
+	struct mdfld_dsi_encoder *dsi_encoder = mdfld_dsi_encoder(encoder);
+	struct mdfld_dsi_config *dsi_config =
+				mdfld_dsi_encoder_get_config(dsi_encoder);
+	struct drm_display_mode *fixed_mode = dsi_config->fixed_mode;
+
+	if (fixed_mode) {
+		adjusted_mode->hdisplay = fixed_mode->hdisplay;
+		adjusted_mode->hsync_start = fixed_mode->hsync_start;
+		adjusted_mode->hsync_end = fixed_mode->hsync_end;
+		adjusted_mode->htotal = fixed_mode->htotal;
+		adjusted_mode->vdisplay = fixed_mode->vdisplay;
+		adjusted_mode->vsync_start = fixed_mode->vsync_start;
+		adjusted_mode->vsync_end = fixed_mode->vsync_end;
+		adjusted_mode->vtotal = fixed_mode->vtotal;
+		adjusted_mode->clock = fixed_mode->clock;
+		drm_mode_set_crtcinfo(adjusted_mode, CRTC_INTERLACE_HALVE_V);
+	}
+	return true;
+}
+
+void mdfld_dsi_dpi_prepare(struct drm_encoder *encoder)
+{
+	mdfld_dsi_dpi_set_power(encoder, false);
+}
+
+void mdfld_dsi_dpi_commit(struct drm_encoder *encoder)
+{
+	mdfld_dsi_dpi_set_power(encoder, true);
+}
+
+/* For TC35876X */
+/* This functionality was implemented in FW in iCDK */
+/* But removed in DV0 and later. So need to add here. */
+static void mipi_set_properties(struct mdfld_dsi_config *dsi_config, int pipe)
+{
+	struct drm_device *dev = dsi_config->dev;
+
+	REG_WRITE(MIPI_CTRL_REG(pipe), 0x00000018);
+	REG_WRITE(MIPI_INTR_EN_REG(pipe), 0xffffffff);
+	REG_WRITE(MIPI_HS_TX_TIMEOUT_REG(pipe), 0xffffff);
+	REG_WRITE(MIPI_LP_RX_TIMEOUT_REG(pipe), 0xffffff);
+	REG_WRITE(MIPI_TURN_AROUND_TIMEOUT_REG(pipe), 0x14);
+	REG_WRITE(MIPI_DEVICE_RESET_TIMER_REG(pipe), 0xff);
+	REG_WRITE(MIPI_HIGH_LOW_SWITCH_COUNT_REG(pipe), 0x25);
+	REG_WRITE(MIPI_INIT_COUNT_REG(pipe), 0xf0);
+	REG_WRITE(MIPI_EOT_DISABLE_REG(pipe), 0x00000000);
+	REG_WRITE(MIPI_LP_BYTECLK_REG(pipe), 0x00000004);
+	REG_WRITE(MIPI_DBI_BW_CTRL_REG(pipe), 0x00000820);
+	REG_WRITE(MIPI_CLK_LANE_SWITCH_TIME_CNT_REG(pipe), (0xa << 16) | 0x14);
+}
+
+static void mdfld_mipi_set_video_timing(struct mdfld_dsi_config *dsi_config,
+					int pipe)
+{
+	struct drm_device *dev = dsi_config->dev;
+	struct mdfld_dsi_dpi_timing dpi_timing;
+	struct drm_display_mode *mode = dsi_config->mode;
+
+	mdfld_dsi_dpi_timing_calculation(mode, &dpi_timing,
+					dsi_config->lane_count,
+					dsi_config->bpp);
+
+	REG_WRITE(MIPI_DPI_RESOLUTION_REG(pipe),
+		mode->vdisplay << 16 | mode->hdisplay);
+	REG_WRITE(MIPI_HSYNC_COUNT_REG(pipe),
+		dpi_timing.hsync_count & DSI_DPI_TIMING_MASK);
+	REG_WRITE(MIPI_HBP_COUNT_REG(pipe),
+		dpi_timing.hbp_count & DSI_DPI_TIMING_MASK);
+	REG_WRITE(MIPI_HFP_COUNT_REG(pipe),
+		dpi_timing.hfp_count & DSI_DPI_TIMING_MASK);
+	REG_WRITE(MIPI_HACTIVE_COUNT_REG(pipe),
+		dpi_timing.hactive_count & DSI_DPI_TIMING_MASK);
+	REG_WRITE(MIPI_VSYNC_COUNT_REG(pipe),
+		dpi_timing.vsync_count & DSI_DPI_TIMING_MASK);
+	REG_WRITE(MIPI_VBP_COUNT_REG(pipe),
+		dpi_timing.vbp_count & DSI_DPI_TIMING_MASK);
+	REG_WRITE(MIPI_VFP_COUNT_REG(pipe),
+		dpi_timing.vfp_count & DSI_DPI_TIMING_MASK);
+}
+
+static void mdfld_mipi_config(struct mdfld_dsi_config *dsi_config, int pipe)
+{
+	struct drm_device *dev = dsi_config->dev;
+	int lane_count = dsi_config->lane_count;
+
+	if (pipe) {
+		REG_WRITE(MIPI_PORT_CONTROL(0), 0x00000002);
+		REG_WRITE(MIPI_PORT_CONTROL(2), 0x80000000);
+	} else {
+		REG_WRITE(MIPI_PORT_CONTROL(0), 0x80010000);
+		REG_WRITE(MIPI_PORT_CONTROL(2), 0x00);
+	}
+
+	REG_WRITE(MIPI_DPHY_PARAM_REG(pipe), 0x150A600F);
+	REG_WRITE(MIPI_VIDEO_MODE_FORMAT_REG(pipe), 0x0000000F);
+
+	/* lane_count = 3 */
+	REG_WRITE(MIPI_DSI_FUNC_PRG_REG(pipe), 0x00000200 | lane_count);
+
+	mdfld_mipi_set_video_timing(dsi_config, pipe);
+}
+
+static void mdfld_set_pipe_timing(struct mdfld_dsi_config *dsi_config, int pipe)
+{
+	struct drm_device *dev = dsi_config->dev;
+	struct drm_display_mode *mode = dsi_config->mode;
+
+	REG_WRITE(HTOTAL_A, ((mode->htotal - 1) << 16) | (mode->hdisplay - 1));
+	REG_WRITE(HBLANK_A, ((mode->htotal - 1) << 16) | (mode->hdisplay - 1));
+	REG_WRITE(HSYNC_A,
+		((mode->hsync_end - 1) << 16) | (mode->hsync_start - 1));
+
+	REG_WRITE(VTOTAL_A, ((mode->vtotal - 1) << 16) | (mode->vdisplay - 1));
+	REG_WRITE(VBLANK_A, ((mode->vtotal - 1) << 16) | (mode->vdisplay - 1));
+	REG_WRITE(VSYNC_A,
+		((mode->vsync_end - 1) << 16) | (mode->vsync_start - 1));
+
+	REG_WRITE(PIPEASRC,
+		((mode->hdisplay - 1) << 16) | (mode->vdisplay - 1));
+}
+/* End for TC35876X */
+
+void mdfld_dsi_dpi_mode_set(struct drm_encoder *encoder,
+				   struct drm_display_mode *mode,
+				   struct drm_display_mode *adjusted_mode)
+{
+	struct mdfld_dsi_encoder *dsi_encoder = mdfld_dsi_encoder(encoder);
+	struct mdfld_dsi_dpi_output *dpi_output =
+					MDFLD_DSI_DPI_OUTPUT(dsi_encoder);
+	struct mdfld_dsi_config *dsi_config =
+				mdfld_dsi_encoder_get_config(dsi_encoder);
+	struct drm_device *dev = dsi_config->dev;
+	struct drm_psb_private *dev_priv = dev->dev_private;
+	int pipe = mdfld_dsi_encoder_get_pipe(dsi_encoder);
+
+	u32 pipeconf_reg = PIPEACONF;
+	u32 dspcntr_reg = DSPACNTR;
+
+	u32 pipeconf = dev_priv->pipeconf[pipe];
+	u32 dspcntr = dev_priv->dspcntr[pipe];
+	u32 mipi = MIPI_PORT_EN | PASS_FROM_SPHY_TO_AFE | SEL_FLOPPED_HSTX;
+
+	if (pipe) {
+		pipeconf_reg = PIPECCONF;
+		dspcntr_reg = DSPCCNTR;
+	} else {
+		if (mdfld_get_panel_type(dev, pipe) == TC35876X)
+			mipi &= (~0x03); /* Use all four lanes */
+		else
+			mipi |= 2;
+	}
+
+	/*start up display island if it was shutdown*/
+	if (!gma_power_begin(dev, true))
+		return;
+
+	if (mdfld_get_panel_type(dev, pipe) == TC35876X) {
+		/*
+		 * The following logic is required to reset the bridge and
+		 * configure. This also starts the DSI clock at 200MHz.
+		 */
+		tc35876x_set_bridge_reset_state(dev, 0);  /*Pull High Reset */
+		tc35876x_toshiba_bridge_panel_on(dev);
+		udelay(100);
+		/* Now start the DSI clock */
+		REG_WRITE(MRST_DPLL_A, 0x00);
+		REG_WRITE(MRST_FPA0, 0xC1);
+		REG_WRITE(MRST_DPLL_A, 0x00800000);
+		udelay(500);
+		REG_WRITE(MRST_DPLL_A, 0x80800000);
+
+		if (REG_BIT_WAIT(pipeconf_reg, 1, 29))
+			dev_err(&dev->pdev->dev, "%s: DSI PLL lock timeout\n",
+				__func__);
+
+		REG_WRITE(MIPI_DPHY_PARAM_REG(pipe), 0x2A0c6008);
+
+		mipi_set_properties(dsi_config, pipe);
+		mdfld_mipi_config(dsi_config, pipe);
+		mdfld_set_pipe_timing(dsi_config, pipe);
+
+		REG_WRITE(DSPABASE, 0x00);
+		REG_WRITE(DSPASTRIDE, (mode->hdisplay * 4));
+		REG_WRITE(DSPASIZE,
+			((mode->vdisplay - 1) << 16) | (mode->hdisplay - 1));
+
+		REG_WRITE(DSPACNTR, 0x98000000);
+		REG_WRITE(DSPASURF, 0x00);
+
+		REG_WRITE(VGACNTRL, 0x80000000);
+		REG_WRITE(DEVICE_READY_REG, 0x00000001);
+
+		REG_WRITE(MIPI_PORT_CONTROL(pipe), 0x80810000);
+	} else {
+		/*set up mipi port FIXME: do at init time */
+		REG_WRITE(MIPI_PORT_CONTROL(pipe), mipi);
+	}
+	REG_READ(MIPI_PORT_CONTROL(pipe));
+
+	if (mdfld_get_panel_type(dev, pipe) == TMD_VID) {
+		/* NOP */
+	} else if (mdfld_get_panel_type(dev, pipe) == TC35876X) {
+		/* set up DSI controller DPI interface */
+		mdfld_dsi_dpi_controller_init(dsi_config, pipe);
+
+		/* Configure MIPI Bridge and Panel */
+		tc35876x_configure_lvds_bridge(dev);
+		dev_priv->dpi_panel_on[pipe] = true;
+	} else {
+		/*turn on DPI interface*/
+		mdfld_dsi_dpi_turn_on(dpi_output, pipe);
+	}
+
+	/*set up pipe*/
+	REG_WRITE(pipeconf_reg, pipeconf);
+	REG_READ(pipeconf_reg);
+
+	/*set up display plane*/
+	REG_WRITE(dspcntr_reg, dspcntr);
+	REG_READ(dspcntr_reg);
+
+	msleep(20); /* FIXME: this should wait for vblank */
+
+	if (mdfld_get_panel_type(dev, pipe) == TMD_VID) {
+		/* NOP */
+	} else if (mdfld_get_panel_type(dev, pipe) == TC35876X) {
+		mdfld_dsi_dpi_turn_on(dpi_output, pipe);
+	} else {
+		/* init driver ic */
+		mdfld_dsi_tpo_ic_init(dsi_config, pipe);
+		/*init backlight*/
+		mdfld_dsi_brightness_init(dsi_config, pipe);
+	}
+
+	gma_power_end(dev);
+}
+
+/*
+ * Init DSI DPI encoder.
+ * Allocate an mdfld_dsi_encoder and attach it to given @dsi_connector
+ * return pointer of newly allocated DPI encoder, NULL on error
+ */
+struct mdfld_dsi_encoder *mdfld_dsi_dpi_init(struct drm_device *dev,
+				struct mdfld_dsi_connector *dsi_connector,
+				const struct panel_funcs *p_funcs)
+{
+	struct mdfld_dsi_dpi_output *dpi_output = NULL;
+	struct mdfld_dsi_config *dsi_config;
+	struct drm_connector *connector = NULL;
+	struct drm_encoder *encoder = NULL;
+	int pipe;
+	u32 data;
+	int ret;
+
+	pipe = dsi_connector->pipe;
+
+	if (mdfld_get_panel_type(dev, pipe) != TC35876X) {
+		dsi_config = mdfld_dsi_get_config(dsi_connector);
+
+		/* panel hard-reset */
+		if (p_funcs->reset) {
+			ret = p_funcs->reset(pipe);
+			if (ret) {
+				DRM_ERROR("Panel %d hard-reset failed\n", pipe);
+				return NULL;
+			}
+		}
+
+		/* panel drvIC init */
+		if (p_funcs->drv_ic_init)
+			p_funcs->drv_ic_init(dsi_config, pipe);
+
+		/* panel power mode detect */
+		ret = mdfld_dsi_get_power_mode(dsi_config, &data, false);
+		if (ret) {
+			DRM_ERROR("Panel %d get power mode failed\n", pipe);
+			dsi_connector->status = connector_status_disconnected;
+		} else {
+			DRM_INFO("pipe %d power mode 0x%x\n", pipe, data);
+			dsi_connector->status = connector_status_connected;
+		}
+	}
+
+	dpi_output = kzalloc(sizeof(struct mdfld_dsi_dpi_output), GFP_KERNEL);
+	if (!dpi_output) {
+		DRM_ERROR("No memory\n");
+		return NULL;
+	}
+
+	if (dsi_connector->pipe)
+		dpi_output->panel_on = 0;
+	else
+		dpi_output->panel_on = 0;
+
+	dpi_output->dev = dev;
+	if (mdfld_get_panel_type(dev, pipe) != TC35876X)
+		dpi_output->p_funcs = p_funcs;
+	dpi_output->first_boot = 1;
+
+	/*get fixed mode*/
+	dsi_config = mdfld_dsi_get_config(dsi_connector);
+
+	/*create drm encoder object*/
+	connector = &dsi_connector->base.base;
+	encoder = &dpi_output->base.base.base;
+	drm_encoder_init(dev,
+			encoder,
+			p_funcs->encoder_funcs,
+			DRM_MODE_ENCODER_LVDS);
+	drm_encoder_helper_add(encoder,
+				p_funcs->encoder_helper_funcs);
+
+	/*attach to given connector*/
+	drm_mode_connector_attach_encoder(connector, encoder);
+
+	/*set possible crtcs and clones*/
+	if (dsi_connector->pipe) {
+		encoder->possible_crtcs = (1 << 2);
+		encoder->possible_clones = (1 << 1);
+	} else {
+		encoder->possible_crtcs = (1 << 0);
+		encoder->possible_clones = (1 << 0);
+	}
+
+	dsi_connector->base.encoder = &dpi_output->base.base;
+
+	return &dpi_output->base;
+}
diff --git a/drivers/gpu/drm/gma500/mdfld_dsi_dpi.h b/drivers/gpu/drm/gma500/mdfld_dsi_dpi.h
new file mode 100644
index 000000000000..6f762478b959
--- /dev/null
+++ b/drivers/gpu/drm/gma500/mdfld_dsi_dpi.h
@@ -0,0 +1,79 @@
+/*
+ * Copyright © 2010 Intel Corporation
+ *
+ * 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:
+ * jim liu <jim.liu@intel.com>
+ * Jackie Li<yaodong.li@intel.com>
+ */
+
+#ifndef __MDFLD_DSI_DPI_H__
+#define __MDFLD_DSI_DPI_H__
+
+#include "mdfld_dsi_output.h"
+#include "mdfld_output.h"
+
+struct mdfld_dsi_dpi_timing {
+	u16 hsync_count;
+	u16 hbp_count;
+	u16 hfp_count;
+	u16 hactive_count;
+	u16 vsync_count;
+	u16 vbp_count;
+	u16 vfp_count;
+};
+
+struct mdfld_dsi_dpi_output {
+	struct mdfld_dsi_encoder base;
+	struct drm_device *dev;
+
+	int panel_on;
+	int first_boot;
+
+	const struct panel_funcs *p_funcs;
+};
+
+#define MDFLD_DSI_DPI_OUTPUT(dsi_encoder)\
+	container_of(dsi_encoder, struct mdfld_dsi_dpi_output, base)
+
+/* Export functions */
+extern int mdfld_dsi_dpi_timing_calculation(struct drm_display_mode *mode,
+				struct mdfld_dsi_dpi_timing *dpi_timing,
+				int num_lane, int bpp);
+extern struct mdfld_dsi_encoder *mdfld_dsi_dpi_init(struct drm_device *dev,
+				struct mdfld_dsi_connector *dsi_connector,
+				const struct panel_funcs *p_funcs);
+
+/* MDFLD DPI helper functions */
+extern void mdfld_dsi_dpi_dpms(struct drm_encoder *encoder, int mode);
+extern bool mdfld_dsi_dpi_mode_fixup(struct drm_encoder *encoder,
+				struct drm_display_mode *mode,
+				struct drm_display_mode *adjusted_mode);
+extern void mdfld_dsi_dpi_prepare(struct drm_encoder *encoder);
+extern void mdfld_dsi_dpi_commit(struct drm_encoder *encoder);
+extern void mdfld_dsi_dpi_mode_set(struct drm_encoder *encoder,
+				struct drm_display_mode *mode,
+				struct drm_display_mode *adjusted_mode);
+extern void mdfld_dsi_dpi_turn_on(struct mdfld_dsi_dpi_output *output,
+				int pipe);
+extern void mdfld_dsi_dpi_controller_init(struct mdfld_dsi_config *dsi_config,
+				int pipe);
+#endif /*__MDFLD_DSI_DPI_H__*/
diff --git a/drivers/gpu/drm/gma500/mdfld_dsi_output.c b/drivers/gpu/drm/gma500/mdfld_dsi_output.c
new file mode 100644
index 000000000000..5675d93b4205
--- /dev/null
+++ b/drivers/gpu/drm/gma500/mdfld_dsi_output.c
@@ -0,0 +1,621 @@
+/*
+ * Copyright © 2010 Intel Corporation
+ *
+ * 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:
+ * jim liu <jim.liu@intel.com>
+ * Jackie Li<yaodong.li@intel.com>
+ */
+
+#include <linux/module.h>
+
+#include "mdfld_dsi_output.h"
+#include "mdfld_dsi_dpi.h"
+#include "mdfld_output.h"
+#include "mdfld_dsi_pkg_sender.h"
+#include "tc35876x-dsi-lvds.h"
+#include <linux/pm_runtime.h>
+#include <asm/intel_scu_ipc.h>
+
+/* get the LABC from command line. */
+static int LABC_control = 1;
+
+#ifdef MODULE
+module_param(LABC_control, int, 0644);
+#else
+
+static int __init parse_LABC_control(char *arg)
+{
+	/* LABC control can be passed in as a cmdline parameter */
+	/* to enable this feature add LABC=1 to cmdline */
+	/* to disable this feature add LABC=0 to cmdline */
+	if (!arg)
+		return -EINVAL;
+
+	if (!strcasecmp(arg, "0"))
+		LABC_control = 0;
+	else if (!strcasecmp(arg, "1"))
+		LABC_control = 1;
+
+	return 0;
+}
+early_param("LABC", parse_LABC_control);
+#endif
+
+/**
+ * Check and see if the generic control or data buffer is empty and ready.
+ */
+void mdfld_dsi_gen_fifo_ready(struct drm_device *dev, u32 gen_fifo_stat_reg,
+							u32 fifo_stat)
+{
+	u32 GEN_BF_time_out_count;
+
+	/* Check MIPI Adatper command registers */
+	for (GEN_BF_time_out_count = 0;
+			GEN_BF_time_out_count < GEN_FB_TIME_OUT;
+			GEN_BF_time_out_count++) {
+		if ((REG_READ(gen_fifo_stat_reg) & fifo_stat) == fifo_stat)
+			break;
+		udelay(100);
+	}
+
+	if (GEN_BF_time_out_count == GEN_FB_TIME_OUT)
+		DRM_ERROR("mdfld_dsi_gen_fifo_ready, Timeout. gen_fifo_stat_reg = 0x%x.\n",
+					gen_fifo_stat_reg);
+}
+
+/**
+ * Manage the DSI MIPI keyboard and display brightness.
+ * FIXME: this is exported to OSPM code. should work out an specific
+ * display interface to OSPM.
+ */
+
+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;
+	u32 gen_ctrl_val;
+
+	if (!sender) {
+		DRM_ERROR("No sender found\n");
+		return;
+	}
+
+	/* Set default display backlight value to 85% (0xd8)*/
+	mdfld_dsi_send_mcs_short(sender, write_display_brightness, 0xd8, 1,
+				true);
+
+	/* Set minimum brightness setting of CABC function to 20% (0x33)*/
+	mdfld_dsi_send_mcs_short(sender, write_cabc_min_bright, 0x33, 1, true);
+
+	/* Enable backlight or/and LABC */
+	gen_ctrl_val = BRIGHT_CNTL_BLOCK_ON | DISPLAY_DIMMING_ON |
+								BACKLIGHT_ON;
+	if (LABC_control == 1)
+		gen_ctrl_val |= DISPLAY_DIMMING_ON | DISPLAY_BRIGHTNESS_AUTO
+								| GAMMA_AUTO;
+
+	if (LABC_control == 1)
+		gen_ctrl_val |= AMBIENT_LIGHT_SENSE_ON;
+
+	dev_priv->mipi_ctrl_display = gen_ctrl_val;
+
+	mdfld_dsi_send_mcs_short(sender, write_ctrl_display, (u8)gen_ctrl_val,
+				1, true);
+
+	mdfld_dsi_send_mcs_short(sender, write_ctrl_cabc, UI_IMAGE, 1, true);
+}
+
+void mdfld_dsi_brightness_control(struct drm_device *dev, int pipe, int level)
+{
+	struct mdfld_dsi_pkg_sender *sender;
+	struct drm_psb_private *dev_priv;
+	struct mdfld_dsi_config *dsi_config;
+	u32 gen_ctrl_val = 0;
+	int p_type = TMD_VID;
+
+	if (!dev || (pipe != 0 && pipe != 2)) {
+		DRM_ERROR("Invalid parameter\n");
+		return;
+	}
+
+	p_type = mdfld_get_panel_type(dev, 0);
+
+	dev_priv = dev->dev_private;
+
+	if (pipe)
+		dsi_config = dev_priv->dsi_configs[1];
+	else
+		dsi_config = dev_priv->dsi_configs[0];
+
+	sender = mdfld_dsi_get_pkg_sender(dsi_config);
+
+	if (!sender) {
+		DRM_ERROR("No sender found\n");
+		return;
+	}
+
+	gen_ctrl_val = (level * 0xff / MDFLD_DSI_BRIGHTNESS_MAX_LEVEL) & 0xff;
+
+	dev_dbg(sender->dev->dev, "pipe = %d, gen_ctrl_val = %d.\n",
+							pipe, gen_ctrl_val);
+
+	if (p_type == TMD_VID) {
+		/* Set display backlight value */
+		mdfld_dsi_send_mcs_short(sender, tmd_write_display_brightness,
+					(u8)gen_ctrl_val, 1, true);
+	} else {
+		/* Set display backlight value */
+		mdfld_dsi_send_mcs_short(sender, write_display_brightness,
+					(u8)gen_ctrl_val, 1, true);
+
+		/* Enable backlight control */
+		if (level == 0)
+			gen_ctrl_val = 0;
+		else
+			gen_ctrl_val = dev_priv->mipi_ctrl_display;
+
+		mdfld_dsi_send_mcs_short(sender, write_ctrl_display,
+					(u8)gen_ctrl_val, 1, true);
+	}
+}
+
+static int mdfld_dsi_get_panel_status(struct mdfld_dsi_config *dsi_config,
+				u8 dcs, u32 *data, bool hs)
+{
+	struct mdfld_dsi_pkg_sender *sender
+		= mdfld_dsi_get_pkg_sender(dsi_config);
+
+	if (!sender || !data) {
+		DRM_ERROR("Invalid parameter\n");
+		return -EINVAL;
+	}
+
+	return mdfld_dsi_read_mcs(sender, dcs, data, 1, hs);
+}
+
+int mdfld_dsi_get_power_mode(struct mdfld_dsi_config *dsi_config, u32 *mode,
+			bool hs)
+{
+	if (!dsi_config || !mode) {
+		DRM_ERROR("Invalid parameter\n");
+		return -EINVAL;
+	}
+
+	return mdfld_dsi_get_panel_status(dsi_config, 0x0a, mode, hs);
+}
+
+/*
+ * NOTE: this function was used by OSPM.
+ * TODO: will be removed later, should work out display interfaces for OSPM
+ */
+void mdfld_dsi_controller_init(struct mdfld_dsi_config *dsi_config, int pipe)
+{
+	if (!dsi_config || ((pipe != 0) && (pipe != 2))) {
+		DRM_ERROR("Invalid parameters\n");
+		return;
+	}
+
+	mdfld_dsi_dpi_controller_init(dsi_config, pipe);
+}
+
+static void mdfld_dsi_connector_save(struct drm_connector *connector)
+{
+}
+
+static void mdfld_dsi_connector_restore(struct drm_connector *connector)
+{
+}
+
+/* FIXME: start using the force parameter */
+static enum drm_connector_status
+mdfld_dsi_connector_detect(struct drm_connector *connector, bool force)
+{
+	struct mdfld_dsi_connector *dsi_connector
+		= mdfld_dsi_connector(connector);
+
+	dsi_connector->status = connector_status_connected;
+
+	return dsi_connector->status;
+}
+
+static int mdfld_dsi_connector_set_property(struct drm_connector *connector,
+				struct drm_property *property,
+				uint64_t value)
+{
+	struct drm_encoder *encoder = connector->encoder;
+
+	if (!strcmp(property->name, "scaling mode") && encoder) {
+		struct psb_intel_crtc *psb_crtc =
+					to_psb_intel_crtc(encoder->crtc);
+		bool centerechange;
+		uint64_t val;
+
+		if (!psb_crtc)
+			goto set_prop_error;
+
+		switch (value) {
+		case DRM_MODE_SCALE_FULLSCREEN:
+			break;
+		case DRM_MODE_SCALE_NO_SCALE:
+			break;
+		case DRM_MODE_SCALE_ASPECT:
+			break;
+		default:
+			goto set_prop_error;
+		}
+
+		if (drm_connector_property_get_value(connector, property, &val))
+			goto set_prop_error;
+
+		if (val == value)
+			goto set_prop_done;
+
+		if (drm_connector_property_set_value(connector,
+							property, value))
+			goto set_prop_error;
+
+		centerechange = (val == DRM_MODE_SCALE_NO_SCALE) ||
+			(value == DRM_MODE_SCALE_NO_SCALE);
+
+		if (psb_crtc->saved_mode.hdisplay != 0 &&
+		    psb_crtc->saved_mode.vdisplay != 0) {
+			if (centerechange) {
+				if (!drm_crtc_helper_set_mode(encoder->crtc,
+						&psb_crtc->saved_mode,
+						encoder->crtc->x,
+						encoder->crtc->y,
+						encoder->crtc->fb))
+					goto set_prop_error;
+			} else {
+				struct drm_encoder_helper_funcs *funcs =
+						encoder->helper_private;
+				funcs->mode_set(encoder,
+					&psb_crtc->saved_mode,
+					&psb_crtc->saved_adjusted_mode);
+			}
+		}
+	} else if (!strcmp(property->name, "backlight") && encoder) {
+		if (drm_connector_property_set_value(connector, property,
+									value))
+			goto set_prop_error;
+		else {
+#ifdef CONFIG_BACKLIGHT_CLASS_DEVICE
+			struct backlight_device *psb_bd;
+
+			psb_bd = mdfld_get_backlight_device();
+			if (psb_bd) {
+				psb_bd->props.brightness = value;
+				mdfld_set_brightness(psb_bd);
+			}
+#endif
+		}
+	}
+set_prop_done:
+	return 0;
+set_prop_error:
+	return -1;
+}
+
+static void mdfld_dsi_connector_destroy(struct drm_connector *connector)
+{
+	struct mdfld_dsi_connector *dsi_connector =
+					mdfld_dsi_connector(connector);
+	struct mdfld_dsi_pkg_sender *sender;
+
+	if (!dsi_connector)
+		return;
+	drm_sysfs_connector_remove(connector);
+	drm_connector_cleanup(connector);
+	sender = dsi_connector->pkg_sender;
+	mdfld_dsi_pkg_sender_destroy(sender);
+	kfree(dsi_connector);
+}
+
+static int mdfld_dsi_connector_get_modes(struct drm_connector *connector)
+{
+	struct mdfld_dsi_connector *dsi_connector =
+				mdfld_dsi_connector(connector);
+	struct mdfld_dsi_config *dsi_config =
+				mdfld_dsi_get_config(dsi_connector);
+	struct drm_display_mode *fixed_mode = dsi_config->fixed_mode;
+	struct drm_display_mode *dup_mode = NULL;
+	struct drm_device *dev = connector->dev;
+
+	connector->display_info.min_vfreq = 0;
+	connector->display_info.max_vfreq = 200;
+	connector->display_info.min_hfreq = 0;
+	connector->display_info.max_hfreq = 200;
+
+	if (fixed_mode) {
+		dev_dbg(dev->dev, "fixed_mode %dx%d\n",
+				fixed_mode->hdisplay, fixed_mode->vdisplay);
+		dup_mode = drm_mode_duplicate(dev, fixed_mode);
+		drm_mode_probed_add(connector, dup_mode);
+		return 1;
+	}
+	DRM_ERROR("Didn't get any modes!\n");
+	return 0;
+}
+
+static int mdfld_dsi_connector_mode_valid(struct drm_connector *connector,
+						struct drm_display_mode *mode)
+{
+	struct mdfld_dsi_connector *dsi_connector =
+					mdfld_dsi_connector(connector);
+	struct mdfld_dsi_config *dsi_config =
+					mdfld_dsi_get_config(dsi_connector);
+	struct drm_display_mode *fixed_mode = dsi_config->fixed_mode;
+
+	if (mode->flags & DRM_MODE_FLAG_DBLSCAN)
+		return MODE_NO_DBLESCAN;
+
+	if (mode->flags & DRM_MODE_FLAG_INTERLACE)
+		return MODE_NO_INTERLACE;
+
+	/**
+	 * FIXME: current DC has no fitting unit, reject any mode setting
+	 * request
+	 * Will figure out a way to do up-scaling(pannel fitting) later.
+	 **/
+	if (fixed_mode) {
+		if (mode->hdisplay != fixed_mode->hdisplay)
+			return MODE_PANEL;
+
+		if (mode->vdisplay != fixed_mode->vdisplay)
+			return MODE_PANEL;
+	}
+
+	return MODE_OK;
+}
+
+static void mdfld_dsi_connector_dpms(struct drm_connector *connector, int mode)
+{
+	if (mode == connector->dpms)
+		return;
+
+	/*first, execute dpms*/
+
+	drm_helper_connector_dpms(connector, mode);
+}
+
+static struct drm_encoder *mdfld_dsi_connector_best_encoder(
+				struct drm_connector *connector)
+{
+	struct mdfld_dsi_connector *dsi_connector =
+				mdfld_dsi_connector(connector);
+	struct mdfld_dsi_config *dsi_config =
+				mdfld_dsi_get_config(dsi_connector);
+	return &dsi_config->encoder->base.base;
+}
+
+/*DSI connector funcs*/
+static const struct drm_connector_funcs mdfld_dsi_connector_funcs = {
+	.dpms = /*drm_helper_connector_dpms*/mdfld_dsi_connector_dpms,
+	.save = mdfld_dsi_connector_save,
+	.restore = mdfld_dsi_connector_restore,
+	.detect = mdfld_dsi_connector_detect,
+	.fill_modes = drm_helper_probe_single_connector_modes,
+	.set_property = mdfld_dsi_connector_set_property,
+	.destroy = mdfld_dsi_connector_destroy,
+};
+
+/*DSI connector helper funcs*/
+static const struct drm_connector_helper_funcs
+	mdfld_dsi_connector_helper_funcs = {
+	.get_modes = mdfld_dsi_connector_get_modes,
+	.mode_valid = mdfld_dsi_connector_mode_valid,
+	.best_encoder = mdfld_dsi_connector_best_encoder,
+};
+
+static int mdfld_dsi_get_default_config(struct drm_device *dev,
+				struct mdfld_dsi_config *config, int pipe)
+{
+	if (!dev || !config) {
+		DRM_ERROR("Invalid parameters");
+		return -EINVAL;
+	}
+
+	config->bpp = 24;
+	if (mdfld_get_panel_type(dev, pipe) == TC35876X)
+		config->lane_count = 4;
+	else
+		config->lane_count = 2;
+	config->channel_num = 0;
+
+	if (mdfld_get_panel_type(dev, pipe) == TMD_VID)
+		config->video_mode = MDFLD_DSI_VIDEO_NON_BURST_MODE_SYNC_PULSE;
+	else if (mdfld_get_panel_type(dev, pipe) == TC35876X)
+		config->video_mode =
+				MDFLD_DSI_VIDEO_NON_BURST_MODE_SYNC_EVENTS;
+	else
+		config->video_mode = MDFLD_DSI_VIDEO_BURST_MODE;
+
+	return 0;
+}
+
+int mdfld_dsi_panel_reset(int pipe)
+{
+	unsigned gpio;
+	int ret = 0;
+
+	switch (pipe) {
+	case 0:
+		gpio = 128;
+		break;
+	case 2:
+		gpio = 34;
+		break;
+	default:
+		DRM_ERROR("Invalid output\n");
+		return -EINVAL;
+	}
+
+	ret = gpio_request(gpio, "gfx");
+	if (ret) {
+		DRM_ERROR("gpio_rqueset failed\n");
+		return ret;
+	}
+
+	ret = gpio_direction_output(gpio, 1);
+	if (ret) {
+		DRM_ERROR("gpio_direction_output failed\n");
+		goto gpio_error;
+	}
+
+	gpio_get_value(128);
+
+gpio_error:
+	if (gpio_is_valid(gpio))
+		gpio_free(gpio);
+
+	return ret;
+}
+
+/*
+ * MIPI output init
+ * @dev drm device
+ * @pipe pipe number. 0 or 2
+ * @config
+ *
+ * Do the initialization of a MIPI output, including create DRM mode objects
+ * initialization of DSI output on @pipe
+ */
+void mdfld_dsi_output_init(struct drm_device *dev,
+			   int pipe,
+			   const struct panel_funcs *p_vid_funcs)
+{
+	struct mdfld_dsi_config *dsi_config;
+	struct mdfld_dsi_connector *dsi_connector;
+	struct drm_connector *connector;
+	struct mdfld_dsi_encoder *encoder;
+	struct drm_psb_private *dev_priv = dev->dev_private;
+	struct panel_info dsi_panel_info;
+	u32 width_mm, height_mm;
+
+	dev_dbg(dev->dev, "init DSI output on pipe %d\n", pipe);
+
+	if (!dev || ((pipe != 0) && (pipe != 2))) {
+		DRM_ERROR("Invalid parameter\n");
+		return;
+	}
+
+	/*create a new connetor*/
+	dsi_connector = kzalloc(sizeof(struct mdfld_dsi_connector), GFP_KERNEL);
+	if (!dsi_connector) {
+		DRM_ERROR("No memory");
+		return;
+	}
+
+	dsi_connector->pipe =  pipe;
+
+	dsi_config = kzalloc(sizeof(struct mdfld_dsi_config),
+			GFP_KERNEL);
+	if (!dsi_config) {
+		DRM_ERROR("cannot allocate memory for DSI config\n");
+		goto dsi_init_err0;
+	}
+	mdfld_dsi_get_default_config(dev, dsi_config, pipe);
+
+	dsi_connector->private = dsi_config;
+
+	dsi_config->changed = 1;
+	dsi_config->dev = dev;
+
+	dsi_config->fixed_mode = p_vid_funcs->get_config_mode(dev);
+	if (p_vid_funcs->get_panel_info(dev, pipe, &dsi_panel_info))
+			goto dsi_init_err0;
+
+	width_mm = dsi_panel_info.width_mm;
+	height_mm = dsi_panel_info.height_mm;
+
+	dsi_config->mode = dsi_config->fixed_mode;
+	dsi_config->connector = dsi_connector;
+
+	if (!dsi_config->fixed_mode) {
+		DRM_ERROR("No pannel fixed mode was found\n");
+		goto dsi_init_err0;
+	}
+
+	if (pipe && dev_priv->dsi_configs[0]) {
+		dsi_config->dvr_ic_inited = 0;
+		dev_priv->dsi_configs[1] = dsi_config;
+	} else if (pipe == 0) {
+		dsi_config->dvr_ic_inited = 1;
+		dev_priv->dsi_configs[0] = dsi_config;
+	} else {
+		DRM_ERROR("Trying to init MIPI1 before MIPI0\n");
+		goto dsi_init_err0;
+	}
+
+
+	connector = &dsi_connector->base.base;
+	drm_connector_init(dev, connector, &mdfld_dsi_connector_funcs,
+						DRM_MODE_CONNECTOR_LVDS);
+	drm_connector_helper_add(connector, &mdfld_dsi_connector_helper_funcs);
+
+	connector->display_info.subpixel_order = SubPixelHorizontalRGB;
+	connector->display_info.width_mm = width_mm;
+	connector->display_info.height_mm = height_mm;
+	connector->interlace_allowed = false;
+	connector->doublescan_allowed = false;
+
+	/*attach properties*/
+	drm_connector_attach_property(connector,
+				dev->mode_config.scaling_mode_property,
+				DRM_MODE_SCALE_FULLSCREEN);
+	drm_connector_attach_property(connector,
+				dev_priv->backlight_property,
+				MDFLD_DSI_BRIGHTNESS_MAX_LEVEL);
+
+	/*init DSI package sender on this output*/
+	if (mdfld_dsi_pkg_sender_init(dsi_connector, pipe)) {
+		DRM_ERROR("Package Sender initialization failed on pipe %d\n",
+									pipe);
+		goto dsi_init_err0;
+	}
+
+	encoder = mdfld_dsi_dpi_init(dev, dsi_connector, p_vid_funcs);
+	if (!encoder) {
+		DRM_ERROR("Create DPI encoder failed\n");
+		goto dsi_init_err1;
+	}
+	encoder->private = dsi_config;
+	dsi_config->encoder = encoder;
+	encoder->base.type = (pipe == 0) ? INTEL_OUTPUT_MIPI :
+		INTEL_OUTPUT_MIPI2;
+	drm_sysfs_connector_add(connector);
+	return;
+
+	/*TODO: add code to destroy outputs on error*/
+dsi_init_err1:
+	/*destroy sender*/
+	mdfld_dsi_pkg_sender_destroy(dsi_connector->pkg_sender);
+
+	drm_connector_cleanup(connector);
+
+	kfree(dsi_config->fixed_mode);
+	kfree(dsi_config);
+dsi_init_err0:
+	kfree(dsi_connector);
+}
diff --git a/drivers/gpu/drm/gma500/mdfld_dsi_output.h b/drivers/gpu/drm/gma500/mdfld_dsi_output.h
new file mode 100644
index 000000000000..21071cef92a4
--- /dev/null
+++ b/drivers/gpu/drm/gma500/mdfld_dsi_output.h
@@ -0,0 +1,378 @@
+/*
+ * Copyright © 2010 Intel Corporation
+ *
+ * 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:
+ * jim liu <jim.liu@intel.com>
+ * Jackie Li<yaodong.li@intel.com>
+ */
+
+#ifndef __MDFLD_DSI_OUTPUT_H__
+#define __MDFLD_DSI_OUTPUT_H__
+
+#include <linux/backlight.h>
+#include <linux/version.h>
+#include <drm/drmP.h>
+#include <drm/drm.h>
+#include <drm/drm_crtc.h>
+#include <drm/drm_edid.h>
+
+#include "psb_drv.h"
+#include "psb_intel_drv.h"
+#include "psb_intel_reg.h"
+#include "mdfld_output.h"
+
+#include <asm/mrst.h>
+
+#define FLD_MASK(start, end)	(((1 << ((start) - (end) + 1)) - 1) << (end))
+#define FLD_VAL(val, start, end) (((val) << (end)) & FLD_MASK(start, end))
+#define FLD_GET(val, start, end) (((val) & FLD_MASK(start, end)) >> (end))
+#define FLD_MOD(orig, val, start, end) \
+	(((orig) & ~FLD_MASK(start, end)) | FLD_VAL(val, start, end))
+
+#define REG_FLD_MOD(reg, val, start, end) \
+	REG_WRITE(reg, FLD_MOD(REG_READ(reg), val, start, end))
+
+static inline int REGISTER_FLD_WAIT(struct drm_device *dev, u32 reg,
+		u32 val, int start, int end)
+{
+	int t = 100000;
+
+	while (FLD_GET(REG_READ(reg), start, end) != val) {
+		if (--t == 0)
+			return 1;
+	}
+
+	return 0;
+}
+
+#define REG_FLD_WAIT(reg, val, start, end) \
+	REGISTER_FLD_WAIT(dev, reg, val, start, end)
+
+#define REG_BIT_WAIT(reg, val, bitnum) \
+	REGISTER_FLD_WAIT(dev, reg, val, bitnum, bitnum)
+
+#define MDFLD_DSI_BRIGHTNESS_MAX_LEVEL 100
+
+#ifdef DEBUG
+#define CHECK_PIPE(pipe) ({			\
+	const typeof(pipe) __pipe = (pipe);	\
+	BUG_ON(__pipe != 0 && __pipe != 2);	\
+	__pipe;	})
+#else
+#define CHECK_PIPE(pipe) (pipe)
+#endif
+
+/*
+ * Actual MIPIA->MIPIC reg offset is 0x800, value 0x400 is valid for 0 and 2
+ */
+#define REG_OFFSET(pipe) (CHECK_PIPE(pipe) * 0x400)
+
+/* mdfld DSI controller registers */
+#define MIPI_DEVICE_READY_REG(pipe)		(0xb000 + REG_OFFSET(pipe))
+#define MIPI_INTR_STAT_REG(pipe)		(0xb004 + REG_OFFSET(pipe))
+#define MIPI_INTR_EN_REG(pipe)			(0xb008 + REG_OFFSET(pipe))
+#define MIPI_DSI_FUNC_PRG_REG(pipe)		(0xb00c + REG_OFFSET(pipe))
+#define MIPI_HS_TX_TIMEOUT_REG(pipe)		(0xb010 + REG_OFFSET(pipe))
+#define MIPI_LP_RX_TIMEOUT_REG(pipe)		(0xb014 + REG_OFFSET(pipe))
+#define MIPI_TURN_AROUND_TIMEOUT_REG(pipe)	(0xb018 + REG_OFFSET(pipe))
+#define MIPI_DEVICE_RESET_TIMER_REG(pipe)	(0xb01c + REG_OFFSET(pipe))
+#define MIPI_DPI_RESOLUTION_REG(pipe)		(0xb020 + REG_OFFSET(pipe))
+#define MIPI_DBI_FIFO_THROTTLE_REG(pipe)	(0xb024 + REG_OFFSET(pipe))
+#define MIPI_HSYNC_COUNT_REG(pipe)		(0xb028 + REG_OFFSET(pipe))
+#define MIPI_HBP_COUNT_REG(pipe)		(0xb02c + REG_OFFSET(pipe))
+#define MIPI_HFP_COUNT_REG(pipe)		(0xb030 + REG_OFFSET(pipe))
+#define MIPI_HACTIVE_COUNT_REG(pipe)		(0xb034 + REG_OFFSET(pipe))
+#define MIPI_VSYNC_COUNT_REG(pipe)		(0xb038 + REG_OFFSET(pipe))
+#define MIPI_VBP_COUNT_REG(pipe)		(0xb03c + REG_OFFSET(pipe))
+#define MIPI_VFP_COUNT_REG(pipe)		(0xb040 + REG_OFFSET(pipe))
+#define MIPI_HIGH_LOW_SWITCH_COUNT_REG(pipe)	(0xb044 + REG_OFFSET(pipe))
+#define MIPI_DPI_CONTROL_REG(pipe)		(0xb048 + REG_OFFSET(pipe))
+#define MIPI_DPI_DATA_REG(pipe)			(0xb04c + REG_OFFSET(pipe))
+#define MIPI_INIT_COUNT_REG(pipe)		(0xb050 + REG_OFFSET(pipe))
+#define MIPI_MAX_RETURN_PACK_SIZE_REG(pipe)	(0xb054 + REG_OFFSET(pipe))
+#define MIPI_VIDEO_MODE_FORMAT_REG(pipe)	(0xb058 + REG_OFFSET(pipe))
+#define MIPI_EOT_DISABLE_REG(pipe)		(0xb05c + REG_OFFSET(pipe))
+#define MIPI_LP_BYTECLK_REG(pipe)		(0xb060 + REG_OFFSET(pipe))
+#define MIPI_LP_GEN_DATA_REG(pipe)		(0xb064 + REG_OFFSET(pipe))
+#define MIPI_HS_GEN_DATA_REG(pipe)		(0xb068 + REG_OFFSET(pipe))
+#define MIPI_LP_GEN_CTRL_REG(pipe)		(0xb06c + REG_OFFSET(pipe))
+#define MIPI_HS_GEN_CTRL_REG(pipe)		(0xb070 + REG_OFFSET(pipe))
+#define MIPI_GEN_FIFO_STAT_REG(pipe)		(0xb074 + REG_OFFSET(pipe))
+#define MIPI_HS_LS_DBI_ENABLE_REG(pipe)		(0xb078 + REG_OFFSET(pipe))
+#define MIPI_DPHY_PARAM_REG(pipe)		(0xb080 + REG_OFFSET(pipe))
+#define MIPI_DBI_BW_CTRL_REG(pipe)		(0xb084 + REG_OFFSET(pipe))
+#define MIPI_CLK_LANE_SWITCH_TIME_CNT_REG(pipe)	(0xb088 + REG_OFFSET(pipe))
+
+#define MIPI_CTRL_REG(pipe)			(0xb104 + REG_OFFSET(pipe))
+#define MIPI_DATA_ADD_REG(pipe)			(0xb108 + REG_OFFSET(pipe))
+#define MIPI_DATA_LEN_REG(pipe)			(0xb10c + REG_OFFSET(pipe))
+#define MIPI_CMD_ADD_REG(pipe)			(0xb110 + REG_OFFSET(pipe))
+#define MIPI_CMD_LEN_REG(pipe)			(0xb114 + REG_OFFSET(pipe))
+
+/* non-uniform reg offset */
+#define MIPI_PORT_CONTROL(pipe)		(CHECK_PIPE(pipe) ? MIPI_C : MIPI)
+
+#define DSI_DEVICE_READY				(0x1)
+#define DSI_POWER_STATE_ULPS_ENTER			(0x2 << 1)
+#define DSI_POWER_STATE_ULPS_EXIT			(0x1 << 1)
+#define DSI_POWER_STATE_ULPS_OFFSET			(0x1)
+
+
+#define DSI_ONE_DATA_LANE					(0x1)
+#define DSI_TWO_DATA_LANE					(0x2)
+#define DSI_THREE_DATA_LANE					(0X3)
+#define DSI_FOUR_DATA_LANE					(0x4)
+#define DSI_DPI_VIRT_CHANNEL_OFFSET			(0x3)
+#define DSI_DBI_VIRT_CHANNEL_OFFSET			(0x5)
+#define DSI_DPI_COLOR_FORMAT_RGB565			(0x01 << 7)
+#define DSI_DPI_COLOR_FORMAT_RGB666			(0x02 << 7)
+#define DSI_DPI_COLOR_FORMAT_RGB666_UNPACK		(0x03 << 7)
+#define DSI_DPI_COLOR_FORMAT_RGB888			(0x04 << 7)
+#define DSI_DBI_COLOR_FORMAT_OPTION2			(0x05 << 13)
+
+#define DSI_INTR_STATE_RXSOTERROR			BIT(0)
+
+#define DSI_INTR_STATE_SPL_PKG_SENT			BIT(30)
+#define DSI_INTR_STATE_TE				BIT(31)
+
+#define DSI_HS_TX_TIMEOUT_MASK				(0xffffff)
+
+#define DSI_LP_RX_TIMEOUT_MASK				(0xffffff)
+
+#define DSI_TURN_AROUND_TIMEOUT_MASK		(0x3f)
+
+#define DSI_RESET_TIMER_MASK				(0xffff)
+
+#define DSI_DBI_FIFO_WM_HALF				(0x0)
+#define DSI_DBI_FIFO_WM_QUARTER				(0x1)
+#define DSI_DBI_FIFO_WM_LOW					(0x2)
+
+#define DSI_DPI_TIMING_MASK					(0xffff)
+
+#define DSI_INIT_TIMER_MASK					(0xffff)
+
+#define DSI_DBI_RETURN_PACK_SIZE_MASK		(0x3ff)
+
+#define DSI_LP_BYTECLK_MASK					(0x0ffff)
+
+#define DSI_HS_CTRL_GEN_SHORT_W0			(0x03)
+#define DSI_HS_CTRL_GEN_SHORT_W1			(0x13)
+#define DSI_HS_CTRL_GEN_SHORT_W2			(0x23)
+#define DSI_HS_CTRL_GEN_R0					(0x04)
+#define DSI_HS_CTRL_GEN_R1					(0x14)
+#define DSI_HS_CTRL_GEN_R2					(0x24)
+#define DSI_HS_CTRL_GEN_LONG_W				(0x29)
+#define DSI_HS_CTRL_MCS_SHORT_W0			(0x05)
+#define DSI_HS_CTRL_MCS_SHORT_W1			(0x15)
+#define DSI_HS_CTRL_MCS_R0					(0x06)
+#define DSI_HS_CTRL_MCS_LONG_W				(0x39)
+#define DSI_HS_CTRL_VC_OFFSET				(0x06)
+#define DSI_HS_CTRL_WC_OFFSET				(0x08)
+
+#define	DSI_FIFO_GEN_HS_DATA_FULL			BIT(0)
+#define DSI_FIFO_GEN_HS_DATA_HALF_EMPTY		BIT(1)
+#define DSI_FIFO_GEN_HS_DATA_EMPTY			BIT(2)
+#define DSI_FIFO_GEN_LP_DATA_FULL			BIT(8)
+#define DSI_FIFO_GEN_LP_DATA_HALF_EMPTY		BIT(9)
+#define DSI_FIFO_GEN_LP_DATA_EMPTY			BIT(10)
+#define DSI_FIFO_GEN_HS_CTRL_FULL			BIT(16)
+#define DSI_FIFO_GEN_HS_CTRL_HALF_EMPTY		BIT(17)
+#define DSI_FIFO_GEN_HS_CTRL_EMPTY			BIT(18)
+#define DSI_FIFO_GEN_LP_CTRL_FULL			BIT(24)
+#define DSI_FIFO_GEN_LP_CTRL_HALF_EMPTY		BIT(25)
+#define DSI_FIFO_GEN_LP_CTRL_EMPTY			BIT(26)
+#define DSI_FIFO_DBI_EMPTY					BIT(27)
+#define DSI_FIFO_DPI_EMPTY					BIT(28)
+
+#define DSI_DBI_HS_LP_SWITCH_MASK			(0x1)
+
+#define DSI_HS_LP_SWITCH_COUNTER_OFFSET		(0x0)
+#define DSI_LP_HS_SWITCH_COUNTER_OFFSET		(0x16)
+
+#define DSI_DPI_CTRL_HS_SHUTDOWN			(0x00000001)
+#define DSI_DPI_CTRL_HS_TURN_ON				(0x00000002)
+
+/*dsi power modes*/
+#define DSI_POWER_MODE_DISPLAY_ON	BIT(2)
+#define DSI_POWER_MODE_NORMAL_ON	BIT(3)
+#define DSI_POWER_MODE_SLEEP_OUT	BIT(4)
+#define DSI_POWER_MODE_PARTIAL_ON	BIT(5)
+#define DSI_POWER_MODE_IDLE_ON		BIT(6)
+
+enum {
+	MDFLD_DSI_VIDEO_NON_BURST_MODE_SYNC_PULSE = 1,
+	MDFLD_DSI_VIDEO_NON_BURST_MODE_SYNC_EVENTS = 2,
+	MDFLD_DSI_VIDEO_BURST_MODE = 3,
+};
+
+#define DSI_DPI_COMPLETE_LAST_LINE			BIT(2)
+#define DSI_DPI_DISABLE_BTA					BIT(3)
+
+struct mdfld_dsi_connector {
+	struct psb_intel_connector base;
+
+	int pipe;
+	void *private;
+	void *pkg_sender;
+
+	/* Connection status */
+	enum drm_connector_status status;
+};
+
+struct mdfld_dsi_encoder {
+	struct psb_intel_encoder base;
+	void *private;
+};
+
+/*
+ * DSI config, consists of one DSI connector, two DSI encoders.
+ * DRM will pick up on DSI encoder basing on differents configs.
+ */
+struct mdfld_dsi_config {
+	struct drm_device *dev;
+	struct drm_display_mode *fixed_mode;
+	struct drm_display_mode *mode;
+
+	struct mdfld_dsi_connector *connector;
+	struct mdfld_dsi_encoder *encoder;
+
+	int changed;
+
+	int bpp;
+	int lane_count;
+	/*Virtual channel number for this encoder*/
+	int channel_num;
+	/*video mode configure*/
+	int video_mode;
+
+	int dvr_ic_inited;
+};
+
+static inline struct mdfld_dsi_connector *mdfld_dsi_connector(
+		struct drm_connector *connector)
+{
+	struct psb_intel_connector *psb_connector;
+
+	psb_connector = to_psb_intel_connector(connector);
+
+	return container_of(psb_connector, struct mdfld_dsi_connector, base);
+}
+
+static inline struct mdfld_dsi_encoder *mdfld_dsi_encoder(
+		struct drm_encoder *encoder)
+{
+	struct psb_intel_encoder *psb_encoder;
+
+	psb_encoder = to_psb_intel_encoder(encoder);
+
+	return container_of(psb_encoder, struct mdfld_dsi_encoder, base);
+}
+
+static inline struct mdfld_dsi_config *
+	mdfld_dsi_get_config(struct mdfld_dsi_connector *connector)
+{
+	if (!connector)
+		return NULL;
+	return (struct mdfld_dsi_config *)connector->private;
+}
+
+static inline void *mdfld_dsi_get_pkg_sender(struct mdfld_dsi_config *config)
+{
+	struct mdfld_dsi_connector *dsi_connector;
+
+	if (!config)
+		return NULL;
+
+	dsi_connector = config->connector;
+
+	if (!dsi_connector)
+		return NULL;
+
+	return dsi_connector->pkg_sender;
+}
+
+static inline struct mdfld_dsi_config *
+	mdfld_dsi_encoder_get_config(struct mdfld_dsi_encoder *encoder)
+{
+	if (!encoder)
+		return NULL;
+	return (struct mdfld_dsi_config *)encoder->private;
+}
+
+static inline struct mdfld_dsi_connector *
+	mdfld_dsi_encoder_get_connector(struct mdfld_dsi_encoder *encoder)
+{
+	struct mdfld_dsi_config *config;
+
+	if (!encoder)
+		return NULL;
+
+	config = mdfld_dsi_encoder_get_config(encoder);
+	if (!config)
+		return NULL;
+
+	return config->connector;
+}
+
+static inline void *mdfld_dsi_encoder_get_pkg_sender(
+				struct mdfld_dsi_encoder *encoder)
+{
+	struct mdfld_dsi_config *dsi_config;
+
+	dsi_config = mdfld_dsi_encoder_get_config(encoder);
+	if (!dsi_config)
+		return NULL;
+
+	return mdfld_dsi_get_pkg_sender(dsi_config);
+}
+
+static inline int mdfld_dsi_encoder_get_pipe(struct mdfld_dsi_encoder *encoder)
+{
+	struct mdfld_dsi_connector *connector;
+
+	if (!encoder)
+		return -1;
+
+	connector = mdfld_dsi_encoder_get_connector(encoder);
+	if (!connector)
+		return -1;
+	return connector->pipe;
+}
+
+/* Export functions */
+extern void mdfld_dsi_gen_fifo_ready(struct drm_device *dev,
+					u32 gen_fifo_stat_reg, u32 fifo_stat);
+extern void mdfld_dsi_brightness_init(struct mdfld_dsi_config *dsi_config,
+					int pipe);
+extern void mdfld_dsi_brightness_control(struct drm_device *dev, int pipe,
+					int level);
+extern void mdfld_dsi_output_init(struct drm_device *dev,
+					int pipe,
+					const struct panel_funcs *p_vid_funcs);
+extern void mdfld_dsi_controller_init(struct mdfld_dsi_config *dsi_config,
+					int pipe);
+
+extern int mdfld_dsi_get_power_mode(struct mdfld_dsi_config *dsi_config,
+					u32 *mode, bool hs);
+extern int mdfld_dsi_panel_reset(int pipe);
+
+#endif /*__MDFLD_DSI_OUTPUT_H__*/
diff --git a/drivers/gpu/drm/gma500/mdfld_dsi_pkg_sender.c b/drivers/gpu/drm/gma500/mdfld_dsi_pkg_sender.c
new file mode 100644
index 000000000000..baa0e14165e0
--- /dev/null
+++ b/drivers/gpu/drm/gma500/mdfld_dsi_pkg_sender.c
@@ -0,0 +1,694 @@
+/*
+ * Copyright © 2010 Intel Corporation
+ *
+ * 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:
+ * Jackie Li<yaodong.li@intel.com>
+ */
+
+#include <linux/freezer.h>
+
+#include "mdfld_dsi_output.h"
+#include "mdfld_dsi_pkg_sender.h"
+#include "mdfld_dsi_dpi.h"
+
+#define MDFLD_DSI_READ_MAX_COUNT		5000
+
+enum data_type {
+	DSI_DT_GENERIC_SHORT_WRITE_0	= 0x03,
+	DSI_DT_GENERIC_SHORT_WRITE_1	= 0x13,
+	DSI_DT_GENERIC_SHORT_WRITE_2	= 0x23,
+	DSI_DT_GENERIC_READ_0		= 0x04,
+	DSI_DT_GENERIC_READ_1		= 0x14,
+	DSI_DT_GENERIC_READ_2		= 0x24,
+	DSI_DT_GENERIC_LONG_WRITE	= 0x29,
+	DSI_DT_DCS_SHORT_WRITE_0	= 0x05,
+	DSI_DT_DCS_SHORT_WRITE_1	= 0x15,
+	DSI_DT_DCS_READ			= 0x06,
+	DSI_DT_DCS_LONG_WRITE		= 0x39,
+};
+
+enum {
+	MDFLD_DSI_PANEL_MODE_SLEEP = 0x1,
+};
+
+enum {
+	MDFLD_DSI_PKG_SENDER_FREE = 0x0,
+	MDFLD_DSI_PKG_SENDER_BUSY = 0x1,
+};
+
+static const char *const dsi_errors[] = {
+	"RX SOT Error",
+	"RX SOT Sync Error",
+	"RX EOT Sync Error",
+	"RX Escape Mode Entry Error",
+	"RX LP TX Sync Error",
+	"RX HS Receive Timeout Error",
+	"RX False Control Error",
+	"RX ECC Single Bit Error",
+	"RX ECC Multibit Error",
+	"RX Checksum Error",
+	"RX DSI Data Type Not Recognised",
+	"RX DSI VC ID Invalid",
+	"TX False Control Error",
+	"TX ECC Single Bit Error",
+	"TX ECC Multibit Error",
+	"TX Checksum Error",
+	"TX DSI Data Type Not Recognised",
+	"TX DSI VC ID invalid",
+	"High Contention",
+	"Low contention",
+	"DPI FIFO Under run",
+	"HS TX Timeout",
+	"LP RX Timeout",
+	"Turn Around ACK Timeout",
+	"ACK With No Error",
+	"RX Invalid TX Length",
+	"RX Prot Violation",
+	"HS Generic Write FIFO Full",
+	"LP Generic Write FIFO Full",
+	"Generic Read Data Avail"
+	"Special Packet Sent",
+	"Tearing Effect",
+};
+
+static inline int wait_for_gen_fifo_empty(struct mdfld_dsi_pkg_sender *sender,
+						u32 mask)
+{
+	struct drm_device *dev = sender->dev;
+	u32 gen_fifo_stat_reg = sender->mipi_gen_fifo_stat_reg;
+	int retry = 0xffff;
+
+	while (retry--) {
+		if ((mask & REG_READ(gen_fifo_stat_reg)) == mask)
+			return 0;
+		udelay(100);
+	}
+	DRM_ERROR("fifo is NOT empty 0x%08x\n", REG_READ(gen_fifo_stat_reg));
+	return -EIO;
+}
+
+static int wait_for_all_fifos_empty(struct mdfld_dsi_pkg_sender *sender)
+{
+	return wait_for_gen_fifo_empty(sender, (BIT(2) | BIT(10) | BIT(18) |
+						BIT(26) | BIT(27) | BIT(28)));
+}
+
+static int wait_for_lp_fifos_empty(struct mdfld_dsi_pkg_sender *sender)
+{
+	return wait_for_gen_fifo_empty(sender, (BIT(10) | BIT(26)));
+}
+
+static int wait_for_hs_fifos_empty(struct mdfld_dsi_pkg_sender *sender)
+{
+	return wait_for_gen_fifo_empty(sender, (BIT(2) | BIT(18)));
+}
+
+static int handle_dsi_error(struct mdfld_dsi_pkg_sender *sender, u32 mask)
+{
+	u32 intr_stat_reg = sender->mipi_intr_stat_reg;
+	struct drm_device *dev = sender->dev;
+
+	dev_dbg(sender->dev->dev, "Handling error 0x%08x\n", mask);
+
+	switch (mask) {
+	case BIT(0):
+	case BIT(1):
+	case BIT(2):
+	case BIT(3):
+	case BIT(4):
+	case BIT(5):
+	case BIT(6):
+	case BIT(7):
+	case BIT(8):
+	case BIT(9):
+	case BIT(10):
+	case BIT(11):
+	case BIT(12):
+	case BIT(13):
+		dev_dbg(sender->dev->dev, "No Action required\n");
+		break;
+	case BIT(14):
+		/*wait for all fifo empty*/
+		/*wait_for_all_fifos_empty(sender)*/;
+		break;
+	case BIT(15):
+		dev_dbg(sender->dev->dev, "No Action required\n");
+		break;
+	case BIT(16):
+		break;
+	case BIT(17):
+		break;
+	case BIT(18):
+	case BIT(19):
+		dev_dbg(sender->dev->dev, "High/Low contention detected\n");
+		/*wait for contention recovery time*/
+		/*mdelay(10);*/
+		/*wait for all fifo empty*/
+		if (0)
+			wait_for_all_fifos_empty(sender);
+		break;
+	case BIT(20):
+		dev_dbg(sender->dev->dev, "No Action required\n");
+		break;
+	case BIT(21):
+		/*wait for all fifo empty*/
+		/*wait_for_all_fifos_empty(sender);*/
+		break;
+	case BIT(22):
+		break;
+	case BIT(23):
+	case BIT(24):
+	case BIT(25):
+	case BIT(26):
+	case BIT(27):
+		dev_dbg(sender->dev->dev, "HS Gen fifo full\n");
+		REG_WRITE(intr_stat_reg, mask);
+		wait_for_hs_fifos_empty(sender);
+		break;
+	case BIT(28):
+		dev_dbg(sender->dev->dev, "LP Gen fifo full\n");
+		REG_WRITE(intr_stat_reg, mask);
+		wait_for_lp_fifos_empty(sender);
+		break;
+	case BIT(29):
+	case BIT(30):
+	case BIT(31):
+		dev_dbg(sender->dev->dev, "No Action required\n");
+		break;
+	}
+
+	if (mask & REG_READ(intr_stat_reg))
+		dev_dbg(sender->dev->dev,
+				"Cannot clean interrupt 0x%08x\n", mask);
+	return 0;
+}
+
+static int dsi_error_handler(struct mdfld_dsi_pkg_sender *sender)
+{
+	struct drm_device *dev = sender->dev;
+	u32 intr_stat_reg = sender->mipi_intr_stat_reg;
+	u32 mask;
+	u32 intr_stat;
+	int i;
+	int err = 0;
+
+	intr_stat = REG_READ(intr_stat_reg);
+
+	for (i = 0; i < 32; i++) {
+		mask = (0x00000001UL) << i;
+		if (intr_stat & mask) {
+			dev_dbg(sender->dev->dev, "[DSI]: %s\n", dsi_errors[i]);
+			err = handle_dsi_error(sender, mask);
+			if (err)
+				DRM_ERROR("Cannot handle error\n");
+		}
+	}
+	return err;
+}
+
+static int send_short_pkg(struct mdfld_dsi_pkg_sender *sender, u8 data_type,
+			u8 cmd, u8 param, bool hs)
+{
+	struct drm_device *dev = sender->dev;
+	u32 ctrl_reg;
+	u32 val;
+	u8 virtual_channel = 0;
+
+	if (hs) {
+		ctrl_reg = sender->mipi_hs_gen_ctrl_reg;
+
+		/* FIXME: wait_for_hs_fifos_empty(sender); */
+	} else {
+		ctrl_reg = sender->mipi_lp_gen_ctrl_reg;
+
+		/* FIXME: wait_for_lp_fifos_empty(sender); */
+	}
+
+	val = FLD_VAL(param, 23, 16) | FLD_VAL(cmd, 15, 8) |
+		FLD_VAL(virtual_channel, 7, 6) | FLD_VAL(data_type, 5, 0);
+
+	REG_WRITE(ctrl_reg, val);
+
+	return 0;
+}
+
+static int send_long_pkg(struct mdfld_dsi_pkg_sender *sender, u8 data_type,
+			u8 *data, int len, bool hs)
+{
+	struct drm_device *dev = sender->dev;
+	u32 ctrl_reg;
+	u32 data_reg;
+	u32 val;
+	u8 *p;
+	u8 b1, b2, b3, b4;
+	u8 virtual_channel = 0;
+	int i;
+
+	if (hs) {
+		ctrl_reg = sender->mipi_hs_gen_ctrl_reg;
+		data_reg = sender->mipi_hs_gen_data_reg;
+
+		/* FIXME: wait_for_hs_fifos_empty(sender); */
+	} else {
+		ctrl_reg = sender->mipi_lp_gen_ctrl_reg;
+		data_reg = sender->mipi_lp_gen_data_reg;
+
+		/* FIXME: wait_for_lp_fifos_empty(sender); */
+	}
+
+	p = data;
+	for (i = 0; i < len / 4; i++) {
+		b1 = *p++;
+		b2 = *p++;
+		b3 = *p++;
+		b4 = *p++;
+
+		REG_WRITE(data_reg, b4 << 24 | b3 << 16 | b2 << 8 | b1);
+	}
+
+	i = len % 4;
+	if (i) {
+		b1 = 0; b2 = 0; b3 = 0;
+
+		switch (i) {
+		case 3:
+			b1 = *p++;
+			b2 = *p++;
+			b3 = *p++;
+			break;
+		case 2:
+			b1 = *p++;
+			b2 = *p++;
+			break;
+		case 1:
+			b1 = *p++;
+			break;
+		}
+
+		REG_WRITE(data_reg, b3 << 16 | b2 << 8 | b1);
+	}
+
+	val = FLD_VAL(len, 23, 8) | FLD_VAL(virtual_channel, 7, 6) |
+		FLD_VAL(data_type, 5, 0);
+
+	REG_WRITE(ctrl_reg, val);
+
+	return 0;
+}
+
+static int send_pkg_prepare(struct mdfld_dsi_pkg_sender *sender, u8 data_type,
+			u8 *data, u16 len)
+{
+	u8 cmd;
+
+	switch (data_type) {
+	case DSI_DT_DCS_SHORT_WRITE_0:
+	case DSI_DT_DCS_SHORT_WRITE_1:
+	case DSI_DT_DCS_LONG_WRITE:
+		cmd = *data;
+		break;
+	default:
+		return 0;
+	}
+
+	/*this prevents other package sending while doing msleep*/
+	sender->status = MDFLD_DSI_PKG_SENDER_BUSY;
+
+	/*wait for 120 milliseconds in case exit_sleep_mode just be sent*/
+	if (unlikely(cmd == DCS_ENTER_SLEEP_MODE)) {
+		/*TODO: replace it with msleep later*/
+		mdelay(120);
+	}
+
+	if (unlikely(cmd == DCS_EXIT_SLEEP_MODE)) {
+		/*TODO: replace it with msleep later*/
+		mdelay(120);
+	}
+	return 0;
+}
+
+static int send_pkg_done(struct mdfld_dsi_pkg_sender *sender, u8 data_type,
+			u8 *data, u16 len)
+{
+	u8 cmd;
+
+	switch (data_type) {
+	case DSI_DT_DCS_SHORT_WRITE_0:
+	case DSI_DT_DCS_SHORT_WRITE_1:
+	case DSI_DT_DCS_LONG_WRITE:
+		cmd = *data;
+		break;
+	default:
+		return 0;
+	}
+
+	/*update panel status*/
+	if (unlikely(cmd == DCS_ENTER_SLEEP_MODE)) {
+		sender->panel_mode |= MDFLD_DSI_PANEL_MODE_SLEEP;
+		/*TODO: replace it with msleep later*/
+		mdelay(120);
+	} else if (unlikely(cmd == DCS_EXIT_SLEEP_MODE)) {
+		sender->panel_mode &= ~MDFLD_DSI_PANEL_MODE_SLEEP;
+		/*TODO: replace it with msleep later*/
+		mdelay(120);
+	} else if (unlikely(cmd == DCS_SOFT_RESET)) {
+		/*TODO: replace it with msleep later*/
+		mdelay(5);
+	}
+
+	sender->status = MDFLD_DSI_PKG_SENDER_FREE;
+
+	return 0;
+}
+
+static int send_pkg(struct mdfld_dsi_pkg_sender *sender, u8 data_type,
+		u8 *data, u16 len, bool hs)
+{
+	int ret;
+
+	/*handle DSI error*/
+	ret = dsi_error_handler(sender);
+	if (ret) {
+		DRM_ERROR("Error handling failed\n");
+		return -EAGAIN;
+	}
+
+	/* send pkg */
+	if (sender->status == MDFLD_DSI_PKG_SENDER_BUSY) {
+		DRM_ERROR("sender is busy\n");
+		return -EAGAIN;
+	}
+
+	ret = send_pkg_prepare(sender, data_type, data, len);
+	if (ret) {
+		DRM_ERROR("send_pkg_prepare error\n");
+		return ret;
+	}
+
+	switch (data_type) {
+	case DSI_DT_GENERIC_SHORT_WRITE_0:
+	case DSI_DT_GENERIC_SHORT_WRITE_1:
+	case DSI_DT_GENERIC_SHORT_WRITE_2:
+	case DSI_DT_GENERIC_READ_0:
+	case DSI_DT_GENERIC_READ_1:
+	case DSI_DT_GENERIC_READ_2:
+	case DSI_DT_DCS_SHORT_WRITE_0:
+	case DSI_DT_DCS_SHORT_WRITE_1:
+	case DSI_DT_DCS_READ:
+		ret = send_short_pkg(sender, data_type, data[0], data[1], hs);
+		break;
+	case DSI_DT_GENERIC_LONG_WRITE:
+	case DSI_DT_DCS_LONG_WRITE:
+		ret = send_long_pkg(sender, data_type, data, len, hs);
+		break;
+	}
+
+	send_pkg_done(sender, data_type, data, len);
+
+	/*FIXME: should I query complete and fifo empty here?*/
+
+	return ret;
+}
+
+int mdfld_dsi_send_mcs_long(struct mdfld_dsi_pkg_sender *sender, u8 *data,
+			u32 len, bool hs)
+{
+	unsigned long flags;
+
+	if (!sender || !data || !len) {
+		DRM_ERROR("Invalid parameters\n");
+		return -EINVAL;
+	}
+
+	spin_lock_irqsave(&sender->lock, flags);
+	send_pkg(sender, DSI_DT_DCS_LONG_WRITE, data, len, hs);
+	spin_unlock_irqrestore(&sender->lock, flags);
+
+	return 0;
+}
+
+int mdfld_dsi_send_mcs_short(struct mdfld_dsi_pkg_sender *sender, u8 cmd,
+			u8 param, u8 param_num, bool hs)
+{
+	u8 data[2];
+	unsigned long flags;
+	u8 data_type;
+
+	if (!sender) {
+		DRM_ERROR("Invalid parameter\n");
+		return -EINVAL;
+	}
+
+	data[0] = cmd;
+
+	if (param_num) {
+		data_type = DSI_DT_DCS_SHORT_WRITE_1;
+		data[1] = param;
+	} else {
+		data_type = DSI_DT_DCS_SHORT_WRITE_0;
+		data[1] = 0;
+	}
+
+	spin_lock_irqsave(&sender->lock, flags);
+	send_pkg(sender, data_type, data, sizeof(data), hs);
+	spin_unlock_irqrestore(&sender->lock, flags);
+
+	return 0;
+}
+
+int mdfld_dsi_send_gen_short(struct mdfld_dsi_pkg_sender *sender, u8 param0,
+			u8 param1, u8 param_num, bool hs)
+{
+	u8 data[2];
+	unsigned long flags;
+	u8 data_type;
+
+	if (!sender || param_num > 2) {
+		DRM_ERROR("Invalid parameter\n");
+		return -EINVAL;
+	}
+
+	switch (param_num) {
+	case 0:
+		data_type = DSI_DT_GENERIC_SHORT_WRITE_0;
+		data[0] = 0;
+		data[1] = 0;
+		break;
+	case 1:
+		data_type = DSI_DT_GENERIC_SHORT_WRITE_1;
+		data[0] = param0;
+		data[1] = 0;
+		break;
+	case 2:
+		data_type = DSI_DT_GENERIC_SHORT_WRITE_2;
+		data[0] = param0;
+		data[1] = param1;
+		break;
+	}
+
+	spin_lock_irqsave(&sender->lock, flags);
+	send_pkg(sender, data_type, data, sizeof(data), hs);
+	spin_unlock_irqrestore(&sender->lock, flags);
+
+	return 0;
+}
+
+int mdfld_dsi_send_gen_long(struct mdfld_dsi_pkg_sender *sender, u8 *data,
+			u32 len, bool hs)
+{
+	unsigned long flags;
+
+	if (!sender || !data || !len) {
+		DRM_ERROR("Invalid parameters\n");
+		return -EINVAL;
+	}
+
+	spin_lock_irqsave(&sender->lock, flags);
+	send_pkg(sender, DSI_DT_GENERIC_LONG_WRITE, data, len, hs);
+	spin_unlock_irqrestore(&sender->lock, flags);
+
+	return 0;
+}
+
+static int __read_panel_data(struct mdfld_dsi_pkg_sender *sender, u8 data_type,
+			u8 *data, u16 len, u32 *data_out, u16 len_out, bool hs)
+{
+	unsigned long flags;
+	struct drm_device *dev = sender->dev;
+	int i;
+	u32 gen_data_reg;
+	int retry = MDFLD_DSI_READ_MAX_COUNT;
+
+	if (!sender || !data_out || !len_out) {
+		DRM_ERROR("Invalid parameters\n");
+		return -EINVAL;
+	}
+
+	/**
+	 * do reading.
+	 * 0) send out generic read request
+	 * 1) polling read data avail interrupt
+	 * 2) read data
+	 */
+	spin_lock_irqsave(&sender->lock, flags);
+
+	REG_WRITE(sender->mipi_intr_stat_reg, BIT(29));
+
+	if ((REG_READ(sender->mipi_intr_stat_reg) & BIT(29)))
+		DRM_ERROR("Can NOT clean read data valid interrupt\n");
+
+	/*send out read request*/
+	send_pkg(sender, data_type, data, len, hs);
+
+	/*polling read data avail interrupt*/
+	while (retry && !(REG_READ(sender->mipi_intr_stat_reg) & BIT(29))) {
+		udelay(100);
+		retry--;
+	}
+
+	if (!retry) {
+		spin_unlock_irqrestore(&sender->lock, flags);
+		return -ETIMEDOUT;
+	}
+
+	REG_WRITE(sender->mipi_intr_stat_reg, BIT(29));
+
+	/*read data*/
+	if (hs)
+		gen_data_reg = sender->mipi_hs_gen_data_reg;
+	else
+		gen_data_reg = sender->mipi_lp_gen_data_reg;
+
+	for (i = 0; i < len_out; i++)
+		*(data_out + i) = REG_READ(gen_data_reg);
+
+	spin_unlock_irqrestore(&sender->lock, flags);
+
+	return 0;
+}
+
+int mdfld_dsi_read_mcs(struct mdfld_dsi_pkg_sender *sender, u8 cmd,
+		u32 *data, u16 len, bool hs)
+{
+	if (!sender || !data || !len) {
+		DRM_ERROR("Invalid parameters\n");
+		return -EINVAL;
+	}
+
+	return __read_panel_data(sender, DSI_DT_DCS_READ, &cmd, 1,
+				data, len, hs);
+}
+
+int mdfld_dsi_pkg_sender_init(struct mdfld_dsi_connector *dsi_connector,
+								int pipe)
+{
+	struct mdfld_dsi_pkg_sender *pkg_sender;
+	struct mdfld_dsi_config *dsi_config =
+				mdfld_dsi_get_config(dsi_connector);
+	struct drm_device *dev = dsi_config->dev;
+	u32 mipi_val = 0;
+
+	if (!dsi_connector) {
+		DRM_ERROR("Invalid parameter\n");
+		return -EINVAL;
+	}
+
+	pkg_sender = dsi_connector->pkg_sender;
+
+	if (!pkg_sender || IS_ERR(pkg_sender)) {
+		pkg_sender = kzalloc(sizeof(struct mdfld_dsi_pkg_sender),
+								GFP_KERNEL);
+		if (!pkg_sender) {
+			DRM_ERROR("Create DSI pkg sender failed\n");
+			return -ENOMEM;
+		}
+		dsi_connector->pkg_sender = (void *)pkg_sender;
+	}
+
+	pkg_sender->dev = dev;
+	pkg_sender->dsi_connector = dsi_connector;
+	pkg_sender->pipe = pipe;
+	pkg_sender->pkg_num = 0;
+	pkg_sender->panel_mode = 0;
+	pkg_sender->status = MDFLD_DSI_PKG_SENDER_FREE;
+
+	/*init regs*/
+	if (pipe == 0) {
+		pkg_sender->dpll_reg = MRST_DPLL_A;
+		pkg_sender->dspcntr_reg = DSPACNTR;
+		pkg_sender->pipeconf_reg = PIPEACONF;
+		pkg_sender->dsplinoff_reg = DSPALINOFF;
+		pkg_sender->dspsurf_reg = DSPASURF;
+		pkg_sender->pipestat_reg = PIPEASTAT;
+	} else if (pipe == 2) {
+		pkg_sender->dpll_reg = MRST_DPLL_A;
+		pkg_sender->dspcntr_reg = DSPCCNTR;
+		pkg_sender->pipeconf_reg = PIPECCONF;
+		pkg_sender->dsplinoff_reg = DSPCLINOFF;
+		pkg_sender->dspsurf_reg = DSPCSURF;
+		pkg_sender->pipestat_reg = PIPECSTAT;
+	}
+
+	pkg_sender->mipi_intr_stat_reg = MIPI_INTR_STAT_REG(pipe);
+	pkg_sender->mipi_lp_gen_data_reg = MIPI_LP_GEN_DATA_REG(pipe);
+	pkg_sender->mipi_hs_gen_data_reg = MIPI_HS_GEN_DATA_REG(pipe);
+	pkg_sender->mipi_lp_gen_ctrl_reg = MIPI_LP_GEN_CTRL_REG(pipe);
+	pkg_sender->mipi_hs_gen_ctrl_reg = MIPI_HS_GEN_CTRL_REG(pipe);
+	pkg_sender->mipi_gen_fifo_stat_reg = MIPI_GEN_FIFO_STAT_REG(pipe);
+	pkg_sender->mipi_data_addr_reg = MIPI_DATA_ADD_REG(pipe);
+	pkg_sender->mipi_data_len_reg = MIPI_DATA_LEN_REG(pipe);
+	pkg_sender->mipi_cmd_addr_reg = MIPI_CMD_ADD_REG(pipe);
+	pkg_sender->mipi_cmd_len_reg = MIPI_CMD_LEN_REG(pipe);
+
+	/*init lock*/
+	spin_lock_init(&pkg_sender->lock);
+
+	if (mdfld_get_panel_type(dev, pipe) != TC35876X) {
+		/**
+		 * For video mode, don't enable DPI timing output here,
+		 * will init the DPI timing output during mode setting.
+		 */
+		mipi_val = PASS_FROM_SPHY_TO_AFE | SEL_FLOPPED_HSTX;
+
+		if (pipe == 0)
+			mipi_val |= 0x2;
+
+		REG_WRITE(MIPI_PORT_CONTROL(pipe), mipi_val);
+		REG_READ(MIPI_PORT_CONTROL(pipe));
+
+		/* do dsi controller init */
+		mdfld_dsi_controller_init(dsi_config, pipe);
+	}
+
+	return 0;
+}
+
+void mdfld_dsi_pkg_sender_destroy(struct mdfld_dsi_pkg_sender *sender)
+{
+	if (!sender || IS_ERR(sender))
+		return;
+
+	/*free*/
+	kfree(sender);
+}
+
+
diff --git a/drivers/gpu/drm/gma500/mdfld_dsi_pkg_sender.h b/drivers/gpu/drm/gma500/mdfld_dsi_pkg_sender.h
new file mode 100644
index 000000000000..459cd7ea8b81
--- /dev/null
+++ b/drivers/gpu/drm/gma500/mdfld_dsi_pkg_sender.h
@@ -0,0 +1,92 @@
+/*
+ * Copyright © 2010 Intel Corporation
+ *
+ * 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:
+ * Jackie Li<yaodong.li@intel.com>
+ */
+#ifndef __MDFLD_DSI_PKG_SENDER_H__
+#define __MDFLD_DSI_PKG_SENDER_H__
+
+#include <linux/kthread.h>
+
+#define MDFLD_MAX_DCS_PARAM	8
+
+struct mdfld_dsi_pkg_sender {
+	struct drm_device *dev;
+	struct mdfld_dsi_connector *dsi_connector;
+	u32 status;
+	u32 panel_mode;
+
+	int pipe;
+
+	spinlock_t lock;
+
+	u32 pkg_num;
+
+	/* Registers */
+	u32 dpll_reg;
+	u32 dspcntr_reg;
+	u32 pipeconf_reg;
+	u32 pipestat_reg;
+	u32 dsplinoff_reg;
+	u32 dspsurf_reg;
+
+	u32 mipi_intr_stat_reg;
+	u32 mipi_lp_gen_data_reg;
+	u32 mipi_hs_gen_data_reg;
+	u32 mipi_lp_gen_ctrl_reg;
+	u32 mipi_hs_gen_ctrl_reg;
+	u32 mipi_gen_fifo_stat_reg;
+	u32 mipi_data_addr_reg;
+	u32 mipi_data_len_reg;
+	u32 mipi_cmd_addr_reg;
+	u32 mipi_cmd_len_reg;
+};
+
+/* DCS definitions */
+#define DCS_SOFT_RESET			0x01
+#define DCS_ENTER_SLEEP_MODE		0x10
+#define DCS_EXIT_SLEEP_MODE		0x11
+#define DCS_SET_DISPLAY_OFF		0x28
+#define DCS_SET_DISPLAY_ON		0x29
+#define DCS_SET_COLUMN_ADDRESS		0x2a
+#define DCS_SET_PAGE_ADDRESS		0x2b
+#define DCS_WRITE_MEM_START		0x2c
+#define DCS_SET_TEAR_OFF		0x34
+#define DCS_SET_TEAR_ON			0x35
+
+extern int mdfld_dsi_pkg_sender_init(struct mdfld_dsi_connector *dsi_connector,
+					int pipe);
+extern void mdfld_dsi_pkg_sender_destroy(struct mdfld_dsi_pkg_sender *sender);
+int mdfld_dsi_send_mcs_short(struct mdfld_dsi_pkg_sender *sender, u8 cmd,
+					u8 param, u8 param_num, bool hs);
+int mdfld_dsi_send_mcs_long(struct mdfld_dsi_pkg_sender *sender, u8 *data,
+					u32 len, bool hs);
+int mdfld_dsi_send_gen_short(struct mdfld_dsi_pkg_sender *sender, u8 param0,
+					u8 param1, u8 param_num, bool hs);
+int mdfld_dsi_send_gen_long(struct mdfld_dsi_pkg_sender *sender, u8 *data,
+					u32 len, bool hs);
+/* Read interfaces */
+int mdfld_dsi_read_mcs(struct mdfld_dsi_pkg_sender *sender, u8 cmd,
+		u32 *data, u16 len, bool hs);
+
+#endif
diff --git a/drivers/gpu/drm/gma500/mdfld_intel_display.c b/drivers/gpu/drm/gma500/mdfld_intel_display.c
new file mode 100644
index 000000000000..a35a2921bdf7
--- /dev/null
+++ b/drivers/gpu/drm/gma500/mdfld_intel_display.c
@@ -0,0 +1,1180 @@
+/*
+ * Copyright © 2006-2007 Intel 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, write to the Free Software Foundation, Inc.,
+ * 51 Franklin St - Fifth Floor, Boston, MA 02110-1301 USA.
+ *
+ * Authors:
+ *	Eric Anholt <eric@anholt.net>
+ */
+
+#include <linux/i2c.h>
+#include <linux/pm_runtime.h>
+
+#include <drm/drmP.h>
+#include "psb_intel_reg.h"
+#include "psb_intel_display.h"
+#include "framebuffer.h"
+#include "mdfld_output.h"
+#include "mdfld_dsi_output.h"
+
+/* Hardcoded currently */
+static int ksel = KSEL_CRYSTAL_19;
+
+struct psb_intel_range_t {
+	int min, max;
+};
+
+struct mrst_limit_t {
+	struct psb_intel_range_t dot, m, p1;
+};
+
+struct mrst_clock_t {
+	/* derived values */
+	int dot;
+	int m;
+	int p1;
+};
+
+#define COUNT_MAX 0x10000000
+
+void mdfldWaitForPipeDisable(struct drm_device *dev, int pipe)
+{
+	int count, temp;
+	u32 pipeconf_reg = PIPEACONF;
+
+	switch (pipe) {
+	case 0:
+		break;
+	case 1:
+		pipeconf_reg = PIPEBCONF;
+		break;
+	case 2:
+		pipeconf_reg = PIPECCONF;
+		break;
+	default:
+		DRM_ERROR("Illegal Pipe Number.\n");
+		return;
+	}
+
+	/* FIXME JLIU7_PO */
+	psb_intel_wait_for_vblank(dev);
+	return;
+
+	/* Wait for for the pipe disable to take effect. */
+	for (count = 0; count < COUNT_MAX; count++) {
+		temp = REG_READ(pipeconf_reg);
+		if ((temp & PIPEACONF_PIPE_STATE) == 0)
+			break;
+	}
+}
+
+void mdfldWaitForPipeEnable(struct drm_device *dev, int pipe)
+{
+	int count, temp;
+	u32 pipeconf_reg = PIPEACONF;
+
+	switch (pipe) {
+	case 0:
+		break;
+	case 1:
+		pipeconf_reg = PIPEBCONF;
+		break;
+	case 2:
+		pipeconf_reg = PIPECCONF;
+		break;
+	default:
+		DRM_ERROR("Illegal Pipe Number.\n");
+		return;
+	}
+
+	/* FIXME JLIU7_PO */
+	psb_intel_wait_for_vblank(dev);
+	return;
+
+	/* Wait for for the pipe enable to take effect. */
+	for (count = 0; count < COUNT_MAX; count++) {
+		temp = REG_READ(pipeconf_reg);
+		if ((temp & PIPEACONF_PIPE_STATE) == 1)
+			break;
+	}
+}
+
+static void psb_intel_crtc_prepare(struct drm_crtc *crtc)
+{
+	struct drm_crtc_helper_funcs *crtc_funcs = crtc->helper_private;
+	crtc_funcs->dpms(crtc, DRM_MODE_DPMS_OFF);
+}
+
+static void psb_intel_crtc_commit(struct drm_crtc *crtc)
+{
+	struct drm_crtc_helper_funcs *crtc_funcs = crtc->helper_private;
+	crtc_funcs->dpms(crtc, DRM_MODE_DPMS_ON);
+}
+
+static bool psb_intel_crtc_mode_fixup(struct drm_crtc *crtc,
+				  struct drm_display_mode *mode,
+				  struct drm_display_mode *adjusted_mode)
+{
+	return true;
+}
+
+/**
+ * Return the pipe currently connected to the panel fitter,
+ * or -1 if the panel fitter is not present or not in use
+ */
+static int psb_intel_panel_fitter_pipe(struct drm_device *dev)
+{
+	u32 pfit_control;
+
+	pfit_control = REG_READ(PFIT_CONTROL);
+
+	/* See if the panel fitter is in use */
+	if ((pfit_control & PFIT_ENABLE) == 0)
+		return -1;
+
+	/* 965 can place panel fitter on either pipe */
+	return (pfit_control >> 29) & 0x3;
+}
+
+static struct drm_device globle_dev;
+
+void mdfld__intel_plane_set_alpha(int enable)
+{
+	struct drm_device *dev = &globle_dev;
+	int dspcntr_reg = DSPACNTR;
+	u32 dspcntr;
+
+	dspcntr = REG_READ(dspcntr_reg);
+
+	if (enable) {
+		dspcntr &= ~DISPPLANE_32BPP_NO_ALPHA;
+		dspcntr |= DISPPLANE_32BPP;
+	} else {
+		dspcntr &= ~DISPPLANE_32BPP;
+		dspcntr |= DISPPLANE_32BPP_NO_ALPHA;
+	}
+
+	REG_WRITE(dspcntr_reg, dspcntr);
+}
+
+static int check_fb(struct drm_framebuffer *fb)
+{
+	if (!fb)
+		return 0;
+
+	switch (fb->bits_per_pixel) {
+	case 8:
+	case 16:
+	case 24:
+	case 32:
+		return 0;
+	default:
+		DRM_ERROR("Unknown color depth\n");
+		return -EINVAL;
+	}
+}
+
+static int mdfld__intel_pipe_set_base(struct drm_crtc *crtc, int x, int y,
+				struct drm_framebuffer *old_fb)
+{
+	struct drm_device *dev = crtc->dev;
+	/* struct drm_i915_master_private *master_priv; */
+	struct psb_intel_crtc *psb_intel_crtc = to_psb_intel_crtc(crtc);
+	struct psb_framebuffer *psbfb = to_psb_fb(crtc->fb);
+	int pipe = psb_intel_crtc->pipe;
+	unsigned long start, offset;
+	int dsplinoff = DSPALINOFF;
+	int dspsurf = DSPASURF;
+	int dspstride = DSPASTRIDE;
+	int dspcntr_reg = DSPACNTR;
+	u32 dspcntr;
+	int ret;
+
+	memcpy(&globle_dev, dev, sizeof(struct drm_device));
+
+	dev_dbg(dev->dev, "pipe = 0x%x.\n", pipe);
+
+	/* no fb bound */
+	if (!crtc->fb) {
+		dev_dbg(dev->dev, "No FB bound\n");
+		return 0;
+	}
+
+	ret = check_fb(crtc->fb);
+	if (ret)
+		return ret;
+
+	switch (pipe) {
+	case 0:
+		dsplinoff = DSPALINOFF;
+		break;
+	case 1:
+		dsplinoff = DSPBLINOFF;
+		dspsurf = DSPBSURF;
+		dspstride = DSPBSTRIDE;
+		dspcntr_reg = DSPBCNTR;
+		break;
+	case 2:
+		dsplinoff = DSPCLINOFF;
+		dspsurf = DSPCSURF;
+		dspstride = DSPCSTRIDE;
+		dspcntr_reg = DSPCCNTR;
+		break;
+	default:
+		DRM_ERROR("Illegal Pipe Number.\n");
+		return -EINVAL;
+	}
+
+	if (!gma_power_begin(dev, true))
+		return 0;
+
+	start = psbfb->gtt->offset;
+	offset = y * crtc->fb->pitches[0] + x * (crtc->fb->bits_per_pixel / 8);
+
+	REG_WRITE(dspstride, crtc->fb->pitches[0]);
+	dspcntr = REG_READ(dspcntr_reg);
+	dspcntr &= ~DISPPLANE_PIXFORMAT_MASK;
+
+	switch (crtc->fb->bits_per_pixel) {
+	case 8:
+		dspcntr |= DISPPLANE_8BPP;
+		break;
+	case 16:
+		if (crtc->fb->depth == 15)
+			dspcntr |= DISPPLANE_15_16BPP;
+		else
+			dspcntr |= DISPPLANE_16BPP;
+		break;
+	case 24:
+	case 32:
+		dspcntr |= DISPPLANE_32BPP_NO_ALPHA;
+		break;
+	}
+	REG_WRITE(dspcntr_reg, dspcntr);
+
+	dev_dbg(dev->dev, "Writing base %08lX %08lX %d %d\n",
+						start, offset, x, y);
+	REG_WRITE(dsplinoff, offset);
+	REG_READ(dsplinoff);
+	REG_WRITE(dspsurf, start);
+	REG_READ(dspsurf);
+
+	gma_power_end(dev);
+
+	return 0;
+}
+
+/*
+ * Disable the pipe, plane and pll.
+ *
+ */
+void mdfld_disable_crtc(struct drm_device *dev, int pipe)
+{
+	int dpll_reg = MRST_DPLL_A;
+	int dspcntr_reg = DSPACNTR;
+	int dspbase_reg = MRST_DSPABASE;
+	int pipeconf_reg = PIPEACONF;
+	u32 temp;
+
+	dev_dbg(dev->dev, "pipe = %d\n", pipe);
+
+
+	switch (pipe) {
+	case 0:
+		break;
+	case 1:
+		dpll_reg = MDFLD_DPLL_B;
+		dspcntr_reg = DSPBCNTR;
+		dspbase_reg = DSPBSURF;
+		pipeconf_reg = PIPEBCONF;
+		break;
+	case 2:
+		dpll_reg = MRST_DPLL_A;
+		dspcntr_reg = DSPCCNTR;
+		dspbase_reg = MDFLD_DSPCBASE;
+		pipeconf_reg = PIPECCONF;
+		break;
+	default:
+		DRM_ERROR("Illegal Pipe Number.\n");
+		return;
+	}
+
+	if (pipe != 1)
+		mdfld_dsi_gen_fifo_ready(dev, MIPI_GEN_FIFO_STAT_REG(pipe),
+				HS_CTRL_FIFO_EMPTY | HS_DATA_FIFO_EMPTY);
+
+	/* Disable display plane */
+	temp = REG_READ(dspcntr_reg);
+	if ((temp & DISPLAY_PLANE_ENABLE) != 0) {
+		REG_WRITE(dspcntr_reg,
+			  temp & ~DISPLAY_PLANE_ENABLE);
+		/* Flush the plane changes */
+		REG_WRITE(dspbase_reg, REG_READ(dspbase_reg));
+		REG_READ(dspbase_reg);
+	}
+
+	/* FIXME_JLIU7 MDFLD_PO revisit */
+
+	/* Next, disable display pipes */
+	temp = REG_READ(pipeconf_reg);
+	if ((temp & PIPEACONF_ENABLE) != 0) {
+		temp &= ~PIPEACONF_ENABLE;
+		temp |= PIPECONF_PLANE_OFF | PIPECONF_CURSOR_OFF;
+		REG_WRITE(pipeconf_reg, temp);
+		REG_READ(pipeconf_reg);
+
+		/* Wait for for the pipe disable to take effect. */
+		mdfldWaitForPipeDisable(dev, pipe);
+	}
+
+	temp = REG_READ(dpll_reg);
+	if (temp & DPLL_VCO_ENABLE) {
+		if ((pipe != 1 &&
+			!((REG_READ(PIPEACONF) | REG_READ(PIPECCONF))
+				& PIPEACONF_ENABLE)) || pipe == 1) {
+			temp &= ~(DPLL_VCO_ENABLE);
+			REG_WRITE(dpll_reg, temp);
+			REG_READ(dpll_reg);
+			/* Wait for the clocks to turn off. */
+			/* FIXME_MDFLD PO may need more delay */
+			udelay(500);
+
+			if (!(temp & MDFLD_PWR_GATE_EN)) {
+				/* gating power of DPLL */
+				REG_WRITE(dpll_reg, temp | MDFLD_PWR_GATE_EN);
+				/* FIXME_MDFLD PO - change 500 to 1 after PO */
+				udelay(5000);
+			}
+		}
+	}
+
+}
+
+/**
+ * Sets the power management mode of the pipe and plane.
+ *
+ * This code should probably grow support for turning the cursor off and back
+ * on appropriately at the same time as we're turning the pipe off/on.
+ */
+static void mdfld_crtc_dpms(struct drm_crtc *crtc, int mode)
+{
+	struct drm_device *dev = crtc->dev;
+	struct drm_psb_private *dev_priv = dev->dev_private;
+	struct psb_intel_crtc *psb_intel_crtc = to_psb_intel_crtc(crtc);
+	int pipe = psb_intel_crtc->pipe;
+	int dpll_reg = MRST_DPLL_A;
+	int dspcntr_reg = DSPACNTR;
+	int dspbase_reg = MRST_DSPABASE;
+	int pipeconf_reg = PIPEACONF;
+	u32 pipestat_reg = PIPEASTAT;
+	u32 pipeconf = dev_priv->pipeconf[pipe];
+	u32 temp;
+	int timeout = 0;
+
+	dev_dbg(dev->dev, "mode = %d, pipe = %d\n", mode, pipe);
+
+/* FIXME_JLIU7 MDFLD_PO replaced w/ the following function */
+/* mdfld_dbi_dpms (struct drm_device *dev, int pipe, bool enabled) */
+
+	switch (pipe) {
+	case 0:
+		break;
+	case 1:
+		dpll_reg = DPLL_B;
+		dspcntr_reg = DSPBCNTR;
+		dspbase_reg = MRST_DSPBBASE;
+		pipeconf_reg = PIPEBCONF;
+		dpll_reg = MDFLD_DPLL_B;
+		break;
+	case 2:
+		dpll_reg = MRST_DPLL_A;
+		dspcntr_reg = DSPCCNTR;
+		dspbase_reg = MDFLD_DSPCBASE;
+		pipeconf_reg = PIPECCONF;
+		pipestat_reg = PIPECSTAT;
+		break;
+	default:
+		DRM_ERROR("Illegal Pipe Number.\n");
+		return;
+	}
+
+	if (!gma_power_begin(dev, true))
+		return;
+
+	/* XXX: When our outputs are all unaware of DPMS modes other than off
+	 * and on, we should map those modes to DRM_MODE_DPMS_OFF in the CRTC.
+	 */
+	switch (mode) {
+	case DRM_MODE_DPMS_ON:
+	case DRM_MODE_DPMS_STANDBY:
+	case DRM_MODE_DPMS_SUSPEND:
+		/* Enable the DPLL */
+		temp = REG_READ(dpll_reg);
+
+		if ((temp & DPLL_VCO_ENABLE) == 0) {
+			/* When ungating power of DPLL, needs to wait 0.5us
+			   before enable the VCO */
+			if (temp & MDFLD_PWR_GATE_EN) {
+				temp &= ~MDFLD_PWR_GATE_EN;
+				REG_WRITE(dpll_reg, temp);
+				/* FIXME_MDFLD PO - change 500 to 1 after PO */
+				udelay(500);
+			}
+
+			REG_WRITE(dpll_reg, temp);
+			REG_READ(dpll_reg);
+			/* FIXME_MDFLD PO - change 500 to 1 after PO */
+			udelay(500);
+
+			REG_WRITE(dpll_reg, temp | DPLL_VCO_ENABLE);
+			REG_READ(dpll_reg);
+
+			/**
+			 * wait for DSI PLL to lock
+			 * NOTE: only need to poll status of pipe 0 and pipe 1,
+			 * since both MIPI pipes share the same PLL.
+			 */
+			while ((pipe != 2) && (timeout < 20000) &&
+			  !(REG_READ(pipeconf_reg) & PIPECONF_DSIPLL_LOCK)) {
+				udelay(150);
+				timeout++;
+			}
+		}
+
+		/* Enable the plane */
+		temp = REG_READ(dspcntr_reg);
+		if ((temp & DISPLAY_PLANE_ENABLE) == 0) {
+			REG_WRITE(dspcntr_reg,
+				temp | DISPLAY_PLANE_ENABLE);
+			/* Flush the plane changes */
+			REG_WRITE(dspbase_reg, REG_READ(dspbase_reg));
+		}
+
+		/* Enable the pipe */
+		temp = REG_READ(pipeconf_reg);
+		if ((temp & PIPEACONF_ENABLE) == 0) {
+			REG_WRITE(pipeconf_reg, pipeconf);
+
+			/* Wait for for the pipe enable to take effect. */
+			mdfldWaitForPipeEnable(dev, pipe);
+		}
+
+		/*workaround for sighting 3741701 Random X blank display*/
+		/*perform w/a in video mode only on pipe A or C*/
+		if (pipe == 0 || pipe == 2) {
+			REG_WRITE(pipestat_reg, REG_READ(pipestat_reg));
+			msleep(100);
+			if (PIPE_VBLANK_STATUS & REG_READ(pipestat_reg))
+				dev_dbg(dev->dev, "OK");
+			else {
+				dev_dbg(dev->dev, "STUCK!!!!");
+				/*shutdown controller*/
+				temp = REG_READ(dspcntr_reg);
+				REG_WRITE(dspcntr_reg,
+						temp & ~DISPLAY_PLANE_ENABLE);
+				REG_WRITE(dspbase_reg, REG_READ(dspbase_reg));
+				/*mdfld_dsi_dpi_shut_down(dev, pipe);*/
+				REG_WRITE(0xb048, 1);
+				msleep(100);
+				temp = REG_READ(pipeconf_reg);
+				temp &= ~PIPEACONF_ENABLE;
+				REG_WRITE(pipeconf_reg, temp);
+				msleep(100); /*wait for pipe disable*/
+				REG_WRITE(MIPI_DEVICE_READY_REG(pipe), 0);
+				msleep(100);
+				REG_WRITE(0xb004, REG_READ(0xb004));
+				/* try to bring the controller back up again*/
+				REG_WRITE(MIPI_DEVICE_READY_REG(pipe), 1);
+				temp = REG_READ(dspcntr_reg);
+				REG_WRITE(dspcntr_reg,
+						temp | DISPLAY_PLANE_ENABLE);
+				REG_WRITE(dspbase_reg, REG_READ(dspbase_reg));
+				/*mdfld_dsi_dpi_turn_on(dev, pipe);*/
+				REG_WRITE(0xb048, 2);
+				msleep(100);
+				temp = REG_READ(pipeconf_reg);
+				temp |= PIPEACONF_ENABLE;
+				REG_WRITE(pipeconf_reg, temp);
+			}
+		}
+
+		psb_intel_crtc_load_lut(crtc);
+
+		/* Give the overlay scaler a chance to enable
+		   if it's on this pipe */
+		/* psb_intel_crtc_dpms_video(crtc, true); TODO */
+
+		break;
+	case DRM_MODE_DPMS_OFF:
+		/* Give the overlay scaler a chance to disable
+		 * if it's on this pipe */
+		/* psb_intel_crtc_dpms_video(crtc, FALSE); TODO */
+		if (pipe != 1)
+			mdfld_dsi_gen_fifo_ready(dev,
+				MIPI_GEN_FIFO_STAT_REG(pipe),
+				HS_CTRL_FIFO_EMPTY | HS_DATA_FIFO_EMPTY);
+
+		/* Disable the VGA plane that we never use */
+		REG_WRITE(VGACNTRL, VGA_DISP_DISABLE);
+
+		/* Disable display plane */
+		temp = REG_READ(dspcntr_reg);
+		if ((temp & DISPLAY_PLANE_ENABLE) != 0) {
+			REG_WRITE(dspcntr_reg,
+				  temp & ~DISPLAY_PLANE_ENABLE);
+			/* Flush the plane changes */
+			REG_WRITE(dspbase_reg, REG_READ(dspbase_reg));
+			REG_READ(dspbase_reg);
+		}
+
+		/* Next, disable display pipes */
+		temp = REG_READ(pipeconf_reg);
+		if ((temp & PIPEACONF_ENABLE) != 0) {
+			temp &= ~PIPEACONF_ENABLE;
+			temp |= PIPECONF_PLANE_OFF | PIPECONF_CURSOR_OFF;
+			REG_WRITE(pipeconf_reg, temp);
+			REG_READ(pipeconf_reg);
+
+			/* Wait for for the pipe disable to take effect. */
+			mdfldWaitForPipeDisable(dev, pipe);
+		}
+
+		temp = REG_READ(dpll_reg);
+		if (temp & DPLL_VCO_ENABLE) {
+			if ((pipe != 1 && !((REG_READ(PIPEACONF)
+				| REG_READ(PIPECCONF)) & PIPEACONF_ENABLE))
+					|| pipe == 1) {
+				temp &= ~(DPLL_VCO_ENABLE);
+				REG_WRITE(dpll_reg, temp);
+				REG_READ(dpll_reg);
+				/* Wait for the clocks to turn off. */
+				/* FIXME_MDFLD PO may need more delay */
+				udelay(500);
+			}
+		}
+		break;
+	}
+	gma_power_end(dev);
+}
+
+
+#define MDFLD_LIMT_DPLL_19	    0
+#define MDFLD_LIMT_DPLL_25	    1
+#define MDFLD_LIMT_DPLL_83	    2
+#define MDFLD_LIMT_DPLL_100	    3
+#define MDFLD_LIMT_DSIPLL_19	    4
+#define MDFLD_LIMT_DSIPLL_25	    5
+#define MDFLD_LIMT_DSIPLL_83	    6
+#define MDFLD_LIMT_DSIPLL_100	    7
+
+#define MDFLD_DOT_MIN		  19750
+#define MDFLD_DOT_MAX		  120000
+#define MDFLD_DPLL_M_MIN_19	    113
+#define MDFLD_DPLL_M_MAX_19	    155
+#define MDFLD_DPLL_P1_MIN_19	    2
+#define MDFLD_DPLL_P1_MAX_19	    10
+#define MDFLD_DPLL_M_MIN_25	    101
+#define MDFLD_DPLL_M_MAX_25	    130
+#define MDFLD_DPLL_P1_MIN_25	    2
+#define MDFLD_DPLL_P1_MAX_25	    10
+#define MDFLD_DPLL_M_MIN_83	    64
+#define MDFLD_DPLL_M_MAX_83	    64
+#define MDFLD_DPLL_P1_MIN_83	    2
+#define MDFLD_DPLL_P1_MAX_83	    2
+#define MDFLD_DPLL_M_MIN_100	    64
+#define MDFLD_DPLL_M_MAX_100	    64
+#define MDFLD_DPLL_P1_MIN_100	    2
+#define MDFLD_DPLL_P1_MAX_100	    2
+#define MDFLD_DSIPLL_M_MIN_19	    131
+#define MDFLD_DSIPLL_M_MAX_19	    175
+#define MDFLD_DSIPLL_P1_MIN_19	    3
+#define MDFLD_DSIPLL_P1_MAX_19	    8
+#define MDFLD_DSIPLL_M_MIN_25	    97
+#define MDFLD_DSIPLL_M_MAX_25	    140
+#define MDFLD_DSIPLL_P1_MIN_25	    3
+#define MDFLD_DSIPLL_P1_MAX_25	    9
+#define MDFLD_DSIPLL_M_MIN_83	    33
+#define MDFLD_DSIPLL_M_MAX_83	    92
+#define MDFLD_DSIPLL_P1_MIN_83	    2
+#define MDFLD_DSIPLL_P1_MAX_83	    3
+#define MDFLD_DSIPLL_M_MIN_100	    97
+#define MDFLD_DSIPLL_M_MAX_100	    140
+#define MDFLD_DSIPLL_P1_MIN_100	    3
+#define MDFLD_DSIPLL_P1_MAX_100	    9
+
+static const struct mrst_limit_t mdfld_limits[] = {
+	{			/* MDFLD_LIMT_DPLL_19 */
+	 .dot = {.min = MDFLD_DOT_MIN, .max = MDFLD_DOT_MAX},
+	 .m = {.min = MDFLD_DPLL_M_MIN_19, .max = MDFLD_DPLL_M_MAX_19},
+	 .p1 = {.min = MDFLD_DPLL_P1_MIN_19, .max = MDFLD_DPLL_P1_MAX_19},
+	 },
+	{			/* MDFLD_LIMT_DPLL_25 */
+	 .dot = {.min = MDFLD_DOT_MIN, .max = MDFLD_DOT_MAX},
+	 .m = {.min = MDFLD_DPLL_M_MIN_25, .max = MDFLD_DPLL_M_MAX_25},
+	 .p1 = {.min = MDFLD_DPLL_P1_MIN_25, .max = MDFLD_DPLL_P1_MAX_25},
+	 },
+	{			/* MDFLD_LIMT_DPLL_83 */
+	 .dot = {.min = MDFLD_DOT_MIN, .max = MDFLD_DOT_MAX},
+	 .m = {.min = MDFLD_DPLL_M_MIN_83, .max = MDFLD_DPLL_M_MAX_83},
+	 .p1 = {.min = MDFLD_DPLL_P1_MIN_83, .max = MDFLD_DPLL_P1_MAX_83},
+	 },
+	{			/* MDFLD_LIMT_DPLL_100 */
+	 .dot = {.min = MDFLD_DOT_MIN, .max = MDFLD_DOT_MAX},
+	 .m = {.min = MDFLD_DPLL_M_MIN_100, .max = MDFLD_DPLL_M_MAX_100},
+	 .p1 = {.min = MDFLD_DPLL_P1_MIN_100, .max = MDFLD_DPLL_P1_MAX_100},
+	 },
+	{			/* MDFLD_LIMT_DSIPLL_19 */
+	 .dot = {.min = MDFLD_DOT_MIN, .max = MDFLD_DOT_MAX},
+	 .m = {.min = MDFLD_DSIPLL_M_MIN_19, .max = MDFLD_DSIPLL_M_MAX_19},
+	 .p1 = {.min = MDFLD_DSIPLL_P1_MIN_19, .max = MDFLD_DSIPLL_P1_MAX_19},
+	 },
+	{			/* MDFLD_LIMT_DSIPLL_25 */
+	 .dot = {.min = MDFLD_DOT_MIN, .max = MDFLD_DOT_MAX},
+	 .m = {.min = MDFLD_DSIPLL_M_MIN_25, .max = MDFLD_DSIPLL_M_MAX_25},
+	 .p1 = {.min = MDFLD_DSIPLL_P1_MIN_25, .max = MDFLD_DSIPLL_P1_MAX_25},
+	 },
+	{			/* MDFLD_LIMT_DSIPLL_83 */
+	 .dot = {.min = MDFLD_DOT_MIN, .max = MDFLD_DOT_MAX},
+	 .m = {.min = MDFLD_DSIPLL_M_MIN_83, .max = MDFLD_DSIPLL_M_MAX_83},
+	 .p1 = {.min = MDFLD_DSIPLL_P1_MIN_83, .max = MDFLD_DSIPLL_P1_MAX_83},
+	 },
+	{			/* MDFLD_LIMT_DSIPLL_100 */
+	 .dot = {.min = MDFLD_DOT_MIN, .max = MDFLD_DOT_MAX},
+	 .m = {.min = MDFLD_DSIPLL_M_MIN_100, .max = MDFLD_DSIPLL_M_MAX_100},
+	 .p1 = {.min = MDFLD_DSIPLL_P1_MIN_100, .max = MDFLD_DSIPLL_P1_MAX_100},
+	 },
+};
+
+#define MDFLD_M_MIN	    21
+#define MDFLD_M_MAX	    180
+static const u32 mdfld_m_converts[] = {
+/* M configuration table from 9-bit LFSR table */
+	224, 368, 440, 220, 366, 439, 219, 365, 182, 347, /* 21 - 30 */
+	173, 342, 171, 85, 298, 149, 74, 37, 18, 265,   /* 31 - 40 */
+	388, 194, 353, 432, 216, 108, 310, 155, 333, 166, /* 41 - 50 */
+	83, 41, 276, 138, 325, 162, 337, 168, 340, 170, /* 51 - 60 */
+	341, 426, 469, 234, 373, 442, 221, 110, 311, 411, /* 61 - 70 */
+	461, 486, 243, 377, 188, 350, 175, 343, 427, 213, /* 71 - 80 */
+	106, 53, 282, 397, 354, 227, 113, 56, 284, 142, /* 81 - 90 */
+	71, 35, 273, 136, 324, 418, 465, 488, 500, 506, /* 91 - 100 */
+	253, 126, 63, 287, 399, 455, 483, 241, 376, 444, /* 101 - 110 */
+	478, 495, 503, 251, 381, 446, 479, 239, 375, 443, /* 111 - 120 */
+	477, 238, 119, 315, 157, 78, 295, 147, 329, 420, /* 121 - 130 */
+	210, 105, 308, 154, 77, 38, 275, 137, 68, 290, /* 131 - 140 */
+	145, 328, 164, 82, 297, 404, 458, 485, 498, 249, /* 141 - 150 */
+	380, 190, 351, 431, 471, 235, 117, 314, 413, 206, /* 151 - 160 */
+	103, 51, 25, 12, 262, 387, 193, 96, 48, 280, /* 161 - 170 */
+	396, 198, 99, 305, 152, 76, 294, 403, 457, 228, /* 171 - 180 */
+};
+
+static const struct mrst_limit_t *mdfld_limit(struct drm_crtc *crtc)
+{
+	const struct mrst_limit_t *limit = NULL;
+	struct drm_device *dev = crtc->dev;
+	struct drm_psb_private *dev_priv = dev->dev_private;
+
+	if (psb_intel_pipe_has_type(crtc, INTEL_OUTPUT_MIPI)
+	    || psb_intel_pipe_has_type(crtc, INTEL_OUTPUT_MIPI2)) {
+		if ((ksel == KSEL_CRYSTAL_19) || (ksel == KSEL_BYPASS_19))
+			limit = &mdfld_limits[MDFLD_LIMT_DSIPLL_19];
+		else if (ksel == KSEL_BYPASS_25)
+			limit = &mdfld_limits[MDFLD_LIMT_DSIPLL_25];
+		else if ((ksel == KSEL_BYPASS_83_100) &&
+				(dev_priv->core_freq == 166))
+			limit = &mdfld_limits[MDFLD_LIMT_DSIPLL_83];
+		else if ((ksel == KSEL_BYPASS_83_100) &&
+			 (dev_priv->core_freq == 100 ||
+				dev_priv->core_freq == 200))
+			limit = &mdfld_limits[MDFLD_LIMT_DSIPLL_100];
+	} else if (psb_intel_pipe_has_type(crtc, INTEL_OUTPUT_HDMI)) {
+		if ((ksel == KSEL_CRYSTAL_19) || (ksel == KSEL_BYPASS_19))
+			limit = &mdfld_limits[MDFLD_LIMT_DPLL_19];
+		else if (ksel == KSEL_BYPASS_25)
+			limit = &mdfld_limits[MDFLD_LIMT_DPLL_25];
+		else if ((ksel == KSEL_BYPASS_83_100) &&
+				(dev_priv->core_freq == 166))
+			limit = &mdfld_limits[MDFLD_LIMT_DPLL_83];
+		else if ((ksel == KSEL_BYPASS_83_100) &&
+				 (dev_priv->core_freq == 100 ||
+				 dev_priv->core_freq == 200))
+			limit = &mdfld_limits[MDFLD_LIMT_DPLL_100];
+	} else {
+		limit = NULL;
+		dev_dbg(dev->dev, "mdfld_limit Wrong display type.\n");
+	}
+
+	return limit;
+}
+
+/** Derive the pixel clock for the given refclk and divisors for 8xx chips. */
+static void mdfld_clock(int refclk, struct mrst_clock_t *clock)
+{
+	clock->dot = (refclk * clock->m) / clock->p1;
+}
+
+/**
+ * Returns a set of divisors for the desired target clock with the given refclk,
+ * or FALSE.  Divisor values are the actual divisors for
+ */
+static bool
+mdfldFindBestPLL(struct drm_crtc *crtc, int target, int refclk,
+		struct mrst_clock_t *best_clock)
+{
+	struct mrst_clock_t clock;
+	const struct mrst_limit_t *limit = mdfld_limit(crtc);
+	int err = target;
+
+	memset(best_clock, 0, sizeof(*best_clock));
+
+	for (clock.m = limit->m.min; clock.m <= limit->m.max; clock.m++) {
+		for (clock.p1 = limit->p1.min; clock.p1 <= limit->p1.max;
+		     clock.p1++) {
+			int this_err;
+
+			mdfld_clock(refclk, &clock);
+
+			this_err = abs(clock.dot - target);
+			if (this_err < err) {
+				*best_clock = clock;
+				err = this_err;
+			}
+		}
+	}
+	return err != target;
+}
+
+static int mdfld_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 psb_intel_crtc *psb_intel_crtc = to_psb_intel_crtc(crtc);
+	struct drm_psb_private *dev_priv = dev->dev_private;
+	int pipe = psb_intel_crtc->pipe;
+	int fp_reg = MRST_FPA0;
+	int dpll_reg = MRST_DPLL_A;
+	int dspcntr_reg = DSPACNTR;
+	int pipeconf_reg = PIPEACONF;
+	int htot_reg = HTOTAL_A;
+	int hblank_reg = HBLANK_A;
+	int hsync_reg = HSYNC_A;
+	int vtot_reg = VTOTAL_A;
+	int vblank_reg = VBLANK_A;
+	int vsync_reg = VSYNC_A;
+	int dspsize_reg = DSPASIZE;
+	int dsppos_reg = DSPAPOS;
+	int pipesrc_reg = PIPEASRC;
+	u32 *pipeconf = &dev_priv->pipeconf[pipe];
+	u32 *dspcntr = &dev_priv->dspcntr[pipe];
+	int refclk = 0;
+	int clk_n = 0, clk_p2 = 0, clk_byte = 1, clk = 0, m_conv = 0,
+								clk_tmp = 0;
+	struct mrst_clock_t clock;
+	bool ok;
+	u32 dpll = 0, fp = 0;
+	bool is_mipi = false, is_mipi2 = false, is_hdmi = false;
+	struct drm_mode_config *mode_config = &dev->mode_config;
+	struct psb_intel_encoder *psb_intel_encoder = NULL;
+	uint64_t scalingType = DRM_MODE_SCALE_FULLSCREEN;
+	struct drm_encoder *encoder;
+	struct drm_connector *connector;
+	int timeout = 0;
+	int ret;
+
+	dev_dbg(dev->dev, "pipe = 0x%x\n", pipe);
+
+#if 0
+	if (pipe == 1) {
+		if (!gma_power_begin(dev, true))
+			return 0;
+		android_hdmi_crtc_mode_set(crtc, mode, adjusted_mode,
+			x, y, old_fb);
+		goto mrst_crtc_mode_set_exit;
+	}
+#endif
+
+	switch (pipe) {
+	case 0:
+		break;
+	case 1:
+		fp_reg = FPB0;
+		dpll_reg = DPLL_B;
+		dspcntr_reg = DSPBCNTR;
+		pipeconf_reg = PIPEBCONF;
+		htot_reg = HTOTAL_B;
+		hblank_reg = HBLANK_B;
+		hsync_reg = HSYNC_B;
+		vtot_reg = VTOTAL_B;
+		vblank_reg = VBLANK_B;
+		vsync_reg = VSYNC_B;
+		dspsize_reg = DSPBSIZE;
+		dsppos_reg = DSPBPOS;
+		pipesrc_reg = PIPEBSRC;
+		fp_reg = MDFLD_DPLL_DIV0;
+		dpll_reg = MDFLD_DPLL_B;
+		break;
+	case 2:
+		dpll_reg = MRST_DPLL_A;
+		dspcntr_reg = DSPCCNTR;
+		pipeconf_reg = PIPECCONF;
+		htot_reg = HTOTAL_C;
+		hblank_reg = HBLANK_C;
+		hsync_reg = HSYNC_C;
+		vtot_reg = VTOTAL_C;
+		vblank_reg = VBLANK_C;
+		vsync_reg = VSYNC_C;
+		dspsize_reg = DSPCSIZE;
+		dsppos_reg = DSPCPOS;
+		pipesrc_reg = PIPECSRC;
+		break;
+	default:
+		DRM_ERROR("Illegal Pipe Number.\n");
+		return 0;
+	}
+
+	ret = check_fb(crtc->fb);
+	if (ret)
+		return ret;
+
+	dev_dbg(dev->dev, "adjusted_hdisplay = %d\n",
+		 adjusted_mode->hdisplay);
+	dev_dbg(dev->dev, "adjusted_vdisplay = %d\n",
+		 adjusted_mode->vdisplay);
+	dev_dbg(dev->dev, "adjusted_hsync_start = %d\n",
+		 adjusted_mode->hsync_start);
+	dev_dbg(dev->dev, "adjusted_hsync_end = %d\n",
+		 adjusted_mode->hsync_end);
+	dev_dbg(dev->dev, "adjusted_htotal = %d\n",
+		 adjusted_mode->htotal);
+	dev_dbg(dev->dev, "adjusted_vsync_start = %d\n",
+		 adjusted_mode->vsync_start);
+	dev_dbg(dev->dev, "adjusted_vsync_end = %d\n",
+		 adjusted_mode->vsync_end);
+	dev_dbg(dev->dev, "adjusted_vtotal = %d\n",
+		 adjusted_mode->vtotal);
+	dev_dbg(dev->dev, "adjusted_clock = %d\n",
+		 adjusted_mode->clock);
+	dev_dbg(dev->dev, "hdisplay = %d\n",
+		 mode->hdisplay);
+	dev_dbg(dev->dev, "vdisplay = %d\n",
+		 mode->vdisplay);
+
+	if (!gma_power_begin(dev, true))
+		return 0;
+
+	memcpy(&psb_intel_crtc->saved_mode, mode,
+					sizeof(struct drm_display_mode));
+	memcpy(&psb_intel_crtc->saved_adjusted_mode, adjusted_mode,
+					sizeof(struct drm_display_mode));
+
+	list_for_each_entry(connector, &mode_config->connector_list, head) {
+		if (!connector)
+			continue;
+
+		encoder = connector->encoder;
+
+		if (!encoder)
+			continue;
+
+		if (encoder->crtc != crtc)
+			continue;
+
+		psb_intel_encoder = psb_intel_attached_encoder(connector);
+
+		switch (psb_intel_encoder->type) {
+		case INTEL_OUTPUT_MIPI:
+			is_mipi = true;
+			break;
+		case INTEL_OUTPUT_MIPI2:
+			is_mipi2 = true;
+			break;
+		case INTEL_OUTPUT_HDMI:
+			is_hdmi = true;
+			break;
+		}
+	}
+
+	/* Disable the VGA plane that we never use */
+	REG_WRITE(VGACNTRL, VGA_DISP_DISABLE);
+
+	/* Disable the panel fitter if it was on our pipe */
+	if (psb_intel_panel_fitter_pipe(dev) == pipe)
+		REG_WRITE(PFIT_CONTROL, 0);
+
+	/* pipesrc and dspsize control the size that is scaled from,
+	 * which should always be the user's requested size.
+	 */
+	if (pipe == 1) {
+		/* FIXME: To make HDMI display with 864x480 (TPO), 480x864
+		 * (PYR) or 480x854 (TMD), set the sprite width/height and
+		 * souce image size registers with the adjusted mode for
+		 * pipe B.
+		 */
+
+		/*
+		 * The defined sprite rectangle must always be completely
+		 * contained within the displayable area of the screen image
+		 * (frame buffer).
+		 */
+		REG_WRITE(dspsize_reg, ((min(mode->crtc_vdisplay, adjusted_mode->crtc_vdisplay) - 1) << 16)
+				| (min(mode->crtc_hdisplay, adjusted_mode->crtc_hdisplay) - 1));
+		/* Set the CRTC with encoder mode. */
+		REG_WRITE(pipesrc_reg, ((mode->crtc_hdisplay - 1) << 16)
+				 | (mode->crtc_vdisplay - 1));
+	} else {
+		REG_WRITE(dspsize_reg,
+				((mode->crtc_vdisplay - 1) << 16) |
+						(mode->crtc_hdisplay - 1));
+		REG_WRITE(pipesrc_reg,
+				((mode->crtc_hdisplay - 1) << 16) |
+						(mode->crtc_vdisplay - 1));
+	}
+
+	REG_WRITE(dsppos_reg, 0);
+
+	if (psb_intel_encoder)
+		drm_connector_property_get_value(connector,
+			dev->mode_config.scaling_mode_property, &scalingType);
+
+	if (scalingType == DRM_MODE_SCALE_NO_SCALE) {
+		/* Medfield doesn't have register support for centering so we
+		 * need to mess with the h/vblank and h/vsync start and ends
+		 * to get centering
+		 */
+		int offsetX = 0, offsetY = 0;
+
+		offsetX = (adjusted_mode->crtc_hdisplay -
+					mode->crtc_hdisplay) / 2;
+		offsetY = (adjusted_mode->crtc_vdisplay -
+					mode->crtc_vdisplay) / 2;
+
+		REG_WRITE(htot_reg, (mode->crtc_hdisplay - 1) |
+			((adjusted_mode->crtc_htotal - 1) << 16));
+		REG_WRITE(vtot_reg, (mode->crtc_vdisplay - 1) |
+			((adjusted_mode->crtc_vtotal - 1) << 16));
+		REG_WRITE(hblank_reg, (adjusted_mode->crtc_hblank_start -
+								offsetX - 1) |
+			((adjusted_mode->crtc_hblank_end - offsetX - 1) << 16));
+		REG_WRITE(hsync_reg, (adjusted_mode->crtc_hsync_start -
+								offsetX - 1) |
+			((adjusted_mode->crtc_hsync_end - offsetX - 1) << 16));
+		REG_WRITE(vblank_reg, (adjusted_mode->crtc_vblank_start -
+								offsetY - 1) |
+			((adjusted_mode->crtc_vblank_end - offsetY - 1) << 16));
+		REG_WRITE(vsync_reg, (adjusted_mode->crtc_vsync_start -
+								offsetY - 1) |
+			((adjusted_mode->crtc_vsync_end - offsetY - 1) << 16));
+	} else {
+		REG_WRITE(htot_reg, (adjusted_mode->crtc_hdisplay - 1) |
+			((adjusted_mode->crtc_htotal - 1) << 16));
+		REG_WRITE(vtot_reg, (adjusted_mode->crtc_vdisplay - 1) |
+			((adjusted_mode->crtc_vtotal - 1) << 16));
+		REG_WRITE(hblank_reg, (adjusted_mode->crtc_hblank_start - 1) |
+			((adjusted_mode->crtc_hblank_end - 1) << 16));
+		REG_WRITE(hsync_reg, (adjusted_mode->crtc_hsync_start - 1) |
+			((adjusted_mode->crtc_hsync_end - 1) << 16));
+		REG_WRITE(vblank_reg, (adjusted_mode->crtc_vblank_start - 1) |
+			((adjusted_mode->crtc_vblank_end - 1) << 16));
+		REG_WRITE(vsync_reg, (adjusted_mode->crtc_vsync_start - 1) |
+			((adjusted_mode->crtc_vsync_end - 1) << 16));
+	}
+
+	/* Flush the plane changes */
+	{
+		struct drm_crtc_helper_funcs *crtc_funcs =
+		    crtc->helper_private;
+		crtc_funcs->mode_set_base(crtc, x, y, old_fb);
+	}
+
+	/* setup pipeconf */
+	*pipeconf = PIPEACONF_ENABLE; /* FIXME_JLIU7 REG_READ(pipeconf_reg); */
+
+	/* Set up the display plane register */
+	*dspcntr = REG_READ(dspcntr_reg);
+	*dspcntr |= pipe << DISPPLANE_SEL_PIPE_POS;
+	*dspcntr |= DISPLAY_PLANE_ENABLE;
+
+	if (is_mipi2)
+		goto mrst_crtc_mode_set_exit;
+	clk = adjusted_mode->clock;
+
+	if (is_hdmi) {
+		if ((ksel == KSEL_CRYSTAL_19) || (ksel == KSEL_BYPASS_19)) {
+			refclk = 19200;
+
+			if (is_mipi || is_mipi2)
+				clk_n = 1, clk_p2 = 8;
+			else if (is_hdmi)
+				clk_n = 1, clk_p2 = 10;
+		} else if (ksel == KSEL_BYPASS_25) {
+			refclk = 25000;
+
+			if (is_mipi || is_mipi2)
+				clk_n = 1, clk_p2 = 8;
+			else if (is_hdmi)
+				clk_n = 1, clk_p2 = 10;
+		} else if ((ksel == KSEL_BYPASS_83_100) &&
+					dev_priv->core_freq == 166) {
+			refclk = 83000;
+
+			if (is_mipi || is_mipi2)
+				clk_n = 4, clk_p2 = 8;
+			else if (is_hdmi)
+				clk_n = 4, clk_p2 = 10;
+		} else if ((ksel == KSEL_BYPASS_83_100) &&
+					(dev_priv->core_freq == 100 ||
+					dev_priv->core_freq == 200)) {
+			refclk = 100000;
+			if (is_mipi || is_mipi2)
+				clk_n = 4, clk_p2 = 8;
+			else if (is_hdmi)
+				clk_n = 4, clk_p2 = 10;
+		}
+
+		if (is_mipi)
+			clk_byte = dev_priv->bpp / 8;
+		else if (is_mipi2)
+			clk_byte = dev_priv->bpp2 / 8;
+
+		clk_tmp = clk * clk_n * clk_p2 * clk_byte;
+
+		dev_dbg(dev->dev, "clk = %d, clk_n = %d, clk_p2 = %d.\n",
+					clk, clk_n, clk_p2);
+		dev_dbg(dev->dev, "adjusted_mode->clock = %d, clk_tmp = %d.\n",
+					adjusted_mode->clock, clk_tmp);
+
+		ok = mdfldFindBestPLL(crtc, clk_tmp, refclk, &clock);
+
+		if (!ok) {
+			DRM_ERROR
+			    ("mdfldFindBestPLL fail in mdfld_crtc_mode_set.\n");
+		} else {
+			m_conv = mdfld_m_converts[(clock.m - MDFLD_M_MIN)];
+
+			dev_dbg(dev->dev, "dot clock = %d,"
+				 "m = %d, p1 = %d, m_conv = %d.\n",
+					clock.dot, clock.m,
+					clock.p1, m_conv);
+		}
+
+		dpll = REG_READ(dpll_reg);
+
+		if (dpll & DPLL_VCO_ENABLE) {
+			dpll &= ~DPLL_VCO_ENABLE;
+			REG_WRITE(dpll_reg, dpll);
+			REG_READ(dpll_reg);
+
+			/* FIXME jliu7 check the DPLL lock bit PIPEACONF[29] */
+			/* FIXME_MDFLD PO - change 500 to 1 after PO */
+			udelay(500);
+
+			/* reset M1, N1 & P1 */
+			REG_WRITE(fp_reg, 0);
+			dpll &= ~MDFLD_P1_MASK;
+			REG_WRITE(dpll_reg, dpll);
+			/* FIXME_MDFLD PO - change 500 to 1 after PO */
+			udelay(500);
+		}
+
+		/* When ungating power of DPLL, needs to wait 0.5us before
+		 * enable the VCO */
+		if (dpll & MDFLD_PWR_GATE_EN) {
+			dpll &= ~MDFLD_PWR_GATE_EN;
+			REG_WRITE(dpll_reg, dpll);
+			/* FIXME_MDFLD PO - change 500 to 1 after PO */
+			udelay(500);
+		}
+		dpll = 0;
+
+#if 0 /* FIXME revisit later */
+		if (ksel == KSEL_CRYSTAL_19 || ksel == KSEL_BYPASS_19 ||
+						ksel == KSEL_BYPASS_25)
+			dpll &= ~MDFLD_INPUT_REF_SEL;
+		else if (ksel == KSEL_BYPASS_83_100)
+			dpll |= MDFLD_INPUT_REF_SEL;
+#endif /* FIXME revisit later */
+
+		if (is_hdmi)
+			dpll |= MDFLD_VCO_SEL;
+
+		fp = (clk_n / 2) << 16;
+		fp |= m_conv;
+
+		/* compute bitmask from p1 value */
+		dpll |= (1 << (clock.p1 - 2)) << 17;
+
+#if 0 /* 1080p30 & 720p */
+		dpll = 0x00050000;
+		fp = 0x000001be;
+#endif
+#if 0 /* 480p */
+		dpll = 0x02010000;
+		fp = 0x000000d2;
+#endif
+	} else {
+#if 0 /*DBI_TPO_480x864*/
+		dpll = 0x00020000;
+		fp = 0x00000156;
+#endif /* DBI_TPO_480x864 */ /* get from spec. */
+
+		dpll = 0x00800000;
+		fp = 0x000000c1;
+	}
+
+	REG_WRITE(fp_reg, fp);
+	REG_WRITE(dpll_reg, dpll);
+	/* FIXME_MDFLD PO - change 500 to 1 after PO */
+	udelay(500);
+
+	dpll |= DPLL_VCO_ENABLE;
+	REG_WRITE(dpll_reg, dpll);
+	REG_READ(dpll_reg);
+
+	/* wait for DSI PLL to lock */
+	while (timeout < 20000 &&
+			!(REG_READ(pipeconf_reg) & PIPECONF_DSIPLL_LOCK)) {
+		udelay(150);
+		timeout++;
+	}
+
+	if (is_mipi)
+		goto mrst_crtc_mode_set_exit;
+
+	dev_dbg(dev->dev, "is_mipi = 0x%x\n", is_mipi);
+
+	REG_WRITE(pipeconf_reg, *pipeconf);
+	REG_READ(pipeconf_reg);
+
+	/* Wait for for the pipe enable to take effect. */
+	REG_WRITE(dspcntr_reg, *dspcntr);
+	psb_intel_wait_for_vblank(dev);
+
+mrst_crtc_mode_set_exit:
+
+	gma_power_end(dev);
+
+	return 0;
+}
+
+const struct drm_crtc_helper_funcs mdfld_helper_funcs = {
+	.dpms = mdfld_crtc_dpms,
+	.mode_fixup = psb_intel_crtc_mode_fixup,
+	.mode_set = mdfld_crtc_mode_set,
+	.mode_set_base = mdfld__intel_pipe_set_base,
+	.prepare = psb_intel_crtc_prepare,
+	.commit = psb_intel_crtc_commit,
+};
+
diff --git a/drivers/gpu/drm/gma500/mdfld_output.c b/drivers/gpu/drm/gma500/mdfld_output.c
new file mode 100644
index 000000000000..c95966bb0c96
--- /dev/null
+++ b/drivers/gpu/drm/gma500/mdfld_output.c
@@ -0,0 +1,74 @@
+/*
+ * Copyright (c)  2010 Intel Corporation
+ *
+ * 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, sublicensen
+ * 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:
+ * Thomas Eaton <thomas.g.eaton@intel.com>
+ * Scott Rowe <scott.m.rowe@intel.com>
+*/
+
+#include "mdfld_output.h"
+#include "mdfld_dsi_dpi.h"
+#include "mdfld_dsi_output.h"
+
+#include "tc35876x-dsi-lvds.h"
+
+int mdfld_get_panel_type(struct drm_device *dev, int pipe)
+{
+	struct drm_psb_private *dev_priv = dev->dev_private;
+	return dev_priv->mdfld_panel_id;
+}
+
+static void mdfld_init_panel(struct drm_device *dev, int mipi_pipe,
+								int p_type)
+{
+	switch (p_type) {
+	case TPO_VID:
+		mdfld_dsi_output_init(dev, mipi_pipe, &mdfld_tpo_vid_funcs);
+		break;
+	case TC35876X:
+		tc35876x_init(dev);
+		mdfld_dsi_output_init(dev, mipi_pipe, &mdfld_tc35876x_funcs);
+		break;
+	case TMD_VID:
+		mdfld_dsi_output_init(dev, mipi_pipe, &mdfld_tmd_vid_funcs);
+		break;
+	case HDMI:
+/*		if (dev_priv->mdfld_hdmi_present)
+			mdfld_hdmi_init(dev, &dev_priv->mode_dev); */
+		break;
+	}
+}
+
+
+int mdfld_output_init(struct drm_device *dev)
+{
+	struct drm_psb_private *dev_priv = dev->dev_private;
+
+	/* FIXME: hardcoded for now */
+	dev_priv->mdfld_panel_id = TC35876X;
+	/* MIPI panel 1 */
+	mdfld_init_panel(dev, 0, dev_priv->mdfld_panel_id);
+	/* HDMI panel */
+	mdfld_init_panel(dev, 1, HDMI);
+	return 0;
+}
+
diff --git a/drivers/gpu/drm/gma500/mdfld_output.h b/drivers/gpu/drm/gma500/mdfld_output.h
new file mode 100644
index 000000000000..ab2b27c0f037
--- /dev/null
+++ b/drivers/gpu/drm/gma500/mdfld_output.h
@@ -0,0 +1,77 @@
+/*
+ * Copyright (c)  2010 Intel Corporation
+ *
+ * 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, sublicensen
+ * 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:
+ * Thomas Eaton <thomas.g.eaton@intel.com>
+ * Scott Rowe <scott.m.rowe@intel.com>
+*/
+
+#ifndef MDFLD_OUTPUT_H
+#define MDFLD_OUTPUT_H
+
+#include "psb_drv.h"
+
+#define TPO_PANEL_WIDTH		84
+#define TPO_PANEL_HEIGHT	46
+#define TMD_PANEL_WIDTH		39
+#define TMD_PANEL_HEIGHT	71
+
+struct mdfld_dsi_config;
+
+enum panel_type {
+	TPO_VID,
+	TMD_VID,
+	HDMI,
+	TC35876X,
+};
+
+struct panel_info {
+	u32 width_mm;
+	u32 height_mm;
+	/* Other info */
+};
+
+struct panel_funcs {
+	const struct drm_encoder_funcs *encoder_funcs;
+	const struct drm_encoder_helper_funcs *encoder_helper_funcs;
+	struct drm_display_mode * (*get_config_mode)(struct drm_device *);
+	int (*get_panel_info)(struct drm_device *, int, struct panel_info *);
+	int (*reset)(int pipe);
+	void (*drv_ic_init)(struct mdfld_dsi_config *dsi_config, int pipe);
+};
+
+int mdfld_output_init(struct drm_device *dev);
+
+struct backlight_device *mdfld_get_backlight_device(void);
+int mdfld_set_brightness(struct backlight_device *bd);
+
+int mdfld_get_panel_type(struct drm_device *dev, int pipe);
+
+extern const struct drm_crtc_helper_funcs mdfld_helper_funcs;
+
+extern const struct panel_funcs mdfld_tmd_vid_funcs;
+extern const struct panel_funcs mdfld_tpo_vid_funcs;
+
+extern void mdfld_disable_crtc(struct drm_device *dev, int pipe);
+extern void mdfldWaitForPipeEnable(struct drm_device *dev, int pipe);
+extern void mdfldWaitForPipeDisable(struct drm_device *dev, int pipe);
+#endif
diff --git a/drivers/gpu/drm/gma500/mdfld_tmd_vid.c b/drivers/gpu/drm/gma500/mdfld_tmd_vid.c
new file mode 100644
index 000000000000..dc0c6c3d3d29
--- /dev/null
+++ b/drivers/gpu/drm/gma500/mdfld_tmd_vid.c
@@ -0,0 +1,201 @@
+/*
+ * Copyright © 2010 Intel Corporation
+ *
+ * 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:
+ * Jim Liu <jim.liu@intel.com>
+ * Jackie Li<yaodong.li@intel.com>
+ * Gideon Eaton <eaton.
+ * Scott Rowe <scott.m.rowe@intel.com>
+ */
+
+#include "mdfld_dsi_dpi.h"
+#include "mdfld_dsi_pkg_sender.h"
+
+static struct drm_display_mode *tmd_vid_get_config_mode(struct drm_device *dev)
+{
+	struct drm_display_mode *mode;
+	struct drm_psb_private *dev_priv = dev->dev_private;
+	struct oaktrail_timing_info *ti = &dev_priv->gct_data.DTD;
+	bool use_gct = false; /*Disable GCT for now*/
+
+	mode = kzalloc(sizeof(*mode), GFP_KERNEL);
+	if (!mode)
+		return NULL;
+
+	if (use_gct) {
+		mode->hdisplay = (ti->hactive_hi << 8) | ti->hactive_lo;
+		mode->vdisplay = (ti->vactive_hi << 8) | ti->vactive_lo;
+		mode->hsync_start = mode->hdisplay + \
+				((ti->hsync_offset_hi << 8) | \
+				ti->hsync_offset_lo);
+		mode->hsync_end = mode->hsync_start + \
+				((ti->hsync_pulse_width_hi << 8) | \
+				ti->hsync_pulse_width_lo);
+		mode->htotal = mode->hdisplay + ((ti->hblank_hi << 8) | \
+								ti->hblank_lo);
+		mode->vsync_start = \
+			mode->vdisplay + ((ti->vsync_offset_hi << 8) | \
+						ti->vsync_offset_lo);
+		mode->vsync_end = \
+			mode->vsync_start + ((ti->vsync_pulse_width_hi << 8) | \
+						ti->vsync_pulse_width_lo);
+		mode->vtotal = mode->vdisplay + \
+				((ti->vblank_hi << 8) | ti->vblank_lo);
+		mode->clock = ti->pixel_clock * 10;
+
+		dev_dbg(dev->dev, "hdisplay is %d\n", mode->hdisplay);
+		dev_dbg(dev->dev, "vdisplay is %d\n", mode->vdisplay);
+		dev_dbg(dev->dev, "HSS is %d\n", mode->hsync_start);
+		dev_dbg(dev->dev, "HSE is %d\n", mode->hsync_end);
+		dev_dbg(dev->dev, "htotal is %d\n", mode->htotal);
+		dev_dbg(dev->dev, "VSS is %d\n", mode->vsync_start);
+		dev_dbg(dev->dev, "VSE is %d\n", mode->vsync_end);
+		dev_dbg(dev->dev, "vtotal is %d\n", mode->vtotal);
+		dev_dbg(dev->dev, "clock is %d\n", mode->clock);
+	} else {
+		mode->hdisplay = 480;
+		mode->vdisplay = 854;
+		mode->hsync_start = 487;
+		mode->hsync_end = 490;
+		mode->htotal = 499;
+		mode->vsync_start = 861;
+		mode->vsync_end = 865;
+		mode->vtotal = 873;
+		mode->clock = 33264;
+	}
+
+	drm_mode_set_name(mode);
+	drm_mode_set_crtcinfo(mode, 0);
+
+	mode->type |= DRM_MODE_TYPE_PREFERRED;
+
+	return mode;
+}
+
+static int tmd_vid_get_panel_info(struct drm_device *dev,
+				int pipe,
+				struct panel_info *pi)
+{
+	if (!dev || !pi)
+		return -EINVAL;
+
+	pi->width_mm = TMD_PANEL_WIDTH;
+	pi->height_mm = TMD_PANEL_HEIGHT;
+
+	return 0;
+}
+
+/* ************************************************************************* *\
+ * FUNCTION: mdfld_init_TMD_MIPI
+ *
+ * DESCRIPTION:  This function is called only by mrst_dsi_mode_set and
+ *               restore_display_registers.  since this function does not
+ *               acquire the mutex, it is important that the calling function
+ *               does!
+\* ************************************************************************* */
+
+/* FIXME: make the below data u8 instead of u32; note byte order! */
+static u32 tmd_cmd_mcap_off[] = {0x000000b2};
+static u32 tmd_cmd_enable_lane_switch[] = {0x000101ef};
+static u32 tmd_cmd_set_lane_num[] = {0x006360ef};
+static u32 tmd_cmd_pushing_clock0[] = {0x00cc2fef};
+static u32 tmd_cmd_pushing_clock1[] = {0x00dd6eef};
+static u32 tmd_cmd_set_mode[] = {0x000000b3};
+static u32 tmd_cmd_set_sync_pulse_mode[] = {0x000961ef};
+static u32 tmd_cmd_set_column[] = {0x0100002a, 0x000000df};
+static u32 tmd_cmd_set_page[] = {0x0300002b, 0x00000055};
+static u32 tmd_cmd_set_video_mode[] = {0x00000153};
+/*no auto_bl,need add in furture*/
+static u32 tmd_cmd_enable_backlight[] = {0x00005ab4};
+static u32 tmd_cmd_set_backlight_dimming[] = {0x00000ebd};
+
+static void mdfld_dsi_tmd_drv_ic_init(struct mdfld_dsi_config *dsi_config,
+				      int pipe)
+{
+	struct mdfld_dsi_pkg_sender *sender
+			= mdfld_dsi_get_pkg_sender(dsi_config);
+
+	DRM_INFO("Enter mdfld init TMD MIPI display.\n");
+
+	if (!sender) {
+		DRM_ERROR("Cannot get sender\n");
+		return;
+	}
+
+	if (dsi_config->dvr_ic_inited)
+		return;
+
+	msleep(3);
+
+	/* FIXME: make the below data u8 instead of u32; note byte order! */
+
+	mdfld_dsi_send_gen_long(sender, (u8 *) tmd_cmd_mcap_off,
+				sizeof(tmd_cmd_mcap_off), false);
+	mdfld_dsi_send_gen_long(sender, (u8 *) tmd_cmd_enable_lane_switch,
+				sizeof(tmd_cmd_enable_lane_switch), false);
+	mdfld_dsi_send_gen_long(sender, (u8 *) tmd_cmd_set_lane_num,
+				sizeof(tmd_cmd_set_lane_num), false);
+	mdfld_dsi_send_gen_long(sender, (u8 *) tmd_cmd_pushing_clock0,
+				sizeof(tmd_cmd_pushing_clock0), false);
+	mdfld_dsi_send_gen_long(sender, (u8 *) tmd_cmd_pushing_clock1,
+				sizeof(tmd_cmd_pushing_clock1), false);
+	mdfld_dsi_send_gen_long(sender, (u8 *) tmd_cmd_set_mode,
+				sizeof(tmd_cmd_set_mode), false);
+	mdfld_dsi_send_gen_long(sender, (u8 *) tmd_cmd_set_sync_pulse_mode,
+				sizeof(tmd_cmd_set_sync_pulse_mode), false);
+	mdfld_dsi_send_mcs_long(sender, (u8 *) tmd_cmd_set_column,
+				sizeof(tmd_cmd_set_column), false);
+	mdfld_dsi_send_mcs_long(sender, (u8 *) tmd_cmd_set_page,
+				sizeof(tmd_cmd_set_page), false);
+	mdfld_dsi_send_gen_long(sender, (u8 *) tmd_cmd_set_video_mode,
+				sizeof(tmd_cmd_set_video_mode), false);
+	mdfld_dsi_send_gen_long(sender, (u8 *) tmd_cmd_enable_backlight,
+				sizeof(tmd_cmd_enable_backlight), false);
+	mdfld_dsi_send_gen_long(sender, (u8 *) tmd_cmd_set_backlight_dimming,
+				sizeof(tmd_cmd_set_backlight_dimming), false);
+
+	dsi_config->dvr_ic_inited = 1;
+}
+
+/*TPO DPI encoder helper funcs*/
+static const struct drm_encoder_helper_funcs
+				mdfld_tpo_dpi_encoder_helper_funcs = {
+	.dpms = mdfld_dsi_dpi_dpms,
+	.mode_fixup = mdfld_dsi_dpi_mode_fixup,
+	.prepare = mdfld_dsi_dpi_prepare,
+	.mode_set = mdfld_dsi_dpi_mode_set,
+	.commit = mdfld_dsi_dpi_commit,
+};
+
+/*TPO DPI encoder funcs*/
+static const struct drm_encoder_funcs mdfld_tpo_dpi_encoder_funcs = {
+	.destroy = drm_encoder_cleanup,
+};
+
+const struct panel_funcs mdfld_tmd_vid_funcs = {
+	.encoder_funcs = &mdfld_tpo_dpi_encoder_funcs,
+	.encoder_helper_funcs = &mdfld_tpo_dpi_encoder_helper_funcs,
+	.get_config_mode = &tmd_vid_get_config_mode,
+	.get_panel_info = tmd_vid_get_panel_info,
+	.reset = mdfld_dsi_panel_reset,
+	.drv_ic_init = mdfld_dsi_tmd_drv_ic_init,
+};
diff --git a/drivers/gpu/drm/gma500/mdfld_tpo_vid.c b/drivers/gpu/drm/gma500/mdfld_tpo_vid.c
new file mode 100644
index 000000000000..d8d4170725b2
--- /dev/null
+++ b/drivers/gpu/drm/gma500/mdfld_tpo_vid.c
@@ -0,0 +1,124 @@
+/*
+ * Copyright © 2010 Intel Corporation
+ *
+ * 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:
+ * jim liu <jim.liu@intel.com>
+ * Jackie Li<yaodong.li@intel.com>
+ */
+
+#include "mdfld_dsi_dpi.h"
+
+static struct drm_display_mode *tpo_vid_get_config_mode(struct drm_device *dev)
+{
+	struct drm_display_mode *mode;
+	struct drm_psb_private *dev_priv = dev->dev_private;
+	struct oaktrail_timing_info *ti = &dev_priv->gct_data.DTD;
+	bool use_gct = false;
+
+	mode = kzalloc(sizeof(*mode), GFP_KERNEL);
+	if (!mode)
+		return NULL;
+
+	if (use_gct) {
+		mode->hdisplay = (ti->hactive_hi << 8) | ti->hactive_lo;
+		mode->vdisplay = (ti->vactive_hi << 8) | ti->vactive_lo;
+		mode->hsync_start = mode->hdisplay +
+				((ti->hsync_offset_hi << 8) |
+				ti->hsync_offset_lo);
+		mode->hsync_end = mode->hsync_start +
+				((ti->hsync_pulse_width_hi << 8) |
+				ti->hsync_pulse_width_lo);
+		mode->htotal = mode->hdisplay + ((ti->hblank_hi << 8) |
+								ti->hblank_lo);
+		mode->vsync_start =
+			mode->vdisplay + ((ti->vsync_offset_hi << 8) |
+						ti->vsync_offset_lo);
+		mode->vsync_end =
+			mode->vsync_start + ((ti->vsync_pulse_width_hi << 8) |
+						ti->vsync_pulse_width_lo);
+		mode->vtotal = mode->vdisplay +
+				((ti->vblank_hi << 8) | ti->vblank_lo);
+		mode->clock = ti->pixel_clock * 10;
+
+		dev_dbg(dev->dev, "hdisplay is %d\n", mode->hdisplay);
+		dev_dbg(dev->dev, "vdisplay is %d\n", mode->vdisplay);
+		dev_dbg(dev->dev, "HSS is %d\n", mode->hsync_start);
+		dev_dbg(dev->dev, "HSE is %d\n", mode->hsync_end);
+		dev_dbg(dev->dev, "htotal is %d\n", mode->htotal);
+		dev_dbg(dev->dev, "VSS is %d\n", mode->vsync_start);
+		dev_dbg(dev->dev, "VSE is %d\n", mode->vsync_end);
+		dev_dbg(dev->dev, "vtotal is %d\n", mode->vtotal);
+		dev_dbg(dev->dev, "clock is %d\n", mode->clock);
+	} else {
+		mode->hdisplay = 864;
+		mode->vdisplay = 480;
+		mode->hsync_start = 873;
+		mode->hsync_end = 876;
+		mode->htotal = 887;
+		mode->vsync_start = 487;
+		mode->vsync_end = 490;
+		mode->vtotal = 499;
+		mode->clock = 33264;
+	}
+
+	drm_mode_set_name(mode);
+	drm_mode_set_crtcinfo(mode, 0);
+
+	mode->type |= DRM_MODE_TYPE_PREFERRED;
+
+	return mode;
+}
+
+static int tpo_vid_get_panel_info(struct drm_device *dev,
+				int pipe,
+				struct panel_info *pi)
+{
+	if (!dev || !pi)
+		return -EINVAL;
+
+	pi->width_mm = TPO_PANEL_WIDTH;
+	pi->height_mm = TPO_PANEL_HEIGHT;
+
+	return 0;
+}
+
+/*TPO DPI encoder helper funcs*/
+static const struct drm_encoder_helper_funcs
+				mdfld_tpo_dpi_encoder_helper_funcs = {
+	.dpms = mdfld_dsi_dpi_dpms,
+	.mode_fixup = mdfld_dsi_dpi_mode_fixup,
+	.prepare = mdfld_dsi_dpi_prepare,
+	.mode_set = mdfld_dsi_dpi_mode_set,
+	.commit = mdfld_dsi_dpi_commit,
+};
+
+/*TPO DPI encoder funcs*/
+static const struct drm_encoder_funcs mdfld_tpo_dpi_encoder_funcs = {
+	.destroy = drm_encoder_cleanup,
+};
+
+const struct panel_funcs mdfld_tpo_vid_funcs = {
+	.encoder_funcs = &mdfld_tpo_dpi_encoder_funcs,
+	.encoder_helper_funcs = &mdfld_tpo_dpi_encoder_helper_funcs,
+	.get_config_mode = &tpo_vid_get_config_mode,
+	.get_panel_info = tpo_vid_get_panel_info,
+};
diff --git a/drivers/gpu/drm/gma500/mmu.c b/drivers/gpu/drm/gma500/mmu.c
index c904d73b1de3..49bac41beefb 100644
--- a/drivers/gpu/drm/gma500/mmu.c
+++ b/drivers/gpu/drm/gma500/mmu.c
@@ -125,14 +125,14 @@ static void psb_page_clflush(struct psb_mmu_driver *driver, struct page* page)
 	int i;
 	uint8_t *clf;
 
-	clf = kmap_atomic(page, KM_USER0);
+	clf = kmap_atomic(page);
 	mb();
 	for (i = 0; i < clflush_count; ++i) {
 		psb_clflush(clf);
 		clf += clflush_add;
 	}
 	mb();
-	kunmap_atomic(clf, KM_USER0);
+	kunmap_atomic(clf);
 }
 
 static void psb_pages_clflush(struct psb_mmu_driver *driver,
@@ -270,7 +270,7 @@ out_err1:
 	return NULL;
 }
 
-void psb_mmu_free_pt(struct psb_mmu_pt *pt)
+static void psb_mmu_free_pt(struct psb_mmu_pt *pt)
 {
 	__free_page(pt->p);
 	kfree(pt);
@@ -325,7 +325,7 @@ static struct psb_mmu_pt *psb_mmu_alloc_pt(struct psb_mmu_pd *pd)
 
 	spin_lock(lock);
 
-	v = kmap_atomic(pt->p, KM_USER0);
+	v = kmap_atomic(pt->p);
 	clf = (uint8_t *) v;
 	ptes = (uint32_t *) v;
 	for (i = 0; i < (PAGE_SIZE / sizeof(uint32_t)); ++i)
@@ -341,7 +341,7 @@ static struct psb_mmu_pt *psb_mmu_alloc_pt(struct psb_mmu_pd *pd)
 		mb();
 	}
 
-	kunmap_atomic(v, KM_USER0);
+	kunmap_atomic(v);
 	spin_unlock(lock);
 
 	pt->count = 0;
@@ -351,7 +351,7 @@ static struct psb_mmu_pt *psb_mmu_alloc_pt(struct psb_mmu_pd *pd)
 	return pt;
 }
 
-struct psb_mmu_pt *psb_mmu_pt_alloc_map_lock(struct psb_mmu_pd *pd,
+static struct psb_mmu_pt *psb_mmu_pt_alloc_map_lock(struct psb_mmu_pd *pd,
 					     unsigned long addr)
 {
 	uint32_t index = psb_mmu_pd_index(addr);
@@ -376,18 +376,18 @@ struct psb_mmu_pt *psb_mmu_pt_alloc_map_lock(struct psb_mmu_pd *pd,
 			continue;
 		}
 
-		v = kmap_atomic(pd->p, KM_USER0);
+		v = kmap_atomic(pd->p);
 		pd->tables[index] = pt;
 		v[index] = (page_to_pfn(pt->p) << 12) | pd->pd_mask;
 		pt->index = index;
-		kunmap_atomic((void *) v, KM_USER0);
+		kunmap_atomic((void *) v);
 
 		if (pd->hw_context != -1) {
 			psb_mmu_clflush(pd->driver, (void *) &v[index]);
 			atomic_set(&pd->driver->needs_tlbflush, 1);
 		}
 	}
-	pt->v = kmap_atomic(pt->p, KM_USER0);
+	pt->v = kmap_atomic(pt->p);
 	return pt;
 }
 
@@ -404,7 +404,7 @@ static struct psb_mmu_pt *psb_mmu_pt_map_lock(struct psb_mmu_pd *pd,
 		spin_unlock(lock);
 		return NULL;
 	}
-	pt->v = kmap_atomic(pt->p, KM_USER0);
+	pt->v = kmap_atomic(pt->p);
 	return pt;
 }
 
@@ -413,9 +413,9 @@ static void psb_mmu_pt_unmap_unlock(struct psb_mmu_pt *pt)
 	struct psb_mmu_pd *pd = pt->pd;
 	uint32_t *v;
 
-	kunmap_atomic(pt->v, KM_USER0);
+	kunmap_atomic(pt->v);
 	if (pt->count == 0) {
-		v = kmap_atomic(pd->p, KM_USER0);
+		v = kmap_atomic(pd->p);
 		v[pt->index] = pd->invalid_pde;
 		pd->tables[pt->index] = NULL;
 
@@ -424,7 +424,7 @@ static void psb_mmu_pt_unmap_unlock(struct psb_mmu_pt *pt)
 					(void *) &v[pt->index]);
 			atomic_set(&pd->driver->needs_tlbflush, 1);
 		}
-		kunmap_atomic(pt->v, KM_USER0);
+		kunmap_atomic(pt->v);
 		spin_unlock(&pd->driver->lock);
 		psb_mmu_free_pt(pt);
 		return;
@@ -457,7 +457,7 @@ void psb_mmu_mirror_gtt(struct psb_mmu_pd *pd,
 	down_read(&driver->sem);
 	spin_lock(&driver->lock);
 
-	v = kmap_atomic(pd->p, KM_USER0);
+	v = kmap_atomic(pd->p);
 	v += start;
 
 	while (gtt_pages--) {
@@ -467,7 +467,7 @@ void psb_mmu_mirror_gtt(struct psb_mmu_pd *pd,
 
 	/*ttm_tt_cache_flush(&pd->p, num_pages);*/
 	psb_pages_clflush(pd->driver, &pd->p, num_pages);
-	kunmap_atomic(v, KM_USER0);
+	kunmap_atomic(v);
 	spin_unlock(&driver->lock);
 
 	if (pd->hw_context != -1)
@@ -488,15 +488,6 @@ struct psb_mmu_pd *psb_mmu_get_default_pd(struct psb_mmu_driver *driver)
 	return pd;
 }
 
-/* Returns the physical address of the PD shared by sgx/msvdx */
-uint32_t psb_get_default_pd_addr(struct psb_mmu_driver *driver)
-{
-	struct psb_mmu_pd *pd;
-
-	pd = psb_mmu_get_default_pd(driver);
-	return page_to_pfn(pd->p) << PAGE_SHIFT;
-}
-
 void psb_mmu_driver_takedown(struct psb_mmu_driver *driver)
 {
 	psb_mmu_free_pagedir(driver->default_pd);
@@ -830,9 +821,9 @@ int psb_mmu_virtual_to_pfn(struct psb_mmu_pd *pd, uint32_t virtual,
 		uint32_t *v;
 
 		spin_lock(lock);
-		v = kmap_atomic(pd->p, KM_USER0);
+		v = kmap_atomic(pd->p);
 		tmp = v[psb_mmu_pd_index(virtual)];
-		kunmap_atomic(v, KM_USER0);
+		kunmap_atomic(v);
 		spin_unlock(lock);
 
 		if (tmp != pd->invalid_pde || !(tmp & PSB_PTE_VALID) ||
diff --git a/drivers/gpu/drm/gma500/oaktrail_crtc.c b/drivers/gpu/drm/gma500/oaktrail_crtc.c
index 9d12a3ee1600..a39b0d0d680f 100644
--- a/drivers/gpu/drm/gma500/oaktrail_crtc.c
+++ b/drivers/gpu/drm/gma500/oaktrail_crtc.c
@@ -115,7 +115,7 @@ static void oaktrail_clock(int refclk, struct oaktrail_clock_t *clock)
 	clock->dot = (refclk * clock->m) / (14 * clock->p1);
 }
 
-void mrstPrintPll(char *prefix, struct oaktrail_clock_t *clock)
+static void mrstPrintPll(char *prefix, struct oaktrail_clock_t *clock)
 {
 	pr_debug("%s: dotclock = %d,  m = %d, p1 = %d.\n",
 	     prefix, clock->dot, clock->m, clock->p1);
@@ -169,7 +169,6 @@ static void oaktrail_crtc_dpms(struct drm_crtc *crtc, int mode)
 	int dspbase_reg = (pipe == 0) ? MRST_DSPABASE : DSPBBASE;
 	int pipeconf_reg = (pipe == 0) ? PIPEACONF : PIPEBCONF;
 	u32 temp;
-	bool enabled;
 
 	if (!gma_power_begin(dev, true))
 		return;
@@ -253,8 +252,6 @@ static void oaktrail_crtc_dpms(struct drm_crtc *crtc, int mode)
 		break;
 	}
 
-	enabled = crtc->enabled && mode != DRM_MODE_DPMS_OFF;
-
 	/*Set FIFO Watermarks*/
 	REG_WRITE(DSPARB, 0x3FFF);
 	REG_WRITE(DSPFW1, 0x3F88080A);
@@ -310,7 +307,7 @@ static int oaktrail_crtc_mode_set(struct drm_crtc *crtc,
 	struct oaktrail_clock_t clock;
 	u32 dpll = 0, fp = 0, dspcntr, pipeconf;
 	bool ok, is_sdvo = false;
-	bool is_crt = false, is_lvds = false, is_tv = false;
+	bool is_lvds = false;
 	bool is_mipi = false;
 	struct drm_mode_config *mode_config = &dev->mode_config;
 	struct psb_intel_encoder *psb_intel_encoder = NULL;
@@ -340,12 +337,6 @@ static int oaktrail_crtc_mode_set(struct drm_crtc *crtc,
 		case INTEL_OUTPUT_SDVO:
 			is_sdvo = true;
 			break;
-		case INTEL_OUTPUT_TVOUT:
-			is_tv = true;
-			break;
-		case INTEL_OUTPUT_ANALOG:
-			is_crt = true;
-			break;
 		case INTEL_OUTPUT_MIPI:
 			is_mipi = true;
 			break;
@@ -428,9 +419,6 @@ static int oaktrail_crtc_mode_set(struct drm_crtc *crtc,
 	else
 		dspcntr |= DISPPLANE_SEL_PIPE_B;
 
-	dev_priv->dspcntr = dspcntr |= DISPLAY_PLANE_ENABLE;
-	dev_priv->pipeconf = pipeconf |= PIPEACONF_ENABLE;
-
 	if (is_mipi)
 		goto oaktrail_crtc_mode_set_exit;
 
@@ -517,7 +505,7 @@ static bool oaktrail_crtc_mode_fixup(struct drm_crtc *crtc,
 	return true;
 }
 
-int oaktrail_pipe_set_base(struct drm_crtc *crtc,
+static int oaktrail_pipe_set_base(struct drm_crtc *crtc,
 			    int x, int y, struct drm_framebuffer *old_fb)
 {
 	struct drm_device *dev = crtc->dev;
diff --git a/drivers/gpu/drm/gma500/oaktrail_device.c b/drivers/gpu/drm/gma500/oaktrail_device.c
index 63aea2f010d9..41d1924ea31e 100644
--- a/drivers/gpu/drm/gma500/oaktrail_device.c
+++ b/drivers/gpu/drm/gma500/oaktrail_device.c
@@ -141,7 +141,7 @@ static const struct backlight_ops oaktrail_ops = {
 	.update_status  = oaktrail_set_brightness,
 };
 
-int oaktrail_backlight_init(struct drm_device *dev)
+static int oaktrail_backlight_init(struct drm_device *dev)
 {
 	struct drm_psb_private *dev_priv = dev->dev_private;
 	int ret;
@@ -176,10 +176,6 @@ int oaktrail_backlight_init(struct drm_device *dev)
  *	for power management
  */
 
-static void oaktrail_init_pm(struct drm_device *dev)
-{
-}
-
 /**
  *	oaktrail_save_display_registers	-	save registers lost on suspend
  *	@dev: our DRM device
@@ -190,81 +186,82 @@ static void oaktrail_init_pm(struct drm_device *dev)
 static int oaktrail_save_display_registers(struct drm_device *dev)
 {
 	struct drm_psb_private *dev_priv = dev->dev_private;
+	struct psb_save_area *regs = &dev_priv->regs;
 	int i;
 	u32 pp_stat;
 
 	/* Display arbitration control + watermarks */
-	dev_priv->saveDSPARB = PSB_RVDC32(DSPARB);
-	dev_priv->saveDSPFW1 = PSB_RVDC32(DSPFW1);
-	dev_priv->saveDSPFW2 = PSB_RVDC32(DSPFW2);
-	dev_priv->saveDSPFW3 = PSB_RVDC32(DSPFW3);
-	dev_priv->saveDSPFW4 = PSB_RVDC32(DSPFW4);
-	dev_priv->saveDSPFW5 = PSB_RVDC32(DSPFW5);
-	dev_priv->saveDSPFW6 = PSB_RVDC32(DSPFW6);
-	dev_priv->saveCHICKENBIT = PSB_RVDC32(DSPCHICKENBIT);
+	regs->psb.saveDSPARB = PSB_RVDC32(DSPARB);
+	regs->psb.saveDSPFW1 = PSB_RVDC32(DSPFW1);
+	regs->psb.saveDSPFW2 = PSB_RVDC32(DSPFW2);
+	regs->psb.saveDSPFW3 = PSB_RVDC32(DSPFW3);
+	regs->psb.saveDSPFW4 = PSB_RVDC32(DSPFW4);
+	regs->psb.saveDSPFW5 = PSB_RVDC32(DSPFW5);
+	regs->psb.saveDSPFW6 = PSB_RVDC32(DSPFW6);
+	regs->psb.saveCHICKENBIT = PSB_RVDC32(DSPCHICKENBIT);
 
 	/* Pipe & plane A info */
-	dev_priv->savePIPEACONF = PSB_RVDC32(PIPEACONF);
-	dev_priv->savePIPEASRC = PSB_RVDC32(PIPEASRC);
-	dev_priv->saveFPA0 = PSB_RVDC32(MRST_FPA0);
-	dev_priv->saveFPA1 = PSB_RVDC32(MRST_FPA1);
-	dev_priv->saveDPLL_A = PSB_RVDC32(MRST_DPLL_A);
-	dev_priv->saveHTOTAL_A = PSB_RVDC32(HTOTAL_A);
-	dev_priv->saveHBLANK_A = PSB_RVDC32(HBLANK_A);
-	dev_priv->saveHSYNC_A = PSB_RVDC32(HSYNC_A);
-	dev_priv->saveVTOTAL_A = PSB_RVDC32(VTOTAL_A);
-	dev_priv->saveVBLANK_A = PSB_RVDC32(VBLANK_A);
-	dev_priv->saveVSYNC_A = PSB_RVDC32(VSYNC_A);
-	dev_priv->saveBCLRPAT_A = PSB_RVDC32(BCLRPAT_A);
-	dev_priv->saveDSPACNTR = PSB_RVDC32(DSPACNTR);
-	dev_priv->saveDSPASTRIDE = PSB_RVDC32(DSPASTRIDE);
-	dev_priv->saveDSPAADDR = PSB_RVDC32(DSPABASE);
-	dev_priv->saveDSPASURF = PSB_RVDC32(DSPASURF);
-	dev_priv->saveDSPALINOFF = PSB_RVDC32(DSPALINOFF);
-	dev_priv->saveDSPATILEOFF = PSB_RVDC32(DSPATILEOFF);
+	regs->psb.savePIPEACONF = PSB_RVDC32(PIPEACONF);
+	regs->psb.savePIPEASRC = PSB_RVDC32(PIPEASRC);
+	regs->psb.saveFPA0 = PSB_RVDC32(MRST_FPA0);
+	regs->psb.saveFPA1 = PSB_RVDC32(MRST_FPA1);
+	regs->psb.saveDPLL_A = PSB_RVDC32(MRST_DPLL_A);
+	regs->psb.saveHTOTAL_A = PSB_RVDC32(HTOTAL_A);
+	regs->psb.saveHBLANK_A = PSB_RVDC32(HBLANK_A);
+	regs->psb.saveHSYNC_A = PSB_RVDC32(HSYNC_A);
+	regs->psb.saveVTOTAL_A = PSB_RVDC32(VTOTAL_A);
+	regs->psb.saveVBLANK_A = PSB_RVDC32(VBLANK_A);
+	regs->psb.saveVSYNC_A = PSB_RVDC32(VSYNC_A);
+	regs->psb.saveBCLRPAT_A = PSB_RVDC32(BCLRPAT_A);
+	regs->psb.saveDSPACNTR = PSB_RVDC32(DSPACNTR);
+	regs->psb.saveDSPASTRIDE = PSB_RVDC32(DSPASTRIDE);
+	regs->psb.saveDSPAADDR = PSB_RVDC32(DSPABASE);
+	regs->psb.saveDSPASURF = PSB_RVDC32(DSPASURF);
+	regs->psb.saveDSPALINOFF = PSB_RVDC32(DSPALINOFF);
+	regs->psb.saveDSPATILEOFF = PSB_RVDC32(DSPATILEOFF);
 
 	/* Save cursor regs */
-	dev_priv->saveDSPACURSOR_CTRL = PSB_RVDC32(CURACNTR);
-	dev_priv->saveDSPACURSOR_BASE = PSB_RVDC32(CURABASE);
-	dev_priv->saveDSPACURSOR_POS = PSB_RVDC32(CURAPOS);
+	regs->psb.saveDSPACURSOR_CTRL = PSB_RVDC32(CURACNTR);
+	regs->psb.saveDSPACURSOR_BASE = PSB_RVDC32(CURABASE);
+	regs->psb.saveDSPACURSOR_POS = PSB_RVDC32(CURAPOS);
 
 	/* Save palette (gamma) */
 	for (i = 0; i < 256; i++)
-		dev_priv->save_palette_a[i] = PSB_RVDC32(PALETTE_A + (i << 2));
+		regs->psb.save_palette_a[i] = PSB_RVDC32(PALETTE_A + (i << 2));
 
 	if (dev_priv->hdmi_priv)
 		oaktrail_hdmi_save(dev);
 
 	/* Save performance state */
-	dev_priv->savePERF_MODE = PSB_RVDC32(MRST_PERF_MODE);
+	regs->psb.savePERF_MODE = PSB_RVDC32(MRST_PERF_MODE);
 
 	/* LVDS state */
-	dev_priv->savePP_CONTROL = PSB_RVDC32(PP_CONTROL);
-	dev_priv->savePFIT_PGM_RATIOS = PSB_RVDC32(PFIT_PGM_RATIOS);
-	dev_priv->savePFIT_AUTO_RATIOS = PSB_RVDC32(PFIT_AUTO_RATIOS);
-	dev_priv->saveBLC_PWM_CTL = PSB_RVDC32(BLC_PWM_CTL);
-	dev_priv->saveBLC_PWM_CTL2 = PSB_RVDC32(BLC_PWM_CTL2);
-	dev_priv->saveLVDS = PSB_RVDC32(LVDS);
-	dev_priv->savePFIT_CONTROL = PSB_RVDC32(PFIT_CONTROL);
-	dev_priv->savePP_ON_DELAYS = PSB_RVDC32(LVDSPP_ON);
-	dev_priv->savePP_OFF_DELAYS = PSB_RVDC32(LVDSPP_OFF);
-	dev_priv->savePP_DIVISOR = PSB_RVDC32(PP_CYCLE);
+	regs->psb.savePP_CONTROL = PSB_RVDC32(PP_CONTROL);
+	regs->psb.savePFIT_PGM_RATIOS = PSB_RVDC32(PFIT_PGM_RATIOS);
+	regs->psb.savePFIT_AUTO_RATIOS = PSB_RVDC32(PFIT_AUTO_RATIOS);
+	regs->saveBLC_PWM_CTL = PSB_RVDC32(BLC_PWM_CTL);
+	regs->saveBLC_PWM_CTL2 = PSB_RVDC32(BLC_PWM_CTL2);
+	regs->psb.saveLVDS = PSB_RVDC32(LVDS);
+	regs->psb.savePFIT_CONTROL = PSB_RVDC32(PFIT_CONTROL);
+	regs->psb.savePP_ON_DELAYS = PSB_RVDC32(LVDSPP_ON);
+	regs->psb.savePP_OFF_DELAYS = PSB_RVDC32(LVDSPP_OFF);
+	regs->psb.savePP_DIVISOR = PSB_RVDC32(PP_CYCLE);
 
 	/* HW overlay */
-	dev_priv->saveOV_OVADD = PSB_RVDC32(OV_OVADD);
-	dev_priv->saveOV_OGAMC0 = PSB_RVDC32(OV_OGAMC0);
-	dev_priv->saveOV_OGAMC1 = PSB_RVDC32(OV_OGAMC1);
-	dev_priv->saveOV_OGAMC2 = PSB_RVDC32(OV_OGAMC2);
-	dev_priv->saveOV_OGAMC3 = PSB_RVDC32(OV_OGAMC3);
-	dev_priv->saveOV_OGAMC4 = PSB_RVDC32(OV_OGAMC4);
-	dev_priv->saveOV_OGAMC5 = PSB_RVDC32(OV_OGAMC5);
+	regs->psb.saveOV_OVADD = PSB_RVDC32(OV_OVADD);
+	regs->psb.saveOV_OGAMC0 = PSB_RVDC32(OV_OGAMC0);
+	regs->psb.saveOV_OGAMC1 = PSB_RVDC32(OV_OGAMC1);
+	regs->psb.saveOV_OGAMC2 = PSB_RVDC32(OV_OGAMC2);
+	regs->psb.saveOV_OGAMC3 = PSB_RVDC32(OV_OGAMC3);
+	regs->psb.saveOV_OGAMC4 = PSB_RVDC32(OV_OGAMC4);
+	regs->psb.saveOV_OGAMC5 = PSB_RVDC32(OV_OGAMC5);
 
 	/* DPST registers */
-	dev_priv->saveHISTOGRAM_INT_CONTROL_REG =
+	regs->psb.saveHISTOGRAM_INT_CONTROL_REG =
 					PSB_RVDC32(HISTOGRAM_INT_CONTROL);
-	dev_priv->saveHISTOGRAM_LOGIC_CONTROL_REG =
+	regs->psb.saveHISTOGRAM_LOGIC_CONTROL_REG =
 					PSB_RVDC32(HISTOGRAM_LOGIC_CONTROL);
-	dev_priv->savePWM_CONTROL_LOGIC = PSB_RVDC32(PWM_CONTROL_LOGIC);
+	regs->psb.savePWM_CONTROL_LOGIC = PSB_RVDC32(PWM_CONTROL_LOGIC);
 
 	if (dev_priv->iLVDS_enable) {
 		/* Shut down the panel */
@@ -302,79 +299,80 @@ static int oaktrail_save_display_registers(struct drm_device *dev)
 static int oaktrail_restore_display_registers(struct drm_device *dev)
 {
 	struct drm_psb_private *dev_priv = dev->dev_private;
+	struct psb_save_area *regs = &dev_priv->regs;
 	u32 pp_stat;
 	int i;
 
 	/* Display arbitration + watermarks */
-	PSB_WVDC32(dev_priv->saveDSPARB, DSPARB);
-	PSB_WVDC32(dev_priv->saveDSPFW1, DSPFW1);
-	PSB_WVDC32(dev_priv->saveDSPFW2, DSPFW2);
-	PSB_WVDC32(dev_priv->saveDSPFW3, DSPFW3);
-	PSB_WVDC32(dev_priv->saveDSPFW4, DSPFW4);
-	PSB_WVDC32(dev_priv->saveDSPFW5, DSPFW5);
-	PSB_WVDC32(dev_priv->saveDSPFW6, DSPFW6);
-	PSB_WVDC32(dev_priv->saveCHICKENBIT, DSPCHICKENBIT);
+	PSB_WVDC32(regs->psb.saveDSPARB, DSPARB);
+	PSB_WVDC32(regs->psb.saveDSPFW1, DSPFW1);
+	PSB_WVDC32(regs->psb.saveDSPFW2, DSPFW2);
+	PSB_WVDC32(regs->psb.saveDSPFW3, DSPFW3);
+	PSB_WVDC32(regs->psb.saveDSPFW4, DSPFW4);
+	PSB_WVDC32(regs->psb.saveDSPFW5, DSPFW5);
+	PSB_WVDC32(regs->psb.saveDSPFW6, DSPFW6);
+	PSB_WVDC32(regs->psb.saveCHICKENBIT, DSPCHICKENBIT);
 
 	/* Make sure VGA plane is off. it initializes to on after reset!*/
 	PSB_WVDC32(0x80000000, VGACNTRL);
 
 	/* set the plls */
-	PSB_WVDC32(dev_priv->saveFPA0, MRST_FPA0);
-	PSB_WVDC32(dev_priv->saveFPA1, MRST_FPA1);
+	PSB_WVDC32(regs->psb.saveFPA0, MRST_FPA0);
+	PSB_WVDC32(regs->psb.saveFPA1, MRST_FPA1);
 
 	/* Actually enable it */
-	PSB_WVDC32(dev_priv->saveDPLL_A, MRST_DPLL_A);
+	PSB_WVDC32(regs->psb.saveDPLL_A, MRST_DPLL_A);
 	DRM_UDELAY(150);
 
 	/* Restore mode */
-	PSB_WVDC32(dev_priv->saveHTOTAL_A, HTOTAL_A);
-	PSB_WVDC32(dev_priv->saveHBLANK_A, HBLANK_A);
-	PSB_WVDC32(dev_priv->saveHSYNC_A, HSYNC_A);
-	PSB_WVDC32(dev_priv->saveVTOTAL_A, VTOTAL_A);
-	PSB_WVDC32(dev_priv->saveVBLANK_A, VBLANK_A);
-	PSB_WVDC32(dev_priv->saveVSYNC_A, VSYNC_A);
-	PSB_WVDC32(dev_priv->savePIPEASRC, PIPEASRC);
-	PSB_WVDC32(dev_priv->saveBCLRPAT_A, BCLRPAT_A);
+	PSB_WVDC32(regs->psb.saveHTOTAL_A, HTOTAL_A);
+	PSB_WVDC32(regs->psb.saveHBLANK_A, HBLANK_A);
+	PSB_WVDC32(regs->psb.saveHSYNC_A, HSYNC_A);
+	PSB_WVDC32(regs->psb.saveVTOTAL_A, VTOTAL_A);
+	PSB_WVDC32(regs->psb.saveVBLANK_A, VBLANK_A);
+	PSB_WVDC32(regs->psb.saveVSYNC_A, VSYNC_A);
+	PSB_WVDC32(regs->psb.savePIPEASRC, PIPEASRC);
+	PSB_WVDC32(regs->psb.saveBCLRPAT_A, BCLRPAT_A);
 
 	/* Restore performance mode*/
-	PSB_WVDC32(dev_priv->savePERF_MODE, MRST_PERF_MODE);
+	PSB_WVDC32(regs->psb.savePERF_MODE, MRST_PERF_MODE);
 
 	/* Enable the pipe*/
 	if (dev_priv->iLVDS_enable)
-		PSB_WVDC32(dev_priv->savePIPEACONF, PIPEACONF);
+		PSB_WVDC32(regs->psb.savePIPEACONF, PIPEACONF);
 
 	/* Set up the plane*/
-	PSB_WVDC32(dev_priv->saveDSPALINOFF, DSPALINOFF);
-	PSB_WVDC32(dev_priv->saveDSPASTRIDE, DSPASTRIDE);
-	PSB_WVDC32(dev_priv->saveDSPATILEOFF, DSPATILEOFF);
+	PSB_WVDC32(regs->psb.saveDSPALINOFF, DSPALINOFF);
+	PSB_WVDC32(regs->psb.saveDSPASTRIDE, DSPASTRIDE);
+	PSB_WVDC32(regs->psb.saveDSPATILEOFF, DSPATILEOFF);
 
 	/* Enable the plane */
-	PSB_WVDC32(dev_priv->saveDSPACNTR, DSPACNTR);
-	PSB_WVDC32(dev_priv->saveDSPASURF, DSPASURF);
+	PSB_WVDC32(regs->psb.saveDSPACNTR, DSPACNTR);
+	PSB_WVDC32(regs->psb.saveDSPASURF, DSPASURF);
 
 	/* Enable Cursor A */
-	PSB_WVDC32(dev_priv->saveDSPACURSOR_CTRL, CURACNTR);
-	PSB_WVDC32(dev_priv->saveDSPACURSOR_POS, CURAPOS);
-	PSB_WVDC32(dev_priv->saveDSPACURSOR_BASE, CURABASE);
+	PSB_WVDC32(regs->psb.saveDSPACURSOR_CTRL, CURACNTR);
+	PSB_WVDC32(regs->psb.saveDSPACURSOR_POS, CURAPOS);
+	PSB_WVDC32(regs->psb.saveDSPACURSOR_BASE, CURABASE);
 
 	/* Restore palette (gamma) */
 	for (i = 0; i < 256; i++)
-		PSB_WVDC32(dev_priv->save_palette_a[i], PALETTE_A + (i << 2));
+		PSB_WVDC32(regs->psb.save_palette_a[i], PALETTE_A + (i << 2));
 
 	if (dev_priv->hdmi_priv)
 		oaktrail_hdmi_restore(dev);
 
 	if (dev_priv->iLVDS_enable) {
-		PSB_WVDC32(dev_priv->saveBLC_PWM_CTL2, BLC_PWM_CTL2);
-		PSB_WVDC32(dev_priv->saveLVDS, LVDS); /*port 61180h*/
-		PSB_WVDC32(dev_priv->savePFIT_CONTROL, PFIT_CONTROL);
-		PSB_WVDC32(dev_priv->savePFIT_PGM_RATIOS, PFIT_PGM_RATIOS);
-		PSB_WVDC32(dev_priv->savePFIT_AUTO_RATIOS, PFIT_AUTO_RATIOS);
-		PSB_WVDC32(dev_priv->saveBLC_PWM_CTL, BLC_PWM_CTL);
-		PSB_WVDC32(dev_priv->savePP_ON_DELAYS, LVDSPP_ON);
-		PSB_WVDC32(dev_priv->savePP_OFF_DELAYS, LVDSPP_OFF);
-		PSB_WVDC32(dev_priv->savePP_DIVISOR, PP_CYCLE);
-		PSB_WVDC32(dev_priv->savePP_CONTROL, PP_CONTROL);
+		PSB_WVDC32(regs->saveBLC_PWM_CTL2, BLC_PWM_CTL2);
+		PSB_WVDC32(regs->psb.saveLVDS, LVDS); /*port 61180h*/
+		PSB_WVDC32(regs->psb.savePFIT_CONTROL, PFIT_CONTROL);
+		PSB_WVDC32(regs->psb.savePFIT_PGM_RATIOS, PFIT_PGM_RATIOS);
+		PSB_WVDC32(regs->psb.savePFIT_AUTO_RATIOS, PFIT_AUTO_RATIOS);
+		PSB_WVDC32(regs->saveBLC_PWM_CTL, BLC_PWM_CTL);
+		PSB_WVDC32(regs->psb.savePP_ON_DELAYS, LVDSPP_ON);
+		PSB_WVDC32(regs->psb.savePP_OFF_DELAYS, LVDSPP_OFF);
+		PSB_WVDC32(regs->psb.savePP_DIVISOR, PP_CYCLE);
+		PSB_WVDC32(regs->psb.savePP_CONTROL, PP_CONTROL);
 	}
 
 	/* Wait for cycle delay */
@@ -388,20 +386,20 @@ static int oaktrail_restore_display_registers(struct drm_device *dev)
 	} while (pp_stat & 0x10000000);
 
 	/* Restore HW overlay */
-	PSB_WVDC32(dev_priv->saveOV_OVADD, OV_OVADD);
-	PSB_WVDC32(dev_priv->saveOV_OGAMC0, OV_OGAMC0);
-	PSB_WVDC32(dev_priv->saveOV_OGAMC1, OV_OGAMC1);
-	PSB_WVDC32(dev_priv->saveOV_OGAMC2, OV_OGAMC2);
-	PSB_WVDC32(dev_priv->saveOV_OGAMC3, OV_OGAMC3);
-	PSB_WVDC32(dev_priv->saveOV_OGAMC4, OV_OGAMC4);
-	PSB_WVDC32(dev_priv->saveOV_OGAMC5, OV_OGAMC5);
+	PSB_WVDC32(regs->psb.saveOV_OVADD, OV_OVADD);
+	PSB_WVDC32(regs->psb.saveOV_OGAMC0, OV_OGAMC0);
+	PSB_WVDC32(regs->psb.saveOV_OGAMC1, OV_OGAMC1);
+	PSB_WVDC32(regs->psb.saveOV_OGAMC2, OV_OGAMC2);
+	PSB_WVDC32(regs->psb.saveOV_OGAMC3, OV_OGAMC3);
+	PSB_WVDC32(regs->psb.saveOV_OGAMC4, OV_OGAMC4);
+	PSB_WVDC32(regs->psb.saveOV_OGAMC5, OV_OGAMC5);
 
 	/* DPST registers */
-	PSB_WVDC32(dev_priv->saveHISTOGRAM_INT_CONTROL_REG,
+	PSB_WVDC32(regs->psb.saveHISTOGRAM_INT_CONTROL_REG,
 						HISTOGRAM_INT_CONTROL);
-	PSB_WVDC32(dev_priv->saveHISTOGRAM_LOGIC_CONTROL_REG,
+	PSB_WVDC32(regs->psb.saveHISTOGRAM_LOGIC_CONTROL_REG,
 						HISTOGRAM_LOGIC_CONTROL);
-	PSB_WVDC32(dev_priv->savePWM_CONTROL_LOGIC, PWM_CONTROL_LOGIC);
+	PSB_WVDC32(regs->psb.savePWM_CONTROL_LOGIC, PWM_CONTROL_LOGIC);
 
 	return 0;
 }
@@ -502,7 +500,6 @@ const struct psb_ops oaktrail_chip_ops = {
 	.backlight_init = oaktrail_backlight_init,
 #endif
 
-	.init_pm = oaktrail_init_pm,
 	.save_regs = oaktrail_save_display_registers,
 	.restore_regs = oaktrail_restore_display_registers,
 	.power_down = oaktrail_power_down,
diff --git a/drivers/gpu/drm/gma500/oaktrail_hdmi.c b/drivers/gpu/drm/gma500/oaktrail_hdmi.c
index 025d30970cc0..f8b367b45f66 100644
--- a/drivers/gpu/drm/gma500/oaktrail_hdmi.c
+++ b/drivers/gpu/drm/gma500/oaktrail_hdmi.c
@@ -125,59 +125,6 @@ static const struct oaktrail_hdmi_limit oaktrail_hdmi_limit = {
 	.nf  = { .min = NF_MIN,			.max = NF_MAX  },
 };
 
-static void wait_for_vblank(struct drm_device *dev)
-{
-	/* FIXME: Can we do this as a sleep ? */
-	/* Wait for 20ms, i.e. one cycle at 50hz. */
-	mdelay(20);
-}
-
-static void scu_busy_loop(void *scu_base)
-{
-	u32 status = 0;
-	u32 loop_count = 0;
-
-	status = readl(scu_base + 0x04);
-	while (status & 1) {
-		udelay(1); /* scu processing time is in few u secods */
-		status = readl(scu_base + 0x04);
-		loop_count++;
-		/* break if scu doesn't reset busy bit after huge retry */
-		if (loop_count > 1000) {
-			DRM_DEBUG_KMS("SCU IPC timed out");
-			return;
-		}
-	}
-}
-
-static void oaktrail_hdmi_reset(struct drm_device *dev)
-{
-	void *base;
-	/* FIXME: at least make these defines */
-	unsigned int scu_ipc_mmio = 0xff11c000;
-	int scu_len = 1024;
-
-	base = ioremap((resource_size_t)scu_ipc_mmio, scu_len);
-	if (base == NULL) {
-		DRM_ERROR("failed to map SCU mmio\n");
-		return;
-	}
-
-	/* scu ipc: assert hdmi controller reset */
-	writel(0xff11d118, base + 0x0c);
-	writel(0x7fffffdf, base + 0x80);
-	writel(0x42005, base + 0x0);
-	scu_busy_loop(base);
-
-	/* scu ipc: de-assert hdmi controller reset */
-	writel(0xff11d118, base + 0x0c);
-	writel(0x7fffffff, base + 0x80);
-	writel(0x42005, base + 0x0);
-	scu_busy_loop(base);
-
-	iounmap(base);
-}
-
 static void oaktrail_hdmi_audio_enable(struct drm_device *dev)
 {
 	struct drm_psb_private *dev_priv = dev->dev_private;
@@ -208,104 +155,6 @@ static void oaktrail_hdmi_audio_disable(struct drm_device *dev)
 	HDMI_READ(HDMI_HCR);
 }
 
-void oaktrail_crtc_hdmi_dpms(struct drm_crtc *crtc, int mode)
-{
-	struct drm_device *dev = crtc->dev;
-	u32 temp;
-
-	switch (mode) {
-	case DRM_MODE_DPMS_OFF:
-		/* Disable VGACNTRL */
-		REG_WRITE(VGACNTRL, 0x80000000);
-
-		/* Disable plane */
-		temp = REG_READ(DSPBCNTR);
-		if ((temp & DISPLAY_PLANE_ENABLE) != 0) {
-			REG_WRITE(DSPBCNTR, temp & ~DISPLAY_PLANE_ENABLE);
-			REG_READ(DSPBCNTR);
-			/* Flush the plane changes */
-			REG_WRITE(DSPBSURF, REG_READ(DSPBSURF));
-			REG_READ(DSPBSURF);
-		}
-
-		/* Disable pipe B */
-		temp = REG_READ(PIPEBCONF);
-		if ((temp & PIPEACONF_ENABLE) != 0) {
-			REG_WRITE(PIPEBCONF, temp & ~PIPEACONF_ENABLE);
-			REG_READ(PIPEBCONF);
-		}
-
-		/* Disable LNW Pipes, etc */
-		temp = REG_READ(PCH_PIPEBCONF);
-		if ((temp & PIPEACONF_ENABLE) != 0) {
-			REG_WRITE(PCH_PIPEBCONF, temp & ~PIPEACONF_ENABLE);
-			REG_READ(PCH_PIPEBCONF);
-		}
-		/* wait for pipe off */
-		udelay(150);
-		/* Disable dpll */
-		temp = REG_READ(DPLL_CTRL);
-		if ((temp & DPLL_PWRDN) == 0) {
-			REG_WRITE(DPLL_CTRL, temp | (DPLL_PWRDN | DPLL_RESET));
-			REG_WRITE(DPLL_STATUS, 0x1);
-		}
-		/* wait for dpll off */
-		udelay(150);
-		break;
-	case DRM_MODE_DPMS_ON:
-	case DRM_MODE_DPMS_STANDBY:
-	case DRM_MODE_DPMS_SUSPEND:
-		/* Enable dpll */
-		temp = REG_READ(DPLL_CTRL);
-		if ((temp & DPLL_PWRDN) != 0) {
-			REG_WRITE(DPLL_CTRL, temp & ~(DPLL_PWRDN | DPLL_RESET));
-			temp = REG_READ(DPLL_CLK_ENABLE);
-			REG_WRITE(DPLL_CLK_ENABLE, temp | DPLL_EN_DISP | DPLL_SEL_HDMI | DPLL_EN_HDMI);
-			REG_READ(DPLL_CLK_ENABLE);
-		}
-		/* wait for dpll warm up */
-		udelay(150);
-
-		/* Enable pipe B */
-		temp = REG_READ(PIPEBCONF);
-		if ((temp & PIPEACONF_ENABLE) == 0) {
-			REG_WRITE(PIPEBCONF, temp | PIPEACONF_ENABLE);
-			REG_READ(PIPEBCONF);
-		}
-
-		/* Enable LNW Pipe B */
-		temp = REG_READ(PCH_PIPEBCONF);
-		if ((temp & PIPEACONF_ENABLE) == 0) {
-			REG_WRITE(PCH_PIPEBCONF, temp | PIPEACONF_ENABLE);
-			REG_READ(PCH_PIPEBCONF);
-		}
-		wait_for_vblank(dev);
-
-		/* Enable plane */
-		temp = REG_READ(DSPBCNTR);
-		if ((temp & DISPLAY_PLANE_ENABLE) == 0) {
-			REG_WRITE(DSPBCNTR, temp | DISPLAY_PLANE_ENABLE);
-			/* Flush the plane changes */
-			REG_WRITE(DSPBSURF, REG_READ(DSPBSURF));
-			REG_READ(DSPBSURF);
-		}
-		psb_intel_crtc_load_lut(crtc);
-	}
-	/* DSPARB */
-	REG_WRITE(DSPARB, 0x00003fbf);
-	/* FW1 */
-	REG_WRITE(0x70034, 0x3f880a0a);
-	/* FW2 */
-	REG_WRITE(0x70038, 0x0b060808);
-	/* FW4 */
-	REG_WRITE(0x70050, 0x08030404);
-	/* FW5 */
-	REG_WRITE(0x70054, 0x04040404);
-	/* LNC Chicken Bits */
-	REG_WRITE(0x70400, 0x4000);
-}
-
-
 static void oaktrail_hdmi_dpms(struct drm_encoder *encoder, int mode)
 {
 	static int dpms_mode = -1;
@@ -327,182 +176,6 @@ static void oaktrail_hdmi_dpms(struct drm_encoder *encoder, int mode)
 	HDMI_WRITE(HDMI_VIDEO_REG, temp);
 }
 
-static unsigned int htotal_calculate(struct drm_display_mode *mode)
-{
-	u32 htotal, new_crtc_htotal;
-
-	htotal = (mode->crtc_hdisplay - 1) | ((mode->crtc_htotal - 1) << 16);
-
-	/*
-	 * 1024 x 768  new_crtc_htotal = 0x1024;
-	 * 1280 x 1024 new_crtc_htotal = 0x0c34;
-	 */
-	new_crtc_htotal = (mode->crtc_htotal - 1) * 200 * 1000 / mode->clock;
-
-	return (mode->crtc_hdisplay - 1) | (new_crtc_htotal << 16);
-}
-
-static void oaktrail_hdmi_find_dpll(struct drm_crtc *crtc, int target,
-				int refclk, struct oaktrail_hdmi_clock *best_clock)
-{
-	int np_min, np_max, nr_min, nr_max;
-	int np, nr, nf;
-
-	np_min = DIV_ROUND_UP(oaktrail_hdmi_limit.vco.min, target * 10);
-	np_max = oaktrail_hdmi_limit.vco.max / (target * 10);
-	if (np_min < oaktrail_hdmi_limit.np.min)
-		np_min = oaktrail_hdmi_limit.np.min;
-	if (np_max > oaktrail_hdmi_limit.np.max)
-		np_max = oaktrail_hdmi_limit.np.max;
-
-	nr_min = DIV_ROUND_UP((refclk * 1000), (target * 10 * np_max));
-	nr_max = DIV_ROUND_UP((refclk * 1000), (target * 10 * np_min));
-	if (nr_min < oaktrail_hdmi_limit.nr.min)
-		nr_min = oaktrail_hdmi_limit.nr.min;
-	if (nr_max > oaktrail_hdmi_limit.nr.max)
-		nr_max = oaktrail_hdmi_limit.nr.max;
-
-	np = DIV_ROUND_UP((refclk * 1000), (target * 10 * nr_max));
-	nr = DIV_ROUND_UP((refclk * 1000), (target * 10 * np));
-	nf = DIV_ROUND_CLOSEST((target * 10 * np * nr), refclk);
-	DRM_DEBUG_KMS("np, nr, nf %d %d %d\n", np, nr, nf);
-
-	/*
-	 * 1024 x 768  np = 1; nr = 0x26; nf = 0x0fd8000;
-	 * 1280 x 1024 np = 1; nr = 0x17; nf = 0x1034000;
-	 */
-	best_clock->np = np;
-	best_clock->nr = nr - 1;
-	best_clock->nf = (nf << 14);
-}
-
-int oaktrail_crtc_hdmi_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 drm_psb_private *dev_priv = dev->dev_private;
-	struct oaktrail_hdmi_dev *hdmi_dev = dev_priv->hdmi_priv;
-	int pipe = 1;
-	int htot_reg = (pipe == 0) ? HTOTAL_A : HTOTAL_B;
-	int hblank_reg = (pipe == 0) ? HBLANK_A : HBLANK_B;
-	int hsync_reg = (pipe == 0) ? HSYNC_A : HSYNC_B;
-	int vtot_reg = (pipe == 0) ? VTOTAL_A : VTOTAL_B;
-	int vblank_reg = (pipe == 0) ? VBLANK_A : VBLANK_B;
-	int vsync_reg = (pipe == 0) ? VSYNC_A : VSYNC_B;
-	int dspsize_reg = (pipe == 0) ? DSPASIZE : DSPBSIZE;
-	int dsppos_reg = (pipe == 0) ? DSPAPOS : DSPBPOS;
-	int pipesrc_reg = (pipe == 0) ? PIPEASRC : PIPEBSRC;
-	int pipeconf_reg = (pipe == 0) ? PIPEACONF : PIPEBCONF;
-	int refclk;
-	struct oaktrail_hdmi_clock clock;
-	u32 dspcntr, pipeconf, dpll, temp;
-	int dspcntr_reg = DSPBCNTR;
-
-	/* Disable the VGA plane that we never use */
-	REG_WRITE(VGACNTRL, VGA_DISP_DISABLE);
-
-	/* XXX: Disable the panel fitter if it was on our pipe */
-
-	/* Disable dpll if necessary */
-	dpll = REG_READ(DPLL_CTRL);
-	if ((dpll & DPLL_PWRDN) == 0) {
-		REG_WRITE(DPLL_CTRL, dpll | (DPLL_PWRDN | DPLL_RESET));
-		REG_WRITE(DPLL_DIV_CTRL, 0x00000000);
-		REG_WRITE(DPLL_STATUS, 0x1);
-	}
-	udelay(150);
-
-	/* reset controller: FIXME - can we sort out the ioremap mess ? */
-	iounmap(hdmi_dev->regs);
-	oaktrail_hdmi_reset(dev);
-
-	/* program and enable dpll */
-	refclk = 25000;
-	oaktrail_hdmi_find_dpll(crtc, adjusted_mode->clock, refclk, &clock);
-
-	/* Setting DPLL */
-	dpll = REG_READ(DPLL_CTRL);
-	dpll &= ~DPLL_PDIV_MASK;
-	dpll &= ~(DPLL_PWRDN | DPLL_RESET);
-	REG_WRITE(DPLL_CTRL, 0x00000008);
-	REG_WRITE(DPLL_DIV_CTRL, ((clock.nf << 6) | clock.nr));
-	REG_WRITE(DPLL_ADJUST, ((clock.nf >> 14) - 1));
-	REG_WRITE(DPLL_CTRL, (dpll | (clock.np << DPLL_PDIV_SHIFT) | DPLL_ENSTAT | DPLL_DITHEN));
-	REG_WRITE(DPLL_UPDATE, 0x80000000);
-	REG_WRITE(DPLL_CLK_ENABLE, 0x80050102);
-	udelay(150);
-
-	hdmi_dev->regs = ioremap(hdmi_dev->mmio, hdmi_dev->mmio_len);
-	if (hdmi_dev->regs == NULL) {
-		DRM_ERROR("failed to do hdmi mmio mapping\n");
-		return -ENOMEM;
-	}
-
-	/* configure HDMI */
-	HDMI_WRITE(0x1004, 0x1fd);
-	HDMI_WRITE(0x2000, 0x1);
-	HDMI_WRITE(0x2008, 0x0);
-	HDMI_WRITE(0x3130, 0x8);
-	HDMI_WRITE(0x101c, 0x1800810);
-
-	temp = htotal_calculate(adjusted_mode);
-	REG_WRITE(htot_reg, temp);
-	REG_WRITE(hblank_reg, (adjusted_mode->crtc_hblank_start - 1) | ((adjusted_mode->crtc_hblank_end - 1) << 16));
-	REG_WRITE(hsync_reg, (adjusted_mode->crtc_hsync_start - 1) | ((adjusted_mode->crtc_hsync_end - 1) << 16));
-	REG_WRITE(vtot_reg, (adjusted_mode->crtc_vdisplay - 1) | ((adjusted_mode->crtc_vtotal - 1) << 16));
-	REG_WRITE(vblank_reg, (adjusted_mode->crtc_vblank_start - 1) | ((adjusted_mode->crtc_vblank_end - 1) << 16));
-	REG_WRITE(vsync_reg, (adjusted_mode->crtc_vsync_start - 1) | ((adjusted_mode->crtc_vsync_end - 1) << 16));
-	REG_WRITE(pipesrc_reg,
-		((mode->crtc_hdisplay - 1) << 16) |  (mode->crtc_vdisplay - 1));
-
-	REG_WRITE(PCH_HTOTAL_B, (adjusted_mode->crtc_hdisplay - 1) | ((adjusted_mode->crtc_htotal - 1) << 16));
-	REG_WRITE(PCH_HBLANK_B, (adjusted_mode->crtc_hblank_start - 1) | ((adjusted_mode->crtc_hblank_end - 1) << 16));
-	REG_WRITE(PCH_HSYNC_B, (adjusted_mode->crtc_hsync_start - 1) | ((adjusted_mode->crtc_hsync_end - 1) << 16));
-	REG_WRITE(PCH_VTOTAL_B, (adjusted_mode->crtc_vdisplay - 1) | ((adjusted_mode->crtc_vtotal - 1) << 16));
-	REG_WRITE(PCH_VBLANK_B, (adjusted_mode->crtc_vblank_start - 1) | ((adjusted_mode->crtc_vblank_end - 1) << 16));
-	REG_WRITE(PCH_VSYNC_B, (adjusted_mode->crtc_vsync_start - 1) | ((adjusted_mode->crtc_vsync_end - 1) << 16));
-	REG_WRITE(PCH_PIPEBSRC,
-		((mode->crtc_hdisplay - 1) << 16) |  (mode->crtc_vdisplay - 1));
-
-	temp = adjusted_mode->crtc_hblank_end - adjusted_mode->crtc_hblank_start;
-	HDMI_WRITE(HDMI_HBLANK_A, ((adjusted_mode->crtc_hdisplay - 1) << 16) |  temp);
-
-	REG_WRITE(dspsize_reg,
-			((mode->vdisplay - 1) << 16) | (mode->hdisplay - 1));
-	REG_WRITE(dsppos_reg, 0);
-
-	/* Flush the plane changes */
-	{
-		struct drm_crtc_helper_funcs *crtc_funcs = crtc->helper_private;
-		crtc_funcs->mode_set_base(crtc, x, y, old_fb);
-	}
-
-	/* Set up the display plane register */
-	dspcntr = REG_READ(dspcntr_reg);
-	dspcntr |= DISPPLANE_GAMMA_ENABLE;
-	dspcntr |= DISPPLANE_SEL_PIPE_B;
-	dspcntr |= DISPLAY_PLANE_ENABLE;
-
-	/* setup pipeconf */
-	pipeconf = REG_READ(pipeconf_reg);
-	pipeconf |= PIPEACONF_ENABLE;
-
-	REG_WRITE(pipeconf_reg, pipeconf);
-	REG_READ(pipeconf_reg);
-
-	REG_WRITE(PCH_PIPEBCONF, pipeconf);
-	REG_READ(PCH_PIPEBCONF);
-	wait_for_vblank(dev);
-
-	REG_WRITE(dspcntr_reg, dspcntr);
-	wait_for_vblank(dev);
-
-	return 0;
-}
-
 static int oaktrail_hdmi_mode_valid(struct drm_connector *connector,
 				struct drm_display_mode *mode)
 {
@@ -692,7 +365,7 @@ failed_connector:
 
 static DEFINE_PCI_DEVICE_TABLE(hdmi_ids) = {
 	{ PCI_DEVICE(PCI_VENDOR_ID_INTEL, 0x080d) },
-	{}
+	{ 0 }
 };
 
 void oaktrail_hdmi_setup(struct drm_device *dev)
@@ -766,6 +439,7 @@ void oaktrail_hdmi_save(struct drm_device *dev)
 {
 	struct drm_psb_private *dev_priv = dev->dev_private;
 	struct oaktrail_hdmi_dev *hdmi_dev = dev_priv->hdmi_priv;
+	struct psb_state *regs = &dev_priv->regs.psb;
 	int i;
 
 	/* dpll */
@@ -776,14 +450,14 @@ void oaktrail_hdmi_save(struct drm_device *dev)
 	hdmi_dev->saveDPLL_CLK_ENABLE = PSB_RVDC32(DPLL_CLK_ENABLE);
 
 	/* pipe B */
-	dev_priv->savePIPEBCONF = PSB_RVDC32(PIPEBCONF);
-	dev_priv->savePIPEBSRC  = PSB_RVDC32(PIPEBSRC);
-	dev_priv->saveHTOTAL_B  = PSB_RVDC32(HTOTAL_B);
-	dev_priv->saveHBLANK_B  = PSB_RVDC32(HBLANK_B);
-	dev_priv->saveHSYNC_B   = PSB_RVDC32(HSYNC_B);
-	dev_priv->saveVTOTAL_B  = PSB_RVDC32(VTOTAL_B);
-	dev_priv->saveVBLANK_B  = PSB_RVDC32(VBLANK_B);
-	dev_priv->saveVSYNC_B   = PSB_RVDC32(VSYNC_B);
+	regs->savePIPEBCONF = PSB_RVDC32(PIPEBCONF);
+	regs->savePIPEBSRC  = PSB_RVDC32(PIPEBSRC);
+	regs->saveHTOTAL_B  = PSB_RVDC32(HTOTAL_B);
+	regs->saveHBLANK_B  = PSB_RVDC32(HBLANK_B);
+	regs->saveHSYNC_B   = PSB_RVDC32(HSYNC_B);
+	regs->saveVTOTAL_B  = PSB_RVDC32(VTOTAL_B);
+	regs->saveVBLANK_B  = PSB_RVDC32(VBLANK_B);
+	regs->saveVSYNC_B   = PSB_RVDC32(VSYNC_B);
 
 	hdmi_dev->savePCH_PIPEBCONF = PSB_RVDC32(PCH_PIPEBCONF);
 	hdmi_dev->savePCH_PIPEBSRC = PSB_RVDC32(PCH_PIPEBSRC);
@@ -795,21 +469,21 @@ void oaktrail_hdmi_save(struct drm_device *dev)
 	hdmi_dev->savePCH_VSYNC_B  = PSB_RVDC32(PCH_VSYNC_B);
 
 	/* plane */
-	dev_priv->saveDSPBCNTR = PSB_RVDC32(DSPBCNTR);
-	dev_priv->saveDSPBSTRIDE = PSB_RVDC32(DSPBSTRIDE);
-	dev_priv->saveDSPBADDR = PSB_RVDC32(DSPBBASE);
-	dev_priv->saveDSPBSURF = PSB_RVDC32(DSPBSURF);
-	dev_priv->saveDSPBLINOFF = PSB_RVDC32(DSPBLINOFF);
-	dev_priv->saveDSPBTILEOFF = PSB_RVDC32(DSPBTILEOFF);
+	regs->saveDSPBCNTR = PSB_RVDC32(DSPBCNTR);
+	regs->saveDSPBSTRIDE = PSB_RVDC32(DSPBSTRIDE);
+	regs->saveDSPBADDR = PSB_RVDC32(DSPBBASE);
+	regs->saveDSPBSURF = PSB_RVDC32(DSPBSURF);
+	regs->saveDSPBLINOFF = PSB_RVDC32(DSPBLINOFF);
+	regs->saveDSPBTILEOFF = PSB_RVDC32(DSPBTILEOFF);
 
 	/* cursor B */
-	dev_priv->saveDSPBCURSOR_CTRL = PSB_RVDC32(CURBCNTR);
-	dev_priv->saveDSPBCURSOR_BASE = PSB_RVDC32(CURBBASE);
-	dev_priv->saveDSPBCURSOR_POS = PSB_RVDC32(CURBPOS);
+	regs->saveDSPBCURSOR_CTRL = PSB_RVDC32(CURBCNTR);
+	regs->saveDSPBCURSOR_BASE = PSB_RVDC32(CURBBASE);
+	regs->saveDSPBCURSOR_POS = PSB_RVDC32(CURBPOS);
 
 	/* save palette */
 	for (i = 0; i < 256; i++)
-		dev_priv->save_palette_b[i] = PSB_RVDC32(PALETTE_B + (i << 2));
+		regs->save_palette_b[i] = PSB_RVDC32(PALETTE_B + (i << 2));
 }
 
 /* restore HDMI register state */
@@ -817,6 +491,7 @@ void oaktrail_hdmi_restore(struct drm_device *dev)
 {
 	struct drm_psb_private *dev_priv = dev->dev_private;
 	struct oaktrail_hdmi_dev *hdmi_dev = dev_priv->hdmi_priv;
+	struct psb_state *regs = &dev_priv->regs.psb;
 	int i;
 
 	/* dpll */
@@ -828,13 +503,13 @@ void oaktrail_hdmi_restore(struct drm_device *dev)
 	DRM_UDELAY(150);
 
 	/* pipe */
-	PSB_WVDC32(dev_priv->savePIPEBSRC, PIPEBSRC);
-	PSB_WVDC32(dev_priv->saveHTOTAL_B, HTOTAL_B);
-	PSB_WVDC32(dev_priv->saveHBLANK_B, HBLANK_B);
-	PSB_WVDC32(dev_priv->saveHSYNC_B,  HSYNC_B);
-	PSB_WVDC32(dev_priv->saveVTOTAL_B, VTOTAL_B);
-	PSB_WVDC32(dev_priv->saveVBLANK_B, VBLANK_B);
-	PSB_WVDC32(dev_priv->saveVSYNC_B,  VSYNC_B);
+	PSB_WVDC32(regs->savePIPEBSRC, PIPEBSRC);
+	PSB_WVDC32(regs->saveHTOTAL_B, HTOTAL_B);
+	PSB_WVDC32(regs->saveHBLANK_B, HBLANK_B);
+	PSB_WVDC32(regs->saveHSYNC_B,  HSYNC_B);
+	PSB_WVDC32(regs->saveVTOTAL_B, VTOTAL_B);
+	PSB_WVDC32(regs->saveVBLANK_B, VBLANK_B);
+	PSB_WVDC32(regs->saveVSYNC_B,  VSYNC_B);
 
 	PSB_WVDC32(hdmi_dev->savePCH_PIPEBSRC, PCH_PIPEBSRC);
 	PSB_WVDC32(hdmi_dev->savePCH_HTOTAL_B, PCH_HTOTAL_B);
@@ -844,22 +519,22 @@ void oaktrail_hdmi_restore(struct drm_device *dev)
 	PSB_WVDC32(hdmi_dev->savePCH_VBLANK_B, PCH_VBLANK_B);
 	PSB_WVDC32(hdmi_dev->savePCH_VSYNC_B,  PCH_VSYNC_B);
 
-	PSB_WVDC32(dev_priv->savePIPEBCONF, PIPEBCONF);
+	PSB_WVDC32(regs->savePIPEBCONF, PIPEBCONF);
 	PSB_WVDC32(hdmi_dev->savePCH_PIPEBCONF, PCH_PIPEBCONF);
 
 	/* plane */
-	PSB_WVDC32(dev_priv->saveDSPBLINOFF, DSPBLINOFF);
-	PSB_WVDC32(dev_priv->saveDSPBSTRIDE, DSPBSTRIDE);
-	PSB_WVDC32(dev_priv->saveDSPBTILEOFF, DSPBTILEOFF);
-	PSB_WVDC32(dev_priv->saveDSPBCNTR, DSPBCNTR);
-	PSB_WVDC32(dev_priv->saveDSPBSURF, DSPBSURF);
+	PSB_WVDC32(regs->saveDSPBLINOFF, DSPBLINOFF);
+	PSB_WVDC32(regs->saveDSPBSTRIDE, DSPBSTRIDE);
+	PSB_WVDC32(regs->saveDSPBTILEOFF, DSPBTILEOFF);
+	PSB_WVDC32(regs->saveDSPBCNTR, DSPBCNTR);
+	PSB_WVDC32(regs->saveDSPBSURF, DSPBSURF);
 
 	/* cursor B */
-	PSB_WVDC32(dev_priv->saveDSPBCURSOR_CTRL, CURBCNTR);
-	PSB_WVDC32(dev_priv->saveDSPBCURSOR_POS, CURBPOS);
-	PSB_WVDC32(dev_priv->saveDSPBCURSOR_BASE, CURBBASE);
+	PSB_WVDC32(regs->saveDSPBCURSOR_CTRL, CURBCNTR);
+	PSB_WVDC32(regs->saveDSPBCURSOR_POS, CURBPOS);
+	PSB_WVDC32(regs->saveDSPBCURSOR_BASE, CURBBASE);
 
 	/* restore palette */
 	for (i = 0; i < 256; i++)
-		PSB_WVDC32(dev_priv->save_palette_b[i], PALETTE_B + (i << 2));
+		PSB_WVDC32(regs->save_palette_b[i], PALETTE_B + (i << 2));
 }
diff --git a/drivers/gpu/drm/gma500/oaktrail_hdmi_i2c.c b/drivers/gpu/drm/gma500/oaktrail_hdmi_i2c.c
index 705440874ac0..5e84fbde749b 100644
--- a/drivers/gpu/drm/gma500/oaktrail_hdmi_i2c.c
+++ b/drivers/gpu/drm/gma500/oaktrail_hdmi_i2c.c
@@ -127,7 +127,7 @@ static int oaktrail_hdmi_i2c_access(struct i2c_adapter *adap,
 {
 	struct oaktrail_hdmi_dev *hdmi_dev = i2c_get_adapdata(adap);
 	struct hdmi_i2c_dev *i2c_dev = hdmi_dev->i2c_dev;
-	int i, err = 0;
+	int i;
 
 	mutex_lock(&i2c_dev->i2c_lock);
 
@@ -139,9 +139,9 @@ static int oaktrail_hdmi_i2c_access(struct i2c_adapter *adap,
 	for (i = 0; i < num; i++) {
 		if (pmsg->len && pmsg->buf) {
 			if (pmsg->flags & I2C_M_RD)
-				err = xfer_read(adap, pmsg);
+				xfer_read(adap, pmsg);
 			else
-				err = xfer_write(adap, pmsg);
+				xfer_write(adap, pmsg);
 		}
 		pmsg++;         /* next message */
 	}
diff --git a/drivers/gpu/drm/gma500/oaktrail_lvds.c b/drivers/gpu/drm/gma500/oaktrail_lvds.c
index 238bbe105304..654f32b22b21 100644
--- a/drivers/gpu/drm/gma500/oaktrail_lvds.c
+++ b/drivers/gpu/drm/gma500/oaktrail_lvds.c
@@ -192,7 +192,7 @@ static u32 oaktrail_lvds_get_max_backlight(struct drm_device *dev)
 
 		gma_power_end(dev);
 	} else
-		ret = ((dev_priv->saveBLC_PWM_CTL &
+		ret = ((dev_priv->regs.saveBLC_PWM_CTL &
 			  BACKLIGHT_MODULATION_FREQ_MASK) >>
 			  BACKLIGHT_MODULATION_FREQ_SHIFT) * 2;
 
@@ -331,7 +331,6 @@ void oaktrail_lvds_init(struct drm_device *dev,
 	struct drm_encoder *encoder;
 	struct drm_psb_private *dev_priv = dev->dev_private;
 	struct edid *edid;
-	int ret = 0;
 	struct i2c_adapter *i2c_adap;
 	struct drm_display_mode *scan;	/* *modes, *bios_mode; */
 
@@ -400,7 +399,7 @@ void oaktrail_lvds_init(struct drm_device *dev,
 		if (edid) {
 			drm_mode_connector_update_edid_property(connector,
 									edid);
-			ret = drm_add_edid_modes(connector, edid);
+			drm_add_edid_modes(connector, edid);
 			kfree(edid);
 		}
 
diff --git a/drivers/gpu/drm/gma500/power.c b/drivers/gpu/drm/gma500/power.c
index 94025693bae1..889b854751da 100644
--- a/drivers/gpu/drm/gma500/power.c
+++ b/drivers/gpu/drm/gma500/power.c
@@ -58,7 +58,8 @@ void gma_power_init(struct drm_device *dev)
 	spin_lock_init(&power_ctrl_lock);
 	mutex_init(&power_mutex);
 
-	dev_priv->ops->init_pm(dev);
+	if (dev_priv->ops->init_pm)
+		dev_priv->ops->init_pm(dev);
 }
 
 /**
@@ -101,9 +102,6 @@ static void gma_resume_display(struct pci_dev *pdev)
 	struct drm_device *dev = pci_get_drvdata(pdev);
 	struct drm_psb_private *dev_priv = dev->dev_private;
 
-	if (dev_priv->suspended == false)
-		return;
-
 	/* turn on the display power island */
 	dev_priv->ops->power_up(dev);
 	dev_priv->suspended = false;
@@ -132,9 +130,9 @@ static void gma_suspend_pci(struct pci_dev *pdev)
 
 	pci_save_state(pdev);
 	pci_read_config_dword(pdev, 0x5C, &bsm);
-	dev_priv->saveBSM = bsm;
+	dev_priv->regs.saveBSM = bsm;
 	pci_read_config_dword(pdev, 0xFC, &vbt);
-	dev_priv->saveVBT = vbt;
+	dev_priv->regs.saveVBT = vbt;
 	pci_read_config_dword(pdev, PSB_PCIx_MSI_ADDR_LOC, &dev_priv->msi_addr);
 	pci_read_config_dword(pdev, PSB_PCIx_MSI_DATA_LOC, &dev_priv->msi_data);
 
@@ -162,8 +160,8 @@ static bool gma_resume_pci(struct pci_dev *pdev)
 
 	pci_set_power_state(pdev, PCI_D0);
 	pci_restore_state(pdev);
-	pci_write_config_dword(pdev, 0x5c, dev_priv->saveBSM);
-	pci_write_config_dword(pdev, 0xFC, dev_priv->saveVBT);
+	pci_write_config_dword(pdev, 0x5c, dev_priv->regs.saveBSM);
+	pci_write_config_dword(pdev, 0xFC, dev_priv->regs.saveVBT);
 	/* restoring MSI address and data in PCIx space */
 	pci_write_config_dword(pdev, PSB_PCIx_MSI_ADDR_LOC, dev_priv->msi_addr);
 	pci_write_config_dword(pdev, PSB_PCIx_MSI_DATA_LOC, dev_priv->msi_data);
@@ -195,6 +193,7 @@ int gma_power_suspend(struct device *_dev)
 	if (!dev_priv->suspended) {
 		if (dev_priv->display_count) {
 			mutex_unlock(&power_mutex);
+			dev_err(dev->dev, "GPU hardware busy, cannot suspend\n");
 			return -EBUSY;
 		}
 		psb_irq_uninstall(dev);
@@ -302,7 +301,7 @@ int psb_runtime_suspend(struct device *dev)
 
 int psb_runtime_resume(struct device *dev)
 {
-	return gma_power_resume(dev);;
+	return gma_power_resume(dev);
 }
 
 int psb_runtime_idle(struct device *dev)
diff --git a/drivers/gpu/drm/gma500/psb_device.c b/drivers/gpu/drm/gma500/psb_device.c
index e5f5906172b0..95d163e4f1f4 100644
--- a/drivers/gpu/drm/gma500/psb_device.c
+++ b/drivers/gpu/drm/gma500/psb_device.c
@@ -177,16 +177,17 @@ static int psb_save_display_registers(struct drm_device *dev)
 	struct drm_psb_private *dev_priv = dev->dev_private;
 	struct drm_crtc *crtc;
 	struct drm_connector *connector;
+	struct psb_state *regs = &dev_priv->regs.psb;
 
 	/* Display arbitration control + watermarks */
-	dev_priv->saveDSPARB = PSB_RVDC32(DSPARB);
-	dev_priv->saveDSPFW1 = PSB_RVDC32(DSPFW1);
-	dev_priv->saveDSPFW2 = PSB_RVDC32(DSPFW2);
-	dev_priv->saveDSPFW3 = PSB_RVDC32(DSPFW3);
-	dev_priv->saveDSPFW4 = PSB_RVDC32(DSPFW4);
-	dev_priv->saveDSPFW5 = PSB_RVDC32(DSPFW5);
-	dev_priv->saveDSPFW6 = PSB_RVDC32(DSPFW6);
-	dev_priv->saveCHICKENBIT = PSB_RVDC32(DSPCHICKENBIT);
+	regs->saveDSPARB = PSB_RVDC32(DSPARB);
+	regs->saveDSPFW1 = PSB_RVDC32(DSPFW1);
+	regs->saveDSPFW2 = PSB_RVDC32(DSPFW2);
+	regs->saveDSPFW3 = PSB_RVDC32(DSPFW3);
+	regs->saveDSPFW4 = PSB_RVDC32(DSPFW4);
+	regs->saveDSPFW5 = PSB_RVDC32(DSPFW5);
+	regs->saveDSPFW6 = PSB_RVDC32(DSPFW6);
+	regs->saveCHICKENBIT = PSB_RVDC32(DSPCHICKENBIT);
 
 	/* Save crtc and output state */
 	mutex_lock(&dev->mode_config.mutex);
@@ -213,16 +214,17 @@ static int psb_restore_display_registers(struct drm_device *dev)
 	struct drm_psb_private *dev_priv = dev->dev_private;
 	struct drm_crtc *crtc;
 	struct drm_connector *connector;
+	struct psb_state *regs = &dev_priv->regs.psb;
 
 	/* Display arbitration + watermarks */
-	PSB_WVDC32(dev_priv->saveDSPARB, DSPARB);
-	PSB_WVDC32(dev_priv->saveDSPFW1, DSPFW1);
-	PSB_WVDC32(dev_priv->saveDSPFW2, DSPFW2);
-	PSB_WVDC32(dev_priv->saveDSPFW3, DSPFW3);
-	PSB_WVDC32(dev_priv->saveDSPFW4, DSPFW4);
-	PSB_WVDC32(dev_priv->saveDSPFW5, DSPFW5);
-	PSB_WVDC32(dev_priv->saveDSPFW6, DSPFW6);
-	PSB_WVDC32(dev_priv->saveCHICKENBIT, DSPCHICKENBIT);
+	PSB_WVDC32(regs->saveDSPARB, DSPARB);
+	PSB_WVDC32(regs->saveDSPFW1, DSPFW1);
+	PSB_WVDC32(regs->saveDSPFW2, DSPFW2);
+	PSB_WVDC32(regs->saveDSPFW3, DSPFW3);
+	PSB_WVDC32(regs->saveDSPFW4, DSPFW4);
+	PSB_WVDC32(regs->saveDSPFW5, DSPFW5);
+	PSB_WVDC32(regs->saveDSPFW6, DSPFW6);
+	PSB_WVDC32(regs->saveCHICKENBIT, DSPCHICKENBIT);
 
 	/*make sure VGA plane is off. it initializes to on after reset!*/
 	PSB_WVDC32(0x80000000, VGACNTRL);
diff --git a/drivers/gpu/drm/gma500/psb_drv.c b/drivers/gpu/drm/gma500/psb_drv.c
index f14768f2b364..c34adf9d910a 100644
--- a/drivers/gpu/drm/gma500/psb_drv.c
+++ b/drivers/gpu/drm/gma500/psb_drv.c
@@ -60,6 +60,16 @@ static DEFINE_PCI_DEVICE_TABLE(pciidlist) = {
 	/* Atom E620 */
 	{ 0x8086, 0x4108, PCI_ANY_ID, PCI_ANY_ID, 0, 0, (long) &oaktrail_chip_ops},
 #endif
+#if defined(CONFIG_DRM_MEDFIELD)
+	{0x8086, 0x0130, PCI_ANY_ID, PCI_ANY_ID, 0, 0, (long) &mdfld_chip_ops},
+	{0x8086, 0x0131, PCI_ANY_ID, PCI_ANY_ID, 0, 0, (long) &mdfld_chip_ops},
+	{0x8086, 0x0132, PCI_ANY_ID, PCI_ANY_ID, 0, 0, (long) &mdfld_chip_ops},
+	{0x8086, 0x0133, PCI_ANY_ID, PCI_ANY_ID, 0, 0, (long) &mdfld_chip_ops},
+	{0x8086, 0x0134, PCI_ANY_ID, PCI_ANY_ID, 0, 0, (long) &mdfld_chip_ops},
+	{0x8086, 0x0135, PCI_ANY_ID, PCI_ANY_ID, 0, 0, (long) &mdfld_chip_ops},
+	{0x8086, 0x0136, PCI_ANY_ID, PCI_ANY_ID, 0, 0, (long) &mdfld_chip_ops},
+	{0x8086, 0x0137, PCI_ANY_ID, PCI_ANY_ID, 0, 0, (long) &mdfld_chip_ops},
+#endif
 #if defined(CONFIG_DRM_GMA3600)
 	{ 0x8086, 0x0be0, PCI_ANY_ID, PCI_ANY_ID, 0, 0, (long) &cdv_chip_ops},
 	{ 0x8086, 0x0be1, PCI_ANY_ID, PCI_ANY_ID, 0, 0, (long) &cdv_chip_ops},
@@ -70,7 +80,7 @@ static DEFINE_PCI_DEVICE_TABLE(pciidlist) = {
 	{ 0x8086, 0x0be6, PCI_ANY_ID, PCI_ANY_ID, 0, 0, (long) &cdv_chip_ops},
 	{ 0x8086, 0x0be7, PCI_ANY_ID, PCI_ANY_ID, 0, 0, (long) &cdv_chip_ops},
 #endif
-	{ 0, 0, 0}
+	{ 0, }
 };
 MODULE_DEVICE_TABLE(pci, pciidlist);
 
@@ -78,27 +88,27 @@ MODULE_DEVICE_TABLE(pci, pciidlist);
  * Standard IOCTLs.
  */
 
-#define DRM_IOCTL_PSB_ADB	\
+#define DRM_IOCTL_GMA_ADB	\
 		DRM_IOWR(DRM_GMA_ADB + DRM_COMMAND_BASE, uint32_t)
-#define DRM_IOCTL_PSB_MODE_OPERATION	\
+#define DRM_IOCTL_GMA_MODE_OPERATION	\
 		DRM_IOWR(DRM_GMA_MODE_OPERATION + DRM_COMMAND_BASE, \
 			 struct drm_psb_mode_operation_arg)
-#define DRM_IOCTL_PSB_STOLEN_MEMORY	\
+#define DRM_IOCTL_GMA_STOLEN_MEMORY	\
 		DRM_IOWR(DRM_GMA_STOLEN_MEMORY + DRM_COMMAND_BASE, \
 			 struct drm_psb_stolen_memory_arg)
-#define DRM_IOCTL_PSB_GAMMA	\
+#define DRM_IOCTL_GMA_GAMMA	\
 		DRM_IOWR(DRM_GMA_GAMMA + DRM_COMMAND_BASE, \
 			 struct drm_psb_dpst_lut_arg)
-#define DRM_IOCTL_PSB_DPST_BL	\
+#define DRM_IOCTL_GMA_DPST_BL	\
 		DRM_IOWR(DRM_GMA_DPST_BL + DRM_COMMAND_BASE, \
 			 uint32_t)
-#define DRM_IOCTL_PSB_GET_PIPE_FROM_CRTC_ID	\
+#define DRM_IOCTL_GMA_GET_PIPE_FROM_CRTC_ID	\
 		DRM_IOWR(DRM_GMA_GET_PIPE_FROM_CRTC_ID + DRM_COMMAND_BASE, \
 			 struct drm_psb_get_pipe_from_crtc_id_arg)
-#define DRM_IOCTL_PSB_GEM_CREATE	\
+#define DRM_IOCTL_GMA_GEM_CREATE	\
 		DRM_IOWR(DRM_GMA_GEM_CREATE + DRM_COMMAND_BASE, \
 			 struct drm_psb_gem_create)
-#define DRM_IOCTL_PSB_GEM_MMAP	\
+#define DRM_IOCTL_GMA_GEM_MMAP	\
 		DRM_IOWR(DRM_GMA_GEM_MMAP + DRM_COMMAND_BASE, \
 			 struct drm_psb_gem_mmap)
 
@@ -113,22 +123,19 @@ static int psb_gamma_ioctl(struct drm_device *dev, void *data,
 static int psb_dpst_bl_ioctl(struct drm_device *dev, void *data,
 			     struct drm_file *file_priv);
 
-#define PSB_IOCTL_DEF(ioctl, func, flags) \
-	[DRM_IOCTL_NR(ioctl) - DRM_COMMAND_BASE] = {ioctl, flags, func}
-
 static struct drm_ioctl_desc psb_ioctls[] = {
-	PSB_IOCTL_DEF(DRM_IOCTL_PSB_ADB, psb_adb_ioctl, DRM_AUTH),
-	PSB_IOCTL_DEF(DRM_IOCTL_PSB_MODE_OPERATION, psb_mode_operation_ioctl,
+	DRM_IOCTL_DEF_DRV(GMA_ADB, psb_adb_ioctl, DRM_AUTH),
+	DRM_IOCTL_DEF_DRV(GMA_MODE_OPERATION, psb_mode_operation_ioctl,
 		      DRM_AUTH),
-	PSB_IOCTL_DEF(DRM_IOCTL_PSB_STOLEN_MEMORY, psb_stolen_memory_ioctl,
+	DRM_IOCTL_DEF_DRV(GMA_STOLEN_MEMORY, psb_stolen_memory_ioctl,
 		      DRM_AUTH),
-	PSB_IOCTL_DEF(DRM_IOCTL_PSB_GAMMA, psb_gamma_ioctl, DRM_AUTH),
-	PSB_IOCTL_DEF(DRM_IOCTL_PSB_DPST_BL, psb_dpst_bl_ioctl, DRM_AUTH),
-	PSB_IOCTL_DEF(DRM_IOCTL_PSB_GET_PIPE_FROM_CRTC_ID,
+	DRM_IOCTL_DEF_DRV(GMA_GAMMA, psb_gamma_ioctl, DRM_AUTH),
+	DRM_IOCTL_DEF_DRV(GMA_DPST_BL, psb_dpst_bl_ioctl, DRM_AUTH),
+	DRM_IOCTL_DEF_DRV(GMA_GET_PIPE_FROM_CRTC_ID,
 					psb_intel_get_pipe_from_crtc_id, 0),
-	PSB_IOCTL_DEF(DRM_IOCTL_PSB_GEM_CREATE, psb_gem_create_ioctl,
+	DRM_IOCTL_DEF_DRV(GMA_GEM_CREATE, psb_gem_create_ioctl,
 						DRM_UNLOCKED | DRM_AUTH),
-	PSB_IOCTL_DEF(DRM_IOCTL_PSB_GEM_MMAP, psb_gem_mmap_ioctl,
+	DRM_IOCTL_DEF_DRV(GMA_GEM_MMAP, psb_gem_mmap_ioctl,
 						DRM_UNLOCKED | DRM_AUTH),
 };
 
@@ -268,10 +275,8 @@ static int psb_driver_load(struct drm_device *dev, unsigned long chipset)
 {
 	struct drm_psb_private *dev_priv;
 	unsigned long resource_start;
-	struct psb_gtt *pg;
 	unsigned long irqflags;
 	int ret = -ENOMEM;
-	uint32_t tt_pages;
 	struct drm_connector *connector;
 	struct psb_intel_encoder *psb_intel_encoder;
 
@@ -283,6 +288,8 @@ static int psb_driver_load(struct drm_device *dev, unsigned long chipset)
 	dev_priv->dev = dev;
 	dev->dev_private = (void *) dev_priv;
 
+	pci_set_master(dev->pdev);
+
 	if (!IS_PSB(dev)) {
 		if (pci_enable_msi(dev->pdev))
 			dev_warn(dev->dev, "Enabling MSI failed!\n");
@@ -327,12 +334,6 @@ static int psb_driver_load(struct drm_device *dev, unsigned long chipset)
 	if (!dev_priv->mmu)
 		goto out_err;
 
-	pg = &dev_priv->gtt;
-
-	tt_pages = (pg->gatt_pages < PSB_TT_PRIV0_PLIMIT) ?
-		(pg->gatt_pages) : PSB_TT_PRIV0_PLIMIT;
-
-
 	dev_priv->pf_pd = psb_mmu_alloc_pd(dev_priv->mmu, 1, 0);
 	if (!dev_priv->pf_pd)
 		goto out_err;
@@ -409,7 +410,7 @@ out_err:
 	return ret;
 }
 
-int psb_driver_device_is_agp(struct drm_device *dev)
+static int psb_driver_device_is_agp(struct drm_device *dev)
 {
 	return 0;
 }
@@ -600,7 +601,7 @@ static long psb_unlocked_ioctl(struct file *filp, unsigned int cmd,
 /* When a client dies:
  *    - Check for and clean up flipped page state
  */
-void psb_driver_preclose(struct drm_device *dev, struct drm_file *priv)
+static void psb_driver_preclose(struct drm_device *dev, struct drm_file *priv)
 {
 }
 
@@ -677,7 +678,9 @@ static struct pci_driver psb_pci_driver = {
 	.id_table = pciidlist,
 	.probe = psb_probe,
 	.remove = psb_remove,
-	.driver.pm = &psb_pm_ops,
+	.driver = {
+		.pm = &psb_pm_ops,
+	}
 };
 
 static int psb_probe(struct pci_dev *pdev, const struct pci_device_id *ent)
diff --git a/drivers/gpu/drm/gma500/psb_drv.h b/drivers/gpu/drm/gma500/psb_drv.h
index eb1568a0da95..40ce2c9bc2e4 100644
--- a/drivers/gpu/drm/gma500/psb_drv.h
+++ b/drivers/gpu/drm/gma500/psb_drv.h
@@ -276,6 +276,217 @@ struct intel_gmbus {
 	u32 reg0;
 };
 
+/*
+ *	Register save state. This is used to hold the context when the
+ *	device is powered off. In the case of Oaktrail this can (but does not
+ *	yet) include screen blank. Operations occuring during the save
+ *	update the register cache instead.
+ */
+struct psb_state {
+	uint32_t saveDSPACNTR;
+	uint32_t saveDSPBCNTR;
+	uint32_t savePIPEACONF;
+	uint32_t savePIPEBCONF;
+	uint32_t savePIPEASRC;
+	uint32_t savePIPEBSRC;
+	uint32_t saveFPA0;
+	uint32_t saveFPA1;
+	uint32_t saveDPLL_A;
+	uint32_t saveDPLL_A_MD;
+	uint32_t saveHTOTAL_A;
+	uint32_t saveHBLANK_A;
+	uint32_t saveHSYNC_A;
+	uint32_t saveVTOTAL_A;
+	uint32_t saveVBLANK_A;
+	uint32_t saveVSYNC_A;
+	uint32_t saveDSPASTRIDE;
+	uint32_t saveDSPASIZE;
+	uint32_t saveDSPAPOS;
+	uint32_t saveDSPABASE;
+	uint32_t saveDSPASURF;
+	uint32_t saveDSPASTATUS;
+	uint32_t saveFPB0;
+	uint32_t saveFPB1;
+	uint32_t saveDPLL_B;
+	uint32_t saveDPLL_B_MD;
+	uint32_t saveHTOTAL_B;
+	uint32_t saveHBLANK_B;
+	uint32_t saveHSYNC_B;
+	uint32_t saveVTOTAL_B;
+	uint32_t saveVBLANK_B;
+	uint32_t saveVSYNC_B;
+	uint32_t saveDSPBSTRIDE;
+	uint32_t saveDSPBSIZE;
+	uint32_t saveDSPBPOS;
+	uint32_t saveDSPBBASE;
+	uint32_t saveDSPBSURF;
+	uint32_t saveDSPBSTATUS;
+	uint32_t saveVCLK_DIVISOR_VGA0;
+	uint32_t saveVCLK_DIVISOR_VGA1;
+	uint32_t saveVCLK_POST_DIV;
+	uint32_t saveVGACNTRL;
+	uint32_t saveADPA;
+	uint32_t saveLVDS;
+	uint32_t saveDVOA;
+	uint32_t saveDVOB;
+	uint32_t saveDVOC;
+	uint32_t savePP_ON;
+	uint32_t savePP_OFF;
+	uint32_t savePP_CONTROL;
+	uint32_t savePP_CYCLE;
+	uint32_t savePFIT_CONTROL;
+	uint32_t savePaletteA[256];
+	uint32_t savePaletteB[256];
+	uint32_t saveCLOCKGATING;
+	uint32_t saveDSPARB;
+	uint32_t saveDSPATILEOFF;
+	uint32_t saveDSPBTILEOFF;
+	uint32_t saveDSPAADDR;
+	uint32_t saveDSPBADDR;
+	uint32_t savePFIT_AUTO_RATIOS;
+	uint32_t savePFIT_PGM_RATIOS;
+	uint32_t savePP_ON_DELAYS;
+	uint32_t savePP_OFF_DELAYS;
+	uint32_t savePP_DIVISOR;
+	uint32_t saveBCLRPAT_A;
+	uint32_t saveBCLRPAT_B;
+	uint32_t saveDSPALINOFF;
+	uint32_t saveDSPBLINOFF;
+	uint32_t savePERF_MODE;
+	uint32_t saveDSPFW1;
+	uint32_t saveDSPFW2;
+	uint32_t saveDSPFW3;
+	uint32_t saveDSPFW4;
+	uint32_t saveDSPFW5;
+	uint32_t saveDSPFW6;
+	uint32_t saveCHICKENBIT;
+	uint32_t saveDSPACURSOR_CTRL;
+	uint32_t saveDSPBCURSOR_CTRL;
+	uint32_t saveDSPACURSOR_BASE;
+	uint32_t saveDSPBCURSOR_BASE;
+	uint32_t saveDSPACURSOR_POS;
+	uint32_t saveDSPBCURSOR_POS;
+	uint32_t save_palette_a[256];
+	uint32_t save_palette_b[256];
+	uint32_t saveOV_OVADD;
+	uint32_t saveOV_OGAMC0;
+	uint32_t saveOV_OGAMC1;
+	uint32_t saveOV_OGAMC2;
+	uint32_t saveOV_OGAMC3;
+	uint32_t saveOV_OGAMC4;
+	uint32_t saveOV_OGAMC5;
+	uint32_t saveOVC_OVADD;
+	uint32_t saveOVC_OGAMC0;
+	uint32_t saveOVC_OGAMC1;
+	uint32_t saveOVC_OGAMC2;
+	uint32_t saveOVC_OGAMC3;
+	uint32_t saveOVC_OGAMC4;
+	uint32_t saveOVC_OGAMC5;
+
+	/* DPST register save */
+	uint32_t saveHISTOGRAM_INT_CONTROL_REG;
+	uint32_t saveHISTOGRAM_LOGIC_CONTROL_REG;
+	uint32_t savePWM_CONTROL_LOGIC;
+};
+
+struct medfield_state {
+	uint32_t saveDPLL_A;
+	uint32_t saveFPA0;
+	uint32_t savePIPEACONF;
+	uint32_t saveHTOTAL_A;
+	uint32_t saveHBLANK_A;
+	uint32_t saveHSYNC_A;
+	uint32_t saveVTOTAL_A;
+	uint32_t saveVBLANK_A;
+	uint32_t saveVSYNC_A;
+	uint32_t savePIPEASRC;
+	uint32_t saveDSPASTRIDE;
+	uint32_t saveDSPALINOFF;
+	uint32_t saveDSPATILEOFF;
+	uint32_t saveDSPASIZE;
+	uint32_t saveDSPAPOS;
+	uint32_t saveDSPASURF;
+	uint32_t saveDSPACNTR;
+	uint32_t saveDSPASTATUS;
+	uint32_t save_palette_a[256];
+	uint32_t saveMIPI;
+
+	uint32_t saveDPLL_B;
+	uint32_t saveFPB0;
+	uint32_t savePIPEBCONF;
+	uint32_t saveHTOTAL_B;
+	uint32_t saveHBLANK_B;
+	uint32_t saveHSYNC_B;
+	uint32_t saveVTOTAL_B;
+	uint32_t saveVBLANK_B;
+	uint32_t saveVSYNC_B;
+	uint32_t savePIPEBSRC;
+	uint32_t saveDSPBSTRIDE;
+	uint32_t saveDSPBLINOFF;
+	uint32_t saveDSPBTILEOFF;
+	uint32_t saveDSPBSIZE;
+	uint32_t saveDSPBPOS;
+	uint32_t saveDSPBSURF;
+	uint32_t saveDSPBCNTR;
+	uint32_t saveDSPBSTATUS;
+	uint32_t save_palette_b[256];
+
+	uint32_t savePIPECCONF;
+	uint32_t saveHTOTAL_C;
+	uint32_t saveHBLANK_C;
+	uint32_t saveHSYNC_C;
+	uint32_t saveVTOTAL_C;
+	uint32_t saveVBLANK_C;
+	uint32_t saveVSYNC_C;
+	uint32_t savePIPECSRC;
+	uint32_t saveDSPCSTRIDE;
+	uint32_t saveDSPCLINOFF;
+	uint32_t saveDSPCTILEOFF;
+	uint32_t saveDSPCSIZE;
+	uint32_t saveDSPCPOS;
+	uint32_t saveDSPCSURF;
+	uint32_t saveDSPCCNTR;
+	uint32_t saveDSPCSTATUS;
+	uint32_t save_palette_c[256];
+	uint32_t saveMIPI_C;
+
+	uint32_t savePFIT_CONTROL;
+	uint32_t savePFIT_PGM_RATIOS;
+	uint32_t saveHDMIPHYMISCCTL;
+	uint32_t saveHDMIB_CONTROL;
+};
+
+struct cdv_state {
+	uint32_t saveDSPCLK_GATE_D;
+	uint32_t saveRAMCLK_GATE_D;
+	uint32_t saveDSPARB;
+	uint32_t saveDSPFW[6];
+	uint32_t saveADPA;
+	uint32_t savePP_CONTROL;
+	uint32_t savePFIT_PGM_RATIOS;
+	uint32_t saveLVDS;
+	uint32_t savePFIT_CONTROL;
+	uint32_t savePP_ON_DELAYS;
+	uint32_t savePP_OFF_DELAYS;
+	uint32_t savePP_CYCLE;
+	uint32_t saveVGACNTRL;
+	uint32_t saveIER;
+	uint32_t saveIMR;
+	u8	 saveLBB;
+};
+
+struct psb_save_area {
+	uint32_t saveBSM;
+	uint32_t saveVBT;
+	union {
+	        struct psb_state psb;
+		struct medfield_state mdfld;
+		struct cdv_state cdv;
+	};
+	uint32_t saveBLC_PWM_CTL2;
+	uint32_t saveBLC_PWM_CTL;
+};
+
 struct psb_ops;
 
 #define PSB_NUM_PIPE		3
@@ -397,216 +608,21 @@ struct drm_psb_private {
 	struct oaktrail_vbt vbt_data;
 	struct oaktrail_gct_data gct_data;
 
-	/* MIPI Panel type etc */
-	int panel_id;
-	bool dual_mipi;		/* dual display - DPI & DBI */
-	bool dpi_panel_on;	/* The DPI panel power is on */
-	bool dpi_panel_on2;	/* The DPI panel power is on */
-	bool dbi_panel_on;	/* The DBI panel power is on */
-	bool dbi_panel_on2;	/* The DBI panel power is on */
-	u32 dsr_fb_update;	/* DSR FB update counter */
-
-	/* Moorestown HDMI state */
+	/* Oaktrail HDMI state */
 	struct oaktrail_hdmi_dev *hdmi_priv;
-
-	/* Moorestown pipe config register value cache */
-	uint32_t pipeconf;
-	uint32_t pipeconf1;
-	uint32_t pipeconf2;
-
-	/* Moorestown plane control register value cache */
-	uint32_t dspcntr;
-	uint32_t dspcntr1;
-	uint32_t dspcntr2;
-
-	/* Moorestown MM backlight cache */
-	uint8_t saveBKLTCNT;
-	uint8_t saveBKLTREQ;
-	uint8_t saveBKLTBRTL;
-
+	
 	/*
 	 * Register state
 	 */
-	uint32_t saveDSPACNTR;
-	uint32_t saveDSPBCNTR;
-	uint32_t savePIPEACONF;
-	uint32_t savePIPEBCONF;
-	uint32_t savePIPEASRC;
-	uint32_t savePIPEBSRC;
-	uint32_t saveFPA0;
-	uint32_t saveFPA1;
-	uint32_t saveDPLL_A;
-	uint32_t saveDPLL_A_MD;
-	uint32_t saveHTOTAL_A;
-	uint32_t saveHBLANK_A;
-	uint32_t saveHSYNC_A;
-	uint32_t saveVTOTAL_A;
-	uint32_t saveVBLANK_A;
-	uint32_t saveVSYNC_A;
-	uint32_t saveDSPASTRIDE;
-	uint32_t saveDSPASIZE;
-	uint32_t saveDSPAPOS;
-	uint32_t saveDSPABASE;
-	uint32_t saveDSPASURF;
-	uint32_t saveDSPASTATUS;
-	uint32_t saveFPB0;
-	uint32_t saveFPB1;
-	uint32_t saveDPLL_B;
-	uint32_t saveDPLL_B_MD;
-	uint32_t saveHTOTAL_B;
-	uint32_t saveHBLANK_B;
-	uint32_t saveHSYNC_B;
-	uint32_t saveVTOTAL_B;
-	uint32_t saveVBLANK_B;
-	uint32_t saveVSYNC_B;
-	uint32_t saveDSPBSTRIDE;
-	uint32_t saveDSPBSIZE;
-	uint32_t saveDSPBPOS;
-	uint32_t saveDSPBBASE;
-	uint32_t saveDSPBSURF;
-	uint32_t saveDSPBSTATUS;
-	uint32_t saveVCLK_DIVISOR_VGA0;
-	uint32_t saveVCLK_DIVISOR_VGA1;
-	uint32_t saveVCLK_POST_DIV;
-	uint32_t saveVGACNTRL;
-	uint32_t saveADPA;
-	uint32_t saveLVDS;
-	uint32_t saveDVOA;
-	uint32_t saveDVOB;
-	uint32_t saveDVOC;
-	uint32_t savePP_ON;
-	uint32_t savePP_OFF;
-	uint32_t savePP_CONTROL;
-	uint32_t savePP_CYCLE;
-	uint32_t savePFIT_CONTROL;
-	uint32_t savePaletteA[256];
-	uint32_t savePaletteB[256];
-	uint32_t saveBLC_PWM_CTL2;
-	uint32_t saveBLC_PWM_CTL;
-	uint32_t saveCLOCKGATING;
-	uint32_t saveDSPARB;
-	uint32_t saveDSPATILEOFF;
-	uint32_t saveDSPBTILEOFF;
-	uint32_t saveDSPAADDR;
-	uint32_t saveDSPBADDR;
-	uint32_t savePFIT_AUTO_RATIOS;
-	uint32_t savePFIT_PGM_RATIOS;
-	uint32_t savePP_ON_DELAYS;
-	uint32_t savePP_OFF_DELAYS;
-	uint32_t savePP_DIVISOR;
-	uint32_t saveBSM;
-	uint32_t saveVBT;
-	uint32_t saveBCLRPAT_A;
-	uint32_t saveBCLRPAT_B;
-	uint32_t saveDSPALINOFF;
-	uint32_t saveDSPBLINOFF;
-	uint32_t savePERF_MODE;
-	uint32_t saveDSPFW1;
-	uint32_t saveDSPFW2;
-	uint32_t saveDSPFW3;
-	uint32_t saveDSPFW4;
-	uint32_t saveDSPFW5;
-	uint32_t saveDSPFW6;
-	uint32_t saveCHICKENBIT;
-	uint32_t saveDSPACURSOR_CTRL;
-	uint32_t saveDSPBCURSOR_CTRL;
-	uint32_t saveDSPACURSOR_BASE;
-	uint32_t saveDSPBCURSOR_BASE;
-	uint32_t saveDSPACURSOR_POS;
-	uint32_t saveDSPBCURSOR_POS;
-	uint32_t save_palette_a[256];
-	uint32_t save_palette_b[256];
-	uint32_t saveOV_OVADD;
-	uint32_t saveOV_OGAMC0;
-	uint32_t saveOV_OGAMC1;
-	uint32_t saveOV_OGAMC2;
-	uint32_t saveOV_OGAMC3;
-	uint32_t saveOV_OGAMC4;
-	uint32_t saveOV_OGAMC5;
-	uint32_t saveOVC_OVADD;
-	uint32_t saveOVC_OGAMC0;
-	uint32_t saveOVC_OGAMC1;
-	uint32_t saveOVC_OGAMC2;
-	uint32_t saveOVC_OGAMC3;
-	uint32_t saveOVC_OGAMC4;
-	uint32_t saveOVC_OGAMC5;
+
+	struct psb_save_area regs;
 
 	/* MSI reg save */
 	uint32_t msi_addr;
 	uint32_t msi_data;
 
-	/* Medfield specific register save state */
-	uint32_t saveHDMIPHYMISCCTL;
-	uint32_t saveHDMIB_CONTROL;
-	uint32_t saveDSPCCNTR;
-	uint32_t savePIPECCONF;
-	uint32_t savePIPECSRC;
-	uint32_t saveHTOTAL_C;
-	uint32_t saveHBLANK_C;
-	uint32_t saveHSYNC_C;
-	uint32_t saveVTOTAL_C;
-	uint32_t saveVBLANK_C;
-	uint32_t saveVSYNC_C;
-	uint32_t saveDSPCSTRIDE;
-	uint32_t saveDSPCSIZE;
-	uint32_t saveDSPCPOS;
-	uint32_t saveDSPCSURF;
-	uint32_t saveDSPCSTATUS;
-	uint32_t saveDSPCLINOFF;
-	uint32_t saveDSPCTILEOFF;
-	uint32_t saveDSPCCURSOR_CTRL;
-	uint32_t saveDSPCCURSOR_BASE;
-	uint32_t saveDSPCCURSOR_POS;
-	uint32_t save_palette_c[256];
-	uint32_t saveOV_OVADD_C;
-	uint32_t saveOV_OGAMC0_C;
-	uint32_t saveOV_OGAMC1_C;
-	uint32_t saveOV_OGAMC2_C;
-	uint32_t saveOV_OGAMC3_C;
-	uint32_t saveOV_OGAMC4_C;
-	uint32_t saveOV_OGAMC5_C;
-
-	/* DSI register save */
-	uint32_t saveDEVICE_READY_REG;
-	uint32_t saveINTR_EN_REG;
-	uint32_t saveDSI_FUNC_PRG_REG;
-	uint32_t saveHS_TX_TIMEOUT_REG;
-	uint32_t saveLP_RX_TIMEOUT_REG;
-	uint32_t saveTURN_AROUND_TIMEOUT_REG;
-	uint32_t saveDEVICE_RESET_REG;
-	uint32_t saveDPI_RESOLUTION_REG;
-	uint32_t saveHORIZ_SYNC_PAD_COUNT_REG;
-	uint32_t saveHORIZ_BACK_PORCH_COUNT_REG;
-	uint32_t saveHORIZ_FRONT_PORCH_COUNT_REG;
-	uint32_t saveHORIZ_ACTIVE_AREA_COUNT_REG;
-	uint32_t saveVERT_SYNC_PAD_COUNT_REG;
-	uint32_t saveVERT_BACK_PORCH_COUNT_REG;
-	uint32_t saveVERT_FRONT_PORCH_COUNT_REG;
-	uint32_t saveHIGH_LOW_SWITCH_COUNT_REG;
-	uint32_t saveINIT_COUNT_REG;
-	uint32_t saveMAX_RET_PAK_REG;
-	uint32_t saveVIDEO_FMT_REG;
-	uint32_t saveEOT_DISABLE_REG;
-	uint32_t saveLP_BYTECLK_REG;
-	uint32_t saveHS_LS_DBI_ENABLE_REG;
-	uint32_t saveTXCLKESC_REG;
-	uint32_t saveDPHY_PARAM_REG;
-	uint32_t saveMIPI_CONTROL_REG;
-	uint32_t saveMIPI;
-	uint32_t saveMIPI_C;
-
-	/* DPST register save */
-	uint32_t saveHISTOGRAM_INT_CONTROL_REG;
-	uint32_t saveHISTOGRAM_LOGIC_CONTROL_REG;
-	uint32_t savePWM_CONTROL_LOGIC;
 
 	/*
-	 * DSI info. 
-	 */
-	void * dbi_dsr_info;	
-	void * dbi_dpu_info;
-	void * dsi_configs[2];
-	/*
 	 * LID-Switch
 	 */
 	spinlock_t lid_lock;
@@ -635,6 +651,24 @@ struct drm_psb_private {
 
 	/* 2D acceleration */
 	spinlock_t lock_2d;
+
+	/*
+	 * Panel brightness
+	 */
+	int brightness;
+	int brightness_adjusted;
+
+	bool dsr_enable;
+	u32 dsr_fb_update;
+	bool dpi_panel_on[3];
+	void *dsi_configs[2];
+	u32 bpp;
+	u32 bpp2;
+
+	u32 pipeconf[3];
+	u32 dspcntr[3];
+
+	int mdfld_panel_id;
 };
 
 
@@ -830,6 +864,9 @@ extern const struct psb_ops psb_chip_ops;
 /* oaktrail_device.c */
 extern const struct psb_ops oaktrail_chip_ops;
 
+/* mdlfd_device.c */
+extern const struct psb_ops mdfld_chip_ops;
+
 /* cdv_device.c */
 extern const struct psb_ops cdv_chip_ops;
 
diff --git a/drivers/gpu/drm/gma500/psb_intel_display.c b/drivers/gpu/drm/gma500/psb_intel_display.c
index 49e983508d5c..2616558457c8 100644
--- a/drivers/gpu/drm/gma500/psb_intel_display.c
+++ b/drivers/gpu/drm/gma500/psb_intel_display.c
@@ -333,7 +333,7 @@ void psb_intel_wait_for_vblank(struct drm_device *dev)
 	mdelay(20);
 }
 
-int psb_intel_pipe_set_base(struct drm_crtc *crtc,
+static int psb_intel_pipe_set_base(struct drm_crtc *crtc,
 			    int x, int y, struct drm_framebuffer *old_fb)
 {
 	struct drm_device *dev = crtc->dev;
@@ -433,7 +433,6 @@ static void psb_intel_crtc_dpms(struct drm_crtc *crtc, int mode)
 	int dspbase_reg = (pipe == 0) ? DSPABASE : DSPBBASE;
 	int pipeconf_reg = (pipe == 0) ? PIPEACONF : PIPEBCONF;
 	u32 temp;
-	bool enabled;
 
 	/* XXX: When our outputs are all unaware of DPMS modes other than off
 	 * and on, we should map those modes to DRM_MODE_DPMS_OFF in the CRTC.
@@ -518,8 +517,6 @@ static void psb_intel_crtc_dpms(struct drm_crtc *crtc, int mode)
 		break;
 	}
 
-	enabled = crtc->enabled && mode != DRM_MODE_DPMS_OFF;
-
 	/*Set FIFO Watermarks*/
 	REG_WRITE(DSPARB, 0x3F3E);
 }
@@ -611,8 +608,8 @@ static int psb_intel_crtc_mode_set(struct drm_crtc *crtc,
 	int refclk;
 	struct psb_intel_clock_t clock;
 	u32 dpll = 0, fp = 0, dspcntr, pipeconf;
-	bool ok, is_sdvo = false, is_dvo = false;
-	bool is_crt = false, is_lvds = false, is_tv = false;
+	bool ok, is_sdvo = false;
+	bool is_lvds = false, is_tv = false;
 	struct drm_mode_config *mode_config = &dev->mode_config;
 	struct drm_connector *connector;
 
@@ -637,15 +634,9 @@ static int psb_intel_crtc_mode_set(struct drm_crtc *crtc,
 		case INTEL_OUTPUT_SDVO:
 			is_sdvo = true;
 			break;
-		case INTEL_OUTPUT_DVO:
-			is_dvo = true;
-			break;
 		case INTEL_OUTPUT_TVOUT:
 			is_tv = true;
 			break;
-		case INTEL_OUTPUT_ANALOG:
-			is_crt = true;
-			break;
 		}
 	}
 
@@ -845,7 +836,7 @@ void psb_intel_crtc_load_lut(struct drm_crtc *crtc)
 		gma_power_end(dev);
 	} else {
 		for (i = 0; i < 256; i++) {
-			dev_priv->save_palette_a[i] =
+			dev_priv->regs.psb.save_palette_a[i] =
 				  ((psb_intel_crtc->lut_r[i] +
 				  psb_intel_crtc->lut_adj[i]) << 16) |
 				  ((psb_intel_crtc->lut_g[i] +
@@ -1141,18 +1132,20 @@ static int psb_intel_crtc_clock_get(struct drm_device *dev,
 		gma_power_end(dev);
 	} else {
 		dpll = (pipe == 0) ?
-			dev_priv->saveDPLL_A : dev_priv->saveDPLL_B;
+			dev_priv->regs.psb.saveDPLL_A :
+			dev_priv->regs.psb.saveDPLL_B;
 
 		if ((dpll & DISPLAY_RATE_SELECT_FPA1) == 0)
 			fp = (pipe == 0) ?
-				dev_priv->saveFPA0 :
-				dev_priv->saveFPB0;
+				dev_priv->regs.psb.saveFPA0 :
+				dev_priv->regs.psb.saveFPB0;
 		else
 			fp = (pipe == 0) ?
-				dev_priv->saveFPA1 :
-				dev_priv->saveFPB1;
+				dev_priv->regs.psb.saveFPA1 :
+				dev_priv->regs.psb.saveFPB1;
 
-		is_lvds = (pipe == 1) && (dev_priv->saveLVDS & LVDS_PORT_EN);
+		is_lvds = (pipe == 1) && (dev_priv->regs.psb.saveLVDS &
+								LVDS_PORT_EN);
 	}
 
 	clock.m1 = (fp & FP_M1_DIV_MASK) >> FP_M1_DIV_SHIFT;
@@ -1218,13 +1211,17 @@ struct drm_display_mode *psb_intel_crtc_mode_get(struct drm_device *dev,
 		gma_power_end(dev);
 	} else {
 		htot = (pipe == 0) ?
-			dev_priv->saveHTOTAL_A : dev_priv->saveHTOTAL_B;
+			dev_priv->regs.psb.saveHTOTAL_A :
+			dev_priv->regs.psb.saveHTOTAL_B;
 		hsync = (pipe == 0) ?
-			dev_priv->saveHSYNC_A : dev_priv->saveHSYNC_B;
+			dev_priv->regs.psb.saveHSYNC_A :
+			dev_priv->regs.psb.saveHSYNC_B;
 		vtot = (pipe == 0) ?
-			dev_priv->saveVTOTAL_A : dev_priv->saveVTOTAL_B;
+			dev_priv->regs.psb.saveVTOTAL_A :
+			dev_priv->regs.psb.saveVTOTAL_B;
 		vsync = (pipe == 0) ?
-			dev_priv->saveVSYNC_A : dev_priv->saveVSYNC_B;
+			dev_priv->regs.psb.saveVSYNC_A :
+			dev_priv->regs.psb.saveVSYNC_B;
 	}
 
 	mode = kzalloc(sizeof(*mode), GFP_KERNEL);
@@ -1419,13 +1416,6 @@ int psb_intel_connector_clones(struct drm_device *dev, int type_mask)
 	return index_mask;
 }
 
-
-void psb_intel_modeset_cleanup(struct drm_device *dev)
-{
-	drm_mode_config_cleanup(dev);
-}
-
-
 /* current intel driver doesn't take advantage of encoders
    always give back the encoder for the connector
 */
diff --git a/drivers/gpu/drm/gma500/psb_intel_lvds.c b/drivers/gpu/drm/gma500/psb_intel_lvds.c
index 0a437586d8cc..c83f5b5d1057 100644
--- a/drivers/gpu/drm/gma500/psb_intel_lvds.c
+++ b/drivers/gpu/drm/gma500/psb_intel_lvds.c
@@ -77,7 +77,7 @@ static u32 psb_intel_lvds_get_max_backlight(struct drm_device *dev)
 		ret = REG_READ(BLC_PWM_CTL);
 		gma_power_end(dev);
 	} else /* Powered off, use the saved value */
-		ret = dev_priv->saveBLC_PWM_CTL;
+		ret = dev_priv->regs.saveBLC_PWM_CTL;
 
 	/* Top 15bits hold the frequency mask */
 	ret = (ret &  BACKLIGHT_MODULATION_FREQ_MASK) >>
@@ -86,7 +86,7 @@ static u32 psb_intel_lvds_get_max_backlight(struct drm_device *dev)
         ret *= 2;	/* Return a 16bit range as needed for setting */
         if (ret == 0)
                 dev_err(dev->dev, "BL bug: Reg %08x save %08X\n",
-                        REG_READ(BLC_PWM_CTL), dev_priv->saveBLC_PWM_CTL);
+                        REG_READ(BLC_PWM_CTL), dev_priv->regs.saveBLC_PWM_CTL);
 	return ret;
 }
 
@@ -203,13 +203,13 @@ static void psb_intel_lvds_set_backlight(struct drm_device *dev, int level)
 		REG_WRITE(BLC_PWM_CTL,
 				(blc_pwm_ctl |
 				(level << BACKLIGHT_DUTY_CYCLE_SHIFT)));
-		dev_priv->saveBLC_PWM_CTL = (blc_pwm_ctl |
+		dev_priv->regs.saveBLC_PWM_CTL = (blc_pwm_ctl |
 					(level << BACKLIGHT_DUTY_CYCLE_SHIFT));
 		gma_power_end(dev);
 	} else {
-		blc_pwm_ctl = dev_priv->saveBLC_PWM_CTL &
+		blc_pwm_ctl = dev_priv->regs.saveBLC_PWM_CTL &
 				~BACKLIGHT_DUTY_CYCLE_MASK;
-		dev_priv->saveBLC_PWM_CTL = (blc_pwm_ctl |
+		dev_priv->regs.saveBLC_PWM_CTL = (blc_pwm_ctl |
 					(level << BACKLIGHT_DUTY_CYCLE_SHIFT));
 	}
 }
@@ -283,7 +283,7 @@ static void psb_intel_lvds_save(struct drm_connector *connector)
 	lvds_priv->savePFIT_PGM_RATIOS = REG_READ(PFIT_PGM_RATIOS);
 
 	/*TODO: move backlight_duty_cycle to psb_intel_lvds_priv*/
-	dev_priv->backlight_duty_cycle = (dev_priv->saveBLC_PWM_CTL &
+	dev_priv->backlight_duty_cycle = (dev_priv->regs.saveBLC_PWM_CTL &
 						BACKLIGHT_DUTY_CYCLE_MASK);
 
 	/*
diff --git a/drivers/gpu/drm/gma500/psb_intel_reg.h b/drivers/gpu/drm/gma500/psb_intel_reg.h
index fcc0af03d685..e89d3a2e8fdc 100644
--- a/drivers/gpu/drm/gma500/psb_intel_reg.h
+++ b/drivers/gpu/drm/gma500/psb_intel_reg.h
@@ -177,6 +177,9 @@
 #define LVDSPP_OFF		0x6120c
 #define PP_CYCLE		0x61210
 
+#define PP_ON_DELAYS		0x61208		/* Cedartrail */
+#define PP_OFF_DELAYS		0x6120c		/* Cedartrail */
+
 #define PFIT_CONTROL		0x61230
 #define PFIT_ENABLE			(1 << 31)
 #define PFIT_PIPE_MASK			(3 << 29)
@@ -1252,6 +1255,12 @@ No status bits are changed.
 # define SB_BYTE_ENABLE_SHIFT                   4
 # define SB_BUSY                                (1 << 0)
 
+#define DSPCLK_GATE_D		0x6200
+# define VRHUNIT_CLOCK_GATE_DISABLE		(1 << 28) /* Fixed value on CDV */
+# define DPOUNIT_CLOCK_GATE_DISABLE		(1 << 11)
+# define DPIOUNIT_CLOCK_GATE_DISABLE		(1 << 6)
+
+#define RAMCLK_GATE_D		0x6210
 
 /* 32-bit value read/written from the DPIO reg. */
 #define SB_DATA		0x02104 /* cedarview */
diff --git a/drivers/gpu/drm/gma500/psb_intel_sdvo.c b/drivers/gpu/drm/gma500/psb_intel_sdvo.c
index 88b42971c0fd..36330cabcea2 100644
--- a/drivers/gpu/drm/gma500/psb_intel_sdvo.c
+++ b/drivers/gpu/drm/gma500/psb_intel_sdvo.c
@@ -1301,7 +1301,7 @@ psb_intel_sdvo_get_analog_edid(struct drm_connector *connector)
 	return NULL;
 }
 
-enum drm_connector_status
+static enum drm_connector_status
 psb_intel_sdvo_hdmi_sink_detect(struct drm_connector *connector)
 {
 	struct psb_intel_sdvo *psb_intel_sdvo = intel_attached_sdvo(connector);
@@ -2312,10 +2312,8 @@ static bool psb_intel_sdvo_tv_create_property(struct psb_intel_sdvo *psb_intel_s
 		psb_intel_sdvo_connector->max_##name = data_value[0]; \
 		psb_intel_sdvo_connector->cur_##name = response; \
 		psb_intel_sdvo_connector->name = \
-			drm_property_create(dev, DRM_MODE_PROP_RANGE, #name, 2); \
+			drm_property_create_range(dev, 0, #name, 0, data_value[0]); \
 		if (!psb_intel_sdvo_connector->name) return false; \
-		psb_intel_sdvo_connector->name->values[0] = 0; \
-		psb_intel_sdvo_connector->name->values[1] = data_value[0]; \
 		drm_connector_attach_property(connector, \
 					      psb_intel_sdvo_connector->name, \
 					      psb_intel_sdvo_connector->cur_##name); \
@@ -2349,25 +2347,19 @@ psb_intel_sdvo_create_enhance_property_tv(struct psb_intel_sdvo *psb_intel_sdvo,
 		psb_intel_sdvo_connector->left_margin = data_value[0] - response;
 		psb_intel_sdvo_connector->right_margin = psb_intel_sdvo_connector->left_margin;
 		psb_intel_sdvo_connector->left =
-			drm_property_create(dev, DRM_MODE_PROP_RANGE,
-					    "left_margin", 2);
+			drm_property_create_range(dev, 0, "left_margin", 0, data_value[0]);
 		if (!psb_intel_sdvo_connector->left)
 			return false;
 
-		psb_intel_sdvo_connector->left->values[0] = 0;
-		psb_intel_sdvo_connector->left->values[1] = data_value[0];
 		drm_connector_attach_property(connector,
 					      psb_intel_sdvo_connector->left,
 					      psb_intel_sdvo_connector->left_margin);
 
 		psb_intel_sdvo_connector->right =
-			drm_property_create(dev, DRM_MODE_PROP_RANGE,
-					    "right_margin", 2);
+			drm_property_create_range(dev, 0, "right_margin", 0, data_value[0]);
 		if (!psb_intel_sdvo_connector->right)
 			return false;
 
-		psb_intel_sdvo_connector->right->values[0] = 0;
-		psb_intel_sdvo_connector->right->values[1] = data_value[0];
 		drm_connector_attach_property(connector,
 					      psb_intel_sdvo_connector->right,
 					      psb_intel_sdvo_connector->right_margin);
@@ -2391,25 +2383,19 @@ psb_intel_sdvo_create_enhance_property_tv(struct psb_intel_sdvo *psb_intel_sdvo,
 		psb_intel_sdvo_connector->top_margin = data_value[0] - response;
 		psb_intel_sdvo_connector->bottom_margin = psb_intel_sdvo_connector->top_margin;
 		psb_intel_sdvo_connector->top =
-			drm_property_create(dev, DRM_MODE_PROP_RANGE,
-					    "top_margin", 2);
+			drm_property_create_range(dev, 0, "top_margin", 0, data_value[0]);
 		if (!psb_intel_sdvo_connector->top)
 			return false;
 
-		psb_intel_sdvo_connector->top->values[0] = 0;
-		psb_intel_sdvo_connector->top->values[1] = data_value[0];
 		drm_connector_attach_property(connector,
 					      psb_intel_sdvo_connector->top,
 					      psb_intel_sdvo_connector->top_margin);
 
 		psb_intel_sdvo_connector->bottom =
-			drm_property_create(dev, DRM_MODE_PROP_RANGE,
-					    "bottom_margin", 2);
+			drm_property_create_range(dev, 0, "bottom_margin", 0, data_value[0]);
 		if (!psb_intel_sdvo_connector->bottom)
 			return false;
 
-		psb_intel_sdvo_connector->bottom->values[0] = 0;
-		psb_intel_sdvo_connector->bottom->values[1] = data_value[0];
 		drm_connector_attach_property(connector,
 					      psb_intel_sdvo_connector->bottom,
 					      psb_intel_sdvo_connector->bottom_margin);
@@ -2438,12 +2424,10 @@ psb_intel_sdvo_create_enhance_property_tv(struct psb_intel_sdvo *psb_intel_sdvo,
 		psb_intel_sdvo_connector->max_dot_crawl = 1;
 		psb_intel_sdvo_connector->cur_dot_crawl = response & 0x1;
 		psb_intel_sdvo_connector->dot_crawl =
-			drm_property_create(dev, DRM_MODE_PROP_RANGE, "dot_crawl", 2);
+			drm_property_create_range(dev, 0, "dot_crawl", 0, 1);
 		if (!psb_intel_sdvo_connector->dot_crawl)
 			return false;
 
-		psb_intel_sdvo_connector->dot_crawl->values[0] = 0;
-		psb_intel_sdvo_connector->dot_crawl->values[1] = 1;
 		drm_connector_attach_property(connector,
 					      psb_intel_sdvo_connector->dot_crawl,
 					      psb_intel_sdvo_connector->cur_dot_crawl);
diff --git a/drivers/gpu/drm/gma500/psb_irq.c b/drivers/gpu/drm/gma500/psb_irq.c
index 7be802baceb5..1869586457b1 100644
--- a/drivers/gpu/drm/gma500/psb_irq.c
+++ b/drivers/gpu/drm/gma500/psb_irq.c
@@ -27,6 +27,8 @@
 #include "psb_reg.h"
 #include "psb_intel_reg.h"
 #include "power.h"
+#include "psb_irq.h"
+#include "mdfld_output.h"
 
 /*
  * inline functions
@@ -113,7 +115,7 @@ psb_disable_pipestat(struct drm_psb_private *dev_priv, int pipe, u32 mask)
 	}
 }
 
-void mid_enable_pipe_event(struct drm_psb_private *dev_priv, int pipe)
+static void mid_enable_pipe_event(struct drm_psb_private *dev_priv, int pipe)
 {
 	if (gma_power_begin(dev_priv->dev, false)) {
 		u32 pipe_event = mid_pipe_event(pipe);
@@ -124,7 +126,7 @@ void mid_enable_pipe_event(struct drm_psb_private *dev_priv, int pipe)
 	}
 }
 
-void mid_disable_pipe_event(struct drm_psb_private *dev_priv, int pipe)
+static void mid_disable_pipe_event(struct drm_psb_private *dev_priv, int pipe)
 {
 	if (dev_priv->pipestat[pipe] == 0) {
 		if (gma_power_begin(dev_priv->dev, false)) {
@@ -453,6 +455,11 @@ int psb_enable_vblank(struct drm_device *dev, int pipe)
 	uint32_t reg_val = 0;
 	uint32_t pipeconf_reg = mid_pipeconf(pipe);
 
+	/* Medfield is different - we should perhaps extract out vblank
+	   and blacklight etc ops */
+	if (IS_MFLD(dev))
+		return mdfld_enable_te(dev, pipe);
+
 	if (gma_power_begin(dev, false)) {
 		reg_val = REG_READ(pipeconf_reg);
 		gma_power_end(dev);
@@ -485,6 +492,8 @@ void psb_disable_vblank(struct drm_device *dev, int pipe)
 	struct drm_psb_private *dev_priv = dev->dev_private;
 	unsigned long irqflags;
 
+	if (IS_MFLD(dev))
+		mdfld_disable_te(dev, pipe);
 	spin_lock_irqsave(&dev_priv->irqmask_lock, irqflags);
 
 	if (pipe == 0)
@@ -499,6 +508,55 @@ void psb_disable_vblank(struct drm_device *dev, int pipe)
 	spin_unlock_irqrestore(&dev_priv->irqmask_lock, irqflags);
 }
 
+/*
+ * It is used to enable TE interrupt
+ */
+int mdfld_enable_te(struct drm_device *dev, int pipe)
+{
+	struct drm_psb_private *dev_priv =
+		(struct drm_psb_private *) dev->dev_private;
+	unsigned long irqflags;
+	uint32_t reg_val = 0;
+	uint32_t pipeconf_reg = mid_pipeconf(pipe);
+
+	if (gma_power_begin(dev, false)) {
+		reg_val = REG_READ(pipeconf_reg);
+		gma_power_end(dev);
+	}
+
+	if (!(reg_val & PIPEACONF_ENABLE))
+		return -EINVAL;
+
+	spin_lock_irqsave(&dev_priv->irqmask_lock, irqflags);
+
+	mid_enable_pipe_event(dev_priv, pipe);
+	psb_enable_pipestat(dev_priv, pipe, PIPE_TE_ENABLE);
+
+	spin_unlock_irqrestore(&dev_priv->irqmask_lock, irqflags);
+
+	return 0;
+}
+
+/*
+ * It is used to disable TE interrupt
+ */
+void mdfld_disable_te(struct drm_device *dev, int pipe)
+{
+	struct drm_psb_private *dev_priv =
+		(struct drm_psb_private *) dev->dev_private;
+	unsigned long irqflags;
+
+	if (!dev_priv->dsr_enable)
+		return;
+
+	spin_lock_irqsave(&dev_priv->irqmask_lock, irqflags);
+
+	mid_disable_pipe_event(dev_priv, pipe);
+	psb_disable_pipestat(dev_priv, pipe, PIPE_TE_ENABLE);
+
+	spin_unlock_irqrestore(&dev_priv->irqmask_lock, irqflags);
+}
+
 /* Called from drm generic code, passed a 'crtc', which
  * we use as a pipe index
  */
diff --git a/drivers/gpu/drm/gma500/psb_irq.h b/drivers/gpu/drm/gma500/psb_irq.h
index 216fda38b57d..603045bee58a 100644
--- a/drivers/gpu/drm/gma500/psb_irq.h
+++ b/drivers/gpu/drm/gma500/psb_irq.h
@@ -42,4 +42,6 @@ int  psb_enable_vblank(struct drm_device *dev, int pipe);
 void psb_disable_vblank(struct drm_device *dev, int pipe);
 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_ */
diff --git a/drivers/gpu/drm/gma500/tc35876x-dsi-lvds.c b/drivers/gpu/drm/gma500/tc35876x-dsi-lvds.c
new file mode 100644
index 000000000000..4a07ab596174
--- /dev/null
+++ b/drivers/gpu/drm/gma500/tc35876x-dsi-lvds.c
@@ -0,0 +1,829 @@
+/*
+ * Copyright © 2011 Intel Corporation
+ *
+ * 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.
+ *
+ */
+
+#include "mdfld_dsi_dpi.h"
+#include "mdfld_output.h"
+#include "mdfld_dsi_pkg_sender.h"
+#include "tc35876x-dsi-lvds.h"
+#include <linux/i2c/tc35876x.h>
+#include <linux/kernel.h>
+#include <linux/module.h>
+#include <asm/intel_scu_ipc.h>
+
+static struct i2c_client *tc35876x_client;
+static struct i2c_client *cmi_lcd_i2c_client;
+
+#define FLD_MASK(start, end)	(((1 << ((start) - (end) + 1)) - 1) << (end))
+#define FLD_VAL(val, start, end) (((val) << (end)) & FLD_MASK(start, end))
+
+/* DSI D-PHY Layer Registers */
+#define D0W_DPHYCONTTX		0x0004
+#define CLW_DPHYCONTRX		0x0020
+#define D0W_DPHYCONTRX		0x0024
+#define D1W_DPHYCONTRX		0x0028
+#define D2W_DPHYCONTRX		0x002C
+#define D3W_DPHYCONTRX		0x0030
+#define COM_DPHYCONTRX		0x0038
+#define CLW_CNTRL		0x0040
+#define D0W_CNTRL		0x0044
+#define D1W_CNTRL		0x0048
+#define D2W_CNTRL		0x004C
+#define D3W_CNTRL		0x0050
+#define DFTMODE_CNTRL		0x0054
+
+/* DSI PPI Layer Registers */
+#define PPI_STARTPPI		0x0104
+#define PPI_BUSYPPI		0x0108
+#define PPI_LINEINITCNT		0x0110
+#define PPI_LPTXTIMECNT		0x0114
+#define PPI_LANEENABLE		0x0134
+#define PPI_TX_RX_TA		0x013C
+#define PPI_CLS_ATMR		0x0140
+#define PPI_D0S_ATMR		0x0144
+#define PPI_D1S_ATMR		0x0148
+#define PPI_D2S_ATMR		0x014C
+#define PPI_D3S_ATMR		0x0150
+#define PPI_D0S_CLRSIPOCOUNT	0x0164
+#define PPI_D1S_CLRSIPOCOUNT	0x0168
+#define PPI_D2S_CLRSIPOCOUNT	0x016C
+#define PPI_D3S_CLRSIPOCOUNT	0x0170
+#define CLS_PRE			0x0180
+#define D0S_PRE			0x0184
+#define D1S_PRE			0x0188
+#define D2S_PRE			0x018C
+#define D3S_PRE			0x0190
+#define CLS_PREP		0x01A0
+#define D0S_PREP		0x01A4
+#define D1S_PREP		0x01A8
+#define D2S_PREP		0x01AC
+#define D3S_PREP		0x01B0
+#define CLS_ZERO		0x01C0
+#define D0S_ZERO		0x01C4
+#define D1S_ZERO		0x01C8
+#define D2S_ZERO		0x01CC
+#define D3S_ZERO		0x01D0
+#define PPI_CLRFLG		0x01E0
+#define PPI_CLRSIPO		0x01E4
+#define HSTIMEOUT		0x01F0
+#define HSTIMEOUTENABLE		0x01F4
+
+/* DSI Protocol Layer Registers */
+#define DSI_STARTDSI		0x0204
+#define DSI_BUSYDSI		0x0208
+#define DSI_LANEENABLE		0x0210
+#define DSI_LANESTATUS0		0x0214
+#define DSI_LANESTATUS1		0x0218
+#define DSI_INTSTATUS		0x0220
+#define DSI_INTMASK		0x0224
+#define DSI_INTCLR		0x0228
+#define DSI_LPTXTO		0x0230
+
+/* DSI General Registers */
+#define DSIERRCNT		0x0300
+
+/* DSI Application Layer Registers */
+#define APLCTRL			0x0400
+#define RDPKTLN			0x0404
+
+/* Video Path Registers */
+#define VPCTRL			0x0450
+#define HTIM1			0x0454
+#define HTIM2			0x0458
+#define VTIM1			0x045C
+#define VTIM2			0x0460
+#define VFUEN			0x0464
+
+/* LVDS Registers */
+#define LVMX0003		0x0480
+#define LVMX0407		0x0484
+#define LVMX0811		0x0488
+#define LVMX1215		0x048C
+#define LVMX1619		0x0490
+#define LVMX2023		0x0494
+#define LVMX2427		0x0498
+#define LVCFG			0x049C
+#define LVPHY0			0x04A0
+#define LVPHY1			0x04A4
+
+/* System Registers */
+#define SYSSTAT			0x0500
+#define SYSRST			0x0504
+
+/* GPIO Registers */
+/*#define GPIOC			0x0520*/
+#define GPIOO			0x0524
+#define GPIOI			0x0528
+
+/* I2C Registers */
+#define I2CTIMCTRL		0x0540
+#define I2CMADDR		0x0544
+#define WDATAQ			0x0548
+#define RDATAQ			0x054C
+
+/* Chip/Rev Registers */
+#define IDREG			0x0580
+
+/* Debug Registers */
+#define DEBUG00			0x05A0
+#define DEBUG01			0x05A4
+
+/* Panel CABC registers */
+#define PANEL_PWM_CONTROL	0x90
+#define PANEL_FREQ_DIVIDER_HI	0x91
+#define PANEL_FREQ_DIVIDER_LO	0x92
+#define PANEL_DUTY_CONTROL	0x93
+#define PANEL_MODIFY_RGB	0x94
+#define PANEL_FRAMERATE_CONTROL	0x96
+#define PANEL_PWM_MIN		0x97
+#define PANEL_PWM_REF		0x98
+#define PANEL_PWM_MAX		0x99
+#define PANEL_ALLOW_DISTORT	0x9A
+#define PANEL_BYPASS_PWMI	0x9B
+
+/* Panel color management registers */
+#define PANEL_CM_ENABLE		0x700
+#define PANEL_CM_HUE		0x701
+#define PANEL_CM_SATURATION	0x702
+#define PANEL_CM_INTENSITY	0x703
+#define PANEL_CM_BRIGHTNESS	0x704
+#define PANEL_CM_CE_ENABLE	0x705
+#define PANEL_CM_PEAK_EN	0x710
+#define PANEL_CM_GAIN		0x711
+#define PANEL_CM_HUETABLE_START	0x730
+#define PANEL_CM_HUETABLE_END	0x747 /* inclusive */
+
+/* Input muxing for registers LVMX0003...LVMX2427 */
+enum {
+	INPUT_R0,	/* 0 */
+	INPUT_R1,
+	INPUT_R2,
+	INPUT_R3,
+	INPUT_R4,
+	INPUT_R5,
+	INPUT_R6,
+	INPUT_R7,
+	INPUT_G0,	/* 8 */
+	INPUT_G1,
+	INPUT_G2,
+	INPUT_G3,
+	INPUT_G4,
+	INPUT_G5,
+	INPUT_G6,
+	INPUT_G7,
+	INPUT_B0,	/* 16 */
+	INPUT_B1,
+	INPUT_B2,
+	INPUT_B3,
+	INPUT_B4,
+	INPUT_B5,
+	INPUT_B6,
+	INPUT_B7,
+	INPUT_HSYNC,	/* 24 */
+	INPUT_VSYNC,
+	INPUT_DE,
+	LOGIC_0,
+	/* 28...31 undefined */
+};
+
+#define INPUT_MUX(lvmx03, lvmx02, lvmx01, lvmx00)		\
+	(FLD_VAL(lvmx03, 29, 24) | FLD_VAL(lvmx02, 20, 16) |	\
+	FLD_VAL(lvmx01, 12, 8) | FLD_VAL(lvmx00, 4, 0))
+
+/**
+ * tc35876x_regw - Write DSI-LVDS bridge register using I2C
+ * @client: struct i2c_client to use
+ * @reg: register address
+ * @value: value to write
+ *
+ * Returns 0 on success, or a negative error value.
+ */
+static int tc35876x_regw(struct i2c_client *client, u16 reg, u32 value)
+{
+	int r;
+	u8 tx_data[] = {
+		/* NOTE: Register address big-endian, data little-endian. */
+		(reg >> 8) & 0xff,
+		reg & 0xff,
+		value & 0xff,
+		(value >> 8) & 0xff,
+		(value >> 16) & 0xff,
+		(value >> 24) & 0xff,
+	};
+	struct i2c_msg msgs[] = {
+		{
+			.addr = client->addr,
+			.flags = 0,
+			.buf = tx_data,
+			.len = ARRAY_SIZE(tx_data),
+		},
+	};
+
+	r = i2c_transfer(client->adapter, msgs, ARRAY_SIZE(msgs));
+	if (r < 0) {
+		dev_err(&client->dev, "%s: reg 0x%04x val 0x%08x error %d\n",
+			__func__, reg, value, r);
+		return r;
+	}
+
+	if (r < ARRAY_SIZE(msgs)) {
+		dev_err(&client->dev, "%s: reg 0x%04x val 0x%08x msgs %d\n",
+			__func__, reg, value, r);
+		return -EAGAIN;
+	}
+
+	dev_dbg(&client->dev, "%s: reg 0x%04x val 0x%08x\n",
+			__func__, reg, value);
+
+	return 0;
+}
+
+/**
+ * tc35876x_regr - Read DSI-LVDS bridge register using I2C
+ * @client: struct i2c_client to use
+ * @reg: register address
+ * @value: pointer for storing the value
+ *
+ * Returns 0 on success, or a negative error value.
+ */
+static int tc35876x_regr(struct i2c_client *client, u16 reg, u32 *value)
+{
+	int r;
+	u8 tx_data[] = {
+		(reg >> 8) & 0xff,
+		reg & 0xff,
+	};
+	u8 rx_data[4];
+	struct i2c_msg msgs[] = {
+		{
+			.addr = client->addr,
+			.flags = 0,
+			.buf = tx_data,
+			.len = ARRAY_SIZE(tx_data),
+		},
+		{
+			.addr = client->addr,
+			.flags = I2C_M_RD,
+			.buf = rx_data,
+			.len = ARRAY_SIZE(rx_data),
+		 },
+	};
+
+	r = i2c_transfer(client->adapter, msgs, ARRAY_SIZE(msgs));
+	if (r < 0) {
+		dev_err(&client->dev, "%s: reg 0x%04x error %d\n", __func__,
+			reg, r);
+		return r;
+	}
+
+	if (r < ARRAY_SIZE(msgs)) {
+		dev_err(&client->dev, "%s: reg 0x%04x msgs %d\n", __func__,
+			reg, r);
+		return -EAGAIN;
+	}
+
+	*value = rx_data[0] << 24 | rx_data[1] << 16 |
+		rx_data[2] << 8 | rx_data[3];
+
+	dev_dbg(&client->dev, "%s: reg 0x%04x value 0x%08x\n", __func__,
+		reg, *value);
+
+	return 0;
+}
+
+void tc35876x_set_bridge_reset_state(struct drm_device *dev, int state)
+{
+	struct tc35876x_platform_data *pdata;
+
+	if (WARN(!tc35876x_client, "%s called before probe", __func__))
+		return;
+
+	dev_dbg(&tc35876x_client->dev, "%s: state %d\n", __func__, state);
+
+	pdata = dev_get_platdata(&tc35876x_client->dev);
+
+	if (pdata->gpio_bridge_reset == -1)
+		return;
+
+	if (state) {
+		gpio_set_value_cansleep(pdata->gpio_bridge_reset, 0);
+		mdelay(10);
+	} else {
+		/* Pull MIPI Bridge reset pin to Low */
+		gpio_set_value_cansleep(pdata->gpio_bridge_reset, 0);
+		mdelay(20);
+		/* Pull MIPI Bridge reset pin to High */
+		gpio_set_value_cansleep(pdata->gpio_bridge_reset, 1);
+		mdelay(40);
+	}
+}
+
+void tc35876x_configure_lvds_bridge(struct drm_device *dev)
+{
+	struct i2c_client *i2c = tc35876x_client;
+	u32 ppi_lptxtimecnt;
+	u32 txtagocnt;
+	u32 txtasurecnt;
+	u32 id;
+
+	if (WARN(!tc35876x_client, "%s called before probe", __func__))
+		return;
+
+	dev_dbg(&tc35876x_client->dev, "%s\n", __func__);
+
+	if (!tc35876x_regr(i2c, IDREG, &id))
+		dev_info(&tc35876x_client->dev, "tc35876x ID 0x%08x\n", id);
+	else
+		dev_err(&tc35876x_client->dev, "Cannot read ID\n");
+
+	ppi_lptxtimecnt = 4;
+	txtagocnt = (5 * ppi_lptxtimecnt - 3) / 4;
+	txtasurecnt = 3 * ppi_lptxtimecnt / 2;
+	tc35876x_regw(i2c, PPI_TX_RX_TA, FLD_VAL(txtagocnt, 26, 16) |
+		FLD_VAL(txtasurecnt, 10, 0));
+	tc35876x_regw(i2c, PPI_LPTXTIMECNT, FLD_VAL(ppi_lptxtimecnt, 10, 0));
+
+	tc35876x_regw(i2c, PPI_D0S_CLRSIPOCOUNT, FLD_VAL(1, 5, 0));
+	tc35876x_regw(i2c, PPI_D1S_CLRSIPOCOUNT, FLD_VAL(1, 5, 0));
+	tc35876x_regw(i2c, PPI_D2S_CLRSIPOCOUNT, FLD_VAL(1, 5, 0));
+	tc35876x_regw(i2c, PPI_D3S_CLRSIPOCOUNT, FLD_VAL(1, 5, 0));
+
+	/* Enabling MIPI & PPI lanes, Enable 4 lanes */
+	tc35876x_regw(i2c, PPI_LANEENABLE,
+		BIT(4) | BIT(3) | BIT(2) | BIT(1) | BIT(0));
+	tc35876x_regw(i2c, DSI_LANEENABLE,
+		BIT(4) | BIT(3) | BIT(2) | BIT(1) | BIT(0));
+	tc35876x_regw(i2c, PPI_STARTPPI, BIT(0));
+	tc35876x_regw(i2c, DSI_STARTDSI, BIT(0));
+
+	/* Setting LVDS output frequency */
+	tc35876x_regw(i2c, LVPHY0, FLD_VAL(1, 20, 16) |
+		FLD_VAL(2, 15, 14) | FLD_VAL(6, 4, 0)); /* 0x00048006 */
+
+	/* Setting video panel control register,0x00000120 VTGen=ON ?!?!? */
+	tc35876x_regw(i2c, VPCTRL, BIT(8) | BIT(5));
+
+	/* Horizontal back porch and horizontal pulse width. 0x00280028 */
+	tc35876x_regw(i2c, HTIM1, FLD_VAL(40, 24, 16) | FLD_VAL(40, 8, 0));
+
+	/* Horizontal front porch and horizontal active video size. 0x00500500*/
+	tc35876x_regw(i2c, HTIM2, FLD_VAL(80, 24, 16) | FLD_VAL(1280, 10, 0));
+
+	/* Vertical back porch and vertical sync pulse width. 0x000e000a */
+	tc35876x_regw(i2c, VTIM1, FLD_VAL(14, 23, 16) | FLD_VAL(10, 7, 0));
+
+	/* Vertical front porch and vertical display size. 0x000e0320 */
+	tc35876x_regw(i2c, VTIM2, FLD_VAL(14, 23, 16) | FLD_VAL(800, 10, 0));
+
+	/* Set above HTIM1, HTIM2, VTIM1, and VTIM2 at next VSYNC. */
+	tc35876x_regw(i2c, VFUEN, BIT(0));
+
+	/* Soft reset LCD controller. */
+	tc35876x_regw(i2c, SYSRST, BIT(2));
+
+	/* LVDS-TX input muxing */
+	tc35876x_regw(i2c, LVMX0003,
+		INPUT_MUX(INPUT_R5, INPUT_R4, INPUT_R3, INPUT_R2));
+	tc35876x_regw(i2c, LVMX0407,
+		INPUT_MUX(INPUT_G2, INPUT_R7, INPUT_R1, INPUT_R6));
+	tc35876x_regw(i2c, LVMX0811,
+		INPUT_MUX(INPUT_G1, INPUT_G0, INPUT_G4, INPUT_G3));
+	tc35876x_regw(i2c, LVMX1215,
+		INPUT_MUX(INPUT_B2, INPUT_G7, INPUT_G6, INPUT_G5));
+	tc35876x_regw(i2c, LVMX1619,
+		INPUT_MUX(INPUT_B4, INPUT_B3, INPUT_B1, INPUT_B0));
+	tc35876x_regw(i2c, LVMX2023,
+		INPUT_MUX(LOGIC_0,  INPUT_B7, INPUT_B6, INPUT_B5));
+	tc35876x_regw(i2c, LVMX2427,
+		INPUT_MUX(INPUT_R0, INPUT_DE, INPUT_VSYNC, INPUT_HSYNC));
+
+	/* Enable LVDS transmitter. */
+	tc35876x_regw(i2c, LVCFG, BIT(0));
+
+	/* Clear notifications. Don't write reserved bits. Was write 0xffffffff
+	 * to 0x0288, must be in error?! */
+	tc35876x_regw(i2c, DSI_INTCLR, FLD_MASK(31, 30) | FLD_MASK(22, 0));
+}
+
+#define GPIOPWMCTRL	0x38F
+#define PWM0CLKDIV0	0x62 /* low byte */
+#define PWM0CLKDIV1	0x61 /* high byte */
+
+#define SYSTEMCLK	19200000UL /* 19.2 MHz */
+#define PWM_FREQUENCY	9600 /* Hz */
+
+/* f = baseclk / (clkdiv + 1) => clkdiv = (baseclk - f) / f */
+static inline u16 calc_clkdiv(unsigned long baseclk, unsigned int f)
+{
+	return (baseclk - f) / f;
+}
+
+static void tc35876x_brightness_init(struct drm_device *dev)
+{
+	int ret;
+	u8 pwmctrl;
+	u16 clkdiv;
+
+	/* Make sure the PWM reference is the 19.2 MHz system clock. Read first
+	 * instead of setting directly to catch potential conflicts between PWM
+	 * users. */
+	ret = intel_scu_ipc_ioread8(GPIOPWMCTRL, &pwmctrl);
+	if (ret || pwmctrl != 0x01) {
+		if (ret)
+			dev_err(&dev->pdev->dev, "GPIOPWMCTRL read failed\n");
+		else
+			dev_warn(&dev->pdev->dev, "GPIOPWMCTRL was not set to system clock (pwmctrl = 0x%02x)\n", pwmctrl);
+
+		ret = intel_scu_ipc_iowrite8(GPIOPWMCTRL, 0x01);
+		if (ret)
+			dev_err(&dev->pdev->dev, "GPIOPWMCTRL set failed\n");
+	}
+
+	clkdiv = calc_clkdiv(SYSTEMCLK, PWM_FREQUENCY);
+
+	ret = intel_scu_ipc_iowrite8(PWM0CLKDIV1, (clkdiv >> 8) & 0xff);
+	if (!ret)
+		ret = intel_scu_ipc_iowrite8(PWM0CLKDIV0, clkdiv & 0xff);
+
+	if (ret)
+		dev_err(&dev->pdev->dev, "PWM0CLKDIV set failed\n");
+	else
+		dev_dbg(&dev->pdev->dev, "PWM0CLKDIV set to 0x%04x (%d Hz)\n",
+			clkdiv, PWM_FREQUENCY);
+}
+
+#define PWM0DUTYCYCLE			0x67
+
+void tc35876x_brightness_control(struct drm_device *dev, int level)
+{
+	int ret;
+	u8 duty_val;
+	u8 panel_duty_val;
+
+	level = clamp(level, 0, MDFLD_DSI_BRIGHTNESS_MAX_LEVEL);
+
+	/* PWM duty cycle 0x00...0x63 corresponds to 0...99% */
+	duty_val = level * 0x63 / MDFLD_DSI_BRIGHTNESS_MAX_LEVEL;
+
+	/* I won't pretend to understand this formula. The panel spec is quite
+	 * bad engrish.
+	 */
+	panel_duty_val = (2 * level - 100) * 0xA9 /
+			 MDFLD_DSI_BRIGHTNESS_MAX_LEVEL + 0x56;
+
+	ret = intel_scu_ipc_iowrite8(PWM0DUTYCYCLE, duty_val);
+	if (ret)
+		dev_err(&tc35876x_client->dev, "%s: ipc write fail\n",
+			__func__);
+
+	if (cmi_lcd_i2c_client) {
+		ret = i2c_smbus_write_byte_data(cmi_lcd_i2c_client,
+						PANEL_PWM_MAX, panel_duty_val);
+		if (ret < 0)
+			dev_err(&cmi_lcd_i2c_client->dev, "%s: i2c write failed\n",
+				__func__);
+	}
+}
+
+void tc35876x_toshiba_bridge_panel_off(struct drm_device *dev)
+{
+	struct tc35876x_platform_data *pdata;
+
+	if (WARN(!tc35876x_client, "%s called before probe", __func__))
+		return;
+
+	dev_dbg(&tc35876x_client->dev, "%s\n", __func__);
+
+	pdata = dev_get_platdata(&tc35876x_client->dev);
+
+	if (pdata->gpio_panel_bl_en != -1)
+		gpio_set_value_cansleep(pdata->gpio_panel_bl_en, 0);
+
+	if (pdata->gpio_panel_vadd != -1)
+		gpio_set_value_cansleep(pdata->gpio_panel_vadd, 0);
+}
+
+void tc35876x_toshiba_bridge_panel_on(struct drm_device *dev)
+{
+	struct tc35876x_platform_data *pdata;
+	struct drm_psb_private *dev_priv = dev->dev_private;
+
+	if (WARN(!tc35876x_client, "%s called before probe", __func__))
+		return;
+
+	dev_dbg(&tc35876x_client->dev, "%s\n", __func__);
+
+	pdata = dev_get_platdata(&tc35876x_client->dev);
+
+	if (pdata->gpio_panel_vadd != -1) {
+		gpio_set_value_cansleep(pdata->gpio_panel_vadd, 1);
+		msleep(260);
+	}
+
+	if (cmi_lcd_i2c_client) {
+		int ret;
+		dev_dbg(&cmi_lcd_i2c_client->dev, "setting TCON\n");
+		/* Bit 4 is average_saving. Setting it to 1, the brightness is
+		 * referenced to the average of the frame content. 0 means
+		 * reference to the maximum of frame contents. Bits 3:0 are
+		 * allow_distort. When set to a nonzero value, all color values
+		 * between 255-allow_distort*2 and 255 are mapped to the
+		 * 255-allow_distort*2 value.
+		 */
+		ret = i2c_smbus_write_byte_data(cmi_lcd_i2c_client,
+						PANEL_ALLOW_DISTORT, 0x10);
+		if (ret < 0)
+			dev_err(&cmi_lcd_i2c_client->dev,
+				"i2c write failed (%d)\n", ret);
+		ret = i2c_smbus_write_byte_data(cmi_lcd_i2c_client,
+						PANEL_BYPASS_PWMI, 0);
+		if (ret < 0)
+			dev_err(&cmi_lcd_i2c_client->dev,
+				"i2c write failed (%d)\n", ret);
+		/* Set minimum brightness value - this is tunable */
+		ret = i2c_smbus_write_byte_data(cmi_lcd_i2c_client,
+						PANEL_PWM_MIN, 0x35);
+		if (ret < 0)
+			dev_err(&cmi_lcd_i2c_client->dev,
+				"i2c write failed (%d)\n", ret);
+	}
+
+	if (pdata->gpio_panel_bl_en != -1)
+		gpio_set_value_cansleep(pdata->gpio_panel_bl_en, 1);
+
+	tc35876x_brightness_control(dev, dev_priv->brightness_adjusted);
+}
+
+static struct drm_display_mode *tc35876x_get_config_mode(struct drm_device *dev)
+{
+	struct drm_display_mode *mode;
+
+	dev_dbg(&dev->pdev->dev, "%s\n", __func__);
+
+	mode = kzalloc(sizeof(*mode), GFP_KERNEL);
+	if (!mode)
+		return NULL;
+
+	/* FIXME: do this properly. */
+	mode->hdisplay = 1280;
+	mode->vdisplay = 800;
+	mode->hsync_start = 1360;
+	mode->hsync_end = 1400;
+	mode->htotal = 1440;
+	mode->vsync_start = 814;
+	mode->vsync_end = 824;
+	mode->vtotal = 838;
+	mode->clock = 33324 << 1;
+
+	dev_info(&dev->pdev->dev, "hdisplay(w) = %d\n", mode->hdisplay);
+	dev_info(&dev->pdev->dev, "vdisplay(h) = %d\n", mode->vdisplay);
+	dev_info(&dev->pdev->dev, "HSS = %d\n", mode->hsync_start);
+	dev_info(&dev->pdev->dev, "HSE = %d\n", mode->hsync_end);
+	dev_info(&dev->pdev->dev, "htotal = %d\n", mode->htotal);
+	dev_info(&dev->pdev->dev, "VSS = %d\n", mode->vsync_start);
+	dev_info(&dev->pdev->dev, "VSE = %d\n", mode->vsync_end);
+	dev_info(&dev->pdev->dev, "vtotal = %d\n", mode->vtotal);
+	dev_info(&dev->pdev->dev, "clock = %d\n", mode->clock);
+
+	drm_mode_set_name(mode);
+	drm_mode_set_crtcinfo(mode, 0);
+
+	mode->type |= DRM_MODE_TYPE_PREFERRED;
+
+	return mode;
+}
+
+/* DV1 Active area 216.96 x 135.6 mm */
+#define DV1_PANEL_WIDTH 217
+#define DV1_PANEL_HEIGHT 136
+
+static int tc35876x_get_panel_info(struct drm_device *dev, int pipe,
+				struct panel_info *pi)
+{
+	if (!dev || !pi)
+		return -EINVAL;
+
+	pi->width_mm = DV1_PANEL_WIDTH;
+	pi->height_mm = DV1_PANEL_HEIGHT;
+
+	return 0;
+}
+
+static int tc35876x_bridge_probe(struct i2c_client *client,
+				const struct i2c_device_id *id)
+{
+	struct tc35876x_platform_data *pdata;
+
+	dev_info(&client->dev, "%s\n", __func__);
+
+	if (!i2c_check_functionality(client->adapter, I2C_FUNC_I2C)) {
+		dev_err(&client->dev, "%s: i2c_check_functionality() failed\n",
+			__func__);
+		return -ENODEV;
+	}
+
+	pdata = dev_get_platdata(&client->dev);
+	if (!pdata) {
+		dev_err(&client->dev, "%s: no platform data\n", __func__);
+		return -ENODEV;
+	}
+
+	if (pdata->gpio_bridge_reset != -1) {
+		gpio_request(pdata->gpio_bridge_reset, "tc35876x bridge reset");
+		gpio_direction_output(pdata->gpio_bridge_reset, 0);
+	}
+
+	if (pdata->gpio_panel_bl_en != -1) {
+		gpio_request(pdata->gpio_panel_bl_en, "tc35876x panel bl en");
+		gpio_direction_output(pdata->gpio_panel_bl_en, 0);
+	}
+
+	if (pdata->gpio_panel_vadd != -1) {
+		gpio_request(pdata->gpio_panel_vadd, "tc35876x panel vadd");
+		gpio_direction_output(pdata->gpio_panel_vadd, 0);
+	}
+
+	tc35876x_client = client;
+
+	return 0;
+}
+
+static int tc35876x_bridge_remove(struct i2c_client *client)
+{
+	struct tc35876x_platform_data *pdata = dev_get_platdata(&client->dev);
+
+	dev_dbg(&client->dev, "%s\n", __func__);
+
+	if (pdata->gpio_bridge_reset != -1)
+		gpio_free(pdata->gpio_bridge_reset);
+
+	if (pdata->gpio_panel_bl_en != -1)
+		gpio_free(pdata->gpio_panel_bl_en);
+
+	if (pdata->gpio_panel_vadd != -1)
+		gpio_free(pdata->gpio_panel_vadd);
+
+	tc35876x_client = NULL;
+
+	return 0;
+}
+
+static const struct i2c_device_id tc35876x_bridge_id[] = {
+	{ "i2c_disp_brig", 0 },
+	{ }
+};
+MODULE_DEVICE_TABLE(i2c, tc35876x_bridge_id);
+
+static struct i2c_driver tc35876x_bridge_i2c_driver = {
+	.driver = {
+		.name = "i2c_disp_brig",
+	},
+	.id_table = tc35876x_bridge_id,
+	.probe = tc35876x_bridge_probe,
+	.remove = __devexit_p(tc35876x_bridge_remove),
+};
+
+/* LCD panel I2C */
+static int cmi_lcd_i2c_probe(struct i2c_client *client,
+			     const struct i2c_device_id *id)
+{
+	dev_info(&client->dev, "%s\n", __func__);
+
+	if (!i2c_check_functionality(client->adapter, I2C_FUNC_I2C)) {
+		dev_err(&client->dev, "%s: i2c_check_functionality() failed\n",
+			__func__);
+		return -ENODEV;
+	}
+
+	cmi_lcd_i2c_client = client;
+
+	return 0;
+}
+
+static int cmi_lcd_i2c_remove(struct i2c_client *client)
+{
+	dev_dbg(&client->dev, "%s\n", __func__);
+
+	cmi_lcd_i2c_client = NULL;
+
+	return 0;
+}
+
+static const struct i2c_device_id cmi_lcd_i2c_id[] = {
+	{ "cmi-lcd", 0 },
+	{ }
+};
+MODULE_DEVICE_TABLE(i2c, cmi_lcd_i2c_id);
+
+static struct i2c_driver cmi_lcd_i2c_driver = {
+	.driver = {
+		.name = "cmi-lcd",
+	},
+	.id_table = cmi_lcd_i2c_id,
+	.probe = cmi_lcd_i2c_probe,
+	.remove = __devexit_p(cmi_lcd_i2c_remove),
+};
+
+/* HACK to create I2C device while it's not created by platform code */
+#define CMI_LCD_I2C_ADAPTER	2
+#define CMI_LCD_I2C_ADDR	0x60
+
+static int cmi_lcd_hack_create_device(void)
+{
+	struct i2c_adapter *adapter;
+	struct i2c_client *client;
+	struct i2c_board_info info = {
+		.type = "cmi-lcd",
+		.addr = CMI_LCD_I2C_ADDR,
+	};
+
+	pr_debug("%s\n", __func__);
+
+	adapter = i2c_get_adapter(CMI_LCD_I2C_ADAPTER);
+	if (!adapter) {
+		pr_err("%s: i2c_get_adapter(%d) failed\n", __func__,
+			CMI_LCD_I2C_ADAPTER);
+		return -EINVAL;
+	}
+
+	client = i2c_new_device(adapter, &info);
+	if (!client) {
+		pr_err("%s: i2c_new_device() failed\n", __func__);
+		i2c_put_adapter(adapter);
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static const struct drm_encoder_helper_funcs tc35876x_encoder_helper_funcs = {
+	.dpms = mdfld_dsi_dpi_dpms,
+	.mode_fixup = mdfld_dsi_dpi_mode_fixup,
+	.prepare = mdfld_dsi_dpi_prepare,
+	.mode_set = mdfld_dsi_dpi_mode_set,
+	.commit = mdfld_dsi_dpi_commit,
+};
+
+static const struct drm_encoder_funcs tc35876x_encoder_funcs = {
+	.destroy = drm_encoder_cleanup,
+};
+
+const struct panel_funcs mdfld_tc35876x_funcs = {
+	.encoder_funcs = &tc35876x_encoder_funcs,
+	.encoder_helper_funcs = &tc35876x_encoder_helper_funcs,
+	.get_config_mode = tc35876x_get_config_mode,
+	.get_panel_info = tc35876x_get_panel_info,
+};
+
+void tc35876x_init(struct drm_device *dev)
+{
+	int r;
+
+	dev_dbg(&dev->pdev->dev, "%s\n", __func__);
+
+	cmi_lcd_hack_create_device();
+
+	r = i2c_add_driver(&cmi_lcd_i2c_driver);
+	if (r < 0)
+		dev_err(&dev->pdev->dev,
+			"%s: i2c_add_driver() for %s failed (%d)\n",
+			__func__, cmi_lcd_i2c_driver.driver.name, r);
+
+	r = i2c_add_driver(&tc35876x_bridge_i2c_driver);
+	if (r < 0)
+		dev_err(&dev->pdev->dev,
+			"%s: i2c_add_driver() for %s failed (%d)\n",
+			__func__, tc35876x_bridge_i2c_driver.driver.name, r);
+
+	tc35876x_brightness_init(dev);
+}
+
+void tc35876x_exit(void)
+{
+	pr_debug("%s\n", __func__);
+
+	i2c_del_driver(&tc35876x_bridge_i2c_driver);
+
+	if (cmi_lcd_i2c_client)
+		i2c_del_driver(&cmi_lcd_i2c_driver);
+}
diff --git a/drivers/gpu/drm/gma500/tc35876x-dsi-lvds.h b/drivers/gpu/drm/gma500/tc35876x-dsi-lvds.h
new file mode 100644
index 000000000000..b14b7f9e7d1e
--- /dev/null
+++ b/drivers/gpu/drm/gma500/tc35876x-dsi-lvds.h
@@ -0,0 +1,38 @@
+/*
+ * Copyright © 2011 Intel Corporation
+ *
+ * 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.
+ *
+ */
+
+#ifndef __MDFLD_DSI_LVDS_BRIDGE_H__
+#define __MDFLD_DSI_LVDS_BRIDGE_H__
+
+void tc35876x_set_bridge_reset_state(struct drm_device *dev, int state);
+void tc35876x_configure_lvds_bridge(struct drm_device *dev);
+void tc35876x_brightness_control(struct drm_device *dev, int level);
+void tc35876x_toshiba_bridge_panel_off(struct drm_device *dev);
+void tc35876x_toshiba_bridge_panel_on(struct drm_device *dev);
+void tc35876x_init(struct drm_device *dev);
+void tc35876x_exit(void);
+
+extern const struct panel_funcs mdfld_tc35876x_funcs;
+
+#endif /*__MDFLD_DSI_LVDS_BRIDGE_H__*/
diff --git a/drivers/gpu/drm/i2c/ch7006_drv.c b/drivers/gpu/drm/i2c/ch7006_drv.c
index 07d55df6623e..d3f2e8785010 100644
--- a/drivers/gpu/drm/i2c/ch7006_drv.c
+++ b/drivers/gpu/drm/i2c/ch7006_drv.c
@@ -252,10 +252,7 @@ static int ch7006_encoder_create_resources(struct drm_encoder *encoder,
 
 	drm_mode_create_tv_properties(dev, NUM_TV_NORMS, ch7006_tv_norm_names);
 
-	priv->scale_property = drm_property_create(dev, DRM_MODE_PROP_RANGE,
-						   "scale", 2);
-	priv->scale_property->values[0] = 0;
-	priv->scale_property->values[1] = 2;
+	priv->scale_property = drm_property_create_range(dev, 0, "scale", 0, 2);
 
 	drm_connector_attach_property(connector, conf->tv_select_subconnector_property,
 				      priv->select_subconnector);
diff --git a/drivers/gpu/drm/i810/i810_dma.c b/drivers/gpu/drm/i810/i810_dma.c
index 7f4b4e10246e..2c8a60c3b98e 100644
--- a/drivers/gpu/drm/i810/i810_dma.c
+++ b/drivers/gpu/drm/i810/i810_dma.c
@@ -99,7 +99,6 @@ static int i810_mmap_buffers(struct file *filp, struct vm_area_struct *vma)
 	buf_priv = buf->dev_private;
 
 	vma->vm_flags |= (VM_IO | VM_DONTCOPY);
-	vma->vm_file = filp;
 
 	buf_priv->currently_mapped = I810_BUF_MAPPED;
 
@@ -1208,6 +1207,8 @@ int i810_driver_load(struct drm_device *dev, unsigned long flags)
 	dev->types[8] = _DRM_STAT_SECONDARY;
 	dev->types[9] = _DRM_STAT_DMA;
 
+	pci_set_master(dev->pdev);
+
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/i915/Makefile b/drivers/gpu/drm/i915/Makefile
index 808b255d7fc6..ce7fc77678b4 100644
--- a/drivers/gpu/drm/i915/Makefile
+++ b/drivers/gpu/drm/i915/Makefile
@@ -3,7 +3,7 @@
 # Direct Rendering Infrastructure (DRI) in XFree86 4.1.0 and higher.
 
 ccflags-y := -Iinclude/drm
-i915-y := i915_drv.o i915_dma.o i915_irq.o i915_mem.o \
+i915-y := i915_drv.o i915_dma.o i915_irq.o \
 	  i915_debugfs.o \
           i915_suspend.o \
 	  i915_gem.o \
diff --git a/drivers/gpu/drm/i915/i915_debugfs.c b/drivers/gpu/drm/i915/i915_debugfs.c
index deaa657292b4..b505b70dba05 100644
--- a/drivers/gpu/drm/i915/i915_debugfs.c
+++ b/drivers/gpu/drm/i915/i915_debugfs.c
@@ -83,6 +83,7 @@ static int i915_capabilities(struct seq_file *m, void *data)
 	B(supports_tv);
 	B(has_bsd_ring);
 	B(has_blt_ring);
+	B(has_llc);
 #undef B
 
 	return 0;
@@ -563,45 +564,6 @@ static int i915_hws_info(struct seq_file *m, void *data)
 	return 0;
 }
 
-static void i915_dump_object(struct seq_file *m,
-			     struct io_mapping *mapping,
-			     struct drm_i915_gem_object *obj)
-{
-	int page, page_count, i;
-
-	page_count = obj->base.size / PAGE_SIZE;
-	for (page = 0; page < page_count; page++) {
-		u32 *mem = io_mapping_map_wc(mapping,
-					     obj->gtt_offset + page * PAGE_SIZE);
-		for (i = 0; i < PAGE_SIZE; i += 4)
-			seq_printf(m, "%08x :  %08x\n", i, mem[i / 4]);
-		io_mapping_unmap(mem);
-	}
-}
-
-static int i915_batchbuffer_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;
-	drm_i915_private_t *dev_priv = dev->dev_private;
-	struct drm_i915_gem_object *obj;
-	int ret;
-
-	ret = mutex_lock_interruptible(&dev->struct_mutex);
-	if (ret)
-		return ret;
-
-	list_for_each_entry(obj, &dev_priv->mm.active_list, mm_list) {
-		if (obj->base.read_domains & I915_GEM_DOMAIN_COMMAND) {
-		    seq_printf(m, "--- gtt_offset = 0x%08x\n", obj->gtt_offset);
-		    i915_dump_object(m, dev_priv->mm.gtt_mapping, obj);
-		}
-	}
-
-	mutex_unlock(&dev->struct_mutex);
-	return 0;
-}
-
 static int i915_ringbuffer_data(struct seq_file *m, void *data)
 {
 	struct drm_info_node *node = (struct drm_info_node *) m->private;
@@ -668,9 +630,9 @@ static int i915_ringbuffer_info(struct seq_file *m, void *data)
 static const char *ring_str(int ring)
 {
 	switch (ring) {
-	case RING_RENDER: return " render";
-	case RING_BSD: return " bsd";
-	case RING_BLT: return " blt";
+	case RCS: return "render";
+	case VCS: return "bsd";
+	case BCS: return "blt";
 	default: return "";
 	}
 }
@@ -713,7 +675,7 @@ static void print_error_buffers(struct seq_file *m,
 	seq_printf(m, "%s [%d]:\n", name, count);
 
 	while (count--) {
-		seq_printf(m, "  %08x %8u %04x %04x %08x%s%s%s%s%s%s",
+		seq_printf(m, "  %08x %8u %04x %04x %08x%s%s%s%s%s%s%s",
 			   err->gtt_offset,
 			   err->size,
 			   err->read_domains,
@@ -723,6 +685,7 @@ static void print_error_buffers(struct seq_file *m,
 			   tiling_flag(err->tiling),
 			   dirty_flag(err->dirty),
 			   purgeable_flag(err->purgeable),
+			   err->ring != -1 ? " " : "",
 			   ring_str(err->ring),
 			   cache_level_str(err->cache_level));
 
@@ -736,6 +699,38 @@ static void print_error_buffers(struct seq_file *m,
 	}
 }
 
+static void i915_ring_error_state(struct seq_file *m,
+				  struct drm_device *dev,
+				  struct drm_i915_error_state *error,
+				  unsigned ring)
+{
+	seq_printf(m, "%s command stream:\n", ring_str(ring));
+	seq_printf(m, "  HEAD: 0x%08x\n", error->head[ring]);
+	seq_printf(m, "  TAIL: 0x%08x\n", error->tail[ring]);
+	seq_printf(m, "  ACTHD: 0x%08x\n", error->acthd[ring]);
+	seq_printf(m, "  IPEIR: 0x%08x\n", error->ipeir[ring]);
+	seq_printf(m, "  IPEHR: 0x%08x\n", error->ipehr[ring]);
+	seq_printf(m, "  INSTDONE: 0x%08x\n", error->instdone[ring]);
+	if (ring == RCS && INTEL_INFO(dev)->gen >= 4) {
+		seq_printf(m, "  INSTDONE1: 0x%08x\n", error->instdone1);
+		seq_printf(m, "  BBADDR: 0x%08llx\n", error->bbaddr);
+	}
+	if (INTEL_INFO(dev)->gen >= 4)
+		seq_printf(m, "  INSTPS: 0x%08x\n", error->instps[ring]);
+	seq_printf(m, "  INSTPM: 0x%08x\n", error->instpm[ring]);
+	if (INTEL_INFO(dev)->gen >= 6) {
+		seq_printf(m, "  FADDR: 0x%08x\n", error->faddr[ring]);
+		seq_printf(m, "  FAULT_REG: 0x%08x\n", error->fault_reg[ring]);
+		seq_printf(m, "  SYNC_0: 0x%08x\n",
+			   error->semaphore_mboxes[ring][0]);
+		seq_printf(m, "  SYNC_1: 0x%08x\n",
+			   error->semaphore_mboxes[ring][1]);
+	}
+	seq_printf(m, "  seqno: 0x%08x\n", error->seqno[ring]);
+	seq_printf(m, "  ring->head: 0x%08x\n", error->cpu_ring_head[ring]);
+	seq_printf(m, "  ring->tail: 0x%08x\n", error->cpu_ring_tail[ring]);
+}
+
 static int i915_error_state(struct seq_file *m, void *unused)
 {
 	struct drm_info_node *node = (struct drm_info_node *) m->private;
@@ -743,7 +738,7 @@ static int i915_error_state(struct seq_file *m, void *unused)
 	drm_i915_private_t *dev_priv = dev->dev_private;
 	struct drm_i915_error_state *error;
 	unsigned long flags;
-	int i, page, offset, elt;
+	int i, j, page, offset, elt;
 
 	spin_lock_irqsave(&dev_priv->error_lock, flags);
 	if (!dev_priv->first_error) {
@@ -758,35 +753,20 @@ static int i915_error_state(struct seq_file *m, void *unused)
 	seq_printf(m, "PCI ID: 0x%04x\n", dev->pci_device);
 	seq_printf(m, "EIR: 0x%08x\n", error->eir);
 	seq_printf(m, "PGTBL_ER: 0x%08x\n", error->pgtbl_er);
+
+	for (i = 0; i < dev_priv->num_fence_regs; i++)
+		seq_printf(m, "  fence[%d] = %08llx\n", i, error->fence[i]);
+
 	if (INTEL_INFO(dev)->gen >= 6) {
 		seq_printf(m, "ERROR: 0x%08x\n", error->error);
-		seq_printf(m, "Blitter command stream:\n");
-		seq_printf(m, "  ACTHD:    0x%08x\n", error->bcs_acthd);
-		seq_printf(m, "  IPEIR:    0x%08x\n", error->bcs_ipeir);
-		seq_printf(m, "  IPEHR:    0x%08x\n", error->bcs_ipehr);
-		seq_printf(m, "  INSTDONE: 0x%08x\n", error->bcs_instdone);
-		seq_printf(m, "  seqno:    0x%08x\n", error->bcs_seqno);
-		seq_printf(m, "Video (BSD) command stream:\n");
-		seq_printf(m, "  ACTHD:    0x%08x\n", error->vcs_acthd);
-		seq_printf(m, "  IPEIR:    0x%08x\n", error->vcs_ipeir);
-		seq_printf(m, "  IPEHR:    0x%08x\n", error->vcs_ipehr);
-		seq_printf(m, "  INSTDONE: 0x%08x\n", error->vcs_instdone);
-		seq_printf(m, "  seqno:    0x%08x\n", error->vcs_seqno);
-	}
-	seq_printf(m, "Render command stream:\n");
-	seq_printf(m, "  ACTHD: 0x%08x\n", error->acthd);
-	seq_printf(m, "  IPEIR: 0x%08x\n", error->ipeir);
-	seq_printf(m, "  IPEHR: 0x%08x\n", error->ipehr);
-	seq_printf(m, "  INSTDONE: 0x%08x\n", error->instdone);
-	if (INTEL_INFO(dev)->gen >= 4) {
-		seq_printf(m, "  INSTDONE1: 0x%08x\n", error->instdone1);
-		seq_printf(m, "  INSTPS: 0x%08x\n", error->instps);
+		seq_printf(m, "DONE_REG: 0x%08x\n", error->done_reg);
 	}
-	seq_printf(m, "  INSTPM: 0x%08x\n", error->instpm);
-	seq_printf(m, "  seqno: 0x%08x\n", error->seqno);
 
-	for (i = 0; i < dev_priv->num_fence_regs; i++)
-		seq_printf(m, "  fence[%d] = %08llx\n", i, error->fence[i]);
+	i915_ring_error_state(m, dev, error, RCS);
+	if (HAS_BLT(dev))
+		i915_ring_error_state(m, dev, error, BCS);
+	if (HAS_BSD(dev))
+		i915_ring_error_state(m, dev, error, VCS);
 
 	if (error->active_bo)
 		print_error_buffers(m, "Active",
@@ -798,10 +778,10 @@ static int i915_error_state(struct seq_file *m, void *unused)
 				    error->pinned_bo,
 				    error->pinned_bo_count);
 
-	for (i = 0; i < ARRAY_SIZE(error->batchbuffer); i++) {
-		if (error->batchbuffer[i]) {
-			struct drm_i915_error_object *obj = error->batchbuffer[i];
+	for (i = 0; i < ARRAY_SIZE(error->ring); i++) {
+		struct drm_i915_error_object *obj;
 
+		if ((obj = error->ring[i].batchbuffer)) {
 			seq_printf(m, "%s --- gtt_offset = 0x%08x\n",
 				   dev_priv->ring[i].name,
 				   obj->gtt_offset);
@@ -813,11 +793,20 @@ static int i915_error_state(struct seq_file *m, void *unused)
 				}
 			}
 		}
-	}
 
-	for (i = 0; i < ARRAY_SIZE(error->ringbuffer); i++) {
-		if (error->ringbuffer[i]) {
-			struct drm_i915_error_object *obj = error->ringbuffer[i];
+		if (error->ring[i].num_requests) {
+			seq_printf(m, "%s --- %d requests\n",
+				   dev_priv->ring[i].name,
+				   error->ring[i].num_requests);
+			for (j = 0; j < error->ring[i].num_requests; j++) {
+				seq_printf(m, "  seqno 0x%08x, emitted %ld, tail 0x%08x\n",
+					   error->ring[i].requests[j].seqno,
+					   error->ring[i].requests[j].jiffies,
+					   error->ring[i].requests[j].tail);
+			}
+		}
+
+		if ((obj = error->ring[i].ringbuffer)) {
 			seq_printf(m, "%s --- ringbuffer = 0x%08x\n",
 				   dev_priv->ring[i].name,
 				   obj->gtt_offset);
@@ -1414,11 +1403,102 @@ static int i915_gen6_forcewake_count_info(struct seq_file *m, void *data)
 	return 0;
 }
 
-static int
-i915_wedged_open(struct inode *inode,
-		 struct file *filp)
+static const char *swizzle_string(unsigned swizzle)
+{
+	switch(swizzle) {
+	case I915_BIT_6_SWIZZLE_NONE:
+		return "none";
+	case I915_BIT_6_SWIZZLE_9:
+		return "bit9";
+	case I915_BIT_6_SWIZZLE_9_10:
+		return "bit9/bit10";
+	case I915_BIT_6_SWIZZLE_9_11:
+		return "bit9/bit11";
+	case I915_BIT_6_SWIZZLE_9_10_11:
+		return "bit9/bit10/bit11";
+	case I915_BIT_6_SWIZZLE_9_17:
+		return "bit9/bit17";
+	case I915_BIT_6_SWIZZLE_9_10_17:
+		return "bit9/bit10/bit17";
+	case I915_BIT_6_SWIZZLE_UNKNOWN:
+		return "unkown";
+	}
+
+	return "bug";
+}
+
+static int i915_swizzle_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 drm_i915_private *dev_priv = dev->dev_private;
+
+	mutex_lock(&dev->struct_mutex);
+	seq_printf(m, "bit6 swizzle for X-tiling = %s\n",
+		   swizzle_string(dev_priv->mm.bit_6_swizzle_x));
+	seq_printf(m, "bit6 swizzle for Y-tiling = %s\n",
+		   swizzle_string(dev_priv->mm.bit_6_swizzle_y));
+
+	if (IS_GEN3(dev) || IS_GEN4(dev)) {
+		seq_printf(m, "DDC = 0x%08x\n",
+			   I915_READ(DCC));
+		seq_printf(m, "C0DRB3 = 0x%04x\n",
+			   I915_READ16(C0DRB3));
+		seq_printf(m, "C1DRB3 = 0x%04x\n",
+			   I915_READ16(C1DRB3));
+	} else if (IS_GEN6(dev) || IS_GEN7(dev)) {
+		seq_printf(m, "MAD_DIMM_C0 = 0x%08x\n",
+			   I915_READ(MAD_DIMM_C0));
+		seq_printf(m, "MAD_DIMM_C1 = 0x%08x\n",
+			   I915_READ(MAD_DIMM_C1));
+		seq_printf(m, "MAD_DIMM_C2 = 0x%08x\n",
+			   I915_READ(MAD_DIMM_C2));
+		seq_printf(m, "TILECTL = 0x%08x\n",
+			   I915_READ(TILECTL));
+		seq_printf(m, "ARB_MODE = 0x%08x\n",
+			   I915_READ(ARB_MODE));
+		seq_printf(m, "DISP_ARB_CTL = 0x%08x\n",
+			   I915_READ(DISP_ARB_CTL));
+	}
+	mutex_unlock(&dev->struct_mutex);
+
+	return 0;
+}
+
+static int i915_ppgtt_info(struct seq_file *m, void *data)
 {
-	filp->private_data = inode->i_private;
+	struct drm_info_node *node = (struct drm_info_node *) m->private;
+	struct drm_device *dev = node->minor->dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct intel_ring_buffer *ring;
+	int i, ret;
+
+
+	ret = mutex_lock_interruptible(&dev->struct_mutex);
+	if (ret)
+		return ret;
+	if (INTEL_INFO(dev)->gen == 6)
+		seq_printf(m, "GFX_MODE: 0x%08x\n", I915_READ(GFX_MODE));
+
+	for (i = 0; i < I915_NUM_RINGS; i++) {
+		ring = &dev_priv->ring[i];
+
+		seq_printf(m, "%s\n", ring->name);
+		if (INTEL_INFO(dev)->gen == 7)
+			seq_printf(m, "GFX_MODE: 0x%08x\n", I915_READ(RING_MODE_GEN7(ring)));
+		seq_printf(m, "PP_DIR_BASE: 0x%08x\n", I915_READ(RING_PP_DIR_BASE(ring)));
+		seq_printf(m, "PP_DIR_BASE_READ: 0x%08x\n", I915_READ(RING_PP_DIR_BASE_READ(ring)));
+		seq_printf(m, "PP_DIR_DCLV: 0x%08x\n", I915_READ(RING_PP_DIR_DCLV(ring)));
+	}
+	if (dev_priv->mm.aliasing_ppgtt) {
+		struct i915_hw_ppgtt *ppgtt = dev_priv->mm.aliasing_ppgtt;
+
+		seq_printf(m, "aliasing PPGTT:\n");
+		seq_printf(m, "pd gtt offset: 0x%08x\n", ppgtt->pd_offset);
+	}
+	seq_printf(m, "ECOCHK: 0x%08x\n", I915_READ(GAM_ECOCHK));
+	mutex_unlock(&dev->struct_mutex);
+
 	return 0;
 }
 
@@ -1472,20 +1552,12 @@ i915_wedged_write(struct file *filp,
 
 static const struct file_operations i915_wedged_fops = {
 	.owner = THIS_MODULE,
-	.open = i915_wedged_open,
+	.open = simple_open,
 	.read = i915_wedged_read,
 	.write = i915_wedged_write,
 	.llseek = default_llseek,
 };
 
-static int
-i915_max_freq_open(struct inode *inode,
-		   struct file *filp)
-{
-	filp->private_data = inode->i_private;
-	return 0;
-}
-
 static ssize_t
 i915_max_freq_read(struct file *filp,
 		   char __user *ubuf,
@@ -1542,20 +1614,12 @@ i915_max_freq_write(struct file *filp,
 
 static const struct file_operations i915_max_freq_fops = {
 	.owner = THIS_MODULE,
-	.open = i915_max_freq_open,
+	.open = simple_open,
 	.read = i915_max_freq_read,
 	.write = i915_max_freq_write,
 	.llseek = default_llseek,
 };
 
-static int
-i915_cache_sharing_open(struct inode *inode,
-		   struct file *filp)
-{
-	filp->private_data = inode->i_private;
-	return 0;
-}
-
 static ssize_t
 i915_cache_sharing_read(struct file *filp,
 		   char __user *ubuf,
@@ -1621,7 +1685,7 @@ i915_cache_sharing_write(struct file *filp,
 
 static const struct file_operations i915_cache_sharing_fops = {
 	.owner = THIS_MODULE,
-	.open = i915_cache_sharing_open,
+	.open = simple_open,
 	.read = i915_cache_sharing_read,
 	.write = i915_cache_sharing_write,
 	.llseek = default_llseek,
@@ -1653,21 +1717,6 @@ drm_add_fake_info_node(struct drm_minor *minor,
 	return 0;
 }
 
-static int i915_wedged_create(struct dentry *root, struct drm_minor *minor)
-{
-	struct drm_device *dev = minor->dev;
-	struct dentry *ent;
-
-	ent = debugfs_create_file("i915_wedged",
-				  S_IRUGO | S_IWUSR,
-				  root, dev,
-				  &i915_wedged_fops);
-	if (IS_ERR(ent))
-		return PTR_ERR(ent);
-
-	return drm_add_fake_info_node(minor, ent, &i915_wedged_fops);
-}
-
 static int i915_forcewake_open(struct inode *inode, struct file *file)
 {
 	struct drm_device *dev = inode->i_private;
@@ -1729,34 +1778,22 @@ static int i915_forcewake_create(struct dentry *root, struct drm_minor *minor)
 	return drm_add_fake_info_node(minor, ent, &i915_forcewake_fops);
 }
 
-static int i915_max_freq_create(struct dentry *root, struct drm_minor *minor)
+static int i915_debugfs_create(struct dentry *root,
+			       struct drm_minor *minor,
+			       const char *name,
+			       const struct file_operations *fops)
 {
 	struct drm_device *dev = minor->dev;
 	struct dentry *ent;
 
-	ent = debugfs_create_file("i915_max_freq",
+	ent = debugfs_create_file(name,
 				  S_IRUGO | S_IWUSR,
 				  root, dev,
-				  &i915_max_freq_fops);
+				  fops);
 	if (IS_ERR(ent))
 		return PTR_ERR(ent);
 
-	return drm_add_fake_info_node(minor, ent, &i915_max_freq_fops);
-}
-
-static int i915_cache_sharing_create(struct dentry *root, struct drm_minor *minor)
-{
-	struct drm_device *dev = minor->dev;
-	struct dentry *ent;
-
-	ent = debugfs_create_file("i915_cache_sharing",
-				  S_IRUGO | S_IWUSR,
-				  root, dev,
-				  &i915_cache_sharing_fops);
-	if (IS_ERR(ent))
-		return PTR_ERR(ent);
-
-	return drm_add_fake_info_node(minor, ent, &i915_cache_sharing_fops);
+	return drm_add_fake_info_node(minor, ent, fops);
 }
 
 static struct drm_info_list i915_debugfs_list[] = {
@@ -1782,7 +1819,6 @@ static struct drm_info_list i915_debugfs_list[] = {
 	{"i915_bsd_ringbuffer_info", i915_ringbuffer_info, 0, (void *)VCS},
 	{"i915_blt_ringbuffer_data", i915_ringbuffer_data, 0, (void *)BCS},
 	{"i915_blt_ringbuffer_info", i915_ringbuffer_info, 0, (void *)BCS},
-	{"i915_batchbuffers", i915_batchbuffer_info, 0},
 	{"i915_error_state", i915_error_state, 0},
 	{"i915_rstdby_delays", i915_rstdby_delays, 0},
 	{"i915_cur_delayinfo", i915_cur_delayinfo, 0},
@@ -1798,6 +1834,8 @@ static struct drm_info_list i915_debugfs_list[] = {
 	{"i915_gem_framebuffer", i915_gem_framebuffer_info, 0},
 	{"i915_context_status", i915_context_status, 0},
 	{"i915_gen6_forcewake_count", i915_gen6_forcewake_count_info, 0},
+	{"i915_swizzle_info", i915_swizzle_info, 0},
+	{"i915_ppgtt_info", i915_ppgtt_info, 0},
 };
 #define I915_DEBUGFS_ENTRIES ARRAY_SIZE(i915_debugfs_list)
 
@@ -1805,17 +1843,25 @@ int i915_debugfs_init(struct drm_minor *minor)
 {
 	int ret;
 
-	ret = i915_wedged_create(minor->debugfs_root, minor);
+	ret = i915_debugfs_create(minor->debugfs_root, minor,
+				  "i915_wedged",
+				  &i915_wedged_fops);
 	if (ret)
 		return ret;
 
 	ret = i915_forcewake_create(minor->debugfs_root, minor);
 	if (ret)
 		return ret;
-	ret = i915_max_freq_create(minor->debugfs_root, minor);
+
+	ret = i915_debugfs_create(minor->debugfs_root, minor,
+				  "i915_max_freq",
+				  &i915_max_freq_fops);
 	if (ret)
 		return ret;
-	ret = i915_cache_sharing_create(minor->debugfs_root, minor);
+
+	ret = i915_debugfs_create(minor->debugfs_root, minor,
+				  "i915_cache_sharing",
+				  &i915_cache_sharing_fops);
 	if (ret)
 		return ret;
 
diff --git a/drivers/gpu/drm/i915/i915_dma.c b/drivers/gpu/drm/i915/i915_dma.c
index ddfe3d902b2a..785f67f963ef 100644
--- a/drivers/gpu/drm/i915/i915_dma.c
+++ b/drivers/gpu/drm/i915/i915_dma.c
@@ -784,6 +784,9 @@ static int i915_getparam(struct drm_device *dev, void *data,
 	case I915_PARAM_HAS_GEN7_SOL_RESET:
 		value = 1;
 		break;
+	case I915_PARAM_HAS_LLC:
+		value = HAS_LLC(dev);
+		break;
 	default:
 		DRM_DEBUG_DRIVER("Unknown parameter %d\n",
 				 param->param);
@@ -1180,6 +1183,21 @@ static bool i915_switcheroo_can_switch(struct pci_dev *pdev)
 	return can_switch;
 }
 
+static bool
+intel_enable_ppgtt(struct drm_device *dev)
+{
+	if (i915_enable_ppgtt >= 0)
+		return i915_enable_ppgtt;
+
+#ifdef CONFIG_INTEL_IOMMU
+	/* Disable ppgtt on SNB if VT-d is on. */
+	if (INTEL_INFO(dev)->gen == 6 && intel_iommu_gfx_mapped)
+		return false;
+#endif
+
+	return true;
+}
+
 static int i915_load_gem_init(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
@@ -1193,22 +1211,41 @@ static int i915_load_gem_init(struct drm_device *dev)
 	/* Basic memrange allocator for stolen space */
 	drm_mm_init(&dev_priv->mm.stolen, 0, prealloc_size);
 
-	/* Let GEM Manage all of the aperture.
-	 *
-	 * However, leave one page at the end still bound to the scratch page.
-	 * There are a number of places where the hardware apparently
-	 * prefetches past the end of the object, and we've seen multiple
-	 * hangs with the GPU head pointer stuck in a batchbuffer bound
-	 * at the last page of the aperture.  One page should be enough to
-	 * keep any prefetching inside of the aperture.
-	 */
-	i915_gem_do_init(dev, 0, mappable_size, gtt_size - PAGE_SIZE);
-
 	mutex_lock(&dev->struct_mutex);
-	ret = i915_gem_init_ringbuffer(dev);
+	if (intel_enable_ppgtt(dev) && HAS_ALIASING_PPGTT(dev)) {
+		/* 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;
+		/* For paranoia keep the guard page in between. */
+		gtt_size -= PAGE_SIZE;
+
+		i915_gem_do_init(dev, 0, mappable_size, gtt_size);
+
+		ret = i915_gem_init_aliasing_ppgtt(dev);
+		if (ret) {
+			mutex_unlock(&dev->struct_mutex);
+			return ret;
+		}
+	} else {
+		/* Let GEM Manage all of the aperture.
+		 *
+		 * However, leave one page at the end still bound to the scratch
+		 * page.  There are a number of places where the hardware
+		 * apparently prefetches past the end of the object, and we've
+		 * seen multiple hangs with the GPU head pointer stuck in a
+		 * batchbuffer bound at the last page of the aperture.  One page
+		 * should be enough to keep any prefetching inside of the
+		 * aperture.
+		 */
+		i915_gem_do_init(dev, 0, mappable_size, gtt_size - PAGE_SIZE);
+	}
+
+	ret = i915_gem_init_hw(dev);
 	mutex_unlock(&dev->struct_mutex);
-	if (ret)
+	if (ret) {
+		i915_gem_cleanup_aliasing_ppgtt(dev);
 		return ret;
+	}
 
 	/* Try to set up FBC with a reasonable compressed buffer size */
 	if (I915_HAS_FBC(dev) && i915_powersave) {
@@ -1295,6 +1332,7 @@ cleanup_gem:
 	mutex_lock(&dev->struct_mutex);
 	i915_gem_cleanup_ringbuffer(dev);
 	mutex_unlock(&dev->struct_mutex);
+	i915_gem_cleanup_aliasing_ppgtt(dev);
 cleanup_vga_switcheroo:
 	vga_switcheroo_unregister_client(dev->pdev);
 cleanup_vga_client:
@@ -1930,6 +1968,8 @@ int i915_driver_load(struct drm_device *dev, unsigned long flags)
 		goto free_priv;
 	}
 
+	pci_set_master(dev->pdev);
+
 	/* overlay on gen2 is broken and can't address above 1G */
 	if (IS_GEN2(dev))
 		dma_set_coherent_mask(&dev->pdev->dev, DMA_BIT_MASK(30));
@@ -2129,7 +2169,7 @@ int i915_driver_unload(struct drm_device *dev)
 		unregister_shrinker(&dev_priv->mm.inactive_shrinker);
 
 	mutex_lock(&dev->struct_mutex);
-	ret = i915_gpu_idle(dev);
+	ret = i915_gpu_idle(dev, true);
 	if (ret)
 		DRM_ERROR("failed to idle hardware: %d\n", ret);
 	mutex_unlock(&dev->struct_mutex);
@@ -2182,6 +2222,7 @@ int i915_driver_unload(struct drm_device *dev)
 		i915_gem_free_all_phys_object(dev);
 		i915_gem_cleanup_ringbuffer(dev);
 		mutex_unlock(&dev->struct_mutex);
+		i915_gem_cleanup_aliasing_ppgtt(dev);
 		if (I915_HAS_FBC(dev) && i915_powersave)
 			i915_cleanup_compression(dev);
 		drm_mm_takedown(&dev_priv->mm.stolen);
@@ -2247,18 +2288,12 @@ void i915_driver_lastclose(struct drm_device * dev)
 
 	i915_gem_lastclose(dev);
 
-	if (dev_priv->agp_heap)
-		i915_mem_takedown(&(dev_priv->agp_heap));
-
 	i915_dma_cleanup(dev);
 }
 
 void i915_driver_preclose(struct drm_device * dev, struct drm_file *file_priv)
 {
-	drm_i915_private_t *dev_priv = dev->dev_private;
 	i915_gem_release(dev, file_priv);
-	if (!drm_core_check_feature(dev, DRIVER_MODESET))
-		i915_mem_release(dev, file_priv, dev_priv->agp_heap);
 }
 
 void i915_driver_postclose(struct drm_device *dev, struct drm_file *file)
@@ -2277,11 +2312,11 @@ struct drm_ioctl_desc i915_ioctls[] = {
 	DRM_IOCTL_DEF_DRV(I915_IRQ_WAIT, i915_irq_wait, DRM_AUTH),
 	DRM_IOCTL_DEF_DRV(I915_GETPARAM, i915_getparam, DRM_AUTH),
 	DRM_IOCTL_DEF_DRV(I915_SETPARAM, i915_setparam, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
-	DRM_IOCTL_DEF_DRV(I915_ALLOC, i915_mem_alloc, DRM_AUTH),
-	DRM_IOCTL_DEF_DRV(I915_FREE, i915_mem_free, DRM_AUTH),
-	DRM_IOCTL_DEF_DRV(I915_INIT_HEAP, i915_mem_init_heap, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
+	DRM_IOCTL_DEF_DRV(I915_ALLOC, drm_noop, DRM_AUTH),
+	DRM_IOCTL_DEF_DRV(I915_FREE, drm_noop, DRM_AUTH),
+	DRM_IOCTL_DEF_DRV(I915_INIT_HEAP, drm_noop, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
 	DRM_IOCTL_DEF_DRV(I915_CMDBUFFER, i915_cmdbuffer, DRM_AUTH),
-	DRM_IOCTL_DEF_DRV(I915_DESTROY_HEAP,  i915_mem_destroy_heap, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
+	DRM_IOCTL_DEF_DRV(I915_DESTROY_HEAP,  drm_noop, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
 	DRM_IOCTL_DEF_DRV(I915_SET_VBLANK_PIPE,  i915_vblank_pipe_set, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
 	DRM_IOCTL_DEF_DRV(I915_GET_VBLANK_PIPE,  i915_vblank_pipe_get, DRM_AUTH),
 	DRM_IOCTL_DEF_DRV(I915_VBLANK_SWAP, i915_vblank_swap, DRM_AUTH),
diff --git a/drivers/gpu/drm/i915/i915_drv.c b/drivers/gpu/drm/i915/i915_drv.c
index 308f81913562..dfa55e7478fb 100644
--- a/drivers/gpu/drm/i915/i915_drv.c
+++ b/drivers/gpu/drm/i915/i915_drv.c
@@ -66,7 +66,11 @@ MODULE_PARM_DESC(semaphores,
 int i915_enable_rc6 __read_mostly = -1;
 module_param_named(i915_enable_rc6, i915_enable_rc6, int, 0600);
 MODULE_PARM_DESC(i915_enable_rc6,
-		"Enable power-saving render C-state 6 (default: -1 (use per-chip default)");
+		"Enable power-saving render C-state 6. "
+		"Different stages can be selected via bitmask values "
+		"(0 = disable; 1 = enable rc6; 2 = enable deep rc6; 4 = enable deepest rc6). "
+		"For example, 3 would enable rc6 and deep rc6, and 7 would enable everything. "
+		"default: -1 (use per-chip default)");
 
 int i915_enable_fbc __read_mostly = -1;
 module_param_named(i915_enable_fbc, i915_enable_fbc, int, 0600);
@@ -103,6 +107,11 @@ MODULE_PARM_DESC(enable_hangcheck,
 		"WARNING: Disabling this can cause system wide hangs. "
 		"(default: true)");
 
+int i915_enable_ppgtt __read_mostly = -1;
+module_param_named(i915_enable_ppgtt, i915_enable_ppgtt, int, 0600);
+MODULE_PARM_DESC(i915_enable_ppgtt,
+		"Enable PPGTT (default: true)");
+
 static struct drm_driver driver;
 extern int intel_agp_enabled;
 
@@ -198,7 +207,7 @@ static const struct intel_device_info intel_pineview_info = {
 
 static const struct intel_device_info intel_ironlake_d_info = {
 	.gen = 5,
-	.need_gfx_hws = 1, .has_pipe_cxsr = 1, .has_hotplug = 1,
+	.need_gfx_hws = 1, .has_hotplug = 1,
 	.has_bsd_ring = 1,
 };
 
@@ -214,6 +223,7 @@ static const struct intel_device_info intel_sandybridge_d_info = {
 	.need_gfx_hws = 1, .has_hotplug = 1,
 	.has_bsd_ring = 1,
 	.has_blt_ring = 1,
+	.has_llc = 1,
 };
 
 static const struct intel_device_info intel_sandybridge_m_info = {
@@ -222,6 +232,7 @@ static const struct intel_device_info intel_sandybridge_m_info = {
 	.has_fbc = 1,
 	.has_bsd_ring = 1,
 	.has_blt_ring = 1,
+	.has_llc = 1,
 };
 
 static const struct intel_device_info intel_ivybridge_d_info = {
@@ -229,6 +240,7 @@ static const struct intel_device_info intel_ivybridge_d_info = {
 	.need_gfx_hws = 1, .has_hotplug = 1,
 	.has_bsd_ring = 1,
 	.has_blt_ring = 1,
+	.has_llc = 1,
 };
 
 static const struct intel_device_info intel_ivybridge_m_info = {
@@ -237,6 +249,7 @@ static const struct intel_device_info intel_ivybridge_m_info = {
 	.has_fbc = 0,	/* FBC is not enabled on Ivybridge mobile yet */
 	.has_bsd_ring = 1,
 	.has_blt_ring = 1,
+	.has_llc = 1,
 };
 
 static const struct pci_device_id pciidlist[] = {		/* aka */
@@ -283,6 +296,7 @@ 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_VGA_DEVICE(0x016a, &intel_ivybridge_d_info), /* GT2 server */
 	{0, 0, 0}
 };
 
@@ -376,16 +390,27 @@ void gen6_gt_force_wake_get(struct drm_i915_private *dev_priv)
 	spin_unlock_irqrestore(&dev_priv->gt_lock, irqflags);
 }
 
+static void gen6_gt_check_fifodbg(struct drm_i915_private *dev_priv)
+{
+	u32 gtfifodbg;
+	gtfifodbg = I915_READ_NOTRACE(GTFIFODBG);
+	if (WARN(gtfifodbg & GT_FIFO_CPU_ERROR_MASK,
+	     "MMIO read or write has been dropped %x\n", gtfifodbg))
+		I915_WRITE_NOTRACE(GTFIFODBG, GT_FIFO_CPU_ERROR_MASK);
+}
+
 void __gen6_gt_force_wake_put(struct drm_i915_private *dev_priv)
 {
 	I915_WRITE_NOTRACE(FORCEWAKE, 0);
-	POSTING_READ(FORCEWAKE);
+	/* The below doubles as a POSTING_READ */
+	gen6_gt_check_fifodbg(dev_priv);
 }
 
 void __gen6_gt_force_wake_mt_put(struct drm_i915_private *dev_priv)
 {
 	I915_WRITE_NOTRACE(FORCEWAKE_MT, (1<<16) | 0);
-	POSTING_READ(FORCEWAKE_MT);
+	/* The below doubles as a POSTING_READ */
+	gen6_gt_check_fifodbg(dev_priv);
 }
 
 /*
@@ -401,8 +426,10 @@ void gen6_gt_force_wake_put(struct drm_i915_private *dev_priv)
 	spin_unlock_irqrestore(&dev_priv->gt_lock, irqflags);
 }
 
-void __gen6_gt_wait_for_fifo(struct drm_i915_private *dev_priv)
+int __gen6_gt_wait_for_fifo(struct drm_i915_private *dev_priv)
 {
+	int ret = 0;
+
 	if (dev_priv->gt_fifo_count < GT_FIFO_NUM_RESERVED_ENTRIES) {
 		int loop = 500;
 		u32 fifo = I915_READ_NOTRACE(GT_FIFO_FREE_ENTRIES);
@@ -410,10 +437,13 @@ void __gen6_gt_wait_for_fifo(struct drm_i915_private *dev_priv)
 			udelay(10);
 			fifo = I915_READ_NOTRACE(GT_FIFO_FREE_ENTRIES);
 		}
-		WARN_ON(loop < 0 && fifo <= GT_FIFO_NUM_RESERVED_ENTRIES);
+		if (WARN_ON(loop < 0 && fifo <= GT_FIFO_NUM_RESERVED_ENTRIES))
+			++ret;
 		dev_priv->gt_fifo_count = fifo;
 	}
 	dev_priv->gt_fifo_count--;
+
+	return ret;
 }
 
 static int i915_drm_freeze(struct drm_device *dev)
@@ -442,6 +472,10 @@ static int i915_drm_freeze(struct drm_device *dev)
 	/* Modeset on resume, not lid events */
 	dev_priv->modeset_on_lid = 0;
 
+	console_lock();
+	intel_fbdev_set_suspend(dev, 1);
+	console_unlock();
+
 	return 0;
 }
 
@@ -494,7 +528,7 @@ static int i915_drm_thaw(struct drm_device *dev)
 		mutex_lock(&dev->struct_mutex);
 		dev_priv->mm.suspended = 0;
 
-		error = i915_gem_init_ringbuffer(dev);
+		error = i915_gem_init_hw(dev);
 		mutex_unlock(&dev->struct_mutex);
 
 		if (HAS_PCH_SPLIT(dev))
@@ -504,7 +538,9 @@ static int i915_drm_thaw(struct drm_device *dev)
 		drm_irq_install(dev);
 
 		/* Resume the modeset for every activated CRTC */
+		mutex_lock(&dev->mode_config.mutex);
 		drm_helper_resume_force_mode(dev);
+		mutex_unlock(&dev->mode_config.mutex);
 
 		if (IS_IRONLAKE_M(dev))
 			ironlake_enable_rc6(dev);
@@ -514,6 +550,9 @@ static int i915_drm_thaw(struct drm_device *dev)
 
 	dev_priv->modeset_on_lid = 0;
 
+	console_lock();
+	intel_fbdev_set_suspend(dev, 0);
+	console_unlock();
 	return error;
 }
 
@@ -633,7 +672,7 @@ static int gen6_do_reset(struct drm_device *dev, u8 flags)
 }
 
 /**
- * i965_reset - reset chip after a hang
+ * i915_reset - reset chip after a hang
  * @dev: drm device to reset
  * @flags: reset domains
  *
@@ -709,12 +748,16 @@ int i915_reset(struct drm_device *dev, u8 flags)
 			!dev_priv->mm.suspended) {
 		dev_priv->mm.suspended = 0;
 
+		i915_gem_init_swizzling(dev);
+
 		dev_priv->ring[RCS].init(&dev_priv->ring[RCS]);
 		if (HAS_BSD(dev))
 		    dev_priv->ring[VCS].init(&dev_priv->ring[VCS]);
 		if (HAS_BLT(dev))
 		    dev_priv->ring[BCS].init(&dev_priv->ring[BCS]);
 
+		i915_gem_init_ppgtt(dev);
+
 		mutex_unlock(&dev->struct_mutex);
 		drm_irq_uninstall(dev);
 		drm_mode_config_reset(dev);
@@ -977,11 +1020,15 @@ __i915_read(64, q)
 
 #define __i915_write(x, y) \
 void i915_write##x(struct drm_i915_private *dev_priv, u32 reg, u##x val) { \
+	u32 __fifo_ret = 0; \
 	trace_i915_reg_rw(true, reg, val, sizeof(val)); \
 	if (NEEDS_FORCE_WAKE((dev_priv), (reg))) { \
-		__gen6_gt_wait_for_fifo(dev_priv); \
+		__fifo_ret = __gen6_gt_wait_for_fifo(dev_priv); \
 	} \
 	write##y(val, dev_priv->regs + reg); \
+	if (unlikely(__fifo_ret)) { \
+		gen6_gt_check_fifodbg(dev_priv); \
+	} \
 }
 __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 9689ca38b2b3..5fabc6c31fec 100644
--- a/drivers/gpu/drm/i915/i915_drv.h
+++ b/drivers/gpu/drm/i915/i915_drv.h
@@ -35,6 +35,7 @@
 #include "intel_ringbuffer.h"
 #include <linux/io-mapping.h>
 #include <linux/i2c.h>
+#include <linux/i2c-algo-bit.h>
 #include <drm/intel-gtt.h>
 #include <linux/backlight.h>
 
@@ -135,6 +136,7 @@ struct drm_i915_fence_reg {
 	struct list_head lru_list;
 	struct drm_i915_gem_object *obj;
 	uint32_t setup_seqno;
+	int pin_count;
 };
 
 struct sdvo_device_mapping {
@@ -152,33 +154,40 @@ struct drm_i915_error_state {
 	u32 eir;
 	u32 pgtbl_er;
 	u32 pipestat[I915_MAX_PIPES];
-	u32 ipeir;
-	u32 ipehr;
-	u32 instdone;
-	u32 acthd;
+	u32 tail[I915_NUM_RINGS];
+	u32 head[I915_NUM_RINGS];
+	u32 ipeir[I915_NUM_RINGS];
+	u32 ipehr[I915_NUM_RINGS];
+	u32 instdone[I915_NUM_RINGS];
+	u32 acthd[I915_NUM_RINGS];
+	u32 semaphore_mboxes[I915_NUM_RINGS][I915_NUM_RINGS - 1];
+	/* our own tracking of ring head and tail */
+	u32 cpu_ring_head[I915_NUM_RINGS];
+	u32 cpu_ring_tail[I915_NUM_RINGS];
 	u32 error; /* gen6+ */
-	u32 bcs_acthd; /* gen6+ blt engine */
-	u32 bcs_ipehr;
-	u32 bcs_ipeir;
-	u32 bcs_instdone;
-	u32 bcs_seqno;
-	u32 vcs_acthd; /* gen6+ bsd engine */
-	u32 vcs_ipehr;
-	u32 vcs_ipeir;
-	u32 vcs_instdone;
-	u32 vcs_seqno;
-	u32 instpm;
-	u32 instps;
+	u32 instpm[I915_NUM_RINGS];
+	u32 instps[I915_NUM_RINGS];
 	u32 instdone1;
-	u32 seqno;
+	u32 seqno[I915_NUM_RINGS];
 	u64 bbaddr;
+	u32 fault_reg[I915_NUM_RINGS];
+	u32 done_reg;
+	u32 faddr[I915_NUM_RINGS];
 	u64 fence[I915_MAX_NUM_FENCES];
 	struct timeval time;
-	struct drm_i915_error_object {
-		int page_count;
-		u32 gtt_offset;
-		u32 *pages[0];
-	} *ringbuffer[I915_NUM_RINGS], *batchbuffer[I915_NUM_RINGS];
+	struct drm_i915_error_ring {
+		struct drm_i915_error_object {
+			int page_count;
+			u32 gtt_offset;
+			u32 *pages[0];
+		} *ringbuffer, *batchbuffer;
+		struct drm_i915_error_request {
+			long jiffies;
+			u32 seqno;
+			u32 tail;
+		} *requests;
+		int num_requests;
+	} ring[I915_NUM_RINGS];
 	struct drm_i915_error_buffer {
 		u32 size;
 		u32 name;
@@ -191,7 +200,7 @@ struct drm_i915_error_state {
 		u32 tiling:2;
 		u32 dirty:1;
 		u32 purgeable:1;
-		u32 ring:4;
+		s32 ring:4;
 		u32 cache_level:2;
 	} *active_bo, *pinned_bo;
 	u32 active_bo_count, pinned_bo_count;
@@ -255,6 +264,17 @@ struct intel_device_info {
 	u8 supports_tv:1;
 	u8 has_bsd_ring:1;
 	u8 has_blt_ring:1;
+	u8 has_llc:1;
+};
+
+#define I915_PPGTT_PD_ENTRIES 512
+#define I915_PPGTT_PT_ENTRIES 1024
+struct i915_hw_ppgtt {
+	unsigned num_pd_entries;
+	struct page **pt_pages;
+	uint32_t pd_offset;
+	dma_addr_t *pt_dma_addr;
+	dma_addr_t scratch_page_dma_addr;
 };
 
 enum no_fbc_reason {
@@ -279,6 +299,16 @@ enum intel_pch {
 struct intel_fbdev;
 struct intel_fbc_work;
 
+struct intel_gmbus {
+	struct i2c_adapter adapter;
+	bool force_bit;
+	bool has_gpio;
+	u32 reg0;
+	u32 gpio_reg;
+	struct i2c_algo_bit_data bit_algo;
+	struct drm_i915_private *dev_priv;
+};
+
 typedef struct drm_i915_private {
 	struct drm_device *dev;
 
@@ -296,11 +326,11 @@ typedef struct drm_i915_private {
 	/** gt_lock is also taken in irq contexts. */
 	struct spinlock gt_lock;
 
-	struct intel_gmbus {
-		struct i2c_adapter adapter;
-		struct i2c_adapter *force_bit;
-		u32 reg0;
-	} *gmbus;
+	struct intel_gmbus *gmbus;
+
+	/** gmbus_mutex protects against concurrent usage of the single hw gmbus
+	 * controller on different i2c buses. */
+	struct mutex gmbus_mutex;
 
 	struct pci_dev *bridge_dev;
 	struct intel_ring_buffer ring[I915_NUM_RINGS];
@@ -335,7 +365,6 @@ typedef struct drm_i915_private {
 
 	int tex_lru_log_granularity;
 	int allow_batchbuffer;
-	struct mem_block *agp_heap;
 	unsigned int sr01, adpa, ppcr, dvob, dvoc, lvds;
 	int vblank_pipe;
 	int num_pipe;
@@ -584,6 +613,9 @@ typedef struct drm_i915_private {
 		struct io_mapping *gtt_mapping;
 		int gtt_mtrr;
 
+		/** PPGTT used for aliasing the PPGTT with the GTT */
+		struct i915_hw_ppgtt *aliasing_ppgtt;
+
 		struct shrinker inactive_shrinker;
 
 		/**
@@ -749,6 +781,13 @@ typedef struct drm_i915_private {
 	struct drm_property *force_audio_property;
 } drm_i915_private_t;
 
+enum hdmi_force_audio {
+	HDMI_AUDIO_OFF_DVI = -2,	/* no aux data for HDMI-DVI converter */
+	HDMI_AUDIO_OFF,			/* force turn off HDMI audio */
+	HDMI_AUDIO_AUTO,		/* trust EDID */
+	HDMI_AUDIO_ON,			/* force turn on HDMI audio */
+};
+
 enum i915_cache_level {
 	I915_CACHE_NONE,
 	I915_CACHE_LLC,
@@ -841,6 +880,8 @@ struct drm_i915_gem_object {
 
 	unsigned int cache_level:2;
 
+	unsigned int has_aliasing_ppgtt_mapping:1;
+
 	struct page **pages;
 
 	/**
@@ -918,6 +959,9 @@ struct drm_i915_gem_request {
 	/** GEM sequence number associated with this request. */
 	uint32_t seqno;
 
+	/** Postion in the ringbuffer of the end of the request */
+	u32 tail;
+
 	/** Time at which this request was emitted, in jiffies. */
 	unsigned long emitted_jiffies;
 
@@ -974,8 +1018,11 @@ struct drm_i915_file_private {
 
 #define HAS_BSD(dev)            (INTEL_INFO(dev)->has_bsd_ring)
 #define HAS_BLT(dev)            (INTEL_INFO(dev)->has_blt_ring)
+#define HAS_LLC(dev)            (INTEL_INFO(dev)->has_llc)
 #define I915_NEED_GFX_HWS(dev)	(INTEL_INFO(dev)->need_gfx_hws)
 
+#define HAS_ALIASING_PPGTT(dev)	(INTEL_INFO(dev)->gen >=6)
+
 #define HAS_OVERLAY(dev)		(INTEL_INFO(dev)->has_overlay)
 #define OVERLAY_NEEDS_PHYSICAL(dev)	(INTEL_INFO(dev)->overlay_needs_physical)
 
@@ -1006,6 +1053,27 @@ struct drm_i915_file_private {
 
 #include "i915_trace.h"
 
+/**
+ * RC6 is a special power stage which allows the GPU to enter an very
+ * low-voltage mode when idle, using down to 0V while at this stage.  This
+ * stage is entered automatically when the GPU is idle when RC6 support is
+ * enabled, and as soon as new workload arises GPU wakes up automatically as well.
+ *
+ * There are different RC6 modes available in Intel GPU, which differentiate
+ * among each other with the latency required to enter and leave RC6 and
+ * voltage consumed by the GPU in different states.
+ *
+ * The combination of the following flags define which states GPU is allowed
+ * to enter, while RC6 is the normal RC6 state, RC6p is the deep RC6, and
+ * RC6pp is deepest RC6. Their support by hardware varies according to the
+ * GPU, BIOS, chipset and platform. RC6 is usually the safest one and the one
+ * which brings the most power savings; deeper states save more power, but
+ * require higher latency to switch to and wake up.
+ */
+#define INTEL_RC6_ENABLE			(1<<0)
+#define INTEL_RC6p_ENABLE			(1<<1)
+#define INTEL_RC6pp_ENABLE			(1<<2)
+
 extern struct drm_ioctl_desc i915_ioctls[];
 extern int i915_max_ioctl;
 extern unsigned int i915_fbpercrtc __always_unused;
@@ -1018,6 +1086,7 @@ extern int i915_vbt_sdvo_panel_type __read_mostly;
 extern int i915_enable_rc6 __read_mostly;
 extern int i915_enable_fbc __read_mostly;
 extern bool i915_enable_hangcheck __read_mostly;
+extern int i915_enable_ppgtt __read_mostly;
 
 extern int i915_suspend(struct drm_device *dev, pm_message_t state);
 extern int i915_resume(struct drm_device *dev);
@@ -1079,18 +1148,6 @@ extern void i915_destroy_error_state(struct drm_device *dev);
 #endif
 
 
-/* i915_mem.c */
-extern int i915_mem_alloc(struct drm_device *dev, void *data,
-			  struct drm_file *file_priv);
-extern int i915_mem_free(struct drm_device *dev, void *data,
-			 struct drm_file *file_priv);
-extern int i915_mem_init_heap(struct drm_device *dev, void *data,
-			      struct drm_file *file_priv);
-extern int i915_mem_destroy_heap(struct drm_device *dev, void *data,
-				 struct drm_file *file_priv);
-extern void i915_mem_takedown(struct mem_block **heap);
-extern void i915_mem_release(struct drm_device * dev,
-			     struct drm_file *file_priv, struct mem_block *heap);
 /* i915_gem.c */
 int i915_gem_init_ioctl(struct drm_device *dev, void *data,
 			struct drm_file *file_priv);
@@ -1170,37 +1227,55 @@ i915_seqno_passed(uint32_t seq1, uint32_t seq2)
 	return (int32_t)(seq1 - seq2) >= 0;
 }
 
-static inline u32
-i915_gem_next_request_seqno(struct intel_ring_buffer *ring)
-{
-	drm_i915_private_t *dev_priv = ring->dev->dev_private;
-	return ring->outstanding_lazy_request = dev_priv->next_seqno;
-}
+u32 i915_gem_next_request_seqno(struct intel_ring_buffer *ring);
 
 int __must_check i915_gem_object_get_fence(struct drm_i915_gem_object *obj,
 					   struct intel_ring_buffer *pipelined);
 int __must_check i915_gem_object_put_fence(struct drm_i915_gem_object *obj);
 
+static inline void
+i915_gem_object_pin_fence(struct drm_i915_gem_object *obj)
+{
+	if (obj->fence_reg != I915_FENCE_REG_NONE) {
+		struct drm_i915_private *dev_priv = obj->base.dev->dev_private;
+		dev_priv->fence_regs[obj->fence_reg].pin_count++;
+	}
+}
+
+static inline void
+i915_gem_object_unpin_fence(struct drm_i915_gem_object *obj)
+{
+	if (obj->fence_reg != I915_FENCE_REG_NONE) {
+		struct drm_i915_private *dev_priv = obj->base.dev->dev_private;
+		dev_priv->fence_regs[obj->fence_reg].pin_count--;
+	}
+}
+
 void i915_gem_retire_requests(struct drm_device *dev);
+void i915_gem_retire_requests_ring(struct intel_ring_buffer *ring);
+
 void i915_gem_reset(struct drm_device *dev);
 void i915_gem_clflush_object(struct drm_i915_gem_object *obj);
 int __must_check i915_gem_object_set_domain(struct drm_i915_gem_object *obj,
 					    uint32_t read_domains,
 					    uint32_t write_domain);
 int __must_check i915_gem_object_finish_gpu(struct drm_i915_gem_object *obj);
-int __must_check i915_gem_init_ringbuffer(struct drm_device *dev);
+int __must_check i915_gem_init_hw(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);
 void i915_gem_do_init(struct drm_device *dev,
 		      unsigned long start,
 		      unsigned long mappable_end,
 		      unsigned long end);
-int __must_check i915_gpu_idle(struct drm_device *dev);
+int __must_check i915_gpu_idle(struct drm_device *dev, bool do_retire);
 int __must_check i915_gem_idle(struct drm_device *dev);
 int __must_check i915_add_request(struct intel_ring_buffer *ring,
 				  struct drm_file *file,
 				  struct drm_i915_gem_request *request);
 int __must_check i915_wait_request(struct intel_ring_buffer *ring,
-				   uint32_t seqno);
+				   uint32_t seqno,
+				   bool do_retire);
 int i915_gem_fault(struct vm_area_struct *vma, struct vm_fault *vmf);
 int __must_check
 i915_gem_object_set_to_gtt_domain(struct drm_i915_gem_object *obj,
@@ -1227,6 +1302,14 @@ int i915_gem_object_set_cache_level(struct drm_i915_gem_object *obj,
 				    enum i915_cache_level cache_level);
 
 /* i915_gem_gtt.c */
+int __must_check i915_gem_init_aliasing_ppgtt(struct drm_device *dev);
+void i915_gem_cleanup_aliasing_ppgtt(struct drm_device *dev);
+void i915_ppgtt_bind_object(struct i915_hw_ppgtt *ppgtt,
+			    struct drm_i915_gem_object *obj,
+			    enum i915_cache_level cache_level);
+void i915_ppgtt_unbind_object(struct i915_hw_ppgtt *ppgtt,
+			      struct drm_i915_gem_object *obj);
+
 void i915_gem_restore_gtt_mappings(struct drm_device *dev);
 int __must_check i915_gem_gtt_bind_object(struct drm_i915_gem_object *obj);
 void i915_gem_gtt_rebind_object(struct drm_i915_gem_object *obj,
@@ -1365,7 +1448,7 @@ extern void intel_display_print_error_state(struct seq_file *m,
  */
 void gen6_gt_force_wake_get(struct drm_i915_private *dev_priv);
 void gen6_gt_force_wake_put(struct drm_i915_private *dev_priv);
-void __gen6_gt_wait_for_fifo(struct drm_i915_private *dev_priv);
+int __gen6_gt_wait_for_fifo(struct drm_i915_private *dev_priv);
 
 /* We give fast paths for the really cool registers */
 #define NEEDS_FORCE_WAKE(dev_priv, reg) \
diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c
index e55badb2d86d..4c65c639f772 100644
--- a/drivers/gpu/drm/i915/i915_gem.c
+++ b/drivers/gpu/drm/i915/i915_gem.c
@@ -58,6 +58,7 @@ static void i915_gem_free_object_tail(struct drm_i915_gem_object *obj);
 
 static int i915_gem_inactive_shrink(struct shrinker *shrinker,
 				    struct shrink_control *sc);
+static void i915_gem_object_truncate(struct drm_i915_gem_object *obj);
 
 /* some bookkeeping */
 static void i915_gem_info_add_obj(struct drm_i915_private *dev_priv,
@@ -258,73 +259,6 @@ static int i915_gem_object_needs_bit17_swizzle(struct drm_i915_gem_object *obj)
 		obj->tiling_mode != I915_TILING_NONE;
 }
 
-static inline void
-slow_shmem_copy(struct page *dst_page,
-		int dst_offset,
-		struct page *src_page,
-		int src_offset,
-		int length)
-{
-	char *dst_vaddr, *src_vaddr;
-
-	dst_vaddr = kmap(dst_page);
-	src_vaddr = kmap(src_page);
-
-	memcpy(dst_vaddr + dst_offset, src_vaddr + src_offset, length);
-
-	kunmap(src_page);
-	kunmap(dst_page);
-}
-
-static inline void
-slow_shmem_bit17_copy(struct page *gpu_page,
-		      int gpu_offset,
-		      struct page *cpu_page,
-		      int cpu_offset,
-		      int length,
-		      int is_read)
-{
-	char *gpu_vaddr, *cpu_vaddr;
-
-	/* Use the unswizzled path if this page isn't affected. */
-	if ((page_to_phys(gpu_page) & (1 << 17)) == 0) {
-		if (is_read)
-			return slow_shmem_copy(cpu_page, cpu_offset,
-					       gpu_page, gpu_offset, length);
-		else
-			return slow_shmem_copy(gpu_page, gpu_offset,
-					       cpu_page, cpu_offset, length);
-	}
-
-	gpu_vaddr = kmap(gpu_page);
-	cpu_vaddr = kmap(cpu_page);
-
-	/* Copy the data, XORing A6 with A17 (1). The user already knows he's
-	 * XORing with the other bits (A9 for Y, A9 and A10 for X)
-	 */
-	while (length > 0) {
-		int cacheline_end = ALIGN(gpu_offset + 1, 64);
-		int this_length = min(cacheline_end - gpu_offset, length);
-		int swizzled_gpu_offset = gpu_offset ^ 64;
-
-		if (is_read) {
-			memcpy(cpu_vaddr + cpu_offset,
-			       gpu_vaddr + swizzled_gpu_offset,
-			       this_length);
-		} else {
-			memcpy(gpu_vaddr + swizzled_gpu_offset,
-			       cpu_vaddr + cpu_offset,
-			       this_length);
-		}
-		cpu_offset += this_length;
-		gpu_offset += this_length;
-		length -= this_length;
-	}
-
-	kunmap(cpu_page);
-	kunmap(gpu_page);
-}
-
 /**
  * This is the fast shmem pread path, which attempts to copy_from_user directly
  * from the backing pages of the object to the user's address space.  On a
@@ -385,6 +319,58 @@ i915_gem_shmem_pread_fast(struct drm_device *dev,
 	return 0;
 }
 
+static inline int
+__copy_to_user_swizzled(char __user *cpu_vaddr,
+			const char *gpu_vaddr, int gpu_offset,
+			int length)
+{
+	int ret, cpu_offset = 0;
+
+	while (length > 0) {
+		int cacheline_end = ALIGN(gpu_offset + 1, 64);
+		int this_length = min(cacheline_end - gpu_offset, length);
+		int swizzled_gpu_offset = gpu_offset ^ 64;
+
+		ret = __copy_to_user(cpu_vaddr + cpu_offset,
+				     gpu_vaddr + swizzled_gpu_offset,
+				     this_length);
+		if (ret)
+			return ret + length;
+
+		cpu_offset += this_length;
+		gpu_offset += this_length;
+		length -= this_length;
+	}
+
+	return 0;
+}
+
+static inline int
+__copy_from_user_swizzled(char __user *gpu_vaddr, int gpu_offset,
+			  const char *cpu_vaddr,
+			  int length)
+{
+	int ret, cpu_offset = 0;
+
+	while (length > 0) {
+		int cacheline_end = ALIGN(gpu_offset + 1, 64);
+		int this_length = min(cacheline_end - gpu_offset, length);
+		int swizzled_gpu_offset = gpu_offset ^ 64;
+
+		ret = __copy_from_user(gpu_vaddr + swizzled_gpu_offset,
+				       cpu_vaddr + cpu_offset,
+				       this_length);
+		if (ret)
+			return ret + length;
+
+		cpu_offset += this_length;
+		gpu_offset += this_length;
+		length -= this_length;
+	}
+
+	return 0;
+}
+
 /**
  * This is the fallback shmem pread path, which allocates temporary storage
  * in kernel space to copy_to_user into outside of the struct_mutex, so we
@@ -398,72 +384,34 @@ i915_gem_shmem_pread_slow(struct drm_device *dev,
 			  struct drm_file *file)
 {
 	struct address_space *mapping = obj->base.filp->f_path.dentry->d_inode->i_mapping;
-	struct mm_struct *mm = current->mm;
-	struct page **user_pages;
+	char __user *user_data;
 	ssize_t remain;
-	loff_t offset, pinned_pages, i;
-	loff_t first_data_page, last_data_page, num_pages;
-	int shmem_page_offset;
-	int data_page_index, data_page_offset;
-	int page_length;
-	int ret;
-	uint64_t data_ptr = args->data_ptr;
-	int do_bit17_swizzling;
+	loff_t offset;
+	int shmem_page_offset, page_length, ret;
+	int obj_do_bit17_swizzling, page_do_bit17_swizzling;
 
+	user_data = (char __user *) (uintptr_t) args->data_ptr;
 	remain = args->size;
 
-	/* Pin the user pages containing the data.  We can't fault while
-	 * holding the struct mutex, yet we want to hold it while
-	 * dereferencing the user data.
-	 */
-	first_data_page = data_ptr / PAGE_SIZE;
-	last_data_page = (data_ptr + args->size - 1) / PAGE_SIZE;
-	num_pages = last_data_page - first_data_page + 1;
+	obj_do_bit17_swizzling = i915_gem_object_needs_bit17_swizzle(obj);
 
-	user_pages = drm_malloc_ab(num_pages, sizeof(struct page *));
-	if (user_pages == NULL)
-		return -ENOMEM;
+	offset = args->offset;
 
 	mutex_unlock(&dev->struct_mutex);
-	down_read(&mm->mmap_sem);
-	pinned_pages = get_user_pages(current, mm, (uintptr_t)args->data_ptr,
-				      num_pages, 1, 0, user_pages, NULL);
-	up_read(&mm->mmap_sem);
-	mutex_lock(&dev->struct_mutex);
-	if (pinned_pages < num_pages) {
-		ret = -EFAULT;
-		goto out;
-	}
-
-	ret = i915_gem_object_set_cpu_read_domain_range(obj,
-							args->offset,
-							args->size);
-	if (ret)
-		goto out;
-
-	do_bit17_swizzling = i915_gem_object_needs_bit17_swizzle(obj);
-
-	offset = args->offset;
 
 	while (remain > 0) {
 		struct page *page;
+		char *vaddr;
 
 		/* Operation in this page
 		 *
 		 * shmem_page_offset = offset within page in shmem file
-		 * data_page_index = page number in get_user_pages return
-		 * data_page_offset = offset with data_page_index page.
 		 * page_length = bytes to copy for this page
 		 */
 		shmem_page_offset = offset_in_page(offset);
-		data_page_index = data_ptr / PAGE_SIZE - first_data_page;
-		data_page_offset = offset_in_page(data_ptr);
-
 		page_length = remain;
 		if ((shmem_page_offset + page_length) > PAGE_SIZE)
 			page_length = PAGE_SIZE - shmem_page_offset;
-		if ((data_page_offset + page_length) > PAGE_SIZE)
-			page_length = PAGE_SIZE - data_page_offset;
 
 		page = shmem_read_mapping_page(mapping, offset >> PAGE_SHIFT);
 		if (IS_ERR(page)) {
@@ -471,36 +419,38 @@ i915_gem_shmem_pread_slow(struct drm_device *dev,
 			goto out;
 		}
 
-		if (do_bit17_swizzling) {
-			slow_shmem_bit17_copy(page,
-					      shmem_page_offset,
-					      user_pages[data_page_index],
-					      data_page_offset,
-					      page_length,
-					      1);
-		} else {
-			slow_shmem_copy(user_pages[data_page_index],
-					data_page_offset,
-					page,
-					shmem_page_offset,
-					page_length);
-		}
+		page_do_bit17_swizzling = obj_do_bit17_swizzling &&
+			(page_to_phys(page) & (1 << 17)) != 0;
+
+		vaddr = kmap(page);
+		if (page_do_bit17_swizzling)
+			ret = __copy_to_user_swizzled(user_data,
+						      vaddr, shmem_page_offset,
+						      page_length);
+		else
+			ret = __copy_to_user(user_data,
+					     vaddr + shmem_page_offset,
+					     page_length);
+		kunmap(page);
 
 		mark_page_accessed(page);
 		page_cache_release(page);
 
+		if (ret) {
+			ret = -EFAULT;
+			goto out;
+		}
+
 		remain -= page_length;
-		data_ptr += page_length;
+		user_data += page_length;
 		offset += page_length;
 	}
 
 out:
-	for (i = 0; i < pinned_pages; i++) {
-		SetPageDirty(user_pages[i]);
-		mark_page_accessed(user_pages[i]);
-		page_cache_release(user_pages[i]);
-	}
-	drm_free_large(user_pages);
+	mutex_lock(&dev->struct_mutex);
+	/* Fixup: Kill any reinstated backing storage pages */
+	if (obj->madv == __I915_MADV_PURGED)
+		i915_gem_object_truncate(obj);
 
 	return ret;
 }
@@ -841,71 +791,36 @@ i915_gem_shmem_pwrite_slow(struct drm_device *dev,
 			   struct drm_file *file)
 {
 	struct address_space *mapping = obj->base.filp->f_path.dentry->d_inode->i_mapping;
-	struct mm_struct *mm = current->mm;
-	struct page **user_pages;
 	ssize_t remain;
-	loff_t offset, pinned_pages, i;
-	loff_t first_data_page, last_data_page, num_pages;
-	int shmem_page_offset;
-	int data_page_index,  data_page_offset;
-	int page_length;
-	int ret;
-	uint64_t data_ptr = args->data_ptr;
-	int do_bit17_swizzling;
+	loff_t offset;
+	char __user *user_data;
+	int shmem_page_offset, page_length, ret;
+	int obj_do_bit17_swizzling, page_do_bit17_swizzling;
 
+	user_data = (char __user *) (uintptr_t) args->data_ptr;
 	remain = args->size;
 
-	/* Pin the user pages containing the data.  We can't fault while
-	 * holding the struct mutex, and all of the pwrite implementations
-	 * want to hold it while dereferencing the user data.
-	 */
-	first_data_page = data_ptr / PAGE_SIZE;
-	last_data_page = (data_ptr + args->size - 1) / PAGE_SIZE;
-	num_pages = last_data_page - first_data_page + 1;
-
-	user_pages = drm_malloc_ab(num_pages, sizeof(struct page *));
-	if (user_pages == NULL)
-		return -ENOMEM;
-
-	mutex_unlock(&dev->struct_mutex);
-	down_read(&mm->mmap_sem);
-	pinned_pages = get_user_pages(current, mm, (uintptr_t)args->data_ptr,
-				      num_pages, 0, 0, user_pages, NULL);
-	up_read(&mm->mmap_sem);
-	mutex_lock(&dev->struct_mutex);
-	if (pinned_pages < num_pages) {
-		ret = -EFAULT;
-		goto out;
-	}
-
-	ret = i915_gem_object_set_to_cpu_domain(obj, 1);
-	if (ret)
-		goto out;
-
-	do_bit17_swizzling = i915_gem_object_needs_bit17_swizzle(obj);
+	obj_do_bit17_swizzling = i915_gem_object_needs_bit17_swizzle(obj);
 
 	offset = args->offset;
 	obj->dirty = 1;
 
+	mutex_unlock(&dev->struct_mutex);
+
 	while (remain > 0) {
 		struct page *page;
+		char *vaddr;
 
 		/* Operation in this page
 		 *
 		 * shmem_page_offset = offset within page in shmem file
-		 * data_page_index = page number in get_user_pages return
-		 * data_page_offset = offset with data_page_index page.
 		 * page_length = bytes to copy for this page
 		 */
 		shmem_page_offset = offset_in_page(offset);
-		data_page_index = data_ptr / PAGE_SIZE - first_data_page;
-		data_page_offset = offset_in_page(data_ptr);
 
 		page_length = remain;
 		if ((shmem_page_offset + page_length) > PAGE_SIZE)
 			page_length = PAGE_SIZE - shmem_page_offset;
-		if ((data_page_offset + page_length) > PAGE_SIZE)
-			page_length = PAGE_SIZE - data_page_offset;
 
 		page = shmem_read_mapping_page(mapping, offset >> PAGE_SHIFT);
 		if (IS_ERR(page)) {
@@ -913,34 +828,45 @@ i915_gem_shmem_pwrite_slow(struct drm_device *dev,
 			goto out;
 		}
 
-		if (do_bit17_swizzling) {
-			slow_shmem_bit17_copy(page,
-					      shmem_page_offset,
-					      user_pages[data_page_index],
-					      data_page_offset,
-					      page_length,
-					      0);
-		} else {
-			slow_shmem_copy(page,
-					shmem_page_offset,
-					user_pages[data_page_index],
-					data_page_offset,
-					page_length);
-		}
+		page_do_bit17_swizzling = obj_do_bit17_swizzling &&
+			(page_to_phys(page) & (1 << 17)) != 0;
+
+		vaddr = kmap(page);
+		if (page_do_bit17_swizzling)
+			ret = __copy_from_user_swizzled(vaddr, shmem_page_offset,
+							user_data,
+							page_length);
+		else
+			ret = __copy_from_user(vaddr + shmem_page_offset,
+					       user_data,
+					       page_length);
+		kunmap(page);
 
 		set_page_dirty(page);
 		mark_page_accessed(page);
 		page_cache_release(page);
 
+		if (ret) {
+			ret = -EFAULT;
+			goto out;
+		}
+
 		remain -= page_length;
-		data_ptr += page_length;
+		user_data += page_length;
 		offset += page_length;
 	}
 
 out:
-	for (i = 0; i < pinned_pages; i++)
-		page_cache_release(user_pages[i]);
-	drm_free_large(user_pages);
+	mutex_lock(&dev->struct_mutex);
+	/* Fixup: Kill any reinstated backing storage pages */
+	if (obj->madv == __I915_MADV_PURGED)
+		i915_gem_object_truncate(obj);
+	/* and flush dirty cachelines in case the object isn't in the cpu write
+	 * domain anymore. */
+	if (obj->base.write_domain != I915_GEM_DOMAIN_CPU) {
+		i915_gem_clflush_object(obj);
+		intel_gtt_chipset_flush();
+	}
 
 	return ret;
 }
@@ -996,10 +922,13 @@ i915_gem_pwrite_ioctl(struct drm_device *dev, void *data,
 	 * pread/pwrite currently are reading and writing from the CPU
 	 * perspective, requiring manual detiling by the client.
 	 */
-	if (obj->phys_obj)
+	if (obj->phys_obj) {
 		ret = i915_gem_phys_pwrite(dev, obj, args, file);
-	else if (obj->gtt_space &&
-		 obj->base.write_domain != I915_GEM_DOMAIN_CPU) {
+		goto out;
+	}
+
+	if (obj->gtt_space &&
+	    obj->base.write_domain != I915_GEM_DOMAIN_CPU) {
 		ret = i915_gem_object_pin(obj, 0, true);
 		if (ret)
 			goto out;
@@ -1018,18 +947,24 @@ i915_gem_pwrite_ioctl(struct drm_device *dev, void *data,
 
 out_unpin:
 		i915_gem_object_unpin(obj);
-	} else {
-		ret = i915_gem_object_set_to_cpu_domain(obj, 1);
-		if (ret)
-			goto out;
 
-		ret = -EFAULT;
-		if (!i915_gem_object_needs_bit17_swizzle(obj))
-			ret = i915_gem_shmem_pwrite_fast(dev, obj, args, file);
-		if (ret == -EFAULT)
-			ret = i915_gem_shmem_pwrite_slow(dev, obj, args, file);
+		if (ret != -EFAULT)
+			goto out;
+		/* Fall through to the shmfs paths because the gtt paths might
+		 * fail with non-page-backed user pointers (e.g. gtt mappings
+		 * when moving data between textures). */
 	}
 
+	ret = i915_gem_object_set_to_cpu_domain(obj, 1);
+	if (ret)
+		goto out;
+
+	ret = -EFAULT;
+	if (!i915_gem_object_needs_bit17_swizzle(obj))
+		ret = i915_gem_shmem_pwrite_fast(dev, obj, args, file);
+	if (ret == -EFAULT)
+		ret = i915_gem_shmem_pwrite_slow(dev, obj, args, file);
+
 out:
 	drm_gem_object_unreference(&obj->base);
 unlock:
@@ -1141,7 +1076,6 @@ int
 i915_gem_mmap_ioctl(struct drm_device *dev, void *data,
 		    struct drm_file *file)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct drm_i915_gem_mmap *args = data;
 	struct drm_gem_object *obj;
 	unsigned long addr;
@@ -1153,11 +1087,6 @@ i915_gem_mmap_ioctl(struct drm_device *dev, void *data,
 	if (obj == NULL)
 		return -ENOENT;
 
-	if (obj->size > dev_priv->mm.gtt_mappable_end) {
-		drm_gem_object_unreference_unlocked(obj);
-		return -E2BIG;
-	}
-
 	down_write(&current->mm->mmap_sem);
 	addr = do_mmap(obj->filp, 0, args->size,
 		       PROT_READ | PROT_WRITE, MAP_SHARED,
@@ -1543,16 +1472,19 @@ i915_gem_object_move_to_active(struct drm_i915_gem_object *obj,
 	list_move_tail(&obj->ring_list, &ring->active_list);
 
 	obj->last_rendering_seqno = seqno;
-	if (obj->fenced_gpu_access) {
-		struct drm_i915_fence_reg *reg;
-
-		BUG_ON(obj->fence_reg == I915_FENCE_REG_NONE);
 
+	if (obj->fenced_gpu_access) {
 		obj->last_fenced_seqno = seqno;
 		obj->last_fenced_ring = ring;
 
-		reg = &dev_priv->fence_regs[obj->fence_reg];
-		list_move_tail(&reg->lru_list, &dev_priv->mm.fence_list);
+		/* Bump MRU to take account of the delayed flush */
+		if (obj->fence_reg != I915_FENCE_REG_NONE) {
+			struct drm_i915_fence_reg *reg;
+
+			reg = &dev_priv->fence_regs[obj->fence_reg];
+			list_move_tail(&reg->lru_list,
+				       &dev_priv->mm.fence_list);
+		}
 	}
 }
 
@@ -1647,6 +1579,28 @@ i915_gem_process_flushing_list(struct intel_ring_buffer *ring,
 	}
 }
 
+static u32
+i915_gem_get_seqno(struct drm_device *dev)
+{
+	drm_i915_private_t *dev_priv = dev->dev_private;
+	u32 seqno = dev_priv->next_seqno;
+
+	/* reserve 0 for non-seqno */
+	if (++dev_priv->next_seqno == 0)
+		dev_priv->next_seqno = 1;
+
+	return seqno;
+}
+
+u32
+i915_gem_next_request_seqno(struct intel_ring_buffer *ring)
+{
+	if (ring->outstanding_lazy_request == 0)
+		ring->outstanding_lazy_request = i915_gem_get_seqno(ring->dev);
+
+	return ring->outstanding_lazy_request;
+}
+
 int
 i915_add_request(struct intel_ring_buffer *ring,
 		 struct drm_file *file,
@@ -1654,10 +1608,19 @@ i915_add_request(struct intel_ring_buffer *ring,
 {
 	drm_i915_private_t *dev_priv = ring->dev->dev_private;
 	uint32_t seqno;
+	u32 request_ring_position;
 	int was_empty;
 	int ret;
 
 	BUG_ON(request == NULL);
+	seqno = i915_gem_next_request_seqno(ring);
+
+	/* Record the position of the start of the request so that
+	 * should we detect the updated seqno part-way through the
+	 * GPU processing the request, we never over-estimate the
+	 * position of the head.
+	 */
+	request_ring_position = intel_ring_get_tail(ring);
 
 	ret = ring->add_request(ring, &seqno);
 	if (ret)
@@ -1667,6 +1630,7 @@ i915_add_request(struct intel_ring_buffer *ring,
 
 	request->seqno = seqno;
 	request->ring = ring;
+	request->tail = request_ring_position;
 	request->emitted_jiffies = jiffies;
 	was_empty = list_empty(&ring->request_list);
 	list_add_tail(&request->list, &ring->request_list);
@@ -1681,7 +1645,7 @@ i915_add_request(struct intel_ring_buffer *ring,
 		spin_unlock(&file_priv->mm.lock);
 	}
 
-	ring->outstanding_lazy_request = false;
+	ring->outstanding_lazy_request = 0;
 
 	if (!dev_priv->mm.suspended) {
 		if (i915_enable_hangcheck) {
@@ -1803,7 +1767,7 @@ void i915_gem_reset(struct drm_device *dev)
 /**
  * This function clears the request list as sequence numbers are passed.
  */
-static void
+void
 i915_gem_retire_requests_ring(struct intel_ring_buffer *ring)
 {
 	uint32_t seqno;
@@ -1831,6 +1795,12 @@ i915_gem_retire_requests_ring(struct intel_ring_buffer *ring)
 			break;
 
 		trace_i915_gem_request_retire(ring, request->seqno);
+		/* We know the GPU must have read the request to have
+		 * sent us the seqno + interrupt, so use the position
+		 * of tail of the request to update the last known position
+		 * of the GPU head.
+		 */
+		ring->last_retired_head = request->tail;
 
 		list_del(&request->list);
 		i915_gem_request_remove_from_client(request);
@@ -1943,7 +1913,8 @@ i915_gem_retire_work_handler(struct work_struct *work)
  */
 int
 i915_wait_request(struct intel_ring_buffer *ring,
-		  uint32_t seqno)
+		  uint32_t seqno,
+		  bool do_retire)
 {
 	drm_i915_private_t *dev_priv = ring->dev->dev_private;
 	u32 ier;
@@ -2017,17 +1988,12 @@ i915_wait_request(struct intel_ring_buffer *ring,
 	if (atomic_read(&dev_priv->mm.wedged))
 		ret = -EAGAIN;
 
-	if (ret && ret != -ERESTARTSYS)
-		DRM_ERROR("%s returns %d (awaiting %d at %d, next %d)\n",
-			  __func__, ret, seqno, ring->get_seqno(ring),
-			  dev_priv->next_seqno);
-
 	/* Directly dispatch request retiring.  While we have the work queue
 	 * to handle this, the waiter on a request often wants an associated
 	 * buffer to have made it to the inactive list, and we would need
 	 * a separate wait queue to handle that.
 	 */
-	if (ret == 0)
+	if (ret == 0 && do_retire)
 		i915_gem_retire_requests_ring(ring);
 
 	return ret;
@@ -2051,7 +2017,8 @@ i915_gem_object_wait_rendering(struct drm_i915_gem_object *obj)
 	 * it.
 	 */
 	if (obj->active) {
-		ret = i915_wait_request(obj->ring, obj->last_rendering_seqno);
+		ret = i915_wait_request(obj->ring, obj->last_rendering_seqno,
+					true);
 		if (ret)
 			return ret;
 	}
@@ -2089,6 +2056,7 @@ static void i915_gem_object_finish_gtt(struct drm_i915_gem_object *obj)
 int
 i915_gem_object_unbind(struct drm_i915_gem_object *obj)
 {
+	drm_i915_private_t *dev_priv = obj->base.dev->dev_private;
 	int ret = 0;
 
 	if (obj->gtt_space == NULL)
@@ -2133,6 +2101,11 @@ i915_gem_object_unbind(struct drm_i915_gem_object *obj)
 	trace_i915_gem_object_unbind(obj);
 
 	i915_gem_gtt_unbind_object(obj);
+	if (obj->has_aliasing_ppgtt_mapping) {
+		i915_ppgtt_unbind_object(dev_priv->mm.aliasing_ppgtt, obj);
+		obj->has_aliasing_ppgtt_mapping = 0;
+	}
+
 	i915_gem_object_put_pages_gtt(obj);
 
 	list_del_init(&obj->gtt_list);
@@ -2172,7 +2145,7 @@ i915_gem_flush_ring(struct intel_ring_buffer *ring,
 	return 0;
 }
 
-static int i915_ring_idle(struct intel_ring_buffer *ring)
+static int i915_ring_idle(struct intel_ring_buffer *ring, bool do_retire)
 {
 	int ret;
 
@@ -2186,18 +2159,18 @@ static int i915_ring_idle(struct intel_ring_buffer *ring)
 			return ret;
 	}
 
-	return i915_wait_request(ring, i915_gem_next_request_seqno(ring));
+	return i915_wait_request(ring, i915_gem_next_request_seqno(ring),
+				 do_retire);
 }
 
-int
-i915_gpu_idle(struct drm_device *dev)
+int i915_gpu_idle(struct drm_device *dev, bool do_retire)
 {
 	drm_i915_private_t *dev_priv = dev->dev_private;
 	int ret, i;
 
 	/* Flush everything onto the inactive list. */
 	for (i = 0; i < I915_NUM_RINGS; i++) {
-		ret = i915_ring_idle(&dev_priv->ring[i]);
+		ret = i915_ring_idle(&dev_priv->ring[i], do_retire);
 		if (ret)
 			return ret;
 	}
@@ -2400,7 +2373,8 @@ i915_gem_object_flush_fence(struct drm_i915_gem_object *obj,
 		if (!ring_passed_seqno(obj->last_fenced_ring,
 				       obj->last_fenced_seqno)) {
 			ret = i915_wait_request(obj->last_fenced_ring,
-						obj->last_fenced_seqno);
+						obj->last_fenced_seqno,
+						true);
 			if (ret)
 				return ret;
 		}
@@ -2432,6 +2406,8 @@ i915_gem_object_put_fence(struct drm_i915_gem_object *obj)
 
 	if (obj->fence_reg != I915_FENCE_REG_NONE) {
 		struct drm_i915_private *dev_priv = obj->base.dev->dev_private;
+
+		WARN_ON(dev_priv->fence_regs[obj->fence_reg].pin_count);
 		i915_gem_clear_fence_reg(obj->base.dev,
 					 &dev_priv->fence_regs[obj->fence_reg]);
 
@@ -2456,7 +2432,7 @@ i915_find_fence_reg(struct drm_device *dev,
 		if (!reg->obj)
 			return reg;
 
-		if (!reg->obj->pin_count)
+		if (!reg->pin_count)
 			avail = reg;
 	}
 
@@ -2466,7 +2442,7 @@ i915_find_fence_reg(struct drm_device *dev,
 	/* None available, try to steal one or wait for a user to finish */
 	avail = first = NULL;
 	list_for_each_entry(reg, &dev_priv->mm.fence_list, lru_list) {
-		if (reg->obj->pin_count)
+		if (reg->pin_count)
 			continue;
 
 		if (first == NULL)
@@ -2541,7 +2517,8 @@ i915_gem_object_get_fence(struct drm_i915_gem_object *obj,
 				if (!ring_passed_seqno(obj->last_fenced_ring,
 						       reg->setup_seqno)) {
 					ret = i915_wait_request(obj->last_fenced_ring,
-								reg->setup_seqno);
+								reg->setup_seqno,
+								true);
 					if (ret)
 						return ret;
 				}
@@ -2560,7 +2537,7 @@ i915_gem_object_get_fence(struct drm_i915_gem_object *obj,
 
 	reg = i915_find_fence_reg(dev, pipelined);
 	if (reg == NULL)
-		return -ENOSPC;
+		return -EDEADLK;
 
 	ret = i915_gem_object_flush_fence(obj, pipelined);
 	if (ret)
@@ -2660,6 +2637,7 @@ i915_gem_clear_fence_reg(struct drm_device *dev,
 	list_del_init(&reg->lru_list);
 	reg->obj = NULL;
 	reg->setup_seqno = 0;
+	reg->pin_count = 0;
 }
 
 /**
@@ -2946,6 +2924,8 @@ i915_gem_object_set_to_gtt_domain(struct drm_i915_gem_object *obj, bool write)
 int i915_gem_object_set_cache_level(struct drm_i915_gem_object *obj,
 				    enum i915_cache_level cache_level)
 {
+	struct drm_device *dev = obj->base.dev;
+	drm_i915_private_t *dev_priv = dev->dev_private;
 	int ret;
 
 	if (obj->cache_level == cache_level)
@@ -2974,6 +2954,9 @@ int i915_gem_object_set_cache_level(struct drm_i915_gem_object *obj,
 		}
 
 		i915_gem_gtt_rebind_object(obj, cache_level);
+		if (obj->has_aliasing_ppgtt_mapping)
+			i915_ppgtt_bind_object(dev_priv->mm.aliasing_ppgtt,
+					       obj, cache_level);
 	}
 
 	if (cache_level == I915_CACHE_NONE) {
@@ -3084,10 +3067,13 @@ i915_gem_object_finish_gpu(struct drm_i915_gem_object *obj)
 			return ret;
 	}
 
+	ret = i915_gem_object_wait_rendering(obj);
+	if (ret)
+		return ret;
+
 	/* Ensure that we invalidate the GPU's caches and TLBs. */
 	obj->base.read_domains &= ~I915_GEM_GPU_DOMAINS;
-
-	return i915_gem_object_wait_rendering(obj);
+	return 0;
 }
 
 /**
@@ -3619,8 +3605,8 @@ struct drm_i915_gem_object *i915_gem_alloc_object(struct drm_device *dev,
 	obj->base.write_domain = I915_GEM_DOMAIN_CPU;
 	obj->base.read_domains = I915_GEM_DOMAIN_CPU;
 
-	if (IS_GEN6(dev) || IS_GEN7(dev)) {
-		/* On Gen6, we can have the GPU use the LLC (the CPU
+	if (HAS_LLC(dev)) {
+		/* On some devices, we can have the GPU use the LLC (the CPU
 		 * cache) for about a 10% performance improvement
 		 * compared to uncached.  Graphics requests other than
 		 * display scanout are coherent with the CPU in
@@ -3710,7 +3696,7 @@ i915_gem_idle(struct drm_device *dev)
 		return 0;
 	}
 
-	ret = i915_gpu_idle(dev);
+	ret = i915_gpu_idle(dev, true);
 	if (ret) {
 		mutex_unlock(&dev->struct_mutex);
 		return ret;
@@ -3745,12 +3731,91 @@ i915_gem_idle(struct drm_device *dev)
 	return 0;
 }
 
+void i915_gem_init_swizzling(struct drm_device *dev)
+{
+	drm_i915_private_t *dev_priv = dev->dev_private;
+
+	if (INTEL_INFO(dev)->gen < 5 ||
+	    dev_priv->mm.bit_6_swizzle_x == I915_BIT_6_SWIZZLE_NONE)
+		return;
+
+	I915_WRITE(DISP_ARB_CTL, I915_READ(DISP_ARB_CTL) |
+				 DISP_TILE_SURFACE_SWIZZLING);
+
+	if (IS_GEN5(dev))
+		return;
+
+	I915_WRITE(TILECTL, I915_READ(TILECTL) | TILECTL_SWZCTL);
+	if (IS_GEN6(dev))
+		I915_WRITE(ARB_MODE, ARB_MODE_ENABLE(ARB_MODE_SWIZZLE_SNB));
+	else
+		I915_WRITE(ARB_MODE, ARB_MODE_ENABLE(ARB_MODE_SWIZZLE_IVB));
+}
+
+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;
+	uint32_t __iomem *pd_addr;
+	uint32_t pd_entry;
+	int i;
+
+	if (!dev_priv->mm.aliasing_ppgtt)
+		return;
+
+
+	pd_addr = dev_priv->mm.gtt->gtt + ppgtt->pd_offset/sizeof(uint32_t);
+	for (i = 0; i < ppgtt->num_pd_entries; i++) {
+		dma_addr_t pt_addr;
+
+		if (dev_priv->mm.gtt->needs_dmar)
+			pt_addr = ppgtt->pt_dma_addr[i];
+		else
+			pt_addr = page_to_phys(ppgtt->pt_pages[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 = I915_READ(GAM_ECOCHK);
+		I915_WRITE(GAM_ECOCHK, ecochk | ECOCHK_SNB_BIT |
+				       ECOCHK_PPGTT_CACHE64B);
+		I915_WRITE(GFX_MODE, GFX_MODE_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 (i = 0; i < I915_NUM_RINGS; i++) {
+		ring = &dev_priv->ring[i];
+
+		if (INTEL_INFO(dev)->gen >= 7)
+			I915_WRITE(RING_MODE_GEN7(ring),
+				   GFX_MODE_ENABLE(GFX_PPGTT_ENABLE));
+
+		I915_WRITE(RING_PP_DIR_DCLV(ring), PP_DIR_DCLV_2G);
+		I915_WRITE(RING_PP_DIR_BASE(ring), pd_offset);
+	}
+}
+
 int
-i915_gem_init_ringbuffer(struct drm_device *dev)
+i915_gem_init_hw(struct drm_device *dev)
 {
 	drm_i915_private_t *dev_priv = dev->dev_private;
 	int ret;
 
+	i915_gem_init_swizzling(dev);
+
 	ret = intel_init_render_ring_buffer(dev);
 	if (ret)
 		return ret;
@@ -3769,6 +3834,8 @@ i915_gem_init_ringbuffer(struct drm_device *dev)
 
 	dev_priv->next_seqno = 1;
 
+	i915_gem_init_ppgtt(dev);
+
 	return 0;
 
 cleanup_bsd_ring:
@@ -3806,7 +3873,7 @@ i915_gem_entervt_ioctl(struct drm_device *dev, void *data,
 	mutex_lock(&dev->struct_mutex);
 	dev_priv->mm.suspended = 0;
 
-	ret = i915_gem_init_ringbuffer(dev);
+	ret = i915_gem_init_hw(dev);
 	if (ret != 0) {
 		mutex_unlock(&dev->struct_mutex);
 		return ret;
@@ -4201,7 +4268,7 @@ rescan:
 		 * This has a dramatic impact to reduce the number of
 		 * OOM-killer events whilst running the GPU aggressively.
 		 */
-		if (i915_gpu_idle(dev) == 0)
+		if (i915_gpu_idle(dev, true) == 0)
 			goto rescan;
 	}
 	mutex_unlock(&dev->struct_mutex);
diff --git a/drivers/gpu/drm/i915/i915_gem_evict.c b/drivers/gpu/drm/i915/i915_gem_evict.c
index ead5d00f91b0..21a82710f4b2 100644
--- a/drivers/gpu/drm/i915/i915_gem_evict.c
+++ b/drivers/gpu/drm/i915/i915_gem_evict.c
@@ -36,7 +36,6 @@ static bool
 mark_free(struct drm_i915_gem_object *obj, struct list_head *unwind)
 {
 	list_add(&obj->exec_list, unwind);
-	drm_gem_object_reference(&obj->base);
 	return drm_mm_scan_add_block(obj->gtt_space);
 }
 
@@ -49,21 +48,6 @@ i915_gem_evict_something(struct drm_device *dev, int min_size,
 	struct drm_i915_gem_object *obj;
 	int ret = 0;
 
-	i915_gem_retire_requests(dev);
-
-	/* Re-check for free space after retiring requests */
-	if (mappable) {
-		if (drm_mm_search_free_in_range(&dev_priv->mm.gtt_space,
-						min_size, alignment, 0,
-						dev_priv->mm.gtt_mappable_end,
-						0))
-			return 0;
-	} else {
-		if (drm_mm_search_free(&dev_priv->mm.gtt_space,
-				       min_size, alignment, 0))
-			return 0;
-	}
-
 	trace_i915_gem_evict(dev, min_size, alignment, mappable);
 
 	/*
@@ -139,7 +123,6 @@ i915_gem_evict_something(struct drm_device *dev, int min_size,
 		BUG_ON(ret);
 
 		list_del_init(&obj->exec_list);
-		drm_gem_object_unreference(&obj->base);
 	}
 
 	/* We expect the caller to unpin, evict all and try again, or give up.
@@ -158,10 +141,10 @@ found:
 				       exec_list);
 		if (drm_mm_scan_remove_block(obj->gtt_space)) {
 			list_move(&obj->exec_list, &eviction_list);
+			drm_gem_object_reference(&obj->base);
 			continue;
 		}
 		list_del_init(&obj->exec_list);
-		drm_gem_object_unreference(&obj->base);
 	}
 
 	/* Unbinding will emit any required flushes */
@@ -195,7 +178,7 @@ i915_gem_evict_everything(struct drm_device *dev, bool purgeable_only)
 	trace_i915_gem_evict_everything(dev, purgeable_only);
 
 	/* Flush everything (on to the inactive lists) and evict */
-	ret = i915_gpu_idle(dev);
+	ret = i915_gpu_idle(dev, true);
 	if (ret)
 		return ret;
 
diff --git a/drivers/gpu/drm/i915/i915_gem_execbuffer.c b/drivers/gpu/drm/i915/i915_gem_execbuffer.c
index 65e1f0043f9d..f51a696486cb 100644
--- a/drivers/gpu/drm/i915/i915_gem_execbuffer.c
+++ b/drivers/gpu/drm/i915/i915_gem_execbuffer.c
@@ -203,9 +203,9 @@ i915_gem_object_set_to_gpu_domain(struct drm_i915_gem_object *obj,
 	cd->invalidate_domains |= invalidate_domains;
 	cd->flush_domains |= flush_domains;
 	if (flush_domains & I915_GEM_GPU_DOMAINS)
-		cd->flush_rings |= obj->ring->id;
+		cd->flush_rings |= intel_ring_flag(obj->ring);
 	if (invalidate_domains & I915_GEM_GPU_DOMAINS)
-		cd->flush_rings |= ring->id;
+		cd->flush_rings |= intel_ring_flag(ring);
 }
 
 struct eb_objects {
@@ -287,14 +287,14 @@ i915_gem_execbuffer_relocate_entry(struct drm_i915_gem_object *obj,
 	 * exec_object list, so it should have a GTT space bound by now.
 	 */
 	if (unlikely(target_offset == 0)) {
-		DRM_ERROR("No GTT space found for object %d\n",
+		DRM_DEBUG("No GTT space found for object %d\n",
 			  reloc->target_handle);
 		return ret;
 	}
 
 	/* Validate that the target is in a valid r/w GPU domain */
 	if (unlikely(reloc->write_domain & (reloc->write_domain - 1))) {
-		DRM_ERROR("reloc with multiple write domains: "
+		DRM_DEBUG("reloc with multiple write domains: "
 			  "obj %p target %d offset %d "
 			  "read %08x write %08x",
 			  obj, reloc->target_handle,
@@ -303,8 +303,9 @@ i915_gem_execbuffer_relocate_entry(struct drm_i915_gem_object *obj,
 			  reloc->write_domain);
 		return ret;
 	}
-	if (unlikely((reloc->write_domain | reloc->read_domains) & I915_GEM_DOMAIN_CPU)) {
-		DRM_ERROR("reloc with read/write CPU domains: "
+	if (unlikely((reloc->write_domain | reloc->read_domains)
+		     & ~I915_GEM_GPU_DOMAINS)) {
+		DRM_DEBUG("reloc with read/write non-GPU domains: "
 			  "obj %p target %d offset %d "
 			  "read %08x write %08x",
 			  obj, reloc->target_handle,
@@ -315,7 +316,7 @@ i915_gem_execbuffer_relocate_entry(struct drm_i915_gem_object *obj,
 	}
 	if (unlikely(reloc->write_domain && target_obj->pending_write_domain &&
 		     reloc->write_domain != target_obj->pending_write_domain)) {
-		DRM_ERROR("Write domain conflict: "
+		DRM_DEBUG("Write domain conflict: "
 			  "obj %p target %d offset %d "
 			  "new %08x old %08x\n",
 			  obj, reloc->target_handle,
@@ -336,7 +337,7 @@ i915_gem_execbuffer_relocate_entry(struct drm_i915_gem_object *obj,
 
 	/* Check that the relocation address is valid... */
 	if (unlikely(reloc->offset > obj->base.size - 4)) {
-		DRM_ERROR("Relocation beyond object bounds: "
+		DRM_DEBUG("Relocation beyond object bounds: "
 			  "obj %p target %d offset %d size %d.\n",
 			  obj, reloc->target_handle,
 			  (int) reloc->offset,
@@ -344,7 +345,7 @@ i915_gem_execbuffer_relocate_entry(struct drm_i915_gem_object *obj,
 		return ret;
 	}
 	if (unlikely(reloc->offset & 3)) {
-		DRM_ERROR("Relocation not 4-byte aligned: "
+		DRM_DEBUG("Relocation not 4-byte aligned: "
 			  "obj %p target %d offset %d.\n",
 			  obj, reloc->target_handle,
 			  (int) reloc->offset);
@@ -461,11 +462,60 @@ i915_gem_execbuffer_relocate(struct drm_device *dev,
 	return ret;
 }
 
+#define  __EXEC_OBJECT_HAS_FENCE (1<<31)
+
+static int
+pin_and_fence_object(struct drm_i915_gem_object *obj,
+		     struct intel_ring_buffer *ring)
+{
+	struct drm_i915_gem_exec_object2 *entry = obj->exec_entry;
+	bool has_fenced_gpu_access = INTEL_INFO(ring->dev)->gen < 4;
+	bool need_fence, need_mappable;
+	int ret;
+
+	need_fence =
+		has_fenced_gpu_access &&
+		entry->flags & EXEC_OBJECT_NEEDS_FENCE &&
+		obj->tiling_mode != I915_TILING_NONE;
+	need_mappable =
+		entry->relocation_count ? true : need_fence;
+
+	ret = i915_gem_object_pin(obj, entry->alignment, need_mappable);
+	if (ret)
+		return ret;
+
+	if (has_fenced_gpu_access) {
+		if (entry->flags & EXEC_OBJECT_NEEDS_FENCE) {
+			if (obj->tiling_mode) {
+				ret = i915_gem_object_get_fence(obj, ring);
+				if (ret)
+					goto err_unpin;
+
+				entry->flags |= __EXEC_OBJECT_HAS_FENCE;
+				i915_gem_object_pin_fence(obj);
+			} else {
+				ret = i915_gem_object_put_fence(obj);
+				if (ret)
+					goto err_unpin;
+			}
+			obj->pending_fenced_gpu_access = true;
+		}
+	}
+
+	entry->offset = obj->gtt_offset;
+	return 0;
+
+err_unpin:
+	i915_gem_object_unpin(obj);
+	return ret;
+}
+
 static int
 i915_gem_execbuffer_reserve(struct intel_ring_buffer *ring,
 			    struct drm_file *file,
 			    struct list_head *objects)
 {
+	drm_i915_private_t *dev_priv = ring->dev->dev_private;
 	struct drm_i915_gem_object *obj;
 	int ret, retry;
 	bool has_fenced_gpu_access = INTEL_INFO(ring->dev)->gen < 4;
@@ -518,6 +568,7 @@ i915_gem_execbuffer_reserve(struct intel_ring_buffer *ring,
 		list_for_each_entry(obj, objects, exec_list) {
 			struct drm_i915_gem_exec_object2 *entry = obj->exec_entry;
 			bool need_fence, need_mappable;
+
 			if (!obj->gtt_space)
 				continue;
 
@@ -532,58 +583,55 @@ i915_gem_execbuffer_reserve(struct intel_ring_buffer *ring,
 			    (need_mappable && !obj->map_and_fenceable))
 				ret = i915_gem_object_unbind(obj);
 			else
-				ret = i915_gem_object_pin(obj,
-							  entry->alignment,
-							  need_mappable);
+				ret = pin_and_fence_object(obj, ring);
 			if (ret)
 				goto err;
-
-			entry++;
 		}
 
 		/* Bind fresh objects */
 		list_for_each_entry(obj, objects, exec_list) {
-			struct drm_i915_gem_exec_object2 *entry = obj->exec_entry;
-			bool need_fence;
+			if (obj->gtt_space)
+				continue;
 
-			need_fence =
-				has_fenced_gpu_access &&
-				entry->flags & EXEC_OBJECT_NEEDS_FENCE &&
-				obj->tiling_mode != I915_TILING_NONE;
+			ret = pin_and_fence_object(obj, ring);
+			if (ret) {
+				int ret_ignore;
+
+				/* This can potentially raise a harmless
+				 * -EINVAL if we failed to bind in the above
+				 * call. It cannot raise -EINTR since we know
+				 * that the bo is freshly bound and so will
+				 * not need to be flushed or waited upon.
+				 */
+				ret_ignore = i915_gem_object_unbind(obj);
+				(void)ret_ignore;
+				WARN_ON(obj->gtt_space);
+				break;
+			}
+		}
 
-			if (!obj->gtt_space) {
-				bool need_mappable =
-					entry->relocation_count ? true : need_fence;
+		/* Decrement pin count for bound objects */
+		list_for_each_entry(obj, objects, exec_list) {
+			struct drm_i915_gem_exec_object2 *entry;
 
-				ret = i915_gem_object_pin(obj,
-							  entry->alignment,
-							  need_mappable);
-				if (ret)
-					break;
-			}
+			if (!obj->gtt_space)
+				continue;
 
-			if (has_fenced_gpu_access) {
-				if (need_fence) {
-					ret = i915_gem_object_get_fence(obj, ring);
-					if (ret)
-						break;
-				} else if (entry->flags & EXEC_OBJECT_NEEDS_FENCE &&
-					   obj->tiling_mode == I915_TILING_NONE) {
-					/* XXX pipelined! */
-					ret = i915_gem_object_put_fence(obj);
-					if (ret)
-						break;
-				}
-				obj->pending_fenced_gpu_access = need_fence;
+			entry = obj->exec_entry;
+			if (entry->flags & __EXEC_OBJECT_HAS_FENCE) {
+				i915_gem_object_unpin_fence(obj);
+				entry->flags &= ~__EXEC_OBJECT_HAS_FENCE;
 			}
 
-			entry->offset = obj->gtt_offset;
-		}
+			i915_gem_object_unpin(obj);
 
-		/* Decrement pin count for bound objects */
-		list_for_each_entry(obj, objects, exec_list) {
-			if (obj->gtt_space)
-				i915_gem_object_unpin(obj);
+			/* ... and ensure ppgtt mapping exist if needed. */
+			if (dev_priv->mm.aliasing_ppgtt && !obj->has_aliasing_ppgtt_mapping) {
+				i915_ppgtt_bind_object(dev_priv->mm.aliasing_ppgtt,
+						       obj, obj->cache_level);
+
+				obj->has_aliasing_ppgtt_mapping = 1;
+			}
 		}
 
 		if (ret != -ENOSPC || retry > 1)
@@ -600,16 +648,19 @@ i915_gem_execbuffer_reserve(struct intel_ring_buffer *ring,
 	} while (1);
 
 err:
-	obj = list_entry(obj->exec_list.prev,
-			 struct drm_i915_gem_object,
-			 exec_list);
-	while (objects != &obj->exec_list) {
-		if (obj->gtt_space)
-			i915_gem_object_unpin(obj);
+	list_for_each_entry_continue_reverse(obj, objects, exec_list) {
+		struct drm_i915_gem_exec_object2 *entry;
+
+		if (!obj->gtt_space)
+			continue;
+
+		entry = obj->exec_entry;
+		if (entry->flags & __EXEC_OBJECT_HAS_FENCE) {
+			i915_gem_object_unpin_fence(obj);
+			entry->flags &= ~__EXEC_OBJECT_HAS_FENCE;
+		}
 
-		obj = list_entry(obj->exec_list.prev,
-				 struct drm_i915_gem_object,
-				 exec_list);
+		i915_gem_object_unpin(obj);
 	}
 
 	return ret;
@@ -682,7 +733,7 @@ i915_gem_execbuffer_relocate_slow(struct drm_device *dev,
 		obj = to_intel_bo(drm_gem_object_lookup(dev, file,
 							exec[i].handle));
 		if (&obj->base == NULL) {
-			DRM_ERROR("Invalid object handle %d at index %d\n",
+			DRM_DEBUG("Invalid object handle %d at index %d\n",
 				   exec[i].handle, i);
 			ret = -ENOENT;
 			goto err;
@@ -1013,7 +1064,7 @@ i915_gem_do_execbuffer(struct drm_device *dev, void *data,
 	int ret, mode, i;
 
 	if (!i915_gem_check_execbuffer(args)) {
-		DRM_ERROR("execbuf with invalid offset/length\n");
+		DRM_DEBUG("execbuf with invalid offset/length\n");
 		return -EINVAL;
 	}
 
@@ -1028,20 +1079,20 @@ i915_gem_do_execbuffer(struct drm_device *dev, void *data,
 		break;
 	case I915_EXEC_BSD:
 		if (!HAS_BSD(dev)) {
-			DRM_ERROR("execbuf with invalid ring (BSD)\n");
+			DRM_DEBUG("execbuf with invalid ring (BSD)\n");
 			return -EINVAL;
 		}
 		ring = &dev_priv->ring[VCS];
 		break;
 	case I915_EXEC_BLT:
 		if (!HAS_BLT(dev)) {
-			DRM_ERROR("execbuf with invalid ring (BLT)\n");
+			DRM_DEBUG("execbuf with invalid ring (BLT)\n");
 			return -EINVAL;
 		}
 		ring = &dev_priv->ring[BCS];
 		break;
 	default:
-		DRM_ERROR("execbuf with unknown ring: %d\n",
+		DRM_DEBUG("execbuf with unknown ring: %d\n",
 			  (int)(args->flags & I915_EXEC_RING_MASK));
 		return -EINVAL;
 	}
@@ -1067,18 +1118,18 @@ i915_gem_do_execbuffer(struct drm_device *dev, void *data,
 		}
 		break;
 	default:
-		DRM_ERROR("execbuf with unknown constants: %d\n", mode);
+		DRM_DEBUG("execbuf with unknown constants: %d\n", mode);
 		return -EINVAL;
 	}
 
 	if (args->buffer_count < 1) {
-		DRM_ERROR("execbuf with %d buffers\n", args->buffer_count);
+		DRM_DEBUG("execbuf with %d buffers\n", args->buffer_count);
 		return -EINVAL;
 	}
 
 	if (args->num_cliprects != 0) {
 		if (ring != &dev_priv->ring[RCS]) {
-			DRM_ERROR("clip rectangles are only valid with the render ring\n");
+			DRM_DEBUG("clip rectangles are only valid with the render ring\n");
 			return -EINVAL;
 		}
 
@@ -1123,7 +1174,7 @@ i915_gem_do_execbuffer(struct drm_device *dev, void *data,
 		obj = to_intel_bo(drm_gem_object_lookup(dev, file,
 							exec[i].handle));
 		if (&obj->base == NULL) {
-			DRM_ERROR("Invalid object handle %d at index %d\n",
+			DRM_DEBUG("Invalid object handle %d at index %d\n",
 				   exec[i].handle, i);
 			/* prevent error path from reading uninitialized data */
 			ret = -ENOENT;
@@ -1131,7 +1182,7 @@ i915_gem_do_execbuffer(struct drm_device *dev, void *data,
 		}
 
 		if (!list_empty(&obj->exec_list)) {
-			DRM_ERROR("Object %p [handle %d, index %d] appears more than once in object list\n",
+			DRM_DEBUG("Object %p [handle %d, index %d] appears more than once in object list\n",
 				   obj, exec[i].handle, i);
 			ret = -EINVAL;
 			goto err;
@@ -1169,7 +1220,7 @@ i915_gem_do_execbuffer(struct drm_device *dev, void *data,
 
 	/* Set the pending read domains for the batch buffer to COMMAND */
 	if (batch_obj->base.pending_write_domain) {
-		DRM_ERROR("Attempting to use self-modifying batch buffer\n");
+		DRM_DEBUG("Attempting to use self-modifying batch buffer\n");
 		ret = -EINVAL;
 		goto err;
 	}
@@ -1186,7 +1237,7 @@ i915_gem_do_execbuffer(struct drm_device *dev, void *data,
 			 * so every billion or so execbuffers, we need to stall
 			 * the GPU in order to reset the counters.
 			 */
-			ret = i915_gpu_idle(dev);
+			ret = i915_gpu_idle(dev, true);
 			if (ret)
 				goto err;
 
@@ -1274,7 +1325,7 @@ i915_gem_execbuffer(struct drm_device *dev, void *data,
 	int ret, i;
 
 	if (args->buffer_count < 1) {
-		DRM_ERROR("execbuf with %d buffers\n", args->buffer_count);
+		DRM_DEBUG("execbuf with %d buffers\n", args->buffer_count);
 		return -EINVAL;
 	}
 
@@ -1282,7 +1333,7 @@ i915_gem_execbuffer(struct drm_device *dev, void *data,
 	exec_list = drm_malloc_ab(sizeof(*exec_list), args->buffer_count);
 	exec2_list = drm_malloc_ab(sizeof(*exec2_list), args->buffer_count);
 	if (exec_list == NULL || exec2_list == NULL) {
-		DRM_ERROR("Failed to allocate exec list for %d buffers\n",
+		DRM_DEBUG("Failed to allocate exec list for %d buffers\n",
 			  args->buffer_count);
 		drm_free_large(exec_list);
 		drm_free_large(exec2_list);
@@ -1293,7 +1344,7 @@ i915_gem_execbuffer(struct drm_device *dev, void *data,
 			     (uintptr_t) args->buffers_ptr,
 			     sizeof(*exec_list) * args->buffer_count);
 	if (ret != 0) {
-		DRM_ERROR("copy %d exec entries failed %d\n",
+		DRM_DEBUG("copy %d exec entries failed %d\n",
 			  args->buffer_count, ret);
 		drm_free_large(exec_list);
 		drm_free_large(exec2_list);
@@ -1334,7 +1385,7 @@ i915_gem_execbuffer(struct drm_device *dev, void *data,
 				   sizeof(*exec_list) * args->buffer_count);
 		if (ret) {
 			ret = -EFAULT;
-			DRM_ERROR("failed to copy %d exec entries "
+			DRM_DEBUG("failed to copy %d exec entries "
 				  "back to user (%d)\n",
 				  args->buffer_count, ret);
 		}
@@ -1354,7 +1405,7 @@ i915_gem_execbuffer2(struct drm_device *dev, void *data,
 	int ret;
 
 	if (args->buffer_count < 1) {
-		DRM_ERROR("execbuf2 with %d buffers\n", args->buffer_count);
+		DRM_DEBUG("execbuf2 with %d buffers\n", args->buffer_count);
 		return -EINVAL;
 	}
 
@@ -1364,7 +1415,7 @@ i915_gem_execbuffer2(struct drm_device *dev, void *data,
 		exec2_list = drm_malloc_ab(sizeof(*exec2_list),
 					   args->buffer_count);
 	if (exec2_list == NULL) {
-		DRM_ERROR("Failed to allocate exec list for %d buffers\n",
+		DRM_DEBUG("Failed to allocate exec list for %d buffers\n",
 			  args->buffer_count);
 		return -ENOMEM;
 	}
@@ -1373,7 +1424,7 @@ i915_gem_execbuffer2(struct drm_device *dev, void *data,
 			     (uintptr_t) args->buffers_ptr,
 			     sizeof(*exec2_list) * args->buffer_count);
 	if (ret != 0) {
-		DRM_ERROR("copy %d exec entries failed %d\n",
+		DRM_DEBUG("copy %d exec entries failed %d\n",
 			  args->buffer_count, ret);
 		drm_free_large(exec2_list);
 		return -EFAULT;
@@ -1388,7 +1439,7 @@ i915_gem_execbuffer2(struct drm_device *dev, void *data,
 				   sizeof(*exec2_list) * args->buffer_count);
 		if (ret) {
 			ret = -EFAULT;
-			DRM_ERROR("failed to copy %d exec entries "
+			DRM_DEBUG("failed to copy %d exec entries "
 				  "back to user (%d)\n",
 				  args->buffer_count, ret);
 		}
diff --git a/drivers/gpu/drm/i915/i915_gem_gtt.c b/drivers/gpu/drm/i915/i915_gem_gtt.c
index 6042c5e6d278..a135c61f4119 100644
--- a/drivers/gpu/drm/i915/i915_gem_gtt.c
+++ b/drivers/gpu/drm/i915/i915_gem_gtt.c
@@ -29,6 +29,270 @@
 #include "i915_trace.h"
 #include "intel_drv.h"
 
+/* PPGTT support for Sandybdrige/Gen6 and later */
+static void i915_ppgtt_clear_range(struct i915_hw_ppgtt *ppgtt,
+				   unsigned first_entry,
+				   unsigned num_entries)
+{
+	uint32_t *pt_vaddr;
+	uint32_t scratch_pte;
+	unsigned act_pd = first_entry / I915_PPGTT_PT_ENTRIES;
+	unsigned first_pte = first_entry % I915_PPGTT_PT_ENTRIES;
+	unsigned last_pte, i;
+
+	scratch_pte = GEN6_PTE_ADDR_ENCODE(ppgtt->scratch_page_dma_addr);
+	scratch_pte |= GEN6_PTE_VALID | GEN6_PTE_CACHE_LLC;
+
+	while (num_entries) {
+		last_pte = first_pte + num_entries;
+		if (last_pte > I915_PPGTT_PT_ENTRIES)
+			last_pte = I915_PPGTT_PT_ENTRIES;
+
+		pt_vaddr = kmap_atomic(ppgtt->pt_pages[act_pd]);
+
+		for (i = first_pte; i < last_pte; i++)
+			pt_vaddr[i] = scratch_pte;
+
+		kunmap_atomic(pt_vaddr);
+
+		num_entries -= last_pte - first_pte;
+		first_pte = 0;
+		act_pd++;
+	}
+}
+
+int i915_gem_init_aliasing_ppgtt(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct i915_hw_ppgtt *ppgtt;
+	unsigned first_pd_entry_in_global_pt;
+	int i;
+	int ret = -ENOMEM;
+
+	/* 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 = 512*1024 - I915_PPGTT_PD_ENTRIES;
+
+	ppgtt = kzalloc(sizeof(*ppgtt), GFP_KERNEL);
+	if (!ppgtt)
+		return ret;
+
+	ppgtt->num_pd_entries = I915_PPGTT_PD_ENTRIES;
+	ppgtt->pt_pages = kzalloc(sizeof(struct page *)*ppgtt->num_pd_entries,
+				  GFP_KERNEL);
+	if (!ppgtt->pt_pages)
+		goto err_ppgtt;
+
+	for (i = 0; i < ppgtt->num_pd_entries; i++) {
+		ppgtt->pt_pages[i] = alloc_page(GFP_KERNEL);
+		if (!ppgtt->pt_pages[i])
+			goto err_pt_alloc;
+	}
+
+	if (dev_priv->mm.gtt->needs_dmar) {
+		ppgtt->pt_dma_addr = kzalloc(sizeof(dma_addr_t)
+						*ppgtt->num_pd_entries,
+					     GFP_KERNEL);
+		if (!ppgtt->pt_dma_addr)
+			goto err_pt_alloc;
+	}
+
+	for (i = 0; i < ppgtt->num_pd_entries; i++) {
+		dma_addr_t pt_addr;
+		if (dev_priv->mm.gtt->needs_dmar) {
+			pt_addr = pci_map_page(dev->pdev, ppgtt->pt_pages[i],
+					       0, 4096,
+					       PCI_DMA_BIDIRECTIONAL);
+
+			if (pci_dma_mapping_error(dev->pdev,
+						  pt_addr)) {
+				ret = -EIO;
+				goto err_pd_pin;
+
+			}
+			ppgtt->pt_dma_addr[i] = pt_addr;
+		} else
+			pt_addr = page_to_phys(ppgtt->pt_pages[i]);
+	}
+
+	ppgtt->scratch_page_dma_addr = dev_priv->mm.gtt->scratch_page_dma;
+
+	i915_ppgtt_clear_range(ppgtt, 0,
+			       ppgtt->num_pd_entries*I915_PPGTT_PT_ENTRIES);
+
+	ppgtt->pd_offset = (first_pd_entry_in_global_pt)*sizeof(uint32_t);
+
+	dev_priv->mm.aliasing_ppgtt = ppgtt;
+
+	return 0;
+
+err_pd_pin:
+	if (ppgtt->pt_dma_addr) {
+		for (i--; i >= 0; i--)
+			pci_unmap_page(dev->pdev, ppgtt->pt_dma_addr[i],
+				       4096, PCI_DMA_BIDIRECTIONAL);
+	}
+err_pt_alloc:
+	kfree(ppgtt->pt_dma_addr);
+	for (i = 0; i < ppgtt->num_pd_entries; i++) {
+		if (ppgtt->pt_pages[i])
+			__free_page(ppgtt->pt_pages[i]);
+	}
+	kfree(ppgtt->pt_pages);
+err_ppgtt:
+	kfree(ppgtt);
+
+	return ret;
+}
+
+void i915_gem_cleanup_aliasing_ppgtt(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct i915_hw_ppgtt *ppgtt = dev_priv->mm.aliasing_ppgtt;
+	int i;
+
+	if (!ppgtt)
+		return;
+
+	if (ppgtt->pt_dma_addr) {
+		for (i = 0; i < ppgtt->num_pd_entries; i++)
+			pci_unmap_page(dev->pdev, ppgtt->pt_dma_addr[i],
+				       4096, PCI_DMA_BIDIRECTIONAL);
+	}
+
+	kfree(ppgtt->pt_dma_addr);
+	for (i = 0; i < ppgtt->num_pd_entries; i++)
+		__free_page(ppgtt->pt_pages[i]);
+	kfree(ppgtt->pt_pages);
+	kfree(ppgtt);
+}
+
+static void i915_ppgtt_insert_sg_entries(struct i915_hw_ppgtt *ppgtt,
+					 struct scatterlist *sg_list,
+					 unsigned sg_len,
+					 unsigned first_entry,
+					 uint32_t pte_flags)
+{
+	uint32_t *pt_vaddr, pte;
+	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 = sg_list;
+	i = 0;
+	segment_len = sg_dma_len(sg) >> PAGE_SHIFT;
+	m = 0;
+
+	while (i < sg_len) {
+		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);
+			pte = GEN6_PTE_ADDR_ENCODE(page_addr);
+			pt_vaddr[j] = pte | pte_flags;
+
+			/* grab the next page */
+			m++;
+			if (m == segment_len) {
+				sg = sg_next(sg);
+				i++;
+				if (i == sg_len)
+					break;
+
+				segment_len = sg_dma_len(sg) >> PAGE_SHIFT;
+				m = 0;
+			}
+		}
+
+		kunmap_atomic(pt_vaddr);
+
+		first_pte = 0;
+		act_pd++;
+	}
+}
+
+static void i915_ppgtt_insert_pages(struct i915_hw_ppgtt *ppgtt,
+				    unsigned first_entry, unsigned num_entries,
+				    struct page **pages, uint32_t pte_flags)
+{
+	uint32_t *pt_vaddr, pte;
+	unsigned act_pd = first_entry / I915_PPGTT_PT_ENTRIES;
+	unsigned first_pte = first_entry % I915_PPGTT_PT_ENTRIES;
+	unsigned last_pte, i;
+	dma_addr_t page_addr;
+
+	while (num_entries) {
+		last_pte = first_pte + num_entries;
+		last_pte = min_t(unsigned, last_pte, I915_PPGTT_PT_ENTRIES);
+
+		pt_vaddr = kmap_atomic(ppgtt->pt_pages[act_pd]);
+
+		for (i = first_pte; i < last_pte; i++) {
+			page_addr = page_to_phys(*pages);
+			pte = GEN6_PTE_ADDR_ENCODE(page_addr);
+			pt_vaddr[i] = pte | pte_flags;
+
+			pages++;
+		}
+
+		kunmap_atomic(pt_vaddr);
+
+		num_entries -= last_pte - first_pte;
+		first_pte = 0;
+		act_pd++;
+	}
+}
+
+void i915_ppgtt_bind_object(struct i915_hw_ppgtt *ppgtt,
+			    struct drm_i915_gem_object *obj,
+			    enum i915_cache_level cache_level)
+{
+	struct drm_device *dev = obj->base.dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	uint32_t pte_flags = GEN6_PTE_VALID;
+
+	switch (cache_level) {
+	case I915_CACHE_LLC_MLC:
+		pte_flags |= GEN6_PTE_CACHE_LLC_MLC;
+		break;
+	case I915_CACHE_LLC:
+		pte_flags |= GEN6_PTE_CACHE_LLC;
+		break;
+	case I915_CACHE_NONE:
+		pte_flags |= GEN6_PTE_UNCACHED;
+		break;
+	default:
+		BUG();
+	}
+
+	if (dev_priv->mm.gtt->needs_dmar) {
+		BUG_ON(!obj->sg_list);
+
+		i915_ppgtt_insert_sg_entries(ppgtt,
+					     obj->sg_list,
+					     obj->num_sg,
+					     obj->gtt_space->start >> PAGE_SHIFT,
+					     pte_flags);
+	} else
+		i915_ppgtt_insert_pages(ppgtt,
+					obj->gtt_space->start >> PAGE_SHIFT,
+					obj->base.size >> PAGE_SHIFT,
+					obj->pages,
+					pte_flags);
+}
+
+void i915_ppgtt_unbind_object(struct i915_hw_ppgtt *ppgtt,
+			      struct drm_i915_gem_object *obj)
+{
+	i915_ppgtt_clear_range(ppgtt,
+			       obj->gtt_space->start >> PAGE_SHIFT,
+			       obj->base.size >> PAGE_SHIFT);
+}
+
 /* XXX kill agp_type! */
 static unsigned int cache_level_to_agp_type(struct drm_device *dev,
 					    enum i915_cache_level cache_level)
@@ -55,7 +319,7 @@ static bool do_idling(struct drm_i915_private *dev_priv)
 
 	if (unlikely(dev_priv->mm.gtt->do_idle_maps)) {
 		dev_priv->mm.interruptible = false;
-		if (i915_gpu_idle(dev_priv->dev)) {
+		if (i915_gpu_idle(dev_priv->dev, false)) {
 			DRM_ERROR("Couldn't idle GPU\n");
 			/* Wait a bit, in hopes it avoids the hang */
 			udelay(10);
diff --git a/drivers/gpu/drm/i915/i915_gem_tiling.c b/drivers/gpu/drm/i915/i915_gem_tiling.c
index 31d334d9d9da..1a9306665987 100644
--- a/drivers/gpu/drm/i915/i915_gem_tiling.c
+++ b/drivers/gpu/drm/i915/i915_gem_tiling.c
@@ -93,8 +93,23 @@ i915_gem_detect_bit_6_swizzle(struct drm_device *dev)
 	uint32_t swizzle_y = I915_BIT_6_SWIZZLE_UNKNOWN;
 
 	if (INTEL_INFO(dev)->gen >= 6) {
-		swizzle_x = I915_BIT_6_SWIZZLE_NONE;
-		swizzle_y = I915_BIT_6_SWIZZLE_NONE;
+		uint32_t dimm_c0, dimm_c1;
+		dimm_c0 = I915_READ(MAD_DIMM_C0);
+		dimm_c1 = I915_READ(MAD_DIMM_C1);
+		dimm_c0 &= MAD_DIMM_A_SIZE_MASK | MAD_DIMM_B_SIZE_MASK;
+		dimm_c1 &= MAD_DIMM_A_SIZE_MASK | MAD_DIMM_B_SIZE_MASK;
+		/* Enable swizzling when the channels are populated with
+		 * identically sized dimms. We don't need to check the 3rd
+		 * channel because no cpu with gpu attached ships in that
+		 * configuration. Also, swizzling only makes sense for 2
+		 * channels anyway. */
+		if (dimm_c0 == dimm_c1) {
+			swizzle_x = I915_BIT_6_SWIZZLE_9_10;
+			swizzle_y = I915_BIT_6_SWIZZLE_9;
+		} else {
+			swizzle_x = I915_BIT_6_SWIZZLE_NONE;
+			swizzle_y = I915_BIT_6_SWIZZLE_NONE;
+		}
 	} else if (IS_GEN5(dev)) {
 		/* On Ironlake whatever DRAM config, GPU always do
 		 * same swizzling setup.
@@ -107,10 +122,10 @@ i915_gem_detect_bit_6_swizzle(struct drm_device *dev)
 		 */
 		swizzle_x = I915_BIT_6_SWIZZLE_NONE;
 		swizzle_y = I915_BIT_6_SWIZZLE_NONE;
-	} else if (IS_MOBILE(dev)) {
+	} else if (IS_MOBILE(dev) || (IS_GEN3(dev) && !IS_G33(dev))) {
 		uint32_t dcc;
 
-		/* On mobile 9xx chipsets, channel interleave by the CPU is
+		/* On 9xx chipsets, channel interleave by the CPU is
 		 * determined by DCC.  For single-channel, neither the CPU
 		 * nor the GPU do swizzling.  For dual channel interleaved,
 		 * the GPU's interleave is bit 9 and 10 for X tiled, and bit
diff --git a/drivers/gpu/drm/i915/i915_irq.c b/drivers/gpu/drm/i915/i915_irq.c
index 5bd4361ea84d..afd4e03e337e 100644
--- a/drivers/gpu/drm/i915/i915_irq.c
+++ b/drivers/gpu/drm/i915/i915_irq.c
@@ -720,7 +720,6 @@ i915_error_object_create(struct drm_i915_private *dev_priv,
 	reloc_offset = src->gtt_offset;
 	for (page = 0; page < page_count; page++) {
 		unsigned long flags;
-		void __iomem *s;
 		void *d;
 
 		d = kmalloc(PAGE_SIZE, GFP_ATOMIC);
@@ -728,10 +727,29 @@ i915_error_object_create(struct drm_i915_private *dev_priv,
 			goto unwind;
 
 		local_irq_save(flags);
-		s = io_mapping_map_atomic_wc(dev_priv->mm.gtt_mapping,
-					     reloc_offset);
-		memcpy_fromio(d, s, PAGE_SIZE);
-		io_mapping_unmap_atomic(s);
+		if (reloc_offset < dev_priv->mm.gtt_mappable_end) {
+			void __iomem *s;
+
+			/* Simply ignore tiling or any overlapping fence.
+			 * It's part of the error state, and this hopefully
+			 * captures what the GPU read.
+			 */
+
+			s = io_mapping_map_atomic_wc(dev_priv->mm.gtt_mapping,
+						     reloc_offset);
+			memcpy_fromio(d, s, PAGE_SIZE);
+			io_mapping_unmap_atomic(s);
+		} else {
+			void *s;
+
+			drm_clflush_pages(&src->pages[page], 1);
+
+			s = kmap_atomic(src->pages[page]);
+			memcpy(d, s, PAGE_SIZE);
+			kunmap_atomic(s);
+
+			drm_clflush_pages(&src->pages[page], 1);
+		}
 		local_irq_restore(flags);
 
 		dst->pages[page] = d;
@@ -770,11 +788,11 @@ i915_error_state_free(struct drm_device *dev,
 {
 	int i;
 
-	for (i = 0; i < ARRAY_SIZE(error->batchbuffer); i++)
-		i915_error_object_free(error->batchbuffer[i]);
-
-	for (i = 0; i < ARRAY_SIZE(error->ringbuffer); i++)
-		i915_error_object_free(error->ringbuffer[i]);
+	for (i = 0; i < ARRAY_SIZE(error->ring); i++) {
+		i915_error_object_free(error->ring[i].batchbuffer);
+		i915_error_object_free(error->ring[i].ringbuffer);
+		kfree(error->ring[i].requests);
+	}
 
 	kfree(error->active_bo);
 	kfree(error->overlay);
@@ -804,7 +822,7 @@ static u32 capture_bo_list(struct drm_i915_error_buffer *err,
 		err->tiling = obj->tiling_mode;
 		err->dirty = obj->dirty;
 		err->purgeable = obj->madv != I915_MADV_WILLNEED;
-		err->ring = obj->ring ? obj->ring->id : 0;
+		err->ring = obj->ring ? obj->ring->id : -1;
 		err->cache_level = obj->cache_level;
 
 		if (++i == count)
@@ -876,6 +894,92 @@ i915_error_first_batchbuffer(struct drm_i915_private *dev_priv,
 	return NULL;
 }
 
+static void i915_record_ring_state(struct drm_device *dev,
+				   struct drm_i915_error_state *error,
+				   struct intel_ring_buffer *ring)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+
+	if (INTEL_INFO(dev)->gen >= 6) {
+		error->faddr[ring->id] = I915_READ(RING_DMA_FADD(ring->mmio_base));
+		error->fault_reg[ring->id] = I915_READ(RING_FAULT_REG(ring));
+		error->semaphore_mboxes[ring->id][0]
+			= I915_READ(RING_SYNC_0(ring->mmio_base));
+		error->semaphore_mboxes[ring->id][1]
+			= I915_READ(RING_SYNC_1(ring->mmio_base));
+	}
+
+	if (INTEL_INFO(dev)->gen >= 4) {
+		error->ipeir[ring->id] = I915_READ(RING_IPEIR(ring->mmio_base));
+		error->ipehr[ring->id] = I915_READ(RING_IPEHR(ring->mmio_base));
+		error->instdone[ring->id] = I915_READ(RING_INSTDONE(ring->mmio_base));
+		error->instps[ring->id] = I915_READ(RING_INSTPS(ring->mmio_base));
+		if (ring->id == RCS) {
+			error->instdone1 = I915_READ(INSTDONE1);
+			error->bbaddr = I915_READ64(BB_ADDR);
+		}
+	} else {
+		error->ipeir[ring->id] = I915_READ(IPEIR);
+		error->ipehr[ring->id] = I915_READ(IPEHR);
+		error->instdone[ring->id] = I915_READ(INSTDONE);
+	}
+
+	error->instpm[ring->id] = I915_READ(RING_INSTPM(ring->mmio_base));
+	error->seqno[ring->id] = ring->get_seqno(ring);
+	error->acthd[ring->id] = intel_ring_get_active_head(ring);
+	error->head[ring->id] = I915_READ_HEAD(ring);
+	error->tail[ring->id] = I915_READ_TAIL(ring);
+
+	error->cpu_ring_head[ring->id] = ring->head;
+	error->cpu_ring_tail[ring->id] = ring->tail;
+}
+
+static void i915_gem_record_rings(struct drm_device *dev,
+				  struct drm_i915_error_state *error)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_gem_request *request;
+	int i, count;
+
+	for (i = 0; i < I915_NUM_RINGS; i++) {
+		struct intel_ring_buffer *ring = &dev_priv->ring[i];
+
+		if (ring->obj == NULL)
+			continue;
+
+		i915_record_ring_state(dev, error, ring);
+
+		error->ring[i].batchbuffer =
+			i915_error_first_batchbuffer(dev_priv, ring);
+
+		error->ring[i].ringbuffer =
+			i915_error_object_create(dev_priv, ring->obj);
+
+		count = 0;
+		list_for_each_entry(request, &ring->request_list, list)
+			count++;
+
+		error->ring[i].num_requests = count;
+		error->ring[i].requests =
+			kmalloc(count*sizeof(struct drm_i915_error_request),
+				GFP_ATOMIC);
+		if (error->ring[i].requests == NULL) {
+			error->ring[i].num_requests = 0;
+			continue;
+		}
+
+		count = 0;
+		list_for_each_entry(request, &ring->request_list, list) {
+			struct drm_i915_error_request *erq;
+
+			erq = &error->ring[i].requests[count++];
+			erq->seqno = request->seqno;
+			erq->jiffies = request->emitted_jiffies;
+			erq->tail = request->tail;
+		}
+	}
+}
+
 /**
  * i915_capture_error_state - capture an error record for later analysis
  * @dev: drm device
@@ -900,7 +1004,7 @@ static void i915_capture_error_state(struct drm_device *dev)
 		return;
 
 	/* Account for pipe specific data like PIPE*STAT */
-	error = kmalloc(sizeof(*error), GFP_ATOMIC);
+	error = kzalloc(sizeof(*error), GFP_ATOMIC);
 	if (!error) {
 		DRM_DEBUG_DRIVER("out of memory, not capturing error state\n");
 		return;
@@ -909,59 +1013,18 @@ static void i915_capture_error_state(struct drm_device *dev)
 	DRM_INFO("capturing error event; look for more information in /debug/dri/%d/i915_error_state\n",
 		 dev->primary->index);
 
-	error->seqno = dev_priv->ring[RCS].get_seqno(&dev_priv->ring[RCS]);
 	error->eir = I915_READ(EIR);
 	error->pgtbl_er = I915_READ(PGTBL_ER);
 	for_each_pipe(pipe)
 		error->pipestat[pipe] = I915_READ(PIPESTAT(pipe));
-	error->instpm = I915_READ(INSTPM);
-	error->error = 0;
+
 	if (INTEL_INFO(dev)->gen >= 6) {
 		error->error = I915_READ(ERROR_GEN6);
-
-		error->bcs_acthd = I915_READ(BCS_ACTHD);
-		error->bcs_ipehr = I915_READ(BCS_IPEHR);
-		error->bcs_ipeir = I915_READ(BCS_IPEIR);
-		error->bcs_instdone = I915_READ(BCS_INSTDONE);
-		error->bcs_seqno = 0;
-		if (dev_priv->ring[BCS].get_seqno)
-			error->bcs_seqno = dev_priv->ring[BCS].get_seqno(&dev_priv->ring[BCS]);
-
-		error->vcs_acthd = I915_READ(VCS_ACTHD);
-		error->vcs_ipehr = I915_READ(VCS_IPEHR);
-		error->vcs_ipeir = I915_READ(VCS_IPEIR);
-		error->vcs_instdone = I915_READ(VCS_INSTDONE);
-		error->vcs_seqno = 0;
-		if (dev_priv->ring[VCS].get_seqno)
-			error->vcs_seqno = dev_priv->ring[VCS].get_seqno(&dev_priv->ring[VCS]);
+		error->done_reg = I915_READ(DONE_REG);
 	}
-	if (INTEL_INFO(dev)->gen >= 4) {
-		error->ipeir = I915_READ(IPEIR_I965);
-		error->ipehr = I915_READ(IPEHR_I965);
-		error->instdone = I915_READ(INSTDONE_I965);
-		error->instps = I915_READ(INSTPS);
-		error->instdone1 = I915_READ(INSTDONE1);
-		error->acthd = I915_READ(ACTHD_I965);
-		error->bbaddr = I915_READ64(BB_ADDR);
-	} else {
-		error->ipeir = I915_READ(IPEIR);
-		error->ipehr = I915_READ(IPEHR);
-		error->instdone = I915_READ(INSTDONE);
-		error->acthd = I915_READ(ACTHD);
-		error->bbaddr = 0;
-	}
-	i915_gem_record_fences(dev, error);
-
-	/* Record the active batch and ring buffers */
-	for (i = 0; i < I915_NUM_RINGS; i++) {
-		error->batchbuffer[i] =
-			i915_error_first_batchbuffer(dev_priv,
-						     &dev_priv->ring[i]);
 
-		error->ringbuffer[i] =
-			i915_error_object_create(dev_priv,
-						 dev_priv->ring[i].obj);
-	}
+	i915_gem_record_fences(dev, error);
+	i915_gem_record_rings(dev, error);
 
 	/* Record buffers on the active and pinned lists. */
 	error->active_bo = NULL;
@@ -1017,11 +1080,12 @@ void i915_destroy_error_state(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct drm_i915_error_state *error;
+	unsigned long flags;
 
-	spin_lock(&dev_priv->error_lock);
+	spin_lock_irqsave(&dev_priv->error_lock, flags);
 	error = dev_priv->first_error;
 	dev_priv->first_error = NULL;
-	spin_unlock(&dev_priv->error_lock);
+	spin_unlock_irqrestore(&dev_priv->error_lock, flags);
 
 	if (error)
 		i915_error_state_free(dev, error);
@@ -1698,6 +1762,7 @@ void i915_hangcheck_elapsed(unsigned long data)
 	    dev_priv->last_instdone1 == instdone1) {
 		if (dev_priv->hangcheck_count++ > 1) {
 			DRM_ERROR("Hangcheck timer elapsed... GPU hung\n");
+			i915_handle_error(dev, true);
 
 			if (!IS_GEN2(dev)) {
 				/* Is the chip hanging on a WAIT_FOR_EVENT?
@@ -1705,7 +1770,6 @@ void i915_hangcheck_elapsed(unsigned long data)
 				 * and break the hang. This should work on
 				 * all but the second generation chipsets.
 				 */
-
 				if (kick_ring(&dev_priv->ring[RCS]))
 					goto repeat;
 
@@ -1718,7 +1782,6 @@ void i915_hangcheck_elapsed(unsigned long data)
 					goto repeat;
 			}
 
-			i915_handle_error(dev, true);
 			return;
 		}
 	} else {
@@ -1752,18 +1815,6 @@ static void ironlake_irq_preinstall(struct drm_device *dev)
 
 	I915_WRITE(HWSTAM, 0xeffe);
 
-	if (IS_GEN6(dev)) {
-		/* Workaround stalls observed on Sandy Bridge GPUs by
-		 * making the blitter command streamer generate a
-		 * write to the Hardware Status Page for
-		 * MI_USER_INTERRUPT.  This appears to serialize the
-		 * previous seqno write out before the interrupt
-		 * happens.
-		 */
-		I915_WRITE(GEN6_BLITTER_HWSTAM, ~GEN6_BLITTER_USER_INTERRUPT);
-		I915_WRITE(GEN6_BSD_HWSTAM, ~GEN6_BSD_USER_INTERRUPT);
-	}
-
 	/* XXX hotplug from PCH */
 
 	I915_WRITE(DEIMR, 0xffffffff);
diff --git a/drivers/gpu/drm/i915/i915_mem.c b/drivers/gpu/drm/i915/i915_mem.c
deleted file mode 100644
index cc8f6d49cf20..000000000000
--- a/drivers/gpu/drm/i915/i915_mem.c
+++ /dev/null
@@ -1,387 +0,0 @@
-/* i915_mem.c -- Simple agp/fb memory manager for i915 -*- linux-c -*-
- */
-/*
- * Copyright 2003 Tungsten Graphics, Inc., Cedar Park, Texas.
- * 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 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 TUNGSTEN GRAPHICS 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.
- *
- */
-
-#include "drmP.h"
-#include "drm.h"
-#include "i915_drm.h"
-#include "i915_drv.h"
-
-/* This memory manager is integrated into the global/local lru
- * mechanisms used by the clients.  Specifically, it operates by
- * setting the 'in_use' fields of the global LRU to indicate whether
- * this region is privately allocated to a client.
- *
- * This does require the client to actually respect that field.
- *
- * Currently no effort is made to allocate 'private' memory in any
- * clever way - the LRU information isn't used to determine which
- * block to allocate, and the ring is drained prior to allocations --
- * in other words allocation is expensive.
- */
-static void mark_block(struct drm_device * dev, struct mem_block *p, int in_use)
-{
-	drm_i915_private_t *dev_priv = dev->dev_private;
-	struct drm_i915_master_private *master_priv = dev->primary->master->driver_priv;
-	drm_i915_sarea_t *sarea_priv = master_priv->sarea_priv;
-	struct drm_tex_region *list;
-	unsigned shift, nr;
-	unsigned start;
-	unsigned end;
-	unsigned i;
-	int age;
-
-	shift = dev_priv->tex_lru_log_granularity;
-	nr = I915_NR_TEX_REGIONS;
-
-	start = p->start >> shift;
-	end = (p->start + p->size - 1) >> shift;
-
-	age = ++sarea_priv->texAge;
-	list = sarea_priv->texList;
-
-	/* Mark the regions with the new flag and update their age.  Move
-	 * them to head of list to preserve LRU semantics.
-	 */
-	for (i = start; i <= end; i++) {
-		list[i].in_use = in_use;
-		list[i].age = age;
-
-		/* remove_from_list(i)
-		 */
-		list[(unsigned)list[i].next].prev = list[i].prev;
-		list[(unsigned)list[i].prev].next = list[i].next;
-
-		/* insert_at_head(list, i)
-		 */
-		list[i].prev = nr;
-		list[i].next = list[nr].next;
-		list[(unsigned)list[nr].next].prev = i;
-		list[nr].next = i;
-	}
-}
-
-/* Very simple allocator for agp memory, working on a static range
- * already mapped into each client's address space.
- */
-
-static struct mem_block *split_block(struct mem_block *p, int start, int size,
-				     struct drm_file *file_priv)
-{
-	/* Maybe cut off the start of an existing block */
-	if (start > p->start) {
-		struct mem_block *newblock = kmalloc(sizeof(*newblock),
-						     GFP_KERNEL);
-		if (!newblock)
-			goto out;
-		newblock->start = start;
-		newblock->size = p->size - (start - p->start);
-		newblock->file_priv = NULL;
-		newblock->next = p->next;
-		newblock->prev = p;
-		p->next->prev = newblock;
-		p->next = newblock;
-		p->size -= newblock->size;
-		p = newblock;
-	}
-
-	/* Maybe cut off the end of an existing block */
-	if (size < p->size) {
-		struct mem_block *newblock = kmalloc(sizeof(*newblock),
-						     GFP_KERNEL);
-		if (!newblock)
-			goto out;
-		newblock->start = start + size;
-		newblock->size = p->size - size;
-		newblock->file_priv = NULL;
-		newblock->next = p->next;
-		newblock->prev = p;
-		p->next->prev = newblock;
-		p->next = newblock;
-		p->size = size;
-	}
-
-      out:
-	/* Our block is in the middle */
-	p->file_priv = file_priv;
-	return p;
-}
-
-static struct mem_block *alloc_block(struct mem_block *heap, int size,
-				     int align2, struct drm_file *file_priv)
-{
-	struct mem_block *p;
-	int mask = (1 << align2) - 1;
-
-	for (p = heap->next; p != heap; p = p->next) {
-		int start = (p->start + mask) & ~mask;
-		if (p->file_priv == NULL && start + size <= p->start + p->size)
-			return split_block(p, start, size, file_priv);
-	}
-
-	return NULL;
-}
-
-static struct mem_block *find_block(struct mem_block *heap, int start)
-{
-	struct mem_block *p;
-
-	for (p = heap->next; p != heap; p = p->next)
-		if (p->start == start)
-			return p;
-
-	return NULL;
-}
-
-static void free_block(struct mem_block *p)
-{
-	p->file_priv = NULL;
-
-	/* Assumes a single contiguous range.  Needs a special file_priv in
-	 * 'heap' to stop it being subsumed.
-	 */
-	if (p->next->file_priv == NULL) {
-		struct mem_block *q = p->next;
-		p->size += q->size;
-		p->next = q->next;
-		p->next->prev = p;
-		kfree(q);
-	}
-
-	if (p->prev->file_priv == NULL) {
-		struct mem_block *q = p->prev;
-		q->size += p->size;
-		q->next = p->next;
-		q->next->prev = q;
-		kfree(p);
-	}
-}
-
-/* Initialize.  How to check for an uninitialized heap?
- */
-static int init_heap(struct mem_block **heap, int start, int size)
-{
-	struct mem_block *blocks = kmalloc(sizeof(*blocks), GFP_KERNEL);
-
-	if (!blocks)
-		return -ENOMEM;
-
-	*heap = kmalloc(sizeof(**heap), GFP_KERNEL);
-	if (!*heap) {
-		kfree(blocks);
-		return -ENOMEM;
-	}
-
-	blocks->start = start;
-	blocks->size = size;
-	blocks->file_priv = NULL;
-	blocks->next = blocks->prev = *heap;
-
-	memset(*heap, 0, sizeof(**heap));
-	(*heap)->file_priv = (struct drm_file *) -1;
-	(*heap)->next = (*heap)->prev = blocks;
-	return 0;
-}
-
-/* Free all blocks associated with the releasing file.
- */
-void i915_mem_release(struct drm_device * dev, struct drm_file *file_priv,
-		      struct mem_block *heap)
-{
-	struct mem_block *p;
-
-	if (!heap || !heap->next)
-		return;
-
-	for (p = heap->next; p != heap; p = p->next) {
-		if (p->file_priv == file_priv) {
-			p->file_priv = NULL;
-			mark_block(dev, p, 0);
-		}
-	}
-
-	/* Assumes a single contiguous range.  Needs a special file_priv in
-	 * 'heap' to stop it being subsumed.
-	 */
-	for (p = heap->next; p != heap; p = p->next) {
-		while (p->file_priv == NULL && p->next->file_priv == NULL) {
-			struct mem_block *q = p->next;
-			p->size += q->size;
-			p->next = q->next;
-			p->next->prev = p;
-			kfree(q);
-		}
-	}
-}
-
-/* Shutdown.
- */
-void i915_mem_takedown(struct mem_block **heap)
-{
-	struct mem_block *p;
-
-	if (!*heap)
-		return;
-
-	for (p = (*heap)->next; p != *heap;) {
-		struct mem_block *q = p;
-		p = p->next;
-		kfree(q);
-	}
-
-	kfree(*heap);
-	*heap = NULL;
-}
-
-static struct mem_block **get_heap(drm_i915_private_t * dev_priv, int region)
-{
-	switch (region) {
-	case I915_MEM_REGION_AGP:
-		return &dev_priv->agp_heap;
-	default:
-		return NULL;
-	}
-}
-
-/* IOCTL HANDLERS */
-
-int i915_mem_alloc(struct drm_device *dev, void *data,
-		   struct drm_file *file_priv)
-{
-	drm_i915_private_t *dev_priv = dev->dev_private;
-	drm_i915_mem_alloc_t *alloc = data;
-	struct mem_block *block, **heap;
-
-	if (!dev_priv) {
-		DRM_ERROR("called with no initialization\n");
-		return -EINVAL;
-	}
-
-	heap = get_heap(dev_priv, alloc->region);
-	if (!heap || !*heap)
-		return -EFAULT;
-
-	/* Make things easier on ourselves: all allocations at least
-	 * 4k aligned.
-	 */
-	if (alloc->alignment < 12)
-		alloc->alignment = 12;
-
-	block = alloc_block(*heap, alloc->size, alloc->alignment, file_priv);
-
-	if (!block)
-		return -ENOMEM;
-
-	mark_block(dev, block, 1);
-
-	if (DRM_COPY_TO_USER(alloc->region_offset, &block->start,
-			     sizeof(int))) {
-		DRM_ERROR("copy_to_user\n");
-		return -EFAULT;
-	}
-
-	return 0;
-}
-
-int i915_mem_free(struct drm_device *dev, void *data,
-		  struct drm_file *file_priv)
-{
-	drm_i915_private_t *dev_priv = dev->dev_private;
-	drm_i915_mem_free_t *memfree = data;
-	struct mem_block *block, **heap;
-
-	if (!dev_priv) {
-		DRM_ERROR("called with no initialization\n");
-		return -EINVAL;
-	}
-
-	heap = get_heap(dev_priv, memfree->region);
-	if (!heap || !*heap)
-		return -EFAULT;
-
-	block = find_block(*heap, memfree->region_offset);
-	if (!block)
-		return -EFAULT;
-
-	if (block->file_priv != file_priv)
-		return -EPERM;
-
-	mark_block(dev, block, 0);
-	free_block(block);
-	return 0;
-}
-
-int i915_mem_init_heap(struct drm_device *dev, void *data,
-		       struct drm_file *file_priv)
-{
-	drm_i915_private_t *dev_priv = dev->dev_private;
-	drm_i915_mem_init_heap_t *initheap = data;
-	struct mem_block **heap;
-
-	if (!dev_priv) {
-		DRM_ERROR("called with no initialization\n");
-		return -EINVAL;
-	}
-
-	heap = get_heap(dev_priv, initheap->region);
-	if (!heap)
-		return -EFAULT;
-
-	if (*heap) {
-		DRM_ERROR("heap already initialized?");
-		return -EFAULT;
-	}
-
-	return init_heap(heap, initheap->start, initheap->size);
-}
-
-int i915_mem_destroy_heap(struct drm_device *dev, void *data,
-			   struct drm_file *file_priv)
-{
-	drm_i915_private_t *dev_priv = dev->dev_private;
-	drm_i915_mem_destroy_heap_t *destroyheap = data;
-	struct mem_block **heap;
-
-	if (!dev_priv) {
-		DRM_ERROR("called with no initialization\n");
-		return -EINVAL;
-	}
-
-	heap = get_heap(dev_priv, destroyheap->region);
-	if (!heap) {
-		DRM_ERROR("get_heap failed");
-		return -EFAULT;
-	}
-
-	if (!*heap) {
-		DRM_ERROR("heap not initialized?");
-		return -EFAULT;
-	}
-
-	i915_mem_takedown(heap);
-	return 0;
-}
diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h
index c3afb783cb9d..2abf4eb94039 100644
--- a/drivers/gpu/drm/i915/i915_reg.h
+++ b/drivers/gpu/drm/i915/i915_reg.h
@@ -86,12 +86,45 @@
 #define   GEN6_MBC_SNPCR_LOW	(2<<21)
 #define   GEN6_MBC_SNPCR_MIN	(3<<21) /* only 1/16th of the cache is shared */
 
+#define GEN6_MBCTL		0x0907c
+#define   GEN6_MBCTL_ENABLE_BOOT_FETCH	(1 << 4)
+#define   GEN6_MBCTL_CTX_FETCH_NEEDED	(1 << 3)
+#define   GEN6_MBCTL_BME_UPDATE_ENABLE	(1 << 2)
+#define   GEN6_MBCTL_MAE_UPDATE_ENABLE	(1 << 1)
+#define   GEN6_MBCTL_BOOT_FETCH_MECH	(1 << 0)
+
 #define GEN6_GDRST	0x941c
 #define  GEN6_GRDOM_FULL		(1 << 0)
 #define  GEN6_GRDOM_RENDER		(1 << 1)
 #define  GEN6_GRDOM_MEDIA		(1 << 2)
 #define  GEN6_GRDOM_BLT			(1 << 3)
 
+/* PPGTT stuff */
+#define GEN6_GTT_ADDR_ENCODE(addr)	((addr) | (((addr) >> 28) & 0xff0))
+
+#define GEN6_PDE_VALID			(1 << 0)
+#define GEN6_PDE_LARGE_PAGE		(2 << 0) /* use 32kb pages */
+/* gen6+ has bit 11-4 for physical addr bit 39-32 */
+#define GEN6_PDE_ADDR_ENCODE(addr)	GEN6_GTT_ADDR_ENCODE(addr)
+
+#define GEN6_PTE_VALID			(1 << 0)
+#define GEN6_PTE_UNCACHED		(1 << 1)
+#define GEN6_PTE_CACHE_LLC		(2 << 1)
+#define GEN6_PTE_CACHE_LLC_MLC		(3 << 1)
+#define GEN6_PTE_CACHE_BITS		(3 << 1)
+#define GEN6_PTE_GFDT			(1 << 3)
+#define GEN6_PTE_ADDR_ENCODE(addr)	GEN6_GTT_ADDR_ENCODE(addr)
+
+#define RING_PP_DIR_BASE(ring)		((ring)->mmio_base+0x228)
+#define RING_PP_DIR_BASE_READ(ring)	((ring)->mmio_base+0x518)
+#define RING_PP_DIR_DCLV(ring)		((ring)->mmio_base+0x220)
+#define   PP_DIR_DCLV_2G		0xffffffff
+
+#define GAM_ECOCHK			0x4090
+#define   ECOCHK_SNB_BIT		(1<<10)
+#define   ECOCHK_PPGTT_CACHE64B		(0x3<<3)
+#define   ECOCHK_PPGTT_CACHE4B		(0x0<<3)
+
 /* VGA stuff */
 
 #define VGA_ST01_MDA 0x3ba
@@ -295,6 +328,12 @@
 #define FENCE_REG_SANDYBRIDGE_0		0x100000
 #define   SANDYBRIDGE_FENCE_PITCH_SHIFT	32
 
+/* control register for cpu gtt access */
+#define TILECTL				0x101000
+#define   TILECTL_SWZCTL			(1 << 0)
+#define   TILECTL_TLB_PREFETCH_DIS	(1 << 2)
+#define   TILECTL_BACKSNOOP_DIS		(1 << 3)
+
 /*
  * Instruction and interrupt control regs
  */
@@ -318,7 +357,14 @@
 #define RING_MAX_IDLE(base)	((base)+0x54)
 #define RING_HWS_PGA(base)	((base)+0x80)
 #define RING_HWS_PGA_GEN6(base)	((base)+0x2080)
+#define ARB_MODE		0x04030
+#define   ARB_MODE_SWIZZLE_SNB	(1<<4)
+#define   ARB_MODE_SWIZZLE_IVB	(1<<5)
+#define   ARB_MODE_ENABLE(x)	GFX_MODE_ENABLE(x)
+#define   ARB_MODE_DISABLE(x)	GFX_MODE_DISABLE(x)
 #define RENDER_HWS_PGA_GEN7	(0x04080)
+#define RING_FAULT_REG(ring)	(0x4094 + 0x100*(ring)->id)
+#define DONE_REG		0x40b0
 #define BSD_HWS_PGA_GEN7	(0x04180)
 #define BLT_HWS_PGA_GEN7	(0x04280)
 #define RING_ACTHD(base)	((base)+0x74)
@@ -352,6 +398,12 @@
 #define IPEIR_I965	0x02064
 #define IPEHR_I965	0x02068
 #define INSTDONE_I965	0x0206c
+#define RING_IPEIR(base)	((base)+0x64)
+#define RING_IPEHR(base)	((base)+0x68)
+#define RING_INSTDONE(base)	((base)+0x6c)
+#define RING_INSTPS(base)	((base)+0x70)
+#define RING_DMA_FADD(base)	((base)+0x78)
+#define RING_INSTPM(base)	((base)+0xc0)
 #define INSTPS		0x02070 /* 965+ only */
 #define INSTDONE1	0x0207c /* 965+ only */
 #define ACTHD_I965	0x02074
@@ -365,14 +417,6 @@
 #define INSTDONE	0x02090
 #define NOPID		0x02094
 #define HWSTAM		0x02098
-#define VCS_INSTDONE	0x1206C
-#define VCS_IPEIR	0x12064
-#define VCS_IPEHR	0x12068
-#define VCS_ACTHD	0x12074
-#define BCS_INSTDONE	0x2206C
-#define BCS_IPEIR	0x22064
-#define BCS_IPEHR	0x22068
-#define BCS_ACTHD	0x22074
 
 #define ERROR_GEN6	0x040a0
 
@@ -391,10 +435,11 @@
 
 #define MI_MODE		0x0209c
 # define VS_TIMER_DISPATCH				(1 << 6)
-# define MI_FLUSH_ENABLE				(1 << 11)
+# define MI_FLUSH_ENABLE				(1 << 12)
 
 #define GFX_MODE	0x02520
 #define GFX_MODE_GEN7	0x0229c
+#define RING_MODE_GEN7(ring)	((ring)->mmio_base+0x29c)
 #define   GFX_RUN_LIST_ENABLE		(1<<15)
 #define   GFX_TLB_INVALIDATE_ALWAYS	(1<<13)
 #define   GFX_SURFACE_FAULT_ENABLE	(1<<12)
@@ -1037,6 +1082,29 @@
 #define C0DRB3			0x10206
 #define C1DRB3			0x10606
 
+/** snb MCH registers for reading the DRAM channel configuration */
+#define MAD_DIMM_C0			(MCHBAR_MIRROR_BASE_SNB + 0x5004)
+#define MAD_DIMM_C1			(MCHBAR_MIRROR_BASE_SNB + 0x5008)
+#define MAD_DIMM_C2			(MCHBAR_MIRROR_BASE_SNB + 0x500C)
+#define   MAD_DIMM_ECC_MASK		(0x3 << 24)
+#define   MAD_DIMM_ECC_OFF		(0x0 << 24)
+#define   MAD_DIMM_ECC_IO_ON_LOGIC_OFF	(0x1 << 24)
+#define   MAD_DIMM_ECC_IO_OFF_LOGIC_ON	(0x2 << 24)
+#define   MAD_DIMM_ECC_ON		(0x3 << 24)
+#define   MAD_DIMM_ENH_INTERLEAVE	(0x1 << 22)
+#define   MAD_DIMM_RANK_INTERLEAVE	(0x1 << 21)
+#define   MAD_DIMM_B_WIDTH_X16		(0x1 << 20) /* X8 chips if unset */
+#define   MAD_DIMM_A_WIDTH_X16		(0x1 << 19) /* X8 chips if unset */
+#define   MAD_DIMM_B_DUAL_RANK		(0x1 << 18)
+#define   MAD_DIMM_A_DUAL_RANK		(0x1 << 17)
+#define   MAD_DIMM_A_SELECT		(0x1 << 16)
+/* DIMM sizes are in multiples of 256mb. */
+#define   MAD_DIMM_B_SIZE_SHIFT		8
+#define   MAD_DIMM_B_SIZE_MASK		(0xff << MAD_DIMM_B_SIZE_SHIFT)
+#define   MAD_DIMM_A_SIZE_SHIFT		0
+#define   MAD_DIMM_A_SIZE_MASK		(0xff << MAD_DIMM_A_SIZE_SHIFT)
+
+
 /* Clocking configuration register */
 #define CLKCFG			0x10c00
 #define CLKCFG_FSB_400					(5 << 0)	/* hrawclk 100 */
@@ -1316,6 +1384,7 @@
 #define _VSYNC_A		0x60014
 #define _PIPEASRC	0x6001c
 #define _BCLRPAT_A	0x60020
+#define _VSYNCSHIFT_A	0x60028
 
 /* Pipe B timing regs */
 #define _HTOTAL_B	0x61000
@@ -1326,6 +1395,8 @@
 #define _VSYNC_B		0x61014
 #define _PIPEBSRC	0x6101c
 #define _BCLRPAT_B	0x61020
+#define _VSYNCSHIFT_B	0x61028
+
 
 #define HTOTAL(pipe) _PIPE(pipe, _HTOTAL_A, _HTOTAL_B)
 #define HBLANK(pipe) _PIPE(pipe, _HBLANK_A, _HBLANK_B)
@@ -1334,6 +1405,7 @@
 #define VBLANK(pipe) _PIPE(pipe, _VBLANK_A, _VBLANK_B)
 #define VSYNC(pipe) _PIPE(pipe, _VSYNC_A, _VSYNC_B)
 #define BCLRPAT(pipe) _PIPE(pipe, _BCLRPAT_A, _BCLRPAT_B)
+#define VSYNCSHIFT(pipe) _PIPE(pipe, _VSYNCSHIFT_A, _VSYNCSHIFT_B)
 
 /* VGA port control */
 #define ADPA			0x61100
@@ -2313,16 +2385,28 @@
 #define   PIPECONF_DISABLE	0
 #define   PIPECONF_DOUBLE_WIDE	(1<<30)
 #define   I965_PIPECONF_ACTIVE	(1<<30)
+#define   PIPECONF_FRAME_START_DELAY_MASK (3<<27)
 #define   PIPECONF_SINGLE_WIDE	0
 #define   PIPECONF_PIPE_UNLOCKED 0
 #define   PIPECONF_PIPE_LOCKED	(1<<25)
 #define   PIPECONF_PALETTE	0
 #define   PIPECONF_GAMMA		(1<<24)
 #define   PIPECONF_FORCE_BORDER	(1<<25)
-#define   PIPECONF_PROGRESSIVE	(0 << 21)
-#define   PIPECONF_INTERLACE_W_FIELD_INDICATION	(6 << 21)
-#define   PIPECONF_INTERLACE_FIELD_0_ONLY		(7 << 21)
 #define   PIPECONF_INTERLACE_MASK	(7 << 21)
+/* Note that pre-gen3 does not support interlaced display directly. Panel
+ * fitting must be disabled on pre-ilk for interlaced. */
+#define   PIPECONF_PROGRESSIVE			(0 << 21)
+#define   PIPECONF_INTERLACE_W_SYNC_SHIFT_PANEL	(4 << 21) /* gen4 only */
+#define   PIPECONF_INTERLACE_W_SYNC_SHIFT	(5 << 21) /* gen4 only */
+#define   PIPECONF_INTERLACE_W_FIELD_INDICATION	(6 << 21)
+#define   PIPECONF_INTERLACE_FIELD_0_ONLY	(7 << 21) /* gen3 only */
+/* Ironlake and later have a complete new set of values for interlaced. PFIT
+ * means panel fitter required, PF means progressive fetch, DBL means power
+ * saving pixel doubling. */
+#define   PIPECONF_PFIT_PF_INTERLACED_ILK	(1 << 21)
+#define   PIPECONF_INTERLACED_ILK		(3 << 21)
+#define   PIPECONF_INTERLACED_DBL_ILK		(4 << 21) /* ilk/snb only */
+#define   PIPECONF_PFIT_PF_INTERLACED_DBL_ILK	(5 << 21) /* ilk/snb only */
 #define   PIPECONF_CXSR_DOWNCLOCK	(1<<16)
 #define   PIPECONF_BPP_MASK	(0x000000e0)
 #define   PIPECONF_BPP_8	(0<<5)
@@ -2689,7 +2773,7 @@
 #define   DVS_FORMAT_RGBX888	(2<<25)
 #define   DVS_FORMAT_RGBX161616	(3<<25)
 #define   DVS_SOURCE_KEY	(1<<22)
-#define   DVS_RGB_ORDER_RGBX	(1<<20)
+#define   DVS_RGB_ORDER_XBGR	(1<<20)
 #define   DVS_YUV_BYTE_ORDER_MASK (3<<16)
 #define   DVS_YUV_ORDER_YUYV	(0<<16)
 #define   DVS_YUV_ORDER_UYVY	(1<<16)
@@ -3028,6 +3112,20 @@
 #define  DISP_TILE_SURFACE_SWIZZLING	(1<<13)
 #define  DISP_FBC_WM_DIS		(1<<15)
 
+/* GEN7 chicken */
+#define GEN7_COMMON_SLICE_CHICKEN1		0x7010
+# define GEN7_CSC1_RHWO_OPT_DISABLE_IN_RCC	((1<<10) | (1<<26))
+
+#define GEN7_L3CNTLREG1				0xB01C
+#define  GEN7_WA_FOR_GEN7_L3_CONTROL			0x3C4FFF8C
+
+#define GEN7_L3_CHICKEN_MODE_REGISTER		0xB030
+#define  GEN7_WA_L3_CHICKEN_MODE				0x20000000
+
+/* WaCatErrorRejectionIssue */
+#define GEN7_SQ_CHICKEN_MBCUNIT_CONFIG		0x9030
+#define  GEN7_SQ_CHICKEN_MBCUNIT_SQINTMOB	(1<<11)
+
 /* PCH */
 
 /* south display engine interrupt */
@@ -3205,6 +3303,7 @@
 #define _TRANS_VSYNC_A           0xe0014
 #define  TRANS_VSYNC_END_SHIFT  16
 #define  TRANS_VSYNC_START_SHIFT 0
+#define _TRANS_VSYNCSHIFT_A	0xe0028
 
 #define _TRANSA_DATA_M1          0xe0030
 #define _TRANSA_DATA_N1          0xe0034
@@ -3235,6 +3334,7 @@
 #define _TRANS_VTOTAL_B          0xe100c
 #define _TRANS_VBLANK_B          0xe1010
 #define _TRANS_VSYNC_B           0xe1014
+#define _TRANS_VSYNCSHIFT_B	 0xe1028
 
 #define TRANS_HTOTAL(pipe) _PIPE(pipe, _TRANS_HTOTAL_A, _TRANS_HTOTAL_B)
 #define TRANS_HBLANK(pipe) _PIPE(pipe, _TRANS_HBLANK_A, _TRANS_HBLANK_B)
@@ -3242,6 +3342,8 @@
 #define TRANS_VTOTAL(pipe) _PIPE(pipe, _TRANS_VTOTAL_A, _TRANS_VTOTAL_B)
 #define TRANS_VBLANK(pipe) _PIPE(pipe, _TRANS_VBLANK_A, _TRANS_VBLANK_B)
 #define TRANS_VSYNC(pipe) _PIPE(pipe, _TRANS_VSYNC_A, _TRANS_VSYNC_B)
+#define TRANS_VSYNCSHIFT(pipe) _PIPE(pipe, _TRANS_VSYNCSHIFT_A, \
+				     _TRANS_VSYNCSHIFT_B)
 
 #define _TRANSB_DATA_M1          0xe1030
 #define _TRANSB_DATA_N1          0xe1034
@@ -3275,7 +3377,10 @@
 #define  TRANS_FSYNC_DELAY_HB4  (3<<27)
 #define  TRANS_DP_AUDIO_ONLY    (1<<26)
 #define  TRANS_DP_VIDEO_AUDIO   (0<<26)
+#define  TRANS_INTERLACE_MASK   (7<<21)
 #define  TRANS_PROGRESSIVE      (0<<21)
+#define  TRANS_INTERLACED       (3<<21)
+#define  TRANS_LEGACY_INTERLACED_ILK (2<<21)
 #define  TRANS_8BPC             (0<<5)
 #define  TRANS_10BPC            (1<<5)
 #define  TRANS_6BPC             (2<<5)
@@ -3614,10 +3719,17 @@
 #define  ECOBUS					0xa180
 #define    FORCEWAKE_MT_ENABLE			(1<<5)
 
+#define  GTFIFODBG				0x120000
+#define    GT_FIFO_CPU_ERROR_MASK		7
+#define    GT_FIFO_OVFERR			(1<<2)
+#define    GT_FIFO_IAWRERR			(1<<1)
+#define    GT_FIFO_IARDERR			(1<<0)
+
 #define  GT_FIFO_FREE_ENTRIES			0x120008
 #define    GT_FIFO_NUM_RESERVED_ENTRIES		20
 
 #define GEN6_UCGCTL2				0x9404
+# define GEN6_RCZUNIT_CLOCK_GATE_DISABLE		(1 << 13)
 # define GEN6_RCPBUNIT_CLOCK_GATE_DISABLE		(1 << 12)
 # define GEN6_RCCUNIT_CLOCK_GATE_DISABLE		(1 << 11)
 
@@ -3742,4 +3854,16 @@
  */
 #define GEN7_SO_WRITE_OFFSET(n)		(0x5280 + (n) * 4)
 
+#define IBX_AUD_CONFIG_A			0xe2000
+#define CPT_AUD_CONFIG_A			0xe5000
+#define   AUD_CONFIG_N_VALUE_INDEX		(1 << 29)
+#define   AUD_CONFIG_N_PROG_ENABLE		(1 << 28)
+#define   AUD_CONFIG_UPPER_N_SHIFT		20
+#define   AUD_CONFIG_UPPER_N_VALUE		(0xff << 20)
+#define   AUD_CONFIG_LOWER_N_SHIFT		4
+#define   AUD_CONFIG_LOWER_N_VALUE		(0xfff << 4)
+#define   AUD_CONFIG_PIXEL_CLOCK_HDMI_SHIFT	16
+#define   AUD_CONFIG_PIXEL_CLOCK_HDMI		(0xf << 16)
+#define   AUD_CONFIG_DISABLE_NCTS		(1 << 3)
+
 #endif /* _I915_REG_H_ */
diff --git a/drivers/gpu/drm/i915/intel_acpi.c b/drivers/gpu/drm/i915/intel_acpi.c
index cb912106d1a2..bae3edf956a4 100644
--- a/drivers/gpu/drm/i915/intel_acpi.c
+++ b/drivers/gpu/drm/i915/intel_acpi.c
@@ -208,7 +208,7 @@ static bool intel_dsm_pci_probe(struct pci_dev *pdev)
 
 	ret = intel_dsm(dhandle, INTEL_DSM_FN_SUPPORTED_FUNCTIONS, 0);
 	if (ret < 0) {
-		DRM_ERROR("failed to get supported _DSM functions\n");
+		DRM_DEBUG_KMS("failed to get supported _DSM functions\n");
 		return false;
 	}
 
diff --git a/drivers/gpu/drm/i915/intel_bios.c b/drivers/gpu/drm/i915/intel_bios.c
index 63880e2e5cfd..b48fc2a8410c 100644
--- a/drivers/gpu/drm/i915/intel_bios.c
+++ b/drivers/gpu/drm/i915/intel_bios.c
@@ -24,6 +24,7 @@
  *    Eric Anholt <eric@anholt.net>
  *
  */
+#include <linux/dmi.h>
 #include <drm/drm_dp_helper.h>
 #include "drmP.h"
 #include "drm.h"
@@ -572,7 +573,7 @@ parse_device_mapping(struct drm_i915_private *dev_priv,
 		DRM_DEBUG_KMS("no child dev is parsed from VBT\n");
 		return;
 	}
-	dev_priv->child_dev = kzalloc(sizeof(*p_child) * count, GFP_KERNEL);
+	dev_priv->child_dev = kcalloc(count, sizeof(*p_child), GFP_KERNEL);
 	if (!dev_priv->child_dev) {
 		DRM_DEBUG_KMS("No memory space for child device\n");
 		return;
@@ -621,6 +622,26 @@ init_vbt_defaults(struct drm_i915_private *dev_priv)
 	dev_priv->edp.bpp = 18;
 }
 
+static int __init intel_no_opregion_vbt_callback(const struct dmi_system_id *id)
+{
+	DRM_DEBUG_KMS("Falling back to manually reading VBT from "
+		      "VBIOS ROM for %s\n",
+		      id->ident);
+	return 1;
+}
+
+static const struct dmi_system_id intel_no_opregion_vbt[] = {
+	{
+		.callback = intel_no_opregion_vbt_callback,
+		.ident = "ThinkCentre A57",
+		.matches = {
+			DMI_MATCH(DMI_SYS_VENDOR, "LENOVO"),
+			DMI_MATCH(DMI_PRODUCT_NAME, "97027RG"),
+		},
+	},
+	{ }
+};
+
 /**
  * intel_parse_bios - find VBT and initialize settings from the BIOS
  * @dev: DRM device
@@ -641,7 +662,7 @@ intel_parse_bios(struct drm_device *dev)
 	init_vbt_defaults(dev_priv);
 
 	/* XXX Should this validation be moved to intel_opregion.c? */
-	if (dev_priv->opregion.vbt) {
+	if (!dmi_check_system(intel_no_opregion_vbt) && dev_priv->opregion.vbt) {
 		struct vbt_header *vbt = dev_priv->opregion.vbt;
 		if (memcmp(vbt->signature, "$VBT", 4) == 0) {
 			DRM_DEBUG_KMS("Using VBT from OpRegion: %20s\n",
@@ -669,7 +690,7 @@ intel_parse_bios(struct drm_device *dev)
 		}
 
 		if (!vbt) {
-			DRM_ERROR("VBT signature missing\n");
+			DRM_DEBUG_DRIVER("VBT signature missing\n");
 			pci_unmap_rom(pdev, bios);
 			return -1;
 		}
diff --git a/drivers/gpu/drm/i915/intel_crt.c b/drivers/gpu/drm/i915/intel_crt.c
index dd729d46a61f..4d3d736a4f56 100644
--- a/drivers/gpu/drm/i915/intel_crt.c
+++ b/drivers/gpu/drm/i915/intel_crt.c
@@ -594,7 +594,10 @@ void intel_crt_init(struct drm_device *dev)
 				1 << INTEL_ANALOG_CLONE_BIT |
 				1 << INTEL_SDVO_LVDS_CLONE_BIT);
 	crt->base.crtc_mask = (1 << 0) | (1 << 1);
-	connector->interlace_allowed = 1;
+	if (IS_GEN2(dev))
+		connector->interlace_allowed = 0;
+	else
+		connector->interlace_allowed = 1;
 	connector->doublescan_allowed = 0;
 
 	drm_encoder_helper_add(&crt->base.base, &intel_crt_helper_funcs);
diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
index b3b51c43dad0..91b35fd1db8c 100644
--- a/drivers/gpu/drm/i915/intel_display.c
+++ b/drivers/gpu/drm/i915/intel_display.c
@@ -75,7 +75,7 @@ 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 *);
+			int, int, intel_clock_t *, intel_clock_t *);
 };
 
 /* FDI */
@@ -83,17 +83,21 @@ struct intel_limit {
 
 static bool
 intel_find_best_PLL(const intel_limit_t *limit, struct drm_crtc *crtc,
-		    int target, int refclk, intel_clock_t *best_clock);
+		    int target, int refclk, intel_clock_t *match_clock,
+		    intel_clock_t *best_clock);
 static bool
 intel_g4x_find_best_PLL(const intel_limit_t *limit, struct drm_crtc *crtc,
-			int target, int refclk, intel_clock_t *best_clock);
+			int target, int refclk, intel_clock_t *match_clock,
+			intel_clock_t *best_clock);
 
 static bool
 intel_find_pll_g4x_dp(const intel_limit_t *, struct drm_crtc *crtc,
-		      int target, int refclk, intel_clock_t *best_clock);
+		      int target, int refclk, intel_clock_t *match_clock,
+		      intel_clock_t *best_clock);
 static bool
 intel_find_pll_ironlake_dp(const intel_limit_t *, struct drm_crtc *crtc,
-			   int target, int refclk, intel_clock_t *best_clock);
+			   int target, int refclk, intel_clock_t *match_clock,
+			   intel_clock_t *best_clock);
 
 static inline u32 /* units of 100MHz */
 intel_fdi_link_freq(struct drm_device *dev)
@@ -515,7 +519,8 @@ static bool intel_PLL_is_valid(struct drm_device *dev,
 
 static bool
 intel_find_best_PLL(const intel_limit_t *limit, struct drm_crtc *crtc,
-		    int target, int refclk, intel_clock_t *best_clock)
+		    int target, int refclk, intel_clock_t *match_clock,
+		    intel_clock_t *best_clock)
 
 {
 	struct drm_device *dev = crtc->dev;
@@ -562,6 +567,9 @@ intel_find_best_PLL(const intel_limit_t *limit, struct drm_crtc *crtc,
 					if (!intel_PLL_is_valid(dev, limit,
 								&clock))
 						continue;
+					if (match_clock &&
+					    clock.p != match_clock->p)
+						continue;
 
 					this_err = abs(clock.dot - target);
 					if (this_err < err) {
@@ -578,7 +586,8 @@ intel_find_best_PLL(const intel_limit_t *limit, struct drm_crtc *crtc,
 
 static bool
 intel_g4x_find_best_PLL(const intel_limit_t *limit, struct drm_crtc *crtc,
-			int target, int refclk, intel_clock_t *best_clock)
+			int target, int refclk, intel_clock_t *match_clock,
+			intel_clock_t *best_clock)
 {
 	struct drm_device *dev = crtc->dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
@@ -625,6 +634,9 @@ intel_g4x_find_best_PLL(const intel_limit_t *limit, struct drm_crtc *crtc,
 					if (!intel_PLL_is_valid(dev, limit,
 								&clock))
 						continue;
+					if (match_clock &&
+					    clock.p != match_clock->p)
+						continue;
 
 					this_err = abs(clock.dot - target);
 					if (this_err < err_most) {
@@ -642,7 +654,8 @@ intel_g4x_find_best_PLL(const intel_limit_t *limit, struct drm_crtc *crtc,
 
 static bool
 intel_find_pll_ironlake_dp(const intel_limit_t *limit, struct drm_crtc *crtc,
-			   int target, int refclk, intel_clock_t *best_clock)
+			   int target, int refclk, intel_clock_t *match_clock,
+			   intel_clock_t *best_clock)
 {
 	struct drm_device *dev = crtc->dev;
 	intel_clock_t clock;
@@ -668,7 +681,8 @@ intel_find_pll_ironlake_dp(const intel_limit_t *limit, struct drm_crtc *crtc,
 /* DisplayPort has only two frequencies, 162MHz and 270MHz */
 static bool
 intel_find_pll_g4x_dp(const intel_limit_t *limit, struct drm_crtc *crtc,
-		      int target, int refclk, intel_clock_t *best_clock)
+		      int target, int refclk, intel_clock_t *match_clock,
+		      intel_clock_t *best_clock)
 {
 	intel_clock_t clock;
 	if (target < 200000) {
@@ -922,6 +936,10 @@ void assert_pipe(struct drm_i915_private *dev_priv,
 	u32 val;
 	bool cur_state;
 
+	/* if we need the pipe A quirk it must be always on */
+	if (pipe == PIPE_A && dev_priv->quirks & QUIRK_PIPEA_FORCE)
+		state = true;
+
 	reg = PIPECONF(pipe);
 	val = I915_READ(reg);
 	cur_state = !!(val & PIPECONF_ENABLE);
@@ -930,19 +948,24 @@ void assert_pipe(struct drm_i915_private *dev_priv,
 	     pipe_name(pipe), state_string(state), state_string(cur_state));
 }
 
-static void assert_plane_enabled(struct drm_i915_private *dev_priv,
-				 enum plane plane)
+static void assert_plane(struct drm_i915_private *dev_priv,
+			 enum plane plane, bool state)
 {
 	int reg;
 	u32 val;
+	bool cur_state;
 
 	reg = DSPCNTR(plane);
 	val = I915_READ(reg);
-	WARN(!(val & DISPLAY_PLANE_ENABLE),
-	     "plane %c assertion failure, should be active but is disabled\n",
-	     plane_name(plane));
+	cur_state = !!(val & DISPLAY_PLANE_ENABLE);
+	WARN(cur_state != state,
+	     "plane %c assertion failure (expected %s, current %s)\n",
+	     plane_name(plane), state_string(state), state_string(cur_state));
 }
 
+#define assert_plane_enabled(d, p) assert_plane(d, p, true)
+#define assert_plane_disabled(d, p) assert_plane(d, p, false)
+
 static void assert_planes_disabled(struct drm_i915_private *dev_priv,
 				   enum pipe pipe)
 {
@@ -951,8 +974,14 @@ 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)) {
+		reg = DSPCNTR(pipe);
+		val = I915_READ(reg);
+		WARN((val & DISPLAY_PLANE_ENABLE),
+		     "plane %c assertion failure, should be disabled but not\n",
+		     plane_name(pipe));
 		return;
+	}
 
 	/* Need to check both planes against the pipe */
 	for (i = 0; i < 2; i++) {
@@ -1071,7 +1100,7 @@ static void assert_pch_hdmi_disabled(struct drm_i915_private *dev_priv,
 {
 	u32 val = I915_READ(reg);
 	WARN(hdmi_pipe_enabled(dev_priv, val, pipe),
-	     "PCH DP (0x%08x) enabled on transcoder %c, should be disabled\n",
+	     "PCH HDMI (0x%08x) enabled on transcoder %c, should be disabled\n",
 	     reg, pipe_name(pipe));
 }
 
@@ -1237,7 +1266,8 @@ static void intel_enable_transcoder(struct drm_i915_private *dev_priv,
 				    enum pipe pipe)
 {
 	int reg;
-	u32 val;
+	u32 val, pipeconf_val;
+	struct drm_crtc *crtc = dev_priv->pipe_to_crtc_mapping[pipe];
 
 	/* PCH only available on ILK+ */
 	BUG_ON(dev_priv->info->gen < 5);
@@ -1251,6 +1281,7 @@ static void intel_enable_transcoder(struct drm_i915_private *dev_priv,
 
 	reg = TRANSCONF(pipe);
 	val = I915_READ(reg);
+	pipeconf_val = I915_READ(PIPECONF(pipe));
 
 	if (HAS_PCH_IBX(dev_priv->dev)) {
 		/*
@@ -1258,8 +1289,19 @@ static void intel_enable_transcoder(struct drm_i915_private *dev_priv,
 		 * that in pipeconf reg.
 		 */
 		val &= ~PIPE_BPC_MASK;
-		val |= I915_READ(PIPECONF(pipe)) & PIPE_BPC_MASK;
+		val |= pipeconf_val & PIPE_BPC_MASK;
 	}
+
+	val &= ~TRANS_INTERLACE_MASK;
+	if ((pipeconf_val & PIPECONF_INTERLACE_MASK) == PIPECONF_INTERLACED_ILK)
+		if (HAS_PCH_IBX(dev_priv->dev) &&
+		    intel_pipe_has_type(crtc, INTEL_OUTPUT_SDVO))
+			val |= TRANS_LEGACY_INTERLACED_ILK;
+		else
+			val |= TRANS_INTERLACED;
+	else
+		val |= TRANS_PROGRESSIVE;
+
 	I915_WRITE(reg, val | TRANS_ENABLE);
 	if (wait_for(I915_READ(reg) & TRANS_STATE_ENABLE, 100))
 		DRM_ERROR("failed to enable transcoder %d\n", pipe);
@@ -1872,7 +1914,7 @@ static void intel_update_fbc(struct drm_device *dev)
 	if (enable_fbc < 0) {
 		DRM_DEBUG_KMS("fbc set to per-chip default\n");
 		enable_fbc = 1;
-		if (INTEL_INFO(dev)->gen <= 5)
+		if (INTEL_INFO(dev)->gen <= 6)
 			enable_fbc = 0;
 	}
 	if (!enable_fbc) {
@@ -2012,6 +2054,8 @@ intel_pin_and_fence_fb_obj(struct drm_device *dev,
 		ret = i915_gem_object_get_fence(obj, pipelined);
 		if (ret)
 			goto err_unpin;
+
+		i915_gem_object_pin_fence(obj);
 	}
 
 	dev_priv->mm.interruptible = true;
@@ -2024,6 +2068,12 @@ err_interruptible:
 	return ret;
 }
 
+void intel_unpin_fb_obj(struct drm_i915_gem_object *obj)
+{
+	i915_gem_object_unpin_fence(obj);
+	i915_gem_object_unpin(obj);
+}
+
 static int i9xx_update_plane(struct drm_crtc *crtc, struct drm_framebuffer *fb,
 			     int x, int y)
 {
@@ -2255,7 +2305,7 @@ intel_pipe_set_base(struct drm_crtc *crtc, int x, int y,
 	ret = intel_pipe_set_base_atomic(crtc, crtc->fb, x, y,
 					 LEAVE_ATOMIC_MODE_SET);
 	if (ret) {
-		i915_gem_object_unpin(to_intel_framebuffer(crtc->fb)->obj);
+		intel_unpin_fb_obj(to_intel_framebuffer(crtc->fb)->obj);
 		mutex_unlock(&dev->struct_mutex);
 		DRM_ERROR("failed to update base address\n");
 		return ret;
@@ -2263,7 +2313,7 @@ intel_pipe_set_base(struct drm_crtc *crtc, int x, int y,
 
 	if (old_fb) {
 		intel_wait_for_vblank(dev, intel_crtc->pipe);
-		i915_gem_object_unpin(to_intel_framebuffer(old_fb)->obj);
+		intel_unpin_fb_obj(to_intel_framebuffer(old_fb)->obj);
 	}
 
 	mutex_unlock(&dev->struct_mutex);
@@ -2936,6 +2986,7 @@ static void ironlake_pch_enable(struct drm_crtc *crtc)
 	I915_WRITE(TRANS_VTOTAL(pipe), I915_READ(VTOTAL(pipe)));
 	I915_WRITE(TRANS_VBLANK(pipe), I915_READ(VBLANK(pipe)));
 	I915_WRITE(TRANS_VSYNC(pipe),  I915_READ(VSYNC(pipe)));
+	I915_WRITE(TRANS_VSYNCSHIFT(pipe),  I915_READ(VSYNCSHIFT(pipe)));
 
 	intel_fdi_normal_train(crtc);
 
@@ -3321,10 +3372,12 @@ static void intel_crtc_disable(struct drm_crtc *crtc)
 	struct drm_device *dev = crtc->dev;
 
 	crtc_funcs->dpms(crtc, DRM_MODE_DPMS_OFF);
+	assert_plane_disabled(dev->dev_private, to_intel_crtc(crtc)->plane);
+	assert_pipe_disabled(dev->dev_private, to_intel_crtc(crtc)->pipe);
 
 	if (crtc->fb) {
 		mutex_lock(&dev->struct_mutex);
-		i915_gem_object_unpin(to_intel_framebuffer(crtc->fb)->obj);
+		intel_unpin_fb_obj(to_intel_framebuffer(crtc->fb)->obj);
 		mutex_unlock(&dev->struct_mutex);
 	}
 }
@@ -3398,11 +3451,8 @@ static bool intel_crtc_mode_fixup(struct drm_crtc *crtc,
 			return false;
 	}
 
-	/* XXX some encoders set the crtcinfo, others don't.
-	 * Obviously we need some form of conflict resolution here...
-	 */
-	if (adjusted_mode->crtc_htotal == 0)
-		drm_mode_set_crtcinfo(adjusted_mode, 0);
+	/* All interlaced capable intel hw wants timings in frames. */
+	drm_mode_set_crtcinfo(adjusted_mode, 0);
 
 	return true;
 }
@@ -4521,6 +4571,7 @@ void sandybridge_update_wm(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	int latency = SNB_READ_WM0_LATENCY() * 100;	/* In unit 0.1us */
+	u32 val;
 	int fbc_wm, plane_wm, cursor_wm;
 	unsigned int enabled;
 
@@ -4529,8 +4580,10 @@ void sandybridge_update_wm(struct drm_device *dev)
 			    &sandybridge_display_wm_info, latency,
 			    &sandybridge_cursor_wm_info, latency,
 			    &plane_wm, &cursor_wm)) {
-		I915_WRITE(WM0_PIPEA_ILK,
-			   (plane_wm << WM0_PIPE_PLANE_SHIFT) | cursor_wm);
+		val = I915_READ(WM0_PIPEA_ILK);
+		val &= ~(WM0_PIPE_PLANE_MASK | WM0_PIPE_CURSOR_MASK);
+		I915_WRITE(WM0_PIPEA_ILK, val |
+			   ((plane_wm << WM0_PIPE_PLANE_SHIFT) | cursor_wm));
 		DRM_DEBUG_KMS("FIFO watermarks For pipe A -"
 			      " plane %d, " "cursor: %d\n",
 			      plane_wm, cursor_wm);
@@ -4541,8 +4594,10 @@ void sandybridge_update_wm(struct drm_device *dev)
 			    &sandybridge_display_wm_info, latency,
 			    &sandybridge_cursor_wm_info, latency,
 			    &plane_wm, &cursor_wm)) {
-		I915_WRITE(WM0_PIPEB_ILK,
-			   (plane_wm << WM0_PIPE_PLANE_SHIFT) | cursor_wm);
+		val = I915_READ(WM0_PIPEB_ILK);
+		val &= ~(WM0_PIPE_PLANE_MASK | WM0_PIPE_CURSOR_MASK);
+		I915_WRITE(WM0_PIPEB_ILK, val |
+			   ((plane_wm << WM0_PIPE_PLANE_SHIFT) | cursor_wm));
 		DRM_DEBUG_KMS("FIFO watermarks For pipe B -"
 			      " plane %d, cursor: %d\n",
 			      plane_wm, cursor_wm);
@@ -4555,8 +4610,10 @@ void sandybridge_update_wm(struct drm_device *dev)
 			    &sandybridge_display_wm_info, latency,
 			    &sandybridge_cursor_wm_info, latency,
 			    &plane_wm, &cursor_wm)) {
-		I915_WRITE(WM0_PIPEC_IVB,
-			   (plane_wm << WM0_PIPE_PLANE_SHIFT) | cursor_wm);
+		val = I915_READ(WM0_PIPEC_IVB);
+		val &= ~(WM0_PIPE_PLANE_MASK | WM0_PIPE_CURSOR_MASK);
+		I915_WRITE(WM0_PIPEC_IVB, val |
+			   ((plane_wm << WM0_PIPE_PLANE_SHIFT) | cursor_wm));
 		DRM_DEBUG_KMS("FIFO watermarks For pipe C -"
 			      " plane %d, cursor: %d\n",
 			      plane_wm, cursor_wm);
@@ -4680,8 +4737,17 @@ sandybridge_compute_sprite_srwm(struct drm_device *dev, int plane,
 
 	crtc = intel_get_crtc_for_plane(dev, plane);
 	clock = crtc->mode.clock;
+	if (!clock) {
+		*sprite_wm = 0;
+		return false;
+	}
 
 	line_time_us = (sprite_width * 1000) / clock;
+	if (!line_time_us) {
+		*sprite_wm = 0;
+		return false;
+	}
+
 	line_count = (latency_ns / line_time_us + 1000) / 1000;
 	line_size = sprite_width * pixel_size;
 
@@ -4700,6 +4766,7 @@ static void sandybridge_update_sprite_wm(struct drm_device *dev, int pipe,
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	int latency = SNB_READ_WM0_LATENCY() * 100;	/* In unit 0.1us */
+	u32 val;
 	int sprite_wm, reg;
 	int ret;
 
@@ -4726,7 +4793,9 @@ static void sandybridge_update_sprite_wm(struct drm_device *dev, int pipe,
 		return;
 	}
 
-	I915_WRITE(reg, I915_READ(reg) | (sprite_wm << WM0_PIPE_SPRITE_SHIFT));
+	val = I915_READ(reg);
+	val &= ~WM0_PIPE_SPRITE_MASK;
+	I915_WRITE(reg, val | (sprite_wm << WM0_PIPE_SPRITE_SHIFT));
 	DRM_DEBUG_KMS("sprite watermarks For pipe %d - %d\n", pipe, sprite_wm);
 
 
@@ -4968,6 +5037,82 @@ static bool intel_choose_pipe_bpp_dither(struct drm_crtc *crtc,
 	return display_bpc != bpc;
 }
 
+static int i9xx_get_refclk(struct drm_crtc *crtc, int num_connectors)
+{
+	struct drm_device *dev = crtc->dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	int refclk;
+
+	if (intel_pipe_has_type(crtc, INTEL_OUTPUT_LVDS) &&
+	    intel_panel_use_ssc(dev_priv) && num_connectors < 2) {
+		refclk = dev_priv->lvds_ssc_freq * 1000;
+		DRM_DEBUG_KMS("using SSC reference clock of %d MHz\n",
+			      refclk / 1000);
+	} else if (!IS_GEN2(dev)) {
+		refclk = 96000;
+	} else {
+		refclk = 48000;
+	}
+
+	return refclk;
+}
+
+static void i9xx_adjust_sdvo_tv_clock(struct drm_display_mode *adjusted_mode,
+				      intel_clock_t *clock)
+{
+	/* SDVO TV has fixed PLL values depend on its clock range,
+	   this mirrors vbios setting. */
+	if (adjusted_mode->clock >= 100000
+	    && adjusted_mode->clock < 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) {
+		clock->p1 = 1;
+		clock->p2 = 10;
+		clock->n = 6;
+		clock->m1 = 12;
+		clock->m2 = 8;
+	}
+}
+
+static void i9xx_update_pll_dividers(struct drm_crtc *crtc,
+				     intel_clock_t *clock,
+				     intel_clock_t *reduced_clock)
+{
+	struct drm_device *dev = crtc->dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
+	int pipe = intel_crtc->pipe;
+	u32 fp, fp2 = 0;
+
+	if (IS_PINEVIEW(dev)) {
+		fp = (1 << clock->n) << 16 | clock->m1 << 8 | clock->m2;
+		if (reduced_clock)
+			fp2 = (1 << reduced_clock->n) << 16 |
+				reduced_clock->m1 << 8 | reduced_clock->m2;
+	} else {
+		fp = clock->n << 16 | clock->m1 << 8 | clock->m2;
+		if (reduced_clock)
+			fp2 = reduced_clock->n << 16 | reduced_clock->m1 << 8 |
+				reduced_clock->m2;
+	}
+
+	I915_WRITE(FP0(pipe), fp);
+
+	intel_crtc->lowfreq_avail = false;
+	if (intel_pipe_has_type(crtc, INTEL_OUTPUT_LVDS) &&
+	    reduced_clock && i915_powersave) {
+		I915_WRITE(FP1(pipe), fp2);
+		intel_crtc->lowfreq_avail = true;
+	} else {
+		I915_WRITE(FP1(pipe), fp);
+	}
+}
+
 static int i9xx_crtc_mode_set(struct drm_crtc *crtc,
 			      struct drm_display_mode *mode,
 			      struct drm_display_mode *adjusted_mode,
@@ -4981,7 +5126,7 @@ static int i9xx_crtc_mode_set(struct drm_crtc *crtc,
 	int plane = intel_crtc->plane;
 	int refclk, num_connectors = 0;
 	intel_clock_t clock, reduced_clock;
-	u32 dpll, fp = 0, fp2 = 0, dspcntr, pipeconf;
+	u32 dpll, dspcntr, pipeconf, vsyncshift;
 	bool ok, has_reduced_clock = false, is_sdvo = false, is_dvo = false;
 	bool is_crt = false, is_lvds = false, is_tv = false, is_dp = false;
 	struct drm_mode_config *mode_config = &dev->mode_config;
@@ -5022,15 +5167,7 @@ static int i9xx_crtc_mode_set(struct drm_crtc *crtc,
 		num_connectors++;
 	}
 
-	if (is_lvds && intel_panel_use_ssc(dev_priv) && num_connectors < 2) {
-		refclk = dev_priv->lvds_ssc_freq * 1000;
-		DRM_DEBUG_KMS("using SSC reference clock of %d MHz\n",
-			      refclk / 1000);
-	} else if (!IS_GEN2(dev)) {
-		refclk = 96000;
-	} else {
-		refclk = 48000;
-	}
+	refclk = i9xx_get_refclk(crtc, num_connectors);
 
 	/*
 	 * Returns a set of divisors for the desired target clock with the given
@@ -5038,7 +5175,8 @@ static int i9xx_crtc_mode_set(struct drm_crtc *crtc,
 	 * reflck * (5 * (m1 + 2) + (m2 + 2)) / (n + 2) / p1 / p2.
 	 */
 	limit = intel_limit(crtc, refclk);
-	ok = limit->find_pll(limit, crtc, adjusted_mode->clock, refclk, &clock);
+	ok = limit->find_pll(limit, crtc, adjusted_mode->clock, refclk, NULL,
+			     &clock);
 	if (!ok) {
 		DRM_ERROR("Couldn't find PLL settings for mode!\n");
 		return -EINVAL;
@@ -5048,53 +5186,24 @@ static int i9xx_crtc_mode_set(struct drm_crtc *crtc,
 	intel_crtc_update_cursor(crtc, true);
 
 	if (is_lvds && dev_priv->lvds_downclock_avail) {
+		/*
+		 * Ensure we match the reduced clock's P to the target clock.
+		 * If the clocks don't match, we can't switch the display clock
+		 * by using the FP0/FP1. In such case we will disable the LVDS
+		 * downclock feature.
+		*/
 		has_reduced_clock = limit->find_pll(limit, crtc,
 						    dev_priv->lvds_downclock,
 						    refclk,
+						    &clock,
 						    &reduced_clock);
-		if (has_reduced_clock && (clock.p != reduced_clock.p)) {
-			/*
-			 * If the different P is found, it means that we can't
-			 * switch the display clock by using the FP0/FP1.
-			 * In such case we will disable the LVDS downclock
-			 * feature.
-			 */
-			DRM_DEBUG_KMS("Different P is found for "
-				      "LVDS clock/downclock\n");
-			has_reduced_clock = 0;
-		}
-	}
-	/* SDVO TV has fixed PLL values depend on its clock range,
-	   this mirrors vbios setting. */
-	if (is_sdvo && is_tv) {
-		if (adjusted_mode->clock >= 100000
-		    && adjusted_mode->clock < 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) {
-			clock.p1 = 1;
-			clock.p2 = 10;
-			clock.n = 6;
-			clock.m1 = 12;
-			clock.m2 = 8;
-		}
 	}
 
-	if (IS_PINEVIEW(dev)) {
-		fp = (1 << clock.n) << 16 | clock.m1 << 8 | clock.m2;
-		if (has_reduced_clock)
-			fp2 = (1 << reduced_clock.n) << 16 |
-				reduced_clock.m1 << 8 | reduced_clock.m2;
-	} else {
-		fp = clock.n << 16 | clock.m1 << 8 | clock.m2;
-		if (has_reduced_clock)
-			fp2 = reduced_clock.n << 16 | reduced_clock.m1 << 8 |
-				reduced_clock.m2;
-	}
+	if (is_sdvo && is_tv)
+		i9xx_adjust_sdvo_tv_clock(adjusted_mode, &clock);
+
+	i9xx_update_pll_dividers(crtc, &clock, has_reduced_clock ?
+				 &reduced_clock : NULL);
 
 	dpll = DPLL_VGA_MODE_DIS;
 
@@ -5168,8 +5277,6 @@ static int i9xx_crtc_mode_set(struct drm_crtc *crtc,
 	/* Set up the display plane register */
 	dspcntr = DISPPLANE_GAMMA_ENABLE;
 
-	/* Ironlake's plane is forced to pipe, bit 24 is to
-	   enable color space conversion */
 	if (pipe == 0)
 		dspcntr &= ~DISPPLANE_SEL_PIPE_MASK;
 	else
@@ -5204,7 +5311,6 @@ static int i9xx_crtc_mode_set(struct drm_crtc *crtc,
 	DRM_DEBUG_KMS("Mode for pipe %c:\n", pipe == 0 ? 'A' : 'B');
 	drm_mode_debug_printmodeline(mode);
 
-	I915_WRITE(FP0(pipe), fp);
 	I915_WRITE(DPLL(pipe), dpll & ~DPLL_VCO_ENABLE);
 
 	POSTING_READ(DPLL(pipe));
@@ -5291,33 +5397,32 @@ static int i9xx_crtc_mode_set(struct drm_crtc *crtc,
 		I915_WRITE(DPLL(pipe), dpll);
 	}
 
-	intel_crtc->lowfreq_avail = false;
-	if (is_lvds && has_reduced_clock && i915_powersave) {
-		I915_WRITE(FP1(pipe), fp2);
-		intel_crtc->lowfreq_avail = true;
-		if (HAS_PIPE_CXSR(dev)) {
+	if (HAS_PIPE_CXSR(dev)) {
+		if (intel_crtc->lowfreq_avail) {
 			DRM_DEBUG_KMS("enabling CxSR downclocking\n");
 			pipeconf |= PIPECONF_CXSR_DOWNCLOCK;
-		}
-	} else {
-		I915_WRITE(FP1(pipe), fp);
-		if (HAS_PIPE_CXSR(dev)) {
+		} else {
 			DRM_DEBUG_KMS("disabling CxSR downclocking\n");
 			pipeconf &= ~PIPECONF_CXSR_DOWNCLOCK;
 		}
 	}
 
-	if (adjusted_mode->flags & DRM_MODE_FLAG_INTERLACE) {
+	pipeconf &= ~PIPECONF_INTERLACE_MASK;
+	if (!IS_GEN2(dev) &&
+	    adjusted_mode->flags & DRM_MODE_FLAG_INTERLACE) {
 		pipeconf |= PIPECONF_INTERLACE_W_FIELD_INDICATION;
 		/* the chip adds 2 halflines automatically */
-		adjusted_mode->crtc_vdisplay -= 1;
 		adjusted_mode->crtc_vtotal -= 1;
-		adjusted_mode->crtc_vblank_start -= 1;
 		adjusted_mode->crtc_vblank_end -= 1;
-		adjusted_mode->crtc_vsync_end -= 1;
-		adjusted_mode->crtc_vsync_start -= 1;
-	} else
-		pipeconf &= ~PIPECONF_INTERLACE_MASK; /* progressive */
+		vsyncshift = adjusted_mode->crtc_hsync_start
+			     - adjusted_mode->crtc_htotal/2;
+	} else {
+		pipeconf |= PIPECONF_PROGRESSIVE;
+		vsyncshift = 0;
+	}
+
+	if (!IS_GEN3(dev))
+		I915_WRITE(VSYNCSHIFT(pipe), vsyncshift);
 
 	I915_WRITE(HTOTAL(pipe),
 		   (adjusted_mode->crtc_hdisplay - 1) |
@@ -5434,7 +5539,8 @@ void ironlake_init_pch_refclk(struct drm_device *dev)
 		if (intel_panel_use_ssc(dev_priv) && can_ssc) {
 			DRM_DEBUG_KMS("Using SSC on panel\n");
 			temp |= DREF_SSC1_ENABLE;
-		}
+		} else
+			temp &= ~DREF_SSC1_ENABLE;
 
 		/* Get SSC going before enabling the outputs */
 		I915_WRITE(PCH_DREF_CONTROL, temp);
@@ -5583,7 +5689,8 @@ static int ironlake_crtc_mode_set(struct drm_crtc *crtc,
 	 * reflck * (5 * (m1 + 2) + (m2 + 2)) / (n + 2) / p1 / p2.
 	 */
 	limit = intel_limit(crtc, refclk);
-	ok = limit->find_pll(limit, crtc, adjusted_mode->clock, refclk, &clock);
+	ok = limit->find_pll(limit, crtc, adjusted_mode->clock, refclk, NULL,
+			     &clock);
 	if (!ok) {
 		DRM_ERROR("Couldn't find PLL settings for mode!\n");
 		return -EINVAL;
@@ -5593,21 +5700,17 @@ static int ironlake_crtc_mode_set(struct drm_crtc *crtc,
 	intel_crtc_update_cursor(crtc, true);
 
 	if (is_lvds && dev_priv->lvds_downclock_avail) {
+		/*
+		 * Ensure we match the reduced clock's P to the target clock.
+		 * If the clocks don't match, we can't switch the display clock
+		 * by using the FP0/FP1. In such case we will disable the LVDS
+		 * downclock feature.
+		*/
 		has_reduced_clock = limit->find_pll(limit, crtc,
 						    dev_priv->lvds_downclock,
 						    refclk,
+						    &clock,
 						    &reduced_clock);
-		if (has_reduced_clock && (clock.p != reduced_clock.p)) {
-			/*
-			 * If the different P is found, it means that we can't
-			 * switch the display clock by using the FP0/FP1.
-			 * In such case we will disable the LVDS downclock
-			 * feature.
-			 */
-			DRM_DEBUG_KMS("Different P is found for "
-				      "LVDS clock/downclock\n");
-			has_reduced_clock = 0;
-		}
 	}
 	/* SDVO TV has fixed PLL values depend on its clock range,
 	   this mirrors vbios setting. */
@@ -5902,17 +6005,19 @@ static int ironlake_crtc_mode_set(struct drm_crtc *crtc,
 		}
 	}
 
+	pipeconf &= ~PIPECONF_INTERLACE_MASK;
 	if (adjusted_mode->flags & DRM_MODE_FLAG_INTERLACE) {
-		pipeconf |= PIPECONF_INTERLACE_W_FIELD_INDICATION;
+		pipeconf |= PIPECONF_INTERLACED_ILK;
 		/* the chip adds 2 halflines automatically */
-		adjusted_mode->crtc_vdisplay -= 1;
 		adjusted_mode->crtc_vtotal -= 1;
-		adjusted_mode->crtc_vblank_start -= 1;
 		adjusted_mode->crtc_vblank_end -= 1;
-		adjusted_mode->crtc_vsync_end -= 1;
-		adjusted_mode->crtc_vsync_start -= 1;
-	} else
-		pipeconf &= ~PIPECONF_INTERLACE_W_FIELD_INDICATION; /* progressive */
+		I915_WRITE(VSYNCSHIFT(pipe),
+			   adjusted_mode->crtc_hsync_start
+			   - adjusted_mode->crtc_htotal/2);
+	} else {
+		pipeconf |= PIPECONF_PROGRESSIVE;
+		I915_WRITE(VSYNCSHIFT(pipe), 0);
+	}
 
 	I915_WRITE(HTOTAL(pipe),
 		   (adjusted_mode->crtc_hdisplay - 1) |
@@ -5955,12 +6060,6 @@ static int ironlake_crtc_mode_set(struct drm_crtc *crtc,
 
 	intel_wait_for_vblank(dev, pipe);
 
-	if (IS_GEN5(dev)) {
-		/* enable address swizzle for tiling buffer */
-		temp = I915_READ(DISP_ARB_CTL);
-		I915_WRITE(DISP_ARB_CTL, temp | DISP_TILE_SURFACE_SWIZZLING);
-	}
-
 	I915_WRITE(DSPCNTR(plane), dspcntr);
 	POSTING_READ(DSPCNTR(plane));
 
@@ -6075,15 +6174,18 @@ static void ironlake_write_eld(struct drm_connector *connector,
 	uint32_t i;
 	int len;
 	int hdmiw_hdmiedid;
+	int aud_config;
 	int aud_cntl_st;
 	int aud_cntrl_st2;
 
 	if (HAS_PCH_IBX(connector->dev)) {
 		hdmiw_hdmiedid = IBX_HDMIW_HDMIEDID_A;
+		aud_config = IBX_AUD_CONFIG_A;
 		aud_cntl_st = IBX_AUD_CNTL_ST_A;
 		aud_cntrl_st2 = IBX_AUD_CNTL_ST2;
 	} else {
 		hdmiw_hdmiedid = CPT_HDMIW_HDMIEDID_A;
+		aud_config = CPT_AUD_CONFIG_A;
 		aud_cntl_st = CPT_AUD_CNTL_ST_A;
 		aud_cntrl_st2 = CPT_AUD_CNTRL_ST2;
 	}
@@ -6091,6 +6193,7 @@ static void ironlake_write_eld(struct drm_connector *connector,
 	i = to_intel_crtc(crtc)->pipe;
 	hdmiw_hdmiedid += i * 0x100;
 	aud_cntl_st += i * 0x100;
+	aud_config += i * 0x100;
 
 	DRM_DEBUG_DRIVER("ELD on pipe %c\n", pipe_name(i));
 
@@ -6110,7 +6213,9 @@ static void ironlake_write_eld(struct drm_connector *connector,
 	if (intel_pipe_has_type(crtc, INTEL_OUTPUT_DISPLAYPORT)) {
 		DRM_DEBUG_DRIVER("ELD: DisplayPort detected\n");
 		eld[5] |= (1 << 2);	/* Conn_Type, 0x1 = DisplayPort */
-	}
+		I915_WRITE(aud_config, AUD_CONFIG_N_VALUE_INDEX); /* 0x1 = DP */
+	} else
+		I915_WRITE(aud_config, 0);
 
 	if (intel_eld_uptodate(connector,
 			       aud_cntrl_st2, eldv,
@@ -6173,7 +6278,7 @@ void intel_crtc_load_lut(struct drm_crtc *crtc)
 	int i;
 
 	/* The clocks have to be on to load the palette. */
-	if (!crtc->enabled)
+	if (!crtc->enabled || !intel_crtc->active)
 		return;
 
 	/* use legacy palette for Ironlake */
@@ -6559,7 +6664,7 @@ intel_framebuffer_create_for_mode(struct drm_device *dev,
 	mode_cmd.height = mode->vdisplay;
 	mode_cmd.pitches[0] = intel_framebuffer_pitch_for_width(mode_cmd.width,
 								bpp);
-	mode_cmd.pixel_format = 0;
+	mode_cmd.pixel_format = drm_mode_legacy_fb_format(bpp, depth);
 
 	return intel_framebuffer_create(dev, &mode_cmd, obj);
 }
@@ -6916,9 +7021,7 @@ static void intel_increase_pllclock(struct drm_crtc *crtc)
 	if (!HAS_PIPE_CXSR(dev) && (dpll & DISPLAY_RATE_SELECT_FPA1)) {
 		DRM_DEBUG_DRIVER("upclocking LVDS\n");
 
-		/* Unlock panel regs */
-		I915_WRITE(PP_CONTROL,
-			   I915_READ(PP_CONTROL) | PANEL_UNLOCK_REGS);
+		assert_panel_unlocked(dev_priv, pipe);
 
 		dpll &= ~DISPLAY_RATE_SELECT_FPA1;
 		I915_WRITE(dpll_reg, dpll);
@@ -6927,9 +7030,6 @@ static void intel_increase_pllclock(struct drm_crtc *crtc)
 		dpll = I915_READ(dpll_reg);
 		if (dpll & DISPLAY_RATE_SELECT_FPA1)
 			DRM_DEBUG_DRIVER("failed to upclock LVDS!\n");
-
-		/* ...and lock them again */
-		I915_WRITE(PP_CONTROL, I915_READ(PP_CONTROL) & 0x3);
 	}
 
 	/* Schedule downclock */
@@ -6959,9 +7059,7 @@ static void intel_decrease_pllclock(struct drm_crtc *crtc)
 	if (!HAS_PIPE_CXSR(dev) && intel_crtc->lowfreq_avail) {
 		DRM_DEBUG_DRIVER("downclocking LVDS\n");
 
-		/* Unlock panel regs */
-		I915_WRITE(PP_CONTROL, I915_READ(PP_CONTROL) |
-			   PANEL_UNLOCK_REGS);
+		assert_panel_unlocked(dev_priv, pipe);
 
 		dpll |= DISPLAY_RATE_SELECT_FPA1;
 		I915_WRITE(dpll_reg, dpll);
@@ -6969,9 +7067,6 @@ static void intel_decrease_pllclock(struct drm_crtc *crtc)
 		dpll = I915_READ(dpll_reg);
 		if (!(dpll & DISPLAY_RATE_SELECT_FPA1))
 			DRM_DEBUG_DRIVER("failed to downclock LVDS!\n");
-
-		/* ...and lock them again */
-		I915_WRITE(PP_CONTROL, I915_READ(PP_CONTROL) & 0x3);
 	}
 
 }
@@ -7086,7 +7181,7 @@ static void intel_unpin_work_fn(struct work_struct *__work)
 		container_of(__work, struct intel_unpin_work, work);
 
 	mutex_lock(&work->dev->struct_mutex);
-	i915_gem_object_unpin(work->old_fb_obj);
+	intel_unpin_fb_obj(work->old_fb_obj);
 	drm_gem_object_unreference(&work->pending_flip_obj->base);
 	drm_gem_object_unreference(&work->old_fb_obj->base);
 
@@ -7236,7 +7331,7 @@ static int intel_gen2_queue_flip(struct drm_device *dev,
 		 MI_DISPLAY_FLIP_PLANE(intel_crtc->plane));
 	OUT_RING(fb->pitches[0]);
 	OUT_RING(obj->gtt_offset + offset);
-	OUT_RING(MI_NOOP);
+	OUT_RING(0); /* aux display base address, unused */
 	ADVANCE_LP_RING();
 out:
 	return ret;
@@ -7486,6 +7581,12 @@ static void intel_sanitize_modesetting(struct drm_device *dev,
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	u32 reg, val;
 
+	/* Clear any frame start delays used for debugging left by the BIOS */
+	for_each_pipe(pipe) {
+		reg = PIPECONF(pipe);
+		I915_WRITE(reg, I915_READ(reg) & ~PIPECONF_FRAME_START_DELAY_MASK);
+	}
+
 	if (HAS_PCH_SPLIT(dev))
 		return;
 
@@ -7670,10 +7771,9 @@ static void intel_setup_outputs(struct drm_device *dev)
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_encoder *encoder;
 	bool dpd_is_edp = false;
-	bool has_lvds = false;
+	bool has_lvds;
 
-	if (IS_MOBILE(dev) && !IS_I830(dev))
-		has_lvds = intel_lvds_init(dev);
+	has_lvds = intel_lvds_init(dev);
 	if (!has_lvds && !HAS_PCH_SPLIT(dev)) {
 		/* disable the panel fitter on everything but LVDS */
 		I915_WRITE(PFIT_CONTROL, 0);
@@ -7817,6 +7917,7 @@ int intel_framebuffer_init(struct drm_device *dev,
 	case DRM_FORMAT_RGB332:
 	case DRM_FORMAT_RGB565:
 	case DRM_FORMAT_XRGB8888:
+	case DRM_FORMAT_XBGR8888:
 	case DRM_FORMAT_ARGB8888:
 	case DRM_FORMAT_XRGB2101010:
 	case DRM_FORMAT_ARGB2101010:
@@ -7828,7 +7929,8 @@ int intel_framebuffer_init(struct drm_device *dev,
 	case DRM_FORMAT_VYUY:
 		break;
 	default:
-		DRM_ERROR("unsupported pixel format\n");
+		DRM_DEBUG_KMS("unsupported pixel format %u\n",
+				mode_cmd->pixel_format);
 		return -EINVAL;
 	}
 
@@ -8120,7 +8222,7 @@ void intel_init_emon(struct drm_device *dev)
 	dev_priv->corr = (lcfuse & LCFUSE_HIV_MASK);
 }
 
-static bool intel_enable_rc6(struct drm_device *dev)
+static int intel_enable_rc6(struct drm_device *dev)
 {
 	/*
 	 * Respect the kernel parameter if it is set
@@ -8138,11 +8240,11 @@ static bool intel_enable_rc6(struct drm_device *dev)
 	 * Disable rc6 on Sandybridge
 	 */
 	if (INTEL_INFO(dev)->gen == 6) {
-		DRM_DEBUG_DRIVER("Sandybridge: RC6 disabled\n");
-		return 0;
+		DRM_DEBUG_DRIVER("Sandybridge: deep RC6 disabled\n");
+		return INTEL_RC6_ENABLE;
 	}
-	DRM_DEBUG_DRIVER("RC6 enabled\n");
-	return 1;
+	DRM_DEBUG_DRIVER("RC6 and deep RC6 enabled\n");
+	return (INTEL_RC6_ENABLE | INTEL_RC6p_ENABLE);
 }
 
 void gen6_enable_rps(struct drm_i915_private *dev_priv)
@@ -8150,7 +8252,9 @@ void gen6_enable_rps(struct drm_i915_private *dev_priv)
 	u32 rp_state_cap = I915_READ(GEN6_RP_STATE_CAP);
 	u32 gt_perf_status = I915_READ(GEN6_GT_PERF_STATUS);
 	u32 pcu_mbox, rc6_mask = 0;
+	u32 gtfifodbg;
 	int cur_freq, min_freq, max_freq;
+	int rc6_mode;
 	int i;
 
 	/* Here begins a magic sequence of register writes to enable
@@ -8161,6 +8265,13 @@ void gen6_enable_rps(struct drm_i915_private *dev_priv)
 	 */
 	I915_WRITE(GEN6_RC_STATE, 0);
 	mutex_lock(&dev_priv->dev->struct_mutex);
+
+	/* Clear the DBG now so we don't confuse earlier errors */
+	if ((gtfifodbg = I915_READ(GTFIFODBG))) {
+		DRM_ERROR("GT fifo had a previous error %x\n", gtfifodbg);
+		I915_WRITE(GTFIFODBG, gtfifodbg);
+	}
+
 	gen6_gt_force_wake_get(dev_priv);
 
 	/* disable the counters and set deterministic thresholds */
@@ -8181,9 +8292,20 @@ void gen6_enable_rps(struct drm_i915_private *dev_priv)
 	I915_WRITE(GEN6_RC6p_THRESHOLD, 100000);
 	I915_WRITE(GEN6_RC6pp_THRESHOLD, 64000); /* unused */
 
-	if (intel_enable_rc6(dev_priv->dev))
-		rc6_mask = GEN6_RC_CTL_RC6p_ENABLE |
-			GEN6_RC_CTL_RC6_ENABLE;
+	rc6_mode = intel_enable_rc6(dev_priv->dev);
+	if (rc6_mode & INTEL_RC6_ENABLE)
+		rc6_mask |= GEN6_RC_CTL_RC6_ENABLE;
+
+	if (rc6_mode & INTEL_RC6p_ENABLE)
+		rc6_mask |= GEN6_RC_CTL_RC6p_ENABLE;
+
+	if (rc6_mode & INTEL_RC6pp_ENABLE)
+		rc6_mask |= GEN6_RC_CTL_RC6pp_ENABLE;
+
+	DRM_INFO("Enabling RC6 states: RC6 %s, RC6p %s, RC6pp %s\n",
+			(rc6_mode & INTEL_RC6_ENABLE) ? "on" : "off",
+			(rc6_mode & INTEL_RC6p_ENABLE) ? "on" : "off",
+			(rc6_mode & INTEL_RC6pp_ENABLE) ? "on" : "off");
 
 	I915_WRITE(GEN6_RC_CONTROL,
 		   rc6_mask |
@@ -8461,12 +8583,32 @@ static void ivybridge_init_clock_gating(struct drm_device *dev)
 	I915_WRITE(WM2_LP_ILK, 0);
 	I915_WRITE(WM1_LP_ILK, 0);
 
+	/* According to the spec, bit 13 (RCZUNIT) must be set on IVB.
+	 * This implements the WaDisableRCZUnitClockGating workaround.
+	 */
+	I915_WRITE(GEN6_UCGCTL2, GEN6_RCZUNIT_CLOCK_GATE_DISABLE);
+
 	I915_WRITE(ILK_DSPCLK_GATE, IVB_VRHUNIT_CLK_GATE);
 
 	I915_WRITE(IVB_CHICKEN3,
 		   CHICKEN3_DGMG_REQ_OUT_FIX_DISABLE |
 		   CHICKEN3_DGMG_DONE_FIX_DISABLE);
 
+	/* Apply the WaDisableRHWOOptimizationForRenderHang workaround. */
+	I915_WRITE(GEN7_COMMON_SLICE_CHICKEN1,
+		   GEN7_CSC1_RHWO_OPT_DISABLE_IN_RCC);
+
+	/* WaApplyL3ControlAndL3ChickenMode requires those two on Ivy Bridge */
+	I915_WRITE(GEN7_L3CNTLREG1,
+			GEN7_WA_FOR_GEN7_L3_CONTROL);
+	I915_WRITE(GEN7_L3_CHICKEN_MODE_REGISTER,
+			GEN7_WA_L3_CHICKEN_MODE);
+
+	/* This is required by WaCatErrorRejectionIssue */
+	I915_WRITE(GEN7_SQ_CHICKEN_MBCUNIT_CONFIG,
+			I915_READ(GEN7_SQ_CHICKEN_MBCUNIT_CONFIG) |
+			GEN7_SQ_CHICKEN_MBCUNIT_SQINTMOB);
+
 	for_each_pipe(pipe) {
 		I915_WRITE(DSPCNTR(pipe),
 			   I915_READ(DSPCNTR(pipe)) |
@@ -8927,8 +9069,6 @@ struct intel_quirk {
 };
 
 struct intel_quirk intel_quirks[] = {
-	/* HP Compaq 2730p needs pipe A force quirk (LP: #291555) */
-	{ 0x2a42, 0x103c, 0x30eb, quirk_pipea_force },
 	/* HP Mini needs pipe A force quirk (LP: #322104) */
 	{ 0x27ae, 0x103c, 0x361a, quirk_pipea_force },
 
@@ -9005,6 +9145,9 @@ void intel_modeset_init(struct drm_device *dev)
 	dev->mode_config.min_width = 0;
 	dev->mode_config.min_height = 0;
 
+	dev->mode_config.preferred_depth = 24;
+	dev->mode_config.prefer_shadow = 1;
+
 	dev->mode_config.funcs = (void *)&intel_mode_funcs;
 
 	intel_init_quirks(dev);
diff --git a/drivers/gpu/drm/i915/intel_dp.c b/drivers/gpu/drm/i915/intel_dp.c
index db3b461ad412..110552ff302c 100644
--- a/drivers/gpu/drm/i915/intel_dp.c
+++ b/drivers/gpu/drm/i915/intel_dp.c
@@ -49,7 +49,7 @@ struct intel_dp {
 	uint32_t DP;
 	uint8_t  link_configuration[DP_LINK_CONFIGURATION_SIZE];
 	bool has_audio;
-	int force_audio;
+	enum hdmi_force_audio force_audio;
 	uint32_t color_range;
 	int dpms_mode;
 	uint8_t link_bw;
@@ -208,17 +208,8 @@ intel_dp_link_clock(uint8_t link_bw)
  */
 
 static int
-intel_dp_link_required(struct intel_dp *intel_dp, int pixel_clock, int check_bpp)
+intel_dp_link_required(int pixel_clock, int bpp)
 {
-	struct drm_crtc *crtc = intel_dp->base.base.crtc;
-	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
-	int bpp = 24;
-
-	if (check_bpp)
-		bpp = check_bpp;
-	else if (intel_crtc)
-		bpp = intel_crtc->bpp;
-
 	return (pixel_clock * bpp + 9) / 10;
 }
 
@@ -245,12 +236,11 @@ intel_dp_mode_valid(struct drm_connector *connector,
 			return MODE_PANEL;
 	}
 
-	mode_rate = intel_dp_link_required(intel_dp, mode->clock, 0);
+	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(intel_dp,
-							   mode->clock, 18);
+			mode_rate = intel_dp_link_required(mode->clock, 18);
 			if (mode_rate > max_rate)
 				return MODE_CLOCK_HIGH;
 			else
@@ -362,7 +352,7 @@ intel_dp_aux_ch(struct intel_dp *intel_dp,
 	int recv_bytes;
 	uint32_t status;
 	uint32_t aux_clock_divider;
-	int try, precharge;
+	int try, precharge = 5;
 
 	intel_dp_check_edp(intel_dp);
 	/* The clock divider is based off the hrawclk,
@@ -378,15 +368,10 @@ intel_dp_aux_ch(struct intel_dp *intel_dp,
 		else
 			aux_clock_divider = 225; /* eDP input clock at 450Mhz */
 	} else if (HAS_PCH_SPLIT(dev))
-		aux_clock_divider = 62; /* IRL input clock fixed at 125Mhz */
+		aux_clock_divider = 63; /* IRL input clock fixed at 125Mhz */
 	else
 		aux_clock_divider = intel_hrawclk(dev) / 2;
 
-	if (IS_GEN6(dev))
-		precharge = 3;
-	else
-		precharge = 5;
-
 	/* Try to wait for any previous AUX channel activity */
 	for (try = 0; try < 3; try++) {
 		status = I915_READ(ch_ctl);
@@ -431,6 +416,10 @@ intel_dp_aux_ch(struct intel_dp *intel_dp,
 			   DP_AUX_CH_CTL_DONE |
 			   DP_AUX_CH_CTL_TIME_OUT_ERROR |
 			   DP_AUX_CH_CTL_RECEIVE_ERROR);
+
+		if (status & (DP_AUX_CH_CTL_TIME_OUT_ERROR |
+			      DP_AUX_CH_CTL_RECEIVE_ERROR))
+			continue;
 		if (status & DP_AUX_CH_CTL_DONE)
 			break;
 	}
@@ -683,7 +672,7 @@ intel_dp_mode_fixup(struct drm_encoder *encoder, struct drm_display_mode *mode,
 	int lane_count, clock;
 	int max_lane_count = intel_dp_max_lane_count(intel_dp);
 	int max_clock = intel_dp_max_link_bw(intel_dp) == DP_LINK_BW_2_7 ? 1 : 0;
-	int bpp = mode->private_flags & INTEL_MODE_DP_FORCE_6BPC ? 18 : 0;
+	int bpp = mode->private_flags & INTEL_MODE_DP_FORCE_6BPC ? 18 : 24;
 	static int bws[2] = { DP_LINK_BW_1_62, DP_LINK_BW_2_7 };
 
 	if (is_edp(intel_dp) && intel_dp->panel_fixed_mode) {
@@ -701,7 +690,7 @@ intel_dp_mode_fixup(struct drm_encoder *encoder, struct drm_display_mode *mode,
 		for (clock = 0; clock <= max_clock; clock++) {
 			int link_avail = intel_dp_max_data_rate(intel_dp_link_clock(bws[clock]), lane_count);
 
-			if (intel_dp_link_required(intel_dp, mode->clock, bpp)
+			if (intel_dp_link_required(mode->clock, bpp)
 					<= link_avail) {
 				intel_dp->link_bw = bws[clock];
 				intel_dp->lane_count = lane_count;
@@ -2127,8 +2116,8 @@ intel_dp_detect(struct drm_connector *connector, bool force)
 	if (status != connector_status_connected)
 		return status;
 
-	if (intel_dp->force_audio) {
-		intel_dp->has_audio = intel_dp->force_audio > 0;
+	if (intel_dp->force_audio != HDMI_AUDIO_AUTO) {
+		intel_dp->has_audio = (intel_dp->force_audio == HDMI_AUDIO_ON);
 	} else {
 		edid = intel_dp_get_edid(connector, &intel_dp->adapter);
 		if (edid) {
@@ -2228,10 +2217,10 @@ intel_dp_set_property(struct drm_connector *connector,
 
 		intel_dp->force_audio = i;
 
-		if (i == 0)
+		if (i == HDMI_AUDIO_AUTO)
 			has_audio = intel_dp_detect_audio(connector);
 		else
-			has_audio = i > 0;
+			has_audio = (i == HDMI_AUDIO_ON);
 
 		if (has_audio == intel_dp->has_audio)
 			return 0;
diff --git a/drivers/gpu/drm/i915/intel_drv.h b/drivers/gpu/drm/i915/intel_drv.h
index 1348705faf6b..5a14149b3794 100644
--- a/drivers/gpu/drm/i915/intel_drv.h
+++ b/drivers/gpu/drm/i915/intel_drv.h
@@ -374,6 +374,7 @@ extern void intel_init_emon(struct drm_device *dev);
 extern int intel_pin_and_fence_fb_obj(struct drm_device *dev,
 				      struct drm_i915_gem_object *obj,
 				      struct intel_ring_buffer *pipelined);
+extern void intel_unpin_fb_obj(struct drm_i915_gem_object *obj);
 
 extern int intel_framebuffer_init(struct drm_device *dev,
 				  struct intel_framebuffer *ifb,
@@ -381,7 +382,7 @@ extern int intel_framebuffer_init(struct drm_device *dev,
 				  struct drm_i915_gem_object *obj);
 extern int intel_fbdev_init(struct drm_device *dev);
 extern void intel_fbdev_fini(struct drm_device *dev);
-
+extern void intel_fbdev_set_suspend(struct drm_device *dev, int state);
 extern void intel_prepare_page_flip(struct drm_device *dev, int plane);
 extern void intel_finish_page_flip(struct drm_device *dev, int pipe);
 extern void intel_finish_page_flip_plane(struct drm_device *dev, int plane);
diff --git a/drivers/gpu/drm/i915/intel_dvo.c b/drivers/gpu/drm/i915/intel_dvo.c
index 6eda1b51c636..020a7d7f744d 100644
--- a/drivers/gpu/drm/i915/intel_dvo.c
+++ b/drivers/gpu/drm/i915/intel_dvo.c
@@ -157,7 +157,6 @@ static bool intel_dvo_mode_fixup(struct drm_encoder *encoder,
 		C(vsync_end);
 		C(vtotal);
 		C(clock);
-		drm_mode_set_crtcinfo(adjusted_mode, CRTC_INTERLACE_HALVE_V);
 #undef C
 	}
 
diff --git a/drivers/gpu/drm/i915/intel_fb.c b/drivers/gpu/drm/i915/intel_fb.c
index 571375a3eef4..19ecd78b8a2c 100644
--- a/drivers/gpu/drm/i915/intel_fb.c
+++ b/drivers/gpu/drm/i915/intel_fb.c
@@ -152,11 +152,7 @@ static int intelfb_create(struct intel_fbdev *ifbdev,
 	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);
 
-	info->pixmap.size = 64*1024;
-	info->pixmap.buf_align = 8;
-	info->pixmap.access_align = 32;
-	info->pixmap.flags = FB_PIXMAP_SYSTEM;
-	info->pixmap.scan_align = 1;
+	/* Use default scratch pixmap (info->pixmap.flags = FB_PIXMAP_SYSTEM) */
 
 	DRM_DEBUG_KMS("allocated %dx%d fb: 0x%08x, bo %p\n",
 		      fb->width, fb->height,
@@ -258,6 +254,16 @@ void intel_fbdev_fini(struct drm_device *dev)
 	kfree(dev_priv->fbdev);
 	dev_priv->fbdev = NULL;
 }
+
+void intel_fbdev_set_suspend(struct drm_device *dev, int state)
+{
+	drm_i915_private_t *dev_priv = dev->dev_private;
+	if (!dev_priv->fbdev)
+		return;
+
+	fb_set_suspend(dev_priv->fbdev->helper.fbdev, state);
+}
+
 MODULE_LICENSE("GPL and additional rights");
 
 void intel_fb_output_poll_changed(struct drm_device *dev)
diff --git a/drivers/gpu/drm/i915/intel_hdmi.c b/drivers/gpu/drm/i915/intel_hdmi.c
index 64541f7ef900..cae3e5f17a49 100644
--- a/drivers/gpu/drm/i915/intel_hdmi.c
+++ b/drivers/gpu/drm/i915/intel_hdmi.c
@@ -44,7 +44,7 @@ struct intel_hdmi {
 	uint32_t color_range;
 	bool has_hdmi_sink;
 	bool has_audio;
-	int force_audio;
+	enum hdmi_force_audio force_audio;
 	void (*write_infoframe)(struct drm_encoder *encoder,
 				struct dip_infoframe *frame);
 };
@@ -339,7 +339,9 @@ intel_hdmi_detect(struct drm_connector *connector, bool force)
 	if (edid) {
 		if (edid->input & DRM_EDID_INPUT_DIGITAL) {
 			status = connector_status_connected;
-			intel_hdmi->has_hdmi_sink = drm_detect_hdmi_monitor(edid);
+			if (intel_hdmi->force_audio != HDMI_AUDIO_OFF_DVI)
+				intel_hdmi->has_hdmi_sink =
+						drm_detect_hdmi_monitor(edid);
 			intel_hdmi->has_audio = drm_detect_monitor_audio(edid);
 		}
 		connector->display_info.raw_edid = NULL;
@@ -347,8 +349,9 @@ intel_hdmi_detect(struct drm_connector *connector, bool force)
 	}
 
 	if (status == connector_status_connected) {
-		if (intel_hdmi->force_audio)
-			intel_hdmi->has_audio = intel_hdmi->force_audio > 0;
+		if (intel_hdmi->force_audio != HDMI_AUDIO_AUTO)
+			intel_hdmi->has_audio =
+				(intel_hdmi->force_audio == HDMI_AUDIO_ON);
 	}
 
 	return status;
@@ -402,7 +405,7 @@ intel_hdmi_set_property(struct drm_connector *connector,
 		return ret;
 
 	if (property == dev_priv->force_audio_property) {
-		int i = val;
+		enum hdmi_force_audio i = val;
 		bool has_audio;
 
 		if (i == intel_hdmi->force_audio)
@@ -410,13 +413,13 @@ intel_hdmi_set_property(struct drm_connector *connector,
 
 		intel_hdmi->force_audio = i;
 
-		if (i == 0)
+		if (i == HDMI_AUDIO_AUTO)
 			has_audio = intel_hdmi_detect_audio(connector);
 		else
-			has_audio = i > 0;
+			has_audio = (i == HDMI_AUDIO_ON);
 
-		if (has_audio == intel_hdmi->has_audio)
-			return 0;
+		if (i == HDMI_AUDIO_OFF_DVI)
+			intel_hdmi->has_hdmi_sink = 0;
 
 		intel_hdmi->has_audio = has_audio;
 		goto done;
@@ -514,7 +517,7 @@ void intel_hdmi_init(struct drm_device *dev, int sdvox_reg)
 	intel_encoder->type = INTEL_OUTPUT_HDMI;
 
 	connector->polled = DRM_CONNECTOR_POLL_HPD;
-	connector->interlace_allowed = 0;
+	connector->interlace_allowed = 1;
 	connector->doublescan_allowed = 0;
 	intel_encoder->crtc_mask = (1 << 0) | (1 << 1) | (1 << 2);
 
diff --git a/drivers/gpu/drm/i915/intel_i2c.c b/drivers/gpu/drm/i915/intel_i2c.c
index d30ccccb9d73..601c86e664af 100644
--- a/drivers/gpu/drm/i915/intel_i2c.c
+++ b/drivers/gpu/drm/i915/intel_i2c.c
@@ -37,7 +37,7 @@
 
 /* Intel GPIO access functions */
 
-#define I2C_RISEFALL_TIME 20
+#define I2C_RISEFALL_TIME 10
 
 static inline struct intel_gmbus *
 to_intel_gmbus(struct i2c_adapter *i2c)
@@ -45,13 +45,6 @@ to_intel_gmbus(struct i2c_adapter *i2c)
 	return container_of(i2c, struct intel_gmbus, adapter);
 }
 
-struct intel_gpio {
-	struct i2c_adapter adapter;
-	struct i2c_algo_bit_data algo;
-	struct drm_i915_private *dev_priv;
-	u32 reg;
-};
-
 void
 intel_i2c_reset(struct drm_device *dev)
 {
@@ -78,15 +71,15 @@ static void intel_i2c_quirk_set(struct drm_i915_private *dev_priv, bool enable)
 	I915_WRITE(DSPCLK_GATE_D, val);
 }
 
-static u32 get_reserved(struct intel_gpio *gpio)
+static u32 get_reserved(struct intel_gmbus *bus)
 {
-	struct drm_i915_private *dev_priv = gpio->dev_priv;
+	struct drm_i915_private *dev_priv = bus->dev_priv;
 	struct drm_device *dev = dev_priv->dev;
 	u32 reserved = 0;
 
 	/* On most chips, these bits must be preserved in software. */
 	if (!IS_I830(dev) && !IS_845G(dev))
-		reserved = I915_READ_NOTRACE(gpio->reg) &
+		reserved = I915_READ_NOTRACE(bus->gpio_reg) &
 					     (GPIO_DATA_PULLUP_DISABLE |
 					      GPIO_CLOCK_PULLUP_DISABLE);
 
@@ -95,29 +88,29 @@ static u32 get_reserved(struct intel_gpio *gpio)
 
 static int get_clock(void *data)
 {
-	struct intel_gpio *gpio = data;
-	struct drm_i915_private *dev_priv = gpio->dev_priv;
-	u32 reserved = get_reserved(gpio);
-	I915_WRITE_NOTRACE(gpio->reg, reserved | GPIO_CLOCK_DIR_MASK);
-	I915_WRITE_NOTRACE(gpio->reg, reserved);
-	return (I915_READ_NOTRACE(gpio->reg) & GPIO_CLOCK_VAL_IN) != 0;
+	struct intel_gmbus *bus = data;
+	struct drm_i915_private *dev_priv = bus->dev_priv;
+	u32 reserved = get_reserved(bus);
+	I915_WRITE_NOTRACE(bus->gpio_reg, reserved | GPIO_CLOCK_DIR_MASK);
+	I915_WRITE_NOTRACE(bus->gpio_reg, reserved);
+	return (I915_READ_NOTRACE(bus->gpio_reg) & GPIO_CLOCK_VAL_IN) != 0;
 }
 
 static int get_data(void *data)
 {
-	struct intel_gpio *gpio = data;
-	struct drm_i915_private *dev_priv = gpio->dev_priv;
-	u32 reserved = get_reserved(gpio);
-	I915_WRITE_NOTRACE(gpio->reg, reserved | GPIO_DATA_DIR_MASK);
-	I915_WRITE_NOTRACE(gpio->reg, reserved);
-	return (I915_READ_NOTRACE(gpio->reg) & GPIO_DATA_VAL_IN) != 0;
+	struct intel_gmbus *bus = data;
+	struct drm_i915_private *dev_priv = bus->dev_priv;
+	u32 reserved = get_reserved(bus);
+	I915_WRITE_NOTRACE(bus->gpio_reg, reserved | GPIO_DATA_DIR_MASK);
+	I915_WRITE_NOTRACE(bus->gpio_reg, reserved);
+	return (I915_READ_NOTRACE(bus->gpio_reg) & GPIO_DATA_VAL_IN) != 0;
 }
 
 static void set_clock(void *data, int state_high)
 {
-	struct intel_gpio *gpio = data;
-	struct drm_i915_private *dev_priv = gpio->dev_priv;
-	u32 reserved = get_reserved(gpio);
+	struct intel_gmbus *bus = data;
+	struct drm_i915_private *dev_priv = bus->dev_priv;
+	u32 reserved = get_reserved(bus);
 	u32 clock_bits;
 
 	if (state_high)
@@ -126,15 +119,15 @@ static void set_clock(void *data, int state_high)
 		clock_bits = GPIO_CLOCK_DIR_OUT | GPIO_CLOCK_DIR_MASK |
 			GPIO_CLOCK_VAL_MASK;
 
-	I915_WRITE_NOTRACE(gpio->reg, reserved | clock_bits);
-	POSTING_READ(gpio->reg);
+	I915_WRITE_NOTRACE(bus->gpio_reg, reserved | clock_bits);
+	POSTING_READ(bus->gpio_reg);
 }
 
 static void set_data(void *data, int state_high)
 {
-	struct intel_gpio *gpio = data;
-	struct drm_i915_private *dev_priv = gpio->dev_priv;
-	u32 reserved = get_reserved(gpio);
+	struct intel_gmbus *bus = data;
+	struct drm_i915_private *dev_priv = bus->dev_priv;
+	u32 reserved = get_reserved(bus);
 	u32 data_bits;
 
 	if (state_high)
@@ -143,13 +136,14 @@ static void set_data(void *data, int state_high)
 		data_bits = GPIO_DATA_DIR_OUT | GPIO_DATA_DIR_MASK |
 			GPIO_DATA_VAL_MASK;
 
-	I915_WRITE_NOTRACE(gpio->reg, reserved | data_bits);
-	POSTING_READ(gpio->reg);
+	I915_WRITE_NOTRACE(bus->gpio_reg, reserved | data_bits);
+	POSTING_READ(bus->gpio_reg);
 }
 
-static struct i2c_adapter *
-intel_gpio_create(struct drm_i915_private *dev_priv, u32 pin)
+static bool
+intel_gpio_setup(struct intel_gmbus *bus, u32 pin)
 {
+	struct drm_i915_private *dev_priv = bus->dev_priv;
 	static const int map_pin_to_reg[] = {
 		0,
 		GPIOB,
@@ -160,65 +154,48 @@ intel_gpio_create(struct drm_i915_private *dev_priv, u32 pin)
 		0,
 		GPIOF,
 	};
-	struct intel_gpio *gpio;
+	struct i2c_algo_bit_data *algo;
 
 	if (pin >= ARRAY_SIZE(map_pin_to_reg) || !map_pin_to_reg[pin])
-		return NULL;
+		return false;
 
-	gpio = kzalloc(sizeof(struct intel_gpio), GFP_KERNEL);
-	if (gpio == NULL)
-		return NULL;
+	algo = &bus->bit_algo;
 
-	gpio->reg = map_pin_to_reg[pin];
+	bus->gpio_reg = map_pin_to_reg[pin];
 	if (HAS_PCH_SPLIT(dev_priv->dev))
-		gpio->reg += PCH_GPIOA - GPIOA;
-	gpio->dev_priv = dev_priv;
-
-	snprintf(gpio->adapter.name, sizeof(gpio->adapter.name),
-		 "i915 GPIO%c", "?BACDE?F"[pin]);
-	gpio->adapter.owner = THIS_MODULE;
-	gpio->adapter.algo_data	= &gpio->algo;
-	gpio->adapter.dev.parent = &dev_priv->dev->pdev->dev;
-	gpio->algo.setsda = set_data;
-	gpio->algo.setscl = set_clock;
-	gpio->algo.getsda = get_data;
-	gpio->algo.getscl = get_clock;
-	gpio->algo.udelay = I2C_RISEFALL_TIME;
-	gpio->algo.timeout = usecs_to_jiffies(2200);
-	gpio->algo.data = gpio;
-
-	if (i2c_bit_add_bus(&gpio->adapter))
-		goto out_free;
-
-	return &gpio->adapter;
-
-out_free:
-	kfree(gpio);
-	return NULL;
+		bus->gpio_reg += PCH_GPIOA - GPIOA;
+
+	bus->adapter.algo_data = algo;
+	algo->setsda = set_data;
+	algo->setscl = set_clock;
+	algo->getsda = get_data;
+	algo->getscl = get_clock;
+	algo->udelay = I2C_RISEFALL_TIME;
+	algo->timeout = usecs_to_jiffies(2200);
+	algo->data = bus;
+
+	return true;
 }
 
 static int
-intel_i2c_quirk_xfer(struct drm_i915_private *dev_priv,
-		     struct i2c_adapter *adapter,
+intel_i2c_quirk_xfer(struct intel_gmbus *bus,
 		     struct i2c_msg *msgs,
 		     int num)
 {
-	struct intel_gpio *gpio = container_of(adapter,
-					       struct intel_gpio,
-					       adapter);
+	struct drm_i915_private *dev_priv = bus->dev_priv;
 	int ret;
 
 	intel_i2c_reset(dev_priv->dev);
 
 	intel_i2c_quirk_set(dev_priv, true);
-	set_data(gpio, 1);
-	set_clock(gpio, 1);
+	set_data(bus, 1);
+	set_clock(bus, 1);
 	udelay(I2C_RISEFALL_TIME);
 
-	ret = adapter->algo->master_xfer(adapter, msgs, num);
+	ret = i2c_bit_algo.master_xfer(&bus->adapter, msgs, num);
 
-	set_data(gpio, 1);
-	set_clock(gpio, 1);
+	set_data(bus, 1);
+	set_clock(bus, 1);
 	intel_i2c_quirk_set(dev_priv, false);
 
 	return ret;
@@ -232,12 +209,15 @@ gmbus_xfer(struct i2c_adapter *adapter,
 	struct intel_gmbus *bus = container_of(adapter,
 					       struct intel_gmbus,
 					       adapter);
-	struct drm_i915_private *dev_priv = adapter->algo_data;
-	int i, reg_offset;
+	struct drm_i915_private *dev_priv = bus->dev_priv;
+	int i, reg_offset, ret;
 
-	if (bus->force_bit)
-		return intel_i2c_quirk_xfer(dev_priv,
-					    bus->force_bit, msgs, num);
+	mutex_lock(&dev_priv->gmbus_mutex);
+
+	if (bus->force_bit) {
+		ret = intel_i2c_quirk_xfer(bus, msgs, num);
+		goto out;
+	}
 
 	reg_offset = HAS_PCH_SPLIT(dev_priv->dev) ? PCH_GMBUS0 - GMBUS0 : 0;
 
@@ -249,7 +229,8 @@ gmbus_xfer(struct i2c_adapter *adapter,
 
 		if (msgs[i].flags & I2C_M_RD) {
 			I915_WRITE(GMBUS1 + reg_offset,
-				   GMBUS_CYCLE_WAIT | (i + 1 == num ? GMBUS_CYCLE_STOP : 0) |
+				   GMBUS_CYCLE_WAIT |
+				   (i + 1 == num ? GMBUS_CYCLE_STOP : 0) |
 				   (len << GMBUS_BYTE_COUNT_SHIFT) |
 				   (msgs[i].addr << GMBUS_SLAVE_ADDR_SHIFT) |
 				   GMBUS_SLAVE_READ | GMBUS_SW_RDY);
@@ -278,7 +259,8 @@ gmbus_xfer(struct i2c_adapter *adapter,
 
 			I915_WRITE(GMBUS3 + reg_offset, val);
 			I915_WRITE(GMBUS1 + reg_offset,
-				   (i + 1 == num ? GMBUS_CYCLE_STOP : GMBUS_CYCLE_WAIT) |
+				   GMBUS_CYCLE_WAIT |
+				   (i + 1 == num ? GMBUS_CYCLE_STOP : 0) |
 				   (msgs[i].len << GMBUS_BYTE_COUNT_SHIFT) |
 				   (msgs[i].addr << GMBUS_SLAVE_ADDR_SHIFT) |
 				   GMBUS_SLAVE_WRITE | GMBUS_SW_RDY);
@@ -317,11 +299,15 @@ clear_err:
 	I915_WRITE(GMBUS1 + reg_offset, 0);
 
 done:
-	/* Mark the GMBUS interface as disabled. We will re-enable it at the
-	 * start of the next xfer, till then let it sleep.
+	/* Mark the GMBUS interface as disabled after waiting for idle.
+	 * We will re-enable it at the start of the next xfer,
+	 * till then let it sleep.
 	 */
+	if (wait_for((I915_READ(GMBUS2 + reg_offset) & GMBUS_ACTIVE) == 0, 10))
+		DRM_INFO("GMBUS timed out waiting for idle\n");
 	I915_WRITE(GMBUS0 + reg_offset, 0);
-	return i;
+	ret = i;
+	goto out;
 
 timeout:
 	DRM_INFO("GMBUS timed out, falling back to bit banging on pin %d [%s]\n",
@@ -329,23 +315,21 @@ timeout:
 	I915_WRITE(GMBUS0 + reg_offset, 0);
 
 	/* Hardware may not support GMBUS over these pins? Try GPIO bitbanging instead. */
-	bus->force_bit = intel_gpio_create(dev_priv, bus->reg0 & 0xff);
-	if (!bus->force_bit)
-		return -ENOMEM;
-
-	return intel_i2c_quirk_xfer(dev_priv, bus->force_bit, msgs, num);
+	if (!bus->has_gpio) {
+		ret = -EIO;
+	} else {
+		bus->force_bit = true;
+		ret = intel_i2c_quirk_xfer(bus, msgs, num);
+	}
+out:
+	mutex_unlock(&dev_priv->gmbus_mutex);
+	return ret;
 }
 
 static u32 gmbus_func(struct i2c_adapter *adapter)
 {
-	struct intel_gmbus *bus = container_of(adapter,
-					       struct intel_gmbus,
-					       adapter);
-
-	if (bus->force_bit)
-		bus->force_bit->algo->functionality(bus->force_bit);
-
-	return (I2C_FUNC_I2C | I2C_FUNC_SMBUS_EMUL |
+	return i2c_bit_algo.functionality(adapter) &
+		(I2C_FUNC_I2C | I2C_FUNC_SMBUS_EMUL |
 		/* I2C_FUNC_10BIT_ADDR | */
 		I2C_FUNC_SMBUS_READ_BLOCK_DATA |
 		I2C_FUNC_SMBUS_BLOCK_PROC_CALL);
@@ -375,11 +359,13 @@ int intel_setup_gmbus(struct drm_device *dev)
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	int ret, i;
 
-	dev_priv->gmbus = kcalloc(sizeof(struct intel_gmbus), GMBUS_NUM_PORTS,
+	dev_priv->gmbus = kcalloc(GMBUS_NUM_PORTS, sizeof(struct intel_gmbus),
 				  GFP_KERNEL);
 	if (dev_priv->gmbus == NULL)
 		return -ENOMEM;
 
+	mutex_init(&dev_priv->gmbus_mutex);
+
 	for (i = 0; i < GMBUS_NUM_PORTS; i++) {
 		struct intel_gmbus *bus = &dev_priv->gmbus[i];
 
@@ -391,7 +377,7 @@ int intel_setup_gmbus(struct drm_device *dev)
 			 names[i]);
 
 		bus->adapter.dev.parent = &dev->pdev->dev;
-		bus->adapter.algo_data	= dev_priv;
+		bus->dev_priv = dev_priv;
 
 		bus->adapter.algo = &gmbus_algorithm;
 		ret = i2c_add_adapter(&bus->adapter);
@@ -401,8 +387,11 @@ int intel_setup_gmbus(struct drm_device *dev)
 		/* By default use a conservative clock rate */
 		bus->reg0 = i | GMBUS_RATE_100KHZ;
 
+		bus->has_gpio = intel_gpio_setup(bus, i);
+
 		/* XXX force bit banging until GMBUS is fully debugged */
-		bus->force_bit = intel_gpio_create(dev_priv, i);
+		if (bus->has_gpio && IS_GEN2(dev))
+			bus->force_bit = true;
 	}
 
 	intel_i2c_reset(dev_priv->dev);
@@ -430,19 +419,8 @@ void intel_gmbus_force_bit(struct i2c_adapter *adapter, bool force_bit)
 {
 	struct intel_gmbus *bus = to_intel_gmbus(adapter);
 
-	if (force_bit) {
-		if (bus->force_bit == NULL) {
-			struct drm_i915_private *dev_priv = adapter->algo_data;
-			bus->force_bit = intel_gpio_create(dev_priv,
-							   bus->reg0 & 0xff);
-		}
-	} else {
-		if (bus->force_bit) {
-			i2c_del_adapter(bus->force_bit);
-			kfree(bus->force_bit);
-			bus->force_bit = NULL;
-		}
-	}
+	if (bus->has_gpio)
+		bus->force_bit = force_bit;
 }
 
 void intel_teardown_gmbus(struct drm_device *dev)
@@ -455,10 +433,6 @@ void intel_teardown_gmbus(struct drm_device *dev)
 
 	for (i = 0; i < GMBUS_NUM_PORTS; i++) {
 		struct intel_gmbus *bus = &dev_priv->gmbus[i];
-		if (bus->force_bit) {
-			i2c_del_adapter(bus->force_bit);
-			kfree(bus->force_bit);
-		}
 		i2c_del_adapter(&bus->adapter);
 	}
 
diff --git a/drivers/gpu/drm/i915/intel_lvds.c b/drivers/gpu/drm/i915/intel_lvds.c
index 798f6e1aa544..95db2e988227 100644
--- a/drivers/gpu/drm/i915/intel_lvds.c
+++ b/drivers/gpu/drm/i915/intel_lvds.c
@@ -694,6 +694,14 @@ static const struct dmi_system_id intel_no_lvds[] = {
 	},
 	{
 		.callback = intel_no_lvds_dmi_callback,
+                .ident = "AOpen i45GMx-I",
+                .matches = {
+                        DMI_MATCH(DMI_BOARD_VENDOR, "AOpen"),
+                        DMI_MATCH(DMI_BOARD_NAME, "i45GMx-I"),
+                },
+        },
+	{
+		.callback = intel_no_lvds_dmi_callback,
 		.ident = "Aopen i945GTt-VFA",
 		.matches = {
 			DMI_MATCH(DMI_PRODUCT_VERSION, "AO00001JW"),
@@ -731,6 +739,30 @@ static const struct dmi_system_id intel_no_lvds[] = {
 			DMI_MATCH(DMI_BOARD_NAME, "AT5NM10T-I"),
 		},
 	},
+	{
+		.callback = intel_no_lvds_dmi_callback,
+		.ident = "Hewlett-Packard t5745",
+		.matches = {
+			DMI_MATCH(DMI_BOARD_VENDOR, "Hewlett-Packard"),
+			DMI_MATCH(DMI_BOARD_NAME, "hp t5745"),
+		},
+	},
+	{
+		.callback = intel_no_lvds_dmi_callback,
+		.ident = "Hewlett-Packard st5747",
+		.matches = {
+			DMI_MATCH(DMI_BOARD_VENDOR, "Hewlett-Packard"),
+			DMI_MATCH(DMI_BOARD_NAME, "hp st5747"),
+		},
+	},
+	{
+		.callback = intel_no_lvds_dmi_callback,
+		.ident = "MSI Wind Box DC500",
+		.matches = {
+			DMI_MATCH(DMI_BOARD_VENDOR, "MICRO-STAR INTERNATIONAL CO., LTD"),
+			DMI_MATCH(DMI_BOARD_NAME, "MS-7469"),
+		},
+	},
 
 	{ }	/* terminating entry */
 };
@@ -836,6 +868,18 @@ static bool lvds_is_present_in_vbt(struct drm_device *dev,
 	return false;
 }
 
+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))
+		return true;
+
+	/* Otherwise LVDS was only attached to mobile products,
+	 * except for the inglorious 830gm */
+	return IS_MOBILE(dev) && !IS_I830(dev);
+}
+
 /**
  * intel_lvds_init - setup LVDS connectors on this device
  * @dev: drm device
@@ -857,6 +901,9 @@ bool intel_lvds_init(struct drm_device *dev)
 	int pipe;
 	u8 pin;
 
+	if (!intel_lvds_supported(dev))
+		return false;
+
 	/* Skip init on machines we know falsely report LVDS */
 	if (dmi_check_system(intel_no_lvds))
 		return false;
diff --git a/drivers/gpu/drm/i915/intel_modes.c b/drivers/gpu/drm/i915/intel_modes.c
index be2c6fe07d12..d1928e79d9b6 100644
--- a/drivers/gpu/drm/i915/intel_modes.c
+++ b/drivers/gpu/drm/i915/intel_modes.c
@@ -28,6 +28,7 @@
 #include <linux/fb.h>
 #include <drm/drm_edid.h>
 #include "drmP.h"
+#include "drm_edid.h"
 #include "intel_drv.h"
 #include "i915_drv.h"
 
@@ -42,13 +43,13 @@ bool intel_ddc_probe(struct intel_encoder *intel_encoder, int ddc_bus)
 	u8 buf[2];
 	struct i2c_msg msgs[] = {
 		{
-			.addr = 0x50,
+			.addr = DDC_ADDR,
 			.flags = 0,
 			.len = 1,
 			.buf = out_buf,
 		},
 		{
-			.addr = 0x50,
+			.addr = DDC_ADDR,
 			.flags = I2C_M_RD,
 			.len = 1,
 			.buf = buf,
@@ -83,10 +84,11 @@ int intel_ddc_get_modes(struct drm_connector *connector,
 	return ret;
 }
 
-static const char *force_audio_names[] = {
-	"off",
-	"auto",
-	"on",
+static const struct drm_prop_enum_list force_audio_names[] = {
+	{ HDMI_AUDIO_OFF_DVI, "force-dvi" },
+	{ HDMI_AUDIO_OFF, "off" },
+	{ HDMI_AUDIO_AUTO, "auto" },
+	{ HDMI_AUDIO_ON, "on" },
 };
 
 void
@@ -95,27 +97,24 @@ intel_attach_force_audio_property(struct drm_connector *connector)
 	struct drm_device *dev = connector->dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct drm_property *prop;
-	int i;
 
 	prop = dev_priv->force_audio_property;
 	if (prop == NULL) {
-		prop = drm_property_create(dev, DRM_MODE_PROP_ENUM,
+		prop = drm_property_create_enum(dev, 0,
 					   "audio",
+					   force_audio_names,
 					   ARRAY_SIZE(force_audio_names));
 		if (prop == NULL)
 			return;
 
-		for (i = 0; i < ARRAY_SIZE(force_audio_names); i++)
-			drm_property_add_enum(prop, i, i-1, force_audio_names[i]);
-
 		dev_priv->force_audio_property = prop;
 	}
 	drm_connector_attach_property(connector, prop, 0);
 }
 
-static const char *broadcast_rgb_names[] = {
-	"Full",
-	"Limited 16:235",
+static const struct drm_prop_enum_list broadcast_rgb_names[] = {
+	{ 0, "Full" },
+	{ 1, "Limited 16:235" },
 };
 
 void
@@ -124,19 +123,16 @@ intel_attach_broadcast_rgb_property(struct drm_connector *connector)
 	struct drm_device *dev = connector->dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct drm_property *prop;
-	int i;
 
 	prop = dev_priv->broadcast_rgb_property;
 	if (prop == NULL) {
-		prop = drm_property_create(dev, DRM_MODE_PROP_ENUM,
+		prop = drm_property_create_enum(dev, DRM_MODE_PROP_ENUM,
 					   "Broadcast RGB",
+					   broadcast_rgb_names,
 					   ARRAY_SIZE(broadcast_rgb_names));
 		if (prop == NULL)
 			return;
 
-		for (i = 0; i < ARRAY_SIZE(broadcast_rgb_names); i++)
-			drm_property_add_enum(prop, i, i, broadcast_rgb_names[i]);
-
 		dev_priv->broadcast_rgb_property = prop;
 	}
 
diff --git a/drivers/gpu/drm/i915/intel_overlay.c b/drivers/gpu/drm/i915/intel_overlay.c
index cdf17d4cc1f7..80b331c322fb 100644
--- a/drivers/gpu/drm/i915/intel_overlay.c
+++ b/drivers/gpu/drm/i915/intel_overlay.c
@@ -25,8 +25,6 @@
  *
  * Derived from Xorg ddx, xf86-video-intel, src/i830_video.c
  */
-
-#include <linux/seq_file.h>
 #include "drmP.h"
 #include "drm.h"
 #include "i915_drm.h"
@@ -227,7 +225,8 @@ static int intel_overlay_do_wait_request(struct intel_overlay *overlay,
 	}
 	overlay->last_flip_req = request->seqno;
 	overlay->flip_tail = tail;
-	ret = i915_wait_request(LP_RING(dev_priv), overlay->last_flip_req);
+	ret = i915_wait_request(LP_RING(dev_priv), overlay->last_flip_req,
+				true);
 	if (ret)
 		return ret;
 
@@ -263,7 +262,7 @@ i830_activate_pipe_a(struct drm_device *dev)
 	DRM_DEBUG_DRIVER("Enabling pipe A in order to enable overlay\n");
 
 	mode = drm_mode_duplicate(dev, &vesa_640x480);
-	drm_mode_set_crtcinfo(mode, CRTC_INTERLACE_HALVE_V);
+	drm_mode_set_crtcinfo(mode, 0);
 	if (!drm_crtc_helper_set_mode(&crtc->base, mode,
 				       crtc->base.x, crtc->base.y,
 				       crtc->base.fb))
@@ -448,7 +447,8 @@ static int intel_overlay_recover_from_interrupt(struct intel_overlay *overlay)
 	if (overlay->last_flip_req == 0)
 		return 0;
 
-	ret = i915_wait_request(LP_RING(dev_priv), overlay->last_flip_req);
+	ret = i915_wait_request(LP_RING(dev_priv), overlay->last_flip_req,
+				true);
 	if (ret)
 		return ret;
 
@@ -935,10 +935,10 @@ static int check_overlay_dst(struct intel_overlay *overlay,
 {
 	struct drm_display_mode *mode = &overlay->crtc->base.mode;
 
-	if (rec->dst_x < mode->crtc_hdisplay &&
-	    rec->dst_x + rec->dst_width <= mode->crtc_hdisplay &&
-	    rec->dst_y < mode->crtc_vdisplay &&
-	    rec->dst_y + rec->dst_height <= mode->crtc_vdisplay)
+	if (rec->dst_x < mode->hdisplay &&
+	    rec->dst_x + rec->dst_width <= mode->hdisplay &&
+	    rec->dst_y < mode->vdisplay &&
+	    rec->dst_y + rec->dst_height <= mode->vdisplay)
 		return 0;
 	else
 		return -EINVAL;
diff --git a/drivers/gpu/drm/i915/intel_panel.c b/drivers/gpu/drm/i915/intel_panel.c
index 04d79fd1dc9d..230a141dbea3 100644
--- a/drivers/gpu/drm/i915/intel_panel.c
+++ b/drivers/gpu/drm/i915/intel_panel.c
@@ -48,7 +48,7 @@ intel_fixed_panel_mode(struct drm_display_mode *fixed_mode,
 
 	adjusted_mode->clock = fixed_mode->clock;
 
-	drm_mode_set_crtcinfo(adjusted_mode, CRTC_INTERLACE_HALVE_V);
+	drm_mode_set_crtcinfo(adjusted_mode, 0);
 }
 
 /* adjusted_mode has been preset to be the panel's fixed mode */
@@ -141,8 +141,8 @@ static u32 i915_read_blc_pwm_ctl(struct drm_i915_private *dev_priv)
 			dev_priv->saveBLC_PWM_CTL2 = val;
 		} else if (val == 0) {
 			I915_WRITE(BLC_PWM_PCH_CTL2,
-				   dev_priv->saveBLC_PWM_CTL);
-			val = dev_priv->saveBLC_PWM_CTL;
+				   dev_priv->saveBLC_PWM_CTL2);
+			val = dev_priv->saveBLC_PWM_CTL2;
 		}
 	} else {
 		val = I915_READ(BLC_PWM_CTL);
diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.c b/drivers/gpu/drm/i915/intel_ringbuffer.c
index 1ab842c6032e..e25581a9f60f 100644
--- a/drivers/gpu/drm/i915/intel_ringbuffer.c
+++ b/drivers/gpu/drm/i915/intel_ringbuffer.c
@@ -52,20 +52,6 @@ static inline int ring_space(struct intel_ring_buffer *ring)
 	return space;
 }
 
-static u32 i915_gem_get_seqno(struct drm_device *dev)
-{
-	drm_i915_private_t *dev_priv = dev->dev_private;
-	u32 seqno;
-
-	seqno = dev_priv->next_seqno;
-
-	/* reserve 0 for non-seqno */
-	if (++dev_priv->next_seqno == 0)
-		dev_priv->next_seqno = 1;
-
-	return seqno;
-}
-
 static int
 render_ring_flush(struct intel_ring_buffer *ring,
 		  u32	invalidate_domains,
@@ -301,7 +287,7 @@ static int init_ring_common(struct intel_ring_buffer *ring)
 
 	I915_WRITE_CTL(ring,
 			((ring->size - PAGE_SIZE) & RING_NR_PAGES)
-			| RING_REPORT_64K | RING_VALID);
+			| RING_VALID);
 
 	/* If the head is still not zero, the ring is dead */
 	if ((I915_READ_CTL(ring) & RING_VALID) == 0 ||
@@ -399,8 +385,6 @@ static int init_render_ring(struct intel_ring_buffer *ring)
 
 	if (INTEL_INFO(dev)->gen > 3) {
 		int mode = VS_TIMER_DISPATCH << 16 | VS_TIMER_DISPATCH;
-		if (IS_GEN6(dev) || IS_GEN7(dev))
-			mode |= MI_FLUSH_ENABLE << 16 | MI_FLUSH_ENABLE;
 		I915_WRITE(MI_MODE, mode);
 		if (IS_GEN7(dev))
 			I915_WRITE(GFX_MODE_GEN7,
@@ -467,7 +451,7 @@ gen6_add_request(struct intel_ring_buffer *ring,
 	mbox1_reg = ring->signal_mbox[0];
 	mbox2_reg = ring->signal_mbox[1];
 
-	*seqno = i915_gem_get_seqno(ring->dev);
+	*seqno = i915_gem_next_request_seqno(ring);
 
 	update_mboxes(ring, *seqno, mbox1_reg);
 	update_mboxes(ring, *seqno, mbox2_reg);
@@ -565,8 +549,7 @@ static int
 pc_render_add_request(struct intel_ring_buffer *ring,
 		      u32 *result)
 {
-	struct drm_device *dev = ring->dev;
-	u32 seqno = i915_gem_get_seqno(dev);
+	u32 seqno = i915_gem_next_request_seqno(ring);
 	struct pipe_control *pc = ring->private;
 	u32 scratch_addr = pc->gtt_offset + 128;
 	int ret;
@@ -600,6 +583,7 @@ pc_render_add_request(struct intel_ring_buffer *ring,
 	PIPE_CONTROL_FLUSH(ring, scratch_addr);
 	scratch_addr += 128;
 	PIPE_CONTROL_FLUSH(ring, scratch_addr);
+
 	intel_ring_emit(ring, GFX_OP_PIPE_CONTROL(4) | PIPE_CONTROL_QW_WRITE |
 			PIPE_CONTROL_WRITE_FLUSH |
 			PIPE_CONTROL_TEXTURE_CACHE_INVALIDATE |
@@ -617,8 +601,7 @@ static int
 render_ring_add_request(struct intel_ring_buffer *ring,
 			u32 *result)
 {
-	struct drm_device *dev = ring->dev;
-	u32 seqno = i915_gem_get_seqno(dev);
+	u32 seqno = i915_gem_next_request_seqno(ring);
 	int ret;
 
 	ret = intel_ring_begin(ring, 4);
@@ -643,7 +626,7 @@ gen6_ring_get_seqno(struct intel_ring_buffer *ring)
 	/* Workaround to force correct ordering between irq and seqno writes on
 	 * ivb (and maybe also on snb) by reading from a CS register (like
 	 * ACTHD) before reading the status page. */
-	if (IS_GEN7(dev))
+	if (IS_GEN6(dev) || IS_GEN7(dev))
 		intel_ring_get_active_head(ring);
 	return intel_read_status_page(ring, I915_GEM_HWS_INDEX);
 }
@@ -744,13 +727,13 @@ void intel_ring_setup_status_page(struct intel_ring_buffer *ring)
 	 */
 	if (IS_GEN7(dev)) {
 		switch (ring->id) {
-		case RING_RENDER:
+		case RCS:
 			mmio = RENDER_HWS_PGA_GEN7;
 			break;
-		case RING_BLT:
+		case BCS:
 			mmio = BLT_HWS_PGA_GEN7;
 			break;
-		case RING_BSD:
+		case VCS:
 			mmio = BSD_HWS_PGA_GEN7;
 			break;
 		}
@@ -792,7 +775,7 @@ ring_add_request(struct intel_ring_buffer *ring,
 	if (ret)
 		return ret;
 
-	seqno = i915_gem_get_seqno(ring->dev);
+	seqno = i915_gem_next_request_seqno(ring);
 
 	intel_ring_emit(ring, MI_STORE_DWORD_INDEX);
 	intel_ring_emit(ring, I915_GEM_HWS_INDEX << MI_STORE_DWORD_INDEX_SHIFT);
@@ -816,8 +799,7 @@ gen6_ring_get_irq(struct intel_ring_buffer *ring, u32 gflag, u32 rflag)
 	/* It looks like we need to prevent the gt from suspending while waiting
 	 * for an notifiy irq, otherwise irqs seem to get lost on at least the
 	 * blt/bsd rings on ivb. */
-	if (IS_GEN7(dev))
-		gen6_gt_force_wake_get(dev_priv);
+	gen6_gt_force_wake_get(dev_priv);
 
 	spin_lock(&ring->irq_lock);
 	if (ring->irq_refcount++ == 0) {
@@ -844,8 +826,7 @@ gen6_ring_put_irq(struct intel_ring_buffer *ring, u32 gflag, u32 rflag)
 	}
 	spin_unlock(&ring->irq_lock);
 
-	if (IS_GEN7(dev))
-		gen6_gt_force_wake_put(dev_priv);
+	gen6_gt_force_wake_put(dev_priv);
 }
 
 static bool
@@ -1127,24 +1108,94 @@ static int intel_wrap_ring_buffer(struct intel_ring_buffer *ring)
 	return 0;
 }
 
-int intel_wait_ring_buffer(struct intel_ring_buffer *ring, int n)
+static int intel_ring_wait_seqno(struct intel_ring_buffer *ring, u32 seqno)
 {
-	struct drm_device *dev = ring->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	unsigned long end;
-	u32 head;
+	struct drm_i915_private *dev_priv = ring->dev->dev_private;
+	bool was_interruptible;
+	int ret;
 
-	/* If the reported head position has wrapped or hasn't advanced,
-	 * fallback to the slow and accurate path.
+	/* XXX As we have not yet audited all the paths to check that
+	 * they are ready for ERESTARTSYS from intel_ring_begin, do not
+	 * allow us to be interruptible by a signal.
 	 */
-	head = intel_read_status_page(ring, 4);
-	if (head > ring->head) {
-		ring->head = head;
+	was_interruptible = dev_priv->mm.interruptible;
+	dev_priv->mm.interruptible = false;
+
+	ret = i915_wait_request(ring, seqno, true);
+
+	dev_priv->mm.interruptible = was_interruptible;
+
+	return ret;
+}
+
+static int intel_ring_wait_request(struct intel_ring_buffer *ring, int n)
+{
+	struct drm_i915_gem_request *request;
+	u32 seqno = 0;
+	int ret;
+
+	i915_gem_retire_requests_ring(ring);
+
+	if (ring->last_retired_head != -1) {
+		ring->head = ring->last_retired_head;
+		ring->last_retired_head = -1;
 		ring->space = ring_space(ring);
 		if (ring->space >= n)
 			return 0;
 	}
 
+	list_for_each_entry(request, &ring->request_list, list) {
+		int space;
+
+		if (request->tail == -1)
+			continue;
+
+		space = request->tail - (ring->tail + 8);
+		if (space < 0)
+			space += ring->size;
+		if (space >= n) {
+			seqno = request->seqno;
+			break;
+		}
+
+		/* Consume this request in case we need more space than
+		 * is available and so need to prevent a race between
+		 * updating last_retired_head and direct reads of
+		 * I915_RING_HEAD. It also provides a nice sanity check.
+		 */
+		request->tail = -1;
+	}
+
+	if (seqno == 0)
+		return -ENOSPC;
+
+	ret = intel_ring_wait_seqno(ring, seqno);
+	if (ret)
+		return ret;
+
+	if (WARN_ON(ring->last_retired_head == -1))
+		return -ENOSPC;
+
+	ring->head = ring->last_retired_head;
+	ring->last_retired_head = -1;
+	ring->space = ring_space(ring);
+	if (WARN_ON(ring->space < n))
+		return -ENOSPC;
+
+	return 0;
+}
+
+int intel_wait_ring_buffer(struct intel_ring_buffer *ring, int n)
+{
+	struct drm_device *dev = ring->dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	unsigned long end;
+	int ret;
+
+	ret = intel_ring_wait_request(ring, n);
+	if (ret != -ENOSPC)
+		return ret;
+
 	trace_i915_ring_wait_begin(ring);
 	if (drm_core_check_feature(dev, DRIVER_GEM))
 		/* With GEM the hangcheck timer should kick us out of the loop,
@@ -1212,7 +1263,7 @@ void intel_ring_advance(struct intel_ring_buffer *ring)
 
 static const struct intel_ring_buffer render_ring = {
 	.name			= "render ring",
-	.id			= RING_RENDER,
+	.id			= RCS,
 	.mmio_base		= RENDER_RING_BASE,
 	.size			= 32 * PAGE_SIZE,
 	.init			= init_render_ring,
@@ -1235,7 +1286,7 @@ static const struct intel_ring_buffer render_ring = {
 
 static const struct intel_ring_buffer bsd_ring = {
 	.name                   = "bsd ring",
-	.id			= RING_BSD,
+	.id			= VCS,
 	.mmio_base		= BSD_RING_BASE,
 	.size			= 32 * PAGE_SIZE,
 	.init			= init_ring_common,
@@ -1345,7 +1396,7 @@ gen6_bsd_ring_put_irq(struct intel_ring_buffer *ring)
 /* ring buffer for Video Codec for Gen6+ */
 static const struct intel_ring_buffer gen6_bsd_ring = {
 	.name			= "gen6 bsd ring",
-	.id			= RING_BSD,
+	.id			= VCS,
 	.mmio_base		= GEN6_BSD_RING_BASE,
 	.size			= 32 * PAGE_SIZE,
 	.init			= init_ring_common,
@@ -1381,79 +1432,13 @@ blt_ring_put_irq(struct intel_ring_buffer *ring)
 			  GEN6_BLITTER_USER_INTERRUPT);
 }
 
-
-/* Workaround for some stepping of SNB,
- * each time when BLT engine ring tail moved,
- * the first command in the ring to be parsed
- * should be MI_BATCH_BUFFER_START
- */
-#define NEED_BLT_WORKAROUND(dev) \
-	(IS_GEN6(dev) && (dev->pdev->revision < 8))
-
-static inline struct drm_i915_gem_object *
-to_blt_workaround(struct intel_ring_buffer *ring)
-{
-	return ring->private;
-}
-
-static int blt_ring_init(struct intel_ring_buffer *ring)
-{
-	if (NEED_BLT_WORKAROUND(ring->dev)) {
-		struct drm_i915_gem_object *obj;
-		u32 *ptr;
-		int ret;
-
-		obj = i915_gem_alloc_object(ring->dev, 4096);
-		if (obj == NULL)
-			return -ENOMEM;
-
-		ret = i915_gem_object_pin(obj, 4096, true);
-		if (ret) {
-			drm_gem_object_unreference(&obj->base);
-			return ret;
-		}
-
-		ptr = kmap(obj->pages[0]);
-		*ptr++ = MI_BATCH_BUFFER_END;
-		*ptr++ = MI_NOOP;
-		kunmap(obj->pages[0]);
-
-		ret = i915_gem_object_set_to_gtt_domain(obj, false);
-		if (ret) {
-			i915_gem_object_unpin(obj);
-			drm_gem_object_unreference(&obj->base);
-			return ret;
-		}
-
-		ring->private = obj;
-	}
-
-	return init_ring_common(ring);
-}
-
-static int blt_ring_begin(struct intel_ring_buffer *ring,
-			  int num_dwords)
-{
-	if (ring->private) {
-		int ret = intel_ring_begin(ring, num_dwords+2);
-		if (ret)
-			return ret;
-
-		intel_ring_emit(ring, MI_BATCH_BUFFER_START);
-		intel_ring_emit(ring, to_blt_workaround(ring)->gtt_offset);
-
-		return 0;
-	} else
-		return intel_ring_begin(ring, 4);
-}
-
 static int blt_ring_flush(struct intel_ring_buffer *ring,
 			  u32 invalidate, u32 flush)
 {
 	uint32_t cmd;
 	int ret;
 
-	ret = blt_ring_begin(ring, 4);
+	ret = intel_ring_begin(ring, 4);
 	if (ret)
 		return ret;
 
@@ -1468,22 +1453,12 @@ static int blt_ring_flush(struct intel_ring_buffer *ring,
 	return 0;
 }
 
-static void blt_ring_cleanup(struct intel_ring_buffer *ring)
-{
-	if (!ring->private)
-		return;
-
-	i915_gem_object_unpin(ring->private);
-	drm_gem_object_unreference(ring->private);
-	ring->private = NULL;
-}
-
 static const struct intel_ring_buffer gen6_blt_ring = {
 	.name			= "blt ring",
-	.id			= RING_BLT,
+	.id			= BCS,
 	.mmio_base		= BLT_RING_BASE,
 	.size			= 32 * PAGE_SIZE,
-	.init			= blt_ring_init,
+	.init			= init_ring_common,
 	.write_tail		= ring_write_tail,
 	.flush			= blt_ring_flush,
 	.add_request		= gen6_add_request,
@@ -1491,7 +1466,6 @@ static const struct intel_ring_buffer gen6_blt_ring = {
 	.irq_get		= blt_ring_get_irq,
 	.irq_put		= blt_ring_put_irq,
 	.dispatch_execbuffer	= gen6_ring_dispatch_execbuffer,
-	.cleanup		= blt_ring_cleanup,
 	.sync_to		= gen6_blt_ring_sync_to,
 	.semaphore_register	= {MI_SEMAPHORE_SYNC_BR,
 				   MI_SEMAPHORE_SYNC_BV,
diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.h b/drivers/gpu/drm/i915/intel_ringbuffer.h
index 68281c96c558..bc0365b8fa4d 100644
--- a/drivers/gpu/drm/i915/intel_ringbuffer.h
+++ b/drivers/gpu/drm/i915/intel_ringbuffer.h
@@ -1,13 +1,6 @@
 #ifndef _INTEL_RINGBUFFER_H_
 #define _INTEL_RINGBUFFER_H_
 
-enum {
-	RCS = 0x0,
-	VCS,
-	BCS,
-	I915_NUM_RINGS,
-};
-
 struct  intel_hw_status_page {
 	u32	__iomem	*page_addr;
 	unsigned int	gfx_addr;
@@ -36,10 +29,11 @@ struct  intel_hw_status_page {
 struct  intel_ring_buffer {
 	const char	*name;
 	enum intel_ring_id {
-		RING_RENDER = 0x1,
-		RING_BSD = 0x2,
-		RING_BLT = 0x4,
+		RCS = 0x0,
+		VCS,
+		BCS,
 	} id;
+#define I915_NUM_RINGS 3
 	u32		mmio_base;
 	void		__iomem *virtual_start;
 	struct		drm_device *dev;
@@ -52,6 +46,16 @@ struct  intel_ring_buffer {
 	int		effective_size;
 	struct intel_hw_status_page status_page;
 
+	/** We track the position of the requests in the ring buffer, and
+	 * when each is retired we increment last_retired_head as the GPU
+	 * must have finished processing the request and so we know we
+	 * can advance the ringbuffer up to that position.
+	 *
+	 * last_retired_head is set to -1 after the value is consumed so
+	 * we can detect new retirements.
+	 */
+	u32		last_retired_head;
+
 	spinlock_t	irq_lock;
 	u32		irq_refcount;
 	u32		irq_mask;
@@ -119,6 +123,12 @@ struct  intel_ring_buffer {
 	void *private;
 };
 
+static inline unsigned
+intel_ring_flag(struct intel_ring_buffer *ring)
+{
+	return 1 << ring->id;
+}
+
 static inline u32
 intel_ring_sync_index(struct intel_ring_buffer *ring,
 		      struct intel_ring_buffer *other)
@@ -193,6 +203,11 @@ int intel_init_blt_ring_buffer(struct drm_device *dev);
 u32 intel_ring_get_active_head(struct intel_ring_buffer *ring);
 void intel_ring_setup_status_page(struct intel_ring_buffer *ring);
 
+static inline u32 intel_ring_get_tail(struct intel_ring_buffer *ring)
+{
+	return ring->tail;
+}
+
 static inline void i915_trace_irq_get(struct intel_ring_buffer *ring, u32 seqno)
 {
 	if (ring->trace_irq_seqno == 0 && ring->irq_get(ring))
diff --git a/drivers/gpu/drm/i915/intel_sdvo.c b/drivers/gpu/drm/i915/intel_sdvo.c
index e334ec33a47d..e36b171c1e7d 100644
--- a/drivers/gpu/drm/i915/intel_sdvo.c
+++ b/drivers/gpu/drm/i915/intel_sdvo.c
@@ -148,7 +148,7 @@ struct intel_sdvo_connector {
 	/* Mark the type of connector */
 	uint16_t output_flag;
 
-	int force_audio;
+	enum hdmi_force_audio force_audio;
 
 	/* This contains all current supported TV format */
 	u8 tv_format_supported[TV_FORMAT_NUM];
@@ -944,7 +944,6 @@ intel_sdvo_set_input_timings_for_mode(struct intel_sdvo *intel_sdvo,
 
 	intel_sdvo_get_mode_from_dtd(adjusted_mode, &intel_sdvo->input_dtd);
 
-	drm_mode_set_crtcinfo(adjusted_mode, 0);
 	return true;
 }
 
@@ -1310,8 +1309,8 @@ intel_sdvo_tmds_sink_detect(struct drm_connector *connector)
 
 	if (status == connector_status_connected) {
 		struct intel_sdvo_connector *intel_sdvo_connector = to_intel_sdvo_connector(connector);
-		if (intel_sdvo_connector->force_audio)
-			intel_sdvo->has_hdmi_audio = intel_sdvo_connector->force_audio > 0;
+		if (intel_sdvo_connector->force_audio != HDMI_AUDIO_AUTO)
+			intel_sdvo->has_hdmi_audio = (intel_sdvo_connector->force_audio == HDMI_AUDIO_ON);
 	}
 
 	return status;
@@ -1684,10 +1683,10 @@ intel_sdvo_set_property(struct drm_connector *connector,
 
 		intel_sdvo_connector->force_audio = i;
 
-		if (i == 0)
+		if (i == HDMI_AUDIO_AUTO)
 			has_audio = intel_sdvo_detect_hdmi_audio(connector);
 		else
-			has_audio = i > 0;
+			has_audio = (i == HDMI_AUDIO_ON);
 
 		if (has_audio == intel_sdvo->has_hdmi_audio)
 			return 0;
@@ -1985,7 +1984,7 @@ intel_sdvo_connector_init(struct intel_sdvo_connector *connector,
 	drm_connector_helper_add(&connector->base.base,
 				 &intel_sdvo_connector_helper_funcs);
 
-	connector->base.base.interlace_allowed = 0;
+	connector->base.base.interlace_allowed = 1;
 	connector->base.base.doublescan_allowed = 0;
 	connector->base.base.display_info.subpixel_order = SubPixelHorizontalRGB;
 
@@ -2277,10 +2276,8 @@ static bool intel_sdvo_tv_create_property(struct intel_sdvo *intel_sdvo,
 		intel_sdvo_connector->max_##name = data_value[0]; \
 		intel_sdvo_connector->cur_##name = response; \
 		intel_sdvo_connector->name = \
-			drm_property_create(dev, DRM_MODE_PROP_RANGE, #name, 2); \
+			drm_property_create_range(dev, 0, #name, 0, data_value[0]); \
 		if (!intel_sdvo_connector->name) return false; \
-		intel_sdvo_connector->name->values[0] = 0; \
-		intel_sdvo_connector->name->values[1] = data_value[0]; \
 		drm_connector_attach_property(connector, \
 					      intel_sdvo_connector->name, \
 					      intel_sdvo_connector->cur_##name); \
@@ -2314,25 +2311,19 @@ intel_sdvo_create_enhance_property_tv(struct intel_sdvo *intel_sdvo,
 		intel_sdvo_connector->left_margin = data_value[0] - response;
 		intel_sdvo_connector->right_margin = intel_sdvo_connector->left_margin;
 		intel_sdvo_connector->left =
-			drm_property_create(dev, DRM_MODE_PROP_RANGE,
-					    "left_margin", 2);
+			drm_property_create_range(dev, 0, "left_margin", 0, data_value[0]);
 		if (!intel_sdvo_connector->left)
 			return false;
 
-		intel_sdvo_connector->left->values[0] = 0;
-		intel_sdvo_connector->left->values[1] = data_value[0];
 		drm_connector_attach_property(connector,
 					      intel_sdvo_connector->left,
 					      intel_sdvo_connector->left_margin);
 
 		intel_sdvo_connector->right =
-			drm_property_create(dev, DRM_MODE_PROP_RANGE,
-					    "right_margin", 2);
+			drm_property_create_range(dev, 0, "right_margin", 0, data_value[0]);
 		if (!intel_sdvo_connector->right)
 			return false;
 
-		intel_sdvo_connector->right->values[0] = 0;
-		intel_sdvo_connector->right->values[1] = data_value[0];
 		drm_connector_attach_property(connector,
 					      intel_sdvo_connector->right,
 					      intel_sdvo_connector->right_margin);
@@ -2356,25 +2347,21 @@ intel_sdvo_create_enhance_property_tv(struct intel_sdvo *intel_sdvo,
 		intel_sdvo_connector->top_margin = data_value[0] - response;
 		intel_sdvo_connector->bottom_margin = intel_sdvo_connector->top_margin;
 		intel_sdvo_connector->top =
-			drm_property_create(dev, DRM_MODE_PROP_RANGE,
-					    "top_margin", 2);
+			drm_property_create_range(dev, 0,
+					    "top_margin", 0, data_value[0]);
 		if (!intel_sdvo_connector->top)
 			return false;
 
-		intel_sdvo_connector->top->values[0] = 0;
-		intel_sdvo_connector->top->values[1] = data_value[0];
 		drm_connector_attach_property(connector,
 					      intel_sdvo_connector->top,
 					      intel_sdvo_connector->top_margin);
 
 		intel_sdvo_connector->bottom =
-			drm_property_create(dev, DRM_MODE_PROP_RANGE,
-					    "bottom_margin", 2);
+			drm_property_create_range(dev, 0,
+					    "bottom_margin", 0, data_value[0]);
 		if (!intel_sdvo_connector->bottom)
 			return false;
 
-		intel_sdvo_connector->bottom->values[0] = 0;
-		intel_sdvo_connector->bottom->values[1] = data_value[0];
 		drm_connector_attach_property(connector,
 					      intel_sdvo_connector->bottom,
 					      intel_sdvo_connector->bottom_margin);
@@ -2403,12 +2390,10 @@ intel_sdvo_create_enhance_property_tv(struct intel_sdvo *intel_sdvo,
 		intel_sdvo_connector->max_dot_crawl = 1;
 		intel_sdvo_connector->cur_dot_crawl = response & 0x1;
 		intel_sdvo_connector->dot_crawl =
-			drm_property_create(dev, DRM_MODE_PROP_RANGE, "dot_crawl", 2);
+			drm_property_create_range(dev, 0, "dot_crawl", 0, 1);
 		if (!intel_sdvo_connector->dot_crawl)
 			return false;
 
-		intel_sdvo_connector->dot_crawl->values[0] = 0;
-		intel_sdvo_connector->dot_crawl->values[1] = 1;
 		drm_connector_attach_property(connector,
 					      intel_sdvo_connector->dot_crawl,
 					      intel_sdvo_connector->cur_dot_crawl);
diff --git a/drivers/gpu/drm/i915/intel_sprite.c b/drivers/gpu/drm/i915/intel_sprite.c
index 2288abf88cce..a464771a7240 100644
--- a/drivers/gpu/drm/i915/intel_sprite.c
+++ b/drivers/gpu/drm/i915/intel_sprite.c
@@ -225,16 +225,16 @@ snb_update_plane(struct drm_plane *plane, struct drm_framebuffer *fb,
 
 	/* Mask out pixel format bits in case we change it */
 	dvscntr &= ~DVS_PIXFORMAT_MASK;
-	dvscntr &= ~DVS_RGB_ORDER_RGBX;
+	dvscntr &= ~DVS_RGB_ORDER_XBGR;
 	dvscntr &= ~DVS_YUV_BYTE_ORDER_MASK;
 
 	switch (fb->pixel_format) {
 	case DRM_FORMAT_XBGR8888:
-		dvscntr |= DVS_FORMAT_RGBX888;
+		dvscntr |= DVS_FORMAT_RGBX888 | DVS_RGB_ORDER_XBGR;
 		pixel_size = 4;
 		break;
 	case DRM_FORMAT_XRGB8888:
-		dvscntr |= DVS_FORMAT_RGBX888 | DVS_RGB_ORDER_RGBX;
+		dvscntr |= DVS_FORMAT_RGBX888;
 		pixel_size = 4;
 		break;
 	case DRM_FORMAT_YUYV:
@@ -411,6 +411,9 @@ intel_update_plane(struct drm_plane *plane, struct drm_crtc *crtc,
 
 	old_obj = intel_plane->obj;
 
+	src_w = src_w >> 16;
+	src_h = src_h >> 16;
+
 	/* Pipe must be running... */
 	if (!(I915_READ(PIPECONF(pipe)) & PIPECONF_ENABLE))
 		return -EINVAL;
@@ -501,7 +504,7 @@ intel_update_plane(struct drm_plane *plane, struct drm_crtc *crtc,
 			intel_wait_for_vblank(dev, to_intel_crtc(crtc)->pipe);
 			mutex_lock(&dev->struct_mutex);
 		}
-		i915_gem_object_unpin(old_obj);
+		intel_unpin_fb_obj(old_obj);
 	}
 
 out_unlock:
@@ -528,7 +531,7 @@ intel_disable_plane(struct drm_plane *plane)
 		goto out;
 
 	mutex_lock(&dev->struct_mutex);
-	i915_gem_object_unpin(intel_plane->obj);
+	intel_unpin_fb_obj(intel_plane->obj);
 	intel_plane->obj = NULL;
 	mutex_unlock(&dev->struct_mutex);
 out:
diff --git a/drivers/gpu/drm/i915/intel_tv.c b/drivers/gpu/drm/i915/intel_tv.c
index 1571be37ce3e..05f765ef5464 100644
--- a/drivers/gpu/drm/i915/intel_tv.c
+++ b/drivers/gpu/drm/i915/intel_tv.c
@@ -1240,7 +1240,7 @@ intel_tv_detect(struct drm_connector *connector, bool force)
 	int type;
 
 	mode = reported_modes[0];
-	drm_mode_set_crtcinfo(&mode, CRTC_INTERLACE_HALVE_V);
+	drm_mode_set_crtcinfo(&mode, 0);
 
 	if (intel_tv->base.base.crtc && intel_tv->base.base.crtc->enabled) {
 		type = intel_tv_detect_type(intel_tv, connector);
diff --git a/drivers/gpu/drm/mga/mga_dma.c b/drivers/gpu/drm/mga/mga_dma.c
index 5ccb65deb83c..507aa3df0168 100644
--- a/drivers/gpu/drm/mga/mga_dma.c
+++ b/drivers/gpu/drm/mga/mga_dma.c
@@ -403,6 +403,8 @@ int mga_driver_load(struct drm_device *dev, unsigned long flags)
 	dev_priv->usec_timeout = MGA_DEFAULT_USEC_TIMEOUT;
 	dev_priv->chipset = flags;
 
+	pci_set_master(dev->pdev);
+
 	dev_priv->mmio_base = pci_resource_start(dev->pdev, 1);
 	dev_priv->mmio_size = pci_resource_len(dev->pdev, 1);
 
diff --git a/drivers/gpu/drm/nouveau/Kconfig b/drivers/gpu/drm/nouveau/Kconfig
index ca1639918f57..97a81260485a 100644
--- a/drivers/gpu/drm/nouveau/Kconfig
+++ b/drivers/gpu/drm/nouveau/Kconfig
@@ -13,6 +13,7 @@ config DRM_NOUVEAU
 	select ACPI_VIDEO if ACPI && X86 && BACKLIGHT_CLASS_DEVICE && VIDEO_OUTPUT_CONTROL && INPUT
 	select ACPI_WMI if ACPI
 	select MXM_WMI if ACPI
+	select POWER_SUPPLY
 	help
 	  Choose this option for open-source nVidia support.
 
diff --git a/drivers/gpu/drm/nouveau/Makefile b/drivers/gpu/drm/nouveau/Makefile
index 9f27e3d9e69a..1a2ad7eb1734 100644
--- a/drivers/gpu/drm/nouveau/Makefile
+++ b/drivers/gpu/drm/nouveau/Makefile
@@ -14,7 +14,8 @@ nouveau-y := nouveau_drv.o nouveau_state.o nouveau_channel.o nouveau_mem.o \
 	     nouveau_mm.o nouveau_vm.o nouveau_mxm.o nouveau_gpio.o \
              nv04_timer.o \
              nv04_mc.o nv40_mc.o nv50_mc.o \
-             nv04_fb.o nv10_fb.o nv30_fb.o nv40_fb.o nv50_fb.o nvc0_fb.o \
+             nv04_fb.o nv10_fb.o nv20_fb.o nv30_fb.o nv40_fb.o \
+             nv50_fb.o nvc0_fb.o \
              nv04_fifo.o nv10_fifo.o nv40_fifo.o nv50_fifo.o nvc0_fifo.o \
              nv04_graph.o nv10_graph.o nv20_graph.o \
              nv40_graph.o nv50_graph.o nvc0_graph.o \
diff --git a/drivers/gpu/drm/nouveau/nouveau_bios.c b/drivers/gpu/drm/nouveau/nouveau_bios.c
index e5cbead85e50..80963d05b54a 100644
--- a/drivers/gpu/drm/nouveau/nouveau_bios.c
+++ b/drivers/gpu/drm/nouveau/nouveau_bios.c
@@ -65,195 +65,233 @@ static bool nv_cksum(const uint8_t *data, unsigned int length)
 }
 
 static int
-score_vbios(struct drm_device *dev, const uint8_t *data, const bool writeable)
+score_vbios(struct nvbios *bios, const bool writeable)
 {
-	if (!(data[0] == 0x55 && data[1] == 0xAA)) {
-		NV_TRACEWARN(dev, "... BIOS signature not found\n");
+	if (!bios->data || bios->data[0] != 0x55 || bios->data[1] != 0xAA) {
+		NV_TRACEWARN(bios->dev, "... BIOS signature not found\n");
 		return 0;
 	}
 
-	if (nv_cksum(data, data[2] * 512)) {
-		NV_TRACEWARN(dev, "... BIOS checksum invalid\n");
+	if (nv_cksum(bios->data, bios->data[2] * 512)) {
+		NV_TRACEWARN(bios->dev, "... BIOS checksum invalid\n");
 		/* if a ro image is somewhat bad, it's probably all rubbish */
 		return writeable ? 2 : 1;
-	} else
-		NV_TRACE(dev, "... appears to be valid\n");
+	}
 
+	NV_TRACE(bios->dev, "... appears to be valid\n");
 	return 3;
 }
 
-static void load_vbios_prom(struct drm_device *dev, uint8_t *data)
+static void
+bios_shadow_prom(struct nvbios *bios)
 {
+	struct drm_device *dev = bios->dev;
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	uint32_t pci_nv_20, save_pci_nv_20;
-	int pcir_ptr;
+	u32 pcireg, access;
+	u16 pcir;
 	int i;
 
+	/* enable access to rom */
 	if (dev_priv->card_type >= NV_50)
-		pci_nv_20 = 0x88050;
+		pcireg = 0x088050;
 	else
-		pci_nv_20 = NV_PBUS_PCI_NV_20;
+		pcireg = NV_PBUS_PCI_NV_20;
+	access = nv_mask(dev, pcireg, 0x00000001, 0x00000000);
 
-	/* enable ROM access */
-	save_pci_nv_20 = nvReadMC(dev, pci_nv_20);
-	nvWriteMC(dev, pci_nv_20,
-		  save_pci_nv_20 & ~NV_PBUS_PCI_NV_20_ROM_SHADOW_ENABLED);
+	/* bail if no rom signature, with a workaround for a PROM reading
+	 * issue on some chipsets.  the first read after a period of
+	 * inactivity returns the wrong result, so retry the first header
+	 * byte a few times before giving up as a workaround
+	 */
+	i = 16;
+	do {
+		if (nv_rd08(dev, NV_PROM_OFFSET + 0) == 0x55)
+			break;
+	} while (i--);
 
-	/* bail if no rom signature */
-	if (nv_rd08(dev, NV_PROM_OFFSET) != 0x55 ||
-	    nv_rd08(dev, NV_PROM_OFFSET + 1) != 0xaa)
+	if (!i || nv_rd08(dev, NV_PROM_OFFSET + 1) != 0xaa)
 		goto out;
 
 	/* additional check (see note below) - read PCI record header */
-	pcir_ptr = nv_rd08(dev, NV_PROM_OFFSET + 0x18) |
-		   nv_rd08(dev, NV_PROM_OFFSET + 0x19) << 8;
-	if (nv_rd08(dev, NV_PROM_OFFSET + pcir_ptr) != 'P' ||
-	    nv_rd08(dev, NV_PROM_OFFSET + pcir_ptr + 1) != 'C' ||
-	    nv_rd08(dev, NV_PROM_OFFSET + pcir_ptr + 2) != 'I' ||
-	    nv_rd08(dev, NV_PROM_OFFSET + pcir_ptr + 3) != 'R')
+	pcir = nv_rd08(dev, NV_PROM_OFFSET + 0x18) |
+	       nv_rd08(dev, NV_PROM_OFFSET + 0x19) << 8;
+	if (nv_rd08(dev, NV_PROM_OFFSET + pcir + 0) != 'P' ||
+	    nv_rd08(dev, NV_PROM_OFFSET + pcir + 1) != 'C' ||
+	    nv_rd08(dev, NV_PROM_OFFSET + pcir + 2) != 'I' ||
+	    nv_rd08(dev, NV_PROM_OFFSET + pcir + 3) != 'R')
 		goto out;
 
-	/* on some 6600GT/6800LE prom reads are messed up.  nvclock alleges a
-	 * a good read may be obtained by waiting or re-reading (cargocult: 5x)
-	 * each byte.  we'll hope pramin has something usable instead
-	 */
-	for (i = 0; i < NV_PROM_SIZE; i++)
-		data[i] = nv_rd08(dev, NV_PROM_OFFSET + i);
+	/* read entire bios image to system memory */
+	bios->length = nv_rd08(dev, NV_PROM_OFFSET + 2) * 512;
+	bios->data = kmalloc(bios->length, GFP_KERNEL);
+	if (bios->data) {
+		for (i = 0; i < bios->length; i++)
+			bios->data[i] = nv_rd08(dev, NV_PROM_OFFSET + i);
+	}
 
 out:
-	/* disable ROM access */
-	nvWriteMC(dev, pci_nv_20,
-		  save_pci_nv_20 | NV_PBUS_PCI_NV_20_ROM_SHADOW_ENABLED);
+	/* disable access to rom */
+	nv_wr32(dev, pcireg, access);
 }
 
-static void load_vbios_pramin(struct drm_device *dev, uint8_t *data)
+static void
+bios_shadow_pramin(struct nvbios *bios)
 {
+	struct drm_device *dev = bios->dev;
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	uint32_t old_bar0_pramin = 0;
+	u32 bar0 = 0;
 	int i;
 
 	if (dev_priv->card_type >= NV_50) {
 		u64 addr = (u64)(nv_rd32(dev, 0x619f04) & 0xffffff00) << 8;
 		if (!addr) {
-			addr  = (u64)nv_rd32(dev, 0x1700) << 16;
+			addr  = (u64)nv_rd32(dev, 0x001700) << 16;
 			addr += 0xf0000;
 		}
 
-		old_bar0_pramin = nv_rd32(dev, 0x1700);
-		nv_wr32(dev, 0x1700, addr >> 16);
+		bar0 = nv_mask(dev, 0x001700, 0xffffffff, addr >> 16);
 	}
 
 	/* bail if no rom signature */
-	if (nv_rd08(dev, NV_PRAMIN_OFFSET) != 0x55 ||
+	if (nv_rd08(dev, NV_PRAMIN_OFFSET + 0) != 0x55 ||
 	    nv_rd08(dev, NV_PRAMIN_OFFSET + 1) != 0xaa)
 		goto out;
 
-	for (i = 0; i < NV_PROM_SIZE; i++)
-		data[i] = nv_rd08(dev, NV_PRAMIN_OFFSET + i);
+	bios->length = nv_rd08(dev, NV_PRAMIN_OFFSET + 2) * 512;
+	bios->data = kmalloc(bios->length, GFP_KERNEL);
+	if (bios->data) {
+		for (i = 0; i < bios->length; i++)
+			bios->data[i] = nv_rd08(dev, NV_PRAMIN_OFFSET + i);
+	}
 
 out:
 	if (dev_priv->card_type >= NV_50)
-		nv_wr32(dev, 0x1700, old_bar0_pramin);
+		nv_wr32(dev, 0x001700, bar0);
 }
 
-static void load_vbios_pci(struct drm_device *dev, uint8_t *data)
+static void
+bios_shadow_pci(struct nvbios *bios)
+{
+	struct pci_dev *pdev = bios->dev->pdev;
+	size_t length;
+
+	if (!pci_enable_rom(pdev)) {
+		void __iomem *rom = pci_map_rom(pdev, &length);
+		if (rom && length) {
+			bios->data = kmalloc(length, GFP_KERNEL);
+			if (bios->data) {
+				memcpy_fromio(bios->data, rom, length);
+				bios->length = length;
+			}
+		}
+		if (rom)
+			pci_unmap_rom(pdev, rom);
+
+		pci_disable_rom(pdev);
+	}
+}
+
+static void
+bios_shadow_acpi(struct nvbios *bios)
 {
-	void __iomem *rom = NULL;
-	size_t rom_len;
-	int ret;
+	struct pci_dev *pdev = bios->dev->pdev;
+	int ptr, len, ret;
+	u8 data[3];
 
-	ret = pci_enable_rom(dev->pdev);
-	if (ret)
+	if (!nouveau_acpi_rom_supported(pdev))
 		return;
 
-	rom = pci_map_rom(dev->pdev, &rom_len);
-	if (!rom)
-		goto out;
-	memcpy_fromio(data, rom, rom_len);
-	pci_unmap_rom(dev->pdev, rom);
+	ret = nouveau_acpi_get_bios_chunk(data, 0, sizeof(data));
+	if (ret != sizeof(data))
+		return;
 
-out:
-	pci_disable_rom(dev->pdev);
-}
+	bios->length = min(data[2] * 512, 65536);
+	bios->data = kmalloc(bios->length, GFP_KERNEL);
+	if (!bios->data)
+		return;
 
-static void load_vbios_acpi(struct drm_device *dev, uint8_t *data)
-{
-	int i;
-	int ret;
-	int size = 64 * 1024;
+	len = bios->length;
+	ptr = 0;
+	while (len) {
+		int size = (len > ROM_BIOS_PAGE) ? ROM_BIOS_PAGE : len;
 
-	if (!nouveau_acpi_rom_supported(dev->pdev))
-		return;
+		ret = nouveau_acpi_get_bios_chunk(bios->data, ptr, size);
+		if (ret != size) {
+			kfree(bios->data);
+			bios->data = NULL;
+			return;
+		}
 
-	for (i = 0; i < (size / ROM_BIOS_PAGE); i++) {
-		ret = nouveau_acpi_get_bios_chunk(data,
-						  (i * ROM_BIOS_PAGE),
-						  ROM_BIOS_PAGE);
-		if (ret <= 0)
-			break;
+		len -= size;
+		ptr += size;
 	}
-	return;
 }
 
 struct methods {
 	const char desc[8];
-	void (*loadbios)(struct drm_device *, uint8_t *);
+	void (*shadow)(struct nvbios *);
 	const bool rw;
+	int score;
+	u32 size;
+	u8 *data;
 };
 
-static struct methods shadow_methods[] = {
-	{ "PRAMIN", load_vbios_pramin, true },
-	{ "PROM", load_vbios_prom, false },
-	{ "PCIROM", load_vbios_pci, true },
-	{ "ACPI", load_vbios_acpi, true },
-};
-#define NUM_SHADOW_METHODS ARRAY_SIZE(shadow_methods)
-
-static bool NVShadowVBIOS(struct drm_device *dev, uint8_t *data)
-{
-	struct methods *methods = shadow_methods;
-	int testscore = 3;
-	int scores[NUM_SHADOW_METHODS], i;
+static bool
+bios_shadow(struct drm_device *dev)
+{
+	struct methods shadow_methods[] = {
+		{ "PRAMIN", bios_shadow_pramin, true, 0, 0, NULL },
+		{ "PROM", bios_shadow_prom, false, 0, 0, NULL },
+		{ "ACPI", bios_shadow_acpi, true, 0, 0, NULL },
+		{ "PCIROM", bios_shadow_pci, true, 0, 0, NULL },
+		{}
+	};
+	struct drm_nouveau_private *dev_priv = dev->dev_private;
+	struct nvbios *bios = &dev_priv->vbios;
+	struct methods *mthd, *best;
 
 	if (nouveau_vbios) {
-		for (i = 0; i < NUM_SHADOW_METHODS; i++)
-			if (!strcasecmp(nouveau_vbios, methods[i].desc))
-				break;
-
-		if (i < NUM_SHADOW_METHODS) {
-			NV_INFO(dev, "Attempting to use BIOS image from %s\n",
-				methods[i].desc);
+		mthd = shadow_methods;
+		do {
+			if (strcasecmp(nouveau_vbios, mthd->desc))
+				continue;
+			NV_INFO(dev, "VBIOS source: %s\n", mthd->desc);
 
-			methods[i].loadbios(dev, data);
-			if (score_vbios(dev, data, methods[i].rw))
+			mthd->shadow(bios);
+			mthd->score = score_vbios(bios, mthd->rw);
+			if (mthd->score)
 				return true;
-		}
+		} while ((++mthd)->shadow);
 
 		NV_ERROR(dev, "VBIOS source \'%s\' invalid\n", nouveau_vbios);
 	}
 
-	for (i = 0; i < NUM_SHADOW_METHODS; i++) {
-		NV_TRACE(dev, "Attempting to load BIOS image from %s\n",
-			 methods[i].desc);
-		data[0] = data[1] = 0;	/* avoid reuse of previous image */
-		methods[i].loadbios(dev, data);
-		scores[i] = score_vbios(dev, data, methods[i].rw);
-		if (scores[i] == testscore)
-			return true;
-	}
-
-	while (--testscore > 0) {
-		for (i = 0; i < NUM_SHADOW_METHODS; i++) {
-			if (scores[i] == testscore) {
-				NV_TRACE(dev, "Using BIOS image from %s\n",
-					 methods[i].desc);
-				methods[i].loadbios(dev, data);
-				return true;
-			}
+	mthd = shadow_methods;
+	do {
+		NV_TRACE(dev, "Checking %s for VBIOS\n", mthd->desc);
+		mthd->shadow(bios);
+		mthd->score = score_vbios(bios, mthd->rw);
+		mthd->size = bios->length;
+		mthd->data = bios->data;
+	} while (mthd->score != 3 && (++mthd)->shadow);
+
+	mthd = shadow_methods;
+	best = mthd;
+	do {
+		if (mthd->score > best->score) {
+			kfree(best->data);
+			best = mthd;
 		}
+	} while ((++mthd)->shadow);
+
+	if (best->score) {
+		NV_TRACE(dev, "Using VBIOS from %s\n", best->desc);
+		bios->length = best->size;
+		bios->data = best->data;
+		return true;
 	}
 
-	NV_ERROR(dev, "No valid BIOS image found\n");
+	NV_ERROR(dev, "No valid VBIOS image found\n");
 	return false;
 }
 
@@ -1107,7 +1145,8 @@ init_dp_condition(struct nvbios *bios, uint16_t offset, struct init_exec *iexec)
 		break;
 	case 1:
 	case 2:
-		if (!(entry[5] & cond))
+		if ((table[0]  < 0x40 && !(entry[5] & cond)) ||
+		    (table[0] == 0x40 && !(entry[4] & cond)))
 			iexec->execute = false;
 		break;
 	case 5:
@@ -6334,11 +6373,7 @@ static bool NVInitVBIOS(struct drm_device *dev)
 	spin_lock_init(&bios->lock);
 	bios->dev = dev;
 
-	if (!NVShadowVBIOS(dev, bios->data))
-		return false;
-
-	bios->length = NV_PROM_SIZE;
-	return true;
+	return bios_shadow(dev);
 }
 
 static int nouveau_parse_vbios_struct(struct drm_device *dev)
@@ -6498,6 +6533,10 @@ nouveau_bios_init(struct drm_device *dev)
 void
 nouveau_bios_takedown(struct drm_device *dev)
 {
+	struct drm_nouveau_private *dev_priv = dev->dev_private;
+
 	nouveau_mxm_fini(dev);
 	nouveau_i2c_fini(dev);
+
+	kfree(dev_priv->vbios.data);
 }
diff --git a/drivers/gpu/drm/nouveau/nouveau_bios.h b/drivers/gpu/drm/nouveau/nouveau_bios.h
index a37c31e358aa..298a3af48d14 100644
--- a/drivers/gpu/drm/nouveau/nouveau_bios.h
+++ b/drivers/gpu/drm/nouveau/nouveau_bios.h
@@ -69,12 +69,16 @@ enum dcb_connector_type {
 	DCB_CONNECTOR_TV_3 = 0x13,
 	DCB_CONNECTOR_DVI_I = 0x30,
 	DCB_CONNECTOR_DVI_D = 0x31,
+	DCB_CONNECTOR_DMS59_0 = 0x38,
+	DCB_CONNECTOR_DMS59_1 = 0x39,
 	DCB_CONNECTOR_LVDS = 0x40,
 	DCB_CONNECTOR_LVDS_SPWG = 0x41,
 	DCB_CONNECTOR_DP = 0x46,
 	DCB_CONNECTOR_eDP = 0x47,
 	DCB_CONNECTOR_HDMI_0 = 0x60,
 	DCB_CONNECTOR_HDMI_1 = 0x61,
+	DCB_CONNECTOR_DMS59_DP0 = 0x64,
+	DCB_CONNECTOR_DMS59_DP1 = 0x65,
 	DCB_CONNECTOR_NONE = 0xff
 };
 
@@ -209,6 +213,8 @@ struct nvbios {
 		NVBIOS_BIT
 	} type;
 	uint16_t offset;
+	uint32_t length;
+	uint8_t *data;
 
 	uint8_t chip_version;
 
@@ -219,8 +225,6 @@ struct nvbios {
 
 	spinlock_t lock;
 
-	uint8_t data[NV_PROM_SIZE];
-	unsigned int length;
 	bool execute;
 
 	uint8_t major_version;
diff --git a/drivers/gpu/drm/nouveau/nouveau_bo.c b/drivers/gpu/drm/nouveau/nouveau_bo.c
index ec54364ac828..7d15a774f9c9 100644
--- a/drivers/gpu/drm/nouveau/nouveau_bo.c
+++ b/drivers/gpu/drm/nouveau/nouveau_bo.c
@@ -693,16 +693,12 @@ nouveau_bo_move_m2mf(struct ttm_buffer_object *bo, int evict, bool intr,
 		     struct ttm_mem_reg *new_mem)
 {
 	struct drm_nouveau_private *dev_priv = nouveau_bdev(bo->bdev);
+	struct nouveau_channel *chan = chan = dev_priv->channel;
 	struct nouveau_bo *nvbo = nouveau_bo(bo);
 	struct ttm_mem_reg *old_mem = &bo->mem;
-	struct nouveau_channel *chan;
 	int ret;
 
-	chan = nvbo->channel;
-	if (!chan) {
-		chan = dev_priv->channel;
-		mutex_lock_nested(&chan->mutex, NOUVEAU_KCHANNEL_MUTEX);
-	}
+	mutex_lock_nested(&chan->mutex, NOUVEAU_KCHANNEL_MUTEX);
 
 	/* create temporary vmas for the transfer and attach them to the
 	 * old nouveau_mem node, these will get cleaned up after ttm has
@@ -734,8 +730,7 @@ nouveau_bo_move_m2mf(struct ttm_buffer_object *bo, int evict, bool intr,
 	}
 
 out:
-	if (chan == dev_priv->channel)
-		mutex_unlock(&chan->mutex);
+	mutex_unlock(&chan->mutex);
 	return ret;
 }
 
diff --git a/drivers/gpu/drm/nouveau/nouveau_channel.c b/drivers/gpu/drm/nouveau/nouveau_channel.c
index a018defb7621..846afb0bfef4 100644
--- a/drivers/gpu/drm/nouveau/nouveau_channel.c
+++ b/drivers/gpu/drm/nouveau/nouveau_channel.c
@@ -122,7 +122,7 @@ nouveau_channel_alloc(struct drm_device *dev, struct nouveau_channel **chan_ret,
 	struct nouveau_fpriv *fpriv = nouveau_fpriv(file_priv);
 	struct nouveau_channel *chan;
 	unsigned long flags;
-	int ret;
+	int ret, i;
 
 	/* allocate and lock channel structure */
 	chan = kzalloc(sizeof(*chan), GFP_KERNEL);
@@ -184,7 +184,7 @@ nouveau_channel_alloc(struct drm_device *dev, struct nouveau_channel **chan_ret,
 		return ret;
 	}
 
-	nouveau_dma_pre_init(chan);
+	nouveau_dma_init(chan);
 	chan->user_put = 0x40;
 	chan->user_get = 0x44;
 	if (dev_priv->card_type >= NV_50)
@@ -202,9 +202,18 @@ nouveau_channel_alloc(struct drm_device *dev, struct nouveau_channel **chan_ret,
 
 	pfifo->reassign(dev, true);
 
-	ret = nouveau_dma_init(chan);
-	if (!ret)
-		ret = nouveau_fence_channel_init(chan);
+	/* Insert NOPs for NOUVEAU_DMA_SKIPS */
+	ret = RING_SPACE(chan, NOUVEAU_DMA_SKIPS);
+	if (ret) {
+		nouveau_channel_put(&chan);
+		return ret;
+	}
+
+	for (i = 0; i < NOUVEAU_DMA_SKIPS; i++)
+		OUT_RING  (chan, 0x00000000);
+	FIRE_RING(chan);
+
+	ret = nouveau_fence_channel_init(chan);
 	if (ret) {
 		nouveau_channel_put(&chan);
 		return ret;
@@ -427,18 +436,11 @@ nouveau_ioctl_fifo_alloc(struct drm_device *dev, void *data,
 	}
 
 	if (dev_priv->card_type < NV_C0) {
-		init->subchan[0].handle = NvM2MF;
-		if (dev_priv->card_type < NV_50)
-			init->subchan[0].grclass = 0x0039;
-		else
-			init->subchan[0].grclass = 0x5039;
+		init->subchan[0].handle = 0x00000000;
+		init->subchan[0].grclass = 0x0000;
 		init->subchan[1].handle = NvSw;
 		init->subchan[1].grclass = NV_SW;
 		init->nr_subchan = 2;
-	} else {
-		init->subchan[0].handle  = 0x9039;
-		init->subchan[0].grclass = 0x9039;
-		init->nr_subchan = 1;
 	}
 
 	/* Named memory object area */
diff --git a/drivers/gpu/drm/nouveau/nouveau_connector.c b/drivers/gpu/drm/nouveau/nouveau_connector.c
index f3ce34be082a..fa860358add1 100644
--- a/drivers/gpu/drm/nouveau/nouveau_connector.c
+++ b/drivers/gpu/drm/nouveau/nouveau_connector.c
@@ -519,6 +519,19 @@ nouveau_connector_set_property(struct drm_connector *connector,
 		return nv_crtc->set_dither(nv_crtc, true);
 	}
 
+	if (nv_crtc && nv_crtc->set_color_vibrance) {
+		/* Hue */
+		if (property == disp->vibrant_hue_property) {
+			nv_crtc->vibrant_hue = value - 90;
+			return nv_crtc->set_color_vibrance(nv_crtc, true);
+		}
+		/* Saturation */
+		if (property == disp->color_vibrance_property) {
+			nv_crtc->color_vibrance = value - 100;
+			return nv_crtc->set_color_vibrance(nv_crtc, true);
+		}
+	}
+
 	if (nv_encoder && nv_encoder->dcb->type == OUTPUT_TV)
 		return get_slave_funcs(encoder)->set_property(
 			encoder, connector, property, value);
@@ -641,10 +654,13 @@ nouveau_connector_detect_depth(struct drm_connector *connector)
 	if (nv_connector->edid && connector->display_info.bpc)
 		return;
 
-	/* if not, we're out of options unless we're LVDS, default to 6bpc */
-	connector->display_info.bpc = 6;
-	if (nv_encoder->dcb->type != OUTPUT_LVDS)
+	/* if not, we're out of options unless we're LVDS, default to 8bpc */
+	if (nv_encoder->dcb->type != OUTPUT_LVDS) {
+		connector->display_info.bpc = 8;
 		return;
+	}
+
+	connector->display_info.bpc = 6;
 
 	/* LVDS: panel straps */
 	if (bios->fp_no_ddc) {
@@ -854,10 +870,14 @@ drm_conntype_from_dcb(enum dcb_connector_type dcb)
 	case DCB_CONNECTOR_TV_0     :
 	case DCB_CONNECTOR_TV_1     :
 	case DCB_CONNECTOR_TV_3     : return DRM_MODE_CONNECTOR_TV;
+	case DCB_CONNECTOR_DMS59_0  :
+	case DCB_CONNECTOR_DMS59_1  :
 	case DCB_CONNECTOR_DVI_I    : return DRM_MODE_CONNECTOR_DVII;
 	case DCB_CONNECTOR_DVI_D    : return DRM_MODE_CONNECTOR_DVID;
 	case DCB_CONNECTOR_LVDS     :
 	case DCB_CONNECTOR_LVDS_SPWG: return DRM_MODE_CONNECTOR_LVDS;
+	case DCB_CONNECTOR_DMS59_DP0:
+	case DCB_CONNECTOR_DMS59_DP1:
 	case DCB_CONNECTOR_DP       : return DRM_MODE_CONNECTOR_DisplayPort;
 	case DCB_CONNECTOR_eDP      : return DRM_MODE_CONNECTOR_eDP;
 	case DCB_CONNECTOR_HDMI_0   :
@@ -998,11 +1018,10 @@ nouveau_connector_create(struct drm_device *dev, int index)
 
 	/* Add overscan compensation options to digital outputs */
 	if (disp->underscan_property &&
-	    (nv_connector->type == DCB_CONNECTOR_DVI_D ||
-	     nv_connector->type == DCB_CONNECTOR_DVI_I ||
-	     nv_connector->type == DCB_CONNECTOR_HDMI_0 ||
-	     nv_connector->type == DCB_CONNECTOR_HDMI_1 ||
-	     nv_connector->type == DCB_CONNECTOR_DP)) {
+	    (type == DRM_MODE_CONNECTOR_DVID ||
+	     type == DRM_MODE_CONNECTOR_DVII ||
+	     type == DRM_MODE_CONNECTOR_HDMIA ||
+	     type == DRM_MODE_CONNECTOR_DisplayPort)) {
 		drm_connector_attach_property(connector,
 					      disp->underscan_property,
 					      UNDERSCAN_OFF);
@@ -1014,6 +1033,16 @@ nouveau_connector_create(struct drm_device *dev, int index)
 					      0);
 	}
 
+	/* Add hue and saturation options */
+	if (disp->vibrant_hue_property)
+		drm_connector_attach_property(connector,
+					      disp->vibrant_hue_property,
+					      90);
+	if (disp->color_vibrance_property)
+		drm_connector_attach_property(connector,
+					      disp->color_vibrance_property,
+					      150);
+
 	switch (nv_connector->type) {
 	case DCB_CONNECTOR_VGA:
 		if (dev_priv->card_type >= NV_50) {
diff --git a/drivers/gpu/drm/nouveau/nouveau_crtc.h b/drivers/gpu/drm/nouveau/nouveau_crtc.h
index 686f6b4a1da3..e6d0d1eb0133 100644
--- a/drivers/gpu/drm/nouveau/nouveau_crtc.h
+++ b/drivers/gpu/drm/nouveau/nouveau_crtc.h
@@ -35,6 +35,8 @@ struct nouveau_crtc {
 	uint32_t dpms_saved_fp_control;
 	uint32_t fp_users;
 	int saturation;
+	int color_vibrance;
+	int vibrant_hue;
 	int sharpness;
 	int last_dpms;
 
@@ -67,6 +69,7 @@ struct nouveau_crtc {
 
 	int (*set_dither)(struct nouveau_crtc *crtc, bool update);
 	int (*set_scale)(struct nouveau_crtc *crtc, bool update);
+	int (*set_color_vibrance)(struct nouveau_crtc *crtc, bool update);
 };
 
 static inline struct nouveau_crtc *nouveau_crtc(struct drm_crtc *crtc)
diff --git a/drivers/gpu/drm/nouveau/nouveau_display.c b/drivers/gpu/drm/nouveau/nouveau_display.c
index 795a9e3c990a..a85e112863d1 100644
--- a/drivers/gpu/drm/nouveau/nouveau_display.c
+++ b/drivers/gpu/drm/nouveau/nouveau_display.c
@@ -155,20 +155,20 @@ static const struct drm_mode_config_funcs nouveau_mode_config_funcs = {
 };
 
 
-struct drm_prop_enum_list {
+struct nouveau_drm_prop_enum_list {
 	u8 gen_mask;
 	int type;
 	char *name;
 };
 
-static struct drm_prop_enum_list underscan[] = {
+static struct nouveau_drm_prop_enum_list underscan[] = {
 	{ 6, UNDERSCAN_AUTO, "auto" },
 	{ 6, UNDERSCAN_OFF, "off" },
 	{ 6, UNDERSCAN_ON, "on" },
 	{}
 };
 
-static struct drm_prop_enum_list dither_mode[] = {
+static struct nouveau_drm_prop_enum_list dither_mode[] = {
 	{ 7, DITHERING_MODE_AUTO, "auto" },
 	{ 7, DITHERING_MODE_OFF, "off" },
 	{ 1, DITHERING_MODE_ON, "on" },
@@ -178,7 +178,7 @@ static struct drm_prop_enum_list dither_mode[] = {
 	{}
 };
 
-static struct drm_prop_enum_list dither_depth[] = {
+static struct nouveau_drm_prop_enum_list dither_depth[] = {
 	{ 6, DITHERING_DEPTH_AUTO, "auto" },
 	{ 6, DITHERING_DEPTH_6BPC, "6 bpc" },
 	{ 6, DITHERING_DEPTH_8BPC, "8 bpc" },
@@ -186,7 +186,7 @@ static struct drm_prop_enum_list dither_depth[] = {
 };
 
 #define PROP_ENUM(p,gen,n,list) do {                                           \
-	struct drm_prop_enum_list *l = (list);                                 \
+	struct nouveau_drm_prop_enum_list *l = (list);                         \
 	int c = 0;                                                             \
 	while (l->gen_mask) {                                                  \
 		if (l->gen_mask & (1 << (gen)))                                \
@@ -281,16 +281,24 @@ nouveau_display_create(struct drm_device *dev)
 	PROP_ENUM(disp->underscan_property, gen, "underscan", underscan);
 
 	disp->underscan_hborder_property =
-		drm_property_create(dev, DRM_MODE_PROP_RANGE,
-				    "underscan hborder", 2);
-	disp->underscan_hborder_property->values[0] = 0;
-	disp->underscan_hborder_property->values[1] = 128;
+		drm_property_create_range(dev, 0, "underscan hborder", 0, 128);
 
 	disp->underscan_vborder_property =
-		drm_property_create(dev, DRM_MODE_PROP_RANGE,
-				    "underscan vborder", 2);
-	disp->underscan_vborder_property->values[0] = 0;
-	disp->underscan_vborder_property->values[1] = 128;
+		drm_property_create_range(dev, 0, "underscan vborder", 0, 128);
+
+	if (gen == 1) {
+		disp->vibrant_hue_property =
+			drm_property_create(dev, DRM_MODE_PROP_RANGE,
+					    "vibrant hue", 2);
+		disp->vibrant_hue_property->values[0] = 0;
+		disp->vibrant_hue_property->values[1] = 180; /* -90..+90 */
+
+		disp->color_vibrance_property =
+			drm_property_create(dev, DRM_MODE_PROP_RANGE,
+					    "color vibrance", 2);
+		disp->color_vibrance_property->values[0] = 0;
+		disp->color_vibrance_property->values[1] = 200; /* -100..+100 */
+	}
 
 	dev->mode_config.funcs = (void *)&nouveau_mode_config_funcs;
 	dev->mode_config.fb_base = pci_resource_start(dev->pdev, 1);
@@ -309,6 +317,9 @@ nouveau_display_create(struct drm_device *dev)
 		dev->mode_config.max_height = 8192;
 	}
 
+	dev->mode_config.preferred_depth = 24;
+	dev->mode_config.prefer_shadow = 1;
+
 	drm_kms_helper_poll_init(dev);
 	drm_kms_helper_poll_disable(dev);
 
@@ -430,15 +441,19 @@ nouveau_page_flip_emit(struct nouveau_channel *chan,
 		goto fail;
 
 	/* Emit the pageflip */
-	ret = RING_SPACE(chan, 2);
+	ret = RING_SPACE(chan, 3);
 	if (ret)
 		goto fail;
 
-	if (dev_priv->card_type < NV_C0)
+	if (dev_priv->card_type < NV_C0) {
 		BEGIN_RING(chan, NvSubSw, NV_SW_PAGE_FLIP, 1);
-	else
-		BEGIN_NVC0(chan, 2, NvSubM2MF, 0x0500, 1);
-	OUT_RING  (chan, 0);
+		OUT_RING  (chan, 0x00000000);
+		OUT_RING  (chan, 0x00000000);
+	} else {
+		BEGIN_NVC0(chan, 2, 0, NV10_SUBCHAN_REF_CNT, 1);
+		OUT_RING  (chan, ++chan->fence.sequence);
+		BEGIN_NVC0(chan, 8, 0, NVSW_SUBCHAN_PAGE_FLIP, 0x0000);
+	}
 	FIRE_RING (chan);
 
 	ret = nouveau_fence_new(chan, pfence, true);
diff --git a/drivers/gpu/drm/nouveau/nouveau_dma.c b/drivers/gpu/drm/nouveau/nouveau_dma.c
index 4c2e4e5925fe..295932e66ac5 100644
--- a/drivers/gpu/drm/nouveau/nouveau_dma.c
+++ b/drivers/gpu/drm/nouveau/nouveau_dma.c
@@ -31,7 +31,7 @@
 #include "nouveau_ramht.h"
 
 void
-nouveau_dma_pre_init(struct nouveau_channel *chan)
+nouveau_dma_init(struct nouveau_channel *chan)
 {
 	struct drm_nouveau_private *dev_priv = chan->dev->dev_private;
 	struct nouveau_bo *pushbuf = chan->pushbuf_bo;
@@ -54,65 +54,6 @@ nouveau_dma_pre_init(struct nouveau_channel *chan)
 	chan->dma.free = chan->dma.max - chan->dma.cur;
 }
 
-int
-nouveau_dma_init(struct nouveau_channel *chan)
-{
-	struct drm_device *dev = chan->dev;
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	int ret, i;
-
-	if (dev_priv->card_type >= NV_C0) {
-		ret = nouveau_gpuobj_gr_new(chan, 0x9039, 0x9039);
-		if (ret)
-			return ret;
-
-		ret = RING_SPACE(chan, 2);
-		if (ret)
-			return ret;
-
-		BEGIN_NVC0(chan, 2, NvSubM2MF, 0x0000, 1);
-		OUT_RING  (chan, 0x00009039);
-		FIRE_RING (chan);
-		return 0;
-	}
-
-	/* Create NV_MEMORY_TO_MEMORY_FORMAT for buffer moves */
-	ret = nouveau_gpuobj_gr_new(chan, NvM2MF, dev_priv->card_type < NV_50 ?
-				    0x0039 : 0x5039);
-	if (ret)
-		return ret;
-
-	/* NV_MEMORY_TO_MEMORY_FORMAT requires a notifier object */
-	ret = nouveau_notifier_alloc(chan, NvNotify0, 32, 0xfe0, 0x1000,
-				     &chan->m2mf_ntfy);
-	if (ret)
-		return ret;
-
-	/* Insert NOPS for NOUVEAU_DMA_SKIPS */
-	ret = RING_SPACE(chan, NOUVEAU_DMA_SKIPS);
-	if (ret)
-		return ret;
-
-	for (i = 0; i < NOUVEAU_DMA_SKIPS; i++)
-		OUT_RING(chan, 0);
-
-	/* Initialise NV_MEMORY_TO_MEMORY_FORMAT */
-	ret = RING_SPACE(chan, 6);
-	if (ret)
-		return ret;
-	BEGIN_RING(chan, NvSubM2MF, NV_MEMORY_TO_MEMORY_FORMAT_NAME, 1);
-	OUT_RING  (chan, NvM2MF);
-	BEGIN_RING(chan, NvSubM2MF, NV_MEMORY_TO_MEMORY_FORMAT_DMA_NOTIFY, 3);
-	OUT_RING  (chan, NvNotify0);
-	OUT_RING  (chan, chan->vram_handle);
-	OUT_RING  (chan, chan->gart_handle);
-
-	/* Sit back and pray the channel works.. */
-	FIRE_RING(chan);
-
-	return 0;
-}
-
 void
 OUT_RINGp(struct nouveau_channel *chan, const void *data, unsigned nr_dwords)
 {
diff --git a/drivers/gpu/drm/nouveau/nouveau_dp.c b/drivers/gpu/drm/nouveau/nouveau_dp.c
index 9b93b703ceab..d996134b1b28 100644
--- a/drivers/gpu/drm/nouveau/nouveau_dp.c
+++ b/drivers/gpu/drm/nouveau/nouveau_dp.c
@@ -161,116 +161,6 @@ out:
 	return ret;
 }
 
-static u32
-dp_link_bw_get(struct drm_device *dev, int or, int link)
-{
-	u32 ctrl = nv_rd32(dev, 0x614300 + (or * 0x800));
-	if (!(ctrl & 0x000c0000))
-		return 162000;
-	return 270000;
-}
-
-static int
-dp_lane_count_get(struct drm_device *dev, int or, int link)
-{
-	u32 ctrl = nv_rd32(dev, NV50_SOR_DP_CTRL(or, link));
-	switch (ctrl & 0x000f0000) {
-	case 0x00010000: return 1;
-	case 0x00030000: return 2;
-	default:
-		return 4;
-	}
-}
-
-void
-nouveau_dp_tu_update(struct drm_device *dev, int or, int link, u32 clk, u32 bpp)
-{
-	const u32 symbol = 100000;
-	int bestTU = 0, bestVTUi = 0, bestVTUf = 0, bestVTUa = 0;
-	int TU, VTUi, VTUf, VTUa;
-	u64 link_data_rate, link_ratio, unk;
-	u32 best_diff = 64 * symbol;
-	u32 link_nr, link_bw, r;
-
-	/* calculate packed data rate for each lane */
-	link_nr = dp_lane_count_get(dev, or, link);
-	link_data_rate = (clk * bpp / 8) / link_nr;
-
-	/* calculate ratio of packed data rate to link symbol rate */
-	link_bw = dp_link_bw_get(dev, or, link);
-	link_ratio = link_data_rate * symbol;
-	r = do_div(link_ratio, link_bw);
-
-	for (TU = 64; TU >= 32; TU--) {
-		/* calculate average number of valid symbols in each TU */
-		u32 tu_valid = link_ratio * TU;
-		u32 calc, diff;
-
-		/* find a hw representation for the fraction.. */
-		VTUi = tu_valid / symbol;
-		calc = VTUi * symbol;
-		diff = tu_valid - calc;
-		if (diff) {
-			if (diff >= (symbol / 2)) {
-				VTUf = symbol / (symbol - diff);
-				if (symbol - (VTUf * diff))
-					VTUf++;
-
-				if (VTUf <= 15) {
-					VTUa  = 1;
-					calc += symbol - (symbol / VTUf);
-				} else {
-					VTUa  = 0;
-					VTUf  = 1;
-					calc += symbol;
-				}
-			} else {
-				VTUa  = 0;
-				VTUf  = min((int)(symbol / diff), 15);
-				calc += symbol / VTUf;
-			}
-
-			diff = calc - tu_valid;
-		} else {
-			/* no remainder, but the hw doesn't like the fractional
-			 * part to be zero.  decrement the integer part and
-			 * have the fraction add a whole symbol back
-			 */
-			VTUa = 0;
-			VTUf = 1;
-			VTUi--;
-		}
-
-		if (diff < best_diff) {
-			best_diff = diff;
-			bestTU = TU;
-			bestVTUa = VTUa;
-			bestVTUf = VTUf;
-			bestVTUi = VTUi;
-			if (diff == 0)
-				break;
-		}
-	}
-
-	if (!bestTU) {
-		NV_ERROR(dev, "DP: unable to find suitable config\n");
-		return;
-	}
-
-	/* XXX close to vbios numbers, but not right */
-	unk  = (symbol - link_ratio) * bestTU;
-	unk *= link_ratio;
-	r = do_div(unk, symbol);
-	r = do_div(unk, symbol);
-	unk += 6;
-
-	nv_mask(dev, NV50_SOR_DP_CTRL(or, link), 0x000001fc, bestTU << 2);
-	nv_mask(dev, NV50_SOR_DP_SCFG(or, link), 0x010f7f3f, bestVTUa << 24 |
-							     bestVTUf << 16 |
-							     bestVTUi << 8 |
-							     unk);
-}
-
 u8 *
 nouveau_dp_bios_data(struct drm_device *dev, struct dcb_entry *dcb, u8 **entry)
 {
@@ -298,6 +188,7 @@ nouveau_dp_bios_data(struct drm_device *dev, struct dcb_entry *dcb, u8 **entry)
 	case 0x20:
 	case 0x21:
 	case 0x30:
+	case 0x40:
 		break;
 	default:
 		NV_ERROR(dev, "displayport table 0x%02x unknown\n", table[0]);
@@ -318,13 +209,10 @@ nouveau_dp_bios_data(struct drm_device *dev, struct dcb_entry *dcb, u8 **entry)
  * link training
  *****************************************************************************/
 struct dp_state {
+	struct dp_train_func *func;
 	struct dcb_entry *dcb;
-	u8 *table;
-	u8 *entry;
 	int auxch;
 	int crtc;
-	int or;
-	int link;
 	u8 *dpcd;
 	int link_nr;
 	u32 link_bw;
@@ -335,142 +223,58 @@ struct dp_state {
 static void
 dp_set_link_config(struct drm_device *dev, struct dp_state *dp)
 {
-	int or = dp->or, link = dp->link;
-	u8 *entry, sink[2];
-	u32 dp_ctrl;
-	u16 script;
+	u8 sink[2];
 
 	NV_DEBUG_KMS(dev, "%d lanes at %d KB/s\n", dp->link_nr, dp->link_bw);
 
-	/* set selected link rate on source */
-	switch (dp->link_bw) {
-	case 270000:
-		nv_mask(dev, 0x614300 + (or * 0x800), 0x000c0000, 0x00040000);
-		sink[0] = DP_LINK_BW_2_7;
-		break;
-	default:
-		nv_mask(dev, 0x614300 + (or * 0x800), 0x000c0000, 0x00000000);
-		sink[0] = DP_LINK_BW_1_62;
-		break;
-	}
-
-	/* offset +0x0a of each dp encoder table entry is a pointer to another
-	 * table, that has (among other things) pointers to more scripts that
-	 * need to be executed, this time depending on link speed.
-	 */
-	entry = ROMPTR(dev, dp->entry[10]);
-	if (entry) {
-		if (dp->table[0] < 0x30) {
-			while (dp->link_bw < (ROM16(entry[0]) * 10))
-				entry += 4;
-			script = ROM16(entry[2]);
-		} else {
-			while (dp->link_bw < (entry[0] * 27000))
-				entry += 3;
-			script = ROM16(entry[1]);
-		}
-
-		nouveau_bios_run_init_table(dev, script, dp->dcb, dp->crtc);
-	}
+	/* set desired link configuration on the source */
+	dp->func->link_set(dev, dp->dcb, dp->crtc, dp->link_nr, dp->link_bw,
+			   dp->dpcd[2] & DP_ENHANCED_FRAME_CAP);
 
-	/* configure lane count on the source */
-	dp_ctrl = ((1 << dp->link_nr) - 1) << 16;
+	/* inform the sink of the new configuration */
+	sink[0] = dp->link_bw / 27000;
 	sink[1] = dp->link_nr;
-	if (dp->dpcd[2] & DP_ENHANCED_FRAME_CAP) {
-		dp_ctrl |= 0x00004000;
+	if (dp->dpcd[2] & DP_ENHANCED_FRAME_CAP)
 		sink[1] |= DP_LANE_COUNT_ENHANCED_FRAME_EN;
-	}
-
-	nv_mask(dev, NV50_SOR_DP_CTRL(or, link), 0x001f4000, dp_ctrl);
 
-	/* inform the sink of the new configuration */
 	auxch_tx(dev, dp->auxch, 8, DP_LINK_BW_SET, sink, 2);
 }
 
 static void
-dp_set_training_pattern(struct drm_device *dev, struct dp_state *dp, u8 tp)
+dp_set_training_pattern(struct drm_device *dev, struct dp_state *dp, u8 pattern)
 {
 	u8 sink_tp;
 
-	NV_DEBUG_KMS(dev, "training pattern %d\n", tp);
+	NV_DEBUG_KMS(dev, "training pattern %d\n", pattern);
 
-	nv_mask(dev, NV50_SOR_DP_CTRL(dp->or, dp->link), 0x0f000000, tp << 24);
+	dp->func->train_set(dev, dp->dcb, pattern);
 
 	auxch_tx(dev, dp->auxch, 9, DP_TRAINING_PATTERN_SET, &sink_tp, 1);
 	sink_tp &= ~DP_TRAINING_PATTERN_MASK;
-	sink_tp |= tp;
+	sink_tp |= pattern;
 	auxch_tx(dev, dp->auxch, 8, DP_TRAINING_PATTERN_SET, &sink_tp, 1);
 }
 
-static const u8 nv50_lane_map[] = { 16, 8, 0, 24 };
-static const u8 nvaf_lane_map[] = { 24, 16, 8, 0 };
-
 static int
 dp_link_train_commit(struct drm_device *dev, struct dp_state *dp)
 {
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	u32 mask = 0, drv = 0, pre = 0, unk = 0;
-	const u8 *shifts;
-	int link = dp->link;
-	int or = dp->or;
 	int i;
 
-	if (dev_priv->chipset != 0xaf)
-		shifts = nv50_lane_map;
-	else
-		shifts = nvaf_lane_map;
-
 	for (i = 0; i < dp->link_nr; i++) {
-		u8 *conf = dp->entry + dp->table[4];
 		u8 lane = (dp->stat[4 + (i >> 1)] >> ((i & 1) * 4)) & 0xf;
 		u8 lpre = (lane & 0x0c) >> 2;
 		u8 lvsw = (lane & 0x03) >> 0;
 
-		mask |= 0xff << shifts[i];
-		unk |= 1 << (shifts[i] >> 3);
-
 		dp->conf[i] = (lpre << 3) | lvsw;
 		if (lvsw == DP_TRAIN_VOLTAGE_SWING_1200)
 			dp->conf[i] |= DP_TRAIN_MAX_SWING_REACHED;
-		if (lpre == DP_TRAIN_PRE_EMPHASIS_9_5)
+		if ((lpre << 3) == DP_TRAIN_PRE_EMPHASIS_9_5)
 			dp->conf[i] |= DP_TRAIN_MAX_PRE_EMPHASIS_REACHED;
 
 		NV_DEBUG_KMS(dev, "config lane %d %02x\n", i, dp->conf[i]);
-
-		if (dp->table[0] < 0x30) {
-			u8 *last = conf + (dp->entry[4] * dp->table[5]);
-			while (lvsw != conf[0] || lpre != conf[1]) {
-				conf += dp->table[5];
-				if (conf >= last)
-					return -EINVAL;
-			}
-
-			conf += 2;
-		} else {
-			/* no lookup table anymore, set entries for each
-			 * combination of voltage swing and pre-emphasis
-			 * level allowed by the DP spec.
-			 */
-			switch (lvsw) {
-			case 0: lpre += 0; break;
-			case 1: lpre += 4; break;
-			case 2: lpre += 7; break;
-			case 3: lpre += 9; break;
-			}
-
-			conf = conf + (lpre * dp->table[5]);
-			conf++;
-		}
-
-		drv |= conf[0] << shifts[i];
-		pre |= conf[1] << shifts[i];
-		unk  = (unk & ~0x0000ff00) | (conf[2] << 8);
+		dp->func->train_adj(dev, dp->dcb, i, lvsw, lpre);
 	}
 
-	nv_mask(dev, NV50_SOR_DP_UNK118(or, link), mask, drv);
-	nv_mask(dev, NV50_SOR_DP_UNK120(or, link), mask, pre);
-	nv_mask(dev, NV50_SOR_DP_UNK130(or, link), 0x0000ff0f, unk);
-
 	return auxch_tx(dev, dp->auxch, 8, DP_TRAINING_LANE0_SET, dp->conf, 4);
 }
 
@@ -554,8 +358,60 @@ dp_link_train_eq(struct drm_device *dev, struct dp_state *dp)
 	return eq_done ? 0 : -1;
 }
 
+static void
+dp_set_downspread(struct drm_device *dev, struct dp_state *dp, bool enable)
+{
+	u16 script = 0x0000;
+	u8 *entry, *table = nouveau_dp_bios_data(dev, dp->dcb, &entry);
+	if (table) {
+		if (table[0] >= 0x20 && table[0] <= 0x30) {
+			if (enable) script = ROM16(entry[12]);
+			else        script = ROM16(entry[14]);
+		} else
+		if (table[0] == 0x40) {
+			if (enable) script = ROM16(entry[11]);
+			else        script = ROM16(entry[13]);
+		}
+	}
+
+	nouveau_bios_run_init_table(dev, script, dp->dcb, dp->crtc);
+}
+
+static void
+dp_link_train_init(struct drm_device *dev, struct dp_state *dp)
+{
+	u16 script = 0x0000;
+	u8 *entry, *table = nouveau_dp_bios_data(dev, dp->dcb, &entry);
+	if (table) {
+		if (table[0] >= 0x20 && table[0] <= 0x30)
+			script = ROM16(entry[6]);
+		else
+		if (table[0] == 0x40)
+			script = ROM16(entry[5]);
+	}
+
+	nouveau_bios_run_init_table(dev, script, dp->dcb, dp->crtc);
+}
+
+static void
+dp_link_train_fini(struct drm_device *dev, struct dp_state *dp)
+{
+	u16 script = 0x0000;
+	u8 *entry, *table = nouveau_dp_bios_data(dev, dp->dcb, &entry);
+	if (table) {
+		if (table[0] >= 0x20 && table[0] <= 0x30)
+			script = ROM16(entry[8]);
+		else
+		if (table[0] == 0x40)
+			script = ROM16(entry[7]);
+	}
+
+	nouveau_bios_run_init_table(dev, script, dp->dcb, dp->crtc);
+}
+
 bool
-nouveau_dp_link_train(struct drm_encoder *encoder, u32 datarate)
+nouveau_dp_link_train(struct drm_encoder *encoder, u32 datarate,
+		      struct dp_train_func *func)
 {
 	struct nouveau_encoder *nv_encoder = nouveau_encoder(encoder);
 	struct nouveau_crtc *nv_crtc = nouveau_crtc(encoder->crtc);
@@ -571,17 +427,15 @@ nouveau_dp_link_train(struct drm_encoder *encoder, u32 datarate)
 	if (!auxch)
 		return false;
 
-	dp.table = nouveau_dp_bios_data(dev, nv_encoder->dcb, &dp.entry);
-	if (!dp.table)
-		return -EINVAL;
-
+	dp.func = func;
 	dp.dcb = nv_encoder->dcb;
 	dp.crtc = nv_crtc->index;
 	dp.auxch = auxch->drive;
-	dp.or = nv_encoder->or;
-	dp.link = !(nv_encoder->dcb->sorconf.link & 1);
 	dp.dpcd = nv_encoder->dp.dpcd;
 
+	/* adjust required bandwidth for 8B/10B coding overhead */
+	datarate = (datarate / 8) * 10;
+
 	/* some sinks toggle hotplug in response to some of the actions
 	 * we take during link training (DP_SET_POWER is one), we need
 	 * to ignore them for the moment to avoid races.
@@ -589,16 +443,10 @@ nouveau_dp_link_train(struct drm_encoder *encoder, u32 datarate)
 	nouveau_gpio_irq(dev, 0, nv_connector->hpd, 0xff, false);
 
 	/* enable down-spreading, if possible */
-	if (dp.table[1] >= 16) {
-		u16 script = ROM16(dp.entry[14]);
-		if (nv_encoder->dp.dpcd[3] & 1)
-			script = ROM16(dp.entry[12]);
-
-		nouveau_bios_run_init_table(dev, script, dp.dcb, dp.crtc);
-	}
+	dp_set_downspread(dev, &dp, nv_encoder->dp.dpcd[3] & 1);
 
 	/* execute pre-train script from vbios */
-	nouveau_bios_run_init_table(dev, ROM16(dp.entry[6]), dp.dcb, dp.crtc);
+	dp_link_train_init(dev, &dp);
 
 	/* start off at highest link rate supported by encoder and display */
 	while (*link_bw > nv_encoder->dp.link_bw)
@@ -632,13 +480,36 @@ nouveau_dp_link_train(struct drm_encoder *encoder, u32 datarate)
 	dp_set_training_pattern(dev, &dp, DP_TRAINING_PATTERN_DISABLE);
 
 	/* execute post-train script from vbios */
-	nouveau_bios_run_init_table(dev, ROM16(dp.entry[8]), dp.dcb, dp.crtc);
+	dp_link_train_fini(dev, &dp);
 
 	/* re-enable hotplug detect */
 	nouveau_gpio_irq(dev, 0, nv_connector->hpd, 0xff, true);
 	return true;
 }
 
+void
+nouveau_dp_dpms(struct drm_encoder *encoder, int mode, u32 datarate,
+		struct dp_train_func *func)
+{
+	struct nouveau_encoder *nv_encoder = nouveau_encoder(encoder);
+	struct nouveau_i2c_chan *auxch;
+	u8 status;
+
+	auxch = nouveau_i2c_find(encoder->dev, nv_encoder->dcb->i2c_index);
+	if (!auxch)
+		return;
+
+	if (mode == DRM_MODE_DPMS_ON)
+		status = DP_SET_POWER_D0;
+	else
+		status = DP_SET_POWER_D3;
+
+	nouveau_dp_auxch(auxch, 8, DP_SET_POWER, &status, 1);
+
+	if (mode == DRM_MODE_DPMS_ON)
+		nouveau_dp_link_train(encoder, datarate, func);
+}
+
 bool
 nouveau_dp_detect(struct drm_encoder *encoder)
 {
diff --git a/drivers/gpu/drm/nouveau/nouveau_drv.c b/drivers/gpu/drm/nouveau/nouveau_drv.c
index 81d7962e7252..4f2030bd5676 100644
--- a/drivers/gpu/drm/nouveau/nouveau_drv.c
+++ b/drivers/gpu/drm/nouveau/nouveau_drv.c
@@ -57,6 +57,10 @@ MODULE_PARM_DESC(vram_notify, "Force DMA notifiers to be in VRAM");
 int nouveau_vram_notify = 0;
 module_param_named(vram_notify, nouveau_vram_notify, int, 0400);
 
+MODULE_PARM_DESC(vram_type, "Override detected VRAM type");
+char *nouveau_vram_type;
+module_param_named(vram_type, nouveau_vram_type, charp, 0400);
+
 MODULE_PARM_DESC(duallink, "Allow dual-link TMDS (>=GeForce 8)");
 int nouveau_duallink = 1;
 module_param_named(duallink, nouveau_duallink, int, 0400);
@@ -89,7 +93,7 @@ MODULE_PARM_DESC(override_conntype, "Ignore DCB connector type");
 int nouveau_override_conntype = 0;
 module_param_named(override_conntype, nouveau_override_conntype, int, 0400);
 
-MODULE_PARM_DESC(tv_disable, "Disable TV-out detection\n");
+MODULE_PARM_DESC(tv_disable, "Disable TV-out detection");
 int nouveau_tv_disable = 0;
 module_param_named(tv_disable, nouveau_tv_disable, int, 0400);
 
@@ -104,27 +108,27 @@ module_param_named(tv_norm, nouveau_tv_norm, charp, 0400);
 MODULE_PARM_DESC(reg_debug, "Register access debug bitmask:\n"
 		"\t\t0x1 mc, 0x2 video, 0x4 fb, 0x8 extdev,\n"
 		"\t\t0x10 crtc, 0x20 ramdac, 0x40 vgacrtc, 0x80 rmvio,\n"
-		"\t\t0x100 vgaattr, 0x200 EVO (G80+). ");
+		"\t\t0x100 vgaattr, 0x200 EVO (G80+)");
 int nouveau_reg_debug;
 module_param_named(reg_debug, nouveau_reg_debug, int, 0600);
 
-MODULE_PARM_DESC(perflvl, "Performance level (default: boot)\n");
+MODULE_PARM_DESC(perflvl, "Performance level (default: boot)");
 char *nouveau_perflvl;
 module_param_named(perflvl, nouveau_perflvl, charp, 0400);
 
-MODULE_PARM_DESC(perflvl_wr, "Allow perflvl changes (warning: dangerous!)\n");
+MODULE_PARM_DESC(perflvl_wr, "Allow perflvl changes (warning: dangerous!)");
 int nouveau_perflvl_wr;
 module_param_named(perflvl_wr, nouveau_perflvl_wr, int, 0400);
 
-MODULE_PARM_DESC(msi, "Enable MSI (default: off)\n");
+MODULE_PARM_DESC(msi, "Enable MSI (default: off)");
 int nouveau_msi;
 module_param_named(msi, nouveau_msi, int, 0400);
 
-MODULE_PARM_DESC(ctxfw, "Use external HUB/GPC ucode (fermi)\n");
+MODULE_PARM_DESC(ctxfw, "Use external HUB/GPC ucode (fermi)");
 int nouveau_ctxfw;
 module_param_named(ctxfw, nouveau_ctxfw, int, 0400);
 
-MODULE_PARM_DESC(mxmdcb, "Santise DCB table according to MXM-SIS\n");
+MODULE_PARM_DESC(mxmdcb, "Santise DCB table according to MXM-SIS");
 int nouveau_mxmdcb = 1;
 module_param_named(mxmdcb, nouveau_mxmdcb, int, 0400);
 
diff --git a/drivers/gpu/drm/nouveau/nouveau_drv.h b/drivers/gpu/drm/nouveau/nouveau_drv.h
index b82709828931..3aef353a926c 100644
--- a/drivers/gpu/drm/nouveau/nouveau_drv.h
+++ b/drivers/gpu/drm/nouveau/nouveau_drv.h
@@ -26,15 +26,15 @@
 #define __NOUVEAU_DRV_H__
 
 #define DRIVER_AUTHOR		"Stephane Marchesin"
-#define DRIVER_EMAIL		"dri-devel@lists.sourceforge.net"
+#define DRIVER_EMAIL		"nouveau@lists.freedesktop.org"
 
 #define DRIVER_NAME		"nouveau"
 #define DRIVER_DESC		"nVidia Riva/TNT/GeForce"
-#define DRIVER_DATE		"20090420"
+#define DRIVER_DATE		"20120316"
 
-#define DRIVER_MAJOR		0
+#define DRIVER_MAJOR		1
 #define DRIVER_MINOR		0
-#define DRIVER_PATCHLEVEL	16
+#define DRIVER_PATCHLEVEL	0
 
 #define NOUVEAU_FAMILY   0x0000FFFF
 #define NOUVEAU_FLAGS    0xFFFF0000
@@ -113,8 +113,6 @@ struct nouveau_bo {
 	int pbbo_index;
 	bool validate_mapped;
 
-	struct nouveau_channel *channel;
-
 	struct list_head vma_list;
 	unsigned page_shift;
 
@@ -296,7 +294,7 @@ struct nouveau_channel {
 
 	uint32_t sw_subchannel[8];
 
-	struct nouveau_vma dispc_vma[2];
+	struct nouveau_vma dispc_vma[4];
 	struct {
 		struct nouveau_gpuobj *vblsem;
 		uint32_t vblsem_head;
@@ -406,6 +404,9 @@ struct nouveau_display_engine {
 	struct drm_property *underscan_property;
 	struct drm_property *underscan_hborder_property;
 	struct drm_property *underscan_vborder_property;
+	/* not really hue and saturation: */
+	struct drm_property *vibrant_hue_property;
+	struct drm_property *color_vibrance_property;
 };
 
 struct nouveau_gpio_engine {
@@ -432,58 +433,85 @@ struct nouveau_pm_voltage {
 	int nr_level;
 };
 
+/* Exclusive upper limits */
+#define NV_MEM_CL_DDR2_MAX 8
+#define NV_MEM_WR_DDR2_MAX 9
+#define NV_MEM_CL_DDR3_MAX 17
+#define NV_MEM_WR_DDR3_MAX 17
+#define NV_MEM_CL_GDDR3_MAX 16
+#define NV_MEM_WR_GDDR3_MAX 18
+#define NV_MEM_CL_GDDR5_MAX 21
+#define NV_MEM_WR_GDDR5_MAX 20
+
 struct nouveau_pm_memtiming {
 	int id;
-	u32 reg_0; /* 0x10f290 on Fermi, 0x100220 for older */
-	u32 reg_1;
-	u32 reg_2;
-	u32 reg_3;
-	u32 reg_4;
-	u32 reg_5;
-	u32 reg_6;
-	u32 reg_7;
-	u32 reg_8;
-	/* To be written to 0x1002c0 */
-	u8 CL;
-	u8 WR;
+
+	u32 reg[9];
+	u32 mr[4];
+
+	u8 tCWL;
+
+	u8 odt;
+	u8 drive_strength;
 };
 
-struct nouveau_pm_tbl_header{
+struct nouveau_pm_tbl_header {
 	u8 version;
 	u8 header_len;
 	u8 entry_cnt;
 	u8 entry_len;
 };
 
-struct nouveau_pm_tbl_entry{
+struct nouveau_pm_tbl_entry {
 	u8 tWR;
-	u8 tUNK_1;
+	u8 tWTR;
 	u8 tCL;
-	u8 tRP;		/* Byte 3 */
+	u8 tRC;
 	u8 empty_4;
-	u8 tRAS;	/* Byte 5 */
+	u8 tRFC;	/* Byte 5 */
 	u8 empty_6;
-	u8 tRFC;	/* Byte 7 */
+	u8 tRAS;	/* Byte 7 */
 	u8 empty_8;
-	u8 tRC;		/* Byte 9 */
-	u8 tUNK_10, tUNK_11, tUNK_12, tUNK_13, tUNK_14;
-	u8 empty_15,empty_16,empty_17;
-	u8 tUNK_18, tUNK_19, tUNK_20, tUNK_21;
+	u8 tRP;		/* Byte 9 */
+	u8 tRCDRD;
+	u8 tRCDWR;
+	u8 tRRD;
+	u8 tUNK_13;
+	u8 RAM_FT1;		/* 14, a bitmask of random RAM features */
+	u8 empty_15;
+	u8 tUNK_16;
+	u8 empty_17;
+	u8 tUNK_18;
+	u8 tCWL;
+	u8 tUNK_20, tUNK_21;
 };
 
-/* nouveau_mem.c */
-void nv30_mem_timing_entry(struct drm_device *dev, struct nouveau_pm_tbl_header *hdr,
-							struct nouveau_pm_tbl_entry *e, uint8_t magic_number,
-							struct nouveau_pm_memtiming *timing);
+struct nouveau_pm_profile;
+struct nouveau_pm_profile_func {
+	void (*destroy)(struct nouveau_pm_profile *);
+	void (*init)(struct nouveau_pm_profile *);
+	void (*fini)(struct nouveau_pm_profile *);
+	struct nouveau_pm_level *(*select)(struct nouveau_pm_profile *);
+};
+
+struct nouveau_pm_profile {
+	const struct nouveau_pm_profile_func *func;
+	struct list_head head;
+	char name[8];
+};
 
 #define NOUVEAU_PM_MAX_LEVEL 8
 struct nouveau_pm_level {
+	struct nouveau_pm_profile profile;
 	struct device_attribute dev_attr;
 	char name[32];
 	int id;
 
-	u32 core;
+	struct nouveau_pm_memtiming timing;
 	u32 memory;
+	u16 memscript;
+
+	u32 core;
 	u32 shader;
 	u32 rop;
 	u32 copy;
@@ -498,9 +526,6 @@ struct nouveau_pm_level {
 	u32 volt_min; /* microvolts */
 	u32 volt_max;
 	u8  fanspeed;
-
-	u16 memscript;
-	struct nouveau_pm_memtiming *timing;
 };
 
 struct nouveau_pm_temp_sensor_constants {
@@ -517,27 +542,26 @@ struct nouveau_pm_threshold_temp {
 	s16 fan_boost;
 };
 
-struct nouveau_pm_memtimings {
-	bool supported;
-	struct nouveau_pm_memtiming *timing;
-	int nr_timing;
-};
-
 struct nouveau_pm_fan {
+	u32 percent;
 	u32 min_duty;
 	u32 max_duty;
 	u32 pwm_freq;
+	u32 pwm_divisor;
 };
 
 struct nouveau_pm_engine {
 	struct nouveau_pm_voltage voltage;
 	struct nouveau_pm_level perflvl[NOUVEAU_PM_MAX_LEVEL];
 	int nr_perflvl;
-	struct nouveau_pm_memtimings memtimings;
 	struct nouveau_pm_temp_sensor_constants sensor_constants;
 	struct nouveau_pm_threshold_temp threshold_temp;
 	struct nouveau_pm_fan fan;
-	u32 pwm_divisor;
+
+	struct nouveau_pm_profile *profile_ac;
+	struct nouveau_pm_profile *profile_dc;
+	struct nouveau_pm_profile *profile;
+	struct list_head profiles;
 
 	struct nouveau_pm_level boot;
 	struct nouveau_pm_level *cur;
@@ -669,14 +693,15 @@ struct nv04_mode_state {
 };
 
 enum nouveau_card_type {
-	NV_04      = 0x00,
+	NV_04      = 0x04,
 	NV_10      = 0x10,
 	NV_20      = 0x20,
 	NV_30      = 0x30,
 	NV_40      = 0x40,
 	NV_50      = 0x50,
 	NV_C0      = 0xc0,
-	NV_D0      = 0xd0
+	NV_D0      = 0xd0,
+	NV_E0      = 0xe0,
 };
 
 struct drm_nouveau_private {
@@ -772,8 +797,22 @@ struct drm_nouveau_private {
 	} tile;
 
 	/* VRAM/fb configuration */
+	enum {
+		NV_MEM_TYPE_UNKNOWN = 0,
+		NV_MEM_TYPE_STOLEN,
+		NV_MEM_TYPE_SGRAM,
+		NV_MEM_TYPE_SDRAM,
+		NV_MEM_TYPE_DDR1,
+		NV_MEM_TYPE_DDR2,
+		NV_MEM_TYPE_DDR3,
+		NV_MEM_TYPE_GDDR2,
+		NV_MEM_TYPE_GDDR3,
+		NV_MEM_TYPE_GDDR4,
+		NV_MEM_TYPE_GDDR5
+	} vram_type;
 	uint64_t vram_size;
 	uint64_t vram_sys_base;
+	bool vram_rank_B;
 
 	uint64_t fb_available_size;
 	uint64_t fb_mappable_pages;
@@ -846,6 +885,7 @@ extern int nouveau_uscript_lvds;
 extern int nouveau_uscript_tmds;
 extern int nouveau_vram_pushbuf;
 extern int nouveau_vram_notify;
+extern char *nouveau_vram_type;
 extern int nouveau_fbpercrtc;
 extern int nouveau_tv_disable;
 extern char *nouveau_tv_norm;
@@ -894,8 +934,12 @@ extern void nouveau_mem_gart_fini(struct drm_device *);
 extern int  nouveau_mem_init_agp(struct drm_device *);
 extern int  nouveau_mem_reset_agp(struct drm_device *);
 extern void nouveau_mem_close(struct drm_device *);
-extern int  nouveau_mem_detect(struct drm_device *);
 extern bool nouveau_mem_flags_valid(struct drm_device *, u32 tile_flags);
+extern int  nouveau_mem_timing_calc(struct drm_device *, u32 freq,
+				    struct nouveau_pm_memtiming *);
+extern void nouveau_mem_timing_read(struct drm_device *,
+				    struct nouveau_pm_memtiming *);
+extern int nouveau_mem_vbios_type(struct drm_device *);
 extern struct nouveau_tile_reg *nv10_mem_set_tiling(
 	struct drm_device *dev, uint32_t addr, uint32_t size,
 	uint32_t pitch, uint32_t flags);
@@ -1046,8 +1090,7 @@ nouveau_debugfs_channel_fini(struct nouveau_channel *chan)
 #endif
 
 /* nouveau_dma.c */
-extern void nouveau_dma_pre_init(struct nouveau_channel *);
-extern int  nouveau_dma_init(struct nouveau_channel *);
+extern void nouveau_dma_init(struct nouveau_channel *);
 extern int  nouveau_dma_wait(struct nouveau_channel *, int slots, int size);
 
 /* nouveau_acpi.c */
@@ -1117,19 +1160,14 @@ int nouveau_ttm_mmap(struct file *, struct vm_area_struct *);
 /* nouveau_hdmi.c */
 void nouveau_hdmi_mode_set(struct drm_encoder *, struct drm_display_mode *);
 
-/* nouveau_dp.c */
-int nouveau_dp_auxch(struct nouveau_i2c_chan *auxch, int cmd, int addr,
-		     uint8_t *data, int data_nr);
-bool nouveau_dp_detect(struct drm_encoder *);
-bool nouveau_dp_link_train(struct drm_encoder *, u32 datarate);
-void nouveau_dp_tu_update(struct drm_device *, int, int, u32, u32);
-u8 *nouveau_dp_bios_data(struct drm_device *, struct dcb_entry *, u8 **);
-
 /* nv04_fb.c */
+extern int  nv04_fb_vram_init(struct drm_device *);
 extern int  nv04_fb_init(struct drm_device *);
 extern void nv04_fb_takedown(struct drm_device *);
 
 /* nv10_fb.c */
+extern int  nv10_fb_vram_init(struct drm_device *dev);
+extern int  nv1a_fb_vram_init(struct drm_device *dev);
 extern int  nv10_fb_init(struct drm_device *);
 extern void nv10_fb_takedown(struct drm_device *);
 extern void nv10_fb_init_tile_region(struct drm_device *dev, int i,
@@ -1138,6 +1176,16 @@ extern void nv10_fb_init_tile_region(struct drm_device *dev, int i,
 extern void nv10_fb_set_tile_region(struct drm_device *dev, int i);
 extern void nv10_fb_free_tile_region(struct drm_device *dev, int i);
 
+/* nv20_fb.c */
+extern int  nv20_fb_vram_init(struct drm_device *dev);
+extern int  nv20_fb_init(struct drm_device *);
+extern void nv20_fb_takedown(struct drm_device *);
+extern void nv20_fb_init_tile_region(struct drm_device *dev, int i,
+				     uint32_t addr, uint32_t size,
+				     uint32_t pitch, uint32_t flags);
+extern void nv20_fb_set_tile_region(struct drm_device *dev, int i);
+extern void nv20_fb_free_tile_region(struct drm_device *dev, int i);
+
 /* nv30_fb.c */
 extern int  nv30_fb_init(struct drm_device *);
 extern void nv30_fb_takedown(struct drm_device *);
@@ -1147,6 +1195,7 @@ extern void nv30_fb_init_tile_region(struct drm_device *dev, int i,
 extern void nv30_fb_free_tile_region(struct drm_device *dev, int i);
 
 /* nv40_fb.c */
+extern int  nv40_fb_vram_init(struct drm_device *dev);
 extern int  nv40_fb_init(struct drm_device *);
 extern void nv40_fb_takedown(struct drm_device *);
 extern void nv40_fb_set_tile_region(struct drm_device *dev, int i);
@@ -1703,6 +1752,7 @@ nv44_graph_class(struct drm_device *dev)
 #define NV_MEM_ACCESS_RW (NV_MEM_ACCESS_RO | NV_MEM_ACCESS_WO)
 #define NV_MEM_ACCESS_SYS 4
 #define NV_MEM_ACCESS_VM  8
+#define NV_MEM_ACCESS_NOSNOOP 16
 
 #define NV_MEM_TARGET_VRAM        0
 #define NV_MEM_TARGET_PCI         1
@@ -1713,13 +1763,27 @@ nv44_graph_class(struct drm_device *dev)
 #define NV_MEM_TYPE_VM 0x7f
 #define NV_MEM_COMP_VM 0x03
 
+/* FIFO methods */
+#define NV01_SUBCHAN_OBJECT                                          0x00000000
+#define NV84_SUBCHAN_SEMAPHORE_ADDRESS_HIGH                          0x00000010
+#define NV84_SUBCHAN_SEMAPHORE_ADDRESS_LOW                           0x00000014
+#define NV84_SUBCHAN_SEMAPHORE_SEQUENCE                              0x00000018
+#define NV84_SUBCHAN_SEMAPHORE_TRIGGER                               0x0000001c
+#define NV84_SUBCHAN_SEMAPHORE_TRIGGER_ACQUIRE_EQUAL                 0x00000001
+#define NV84_SUBCHAN_SEMAPHORE_TRIGGER_WRITE_LONG                    0x00000002
+#define NV84_SUBCHAN_SEMAPHORE_TRIGGER_ACQUIRE_GEQUAL                0x00000004
+#define NV84_SUBCHAN_NOTIFY_INTR                                     0x00000020
+#define NV84_SUBCHAN_WRCACHE_FLUSH                                   0x00000024
+#define NV10_SUBCHAN_REF_CNT                                         0x00000050
+#define NVSW_SUBCHAN_PAGE_FLIP                                       0x00000054
+#define NV11_SUBCHAN_DMA_SEMAPHORE                                   0x00000060
+#define NV11_SUBCHAN_SEMAPHORE_OFFSET                                0x00000064
+#define NV11_SUBCHAN_SEMAPHORE_ACQUIRE                               0x00000068
+#define NV11_SUBCHAN_SEMAPHORE_RELEASE                               0x0000006c
+#define NV40_SUBCHAN_YIELD                                           0x00000080
+
 /* NV_SW object class */
 #define NV_SW                                                        0x0000506e
-#define NV_SW_DMA_SEMAPHORE                                          0x00000060
-#define NV_SW_SEMAPHORE_OFFSET                                       0x00000064
-#define NV_SW_SEMAPHORE_ACQUIRE                                      0x00000068
-#define NV_SW_SEMAPHORE_RELEASE                                      0x0000006c
-#define NV_SW_YIELD                                                  0x00000080
 #define NV_SW_DMA_VBLSEM                                             0x0000018c
 #define NV_SW_VBLSEM_OFFSET                                          0x00000400
 #define NV_SW_VBLSEM_RELEASE_VALUE                                   0x00000404
diff --git a/drivers/gpu/drm/nouveau/nouveau_encoder.h b/drivers/gpu/drm/nouveau/nouveau_encoder.h
index e5d6e3faff3d..3dc14a3dcc4c 100644
--- a/drivers/gpu/drm/nouveau/nouveau_encoder.h
+++ b/drivers/gpu/drm/nouveau/nouveau_encoder.h
@@ -32,6 +32,14 @@
 
 #define NV_DPMS_CLEARED 0x80
 
+struct dp_train_func {
+	void (*link_set)(struct drm_device *, struct dcb_entry *, int crtc,
+			 int nr, u32 bw, bool enhframe);
+	void (*train_set)(struct drm_device *, struct dcb_entry *, u8 pattern);
+	void (*train_adj)(struct drm_device *, struct dcb_entry *,
+			  u8 lane, u8 swing, u8 preem);
+};
+
 struct nouveau_encoder {
 	struct drm_encoder_slave base;
 
@@ -78,9 +86,19 @@ get_slave_funcs(struct drm_encoder *enc)
 	return to_encoder_slave(enc)->slave_funcs;
 }
 
+/* nouveau_dp.c */
+int nouveau_dp_auxch(struct nouveau_i2c_chan *auxch, int cmd, int addr,
+		     uint8_t *data, int data_nr);
+bool nouveau_dp_detect(struct drm_encoder *);
+void nouveau_dp_dpms(struct drm_encoder *, int mode, u32 datarate,
+		     struct dp_train_func *);
+u8 *nouveau_dp_bios_data(struct drm_device *, struct dcb_entry *, u8 **);
+
 struct nouveau_connector *
 nouveau_encoder_connector_get(struct nouveau_encoder *encoder);
 int nv50_sor_create(struct drm_connector *, struct dcb_entry *);
+void nv50_sor_dp_calc_tu(struct drm_device *, int, int, u32, u32);
 int nv50_dac_create(struct drm_connector *, struct dcb_entry *);
 
+
 #endif /* __NOUVEAU_ENCODER_H__ */
diff --git a/drivers/gpu/drm/nouveau/nouveau_fbcon.c b/drivers/gpu/drm/nouveau/nouveau_fbcon.c
index 9892218d7452..8113e9201ed9 100644
--- a/drivers/gpu/drm/nouveau/nouveau_fbcon.c
+++ b/drivers/gpu/drm/nouveau/nouveau_fbcon.c
@@ -381,11 +381,7 @@ nouveau_fbcon_create(struct nouveau_fbdev *nfbdev,
 		goto out_unref;
 	}
 
-	info->pixmap.size = 64*1024;
-	info->pixmap.buf_align = 8;
-	info->pixmap.access_align = 32;
-	info->pixmap.flags = FB_PIXMAP_SYSTEM;
-	info->pixmap.scan_align = 1;
+	/* Use default scratch pixmap (info->pixmap.flags = FB_PIXMAP_SYSTEM) */
 
 	mutex_unlock(&dev->struct_mutex);
 
diff --git a/drivers/gpu/drm/nouveau/nouveau_fence.c b/drivers/gpu/drm/nouveau/nouveau_fence.c
index 2f6daae68b9d..c1dc20f6cb85 100644
--- a/drivers/gpu/drm/nouveau/nouveau_fence.c
+++ b/drivers/gpu/drm/nouveau/nouveau_fence.c
@@ -93,18 +93,17 @@ nouveau_fence_update(struct nouveau_channel *chan)
 	}
 
 	list_for_each_entry_safe(fence, tmp, &chan->fence.pending, entry) {
-		sequence = fence->sequence;
+		if (fence->sequence > chan->fence.sequence_ack)
+			break;
+
 		fence->signalled = true;
 		list_del(&fence->entry);
-
-		if (unlikely(fence->work))
+		if (fence->work)
 			fence->work(fence->priv, true);
 
 		kref_put(&fence->refcount, nouveau_fence_del);
-
-		if (sequence == chan->fence.sequence_ack)
-			break;
 	}
+
 out:
 	spin_unlock(&chan->fence.lock);
 }
@@ -165,9 +164,9 @@ nouveau_fence_emit(struct nouveau_fence *fence)
 
 	if (USE_REFCNT(dev)) {
 		if (dev_priv->card_type < NV_C0)
-			BEGIN_RING(chan, NvSubSw, 0x0050, 1);
+			BEGIN_RING(chan, 0, NV10_SUBCHAN_REF_CNT, 1);
 		else
-			BEGIN_NVC0(chan, 2, NvSubM2MF, 0x0050, 1);
+			BEGIN_NVC0(chan, 2, 0, NV10_SUBCHAN_REF_CNT, 1);
 	} else {
 		BEGIN_RING(chan, NvSubSw, 0x0150, 1);
 	}
@@ -344,7 +343,7 @@ semaphore_acquire(struct nouveau_channel *chan, struct nouveau_semaphore *sema)
 		if (ret)
 			return ret;
 
-		BEGIN_RING(chan, NvSubSw, NV_SW_DMA_SEMAPHORE, 3);
+		BEGIN_RING(chan, 0, NV11_SUBCHAN_DMA_SEMAPHORE, 3);
 		OUT_RING  (chan, NvSema);
 		OUT_RING  (chan, offset);
 		OUT_RING  (chan, 1);
@@ -354,9 +353,9 @@ semaphore_acquire(struct nouveau_channel *chan, struct nouveau_semaphore *sema)
 		if (ret)
 			return ret;
 
-		BEGIN_RING(chan, NvSubSw, NV_SW_DMA_SEMAPHORE, 1);
+		BEGIN_RING(chan, 0, NV11_SUBCHAN_DMA_SEMAPHORE, 1);
 		OUT_RING  (chan, chan->vram_handle);
-		BEGIN_RING(chan, NvSubSw, 0x0010, 4);
+		BEGIN_RING(chan, 0, NV84_SUBCHAN_SEMAPHORE_ADDRESS_HIGH, 4);
 		OUT_RING  (chan, upper_32_bits(offset));
 		OUT_RING  (chan, lower_32_bits(offset));
 		OUT_RING  (chan, 1);
@@ -366,7 +365,7 @@ semaphore_acquire(struct nouveau_channel *chan, struct nouveau_semaphore *sema)
 		if (ret)
 			return ret;
 
-		BEGIN_NVC0(chan, 2, NvSubM2MF, 0x0010, 4);
+		BEGIN_NVC0(chan, 2, 0, NV84_SUBCHAN_SEMAPHORE_ADDRESS_HIGH, 4);
 		OUT_RING  (chan, upper_32_bits(offset));
 		OUT_RING  (chan, lower_32_bits(offset));
 		OUT_RING  (chan, 1);
@@ -397,10 +396,10 @@ semaphore_release(struct nouveau_channel *chan, struct nouveau_semaphore *sema)
 		if (ret)
 			return ret;
 
-		BEGIN_RING(chan, NvSubSw, NV_SW_DMA_SEMAPHORE, 2);
+		BEGIN_RING(chan, 0, NV11_SUBCHAN_DMA_SEMAPHORE, 2);
 		OUT_RING  (chan, NvSema);
 		OUT_RING  (chan, offset);
-		BEGIN_RING(chan, NvSubSw, NV_SW_SEMAPHORE_RELEASE, 1);
+		BEGIN_RING(chan, 0, NV11_SUBCHAN_SEMAPHORE_RELEASE, 1);
 		OUT_RING  (chan, 1);
 	} else
 	if (dev_priv->chipset < 0xc0) {
@@ -408,9 +407,9 @@ semaphore_release(struct nouveau_channel *chan, struct nouveau_semaphore *sema)
 		if (ret)
 			return ret;
 
-		BEGIN_RING(chan, NvSubSw, NV_SW_DMA_SEMAPHORE, 1);
+		BEGIN_RING(chan, 0, NV11_SUBCHAN_DMA_SEMAPHORE, 1);
 		OUT_RING  (chan, chan->vram_handle);
-		BEGIN_RING(chan, NvSubSw, 0x0010, 4);
+		BEGIN_RING(chan, 0, NV84_SUBCHAN_SEMAPHORE_ADDRESS_HIGH, 4);
 		OUT_RING  (chan, upper_32_bits(offset));
 		OUT_RING  (chan, lower_32_bits(offset));
 		OUT_RING  (chan, 1);
@@ -420,7 +419,7 @@ semaphore_release(struct nouveau_channel *chan, struct nouveau_semaphore *sema)
 		if (ret)
 			return ret;
 
-		BEGIN_NVC0(chan, 2, NvSubM2MF, 0x0010, 4);
+		BEGIN_NVC0(chan, 2, 0, NV84_SUBCHAN_SEMAPHORE_ADDRESS_HIGH, 4);
 		OUT_RING  (chan, upper_32_bits(offset));
 		OUT_RING  (chan, lower_32_bits(offset));
 		OUT_RING  (chan, 1);
@@ -510,7 +509,7 @@ nouveau_fence_channel_init(struct nouveau_channel *chan)
 		if (ret)
 			return ret;
 
-		BEGIN_RING(chan, NvSubSw, 0, 1);
+		BEGIN_RING(chan, NvSubSw, NV01_SUBCHAN_OBJECT, 1);
 		OUT_RING  (chan, NvSw);
 		FIRE_RING (chan);
 	}
diff --git a/drivers/gpu/drm/nouveau/nouveau_gem.c b/drivers/gpu/drm/nouveau/nouveau_gem.c
index 7ce3fde40743..ed52a6f41613 100644
--- a/drivers/gpu/drm/nouveau/nouveau_gem.c
+++ b/drivers/gpu/drm/nouveau/nouveau_gem.c
@@ -426,9 +426,7 @@ validate_list(struct nouveau_channel *chan, struct list_head *list,
 			return ret;
 		}
 
-		nvbo->channel = (b->read_domains & (1 << 31)) ? NULL : chan;
 		ret = nouveau_bo_validate(nvbo, true, false, false);
-		nvbo->channel = NULL;
 		if (unlikely(ret)) {
 			if (ret != -ERESTARTSYS)
 				NV_ERROR(dev, "fail ttm_validate\n");
@@ -678,19 +676,13 @@ nouveau_gem_ioctl_pushbuf(struct drm_device *dev, void *data,
 		return PTR_ERR(bo);
 	}
 
-	/* Mark push buffers as being used on PFIFO, the validation code
-	 * will then make sure that if the pushbuf bo moves, that they
-	 * happen on the kernel channel, which will in turn cause a sync
-	 * to happen before we try and submit the push buffer.
-	 */
+	/* Ensure all push buffers are on validate list */
 	for (i = 0; i < req->nr_push; i++) {
 		if (push[i].bo_index >= req->nr_buffers) {
 			NV_ERROR(dev, "push %d buffer not in list\n", i);
 			ret = -EINVAL;
 			goto out_prevalid;
 		}
-
-		bo[push[i].bo_index].read_domains |= (1 << 31);
 	}
 
 	/* Validate buffer list */
diff --git a/drivers/gpu/drm/nouveau/nouveau_i2c.c b/drivers/gpu/drm/nouveau/nouveau_i2c.c
index 820ae7f52044..e2be95af2e52 100644
--- a/drivers/gpu/drm/nouveau/nouveau_i2c.c
+++ b/drivers/gpu/drm/nouveau/nouveau_i2c.c
@@ -277,7 +277,7 @@ i2c_bit_func(struct i2c_adapter *adap)
 	return I2C_FUNC_I2C | I2C_FUNC_SMBUS_EMUL;
 }
 
-const struct i2c_algorithm i2c_bit_algo = {
+const struct i2c_algorithm nouveau_i2c_bit_algo = {
 	.master_xfer = i2c_bit_xfer,
 	.functionality = i2c_bit_func
 };
@@ -315,8 +315,8 @@ nouveau_i2c_init(struct drm_device *dev)
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
 	struct nvbios *bios = &dev_priv->vbios;
 	struct nouveau_i2c_chan *port;
+	u8 version = 0x00, entries, recordlen;
 	u8 *i2c, *entry, legacy[2][4] = {};
-	u8 version, entries, recordlen;
 	int ret, i;
 
 	INIT_LIST_HEAD(&dev_priv->i2c_ports);
@@ -346,12 +346,12 @@ nouveau_i2c_init(struct drm_device *dev)
 		if (i2c[7]) legacy[1][1] = i2c[7];
 	}
 
-	if (i2c && version >= 0x30) {
+	if (version >= 0x30) {
 		entry     = i2c[1] + i2c;
 		entries   = i2c[2];
 		recordlen = i2c[3];
 	} else
-	if (i2c) {
+	if (version) {
 		entry     = i2c;
 		entries   = 16;
 		recordlen = 4;
@@ -384,12 +384,12 @@ nouveau_i2c_init(struct drm_device *dev)
 		case 0: /* NV04:NV50 */
 			port->drive = entry[0];
 			port->sense = entry[1];
-			port->adapter.algo = &i2c_bit_algo;
+			port->adapter.algo = &nouveau_i2c_bit_algo;
 			break;
 		case 4: /* NV4E */
 			port->drive = 0x600800 + entry[1];
 			port->sense = port->drive;
-			port->adapter.algo = &i2c_bit_algo;
+			port->adapter.algo = &nouveau_i2c_bit_algo;
 			break;
 		case 5: /* NV50- */
 			port->drive = entry[0] & 0x0f;
@@ -402,7 +402,7 @@ nouveau_i2c_init(struct drm_device *dev)
 				port->drive = 0x00d014 + (port->drive * 0x20);
 				port->sense = port->drive;
 			}
-			port->adapter.algo = &i2c_bit_algo;
+			port->adapter.algo = &nouveau_i2c_bit_algo;
 			break;
 		case 6: /* NV50- DP AUX */
 			port->drive = entry[0];
diff --git a/drivers/gpu/drm/nouveau/nouveau_mem.c b/drivers/gpu/drm/nouveau/nouveau_mem.c
index c3a5745e9c79..b08065f981df 100644
--- a/drivers/gpu/drm/nouveau/nouveau_mem.c
+++ b/drivers/gpu/drm/nouveau/nouveau_mem.c
@@ -26,7 +26,8 @@
  * DEALINGS IN THE SOFTWARE.
  *
  * Authors:
- *    Keith Whitwell <keith@tungstengraphics.com>
+ *    Ben Skeggs <bskeggs@redhat.com>
+ *    Roy Spliet <r.spliet@student.tudelft.nl>
  */
 
 
@@ -192,75 +193,6 @@ nouveau_mem_gart_fini(struct drm_device *dev)
 	}
 }
 
-static uint32_t
-nouveau_mem_detect_nv04(struct drm_device *dev)
-{
-	uint32_t boot0 = nv_rd32(dev, NV04_PFB_BOOT_0);
-
-	if (boot0 & 0x00000100)
-		return (((boot0 >> 12) & 0xf) * 2 + 2) * 1024 * 1024;
-
-	switch (boot0 & NV04_PFB_BOOT_0_RAM_AMOUNT) {
-	case NV04_PFB_BOOT_0_RAM_AMOUNT_32MB:
-		return 32 * 1024 * 1024;
-	case NV04_PFB_BOOT_0_RAM_AMOUNT_16MB:
-		return 16 * 1024 * 1024;
-	case NV04_PFB_BOOT_0_RAM_AMOUNT_8MB:
-		return 8 * 1024 * 1024;
-	case NV04_PFB_BOOT_0_RAM_AMOUNT_4MB:
-		return 4 * 1024 * 1024;
-	}
-
-	return 0;
-}
-
-static uint32_t
-nouveau_mem_detect_nforce(struct drm_device *dev)
-{
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	struct pci_dev *bridge;
-	uint32_t mem;
-
-	bridge = pci_get_bus_and_slot(0, PCI_DEVFN(0, 1));
-	if (!bridge) {
-		NV_ERROR(dev, "no bridge device\n");
-		return 0;
-	}
-
-	if (dev_priv->flags & NV_NFORCE) {
-		pci_read_config_dword(bridge, 0x7C, &mem);
-		return (uint64_t)(((mem >> 6) & 31) + 1)*1024*1024;
-	} else
-	if (dev_priv->flags & NV_NFORCE2) {
-		pci_read_config_dword(bridge, 0x84, &mem);
-		return (uint64_t)(((mem >> 4) & 127) + 1)*1024*1024;
-	}
-
-	NV_ERROR(dev, "impossible!\n");
-	return 0;
-}
-
-int
-nouveau_mem_detect(struct drm_device *dev)
-{
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-
-	if (dev_priv->card_type == NV_04) {
-		dev_priv->vram_size = nouveau_mem_detect_nv04(dev);
-	} else
-	if (dev_priv->flags & (NV_NFORCE | NV_NFORCE2)) {
-		dev_priv->vram_size = nouveau_mem_detect_nforce(dev);
-	} else
-	if (dev_priv->card_type < NV_50) {
-		dev_priv->vram_size  = nv_rd32(dev, NV04_PFB_FIFO_DATA);
-		dev_priv->vram_size &= NV10_PFB_FIFO_DATA_RAM_AMOUNT_MB_MASK;
-	}
-
-	if (dev_priv->vram_size)
-		return 0;
-	return -ENOMEM;
-}
-
 bool
 nouveau_mem_flags_valid(struct drm_device *dev, u32 tile_flags)
 {
@@ -385,11 +317,29 @@ nouveau_mem_init_agp(struct drm_device *dev)
 	return 0;
 }
 
+static const struct vram_types {
+	int value;
+	const char *name;
+} vram_type_map[] = {
+	{ NV_MEM_TYPE_STOLEN , "stolen system memory" },
+	{ NV_MEM_TYPE_SGRAM  , "SGRAM" },
+	{ NV_MEM_TYPE_SDRAM  , "SDRAM" },
+	{ NV_MEM_TYPE_DDR1   , "DDR1" },
+	{ NV_MEM_TYPE_DDR2   , "DDR2" },
+	{ NV_MEM_TYPE_DDR3   , "DDR3" },
+	{ NV_MEM_TYPE_GDDR2  , "GDDR2" },
+	{ NV_MEM_TYPE_GDDR3  , "GDDR3" },
+	{ NV_MEM_TYPE_GDDR4  , "GDDR4" },
+	{ NV_MEM_TYPE_GDDR5  , "GDDR5" },
+	{ NV_MEM_TYPE_UNKNOWN, "unknown type" }
+};
+
 int
 nouveau_mem_vram_init(struct drm_device *dev)
 {
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
 	struct ttm_bo_device *bdev = &dev_priv->ttm.bdev;
+	const struct vram_types *vram_type;
 	int ret, dma_bits;
 
 	dma_bits = 32;
@@ -427,7 +377,21 @@ nouveau_mem_vram_init(struct drm_device *dev)
 		return ret;
 	}
 
-	NV_INFO(dev, "Detected %dMiB VRAM\n", (int)(dev_priv->vram_size >> 20));
+	vram_type = vram_type_map;
+	while (vram_type->value != NV_MEM_TYPE_UNKNOWN) {
+		if (nouveau_vram_type) {
+			if (!strcasecmp(nouveau_vram_type, vram_type->name))
+				break;
+			dev_priv->vram_type = vram_type->value;
+		} else {
+			if (vram_type->value == dev_priv->vram_type)
+				break;
+		}
+		vram_type++;
+	}
+
+	NV_INFO(dev, "Detected %dMiB VRAM (%s)\n",
+		(int)(dev_priv->vram_size >> 20), vram_type->name);
 	if (dev_priv->vram_sys_base) {
 		NV_INFO(dev, "Stolen system memory at: 0x%010llx\n",
 			dev_priv->vram_sys_base);
@@ -508,216 +472,617 @@ nouveau_mem_gart_init(struct drm_device *dev)
 	return 0;
 }
 
-/* XXX: For now a dummy. More samples required, possibly even a card
- * Called from nouveau_perf.c */
-void nv30_mem_timing_entry(struct drm_device *dev, struct nouveau_pm_tbl_header *hdr,
-							struct nouveau_pm_tbl_entry *e, uint8_t magic_number,
-							struct nouveau_pm_memtiming *timing) {
-
-	NV_DEBUG(dev,"Timing entry format unknown, please contact nouveau developers");
-}
-
-void nv40_mem_timing_entry(struct drm_device *dev, struct nouveau_pm_tbl_header *hdr,
-							struct nouveau_pm_tbl_entry *e, uint8_t magic_number,
-							struct nouveau_pm_memtiming *timing) {
-
-	timing->reg_0 = (e->tRC << 24 | e->tRFC << 16 | e->tRAS << 8 | e->tRP);
+static int
+nv40_mem_timing_calc(struct drm_device *dev, u32 freq,
+		     struct nouveau_pm_tbl_entry *e, u8 len,
+		     struct nouveau_pm_memtiming *boot,
+		     struct nouveau_pm_memtiming *t)
+{
+	t->reg[0] = (e->tRP << 24 | e->tRAS << 16 | e->tRFC << 8 | e->tRC);
 
 	/* XXX: I don't trust the -1's and +1's... they must come
 	 *      from somewhere! */
-	timing->reg_1 = (e->tWR + 2 + magic_number) << 24 |
-				  1 << 16 |
-				  (e->tUNK_1 + 2 + magic_number) << 8 |
-				  (e->tCL + 2 - magic_number);
-	timing->reg_2 = (magic_number << 24 | e->tUNK_12 << 16 | e->tUNK_11 << 8 | e->tUNK_10);
-	timing->reg_2 |= 0x20200000;
-
-	NV_DEBUG(dev, "Entry %d: 220: %08x %08x %08x\n", timing->id,
-		 timing->reg_0, timing->reg_1,timing->reg_2);
+	t->reg[1] = (e->tWR + 2 + (t->tCWL - 1)) << 24 |
+		    1 << 16 |
+		    (e->tWTR + 2 + (t->tCWL - 1)) << 8 |
+		    (e->tCL + 2 - (t->tCWL - 1));
+
+	t->reg[2] = 0x20200000 |
+		    ((t->tCWL - 1) << 24 |
+		     e->tRRD << 16 |
+		     e->tRCDWR << 8 |
+		     e->tRCDRD);
+
+	NV_DEBUG(dev, "Entry %d: 220: %08x %08x %08x\n", t->id,
+		 t->reg[0], t->reg[1], t->reg[2]);
+	return 0;
 }
 
-void nv50_mem_timing_entry(struct drm_device *dev, struct bit_entry *P, struct nouveau_pm_tbl_header *hdr,
-							struct nouveau_pm_tbl_entry *e, uint8_t magic_number,struct nouveau_pm_memtiming *timing) {
+static int
+nv50_mem_timing_calc(struct drm_device *dev, u32 freq,
+		     struct nouveau_pm_tbl_entry *e, u8 len,
+		     struct nouveau_pm_memtiming *boot,
+		     struct nouveau_pm_memtiming *t)
+{
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
+	struct bit_entry P;
+	uint8_t unk18 = 1, unk20 = 0, unk21 = 0, tmp7_3;
 
-	uint8_t unk18 = 1,
-		unk19 = 1,
-		unk20 = 0,
-		unk21 = 0;
+	if (bit_table(dev, 'P', &P))
+		return -EINVAL;
 
-	switch (min(hdr->entry_len, (u8) 22)) {
+	switch (min(len, (u8) 22)) {
 	case 22:
 		unk21 = e->tUNK_21;
 	case 21:
 		unk20 = e->tUNK_20;
 	case 20:
-		unk19 = e->tUNK_19;
+		if (e->tCWL > 0)
+			t->tCWL = e->tCWL;
 	case 19:
 		unk18 = e->tUNK_18;
 		break;
 	}
 
-	timing->reg_0 = (e->tRC << 24 | e->tRFC << 16 | e->tRAS << 8 | e->tRP);
+	t->reg[0] = (e->tRP << 24 | e->tRAS << 16 | e->tRFC << 8 | e->tRC);
 
-	/* XXX: I don't trust the -1's and +1's... they must come
-	 *      from somewhere! */
-	timing->reg_1 = (e->tWR + unk19 + 1 + magic_number) << 24 |
-				  max(unk18, (u8) 1) << 16 |
-				  (e->tUNK_1 + unk19 + 1 + magic_number) << 8;
-	if (dev_priv->chipset == 0xa8) {
-		timing->reg_1 |= (e->tCL - 1);
-	} else {
-		timing->reg_1 |= (e->tCL + 2 - magic_number);
-	}
-	timing->reg_2 = (e->tUNK_12 << 16 | e->tUNK_11 << 8 | e->tUNK_10);
-
-	timing->reg_5 = (e->tRAS << 24 | e->tRC);
-	timing->reg_5 += max(e->tUNK_10, e->tUNK_11) << 16;
-
-	if (P->version == 1) {
-		timing->reg_2 |= magic_number << 24;
-		timing->reg_3 = (0x14 + e->tCL) << 24 |
-						0x16 << 16 |
-						(e->tCL - 1) << 8 |
-						(e->tCL - 1);
-		timing->reg_4 = (nv_rd32(dev,0x10022c) & 0xffff0000) | e->tUNK_13 << 8  | e->tUNK_13;
-		timing->reg_5 |= (e->tCL + 2) << 8;
-		timing->reg_7 = 0x4000202 | (e->tCL - 1) << 16;
+	t->reg[1] = (e->tWR + 2 + (t->tCWL - 1)) << 24 |
+				max(unk18, (u8) 1) << 16 |
+				(e->tWTR + 2 + (t->tCWL - 1)) << 8;
+
+	t->reg[2] = ((t->tCWL - 1) << 24 |
+		    e->tRRD << 16 |
+		    e->tRCDWR << 8 |
+		    e->tRCDRD);
+
+	t->reg[4] = e->tUNK_13 << 8  | e->tUNK_13;
+
+	t->reg[5] = (e->tRFC << 24 | max(e->tRCDRD, e->tRCDWR) << 16 | e->tRP);
+
+	t->reg[8] = boot->reg[8] & 0xffffff00;
+
+	if (P.version == 1) {
+		t->reg[1] |= (e->tCL + 2 - (t->tCWL - 1));
+
+		t->reg[3] = (0x14 + e->tCL) << 24 |
+			    0x16 << 16 |
+			    (e->tCL - 1) << 8 |
+			    (e->tCL - 1);
+
+		t->reg[4] |= boot->reg[4] & 0xffff0000;
+
+		t->reg[6] = (0x33 - t->tCWL) << 16 |
+			    t->tCWL << 8 |
+			    (0x2e + e->tCL - t->tCWL);
+
+		t->reg[7] = 0x4000202 | (e->tCL - 1) << 16;
+
+		/* XXX: P.version == 1 only has DDR2 and GDDR3? */
+		if (dev_priv->vram_type == NV_MEM_TYPE_DDR2) {
+			t->reg[5] |= (e->tCL + 3) << 8;
+			t->reg[6] |= (t->tCWL - 2) << 8;
+			t->reg[8] |= (e->tCL - 4);
+		} else {
+			t->reg[5] |= (e->tCL + 2) << 8;
+			t->reg[6] |= t->tCWL << 8;
+			t->reg[8] |= (e->tCL - 2);
+		}
 	} else {
-		timing->reg_2 |= (unk19 - 1) << 24;
-		/* XXX: reg_10022c for recentish cards pretty much unknown*/
-		timing->reg_3 = e->tCL - 1;
-		timing->reg_4 = (unk20 << 24 | unk21 << 16 |
-							e->tUNK_13 << 8  | e->tUNK_13);
+		t->reg[1] |= (5 + e->tCL - (t->tCWL));
+
+		/* XXX: 0xb? 0x30? */
+		t->reg[3] = (0x30 + e->tCL) << 24 |
+			    (boot->reg[3] & 0x00ff0000)|
+			    (0xb + e->tCL) << 8 |
+			    (e->tCL - 1);
+
+		t->reg[4] |= (unk20 << 24 | unk21 << 16);
+
 		/* XXX: +6? */
-		timing->reg_5 |= (unk19 + 6) << 8;
+		t->reg[5] |= (t->tCWL + 6) << 8;
 
-		/* XXX: reg_10023c currently unknown
-		 * 10023c seen as 06xxxxxx, 0bxxxxxx or 0fxxxxxx */
-		timing->reg_7 = 0x202;
+		t->reg[6] = (0x5a + e->tCL) << 16 |
+			    (6 - e->tCL + t->tCWL) << 8 |
+			    (0x50 + e->tCL - t->tCWL);
+
+		tmp7_3 = (boot->reg[7] & 0xff000000) >> 24;
+		t->reg[7] = (tmp7_3 << 24) |
+			    ((tmp7_3 - 6 + e->tCL) << 16) |
+			    0x202;
 	}
 
-	NV_DEBUG(dev, "Entry %d: 220: %08x %08x %08x %08x\n", timing->id,
-		 timing->reg_0, timing->reg_1,
-		 timing->reg_2, timing->reg_3);
+	NV_DEBUG(dev, "Entry %d: 220: %08x %08x %08x %08x\n", t->id,
+		 t->reg[0], t->reg[1], t->reg[2], t->reg[3]);
 	NV_DEBUG(dev, "         230: %08x %08x %08x %08x\n",
-		 timing->reg_4, timing->reg_5,
-		 timing->reg_6, timing->reg_7);
-	NV_DEBUG(dev, "         240: %08x\n", timing->reg_8);
-}
-
-void nvc0_mem_timing_entry(struct drm_device *dev, struct nouveau_pm_tbl_header *hdr,
-							struct nouveau_pm_tbl_entry *e, struct nouveau_pm_memtiming *timing) {
-	timing->reg_0 = (e->tRC << 24 | (e->tRFC & 0x7f) << 17 | e->tRAS << 8 | e->tRP);
-	timing->reg_1 = (nv_rd32(dev,0x10f294) & 0xff000000) | (e->tUNK_11&0x0f) << 20 | (e->tUNK_19 << 7) | (e->tCL & 0x0f);
-	timing->reg_2 = (nv_rd32(dev,0x10f298) & 0xff0000ff) | e->tWR << 16 | e->tUNK_1 << 8;
-	timing->reg_3 = e->tUNK_20 << 9 | e->tUNK_13;
-	timing->reg_4 = (nv_rd32(dev,0x10f2a0) & 0xfff000ff) | e->tUNK_12 << 15;
-	NV_DEBUG(dev, "Entry %d: 290: %08x %08x %08x %08x\n", timing->id,
-		 timing->reg_0, timing->reg_1,
-		 timing->reg_2, timing->reg_3);
-	NV_DEBUG(dev, "         2a0: %08x %08x %08x %08x\n",
-		 timing->reg_4, timing->reg_5,
-		 timing->reg_6, timing->reg_7);
+		 t->reg[4], t->reg[5], t->reg[6], t->reg[7]);
+	NV_DEBUG(dev, "         240: %08x\n", t->reg[8]);
+	return 0;
+}
+
+static int
+nvc0_mem_timing_calc(struct drm_device *dev, u32 freq,
+		     struct nouveau_pm_tbl_entry *e, u8 len,
+		     struct nouveau_pm_memtiming *boot,
+		     struct nouveau_pm_memtiming *t)
+{
+	if (e->tCWL > 0)
+		t->tCWL = e->tCWL;
+
+	t->reg[0] = (e->tRP << 24 | (e->tRAS & 0x7f) << 17 |
+		     e->tRFC << 8 | e->tRC);
+
+	t->reg[1] = (boot->reg[1] & 0xff000000) |
+		    (e->tRCDWR & 0x0f) << 20 |
+		    (e->tRCDRD & 0x0f) << 14 |
+		    (t->tCWL << 7) |
+		    (e->tCL & 0x0f);
+
+	t->reg[2] = (boot->reg[2] & 0xff0000ff) |
+		    e->tWR << 16 | e->tWTR << 8;
+
+	t->reg[3] = (e->tUNK_20 & 0x1f) << 9 |
+		    (e->tUNK_21 & 0xf) << 5 |
+		    (e->tUNK_13 & 0x1f);
+
+	t->reg[4] = (boot->reg[4] & 0xfff00fff) |
+		    (e->tRRD&0x1f) << 15;
+
+	NV_DEBUG(dev, "Entry %d: 290: %08x %08x %08x %08x\n", t->id,
+		 t->reg[0], t->reg[1], t->reg[2], t->reg[3]);
+	NV_DEBUG(dev, "         2a0: %08x\n", t->reg[4]);
+	return 0;
 }
 
 /**
- * Processes the Memory Timing BIOS table, stores generated
- * register values
- * @pre init scripts were run, memtiming regs are initialized
+ * MR generation methods
  */
-void
-nouveau_mem_timing_init(struct drm_device *dev)
+
+static int
+nouveau_mem_ddr2_mr(struct drm_device *dev, u32 freq,
+		    struct nouveau_pm_tbl_entry *e, u8 len,
+		    struct nouveau_pm_memtiming *boot,
+		    struct nouveau_pm_memtiming *t)
+{
+	t->drive_strength = 0;
+	if (len < 15) {
+		t->odt = boot->odt;
+	} else {
+		t->odt = e->RAM_FT1 & 0x07;
+	}
+
+	if (e->tCL >= NV_MEM_CL_DDR2_MAX) {
+		NV_WARN(dev, "(%u) Invalid tCL: %u", t->id, e->tCL);
+		return -ERANGE;
+	}
+
+	if (e->tWR >= NV_MEM_WR_DDR2_MAX) {
+		NV_WARN(dev, "(%u) Invalid tWR: %u", t->id, e->tWR);
+		return -ERANGE;
+	}
+
+	if (t->odt > 3) {
+		NV_WARN(dev, "(%u) Invalid odt value, assuming disabled: %x",
+			t->id, t->odt);
+		t->odt = 0;
+	}
+
+	t->mr[0] = (boot->mr[0] & 0x100f) |
+		   (e->tCL) << 4 |
+		   (e->tWR - 1) << 9;
+	t->mr[1] = (boot->mr[1] & 0x101fbb) |
+		   (t->odt & 0x1) << 2 |
+		   (t->odt & 0x2) << 5;
+
+	NV_DEBUG(dev, "(%u) MR: %08x", t->id, t->mr[0]);
+	return 0;
+}
+
+uint8_t nv_mem_wr_lut_ddr3[NV_MEM_WR_DDR3_MAX] = {
+	0, 0, 0, 0, 0, 1, 2, 3, 4, 5, 5, 6, 6, 7, 7, 0, 0};
+
+static int
+nouveau_mem_ddr3_mr(struct drm_device *dev, u32 freq,
+		    struct nouveau_pm_tbl_entry *e, u8 len,
+		    struct nouveau_pm_memtiming *boot,
+		    struct nouveau_pm_memtiming *t)
+{
+	u8 cl = e->tCL - 4;
+
+	t->drive_strength = 0;
+	if (len < 15) {
+		t->odt = boot->odt;
+	} else {
+		t->odt = e->RAM_FT1 & 0x07;
+	}
+
+	if (e->tCL >= NV_MEM_CL_DDR3_MAX || e->tCL < 4) {
+		NV_WARN(dev, "(%u) Invalid tCL: %u", t->id, e->tCL);
+		return -ERANGE;
+	}
+
+	if (e->tWR >= NV_MEM_WR_DDR3_MAX || e->tWR < 4) {
+		NV_WARN(dev, "(%u) Invalid tWR: %u", t->id, e->tWR);
+		return -ERANGE;
+	}
+
+	if (e->tCWL < 5) {
+		NV_WARN(dev, "(%u) Invalid tCWL: %u", t->id, e->tCWL);
+		return -ERANGE;
+	}
+
+	t->mr[0] = (boot->mr[0] & 0x180b) |
+		   /* CAS */
+		   (cl & 0x7) << 4 |
+		   (cl & 0x8) >> 1 |
+		   (nv_mem_wr_lut_ddr3[e->tWR]) << 9;
+	t->mr[1] = (boot->mr[1] & 0x101dbb) |
+		   (t->odt & 0x1) << 2 |
+		   (t->odt & 0x2) << 5 |
+		   (t->odt & 0x4) << 7;
+	t->mr[2] = (boot->mr[2] & 0x20ffb7) | (e->tCWL - 5) << 3;
+
+	NV_DEBUG(dev, "(%u) MR: %08x %08x", t->id, t->mr[0], t->mr[2]);
+	return 0;
+}
+
+uint8_t nv_mem_cl_lut_gddr3[NV_MEM_CL_GDDR3_MAX] = {
+	0, 0, 0, 0, 4, 5, 6, 7, 0, 1, 2, 3, 8, 9, 10, 11};
+uint8_t nv_mem_wr_lut_gddr3[NV_MEM_WR_GDDR3_MAX] = {
+	0, 0, 0, 0, 0, 2, 3, 8, 9, 10, 11, 0, 0, 1, 1, 0, 3};
+
+static int
+nouveau_mem_gddr3_mr(struct drm_device *dev, u32 freq,
+		     struct nouveau_pm_tbl_entry *e, u8 len,
+		     struct nouveau_pm_memtiming *boot,
+		     struct nouveau_pm_memtiming *t)
+{
+	if (len < 15) {
+		t->drive_strength = boot->drive_strength;
+		t->odt = boot->odt;
+	} else {
+		t->drive_strength = (e->RAM_FT1 & 0x30) >> 4;
+		t->odt = e->RAM_FT1 & 0x07;
+	}
+
+	if (e->tCL >= NV_MEM_CL_GDDR3_MAX) {
+		NV_WARN(dev, "(%u) Invalid tCL: %u", t->id, e->tCL);
+		return -ERANGE;
+	}
+
+	if (e->tWR >= NV_MEM_WR_GDDR3_MAX) {
+		NV_WARN(dev, "(%u) Invalid tWR: %u", t->id, e->tWR);
+		return -ERANGE;
+	}
+
+	if (t->odt > 3) {
+		NV_WARN(dev, "(%u) Invalid odt value, assuming autocal: %x",
+			t->id, t->odt);
+		t->odt = 0;
+	}
+
+	t->mr[0] = (boot->mr[0] & 0xe0b) |
+		   /* CAS */
+		   ((nv_mem_cl_lut_gddr3[e->tCL] & 0x7) << 4) |
+		   ((nv_mem_cl_lut_gddr3[e->tCL] & 0x8) >> 2);
+	t->mr[1] = (boot->mr[1] & 0x100f40) | t->drive_strength |
+		   (t->odt << 2) |
+		   (nv_mem_wr_lut_gddr3[e->tWR] & 0xf) << 4;
+	t->mr[2] = boot->mr[2];
+
+	NV_DEBUG(dev, "(%u) MR: %08x %08x %08x", t->id,
+		      t->mr[0], t->mr[1], t->mr[2]);
+	return 0;
+}
+
+static int
+nouveau_mem_gddr5_mr(struct drm_device *dev, u32 freq,
+		     struct nouveau_pm_tbl_entry *e, u8 len,
+		     struct nouveau_pm_memtiming *boot,
+		     struct nouveau_pm_memtiming *t)
+{
+	if (len < 15) {
+		t->drive_strength = boot->drive_strength;
+		t->odt = boot->odt;
+	} else {
+		t->drive_strength = (e->RAM_FT1 & 0x30) >> 4;
+		t->odt = e->RAM_FT1 & 0x03;
+	}
+
+	if (e->tCL >= NV_MEM_CL_GDDR5_MAX) {
+		NV_WARN(dev, "(%u) Invalid tCL: %u", t->id, e->tCL);
+		return -ERANGE;
+	}
+
+	if (e->tWR >= NV_MEM_WR_GDDR5_MAX) {
+		NV_WARN(dev, "(%u) Invalid tWR: %u", t->id, e->tWR);
+		return -ERANGE;
+	}
+
+	if (t->odt > 3) {
+		NV_WARN(dev, "(%u) Invalid odt value, assuming autocal: %x",
+			t->id, t->odt);
+		t->odt = 0;
+	}
+
+	t->mr[0] = (boot->mr[0] & 0x007) |
+		   ((e->tCL - 5) << 3) |
+		   ((e->tWR - 4) << 8);
+	t->mr[1] = (boot->mr[1] & 0x1007f0) |
+		   t->drive_strength |
+		   (t->odt << 2);
+
+	NV_DEBUG(dev, "(%u) MR: %08x %08x", t->id, t->mr[0], t->mr[1]);
+	return 0;
+}
+
+int
+nouveau_mem_timing_calc(struct drm_device *dev, u32 freq,
+			struct nouveau_pm_memtiming *t)
 {
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
 	struct nouveau_pm_engine *pm = &dev_priv->engine.pm;
-	struct nouveau_pm_memtimings *memtimings = &pm->memtimings;
-	struct nvbios *bios = &dev_priv->vbios;
-	struct bit_entry P;
-	struct nouveau_pm_tbl_header *hdr = NULL;
-	uint8_t magic_number;
-	u8 *entry;
-	int i;
+	struct nouveau_pm_memtiming *boot = &pm->boot.timing;
+	struct nouveau_pm_tbl_entry *e;
+	u8 ver, len, *ptr, *ramcfg;
+	int ret;
+
+	ptr = nouveau_perf_timing(dev, freq, &ver, &len);
+	if (!ptr || ptr[0] == 0x00) {
+		*t = *boot;
+		return 0;
+	}
+	e = (struct nouveau_pm_tbl_entry *)ptr;
+
+	t->tCWL = boot->tCWL;
+
+	switch (dev_priv->card_type) {
+	case NV_40:
+		ret = nv40_mem_timing_calc(dev, freq, e, len, boot, t);
+		break;
+	case NV_50:
+		ret = nv50_mem_timing_calc(dev, freq, e, len, boot, t);
+		break;
+	case NV_C0:
+		ret = nvc0_mem_timing_calc(dev, freq, e, len, boot, t);
+		break;
+	default:
+		ret = -ENODEV;
+		break;
+	}
 
-	if (bios->type == NVBIOS_BIT) {
-		if (bit_table(dev, 'P', &P))
-			return;
+	switch (dev_priv->vram_type * !ret) {
+	case NV_MEM_TYPE_GDDR3:
+		ret = nouveau_mem_gddr3_mr(dev, freq, e, len, boot, t);
+		break;
+	case NV_MEM_TYPE_GDDR5:
+		ret = nouveau_mem_gddr5_mr(dev, freq, e, len, boot, t);
+		break;
+	case NV_MEM_TYPE_DDR2:
+		ret = nouveau_mem_ddr2_mr(dev, freq, e, len, boot, t);
+		break;
+	case NV_MEM_TYPE_DDR3:
+		ret = nouveau_mem_ddr3_mr(dev, freq, e, len, boot, t);
+		break;
+	default:
+		ret = -EINVAL;
+		break;
+	}
+
+	ramcfg = nouveau_perf_ramcfg(dev, freq, &ver, &len);
+	if (ramcfg) {
+		int dll_off;
 
-		if (P.version == 1)
-			hdr = (struct nouveau_pm_tbl_header *) ROMPTR(dev, P.data[4]);
+		if (ver == 0x00)
+			dll_off = !!(ramcfg[3] & 0x04);
 		else
-		if (P.version == 2)
-			hdr = (struct nouveau_pm_tbl_header *) ROMPTR(dev, P.data[8]);
-		else {
-			NV_WARN(dev, "unknown mem for BIT P %d\n", P.version);
+			dll_off = !!(ramcfg[2] & 0x40);
+
+		switch (dev_priv->vram_type) {
+		case NV_MEM_TYPE_GDDR3:
+			t->mr[1] &= ~0x00000040;
+			t->mr[1] |=  0x00000040 * dll_off;
+			break;
+		default:
+			t->mr[1] &= ~0x00000001;
+			t->mr[1] |=  0x00000001 * dll_off;
+			break;
 		}
+	}
+
+	return ret;
+}
+
+void
+nouveau_mem_timing_read(struct drm_device *dev, struct nouveau_pm_memtiming *t)
+{
+	struct drm_nouveau_private *dev_priv = dev->dev_private;
+	u32 timing_base, timing_regs, mr_base;
+	int i;
+
+	if (dev_priv->card_type >= 0xC0) {
+		timing_base = 0x10f290;
+		mr_base = 0x10f300;
 	} else {
-		NV_DEBUG(dev, "BMP version too old for memory\n");
-		return;
+		timing_base = 0x100220;
+		mr_base = 0x1002c0;
 	}
 
-	if (!hdr) {
-		NV_DEBUG(dev, "memory timing table pointer invalid\n");
+	t->id = -1;
+
+	switch (dev_priv->card_type) {
+	case NV_50:
+		timing_regs = 9;
+		break;
+	case NV_C0:
+	case NV_D0:
+		timing_regs = 5;
+		break;
+	case NV_30:
+	case NV_40:
+		timing_regs = 3;
+		break;
+	default:
+		timing_regs = 0;
 		return;
 	}
+	for(i = 0; i < timing_regs; i++)
+		t->reg[i] = nv_rd32(dev, timing_base + (0x04 * i));
+
+	t->tCWL = 0;
+	if (dev_priv->card_type < NV_C0) {
+		t->tCWL = ((nv_rd32(dev, 0x100228) & 0x0f000000) >> 24) + 1;
+	} else if (dev_priv->card_type <= NV_D0) {
+		t->tCWL = ((nv_rd32(dev, 0x10f294) & 0x00000f80) >> 7);
+	}
 
-	if (hdr->version != 0x10) {
-		NV_WARN(dev, "memory timing table 0x%02x unknown\n", hdr->version);
-		return;
+	t->mr[0] = nv_rd32(dev, mr_base);
+	t->mr[1] = nv_rd32(dev, mr_base + 0x04);
+	t->mr[2] = nv_rd32(dev, mr_base + 0x20);
+	t->mr[3] = nv_rd32(dev, mr_base + 0x24);
+
+	t->odt = 0;
+	t->drive_strength = 0;
+
+	switch (dev_priv->vram_type) {
+	case NV_MEM_TYPE_DDR3:
+		t->odt |= (t->mr[1] & 0x200) >> 7;
+	case NV_MEM_TYPE_DDR2:
+		t->odt |= (t->mr[1] & 0x04) >> 2 |
+			  (t->mr[1] & 0x40) >> 5;
+		break;
+	case NV_MEM_TYPE_GDDR3:
+	case NV_MEM_TYPE_GDDR5:
+		t->drive_strength = t->mr[1] & 0x03;
+		t->odt = (t->mr[1] & 0x0c) >> 2;
+		break;
+	default:
+		break;
 	}
+}
 
-	/* validate record length */
-	if (hdr->entry_len < 15) {
-		NV_ERROR(dev, "mem timing table length unknown: %d\n", hdr->entry_len);
-		return;
+int
+nouveau_mem_exec(struct nouveau_mem_exec_func *exec,
+		 struct nouveau_pm_level *perflvl)
+{
+	struct drm_nouveau_private *dev_priv = exec->dev->dev_private;
+	struct nouveau_pm_memtiming *info = &perflvl->timing;
+	u32 tMRD = 1000, tCKSRE = 0, tCKSRX = 0, tXS = 0, tDLLK = 0;
+	u32 mr[3] = { info->mr[0], info->mr[1], info->mr[2] };
+	u32 mr1_dlloff;
+
+	switch (dev_priv->vram_type) {
+	case NV_MEM_TYPE_DDR2:
+		tDLLK = 2000;
+		mr1_dlloff = 0x00000001;
+		break;
+	case NV_MEM_TYPE_DDR3:
+		tDLLK = 12000;
+		mr1_dlloff = 0x00000001;
+		break;
+	case NV_MEM_TYPE_GDDR3:
+		tDLLK = 40000;
+		mr1_dlloff = 0x00000040;
+		break;
+	default:
+		NV_ERROR(exec->dev, "cannot reclock unsupported memtype\n");
+		return -ENODEV;
 	}
 
-	/* parse vbios entries into common format */
-	memtimings->timing =
-		kcalloc(hdr->entry_cnt, sizeof(*memtimings->timing), GFP_KERNEL);
-	if (!memtimings->timing)
-		return;
+	/* fetch current MRs */
+	switch (dev_priv->vram_type) {
+	case NV_MEM_TYPE_GDDR3:
+	case NV_MEM_TYPE_DDR3:
+		mr[2] = exec->mrg(exec, 2);
+	default:
+		mr[1] = exec->mrg(exec, 1);
+		mr[0] = exec->mrg(exec, 0);
+		break;
+	}
 
-	/* Get "some number" from the timing reg for NV_40 and NV_50
-	 * Used in calculations later... source unknown */
-	magic_number = 0;
-	if (P.version == 1) {
-		magic_number = (nv_rd32(dev, 0x100228) & 0x0f000000) >> 24;
+	/* DLL 'on' -> DLL 'off' mode, disable before entering self-refresh  */
+	if (!(mr[1] & mr1_dlloff) && (info->mr[1] & mr1_dlloff)) {
+		exec->precharge(exec);
+		exec->mrs (exec, 1, mr[1] | mr1_dlloff);
+		exec->wait(exec, tMRD);
 	}
 
-	entry = (u8*) hdr + hdr->header_len;
-	for (i = 0; i < hdr->entry_cnt; i++, entry += hdr->entry_len) {
-		struct nouveau_pm_memtiming *timing = &pm->memtimings.timing[i];
-		if (entry[0] == 0)
-			continue;
+	/* enter self-refresh mode */
+	exec->precharge(exec);
+	exec->refresh(exec);
+	exec->refresh(exec);
+	exec->refresh_auto(exec, false);
+	exec->refresh_self(exec, true);
+	exec->wait(exec, tCKSRE);
+
+	/* modify input clock frequency */
+	exec->clock_set(exec);
+
+	/* exit self-refresh mode */
+	exec->wait(exec, tCKSRX);
+	exec->precharge(exec);
+	exec->refresh_self(exec, false);
+	exec->refresh_auto(exec, true);
+	exec->wait(exec, tXS);
+
+	/* update MRs */
+	if (mr[2] != info->mr[2]) {
+		exec->mrs (exec, 2, info->mr[2]);
+		exec->wait(exec, tMRD);
+	}
+
+	if (mr[1] != info->mr[1]) {
+		/* need to keep DLL off until later, at least on GDDR3 */
+		exec->mrs (exec, 1, info->mr[1] | (mr[1] & mr1_dlloff));
+		exec->wait(exec, tMRD);
+	}
+
+	if (mr[0] != info->mr[0]) {
+		exec->mrs (exec, 0, info->mr[0]);
+		exec->wait(exec, tMRD);
+	}
 
-		timing->id = i;
-		timing->WR = entry[0];
-		timing->CL = entry[2];
+	/* update PFB timing registers */
+	exec->timing_set(exec);
 
-		if(dev_priv->card_type <= NV_40) {
-			nv40_mem_timing_entry(dev,hdr,(struct nouveau_pm_tbl_entry*) entry,magic_number,&pm->memtimings.timing[i]);
-		} else if(dev_priv->card_type == NV_50){
-			nv50_mem_timing_entry(dev,&P,hdr,(struct nouveau_pm_tbl_entry*) entry,magic_number,&pm->memtimings.timing[i]);
-		} else if(dev_priv->card_type == NV_C0) {
-			nvc0_mem_timing_entry(dev,hdr,(struct nouveau_pm_tbl_entry*) entry,&pm->memtimings.timing[i]);
+	/* DLL (enable + ) reset */
+	if (!(info->mr[1] & mr1_dlloff)) {
+		if (mr[1] & mr1_dlloff) {
+			exec->mrs (exec, 1, info->mr[1]);
+			exec->wait(exec, tMRD);
 		}
+		exec->mrs (exec, 0, info->mr[0] | 0x00000100);
+		exec->wait(exec, tMRD);
+		exec->mrs (exec, 0, info->mr[0] | 0x00000000);
+		exec->wait(exec, tMRD);
+		exec->wait(exec, tDLLK);
+		if (dev_priv->vram_type == NV_MEM_TYPE_GDDR3)
+			exec->precharge(exec);
 	}
 
-	memtimings->nr_timing = hdr->entry_cnt;
-	memtimings->supported = P.version == 1;
+	return 0;
 }
 
-void
-nouveau_mem_timing_fini(struct drm_device *dev)
+int
+nouveau_mem_vbios_type(struct drm_device *dev)
 {
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	struct nouveau_pm_memtimings *mem = &dev_priv->engine.pm.memtimings;
+	struct bit_entry M;
+	u8 ramcfg = (nv_rd32(dev, 0x101000) & 0x0000003c) >> 2;
+	if (!bit_table(dev, 'M', &M) || M.version != 2 || M.length < 5) {
+		u8 *table = ROMPTR(dev, M.data[3]);
+		if (table && table[0] == 0x10 && ramcfg < table[3]) {
+			u8 *entry = table + table[1] + (ramcfg * table[2]);
+			switch (entry[0] & 0x0f) {
+			case 0: return NV_MEM_TYPE_DDR2;
+			case 1: return NV_MEM_TYPE_DDR3;
+			case 2: return NV_MEM_TYPE_GDDR3;
+			case 3: return NV_MEM_TYPE_GDDR5;
+			default:
+				break;
+			}
 
-	if(mem->timing) {
-		kfree(mem->timing);
-		mem->timing = NULL;
+		}
 	}
+	return NV_MEM_TYPE_UNKNOWN;
 }
 
 static int
diff --git a/drivers/gpu/drm/nouveau/nouveau_mxm.c b/drivers/gpu/drm/nouveau/nouveau_mxm.c
index e5a64f0f4cb7..07d0d1e03690 100644
--- a/drivers/gpu/drm/nouveau/nouveau_mxm.c
+++ b/drivers/gpu/drm/nouveau/nouveau_mxm.c
@@ -582,6 +582,35 @@ mxm_shadow_dsm(struct drm_device *dev, u8 version)
 
 #define WMI_WMMX_GUID "F6CB5C3C-9CAE-4EBD-B577-931EA32A2CC0"
 
+static u8
+wmi_wmmx_mxmi(struct drm_device *dev, u8 version)
+{
+	u32 mxmi_args[] = { 0x494D584D /* MXMI */, version, 0 };
+	struct acpi_buffer args = { sizeof(mxmi_args), mxmi_args };
+	struct acpi_buffer retn = { ACPI_ALLOCATE_BUFFER, NULL };
+	union acpi_object *obj;
+	acpi_status status;
+
+	status = wmi_evaluate_method(WMI_WMMX_GUID, 0, 0, &args, &retn);
+	if (ACPI_FAILURE(status)) {
+		MXM_DBG(dev, "WMMX MXMI returned %d\n", status);
+		return 0x00;
+	}
+
+	obj = retn.pointer;
+	if (obj->type == ACPI_TYPE_INTEGER) {
+		version = obj->integer.value;
+		MXM_DBG(dev, "WMMX MXMI version %d.%d\n",
+			     (version >> 4), version & 0x0f);
+	} else {
+		version = 0;
+		MXM_DBG(dev, "WMMX MXMI returned non-integer\n");
+	}
+
+	kfree(obj);
+	return version;
+}
+
 static bool
 mxm_shadow_wmi(struct drm_device *dev, u8 version)
 {
@@ -592,7 +621,15 @@ mxm_shadow_wmi(struct drm_device *dev, u8 version)
 	union acpi_object *obj;
 	acpi_status status;
 
-	if (!wmi_has_guid(WMI_WMMX_GUID))
+	if (!wmi_has_guid(WMI_WMMX_GUID)) {
+		MXM_DBG(dev, "WMMX GUID not found\n");
+		return false;
+	}
+
+	mxms_args[1] = wmi_wmmx_mxmi(dev, 0x00);
+	if (!mxms_args[1])
+		mxms_args[1] = wmi_wmmx_mxmi(dev, version);
+	if (!mxms_args[1])
 		return false;
 
 	status = wmi_evaluate_method(WMI_WMMX_GUID, 0, 0, &args, &retn);
diff --git a/drivers/gpu/drm/nouveau/nouveau_perf.c b/drivers/gpu/drm/nouveau/nouveau_perf.c
index 58f497343cec..69a528d106e6 100644
--- a/drivers/gpu/drm/nouveau/nouveau_perf.c
+++ b/drivers/gpu/drm/nouveau/nouveau_perf.c
@@ -27,6 +27,178 @@
 #include "nouveau_drv.h"
 #include "nouveau_pm.h"
 
+static u8 *
+nouveau_perf_table(struct drm_device *dev, u8 *ver)
+{
+	struct drm_nouveau_private *dev_priv = dev->dev_private;
+	struct nvbios *bios = &dev_priv->vbios;
+	struct bit_entry P;
+
+	if (!bit_table(dev, 'P', &P) && P.version && P.version <= 2) {
+		u8 *perf = ROMPTR(dev, P.data[0]);
+		if (perf) {
+			*ver = perf[0];
+			return perf;
+		}
+	}
+
+	if (bios->type == NVBIOS_BMP) {
+		if (bios->data[bios->offset + 6] >= 0x25) {
+			u8 *perf = ROMPTR(dev, bios->data[bios->offset + 0x94]);
+			if (perf) {
+				*ver = perf[1];
+				return perf;
+			}
+		}
+	}
+
+	return NULL;
+}
+
+static u8 *
+nouveau_perf_entry(struct drm_device *dev, int idx,
+		   u8 *ver, u8 *hdr, u8 *cnt, u8 *len)
+{
+	u8 *perf = nouveau_perf_table(dev, ver);
+	if (perf) {
+		if (*ver >= 0x12 && *ver < 0x20 && idx < perf[2]) {
+			*hdr = perf[3];
+			*cnt = 0;
+			*len = 0;
+			return perf + perf[0] + idx * perf[3];
+		} else
+		if (*ver >= 0x20 && *ver < 0x40 && idx < perf[2]) {
+			*hdr = perf[3];
+			*cnt = perf[4];
+			*len = perf[5];
+			return perf + perf[1] + idx * (*hdr + (*cnt * *len));
+		} else
+		if (*ver >= 0x40 && *ver < 0x41 && idx < perf[5]) {
+			*hdr = perf[2];
+			*cnt = perf[4];
+			*len = perf[3];
+			return perf + perf[1] + idx * (*hdr + (*cnt * *len));
+		}
+	}
+	return NULL;
+}
+
+static u8 *
+nouveau_perf_rammap(struct drm_device *dev, u32 freq,
+		    u8 *ver, u8 *hdr, u8 *cnt, u8 *len)
+{
+	struct drm_nouveau_private *dev_priv = dev->dev_private;
+	struct bit_entry P;
+	u8 *perf, i = 0;
+
+	if (!bit_table(dev, 'P', &P) && P.version == 2) {
+		u8 *rammap = ROMPTR(dev, P.data[4]);
+		if (rammap) {
+			u8 *ramcfg = rammap + rammap[1];
+
+			*ver = rammap[0];
+			*hdr = rammap[2];
+			*cnt = rammap[4];
+			*len = rammap[3];
+
+			freq /= 1000;
+			for (i = 0; i < rammap[5]; i++) {
+				if (freq >= ROM16(ramcfg[0]) &&
+				    freq <= ROM16(ramcfg[2]))
+					return ramcfg;
+
+				ramcfg += *hdr + (*cnt * *len);
+			}
+		}
+
+		return NULL;
+	}
+
+	if (dev_priv->chipset == 0x49 ||
+	    dev_priv->chipset == 0x4b)
+		freq /= 2;
+
+	while ((perf = nouveau_perf_entry(dev, i++, ver, hdr, cnt, len))) {
+		if (*ver >= 0x20 && *ver < 0x25) {
+			if (perf[0] != 0xff && freq <= ROM16(perf[11]) * 1000)
+				break;
+		} else
+		if (*ver >= 0x25 && *ver < 0x40) {
+			if (perf[0] != 0xff && freq <= ROM16(perf[12]) * 1000)
+				break;
+		}
+	}
+
+	if (perf) {
+		u8 *ramcfg = perf + *hdr;
+		*ver = 0x00;
+		*hdr = 0;
+		return ramcfg;
+	}
+
+	return NULL;
+}
+
+u8 *
+nouveau_perf_ramcfg(struct drm_device *dev, u32 freq, u8 *ver, u8 *len)
+{
+	struct drm_nouveau_private *dev_priv = dev->dev_private;
+	struct nvbios *bios = &dev_priv->vbios;
+	u8 strap, hdr, cnt;
+	u8 *rammap;
+
+	strap = (nv_rd32(dev, 0x101000) & 0x0000003c) >> 2;
+	if (bios->ram_restrict_tbl_ptr)
+		strap = bios->data[bios->ram_restrict_tbl_ptr + strap];
+
+	rammap = nouveau_perf_rammap(dev, freq, ver, &hdr, &cnt, len);
+	if (rammap && strap < cnt)
+		return rammap + hdr + (strap * *len);
+
+	return NULL;
+}
+
+u8 *
+nouveau_perf_timing(struct drm_device *dev, u32 freq, u8 *ver, u8 *len)
+{
+	struct drm_nouveau_private *dev_priv = dev->dev_private;
+	struct nvbios *bios = &dev_priv->vbios;
+	struct bit_entry P;
+	u8 *perf, *timing = NULL;
+	u8 i = 0, hdr, cnt;
+
+	if (bios->type == NVBIOS_BMP) {
+		while ((perf = nouveau_perf_entry(dev, i++, ver, &hdr, &cnt,
+						  len)) && *ver == 0x15) {
+			if (freq <= ROM32(perf[5]) * 20) {
+				*ver = 0x00;
+				*len = 14;
+				return perf + 41;
+			}
+		}
+		return NULL;
+	}
+
+	if (!bit_table(dev, 'P', &P)) {
+		if (P.version == 1)
+			timing = ROMPTR(dev, P.data[4]);
+		else
+		if (P.version == 2)
+			timing = ROMPTR(dev, P.data[8]);
+	}
+
+	if (timing && timing[0] == 0x10) {
+		u8 *ramcfg = nouveau_perf_ramcfg(dev, freq, ver, len);
+		if (ramcfg && ramcfg[1] < timing[2]) {
+			*ver = timing[0];
+			*len = timing[3];
+			return timing + timing[1] + (ramcfg[1] * timing[3]);
+		}
+	}
+
+	return NULL;
+}
+
 static void
 legacy_perf_init(struct drm_device *dev)
 {
@@ -72,74 +244,11 @@ legacy_perf_init(struct drm_device *dev)
 	pm->nr_perflvl = 1;
 }
 
-static struct nouveau_pm_memtiming *
-nouveau_perf_timing(struct drm_device *dev, struct bit_entry *P,
-		    u16 memclk, u8 *entry, u8 recordlen, u8 entries)
-{
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	struct nouveau_pm_engine *pm = &dev_priv->engine.pm;
-	struct nvbios *bios = &dev_priv->vbios;
-	u8 ramcfg;
-	int i;
-
-	/* perf v2 has a separate "timing map" table, we have to match
-	 * the target memory clock to a specific entry, *then* use
-	 * ramcfg to select the correct subentry
-	 */
-	if (P->version == 2) {
-		u8 *tmap = ROMPTR(dev, P->data[4]);
-		if (!tmap) {
-			NV_DEBUG(dev, "no timing map pointer\n");
-			return NULL;
-		}
-
-		if (tmap[0] != 0x10) {
-			NV_WARN(dev, "timing map 0x%02x unknown\n", tmap[0]);
-			return NULL;
-		}
-
-		entry = tmap + tmap[1];
-		recordlen = tmap[2] + (tmap[4] * tmap[3]);
-		for (i = 0; i < tmap[5]; i++, entry += recordlen) {
-			if (memclk >= ROM16(entry[0]) &&
-			    memclk <= ROM16(entry[2]))
-				break;
-		}
-
-		if (i == tmap[5]) {
-			NV_WARN(dev, "no match in timing map table\n");
-			return NULL;
-		}
-
-		entry += tmap[2];
-		recordlen = tmap[3];
-		entries   = tmap[4];
-	}
-
-	ramcfg = (nv_rd32(dev, NV_PEXTDEV_BOOT_0) & 0x0000003c) >> 2;
-	if (bios->ram_restrict_tbl_ptr)
-		ramcfg = bios->data[bios->ram_restrict_tbl_ptr + ramcfg];
-
-	if (ramcfg >= entries) {
-		NV_WARN(dev, "ramcfg strap out of bounds!\n");
-		return NULL;
-	}
-
-	entry += ramcfg * recordlen;
-	if (entry[1] >= pm->memtimings.nr_timing) {
-		if (entry[1] != 0xff)
-			NV_WARN(dev, "timingset %d does not exist\n", entry[1]);
-		return NULL;
-	}
-
-	return &pm->memtimings.timing[entry[1]];
-}
-
 static void
-nouveau_perf_voltage(struct drm_device *dev, struct bit_entry *P,
-		     struct nouveau_pm_level *perflvl)
+nouveau_perf_voltage(struct drm_device *dev, struct nouveau_pm_level *perflvl)
 {
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
+	struct bit_entry P;
 	u8 *vmap;
 	int id;
 
@@ -158,13 +267,13 @@ nouveau_perf_voltage(struct drm_device *dev, struct bit_entry *P,
 	/* on newer ones, the perflvl stores an index into yet another
 	 * vbios table containing a min/max voltage value for the perflvl
 	 */
-	if (P->version != 2 || P->length < 34) {
+	if (bit_table(dev, 'P', &P) || P.version != 2 || P.length < 34) {
 		NV_DEBUG(dev, "where's our volt map table ptr? %d %d\n",
-			 P->version, P->length);
+			 P.version, P.length);
 		return;
 	}
 
-	vmap = ROMPTR(dev, P->data[32]);
+	vmap = ROMPTR(dev, P.data[32]);
 	if (!vmap) {
 		NV_DEBUG(dev, "volt map table pointer invalid\n");
 		return;
@@ -183,130 +292,70 @@ nouveau_perf_init(struct drm_device *dev)
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
 	struct nouveau_pm_engine *pm = &dev_priv->engine.pm;
 	struct nvbios *bios = &dev_priv->vbios;
-	struct bit_entry P;
-	struct nouveau_pm_memtimings *memtimings = &pm->memtimings;
-	struct nouveau_pm_tbl_header mt_hdr;
-	u8 version, headerlen, recordlen, entries;
-	u8 *perf, *entry;
-	int vid, i;
-
-	if (bios->type == NVBIOS_BIT) {
-		if (bit_table(dev, 'P', &P))
-			return;
-
-		if (P.version != 1 && P.version != 2) {
-			NV_WARN(dev, "unknown perf for BIT P %d\n", P.version);
-			return;
-		}
-
-		perf = ROMPTR(dev, P.data[0]);
-		version   = perf[0];
-		headerlen = perf[1];
-		if (version < 0x40) {
-			recordlen = perf[3] + (perf[4] * perf[5]);
-			entries   = perf[2];
-
-			pm->pwm_divisor = ROM16(perf[6]);
-		} else {
-			recordlen = perf[2] + (perf[3] * perf[4]);
-			entries   = perf[5];
-		}
-	} else {
-		if (bios->data[bios->offset + 6] < 0x25) {
-			legacy_perf_init(dev);
-			return;
-		}
+	u8 *perf, ver, hdr, cnt, len;
+	int ret, vid, i = -1;
 
-		perf = ROMPTR(dev, bios->data[bios->offset + 0x94]);
-		if (!perf) {
-			NV_DEBUG(dev, "perf table pointer invalid\n");
-			return;
-		}
-
-		version   = perf[1];
-		headerlen = perf[0];
-		recordlen = perf[3];
-		entries   = perf[2];
-	}
-
-	if (entries > NOUVEAU_PM_MAX_LEVEL) {
-		NV_DEBUG(dev, "perf table has too many entries - buggy vbios?\n");
-		entries = NOUVEAU_PM_MAX_LEVEL;
+	if (bios->type == NVBIOS_BMP && bios->data[bios->offset + 6] < 0x25) {
+		legacy_perf_init(dev);
+		return;
 	}
 
-	entry = perf + headerlen;
-
-	/* For version 0x15, initialize memtiming table */
-	if(version == 0x15) {
-		memtimings->timing =
-				kcalloc(entries, sizeof(*memtimings->timing), GFP_KERNEL);
-		if (!memtimings->timing) {
-			NV_WARN(dev,"Could not allocate memtiming table\n");
-			return;
-		}
-
-		mt_hdr.entry_cnt = entries;
-		mt_hdr.entry_len = 14;
-		mt_hdr.version = version;
-		mt_hdr.header_len = 4;
-	}
+	perf = nouveau_perf_table(dev, &ver);
+	if (ver >= 0x20 && ver < 0x40)
+		pm->fan.pwm_divisor = ROM16(perf[6]);
 
-	for (i = 0; i < entries; i++) {
+	while ((perf = nouveau_perf_entry(dev, ++i, &ver, &hdr, &cnt, &len))) {
 		struct nouveau_pm_level *perflvl = &pm->perflvl[pm->nr_perflvl];
 
-		perflvl->timing = NULL;
-
-		if (entry[0] == 0xff) {
-			entry += recordlen;
+		if (perf[0] == 0xff)
 			continue;
-		}
 
-		switch (version) {
+		switch (ver) {
 		case 0x12:
 		case 0x13:
 		case 0x15:
-			perflvl->fanspeed = entry[55];
-			if (recordlen > 56)
-				perflvl->volt_min = entry[56];
-			perflvl->core = ROM32(entry[1]) * 10;
-			perflvl->memory = ROM32(entry[5]) * 20;
+			perflvl->fanspeed = perf[55];
+			if (hdr > 56)
+				perflvl->volt_min = perf[56];
+			perflvl->core = ROM32(perf[1]) * 10;
+			perflvl->memory = ROM32(perf[5]) * 20;
 			break;
 		case 0x21:
 		case 0x23:
 		case 0x24:
-			perflvl->fanspeed = entry[4];
-			perflvl->volt_min = entry[5];
-			perflvl->shader = ROM16(entry[6]) * 1000;
+			perflvl->fanspeed = perf[4];
+			perflvl->volt_min = perf[5];
+			perflvl->shader = ROM16(perf[6]) * 1000;
 			perflvl->core = perflvl->shader;
-			perflvl->core += (signed char)entry[8] * 1000;
+			perflvl->core += (signed char)perf[8] * 1000;
 			if (dev_priv->chipset == 0x49 ||
 			    dev_priv->chipset == 0x4b)
-				perflvl->memory = ROM16(entry[11]) * 1000;
+				perflvl->memory = ROM16(perf[11]) * 1000;
 			else
-				perflvl->memory = ROM16(entry[11]) * 2000;
+				perflvl->memory = ROM16(perf[11]) * 2000;
 			break;
 		case 0x25:
-			perflvl->fanspeed = entry[4];
-			perflvl->volt_min = entry[5];
-			perflvl->core = ROM16(entry[6]) * 1000;
-			perflvl->shader = ROM16(entry[10]) * 1000;
-			perflvl->memory = ROM16(entry[12]) * 1000;
+			perflvl->fanspeed = perf[4];
+			perflvl->volt_min = perf[5];
+			perflvl->core = ROM16(perf[6]) * 1000;
+			perflvl->shader = ROM16(perf[10]) * 1000;
+			perflvl->memory = ROM16(perf[12]) * 1000;
 			break;
 		case 0x30:
-			perflvl->memscript = ROM16(entry[2]);
+			perflvl->memscript = ROM16(perf[2]);
 		case 0x35:
-			perflvl->fanspeed = entry[6];
-			perflvl->volt_min = entry[7];
-			perflvl->core = ROM16(entry[8]) * 1000;
-			perflvl->shader = ROM16(entry[10]) * 1000;
-			perflvl->memory = ROM16(entry[12]) * 1000;
-			perflvl->vdec = ROM16(entry[16]) * 1000;
-			perflvl->dom6 = ROM16(entry[20]) * 1000;
+			perflvl->fanspeed = perf[6];
+			perflvl->volt_min = perf[7];
+			perflvl->core = ROM16(perf[8]) * 1000;
+			perflvl->shader = ROM16(perf[10]) * 1000;
+			perflvl->memory = ROM16(perf[12]) * 1000;
+			perflvl->vdec = ROM16(perf[16]) * 1000;
+			perflvl->dom6 = ROM16(perf[20]) * 1000;
 			break;
 		case 0x40:
-#define subent(n) (ROM16(entry[perf[2] + ((n) * perf[3])]) & 0xfff) * 1000
+#define subent(n) ((ROM16(perf[hdr + (n) * len]) & 0xfff) * 1000)
 			perflvl->fanspeed = 0; /*XXX*/
-			perflvl->volt_min = entry[2];
+			perflvl->volt_min = perf[2];
 			if (dev_priv->card_type == NV_50) {
 				perflvl->core   = subent(0);
 				perflvl->shader = subent(1);
@@ -329,36 +378,34 @@ nouveau_perf_init(struct drm_device *dev)
 		}
 
 		/* make sure vid is valid */
-		nouveau_perf_voltage(dev, &P, perflvl);
+		nouveau_perf_voltage(dev, perflvl);
 		if (pm->voltage.supported && perflvl->volt_min) {
 			vid = nouveau_volt_vid_lookup(dev, perflvl->volt_min);
 			if (vid < 0) {
-				NV_DEBUG(dev, "drop perflvl %d, bad vid\n", i);
-				entry += recordlen;
+				NV_DEBUG(dev, "perflvl %d, bad vid\n", i);
 				continue;
 			}
 		}
 
 		/* get the corresponding memory timings */
-		if (version == 0x15) {
-			memtimings->timing[i].id = i;
-			nv30_mem_timing_entry(dev,&mt_hdr,(struct nouveau_pm_tbl_entry*) &entry[41],0,&memtimings->timing[i]);
-			perflvl->timing = &memtimings->timing[i];
-		} else if (version > 0x15) {
-			/* last 3 args are for < 0x40, ignored for >= 0x40 */
-			perflvl->timing =
-				nouveau_perf_timing(dev, &P,
-						    perflvl->memory / 1000,
-						    entry + perf[3],
-						    perf[5], perf[4]);
+		ret = nouveau_mem_timing_calc(dev, perflvl->memory,
+					          &perflvl->timing);
+		if (ret) {
+			NV_DEBUG(dev, "perflvl %d, bad timing: %d\n", i, ret);
+			continue;
 		}
 
 		snprintf(perflvl->name, sizeof(perflvl->name),
 			 "performance_level_%d", i);
 		perflvl->id = i;
-		pm->nr_perflvl++;
 
-		entry += recordlen;
+		snprintf(perflvl->profile.name, sizeof(perflvl->profile.name),
+			 "%d", perflvl->id);
+		perflvl->profile.func = &nouveau_pm_static_profile_func;
+		list_add_tail(&perflvl->profile.head, &pm->profiles);
+
+
+		pm->nr_perflvl++;
 	}
 }
 
diff --git a/drivers/gpu/drm/nouveau/nouveau_pm.c b/drivers/gpu/drm/nouveau/nouveau_pm.c
index 9064d7f19794..34d591b7d4ef 100644
--- a/drivers/gpu/drm/nouveau/nouveau_pm.c
+++ b/drivers/gpu/drm/nouveau/nouveau_pm.c
@@ -50,7 +50,7 @@ nouveau_pwmfan_get(struct drm_device *dev)
 	ret = nouveau_gpio_find(dev, 0, DCB_GPIO_PWM_FAN, 0xff, &gpio);
 	if (ret == 0) {
 		ret = pm->pwm_get(dev, gpio.line, &divs, &duty);
-		if (ret == 0) {
+		if (ret == 0 && divs) {
 			divs = max(divs, duty);
 			if (dev_priv->card_type <= NV_40 || (gpio.log[0] & 1))
 				duty = divs - duty;
@@ -77,7 +77,7 @@ nouveau_pwmfan_set(struct drm_device *dev, int percent)
 
 	ret = nouveau_gpio_find(dev, 0, DCB_GPIO_PWM_FAN, 0xff, &gpio);
 	if (ret == 0) {
-		divs = pm->pwm_divisor;
+		divs = pm->fan.pwm_divisor;
 		if (pm->fan.pwm_freq) {
 			/*XXX: PNVIO clock more than likely... */
 			divs = 135000 / pm->fan.pwm_freq;
@@ -89,7 +89,10 @@ nouveau_pwmfan_set(struct drm_device *dev, int percent)
 		if (dev_priv->card_type <= NV_40 || (gpio.log[0] & 1))
 			duty = divs - duty;
 
-		return pm->pwm_set(dev, gpio.line, divs, duty);
+		ret = pm->pwm_set(dev, gpio.line, divs, duty);
+		if (!ret)
+			pm->fan.percent = percent;
+		return ret;
 	}
 
 	return -ENODEV;
@@ -144,9 +147,13 @@ nouveau_pm_perflvl_set(struct drm_device *dev, struct nouveau_pm_level *perflvl)
 		return ret;
 
 	state = pm->clocks_pre(dev, perflvl);
-	if (IS_ERR(state))
-		return PTR_ERR(state);
-	pm->clocks_set(dev, state);
+	if (IS_ERR(state)) {
+		ret = PTR_ERR(state);
+		goto error;
+	}
+	ret = pm->clocks_set(dev, state);
+	if (ret)
+		goto error;
 
 	ret = nouveau_pm_perflvl_aux(dev, perflvl, perflvl, pm->cur);
 	if (ret)
@@ -154,6 +161,65 @@ nouveau_pm_perflvl_set(struct drm_device *dev, struct nouveau_pm_level *perflvl)
 
 	pm->cur = perflvl;
 	return 0;
+
+error:
+	/* restore the fan speed and voltage before leaving */
+	nouveau_pm_perflvl_aux(dev, perflvl, perflvl, pm->cur);
+	return ret;
+}
+
+void
+nouveau_pm_trigger(struct drm_device *dev)
+{
+	struct drm_nouveau_private *dev_priv = dev->dev_private;
+	struct nouveau_pm_engine *pm = &dev_priv->engine.pm;
+	struct nouveau_pm_profile *profile = NULL;
+	struct nouveau_pm_level *perflvl = NULL;
+	int ret;
+
+	/* select power profile based on current power source */
+	if (power_supply_is_system_supplied())
+		profile = pm->profile_ac;
+	else
+		profile = pm->profile_dc;
+
+	if (profile != pm->profile) {
+		pm->profile->func->fini(pm->profile);
+		pm->profile = profile;
+		pm->profile->func->init(pm->profile);
+	}
+
+	/* select performance level based on profile */
+	perflvl = profile->func->select(profile);
+
+	/* change perflvl, if necessary */
+	if (perflvl != pm->cur) {
+		struct nouveau_timer_engine *ptimer = &dev_priv->engine.timer;
+		u64 time0 = ptimer->read(dev);
+
+		NV_INFO(dev, "setting performance level: %d", perflvl->id);
+		ret = nouveau_pm_perflvl_set(dev, perflvl);
+		if (ret)
+			NV_INFO(dev, "> reclocking failed: %d\n\n", ret);
+
+		NV_INFO(dev, "> reclocking took %lluns\n\n",
+			     ptimer->read(dev) - time0);
+	}
+}
+
+static struct nouveau_pm_profile *
+profile_find(struct drm_device *dev, const char *string)
+{
+	struct drm_nouveau_private *dev_priv = dev->dev_private;
+	struct nouveau_pm_engine *pm = &dev_priv->engine.pm;
+	struct nouveau_pm_profile *profile;
+
+	list_for_each_entry(profile, &pm->profiles, head) {
+		if (!strncmp(profile->name, string, sizeof(profile->name)))
+			return profile;
+	}
+
+	return NULL;
 }
 
 static int
@@ -161,33 +227,54 @@ nouveau_pm_profile_set(struct drm_device *dev, const char *profile)
 {
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
 	struct nouveau_pm_engine *pm = &dev_priv->engine.pm;
-	struct nouveau_pm_level *perflvl = NULL;
+	struct nouveau_pm_profile *ac = NULL, *dc = NULL;
+	char string[16], *cur = string, *ptr;
 
 	/* safety precaution, for now */
 	if (nouveau_perflvl_wr != 7777)
 		return -EPERM;
 
-	if (!strncmp(profile, "boot", 4))
-		perflvl = &pm->boot;
-	else {
-		int pl = simple_strtol(profile, NULL, 10);
-		int i;
+	strncpy(string, profile, sizeof(string));
+	if ((ptr = strchr(string, '\n')))
+		*ptr = '\0';
 
-		for (i = 0; i < pm->nr_perflvl; i++) {
-			if (pm->perflvl[i].id == pl) {
-				perflvl = &pm->perflvl[i];
-				break;
-			}
-		}
+	ptr = strsep(&cur, ",");
+	if (ptr)
+		ac = profile_find(dev, ptr);
 
-		if (!perflvl)
-			return -EINVAL;
-	}
+	ptr = strsep(&cur, ",");
+	if (ptr)
+		dc = profile_find(dev, ptr);
+	else
+		dc = ac;
+
+	if (ac == NULL || dc == NULL)
+		return -EINVAL;
+
+	pm->profile_ac = ac;
+	pm->profile_dc = dc;
+	nouveau_pm_trigger(dev);
+	return 0;
+}
+
+static void
+nouveau_pm_static_dummy(struct nouveau_pm_profile *profile)
+{
+}
 
-	NV_INFO(dev, "setting performance level: %s\n", profile);
-	return nouveau_pm_perflvl_set(dev, perflvl);
+static struct nouveau_pm_level *
+nouveau_pm_static_select(struct nouveau_pm_profile *profile)
+{
+	return container_of(profile, struct nouveau_pm_level, profile);
 }
 
+const struct nouveau_pm_profile_func nouveau_pm_static_profile_func = {
+	.destroy = nouveau_pm_static_dummy,
+	.init = nouveau_pm_static_dummy,
+	.fini = nouveau_pm_static_dummy,
+	.select = nouveau_pm_static_select,
+};
+
 static int
 nouveau_pm_perflvl_get(struct drm_device *dev, struct nouveau_pm_level *perflvl)
 {
@@ -197,9 +284,11 @@ nouveau_pm_perflvl_get(struct drm_device *dev, struct nouveau_pm_level *perflvl)
 
 	memset(perflvl, 0, sizeof(*perflvl));
 
-	ret = pm->clocks_get(dev, perflvl);
-	if (ret)
-		return ret;
+	if (pm->clocks_get) {
+		ret = pm->clocks_get(dev, perflvl);
+		if (ret)
+			return ret;
+	}
 
 	if (pm->voltage.supported && pm->voltage_get) {
 		ret = pm->voltage_get(dev);
@@ -213,13 +302,14 @@ nouveau_pm_perflvl_get(struct drm_device *dev, struct nouveau_pm_level *perflvl)
 	if (ret > 0)
 		perflvl->fanspeed = ret;
 
+	nouveau_mem_timing_read(dev, &perflvl->timing);
 	return 0;
 }
 
 static void
 nouveau_pm_perflvl_info(struct nouveau_pm_level *perflvl, char *ptr, int len)
 {
-	char c[16], s[16], v[32], f[16], t[16], m[16];
+	char c[16], s[16], v[32], f[16], m[16];
 
 	c[0] = '\0';
 	if (perflvl->core)
@@ -247,18 +337,15 @@ nouveau_pm_perflvl_info(struct nouveau_pm_level *perflvl, char *ptr, int len)
 	if (perflvl->fanspeed)
 		snprintf(f, sizeof(f), " fanspeed %d%%", perflvl->fanspeed);
 
-	t[0] = '\0';
-	if (perflvl->timing)
-		snprintf(t, sizeof(t), " timing %d", perflvl->timing->id);
-
-	snprintf(ptr, len, "%s%s%s%s%s%s\n", c, s, m, t, v, f);
+	snprintf(ptr, len, "%s%s%s%s%s\n", c, s, m, v, f);
 }
 
 static ssize_t
 nouveau_pm_get_perflvl_info(struct device *d,
 			    struct device_attribute *a, char *buf)
 {
-	struct nouveau_pm_level *perflvl = (struct nouveau_pm_level *)a;
+	struct nouveau_pm_level *perflvl =
+		container_of(a, struct nouveau_pm_level, dev_attr);
 	char *ptr = buf;
 	int len = PAGE_SIZE;
 
@@ -280,12 +367,8 @@ nouveau_pm_get_perflvl(struct device *d, struct device_attribute *a, char *buf)
 	int len = PAGE_SIZE, ret;
 	char *ptr = buf;
 
-	if (!pm->cur)
-		snprintf(ptr, len, "setting: boot\n");
-	else if (pm->cur == &pm->boot)
-		snprintf(ptr, len, "setting: boot\nc:");
-	else
-		snprintf(ptr, len, "setting: static %d\nc:", pm->cur->id);
+	snprintf(ptr, len, "profile: %s, %s\nc:",
+		 pm->profile_ac->name, pm->profile_dc->name);
 	ptr += strlen(buf);
 	len -= strlen(buf);
 
@@ -397,7 +480,7 @@ nouveau_hwmon_set_max_temp(struct device *d, struct device_attribute *a,
 	struct nouveau_pm_threshold_temp *temp = &pm->threshold_temp;
 	long value;
 
-	if (strict_strtol(buf, 10, &value) == -EINVAL)
+	if (kstrtol(buf, 10, &value) == -EINVAL)
 		return count;
 
 	temp->down_clock = value/1000;
@@ -432,7 +515,7 @@ nouveau_hwmon_set_critical_temp(struct device *d, struct device_attribute *a,
 	struct nouveau_pm_threshold_temp *temp = &pm->threshold_temp;
 	long value;
 
-	if (strict_strtol(buf, 10, &value) == -EINVAL)
+	if (kstrtol(buf, 10, &value) == -EINVAL)
 		return count;
 
 	temp->critical = value/1000;
@@ -529,7 +612,7 @@ nouveau_hwmon_set_pwm0(struct device *d, struct device_attribute *a,
 	if (nouveau_perflvl_wr != 7777)
 		return -EPERM;
 
-	if (strict_strtol(buf, 10, &value) == -EINVAL)
+	if (kstrtol(buf, 10, &value) == -EINVAL)
 		return -EINVAL;
 
 	if (value < pm->fan.min_duty)
@@ -568,7 +651,7 @@ nouveau_hwmon_set_pwm0_min(struct device *d, struct device_attribute *a,
 	struct nouveau_pm_engine *pm = &dev_priv->engine.pm;
 	long value;
 
-	if (strict_strtol(buf, 10, &value) == -EINVAL)
+	if (kstrtol(buf, 10, &value) == -EINVAL)
 		return -EINVAL;
 
 	if (value < 0)
@@ -609,7 +692,7 @@ nouveau_hwmon_set_pwm0_max(struct device *d, struct device_attribute *a,
 	struct nouveau_pm_engine *pm = &dev_priv->engine.pm;
 	long value;
 
-	if (strict_strtol(buf, 10, &value) == -EINVAL)
+	if (kstrtol(buf, 10, &value) == -EINVAL)
 		return -EINVAL;
 
 	if (value < 0)
@@ -731,8 +814,10 @@ nouveau_hwmon_fini(struct drm_device *dev)
 
 	if (pm->hwmon) {
 		sysfs_remove_group(&dev->pdev->dev.kobj, &hwmon_attrgroup);
-		sysfs_remove_group(&dev->pdev->dev.kobj, &hwmon_pwm_fan_attrgroup);
-		sysfs_remove_group(&dev->pdev->dev.kobj, &hwmon_fan_rpm_attrgroup);
+		sysfs_remove_group(&dev->pdev->dev.kobj,
+				   &hwmon_pwm_fan_attrgroup);
+		sysfs_remove_group(&dev->pdev->dev.kobj,
+				   &hwmon_fan_rpm_attrgroup);
 
 		hwmon_device_unregister(pm->hwmon);
 	}
@@ -752,6 +837,7 @@ nouveau_pm_acpi_event(struct notifier_block *nb, unsigned long val, void *data)
 		bool ac = power_supply_is_system_supplied();
 
 		NV_DEBUG(dev, "power supply changed: %s\n", ac ? "AC" : "DC");
+		nouveau_pm_trigger(dev);
 	}
 
 	return NOTIFY_OK;
@@ -766,35 +852,48 @@ nouveau_pm_init(struct drm_device *dev)
 	char info[256];
 	int ret, i;
 
-	nouveau_mem_timing_init(dev);
+	/* parse aux tables from vbios */
 	nouveau_volt_init(dev);
-	nouveau_perf_init(dev);
 	nouveau_temp_init(dev);
 
+	/* determine current ("boot") performance level */
+	ret = nouveau_pm_perflvl_get(dev, &pm->boot);
+	if (ret) {
+		NV_ERROR(dev, "failed to determine boot perflvl\n");
+		return ret;
+	}
+
+	strncpy(pm->boot.name, "boot", 4);
+	strncpy(pm->boot.profile.name, "boot", 4);
+	pm->boot.profile.func = &nouveau_pm_static_profile_func;
+
+	INIT_LIST_HEAD(&pm->profiles);
+	list_add(&pm->boot.profile.head, &pm->profiles);
+
+	pm->profile_ac = &pm->boot.profile;
+	pm->profile_dc = &pm->boot.profile;
+	pm->profile = &pm->boot.profile;
+	pm->cur = &pm->boot;
+
+	/* add performance levels from vbios */
+	nouveau_perf_init(dev);
+
+	/* display available performance levels */
 	NV_INFO(dev, "%d available performance level(s)\n", pm->nr_perflvl);
 	for (i = 0; i < pm->nr_perflvl; i++) {
 		nouveau_pm_perflvl_info(&pm->perflvl[i], info, sizeof(info));
 		NV_INFO(dev, "%d:%s", pm->perflvl[i].id, info);
 	}
 
-	/* determine current ("boot") performance level */
-	ret = nouveau_pm_perflvl_get(dev, &pm->boot);
-	if (ret == 0) {
-		strncpy(pm->boot.name, "boot", 4);
-		pm->cur = &pm->boot;
-
-		nouveau_pm_perflvl_info(&pm->boot, info, sizeof(info));
-		NV_INFO(dev, "c:%s", info);
-	}
+	nouveau_pm_perflvl_info(&pm->boot, info, sizeof(info));
+	NV_INFO(dev, "c:%s", info);
 
 	/* switch performance levels now if requested */
-	if (nouveau_perflvl != NULL) {
-		ret = nouveau_pm_profile_set(dev, nouveau_perflvl);
-		if (ret) {
-			NV_ERROR(dev, "error setting perflvl \"%s\": %d\n",
-				 nouveau_perflvl, ret);
-		}
-	}
+	if (nouveau_perflvl != NULL)
+		nouveau_pm_profile_set(dev, nouveau_perflvl);
+
+	/* determine the current fan speed */
+	pm->fan.percent = nouveau_pwmfan_get(dev);
 
 	nouveau_sysfs_init(dev);
 	nouveau_hwmon_init(dev);
@@ -811,6 +910,12 @@ nouveau_pm_fini(struct drm_device *dev)
 {
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
 	struct nouveau_pm_engine *pm = &dev_priv->engine.pm;
+	struct nouveau_pm_profile *profile, *tmp;
+
+	list_for_each_entry_safe(profile, tmp, &pm->profiles, head) {
+		list_del(&profile->head);
+		profile->func->destroy(profile);
+	}
 
 	if (pm->cur != &pm->boot)
 		nouveau_pm_perflvl_set(dev, &pm->boot);
@@ -818,7 +923,6 @@ nouveau_pm_fini(struct drm_device *dev)
 	nouveau_temp_fini(dev);
 	nouveau_perf_fini(dev);
 	nouveau_volt_fini(dev);
-	nouveau_mem_timing_fini(dev);
 
 #if defined(CONFIG_ACPI) && defined(CONFIG_POWER_SUPPLY)
 	unregister_acpi_notifier(&pm->acpi_nb);
@@ -840,4 +944,5 @@ nouveau_pm_resume(struct drm_device *dev)
 	perflvl = pm->cur;
 	pm->cur = &pm->boot;
 	nouveau_pm_perflvl_set(dev, perflvl);
+	nouveau_pwmfan_set(dev, pm->fan.percent);
 }
diff --git a/drivers/gpu/drm/nouveau/nouveau_pm.h b/drivers/gpu/drm/nouveau/nouveau_pm.h
index 2f8e14fbcff8..3f82dfea61dd 100644
--- a/drivers/gpu/drm/nouveau/nouveau_pm.h
+++ b/drivers/gpu/drm/nouveau/nouveau_pm.h
@@ -25,10 +25,30 @@
 #ifndef __NOUVEAU_PM_H__
 #define __NOUVEAU_PM_H__
 
+struct nouveau_mem_exec_func {
+	struct drm_device *dev;
+	void (*precharge)(struct nouveau_mem_exec_func *);
+	void (*refresh)(struct nouveau_mem_exec_func *);
+	void (*refresh_auto)(struct nouveau_mem_exec_func *, bool);
+	void (*refresh_self)(struct nouveau_mem_exec_func *, bool);
+	void (*wait)(struct nouveau_mem_exec_func *, u32 nsec);
+	u32  (*mrg)(struct nouveau_mem_exec_func *, int mr);
+	void (*mrs)(struct nouveau_mem_exec_func *, int mr, u32 data);
+	void (*clock_set)(struct nouveau_mem_exec_func *);
+	void (*timing_set)(struct nouveau_mem_exec_func *);
+	void *priv;
+};
+
+/* nouveau_mem.c */
+int  nouveau_mem_exec(struct nouveau_mem_exec_func *,
+		      struct nouveau_pm_level *);
+
 /* nouveau_pm.c */
 int  nouveau_pm_init(struct drm_device *dev);
 void nouveau_pm_fini(struct drm_device *dev);
 void nouveau_pm_resume(struct drm_device *dev);
+extern const struct nouveau_pm_profile_func nouveau_pm_static_profile_func;
+void nouveau_pm_trigger(struct drm_device *dev);
 
 /* nouveau_volt.c */
 void nouveau_volt_init(struct drm_device *);
@@ -41,6 +61,8 @@ int  nouveau_voltage_gpio_set(struct drm_device *, int voltage);
 /* nouveau_perf.c */
 void nouveau_perf_init(struct drm_device *);
 void nouveau_perf_fini(struct drm_device *);
+u8 *nouveau_perf_timing(struct drm_device *, u32 freq, u8 *ver, u8 *len);
+u8 *nouveau_perf_ramcfg(struct drm_device *, u32 freq, u8 *ver, u8 *len);
 
 /* nouveau_mem.c */
 void nouveau_mem_timing_init(struct drm_device *);
diff --git a/drivers/gpu/drm/nouveau/nouveau_state.c b/drivers/gpu/drm/nouveau/nouveau_state.c
index f80c5e0762ff..c2a8511e855a 100644
--- a/drivers/gpu/drm/nouveau/nouveau_state.c
+++ b/drivers/gpu/drm/nouveau/nouveau_state.c
@@ -87,7 +87,7 @@ static int nouveau_init_engine_ptrs(struct drm_device *dev)
 		engine->pm.clocks_get		= nv04_pm_clocks_get;
 		engine->pm.clocks_pre		= nv04_pm_clocks_pre;
 		engine->pm.clocks_set		= nv04_pm_clocks_set;
-		engine->vram.init		= nouveau_mem_detect;
+		engine->vram.init		= nv04_fb_vram_init;
 		engine->vram.takedown		= nouveau_stub_takedown;
 		engine->vram.flags_valid	= nouveau_mem_flags_valid;
 		break;
@@ -134,7 +134,11 @@ static int nouveau_init_engine_ptrs(struct drm_device *dev)
 		engine->pm.clocks_get		= nv04_pm_clocks_get;
 		engine->pm.clocks_pre		= nv04_pm_clocks_pre;
 		engine->pm.clocks_set		= nv04_pm_clocks_set;
-		engine->vram.init		= nouveau_mem_detect;
+		if (dev_priv->chipset == 0x1a ||
+		    dev_priv->chipset == 0x1f)
+			engine->vram.init	= nv1a_fb_vram_init;
+		else
+			engine->vram.init	= nv10_fb_vram_init;
 		engine->vram.takedown		= nouveau_stub_takedown;
 		engine->vram.flags_valid	= nouveau_mem_flags_valid;
 		break;
@@ -153,11 +157,11 @@ static int nouveau_init_engine_ptrs(struct drm_device *dev)
 		engine->timer.init		= nv04_timer_init;
 		engine->timer.read		= nv04_timer_read;
 		engine->timer.takedown		= nv04_timer_takedown;
-		engine->fb.init			= nv10_fb_init;
-		engine->fb.takedown		= nv10_fb_takedown;
-		engine->fb.init_tile_region	= nv10_fb_init_tile_region;
-		engine->fb.set_tile_region	= nv10_fb_set_tile_region;
-		engine->fb.free_tile_region	= nv10_fb_free_tile_region;
+		engine->fb.init			= nv20_fb_init;
+		engine->fb.takedown		= nv20_fb_takedown;
+		engine->fb.init_tile_region	= nv20_fb_init_tile_region;
+		engine->fb.set_tile_region	= nv20_fb_set_tile_region;
+		engine->fb.free_tile_region	= nv20_fb_free_tile_region;
 		engine->fifo.channels		= 32;
 		engine->fifo.init		= nv10_fifo_init;
 		engine->fifo.takedown		= nv04_fifo_fini;
@@ -181,7 +185,7 @@ static int nouveau_init_engine_ptrs(struct drm_device *dev)
 		engine->pm.clocks_get		= nv04_pm_clocks_get;
 		engine->pm.clocks_pre		= nv04_pm_clocks_pre;
 		engine->pm.clocks_set		= nv04_pm_clocks_set;
-		engine->vram.init		= nouveau_mem_detect;
+		engine->vram.init		= nv20_fb_vram_init;
 		engine->vram.takedown		= nouveau_stub_takedown;
 		engine->vram.flags_valid	= nouveau_mem_flags_valid;
 		break;
@@ -230,7 +234,7 @@ static int nouveau_init_engine_ptrs(struct drm_device *dev)
 		engine->pm.clocks_set		= nv04_pm_clocks_set;
 		engine->pm.voltage_get		= nouveau_voltage_gpio_get;
 		engine->pm.voltage_set		= nouveau_voltage_gpio_set;
-		engine->vram.init		= nouveau_mem_detect;
+		engine->vram.init		= nv20_fb_vram_init;
 		engine->vram.takedown		= nouveau_stub_takedown;
 		engine->vram.flags_valid	= nouveau_mem_flags_valid;
 		break;
@@ -286,7 +290,7 @@ static int nouveau_init_engine_ptrs(struct drm_device *dev)
 		engine->pm.temp_get		= nv40_temp_get;
 		engine->pm.pwm_get		= nv40_pm_pwm_get;
 		engine->pm.pwm_set		= nv40_pm_pwm_set;
-		engine->vram.init		= nouveau_mem_detect;
+		engine->vram.init		= nv40_fb_vram_init;
 		engine->vram.takedown		= nouveau_stub_takedown;
 		engine->vram.flags_valid	= nouveau_mem_flags_valid;
 		break;
@@ -475,6 +479,47 @@ static int nouveau_init_engine_ptrs(struct drm_device *dev)
 		engine->pm.voltage_get		= nouveau_voltage_gpio_get;
 		engine->pm.voltage_set		= nouveau_voltage_gpio_set;
 		break;
+	case 0xe0:
+		engine->instmem.init		= nvc0_instmem_init;
+		engine->instmem.takedown	= nvc0_instmem_takedown;
+		engine->instmem.suspend		= nvc0_instmem_suspend;
+		engine->instmem.resume		= nvc0_instmem_resume;
+		engine->instmem.get		= nv50_instmem_get;
+		engine->instmem.put		= nv50_instmem_put;
+		engine->instmem.map		= nv50_instmem_map;
+		engine->instmem.unmap		= nv50_instmem_unmap;
+		engine->instmem.flush		= nv84_instmem_flush;
+		engine->mc.init			= nv50_mc_init;
+		engine->mc.takedown		= nv50_mc_takedown;
+		engine->timer.init		= nv04_timer_init;
+		engine->timer.read		= nv04_timer_read;
+		engine->timer.takedown		= nv04_timer_takedown;
+		engine->fb.init			= nvc0_fb_init;
+		engine->fb.takedown		= nvc0_fb_takedown;
+		engine->fifo.channels		= 0;
+		engine->fifo.init		= nouveau_stub_init;
+		engine->fifo.takedown		= nouveau_stub_takedown;
+		engine->fifo.disable		= nvc0_fifo_disable;
+		engine->fifo.enable		= nvc0_fifo_enable;
+		engine->fifo.reassign		= nvc0_fifo_reassign;
+		engine->fifo.unload_context	= nouveau_stub_init;
+		engine->display.early_init	= nouveau_stub_init;
+		engine->display.late_takedown	= nouveau_stub_takedown;
+		engine->display.create		= nvd0_display_create;
+		engine->display.destroy		= nvd0_display_destroy;
+		engine->display.init		= nvd0_display_init;
+		engine->display.fini		= nvd0_display_fini;
+		engine->gpio.init		= nv50_gpio_init;
+		engine->gpio.fini		= nv50_gpio_fini;
+		engine->gpio.drive		= nvd0_gpio_drive;
+		engine->gpio.sense		= nvd0_gpio_sense;
+		engine->gpio.irq_enable		= nv50_gpio_irq_enable;
+		engine->vram.init		= nvc0_vram_init;
+		engine->vram.takedown		= nv50_vram_fini;
+		engine->vram.get		= nvc0_vram_new;
+		engine->vram.put		= nv50_vram_del;
+		engine->vram.flags_valid	= nvc0_vram_flags_valid;
+		break;
 	default:
 		NV_ERROR(dev, "NV%02x unsupported\n", dev_priv->chipset);
 		return 1;
@@ -548,6 +593,75 @@ static bool nouveau_switcheroo_can_switch(struct pci_dev *pdev)
 	return can_switch;
 }
 
+static void
+nouveau_card_channel_fini(struct drm_device *dev)
+{
+	struct drm_nouveau_private *dev_priv = dev->dev_private;
+
+	if (dev_priv->channel)
+		nouveau_channel_put_unlocked(&dev_priv->channel);
+}
+
+static int
+nouveau_card_channel_init(struct drm_device *dev)
+{
+	struct drm_nouveau_private *dev_priv = dev->dev_private;
+	struct nouveau_channel *chan;
+	int ret, oclass;
+
+	ret = nouveau_channel_alloc(dev, &chan, NULL, NvDmaFB, NvDmaTT);
+	dev_priv->channel = chan;
+	if (ret)
+		return ret;
+
+	mutex_unlock(&dev_priv->channel->mutex);
+
+	if (dev_priv->card_type <= NV_50) {
+		if (dev_priv->card_type < NV_50)
+			oclass = 0x0039;
+		else
+			oclass = 0x5039;
+
+		ret = nouveau_gpuobj_gr_new(chan, NvM2MF, oclass);
+		if (ret)
+			goto error;
+
+		ret = nouveau_notifier_alloc(chan, NvNotify0, 32, 0xfe0, 0x1000,
+					     &chan->m2mf_ntfy);
+		if (ret)
+			goto error;
+
+		ret = RING_SPACE(chan, 6);
+		if (ret)
+			goto error;
+
+		BEGIN_RING(chan, NvSubM2MF, NV_MEMORY_TO_MEMORY_FORMAT_NAME, 1);
+		OUT_RING  (chan, NvM2MF);
+		BEGIN_RING(chan, NvSubM2MF, NV_MEMORY_TO_MEMORY_FORMAT_DMA_NOTIFY, 3);
+		OUT_RING  (chan, NvNotify0);
+		OUT_RING  (chan, chan->vram_handle);
+		OUT_RING  (chan, chan->gart_handle);
+	} else
+	if (dev_priv->card_type <= NV_D0) {
+		ret = nouveau_gpuobj_gr_new(chan, 0x9039, 0x9039);
+		if (ret)
+			goto error;
+
+		ret = RING_SPACE(chan, 2);
+		if (ret)
+			goto error;
+
+		BEGIN_NVC0(chan, 2, NvSubM2MF, 0x0000, 1);
+		OUT_RING  (chan, 0x00009039);
+	}
+
+	FIRE_RING (chan);
+error:
+	if (ret)
+		nouveau_card_channel_fini(dev);
+	return ret;
+}
+
 int
 nouveau_card_init(struct drm_device *dev)
 {
@@ -588,47 +702,45 @@ nouveau_card_init(struct drm_device *dev)
 		nv_mask(dev, 0x00088080, 0x00000800, 0x00000000);
 	}
 
-	nouveau_pm_init(dev);
-
-	ret = engine->vram.init(dev);
+	/* PMC */
+	ret = engine->mc.init(dev);
 	if (ret)
 		goto out_bios;
 
-	ret = nouveau_gpuobj_init(dev);
+	/* PTIMER */
+	ret = engine->timer.init(dev);
 	if (ret)
-		goto out_vram;
+		goto out_mc;
 
-	ret = engine->instmem.init(dev);
+	/* PFB */
+	ret = engine->fb.init(dev);
 	if (ret)
-		goto out_gpuobj;
+		goto out_timer;
 
-	ret = nouveau_mem_vram_init(dev);
+	ret = engine->vram.init(dev);
 	if (ret)
-		goto out_instmem;
+		goto out_fb;
 
-	ret = nouveau_mem_gart_init(dev);
+	/* PGPIO */
+	ret = nouveau_gpio_create(dev);
 	if (ret)
-		goto out_ttmvram;
+		goto out_vram;
 
-	/* PMC */
-	ret = engine->mc.init(dev);
+	ret = nouveau_gpuobj_init(dev);
 	if (ret)
-		goto out_gart;
+		goto out_gpio;
 
-	/* PGPIO */
-	ret = nouveau_gpio_create(dev);
+	ret = engine->instmem.init(dev);
 	if (ret)
-		goto out_mc;
+		goto out_gpuobj;
 
-	/* PTIMER */
-	ret = engine->timer.init(dev);
+	ret = nouveau_mem_vram_init(dev);
 	if (ret)
-		goto out_gpio;
+		goto out_instmem;
 
-	/* PFB */
-	ret = engine->fb.init(dev);
+	ret = nouveau_mem_gart_init(dev);
 	if (ret)
-		goto out_timer;
+		goto out_ttmvram;
 
 	if (!dev_priv->noaccel) {
 		switch (dev_priv->card_type) {
@@ -734,18 +846,16 @@ nouveau_card_init(struct drm_device *dev)
 		goto out_irq;
 
 	nouveau_backlight_init(dev);
+	nouveau_pm_init(dev);
 
-	if (dev_priv->eng[NVOBJ_ENGINE_GR]) {
-		ret = nouveau_fence_init(dev);
-		if (ret)
-			goto out_disp;
+	ret = nouveau_fence_init(dev);
+	if (ret)
+		goto out_pm;
 
-		ret = nouveau_channel_alloc(dev, &dev_priv->channel, NULL,
-					    NvDmaFB, NvDmaTT);
+	if (dev_priv->eng[NVOBJ_ENGINE_GR]) {
+		ret = nouveau_card_channel_init(dev);
 		if (ret)
 			goto out_fence;
-
-		mutex_unlock(&dev_priv->channel->mutex);
 	}
 
 	if (dev->mode_config.num_crtc) {
@@ -759,10 +869,11 @@ nouveau_card_init(struct drm_device *dev)
 	return 0;
 
 out_chan:
-	nouveau_channel_put_unlocked(&dev_priv->channel);
+	nouveau_card_channel_fini(dev);
 out_fence:
 	nouveau_fence_fini(dev);
-out_disp:
+out_pm:
+	nouveau_pm_fini(dev);
 	nouveau_backlight_exit(dev);
 	nouveau_display_destroy(dev);
 out_irq:
@@ -779,15 +890,6 @@ out_engine:
 			dev_priv->eng[e]->destroy(dev,e );
 		}
 	}
-
-	engine->fb.takedown(dev);
-out_timer:
-	engine->timer.takedown(dev);
-out_gpio:
-	nouveau_gpio_destroy(dev);
-out_mc:
-	engine->mc.takedown(dev);
-out_gart:
 	nouveau_mem_gart_fini(dev);
 out_ttmvram:
 	nouveau_mem_vram_fini(dev);
@@ -795,10 +897,17 @@ out_instmem:
 	engine->instmem.takedown(dev);
 out_gpuobj:
 	nouveau_gpuobj_takedown(dev);
+out_gpio:
+	nouveau_gpio_destroy(dev);
 out_vram:
 	engine->vram.takedown(dev);
+out_fb:
+	engine->fb.takedown(dev);
+out_timer:
+	engine->timer.takedown(dev);
+out_mc:
+	engine->mc.takedown(dev);
 out_bios:
-	nouveau_pm_fini(dev);
 	nouveau_bios_takedown(dev);
 out_display_early:
 	engine->display.late_takedown(dev);
@@ -818,11 +927,9 @@ static void nouveau_card_takedown(struct drm_device *dev)
 		nouveau_display_fini(dev);
 	}
 
-	if (dev_priv->channel) {
-		nouveau_channel_put_unlocked(&dev_priv->channel);
-		nouveau_fence_fini(dev);
-	}
-
+	nouveau_card_channel_fini(dev);
+	nouveau_fence_fini(dev);
+	nouveau_pm_fini(dev);
 	nouveau_backlight_exit(dev);
 	nouveau_display_destroy(dev);
 
@@ -835,11 +942,6 @@ static void nouveau_card_takedown(struct drm_device *dev)
 			}
 		}
 	}
-	engine->fb.takedown(dev);
-	engine->timer.takedown(dev);
-	nouveau_gpio_destroy(dev);
-	engine->mc.takedown(dev);
-	engine->display.late_takedown(dev);
 
 	if (dev_priv->vga_ram) {
 		nouveau_bo_unpin(dev_priv->vga_ram);
@@ -855,12 +957,17 @@ static void nouveau_card_takedown(struct drm_device *dev)
 
 	engine->instmem.takedown(dev);
 	nouveau_gpuobj_takedown(dev);
-	engine->vram.takedown(dev);
 
-	nouveau_irq_fini(dev);
+	nouveau_gpio_destroy(dev);
+	engine->vram.takedown(dev);
+	engine->fb.takedown(dev);
+	engine->timer.takedown(dev);
+	engine->mc.takedown(dev);
 
-	nouveau_pm_fini(dev);
 	nouveau_bios_takedown(dev);
+	engine->display.late_takedown(dev);
+
+	nouveau_irq_fini(dev);
 
 	vga_client_register(dev->pdev, NULL, NULL, NULL);
 }
@@ -990,8 +1097,8 @@ static int nouveau_remove_conflicting_drivers(struct drm_device *dev)
 int nouveau_load(struct drm_device *dev, unsigned long flags)
 {
 	struct drm_nouveau_private *dev_priv;
-	uint32_t reg0, strap;
-	resource_size_t mmio_start_offs;
+	unsigned long long offset, length;
+	uint32_t reg0 = ~0, strap;
 	int ret;
 
 	dev_priv = kzalloc(sizeof(*dev_priv), GFP_KERNEL);
@@ -1002,83 +1109,90 @@ int nouveau_load(struct drm_device *dev, unsigned long flags)
 	dev->dev_private = dev_priv;
 	dev_priv->dev = dev;
 
+	pci_set_master(dev->pdev);
+
 	dev_priv->flags = flags & NOUVEAU_FLAGS;
 
 	NV_DEBUG(dev, "vendor: 0x%X device: 0x%X class: 0x%X\n",
 		 dev->pci_vendor, dev->pci_device, dev->pdev->class);
 
-	/* resource 0 is mmio regs */
-	/* resource 1 is linear FB */
-	/* resource 2 is RAMIN (mmio regs + 0x1000000) */
-	/* resource 6 is bios */
+	/* first up, map the start of mmio and determine the chipset */
+	dev_priv->mmio = ioremap(pci_resource_start(dev->pdev, 0), PAGE_SIZE);
+	if (dev_priv->mmio) {
+#ifdef __BIG_ENDIAN
+		/* put the card into big-endian mode if it's not */
+		if (nv_rd32(dev, NV03_PMC_BOOT_1) != 0x01000001)
+			nv_wr32(dev, NV03_PMC_BOOT_1, 0x01000001);
+		DRM_MEMORYBARRIER();
+#endif
 
-	/* map the mmio regs */
-	mmio_start_offs = pci_resource_start(dev->pdev, 0);
-	dev_priv->mmio = ioremap(mmio_start_offs, 0x00800000);
-	if (!dev_priv->mmio) {
-		NV_ERROR(dev, "Unable to initialize the mmio mapping. "
-			 "Please report your setup to " DRIVER_EMAIL "\n");
+		/* determine chipset and derive architecture from it */
+		reg0 = nv_rd32(dev, NV03_PMC_BOOT_0);
+		if ((reg0 & 0x0f000000) > 0) {
+			dev_priv->chipset = (reg0 & 0xff00000) >> 20;
+			switch (dev_priv->chipset & 0xf0) {
+			case 0x10:
+			case 0x20:
+			case 0x30:
+				dev_priv->card_type = dev_priv->chipset & 0xf0;
+				break;
+			case 0x40:
+			case 0x60:
+				dev_priv->card_type = NV_40;
+				break;
+			case 0x50:
+			case 0x80:
+			case 0x90:
+			case 0xa0:
+				dev_priv->card_type = NV_50;
+				break;
+			case 0xc0:
+				dev_priv->card_type = NV_C0;
+				break;
+			case 0xd0:
+				dev_priv->card_type = NV_D0;
+				break;
+			case 0xe0:
+				dev_priv->card_type = NV_E0;
+				break;
+			default:
+				break;
+			}
+		} else
+		if ((reg0 & 0xff00fff0) == 0x20004000) {
+			if (reg0 & 0x00f00000)
+				dev_priv->chipset = 0x05;
+			else
+				dev_priv->chipset = 0x04;
+			dev_priv->card_type = NV_04;
+		}
+
+		iounmap(dev_priv->mmio);
+	}
+
+	if (!dev_priv->card_type) {
+		NV_ERROR(dev, "unsupported chipset 0x%08x\n", reg0);
 		ret = -EINVAL;
 		goto err_priv;
 	}
-	NV_DEBUG(dev, "regs mapped ok at 0x%llx\n",
-					(unsigned long long)mmio_start_offs);
-
-#ifdef __BIG_ENDIAN
-	/* Put the card in BE mode if it's not */
-	if (nv_rd32(dev, NV03_PMC_BOOT_1) != 0x01000001)
-		nv_wr32(dev, NV03_PMC_BOOT_1, 0x01000001);
 
-	DRM_MEMORYBARRIER();
-#endif
+	NV_INFO(dev, "Detected an NV%2x generation card (0x%08x)\n",
+		     dev_priv->card_type, reg0);
 
-	/* Time to determine the card architecture */
-	reg0 = nv_rd32(dev, NV03_PMC_BOOT_0);
-
-	/* We're dealing with >=NV10 */
-	if ((reg0 & 0x0f000000) > 0) {
-		/* Bit 27-20 contain the architecture in hex */
-		dev_priv->chipset = (reg0 & 0xff00000) >> 20;
-	/* NV04 or NV05 */
-	} else if ((reg0 & 0xff00fff0) == 0x20004000) {
-		if (reg0 & 0x00f00000)
-			dev_priv->chipset = 0x05;
-		else
-			dev_priv->chipset = 0x04;
-	} else
-		dev_priv->chipset = 0xff;
+	/* map the mmio regs, limiting the amount to preserve vmap space */
+	offset = pci_resource_start(dev->pdev, 0);
+	length = pci_resource_len(dev->pdev, 0);
+	if (dev_priv->card_type < NV_E0)
+		length = min(length, (unsigned long long)0x00800000);
 
-	switch (dev_priv->chipset & 0xf0) {
-	case 0x00:
-	case 0x10:
-	case 0x20:
-	case 0x30:
-		dev_priv->card_type = dev_priv->chipset & 0xf0;
-		break;
-	case 0x40:
-	case 0x60:
-		dev_priv->card_type = NV_40;
-		break;
-	case 0x50:
-	case 0x80:
-	case 0x90:
-	case 0xa0:
-		dev_priv->card_type = NV_50;
-		break;
-	case 0xc0:
-		dev_priv->card_type = NV_C0;
-		break;
-	case 0xd0:
-		dev_priv->card_type = NV_D0;
-		break;
-	default:
-		NV_INFO(dev, "Unsupported chipset 0x%08x\n", reg0);
+	dev_priv->mmio = ioremap(offset, length);
+	if (!dev_priv->mmio) {
+		NV_ERROR(dev, "Unable to initialize the mmio mapping. "
+			 "Please report your setup to " DRIVER_EMAIL "\n");
 		ret = -EINVAL;
-		goto err_mmio;
+		goto err_priv;
 	}
-
-	NV_INFO(dev, "Detected an NV%2x generation card (0x%08x)\n",
-		dev_priv->card_type, reg0);
+	NV_DEBUG(dev, "regs mapped ok at 0x%llx\n", offset);
 
 	/* determine frequency of timing crystal */
 	strap = nv_rd32(dev, 0x101000);
@@ -1136,7 +1250,7 @@ int nouveau_load(struct drm_device *dev, unsigned long flags)
 		}
 	} else {
 		dev_priv->ramin_size = 1 * 1024 * 1024;
-		dev_priv->ramin = ioremap(mmio_start_offs + NV_RAMIN,
+		dev_priv->ramin = ioremap(offset + NV_RAMIN,
 					  dev_priv->ramin_size);
 		if (!dev_priv->ramin) {
 			NV_ERROR(dev, "Failed to map BAR0 PRAMIN.\n");
diff --git a/drivers/gpu/drm/nouveau/nv04_fb.c b/drivers/gpu/drm/nouveau/nv04_fb.c
index 638cf601c427..d5eedd67afe5 100644
--- a/drivers/gpu/drm/nouveau/nv04_fb.c
+++ b/drivers/gpu/drm/nouveau/nv04_fb.c
@@ -4,6 +4,40 @@
 #include "nouveau_drm.h"
 
 int
+nv04_fb_vram_init(struct drm_device *dev)
+{
+	struct drm_nouveau_private *dev_priv = dev->dev_private;
+	u32 boot0 = nv_rd32(dev, NV04_PFB_BOOT_0);
+
+	if (boot0 & 0x00000100) {
+		dev_priv->vram_size  = ((boot0 >> 12) & 0xf) * 2 + 2;
+		dev_priv->vram_size *= 1024 * 1024;
+	} else {
+		switch (boot0 & NV04_PFB_BOOT_0_RAM_AMOUNT) {
+		case NV04_PFB_BOOT_0_RAM_AMOUNT_32MB:
+			dev_priv->vram_size = 32 * 1024 * 1024;
+			break;
+		case NV04_PFB_BOOT_0_RAM_AMOUNT_16MB:
+			dev_priv->vram_size = 16 * 1024 * 1024;
+			break;
+		case NV04_PFB_BOOT_0_RAM_AMOUNT_8MB:
+			dev_priv->vram_size = 8 * 1024 * 1024;
+			break;
+		case NV04_PFB_BOOT_0_RAM_AMOUNT_4MB:
+			dev_priv->vram_size = 4 * 1024 * 1024;
+			break;
+		}
+	}
+
+	if ((boot0 & 0x00000038) <= 0x10)
+		dev_priv->vram_type = NV_MEM_TYPE_SGRAM;
+	else
+		dev_priv->vram_type = NV_MEM_TYPE_SDRAM;
+
+	return 0;
+}
+
+int
 nv04_fb_init(struct drm_device *dev)
 {
 	/* This is what the DDX did for NV_ARCH_04, but a mmio-trace shows
diff --git a/drivers/gpu/drm/nouveau/nv10_fb.c b/drivers/gpu/drm/nouveau/nv10_fb.c
index f78181a59b4a..420b1608536d 100644
--- a/drivers/gpu/drm/nouveau/nv10_fb.c
+++ b/drivers/gpu/drm/nouveau/nv10_fb.c
@@ -3,81 +3,16 @@
 #include "nouveau_drv.h"
 #include "nouveau_drm.h"
 
-static struct drm_mm_node *
-nv20_fb_alloc_tag(struct drm_device *dev, uint32_t size)
-{
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	struct nouveau_fb_engine *pfb = &dev_priv->engine.fb;
-	struct drm_mm_node *mem;
-	int ret;
-
-	ret = drm_mm_pre_get(&pfb->tag_heap);
-	if (ret)
-		return NULL;
-
-	spin_lock(&dev_priv->tile.lock);
-	mem = drm_mm_search_free(&pfb->tag_heap, size, 0, 0);
-	if (mem)
-		mem = drm_mm_get_block_atomic(mem, size, 0);
-	spin_unlock(&dev_priv->tile.lock);
-
-	return mem;
-}
-
-static void
-nv20_fb_free_tag(struct drm_device *dev, struct drm_mm_node *mem)
-{
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-
-	spin_lock(&dev_priv->tile.lock);
-	drm_mm_put_block(mem);
-	spin_unlock(&dev_priv->tile.lock);
-}
-
 void
 nv10_fb_init_tile_region(struct drm_device *dev, int i, uint32_t addr,
 			 uint32_t size, uint32_t pitch, uint32_t flags)
 {
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
 	struct nouveau_tile_reg *tile = &dev_priv->tile.reg[i];
-	int bpp = (flags & NOUVEAU_GEM_TILE_32BPP ? 32 : 16);
 
-	tile->addr = addr;
+	tile->addr  = 0x80000000 | addr;
 	tile->limit = max(1u, addr + size) - 1;
 	tile->pitch = pitch;
-
-	if (dev_priv->card_type == NV_20) {
-		if (flags & NOUVEAU_GEM_TILE_ZETA) {
-			/*
-			 * Allocate some of the on-die tag memory,
-			 * used to store Z compression meta-data (most
-			 * likely just a bitmap determining if a given
-			 * tile is compressed or not).
-			 */
-			tile->tag_mem = nv20_fb_alloc_tag(dev, size / 256);
-
-			if (tile->tag_mem) {
-				/* Enable Z compression */
-				if (dev_priv->chipset >= 0x25)
-					tile->zcomp = tile->tag_mem->start |
-						(bpp == 16 ?
-						 NV25_PFB_ZCOMP_MODE_16 :
-						 NV25_PFB_ZCOMP_MODE_32);
-				else
-					tile->zcomp = tile->tag_mem->start |
-						NV20_PFB_ZCOMP_EN |
-						(bpp == 16 ? 0 :
-						 NV20_PFB_ZCOMP_MODE_32);
-			}
-
-			tile->addr |= 3;
-		} else {
-			tile->addr |= 1;
-		}
-
-	} else {
-		tile->addr |= 1 << 31;
-	}
 }
 
 void
@@ -86,11 +21,6 @@ nv10_fb_free_tile_region(struct drm_device *dev, int i)
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
 	struct nouveau_tile_reg *tile = &dev_priv->tile.reg[i];
 
-	if (tile->tag_mem) {
-		nv20_fb_free_tag(dev, tile->tag_mem);
-		tile->tag_mem = NULL;
-	}
-
 	tile->addr = tile->limit = tile->pitch = tile->zcomp = 0;
 }
 
@@ -103,9 +33,48 @@ nv10_fb_set_tile_region(struct drm_device *dev, int i)
 	nv_wr32(dev, NV10_PFB_TLIMIT(i), tile->limit);
 	nv_wr32(dev, NV10_PFB_TSIZE(i), tile->pitch);
 	nv_wr32(dev, NV10_PFB_TILE(i), tile->addr);
+}
+
+int
+nv1a_fb_vram_init(struct drm_device *dev)
+{
+	struct drm_nouveau_private *dev_priv = dev->dev_private;
+	struct pci_dev *bridge;
+	uint32_t mem, mib;
+
+	bridge = pci_get_bus_and_slot(0, PCI_DEVFN(0, 1));
+	if (!bridge) {
+		NV_ERROR(dev, "no bridge device\n");
+		return 0;
+	}
+
+	if (dev_priv->chipset == 0x1a) {
+		pci_read_config_dword(bridge, 0x7c, &mem);
+		mib = ((mem >> 6) & 31) + 1;
+	} else {
+		pci_read_config_dword(bridge, 0x84, &mem);
+		mib = ((mem >> 4) & 127) + 1;
+	}
+
+	dev_priv->vram_size = mib * 1024 * 1024;
+	return 0;
+}
+
+int
+nv10_fb_vram_init(struct drm_device *dev)
+{
+	struct drm_nouveau_private *dev_priv = dev->dev_private;
+	u32 fifo_data = nv_rd32(dev, NV04_PFB_FIFO_DATA);
+	u32 cfg0 = nv_rd32(dev, 0x100200);
 
-	if (dev_priv->card_type == NV_20)
-		nv_wr32(dev, NV20_PFB_ZCOMP(i), tile->zcomp);
+	dev_priv->vram_size = fifo_data & NV10_PFB_FIFO_DATA_RAM_AMOUNT_MB_MASK;
+
+	if (cfg0 & 0x00000001)
+		dev_priv->vram_type = NV_MEM_TYPE_DDR1;
+	else
+		dev_priv->vram_type = NV_MEM_TYPE_SDRAM;
+
+	return 0;
 }
 
 int
@@ -115,14 +84,8 @@ nv10_fb_init(struct drm_device *dev)
 	struct nouveau_fb_engine *pfb = &dev_priv->engine.fb;
 	int i;
 
-	pfb->num_tiles = NV10_PFB_TILE__SIZE;
-
-	if (dev_priv->card_type == NV_20)
-		drm_mm_init(&pfb->tag_heap, 0,
-			    (dev_priv->chipset >= 0x25 ?
-			     64 * 1024 : 32 * 1024));
-
 	/* Turn all the tiling regions off. */
+	pfb->num_tiles = NV10_PFB_TILE__SIZE;
 	for (i = 0; i < pfb->num_tiles; i++)
 		pfb->set_tile_region(dev, i);
 
@@ -138,7 +101,4 @@ nv10_fb_takedown(struct drm_device *dev)
 
 	for (i = 0; i < pfb->num_tiles; i++)
 		pfb->free_tile_region(dev, i);
-
-	if (dev_priv->card_type == NV_20)
-		drm_mm_takedown(&pfb->tag_heap);
 }
diff --git a/drivers/gpu/drm/nouveau/nv20_fb.c b/drivers/gpu/drm/nouveau/nv20_fb.c
new file mode 100644
index 000000000000..19bd64059a66
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/nv20_fb.c
@@ -0,0 +1,148 @@
+#include "drmP.h"
+#include "drm.h"
+#include "nouveau_drv.h"
+#include "nouveau_drm.h"
+
+static struct drm_mm_node *
+nv20_fb_alloc_tag(struct drm_device *dev, uint32_t size)
+{
+	struct drm_nouveau_private *dev_priv = dev->dev_private;
+	struct nouveau_fb_engine *pfb = &dev_priv->engine.fb;
+	struct drm_mm_node *mem;
+	int ret;
+
+	ret = drm_mm_pre_get(&pfb->tag_heap);
+	if (ret)
+		return NULL;
+
+	spin_lock(&dev_priv->tile.lock);
+	mem = drm_mm_search_free(&pfb->tag_heap, size, 0, 0);
+	if (mem)
+		mem = drm_mm_get_block_atomic(mem, size, 0);
+	spin_unlock(&dev_priv->tile.lock);
+
+	return mem;
+}
+
+static void
+nv20_fb_free_tag(struct drm_device *dev, struct drm_mm_node **pmem)
+{
+	struct drm_nouveau_private *dev_priv = dev->dev_private;
+	struct drm_mm_node *mem = *pmem;
+	if (mem) {
+		spin_lock(&dev_priv->tile.lock);
+		drm_mm_put_block(mem);
+		spin_unlock(&dev_priv->tile.lock);
+		*pmem = NULL;
+	}
+}
+
+void
+nv20_fb_init_tile_region(struct drm_device *dev, int i, uint32_t addr,
+			 uint32_t size, uint32_t pitch, uint32_t flags)
+{
+	struct drm_nouveau_private *dev_priv = dev->dev_private;
+	struct nouveau_tile_reg *tile = &dev_priv->tile.reg[i];
+	int bpp = (flags & NOUVEAU_GEM_TILE_32BPP ? 32 : 16);
+
+	tile->addr  = 0x00000001 | addr;
+	tile->limit = max(1u, addr + size) - 1;
+	tile->pitch = pitch;
+
+	/* Allocate some of the on-die tag memory, used to store Z
+	 * compression meta-data (most likely just a bitmap determining
+	 * if a given tile is compressed or not).
+	 */
+	if (flags & NOUVEAU_GEM_TILE_ZETA) {
+		tile->tag_mem = nv20_fb_alloc_tag(dev, size / 256);
+		if (tile->tag_mem) {
+			/* Enable Z compression */
+			tile->zcomp = tile->tag_mem->start;
+			if (dev_priv->chipset >= 0x25) {
+				if (bpp == 16)
+					tile->zcomp |= NV25_PFB_ZCOMP_MODE_16;
+				else
+					tile->zcomp |= NV25_PFB_ZCOMP_MODE_32;
+			} else {
+				tile->zcomp |= NV20_PFB_ZCOMP_EN;
+				if (bpp != 16)
+					tile->zcomp |= NV20_PFB_ZCOMP_MODE_32;
+			}
+		}
+
+		tile->addr |= 2;
+	}
+}
+
+void
+nv20_fb_free_tile_region(struct drm_device *dev, int i)
+{
+	struct drm_nouveau_private *dev_priv = dev->dev_private;
+	struct nouveau_tile_reg *tile = &dev_priv->tile.reg[i];
+
+	tile->addr = tile->limit = tile->pitch = tile->zcomp = 0;
+	nv20_fb_free_tag(dev, &tile->tag_mem);
+}
+
+void
+nv20_fb_set_tile_region(struct drm_device *dev, int i)
+{
+	struct drm_nouveau_private *dev_priv = dev->dev_private;
+	struct nouveau_tile_reg *tile = &dev_priv->tile.reg[i];
+
+	nv_wr32(dev, NV10_PFB_TLIMIT(i), tile->limit);
+	nv_wr32(dev, NV10_PFB_TSIZE(i), tile->pitch);
+	nv_wr32(dev, NV10_PFB_TILE(i), tile->addr);
+	nv_wr32(dev, NV20_PFB_ZCOMP(i), tile->zcomp);
+}
+
+int
+nv20_fb_vram_init(struct drm_device *dev)
+{
+	struct drm_nouveau_private *dev_priv = dev->dev_private;
+	u32 mem_size = nv_rd32(dev, 0x10020c);
+	u32 pbus1218 = nv_rd32(dev, 0x001218);
+
+	dev_priv->vram_size = mem_size & 0xff000000;
+	switch (pbus1218 & 0x00000300) {
+	case 0x00000000: dev_priv->vram_type = NV_MEM_TYPE_SDRAM; break;
+	case 0x00000100: dev_priv->vram_type = NV_MEM_TYPE_DDR1; break;
+	case 0x00000200: dev_priv->vram_type = NV_MEM_TYPE_GDDR3; break;
+	case 0x00000300: dev_priv->vram_type = NV_MEM_TYPE_GDDR2; break;
+	}
+
+	return 0;
+}
+
+int
+nv20_fb_init(struct drm_device *dev)
+{
+	struct drm_nouveau_private *dev_priv = dev->dev_private;
+	struct nouveau_fb_engine *pfb = &dev_priv->engine.fb;
+	int i;
+
+	if (dev_priv->chipset >= 0x25)
+		drm_mm_init(&pfb->tag_heap, 0, 64 * 1024);
+	else
+		drm_mm_init(&pfb->tag_heap, 0, 32 * 1024);
+
+	/* Turn all the tiling regions off. */
+	pfb->num_tiles = NV10_PFB_TILE__SIZE;
+	for (i = 0; i < pfb->num_tiles; i++)
+		pfb->set_tile_region(dev, i);
+
+	return 0;
+}
+
+void
+nv20_fb_takedown(struct drm_device *dev)
+{
+	struct drm_nouveau_private *dev_priv = dev->dev_private;
+	struct nouveau_fb_engine *pfb = &dev_priv->engine.fb;
+	int i;
+
+	for (i = 0; i < pfb->num_tiles; i++)
+		pfb->free_tile_region(dev, i);
+
+	drm_mm_takedown(&pfb->tag_heap);
+}
diff --git a/drivers/gpu/drm/nouveau/nv40_fb.c b/drivers/gpu/drm/nouveau/nv40_fb.c
index f0ac2a768c67..7fbcb334c096 100644
--- a/drivers/gpu/drm/nouveau/nv40_fb.c
+++ b/drivers/gpu/drm/nouveau/nv40_fb.c
@@ -72,6 +72,51 @@ nv44_fb_init_gart(struct drm_device *dev)
 }
 
 int
+nv40_fb_vram_init(struct drm_device *dev)
+{
+	struct drm_nouveau_private *dev_priv = dev->dev_private;
+
+	/* 0x001218 is actually present on a few other NV4X I looked at,
+	 * and even contains sane values matching 0x100474.  From looking
+	 * at various vbios images however, this isn't the case everywhere.
+	 * So, I chose to use the same regs I've seen NVIDIA reading around
+	 * the memory detection, hopefully that'll get us the right numbers
+	 */
+	if (dev_priv->chipset == 0x40) {
+		u32 pbus1218 = nv_rd32(dev, 0x001218);
+		switch (pbus1218 & 0x00000300) {
+		case 0x00000000: dev_priv->vram_type = NV_MEM_TYPE_SDRAM; break;
+		case 0x00000100: dev_priv->vram_type = NV_MEM_TYPE_DDR1; break;
+		case 0x00000200: dev_priv->vram_type = NV_MEM_TYPE_GDDR3; break;
+		case 0x00000300: dev_priv->vram_type = NV_MEM_TYPE_DDR2; break;
+		}
+	} else
+	if (dev_priv->chipset == 0x49 || dev_priv->chipset == 0x4b) {
+		u32 pfb914 = nv_rd32(dev, 0x100914);
+		switch (pfb914 & 0x00000003) {
+		case 0x00000000: dev_priv->vram_type = NV_MEM_TYPE_DDR1; break;
+		case 0x00000001: dev_priv->vram_type = NV_MEM_TYPE_DDR2; break;
+		case 0x00000002: dev_priv->vram_type = NV_MEM_TYPE_GDDR3; break;
+		case 0x00000003: break;
+		}
+	} else
+	if (dev_priv->chipset != 0x4e) {
+		u32 pfb474 = nv_rd32(dev, 0x100474);
+		if (pfb474 & 0x00000004)
+			dev_priv->vram_type = NV_MEM_TYPE_GDDR3;
+		if (pfb474 & 0x00000002)
+			dev_priv->vram_type = NV_MEM_TYPE_DDR2;
+		if (pfb474 & 0x00000001)
+			dev_priv->vram_type = NV_MEM_TYPE_DDR1;
+	} else {
+		dev_priv->vram_type = NV_MEM_TYPE_STOLEN;
+	}
+
+	dev_priv->vram_size = nv_rd32(dev, 0x10020c) & 0xff000000;
+	return 0;
+}
+
+int
 nv40_fb_init(struct drm_device *dev)
 {
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
diff --git a/drivers/gpu/drm/nouveau/nv50_crtc.c b/drivers/gpu/drm/nouveau/nv50_crtc.c
index 8f6c2ace3adf..701b927998bf 100644
--- a/drivers/gpu/drm/nouveau/nv50_crtc.c
+++ b/drivers/gpu/drm/nouveau/nv50_crtc.c
@@ -170,6 +170,41 @@ nv50_crtc_set_dither(struct nouveau_crtc *nv_crtc, bool update)
 	return ret;
 }
 
+static int
+nv50_crtc_set_color_vibrance(struct nouveau_crtc *nv_crtc, bool update)
+{
+	struct drm_device *dev = nv_crtc->base.dev;
+	struct nouveau_channel *evo = nv50_display(dev)->master;
+	int ret;
+	int adj;
+	u32 hue, vib;
+
+	NV_DEBUG_KMS(dev, "vibrance = %i, hue = %i\n",
+		     nv_crtc->color_vibrance, nv_crtc->vibrant_hue);
+
+	ret = RING_SPACE(evo, 2 + (update ? 2 : 0));
+	if (ret) {
+		NV_ERROR(dev, "no space while setting color vibrance\n");
+		return ret;
+	}
+
+	adj = (nv_crtc->color_vibrance > 0) ? 50 : 0;
+	vib = ((nv_crtc->color_vibrance * 2047 + adj) / 100) & 0xfff;
+
+	hue = ((nv_crtc->vibrant_hue * 2047) / 100) & 0xfff;
+
+	BEGIN_RING(evo, 0, NV50_EVO_CRTC(nv_crtc->index, COLOR_CTRL), 1);
+	OUT_RING  (evo, (hue << 20) | (vib << 8));
+
+	if (update) {
+		BEGIN_RING(evo, 0, NV50_EVO_UPDATE, 1);
+		OUT_RING  (evo, 0);
+		FIRE_RING (evo);
+	}
+
+	return 0;
+}
+
 struct nouveau_connector *
 nouveau_crtc_connector_get(struct nouveau_crtc *nv_crtc)
 {
@@ -577,8 +612,6 @@ nv50_crtc_do_mode_set_base(struct drm_crtc *crtc,
 	OUT_RING  (evo, fb->base.depth == 8 ?
 		   NV50_EVO_CRTC_CLUT_MODE_OFF : NV50_EVO_CRTC_CLUT_MODE_ON);
 
-	BEGIN_RING(evo, 0, NV50_EVO_CRTC(nv_crtc->index, COLOR_CTRL), 1);
-	OUT_RING  (evo, NV50_EVO_CRTC_COLOR_CTRL_COLOR);
 	BEGIN_RING(evo, 0, NV50_EVO_CRTC(nv_crtc->index, FB_POS), 1);
 	OUT_RING  (evo, (y << 16) | x);
 
@@ -661,6 +694,7 @@ nv50_crtc_mode_set(struct drm_crtc *crtc, struct drm_display_mode *umode,
 
 	nv_crtc->set_dither(nv_crtc, false);
 	nv_crtc->set_scale(nv_crtc, false);
+	nv_crtc->set_color_vibrance(nv_crtc, false);
 
 	return nv50_crtc_do_mode_set_base(crtc, old_fb, x, y, false);
 }
@@ -721,6 +755,9 @@ nv50_crtc_create(struct drm_device *dev, int index)
 	if (!nv_crtc)
 		return -ENOMEM;
 
+	nv_crtc->color_vibrance = 50;
+	nv_crtc->vibrant_hue = 0;
+
 	/* Default CLUT parameters, will be activated on the hw upon
 	 * first mode set.
 	 */
@@ -751,6 +788,7 @@ nv50_crtc_create(struct drm_device *dev, int index)
 	/* set function pointers */
 	nv_crtc->set_dither = nv50_crtc_set_dither;
 	nv_crtc->set_scale = nv50_crtc_set_scale;
+	nv_crtc->set_color_vibrance = nv50_crtc_set_color_vibrance;
 
 	drm_crtc_init(dev, &nv_crtc->base, &nv50_crtc_funcs);
 	drm_crtc_helper_add(&nv_crtc->base, &nv50_crtc_helper_funcs);
diff --git a/drivers/gpu/drm/nouveau/nv50_dac.c b/drivers/gpu/drm/nouveau/nv50_dac.c
index a0f2bebf49e3..55c56330be6d 100644
--- a/drivers/gpu/drm/nouveau/nv50_dac.c
+++ b/drivers/gpu/drm/nouveau/nv50_dac.c
@@ -190,11 +190,8 @@ nv50_dac_mode_fixup(struct drm_encoder *encoder, struct drm_display_mode *mode,
 	}
 
 	if (connector->scaling_mode != DRM_MODE_SCALE_NONE &&
-	     connector->native_mode) {
-		int id = adjusted_mode->base.id;
-		*adjusted_mode = *connector->native_mode;
-		adjusted_mode->base.id = id;
-	}
+	     connector->native_mode)
+		drm_mode_copy(adjusted_mode, connector->native_mode);
 
 	return true;
 }
diff --git a/drivers/gpu/drm/nouveau/nv50_display.c b/drivers/gpu/drm/nouveau/nv50_display.c
index 7ba28e08ee31..8b78b9cfa383 100644
--- a/drivers/gpu/drm/nouveau/nv50_display.c
+++ b/drivers/gpu/drm/nouveau/nv50_display.c
@@ -50,6 +50,29 @@ nv50_sor_nr(struct drm_device *dev)
 	return 4;
 }
 
+u32
+nv50_display_active_crtcs(struct drm_device *dev)
+{
+	struct drm_nouveau_private *dev_priv = dev->dev_private;
+	u32 mask = 0;
+	int i;
+
+	if (dev_priv->chipset  < 0x90 ||
+	    dev_priv->chipset == 0x92 ||
+	    dev_priv->chipset == 0xa0) {
+		for (i = 0; i < 2; i++)
+			mask |= nv_rd32(dev, NV50_PDISPLAY_SOR_MODE_CTRL_C(i));
+	} else {
+		for (i = 0; i < 4; i++)
+			mask |= nv_rd32(dev, NV90_PDISPLAY_SOR_MODE_CTRL_C(i));
+	}
+
+	for (i = 0; i < 3; i++)
+		mask |= nv_rd32(dev, NV50_PDISPLAY_DAC_MODE_CTRL_C(i));
+
+	return mask & 3;
+}
+
 static int
 evo_icmd(struct drm_device *dev, int ch, u32 mthd, u32 data)
 {
@@ -451,15 +474,15 @@ nv50_display_flip_next(struct drm_crtc *crtc, struct drm_framebuffer *fb,
 		}
 
 		if (dev_priv->chipset < 0xc0) {
-			BEGIN_RING(chan, NvSubSw, 0x0060, 2);
+			BEGIN_RING(chan, 0, 0x0060, 2);
 			OUT_RING  (chan, NvEvoSema0 + nv_crtc->index);
 			OUT_RING  (chan, dispc->sem.offset);
-			BEGIN_RING(chan, NvSubSw, 0x006c, 1);
+			BEGIN_RING(chan, 0, 0x006c, 1);
 			OUT_RING  (chan, 0xf00d0000 | dispc->sem.value);
-			BEGIN_RING(chan, NvSubSw, 0x0064, 2);
+			BEGIN_RING(chan, 0, 0x0064, 2);
 			OUT_RING  (chan, dispc->sem.offset ^ 0x10);
 			OUT_RING  (chan, 0x74b1e000);
-			BEGIN_RING(chan, NvSubSw, 0x0060, 1);
+			BEGIN_RING(chan, 0, 0x0060, 1);
 			if (dev_priv->chipset < 0x84)
 				OUT_RING  (chan, NvSema);
 			else
@@ -467,12 +490,12 @@ nv50_display_flip_next(struct drm_crtc *crtc, struct drm_framebuffer *fb,
 		} else {
 			u64 offset = chan->dispc_vma[nv_crtc->index].offset;
 			offset += dispc->sem.offset;
-			BEGIN_NVC0(chan, 2, NvSubM2MF, 0x0010, 4);
+			BEGIN_NVC0(chan, 2, 0, 0x0010, 4);
 			OUT_RING  (chan, upper_32_bits(offset));
 			OUT_RING  (chan, lower_32_bits(offset));
 			OUT_RING  (chan, 0xf00d0000 | dispc->sem.value);
 			OUT_RING  (chan, 0x1002);
-			BEGIN_NVC0(chan, 2, NvSubM2MF, 0x0010, 4);
+			BEGIN_NVC0(chan, 2, 0, 0x0010, 4);
 			OUT_RING  (chan, upper_32_bits(offset));
 			OUT_RING  (chan, lower_32_bits(offset ^ 0x10));
 			OUT_RING  (chan, 0x74b1e000);
@@ -840,9 +863,9 @@ nv50_display_unk20_handler(struct drm_device *dev)
 	if (type == OUTPUT_DP) {
 		int link = !(dcb->dpconf.sor.link & 1);
 		if ((mc & 0x000f0000) == 0x00020000)
-			nouveau_dp_tu_update(dev, or, link, pclk, 18);
+			nv50_sor_dp_calc_tu(dev, or, link, pclk, 18);
 		else
-			nouveau_dp_tu_update(dev, or, link, pclk, 24);
+			nv50_sor_dp_calc_tu(dev, or, link, pclk, 24);
 	}
 
 	if (dcb->type != OUTPUT_ANALOG) {
diff --git a/drivers/gpu/drm/nouveau/nv50_display.h b/drivers/gpu/drm/nouveau/nv50_display.h
index 95874f7c043c..5d3dd14d2837 100644
--- a/drivers/gpu/drm/nouveau/nv50_display.h
+++ b/drivers/gpu/drm/nouveau/nv50_display.h
@@ -74,6 +74,8 @@ void nv50_display_destroy(struct drm_device *dev);
 int nv50_crtc_blank(struct nouveau_crtc *, bool blank);
 int nv50_crtc_set_clock(struct drm_device *, int head, int pclk);
 
+u32  nv50_display_active_crtcs(struct drm_device *);
+
 int  nv50_display_sync(struct drm_device *);
 int  nv50_display_flip_next(struct drm_crtc *, struct drm_framebuffer *,
 			    struct nouveau_channel *chan);
diff --git a/drivers/gpu/drm/nouveau/nv50_evo.h b/drivers/gpu/drm/nouveau/nv50_evo.h
index 3860ca62cb19..771d879bc834 100644
--- a/drivers/gpu/drm/nouveau/nv50_evo.h
+++ b/drivers/gpu/drm/nouveau/nv50_evo.h
@@ -104,7 +104,8 @@
 #define NV50_EVO_CRTC_SCALE_CTRL_INACTIVE                            0x00000000
 #define NV50_EVO_CRTC_SCALE_CTRL_ACTIVE                              0x00000009
 #define NV50_EVO_CRTC_COLOR_CTRL                                     0x000008a8
-#define NV50_EVO_CRTC_COLOR_CTRL_COLOR                               0x00040000
+#define NV50_EVO_CRTC_COLOR_CTRL_VIBRANCE                            0x000fff00
+#define NV50_EVO_CRTC_COLOR_CTRL_HUE                                 0xfff00000
 #define NV50_EVO_CRTC_FB_POS                                         0x000008c0
 #define NV50_EVO_CRTC_REAL_RES                                       0x000008c8
 #define NV50_EVO_CRTC_SCALE_CENTER_OFFSET                            0x000008d4
diff --git a/drivers/gpu/drm/nouveau/nv50_pm.c b/drivers/gpu/drm/nouveau/nv50_pm.c
index ec5481dfcd82..d020ed4979b4 100644
--- a/drivers/gpu/drm/nouveau/nv50_pm.c
+++ b/drivers/gpu/drm/nouveau/nv50_pm.c
@@ -28,6 +28,7 @@
 #include "nouveau_hw.h"
 #include "nouveau_pm.h"
 #include "nouveau_hwsq.h"
+#include "nv50_display.h"
 
 enum clk_src {
 	clk_src_crystal,
@@ -352,17 +353,13 @@ nv50_pm_clocks_get(struct drm_device *dev, struct nouveau_pm_level *perflvl)
 }
 
 struct nv50_pm_state {
+	struct nouveau_pm_level *perflvl;
+	struct hwsq_ucode eclk_hwsq;
 	struct hwsq_ucode mclk_hwsq;
 	u32 mscript;
-
-	u32 emast;
-	u32 nctrl;
-	u32 ncoef;
-	u32 sctrl;
-	u32 scoef;
-
-	u32 amast;
-	u32 pdivs;
+	u32 mmast;
+	u32 mctrl;
+	u32 mcoef;
 };
 
 static u32
@@ -415,40 +412,153 @@ clk_same(u32 a, u32 b)
 	return ((a / 1000) == (b / 1000));
 }
 
+static void
+mclk_precharge(struct nouveau_mem_exec_func *exec)
+{
+	struct nv50_pm_state *info = exec->priv;
+	struct hwsq_ucode *hwsq = &info->mclk_hwsq;
+
+	hwsq_wr32(hwsq, 0x1002d4, 0x00000001);
+}
+
+static void
+mclk_refresh(struct nouveau_mem_exec_func *exec)
+{
+	struct nv50_pm_state *info = exec->priv;
+	struct hwsq_ucode *hwsq = &info->mclk_hwsq;
+
+	hwsq_wr32(hwsq, 0x1002d0, 0x00000001);
+}
+
+static void
+mclk_refresh_auto(struct nouveau_mem_exec_func *exec, bool enable)
+{
+	struct nv50_pm_state *info = exec->priv;
+	struct hwsq_ucode *hwsq = &info->mclk_hwsq;
+
+	hwsq_wr32(hwsq, 0x100210, enable ? 0x80000000 : 0x00000000);
+}
+
+static void
+mclk_refresh_self(struct nouveau_mem_exec_func *exec, bool enable)
+{
+	struct nv50_pm_state *info = exec->priv;
+	struct hwsq_ucode *hwsq = &info->mclk_hwsq;
+
+	hwsq_wr32(hwsq, 0x1002dc, enable ? 0x00000001 : 0x00000000);
+}
+
+static void
+mclk_wait(struct nouveau_mem_exec_func *exec, u32 nsec)
+{
+	struct nv50_pm_state *info = exec->priv;
+	struct hwsq_ucode *hwsq = &info->mclk_hwsq;
+
+	if (nsec > 1000)
+		hwsq_usec(hwsq, (nsec + 500) / 1000);
+}
+
+static u32
+mclk_mrg(struct nouveau_mem_exec_func *exec, int mr)
+{
+	if (mr <= 1)
+		return nv_rd32(exec->dev, 0x1002c0 + ((mr - 0) * 4));
+	if (mr <= 3)
+		return nv_rd32(exec->dev, 0x1002e0 + ((mr - 2) * 4));
+	return 0;
+}
+
+static void
+mclk_mrs(struct nouveau_mem_exec_func *exec, int mr, u32 data)
+{
+	struct drm_nouveau_private *dev_priv = exec->dev->dev_private;
+	struct nv50_pm_state *info = exec->priv;
+	struct hwsq_ucode *hwsq = &info->mclk_hwsq;
+
+	if (mr <= 1) {
+		if (dev_priv->vram_rank_B)
+			hwsq_wr32(hwsq, 0x1002c8 + ((mr - 0) * 4), data);
+		hwsq_wr32(hwsq, 0x1002c0 + ((mr - 0) * 4), data);
+	} else
+	if (mr <= 3) {
+		if (dev_priv->vram_rank_B)
+			hwsq_wr32(hwsq, 0x1002e8 + ((mr - 2) * 4), data);
+		hwsq_wr32(hwsq, 0x1002e0 + ((mr - 2) * 4), data);
+	}
+}
+
+static void
+mclk_clock_set(struct nouveau_mem_exec_func *exec)
+{
+	struct nv50_pm_state *info = exec->priv;
+	struct hwsq_ucode *hwsq = &info->mclk_hwsq;
+	u32 ctrl = nv_rd32(exec->dev, 0x004008);
+
+	info->mmast = nv_rd32(exec->dev, 0x00c040);
+	info->mmast &= ~0xc0000000; /* get MCLK_2 from HREF */
+	info->mmast |=  0x0000c000; /* use MCLK_2 as MPLL_BYPASS clock */
+
+	hwsq_wr32(hwsq, 0xc040, info->mmast);
+	hwsq_wr32(hwsq, 0x4008, ctrl | 0x00000200); /* bypass MPLL */
+	if (info->mctrl & 0x80000000)
+		hwsq_wr32(hwsq, 0x400c, info->mcoef);
+	hwsq_wr32(hwsq, 0x4008, info->mctrl);
+}
+
+static void
+mclk_timing_set(struct nouveau_mem_exec_func *exec)
+{
+	struct drm_device *dev = exec->dev;
+	struct nv50_pm_state *info = exec->priv;
+	struct nouveau_pm_level *perflvl = info->perflvl;
+	struct hwsq_ucode *hwsq = &info->mclk_hwsq;
+	int i;
+
+	for (i = 0; i < 9; i++) {
+		u32 reg = 0x100220 + (i * 4);
+		u32 val = nv_rd32(dev, reg);
+		if (val != perflvl->timing.reg[i])
+			hwsq_wr32(hwsq, reg, perflvl->timing.reg[i]);
+	}
+}
+
 static int
-calc_mclk(struct drm_device *dev, u32 freq, struct hwsq_ucode *hwsq)
+calc_mclk(struct drm_device *dev, struct nouveau_pm_level *perflvl,
+	  struct nv50_pm_state *info)
 {
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
+	u32 crtc_mask = nv50_display_active_crtcs(dev);
+	struct nouveau_mem_exec_func exec = {
+		.dev = dev,
+		.precharge = mclk_precharge,
+		.refresh = mclk_refresh,
+		.refresh_auto = mclk_refresh_auto,
+		.refresh_self = mclk_refresh_self,
+		.wait = mclk_wait,
+		.mrg = mclk_mrg,
+		.mrs = mclk_mrs,
+		.clock_set = mclk_clock_set,
+		.timing_set = mclk_timing_set,
+		.priv = info
+	};
+	struct hwsq_ucode *hwsq = &info->mclk_hwsq;
 	struct pll_lims pll;
-	u32 mast = nv_rd32(dev, 0x00c040);
-	u32 ctrl = nv_rd32(dev, 0x004008);
-	u32 coef = nv_rd32(dev, 0x00400c);
-	u32 orig = ctrl;
-	u32 crtc_mask = 0;
 	int N, M, P;
-	int ret, i;
+	int ret;
 
 	/* use pcie refclock if possible, otherwise use mpll */
-	ctrl &= ~0x81ff0200;
-	if (clk_same(freq, read_clk(dev, clk_src_href))) {
-		ctrl |= 0x00000200 | (pll.log2p_bias << 19);
+	info->mctrl  = nv_rd32(dev, 0x004008);
+	info->mctrl &= ~0x81ff0200;
+	if (clk_same(perflvl->memory, read_clk(dev, clk_src_href))) {
+		info->mctrl |= 0x00000200 | (pll.log2p_bias << 19);
 	} else {
-		ret = calc_pll(dev, 0x4008, &pll, freq, &N, &M, &P);
+		ret = calc_pll(dev, 0x4008, &pll, perflvl->memory, &N, &M, &P);
 		if (ret == 0)
 			return -EINVAL;
 
-		ctrl |= 0x80000000 | (P << 22) | (P << 16);
-		ctrl |= pll.log2p_bias << 19;
-		coef  = (N << 8) | M;
-	}
-
-	mast &= ~0xc0000000; /* get MCLK_2 from HREF */
-	mast |=  0x0000c000; /* use MCLK_2 as MPLL_BYPASS clock */
-
-	/* determine active crtcs */
-	for (i = 0; i < 2; i++) {
-		if (nv_rd32(dev, NV50_PDISPLAY_CRTC_C(i, CLOCK)))
-			crtc_mask |= (1 << i);
+		info->mctrl |= 0x80000000 | (P << 22) | (P << 16);
+		info->mctrl |= pll.log2p_bias << 19;
+		info->mcoef  = (N << 8) | M;
 	}
 
 	/* build the ucode which will reclock the memory for us */
@@ -462,25 +572,10 @@ calc_mclk(struct drm_device *dev, u32 freq, struct hwsq_ucode *hwsq)
 	hwsq_setf(hwsq, 0x10, 0); /* disable bus access */
 	hwsq_op5f(hwsq, 0x00, 0x01); /* no idea :s */
 
-	/* prepare memory controller */
-	hwsq_wr32(hwsq, 0x1002d4, 0x00000001); /* precharge banks and idle */
-	hwsq_wr32(hwsq, 0x1002d0, 0x00000001); /* force refresh */
-	hwsq_wr32(hwsq, 0x100210, 0x00000000); /* stop the automatic refresh */
-	hwsq_wr32(hwsq, 0x1002dc, 0x00000001); /* start self refresh mode */
-
-	/* reclock memory */
-	hwsq_wr32(hwsq, 0xc040, mast);
-	hwsq_wr32(hwsq, 0x4008, orig | 0x00000200); /* bypass MPLL */
-	hwsq_wr32(hwsq, 0x400c, coef);
-	hwsq_wr32(hwsq, 0x4008, ctrl);
-
-	/* restart memory controller */
-	hwsq_wr32(hwsq, 0x1002d4, 0x00000001); /* precharge banks and idle */
-	hwsq_wr32(hwsq, 0x1002dc, 0x00000000); /* stop self refresh mode */
-	hwsq_wr32(hwsq, 0x100210, 0x80000000); /* restart automatic refresh */
-	hwsq_usec(hwsq, 12); /* wait for the PLL to stabilize */
-
-	hwsq_usec(hwsq, 48); /* may be unnecessary: causes flickering */
+	ret = nouveau_mem_exec(&exec, perflvl);
+	if (ret)
+		return ret;
+
 	hwsq_setf(hwsq, 0x10, 1); /* enable bus access */
 	hwsq_op5f(hwsq, 0x00, 0x00); /* no idea, reverse of 0x00, 0x01? */
 	if (dev_priv->chipset >= 0x92)
@@ -494,10 +589,11 @@ nv50_pm_clocks_pre(struct drm_device *dev, struct nouveau_pm_level *perflvl)
 {
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
 	struct nv50_pm_state *info;
+	struct hwsq_ucode *hwsq;
 	struct pll_lims pll;
+	u32 out, mast, divs, ctrl;
 	int clk, ret = -EINVAL;
 	int N, M, P1, P2;
-	u32 out;
 
 	if (dev_priv->chipset == 0xaa ||
 	    dev_priv->chipset == 0xac)
@@ -506,54 +602,44 @@ nv50_pm_clocks_pre(struct drm_device *dev, struct nouveau_pm_level *perflvl)
 	info = kmalloc(sizeof(*info), GFP_KERNEL);
 	if (!info)
 		return ERR_PTR(-ENOMEM);
+	info->perflvl = perflvl;
 
-	/* core: for the moment at least, always use nvpll */
-	clk = calc_pll(dev, 0x4028, &pll, perflvl->core, &N, &M, &P1);
-	if (clk == 0)
-		goto error;
+	/* memory: build hwsq ucode which we'll use to reclock memory.
+	 *         use pcie refclock if possible, otherwise use mpll */
+	info->mclk_hwsq.len = 0;
+	if (perflvl->memory) {
+		ret = calc_mclk(dev, perflvl, info);
+		if (ret)
+			goto error;
+		info->mscript = perflvl->memscript;
+	}
 
-	info->emast = 0x00000003;
-	info->nctrl = 0x80000000 | (P1 << 19) | (P1 << 16);
-	info->ncoef = (N << 8) | M;
+	divs = read_div(dev);
+	mast = info->mmast;
 
-	/* shader: tie to nvclk if possible, otherwise use spll.  have to be
-	 * very careful that the shader clock is at least twice the core, or
-	 * some chipsets will be very unhappy.  i expect most or all of these
-	 * cases will be handled by tying to nvclk, but it's possible there's
-	 * corners
-	 */
-	if (P1-- && perflvl->shader == (perflvl->core << 1)) {
-		info->emast |= 0x00000020;
-		info->sctrl  = 0x00000000 | (P1 << 19) | (P1 << 16);
-		info->scoef  = nv_rd32(dev, 0x004024);
-	} else {
-		clk = calc_pll(dev, 0x4020, &pll, perflvl->shader, &N, &M, &P1);
-		if (clk == 0)
-			goto error;
+	/* start building HWSQ script for engine reclocking */
+	hwsq = &info->eclk_hwsq;
+	hwsq_init(hwsq);
+	hwsq_setf(hwsq, 0x10, 0); /* disable bus access */
+	hwsq_op5f(hwsq, 0x00, 0x01); /* wait for access disabled? */
 
-		info->emast |= 0x00000030;
-		info->sctrl  = 0x80000000 | (P1 << 19) | (P1 << 16);
-		info->scoef  = (N << 8) | M;
+	/* vdec/dom6: switch to "safe" clocks temporarily */
+	if (perflvl->vdec) {
+		mast &= ~0x00000c00;
+		divs &= ~0x00000700;
 	}
 
-	/* memory: build hwsq ucode which we'll use to reclock memory */
-	info->mclk_hwsq.len = 0;
-	if (perflvl->memory) {
-		clk = calc_mclk(dev, perflvl->memory, &info->mclk_hwsq);
-		if (clk < 0) {
-			ret = clk;
-			goto error;
-		}
-
-		info->mscript = perflvl->memscript;
+	if (perflvl->dom6) {
+		mast &= ~0x0c000000;
+		divs &= ~0x00000007;
 	}
 
+	hwsq_wr32(hwsq, 0x00c040, mast);
+
 	/* vdec: avoid modifying xpll until we know exactly how the other
 	 * clock domains work, i suspect at least some of them can also be
 	 * tied to xpll...
 	 */
-	info->amast = nv_rd32(dev, 0x00c040);
-	info->pdivs = read_div(dev);
 	if (perflvl->vdec) {
 		/* see how close we can get using nvclk as a source */
 		clk = calc_div(perflvl->core, perflvl->vdec, &P1);
@@ -566,16 +652,14 @@ nv50_pm_clocks_pre(struct drm_device *dev, struct nouveau_pm_level *perflvl)
 		out = calc_div(out, perflvl->vdec, &P2);
 
 		/* select whichever gets us closest */
-		info->amast &= ~0x00000c00;
-		info->pdivs &= ~0x00000700;
 		if (abs((int)perflvl->vdec - clk) <=
 		    abs((int)perflvl->vdec - out)) {
 			if (dev_priv->chipset != 0x98)
-				info->amast |= 0x00000c00;
-			info->pdivs |= P1 << 8;
+				mast |= 0x00000c00;
+			divs |= P1 << 8;
 		} else {
-			info->amast |= 0x00000800;
-			info->pdivs |= P2 << 8;
+			mast |= 0x00000800;
+			divs |= P2 << 8;
 		}
 	}
 
@@ -583,21 +667,82 @@ nv50_pm_clocks_pre(struct drm_device *dev, struct nouveau_pm_level *perflvl)
 	 * of the host clock frequency
 	 */
 	if (perflvl->dom6) {
-		info->amast &= ~0x0c000000;
 		if (clk_same(perflvl->dom6, read_clk(dev, clk_src_href))) {
-			info->amast |= 0x00000000;
+			mast |= 0x00000000;
 		} else
 		if (clk_same(perflvl->dom6, read_clk(dev, clk_src_hclk))) {
-			info->amast |= 0x08000000;
+			mast |= 0x08000000;
 		} else {
 			clk = read_clk(dev, clk_src_hclk) * 3;
 			clk = calc_div(clk, perflvl->dom6, &P1);
 
-			info->amast |= 0x0c000000;
-			info->pdivs  = (info->pdivs & ~0x00000007) | P1;
+			mast |= 0x0c000000;
+			divs |= P1;
 		}
 	}
 
+	/* vdec/dom6: complete switch to new clocks */
+	switch (dev_priv->chipset) {
+	case 0x92:
+	case 0x94:
+	case 0x96:
+		hwsq_wr32(hwsq, 0x004800, divs);
+		break;
+	default:
+		hwsq_wr32(hwsq, 0x004700, divs);
+		break;
+	}
+
+	hwsq_wr32(hwsq, 0x00c040, mast);
+
+	/* core/shader: make sure sclk/nvclk are disconnected from their
+	 * PLLs (nvclk to dom6, sclk to hclk)
+	 */
+	if (dev_priv->chipset < 0x92)
+		mast = (mast & ~0x001000b0) | 0x00100080;
+	else
+		mast = (mast & ~0x000000b3) | 0x00000081;
+
+	hwsq_wr32(hwsq, 0x00c040, mast);
+
+	/* core: for the moment at least, always use nvpll */
+	clk = calc_pll(dev, 0x4028, &pll, perflvl->core, &N, &M, &P1);
+	if (clk == 0)
+		goto error;
+
+	ctrl  = nv_rd32(dev, 0x004028) & ~0xc03f0100;
+	mast &= ~0x00100000;
+	mast |= 3;
+
+	hwsq_wr32(hwsq, 0x004028, 0x80000000 | (P1 << 19) | (P1 << 16) | ctrl);
+	hwsq_wr32(hwsq, 0x00402c, (N << 8) | M);
+
+	/* shader: tie to nvclk if possible, otherwise use spll.  have to be
+	 * very careful that the shader clock is at least twice the core, or
+	 * some chipsets will be very unhappy.  i expect most or all of these
+	 * cases will be handled by tying to nvclk, but it's possible there's
+	 * corners
+	 */
+	ctrl = nv_rd32(dev, 0x004020) & ~0xc03f0100;
+
+	if (P1-- && perflvl->shader == (perflvl->core << 1)) {
+		hwsq_wr32(hwsq, 0x004020, (P1 << 19) | (P1 << 16) | ctrl);
+		hwsq_wr32(hwsq, 0x00c040, 0x00000020 | mast);
+	} else {
+		clk = calc_pll(dev, 0x4020, &pll, perflvl->shader, &N, &M, &P1);
+		if (clk == 0)
+			goto error;
+		ctrl |= 0x80000000;
+
+		hwsq_wr32(hwsq, 0x004020, (P1 << 19) | (P1 << 16) | ctrl);
+		hwsq_wr32(hwsq, 0x004024, (N << 8) | M);
+		hwsq_wr32(hwsq, 0x00c040, 0x00000030 | mast);
+	}
+
+	hwsq_setf(hwsq, 0x10, 1); /* enable bus access */
+	hwsq_op5f(hwsq, 0x00, 0x00); /* wait for access enabled? */
+	hwsq_fini(hwsq);
+
 	return info;
 error:
 	kfree(info);
@@ -605,23 +750,24 @@ error:
 }
 
 static int
-prog_mclk(struct drm_device *dev, struct hwsq_ucode *hwsq)
+prog_hwsq(struct drm_device *dev, struct hwsq_ucode *hwsq)
 {
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
 	u32 hwsq_data, hwsq_kick;
 	int i;
 
-	if (dev_priv->chipset < 0x90) {
+	if (dev_priv->chipset < 0x94) {
 		hwsq_data = 0x001400;
 		hwsq_kick = 0x00000003;
 	} else {
 		hwsq_data = 0x080000;
 		hwsq_kick = 0x00000001;
 	}
-
 	/* upload hwsq ucode */
 	nv_mask(dev, 0x001098, 0x00000008, 0x00000000);
 	nv_wr32(dev, 0x001304, 0x00000000);
+	if (dev_priv->chipset >= 0x92)
+		nv_wr32(dev, 0x001318, 0x00000000);
 	for (i = 0; i < hwsq->len / 4; i++)
 		nv_wr32(dev, hwsq_data + (i * 4), hwsq->ptr.u32[i]);
 	nv_mask(dev, 0x001098, 0x00000018, 0x00000018);
@@ -645,20 +791,19 @@ prog_mclk(struct drm_device *dev, struct hwsq_ucode *hwsq)
 int
 nv50_pm_clocks_set(struct drm_device *dev, void *data)
 {
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
 	struct nv50_pm_state *info = data;
 	struct bit_entry M;
-	int ret = 0;
+	int ret = -EBUSY;
 
 	/* halt and idle execution engines */
 	nv_mask(dev, 0x002504, 0x00000001, 0x00000001);
 	if (!nv_wait(dev, 0x002504, 0x00000010, 0x00000010))
-		goto error;
+		goto resume;
+	if (!nv_wait(dev, 0x00251c, 0x0000003f, 0x0000003f))
+		goto resume;
 
-	/* memory: it is *very* important we change this first, the ucode
-	 * we build in pre() now has hardcoded 0xc040 values, which can't
-	 * change before we execute it or the engine clocks may end up
-	 * messed up.
+	/* program memory clock, if necessary - must come before engine clock
+	 * reprogramming due to how we construct the hwsq scripts in pre()
 	 */
 	if (info->mclk_hwsq.len) {
 		/* execute some scripts that do ??? from the vbios.. */
@@ -672,42 +817,14 @@ nv50_pm_clocks_set(struct drm_device *dev, void *data)
 			nouveau_bios_init_exec(dev, info->mscript);
 		}
 
-		ret = prog_mclk(dev, &info->mclk_hwsq);
+		ret = prog_hwsq(dev, &info->mclk_hwsq);
 		if (ret)
 			goto resume;
 	}
 
-	/* reclock vdec/dom6 */
-	nv_mask(dev, 0x00c040, 0x00000c00, 0x00000000);
-	switch (dev_priv->chipset) {
-	case 0x92:
-	case 0x94:
-	case 0x96:
-		nv_mask(dev, 0x004800, 0x00000707, info->pdivs);
-		break;
-	default:
-		nv_mask(dev, 0x004700, 0x00000707, info->pdivs);
-		break;
-	}
-	nv_mask(dev, 0x00c040, 0x0c000c00, info->amast);
+	/* program engine clocks */
+	ret = prog_hwsq(dev, &info->eclk_hwsq);
 
-	/* core/shader: make sure sclk/nvclk are disconnected from their
-	 * plls (nvclk to dom6, sclk to hclk), modify the plls, and
-	 * reconnect sclk/nvclk to their new clock source
-	 */
-	if (dev_priv->chipset < 0x92)
-		nv_mask(dev, 0x00c040, 0x001000b0, 0x00100080); /* grrr! */
-	else
-		nv_mask(dev, 0x00c040, 0x000000b3, 0x00000081);
-	nv_mask(dev, 0x004020, 0xc03f0100, info->sctrl);
-	nv_wr32(dev, 0x004024, info->scoef);
-	nv_mask(dev, 0x004028, 0xc03f0100, info->nctrl);
-	nv_wr32(dev, 0x00402c, info->ncoef);
-	nv_mask(dev, 0x00c040, 0x00100033, info->emast);
-
-	goto resume;
-error:
-	ret = -EBUSY;
 resume:
 	nv_mask(dev, 0x002504, 0x00000001, 0x00000000);
 	kfree(info);
diff --git a/drivers/gpu/drm/nouveau/nv50_sor.c b/drivers/gpu/drm/nouveau/nv50_sor.c
index c4423ba9c9bf..a7844ab6a50c 100644
--- a/drivers/gpu/drm/nouveau/nv50_sor.c
+++ b/drivers/gpu/drm/nouveau/nv50_sor.c
@@ -36,6 +36,193 @@
 #include "nouveau_crtc.h"
 #include "nv50_display.h"
 
+static u32
+nv50_sor_dp_lane_map(struct drm_device *dev, struct dcb_entry *dcb, u8 lane)
+{
+	struct drm_nouveau_private *dev_priv = dev->dev_private;
+	static const u8 nvaf[] = { 24, 16, 8, 0 }; /* thanks, apple.. */
+	static const u8 nv50[] = { 16, 8, 0, 24 };
+	if (dev_priv->card_type == 0xaf)
+		return nvaf[lane];
+	return nv50[lane];
+}
+
+static void
+nv50_sor_dp_train_set(struct drm_device *dev, struct dcb_entry *dcb, u8 pattern)
+{
+	u32 or = ffs(dcb->or) - 1, link = !(dcb->sorconf.link & 1);
+	nv_mask(dev, NV50_SOR_DP_CTRL(or, link), 0x0f000000, pattern << 24);
+}
+
+static void
+nv50_sor_dp_train_adj(struct drm_device *dev, struct dcb_entry *dcb,
+		      u8 lane, u8 swing, u8 preem)
+{
+	u32 or = ffs(dcb->or) - 1, link = !(dcb->sorconf.link & 1);
+	u32 shift = nv50_sor_dp_lane_map(dev, dcb, lane);
+	u32 mask = 0x000000ff << shift;
+	u8 *table, *entry, *config;
+
+	table = nouveau_dp_bios_data(dev, dcb, &entry);
+	if (!table || (table[0] != 0x20 && table[0] != 0x21)) {
+		NV_ERROR(dev, "PDISP: unsupported DP table for chipset\n");
+		return;
+	}
+
+	config = entry + table[4];
+	while (config[0] != swing || config[1] != preem) {
+		config += table[5];
+		if (config >= entry + table[4] + entry[4] * table[5])
+			return;
+	}
+
+	nv_mask(dev, NV50_SOR_DP_UNK118(or, link), mask, config[2] << shift);
+	nv_mask(dev, NV50_SOR_DP_UNK120(or, link), mask, config[3] << shift);
+	nv_mask(dev, NV50_SOR_DP_UNK130(or, link), 0x0000ff00, config[4] << 8);
+}
+
+static void
+nv50_sor_dp_link_set(struct drm_device *dev, struct dcb_entry *dcb, int crtc,
+		     int link_nr, u32 link_bw, bool enhframe)
+{
+	u32 or = ffs(dcb->or) - 1, link = !(dcb->sorconf.link & 1);
+	u32 dpctrl = nv_rd32(dev, NV50_SOR_DP_CTRL(or, link)) & ~0x001f4000;
+	u32 clksor = nv_rd32(dev, 0x614300 + (or * 0x800)) & ~0x000c0000;
+	u8 *table, *entry, mask;
+	int i;
+
+	table = nouveau_dp_bios_data(dev, dcb, &entry);
+	if (!table || (table[0] != 0x20 && table[0] != 0x21)) {
+		NV_ERROR(dev, "PDISP: unsupported DP table for chipset\n");
+		return;
+	}
+
+	entry = ROMPTR(dev, entry[10]);
+	if (entry) {
+		while (link_bw < ROM16(entry[0]) * 10)
+			entry += 4;
+
+		nouveau_bios_run_init_table(dev, ROM16(entry[2]), dcb, crtc);
+	}
+
+	dpctrl |= ((1 << link_nr) - 1) << 16;
+	if (enhframe)
+		dpctrl |= 0x00004000;
+
+	if (link_bw > 162000)
+		clksor |= 0x00040000;
+
+	nv_wr32(dev, 0x614300 + (or * 0x800), clksor);
+	nv_wr32(dev, NV50_SOR_DP_CTRL(or, link), dpctrl);
+
+	mask = 0;
+	for (i = 0; i < link_nr; i++)
+		mask |= 1 << (nv50_sor_dp_lane_map(dev, dcb, i) >> 3);
+	nv_mask(dev, NV50_SOR_DP_UNK130(or, link), 0x0000000f, mask);
+}
+
+static void
+nv50_sor_dp_link_get(struct drm_device *dev, u32 or, u32 link, u32 *nr, u32 *bw)
+{
+	u32 dpctrl = nv_rd32(dev, NV50_SOR_DP_CTRL(or, link)) & 0x000f0000;
+	u32 clksor = nv_rd32(dev, 0x614300 + (or * 0x800));
+	if (clksor & 0x000c0000)
+		*bw = 270000;
+	else
+		*bw = 162000;
+
+	if      (dpctrl > 0x00030000) *nr = 4;
+	else if (dpctrl > 0x00010000) *nr = 2;
+	else			      *nr = 1;
+}
+
+void
+nv50_sor_dp_calc_tu(struct drm_device *dev, int or, int link, u32 clk, u32 bpp)
+{
+	const u32 symbol = 100000;
+	int bestTU = 0, bestVTUi = 0, bestVTUf = 0, bestVTUa = 0;
+	int TU, VTUi, VTUf, VTUa;
+	u64 link_data_rate, link_ratio, unk;
+	u32 best_diff = 64 * symbol;
+	u32 link_nr, link_bw, r;
+
+	/* calculate packed data rate for each lane */
+	nv50_sor_dp_link_get(dev, or, link, &link_nr, &link_bw);
+	link_data_rate = (clk * bpp / 8) / link_nr;
+
+	/* calculate ratio of packed data rate to link symbol rate */
+	link_ratio = link_data_rate * symbol;
+	r = do_div(link_ratio, link_bw);
+
+	for (TU = 64; TU >= 32; TU--) {
+		/* calculate average number of valid symbols in each TU */
+		u32 tu_valid = link_ratio * TU;
+		u32 calc, diff;
+
+		/* find a hw representation for the fraction.. */
+		VTUi = tu_valid / symbol;
+		calc = VTUi * symbol;
+		diff = tu_valid - calc;
+		if (diff) {
+			if (diff >= (symbol / 2)) {
+				VTUf = symbol / (symbol - diff);
+				if (symbol - (VTUf * diff))
+					VTUf++;
+
+				if (VTUf <= 15) {
+					VTUa  = 1;
+					calc += symbol - (symbol / VTUf);
+				} else {
+					VTUa  = 0;
+					VTUf  = 1;
+					calc += symbol;
+				}
+			} else {
+				VTUa  = 0;
+				VTUf  = min((int)(symbol / diff), 15);
+				calc += symbol / VTUf;
+			}
+
+			diff = calc - tu_valid;
+		} else {
+			/* no remainder, but the hw doesn't like the fractional
+			 * part to be zero.  decrement the integer part and
+			 * have the fraction add a whole symbol back
+			 */
+			VTUa = 0;
+			VTUf = 1;
+			VTUi--;
+		}
+
+		if (diff < best_diff) {
+			best_diff = diff;
+			bestTU = TU;
+			bestVTUa = VTUa;
+			bestVTUf = VTUf;
+			bestVTUi = VTUi;
+			if (diff == 0)
+				break;
+		}
+	}
+
+	if (!bestTU) {
+		NV_ERROR(dev, "DP: unable to find suitable config\n");
+		return;
+	}
+
+	/* XXX close to vbios numbers, but not right */
+	unk  = (symbol - link_ratio) * bestTU;
+	unk *= link_ratio;
+	r = do_div(unk, symbol);
+	r = do_div(unk, symbol);
+	unk += 6;
+
+	nv_mask(dev, NV50_SOR_DP_CTRL(or, link), 0x000001fc, bestTU << 2);
+	nv_mask(dev, NV50_SOR_DP_SCFG(or, link), 0x010f7f3f, bestVTUa << 24 |
+							     bestVTUf << 16 |
+							     bestVTUi << 8 |
+							     unk);
+}
 static void
 nv50_sor_disconnect(struct drm_encoder *encoder)
 {
@@ -117,20 +304,13 @@ nv50_sor_dpms(struct drm_encoder *encoder, int mode)
 	}
 
 	if (nv_encoder->dcb->type == OUTPUT_DP) {
-		struct nouveau_i2c_chan *auxch;
-
-		auxch = nouveau_i2c_find(dev, nv_encoder->dcb->i2c_index);
-		if (!auxch)
-			return;
+		struct dp_train_func func = {
+			.link_set = nv50_sor_dp_link_set,
+			.train_set = nv50_sor_dp_train_set,
+			.train_adj = nv50_sor_dp_train_adj
+		};
 
-		if (mode == DRM_MODE_DPMS_ON) {
-			u8 status = DP_SET_POWER_D0;
-			nouveau_dp_auxch(auxch, 8, DP_SET_POWER, &status, 1);
-			nouveau_dp_link_train(encoder, nv_encoder->dp.datarate);
-		} else {
-			u8 status = DP_SET_POWER_D3;
-			nouveau_dp_auxch(auxch, 8, DP_SET_POWER, &status, 1);
-		}
+		nouveau_dp_dpms(encoder, mode, nv_encoder->dp.datarate, &func);
 	}
 }
 
@@ -162,11 +342,8 @@ nv50_sor_mode_fixup(struct drm_encoder *encoder, struct drm_display_mode *mode,
 	}
 
 	if (connector->scaling_mode != DRM_MODE_SCALE_NONE &&
-	     connector->native_mode) {
-		int id = adjusted_mode->base.id;
-		*adjusted_mode = *connector->native_mode;
-		adjusted_mode->base.id = id;
-	}
+	     connector->native_mode)
+		drm_mode_copy(adjusted_mode, connector->native_mode);
 
 	return true;
 }
diff --git a/drivers/gpu/drm/nouveau/nv50_vm.c b/drivers/gpu/drm/nouveau/nv50_vm.c
index 6f38ceae3aa4..44fbac9c7d93 100644
--- a/drivers/gpu/drm/nouveau/nv50_vm.c
+++ b/drivers/gpu/drm/nouveau/nv50_vm.c
@@ -57,27 +57,15 @@ nv50_vm_map_pgt(struct nouveau_gpuobj *pgd, u32 pde,
 }
 
 static inline u64
-nv50_vm_addr(struct nouveau_vma *vma, u64 phys, u32 memtype, u32 target)
+vm_addr(struct nouveau_vma *vma, u64 phys, u32 memtype, u32 target)
 {
-	struct drm_nouveau_private *dev_priv = vma->vm->dev->dev_private;
-
 	phys |= 1; /* present */
 	phys |= (u64)memtype << 40;
-
-	/* IGPs don't have real VRAM, re-target to stolen system memory */
-	if (target == 0 && dev_priv->vram_sys_base) {
-		phys  += dev_priv->vram_sys_base;
-		target = 3;
-	}
-
 	phys |= target << 4;
-
 	if (vma->access & NV_MEM_ACCESS_SYS)
 		phys |= (1 << 6);
-
 	if (!(vma->access & NV_MEM_ACCESS_WO))
 		phys |= (1 << 3);
-
 	return phys;
 }
 
@@ -85,11 +73,19 @@ void
 nv50_vm_map(struct nouveau_vma *vma, struct nouveau_gpuobj *pgt,
 	    struct nouveau_mem *mem, u32 pte, u32 cnt, u64 phys, u64 delta)
 {
+	struct drm_nouveau_private *dev_priv = vma->vm->dev->dev_private;
 	u32 comp = (mem->memtype & 0x180) >> 7;
-	u32 block;
+	u32 block, target;
 	int i;
 
-	phys  = nv50_vm_addr(vma, phys, mem->memtype, 0);
+	/* IGPs don't have real VRAM, re-target to stolen system memory */
+	target = 0;
+	if (dev_priv->vram_sys_base) {
+		phys += dev_priv->vram_sys_base;
+		target = 3;
+	}
+
+	phys  = vm_addr(vma, phys, mem->memtype, target);
 	pte <<= 3;
 	cnt <<= 3;
 
@@ -125,9 +121,10 @@ void
 nv50_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) ? 3 : 2;
 	pte <<= 3;
 	while (cnt--) {
-		u64 phys = nv50_vm_addr(vma, (u64)*list++, mem->memtype, 2);
+		u64 phys = vm_addr(vma, (u64)*list++, mem->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/nv50_vram.c b/drivers/gpu/drm/nouveau/nv50_vram.c
index 2e45e57fd869..9ed9ae397d75 100644
--- a/drivers/gpu/drm/nouveau/nv50_vram.c
+++ b/drivers/gpu/drm/nouveau/nv50_vram.c
@@ -189,8 +189,25 @@ nv50_vram_init(struct drm_device *dev)
 	struct nouveau_vram_engine *vram = &dev_priv->engine.vram;
 	const u32 rsvd_head = ( 256 * 1024) >> 12; /* vga memory */
 	const u32 rsvd_tail = (1024 * 1024) >> 12; /* vbios etc */
+	u32 pfb714 = nv_rd32(dev, 0x100714);
 	u32 rblock, length;
 
+	switch (pfb714 & 0x00000007) {
+	case 0: dev_priv->vram_type = NV_MEM_TYPE_DDR1; break;
+	case 1:
+		if (nouveau_mem_vbios_type(dev) == NV_MEM_TYPE_DDR3)
+			dev_priv->vram_type = NV_MEM_TYPE_DDR3;
+		else
+			dev_priv->vram_type = NV_MEM_TYPE_DDR2;
+		break;
+	case 2: dev_priv->vram_type = NV_MEM_TYPE_GDDR3; break;
+	case 3: dev_priv->vram_type = NV_MEM_TYPE_GDDR4; break;
+	case 4: dev_priv->vram_type = NV_MEM_TYPE_GDDR5; break;
+	default:
+		break;
+	}
+
+	dev_priv->vram_rank_B = !!(nv_rd32(dev, 0x100200) & 0x4);
 	dev_priv->vram_size  = nv_rd32(dev, 0x10020c);
 	dev_priv->vram_size |= (dev_priv->vram_size & 0xff) << 32;
 	dev_priv->vram_size &= 0xffffffff00ULL;
diff --git a/drivers/gpu/drm/nouveau/nvc0_fifo.c b/drivers/gpu/drm/nouveau/nvc0_fifo.c
index dcbe0d5d0241..50d68a7a1379 100644
--- a/drivers/gpu/drm/nouveau/nvc0_fifo.c
+++ b/drivers/gpu/drm/nouveau/nvc0_fifo.c
@@ -436,6 +436,24 @@ nvc0_fifo_isr_vm_fault(struct drm_device *dev, int unit)
 	printk(" on channel 0x%010llx\n", (u64)inst << 12);
 }
 
+static int
+nvc0_fifo_page_flip(struct drm_device *dev, u32 chid)
+{
+	struct drm_nouveau_private *dev_priv = dev->dev_private;
+	struct nouveau_channel *chan = NULL;
+	unsigned long flags;
+	int ret = -EINVAL;
+
+	spin_lock_irqsave(&dev_priv->channels.lock, flags);
+	if (likely(chid >= 0 && chid < dev_priv->engine.fifo.channels)) {
+		chan = dev_priv->channels.ptr[chid];
+		if (likely(chan))
+			ret = nouveau_finish_page_flip(chan, NULL);
+	}
+	spin_unlock_irqrestore(&dev_priv->channels.lock, flags);
+	return ret;
+}
+
 static void
 nvc0_fifo_isr_subfifo_intr(struct drm_device *dev, int unit)
 {
@@ -445,11 +463,21 @@ nvc0_fifo_isr_subfifo_intr(struct drm_device *dev, int unit)
 	u32 chid = nv_rd32(dev, 0x040120 + (unit * 0x2000)) & 0x7f;
 	u32 subc = (addr & 0x00070000);
 	u32 mthd = (addr & 0x00003ffc);
+	u32 show = stat;
 
-	NV_INFO(dev, "PSUBFIFO %d:", unit);
-	nouveau_bitfield_print(nvc0_fifo_subfifo_intr, stat);
-	NV_INFO(dev, "PSUBFIFO %d: ch %d subc %d mthd 0x%04x data 0x%08x\n",
-		unit, chid, subc, mthd, data);
+	if (stat & 0x00200000) {
+		if (mthd == 0x0054) {
+			if (!nvc0_fifo_page_flip(dev, chid))
+				show &= ~0x00200000;
+		}
+	}
+
+	if (show) {
+		NV_INFO(dev, "PFIFO%d:", unit);
+		nouveau_bitfield_print(nvc0_fifo_subfifo_intr, show);
+		NV_INFO(dev, "PFIFO%d: ch %d subc %d mthd 0x%04x data 0x%08x\n",
+			     unit, chid, subc, mthd, data);
+	}
 
 	nv_wr32(dev, 0x0400c0 + (unit * 0x2000), 0x80600008);
 	nv_wr32(dev, 0x040108 + (unit * 0x2000), stat);
diff --git a/drivers/gpu/drm/nouveau/nvc0_graph.c b/drivers/gpu/drm/nouveau/nvc0_graph.c
index 8ee3963f9030..9066102d1159 100644
--- a/drivers/gpu/drm/nouveau/nvc0_graph.c
+++ b/drivers/gpu/drm/nouveau/nvc0_graph.c
@@ -333,14 +333,6 @@ nvc0_graph_fini(struct drm_device *dev, int engine, bool suspend)
 	return 0;
 }
 
-static int
-nvc0_graph_mthd_page_flip(struct nouveau_channel *chan,
-			  u32 class, u32 mthd, u32 data)
-{
-	nouveau_finish_page_flip(chan, NULL);
-	return 0;
-}
-
 static void
 nvc0_graph_init_obj418880(struct drm_device *dev)
 {
@@ -889,7 +881,6 @@ nvc0_graph_create(struct drm_device *dev)
 
 	NVOBJ_CLASS(dev, 0x902d, GR); /* 2D */
 	NVOBJ_CLASS(dev, 0x9039, GR); /* M2MF */
-	NVOBJ_MTHD (dev, 0x9039, 0x0500, nvc0_graph_mthd_page_flip);
 	NVOBJ_CLASS(dev, 0x9097, GR); /* 3D */
 	if (fermi >= 0x9197)
 		NVOBJ_CLASS(dev, 0x9197, GR); /* 3D (NVC1-) */
diff --git a/drivers/gpu/drm/nouveau/nvc0_pm.c b/drivers/gpu/drm/nouveau/nvc0_pm.c
index e9992f62c1c0..ce65f81bb871 100644
--- a/drivers/gpu/drm/nouveau/nvc0_pm.c
+++ b/drivers/gpu/drm/nouveau/nvc0_pm.c
@@ -269,7 +269,7 @@ calc_clk(struct drm_device *dev, int clk, struct nvc0_pm_clock *info, u32 freq)
 	clk0 = calc_div(dev, clk, clk0, freq, &div1D);
 
 	/* see if we can get any closer using PLLs */
-	if (clk0 != freq) {
+	if (clk0 != freq && (0x00004387 & (1 << clk))) {
 		if (clk < 7)
 			clk1 = calc_pll(dev, clk, freq, &info->coef);
 		else
diff --git a/drivers/gpu/drm/nouveau/nvc0_vm.c b/drivers/gpu/drm/nouveau/nvc0_vm.c
index 9e352944a35a..30d2bd58828f 100644
--- a/drivers/gpu/drm/nouveau/nvc0_vm.c
+++ b/drivers/gpu/drm/nouveau/nvc0_vm.c
@@ -77,9 +77,11 @@ void
 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;
+
 	pte <<= 3;
 	while (cnt--) {
-		u64 phys = nvc0_vm_addr(vma, *list++, mem->memtype, 5);
+		u64 phys = nvc0_vm_addr(vma, *list++, mem->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/nvc0_vram.c b/drivers/gpu/drm/nouveau/nvc0_vram.c
index ce984d573a51..a7eef8934c07 100644
--- a/drivers/gpu/drm/nouveau/nvc0_vram.c
+++ b/drivers/gpu/drm/nouveau/nvc0_vram.c
@@ -106,31 +106,32 @@ nvc0_vram_init(struct drm_device *dev)
 	struct nouveau_vram_engine *vram = &dev_priv->engine.vram;
 	const u32 rsvd_head = ( 256 * 1024) >> 12; /* vga memory */
 	const u32 rsvd_tail = (1024 * 1024) >> 12; /* vbios etc */
-	u32 parts = nv_rd32(dev, 0x121c74);
+	u32 parts = nv_rd32(dev, 0x022438);
+	u32 pmask = nv_rd32(dev, 0x022554);
 	u32 bsize = nv_rd32(dev, 0x10f20c);
 	u32 offset, length;
 	bool uniform = true;
 	int ret, part;
 
 	NV_DEBUG(dev, "0x100800: 0x%08x\n", nv_rd32(dev, 0x100800));
-	NV_DEBUG(dev, "parts 0x%08x bcast_mem_amount 0x%08x\n", parts, bsize);
+	NV_DEBUG(dev, "parts 0x%08x mask 0x%08x\n", parts, pmask);
+
+	dev_priv->vram_type = nouveau_mem_vbios_type(dev);
+	dev_priv->vram_rank_B = !!(nv_rd32(dev, 0x10f200) & 0x00000004);
 
 	/* read amount of vram attached to each memory controller */
-	part = 0;
-	while (parts) {
-		u32 psize = nv_rd32(dev, 0x11020c + (part++ * 0x1000));
-		if (psize == 0)
-			continue;
-		parts--;
-
-		if (psize != bsize) {
-			if (psize < bsize)
-				bsize = psize;
-			uniform = false;
+	for (part = 0; part < parts; part++) {
+		if (!(pmask & (1 << part))) {
+			u32 psize = nv_rd32(dev, 0x11020c + (part * 0x1000));
+			if (psize != bsize) {
+				if (psize < bsize)
+					bsize = psize;
+				uniform = false;
+			}
+
+			NV_DEBUG(dev, "%d: mem_amount 0x%08x\n", part, psize);
+			dev_priv->vram_size += (u64)psize << 20;
 		}
-
-		NV_DEBUG(dev, "%d: mem_amount 0x%08x\n", part, psize);
-		dev_priv->vram_size += (u64)psize << 20;
 	}
 
 	/* if all controllers have the same amount attached, there's no holes */
diff --git a/drivers/gpu/drm/nouveau/nvd0_display.c b/drivers/gpu/drm/nouveau/nvd0_display.c
index d2ba2f07400b..0247250939e8 100644
--- a/drivers/gpu/drm/nouveau/nvd0_display.c
+++ b/drivers/gpu/drm/nouveau/nvd0_display.c
@@ -284,6 +284,8 @@ nvd0_display_flip_next(struct drm_crtc *crtc, struct drm_framebuffer *fb,
 	u32 *push;
 	int ret;
 
+	evo_sync(crtc->dev, EVO_MASTER);
+
 	swap_interval <<= 4;
 	if (swap_interval == 0)
 		swap_interval |= 0x100;
@@ -301,12 +303,12 @@ nvd0_display_flip_next(struct drm_crtc *crtc, struct drm_framebuffer *fb,
 		offset  = chan->dispc_vma[nv_crtc->index].offset;
 		offset += evo->sem.offset;
 
-		BEGIN_NVC0(chan, 2, NvSubM2MF, 0x0010, 4);
+		BEGIN_NVC0(chan, 2, 0, NV84_SUBCHAN_SEMAPHORE_ADDRESS_HIGH, 4);
 		OUT_RING  (chan, upper_32_bits(offset));
 		OUT_RING  (chan, lower_32_bits(offset));
 		OUT_RING  (chan, 0xf00d0000 | evo->sem.value);
 		OUT_RING  (chan, 0x1002);
-		BEGIN_NVC0(chan, 2, NvSubM2MF, 0x0010, 4);
+		BEGIN_NVC0(chan, 2, 0, NV84_SUBCHAN_SEMAPHORE_ADDRESS_HIGH, 4);
 		OUT_RING  (chan, upper_32_bits(offset));
 		OUT_RING  (chan, lower_32_bits(offset ^ 0x10));
 		OUT_RING  (chan, 0x74b1e000);
@@ -361,10 +363,12 @@ nvd0_display_flip_next(struct drm_crtc *crtc, struct drm_framebuffer *fb,
 static int
 nvd0_crtc_set_dither(struct nouveau_crtc *nv_crtc, bool update)
 {
+	struct drm_nouveau_private *dev_priv = nv_crtc->base.dev->dev_private;
 	struct drm_device *dev = nv_crtc->base.dev;
 	struct nouveau_connector *nv_connector;
 	struct drm_connector *connector;
 	u32 *push, mode = 0x00;
+	u32 mthd;
 
 	nv_connector = nouveau_crtc_connector_get(nv_crtc);
 	connector = &nv_connector->base;
@@ -382,9 +386,14 @@ nvd0_crtc_set_dither(struct nouveau_crtc *nv_crtc, bool update)
 		mode |= nv_connector->dithering_depth;
 	}
 
+	if (dev_priv->card_type < NV_E0)
+		mthd = 0x0490 + (nv_crtc->index * 0x0300);
+	else
+		mthd = 0x04a0 + (nv_crtc->index * 0x0300);
+
 	push = evo_wait(dev, EVO_MASTER, 4);
 	if (push) {
-		evo_mthd(push, 0x0490 + (nv_crtc->index * 0x300), 1);
+		evo_mthd(push, mthd, 1);
 		evo_data(push, mode);
 		if (update) {
 			evo_mthd(push, 0x0080, 1);
@@ -593,7 +602,7 @@ nvd0_crtc_commit(struct drm_crtc *crtc)
 		evo_kick(push, crtc->dev, EVO_MASTER);
 	}
 
-	nvd0_crtc_cursor_show(nv_crtc, nv_crtc->cursor.visible, false);
+	nvd0_crtc_cursor_show(nv_crtc, nv_crtc->cursor.visible, true);
 	nvd0_display_flip_next(crtc, crtc->fb, NULL, 1);
 }
 
@@ -634,8 +643,7 @@ nvd0_crtc_mode_set(struct drm_crtc *crtc, struct drm_display_mode *umode,
 	u32 hactive, hsynce, hbackp, hfrontp, hblanke, hblanks;
 	u32 vactive, vsynce, vbackp, vfrontp, vblanke, vblanks;
 	u32 vblan2e = 0, vblan2s = 1;
-	u32 magic = 0x31ec6000;
-	u32 syncs, *push;
+	u32 *push;
 	int ret;
 
 	hactive = mode->htotal;
@@ -655,15 +663,8 @@ nvd0_crtc_mode_set(struct drm_crtc *crtc, struct drm_display_mode *umode,
 		vblan2e = vactive + vsynce + vbackp;
 		vblan2s = vblan2e + (mode->vdisplay * vscan / ilace);
 		vactive = (vactive * 2) + 1;
-		magic  |= 0x00000001;
 	}
 
-	syncs = 0x00000001;
-	if (mode->flags & DRM_MODE_FLAG_NHSYNC)
-		syncs |= 0x00000008;
-	if (mode->flags & DRM_MODE_FLAG_NVSYNC)
-		syncs |= 0x00000010;
-
 	ret = nvd0_crtc_swap_fbs(crtc, old_fb);
 	if (ret)
 		return ret;
@@ -683,9 +684,6 @@ nvd0_crtc_mode_set(struct drm_crtc *crtc, struct drm_display_mode *umode,
 		evo_data(push, mode->clock * 1000);
 		evo_data(push, 0x00200000); /* ??? */
 		evo_data(push, mode->clock * 1000);
-		evo_mthd(push, 0x0404 + (nv_crtc->index * 0x300), 2);
-		evo_data(push, syncs);
-		evo_data(push, magic);
 		evo_mthd(push, 0x04d0 + (nv_crtc->index * 0x300), 2);
 		evo_data(push, 0x00000311);
 		evo_data(push, 0x00000100);
@@ -959,11 +957,6 @@ nvd0_dac_mode_fixup(struct drm_encoder *encoder, struct drm_display_mode *mode,
 }
 
 static void
-nvd0_dac_prepare(struct drm_encoder *encoder)
-{
-}
-
-static void
 nvd0_dac_commit(struct drm_encoder *encoder)
 {
 }
@@ -974,13 +967,26 @@ nvd0_dac_mode_set(struct drm_encoder *encoder, struct drm_display_mode *mode,
 {
 	struct nouveau_encoder *nv_encoder = nouveau_encoder(encoder);
 	struct nouveau_crtc *nv_crtc = nouveau_crtc(encoder->crtc);
-	u32 *push;
+	u32 syncs, magic, *push;
+
+	syncs = 0x00000001;
+	if (mode->flags & DRM_MODE_FLAG_NHSYNC)
+		syncs |= 0x00000008;
+	if (mode->flags & DRM_MODE_FLAG_NVSYNC)
+		syncs |= 0x00000010;
+
+	magic = 0x31ec6000 | (nv_crtc->index << 25);
+	if (mode->flags & DRM_MODE_FLAG_INTERLACE)
+		magic |= 0x00000001;
 
 	nvd0_dac_dpms(encoder, DRM_MODE_DPMS_ON);
 
-	push = evo_wait(encoder->dev, EVO_MASTER, 4);
+	push = evo_wait(encoder->dev, EVO_MASTER, 8);
 	if (push) {
-		evo_mthd(push, 0x0180 + (nv_encoder->or * 0x20), 2);
+		evo_mthd(push, 0x0404 + (nv_crtc->index * 0x300), 2);
+		evo_data(push, syncs);
+		evo_data(push, magic);
+		evo_mthd(push, 0x0180 + (nv_encoder->or * 0x020), 2);
 		evo_data(push, 1 << nv_crtc->index);
 		evo_data(push, 0x00ff);
 		evo_kick(push, encoder->dev, EVO_MASTER);
@@ -1043,7 +1049,7 @@ nvd0_dac_destroy(struct drm_encoder *encoder)
 static const struct drm_encoder_helper_funcs nvd0_dac_hfunc = {
 	.dpms = nvd0_dac_dpms,
 	.mode_fixup = nvd0_dac_mode_fixup,
-	.prepare = nvd0_dac_prepare,
+	.prepare = nvd0_dac_disconnect,
 	.commit = nvd0_dac_commit,
 	.mode_set = nvd0_dac_mode_set,
 	.disable = nvd0_dac_disconnect,
@@ -1183,6 +1189,149 @@ nvd0_hdmi_disconnect(struct drm_encoder *encoder)
 /******************************************************************************
  * SOR
  *****************************************************************************/
+static inline u32
+nvd0_sor_dp_lane_map(struct drm_device *dev, struct dcb_entry *dcb, u8 lane)
+{
+	static const u8 nvd0[] = { 16, 8, 0, 24 };
+	return nvd0[lane];
+}
+
+static void
+nvd0_sor_dp_train_set(struct drm_device *dev, struct dcb_entry *dcb, u8 pattern)
+{
+	const u32 or = ffs(dcb->or) - 1, link = !(dcb->sorconf.link & 1);
+	const u32 loff = (or * 0x800) + (link * 0x80);
+	nv_mask(dev, 0x61c110 + loff, 0x0f0f0f0f, 0x01010101 * pattern);
+}
+
+static void
+nvd0_sor_dp_train_adj(struct drm_device *dev, struct dcb_entry *dcb,
+		      u8 lane, u8 swing, u8 preem)
+{
+	const u32 or = ffs(dcb->or) - 1, link = !(dcb->sorconf.link & 1);
+	const u32 loff = (or * 0x800) + (link * 0x80);
+	u32 shift = nvd0_sor_dp_lane_map(dev, dcb, lane);
+	u32 mask = 0x000000ff << shift;
+	u8 *table, *entry, *config = NULL;
+
+	switch (swing) {
+	case 0: preem += 0; break;
+	case 1: preem += 4; break;
+	case 2: preem += 7; break;
+	case 3: preem += 9; break;
+	}
+
+	table = nouveau_dp_bios_data(dev, dcb, &entry);
+	if (table) {
+		if (table[0] == 0x30) {
+			config  = entry + table[4];
+			config += table[5] * preem;
+		} else
+		if (table[0] == 0x40) {
+			config  = table + table[1];
+			config += table[2] * table[3];
+			config += table[6] * preem;
+		}
+	}
+
+	if (!config) {
+		NV_ERROR(dev, "PDISP: unsupported DP table for chipset\n");
+		return;
+	}
+
+	nv_mask(dev, 0x61c118 + loff, mask, config[1] << shift);
+	nv_mask(dev, 0x61c120 + loff, mask, config[2] << shift);
+	nv_mask(dev, 0x61c130 + loff, 0x0000ff00, config[3] << 8);
+	nv_mask(dev, 0x61c13c + loff, 0x00000000, 0x00000000);
+}
+
+static void
+nvd0_sor_dp_link_set(struct drm_device *dev, struct dcb_entry *dcb, int crtc,
+		     int link_nr, u32 link_bw, bool enhframe)
+{
+	const u32 or = ffs(dcb->or) - 1, link = !(dcb->sorconf.link & 1);
+	const u32 loff = (or * 0x800) + (link * 0x80);
+	const u32 soff = (or * 0x800);
+	u32 dpctrl = nv_rd32(dev, 0x61c10c + loff) & ~0x001f4000;
+	u32 clksor = nv_rd32(dev, 0x612300 + soff) & ~0x007c0000;
+	u32 script = 0x0000, lane_mask = 0;
+	u8 *table, *entry;
+	int i;
+
+	link_bw /= 27000;
+
+	table = nouveau_dp_bios_data(dev, dcb, &entry);
+	if (table) {
+		if      (table[0] == 0x30) entry = ROMPTR(dev, entry[10]);
+		else if (table[0] == 0x40) entry = ROMPTR(dev, entry[9]);
+		else                       entry = NULL;
+
+		while (entry) {
+			if (entry[0] >= link_bw)
+				break;
+			entry += 3;
+		}
+
+		nouveau_bios_run_init_table(dev, script, dcb, crtc);
+	}
+
+	clksor |= link_bw << 18;
+	dpctrl |= ((1 << link_nr) - 1) << 16;
+	if (enhframe)
+		dpctrl |= 0x00004000;
+
+	for (i = 0; i < link_nr; i++)
+		lane_mask |= 1 << (nvd0_sor_dp_lane_map(dev, dcb, i) >> 3);
+
+	nv_wr32(dev, 0x612300 + soff, clksor);
+	nv_wr32(dev, 0x61c10c + loff, dpctrl);
+	nv_mask(dev, 0x61c130 + loff, 0x0000000f, lane_mask);
+}
+
+static void
+nvd0_sor_dp_link_get(struct drm_device *dev, struct dcb_entry *dcb,
+		     u32 *link_nr, u32 *link_bw)
+{
+	const u32 or = ffs(dcb->or) - 1, link = !(dcb->sorconf.link & 1);
+	const u32 loff = (or * 0x800) + (link * 0x80);
+	const u32 soff = (or * 0x800);
+	u32 dpctrl = nv_rd32(dev, 0x61c10c + loff) & 0x000f0000;
+	u32 clksor = nv_rd32(dev, 0x612300 + soff);
+
+	if      (dpctrl > 0x00030000) *link_nr = 4;
+	else if (dpctrl > 0x00010000) *link_nr = 2;
+	else			      *link_nr = 1;
+
+	*link_bw  = (clksor & 0x007c0000) >> 18;
+	*link_bw *= 27000;
+}
+
+static void
+nvd0_sor_dp_calc_tu(struct drm_device *dev, struct dcb_entry *dcb,
+		    u32 crtc, u32 datarate)
+{
+	const u32 symbol = 100000;
+	const u32 TU = 64;
+	u32 link_nr, link_bw;
+	u64 ratio, value;
+
+	nvd0_sor_dp_link_get(dev, dcb, &link_nr, &link_bw);
+
+	ratio  = datarate;
+	ratio *= symbol;
+	do_div(ratio, link_nr * link_bw);
+
+	value  = (symbol - ratio) * TU;
+	value *= ratio;
+	do_div(value, symbol);
+	do_div(value, symbol);
+
+	value += 5;
+	value |= 0x08000000;
+
+	nv_wr32(dev, 0x616610 + (crtc * 0x800), value);
+}
+
 static void
 nvd0_sor_dpms(struct drm_encoder *encoder, int mode)
 {
@@ -1215,6 +1364,16 @@ nvd0_sor_dpms(struct drm_encoder *encoder, int mode)
 	nv_mask(dev, 0x61c004 + (or * 0x0800), 0x80000001, dpms_ctrl);
 	nv_wait(dev, 0x61c004 + (or * 0x0800), 0x80000000, 0x00000000);
 	nv_wait(dev, 0x61c030 + (or * 0x0800), 0x10000000, 0x00000000);
+
+	if (nv_encoder->dcb->type == OUTPUT_DP) {
+		struct dp_train_func func = {
+			.link_set = nvd0_sor_dp_link_set,
+			.train_set = nvd0_sor_dp_train_set,
+			.train_adj = nvd0_sor_dp_train_adj
+		};
+
+		nouveau_dp_dpms(encoder, mode, nv_encoder->dp.datarate, &func);
+	}
 }
 
 static bool
@@ -1237,8 +1396,37 @@ nvd0_sor_mode_fixup(struct drm_encoder *encoder, struct drm_display_mode *mode,
 }
 
 static void
+nvd0_sor_disconnect(struct drm_encoder *encoder)
+{
+	struct nouveau_encoder *nv_encoder = nouveau_encoder(encoder);
+	struct drm_device *dev = encoder->dev;
+	u32 *push;
+
+	if (nv_encoder->crtc) {
+		nvd0_crtc_prepare(nv_encoder->crtc);
+
+		push = evo_wait(dev, EVO_MASTER, 4);
+		if (push) {
+			evo_mthd(push, 0x0200 + (nv_encoder->or * 0x20), 1);
+			evo_data(push, 0x00000000);
+			evo_mthd(push, 0x0080, 1);
+			evo_data(push, 0x00000000);
+			evo_kick(push, dev, EVO_MASTER);
+		}
+
+		nvd0_hdmi_disconnect(encoder);
+
+		nv_encoder->crtc = NULL;
+		nv_encoder->last_dpms = DRM_MODE_DPMS_OFF;
+	}
+}
+
+static void
 nvd0_sor_prepare(struct drm_encoder *encoder)
 {
+	nvd0_sor_disconnect(encoder);
+	if (nouveau_encoder(encoder)->dcb->type == OUTPUT_DP)
+		evo_sync(encoder->dev, EVO_MASTER);
 }
 
 static void
@@ -1257,7 +1445,18 @@ nvd0_sor_mode_set(struct drm_encoder *encoder, struct drm_display_mode *umode,
 	struct nouveau_connector *nv_connector;
 	struct nvbios *bios = &dev_priv->vbios;
 	u32 mode_ctrl = (1 << nv_crtc->index);
-	u32 *push, or_config;
+	u32 syncs, magic, *push;
+	u32 or_config;
+
+	syncs = 0x00000001;
+	if (mode->flags & DRM_MODE_FLAG_NHSYNC)
+		syncs |= 0x00000008;
+	if (mode->flags & DRM_MODE_FLAG_NVSYNC)
+		syncs |= 0x00000010;
+
+	magic = 0x31ec6000 | (nv_crtc->index << 25);
+	if (mode->flags & DRM_MODE_FLAG_INTERLACE)
+		magic |= 0x00000001;
 
 	nv_connector = nouveau_encoder_connector_get(nv_encoder);
 	switch (nv_encoder->dcb->type) {
@@ -1306,6 +1505,22 @@ nvd0_sor_mode_set(struct drm_encoder *encoder, struct drm_display_mode *umode,
 
 		}
 		break;
+	case OUTPUT_DP:
+		if (nv_connector->base.display_info.bpc == 6) {
+			nv_encoder->dp.datarate = mode->clock * 18 / 8;
+			syncs |= 0x00000140;
+		} else {
+			nv_encoder->dp.datarate = mode->clock * 24 / 8;
+			syncs |= 0x00000180;
+		}
+
+		if (nv_encoder->dcb->sorconf.link & 1)
+			mode_ctrl |= 0x00000800;
+		else
+			mode_ctrl |= 0x00000900;
+
+		or_config = (mode_ctrl & 0x00000f00) >> 8;
+		break;
 	default:
 		BUG_ON(1);
 		break;
@@ -1313,9 +1528,17 @@ nvd0_sor_mode_set(struct drm_encoder *encoder, struct drm_display_mode *umode,
 
 	nvd0_sor_dpms(encoder, DRM_MODE_DPMS_ON);
 
-	push = evo_wait(dev, EVO_MASTER, 4);
+	if (nv_encoder->dcb->type == OUTPUT_DP) {
+		nvd0_sor_dp_calc_tu(dev, nv_encoder->dcb, nv_crtc->index,
+					 nv_encoder->dp.datarate);
+	}
+
+	push = evo_wait(dev, EVO_MASTER, 8);
 	if (push) {
-		evo_mthd(push, 0x0200 + (nv_encoder->or * 0x20), 2);
+		evo_mthd(push, 0x0404 + (nv_crtc->index * 0x300), 2);
+		evo_data(push, syncs);
+		evo_data(push, magic);
+		evo_mthd(push, 0x0200 + (nv_encoder->or * 0x020), 2);
 		evo_data(push, mode_ctrl);
 		evo_data(push, or_config);
 		evo_kick(push, dev, EVO_MASTER);
@@ -1325,32 +1548,6 @@ nvd0_sor_mode_set(struct drm_encoder *encoder, struct drm_display_mode *umode,
 }
 
 static void
-nvd0_sor_disconnect(struct drm_encoder *encoder)
-{
-	struct nouveau_encoder *nv_encoder = nouveau_encoder(encoder);
-	struct drm_device *dev = encoder->dev;
-	u32 *push;
-
-	if (nv_encoder->crtc) {
-		nvd0_crtc_prepare(nv_encoder->crtc);
-
-		push = evo_wait(dev, EVO_MASTER, 4);
-		if (push) {
-			evo_mthd(push, 0x0200 + (nv_encoder->or * 0x20), 1);
-			evo_data(push, 0x00000000);
-			evo_mthd(push, 0x0080, 1);
-			evo_data(push, 0x00000000);
-			evo_kick(push, dev, EVO_MASTER);
-		}
-
-		nvd0_hdmi_disconnect(encoder);
-
-		nv_encoder->crtc = NULL;
-		nv_encoder->last_dpms = DRM_MODE_DPMS_OFF;
-	}
-}
-
-static void
 nvd0_sor_destroy(struct drm_encoder *encoder)
 {
 	drm_encoder_cleanup(encoder);
@@ -1402,17 +1599,19 @@ static struct dcb_entry *
 lookup_dcb(struct drm_device *dev, int id, u32 mc)
 {
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	int type, or, i;
+	int type, or, i, link = -1;
 
 	if (id < 4) {
 		type = OUTPUT_ANALOG;
 		or   = id;
 	} else {
 		switch (mc & 0x00000f00) {
-		case 0x00000000: type = OUTPUT_LVDS; break;
-		case 0x00000100: type = OUTPUT_TMDS; break;
-		case 0x00000200: type = OUTPUT_TMDS; break;
-		case 0x00000500: type = OUTPUT_TMDS; break;
+		case 0x00000000: link = 0; type = OUTPUT_LVDS; break;
+		case 0x00000100: link = 0; type = OUTPUT_TMDS; break;
+		case 0x00000200: link = 1; type = OUTPUT_TMDS; break;
+		case 0x00000500: link = 0; type = OUTPUT_TMDS; break;
+		case 0x00000800: link = 0; type = OUTPUT_DP; break;
+		case 0x00000900: link = 1; type = OUTPUT_DP; break;
 		default:
 			NV_ERROR(dev, "PDISP: unknown SOR mc 0x%08x\n", mc);
 			return NULL;
@@ -1423,7 +1622,8 @@ lookup_dcb(struct drm_device *dev, int id, u32 mc)
 
 	for (i = 0; i < dev_priv->vbios.dcb.entries; i++) {
 		struct dcb_entry *dcb = &dev_priv->vbios.dcb.entry[i];
-		if (dcb->type == type && (dcb->or & (1 << or)))
+		if (dcb->type == type && (dcb->or & (1 << or)) &&
+		    (link < 0 || link == !(dcb->sorconf.link & 1)))
 			return dcb;
 	}
 
@@ -1474,7 +1674,9 @@ nvd0_display_unk2_handler(struct drm_device *dev, u32 crtc, u32 mask)
 	}
 
 	pclk = nv_rd32(dev, 0x660450 + (crtc * 0x300)) / 1000;
-	if (mask & 0x00010000) {
+	NV_DEBUG_KMS(dev, "PDISP: crtc %d pclk %d mask 0x%08x\n",
+			  crtc, pclk, mask);
+	if (pclk && (mask & 0x00010000)) {
 		nv50_crtc_set_clock(dev, crtc, pclk);
 	}
 
@@ -1498,6 +1700,7 @@ nvd0_display_unk2_handler(struct drm_device *dev, u32 crtc, u32 mask)
 			break;
 		case OUTPUT_TMDS:
 		case OUTPUT_LVDS:
+		case OUTPUT_DP:
 			if (cfg & 0x00000100)
 				tmp = 0x00000101;
 			else
@@ -1548,7 +1751,7 @@ nvd0_display_bh(unsigned long data)
 {
 	struct drm_device *dev = (struct drm_device *)data;
 	struct nvd0_display *disp = nvd0_display(dev);
-	u32 mask, crtc;
+	u32 mask = 0, crtc = ~0;
 	int i;
 
 	if (drm_debug & (DRM_UT_DRIVER | DRM_UT_KMS)) {
@@ -1564,12 +1767,8 @@ nvd0_display_bh(unsigned long data)
 		}
 	}
 
-	mask = nv_rd32(dev, 0x6101d4);
-	crtc = 0;
-	if (!mask) {
-		mask = nv_rd32(dev, 0x6109d4);
-		crtc = 1;
-	}
+	while (!mask && ++crtc < dev->mode_config.num_crtc)
+		mask = nv_rd32(dev, 0x6101d4 + (crtc * 0x800));
 
 	if (disp->modeset & 0x00000001)
 		nvd0_display_unk1_handler(dev, crtc, mask);
@@ -1584,6 +1783,7 @@ nvd0_display_intr(struct drm_device *dev)
 {
 	struct nvd0_display *disp = nvd0_display(dev);
 	u32 intr = nv_rd32(dev, 0x610088);
+	int i;
 
 	if (intr & 0x00000001) {
 		u32 stat = nv_rd32(dev, 0x61008c);
@@ -1628,16 +1828,13 @@ nvd0_display_intr(struct drm_device *dev)
 		intr &= ~0x00100000;
 	}
 
-	if (intr & 0x01000000) {
-		u32 stat = nv_rd32(dev, 0x6100bc);
-		nv_wr32(dev, 0x6100bc, stat);
-		intr &= ~0x01000000;
-	}
-
-	if (intr & 0x02000000) {
-		u32 stat = nv_rd32(dev, 0x6108bc);
-		nv_wr32(dev, 0x6108bc, stat);
-		intr &= ~0x02000000;
+	for (i = 0; i < dev->mode_config.num_crtc; i++) {
+		u32 mask = 0x01000000 << i;
+		if (intr & mask) {
+			u32 stat = nv_rd32(dev, 0x6100bc + (i * 0x800));
+			nv_wr32(dev, 0x6100bc + (i * 0x800), stat);
+			intr &= ~mask;
+		}
 	}
 
 	if (intr)
@@ -1774,7 +1971,7 @@ nvd0_display_create(struct drm_device *dev)
 	struct pci_dev *pdev = dev->pdev;
 	struct nvd0_display *disp;
 	struct dcb_entry *dcbe;
-	int ret, i;
+	int crtcs, ret, i;
 
 	disp = kzalloc(sizeof(*disp), GFP_KERNEL);
 	if (!disp)
@@ -1782,7 +1979,8 @@ nvd0_display_create(struct drm_device *dev)
 	dev_priv->engine.display.priv = disp;
 
 	/* create crtc objects to represent the hw heads */
-	for (i = 0; i < 2; i++) {
+	crtcs = nv_rd32(dev, 0x022448);
+	for (i = 0; i < crtcs; i++) {
 		ret = nvd0_crtc_create(dev, i);
 		if (ret)
 			goto out;
@@ -1803,6 +2001,7 @@ nvd0_display_create(struct drm_device *dev)
 		switch (dcbe->type) {
 		case OUTPUT_TMDS:
 		case OUTPUT_LVDS:
+		case OUTPUT_DP:
 			nvd0_sor_create(connector, dcbe);
 			break;
 		case OUTPUT_ANALOG:
diff --git a/drivers/gpu/drm/r128/r128_drv.c b/drivers/gpu/drm/r128/r128_drv.c
index 6a5f4395838f..88718fad5d6d 100644
--- a/drivers/gpu/drm/r128/r128_drv.c
+++ b/drivers/gpu/drm/r128/r128_drv.c
@@ -85,6 +85,7 @@ static struct drm_driver driver = {
 
 int r128_driver_load(struct drm_device *dev, unsigned long flags)
 {
+	pci_set_master(dev->pdev);
 	return drm_vblank_init(dev, 1);
 }
 
diff --git a/drivers/gpu/drm/radeon/Makefile b/drivers/gpu/drm/radeon/Makefile
index 2139fe893ec5..9d83729956ff 100644
--- a/drivers/gpu/drm/radeon/Makefile
+++ b/drivers/gpu/drm/radeon/Makefile
@@ -71,7 +71,7 @@ radeon-y += radeon_device.o radeon_asic.o radeon_kms.o \
 	r600_blit_kms.o radeon_pm.o atombios_dp.o r600_audio.o r600_hdmi.o \
 	evergreen.o evergreen_cs.o evergreen_blit_shaders.o evergreen_blit_kms.o \
 	radeon_trace_points.o ni.o cayman_blit_shaders.o atombios_encoders.o \
-	radeon_semaphore.o radeon_sa.o
+	radeon_semaphore.o radeon_sa.o atombios_i2c.o si.o si_blit_shaders.o
 
 radeon-$(CONFIG_COMPAT) += radeon_ioc32.o
 radeon-$(CONFIG_VGA_SWITCHEROO) += radeon_atpx_handler.o
diff --git a/drivers/gpu/drm/radeon/ObjectID.h b/drivers/gpu/drm/radeon/ObjectID.h
index c61c3fe9fb98..ca4b038050d2 100644
--- a/drivers/gpu/drm/radeon/ObjectID.h
+++ b/drivers/gpu/drm/radeon/ObjectID.h
@@ -85,6 +85,7 @@
 #define ENCODER_OBJECT_ID_INTERNAL_KLDSCP_LVTMA   0x1F
 #define ENCODER_OBJECT_ID_INTERNAL_UNIPHY1        0x20
 #define ENCODER_OBJECT_ID_INTERNAL_UNIPHY2        0x21
+#define ENCODER_OBJECT_ID_INTERNAL_VCE            0x24
 
 #define ENCODER_OBJECT_ID_GENERAL_EXTERNAL_DVO    0xFF
 
@@ -387,6 +388,10 @@
                                                   GRAPH_OBJECT_ENUM_ID1 << ENUM_ID_SHIFT |\
                                                   ENCODER_OBJECT_ID_NUTMEG << OBJECT_ID_SHIFT)
 
+#define ENCODER_VCE_ENUM_ID1                     ( GRAPH_OBJECT_TYPE_ENCODER << OBJECT_TYPE_SHIFT |\
+                                                  GRAPH_OBJECT_ENUM_ID1 << ENUM_ID_SHIFT |\
+                                                  ENCODER_OBJECT_ID_INTERNAL_VCE << OBJECT_ID_SHIFT)
+
 /****************************************************/
 /* Connector Object ID definition - Shared with BIOS */
 /****************************************************/
diff --git a/drivers/gpu/drm/radeon/atom.c b/drivers/gpu/drm/radeon/atom.c
index d1bd239cd9e9..5ce9bf51a8de 100644
--- a/drivers/gpu/drm/radeon/atom.c
+++ b/drivers/gpu/drm/radeon/atom.c
@@ -1306,8 +1306,11 @@ struct atom_context *atom_parse(struct card_info *card, void *bios)
 
 int atom_asic_init(struct atom_context *ctx)
 {
+	struct radeon_device *rdev = ctx->card->dev->dev_private;
 	int hwi = CU16(ctx->data_table + ATOM_DATA_FWI_PTR);
 	uint32_t ps[16];
+	int ret;
+
 	memset(ps, 0, 64);
 
 	ps[0] = cpu_to_le32(CU32(hwi + ATOM_FWI_DEFSCLK_PTR));
@@ -1317,7 +1320,17 @@ int atom_asic_init(struct atom_context *ctx)
 
 	if (!CU16(ctx->cmd_table + 4 + 2 * ATOM_CMD_INIT))
 		return 1;
-	return atom_execute_table(ctx, ATOM_CMD_INIT, ps);
+	ret = atom_execute_table(ctx, ATOM_CMD_INIT, ps);
+	if (ret)
+		return ret;
+
+	memset(ps, 0, 64);
+
+	if (rdev->family < CHIP_R600) {
+		if (CU16(ctx->cmd_table + 4 + 2 * ATOM_CMD_SPDFANCNTL))
+			atom_execute_table(ctx, ATOM_CMD_SPDFANCNTL, ps);
+	}
+	return ret;
 }
 
 void atom_destroy(struct atom_context *ctx)
diff --git a/drivers/gpu/drm/radeon/atom.h b/drivers/gpu/drm/radeon/atom.h
index 93cfe2086ba0..25fea631dad2 100644
--- a/drivers/gpu/drm/radeon/atom.h
+++ b/drivers/gpu/drm/radeon/atom.h
@@ -44,6 +44,7 @@
 #define ATOM_CMD_SETSCLK	0x0A
 #define ATOM_CMD_SETMCLK	0x0B
 #define ATOM_CMD_SETPCLK	0x0C
+#define ATOM_CMD_SPDFANCNTL	0x39
 
 #define ATOM_DATA_FWI_PTR	0xC
 #define ATOM_DATA_IIO_PTR	0x32
diff --git a/drivers/gpu/drm/radeon/atombios.h b/drivers/gpu/drm/radeon/atombios.h
index 1b50ad8919d5..4b04ba3828e8 100644
--- a/drivers/gpu/drm/radeon/atombios.h
+++ b/drivers/gpu/drm/radeon/atombios.h
@@ -101,6 +101,7 @@
 #define ATOM_LCD_SELFTEST_START									(ATOM_DISABLE+5)
 #define ATOM_LCD_SELFTEST_STOP									(ATOM_ENABLE+5)
 #define ATOM_ENCODER_INIT			                  (ATOM_DISABLE+7)
+#define ATOM_INIT			                          (ATOM_DISABLE+7)
 #define ATOM_GET_STATUS                         (ATOM_DISABLE+8)
 
 #define ATOM_BLANKING         1
@@ -251,25 +252,25 @@ typedef struct _ATOM_MASTER_LIST_OF_COMMAND_TABLES{
   USHORT SetEngineClock;                         //Function Table,directly used by various SW components,latest version 1.1
   USHORT SetMemoryClock;                         //Function Table,directly used by various SW components,latest version 1.1
   USHORT SetPixelClock;                          //Function Table,directly used by various SW components,latest version 1.2  
-  USHORT DynamicClockGating;                     //Atomic Table,  indirectly used by various SW components,called from ASIC_Init
+  USHORT EnableDispPowerGating;                  //Atomic Table,  indirectly used by various SW components,called from ASIC_Init
   USHORT ResetMemoryDLL;                         //Atomic Table,  indirectly used by various SW components,called from SetMemoryClock
   USHORT ResetMemoryDevice;                      //Atomic Table,  indirectly used by various SW components,called from SetMemoryClock
-  USHORT MemoryPLLInit;
-  USHORT AdjustDisplayPll;												//only used by Bios
+  USHORT MemoryPLLInit;                          //Atomic Table,  used only by Bios
+  USHORT AdjustDisplayPll;											 //Atomic Table,  used by various SW componentes. 
   USHORT AdjustMemoryController;                 //Atomic Table,  indirectly used by various SW components,called from SetMemoryClock                
   USHORT EnableASIC_StaticPwrMgt;                //Atomic Table,  only used by Bios
   USHORT ASIC_StaticPwrMgtStatusChange;          //Obsolete ,     only used by Bios   
   USHORT DAC_LoadDetection;                      //Atomic Table,  directly used by various SW components,latest version 1.2  
   USHORT LVTMAEncoderControl;                    //Atomic Table,directly used by various SW components,latest version 1.3
-  USHORT LCD1OutputControl;                      //Atomic Table,  directly used by various SW components,latest version 1.1 
+  USHORT HW_Misc_Operation;                      //Atomic Table,  directly used by various SW components,latest version 1.1 
   USHORT DAC1EncoderControl;                     //Atomic Table,  directly used by various SW components,latest version 1.1  
   USHORT DAC2EncoderControl;                     //Atomic Table,  directly used by various SW components,latest version 1.1 
   USHORT DVOOutputControl;                       //Atomic Table,  directly used by various SW components,latest version 1.1 
   USHORT CV1OutputControl;                       //Atomic Table,  Atomic Table,  Obsolete from Ry6xx, use DAC2 Output instead 
-  USHORT GetConditionalGoldenSetting;            //only used by Bios
+  USHORT GetConditionalGoldenSetting;            //Only used by Bios
   USHORT TVEncoderControl;                       //Function Table,directly used by various SW components,latest version 1.1
-  USHORT TMDSAEncoderControl;                    //Atomic Table,  directly used by various SW components,latest version 1.3
-  USHORT LVDSEncoderControl;                     //Atomic Table,  directly used by various SW components,latest version 1.3
+  USHORT PatchMCSetting;                         //only used by BIOS
+  USHORT MC_SEQ_Control;                         //only used by BIOS
   USHORT TV1OutputControl;                       //Atomic Table,  Obsolete from Ry6xx, use DAC2 Output instead
   USHORT EnableScaler;                           //Atomic Table,  used only by Bios
   USHORT BlankCRTC;                              //Atomic Table,  directly used by various SW components,latest version 1.1 
@@ -282,7 +283,7 @@ typedef struct _ATOM_MASTER_LIST_OF_COMMAND_TABLES{
   USHORT SetCRTC_Replication;                    //Atomic Table,  used only by Bios
   USHORT SelectCRTC_Source;                      //Atomic Table,  directly used by various SW components,latest version 1.1 
   USHORT EnableGraphSurfaces;                    //Atomic Table,  used only by Bios
-  USHORT UpdateCRTC_DoubleBufferRegisters;
+  USHORT UpdateCRTC_DoubleBufferRegisters;			 //Atomic Table,  used only by Bios
   USHORT LUT_AutoFill;                           //Atomic Table,  only used by Bios
   USHORT EnableHW_IconCursor;                    //Atomic Table,  only used by Bios
   USHORT GetMemoryClock;                         //Atomic Table,  directly used by various SW components,latest version 1.1 
@@ -308,27 +309,36 @@ typedef struct _ATOM_MASTER_LIST_OF_COMMAND_TABLES{
   USHORT SetVoltage;                             //Function Table,directly and/or indirectly used by various SW components,latest version 1.1
   USHORT DAC1OutputControl;                      //Atomic Table,  directly used by various SW components,latest version 1.1
   USHORT DAC2OutputControl;                      //Atomic Table,  directly used by various SW components,latest version 1.1
-  USHORT SetupHWAssistedI2CStatus;               //Function Table,only used by Bios, obsolete soon.Switch to use "ReadEDIDFromHWAssistedI2C"
+  USHORT ComputeMemoryClockParam;                //Function Table,only used by Bios, obsolete soon.Switch to use "ReadEDIDFromHWAssistedI2C"
   USHORT ClockSource;                            //Atomic Table,  indirectly used by various SW components,called from ASIC_Init
   USHORT MemoryDeviceInit;                       //Atomic Table,  indirectly used by various SW components,called from SetMemoryClock
-  USHORT EnableYUV;                              //Atomic Table,  indirectly used by various SW components,called from EnableVGARender
+  USHORT GetDispObjectInfo;                      //Atomic Table,  indirectly used by various SW components,called from EnableVGARender
   USHORT DIG1EncoderControl;                     //Atomic Table,directly used by various SW components,latest version 1.1
   USHORT DIG2EncoderControl;                     //Atomic Table,directly used by various SW components,latest version 1.1
   USHORT DIG1TransmitterControl;                 //Atomic Table,directly used by various SW components,latest version 1.1
   USHORT DIG2TransmitterControl;	               //Atomic Table,directly used by various SW components,latest version 1.1 
   USHORT ProcessAuxChannelTransaction;					 //Function Table,only used by Bios
   USHORT DPEncoderService;											 //Function Table,only used by Bios
+  USHORT GetVoltageInfo;                         //Function Table,only used by Bios since SI
 }ATOM_MASTER_LIST_OF_COMMAND_TABLES;   
 
 // For backward compatible 
 #define ReadEDIDFromHWAssistedI2C                ProcessI2cChannelTransaction
-#define UNIPHYTransmitterControl						     DIG1TransmitterControl
-#define LVTMATransmitterControl							     DIG2TransmitterControl
+#define DPTranslatorControl                      DIG2EncoderControl
+#define UNIPHYTransmitterControl			     DIG1TransmitterControl
+#define LVTMATransmitterControl				     DIG2TransmitterControl
 #define SetCRTC_DPM_State                        GetConditionalGoldenSetting
 #define SetUniphyInstance                        ASIC_StaticPwrMgtStatusChange
 #define HPDInterruptService                      ReadHWAssistedI2CStatus
 #define EnableVGA_Access                         GetSCLKOverMCLKRatio
-#define GetDispObjectInfo                        EnableYUV 
+#define EnableYUV                                GetDispObjectInfo                         
+#define DynamicClockGating                       EnableDispPowerGating
+#define SetupHWAssistedI2CStatus                 ComputeMemoryClockParam
+
+#define TMDSAEncoderControl                      PatchMCSetting
+#define LVDSEncoderControl                       MC_SEQ_Control
+#define LCD1OutputControl                        HW_Misc_Operation
+
 
 typedef struct _ATOM_MASTER_COMMAND_TABLE
 {
@@ -495,6 +505,34 @@ typedef struct _COMPUTE_MEMORY_ENGINE_PLL_PARAMETERS_V5
 // ucInputFlag
 #define ATOM_PLL_INPUT_FLAG_PLL_STROBE_MODE_EN  1   // 1-StrobeMode, 0-PerformanceMode
 
+// use for ComputeMemoryClockParamTable
+typedef struct _COMPUTE_MEMORY_CLOCK_PARAM_PARAMETERS_V2_1
+{
+  union
+  {
+    ULONG  ulClock;         
+    ATOM_S_MPLL_FB_DIVIDER   ulFbDiv;         //Output:UPPER_WORD=FB_DIV_INTEGER,  LOWER_WORD=FB_DIV_FRAC shl (16-FB_FRACTION_BITS)
+  };
+  UCHAR   ucDllSpeed;                         //Output 
+  UCHAR   ucPostDiv;                          //Output
+  union{
+    UCHAR   ucInputFlag;                      //Input : ATOM_PLL_INPUT_FLAG_PLL_STROBE_MODE_EN: 1-StrobeMode, 0-PerformanceMode
+    UCHAR   ucPllCntlFlag;                    //Output: 
+  };
+  UCHAR   ucBWCntl;                       
+}COMPUTE_MEMORY_CLOCK_PARAM_PARAMETERS_V2_1;
+
+// definition of ucInputFlag
+#define MPLL_INPUT_FLAG_STROBE_MODE_EN          0x01
+// definition of ucPllCntlFlag
+#define MPLL_CNTL_FLAG_VCO_MODE_MASK            0x03 
+#define MPLL_CNTL_FLAG_BYPASS_DQ_PLL            0x04
+#define MPLL_CNTL_FLAG_QDR_ENABLE               0x08
+#define MPLL_CNTL_FLAG_AD_HALF_RATE             0x10
+
+//MPLL_CNTL_FLAG_BYPASS_AD_PLL has a wrong name, should be BYPASS_DQ_PLL
+#define MPLL_CNTL_FLAG_BYPASS_AD_PLL            0x04
+
 typedef struct _DYNAMICE_MEMORY_SETTINGS_PARAMETER
 {
   ATOM_COMPUTE_CLOCK_FREQ ulClock;
@@ -562,6 +600,16 @@ typedef struct _DYNAMIC_CLOCK_GATING_PARAMETERS
 #define  DYNAMIC_CLOCK_GATING_PS_ALLOCATION  DYNAMIC_CLOCK_GATING_PARAMETERS
 
 /****************************************************************************/	
+// Structure used by EnableDispPowerGatingTable.ctb
+/****************************************************************************/	
+typedef struct _ENABLE_DISP_POWER_GATING_PARAMETERS_V2_1 
+{
+  UCHAR ucDispPipeId;                 // ATOM_CRTC1, ATOM_CRTC2, ...
+  UCHAR ucEnable;                     // ATOM_ENABLE or ATOM_DISABLE
+  UCHAR ucPadding[2];
+}ENABLE_DISP_POWER_GATING_PARAMETERS_V2_1;
+
+/****************************************************************************/	
 // Structure used by EnableASIC_StaticPwrMgtTable.ctb
 /****************************************************************************/	
 typedef struct _ENABLE_ASIC_STATIC_PWR_MGT_PARAMETERS
@@ -807,6 +855,7 @@ typedef struct _ATOM_DIG_ENCODER_CONFIG_V4
 #define ATOM_ENCODER_CONFIG_V4_DPLINKRATE_1_62GHZ		  0x00
 #define ATOM_ENCODER_CONFIG_V4_DPLINKRATE_2_70GHZ		  0x01
 #define ATOM_ENCODER_CONFIG_V4_DPLINKRATE_5_40GHZ		  0x02
+#define ATOM_ENCODER_CONFIG_V4_DPLINKRATE_3_24GHZ		  0x03
 #define ATOM_ENCODER_CONFIG_V4_ENCODER_SEL					  0x70
 #define ATOM_ENCODER_CONFIG_V4_DIG0_ENCODER					  0x00
 #define ATOM_ENCODER_CONFIG_V4_DIG1_ENCODER					  0x10
@@ -814,6 +863,7 @@ typedef struct _ATOM_DIG_ENCODER_CONFIG_V4
 #define ATOM_ENCODER_CONFIG_V4_DIG3_ENCODER					  0x30
 #define ATOM_ENCODER_CONFIG_V4_DIG4_ENCODER					  0x40
 #define ATOM_ENCODER_CONFIG_V4_DIG5_ENCODER					  0x50
+#define ATOM_ENCODER_CONFIG_V4_DIG6_ENCODER					  0x60
 
 typedef struct _DIG_ENCODER_CONTROL_PARAMETERS_V4
 {
@@ -1171,6 +1221,106 @@ typedef struct _DIG_TRANSMITTER_CONTROL_PARAMETERS_V4
 #define ATOM_TRANSMITTER_CONFIG_V4_TRANSMITTER3           	0x80	//EF
 
 
+typedef struct _ATOM_DIG_TRANSMITTER_CONFIG_V5
+{
+#if ATOM_BIG_ENDIAN
+  UCHAR ucReservd1:1;
+  UCHAR ucHPDSel:3;
+  UCHAR ucPhyClkSrcId:2;            
+  UCHAR ucCoherentMode:1;            
+  UCHAR ucReserved:1;
+#else
+  UCHAR ucReserved:1;
+  UCHAR ucCoherentMode:1;            
+  UCHAR ucPhyClkSrcId:2;            
+  UCHAR ucHPDSel:3;
+  UCHAR ucReservd1:1;
+#endif
+}ATOM_DIG_TRANSMITTER_CONFIG_V5;
+
+typedef struct _DIG_TRANSMITTER_CONTROL_PARAMETERS_V1_5
+{
+  USHORT usSymClock;		        // Encoder Clock in 10kHz,(DP mode)= linkclock/10, (TMDS/LVDS/HDMI)= pixel clock,  (HDMI deep color), =pixel clock * deep_color_ratio
+  UCHAR  ucPhyId;                   // 0=UNIPHYA, 1=UNIPHYB, 2=UNIPHYC, 3=UNIPHYD, 4= UNIPHYE 5=UNIPHYF
+  UCHAR  ucAction;				    // define as ATOM_TRANSMITER_ACTION_xxx
+  UCHAR  ucLaneNum;                 // indicate lane number 1-8
+  UCHAR  ucConnObjId;               // Connector Object Id defined in ObjectId.h
+  UCHAR  ucDigMode;                 // indicate DIG mode
+  union{
+  ATOM_DIG_TRANSMITTER_CONFIG_V5 asConfig;
+  UCHAR ucConfig;
+  };
+  UCHAR  ucDigEncoderSel;           // indicate DIG front end encoder 
+  UCHAR  ucDPLaneSet;
+  UCHAR  ucReserved;
+  UCHAR  ucReserved1;
+}DIG_TRANSMITTER_CONTROL_PARAMETERS_V1_5;
+
+//ucPhyId
+#define ATOM_PHY_ID_UNIPHYA                                 0  
+#define ATOM_PHY_ID_UNIPHYB                                 1
+#define ATOM_PHY_ID_UNIPHYC                                 2
+#define ATOM_PHY_ID_UNIPHYD                                 3
+#define ATOM_PHY_ID_UNIPHYE                                 4
+#define ATOM_PHY_ID_UNIPHYF                                 5
+#define ATOM_PHY_ID_UNIPHYG                                 6
+
+// ucDigEncoderSel
+#define ATOM_TRANMSITTER_V5__DIGA_SEL                       0x01
+#define ATOM_TRANMSITTER_V5__DIGB_SEL                       0x02
+#define ATOM_TRANMSITTER_V5__DIGC_SEL                       0x04
+#define ATOM_TRANMSITTER_V5__DIGD_SEL                       0x08
+#define ATOM_TRANMSITTER_V5__DIGE_SEL                       0x10
+#define ATOM_TRANMSITTER_V5__DIGF_SEL                       0x20
+#define ATOM_TRANMSITTER_V5__DIGG_SEL                       0x40
+
+// ucDigMode
+#define ATOM_TRANSMITTER_DIGMODE_V5_DP                      0
+#define ATOM_TRANSMITTER_DIGMODE_V5_LVDS                    1
+#define ATOM_TRANSMITTER_DIGMODE_V5_DVI                     2
+#define ATOM_TRANSMITTER_DIGMODE_V5_HDMI                    3
+#define ATOM_TRANSMITTER_DIGMODE_V5_SDVO                    4
+#define ATOM_TRANSMITTER_DIGMODE_V5_DP_MST                  5
+
+// ucDPLaneSet
+#define DP_LANE_SET__0DB_0_4V                               0x00
+#define DP_LANE_SET__0DB_0_6V                               0x01
+#define DP_LANE_SET__0DB_0_8V                               0x02
+#define DP_LANE_SET__0DB_1_2V                               0x03
+#define DP_LANE_SET__3_5DB_0_4V                             0x08  
+#define DP_LANE_SET__3_5DB_0_6V                             0x09
+#define DP_LANE_SET__3_5DB_0_8V                             0x0a
+#define DP_LANE_SET__6DB_0_4V                               0x10
+#define DP_LANE_SET__6DB_0_6V                               0x11
+#define DP_LANE_SET__9_5DB_0_4V                             0x18  
+
+// ATOM_DIG_TRANSMITTER_CONFIG_V5 asConfig;
+// Bit1
+#define ATOM_TRANSMITTER_CONFIG_V5_COHERENT				          0x02
+
+// Bit3:2
+#define ATOM_TRANSMITTER_CONFIG_V5_REFCLK_SEL_MASK 	        0x0c
+#define ATOM_TRANSMITTER_CONFIG_V5_REFCLK_SEL_SHIFT		    0x02
+
+#define ATOM_TRANSMITTER_CONFIG_V5_P1PLL         		        0x00
+#define ATOM_TRANSMITTER_CONFIG_V5_P2PLL		                0x04
+#define ATOM_TRANSMITTER_CONFIG_V5_P0PLL		                0x08   
+#define ATOM_TRANSMITTER_CONFIG_V5_REFCLK_SRC_EXT           0x0c
+// Bit6:4
+#define ATOM_TRANSMITTER_CONFIG_V5_HPD_SEL_MASK		          0x70
+#define ATOM_TRANSMITTER_CONFIG_V5_HPD_SEL_SHIFT		      0x04
+
+#define ATOM_TRANSMITTER_CONFIG_V5_NO_HPD_SEL				        0x00
+#define ATOM_TRANSMITTER_CONFIG_V5_HPD1_SEL				          0x10
+#define ATOM_TRANSMITTER_CONFIG_V5_HPD2_SEL				          0x20
+#define ATOM_TRANSMITTER_CONFIG_V5_HPD3_SEL				          0x30
+#define ATOM_TRANSMITTER_CONFIG_V5_HPD4_SEL				          0x40
+#define ATOM_TRANSMITTER_CONFIG_V5_HPD5_SEL				          0x50
+#define ATOM_TRANSMITTER_CONFIG_V5_HPD6_SEL				          0x60
+
+#define DIG_TRANSMITTER_CONTROL_PS_ALLOCATION_V1_5            DIG_TRANSMITTER_CONTROL_PARAMETERS_V1_5
+
+
 /****************************************************************************/	
 // Structures used by ExternalEncoderControlTable V1.3
 // ASIC Families: Evergreen, Llano, NI
@@ -1793,6 +1943,7 @@ typedef struct _ENABLE_SPREAD_SPECTRUM_ON_PPLL_V2
 #define ATOM_PPLL_SS_TYPE_V3_P1PLL            0x00
 #define ATOM_PPLL_SS_TYPE_V3_P2PLL            0x04
 #define ATOM_PPLL_SS_TYPE_V3_DCPLL            0x08
+#define ATOM_PPLL_SS_TYPE_V3_P0PLL            ATOM_PPLL_SS_TYPE_V3_DCPLL
 #define ATOM_PPLL_SS_AMOUNT_V3_FBDIV_MASK     0x00FF
 #define ATOM_PPLL_SS_AMOUNT_V3_FBDIV_SHIFT    0
 #define ATOM_PPLL_SS_AMOUNT_V3_NFRAC_MASK     0x0F00
@@ -2030,12 +2181,77 @@ typedef struct	_SET_VOLTAGE_PARAMETERS_V2
   USHORT   usVoltageLevel;              // real voltage level
 }SET_VOLTAGE_PARAMETERS_V2;
 
+
+typedef struct	_SET_VOLTAGE_PARAMETERS_V1_3
+{
+  UCHAR    ucVoltageType;               // To tell which voltage to set up, VDDC/MVDDC/MVDDQ/VDDCI
+  UCHAR    ucVoltageMode;               // Indicate action: Set voltage level
+  USHORT   usVoltageLevel;              // real voltage level in unit of mv or Voltage Phase (0, 1, 2, .. )
+}SET_VOLTAGE_PARAMETERS_V1_3;
+
+//ucVoltageType
+#define VOLTAGE_TYPE_VDDC                    1
+#define VOLTAGE_TYPE_MVDDC                   2
+#define VOLTAGE_TYPE_MVDDQ                   3
+#define VOLTAGE_TYPE_VDDCI                   4
+
+//SET_VOLTAGE_PARAMETERS_V3.ucVoltageMode
+#define ATOM_SET_VOLTAGE                     0        //Set voltage Level
+#define ATOM_INIT_VOLTAGE_REGULATOR          3        //Init Regulator
+#define ATOM_SET_VOLTAGE_PHASE               4        //Set Vregulator Phase
+#define ATOM_GET_MAX_VOLTAGE                 6        //Get Max Voltage, not used in SetVoltageTable v1.3
+#define ATOM_GET_VOLTAGE_LEVEL               6        //Get Voltage level from vitual voltage ID
+
+// define vitual voltage id in usVoltageLevel
+#define ATOM_VIRTUAL_VOLTAGE_ID0             0xff01
+#define ATOM_VIRTUAL_VOLTAGE_ID1             0xff02
+#define ATOM_VIRTUAL_VOLTAGE_ID2             0xff03
+#define ATOM_VIRTUAL_VOLTAGE_ID3             0xff04
+
 typedef struct _SET_VOLTAGE_PS_ALLOCATION
 {
   SET_VOLTAGE_PARAMETERS sASICSetVoltage;
   WRITE_ONE_BYTE_HW_I2C_DATA_PS_ALLOCATION sReserved;
 }SET_VOLTAGE_PS_ALLOCATION;
 
+// New Added from SI for GetVoltageInfoTable, input parameter structure
+typedef struct  _GET_VOLTAGE_INFO_INPUT_PARAMETER_V1_1
+{
+  UCHAR    ucVoltageType;               // Input: To tell which voltage to set up, VDDC/MVDDC/MVDDQ/VDDCI
+  UCHAR    ucVoltageMode;               // Input: Indicate action: Get voltage info
+  USHORT   usVoltageLevel;              // Input: real voltage level in unit of mv or Voltage Phase (0, 1, 2, .. ) or Leakage Id 
+  ULONG    ulReserved;
+}GET_VOLTAGE_INFO_INPUT_PARAMETER_V1_1;
+
+// New Added from SI for GetVoltageInfoTable, output parameter structure when ucVotlageMode == ATOM_GET_VOLTAGE_VID
+typedef struct  _GET_VOLTAGE_INFO_OUTPUT_PARAMETER_V1_1
+{
+  ULONG    ulVotlageGpioState;
+  ULONG    ulVoltageGPioMask;
+}GET_VOLTAGE_INFO_OUTPUT_PARAMETER_V1_1;
+
+// New Added from SI for GetVoltageInfoTable, output parameter structure when ucVotlageMode == ATOM_GET_VOLTAGE_STATEx_LEAKAGE_VID
+typedef struct  _GET_LEAKAGE_VOLTAGE_INFO_OUTPUT_PARAMETER_V1_1
+{
+  USHORT   usVoltageLevel;
+  USHORT   usVoltageId;                                  // Voltage Id programmed in Voltage Regulator
+  ULONG    ulReseved;
+}GET_LEAKAGE_VOLTAGE_INFO_OUTPUT_PARAMETER_V1_1;
+
+
+// GetVoltageInfo v1.1 ucVoltageMode
+#define	ATOM_GET_VOLTAGE_VID                0x00
+#define ATOM_GET_VOTLAGE_INIT_SEQ           0x03
+#define ATOM_GET_VOLTTAGE_PHASE_PHASE_VID   0x04
+// for SI, this state map to 0xff02 voltage state in Power Play table, which is power boost state
+#define	ATOM_GET_VOLTAGE_STATE0_LEAKAGE_VID 0x10
+
+// for SI, this state map to 0xff01 voltage state in Power Play table, which is performance state
+#define	ATOM_GET_VOLTAGE_STATE1_LEAKAGE_VID 0x11
+// undefined power state
+#define	ATOM_GET_VOLTAGE_STATE2_LEAKAGE_VID 0x12
+#define	ATOM_GET_VOLTAGE_STATE3_LEAKAGE_VID 0x13
+
 /****************************************************************************/	
 // Structures used by TVEncoderControlTable
 /****************************************************************************/	
@@ -2065,9 +2281,9 @@ typedef struct _ATOM_MASTER_LIST_OF_DATA_TABLES
   USHORT        MultimediaConfigInfo;     // Only used by MM Lib,latest version 2.1, not configuable from Bios, need to include the table to build Bios
   USHORT        StandardVESA_Timing;      // Only used by Bios
   USHORT        FirmwareInfo;             // Shared by various SW components,latest version 1.4
-  USHORT        DAC_Info;                 // Will be obsolete from R600
+  USHORT        PaletteData;              // Only used by BIOS
   USHORT        LCD_Info;                 // Shared by various SW components,latest version 1.3, was called LVDS_Info 
-  USHORT        TMDS_Info;                // Will be obsolete from R600
+  USHORT        DIGTransmitterInfo;       // Internal used by VBIOS only version 3.1
   USHORT        AnalogTV_Info;            // Shared by various SW components,latest version 1.1 
   USHORT        SupportedDevicesInfo;     // Will be obsolete from R600
   USHORT        GPIO_I2C_Info;            // Shared by various SW components,latest version 1.2 will be used from R600           
@@ -2096,15 +2312,16 @@ typedef struct _ATOM_MASTER_LIST_OF_DATA_TABLES
 	USHORT				PowerSourceInfo;					// Shared by various SW components, latest versoin 1.1
 }ATOM_MASTER_LIST_OF_DATA_TABLES;
 
-// For backward compatible 
-#define LVDS_Info                LCD_Info
-
 typedef struct _ATOM_MASTER_DATA_TABLE
 { 
   ATOM_COMMON_TABLE_HEADER sHeader;  
   ATOM_MASTER_LIST_OF_DATA_TABLES   ListOfDataTables;
 }ATOM_MASTER_DATA_TABLE;
 
+// For backward compatible 
+#define LVDS_Info                LCD_Info
+#define DAC_Info                 PaletteData
+#define TMDS_Info                DIGTransmitterInfo
 
 /****************************************************************************/	
 // Structure used in MultimediaCapabilityInfoTable
@@ -2171,7 +2388,9 @@ typedef struct _ATOM_MULTIMEDIA_CONFIG_INFO
 typedef struct _ATOM_FIRMWARE_CAPABILITY
 {
 #if ATOM_BIG_ENDIAN
-  USHORT Reserved:3;
+  USHORT Reserved:1;
+  USHORT SCL2Redefined:1;
+  USHORT PostWithoutModeSet:1;
   USHORT HyperMemory_Size:4;
   USHORT HyperMemory_Support:1;
   USHORT PPMode_Assigned:1;
@@ -2193,7 +2412,9 @@ typedef struct _ATOM_FIRMWARE_CAPABILITY
   USHORT PPMode_Assigned:1;
   USHORT HyperMemory_Support:1;
   USHORT HyperMemory_Size:4;
-  USHORT Reserved:3;
+  USHORT PostWithoutModeSet:1;
+  USHORT SCL2Redefined:1;
+  USHORT Reserved:1;
 #endif
 }ATOM_FIRMWARE_CAPABILITY;
 
@@ -2418,7 +2639,8 @@ typedef struct _ATOM_FIRMWARE_INFO_V2_2
   USHORT                          usLcdMaxPixelClockPLL_Output; // In MHz unit
   ULONG                           ulReserved4;                //Was ulAsicMaximumVoltage
   ULONG                           ulMinPixelClockPLL_Output;  //In 10Khz unit
-  ULONG                           ulReserved5;                //Was usMinEngineClockPLL_Input and usMaxEngineClockPLL_Input
+  UCHAR                           ucRemoteDisplayConfig;
+  UCHAR                           ucReserved5[3];             //Was usMinEngineClockPLL_Input and usMaxEngineClockPLL_Input
   ULONG                           ulReserved6;                //Was usMinEngineClockPLL_Output and usMinMemoryClockPLL_Input
   ULONG                           ulReserved7;                //Was usMaxMemoryClockPLL_Input and usMinMemoryClockPLL_Output
   USHORT                          usReserved11;               //Was usMaxPixelClock;  //In 10Khz unit, Max.  Pclk used only for DAC
@@ -2438,6 +2660,11 @@ typedef struct _ATOM_FIRMWARE_INFO_V2_2
 
 #define ATOM_FIRMWARE_INFO_LAST  ATOM_FIRMWARE_INFO_V2_2
 
+
+// definition of ucRemoteDisplayConfig
+#define REMOTE_DISPLAY_DISABLE                   0x00
+#define REMOTE_DISPLAY_ENABLE                    0x01
+
 /****************************************************************************/	
 // Structures used in IntegratedSystemInfoTable
 /****************************************************************************/	
@@ -2660,8 +2887,9 @@ usMinDownStreamHTLinkWidth:  same as above.
 #define    INTEGRATED_SYSTEM_INFO__AMD_CPU__GREYHOUND      2
 #define    INTEGRATED_SYSTEM_INFO__AMD_CPU__K8             3
 #define    INTEGRATED_SYSTEM_INFO__AMD_CPU__PHARAOH        4
+#define    INTEGRATED_SYSTEM_INFO__AMD_CPU__OROCHI         5
 
-#define    INTEGRATED_SYSTEM_INFO__AMD_CPU__MAX_CODE       INTEGRATED_SYSTEM_INFO__AMD_CPU__PHARAOH    // this deff reflects max defined CPU code
+#define    INTEGRATED_SYSTEM_INFO__AMD_CPU__MAX_CODE       INTEGRATED_SYSTEM_INFO__AMD_CPU__OROCHI    // this deff reflects max defined CPU code
 
 #define SYSTEM_CONFIG_POWEREXPRESS_ENABLE                 0x00000001
 #define SYSTEM_CONFIG_RUN_AT_OVERDRIVE_ENGINE             0x00000002
@@ -2753,6 +2981,7 @@ typedef struct _ATOM_INTEGRATED_SYSTEM_INFO_V5
 #define ASIC_INT_DIG4_ENCODER_ID													0x0b
 #define ASIC_INT_DIG5_ENCODER_ID													0x0c
 #define ASIC_INT_DIG6_ENCODER_ID													0x0d
+#define ASIC_INT_DIG7_ENCODER_ID													0x0e
 
 //define Encoder attribute
 #define ATOM_ANALOG_ENCODER																0
@@ -3226,15 +3455,23 @@ typedef struct _ATOM_LCD_INFO_V13
 
   UCHAR               ucPowerSequenceDIGONtoDE_in4Ms;
   UCHAR               ucPowerSequenceDEtoVARY_BL_in4Ms;
-  UCHAR               ucPowerSequenceDEtoDIGON_in4Ms;
   UCHAR               ucPowerSequenceVARY_BLtoDE_in4Ms;
+  UCHAR               ucPowerSequenceDEtoDIGON_in4Ms;
 
   UCHAR               ucOffDelay_in4Ms;
   UCHAR               ucPowerSequenceVARY_BLtoBLON_in4Ms;
   UCHAR               ucPowerSequenceBLONtoVARY_BL_in4Ms;
   UCHAR               ucReserved1;
 
-  ULONG               ulReserved[4];
+  UCHAR               ucDPCD_eDP_CONFIGURATION_CAP;     // dpcd 0dh
+  UCHAR               ucDPCD_MAX_LINK_RATE;             // dpcd 01h
+  UCHAR               ucDPCD_MAX_LANE_COUNT;            // dpcd 02h
+  UCHAR               ucDPCD_MAX_DOWNSPREAD;            // dpcd 03h
+
+  USHORT              usMaxPclkFreqInSingleLink;        // Max PixelClock frequency in single link mode. 
+  UCHAR               uceDPToLVDSRxId;
+  UCHAR               ucLcdReservd;
+  ULONG               ulReserved[2];
 }ATOM_LCD_INFO_V13;  
 
 #define ATOM_LCD_INFO_LAST  ATOM_LCD_INFO_V13    
@@ -3273,6 +3510,11 @@ typedef struct _ATOM_LCD_INFO_V13
 //Use this cap bit for a quick reference whether an embadded panel (LCD1 ) is LVDS or eDP.
 #define	LCDPANEL_CAP_V13_eDP                    0x4        // = LCDPANEL_CAP_eDP no change comparing to previous version
 
+//uceDPToLVDSRxId
+#define eDP_TO_LVDS_RX_DISABLE                  0x00       // no eDP->LVDS translator chip 
+#define eDP_TO_LVDS_COMMON_ID                   0x01       // common eDP->LVDS translator chip without AMD SW init
+#define eDP_TO_LVDS_RT_ID                       0x02       // RT tanslator which require AMD SW init
+
 typedef struct  _ATOM_PATCH_RECORD_MODE
 {
   UCHAR     ucRecordType;
@@ -3317,6 +3559,7 @@ typedef struct  _ATOM_PANEL_RESOLUTION_PATCH_RECORD
 #define LCD_CAP_RECORD_TYPE                   3
 #define LCD_FAKE_EDID_PATCH_RECORD_TYPE       4
 #define LCD_PANEL_RESOLUTION_RECORD_TYPE      5
+#define LCD_EDID_OFFSET_PATCH_RECORD_TYPE     6
 #define ATOM_RECORD_END_TYPE                  0xFF
 
 /****************************Spread Spectrum Info Table Definitions **********************/
@@ -3528,6 +3771,7 @@ else	//Non VGA case
 
 CAIL needs to claim an reserved area defined by FBAccessAreaOffset and usFBUsedbyDrvInKB in non VGA case.*/
 
+/***********************************************************************************/	
 #define ATOM_MAX_FIRMWARE_VRAM_USAGE_INFO			1
 
 typedef struct _ATOM_FIRMWARE_VRAM_RESERVE_INFO
@@ -3818,13 +4062,17 @@ typedef struct _EXT_DISPLAY_PATH
     ATOM_DP_CONN_CHANNEL_MAPPING asDPMapping;
     ATOM_DVI_CONN_CHANNEL_MAPPING asDVIMapping;
   };
-  UCHAR   ucReserved;
-  USHORT  usReserved[2]; 
+  UCHAR   ucChPNInvert;                   // bit vector for up to 8 lanes, =0: P and N is not invert, =1 P and N is inverted
+  USHORT  usCaps;
+  USHORT  usReserved; 
 }EXT_DISPLAY_PATH;
    
 #define NUMBER_OF_UCHAR_FOR_GUID          16
 #define MAX_NUMBER_OF_EXT_DISPLAY_PATH    7
 
+//usCaps
+#define  EXT_DISPLAY_PATH_CAPS__HBR2_DISABLE          0x01
+
 typedef  struct _ATOM_EXTERNAL_DISPLAY_CONNECTION_INFO
 {
   ATOM_COMMON_TABLE_HEADER sHeader;
@@ -3832,7 +4080,9 @@ typedef  struct _ATOM_EXTERNAL_DISPLAY_CONNECTION_INFO
   EXT_DISPLAY_PATH         sPath[MAX_NUMBER_OF_EXT_DISPLAY_PATH]; // total of fixed 7 entries.
   UCHAR                    ucChecksum;                            // a  simple Checksum of the sum of whole structure equal to 0x0. 
   UCHAR                    uc3DStereoPinId;                       // use for eDP panel
-  UCHAR                    Reserved [6];                          // for potential expansion
+  UCHAR                    ucRemoteDisplayConfig;
+  UCHAR                    uceDPToLVDSRxId;
+  UCHAR                    Reserved[4];                           // for potential expansion
 }ATOM_EXTERNAL_DISPLAY_CONNECTION_INFO;
 
 //Related definitions, all records are different but they have a commond header
@@ -3977,6 +4227,7 @@ typedef struct  _ATOM_OBJECT_GPIO_CNTL_RECORD
 #define GPIO_PIN_STATE_ACTIVE_HIGH      0x1
 
 // Indexes to GPIO array in GLSync record 
+// GLSync record is for Frame Lock/Gen Lock feature.
 #define ATOM_GPIO_INDEX_GLSYNC_REFCLK    0
 #define ATOM_GPIO_INDEX_GLSYNC_HSYNC     1
 #define ATOM_GPIO_INDEX_GLSYNC_VSYNC     2
@@ -3984,7 +4235,9 @@ typedef struct  _ATOM_OBJECT_GPIO_CNTL_RECORD
 #define ATOM_GPIO_INDEX_GLSYNC_SWAP_GNT  4
 #define ATOM_GPIO_INDEX_GLSYNC_INTERRUPT 5
 #define ATOM_GPIO_INDEX_GLSYNC_V_RESET   6
-#define ATOM_GPIO_INDEX_GLSYNC_MAX       7
+#define ATOM_GPIO_INDEX_GLSYNC_SWAP_CNTL 7
+#define ATOM_GPIO_INDEX_GLSYNC_SWAP_SEL  8
+#define ATOM_GPIO_INDEX_GLSYNC_MAX       9
 
 typedef struct  _ATOM_ENCODER_DVO_CF_RECORD
 {
@@ -3994,7 +4247,8 @@ typedef struct  _ATOM_ENCODER_DVO_CF_RECORD
 }ATOM_ENCODER_DVO_CF_RECORD;
 
 // Bit maps for ATOM_ENCODER_CAP_RECORD.ucEncoderCap
-#define ATOM_ENCODER_CAP_RECORD_HBR2     0x01         // DP1.2 HBR2 is supported by this path
+#define ATOM_ENCODER_CAP_RECORD_HBR2                  0x01         // DP1.2 HBR2 is supported by HW encoder
+#define ATOM_ENCODER_CAP_RECORD_HBR2_EN               0x02         // DP1.2 HBR2 setting is qualified and HBR2 can be enabled 
 
 typedef struct  _ATOM_ENCODER_CAP_RECORD
 {
@@ -4003,11 +4257,13 @@ typedef struct  _ATOM_ENCODER_CAP_RECORD
     USHORT                    usEncoderCap;         
     struct {
 #if ATOM_BIG_ENDIAN
-      USHORT                  usReserved:15;        // Bit1-15 may be defined for other capability in future
+      USHORT                  usReserved:14;        // Bit1-15 may be defined for other capability in future
+      USHORT                  usHBR2En:1;           // Bit1 is for DP1.2 HBR2 enable
       USHORT                  usHBR2Cap:1;          // Bit0 is for DP1.2 HBR2 capability. 
 #else
       USHORT                  usHBR2Cap:1;          // Bit0 is for DP1.2 HBR2 capability. 
-      USHORT                  usReserved:15;        // Bit1-15 may be defined for other capability in future
+      USHORT                  usHBR2En:1;           // Bit1 is for DP1.2 HBR2 enable
+      USHORT                  usReserved:14;        // Bit1-15 may be defined for other capability in future
 #endif
     };
   }; 
@@ -4157,6 +4413,7 @@ typedef struct _ATOM_VOLTAGE_CONTROL
 #define	VOLTAGE_CONTROL_ID_VT1556M						0x07									
 #define	VOLTAGE_CONTROL_ID_CHL822x						0x08									
 #define	VOLTAGE_CONTROL_ID_VT1586M						0x09
+#define VOLTAGE_CONTROL_ID_UP1637 						0x0A
 
 typedef struct  _ATOM_VOLTAGE_OBJECT
 {
@@ -4193,6 +4450,69 @@ typedef struct  _ATOM_LEAKID_VOLTAGE
 	USHORT	usVoltage;
 }ATOM_LEAKID_VOLTAGE;
 
+typedef struct _ATOM_VOLTAGE_OBJECT_HEADER_V3{
+ 	 UCHAR		ucVoltageType;									//Indicate Voltage Source: VDDC, MVDDC, MVDDQ or MVDDCI	 
+   UCHAR		ucVoltageMode;							    //Indicate voltage control mode: Init/Set/Leakage/Set phase 
+	 USHORT		usSize;													//Size of Object	
+}ATOM_VOLTAGE_OBJECT_HEADER_V3;
+
+typedef struct  _VOLTAGE_LUT_ENTRY_V2
+{
+	 ULONG		ulVoltageId;									  // The Voltage ID which is used to program GPIO register
+	 USHORT		usVoltageValue;									// The corresponding Voltage Value, in mV
+}VOLTAGE_LUT_ENTRY_V2;
+
+typedef struct  _LEAKAGE_VOLTAGE_LUT_ENTRY_V2
+{
+  USHORT	usVoltageLevel; 							  // The Voltage ID which is used to program GPIO register
+  USHORT  usVoltageId;                    
+	USHORT	usLeakageId;									  // The corresponding Voltage Value, in mV
+}LEAKAGE_VOLTAGE_LUT_ENTRY_V2;
+
+typedef struct  _ATOM_I2C_VOLTAGE_OBJECT_V3
+{
+   ATOM_VOLTAGE_OBJECT_HEADER_V3 sHeader;
+   UCHAR	ucVoltageRegulatorId;					  //Indicate Voltage Regulator Id
+   UCHAR    ucVoltageControlI2cLine;
+   UCHAR    ucVoltageControlAddress;
+   UCHAR    ucVoltageControlOffset;	 	
+   ULONG    ulReserved;
+   VOLTAGE_LUT_ENTRY asVolI2cLut[1];        // end with 0xff
+}ATOM_I2C_VOLTAGE_OBJECT_V3;
+
+typedef struct  _ATOM_GPIO_VOLTAGE_OBJECT_V3
+{
+   ATOM_VOLTAGE_OBJECT_HEADER_V3 sHeader;   
+   UCHAR    ucVoltageGpioCntlId;         // default is 0 which indicate control through CG VID mode 
+   UCHAR    ucGpioEntryNum;              // indiate the entry numbers of Votlage/Gpio value Look up table
+   UCHAR    ucPhaseDelay;                // phase delay in unit of micro second
+   UCHAR    ucReserved;   
+   ULONG    ulGpioMaskVal;               // GPIO Mask value
+   VOLTAGE_LUT_ENTRY_V2 asVolGpioLut[1];   
+}ATOM_GPIO_VOLTAGE_OBJECT_V3;
+
+typedef struct  _ATOM_LEAKAGE_VOLTAGE_OBJECT_V3
+{
+   ATOM_VOLTAGE_OBJECT_HEADER_V3 sHeader;
+   UCHAR    ucLeakageCntlId;             // default is 0
+   UCHAR    ucLeakageEntryNum;           // indicate the entry number of LeakageId/Voltage Lut table
+   UCHAR    ucReserved[2];               
+   ULONG    ulMaxVoltageLevel;
+   LEAKAGE_VOLTAGE_LUT_ENTRY_V2 asLeakageIdLut[1];   
+}ATOM_LEAKAGE_VOLTAGE_OBJECT_V3;
+
+typedef union _ATOM_VOLTAGE_OBJECT_V3{
+  ATOM_GPIO_VOLTAGE_OBJECT_V3 asGpioVoltageObj;
+  ATOM_I2C_VOLTAGE_OBJECT_V3 asI2cVoltageObj;
+  ATOM_LEAKAGE_VOLTAGE_OBJECT_V3 asLeakageObj;
+}ATOM_VOLTAGE_OBJECT_V3;
+
+typedef struct  _ATOM_VOLTAGE_OBJECT_INFO_V3_1
+{
+   ATOM_COMMON_TABLE_HEADER	sHeader; 
+	 ATOM_VOLTAGE_OBJECT_V3			asVoltageObj[3];	//Info for Voltage control	  	 
+}ATOM_VOLTAGE_OBJECT_INFO_V3_1;
+
 typedef struct  _ATOM_ASIC_PROFILE_VOLTAGE
 {
 	UCHAR		ucProfileId;
@@ -4305,7 +4625,18 @@ typedef struct _ATOM_INTEGRATED_SYSTEM_INFO_V6
   USHORT usHDMISSpreadRateIn10Hz;
   USHORT usDVISSPercentage;
   USHORT usDVISSpreadRateIn10Hz;
-  ULONG  ulReserved3[21]; 
+  ULONG  SclkDpmBoostMargin;
+  ULONG  SclkDpmThrottleMargin;
+  USHORT SclkDpmTdpLimitPG; 
+  USHORT SclkDpmTdpLimitBoost;
+  ULONG  ulBoostEngineCLock;
+  UCHAR  ulBoostVid_2bit;  
+  UCHAR  EnableBoost;
+  USHORT GnbTdpLimit;
+  USHORT usMaxLVDSPclkFreqInSingleLink;
+  UCHAR  ucLvdsMisc;
+  UCHAR  ucLVDSReserved;
+  ULONG  ulReserved3[15]; 
   ATOM_EXTERNAL_DISPLAY_CONNECTION_INFO sExtDispConnInfo;   
 }ATOM_INTEGRATED_SYSTEM_INFO_V6;   
 
@@ -4313,9 +4644,16 @@ typedef struct _ATOM_INTEGRATED_SYSTEM_INFO_V6
 #define INTEGRATED_SYSTEM_INFO_V6_GPUCAPINFO__TMDSHDMI_COHERENT_SINGLEPLL_MODE       0x01
 #define INTEGRATED_SYSTEM_INFO_V6_GPUCAPINFO__DISABLE_AUX_HW_MODE_DETECTION          0x08
 
-// ulOtherDisplayMisc
-#define INTEGRATED_SYSTEM_INFO__GET_EDID_CALLBACK_FUNC_SUPPORT                       0x01
+//ucLVDSMisc:                   
+#define SYS_INFO_LVDSMISC__888_FPDI_MODE                                             0x01
+#define SYS_INFO_LVDSMISC__DL_CH_SWAP                                                0x02
+#define SYS_INFO_LVDSMISC__888_BPC                                                   0x04
+#define SYS_INFO_LVDSMISC__OVERRIDE_EN                                               0x08
+#define SYS_INFO_LVDSMISC__BLON_ACTIVE_LOW                                           0x10
 
+// not used any more
+#define SYS_INFO_LVDSMISC__VSYNC_ACTIVE_LOW                                          0x04
+#define SYS_INFO_LVDSMISC__HSYNC_ACTIVE_LOW                                          0x08
 
 /**********************************************************************************************************************
   ATOM_INTEGRATED_SYSTEM_INFO_V6 Description
@@ -4384,7 +4722,208 @@ ucUMAChannelNumber:      	        System memory channel numbers.
 ulCSR_M3_ARB_CNTL_DEFAULT[10]:    Arrays with values for CSR M3 arbiter for default
 ulCSR_M3_ARB_CNTL_UVD[10]:        Arrays with values for CSR M3 arbiter for UVD playback.
 ulCSR_M3_ARB_CNTL_FS3D[10]:       Arrays with values for CSR M3 arbiter for Full Screen 3D applications.
-sAvail_SCLK[5]:                   Arrays to provide available list of SLCK and corresponding voltage, order from low to high  
+sAvail_SCLK[5]:                   Arrays to provide availabe list of SLCK and corresponding voltage, order from low to high  
+ulGMCRestoreResetTime:            GMC power restore and GMC reset time to calculate data reconnection latency. Unit in ns. 
+ulMinimumNClk:                    Minimum NCLK speed among all NB-Pstates to calcualte data reconnection latency. Unit in 10kHz. 
+ulIdleNClk:                       NCLK speed while memory runs in self-refresh state. Unit in 10kHz.
+ulDDR_DLL_PowerUpTime:            DDR PHY DLL power up time. Unit in ns.
+ulDDR_PLL_PowerUpTime:            DDR PHY PLL power up time. Unit in ns.
+usPCIEClkSSPercentage:            PCIE Clock Spred Spectrum Percentage in unit 0.01%; 100 mean 1%.
+usPCIEClkSSType:                  PCIE Clock Spred Spectrum Type. 0 for Down spread(default); 1 for Center spread.
+usLvdsSSPercentage:               LVDS panel ( not include eDP ) Spread Spectrum Percentage in unit of 0.01%, =0, use VBIOS default setting. 
+usLvdsSSpreadRateIn10Hz:          LVDS panel ( not include eDP ) Spread Spectrum frequency in unit of 10Hz, =0, use VBIOS default setting. 
+usHDMISSPercentage:               HDMI Spread Spectrum Percentage in unit 0.01%; 100 mean 1%,  =0, use VBIOS default setting. 
+usHDMISSpreadRateIn10Hz:          HDMI Spread Spectrum frequency in unit of 10Hz,  =0, use VBIOS default setting. 
+usDVISSPercentage:                DVI Spread Spectrum Percentage in unit 0.01%; 100 mean 1%,  =0, use VBIOS default setting. 
+usDVISSpreadRateIn10Hz:           DVI Spread Spectrum frequency in unit of 10Hz,  =0, use VBIOS default setting. 
+usMaxLVDSPclkFreqInSingleLink:    Max pixel clock LVDS panel single link, if=0 means VBIOS use default threhold, right now it is 85Mhz
+ucLVDSMisc:                       [bit0] LVDS 888bit panel mode =0: LVDS 888 panel in LDI mode, =1: LVDS 888 panel in FPDI mode
+                                  [bit1] LVDS panel lower and upper link mapping =0: lower link and upper link not swap, =1: lower link and upper link are swapped
+                                  [bit2] LVDS 888bit per color mode  =0: 666 bit per color =1:888 bit per color
+                                  [bit3] LVDS parameter override enable  =0: ucLvdsMisc parameter are not used =1: ucLvdsMisc parameter should be used
+                                  [bit4] Polarity of signal sent to digital BLON output pin. =0: not inverted(active high) =1: inverted ( active low )
+**********************************************************************************************************************/
+
+// this Table is used for Liano/Ontario APU
+typedef struct _ATOM_FUSION_SYSTEM_INFO_V1
+{
+  ATOM_INTEGRATED_SYSTEM_INFO_V6    sIntegratedSysInfo;   
+  ULONG  ulPowerplayTable[128];  
+}ATOM_FUSION_SYSTEM_INFO_V1; 
+/**********************************************************************************************************************
+  ATOM_FUSION_SYSTEM_INFO_V1 Description
+sIntegratedSysInfo:               refer to ATOM_INTEGRATED_SYSTEM_INFO_V6 definition.
+ulPowerplayTable[128]:            This 512 bytes memory is used to save ATOM_PPLIB_POWERPLAYTABLE3, starting form ulPowerplayTable[0]    
+**********************************************************************************************************************/ 
+
+// this IntegrateSystemInfoTable is used for Trinity APU
+typedef struct _ATOM_INTEGRATED_SYSTEM_INFO_V1_7
+{
+  ATOM_COMMON_TABLE_HEADER   sHeader;
+  ULONG  ulBootUpEngineClock;
+  ULONG  ulDentistVCOFreq;
+  ULONG  ulBootUpUMAClock;
+  ATOM_CLK_VOLT_CAPABILITY   sDISPCLK_Voltage[4];
+  ULONG  ulBootUpReqDisplayVector;
+  ULONG  ulOtherDisplayMisc;
+  ULONG  ulGPUCapInfo;
+  ULONG  ulSB_MMIO_Base_Addr;
+  USHORT usRequestedPWMFreqInHz;
+  UCHAR  ucHtcTmpLmt;
+  UCHAR  ucHtcHystLmt;
+  ULONG  ulMinEngineClock;
+  ULONG  ulSystemConfig;            
+  ULONG  ulCPUCapInfo;
+  USHORT usNBP0Voltage;               
+  USHORT usNBP1Voltage;
+  USHORT usBootUpNBVoltage;                       
+  USHORT usExtDispConnInfoOffset;
+  USHORT usPanelRefreshRateRange;     
+  UCHAR  ucMemoryType;  
+  UCHAR  ucUMAChannelNumber;
+  UCHAR  strVBIOSMsg[40];
+  ULONG  ulReserved[20];
+  ATOM_AVAILABLE_SCLK_LIST   sAvail_SCLK[5];
+  ULONG  ulGMCRestoreResetTime;
+  ULONG  ulMinimumNClk;
+  ULONG  ulIdleNClk;
+  ULONG  ulDDR_DLL_PowerUpTime;
+  ULONG  ulDDR_PLL_PowerUpTime;
+  USHORT usPCIEClkSSPercentage;
+  USHORT usPCIEClkSSType;
+  USHORT usLvdsSSPercentage;
+  USHORT usLvdsSSpreadRateIn10Hz;
+  USHORT usHDMISSPercentage;
+  USHORT usHDMISSpreadRateIn10Hz;
+  USHORT usDVISSPercentage;
+  USHORT usDVISSpreadRateIn10Hz;
+  ULONG  SclkDpmBoostMargin;
+  ULONG  SclkDpmThrottleMargin;
+  USHORT SclkDpmTdpLimitPG; 
+  USHORT SclkDpmTdpLimitBoost;
+  ULONG  ulBoostEngineCLock;
+  UCHAR  ulBoostVid_2bit;  
+  UCHAR  EnableBoost;
+  USHORT GnbTdpLimit;
+  USHORT usMaxLVDSPclkFreqInSingleLink;
+  UCHAR  ucLvdsMisc;
+  UCHAR  ucLVDSReserved;
+  UCHAR  ucLVDSPwrOnSeqDIGONtoDE_in4Ms;
+  UCHAR  ucLVDSPwrOnSeqDEtoVARY_BL_in4Ms;
+  UCHAR  ucLVDSPwrOffSeqVARY_BLtoDE_in4Ms;
+  UCHAR  ucLVDSPwrOffSeqDEtoDIGON_in4Ms;
+  UCHAR  ucLVDSOffToOnDelay_in4Ms;
+  UCHAR  ucLVDSPwrOnSeqVARY_BLtoBLON_in4Ms;
+  UCHAR  ucLVDSPwrOffSeqBLONtoVARY_BL_in4Ms;
+  UCHAR  ucLVDSReserved1;
+  ULONG  ulLCDBitDepthControlVal;
+  ULONG  ulNbpStateMemclkFreq[4];
+  USHORT usNBP2Voltage;               
+  USHORT usNBP3Voltage;
+  ULONG  ulNbpStateNClkFreq[4];
+  UCHAR  ucNBDPMEnable;
+  UCHAR  ucReserved[3];
+  UCHAR  ucDPMState0VclkFid;
+  UCHAR  ucDPMState0DclkFid;
+  UCHAR  ucDPMState1VclkFid;
+  UCHAR  ucDPMState1DclkFid;
+  UCHAR  ucDPMState2VclkFid;
+  UCHAR  ucDPMState2DclkFid;
+  UCHAR  ucDPMState3VclkFid;
+  UCHAR  ucDPMState3DclkFid;
+  ATOM_EXTERNAL_DISPLAY_CONNECTION_INFO sExtDispConnInfo;
+}ATOM_INTEGRATED_SYSTEM_INFO_V1_7;
+
+// ulOtherDisplayMisc
+#define INTEGRATED_SYSTEM_INFO__GET_EDID_CALLBACK_FUNC_SUPPORT            0x01
+#define INTEGRATED_SYSTEM_INFO__GET_BOOTUP_DISPLAY_CALLBACK_FUNC_SUPPORT  0x02
+#define INTEGRATED_SYSTEM_INFO__GET_EXPANSION_CALLBACK_FUNC_SUPPORT       0x04
+#define INTEGRATED_SYSTEM_INFO__FAST_BOOT_SUPPORT                         0x08
+
+// ulGPUCapInfo
+#define SYS_INFO_GPUCAPS__TMDSHDMI_COHERENT_SINGLEPLL_MODE                0x01
+#define SYS_INFO_GPUCAPS__DP_SINGLEPLL_MODE                               0x02
+#define SYS_INFO_GPUCAPS__DISABLE_AUX_MODE_DETECT                         0x08
+
+/**********************************************************************************************************************
+  ATOM_INTEGRATED_SYSTEM_INFO_V1_7 Description
+ulBootUpEngineClock:              VBIOS bootup Engine clock frequency, in 10kHz unit. if it is equal 0, then VBIOS use pre-defined bootup engine clock
+ulDentistVCOFreq:                 Dentist VCO clock in 10kHz unit. 
+ulBootUpUMAClock:                 System memory boot up clock frequency in 10Khz unit. 
+sDISPCLK_Voltage:                 Report Display clock voltage requirement.
+ 
+ulBootUpReqDisplayVector:         VBIOS boot up display IDs, following are supported devices in Trinity projects:
+                                  ATOM_DEVICE_CRT1_SUPPORT                  0x0001
+                                  ATOM_DEVICE_DFP1_SUPPORT                  0x0008 
+                                  ATOM_DEVICE_DFP6_SUPPORT                  0x0040 
+                                  ATOM_DEVICE_DFP2_SUPPORT                  0x0080       
+                                  ATOM_DEVICE_DFP3_SUPPORT                  0x0200       
+                                  ATOM_DEVICE_DFP4_SUPPORT                  0x0400        
+                                  ATOM_DEVICE_DFP5_SUPPORT                  0x0800
+                                  ATOM_DEVICE_LCD1_SUPPORT                  0x0002
+ulOtherDisplayMisc:      	        bit[0]=0: INT15 callback function Get LCD EDID ( ax=4e08, bl=1b ) is not supported by SBIOS. 
+                                        =1: INT15 callback function Get LCD EDID ( ax=4e08, bl=1b ) is supported by SBIOS. 
+                                  bit[1]=0: INT15 callback function Get boot display( ax=4e08, bl=01h) is not supported by SBIOS
+                                        =1: INT15 callback function Get boot display( ax=4e08, bl=01h) is supported by SBIOS
+                                  bit[2]=0: INT15 callback function Get panel Expansion ( ax=4e08, bl=02h) is not supported by SBIOS
+                                        =1: INT15 callback function Get panel Expansion ( ax=4e08, bl=02h) is supported by SBIOS
+                                  bit[3]=0: VBIOS fast boot is disable
+                                        =1: VBIOS fast boot is enable. ( VBIOS skip display device detection in every set mode if LCD panel is connect and LID is open)
+ulGPUCapInfo:                     bit[0]=0: TMDS/HDMI Coherent Mode use cascade PLL mode.
+                                        =1: TMDS/HDMI Coherent Mode use signel PLL mode.
+                                  bit[1]=0: DP mode use cascade PLL mode ( New for Trinity )
+                                        =1: DP mode use single PLL mode
+                                  bit[3]=0: Enable AUX HW mode detection logic
+                                        =1: Disable AUX HW mode detection logic
+                                      
+ulSB_MMIO_Base_Addr:              Physical Base address to SB MMIO space. Driver needs to initialize it for SMU usage.
+
+usRequestedPWMFreqInHz:           When it's set to 0x0 by SBIOS: the LCD BackLight is not controlled by GPU(SW). 
+                                  Any attempt to change BL using VBIOS function or enable VariBri from PP table is not effective since ATOM_BIOS_INFO_BL_CONTROLLED_BY_GPU==0;
+                                  
+                                  When it's set to a non-zero frequency, the BackLight is controlled by GPU (SW) in one of two ways below:
+                                  1. SW uses the GPU BL PWM output to control the BL, in chis case, this non-zero frequency determines what freq GPU should use;
+                                  VBIOS will set up proper PWM frequency and ATOM_BIOS_INFO_BL_CONTROLLED_BY_GPU==1,as the result,
+                                  Changing BL using VBIOS function is functional in both driver and non-driver present environment; 
+                                  and enabling VariBri under the driver environment from PP table is optional.
+
+                                  2. SW uses other means to control BL (like DPCD),this non-zero frequency serves as a flag only indicating
+                                  that BL control from GPU is expected.
+                                  VBIOS will NOT set up PWM frequency but make ATOM_BIOS_INFO_BL_CONTROLLED_BY_GPU==1
+                                  Changing BL using VBIOS function could be functional in both driver and non-driver present environment,but
+                                  it's per platform 
+                                  and enabling VariBri under the driver environment from PP table is optional.
+
+ucHtcTmpLmt:                      Refer to D18F3x64 bit[22:16], HtcTmpLmt. 
+                                  Threshold on value to enter HTC_active state.
+ucHtcHystLmt:                     Refer to D18F3x64 bit[27:24], HtcHystLmt. 
+                                  To calculate threshold off value to exit HTC_active state, which is Threshold on vlaue minus ucHtcHystLmt.
+ulMinEngineClock:                 Minimum SCLK allowed in 10kHz unit. This is calculated based on WRCK Fuse settings.
+ulSystemConfig:                   Bit[0]=0: PCIE Power Gating Disabled 
+                                        =1: PCIE Power Gating Enabled
+                                  Bit[1]=0: DDR-DLL shut-down feature disabled.
+                                         1: DDR-DLL shut-down feature enabled.
+                                  Bit[2]=0: DDR-PLL Power down feature disabled.
+                                         1: DDR-PLL Power down feature enabled.                                 
+ulCPUCapInfo:                     TBD
+usNBP0Voltage:                    VID for voltage on NB P0 State
+usNBP1Voltage:                    VID for voltage on NB P1 State  
+usNBP2Voltage:                    VID for voltage on NB P2 State
+usNBP3Voltage:                    VID for voltage on NB P3 State  
+usBootUpNBVoltage:                Voltage Index of GNB voltage configured by SBIOS, which is suffcient to support VBIOS DISPCLK requirement.
+usExtDispConnInfoOffset:          Offset to sExtDispConnInfo inside the structure
+usPanelRefreshRateRange:          Bit vector for LCD supported refresh rate range. If DRR is requestd by the platform, at least two bits need to be set
+                                  to indicate a range.
+                                  SUPPORTED_LCD_REFRESHRATE_30Hz          0x0004
+                                  SUPPORTED_LCD_REFRESHRATE_40Hz          0x0008
+                                  SUPPORTED_LCD_REFRESHRATE_50Hz          0x0010
+                                  SUPPORTED_LCD_REFRESHRATE_60Hz          0x0020
+ucMemoryType:                     [3:0]=1:DDR1;=2:DDR2;=3:DDR3.[7:4] is reserved.
+ucUMAChannelNumber:      	        System memory channel numbers. 
+ulCSR_M3_ARB_CNTL_DEFAULT[10]:    Arrays with values for CSR M3 arbiter for default
+ulCSR_M3_ARB_CNTL_UVD[10]:        Arrays with values for CSR M3 arbiter for UVD playback.
+ulCSR_M3_ARB_CNTL_FS3D[10]:       Arrays with values for CSR M3 arbiter for Full Screen 3D applications.
+sAvail_SCLK[5]:                   Arrays to provide availabe list of SLCK and corresponding voltage, order from low to high  
 ulGMCRestoreResetTime:            GMC power restore and GMC reset time to calculate data reconnection latency. Unit in ns. 
 ulMinimumNClk:                    Minimum NCLK speed among all NB-Pstates to calcualte data reconnection latency. Unit in 10kHz. 
 ulIdleNClk:                       NCLK speed while memory runs in self-refresh state. Unit in 10kHz.
@@ -4398,6 +4937,41 @@ usHDMISSPercentage:               HDMI Spread Spectrum Percentage in unit 0.01%;
 usHDMISSpreadRateIn10Hz:          HDMI Spread Spectrum frequency in unit of 10Hz,  =0, use VBIOS default setting. 
 usDVISSPercentage:                DVI Spread Spectrum Percentage in unit 0.01%; 100 mean 1%,  =0, use VBIOS default setting. 
 usDVISSpreadRateIn10Hz:           DVI Spread Spectrum frequency in unit of 10Hz,  =0, use VBIOS default setting. 
+usMaxLVDSPclkFreqInSingleLink:    Max pixel clock LVDS panel single link, if=0 means VBIOS use default threhold, right now it is 85Mhz
+ucLVDSMisc:                       [bit0] LVDS 888bit panel mode =0: LVDS 888 panel in LDI mode, =1: LVDS 888 panel in FPDI mode
+                                  [bit1] LVDS panel lower and upper link mapping =0: lower link and upper link not swap, =1: lower link and upper link are swapped
+                                  [bit2] LVDS 888bit per color mode  =0: 666 bit per color =1:888 bit per color
+                                  [bit3] LVDS parameter override enable  =0: ucLvdsMisc parameter are not used =1: ucLvdsMisc parameter should be used
+                                  [bit4] Polarity of signal sent to digital BLON output pin. =0: not inverted(active high) =1: inverted ( active low )
+ucLVDSPwrOnSeqDIGONtoDE_in4Ms:    LVDS power up sequence time in unit of 4ms, time delay from DIGON signal active to data enable signal active( DE ).
+                                  =0 mean use VBIOS default which is 8 ( 32ms ). The LVDS power up sequence is as following: DIGON->DE->VARY_BL->BLON. 
+                                  This parameter is used by VBIOS only. VBIOS will patch LVDS_InfoTable.
+ucLVDSPwrOnDEtoVARY_BL_in4Ms:     LVDS power up sequence time in unit of 4ms., time delay from DE( data enable ) active to Vary Brightness enable signal active( VARY_BL ).  
+                                  =0 mean use VBIOS default which is 90 ( 360ms ). The LVDS power up sequence is as following: DIGON->DE->VARY_BL->BLON. 
+                                  This parameter is used by VBIOS only. VBIOS will patch LVDS_InfoTable.
+
+ucLVDSPwrOffVARY_BLtoDE_in4Ms:    LVDS power down sequence time in unit of 4ms, time delay from data enable ( DE ) signal off to LCDVCC (DIGON) off. 
+                                  =0 mean use VBIOS default delay which is 8 ( 32ms ). The LVDS power down sequence is as following: BLON->VARY_BL->DE->DIGON
+                                  This parameter is used by VBIOS only. VBIOS will patch LVDS_InfoTable.
+
+ucLVDSPwrOffDEtoDIGON_in4Ms:      LVDS power down sequence time in unit of 4ms, time delay from vary brightness enable signal( VARY_BL) off to data enable ( DE ) signal off. 
+                                  =0 mean use VBIOS default which is 90 ( 360ms ). The LVDS power down sequence is as following: BLON->VARY_BL->DE->DIGON
+                                  This parameter is used by VBIOS only. VBIOS will patch LVDS_InfoTable.
+
+ucLVDSOffToOnDelay_in4Ms:         LVDS power down sequence time in unit of 4ms. Time delay from DIGON signal off to DIGON signal active. 
+                                  =0 means to use VBIOS default delay which is 125 ( 500ms ).
+                                  This parameter is used by VBIOS only. VBIOS will patch LVDS_InfoTable.
+
+ucLVDSPwrOnVARY_BLtoBLON_in4Ms:   LVDS power up sequence time in unit of 4ms. Time delay from VARY_BL signal on to DLON signal active. 
+                                  =0 means to use VBIOS default delay which is 0 ( 0ms ).
+                                  This parameter is used by VBIOS only. VBIOS will patch LVDS_InfoTable.
+
+ucLVDSPwrOffBLONtoVARY_BL_in4Ms:  LVDS power down sequence time in unit of 4ms. Time delay from BLON signal off to VARY_BL signal off. 
+                                  =0 means to use VBIOS default delay which is 0 ( 0ms ).
+                                  This parameter is used by VBIOS only. VBIOS will patch LVDS_InfoTable.
+
+ulNbpStateMemclkFreq[4]:          system memory clock frequncey in unit of 10Khz in different NB pstate. 
+
 **********************************************************************************************************************/
 
 /**************************************************************************/
@@ -4459,6 +5033,7 @@ typedef struct _ATOM_ASIC_SS_ASSIGNMENT
 #define ASIC_INTERNAL_SS_ON_DP      7
 #define ASIC_INTERNAL_SS_ON_DCPLL   8
 #define ASIC_EXTERNAL_SS_ON_DP_CLOCK 9
+#define ASIC_INTERNAL_VCE_SS        10
 
 typedef struct _ATOM_ASIC_SS_ASSIGNMENT_V2
 {
@@ -4520,7 +5095,7 @@ typedef struct _ATOM_ASIC_INTERNAL_SS_INFO_V3
 #define ATOM_DOS_MODE_INFO_DEF        7
 #define ATOM_I2C_CHANNEL_STATUS_DEF   8
 #define ATOM_I2C_CHANNEL_STATUS1_DEF  9
-
+#define ATOM_INTERNAL_TIMER_DEF       10
 
 // BIOS_0_SCRATCH Definition 
 #define ATOM_S0_CRT1_MONO               0x00000001L
@@ -4648,6 +5223,7 @@ typedef struct _ATOM_ASIC_INTERNAL_SS_INFO_V3
 #define ATOM_S2_DEVICE_DPMS_MASKw1      0x3FF
 #define ATOM_S2_FORCEDLOWPWRMODE_STATE_MASKb3     0x0C
 #define ATOM_S2_FORCEDLOWPWRMODE_STATE_CHANGEb3   0x10
+#define ATOM_S2_TMDS_COHERENT_MODEb3    0x10          // used by VBIOS code only, use coherent mode for TMDS/HDMI mode
 #define ATOM_S2_VRI_BRIGHT_ENABLEb3     0x20
 #define ATOM_S2_ROTATION_STATE_MASKb3   0xC0
 
@@ -5038,6 +5614,23 @@ typedef struct _ENABLE_GRAPH_SURFACE_PARAMETERS_V1_3
   USHORT usDeviceId;                  // Active Device Id for this surface. If no device, set to 0. 
 }ENABLE_GRAPH_SURFACE_PARAMETERS_V1_3;
 
+typedef struct _ENABLE_GRAPH_SURFACE_PARAMETERS_V1_4
+{
+  USHORT usHight;                     // Image Hight
+  USHORT usWidth;                     // Image Width
+  USHORT usGraphPitch;
+  UCHAR  ucColorDepth;
+  UCHAR  ucPixelFormat;
+  UCHAR  ucSurface;                   // Surface 1 or 2
+  UCHAR  ucEnable;                    // ATOM_ENABLE or ATOM_DISABLE
+  UCHAR  ucModeType;
+  UCHAR  ucReserved;
+}ENABLE_GRAPH_SURFACE_PARAMETERS_V1_4;
+
+// ucEnable
+#define ATOM_GRAPH_CONTROL_SET_PITCH             0x0f
+#define ATOM_GRAPH_CONTROL_SET_DISP_START        0x10
+
 typedef struct _ENABLE_GRAPH_SURFACE_PS_ALLOCATION
 {
   ENABLE_GRAPH_SURFACE_PARAMETERS sSetSurface;          
@@ -5057,6 +5650,58 @@ typedef struct  _GET_DISPLAY_SURFACE_SIZE_PARAMETERS
   USHORT  usY_Size;
 }GET_DISPLAY_SURFACE_SIZE_PARAMETERS; 
 
+typedef struct  _GET_DISPLAY_SURFACE_SIZE_PARAMETERS_V2
+{
+  union{
+    USHORT  usX_Size;                     //When use as input parameter, usX_Size indicates which CRTC                 
+    USHORT  usSurface; 
+  };
+  USHORT usY_Size;
+  USHORT usDispXStart;               
+  USHORT usDispYStart;
+}GET_DISPLAY_SURFACE_SIZE_PARAMETERS_V2; 
+
+
+typedef struct _PALETTE_DATA_CONTROL_PARAMETERS_V3 
+{
+  UCHAR  ucLutId;
+  UCHAR  ucAction;
+  USHORT usLutStartIndex;
+  USHORT usLutLength;
+  USHORT usLutOffsetInVram;
+}PALETTE_DATA_CONTROL_PARAMETERS_V3;
+
+// ucAction:
+#define PALETTE_DATA_AUTO_FILL            1
+#define PALETTE_DATA_READ                 2
+#define PALETTE_DATA_WRITE                3
+
+
+typedef struct _INTERRUPT_SERVICE_PARAMETERS_V2
+{
+  UCHAR  ucInterruptId;
+  UCHAR  ucServiceId;
+  UCHAR  ucStatus;
+  UCHAR  ucReserved;
+}INTERRUPT_SERVICE_PARAMETER_V2;
+
+// ucInterruptId
+#define HDP1_INTERRUPT_ID                 1
+#define HDP2_INTERRUPT_ID                 2
+#define HDP3_INTERRUPT_ID                 3
+#define HDP4_INTERRUPT_ID                 4
+#define HDP5_INTERRUPT_ID                 5
+#define HDP6_INTERRUPT_ID                 6
+#define SW_INTERRUPT_ID                   11   
+
+// ucAction
+#define INTERRUPT_SERVICE_GEN_SW_INT      1
+#define INTERRUPT_SERVICE_GET_STATUS      2
+
+ // ucStatus
+#define INTERRUPT_STATUS__INT_TRIGGER     1
+#define INTERRUPT_STATUS__HPD_HIGH        2
+
 typedef struct _INDIRECT_IO_ACCESS
 {
   ATOM_COMMON_TABLE_HEADER sHeader;  
@@ -5189,7 +5834,7 @@ typedef struct _ATOM_INIT_REG_BLOCK{
 
 #define END_OF_REG_INDEX_BLOCK  0x0ffff
 #define END_OF_REG_DATA_BLOCK   0x00000000
-#define ATOM_INIT_REG_MASK_FLAG 0x80
+#define ATOM_INIT_REG_MASK_FLAG 0x80               //Not used in BIOS
 #define	CLOCK_RANGE_HIGHEST			0x00ffffff
 
 #define VALUE_DWORD             SIZEOF ULONG
@@ -5229,6 +5874,7 @@ typedef struct _ATOM_MC_INIT_PARAM_TABLE
 #define _128Mx8             0x51
 #define _128Mx16            0x52
 #define _256Mx8             0x61
+#define _256Mx16            0x62
 
 #define SAMSUNG             0x1
 #define INFINEON            0x2
@@ -5585,7 +6231,7 @@ typedef struct _ATOM_VRAM_MODULE_V7
   ULONG	  ulChannelMapCfg;	                // mmMC_SHARED_CHREMAP
   USHORT  usModuleSize;                     // Size of ATOM_VRAM_MODULE_V7
   USHORT  usPrivateReserved;                // MC_ARB_RAMCFG (includes NOOFBANK,NOOFRANKS,NOOFROWS,NOOFCOLS)
-  USHORT  usReserved;
+  USHORT  usEnableChannels;                 // bit vector which indicate which channels are enabled
   UCHAR   ucExtMemoryID;                    // Current memory module ID
   UCHAR   ucMemoryType;                     // MEM_TYPE_DDR2/DDR3/GDDR3/GDDR5
   UCHAR   ucChannelNum;                     // Number of mem. channels supported in this module
@@ -5597,7 +6243,8 @@ typedef struct _ATOM_VRAM_MODULE_V7
   UCHAR   ucNPL_RT;                         // Round trip delay (MC_SEQ_CAS_TIMING [28:24]:TCL=CL+NPL_RT-2). Always 2.
   UCHAR	  ucPreamble;                       // [7:4] Write Preamble, [3:0] Read Preamble
   UCHAR   ucMemorySize;                     // Total memory size in unit of 16MB for CONFIG_MEMSIZE - bit[23:0] zeros
-  UCHAR   ucReserved[3];
+  USHORT  usSEQSettingOffset;
+  UCHAR   ucReserved;
 // Memory Module specific values
   USHORT  usEMRS2Value;                     // EMRS2/MR2 Value. 
   USHORT  usEMRS3Value;                     // EMRS3/MR3 Value.
@@ -5633,10 +6280,10 @@ typedef struct _ATOM_VRAM_INFO_V3
 typedef struct _ATOM_VRAM_INFO_V4
 {
   ATOM_COMMON_TABLE_HEADER   sHeader;
-	USHORT										 usMemAdjustTblOffset;													 // offset of ATOM_INIT_REG_BLOCK structure for memory vendor specific MC adjust setting
-	USHORT										 usMemClkPatchTblOffset;												 //	offset of ATOM_INIT_REG_BLOCK structure for memory clock specific MC setting
-	USHORT										 usRerseved;
-	UCHAR           	         ucMemDQ7_0ByteRemap;													   // DQ line byte remap, =0: Memory Data line BYTE0, =1: BYTE1, =2: BYTE2, =3: BYTE3
+  USHORT                     usMemAdjustTblOffset;													 // offset of ATOM_INIT_REG_BLOCK structure for memory vendor specific MC adjust setting
+  USHORT                     usMemClkPatchTblOffset;												 //	offset of ATOM_INIT_REG_BLOCK structure for memory clock specific MC setting
+  USHORT										 usRerseved;
+  UCHAR           	         ucMemDQ7_0ByteRemap;													   // DQ line byte remap, =0: Memory Data line BYTE0, =1: BYTE1, =2: BYTE2, =3: BYTE3
   ULONG                      ulMemDQ7_0BitRemap;                             // each DQ line ( 7~0) use 3bits, like: DQ0=Bit[2:0], DQ1:[5:3], ... DQ7:[23:21]
   UCHAR                      ucReservde[4]; 
   UCHAR                      ucNumOfVRAMModule;
@@ -5648,9 +6295,10 @@ typedef struct _ATOM_VRAM_INFO_V4
 typedef struct _ATOM_VRAM_INFO_HEADER_V2_1
 {
   ATOM_COMMON_TABLE_HEADER   sHeader;
-	USHORT										 usMemAdjustTblOffset;													 // offset of ATOM_INIT_REG_BLOCK structure for memory vendor specific MC adjust setting
-	USHORT										 usMemClkPatchTblOffset;												 //	offset of ATOM_INIT_REG_BLOCK structure for memory clock specific MC setting
-	USHORT										 usReserved[4];
+  USHORT                     usMemAdjustTblOffset;													 // offset of ATOM_INIT_REG_BLOCK structure for memory vendor specific MC adjust setting
+  USHORT                     usMemClkPatchTblOffset;												 //	offset of ATOM_INIT_REG_BLOCK structure for memory clock specific MC setting
+  USHORT                     usPerBytePresetOffset;                          // offset of ATOM_INIT_REG_BLOCK structure for Per Byte Offset Preset Settings
+  USHORT                     usReserved[3];
   UCHAR                      ucNumOfVRAMModule;                              // indicate number of VRAM module
   UCHAR                      ucMemoryClkPatchTblVer;                         // version of memory AC timing register list
   UCHAR                      ucVramModuleVer;                                // indicate ATOM_VRAM_MODUE version
@@ -5935,6 +6583,52 @@ typedef struct _ATOM_DISP_OUT_INFO_V2
 	ASIC_ENCODER_INFO      asEncoderInfo[1];
 }ATOM_DISP_OUT_INFO_V2;
 
+
+typedef struct _ATOM_DISP_CLOCK_ID {
+  UCHAR ucPpllId; 
+  UCHAR ucPpllAttribute;
+}ATOM_DISP_CLOCK_ID;
+
+// ucPpllAttribute
+#define CLOCK_SOURCE_SHAREABLE            0x01
+#define CLOCK_SOURCE_DP_MODE              0x02
+#define CLOCK_SOURCE_NONE_DP_MODE         0x04
+
+//DispOutInfoTable
+typedef struct _ASIC_TRANSMITTER_INFO_V2
+{
+	USHORT usTransmitterObjId;
+	USHORT usDispClkIdOffset;    // point to clock source id list supported by Encoder Object
+  UCHAR  ucTransmitterCmdTblId;
+	UCHAR  ucConfig;
+	UCHAR  ucEncoderID;					 // available 1st encoder ( default )
+	UCHAR  ucOptionEncoderID;    // available 2nd encoder ( optional )
+	UCHAR  uc2ndEncoderID;
+	UCHAR  ucReserved;
+}ASIC_TRANSMITTER_INFO_V2;
+
+typedef struct _ATOM_DISP_OUT_INFO_V3
+{
+  ATOM_COMMON_TABLE_HEADER sHeader;  
+	USHORT ptrTransmitterInfo;
+	USHORT ptrEncoderInfo;
+  USHORT ptrMainCallParserFar;                  // direct address of main parser call in VBIOS binary. 
+  USHORT usReserved;
+  UCHAR  ucDCERevision;   
+  UCHAR  ucMaxDispEngineNum;
+  UCHAR  ucMaxActiveDispEngineNum;
+  UCHAR  ucMaxPPLLNum;
+  UCHAR  ucCoreRefClkSource;                          // value of CORE_REF_CLK_SOURCE
+  UCHAR  ucReserved[3];
+	ASIC_TRANSMITTER_INFO_V2  asTransmitterInfo[1];     // for alligment only
+}ATOM_DISP_OUT_INFO_V3;
+
+typedef enum CORE_REF_CLK_SOURCE{
+  CLOCK_SRC_XTALIN=0,
+  CLOCK_SRC_XO_IN=1,
+  CLOCK_SRC_XO_IN2=2,
+}CORE_REF_CLK_SOURCE;
+
 // DispDevicePriorityInfo
 typedef struct _ATOM_DISPLAY_DEVICE_PRIORITY_INFO
 {
@@ -6070,6 +6764,39 @@ typedef struct _PROCESS_I2C_CHANNEL_TRANSACTION_PARAMETERS
 #define HW_I2C_READ         0
 #define I2C_2BYTE_ADDR      0x02
 
+/****************************************************************************/	
+// Structures used by HW_Misc_OperationTable
+/****************************************************************************/	
+typedef struct  _ATOM_HW_MISC_OPERATION_INPUT_PARAMETER_V1_1 
+{
+  UCHAR  ucCmd;                //  Input: To tell which action to take
+  UCHAR  ucReserved[3];
+  ULONG  ulReserved;
+}ATOM_HW_MISC_OPERATION_INPUT_PARAMETER_V1_1; 
+
+typedef struct  _ATOM_HW_MISC_OPERATION_OUTPUT_PARAMETER_V1_1 
+{
+  UCHAR  ucReturnCode;        // Output: Return value base on action was taken
+  UCHAR  ucReserved[3];
+  ULONG  ulReserved;
+}ATOM_HW_MISC_OPERATION_OUTPUT_PARAMETER_V1_1;
+
+// Actions code
+#define  ATOM_GET_SDI_SUPPORT              0xF0
+
+// Return code 
+#define  ATOM_UNKNOWN_CMD                   0
+#define  ATOM_FEATURE_NOT_SUPPORTED         1
+#define  ATOM_FEATURE_SUPPORTED             2
+
+typedef struct _ATOM_HW_MISC_OPERATION_PS_ALLOCATION
+{
+	ATOM_HW_MISC_OPERATION_INPUT_PARAMETER_V1_1        sInput_Output;
+	PROCESS_I2C_CHANNEL_TRANSACTION_PARAMETERS         sReserved; 
+}ATOM_HW_MISC_OPERATION_PS_ALLOCATION;
+
+/****************************************************************************/	
+
 typedef struct _SET_HWBLOCK_INSTANCE_PARAMETER_V2
 {
    UCHAR ucHWBlkInst;                // HW block instance, 0, 1, 2, ...
@@ -6090,6 +6817,52 @@ typedef struct _SET_HWBLOCK_INSTANCE_PARAMETER_V2
 #define SELECT_CRTC_PIXEL_RATE        7
 #define SELECT_VGA_BLK                8
 
+// DIGTransmitterInfoTable structure used to program UNIPHY settings 
+typedef struct _DIG_TRANSMITTER_INFO_HEADER_V3_1{  
+  ATOM_COMMON_TABLE_HEADER sHeader;  
+  USHORT usDPVsPreEmphSettingOffset;     // offset of PHY_ANALOG_SETTING_INFO * with DP Voltage Swing and Pre-Emphasis for each Link clock 
+  USHORT usPhyAnalogRegListOffset;       // offset of CLOCK_CONDITION_REGESTER_INFO* with None-DP mode Analog Setting's register Info 
+  USHORT usPhyAnalogSettingOffset;       // offset of CLOCK_CONDITION_SETTING_ENTRY* with None-DP mode Analog Setting for each link clock range
+  USHORT usPhyPllRegListOffset;          // offset of CLOCK_CONDITION_REGESTER_INFO* with Phy Pll register Info 
+  USHORT usPhyPllSettingOffset;          // offset of CLOCK_CONDITION_SETTING_ENTRY* with Phy Pll Settings
+}DIG_TRANSMITTER_INFO_HEADER_V3_1;
+
+typedef struct _CLOCK_CONDITION_REGESTER_INFO{
+  USHORT usRegisterIndex;
+  UCHAR  ucStartBit;
+  UCHAR  ucEndBit;
+}CLOCK_CONDITION_REGESTER_INFO;
+
+typedef struct _CLOCK_CONDITION_SETTING_ENTRY{
+  USHORT usMaxClockFreq;
+  UCHAR  ucEncodeMode;
+  UCHAR  ucPhySel;
+  ULONG  ulAnalogSetting[1];
+}CLOCK_CONDITION_SETTING_ENTRY;
+
+typedef struct _CLOCK_CONDITION_SETTING_INFO{
+  USHORT usEntrySize;
+  CLOCK_CONDITION_SETTING_ENTRY asClkCondSettingEntry[1];
+}CLOCK_CONDITION_SETTING_INFO;
+
+typedef struct _PHY_CONDITION_REG_VAL{
+  ULONG  ulCondition;
+  ULONG  ulRegVal;
+}PHY_CONDITION_REG_VAL;
+
+typedef struct _PHY_CONDITION_REG_INFO{
+  USHORT usRegIndex;
+  USHORT usSize;
+  PHY_CONDITION_REG_VAL asRegVal[1];
+}PHY_CONDITION_REG_INFO;
+
+typedef struct _PHY_ANALOG_SETTING_INFO{
+  UCHAR  ucEncodeMode;
+  UCHAR  ucPhySel;
+  USHORT usSize;
+  PHY_CONDITION_REG_INFO  asAnalogSetting[1];
+}PHY_ANALOG_SETTING_INFO;
+
 /****************************************************************************/	
 //Portion VI: Definitinos for vbios MC scratch registers that driver used
 /****************************************************************************/
@@ -6497,6 +7270,8 @@ typedef struct _ATOM_PPLIB_THERMALCONTROLLER
 #define ATOM_PP_THERMALCONTROLLER_EMC2103   13  /* 0x0D */ // Only fan control will be implemented, do NOT show this in PPGen.
 #define ATOM_PP_THERMALCONTROLLER_SUMO      14  /* 0x0E */ // Sumo type, used internally
 #define ATOM_PP_THERMALCONTROLLER_NISLANDS  15
+#define ATOM_PP_THERMALCONTROLLER_SISLANDS  16
+#define ATOM_PP_THERMALCONTROLLER_LM96163   17
 
 // Thermal controller 'combo type' to use an external controller for Fan control and an internal controller for thermal.
 // We probably should reserve the bit 0x80 for this use.
@@ -6512,6 +7287,7 @@ typedef struct _ATOM_PPLIB_STATE
     UCHAR ucClockStateIndices[1]; // variable-sized
 } ATOM_PPLIB_STATE;
 
+
 typedef struct _ATOM_PPLIB_FANTABLE
 {
     UCHAR   ucFanTableFormat;                // Change this if the table format changes or version changes so that the other fields are not the same.
@@ -6524,12 +7300,20 @@ typedef struct _ATOM_PPLIB_FANTABLE
     USHORT  usPWMHigh;                       // The PWM value at THigh.
 } ATOM_PPLIB_FANTABLE;
 
+typedef struct _ATOM_PPLIB_FANTABLE2
+{
+    ATOM_PPLIB_FANTABLE basicTable;
+    USHORT  usTMax;                          // The max temperature
+} ATOM_PPLIB_FANTABLE2;
+
 typedef struct _ATOM_PPLIB_EXTENDEDHEADER
 {
     USHORT  usSize;
     ULONG   ulMaxEngineClock;   // For Overdrive.
     ULONG   ulMaxMemoryClock;   // For Overdrive.
     // Add extra system parameters here, always adjust size to include all fields.
+    USHORT  usVCETableOffset; //points to ATOM_PPLIB_VCE_Table
+    USHORT  usUVDTableOffset;   //points to ATOM_PPLIB_UVD_Table
 } ATOM_PPLIB_EXTENDEDHEADER;
 
 //// ATOM_PPLIB_POWERPLAYTABLE::ulPlatformCaps
@@ -6552,6 +7336,7 @@ typedef struct _ATOM_PPLIB_EXTENDEDHEADER
 #define ATOM_PP_PLATFORM_CAP_REGULATOR_HOT 0x00010000               // Enable the 'regulator hot' feature.
 #define ATOM_PP_PLATFORM_CAP_BACO          0x00020000               // Does the driver supports BACO state.
 
+
 typedef struct _ATOM_PPLIB_POWERPLAYTABLE
 {
       ATOM_COMMON_TABLE_HEADER sHeader;
@@ -6610,7 +7395,8 @@ typedef struct _ATOM_PPLIB_POWERPLAYTABLE4
     USHORT                     usVddciDependencyOnMCLKOffset;
     USHORT                     usVddcDependencyOnMCLKOffset;
     USHORT                     usMaxClockVoltageOnDCOffset;
-    USHORT                     usReserved[2];  
+    USHORT                     usVddcPhaseShedLimitsTableOffset;    // Points to ATOM_PPLIB_PhaseSheddingLimits_Table
+    USHORT                     usReserved;  
 } ATOM_PPLIB_POWERPLAYTABLE4, *LPATOM_PPLIB_POWERPLAYTABLE4;
 
 typedef struct _ATOM_PPLIB_POWERPLAYTABLE5
@@ -6620,8 +7406,9 @@ typedef struct _ATOM_PPLIB_POWERPLAYTABLE5
     ULONG                      ulNearTDPLimit;
     ULONG                      ulSQRampingThreshold;
     USHORT                     usCACLeakageTableOffset;         // Points to ATOM_PPLIB_CAC_Leakage_Table
-    ULONG                      ulCACLeakage;                    // TBD, this parameter is still under discussion.  Change to ulReserved if not needed.
-    ULONG                      ulReserved;
+    ULONG                      ulCACLeakage;                    // The iLeakage for driver calculated CAC leakage table
+    USHORT                     usTDPODLimit;
+    USHORT                     usLoadLineSlope;                 // in milliOhms * 100
 } ATOM_PPLIB_POWERPLAYTABLE5, *LPATOM_PPLIB_POWERPLAYTABLE5;
 
 //// ATOM_PPLIB_NONCLOCK_INFO::usClassification
@@ -6650,6 +7437,7 @@ typedef struct _ATOM_PPLIB_POWERPLAYTABLE5
 //// ATOM_PPLIB_NONCLOCK_INFO::usClassification2
 #define ATOM_PPLIB_CLASSIFICATION2_LIMITEDPOWERSOURCE_2     0x0001
 #define ATOM_PPLIB_CLASSIFICATION2_ULV                      0x0002
+#define ATOM_PPLIB_CLASSIFICATION2_MVC                      0x0004   //Multi-View Codec (BD-3D)
 
 //// ATOM_PPLIB_NONCLOCK_INFO::ulCapsAndSettings
 #define ATOM_PPLIB_SINGLE_DISPLAY_ONLY           0x00000001
@@ -6673,7 +7461,9 @@ typedef struct _ATOM_PPLIB_POWERPLAYTABLE5
 
 #define ATOM_PPLIB_SOFTWARE_DISABLE_LOADBALANCING        0x00001000
 #define ATOM_PPLIB_SOFTWARE_ENABLE_SLEEP_FOR_TIMESTAMPS  0x00002000
-#define ATOM_PPLIB_DISALLOW_ON_DC                        0x00004000
+
+#define ATOM_PPLIB_DISALLOW_ON_DC                       0x00004000
+
 #define ATOM_PPLIB_ENABLE_VARIBRIGHT                     0x00008000
 
 //memory related flags
@@ -6735,7 +7525,7 @@ typedef struct _ATOM_PPLIB_R600_CLOCK_INFO
 #define ATOM_PPLIB_R600_FLAGS_UVDSAFE           2
 #define ATOM_PPLIB_R600_FLAGS_BACKBIASENABLE    4
 #define ATOM_PPLIB_R600_FLAGS_MEMORY_ODT_OFF    8
-#define ATOM_PPLIB_R600_FLAGS_MEMORY_DLL_OFF    16
+#define ATOM_PPLIB_R600_FLAGS_MEMORY_DLL_OFF   16
 #define ATOM_PPLIB_R600_FLAGS_LOWPOWER         32   // On the RV770 use 'low power' setting (sequencer S0).
 
 typedef struct _ATOM_PPLIB_EVERGREEN_CLOCK_INFO
@@ -6754,6 +7544,24 @@ typedef struct _ATOM_PPLIB_EVERGREEN_CLOCK_INFO
 
 } ATOM_PPLIB_EVERGREEN_CLOCK_INFO;
 
+typedef struct _ATOM_PPLIB_SI_CLOCK_INFO
+{
+      USHORT usEngineClockLow;
+      UCHAR  ucEngineClockHigh;
+
+      USHORT usMemoryClockLow;
+      UCHAR  ucMemoryClockHigh;
+
+      USHORT usVDDC;
+      USHORT usVDDCI;
+      UCHAR  ucPCIEGen;
+      UCHAR  ucUnused1;
+
+      ULONG ulFlags; // ATOM_PPLIB_SI_FLAGS_*, no flag is necessary for now
+
+} ATOM_PPLIB_SI_CLOCK_INFO;
+
+
 typedef struct _ATOM_PPLIB_RS780_CLOCK_INFO
 
 {
@@ -6766,7 +7574,7 @@ typedef struct _ATOM_PPLIB_RS780_CLOCK_INFO
       UCHAR  ucPadding;                   // For proper alignment and size.
       USHORT usVDDC;                      // For the 780, use: None, Low, High, Variable
       UCHAR  ucMaxHTLinkWidth;            // From SBIOS - {2, 4, 8, 16}
-      UCHAR  ucMinHTLinkWidth;            // From SBIOS - {2, 4, 8, 16}. Effective only if CDLW enabled. Minimum down stream width could be bigger as display BW requirement.
+      UCHAR  ucMinHTLinkWidth;            // From SBIOS - {2, 4, 8, 16}. Effective only if CDLW enabled. Minimum down stream width could be bigger as display BW requriement.
       USHORT usHTLinkFreq;                // See definition ATOM_PPLIB_RS780_HTLINKFREQ_xxx or in MHz(>=200).
       ULONG  ulFlags; 
 } ATOM_PPLIB_RS780_CLOCK_INFO;
@@ -6788,9 +7596,7 @@ typedef struct _ATOM_PPLIB_SUMO_CLOCK_INFO{
       USHORT usEngineClockLow;  //clockfrequency & 0xFFFF. The unit is in 10khz
       UCHAR  ucEngineClockHigh; //clockfrequency >> 16. 
       UCHAR  vddcIndex;         //2-bit vddc index;
-      UCHAR  leakage;          //please use 8-bit absolute value, not the 6-bit % value 
-      //please initalize to 0
-      UCHAR  rsv;
+      USHORT tdpLimit;
       //please initalize to 0
       USHORT rsv1;
       //please initialize to 0s
@@ -6813,7 +7619,7 @@ typedef struct _ATOM_PPLIB_STATE_V2
       UCHAR clockInfoIndex[1];
 } ATOM_PPLIB_STATE_V2;
 
-typedef struct StateArray{
+typedef struct _StateArray{
     //how many states we have 
     UCHAR ucNumEntries;
     
@@ -6821,18 +7627,17 @@ typedef struct StateArray{
 }StateArray;
 
 
-typedef struct ClockInfoArray{
+typedef struct _ClockInfoArray{
     //how many clock levels we have
     UCHAR ucNumEntries;
     
-    //sizeof(ATOM_PPLIB_SUMO_CLOCK_INFO)
+    //sizeof(ATOM_PPLIB_CLOCK_INFO)
     UCHAR ucEntrySize;
     
-    //this is for Sumo
-    ATOM_PPLIB_SUMO_CLOCK_INFO clockInfo[1];
+    UCHAR clockInfo[1];
 }ClockInfoArray;
 
-typedef struct NonClockInfoArray{
+typedef struct _NonClockInfoArray{
 
     //how many non-clock levels we have. normally should be same as number of states
     UCHAR ucNumEntries;
@@ -6871,6 +7676,124 @@ typedef struct _ATOM_PPLIB_Clock_Voltage_Limit_Table
     ATOM_PPLIB_Clock_Voltage_Limit_Record entries[1];                  // Dynamically allocate entries.
 }ATOM_PPLIB_Clock_Voltage_Limit_Table;
 
+typedef struct _ATOM_PPLIB_CAC_Leakage_Record
+{
+    USHORT usVddc;  // We use this field for the "fake" standardized VDDC for power calculations                                                  
+    ULONG  ulLeakageValue;
+}ATOM_PPLIB_CAC_Leakage_Record;
+
+typedef struct _ATOM_PPLIB_CAC_Leakage_Table
+{
+    UCHAR ucNumEntries;                                                 // Number of entries.
+    ATOM_PPLIB_CAC_Leakage_Record entries[1];                           // Dynamically allocate entries.
+}ATOM_PPLIB_CAC_Leakage_Table;
+
+typedef struct _ATOM_PPLIB_PhaseSheddingLimits_Record
+{
+    USHORT usVoltage;
+    USHORT usSclkLow;
+    UCHAR  ucSclkHigh;
+    USHORT usMclkLow;
+    UCHAR  ucMclkHigh;
+}ATOM_PPLIB_PhaseSheddingLimits_Record;
+
+typedef struct _ATOM_PPLIB_PhaseSheddingLimits_Table
+{
+    UCHAR ucNumEntries;                                                 // Number of entries.
+    ATOM_PPLIB_PhaseSheddingLimits_Record entries[1];                   // Dynamically allocate entries.
+}ATOM_PPLIB_PhaseSheddingLimits_Table;
+
+typedef struct _VCEClockInfo{
+    USHORT usEVClkLow;
+    UCHAR  ucEVClkHigh;
+    USHORT usECClkLow;
+    UCHAR  ucECClkHigh;
+}VCEClockInfo;
+
+typedef struct _VCEClockInfoArray{
+    UCHAR ucNumEntries;
+    VCEClockInfo entries[1];
+}VCEClockInfoArray;
+
+typedef struct _ATOM_PPLIB_VCE_Clock_Voltage_Limit_Record
+{
+    USHORT usVoltage;
+    UCHAR  ucVCEClockInfoIndex;
+}ATOM_PPLIB_VCE_Clock_Voltage_Limit_Record;
+
+typedef struct _ATOM_PPLIB_VCE_Clock_Voltage_Limit_Table
+{
+    UCHAR numEntries;
+    ATOM_PPLIB_VCE_Clock_Voltage_Limit_Record entries[1];
+}ATOM_PPLIB_VCE_Clock_Voltage_Limit_Table;
+
+typedef struct _ATOM_PPLIB_VCE_State_Record
+{
+    UCHAR  ucVCEClockInfoIndex;
+    UCHAR  ucClockInfoIndex; //highest 2 bits indicates memory p-states, lower 6bits indicates index to ClockInfoArrary
+}ATOM_PPLIB_VCE_State_Record;
+
+typedef struct _ATOM_PPLIB_VCE_State_Table
+{
+    UCHAR numEntries;
+    ATOM_PPLIB_VCE_State_Record entries[1];
+}ATOM_PPLIB_VCE_State_Table;
+
+
+typedef struct _ATOM_PPLIB_VCE_Table
+{
+      UCHAR revid;
+//    VCEClockInfoArray array;
+//    ATOM_PPLIB_VCE_Clock_Voltage_Limit_Table limits;
+//    ATOM_PPLIB_VCE_State_Table states;
+}ATOM_PPLIB_VCE_Table;
+
+
+typedef struct _UVDClockInfo{
+    USHORT usVClkLow;
+    UCHAR  ucVClkHigh;
+    USHORT usDClkLow;
+    UCHAR  ucDClkHigh;
+}UVDClockInfo;
+
+typedef struct _UVDClockInfoArray{
+    UCHAR ucNumEntries;
+    UVDClockInfo entries[1];
+}UVDClockInfoArray;
+
+typedef struct _ATOM_PPLIB_UVD_Clock_Voltage_Limit_Record
+{
+    USHORT usVoltage;
+    UCHAR  ucUVDClockInfoIndex;
+}ATOM_PPLIB_UVD_Clock_Voltage_Limit_Record;
+
+typedef struct _ATOM_PPLIB_UVD_Clock_Voltage_Limit_Table
+{
+    UCHAR numEntries;
+    ATOM_PPLIB_UVD_Clock_Voltage_Limit_Record entries[1];
+}ATOM_PPLIB_UVD_Clock_Voltage_Limit_Table;
+
+typedef struct _ATOM_PPLIB_UVD_State_Record
+{
+    UCHAR  ucUVDClockInfoIndex;
+    UCHAR  ucClockInfoIndex; //highest 2 bits indicates memory p-states, lower 6bits indicates index to ClockInfoArrary
+}ATOM_PPLIB_UVD_State_Record;
+
+typedef struct _ATOM_PPLIB_UVD_State_Table
+{
+    UCHAR numEntries;
+    ATOM_PPLIB_UVD_State_Record entries[1];
+}ATOM_PPLIB_UVD_State_Table;
+
+
+typedef struct _ATOM_PPLIB_UVD_Table
+{
+      UCHAR revid;
+//    UVDClockInfoArray array;
+//    ATOM_PPLIB_UVD_Clock_Voltage_Limit_Table limits;
+//    ATOM_PPLIB_UVD_State_Table states;
+}ATOM_PPLIB_UVD_Table;
+
 /**************************************************************************/
 
 
@@ -7020,4 +7943,68 @@ typedef struct _ATOM_PPLIB_Clock_Voltage_Limit_Table
 
 #pragma pack() // BIOS data must use byte aligment
 
+//
+// AMD ACPI Table
+//
+#pragma pack(1)
+
+typedef struct {
+  ULONG Signature;
+  ULONG TableLength;      //Length
+  UCHAR Revision;
+  UCHAR Checksum;
+  UCHAR OemId[6];
+  UCHAR OemTableId[8];    //UINT64  OemTableId;
+  ULONG OemRevision;
+  ULONG CreatorId;
+  ULONG CreatorRevision;
+} AMD_ACPI_DESCRIPTION_HEADER;
+/*
+//EFI_ACPI_DESCRIPTION_HEADER from AcpiCommon.h
+typedef struct {
+  UINT32  Signature;       //0x0
+  UINT32  Length;          //0x4
+  UINT8   Revision;        //0x8
+  UINT8   Checksum;        //0x9
+  UINT8   OemId[6];        //0xA
+  UINT64  OemTableId;      //0x10
+  UINT32  OemRevision;     //0x18
+  UINT32  CreatorId;       //0x1C
+  UINT32  CreatorRevision; //0x20
+}EFI_ACPI_DESCRIPTION_HEADER;
+*/
+typedef struct {
+  AMD_ACPI_DESCRIPTION_HEADER SHeader;
+  UCHAR TableUUID[16];    //0x24
+  ULONG VBIOSImageOffset; //0x34. Offset to the first GOP_VBIOS_CONTENT block from the beginning of the stucture.
+  ULONG Lib1ImageOffset;  //0x38. Offset to the first GOP_LIB1_CONTENT block from the beginning of the stucture.
+  ULONG Reserved[4];      //0x3C
+}UEFI_ACPI_VFCT;
+
+typedef struct {
+  ULONG  PCIBus;          //0x4C
+  ULONG  PCIDevice;       //0x50
+  ULONG  PCIFunction;     //0x54
+  USHORT VendorID;        //0x58
+  USHORT DeviceID;        //0x5A
+  USHORT SSVID;           //0x5C
+  USHORT SSID;            //0x5E
+  ULONG  Revision;        //0x60
+  ULONG  ImageLength;     //0x64
+}VFCT_IMAGE_HEADER;
+
+
+typedef struct {
+  VFCT_IMAGE_HEADER	VbiosHeader;
+  UCHAR	VbiosContent[1];
+}GOP_VBIOS_CONTENT;
+
+typedef struct {
+  VFCT_IMAGE_HEADER	Lib1Header;
+  UCHAR	Lib1Content[1];
+}GOP_LIB1_CONTENT;
+
+#pragma pack()
+
+
 #endif /* _ATOMBIOS_H */
diff --git a/drivers/gpu/drm/radeon/atombios_crtc.c b/drivers/gpu/drm/radeon/atombios_crtc.c
index 742f17f009a9..b5ff1f7b6f7e 100644
--- a/drivers/gpu/drm/radeon/atombios_crtc.c
+++ b/drivers/gpu/drm/radeon/atombios_crtc.c
@@ -231,6 +231,22 @@ static void atombios_blank_crtc(struct drm_crtc *crtc, int state)
 	atom_execute_table(rdev->mode_info.atom_context, index, (uint32_t *)&args);
 }
 
+static void atombios_powergate_crtc(struct drm_crtc *crtc, int state)
+{
+	struct radeon_crtc *radeon_crtc = to_radeon_crtc(crtc);
+	struct drm_device *dev = crtc->dev;
+	struct radeon_device *rdev = dev->dev_private;
+	int index = GetIndexIntoMasterTable(COMMAND, EnableDispPowerGating);
+	ENABLE_DISP_POWER_GATING_PARAMETERS_V2_1 args;
+
+	memset(&args, 0, sizeof(args));
+
+	args.ucDispPipeId = radeon_crtc->crtc_id;
+	args.ucEnable = state;
+
+	atom_execute_table(rdev->mode_info.atom_context, index, (uint32_t *)&args);
+}
+
 void atombios_crtc_dpms(struct drm_crtc *crtc, int mode)
 {
 	struct drm_device *dev = crtc->dev;
@@ -242,8 +258,11 @@ void atombios_crtc_dpms(struct drm_crtc *crtc, int mode)
 		radeon_crtc->enabled = true;
 		/* adjust pm to dpms changes BEFORE enabling crtcs */
 		radeon_pm_compute_clocks(rdev);
+		/* disable crtc pair power gating before programming */
+		if (ASIC_IS_DCE6(rdev))
+			atombios_powergate_crtc(crtc, ATOM_DISABLE);
 		atombios_enable_crtc(crtc, ATOM_ENABLE);
-		if (ASIC_IS_DCE3(rdev))
+		if (ASIC_IS_DCE3(rdev) && !ASIC_IS_DCE6(rdev))
 			atombios_enable_crtc_memreq(crtc, ATOM_ENABLE);
 		atombios_blank_crtc(crtc, ATOM_DISABLE);
 		drm_vblank_post_modeset(dev, radeon_crtc->crtc_id);
@@ -255,10 +274,29 @@ void atombios_crtc_dpms(struct drm_crtc *crtc, int mode)
 		drm_vblank_pre_modeset(dev, radeon_crtc->crtc_id);
 		if (radeon_crtc->enabled)
 			atombios_blank_crtc(crtc, ATOM_ENABLE);
-		if (ASIC_IS_DCE3(rdev))
+		if (ASIC_IS_DCE3(rdev) && !ASIC_IS_DCE6(rdev))
 			atombios_enable_crtc_memreq(crtc, ATOM_DISABLE);
 		atombios_enable_crtc(crtc, ATOM_DISABLE);
 		radeon_crtc->enabled = false;
+		/* power gating is per-pair */
+		if (ASIC_IS_DCE6(rdev)) {
+			struct drm_crtc *other_crtc;
+			struct radeon_crtc *other_radeon_crtc;
+			list_for_each_entry(other_crtc, &rdev->ddev->mode_config.crtc_list, head) {
+				other_radeon_crtc = to_radeon_crtc(other_crtc);
+				if (((radeon_crtc->crtc_id == 0) && (other_radeon_crtc->crtc_id == 1)) ||
+				    ((radeon_crtc->crtc_id == 1) && (other_radeon_crtc->crtc_id == 0)) ||
+				    ((radeon_crtc->crtc_id == 2) && (other_radeon_crtc->crtc_id == 3)) ||
+				    ((radeon_crtc->crtc_id == 3) && (other_radeon_crtc->crtc_id == 2)) ||
+				    ((radeon_crtc->crtc_id == 4) && (other_radeon_crtc->crtc_id == 5)) ||
+				    ((radeon_crtc->crtc_id == 5) && (other_radeon_crtc->crtc_id == 4))) {
+					/* if both crtcs in the pair are off, enable power gating */
+					if (other_radeon_crtc->enabled == false)
+						atombios_powergate_crtc(crtc, ATOM_ENABLE);
+					break;
+				}
+			}
+		}
 		/* adjust pm to dpms changes AFTER disabling crtcs */
 		radeon_pm_compute_clocks(rdev);
 		break;
@@ -436,7 +474,7 @@ static void atombios_crtc_program_ss(struct radeon_device *rdev,
 			return;
 		}
 		args.v3.ucEnable = enable;
-		if ((ss->percentage == 0) || (ss->type & ATOM_EXTERNAL_SS_MASK))
+		if ((ss->percentage == 0) || (ss->type & ATOM_EXTERNAL_SS_MASK) || ASIC_IS_DCE61(rdev))
 			args.v3.ucEnable = ATOM_DISABLE;
 	} else if (ASIC_IS_DCE4(rdev)) {
 		args.v2.usSpreadSpectrumPercentage = cpu_to_le16(ss->percentage);
@@ -550,8 +588,8 @@ static u32 atombios_adjust_pll(struct drm_crtc *crtc,
 		if (encoder->crtc == crtc) {
 			radeon_encoder = to_radeon_encoder(encoder);
 			connector = radeon_get_connector_for_encoder(encoder);
-			if (connector && connector->display_info.bpc)
-				bpc = connector->display_info.bpc;
+			/* if (connector && connector->display_info.bpc)
+				bpc = connector->display_info.bpc; */
 			encoder_mode = atombios_get_encoder_mode(encoder);
 			is_duallink = radeon_dig_monitor_is_duallink(encoder, mode->clock);
 			if ((radeon_encoder->devices & (ATOM_DEVICE_LCD_SUPPORT | ATOM_DEVICE_DFP_SUPPORT)) ||
@@ -699,7 +737,7 @@ union set_pixel_clock {
 /* on DCE5, make sure the voltage is high enough to support the
  * required disp clk.
  */
-static void atombios_crtc_set_dcpll(struct radeon_device *rdev,
+static void atombios_crtc_set_disp_eng_pll(struct radeon_device *rdev,
 				    u32 dispclk)
 {
 	u8 frev, crev;
@@ -729,7 +767,12 @@ static void atombios_crtc_set_dcpll(struct radeon_device *rdev,
 			 * SetPixelClock provides the dividers
 			 */
 			args.v6.ulDispEngClkFreq = cpu_to_le32(dispclk);
-			args.v6.ucPpll = ATOM_DCPLL;
+			if (ASIC_IS_DCE61(rdev))
+				args.v6.ucPpll = ATOM_EXT_PLL1;
+			else if (ASIC_IS_DCE6(rdev))
+				args.v6.ucPpll = ATOM_PPLL0;
+			else
+				args.v6.ucPpll = ATOM_DCPLL;
 			break;
 		default:
 			DRM_ERROR("Unknown table version %d %d\n", frev, crev);
@@ -922,7 +965,9 @@ static void atombios_crtc_set_pll(struct drm_crtc *crtc, struct drm_display_mode
 		struct radeon_connector_atom_dig *dig_connector =
 			radeon_connector->con_priv;
 		int dp_clock;
-		bpc = connector->display_info.bpc;
+
+		/* if (connector->display_info.bpc)
+			bpc = connector->display_info.bpc; */
 
 		switch (encoder_mode) {
 		case ATOM_ENCODER_MODE_DP_MST:
@@ -1031,6 +1076,7 @@ static int dce4_crtc_do_set_base(struct drm_crtc *crtc,
 	struct radeon_bo *rbo;
 	uint64_t fb_location;
 	uint32_t fb_format, fb_pitch_pixels, tiling_flags;
+	unsigned bankw, bankh, mtaspect, tile_split;
 	u32 fb_swap = EVERGREEN_GRPH_ENDIAN_SWAP(EVERGREEN_GRPH_ENDIAN_NONE);
 	u32 tmp, viewport_w, viewport_h;
 	int r;
@@ -1121,20 +1167,13 @@ static int dce4_crtc_do_set_base(struct drm_crtc *crtc,
 			break;
 		}
 
-		switch ((tmp & 0xf000) >> 12) {
-		case 0: /* 1KB rows */
-		default:
-			fb_format |= EVERGREEN_GRPH_TILE_SPLIT(EVERGREEN_ADDR_SURF_TILE_SPLIT_1KB);
-			break;
-		case 1: /* 2KB rows */
-			fb_format |= EVERGREEN_GRPH_TILE_SPLIT(EVERGREEN_ADDR_SURF_TILE_SPLIT_2KB);
-			break;
-		case 2: /* 4KB rows */
-			fb_format |= EVERGREEN_GRPH_TILE_SPLIT(EVERGREEN_ADDR_SURF_TILE_SPLIT_4KB);
-			break;
-		}
-
 		fb_format |= EVERGREEN_GRPH_ARRAY_MODE(EVERGREEN_GRPH_ARRAY_2D_TILED_THIN1);
+
+		evergreen_tiling_fields(tiling_flags, &bankw, &bankh, &mtaspect, &tile_split);
+		fb_format |= EVERGREEN_GRPH_TILE_SPLIT(tile_split);
+		fb_format |= EVERGREEN_GRPH_BANK_WIDTH(bankw);
+		fb_format |= EVERGREEN_GRPH_BANK_HEIGHT(bankh);
+		fb_format |= EVERGREEN_GRPH_MACRO_TILE_ASPECT(mtaspect);
 	} else if (tiling_flags & RADEON_TILING_MICRO)
 		fb_format |= EVERGREEN_GRPH_ARRAY_MODE(EVERGREEN_GRPH_ARRAY_1D_TILED_THIN1);
 
@@ -1450,7 +1489,36 @@ static int radeon_atom_pick_pll(struct drm_crtc *crtc)
 	struct drm_crtc *test_crtc;
 	uint32_t pll_in_use = 0;
 
-	if (ASIC_IS_DCE4(rdev)) {
+	if (ASIC_IS_DCE61(rdev)) {
+		list_for_each_entry(test_encoder, &dev->mode_config.encoder_list, head) {
+			if (test_encoder->crtc && (test_encoder->crtc == crtc)) {
+				struct radeon_encoder *test_radeon_encoder =
+					to_radeon_encoder(test_encoder);
+				struct radeon_encoder_atom_dig *dig =
+					test_radeon_encoder->enc_priv;
+
+				if ((test_radeon_encoder->encoder_id ==
+				     ENCODER_OBJECT_ID_INTERNAL_UNIPHY) &&
+				    (dig->linkb == false)) /* UNIPHY A uses PPLL2 */
+					return ATOM_PPLL2;
+			}
+		}
+		/* UNIPHY B/C/D/E/F */
+		list_for_each_entry(test_crtc, &dev->mode_config.crtc_list, head) {
+			struct radeon_crtc *radeon_test_crtc;
+
+			if (crtc == test_crtc)
+				continue;
+
+			radeon_test_crtc = to_radeon_crtc(test_crtc);
+			if ((radeon_test_crtc->pll_id == ATOM_PPLL0) ||
+			    (radeon_test_crtc->pll_id == ATOM_PPLL1))
+				pll_in_use |= (1 << radeon_test_crtc->pll_id);
+		}
+		if (!(pll_in_use & 4))
+			return ATOM_PPLL0;
+		return ATOM_PPLL1;
+	} else if (ASIC_IS_DCE4(rdev)) {
 		list_for_each_entry(test_encoder, &dev->mode_config.encoder_list, head) {
 			if (test_encoder->crtc && (test_encoder->crtc == crtc)) {
 				/* in DP mode, the DP ref clock can come from PPLL, DCPLL, or ext clock,
@@ -1489,10 +1557,12 @@ static int radeon_atom_pick_pll(struct drm_crtc *crtc)
 
 }
 
-void radeon_atom_dcpll_init(struct radeon_device *rdev)
+void radeon_atom_disp_eng_pll_init(struct radeon_device *rdev)
 {
 	/* always set DCPLL */
-	if (ASIC_IS_DCE4(rdev)) {
+	if (ASIC_IS_DCE6(rdev))
+		atombios_crtc_set_disp_eng_pll(rdev, rdev->clock.default_dispclk);
+	else if (ASIC_IS_DCE4(rdev)) {
 		struct radeon_atom_ss ss;
 		bool ss_enabled = radeon_atombios_get_asic_ss_info(rdev, &ss,
 								   ASIC_INTERNAL_SS_ON_DCPLL,
@@ -1500,7 +1570,7 @@ void radeon_atom_dcpll_init(struct radeon_device *rdev)
 		if (ss_enabled)
 			atombios_crtc_program_ss(rdev, ATOM_DISABLE, ATOM_DCPLL, &ss);
 		/* XXX: DCE5, make sure voltage, dispclk is high enough */
-		atombios_crtc_set_dcpll(rdev, rdev->clock.default_dispclk);
+		atombios_crtc_set_disp_eng_pll(rdev, rdev->clock.default_dispclk);
 		if (ss_enabled)
 			atombios_crtc_program_ss(rdev, ATOM_ENABLE, ATOM_DCPLL, &ss);
 	}
@@ -1578,6 +1648,8 @@ static void atombios_crtc_commit(struct drm_crtc *crtc)
 static void atombios_crtc_disable(struct drm_crtc *crtc)
 {
 	struct radeon_crtc *radeon_crtc = to_radeon_crtc(crtc);
+	struct drm_device *dev = crtc->dev;
+	struct radeon_device *rdev = dev->dev_private;
 	struct radeon_atom_ss ss;
 
 	atombios_crtc_dpms(crtc, DRM_MODE_DPMS_OFF);
@@ -1589,6 +1661,12 @@ static void atombios_crtc_disable(struct drm_crtc *crtc)
 		atombios_crtc_program_pll(crtc, radeon_crtc->crtc_id, radeon_crtc->pll_id,
 					  0, 0, ATOM_DISABLE, 0, 0, 0, 0, 0, false, &ss);
 		break;
+	case ATOM_PPLL0:
+		/* disable the ppll */
+		if (ASIC_IS_DCE61(rdev))
+			atombios_crtc_program_pll(crtc, radeon_crtc->crtc_id, radeon_crtc->pll_id,
+						  0, 0, ATOM_DISABLE, 0, 0, 0, 0, 0, false, &ss);
+		break;
 	default:
 		break;
 	}
diff --git a/drivers/gpu/drm/radeon/atombios_dp.c b/drivers/gpu/drm/radeon/atombios_dp.c
index 552b436451fd..c57d85664e77 100644
--- a/drivers/gpu/drm/radeon/atombios_dp.c
+++ b/drivers/gpu/drm/radeon/atombios_dp.c
@@ -63,12 +63,12 @@ static int radeon_process_aux_ch(struct radeon_i2c_chan *chan,
 
 	memset(&args, 0, sizeof(args));
 
-	base = (unsigned char *)rdev->mode_info.atom_context->scratch;
+	base = (unsigned char *)(rdev->mode_info.atom_context->scratch + 1);
 
 	memcpy(base, send, send_bytes);
 
-	args.v1.lpAuxRequest = 0;
-	args.v1.lpDataOut = 16;
+	args.v1.lpAuxRequest = 0 + 4;
+	args.v1.lpDataOut = 16 + 4;
 	args.v1.ucDataOutLen = 0;
 	args.v1.ucChannelID = chan->rec.i2c_id;
 	args.v1.ucDelay = delay / 10;
@@ -405,10 +405,13 @@ static void dp_get_adjust_train(u8 link_status[DP_LINK_STATUS_SIZE],
 /* get bpc from the EDID */
 static int convert_bpc_to_bpp(int bpc)
 {
+#if 0
 	if (bpc == 0)
 		return 24;
 	else
 		return bpc * 3;
+#endif
+	return 24;
 }
 
 /* get the max pix clock supported by the link rate and lane num */
@@ -746,7 +749,8 @@ static int radeon_dp_link_train_init(struct radeon_dp_link_train_info *dp_info)
 
 	/* set the lane count on the sink */
 	tmp = dp_info->dp_lane_count;
-	if (dp_info->dpcd[0] >= 0x11)
+	if (dp_info->dpcd[DP_DPCD_REV] >= 0x11 &&
+	    dp_info->dpcd[DP_MAX_LANE_COUNT] & DP_ENHANCED_FRAME_CAP)
 		tmp |= DP_LANE_COUNT_ENHANCED_FRAME_EN;
 	radeon_write_dpcd_reg(dp_info->radeon_connector, DP_LANE_COUNT_SET, tmp);
 
diff --git a/drivers/gpu/drm/radeon/atombios_encoders.c b/drivers/gpu/drm/radeon/atombios_encoders.c
index b88c4608731b..e607c4d7dd98 100644
--- a/drivers/gpu/drm/radeon/atombios_encoders.c
+++ b/drivers/gpu/drm/radeon/atombios_encoders.c
@@ -479,7 +479,7 @@ atombios_get_encoder_mode(struct drm_encoder *encoder)
  * - 2 DIG encoder blocks.
  * DIG1/2 can drive UNIPHY0/1/2 link A or link B
  *
- * DCE 4.0/5.0
+ * DCE 4.0/5.0/6.0
  * - 3 DIG transmitter blocks UNIPHY0/1/2 (links A and B).
  * Supports up to 6 digital outputs
  * - 6 DIG encoder blocks.
@@ -495,7 +495,11 @@ atombios_get_encoder_mode(struct drm_encoder *encoder)
  * - 3 DIG transmitter blocks UNIPHY0/1/2 (links A and B).
  * Supports up to 6 digital outputs
  * - 2 DIG encoder blocks.
+ * llano
  * DIG1/2 can drive UNIPHY0/1/2 link A or link B
+ * ontario
+ * DIG1 drives UNIPHY0/1/2 link A
+ * DIG2 drives UNIPHY0/1/2 link B
  *
  * Routing
  * crtc -> dig encoder -> UNIPHY/LVTMA (1 or 2 links)
@@ -537,7 +541,7 @@ atombios_dig_encoder_setup(struct drm_encoder *encoder, int action, int panel_mo
 		dp_clock = dig_connector->dp_clock;
 		dp_lane_count = dig_connector->dp_lane_count;
 		hpd_id = radeon_connector->hpd.hpd;
-		bpc = connector->display_info.bpc;
+		/* bpc = connector->display_info.bpc; */
 	}
 
 	/* no dig encoder assigned */
@@ -703,6 +707,7 @@ union dig_transmitter_control {
 	DIG_TRANSMITTER_CONTROL_PARAMETERS_V2 v2;
 	DIG_TRANSMITTER_CONTROL_PARAMETERS_V3 v3;
 	DIG_TRANSMITTER_CONTROL_PARAMETERS_V4 v4;
+	DIG_TRANSMITTER_CONTROL_PARAMETERS_V1_5 v5;
 };
 
 void
@@ -723,6 +728,7 @@ atombios_dig_transmitter_setup(struct drm_encoder *encoder, int action, uint8_t
 	int connector_object_id = 0;
 	int igp_lane_info = 0;
 	int dig_encoder = dig->dig_encoder;
+	int hpd_id = RADEON_HPD_NONE;
 
 	if (action == ATOM_TRANSMITTER_ACTION_INIT) {
 		connector = radeon_get_connector_for_encoder_init(encoder);
@@ -738,6 +744,7 @@ atombios_dig_transmitter_setup(struct drm_encoder *encoder, int action, uint8_t
 		struct radeon_connector_atom_dig *dig_connector =
 			radeon_connector->con_priv;
 
+		hpd_id = radeon_connector->hpd.hpd;
 		dp_clock = dig_connector->dp_clock;
 		dp_lane_count = dig_connector->dp_lane_count;
 		connector_object_id =
@@ -1003,6 +1010,60 @@ atombios_dig_transmitter_setup(struct drm_encoder *encoder, int action, uint8_t
 					args.v4.acConfig.fDualLinkConnector = 1;
 			}
 			break;
+		case 5:
+			args.v5.ucAction = action;
+			if (is_dp)
+				args.v5.usSymClock = cpu_to_le16(dp_clock / 10);
+			else
+				args.v5.usSymClock = cpu_to_le16(radeon_encoder->pixel_clock / 10);
+
+			switch (radeon_encoder->encoder_id) {
+			case ENCODER_OBJECT_ID_INTERNAL_UNIPHY:
+				if (dig->linkb)
+					args.v5.ucPhyId = ATOM_PHY_ID_UNIPHYB;
+				else
+					args.v5.ucPhyId = ATOM_PHY_ID_UNIPHYA;
+				break;
+			case ENCODER_OBJECT_ID_INTERNAL_UNIPHY1:
+				if (dig->linkb)
+					args.v5.ucPhyId = ATOM_PHY_ID_UNIPHYD;
+				else
+					args.v5.ucPhyId = ATOM_PHY_ID_UNIPHYC;
+				break;
+			case ENCODER_OBJECT_ID_INTERNAL_UNIPHY2:
+				if (dig->linkb)
+					args.v5.ucPhyId = ATOM_PHY_ID_UNIPHYF;
+				else
+					args.v5.ucPhyId = ATOM_PHY_ID_UNIPHYE;
+				break;
+			}
+			if (is_dp)
+				args.v5.ucLaneNum = dp_lane_count;
+			else if (radeon_encoder->pixel_clock > 165000)
+				args.v5.ucLaneNum = 8;
+			else
+				args.v5.ucLaneNum = 4;
+			args.v5.ucConnObjId = connector_object_id;
+			args.v5.ucDigMode = atombios_get_encoder_mode(encoder);
+
+			if (is_dp && rdev->clock.dp_extclk)
+				args.v5.asConfig.ucPhyClkSrcId = ENCODER_REFCLK_SRC_EXTCLK;
+			else
+				args.v5.asConfig.ucPhyClkSrcId = pll_id;
+
+			if (is_dp)
+				args.v5.asConfig.ucCoherentMode = 1; /* DP requires coherent */
+			else if (radeon_encoder->devices & (ATOM_DEVICE_DFP_SUPPORT)) {
+				if (dig->coherent_mode)
+					args.v5.asConfig.ucCoherentMode = 1;
+			}
+			if (hpd_id == RADEON_HPD_NONE)
+				args.v5.asConfig.ucHPDSel = 0;
+			else
+				args.v5.asConfig.ucHPDSel = hpd_id + 1;
+			args.v5.ucDigEncoderSel = 1 << dig_encoder;
+			args.v5.ucDPLaneSet = lane_set;
+			break;
 		default:
 			DRM_ERROR("Unknown table version %d, %d\n", frev, crev);
 			break;
@@ -1098,7 +1159,7 @@ atombios_external_encoder_setup(struct drm_encoder *encoder,
 		dp_lane_count = dig_connector->dp_lane_count;
 		connector_object_id =
 			(radeon_connector->connector_object_id & OBJECT_ID_MASK) >> OBJECT_ID_SHIFT;
-		bpc = connector->display_info.bpc;
+		/* bpc = connector->display_info.bpc; */
 	}
 
 	memset(&args, 0, sizeof(args));
@@ -1377,7 +1438,7 @@ radeon_atom_encoder_dpms_ext(struct drm_encoder *encoder,
 	switch (mode) {
 	case DRM_MODE_DPMS_ON:
 	default:
-		if (ASIC_IS_DCE41(rdev)) {
+		if (ASIC_IS_DCE41(rdev) || ASIC_IS_DCE61(rdev)) {
 			atombios_external_encoder_setup(encoder, ext_encoder,
 							EXTERNAL_ENCODER_ACTION_V3_ENABLE_OUTPUT);
 			atombios_external_encoder_setup(encoder, ext_encoder,
@@ -1388,7 +1449,7 @@ radeon_atom_encoder_dpms_ext(struct drm_encoder *encoder,
 	case DRM_MODE_DPMS_STANDBY:
 	case DRM_MODE_DPMS_SUSPEND:
 	case DRM_MODE_DPMS_OFF:
-		if (ASIC_IS_DCE41(rdev)) {
+		if (ASIC_IS_DCE41(rdev) || ASIC_IS_DCE61(rdev)) {
 			atombios_external_encoder_setup(encoder, ext_encoder,
 							EXTERNAL_ENCODER_ACTION_V3_ENCODER_BLANKING);
 			atombios_external_encoder_setup(encoder, ext_encoder,
@@ -1761,7 +1822,7 @@ radeon_atom_encoder_init(struct radeon_device *rdev)
 			break;
 		}
 
-		if (ext_encoder && ASIC_IS_DCE41(rdev))
+		if (ext_encoder && (ASIC_IS_DCE41(rdev) || ASIC_IS_DCE61(rdev)))
 			atombios_external_encoder_setup(encoder, ext_encoder,
 							EXTERNAL_ENCODER_ACTION_V3_ENCODER_INIT);
 	}
@@ -1850,7 +1911,7 @@ radeon_atom_encoder_mode_set(struct drm_encoder *encoder,
 	}
 
 	if (ext_encoder) {
-		if (ASIC_IS_DCE41(rdev))
+		if (ASIC_IS_DCE41(rdev) || ASIC_IS_DCE61(rdev))
 			atombios_external_encoder_setup(encoder, ext_encoder,
 							EXTERNAL_ENCODER_ACTION_V3_ENCODER_SETUP);
 		else
diff --git a/drivers/gpu/drm/radeon/atombios_i2c.c b/drivers/gpu/drm/radeon/atombios_i2c.c
new file mode 100644
index 000000000000..44d87b6b4220
--- /dev/null
+++ b/drivers/gpu/drm/radeon/atombios_i2c.c
@@ -0,0 +1,139 @@
+/*
+ * Copyright 2011 Advanced Micro Devices, 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: Alex Deucher
+ *
+ */
+#include "drmP.h"
+#include "radeon_drm.h"
+#include "radeon.h"
+#include "atom.h"
+
+#define TARGET_HW_I2C_CLOCK 50
+
+/* these are a limitation of ProcessI2cChannelTransaction not the hw */
+#define ATOM_MAX_HW_I2C_WRITE 2
+#define ATOM_MAX_HW_I2C_READ  255
+
+static int radeon_process_i2c_ch(struct radeon_i2c_chan *chan,
+				 u8 slave_addr, u8 flags,
+				 u8 *buf, u8 num)
+{
+	struct drm_device *dev = chan->dev;
+	struct radeon_device *rdev = dev->dev_private;
+	PROCESS_I2C_CHANNEL_TRANSACTION_PS_ALLOCATION args;
+	int index = GetIndexIntoMasterTable(COMMAND, ProcessI2cChannelTransaction);
+	unsigned char *base;
+	u16 out;
+
+	memset(&args, 0, sizeof(args));
+
+	base = (unsigned char *)rdev->mode_info.atom_context->scratch;
+
+	if (flags & HW_I2C_WRITE) {
+		if (num > ATOM_MAX_HW_I2C_WRITE) {
+			DRM_ERROR("hw i2c: tried to write too many bytes (%d vs 2)\n", num);
+			return -EINVAL;
+		}
+		memcpy(&out, buf, num);
+		args.lpI2CDataOut = cpu_to_le16(out);
+	} else {
+		if (num > ATOM_MAX_HW_I2C_READ) {
+			DRM_ERROR("hw i2c: tried to read too many bytes (%d vs 255)\n", num);
+			return -EINVAL;
+		}
+	}
+
+	args.ucI2CSpeed = TARGET_HW_I2C_CLOCK;
+	args.ucRegIndex = 0;
+	args.ucTransBytes = num;
+	args.ucSlaveAddr = slave_addr << 1;
+	args.ucLineNumber = chan->rec.i2c_id;
+
+	atom_execute_table(rdev->mode_info.atom_context, index, (uint32_t *)&args);
+
+	/* error */
+	if (args.ucStatus != HW_ASSISTED_I2C_STATUS_SUCCESS) {
+		DRM_DEBUG_KMS("hw_i2c error\n");
+		return -EIO;
+	}
+
+	if (!(flags & HW_I2C_WRITE))
+		memcpy(buf, base, num);
+
+	return 0;
+}
+
+int radeon_atom_hw_i2c_xfer(struct i2c_adapter *i2c_adap,
+			    struct i2c_msg *msgs, int num)
+{
+	struct radeon_i2c_chan *i2c = i2c_get_adapdata(i2c_adap);
+	struct i2c_msg *p;
+	int i, remaining, current_count, buffer_offset, max_bytes, ret;
+	u8 buf = 0, flags;
+
+	/* check for bus probe */
+	p = &msgs[0];
+	if ((num == 1) && (p->len == 0)) {
+		ret = radeon_process_i2c_ch(i2c,
+					    p->addr, HW_I2C_WRITE,
+					    &buf, 1);
+		if (ret)
+			return ret;
+		else
+			return num;
+	}
+
+	for (i = 0; i < num; i++) {
+		p = &msgs[i];
+		remaining = p->len;
+		buffer_offset = 0;
+		/* max_bytes are a limitation of ProcessI2cChannelTransaction not the hw */
+		if (p->flags & I2C_M_RD) {
+			max_bytes = ATOM_MAX_HW_I2C_READ;
+			flags = HW_I2C_READ;
+		} else {
+			max_bytes = ATOM_MAX_HW_I2C_WRITE;
+			flags = HW_I2C_WRITE;
+		}
+		while (remaining) {
+			if (remaining > max_bytes)
+				current_count = max_bytes;
+			else
+				current_count = remaining;
+			ret = radeon_process_i2c_ch(i2c,
+						    p->addr, flags,
+						    &p->buf[buffer_offset], current_count);
+			if (ret)
+				return ret;
+			remaining -= current_count;
+			buffer_offset += current_count;
+		}
+	}
+
+	return num;
+}
+
+u32 radeon_atom_hw_i2c_func(struct i2c_adapter *adap)
+{
+	return I2C_FUNC_I2C | I2C_FUNC_SMBUS_EMUL;
+}
+
diff --git a/drivers/gpu/drm/radeon/cayman_blit_shaders.c b/drivers/gpu/drm/radeon/cayman_blit_shaders.c
index 7b4eeb7b4a8c..19a0114d2e3b 100644
--- a/drivers/gpu/drm/radeon/cayman_blit_shaders.c
+++ b/drivers/gpu/drm/radeon/cayman_blit_shaders.c
@@ -24,6 +24,7 @@
  *     Alex Deucher <alexander.deucher@amd.com>
  */
 
+#include <linux/bug.h>
 #include <linux/types.h>
 #include <linux/kernel.h>
 
diff --git a/drivers/gpu/drm/radeon/evergreen.c b/drivers/gpu/drm/radeon/evergreen.c
index ae09fe82afbc..cfa372cb1cb3 100644
--- a/drivers/gpu/drm/radeon/evergreen.c
+++ b/drivers/gpu/drm/radeon/evergreen.c
@@ -43,6 +43,37 @@ 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);
 
+void evergreen_tiling_fields(unsigned tiling_flags, unsigned *bankw,
+			     unsigned *bankh, unsigned *mtaspect,
+			     unsigned *tile_split)
+{
+	*bankw = (tiling_flags >> RADEON_TILING_EG_BANKW_SHIFT) & RADEON_TILING_EG_BANKW_MASK;
+	*bankh = (tiling_flags >> RADEON_TILING_EG_BANKH_SHIFT) & RADEON_TILING_EG_BANKH_MASK;
+	*mtaspect = (tiling_flags >> RADEON_TILING_EG_MACRO_TILE_ASPECT_SHIFT) & RADEON_TILING_EG_MACRO_TILE_ASPECT_MASK;
+	*tile_split = (tiling_flags >> RADEON_TILING_EG_TILE_SPLIT_SHIFT) & RADEON_TILING_EG_TILE_SPLIT_MASK;
+	switch (*bankw) {
+	default:
+	case 1: *bankw = EVERGREEN_ADDR_SURF_BANK_WIDTH_1; break;
+	case 2: *bankw = EVERGREEN_ADDR_SURF_BANK_WIDTH_2; break;
+	case 4: *bankw = EVERGREEN_ADDR_SURF_BANK_WIDTH_4; break;
+	case 8: *bankw = EVERGREEN_ADDR_SURF_BANK_WIDTH_8; break;
+	}
+	switch (*bankh) {
+	default:
+	case 1: *bankh = EVERGREEN_ADDR_SURF_BANK_HEIGHT_1; break;
+	case 2: *bankh = EVERGREEN_ADDR_SURF_BANK_HEIGHT_2; break;
+	case 4: *bankh = EVERGREEN_ADDR_SURF_BANK_HEIGHT_4; break;
+	case 8: *bankh = EVERGREEN_ADDR_SURF_BANK_HEIGHT_8; break;
+	}
+	switch (*mtaspect) {
+	default:
+	case 1: *mtaspect = EVERGREEN_ADDR_SURF_MACRO_TILE_ASPECT_1; break;
+	case 2: *mtaspect = EVERGREEN_ADDR_SURF_MACRO_TILE_ASPECT_2; break;
+	case 4: *mtaspect = EVERGREEN_ADDR_SURF_MACRO_TILE_ASPECT_4; break;
+	case 8: *mtaspect = EVERGREEN_ADDR_SURF_MACRO_TILE_ASPECT_8; break;
+	}
+}
+
 void evergreen_fix_pci_max_read_req_size(struct radeon_device *rdev)
 {
 	u16 ctl, v;
@@ -68,6 +99,25 @@ void evergreen_fix_pci_max_read_req_size(struct radeon_device *rdev)
 	}
 }
 
+void dce4_wait_for_vblank(struct radeon_device *rdev, int crtc)
+{
+	struct radeon_crtc *radeon_crtc = rdev->mode_info.crtcs[crtc];
+	int i;
+
+	if (RREG32(EVERGREEN_CRTC_CONTROL + radeon_crtc->crtc_offset) & EVERGREEN_CRTC_MASTER_EN) {
+		for (i = 0; i < rdev->usec_timeout; i++) {
+			if (!(RREG32(EVERGREEN_CRTC_STATUS + radeon_crtc->crtc_offset) & EVERGREEN_CRTC_V_BLANK))
+				break;
+			udelay(1);
+		}
+		for (i = 0; i < rdev->usec_timeout; i++) {
+			if (RREG32(EVERGREEN_CRTC_STATUS + radeon_crtc->crtc_offset) & EVERGREEN_CRTC_V_BLANK)
+				break;
+			udelay(1);
+		}
+	}
+}
+
 void evergreen_pre_page_flip(struct radeon_device *rdev, int crtc)
 {
 	/* enable the pflip int */
@@ -531,7 +581,7 @@ static u32 evergreen_line_buffer_adjust(struct radeon_device *rdev,
 	return 0;
 }
 
-static u32 evergreen_get_number_of_dram_channels(struct radeon_device *rdev)
+u32 evergreen_get_number_of_dram_channels(struct radeon_device *rdev)
 {
 	u32 tmp = RREG32(MC_SHARED_CHMAP);
 
@@ -1278,7 +1328,10 @@ void evergreen_mc_program(struct radeon_device *rdev)
 			rdev->mc.vram_end >> 12);
 	}
 	WREG32(MC_VM_SYSTEM_APERTURE_DEFAULT_ADDR, rdev->vram_scratch.gpu_addr >> 12);
-	if (rdev->flags & RADEON_IS_IGP) {
+	/* llano/ontario only */
+	if ((rdev->family == CHIP_PALM) ||
+	    (rdev->family == CHIP_SUMO) ||
+	    (rdev->family == CHIP_SUMO2)) {
 		tmp = RREG32(MC_FUS_VM_FB_OFFSET) & 0x000FFFFF;
 		tmp |= ((rdev->mc.vram_end >> 20) & 0xF) << 24;
 		tmp |= ((rdev->mc.vram_start >> 20) & 0xF) << 20;
@@ -1489,7 +1542,7 @@ int evergreen_cp_resume(struct radeon_device *rdev)
 
 	evergreen_cp_start(rdev);
 	ring->ready = true;
-	r = radeon_ring_test(rdev, ring);
+	r = radeon_ring_test(rdev, RADEON_RING_TYPE_GFX_INDEX, ring);
 	if (r) {
 		ring->ready = false;
 		return r;
@@ -1922,7 +1975,9 @@ static void evergreen_gpu_init(struct radeon_device *rdev)
 
 
 	mc_shared_chmap = RREG32(MC_SHARED_CHMAP);
-	if (rdev->flags & RADEON_IS_IGP)
+	if ((rdev->family == CHIP_PALM) ||
+	    (rdev->family == CHIP_SUMO) ||
+	    (rdev->family == CHIP_SUMO2))
 		mc_arb_ramcfg = RREG32(FUS_MC_ARB_RAMCFG);
 	else
 		mc_arb_ramcfg = RREG32(MC_ARB_RAMCFG);
@@ -2312,7 +2367,9 @@ int evergreen_mc_init(struct radeon_device *rdev)
 
 	/* Get VRAM informations */
 	rdev->mc.vram_is_ddr = true;
-	if (rdev->flags & RADEON_IS_IGP)
+	if ((rdev->family == CHIP_PALM) ||
+	    (rdev->family == CHIP_SUMO) ||
+	    (rdev->family == CHIP_SUMO2))
 		tmp = RREG32(FUS_MC_ARB_RAMCFG);
 	else
 		tmp = RREG32(MC_ARB_RAMCFG);
@@ -2344,12 +2401,14 @@ int evergreen_mc_init(struct radeon_device *rdev)
 	rdev->mc.aper_base = pci_resource_start(rdev->pdev, 0);
 	rdev->mc.aper_size = pci_resource_len(rdev->pdev, 0);
 	/* Setup GPU memory space */
-	if (rdev->flags & RADEON_IS_IGP) {
+	if ((rdev->family == CHIP_PALM) ||
+	    (rdev->family == CHIP_SUMO) ||
+	    (rdev->family == CHIP_SUMO2)) {
 		/* size in bytes on fusion */
 		rdev->mc.mc_vram_size = RREG32(CONFIG_MEMSIZE);
 		rdev->mc.real_vram_size = RREG32(CONFIG_MEMSIZE);
 	} else {
-		/* size in MB on evergreen */
+		/* size in MB on evergreen/cayman/tn */
 		rdev->mc.mc_vram_size = RREG32(CONFIG_MEMSIZE) * 1024 * 1024;
 		rdev->mc.real_vram_size = RREG32(CONFIG_MEMSIZE) * 1024 * 1024;
 	}
@@ -2507,7 +2566,9 @@ void evergreen_disable_interrupt_state(struct radeon_device *rdev)
 		WREG32(GRPH_INT_CONTROL + EVERGREEN_CRTC5_REGISTER_OFFSET, 0);
 	}
 
-	WREG32(DACA_AUTODETECT_INT_CONTROL, 0);
+	/* only one DAC on DCE6 */
+	if (!ASIC_IS_DCE6(rdev))
+		WREG32(DACA_AUTODETECT_INT_CONTROL, 0);
 	WREG32(DACB_AUTODETECT_INT_CONTROL, 0);
 
 	tmp = RREG32(DC_HPD1_INT_CONTROL) & DC_HPDx_INT_POLARITY;
@@ -3147,7 +3208,7 @@ static int evergreen_startup(struct radeon_device *rdev)
 	r = evergreen_blit_init(rdev);
 	if (r) {
 		r600_blit_fini(rdev);
-		rdev->asic->copy = NULL;
+		rdev->asic->copy.copy = NULL;
 		dev_warn(rdev->dev, "failed blitter (%d) falling back to memcpy\n", r);
 	}
 
@@ -3187,10 +3248,11 @@ static int evergreen_startup(struct radeon_device *rdev)
 	if (r)
 		return r;
 
-	r = r600_ib_test(rdev, RADEON_RING_TYPE_GFX_INDEX);
+	r = radeon_ib_test(rdev, RADEON_RING_TYPE_GFX_INDEX, &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]);
 	if (r) {
 		DRM_ERROR("radeon: failed testing IB (%d).\n", r);
 		rdev->accel_working = false;
+		return r;
 	}
 
 	r = r600_audio_init(rdev);
@@ -3222,6 +3284,7 @@ int evergreen_resume(struct radeon_device *rdev)
 	r = evergreen_startup(rdev);
 	if (r) {
 		DRM_ERROR("evergreen startup failed on resume\n");
+		rdev->accel_working = false;
 		return r;
 	}
 
diff --git a/drivers/gpu/drm/radeon/evergreen_blit_kms.c b/drivers/gpu/drm/radeon/evergreen_blit_kms.c
index 2379849515c7..222acd2d33df 100644
--- a/drivers/gpu/drm/radeon/evergreen_blit_kms.c
+++ b/drivers/gpu/drm/radeon/evergreen_blit_kms.c
@@ -32,17 +32,7 @@
 #include "evergreend.h"
 #include "evergreen_blit_shaders.h"
 #include "cayman_blit_shaders.h"
-
-#define DI_PT_RECTLIST        0x11
-#define DI_INDEX_SIZE_16_BIT  0x0
-#define DI_SRC_SEL_AUTO_INDEX 0x2
-
-#define FMT_8                 0x1
-#define FMT_5_6_5             0x8
-#define FMT_8_8_8_8           0x1a
-#define COLOR_8               0x1
-#define COLOR_5_6_5           0x8
-#define COLOR_8_8_8_8         0x1a
+#include "radeon_blit_common.h"
 
 /* emits 17 */
 static void
@@ -236,7 +226,7 @@ set_scissors(struct radeon_device *rdev, int x1, int y1,
 		x1 = 1;
 	if (y2 == 0)
 		y1 = 1;
-	if (rdev->family == CHIP_CAYMAN) {
+	if (rdev->family >= CHIP_CAYMAN) {
 		if ((x2 == 1) && (y2 == 1))
 			x2 = 2;
 	}
diff --git a/drivers/gpu/drm/radeon/evergreen_blit_shaders.c b/drivers/gpu/drm/radeon/evergreen_blit_shaders.c
index 3a10399e0066..f85c0af115b5 100644
--- a/drivers/gpu/drm/radeon/evergreen_blit_shaders.c
+++ b/drivers/gpu/drm/radeon/evergreen_blit_shaders.c
@@ -24,6 +24,7 @@
  *     Alex Deucher <alexander.deucher@amd.com>
  */
 
+#include <linux/bug.h>
 #include <linux/types.h>
 #include <linux/kernel.h>
 
diff --git a/drivers/gpu/drm/radeon/evergreen_cs.c b/drivers/gpu/drm/radeon/evergreen_cs.c
index 8e8cd85e5c00..70089d32b80f 100644
--- a/drivers/gpu/drm/radeon/evergreen_cs.c
+++ b/drivers/gpu/drm/radeon/evergreen_cs.c
@@ -31,6 +31,9 @@
 #include "evergreen_reg_safe.h"
 #include "cayman_reg_safe.h"
 
+#define MAX(a,b)                   (((a)>(b))?(a):(b))
+#define MIN(a,b)                   (((a)<(b))?(a):(b))
+
 static int evergreen_cs_packet_next_reloc(struct radeon_cs_parser *p,
 					  struct radeon_cs_reloc **cs_reloc);
 
@@ -40,42 +43,46 @@ struct evergreen_cs_track {
 	u32			npipes;
 	u32			row_size;
 	/* value we track */
-	u32			nsamples;
-	u32			cb_color_base_last[12];
+	u32			nsamples;		/* unused */
 	struct radeon_bo	*cb_color_bo[12];
 	u32			cb_color_bo_offset[12];
-	struct radeon_bo	*cb_color_fmask_bo[8];
-	struct radeon_bo	*cb_color_cmask_bo[8];
+	struct radeon_bo	*cb_color_fmask_bo[8];	/* unused */
+	struct radeon_bo	*cb_color_cmask_bo[8];	/* unused */
 	u32			cb_color_info[12];
 	u32			cb_color_view[12];
-	u32			cb_color_pitch_idx[12];
-	u32			cb_color_slice_idx[12];
-	u32			cb_color_dim_idx[12];
-	u32			cb_color_dim[12];
 	u32			cb_color_pitch[12];
 	u32			cb_color_slice[12];
-	u32			cb_color_cmask_slice[8];
-	u32			cb_color_fmask_slice[8];
+	u32			cb_color_attrib[12];
+	u32			cb_color_cmask_slice[8];/* unused */
+	u32			cb_color_fmask_slice[8];/* unused */
 	u32			cb_target_mask;
-	u32			cb_shader_mask;
+	u32			cb_shader_mask; /* unused */
 	u32			vgt_strmout_config;
 	u32			vgt_strmout_buffer_config;
+	struct radeon_bo	*vgt_strmout_bo[4];
+	u32			vgt_strmout_bo_offset[4];
+	u32			vgt_strmout_size[4];
 	u32			db_depth_control;
 	u32			db_depth_view;
+	u32			db_depth_slice;
 	u32			db_depth_size;
-	u32			db_depth_size_idx;
 	u32			db_z_info;
-	u32			db_z_idx;
 	u32			db_z_read_offset;
 	u32			db_z_write_offset;
 	struct radeon_bo	*db_z_read_bo;
 	struct radeon_bo	*db_z_write_bo;
 	u32			db_s_info;
-	u32			db_s_idx;
 	u32			db_s_read_offset;
 	u32			db_s_write_offset;
 	struct radeon_bo	*db_s_read_bo;
 	struct radeon_bo	*db_s_write_bo;
+	bool			sx_misc_kill_all_prims;
+	bool			cb_dirty;
+	bool			db_dirty;
+	bool			streamout_dirty;
+	u32			htile_offset;
+	u32			htile_surface;
+	struct radeon_bo	*htile_bo;
 };
 
 static u32 evergreen_cs_get_aray_mode(u32 tiling_flags)
@@ -103,19 +110,6 @@ static u32 evergreen_cs_get_num_banks(u32 nbanks)
 	}
 }
 
-static u32 evergreen_cs_get_tile_split(u32 row_size)
-{
-	switch (row_size) {
-	case 1:
-	default:
-		return ADDR_SURF_TILE_SPLIT_1KB;
-	case 2:
-		return ADDR_SURF_TILE_SPLIT_2KB;
-	case 4:
-		return ADDR_SURF_TILE_SPLIT_4KB;
-	}
-}
-
 static void evergreen_cs_track_init(struct evergreen_cs_track *track)
 {
 	int i;
@@ -128,50 +122,820 @@ static void evergreen_cs_track_init(struct evergreen_cs_track *track)
 	}
 
 	for (i = 0; i < 12; i++) {
-		track->cb_color_base_last[i] = 0;
 		track->cb_color_bo[i] = NULL;
 		track->cb_color_bo_offset[i] = 0xFFFFFFFF;
 		track->cb_color_info[i] = 0;
-		track->cb_color_view[i] = 0;
-		track->cb_color_pitch_idx[i] = 0;
-		track->cb_color_slice_idx[i] = 0;
-		track->cb_color_dim[i] = 0;
+		track->cb_color_view[i] = 0xFFFFFFFF;
 		track->cb_color_pitch[i] = 0;
 		track->cb_color_slice[i] = 0;
-		track->cb_color_dim[i] = 0;
 	}
 	track->cb_target_mask = 0xFFFFFFFF;
 	track->cb_shader_mask = 0xFFFFFFFF;
+	track->cb_dirty = true;
 
 	track->db_depth_view = 0xFFFFC000;
 	track->db_depth_size = 0xFFFFFFFF;
-	track->db_depth_size_idx = 0;
 	track->db_depth_control = 0xFFFFFFFF;
 	track->db_z_info = 0xFFFFFFFF;
-	track->db_z_idx = 0xFFFFFFFF;
 	track->db_z_read_offset = 0xFFFFFFFF;
 	track->db_z_write_offset = 0xFFFFFFFF;
 	track->db_z_read_bo = NULL;
 	track->db_z_write_bo = NULL;
 	track->db_s_info = 0xFFFFFFFF;
-	track->db_s_idx = 0xFFFFFFFF;
 	track->db_s_read_offset = 0xFFFFFFFF;
 	track->db_s_write_offset = 0xFFFFFFFF;
 	track->db_s_read_bo = NULL;
 	track->db_s_write_bo = NULL;
+	track->db_dirty = true;
+	track->htile_bo = NULL;
+	track->htile_offset = 0xFFFFFFFF;
+	track->htile_surface = 0;
+
+	for (i = 0; i < 4; i++) {
+		track->vgt_strmout_size[i] = 0;
+		track->vgt_strmout_bo[i] = NULL;
+		track->vgt_strmout_bo_offset[i] = 0xFFFFFFFF;
+	}
+	track->streamout_dirty = true;
+	track->sx_misc_kill_all_prims = false;
 }
 
-static int evergreen_cs_track_check(struct radeon_cs_parser *p)
+struct eg_surface {
+	/* value gathered from cs */
+	unsigned	nbx;
+	unsigned	nby;
+	unsigned	format;
+	unsigned	mode;
+	unsigned	nbanks;
+	unsigned	bankw;
+	unsigned	bankh;
+	unsigned	tsplit;
+	unsigned	mtilea;
+	unsigned	nsamples;
+	/* output value */
+	unsigned	bpe;
+	unsigned	layer_size;
+	unsigned	palign;
+	unsigned	halign;
+	unsigned long	base_align;
+};
+
+static int evergreen_surface_check_linear(struct radeon_cs_parser *p,
+					  struct eg_surface *surf,
+					  const char *prefix)
+{
+	surf->layer_size = surf->nbx * surf->nby * surf->bpe * surf->nsamples;
+	surf->base_align = surf->bpe;
+	surf->palign = 1;
+	surf->halign = 1;
+	return 0;
+}
+
+static int evergreen_surface_check_linear_aligned(struct radeon_cs_parser *p,
+						  struct eg_surface *surf,
+						  const char *prefix)
+{
+	struct evergreen_cs_track *track = p->track;
+	unsigned palign;
+
+	palign = MAX(64, track->group_size / surf->bpe);
+	surf->layer_size = surf->nbx * surf->nby * surf->bpe * surf->nsamples;
+	surf->base_align = track->group_size;
+	surf->palign = palign;
+	surf->halign = 1;
+	if (surf->nbx & (palign - 1)) {
+		if (prefix) {
+			dev_warn(p->dev, "%s:%d %s pitch %d invalid must be aligned with %d\n",
+				 __func__, __LINE__, prefix, surf->nbx, palign);
+		}
+		return -EINVAL;
+	}
+	return 0;
+}
+
+static int evergreen_surface_check_1d(struct radeon_cs_parser *p,
+				      struct eg_surface *surf,
+				      const char *prefix)
+{
+	struct evergreen_cs_track *track = p->track;
+	unsigned palign;
+
+	palign = track->group_size / (8 * surf->bpe * surf->nsamples);
+	palign = MAX(8, palign);
+	surf->layer_size = surf->nbx * surf->nby * surf->bpe;
+	surf->base_align = track->group_size;
+	surf->palign = palign;
+	surf->halign = 8;
+	if ((surf->nbx & (palign - 1))) {
+		if (prefix) {
+			dev_warn(p->dev, "%s:%d %s pitch %d invalid must be aligned with %d (%d %d %d)\n",
+				 __func__, __LINE__, prefix, surf->nbx, palign,
+				 track->group_size, surf->bpe, surf->nsamples);
+		}
+		return -EINVAL;
+	}
+	if ((surf->nby & (8 - 1))) {
+		if (prefix) {
+			dev_warn(p->dev, "%s:%d %s height %d invalid must be aligned with 8\n",
+				 __func__, __LINE__, prefix, surf->nby);
+		}
+		return -EINVAL;
+	}
+	return 0;
+}
+
+static int evergreen_surface_check_2d(struct radeon_cs_parser *p,
+				      struct eg_surface *surf,
+				      const char *prefix)
+{
+	struct evergreen_cs_track *track = p->track;
+	unsigned palign, halign, tileb, slice_pt;
+
+	tileb = 64 * surf->bpe * surf->nsamples;
+	palign = track->group_size / (8 * surf->bpe * surf->nsamples);
+	palign = MAX(8, palign);
+	slice_pt = 1;
+	if (tileb > surf->tsplit) {
+		slice_pt = tileb / surf->tsplit;
+	}
+	tileb = tileb / slice_pt;
+	/* macro tile width & height */
+	palign = (8 * surf->bankw * track->npipes) * surf->mtilea;
+	halign = (8 * surf->bankh * surf->nbanks) / surf->mtilea;
+	surf->layer_size = surf->nbx * surf->nby * surf->bpe * slice_pt;
+	surf->base_align = (palign / 8) * (halign / 8) * tileb;
+	surf->palign = palign;
+	surf->halign = halign;
+
+	if ((surf->nbx & (palign - 1))) {
+		if (prefix) {
+			dev_warn(p->dev, "%s:%d %s pitch %d invalid must be aligned with %d\n",
+				 __func__, __LINE__, prefix, surf->nbx, palign);
+		}
+		return -EINVAL;
+	}
+	if ((surf->nby & (halign - 1))) {
+		if (prefix) {
+			dev_warn(p->dev, "%s:%d %s height %d invalid must be aligned with %d\n",
+				 __func__, __LINE__, prefix, surf->nby, halign);
+		}
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static int evergreen_surface_check(struct radeon_cs_parser *p,
+				   struct eg_surface *surf,
+				   const char *prefix)
+{
+	/* some common value computed here */
+	surf->bpe = r600_fmt_get_blocksize(surf->format);
+
+	switch (surf->mode) {
+	case ARRAY_LINEAR_GENERAL:
+		return evergreen_surface_check_linear(p, surf, prefix);
+	case ARRAY_LINEAR_ALIGNED:
+		return evergreen_surface_check_linear_aligned(p, surf, prefix);
+	case ARRAY_1D_TILED_THIN1:
+		return evergreen_surface_check_1d(p, surf, prefix);
+	case ARRAY_2D_TILED_THIN1:
+		return evergreen_surface_check_2d(p, surf, prefix);
+	default:
+		dev_warn(p->dev, "%s:%d %s invalid array mode %d\n",
+				__func__, __LINE__, prefix, surf->mode);
+		return -EINVAL;
+	}
+	return -EINVAL;
+}
+
+static int evergreen_surface_value_conv_check(struct radeon_cs_parser *p,
+					      struct eg_surface *surf,
+					      const char *prefix)
+{
+	switch (surf->mode) {
+	case ARRAY_2D_TILED_THIN1:
+		break;
+	case ARRAY_LINEAR_GENERAL:
+	case ARRAY_LINEAR_ALIGNED:
+	case ARRAY_1D_TILED_THIN1:
+		return 0;
+	default:
+		dev_warn(p->dev, "%s:%d %s invalid array mode %d\n",
+				__func__, __LINE__, prefix, surf->mode);
+		return -EINVAL;
+	}
+
+	switch (surf->nbanks) {
+	case 0: surf->nbanks = 2; break;
+	case 1: surf->nbanks = 4; break;
+	case 2: surf->nbanks = 8; break;
+	case 3: surf->nbanks = 16; break;
+	default:
+		dev_warn(p->dev, "%s:%d %s invalid number of banks %d\n",
+			 __func__, __LINE__, prefix, surf->nbanks);
+		return -EINVAL;
+	}
+	switch (surf->bankw) {
+	case 0: surf->bankw = 1; break;
+	case 1: surf->bankw = 2; break;
+	case 2: surf->bankw = 4; break;
+	case 3: surf->bankw = 8; break;
+	default:
+		dev_warn(p->dev, "%s:%d %s invalid bankw %d\n",
+			 __func__, __LINE__, prefix, surf->bankw);
+		return -EINVAL;
+	}
+	switch (surf->bankh) {
+	case 0: surf->bankh = 1; break;
+	case 1: surf->bankh = 2; break;
+	case 2: surf->bankh = 4; break;
+	case 3: surf->bankh = 8; break;
+	default:
+		dev_warn(p->dev, "%s:%d %s invalid bankh %d\n",
+			 __func__, __LINE__, prefix, surf->bankh);
+		return -EINVAL;
+	}
+	switch (surf->mtilea) {
+	case 0: surf->mtilea = 1; break;
+	case 1: surf->mtilea = 2; break;
+	case 2: surf->mtilea = 4; break;
+	case 3: surf->mtilea = 8; break;
+	default:
+		dev_warn(p->dev, "%s:%d %s invalid macro tile aspect %d\n",
+			 __func__, __LINE__, prefix, surf->mtilea);
+		return -EINVAL;
+	}
+	switch (surf->tsplit) {
+	case 0: surf->tsplit = 64; break;
+	case 1: surf->tsplit = 128; break;
+	case 2: surf->tsplit = 256; break;
+	case 3: surf->tsplit = 512; break;
+	case 4: surf->tsplit = 1024; break;
+	case 5: surf->tsplit = 2048; break;
+	case 6: surf->tsplit = 4096; break;
+	default:
+		dev_warn(p->dev, "%s:%d %s invalid tile split %d\n",
+			 __func__, __LINE__, prefix, surf->tsplit);
+		return -EINVAL;
+	}
+	return 0;
+}
+
+static int evergreen_cs_track_validate_cb(struct radeon_cs_parser *p, unsigned id)
 {
 	struct evergreen_cs_track *track = p->track;
+	struct eg_surface surf;
+	unsigned pitch, slice, mslice;
+	unsigned long offset;
+	int r;
 
-	/* we don't support stream out buffer yet */
-	if (track->vgt_strmout_config || track->vgt_strmout_buffer_config) {
-		dev_warn(p->dev, "this kernel doesn't support SMX output buffer\n");
+	mslice = G_028C6C_SLICE_MAX(track->cb_color_view[id]) + 1;
+	pitch = track->cb_color_pitch[id];
+	slice = track->cb_color_slice[id];
+	surf.nbx = (pitch + 1) * 8;
+	surf.nby = ((slice + 1) * 64) / surf.nbx;
+	surf.mode = G_028C70_ARRAY_MODE(track->cb_color_info[id]);
+	surf.format = G_028C70_FORMAT(track->cb_color_info[id]);
+	surf.tsplit = G_028C74_TILE_SPLIT(track->cb_color_attrib[id]);
+	surf.nbanks = G_028C74_NUM_BANKS(track->cb_color_attrib[id]);
+	surf.bankw = G_028C74_BANK_WIDTH(track->cb_color_attrib[id]);
+	surf.bankh = G_028C74_BANK_HEIGHT(track->cb_color_attrib[id]);
+	surf.mtilea = G_028C74_MACRO_TILE_ASPECT(track->cb_color_attrib[id]);
+	surf.nsamples = 1;
+
+	if (!r600_fmt_is_valid_color(surf.format)) {
+		dev_warn(p->dev, "%s:%d cb invalid format %d for %d (0x%08x)\n",
+			 __func__, __LINE__, surf.format,
+			id, track->cb_color_info[id]);
+		return -EINVAL;
+	}
+
+	r = evergreen_surface_value_conv_check(p, &surf, "cb");
+	if (r) {
+		return r;
+	}
+
+	r = evergreen_surface_check(p, &surf, "cb");
+	if (r) {
+		dev_warn(p->dev, "%s:%d cb[%d] invalid (0x%08x 0x%08x 0x%08x 0x%08x)\n",
+			 __func__, __LINE__, id, track->cb_color_pitch[id],
+			 track->cb_color_slice[id], track->cb_color_attrib[id],
+			 track->cb_color_info[id]);
+		return r;
+	}
+
+	offset = track->cb_color_bo_offset[id] << 8;
+	if (offset & (surf.base_align - 1)) {
+		dev_warn(p->dev, "%s:%d cb[%d] bo base %ld not aligned with %ld\n",
+			 __func__, __LINE__, id, offset, surf.base_align);
+		return -EINVAL;
+	}
+
+	offset += surf.layer_size * mslice;
+	if (offset > radeon_bo_size(track->cb_color_bo[id])) {
+		dev_warn(p->dev, "%s:%d cb[%d] bo too small (layer size %d, "
+			 "offset %d, max layer %d, bo size %ld, slice %d)\n",
+			 __func__, __LINE__, id, surf.layer_size,
+			track->cb_color_bo_offset[id] << 8, mslice,
+			radeon_bo_size(track->cb_color_bo[id]), slice);
+		dev_warn(p->dev, "%s:%d problematic surf: (%d %d) (%d %d %d %d %d %d %d)\n",
+			 __func__, __LINE__, surf.nbx, surf.nby,
+			surf.mode, surf.bpe, surf.nsamples,
+			surf.bankw, surf.bankh,
+			surf.tsplit, surf.mtilea);
 		return -EINVAL;
 	}
 
-	/* XXX fill in */
+	return 0;
+}
+
+static int evergreen_cs_track_validate_htile(struct radeon_cs_parser *p,
+						unsigned nbx, unsigned nby)
+{
+	struct evergreen_cs_track *track = p->track;
+	unsigned long size;
+
+	if (track->htile_bo == NULL) {
+		dev_warn(p->dev, "%s:%d htile enabled without htile surface 0x%08x\n",
+				__func__, __LINE__, track->db_z_info);
+		return -EINVAL;
+	}
+
+	if (G_028ABC_LINEAR(track->htile_surface)) {
+		/* pitch must be 16 htiles aligned == 16 * 8 pixel aligned */
+		nbx = round_up(nbx, 16 * 8);
+		/* height is npipes htiles aligned == npipes * 8 pixel aligned */
+		nby = round_up(nby, track->npipes * 8);
+	} else {
+		switch (track->npipes) {
+		case 8:
+			nbx = round_up(nbx, 64 * 8);
+			nby = round_up(nby, 64 * 8);
+			break;
+		case 4:
+			nbx = round_up(nbx, 64 * 8);
+			nby = round_up(nby, 32 * 8);
+			break;
+		case 2:
+			nbx = round_up(nbx, 32 * 8);
+			nby = round_up(nby, 32 * 8);
+			break;
+		case 1:
+			nbx = round_up(nbx, 32 * 8);
+			nby = round_up(nby, 16 * 8);
+			break;
+		default:
+			dev_warn(p->dev, "%s:%d invalid num pipes %d\n",
+					__func__, __LINE__, track->npipes);
+			return -EINVAL;
+		}
+	}
+	/* compute number of htile */
+	nbx = nbx / 8;
+	nby = nby / 8;
+	size = nbx * nby * 4;
+	size += track->htile_offset;
+
+	if (size > radeon_bo_size(track->htile_bo)) {
+		dev_warn(p->dev, "%s:%d htile surface too small %ld for %ld (%d %d)\n",
+				__func__, __LINE__, radeon_bo_size(track->htile_bo),
+				size, nbx, nby);
+		return -EINVAL;
+	}
+	return 0;
+}
+
+static int evergreen_cs_track_validate_stencil(struct radeon_cs_parser *p)
+{
+	struct evergreen_cs_track *track = p->track;
+	struct eg_surface surf;
+	unsigned pitch, slice, mslice;
+	unsigned long offset;
+	int r;
+
+	mslice = G_028008_SLICE_MAX(track->db_depth_view) + 1;
+	pitch = G_028058_PITCH_TILE_MAX(track->db_depth_size);
+	slice = track->db_depth_slice;
+	surf.nbx = (pitch + 1) * 8;
+	surf.nby = ((slice + 1) * 64) / surf.nbx;
+	surf.mode = G_028040_ARRAY_MODE(track->db_z_info);
+	surf.format = G_028044_FORMAT(track->db_s_info);
+	surf.tsplit = G_028044_TILE_SPLIT(track->db_s_info);
+	surf.nbanks = G_028040_NUM_BANKS(track->db_z_info);
+	surf.bankw = G_028040_BANK_WIDTH(track->db_z_info);
+	surf.bankh = G_028040_BANK_HEIGHT(track->db_z_info);
+	surf.mtilea = G_028040_MACRO_TILE_ASPECT(track->db_z_info);
+	surf.nsamples = 1;
+
+	if (surf.format != 1) {
+		dev_warn(p->dev, "%s:%d stencil invalid format %d\n",
+			 __func__, __LINE__, surf.format);
+		return -EINVAL;
+	}
+	/* replace by color format so we can use same code */
+	surf.format = V_028C70_COLOR_8;
+
+	r = evergreen_surface_value_conv_check(p, &surf, "stencil");
+	if (r) {
+		return r;
+	}
+
+	r = evergreen_surface_check(p, &surf, NULL);
+	if (r) {
+		/* old userspace doesn't compute proper depth/stencil alignment
+		 * check that alignment against a bigger byte per elements and
+		 * only report if that alignment is wrong too.
+		 */
+		surf.format = V_028C70_COLOR_8_8_8_8;
+		r = evergreen_surface_check(p, &surf, "stencil");
+		if (r) {
+			dev_warn(p->dev, "%s:%d stencil invalid (0x%08x 0x%08x 0x%08x 0x%08x)\n",
+				 __func__, __LINE__, track->db_depth_size,
+				 track->db_depth_slice, track->db_s_info, track->db_z_info);
+		}
+		return r;
+	}
+
+	offset = track->db_s_read_offset << 8;
+	if (offset & (surf.base_align - 1)) {
+		dev_warn(p->dev, "%s:%d stencil read bo base %ld not aligned with %ld\n",
+			 __func__, __LINE__, offset, surf.base_align);
+		return -EINVAL;
+	}
+	offset += surf.layer_size * mslice;
+	if (offset > radeon_bo_size(track->db_s_read_bo)) {
+		dev_warn(p->dev, "%s:%d stencil read bo too small (layer size %d, "
+			 "offset %ld, max layer %d, bo size %ld)\n",
+			 __func__, __LINE__, surf.layer_size,
+			(unsigned long)track->db_s_read_offset << 8, mslice,
+			radeon_bo_size(track->db_s_read_bo));
+		dev_warn(p->dev, "%s:%d stencil invalid (0x%08x 0x%08x 0x%08x 0x%08x)\n",
+			 __func__, __LINE__, track->db_depth_size,
+			 track->db_depth_slice, track->db_s_info, track->db_z_info);
+		return -EINVAL;
+	}
+
+	offset = track->db_s_write_offset << 8;
+	if (offset & (surf.base_align - 1)) {
+		dev_warn(p->dev, "%s:%d stencil write bo base %ld not aligned with %ld\n",
+			 __func__, __LINE__, offset, surf.base_align);
+		return -EINVAL;
+	}
+	offset += surf.layer_size * mslice;
+	if (offset > radeon_bo_size(track->db_s_write_bo)) {
+		dev_warn(p->dev, "%s:%d stencil write bo too small (layer size %d, "
+			 "offset %ld, max layer %d, bo size %ld)\n",
+			 __func__, __LINE__, surf.layer_size,
+			(unsigned long)track->db_s_write_offset << 8, mslice,
+			radeon_bo_size(track->db_s_write_bo));
+		return -EINVAL;
+	}
+
+	/* hyperz */
+	if (G_028040_TILE_SURFACE_ENABLE(track->db_z_info)) {
+		r = evergreen_cs_track_validate_htile(p, surf.nbx, surf.nby);
+		if (r) {
+			return r;
+		}
+	}
+
+	return 0;
+}
+
+static int evergreen_cs_track_validate_depth(struct radeon_cs_parser *p)
+{
+	struct evergreen_cs_track *track = p->track;
+	struct eg_surface surf;
+	unsigned pitch, slice, mslice;
+	unsigned long offset;
+	int r;
+
+	mslice = G_028008_SLICE_MAX(track->db_depth_view) + 1;
+	pitch = G_028058_PITCH_TILE_MAX(track->db_depth_size);
+	slice = track->db_depth_slice;
+	surf.nbx = (pitch + 1) * 8;
+	surf.nby = ((slice + 1) * 64) / surf.nbx;
+	surf.mode = G_028040_ARRAY_MODE(track->db_z_info);
+	surf.format = G_028040_FORMAT(track->db_z_info);
+	surf.tsplit = G_028040_TILE_SPLIT(track->db_z_info);
+	surf.nbanks = G_028040_NUM_BANKS(track->db_z_info);
+	surf.bankw = G_028040_BANK_WIDTH(track->db_z_info);
+	surf.bankh = G_028040_BANK_HEIGHT(track->db_z_info);
+	surf.mtilea = G_028040_MACRO_TILE_ASPECT(track->db_z_info);
+	surf.nsamples = 1;
+
+	switch (surf.format) {
+	case V_028040_Z_16:
+		surf.format = V_028C70_COLOR_16;
+		break;
+	case V_028040_Z_24:
+	case V_028040_Z_32_FLOAT:
+		surf.format = V_028C70_COLOR_8_8_8_8;
+		break;
+	default:
+		dev_warn(p->dev, "%s:%d depth invalid format %d\n",
+			 __func__, __LINE__, surf.format);
+		return -EINVAL;
+	}
+
+	r = evergreen_surface_value_conv_check(p, &surf, "depth");
+	if (r) {
+		dev_warn(p->dev, "%s:%d depth invalid (0x%08x 0x%08x 0x%08x)\n",
+			 __func__, __LINE__, track->db_depth_size,
+			 track->db_depth_slice, track->db_z_info);
+		return r;
+	}
+
+	r = evergreen_surface_check(p, &surf, "depth");
+	if (r) {
+		dev_warn(p->dev, "%s:%d depth invalid (0x%08x 0x%08x 0x%08x)\n",
+			 __func__, __LINE__, track->db_depth_size,
+			 track->db_depth_slice, track->db_z_info);
+		return r;
+	}
+
+	offset = track->db_z_read_offset << 8;
+	if (offset & (surf.base_align - 1)) {
+		dev_warn(p->dev, "%s:%d stencil read bo base %ld not aligned with %ld\n",
+			 __func__, __LINE__, offset, surf.base_align);
+		return -EINVAL;
+	}
+	offset += surf.layer_size * mslice;
+	if (offset > radeon_bo_size(track->db_z_read_bo)) {
+		dev_warn(p->dev, "%s:%d depth read bo too small (layer size %d, "
+			 "offset %ld, max layer %d, bo size %ld)\n",
+			 __func__, __LINE__, surf.layer_size,
+			(unsigned long)track->db_z_read_offset << 8, mslice,
+			radeon_bo_size(track->db_z_read_bo));
+		return -EINVAL;
+	}
+
+	offset = track->db_z_write_offset << 8;
+	if (offset & (surf.base_align - 1)) {
+		dev_warn(p->dev, "%s:%d stencil write bo base %ld not aligned with %ld\n",
+			 __func__, __LINE__, offset, surf.base_align);
+		return -EINVAL;
+	}
+	offset += surf.layer_size * mslice;
+	if (offset > radeon_bo_size(track->db_z_write_bo)) {
+		dev_warn(p->dev, "%s:%d depth write bo too small (layer size %d, "
+			 "offset %ld, max layer %d, bo size %ld)\n",
+			 __func__, __LINE__, surf.layer_size,
+			(unsigned long)track->db_z_write_offset << 8, mslice,
+			radeon_bo_size(track->db_z_write_bo));
+		return -EINVAL;
+	}
+
+	/* hyperz */
+	if (G_028040_TILE_SURFACE_ENABLE(track->db_z_info)) {
+		r = evergreen_cs_track_validate_htile(p, surf.nbx, surf.nby);
+		if (r) {
+			return r;
+		}
+	}
+
+	return 0;
+}
+
+static int evergreen_cs_track_validate_texture(struct radeon_cs_parser *p,
+					       struct radeon_bo *texture,
+					       struct radeon_bo *mipmap,
+					       unsigned idx)
+{
+	struct eg_surface surf;
+	unsigned long toffset, moffset;
+	unsigned dim, llevel, mslice, width, height, depth, i;
+	u32 texdw[8];
+	int r;
+
+	texdw[0] = radeon_get_ib_value(p, idx + 0);
+	texdw[1] = radeon_get_ib_value(p, idx + 1);
+	texdw[2] = radeon_get_ib_value(p, idx + 2);
+	texdw[3] = radeon_get_ib_value(p, idx + 3);
+	texdw[4] = radeon_get_ib_value(p, idx + 4);
+	texdw[5] = radeon_get_ib_value(p, idx + 5);
+	texdw[6] = radeon_get_ib_value(p, idx + 6);
+	texdw[7] = radeon_get_ib_value(p, idx + 7);
+	dim = G_030000_DIM(texdw[0]);
+	llevel = G_030014_LAST_LEVEL(texdw[5]);
+	mslice = G_030014_LAST_ARRAY(texdw[5]) + 1;
+	width = G_030000_TEX_WIDTH(texdw[0]) + 1;
+	height =  G_030004_TEX_HEIGHT(texdw[1]) + 1;
+	depth = G_030004_TEX_DEPTH(texdw[1]) + 1;
+	surf.format = G_03001C_DATA_FORMAT(texdw[7]);
+	surf.nbx = (G_030000_PITCH(texdw[0]) + 1) * 8;
+	surf.nbx = r600_fmt_get_nblocksx(surf.format, surf.nbx);
+	surf.nby = r600_fmt_get_nblocksy(surf.format, height);
+	surf.mode = G_030004_ARRAY_MODE(texdw[1]);
+	surf.tsplit = G_030018_TILE_SPLIT(texdw[6]);
+	surf.nbanks = G_03001C_NUM_BANKS(texdw[7]);
+	surf.bankw = G_03001C_BANK_WIDTH(texdw[7]);
+	surf.bankh = G_03001C_BANK_HEIGHT(texdw[7]);
+	surf.mtilea = G_03001C_MACRO_TILE_ASPECT(texdw[7]);
+	surf.nsamples = 1;
+	toffset = texdw[2] << 8;
+	moffset = texdw[3] << 8;
+
+	if (!r600_fmt_is_valid_texture(surf.format, p->family)) {
+		dev_warn(p->dev, "%s:%d texture invalid format %d\n",
+			 __func__, __LINE__, surf.format);
+		return -EINVAL;
+	}
+	switch (dim) {
+	case V_030000_SQ_TEX_DIM_1D:
+	case V_030000_SQ_TEX_DIM_2D:
+	case V_030000_SQ_TEX_DIM_CUBEMAP:
+	case V_030000_SQ_TEX_DIM_1D_ARRAY:
+	case V_030000_SQ_TEX_DIM_2D_ARRAY:
+		depth = 1;
+	case V_030000_SQ_TEX_DIM_3D:
+		break;
+	default:
+		dev_warn(p->dev, "%s:%d texture invalid dimension %d\n",
+			 __func__, __LINE__, dim);
+		return -EINVAL;
+	}
+
+	r = evergreen_surface_value_conv_check(p, &surf, "texture");
+	if (r) {
+		return r;
+	}
+
+	/* align height */
+	evergreen_surface_check(p, &surf, NULL);
+	surf.nby = ALIGN(surf.nby, surf.halign);
+
+	r = evergreen_surface_check(p, &surf, "texture");
+	if (r) {
+		dev_warn(p->dev, "%s:%d texture invalid 0x%08x 0x%08x 0x%08x 0x%08x 0x%08x 0x%08x\n",
+			 __func__, __LINE__, texdw[0], texdw[1], texdw[4],
+			 texdw[5], texdw[6], texdw[7]);
+		return r;
+	}
+
+	/* check texture size */
+	if (toffset & (surf.base_align - 1)) {
+		dev_warn(p->dev, "%s:%d texture bo base %ld not aligned with %ld\n",
+			 __func__, __LINE__, toffset, surf.base_align);
+		return -EINVAL;
+	}
+	if (moffset & (surf.base_align - 1)) {
+		dev_warn(p->dev, "%s:%d mipmap bo base %ld not aligned with %ld\n",
+			 __func__, __LINE__, moffset, surf.base_align);
+		return -EINVAL;
+	}
+	if (dim == SQ_TEX_DIM_3D) {
+		toffset += surf.layer_size * depth;
+	} else {
+		toffset += surf.layer_size * mslice;
+	}
+	if (toffset > radeon_bo_size(texture)) {
+		dev_warn(p->dev, "%s:%d texture bo too small (layer size %d, "
+			 "offset %ld, max layer %d, depth %d, bo size %ld) (%d %d)\n",
+			 __func__, __LINE__, surf.layer_size,
+			(unsigned long)texdw[2] << 8, mslice,
+			depth, radeon_bo_size(texture),
+			surf.nbx, surf.nby);
+		return -EINVAL;
+	}
+
+	/* check mipmap size */
+	for (i = 1; i <= llevel; i++) {
+		unsigned w, h, d;
+
+		w = r600_mip_minify(width, i);
+		h = r600_mip_minify(height, i);
+		d = r600_mip_minify(depth, i);
+		surf.nbx = r600_fmt_get_nblocksx(surf.format, w);
+		surf.nby = r600_fmt_get_nblocksy(surf.format, h);
+
+		switch (surf.mode) {
+		case ARRAY_2D_TILED_THIN1:
+			if (surf.nbx < surf.palign || surf.nby < surf.halign) {
+				surf.mode = ARRAY_1D_TILED_THIN1;
+			}
+			/* recompute alignment */
+			evergreen_surface_check(p, &surf, NULL);
+			break;
+		case ARRAY_LINEAR_GENERAL:
+		case ARRAY_LINEAR_ALIGNED:
+		case ARRAY_1D_TILED_THIN1:
+			break;
+		default:
+			dev_warn(p->dev, "%s:%d invalid array mode %d\n",
+				 __func__, __LINE__, surf.mode);
+			return -EINVAL;
+		}
+		surf.nbx = ALIGN(surf.nbx, surf.palign);
+		surf.nby = ALIGN(surf.nby, surf.halign);
+
+		r = evergreen_surface_check(p, &surf, "mipmap");
+		if (r) {
+			return r;
+		}
+
+		if (dim == SQ_TEX_DIM_3D) {
+			moffset += surf.layer_size * d;
+		} else {
+			moffset += surf.layer_size * mslice;
+		}
+		if (moffset > radeon_bo_size(mipmap)) {
+			dev_warn(p->dev, "%s:%d mipmap [%d] bo too small (layer size %d, "
+					"offset %ld, coffset %ld, max layer %d, depth %d, "
+					"bo size %ld) level0 (%d %d %d)\n",
+					__func__, __LINE__, i, surf.layer_size,
+					(unsigned long)texdw[3] << 8, moffset, mslice,
+					d, radeon_bo_size(mipmap),
+					width, height, depth);
+			dev_warn(p->dev, "%s:%d problematic surf: (%d %d) (%d %d %d %d %d %d %d)\n",
+				 __func__, __LINE__, surf.nbx, surf.nby,
+				surf.mode, surf.bpe, surf.nsamples,
+				surf.bankw, surf.bankh,
+				surf.tsplit, surf.mtilea);
+			return -EINVAL;
+		}
+	}
+
+	return 0;
+}
+
+static int evergreen_cs_track_check(struct radeon_cs_parser *p)
+{
+	struct evergreen_cs_track *track = p->track;
+	unsigned tmp, i;
+	int r;
+	unsigned buffer_mask = 0;
+
+	/* check streamout */
+	if (track->streamout_dirty && track->vgt_strmout_config) {
+		for (i = 0; i < 4; i++) {
+			if (track->vgt_strmout_config & (1 << i)) {
+				buffer_mask |= (track->vgt_strmout_buffer_config >> (i * 4)) & 0xf;
+			}
+		}
+
+		for (i = 0; i < 4; i++) {
+			if (buffer_mask & (1 << i)) {
+				if (track->vgt_strmout_bo[i]) {
+					u64 offset = (u64)track->vgt_strmout_bo_offset[i] +
+							(u64)track->vgt_strmout_size[i];
+					if (offset > radeon_bo_size(track->vgt_strmout_bo[i])) {
+						DRM_ERROR("streamout %d bo too small: 0x%llx, 0x%lx\n",
+							  i, offset,
+							  radeon_bo_size(track->vgt_strmout_bo[i]));
+						return -EINVAL;
+					}
+				} else {
+					dev_warn(p->dev, "No buffer for streamout %d\n", i);
+					return -EINVAL;
+				}
+			}
+		}
+		track->streamout_dirty = false;
+	}
+
+	if (track->sx_misc_kill_all_prims)
+		return 0;
+
+	/* check that we have a cb for each enabled target
+	 */
+	if (track->cb_dirty) {
+		tmp = track->cb_target_mask;
+		for (i = 0; i < 8; i++) {
+			if ((tmp >> (i * 4)) & 0xF) {
+				/* at least one component is enabled */
+				if (track->cb_color_bo[i] == NULL) {
+					dev_warn(p->dev, "%s:%d mask 0x%08X | 0x%08X no cb for %d\n",
+						__func__, __LINE__, track->cb_target_mask, track->cb_shader_mask, i);
+					return -EINVAL;
+				}
+				/* check cb */
+				r = evergreen_cs_track_validate_cb(p, i);
+				if (r) {
+					return r;
+				}
+			}
+		}
+		track->cb_dirty = false;
+	}
+
+	if (track->db_dirty) {
+		/* Check stencil buffer */
+		if (G_028800_STENCIL_ENABLE(track->db_depth_control)) {
+			r = evergreen_cs_track_validate_stencil(p);
+			if (r)
+				return r;
+		}
+		/* Check depth buffer */
+		if (G_028800_Z_ENABLE(track->db_depth_control)) {
+			r = evergreen_cs_track_validate_depth(p);
+			if (r)
+				return r;
+		}
+		track->db_dirty = false;
+	}
+
 	return 0;
 }
 
@@ -503,6 +1267,7 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
 		break;
 	case DB_DEPTH_CONTROL:
 		track->db_depth_control = radeon_get_ib_value(p, idx);
+		track->db_dirty = true;
 		break;
 	case CAYMAN_DB_EQAA:
 		if (p->rdev->family < CHIP_CAYMAN) {
@@ -532,20 +1297,35 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
 			ib[idx] |= Z_ARRAY_MODE(evergreen_cs_get_aray_mode(reloc->lobj.tiling_flags));
 			track->db_z_info |= Z_ARRAY_MODE(evergreen_cs_get_aray_mode(reloc->lobj.tiling_flags));
 			if (reloc->lobj.tiling_flags & RADEON_TILING_MACRO) {
+				unsigned bankw, bankh, mtaspect, tile_split;
+
+				evergreen_tiling_fields(reloc->lobj.tiling_flags,
+							&bankw, &bankh, &mtaspect,
+							&tile_split);
 				ib[idx] |= DB_NUM_BANKS(evergreen_cs_get_num_banks(track->nbanks));
-				ib[idx] |= DB_TILE_SPLIT(evergreen_cs_get_tile_split(track->row_size));
+				ib[idx] |= DB_TILE_SPLIT(tile_split) |
+						DB_BANK_WIDTH(bankw) |
+						DB_BANK_HEIGHT(bankh) |
+						DB_MACRO_TILE_ASPECT(mtaspect);
 			}
 		}
+		track->db_dirty = true;
 		break;
 	case DB_STENCIL_INFO:
 		track->db_s_info = radeon_get_ib_value(p, idx);
+		track->db_dirty = true;
 		break;
 	case DB_DEPTH_VIEW:
 		track->db_depth_view = radeon_get_ib_value(p, idx);
+		track->db_dirty = true;
 		break;
 	case DB_DEPTH_SIZE:
 		track->db_depth_size = radeon_get_ib_value(p, idx);
-		track->db_depth_size_idx = idx;
+		track->db_dirty = true;
+		break;
+	case R_02805C_DB_DEPTH_SLICE:
+		track->db_depth_slice = radeon_get_ib_value(p, idx);
+		track->db_dirty = true;
 		break;
 	case DB_Z_READ_BASE:
 		r = evergreen_cs_packet_next_reloc(p, &reloc);
@@ -557,6 +1337,7 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
 		track->db_z_read_offset = radeon_get_ib_value(p, idx);
 		ib[idx] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff);
 		track->db_z_read_bo = reloc->robj;
+		track->db_dirty = true;
 		break;
 	case DB_Z_WRITE_BASE:
 		r = evergreen_cs_packet_next_reloc(p, &reloc);
@@ -568,6 +1349,7 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
 		track->db_z_write_offset = radeon_get_ib_value(p, idx);
 		ib[idx] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff);
 		track->db_z_write_bo = reloc->robj;
+		track->db_dirty = true;
 		break;
 	case DB_STENCIL_READ_BASE:
 		r = evergreen_cs_packet_next_reloc(p, &reloc);
@@ -579,6 +1361,7 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
 		track->db_s_read_offset = radeon_get_ib_value(p, idx);
 		ib[idx] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff);
 		track->db_s_read_bo = reloc->robj;
+		track->db_dirty = true;
 		break;
 	case DB_STENCIL_WRITE_BASE:
 		r = evergreen_cs_packet_next_reloc(p, &reloc);
@@ -590,18 +1373,56 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
 		track->db_s_write_offset = radeon_get_ib_value(p, idx);
 		ib[idx] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff);
 		track->db_s_write_bo = reloc->robj;
+		track->db_dirty = true;
 		break;
 	case VGT_STRMOUT_CONFIG:
 		track->vgt_strmout_config = radeon_get_ib_value(p, idx);
+		track->streamout_dirty = true;
 		break;
 	case VGT_STRMOUT_BUFFER_CONFIG:
 		track->vgt_strmout_buffer_config = radeon_get_ib_value(p, idx);
+		track->streamout_dirty = true;
 		break;
+	case VGT_STRMOUT_BUFFER_BASE_0:
+	case VGT_STRMOUT_BUFFER_BASE_1:
+	case VGT_STRMOUT_BUFFER_BASE_2:
+	case VGT_STRMOUT_BUFFER_BASE_3:
+		r = evergreen_cs_packet_next_reloc(p, &reloc);
+		if (r) {
+			dev_warn(p->dev, "bad SET_CONTEXT_REG "
+					"0x%04X\n", reg);
+			return -EINVAL;
+		}
+		tmp = (reg - VGT_STRMOUT_BUFFER_BASE_0) / 16;
+		track->vgt_strmout_bo_offset[tmp] = radeon_get_ib_value(p, idx) << 8;
+		ib[idx] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff);
+		track->vgt_strmout_bo[tmp] = reloc->robj;
+		track->streamout_dirty = true;
+		break;
+	case VGT_STRMOUT_BUFFER_SIZE_0:
+	case VGT_STRMOUT_BUFFER_SIZE_1:
+	case VGT_STRMOUT_BUFFER_SIZE_2:
+	case VGT_STRMOUT_BUFFER_SIZE_3:
+		tmp = (reg - VGT_STRMOUT_BUFFER_SIZE_0) / 16;
+		/* size in register is DWs, convert to bytes */
+		track->vgt_strmout_size[tmp] = radeon_get_ib_value(p, idx) * 4;
+		track->streamout_dirty = true;
+		break;
+	case CP_COHER_BASE:
+		r = evergreen_cs_packet_next_reloc(p, &reloc);
+		if (r) {
+			dev_warn(p->dev, "missing reloc for CP_COHER_BASE "
+					"0x%04X\n", reg);
+			return -EINVAL;
+		}
+		ib[idx] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff);
 	case CB_TARGET_MASK:
 		track->cb_target_mask = radeon_get_ib_value(p, idx);
+		track->cb_dirty = true;
 		break;
 	case CB_SHADER_MASK:
 		track->cb_shader_mask = radeon_get_ib_value(p, idx);
+		track->cb_dirty = true;
 		break;
 	case PA_SC_AA_CONFIG:
 		if (p->rdev->family >= CHIP_CAYMAN) {
@@ -631,6 +1452,7 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
 	case CB_COLOR7_VIEW:
 		tmp = (reg - CB_COLOR0_VIEW) / 0x3c;
 		track->cb_color_view[tmp] = radeon_get_ib_value(p, idx);
+		track->cb_dirty = true;
 		break;
 	case CB_COLOR8_VIEW:
 	case CB_COLOR9_VIEW:
@@ -638,6 +1460,7 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
 	case CB_COLOR11_VIEW:
 		tmp = ((reg - CB_COLOR8_VIEW) / 0x1c) + 8;
 		track->cb_color_view[tmp] = radeon_get_ib_value(p, idx);
+		track->cb_dirty = true;
 		break;
 	case CB_COLOR0_INFO:
 	case CB_COLOR1_INFO:
@@ -659,6 +1482,7 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
 			ib[idx] |= CB_ARRAY_MODE(evergreen_cs_get_aray_mode(reloc->lobj.tiling_flags));
 			track->cb_color_info[tmp] |= CB_ARRAY_MODE(evergreen_cs_get_aray_mode(reloc->lobj.tiling_flags));
 		}
+		track->cb_dirty = true;
 		break;
 	case CB_COLOR8_INFO:
 	case CB_COLOR9_INFO:
@@ -676,6 +1500,7 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
 			ib[idx] |= CB_ARRAY_MODE(evergreen_cs_get_aray_mode(reloc->lobj.tiling_flags));
 			track->cb_color_info[tmp] |= CB_ARRAY_MODE(evergreen_cs_get_aray_mode(reloc->lobj.tiling_flags));
 		}
+		track->cb_dirty = true;
 		break;
 	case CB_COLOR0_PITCH:
 	case CB_COLOR1_PITCH:
@@ -687,7 +1512,7 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
 	case CB_COLOR7_PITCH:
 		tmp = (reg - CB_COLOR0_PITCH) / 0x3c;
 		track->cb_color_pitch[tmp] = radeon_get_ib_value(p, idx);
-		track->cb_color_pitch_idx[tmp] = idx;
+		track->cb_dirty = true;
 		break;
 	case CB_COLOR8_PITCH:
 	case CB_COLOR9_PITCH:
@@ -695,7 +1520,7 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
 	case CB_COLOR11_PITCH:
 		tmp = ((reg - CB_COLOR8_PITCH) / 0x1c) + 8;
 		track->cb_color_pitch[tmp] = radeon_get_ib_value(p, idx);
-		track->cb_color_pitch_idx[tmp] = idx;
+		track->cb_dirty = true;
 		break;
 	case CB_COLOR0_SLICE:
 	case CB_COLOR1_SLICE:
@@ -707,7 +1532,7 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
 	case CB_COLOR7_SLICE:
 		tmp = (reg - CB_COLOR0_SLICE) / 0x3c;
 		track->cb_color_slice[tmp] = radeon_get_ib_value(p, idx);
-		track->cb_color_slice_idx[tmp] = idx;
+		track->cb_dirty = true;
 		break;
 	case CB_COLOR8_SLICE:
 	case CB_COLOR9_SLICE:
@@ -715,7 +1540,7 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
 	case CB_COLOR11_SLICE:
 		tmp = ((reg - CB_COLOR8_SLICE) / 0x1c) + 8;
 		track->cb_color_slice[tmp] = radeon_get_ib_value(p, idx);
-		track->cb_color_slice_idx[tmp] = idx;
+		track->cb_dirty = true;
 		break;
 	case CB_COLOR0_ATTRIB:
 	case CB_COLOR1_ATTRIB:
@@ -725,6 +1550,30 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
 	case CB_COLOR5_ATTRIB:
 	case CB_COLOR6_ATTRIB:
 	case CB_COLOR7_ATTRIB:
+		r = evergreen_cs_packet_next_reloc(p, &reloc);
+		if (r) {
+			dev_warn(p->dev, "bad SET_CONTEXT_REG "
+					"0x%04X\n", reg);
+			return -EINVAL;
+		}
+		if (!(p->cs_flags & RADEON_CS_KEEP_TILING_FLAGS)) {
+			if (reloc->lobj.tiling_flags & RADEON_TILING_MACRO) {
+				unsigned bankw, bankh, mtaspect, tile_split;
+
+				evergreen_tiling_fields(reloc->lobj.tiling_flags,
+							&bankw, &bankh, &mtaspect,
+							&tile_split);
+				ib[idx] |= CB_NUM_BANKS(evergreen_cs_get_num_banks(track->nbanks));
+				ib[idx] |= CB_TILE_SPLIT(tile_split) |
+					   CB_BANK_WIDTH(bankw) |
+					   CB_BANK_HEIGHT(bankh) |
+					   CB_MACRO_TILE_ASPECT(mtaspect);
+			}
+		}
+		tmp = ((reg - CB_COLOR0_ATTRIB) / 0x3c);
+		track->cb_color_attrib[tmp] = ib[idx];
+		track->cb_dirty = true;
+		break;
 	case CB_COLOR8_ATTRIB:
 	case CB_COLOR9_ATTRIB:
 	case CB_COLOR10_ATTRIB:
@@ -735,30 +1584,23 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
 					"0x%04X\n", reg);
 			return -EINVAL;
 		}
-		if (reloc->lobj.tiling_flags & RADEON_TILING_MACRO) {
-			ib[idx] |= CB_NUM_BANKS(evergreen_cs_get_num_banks(track->nbanks));
-			ib[idx] |= CB_TILE_SPLIT(evergreen_cs_get_tile_split(track->row_size));
+		if (!(p->cs_flags & RADEON_CS_KEEP_TILING_FLAGS)) {
+			if (reloc->lobj.tiling_flags & RADEON_TILING_MACRO) {
+				unsigned bankw, bankh, mtaspect, tile_split;
+
+				evergreen_tiling_fields(reloc->lobj.tiling_flags,
+							&bankw, &bankh, &mtaspect,
+							&tile_split);
+				ib[idx] |= CB_NUM_BANKS(evergreen_cs_get_num_banks(track->nbanks));
+				ib[idx] |= CB_TILE_SPLIT(tile_split) |
+					   CB_BANK_WIDTH(bankw) |
+					   CB_BANK_HEIGHT(bankh) |
+					   CB_MACRO_TILE_ASPECT(mtaspect);
+			}
 		}
-		break;
-	case CB_COLOR0_DIM:
-	case CB_COLOR1_DIM:
-	case CB_COLOR2_DIM:
-	case CB_COLOR3_DIM:
-	case CB_COLOR4_DIM:
-	case CB_COLOR5_DIM:
-	case CB_COLOR6_DIM:
-	case CB_COLOR7_DIM:
-		tmp = (reg - CB_COLOR0_DIM) / 0x3c;
-		track->cb_color_dim[tmp] = radeon_get_ib_value(p, idx);
-		track->cb_color_dim_idx[tmp] = idx;
-		break;
-	case CB_COLOR8_DIM:
-	case CB_COLOR9_DIM:
-	case CB_COLOR10_DIM:
-	case CB_COLOR11_DIM:
-		tmp = ((reg - CB_COLOR8_DIM) / 0x1c) + 8;
-		track->cb_color_dim[tmp] = radeon_get_ib_value(p, idx);
-		track->cb_color_dim_idx[tmp] = idx;
+		tmp = ((reg - CB_COLOR8_ATTRIB) / 0x1c) + 8;
+		track->cb_color_attrib[tmp] = ib[idx];
+		track->cb_dirty = true;
 		break;
 	case CB_COLOR0_FMASK:
 	case CB_COLOR1_FMASK:
@@ -833,8 +1675,8 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
 		tmp = (reg - CB_COLOR0_BASE) / 0x3c;
 		track->cb_color_bo_offset[tmp] = radeon_get_ib_value(p, idx);
 		ib[idx] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff);
-		track->cb_color_base_last[tmp] = ib[idx];
 		track->cb_color_bo[tmp] = reloc->robj;
+		track->cb_dirty = true;
 		break;
 	case CB_COLOR8_BASE:
 	case CB_COLOR9_BASE:
@@ -849,8 +1691,25 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
 		tmp = ((reg - CB_COLOR8_BASE) / 0x1c) + 8;
 		track->cb_color_bo_offset[tmp] = radeon_get_ib_value(p, idx);
 		ib[idx] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff);
-		track->cb_color_base_last[tmp] = ib[idx];
 		track->cb_color_bo[tmp] = reloc->robj;
+		track->cb_dirty = true;
+		break;
+	case DB_HTILE_DATA_BASE:
+		r = evergreen_cs_packet_next_reloc(p, &reloc);
+		if (r) {
+			dev_warn(p->dev, "bad SET_CONTEXT_REG "
+					"0x%04X\n", reg);
+			return -EINVAL;
+		}
+		track->htile_offset = radeon_get_ib_value(p, idx);
+		ib[idx] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff);
+		track->htile_bo = reloc->robj;
+		track->db_dirty = true;
+		break;
+	case DB_HTILE_SURFACE:
+		/* 8x8 only */
+		track->htile_surface = radeon_get_ib_value(p, idx);
+		track->db_dirty = true;
 		break;
 	case CB_IMMED0_BASE:
 	case CB_IMMED1_BASE:
@@ -864,7 +1723,6 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
 	case CB_IMMED9_BASE:
 	case CB_IMMED10_BASE:
 	case CB_IMMED11_BASE:
-	case DB_HTILE_DATA_BASE:
 	case SQ_PGM_START_FS:
 	case SQ_PGM_START_ES:
 	case SQ_PGM_START_VS:
@@ -989,6 +1847,9 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
 		}
 		ib[idx] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff);
 		break;
+	case SX_MISC:
+		track->sx_misc_kill_all_prims = (radeon_get_ib_value(p, idx) & 0x1) != 0;
+		break;
 	default:
 		dev_warn(p->dev, "forbidden register 0x%08x at %d\n", reg, idx);
 		return -EINVAL;
@@ -996,22 +1857,30 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
 	return 0;
 }
 
-/**
- * evergreen_check_texture_resource() - check if register is authorized or not
- * @p: parser structure holding parsing context
- * @idx: index into the cs buffer
- * @texture: texture's bo structure
- * @mipmap: mipmap's bo structure
- *
- * This function will check that the resource has valid field and that
- * the texture and mipmap bo object are big enough to cover this resource.
- */
-static int evergreen_check_texture_resource(struct radeon_cs_parser *p,  u32 idx,
-						   struct radeon_bo *texture,
-						   struct radeon_bo *mipmap)
+static bool evergreen_is_safe_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
 {
-	/* XXX fill in */
-	return 0;
+	u32 last_reg, m, i;
+
+	if (p->rdev->family >= CHIP_CAYMAN)
+		last_reg = ARRAY_SIZE(cayman_reg_safe_bm);
+	else
+		last_reg = ARRAY_SIZE(evergreen_reg_safe_bm);
+
+	i = (reg >> 7);
+	if (i >= last_reg) {
+		dev_warn(p->dev, "forbidden register 0x%08x at %d\n", reg, idx);
+		return false;
+	}
+	m = 1 << ((reg >> 2) & 31);
+	if (p->rdev->family >= CHIP_CAYMAN) {
+		if (!(cayman_reg_safe_bm[i] & m))
+			return true;
+	} else {
+		if (!(evergreen_reg_safe_bm[i] & m))
+			return true;
+	}
+	dev_warn(p->dev, "forbidden register 0x%08x at %d\n", reg, idx);
+	return false;
 }
 
 static int evergreen_packet3_check(struct radeon_cs_parser *p,
@@ -1036,6 +1905,8 @@ static int evergreen_packet3_check(struct radeon_cs_parser *p,
 	{
 		int pred_op;
 		int tmp;
+		uint64_t offset;
+
 		if (pkt->count != 1) {
 			DRM_ERROR("bad SET PREDICATION\n");
 			return -EINVAL;
@@ -1059,8 +1930,12 @@ static int evergreen_packet3_check(struct radeon_cs_parser *p,
 			return -EINVAL;
 		}
 
-		ib[idx + 0] = idx_value + (u32)(reloc->lobj.gpu_offset & 0xffffffff);
-		ib[idx + 1] = tmp + (upper_32_bits(reloc->lobj.gpu_offset) & 0xff);
+		offset = reloc->lobj.gpu_offset +
+		         (idx_value & 0xfffffff0) +
+		         ((u64)(tmp & 0xff) << 32);
+
+		ib[idx + 0] = offset;
+		ib[idx + 1] = (tmp & 0xffffff00) | (upper_32_bits(offset) & 0xff);
 	}
 	break;
 	case PACKET3_CONTEXT_CONTROL:
@@ -1088,6 +1963,9 @@ static int evergreen_packet3_check(struct radeon_cs_parser *p,
 		}
 		break;
 	case PACKET3_INDEX_BASE:
+	{
+		uint64_t offset;
+
 		if (pkt->count != 1) {
 			DRM_ERROR("bad INDEX_BASE\n");
 			return -EINVAL;
@@ -1097,15 +1975,24 @@ static int evergreen_packet3_check(struct radeon_cs_parser *p,
 			DRM_ERROR("bad INDEX_BASE\n");
 			return -EINVAL;
 		}
-		ib[idx+0] = idx_value + (u32)(reloc->lobj.gpu_offset & 0xffffffff);
-		ib[idx+1] += upper_32_bits(reloc->lobj.gpu_offset) & 0xff;
+
+		offset = reloc->lobj.gpu_offset +
+		         idx_value +
+		         ((u64)(radeon_get_ib_value(p, idx+1) & 0xff) << 32);
+
+		ib[idx+0] = offset;
+		ib[idx+1] = upper_32_bits(offset) & 0xff;
+
 		r = evergreen_cs_track_check(p);
 		if (r) {
 			dev_warn(p->dev, "%s:%d invalid cmd stream\n", __func__, __LINE__);
 			return r;
 		}
 		break;
+	}
 	case PACKET3_DRAW_INDEX:
+	{
+		uint64_t offset;
 		if (pkt->count != 3) {
 			DRM_ERROR("bad DRAW_INDEX\n");
 			return -EINVAL;
@@ -1115,15 +2002,25 @@ static int evergreen_packet3_check(struct radeon_cs_parser *p,
 			DRM_ERROR("bad DRAW_INDEX\n");
 			return -EINVAL;
 		}
-		ib[idx+0] = idx_value + (u32)(reloc->lobj.gpu_offset & 0xffffffff);
-		ib[idx+1] += upper_32_bits(reloc->lobj.gpu_offset) & 0xff;
+
+		offset = reloc->lobj.gpu_offset +
+		         idx_value +
+		         ((u64)(radeon_get_ib_value(p, idx+1) & 0xff) << 32);
+
+		ib[idx+0] = offset;
+		ib[idx+1] = upper_32_bits(offset) & 0xff;
+
 		r = evergreen_cs_track_check(p);
 		if (r) {
 			dev_warn(p->dev, "%s:%d invalid cmd stream\n", __func__, __LINE__);
 			return r;
 		}
 		break;
+	}
 	case PACKET3_DRAW_INDEX_2:
+	{
+		uint64_t offset;
+
 		if (pkt->count != 4) {
 			DRM_ERROR("bad DRAW_INDEX_2\n");
 			return -EINVAL;
@@ -1133,14 +2030,21 @@ static int evergreen_packet3_check(struct radeon_cs_parser *p,
 			DRM_ERROR("bad DRAW_INDEX_2\n");
 			return -EINVAL;
 		}
-		ib[idx+1] = idx_value + (u32)(reloc->lobj.gpu_offset & 0xffffffff);
-		ib[idx+2] += upper_32_bits(reloc->lobj.gpu_offset) & 0xff;
+
+		offset = reloc->lobj.gpu_offset +
+		         radeon_get_ib_value(p, idx+1) +
+		         ((u64)(radeon_get_ib_value(p, idx+2) & 0xff) << 32);
+
+		ib[idx+1] = offset;
+		ib[idx+2] = upper_32_bits(offset) & 0xff;
+
 		r = evergreen_cs_track_check(p);
 		if (r) {
 			dev_warn(p->dev, "%s:%d invalid cmd stream\n", __func__, __LINE__);
 			return r;
 		}
 		break;
+	}
 	case PACKET3_DRAW_INDEX_AUTO:
 		if (pkt->count != 1) {
 			DRM_ERROR("bad DRAW_INDEX_AUTO\n");
@@ -1231,13 +2135,20 @@ static int evergreen_packet3_check(struct radeon_cs_parser *p,
 		}
 		/* bit 4 is reg (0) or mem (1) */
 		if (idx_value & 0x10) {
+			uint64_t offset;
+
 			r = evergreen_cs_packet_next_reloc(p, &reloc);
 			if (r) {
 				DRM_ERROR("bad WAIT_REG_MEM\n");
 				return -EINVAL;
 			}
-			ib[idx+1] += (u32)(reloc->lobj.gpu_offset & 0xffffffff);
-			ib[idx+2] += upper_32_bits(reloc->lobj.gpu_offset) & 0xff;
+
+			offset = reloc->lobj.gpu_offset +
+			         (radeon_get_ib_value(p, idx+1) & 0xfffffffc) +
+			         ((u64)(radeon_get_ib_value(p, idx+2) & 0xff) << 32);
+
+			ib[idx+1] = (ib[idx+1] & 0x3) | (offset & 0xfffffffc);
+			ib[idx+2] = upper_32_bits(offset) & 0xff;
 		}
 		break;
 	case PACKET3_SURFACE_SYNC:
@@ -1262,16 +2173,25 @@ static int evergreen_packet3_check(struct radeon_cs_parser *p,
 			return -EINVAL;
 		}
 		if (pkt->count) {
+			uint64_t offset;
+
 			r = evergreen_cs_packet_next_reloc(p, &reloc);
 			if (r) {
 				DRM_ERROR("bad EVENT_WRITE\n");
 				return -EINVAL;
 			}
-			ib[idx+1] += (u32)(reloc->lobj.gpu_offset & 0xffffffff);
-			ib[idx+2] += upper_32_bits(reloc->lobj.gpu_offset) & 0xff;
+			offset = reloc->lobj.gpu_offset +
+			         (radeon_get_ib_value(p, idx+1) & 0xfffffff8) +
+			         ((u64)(radeon_get_ib_value(p, idx+2) & 0xff) << 32);
+
+			ib[idx+1] = offset & 0xfffffff8;
+			ib[idx+2] = upper_32_bits(offset) & 0xff;
 		}
 		break;
 	case PACKET3_EVENT_WRITE_EOP:
+	{
+		uint64_t offset;
+
 		if (pkt->count != 4) {
 			DRM_ERROR("bad EVENT_WRITE_EOP\n");
 			return -EINVAL;
@@ -1281,10 +2201,19 @@ static int evergreen_packet3_check(struct radeon_cs_parser *p,
 			DRM_ERROR("bad EVENT_WRITE_EOP\n");
 			return -EINVAL;
 		}
-		ib[idx+1] += (u32)(reloc->lobj.gpu_offset & 0xffffffff);
-		ib[idx+2] += upper_32_bits(reloc->lobj.gpu_offset) & 0xff;
+
+		offset = reloc->lobj.gpu_offset +
+		         (radeon_get_ib_value(p, idx+1) & 0xfffffffc) +
+		         ((u64)(radeon_get_ib_value(p, idx+2) & 0xff) << 32);
+
+		ib[idx+1] = offset & 0xfffffffc;
+		ib[idx+2] = (ib[idx+2] & 0xffffff00) | (upper_32_bits(offset) & 0xff);
 		break;
+	}
 	case PACKET3_EVENT_WRITE_EOS:
+	{
+		uint64_t offset;
+
 		if (pkt->count != 3) {
 			DRM_ERROR("bad EVENT_WRITE_EOS\n");
 			return -EINVAL;
@@ -1294,9 +2223,15 @@ static int evergreen_packet3_check(struct radeon_cs_parser *p,
 			DRM_ERROR("bad EVENT_WRITE_EOS\n");
 			return -EINVAL;
 		}
-		ib[idx+1] += (u32)(reloc->lobj.gpu_offset & 0xffffffff);
-		ib[idx+2] += upper_32_bits(reloc->lobj.gpu_offset) & 0xff;
+
+		offset = reloc->lobj.gpu_offset +
+		         (radeon_get_ib_value(p, idx+1) & 0xfffffffc) +
+		         ((u64)(radeon_get_ib_value(p, idx+2) & 0xff) << 32);
+
+		ib[idx+1] = offset & 0xfffffffc;
+		ib[idx+2] = (ib[idx+2] & 0xffffff00) | (upper_32_bits(offset) & 0xff);
 		break;
+	}
 	case PACKET3_SET_CONFIG_REG:
 		start_reg = (idx_value << 2) + PACKET3_SET_CONFIG_REG_START;
 		end_reg = 4 * pkt->count + start_reg - 4;
@@ -1344,6 +2279,7 @@ static int evergreen_packet3_check(struct radeon_cs_parser *p,
 		}
 		for (i = 0; i < (pkt->count / 8); i++) {
 			struct radeon_bo *texture, *mipmap;
+			u32 toffset, moffset;
 			u32 size, offset;
 
 			switch (G__SQ_CONSTANT_TYPE(radeon_get_ib_value(p, idx+1+(i*8)+7))) {
@@ -1354,32 +2290,42 @@ static int evergreen_packet3_check(struct radeon_cs_parser *p,
 					DRM_ERROR("bad SET_RESOURCE (tex)\n");
 					return -EINVAL;
 				}
-				ib[idx+1+(i*8)+2] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff);
 				if (!(p->cs_flags & RADEON_CS_KEEP_TILING_FLAGS)) {
 					ib[idx+1+(i*8)+1] |=
 						TEX_ARRAY_MODE(evergreen_cs_get_aray_mode(reloc->lobj.tiling_flags));
 					if (reloc->lobj.tiling_flags & RADEON_TILING_MACRO) {
-						ib[idx+1+(i*8)+6] |=
-							TEX_TILE_SPLIT(evergreen_cs_get_tile_split(track->row_size));
+						unsigned bankw, bankh, mtaspect, tile_split;
+
+						evergreen_tiling_fields(reloc->lobj.tiling_flags,
+									&bankw, &bankh, &mtaspect,
+									&tile_split);
+						ib[idx+1+(i*8)+6] |= TEX_TILE_SPLIT(tile_split);
 						ib[idx+1+(i*8)+7] |=
+							TEX_BANK_WIDTH(bankw) |
+							TEX_BANK_HEIGHT(bankh) |
+							MACRO_TILE_ASPECT(mtaspect) |
 							TEX_NUM_BANKS(evergreen_cs_get_num_banks(track->nbanks));
 					}
 				}
 				texture = reloc->robj;
+				toffset = (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff);
 				/* tex mip base */
 				r = evergreen_cs_packet_next_reloc(p, &reloc);
 				if (r) {
 					DRM_ERROR("bad SET_RESOURCE (tex)\n");
 					return -EINVAL;
 				}
-				ib[idx+1+(i*8)+3] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff);
+				moffset = (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff);
 				mipmap = reloc->robj;
-				r = evergreen_check_texture_resource(p,  idx+1+(i*8),
-						texture, mipmap);
+				r = evergreen_cs_track_validate_texture(p, texture, mipmap, idx+1+(i*8));
 				if (r)
 					return r;
+				ib[idx+1+(i*8)+2] += toffset;
+				ib[idx+1+(i*8)+3] += moffset;
 				break;
 			case SQ_TEX_VTX_VALID_BUFFER:
+			{
+				uint64_t offset64;
 				/* vtx base */
 				r = evergreen_cs_packet_next_reloc(p, &reloc);
 				if (r) {
@@ -1391,11 +2337,15 @@ static int evergreen_packet3_check(struct radeon_cs_parser *p,
 				if (p->rdev && (size + offset) > radeon_bo_size(reloc->robj)) {
 					/* force size to size of the buffer */
 					dev_warn(p->dev, "vbo resource seems too big for the bo\n");
-					ib[idx+1+(i*8)+1] = radeon_bo_size(reloc->robj);
+					ib[idx+1+(i*8)+1] = radeon_bo_size(reloc->robj) - offset;
 				}
-				ib[idx+1+(i*8)+0] += (u32)((reloc->lobj.gpu_offset) & 0xffffffff);
-				ib[idx+1+(i*8)+2] += upper_32_bits(reloc->lobj.gpu_offset) & 0xff;
+
+				offset64 = reloc->lobj.gpu_offset + offset;
+				ib[idx+1+(i*8)+0] = offset64;
+				ib[idx+1+(i*8)+2] = (ib[idx+1+(i*8)+2] & 0xffffff00) |
+						    (upper_32_bits(offset64) & 0xff);
 				break;
+			}
 			case SQ_TEX_VTX_INVALID_TEXTURE:
 			case SQ_TEX_VTX_INVALID_BUFFER:
 			default:
@@ -1451,6 +2401,104 @@ static int evergreen_packet3_check(struct radeon_cs_parser *p,
 			return -EINVAL;
 		}
 		break;
+	case PACKET3_STRMOUT_BUFFER_UPDATE:
+		if (pkt->count != 4) {
+			DRM_ERROR("bad STRMOUT_BUFFER_UPDATE (invalid count)\n");
+			return -EINVAL;
+		}
+		/* Updating memory at DST_ADDRESS. */
+		if (idx_value & 0x1) {
+			u64 offset;
+			r = evergreen_cs_packet_next_reloc(p, &reloc);
+			if (r) {
+				DRM_ERROR("bad STRMOUT_BUFFER_UPDATE (missing dst reloc)\n");
+				return -EINVAL;
+			}
+			offset = radeon_get_ib_value(p, idx+1);
+			offset += ((u64)(radeon_get_ib_value(p, idx+2) & 0xff)) << 32;
+			if ((offset + 4) > radeon_bo_size(reloc->robj)) {
+				DRM_ERROR("bad STRMOUT_BUFFER_UPDATE dst bo too small: 0x%llx, 0x%lx\n",
+					  offset + 4, radeon_bo_size(reloc->robj));
+				return -EINVAL;
+			}
+			offset += reloc->lobj.gpu_offset;
+			ib[idx+1] = offset;
+			ib[idx+2] = upper_32_bits(offset) & 0xff;
+		}
+		/* Reading data from SRC_ADDRESS. */
+		if (((idx_value >> 1) & 0x3) == 2) {
+			u64 offset;
+			r = evergreen_cs_packet_next_reloc(p, &reloc);
+			if (r) {
+				DRM_ERROR("bad STRMOUT_BUFFER_UPDATE (missing src reloc)\n");
+				return -EINVAL;
+			}
+			offset = radeon_get_ib_value(p, idx+3);
+			offset += ((u64)(radeon_get_ib_value(p, idx+4) & 0xff)) << 32;
+			if ((offset + 4) > radeon_bo_size(reloc->robj)) {
+				DRM_ERROR("bad STRMOUT_BUFFER_UPDATE src bo too small: 0x%llx, 0x%lx\n",
+					  offset + 4, radeon_bo_size(reloc->robj));
+				return -EINVAL;
+			}
+			offset += reloc->lobj.gpu_offset;
+			ib[idx+3] = offset;
+			ib[idx+4] = upper_32_bits(offset) & 0xff;
+		}
+		break;
+	case PACKET3_COPY_DW:
+		if (pkt->count != 4) {
+			DRM_ERROR("bad COPY_DW (invalid count)\n");
+			return -EINVAL;
+		}
+		if (idx_value & 0x1) {
+			u64 offset;
+			/* SRC is memory. */
+			r = evergreen_cs_packet_next_reloc(p, &reloc);
+			if (r) {
+				DRM_ERROR("bad COPY_DW (missing src reloc)\n");
+				return -EINVAL;
+			}
+			offset = radeon_get_ib_value(p, idx+1);
+			offset += ((u64)(radeon_get_ib_value(p, idx+2) & 0xff)) << 32;
+			if ((offset + 4) > radeon_bo_size(reloc->robj)) {
+				DRM_ERROR("bad COPY_DW src bo too small: 0x%llx, 0x%lx\n",
+					  offset + 4, radeon_bo_size(reloc->robj));
+				return -EINVAL;
+			}
+			offset += reloc->lobj.gpu_offset;
+			ib[idx+1] = offset;
+			ib[idx+2] = upper_32_bits(offset) & 0xff;
+		} else {
+			/* SRC is a reg. */
+			reg = radeon_get_ib_value(p, idx+1) << 2;
+			if (!evergreen_is_safe_reg(p, reg, idx+1))
+				return -EINVAL;
+		}
+		if (idx_value & 0x2) {
+			u64 offset;
+			/* DST is memory. */
+			r = evergreen_cs_packet_next_reloc(p, &reloc);
+			if (r) {
+				DRM_ERROR("bad COPY_DW (missing dst reloc)\n");
+				return -EINVAL;
+			}
+			offset = radeon_get_ib_value(p, idx+3);
+			offset += ((u64)(radeon_get_ib_value(p, idx+4) & 0xff)) << 32;
+			if ((offset + 4) > radeon_bo_size(reloc->robj)) {
+				DRM_ERROR("bad COPY_DW dst bo too small: 0x%llx, 0x%lx\n",
+					  offset + 4, radeon_bo_size(reloc->robj));
+				return -EINVAL;
+			}
+			offset += reloc->lobj.gpu_offset;
+			ib[idx+3] = offset;
+			ib[idx+4] = upper_32_bits(offset) & 0xff;
+		} else {
+			/* DST is a reg. */
+			reg = radeon_get_ib_value(p, idx+3) << 2;
+			if (!evergreen_is_safe_reg(p, reg, idx+3))
+				return -EINVAL;
+		}
+		break;
 	case PACKET3_NOP:
 		break;
 	default:
diff --git a/drivers/gpu/drm/radeon/evergreen_reg.h b/drivers/gpu/drm/radeon/evergreen_reg.h
index 4215de95477e..96c10b3991aa 100644
--- a/drivers/gpu/drm/radeon/evergreen_reg.h
+++ b/drivers/gpu/drm/radeon/evergreen_reg.h
@@ -219,6 +219,7 @@
 #       define EVERGREEN_CRTC_MASTER_EN                 (1 << 0)
 #       define EVERGREEN_CRTC_DISP_READ_REQUEST_DISABLE (1 << 24)
 #define EVERGREEN_CRTC_STATUS                           0x6e8c
+#       define EVERGREEN_CRTC_V_BLANK                   (1 << 0)
 #define EVERGREEN_CRTC_STATUS_POSITION                  0x6e90
 #define EVERGREEN_MASTER_UPDATE_MODE                    0x6ef8
 #define EVERGREEN_CRTC_UPDATE_LOCK                      0x6ed4
diff --git a/drivers/gpu/drm/radeon/evergreend.h b/drivers/gpu/drm/radeon/evergreend.h
index 74713d42df29..b4eefc355f16 100644
--- a/drivers/gpu/drm/radeon/evergreend.h
+++ b/drivers/gpu/drm/radeon/evergreend.h
@@ -77,6 +77,7 @@
 
 #define	CONFIG_MEMSIZE					0x5428
 
+#define	CP_COHER_BASE					0x85F8
 #define CP_ME_CNTL					0x86D8
 #define		CP_ME_HALT					(1 << 28)
 #define		CP_PFP_HALT					(1 << 26)
@@ -925,20 +926,138 @@
 #define DB_DEBUG4					0x983C
 #define DB_WATERMARKS					0x9854
 #define DB_DEPTH_CONTROL				0x28800
+#define R_028800_DB_DEPTH_CONTROL                    0x028800
+#define   S_028800_STENCIL_ENABLE(x)                   (((x) & 0x1) << 0)
+#define   G_028800_STENCIL_ENABLE(x)                   (((x) >> 0) & 0x1)
+#define   C_028800_STENCIL_ENABLE                      0xFFFFFFFE
+#define   S_028800_Z_ENABLE(x)                         (((x) & 0x1) << 1)
+#define   G_028800_Z_ENABLE(x)                         (((x) >> 1) & 0x1)
+#define   C_028800_Z_ENABLE                            0xFFFFFFFD
+#define   S_028800_Z_WRITE_ENABLE(x)                   (((x) & 0x1) << 2)
+#define   G_028800_Z_WRITE_ENABLE(x)                   (((x) >> 2) & 0x1)
+#define   C_028800_Z_WRITE_ENABLE                      0xFFFFFFFB
+#define   S_028800_ZFUNC(x)                            (((x) & 0x7) << 4)
+#define   G_028800_ZFUNC(x)                            (((x) >> 4) & 0x7)
+#define   C_028800_ZFUNC                               0xFFFFFF8F
+#define   S_028800_BACKFACE_ENABLE(x)                  (((x) & 0x1) << 7)
+#define   G_028800_BACKFACE_ENABLE(x)                  (((x) >> 7) & 0x1)
+#define   C_028800_BACKFACE_ENABLE                     0xFFFFFF7F
+#define   S_028800_STENCILFUNC(x)                      (((x) & 0x7) << 8)
+#define   G_028800_STENCILFUNC(x)                      (((x) >> 8) & 0x7)
+#define   C_028800_STENCILFUNC                         0xFFFFF8FF
+#define     V_028800_STENCILFUNC_NEVER                 0x00000000
+#define     V_028800_STENCILFUNC_LESS                  0x00000001
+#define     V_028800_STENCILFUNC_EQUAL                 0x00000002
+#define     V_028800_STENCILFUNC_LEQUAL                0x00000003
+#define     V_028800_STENCILFUNC_GREATER               0x00000004
+#define     V_028800_STENCILFUNC_NOTEQUAL              0x00000005
+#define     V_028800_STENCILFUNC_GEQUAL                0x00000006
+#define     V_028800_STENCILFUNC_ALWAYS                0x00000007
+#define   S_028800_STENCILFAIL(x)                      (((x) & 0x7) << 11)
+#define   G_028800_STENCILFAIL(x)                      (((x) >> 11) & 0x7)
+#define   C_028800_STENCILFAIL                         0xFFFFC7FF
+#define     V_028800_STENCIL_KEEP                      0x00000000
+#define     V_028800_STENCIL_ZERO                      0x00000001
+#define     V_028800_STENCIL_REPLACE                   0x00000002
+#define     V_028800_STENCIL_INCR                      0x00000003
+#define     V_028800_STENCIL_DECR                      0x00000004
+#define     V_028800_STENCIL_INVERT                    0x00000005
+#define     V_028800_STENCIL_INCR_WRAP                 0x00000006
+#define     V_028800_STENCIL_DECR_WRAP                 0x00000007
+#define   S_028800_STENCILZPASS(x)                     (((x) & 0x7) << 14)
+#define   G_028800_STENCILZPASS(x)                     (((x) >> 14) & 0x7)
+#define   C_028800_STENCILZPASS                        0xFFFE3FFF
+#define   S_028800_STENCILZFAIL(x)                     (((x) & 0x7) << 17)
+#define   G_028800_STENCILZFAIL(x)                     (((x) >> 17) & 0x7)
+#define   C_028800_STENCILZFAIL                        0xFFF1FFFF
+#define   S_028800_STENCILFUNC_BF(x)                   (((x) & 0x7) << 20)
+#define   G_028800_STENCILFUNC_BF(x)                   (((x) >> 20) & 0x7)
+#define   C_028800_STENCILFUNC_BF                      0xFF8FFFFF
+#define   S_028800_STENCILFAIL_BF(x)                   (((x) & 0x7) << 23)
+#define   G_028800_STENCILFAIL_BF(x)                   (((x) >> 23) & 0x7)
+#define   C_028800_STENCILFAIL_BF                      0xFC7FFFFF
+#define   S_028800_STENCILZPASS_BF(x)                  (((x) & 0x7) << 26)
+#define   G_028800_STENCILZPASS_BF(x)                  (((x) >> 26) & 0x7)
+#define   C_028800_STENCILZPASS_BF                     0xE3FFFFFF
+#define   S_028800_STENCILZFAIL_BF(x)                  (((x) & 0x7) << 29)
+#define   G_028800_STENCILZFAIL_BF(x)                  (((x) >> 29) & 0x7)
+#define   C_028800_STENCILZFAIL_BF                     0x1FFFFFFF
 #define DB_DEPTH_VIEW					0x28008
+#define R_028008_DB_DEPTH_VIEW                       0x00028008
+#define   S_028008_SLICE_START(x)                      (((x) & 0x7FF) << 0)
+#define   G_028008_SLICE_START(x)                      (((x) >> 0) & 0x7FF)
+#define   C_028008_SLICE_START                         0xFFFFF800
+#define   S_028008_SLICE_MAX(x)                        (((x) & 0x7FF) << 13)
+#define   G_028008_SLICE_MAX(x)                        (((x) >> 13) & 0x7FF)
+#define   C_028008_SLICE_MAX                           0xFF001FFF
 #define DB_HTILE_DATA_BASE				0x28014
+#define DB_HTILE_SURFACE				0x28abc
+#define   S_028ABC_HTILE_WIDTH(x)                      (((x) & 0x1) << 0)
+#define   G_028ABC_HTILE_WIDTH(x)                      (((x) >> 0) & 0x1)
+#define   C_028ABC_HTILE_WIDTH                         0xFFFFFFFE
+#define   S_028ABC_HTILE_HEIGHT(x)                      (((x) & 0x1) << 1)
+#define   G_028ABC_HTILE_HEIGHT(x)                      (((x) >> 1) & 0x1)
+#define   C_028ABC_HTILE_HEIGHT                         0xFFFFFFFD
+#define   G_028ABC_LINEAR(x)                           (((x) >> 2) & 0x1)
 #define DB_Z_INFO					0x28040
 #       define Z_ARRAY_MODE(x)                          ((x) << 4)
 #       define DB_TILE_SPLIT(x)                         (((x) & 0x7) << 8)
 #       define DB_NUM_BANKS(x)                          (((x) & 0x3) << 12)
 #       define DB_BANK_WIDTH(x)                         (((x) & 0x3) << 16)
 #       define DB_BANK_HEIGHT(x)                        (((x) & 0x3) << 20)
+#       define DB_MACRO_TILE_ASPECT(x)                  (((x) & 0x3) << 24)
+#define R_028040_DB_Z_INFO                       0x028040
+#define   S_028040_FORMAT(x)                           (((x) & 0x3) << 0)
+#define   G_028040_FORMAT(x)                           (((x) >> 0) & 0x3)
+#define   C_028040_FORMAT                              0xFFFFFFFC
+#define     V_028040_Z_INVALID                     0x00000000
+#define     V_028040_Z_16                          0x00000001
+#define     V_028040_Z_24                          0x00000002
+#define     V_028040_Z_32_FLOAT                    0x00000003
+#define   S_028040_ARRAY_MODE(x)                       (((x) & 0xF) << 4)
+#define   G_028040_ARRAY_MODE(x)                       (((x) >> 4) & 0xF)
+#define   C_028040_ARRAY_MODE                          0xFFFFFF0F
+#define   S_028040_READ_SIZE(x)                        (((x) & 0x1) << 28)
+#define   G_028040_READ_SIZE(x)                        (((x) >> 28) & 0x1)
+#define   C_028040_READ_SIZE                           0xEFFFFFFF
+#define   S_028040_TILE_SURFACE_ENABLE(x)              (((x) & 0x1) << 29)
+#define   G_028040_TILE_SURFACE_ENABLE(x)              (((x) >> 29) & 0x1)
+#define   C_028040_TILE_SURFACE_ENABLE                 0xDFFFFFFF
+#define   S_028040_ZRANGE_PRECISION(x)                 (((x) & 0x1) << 31)
+#define   G_028040_ZRANGE_PRECISION(x)                 (((x) >> 31) & 0x1)
+#define   C_028040_ZRANGE_PRECISION                    0x7FFFFFFF
+#define   S_028040_TILE_SPLIT(x)                       (((x) & 0x7) << 8)
+#define   G_028040_TILE_SPLIT(x)                       (((x) >> 8) & 0x7)
+#define   S_028040_NUM_BANKS(x)                        (((x) & 0x3) << 12)
+#define   G_028040_NUM_BANKS(x)                        (((x) >> 12) & 0x3)
+#define   S_028040_BANK_WIDTH(x)                       (((x) & 0x3) << 16)
+#define   G_028040_BANK_WIDTH(x)                       (((x) >> 16) & 0x3)
+#define   S_028040_BANK_HEIGHT(x)                      (((x) & 0x3) << 20)
+#define   G_028040_BANK_HEIGHT(x)                      (((x) >> 20) & 0x3)
+#define   S_028040_MACRO_TILE_ASPECT(x)                (((x) & 0x3) << 24)
+#define   G_028040_MACRO_TILE_ASPECT(x)                (((x) >> 24) & 0x3)
 #define DB_STENCIL_INFO					0x28044
+#define R_028044_DB_STENCIL_INFO                     0x028044
+#define   S_028044_FORMAT(x)                           (((x) & 0x1) << 0)
+#define   G_028044_FORMAT(x)                           (((x) >> 0) & 0x1)
+#define   C_028044_FORMAT                              0xFFFFFFFE
+#define   G_028044_TILE_SPLIT(x)                       (((x) >> 8) & 0x7)
 #define DB_Z_READ_BASE					0x28048
 #define DB_STENCIL_READ_BASE				0x2804c
 #define DB_Z_WRITE_BASE					0x28050
 #define DB_STENCIL_WRITE_BASE				0x28054
 #define DB_DEPTH_SIZE					0x28058
+#define R_028058_DB_DEPTH_SIZE                       0x028058
+#define   S_028058_PITCH_TILE_MAX(x)                   (((x) & 0x7FF) << 0)
+#define   G_028058_PITCH_TILE_MAX(x)                   (((x) >> 0) & 0x7FF)
+#define   C_028058_PITCH_TILE_MAX                      0xFFFFF800
+#define   S_028058_HEIGHT_TILE_MAX(x)                   (((x) & 0x7FF) << 11)
+#define   G_028058_HEIGHT_TILE_MAX(x)                   (((x) >> 11) & 0x7FF)
+#define   C_028058_HEIGHT_TILE_MAX                      0xFFC007FF
+#define R_02805C_DB_DEPTH_SLICE                      0x02805C
+#define   S_02805C_SLICE_TILE_MAX(x)                   (((x) & 0x3FFFFF) << 0)
+#define   G_02805C_SLICE_TILE_MAX(x)                   (((x) >> 0) & 0x3FFFFF)
+#define   C_02805C_SLICE_TILE_MAX                      0xFFC00000
 
 #define SQ_PGM_START_PS					0x28840
 #define SQ_PGM_START_VS					0x2885c
@@ -948,6 +1067,14 @@
 #define SQ_PGM_START_HS					0x288b8
 #define SQ_PGM_START_LS					0x288d0
 
+#define	VGT_STRMOUT_BUFFER_BASE_0			0x28AD8
+#define	VGT_STRMOUT_BUFFER_BASE_1			0x28AE8
+#define	VGT_STRMOUT_BUFFER_BASE_2			0x28AF8
+#define	VGT_STRMOUT_BUFFER_BASE_3			0x28B08
+#define VGT_STRMOUT_BUFFER_SIZE_0			0x28AD0
+#define VGT_STRMOUT_BUFFER_SIZE_1			0x28AE0
+#define VGT_STRMOUT_BUFFER_SIZE_2			0x28AF0
+#define VGT_STRMOUT_BUFFER_SIZE_3			0x28B00
 #define VGT_STRMOUT_CONFIG				0x28b94
 #define VGT_STRMOUT_BUFFER_CONFIG			0x28b98
 
@@ -974,6 +1101,114 @@
 #define	CB_COLOR0_PITCH					0x28c64
 #define	CB_COLOR0_SLICE					0x28c68
 #define	CB_COLOR0_VIEW					0x28c6c
+#define R_028C6C_CB_COLOR0_VIEW                      0x00028C6C
+#define   S_028C6C_SLICE_START(x)                      (((x) & 0x7FF) << 0)
+#define   G_028C6C_SLICE_START(x)                      (((x) >> 0) & 0x7FF)
+#define   C_028C6C_SLICE_START                         0xFFFFF800
+#define   S_028C6C_SLICE_MAX(x)                        (((x) & 0x7FF) << 13)
+#define   G_028C6C_SLICE_MAX(x)                        (((x) >> 13) & 0x7FF)
+#define   C_028C6C_SLICE_MAX                           0xFF001FFF
+#define R_028C70_CB_COLOR0_INFO                      0x028C70
+#define   S_028C70_ENDIAN(x)                           (((x) & 0x3) << 0)
+#define   G_028C70_ENDIAN(x)                           (((x) >> 0) & 0x3)
+#define   C_028C70_ENDIAN                              0xFFFFFFFC
+#define   S_028C70_FORMAT(x)                           (((x) & 0x3F) << 2)
+#define   G_028C70_FORMAT(x)                           (((x) >> 2) & 0x3F)
+#define   C_028C70_FORMAT                              0xFFFFFF03
+#define     V_028C70_COLOR_INVALID                     0x00000000
+#define     V_028C70_COLOR_8                           0x00000001
+#define     V_028C70_COLOR_4_4                         0x00000002
+#define     V_028C70_COLOR_3_3_2                       0x00000003
+#define     V_028C70_COLOR_16                          0x00000005
+#define     V_028C70_COLOR_16_FLOAT                    0x00000006
+#define     V_028C70_COLOR_8_8                         0x00000007
+#define     V_028C70_COLOR_5_6_5                       0x00000008
+#define     V_028C70_COLOR_6_5_5                       0x00000009
+#define     V_028C70_COLOR_1_5_5_5                     0x0000000A
+#define     V_028C70_COLOR_4_4_4_4                     0x0000000B
+#define     V_028C70_COLOR_5_5_5_1                     0x0000000C
+#define     V_028C70_COLOR_32                          0x0000000D
+#define     V_028C70_COLOR_32_FLOAT                    0x0000000E
+#define     V_028C70_COLOR_16_16                       0x0000000F
+#define     V_028C70_COLOR_16_16_FLOAT                 0x00000010
+#define     V_028C70_COLOR_8_24                        0x00000011
+#define     V_028C70_COLOR_8_24_FLOAT                  0x00000012
+#define     V_028C70_COLOR_24_8                        0x00000013
+#define     V_028C70_COLOR_24_8_FLOAT                  0x00000014
+#define     V_028C70_COLOR_10_11_11                    0x00000015
+#define     V_028C70_COLOR_10_11_11_FLOAT              0x00000016
+#define     V_028C70_COLOR_11_11_10                    0x00000017
+#define     V_028C70_COLOR_11_11_10_FLOAT              0x00000018
+#define     V_028C70_COLOR_2_10_10_10                  0x00000019
+#define     V_028C70_COLOR_8_8_8_8                     0x0000001A
+#define     V_028C70_COLOR_10_10_10_2                  0x0000001B
+#define     V_028C70_COLOR_X24_8_32_FLOAT              0x0000001C
+#define     V_028C70_COLOR_32_32                       0x0000001D
+#define     V_028C70_COLOR_32_32_FLOAT                 0x0000001E
+#define     V_028C70_COLOR_16_16_16_16                 0x0000001F
+#define     V_028C70_COLOR_16_16_16_16_FLOAT           0x00000020
+#define     V_028C70_COLOR_32_32_32_32                 0x00000022
+#define     V_028C70_COLOR_32_32_32_32_FLOAT           0x00000023
+#define     V_028C70_COLOR_32_32_32_FLOAT              0x00000030
+#define   S_028C70_ARRAY_MODE(x)                       (((x) & 0xF) << 8)
+#define   G_028C70_ARRAY_MODE(x)                       (((x) >> 8) & 0xF)
+#define   C_028C70_ARRAY_MODE                          0xFFFFF0FF
+#define     V_028C70_ARRAY_LINEAR_GENERAL              0x00000000
+#define     V_028C70_ARRAY_LINEAR_ALIGNED              0x00000001
+#define     V_028C70_ARRAY_1D_TILED_THIN1              0x00000002
+#define     V_028C70_ARRAY_2D_TILED_THIN1              0x00000004
+#define   S_028C70_NUMBER_TYPE(x)                      (((x) & 0x7) << 12)
+#define   G_028C70_NUMBER_TYPE(x)                      (((x) >> 12) & 0x7)
+#define   C_028C70_NUMBER_TYPE                         0xFFFF8FFF
+#define     V_028C70_NUMBER_UNORM                      0x00000000
+#define     V_028C70_NUMBER_SNORM                      0x00000001
+#define     V_028C70_NUMBER_USCALED                    0x00000002
+#define     V_028C70_NUMBER_SSCALED                    0x00000003
+#define     V_028C70_NUMBER_UINT                       0x00000004
+#define     V_028C70_NUMBER_SINT                       0x00000005
+#define     V_028C70_NUMBER_SRGB                       0x00000006
+#define     V_028C70_NUMBER_FLOAT                      0x00000007
+#define   S_028C70_COMP_SWAP(x)                        (((x) & 0x3) << 15)
+#define   G_028C70_COMP_SWAP(x)                        (((x) >> 15) & 0x3)
+#define   C_028C70_COMP_SWAP                           0xFFFE7FFF
+#define     V_028C70_SWAP_STD                          0x00000000
+#define     V_028C70_SWAP_ALT                          0x00000001
+#define     V_028C70_SWAP_STD_REV                      0x00000002
+#define     V_028C70_SWAP_ALT_REV                      0x00000003
+#define   S_028C70_FAST_CLEAR(x)                       (((x) & 0x1) << 17)
+#define   G_028C70_FAST_CLEAR(x)                       (((x) >> 17) & 0x1)
+#define   C_028C70_FAST_CLEAR                          0xFFFDFFFF
+#define   S_028C70_COMPRESSION(x)                      (((x) & 0x3) << 18)
+#define   G_028C70_COMPRESSION(x)                      (((x) >> 18) & 0x3)
+#define   C_028C70_COMPRESSION                         0xFFF3FFFF
+#define   S_028C70_BLEND_CLAMP(x)                      (((x) & 0x1) << 19)
+#define   G_028C70_BLEND_CLAMP(x)                      (((x) >> 19) & 0x1)
+#define   C_028C70_BLEND_CLAMP                         0xFFF7FFFF
+#define   S_028C70_BLEND_BYPASS(x)                     (((x) & 0x1) << 20)
+#define   G_028C70_BLEND_BYPASS(x)                     (((x) >> 20) & 0x1)
+#define   C_028C70_BLEND_BYPASS                        0xFFEFFFFF
+#define   S_028C70_SIMPLE_FLOAT(x)                     (((x) & 0x1) << 21)
+#define   G_028C70_SIMPLE_FLOAT(x)                     (((x) >> 21) & 0x1)
+#define   C_028C70_SIMPLE_FLOAT                        0xFFDFFFFF
+#define   S_028C70_ROUND_MODE(x)                       (((x) & 0x1) << 22)
+#define   G_028C70_ROUND_MODE(x)                       (((x) >> 22) & 0x1)
+#define   C_028C70_ROUND_MODE                          0xFFBFFFFF
+#define   S_028C70_TILE_COMPACT(x)                     (((x) & 0x1) << 23)
+#define   G_028C70_TILE_COMPACT(x)                     (((x) >> 23) & 0x1)
+#define   C_028C70_TILE_COMPACT                        0xFF7FFFFF
+#define   S_028C70_SOURCE_FORMAT(x)                    (((x) & 0x3) << 24)
+#define   G_028C70_SOURCE_FORMAT(x)                    (((x) >> 24) & 0x3)
+#define   C_028C70_SOURCE_FORMAT                       0xFCFFFFFF
+#define     V_028C70_EXPORT_4C_32BPC                   0x0
+#define     V_028C70_EXPORT_4C_16BPC                   0x1
+#define     V_028C70_EXPORT_2C_32BPC                   0x2 /* Do not use */
+#define   S_028C70_RAT(x)                              (((x) & 0x1) << 26)
+#define   G_028C70_RAT(x)                              (((x) >> 26) & 0x1)
+#define   C_028C70_RAT                                 0xFBFFFFFF
+#define   S_028C70_RESOURCE_TYPE(x)                    (((x) & 0x7) << 27)
+#define   G_028C70_RESOURCE_TYPE(x)                    (((x) >> 27) & 0x7)
+#define   C_028C70_RESOURCE_TYPE                       0xC7FFFFFF
+
 #define	CB_COLOR0_INFO					0x28c70
 #	define CB_FORMAT(x)				((x) << 2)
 #       define CB_ARRAY_MODE(x)                         ((x) << 8)
@@ -984,6 +1219,20 @@
 #	define CB_SOURCE_FORMAT(x)			((x) << 24)
 #	define CB_SF_EXPORT_FULL			0
 #	define CB_SF_EXPORT_NORM			1
+#define R_028C74_CB_COLOR0_ATTRIB                      0x028C74
+#define   S_028C74_NON_DISP_TILING_ORDER(x)            (((x) & 0x1) << 4)
+#define   G_028C74_NON_DISP_TILING_ORDER(x)            (((x) >> 4) & 0x1)
+#define   C_028C74_NON_DISP_TILING_ORDER               0xFFFFFFEF
+#define   S_028C74_TILE_SPLIT(x)                       (((x) & 0xf) << 5)
+#define   G_028C74_TILE_SPLIT(x)                       (((x) >> 5) & 0xf)
+#define   S_028C74_NUM_BANKS(x)                        (((x) & 0x3) << 10)
+#define   G_028C74_NUM_BANKS(x)                        (((x) >> 10) & 0x3)
+#define   S_028C74_BANK_WIDTH(x)                       (((x) & 0x3) << 13)
+#define   G_028C74_BANK_WIDTH(x)                       (((x) >> 13) & 0x3)
+#define   S_028C74_BANK_HEIGHT(x)                      (((x) & 0x3) << 16)
+#define   G_028C74_BANK_HEIGHT(x)                      (((x) >> 16) & 0x3)
+#define   S_028C74_MACRO_TILE_ASPECT(x)                (((x) & 0x3) << 19)
+#define   G_028C74_MACRO_TILE_ASPECT(x)                (((x) >> 19) & 0x3)
 #define	CB_COLOR0_ATTRIB				0x28c74
 #       define CB_TILE_SPLIT(x)                         (((x) & 0x7) << 5)
 #       define ADDR_SURF_TILE_SPLIT_64B                 0
@@ -1008,6 +1257,7 @@
 #       define ADDR_SURF_BANK_HEIGHT_2                  1
 #       define ADDR_SURF_BANK_HEIGHT_4                  2
 #       define ADDR_SURF_BANK_HEIGHT_8                  3
+#       define CB_MACRO_TILE_ASPECT(x)                  (((x) & 0x3) << 19)
 #define	CB_COLOR0_DIM					0x28c78
 /* only CB0-7 blocks have these regs */
 #define	CB_COLOR0_CMASK					0x28c7c
@@ -1196,9 +1446,144 @@
 #define SQ_TEX_RESOURCE_WORD6_0                         0x30018
 #       define TEX_TILE_SPLIT(x)                        (((x) & 0x7) << 29)
 #define SQ_TEX_RESOURCE_WORD7_0                         0x3001c
+#       define MACRO_TILE_ASPECT(x)                     (((x) & 0x3) << 6)
 #       define TEX_BANK_WIDTH(x)                        (((x) & 0x3) << 8)
 #       define TEX_BANK_HEIGHT(x)                       (((x) & 0x3) << 10)
 #       define TEX_NUM_BANKS(x)                         (((x) & 0x3) << 16)
+#define R_030000_SQ_TEX_RESOURCE_WORD0_0             0x030000
+#define   S_030000_DIM(x)                              (((x) & 0x7) << 0)
+#define   G_030000_DIM(x)                              (((x) >> 0) & 0x7)
+#define   C_030000_DIM                                 0xFFFFFFF8
+#define     V_030000_SQ_TEX_DIM_1D                     0x00000000
+#define     V_030000_SQ_TEX_DIM_2D                     0x00000001
+#define     V_030000_SQ_TEX_DIM_3D                     0x00000002
+#define     V_030000_SQ_TEX_DIM_CUBEMAP                0x00000003
+#define     V_030000_SQ_TEX_DIM_1D_ARRAY               0x00000004
+#define     V_030000_SQ_TEX_DIM_2D_ARRAY               0x00000005
+#define     V_030000_SQ_TEX_DIM_2D_MSAA                0x00000006
+#define     V_030000_SQ_TEX_DIM_2D_ARRAY_MSAA          0x00000007
+#define   S_030000_NON_DISP_TILING_ORDER(x)            (((x) & 0x1) << 5)
+#define   G_030000_NON_DISP_TILING_ORDER(x)            (((x) >> 5) & 0x1)
+#define   C_030000_NON_DISP_TILING_ORDER               0xFFFFFFDF
+#define   S_030000_PITCH(x)                            (((x) & 0xFFF) << 6)
+#define   G_030000_PITCH(x)                            (((x) >> 6) & 0xFFF)
+#define   C_030000_PITCH                               0xFFFC003F
+#define   S_030000_TEX_WIDTH(x)                        (((x) & 0x3FFF) << 18)
+#define   G_030000_TEX_WIDTH(x)                        (((x) >> 18) & 0x3FFF)
+#define   C_030000_TEX_WIDTH                           0x0003FFFF
+#define R_030004_SQ_TEX_RESOURCE_WORD1_0             0x030004
+#define   S_030004_TEX_HEIGHT(x)                       (((x) & 0x3FFF) << 0)
+#define   G_030004_TEX_HEIGHT(x)                       (((x) >> 0) & 0x3FFF)
+#define   C_030004_TEX_HEIGHT                          0xFFFFC000
+#define   S_030004_TEX_DEPTH(x)                        (((x) & 0x1FFF) << 14)
+#define   G_030004_TEX_DEPTH(x)                        (((x) >> 14) & 0x1FFF)
+#define   C_030004_TEX_DEPTH                           0xF8003FFF
+#define   S_030004_ARRAY_MODE(x)                       (((x) & 0xF) << 28)
+#define   G_030004_ARRAY_MODE(x)                       (((x) >> 28) & 0xF)
+#define   C_030004_ARRAY_MODE                          0x0FFFFFFF
+#define R_030008_SQ_TEX_RESOURCE_WORD2_0             0x030008
+#define   S_030008_BASE_ADDRESS(x)                     (((x) & 0xFFFFFFFF) << 0)
+#define   G_030008_BASE_ADDRESS(x)                     (((x) >> 0) & 0xFFFFFFFF)
+#define   C_030008_BASE_ADDRESS                        0x00000000
+#define R_03000C_SQ_TEX_RESOURCE_WORD3_0             0x03000C
+#define   S_03000C_MIP_ADDRESS(x)                      (((x) & 0xFFFFFFFF) << 0)
+#define   G_03000C_MIP_ADDRESS(x)                      (((x) >> 0) & 0xFFFFFFFF)
+#define   C_03000C_MIP_ADDRESS                         0x00000000
+#define R_030010_SQ_TEX_RESOURCE_WORD4_0             0x030010
+#define   S_030010_FORMAT_COMP_X(x)                    (((x) & 0x3) << 0)
+#define   G_030010_FORMAT_COMP_X(x)                    (((x) >> 0) & 0x3)
+#define   C_030010_FORMAT_COMP_X                       0xFFFFFFFC
+#define     V_030010_SQ_FORMAT_COMP_UNSIGNED           0x00000000
+#define     V_030010_SQ_FORMAT_COMP_SIGNED             0x00000001
+#define     V_030010_SQ_FORMAT_COMP_UNSIGNED_BIASED    0x00000002
+#define   S_030010_FORMAT_COMP_Y(x)                    (((x) & 0x3) << 2)
+#define   G_030010_FORMAT_COMP_Y(x)                    (((x) >> 2) & 0x3)
+#define   C_030010_FORMAT_COMP_Y                       0xFFFFFFF3
+#define   S_030010_FORMAT_COMP_Z(x)                    (((x) & 0x3) << 4)
+#define   G_030010_FORMAT_COMP_Z(x)                    (((x) >> 4) & 0x3)
+#define   C_030010_FORMAT_COMP_Z                       0xFFFFFFCF
+#define   S_030010_FORMAT_COMP_W(x)                    (((x) & 0x3) << 6)
+#define   G_030010_FORMAT_COMP_W(x)                    (((x) >> 6) & 0x3)
+#define   C_030010_FORMAT_COMP_W                       0xFFFFFF3F
+#define   S_030010_NUM_FORMAT_ALL(x)                   (((x) & 0x3) << 8)
+#define   G_030010_NUM_FORMAT_ALL(x)                   (((x) >> 8) & 0x3)
+#define   C_030010_NUM_FORMAT_ALL                      0xFFFFFCFF
+#define     V_030010_SQ_NUM_FORMAT_NORM                0x00000000
+#define     V_030010_SQ_NUM_FORMAT_INT                 0x00000001
+#define     V_030010_SQ_NUM_FORMAT_SCALED              0x00000002
+#define   S_030010_SRF_MODE_ALL(x)                     (((x) & 0x1) << 10)
+#define   G_030010_SRF_MODE_ALL(x)                     (((x) >> 10) & 0x1)
+#define   C_030010_SRF_MODE_ALL                        0xFFFFFBFF
+#define     V_030010_SRF_MODE_ZERO_CLAMP_MINUS_ONE     0x00000000
+#define     V_030010_SRF_MODE_NO_ZERO                  0x00000001
+#define   S_030010_FORCE_DEGAMMA(x)                    (((x) & 0x1) << 11)
+#define   G_030010_FORCE_DEGAMMA(x)                    (((x) >> 11) & 0x1)
+#define   C_030010_FORCE_DEGAMMA                       0xFFFFF7FF
+#define   S_030010_ENDIAN_SWAP(x)                      (((x) & 0x3) << 12)
+#define   G_030010_ENDIAN_SWAP(x)                      (((x) >> 12) & 0x3)
+#define   C_030010_ENDIAN_SWAP                         0xFFFFCFFF
+#define   S_030010_DST_SEL_X(x)                        (((x) & 0x7) << 16)
+#define   G_030010_DST_SEL_X(x)                        (((x) >> 16) & 0x7)
+#define   C_030010_DST_SEL_X                           0xFFF8FFFF
+#define     V_030010_SQ_SEL_X                          0x00000000
+#define     V_030010_SQ_SEL_Y                          0x00000001
+#define     V_030010_SQ_SEL_Z                          0x00000002
+#define     V_030010_SQ_SEL_W                          0x00000003
+#define     V_030010_SQ_SEL_0                          0x00000004
+#define     V_030010_SQ_SEL_1                          0x00000005
+#define   S_030010_DST_SEL_Y(x)                        (((x) & 0x7) << 19)
+#define   G_030010_DST_SEL_Y(x)                        (((x) >> 19) & 0x7)
+#define   C_030010_DST_SEL_Y                           0xFFC7FFFF
+#define   S_030010_DST_SEL_Z(x)                        (((x) & 0x7) << 22)
+#define   G_030010_DST_SEL_Z(x)                        (((x) >> 22) & 0x7)
+#define   C_030010_DST_SEL_Z                           0xFE3FFFFF
+#define   S_030010_DST_SEL_W(x)                        (((x) & 0x7) << 25)
+#define   G_030010_DST_SEL_W(x)                        (((x) >> 25) & 0x7)
+#define   C_030010_DST_SEL_W                           0xF1FFFFFF
+#define   S_030010_BASE_LEVEL(x)                       (((x) & 0xF) << 28)
+#define   G_030010_BASE_LEVEL(x)                       (((x) >> 28) & 0xF)
+#define   C_030010_BASE_LEVEL                          0x0FFFFFFF
+#define R_030014_SQ_TEX_RESOURCE_WORD5_0             0x030014
+#define   S_030014_LAST_LEVEL(x)                       (((x) & 0xF) << 0)
+#define   G_030014_LAST_LEVEL(x)                       (((x) >> 0) & 0xF)
+#define   C_030014_LAST_LEVEL                          0xFFFFFFF0
+#define   S_030014_BASE_ARRAY(x)                       (((x) & 0x1FFF) << 4)
+#define   G_030014_BASE_ARRAY(x)                       (((x) >> 4) & 0x1FFF)
+#define   C_030014_BASE_ARRAY                          0xFFFE000F
+#define   S_030014_LAST_ARRAY(x)                       (((x) & 0x1FFF) << 17)
+#define   G_030014_LAST_ARRAY(x)                       (((x) >> 17) & 0x1FFF)
+#define   C_030014_LAST_ARRAY                          0xC001FFFF
+#define R_030018_SQ_TEX_RESOURCE_WORD6_0             0x030018
+#define   S_030018_MAX_ANISO(x)                        (((x) & 0x7) << 0)
+#define   G_030018_MAX_ANISO(x)                        (((x) >> 0) & 0x7)
+#define   C_030018_MAX_ANISO                           0xFFFFFFF8
+#define   S_030018_PERF_MODULATION(x)                  (((x) & 0x7) << 3)
+#define   G_030018_PERF_MODULATION(x)                  (((x) >> 3) & 0x7)
+#define   C_030018_PERF_MODULATION                     0xFFFFFFC7
+#define   S_030018_INTERLACED(x)                       (((x) & 0x1) << 6)
+#define   G_030018_INTERLACED(x)                       (((x) >> 6) & 0x1)
+#define   C_030018_INTERLACED                          0xFFFFFFBF
+#define   S_030018_TILE_SPLIT(x)                       (((x) & 0x7) << 29)
+#define   G_030018_TILE_SPLIT(x)                       (((x) >> 29) & 0x7)
+#define R_03001C_SQ_TEX_RESOURCE_WORD7_0             0x03001C
+#define   S_03001C_MACRO_TILE_ASPECT(x)                (((x) & 0x3) << 6)
+#define   G_03001C_MACRO_TILE_ASPECT(x)                (((x) >> 6) & 0x3)
+#define   S_03001C_BANK_WIDTH(x)                       (((x) & 0x3) << 8)
+#define   G_03001C_BANK_WIDTH(x)                       (((x) >> 8) & 0x3)
+#define   S_03001C_BANK_HEIGHT(x)                      (((x) & 0x3) << 10)
+#define   G_03001C_BANK_HEIGHT(x)                      (((x) >> 10) & 0x3)
+#define   S_03001C_NUM_BANKS(x)                        (((x) & 0x3) << 16)
+#define   G_03001C_NUM_BANKS(x)                        (((x) >> 16) & 0x3)
+#define   S_03001C_TYPE(x)                             (((x) & 0x3) << 30)
+#define   G_03001C_TYPE(x)                             (((x) >> 30) & 0x3)
+#define   C_03001C_TYPE                                0x3FFFFFFF
+#define     V_03001C_SQ_TEX_VTX_INVALID_TEXTURE        0x00000000
+#define     V_03001C_SQ_TEX_VTX_INVALID_BUFFER         0x00000001
+#define     V_03001C_SQ_TEX_VTX_VALID_TEXTURE          0x00000002
+#define     V_03001C_SQ_TEX_VTX_VALID_BUFFER           0x00000003
+#define   S_03001C_DATA_FORMAT(x)                      (((x) & 0x3F) << 0)
+#define   G_03001C_DATA_FORMAT(x)                      (((x) >> 0) & 0x3F)
+#define   C_03001C_DATA_FORMAT                         0xFFFFFFC0
 
 #define SQ_VTX_CONSTANT_WORD0_0				0x30000
 #define SQ_VTX_CONSTANT_WORD1_0				0x30004
diff --git a/drivers/gpu/drm/radeon/ni.c b/drivers/gpu/drm/radeon/ni.c
index db09065e68fd..a48ca53fcd6a 100644
--- a/drivers/gpu/drm/radeon/ni.c
+++ b/drivers/gpu/drm/radeon/ni.c
@@ -42,6 +42,8 @@ extern void evergreen_irq_suspend(struct radeon_device *rdev);
 extern int evergreen_mc_init(struct radeon_device *rdev);
 extern void evergreen_fix_pci_max_read_req_size(struct radeon_device *rdev);
 extern void evergreen_pcie_gen2_enable(struct radeon_device *rdev);
+extern void si_rlc_fini(struct radeon_device *rdev);
+extern int si_rlc_init(struct radeon_device *rdev);
 
 #define EVERGREEN_PFP_UCODE_SIZE 1120
 #define EVERGREEN_PM4_UCODE_SIZE 1376
@@ -53,6 +55,8 @@ extern void evergreen_pcie_gen2_enable(struct radeon_device *rdev);
 #define CAYMAN_RLC_UCODE_SIZE 1024
 #define CAYMAN_MC_UCODE_SIZE 6037
 
+#define ARUBA_RLC_UCODE_SIZE 1536
+
 /* Firmware Names */
 MODULE_FIRMWARE("radeon/BARTS_pfp.bin");
 MODULE_FIRMWARE("radeon/BARTS_me.bin");
@@ -68,6 +72,9 @@ MODULE_FIRMWARE("radeon/CAYMAN_pfp.bin");
 MODULE_FIRMWARE("radeon/CAYMAN_me.bin");
 MODULE_FIRMWARE("radeon/CAYMAN_mc.bin");
 MODULE_FIRMWARE("radeon/CAYMAN_rlc.bin");
+MODULE_FIRMWARE("radeon/ARUBA_pfp.bin");
+MODULE_FIRMWARE("radeon/ARUBA_me.bin");
+MODULE_FIRMWARE("radeon/ARUBA_rlc.bin");
 
 #define BTC_IO_MC_REGS_SIZE 29
 
@@ -326,6 +333,15 @@ int ni_init_microcode(struct radeon_device *rdev)
 		rlc_req_size = CAYMAN_RLC_UCODE_SIZE * 4;
 		mc_req_size = CAYMAN_MC_UCODE_SIZE * 4;
 		break;
+	case CHIP_ARUBA:
+		chip_name = "ARUBA";
+		rlc_chip_name = "ARUBA";
+		/* pfp/me same size as CAYMAN */
+		pfp_req_size = CAYMAN_PFP_UCODE_SIZE * 4;
+		me_req_size = CAYMAN_PM4_UCODE_SIZE * 4;
+		rlc_req_size = ARUBA_RLC_UCODE_SIZE * 4;
+		mc_req_size = 0;
+		break;
 	default: BUG();
 	}
 
@@ -365,15 +381,18 @@ int ni_init_microcode(struct radeon_device *rdev)
 		err = -EINVAL;
 	}
 
-	snprintf(fw_name, sizeof(fw_name), "radeon/%s_mc.bin", chip_name);
-	err = request_firmware(&rdev->mc_fw, fw_name, &pdev->dev);
-	if (err)
-		goto out;
-	if (rdev->mc_fw->size != mc_req_size) {
-		printk(KERN_ERR
-		       "ni_mc: Bogus length %zu in firmware \"%s\"\n",
-		       rdev->mc_fw->size, fw_name);
-		err = -EINVAL;
+	/* no MC ucode on TN */
+	if (!(rdev->flags & RADEON_IS_IGP)) {
+		snprintf(fw_name, sizeof(fw_name), "radeon/%s_mc.bin", chip_name);
+		err = request_firmware(&rdev->mc_fw, fw_name, &pdev->dev);
+		if (err)
+			goto out;
+		if (rdev->mc_fw->size != mc_req_size) {
+			printk(KERN_ERR
+			       "ni_mc: Bogus length %zu in firmware \"%s\"\n",
+			       rdev->mc_fw->size, fw_name);
+			err = -EINVAL;
+		}
 	}
 out:
 	platform_device_unregister(pdev);
@@ -478,6 +497,7 @@ static u32 cayman_get_tile_pipe_to_backend_map(struct radeon_device *rdev,
 	memset((uint8_t *)&swizzle_pipe[0], 0, sizeof(u32) * CAYMAN_MAX_PIPES);
 	switch (rdev->family) {
 	case CHIP_CAYMAN:
+	case CHIP_ARUBA:
 		force_no_swizzle = true;
 		break;
 	default:
@@ -610,7 +630,6 @@ static void cayman_gpu_init(struct radeon_device *rdev)
 
 	switch (rdev->family) {
 	case CHIP_CAYMAN:
-	default:
 		rdev->config.cayman.max_shader_engines = 2;
 		rdev->config.cayman.max_pipes_per_simd = 4;
 		rdev->config.cayman.max_tile_pipes = 8;
@@ -632,6 +651,43 @@ static void cayman_gpu_init(struct radeon_device *rdev)
 		rdev->config.cayman.sc_hiz_tile_fifo_size = 0x30;
 		rdev->config.cayman.sc_earlyz_tile_fifo_size = 0x130;
 		break;
+	case CHIP_ARUBA:
+	default:
+		rdev->config.cayman.max_shader_engines = 1;
+		rdev->config.cayman.max_pipes_per_simd = 4;
+		rdev->config.cayman.max_tile_pipes = 2;
+		if ((rdev->pdev->device == 0x9900) ||
+		    (rdev->pdev->device == 0x9901)) {
+			rdev->config.cayman.max_simds_per_se = 6;
+			rdev->config.cayman.max_backends_per_se = 2;
+		} else if ((rdev->pdev->device == 0x9903) ||
+			   (rdev->pdev->device == 0x9904)) {
+			rdev->config.cayman.max_simds_per_se = 4;
+			rdev->config.cayman.max_backends_per_se = 2;
+		} else if ((rdev->pdev->device == 0x9990) ||
+			   (rdev->pdev->device == 0x9991)) {
+			rdev->config.cayman.max_simds_per_se = 3;
+			rdev->config.cayman.max_backends_per_se = 1;
+		} else {
+			rdev->config.cayman.max_simds_per_se = 2;
+			rdev->config.cayman.max_backends_per_se = 1;
+		}
+		rdev->config.cayman.max_texture_channel_caches = 2;
+		rdev->config.cayman.max_gprs = 256;
+		rdev->config.cayman.max_threads = 256;
+		rdev->config.cayman.max_gs_threads = 32;
+		rdev->config.cayman.max_stack_entries = 512;
+		rdev->config.cayman.sx_num_of_sets = 8;
+		rdev->config.cayman.sx_max_export_size = 256;
+		rdev->config.cayman.sx_max_export_pos_size = 64;
+		rdev->config.cayman.sx_max_export_smx_size = 192;
+		rdev->config.cayman.max_hw_contexts = 8;
+		rdev->config.cayman.sq_num_cf_insts = 2;
+
+		rdev->config.cayman.sc_prim_fifo_size = 0x40;
+		rdev->config.cayman.sc_hiz_tile_fifo_size = 0x30;
+		rdev->config.cayman.sc_earlyz_tile_fifo_size = 0x130;
+		break;
 	}
 
 	/* Initialize HDP */
@@ -652,7 +708,9 @@ static void cayman_gpu_init(struct radeon_device *rdev)
 
 	cc_rb_backend_disable = RREG32(CC_RB_BACKEND_DISABLE);
 	cc_gc_shader_pipe_config = RREG32(CC_GC_SHADER_PIPE_CONFIG);
-	cgts_tcc_disable = 0xff000000;
+	cgts_tcc_disable = 0xffff0000;
+	for (i = 0; i < rdev->config.cayman.max_texture_channel_caches; i++)
+		cgts_tcc_disable &= ~(1 << (16 + i));
 	gc_user_rb_backend_disable = RREG32(GC_USER_RB_BACKEND_DISABLE);
 	gc_user_shader_pipe_config = RREG32(GC_USER_SHADER_PIPE_CONFIG);
 	cgts_user_tcc_disable = RREG32(CGTS_USER_TCC_DISABLE);
@@ -804,8 +862,13 @@ static void cayman_gpu_init(struct radeon_device *rdev)
 		rdev->config.cayman.tile_config |= (3 << 0);
 		break;
 	}
-	rdev->config.cayman.tile_config |=
-		((mc_arb_ramcfg & NOOFBANK_MASK) >> NOOFBANK_SHIFT) << 4;
+
+	/* num banks is 8 on all fusion asics. 0 = 4, 1 = 8, 2 = 16 */
+	if (rdev->flags & RADEON_IS_IGP)
+		rdev->config.evergreen.tile_config |= 1 << 4;
+	else
+		rdev->config.cayman.tile_config |=
+			((mc_arb_ramcfg & NOOFBANK_MASK) >> NOOFBANK_SHIFT) << 4;
 	rdev->config.cayman.tile_config |=
 		((gb_addr_config & PIPE_INTERLEAVE_SIZE_MASK) >> PIPE_INTERLEAVE_SIZE_SHIFT) << 8;
 	rdev->config.cayman.tile_config |=
@@ -1318,7 +1381,7 @@ int cayman_cp_resume(struct radeon_device *rdev)
 	rdev->ring[CAYMAN_RING_TYPE_CP1_INDEX].ready = false;
 	rdev->ring[CAYMAN_RING_TYPE_CP2_INDEX].ready = false;
 	/* this only test cp0 */
-	r = radeon_ring_test(rdev, &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]);
+	r = radeon_ring_test(rdev, RADEON_RING_TYPE_GFX_INDEX, &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]);
 	if (r) {
 		rdev->ring[RADEON_RING_TYPE_GFX_INDEX].ready = false;
 		rdev->ring[CAYMAN_RING_TYPE_CP1_INDEX].ready = false;
@@ -1440,18 +1503,29 @@ static int cayman_startup(struct radeon_device *rdev)
 	/* enable pcie gen2 link */
 	evergreen_pcie_gen2_enable(rdev);
 
-	if (!rdev->me_fw || !rdev->pfp_fw || !rdev->rlc_fw || !rdev->mc_fw) {
-		r = ni_init_microcode(rdev);
+	if (rdev->flags & RADEON_IS_IGP) {
+		if (!rdev->me_fw || !rdev->pfp_fw || !rdev->rlc_fw) {
+			r = ni_init_microcode(rdev);
+			if (r) {
+				DRM_ERROR("Failed to load firmware!\n");
+				return r;
+			}
+		}
+	} else {
+		if (!rdev->me_fw || !rdev->pfp_fw || !rdev->rlc_fw || !rdev->mc_fw) {
+			r = ni_init_microcode(rdev);
+			if (r) {
+				DRM_ERROR("Failed to load firmware!\n");
+				return r;
+			}
+		}
+
+		r = ni_mc_load_microcode(rdev);
 		if (r) {
-			DRM_ERROR("Failed to load firmware!\n");
+			DRM_ERROR("Failed to load MC firmware!\n");
 			return r;
 		}
 	}
-	r = ni_mc_load_microcode(rdev);
-	if (r) {
-		DRM_ERROR("Failed to load MC firmware!\n");
-		return r;
-	}
 
 	r = r600_vram_scratch_init(rdev);
 	if (r)
@@ -1466,10 +1540,19 @@ static int cayman_startup(struct radeon_device *rdev)
 	r = evergreen_blit_init(rdev);
 	if (r) {
 		r600_blit_fini(rdev);
-		rdev->asic->copy = NULL;
+		rdev->asic->copy.copy = NULL;
 		dev_warn(rdev->dev, "failed blitter (%d) falling back to memcpy\n", r);
 	}
 
+	/* allocate rlc buffers */
+	if (rdev->flags & RADEON_IS_IGP) {
+		r = si_rlc_init(rdev);
+		if (r) {
+			DRM_ERROR("Failed to init rlc BOs!\n");
+			return r;
+		}
+	}
+
 	/* allocate wb buffer */
 	r = radeon_wb_init(rdev);
 	if (r)
@@ -1518,7 +1601,7 @@ static int cayman_startup(struct radeon_device *rdev)
 	if (r)
 		return r;
 
-	r = r600_ib_test(rdev, RADEON_RING_TYPE_GFX_INDEX);
+	r = radeon_ib_test(rdev, RADEON_RING_TYPE_GFX_INDEX, &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]);
 	if (r) {
 		DRM_ERROR("radeon: failed testing IB (%d).\n", r);
 		rdev->accel_working = false;
@@ -1547,6 +1630,7 @@ int cayman_resume(struct radeon_device *rdev)
 	r = cayman_startup(rdev);
 	if (r) {
 		DRM_ERROR("cayman startup failed on resume\n");
+		rdev->accel_working = false;
 		return r;
 	}
 	return r;
@@ -1653,6 +1737,8 @@ int cayman_init(struct radeon_device *rdev)
 		dev_err(rdev->dev, "disabling GPU acceleration\n");
 		cayman_cp_fini(rdev);
 		r600_irq_fini(rdev);
+		if (rdev->flags & RADEON_IS_IGP)
+			si_rlc_fini(rdev);
 		radeon_wb_fini(rdev);
 		r100_ib_fini(rdev);
 		radeon_vm_manager_fini(rdev);
@@ -1664,8 +1750,11 @@ int cayman_init(struct radeon_device *rdev)
 	/* Don't start up if the MC ucode is missing.
 	 * The default clocks and voltages before the MC ucode
 	 * is loaded are not suffient for advanced operations.
+	 *
+	 * We can skip this check for TN, because there is no MC
+	 * ucode.
 	 */
-	if (!rdev->mc_fw) {
+	if (!rdev->mc_fw && !(rdev->flags & RADEON_IS_IGP)) {
 		DRM_ERROR("radeon: MC ucode required for NI+.\n");
 		return -EINVAL;
 	}
@@ -1678,6 +1767,8 @@ void cayman_fini(struct radeon_device *rdev)
 	r600_blit_fini(rdev);
 	cayman_cp_fini(rdev);
 	r600_irq_fini(rdev);
+	if (rdev->flags & RADEON_IS_IGP)
+		si_rlc_fini(rdev);
 	radeon_wb_fini(rdev);
 	radeon_vm_manager_fini(rdev);
 	r100_ib_fini(rdev);
@@ -1701,7 +1792,12 @@ int cayman_vm_init(struct radeon_device *rdev)
 	/* number of VMs */
 	rdev->vm_manager.nvm = 8;
 	/* base offset of vram pages */
-	rdev->vm_manager.vram_base_offset = 0;
+	if (rdev->flags & RADEON_IS_IGP) {
+		u64 tmp = RREG32(FUS_MC_VM_FB_OFFSET);
+		tmp <<= 22;
+		rdev->vm_manager.vram_base_offset = tmp;
+	} else
+		rdev->vm_manager.vram_base_offset = 0;
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/radeon/nid.h b/drivers/gpu/drm/radeon/nid.h
index 9a7f3b6e02de..2aa7046ada56 100644
--- a/drivers/gpu/drm/radeon/nid.h
+++ b/drivers/gpu/drm/radeon/nid.h
@@ -106,6 +106,7 @@
 #define		SYSTEM_ACCESS_MODE_NOT_IN_SYS			(3 << 3)
 #define		SYSTEM_APERTURE_UNMAPPED_ACCESS_PASS_THRU	(0 << 5)
 #define		ENABLE_ADVANCED_DRIVER_MODEL			(1 << 6)
+#define	FUS_MC_VM_FB_OFFSET				0x2068
 
 #define MC_SHARED_BLACKOUT_CNTL           		0x20ac
 #define	MC_ARB_RAMCFG					0x2760
diff --git a/drivers/gpu/drm/radeon/r100.c b/drivers/gpu/drm/radeon/r100.c
index bfd36ab643a6..81801c176aa5 100644
--- a/drivers/gpu/drm/radeon/r100.c
+++ b/drivers/gpu/drm/radeon/r100.c
@@ -65,6 +65,40 @@ MODULE_FIRMWARE(FIRMWARE_R520);
 
 #include "r100_track.h"
 
+void r100_wait_for_vblank(struct radeon_device *rdev, int crtc)
+{
+	struct radeon_crtc *radeon_crtc = rdev->mode_info.crtcs[crtc];
+	int i;
+
+	if (radeon_crtc->crtc_id == 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);
+			}
+		}
+	} 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);
+			}
+		}
+	}
+}
+
 /* This files gather functions specifics to:
  * r100,rv100,rs100,rv200,rs200,r200,rv250,rs300,rv280
  */
@@ -87,23 +121,27 @@ int r100_reloc_pitch_offset(struct radeon_cs_parser *p,
 		r100_cs_dump_packet(p, pkt);
 		return r;
 	}
+
 	value = radeon_get_ib_value(p, idx);
 	tmp = value & 0x003fffff;
 	tmp += (((u32)reloc->lobj.gpu_offset) >> 10);
 
-	if (reloc->lobj.tiling_flags & RADEON_TILING_MACRO)
-		tile_flags |= RADEON_DST_TILE_MACRO;
-	if (reloc->lobj.tiling_flags & RADEON_TILING_MICRO) {
-		if (reg == RADEON_SRC_PITCH_OFFSET) {
-			DRM_ERROR("Cannot src blit from microtiled surface\n");
-			r100_cs_dump_packet(p, pkt);
-			return -EINVAL;
+	if (!(p->cs_flags & RADEON_CS_KEEP_TILING_FLAGS)) {
+		if (reloc->lobj.tiling_flags & RADEON_TILING_MACRO)
+			tile_flags |= RADEON_DST_TILE_MACRO;
+		if (reloc->lobj.tiling_flags & RADEON_TILING_MICRO) {
+			if (reg == RADEON_SRC_PITCH_OFFSET) {
+				DRM_ERROR("Cannot src blit from microtiled surface\n");
+				r100_cs_dump_packet(p, pkt);
+				return -EINVAL;
+			}
+			tile_flags |= RADEON_DST_TILE_MICRO;
 		}
-		tile_flags |= RADEON_DST_TILE_MICRO;
-	}
 
-	tmp |= tile_flags;
-	p->ib->ptr[idx] = (value & 0x3fc00000) | tmp;
+		tmp |= tile_flags;
+		p->ib->ptr[idx] = (value & 0x3fc00000) | tmp;
+	} else
+		p->ib->ptr[idx] = (value & 0xffc00000) | tmp;
 	return 0;
 }
 
@@ -412,7 +450,7 @@ void r100_pm_misc(struct radeon_device *rdev)
 	/* set pcie lanes */
 	if ((rdev->flags & RADEON_IS_PCIE) &&
 	    !(rdev->flags & RADEON_IS_IGP) &&
-	    rdev->asic->set_pcie_lanes &&
+	    rdev->asic->pm.set_pcie_lanes &&
 	    (ps->pcie_lanes !=
 	     rdev->pm.power_state[rdev->pm.current_power_state_index].pcie_lanes)) {
 		radeon_set_pcie_lanes(rdev,
@@ -592,8 +630,8 @@ int r100_pci_gart_init(struct radeon_device *rdev)
 	if (r)
 		return r;
 	rdev->gart.table_size = rdev->gart.num_gpu_pages * 4;
-	rdev->asic->gart_tlb_flush = &r100_pci_gart_tlb_flush;
-	rdev->asic->gart_set_page = &r100_pci_gart_set_page;
+	rdev->asic->gart.tlb_flush = &r100_pci_gart_tlb_flush;
+	rdev->asic->gart.set_page = &r100_pci_gart_set_page;
 	return radeon_gart_table_ram_alloc(rdev);
 }
 
@@ -789,9 +827,7 @@ int r100_irq_process(struct radeon_device *rdev)
 			WREG32(RADEON_AIC_CNTL, msi_rearm | RS400_MSI_REARM);
 			break;
 		default:
-			msi_rearm = RREG32(RADEON_MSI_REARM_EN) & ~RV370_MSI_REARM_EN;
-			WREG32(RADEON_MSI_REARM_EN, msi_rearm);
-			WREG32(RADEON_MSI_REARM_EN, msi_rearm | RV370_MSI_REARM_EN);
+			WREG32(RADEON_MSI_REARM_EN, RV370_MSI_REARM_EN);
 			break;
 		}
 	}
@@ -932,9 +968,8 @@ static int r100_cp_wait_for_idle(struct radeon_device *rdev)
 	return -1;
 }
 
-void r100_ring_start(struct radeon_device *rdev)
+void r100_ring_start(struct radeon_device *rdev, struct radeon_ring *ring)
 {
-	struct radeon_ring *ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX];
 	int r;
 
 	r = radeon_ring_lock(rdev, ring, 2);
@@ -1145,8 +1180,8 @@ int r100_cp_init(struct radeon_device *rdev, unsigned ring_size)
 	WREG32(RADEON_CP_RB_WPTR_DELAY, 0);
 	WREG32(RADEON_CP_CSQ_MODE, 0x00004D4D);
 	WREG32(RADEON_CP_CSQ_CNTL, RADEON_CSQ_PRIBM_INDBM);
-	radeon_ring_start(rdev);
-	r = radeon_ring_test(rdev, ring);
+	radeon_ring_start(rdev, RADEON_RING_TYPE_GFX_INDEX, &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]);
+	r = radeon_ring_test(rdev, RADEON_RING_TYPE_GFX_INDEX, ring);
 	if (r) {
 		DRM_ERROR("radeon: cp isn't working (%d).\n", r);
 		return r;
@@ -1554,7 +1589,17 @@ static int r100_packet0_check(struct radeon_cs_parser *p,
 			r100_cs_dump_packet(p, pkt);
 			return r;
 		}
-		ib[idx] = idx_value + ((u32)reloc->lobj.gpu_offset);
+		if (!(p->cs_flags & RADEON_CS_KEEP_TILING_FLAGS)) {
+			if (reloc->lobj.tiling_flags & RADEON_TILING_MACRO)
+				tile_flags |= RADEON_TXO_MACRO_TILE;
+			if (reloc->lobj.tiling_flags & RADEON_TILING_MICRO)
+				tile_flags |= RADEON_TXO_MICRO_TILE_X2;
+
+			tmp = idx_value & ~(0x7 << 2);
+			tmp |= tile_flags;
+			ib[idx] = tmp + ((u32)reloc->lobj.gpu_offset);
+		} else
+			ib[idx] = idx_value + ((u32)reloc->lobj.gpu_offset);
 		track->textures[i].robj = reloc->robj;
 		track->tex_dirty = true;
 		break;
@@ -1625,15 +1670,17 @@ static int r100_packet0_check(struct radeon_cs_parser *p,
 			r100_cs_dump_packet(p, pkt);
 			return r;
 		}
-
-		if (reloc->lobj.tiling_flags & RADEON_TILING_MACRO)
-			tile_flags |= RADEON_COLOR_TILE_ENABLE;
-		if (reloc->lobj.tiling_flags & RADEON_TILING_MICRO)
-			tile_flags |= RADEON_COLOR_MICROTILE_ENABLE;
-
-		tmp = idx_value & ~(0x7 << 16);
-		tmp |= tile_flags;
-		ib[idx] = tmp;
+		if (!(p->cs_flags & RADEON_CS_KEEP_TILING_FLAGS)) {
+			if (reloc->lobj.tiling_flags & RADEON_TILING_MACRO)
+				tile_flags |= RADEON_COLOR_TILE_ENABLE;
+			if (reloc->lobj.tiling_flags & RADEON_TILING_MICRO)
+				tile_flags |= RADEON_COLOR_MICROTILE_ENABLE;
+
+			tmp = idx_value & ~(0x7 << 16);
+			tmp |= tile_flags;
+			ib[idx] = tmp;
+		} else
+			ib[idx] = idx_value;
 
 		track->cb[0].pitch = idx_value & RADEON_COLORPITCH_MASK;
 		track->cb_dirty = true;
@@ -3693,7 +3740,7 @@ void r100_ring_ib_execute(struct radeon_device *rdev, struct radeon_ib *ib)
 	radeon_ring_write(ring, ib->length_dw);
 }
 
-int r100_ib_test(struct radeon_device *rdev)
+int r100_ib_test(struct radeon_device *rdev, struct radeon_ring *ring)
 {
 	struct radeon_ib *ib;
 	uint32_t scratch;
@@ -3918,7 +3965,7 @@ static int r100_startup(struct radeon_device *rdev)
 	if (r)
 		return r;
 
-	r = r100_ib_test(rdev);
+	r = radeon_ib_test(rdev, RADEON_RING_TYPE_GFX_INDEX, &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]);
 	if (r) {
 		dev_err(rdev->dev, "failed testing IB (%d).\n", r);
 		rdev->accel_working = false;
@@ -3930,6 +3977,8 @@ static int r100_startup(struct radeon_device *rdev)
 
 int r100_resume(struct radeon_device *rdev)
 {
+	int r;
+
 	/* Make sur GART are not working */
 	if (rdev->flags & RADEON_IS_PCI)
 		r100_pci_gart_disable(rdev);
@@ -3949,7 +3998,11 @@ int r100_resume(struct radeon_device *rdev)
 	radeon_surface_init(rdev);
 
 	rdev->accel_working = true;
-	return r100_startup(rdev);
+	r = r100_startup(rdev);
+	if (r) {
+		rdev->accel_working = false;
+	}
+	return r;
 }
 
 int r100_suspend(struct radeon_device *rdev)
diff --git a/drivers/gpu/drm/radeon/r200.c b/drivers/gpu/drm/radeon/r200.c
index eba4cbfa78f6..a59cc474d537 100644
--- a/drivers/gpu/drm/radeon/r200.c
+++ b/drivers/gpu/drm/radeon/r200.c
@@ -215,7 +215,17 @@ int r200_packet0_check(struct radeon_cs_parser *p,
 			r100_cs_dump_packet(p, pkt);
 			return r;
 		}
-		ib[idx] = idx_value + ((u32)reloc->lobj.gpu_offset);
+		if (!(p->cs_flags & RADEON_CS_KEEP_TILING_FLAGS)) {
+			if (reloc->lobj.tiling_flags & RADEON_TILING_MACRO)
+				tile_flags |= R200_TXO_MACRO_TILE;
+			if (reloc->lobj.tiling_flags & RADEON_TILING_MICRO)
+				tile_flags |= R200_TXO_MICRO_TILE;
+
+			tmp = idx_value & ~(0x7 << 2);
+			tmp |= tile_flags;
+			ib[idx] = tmp + ((u32)reloc->lobj.gpu_offset);
+		} else
+			ib[idx] = idx_value + ((u32)reloc->lobj.gpu_offset);
 		track->textures[i].robj = reloc->robj;
 		track->tex_dirty = true;
 		break;
@@ -277,14 +287,17 @@ int r200_packet0_check(struct radeon_cs_parser *p,
 			return r;
 		}
 
-		if (reloc->lobj.tiling_flags & RADEON_TILING_MACRO)
-			tile_flags |= RADEON_COLOR_TILE_ENABLE;
-		if (reloc->lobj.tiling_flags & RADEON_TILING_MICRO)
-			tile_flags |= RADEON_COLOR_MICROTILE_ENABLE;
+		if (!(p->cs_flags & RADEON_CS_KEEP_TILING_FLAGS)) {
+			if (reloc->lobj.tiling_flags & RADEON_TILING_MACRO)
+				tile_flags |= RADEON_COLOR_TILE_ENABLE;
+			if (reloc->lobj.tiling_flags & RADEON_TILING_MICRO)
+				tile_flags |= RADEON_COLOR_MICROTILE_ENABLE;
 
-		tmp = idx_value & ~(0x7 << 16);
-		tmp |= tile_flags;
-		ib[idx] = tmp;
+			tmp = idx_value & ~(0x7 << 16);
+			tmp |= tile_flags;
+			ib[idx] = tmp;
+		} else
+			ib[idx] = idx_value;
 
 		track->cb[0].pitch = idx_value & RADEON_COLORPITCH_MASK;
 		track->cb_dirty = true;
diff --git a/drivers/gpu/drm/radeon/r300.c b/drivers/gpu/drm/radeon/r300.c
index 3fc0d29a5f39..fa14383f9ca0 100644
--- a/drivers/gpu/drm/radeon/r300.c
+++ b/drivers/gpu/drm/radeon/r300.c
@@ -105,8 +105,8 @@ int rv370_pcie_gart_init(struct radeon_device *rdev)
 	if (r)
 		DRM_ERROR("Failed to register debugfs file for PCIE gart !\n");
 	rdev->gart.table_size = rdev->gart.num_gpu_pages * 4;
-	rdev->asic->gart_tlb_flush = &rv370_pcie_gart_tlb_flush;
-	rdev->asic->gart_set_page = &rv370_pcie_gart_set_page;
+	rdev->asic->gart.tlb_flush = &rv370_pcie_gart_tlb_flush;
+	rdev->asic->gart.set_page = &rv370_pcie_gart_set_page;
 	return radeon_gart_table_vram_alloc(rdev);
 }
 
@@ -206,9 +206,8 @@ void r300_fence_ring_emit(struct radeon_device *rdev,
 	radeon_ring_write(ring, RADEON_SW_INT_FIRE);
 }
 
-void r300_ring_start(struct radeon_device *rdev)
+void r300_ring_start(struct radeon_device *rdev, struct radeon_ring *ring)
 {
-	struct radeon_ring *ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX];
 	unsigned gb_tile_config;
 	int r;
 
@@ -1419,7 +1418,7 @@ static int r300_startup(struct radeon_device *rdev)
 	if (r)
 		return r;
 
-	r = r100_ib_test(rdev);
+	r = radeon_ib_test(rdev, RADEON_RING_TYPE_GFX_INDEX, &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]);
 	if (r) {
 		dev_err(rdev->dev, "failed testing IB (%d).\n", r);
 		rdev->accel_working = false;
@@ -1431,6 +1430,8 @@ static int r300_startup(struct radeon_device *rdev)
 
 int r300_resume(struct radeon_device *rdev)
 {
+	int r;
+
 	/* Make sur GART are not working */
 	if (rdev->flags & RADEON_IS_PCIE)
 		rv370_pcie_gart_disable(rdev);
@@ -1452,7 +1453,11 @@ int r300_resume(struct radeon_device *rdev)
 	radeon_surface_init(rdev);
 
 	rdev->accel_working = true;
-	return r300_startup(rdev);
+	r = r300_startup(rdev);
+	if (r) {
+		rdev->accel_working = false;
+	}
+	return r;
 }
 
 int r300_suspend(struct radeon_device *rdev)
diff --git a/drivers/gpu/drm/radeon/r420.c b/drivers/gpu/drm/radeon/r420.c
index 666e28fe509c..f3fcaacfea01 100644
--- a/drivers/gpu/drm/radeon/r420.c
+++ b/drivers/gpu/drm/radeon/r420.c
@@ -279,7 +279,7 @@ static int r420_startup(struct radeon_device *rdev)
 	if (r)
 		return r;
 
-	r = r100_ib_test(rdev);
+	r = radeon_ib_test(rdev, RADEON_RING_TYPE_GFX_INDEX, &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]);
 	if (r) {
 		dev_err(rdev->dev, "failed testing IB (%d).\n", r);
 		rdev->accel_working = false;
@@ -291,6 +291,8 @@ static int r420_startup(struct radeon_device *rdev)
 
 int r420_resume(struct radeon_device *rdev)
 {
+	int r;
+
 	/* Make sur GART are not working */
 	if (rdev->flags & RADEON_IS_PCIE)
 		rv370_pcie_gart_disable(rdev);
@@ -316,7 +318,11 @@ int r420_resume(struct radeon_device *rdev)
 	radeon_surface_init(rdev);
 
 	rdev->accel_working = true;
-	return r420_startup(rdev);
+	r = r420_startup(rdev);
+	if (r) {
+		rdev->accel_working = false;
+	}
+	return r;
 }
 
 int r420_suspend(struct radeon_device *rdev)
diff --git a/drivers/gpu/drm/radeon/r500_reg.h b/drivers/gpu/drm/radeon/r500_reg.h
index 3bd8f1b1c606..ec576aaafb73 100644
--- a/drivers/gpu/drm/radeon/r500_reg.h
+++ b/drivers/gpu/drm/radeon/r500_reg.h
@@ -351,6 +351,8 @@
 #define AVIVO_D1CRTC_BLANK_CONTROL                              0x6084
 #define AVIVO_D1CRTC_INTERLACE_CONTROL                          0x6088
 #define AVIVO_D1CRTC_INTERLACE_STATUS                           0x608c
+#define AVIVO_D1CRTC_STATUS                                     0x609c
+#       define AVIVO_D1CRTC_V_BLANK                             (1 << 0)
 #define AVIVO_D1CRTC_STATUS_POSITION                            0x60a0
 #define AVIVO_D1CRTC_FRAME_COUNT                                0x60a4
 #define AVIVO_D1CRTC_STEREO_CONTROL                             0x60c4
diff --git a/drivers/gpu/drm/radeon/r520.c b/drivers/gpu/drm/radeon/r520.c
index 4ae1615e752f..ebcc15b03c9f 100644
--- a/drivers/gpu/drm/radeon/r520.c
+++ b/drivers/gpu/drm/radeon/r520.c
@@ -33,7 +33,7 @@
 
 /* This files gather functions specifics to: r520,rv530,rv560,rv570,r580 */
 
-static int r520_mc_wait_for_idle(struct radeon_device *rdev)
+int r520_mc_wait_for_idle(struct radeon_device *rdev)
 {
 	unsigned i;
 	uint32_t tmp;
@@ -207,7 +207,7 @@ static int r520_startup(struct radeon_device *rdev)
 	if (r)
 		return r;
 
-	r = r100_ib_test(rdev);
+	r = radeon_ib_test(rdev, RADEON_RING_TYPE_GFX_INDEX, &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]);
 	if (r) {
 		dev_err(rdev->dev, "failed testing IB (%d).\n", r);
 		rdev->accel_working = false;
@@ -218,6 +218,8 @@ static int r520_startup(struct radeon_device *rdev)
 
 int r520_resume(struct radeon_device *rdev)
 {
+	int r;
+
 	/* Make sur GART are not working */
 	if (rdev->flags & RADEON_IS_PCIE)
 		rv370_pcie_gart_disable(rdev);
@@ -237,7 +239,11 @@ int r520_resume(struct radeon_device *rdev)
 	radeon_surface_init(rdev);
 
 	rdev->accel_working = true;
-	return r520_startup(rdev);
+	r = r520_startup(rdev);
+	if (r) {
+		rdev->accel_working = false;
+	}
+	return r;
 }
 
 int r520_init(struct radeon_device *rdev)
diff --git a/drivers/gpu/drm/radeon/r600.c b/drivers/gpu/drm/radeon/r600.c
index 4f08e5e6ee9d..391bd2636a80 100644
--- a/drivers/gpu/drm/radeon/r600.c
+++ b/drivers/gpu/drm/radeon/r600.c
@@ -49,6 +49,7 @@
 #define EVERGREEN_PM4_UCODE_SIZE 1376
 #define EVERGREEN_RLC_UCODE_SIZE 768
 #define CAYMAN_RLC_UCODE_SIZE 1024
+#define ARUBA_RLC_UCODE_SIZE 1536
 
 /* Firmware Names */
 MODULE_FIRMWARE("radeon/R600_pfp.bin");
@@ -2226,7 +2227,7 @@ int r600_cp_resume(struct radeon_device *rdev)
 
 	r600_cp_start(rdev);
 	ring->ready = true;
-	r = radeon_ring_test(rdev, ring);
+	r = radeon_ring_test(rdev, RADEON_RING_TYPE_GFX_INDEX, ring);
 	if (r) {
 		ring->ready = false;
 		return r;
@@ -2362,6 +2363,9 @@ void r600_semaphore_ring_emit(struct radeon_device *rdev,
 	uint64_t addr = semaphore->gpu_addr;
 	unsigned sel = emit_wait ? PACKET3_SEM_SEL_WAIT : PACKET3_SEM_SEL_SIGNAL;
 
+	if (rdev->family < CHIP_CAYMAN)
+		sel |= PACKET3_SEM_WAIT_ON_SIGNAL;
+
 	radeon_ring_write(ring, PACKET3(PACKET3_MEM_SEMAPHORE, 1));
 	radeon_ring_write(ring, addr & 0xffffffff);
 	radeon_ring_write(ring, (upper_32_bits(addr) & 0xff) | sel);
@@ -2449,7 +2453,7 @@ int r600_startup(struct radeon_device *rdev)
 	r = r600_blit_init(rdev);
 	if (r) {
 		r600_blit_fini(rdev);
-		rdev->asic->copy = NULL;
+		rdev->asic->copy.copy = NULL;
 		dev_warn(rdev->dev, "failed blitter (%d) falling back to memcpy\n", r);
 	}
 
@@ -2490,7 +2494,7 @@ int r600_startup(struct radeon_device *rdev)
 	if (r)
 		return r;
 
-	r = r600_ib_test(rdev, RADEON_RING_TYPE_GFX_INDEX);
+	r = radeon_ib_test(rdev, RADEON_RING_TYPE_GFX_INDEX, &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]);
 	if (r) {
 		DRM_ERROR("radeon: failed testing IB (%d).\n", r);
 		rdev->accel_working = false;
@@ -2529,6 +2533,7 @@ int r600_resume(struct radeon_device *rdev)
 	r = r600_startup(rdev);
 	if (r) {
 		DRM_ERROR("r600 startup failed on resume\n");
+		rdev->accel_working = false;
 		return r;
 	}
 
@@ -2697,13 +2702,14 @@ void r600_ring_ib_execute(struct radeon_device *rdev, struct radeon_ib *ib)
 	radeon_ring_write(ring, ib->length_dw);
 }
 
-int r600_ib_test(struct radeon_device *rdev, int ring)
+int r600_ib_test(struct radeon_device *rdev, struct radeon_ring *ring)
 {
 	struct radeon_ib *ib;
 	uint32_t scratch;
 	uint32_t tmp = 0;
 	unsigned i;
 	int r;
+	int ring_index = radeon_ring_index(rdev, ring);
 
 	r = radeon_scratch_get(rdev, &scratch);
 	if (r) {
@@ -2711,7 +2717,7 @@ int r600_ib_test(struct radeon_device *rdev, int ring)
 		return r;
 	}
 	WREG32(scratch, 0xCAFEDEAD);
-	r = radeon_ib_get(rdev, ring, &ib, 256);
+	r = radeon_ib_get(rdev, ring_index, &ib, 256);
 	if (r) {
 		DRM_ERROR("radeon: failed to get ib (%d).\n", r);
 		return r;
@@ -2719,20 +2725,7 @@ int r600_ib_test(struct radeon_device *rdev, int ring)
 	ib->ptr[0] = PACKET3(PACKET3_SET_CONFIG_REG, 1);
 	ib->ptr[1] = ((scratch - PACKET3_SET_CONFIG_REG_OFFSET) >> 2);
 	ib->ptr[2] = 0xDEADBEEF;
-	ib->ptr[3] = PACKET2(0);
-	ib->ptr[4] = PACKET2(0);
-	ib->ptr[5] = PACKET2(0);
-	ib->ptr[6] = PACKET2(0);
-	ib->ptr[7] = PACKET2(0);
-	ib->ptr[8] = PACKET2(0);
-	ib->ptr[9] = PACKET2(0);
-	ib->ptr[10] = PACKET2(0);
-	ib->ptr[11] = PACKET2(0);
-	ib->ptr[12] = PACKET2(0);
-	ib->ptr[13] = PACKET2(0);
-	ib->ptr[14] = PACKET2(0);
-	ib->ptr[15] = PACKET2(0);
-	ib->length_dw = 16;
+	ib->length_dw = 3;
 	r = radeon_ib_schedule(rdev, ib);
 	if (r) {
 		radeon_scratch_free(rdev, scratch);
@@ -2786,7 +2779,7 @@ void r600_ih_ring_init(struct radeon_device *rdev, unsigned ring_size)
 	rdev->ih.rptr = 0;
 }
 
-static int r600_ih_ring_alloc(struct radeon_device *rdev)
+int r600_ih_ring_alloc(struct radeon_device *rdev)
 {
 	int r;
 
@@ -2822,7 +2815,7 @@ static int r600_ih_ring_alloc(struct radeon_device *rdev)
 	return 0;
 }
 
-static void r600_ih_ring_fini(struct radeon_device *rdev)
+void r600_ih_ring_fini(struct radeon_device *rdev)
 {
 	int r;
 	if (rdev->ih.ring_obj) {
@@ -2869,10 +2862,17 @@ static int r600_rlc_init(struct radeon_device *rdev)
 
 	r600_rlc_stop(rdev);
 
-	WREG32(RLC_HB_BASE, 0);
 	WREG32(RLC_HB_CNTL, 0);
-	WREG32(RLC_HB_RPTR, 0);
-	WREG32(RLC_HB_WPTR, 0);
+
+	if (rdev->family == CHIP_ARUBA) {
+		WREG32(TN_RLC_SAVE_AND_RESTORE_BASE, rdev->rlc.save_restore_gpu_addr >> 8);
+		WREG32(TN_RLC_CLEAR_STATE_RESTORE_BASE, rdev->rlc.clear_state_gpu_addr >> 8);
+	}
+	if (rdev->family <= CHIP_CAYMAN) {
+		WREG32(RLC_HB_BASE, 0);
+		WREG32(RLC_HB_RPTR, 0);
+		WREG32(RLC_HB_WPTR, 0);
+	}
 	if (rdev->family <= CHIP_CAICOS) {
 		WREG32(RLC_HB_WPTR_LSB_ADDR, 0);
 		WREG32(RLC_HB_WPTR_MSB_ADDR, 0);
@@ -2881,7 +2881,12 @@ static int r600_rlc_init(struct radeon_device *rdev)
 	WREG32(RLC_UCODE_CNTL, 0);
 
 	fw_data = (const __be32 *)rdev->rlc_fw->data;
-	if (rdev->family >= CHIP_CAYMAN) {
+	if (rdev->family >= CHIP_ARUBA) {
+		for (i = 0; i < ARUBA_RLC_UCODE_SIZE; i++) {
+			WREG32(RLC_UCODE_ADDR, i);
+			WREG32(RLC_UCODE_DATA, be32_to_cpup(fw_data++));
+		}
+	} else if (rdev->family >= CHIP_CAYMAN) {
 		for (i = 0; i < CAYMAN_RLC_UCODE_SIZE; i++) {
 			WREG32(RLC_UCODE_ADDR, i);
 			WREG32(RLC_UCODE_DATA, be32_to_cpup(fw_data++));
diff --git a/drivers/gpu/drm/radeon/r600_blit_kms.c b/drivers/gpu/drm/radeon/r600_blit_kms.c
index accc032c103f..db38f587f27a 100644
--- a/drivers/gpu/drm/radeon/r600_blit_kms.c
+++ b/drivers/gpu/drm/radeon/r600_blit_kms.c
@@ -30,20 +30,7 @@
 
 #include "r600d.h"
 #include "r600_blit_shaders.h"
-
-#define DI_PT_RECTLIST        0x11
-#define DI_INDEX_SIZE_16_BIT  0x0
-#define DI_SRC_SEL_AUTO_INDEX 0x2
-
-#define FMT_8                 0x1
-#define FMT_5_6_5             0x8
-#define FMT_8_8_8_8           0x1a
-#define COLOR_8               0x1
-#define COLOR_5_6_5           0x8
-#define COLOR_8_8_8_8         0x1a
-
-#define RECT_UNIT_H           32
-#define RECT_UNIT_W           (RADEON_GPU_PAGE_SIZE / 4 / RECT_UNIT_H)
+#include "radeon_blit_common.h"
 
 /* emits 21 on rv770+, 23 on r600 */
 static void
diff --git a/drivers/gpu/drm/radeon/r600_blit_shaders.c b/drivers/gpu/drm/radeon/r600_blit_shaders.c
index 2d1f6c5ee2a7..34c8b2340f33 100644
--- a/drivers/gpu/drm/radeon/r600_blit_shaders.c
+++ b/drivers/gpu/drm/radeon/r600_blit_shaders.c
@@ -24,6 +24,7 @@
  *     Alex Deucher <alexander.deucher@amd.com>
  */
 
+#include <linux/bug.h>
 #include <linux/types.h>
 #include <linux/kernel.h>
 
@@ -314,6 +315,10 @@ const u32 r6xx_default_state[] =
 	0x00000000, /* VGT_VTX_CNT_EN */
 
 	0xc0016900,
+	0x000000d4,
+	0x00000000, /* SX_MISC */
+
+	0xc0016900,
 	0x000002c8,
 	0x00000000, /* VGT_STRMOUT_BUFFER_EN */
 
@@ -626,6 +631,10 @@ const u32 r7xx_default_state[] =
 	0x00000000, /* VGT_VTX_CNT_EN */
 
 	0xc0016900,
+	0x000000d4,
+	0x00000000, /* SX_MISC */
+
+	0xc0016900,
 	0x000002c8,
 	0x00000000, /* VGT_STRMOUT_BUFFER_EN */
 
diff --git a/drivers/gpu/drm/radeon/r600_cs.c b/drivers/gpu/drm/radeon/r600_cs.c
index 38ce5d0427e3..b8e12af304a9 100644
--- a/drivers/gpu/drm/radeon/r600_cs.c
+++ b/drivers/gpu/drm/radeon/r600_cs.c
@@ -52,15 +52,20 @@ struct r600_cs_track {
 	struct radeon_bo	*cb_color_bo[8];
 	u64			cb_color_bo_mc[8];
 	u32			cb_color_bo_offset[8];
-	struct radeon_bo	*cb_color_frag_bo[8];
-	struct radeon_bo	*cb_color_tile_bo[8];
+	struct radeon_bo	*cb_color_frag_bo[8]; /* unused */
+	struct radeon_bo	*cb_color_tile_bo[8]; /* unused */
 	u32			cb_color_info[8];
-	u32			cb_color_size_idx[8];
+	u32			cb_color_view[8];
+	u32			cb_color_size_idx[8]; /* unused */
 	u32			cb_target_mask;
-	u32			cb_shader_mask;
+	u32			cb_shader_mask;  /* unused */
 	u32			cb_color_size[8];
 	u32			vgt_strmout_en;
 	u32			vgt_strmout_buffer_en;
+	struct radeon_bo	*vgt_strmout_bo[4];
+	u64			vgt_strmout_bo_mc[4]; /* unused */
+	u32			vgt_strmout_bo_offset[4];
+	u32			vgt_strmout_size[4];
 	u32			db_depth_control;
 	u32			db_depth_info;
 	u32			db_depth_size_idx;
@@ -69,13 +74,20 @@ struct r600_cs_track {
 	u32			db_offset;
 	struct radeon_bo	*db_bo;
 	u64			db_bo_mc;
+	bool			sx_misc_kill_all_prims;
+	bool			cb_dirty;
+	bool			db_dirty;
+	bool			streamout_dirty;
+	struct radeon_bo	*htile_bo;
+	u64			htile_offset;
+	u32			htile_surface;
 };
 
 #define FMT_8_BIT(fmt, vc)   [fmt] = { 1, 1, 1, vc, CHIP_R600 }
 #define FMT_16_BIT(fmt, vc)  [fmt] = { 1, 1, 2, vc, CHIP_R600 }
-#define FMT_24_BIT(fmt)      [fmt] = { 1, 1, 3,  0, CHIP_R600 }
+#define FMT_24_BIT(fmt)      [fmt] = { 1, 1, 4,  0, CHIP_R600 }
 #define FMT_32_BIT(fmt, vc)  [fmt] = { 1, 1, 4, vc, CHIP_R600 }
-#define FMT_48_BIT(fmt)      [fmt] = { 1, 1, 6,  0, CHIP_R600 }
+#define FMT_48_BIT(fmt)      [fmt] = { 1, 1, 8,  0, CHIP_R600 }
 #define FMT_64_BIT(fmt, vc)  [fmt] = { 1, 1, 8, vc, CHIP_R600 }
 #define FMT_96_BIT(fmt)      [fmt] = { 1, 1, 12, 0, CHIP_R600 }
 #define FMT_128_BIT(fmt, vc) [fmt] = { 1, 1, 16,vc, CHIP_R600 }
@@ -107,7 +119,7 @@ static const struct gpu_formats color_formats_table[] = {
 
 	/* 24-bit */
 	FMT_24_BIT(V_038004_FMT_8_8_8),
-					       
+
 	/* 32-bit */
 	FMT_32_BIT(V_038004_COLOR_32, 1),
 	FMT_32_BIT(V_038004_COLOR_32_FLOAT, 1),
@@ -162,22 +174,22 @@ static const struct gpu_formats color_formats_table[] = {
 	[V_038004_FMT_32_AS_32_32_32_32] = { 1, 1, 4, 0, CHIP_CEDAR},
 };
 
-static bool fmt_is_valid_color(u32 format)
+bool r600_fmt_is_valid_color(u32 format)
 {
 	if (format >= ARRAY_SIZE(color_formats_table))
 		return false;
-	
+
 	if (color_formats_table[format].valid_color)
 		return true;
 
 	return false;
 }
 
-static bool fmt_is_valid_texture(u32 format, enum radeon_family family)
+bool r600_fmt_is_valid_texture(u32 format, enum radeon_family family)
 {
 	if (format >= ARRAY_SIZE(color_formats_table))
 		return false;
-	
+
 	if (family < color_formats_table[format].min_family)
 		return false;
 
@@ -187,7 +199,7 @@ static bool fmt_is_valid_texture(u32 format, enum radeon_family family)
 	return false;
 }
 
-static int fmt_get_blocksize(u32 format)
+int r600_fmt_get_blocksize(u32 format)
 {
 	if (format >= ARRAY_SIZE(color_formats_table))
 		return 0;
@@ -195,7 +207,7 @@ static int fmt_get_blocksize(u32 format)
 	return color_formats_table[format].blocksize;
 }
 
-static int fmt_get_nblocksx(u32 format, u32 w)
+int r600_fmt_get_nblocksx(u32 format, u32 w)
 {
 	unsigned bw;
 
@@ -209,7 +221,7 @@ static int fmt_get_nblocksx(u32 format, u32 w)
 	return (w + bw - 1) / bw;
 }
 
-static int fmt_get_nblocksy(u32 format, u32 h)
+int r600_fmt_get_nblocksy(u32 format, u32 h)
 {
 	unsigned bh;
 
@@ -256,7 +268,7 @@ static int r600_get_array_mode_alignment(struct array_mode_checker *values,
 		break;
 	case ARRAY_LINEAR_ALIGNED:
 		*pitch_align = max((u32)64, (u32)(values->group_size / values->blocksize));
-		*height_align = tile_height;
+		*height_align = 1;
 		*depth_align = 1;
 		*base_align = values->group_size;
 		break;
@@ -269,10 +281,9 @@ static int r600_get_array_mode_alignment(struct array_mode_checker *values,
 		*base_align = values->group_size;
 		break;
 	case ARRAY_2D_TILED_THIN1:
-		*pitch_align = max((u32)macro_tile_width,
-				  (u32)(((values->group_size / tile_height) /
-					 (values->blocksize * values->nsamples)) *
-					values->nbanks)) * tile_width;
+		*pitch_align = max((u32)macro_tile_width * tile_width,
+				(u32)((values->group_size * values->nbanks) /
+				(values->blocksize * values->nsamples * tile_width)));
 		*height_align = macro_tile_height * tile_height;
 		*depth_align = 1;
 		*base_align = max(macro_tile_bytes,
@@ -296,12 +307,14 @@ static void r600_cs_track_init(struct r600_cs_track *track)
 		track->cb_color_size[i] = 0;
 		track->cb_color_size_idx[i] = 0;
 		track->cb_color_info[i] = 0;
+		track->cb_color_view[i] = 0xFFFFFFFF;
 		track->cb_color_bo[i] = NULL;
 		track->cb_color_bo_offset[i] = 0xFFFFFFFF;
 		track->cb_color_bo_mc[i] = 0xFFFFFFFF;
 	}
 	track->cb_target_mask = 0xFFFFFFFF;
 	track->cb_shader_mask = 0xFFFFFFFF;
+	track->cb_dirty = true;
 	track->db_bo = NULL;
 	track->db_bo_mc = 0xFFFFFFFF;
 	/* assume the biggest format and that htile is enabled */
@@ -310,6 +323,19 @@ static void r600_cs_track_init(struct r600_cs_track *track)
 	track->db_depth_size = 0xFFFFFFFF;
 	track->db_depth_size_idx = 0;
 	track->db_depth_control = 0xFFFFFFFF;
+	track->db_dirty = true;
+	track->htile_bo = NULL;
+	track->htile_offset = 0xFFFFFFFF;
+	track->htile_surface = 0;
+
+	for (i = 0; i < 4; i++) {
+		track->vgt_strmout_size[i] = 0;
+		track->vgt_strmout_bo[i] = NULL;
+		track->vgt_strmout_bo_offset[i] = 0xFFFFFFFF;
+		track->vgt_strmout_bo_mc[i] = 0xFFFFFFFF;
+	}
+	track->streamout_dirty = true;
+	track->sx_misc_kill_all_prims = false;
 }
 
 static int r600_cs_track_validate_cb(struct radeon_cs_parser *p, int i)
@@ -322,13 +348,14 @@ static int r600_cs_track_validate_cb(struct radeon_cs_parser *p, int i)
 	volatile u32 *ib = p->ib->ptr;
 	unsigned array_mode;
 	u32 format;
+
 	if (G_0280A0_TILE_MODE(track->cb_color_info[i])) {
 		dev_warn(p->dev, "FMASK or CMASK buffer are not supported by this kernel\n");
 		return -EINVAL;
 	}
 	size = radeon_bo_size(track->cb_color_bo[i]) - track->cb_color_bo_offset[i];
 	format = G_0280A0_FORMAT(track->cb_color_info[i]);
-	if (!fmt_is_valid_color(format)) {
+	if (!r600_fmt_is_valid_color(format)) {
 		dev_warn(p->dev, "%s:%d cb invalid format %d for %d (0x%08X)\n",
 			 __func__, __LINE__, format,
 			i, track->cb_color_info[i]);
@@ -349,7 +376,7 @@ static int r600_cs_track_validate_cb(struct radeon_cs_parser *p, int i)
 	array_check.nbanks = track->nbanks;
 	array_check.npipes = track->npipes;
 	array_check.nsamples = track->nsamples;
-	array_check.blocksize = fmt_get_blocksize(format);
+	array_check.blocksize = r600_fmt_get_blocksize(format);
 	if (r600_get_array_mode_alignment(&array_check,
 					  &pitch_align, &height_align, &depth_align, &base_align)) {
 		dev_warn(p->dev, "%s invalid tiling %d for %d (0x%08X)\n", __func__,
@@ -393,7 +420,18 @@ static int r600_cs_track_validate_cb(struct radeon_cs_parser *p, int i)
 	}
 
 	/* check offset */
-	tmp = fmt_get_nblocksy(format, height) * fmt_get_nblocksx(format, pitch) * fmt_get_blocksize(format);
+	tmp = r600_fmt_get_nblocksy(format, height) * r600_fmt_get_nblocksx(format, pitch) * r600_fmt_get_blocksize(format);
+	switch (array_mode) {
+	default:
+	case V_0280A0_ARRAY_LINEAR_GENERAL:
+	case V_0280A0_ARRAY_LINEAR_ALIGNED:
+		tmp += track->cb_color_view[i] & 0xFF;
+		break;
+	case V_0280A0_ARRAY_1D_TILED_THIN1:
+	case V_0280A0_ARRAY_2D_TILED_THIN1:
+		tmp += G_028080_SLICE_MAX(track->cb_color_view[i]) * tmp;
+		break;
+	}
 	if ((tmp + track->cb_color_bo_offset[i]) > radeon_bo_size(track->cb_color_bo[i])) {
 		if (array_mode == V_0280A0_ARRAY_LINEAR_GENERAL) {
 			/* the initial DDX does bad things with the CB size occasionally */
@@ -403,10 +441,13 @@ static int r600_cs_track_validate_cb(struct radeon_cs_parser *p, int i)
 			 * broken userspace.
 			 */
 		} else {
-			dev_warn(p->dev, "%s offset[%d] %d %d %d %lu too big\n", __func__, i,
-				 array_mode,
+			dev_warn(p->dev, "%s offset[%d] %d %d %d %lu too big (%d %d) (%d %d %d)\n",
+				 __func__, i, array_mode,
 				 track->cb_color_bo_offset[i], tmp,
-				 radeon_bo_size(track->cb_color_bo[i]));
+				 radeon_bo_size(track->cb_color_bo[i]),
+				 pitch, height, r600_fmt_get_nblocksx(format, pitch),
+				 r600_fmt_get_nblocksy(format, height),
+				 r600_fmt_get_blocksize(format));
 			return -EINVAL;
 		}
 	}
@@ -420,154 +461,316 @@ static int r600_cs_track_validate_cb(struct radeon_cs_parser *p, int i)
 	return 0;
 }
 
-static int r600_cs_track_check(struct radeon_cs_parser *p)
+static int r600_cs_track_validate_db(struct radeon_cs_parser *p)
 {
 	struct r600_cs_track *track = p->track;
-	u32 tmp;
-	int r, i;
+	u32 nviews, bpe, ntiles, size, slice_tile_max, tmp;
+	u32 height_align, pitch_align, depth_align;
+	u32 pitch = 8192;
+	u32 height = 8192;
+	u64 base_offset, base_align;
+	struct array_mode_checker array_check;
+	int array_mode;
 	volatile u32 *ib = p->ib->ptr;
 
-	/* on legacy kernel we don't perform advanced check */
-	if (p->rdev == NULL)
-		return 0;
-	/* we don't support out buffer yet */
-	if (track->vgt_strmout_en || track->vgt_strmout_buffer_en) {
-		dev_warn(p->dev, "this kernel doesn't support SMX output buffer\n");
+
+	if (track->db_bo == NULL) {
+		dev_warn(p->dev, "z/stencil with no depth buffer\n");
 		return -EINVAL;
 	}
-	/* check that we have a cb for each enabled target, we don't check
-	 * shader_mask because it seems mesa isn't always setting it :(
-	 */
-	tmp = track->cb_target_mask;
-	for (i = 0; i < 8; i++) {
-		if ((tmp >> (i * 4)) & 0xF) {
-			/* at least one component is enabled */
-			if (track->cb_color_bo[i] == NULL) {
-				dev_warn(p->dev, "%s:%d mask 0x%08X | 0x%08X no cb for %d\n",
-					__func__, __LINE__, track->cb_target_mask, track->cb_shader_mask, i);
-				return -EINVAL;
-			}
-			/* perform rewrite of CB_COLOR[0-7]_SIZE */
-			r = r600_cs_track_validate_cb(p, i);
-			if (r)
-				return r;
-		}
+	switch (G_028010_FORMAT(track->db_depth_info)) {
+	case V_028010_DEPTH_16:
+		bpe = 2;
+		break;
+	case V_028010_DEPTH_X8_24:
+	case V_028010_DEPTH_8_24:
+	case V_028010_DEPTH_X8_24_FLOAT:
+	case V_028010_DEPTH_8_24_FLOAT:
+	case V_028010_DEPTH_32_FLOAT:
+		bpe = 4;
+		break;
+	case V_028010_DEPTH_X24_8_32_FLOAT:
+		bpe = 8;
+		break;
+	default:
+		dev_warn(p->dev, "z/stencil with invalid format %d\n", G_028010_FORMAT(track->db_depth_info));
+		return -EINVAL;
 	}
-	/* Check depth buffer */
-	if (G_028800_STENCIL_ENABLE(track->db_depth_control) ||
-		G_028800_Z_ENABLE(track->db_depth_control)) {
-		u32 nviews, bpe, ntiles, size, slice_tile_max;
-		u32 height, height_align, pitch, pitch_align, depth_align;
-		u64 base_offset, base_align;
-		struct array_mode_checker array_check;
-		int array_mode;
-
-		if (track->db_bo == NULL) {
-			dev_warn(p->dev, "z/stencil with no depth buffer\n");
+	if ((track->db_depth_size & 0xFFFFFC00) == 0xFFFFFC00) {
+		if (!track->db_depth_size_idx) {
+			dev_warn(p->dev, "z/stencil buffer size not set\n");
 			return -EINVAL;
 		}
-		if (G_028010_TILE_SURFACE_ENABLE(track->db_depth_info)) {
-			dev_warn(p->dev, "this kernel doesn't support z/stencil htile\n");
+		tmp = radeon_bo_size(track->db_bo) - track->db_offset;
+		tmp = (tmp / bpe) >> 6;
+		if (!tmp) {
+			dev_warn(p->dev, "z/stencil buffer too small (0x%08X %d %d %ld)\n",
+					track->db_depth_size, bpe, track->db_offset,
+					radeon_bo_size(track->db_bo));
 			return -EINVAL;
 		}
-		switch (G_028010_FORMAT(track->db_depth_info)) {
-		case V_028010_DEPTH_16:
-			bpe = 2;
-			break;
-		case V_028010_DEPTH_X8_24:
-		case V_028010_DEPTH_8_24:
-		case V_028010_DEPTH_X8_24_FLOAT:
-		case V_028010_DEPTH_8_24_FLOAT:
-		case V_028010_DEPTH_32_FLOAT:
-			bpe = 4;
+		ib[track->db_depth_size_idx] = S_028000_SLICE_TILE_MAX(tmp - 1) | (track->db_depth_size & 0x3FF);
+	} else {
+		size = radeon_bo_size(track->db_bo);
+		/* pitch in pixels */
+		pitch = (G_028000_PITCH_TILE_MAX(track->db_depth_size) + 1) * 8;
+		slice_tile_max = G_028000_SLICE_TILE_MAX(track->db_depth_size) + 1;
+		slice_tile_max *= 64;
+		height = slice_tile_max / pitch;
+		if (height > 8192)
+			height = 8192;
+		base_offset = track->db_bo_mc + track->db_offset;
+		array_mode = G_028010_ARRAY_MODE(track->db_depth_info);
+		array_check.array_mode = array_mode;
+		array_check.group_size = track->group_size;
+		array_check.nbanks = track->nbanks;
+		array_check.npipes = track->npipes;
+		array_check.nsamples = track->nsamples;
+		array_check.blocksize = bpe;
+		if (r600_get_array_mode_alignment(&array_check,
+					&pitch_align, &height_align, &depth_align, &base_align)) {
+			dev_warn(p->dev, "%s invalid tiling %d (0x%08X)\n", __func__,
+					G_028010_ARRAY_MODE(track->db_depth_info),
+					track->db_depth_info);
+			return -EINVAL;
+		}
+		switch (array_mode) {
+		case V_028010_ARRAY_1D_TILED_THIN1:
+			/* don't break userspace */
+			height &= ~0x7;
 			break;
-		case V_028010_DEPTH_X24_8_32_FLOAT:
-			bpe = 8;
+		case V_028010_ARRAY_2D_TILED_THIN1:
 			break;
 		default:
-			dev_warn(p->dev, "z/stencil with invalid format %d\n", G_028010_FORMAT(track->db_depth_info));
+			dev_warn(p->dev, "%s invalid tiling %d (0x%08X)\n", __func__,
+					G_028010_ARRAY_MODE(track->db_depth_info),
+					track->db_depth_info);
+			return -EINVAL;
+		}
+
+		if (!IS_ALIGNED(pitch, pitch_align)) {
+			dev_warn(p->dev, "%s:%d db pitch (%d, 0x%x, %d) invalid\n",
+					__func__, __LINE__, pitch, pitch_align, array_mode);
+			return -EINVAL;
+		}
+		if (!IS_ALIGNED(height, height_align)) {
+			dev_warn(p->dev, "%s:%d db height (%d, 0x%x, %d) invalid\n",
+					__func__, __LINE__, height, height_align, array_mode);
+			return -EINVAL;
+		}
+		if (!IS_ALIGNED(base_offset, base_align)) {
+			dev_warn(p->dev, "%s offset 0x%llx, 0x%llx, %d not aligned\n", __func__,
+					base_offset, base_align, array_mode);
+			return -EINVAL;
+		}
+
+		ntiles = G_028000_SLICE_TILE_MAX(track->db_depth_size) + 1;
+		nviews = G_028004_SLICE_MAX(track->db_depth_view) + 1;
+		tmp = ntiles * bpe * 64 * nviews;
+		if ((tmp + track->db_offset) > radeon_bo_size(track->db_bo)) {
+			dev_warn(p->dev, "z/stencil buffer (%d) too small (0x%08X %d %d %d -> %u have %lu)\n",
+					array_mode,
+					track->db_depth_size, ntiles, nviews, bpe, tmp + track->db_offset,
+					radeon_bo_size(track->db_bo));
+			return -EINVAL;
+		}
+	}
+
+	/* hyperz */
+	if (G_028010_TILE_SURFACE_ENABLE(track->db_depth_info)) {
+		unsigned long size;
+		unsigned nbx, nby;
+
+		if (track->htile_bo == NULL) {
+			dev_warn(p->dev, "%s:%d htile enabled without htile surface 0x%08x\n",
+				 __func__, __LINE__, track->db_depth_info);
 			return -EINVAL;
 		}
 		if ((track->db_depth_size & 0xFFFFFC00) == 0xFFFFFC00) {
-			if (!track->db_depth_size_idx) {
-				dev_warn(p->dev, "z/stencil buffer size not set\n");
-				return -EINVAL;
-			}
-			tmp = radeon_bo_size(track->db_bo) - track->db_offset;
-			tmp = (tmp / bpe) >> 6;
-			if (!tmp) {
-				dev_warn(p->dev, "z/stencil buffer too small (0x%08X %d %d %ld)\n",
-						track->db_depth_size, bpe, track->db_offset,
-						radeon_bo_size(track->db_bo));
-				return -EINVAL;
-			}
-			ib[track->db_depth_size_idx] = S_028000_SLICE_TILE_MAX(tmp - 1) | (track->db_depth_size & 0x3FF);
+			dev_warn(p->dev, "%s:%d htile can't be enabled with bogus db_depth_size 0x%08x\n",
+				 __func__, __LINE__, track->db_depth_size);
+			return -EINVAL;
+		}
+
+		nbx = pitch;
+		nby = height;
+		if (G_028D24_LINEAR(track->htile_surface)) {
+			/* nbx must be 16 htiles aligned == 16 * 8 pixel aligned */
+			nbx = round_up(nbx, 16 * 8);
+			/* nby is npipes htiles aligned == npipes * 8 pixel aligned */
+			nby = round_up(nby, track->npipes * 8);
 		} else {
-			size = radeon_bo_size(track->db_bo);
-			/* pitch in pixels */
-			pitch = (G_028000_PITCH_TILE_MAX(track->db_depth_size) + 1) * 8;
-			slice_tile_max = G_028000_SLICE_TILE_MAX(track->db_depth_size) + 1;
-			slice_tile_max *= 64;
-			height = slice_tile_max / pitch;
-			if (height > 8192)
-				height = 8192;
-			base_offset = track->db_bo_mc + track->db_offset;
-			array_mode = G_028010_ARRAY_MODE(track->db_depth_info);
-			array_check.array_mode = array_mode;
-			array_check.group_size = track->group_size;
-			array_check.nbanks = track->nbanks;
-			array_check.npipes = track->npipes;
-			array_check.nsamples = track->nsamples;
-			array_check.blocksize = bpe;
-			if (r600_get_array_mode_alignment(&array_check,
-							  &pitch_align, &height_align, &depth_align, &base_align)) {
-				dev_warn(p->dev, "%s invalid tiling %d (0x%08X)\n", __func__,
-					 G_028010_ARRAY_MODE(track->db_depth_info),
-					 track->db_depth_info);
-				return -EINVAL;
-			}
-			switch (array_mode) {
-			case V_028010_ARRAY_1D_TILED_THIN1:
-				/* don't break userspace */
-				height &= ~0x7;
+			/* htile widht & nby (8 or 4) make 2 bits number */
+			tmp = track->htile_surface & 3;
+			/* align is htile align * 8, htile align vary according to
+			 * number of pipe and tile width and nby
+			 */
+			switch (track->npipes) {
+			case 8:
+				switch (tmp) {
+				case 3:	/* HTILE_WIDTH = 8 & HTILE_HEIGHT = 8*/
+					nbx = round_up(nbx, 64 * 8);
+					nby = round_up(nby, 64 * 8);
+					break;
+				case 2:	/* HTILE_WIDTH = 4 & HTILE_HEIGHT = 8*/
+				case 1:	/* HTILE_WIDTH = 8 & HTILE_HEIGHT = 4*/
+					nbx = round_up(nbx, 64 * 8);
+					nby = round_up(nby, 32 * 8);
+					break;
+				case 0:	/* HTILE_WIDTH = 4 & HTILE_HEIGHT = 4*/
+					nbx = round_up(nbx, 32 * 8);
+					nby = round_up(nby, 32 * 8);
+					break;
+				default:
+					return -EINVAL;
+				}
+				break;
+			case 4:
+				switch (tmp) {
+				case 3:	/* HTILE_WIDTH = 8 & HTILE_HEIGHT = 8*/
+					nbx = round_up(nbx, 64 * 8);
+					nby = round_up(nby, 32 * 8);
+					break;
+				case 2:	/* HTILE_WIDTH = 4 & HTILE_HEIGHT = 8*/
+				case 1:	/* HTILE_WIDTH = 8 & HTILE_HEIGHT = 4*/
+					nbx = round_up(nbx, 32 * 8);
+					nby = round_up(nby, 32 * 8);
+					break;
+				case 0:	/* HTILE_WIDTH = 4 & HTILE_HEIGHT = 4*/
+					nbx = round_up(nbx, 32 * 8);
+					nby = round_up(nby, 16 * 8);
+					break;
+				default:
+					return -EINVAL;
+				}
+				break;
+			case 2:
+				switch (tmp) {
+				case 3:	/* HTILE_WIDTH = 8 & HTILE_HEIGHT = 8*/
+					nbx = round_up(nbx, 32 * 8);
+					nby = round_up(nby, 32 * 8);
+					break;
+				case 2:	/* HTILE_WIDTH = 4 & HTILE_HEIGHT = 8*/
+				case 1:	/* HTILE_WIDTH = 8 & HTILE_HEIGHT = 4*/
+					nbx = round_up(nbx, 32 * 8);
+					nby = round_up(nby, 16 * 8);
+					break;
+				case 0:	/* HTILE_WIDTH = 4 & HTILE_HEIGHT = 4*/
+					nbx = round_up(nbx, 16 * 8);
+					nby = round_up(nby, 16 * 8);
+					break;
+				default:
+					return -EINVAL;
+				}
 				break;
-			case V_028010_ARRAY_2D_TILED_THIN1:
+			case 1:
+				switch (tmp) {
+				case 3:	/* HTILE_WIDTH = 8 & HTILE_HEIGHT = 8*/
+					nbx = round_up(nbx, 32 * 8);
+					nby = round_up(nby, 16 * 8);
+					break;
+				case 2:	/* HTILE_WIDTH = 4 & HTILE_HEIGHT = 8*/
+				case 1:	/* HTILE_WIDTH = 8 & HTILE_HEIGHT = 4*/
+					nbx = round_up(nbx, 16 * 8);
+					nby = round_up(nby, 16 * 8);
+					break;
+				case 0:	/* HTILE_WIDTH = 4 & HTILE_HEIGHT = 4*/
+					nbx = round_up(nbx, 16 * 8);
+					nby = round_up(nby, 8 * 8);
+					break;
+				default:
+					return -EINVAL;
+				}
 				break;
 			default:
-				dev_warn(p->dev, "%s invalid tiling %d (0x%08X)\n", __func__,
-					 G_028010_ARRAY_MODE(track->db_depth_info),
-					 track->db_depth_info);
+				dev_warn(p->dev, "%s:%d invalid num pipes %d\n",
+					 __func__, __LINE__, track->npipes);
 				return -EINVAL;
 			}
+		}
+		/* compute number of htile */
+		nbx = G_028D24_HTILE_WIDTH(track->htile_surface) ? nbx / 8 : nbx / 4;
+		nby = G_028D24_HTILE_HEIGHT(track->htile_surface) ? nby / 8 : nby / 4;
+		size = nbx * nby * 4;
+		size += track->htile_offset;
+
+		if (size > radeon_bo_size(track->htile_bo)) {
+			dev_warn(p->dev, "%s:%d htile surface too small %ld for %ld (%d %d)\n",
+				 __func__, __LINE__, radeon_bo_size(track->htile_bo),
+				 size, nbx, nby);
+			return -EINVAL;
+		}
+	}
 
-			if (!IS_ALIGNED(pitch, pitch_align)) {
-				dev_warn(p->dev, "%s:%d db pitch (%d, 0x%x, %d) invalid\n",
-					 __func__, __LINE__, pitch, pitch_align, array_mode);
-				return -EINVAL;
-			}
-			if (!IS_ALIGNED(height, height_align)) {
-				dev_warn(p->dev, "%s:%d db height (%d, 0x%x, %d) invalid\n",
-					 __func__, __LINE__, height, height_align, array_mode);
-				return -EINVAL;
-			}
-			if (!IS_ALIGNED(base_offset, base_align)) {
-				dev_warn(p->dev, "%s offset[%d] 0x%llx, 0x%llx, %d not aligned\n", __func__, i,
-					 base_offset, base_align, array_mode);
-				return -EINVAL;
+	track->db_dirty = false;
+	return 0;
+}
+
+static int r600_cs_track_check(struct radeon_cs_parser *p)
+{
+	struct r600_cs_track *track = p->track;
+	u32 tmp;
+	int r, i;
+
+	/* on legacy kernel we don't perform advanced check */
+	if (p->rdev == NULL)
+		return 0;
+
+	/* check streamout */
+	if (track->streamout_dirty && track->vgt_strmout_en) {
+		for (i = 0; i < 4; i++) {
+			if (track->vgt_strmout_buffer_en & (1 << i)) {
+				if (track->vgt_strmout_bo[i]) {
+					u64 offset = (u64)track->vgt_strmout_bo_offset[i] +
+						(u64)track->vgt_strmout_size[i];
+					if (offset > radeon_bo_size(track->vgt_strmout_bo[i])) {
+						DRM_ERROR("streamout %d bo too small: 0x%llx, 0x%lx\n",
+							  i, offset,
+							  radeon_bo_size(track->vgt_strmout_bo[i]));
+						return -EINVAL;
+					}
+				} else {
+					dev_warn(p->dev, "No buffer for streamout %d\n", i);
+					return -EINVAL;
+				}
 			}
+		}
+		track->streamout_dirty = false;
+	}
 
-			ntiles = G_028000_SLICE_TILE_MAX(track->db_depth_size) + 1;
-			nviews = G_028004_SLICE_MAX(track->db_depth_view) + 1;
-			tmp = ntiles * bpe * 64 * nviews;
-			if ((tmp + track->db_offset) > radeon_bo_size(track->db_bo)) {
-				dev_warn(p->dev, "z/stencil buffer (%d) too small (0x%08X %d %d %d -> %u have %lu)\n",
-					 array_mode,
-					 track->db_depth_size, ntiles, nviews, bpe, tmp + track->db_offset,
-					 radeon_bo_size(track->db_bo));
-				return -EINVAL;
+	if (track->sx_misc_kill_all_prims)
+		return 0;
+
+	/* check that we have a cb for each enabled target, we don't check
+	 * shader_mask because it seems mesa isn't always setting it :(
+	 */
+	if (track->cb_dirty) {
+		tmp = track->cb_target_mask;
+		for (i = 0; i < 8; i++) {
+			if ((tmp >> (i * 4)) & 0xF) {
+				/* at least one component is enabled */
+				if (track->cb_color_bo[i] == NULL) {
+					dev_warn(p->dev, "%s:%d mask 0x%08X | 0x%08X no cb for %d\n",
+						__func__, __LINE__, track->cb_target_mask, track->cb_shader_mask, i);
+					return -EINVAL;
+				}
+				/* perform rewrite of CB_COLOR[0-7]_SIZE */
+				r = r600_cs_track_validate_cb(p, i);
+				if (r)
+					return r;
 			}
 		}
+		track->cb_dirty = false;
+	}
+
+	/* Check depth buffer */
+	if (track->db_dirty && (G_028800_STENCIL_ENABLE(track->db_depth_control) ||
+		G_028800_Z_ENABLE(track->db_depth_control))) {
+		r = r600_cs_track_validate_db(p);
+		if (r)
+			return r;
 	}
+
 	return 0;
 }
 
@@ -939,6 +1142,7 @@ static int r600_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
 		break;
 	case R_028800_DB_DEPTH_CONTROL:
 		track->db_depth_control = radeon_get_ib_value(p, idx);
+		track->db_dirty = true;
 		break;
 	case R_028010_DB_DEPTH_INFO:
 		if (!(p->cs_flags & RADEON_CS_KEEP_TILING_FLAGS) &&
@@ -959,24 +1163,66 @@ static int r600_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
 				ib[idx] |= S_028010_ARRAY_MODE(V_028010_ARRAY_1D_TILED_THIN1);
 				track->db_depth_info |= S_028010_ARRAY_MODE(V_028010_ARRAY_1D_TILED_THIN1);
 			}
-		} else
+		} else {
 			track->db_depth_info = radeon_get_ib_value(p, idx);
+		}
+		track->db_dirty = true;
 		break;
 	case R_028004_DB_DEPTH_VIEW:
 		track->db_depth_view = radeon_get_ib_value(p, idx);
+		track->db_dirty = true;
 		break;
 	case R_028000_DB_DEPTH_SIZE:
 		track->db_depth_size = radeon_get_ib_value(p, idx);
 		track->db_depth_size_idx = idx;
+		track->db_dirty = true;
 		break;
 	case R_028AB0_VGT_STRMOUT_EN:
 		track->vgt_strmout_en = radeon_get_ib_value(p, idx);
+		track->streamout_dirty = true;
 		break;
 	case R_028B20_VGT_STRMOUT_BUFFER_EN:
 		track->vgt_strmout_buffer_en = radeon_get_ib_value(p, idx);
+		track->streamout_dirty = true;
+		break;
+	case VGT_STRMOUT_BUFFER_BASE_0:
+	case VGT_STRMOUT_BUFFER_BASE_1:
+	case VGT_STRMOUT_BUFFER_BASE_2:
+	case VGT_STRMOUT_BUFFER_BASE_3:
+		r = r600_cs_packet_next_reloc(p, &reloc);
+		if (r) {
+			dev_warn(p->dev, "bad SET_CONTEXT_REG "
+					"0x%04X\n", reg);
+			return -EINVAL;
+		}
+		tmp = (reg - VGT_STRMOUT_BUFFER_BASE_0) / 16;
+		track->vgt_strmout_bo_offset[tmp] = radeon_get_ib_value(p, idx) << 8;
+		ib[idx] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff);
+		track->vgt_strmout_bo[tmp] = reloc->robj;
+		track->vgt_strmout_bo_mc[tmp] = reloc->lobj.gpu_offset;
+		track->streamout_dirty = true;
+		break;
+	case VGT_STRMOUT_BUFFER_SIZE_0:
+	case VGT_STRMOUT_BUFFER_SIZE_1:
+	case VGT_STRMOUT_BUFFER_SIZE_2:
+	case VGT_STRMOUT_BUFFER_SIZE_3:
+		tmp = (reg - VGT_STRMOUT_BUFFER_SIZE_0) / 16;
+		/* size in register is DWs, convert to bytes */
+		track->vgt_strmout_size[tmp] = radeon_get_ib_value(p, idx) * 4;
+		track->streamout_dirty = true;
+		break;
+	case CP_COHER_BASE:
+		r = r600_cs_packet_next_reloc(p, &reloc);
+		if (r) {
+			dev_warn(p->dev, "missing reloc for CP_COHER_BASE "
+					"0x%04X\n", reg);
+			return -EINVAL;
+		}
+		ib[idx] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff);
 		break;
 	case R_028238_CB_TARGET_MASK:
 		track->cb_target_mask = radeon_get_ib_value(p, idx);
+		track->cb_dirty = true;
 		break;
 	case R_02823C_CB_SHADER_MASK:
 		track->cb_shader_mask = radeon_get_ib_value(p, idx);
@@ -984,6 +1230,7 @@ static int r600_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
 	case R_028C04_PA_SC_AA_CONFIG:
 		tmp = G_028C04_MSAA_NUM_SAMPLES(radeon_get_ib_value(p, idx));
 		track->nsamples = 1 << tmp;
+		track->cb_dirty = true;
 		break;
 	case R_0280A0_CB_COLOR0_INFO:
 	case R_0280A4_CB_COLOR1_INFO:
@@ -1013,6 +1260,19 @@ static int r600_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
 			tmp = (reg - R_0280A0_CB_COLOR0_INFO) / 4;
 			track->cb_color_info[tmp] = radeon_get_ib_value(p, idx);
 		}
+		track->cb_dirty = true;
+		break;
+	case R_028080_CB_COLOR0_VIEW:
+	case R_028084_CB_COLOR1_VIEW:
+	case R_028088_CB_COLOR2_VIEW:
+	case R_02808C_CB_COLOR3_VIEW:
+	case R_028090_CB_COLOR4_VIEW:
+	case R_028094_CB_COLOR5_VIEW:
+	case R_028098_CB_COLOR6_VIEW:
+	case R_02809C_CB_COLOR7_VIEW:
+		tmp = (reg - R_028080_CB_COLOR0_VIEW) / 4;
+		track->cb_color_view[tmp] = radeon_get_ib_value(p, idx);
+		track->cb_dirty = true;
 		break;
 	case R_028060_CB_COLOR0_SIZE:
 	case R_028064_CB_COLOR1_SIZE:
@@ -1025,6 +1285,7 @@ static int r600_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
 		tmp = (reg - R_028060_CB_COLOR0_SIZE) / 4;
 		track->cb_color_size[tmp] = radeon_get_ib_value(p, idx);
 		track->cb_color_size_idx[tmp] = idx;
+		track->cb_dirty = true;
 		break;
 		/* This register were added late, there is userspace
 		 * which does provide relocation for those but set
@@ -1107,6 +1368,7 @@ static int r600_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
 		track->cb_color_base_last[tmp] = ib[idx];
 		track->cb_color_bo[tmp] = reloc->robj;
 		track->cb_color_bo_mc[tmp] = reloc->lobj.gpu_offset;
+		track->cb_dirty = true;
 		break;
 	case DB_DEPTH_BASE:
 		r = r600_cs_packet_next_reloc(p, &reloc);
@@ -1119,8 +1381,24 @@ static int r600_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
 		ib[idx] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff);
 		track->db_bo = reloc->robj;
 		track->db_bo_mc = reloc->lobj.gpu_offset;
+		track->db_dirty = true;
 		break;
 	case DB_HTILE_DATA_BASE:
+		r = r600_cs_packet_next_reloc(p, &reloc);
+		if (r) {
+			dev_warn(p->dev, "bad SET_CONTEXT_REG "
+					"0x%04X\n", reg);
+			return -EINVAL;
+		}
+		track->htile_offset = radeon_get_ib_value(p, idx) << 8;
+		ib[idx] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff);
+		track->htile_bo = reloc->robj;
+		track->db_dirty = true;
+		break;
+	case DB_HTILE_SURFACE:
+		track->htile_surface = radeon_get_ib_value(p, idx);
+		track->db_dirty = true;
+		break;
 	case SQ_PGM_START_FS:
 	case SQ_PGM_START_ES:
 	case SQ_PGM_START_VS:
@@ -1191,6 +1469,9 @@ static int r600_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
 		}
 		ib[idx] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff);
 		break;
+	case SX_MISC:
+		track->sx_misc_kill_all_prims = (radeon_get_ib_value(p, idx) & 0x1) != 0;
+		break;
 	default:
 		dev_warn(p->dev, "forbidden register 0x%08x at %d\n", reg, idx);
 		return -EINVAL;
@@ -1198,7 +1479,7 @@ static int r600_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
 	return 0;
 }
 
-static unsigned mip_minify(unsigned size, unsigned level)
+unsigned r600_mip_minify(unsigned size, unsigned level)
 {
 	unsigned val;
 
@@ -1220,22 +1501,22 @@ static void r600_texture_size(unsigned nfaces, unsigned blevel, unsigned llevel,
 	unsigned nlevels = llevel - blevel + 1;
 
 	*l0_size = -1;
-	blocksize = fmt_get_blocksize(format);
+	blocksize = r600_fmt_get_blocksize(format);
 
-	w0 = mip_minify(w0, 0);
-	h0 = mip_minify(h0, 0);
-	d0 = mip_minify(d0, 0);
+	w0 = r600_mip_minify(w0, 0);
+	h0 = r600_mip_minify(h0, 0);
+	d0 = r600_mip_minify(d0, 0);
 	for(i = 0, offset = 0, level = blevel; i < nlevels; i++, level++) {
-		width = mip_minify(w0, i);
-		nbx = fmt_get_nblocksx(format, width);
+		width = r600_mip_minify(w0, i);
+		nbx = r600_fmt_get_nblocksx(format, width);
 
 		nbx = round_up(nbx, block_align);
 
-		height = mip_minify(h0, i);
-		nby = fmt_get_nblocksy(format, height);
+		height = r600_mip_minify(h0, i);
+		nby = r600_fmt_get_nblocksy(format, height);
 		nby = round_up(nby, height_align);
 
-		depth = mip_minify(d0, i);
+		depth = r600_mip_minify(d0, i);
 
 		size = nbx * nby * blocksize;
 		if (nfaces)
@@ -1304,6 +1585,7 @@ static int r600_check_texture_resource(struct radeon_cs_parser *p,  u32 idx,
 	h0 = G_038004_TEX_HEIGHT(word1) + 1;
 	d0 = G_038004_TEX_DEPTH(word1);
 	nfaces = 1;
+	array = 0;
 	switch (G_038000_DIM(word0)) {
 	case V_038000_SQ_TEX_DIM_1D:
 	case V_038000_SQ_TEX_DIM_2D:
@@ -1326,7 +1608,7 @@ static int r600_check_texture_resource(struct radeon_cs_parser *p,  u32 idx,
 		return -EINVAL;
 	}
 	format = G_038004_DATA_FORMAT(word1);
-	if (!fmt_is_valid_texture(format, p->family)) {
+	if (!r600_fmt_is_valid_texture(format, p->family)) {
 		dev_warn(p->dev, "%s:%d texture invalid format %d\n",
 			 __func__, __LINE__, format);
 		return -EINVAL;
@@ -1339,7 +1621,7 @@ static int r600_check_texture_resource(struct radeon_cs_parser *p,  u32 idx,
 	array_check.nbanks = track->nbanks;
 	array_check.npipes = track->npipes;
 	array_check.nsamples = 1;
-	array_check.blocksize = fmt_get_blocksize(format);
+	array_check.blocksize = r600_fmt_get_blocksize(format);
 	if (r600_get_array_mode_alignment(&array_check,
 					  &pitch_align, &height_align, &depth_align, &base_align)) {
 		dev_warn(p->dev, "%s:%d tex array mode (%d) invalid\n",
@@ -1372,6 +1654,10 @@ static int r600_check_texture_resource(struct radeon_cs_parser *p,  u32 idx,
 	word1 = radeon_get_ib_value(p, idx + 5);
 	blevel = G_038010_BASE_LEVEL(word0);
 	llevel = G_038014_LAST_LEVEL(word1);
+	if (blevel > llevel) {
+		dev_warn(p->dev, "texture blevel %d > llevel %d\n",
+			 blevel, llevel);
+	}
 	if (array == 1) {
 		barray = G_038014_BASE_ARRAY(word1);
 		larray = G_038014_LAST_ARRAY(word1);
@@ -1383,8 +1669,10 @@ static int r600_check_texture_resource(struct radeon_cs_parser *p,  u32 idx,
 			  &l0_size, &mipmap_size);
 	/* using get ib will give us the offset into the texture bo */
 	if ((l0_size + word2) > radeon_bo_size(texture)) {
-		dev_warn(p->dev, "texture bo too small (%d %d %d %d -> %d have %ld)\n",
-			w0, h0, format, word2, l0_size, radeon_bo_size(texture));
+		dev_warn(p->dev, "texture bo too small ((%d %d) (%d %d) %d %d %d -> %d have %ld)\n",
+			 w0, h0, pitch_align, height_align,
+			 array_check.array_mode, format, word2,
+			 l0_size, radeon_bo_size(texture));
 		dev_warn(p->dev, "alignments %d %d %d %lld\n", pitch, pitch_align, height_align, base_align);
 		return -EINVAL;
 	}
@@ -1397,6 +1685,22 @@ static int r600_check_texture_resource(struct radeon_cs_parser *p,  u32 idx,
 	return 0;
 }
 
+static bool r600_is_safe_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
+{
+	u32 m, i;
+
+	i = (reg >> 7);
+	if (i >= ARRAY_SIZE(r600_reg_safe_bm)) {
+		dev_warn(p->dev, "forbidden register 0x%08x at %d\n", reg, idx);
+		return false;
+	}
+	m = 1 << ((reg >> 2) & 31);
+	if (!(r600_reg_safe_bm[i] & m))
+		return true;
+	dev_warn(p->dev, "forbidden register 0x%08x at %d\n", reg, idx);
+	return false;
+}
+
 static int r600_packet3_check(struct radeon_cs_parser *p,
 				struct radeon_cs_packet *pkt)
 {
@@ -1419,6 +1723,8 @@ static int r600_packet3_check(struct radeon_cs_parser *p,
 	{
 		int pred_op;
 		int tmp;
+		uint64_t offset;
+
 		if (pkt->count != 1) {
 			DRM_ERROR("bad SET PREDICATION\n");
 			return -EINVAL;
@@ -1442,8 +1748,12 @@ static int r600_packet3_check(struct radeon_cs_parser *p,
 			return -EINVAL;
 		}
 
-		ib[idx + 0] = idx_value + (u32)(reloc->lobj.gpu_offset & 0xffffffff);
-		ib[idx + 1] = tmp + (upper_32_bits(reloc->lobj.gpu_offset) & 0xff);
+		offset = reloc->lobj.gpu_offset +
+		         (idx_value & 0xfffffff0) +
+		         ((u64)(tmp & 0xff) << 32);
+
+		ib[idx + 0] = offset;
+		ib[idx + 1] = (tmp & 0xffffff00) | (upper_32_bits(offset) & 0xff);
 	}
 	break;
 
@@ -1467,6 +1777,8 @@ static int r600_packet3_check(struct radeon_cs_parser *p,
 		}
 		break;
 	case PACKET3_DRAW_INDEX:
+	{
+		uint64_t offset;
 		if (pkt->count != 3) {
 			DRM_ERROR("bad DRAW_INDEX\n");
 			return -EINVAL;
@@ -1476,14 +1788,21 @@ static int r600_packet3_check(struct radeon_cs_parser *p,
 			DRM_ERROR("bad DRAW_INDEX\n");
 			return -EINVAL;
 		}
-		ib[idx+0] = idx_value + (u32)(reloc->lobj.gpu_offset & 0xffffffff);
-		ib[idx+1] += upper_32_bits(reloc->lobj.gpu_offset) & 0xff;
+
+		offset = reloc->lobj.gpu_offset +
+		         idx_value +
+		         ((u64)(radeon_get_ib_value(p, idx+1) & 0xff) << 32);
+
+		ib[idx+0] = offset;
+		ib[idx+1] = upper_32_bits(offset) & 0xff;
+
 		r = r600_cs_track_check(p);
 		if (r) {
 			dev_warn(p->dev, "%s:%d invalid cmd stream\n", __func__, __LINE__);
 			return r;
 		}
 		break;
+	}
 	case PACKET3_DRAW_INDEX_AUTO:
 		if (pkt->count != 1) {
 			DRM_ERROR("bad DRAW_INDEX_AUTO\n");
@@ -1514,13 +1833,20 @@ static int r600_packet3_check(struct radeon_cs_parser *p,
 		}
 		/* bit 4 is reg (0) or mem (1) */
 		if (idx_value & 0x10) {
+			uint64_t offset;
+
 			r = r600_cs_packet_next_reloc(p, &reloc);
 			if (r) {
 				DRM_ERROR("bad WAIT_REG_MEM\n");
 				return -EINVAL;
 			}
-			ib[idx+1] += (u32)(reloc->lobj.gpu_offset & 0xffffffff);
-			ib[idx+2] += upper_32_bits(reloc->lobj.gpu_offset) & 0xff;
+
+			offset = reloc->lobj.gpu_offset +
+			         (radeon_get_ib_value(p, idx+1) & 0xfffffff0) +
+			         ((u64)(radeon_get_ib_value(p, idx+2) & 0xff) << 32);
+
+			ib[idx+1] = (ib[idx+1] & 0x3) | (offset & 0xfffffff0);
+			ib[idx+2] = upper_32_bits(offset) & 0xff;
 		}
 		break;
 	case PACKET3_SURFACE_SYNC:
@@ -1545,16 +1871,25 @@ static int r600_packet3_check(struct radeon_cs_parser *p,
 			return -EINVAL;
 		}
 		if (pkt->count) {
+			uint64_t offset;
+
 			r = r600_cs_packet_next_reloc(p, &reloc);
 			if (r) {
 				DRM_ERROR("bad EVENT_WRITE\n");
 				return -EINVAL;
 			}
-			ib[idx+1] += (u32)(reloc->lobj.gpu_offset & 0xffffffff);
-			ib[idx+2] += upper_32_bits(reloc->lobj.gpu_offset) & 0xff;
+			offset = reloc->lobj.gpu_offset +
+			         (radeon_get_ib_value(p, idx+1) & 0xfffffff8) +
+			         ((u64)(radeon_get_ib_value(p, idx+2) & 0xff) << 32);
+
+			ib[idx+1] = offset & 0xfffffff8;
+			ib[idx+2] = upper_32_bits(offset) & 0xff;
 		}
 		break;
 	case PACKET3_EVENT_WRITE_EOP:
+	{
+		uint64_t offset;
+
 		if (pkt->count != 4) {
 			DRM_ERROR("bad EVENT_WRITE_EOP\n");
 			return -EINVAL;
@@ -1564,9 +1899,15 @@ static int r600_packet3_check(struct radeon_cs_parser *p,
 			DRM_ERROR("bad EVENT_WRITE\n");
 			return -EINVAL;
 		}
-		ib[idx+1] += (u32)(reloc->lobj.gpu_offset & 0xffffffff);
-		ib[idx+2] += upper_32_bits(reloc->lobj.gpu_offset) & 0xff;
+
+		offset = reloc->lobj.gpu_offset +
+		         (radeon_get_ib_value(p, idx+1) & 0xfffffffc) +
+		         ((u64)(radeon_get_ib_value(p, idx+2) & 0xff) << 32);
+
+		ib[idx+1] = offset & 0xfffffffc;
+		ib[idx+2] = (ib[idx+2] & 0xffffff00) | (upper_32_bits(offset) & 0xff);
 		break;
+	}
 	case PACKET3_SET_CONFIG_REG:
 		start_reg = (idx_value << 2) + PACKET3_SET_CONFIG_REG_OFFSET;
 		end_reg = 4 * pkt->count + start_reg - 4;
@@ -1651,6 +1992,8 @@ static int r600_packet3_check(struct radeon_cs_parser *p,
 				ib[idx+1+(i*7)+3] += mip_offset;
 				break;
 			case SQ_TEX_VTX_VALID_BUFFER:
+			{
+				uint64_t offset64;
 				/* vtx base */
 				r = r600_cs_packet_next_reloc(p, &reloc);
 				if (r) {
@@ -1663,11 +2006,15 @@ static int r600_packet3_check(struct radeon_cs_parser *p,
 					/* force size to size of the buffer */
 					dev_warn(p->dev, "vbo resource seems too big (%d) for the bo (%ld)\n",
 						 size + offset, radeon_bo_size(reloc->robj));
-					ib[idx+1+(i*7)+1] = radeon_bo_size(reloc->robj);
+					ib[idx+1+(i*7)+1] = radeon_bo_size(reloc->robj) - offset;
 				}
-				ib[idx+1+(i*7)+0] += (u32)((reloc->lobj.gpu_offset) & 0xffffffff);
-				ib[idx+1+(i*7)+2] += upper_32_bits(reloc->lobj.gpu_offset) & 0xff;
+
+				offset64 = reloc->lobj.gpu_offset + offset;
+				ib[idx+1+(i*8)+0] = offset64;
+				ib[idx+1+(i*8)+2] = (ib[idx+1+(i*8)+2] & 0xffffff00) |
+						    (upper_32_bits(offset64) & 0xff);
 				break;
+			}
 			case SQ_TEX_VTX_INVALID_TEXTURE:
 			case SQ_TEX_VTX_INVALID_BUFFER:
 			default:
@@ -1742,6 +2089,104 @@ static int r600_packet3_check(struct radeon_cs_parser *p,
 			return -EINVAL;
 		}
 		break;
+	case PACKET3_STRMOUT_BUFFER_UPDATE:
+		if (pkt->count != 4) {
+			DRM_ERROR("bad STRMOUT_BUFFER_UPDATE (invalid count)\n");
+			return -EINVAL;
+		}
+		/* Updating memory at DST_ADDRESS. */
+		if (idx_value & 0x1) {
+			u64 offset;
+			r = r600_cs_packet_next_reloc(p, &reloc);
+			if (r) {
+				DRM_ERROR("bad STRMOUT_BUFFER_UPDATE (missing dst reloc)\n");
+				return -EINVAL;
+			}
+			offset = radeon_get_ib_value(p, idx+1);
+			offset += ((u64)(radeon_get_ib_value(p, idx+2) & 0xff)) << 32;
+			if ((offset + 4) > radeon_bo_size(reloc->robj)) {
+				DRM_ERROR("bad STRMOUT_BUFFER_UPDATE dst bo too small: 0x%llx, 0x%lx\n",
+					  offset + 4, radeon_bo_size(reloc->robj));
+				return -EINVAL;
+			}
+			offset += reloc->lobj.gpu_offset;
+			ib[idx+1] = offset;
+			ib[idx+2] = upper_32_bits(offset) & 0xff;
+		}
+		/* Reading data from SRC_ADDRESS. */
+		if (((idx_value >> 1) & 0x3) == 2) {
+			u64 offset;
+			r = r600_cs_packet_next_reloc(p, &reloc);
+			if (r) {
+				DRM_ERROR("bad STRMOUT_BUFFER_UPDATE (missing src reloc)\n");
+				return -EINVAL;
+			}
+			offset = radeon_get_ib_value(p, idx+3);
+			offset += ((u64)(radeon_get_ib_value(p, idx+4) & 0xff)) << 32;
+			if ((offset + 4) > radeon_bo_size(reloc->robj)) {
+				DRM_ERROR("bad STRMOUT_BUFFER_UPDATE src bo too small: 0x%llx, 0x%lx\n",
+					  offset + 4, radeon_bo_size(reloc->robj));
+				return -EINVAL;
+			}
+			offset += reloc->lobj.gpu_offset;
+			ib[idx+3] = offset;
+			ib[idx+4] = upper_32_bits(offset) & 0xff;
+		}
+		break;
+	case PACKET3_COPY_DW:
+		if (pkt->count != 4) {
+			DRM_ERROR("bad COPY_DW (invalid count)\n");
+			return -EINVAL;
+		}
+		if (idx_value & 0x1) {
+			u64 offset;
+			/* SRC is memory. */
+			r = r600_cs_packet_next_reloc(p, &reloc);
+			if (r) {
+				DRM_ERROR("bad COPY_DW (missing src reloc)\n");
+				return -EINVAL;
+			}
+			offset = radeon_get_ib_value(p, idx+1);
+			offset += ((u64)(radeon_get_ib_value(p, idx+2) & 0xff)) << 32;
+			if ((offset + 4) > radeon_bo_size(reloc->robj)) {
+				DRM_ERROR("bad COPY_DW src bo too small: 0x%llx, 0x%lx\n",
+					  offset + 4, radeon_bo_size(reloc->robj));
+				return -EINVAL;
+			}
+			offset += reloc->lobj.gpu_offset;
+			ib[idx+1] = offset;
+			ib[idx+2] = upper_32_bits(offset) & 0xff;
+		} else {
+			/* SRC is a reg. */
+			reg = radeon_get_ib_value(p, idx+1) << 2;
+			if (!r600_is_safe_reg(p, reg, idx+1))
+				return -EINVAL;
+		}
+		if (idx_value & 0x2) {
+			u64 offset;
+			/* DST is memory. */
+			r = r600_cs_packet_next_reloc(p, &reloc);
+			if (r) {
+				DRM_ERROR("bad COPY_DW (missing dst reloc)\n");
+				return -EINVAL;
+			}
+			offset = radeon_get_ib_value(p, idx+3);
+			offset += ((u64)(radeon_get_ib_value(p, idx+4) & 0xff)) << 32;
+			if ((offset + 4) > radeon_bo_size(reloc->robj)) {
+				DRM_ERROR("bad COPY_DW dst bo too small: 0x%llx, 0x%lx\n",
+					  offset + 4, radeon_bo_size(reloc->robj));
+				return -EINVAL;
+			}
+			offset += reloc->lobj.gpu_offset;
+			ib[idx+3] = offset;
+			ib[idx+4] = upper_32_bits(offset) & 0xff;
+		} else {
+			/* DST is a reg. */
+			reg = radeon_get_ib_value(p, idx+3) << 2;
+			if (!r600_is_safe_reg(p, reg, idx+3))
+				return -EINVAL;
+		}
+		break;
 	case PACKET3_NOP:
 		break;
 	default:
diff --git a/drivers/gpu/drm/radeon/r600d.h b/drivers/gpu/drm/radeon/r600d.h
index 3ee1fd7ef394..59f9c993cc31 100644
--- a/drivers/gpu/drm/radeon/r600d.h
+++ b/drivers/gpu/drm/radeon/r600d.h
@@ -78,6 +78,20 @@
 
 #define CB_COLOR0_SIZE                                  0x28060
 #define CB_COLOR0_VIEW                                  0x28080
+#define R_028080_CB_COLOR0_VIEW                      0x028080
+#define   S_028080_SLICE_START(x)                      (((x) & 0x7FF) << 0)
+#define   G_028080_SLICE_START(x)                      (((x) >> 0) & 0x7FF)
+#define   C_028080_SLICE_START                         0xFFFFF800
+#define   S_028080_SLICE_MAX(x)                        (((x) & 0x7FF) << 13)
+#define   G_028080_SLICE_MAX(x)                        (((x) >> 13) & 0x7FF)
+#define   C_028080_SLICE_MAX                           0xFF001FFF
+#define R_028084_CB_COLOR1_VIEW                      0x028084
+#define R_028088_CB_COLOR2_VIEW                      0x028088
+#define R_02808C_CB_COLOR3_VIEW                      0x02808C
+#define R_028090_CB_COLOR4_VIEW                      0x028090
+#define R_028094_CB_COLOR5_VIEW                      0x028094
+#define R_028098_CB_COLOR6_VIEW                      0x028098
+#define R_02809C_CB_COLOR7_VIEW                      0x02809C
 #define CB_COLOR0_INFO                                  0x280a0
 #	define CB_FORMAT(x)				((x) << 2)
 #       define CB_ARRAY_MODE(x)                         ((x) << 8)
@@ -181,6 +195,14 @@
 #define		PREZ_MUST_WAIT_FOR_POSTZ_DONE			(1 << 31)
 #define	DB_DEPTH_BASE					0x2800C
 #define	DB_HTILE_DATA_BASE				0x28014
+#define	DB_HTILE_SURFACE				0x28D24
+#define   S_028D24_HTILE_WIDTH(x)                      (((x) & 0x1) << 0)
+#define   G_028D24_HTILE_WIDTH(x)                      (((x) >> 0) & 0x1)
+#define   C_028D24_HTILE_WIDTH                         0xFFFFFFFE
+#define   S_028D24_HTILE_HEIGHT(x)                      (((x) & 0x1) << 1)
+#define   G_028D24_HTILE_HEIGHT(x)                      (((x) >> 1) & 0x1)
+#define   C_028D24_HTILE_HEIGHT                         0xFFFFFFFD
+#define   G_028D24_LINEAR(x)                           (((x) >> 2) & 0x1)
 #define	DB_WATERMARKS					0x9838
 #define		DEPTH_FREE(x)					((x) << 0)
 #define		DEPTH_FLUSH(x)					((x) << 5)
@@ -493,6 +515,11 @@
 #define	VGT_STRMOUT_BUFFER_OFFSET_1			0x28AEC
 #define	VGT_STRMOUT_BUFFER_OFFSET_2			0x28AFC
 #define	VGT_STRMOUT_BUFFER_OFFSET_3			0x28B0C
+#define VGT_STRMOUT_BUFFER_SIZE_0			0x28AD0
+#define VGT_STRMOUT_BUFFER_SIZE_1			0x28AE0
+#define VGT_STRMOUT_BUFFER_SIZE_2			0x28AF0
+#define VGT_STRMOUT_BUFFER_SIZE_3			0x28B00
+
 #define	VGT_STRMOUT_EN					0x28AB0
 #define	VGT_VERTEX_REUSE_BLOCK_CNTL			0x28C58
 #define		VTX_REUSE_DEPTH_MASK				0x000000FF
@@ -574,6 +601,10 @@
 #define RLC_UCODE_ADDR                                    0x3f2c
 #define RLC_UCODE_DATA                                    0x3f30
 
+/* new for TN */
+#define TN_RLC_SAVE_AND_RESTORE_BASE                      0x3f10
+#define TN_RLC_CLEAR_STATE_RESTORE_BASE                   0x3f20
+
 #define SRBM_SOFT_RESET                                   0xe60
 #       define SOFT_RESET_RLC                             (1 << 13)
 
@@ -831,9 +862,11 @@
 #define	PACKET3_STRMOUT_BUFFER_UPDATE			0x34
 #define	PACKET3_INDIRECT_BUFFER_MP			0x38
 #define	PACKET3_MEM_SEMAPHORE				0x39
+#              define PACKET3_SEM_WAIT_ON_SIGNAL    (0x1 << 12)
 #              define PACKET3_SEM_SEL_SIGNAL	    (0x6 << 29)
 #              define PACKET3_SEM_SEL_WAIT	    (0x7 << 29)
 #define	PACKET3_MPEG_INDEX				0x3A
+#define	PACKET3_COPY_DW					0x3B
 #define	PACKET3_WAIT_REG_MEM				0x3C
 #define	PACKET3_MEM_WRITE				0x3D
 #define	PACKET3_INDIRECT_BUFFER				0x32
diff --git a/drivers/gpu/drm/radeon/radeon.h b/drivers/gpu/drm/radeon/radeon.h
index 1668ec1ee770..138b95216d8d 100644
--- a/drivers/gpu/drm/radeon/radeon.h
+++ b/drivers/gpu/drm/radeon/radeon.h
@@ -236,12 +236,15 @@ 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);
 void radeon_atom_set_voltage(struct radeon_device *rdev, u16 voltage_level, u8 voltage_type);
-int radeon_atom_get_max_vddc(struct radeon_device *rdev, u16 *voltage);
 void rs690_pm_info(struct radeon_device *rdev);
 extern int rv6xx_get_temp(struct radeon_device *rdev);
 extern int rv770_get_temp(struct radeon_device *rdev);
 extern int evergreen_get_temp(struct radeon_device *rdev);
 extern int sumo_get_temp(struct radeon_device *rdev);
+extern int si_get_temp(struct radeon_device *rdev);
+extern void evergreen_tiling_fields(unsigned tiling_flags, unsigned *bankw,
+				    unsigned *bankh, unsigned *mtaspect,
+				    unsigned *tile_split);
 
 /*
  * Fences.
@@ -411,9 +414,6 @@ int radeon_gem_object_create(struct radeon_device *rdev, int size,
 				int alignment, int initial_domain,
 				bool discardable, bool kernel,
 				struct drm_gem_object **obj);
-int radeon_gem_object_pin(struct drm_gem_object *obj, uint32_t pin_domain,
-			  uint64_t *gpu_addr);
-void radeon_gem_object_unpin(struct drm_gem_object *obj);
 
 int radeon_mode_dumb_create(struct drm_file *file_priv,
 			    struct drm_device *dev,
@@ -632,6 +632,7 @@ struct radeon_ib {
 	uint32_t		*ptr;
 	struct radeon_fence	*fence;
 	unsigned		vm_id;
+	bool			is_const_ib;
 };
 
 /*
@@ -771,6 +772,18 @@ struct r600_blit {
 
 void r600_blit_suspend(struct radeon_device *rdev);
 
+/*
+ * SI RLC stuff
+ */
+struct si_rlc {
+	/* for power gating */
+	struct radeon_bo	*save_restore_obj;
+	uint64_t		save_restore_gpu_addr;
+	/* for clear state */
+	struct radeon_bo	*clear_state_obj;
+	uint64_t		clear_state_gpu_addr;
+};
+
 int radeon_ib_get(struct radeon_device *rdev, int ring,
 		  struct radeon_ib **ib, unsigned size);
 void radeon_ib_free(struct radeon_device *rdev, struct radeon_ib **ib);
@@ -780,7 +793,6 @@ int radeon_ib_pool_init(struct radeon_device *rdev);
 void radeon_ib_pool_fini(struct radeon_device *rdev);
 int radeon_ib_pool_start(struct radeon_device *rdev);
 int radeon_ib_pool_suspend(struct radeon_device *rdev);
-int radeon_ib_test(struct radeon_device *rdev);
 /* Ring access between begin & end cannot sleep */
 int radeon_ring_index(struct radeon_device *rdev, struct radeon_ring *cp);
 void radeon_ring_free_size(struct radeon_device *rdev, struct radeon_ring *cp);
@@ -833,12 +845,13 @@ struct radeon_cs_parser {
 	struct radeon_cs_reloc	*relocs;
 	struct radeon_cs_reloc	**relocs_ptr;
 	struct list_head	validated;
-	bool			sync_to_ring[RADEON_NUM_RINGS];
 	/* indices of various chunks */
 	int			chunk_ib_idx;
 	int			chunk_relocs_idx;
 	int			chunk_flags_idx;
+	int			chunk_const_ib_idx;
 	struct radeon_ib	*ib;
+	struct radeon_ib	*const_ib;
 	void			*track;
 	unsigned		family;
 	int			parser_error;
@@ -980,6 +993,7 @@ enum radeon_int_thermal_type {
 	THERMAL_TYPE_EVERGREEN,
 	THERMAL_TYPE_SUMO,
 	THERMAL_TYPE_NI,
+	THERMAL_TYPE_SI,
 };
 
 struct radeon_voltage {
@@ -1132,57 +1146,6 @@ struct radeon_asic {
 	void (*vga_set_state)(struct radeon_device *rdev, bool state);
 	bool (*gpu_is_lockup)(struct radeon_device *rdev, struct radeon_ring *cp);
 	int (*asic_reset)(struct radeon_device *rdev);
-	void (*gart_tlb_flush)(struct radeon_device *rdev);
-	int (*gart_set_page)(struct radeon_device *rdev, int i, uint64_t addr);
-	int (*cp_init)(struct radeon_device *rdev, unsigned ring_size);
-	void (*cp_fini)(struct radeon_device *rdev);
-	void (*cp_disable)(struct radeon_device *rdev);
-	void (*ring_start)(struct radeon_device *rdev);
-
-	struct {
-		void (*ib_execute)(struct radeon_device *rdev, struct radeon_ib *ib);
-		int (*ib_parse)(struct radeon_device *rdev, struct radeon_ib *ib);
-		void (*emit_fence)(struct radeon_device *rdev, struct radeon_fence *fence);
-		void (*emit_semaphore)(struct radeon_device *rdev, struct radeon_ring *cp,
-				       struct radeon_semaphore *semaphore, bool emit_wait);
-	} ring[RADEON_NUM_RINGS];
-
-	int (*ring_test)(struct radeon_device *rdev, struct radeon_ring *cp);
-	int (*irq_set)(struct radeon_device *rdev);
-	int (*irq_process)(struct radeon_device *rdev);
-	u32 (*get_vblank_counter)(struct radeon_device *rdev, int crtc);
-	int (*cs_parse)(struct radeon_cs_parser *p);
-	int (*copy_blit)(struct radeon_device *rdev,
-			 uint64_t src_offset,
-			 uint64_t dst_offset,
-			 unsigned num_gpu_pages,
-			 struct radeon_fence *fence);
-	int (*copy_dma)(struct radeon_device *rdev,
-			uint64_t src_offset,
-			uint64_t dst_offset,
-			unsigned num_gpu_pages,
-			struct radeon_fence *fence);
-	int (*copy)(struct radeon_device *rdev,
-		    uint64_t src_offset,
-		    uint64_t dst_offset,
-		    unsigned num_gpu_pages,
-		    struct radeon_fence *fence);
-	uint32_t (*get_engine_clock)(struct radeon_device *rdev);
-	void (*set_engine_clock)(struct radeon_device *rdev, uint32_t eng_clock);
-	uint32_t (*get_memory_clock)(struct radeon_device *rdev);
-	void (*set_memory_clock)(struct radeon_device *rdev, uint32_t mem_clock);
-	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_surface_reg)(struct radeon_device *rdev, int reg,
-			       uint32_t tiling_flags, uint32_t pitch,
-			       uint32_t offset, uint32_t obj_size);
-	void (*clear_surface_reg)(struct radeon_device *rdev, int reg);
-	void (*bandwidth_update)(struct radeon_device *rdev);
-	void (*hpd_init)(struct radeon_device *rdev);
-	void (*hpd_fini)(struct radeon_device *rdev);
-	bool (*hpd_sense)(struct radeon_device *rdev, enum radeon_hpd_id hpd);
-	void (*hpd_set_polarity)(struct radeon_device *rdev, enum radeon_hpd_id hpd);
 	/* ioctl hw specific callback. Some hw might want to perform special
 	 * operation on specific ioctl. For instance on wait idle some hw
 	 * might want to perform and HDP flush through MMIO as it seems that
@@ -1190,17 +1153,99 @@ struct radeon_asic {
 	 * through ring.
 	 */
 	void (*ioctl_wait_idle)(struct radeon_device *rdev, struct radeon_bo *bo);
+	/* check if 3D engine is idle */
 	bool (*gui_idle)(struct radeon_device *rdev);
+	/* wait for mc_idle */
+	int (*mc_wait_for_idle)(struct radeon_device *rdev);
+	/* gart */
+	struct {
+		void (*tlb_flush)(struct radeon_device *rdev);
+		int (*set_page)(struct radeon_device *rdev, int i, uint64_t addr);
+	} gart;
+	/* ring specific callbacks */
+	struct {
+		void (*ib_execute)(struct radeon_device *rdev, struct radeon_ib *ib);
+		int (*ib_parse)(struct radeon_device *rdev, struct radeon_ib *ib);
+		void (*emit_fence)(struct radeon_device *rdev, struct radeon_fence *fence);
+		void (*emit_semaphore)(struct radeon_device *rdev, struct radeon_ring *cp,
+				       struct radeon_semaphore *semaphore, bool emit_wait);
+		int (*cs_parse)(struct radeon_cs_parser *p);
+		void (*ring_start)(struct radeon_device *rdev, struct radeon_ring *cp);
+		int (*ring_test)(struct radeon_device *rdev, struct radeon_ring *cp);
+		int (*ib_test)(struct radeon_device *rdev, struct radeon_ring *cp);
+	} ring[RADEON_NUM_RINGS];
+	/* irqs */
+	struct {
+		int (*set)(struct radeon_device *rdev);
+		int (*process)(struct radeon_device *rdev);
+	} irq;
+	/* displays */
+	struct {
+		/* display watermarks */
+		void (*bandwidth_update)(struct radeon_device *rdev);
+		/* get frame count */
+		u32 (*get_vblank_counter)(struct radeon_device *rdev, int crtc);
+		/* wait for vblank */
+		void (*wait_for_vblank)(struct radeon_device *rdev, int crtc);
+	} display;
+	/* copy functions for bo handling */
+	struct {
+		int (*blit)(struct radeon_device *rdev,
+			    uint64_t src_offset,
+			    uint64_t dst_offset,
+			    unsigned num_gpu_pages,
+			    struct radeon_fence *fence);
+		u32 blit_ring_index;
+		int (*dma)(struct radeon_device *rdev,
+			   uint64_t src_offset,
+			   uint64_t dst_offset,
+			   unsigned num_gpu_pages,
+			   struct radeon_fence *fence);
+		u32 dma_ring_index;
+		/* method used for bo copy */
+		int (*copy)(struct radeon_device *rdev,
+			    uint64_t src_offset,
+			    uint64_t dst_offset,
+			    unsigned num_gpu_pages,
+			    struct radeon_fence *fence);
+		/* ring used for bo copies */
+		u32 copy_ring_index;
+	} copy;
+	/* surfaces */
+	struct {
+		int (*set_reg)(struct radeon_device *rdev, int reg,
+				       uint32_t tiling_flags, uint32_t pitch,
+				       uint32_t offset, uint32_t obj_size);
+		void (*clear_reg)(struct radeon_device *rdev, int reg);
+	} surface;
+	/* hotplug detect */
+	struct {
+		void (*init)(struct radeon_device *rdev);
+		void (*fini)(struct radeon_device *rdev);
+		bool (*sense)(struct radeon_device *rdev, enum radeon_hpd_id hpd);
+		void (*set_polarity)(struct radeon_device *rdev, enum radeon_hpd_id hpd);
+	} hpd;
 	/* power management */
-	void (*pm_misc)(struct radeon_device *rdev);
-	void (*pm_prepare)(struct radeon_device *rdev);
-	void (*pm_finish)(struct radeon_device *rdev);
-	void (*pm_init_profile)(struct radeon_device *rdev);
-	void (*pm_get_dynpm_state)(struct radeon_device *rdev);
+	struct {
+		void (*misc)(struct radeon_device *rdev);
+		void (*prepare)(struct radeon_device *rdev);
+		void (*finish)(struct radeon_device *rdev);
+		void (*init_profile)(struct radeon_device *rdev);
+		void (*get_dynpm_state)(struct radeon_device *rdev);
+		uint32_t (*get_engine_clock)(struct radeon_device *rdev);
+		void (*set_engine_clock)(struct radeon_device *rdev, uint32_t eng_clock);
+		uint32_t (*get_memory_clock)(struct radeon_device *rdev);
+		void (*set_memory_clock)(struct radeon_device *rdev, uint32_t mem_clock);
+		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);
+	} pm;
 	/* pageflipping */
-	void (*pre_page_flip)(struct radeon_device *rdev, int crtc);
-	u32 (*page_flip)(struct radeon_device *rdev, int crtc, u64 crtc_base);
-	void (*post_page_flip)(struct radeon_device *rdev, int crtc);
+	struct {
+		void (*pre_page_flip)(struct radeon_device *rdev, int crtc);
+		u32 (*page_flip)(struct radeon_device *rdev, int crtc, u64 crtc_base);
+		void (*post_page_flip)(struct radeon_device *rdev, int crtc);
+	} pflip;
 };
 
 /*
@@ -1340,6 +1385,37 @@ struct cayman_asic {
 	struct r100_gpu_lockup	lockup;
 };
 
+struct si_asic {
+	unsigned max_shader_engines;
+	unsigned max_pipes_per_simd;
+	unsigned max_tile_pipes;
+	unsigned max_simds_per_se;
+	unsigned max_backends_per_se;
+	unsigned max_texture_channel_caches;
+	unsigned max_gprs;
+	unsigned max_gs_threads;
+	unsigned max_hw_contexts;
+	unsigned sc_prim_fifo_size_frontend;
+	unsigned sc_prim_fifo_size_backend;
+	unsigned sc_hiz_tile_fifo_size;
+	unsigned sc_earlyz_tile_fifo_size;
+
+	unsigned num_shader_engines;
+	unsigned num_tile_pipes;
+	unsigned num_backends_per_se;
+	unsigned backend_disable_mask_per_asic;
+	unsigned backend_map;
+	unsigned num_texture_channel_caches;
+	unsigned mem_max_burst_length_bytes;
+	unsigned mem_row_size_in_kb;
+	unsigned shader_engine_tile_size;
+	unsigned num_gpus;
+	unsigned multi_gpu_tile_size;
+
+	unsigned tile_config;
+	struct r100_gpu_lockup	lockup;
+};
+
 union radeon_asic_config {
 	struct r300_asic	r300;
 	struct r100_asic	r100;
@@ -1347,6 +1423,7 @@ union radeon_asic_config {
 	struct rv770_asic	rv770;
 	struct evergreen_asic	evergreen;
 	struct cayman_asic	cayman;
+	struct si_asic		si;
 };
 
 /*
@@ -1462,10 +1539,12 @@ struct radeon_device {
 	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 */
 	struct r600_blit r600_blit;
 	struct r600_vram_scratch vram_scratch;
 	int msi_enabled; /* msi enabled */
 	struct r600_ih ih; /* r6/700 interrupt ring */
+	struct si_rlc rlc;
 	struct work_struct hotplug_work;
 	int num_crtc; /* number of crtcs */
 	struct mutex dc_hw_i2c_mutex; /* display controller hw i2c mutex */
@@ -1491,8 +1570,6 @@ struct radeon_device {
 	unsigned 		debugfs_count;
 	/* virtual memory */
 	struct radeon_vm_manager	vm_manager;
-	/* ring used for bo copies */
-	u32				copy_ring;
 };
 
 int radeon_device_init(struct radeon_device *rdev,
@@ -1611,6 +1688,9 @@ void r100_pll_errata_after_index(struct radeon_device *rdev);
 #define ASIC_IS_DCE41(rdev) ((rdev->family >= CHIP_PALM) && \
 			     (rdev->flags & RADEON_IS_IGP))
 #define ASIC_IS_DCE5(rdev) ((rdev->family >= CHIP_BARTS))
+#define ASIC_IS_DCE6(rdev) ((rdev->family >= CHIP_ARUBA))
+#define ASIC_IS_DCE61(rdev) ((rdev->family >= CHIP_ARUBA) && \
+			     (rdev->flags & RADEON_IS_IGP))
 
 /*
  * BIOS helpers.
@@ -1648,47 +1728,53 @@ void radeon_ring_write(struct radeon_ring *ring, uint32_t v);
 #define radeon_fini(rdev) (rdev)->asic->fini((rdev))
 #define radeon_resume(rdev) (rdev)->asic->resume((rdev))
 #define radeon_suspend(rdev) (rdev)->asic->suspend((rdev))
-#define radeon_cs_parse(p) rdev->asic->cs_parse((p))
+#define radeon_cs_parse(rdev, r, p) (rdev)->asic->ring[(r)].cs_parse((p))
 #define radeon_vga_set_state(rdev, state) (rdev)->asic->vga_set_state((rdev), (state))
 #define radeon_gpu_is_lockup(rdev, cp) (rdev)->asic->gpu_is_lockup((rdev), (cp))
 #define radeon_asic_reset(rdev) (rdev)->asic->asic_reset((rdev))
-#define radeon_gart_tlb_flush(rdev) (rdev)->asic->gart_tlb_flush((rdev))
-#define radeon_gart_set_page(rdev, i, p) (rdev)->asic->gart_set_page((rdev), (i), (p))
-#define radeon_ring_start(rdev) (rdev)->asic->ring_start((rdev))
-#define radeon_ring_test(rdev, cp) (rdev)->asic->ring_test((rdev), (cp))
+#define radeon_gart_tlb_flush(rdev) (rdev)->asic->gart.tlb_flush((rdev))
+#define radeon_gart_set_page(rdev, i, p) (rdev)->asic->gart.set_page((rdev), (i), (p))
+#define radeon_ring_start(rdev, r, cp) (rdev)->asic->ring[(r)].ring_start((rdev), (cp))
+#define radeon_ring_test(rdev, r, cp) (rdev)->asic->ring[(r)].ring_test((rdev), (cp))
+#define radeon_ib_test(rdev, r, cp) (rdev)->asic->ring[(r)].ib_test((rdev), (cp))
 #define radeon_ring_ib_execute(rdev, r, ib) (rdev)->asic->ring[(r)].ib_execute((rdev), (ib))
 #define radeon_ring_ib_parse(rdev, r, ib) (rdev)->asic->ring[(r)].ib_parse((rdev), (ib))
-#define radeon_irq_set(rdev) (rdev)->asic->irq_set((rdev))
-#define radeon_irq_process(rdev) (rdev)->asic->irq_process((rdev))
-#define radeon_get_vblank_counter(rdev, crtc) (rdev)->asic->get_vblank_counter((rdev), (crtc))
+#define radeon_irq_set(rdev) (rdev)->asic->irq.set((rdev))
+#define radeon_irq_process(rdev) (rdev)->asic->irq.process((rdev))
+#define radeon_get_vblank_counter(rdev, crtc) (rdev)->asic->display.get_vblank_counter((rdev), (crtc))
 #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))
-#define radeon_copy_dma(rdev, s, d, np, f) (rdev)->asic->copy_dma((rdev), (s), (d), (np), (f))
-#define radeon_copy(rdev, s, d, np, f) (rdev)->asic->copy((rdev), (s), (d), (np), (f))
-#define radeon_get_engine_clock(rdev) (rdev)->asic->get_engine_clock((rdev))
-#define radeon_set_engine_clock(rdev, e) (rdev)->asic->set_engine_clock((rdev), (e))
-#define radeon_get_memory_clock(rdev) (rdev)->asic->get_memory_clock((rdev))
-#define radeon_set_memory_clock(rdev, e) (rdev)->asic->set_memory_clock((rdev), (e))
-#define radeon_get_pcie_lanes(rdev) (rdev)->asic->get_pcie_lanes((rdev))
-#define radeon_set_pcie_lanes(rdev, l) (rdev)->asic->set_pcie_lanes((rdev), (l))
-#define radeon_set_clock_gating(rdev, e) (rdev)->asic->set_clock_gating((rdev), (e))
-#define radeon_set_surface_reg(rdev, r, f, p, o, s) ((rdev)->asic->set_surface_reg((rdev), (r), (f), (p), (o), (s)))
-#define radeon_clear_surface_reg(rdev, r) ((rdev)->asic->clear_surface_reg((rdev), (r)))
-#define radeon_bandwidth_update(rdev) (rdev)->asic->bandwidth_update((rdev))
-#define radeon_hpd_init(rdev) (rdev)->asic->hpd_init((rdev))
-#define radeon_hpd_fini(rdev) (rdev)->asic->hpd_fini((rdev))
-#define radeon_hpd_sense(rdev, hpd) (rdev)->asic->hpd_sense((rdev), (hpd))
-#define radeon_hpd_set_polarity(rdev, hpd) (rdev)->asic->hpd_set_polarity((rdev), (hpd))
+#define radeon_copy_blit(rdev, s, d, np, f) (rdev)->asic->copy.blit((rdev), (s), (d), (np), (f))
+#define radeon_copy_dma(rdev, s, d, np, f) (rdev)->asic->copy.dma((rdev), (s), (d), (np), (f))
+#define radeon_copy(rdev, s, d, np, f) (rdev)->asic->copy.copy((rdev), (s), (d), (np), (f))
+#define radeon_copy_blit_ring_index(rdev) (rdev)->asic->copy.blit_ring_index
+#define radeon_copy_dma_ring_index(rdev) (rdev)->asic->copy.dma_ring_index
+#define radeon_copy_ring_index(rdev) (rdev)->asic->copy.copy_ring_index
+#define radeon_get_engine_clock(rdev) (rdev)->asic->pm.get_engine_clock((rdev))
+#define radeon_set_engine_clock(rdev, e) (rdev)->asic->pm.set_engine_clock((rdev), (e))
+#define radeon_get_memory_clock(rdev) (rdev)->asic->pm.get_memory_clock((rdev))
+#define radeon_set_memory_clock(rdev, e) (rdev)->asic->pm.set_memory_clock((rdev), (e))
+#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_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))
+#define radeon_hpd_init(rdev) (rdev)->asic->hpd.init((rdev))
+#define radeon_hpd_fini(rdev) (rdev)->asic->hpd.fini((rdev))
+#define radeon_hpd_sense(rdev, h) (rdev)->asic->hpd.sense((rdev), (h))
+#define radeon_hpd_set_polarity(rdev, h) (rdev)->asic->hpd.set_polarity((rdev), (h))
 #define radeon_gui_idle(rdev) (rdev)->asic->gui_idle((rdev))
-#define radeon_pm_misc(rdev) (rdev)->asic->pm_misc((rdev))
-#define radeon_pm_prepare(rdev) (rdev)->asic->pm_prepare((rdev))
-#define radeon_pm_finish(rdev) (rdev)->asic->pm_finish((rdev))
-#define radeon_pm_init_profile(rdev) (rdev)->asic->pm_init_profile((rdev))
-#define radeon_pm_get_dynpm_state(rdev) (rdev)->asic->pm_get_dynpm_state((rdev))
-#define radeon_pre_page_flip(rdev, crtc) rdev->asic->pre_page_flip((rdev), (crtc))
-#define radeon_page_flip(rdev, crtc, base) rdev->asic->page_flip((rdev), (crtc), (base))
-#define radeon_post_page_flip(rdev, crtc) rdev->asic->post_page_flip((rdev), (crtc))
+#define radeon_pm_misc(rdev) (rdev)->asic->pm.misc((rdev))
+#define radeon_pm_prepare(rdev) (rdev)->asic->pm.prepare((rdev))
+#define radeon_pm_finish(rdev) (rdev)->asic->pm.finish((rdev))
+#define radeon_pm_init_profile(rdev) (rdev)->asic->pm.init_profile((rdev))
+#define radeon_pm_get_dynpm_state(rdev) (rdev)->asic->pm.get_dynpm_state((rdev))
+#define radeon_pre_page_flip(rdev, crtc) rdev->asic->pflip.pre_page_flip((rdev), (crtc))
+#define radeon_page_flip(rdev, crtc, base) rdev->asic->pflip.page_flip((rdev), (crtc), (base))
+#define radeon_post_page_flip(rdev, crtc) rdev->asic->pflip.post_page_flip((rdev), (crtc))
+#define radeon_wait_for_vblank(rdev, crtc) rdev->asic->display.wait_for_vblank((rdev), (crtc))
+#define radeon_mc_wait_for_idle(rdev) rdev->asic->mc_wait_for_idle((rdev))
 
 /* Common functions */
 /* AGP */
@@ -1750,6 +1836,16 @@ int r600_vram_scratch_init(struct radeon_device *rdev);
 void r600_vram_scratch_fini(struct radeon_device *rdev);
 
 /*
+ * r600 cs checking helper
+ */
+unsigned r600_mip_minify(unsigned size, unsigned level);
+bool r600_fmt_is_valid_color(u32 format);
+bool r600_fmt_is_valid_texture(u32 format, enum radeon_family family);
+int r600_fmt_get_blocksize(u32 format);
+int r600_fmt_get_nblocksx(u32 format, u32 w);
+int r600_fmt_get_nblocksy(u32 format, u32 h);
+
+/*
  * r600 functions used by radeon_encoder.c
  */
 extern void r600_hdmi_enable(struct drm_encoder *encoder);
diff --git a/drivers/gpu/drm/radeon/radeon_asic.c b/drivers/gpu/drm/radeon/radeon_asic.c
index 36a6192ce862..be4dc2ff0e40 100644
--- a/drivers/gpu/drm/radeon/radeon_asic.c
+++ b/drivers/gpu/drm/radeon/radeon_asic.c
@@ -114,13 +114,13 @@ void radeon_agp_disable(struct radeon_device *rdev)
 			rdev->family == CHIP_R423) {
 		DRM_INFO("Forcing AGP to PCIE mode\n");
 		rdev->flags |= RADEON_IS_PCIE;
-		rdev->asic->gart_tlb_flush = &rv370_pcie_gart_tlb_flush;
-		rdev->asic->gart_set_page = &rv370_pcie_gart_set_page;
+		rdev->asic->gart.tlb_flush = &rv370_pcie_gart_tlb_flush;
+		rdev->asic->gart.set_page = &rv370_pcie_gart_set_page;
 	} else {
 		DRM_INFO("Forcing AGP to PCI mode\n");
 		rdev->flags |= RADEON_IS_PCI;
-		rdev->asic->gart_tlb_flush = &r100_pci_gart_tlb_flush;
-		rdev->asic->gart_set_page = &r100_pci_gart_set_page;
+		rdev->asic->gart.tlb_flush = &r100_pci_gart_tlb_flush;
+		rdev->asic->gart.set_page = &r100_pci_gart_set_page;
 	}
 	rdev->mc.gtt_size = radeon_gart_size * 1024 * 1024;
 }
@@ -136,48 +136,70 @@ static struct radeon_asic r100_asic = {
 	.vga_set_state = &r100_vga_set_state,
 	.gpu_is_lockup = &r100_gpu_is_lockup,
 	.asic_reset = &r100_asic_reset,
-	.gart_tlb_flush = &r100_pci_gart_tlb_flush,
-	.gart_set_page = &r100_pci_gart_set_page,
-	.ring_start = &r100_ring_start,
-	.ring_test = &r100_ring_test,
+	.ioctl_wait_idle = NULL,
+	.gui_idle = &r100_gui_idle,
+	.mc_wait_for_idle = &r100_mc_wait_for_idle,
+	.gart = {
+		.tlb_flush = &r100_pci_gart_tlb_flush,
+		.set_page = &r100_pci_gart_set_page,
+	},
 	.ring = {
 		[RADEON_RING_TYPE_GFX_INDEX] = {
 			.ib_execute = &r100_ring_ib_execute,
 			.emit_fence = &r100_fence_ring_emit,
 			.emit_semaphore = &r100_semaphore_ring_emit,
+			.cs_parse = &r100_cs_parse,
+			.ring_start = &r100_ring_start,
+			.ring_test = &r100_ring_test,
+			.ib_test = &r100_ib_test,
 		}
 	},
-	.irq_set = &r100_irq_set,
-	.irq_process = &r100_irq_process,
-	.get_vblank_counter = &r100_get_vblank_counter,
-	.cs_parse = &r100_cs_parse,
-	.copy_blit = &r100_copy_blit,
-	.copy_dma = NULL,
-	.copy = &r100_copy_blit,
-	.get_engine_clock = &radeon_legacy_get_engine_clock,
-	.set_engine_clock = &radeon_legacy_set_engine_clock,
-	.get_memory_clock = &radeon_legacy_get_memory_clock,
-	.set_memory_clock = NULL,
-	.get_pcie_lanes = NULL,
-	.set_pcie_lanes = NULL,
-	.set_clock_gating = &radeon_legacy_set_clock_gating,
-	.set_surface_reg = r100_set_surface_reg,
-	.clear_surface_reg = r100_clear_surface_reg,
-	.bandwidth_update = &r100_bandwidth_update,
-	.hpd_init = &r100_hpd_init,
-	.hpd_fini = &r100_hpd_fini,
-	.hpd_sense = &r100_hpd_sense,
-	.hpd_set_polarity = &r100_hpd_set_polarity,
-	.ioctl_wait_idle = NULL,
-	.gui_idle = &r100_gui_idle,
-	.pm_misc = &r100_pm_misc,
-	.pm_prepare = &r100_pm_prepare,
-	.pm_finish = &r100_pm_finish,
-	.pm_init_profile = &r100_pm_init_profile,
-	.pm_get_dynpm_state = &r100_pm_get_dynpm_state,
-	.pre_page_flip = &r100_pre_page_flip,
-	.page_flip = &r100_page_flip,
-	.post_page_flip = &r100_post_page_flip,
+	.irq = {
+		.set = &r100_irq_set,
+		.process = &r100_irq_process,
+	},
+	.display = {
+		.bandwidth_update = &r100_bandwidth_update,
+		.get_vblank_counter = &r100_get_vblank_counter,
+		.wait_for_vblank = &r100_wait_for_vblank,
+	},
+	.copy = {
+		.blit = &r100_copy_blit,
+		.blit_ring_index = RADEON_RING_TYPE_GFX_INDEX,
+		.dma = NULL,
+		.dma_ring_index = RADEON_RING_TYPE_GFX_INDEX,
+		.copy = &r100_copy_blit,
+		.copy_ring_index = RADEON_RING_TYPE_GFX_INDEX,
+	},
+	.surface = {
+		.set_reg = r100_set_surface_reg,
+		.clear_reg = r100_clear_surface_reg,
+	},
+	.hpd = {
+		.init = &r100_hpd_init,
+		.fini = &r100_hpd_fini,
+		.sense = &r100_hpd_sense,
+		.set_polarity = &r100_hpd_set_polarity,
+	},
+	.pm = {
+		.misc = &r100_pm_misc,
+		.prepare = &r100_pm_prepare,
+		.finish = &r100_pm_finish,
+		.init_profile = &r100_pm_init_profile,
+		.get_dynpm_state = &r100_pm_get_dynpm_state,
+		.get_engine_clock = &radeon_legacy_get_engine_clock,
+		.set_engine_clock = &radeon_legacy_set_engine_clock,
+		.get_memory_clock = &radeon_legacy_get_memory_clock,
+		.set_memory_clock = NULL,
+		.get_pcie_lanes = NULL,
+		.set_pcie_lanes = NULL,
+		.set_clock_gating = &radeon_legacy_set_clock_gating,
+	},
+	.pflip = {
+		.pre_page_flip = &r100_pre_page_flip,
+		.page_flip = &r100_page_flip,
+		.post_page_flip = &r100_post_page_flip,
+	},
 };
 
 static struct radeon_asic r200_asic = {
@@ -188,47 +210,70 @@ static struct radeon_asic r200_asic = {
 	.vga_set_state = &r100_vga_set_state,
 	.gpu_is_lockup = &r100_gpu_is_lockup,
 	.asic_reset = &r100_asic_reset,
-	.gart_tlb_flush = &r100_pci_gart_tlb_flush,
-	.gart_set_page = &r100_pci_gart_set_page,
-	.ring_start = &r100_ring_start,
-	.ring_test = &r100_ring_test,
+	.ioctl_wait_idle = NULL,
+	.gui_idle = &r100_gui_idle,
+	.mc_wait_for_idle = &r100_mc_wait_for_idle,
+	.gart = {
+		.tlb_flush = &r100_pci_gart_tlb_flush,
+		.set_page = &r100_pci_gart_set_page,
+	},
 	.ring = {
 		[RADEON_RING_TYPE_GFX_INDEX] = {
 			.ib_execute = &r100_ring_ib_execute,
 			.emit_fence = &r100_fence_ring_emit,
 			.emit_semaphore = &r100_semaphore_ring_emit,
+			.cs_parse = &r100_cs_parse,
+			.ring_start = &r100_ring_start,
+			.ring_test = &r100_ring_test,
+			.ib_test = &r100_ib_test,
 		}
 	},
-	.irq_set = &r100_irq_set,
-	.irq_process = &r100_irq_process,
-	.get_vblank_counter = &r100_get_vblank_counter,
-	.cs_parse = &r100_cs_parse,
-	.copy_blit = &r100_copy_blit,
-	.copy_dma = &r200_copy_dma,
-	.copy = &r100_copy_blit,
-	.get_engine_clock = &radeon_legacy_get_engine_clock,
-	.set_engine_clock = &radeon_legacy_set_engine_clock,
-	.get_memory_clock = &radeon_legacy_get_memory_clock,
-	.set_memory_clock = NULL,
-	.set_pcie_lanes = NULL,
-	.set_clock_gating = &radeon_legacy_set_clock_gating,
-	.set_surface_reg = r100_set_surface_reg,
-	.clear_surface_reg = r100_clear_surface_reg,
-	.bandwidth_update = &r100_bandwidth_update,
-	.hpd_init = &r100_hpd_init,
-	.hpd_fini = &r100_hpd_fini,
-	.hpd_sense = &r100_hpd_sense,
-	.hpd_set_polarity = &r100_hpd_set_polarity,
-	.ioctl_wait_idle = NULL,
-	.gui_idle = &r100_gui_idle,
-	.pm_misc = &r100_pm_misc,
-	.pm_prepare = &r100_pm_prepare,
-	.pm_finish = &r100_pm_finish,
-	.pm_init_profile = &r100_pm_init_profile,
-	.pm_get_dynpm_state = &r100_pm_get_dynpm_state,
-	.pre_page_flip = &r100_pre_page_flip,
-	.page_flip = &r100_page_flip,
-	.post_page_flip = &r100_post_page_flip,
+	.irq = {
+		.set = &r100_irq_set,
+		.process = &r100_irq_process,
+	},
+	.display = {
+		.bandwidth_update = &r100_bandwidth_update,
+		.get_vblank_counter = &r100_get_vblank_counter,
+		.wait_for_vblank = &r100_wait_for_vblank,
+	},
+	.copy = {
+		.blit = &r100_copy_blit,
+		.blit_ring_index = RADEON_RING_TYPE_GFX_INDEX,
+		.dma = &r200_copy_dma,
+		.dma_ring_index = RADEON_RING_TYPE_GFX_INDEX,
+		.copy = &r100_copy_blit,
+		.copy_ring_index = RADEON_RING_TYPE_GFX_INDEX,
+	},
+	.surface = {
+		.set_reg = r100_set_surface_reg,
+		.clear_reg = r100_clear_surface_reg,
+	},
+	.hpd = {
+		.init = &r100_hpd_init,
+		.fini = &r100_hpd_fini,
+		.sense = &r100_hpd_sense,
+		.set_polarity = &r100_hpd_set_polarity,
+	},
+	.pm = {
+		.misc = &r100_pm_misc,
+		.prepare = &r100_pm_prepare,
+		.finish = &r100_pm_finish,
+		.init_profile = &r100_pm_init_profile,
+		.get_dynpm_state = &r100_pm_get_dynpm_state,
+		.get_engine_clock = &radeon_legacy_get_engine_clock,
+		.set_engine_clock = &radeon_legacy_set_engine_clock,
+		.get_memory_clock = &radeon_legacy_get_memory_clock,
+		.set_memory_clock = NULL,
+		.get_pcie_lanes = NULL,
+		.set_pcie_lanes = NULL,
+		.set_clock_gating = &radeon_legacy_set_clock_gating,
+	},
+	.pflip = {
+		.pre_page_flip = &r100_pre_page_flip,
+		.page_flip = &r100_page_flip,
+		.post_page_flip = &r100_post_page_flip,
+	},
 };
 
 static struct radeon_asic r300_asic = {
@@ -239,48 +284,70 @@ static struct radeon_asic r300_asic = {
 	.vga_set_state = &r100_vga_set_state,
 	.gpu_is_lockup = &r300_gpu_is_lockup,
 	.asic_reset = &r300_asic_reset,
-	.gart_tlb_flush = &r100_pci_gart_tlb_flush,
-	.gart_set_page = &r100_pci_gart_set_page,
-	.ring_start = &r300_ring_start,
-	.ring_test = &r100_ring_test,
+	.ioctl_wait_idle = NULL,
+	.gui_idle = &r100_gui_idle,
+	.mc_wait_for_idle = &r300_mc_wait_for_idle,
+	.gart = {
+		.tlb_flush = &r100_pci_gart_tlb_flush,
+		.set_page = &r100_pci_gart_set_page,
+	},
 	.ring = {
 		[RADEON_RING_TYPE_GFX_INDEX] = {
 			.ib_execute = &r100_ring_ib_execute,
 			.emit_fence = &r300_fence_ring_emit,
 			.emit_semaphore = &r100_semaphore_ring_emit,
+			.cs_parse = &r300_cs_parse,
+			.ring_start = &r300_ring_start,
+			.ring_test = &r100_ring_test,
+			.ib_test = &r100_ib_test,
 		}
 	},
-	.irq_set = &r100_irq_set,
-	.irq_process = &r100_irq_process,
-	.get_vblank_counter = &r100_get_vblank_counter,
-	.cs_parse = &r300_cs_parse,
-	.copy_blit = &r100_copy_blit,
-	.copy_dma = &r200_copy_dma,
-	.copy = &r100_copy_blit,
-	.get_engine_clock = &radeon_legacy_get_engine_clock,
-	.set_engine_clock = &radeon_legacy_set_engine_clock,
-	.get_memory_clock = &radeon_legacy_get_memory_clock,
-	.set_memory_clock = NULL,
-	.get_pcie_lanes = &rv370_get_pcie_lanes,
-	.set_pcie_lanes = &rv370_set_pcie_lanes,
-	.set_clock_gating = &radeon_legacy_set_clock_gating,
-	.set_surface_reg = r100_set_surface_reg,
-	.clear_surface_reg = r100_clear_surface_reg,
-	.bandwidth_update = &r100_bandwidth_update,
-	.hpd_init = &r100_hpd_init,
-	.hpd_fini = &r100_hpd_fini,
-	.hpd_sense = &r100_hpd_sense,
-	.hpd_set_polarity = &r100_hpd_set_polarity,
-	.ioctl_wait_idle = NULL,
-	.gui_idle = &r100_gui_idle,
-	.pm_misc = &r100_pm_misc,
-	.pm_prepare = &r100_pm_prepare,
-	.pm_finish = &r100_pm_finish,
-	.pm_init_profile = &r100_pm_init_profile,
-	.pm_get_dynpm_state = &r100_pm_get_dynpm_state,
-	.pre_page_flip = &r100_pre_page_flip,
-	.page_flip = &r100_page_flip,
-	.post_page_flip = &r100_post_page_flip,
+	.irq = {
+		.set = &r100_irq_set,
+		.process = &r100_irq_process,
+	},
+	.display = {
+		.bandwidth_update = &r100_bandwidth_update,
+		.get_vblank_counter = &r100_get_vblank_counter,
+		.wait_for_vblank = &r100_wait_for_vblank,
+	},
+	.copy = {
+		.blit = &r100_copy_blit,
+		.blit_ring_index = RADEON_RING_TYPE_GFX_INDEX,
+		.dma = &r200_copy_dma,
+		.dma_ring_index = RADEON_RING_TYPE_GFX_INDEX,
+		.copy = &r100_copy_blit,
+		.copy_ring_index = RADEON_RING_TYPE_GFX_INDEX,
+	},
+	.surface = {
+		.set_reg = r100_set_surface_reg,
+		.clear_reg = r100_clear_surface_reg,
+	},
+	.hpd = {
+		.init = &r100_hpd_init,
+		.fini = &r100_hpd_fini,
+		.sense = &r100_hpd_sense,
+		.set_polarity = &r100_hpd_set_polarity,
+	},
+	.pm = {
+		.misc = &r100_pm_misc,
+		.prepare = &r100_pm_prepare,
+		.finish = &r100_pm_finish,
+		.init_profile = &r100_pm_init_profile,
+		.get_dynpm_state = &r100_pm_get_dynpm_state,
+		.get_engine_clock = &radeon_legacy_get_engine_clock,
+		.set_engine_clock = &radeon_legacy_set_engine_clock,
+		.get_memory_clock = &radeon_legacy_get_memory_clock,
+		.set_memory_clock = NULL,
+		.get_pcie_lanes = &rv370_get_pcie_lanes,
+		.set_pcie_lanes = &rv370_set_pcie_lanes,
+		.set_clock_gating = &radeon_legacy_set_clock_gating,
+	},
+	.pflip = {
+		.pre_page_flip = &r100_pre_page_flip,
+		.page_flip = &r100_page_flip,
+		.post_page_flip = &r100_post_page_flip,
+	},
 };
 
 static struct radeon_asic r300_asic_pcie = {
@@ -291,47 +358,70 @@ static struct radeon_asic r300_asic_pcie = {
 	.vga_set_state = &r100_vga_set_state,
 	.gpu_is_lockup = &r300_gpu_is_lockup,
 	.asic_reset = &r300_asic_reset,
-	.gart_tlb_flush = &rv370_pcie_gart_tlb_flush,
-	.gart_set_page = &rv370_pcie_gart_set_page,
-	.ring_start = &r300_ring_start,
-	.ring_test = &r100_ring_test,
+	.ioctl_wait_idle = NULL,
+	.gui_idle = &r100_gui_idle,
+	.mc_wait_for_idle = &r300_mc_wait_for_idle,
+	.gart = {
+		.tlb_flush = &rv370_pcie_gart_tlb_flush,
+		.set_page = &rv370_pcie_gart_set_page,
+	},
 	.ring = {
 		[RADEON_RING_TYPE_GFX_INDEX] = {
 			.ib_execute = &r100_ring_ib_execute,
 			.emit_fence = &r300_fence_ring_emit,
 			.emit_semaphore = &r100_semaphore_ring_emit,
+			.cs_parse = &r300_cs_parse,
+			.ring_start = &r300_ring_start,
+			.ring_test = &r100_ring_test,
+			.ib_test = &r100_ib_test,
 		}
 	},
-	.irq_set = &r100_irq_set,
-	.irq_process = &r100_irq_process,
-	.get_vblank_counter = &r100_get_vblank_counter,
-	.cs_parse = &r300_cs_parse,
-	.copy_blit = &r100_copy_blit,
-	.copy_dma = &r200_copy_dma,
-	.copy = &r100_copy_blit,
-	.get_engine_clock = &radeon_legacy_get_engine_clock,
-	.set_engine_clock = &radeon_legacy_set_engine_clock,
-	.get_memory_clock = &radeon_legacy_get_memory_clock,
-	.set_memory_clock = NULL,
-	.set_pcie_lanes = &rv370_set_pcie_lanes,
-	.set_clock_gating = &radeon_legacy_set_clock_gating,
-	.set_surface_reg = r100_set_surface_reg,
-	.clear_surface_reg = r100_clear_surface_reg,
-	.bandwidth_update = &r100_bandwidth_update,
-	.hpd_init = &r100_hpd_init,
-	.hpd_fini = &r100_hpd_fini,
-	.hpd_sense = &r100_hpd_sense,
-	.hpd_set_polarity = &r100_hpd_set_polarity,
-	.ioctl_wait_idle = NULL,
-	.gui_idle = &r100_gui_idle,
-	.pm_misc = &r100_pm_misc,
-	.pm_prepare = &r100_pm_prepare,
-	.pm_finish = &r100_pm_finish,
-	.pm_init_profile = &r100_pm_init_profile,
-	.pm_get_dynpm_state = &r100_pm_get_dynpm_state,
-	.pre_page_flip = &r100_pre_page_flip,
-	.page_flip = &r100_page_flip,
-	.post_page_flip = &r100_post_page_flip,
+	.irq = {
+		.set = &r100_irq_set,
+		.process = &r100_irq_process,
+	},
+	.display = {
+		.bandwidth_update = &r100_bandwidth_update,
+		.get_vblank_counter = &r100_get_vblank_counter,
+		.wait_for_vblank = &r100_wait_for_vblank,
+	},
+	.copy = {
+		.blit = &r100_copy_blit,
+		.blit_ring_index = RADEON_RING_TYPE_GFX_INDEX,
+		.dma = &r200_copy_dma,
+		.dma_ring_index = RADEON_RING_TYPE_GFX_INDEX,
+		.copy = &r100_copy_blit,
+		.copy_ring_index = RADEON_RING_TYPE_GFX_INDEX,
+	},
+	.surface = {
+		.set_reg = r100_set_surface_reg,
+		.clear_reg = r100_clear_surface_reg,
+	},
+	.hpd = {
+		.init = &r100_hpd_init,
+		.fini = &r100_hpd_fini,
+		.sense = &r100_hpd_sense,
+		.set_polarity = &r100_hpd_set_polarity,
+	},
+	.pm = {
+		.misc = &r100_pm_misc,
+		.prepare = &r100_pm_prepare,
+		.finish = &r100_pm_finish,
+		.init_profile = &r100_pm_init_profile,
+		.get_dynpm_state = &r100_pm_get_dynpm_state,
+		.get_engine_clock = &radeon_legacy_get_engine_clock,
+		.set_engine_clock = &radeon_legacy_set_engine_clock,
+		.get_memory_clock = &radeon_legacy_get_memory_clock,
+		.set_memory_clock = NULL,
+		.get_pcie_lanes = &rv370_get_pcie_lanes,
+		.set_pcie_lanes = &rv370_set_pcie_lanes,
+		.set_clock_gating = &radeon_legacy_set_clock_gating,
+	},
+	.pflip = {
+		.pre_page_flip = &r100_pre_page_flip,
+		.page_flip = &r100_page_flip,
+		.post_page_flip = &r100_post_page_flip,
+	},
 };
 
 static struct radeon_asic r420_asic = {
@@ -342,48 +432,70 @@ static struct radeon_asic r420_asic = {
 	.vga_set_state = &r100_vga_set_state,
 	.gpu_is_lockup = &r300_gpu_is_lockup,
 	.asic_reset = &r300_asic_reset,
-	.gart_tlb_flush = &rv370_pcie_gart_tlb_flush,
-	.gart_set_page = &rv370_pcie_gart_set_page,
-	.ring_start = &r300_ring_start,
-	.ring_test = &r100_ring_test,
+	.ioctl_wait_idle = NULL,
+	.gui_idle = &r100_gui_idle,
+	.mc_wait_for_idle = &r300_mc_wait_for_idle,
+	.gart = {
+		.tlb_flush = &rv370_pcie_gart_tlb_flush,
+		.set_page = &rv370_pcie_gart_set_page,
+	},
 	.ring = {
 		[RADEON_RING_TYPE_GFX_INDEX] = {
 			.ib_execute = &r100_ring_ib_execute,
 			.emit_fence = &r300_fence_ring_emit,
 			.emit_semaphore = &r100_semaphore_ring_emit,
+			.cs_parse = &r300_cs_parse,
+			.ring_start = &r300_ring_start,
+			.ring_test = &r100_ring_test,
+			.ib_test = &r100_ib_test,
 		}
 	},
-	.irq_set = &r100_irq_set,
-	.irq_process = &r100_irq_process,
-	.get_vblank_counter = &r100_get_vblank_counter,
-	.cs_parse = &r300_cs_parse,
-	.copy_blit = &r100_copy_blit,
-	.copy_dma = &r200_copy_dma,
-	.copy = &r100_copy_blit,
-	.get_engine_clock = &radeon_atom_get_engine_clock,
-	.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 = &rv370_get_pcie_lanes,
-	.set_pcie_lanes = &rv370_set_pcie_lanes,
-	.set_clock_gating = &radeon_atom_set_clock_gating,
-	.set_surface_reg = r100_set_surface_reg,
-	.clear_surface_reg = r100_clear_surface_reg,
-	.bandwidth_update = &r100_bandwidth_update,
-	.hpd_init = &r100_hpd_init,
-	.hpd_fini = &r100_hpd_fini,
-	.hpd_sense = &r100_hpd_sense,
-	.hpd_set_polarity = &r100_hpd_set_polarity,
-	.ioctl_wait_idle = NULL,
-	.gui_idle = &r100_gui_idle,
-	.pm_misc = &r100_pm_misc,
-	.pm_prepare = &r100_pm_prepare,
-	.pm_finish = &r100_pm_finish,
-	.pm_init_profile = &r420_pm_init_profile,
-	.pm_get_dynpm_state = &r100_pm_get_dynpm_state,
-	.pre_page_flip = &r100_pre_page_flip,
-	.page_flip = &r100_page_flip,
-	.post_page_flip = &r100_post_page_flip,
+	.irq = {
+		.set = &r100_irq_set,
+		.process = &r100_irq_process,
+	},
+	.display = {
+		.bandwidth_update = &r100_bandwidth_update,
+		.get_vblank_counter = &r100_get_vblank_counter,
+		.wait_for_vblank = &r100_wait_for_vblank,
+	},
+	.copy = {
+		.blit = &r100_copy_blit,
+		.blit_ring_index = RADEON_RING_TYPE_GFX_INDEX,
+		.dma = &r200_copy_dma,
+		.dma_ring_index = RADEON_RING_TYPE_GFX_INDEX,
+		.copy = &r100_copy_blit,
+		.copy_ring_index = RADEON_RING_TYPE_GFX_INDEX,
+	},
+	.surface = {
+		.set_reg = r100_set_surface_reg,
+		.clear_reg = r100_clear_surface_reg,
+	},
+	.hpd = {
+		.init = &r100_hpd_init,
+		.fini = &r100_hpd_fini,
+		.sense = &r100_hpd_sense,
+		.set_polarity = &r100_hpd_set_polarity,
+	},
+	.pm = {
+		.misc = &r100_pm_misc,
+		.prepare = &r100_pm_prepare,
+		.finish = &r100_pm_finish,
+		.init_profile = &r420_pm_init_profile,
+		.get_dynpm_state = &r100_pm_get_dynpm_state,
+		.get_engine_clock = &radeon_atom_get_engine_clock,
+		.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 = &rv370_get_pcie_lanes,
+		.set_pcie_lanes = &rv370_set_pcie_lanes,
+		.set_clock_gating = &radeon_atom_set_clock_gating,
+	},
+	.pflip = {
+		.pre_page_flip = &r100_pre_page_flip,
+		.page_flip = &r100_page_flip,
+		.post_page_flip = &r100_post_page_flip,
+	},
 };
 
 static struct radeon_asic rs400_asic = {
@@ -394,48 +506,70 @@ static struct radeon_asic rs400_asic = {
 	.vga_set_state = &r100_vga_set_state,
 	.gpu_is_lockup = &r300_gpu_is_lockup,
 	.asic_reset = &r300_asic_reset,
-	.gart_tlb_flush = &rs400_gart_tlb_flush,
-	.gart_set_page = &rs400_gart_set_page,
-	.ring_start = &r300_ring_start,
-	.ring_test = &r100_ring_test,
+	.ioctl_wait_idle = NULL,
+	.gui_idle = &r100_gui_idle,
+	.mc_wait_for_idle = &rs400_mc_wait_for_idle,
+	.gart = {
+		.tlb_flush = &rs400_gart_tlb_flush,
+		.set_page = &rs400_gart_set_page,
+	},
 	.ring = {
 		[RADEON_RING_TYPE_GFX_INDEX] = {
 			.ib_execute = &r100_ring_ib_execute,
 			.emit_fence = &r300_fence_ring_emit,
 			.emit_semaphore = &r100_semaphore_ring_emit,
+			.cs_parse = &r300_cs_parse,
+			.ring_start = &r300_ring_start,
+			.ring_test = &r100_ring_test,
+			.ib_test = &r100_ib_test,
 		}
 	},
-	.irq_set = &r100_irq_set,
-	.irq_process = &r100_irq_process,
-	.get_vblank_counter = &r100_get_vblank_counter,
-	.cs_parse = &r300_cs_parse,
-	.copy_blit = &r100_copy_blit,
-	.copy_dma = &r200_copy_dma,
-	.copy = &r100_copy_blit,
-	.get_engine_clock = &radeon_legacy_get_engine_clock,
-	.set_engine_clock = &radeon_legacy_set_engine_clock,
-	.get_memory_clock = &radeon_legacy_get_memory_clock,
-	.set_memory_clock = NULL,
-	.get_pcie_lanes = NULL,
-	.set_pcie_lanes = NULL,
-	.set_clock_gating = &radeon_legacy_set_clock_gating,
-	.set_surface_reg = r100_set_surface_reg,
-	.clear_surface_reg = r100_clear_surface_reg,
-	.bandwidth_update = &r100_bandwidth_update,
-	.hpd_init = &r100_hpd_init,
-	.hpd_fini = &r100_hpd_fini,
-	.hpd_sense = &r100_hpd_sense,
-	.hpd_set_polarity = &r100_hpd_set_polarity,
-	.ioctl_wait_idle = NULL,
-	.gui_idle = &r100_gui_idle,
-	.pm_misc = &r100_pm_misc,
-	.pm_prepare = &r100_pm_prepare,
-	.pm_finish = &r100_pm_finish,
-	.pm_init_profile = &r100_pm_init_profile,
-	.pm_get_dynpm_state = &r100_pm_get_dynpm_state,
-	.pre_page_flip = &r100_pre_page_flip,
-	.page_flip = &r100_page_flip,
-	.post_page_flip = &r100_post_page_flip,
+	.irq = {
+		.set = &r100_irq_set,
+		.process = &r100_irq_process,
+	},
+	.display = {
+		.bandwidth_update = &r100_bandwidth_update,
+		.get_vblank_counter = &r100_get_vblank_counter,
+		.wait_for_vblank = &r100_wait_for_vblank,
+	},
+	.copy = {
+		.blit = &r100_copy_blit,
+		.blit_ring_index = RADEON_RING_TYPE_GFX_INDEX,
+		.dma = &r200_copy_dma,
+		.dma_ring_index = RADEON_RING_TYPE_GFX_INDEX,
+		.copy = &r100_copy_blit,
+		.copy_ring_index = RADEON_RING_TYPE_GFX_INDEX,
+	},
+	.surface = {
+		.set_reg = r100_set_surface_reg,
+		.clear_reg = r100_clear_surface_reg,
+	},
+	.hpd = {
+		.init = &r100_hpd_init,
+		.fini = &r100_hpd_fini,
+		.sense = &r100_hpd_sense,
+		.set_polarity = &r100_hpd_set_polarity,
+	},
+	.pm = {
+		.misc = &r100_pm_misc,
+		.prepare = &r100_pm_prepare,
+		.finish = &r100_pm_finish,
+		.init_profile = &r100_pm_init_profile,
+		.get_dynpm_state = &r100_pm_get_dynpm_state,
+		.get_engine_clock = &radeon_legacy_get_engine_clock,
+		.set_engine_clock = &radeon_legacy_set_engine_clock,
+		.get_memory_clock = &radeon_legacy_get_memory_clock,
+		.set_memory_clock = NULL,
+		.get_pcie_lanes = NULL,
+		.set_pcie_lanes = NULL,
+		.set_clock_gating = &radeon_legacy_set_clock_gating,
+	},
+	.pflip = {
+		.pre_page_flip = &r100_pre_page_flip,
+		.page_flip = &r100_page_flip,
+		.post_page_flip = &r100_post_page_flip,
+	},
 };
 
 static struct radeon_asic rs600_asic = {
@@ -446,48 +580,70 @@ static struct radeon_asic rs600_asic = {
 	.vga_set_state = &r100_vga_set_state,
 	.gpu_is_lockup = &r300_gpu_is_lockup,
 	.asic_reset = &rs600_asic_reset,
-	.gart_tlb_flush = &rs600_gart_tlb_flush,
-	.gart_set_page = &rs600_gart_set_page,
-	.ring_start = &r300_ring_start,
-	.ring_test = &r100_ring_test,
+	.ioctl_wait_idle = NULL,
+	.gui_idle = &r100_gui_idle,
+	.mc_wait_for_idle = &rs600_mc_wait_for_idle,
+	.gart = {
+		.tlb_flush = &rs600_gart_tlb_flush,
+		.set_page = &rs600_gart_set_page,
+	},
 	.ring = {
 		[RADEON_RING_TYPE_GFX_INDEX] = {
 			.ib_execute = &r100_ring_ib_execute,
 			.emit_fence = &r300_fence_ring_emit,
 			.emit_semaphore = &r100_semaphore_ring_emit,
+			.cs_parse = &r300_cs_parse,
+			.ring_start = &r300_ring_start,
+			.ring_test = &r100_ring_test,
+			.ib_test = &r100_ib_test,
 		}
 	},
-	.irq_set = &rs600_irq_set,
-	.irq_process = &rs600_irq_process,
-	.get_vblank_counter = &rs600_get_vblank_counter,
-	.cs_parse = &r300_cs_parse,
-	.copy_blit = &r100_copy_blit,
-	.copy_dma = &r200_copy_dma,
-	.copy = &r100_copy_blit,
-	.get_engine_clock = &radeon_atom_get_engine_clock,
-	.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,
-	.set_clock_gating = &radeon_atom_set_clock_gating,
-	.set_surface_reg = r100_set_surface_reg,
-	.clear_surface_reg = r100_clear_surface_reg,
-	.bandwidth_update = &rs600_bandwidth_update,
-	.hpd_init = &rs600_hpd_init,
-	.hpd_fini = &rs600_hpd_fini,
-	.hpd_sense = &rs600_hpd_sense,
-	.hpd_set_polarity = &rs600_hpd_set_polarity,
-	.ioctl_wait_idle = NULL,
-	.gui_idle = &r100_gui_idle,
-	.pm_misc = &rs600_pm_misc,
-	.pm_prepare = &rs600_pm_prepare,
-	.pm_finish = &rs600_pm_finish,
-	.pm_init_profile = &r420_pm_init_profile,
-	.pm_get_dynpm_state = &r100_pm_get_dynpm_state,
-	.pre_page_flip = &rs600_pre_page_flip,
-	.page_flip = &rs600_page_flip,
-	.post_page_flip = &rs600_post_page_flip,
+	.irq = {
+		.set = &rs600_irq_set,
+		.process = &rs600_irq_process,
+	},
+	.display = {
+		.bandwidth_update = &rs600_bandwidth_update,
+		.get_vblank_counter = &rs600_get_vblank_counter,
+		.wait_for_vblank = &avivo_wait_for_vblank,
+	},
+	.copy = {
+		.blit = &r100_copy_blit,
+		.blit_ring_index = RADEON_RING_TYPE_GFX_INDEX,
+		.dma = &r200_copy_dma,
+		.dma_ring_index = RADEON_RING_TYPE_GFX_INDEX,
+		.copy = &r100_copy_blit,
+		.copy_ring_index = RADEON_RING_TYPE_GFX_INDEX,
+	},
+	.surface = {
+		.set_reg = r100_set_surface_reg,
+		.clear_reg = r100_clear_surface_reg,
+	},
+	.hpd = {
+		.init = &rs600_hpd_init,
+		.fini = &rs600_hpd_fini,
+		.sense = &rs600_hpd_sense,
+		.set_polarity = &rs600_hpd_set_polarity,
+	},
+	.pm = {
+		.misc = &rs600_pm_misc,
+		.prepare = &rs600_pm_prepare,
+		.finish = &rs600_pm_finish,
+		.init_profile = &r420_pm_init_profile,
+		.get_dynpm_state = &r100_pm_get_dynpm_state,
+		.get_engine_clock = &radeon_atom_get_engine_clock,
+		.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,
+		.set_clock_gating = &radeon_atom_set_clock_gating,
+	},
+	.pflip = {
+		.pre_page_flip = &rs600_pre_page_flip,
+		.page_flip = &rs600_page_flip,
+		.post_page_flip = &rs600_post_page_flip,
+	},
 };
 
 static struct radeon_asic rs690_asic = {
@@ -498,48 +654,70 @@ static struct radeon_asic rs690_asic = {
 	.vga_set_state = &r100_vga_set_state,
 	.gpu_is_lockup = &r300_gpu_is_lockup,
 	.asic_reset = &rs600_asic_reset,
-	.gart_tlb_flush = &rs400_gart_tlb_flush,
-	.gart_set_page = &rs400_gart_set_page,
-	.ring_start = &r300_ring_start,
-	.ring_test = &r100_ring_test,
+	.ioctl_wait_idle = NULL,
+	.gui_idle = &r100_gui_idle,
+	.mc_wait_for_idle = &rs690_mc_wait_for_idle,
+	.gart = {
+		.tlb_flush = &rs400_gart_tlb_flush,
+		.set_page = &rs400_gart_set_page,
+	},
 	.ring = {
 		[RADEON_RING_TYPE_GFX_INDEX] = {
 			.ib_execute = &r100_ring_ib_execute,
 			.emit_fence = &r300_fence_ring_emit,
 			.emit_semaphore = &r100_semaphore_ring_emit,
+			.cs_parse = &r300_cs_parse,
+			.ring_start = &r300_ring_start,
+			.ring_test = &r100_ring_test,
+			.ib_test = &r100_ib_test,
 		}
 	},
-	.irq_set = &rs600_irq_set,
-	.irq_process = &rs600_irq_process,
-	.get_vblank_counter = &rs600_get_vblank_counter,
-	.cs_parse = &r300_cs_parse,
-	.copy_blit = &r100_copy_blit,
-	.copy_dma = &r200_copy_dma,
-	.copy = &r200_copy_dma,
-	.get_engine_clock = &radeon_atom_get_engine_clock,
-	.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,
-	.set_clock_gating = &radeon_atom_set_clock_gating,
-	.set_surface_reg = r100_set_surface_reg,
-	.clear_surface_reg = r100_clear_surface_reg,
-	.bandwidth_update = &rs690_bandwidth_update,
-	.hpd_init = &rs600_hpd_init,
-	.hpd_fini = &rs600_hpd_fini,
-	.hpd_sense = &rs600_hpd_sense,
-	.hpd_set_polarity = &rs600_hpd_set_polarity,
-	.ioctl_wait_idle = NULL,
-	.gui_idle = &r100_gui_idle,
-	.pm_misc = &rs600_pm_misc,
-	.pm_prepare = &rs600_pm_prepare,
-	.pm_finish = &rs600_pm_finish,
-	.pm_init_profile = &r420_pm_init_profile,
-	.pm_get_dynpm_state = &r100_pm_get_dynpm_state,
-	.pre_page_flip = &rs600_pre_page_flip,
-	.page_flip = &rs600_page_flip,
-	.post_page_flip = &rs600_post_page_flip,
+	.irq = {
+		.set = &rs600_irq_set,
+		.process = &rs600_irq_process,
+	},
+	.display = {
+		.get_vblank_counter = &rs600_get_vblank_counter,
+		.bandwidth_update = &rs690_bandwidth_update,
+		.wait_for_vblank = &avivo_wait_for_vblank,
+	},
+	.copy = {
+		.blit = &r100_copy_blit,
+		.blit_ring_index = RADEON_RING_TYPE_GFX_INDEX,
+		.dma = &r200_copy_dma,
+		.dma_ring_index = RADEON_RING_TYPE_GFX_INDEX,
+		.copy = &r200_copy_dma,
+		.copy_ring_index = RADEON_RING_TYPE_GFX_INDEX,
+	},
+	.surface = {
+		.set_reg = r100_set_surface_reg,
+		.clear_reg = r100_clear_surface_reg,
+	},
+	.hpd = {
+		.init = &rs600_hpd_init,
+		.fini = &rs600_hpd_fini,
+		.sense = &rs600_hpd_sense,
+		.set_polarity = &rs600_hpd_set_polarity,
+	},
+	.pm = {
+		.misc = &rs600_pm_misc,
+		.prepare = &rs600_pm_prepare,
+		.finish = &rs600_pm_finish,
+		.init_profile = &r420_pm_init_profile,
+		.get_dynpm_state = &r100_pm_get_dynpm_state,
+		.get_engine_clock = &radeon_atom_get_engine_clock,
+		.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,
+		.set_clock_gating = &radeon_atom_set_clock_gating,
+	},
+	.pflip = {
+		.pre_page_flip = &rs600_pre_page_flip,
+		.page_flip = &rs600_page_flip,
+		.post_page_flip = &rs600_post_page_flip,
+	},
 };
 
 static struct radeon_asic rv515_asic = {
@@ -550,48 +728,70 @@ static struct radeon_asic rv515_asic = {
 	.vga_set_state = &r100_vga_set_state,
 	.gpu_is_lockup = &r300_gpu_is_lockup,
 	.asic_reset = &rs600_asic_reset,
-	.gart_tlb_flush = &rv370_pcie_gart_tlb_flush,
-	.gart_set_page = &rv370_pcie_gart_set_page,
-	.ring_start = &rv515_ring_start,
-	.ring_test = &r100_ring_test,
+	.ioctl_wait_idle = NULL,
+	.gui_idle = &r100_gui_idle,
+	.mc_wait_for_idle = &rv515_mc_wait_for_idle,
+	.gart = {
+		.tlb_flush = &rv370_pcie_gart_tlb_flush,
+		.set_page = &rv370_pcie_gart_set_page,
+	},
 	.ring = {
 		[RADEON_RING_TYPE_GFX_INDEX] = {
 			.ib_execute = &r100_ring_ib_execute,
 			.emit_fence = &r300_fence_ring_emit,
 			.emit_semaphore = &r100_semaphore_ring_emit,
+			.cs_parse = &r300_cs_parse,
+			.ring_start = &rv515_ring_start,
+			.ring_test = &r100_ring_test,
+			.ib_test = &r100_ib_test,
 		}
 	},
-	.irq_set = &rs600_irq_set,
-	.irq_process = &rs600_irq_process,
-	.get_vblank_counter = &rs600_get_vblank_counter,
-	.cs_parse = &r300_cs_parse,
-	.copy_blit = &r100_copy_blit,
-	.copy_dma = &r200_copy_dma,
-	.copy = &r100_copy_blit,
-	.get_engine_clock = &radeon_atom_get_engine_clock,
-	.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 = &rv370_get_pcie_lanes,
-	.set_pcie_lanes = &rv370_set_pcie_lanes,
-	.set_clock_gating = &radeon_atom_set_clock_gating,
-	.set_surface_reg = r100_set_surface_reg,
-	.clear_surface_reg = r100_clear_surface_reg,
-	.bandwidth_update = &rv515_bandwidth_update,
-	.hpd_init = &rs600_hpd_init,
-	.hpd_fini = &rs600_hpd_fini,
-	.hpd_sense = &rs600_hpd_sense,
-	.hpd_set_polarity = &rs600_hpd_set_polarity,
-	.ioctl_wait_idle = NULL,
-	.gui_idle = &r100_gui_idle,
-	.pm_misc = &rs600_pm_misc,
-	.pm_prepare = &rs600_pm_prepare,
-	.pm_finish = &rs600_pm_finish,
-	.pm_init_profile = &r420_pm_init_profile,
-	.pm_get_dynpm_state = &r100_pm_get_dynpm_state,
-	.pre_page_flip = &rs600_pre_page_flip,
-	.page_flip = &rs600_page_flip,
-	.post_page_flip = &rs600_post_page_flip,
+	.irq = {
+		.set = &rs600_irq_set,
+		.process = &rs600_irq_process,
+	},
+	.display = {
+		.get_vblank_counter = &rs600_get_vblank_counter,
+		.bandwidth_update = &rv515_bandwidth_update,
+		.wait_for_vblank = &avivo_wait_for_vblank,
+	},
+	.copy = {
+		.blit = &r100_copy_blit,
+		.blit_ring_index = RADEON_RING_TYPE_GFX_INDEX,
+		.dma = &r200_copy_dma,
+		.dma_ring_index = RADEON_RING_TYPE_GFX_INDEX,
+		.copy = &r100_copy_blit,
+		.copy_ring_index = RADEON_RING_TYPE_GFX_INDEX,
+	},
+	.surface = {
+		.set_reg = r100_set_surface_reg,
+		.clear_reg = r100_clear_surface_reg,
+	},
+	.hpd = {
+		.init = &rs600_hpd_init,
+		.fini = &rs600_hpd_fini,
+		.sense = &rs600_hpd_sense,
+		.set_polarity = &rs600_hpd_set_polarity,
+	},
+	.pm = {
+		.misc = &rs600_pm_misc,
+		.prepare = &rs600_pm_prepare,
+		.finish = &rs600_pm_finish,
+		.init_profile = &r420_pm_init_profile,
+		.get_dynpm_state = &r100_pm_get_dynpm_state,
+		.get_engine_clock = &radeon_atom_get_engine_clock,
+		.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 = &rv370_get_pcie_lanes,
+		.set_pcie_lanes = &rv370_set_pcie_lanes,
+		.set_clock_gating = &radeon_atom_set_clock_gating,
+	},
+	.pflip = {
+		.pre_page_flip = &rs600_pre_page_flip,
+		.page_flip = &rs600_page_flip,
+		.post_page_flip = &rs600_post_page_flip,
+	},
 };
 
 static struct radeon_asic r520_asic = {
@@ -602,48 +802,70 @@ static struct radeon_asic r520_asic = {
 	.vga_set_state = &r100_vga_set_state,
 	.gpu_is_lockup = &r300_gpu_is_lockup,
 	.asic_reset = &rs600_asic_reset,
-	.gart_tlb_flush = &rv370_pcie_gart_tlb_flush,
-	.gart_set_page = &rv370_pcie_gart_set_page,
-	.ring_start = &rv515_ring_start,
-	.ring_test = &r100_ring_test,
+	.ioctl_wait_idle = NULL,
+	.gui_idle = &r100_gui_idle,
+	.mc_wait_for_idle = &r520_mc_wait_for_idle,
+	.gart = {
+		.tlb_flush = &rv370_pcie_gart_tlb_flush,
+		.set_page = &rv370_pcie_gart_set_page,
+	},
 	.ring = {
 		[RADEON_RING_TYPE_GFX_INDEX] = {
 			.ib_execute = &r100_ring_ib_execute,
 			.emit_fence = &r300_fence_ring_emit,
 			.emit_semaphore = &r100_semaphore_ring_emit,
+			.cs_parse = &r300_cs_parse,
+			.ring_start = &rv515_ring_start,
+			.ring_test = &r100_ring_test,
+			.ib_test = &r100_ib_test,
 		}
 	},
-	.irq_set = &rs600_irq_set,
-	.irq_process = &rs600_irq_process,
-	.get_vblank_counter = &rs600_get_vblank_counter,
-	.cs_parse = &r300_cs_parse,
-	.copy_blit = &r100_copy_blit,
-	.copy_dma = &r200_copy_dma,
-	.copy = &r100_copy_blit,
-	.get_engine_clock = &radeon_atom_get_engine_clock,
-	.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 = &rv370_get_pcie_lanes,
-	.set_pcie_lanes = &rv370_set_pcie_lanes,
-	.set_clock_gating = &radeon_atom_set_clock_gating,
-	.set_surface_reg = r100_set_surface_reg,
-	.clear_surface_reg = r100_clear_surface_reg,
-	.bandwidth_update = &rv515_bandwidth_update,
-	.hpd_init = &rs600_hpd_init,
-	.hpd_fini = &rs600_hpd_fini,
-	.hpd_sense = &rs600_hpd_sense,
-	.hpd_set_polarity = &rs600_hpd_set_polarity,
-	.ioctl_wait_idle = NULL,
-	.gui_idle = &r100_gui_idle,
-	.pm_misc = &rs600_pm_misc,
-	.pm_prepare = &rs600_pm_prepare,
-	.pm_finish = &rs600_pm_finish,
-	.pm_init_profile = &r420_pm_init_profile,
-	.pm_get_dynpm_state = &r100_pm_get_dynpm_state,
-	.pre_page_flip = &rs600_pre_page_flip,
-	.page_flip = &rs600_page_flip,
-	.post_page_flip = &rs600_post_page_flip,
+	.irq = {
+		.set = &rs600_irq_set,
+		.process = &rs600_irq_process,
+	},
+	.display = {
+		.bandwidth_update = &rv515_bandwidth_update,
+		.get_vblank_counter = &rs600_get_vblank_counter,
+		.wait_for_vblank = &avivo_wait_for_vblank,
+	},
+	.copy = {
+		.blit = &r100_copy_blit,
+		.blit_ring_index = RADEON_RING_TYPE_GFX_INDEX,
+		.dma = &r200_copy_dma,
+		.dma_ring_index = RADEON_RING_TYPE_GFX_INDEX,
+		.copy = &r100_copy_blit,
+		.copy_ring_index = RADEON_RING_TYPE_GFX_INDEX,
+	},
+	.surface = {
+		.set_reg = r100_set_surface_reg,
+		.clear_reg = r100_clear_surface_reg,
+	},
+	.hpd = {
+		.init = &rs600_hpd_init,
+		.fini = &rs600_hpd_fini,
+		.sense = &rs600_hpd_sense,
+		.set_polarity = &rs600_hpd_set_polarity,
+	},
+	.pm = {
+		.misc = &rs600_pm_misc,
+		.prepare = &rs600_pm_prepare,
+		.finish = &rs600_pm_finish,
+		.init_profile = &r420_pm_init_profile,
+		.get_dynpm_state = &r100_pm_get_dynpm_state,
+		.get_engine_clock = &radeon_atom_get_engine_clock,
+		.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 = &rv370_get_pcie_lanes,
+		.set_pcie_lanes = &rv370_set_pcie_lanes,
+		.set_clock_gating = &radeon_atom_set_clock_gating,
+	},
+	.pflip = {
+		.pre_page_flip = &rs600_pre_page_flip,
+		.page_flip = &rs600_page_flip,
+		.post_page_flip = &rs600_post_page_flip,
+	},
 };
 
 static struct radeon_asic r600_asic = {
@@ -654,47 +876,69 @@ static struct radeon_asic r600_asic = {
 	.vga_set_state = &r600_vga_set_state,
 	.gpu_is_lockup = &r600_gpu_is_lockup,
 	.asic_reset = &r600_asic_reset,
-	.gart_tlb_flush = &r600_pcie_gart_tlb_flush,
-	.gart_set_page = &rs600_gart_set_page,
-	.ring_test = &r600_ring_test,
+	.ioctl_wait_idle = r600_ioctl_wait_idle,
+	.gui_idle = &r600_gui_idle,
+	.mc_wait_for_idle = &r600_mc_wait_for_idle,
+	.gart = {
+		.tlb_flush = &r600_pcie_gart_tlb_flush,
+		.set_page = &rs600_gart_set_page,
+	},
 	.ring = {
 		[RADEON_RING_TYPE_GFX_INDEX] = {
 			.ib_execute = &r600_ring_ib_execute,
 			.emit_fence = &r600_fence_ring_emit,
 			.emit_semaphore = &r600_semaphore_ring_emit,
+			.cs_parse = &r600_cs_parse,
+			.ring_test = &r600_ring_test,
+			.ib_test = &r600_ib_test,
 		}
 	},
-	.irq_set = &r600_irq_set,
-	.irq_process = &r600_irq_process,
-	.get_vblank_counter = &rs600_get_vblank_counter,
-	.cs_parse = &r600_cs_parse,
-	.copy_blit = &r600_copy_blit,
-	.copy_dma = NULL,
-	.copy = &r600_copy_blit,
-	.get_engine_clock = &radeon_atom_get_engine_clock,
-	.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 = &r600_get_pcie_lanes,
-	.set_pcie_lanes = &r600_set_pcie_lanes,
-	.set_clock_gating = NULL,
-	.set_surface_reg = r600_set_surface_reg,
-	.clear_surface_reg = r600_clear_surface_reg,
-	.bandwidth_update = &rv515_bandwidth_update,
-	.hpd_init = &r600_hpd_init,
-	.hpd_fini = &r600_hpd_fini,
-	.hpd_sense = &r600_hpd_sense,
-	.hpd_set_polarity = &r600_hpd_set_polarity,
-	.ioctl_wait_idle = r600_ioctl_wait_idle,
-	.gui_idle = &r600_gui_idle,
-	.pm_misc = &r600_pm_misc,
-	.pm_prepare = &rs600_pm_prepare,
-	.pm_finish = &rs600_pm_finish,
-	.pm_init_profile = &r600_pm_init_profile,
-	.pm_get_dynpm_state = &r600_pm_get_dynpm_state,
-	.pre_page_flip = &rs600_pre_page_flip,
-	.page_flip = &rs600_page_flip,
-	.post_page_flip = &rs600_post_page_flip,
+	.irq = {
+		.set = &r600_irq_set,
+		.process = &r600_irq_process,
+	},
+	.display = {
+		.bandwidth_update = &rv515_bandwidth_update,
+		.get_vblank_counter = &rs600_get_vblank_counter,
+		.wait_for_vblank = &avivo_wait_for_vblank,
+	},
+	.copy = {
+		.blit = &r600_copy_blit,
+		.blit_ring_index = RADEON_RING_TYPE_GFX_INDEX,
+		.dma = NULL,
+		.dma_ring_index = RADEON_RING_TYPE_GFX_INDEX,
+		.copy = &r600_copy_blit,
+		.copy_ring_index = RADEON_RING_TYPE_GFX_INDEX,
+	},
+	.surface = {
+		.set_reg = r600_set_surface_reg,
+		.clear_reg = r600_clear_surface_reg,
+	},
+	.hpd = {
+		.init = &r600_hpd_init,
+		.fini = &r600_hpd_fini,
+		.sense = &r600_hpd_sense,
+		.set_polarity = &r600_hpd_set_polarity,
+	},
+	.pm = {
+		.misc = &r600_pm_misc,
+		.prepare = &rs600_pm_prepare,
+		.finish = &rs600_pm_finish,
+		.init_profile = &r600_pm_init_profile,
+		.get_dynpm_state = &r600_pm_get_dynpm_state,
+		.get_engine_clock = &radeon_atom_get_engine_clock,
+		.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 = &r600_get_pcie_lanes,
+		.set_pcie_lanes = &r600_set_pcie_lanes,
+		.set_clock_gating = NULL,
+	},
+	.pflip = {
+		.pre_page_flip = &rs600_pre_page_flip,
+		.page_flip = &rs600_page_flip,
+		.post_page_flip = &rs600_post_page_flip,
+	},
 };
 
 static struct radeon_asic rs780_asic = {
@@ -705,47 +949,69 @@ static struct radeon_asic rs780_asic = {
 	.gpu_is_lockup = &r600_gpu_is_lockup,
 	.vga_set_state = &r600_vga_set_state,
 	.asic_reset = &r600_asic_reset,
-	.gart_tlb_flush = &r600_pcie_gart_tlb_flush,
-	.gart_set_page = &rs600_gart_set_page,
-	.ring_test = &r600_ring_test,
+	.ioctl_wait_idle = r600_ioctl_wait_idle,
+	.gui_idle = &r600_gui_idle,
+	.mc_wait_for_idle = &r600_mc_wait_for_idle,
+	.gart = {
+		.tlb_flush = &r600_pcie_gart_tlb_flush,
+		.set_page = &rs600_gart_set_page,
+	},
 	.ring = {
 		[RADEON_RING_TYPE_GFX_INDEX] = {
 			.ib_execute = &r600_ring_ib_execute,
 			.emit_fence = &r600_fence_ring_emit,
 			.emit_semaphore = &r600_semaphore_ring_emit,
+			.cs_parse = &r600_cs_parse,
+			.ring_test = &r600_ring_test,
+			.ib_test = &r600_ib_test,
 		}
 	},
-	.irq_set = &r600_irq_set,
-	.irq_process = &r600_irq_process,
-	.get_vblank_counter = &rs600_get_vblank_counter,
-	.cs_parse = &r600_cs_parse,
-	.copy_blit = &r600_copy_blit,
-	.copy_dma = NULL,
-	.copy = &r600_copy_blit,
-	.get_engine_clock = &radeon_atom_get_engine_clock,
-	.set_engine_clock = &radeon_atom_set_engine_clock,
-	.get_memory_clock = NULL,
-	.set_memory_clock = NULL,
-	.get_pcie_lanes = NULL,
-	.set_pcie_lanes = NULL,
-	.set_clock_gating = NULL,
-	.set_surface_reg = r600_set_surface_reg,
-	.clear_surface_reg = r600_clear_surface_reg,
-	.bandwidth_update = &rs690_bandwidth_update,
-	.hpd_init = &r600_hpd_init,
-	.hpd_fini = &r600_hpd_fini,
-	.hpd_sense = &r600_hpd_sense,
-	.hpd_set_polarity = &r600_hpd_set_polarity,
-	.ioctl_wait_idle = r600_ioctl_wait_idle,
-	.gui_idle = &r600_gui_idle,
-	.pm_misc = &r600_pm_misc,
-	.pm_prepare = &rs600_pm_prepare,
-	.pm_finish = &rs600_pm_finish,
-	.pm_init_profile = &rs780_pm_init_profile,
-	.pm_get_dynpm_state = &r600_pm_get_dynpm_state,
-	.pre_page_flip = &rs600_pre_page_flip,
-	.page_flip = &rs600_page_flip,
-	.post_page_flip = &rs600_post_page_flip,
+	.irq = {
+		.set = &r600_irq_set,
+		.process = &r600_irq_process,
+	},
+	.display = {
+		.bandwidth_update = &rs690_bandwidth_update,
+		.get_vblank_counter = &rs600_get_vblank_counter,
+		.wait_for_vblank = &avivo_wait_for_vblank,
+	},
+	.copy = {
+		.blit = &r600_copy_blit,
+		.blit_ring_index = RADEON_RING_TYPE_GFX_INDEX,
+		.dma = NULL,
+		.dma_ring_index = RADEON_RING_TYPE_GFX_INDEX,
+		.copy = &r600_copy_blit,
+		.copy_ring_index = RADEON_RING_TYPE_GFX_INDEX,
+	},
+	.surface = {
+		.set_reg = r600_set_surface_reg,
+		.clear_reg = r600_clear_surface_reg,
+	},
+	.hpd = {
+		.init = &r600_hpd_init,
+		.fini = &r600_hpd_fini,
+		.sense = &r600_hpd_sense,
+		.set_polarity = &r600_hpd_set_polarity,
+	},
+	.pm = {
+		.misc = &r600_pm_misc,
+		.prepare = &rs600_pm_prepare,
+		.finish = &rs600_pm_finish,
+		.init_profile = &rs780_pm_init_profile,
+		.get_dynpm_state = &r600_pm_get_dynpm_state,
+		.get_engine_clock = &radeon_atom_get_engine_clock,
+		.set_engine_clock = &radeon_atom_set_engine_clock,
+		.get_memory_clock = NULL,
+		.set_memory_clock = NULL,
+		.get_pcie_lanes = NULL,
+		.set_pcie_lanes = NULL,
+		.set_clock_gating = NULL,
+	},
+	.pflip = {
+		.pre_page_flip = &rs600_pre_page_flip,
+		.page_flip = &rs600_page_flip,
+		.post_page_flip = &rs600_post_page_flip,
+	},
 };
 
 static struct radeon_asic rv770_asic = {
@@ -756,47 +1022,69 @@ static struct radeon_asic rv770_asic = {
 	.asic_reset = &r600_asic_reset,
 	.gpu_is_lockup = &r600_gpu_is_lockup,
 	.vga_set_state = &r600_vga_set_state,
-	.gart_tlb_flush = &r600_pcie_gart_tlb_flush,
-	.gart_set_page = &rs600_gart_set_page,
-	.ring_test = &r600_ring_test,
+	.ioctl_wait_idle = r600_ioctl_wait_idle,
+	.gui_idle = &r600_gui_idle,
+	.mc_wait_for_idle = &r600_mc_wait_for_idle,
+	.gart = {
+		.tlb_flush = &r600_pcie_gart_tlb_flush,
+		.set_page = &rs600_gart_set_page,
+	},
 	.ring = {
 		[RADEON_RING_TYPE_GFX_INDEX] = {
 			.ib_execute = &r600_ring_ib_execute,
 			.emit_fence = &r600_fence_ring_emit,
 			.emit_semaphore = &r600_semaphore_ring_emit,
+			.cs_parse = &r600_cs_parse,
+			.ring_test = &r600_ring_test,
+			.ib_test = &r600_ib_test,
 		}
 	},
-	.irq_set = &r600_irq_set,
-	.irq_process = &r600_irq_process,
-	.get_vblank_counter = &rs600_get_vblank_counter,
-	.cs_parse = &r600_cs_parse,
-	.copy_blit = &r600_copy_blit,
-	.copy_dma = NULL,
-	.copy = &r600_copy_blit,
-	.get_engine_clock = &radeon_atom_get_engine_clock,
-	.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 = &r600_get_pcie_lanes,
-	.set_pcie_lanes = &r600_set_pcie_lanes,
-	.set_clock_gating = &radeon_atom_set_clock_gating,
-	.set_surface_reg = r600_set_surface_reg,
-	.clear_surface_reg = r600_clear_surface_reg,
-	.bandwidth_update = &rv515_bandwidth_update,
-	.hpd_init = &r600_hpd_init,
-	.hpd_fini = &r600_hpd_fini,
-	.hpd_sense = &r600_hpd_sense,
-	.hpd_set_polarity = &r600_hpd_set_polarity,
-	.ioctl_wait_idle = r600_ioctl_wait_idle,
-	.gui_idle = &r600_gui_idle,
-	.pm_misc = &rv770_pm_misc,
-	.pm_prepare = &rs600_pm_prepare,
-	.pm_finish = &rs600_pm_finish,
-	.pm_init_profile = &r600_pm_init_profile,
-	.pm_get_dynpm_state = &r600_pm_get_dynpm_state,
-	.pre_page_flip = &rs600_pre_page_flip,
-	.page_flip = &rv770_page_flip,
-	.post_page_flip = &rs600_post_page_flip,
+	.irq = {
+		.set = &r600_irq_set,
+		.process = &r600_irq_process,
+	},
+	.display = {
+		.bandwidth_update = &rv515_bandwidth_update,
+		.get_vblank_counter = &rs600_get_vblank_counter,
+		.wait_for_vblank = &avivo_wait_for_vblank,
+	},
+	.copy = {
+		.blit = &r600_copy_blit,
+		.blit_ring_index = RADEON_RING_TYPE_GFX_INDEX,
+		.dma = NULL,
+		.dma_ring_index = RADEON_RING_TYPE_GFX_INDEX,
+		.copy = &r600_copy_blit,
+		.copy_ring_index = RADEON_RING_TYPE_GFX_INDEX,
+	},
+	.surface = {
+		.set_reg = r600_set_surface_reg,
+		.clear_reg = r600_clear_surface_reg,
+	},
+	.hpd = {
+		.init = &r600_hpd_init,
+		.fini = &r600_hpd_fini,
+		.sense = &r600_hpd_sense,
+		.set_polarity = &r600_hpd_set_polarity,
+	},
+	.pm = {
+		.misc = &rv770_pm_misc,
+		.prepare = &rs600_pm_prepare,
+		.finish = &rs600_pm_finish,
+		.init_profile = &r600_pm_init_profile,
+		.get_dynpm_state = &r600_pm_get_dynpm_state,
+		.get_engine_clock = &radeon_atom_get_engine_clock,
+		.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 = &r600_get_pcie_lanes,
+		.set_pcie_lanes = &r600_set_pcie_lanes,
+		.set_clock_gating = &radeon_atom_set_clock_gating,
+	},
+	.pflip = {
+		.pre_page_flip = &rs600_pre_page_flip,
+		.page_flip = &rv770_page_flip,
+		.post_page_flip = &rs600_post_page_flip,
+	},
 };
 
 static struct radeon_asic evergreen_asic = {
@@ -807,47 +1095,69 @@ static struct radeon_asic evergreen_asic = {
 	.gpu_is_lockup = &evergreen_gpu_is_lockup,
 	.asic_reset = &evergreen_asic_reset,
 	.vga_set_state = &r600_vga_set_state,
-	.gart_tlb_flush = &evergreen_pcie_gart_tlb_flush,
-	.gart_set_page = &rs600_gart_set_page,
-	.ring_test = &r600_ring_test,
+	.ioctl_wait_idle = r600_ioctl_wait_idle,
+	.gui_idle = &r600_gui_idle,
+	.mc_wait_for_idle = &evergreen_mc_wait_for_idle,
+	.gart = {
+		.tlb_flush = &evergreen_pcie_gart_tlb_flush,
+		.set_page = &rs600_gart_set_page,
+	},
 	.ring = {
 		[RADEON_RING_TYPE_GFX_INDEX] = {
 			.ib_execute = &evergreen_ring_ib_execute,
 			.emit_fence = &r600_fence_ring_emit,
 			.emit_semaphore = &r600_semaphore_ring_emit,
+			.cs_parse = &evergreen_cs_parse,
+			.ring_test = &r600_ring_test,
+			.ib_test = &r600_ib_test,
 		}
 	},
-	.irq_set = &evergreen_irq_set,
-	.irq_process = &evergreen_irq_process,
-	.get_vblank_counter = &evergreen_get_vblank_counter,
-	.cs_parse = &evergreen_cs_parse,
-	.copy_blit = &r600_copy_blit,
-	.copy_dma = NULL,
-	.copy = &r600_copy_blit,
-	.get_engine_clock = &radeon_atom_get_engine_clock,
-	.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 = &r600_get_pcie_lanes,
-	.set_pcie_lanes = &r600_set_pcie_lanes,
-	.set_clock_gating = NULL,
-	.set_surface_reg = r600_set_surface_reg,
-	.clear_surface_reg = r600_clear_surface_reg,
-	.bandwidth_update = &evergreen_bandwidth_update,
-	.hpd_init = &evergreen_hpd_init,
-	.hpd_fini = &evergreen_hpd_fini,
-	.hpd_sense = &evergreen_hpd_sense,
-	.hpd_set_polarity = &evergreen_hpd_set_polarity,
-	.ioctl_wait_idle = r600_ioctl_wait_idle,
-	.gui_idle = &r600_gui_idle,
-	.pm_misc = &evergreen_pm_misc,
-	.pm_prepare = &evergreen_pm_prepare,
-	.pm_finish = &evergreen_pm_finish,
-	.pm_init_profile = &r600_pm_init_profile,
-	.pm_get_dynpm_state = &r600_pm_get_dynpm_state,
-	.pre_page_flip = &evergreen_pre_page_flip,
-	.page_flip = &evergreen_page_flip,
-	.post_page_flip = &evergreen_post_page_flip,
+	.irq = {
+		.set = &evergreen_irq_set,
+		.process = &evergreen_irq_process,
+	},
+	.display = {
+		.bandwidth_update = &evergreen_bandwidth_update,
+		.get_vblank_counter = &evergreen_get_vblank_counter,
+		.wait_for_vblank = &dce4_wait_for_vblank,
+	},
+	.copy = {
+		.blit = &r600_copy_blit,
+		.blit_ring_index = RADEON_RING_TYPE_GFX_INDEX,
+		.dma = NULL,
+		.dma_ring_index = RADEON_RING_TYPE_GFX_INDEX,
+		.copy = &r600_copy_blit,
+		.copy_ring_index = RADEON_RING_TYPE_GFX_INDEX,
+	},
+	.surface = {
+		.set_reg = r600_set_surface_reg,
+		.clear_reg = r600_clear_surface_reg,
+	},
+	.hpd = {
+		.init = &evergreen_hpd_init,
+		.fini = &evergreen_hpd_fini,
+		.sense = &evergreen_hpd_sense,
+		.set_polarity = &evergreen_hpd_set_polarity,
+	},
+	.pm = {
+		.misc = &evergreen_pm_misc,
+		.prepare = &evergreen_pm_prepare,
+		.finish = &evergreen_pm_finish,
+		.init_profile = &r600_pm_init_profile,
+		.get_dynpm_state = &r600_pm_get_dynpm_state,
+		.get_engine_clock = &radeon_atom_get_engine_clock,
+		.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 = &r600_get_pcie_lanes,
+		.set_pcie_lanes = &r600_set_pcie_lanes,
+		.set_clock_gating = NULL,
+	},
+	.pflip = {
+		.pre_page_flip = &evergreen_pre_page_flip,
+		.page_flip = &evergreen_page_flip,
+		.post_page_flip = &evergreen_post_page_flip,
+	},
 };
 
 static struct radeon_asic sumo_asic = {
@@ -858,47 +1168,69 @@ static struct radeon_asic sumo_asic = {
 	.gpu_is_lockup = &evergreen_gpu_is_lockup,
 	.asic_reset = &evergreen_asic_reset,
 	.vga_set_state = &r600_vga_set_state,
-	.gart_tlb_flush = &evergreen_pcie_gart_tlb_flush,
-	.gart_set_page = &rs600_gart_set_page,
-	.ring_test = &r600_ring_test,
+	.ioctl_wait_idle = r600_ioctl_wait_idle,
+	.gui_idle = &r600_gui_idle,
+	.mc_wait_for_idle = &evergreen_mc_wait_for_idle,
+	.gart = {
+		.tlb_flush = &evergreen_pcie_gart_tlb_flush,
+		.set_page = &rs600_gart_set_page,
+	},
 	.ring = {
 		[RADEON_RING_TYPE_GFX_INDEX] = {
 			.ib_execute = &evergreen_ring_ib_execute,
 			.emit_fence = &r600_fence_ring_emit,
 			.emit_semaphore = &r600_semaphore_ring_emit,
-		}
+			.cs_parse = &evergreen_cs_parse,
+			.ring_test = &r600_ring_test,
+			.ib_test = &r600_ib_test,
+		},
+	},
+	.irq = {
+		.set = &evergreen_irq_set,
+		.process = &evergreen_irq_process,
+	},
+	.display = {
+		.bandwidth_update = &evergreen_bandwidth_update,
+		.get_vblank_counter = &evergreen_get_vblank_counter,
+		.wait_for_vblank = &dce4_wait_for_vblank,
+	},
+	.copy = {
+		.blit = &r600_copy_blit,
+		.blit_ring_index = RADEON_RING_TYPE_GFX_INDEX,
+		.dma = NULL,
+		.dma_ring_index = RADEON_RING_TYPE_GFX_INDEX,
+		.copy = &r600_copy_blit,
+		.copy_ring_index = RADEON_RING_TYPE_GFX_INDEX,
+	},
+	.surface = {
+		.set_reg = r600_set_surface_reg,
+		.clear_reg = r600_clear_surface_reg,
+	},
+	.hpd = {
+		.init = &evergreen_hpd_init,
+		.fini = &evergreen_hpd_fini,
+		.sense = &evergreen_hpd_sense,
+		.set_polarity = &evergreen_hpd_set_polarity,
+	},
+	.pm = {
+		.misc = &evergreen_pm_misc,
+		.prepare = &evergreen_pm_prepare,
+		.finish = &evergreen_pm_finish,
+		.init_profile = &sumo_pm_init_profile,
+		.get_dynpm_state = &r600_pm_get_dynpm_state,
+		.get_engine_clock = &radeon_atom_get_engine_clock,
+		.set_engine_clock = &radeon_atom_set_engine_clock,
+		.get_memory_clock = NULL,
+		.set_memory_clock = NULL,
+		.get_pcie_lanes = NULL,
+		.set_pcie_lanes = NULL,
+		.set_clock_gating = NULL,
+	},
+	.pflip = {
+		.pre_page_flip = &evergreen_pre_page_flip,
+		.page_flip = &evergreen_page_flip,
+		.post_page_flip = &evergreen_post_page_flip,
 	},
-	.irq_set = &evergreen_irq_set,
-	.irq_process = &evergreen_irq_process,
-	.get_vblank_counter = &evergreen_get_vblank_counter,
-	.cs_parse = &evergreen_cs_parse,
-	.copy_blit = &r600_copy_blit,
-	.copy_dma = NULL,
-	.copy = &r600_copy_blit,
-	.get_engine_clock = &radeon_atom_get_engine_clock,
-	.set_engine_clock = &radeon_atom_set_engine_clock,
-	.get_memory_clock = NULL,
-	.set_memory_clock = NULL,
-	.get_pcie_lanes = NULL,
-	.set_pcie_lanes = NULL,
-	.set_clock_gating = NULL,
-	.set_surface_reg = r600_set_surface_reg,
-	.clear_surface_reg = r600_clear_surface_reg,
-	.bandwidth_update = &evergreen_bandwidth_update,
-	.hpd_init = &evergreen_hpd_init,
-	.hpd_fini = &evergreen_hpd_fini,
-	.hpd_sense = &evergreen_hpd_sense,
-	.hpd_set_polarity = &evergreen_hpd_set_polarity,
-	.ioctl_wait_idle = r600_ioctl_wait_idle,
-	.gui_idle = &r600_gui_idle,
-	.pm_misc = &evergreen_pm_misc,
-	.pm_prepare = &evergreen_pm_prepare,
-	.pm_finish = &evergreen_pm_finish,
-	.pm_init_profile = &sumo_pm_init_profile,
-	.pm_get_dynpm_state = &r600_pm_get_dynpm_state,
-	.pre_page_flip = &evergreen_pre_page_flip,
-	.page_flip = &evergreen_page_flip,
-	.post_page_flip = &evergreen_post_page_flip,
 };
 
 static struct radeon_asic btc_asic = {
@@ -909,47 +1241,69 @@ static struct radeon_asic btc_asic = {
 	.gpu_is_lockup = &evergreen_gpu_is_lockup,
 	.asic_reset = &evergreen_asic_reset,
 	.vga_set_state = &r600_vga_set_state,
-	.gart_tlb_flush = &evergreen_pcie_gart_tlb_flush,
-	.gart_set_page = &rs600_gart_set_page,
-	.ring_test = &r600_ring_test,
+	.ioctl_wait_idle = r600_ioctl_wait_idle,
+	.gui_idle = &r600_gui_idle,
+	.mc_wait_for_idle = &evergreen_mc_wait_for_idle,
+	.gart = {
+		.tlb_flush = &evergreen_pcie_gart_tlb_flush,
+		.set_page = &rs600_gart_set_page,
+	},
 	.ring = {
 		[RADEON_RING_TYPE_GFX_INDEX] = {
 			.ib_execute = &evergreen_ring_ib_execute,
 			.emit_fence = &r600_fence_ring_emit,
 			.emit_semaphore = &r600_semaphore_ring_emit,
+			.cs_parse = &evergreen_cs_parse,
+			.ring_test = &r600_ring_test,
+			.ib_test = &r600_ib_test,
 		}
 	},
-	.irq_set = &evergreen_irq_set,
-	.irq_process = &evergreen_irq_process,
-	.get_vblank_counter = &evergreen_get_vblank_counter,
-	.cs_parse = &evergreen_cs_parse,
-	.copy_blit = &r600_copy_blit,
-	.copy_dma = NULL,
-	.copy = &r600_copy_blit,
-	.get_engine_clock = &radeon_atom_get_engine_clock,
-	.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,
-	.set_clock_gating = NULL,
-	.set_surface_reg = r600_set_surface_reg,
-	.clear_surface_reg = r600_clear_surface_reg,
-	.bandwidth_update = &evergreen_bandwidth_update,
-	.hpd_init = &evergreen_hpd_init,
-	.hpd_fini = &evergreen_hpd_fini,
-	.hpd_sense = &evergreen_hpd_sense,
-	.hpd_set_polarity = &evergreen_hpd_set_polarity,
-	.ioctl_wait_idle = r600_ioctl_wait_idle,
-	.gui_idle = &r600_gui_idle,
-	.pm_misc = &evergreen_pm_misc,
-	.pm_prepare = &evergreen_pm_prepare,
-	.pm_finish = &evergreen_pm_finish,
-	.pm_init_profile = &r600_pm_init_profile,
-	.pm_get_dynpm_state = &r600_pm_get_dynpm_state,
-	.pre_page_flip = &evergreen_pre_page_flip,
-	.page_flip = &evergreen_page_flip,
-	.post_page_flip = &evergreen_post_page_flip,
+	.irq = {
+		.set = &evergreen_irq_set,
+		.process = &evergreen_irq_process,
+	},
+	.display = {
+		.bandwidth_update = &evergreen_bandwidth_update,
+		.get_vblank_counter = &evergreen_get_vblank_counter,
+		.wait_for_vblank = &dce4_wait_for_vblank,
+	},
+	.copy = {
+		.blit = &r600_copy_blit,
+		.blit_ring_index = RADEON_RING_TYPE_GFX_INDEX,
+		.dma = NULL,
+		.dma_ring_index = RADEON_RING_TYPE_GFX_INDEX,
+		.copy = &r600_copy_blit,
+		.copy_ring_index = RADEON_RING_TYPE_GFX_INDEX,
+	},
+	.surface = {
+		.set_reg = r600_set_surface_reg,
+		.clear_reg = r600_clear_surface_reg,
+	},
+	.hpd = {
+		.init = &evergreen_hpd_init,
+		.fini = &evergreen_hpd_fini,
+		.sense = &evergreen_hpd_sense,
+		.set_polarity = &evergreen_hpd_set_polarity,
+	},
+	.pm = {
+		.misc = &evergreen_pm_misc,
+		.prepare = &evergreen_pm_prepare,
+		.finish = &evergreen_pm_finish,
+		.init_profile = &r600_pm_init_profile,
+		.get_dynpm_state = &r600_pm_get_dynpm_state,
+		.get_engine_clock = &radeon_atom_get_engine_clock,
+		.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,
+		.set_clock_gating = NULL,
+	},
+	.pflip = {
+		.pre_page_flip = &evergreen_pre_page_flip,
+		.page_flip = &evergreen_page_flip,
+		.post_page_flip = &evergreen_post_page_flip,
+	},
 };
 
 static const struct radeon_vm_funcs cayman_vm_funcs = {
@@ -970,60 +1324,282 @@ static struct radeon_asic cayman_asic = {
 	.gpu_is_lockup = &cayman_gpu_is_lockup,
 	.asic_reset = &cayman_asic_reset,
 	.vga_set_state = &r600_vga_set_state,
-	.gart_tlb_flush = &cayman_pcie_gart_tlb_flush,
-	.gart_set_page = &rs600_gart_set_page,
-	.ring_test = &r600_ring_test,
+	.ioctl_wait_idle = r600_ioctl_wait_idle,
+	.gui_idle = &r600_gui_idle,
+	.mc_wait_for_idle = &evergreen_mc_wait_for_idle,
+	.gart = {
+		.tlb_flush = &cayman_pcie_gart_tlb_flush,
+		.set_page = &rs600_gart_set_page,
+	},
+	.ring = {
+		[RADEON_RING_TYPE_GFX_INDEX] = {
+			.ib_execute = &cayman_ring_ib_execute,
+			.ib_parse = &evergreen_ib_parse,
+			.emit_fence = &cayman_fence_ring_emit,
+			.emit_semaphore = &r600_semaphore_ring_emit,
+			.cs_parse = &evergreen_cs_parse,
+			.ring_test = &r600_ring_test,
+			.ib_test = &r600_ib_test,
+		},
+		[CAYMAN_RING_TYPE_CP1_INDEX] = {
+			.ib_execute = &cayman_ring_ib_execute,
+			.ib_parse = &evergreen_ib_parse,
+			.emit_fence = &cayman_fence_ring_emit,
+			.emit_semaphore = &r600_semaphore_ring_emit,
+			.cs_parse = &evergreen_cs_parse,
+			.ring_test = &r600_ring_test,
+			.ib_test = &r600_ib_test,
+		},
+		[CAYMAN_RING_TYPE_CP2_INDEX] = {
+			.ib_execute = &cayman_ring_ib_execute,
+			.ib_parse = &evergreen_ib_parse,
+			.emit_fence = &cayman_fence_ring_emit,
+			.emit_semaphore = &r600_semaphore_ring_emit,
+			.cs_parse = &evergreen_cs_parse,
+			.ring_test = &r600_ring_test,
+			.ib_test = &r600_ib_test,
+		}
+	},
+	.irq = {
+		.set = &evergreen_irq_set,
+		.process = &evergreen_irq_process,
+	},
+	.display = {
+		.bandwidth_update = &evergreen_bandwidth_update,
+		.get_vblank_counter = &evergreen_get_vblank_counter,
+		.wait_for_vblank = &dce4_wait_for_vblank,
+	},
+	.copy = {
+		.blit = &r600_copy_blit,
+		.blit_ring_index = RADEON_RING_TYPE_GFX_INDEX,
+		.dma = NULL,
+		.dma_ring_index = RADEON_RING_TYPE_GFX_INDEX,
+		.copy = &r600_copy_blit,
+		.copy_ring_index = RADEON_RING_TYPE_GFX_INDEX,
+	},
+	.surface = {
+		.set_reg = r600_set_surface_reg,
+		.clear_reg = r600_clear_surface_reg,
+	},
+	.hpd = {
+		.init = &evergreen_hpd_init,
+		.fini = &evergreen_hpd_fini,
+		.sense = &evergreen_hpd_sense,
+		.set_polarity = &evergreen_hpd_set_polarity,
+	},
+	.pm = {
+		.misc = &evergreen_pm_misc,
+		.prepare = &evergreen_pm_prepare,
+		.finish = &evergreen_pm_finish,
+		.init_profile = &r600_pm_init_profile,
+		.get_dynpm_state = &r600_pm_get_dynpm_state,
+		.get_engine_clock = &radeon_atom_get_engine_clock,
+		.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,
+		.set_clock_gating = NULL,
+	},
+	.pflip = {
+		.pre_page_flip = &evergreen_pre_page_flip,
+		.page_flip = &evergreen_page_flip,
+		.post_page_flip = &evergreen_post_page_flip,
+	},
+};
+
+static struct radeon_asic trinity_asic = {
+	.init = &cayman_init,
+	.fini = &cayman_fini,
+	.suspend = &cayman_suspend,
+	.resume = &cayman_resume,
+	.gpu_is_lockup = &cayman_gpu_is_lockup,
+	.asic_reset = &cayman_asic_reset,
+	.vga_set_state = &r600_vga_set_state,
+	.ioctl_wait_idle = r600_ioctl_wait_idle,
+	.gui_idle = &r600_gui_idle,
+	.mc_wait_for_idle = &evergreen_mc_wait_for_idle,
+	.gart = {
+		.tlb_flush = &cayman_pcie_gart_tlb_flush,
+		.set_page = &rs600_gart_set_page,
+	},
 	.ring = {
 		[RADEON_RING_TYPE_GFX_INDEX] = {
 			.ib_execute = &cayman_ring_ib_execute,
 			.ib_parse = &evergreen_ib_parse,
 			.emit_fence = &cayman_fence_ring_emit,
 			.emit_semaphore = &r600_semaphore_ring_emit,
+			.cs_parse = &evergreen_cs_parse,
+			.ring_test = &r600_ring_test,
+			.ib_test = &r600_ib_test,
 		},
 		[CAYMAN_RING_TYPE_CP1_INDEX] = {
 			.ib_execute = &cayman_ring_ib_execute,
 			.ib_parse = &evergreen_ib_parse,
 			.emit_fence = &cayman_fence_ring_emit,
 			.emit_semaphore = &r600_semaphore_ring_emit,
+			.cs_parse = &evergreen_cs_parse,
+			.ring_test = &r600_ring_test,
+			.ib_test = &r600_ib_test,
 		},
 		[CAYMAN_RING_TYPE_CP2_INDEX] = {
 			.ib_execute = &cayman_ring_ib_execute,
 			.ib_parse = &evergreen_ib_parse,
 			.emit_fence = &cayman_fence_ring_emit,
 			.emit_semaphore = &r600_semaphore_ring_emit,
+			.cs_parse = &evergreen_cs_parse,
+			.ring_test = &r600_ring_test,
+			.ib_test = &r600_ib_test,
 		}
 	},
-	.irq_set = &evergreen_irq_set,
-	.irq_process = &evergreen_irq_process,
-	.get_vblank_counter = &evergreen_get_vblank_counter,
-	.cs_parse = &evergreen_cs_parse,
-	.copy_blit = &r600_copy_blit,
-	.copy_dma = NULL,
-	.copy = &r600_copy_blit,
-	.get_engine_clock = &radeon_atom_get_engine_clock,
-	.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,
-	.set_clock_gating = NULL,
-	.set_surface_reg = r600_set_surface_reg,
-	.clear_surface_reg = r600_clear_surface_reg,
-	.bandwidth_update = &evergreen_bandwidth_update,
-	.hpd_init = &evergreen_hpd_init,
-	.hpd_fini = &evergreen_hpd_fini,
-	.hpd_sense = &evergreen_hpd_sense,
-	.hpd_set_polarity = &evergreen_hpd_set_polarity,
+	.irq = {
+		.set = &evergreen_irq_set,
+		.process = &evergreen_irq_process,
+	},
+	.display = {
+		.bandwidth_update = &dce6_bandwidth_update,
+		.get_vblank_counter = &evergreen_get_vblank_counter,
+		.wait_for_vblank = &dce4_wait_for_vblank,
+	},
+	.copy = {
+		.blit = &r600_copy_blit,
+		.blit_ring_index = RADEON_RING_TYPE_GFX_INDEX,
+		.dma = NULL,
+		.dma_ring_index = RADEON_RING_TYPE_GFX_INDEX,
+		.copy = &r600_copy_blit,
+		.copy_ring_index = RADEON_RING_TYPE_GFX_INDEX,
+	},
+	.surface = {
+		.set_reg = r600_set_surface_reg,
+		.clear_reg = r600_clear_surface_reg,
+	},
+	.hpd = {
+		.init = &evergreen_hpd_init,
+		.fini = &evergreen_hpd_fini,
+		.sense = &evergreen_hpd_sense,
+		.set_polarity = &evergreen_hpd_set_polarity,
+	},
+	.pm = {
+		.misc = &evergreen_pm_misc,
+		.prepare = &evergreen_pm_prepare,
+		.finish = &evergreen_pm_finish,
+		.init_profile = &sumo_pm_init_profile,
+		.get_dynpm_state = &r600_pm_get_dynpm_state,
+		.get_engine_clock = &radeon_atom_get_engine_clock,
+		.set_engine_clock = &radeon_atom_set_engine_clock,
+		.get_memory_clock = NULL,
+		.set_memory_clock = NULL,
+		.get_pcie_lanes = NULL,
+		.set_pcie_lanes = NULL,
+		.set_clock_gating = NULL,
+	},
+	.pflip = {
+		.pre_page_flip = &evergreen_pre_page_flip,
+		.page_flip = &evergreen_page_flip,
+		.post_page_flip = &evergreen_post_page_flip,
+	},
+};
+
+static const struct radeon_vm_funcs si_vm_funcs = {
+	.init = &si_vm_init,
+	.fini = &si_vm_fini,
+	.bind = &si_vm_bind,
+	.unbind = &si_vm_unbind,
+	.tlb_flush = &si_vm_tlb_flush,
+	.page_flags = &cayman_vm_page_flags,
+	.set_page = &cayman_vm_set_page,
+};
+
+static struct radeon_asic si_asic = {
+	.init = &si_init,
+	.fini = &si_fini,
+	.suspend = &si_suspend,
+	.resume = &si_resume,
+	.gpu_is_lockup = &si_gpu_is_lockup,
+	.asic_reset = &si_asic_reset,
+	.vga_set_state = &r600_vga_set_state,
 	.ioctl_wait_idle = r600_ioctl_wait_idle,
 	.gui_idle = &r600_gui_idle,
-	.pm_misc = &evergreen_pm_misc,
-	.pm_prepare = &evergreen_pm_prepare,
-	.pm_finish = &evergreen_pm_finish,
-	.pm_init_profile = &r600_pm_init_profile,
-	.pm_get_dynpm_state = &r600_pm_get_dynpm_state,
-	.pre_page_flip = &evergreen_pre_page_flip,
-	.page_flip = &evergreen_page_flip,
-	.post_page_flip = &evergreen_post_page_flip,
+	.mc_wait_for_idle = &evergreen_mc_wait_for_idle,
+	.gart = {
+		.tlb_flush = &si_pcie_gart_tlb_flush,
+		.set_page = &rs600_gart_set_page,
+	},
+	.ring = {
+		[RADEON_RING_TYPE_GFX_INDEX] = {
+			.ib_execute = &si_ring_ib_execute,
+			.ib_parse = &si_ib_parse,
+			.emit_fence = &si_fence_ring_emit,
+			.emit_semaphore = &r600_semaphore_ring_emit,
+			.cs_parse = NULL,
+			.ring_test = &r600_ring_test,
+			.ib_test = &r600_ib_test,
+		},
+		[CAYMAN_RING_TYPE_CP1_INDEX] = {
+			.ib_execute = &si_ring_ib_execute,
+			.ib_parse = &si_ib_parse,
+			.emit_fence = &si_fence_ring_emit,
+			.emit_semaphore = &r600_semaphore_ring_emit,
+			.cs_parse = NULL,
+			.ring_test = &r600_ring_test,
+			.ib_test = &r600_ib_test,
+		},
+		[CAYMAN_RING_TYPE_CP2_INDEX] = {
+			.ib_execute = &si_ring_ib_execute,
+			.ib_parse = &si_ib_parse,
+			.emit_fence = &si_fence_ring_emit,
+			.emit_semaphore = &r600_semaphore_ring_emit,
+			.cs_parse = NULL,
+			.ring_test = &r600_ring_test,
+			.ib_test = &r600_ib_test,
+		}
+	},
+	.irq = {
+		.set = &si_irq_set,
+		.process = &si_irq_process,
+	},
+	.display = {
+		.bandwidth_update = &dce6_bandwidth_update,
+		.get_vblank_counter = &evergreen_get_vblank_counter,
+		.wait_for_vblank = &dce4_wait_for_vblank,
+	},
+	.copy = {
+		.blit = NULL,
+		.blit_ring_index = RADEON_RING_TYPE_GFX_INDEX,
+		.dma = NULL,
+		.dma_ring_index = RADEON_RING_TYPE_GFX_INDEX,
+		.copy = NULL,
+		.copy_ring_index = RADEON_RING_TYPE_GFX_INDEX,
+	},
+	.surface = {
+		.set_reg = r600_set_surface_reg,
+		.clear_reg = r600_clear_surface_reg,
+	},
+	.hpd = {
+		.init = &evergreen_hpd_init,
+		.fini = &evergreen_hpd_fini,
+		.sense = &evergreen_hpd_sense,
+		.set_polarity = &evergreen_hpd_set_polarity,
+	},
+	.pm = {
+		.misc = &evergreen_pm_misc,
+		.prepare = &evergreen_pm_prepare,
+		.finish = &evergreen_pm_finish,
+		.init_profile = &sumo_pm_init_profile,
+		.get_dynpm_state = &r600_pm_get_dynpm_state,
+		.get_engine_clock = &radeon_atom_get_engine_clock,
+		.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,
+		.set_clock_gating = NULL,
+	},
+	.pflip = {
+		.pre_page_flip = &evergreen_pre_page_flip,
+		.page_flip = &evergreen_page_flip,
+		.post_page_flip = &evergreen_post_page_flip,
+	},
 };
 
 int radeon_asic_init(struct radeon_device *rdev)
@@ -1036,9 +1612,6 @@ int radeon_asic_init(struct radeon_device *rdev)
 	else
 		rdev->num_crtc = 2;
 
-	/* set the ring used for bo copies */
-	rdev->copy_ring = RADEON_RING_TYPE_GFX_INDEX;
-
 	switch (rdev->family) {
 	case CHIP_R100:
 	case CHIP_RV100:
@@ -1068,10 +1641,10 @@ int radeon_asic_init(struct radeon_device *rdev)
 		rdev->asic = &r420_asic;
 		/* handle macs */
 		if (rdev->bios == NULL) {
-			rdev->asic->get_engine_clock = &radeon_legacy_get_engine_clock;
-			rdev->asic->set_engine_clock = &radeon_legacy_set_engine_clock;
-			rdev->asic->get_memory_clock = &radeon_legacy_get_memory_clock;
-			rdev->asic->set_memory_clock = NULL;
+			rdev->asic->pm.get_engine_clock = &radeon_legacy_get_engine_clock;
+			rdev->asic->pm.set_engine_clock = &radeon_legacy_set_engine_clock;
+			rdev->asic->pm.get_memory_clock = &radeon_legacy_get_memory_clock;
+			rdev->asic->pm.set_memory_clock = NULL;
 		}
 		break;
 	case CHIP_RS400:
@@ -1146,14 +1719,28 @@ int radeon_asic_init(struct radeon_device *rdev)
 		rdev->num_crtc = 6;
 		rdev->vm_manager.funcs = &cayman_vm_funcs;
 		break;
+	case CHIP_ARUBA:
+		rdev->asic = &trinity_asic;
+		/* set num crtcs */
+		rdev->num_crtc = 4;
+		rdev->vm_manager.funcs = &cayman_vm_funcs;
+		break;
+	case CHIP_TAHITI:
+	case CHIP_PITCAIRN:
+	case CHIP_VERDE:
+		rdev->asic = &si_asic;
+		/* set num crtcs */
+		rdev->num_crtc = 6;
+		rdev->vm_manager.funcs = &si_vm_funcs;
+		break;
 	default:
 		/* FIXME: not supported yet */
 		return -EINVAL;
 	}
 
 	if (rdev->flags & RADEON_IS_IGP) {
-		rdev->asic->get_memory_clock = NULL;
-		rdev->asic->set_memory_clock = NULL;
+		rdev->asic->pm.get_memory_clock = NULL;
+		rdev->asic->pm.set_memory_clock = NULL;
 	}
 
 	return 0;
diff --git a/drivers/gpu/drm/radeon/radeon_asic.h b/drivers/gpu/drm/radeon/radeon_asic.h
index 6304aef0d9b2..3d9f9f1d8f90 100644
--- a/drivers/gpu/drm/radeon/radeon_asic.h
+++ b/drivers/gpu/drm/radeon/radeon_asic.h
@@ -63,7 +63,7 @@ int r100_asic_reset(struct radeon_device *rdev);
 u32 r100_get_vblank_counter(struct radeon_device *rdev, int crtc);
 void r100_pci_gart_tlb_flush(struct radeon_device *rdev);
 int r100_pci_gart_set_page(struct radeon_device *rdev, int i, uint64_t addr);
-void r100_ring_start(struct radeon_device *rdev);
+void r100_ring_start(struct radeon_device *rdev, struct radeon_ring *ring);
 int r100_irq_set(struct radeon_device *rdev);
 int r100_irq_process(struct radeon_device *rdev);
 void r100_fence_ring_emit(struct radeon_device *rdev,
@@ -109,7 +109,7 @@ bool r100_gpu_cp_is_lockup(struct radeon_device *rdev,
 			   struct r100_gpu_lockup *lockup,
 			   struct radeon_ring *cp);
 void r100_ib_fini(struct radeon_device *rdev);
-int r100_ib_test(struct radeon_device *rdev);
+int r100_ib_test(struct radeon_device *rdev, struct radeon_ring *ring);
 void r100_irq_disable(struct radeon_device *rdev);
 void r100_mc_stop(struct radeon_device *rdev, struct r100_mc_save *save);
 void r100_mc_resume(struct radeon_device *rdev, struct r100_mc_save *save);
@@ -139,6 +139,8 @@ extern void r100_pm_get_dynpm_state(struct radeon_device *rdev);
 extern void r100_pre_page_flip(struct radeon_device *rdev, int crtc);
 extern u32 r100_page_flip(struct radeon_device *rdev, int crtc, u64 crtc_base);
 extern void r100_post_page_flip(struct radeon_device *rdev, int crtc);
+extern void r100_wait_for_vblank(struct radeon_device *rdev, int crtc);
+extern int r100_mc_wait_for_idle(struct radeon_device *rdev);
 
 /*
  * r200,rv250,rs300,rv280
@@ -159,7 +161,7 @@ extern int r300_suspend(struct radeon_device *rdev);
 extern int r300_resume(struct radeon_device *rdev);
 extern bool r300_gpu_is_lockup(struct radeon_device *rdev, struct radeon_ring *cp);
 extern int r300_asic_reset(struct radeon_device *rdev);
-extern void r300_ring_start(struct radeon_device *rdev);
+extern void r300_ring_start(struct radeon_device *rdev, struct radeon_ring *ring);
 extern void r300_fence_ring_emit(struct radeon_device *rdev,
 				struct radeon_fence *fence);
 extern int r300_cs_parse(struct radeon_cs_parser *p);
@@ -176,6 +178,7 @@ extern int rv370_pcie_gart_init(struct radeon_device *rdev);
 extern void rv370_pcie_gart_fini(struct radeon_device *rdev);
 extern int rv370_pcie_gart_enable(struct radeon_device *rdev);
 extern void rv370_pcie_gart_disable(struct radeon_device *rdev);
+extern int r300_mc_wait_for_idle(struct radeon_device *rdev);
 
 /*
  * r420,r423,rv410
@@ -206,6 +209,7 @@ int rs400_gart_enable(struct radeon_device *rdev);
 void rs400_gart_adjust_size(struct radeon_device *rdev);
 void rs400_gart_disable(struct radeon_device *rdev);
 void rs400_gart_fini(struct radeon_device *rdev);
+extern int rs400_mc_wait_for_idle(struct radeon_device *rdev);
 
 /*
  * rs600.
@@ -236,7 +240,8 @@ extern void rs600_pre_page_flip(struct radeon_device *rdev, int crtc);
 extern u32 rs600_page_flip(struct radeon_device *rdev, int crtc, u64 crtc_base);
 extern void rs600_post_page_flip(struct radeon_device *rdev, int crtc);
 void rs600_set_safe_registers(struct radeon_device *rdev);
-
+extern void avivo_wait_for_vblank(struct radeon_device *rdev, int crtc);
+extern int rs600_mc_wait_for_idle(struct radeon_device *rdev);
 
 /*
  * rs690,rs740
@@ -251,6 +256,7 @@ void rs690_bandwidth_update(struct radeon_device *rdev);
 void rs690_line_buffer_adjust(struct radeon_device *rdev,
 					struct drm_display_mode *mode1,
 					struct drm_display_mode *mode2);
+extern int rs690_mc_wait_for_idle(struct radeon_device *rdev);
 
 /*
  * rv515
@@ -267,7 +273,7 @@ int rv515_init(struct radeon_device *rdev);
 void rv515_fini(struct radeon_device *rdev);
 uint32_t rv515_mc_rreg(struct radeon_device *rdev, uint32_t reg);
 void rv515_mc_wreg(struct radeon_device *rdev, uint32_t reg, uint32_t v);
-void rv515_ring_start(struct radeon_device *rdev);
+void rv515_ring_start(struct radeon_device *rdev, struct radeon_ring *ring);
 void rv515_bandwidth_update(struct radeon_device *rdev);
 int rv515_resume(struct radeon_device *rdev);
 int rv515_suspend(struct radeon_device *rdev);
@@ -278,13 +284,14 @@ void rv515_mc_stop(struct radeon_device *rdev, struct rv515_mc_save *save);
 void rv515_mc_resume(struct radeon_device *rdev, struct rv515_mc_save *save);
 void rv515_clock_startup(struct radeon_device *rdev);
 void rv515_debugfs(struct radeon_device *rdev);
-
+int rv515_mc_wait_for_idle(struct radeon_device *rdev);
 
 /*
  * r520,rv530,rv560,rv570,r580
  */
 int r520_init(struct radeon_device *rdev);
 int r520_resume(struct radeon_device *rdev);
+int r520_mc_wait_for_idle(struct radeon_device *rdev);
 
 /*
  * r600,rv610,rv630,rv620,rv635,rv670,rs780,rs880
@@ -312,7 +319,7 @@ int r600_set_surface_reg(struct radeon_device *rdev, int reg,
 			 uint32_t tiling_flags, uint32_t pitch,
 			 uint32_t offset, uint32_t obj_size);
 void r600_clear_surface_reg(struct radeon_device *rdev, int reg);
-int r600_ib_test(struct radeon_device *rdev, int ring);
+int r600_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_copy_blit(struct radeon_device *rdev,
@@ -375,6 +382,7 @@ void r600_blit_done_copy(struct radeon_device *rdev, struct radeon_fence *fence)
 void r600_kms_blit_copy(struct radeon_device *rdev,
 			u64 src_gpu_addr, u64 dst_gpu_addr,
 			unsigned num_gpu_pages);
+int r600_mc_wait_for_idle(struct radeon_device *rdev);
 
 /*
  * rv770,rv730,rv710,rv740
@@ -423,8 +431,10 @@ extern void sumo_pm_init_profile(struct radeon_device *rdev);
 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);
+extern void dce4_wait_for_vblank(struct radeon_device *rdev, int crtc);
 void evergreen_disable_interrupt_state(struct radeon_device *rdev);
 int evergreen_blit_init(struct radeon_device *rdev);
+int evergreen_mc_wait_for_idle(struct radeon_device *rdev);
 
 /*
  * cayman
@@ -451,4 +461,29 @@ void cayman_vm_set_page(struct radeon_device *rdev, struct radeon_vm *vm,
 			unsigned pfn, uint64_t addr, uint32_t flags);
 int evergreen_ib_parse(struct radeon_device *rdev, struct radeon_ib *ib);
 
+/* DCE6 - SI */
+void dce6_bandwidth_update(struct radeon_device *rdev);
+
+/*
+ * si
+ */
+void si_fence_ring_emit(struct radeon_device *rdev,
+			struct radeon_fence *fence);
+void si_pcie_gart_tlb_flush(struct radeon_device *rdev);
+int si_init(struct radeon_device *rdev);
+void si_fini(struct radeon_device *rdev);
+int si_suspend(struct radeon_device *rdev);
+int si_resume(struct radeon_device *rdev);
+bool si_gpu_is_lockup(struct radeon_device *rdev, struct radeon_ring *cp);
+int si_asic_reset(struct radeon_device *rdev);
+void si_ring_ib_execute(struct radeon_device *rdev, struct radeon_ib *ib);
+int si_irq_set(struct radeon_device *rdev);
+int si_irq_process(struct radeon_device *rdev);
+int si_vm_init(struct radeon_device *rdev);
+void si_vm_fini(struct radeon_device *rdev);
+int si_vm_bind(struct radeon_device *rdev, struct radeon_vm *vm, int id);
+void si_vm_unbind(struct radeon_device *rdev, struct radeon_vm *vm);
+void si_vm_tlb_flush(struct radeon_device *rdev, struct radeon_vm *vm);
+int si_ib_parse(struct radeon_device *rdev, struct radeon_ib *ib);
+
 #endif
diff --git a/drivers/gpu/drm/radeon/radeon_atombios.c b/drivers/gpu/drm/radeon/radeon_atombios.c
index 5082d17d14dc..f6e69b8c06c6 100644
--- a/drivers/gpu/drm/radeon/radeon_atombios.c
+++ b/drivers/gpu/drm/radeon/radeon_atombios.c
@@ -56,6 +56,10 @@ extern void
 radeon_add_legacy_encoder(struct drm_device *dev, uint32_t encoder_enum,
 			  uint32_t supported_device);
 
+/* local */
+static int radeon_atom_get_max_vddc(struct radeon_device *rdev, u8 voltage_type,
+				    u16 voltage_id, u16 *voltage);
+
 union atom_supported_devices {
 	struct _ATOM_SUPPORTED_DEVICES_INFO info;
 	struct _ATOM_SUPPORTED_DEVICES_INFO_2 info_2;
@@ -253,7 +257,9 @@ static struct radeon_hpd radeon_atom_get_hpd_info_from_gpio(struct radeon_device
 
 	memset(&hpd, 0, sizeof(struct radeon_hpd));
 
-	if (ASIC_IS_DCE4(rdev))
+	if (ASIC_IS_DCE6(rdev))
+		reg = SI_DC_GPIO_HPD_A;
+	else if (ASIC_IS_DCE4(rdev))
 		reg = EVERGREEN_DC_GPIO_HPD_A;
 	else
 		reg = AVIVO_DC_GPIO_HPD_A;
@@ -442,6 +448,20 @@ static bool radeon_atom_apply_quirks(struct drm_device *dev,
 		struct radeon_device *rdev = dev->dev_private;
 		*i2c_bus = radeon_lookup_i2c_gpio(rdev, 0x93);
 	}
+
+	/* Fujitsu D3003-S2 board lists DVI-I as DVI-D and VGA */
+	if ((dev->pdev->device == 0x9802) &&
+	    (dev->pdev->subsystem_vendor == 0x1734) &&
+	    (dev->pdev->subsystem_device == 0x11bd)) {
+		if (*connector_type == DRM_MODE_CONNECTOR_VGA) {
+			*connector_type = DRM_MODE_CONNECTOR_DVII;
+			*line_mux = 0x3103;
+		} else if (*connector_type == DRM_MODE_CONNECTOR_DVID) {
+			*connector_type = DRM_MODE_CONNECTOR_DVII;
+		}
+	}
+
+
 	return true;
 }
 
@@ -1874,6 +1894,8 @@ static const char *pp_lib_thermal_controller_names[] = {
 	"emc2103",
 	"Sumo",
 	"Northern Islands",
+	"Southern Islands",
+	"lm96163",
 };
 
 union power_info {
@@ -1890,6 +1912,7 @@ union pplib_clock_info {
 	struct _ATOM_PPLIB_RS780_CLOCK_INFO rs780;
 	struct _ATOM_PPLIB_EVERGREEN_CLOCK_INFO evergreen;
 	struct _ATOM_PPLIB_SUMO_CLOCK_INFO sumo;
+	struct _ATOM_PPLIB_SI_CLOCK_INFO si;
 };
 
 union pplib_power_state {
@@ -2147,6 +2170,11 @@ static void radeon_atombios_add_pplib_thermal_controller(struct radeon_device *r
 				 (controller->ucFanParameters &
 				  ATOM_PP_FANPARAMETERS_NOFAN) ? "without" : "with");
 			rdev->pm.int_thermal_type = THERMAL_TYPE_NI;
+		} else if (controller->ucType == ATOM_PP_THERMALCONTROLLER_SISLANDS) {
+			DRM_INFO("Internal thermal controller %s fan control\n",
+				 (controller->ucFanParameters &
+				  ATOM_PP_FANPARAMETERS_NOFAN) ? "without" : "with");
+			rdev->pm.int_thermal_type = THERMAL_TYPE_SI;
 		} else if ((controller->ucType ==
 			    ATOM_PP_THERMALCONTROLLER_EXTERNAL_GPIO) ||
 			   (controller->ucType ==
@@ -2267,6 +2295,7 @@ static bool radeon_atombios_parse_pplib_clock_info(struct radeon_device *rdev,
 						   union pplib_clock_info *clock_info)
 {
 	u32 sclk, mclk;
+	u16 vddc;
 
 	if (rdev->flags & RADEON_IS_IGP) {
 		if (rdev->family >= CHIP_PALM) {
@@ -2278,6 +2307,19 @@ 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)) {
+		sclk = le16_to_cpu(clock_info->si.usEngineClockLow);
+		sclk |= clock_info->si.ucEngineClockHigh << 16;
+		mclk = le16_to_cpu(clock_info->si.usMemoryClockLow);
+		mclk |= clock_info->si.ucMemoryClockHigh << 16;
+		rdev->pm.power_state[state_index].clock_info[mode_index].mclk = mclk;
+		rdev->pm.power_state[state_index].clock_info[mode_index].sclk = sclk;
+		rdev->pm.power_state[state_index].clock_info[mode_index].voltage.type =
+			VOLTAGE_SW;
+		rdev->pm.power_state[state_index].clock_info[mode_index].voltage.voltage =
+			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)) {
 		sclk = le16_to_cpu(clock_info->evergreen.usEngineClockLow);
 		sclk |= clock_info->evergreen.ucEngineClockHigh << 16;
@@ -2305,11 +2347,18 @@ static bool radeon_atombios_parse_pplib_clock_info(struct radeon_device *rdev,
 	}
 
 	/* patch up vddc if necessary */
-	if (rdev->pm.power_state[state_index].clock_info[mode_index].voltage.voltage == 0xff01) {
-		u16 vddc;
-
-		if (radeon_atom_get_max_vddc(rdev, &vddc) == 0)
+	switch (rdev->pm.power_state[state_index].clock_info[mode_index].voltage.voltage) {
+	case ATOM_VIRTUAL_VOLTAGE_ID0:
+	case ATOM_VIRTUAL_VOLTAGE_ID1:
+	case ATOM_VIRTUAL_VOLTAGE_ID2:
+	case ATOM_VIRTUAL_VOLTAGE_ID3:
+		if (radeon_atom_get_max_vddc(rdev, VOLTAGE_TYPE_VDDC,
+					     rdev->pm.power_state[state_index].clock_info[mode_index].voltage.voltage,
+					     &vddc) == 0)
 			rdev->pm.power_state[state_index].clock_info[mode_index].voltage.voltage = vddc;
+		break;
+	default:
+		break;
 	}
 
 	if (rdev->flags & RADEON_IS_IGP) {
@@ -2419,9 +2468,9 @@ static int radeon_atombios_parse_power_table_6(struct radeon_device *rdev)
 	int i, j, non_clock_array_index, clock_array_index;
 	int state_index = 0, mode_index = 0;
 	union pplib_clock_info *clock_info;
-	struct StateArray *state_array;
-	struct ClockInfoArray *clock_info_array;
-	struct NonClockInfoArray *non_clock_info_array;
+	struct _StateArray *state_array;
+	struct _ClockInfoArray *clock_info_array;
+	struct _NonClockInfoArray *non_clock_info_array;
 	bool valid;
 	union power_info *power_info;
 	int index = GetIndexIntoMasterTable(DATA, PowerPlayInfo);
@@ -2434,13 +2483,13 @@ static int radeon_atombios_parse_power_table_6(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);
-	state_array = (struct StateArray *)
+	state_array = (struct _StateArray *)
 		(mode_info->atom_context->bios + data_offset +
 		 le16_to_cpu(power_info->pplib.usStateArrayOffset));
-	clock_info_array = (struct ClockInfoArray *)
+	clock_info_array = (struct _ClockInfoArray *)
 		(mode_info->atom_context->bios + data_offset +
 		 le16_to_cpu(power_info->pplib.usClockInfoArrayOffset));
-	non_clock_info_array = (struct NonClockInfoArray *)
+	non_clock_info_array = (struct _NonClockInfoArray *)
 		(mode_info->atom_context->bios + data_offset +
 		 le16_to_cpu(power_info->pplib.usNonClockInfoArrayOffset));
 	rdev->pm.power_state = kzalloc(sizeof(struct radeon_power_state) *
@@ -2467,7 +2516,7 @@ static int radeon_atombios_parse_power_table_6(struct radeon_device *rdev)
 				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->clockInfo[clock_array_index * clock_info_array->ucEntrySize];
 				valid = radeon_atombios_parse_pplib_clock_info(rdev,
 									       state_index, mode_index,
 									       clock_info);
@@ -2624,6 +2673,7 @@ union set_voltage {
 	struct _SET_VOLTAGE_PS_ALLOCATION alloc;
 	struct _SET_VOLTAGE_PARAMETERS v1;
 	struct _SET_VOLTAGE_PARAMETERS_V2 v2;
+	struct _SET_VOLTAGE_PARAMETERS_V1_3 v3;
 };
 
 void radeon_atom_set_voltage(struct radeon_device *rdev, u16 voltage_level, u8 voltage_type)
@@ -2650,6 +2700,11 @@ void radeon_atom_set_voltage(struct radeon_device *rdev, u16 voltage_level, u8 v
 		args.v2.ucVoltageMode = SET_ASIC_VOLTAGE_MODE_SET_VOLTAGE;
 		args.v2.usVoltageLevel = cpu_to_le16(voltage_level);
 		break;
+	case 3:
+		args.v3.ucVoltageType = voltage_type;
+		args.v3.ucVoltageMode = ATOM_SET_VOLTAGE;
+		args.v3.usVoltageLevel = cpu_to_le16(voltage_level);
+		break;
 	default:
 		DRM_ERROR("Unknown table version %d, %d\n", frev, crev);
 		return;
@@ -2658,8 +2713,8 @@ void radeon_atom_set_voltage(struct radeon_device *rdev, u16 voltage_level, u8 v
 	atom_execute_table(rdev->mode_info.atom_context, index, (uint32_t *)&args);
 }
 
-int radeon_atom_get_max_vddc(struct radeon_device *rdev,
-			     u16 *voltage)
+static int radeon_atom_get_max_vddc(struct radeon_device *rdev, u8 voltage_type,
+				    u16 voltage_id, u16 *voltage)
 {
 	union set_voltage args;
 	int index = GetIndexIntoMasterTable(COMMAND, SetVoltage);
@@ -2680,6 +2735,15 @@ int radeon_atom_get_max_vddc(struct radeon_device *rdev,
 
 		*voltage = le16_to_cpu(args.v2.usVoltageLevel);
 		break;
+	case 3:
+		args.v3.ucVoltageType = voltage_type;
+		args.v3.ucVoltageMode = ATOM_GET_VOLTAGE_LEVEL;
+		args.v3.usVoltageLevel = cpu_to_le16(voltage_id);
+
+		atom_execute_table(rdev->mode_info.atom_context, index, (uint32_t *)&args);
+
+		*voltage = le16_to_cpu(args.v3.usVoltageLevel);
+		break;
 	default:
 		DRM_ERROR("Unknown table version %d, %d\n", frev, crev);
 		return -EINVAL;
@@ -2931,6 +2995,20 @@ radeon_atombios_connected_scratch_regs(struct drm_connector *connector,
 			bios_6_scratch &= ~ATOM_S6_ACC_REQ_DFP5;
 		}
 	}
+	if ((radeon_encoder->devices & ATOM_DEVICE_DFP6_SUPPORT) &&
+	    (radeon_connector->devices & ATOM_DEVICE_DFP6_SUPPORT)) {
+		if (connected) {
+			DRM_DEBUG_KMS("DFP6 connected\n");
+			bios_0_scratch |= ATOM_S0_DFP6;
+			bios_3_scratch |= ATOM_S3_DFP6_ACTIVE;
+			bios_6_scratch |= ATOM_S6_ACC_REQ_DFP6;
+		} else {
+			DRM_DEBUG_KMS("DFP6 disconnected\n");
+			bios_0_scratch &= ~ATOM_S0_DFP6;
+			bios_3_scratch &= ~ATOM_S3_DFP6_ACTIVE;
+			bios_6_scratch &= ~ATOM_S6_ACC_REQ_DFP6;
+		}
+	}
 
 	if (rdev->family >= CHIP_R600) {
 		WREG32(R600_BIOS_0_SCRATCH, bios_0_scratch);
@@ -2951,6 +3029,9 @@ radeon_atombios_encoder_crtc_scratch_regs(struct drm_encoder *encoder, int crtc)
 	struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder);
 	uint32_t bios_3_scratch;
 
+	if (ASIC_IS_DCE4(rdev))
+		return;
+
 	if (rdev->family >= CHIP_R600)
 		bios_3_scratch = RREG32(R600_BIOS_3_SCRATCH);
 	else
@@ -3003,6 +3084,9 @@ radeon_atombios_encoder_dpms_scratch_regs(struct drm_encoder *encoder, bool on)
 	struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder);
 	uint32_t bios_2_scratch;
 
+	if (ASIC_IS_DCE4(rdev))
+		return;
+
 	if (rdev->family >= CHIP_R600)
 		bios_2_scratch = RREG32(R600_BIOS_2_SCRATCH);
 	else
diff --git a/drivers/gpu/drm/radeon/radeon_benchmark.c b/drivers/gpu/drm/radeon/radeon_benchmark.c
index 815f2341ab94..fef7b722b05d 100644
--- a/drivers/gpu/drm/radeon/radeon_benchmark.c
+++ b/drivers/gpu/drm/radeon/radeon_benchmark.c
@@ -43,17 +43,19 @@ static int radeon_benchmark_do_move(struct radeon_device *rdev, unsigned size,
 
 	start_jiffies = jiffies;
 	for (i = 0; i < n; i++) {
-		r = radeon_fence_create(rdev, &fence, RADEON_RING_TYPE_GFX_INDEX);
-		if (r)
-			return r;
-
 		switch (flag) {
 		case RADEON_BENCHMARK_COPY_DMA:
+			r = radeon_fence_create(rdev, &fence, radeon_copy_dma_ring_index(rdev));
+			if (r)
+				return r;
 			r = radeon_copy_dma(rdev, saddr, daddr,
 					    size / RADEON_GPU_PAGE_SIZE,
 					    fence);
 			break;
 		case RADEON_BENCHMARK_COPY_BLIT:
+			r = radeon_fence_create(rdev, &fence, radeon_copy_blit_ring_index(rdev));
+			if (r)
+				return r;
 			r = radeon_copy_blit(rdev, saddr, daddr,
 					     size / RADEON_GPU_PAGE_SIZE,
 					     fence);
@@ -129,7 +131,7 @@ static void radeon_benchmark_move(struct radeon_device *rdev, unsigned size,
 	/* r100 doesn't have dma engine so skip the test */
 	/* also, VRAM-to-VRAM test doesn't make much sense for DMA */
 	/* skip it as well if domains are the same */
-	if ((rdev->asic->copy_dma) && (sdomain != ddomain)) {
+	if ((rdev->asic->copy.dma) && (sdomain != ddomain)) {
 		time = radeon_benchmark_do_move(rdev, size, saddr, daddr,
 						RADEON_BENCHMARK_COPY_DMA, n);
 		if (time < 0)
@@ -208,22 +210,22 @@ void radeon_benchmark(struct radeon_device *rdev, int test_number)
 		break;
 	case 3:
 		/* GTT to VRAM, buffer size sweep, powers of 2 */
-		for (i = 1; i <= 65536; i <<= 1)
-			radeon_benchmark_move(rdev, i*1024,
+		for (i = 1; i <= 16384; i <<= 1)
+			radeon_benchmark_move(rdev, i * RADEON_GPU_PAGE_SIZE,
 					      RADEON_GEM_DOMAIN_GTT,
 					      RADEON_GEM_DOMAIN_VRAM);
 		break;
 	case 4:
 		/* VRAM to GTT, buffer size sweep, powers of 2 */
-		for (i = 1; i <= 65536; i <<= 1)
-			radeon_benchmark_move(rdev, i*1024,
+		for (i = 1; i <= 16384; i <<= 1)
+			radeon_benchmark_move(rdev, i * RADEON_GPU_PAGE_SIZE,
 					      RADEON_GEM_DOMAIN_VRAM,
 					      RADEON_GEM_DOMAIN_GTT);
 		break;
 	case 5:
 		/* VRAM to VRAM, buffer size sweep, powers of 2 */
-		for (i = 1; i <= 65536; i <<= 1)
-			radeon_benchmark_move(rdev, i*1024,
+		for (i = 1; i <= 16384; i <<= 1)
+			radeon_benchmark_move(rdev, i * RADEON_GPU_PAGE_SIZE,
 					      RADEON_GEM_DOMAIN_VRAM,
 					      RADEON_GEM_DOMAIN_VRAM);
 		break;
diff --git a/drivers/gpu/drm/radeon/radeon_blit_common.h b/drivers/gpu/drm/radeon/radeon_blit_common.h
new file mode 100644
index 000000000000..4ecbe72c9d2d
--- /dev/null
+++ b/drivers/gpu/drm/radeon/radeon_blit_common.h
@@ -0,0 +1,44 @@
+/*
+ * Copyright 2009 Advanced Micro Devices, Inc.
+ * Copyright 2009 Red Hat Inc.
+ * Copyright 2012 Alcatel-Lucent, 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 (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 HOLDER(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.
+ *
+ */
+
+#ifndef __RADEON_BLIT_COMMON_H__
+
+#define DI_PT_RECTLIST        0x11
+#define DI_INDEX_SIZE_16_BIT  0x0
+#define DI_SRC_SEL_AUTO_INDEX 0x2
+
+#define FMT_8                 0x1
+#define FMT_5_6_5             0x8
+#define FMT_8_8_8_8           0x1a
+#define COLOR_8               0x1
+#define COLOR_5_6_5           0x8
+#define COLOR_8_8_8_8         0x1a
+
+#define RECT_UNIT_H           32
+#define RECT_UNIT_W           (RADEON_GPU_PAGE_SIZE / 4 / RECT_UNIT_H)
+
+#define __RADEON_BLIT_COMMON_H__
+#endif
diff --git a/drivers/gpu/drm/radeon/radeon_clocks.c b/drivers/gpu/drm/radeon/radeon_clocks.c
index b6e18c8db9f5..6ae0c75f016a 100644
--- a/drivers/gpu/drm/radeon/radeon_clocks.c
+++ b/drivers/gpu/drm/radeon/radeon_clocks.c
@@ -334,7 +334,7 @@ void radeon_get_clock_info(struct drm_device *dev)
 
 	if (!rdev->clock.default_sclk)
 		rdev->clock.default_sclk = radeon_get_engine_clock(rdev);
-	if ((!rdev->clock.default_mclk) && rdev->asic->get_memory_clock)
+	if ((!rdev->clock.default_mclk) && rdev->asic->pm.get_memory_clock)
 		rdev->clock.default_mclk = radeon_get_memory_clock(rdev);
 
 	rdev->pm.current_sclk = rdev->clock.default_sclk;
diff --git a/drivers/gpu/drm/radeon/radeon_connectors.c b/drivers/gpu/drm/radeon/radeon_connectors.c
index e7cb3ab09243..bd05156edbdb 100644
--- a/drivers/gpu/drm/radeon/radeon_connectors.c
+++ b/drivers/gpu/drm/radeon/radeon_connectors.c
@@ -827,6 +827,27 @@ static int radeon_dvi_get_modes(struct drm_connector *connector)
 	return ret;
 }
 
+static bool radeon_check_hpd_status_unchanged(struct drm_connector *connector)
+{
+	struct drm_device *dev = connector->dev;
+	struct radeon_device *rdev = dev->dev_private;
+	struct radeon_connector *radeon_connector = to_radeon_connector(connector);
+	enum drm_connector_status status;
+
+	/* We only trust HPD on R600 and newer ASICS. */
+	if (rdev->family >= CHIP_R600
+	  && radeon_connector->hpd.hpd != RADEON_HPD_NONE) {
+		if (radeon_hpd_sense(rdev, radeon_connector->hpd.hpd))
+			status = connector_status_connected;
+		else
+			status = connector_status_disconnected;
+		if (connector->status == status)
+			return true;
+	}
+
+	return false;
+}
+
 /*
  * DVI is complicated
  * Do a DDC probe, if DDC probe passes, get the full EDID so
@@ -851,6 +872,9 @@ radeon_dvi_detect(struct drm_connector *connector, bool force)
 	enum drm_connector_status ret = connector_status_disconnected;
 	bool dret = false;
 
+	if (!force && radeon_check_hpd_status_unchanged(connector))
+		return connector->status;
+
 	if (radeon_connector->ddc_bus)
 		dret = radeon_ddc_probe(radeon_connector);
 	if (dret) {
@@ -946,6 +970,10 @@ radeon_dvi_detect(struct drm_connector *connector, bool force)
 
 			encoder = obj_to_encoder(obj);
 
+			if (encoder->encoder_type != DRM_MODE_ENCODER_DAC ||
+			    encoder->encoder_type != DRM_MODE_ENCODER_TVDAC)
+				continue;
+
 			encoder_funcs = encoder->helper_private;
 			if (encoder_funcs->detect) {
 				if (ret != connector_status_connected) {
@@ -1057,7 +1085,7 @@ static int radeon_dvi_mode_valid(struct drm_connector *connector,
 		    (radeon_connector->connector_object_id == CONNECTOR_OBJECT_ID_HDMI_TYPE_B))
 			return MODE_OK;
 		else if (radeon_connector->connector_object_id == CONNECTOR_OBJECT_ID_HDMI_TYPE_A) {
-			if (ASIC_IS_DCE3(rdev)) {
+			if (ASIC_IS_DCE6(rdev)) {
 				/* HDMI 1.3+ supports max clock of 340 Mhz */
 				if (mode->clock > 340000)
 					return MODE_CLOCK_HIGH;
@@ -1117,13 +1145,23 @@ static int radeon_dp_get_modes(struct drm_connector *connector)
 	    (connector->connector_type == DRM_MODE_CONNECTOR_LVDS)) {
 		struct drm_display_mode *mode;
 
-		if (!radeon_dig_connector->edp_on)
-			atombios_set_edp_panel_power(connector,
-						     ATOM_TRANSMITTER_ACTION_POWER_ON);
-		ret = radeon_ddc_get_modes(radeon_connector);
-		if (!radeon_dig_connector->edp_on)
-			atombios_set_edp_panel_power(connector,
-						     ATOM_TRANSMITTER_ACTION_POWER_OFF);
+		if (connector->connector_type == DRM_MODE_CONNECTOR_eDP) {
+			if (!radeon_dig_connector->edp_on)
+				atombios_set_edp_panel_power(connector,
+							     ATOM_TRANSMITTER_ACTION_POWER_ON);
+			ret = radeon_ddc_get_modes(radeon_connector);
+			if (!radeon_dig_connector->edp_on)
+				atombios_set_edp_panel_power(connector,
+							     ATOM_TRANSMITTER_ACTION_POWER_OFF);
+		} else {
+			/* need to setup ddc on the bridge */
+			if (radeon_connector_encoder_get_dp_bridge_encoder_id(connector) !=
+			    ENCODER_OBJECT_ID_NONE) {
+				if (encoder)
+					radeon_atom_ext_encoder_setup_ddc(encoder);
+			}
+			ret = radeon_ddc_get_modes(radeon_connector);
+		}
 
 		if (ret > 0) {
 			if (encoder) {
@@ -1134,7 +1172,6 @@ static int radeon_dp_get_modes(struct drm_connector *connector)
 			return ret;
 		}
 
-		encoder = radeon_best_single_encoder(connector);
 		if (!encoder)
 			return 0;
 
@@ -1241,6 +1278,9 @@ radeon_dp_detect(struct drm_connector *connector, bool force)
 	struct radeon_connector_atom_dig *radeon_dig_connector = radeon_connector->con_priv;
 	struct drm_encoder *encoder = radeon_best_single_encoder(connector);
 
+	if (!force && radeon_check_hpd_status_unchanged(connector))
+		return connector->status;
+
 	if (radeon_connector->edid) {
 		kfree(radeon_connector->edid);
 		radeon_connector->edid = NULL;
diff --git a/drivers/gpu/drm/radeon/radeon_cp.c b/drivers/gpu/drm/radeon/radeon_cp.c
index 72ae8266b8e9..0ebb7d4796fa 100644
--- a/drivers/gpu/drm/radeon/radeon_cp.c
+++ b/drivers/gpu/drm/radeon/radeon_cp.c
@@ -2115,6 +2115,8 @@ int radeon_driver_load(struct drm_device *dev, unsigned long flags)
 		break;
 	}
 
+	pci_set_master(dev->pdev);
+
 	if (drm_pci_device_is_agp(dev))
 		dev_priv->flags |= RADEON_IS_AGP;
 	else if (pci_is_pcie(dev->pdev))
diff --git a/drivers/gpu/drm/radeon/radeon_cs.c b/drivers/gpu/drm/radeon/radeon_cs.c
index 435a3d970ab8..5cac83278338 100644
--- a/drivers/gpu/drm/radeon/radeon_cs.c
+++ b/drivers/gpu/drm/radeon/radeon_cs.c
@@ -85,12 +85,6 @@ int radeon_cs_parser_relocs(struct radeon_cs_parser *p)
 			radeon_bo_list_add_object(&p->relocs[i].lobj,
 						  &p->validated);
 
-			if (p->relocs[i].robj->tbo.sync_obj && !(r->flags & RADEON_RELOC_DONT_SYNC)) {
-				struct radeon_fence *fence = p->relocs[i].robj->tbo.sync_obj;
-				if (!radeon_fence_signaled(fence)) {
-					p->sync_to_ring[fence->ring] = true;
-				}
-			}
 		} else
 			p->relocs[i].handle = 0;
 	}
@@ -109,8 +103,13 @@ static int radeon_cs_get_ring(struct radeon_cs_parser *p, u32 ring, s32 priority
 		p->ring = RADEON_RING_TYPE_GFX_INDEX;
 		break;
 	case RADEON_CS_RING_COMPUTE:
-		/* for now */
-		p->ring = RADEON_RING_TYPE_GFX_INDEX;
+		if (p->rdev->family >= CHIP_TAHITI) {
+			if (p->priority > 0)
+				p->ring = CAYMAN_RING_TYPE_CP1_INDEX;
+			else
+				p->ring = CAYMAN_RING_TYPE_CP2_INDEX;
+		} else
+			p->ring = RADEON_RING_TYPE_GFX_INDEX;
 		break;
 	}
 	return 0;
@@ -118,11 +117,24 @@ static int radeon_cs_get_ring(struct radeon_cs_parser *p, u32 ring, s32 priority
 
 static int radeon_cs_sync_rings(struct radeon_cs_parser *p)
 {
+	bool sync_to_ring[RADEON_NUM_RINGS] = { };
 	int i, r;
 
+	for (i = 0; i < p->nrelocs; i++) {
+		if (!p->relocs[i].robj || !p->relocs[i].robj->tbo.sync_obj)
+			continue;
+
+		if (!(p->relocs[i].flags & RADEON_RELOC_DONT_SYNC)) {
+			struct radeon_fence *fence = p->relocs[i].robj->tbo.sync_obj;
+			if (!radeon_fence_signaled(fence)) {
+				sync_to_ring[fence->ring] = true;
+			}
+		}
+	}
+
 	for (i = 0; i < RADEON_NUM_RINGS; ++i) {
 		/* no need to sync to our own or unused rings */
-		if (i == p->ring || !p->sync_to_ring[i] || !p->rdev->ring[i].ready)
+		if (i == p->ring || !sync_to_ring[i] || !p->rdev->ring[i].ready)
 			continue;
 
 		if (!p->ib->fence->semaphore) {
@@ -163,6 +175,7 @@ int radeon_cs_parser_init(struct radeon_cs_parser *p, void *data)
 	p->chunk_ib_idx = -1;
 	p->chunk_relocs_idx = -1;
 	p->chunk_flags_idx = -1;
+	p->chunk_const_ib_idx = -1;
 	p->chunks_array = kcalloc(cs->num_chunks, sizeof(uint64_t), GFP_KERNEL);
 	if (p->chunks_array == NULL) {
 		return -ENOMEM;
@@ -201,6 +214,12 @@ int radeon_cs_parser_init(struct radeon_cs_parser *p, void *data)
 			if (p->chunks[i].length_dw == 0)
 				return -EINVAL;
 		}
+		if (p->chunks[i].chunk_id == RADEON_CHUNK_ID_CONST_IB) {
+			p->chunk_const_ib_idx = i;
+			/* zero length CONST IB isn't useful */
+			if (p->chunks[i].length_dw == 0)
+				return -EINVAL;
+		}
 		if (p->chunks[i].chunk_id == RADEON_CHUNK_ID_FLAGS) {
 			p->chunk_flags_idx = i;
 			/* zero length flags aren't useful */
@@ -236,21 +255,19 @@ int radeon_cs_parser_init(struct radeon_cs_parser *p, void *data)
 	if ((p->cs_flags & RADEON_CS_USE_VM) &&
 	    !p->rdev->vm_manager.enabled) {
 		DRM_ERROR("VM not active on asic!\n");
-		if (p->chunk_relocs_idx != -1)
-			kfree(p->chunks[p->chunk_relocs_idx].kdata);
-		if (p->chunk_flags_idx != -1)
-			kfree(p->chunks[p->chunk_flags_idx].kdata);
 		return -EINVAL;
 	}
 
-	if (radeon_cs_get_ring(p, ring, priority)) {
-		if (p->chunk_relocs_idx != -1)
-			kfree(p->chunks[p->chunk_relocs_idx].kdata);
-		if (p->chunk_flags_idx != -1)
-			kfree(p->chunks[p->chunk_flags_idx].kdata);
+	/* 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");
 		return -EINVAL;
 	}
 
+	if (radeon_cs_get_ring(p, ring, priority))
+		return -EINVAL;
+
 
 	/* deal with non-vm */
 	if ((p->chunk_ib_idx != -1) &&
@@ -264,11 +281,8 @@ int radeon_cs_parser_init(struct radeon_cs_parser *p, void *data)
 		p->chunks[p->chunk_ib_idx].kpage[0] = kmalloc(PAGE_SIZE, GFP_KERNEL);
 		p->chunks[p->chunk_ib_idx].kpage[1] = kmalloc(PAGE_SIZE, GFP_KERNEL);
 		if (p->chunks[p->chunk_ib_idx].kpage[0] == NULL ||
-		    p->chunks[p->chunk_ib_idx].kpage[1] == NULL) {
-			kfree(p->chunks[p->chunk_ib_idx].kpage[0]);
-			kfree(p->chunks[p->chunk_ib_idx].kpage[1]);
+		    p->chunks[p->chunk_ib_idx].kpage[1] == NULL)
 			return -ENOMEM;
-		}
 		p->chunks[p->chunk_ib_idx].kpage_idx[0] = -1;
 		p->chunks[p->chunk_ib_idx].kpage_idx[1] = -1;
 		p->chunks[p->chunk_ib_idx].last_copied_page = -1;
@@ -341,7 +355,7 @@ static int radeon_cs_ib_chunk(struct radeon_device *rdev,
 		return r;
 	}
 	parser->ib->length_dw = ib_chunk->length_dw;
-	r = radeon_cs_parse(parser);
+	r = radeon_cs_parse(rdev, parser->ring, parser);
 	if (r || parser->parser_error) {
 		DRM_ERROR("Invalid command stream !\n");
 		return r;
@@ -394,6 +408,32 @@ static int radeon_cs_ib_vm_chunk(struct radeon_device *rdev,
 	if ((parser->cs_flags & RADEON_CS_USE_VM) == 0)
 		return 0;
 
+	if ((rdev->family >= CHIP_TAHITI) &&
+	    (parser->chunk_const_ib_idx != -1)) {
+		ib_chunk = &parser->chunks[parser->chunk_const_ib_idx];
+		if (ib_chunk->length_dw > RADEON_IB_VM_MAX_SIZE) {
+			DRM_ERROR("cs IB CONST too big: %d\n", ib_chunk->length_dw);
+			return -EINVAL;
+		}
+		r =  radeon_ib_get(rdev, parser->ring, &parser->const_ib,
+				   ib_chunk->length_dw * 4);
+		if (r) {
+			DRM_ERROR("Failed to get const ib !\n");
+			return r;
+		}
+		parser->const_ib->is_const_ib = true;
+		parser->const_ib->length_dw = ib_chunk->length_dw;
+		/* Copy the packet into the IB */
+		if (DRM_COPY_FROM_USER(parser->const_ib->ptr, ib_chunk->user_ptr,
+				       ib_chunk->length_dw * 4)) {
+			return -EFAULT;
+		}
+		r = radeon_ring_ib_parse(rdev, parser->ring, parser->const_ib);
+		if (r) {
+			return r;
+		}
+	}
+
 	ib_chunk = &parser->chunks[parser->chunk_ib_idx];
 	if (ib_chunk->length_dw > RADEON_IB_VM_MAX_SIZE) {
 		DRM_ERROR("cs IB too big: %d\n", ib_chunk->length_dw);
@@ -429,11 +469,25 @@ static int radeon_cs_ib_vm_chunk(struct radeon_device *rdev,
 	if (r) {
 		DRM_ERROR("Failed to synchronize rings !\n");
 	}
+
+	if ((rdev->family >= CHIP_TAHITI) &&
+	    (parser->chunk_const_ib_idx != -1)) {
+		parser->const_ib->vm_id = vm->id;
+		/* ib pool is bind at 0 in virtual address space to gpu_addr is the
+		 * offset inside the pool bo
+		 */
+		parser->const_ib->gpu_addr = parser->const_ib->sa_bo.offset;
+		r = radeon_ib_schedule(rdev, parser->const_ib);
+		if (r)
+			goto out;
+	}
+
 	parser->ib->vm_id = vm->id;
 	/* ib pool is bind at 0 in virtual address space to gpu_addr is the
 	 * offset inside the pool bo
 	 */
 	parser->ib->gpu_addr = parser->ib->sa_bo.offset;
+	parser->ib->is_const_ib = false;
 	r = radeon_ib_schedule(rdev, parser->ib);
 out:
 	if (!r) {
@@ -453,6 +507,10 @@ int radeon_cs_ioctl(struct drm_device *dev, void *data, struct drm_file *filp)
 	int r;
 
 	radeon_mutex_lock(&rdev->cs_mutex);
+	if (!rdev->accel_working) {
+		radeon_mutex_unlock(&rdev->cs_mutex);
+		return -EBUSY;
+	}
 	/* initialize parser */
 	memset(&parser, 0, sizeof(struct radeon_cs_parser));
 	parser.filp = filp;
diff --git a/drivers/gpu/drm/radeon/radeon_cursor.c b/drivers/gpu/drm/radeon/radeon_cursor.c
index fde25c0d65a0..42acc6449dd6 100644
--- a/drivers/gpu/drm/radeon/radeon_cursor.c
+++ b/drivers/gpu/drm/radeon/radeon_cursor.c
@@ -151,7 +151,9 @@ int radeon_crtc_cursor_set(struct drm_crtc *crtc,
 			   uint32_t height)
 {
 	struct radeon_crtc *radeon_crtc = to_radeon_crtc(crtc);
+	struct radeon_device *rdev = crtc->dev->dev_private;
 	struct drm_gem_object *obj;
+	struct radeon_bo *robj;
 	uint64_t gpu_addr;
 	int ret;
 
@@ -173,7 +175,15 @@ int radeon_crtc_cursor_set(struct drm_crtc *crtc,
 		return -ENOENT;
 	}
 
-	ret = radeon_gem_object_pin(obj, RADEON_GEM_DOMAIN_VRAM, &gpu_addr);
+	robj = gem_to_radeon_bo(obj);
+	ret = radeon_bo_reserve(robj, false);
+	if (unlikely(ret != 0))
+		goto fail;
+	/* Only 27 bit offset for legacy cursor */
+	ret = radeon_bo_pin_restricted(robj, RADEON_GEM_DOMAIN_VRAM,
+				       ASIC_IS_AVIVO(rdev) ? 0 : 1 << 27,
+				       &gpu_addr);
+	radeon_bo_unreserve(robj);
 	if (ret)
 		goto fail;
 
@@ -181,14 +191,18 @@ int radeon_crtc_cursor_set(struct drm_crtc *crtc,
 	radeon_crtc->cursor_height = height;
 
 	radeon_lock_cursor(crtc, true);
-	/* XXX only 27 bit offset for legacy cursor */
 	radeon_set_cursor(crtc, obj, gpu_addr);
 	radeon_show_cursor(crtc);
 	radeon_lock_cursor(crtc, false);
 
 unpin:
 	if (radeon_crtc->cursor_bo) {
-		radeon_gem_object_unpin(radeon_crtc->cursor_bo);
+		robj = gem_to_radeon_bo(radeon_crtc->cursor_bo);
+		ret = radeon_bo_reserve(robj, false);
+		if (likely(ret == 0)) {
+			radeon_bo_unpin(robj);
+			radeon_bo_unreserve(robj);
+		}
 		drm_gem_object_unreference_unlocked(radeon_crtc->cursor_bo);
 	}
 
diff --git a/drivers/gpu/drm/radeon/radeon_device.c b/drivers/gpu/drm/radeon/radeon_device.c
index 49f7cb7e226b..ea7df16e2f84 100644
--- a/drivers/gpu/drm/radeon/radeon_device.c
+++ b/drivers/gpu/drm/radeon/radeon_device.c
@@ -89,6 +89,10 @@ static const char radeon_family_name[][16] = {
 	"TURKS",
 	"CAICOS",
 	"CAYMAN",
+	"ARUBA",
+	"TAHITI",
+	"PITCAIRN",
+	"VERDE",
 	"LAST",
 };
 
@@ -964,7 +968,7 @@ int radeon_resume_kms(struct drm_device *dev)
 	/* init dig PHYs, disp eng pll */
 	if (rdev->is_atom_bios) {
 		radeon_atom_encoder_init(rdev);
-		radeon_atom_dcpll_init(rdev);
+		radeon_atom_disp_eng_pll_init(rdev);
 	}
 	/* reset hpd state */
 	radeon_hpd_init(rdev);
diff --git a/drivers/gpu/drm/radeon/radeon_display.c b/drivers/gpu/drm/radeon/radeon_display.c
index 8c49fef1ce78..8086c96e0b06 100644
--- a/drivers/gpu/drm/radeon/radeon_display.c
+++ b/drivers/gpu/drm/radeon/radeon_display.c
@@ -303,8 +303,17 @@ void radeon_crtc_handle_flip(struct radeon_device *rdev, int crtc_id)
 	if (update_pending &&
 	    (DRM_SCANOUTPOS_VALID & radeon_get_crtc_scanoutpos(rdev->ddev, crtc_id,
 							       &vpos, &hpos)) &&
-	    (vpos >=0) &&
-	    (vpos < (99 * rdev->mode_info.crtcs[crtc_id]->base.hwmode.crtc_vdisplay)/100)) {
+	    ((vpos >= (99 * rdev->mode_info.crtcs[crtc_id]->base.hwmode.crtc_vdisplay)/100) ||
+	     (vpos < 0 && !ASIC_IS_AVIVO(rdev)))) {
+		/* crtc didn't flip in this target vblank interval,
+		 * but flip is pending in crtc. Based on the current
+		 * scanout position we know that the current frame is
+		 * (nearly) complete and the flip will (likely)
+		 * complete before the start of the next frame.
+		 */
+		update_pending = 0;
+	}
+	if (update_pending) {
 		/* crtc didn't flip in this target vblank interval,
 		 * but flip is pending in crtc. It will complete it
 		 * in next vblank interval, so complete the flip at
@@ -393,7 +402,9 @@ static int radeon_crtc_page_flip(struct drm_crtc *crtc,
 		DRM_ERROR("failed to reserve new rbo buffer before flip\n");
 		goto pflip_cleanup;
 	}
-	r = radeon_bo_pin(rbo, RADEON_GEM_DOMAIN_VRAM, &base);
+	/* Only 27 bit offset for legacy CRTC */
+	r = radeon_bo_pin_restricted(rbo, RADEON_GEM_DOMAIN_VRAM,
+				     ASIC_IS_AVIVO(rdev) ? 0 : 1 << 27, &base);
 	if (unlikely(r != 0)) {
 		radeon_bo_unreserve(rbo);
 		r = -EINVAL;
@@ -1078,15 +1089,21 @@ static const struct drm_framebuffer_funcs radeon_fb_funcs = {
 	.create_handle = radeon_user_framebuffer_create_handle,
 };
 
-void
+int
 radeon_framebuffer_init(struct drm_device *dev,
 			struct radeon_framebuffer *rfb,
 			struct drm_mode_fb_cmd2 *mode_cmd,
 			struct drm_gem_object *obj)
 {
+	int ret;
 	rfb->obj = obj;
-	drm_framebuffer_init(dev, &rfb->base, &radeon_fb_funcs);
+	ret = drm_framebuffer_init(dev, &rfb->base, &radeon_fb_funcs);
+	if (ret) {
+		rfb->obj = NULL;
+		return ret;
+	}
 	drm_helper_mode_fill_fb_struct(&rfb->base, mode_cmd);
+	return 0;
 }
 
 static struct drm_framebuffer *
@@ -1096,6 +1113,7 @@ radeon_user_framebuffer_create(struct drm_device *dev,
 {
 	struct drm_gem_object *obj;
 	struct radeon_framebuffer *radeon_fb;
+	int ret;
 
 	obj = drm_gem_object_lookup(dev, file_priv, mode_cmd->handles[0]);
 	if (obj ==  NULL) {
@@ -1108,7 +1126,12 @@ radeon_user_framebuffer_create(struct drm_device *dev,
 	if (radeon_fb == NULL)
 		return ERR_PTR(-ENOMEM);
 
-	radeon_framebuffer_init(dev, radeon_fb, mode_cmd, obj);
+	ret = radeon_framebuffer_init(dev, radeon_fb, mode_cmd, obj);
+	if (ret) {
+		kfree(radeon_fb);
+		drm_gem_object_unreference_unlocked(obj);
+		return NULL;
+	}
 
 	return &radeon_fb->base;
 }
@@ -1124,11 +1147,6 @@ static const struct drm_mode_config_funcs radeon_mode_funcs = {
 	.output_poll_changed = radeon_output_poll_changed
 };
 
-struct drm_prop_enum_list {
-	int type;
-	char *name;
-};
-
 static struct drm_prop_enum_list radeon_tmds_pll_enum_list[] =
 {	{ 0, "driver" },
 	{ 1, "bios" },
@@ -1153,86 +1171,53 @@ static struct drm_prop_enum_list radeon_underscan_enum_list[] =
 
 static int radeon_modeset_create_props(struct radeon_device *rdev)
 {
-	int i, sz;
+	int sz;
 
 	if (rdev->is_atom_bios) {
 		rdev->mode_info.coherent_mode_property =
-			drm_property_create(rdev->ddev,
-					    DRM_MODE_PROP_RANGE,
-					    "coherent", 2);
+			drm_property_create_range(rdev->ddev, 0 , "coherent", 0, 1);
 		if (!rdev->mode_info.coherent_mode_property)
 			return -ENOMEM;
-
-		rdev->mode_info.coherent_mode_property->values[0] = 0;
-		rdev->mode_info.coherent_mode_property->values[1] = 1;
 	}
 
 	if (!ASIC_IS_AVIVO(rdev)) {
 		sz = ARRAY_SIZE(radeon_tmds_pll_enum_list);
 		rdev->mode_info.tmds_pll_property =
-			drm_property_create(rdev->ddev,
-					    DRM_MODE_PROP_ENUM,
-					    "tmds_pll", sz);
-		for (i = 0; i < sz; i++) {
-			drm_property_add_enum(rdev->mode_info.tmds_pll_property,
-					      i,
-					      radeon_tmds_pll_enum_list[i].type,
-					      radeon_tmds_pll_enum_list[i].name);
-		}
+			drm_property_create_enum(rdev->ddev, 0,
+					    "tmds_pll",
+					    radeon_tmds_pll_enum_list, sz);
 	}
 
 	rdev->mode_info.load_detect_property =
-		drm_property_create(rdev->ddev,
-				    DRM_MODE_PROP_RANGE,
-				    "load detection", 2);
+		drm_property_create_range(rdev->ddev, 0, "load detection", 0, 1);
 	if (!rdev->mode_info.load_detect_property)
 		return -ENOMEM;
-	rdev->mode_info.load_detect_property->values[0] = 0;
-	rdev->mode_info.load_detect_property->values[1] = 1;
 
 	drm_mode_create_scaling_mode_property(rdev->ddev);
 
 	sz = ARRAY_SIZE(radeon_tv_std_enum_list);
 	rdev->mode_info.tv_std_property =
-		drm_property_create(rdev->ddev,
-				    DRM_MODE_PROP_ENUM,
-				    "tv standard", sz);
-	for (i = 0; i < sz; i++) {
-		drm_property_add_enum(rdev->mode_info.tv_std_property,
-				      i,
-				      radeon_tv_std_enum_list[i].type,
-				      radeon_tv_std_enum_list[i].name);
-	}
+		drm_property_create_enum(rdev->ddev, 0,
+				    "tv standard",
+				    radeon_tv_std_enum_list, sz);
 
 	sz = ARRAY_SIZE(radeon_underscan_enum_list);
 	rdev->mode_info.underscan_property =
-		drm_property_create(rdev->ddev,
-				    DRM_MODE_PROP_ENUM,
-				    "underscan", sz);
-	for (i = 0; i < sz; i++) {
-		drm_property_add_enum(rdev->mode_info.underscan_property,
-				      i,
-				      radeon_underscan_enum_list[i].type,
-				      radeon_underscan_enum_list[i].name);
-	}
+		drm_property_create_enum(rdev->ddev, 0,
+				    "underscan",
+				    radeon_underscan_enum_list, sz);
 
 	rdev->mode_info.underscan_hborder_property =
-		drm_property_create(rdev->ddev,
-					DRM_MODE_PROP_RANGE,
-					"underscan hborder", 2);
+		drm_property_create_range(rdev->ddev, 0,
+					"underscan hborder", 0, 128);
 	if (!rdev->mode_info.underscan_hborder_property)
 		return -ENOMEM;
-	rdev->mode_info.underscan_hborder_property->values[0] = 0;
-	rdev->mode_info.underscan_hborder_property->values[1] = 128;
 
 	rdev->mode_info.underscan_vborder_property =
-		drm_property_create(rdev->ddev,
-					DRM_MODE_PROP_RANGE,
-					"underscan vborder", 2);
+		drm_property_create_range(rdev->ddev, 0,
+					"underscan vborder", 0, 128);
 	if (!rdev->mode_info.underscan_vborder_property)
 		return -ENOMEM;
-	rdev->mode_info.underscan_vborder_property->values[0] = 0;
-	rdev->mode_info.underscan_vborder_property->values[1] = 128;
 
 	return 0;
 }
@@ -1278,6 +1263,9 @@ int radeon_modeset_init(struct radeon_device *rdev)
 		rdev->ddev->mode_config.max_height = 4096;
 	}
 
+	rdev->ddev->mode_config.preferred_depth = 24;
+	rdev->ddev->mode_config.prefer_shadow = 1;
+
 	rdev->ddev->mode_config.fb_base = rdev->mc.aper_base;
 
 	ret = radeon_modeset_create_props(rdev);
@@ -1308,7 +1296,7 @@ int radeon_modeset_init(struct radeon_device *rdev)
 	/* init dig PHYs, disp eng pll */
 	if (rdev->is_atom_bios) {
 		radeon_atom_encoder_init(rdev);
-		radeon_atom_dcpll_init(rdev);
+		radeon_atom_disp_eng_pll_init(rdev);
 	}
 
 	/* initialize hpd */
diff --git a/drivers/gpu/drm/radeon/radeon_drv.c b/drivers/gpu/drm/radeon/radeon_drv.c
index 8032f1fedb11..ef7bb3f6ecae 100644
--- a/drivers/gpu/drm/radeon/radeon_drv.c
+++ b/drivers/gpu/drm/radeon/radeon_drv.c
@@ -54,10 +54,12 @@
  *   2.10.0 - fusion 2D tiling
  *   2.11.0 - backend map, initial compute support for the CS checker
  *   2.12.0 - RADEON_CS_KEEP_TILING_FLAGS
- *   2.13.0 - virtual memory support
+ *   2.13.0 - virtual memory support, streamout
+ *   2.14.0 - add evergreen tiling informations
+ *   2.15.0 - add max_pipes query
  */
 #define KMS_DRIVER_MAJOR	2
-#define KMS_DRIVER_MINOR	13
+#define KMS_DRIVER_MINOR	15
 #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);
diff --git a/drivers/gpu/drm/radeon/radeon_encoders.c b/drivers/gpu/drm/radeon/radeon_encoders.c
index 9419c51bcf50..74670696277d 100644
--- a/drivers/gpu/drm/radeon/radeon_encoders.c
+++ b/drivers/gpu/drm/radeon/radeon_encoders.c
@@ -326,7 +326,7 @@ bool radeon_dig_monitor_is_duallink(struct drm_encoder *encoder,
 	case DRM_MODE_CONNECTOR_HDMIB:
 		if (radeon_connector->use_digital) {
 			/* HDMI 1.3 supports up to 340 Mhz over single link */
-			if (ASIC_IS_DCE3(rdev) && drm_detect_hdmi_monitor(radeon_connector->edid)) {
+			if (ASIC_IS_DCE6(rdev) && drm_detect_hdmi_monitor(radeon_connector->edid)) {
 				if (pixel_clock > 340000)
 					return true;
 				else
@@ -348,7 +348,7 @@ bool radeon_dig_monitor_is_duallink(struct drm_encoder *encoder,
 			return false;
 		else {
 			/* HDMI 1.3 supports up to 340 Mhz over single link */
-			if (ASIC_IS_DCE3(rdev) && drm_detect_hdmi_monitor(radeon_connector->edid)) {
+			if (ASIC_IS_DCE6(rdev) && drm_detect_hdmi_monitor(radeon_connector->edid)) {
 				if (pixel_clock > 340000)
 					return true;
 				else
diff --git a/drivers/gpu/drm/radeon/radeon_family.h b/drivers/gpu/drm/radeon/radeon_family.h
index ec2f1ea84f81..d1fafeabea09 100644
--- a/drivers/gpu/drm/radeon/radeon_family.h
+++ b/drivers/gpu/drm/radeon/radeon_family.h
@@ -87,6 +87,10 @@ enum radeon_family {
 	CHIP_TURKS,
 	CHIP_CAICOS,
 	CHIP_CAYMAN,
+	CHIP_ARUBA,
+	CHIP_TAHITI,
+	CHIP_PITCAIRN,
+	CHIP_VERDE,
 	CHIP_LAST,
 };
 
diff --git a/drivers/gpu/drm/radeon/radeon_fb.c b/drivers/gpu/drm/radeon/radeon_fb.c
index cf2bf35b56b8..5906914a78bc 100644
--- a/drivers/gpu/drm/radeon/radeon_fb.c
+++ b/drivers/gpu/drm/radeon/radeon_fb.c
@@ -164,7 +164,10 @@ static int radeonfb_create_pinned_object(struct radeon_fbdev *rfbdev,
 	ret = radeon_bo_reserve(rbo, false);
 	if (unlikely(ret != 0))
 		goto out_unref;
-	ret = radeon_bo_pin(rbo, RADEON_GEM_DOMAIN_VRAM, NULL);
+	/* Only 27 bit offset for legacy CRTC */
+	ret = radeon_bo_pin_restricted(rbo, RADEON_GEM_DOMAIN_VRAM,
+				       ASIC_IS_AVIVO(rdev) ? 0 : 1 << 27,
+				       NULL);
 	if (ret) {
 		radeon_bo_unreserve(rbo);
 		goto out_unref;
@@ -209,6 +212,11 @@ static int radeonfb_create(struct radeon_fbdev *rfbdev,
 							  sizes->surface_depth);
 
 	ret = radeonfb_create_pinned_object(rfbdev, &mode_cmd, &gobj);
+	if (ret) {
+		DRM_ERROR("failed to create fbcon object %d\n", ret);
+		return ret;
+	}
+
 	rbo = gem_to_radeon_bo(gobj);
 
 	/* okay we have an object now allocate the framebuffer */
@@ -220,7 +228,11 @@ static int radeonfb_create(struct radeon_fbdev *rfbdev,
 
 	info->par = rfbdev;
 
-	radeon_framebuffer_init(rdev->ddev, &rfbdev->rfb, &mode_cmd, gobj);
+	ret = radeon_framebuffer_init(rdev->ddev, &rfbdev->rfb, &mode_cmd, gobj);
+	if (ret) {
+		DRM_ERROR("failed to initalise framebuffer %d\n", ret);
+		goto out_unref;
+	}
 
 	fb = &rfbdev->rfb.base;
 
@@ -254,11 +266,7 @@ static int radeonfb_create(struct radeon_fbdev *rfbdev,
 	info->apertures->ranges[0].base = rdev->ddev->mode_config.fb_base;
 	info->apertures->ranges[0].size = rdev->mc.aper_size;
 
-	info->pixmap.size = 64*1024;
-	info->pixmap.buf_align = 8;
-	info->pixmap.access_align = 32;
-	info->pixmap.flags = FB_PIXMAP_SYSTEM;
-	info->pixmap.scan_align = 1;
+	/* Use default scratch pixmap (info->pixmap.flags = FB_PIXMAP_SYSTEM) */
 
 	if (info->screen_base == NULL) {
 		ret = -ENOSPC;
diff --git a/drivers/gpu/drm/radeon/radeon_fence.c b/drivers/gpu/drm/radeon/radeon_fence.c
index 64ea3dd9e6ff..4bd36a354fbe 100644
--- a/drivers/gpu/drm/radeon/radeon_fence.c
+++ b/drivers/gpu/drm/radeon/radeon_fence.c
@@ -364,8 +364,10 @@ int radeon_fence_count_emitted(struct radeon_device *rdev, int ring)
 	int not_processed = 0;
 
 	read_lock_irqsave(&rdev->fence_lock, irq_flags);
-	if (!rdev->fence_drv[ring].initialized)
+	if (!rdev->fence_drv[ring].initialized) {
+		read_unlock_irqrestore(&rdev->fence_lock, irq_flags);
 		return 0;
+	}
 
 	if (!list_empty(&rdev->fence_drv[ring].emitted)) {
 		struct list_head *ptr;
diff --git a/drivers/gpu/drm/radeon/radeon_gart.c b/drivers/gpu/drm/radeon/radeon_gart.c
index d5b1adc457af..456a77cf4b7f 100644
--- a/drivers/gpu/drm/radeon/radeon_gart.c
+++ b/drivers/gpu/drm/radeon/radeon_gart.c
@@ -597,13 +597,13 @@ int radeon_vm_bo_rmv(struct radeon_device *rdev,
 	if (bo_va == NULL)
 		return 0;
 
-	list_del(&bo_va->bo_list);
 	mutex_lock(&vm->mutex);
 	radeon_mutex_lock(&rdev->cs_mutex);
 	radeon_vm_bo_update_pte(rdev, vm, bo, NULL);
 	radeon_mutex_unlock(&rdev->cs_mutex);
 	list_del(&bo_va->vm_list);
 	mutex_unlock(&vm->mutex);
+	list_del(&bo_va->bo_list);
 
 	kfree(bo_va);
 	return 0;
diff --git a/drivers/gpu/drm/radeon/radeon_gem.c b/drivers/gpu/drm/radeon/radeon_gem.c
index 7337850af2fa..c7008b5210f7 100644
--- a/drivers/gpu/drm/radeon/radeon_gem.c
+++ b/drivers/gpu/drm/radeon/radeon_gem.c
@@ -75,32 +75,6 @@ int radeon_gem_object_create(struct radeon_device *rdev, int size,
 	return 0;
 }
 
-int radeon_gem_object_pin(struct drm_gem_object *obj, uint32_t pin_domain,
-			  uint64_t *gpu_addr)
-{
-	struct radeon_bo *robj = gem_to_radeon_bo(obj);
-	int r;
-
-	r = radeon_bo_reserve(robj, false);
-	if (unlikely(r != 0))
-		return r;
-	r = radeon_bo_pin(robj, pin_domain, gpu_addr);
-	radeon_bo_unreserve(robj);
-	return r;
-}
-
-void radeon_gem_object_unpin(struct drm_gem_object *obj)
-{
-	struct radeon_bo *robj = gem_to_radeon_bo(obj);
-	int r;
-
-	r = radeon_bo_reserve(robj, false);
-	if (likely(r == 0)) {
-		radeon_bo_unpin(robj);
-		radeon_bo_unreserve(robj);
-	}
-}
-
 int radeon_gem_set_domain(struct drm_gem_object *gobj,
 			  uint32_t rdomain, uint32_t wdomain)
 {
diff --git a/drivers/gpu/drm/radeon/radeon_i2c.c b/drivers/gpu/drm/radeon/radeon_i2c.c
index 98a8ad680109..85bcfc8923a7 100644
--- a/drivers/gpu/drm/radeon/radeon_i2c.c
+++ b/drivers/gpu/drm/radeon/radeon_i2c.c
@@ -26,10 +26,15 @@
 #include <linux/export.h>
 
 #include "drmP.h"
+#include "drm_edid.h"
 #include "radeon_drm.h"
 #include "radeon.h"
 #include "atom.h"
 
+extern int radeon_atom_hw_i2c_xfer(struct i2c_adapter *i2c_adap,
+				   struct i2c_msg *msgs, int num);
+extern u32 radeon_atom_hw_i2c_func(struct i2c_adapter *adap);
+
 /**
  * radeon_ddc_probe
  *
@@ -41,13 +46,13 @@ bool radeon_ddc_probe(struct radeon_connector *radeon_connector)
 	int ret;
 	struct i2c_msg msgs[] = {
 		{
-			.addr = 0x50,
+			.addr = DDC_ADDR,
 			.flags = 0,
 			.len = 1,
 			.buf = &out,
 		},
 		{
-			.addr = 0x50,
+			.addr = DDC_ADDR,
 			.flags = I2C_M_RD,
 			.len = 8,
 			.buf = buf,
@@ -882,6 +887,11 @@ static const struct i2c_algorithm radeon_i2c_algo = {
 	.functionality = radeon_hw_i2c_func,
 };
 
+static const struct i2c_algorithm radeon_atom_i2c_algo = {
+	.master_xfer = radeon_atom_hw_i2c_xfer,
+	.functionality = radeon_atom_hw_i2c_func,
+};
+
 struct radeon_i2c_chan *radeon_i2c_create(struct drm_device *dev,
 					  struct radeon_i2c_bus_rec *rec,
 					  const char *name)
@@ -914,6 +924,18 @@ struct radeon_i2c_chan *radeon_i2c_create(struct drm_device *dev,
 			DRM_ERROR("Failed to register hw i2c %s\n", name);
 			goto out_free;
 		}
+	} else if (rec->hw_capable &&
+		   radeon_hw_i2c &&
+		   ASIC_IS_DCE3(rdev)) {
+		/* hw i2c using atom */
+		snprintf(i2c->adapter.name, sizeof(i2c->adapter.name),
+			 "Radeon i2c hw bus %s", name);
+		i2c->adapter.algo = &radeon_atom_i2c_algo;
+		ret = i2c_add_adapter(&i2c->adapter);
+		if (ret) {
+			DRM_ERROR("Failed to register hw i2c %s\n", name);
+			goto out_free;
+		}
 	} else {
 		/* set the radeon bit adapter */
 		snprintf(i2c->adapter.name, sizeof(i2c->adapter.name),
@@ -925,10 +947,8 @@ struct radeon_i2c_chan *radeon_i2c_create(struct drm_device *dev,
 		i2c->algo.bit.setscl = set_clock;
 		i2c->algo.bit.getsda = get_data;
 		i2c->algo.bit.getscl = get_clock;
-		i2c->algo.bit.udelay = 20;
-		/* vesa says 2.2 ms is enough, 1 jiffy doesn't seem to always
-		 * make this, 2 jiffies is a lot more reliable */
-		i2c->algo.bit.timeout = 2;
+		i2c->algo.bit.udelay = 10;
+		i2c->algo.bit.timeout = usecs_to_jiffies(2200);	/* from VESA */
 		i2c->algo.bit.data = i2c;
 		ret = i2c_bit_add_bus(&i2c->adapter);
 		if (ret) {
diff --git a/drivers/gpu/drm/radeon/radeon_kms.c b/drivers/gpu/drm/radeon/radeon_kms.c
index d3352889a870..3c2628b14d56 100644
--- a/drivers/gpu/drm/radeon/radeon_kms.c
+++ b/drivers/gpu/drm/radeon/radeon_kms.c
@@ -57,6 +57,8 @@ int radeon_driver_load_kms(struct drm_device *dev, unsigned long flags)
 	}
 	dev->dev_private = (void *)rdev;
 
+	pci_set_master(dev->pdev);
+
 	/* update BUS flag */
 	if (drm_pci_device_is_agp(dev)) {
 		flags |= RADEON_IS_AGP;
@@ -169,7 +171,9 @@ int radeon_info_ioctl(struct drm_device *dev, void *data, struct drm_file *filp)
 		value = rdev->accel_working;
 		break;
 	case RADEON_INFO_TILING_CONFIG:
-		if (rdev->family >= CHIP_CAYMAN)
+		if (rdev->family >= CHIP_TAHITI)
+			value = rdev->config.si.tile_config;
+		else if (rdev->family >= CHIP_CAYMAN)
 			value = rdev->config.cayman.tile_config;
 		else if (rdev->family >= CHIP_CEDAR)
 			value = rdev->config.evergreen.tile_config;
@@ -208,7 +212,10 @@ int radeon_info_ioctl(struct drm_device *dev, void *data, struct drm_file *filp)
 		value = rdev->clock.spll.reference_freq * 10;
 		break;
 	case RADEON_INFO_NUM_BACKENDS:
-		if (rdev->family >= CHIP_CAYMAN)
+		if (rdev->family >= CHIP_TAHITI)
+			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 *
 				rdev->config.cayman.max_shader_engines;
 		else if (rdev->family >= CHIP_CEDAR)
@@ -222,7 +229,9 @@ int radeon_info_ioctl(struct drm_device *dev, void *data, struct drm_file *filp)
 		}
 		break;
 	case RADEON_INFO_NUM_TILE_PIPES:
-		if (rdev->family >= CHIP_CAYMAN)
+		if (rdev->family >= CHIP_TAHITI)
+			value = rdev->config.si.max_tile_pipes;
+		else if (rdev->family >= CHIP_CAYMAN)
 			value = rdev->config.cayman.max_tile_pipes;
 		else if (rdev->family >= CHIP_CEDAR)
 			value = rdev->config.evergreen.max_tile_pipes;
@@ -238,7 +247,9 @@ int radeon_info_ioctl(struct drm_device *dev, void *data, struct drm_file *filp)
 		value = 1;
 		break;
 	case RADEON_INFO_BACKEND_MAP:
-		if (rdev->family >= CHIP_CAYMAN)
+		if (rdev->family >= CHIP_TAHITI)
+			value = rdev->config.si.backend_map;
+		else if (rdev->family >= CHIP_CAYMAN)
 			value = rdev->config.cayman.backend_map;
 		else if (rdev->family >= CHIP_CEDAR)
 			value = rdev->config.evergreen.backend_map;
@@ -262,6 +273,21 @@ int radeon_info_ioctl(struct drm_device *dev, void *data, struct drm_file *filp)
 			return -EINVAL;
 		value = RADEON_IB_VM_MAX_SIZE;
 		break;
+	case RADEON_INFO_MAX_PIPES:
+		if (rdev->family >= CHIP_TAHITI)
+			value = rdev->config.si.max_pipes_per_simd;
+		else if (rdev->family >= CHIP_CAYMAN)
+			value = rdev->config.cayman.max_pipes_per_simd;
+		else if (rdev->family >= CHIP_CEDAR)
+			value = rdev->config.evergreen.max_pipes;
+		else if (rdev->family >= CHIP_RV770)
+			value = rdev->config.rv770.max_pipes;
+		else if (rdev->family >= CHIP_R600)
+			value = rdev->config.r600.max_pipes;
+		else {
+			return -EINVAL;
+		}
+		break;
 	default:
 		DRM_DEBUG_KMS("Invalid request %d\n", info->request);
 		return -EINVAL;
diff --git a/drivers/gpu/drm/radeon/radeon_legacy_crtc.c b/drivers/gpu/drm/radeon/radeon_legacy_crtc.c
index 25a19c483075..210317c7045e 100644
--- a/drivers/gpu/drm/radeon/radeon_legacy_crtc.c
+++ b/drivers/gpu/drm/radeon/radeon_legacy_crtc.c
@@ -419,7 +419,9 @@ int radeon_crtc_do_set_base(struct drm_crtc *crtc,
 	r = radeon_bo_reserve(rbo, false);
 	if (unlikely(r != 0))
 		return r;
-	r = radeon_bo_pin(rbo, RADEON_GEM_DOMAIN_VRAM, &base);
+	/* Only 27 bit offset for legacy CRTC */
+	r = radeon_bo_pin_restricted(rbo, RADEON_GEM_DOMAIN_VRAM, 1 << 27,
+				     &base);
 	if (unlikely(r != 0)) {
 		radeon_bo_unreserve(rbo);
 		return -EINVAL;
diff --git a/drivers/gpu/drm/radeon/radeon_mode.h b/drivers/gpu/drm/radeon/radeon_mode.h
index 4330e3253573..f7eb5d8b9fd3 100644
--- a/drivers/gpu/drm/radeon/radeon_mode.h
+++ b/drivers/gpu/drm/radeon/radeon_mode.h
@@ -491,7 +491,7 @@ extern int radeon_dp_get_panel_mode(struct drm_encoder *encoder,
 				    struct drm_connector *connector);
 extern void atombios_dig_encoder_setup(struct drm_encoder *encoder, int action, int panel_mode);
 extern void radeon_atom_encoder_init(struct radeon_device *rdev);
-extern void radeon_atom_dcpll_init(struct radeon_device *rdev);
+extern void radeon_atom_disp_eng_pll_init(struct radeon_device *rdev);
 extern void atombios_dig_transmitter_setup(struct drm_encoder *encoder,
 					   int action, uint8_t lane_num,
 					   uint8_t lane_set);
@@ -649,7 +649,7 @@ extern void radeon_crtc_fb_gamma_set(struct drm_crtc *crtc, u16 red, u16 green,
 				     u16 blue, int regno);
 extern void radeon_crtc_fb_gamma_get(struct drm_crtc *crtc, u16 *red, u16 *green,
 				     u16 *blue, int regno);
-void radeon_framebuffer_init(struct drm_device *dev,
+int radeon_framebuffer_init(struct drm_device *dev,
 			     struct radeon_framebuffer *rfb,
 			     struct drm_mode_fb_cmd2 *mode_cmd,
 			     struct drm_gem_object *obj);
diff --git a/drivers/gpu/drm/radeon/radeon_object.c b/drivers/gpu/drm/radeon/radeon_object.c
index d45df1763598..df6a4dbd93f8 100644
--- a/drivers/gpu/drm/radeon/radeon_object.c
+++ b/drivers/gpu/drm/radeon/radeon_object.c
@@ -224,7 +224,8 @@ void radeon_bo_unref(struct radeon_bo **bo)
 		*bo = NULL;
 }
 
-int radeon_bo_pin(struct radeon_bo *bo, u32 domain, u64 *gpu_addr)
+int radeon_bo_pin_restricted(struct radeon_bo *bo, u32 domain, u64 max_offset,
+			     u64 *gpu_addr)
 {
 	int r, i;
 
@@ -232,6 +233,18 @@ int radeon_bo_pin(struct radeon_bo *bo, u32 domain, u64 *gpu_addr)
 		bo->pin_count++;
 		if (gpu_addr)
 			*gpu_addr = radeon_bo_gpu_offset(bo);
+
+		if (max_offset != 0) {
+			u64 domain_start;
+
+			if (domain == RADEON_GEM_DOMAIN_VRAM)
+				domain_start = bo->rdev->mc.vram_start;
+			else
+				domain_start = bo->rdev->mc.gtt_start;
+			WARN_ON_ONCE(max_offset <
+				     (radeon_bo_gpu_offset(bo) - domain_start));
+		}
+
 		return 0;
 	}
 	radeon_ttm_placement_from_domain(bo, domain);
@@ -239,6 +252,15 @@ int radeon_bo_pin(struct radeon_bo *bo, u32 domain, u64 *gpu_addr)
 		/* force to pin into visible video ram */
 		bo->placement.lpfn = bo->rdev->mc.visible_vram_size >> PAGE_SHIFT;
 	}
+	if (max_offset) {
+		u64 lpfn = max_offset >> PAGE_SHIFT;
+
+		if (!bo->placement.lpfn)
+			bo->placement.lpfn = bo->rdev->mc.gtt_size >> PAGE_SHIFT;
+
+		if (lpfn < bo->placement.lpfn)
+			bo->placement.lpfn = lpfn;
+	}
 	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, false);
@@ -252,6 +274,11 @@ int radeon_bo_pin(struct radeon_bo *bo, u32 domain, u64 *gpu_addr)
 	return r;
 }
 
+int radeon_bo_pin(struct radeon_bo *bo, u32 domain, u64 *gpu_addr)
+{
+	return radeon_bo_pin_restricted(bo, domain, 0, gpu_addr);
+}
+
 int radeon_bo_unpin(struct radeon_bo *bo)
 {
 	int r, i;
@@ -445,8 +472,54 @@ static void radeon_bo_clear_surface_reg(struct radeon_bo *bo)
 int radeon_bo_set_tiling_flags(struct radeon_bo *bo,
 				uint32_t tiling_flags, uint32_t pitch)
 {
+	struct radeon_device *rdev = bo->rdev;
 	int r;
 
+	if (rdev->family >= CHIP_CEDAR) {
+		unsigned bankw, bankh, mtaspect, tilesplit, stilesplit;
+
+		bankw = (tiling_flags >> RADEON_TILING_EG_BANKW_SHIFT) & RADEON_TILING_EG_BANKW_MASK;
+		bankh = (tiling_flags >> RADEON_TILING_EG_BANKH_SHIFT) & RADEON_TILING_EG_BANKH_MASK;
+		mtaspect = (tiling_flags >> RADEON_TILING_EG_MACRO_TILE_ASPECT_SHIFT) & RADEON_TILING_EG_MACRO_TILE_ASPECT_MASK;
+		tilesplit = (tiling_flags >> RADEON_TILING_EG_TILE_SPLIT_SHIFT) & RADEON_TILING_EG_TILE_SPLIT_MASK;
+		stilesplit = (tiling_flags >> RADEON_TILING_EG_STENCIL_TILE_SPLIT_SHIFT) & RADEON_TILING_EG_STENCIL_TILE_SPLIT_MASK;
+		switch (bankw) {
+		case 0:
+		case 1:
+		case 2:
+		case 4:
+		case 8:
+			break;
+		default:
+			return -EINVAL;
+		}
+		switch (bankh) {
+		case 0:
+		case 1:
+		case 2:
+		case 4:
+		case 8:
+			break;
+		default:
+			return -EINVAL;
+		}
+		switch (mtaspect) {
+		case 0:
+		case 1:
+		case 2:
+		case 4:
+		case 8:
+			break;
+		default:
+			return -EINVAL;
+		}
+		if (tilesplit > 6) {
+			return -EINVAL;
+		}
+		if (stilesplit > 6) {
+			return -EINVAL;
+		}
+	}
 	r = radeon_bo_reserve(bo, false);
 	if (unlikely(r != 0))
 		return r;
diff --git a/drivers/gpu/drm/radeon/radeon_object.h b/drivers/gpu/drm/radeon/radeon_object.h
index cde430308870..f9104be88d7c 100644
--- a/drivers/gpu/drm/radeon/radeon_object.h
+++ b/drivers/gpu/drm/radeon/radeon_object.h
@@ -118,6 +118,8 @@ extern int radeon_bo_kmap(struct radeon_bo *bo, void **ptr);
 extern void radeon_bo_kunmap(struct radeon_bo *bo);
 extern void radeon_bo_unref(struct radeon_bo **bo);
 extern int radeon_bo_pin(struct radeon_bo *bo, u32 domain, u64 *gpu_addr);
+extern int radeon_bo_pin_restricted(struct radeon_bo *bo, u32 domain,
+				    u64 max_offset, u64 *gpu_addr);
 extern int radeon_bo_unpin(struct radeon_bo *bo);
 extern int radeon_bo_evict_vram(struct radeon_device *rdev);
 extern void radeon_bo_force_delete(struct radeon_device *rdev);
diff --git a/drivers/gpu/drm/radeon/radeon_pm.c b/drivers/gpu/drm/radeon/radeon_pm.c
index 095148e29a1f..caa55d68f319 100644
--- a/drivers/gpu/drm/radeon/radeon_pm.c
+++ b/drivers/gpu/drm/radeon/radeon_pm.c
@@ -221,7 +221,7 @@ static void radeon_set_power_state(struct radeon_device *rdev)
 		}
 
 		/* set memory clock */
-		if (rdev->asic->set_memory_clock && (mclk != rdev->pm.current_mclk)) {
+		if (rdev->asic->pm.set_memory_clock && (mclk != rdev->pm.current_mclk)) {
 			radeon_pm_debug_check_in_vbl(rdev, false);
 			radeon_set_memory_clock(rdev, mclk);
 			radeon_pm_debug_check_in_vbl(rdev, true);
@@ -474,6 +474,9 @@ static ssize_t radeon_hwmon_show_temp(struct device *dev,
 	case THERMAL_TYPE_SUMO:
 		temp = sumo_get_temp(rdev);
 		break;
+	case THERMAL_TYPE_SI:
+		temp = si_get_temp(rdev);
+		break;
 	default:
 		temp = 0;
 		break;
@@ -514,6 +517,10 @@ static int radeon_hwmon_init(struct radeon_device *rdev)
 	case THERMAL_TYPE_EVERGREEN:
 	case THERMAL_TYPE_NI:
 	case THERMAL_TYPE_SUMO:
+	case THERMAL_TYPE_SI:
+		/* No support for TN yet */
+		if (rdev->family == CHIP_ARUBA)
+			return err;
 		rdev->pm.int_hwmon_dev = hwmon_device_register(rdev->dev);
 		if (IS_ERR(rdev->pm.int_hwmon_dev)) {
 			err = PTR_ERR(rdev->pm.int_hwmon_dev);
@@ -863,11 +870,11 @@ static int radeon_debugfs_pm_info(struct seq_file *m, void *data)
 	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));
 	seq_printf(m, "default memory clock: %u0 kHz\n", rdev->pm.default_mclk);
-	if (rdev->asic->get_memory_clock)
+	if (rdev->asic->pm.get_memory_clock)
 		seq_printf(m, "current memory clock: %u0 kHz\n", radeon_get_memory_clock(rdev));
 	if (rdev->pm.current_vddc)
 		seq_printf(m, "voltage: %u mV\n", rdev->pm.current_vddc);
-	if (rdev->asic->get_pcie_lanes)
+	if (rdev->asic->pm.get_pcie_lanes)
 		seq_printf(m, "PCIE lanes: %d\n", radeon_get_pcie_lanes(rdev));
 
 	return 0;
diff --git a/drivers/gpu/drm/radeon/radeon_reg.h b/drivers/gpu/drm/radeon/radeon_reg.h
index b4ce86455707..5d8f735d6aaf 100644
--- a/drivers/gpu/drm/radeon/radeon_reg.h
+++ b/drivers/gpu/drm/radeon/radeon_reg.h
@@ -56,6 +56,7 @@
 #include "r600_reg.h"
 #include "evergreen_reg.h"
 #include "ni_reg.h"
+#include "si_reg.h"
 
 #define RADEON_MC_AGP_LOCATION		0x014c
 #define		RADEON_MC_AGP_START_MASK	0x0000FFFF
@@ -539,9 +540,11 @@
 
 #define RADEON_CRTC2_PITCH                  0x032c
 #define RADEON_CRTC_STATUS                  0x005c
+#       define RADEON_CRTC_VBLANK_CUR       (1 <<  0)
 #       define RADEON_CRTC_VBLANK_SAVE      (1 <<  1)
 #       define RADEON_CRTC_VBLANK_SAVE_CLEAR  (1 <<  1)
 #define RADEON_CRTC2_STATUS                  0x03fc
+#       define RADEON_CRTC2_VBLANK_CUR       (1 <<  0)
 #       define RADEON_CRTC2_VBLANK_SAVE      (1 <<  1)
 #       define RADEON_CRTC2_VBLANK_SAVE_CLEAR  (1 <<  1)
 #define RADEON_CRTC_V_SYNC_STRT_WID         0x020c
diff --git a/drivers/gpu/drm/radeon/radeon_ring.c b/drivers/gpu/drm/radeon/radeon_ring.c
index 30a4c5014c8b..cc33b3d7c33b 100644
--- a/drivers/gpu/drm/radeon/radeon_ring.c
+++ b/drivers/gpu/drm/radeon/radeon_ring.c
@@ -133,6 +133,7 @@ retry:
 				(*ib)->gpu_addr += (*ib)->sa_bo.offset;
 				(*ib)->fence = fence;
 				(*ib)->vm_id = 0;
+				(*ib)->is_const_ib = false;
 				/* ib are most likely to be allocated in a ring fashion
 				 * thus rdev->ib_pool.head_id should be the id of the
 				 * oldest ib
@@ -478,7 +479,9 @@ static struct drm_info_list radeon_debugfs_ring_info_list[] = {
 static int radeon_debugfs_ib_info(struct seq_file *m, void *data)
 {
 	struct drm_info_node *node = (struct drm_info_node *) m->private;
-	struct radeon_ib *ib = node->info_ent->data;
+	struct drm_device *dev = node->minor->dev;
+	struct radeon_device *rdev = dev->dev_private;
+	struct radeon_ib *ib = &rdev->ib_pool.ibs[*((unsigned*)node->info_ent->data)];
 	unsigned i;
 
 	if (ib == NULL) {
@@ -495,13 +498,17 @@ static int radeon_debugfs_ib_info(struct seq_file *m, void *data)
 
 static struct drm_info_list radeon_debugfs_ib_list[RADEON_IB_POOL_SIZE];
 static char radeon_debugfs_ib_names[RADEON_IB_POOL_SIZE][32];
+static unsigned radeon_debugfs_ib_idx[RADEON_IB_POOL_SIZE];
 #endif
 
 int radeon_debugfs_ring_init(struct radeon_device *rdev)
 {
 #if defined(CONFIG_DEBUG_FS)
-	return radeon_debugfs_add_files(rdev, radeon_debugfs_ring_info_list,
-					ARRAY_SIZE(radeon_debugfs_ring_info_list));
+	if (rdev->family >= CHIP_CAYMAN)
+		return radeon_debugfs_add_files(rdev, radeon_debugfs_ring_info_list,
+						ARRAY_SIZE(radeon_debugfs_ring_info_list));
+	else
+		return radeon_debugfs_add_files(rdev, radeon_debugfs_ring_info_list, 1);
 #else
 	return 0;
 #endif
@@ -514,10 +521,11 @@ int radeon_debugfs_ib_init(struct radeon_device *rdev)
 
 	for (i = 0; i < RADEON_IB_POOL_SIZE; i++) {
 		sprintf(radeon_debugfs_ib_names[i], "radeon_ib_%04u", i);
+		radeon_debugfs_ib_idx[i] = i;
 		radeon_debugfs_ib_list[i].name = radeon_debugfs_ib_names[i];
 		radeon_debugfs_ib_list[i].show = &radeon_debugfs_ib_info;
 		radeon_debugfs_ib_list[i].driver_features = 0;
-		radeon_debugfs_ib_list[i].data = &rdev->ib_pool.ibs[i];
+		radeon_debugfs_ib_list[i].data = &radeon_debugfs_ib_idx[i];
 	}
 	return radeon_debugfs_add_files(rdev, radeon_debugfs_ib_list,
 					RADEON_IB_POOL_SIZE);
diff --git a/drivers/gpu/drm/radeon/radeon_ttm.c b/drivers/gpu/drm/radeon/radeon_ttm.c
index c421e77ace71..f493c6403af5 100644
--- a/drivers/gpu/drm/radeon/radeon_ttm.c
+++ b/drivers/gpu/drm/radeon/radeon_ttm.c
@@ -226,7 +226,7 @@ static int radeon_move_blit(struct ttm_buffer_object *bo,
 	int r, i;
 
 	rdev = radeon_get_rdev(bo->bdev);
-	r = radeon_fence_create(rdev, &fence, rdev->copy_ring);
+	r = radeon_fence_create(rdev, &fence, radeon_copy_ring_index(rdev));
 	if (unlikely(r)) {
 		return r;
 	}
@@ -255,7 +255,7 @@ static int radeon_move_blit(struct ttm_buffer_object *bo,
 		DRM_ERROR("Unknown placement %d\n", old_mem->mem_type);
 		return -EINVAL;
 	}
-	if (!rdev->ring[rdev->copy_ring].ready) {
+	if (!rdev->ring[radeon_copy_ring_index(rdev)].ready) {
 		DRM_ERROR("Trying to move memory with ring turned off.\n");
 		return -EINVAL;
 	}
@@ -266,7 +266,7 @@ static int radeon_move_blit(struct ttm_buffer_object *bo,
 	if (rdev->family >= CHIP_R600) {
 		for (i = 0; i < RADEON_NUM_RINGS; ++i) {
 			/* no need to sync to our own or unused rings */
-			if (i == rdev->copy_ring || !rdev->ring[i].ready)
+			if (i == radeon_copy_ring_index(rdev) || !rdev->ring[i].ready)
 				continue;
 
 			if (!fence->semaphore) {
@@ -283,12 +283,12 @@ static int radeon_move_blit(struct ttm_buffer_object *bo,
 			radeon_semaphore_emit_signal(rdev, i, fence->semaphore);
 			radeon_ring_unlock_commit(rdev, &rdev->ring[i]);
 
-			r = radeon_ring_lock(rdev, &rdev->ring[rdev->copy_ring], 3);
+			r = radeon_ring_lock(rdev, &rdev->ring[radeon_copy_ring_index(rdev)], 3);
 			/* FIXME: handle ring lock error */
 			if (r)
 				continue;
-			radeon_semaphore_emit_wait(rdev, rdev->copy_ring, fence->semaphore);
-			radeon_ring_unlock_commit(rdev, &rdev->ring[rdev->copy_ring]);
+			radeon_semaphore_emit_wait(rdev, radeon_copy_ring_index(rdev), fence->semaphore);
+			radeon_ring_unlock_commit(rdev, &rdev->ring[radeon_copy_ring_index(rdev)]);
 		}
 	}
 
@@ -410,7 +410,8 @@ static int radeon_bo_move(struct ttm_buffer_object *bo,
 		radeon_move_null(bo, new_mem);
 		return 0;
 	}
-	if (!rdev->ring[RADEON_RING_TYPE_GFX_INDEX].ready || rdev->asic->copy == NULL) {
+	if (!rdev->ring[radeon_copy_ring_index(rdev)].ready ||
+	    rdev->asic->copy.copy == NULL) {
 		/* use memcpy */
 		goto memcpy;
 	}
diff --git a/drivers/gpu/drm/radeon/reg_srcs/cayman b/drivers/gpu/drm/radeon/reg_srcs/cayman
index 2316977eb924..0f656b111c15 100644
--- a/drivers/gpu/drm/radeon/reg_srcs/cayman
+++ b/drivers/gpu/drm/radeon/reg_srcs/cayman
@@ -1,5 +1,8 @@
 cayman 0x9400
 0x0000802C GRBM_GFX_INDEX
+0x000084FC CP_STRMOUT_CNTL
+0x000085F0 CP_COHER_CNTL
+0x000085F4 CP_COHER_SIZE
 0x000088B0 VGT_VTX_VECT_EJECT_REG
 0x000088C4 VGT_CACHE_INVALIDATION
 0x000088D4 VGT_GS_VERTEX_REUSE
@@ -77,7 +80,6 @@ cayman 0x9400
 0x0002802C DB_DEPTH_CLEAR
 0x00028030 PA_SC_SCREEN_SCISSOR_TL
 0x00028034 PA_SC_SCREEN_SCISSOR_BR
-0x0002805C DB_DEPTH_SLICE
 0x00028140 SQ_ALU_CONST_BUFFER_SIZE_PS_0
 0x00028144 SQ_ALU_CONST_BUFFER_SIZE_PS_1
 0x00028148 SQ_ALU_CONST_BUFFER_SIZE_PS_2
@@ -206,7 +208,6 @@ cayman 0x9400
 0x00028344 PA_SC_VPORT_ZMAX_14
 0x00028348 PA_SC_VPORT_ZMIN_15
 0x0002834C PA_SC_VPORT_ZMAX_15
-0x00028350 SX_MISC
 0x00028354 SX_SURFACE_SYNC
 0x0002835C SX_SCATTER_EXPORT_SIZE
 0x00028380 SQ_VTX_SEMANTIC_0
@@ -508,10 +509,16 @@ cayman 0x9400
 0x00028AA8 IA_MULTI_VGT_PARAM
 0x00028AB4 VGT_REUSE_OFF
 0x00028AB8 VGT_VTX_CNT_EN
-0x00028ABC DB_HTILE_SURFACE
 0x00028AC0 DB_SRESULTS_COMPARE_STATE0
 0x00028AC4 DB_SRESULTS_COMPARE_STATE1
 0x00028AC8 DB_PRELOAD_CONTROL
+0x00028AD4 VGT_STRMOUT_VTX_STRIDE_0
+0x00028AE4 VGT_STRMOUT_VTX_STRIDE_1
+0x00028AF4 VGT_STRMOUT_VTX_STRIDE_2
+0x00028B04 VGT_STRMOUT_VTX_STRIDE_3
+0x00028B28 VGT_STRMOUT_DRAW_OPAQUE_OFFSET
+0x00028B2C VGT_STRMOUT_DRAW_OPAQUE_BUFFER_FILLED_SIZE
+0x00028B30 VGT_STRMOUT_DRAW_OPAQUE_VERTEX_STRIDE
 0x00028B38 VGT_GS_MAX_VERT_OUT
 0x00028B54 VGT_SHADER_STAGES_EN
 0x00028B58 VGT_LS_HS_CONFIG
@@ -551,6 +558,18 @@ cayman 0x9400
 0x00028C34 PA_SC_AA_SAMPLE_LOCS_PIXEL_X1_Y1_3
 0x00028C38 PA_SC_AA_MASK_X0_Y0_X1_Y0
 0x00028C3C PA_SC_AA_MASK_X0_Y1_X1_Y1
+0x00028C78 CB_COLOR0_DIM
+0x00028CB4 CB_COLOR1_DIM
+0x00028CF0 CB_COLOR2_DIM
+0x00028D2C CB_COLOR3_DIM
+0x00028D68 CB_COLOR4_DIM
+0x00028DA4 CB_COLOR5_DIM
+0x00028DE0 CB_COLOR6_DIM
+0x00028E1C CB_COLOR7_DIM
+0x00028E58 CB_COLOR8_DIM
+0x00028E74 CB_COLOR9_DIM
+0x00028E90 CB_COLOR10_DIM
+0x00028EAC CB_COLOR11_DIM
 0x00028C8C CB_COLOR0_CLEAR_WORD0
 0x00028C90 CB_COLOR0_CLEAR_WORD1
 0x00028C94 CB_COLOR0_CLEAR_WORD2
diff --git a/drivers/gpu/drm/radeon/reg_srcs/evergreen b/drivers/gpu/drm/radeon/reg_srcs/evergreen
index 161737a28c23..b912a37689bf 100644
--- a/drivers/gpu/drm/radeon/reg_srcs/evergreen
+++ b/drivers/gpu/drm/radeon/reg_srcs/evergreen
@@ -4,6 +4,9 @@ evergreen 0x9400
 0x00008044 WAIT_UNTIL_POLL_CNTL
 0x00008048 WAIT_UNTIL_POLL_MASK
 0x0000804c WAIT_UNTIL_POLL_REFDATA
+0x000084FC CP_STRMOUT_CNTL
+0x000085F0 CP_COHER_CNTL
+0x000085F4 CP_COHER_SIZE
 0x000088B0 VGT_VTX_VECT_EJECT_REG
 0x000088C4 VGT_CACHE_INVALIDATION
 0x000088D4 VGT_GS_VERTEX_REUSE
@@ -93,7 +96,6 @@ evergreen 0x9400
 0x0002802C DB_DEPTH_CLEAR
 0x00028030 PA_SC_SCREEN_SCISSOR_TL
 0x00028034 PA_SC_SCREEN_SCISSOR_BR
-0x0002805C DB_DEPTH_SLICE
 0x00028140 SQ_ALU_CONST_BUFFER_SIZE_PS_0
 0x00028144 SQ_ALU_CONST_BUFFER_SIZE_PS_1
 0x00028148 SQ_ALU_CONST_BUFFER_SIZE_PS_2
@@ -222,7 +224,6 @@ evergreen 0x9400
 0x00028344 PA_SC_VPORT_ZMAX_14
 0x00028348 PA_SC_VPORT_ZMIN_15
 0x0002834C PA_SC_VPORT_ZMAX_15
-0x00028350 SX_MISC
 0x00028354 SX_SURFACE_SYNC
 0x00028380 SQ_VTX_SEMANTIC_0
 0x00028384 SQ_VTX_SEMANTIC_1
@@ -518,10 +519,16 @@ evergreen 0x9400
 0x00028AA4 VGT_INSTANCE_STEP_RATE_1
 0x00028AB4 VGT_REUSE_OFF
 0x00028AB8 VGT_VTX_CNT_EN
-0x00028ABC DB_HTILE_SURFACE
 0x00028AC0 DB_SRESULTS_COMPARE_STATE0
 0x00028AC4 DB_SRESULTS_COMPARE_STATE1
 0x00028AC8 DB_PRELOAD_CONTROL
+0x00028AD4 VGT_STRMOUT_VTX_STRIDE_0
+0x00028AE4 VGT_STRMOUT_VTX_STRIDE_1
+0x00028AF4 VGT_STRMOUT_VTX_STRIDE_2
+0x00028B04 VGT_STRMOUT_VTX_STRIDE_3
+0x00028B28 VGT_STRMOUT_DRAW_OPAQUE_OFFSET
+0x00028B2C VGT_STRMOUT_DRAW_OPAQUE_BUFFER_FILLED_SIZE
+0x00028B30 VGT_STRMOUT_DRAW_OPAQUE_VERTEX_STRIDE
 0x00028B38 VGT_GS_MAX_VERT_OUT
 0x00028B54 VGT_SHADER_STAGES_EN
 0x00028B58 VGT_LS_HS_CONFIG
@@ -554,6 +561,18 @@ evergreen 0x9400
 0x00028C34 PA_SC_AA_SAMPLE_LOCS_6
 0x00028C38 PA_SC_AA_SAMPLE_LOCS_7
 0x00028C3C PA_SC_AA_MASK
+0x00028C78 CB_COLOR0_DIM
+0x00028CB4 CB_COLOR1_DIM
+0x00028CF0 CB_COLOR2_DIM
+0x00028D2C CB_COLOR3_DIM
+0x00028D68 CB_COLOR4_DIM
+0x00028DA4 CB_COLOR5_DIM
+0x00028DE0 CB_COLOR6_DIM
+0x00028E1C CB_COLOR7_DIM
+0x00028E58 CB_COLOR8_DIM
+0x00028E74 CB_COLOR9_DIM
+0x00028E90 CB_COLOR10_DIM
+0x00028EAC CB_COLOR11_DIM
 0x00028C8C CB_COLOR0_CLEAR_WORD0
 0x00028C90 CB_COLOR0_CLEAR_WORD1
 0x00028C94 CB_COLOR0_CLEAR_WORD2
diff --git a/drivers/gpu/drm/radeon/reg_srcs/r600 b/drivers/gpu/drm/radeon/reg_srcs/r600
index 0380c5c15f80..5e659b034d9a 100644
--- a/drivers/gpu/drm/radeon/reg_srcs/r600
+++ b/drivers/gpu/drm/radeon/reg_srcs/r600
@@ -3,6 +3,9 @@ r600 0x9400
 0x00028230 R7xx_PA_SC_EDGERULE
 0x000286C8 R7xx_SPI_THREAD_GROUPING
 0x00008D8C R7xx_SQ_DYN_GPR_CNTL_PS_FLUSH_REQ
+0x00008490 CP_STRMOUT_CNTL
+0x000085F0 CP_COHER_CNTL
+0x000085F4 CP_COHER_SIZE
 0x000088C4 VGT_CACHE_INVALIDATION
 0x00028A50 VGT_ENHANCE
 0x000088CC VGT_ES_PER_GS
@@ -38,6 +41,13 @@ r600 0x9400
 0x00028AB4 VGT_REUSE_OFF
 0x00028AB8 VGT_VTX_CNT_EN
 0x000088B0 VGT_VTX_VECT_EJECT_REG
+0x00028AD4 VGT_STRMOUT_VTX_STRIDE_0
+0x00028AE4 VGT_STRMOUT_VTX_STRIDE_1
+0x00028AF4 VGT_STRMOUT_VTX_STRIDE_2
+0x00028B04 VGT_STRMOUT_VTX_STRIDE_3
+0x00028B28 VGT_STRMOUT_DRAW_OPAQUE_OFFSET
+0x00028B2C VGT_STRMOUT_DRAW_OPAQUE_BUFFER_FILLED_SIZE
+0x00028B30 VGT_STRMOUT_DRAW_OPAQUE_VERTEX_STRIDE
 0x00028810 PA_CL_CLIP_CNTL
 0x00008A14 PA_CL_ENHANCE
 0x00028C14 PA_CL_GB_HORZ_CLIP_ADJ
@@ -428,7 +438,7 @@ r600 0x9400
 0x00028638 SPI_VS_OUT_ID_9
 0x00028438 SX_ALPHA_REF
 0x00028410 SX_ALPHA_TEST_CONTROL
-0x00028350 SX_MISC
+0x00028354 SX_SURFACE_SYNC
 0x00009014 SX_MEMORY_EXPORT_SIZE
 0x00009604 TC_INVALIDATE
 0x00009400 TD_FILTER4
@@ -703,7 +713,6 @@ r600 0x9400
 0x0000A710 TD_VS_SAMPLER17_BORDER_RED
 0x00009508 TA_CNTL_AUX
 0x0002802C DB_DEPTH_CLEAR
-0x00028D24 DB_HTILE_SURFACE
 0x00028D34 DB_PREFETCH_LIMIT
 0x00028D30 DB_PRELOAD_CONTROL
 0x00028D0C DB_RENDER_CONTROL
@@ -743,14 +752,6 @@ r600 0x9400
 0x00028114 CB_COLOR5_MASK
 0x00028118 CB_COLOR6_MASK
 0x0002811C CB_COLOR7_MASK
-0x00028080 CB_COLOR0_VIEW
-0x00028084 CB_COLOR1_VIEW
-0x00028088 CB_COLOR2_VIEW
-0x0002808C CB_COLOR3_VIEW
-0x00028090 CB_COLOR4_VIEW
-0x00028094 CB_COLOR5_VIEW
-0x00028098 CB_COLOR6_VIEW
-0x0002809C CB_COLOR7_VIEW
 0x00028808 CB_COLOR_CONTROL
 0x0002842C CB_FOG_BLUE
 0x00028428 CB_FOG_GREEN
diff --git a/drivers/gpu/drm/radeon/rs400.c b/drivers/gpu/drm/radeon/rs400.c
index b0ce84a20a68..4cf381b3a6d8 100644
--- a/drivers/gpu/drm/radeon/rs400.c
+++ b/drivers/gpu/drm/radeon/rs400.c
@@ -430,7 +430,7 @@ static int rs400_startup(struct radeon_device *rdev)
 	if (r)
 		return r;
 
-	r = r100_ib_test(rdev);
+	r = radeon_ib_test(rdev, RADEON_RING_TYPE_GFX_INDEX, &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]);
 	if (r) {
 		dev_err(rdev->dev, "failed testing IB (%d).\n", r);
 		rdev->accel_working = false;
@@ -442,6 +442,8 @@ static int rs400_startup(struct radeon_device *rdev)
 
 int rs400_resume(struct radeon_device *rdev)
 {
+	int r;
+
 	/* Make sur GART are not working */
 	rs400_gart_disable(rdev);
 	/* Resume clock before doing reset */
@@ -462,7 +464,11 @@ int rs400_resume(struct radeon_device *rdev)
 	radeon_surface_init(rdev);
 
 	rdev->accel_working = true;
-	return rs400_startup(rdev);
+	r = rs400_startup(rdev);
+	if (r) {
+		rdev->accel_working = false;
+	}
+	return r;
 }
 
 int rs400_suspend(struct radeon_device *rdev)
diff --git a/drivers/gpu/drm/radeon/rs600.c b/drivers/gpu/drm/radeon/rs600.c
index ec46eb45e34c..d25cf869d08d 100644
--- a/drivers/gpu/drm/radeon/rs600.c
+++ b/drivers/gpu/drm/radeon/rs600.c
@@ -46,6 +46,25 @@
 void rs600_gpu_init(struct radeon_device *rdev);
 int rs600_mc_wait_for_idle(struct radeon_device *rdev);
 
+void avivo_wait_for_vblank(struct radeon_device *rdev, int crtc)
+{
+	struct radeon_crtc *radeon_crtc = rdev->mode_info.crtcs[crtc];
+	int i;
+
+	if (RREG32(AVIVO_D1CRTC_CONTROL + radeon_crtc->crtc_offset) & AVIVO_CRTC_EN) {
+		for (i = 0; i < rdev->usec_timeout; i++) {
+			if (!(RREG32(AVIVO_D1CRTC_STATUS + radeon_crtc->crtc_offset) & AVIVO_D1CRTC_V_BLANK))
+				break;
+			udelay(1);
+		}
+		for (i = 0; i < rdev->usec_timeout; i++) {
+			if (RREG32(AVIVO_D1CRTC_STATUS + radeon_crtc->crtc_offset) & AVIVO_D1CRTC_V_BLANK)
+				break;
+			udelay(1);
+		}
+	}
+}
+
 void rs600_pre_page_flip(struct radeon_device *rdev, int crtc)
 {
 	/* enable the pflip int */
@@ -175,7 +194,7 @@ void rs600_pm_misc(struct radeon_device *rdev)
 	/* set pcie lanes */
 	if ((rdev->flags & RADEON_IS_PCIE) &&
 	    !(rdev->flags & RADEON_IS_IGP) &&
-	    rdev->asic->set_pcie_lanes &&
+	    rdev->asic->pm.set_pcie_lanes &&
 	    (ps->pcie_lanes !=
 	     rdev->pm.power_state[rdev->pm.current_power_state_index].pcie_lanes)) {
 		radeon_set_pcie_lanes(rdev,
@@ -684,9 +703,7 @@ int rs600_irq_process(struct radeon_device *rdev)
 			WREG32(RADEON_BUS_CNTL, msi_rearm | RS600_MSI_REARM);
 			break;
 		default:
-			msi_rearm = RREG32(RADEON_MSI_REARM_EN) & ~RV370_MSI_REARM_EN;
-			WREG32(RADEON_MSI_REARM_EN, msi_rearm);
-			WREG32(RADEON_MSI_REARM_EN, msi_rearm | RV370_MSI_REARM_EN);
+			WREG32(RADEON_MSI_REARM_EN, RV370_MSI_REARM_EN);
 			break;
 		}
 	}
@@ -866,7 +883,7 @@ static int rs600_startup(struct radeon_device *rdev)
 	if (r)
 		return r;
 
-	r = r100_ib_test(rdev);
+	r = radeon_ib_test(rdev, RADEON_RING_TYPE_GFX_INDEX, &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]);
 	if (r) {
 		dev_err(rdev->dev, "failed testing IB (%d).\n", r);
 		rdev->accel_working = false;
@@ -878,6 +895,8 @@ static int rs600_startup(struct radeon_device *rdev)
 
 int rs600_resume(struct radeon_device *rdev)
 {
+	int r;
+
 	/* Make sur GART are not working */
 	rs600_gart_disable(rdev);
 	/* Resume clock before doing reset */
@@ -896,7 +915,11 @@ int rs600_resume(struct radeon_device *rdev)
 	radeon_surface_init(rdev);
 
 	rdev->accel_working = true;
-	return rs600_startup(rdev);
+	r = rs600_startup(rdev);
+	if (r) {
+		rdev->accel_working = false;
+	}
+	return r;
 }
 
 int rs600_suspend(struct radeon_device *rdev)
diff --git a/drivers/gpu/drm/radeon/rs690.c b/drivers/gpu/drm/radeon/rs690.c
index 4f24a0fa8c82..f2c3b9d75f18 100644
--- a/drivers/gpu/drm/radeon/rs690.c
+++ b/drivers/gpu/drm/radeon/rs690.c
@@ -31,7 +31,7 @@
 #include "atom.h"
 #include "rs690d.h"
 
-static int rs690_mc_wait_for_idle(struct radeon_device *rdev)
+int rs690_mc_wait_for_idle(struct radeon_device *rdev)
 {
 	unsigned i;
 	uint32_t tmp;
@@ -647,7 +647,7 @@ static int rs690_startup(struct radeon_device *rdev)
 	if (r)
 		return r;
 
-	r = r100_ib_test(rdev);
+	r = radeon_ib_test(rdev, RADEON_RING_TYPE_GFX_INDEX, &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]);
 	if (r) {
 		dev_err(rdev->dev, "failed testing IB (%d).\n", r);
 		rdev->accel_working = false;
@@ -659,6 +659,8 @@ static int rs690_startup(struct radeon_device *rdev)
 
 int rs690_resume(struct radeon_device *rdev)
 {
+	int r;
+
 	/* Make sur GART are not working */
 	rs400_gart_disable(rdev);
 	/* Resume clock before doing reset */
@@ -677,7 +679,11 @@ int rs690_resume(struct radeon_device *rdev)
 	radeon_surface_init(rdev);
 
 	rdev->accel_working = true;
-	return rs690_startup(rdev);
+	r = rs690_startup(rdev);
+	if (r) {
+		rdev->accel_working = false;
+	}
+	return r;
 }
 
 int rs690_suspend(struct radeon_device *rdev)
diff --git a/drivers/gpu/drm/radeon/rv515.c b/drivers/gpu/drm/radeon/rv515.c
index 8787b18d6372..d8d78fe17946 100644
--- a/drivers/gpu/drm/radeon/rv515.c
+++ b/drivers/gpu/drm/radeon/rv515.c
@@ -53,9 +53,8 @@ void rv515_debugfs(struct radeon_device *rdev)
 	}
 }
 
-void rv515_ring_start(struct radeon_device *rdev)
+void rv515_ring_start(struct radeon_device *rdev, struct radeon_ring *ring)
 {
-	struct radeon_ring *ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX];
 	int r;
 
 	r = radeon_ring_lock(rdev, ring, 64);
@@ -413,7 +412,7 @@ static int rv515_startup(struct radeon_device *rdev)
 	if (r)
 		return r;
 
-	r = r100_ib_test(rdev);
+	r = radeon_ib_test(rdev, RADEON_RING_TYPE_GFX_INDEX, &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]);
 	if (r) {
 		dev_err(rdev->dev, "failed testing IB (%d).\n", r);
 		rdev->accel_working = false;
@@ -424,6 +423,8 @@ static int rv515_startup(struct radeon_device *rdev)
 
 int rv515_resume(struct radeon_device *rdev)
 {
+	int r;
+
 	/* Make sur GART are not working */
 	if (rdev->flags & RADEON_IS_PCIE)
 		rv370_pcie_gart_disable(rdev);
@@ -443,7 +444,11 @@ int rv515_resume(struct radeon_device *rdev)
 	radeon_surface_init(rdev);
 
 	rdev->accel_working = true;
-	return rv515_startup(rdev);
+	r =  rv515_startup(rdev);
+	if (r) {
+		rdev->accel_working = false;
+	}
+	return r;
 }
 
 int rv515_suspend(struct radeon_device *rdev)
diff --git a/drivers/gpu/drm/radeon/rv770.c b/drivers/gpu/drm/radeon/rv770.c
index a1668b659ddd..c62ae4be3845 100644
--- a/drivers/gpu/drm/radeon/rv770.c
+++ b/drivers/gpu/drm/radeon/rv770.c
@@ -1074,7 +1074,7 @@ static int rv770_startup(struct radeon_device *rdev)
 	r = r600_blit_init(rdev);
 	if (r) {
 		r600_blit_fini(rdev);
-		rdev->asic->copy = NULL;
+		rdev->asic->copy.copy = NULL;
 		dev_warn(rdev->dev, "failed blitter (%d) falling back to memcpy\n", r);
 	}
 
@@ -1114,7 +1114,7 @@ static int rv770_startup(struct radeon_device *rdev)
 	if (r)
 		return r;
 
-	r = r600_ib_test(rdev, RADEON_RING_TYPE_GFX_INDEX);
+	r = radeon_ib_test(rdev, RADEON_RING_TYPE_GFX_INDEX, &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]);
 	if (r) {
 		dev_err(rdev->dev, "IB test failed (%d).\n", r);
 		rdev->accel_working = false;
@@ -1139,6 +1139,7 @@ int rv770_resume(struct radeon_device *rdev)
 	r = rv770_startup(rdev);
 	if (r) {
 		DRM_ERROR("r600 startup failed on resume\n");
+		rdev->accel_working = false;
 		return r;
 	}
 
diff --git a/drivers/gpu/drm/radeon/si.c b/drivers/gpu/drm/radeon/si.c
new file mode 100644
index 000000000000..ac7a199ffece
--- /dev/null
+++ b/drivers/gpu/drm/radeon/si.c
@@ -0,0 +1,4128 @@
+/*
+ * Copyright 2011 Advanced Micro Devices, 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: Alex Deucher
+ */
+#include <linux/firmware.h>
+#include <linux/platform_device.h>
+#include <linux/slab.h>
+#include <linux/module.h>
+#include "drmP.h"
+#include "radeon.h"
+#include "radeon_asic.h"
+#include "radeon_drm.h"
+#include "sid.h"
+#include "atom.h"
+#include "si_blit_shaders.h"
+
+#define SI_PFP_UCODE_SIZE 2144
+#define SI_PM4_UCODE_SIZE 2144
+#define SI_CE_UCODE_SIZE 2144
+#define SI_RLC_UCODE_SIZE 2048
+#define SI_MC_UCODE_SIZE 7769
+
+MODULE_FIRMWARE("radeon/TAHITI_pfp.bin");
+MODULE_FIRMWARE("radeon/TAHITI_me.bin");
+MODULE_FIRMWARE("radeon/TAHITI_ce.bin");
+MODULE_FIRMWARE("radeon/TAHITI_mc.bin");
+MODULE_FIRMWARE("radeon/TAHITI_rlc.bin");
+MODULE_FIRMWARE("radeon/PITCAIRN_pfp.bin");
+MODULE_FIRMWARE("radeon/PITCAIRN_me.bin");
+MODULE_FIRMWARE("radeon/PITCAIRN_ce.bin");
+MODULE_FIRMWARE("radeon/PITCAIRN_mc.bin");
+MODULE_FIRMWARE("radeon/PITCAIRN_rlc.bin");
+MODULE_FIRMWARE("radeon/VERDE_pfp.bin");
+MODULE_FIRMWARE("radeon/VERDE_me.bin");
+MODULE_FIRMWARE("radeon/VERDE_ce.bin");
+MODULE_FIRMWARE("radeon/VERDE_mc.bin");
+MODULE_FIRMWARE("radeon/VERDE_rlc.bin");
+
+extern int r600_ih_ring_alloc(struct radeon_device *rdev);
+extern void r600_ih_ring_fini(struct radeon_device *rdev);
+extern void evergreen_fix_pci_max_read_req_size(struct radeon_device *rdev);
+extern void evergreen_mc_stop(struct radeon_device *rdev, struct evergreen_mc_save *save);
+extern void evergreen_mc_resume(struct radeon_device *rdev, struct evergreen_mc_save *save);
+extern u32 evergreen_get_number_of_dram_channels(struct radeon_device *rdev);
+
+/* get temperature in millidegrees */
+int si_get_temp(struct radeon_device *rdev)
+{
+	u32 temp;
+	int actual_temp = 0;
+
+	temp = (RREG32(CG_MULT_THERMAL_STATUS) & CTF_TEMP_MASK) >>
+		CTF_TEMP_SHIFT;
+
+	if (temp & 0x200)
+		actual_temp = 255;
+	else
+		actual_temp = temp & 0x1ff;
+
+	actual_temp = (actual_temp * 1000);
+
+	return actual_temp;
+}
+
+#define TAHITI_IO_MC_REGS_SIZE 36
+
+static const u32 tahiti_io_mc_regs[TAHITI_IO_MC_REGS_SIZE][2] = {
+	{0x0000006f, 0x03044000},
+	{0x00000070, 0x0480c018},
+	{0x00000071, 0x00000040},
+	{0x00000072, 0x01000000},
+	{0x00000074, 0x000000ff},
+	{0x00000075, 0x00143400},
+	{0x00000076, 0x08ec0800},
+	{0x00000077, 0x040000cc},
+	{0x00000079, 0x00000000},
+	{0x0000007a, 0x21000409},
+	{0x0000007c, 0x00000000},
+	{0x0000007d, 0xe8000000},
+	{0x0000007e, 0x044408a8},
+	{0x0000007f, 0x00000003},
+	{0x00000080, 0x00000000},
+	{0x00000081, 0x01000000},
+	{0x00000082, 0x02000000},
+	{0x00000083, 0x00000000},
+	{0x00000084, 0xe3f3e4f4},
+	{0x00000085, 0x00052024},
+	{0x00000087, 0x00000000},
+	{0x00000088, 0x66036603},
+	{0x00000089, 0x01000000},
+	{0x0000008b, 0x1c0a0000},
+	{0x0000008c, 0xff010000},
+	{0x0000008e, 0xffffefff},
+	{0x0000008f, 0xfff3efff},
+	{0x00000090, 0xfff3efbf},
+	{0x00000094, 0x00101101},
+	{0x00000095, 0x00000fff},
+	{0x00000096, 0x00116fff},
+	{0x00000097, 0x60010000},
+	{0x00000098, 0x10010000},
+	{0x00000099, 0x00006000},
+	{0x0000009a, 0x00001000},
+	{0x0000009f, 0x00a77400}
+};
+
+static const u32 pitcairn_io_mc_regs[TAHITI_IO_MC_REGS_SIZE][2] = {
+	{0x0000006f, 0x03044000},
+	{0x00000070, 0x0480c018},
+	{0x00000071, 0x00000040},
+	{0x00000072, 0x01000000},
+	{0x00000074, 0x000000ff},
+	{0x00000075, 0x00143400},
+	{0x00000076, 0x08ec0800},
+	{0x00000077, 0x040000cc},
+	{0x00000079, 0x00000000},
+	{0x0000007a, 0x21000409},
+	{0x0000007c, 0x00000000},
+	{0x0000007d, 0xe8000000},
+	{0x0000007e, 0x044408a8},
+	{0x0000007f, 0x00000003},
+	{0x00000080, 0x00000000},
+	{0x00000081, 0x01000000},
+	{0x00000082, 0x02000000},
+	{0x00000083, 0x00000000},
+	{0x00000084, 0xe3f3e4f4},
+	{0x00000085, 0x00052024},
+	{0x00000087, 0x00000000},
+	{0x00000088, 0x66036603},
+	{0x00000089, 0x01000000},
+	{0x0000008b, 0x1c0a0000},
+	{0x0000008c, 0xff010000},
+	{0x0000008e, 0xffffefff},
+	{0x0000008f, 0xfff3efff},
+	{0x00000090, 0xfff3efbf},
+	{0x00000094, 0x00101101},
+	{0x00000095, 0x00000fff},
+	{0x00000096, 0x00116fff},
+	{0x00000097, 0x60010000},
+	{0x00000098, 0x10010000},
+	{0x00000099, 0x00006000},
+	{0x0000009a, 0x00001000},
+	{0x0000009f, 0x00a47400}
+};
+
+static const u32 verde_io_mc_regs[TAHITI_IO_MC_REGS_SIZE][2] = {
+	{0x0000006f, 0x03044000},
+	{0x00000070, 0x0480c018},
+	{0x00000071, 0x00000040},
+	{0x00000072, 0x01000000},
+	{0x00000074, 0x000000ff},
+	{0x00000075, 0x00143400},
+	{0x00000076, 0x08ec0800},
+	{0x00000077, 0x040000cc},
+	{0x00000079, 0x00000000},
+	{0x0000007a, 0x21000409},
+	{0x0000007c, 0x00000000},
+	{0x0000007d, 0xe8000000},
+	{0x0000007e, 0x044408a8},
+	{0x0000007f, 0x00000003},
+	{0x00000080, 0x00000000},
+	{0x00000081, 0x01000000},
+	{0x00000082, 0x02000000},
+	{0x00000083, 0x00000000},
+	{0x00000084, 0xe3f3e4f4},
+	{0x00000085, 0x00052024},
+	{0x00000087, 0x00000000},
+	{0x00000088, 0x66036603},
+	{0x00000089, 0x01000000},
+	{0x0000008b, 0x1c0a0000},
+	{0x0000008c, 0xff010000},
+	{0x0000008e, 0xffffefff},
+	{0x0000008f, 0xfff3efff},
+	{0x00000090, 0xfff3efbf},
+	{0x00000094, 0x00101101},
+	{0x00000095, 0x00000fff},
+	{0x00000096, 0x00116fff},
+	{0x00000097, 0x60010000},
+	{0x00000098, 0x10010000},
+	{0x00000099, 0x00006000},
+	{0x0000009a, 0x00001000},
+	{0x0000009f, 0x00a37400}
+};
+
+/* ucode loading */
+static int si_mc_load_microcode(struct radeon_device *rdev)
+{
+	const __be32 *fw_data;
+	u32 running, blackout = 0;
+	u32 *io_mc_regs;
+	int i, ucode_size, regs_size;
+
+	if (!rdev->mc_fw)
+		return -EINVAL;
+
+	switch (rdev->family) {
+	case CHIP_TAHITI:
+		io_mc_regs = (u32 *)&tahiti_io_mc_regs;
+		ucode_size = SI_MC_UCODE_SIZE;
+		regs_size = TAHITI_IO_MC_REGS_SIZE;
+		break;
+	case CHIP_PITCAIRN:
+		io_mc_regs = (u32 *)&pitcairn_io_mc_regs;
+		ucode_size = SI_MC_UCODE_SIZE;
+		regs_size = TAHITI_IO_MC_REGS_SIZE;
+		break;
+	case CHIP_VERDE:
+	default:
+		io_mc_regs = (u32 *)&verde_io_mc_regs;
+		ucode_size = SI_MC_UCODE_SIZE;
+		regs_size = TAHITI_IO_MC_REGS_SIZE;
+		break;
+	}
+
+	running = RREG32(MC_SEQ_SUP_CNTL) & RUN_MASK;
+
+	if (running == 0) {
+		if (running) {
+			blackout = RREG32(MC_SHARED_BLACKOUT_CNTL);
+			WREG32(MC_SHARED_BLACKOUT_CNTL, blackout | 1);
+		}
+
+		/* reset the engine and set to writable */
+		WREG32(MC_SEQ_SUP_CNTL, 0x00000008);
+		WREG32(MC_SEQ_SUP_CNTL, 0x00000010);
+
+		/* load mc io regs */
+		for (i = 0; i < regs_size; i++) {
+			WREG32(MC_SEQ_IO_DEBUG_INDEX, io_mc_regs[(i << 1)]);
+			WREG32(MC_SEQ_IO_DEBUG_DATA, io_mc_regs[(i << 1) + 1]);
+		}
+		/* load the MC ucode */
+		fw_data = (const __be32 *)rdev->mc_fw->data;
+		for (i = 0; i < ucode_size; i++)
+			WREG32(MC_SEQ_SUP_PGM, be32_to_cpup(fw_data++));
+
+		/* put the engine back into the active state */
+		WREG32(MC_SEQ_SUP_CNTL, 0x00000008);
+		WREG32(MC_SEQ_SUP_CNTL, 0x00000004);
+		WREG32(MC_SEQ_SUP_CNTL, 0x00000001);
+
+		/* wait for training to complete */
+		for (i = 0; i < rdev->usec_timeout; i++) {
+			if (RREG32(MC_SEQ_TRAIN_WAKEUP_CNTL) & TRAIN_DONE_D0)
+				break;
+			udelay(1);
+		}
+		for (i = 0; i < rdev->usec_timeout; i++) {
+			if (RREG32(MC_SEQ_TRAIN_WAKEUP_CNTL) & TRAIN_DONE_D1)
+				break;
+			udelay(1);
+		}
+
+		if (running)
+			WREG32(MC_SHARED_BLACKOUT_CNTL, blackout);
+	}
+
+	return 0;
+}
+
+static int si_init_microcode(struct radeon_device *rdev)
+{
+	struct platform_device *pdev;
+	const char *chip_name;
+	const char *rlc_chip_name;
+	size_t pfp_req_size, me_req_size, ce_req_size, rlc_req_size, mc_req_size;
+	char fw_name[30];
+	int err;
+
+	DRM_DEBUG("\n");
+
+	pdev = platform_device_register_simple("radeon_cp", 0, NULL, 0);
+	err = IS_ERR(pdev);
+	if (err) {
+		printk(KERN_ERR "radeon_cp: Failed to register firmware\n");
+		return -EINVAL;
+	}
+
+	switch (rdev->family) {
+	case CHIP_TAHITI:
+		chip_name = "TAHITI";
+		rlc_chip_name = "TAHITI";
+		pfp_req_size = SI_PFP_UCODE_SIZE * 4;
+		me_req_size = SI_PM4_UCODE_SIZE * 4;
+		ce_req_size = SI_CE_UCODE_SIZE * 4;
+		rlc_req_size = SI_RLC_UCODE_SIZE * 4;
+		mc_req_size = SI_MC_UCODE_SIZE * 4;
+		break;
+	case CHIP_PITCAIRN:
+		chip_name = "PITCAIRN";
+		rlc_chip_name = "PITCAIRN";
+		pfp_req_size = SI_PFP_UCODE_SIZE * 4;
+		me_req_size = SI_PM4_UCODE_SIZE * 4;
+		ce_req_size = SI_CE_UCODE_SIZE * 4;
+		rlc_req_size = SI_RLC_UCODE_SIZE * 4;
+		mc_req_size = SI_MC_UCODE_SIZE * 4;
+		break;
+	case CHIP_VERDE:
+		chip_name = "VERDE";
+		rlc_chip_name = "VERDE";
+		pfp_req_size = SI_PFP_UCODE_SIZE * 4;
+		me_req_size = SI_PM4_UCODE_SIZE * 4;
+		ce_req_size = SI_CE_UCODE_SIZE * 4;
+		rlc_req_size = SI_RLC_UCODE_SIZE * 4;
+		mc_req_size = SI_MC_UCODE_SIZE * 4;
+		break;
+	default: BUG();
+	}
+
+	DRM_INFO("Loading %s Microcode\n", chip_name);
+
+	snprintf(fw_name, sizeof(fw_name), "radeon/%s_pfp.bin", chip_name);
+	err = request_firmware(&rdev->pfp_fw, fw_name, &pdev->dev);
+	if (err)
+		goto out;
+	if (rdev->pfp_fw->size != pfp_req_size) {
+		printk(KERN_ERR
+		       "si_cp: Bogus length %zu in firmware \"%s\"\n",
+		       rdev->pfp_fw->size, fw_name);
+		err = -EINVAL;
+		goto out;
+	}
+
+	snprintf(fw_name, sizeof(fw_name), "radeon/%s_me.bin", chip_name);
+	err = request_firmware(&rdev->me_fw, fw_name, &pdev->dev);
+	if (err)
+		goto out;
+	if (rdev->me_fw->size != me_req_size) {
+		printk(KERN_ERR
+		       "si_cp: Bogus length %zu in firmware \"%s\"\n",
+		       rdev->me_fw->size, fw_name);
+		err = -EINVAL;
+	}
+
+	snprintf(fw_name, sizeof(fw_name), "radeon/%s_ce.bin", chip_name);
+	err = request_firmware(&rdev->ce_fw, fw_name, &pdev->dev);
+	if (err)
+		goto out;
+	if (rdev->ce_fw->size != ce_req_size) {
+		printk(KERN_ERR
+		       "si_cp: Bogus length %zu in firmware \"%s\"\n",
+		       rdev->ce_fw->size, fw_name);
+		err = -EINVAL;
+	}
+
+	snprintf(fw_name, sizeof(fw_name), "radeon/%s_rlc.bin", rlc_chip_name);
+	err = request_firmware(&rdev->rlc_fw, fw_name, &pdev->dev);
+	if (err)
+		goto out;
+	if (rdev->rlc_fw->size != rlc_req_size) {
+		printk(KERN_ERR
+		       "si_rlc: Bogus length %zu in firmware \"%s\"\n",
+		       rdev->rlc_fw->size, fw_name);
+		err = -EINVAL;
+	}
+
+	snprintf(fw_name, sizeof(fw_name), "radeon/%s_mc.bin", chip_name);
+	err = request_firmware(&rdev->mc_fw, fw_name, &pdev->dev);
+	if (err)
+		goto out;
+	if (rdev->mc_fw->size != mc_req_size) {
+		printk(KERN_ERR
+		       "si_mc: Bogus length %zu in firmware \"%s\"\n",
+		       rdev->mc_fw->size, fw_name);
+		err = -EINVAL;
+	}
+
+out:
+	platform_device_unregister(pdev);
+
+	if (err) {
+		if (err != -EINVAL)
+			printk(KERN_ERR
+			       "si_cp: Failed to load firmware \"%s\"\n",
+			       fw_name);
+		release_firmware(rdev->pfp_fw);
+		rdev->pfp_fw = NULL;
+		release_firmware(rdev->me_fw);
+		rdev->me_fw = NULL;
+		release_firmware(rdev->ce_fw);
+		rdev->ce_fw = NULL;
+		release_firmware(rdev->rlc_fw);
+		rdev->rlc_fw = NULL;
+		release_firmware(rdev->mc_fw);
+		rdev->mc_fw = NULL;
+	}
+	return err;
+}
+
+/* watermark setup */
+static u32 dce6_line_buffer_adjust(struct radeon_device *rdev,
+				   struct radeon_crtc *radeon_crtc,
+				   struct drm_display_mode *mode,
+				   struct drm_display_mode *other_mode)
+{
+	u32 tmp;
+	/*
+	 * Line Buffer Setup
+	 * There are 3 line buffers, each one shared by 2 display controllers.
+	 * DC_LB_MEMORY_SPLIT controls how that line buffer is shared between
+	 * the display controllers.  The paritioning is done via one of four
+	 * preset allocations specified in bits 21:20:
+	 *  0 - half lb
+	 *  2 - whole lb, other crtc must be disabled
+	 */
+	/* this can get tricky if we have two large displays on a paired group
+	 * of crtcs.  Ideally for multiple large displays we'd assign them to
+	 * non-linked crtcs for maximum line buffer allocation.
+	 */
+	if (radeon_crtc->base.enabled && mode) {
+		if (other_mode)
+			tmp = 0; /* 1/2 */
+		else
+			tmp = 2; /* whole */
+	} else
+		tmp = 0;
+
+	WREG32(DC_LB_MEMORY_SPLIT + radeon_crtc->crtc_offset,
+	       DC_LB_MEMORY_CONFIG(tmp));
+
+	if (radeon_crtc->base.enabled && mode) {
+		switch (tmp) {
+		case 0:
+		default:
+			return 4096 * 2;
+		case 2:
+			return 8192 * 2;
+		}
+	}
+
+	/* controller not enabled, so no lb used */
+	return 0;
+}
+
+static u32 si_get_number_of_dram_channels(struct radeon_device *rdev)
+{
+	u32 tmp = RREG32(MC_SHARED_CHMAP);
+
+	switch ((tmp & NOOFCHAN_MASK) >> NOOFCHAN_SHIFT) {
+	case 0:
+	default:
+		return 1;
+	case 1:
+		return 2;
+	case 2:
+		return 4;
+	case 3:
+		return 8;
+	case 4:
+		return 3;
+	case 5:
+		return 6;
+	case 6:
+		return 10;
+	case 7:
+		return 12;
+	case 8:
+		return 16;
+	}
+}
+
+struct dce6_wm_params {
+	u32 dram_channels; /* number of dram channels */
+	u32 yclk;          /* bandwidth per dram data pin in kHz */
+	u32 sclk;          /* engine clock in kHz */
+	u32 disp_clk;      /* display clock in kHz */
+	u32 src_width;     /* viewport width */
+	u32 active_time;   /* active display time in ns */
+	u32 blank_time;    /* blank time in ns */
+	bool interlaced;    /* mode is interlaced */
+	fixed20_12 vsc;    /* vertical scale ratio */
+	u32 num_heads;     /* number of active crtcs */
+	u32 bytes_per_pixel; /* bytes per pixel display + overlay */
+	u32 lb_size;       /* line buffer allocated to pipe */
+	u32 vtaps;         /* vertical scaler taps */
+};
+
+static u32 dce6_dram_bandwidth(struct dce6_wm_params *wm)
+{
+	/* Calculate raw DRAM Bandwidth */
+	fixed20_12 dram_efficiency; /* 0.7 */
+	fixed20_12 yclk, dram_channels, bandwidth;
+	fixed20_12 a;
+
+	a.full = dfixed_const(1000);
+	yclk.full = dfixed_const(wm->yclk);
+	yclk.full = dfixed_div(yclk, a);
+	dram_channels.full = dfixed_const(wm->dram_channels * 4);
+	a.full = dfixed_const(10);
+	dram_efficiency.full = dfixed_const(7);
+	dram_efficiency.full = dfixed_div(dram_efficiency, a);
+	bandwidth.full = dfixed_mul(dram_channels, yclk);
+	bandwidth.full = dfixed_mul(bandwidth, dram_efficiency);
+
+	return dfixed_trunc(bandwidth);
+}
+
+static u32 dce6_dram_bandwidth_for_display(struct dce6_wm_params *wm)
+{
+	/* Calculate DRAM Bandwidth and the part allocated to display. */
+	fixed20_12 disp_dram_allocation; /* 0.3 to 0.7 */
+	fixed20_12 yclk, dram_channels, bandwidth;
+	fixed20_12 a;
+
+	a.full = dfixed_const(1000);
+	yclk.full = dfixed_const(wm->yclk);
+	yclk.full = dfixed_div(yclk, a);
+	dram_channels.full = dfixed_const(wm->dram_channels * 4);
+	a.full = dfixed_const(10);
+	disp_dram_allocation.full = dfixed_const(3); /* XXX worse case value 0.3 */
+	disp_dram_allocation.full = dfixed_div(disp_dram_allocation, a);
+	bandwidth.full = dfixed_mul(dram_channels, yclk);
+	bandwidth.full = dfixed_mul(bandwidth, disp_dram_allocation);
+
+	return dfixed_trunc(bandwidth);
+}
+
+static u32 dce6_data_return_bandwidth(struct dce6_wm_params *wm)
+{
+	/* Calculate the display Data return Bandwidth */
+	fixed20_12 return_efficiency; /* 0.8 */
+	fixed20_12 sclk, bandwidth;
+	fixed20_12 a;
+
+	a.full = dfixed_const(1000);
+	sclk.full = dfixed_const(wm->sclk);
+	sclk.full = dfixed_div(sclk, a);
+	a.full = dfixed_const(10);
+	return_efficiency.full = dfixed_const(8);
+	return_efficiency.full = dfixed_div(return_efficiency, a);
+	a.full = dfixed_const(32);
+	bandwidth.full = dfixed_mul(a, sclk);
+	bandwidth.full = dfixed_mul(bandwidth, return_efficiency);
+
+	return dfixed_trunc(bandwidth);
+}
+
+static u32 dce6_get_dmif_bytes_per_request(struct dce6_wm_params *wm)
+{
+	return 32;
+}
+
+static u32 dce6_dmif_request_bandwidth(struct dce6_wm_params *wm)
+{
+	/* Calculate the DMIF Request Bandwidth */
+	fixed20_12 disp_clk_request_efficiency; /* 0.8 */
+	fixed20_12 disp_clk, sclk, bandwidth;
+	fixed20_12 a, b1, b2;
+	u32 min_bandwidth;
+
+	a.full = dfixed_const(1000);
+	disp_clk.full = dfixed_const(wm->disp_clk);
+	disp_clk.full = dfixed_div(disp_clk, a);
+	a.full = dfixed_const(dce6_get_dmif_bytes_per_request(wm) / 2);
+	b1.full = dfixed_mul(a, disp_clk);
+
+	a.full = dfixed_const(1000);
+	sclk.full = dfixed_const(wm->sclk);
+	sclk.full = dfixed_div(sclk, a);
+	a.full = dfixed_const(dce6_get_dmif_bytes_per_request(wm));
+	b2.full = dfixed_mul(a, sclk);
+
+	a.full = dfixed_const(10);
+	disp_clk_request_efficiency.full = dfixed_const(8);
+	disp_clk_request_efficiency.full = dfixed_div(disp_clk_request_efficiency, a);
+
+	min_bandwidth = min(dfixed_trunc(b1), dfixed_trunc(b2));
+
+	a.full = dfixed_const(min_bandwidth);
+	bandwidth.full = dfixed_mul(a, disp_clk_request_efficiency);
+
+	return dfixed_trunc(bandwidth);
+}
+
+static u32 dce6_available_bandwidth(struct dce6_wm_params *wm)
+{
+	/* Calculate the Available bandwidth. Display can use this temporarily but not in average. */
+	u32 dram_bandwidth = dce6_dram_bandwidth(wm);
+	u32 data_return_bandwidth = dce6_data_return_bandwidth(wm);
+	u32 dmif_req_bandwidth = dce6_dmif_request_bandwidth(wm);
+
+	return min(dram_bandwidth, min(data_return_bandwidth, dmif_req_bandwidth));
+}
+
+static u32 dce6_average_bandwidth(struct dce6_wm_params *wm)
+{
+	/* Calculate the display mode Average Bandwidth
+	 * DisplayMode should contain the source and destination dimensions,
+	 * timing, etc.
+	 */
+	fixed20_12 bpp;
+	fixed20_12 line_time;
+	fixed20_12 src_width;
+	fixed20_12 bandwidth;
+	fixed20_12 a;
+
+	a.full = dfixed_const(1000);
+	line_time.full = dfixed_const(wm->active_time + wm->blank_time);
+	line_time.full = dfixed_div(line_time, a);
+	bpp.full = dfixed_const(wm->bytes_per_pixel);
+	src_width.full = dfixed_const(wm->src_width);
+	bandwidth.full = dfixed_mul(src_width, bpp);
+	bandwidth.full = dfixed_mul(bandwidth, wm->vsc);
+	bandwidth.full = dfixed_div(bandwidth, line_time);
+
+	return dfixed_trunc(bandwidth);
+}
+
+static u32 dce6_latency_watermark(struct dce6_wm_params *wm)
+{
+	/* First calcualte the latency in ns */
+	u32 mc_latency = 2000; /* 2000 ns. */
+	u32 available_bandwidth = dce6_available_bandwidth(wm);
+	u32 worst_chunk_return_time = (512 * 8 * 1000) / available_bandwidth;
+	u32 cursor_line_pair_return_time = (128 * 4 * 1000) / available_bandwidth;
+	u32 dc_latency = 40000000 / wm->disp_clk; /* dc pipe latency */
+	u32 other_heads_data_return_time = ((wm->num_heads + 1) * worst_chunk_return_time) +
+		(wm->num_heads * cursor_line_pair_return_time);
+	u32 latency = mc_latency + other_heads_data_return_time + dc_latency;
+	u32 max_src_lines_per_dst_line, lb_fill_bw, line_fill_time;
+	u32 tmp, dmif_size = 12288;
+	fixed20_12 a, b, c;
+
+	if (wm->num_heads == 0)
+		return 0;
+
+	a.full = dfixed_const(2);
+	b.full = dfixed_const(1);
+	if ((wm->vsc.full > a.full) ||
+	    ((wm->vsc.full > b.full) && (wm->vtaps >= 3)) ||
+	    (wm->vtaps >= 5) ||
+	    ((wm->vsc.full >= a.full) && wm->interlaced))
+		max_src_lines_per_dst_line = 4;
+	else
+		max_src_lines_per_dst_line = 2;
+
+	a.full = dfixed_const(available_bandwidth);
+	b.full = dfixed_const(wm->num_heads);
+	a.full = dfixed_div(a, b);
+
+	b.full = dfixed_const(mc_latency + 512);
+	c.full = dfixed_const(wm->disp_clk);
+	b.full = dfixed_div(b, c);
+
+	c.full = dfixed_const(dmif_size);
+	b.full = dfixed_div(c, b);
+
+	tmp = min(dfixed_trunc(a), dfixed_trunc(b));
+
+	b.full = dfixed_const(1000);
+	c.full = dfixed_const(wm->disp_clk);
+	b.full = dfixed_div(c, b);
+	c.full = dfixed_const(wm->bytes_per_pixel);
+	b.full = dfixed_mul(b, c);
+
+	lb_fill_bw = min(tmp, dfixed_trunc(b));
+
+	a.full = dfixed_const(max_src_lines_per_dst_line * wm->src_width * wm->bytes_per_pixel);
+	b.full = dfixed_const(1000);
+	c.full = dfixed_const(lb_fill_bw);
+	b.full = dfixed_div(c, b);
+	a.full = dfixed_div(a, b);
+	line_fill_time = dfixed_trunc(a);
+
+	if (line_fill_time < wm->active_time)
+		return latency;
+	else
+		return latency + (line_fill_time - wm->active_time);
+
+}
+
+static bool dce6_average_bandwidth_vs_dram_bandwidth_for_display(struct dce6_wm_params *wm)
+{
+	if (dce6_average_bandwidth(wm) <=
+	    (dce6_dram_bandwidth_for_display(wm) / wm->num_heads))
+		return true;
+	else
+		return false;
+};
+
+static bool dce6_average_bandwidth_vs_available_bandwidth(struct dce6_wm_params *wm)
+{
+	if (dce6_average_bandwidth(wm) <=
+	    (dce6_available_bandwidth(wm) / wm->num_heads))
+		return true;
+	else
+		return false;
+};
+
+static bool dce6_check_latency_hiding(struct dce6_wm_params *wm)
+{
+	u32 lb_partitions = wm->lb_size / wm->src_width;
+	u32 line_time = wm->active_time + wm->blank_time;
+	u32 latency_tolerant_lines;
+	u32 latency_hiding;
+	fixed20_12 a;
+
+	a.full = dfixed_const(1);
+	if (wm->vsc.full > a.full)
+		latency_tolerant_lines = 1;
+	else {
+		if (lb_partitions <= (wm->vtaps + 1))
+			latency_tolerant_lines = 1;
+		else
+			latency_tolerant_lines = 2;
+	}
+
+	latency_hiding = (latency_tolerant_lines * line_time + wm->blank_time);
+
+	if (dce6_latency_watermark(wm) <= latency_hiding)
+		return true;
+	else
+		return false;
+}
+
+static void dce6_program_watermarks(struct radeon_device *rdev,
+					 struct radeon_crtc *radeon_crtc,
+					 u32 lb_size, u32 num_heads)
+{
+	struct drm_display_mode *mode = &radeon_crtc->base.mode;
+	struct dce6_wm_params wm;
+	u32 pixel_period;
+	u32 line_time = 0;
+	u32 latency_watermark_a = 0, latency_watermark_b = 0;
+	u32 priority_a_mark = 0, priority_b_mark = 0;
+	u32 priority_a_cnt = PRIORITY_OFF;
+	u32 priority_b_cnt = PRIORITY_OFF;
+	u32 tmp, arb_control3;
+	fixed20_12 a, b, c;
+
+	if (radeon_crtc->base.enabled && num_heads && mode) {
+		pixel_period = 1000000 / (u32)mode->clock;
+		line_time = min((u32)mode->crtc_htotal * pixel_period, (u32)65535);
+		priority_a_cnt = 0;
+		priority_b_cnt = 0;
+
+		wm.yclk = rdev->pm.current_mclk * 10;
+		wm.sclk = rdev->pm.current_sclk * 10;
+		wm.disp_clk = mode->clock;
+		wm.src_width = mode->crtc_hdisplay;
+		wm.active_time = mode->crtc_hdisplay * pixel_period;
+		wm.blank_time = line_time - wm.active_time;
+		wm.interlaced = false;
+		if (mode->flags & DRM_MODE_FLAG_INTERLACE)
+			wm.interlaced = true;
+		wm.vsc = radeon_crtc->vsc;
+		wm.vtaps = 1;
+		if (radeon_crtc->rmx_type != RMX_OFF)
+			wm.vtaps = 2;
+		wm.bytes_per_pixel = 4; /* XXX: get this from fb config */
+		wm.lb_size = lb_size;
+		if (rdev->family == CHIP_ARUBA)
+			wm.dram_channels = evergreen_get_number_of_dram_channels(rdev);
+		else
+			wm.dram_channels = si_get_number_of_dram_channels(rdev);
+		wm.num_heads = num_heads;
+
+		/* set for high clocks */
+		latency_watermark_a = min(dce6_latency_watermark(&wm), (u32)65535);
+		/* set for low clocks */
+		/* wm.yclk = low clk; wm.sclk = low clk */
+		latency_watermark_b = min(dce6_latency_watermark(&wm), (u32)65535);
+
+		/* possibly force display priority to high */
+		/* should really do this at mode validation time... */
+		if (!dce6_average_bandwidth_vs_dram_bandwidth_for_display(&wm) ||
+		    !dce6_average_bandwidth_vs_available_bandwidth(&wm) ||
+		    !dce6_check_latency_hiding(&wm) ||
+		    (rdev->disp_priority == 2)) {
+			DRM_DEBUG_KMS("force priority to high\n");
+			priority_a_cnt |= PRIORITY_ALWAYS_ON;
+			priority_b_cnt |= PRIORITY_ALWAYS_ON;
+		}
+
+		a.full = dfixed_const(1000);
+		b.full = dfixed_const(mode->clock);
+		b.full = dfixed_div(b, a);
+		c.full = dfixed_const(latency_watermark_a);
+		c.full = dfixed_mul(c, b);
+		c.full = dfixed_mul(c, radeon_crtc->hsc);
+		c.full = dfixed_div(c, a);
+		a.full = dfixed_const(16);
+		c.full = dfixed_div(c, a);
+		priority_a_mark = dfixed_trunc(c);
+		priority_a_cnt |= priority_a_mark & PRIORITY_MARK_MASK;
+
+		a.full = dfixed_const(1000);
+		b.full = dfixed_const(mode->clock);
+		b.full = dfixed_div(b, a);
+		c.full = dfixed_const(latency_watermark_b);
+		c.full = dfixed_mul(c, b);
+		c.full = dfixed_mul(c, radeon_crtc->hsc);
+		c.full = dfixed_div(c, a);
+		a.full = dfixed_const(16);
+		c.full = dfixed_div(c, a);
+		priority_b_mark = dfixed_trunc(c);
+		priority_b_cnt |= priority_b_mark & PRIORITY_MARK_MASK;
+	}
+
+	/* select wm A */
+	arb_control3 = RREG32(DPG_PIPE_ARBITRATION_CONTROL3 + radeon_crtc->crtc_offset);
+	tmp = arb_control3;
+	tmp &= ~LATENCY_WATERMARK_MASK(3);
+	tmp |= LATENCY_WATERMARK_MASK(1);
+	WREG32(DPG_PIPE_ARBITRATION_CONTROL3 + radeon_crtc->crtc_offset, tmp);
+	WREG32(DPG_PIPE_LATENCY_CONTROL + radeon_crtc->crtc_offset,
+	       (LATENCY_LOW_WATERMARK(latency_watermark_a) |
+		LATENCY_HIGH_WATERMARK(line_time)));
+	/* select wm B */
+	tmp = RREG32(DPG_PIPE_ARBITRATION_CONTROL3 + radeon_crtc->crtc_offset);
+	tmp &= ~LATENCY_WATERMARK_MASK(3);
+	tmp |= LATENCY_WATERMARK_MASK(2);
+	WREG32(DPG_PIPE_ARBITRATION_CONTROL3 + radeon_crtc->crtc_offset, tmp);
+	WREG32(DPG_PIPE_LATENCY_CONTROL + radeon_crtc->crtc_offset,
+	       (LATENCY_LOW_WATERMARK(latency_watermark_b) |
+		LATENCY_HIGH_WATERMARK(line_time)));
+	/* restore original selection */
+	WREG32(DPG_PIPE_ARBITRATION_CONTROL3 + radeon_crtc->crtc_offset, arb_control3);
+
+	/* write the priority marks */
+	WREG32(PRIORITY_A_CNT + radeon_crtc->crtc_offset, priority_a_cnt);
+	WREG32(PRIORITY_B_CNT + radeon_crtc->crtc_offset, priority_b_cnt);
+
+}
+
+void dce6_bandwidth_update(struct radeon_device *rdev)
+{
+	struct drm_display_mode *mode0 = NULL;
+	struct drm_display_mode *mode1 = NULL;
+	u32 num_heads = 0, lb_size;
+	int i;
+
+	radeon_update_display_priority(rdev);
+
+	for (i = 0; i < rdev->num_crtc; i++) {
+		if (rdev->mode_info.crtcs[i]->base.enabled)
+			num_heads++;
+	}
+	for (i = 0; i < rdev->num_crtc; i += 2) {
+		mode0 = &rdev->mode_info.crtcs[i]->base.mode;
+		mode1 = &rdev->mode_info.crtcs[i+1]->base.mode;
+		lb_size = dce6_line_buffer_adjust(rdev, rdev->mode_info.crtcs[i], mode0, mode1);
+		dce6_program_watermarks(rdev, rdev->mode_info.crtcs[i], lb_size, num_heads);
+		lb_size = dce6_line_buffer_adjust(rdev, rdev->mode_info.crtcs[i+1], mode1, mode0);
+		dce6_program_watermarks(rdev, rdev->mode_info.crtcs[i+1], lb_size, num_heads);
+	}
+}
+
+/*
+ * Core functions
+ */
+static u32 si_get_tile_pipe_to_backend_map(struct radeon_device *rdev,
+					   u32 num_tile_pipes,
+					   u32 num_backends_per_asic,
+					   u32 *backend_disable_mask_per_asic,
+					   u32 num_shader_engines)
+{
+	u32 backend_map = 0;
+	u32 enabled_backends_mask = 0;
+	u32 enabled_backends_count = 0;
+	u32 num_backends_per_se;
+	u32 cur_pipe;
+	u32 swizzle_pipe[SI_MAX_PIPES];
+	u32 cur_backend = 0;
+	u32 i;
+	bool force_no_swizzle;
+
+	/* force legal values */
+	if (num_tile_pipes < 1)
+		num_tile_pipes = 1;
+	if (num_tile_pipes > rdev->config.si.max_tile_pipes)
+		num_tile_pipes = rdev->config.si.max_tile_pipes;
+	if (num_shader_engines < 1)
+		num_shader_engines = 1;
+	if (num_shader_engines > rdev->config.si.max_shader_engines)
+		num_shader_engines = rdev->config.si.max_shader_engines;
+	if (num_backends_per_asic < num_shader_engines)
+		num_backends_per_asic = num_shader_engines;
+	if (num_backends_per_asic > (rdev->config.si.max_backends_per_se * num_shader_engines))
+		num_backends_per_asic = rdev->config.si.max_backends_per_se * num_shader_engines;
+
+	/* make sure we have the same number of backends per se */
+	num_backends_per_asic = ALIGN(num_backends_per_asic, num_shader_engines);
+	/* set up the number of backends per se */
+	num_backends_per_se = num_backends_per_asic / num_shader_engines;
+	if (num_backends_per_se > rdev->config.si.max_backends_per_se) {
+		num_backends_per_se = rdev->config.si.max_backends_per_se;
+		num_backends_per_asic = num_backends_per_se * num_shader_engines;
+	}
+
+	/* create enable mask and count for enabled backends */
+	for (i = 0; i < SI_MAX_BACKENDS; ++i) {
+		if (((*backend_disable_mask_per_asic >> i) & 1) == 0) {
+			enabled_backends_mask |= (1 << i);
+			++enabled_backends_count;
+		}
+		if (enabled_backends_count == num_backends_per_asic)
+			break;
+	}
+
+	/* force the backends mask to match the current number of backends */
+	if (enabled_backends_count != num_backends_per_asic) {
+		u32 this_backend_enabled;
+		u32 shader_engine;
+		u32 backend_per_se;
+
+		enabled_backends_mask = 0;
+		enabled_backends_count = 0;
+		*backend_disable_mask_per_asic = SI_MAX_BACKENDS_MASK;
+		for (i = 0; i < SI_MAX_BACKENDS; ++i) {
+			/* calc the current se */
+			shader_engine = i / rdev->config.si.max_backends_per_se;
+			/* calc the backend per se */
+			backend_per_se = i % rdev->config.si.max_backends_per_se;
+			/* default to not enabled */
+			this_backend_enabled = 0;
+			if ((shader_engine < num_shader_engines) &&
+			    (backend_per_se < num_backends_per_se))
+				this_backend_enabled = 1;
+			if (this_backend_enabled) {
+				enabled_backends_mask |= (1 << i);
+				*backend_disable_mask_per_asic &= ~(1 << i);
+				++enabled_backends_count;
+			}
+		}
+	}
+
+
+	memset((uint8_t *)&swizzle_pipe[0], 0, sizeof(u32) * SI_MAX_PIPES);
+	switch (rdev->family) {
+	case CHIP_TAHITI:
+	case CHIP_PITCAIRN:
+	case CHIP_VERDE:
+		force_no_swizzle = true;
+		break;
+	default:
+		force_no_swizzle = false;
+		break;
+	}
+	if (force_no_swizzle) {
+		bool last_backend_enabled = false;
+
+		force_no_swizzle = false;
+		for (i = 0; i < SI_MAX_BACKENDS; ++i) {
+			if (((enabled_backends_mask >> i) & 1) == 1) {
+				if (last_backend_enabled)
+					force_no_swizzle = true;
+				last_backend_enabled = true;
+			} else
+				last_backend_enabled = false;
+		}
+	}
+
+	switch (num_tile_pipes) {
+	case 1:
+	case 3:
+	case 5:
+	case 7:
+		DRM_ERROR("odd number of pipes!\n");
+		break;
+	case 2:
+		swizzle_pipe[0] = 0;
+		swizzle_pipe[1] = 1;
+		break;
+	case 4:
+		if (force_no_swizzle) {
+			swizzle_pipe[0] = 0;
+			swizzle_pipe[1] = 1;
+			swizzle_pipe[2] = 2;
+			swizzle_pipe[3] = 3;
+		} else {
+			swizzle_pipe[0] = 0;
+			swizzle_pipe[1] = 2;
+			swizzle_pipe[2] = 1;
+			swizzle_pipe[3] = 3;
+		}
+		break;
+	case 6:
+		if (force_no_swizzle) {
+			swizzle_pipe[0] = 0;
+			swizzle_pipe[1] = 1;
+			swizzle_pipe[2] = 2;
+			swizzle_pipe[3] = 3;
+			swizzle_pipe[4] = 4;
+			swizzle_pipe[5] = 5;
+		} else {
+			swizzle_pipe[0] = 0;
+			swizzle_pipe[1] = 2;
+			swizzle_pipe[2] = 4;
+			swizzle_pipe[3] = 1;
+			swizzle_pipe[4] = 3;
+			swizzle_pipe[5] = 5;
+		}
+		break;
+	case 8:
+		if (force_no_swizzle) {
+			swizzle_pipe[0] = 0;
+			swizzle_pipe[1] = 1;
+			swizzle_pipe[2] = 2;
+			swizzle_pipe[3] = 3;
+			swizzle_pipe[4] = 4;
+			swizzle_pipe[5] = 5;
+			swizzle_pipe[6] = 6;
+			swizzle_pipe[7] = 7;
+		} else {
+			swizzle_pipe[0] = 0;
+			swizzle_pipe[1] = 2;
+			swizzle_pipe[2] = 4;
+			swizzle_pipe[3] = 6;
+			swizzle_pipe[4] = 1;
+			swizzle_pipe[5] = 3;
+			swizzle_pipe[6] = 5;
+			swizzle_pipe[7] = 7;
+		}
+		break;
+	}
+
+	for (cur_pipe = 0; cur_pipe < num_tile_pipes; ++cur_pipe) {
+		while (((1 << cur_backend) & enabled_backends_mask) == 0)
+			cur_backend = (cur_backend + 1) % SI_MAX_BACKENDS;
+
+		backend_map |= (((cur_backend & 0xf) << (swizzle_pipe[cur_pipe] * 4)));
+
+		cur_backend = (cur_backend + 1) % SI_MAX_BACKENDS;
+	}
+
+	return backend_map;
+}
+
+static u32 si_get_disable_mask_per_asic(struct radeon_device *rdev,
+					u32 disable_mask_per_se,
+					u32 max_disable_mask_per_se,
+					u32 num_shader_engines)
+{
+	u32 disable_field_width_per_se = r600_count_pipe_bits(disable_mask_per_se);
+	u32 disable_mask_per_asic = disable_mask_per_se & max_disable_mask_per_se;
+
+	if (num_shader_engines == 1)
+		return disable_mask_per_asic;
+	else if (num_shader_engines == 2)
+		return disable_mask_per_asic | (disable_mask_per_asic << disable_field_width_per_se);
+	else
+		return 0xffffffff;
+}
+
+static void si_tiling_mode_table_init(struct radeon_device *rdev)
+{
+	const u32 num_tile_mode_states = 32;
+	u32 reg_offset, gb_tile_moden, split_equal_to_row_size;
+
+	switch (rdev->config.si.mem_row_size_in_kb) {
+	case 1:
+		split_equal_to_row_size = ADDR_SURF_TILE_SPLIT_1KB;
+		break;
+	case 2:
+	default:
+		split_equal_to_row_size = ADDR_SURF_TILE_SPLIT_2KB;
+		break;
+	case 4:
+		split_equal_to_row_size = ADDR_SURF_TILE_SPLIT_4KB;
+		break;
+	}
+
+	if ((rdev->family == CHIP_TAHITI) ||
+	    (rdev->family == CHIP_PITCAIRN)) {
+		for (reg_offset = 0; reg_offset < num_tile_mode_states; reg_offset++) {
+			switch (reg_offset) {
+			case 0:  /* non-AA compressed depth or any compressed stencil */
+				gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE(ADDR_SURF_DEPTH_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) |
+						 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_64B) |
+						 NUM_BANKS(ADDR_SURF_16_BANK) |
+						 BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2));
+				break;
+			case 1:  /* 2xAA/4xAA compressed depth only */
+				gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE(ADDR_SURF_DEPTH_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) |
+						 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_128B) |
+						 NUM_BANKS(ADDR_SURF_16_BANK) |
+						 BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2));
+				break;
+			case 2:  /* 8xAA compressed depth only */
+				gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE(ADDR_SURF_DEPTH_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) |
+						 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_256B) |
+						 NUM_BANKS(ADDR_SURF_16_BANK) |
+						 BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2));
+				break;
+			case 3:  /* 2xAA/4xAA compressed depth with stencil (for depth buffer) */
+				gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE(ADDR_SURF_DEPTH_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) |
+						 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_128B) |
+						 NUM_BANKS(ADDR_SURF_16_BANK) |
+						 BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2));
+				break;
+			case 4:  /* Maps w/ a dimension less than the 2D macro-tile dimensions (for mipmapped depth textures) */
+				gb_tile_moden = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) |
+						 MICRO_TILE_MODE(ADDR_SURF_DEPTH_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) |
+						 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_64B) |
+						 NUM_BANKS(ADDR_SURF_16_BANK) |
+						 BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2));
+				break;
+			case 5:  /* Uncompressed 16bpp depth - and stencil buffer allocated with it */
+				gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE(ADDR_SURF_DEPTH_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) |
+						 TILE_SPLIT(split_equal_to_row_size) |
+						 NUM_BANKS(ADDR_SURF_16_BANK) |
+						 BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2));
+				break;
+			case 6:  /* Uncompressed 32bpp depth - and stencil buffer allocated with it */
+				gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE(ADDR_SURF_DEPTH_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) |
+						 TILE_SPLIT(split_equal_to_row_size) |
+						 NUM_BANKS(ADDR_SURF_16_BANK) |
+						 BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_1));
+				break;
+			case 7:  /* Uncompressed 8bpp stencil without depth (drivers typically do not use) */
+				gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE(ADDR_SURF_DEPTH_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) |
+						 TILE_SPLIT(split_equal_to_row_size) |
+						 NUM_BANKS(ADDR_SURF_16_BANK) |
+						 BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2));
+				break;
+			case 8:  /* 1D and 1D Array Surfaces */
+				gb_tile_moden = (ARRAY_MODE(ARRAY_LINEAR_ALIGNED) |
+						 MICRO_TILE_MODE(ADDR_SURF_DISPLAY_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) |
+						 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_64B) |
+						 NUM_BANKS(ADDR_SURF_16_BANK) |
+						 BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2));
+				break;
+			case 9:  /* Displayable maps. */
+				gb_tile_moden = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) |
+						 MICRO_TILE_MODE(ADDR_SURF_DISPLAY_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) |
+						 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_64B) |
+						 NUM_BANKS(ADDR_SURF_16_BANK) |
+						 BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2));
+				break;
+			case 10:  /* Display 8bpp. */
+				gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE(ADDR_SURF_DISPLAY_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) |
+						 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_256B) |
+						 NUM_BANKS(ADDR_SURF_16_BANK) |
+						 BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2));
+				break;
+			case 11:  /* Display 16bpp. */
+				gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE(ADDR_SURF_DISPLAY_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) |
+						 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_256B) |
+						 NUM_BANKS(ADDR_SURF_16_BANK) |
+						 BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2));
+				break;
+			case 12:  /* Display 32bpp. */
+				gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE(ADDR_SURF_DISPLAY_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) |
+						 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_512B) |
+						 NUM_BANKS(ADDR_SURF_16_BANK) |
+						 BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_1));
+				break;
+			case 13:  /* Thin. */
+				gb_tile_moden = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) |
+						 MICRO_TILE_MODE(ADDR_SURF_THIN_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) |
+						 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_64B) |
+						 NUM_BANKS(ADDR_SURF_16_BANK) |
+						 BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2));
+				break;
+			case 14:  /* Thin 8 bpp. */
+				gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE(ADDR_SURF_THIN_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) |
+						 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_256B) |
+						 NUM_BANKS(ADDR_SURF_16_BANK) |
+						 BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_1));
+				break;
+			case 15:  /* Thin 16 bpp. */
+				gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE(ADDR_SURF_THIN_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) |
+						 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_256B) |
+						 NUM_BANKS(ADDR_SURF_16_BANK) |
+						 BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_1));
+				break;
+			case 16:  /* Thin 32 bpp. */
+				gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE(ADDR_SURF_THIN_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) |
+						 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_512B) |
+						 NUM_BANKS(ADDR_SURF_16_BANK) |
+						 BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_1));
+				break;
+			case 17:  /* Thin 64 bpp. */
+				gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE(ADDR_SURF_THIN_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) |
+						 TILE_SPLIT(split_equal_to_row_size) |
+						 NUM_BANKS(ADDR_SURF_16_BANK) |
+						 BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_1));
+				break;
+			case 21:  /* 8 bpp PRT. */
+				gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE(ADDR_SURF_THIN_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) |
+						 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_256B) |
+						 NUM_BANKS(ADDR_SURF_16_BANK) |
+						 BANK_WIDTH(ADDR_SURF_BANK_WIDTH_2) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2));
+				break;
+			case 22:  /* 16 bpp PRT */
+				gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE(ADDR_SURF_THIN_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) |
+						 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_256B) |
+						 NUM_BANKS(ADDR_SURF_16_BANK) |
+						 BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4));
+				break;
+			case 23:  /* 32 bpp PRT */
+				gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE(ADDR_SURF_THIN_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) |
+						 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_256B) |
+						 NUM_BANKS(ADDR_SURF_16_BANK) |
+						 BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2));
+				break;
+			case 24:  /* 64 bpp PRT */
+				gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE(ADDR_SURF_THIN_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) |
+						 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_512B) |
+						 NUM_BANKS(ADDR_SURF_16_BANK) |
+						 BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2));
+				break;
+			case 25:  /* 128 bpp PRT */
+				gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE(ADDR_SURF_THIN_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) |
+						 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_1KB) |
+						 NUM_BANKS(ADDR_SURF_8_BANK) |
+						 BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_1));
+				break;
+			default:
+				gb_tile_moden = 0;
+				break;
+			}
+			WREG32(GB_TILE_MODE0 + (reg_offset * 4), gb_tile_moden);
+		}
+	} else if (rdev->family == CHIP_VERDE) {
+		for (reg_offset = 0; reg_offset < num_tile_mode_states; reg_offset++) {
+			switch (reg_offset) {
+			case 0:  /* non-AA compressed depth or any compressed stencil */
+				gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE(ADDR_SURF_DEPTH_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P4_8x16) |
+						 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_64B) |
+						 NUM_BANKS(ADDR_SURF_16_BANK) |
+						 BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4));
+				break;
+			case 1:  /* 2xAA/4xAA compressed depth only */
+				gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE(ADDR_SURF_DEPTH_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P4_8x16) |
+						 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_128B) |
+						 NUM_BANKS(ADDR_SURF_16_BANK) |
+						 BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4));
+				break;
+			case 2:  /* 8xAA compressed depth only */
+				gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE(ADDR_SURF_DEPTH_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P4_8x16) |
+						 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_256B) |
+						 NUM_BANKS(ADDR_SURF_16_BANK) |
+						 BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4));
+				break;
+			case 3:  /* 2xAA/4xAA compressed depth with stencil (for depth buffer) */
+				gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE(ADDR_SURF_DEPTH_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P4_8x16) |
+						 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_128B) |
+						 NUM_BANKS(ADDR_SURF_16_BANK) |
+						 BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4));
+				break;
+			case 4:  /* Maps w/ a dimension less than the 2D macro-tile dimensions (for mipmapped depth textures) */
+				gb_tile_moden = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) |
+						 MICRO_TILE_MODE(ADDR_SURF_DEPTH_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P4_8x16) |
+						 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_64B) |
+						 NUM_BANKS(ADDR_SURF_16_BANK) |
+						 BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2));
+				break;
+			case 5:  /* Uncompressed 16bpp depth - and stencil buffer allocated with it */
+				gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE(ADDR_SURF_DEPTH_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P4_8x16) |
+						 TILE_SPLIT(split_equal_to_row_size) |
+						 NUM_BANKS(ADDR_SURF_16_BANK) |
+						 BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2));
+				break;
+			case 6:  /* Uncompressed 32bpp depth - and stencil buffer allocated with it */
+				gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE(ADDR_SURF_DEPTH_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P4_8x16) |
+						 TILE_SPLIT(split_equal_to_row_size) |
+						 NUM_BANKS(ADDR_SURF_16_BANK) |
+						 BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2));
+				break;
+			case 7:  /* Uncompressed 8bpp stencil without depth (drivers typically do not use) */
+				gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE(ADDR_SURF_DEPTH_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P4_8x16) |
+						 TILE_SPLIT(split_equal_to_row_size) |
+						 NUM_BANKS(ADDR_SURF_16_BANK) |
+						 BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4));
+				break;
+			case 8:  /* 1D and 1D Array Surfaces */
+				gb_tile_moden = (ARRAY_MODE(ARRAY_LINEAR_ALIGNED) |
+						 MICRO_TILE_MODE(ADDR_SURF_DISPLAY_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P4_8x16) |
+						 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_64B) |
+						 NUM_BANKS(ADDR_SURF_16_BANK) |
+						 BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2));
+				break;
+			case 9:  /* Displayable maps. */
+				gb_tile_moden = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) |
+						 MICRO_TILE_MODE(ADDR_SURF_DISPLAY_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P4_8x16) |
+						 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_64B) |
+						 NUM_BANKS(ADDR_SURF_16_BANK) |
+						 BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2));
+				break;
+			case 10:  /* Display 8bpp. */
+				gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE(ADDR_SURF_DISPLAY_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P4_8x16) |
+						 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_256B) |
+						 NUM_BANKS(ADDR_SURF_16_BANK) |
+						 BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4));
+				break;
+			case 11:  /* Display 16bpp. */
+				gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE(ADDR_SURF_DISPLAY_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P4_8x16) |
+						 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_256B) |
+						 NUM_BANKS(ADDR_SURF_16_BANK) |
+						 BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2));
+				break;
+			case 12:  /* Display 32bpp. */
+				gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE(ADDR_SURF_DISPLAY_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P4_8x16) |
+						 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_512B) |
+						 NUM_BANKS(ADDR_SURF_16_BANK) |
+						 BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2));
+				break;
+			case 13:  /* Thin. */
+				gb_tile_moden = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) |
+						 MICRO_TILE_MODE(ADDR_SURF_THIN_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P4_8x16) |
+						 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_64B) |
+						 NUM_BANKS(ADDR_SURF_16_BANK) |
+						 BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2));
+				break;
+			case 14:  /* Thin 8 bpp. */
+				gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE(ADDR_SURF_THIN_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P4_8x16) |
+						 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_256B) |
+						 NUM_BANKS(ADDR_SURF_16_BANK) |
+						 BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2));
+				break;
+			case 15:  /* Thin 16 bpp. */
+				gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE(ADDR_SURF_THIN_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P4_8x16) |
+						 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_256B) |
+						 NUM_BANKS(ADDR_SURF_16_BANK) |
+						 BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2));
+				break;
+			case 16:  /* Thin 32 bpp. */
+				gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE(ADDR_SURF_THIN_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P4_8x16) |
+						 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_512B) |
+						 NUM_BANKS(ADDR_SURF_16_BANK) |
+						 BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2));
+				break;
+			case 17:  /* Thin 64 bpp. */
+				gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE(ADDR_SURF_THIN_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P4_8x16) |
+						 TILE_SPLIT(split_equal_to_row_size) |
+						 NUM_BANKS(ADDR_SURF_16_BANK) |
+						 BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2));
+				break;
+			case 21:  /* 8 bpp PRT. */
+				gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE(ADDR_SURF_THIN_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) |
+						 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_256B) |
+						 NUM_BANKS(ADDR_SURF_16_BANK) |
+						 BANK_WIDTH(ADDR_SURF_BANK_WIDTH_2) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2));
+				break;
+			case 22:  /* 16 bpp PRT */
+				gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE(ADDR_SURF_THIN_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) |
+						 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_256B) |
+						 NUM_BANKS(ADDR_SURF_16_BANK) |
+						 BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4));
+				break;
+			case 23:  /* 32 bpp PRT */
+				gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE(ADDR_SURF_THIN_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) |
+						 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_256B) |
+						 NUM_BANKS(ADDR_SURF_16_BANK) |
+						 BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2));
+				break;
+			case 24:  /* 64 bpp PRT */
+				gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE(ADDR_SURF_THIN_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) |
+						 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_512B) |
+						 NUM_BANKS(ADDR_SURF_16_BANK) |
+						 BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2));
+				break;
+			case 25:  /* 128 bpp PRT */
+				gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE(ADDR_SURF_THIN_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) |
+						 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_1KB) |
+						 NUM_BANKS(ADDR_SURF_8_BANK) |
+						 BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_1));
+				break;
+			default:
+				gb_tile_moden = 0;
+				break;
+			}
+			WREG32(GB_TILE_MODE0 + (reg_offset * 4), gb_tile_moden);
+		}
+	} else
+		DRM_ERROR("unknown asic: 0x%x\n", rdev->family);
+}
+
+static void si_gpu_init(struct radeon_device *rdev)
+{
+	u32 cc_rb_backend_disable = 0;
+	u32 cc_gc_shader_array_config;
+	u32 gb_addr_config = 0;
+	u32 mc_shared_chmap, mc_arb_ramcfg;
+	u32 gb_backend_map;
+	u32 cgts_tcc_disable;
+	u32 sx_debug_1;
+	u32 gc_user_shader_array_config;
+	u32 gc_user_rb_backend_disable;
+	u32 cgts_user_tcc_disable;
+	u32 hdp_host_path_cntl;
+	u32 tmp;
+	int i, j;
+
+	switch (rdev->family) {
+	case CHIP_TAHITI:
+		rdev->config.si.max_shader_engines = 2;
+		rdev->config.si.max_pipes_per_simd = 4;
+		rdev->config.si.max_tile_pipes = 12;
+		rdev->config.si.max_simds_per_se = 8;
+		rdev->config.si.max_backends_per_se = 4;
+		rdev->config.si.max_texture_channel_caches = 12;
+		rdev->config.si.max_gprs = 256;
+		rdev->config.si.max_gs_threads = 32;
+		rdev->config.si.max_hw_contexts = 8;
+
+		rdev->config.si.sc_prim_fifo_size_frontend = 0x20;
+		rdev->config.si.sc_prim_fifo_size_backend = 0x100;
+		rdev->config.si.sc_hiz_tile_fifo_size = 0x30;
+		rdev->config.si.sc_earlyz_tile_fifo_size = 0x130;
+		break;
+	case CHIP_PITCAIRN:
+		rdev->config.si.max_shader_engines = 2;
+		rdev->config.si.max_pipes_per_simd = 4;
+		rdev->config.si.max_tile_pipes = 8;
+		rdev->config.si.max_simds_per_se = 5;
+		rdev->config.si.max_backends_per_se = 4;
+		rdev->config.si.max_texture_channel_caches = 8;
+		rdev->config.si.max_gprs = 256;
+		rdev->config.si.max_gs_threads = 32;
+		rdev->config.si.max_hw_contexts = 8;
+
+		rdev->config.si.sc_prim_fifo_size_frontend = 0x20;
+		rdev->config.si.sc_prim_fifo_size_backend = 0x100;
+		rdev->config.si.sc_hiz_tile_fifo_size = 0x30;
+		rdev->config.si.sc_earlyz_tile_fifo_size = 0x130;
+		break;
+	case CHIP_VERDE:
+	default:
+		rdev->config.si.max_shader_engines = 1;
+		rdev->config.si.max_pipes_per_simd = 4;
+		rdev->config.si.max_tile_pipes = 4;
+		rdev->config.si.max_simds_per_se = 2;
+		rdev->config.si.max_backends_per_se = 4;
+		rdev->config.si.max_texture_channel_caches = 4;
+		rdev->config.si.max_gprs = 256;
+		rdev->config.si.max_gs_threads = 32;
+		rdev->config.si.max_hw_contexts = 8;
+
+		rdev->config.si.sc_prim_fifo_size_frontend = 0x20;
+		rdev->config.si.sc_prim_fifo_size_backend = 0x40;
+		rdev->config.si.sc_hiz_tile_fifo_size = 0x30;
+		rdev->config.si.sc_earlyz_tile_fifo_size = 0x130;
+		break;
+	}
+
+	/* Initialize HDP */
+	for (i = 0, j = 0; i < 32; i++, j += 0x18) {
+		WREG32((0x2c14 + j), 0x00000000);
+		WREG32((0x2c18 + j), 0x00000000);
+		WREG32((0x2c1c + j), 0x00000000);
+		WREG32((0x2c20 + j), 0x00000000);
+		WREG32((0x2c24 + j), 0x00000000);
+	}
+
+	WREG32(GRBM_CNTL, GRBM_READ_TIMEOUT(0xff));
+
+	evergreen_fix_pci_max_read_req_size(rdev);
+
+	WREG32(BIF_FB_EN, FB_READ_EN | FB_WRITE_EN);
+
+	mc_shared_chmap = RREG32(MC_SHARED_CHMAP);
+	mc_arb_ramcfg = RREG32(MC_ARB_RAMCFG);
+
+	cc_rb_backend_disable = RREG32(CC_RB_BACKEND_DISABLE);
+	cc_gc_shader_array_config = RREG32(CC_GC_SHADER_ARRAY_CONFIG);
+	cgts_tcc_disable = 0xffff0000;
+	for (i = 0; i < rdev->config.si.max_texture_channel_caches; i++)
+		cgts_tcc_disable &= ~(1 << (16 + i));
+	gc_user_rb_backend_disable = RREG32(GC_USER_RB_BACKEND_DISABLE);
+	gc_user_shader_array_config = RREG32(GC_USER_SHADER_ARRAY_CONFIG);
+	cgts_user_tcc_disable = RREG32(CGTS_USER_TCC_DISABLE);
+
+	rdev->config.si.num_shader_engines = rdev->config.si.max_shader_engines;
+	rdev->config.si.num_tile_pipes = rdev->config.si.max_tile_pipes;
+	tmp = ((~gc_user_rb_backend_disable) & BACKEND_DISABLE_MASK) >> BACKEND_DISABLE_SHIFT;
+	rdev->config.si.num_backends_per_se = r600_count_pipe_bits(tmp);
+	tmp = (gc_user_rb_backend_disable & BACKEND_DISABLE_MASK) >> BACKEND_DISABLE_SHIFT;
+	rdev->config.si.backend_disable_mask_per_asic =
+		si_get_disable_mask_per_asic(rdev, tmp, SI_MAX_BACKENDS_PER_SE_MASK,
+					     rdev->config.si.num_shader_engines);
+	rdev->config.si.backend_map =
+		si_get_tile_pipe_to_backend_map(rdev, rdev->config.si.num_tile_pipes,
+						rdev->config.si.num_backends_per_se *
+						rdev->config.si.num_shader_engines,
+						&rdev->config.si.backend_disable_mask_per_asic,
+						rdev->config.si.num_shader_engines);
+	tmp = ((~cgts_user_tcc_disable) & TCC_DISABLE_MASK) >> TCC_DISABLE_SHIFT;
+	rdev->config.si.num_texture_channel_caches = r600_count_pipe_bits(tmp);
+	rdev->config.si.mem_max_burst_length_bytes = 256;
+	tmp = (mc_arb_ramcfg & NOOFCOLS_MASK) >> NOOFCOLS_SHIFT;
+	rdev->config.si.mem_row_size_in_kb = (4 * (1 << (8 + tmp))) / 1024;
+	if (rdev->config.si.mem_row_size_in_kb > 4)
+		rdev->config.si.mem_row_size_in_kb = 4;
+	/* XXX use MC settings? */
+	rdev->config.si.shader_engine_tile_size = 32;
+	rdev->config.si.num_gpus = 1;
+	rdev->config.si.multi_gpu_tile_size = 64;
+
+	gb_addr_config = 0;
+	switch (rdev->config.si.num_tile_pipes) {
+	case 1:
+		gb_addr_config |= NUM_PIPES(0);
+		break;
+	case 2:
+		gb_addr_config |= NUM_PIPES(1);
+		break;
+	case 4:
+		gb_addr_config |= NUM_PIPES(2);
+		break;
+	case 8:
+	default:
+		gb_addr_config |= NUM_PIPES(3);
+		break;
+	}
+
+	tmp = (rdev->config.si.mem_max_burst_length_bytes / 256) - 1;
+	gb_addr_config |= PIPE_INTERLEAVE_SIZE(tmp);
+	gb_addr_config |= NUM_SHADER_ENGINES(rdev->config.si.num_shader_engines - 1);
+	tmp = (rdev->config.si.shader_engine_tile_size / 16) - 1;
+	gb_addr_config |= SHADER_ENGINE_TILE_SIZE(tmp);
+	switch (rdev->config.si.num_gpus) {
+	case 1:
+	default:
+		gb_addr_config |= NUM_GPUS(0);
+		break;
+	case 2:
+		gb_addr_config |= NUM_GPUS(1);
+		break;
+	case 4:
+		gb_addr_config |= NUM_GPUS(2);
+		break;
+	}
+	switch (rdev->config.si.multi_gpu_tile_size) {
+	case 16:
+		gb_addr_config |= MULTI_GPU_TILE_SIZE(0);
+		break;
+	case 32:
+	default:
+		gb_addr_config |= MULTI_GPU_TILE_SIZE(1);
+		break;
+	case 64:
+		gb_addr_config |= MULTI_GPU_TILE_SIZE(2);
+		break;
+	case 128:
+		gb_addr_config |= MULTI_GPU_TILE_SIZE(3);
+		break;
+	}
+	switch (rdev->config.si.mem_row_size_in_kb) {
+	case 1:
+	default:
+		gb_addr_config |= ROW_SIZE(0);
+		break;
+	case 2:
+		gb_addr_config |= ROW_SIZE(1);
+		break;
+	case 4:
+		gb_addr_config |= ROW_SIZE(2);
+		break;
+	}
+
+	tmp = (gb_addr_config & NUM_PIPES_MASK) >> NUM_PIPES_SHIFT;
+	rdev->config.si.num_tile_pipes = (1 << tmp);
+	tmp = (gb_addr_config & PIPE_INTERLEAVE_SIZE_MASK) >> PIPE_INTERLEAVE_SIZE_SHIFT;
+	rdev->config.si.mem_max_burst_length_bytes = (tmp + 1) * 256;
+	tmp = (gb_addr_config & NUM_SHADER_ENGINES_MASK) >> NUM_SHADER_ENGINES_SHIFT;
+	rdev->config.si.num_shader_engines = tmp + 1;
+	tmp = (gb_addr_config & NUM_GPUS_MASK) >> NUM_GPUS_SHIFT;
+	rdev->config.si.num_gpus = tmp + 1;
+	tmp = (gb_addr_config & MULTI_GPU_TILE_SIZE_MASK) >> MULTI_GPU_TILE_SIZE_SHIFT;
+	rdev->config.si.multi_gpu_tile_size = 1 << tmp;
+	tmp = (gb_addr_config & ROW_SIZE_MASK) >> ROW_SIZE_SHIFT;
+	rdev->config.si.mem_row_size_in_kb = 1 << tmp;
+
+	gb_backend_map =
+		si_get_tile_pipe_to_backend_map(rdev, rdev->config.si.num_tile_pipes,
+						rdev->config.si.num_backends_per_se *
+						rdev->config.si.num_shader_engines,
+						&rdev->config.si.backend_disable_mask_per_asic,
+						rdev->config.si.num_shader_engines);
+
+	/* setup tiling info dword.  gb_addr_config is not adequate since it does
+	 * not have bank info, so create a custom tiling dword.
+	 * bits 3:0   num_pipes
+	 * bits 7:4   num_banks
+	 * bits 11:8  group_size
+	 * bits 15:12 row_size
+	 */
+	rdev->config.si.tile_config = 0;
+	switch (rdev->config.si.num_tile_pipes) {
+	case 1:
+		rdev->config.si.tile_config |= (0 << 0);
+		break;
+	case 2:
+		rdev->config.si.tile_config |= (1 << 0);
+		break;
+	case 4:
+		rdev->config.si.tile_config |= (2 << 0);
+		break;
+	case 8:
+	default:
+		/* XXX what about 12? */
+		rdev->config.si.tile_config |= (3 << 0);
+		break;
+	}
+	rdev->config.si.tile_config |=
+		((mc_arb_ramcfg & NOOFBANK_MASK) >> NOOFBANK_SHIFT) << 4;
+	rdev->config.si.tile_config |=
+		((gb_addr_config & PIPE_INTERLEAVE_SIZE_MASK) >> PIPE_INTERLEAVE_SIZE_SHIFT) << 8;
+	rdev->config.si.tile_config |=
+		((gb_addr_config & ROW_SIZE_MASK) >> ROW_SIZE_SHIFT) << 12;
+
+	rdev->config.si.backend_map = gb_backend_map;
+	WREG32(GB_ADDR_CONFIG, gb_addr_config);
+	WREG32(DMIF_ADDR_CONFIG, gb_addr_config);
+	WREG32(HDP_ADDR_CONFIG, gb_addr_config);
+
+	/* primary versions */
+	WREG32(CC_RB_BACKEND_DISABLE, cc_rb_backend_disable);
+	WREG32(CC_SYS_RB_BACKEND_DISABLE, cc_rb_backend_disable);
+	WREG32(CC_GC_SHADER_ARRAY_CONFIG, cc_gc_shader_array_config);
+
+	WREG32(CGTS_TCC_DISABLE, cgts_tcc_disable);
+
+	/* user versions */
+	WREG32(GC_USER_RB_BACKEND_DISABLE, cc_rb_backend_disable);
+	WREG32(GC_USER_SYS_RB_BACKEND_DISABLE, cc_rb_backend_disable);
+	WREG32(GC_USER_SHADER_ARRAY_CONFIG, cc_gc_shader_array_config);
+
+	WREG32(CGTS_USER_TCC_DISABLE, cgts_tcc_disable);
+
+	si_tiling_mode_table_init(rdev);
+
+	/* set HW defaults for 3D engine */
+	WREG32(CP_QUEUE_THRESHOLDS, (ROQ_IB1_START(0x16) |
+				     ROQ_IB2_START(0x2b)));
+	WREG32(CP_MEQ_THRESHOLDS, MEQ1_START(0x30) | MEQ2_START(0x60));
+
+	sx_debug_1 = RREG32(SX_DEBUG_1);
+	WREG32(SX_DEBUG_1, sx_debug_1);
+
+	WREG32(SPI_CONFIG_CNTL_1, VTX_DONE_DELAY(4));
+
+	WREG32(PA_SC_FIFO_SIZE, (SC_FRONTEND_PRIM_FIFO_SIZE(rdev->config.si.sc_prim_fifo_size_frontend) |
+				 SC_BACKEND_PRIM_FIFO_SIZE(rdev->config.si.sc_prim_fifo_size_backend) |
+				 SC_HIZ_TILE_FIFO_SIZE(rdev->config.si.sc_hiz_tile_fifo_size) |
+				 SC_EARLYZ_TILE_FIFO_SIZE(rdev->config.si.sc_earlyz_tile_fifo_size)));
+
+	WREG32(VGT_NUM_INSTANCES, 1);
+
+	WREG32(CP_PERFMON_CNTL, 0);
+
+	WREG32(SQ_CONFIG, 0);
+
+	WREG32(PA_SC_FORCE_EOV_MAX_CNTS, (FORCE_EOV_MAX_CLK_CNT(4095) |
+					  FORCE_EOV_MAX_REZ_CNT(255)));
+
+	WREG32(VGT_CACHE_INVALIDATION, CACHE_INVALIDATION(VC_AND_TC) |
+	       AUTO_INVLD_EN(ES_AND_GS_AUTO));
+
+	WREG32(VGT_GS_VERTEX_REUSE, 16);
+	WREG32(PA_SC_LINE_STIPPLE_STATE, 0);
+
+	WREG32(CB_PERFCOUNTER0_SELECT0, 0);
+	WREG32(CB_PERFCOUNTER0_SELECT1, 0);
+	WREG32(CB_PERFCOUNTER1_SELECT0, 0);
+	WREG32(CB_PERFCOUNTER1_SELECT1, 0);
+	WREG32(CB_PERFCOUNTER2_SELECT0, 0);
+	WREG32(CB_PERFCOUNTER2_SELECT1, 0);
+	WREG32(CB_PERFCOUNTER3_SELECT0, 0);
+	WREG32(CB_PERFCOUNTER3_SELECT1, 0);
+
+	tmp = RREG32(HDP_MISC_CNTL);
+	tmp |= HDP_FLUSH_INVALIDATE_CACHE;
+	WREG32(HDP_MISC_CNTL, tmp);
+
+	hdp_host_path_cntl = RREG32(HDP_HOST_PATH_CNTL);
+	WREG32(HDP_HOST_PATH_CNTL, hdp_host_path_cntl);
+
+	WREG32(PA_CL_ENHANCE, CLIP_VTX_REORDER_ENA | NUM_CLIP_SEQ(3));
+
+	udelay(50);
+}
+
+/*
+ * GPU scratch registers helpers function.
+ */
+static void si_scratch_init(struct radeon_device *rdev)
+{
+	int i;
+
+	rdev->scratch.num_reg = 7;
+	rdev->scratch.reg_base = SCRATCH_REG0;
+	for (i = 0; i < rdev->scratch.num_reg; i++) {
+		rdev->scratch.free[i] = true;
+		rdev->scratch.reg[i] = rdev->scratch.reg_base + (i * 4);
+	}
+}
+
+void si_fence_ring_emit(struct radeon_device *rdev,
+			struct radeon_fence *fence)
+{
+	struct radeon_ring *ring = &rdev->ring[fence->ring];
+	u64 addr = rdev->fence_drv[fence->ring].gpu_addr;
+
+	/* flush read cache over gart */
+	radeon_ring_write(ring, PACKET3(PACKET3_SET_CONFIG_REG, 1));
+	radeon_ring_write(ring, (CP_COHER_CNTL2 - PACKET3_SET_CONFIG_REG_START) >> 2);
+	radeon_ring_write(ring, 0);
+	radeon_ring_write(ring, PACKET3(PACKET3_SURFACE_SYNC, 3));
+	radeon_ring_write(ring, PACKET3_TCL1_ACTION_ENA |
+			  PACKET3_TC_ACTION_ENA |
+			  PACKET3_SH_KCACHE_ACTION_ENA |
+			  PACKET3_SH_ICACHE_ACTION_ENA);
+	radeon_ring_write(ring, 0xFFFFFFFF);
+	radeon_ring_write(ring, 0);
+	radeon_ring_write(ring, 10); /* poll interval */
+	/* EVENT_WRITE_EOP - flush caches, send int */
+	radeon_ring_write(ring, PACKET3(PACKET3_EVENT_WRITE_EOP, 4));
+	radeon_ring_write(ring, EVENT_TYPE(CACHE_FLUSH_AND_INV_TS_EVENT) | EVENT_INDEX(5));
+	radeon_ring_write(ring, addr & 0xffffffff);
+	radeon_ring_write(ring, (upper_32_bits(addr) & 0xff) | DATA_SEL(1) | INT_SEL(2));
+	radeon_ring_write(ring, fence->seq);
+	radeon_ring_write(ring, 0);
+}
+
+/*
+ * IB stuff
+ */
+void si_ring_ib_execute(struct radeon_device *rdev, struct radeon_ib *ib)
+{
+	struct radeon_ring *ring = &rdev->ring[ib->fence->ring];
+	u32 header;
+
+	if (ib->is_const_ib)
+		header = PACKET3(PACKET3_INDIRECT_BUFFER_CONST, 2);
+	else
+		header = PACKET3(PACKET3_INDIRECT_BUFFER, 2);
+
+	radeon_ring_write(ring, header);
+	radeon_ring_write(ring,
+#ifdef __BIG_ENDIAN
+			  (2 << 0) |
+#endif
+			  (ib->gpu_addr & 0xFFFFFFFC));
+	radeon_ring_write(ring, upper_32_bits(ib->gpu_addr) & 0xFFFF);
+	radeon_ring_write(ring, ib->length_dw | (ib->vm_id << 24));
+
+	/* flush read cache over gart for this vmid */
+	radeon_ring_write(ring, PACKET3(PACKET3_SET_CONFIG_REG, 1));
+	radeon_ring_write(ring, (CP_COHER_CNTL2 - PACKET3_SET_CONFIG_REG_START) >> 2);
+	radeon_ring_write(ring, ib->vm_id);
+	radeon_ring_write(ring, PACKET3(PACKET3_SURFACE_SYNC, 3));
+	radeon_ring_write(ring, PACKET3_TCL1_ACTION_ENA |
+			  PACKET3_TC_ACTION_ENA |
+			  PACKET3_SH_KCACHE_ACTION_ENA |
+			  PACKET3_SH_ICACHE_ACTION_ENA);
+	radeon_ring_write(ring, 0xFFFFFFFF);
+	radeon_ring_write(ring, 0);
+	radeon_ring_write(ring, 10); /* poll interval */
+}
+
+/*
+ * CP.
+ */
+static void si_cp_enable(struct radeon_device *rdev, bool enable)
+{
+	if (enable)
+		WREG32(CP_ME_CNTL, 0);
+	else {
+		radeon_ttm_set_active_vram_size(rdev, rdev->mc.visible_vram_size);
+		WREG32(CP_ME_CNTL, (CP_ME_HALT | CP_PFP_HALT | CP_CE_HALT));
+		WREG32(SCRATCH_UMSK, 0);
+	}
+	udelay(50);
+}
+
+static int si_cp_load_microcode(struct radeon_device *rdev)
+{
+	const __be32 *fw_data;
+	int i;
+
+	if (!rdev->me_fw || !rdev->pfp_fw)
+		return -EINVAL;
+
+	si_cp_enable(rdev, false);
+
+	/* PFP */
+	fw_data = (const __be32 *)rdev->pfp_fw->data;
+	WREG32(CP_PFP_UCODE_ADDR, 0);
+	for (i = 0; i < SI_PFP_UCODE_SIZE; i++)
+		WREG32(CP_PFP_UCODE_DATA, be32_to_cpup(fw_data++));
+	WREG32(CP_PFP_UCODE_ADDR, 0);
+
+	/* CE */
+	fw_data = (const __be32 *)rdev->ce_fw->data;
+	WREG32(CP_CE_UCODE_ADDR, 0);
+	for (i = 0; i < SI_CE_UCODE_SIZE; i++)
+		WREG32(CP_CE_UCODE_DATA, be32_to_cpup(fw_data++));
+	WREG32(CP_CE_UCODE_ADDR, 0);
+
+	/* ME */
+	fw_data = (const __be32 *)rdev->me_fw->data;
+	WREG32(CP_ME_RAM_WADDR, 0);
+	for (i = 0; i < SI_PM4_UCODE_SIZE; i++)
+		WREG32(CP_ME_RAM_DATA, be32_to_cpup(fw_data++));
+	WREG32(CP_ME_RAM_WADDR, 0);
+
+	WREG32(CP_PFP_UCODE_ADDR, 0);
+	WREG32(CP_CE_UCODE_ADDR, 0);
+	WREG32(CP_ME_RAM_WADDR, 0);
+	WREG32(CP_ME_RAM_RADDR, 0);
+	return 0;
+}
+
+static int si_cp_start(struct radeon_device *rdev)
+{
+	struct radeon_ring *ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX];
+	int r, i;
+
+	r = radeon_ring_lock(rdev, ring, 7 + 4);
+	if (r) {
+		DRM_ERROR("radeon: cp failed to lock ring (%d).\n", r);
+		return r;
+	}
+	/* init the CP */
+	radeon_ring_write(ring, PACKET3(PACKET3_ME_INITIALIZE, 5));
+	radeon_ring_write(ring, 0x1);
+	radeon_ring_write(ring, 0x0);
+	radeon_ring_write(ring, rdev->config.si.max_hw_contexts - 1);
+	radeon_ring_write(ring, PACKET3_ME_INITIALIZE_DEVICE_ID(1));
+	radeon_ring_write(ring, 0);
+	radeon_ring_write(ring, 0);
+
+	/* init the CE partitions */
+	radeon_ring_write(ring, PACKET3(PACKET3_SET_BASE, 2));
+	radeon_ring_write(ring, PACKET3_BASE_INDEX(CE_PARTITION_BASE));
+	radeon_ring_write(ring, 0xc000);
+	radeon_ring_write(ring, 0xe000);
+	radeon_ring_unlock_commit(rdev, ring);
+
+	si_cp_enable(rdev, true);
+
+	r = radeon_ring_lock(rdev, ring, si_default_size + 10);
+	if (r) {
+		DRM_ERROR("radeon: cp failed to lock ring (%d).\n", r);
+		return r;
+	}
+
+	/* setup clear context state */
+	radeon_ring_write(ring, PACKET3(PACKET3_PREAMBLE_CNTL, 0));
+	radeon_ring_write(ring, PACKET3_PREAMBLE_BEGIN_CLEAR_STATE);
+
+	for (i = 0; i < si_default_size; i++)
+		radeon_ring_write(ring, si_default_state[i]);
+
+	radeon_ring_write(ring, PACKET3(PACKET3_PREAMBLE_CNTL, 0));
+	radeon_ring_write(ring, PACKET3_PREAMBLE_END_CLEAR_STATE);
+
+	/* set clear context state */
+	radeon_ring_write(ring, PACKET3(PACKET3_CLEAR_STATE, 0));
+	radeon_ring_write(ring, 0);
+
+	radeon_ring_write(ring, PACKET3(PACKET3_SET_CONTEXT_REG, 2));
+	radeon_ring_write(ring, 0x00000316);
+	radeon_ring_write(ring, 0x0000000e); /* VGT_VERTEX_REUSE_BLOCK_CNTL */
+	radeon_ring_write(ring, 0x00000010); /* VGT_OUT_DEALLOC_CNTL */
+
+	radeon_ring_unlock_commit(rdev, ring);
+
+	for (i = RADEON_RING_TYPE_GFX_INDEX; i <= CAYMAN_RING_TYPE_CP2_INDEX; ++i) {
+		ring = &rdev->ring[i];
+		r = radeon_ring_lock(rdev, ring, 2);
+
+		/* clear the compute context state */
+		radeon_ring_write(ring, PACKET3_COMPUTE(PACKET3_CLEAR_STATE, 0));
+		radeon_ring_write(ring, 0);
+
+		radeon_ring_unlock_commit(rdev, ring);
+	}
+
+	return 0;
+}
+
+static void si_cp_fini(struct radeon_device *rdev)
+{
+	si_cp_enable(rdev, false);
+	radeon_ring_fini(rdev, &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]);
+	radeon_ring_fini(rdev, &rdev->ring[CAYMAN_RING_TYPE_CP1_INDEX]);
+	radeon_ring_fini(rdev, &rdev->ring[CAYMAN_RING_TYPE_CP2_INDEX]);
+}
+
+static int si_cp_resume(struct radeon_device *rdev)
+{
+	struct radeon_ring *ring;
+	u32 tmp;
+	u32 rb_bufsz;
+	int r;
+
+	/* Reset cp; if cp is reset, then PA, SH, VGT also need to be reset */
+	WREG32(GRBM_SOFT_RESET, (SOFT_RESET_CP |
+				 SOFT_RESET_PA |
+				 SOFT_RESET_VGT |
+				 SOFT_RESET_SPI |
+				 SOFT_RESET_SX));
+	RREG32(GRBM_SOFT_RESET);
+	mdelay(15);
+	WREG32(GRBM_SOFT_RESET, 0);
+	RREG32(GRBM_SOFT_RESET);
+
+	WREG32(CP_SEM_WAIT_TIMER, 0x0);
+	WREG32(CP_SEM_INCOMPLETE_TIMER_CNTL, 0x0);
+
+	/* Set the write pointer delay */
+	WREG32(CP_RB_WPTR_DELAY, 0);
+
+	WREG32(CP_DEBUG, 0);
+	WREG32(SCRATCH_ADDR, ((rdev->wb.gpu_addr + RADEON_WB_SCRATCH_OFFSET) >> 8) & 0xFFFFFFFF);
+
+	/* ring 0 - compute and gfx */
+	/* Set ring buffer size */
+	ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX];
+	rb_bufsz = drm_order(ring->ring_size / 8);
+	tmp = (drm_order(RADEON_GPU_PAGE_SIZE/8) << 8) | rb_bufsz;
+#ifdef __BIG_ENDIAN
+	tmp |= BUF_SWAP_32BIT;
+#endif
+	WREG32(CP_RB0_CNTL, tmp);
+
+	/* Initialize the ring buffer's read and write pointers */
+	WREG32(CP_RB0_CNTL, tmp | RB_RPTR_WR_ENA);
+	ring->wptr = 0;
+	WREG32(CP_RB0_WPTR, ring->wptr);
+
+	/* set the wb address wether it's enabled or not */
+	WREG32(CP_RB0_RPTR_ADDR, (rdev->wb.gpu_addr + RADEON_WB_CP_RPTR_OFFSET) & 0xFFFFFFFC);
+	WREG32(CP_RB0_RPTR_ADDR_HI, upper_32_bits(rdev->wb.gpu_addr + RADEON_WB_CP_RPTR_OFFSET) & 0xFF);
+
+	if (rdev->wb.enabled)
+		WREG32(SCRATCH_UMSK, 0xff);
+	else {
+		tmp |= RB_NO_UPDATE;
+		WREG32(SCRATCH_UMSK, 0);
+	}
+
+	mdelay(1);
+	WREG32(CP_RB0_CNTL, tmp);
+
+	WREG32(CP_RB0_BASE, ring->gpu_addr >> 8);
+
+	ring->rptr = RREG32(CP_RB0_RPTR);
+
+	/* ring1  - compute only */
+	/* Set ring buffer size */
+	ring = &rdev->ring[CAYMAN_RING_TYPE_CP1_INDEX];
+	rb_bufsz = drm_order(ring->ring_size / 8);
+	tmp = (drm_order(RADEON_GPU_PAGE_SIZE/8) << 8) | rb_bufsz;
+#ifdef __BIG_ENDIAN
+	tmp |= BUF_SWAP_32BIT;
+#endif
+	WREG32(CP_RB1_CNTL, tmp);
+
+	/* Initialize the ring buffer's read and write pointers */
+	WREG32(CP_RB1_CNTL, tmp | RB_RPTR_WR_ENA);
+	ring->wptr = 0;
+	WREG32(CP_RB1_WPTR, ring->wptr);
+
+	/* set the wb address wether it's enabled or not */
+	WREG32(CP_RB1_RPTR_ADDR, (rdev->wb.gpu_addr + RADEON_WB_CP1_RPTR_OFFSET) & 0xFFFFFFFC);
+	WREG32(CP_RB1_RPTR_ADDR_HI, upper_32_bits(rdev->wb.gpu_addr + RADEON_WB_CP1_RPTR_OFFSET) & 0xFF);
+
+	mdelay(1);
+	WREG32(CP_RB1_CNTL, tmp);
+
+	WREG32(CP_RB1_BASE, ring->gpu_addr >> 8);
+
+	ring->rptr = RREG32(CP_RB1_RPTR);
+
+	/* ring2 - compute only */
+	/* Set ring buffer size */
+	ring = &rdev->ring[CAYMAN_RING_TYPE_CP2_INDEX];
+	rb_bufsz = drm_order(ring->ring_size / 8);
+	tmp = (drm_order(RADEON_GPU_PAGE_SIZE/8) << 8) | rb_bufsz;
+#ifdef __BIG_ENDIAN
+	tmp |= BUF_SWAP_32BIT;
+#endif
+	WREG32(CP_RB2_CNTL, tmp);
+
+	/* Initialize the ring buffer's read and write pointers */
+	WREG32(CP_RB2_CNTL, tmp | RB_RPTR_WR_ENA);
+	ring->wptr = 0;
+	WREG32(CP_RB2_WPTR, ring->wptr);
+
+	/* set the wb address wether it's enabled or not */
+	WREG32(CP_RB2_RPTR_ADDR, (rdev->wb.gpu_addr + RADEON_WB_CP2_RPTR_OFFSET) & 0xFFFFFFFC);
+	WREG32(CP_RB2_RPTR_ADDR_HI, upper_32_bits(rdev->wb.gpu_addr + RADEON_WB_CP2_RPTR_OFFSET) & 0xFF);
+
+	mdelay(1);
+	WREG32(CP_RB2_CNTL, tmp);
+
+	WREG32(CP_RB2_BASE, ring->gpu_addr >> 8);
+
+	ring->rptr = RREG32(CP_RB2_RPTR);
+
+	/* start the rings */
+	si_cp_start(rdev);
+	rdev->ring[RADEON_RING_TYPE_GFX_INDEX].ready = true;
+	rdev->ring[CAYMAN_RING_TYPE_CP1_INDEX].ready = true;
+	rdev->ring[CAYMAN_RING_TYPE_CP2_INDEX].ready = true;
+	r = radeon_ring_test(rdev, RADEON_RING_TYPE_GFX_INDEX, &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]);
+	if (r) {
+		rdev->ring[RADEON_RING_TYPE_GFX_INDEX].ready = false;
+		rdev->ring[CAYMAN_RING_TYPE_CP1_INDEX].ready = false;
+		rdev->ring[CAYMAN_RING_TYPE_CP2_INDEX].ready = false;
+		return r;
+	}
+	r = radeon_ring_test(rdev, CAYMAN_RING_TYPE_CP1_INDEX, &rdev->ring[CAYMAN_RING_TYPE_CP1_INDEX]);
+	if (r) {
+		rdev->ring[CAYMAN_RING_TYPE_CP1_INDEX].ready = false;
+	}
+	r = radeon_ring_test(rdev, CAYMAN_RING_TYPE_CP2_INDEX, &rdev->ring[CAYMAN_RING_TYPE_CP2_INDEX]);
+	if (r) {
+		rdev->ring[CAYMAN_RING_TYPE_CP2_INDEX].ready = false;
+	}
+
+	return 0;
+}
+
+bool si_gpu_is_lockup(struct radeon_device *rdev, struct radeon_ring *ring)
+{
+	u32 srbm_status;
+	u32 grbm_status, grbm_status2;
+	u32 grbm_status_se0, grbm_status_se1;
+	struct r100_gpu_lockup *lockup = &rdev->config.si.lockup;
+	int r;
+
+	srbm_status = RREG32(SRBM_STATUS);
+	grbm_status = RREG32(GRBM_STATUS);
+	grbm_status2 = RREG32(GRBM_STATUS2);
+	grbm_status_se0 = RREG32(GRBM_STATUS_SE0);
+	grbm_status_se1 = RREG32(GRBM_STATUS_SE1);
+	if (!(grbm_status & GUI_ACTIVE)) {
+		r100_gpu_lockup_update(lockup, ring);
+		return false;
+	}
+	/* force CP activities */
+	r = radeon_ring_lock(rdev, ring, 2);
+	if (!r) {
+		/* PACKET2 NOP */
+		radeon_ring_write(ring, 0x80000000);
+		radeon_ring_write(ring, 0x80000000);
+		radeon_ring_unlock_commit(rdev, ring);
+	}
+	/* XXX deal with CP0,1,2 */
+	ring->rptr = RREG32(ring->rptr_reg);
+	return r100_gpu_cp_is_lockup(rdev, lockup, ring);
+}
+
+static int si_gpu_soft_reset(struct radeon_device *rdev)
+{
+	struct evergreen_mc_save save;
+	u32 grbm_reset = 0;
+
+	if (!(RREG32(GRBM_STATUS) & GUI_ACTIVE))
+		return 0;
+
+	dev_info(rdev->dev, "GPU softreset \n");
+	dev_info(rdev->dev, "  GRBM_STATUS=0x%08X\n",
+		RREG32(GRBM_STATUS));
+	dev_info(rdev->dev, "  GRBM_STATUS2=0x%08X\n",
+		RREG32(GRBM_STATUS2));
+	dev_info(rdev->dev, "  GRBM_STATUS_SE0=0x%08X\n",
+		RREG32(GRBM_STATUS_SE0));
+	dev_info(rdev->dev, "  GRBM_STATUS_SE1=0x%08X\n",
+		RREG32(GRBM_STATUS_SE1));
+	dev_info(rdev->dev, "  SRBM_STATUS=0x%08X\n",
+		RREG32(SRBM_STATUS));
+	evergreen_mc_stop(rdev, &save);
+	if (radeon_mc_wait_for_idle(rdev)) {
+		dev_warn(rdev->dev, "Wait for MC idle timedout !\n");
+	}
+	/* Disable CP parsing/prefetching */
+	WREG32(CP_ME_CNTL, CP_ME_HALT | CP_PFP_HALT | CP_CE_HALT);
+
+	/* reset all the gfx blocks */
+	grbm_reset = (SOFT_RESET_CP |
+		      SOFT_RESET_CB |
+		      SOFT_RESET_DB |
+		      SOFT_RESET_GDS |
+		      SOFT_RESET_PA |
+		      SOFT_RESET_SC |
+		      SOFT_RESET_SPI |
+		      SOFT_RESET_SX |
+		      SOFT_RESET_TC |
+		      SOFT_RESET_TA |
+		      SOFT_RESET_VGT |
+		      SOFT_RESET_IA);
+
+	dev_info(rdev->dev, "  GRBM_SOFT_RESET=0x%08X\n", grbm_reset);
+	WREG32(GRBM_SOFT_RESET, grbm_reset);
+	(void)RREG32(GRBM_SOFT_RESET);
+	udelay(50);
+	WREG32(GRBM_SOFT_RESET, 0);
+	(void)RREG32(GRBM_SOFT_RESET);
+	/* Wait a little for things to settle down */
+	udelay(50);
+	dev_info(rdev->dev, "  GRBM_STATUS=0x%08X\n",
+		RREG32(GRBM_STATUS));
+	dev_info(rdev->dev, "  GRBM_STATUS2=0x%08X\n",
+		RREG32(GRBM_STATUS2));
+	dev_info(rdev->dev, "  GRBM_STATUS_SE0=0x%08X\n",
+		RREG32(GRBM_STATUS_SE0));
+	dev_info(rdev->dev, "  GRBM_STATUS_SE1=0x%08X\n",
+		RREG32(GRBM_STATUS_SE1));
+	dev_info(rdev->dev, "  SRBM_STATUS=0x%08X\n",
+		RREG32(SRBM_STATUS));
+	evergreen_mc_resume(rdev, &save);
+	return 0;
+}
+
+int si_asic_reset(struct radeon_device *rdev)
+{
+	return si_gpu_soft_reset(rdev);
+}
+
+/* MC */
+static void si_mc_program(struct radeon_device *rdev)
+{
+	struct evergreen_mc_save save;
+	u32 tmp;
+	int i, j;
+
+	/* Initialize HDP */
+	for (i = 0, j = 0; i < 32; i++, j += 0x18) {
+		WREG32((0x2c14 + j), 0x00000000);
+		WREG32((0x2c18 + j), 0x00000000);
+		WREG32((0x2c1c + j), 0x00000000);
+		WREG32((0x2c20 + j), 0x00000000);
+		WREG32((0x2c24 + j), 0x00000000);
+	}
+	WREG32(HDP_REG_COHERENCY_FLUSH_CNTL, 0);
+
+	evergreen_mc_stop(rdev, &save);
+	if (radeon_mc_wait_for_idle(rdev)) {
+		dev_warn(rdev->dev, "Wait for MC idle timedout !\n");
+	}
+	/* Lockout access through VGA aperture*/
+	WREG32(VGA_HDP_CONTROL, VGA_MEMORY_DISABLE);
+	/* Update configuration */
+	WREG32(MC_VM_SYSTEM_APERTURE_LOW_ADDR,
+	       rdev->mc.vram_start >> 12);
+	WREG32(MC_VM_SYSTEM_APERTURE_HIGH_ADDR,
+	       rdev->mc.vram_end >> 12);
+	WREG32(MC_VM_SYSTEM_APERTURE_DEFAULT_ADDR,
+	       rdev->vram_scratch.gpu_addr >> 12);
+	tmp = ((rdev->mc.vram_end >> 24) & 0xFFFF) << 16;
+	tmp |= ((rdev->mc.vram_start >> 24) & 0xFFFF);
+	WREG32(MC_VM_FB_LOCATION, tmp);
+	/* XXX double check these! */
+	WREG32(HDP_NONSURFACE_BASE, (rdev->mc.vram_start >> 8));
+	WREG32(HDP_NONSURFACE_INFO, (2 << 7) | (1 << 30));
+	WREG32(HDP_NONSURFACE_SIZE, 0x3FFFFFFF);
+	WREG32(MC_VM_AGP_BASE, 0);
+	WREG32(MC_VM_AGP_TOP, 0x0FFFFFFF);
+	WREG32(MC_VM_AGP_BOT, 0x0FFFFFFF);
+	if (radeon_mc_wait_for_idle(rdev)) {
+		dev_warn(rdev->dev, "Wait for MC idle timedout !\n");
+	}
+	evergreen_mc_resume(rdev, &save);
+	/* we need to own VRAM, so turn off the VGA renderer here
+	 * to stop it overwriting our objects */
+	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)
+{
+	if (mc->mc_vram_size > 0xFFC0000000ULL) {
+		/* leave room for at least 1024M GTT */
+		dev_warn(rdev->dev, "limiting VRAM\n");
+		mc->real_vram_size = 0xFFC0000000ULL;
+		mc->mc_vram_size = 0xFFC0000000ULL;
+	}
+	si_vram_location(rdev, &rdev->mc, 0);
+	rdev->mc.gtt_base_align = 0;
+	si_gtt_location(rdev, mc);
+}
+
+static int si_mc_init(struct radeon_device *rdev)
+{
+	u32 tmp;
+	int chansize, numchan;
+
+	/* Get VRAM informations */
+	rdev->mc.vram_is_ddr = true;
+	tmp = RREG32(MC_ARB_RAMCFG);
+	if (tmp & CHANSIZE_OVERRIDE) {
+		chansize = 16;
+	} else if (tmp & CHANSIZE_MASK) {
+		chansize = 64;
+	} else {
+		chansize = 32;
+	}
+	tmp = RREG32(MC_SHARED_CHMAP);
+	switch ((tmp & NOOFCHAN_MASK) >> NOOFCHAN_SHIFT) {
+	case 0:
+	default:
+		numchan = 1;
+		break;
+	case 1:
+		numchan = 2;
+		break;
+	case 2:
+		numchan = 4;
+		break;
+	case 3:
+		numchan = 8;
+		break;
+	case 4:
+		numchan = 3;
+		break;
+	case 5:
+		numchan = 6;
+		break;
+	case 6:
+		numchan = 10;
+		break;
+	case 7:
+		numchan = 12;
+		break;
+	case 8:
+		numchan = 16;
+		break;
+	}
+	rdev->mc.vram_width = numchan * chansize;
+	/* Could aper size report 0 ? */
+	rdev->mc.aper_base = pci_resource_start(rdev->pdev, 0);
+	rdev->mc.aper_size = pci_resource_len(rdev->pdev, 0);
+	/* size in MB on si */
+	rdev->mc.mc_vram_size = RREG32(CONFIG_MEMSIZE) * 1024 * 1024;
+	rdev->mc.real_vram_size = RREG32(CONFIG_MEMSIZE) * 1024 * 1024;
+	rdev->mc.visible_vram_size = rdev->mc.aper_size;
+	si_vram_gtt_location(rdev, &rdev->mc);
+	radeon_update_bandwidth_info(rdev);
+
+	return 0;
+}
+
+/*
+ * GART
+ */
+void si_pcie_gart_tlb_flush(struct radeon_device *rdev)
+{
+	/* flush hdp cache */
+	WREG32(HDP_MEM_COHERENCY_FLUSH_CNTL, 0x1);
+
+	/* bits 0-15 are the VM contexts0-15 */
+	WREG32(VM_INVALIDATE_REQUEST, 1);
+}
+
+int si_pcie_gart_enable(struct radeon_device *rdev)
+{
+	int r, i;
+
+	if (rdev->gart.robj == NULL) {
+		dev_err(rdev->dev, "No VRAM object for PCIE GART.\n");
+		return -EINVAL;
+	}
+	r = radeon_gart_table_vram_pin(rdev);
+	if (r)
+		return r;
+	radeon_gart_restore(rdev);
+	/* Setup TLB control */
+	WREG32(MC_VM_MX_L1_TLB_CNTL,
+	       (0xA << 7) |
+	       ENABLE_L1_TLB |
+	       SYSTEM_ACCESS_MODE_NOT_IN_SYS |
+	       ENABLE_ADVANCED_DRIVER_MODEL |
+	       SYSTEM_APERTURE_UNMAPPED_ACCESS_PASS_THRU);
+	/* Setup L2 cache */
+	WREG32(VM_L2_CNTL, ENABLE_L2_CACHE |
+	       ENABLE_L2_PTE_CACHE_LRU_UPDATE_BY_WRITE |
+	       ENABLE_L2_PDE0_CACHE_LRU_UPDATE_BY_WRITE |
+	       EFFECTIVE_L2_QUEUE_SIZE(7) |
+	       CONTEXT1_IDENTITY_ACCESS_MODE(1));
+	WREG32(VM_L2_CNTL2, INVALIDATE_ALL_L1_TLBS | INVALIDATE_L2_CACHE);
+	WREG32(VM_L2_CNTL3, L2_CACHE_BIGK_ASSOCIATIVITY |
+	       L2_CACHE_BIGK_FRAGMENT_SIZE(0));
+	/* setup context0 */
+	WREG32(VM_CONTEXT0_PAGE_TABLE_START_ADDR, rdev->mc.gtt_start >> 12);
+	WREG32(VM_CONTEXT0_PAGE_TABLE_END_ADDR, rdev->mc.gtt_end >> 12);
+	WREG32(VM_CONTEXT0_PAGE_TABLE_BASE_ADDR, rdev->gart.table_addr >> 12);
+	WREG32(VM_CONTEXT0_PROTECTION_FAULT_DEFAULT_ADDR,
+			(u32)(rdev->dummy_page.addr >> 12));
+	WREG32(VM_CONTEXT0_CNTL2, 0);
+	WREG32(VM_CONTEXT0_CNTL, (ENABLE_CONTEXT | PAGE_TABLE_DEPTH(0) |
+				  RANGE_PROTECTION_FAULT_ENABLE_DEFAULT));
+
+	WREG32(0x15D4, 0);
+	WREG32(0x15D8, 0);
+	WREG32(0x15DC, 0);
+
+	/* empty context1-15 */
+	/* FIXME start with 1G, once using 2 level pt switch to full
+	 * vm size space
+	 */
+	/* set vm size, must be a multiple of 4 */
+	WREG32(VM_CONTEXT1_PAGE_TABLE_START_ADDR, 0);
+	WREG32(VM_CONTEXT1_PAGE_TABLE_END_ADDR, (1 << 30) / RADEON_GPU_PAGE_SIZE);
+	for (i = 1; i < 16; i++) {
+		if (i < 8)
+			WREG32(VM_CONTEXT0_PAGE_TABLE_BASE_ADDR + (i << 2),
+			       rdev->gart.table_addr >> 12);
+		else
+			WREG32(VM_CONTEXT8_PAGE_TABLE_BASE_ADDR + ((i - 8) << 2),
+			       rdev->gart.table_addr >> 12);
+	}
+
+	/* enable context1-15 */
+	WREG32(VM_CONTEXT1_PROTECTION_FAULT_DEFAULT_ADDR,
+	       (u32)(rdev->dummy_page.addr >> 12));
+	WREG32(VM_CONTEXT1_CNTL2, 0);
+	WREG32(VM_CONTEXT1_CNTL, ENABLE_CONTEXT | PAGE_TABLE_DEPTH(0) |
+				RANGE_PROTECTION_FAULT_ENABLE_DEFAULT);
+
+	si_pcie_gart_tlb_flush(rdev);
+	DRM_INFO("PCIE GART of %uM enabled (table at 0x%016llX).\n",
+		 (unsigned)(rdev->mc.gtt_size >> 20),
+		 (unsigned long long)rdev->gart.table_addr);
+	rdev->gart.ready = true;
+	return 0;
+}
+
+void si_pcie_gart_disable(struct radeon_device *rdev)
+{
+	/* Disable all tables */
+	WREG32(VM_CONTEXT0_CNTL, 0);
+	WREG32(VM_CONTEXT1_CNTL, 0);
+	/* Setup TLB control */
+	WREG32(MC_VM_MX_L1_TLB_CNTL, SYSTEM_ACCESS_MODE_NOT_IN_SYS |
+	       SYSTEM_APERTURE_UNMAPPED_ACCESS_PASS_THRU);
+	/* Setup L2 cache */
+	WREG32(VM_L2_CNTL, ENABLE_L2_PTE_CACHE_LRU_UPDATE_BY_WRITE |
+	       ENABLE_L2_PDE0_CACHE_LRU_UPDATE_BY_WRITE |
+	       EFFECTIVE_L2_QUEUE_SIZE(7) |
+	       CONTEXT1_IDENTITY_ACCESS_MODE(1));
+	WREG32(VM_L2_CNTL2, 0);
+	WREG32(VM_L2_CNTL3, L2_CACHE_BIGK_ASSOCIATIVITY |
+	       L2_CACHE_BIGK_FRAGMENT_SIZE(0));
+	radeon_gart_table_vram_unpin(rdev);
+}
+
+void si_pcie_gart_fini(struct radeon_device *rdev)
+{
+	si_pcie_gart_disable(rdev);
+	radeon_gart_table_vram_free(rdev);
+	radeon_gart_fini(rdev);
+}
+
+/* vm parser */
+static bool si_vm_reg_valid(u32 reg)
+{
+	/* context regs are fine */
+	if (reg >= 0x28000)
+		return true;
+
+	/* check config regs */
+	switch (reg) {
+	case GRBM_GFX_INDEX:
+	case VGT_VTX_VECT_EJECT_REG:
+	case VGT_CACHE_INVALIDATION:
+	case VGT_ESGS_RING_SIZE:
+	case VGT_GSVS_RING_SIZE:
+	case VGT_GS_VERTEX_REUSE:
+	case VGT_PRIMITIVE_TYPE:
+	case VGT_INDEX_TYPE:
+	case VGT_NUM_INDICES:
+	case VGT_NUM_INSTANCES:
+	case VGT_TF_RING_SIZE:
+	case VGT_HS_OFFCHIP_PARAM:
+	case VGT_TF_MEMORY_BASE:
+	case PA_CL_ENHANCE:
+	case PA_SU_LINE_STIPPLE_VALUE:
+	case PA_SC_LINE_STIPPLE_STATE:
+	case PA_SC_ENHANCE:
+	case SQC_CACHES:
+	case SPI_STATIC_THREAD_MGMT_1:
+	case SPI_STATIC_THREAD_MGMT_2:
+	case SPI_STATIC_THREAD_MGMT_3:
+	case SPI_PS_MAX_WAVE_ID:
+	case SPI_CONFIG_CNTL:
+	case SPI_CONFIG_CNTL_1:
+	case TA_CNTL_AUX:
+		return true;
+	default:
+		DRM_ERROR("Invalid register 0x%x in CS\n", reg);
+		return false;
+	}
+}
+
+static int si_vm_packet3_ce_check(struct radeon_device *rdev,
+				  u32 *ib, struct radeon_cs_packet *pkt)
+{
+	switch (pkt->opcode) {
+	case PACKET3_NOP:
+	case PACKET3_SET_BASE:
+	case PACKET3_SET_CE_DE_COUNTERS:
+	case PACKET3_LOAD_CONST_RAM:
+	case PACKET3_WRITE_CONST_RAM:
+	case PACKET3_WRITE_CONST_RAM_OFFSET:
+	case PACKET3_DUMP_CONST_RAM:
+	case PACKET3_INCREMENT_CE_COUNTER:
+	case PACKET3_WAIT_ON_DE_COUNTER:
+	case PACKET3_CE_WRITE:
+		break;
+	default:
+		DRM_ERROR("Invalid CE packet3: 0x%x\n", pkt->opcode);
+		return -EINVAL;
+	}
+	return 0;
+}
+
+static int si_vm_packet3_gfx_check(struct radeon_device *rdev,
+				   u32 *ib, struct radeon_cs_packet *pkt)
+{
+	u32 idx = pkt->idx + 1;
+	u32 idx_value = ib[idx];
+	u32 start_reg, end_reg, reg, i;
+
+	switch (pkt->opcode) {
+	case PACKET3_NOP:
+	case PACKET3_SET_BASE:
+	case PACKET3_CLEAR_STATE:
+	case PACKET3_INDEX_BUFFER_SIZE:
+	case PACKET3_DISPATCH_DIRECT:
+	case PACKET3_DISPATCH_INDIRECT:
+	case PACKET3_ALLOC_GDS:
+	case PACKET3_WRITE_GDS_RAM:
+	case PACKET3_ATOMIC_GDS:
+	case PACKET3_ATOMIC:
+	case PACKET3_OCCLUSION_QUERY:
+	case PACKET3_SET_PREDICATION:
+	case PACKET3_COND_EXEC:
+	case PACKET3_PRED_EXEC:
+	case PACKET3_DRAW_INDIRECT:
+	case PACKET3_DRAW_INDEX_INDIRECT:
+	case PACKET3_INDEX_BASE:
+	case PACKET3_DRAW_INDEX_2:
+	case PACKET3_CONTEXT_CONTROL:
+	case PACKET3_INDEX_TYPE:
+	case PACKET3_DRAW_INDIRECT_MULTI:
+	case PACKET3_DRAW_INDEX_AUTO:
+	case PACKET3_DRAW_INDEX_IMMD:
+	case PACKET3_NUM_INSTANCES:
+	case PACKET3_DRAW_INDEX_MULTI_AUTO:
+	case PACKET3_STRMOUT_BUFFER_UPDATE:
+	case PACKET3_DRAW_INDEX_OFFSET_2:
+	case PACKET3_DRAW_INDEX_MULTI_ELEMENT:
+	case PACKET3_DRAW_INDEX_INDIRECT_MULTI:
+	case PACKET3_MPEG_INDEX:
+	case PACKET3_WAIT_REG_MEM:
+	case PACKET3_MEM_WRITE:
+	case PACKET3_PFP_SYNC_ME:
+	case PACKET3_SURFACE_SYNC:
+	case PACKET3_EVENT_WRITE:
+	case PACKET3_EVENT_WRITE_EOP:
+	case PACKET3_EVENT_WRITE_EOS:
+	case PACKET3_SET_CONTEXT_REG:
+	case PACKET3_SET_CONTEXT_REG_INDIRECT:
+	case PACKET3_SET_SH_REG:
+	case PACKET3_SET_SH_REG_OFFSET:
+	case PACKET3_INCREMENT_DE_COUNTER:
+	case PACKET3_WAIT_ON_CE_COUNTER:
+	case PACKET3_WAIT_ON_AVAIL_BUFFER:
+	case PACKET3_ME_WRITE:
+		break;
+	case PACKET3_COPY_DATA:
+		if ((idx_value & 0xf00) == 0) {
+			reg = ib[idx + 3] * 4;
+			if (!si_vm_reg_valid(reg))
+				return -EINVAL;
+		}
+		break;
+	case PACKET3_WRITE_DATA:
+		if ((idx_value & 0xf00) == 0) {
+			start_reg = ib[idx + 1] * 4;
+			if (idx_value & 0x10000) {
+				if (!si_vm_reg_valid(start_reg))
+					return -EINVAL;
+			} else {
+				for (i = 0; i < (pkt->count - 2); i++) {
+					reg = start_reg + (4 * i);
+					if (!si_vm_reg_valid(reg))
+						return -EINVAL;
+				}
+			}
+		}
+		break;
+	case PACKET3_COND_WRITE:
+		if (idx_value & 0x100) {
+			reg = ib[idx + 5] * 4;
+			if (!si_vm_reg_valid(reg))
+				return -EINVAL;
+		}
+		break;
+	case PACKET3_COPY_DW:
+		if (idx_value & 0x2) {
+			reg = ib[idx + 3] * 4;
+			if (!si_vm_reg_valid(reg))
+				return -EINVAL;
+		}
+		break;
+	case PACKET3_SET_CONFIG_REG:
+		start_reg = (idx_value << 2) + PACKET3_SET_CONFIG_REG_START;
+		end_reg = 4 * pkt->count + start_reg - 4;
+		if ((start_reg < PACKET3_SET_CONFIG_REG_START) ||
+		    (start_reg >= PACKET3_SET_CONFIG_REG_END) ||
+		    (end_reg >= PACKET3_SET_CONFIG_REG_END)) {
+			DRM_ERROR("bad PACKET3_SET_CONFIG_REG\n");
+			return -EINVAL;
+		}
+		for (i = 0; i < pkt->count; i++) {
+			reg = start_reg + (4 * i);
+			if (!si_vm_reg_valid(reg))
+				return -EINVAL;
+		}
+		break;
+	default:
+		DRM_ERROR("Invalid GFX packet3: 0x%x\n", pkt->opcode);
+		return -EINVAL;
+	}
+	return 0;
+}
+
+static int si_vm_packet3_compute_check(struct radeon_device *rdev,
+				       u32 *ib, struct radeon_cs_packet *pkt)
+{
+	u32 idx = pkt->idx + 1;
+	u32 idx_value = ib[idx];
+	u32 start_reg, reg, i;
+
+	switch (pkt->opcode) {
+	case PACKET3_NOP:
+	case PACKET3_SET_BASE:
+	case PACKET3_CLEAR_STATE:
+	case PACKET3_DISPATCH_DIRECT:
+	case PACKET3_DISPATCH_INDIRECT:
+	case PACKET3_ALLOC_GDS:
+	case PACKET3_WRITE_GDS_RAM:
+	case PACKET3_ATOMIC_GDS:
+	case PACKET3_ATOMIC:
+	case PACKET3_OCCLUSION_QUERY:
+	case PACKET3_SET_PREDICATION:
+	case PACKET3_COND_EXEC:
+	case PACKET3_PRED_EXEC:
+	case PACKET3_CONTEXT_CONTROL:
+	case PACKET3_STRMOUT_BUFFER_UPDATE:
+	case PACKET3_WAIT_REG_MEM:
+	case PACKET3_MEM_WRITE:
+	case PACKET3_PFP_SYNC_ME:
+	case PACKET3_SURFACE_SYNC:
+	case PACKET3_EVENT_WRITE:
+	case PACKET3_EVENT_WRITE_EOP:
+	case PACKET3_EVENT_WRITE_EOS:
+	case PACKET3_SET_CONTEXT_REG:
+	case PACKET3_SET_CONTEXT_REG_INDIRECT:
+	case PACKET3_SET_SH_REG:
+	case PACKET3_SET_SH_REG_OFFSET:
+	case PACKET3_INCREMENT_DE_COUNTER:
+	case PACKET3_WAIT_ON_CE_COUNTER:
+	case PACKET3_WAIT_ON_AVAIL_BUFFER:
+	case PACKET3_ME_WRITE:
+		break;
+	case PACKET3_COPY_DATA:
+		if ((idx_value & 0xf00) == 0) {
+			reg = ib[idx + 3] * 4;
+			if (!si_vm_reg_valid(reg))
+				return -EINVAL;
+		}
+		break;
+	case PACKET3_WRITE_DATA:
+		if ((idx_value & 0xf00) == 0) {
+			start_reg = ib[idx + 1] * 4;
+			if (idx_value & 0x10000) {
+				if (!si_vm_reg_valid(start_reg))
+					return -EINVAL;
+			} else {
+				for (i = 0; i < (pkt->count - 2); i++) {
+					reg = start_reg + (4 * i);
+					if (!si_vm_reg_valid(reg))
+						return -EINVAL;
+				}
+			}
+		}
+		break;
+	case PACKET3_COND_WRITE:
+		if (idx_value & 0x100) {
+			reg = ib[idx + 5] * 4;
+			if (!si_vm_reg_valid(reg))
+				return -EINVAL;
+		}
+		break;
+	case PACKET3_COPY_DW:
+		if (idx_value & 0x2) {
+			reg = ib[idx + 3] * 4;
+			if (!si_vm_reg_valid(reg))
+				return -EINVAL;
+		}
+		break;
+	default:
+		DRM_ERROR("Invalid Compute packet3: 0x%x\n", pkt->opcode);
+		return -EINVAL;
+	}
+	return 0;
+}
+
+int si_ib_parse(struct radeon_device *rdev, struct radeon_ib *ib)
+{
+	int ret = 0;
+	u32 idx = 0;
+	struct radeon_cs_packet pkt;
+
+	do {
+		pkt.idx = idx;
+		pkt.type = CP_PACKET_GET_TYPE(ib->ptr[idx]);
+		pkt.count = CP_PACKET_GET_COUNT(ib->ptr[idx]);
+		pkt.one_reg_wr = 0;
+		switch (pkt.type) {
+		case PACKET_TYPE0:
+			dev_err(rdev->dev, "Packet0 not allowed!\n");
+			ret = -EINVAL;
+			break;
+		case PACKET_TYPE2:
+			idx += 1;
+			break;
+		case PACKET_TYPE3:
+			pkt.opcode = CP_PACKET3_GET_OPCODE(ib->ptr[idx]);
+			if (ib->is_const_ib)
+				ret = si_vm_packet3_ce_check(rdev, ib->ptr, &pkt);
+			else {
+				switch (ib->fence->ring) {
+				case RADEON_RING_TYPE_GFX_INDEX:
+					ret = si_vm_packet3_gfx_check(rdev, ib->ptr, &pkt);
+					break;
+				case CAYMAN_RING_TYPE_CP1_INDEX:
+				case CAYMAN_RING_TYPE_CP2_INDEX:
+					ret = si_vm_packet3_compute_check(rdev, ib->ptr, &pkt);
+					break;
+				default:
+					dev_err(rdev->dev, "Non-PM4 ring %d !\n", ib->fence->ring);
+					ret = -EINVAL;
+					break;
+				}
+			}
+			idx += pkt.count + 2;
+			break;
+		default:
+			dev_err(rdev->dev, "Unknown packet type %d !\n", pkt.type);
+			ret = -EINVAL;
+			break;
+		}
+		if (ret)
+			break;
+	} while (idx < ib->length_dw);
+
+	return ret;
+}
+
+/*
+ * vm
+ */
+int si_vm_init(struct radeon_device *rdev)
+{
+	/* number of VMs */
+	rdev->vm_manager.nvm = 16;
+	/* base offset of vram pages */
+	rdev->vm_manager.vram_base_offset = 0;
+
+	return 0;
+}
+
+void si_vm_fini(struct radeon_device *rdev)
+{
+}
+
+int si_vm_bind(struct radeon_device *rdev, struct radeon_vm *vm, int id)
+{
+	if (id < 8)
+		WREG32(VM_CONTEXT0_PAGE_TABLE_BASE_ADDR + (id << 2), vm->pt_gpu_addr >> 12);
+	else
+		WREG32(VM_CONTEXT8_PAGE_TABLE_BASE_ADDR + ((id - 8) << 2),
+		       vm->pt_gpu_addr >> 12);
+	/* flush hdp cache */
+	WREG32(HDP_MEM_COHERENCY_FLUSH_CNTL, 0x1);
+	/* bits 0-15 are the VM contexts0-15 */
+	WREG32(VM_INVALIDATE_REQUEST, 1 << id);
+	return 0;
+}
+
+void si_vm_unbind(struct radeon_device *rdev, struct radeon_vm *vm)
+{
+	if (vm->id < 8)
+		WREG32(VM_CONTEXT0_PAGE_TABLE_BASE_ADDR + (vm->id << 2), 0);
+	else
+		WREG32(VM_CONTEXT8_PAGE_TABLE_BASE_ADDR + ((vm->id - 8) << 2), 0);
+	/* flush hdp cache */
+	WREG32(HDP_MEM_COHERENCY_FLUSH_CNTL, 0x1);
+	/* bits 0-15 are the VM contexts0-15 */
+	WREG32(VM_INVALIDATE_REQUEST, 1 << vm->id);
+}
+
+void si_vm_tlb_flush(struct radeon_device *rdev, struct radeon_vm *vm)
+{
+	if (vm->id == -1)
+		return;
+
+	/* flush hdp cache */
+	WREG32(HDP_MEM_COHERENCY_FLUSH_CNTL, 0x1);
+	/* bits 0-15 are the VM contexts0-15 */
+	WREG32(VM_INVALIDATE_REQUEST, 1 << vm->id);
+}
+
+/*
+ * RLC
+ */
+void si_rlc_fini(struct radeon_device *rdev)
+{
+	int r;
+
+	/* save restore block */
+	if (rdev->rlc.save_restore_obj) {
+		r = radeon_bo_reserve(rdev->rlc.save_restore_obj, false);
+		if (unlikely(r != 0))
+			dev_warn(rdev->dev, "(%d) reserve RLC sr bo failed\n", r);
+		radeon_bo_unpin(rdev->rlc.save_restore_obj);
+		radeon_bo_unreserve(rdev->rlc.save_restore_obj);
+
+		radeon_bo_unref(&rdev->rlc.save_restore_obj);
+		rdev->rlc.save_restore_obj = NULL;
+	}
+
+	/* clear state block */
+	if (rdev->rlc.clear_state_obj) {
+		r = radeon_bo_reserve(rdev->rlc.clear_state_obj, false);
+		if (unlikely(r != 0))
+			dev_warn(rdev->dev, "(%d) reserve RLC c bo failed\n", r);
+		radeon_bo_unpin(rdev->rlc.clear_state_obj);
+		radeon_bo_unreserve(rdev->rlc.clear_state_obj);
+
+		radeon_bo_unref(&rdev->rlc.clear_state_obj);
+		rdev->rlc.clear_state_obj = NULL;
+	}
+}
+
+int si_rlc_init(struct radeon_device *rdev)
+{
+	int r;
+
+	/* save restore block */
+	if (rdev->rlc.save_restore_obj == NULL) {
+		r = radeon_bo_create(rdev, RADEON_GPU_PAGE_SIZE, PAGE_SIZE, true,
+				RADEON_GEM_DOMAIN_VRAM, &rdev->rlc.save_restore_obj);
+		if (r) {
+			dev_warn(rdev->dev, "(%d) create RLC sr bo failed\n", r);
+			return r;
+		}
+	}
+
+	r = radeon_bo_reserve(rdev->rlc.save_restore_obj, false);
+	if (unlikely(r != 0)) {
+		si_rlc_fini(rdev);
+		return r;
+	}
+	r = radeon_bo_pin(rdev->rlc.save_restore_obj, RADEON_GEM_DOMAIN_VRAM,
+			  &rdev->rlc.save_restore_gpu_addr);
+	if (r) {
+		radeon_bo_unreserve(rdev->rlc.save_restore_obj);
+		dev_warn(rdev->dev, "(%d) pin RLC sr bo failed\n", r);
+		si_rlc_fini(rdev);
+		return r;
+	}
+
+	/* clear state block */
+	if (rdev->rlc.clear_state_obj == NULL) {
+		r = radeon_bo_create(rdev, RADEON_GPU_PAGE_SIZE, PAGE_SIZE, true,
+				RADEON_GEM_DOMAIN_VRAM, &rdev->rlc.clear_state_obj);
+		if (r) {
+			dev_warn(rdev->dev, "(%d) create RLC c bo failed\n", r);
+			si_rlc_fini(rdev);
+			return r;
+		}
+	}
+	r = radeon_bo_reserve(rdev->rlc.clear_state_obj, false);
+	if (unlikely(r != 0)) {
+		si_rlc_fini(rdev);
+		return r;
+	}
+	r = radeon_bo_pin(rdev->rlc.clear_state_obj, RADEON_GEM_DOMAIN_VRAM,
+			  &rdev->rlc.clear_state_gpu_addr);
+	if (r) {
+
+		radeon_bo_unreserve(rdev->rlc.clear_state_obj);
+		dev_warn(rdev->dev, "(%d) pin RLC c bo failed\n", r);
+		si_rlc_fini(rdev);
+		return r;
+	}
+
+	return 0;
+}
+
+static void si_rlc_stop(struct radeon_device *rdev)
+{
+	WREG32(RLC_CNTL, 0);
+}
+
+static void si_rlc_start(struct radeon_device *rdev)
+{
+	WREG32(RLC_CNTL, RLC_ENABLE);
+}
+
+static int si_rlc_resume(struct radeon_device *rdev)
+{
+	u32 i;
+	const __be32 *fw_data;
+
+	if (!rdev->rlc_fw)
+		return -EINVAL;
+
+	si_rlc_stop(rdev);
+
+	WREG32(RLC_RL_BASE, 0);
+	WREG32(RLC_RL_SIZE, 0);
+	WREG32(RLC_LB_CNTL, 0);
+	WREG32(RLC_LB_CNTR_MAX, 0xffffffff);
+	WREG32(RLC_LB_CNTR_INIT, 0);
+
+	WREG32(RLC_SAVE_AND_RESTORE_BASE, rdev->rlc.save_restore_gpu_addr >> 8);
+	WREG32(RLC_CLEAR_STATE_RESTORE_BASE, rdev->rlc.clear_state_gpu_addr >> 8);
+
+	WREG32(RLC_MC_CNTL, 0);
+	WREG32(RLC_UCODE_CNTL, 0);
+
+	fw_data = (const __be32 *)rdev->rlc_fw->data;
+	for (i = 0; i < SI_RLC_UCODE_SIZE; i++) {
+		WREG32(RLC_UCODE_ADDR, i);
+		WREG32(RLC_UCODE_DATA, be32_to_cpup(fw_data++));
+	}
+	WREG32(RLC_UCODE_ADDR, 0);
+
+	si_rlc_start(rdev);
+
+	return 0;
+}
+
+static void si_enable_interrupts(struct radeon_device *rdev)
+{
+	u32 ih_cntl = RREG32(IH_CNTL);
+	u32 ih_rb_cntl = RREG32(IH_RB_CNTL);
+
+	ih_cntl |= ENABLE_INTR;
+	ih_rb_cntl |= IH_RB_ENABLE;
+	WREG32(IH_CNTL, ih_cntl);
+	WREG32(IH_RB_CNTL, ih_rb_cntl);
+	rdev->ih.enabled = true;
+}
+
+static void si_disable_interrupts(struct radeon_device *rdev)
+{
+	u32 ih_rb_cntl = RREG32(IH_RB_CNTL);
+	u32 ih_cntl = RREG32(IH_CNTL);
+
+	ih_rb_cntl &= ~IH_RB_ENABLE;
+	ih_cntl &= ~ENABLE_INTR;
+	WREG32(IH_RB_CNTL, ih_rb_cntl);
+	WREG32(IH_CNTL, ih_cntl);
+	/* set rptr, wptr to 0 */
+	WREG32(IH_RB_RPTR, 0);
+	WREG32(IH_RB_WPTR, 0);
+	rdev->ih.enabled = false;
+	rdev->ih.wptr = 0;
+	rdev->ih.rptr = 0;
+}
+
+static void si_disable_interrupt_state(struct radeon_device *rdev)
+{
+	u32 tmp;
+
+	WREG32(CP_INT_CNTL_RING0, CNTX_BUSY_INT_ENABLE | CNTX_EMPTY_INT_ENABLE);
+	WREG32(CP_INT_CNTL_RING1, 0);
+	WREG32(CP_INT_CNTL_RING2, 0);
+	WREG32(GRBM_INT_CNTL, 0);
+	WREG32(INT_MASK + EVERGREEN_CRTC0_REGISTER_OFFSET, 0);
+	WREG32(INT_MASK + EVERGREEN_CRTC1_REGISTER_OFFSET, 0);
+	if (rdev->num_crtc >= 4) {
+		WREG32(INT_MASK + EVERGREEN_CRTC2_REGISTER_OFFSET, 0);
+		WREG32(INT_MASK + EVERGREEN_CRTC3_REGISTER_OFFSET, 0);
+	}
+	if (rdev->num_crtc >= 6) {
+		WREG32(INT_MASK + EVERGREEN_CRTC4_REGISTER_OFFSET, 0);
+		WREG32(INT_MASK + EVERGREEN_CRTC5_REGISTER_OFFSET, 0);
+	}
+
+	WREG32(GRPH_INT_CONTROL + EVERGREEN_CRTC0_REGISTER_OFFSET, 0);
+	WREG32(GRPH_INT_CONTROL + EVERGREEN_CRTC1_REGISTER_OFFSET, 0);
+	if (rdev->num_crtc >= 4) {
+		WREG32(GRPH_INT_CONTROL + EVERGREEN_CRTC2_REGISTER_OFFSET, 0);
+		WREG32(GRPH_INT_CONTROL + EVERGREEN_CRTC3_REGISTER_OFFSET, 0);
+	}
+	if (rdev->num_crtc >= 6) {
+		WREG32(GRPH_INT_CONTROL + EVERGREEN_CRTC4_REGISTER_OFFSET, 0);
+		WREG32(GRPH_INT_CONTROL + EVERGREEN_CRTC5_REGISTER_OFFSET, 0);
+	}
+
+	WREG32(DACA_AUTODETECT_INT_CONTROL, 0);
+
+	tmp = RREG32(DC_HPD1_INT_CONTROL) & DC_HPDx_INT_POLARITY;
+	WREG32(DC_HPD1_INT_CONTROL, tmp);
+	tmp = RREG32(DC_HPD2_INT_CONTROL) & DC_HPDx_INT_POLARITY;
+	WREG32(DC_HPD2_INT_CONTROL, tmp);
+	tmp = RREG32(DC_HPD3_INT_CONTROL) & DC_HPDx_INT_POLARITY;
+	WREG32(DC_HPD3_INT_CONTROL, tmp);
+	tmp = RREG32(DC_HPD4_INT_CONTROL) & DC_HPDx_INT_POLARITY;
+	WREG32(DC_HPD4_INT_CONTROL, tmp);
+	tmp = RREG32(DC_HPD5_INT_CONTROL) & DC_HPDx_INT_POLARITY;
+	WREG32(DC_HPD5_INT_CONTROL, tmp);
+	tmp = RREG32(DC_HPD6_INT_CONTROL) & DC_HPDx_INT_POLARITY;
+	WREG32(DC_HPD6_INT_CONTROL, tmp);
+
+}
+
+static int si_irq_init(struct radeon_device *rdev)
+{
+	int ret = 0;
+	int rb_bufsz;
+	u32 interrupt_cntl, ih_cntl, ih_rb_cntl;
+
+	/* allocate ring */
+	ret = r600_ih_ring_alloc(rdev);
+	if (ret)
+		return ret;
+
+	/* disable irqs */
+	si_disable_interrupts(rdev);
+
+	/* init rlc */
+	ret = si_rlc_resume(rdev);
+	if (ret) {
+		r600_ih_ring_fini(rdev);
+		return ret;
+	}
+
+	/* setup interrupt control */
+	/* set dummy read address to ring address */
+	WREG32(INTERRUPT_CNTL2, rdev->ih.gpu_addr >> 8);
+	interrupt_cntl = RREG32(INTERRUPT_CNTL);
+	/* IH_DUMMY_RD_OVERRIDE=0 - dummy read disabled with msi, enabled without msi
+	 * IH_DUMMY_RD_OVERRIDE=1 - dummy read controlled by IH_DUMMY_RD_EN
+	 */
+	interrupt_cntl &= ~IH_DUMMY_RD_OVERRIDE;
+	/* IH_REQ_NONSNOOP_EN=1 if ring is in non-cacheable memory, e.g., vram */
+	interrupt_cntl &= ~IH_REQ_NONSNOOP_EN;
+	WREG32(INTERRUPT_CNTL, interrupt_cntl);
+
+	WREG32(IH_RB_BASE, rdev->ih.gpu_addr >> 8);
+	rb_bufsz = drm_order(rdev->ih.ring_size / 4);
+
+	ih_rb_cntl = (IH_WPTR_OVERFLOW_ENABLE |
+		      IH_WPTR_OVERFLOW_CLEAR |
+		      (rb_bufsz << 1));
+
+	if (rdev->wb.enabled)
+		ih_rb_cntl |= IH_WPTR_WRITEBACK_ENABLE;
+
+	/* set the writeback address whether it's enabled or not */
+	WREG32(IH_RB_WPTR_ADDR_LO, (rdev->wb.gpu_addr + R600_WB_IH_WPTR_OFFSET) & 0xFFFFFFFC);
+	WREG32(IH_RB_WPTR_ADDR_HI, upper_32_bits(rdev->wb.gpu_addr + R600_WB_IH_WPTR_OFFSET) & 0xFF);
+
+	WREG32(IH_RB_CNTL, ih_rb_cntl);
+
+	/* set rptr, wptr to 0 */
+	WREG32(IH_RB_RPTR, 0);
+	WREG32(IH_RB_WPTR, 0);
+
+	/* Default settings for IH_CNTL (disabled at first) */
+	ih_cntl = MC_WRREQ_CREDIT(0x10) | MC_WR_CLEAN_CNT(0x10) | MC_VMID(0);
+	/* RPTR_REARM only works if msi's are enabled */
+	if (rdev->msi_enabled)
+		ih_cntl |= RPTR_REARM;
+	WREG32(IH_CNTL, ih_cntl);
+
+	/* force the active interrupt state to all disabled */
+	si_disable_interrupt_state(rdev);
+
+	/* enable irqs */
+	si_enable_interrupts(rdev);
+
+	return ret;
+}
+
+int si_irq_set(struct radeon_device *rdev)
+{
+	u32 cp_int_cntl = CNTX_BUSY_INT_ENABLE | CNTX_EMPTY_INT_ENABLE;
+	u32 cp_int_cntl1 = 0, cp_int_cntl2 = 0;
+	u32 crtc1 = 0, crtc2 = 0, crtc3 = 0, crtc4 = 0, crtc5 = 0, crtc6 = 0;
+	u32 hpd1, hpd2, hpd3, hpd4, hpd5, hpd6;
+	u32 grbm_int_cntl = 0;
+	u32 grph1 = 0, grph2 = 0, grph3 = 0, grph4 = 0, grph5 = 0, grph6 = 0;
+
+	if (!rdev->irq.installed) {
+		WARN(1, "Can't enable IRQ/MSI because no handler is installed\n");
+		return -EINVAL;
+	}
+	/* don't enable anything if the ih is disabled */
+	if (!rdev->ih.enabled) {
+		si_disable_interrupts(rdev);
+		/* force the active interrupt state to all disabled */
+		si_disable_interrupt_state(rdev);
+		return 0;
+	}
+
+	hpd1 = RREG32(DC_HPD1_INT_CONTROL) & ~DC_HPDx_INT_EN;
+	hpd2 = RREG32(DC_HPD2_INT_CONTROL) & ~DC_HPDx_INT_EN;
+	hpd3 = RREG32(DC_HPD3_INT_CONTROL) & ~DC_HPDx_INT_EN;
+	hpd4 = RREG32(DC_HPD4_INT_CONTROL) & ~DC_HPDx_INT_EN;
+	hpd5 = RREG32(DC_HPD5_INT_CONTROL) & ~DC_HPDx_INT_EN;
+	hpd6 = RREG32(DC_HPD6_INT_CONTROL) & ~DC_HPDx_INT_EN;
+
+	/* enable CP interrupts on all rings */
+	if (rdev->irq.sw_int[RADEON_RING_TYPE_GFX_INDEX]) {
+		DRM_DEBUG("si_irq_set: sw int gfx\n");
+		cp_int_cntl |= TIME_STAMP_INT_ENABLE;
+	}
+	if (rdev->irq.sw_int[CAYMAN_RING_TYPE_CP1_INDEX]) {
+		DRM_DEBUG("si_irq_set: sw int cp1\n");
+		cp_int_cntl1 |= TIME_STAMP_INT_ENABLE;
+	}
+	if (rdev->irq.sw_int[CAYMAN_RING_TYPE_CP2_INDEX]) {
+		DRM_DEBUG("si_irq_set: sw int cp2\n");
+		cp_int_cntl2 |= TIME_STAMP_INT_ENABLE;
+	}
+	if (rdev->irq.crtc_vblank_int[0] ||
+	    rdev->irq.pflip[0]) {
+		DRM_DEBUG("si_irq_set: vblank 0\n");
+		crtc1 |= VBLANK_INT_MASK;
+	}
+	if (rdev->irq.crtc_vblank_int[1] ||
+	    rdev->irq.pflip[1]) {
+		DRM_DEBUG("si_irq_set: vblank 1\n");
+		crtc2 |= VBLANK_INT_MASK;
+	}
+	if (rdev->irq.crtc_vblank_int[2] ||
+	    rdev->irq.pflip[2]) {
+		DRM_DEBUG("si_irq_set: vblank 2\n");
+		crtc3 |= VBLANK_INT_MASK;
+	}
+	if (rdev->irq.crtc_vblank_int[3] ||
+	    rdev->irq.pflip[3]) {
+		DRM_DEBUG("si_irq_set: vblank 3\n");
+		crtc4 |= VBLANK_INT_MASK;
+	}
+	if (rdev->irq.crtc_vblank_int[4] ||
+	    rdev->irq.pflip[4]) {
+		DRM_DEBUG("si_irq_set: vblank 4\n");
+		crtc5 |= VBLANK_INT_MASK;
+	}
+	if (rdev->irq.crtc_vblank_int[5] ||
+	    rdev->irq.pflip[5]) {
+		DRM_DEBUG("si_irq_set: vblank 5\n");
+		crtc6 |= VBLANK_INT_MASK;
+	}
+	if (rdev->irq.hpd[0]) {
+		DRM_DEBUG("si_irq_set: hpd 1\n");
+		hpd1 |= DC_HPDx_INT_EN;
+	}
+	if (rdev->irq.hpd[1]) {
+		DRM_DEBUG("si_irq_set: hpd 2\n");
+		hpd2 |= DC_HPDx_INT_EN;
+	}
+	if (rdev->irq.hpd[2]) {
+		DRM_DEBUG("si_irq_set: hpd 3\n");
+		hpd3 |= DC_HPDx_INT_EN;
+	}
+	if (rdev->irq.hpd[3]) {
+		DRM_DEBUG("si_irq_set: hpd 4\n");
+		hpd4 |= DC_HPDx_INT_EN;
+	}
+	if (rdev->irq.hpd[4]) {
+		DRM_DEBUG("si_irq_set: hpd 5\n");
+		hpd5 |= DC_HPDx_INT_EN;
+	}
+	if (rdev->irq.hpd[5]) {
+		DRM_DEBUG("si_irq_set: hpd 6\n");
+		hpd6 |= DC_HPDx_INT_EN;
+	}
+	if (rdev->irq.gui_idle) {
+		DRM_DEBUG("gui idle\n");
+		grbm_int_cntl |= GUI_IDLE_INT_ENABLE;
+	}
+
+	WREG32(CP_INT_CNTL_RING0, cp_int_cntl);
+	WREG32(CP_INT_CNTL_RING1, cp_int_cntl1);
+	WREG32(CP_INT_CNTL_RING2, cp_int_cntl2);
+
+	WREG32(GRBM_INT_CNTL, grbm_int_cntl);
+
+	WREG32(INT_MASK + EVERGREEN_CRTC0_REGISTER_OFFSET, crtc1);
+	WREG32(INT_MASK + EVERGREEN_CRTC1_REGISTER_OFFSET, crtc2);
+	if (rdev->num_crtc >= 4) {
+		WREG32(INT_MASK + EVERGREEN_CRTC2_REGISTER_OFFSET, crtc3);
+		WREG32(INT_MASK + EVERGREEN_CRTC3_REGISTER_OFFSET, crtc4);
+	}
+	if (rdev->num_crtc >= 6) {
+		WREG32(INT_MASK + EVERGREEN_CRTC4_REGISTER_OFFSET, crtc5);
+		WREG32(INT_MASK + EVERGREEN_CRTC5_REGISTER_OFFSET, crtc6);
+	}
+
+	WREG32(GRPH_INT_CONTROL + EVERGREEN_CRTC0_REGISTER_OFFSET, grph1);
+	WREG32(GRPH_INT_CONTROL + EVERGREEN_CRTC1_REGISTER_OFFSET, grph2);
+	if (rdev->num_crtc >= 4) {
+		WREG32(GRPH_INT_CONTROL + EVERGREEN_CRTC2_REGISTER_OFFSET, grph3);
+		WREG32(GRPH_INT_CONTROL + EVERGREEN_CRTC3_REGISTER_OFFSET, grph4);
+	}
+	if (rdev->num_crtc >= 6) {
+		WREG32(GRPH_INT_CONTROL + EVERGREEN_CRTC4_REGISTER_OFFSET, grph5);
+		WREG32(GRPH_INT_CONTROL + EVERGREEN_CRTC5_REGISTER_OFFSET, grph6);
+	}
+
+	WREG32(DC_HPD1_INT_CONTROL, hpd1);
+	WREG32(DC_HPD2_INT_CONTROL, hpd2);
+	WREG32(DC_HPD3_INT_CONTROL, hpd3);
+	WREG32(DC_HPD4_INT_CONTROL, hpd4);
+	WREG32(DC_HPD5_INT_CONTROL, hpd5);
+	WREG32(DC_HPD6_INT_CONTROL, hpd6);
+
+	return 0;
+}
+
+static inline void si_irq_ack(struct radeon_device *rdev)
+{
+	u32 tmp;
+
+	rdev->irq.stat_regs.evergreen.disp_int = RREG32(DISP_INTERRUPT_STATUS);
+	rdev->irq.stat_regs.evergreen.disp_int_cont = RREG32(DISP_INTERRUPT_STATUS_CONTINUE);
+	rdev->irq.stat_regs.evergreen.disp_int_cont2 = RREG32(DISP_INTERRUPT_STATUS_CONTINUE2);
+	rdev->irq.stat_regs.evergreen.disp_int_cont3 = RREG32(DISP_INTERRUPT_STATUS_CONTINUE3);
+	rdev->irq.stat_regs.evergreen.disp_int_cont4 = RREG32(DISP_INTERRUPT_STATUS_CONTINUE4);
+	rdev->irq.stat_regs.evergreen.disp_int_cont5 = RREG32(DISP_INTERRUPT_STATUS_CONTINUE5);
+	rdev->irq.stat_regs.evergreen.d1grph_int = RREG32(GRPH_INT_STATUS + EVERGREEN_CRTC0_REGISTER_OFFSET);
+	rdev->irq.stat_regs.evergreen.d2grph_int = RREG32(GRPH_INT_STATUS + EVERGREEN_CRTC1_REGISTER_OFFSET);
+	if (rdev->num_crtc >= 4) {
+		rdev->irq.stat_regs.evergreen.d3grph_int = RREG32(GRPH_INT_STATUS + EVERGREEN_CRTC2_REGISTER_OFFSET);
+		rdev->irq.stat_regs.evergreen.d4grph_int = RREG32(GRPH_INT_STATUS + EVERGREEN_CRTC3_REGISTER_OFFSET);
+	}
+	if (rdev->num_crtc >= 6) {
+		rdev->irq.stat_regs.evergreen.d5grph_int = RREG32(GRPH_INT_STATUS + EVERGREEN_CRTC4_REGISTER_OFFSET);
+		rdev->irq.stat_regs.evergreen.d6grph_int = RREG32(GRPH_INT_STATUS + EVERGREEN_CRTC5_REGISTER_OFFSET);
+	}
+
+	if (rdev->irq.stat_regs.evergreen.d1grph_int & GRPH_PFLIP_INT_OCCURRED)
+		WREG32(GRPH_INT_STATUS + EVERGREEN_CRTC0_REGISTER_OFFSET, GRPH_PFLIP_INT_CLEAR);
+	if (rdev->irq.stat_regs.evergreen.d2grph_int & GRPH_PFLIP_INT_OCCURRED)
+		WREG32(GRPH_INT_STATUS + EVERGREEN_CRTC1_REGISTER_OFFSET, GRPH_PFLIP_INT_CLEAR);
+	if (rdev->irq.stat_regs.evergreen.disp_int & LB_D1_VBLANK_INTERRUPT)
+		WREG32(VBLANK_STATUS + EVERGREEN_CRTC0_REGISTER_OFFSET, VBLANK_ACK);
+	if (rdev->irq.stat_regs.evergreen.disp_int & LB_D1_VLINE_INTERRUPT)
+		WREG32(VLINE_STATUS + EVERGREEN_CRTC0_REGISTER_OFFSET, VLINE_ACK);
+	if (rdev->irq.stat_regs.evergreen.disp_int_cont & LB_D2_VBLANK_INTERRUPT)
+		WREG32(VBLANK_STATUS + EVERGREEN_CRTC1_REGISTER_OFFSET, VBLANK_ACK);
+	if (rdev->irq.stat_regs.evergreen.disp_int_cont & LB_D2_VLINE_INTERRUPT)
+		WREG32(VLINE_STATUS + EVERGREEN_CRTC1_REGISTER_OFFSET, VLINE_ACK);
+
+	if (rdev->num_crtc >= 4) {
+		if (rdev->irq.stat_regs.evergreen.d3grph_int & GRPH_PFLIP_INT_OCCURRED)
+			WREG32(GRPH_INT_STATUS + EVERGREEN_CRTC2_REGISTER_OFFSET, GRPH_PFLIP_INT_CLEAR);
+		if (rdev->irq.stat_regs.evergreen.d4grph_int & GRPH_PFLIP_INT_OCCURRED)
+			WREG32(GRPH_INT_STATUS + EVERGREEN_CRTC3_REGISTER_OFFSET, GRPH_PFLIP_INT_CLEAR);
+		if (rdev->irq.stat_regs.evergreen.disp_int_cont2 & LB_D3_VBLANK_INTERRUPT)
+			WREG32(VBLANK_STATUS + EVERGREEN_CRTC2_REGISTER_OFFSET, VBLANK_ACK);
+		if (rdev->irq.stat_regs.evergreen.disp_int_cont2 & LB_D3_VLINE_INTERRUPT)
+			WREG32(VLINE_STATUS + EVERGREEN_CRTC2_REGISTER_OFFSET, VLINE_ACK);
+		if (rdev->irq.stat_regs.evergreen.disp_int_cont3 & LB_D4_VBLANK_INTERRUPT)
+			WREG32(VBLANK_STATUS + EVERGREEN_CRTC3_REGISTER_OFFSET, VBLANK_ACK);
+		if (rdev->irq.stat_regs.evergreen.disp_int_cont3 & LB_D4_VLINE_INTERRUPT)
+			WREG32(VLINE_STATUS + EVERGREEN_CRTC3_REGISTER_OFFSET, VLINE_ACK);
+	}
+
+	if (rdev->num_crtc >= 6) {
+		if (rdev->irq.stat_regs.evergreen.d5grph_int & GRPH_PFLIP_INT_OCCURRED)
+			WREG32(GRPH_INT_STATUS + EVERGREEN_CRTC4_REGISTER_OFFSET, GRPH_PFLIP_INT_CLEAR);
+		if (rdev->irq.stat_regs.evergreen.d6grph_int & GRPH_PFLIP_INT_OCCURRED)
+			WREG32(GRPH_INT_STATUS + EVERGREEN_CRTC5_REGISTER_OFFSET, GRPH_PFLIP_INT_CLEAR);
+		if (rdev->irq.stat_regs.evergreen.disp_int_cont4 & LB_D5_VBLANK_INTERRUPT)
+			WREG32(VBLANK_STATUS + EVERGREEN_CRTC4_REGISTER_OFFSET, VBLANK_ACK);
+		if (rdev->irq.stat_regs.evergreen.disp_int_cont4 & LB_D5_VLINE_INTERRUPT)
+			WREG32(VLINE_STATUS + EVERGREEN_CRTC4_REGISTER_OFFSET, VLINE_ACK);
+		if (rdev->irq.stat_regs.evergreen.disp_int_cont5 & LB_D6_VBLANK_INTERRUPT)
+			WREG32(VBLANK_STATUS + EVERGREEN_CRTC5_REGISTER_OFFSET, VBLANK_ACK);
+		if (rdev->irq.stat_regs.evergreen.disp_int_cont5 & LB_D6_VLINE_INTERRUPT)
+			WREG32(VLINE_STATUS + EVERGREEN_CRTC5_REGISTER_OFFSET, VLINE_ACK);
+	}
+
+	if (rdev->irq.stat_regs.evergreen.disp_int & DC_HPD1_INTERRUPT) {
+		tmp = RREG32(DC_HPD1_INT_CONTROL);
+		tmp |= DC_HPDx_INT_ACK;
+		WREG32(DC_HPD1_INT_CONTROL, tmp);
+	}
+	if (rdev->irq.stat_regs.evergreen.disp_int_cont & DC_HPD2_INTERRUPT) {
+		tmp = RREG32(DC_HPD2_INT_CONTROL);
+		tmp |= DC_HPDx_INT_ACK;
+		WREG32(DC_HPD2_INT_CONTROL, tmp);
+	}
+	if (rdev->irq.stat_regs.evergreen.disp_int_cont2 & DC_HPD3_INTERRUPT) {
+		tmp = RREG32(DC_HPD3_INT_CONTROL);
+		tmp |= DC_HPDx_INT_ACK;
+		WREG32(DC_HPD3_INT_CONTROL, tmp);
+	}
+	if (rdev->irq.stat_regs.evergreen.disp_int_cont3 & DC_HPD4_INTERRUPT) {
+		tmp = RREG32(DC_HPD4_INT_CONTROL);
+		tmp |= DC_HPDx_INT_ACK;
+		WREG32(DC_HPD4_INT_CONTROL, tmp);
+	}
+	if (rdev->irq.stat_regs.evergreen.disp_int_cont4 & DC_HPD5_INTERRUPT) {
+		tmp = RREG32(DC_HPD5_INT_CONTROL);
+		tmp |= DC_HPDx_INT_ACK;
+		WREG32(DC_HPD5_INT_CONTROL, tmp);
+	}
+	if (rdev->irq.stat_regs.evergreen.disp_int_cont5 & DC_HPD6_INTERRUPT) {
+		tmp = RREG32(DC_HPD5_INT_CONTROL);
+		tmp |= DC_HPDx_INT_ACK;
+		WREG32(DC_HPD6_INT_CONTROL, tmp);
+	}
+}
+
+static void si_irq_disable(struct radeon_device *rdev)
+{
+	si_disable_interrupts(rdev);
+	/* Wait and acknowledge irq */
+	mdelay(1);
+	si_irq_ack(rdev);
+	si_disable_interrupt_state(rdev);
+}
+
+static void si_irq_suspend(struct radeon_device *rdev)
+{
+	si_irq_disable(rdev);
+	si_rlc_stop(rdev);
+}
+
+static void si_irq_fini(struct radeon_device *rdev)
+{
+	si_irq_suspend(rdev);
+	r600_ih_ring_fini(rdev);
+}
+
+static inline u32 si_get_ih_wptr(struct radeon_device *rdev)
+{
+	u32 wptr, tmp;
+
+	if (rdev->wb.enabled)
+		wptr = le32_to_cpu(rdev->wb.wb[R600_WB_IH_WPTR_OFFSET/4]);
+	else
+		wptr = RREG32(IH_RB_WPTR);
+
+	if (wptr & RB_OVERFLOW) {
+		/* When a ring buffer overflow happen start parsing interrupt
+		 * from the last not overwritten vector (wptr + 16). Hopefully
+		 * this should allow us to catchup.
+		 */
+		dev_warn(rdev->dev, "IH ring buffer overflow (0x%08X, %d, %d)\n",
+			wptr, rdev->ih.rptr, (wptr + 16) + rdev->ih.ptr_mask);
+		rdev->ih.rptr = (wptr + 16) & rdev->ih.ptr_mask;
+		tmp = RREG32(IH_RB_CNTL);
+		tmp |= IH_WPTR_OVERFLOW_CLEAR;
+		WREG32(IH_RB_CNTL, tmp);
+	}
+	return (wptr & rdev->ih.ptr_mask);
+}
+
+/*        SI IV Ring
+ * Each IV ring entry is 128 bits:
+ * [7:0]    - interrupt source id
+ * [31:8]   - reserved
+ * [59:32]  - interrupt source data
+ * [63:60]  - reserved
+ * [71:64]  - RINGID
+ * [79:72]  - VMID
+ * [127:80] - reserved
+ */
+int si_irq_process(struct radeon_device *rdev)
+{
+	u32 wptr;
+	u32 rptr;
+	u32 src_id, src_data, ring_id;
+	u32 ring_index;
+	unsigned long flags;
+	bool queue_hotplug = false;
+
+	if (!rdev->ih.enabled || rdev->shutdown)
+		return IRQ_NONE;
+
+	wptr = si_get_ih_wptr(rdev);
+	rptr = rdev->ih.rptr;
+	DRM_DEBUG("si_irq_process start: rptr %d, wptr %d\n", rptr, wptr);
+
+	spin_lock_irqsave(&rdev->ih.lock, flags);
+	if (rptr == wptr) {
+		spin_unlock_irqrestore(&rdev->ih.lock, flags);
+		return IRQ_NONE;
+	}
+restart_ih:
+	/* Order reading of wptr vs. reading of IH ring data */
+	rmb();
+
+	/* display interrupts */
+	si_irq_ack(rdev);
+
+	rdev->ih.wptr = wptr;
+	while (rptr != wptr) {
+		/* wptr/rptr are in bytes! */
+		ring_index = rptr / 4;
+		src_id =  le32_to_cpu(rdev->ih.ring[ring_index]) & 0xff;
+		src_data = le32_to_cpu(rdev->ih.ring[ring_index + 1]) & 0xfffffff;
+		ring_id = le32_to_cpu(rdev->ih.ring[ring_index + 2]) & 0xff;
+
+		switch (src_id) {
+		case 1: /* D1 vblank/vline */
+			switch (src_data) {
+			case 0: /* D1 vblank */
+				if (rdev->irq.stat_regs.evergreen.disp_int & LB_D1_VBLANK_INTERRUPT) {
+					if (rdev->irq.crtc_vblank_int[0]) {
+						drm_handle_vblank(rdev->ddev, 0);
+						rdev->pm.vblank_sync = true;
+						wake_up(&rdev->irq.vblank_queue);
+					}
+					if (rdev->irq.pflip[0])
+						radeon_crtc_handle_flip(rdev, 0);
+					rdev->irq.stat_regs.evergreen.disp_int &= ~LB_D1_VBLANK_INTERRUPT;
+					DRM_DEBUG("IH: D1 vblank\n");
+				}
+				break;
+			case 1: /* D1 vline */
+				if (rdev->irq.stat_regs.evergreen.disp_int & LB_D1_VLINE_INTERRUPT) {
+					rdev->irq.stat_regs.evergreen.disp_int &= ~LB_D1_VLINE_INTERRUPT;
+					DRM_DEBUG("IH: D1 vline\n");
+				}
+				break;
+			default:
+				DRM_DEBUG("Unhandled interrupt: %d %d\n", src_id, src_data);
+				break;
+			}
+			break;
+		case 2: /* D2 vblank/vline */
+			switch (src_data) {
+			case 0: /* D2 vblank */
+				if (rdev->irq.stat_regs.evergreen.disp_int_cont & LB_D2_VBLANK_INTERRUPT) {
+					if (rdev->irq.crtc_vblank_int[1]) {
+						drm_handle_vblank(rdev->ddev, 1);
+						rdev->pm.vblank_sync = true;
+						wake_up(&rdev->irq.vblank_queue);
+					}
+					if (rdev->irq.pflip[1])
+						radeon_crtc_handle_flip(rdev, 1);
+					rdev->irq.stat_regs.evergreen.disp_int_cont &= ~LB_D2_VBLANK_INTERRUPT;
+					DRM_DEBUG("IH: D2 vblank\n");
+				}
+				break;
+			case 1: /* D2 vline */
+				if (rdev->irq.stat_regs.evergreen.disp_int_cont & LB_D2_VLINE_INTERRUPT) {
+					rdev->irq.stat_regs.evergreen.disp_int_cont &= ~LB_D2_VLINE_INTERRUPT;
+					DRM_DEBUG("IH: D2 vline\n");
+				}
+				break;
+			default:
+				DRM_DEBUG("Unhandled interrupt: %d %d\n", src_id, src_data);
+				break;
+			}
+			break;
+		case 3: /* D3 vblank/vline */
+			switch (src_data) {
+			case 0: /* D3 vblank */
+				if (rdev->irq.stat_regs.evergreen.disp_int_cont2 & LB_D3_VBLANK_INTERRUPT) {
+					if (rdev->irq.crtc_vblank_int[2]) {
+						drm_handle_vblank(rdev->ddev, 2);
+						rdev->pm.vblank_sync = true;
+						wake_up(&rdev->irq.vblank_queue);
+					}
+					if (rdev->irq.pflip[2])
+						radeon_crtc_handle_flip(rdev, 2);
+					rdev->irq.stat_regs.evergreen.disp_int_cont2 &= ~LB_D3_VBLANK_INTERRUPT;
+					DRM_DEBUG("IH: D3 vblank\n");
+				}
+				break;
+			case 1: /* D3 vline */
+				if (rdev->irq.stat_regs.evergreen.disp_int_cont2 & LB_D3_VLINE_INTERRUPT) {
+					rdev->irq.stat_regs.evergreen.disp_int_cont2 &= ~LB_D3_VLINE_INTERRUPT;
+					DRM_DEBUG("IH: D3 vline\n");
+				}
+				break;
+			default:
+				DRM_DEBUG("Unhandled interrupt: %d %d\n", src_id, src_data);
+				break;
+			}
+			break;
+		case 4: /* D4 vblank/vline */
+			switch (src_data) {
+			case 0: /* D4 vblank */
+				if (rdev->irq.stat_regs.evergreen.disp_int_cont3 & LB_D4_VBLANK_INTERRUPT) {
+					if (rdev->irq.crtc_vblank_int[3]) {
+						drm_handle_vblank(rdev->ddev, 3);
+						rdev->pm.vblank_sync = true;
+						wake_up(&rdev->irq.vblank_queue);
+					}
+					if (rdev->irq.pflip[3])
+						radeon_crtc_handle_flip(rdev, 3);
+					rdev->irq.stat_regs.evergreen.disp_int_cont3 &= ~LB_D4_VBLANK_INTERRUPT;
+					DRM_DEBUG("IH: D4 vblank\n");
+				}
+				break;
+			case 1: /* D4 vline */
+				if (rdev->irq.stat_regs.evergreen.disp_int_cont3 & LB_D4_VLINE_INTERRUPT) {
+					rdev->irq.stat_regs.evergreen.disp_int_cont3 &= ~LB_D4_VLINE_INTERRUPT;
+					DRM_DEBUG("IH: D4 vline\n");
+				}
+				break;
+			default:
+				DRM_DEBUG("Unhandled interrupt: %d %d\n", src_id, src_data);
+				break;
+			}
+			break;
+		case 5: /* D5 vblank/vline */
+			switch (src_data) {
+			case 0: /* D5 vblank */
+				if (rdev->irq.stat_regs.evergreen.disp_int_cont4 & LB_D5_VBLANK_INTERRUPT) {
+					if (rdev->irq.crtc_vblank_int[4]) {
+						drm_handle_vblank(rdev->ddev, 4);
+						rdev->pm.vblank_sync = true;
+						wake_up(&rdev->irq.vblank_queue);
+					}
+					if (rdev->irq.pflip[4])
+						radeon_crtc_handle_flip(rdev, 4);
+					rdev->irq.stat_regs.evergreen.disp_int_cont4 &= ~LB_D5_VBLANK_INTERRUPT;
+					DRM_DEBUG("IH: D5 vblank\n");
+				}
+				break;
+			case 1: /* D5 vline */
+				if (rdev->irq.stat_regs.evergreen.disp_int_cont4 & LB_D5_VLINE_INTERRUPT) {
+					rdev->irq.stat_regs.evergreen.disp_int_cont4 &= ~LB_D5_VLINE_INTERRUPT;
+					DRM_DEBUG("IH: D5 vline\n");
+				}
+				break;
+			default:
+				DRM_DEBUG("Unhandled interrupt: %d %d\n", src_id, src_data);
+				break;
+			}
+			break;
+		case 6: /* D6 vblank/vline */
+			switch (src_data) {
+			case 0: /* D6 vblank */
+				if (rdev->irq.stat_regs.evergreen.disp_int_cont5 & LB_D6_VBLANK_INTERRUPT) {
+					if (rdev->irq.crtc_vblank_int[5]) {
+						drm_handle_vblank(rdev->ddev, 5);
+						rdev->pm.vblank_sync = true;
+						wake_up(&rdev->irq.vblank_queue);
+					}
+					if (rdev->irq.pflip[5])
+						radeon_crtc_handle_flip(rdev, 5);
+					rdev->irq.stat_regs.evergreen.disp_int_cont5 &= ~LB_D6_VBLANK_INTERRUPT;
+					DRM_DEBUG("IH: D6 vblank\n");
+				}
+				break;
+			case 1: /* D6 vline */
+				if (rdev->irq.stat_regs.evergreen.disp_int_cont5 & LB_D6_VLINE_INTERRUPT) {
+					rdev->irq.stat_regs.evergreen.disp_int_cont5 &= ~LB_D6_VLINE_INTERRUPT;
+					DRM_DEBUG("IH: D6 vline\n");
+				}
+				break;
+			default:
+				DRM_DEBUG("Unhandled interrupt: %d %d\n", src_id, src_data);
+				break;
+			}
+			break;
+		case 42: /* HPD hotplug */
+			switch (src_data) {
+			case 0:
+				if (rdev->irq.stat_regs.evergreen.disp_int & DC_HPD1_INTERRUPT) {
+					rdev->irq.stat_regs.evergreen.disp_int &= ~DC_HPD1_INTERRUPT;
+					queue_hotplug = true;
+					DRM_DEBUG("IH: HPD1\n");
+				}
+				break;
+			case 1:
+				if (rdev->irq.stat_regs.evergreen.disp_int_cont & DC_HPD2_INTERRUPT) {
+					rdev->irq.stat_regs.evergreen.disp_int_cont &= ~DC_HPD2_INTERRUPT;
+					queue_hotplug = true;
+					DRM_DEBUG("IH: HPD2\n");
+				}
+				break;
+			case 2:
+				if (rdev->irq.stat_regs.evergreen.disp_int_cont2 & DC_HPD3_INTERRUPT) {
+					rdev->irq.stat_regs.evergreen.disp_int_cont2 &= ~DC_HPD3_INTERRUPT;
+					queue_hotplug = true;
+					DRM_DEBUG("IH: HPD3\n");
+				}
+				break;
+			case 3:
+				if (rdev->irq.stat_regs.evergreen.disp_int_cont3 & DC_HPD4_INTERRUPT) {
+					rdev->irq.stat_regs.evergreen.disp_int_cont3 &= ~DC_HPD4_INTERRUPT;
+					queue_hotplug = true;
+					DRM_DEBUG("IH: HPD4\n");
+				}
+				break;
+			case 4:
+				if (rdev->irq.stat_regs.evergreen.disp_int_cont4 & DC_HPD5_INTERRUPT) {
+					rdev->irq.stat_regs.evergreen.disp_int_cont4 &= ~DC_HPD5_INTERRUPT;
+					queue_hotplug = true;
+					DRM_DEBUG("IH: HPD5\n");
+				}
+				break;
+			case 5:
+				if (rdev->irq.stat_regs.evergreen.disp_int_cont5 & DC_HPD6_INTERRUPT) {
+					rdev->irq.stat_regs.evergreen.disp_int_cont5 &= ~DC_HPD6_INTERRUPT;
+					queue_hotplug = true;
+					DRM_DEBUG("IH: HPD6\n");
+				}
+				break;
+			default:
+				DRM_DEBUG("Unhandled interrupt: %d %d\n", src_id, src_data);
+				break;
+			}
+			break;
+		case 176: /* RINGID0 CP_INT */
+			radeon_fence_process(rdev, RADEON_RING_TYPE_GFX_INDEX);
+			break;
+		case 177: /* RINGID1 CP_INT */
+			radeon_fence_process(rdev, CAYMAN_RING_TYPE_CP1_INDEX);
+			break;
+		case 178: /* RINGID2 CP_INT */
+			radeon_fence_process(rdev, CAYMAN_RING_TYPE_CP2_INDEX);
+			break;
+		case 181: /* CP EOP event */
+			DRM_DEBUG("IH: CP EOP\n");
+			switch (ring_id) {
+			case 0:
+				radeon_fence_process(rdev, RADEON_RING_TYPE_GFX_INDEX);
+				break;
+			case 1:
+				radeon_fence_process(rdev, CAYMAN_RING_TYPE_CP1_INDEX);
+				break;
+			case 2:
+				radeon_fence_process(rdev, CAYMAN_RING_TYPE_CP2_INDEX);
+				break;
+			}
+			break;
+		case 233: /* GUI IDLE */
+			DRM_DEBUG("IH: GUI idle\n");
+			rdev->pm.gui_idle = true;
+			wake_up(&rdev->irq.idle_queue);
+			break;
+		default:
+			DRM_DEBUG("Unhandled interrupt: %d %d\n", src_id, src_data);
+			break;
+		}
+
+		/* wptr/rptr are in bytes! */
+		rptr += 16;
+		rptr &= rdev->ih.ptr_mask;
+	}
+	/* make sure wptr hasn't changed while processing */
+	wptr = si_get_ih_wptr(rdev);
+	if (wptr != rdev->ih.wptr)
+		goto restart_ih;
+	if (queue_hotplug)
+		schedule_work(&rdev->hotplug_work);
+	rdev->ih.rptr = rptr;
+	WREG32(IH_RB_RPTR, rdev->ih.rptr);
+	spin_unlock_irqrestore(&rdev->ih.lock, flags);
+	return IRQ_HANDLED;
+}
+
+/*
+ * startup/shutdown callbacks
+ */
+static int si_startup(struct radeon_device *rdev)
+{
+	struct radeon_ring *ring;
+	int r;
+
+	if (!rdev->me_fw || !rdev->pfp_fw || !rdev->ce_fw ||
+	    !rdev->rlc_fw || !rdev->mc_fw) {
+		r = si_init_microcode(rdev);
+		if (r) {
+			DRM_ERROR("Failed to load firmware!\n");
+			return r;
+		}
+	}
+
+	r = si_mc_load_microcode(rdev);
+	if (r) {
+		DRM_ERROR("Failed to load MC firmware!\n");
+		return r;
+	}
+
+	r = r600_vram_scratch_init(rdev);
+	if (r)
+		return r;
+
+	si_mc_program(rdev);
+	r = si_pcie_gart_enable(rdev);
+	if (r)
+		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) {
+		DRM_ERROR("Failed to init rlc BOs!\n");
+		return r;
+	}
+
+	/* allocate wb buffer */
+	r = radeon_wb_init(rdev);
+	if (r)
+		return r;
+
+	r = radeon_fence_driver_start_ring(rdev, RADEON_RING_TYPE_GFX_INDEX);
+	if (r) {
+		dev_err(rdev->dev, "failed initializing CP fences (%d).\n", r);
+		return r;
+	}
+
+	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);
+		return r;
+	}
+
+	r = radeon_fence_driver_start_ring(rdev, CAYMAN_RING_TYPE_CP2_INDEX);
+	if (r) {
+		dev_err(rdev->dev, "failed initializing CP fences (%d).\n", r);
+		return r;
+	}
+
+	/* Enable IRQ */
+	r = si_irq_init(rdev);
+	if (r) {
+		DRM_ERROR("radeon: IH init failed (%d).\n", r);
+		radeon_irq_kms_fini(rdev);
+		return r;
+	}
+	si_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,
+			     CP_RB0_RPTR, CP_RB0_WPTR,
+			     0, 0xfffff, RADEON_CP_PACKET2);
+	if (r)
+		return r;
+
+	ring = &rdev->ring[CAYMAN_RING_TYPE_CP1_INDEX];
+	r = radeon_ring_init(rdev, ring, ring->ring_size, RADEON_WB_CP1_RPTR_OFFSET,
+			     CP_RB1_RPTR, CP_RB1_WPTR,
+			     0, 0xfffff, RADEON_CP_PACKET2);
+	if (r)
+		return r;
+
+	ring = &rdev->ring[CAYMAN_RING_TYPE_CP2_INDEX];
+	r = radeon_ring_init(rdev, ring, ring->ring_size, RADEON_WB_CP2_RPTR_OFFSET,
+			     CP_RB2_RPTR, CP_RB2_WPTR,
+			     0, 0xfffff, RADEON_CP_PACKET2);
+	if (r)
+		return r;
+
+	r = si_cp_load_microcode(rdev);
+	if (r)
+		return r;
+	r = si_cp_resume(rdev);
+	if (r)
+		return r;
+
+	r = radeon_ib_pool_start(rdev);
+	if (r)
+		return r;
+
+	r = radeon_ib_test(rdev, RADEON_RING_TYPE_GFX_INDEX, &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]);
+	if (r) {
+		DRM_ERROR("radeon: failed testing IB (%d) on CP ring 0\n", r);
+		rdev->accel_working = false;
+		return r;
+	}
+
+	r = radeon_ib_test(rdev, CAYMAN_RING_TYPE_CP1_INDEX, &rdev->ring[CAYMAN_RING_TYPE_CP1_INDEX]);
+	if (r) {
+		DRM_ERROR("radeon: failed testing IB (%d) on CP ring 1\n", r);
+		rdev->accel_working = false;
+		return r;
+	}
+
+	r = radeon_ib_test(rdev, CAYMAN_RING_TYPE_CP2_INDEX, &rdev->ring[CAYMAN_RING_TYPE_CP2_INDEX]);
+	if (r) {
+		DRM_ERROR("radeon: failed testing IB (%d) on CP ring 2\n", r);
+		rdev->accel_working = false;
+		return r;
+	}
+
+	r = radeon_vm_manager_start(rdev);
+	if (r)
+		return r;
+
+	return 0;
+}
+
+int si_resume(struct radeon_device *rdev)
+{
+	int r;
+
+	/* Do not reset GPU before posting, on rv770 hw unlike on r500 hw,
+	 * posting will perform necessary task to bring back GPU into good
+	 * shape.
+	 */
+	/* post card */
+	atom_asic_init(rdev->mode_info.atom_context);
+
+	rdev->accel_working = true;
+	r = si_startup(rdev);
+	if (r) {
+		DRM_ERROR("si startup failed on resume\n");
+		rdev->accel_working = false;
+		return r;
+	}
+
+	return r;
+
+}
+
+int si_suspend(struct radeon_device *rdev)
+{
+	/* FIXME: we should wait for ring to be empty */
+	radeon_ib_pool_suspend(rdev);
+	radeon_vm_manager_suspend(rdev);
+#if 0
+	r600_blit_suspend(rdev);
+#endif
+	si_cp_enable(rdev, false);
+	rdev->ring[RADEON_RING_TYPE_GFX_INDEX].ready = false;
+	rdev->ring[CAYMAN_RING_TYPE_CP1_INDEX].ready = false;
+	rdev->ring[CAYMAN_RING_TYPE_CP2_INDEX].ready = false;
+	si_irq_suspend(rdev);
+	radeon_wb_disable(rdev);
+	si_pcie_gart_disable(rdev);
+	return 0;
+}
+
+/* Plan is to move initialization in that function and use
+ * helper function so that radeon_device_init pretty much
+ * do nothing more than calling asic specific function. This
+ * should also allow to remove a bunch of callback function
+ * like vram_info.
+ */
+int si_init(struct radeon_device *rdev)
+{
+	struct radeon_ring *ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX];
+	int r;
+
+	/* This don't do much */
+	r = radeon_gem_init(rdev);
+	if (r)
+		return r;
+	/* Read BIOS */
+	if (!radeon_get_bios(rdev)) {
+		if (ASIC_IS_AVIVO(rdev))
+			return -EINVAL;
+	}
+	/* Must be an ATOMBIOS */
+	if (!rdev->is_atom_bios) {
+		dev_err(rdev->dev, "Expecting atombios for cayman GPU\n");
+		return -EINVAL;
+	}
+	r = radeon_atombios_init(rdev);
+	if (r)
+		return r;
+
+	/* Post card if necessary */
+	if (!radeon_card_posted(rdev)) {
+		if (!rdev->bios) {
+			dev_err(rdev->dev, "Card not posted and no BIOS - ignoring\n");
+			return -EINVAL;
+		}
+		DRM_INFO("GPU not posted. posting now...\n");
+		atom_asic_init(rdev->mode_info.atom_context);
+	}
+	/* Initialize scratch registers */
+	si_scratch_init(rdev);
+	/* Initialize surface registers */
+	radeon_surface_init(rdev);
+	/* Initialize clocks */
+	radeon_get_clock_info(rdev->ddev);
+
+	/* Fence driver */
+	r = radeon_fence_driver_init(rdev);
+	if (r)
+		return r;
+
+	/* initialize memory controller */
+	r = si_mc_init(rdev);
+	if (r)
+		return r;
+	/* Memory manager */
+	r = radeon_bo_init(rdev);
+	if (r)
+		return r;
+
+	r = radeon_irq_kms_init(rdev);
+	if (r)
+		return r;
+
+	ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX];
+	ring->ring_obj = NULL;
+	r600_ring_init(rdev, ring, 1024 * 1024);
+
+	ring = &rdev->ring[CAYMAN_RING_TYPE_CP1_INDEX];
+	ring->ring_obj = NULL;
+	r600_ring_init(rdev, ring, 1024 * 1024);
+
+	ring = &rdev->ring[CAYMAN_RING_TYPE_CP2_INDEX];
+	ring->ring_obj = NULL;
+	r600_ring_init(rdev, ring, 1024 * 1024);
+
+	rdev->ih.ring_obj = NULL;
+	r600_ih_ring_init(rdev, 64 * 1024);
+
+	r = r600_pcie_gart_init(rdev);
+	if (r)
+		return r;
+
+	r = radeon_ib_pool_init(rdev);
+	rdev->accel_working = true;
+	if (r) {
+		dev_err(rdev->dev, "IB initialization failed (%d).\n", r);
+		rdev->accel_working = false;
+	}
+	r = radeon_vm_manager_init(rdev);
+	if (r) {
+		dev_err(rdev->dev, "vm manager initialization failed (%d).\n", r);
+	}
+
+	r = si_startup(rdev);
+	if (r) {
+		dev_err(rdev->dev, "disabling GPU acceleration\n");
+		si_cp_fini(rdev);
+		si_irq_fini(rdev);
+		si_rlc_fini(rdev);
+		radeon_wb_fini(rdev);
+		r100_ib_fini(rdev);
+		radeon_vm_manager_fini(rdev);
+		radeon_irq_kms_fini(rdev);
+		si_pcie_gart_fini(rdev);
+		rdev->accel_working = false;
+	}
+
+	/* Don't start up if the MC ucode is missing.
+	 * The default clocks and voltages before the MC ucode
+	 * is loaded are not suffient for advanced operations.
+	 */
+	if (!rdev->mc_fw) {
+		DRM_ERROR("radeon: MC ucode required for NI+.\n");
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+void si_fini(struct radeon_device *rdev)
+{
+#if 0
+	r600_blit_fini(rdev);
+#endif
+	si_cp_fini(rdev);
+	si_irq_fini(rdev);
+	si_rlc_fini(rdev);
+	radeon_wb_fini(rdev);
+	radeon_vm_manager_fini(rdev);
+	r100_ib_fini(rdev);
+	radeon_irq_kms_fini(rdev);
+	si_pcie_gart_fini(rdev);
+	r600_vram_scratch_fini(rdev);
+	radeon_gem_fini(rdev);
+	radeon_semaphore_driver_fini(rdev);
+	radeon_fence_driver_fini(rdev);
+	radeon_bo_fini(rdev);
+	radeon_atombios_fini(rdev);
+	kfree(rdev->bios);
+	rdev->bios = NULL;
+}
+
diff --git a/drivers/gpu/drm/radeon/si_blit_shaders.c b/drivers/gpu/drm/radeon/si_blit_shaders.c
new file mode 100644
index 000000000000..ec415e7dfa4b
--- /dev/null
+++ b/drivers/gpu/drm/radeon/si_blit_shaders.c
@@ -0,0 +1,253 @@
+/*
+ * Copyright 2011 Advanced Micro Devices, 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 (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 HOLDER(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:
+ *     Alex Deucher <alexander.deucher@amd.com>
+ */
+
+#include <linux/types.h>
+#include <linux/bug.h>
+#include <linux/kernel.h>
+
+const u32 si_default_state[] =
+{
+	0xc0066900,
+	0x00000000,
+	0x00000060, /* DB_RENDER_CONTROL */
+	0x00000000, /* DB_COUNT_CONTROL */
+	0x00000000, /* DB_DEPTH_VIEW */
+	0x0000002a, /* DB_RENDER_OVERRIDE */
+	0x00000000, /* DB_RENDER_OVERRIDE2 */
+	0x00000000, /* DB_HTILE_DATA_BASE */
+
+	0xc0046900,
+	0x00000008,
+	0x00000000, /* DB_DEPTH_BOUNDS_MIN */
+	0x00000000, /* DB_DEPTH_BOUNDS_MAX */
+	0x00000000, /* DB_STENCIL_CLEAR */
+	0x00000000, /* DB_DEPTH_CLEAR */
+
+	0xc0036900,
+	0x0000000f,
+	0x00000000, /* DB_DEPTH_INFO */
+	0x00000000, /* DB_Z_INFO */
+	0x00000000, /* DB_STENCIL_INFO */
+
+	0xc0016900,
+	0x00000080,
+	0x00000000, /* PA_SC_WINDOW_OFFSET */
+
+	0xc00d6900,
+	0x00000083,
+	0x0000ffff, /* PA_SC_CLIPRECT_RULE */
+	0x00000000, /* PA_SC_CLIPRECT_0_TL */
+	0x20002000, /* PA_SC_CLIPRECT_0_BR */
+	0x00000000,
+	0x20002000,
+	0x00000000,
+	0x20002000,
+	0x00000000,
+	0x20002000,
+	0xaaaaaaaa, /* PA_SC_EDGERULE */
+	0x00000000, /* PA_SU_HARDWARE_SCREEN_OFFSET */
+	0x0000000f, /* CB_TARGET_MASK */
+	0x0000000f, /* CB_SHADER_MASK */
+
+	0xc0226900,
+	0x00000094,
+	0x80000000, /* PA_SC_VPORT_SCISSOR_0_TL */
+	0x20002000, /* PA_SC_VPORT_SCISSOR_0_BR */
+	0x80000000,
+	0x20002000,
+	0x80000000,
+	0x20002000,
+	0x80000000,
+	0x20002000,
+	0x80000000,
+	0x20002000,
+	0x80000000,
+	0x20002000,
+	0x80000000,
+	0x20002000,
+	0x80000000,
+	0x20002000,
+	0x80000000,
+	0x20002000,
+	0x80000000,
+	0x20002000,
+	0x80000000,
+	0x20002000,
+	0x80000000,
+	0x20002000,
+	0x80000000,
+	0x20002000,
+	0x80000000,
+	0x20002000,
+	0x80000000,
+	0x20002000,
+	0x80000000,
+	0x20002000,
+	0x00000000, /* PA_SC_VPORT_ZMIN_0 */
+	0x3f800000, /* PA_SC_VPORT_ZMAX_0 */
+
+	0xc0026900,
+	0x000000d9,
+	0x00000000, /* CP_RINGID */
+	0x00000000, /* CP_VMID */
+
+	0xc0046900,
+	0x00000100,
+	0xffffffff, /* VGT_MAX_VTX_INDX */
+	0x00000000, /* VGT_MIN_VTX_INDX */
+	0x00000000, /* VGT_INDX_OFFSET */
+	0x00000000, /* VGT_MULTI_PRIM_IB_RESET_INDX */
+
+	0xc0046900,
+	0x00000105,
+	0x00000000, /* CB_BLEND_RED */
+	0x00000000, /* CB_BLEND_GREEN */
+	0x00000000, /* CB_BLEND_BLUE */
+	0x00000000, /* CB_BLEND_ALPHA */
+
+	0xc0016900,
+	0x000001e0,
+	0x00000000, /* CB_BLEND0_CONTROL */
+
+	0xc00e6900,
+	0x00000200,
+	0x00000000, /* DB_DEPTH_CONTROL */
+	0x00000000, /* DB_EQAA */
+	0x00cc0010, /* CB_COLOR_CONTROL */
+	0x00000210, /* DB_SHADER_CONTROL */
+	0x00010000, /* PA_CL_CLIP_CNTL */
+	0x00000004, /* PA_SU_SC_MODE_CNTL */
+	0x00000100, /* PA_CL_VTE_CNTL */
+	0x00000000, /* PA_CL_VS_OUT_CNTL */
+	0x00000000, /* PA_CL_NANINF_CNTL */
+	0x00000000, /* PA_SU_LINE_STIPPLE_CNTL */
+	0x00000000, /* PA_SU_LINE_STIPPLE_SCALE */
+	0x00000000, /* PA_SU_PRIM_FILTER_CNTL */
+	0x00000000, /*  */
+	0x00000000, /*  */
+
+	0xc0116900,
+	0x00000280,
+	0x00000000, /* PA_SU_POINT_SIZE */
+	0x00000000, /* PA_SU_POINT_MINMAX */
+	0x00000008, /* PA_SU_LINE_CNTL */
+	0x00000000, /* PA_SC_LINE_STIPPLE */
+	0x00000000, /* VGT_OUTPUT_PATH_CNTL */
+	0x00000000, /* VGT_HOS_CNTL */
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000, /* VGT_GS_MODE */
+
+	0xc0026900,
+	0x00000292,
+	0x00000000, /* PA_SC_MODE_CNTL_0 */
+	0x00000000, /* PA_SC_MODE_CNTL_1 */
+
+	0xc0016900,
+	0x000002a1,
+	0x00000000, /* VGT_PRIMITIVEID_EN */
+
+	0xc0016900,
+	0x000002a5,
+	0x00000000, /* VGT_MULTI_PRIM_IB_RESET_EN */
+
+	0xc0026900,
+	0x000002a8,
+	0x00000000, /* VGT_INSTANCE_STEP_RATE_0 */
+	0x00000000,
+
+	0xc0026900,
+	0x000002ad,
+	0x00000000, /* VGT_REUSE_OFF */
+	0x00000000,
+
+	0xc0016900,
+	0x000002d5,
+	0x00000000, /* VGT_SHADER_STAGES_EN */
+
+	0xc0016900,
+	0x000002dc,
+	0x0000aa00, /* DB_ALPHA_TO_MASK */
+
+	0xc0066900,
+	0x000002de,
+	0x00000000, /* PA_SU_POLY_OFFSET_DB_FMT_CNTL */
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+
+	0xc0026900,
+	0x000002e5,
+	0x00000000, /* VGT_STRMOUT_CONFIG */
+	0x00000000,
+
+	0xc01b6900,
+	0x000002f5,
+	0x76543210, /* PA_SC_CENTROID_PRIORITY_0 */
+	0xfedcba98, /* PA_SC_CENTROID_PRIORITY_1 */
+	0x00000000, /* PA_SC_LINE_CNTL */
+	0x00000000, /* PA_SC_AA_CONFIG */
+	0x00000005, /* PA_SU_VTX_CNTL */
+	0x3f800000, /* PA_CL_GB_VERT_CLIP_ADJ */
+	0x3f800000, /* PA_CL_GB_VERT_DISC_ADJ */
+	0x3f800000, /* PA_CL_GB_HORZ_CLIP_ADJ */
+	0x3f800000, /* PA_CL_GB_HORZ_DISC_ADJ */
+	0x00000000, /* PA_SC_AA_SAMPLE_LOCS_PIXEL_X0Y0_0 */
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0xffffffff, /* PA_SC_AA_MASK_X0Y0_X1Y0 */
+	0xffffffff,
+
+	0xc0026900,
+	0x00000316,
+	0x0000000e, /* VGT_VERTEX_REUSE_BLOCK_CNTL */
+	0x00000010, /*  */
+};
+
+const u32 si_default_size = ARRAY_SIZE(si_default_state);
diff --git a/drivers/gpu/drm/radeon/si_blit_shaders.h b/drivers/gpu/drm/radeon/si_blit_shaders.h
new file mode 100644
index 000000000000..c739e51e3961
--- /dev/null
+++ b/drivers/gpu/drm/radeon/si_blit_shaders.h
@@ -0,0 +1,32 @@
+/*
+ * Copyright 2011 Advanced Micro Devices, 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 (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 HOLDER(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.
+ *
+ */
+
+#ifndef SI_BLIT_SHADERS_H
+#define SI_BLIT_SHADERS_H
+
+extern const u32 si_default_state[];
+
+extern const u32 si_default_size;
+
+#endif
diff --git a/drivers/gpu/drm/radeon/si_reg.h b/drivers/gpu/drm/radeon/si_reg.h
new file mode 100644
index 000000000000..eda938a7cb6e
--- /dev/null
+++ b/drivers/gpu/drm/radeon/si_reg.h
@@ -0,0 +1,33 @@
+/*
+ * Copyright 2010 Advanced Micro Devices, 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: Alex Deucher
+ */
+#ifndef __SI_REG_H__
+#define __SI_REG_H__
+
+/* SI */
+#define SI_DC_GPIO_HPD_MASK                      0x65b0
+#define SI_DC_GPIO_HPD_A                         0x65b4
+#define SI_DC_GPIO_HPD_EN                        0x65b8
+#define SI_DC_GPIO_HPD_Y                         0x65bc
+
+#endif
diff --git a/drivers/gpu/drm/radeon/sid.h b/drivers/gpu/drm/radeon/sid.h
new file mode 100644
index 000000000000..53ea2c42dbd6
--- /dev/null
+++ b/drivers/gpu/drm/radeon/sid.h
@@ -0,0 +1,886 @@
+/*
+ * Copyright 2011 Advanced Micro Devices, 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: Alex Deucher
+ */
+#ifndef SI_H
+#define SI_H
+
+#define	CG_MULT_THERMAL_STATUS					0x714
+#define		ASIC_MAX_TEMP(x)				((x) << 0)
+#define		ASIC_MAX_TEMP_MASK				0x000001ff
+#define		ASIC_MAX_TEMP_SHIFT				0
+#define		CTF_TEMP(x)					((x) << 9)
+#define		CTF_TEMP_MASK					0x0003fe00
+#define		CTF_TEMP_SHIFT					9
+
+#define SI_MAX_SH_GPRS           256
+#define SI_MAX_TEMP_GPRS         16
+#define SI_MAX_SH_THREADS        256
+#define SI_MAX_SH_STACK_ENTRIES  4096
+#define SI_MAX_FRC_EOV_CNT       16384
+#define SI_MAX_BACKENDS          8
+#define SI_MAX_BACKENDS_MASK     0xFF
+#define SI_MAX_BACKENDS_PER_SE_MASK     0x0F
+#define SI_MAX_SIMDS             12
+#define SI_MAX_SIMDS_MASK        0x0FFF
+#define SI_MAX_SIMDS_PER_SE_MASK        0x00FF
+#define SI_MAX_PIPES             8
+#define SI_MAX_PIPES_MASK        0xFF
+#define SI_MAX_PIPES_PER_SIMD_MASK      0x3F
+#define SI_MAX_LDS_NUM           0xFFFF
+#define SI_MAX_TCC               16
+#define SI_MAX_TCC_MASK          0xFFFF
+
+#define VGA_HDP_CONTROL  				0x328
+#define		VGA_MEMORY_DISABLE				(1 << 4)
+
+#define DMIF_ADDR_CONFIG  				0xBD4
+
+#define	SRBM_STATUS				        0xE50
+
+#define	CC_SYS_RB_BACKEND_DISABLE			0xe80
+#define	GC_USER_SYS_RB_BACKEND_DISABLE			0xe84
+
+#define VM_L2_CNTL					0x1400
+#define		ENABLE_L2_CACHE					(1 << 0)
+#define		ENABLE_L2_FRAGMENT_PROCESSING			(1 << 1)
+#define		L2_CACHE_PTE_ENDIAN_SWAP_MODE(x)		((x) << 2)
+#define		L2_CACHE_PDE_ENDIAN_SWAP_MODE(x)		((x) << 4)
+#define		ENABLE_L2_PTE_CACHE_LRU_UPDATE_BY_WRITE		(1 << 9)
+#define		ENABLE_L2_PDE0_CACHE_LRU_UPDATE_BY_WRITE	(1 << 10)
+#define		EFFECTIVE_L2_QUEUE_SIZE(x)			(((x) & 7) << 15)
+#define		CONTEXT1_IDENTITY_ACCESS_MODE(x)		(((x) & 3) << 19)
+#define VM_L2_CNTL2					0x1404
+#define		INVALIDATE_ALL_L1_TLBS				(1 << 0)
+#define		INVALIDATE_L2_CACHE				(1 << 1)
+#define		INVALIDATE_CACHE_MODE(x)			((x) << 26)
+#define			INVALIDATE_PTE_AND_PDE_CACHES		0
+#define			INVALIDATE_ONLY_PTE_CACHES		1
+#define			INVALIDATE_ONLY_PDE_CACHES		2
+#define VM_L2_CNTL3					0x1408
+#define		BANK_SELECT(x)					((x) << 0)
+#define		L2_CACHE_UPDATE_MODE(x)				((x) << 6)
+#define		L2_CACHE_BIGK_FRAGMENT_SIZE(x)			((x) << 15)
+#define		L2_CACHE_BIGK_ASSOCIATIVITY			(1 << 20)
+#define	VM_L2_STATUS					0x140C
+#define		L2_BUSY						(1 << 0)
+#define VM_CONTEXT0_CNTL				0x1410
+#define		ENABLE_CONTEXT					(1 << 0)
+#define		PAGE_TABLE_DEPTH(x)				(((x) & 3) << 1)
+#define		RANGE_PROTECTION_FAULT_ENABLE_DEFAULT		(1 << 4)
+#define VM_CONTEXT1_CNTL				0x1414
+#define VM_CONTEXT0_CNTL2				0x1430
+#define VM_CONTEXT1_CNTL2				0x1434
+#define	VM_CONTEXT8_PAGE_TABLE_BASE_ADDR		0x1438
+#define	VM_CONTEXT9_PAGE_TABLE_BASE_ADDR		0x143c
+#define	VM_CONTEXT10_PAGE_TABLE_BASE_ADDR		0x1440
+#define	VM_CONTEXT11_PAGE_TABLE_BASE_ADDR		0x1444
+#define	VM_CONTEXT12_PAGE_TABLE_BASE_ADDR		0x1448
+#define	VM_CONTEXT13_PAGE_TABLE_BASE_ADDR		0x144c
+#define	VM_CONTEXT14_PAGE_TABLE_BASE_ADDR		0x1450
+#define	VM_CONTEXT15_PAGE_TABLE_BASE_ADDR		0x1454
+
+#define VM_INVALIDATE_REQUEST				0x1478
+#define VM_INVALIDATE_RESPONSE				0x147c
+
+#define VM_CONTEXT0_PROTECTION_FAULT_DEFAULT_ADDR	0x1518
+#define VM_CONTEXT1_PROTECTION_FAULT_DEFAULT_ADDR	0x151c
+
+#define	VM_CONTEXT0_PAGE_TABLE_BASE_ADDR		0x153c
+#define	VM_CONTEXT1_PAGE_TABLE_BASE_ADDR		0x1540
+#define	VM_CONTEXT2_PAGE_TABLE_BASE_ADDR		0x1544
+#define	VM_CONTEXT3_PAGE_TABLE_BASE_ADDR		0x1548
+#define	VM_CONTEXT4_PAGE_TABLE_BASE_ADDR		0x154c
+#define	VM_CONTEXT5_PAGE_TABLE_BASE_ADDR		0x1550
+#define	VM_CONTEXT6_PAGE_TABLE_BASE_ADDR		0x1554
+#define	VM_CONTEXT7_PAGE_TABLE_BASE_ADDR		0x1558
+#define	VM_CONTEXT0_PAGE_TABLE_START_ADDR		0x155c
+#define	VM_CONTEXT1_PAGE_TABLE_START_ADDR		0x1560
+
+#define	VM_CONTEXT0_PAGE_TABLE_END_ADDR			0x157C
+#define	VM_CONTEXT1_PAGE_TABLE_END_ADDR			0x1580
+
+#define MC_SHARED_CHMAP						0x2004
+#define		NOOFCHAN_SHIFT					12
+#define		NOOFCHAN_MASK					0x0000f000
+#define MC_SHARED_CHREMAP					0x2008
+
+#define	MC_VM_FB_LOCATION				0x2024
+#define	MC_VM_AGP_TOP					0x2028
+#define	MC_VM_AGP_BOT					0x202C
+#define	MC_VM_AGP_BASE					0x2030
+#define	MC_VM_SYSTEM_APERTURE_LOW_ADDR			0x2034
+#define	MC_VM_SYSTEM_APERTURE_HIGH_ADDR			0x2038
+#define	MC_VM_SYSTEM_APERTURE_DEFAULT_ADDR		0x203C
+
+#define	MC_VM_MX_L1_TLB_CNTL				0x2064
+#define		ENABLE_L1_TLB					(1 << 0)
+#define		ENABLE_L1_FRAGMENT_PROCESSING			(1 << 1)
+#define		SYSTEM_ACCESS_MODE_PA_ONLY			(0 << 3)
+#define		SYSTEM_ACCESS_MODE_USE_SYS_MAP			(1 << 3)
+#define		SYSTEM_ACCESS_MODE_IN_SYS			(2 << 3)
+#define		SYSTEM_ACCESS_MODE_NOT_IN_SYS			(3 << 3)
+#define		SYSTEM_APERTURE_UNMAPPED_ACCESS_PASS_THRU	(0 << 5)
+#define		ENABLE_ADVANCED_DRIVER_MODEL			(1 << 6)
+
+#define MC_SHARED_BLACKOUT_CNTL           		0x20ac
+
+#define	MC_ARB_RAMCFG					0x2760
+#define		NOOFBANK_SHIFT					0
+#define		NOOFBANK_MASK					0x00000003
+#define		NOOFRANK_SHIFT					2
+#define		NOOFRANK_MASK					0x00000004
+#define		NOOFROWS_SHIFT					3
+#define		NOOFROWS_MASK					0x00000038
+#define		NOOFCOLS_SHIFT					6
+#define		NOOFCOLS_MASK					0x000000C0
+#define		CHANSIZE_SHIFT					8
+#define		CHANSIZE_MASK					0x00000100
+#define		CHANSIZE_OVERRIDE				(1 << 11)
+#define		NOOFGROUPS_SHIFT				12
+#define		NOOFGROUPS_MASK					0x00001000
+
+#define	MC_SEQ_TRAIN_WAKEUP_CNTL			0x2808
+#define		TRAIN_DONE_D0      			(1 << 30)
+#define		TRAIN_DONE_D1      			(1 << 31)
+
+#define MC_SEQ_SUP_CNTL           			0x28c8
+#define		RUN_MASK      				(1 << 0)
+#define MC_SEQ_SUP_PGM           			0x28cc
+
+#define MC_IO_PAD_CNTL_D0           			0x29d0
+#define		MEM_FALL_OUT_CMD      			(1 << 8)
+
+#define MC_SEQ_IO_DEBUG_INDEX           		0x2a44
+#define MC_SEQ_IO_DEBUG_DATA           			0x2a48
+
+#define	HDP_HOST_PATH_CNTL				0x2C00
+#define	HDP_NONSURFACE_BASE				0x2C04
+#define	HDP_NONSURFACE_INFO				0x2C08
+#define	HDP_NONSURFACE_SIZE				0x2C0C
+
+#define HDP_ADDR_CONFIG  				0x2F48
+#define HDP_MISC_CNTL					0x2F4C
+#define 	HDP_FLUSH_INVALIDATE_CACHE			(1 << 0)
+
+#define IH_RB_CNTL                                        0x3e00
+#       define IH_RB_ENABLE                               (1 << 0)
+#       define IH_IB_SIZE(x)                              ((x) << 1) /* log2 */
+#       define IH_RB_FULL_DRAIN_ENABLE                    (1 << 6)
+#       define IH_WPTR_WRITEBACK_ENABLE                   (1 << 8)
+#       define IH_WPTR_WRITEBACK_TIMER(x)                 ((x) << 9) /* log2 */
+#       define IH_WPTR_OVERFLOW_ENABLE                    (1 << 16)
+#       define IH_WPTR_OVERFLOW_CLEAR                     (1 << 31)
+#define IH_RB_BASE                                        0x3e04
+#define IH_RB_RPTR                                        0x3e08
+#define IH_RB_WPTR                                        0x3e0c
+#       define RB_OVERFLOW                                (1 << 0)
+#       define WPTR_OFFSET_MASK                           0x3fffc
+#define IH_RB_WPTR_ADDR_HI                                0x3e10
+#define IH_RB_WPTR_ADDR_LO                                0x3e14
+#define IH_CNTL                                           0x3e18
+#       define ENABLE_INTR                                (1 << 0)
+#       define IH_MC_SWAP(x)                              ((x) << 1)
+#       define IH_MC_SWAP_NONE                            0
+#       define IH_MC_SWAP_16BIT                           1
+#       define IH_MC_SWAP_32BIT                           2
+#       define IH_MC_SWAP_64BIT                           3
+#       define RPTR_REARM                                 (1 << 4)
+#       define MC_WRREQ_CREDIT(x)                         ((x) << 15)
+#       define MC_WR_CLEAN_CNT(x)                         ((x) << 20)
+#       define MC_VMID(x)                                 ((x) << 25)
+
+#define	CONFIG_MEMSIZE					0x5428
+
+#define INTERRUPT_CNTL                                    0x5468
+#       define IH_DUMMY_RD_OVERRIDE                       (1 << 0)
+#       define IH_DUMMY_RD_EN                             (1 << 1)
+#       define IH_REQ_NONSNOOP_EN                         (1 << 3)
+#       define GEN_IH_INT_EN                              (1 << 8)
+#define INTERRUPT_CNTL2                                   0x546c
+
+#define HDP_MEM_COHERENCY_FLUSH_CNTL			0x5480
+
+#define	BIF_FB_EN						0x5490
+#define		FB_READ_EN					(1 << 0)
+#define		FB_WRITE_EN					(1 << 1)
+
+#define HDP_REG_COHERENCY_FLUSH_CNTL			0x54A0
+
+#define	DC_LB_MEMORY_SPLIT					0x6b0c
+#define		DC_LB_MEMORY_CONFIG(x)				((x) << 20)
+
+#define	PRIORITY_A_CNT						0x6b18
+#define		PRIORITY_MARK_MASK				0x7fff
+#define		PRIORITY_OFF					(1 << 16)
+#define		PRIORITY_ALWAYS_ON				(1 << 20)
+#define	PRIORITY_B_CNT						0x6b1c
+
+#define	DPG_PIPE_ARBITRATION_CONTROL3				0x6cc8
+#       define LATENCY_WATERMARK_MASK(x)			((x) << 16)
+#define	DPG_PIPE_LATENCY_CONTROL				0x6ccc
+#       define LATENCY_LOW_WATERMARK(x)				((x) << 0)
+#       define LATENCY_HIGH_WATERMARK(x)			((x) << 16)
+
+/* 0x6bb8, 0x77b8, 0x103b8, 0x10fb8, 0x11bb8, 0x127b8 */
+#define VLINE_STATUS                                    0x6bb8
+#       define VLINE_OCCURRED                           (1 << 0)
+#       define VLINE_ACK                                (1 << 4)
+#       define VLINE_STAT                               (1 << 12)
+#       define VLINE_INTERRUPT                          (1 << 16)
+#       define VLINE_INTERRUPT_TYPE                     (1 << 17)
+/* 0x6bbc, 0x77bc, 0x103bc, 0x10fbc, 0x11bbc, 0x127bc */
+#define VBLANK_STATUS                                   0x6bbc
+#       define VBLANK_OCCURRED                          (1 << 0)
+#       define VBLANK_ACK                               (1 << 4)
+#       define VBLANK_STAT                              (1 << 12)
+#       define VBLANK_INTERRUPT                         (1 << 16)
+#       define VBLANK_INTERRUPT_TYPE                    (1 << 17)
+
+/* 0x6b40, 0x7740, 0x10340, 0x10f40, 0x11b40, 0x12740 */
+#define INT_MASK                                        0x6b40
+#       define VBLANK_INT_MASK                          (1 << 0)
+#       define VLINE_INT_MASK                           (1 << 4)
+
+#define DISP_INTERRUPT_STATUS                           0x60f4
+#       define LB_D1_VLINE_INTERRUPT                    (1 << 2)
+#       define LB_D1_VBLANK_INTERRUPT                   (1 << 3)
+#       define DC_HPD1_INTERRUPT                        (1 << 17)
+#       define DC_HPD1_RX_INTERRUPT                     (1 << 18)
+#       define DACA_AUTODETECT_INTERRUPT                (1 << 22)
+#       define DACB_AUTODETECT_INTERRUPT                (1 << 23)
+#       define DC_I2C_SW_DONE_INTERRUPT                 (1 << 24)
+#       define DC_I2C_HW_DONE_INTERRUPT                 (1 << 25)
+#define DISP_INTERRUPT_STATUS_CONTINUE                  0x60f8
+#       define LB_D2_VLINE_INTERRUPT                    (1 << 2)
+#       define LB_D2_VBLANK_INTERRUPT                   (1 << 3)
+#       define DC_HPD2_INTERRUPT                        (1 << 17)
+#       define DC_HPD2_RX_INTERRUPT                     (1 << 18)
+#       define DISP_TIMER_INTERRUPT                     (1 << 24)
+#define DISP_INTERRUPT_STATUS_CONTINUE2                 0x60fc
+#       define LB_D3_VLINE_INTERRUPT                    (1 << 2)
+#       define LB_D3_VBLANK_INTERRUPT                   (1 << 3)
+#       define DC_HPD3_INTERRUPT                        (1 << 17)
+#       define DC_HPD3_RX_INTERRUPT                     (1 << 18)
+#define DISP_INTERRUPT_STATUS_CONTINUE3                 0x6100
+#       define LB_D4_VLINE_INTERRUPT                    (1 << 2)
+#       define LB_D4_VBLANK_INTERRUPT                   (1 << 3)
+#       define DC_HPD4_INTERRUPT                        (1 << 17)
+#       define DC_HPD4_RX_INTERRUPT                     (1 << 18)
+#define DISP_INTERRUPT_STATUS_CONTINUE4                 0x614c
+#       define LB_D5_VLINE_INTERRUPT                    (1 << 2)
+#       define LB_D5_VBLANK_INTERRUPT                   (1 << 3)
+#       define DC_HPD5_INTERRUPT                        (1 << 17)
+#       define DC_HPD5_RX_INTERRUPT                     (1 << 18)
+#define DISP_INTERRUPT_STATUS_CONTINUE5                 0x6150
+#       define LB_D6_VLINE_INTERRUPT                    (1 << 2)
+#       define LB_D6_VBLANK_INTERRUPT                   (1 << 3)
+#       define DC_HPD6_INTERRUPT                        (1 << 17)
+#       define DC_HPD6_RX_INTERRUPT                     (1 << 18)
+
+/* 0x6858, 0x7458, 0x10058, 0x10c58, 0x11858, 0x12458 */
+#define GRPH_INT_STATUS                                 0x6858
+#       define GRPH_PFLIP_INT_OCCURRED                  (1 << 0)
+#       define GRPH_PFLIP_INT_CLEAR                     (1 << 8)
+/* 0x685c, 0x745c, 0x1005c, 0x10c5c, 0x1185c, 0x1245c */
+#define	GRPH_INT_CONTROL			        0x685c
+#       define GRPH_PFLIP_INT_MASK                      (1 << 0)
+#       define GRPH_PFLIP_INT_TYPE                      (1 << 8)
+
+#define	DACA_AUTODETECT_INT_CONTROL			0x66c8
+
+#define DC_HPD1_INT_STATUS                              0x601c
+#define DC_HPD2_INT_STATUS                              0x6028
+#define DC_HPD3_INT_STATUS                              0x6034
+#define DC_HPD4_INT_STATUS                              0x6040
+#define DC_HPD5_INT_STATUS                              0x604c
+#define DC_HPD6_INT_STATUS                              0x6058
+#       define DC_HPDx_INT_STATUS                       (1 << 0)
+#       define DC_HPDx_SENSE                            (1 << 1)
+#       define DC_HPDx_RX_INT_STATUS                    (1 << 8)
+
+#define DC_HPD1_INT_CONTROL                             0x6020
+#define DC_HPD2_INT_CONTROL                             0x602c
+#define DC_HPD3_INT_CONTROL                             0x6038
+#define DC_HPD4_INT_CONTROL                             0x6044
+#define DC_HPD5_INT_CONTROL                             0x6050
+#define DC_HPD6_INT_CONTROL                             0x605c
+#       define DC_HPDx_INT_ACK                          (1 << 0)
+#       define DC_HPDx_INT_POLARITY                     (1 << 8)
+#       define DC_HPDx_INT_EN                           (1 << 16)
+#       define DC_HPDx_RX_INT_ACK                       (1 << 20)
+#       define DC_HPDx_RX_INT_EN                        (1 << 24)
+
+#define DC_HPD1_CONTROL                                   0x6024
+#define DC_HPD2_CONTROL                                   0x6030
+#define DC_HPD3_CONTROL                                   0x603c
+#define DC_HPD4_CONTROL                                   0x6048
+#define DC_HPD5_CONTROL                                   0x6054
+#define DC_HPD6_CONTROL                                   0x6060
+#       define DC_HPDx_CONNECTION_TIMER(x)                ((x) << 0)
+#       define DC_HPDx_RX_INT_TIMER(x)                    ((x) << 16)
+#       define DC_HPDx_EN                                 (1 << 28)
+
+/* 0x6e98, 0x7a98, 0x10698, 0x11298, 0x11e98, 0x12a98 */
+#define CRTC_STATUS_FRAME_COUNT                         0x6e98
+
+#define	GRBM_CNTL					0x8000
+#define		GRBM_READ_TIMEOUT(x)				((x) << 0)
+
+#define	GRBM_STATUS2					0x8008
+#define		RLC_RQ_PENDING 					(1 << 0)
+#define		RLC_BUSY 					(1 << 8)
+#define		TC_BUSY 					(1 << 9)
+
+#define	GRBM_STATUS					0x8010
+#define		CMDFIFO_AVAIL_MASK				0x0000000F
+#define		RING2_RQ_PENDING				(1 << 4)
+#define		SRBM_RQ_PENDING					(1 << 5)
+#define		RING1_RQ_PENDING				(1 << 6)
+#define		CF_RQ_PENDING					(1 << 7)
+#define		PF_RQ_PENDING					(1 << 8)
+#define		GDS_DMA_RQ_PENDING				(1 << 9)
+#define		GRBM_EE_BUSY					(1 << 10)
+#define		DB_CLEAN					(1 << 12)
+#define		CB_CLEAN					(1 << 13)
+#define		TA_BUSY 					(1 << 14)
+#define		GDS_BUSY 					(1 << 15)
+#define		VGT_BUSY					(1 << 17)
+#define		IA_BUSY_NO_DMA					(1 << 18)
+#define		IA_BUSY						(1 << 19)
+#define		SX_BUSY 					(1 << 20)
+#define		SPI_BUSY					(1 << 22)
+#define		BCI_BUSY					(1 << 23)
+#define		SC_BUSY 					(1 << 24)
+#define		PA_BUSY 					(1 << 25)
+#define		DB_BUSY 					(1 << 26)
+#define		CP_COHERENCY_BUSY      				(1 << 28)
+#define		CP_BUSY 					(1 << 29)
+#define		CB_BUSY 					(1 << 30)
+#define		GUI_ACTIVE					(1 << 31)
+#define	GRBM_STATUS_SE0					0x8014
+#define	GRBM_STATUS_SE1					0x8018
+#define		SE_DB_CLEAN					(1 << 1)
+#define		SE_CB_CLEAN					(1 << 2)
+#define		SE_BCI_BUSY					(1 << 22)
+#define		SE_VGT_BUSY					(1 << 23)
+#define		SE_PA_BUSY					(1 << 24)
+#define		SE_TA_BUSY					(1 << 25)
+#define		SE_SX_BUSY					(1 << 26)
+#define		SE_SPI_BUSY					(1 << 27)
+#define		SE_SC_BUSY					(1 << 29)
+#define		SE_DB_BUSY					(1 << 30)
+#define		SE_CB_BUSY					(1 << 31)
+
+#define	GRBM_SOFT_RESET					0x8020
+#define		SOFT_RESET_CP					(1 << 0)
+#define		SOFT_RESET_CB					(1 << 1)
+#define		SOFT_RESET_RLC					(1 << 2)
+#define		SOFT_RESET_DB					(1 << 3)
+#define		SOFT_RESET_GDS					(1 << 4)
+#define		SOFT_RESET_PA					(1 << 5)
+#define		SOFT_RESET_SC					(1 << 6)
+#define		SOFT_RESET_BCI					(1 << 7)
+#define		SOFT_RESET_SPI					(1 << 8)
+#define		SOFT_RESET_SX					(1 << 10)
+#define		SOFT_RESET_TC					(1 << 11)
+#define		SOFT_RESET_TA					(1 << 12)
+#define		SOFT_RESET_VGT					(1 << 14)
+#define		SOFT_RESET_IA					(1 << 15)
+
+#define GRBM_GFX_INDEX          			0x802C
+
+#define GRBM_INT_CNTL                                   0x8060
+#       define RDERR_INT_ENABLE                         (1 << 0)
+#       define GUI_IDLE_INT_ENABLE                      (1 << 19)
+
+#define	SCRATCH_REG0					0x8500
+#define	SCRATCH_REG1					0x8504
+#define	SCRATCH_REG2					0x8508
+#define	SCRATCH_REG3					0x850C
+#define	SCRATCH_REG4					0x8510
+#define	SCRATCH_REG5					0x8514
+#define	SCRATCH_REG6					0x8518
+#define	SCRATCH_REG7					0x851C
+
+#define	SCRATCH_UMSK					0x8540
+#define	SCRATCH_ADDR					0x8544
+
+#define	CP_SEM_WAIT_TIMER				0x85BC
+
+#define	CP_SEM_INCOMPLETE_TIMER_CNTL			0x85C8
+
+#define CP_ME_CNTL					0x86D8
+#define		CP_CE_HALT					(1 << 24)
+#define		CP_PFP_HALT					(1 << 26)
+#define		CP_ME_HALT					(1 << 28)
+
+#define	CP_COHER_CNTL2					0x85E8
+
+#define	CP_RB2_RPTR					0x86f8
+#define	CP_RB1_RPTR					0x86fc
+#define	CP_RB0_RPTR					0x8700
+#define	CP_RB_WPTR_DELAY				0x8704
+
+#define	CP_QUEUE_THRESHOLDS				0x8760
+#define		ROQ_IB1_START(x)				((x) << 0)
+#define		ROQ_IB2_START(x)				((x) << 8)
+#define CP_MEQ_THRESHOLDS				0x8764
+#define		MEQ1_START(x)				((x) << 0)
+#define		MEQ2_START(x)				((x) << 8)
+
+#define	CP_PERFMON_CNTL					0x87FC
+
+#define	VGT_VTX_VECT_EJECT_REG				0x88B0
+
+#define	VGT_CACHE_INVALIDATION				0x88C4
+#define		CACHE_INVALIDATION(x)				((x) << 0)
+#define			VC_ONLY						0
+#define			TC_ONLY						1
+#define			VC_AND_TC					2
+#define		AUTO_INVLD_EN(x)				((x) << 6)
+#define			NO_AUTO						0
+#define			ES_AUTO						1
+#define			GS_AUTO						2
+#define			ES_AND_GS_AUTO					3
+#define	VGT_ESGS_RING_SIZE				0x88C8
+#define	VGT_GSVS_RING_SIZE				0x88CC
+
+#define	VGT_GS_VERTEX_REUSE				0x88D4
+
+#define	VGT_PRIMITIVE_TYPE				0x8958
+#define	VGT_INDEX_TYPE					0x895C
+
+#define	VGT_NUM_INDICES					0x8970
+#define	VGT_NUM_INSTANCES				0x8974
+
+#define	VGT_TF_RING_SIZE				0x8988
+
+#define	VGT_HS_OFFCHIP_PARAM				0x89B0
+
+#define	VGT_TF_MEMORY_BASE				0x89B8
+
+#define CC_GC_SHADER_ARRAY_CONFIG			0x89bc
+#define GC_USER_SHADER_ARRAY_CONFIG			0x89c0
+
+#define	PA_CL_ENHANCE					0x8A14
+#define		CLIP_VTX_REORDER_ENA				(1 << 0)
+#define		NUM_CLIP_SEQ(x)					((x) << 1)
+
+#define	PA_SU_LINE_STIPPLE_VALUE			0x8A60
+
+#define	PA_SC_LINE_STIPPLE_STATE			0x8B10
+
+#define	PA_SC_FORCE_EOV_MAX_CNTS			0x8B24
+#define		FORCE_EOV_MAX_CLK_CNT(x)			((x) << 0)
+#define		FORCE_EOV_MAX_REZ_CNT(x)			((x) << 16)
+
+#define	PA_SC_FIFO_SIZE					0x8BCC
+#define		SC_FRONTEND_PRIM_FIFO_SIZE(x)			((x) << 0)
+#define		SC_BACKEND_PRIM_FIFO_SIZE(x)			((x) << 6)
+#define		SC_HIZ_TILE_FIFO_SIZE(x)			((x) << 15)
+#define		SC_EARLYZ_TILE_FIFO_SIZE(x)			((x) << 23)
+
+#define	PA_SC_ENHANCE					0x8BF0
+
+#define	SQ_CONFIG					0x8C00
+
+#define	SQC_CACHES					0x8C08
+
+#define	SX_DEBUG_1					0x9060
+
+#define	SPI_STATIC_THREAD_MGMT_1			0x90E0
+#define	SPI_STATIC_THREAD_MGMT_2			0x90E4
+#define	SPI_STATIC_THREAD_MGMT_3			0x90E8
+#define	SPI_PS_MAX_WAVE_ID				0x90EC
+
+#define	SPI_CONFIG_CNTL					0x9100
+
+#define	SPI_CONFIG_CNTL_1				0x913C
+#define		VTX_DONE_DELAY(x)				((x) << 0)
+#define		INTERP_ONE_PRIM_PER_ROW				(1 << 4)
+
+#define	CGTS_TCC_DISABLE				0x9148
+#define	CGTS_USER_TCC_DISABLE				0x914C
+#define		TCC_DISABLE_MASK				0xFFFF0000
+#define		TCC_DISABLE_SHIFT				16
+
+#define	TA_CNTL_AUX					0x9508
+
+#define CC_RB_BACKEND_DISABLE				0x98F4
+#define		BACKEND_DISABLE(x)     			((x) << 16)
+#define GB_ADDR_CONFIG  				0x98F8
+#define		NUM_PIPES(x)				((x) << 0)
+#define		NUM_PIPES_MASK				0x00000007
+#define		NUM_PIPES_SHIFT				0
+#define		PIPE_INTERLEAVE_SIZE(x)			((x) << 4)
+#define		PIPE_INTERLEAVE_SIZE_MASK		0x00000070
+#define		PIPE_INTERLEAVE_SIZE_SHIFT		4
+#define		NUM_SHADER_ENGINES(x)			((x) << 12)
+#define		NUM_SHADER_ENGINES_MASK			0x00003000
+#define		NUM_SHADER_ENGINES_SHIFT		12
+#define		SHADER_ENGINE_TILE_SIZE(x)     		((x) << 16)
+#define		SHADER_ENGINE_TILE_SIZE_MASK		0x00070000
+#define		SHADER_ENGINE_TILE_SIZE_SHIFT		16
+#define		NUM_GPUS(x)     			((x) << 20)
+#define		NUM_GPUS_MASK				0x00700000
+#define		NUM_GPUS_SHIFT				20
+#define		MULTI_GPU_TILE_SIZE(x)     		((x) << 24)
+#define		MULTI_GPU_TILE_SIZE_MASK		0x03000000
+#define		MULTI_GPU_TILE_SIZE_SHIFT		24
+#define		ROW_SIZE(x)             		((x) << 28)
+#define		ROW_SIZE_MASK				0x30000000
+#define		ROW_SIZE_SHIFT				28
+
+#define	GB_TILE_MODE0					0x9910
+#       define MICRO_TILE_MODE(x)				((x) << 0)
+#              define	ADDR_SURF_DISPLAY_MICRO_TILING		0
+#              define	ADDR_SURF_THIN_MICRO_TILING		1
+#              define	ADDR_SURF_DEPTH_MICRO_TILING		2
+#       define ARRAY_MODE(x)					((x) << 2)
+#              define	ARRAY_LINEAR_GENERAL			0
+#              define	ARRAY_LINEAR_ALIGNED			1
+#              define	ARRAY_1D_TILED_THIN1			2
+#              define	ARRAY_2D_TILED_THIN1			4
+#       define PIPE_CONFIG(x)					((x) << 6)
+#              define	ADDR_SURF_P2				0
+#              define	ADDR_SURF_P4_8x16			4
+#              define	ADDR_SURF_P4_16x16			5
+#              define	ADDR_SURF_P4_16x32			6
+#              define	ADDR_SURF_P4_32x32			7
+#              define	ADDR_SURF_P8_16x16_8x16			8
+#              define	ADDR_SURF_P8_16x32_8x16			9
+#              define	ADDR_SURF_P8_32x32_8x16			10
+#              define	ADDR_SURF_P8_16x32_16x16		11
+#              define	ADDR_SURF_P8_32x32_16x16		12
+#              define	ADDR_SURF_P8_32x32_16x32		13
+#              define	ADDR_SURF_P8_32x64_32x32		14
+#       define TILE_SPLIT(x)					((x) << 11)
+#              define	ADDR_SURF_TILE_SPLIT_64B		0
+#              define	ADDR_SURF_TILE_SPLIT_128B		1
+#              define	ADDR_SURF_TILE_SPLIT_256B		2
+#              define	ADDR_SURF_TILE_SPLIT_512B		3
+#              define	ADDR_SURF_TILE_SPLIT_1KB		4
+#              define	ADDR_SURF_TILE_SPLIT_2KB		5
+#              define	ADDR_SURF_TILE_SPLIT_4KB		6
+#       define BANK_WIDTH(x)					((x) << 14)
+#              define	ADDR_SURF_BANK_WIDTH_1			0
+#              define	ADDR_SURF_BANK_WIDTH_2			1
+#              define	ADDR_SURF_BANK_WIDTH_4			2
+#              define	ADDR_SURF_BANK_WIDTH_8			3
+#       define BANK_HEIGHT(x)					((x) << 16)
+#              define	ADDR_SURF_BANK_HEIGHT_1			0
+#              define	ADDR_SURF_BANK_HEIGHT_2			1
+#              define	ADDR_SURF_BANK_HEIGHT_4			2
+#              define	ADDR_SURF_BANK_HEIGHT_8			3
+#       define MACRO_TILE_ASPECT(x)				((x) << 18)
+#              define	ADDR_SURF_MACRO_ASPECT_1		0
+#              define	ADDR_SURF_MACRO_ASPECT_2		1
+#              define	ADDR_SURF_MACRO_ASPECT_4		2
+#              define	ADDR_SURF_MACRO_ASPECT_8		3
+#       define NUM_BANKS(x)					((x) << 20)
+#              define	ADDR_SURF_2_BANK			0
+#              define	ADDR_SURF_4_BANK			1
+#              define	ADDR_SURF_8_BANK			2
+#              define	ADDR_SURF_16_BANK			3
+
+#define	CB_PERFCOUNTER0_SELECT0				0x9a20
+#define	CB_PERFCOUNTER0_SELECT1				0x9a24
+#define	CB_PERFCOUNTER1_SELECT0				0x9a28
+#define	CB_PERFCOUNTER1_SELECT1				0x9a2c
+#define	CB_PERFCOUNTER2_SELECT0				0x9a30
+#define	CB_PERFCOUNTER2_SELECT1				0x9a34
+#define	CB_PERFCOUNTER3_SELECT0				0x9a38
+#define	CB_PERFCOUNTER3_SELECT1				0x9a3c
+
+#define	GC_USER_RB_BACKEND_DISABLE			0x9B7C
+#define		BACKEND_DISABLE_MASK			0x00FF0000
+#define		BACKEND_DISABLE_SHIFT			16
+
+#define	TCP_CHAN_STEER_LO				0xac0c
+#define	TCP_CHAN_STEER_HI				0xac10
+
+#define	CP_RB0_BASE					0xC100
+#define	CP_RB0_CNTL					0xC104
+#define		RB_BUFSZ(x)					((x) << 0)
+#define		RB_BLKSZ(x)					((x) << 8)
+#define		BUF_SWAP_32BIT					(2 << 16)
+#define		RB_NO_UPDATE					(1 << 27)
+#define		RB_RPTR_WR_ENA					(1 << 31)
+
+#define	CP_RB0_RPTR_ADDR				0xC10C
+#define	CP_RB0_RPTR_ADDR_HI				0xC110
+#define	CP_RB0_WPTR					0xC114
+
+#define	CP_PFP_UCODE_ADDR				0xC150
+#define	CP_PFP_UCODE_DATA				0xC154
+#define	CP_ME_RAM_RADDR					0xC158
+#define	CP_ME_RAM_WADDR					0xC15C
+#define	CP_ME_RAM_DATA					0xC160
+
+#define	CP_CE_UCODE_ADDR				0xC168
+#define	CP_CE_UCODE_DATA				0xC16C
+
+#define	CP_RB1_BASE					0xC180
+#define	CP_RB1_CNTL					0xC184
+#define	CP_RB1_RPTR_ADDR				0xC188
+#define	CP_RB1_RPTR_ADDR_HI				0xC18C
+#define	CP_RB1_WPTR					0xC190
+#define	CP_RB2_BASE					0xC194
+#define	CP_RB2_CNTL					0xC198
+#define	CP_RB2_RPTR_ADDR				0xC19C
+#define	CP_RB2_RPTR_ADDR_HI				0xC1A0
+#define	CP_RB2_WPTR					0xC1A4
+#define CP_INT_CNTL_RING0                               0xC1A8
+#define CP_INT_CNTL_RING1                               0xC1AC
+#define CP_INT_CNTL_RING2                               0xC1B0
+#       define CNTX_BUSY_INT_ENABLE                     (1 << 19)
+#       define CNTX_EMPTY_INT_ENABLE                    (1 << 20)
+#       define WAIT_MEM_SEM_INT_ENABLE                  (1 << 21)
+#       define TIME_STAMP_INT_ENABLE                    (1 << 26)
+#       define CP_RINGID2_INT_ENABLE                    (1 << 29)
+#       define CP_RINGID1_INT_ENABLE                    (1 << 30)
+#       define CP_RINGID0_INT_ENABLE                    (1 << 31)
+#define CP_INT_STATUS_RING0                             0xC1B4
+#define CP_INT_STATUS_RING1                             0xC1B8
+#define CP_INT_STATUS_RING2                             0xC1BC
+#       define WAIT_MEM_SEM_INT_STAT                    (1 << 21)
+#       define TIME_STAMP_INT_STAT                      (1 << 26)
+#       define CP_RINGID2_INT_STAT                      (1 << 29)
+#       define CP_RINGID1_INT_STAT                      (1 << 30)
+#       define CP_RINGID0_INT_STAT                      (1 << 31)
+
+#define	CP_DEBUG					0xC1FC
+
+#define RLC_CNTL                                          0xC300
+#       define RLC_ENABLE                                 (1 << 0)
+#define RLC_RL_BASE                                       0xC304
+#define RLC_RL_SIZE                                       0xC308
+#define RLC_LB_CNTL                                       0xC30C
+#define RLC_SAVE_AND_RESTORE_BASE                         0xC310
+#define RLC_LB_CNTR_MAX                                   0xC314
+#define RLC_LB_CNTR_INIT                                  0xC318
+
+#define RLC_CLEAR_STATE_RESTORE_BASE                      0xC320
+
+#define RLC_UCODE_ADDR                                    0xC32C
+#define RLC_UCODE_DATA                                    0xC330
+
+#define RLC_MC_CNTL                                       0xC344
+#define RLC_UCODE_CNTL                                    0xC348
+
+#define VGT_EVENT_INITIATOR                             0x28a90
+#       define SAMPLE_STREAMOUTSTATS1                   (1 << 0)
+#       define SAMPLE_STREAMOUTSTATS2                   (2 << 0)
+#       define SAMPLE_STREAMOUTSTATS3                   (3 << 0)
+#       define CACHE_FLUSH_TS                           (4 << 0)
+#       define CACHE_FLUSH                              (6 << 0)
+#       define CS_PARTIAL_FLUSH                         (7 << 0)
+#       define VGT_STREAMOUT_RESET                      (10 << 0)
+#       define END_OF_PIPE_INCR_DE                      (11 << 0)
+#       define END_OF_PIPE_IB_END                       (12 << 0)
+#       define RST_PIX_CNT                              (13 << 0)
+#       define VS_PARTIAL_FLUSH                         (15 << 0)
+#       define PS_PARTIAL_FLUSH                         (16 << 0)
+#       define CACHE_FLUSH_AND_INV_TS_EVENT             (20 << 0)
+#       define ZPASS_DONE                               (21 << 0)
+#       define CACHE_FLUSH_AND_INV_EVENT                (22 << 0)
+#       define PERFCOUNTER_START                        (23 << 0)
+#       define PERFCOUNTER_STOP                         (24 << 0)
+#       define PIPELINESTAT_START                       (25 << 0)
+#       define PIPELINESTAT_STOP                        (26 << 0)
+#       define PERFCOUNTER_SAMPLE                       (27 << 0)
+#       define SAMPLE_PIPELINESTAT                      (30 << 0)
+#       define SAMPLE_STREAMOUTSTATS                    (32 << 0)
+#       define RESET_VTX_CNT                            (33 << 0)
+#       define VGT_FLUSH                                (36 << 0)
+#       define BOTTOM_OF_PIPE_TS                        (40 << 0)
+#       define DB_CACHE_FLUSH_AND_INV                   (42 << 0)
+#       define FLUSH_AND_INV_DB_DATA_TS                 (43 << 0)
+#       define FLUSH_AND_INV_DB_META                    (44 << 0)
+#       define FLUSH_AND_INV_CB_DATA_TS                 (45 << 0)
+#       define FLUSH_AND_INV_CB_META                    (46 << 0)
+#       define CS_DONE                                  (47 << 0)
+#       define PS_DONE                                  (48 << 0)
+#       define FLUSH_AND_INV_CB_PIXEL_DATA              (49 << 0)
+#       define THREAD_TRACE_START                       (51 << 0)
+#       define THREAD_TRACE_STOP                        (52 << 0)
+#       define THREAD_TRACE_FLUSH                       (54 << 0)
+#       define THREAD_TRACE_FINISH                      (55 << 0)
+
+/*
+ * PM4
+ */
+#define	PACKET_TYPE0	0
+#define	PACKET_TYPE1	1
+#define	PACKET_TYPE2	2
+#define	PACKET_TYPE3	3
+
+#define CP_PACKET_GET_TYPE(h) (((h) >> 30) & 3)
+#define CP_PACKET_GET_COUNT(h) (((h) >> 16) & 0x3FFF)
+#define CP_PACKET0_GET_REG(h) (((h) & 0xFFFF) << 2)
+#define CP_PACKET3_GET_OPCODE(h) (((h) >> 8) & 0xFF)
+#define PACKET0(reg, n)	((PACKET_TYPE0 << 30) |				\
+			 (((reg) >> 2) & 0xFFFF) |			\
+			 ((n) & 0x3FFF) << 16)
+#define CP_PACKET2			0x80000000
+#define		PACKET2_PAD_SHIFT		0
+#define		PACKET2_PAD_MASK		(0x3fffffff << 0)
+
+#define PACKET2(v)	(CP_PACKET2 | REG_SET(PACKET2_PAD, (v)))
+
+#define PACKET3(op, n)	((PACKET_TYPE3 << 30) |				\
+			 (((op) & 0xFF) << 8) |				\
+			 ((n) & 0x3FFF) << 16)
+
+#define PACKET3_COMPUTE(op, n) (PACKET3(op, n) | 1 << 1)
+
+/* Packet 3 types */
+#define	PACKET3_NOP					0x10
+#define	PACKET3_SET_BASE				0x11
+#define		PACKET3_BASE_INDEX(x)                  ((x) << 0)
+#define			GDS_PARTITION_BASE		2
+#define			CE_PARTITION_BASE		3
+#define	PACKET3_CLEAR_STATE				0x12
+#define	PACKET3_INDEX_BUFFER_SIZE			0x13
+#define	PACKET3_DISPATCH_DIRECT				0x15
+#define	PACKET3_DISPATCH_INDIRECT			0x16
+#define	PACKET3_ALLOC_GDS				0x1B
+#define	PACKET3_WRITE_GDS_RAM				0x1C
+#define	PACKET3_ATOMIC_GDS				0x1D
+#define	PACKET3_ATOMIC					0x1E
+#define	PACKET3_OCCLUSION_QUERY				0x1F
+#define	PACKET3_SET_PREDICATION				0x20
+#define	PACKET3_REG_RMW					0x21
+#define	PACKET3_COND_EXEC				0x22
+#define	PACKET3_PRED_EXEC				0x23
+#define	PACKET3_DRAW_INDIRECT				0x24
+#define	PACKET3_DRAW_INDEX_INDIRECT			0x25
+#define	PACKET3_INDEX_BASE				0x26
+#define	PACKET3_DRAW_INDEX_2				0x27
+#define	PACKET3_CONTEXT_CONTROL				0x28
+#define	PACKET3_INDEX_TYPE				0x2A
+#define	PACKET3_DRAW_INDIRECT_MULTI			0x2C
+#define	PACKET3_DRAW_INDEX_AUTO				0x2D
+#define	PACKET3_DRAW_INDEX_IMMD				0x2E
+#define	PACKET3_NUM_INSTANCES				0x2F
+#define	PACKET3_DRAW_INDEX_MULTI_AUTO			0x30
+#define	PACKET3_INDIRECT_BUFFER_CONST			0x31
+#define	PACKET3_INDIRECT_BUFFER				0x32
+#define	PACKET3_STRMOUT_BUFFER_UPDATE			0x34
+#define	PACKET3_DRAW_INDEX_OFFSET_2			0x35
+#define	PACKET3_DRAW_INDEX_MULTI_ELEMENT		0x36
+#define	PACKET3_WRITE_DATA				0x37
+#define	PACKET3_DRAW_INDEX_INDIRECT_MULTI		0x38
+#define	PACKET3_MEM_SEMAPHORE				0x39
+#define	PACKET3_MPEG_INDEX				0x3A
+#define	PACKET3_COPY_DW					0x3B
+#define	PACKET3_WAIT_REG_MEM				0x3C
+#define	PACKET3_MEM_WRITE				0x3D
+#define	PACKET3_COPY_DATA				0x40
+#define	PACKET3_PFP_SYNC_ME				0x42
+#define	PACKET3_SURFACE_SYNC				0x43
+#              define PACKET3_DEST_BASE_0_ENA      (1 << 0)
+#              define PACKET3_DEST_BASE_1_ENA      (1 << 1)
+#              define PACKET3_CB0_DEST_BASE_ENA    (1 << 6)
+#              define PACKET3_CB1_DEST_BASE_ENA    (1 << 7)
+#              define PACKET3_CB2_DEST_BASE_ENA    (1 << 8)
+#              define PACKET3_CB3_DEST_BASE_ENA    (1 << 9)
+#              define PACKET3_CB4_DEST_BASE_ENA    (1 << 10)
+#              define PACKET3_CB5_DEST_BASE_ENA    (1 << 11)
+#              define PACKET3_CB6_DEST_BASE_ENA    (1 << 12)
+#              define PACKET3_CB7_DEST_BASE_ENA    (1 << 13)
+#              define PACKET3_DB_DEST_BASE_ENA     (1 << 14)
+#              define PACKET3_DEST_BASE_2_ENA      (1 << 19)
+#              define PACKET3_DEST_BASE_3_ENA      (1 << 21)
+#              define PACKET3_TCL1_ACTION_ENA      (1 << 22)
+#              define PACKET3_TC_ACTION_ENA        (1 << 23)
+#              define PACKET3_CB_ACTION_ENA        (1 << 25)
+#              define PACKET3_DB_ACTION_ENA        (1 << 26)
+#              define PACKET3_SH_KCACHE_ACTION_ENA (1 << 27)
+#              define PACKET3_SH_ICACHE_ACTION_ENA (1 << 29)
+#define	PACKET3_ME_INITIALIZE				0x44
+#define		PACKET3_ME_INITIALIZE_DEVICE_ID(x) ((x) << 16)
+#define	PACKET3_COND_WRITE				0x45
+#define	PACKET3_EVENT_WRITE				0x46
+#define		EVENT_TYPE(x)                           ((x) << 0)
+#define		EVENT_INDEX(x)                          ((x) << 8)
+                /* 0 - any non-TS event
+		 * 1 - ZPASS_DONE
+		 * 2 - SAMPLE_PIPELINESTAT
+		 * 3 - SAMPLE_STREAMOUTSTAT*
+		 * 4 - *S_PARTIAL_FLUSH
+		 * 5 - EOP events
+		 * 6 - EOS events
+		 * 7 - CACHE_FLUSH, CACHE_FLUSH_AND_INV_EVENT
+		 */
+#define		INV_L2                                  (1 << 20)
+                /* INV TC L2 cache when EVENT_INDEX = 7 */
+#define	PACKET3_EVENT_WRITE_EOP				0x47
+#define		DATA_SEL(x)                             ((x) << 29)
+                /* 0 - discard
+		 * 1 - send low 32bit data
+		 * 2 - send 64bit data
+		 * 3 - send 64bit counter value
+		 */
+#define		INT_SEL(x)                              ((x) << 24)
+                /* 0 - none
+		 * 1 - interrupt only (DATA_SEL = 0)
+		 * 2 - interrupt when data write is confirmed
+		 */
+#define	PACKET3_EVENT_WRITE_EOS				0x48
+#define	PACKET3_PREAMBLE_CNTL				0x4A
+#              define PACKET3_PREAMBLE_BEGIN_CLEAR_STATE     (2 << 28)
+#              define PACKET3_PREAMBLE_END_CLEAR_STATE       (3 << 28)
+#define	PACKET3_ONE_REG_WRITE				0x57
+#define	PACKET3_LOAD_CONFIG_REG				0x5F
+#define	PACKET3_LOAD_CONTEXT_REG			0x60
+#define	PACKET3_LOAD_SH_REG				0x61
+#define	PACKET3_SET_CONFIG_REG				0x68
+#define		PACKET3_SET_CONFIG_REG_START			0x00008000
+#define		PACKET3_SET_CONFIG_REG_END			0x0000b000
+#define	PACKET3_SET_CONTEXT_REG				0x69
+#define		PACKET3_SET_CONTEXT_REG_START			0x00028000
+#define		PACKET3_SET_CONTEXT_REG_END			0x00029000
+#define	PACKET3_SET_CONTEXT_REG_INDIRECT		0x73
+#define	PACKET3_SET_RESOURCE_INDIRECT			0x74
+#define	PACKET3_SET_SH_REG				0x76
+#define		PACKET3_SET_SH_REG_START			0x0000b000
+#define		PACKET3_SET_SH_REG_END				0x0000c000
+#define	PACKET3_SET_SH_REG_OFFSET			0x77
+#define	PACKET3_ME_WRITE				0x7A
+#define	PACKET3_SCRATCH_RAM_WRITE			0x7D
+#define	PACKET3_SCRATCH_RAM_READ			0x7E
+#define	PACKET3_CE_WRITE				0x7F
+#define	PACKET3_LOAD_CONST_RAM				0x80
+#define	PACKET3_WRITE_CONST_RAM				0x81
+#define	PACKET3_WRITE_CONST_RAM_OFFSET			0x82
+#define	PACKET3_DUMP_CONST_RAM				0x83
+#define	PACKET3_INCREMENT_CE_COUNTER			0x84
+#define	PACKET3_INCREMENT_DE_COUNTER			0x85
+#define	PACKET3_WAIT_ON_CE_COUNTER			0x86
+#define	PACKET3_WAIT_ON_DE_COUNTER			0x87
+#define	PACKET3_WAIT_ON_DE_COUNTER_DIFF			0x88
+#define	PACKET3_SET_CE_DE_COUNTERS			0x89
+#define	PACKET3_WAIT_ON_AVAIL_BUFFER			0x8A
+
+#endif
diff --git a/drivers/gpu/drm/savage/savage_state.c b/drivers/gpu/drm/savage/savage_state.c
index 8a3e31599c94..031aaaf79ac2 100644
--- a/drivers/gpu/drm/savage/savage_state.c
+++ b/drivers/gpu/drm/savage/savage_state.c
@@ -1057,7 +1057,8 @@ int savage_bci_cmdbuf(struct drm_device *dev, void *data, struct drm_file *file_
 				DRM_ERROR("indexed drawing command extends "
 					  "beyond end of command buffer\n");
 				DMA_FLUSH();
-				return -EINVAL;
+				ret = -EINVAL;
+				goto done;
 			}
 			/* fall through */
 		case SAVAGE_CMD_DMA_PRIM:
@@ -1076,7 +1077,7 @@ int savage_bci_cmdbuf(struct drm_device *dev, void *data, struct drm_file *file_
 				      cmdbuf->vb_stride,
 				      cmdbuf->nbox, cmdbuf->box_addr);
 				if (ret != 0)
-					return ret;
+					goto done;
 				first_draw_cmd = NULL;
 			}
 		}
diff --git a/drivers/gpu/drm/sis/sis_drv.c b/drivers/gpu/drm/sis/sis_drv.c
index 573220cc5269..30d98d14b5c5 100644
--- a/drivers/gpu/drm/sis/sis_drv.c
+++ b/drivers/gpu/drm/sis/sis_drv.c
@@ -41,6 +41,8 @@ static int sis_driver_load(struct drm_device *dev, unsigned long chipset)
 {
 	drm_sis_private_t *dev_priv;
 
+	pci_set_master(dev->pdev);
+
 	dev_priv = kzalloc(sizeof(drm_sis_private_t), GFP_KERNEL);
 	if (dev_priv == NULL)
 		return -ENOMEM;
diff --git a/drivers/gpu/drm/ttm/ttm_agp_backend.c b/drivers/gpu/drm/ttm/ttm_agp_backend.c
index 747c1413fc95..4a8728291361 100644
--- a/drivers/gpu/drm/ttm/ttm_agp_backend.c
+++ b/drivers/gpu/drm/ttm/ttm_agp_backend.c
@@ -29,6 +29,8 @@
  *          Keith Packard.
  */
 
+#define pr_fmt(fmt) "[TTM] " fmt
+
 #include "ttm/ttm_module.h"
 #include "ttm/ttm_bo_driver.h"
 #include "ttm/ttm_page_alloc.h"
@@ -74,7 +76,7 @@ static int ttm_agp_bind(struct ttm_tt *ttm, struct ttm_mem_reg *bo_mem)
 
 	ret = agp_bind_memory(mem, node->start);
 	if (ret)
-		printk(KERN_ERR TTM_PFX "AGP Bind memory failed.\n");
+		pr_err("AGP Bind memory failed\n");
 
 	return ret;
 }
diff --git a/drivers/gpu/drm/ttm/ttm_bo.c b/drivers/gpu/drm/ttm/ttm_bo.c
index 7c3a57de8187..1f5c67c579cf 100644
--- a/drivers/gpu/drm/ttm/ttm_bo.c
+++ b/drivers/gpu/drm/ttm/ttm_bo.c
@@ -28,6 +28,8 @@
  * Authors: Thomas Hellstrom <thellstrom-at-vmware-dot-com>
  */
 
+#define pr_fmt(fmt) "[TTM] " fmt
+
 #include "ttm/ttm_module.h"
 #include "ttm/ttm_bo_driver.h"
 #include "ttm/ttm_placement.h"
@@ -68,15 +70,13 @@ static void ttm_mem_type_debug(struct ttm_bo_device *bdev, int mem_type)
 {
 	struct ttm_mem_type_manager *man = &bdev->man[mem_type];
 
-	printk(KERN_ERR TTM_PFX "    has_type: %d\n", man->has_type);
-	printk(KERN_ERR TTM_PFX "    use_type: %d\n", man->use_type);
-	printk(KERN_ERR TTM_PFX "    flags: 0x%08X\n", man->flags);
-	printk(KERN_ERR TTM_PFX "    gpu_offset: 0x%08lX\n", man->gpu_offset);
-	printk(KERN_ERR TTM_PFX "    size: %llu\n", man->size);
-	printk(KERN_ERR TTM_PFX "    available_caching: 0x%08X\n",
-		man->available_caching);
-	printk(KERN_ERR TTM_PFX "    default_caching: 0x%08X\n",
-		man->default_caching);
+	pr_err("    has_type: %d\n", man->has_type);
+	pr_err("    use_type: %d\n", man->use_type);
+	pr_err("    flags: 0x%08X\n", man->flags);
+	pr_err("    gpu_offset: 0x%08lX\n", man->gpu_offset);
+	pr_err("    size: %llu\n", man->size);
+	pr_err("    available_caching: 0x%08X\n", man->available_caching);
+	pr_err("    default_caching: 0x%08X\n", man->default_caching);
 	if (mem_type != TTM_PL_SYSTEM)
 		(*man->func->debug)(man, TTM_PFX);
 }
@@ -86,16 +86,16 @@ static void ttm_bo_mem_space_debug(struct ttm_buffer_object *bo,
 {
 	int i, ret, mem_type;
 
-	printk(KERN_ERR TTM_PFX "No space for %p (%lu pages, %luK, %luM)\n",
-		bo, bo->mem.num_pages, bo->mem.size >> 10,
-		bo->mem.size >> 20);
+	pr_err("No space for %p (%lu pages, %luK, %luM)\n",
+	       bo, bo->mem.num_pages, bo->mem.size >> 10,
+	       bo->mem.size >> 20);
 	for (i = 0; i < placement->num_placement; i++) {
 		ret = ttm_mem_type_from_flags(placement->placement[i],
 						&mem_type);
 		if (ret)
 			return;
-		printk(KERN_ERR TTM_PFX "  placement[%d]=0x%08X (%d)\n",
-			i, placement->placement[i], mem_type);
+		pr_err("  placement[%d]=0x%08X (%d)\n",
+		       i, placement->placement[i], mem_type);
 		ttm_mem_type_debug(bo->bdev, mem_type);
 	}
 }
@@ -344,7 +344,7 @@ static int ttm_bo_add_ttm(struct ttm_buffer_object *bo, bool zero_alloc)
 			ret = -ENOMEM;
 		break;
 	default:
-		printk(KERN_ERR TTM_PFX "Illegal buffer object type\n");
+		pr_err("Illegal buffer object type\n");
 		ret = -EINVAL;
 		break;
 	}
@@ -432,7 +432,7 @@ moved:
 	if (bo->evicted) {
 		ret = bdev->driver->invalidate_caches(bdev, bo->mem.placement);
 		if (ret)
-			printk(KERN_ERR TTM_PFX "Can not flush read caches\n");
+			pr_err("Can not flush read caches\n");
 		bo->evicted = false;
 	}
 
@@ -734,9 +734,7 @@ static int ttm_bo_evict(struct ttm_buffer_object *bo, bool interruptible,
 
 	if (unlikely(ret != 0)) {
 		if (ret != -ERESTARTSYS) {
-			printk(KERN_ERR TTM_PFX
-			       "Failed to expire sync object before "
-			       "buffer eviction.\n");
+			pr_err("Failed to expire sync object before buffer eviction\n");
 		}
 		goto out;
 	}
@@ -757,9 +755,8 @@ static int ttm_bo_evict(struct ttm_buffer_object *bo, bool interruptible,
 				no_wait_reserve, no_wait_gpu);
 	if (ret) {
 		if (ret != -ERESTARTSYS) {
-			printk(KERN_ERR TTM_PFX
-			       "Failed to find memory space for "
-			       "buffer 0x%p eviction.\n", bo);
+			pr_err("Failed to find memory space for buffer 0x%p eviction\n",
+			       bo);
 			ttm_bo_mem_space_debug(bo, &placement);
 		}
 		goto out;
@@ -769,7 +766,7 @@ static int ttm_bo_evict(struct ttm_buffer_object *bo, bool interruptible,
 				     no_wait_reserve, no_wait_gpu);
 	if (ret) {
 		if (ret != -ERESTARTSYS)
-			printk(KERN_ERR TTM_PFX "Buffer eviction failed\n");
+			pr_err("Buffer eviction failed\n");
 		ttm_bo_mem_put(bo, &evict_mem);
 		goto out;
 	}
@@ -1180,7 +1177,7 @@ int ttm_bo_init(struct ttm_bo_device *bdev,
 
 	ret = ttm_mem_global_alloc(mem_glob, acc_size, false, false);
 	if (ret) {
-		printk(KERN_ERR TTM_PFX "Out of kernel memory.\n");
+		pr_err("Out of kernel memory\n");
 		if (destroy)
 			(*destroy)(bo);
 		else
@@ -1191,7 +1188,7 @@ int ttm_bo_init(struct ttm_bo_device *bdev,
 	size += buffer_start & ~PAGE_MASK;
 	num_pages = (size + PAGE_SIZE - 1) >> PAGE_SHIFT;
 	if (num_pages == 0) {
-		printk(KERN_ERR TTM_PFX "Illegal buffer object size.\n");
+		pr_err("Illegal buffer object size\n");
 		if (destroy)
 			(*destroy)(bo);
 		else
@@ -1342,8 +1339,7 @@ static int ttm_bo_force_list_clean(struct ttm_bo_device *bdev,
 			if (allow_errors) {
 				return ret;
 			} else {
-				printk(KERN_ERR TTM_PFX
-					"Cleanup eviction failed\n");
+				pr_err("Cleanup eviction failed\n");
 			}
 		}
 		spin_lock(&glob->lru_lock);
@@ -1358,14 +1354,14 @@ int ttm_bo_clean_mm(struct ttm_bo_device *bdev, unsigned mem_type)
 	int ret = -EINVAL;
 
 	if (mem_type >= TTM_NUM_MEM_TYPES) {
-		printk(KERN_ERR TTM_PFX "Illegal memory type %d\n", mem_type);
+		pr_err("Illegal memory type %d\n", mem_type);
 		return ret;
 	}
 	man = &bdev->man[mem_type];
 
 	if (!man->has_type) {
-		printk(KERN_ERR TTM_PFX "Trying to take down uninitialized "
-		       "memory manager type %u\n", mem_type);
+		pr_err("Trying to take down uninitialized memory manager type %u\n",
+		       mem_type);
 		return ret;
 	}
 
@@ -1388,16 +1384,12 @@ int ttm_bo_evict_mm(struct ttm_bo_device *bdev, unsigned mem_type)
 	struct ttm_mem_type_manager *man = &bdev->man[mem_type];
 
 	if (mem_type == 0 || mem_type >= TTM_NUM_MEM_TYPES) {
-		printk(KERN_ERR TTM_PFX
-		       "Illegal memory manager memory type %u.\n",
-		       mem_type);
+		pr_err("Illegal memory manager memory type %u\n", mem_type);
 		return -EINVAL;
 	}
 
 	if (!man->has_type) {
-		printk(KERN_ERR TTM_PFX
-		       "Memory type %u has not been initialized.\n",
-		       mem_type);
+		pr_err("Memory type %u has not been initialized\n", mem_type);
 		return 0;
 	}
 
@@ -1482,8 +1474,7 @@ int ttm_bo_global_init(struct drm_global_reference *ref)
 	ttm_mem_init_shrink(&glob->shrink, ttm_bo_swapout);
 	ret = ttm_mem_register_shrink(glob->mem_glob, &glob->shrink);
 	if (unlikely(ret != 0)) {
-		printk(KERN_ERR TTM_PFX
-		       "Could not register buffer object swapout.\n");
+		pr_err("Could not register buffer object swapout\n");
 		goto out_no_shrink;
 	}
 
@@ -1516,9 +1507,8 @@ int ttm_bo_device_release(struct ttm_bo_device *bdev)
 			man->use_type = false;
 			if ((i != TTM_PL_SYSTEM) && ttm_bo_clean_mm(bdev, i)) {
 				ret = -EBUSY;
-				printk(KERN_ERR TTM_PFX
-				       "DRM memory manager type %d "
-				       "is not clean.\n", i);
+				pr_err("DRM memory manager type %d is not clean\n",
+				       i);
 			}
 			man->has_type = false;
 		}
diff --git a/drivers/gpu/drm/ttm/ttm_bo_vm.c b/drivers/gpu/drm/ttm/ttm_bo_vm.c
index 54412848de88..a877813571a4 100644
--- a/drivers/gpu/drm/ttm/ttm_bo_vm.c
+++ b/drivers/gpu/drm/ttm/ttm_bo_vm.c
@@ -28,6 +28,8 @@
  * Authors: Thomas Hellstrom <thellstrom-at-vmware-dot-com>
  */
 
+#define pr_fmt(fmt) "[TTM] " fmt
+
 #include <ttm/ttm_module.h>
 #include <ttm/ttm_bo_driver.h>
 #include <ttm/ttm_placement.h>
@@ -262,8 +264,7 @@ int ttm_bo_mmap(struct file *filp, struct vm_area_struct *vma,
 	read_unlock(&bdev->vm_lock);
 
 	if (unlikely(bo == NULL)) {
-		printk(KERN_ERR TTM_PFX
-		       "Could not find buffer object to map.\n");
+		pr_err("Could not find buffer object to map\n");
 		return -EINVAL;
 	}
 
diff --git a/drivers/gpu/drm/ttm/ttm_memory.c b/drivers/gpu/drm/ttm/ttm_memory.c
index 9eba8e9a4e9c..23d2ecbaed59 100644
--- a/drivers/gpu/drm/ttm/ttm_memory.c
+++ b/drivers/gpu/drm/ttm/ttm_memory.c
@@ -25,6 +25,8 @@
  *
  **************************************************************************/
 
+#define pr_fmt(fmt) "[TTM] " fmt
+
 #include "ttm/ttm_memory.h"
 #include "ttm/ttm_module.h"
 #include "ttm/ttm_page_alloc.h"
@@ -74,9 +76,8 @@ static void ttm_mem_zone_kobj_release(struct kobject *kobj)
 	struct ttm_mem_zone *zone =
 		container_of(kobj, struct ttm_mem_zone, kobj);
 
-	printk(KERN_INFO TTM_PFX
-	       "Zone %7s: Used memory at exit: %llu kiB.\n",
-	       zone->name, (unsigned long long) zone->used_mem >> 10);
+	pr_info("Zone %7s: Used memory at exit: %llu kiB\n",
+		zone->name, (unsigned long long)zone->used_mem >> 10);
 	kfree(zone);
 }
 
@@ -390,9 +391,8 @@ int ttm_mem_global_init(struct ttm_mem_global *glob)
 #endif
 	for (i = 0; i < glob->num_zones; ++i) {
 		zone = glob->zones[i];
-		printk(KERN_INFO TTM_PFX
-		       "Zone %7s: Available graphics memory: %llu kiB.\n",
-		       zone->name, (unsigned long long) zone->max_mem >> 10);
+		pr_info("Zone %7s: Available graphics memory: %llu kiB\n",
+			zone->name, (unsigned long long)zone->max_mem >> 10);
 	}
 	ttm_page_alloc_init(glob, glob->zone_kernel->max_mem/(2*PAGE_SIZE));
 	ttm_dma_page_alloc_init(glob, glob->zone_kernel->max_mem/(2*PAGE_SIZE));
diff --git a/drivers/gpu/drm/ttm/ttm_object.c b/drivers/gpu/drm/ttm/ttm_object.c
index 93577f2e2954..68daca412cbd 100644
--- a/drivers/gpu/drm/ttm/ttm_object.c
+++ b/drivers/gpu/drm/ttm/ttm_object.c
@@ -49,6 +49,8 @@
  * for fast lookup of ref objects given a base object.
  */
 
+#define pr_fmt(fmt) "[TTM] " fmt
+
 #include "ttm/ttm_object.h"
 #include "ttm/ttm_module.h"
 #include <linux/list.h>
@@ -232,8 +234,7 @@ struct ttm_base_object *ttm_base_object_lookup(struct ttm_object_file *tfile,
 		return NULL;
 
 	if (tfile != base->tfile && !base->shareable) {
-		printk(KERN_ERR TTM_PFX
-		       "Attempted access of non-shareable object.\n");
+		pr_err("Attempted access of non-shareable object\n");
 		ttm_base_object_unref(&base);
 		return NULL;
 	}
diff --git a/drivers/gpu/drm/ttm/ttm_page_alloc.c b/drivers/gpu/drm/ttm/ttm_page_alloc.c
index 499debda791e..ebc6fac96e36 100644
--- a/drivers/gpu/drm/ttm/ttm_page_alloc.c
+++ b/drivers/gpu/drm/ttm/ttm_page_alloc.c
@@ -30,6 +30,9 @@
  * - Use page->lru to keep a free list
  * - doesn't track currently in use pages
  */
+
+#define pr_fmt(fmt) "[TTM] " fmt
+
 #include <linux/list.h>
 #include <linux/spinlock.h>
 #include <linux/highmem.h>
@@ -167,18 +170,13 @@ static ssize_t ttm_pool_store(struct kobject *kobj,
 		m->options.small = val;
 	else if (attr == &ttm_page_pool_alloc_size) {
 		if (val > NUM_PAGES_TO_ALLOC*8) {
-			printk(KERN_ERR TTM_PFX
-			       "Setting allocation size to %lu "
-			       "is not allowed. Recommended size is "
-			       "%lu\n",
+			pr_err("Setting allocation size to %lu is not allowed. Recommended size is %lu\n",
 			       NUM_PAGES_TO_ALLOC*(PAGE_SIZE >> 7),
 			       NUM_PAGES_TO_ALLOC*(PAGE_SIZE >> 10));
 			return size;
 		} else if (val > NUM_PAGES_TO_ALLOC) {
-			printk(KERN_WARNING TTM_PFX
-			       "Setting allocation size to "
-			       "larger than %lu is not recommended.\n",
-			       NUM_PAGES_TO_ALLOC*(PAGE_SIZE >> 10));
+			pr_warn("Setting allocation size to larger than %lu is not recommended\n",
+				NUM_PAGES_TO_ALLOC*(PAGE_SIZE >> 10));
 		}
 		m->options.alloc_size = val;
 	}
@@ -279,8 +277,7 @@ static void ttm_pages_put(struct page *pages[], unsigned npages)
 {
 	unsigned i;
 	if (set_pages_array_wb(pages, npages))
-		printk(KERN_ERR TTM_PFX "Failed to set %d pages to wb!\n",
-				npages);
+		pr_err("Failed to set %d pages to wb!\n", npages);
 	for (i = 0; i < npages; ++i)
 		__free_page(pages[i]);
 }
@@ -315,8 +312,7 @@ static int ttm_page_pool_free(struct ttm_page_pool *pool, unsigned nr_free)
 	pages_to_free = kmalloc(npages_to_free * sizeof(struct page *),
 			GFP_KERNEL);
 	if (!pages_to_free) {
-		printk(KERN_ERR TTM_PFX
-		       "Failed to allocate memory for pool free operation.\n");
+		pr_err("Failed to allocate memory for pool free operation\n");
 		return 0;
 	}
 
@@ -438,16 +434,12 @@ static int ttm_set_pages_caching(struct page **pages,
 	case tt_uncached:
 		r = set_pages_array_uc(pages, cpages);
 		if (r)
-			printk(KERN_ERR TTM_PFX
-			       "Failed to set %d pages to uc!\n",
-			       cpages);
+			pr_err("Failed to set %d pages to uc!\n", cpages);
 		break;
 	case tt_wc:
 		r = set_pages_array_wc(pages, cpages);
 		if (r)
-			printk(KERN_ERR TTM_PFX
-			       "Failed to set %d pages to wc!\n",
-			       cpages);
+			pr_err("Failed to set %d pages to wc!\n", cpages);
 		break;
 	default:
 		break;
@@ -492,8 +484,7 @@ static int ttm_alloc_new_pages(struct list_head *pages, gfp_t gfp_flags,
 	caching_array = kmalloc(max_cpages*sizeof(struct page *), GFP_KERNEL);
 
 	if (!caching_array) {
-		printk(KERN_ERR TTM_PFX
-		       "Unable to allocate table for new pages.");
+		pr_err("Unable to allocate table for new pages\n");
 		return -ENOMEM;
 	}
 
@@ -501,7 +492,7 @@ static int ttm_alloc_new_pages(struct list_head *pages, gfp_t gfp_flags,
 		p = alloc_page(gfp_flags);
 
 		if (!p) {
-			printk(KERN_ERR TTM_PFX "Unable to get page %u.\n", i);
+			pr_err("Unable to get page %u\n", i);
 
 			/* store already allocated pages in the pool after
 			 * setting the caching state */
@@ -599,8 +590,7 @@ static void ttm_page_pool_fill_locked(struct ttm_page_pool *pool,
 			++pool->nrefills;
 			pool->npages += alloc_size;
 		} else {
-			printk(KERN_ERR TTM_PFX
-			       "Failed to fill pool (%p).", pool);
+			pr_err("Failed to fill pool (%p)\n", pool);
 			/* If we have any pages left put them to the pool. */
 			list_for_each_entry(p, &pool->list, lru) {
 				++cpages;
@@ -675,9 +665,7 @@ static void ttm_put_pages(struct page **pages, unsigned npages, int flags,
 		for (i = 0; i < npages; i++) {
 			if (pages[i]) {
 				if (page_count(pages[i]) != 1)
-					printk(KERN_ERR TTM_PFX
-					       "Erroneous page count. "
-					       "Leaking pages.\n");
+					pr_err("Erroneous page count. Leaking pages.\n");
 				__free_page(pages[i]);
 				pages[i] = NULL;
 			}
@@ -689,9 +677,7 @@ static void ttm_put_pages(struct page **pages, unsigned npages, int flags,
 	for (i = 0; i < npages; i++) {
 		if (pages[i]) {
 			if (page_count(pages[i]) != 1)
-				printk(KERN_ERR TTM_PFX
-				       "Erroneous page count. "
-				       "Leaking pages.\n");
+				pr_err("Erroneous page count. Leaking pages.\n");
 			list_add_tail(&pages[i]->lru, &pool->list);
 			pages[i] = NULL;
 			pool->npages++;
@@ -740,8 +726,7 @@ static int ttm_get_pages(struct page **pages, unsigned npages, int flags,
 			p = alloc_page(gfp_flags);
 			if (!p) {
 
-				printk(KERN_ERR TTM_PFX
-				       "Unable to allocate page.");
+				pr_err("Unable to allocate page\n");
 				return -ENOMEM;
 			}
 
@@ -781,9 +766,7 @@ static int ttm_get_pages(struct page **pages, unsigned npages, int flags,
 		if (r) {
 			/* If there is any pages in the list put them back to
 			 * the pool. */
-			printk(KERN_ERR TTM_PFX
-			       "Failed to allocate extra pages "
-			       "for large request.");
+			pr_err("Failed to allocate extra pages for large request\n");
 			ttm_put_pages(pages, count, flags, cstate);
 			return r;
 		}
@@ -809,7 +792,7 @@ int ttm_page_alloc_init(struct ttm_mem_global *glob, unsigned max_pages)
 
 	WARN_ON(_manager);
 
-	printk(KERN_INFO TTM_PFX "Initializing pool allocator.\n");
+	pr_info("Initializing pool allocator\n");
 
 	_manager = kzalloc(sizeof(*_manager), GFP_KERNEL);
 
@@ -844,7 +827,7 @@ void ttm_page_alloc_fini(void)
 {
 	int i;
 
-	printk(KERN_INFO TTM_PFX "Finalizing pool allocator.\n");
+	pr_info("Finalizing pool allocator\n");
 	ttm_pool_mm_shrink_fini(_manager);
 
 	for (i = 0; i < NUM_POOLS; ++i)
diff --git a/drivers/gpu/drm/ttm/ttm_page_alloc_dma.c b/drivers/gpu/drm/ttm/ttm_page_alloc_dma.c
index 0c46d8cdc6ea..4f9e548b2eec 100644
--- a/drivers/gpu/drm/ttm/ttm_page_alloc_dma.c
+++ b/drivers/gpu/drm/ttm/ttm_page_alloc_dma.c
@@ -33,6 +33,8 @@
  *   when freed).
  */
 
+#define pr_fmt(fmt) "[TTM] " fmt
+
 #include <linux/dma-mapping.h>
 #include <linux/list.h>
 #include <linux/seq_file.h> /* for seq_printf */
@@ -221,18 +223,13 @@ static ssize_t ttm_pool_store(struct kobject *kobj, struct attribute *attr,
 		m->options.small = val;
 	else if (attr == &ttm_page_pool_alloc_size) {
 		if (val > NUM_PAGES_TO_ALLOC*8) {
-			printk(KERN_ERR TTM_PFX
-			       "Setting allocation size to %lu "
-			       "is not allowed. Recommended size is "
-			       "%lu\n",
+			pr_err("Setting allocation size to %lu is not allowed. Recommended size is %lu\n",
 			       NUM_PAGES_TO_ALLOC*(PAGE_SIZE >> 7),
 			       NUM_PAGES_TO_ALLOC*(PAGE_SIZE >> 10));
 			return size;
 		} else if (val > NUM_PAGES_TO_ALLOC) {
-			printk(KERN_WARNING TTM_PFX
-			       "Setting allocation size to "
-			       "larger than %lu is not recommended.\n",
-			       NUM_PAGES_TO_ALLOC*(PAGE_SIZE >> 10));
+			pr_warn("Setting allocation size to larger than %lu is not recommended\n",
+				NUM_PAGES_TO_ALLOC*(PAGE_SIZE >> 10));
 		}
 		m->options.alloc_size = val;
 	}
@@ -313,15 +310,13 @@ static int ttm_set_pages_caching(struct dma_pool *pool,
 	if (pool->type & IS_UC) {
 		r = set_pages_array_uc(pages, cpages);
 		if (r)
-			pr_err(TTM_PFX
-			       "%s: Failed to set %d pages to uc!\n",
+			pr_err("%s: Failed to set %d pages to uc!\n",
 			       pool->dev_name, cpages);
 	}
 	if (pool->type & IS_WC) {
 		r = set_pages_array_wc(pages, cpages);
 		if (r)
-			pr_err(TTM_PFX
-			       "%s: Failed to set %d pages to wc!\n",
+			pr_err("%s: Failed to set %d pages to wc!\n",
 			       pool->dev_name, cpages);
 	}
 	return r;
@@ -387,8 +382,8 @@ static void ttm_dma_pages_put(struct dma_pool *pool, struct list_head *d_pages,
 	/* Don't set WB on WB page pool. */
 	if (npages && !(pool->type & IS_CACHED) &&
 	    set_pages_array_wb(pages, npages))
-		pr_err(TTM_PFX "%s: Failed to set %d pages to wb!\n",
-			pool->dev_name, npages);
+		pr_err("%s: Failed to set %d pages to wb!\n",
+		       pool->dev_name, npages);
 
 	list_for_each_entry_safe(d_page, tmp, d_pages, page_list) {
 		list_del(&d_page->page_list);
@@ -400,8 +395,8 @@ static void ttm_dma_page_put(struct dma_pool *pool, struct dma_page *d_page)
 {
 	/* Don't set WB on WB page pool. */
 	if (!(pool->type & IS_CACHED) && set_pages_array_wb(&d_page->p, 1))
-		pr_err(TTM_PFX "%s: Failed to set %d pages to wb!\n",
-			pool->dev_name, 1);
+		pr_err("%s: Failed to set %d pages to wb!\n",
+		       pool->dev_name, 1);
 
 	list_del(&d_page->page_list);
 	__ttm_dma_free_page(pool, d_page);
@@ -430,17 +425,16 @@ static unsigned ttm_dma_page_pool_free(struct dma_pool *pool, unsigned nr_free)
 #if 0
 	if (nr_free > 1) {
 		pr_debug("%s: (%s:%d) Attempting to free %d (%d) pages\n",
-			pool->dev_name, pool->name, current->pid,
-			npages_to_free, nr_free);
+			 pool->dev_name, pool->name, current->pid,
+			 npages_to_free, nr_free);
 	}
 #endif
 	pages_to_free = kmalloc(npages_to_free * sizeof(struct page *),
 			GFP_KERNEL);
 
 	if (!pages_to_free) {
-		pr_err(TTM_PFX
-		       "%s: Failed to allocate memory for pool free operation.\n",
-			pool->dev_name);
+		pr_err("%s: Failed to allocate memory for pool free operation\n",
+		       pool->dev_name);
 		return 0;
 	}
 	INIT_LIST_HEAD(&d_pages);
@@ -723,23 +717,21 @@ static int ttm_dma_pool_alloc_new_pages(struct dma_pool *pool,
 	caching_array = kmalloc(max_cpages*sizeof(struct page *), GFP_KERNEL);
 
 	if (!caching_array) {
-		pr_err(TTM_PFX
-		       "%s: Unable to allocate table for new pages.",
-			pool->dev_name);
+		pr_err("%s: Unable to allocate table for new pages\n",
+		       pool->dev_name);
 		return -ENOMEM;
 	}
 
 	if (count > 1) {
 		pr_debug("%s: (%s:%d) Getting %d pages\n",
-			pool->dev_name, pool->name, current->pid,
-			count);
+			 pool->dev_name, pool->name, current->pid, count);
 	}
 
 	for (i = 0, cpages = 0; i < count; ++i) {
 		dma_p = __ttm_dma_alloc_page(pool);
 		if (!dma_p) {
-			pr_err(TTM_PFX "%s: Unable to get page %u.\n",
-				pool->dev_name, i);
+			pr_err("%s: Unable to get page %u\n",
+			       pool->dev_name, i);
 
 			/* store already allocated pages in the pool after
 			 * setting the caching state */
@@ -821,8 +813,8 @@ static int ttm_dma_page_pool_fill_locked(struct dma_pool *pool,
 			struct dma_page *d_page;
 			unsigned cpages = 0;
 
-			pr_err(TTM_PFX "%s: Failed to fill %s pool (r:%d)!\n",
-				pool->dev_name, pool->name, r);
+			pr_err("%s: Failed to fill %s pool (r:%d)!\n",
+			       pool->dev_name, pool->name, r);
 
 			list_for_each_entry(d_page, &d_pages, page_list) {
 				cpages++;
@@ -1038,8 +1030,8 @@ static int ttm_dma_pool_mm_shrink(struct shrinker *shrink,
 		nr_free = shrink_pages;
 		shrink_pages = ttm_dma_page_pool_free(p->pool, nr_free);
 		pr_debug("%s: (%s:%d) Asked to shrink %d, have %d more to go\n",
-			p->pool->dev_name, p->pool->name, current->pid, nr_free,
-			shrink_pages);
+			 p->pool->dev_name, p->pool->name, current->pid,
+			 nr_free, shrink_pages);
 	}
 	mutex_unlock(&_manager->lock);
 	/* return estimated number of unused pages in pool */
@@ -1064,7 +1056,7 @@ int ttm_dma_page_alloc_init(struct ttm_mem_global *glob, unsigned max_pages)
 
 	WARN_ON(_manager);
 
-	printk(KERN_INFO TTM_PFX "Initializing DMA pool allocator.\n");
+	pr_info("Initializing DMA pool allocator\n");
 
 	_manager = kzalloc(sizeof(*_manager), GFP_KERNEL);
 	if (!_manager)
@@ -1097,7 +1089,7 @@ void ttm_dma_page_alloc_fini(void)
 {
 	struct device_pools *p, *t;
 
-	printk(KERN_INFO TTM_PFX "Finalizing DMA pool allocator.\n");
+	pr_info("Finalizing DMA pool allocator\n");
 	ttm_dma_pool_mm_shrink_fini(_manager);
 
 	list_for_each_entry_safe_reverse(p, t, &_manager->pools, pools) {
diff --git a/drivers/gpu/drm/ttm/ttm_tt.c b/drivers/gpu/drm/ttm/ttm_tt.c
index 2f75d203a2bf..fa09daf9a50c 100644
--- a/drivers/gpu/drm/ttm/ttm_tt.c
+++ b/drivers/gpu/drm/ttm/ttm_tt.c
@@ -28,6 +28,8 @@
  * Authors: Thomas Hellstrom <thellstrom-at-vmware-dot-com>
  */
 
+#define pr_fmt(fmt) "[TTM] " fmt
+
 #include <linux/sched.h>
 #include <linux/highmem.h>
 #include <linux/pagemap.h>
@@ -196,7 +198,7 @@ int ttm_tt_init(struct ttm_tt *ttm, struct ttm_bo_device *bdev,
 	ttm_tt_alloc_page_directory(ttm);
 	if (!ttm->pages) {
 		ttm_tt_destroy(ttm);
-		printk(KERN_ERR TTM_PFX "Failed allocating page table\n");
+		pr_err("Failed allocating page table\n");
 		return -ENOMEM;
 	}
 	return 0;
@@ -229,7 +231,7 @@ int ttm_dma_tt_init(struct ttm_dma_tt *ttm_dma, struct ttm_bo_device *bdev,
 	ttm_dma_tt_alloc_page_directory(ttm_dma);
 	if (!ttm->pages || !ttm_dma->dma_address) {
 		ttm_tt_destroy(ttm);
-		printk(KERN_ERR TTM_PFX "Failed allocating page table\n");
+		pr_err("Failed allocating page table\n");
 		return -ENOMEM;
 	}
 	return 0;
@@ -309,11 +311,11 @@ int ttm_tt_swapin(struct ttm_tt *ttm)
 			goto out_err;
 
 		preempt_disable();
-		from_virtual = kmap_atomic(from_page, KM_USER0);
-		to_virtual = kmap_atomic(to_page, KM_USER1);
+		from_virtual = kmap_atomic(from_page);
+		to_virtual = kmap_atomic(to_page);
 		memcpy(to_virtual, from_virtual, PAGE_SIZE);
-		kunmap_atomic(to_virtual, KM_USER1);
-		kunmap_atomic(from_virtual, KM_USER0);
+		kunmap_atomic(to_virtual);
+		kunmap_atomic(from_virtual);
 		preempt_enable();
 		page_cache_release(from_page);
 	}
@@ -347,7 +349,7 @@ int ttm_tt_swapout(struct ttm_tt *ttm, struct file *persistent_swap_storage)
 						ttm->num_pages << PAGE_SHIFT,
 						0);
 		if (unlikely(IS_ERR(swap_storage))) {
-			printk(KERN_ERR "Failed allocating swap storage.\n");
+			pr_err("Failed allocating swap storage\n");
 			return PTR_ERR(swap_storage);
 		}
 	} else
@@ -365,11 +367,11 @@ int ttm_tt_swapout(struct ttm_tt *ttm, struct file *persistent_swap_storage)
 			goto out_err;
 		}
 		preempt_disable();
-		from_virtual = kmap_atomic(from_page, KM_USER0);
-		to_virtual = kmap_atomic(to_page, KM_USER1);
+		from_virtual = kmap_atomic(from_page);
+		to_virtual = kmap_atomic(to_page);
 		memcpy(to_virtual, from_virtual, PAGE_SIZE);
-		kunmap_atomic(to_virtual, KM_USER1);
-		kunmap_atomic(from_virtual, KM_USER0);
+		kunmap_atomic(to_virtual);
+		kunmap_atomic(from_virtual);
 		preempt_enable();
 		set_page_dirty(to_page);
 		mark_page_accessed(to_page);
diff --git a/drivers/gpu/drm/udl/Kconfig b/drivers/gpu/drm/udl/Kconfig
new file mode 100644
index 000000000000..0b5e096d39a6
--- /dev/null
+++ b/drivers/gpu/drm/udl/Kconfig
@@ -0,0 +1,12 @@
+config DRM_UDL
+	tristate "DisplayLink"
+	depends on DRM && EXPERIMENTAL
+	select DRM_USB
+	select FB_SYS_FILLRECT
+	select FB_SYS_COPYAREA
+	select FB_SYS_IMAGEBLIT
+	select FB_DEFERRED_IO
+	select DRM_KMS_HELPER
+	help
+	  This is a KMS driver for the USB displaylink video adapters.
+          Say M/Y to add support for these devices via drm/kms interfaces.
diff --git a/drivers/gpu/drm/udl/Makefile b/drivers/gpu/drm/udl/Makefile
new file mode 100644
index 000000000000..05c7481bfd40
--- /dev/null
+++ b/drivers/gpu/drm/udl/Makefile
@@ -0,0 +1,6 @@
+
+ccflags-y := -Iinclude/drm
+
+udl-y := udl_drv.o udl_modeset.o udl_connector.o udl_encoder.o udl_main.o udl_fb.o udl_transfer.o udl_gem.o
+
+obj-$(CONFIG_DRM_UDL) := udl.o
diff --git a/drivers/gpu/drm/udl/udl_connector.c b/drivers/gpu/drm/udl/udl_connector.c
new file mode 100644
index 000000000000..ba055e9ca007
--- /dev/null
+++ b/drivers/gpu/drm/udl/udl_connector.c
@@ -0,0 +1,141 @@
+/*
+ * Copyright (C) 2012 Red Hat
+ * based in parts on udlfb.c:
+ * Copyright (C) 2009 Roberto De Ioris <roberto@unbit.it>
+ * Copyright (C) 2009 Jaya Kumar <jayakumar.lkml@gmail.com>
+ * Copyright (C) 2009 Bernie Thompson <bernie@plugable.com>
+ *
+ * This file is subject to the terms and conditions of the GNU General Public
+ * License v2. See the file COPYING in the main directory of this archive for
+ * more details.
+ */
+
+#include "drmP.h"
+#include "drm_crtc.h"
+#include "drm_edid.h"
+#include "drm_crtc_helper.h"
+#include "udl_drv.h"
+
+/* dummy connector to just get EDID,
+   all UDL appear to have a DVI-D */
+
+static u8 *udl_get_edid(struct udl_device *udl)
+{
+	u8 *block;
+	char rbuf[3];
+	int ret, i;
+
+	block = kmalloc(EDID_LENGTH, GFP_KERNEL);
+	if (block == NULL)
+		return NULL;
+
+	for (i = 0; i < EDID_LENGTH; i++) {
+		ret = usb_control_msg(udl->ddev->usbdev,
+				      usb_rcvctrlpipe(udl->ddev->usbdev, 0), (0x02),
+				      (0x80 | (0x02 << 5)), i << 8, 0xA1, rbuf, 2,
+				      HZ);
+		if (ret < 1) {
+			DRM_ERROR("Read EDID byte %d failed err %x\n", i, ret);
+			i--;
+			goto error;
+		}
+		block[i] = rbuf[1];
+	}
+
+	return block;
+
+error:
+	kfree(block);
+	return NULL;
+}
+
+static int udl_get_modes(struct drm_connector *connector)
+{
+	struct udl_device *udl = connector->dev->dev_private;
+	struct edid *edid;
+	int ret;
+
+	edid = (struct edid *)udl_get_edid(udl);
+
+	connector->display_info.raw_edid = (char *)edid;
+
+	drm_mode_connector_update_edid_property(connector, edid);
+	ret = drm_add_edid_modes(connector, edid);
+	connector->display_info.raw_edid = NULL;
+	kfree(edid);
+	return ret;
+}
+
+static int udl_mode_valid(struct drm_connector *connector,
+			  struct drm_display_mode *mode)
+{
+	return 0;
+}
+
+static enum drm_connector_status
+udl_detect(struct drm_connector *connector, bool force)
+{
+	if (drm_device_is_unplugged(connector->dev))
+		return connector_status_disconnected;
+	return connector_status_connected;
+}
+
+struct drm_encoder *udl_best_single_encoder(struct drm_connector *connector)
+{
+	int enc_id = connector->encoder_ids[0];
+	struct drm_mode_object *obj;
+	struct drm_encoder *encoder;
+
+	obj = drm_mode_object_find(connector->dev, enc_id, DRM_MODE_OBJECT_ENCODER);
+	if (!obj)
+		return NULL;
+	encoder = obj_to_encoder(obj);
+	return encoder;
+}
+
+int udl_connector_set_property(struct drm_connector *connector, struct drm_property *property,
+			       uint64_t val)
+{
+	return 0;
+}
+
+static void udl_connector_destroy(struct drm_connector *connector)
+{
+	drm_sysfs_connector_remove(connector);
+	drm_connector_cleanup(connector);
+	kfree(connector);
+}
+
+struct drm_connector_helper_funcs udl_connector_helper_funcs = {
+	.get_modes = udl_get_modes,
+	.mode_valid = udl_mode_valid,
+	.best_encoder = udl_best_single_encoder,
+};
+
+struct drm_connector_funcs udl_connector_funcs = {
+	.dpms = drm_helper_connector_dpms,
+	.detect = udl_detect,
+	.fill_modes = drm_helper_probe_single_connector_modes,
+	.destroy = udl_connector_destroy,
+	.set_property = udl_connector_set_property,
+};
+
+int udl_connector_init(struct drm_device *dev, struct drm_encoder *encoder)
+{
+	struct drm_connector *connector;
+
+	connector = kzalloc(sizeof(struct drm_connector), GFP_KERNEL);
+	if (!connector)
+		return -ENOMEM;
+
+	drm_connector_init(dev, connector, &udl_connector_funcs, DRM_MODE_CONNECTOR_DVII);
+	drm_connector_helper_add(connector, &udl_connector_helper_funcs);
+
+	drm_sysfs_connector_add(connector);
+	drm_mode_connector_attach_encoder(connector, encoder);
+
+	drm_connector_attach_property(connector,
+				      dev->mode_config.dirty_info_property,
+				      1);
+	return 0;
+}
diff --git a/drivers/gpu/drm/udl/udl_drv.c b/drivers/gpu/drm/udl/udl_drv.c
new file mode 100644
index 000000000000..53673907a6a0
--- /dev/null
+++ b/drivers/gpu/drm/udl/udl_drv.c
@@ -0,0 +1,99 @@
+/*
+ * Copyright (C) 2012 Red Hat
+ *
+ * This file is subject to the terms and conditions of the GNU General Public
+ * License v2. See the file COPYING in the main directory of this archive for
+ * more details.
+ */
+
+#include <linux/module.h>
+#include "drm_usb.h"
+#include "drm_crtc_helper.h"
+#include "udl_drv.h"
+
+static struct drm_driver driver;
+
+static struct usb_device_id id_table[] = {
+	{.idVendor = 0x17e9, .match_flags = USB_DEVICE_ID_MATCH_VENDOR,},
+	{},
+};
+MODULE_DEVICE_TABLE(usb, id_table);
+
+MODULE_LICENSE("GPL");
+
+static int udl_usb_probe(struct usb_interface *interface,
+			 const struct usb_device_id *id)
+{
+	return drm_get_usb_dev(interface, id, &driver);
+}
+
+static void udl_usb_disconnect(struct usb_interface *interface)
+{
+	struct drm_device *dev = usb_get_intfdata(interface);
+
+	drm_kms_helper_poll_disable(dev);
+	drm_connector_unplug_all(dev);
+	udl_fbdev_unplug(dev);
+	udl_drop_usb(dev);
+	drm_unplug_dev(dev);
+}
+
+static struct vm_operations_struct udl_gem_vm_ops = {
+	.fault = udl_gem_fault,
+	.open = drm_gem_vm_open,
+	.close = drm_gem_vm_close,
+};
+
+static const struct file_operations udl_driver_fops = {
+	.owner = THIS_MODULE,
+	.open = drm_open,
+	.mmap = udl_drm_gem_mmap,
+	.poll = drm_poll,
+	.read = drm_read,
+	.unlocked_ioctl	= drm_ioctl,
+	.release = drm_release,
+	.fasync = drm_fasync,
+	.llseek = noop_llseek,
+};
+
+static struct drm_driver driver = {
+	.driver_features = DRIVER_MODESET | DRIVER_GEM,
+	.load = udl_driver_load,
+	.unload = udl_driver_unload,
+
+	/* gem hooks */
+	.gem_init_object = udl_gem_init_object,
+	.gem_free_object = udl_gem_free_object,
+	.gem_vm_ops = &udl_gem_vm_ops,
+
+	.dumb_create = udl_dumb_create,
+	.dumb_map_offset = udl_gem_mmap,
+	.dumb_destroy = udl_dumb_destroy,
+	.fops = &udl_driver_fops,
+	.name = DRIVER_NAME,
+	.desc = DRIVER_DESC,
+	.date = DRIVER_DATE,
+	.major = DRIVER_MAJOR,
+	.minor = DRIVER_MINOR,
+	.patchlevel = DRIVER_PATCHLEVEL,
+};
+
+static struct usb_driver udl_driver = {
+	.name = "udl",
+	.probe = udl_usb_probe,
+	.disconnect = udl_usb_disconnect,
+	.id_table = id_table,
+};
+
+static int __init udl_init(void)
+{
+	return drm_usb_init(&driver, &udl_driver);
+}
+
+static void __exit udl_exit(void)
+{
+	drm_usb_exit(&driver, &udl_driver);
+}
+
+module_init(udl_init);
+module_exit(udl_exit);
diff --git a/drivers/gpu/drm/udl/udl_drv.h b/drivers/gpu/drm/udl/udl_drv.h
new file mode 100644
index 000000000000..96820d03a303
--- /dev/null
+++ b/drivers/gpu/drm/udl/udl_drv.h
@@ -0,0 +1,142 @@
+/*
+ * Copyright (C) 2012 Red Hat
+ *
+ * based in parts on udlfb.c:
+ * Copyright (C) 2009 Roberto De Ioris <roberto@unbit.it>
+ * Copyright (C) 2009 Jaya Kumar <jayakumar.lkml@gmail.com>
+ * Copyright (C) 2009 Bernie Thompson <bernie@plugable.com>
+ *
+ * This file is subject to the terms and conditions of the GNU General Public
+ * License v2. See the file COPYING in the main directory of this archive for
+ * more details.
+ */
+
+#ifndef UDL_DRV_H
+#define UDL_DRV_H
+
+#include <linux/usb.h>
+
+#define DRIVER_NAME		"udl"
+#define DRIVER_DESC		"DisplayLink"
+#define DRIVER_DATE		"20120220"
+
+#define DRIVER_MAJOR		0
+#define DRIVER_MINOR		0
+#define DRIVER_PATCHLEVEL	1
+
+struct udl_device;
+
+struct urb_node {
+	struct list_head entry;
+	struct udl_device *dev;
+	struct delayed_work release_urb_work;
+	struct urb *urb;
+};
+
+struct urb_list {
+	struct list_head list;
+	spinlock_t lock;
+	struct semaphore limit_sem;
+	int available;
+	int count;
+	size_t size;
+};
+
+struct udl_fbdev;
+
+struct udl_device {
+	struct device *dev;
+	struct drm_device *ddev;
+
+	int sku_pixel_limit;
+
+	struct urb_list urbs;
+	atomic_t lost_pixels; /* 1 = a render op failed. Need screen refresh */
+
+	struct udl_fbdev *fbdev;
+	char mode_buf[1024];
+	uint32_t mode_buf_len;
+	atomic_t bytes_rendered; /* raw pixel-bytes driver asked to render */
+	atomic_t bytes_identical; /* saved effort with backbuffer comparison */
+	atomic_t bytes_sent; /* to usb, after compression including overhead */
+	atomic_t cpu_kcycles_used; /* transpired during pixel processing */
+};
+
+struct udl_gem_object {
+	struct drm_gem_object base;
+	struct page **pages;
+	void *vmapping;
+};
+
+#define to_udl_bo(x) container_of(x, struct udl_gem_object, base)
+
+struct udl_framebuffer {
+	struct drm_framebuffer base;
+	struct udl_gem_object *obj;
+	bool active_16; /* active on the 16-bit channel */
+};
+
+#define to_udl_fb(x) container_of(x, struct udl_framebuffer, base)
+
+/* modeset */
+int udl_modeset_init(struct drm_device *dev);
+void udl_modeset_cleanup(struct drm_device *dev);
+int udl_connector_init(struct drm_device *dev, struct drm_encoder *encoder);
+
+struct drm_encoder *udl_encoder_init(struct drm_device *dev);
+
+struct urb *udl_get_urb(struct drm_device *dev);
+
+int udl_submit_urb(struct drm_device *dev, struct urb *urb, size_t len);
+void udl_urb_completion(struct urb *urb);
+
+int udl_driver_load(struct drm_device *dev, unsigned long flags);
+int udl_driver_unload(struct drm_device *dev);
+
+int udl_fbdev_init(struct drm_device *dev);
+void udl_fbdev_cleanup(struct drm_device *dev);
+void udl_fbdev_unplug(struct drm_device *dev);
+struct drm_framebuffer *
+udl_fb_user_fb_create(struct drm_device *dev,
+		      struct drm_file *file,
+		      struct drm_mode_fb_cmd2 *mode_cmd);
+
+int udl_render_hline(struct drm_device *dev, int bpp, struct urb **urb_ptr,
+		     const char *front, char **urb_buf_ptr,
+		     u32 byte_offset, u32 byte_width,
+		     int *ident_ptr, int *sent_ptr);
+
+int udl_dumb_create(struct drm_file *file_priv,
+		    struct drm_device *dev,
+		    struct drm_mode_create_dumb *args);
+int udl_gem_mmap(struct drm_file *file_priv, struct drm_device *dev,
+		 uint32_t handle, uint64_t *offset);
+int udl_dumb_destroy(struct drm_file *file_priv, struct drm_device *dev,
+		     uint32_t handle);
+
+int udl_gem_init_object(struct drm_gem_object *obj);
+void udl_gem_free_object(struct drm_gem_object *gem_obj);
+struct udl_gem_object *udl_gem_alloc_object(struct drm_device *dev,
+					    size_t size);
+
+int udl_gem_vmap(struct udl_gem_object *obj);
+void udl_gem_vunmap(struct udl_gem_object *obj);
+int udl_drm_gem_mmap(struct file *filp, struct vm_area_struct *vma);
+int udl_gem_fault(struct vm_area_struct *vma, struct vm_fault *vmf);
+
+int udl_handle_damage(struct udl_framebuffer *fb, int x, int y,
+		      int width, int height);
+
+int udl_drop_usb(struct drm_device *dev);
+
+#define CMD_WRITE_RAW8   "\xAF\x60" /**< 8 bit raw write command. */
+#define CMD_WRITE_RL8    "\xAF\x61" /**< 8 bit run length command. */
+#define CMD_WRITE_COPY8  "\xAF\x62" /**< 8 bit copy command. */
+#define CMD_WRITE_RLX8   "\xAF\x63" /**< 8 bit extended run length command. */
+
+#define CMD_WRITE_RAW16  "\xAF\x68" /**< 16 bit raw write command. */
+#define CMD_WRITE_RL16   "\xAF\x69" /**< 16 bit run length command. */
+#define CMD_WRITE_COPY16 "\xAF\x6A" /**< 16 bit copy command. */
+#define CMD_WRITE_RLX16  "\xAF\x6B" /**< 16 bit extended run length command. */
+
+#endif
diff --git a/drivers/gpu/drm/udl/udl_encoder.c b/drivers/gpu/drm/udl/udl_encoder.c
new file mode 100644
index 000000000000..56e75f0f1df5
--- /dev/null
+++ b/drivers/gpu/drm/udl/udl_encoder.c
@@ -0,0 +1,80 @@
+/*
+ * Copyright (C) 2012 Red Hat
+ * based in parts on udlfb.c:
+ * Copyright (C) 2009 Roberto De Ioris <roberto@unbit.it>
+ * Copyright (C) 2009 Jaya Kumar <jayakumar.lkml@gmail.com>
+ * Copyright (C) 2009 Bernie Thompson <bernie@plugable.com>
+ *
+ * This file is subject to the terms and conditions of the GNU General Public
+ * License v2. See the file COPYING in the main directory of this archive for
+ * more details.
+ */
+
+#include "drmP.h"
+#include "drm_crtc.h"
+#include "drm_crtc_helper.h"
+#include "udl_drv.h"
+
+/* dummy encoder */
+void udl_enc_destroy(struct drm_encoder *encoder)
+{
+	drm_encoder_cleanup(encoder);
+	kfree(encoder);
+}
+
+static void udl_encoder_disable(struct drm_encoder *encoder)
+{
+}
+
+static bool udl_mode_fixup(struct drm_encoder *encoder,
+			   struct drm_display_mode *mode,
+			   struct drm_display_mode *adjusted_mode)
+{
+	return true;
+}
+
+static void udl_encoder_prepare(struct drm_encoder *encoder)
+{
+}
+
+static void udl_encoder_commit(struct drm_encoder *encoder)
+{
+}
+
+static void udl_encoder_mode_set(struct drm_encoder *encoder,
+				 struct drm_display_mode *mode,
+				 struct drm_display_mode *adjusted_mode)
+{
+}
+
+static void
+udl_encoder_dpms(struct drm_encoder *encoder, int mode)
+{
+}
+
+static const struct drm_encoder_helper_funcs udl_helper_funcs = {
+	.dpms = udl_encoder_dpms,
+	.mode_fixup = udl_mode_fixup,
+	.prepare = udl_encoder_prepare,
+	.mode_set = udl_encoder_mode_set,
+	.commit = udl_encoder_commit,
+	.disable = udl_encoder_disable,
+};
+
+static const struct drm_encoder_funcs udl_enc_funcs = {
+	.destroy = udl_enc_destroy,
+};
+
+struct drm_encoder *udl_encoder_init(struct drm_device *dev)
+{
+	struct drm_encoder *encoder;
+
+	encoder = kzalloc(sizeof(struct drm_encoder), GFP_KERNEL);
+	if (!encoder)
+		return NULL;
+
+	drm_encoder_init(dev, encoder, &udl_enc_funcs, DRM_MODE_ENCODER_TMDS);
+	drm_encoder_helper_add(encoder, &udl_helper_funcs);
+	encoder->possible_crtcs = 1;
+	return encoder;
+}
diff --git a/drivers/gpu/drm/udl/udl_fb.c b/drivers/gpu/drm/udl/udl_fb.c
new file mode 100644
index 000000000000..4d9c3a5d8a45
--- /dev/null
+++ b/drivers/gpu/drm/udl/udl_fb.c
@@ -0,0 +1,611 @@
+/*
+ * Copyright (C) 2012 Red Hat
+ *
+ * based in parts on udlfb.c:
+ * Copyright (C) 2009 Roberto De Ioris <roberto@unbit.it>
+ * Copyright (C) 2009 Jaya Kumar <jayakumar.lkml@gmail.com>
+ * Copyright (C) 2009 Bernie Thompson <bernie@plugable.com>
+ *
+ * This file is subject to the terms and conditions of the GNU General Public
+ * License v2. See the file COPYING in the main directory of this archive for
+ * more details.
+ */
+#include <linux/module.h>
+#include <linux/slab.h>
+#include <linux/fb.h>
+
+#include "drmP.h"
+#include "drm.h"
+#include "drm_crtc.h"
+#include "drm_crtc_helper.h"
+#include "udl_drv.h"
+
+#include "drm_fb_helper.h"
+
+#define DL_DEFIO_WRITE_DELAY    5 /* fb_deferred_io.delay in jiffies */
+
+static int fb_defio = 1;  /* Optionally enable experimental fb_defio mmap support */
+static int fb_bpp = 16;
+
+module_param(fb_bpp, int, S_IWUSR | S_IRUSR | S_IWGRP | S_IRGRP);
+module_param(fb_defio, int, S_IWUSR | S_IRUSR | S_IWGRP | S_IRGRP);
+
+struct udl_fbdev {
+	struct drm_fb_helper helper;
+	struct udl_framebuffer ufb;
+	struct list_head fbdev_list;
+	int fb_count;
+};
+
+#define DL_ALIGN_UP(x, a) ALIGN(x, a)
+#define DL_ALIGN_DOWN(x, a) ALIGN(x-(a-1), a)
+
+/** Read the red component (0..255) of a 32 bpp colour. */
+#define DLO_RGB_GETRED(col) (uint8_t)((col) & 0xFF)
+
+/** Read the green component (0..255) of a 32 bpp colour. */
+#define DLO_RGB_GETGRN(col) (uint8_t)(((col) >> 8) & 0xFF)
+
+/** Read the blue component (0..255) of a 32 bpp colour. */
+#define DLO_RGB_GETBLU(col) (uint8_t)(((col) >> 16) & 0xFF)
+
+/** Return red/green component of a 16 bpp colour number. */
+#define DLO_RG16(red, grn) (uint8_t)((((red) & 0xF8) | ((grn) >> 5)) & 0xFF)
+
+/** Return green/blue component of a 16 bpp colour number. */
+#define DLO_GB16(grn, blu) (uint8_t)(((((grn) & 0x1C) << 3) | ((blu) >> 3)) & 0xFF)
+
+/** Return 8 bpp colour number from red, green and blue components. */
+#define DLO_RGB8(red, grn, blu) ((((red) << 5) | (((grn) & 3) << 3) | ((blu) & 7)) & 0xFF)
+
+#if 0
+static uint8_t rgb8(uint32_t col)
+{
+	uint8_t red = DLO_RGB_GETRED(col);
+	uint8_t grn = DLO_RGB_GETGRN(col);
+	uint8_t blu = DLO_RGB_GETBLU(col);
+
+	return DLO_RGB8(red, grn, blu);
+}
+
+static uint16_t rgb16(uint32_t col)
+{
+	uint8_t red = DLO_RGB_GETRED(col);
+	uint8_t grn = DLO_RGB_GETGRN(col);
+	uint8_t blu = DLO_RGB_GETBLU(col);
+
+	return (DLO_RG16(red, grn) << 8) + DLO_GB16(grn, blu);
+}
+#endif
+
+/*
+ * NOTE: fb_defio.c is holding info->fbdefio.mutex
+ *   Touching ANY framebuffer memory that triggers a page fault
+ *   in fb_defio will cause a deadlock, when it also tries to
+ *   grab the same mutex.
+ */
+static void udlfb_dpy_deferred_io(struct fb_info *info,
+				  struct list_head *pagelist)
+{
+	struct page *cur;
+	struct fb_deferred_io *fbdefio = info->fbdefio;
+	struct udl_fbdev *ufbdev = info->par;
+	struct drm_device *dev = ufbdev->ufb.base.dev;
+	struct udl_device *udl = dev->dev_private;
+	struct urb *urb;
+	char *cmd;
+	cycles_t start_cycles, end_cycles;
+	int bytes_sent = 0;
+	int bytes_identical = 0;
+	int bytes_rendered = 0;
+
+	if (!fb_defio)
+		return;
+
+	start_cycles = get_cycles();
+
+	urb = udl_get_urb(dev);
+	if (!urb)
+		return;
+
+	cmd = urb->transfer_buffer;
+
+	/* walk the written page list and render each to device */
+	list_for_each_entry(cur, &fbdefio->pagelist, lru) {
+
+		if (udl_render_hline(dev, (ufbdev->ufb.base.bits_per_pixel / 8),
+				  &urb, (char *) info->fix.smem_start,
+				  &cmd, cur->index << PAGE_SHIFT,
+				  PAGE_SIZE, &bytes_identical, &bytes_sent))
+			goto error;
+		bytes_rendered += PAGE_SIZE;
+	}
+
+	if (cmd > (char *) urb->transfer_buffer) {
+		/* Send partial buffer remaining before exiting */
+		int len = cmd - (char *) urb->transfer_buffer;
+		udl_submit_urb(dev, urb, len);
+		bytes_sent += len;
+	} else
+		udl_urb_completion(urb);
+
+error:
+	atomic_add(bytes_sent, &udl->bytes_sent);
+	atomic_add(bytes_identical, &udl->bytes_identical);
+	atomic_add(bytes_rendered, &udl->bytes_rendered);
+	end_cycles = get_cycles();
+	atomic_add(((unsigned int) ((end_cycles - start_cycles)
+		    >> 10)), /* Kcycles */
+		   &udl->cpu_kcycles_used);
+}
+
+int udl_handle_damage(struct udl_framebuffer *fb, int x, int y,
+		      int width, int height)
+{
+	struct drm_device *dev = fb->base.dev;
+	struct udl_device *udl = dev->dev_private;
+	int i, ret;
+	char *cmd;
+	cycles_t start_cycles, end_cycles;
+	int bytes_sent = 0;
+	int bytes_identical = 0;
+	struct urb *urb;
+	int aligned_x;
+	int bpp = (fb->base.bits_per_pixel / 8);
+
+	if (!fb->active_16)
+		return 0;
+
+	if (!fb->obj->vmapping)
+		udl_gem_vmap(fb->obj);
+
+	start_cycles = get_cycles();
+
+	aligned_x = DL_ALIGN_DOWN(x, sizeof(unsigned long));
+	width = DL_ALIGN_UP(width + (x-aligned_x), sizeof(unsigned long));
+	x = aligned_x;
+
+	if ((width <= 0) ||
+	    (x + width > fb->base.width) ||
+	    (y + height > fb->base.height))
+		return -EINVAL;
+
+	urb = udl_get_urb(dev);
+	if (!urb)
+		return 0;
+	cmd = urb->transfer_buffer;
+
+	for (i = y; i < y + height ; i++) {
+		const int line_offset = fb->base.pitches[0] * i;
+		const int byte_offset = line_offset + (x * bpp);
+
+		if (udl_render_hline(dev, bpp, &urb,
+				     (char *) fb->obj->vmapping,
+				     &cmd, byte_offset, width * bpp,
+				     &bytes_identical, &bytes_sent))
+			goto error;
+	}
+
+	if (cmd > (char *) urb->transfer_buffer) {
+		/* Send partial buffer remaining before exiting */
+		int len = cmd - (char *) urb->transfer_buffer;
+		ret = udl_submit_urb(dev, urb, len);
+		bytes_sent += len;
+	} else
+		udl_urb_completion(urb);
+
+error:
+	atomic_add(bytes_sent, &udl->bytes_sent);
+	atomic_add(bytes_identical, &udl->bytes_identical);
+	atomic_add(width*height*bpp, &udl->bytes_rendered);
+	end_cycles = get_cycles();
+	atomic_add(((unsigned int) ((end_cycles - start_cycles)
+		    >> 10)), /* Kcycles */
+		   &udl->cpu_kcycles_used);
+
+	return 0;
+}
+
+static int udl_fb_mmap(struct fb_info *info, struct vm_area_struct *vma)
+{
+	unsigned long start = vma->vm_start;
+	unsigned long size = vma->vm_end - vma->vm_start;
+	unsigned long offset = vma->vm_pgoff << PAGE_SHIFT;
+	unsigned long page, pos;
+
+	if (offset + size > info->fix.smem_len)
+		return -EINVAL;
+
+	pos = (unsigned long)info->fix.smem_start + offset;
+
+	pr_notice("mmap() framebuffer addr:%lu size:%lu\n",
+		  pos, size);
+
+	while (size > 0) {
+		page = vmalloc_to_pfn((void *)pos);
+		if (remap_pfn_range(vma, start, page, PAGE_SIZE, PAGE_SHARED))
+			return -EAGAIN;
+
+		start += PAGE_SIZE;
+		pos += PAGE_SIZE;
+		if (size > PAGE_SIZE)
+			size -= PAGE_SIZE;
+		else
+			size = 0;
+	}
+
+	vma->vm_flags |= VM_RESERVED;	/* avoid to swap out this VMA */
+	return 0;
+}
+
+static void udl_fb_fillrect(struct fb_info *info, const struct fb_fillrect *rect)
+{
+	struct udl_fbdev *ufbdev = info->par;
+
+	sys_fillrect(info, rect);
+
+	udl_handle_damage(&ufbdev->ufb, rect->dx, rect->dy, rect->width,
+			  rect->height);
+}
+
+static void udl_fb_copyarea(struct fb_info *info, const struct fb_copyarea *region)
+{
+	struct udl_fbdev *ufbdev = info->par;
+
+	sys_copyarea(info, region);
+
+	udl_handle_damage(&ufbdev->ufb, region->dx, region->dy, region->width,
+			  region->height);
+}
+
+static void udl_fb_imageblit(struct fb_info *info, const struct fb_image *image)
+{
+	struct udl_fbdev *ufbdev = info->par;
+
+	sys_imageblit(info, image);
+
+	udl_handle_damage(&ufbdev->ufb, image->dx, image->dy, image->width,
+			  image->height);
+}
+
+/*
+ * It's common for several clients to have framebuffer open simultaneously.
+ * e.g. both fbcon and X. Makes things interesting.
+ * Assumes caller is holding info->lock (for open and release at least)
+ */
+static int udl_fb_open(struct fb_info *info, int user)
+{
+	struct udl_fbdev *ufbdev = info->par;
+	struct drm_device *dev = ufbdev->ufb.base.dev;
+	struct udl_device *udl = dev->dev_private;
+
+	/* If the USB device is gone, we don't accept new opens */
+	if (drm_device_is_unplugged(udl->ddev))
+		return -ENODEV;
+
+	ufbdev->fb_count++;
+
+	if (fb_defio && (info->fbdefio == NULL)) {
+		/* enable defio at last moment if not disabled by client */
+
+		struct fb_deferred_io *fbdefio;
+
+		fbdefio = kmalloc(sizeof(struct fb_deferred_io), GFP_KERNEL);
+
+		if (fbdefio) {
+			fbdefio->delay = DL_DEFIO_WRITE_DELAY;
+			fbdefio->deferred_io = udlfb_dpy_deferred_io;
+		}
+
+		info->fbdefio = fbdefio;
+		fb_deferred_io_init(info);
+	}
+
+	pr_notice("open /dev/fb%d user=%d fb_info=%p count=%d\n",
+		  info->node, user, info, ufbdev->fb_count);
+
+	return 0;
+}
+
+
+/*
+ * Assumes caller is holding info->lock mutex (for open and release at least)
+ */
+static int udl_fb_release(struct fb_info *info, int user)
+{
+	struct udl_fbdev *ufbdev = info->par;
+
+	ufbdev->fb_count--;
+
+	if ((ufbdev->fb_count == 0) && (info->fbdefio)) {
+		fb_deferred_io_cleanup(info);
+		kfree(info->fbdefio);
+		info->fbdefio = NULL;
+		info->fbops->fb_mmap = udl_fb_mmap;
+	}
+
+	pr_warn("released /dev/fb%d user=%d count=%d\n",
+		info->node, user, ufbdev->fb_count);
+
+	return 0;
+}
+
+static struct fb_ops udlfb_ops = {
+	.owner = THIS_MODULE,
+	.fb_check_var = drm_fb_helper_check_var,
+	.fb_set_par = drm_fb_helper_set_par,
+	.fb_fillrect = udl_fb_fillrect,
+	.fb_copyarea = udl_fb_copyarea,
+	.fb_imageblit = udl_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,
+	.fb_mmap = udl_fb_mmap,
+	.fb_open = udl_fb_open,
+	.fb_release = udl_fb_release,
+};
+
+void udl_crtc_fb_gamma_set(struct drm_crtc *crtc, u16 red, u16 green,
+			   u16 blue, int regno)
+{
+}
+
+void udl_crtc_fb_gamma_get(struct drm_crtc *crtc, u16 *red, u16 *green,
+			     u16 *blue, int regno)
+{
+	*red = 0;
+	*green = 0;
+	*blue = 0;
+}
+
+static int udl_user_framebuffer_dirty(struct drm_framebuffer *fb,
+				      struct drm_file *file,
+				      unsigned flags, unsigned color,
+				      struct drm_clip_rect *clips,
+				      unsigned num_clips)
+{
+	struct udl_framebuffer *ufb = to_udl_fb(fb);
+	int i;
+
+	if (!ufb->active_16)
+		return 0;
+
+	for (i = 0; i < num_clips; i++) {
+		udl_handle_damage(ufb, clips[i].x1, clips[i].y1,
+				  clips[i].x2 - clips[i].x1,
+				  clips[i].y2 - clips[i].y1);
+	}
+	return 0;
+}
+
+static void udl_user_framebuffer_destroy(struct drm_framebuffer *fb)
+{
+	struct udl_framebuffer *ufb = to_udl_fb(fb);
+
+	if (ufb->obj)
+		drm_gem_object_unreference_unlocked(&ufb->obj->base);
+
+	drm_framebuffer_cleanup(fb);
+	kfree(ufb);
+}
+
+static const struct drm_framebuffer_funcs udlfb_funcs = {
+	.destroy = udl_user_framebuffer_destroy,
+	.dirty = udl_user_framebuffer_dirty,
+	.create_handle = NULL,
+};
+
+
+static int
+udl_framebuffer_init(struct drm_device *dev,
+		     struct udl_framebuffer *ufb,
+		     struct drm_mode_fb_cmd2 *mode_cmd,
+		     struct udl_gem_object *obj)
+{
+	int ret;
+
+	ufb->obj = obj;
+	ret = drm_framebuffer_init(dev, &ufb->base, &udlfb_funcs);
+	drm_helper_mode_fill_fb_struct(&ufb->base, mode_cmd);
+	return ret;
+}
+
+
+static int udlfb_create(struct udl_fbdev *ufbdev,
+			struct drm_fb_helper_surface_size *sizes)
+{
+	struct drm_device *dev = ufbdev->helper.dev;
+	struct fb_info *info;
+	struct device *device = &dev->usbdev->dev;
+	struct drm_framebuffer *fb;
+	struct drm_mode_fb_cmd2 mode_cmd;
+	struct udl_gem_object *obj;
+	uint32_t size;
+	int ret = 0;
+
+	if (sizes->surface_bpp == 24)
+		sizes->surface_bpp = 32;
+
+	mode_cmd.width = sizes->surface_width;
+	mode_cmd.height = sizes->surface_height;
+	mode_cmd.pitches[0] = mode_cmd.width * ((sizes->surface_bpp + 7) / 8);
+
+	mode_cmd.pixel_format = drm_mode_legacy_fb_format(sizes->surface_bpp,
+							  sizes->surface_depth);
+
+	size = mode_cmd.pitches[0] * mode_cmd.height;
+	size = ALIGN(size, PAGE_SIZE);
+
+	obj = udl_gem_alloc_object(dev, size);
+	if (!obj)
+		goto out;
+
+	ret = udl_gem_vmap(obj);
+	if (ret) {
+		DRM_ERROR("failed to vmap fb\n");
+		goto out_gfree;
+	}
+
+	info = framebuffer_alloc(0, device);
+	if (!info) {
+		ret = -ENOMEM;
+		goto out_gfree;
+	}
+	info->par = ufbdev;
+
+	ret = udl_framebuffer_init(dev, &ufbdev->ufb, &mode_cmd, obj);
+	if (ret)
+		goto out_gfree;
+
+	fb = &ufbdev->ufb.base;
+
+	ufbdev->helper.fb = fb;
+	ufbdev->helper.fbdev = info;
+
+	strcpy(info->fix.id, "udldrmfb");
+
+	info->screen_base = ufbdev->ufb.obj->vmapping;
+	info->fix.smem_len = size;
+	info->fix.smem_start = (unsigned long)ufbdev->ufb.obj->vmapping;
+
+	info->flags = FBINFO_DEFAULT | FBINFO_CAN_FORCE_OUTPUT;
+	info->fbops = &udlfb_ops;
+	drm_fb_helper_fill_fix(info, fb->pitches[0], fb->depth);
+	drm_fb_helper_fill_var(info, &ufbdev->helper, sizes->fb_width, sizes->fb_height);
+
+	ret = fb_alloc_cmap(&info->cmap, 256, 0);
+	if (ret) {
+		ret = -ENOMEM;
+		goto out_gfree;
+	}
+
+
+	DRM_DEBUG_KMS("allocated %dx%d vmal %p\n",
+		      fb->width, fb->height,
+		      ufbdev->ufb.obj->vmapping);
+
+	return ret;
+out_gfree:
+	drm_gem_object_unreference(&ufbdev->ufb.obj->base);
+out:
+	return ret;
+}
+
+static int udl_fb_find_or_create_single(struct drm_fb_helper *helper,
+					struct drm_fb_helper_surface_size *sizes)
+{
+	struct udl_fbdev *ufbdev = (struct udl_fbdev *)helper;
+	int new_fb = 0;
+	int ret;
+
+	if (!helper->fb) {
+		ret = udlfb_create(ufbdev, sizes);
+		if (ret)
+			return ret;
+
+		new_fb = 1;
+	}
+	return new_fb;
+}
+
+static struct drm_fb_helper_funcs udl_fb_helper_funcs = {
+	.gamma_set = udl_crtc_fb_gamma_set,
+	.gamma_get = udl_crtc_fb_gamma_get,
+	.fb_probe = udl_fb_find_or_create_single,
+};
+
+static void udl_fbdev_destroy(struct drm_device *dev,
+			      struct udl_fbdev *ufbdev)
+{
+	struct fb_info *info;
+	if (ufbdev->helper.fbdev) {
+		info = ufbdev->helper.fbdev;
+		unregister_framebuffer(info);
+		if (info->cmap.len)
+			fb_dealloc_cmap(&info->cmap);
+		framebuffer_release(info);
+	}
+	drm_fb_helper_fini(&ufbdev->helper);
+	drm_framebuffer_cleanup(&ufbdev->ufb.base);
+	drm_gem_object_unreference_unlocked(&ufbdev->ufb.obj->base);
+}
+
+int udl_fbdev_init(struct drm_device *dev)
+{
+	struct udl_device *udl = dev->dev_private;
+	int bpp_sel = fb_bpp;
+	struct udl_fbdev *ufbdev;
+	int ret;
+
+	ufbdev = kzalloc(sizeof(struct udl_fbdev), GFP_KERNEL);
+	if (!ufbdev)
+		return -ENOMEM;
+
+	udl->fbdev = ufbdev;
+	ufbdev->helper.funcs = &udl_fb_helper_funcs;
+
+	ret = drm_fb_helper_init(dev, &ufbdev->helper,
+				 1, 1);
+	if (ret) {
+		kfree(ufbdev);
+		return ret;
+
+	}
+
+	drm_fb_helper_single_add_all_connectors(&ufbdev->helper);
+	drm_fb_helper_initial_config(&ufbdev->helper, bpp_sel);
+	return 0;
+}
+
+void udl_fbdev_cleanup(struct drm_device *dev)
+{
+	struct udl_device *udl = dev->dev_private;
+	if (!udl->fbdev)
+		return;
+
+	udl_fbdev_destroy(dev, udl->fbdev);
+	kfree(udl->fbdev);
+	udl->fbdev = NULL;
+}
+
+void udl_fbdev_unplug(struct drm_device *dev)
+{
+	struct udl_device *udl = dev->dev_private;
+	struct udl_fbdev *ufbdev;
+	if (!udl->fbdev)
+		return;
+
+	ufbdev = udl->fbdev;
+	if (ufbdev->helper.fbdev) {
+		struct fb_info *info;
+		info = ufbdev->helper.fbdev;
+		unlink_framebuffer(info);
+	}
+}
+
+struct drm_framebuffer *
+udl_fb_user_fb_create(struct drm_device *dev,
+		   struct drm_file *file,
+		   struct drm_mode_fb_cmd2 *mode_cmd)
+{
+	struct drm_gem_object *obj;
+	struct udl_framebuffer *ufb;
+	int ret;
+
+	obj = drm_gem_object_lookup(dev, file, mode_cmd->handles[0]);
+	if (obj == NULL)
+		return ERR_PTR(-ENOENT);
+
+	ufb = kzalloc(sizeof(*ufb), GFP_KERNEL);
+	if (ufb == NULL)
+		return ERR_PTR(-ENOMEM);
+
+	ret = udl_framebuffer_init(dev, ufb, mode_cmd, to_udl_bo(obj));
+	if (ret) {
+		kfree(ufb);
+		return ERR_PTR(-EINVAL);
+	}
+	return &ufb->base;
+}
diff --git a/drivers/gpu/drm/udl/udl_gem.c b/drivers/gpu/drm/udl/udl_gem.c
new file mode 100644
index 000000000000..92f19ef329b0
--- /dev/null
+++ b/drivers/gpu/drm/udl/udl_gem.c
@@ -0,0 +1,241 @@
+/*
+ * Copyright (C) 2012 Red Hat
+ *
+ * This file is subject to the terms and conditions of the GNU General Public
+ * License v2. See the file COPYING in the main directory of this archive for
+ * more details.
+ */
+
+#include "drmP.h"
+#include "udl_drv.h"
+#include <linux/shmem_fs.h>
+
+struct udl_gem_object *udl_gem_alloc_object(struct drm_device *dev,
+					    size_t size)
+{
+	struct udl_gem_object *obj;
+
+	obj = kzalloc(sizeof(*obj), GFP_KERNEL);
+	if (obj == NULL)
+		return NULL;
+
+	if (drm_gem_object_init(dev, &obj->base, size) != 0) {
+		kfree(obj);
+		return NULL;
+	}
+
+	return obj;
+}
+
+static int
+udl_gem_create(struct drm_file *file,
+	       struct drm_device *dev,
+	       uint64_t size,
+	       uint32_t *handle_p)
+{
+	struct udl_gem_object *obj;
+	int ret;
+	u32 handle;
+
+	size = roundup(size, PAGE_SIZE);
+
+	obj = udl_gem_alloc_object(dev, size);
+	if (obj == NULL)
+		return -ENOMEM;
+
+	ret = drm_gem_handle_create(file, &obj->base, &handle);
+	if (ret) {
+		drm_gem_object_release(&obj->base);
+		kfree(obj);
+		return ret;
+	}
+
+	drm_gem_object_unreference(&obj->base);
+	*handle_p = handle;
+	return 0;
+}
+
+int udl_dumb_create(struct drm_file *file,
+		    struct drm_device *dev,
+		    struct drm_mode_create_dumb *args)
+{
+	args->pitch = args->width * ((args->bpp + 1) / 8);
+	args->size = args->pitch * args->height;
+	return udl_gem_create(file, dev,
+			      args->size, &args->handle);
+}
+
+int udl_dumb_destroy(struct drm_file *file, struct drm_device *dev,
+		     uint32_t handle)
+{
+	return drm_gem_handle_delete(file, handle);
+}
+
+int udl_drm_gem_mmap(struct file *filp, struct vm_area_struct *vma)
+{
+	int ret;
+
+	ret = drm_gem_mmap(filp, vma);
+	if (ret)
+		return ret;
+
+	vma->vm_flags &= ~VM_PFNMAP;
+	vma->vm_flags |= VM_MIXEDMAP;
+
+	return ret;
+}
+
+int udl_gem_fault(struct vm_area_struct *vma, struct vm_fault *vmf)
+{
+	struct udl_gem_object *obj = to_udl_bo(vma->vm_private_data);
+	struct page *page;
+	unsigned int page_offset;
+	int ret = 0;
+
+	page_offset = ((unsigned long)vmf->virtual_address - vma->vm_start) >>
+		PAGE_SHIFT;
+
+	if (!obj->pages)
+		return VM_FAULT_SIGBUS;
+
+	page = obj->pages[page_offset];
+	ret = vm_insert_page(vma, (unsigned long)vmf->virtual_address, page);
+	switch (ret) {
+	case -EAGAIN:
+		set_need_resched();
+	case 0:
+	case -ERESTARTSYS:
+		return VM_FAULT_NOPAGE;
+	case -ENOMEM:
+		return VM_FAULT_OOM;
+	default:
+		return VM_FAULT_SIGBUS;
+	}
+}
+
+int udl_gem_init_object(struct drm_gem_object *obj)
+{
+	BUG();
+
+	return 0;
+}
+
+static int udl_gem_get_pages(struct udl_gem_object *obj, gfp_t gfpmask)
+{
+	int page_count, i;
+	struct page *page;
+	struct inode *inode;
+	struct address_space *mapping;
+
+	if (obj->pages)
+		return 0;
+
+	page_count = obj->base.size / PAGE_SIZE;
+	BUG_ON(obj->pages != NULL);
+	obj->pages = drm_malloc_ab(page_count, sizeof(struct page *));
+	if (obj->pages == NULL)
+		return -ENOMEM;
+
+	inode = obj->base.filp->f_path.dentry->d_inode;
+	mapping = inode->i_mapping;
+	gfpmask |= mapping_gfp_mask(mapping);
+
+	for (i = 0; i < page_count; i++) {
+		page = shmem_read_mapping_page_gfp(mapping, i, gfpmask);
+		if (IS_ERR(page))
+			goto err_pages;
+		obj->pages[i] = page;
+	}
+
+	return 0;
+err_pages:
+	while (i--)
+		page_cache_release(obj->pages[i]);
+	drm_free_large(obj->pages);
+	obj->pages = NULL;
+	return PTR_ERR(page);
+}
+
+static void udl_gem_put_pages(struct udl_gem_object *obj)
+{
+	int page_count = obj->base.size / PAGE_SIZE;
+	int i;
+
+	for (i = 0; i < page_count; i++)
+		page_cache_release(obj->pages[i]);
+
+	drm_free_large(obj->pages);
+	obj->pages = NULL;
+}
+
+int udl_gem_vmap(struct udl_gem_object *obj)
+{
+	int page_count = obj->base.size / PAGE_SIZE;
+	int ret;
+
+	ret = udl_gem_get_pages(obj, GFP_KERNEL);
+	if (ret)
+		return ret;
+
+	obj->vmapping = vmap(obj->pages, page_count, 0, PAGE_KERNEL);
+	if (!obj->vmapping)
+		return -ENOMEM;
+	return 0;
+}
+
+void udl_gem_vunmap(struct udl_gem_object *obj)
+{
+	if (obj->vmapping)
+		vunmap(obj->vmapping);
+
+	udl_gem_put_pages(obj);
+}
+
+void udl_gem_free_object(struct drm_gem_object *gem_obj)
+{
+	struct udl_gem_object *obj = to_udl_bo(gem_obj);
+
+	if (obj->vmapping)
+		udl_gem_vunmap(obj);
+
+	if (obj->pages)
+		udl_gem_put_pages(obj);
+
+	if (gem_obj->map_list.map)
+		drm_gem_free_mmap_offset(gem_obj);
+}
+
+/* the dumb interface doesn't work with the GEM straight MMAP
+   interface, it expects to do MMAP on the drm fd, like normal */
+int udl_gem_mmap(struct drm_file *file, struct drm_device *dev,
+		 uint32_t handle, uint64_t *offset)
+{
+	struct udl_gem_object *gobj;
+	struct drm_gem_object *obj;
+	int ret = 0;
+
+	mutex_lock(&dev->struct_mutex);
+	obj = drm_gem_object_lookup(dev, file, handle);
+	if (obj == NULL) {
+		ret = -ENOENT;
+		goto unlock;
+	}
+	gobj = to_udl_bo(obj);
+
+	ret = udl_gem_get_pages(gobj, GFP_KERNEL);
+	if (ret)
+		return ret;
+	if (!gobj->base.map_list.map) {
+		ret = drm_gem_create_mmap_offset(obj);
+		if (ret)
+			goto out;
+	}
+
+	*offset = (u64)gobj->base.map_list.hash.key << PAGE_SHIFT;
+
+out:
+	drm_gem_object_unreference(&gobj->base);
+unlock:
+	mutex_unlock(&dev->struct_mutex);
+	return ret;
+}
diff --git a/drivers/gpu/drm/udl/udl_main.c b/drivers/gpu/drm/udl/udl_main.c
new file mode 100644
index 000000000000..a8d5f09428c7
--- /dev/null
+++ b/drivers/gpu/drm/udl/udl_main.c
@@ -0,0 +1,338 @@
+/*
+ * Copyright (C) 2012 Red Hat
+ *
+ * based in parts on udlfb.c:
+ * Copyright (C) 2009 Roberto De Ioris <roberto@unbit.it>
+ * Copyright (C) 2009 Jaya Kumar <jayakumar.lkml@gmail.com>
+ * Copyright (C) 2009 Bernie Thompson <bernie@plugable.com>
+ *
+ * This file is subject to the terms and conditions of the GNU General Public
+ * License v2. See the file COPYING in the main directory of this archive for
+ * more details.
+ */
+#include "drmP.h"
+#include "udl_drv.h"
+
+/* -BULK_SIZE as per usb-skeleton. Can we get full page and avoid overhead? */
+#define BULK_SIZE 512
+
+#define MAX_TRANSFER (PAGE_SIZE*16 - BULK_SIZE)
+#define WRITES_IN_FLIGHT (4)
+#define MAX_VENDOR_DESCRIPTOR_SIZE 256
+
+#define GET_URB_TIMEOUT	HZ
+#define FREE_URB_TIMEOUT (HZ*2)
+
+static int udl_parse_vendor_descriptor(struct drm_device *dev,
+				       struct usb_device *usbdev)
+{
+	struct udl_device *udl = dev->dev_private;
+	char *desc;
+	char *buf;
+	char *desc_end;
+
+	u8 total_len = 0;
+
+	buf = kzalloc(MAX_VENDOR_DESCRIPTOR_SIZE, GFP_KERNEL);
+	if (!buf)
+		return false;
+	desc = buf;
+
+	total_len = usb_get_descriptor(usbdev, 0x5f, /* vendor specific */
+				    0, desc, MAX_VENDOR_DESCRIPTOR_SIZE);
+	if (total_len > 5) {
+		DRM_INFO("vendor descriptor length:%x data:%02x %02x %02x %02x" \
+			"%02x %02x %02x %02x %02x %02x %02x\n",
+			total_len, desc[0],
+			desc[1], desc[2], desc[3], desc[4], desc[5], desc[6],
+			desc[7], desc[8], desc[9], desc[10]);
+
+		if ((desc[0] != total_len) || /* descriptor length */
+		    (desc[1] != 0x5f) ||   /* vendor descriptor type */
+		    (desc[2] != 0x01) ||   /* version (2 bytes) */
+		    (desc[3] != 0x00) ||
+		    (desc[4] != total_len - 2)) /* length after type */
+			goto unrecognized;
+
+		desc_end = desc + total_len;
+		desc += 5; /* the fixed header we've already parsed */
+
+		while (desc < desc_end) {
+			u8 length;
+			u16 key;
+
+			key = *((u16 *) desc);
+			desc += sizeof(u16);
+			length = *desc;
+			desc++;
+
+			switch (key) {
+			case 0x0200: { /* max_area */
+				u32 max_area;
+				max_area = le32_to_cpu(*((u32 *)desc));
+				DRM_DEBUG("DL chip limited to %d pixel modes\n",
+					max_area);
+				udl->sku_pixel_limit = max_area;
+				break;
+			}
+			default:
+				break;
+			}
+			desc += length;
+		}
+	}
+
+	goto success;
+
+unrecognized:
+	/* allow udlfb to load for now even if firmware unrecognized */
+	DRM_ERROR("Unrecognized vendor firmware descriptor\n");
+
+success:
+	kfree(buf);
+	return true;
+}
+
+static void udl_release_urb_work(struct work_struct *work)
+{
+	struct urb_node *unode = container_of(work, struct urb_node,
+					      release_urb_work.work);
+
+	up(&unode->dev->urbs.limit_sem);
+}
+
+void udl_urb_completion(struct urb *urb)
+{
+	struct urb_node *unode = urb->context;
+	struct udl_device *udl = unode->dev;
+	unsigned long flags;
+
+	/* sync/async unlink faults aren't errors */
+	if (urb->status) {
+		if (!(urb->status == -ENOENT ||
+		    urb->status == -ECONNRESET ||
+		    urb->status == -ESHUTDOWN)) {
+			DRM_ERROR("%s - nonzero write bulk status received: %d\n",
+				__func__, urb->status);
+			atomic_set(&udl->lost_pixels, 1);
+		}
+	}
+
+	urb->transfer_buffer_length = udl->urbs.size; /* reset to actual */
+
+	spin_lock_irqsave(&udl->urbs.lock, flags);
+	list_add_tail(&unode->entry, &udl->urbs.list);
+	udl->urbs.available++;
+	spin_unlock_irqrestore(&udl->urbs.lock, flags);
+
+#if 0
+	/*
+	 * When using fb_defio, we deadlock if up() is called
+	 * while another is waiting. So queue to another process.
+	 */
+	if (fb_defio)
+		schedule_delayed_work(&unode->release_urb_work, 0);
+	else
+#endif
+		up(&udl->urbs.limit_sem);
+}
+
+static void udl_free_urb_list(struct drm_device *dev)
+{
+	struct udl_device *udl = dev->dev_private;
+	int count = udl->urbs.count;
+	struct list_head *node;
+	struct urb_node *unode;
+	struct urb *urb;
+	int ret;
+	unsigned long flags;
+
+	DRM_DEBUG("Waiting for completes and freeing all render urbs\n");
+
+	/* keep waiting and freeing, until we've got 'em all */
+	while (count--) {
+
+		/* Getting interrupted means a leak, but ok at shutdown*/
+		ret = down_interruptible(&udl->urbs.limit_sem);
+		if (ret)
+			break;
+
+		spin_lock_irqsave(&udl->urbs.lock, flags);
+
+		node = udl->urbs.list.next; /* have reserved one with sem */
+		list_del_init(node);
+
+		spin_unlock_irqrestore(&udl->urbs.lock, flags);
+
+		unode = list_entry(node, struct urb_node, entry);
+		urb = unode->urb;
+
+		/* Free each separately allocated piece */
+		usb_free_coherent(urb->dev, udl->urbs.size,
+				  urb->transfer_buffer, urb->transfer_dma);
+		usb_free_urb(urb);
+		kfree(node);
+	}
+	udl->urbs.count = 0;
+}
+
+static int udl_alloc_urb_list(struct drm_device *dev, int count, size_t size)
+{
+	struct udl_device *udl = dev->dev_private;
+	int i = 0;
+	struct urb *urb;
+	struct urb_node *unode;
+	char *buf;
+
+	spin_lock_init(&udl->urbs.lock);
+
+	udl->urbs.size = size;
+	INIT_LIST_HEAD(&udl->urbs.list);
+
+	while (i < count) {
+		unode = kzalloc(sizeof(struct urb_node), GFP_KERNEL);
+		if (!unode)
+			break;
+		unode->dev = udl;
+
+		INIT_DELAYED_WORK(&unode->release_urb_work,
+			  udl_release_urb_work);
+
+		urb = usb_alloc_urb(0, GFP_KERNEL);
+		if (!urb) {
+			kfree(unode);
+			break;
+		}
+		unode->urb = urb;
+
+		buf = usb_alloc_coherent(udl->ddev->usbdev, MAX_TRANSFER, GFP_KERNEL,
+					 &urb->transfer_dma);
+		if (!buf) {
+			kfree(unode);
+			usb_free_urb(urb);
+			break;
+		}
+
+		/* urb->transfer_buffer_length set to actual before submit */
+		usb_fill_bulk_urb(urb, udl->ddev->usbdev, usb_sndbulkpipe(udl->ddev->usbdev, 1),
+			buf, size, udl_urb_completion, unode);
+		urb->transfer_flags |= URB_NO_TRANSFER_DMA_MAP;
+
+		list_add_tail(&unode->entry, &udl->urbs.list);
+
+		i++;
+	}
+
+	sema_init(&udl->urbs.limit_sem, i);
+	udl->urbs.count = i;
+	udl->urbs.available = i;
+
+	DRM_DEBUG("allocated %d %d byte urbs\n", i, (int) size);
+
+	return i;
+}
+
+struct urb *udl_get_urb(struct drm_device *dev)
+{
+	struct udl_device *udl = dev->dev_private;
+	int ret = 0;
+	struct list_head *entry;
+	struct urb_node *unode;
+	struct urb *urb = NULL;
+	unsigned long flags;
+
+	/* Wait for an in-flight buffer to complete and get re-queued */
+	ret = down_timeout(&udl->urbs.limit_sem, GET_URB_TIMEOUT);
+	if (ret) {
+		atomic_set(&udl->lost_pixels, 1);
+		DRM_INFO("wait for urb interrupted: %x available: %d\n",
+		       ret, udl->urbs.available);
+		goto error;
+	}
+
+	spin_lock_irqsave(&udl->urbs.lock, flags);
+
+	BUG_ON(list_empty(&udl->urbs.list)); /* reserved one with limit_sem */
+	entry = udl->urbs.list.next;
+	list_del_init(entry);
+	udl->urbs.available--;
+
+	spin_unlock_irqrestore(&udl->urbs.lock, flags);
+
+	unode = list_entry(entry, struct urb_node, entry);
+	urb = unode->urb;
+
+error:
+	return urb;
+}
+
+int udl_submit_urb(struct drm_device *dev, struct urb *urb, size_t len)
+{
+	struct udl_device *udl = dev->dev_private;
+	int ret;
+
+	BUG_ON(len > udl->urbs.size);
+
+	urb->transfer_buffer_length = len; /* set to actual payload len */
+	ret = usb_submit_urb(urb, GFP_ATOMIC);
+	if (ret) {
+		udl_urb_completion(urb); /* because no one else will */
+		atomic_set(&udl->lost_pixels, 1);
+		DRM_ERROR("usb_submit_urb error %x\n", ret);
+	}
+	return ret;
+}
+
+int udl_driver_load(struct drm_device *dev, unsigned long flags)
+{
+	struct udl_device *udl;
+	int ret;
+
+	DRM_DEBUG("\n");
+	udl = kzalloc(sizeof(struct udl_device), GFP_KERNEL);
+	if (!udl)
+		return -ENOMEM;
+
+	udl->ddev = dev;
+	dev->dev_private = udl;
+
+	if (!udl_parse_vendor_descriptor(dev, dev->usbdev)) {
+		DRM_ERROR("firmware not recognized. Assume incompatible device\n");
+		goto err;
+	}
+
+	if (!udl_alloc_urb_list(dev, WRITES_IN_FLIGHT, MAX_TRANSFER)) {
+		ret = -ENOMEM;
+		DRM_ERROR("udl_alloc_urb_list failed\n");
+		goto err;
+	}
+
+	DRM_DEBUG("\n");
+	ret = udl_modeset_init(dev);
+
+	ret = udl_fbdev_init(dev);
+	return 0;
+err:
+	kfree(udl);
+	DRM_ERROR("%d\n", ret);
+	return ret;
+}
+
+int udl_drop_usb(struct drm_device *dev)
+{
+	udl_free_urb_list(dev);
+	return 0;
+}
+
+int udl_driver_unload(struct drm_device *dev)
+{
+	struct udl_device *udl = dev->dev_private;
+
+	if (udl->urbs.count)
+		udl_free_urb_list(dev);
+
+	udl_fbdev_cleanup(dev);
+	udl_modeset_cleanup(dev);
+	kfree(udl);
+	return 0;
+}
diff --git a/drivers/gpu/drm/udl/udl_modeset.c b/drivers/gpu/drm/udl/udl_modeset.c
new file mode 100644
index 000000000000..b3ecb3d12a1d
--- /dev/null
+++ b/drivers/gpu/drm/udl/udl_modeset.c
@@ -0,0 +1,414 @@
+/*
+ * Copyright (C) 2012 Red Hat
+ *
+ * based in parts on udlfb.c:
+ * Copyright (C) 2009 Roberto De Ioris <roberto@unbit.it>
+ * Copyright (C) 2009 Jaya Kumar <jayakumar.lkml@gmail.com>
+ * Copyright (C) 2009 Bernie Thompson <bernie@plugable.com>
+
+ * This file is subject to the terms and conditions of the GNU General Public
+ * License v2. See the file COPYING in the main directory of this archive for
+ * more details.
+ */
+
+#include "drmP.h"
+#include "drm_crtc.h"
+#include "drm_crtc_helper.h"
+#include "udl_drv.h"
+
+/*
+ * All DisplayLink bulk operations start with 0xAF, followed by specific code
+ * All operations are written to buffers which then later get sent to device
+ */
+static char *udl_set_register(char *buf, u8 reg, u8 val)
+{
+	*buf++ = 0xAF;
+	*buf++ = 0x20;
+	*buf++ = reg;
+	*buf++ = val;
+	return buf;
+}
+
+static char *udl_vidreg_lock(char *buf)
+{
+	return udl_set_register(buf, 0xFF, 0x00);
+}
+
+static char *udl_vidreg_unlock(char *buf)
+{
+	return udl_set_register(buf, 0xFF, 0xFF);
+}
+
+/*
+ * On/Off for driving the DisplayLink framebuffer to the display
+ *  0x00 H and V sync on
+ *  0x01 H and V sync off (screen blank but powered)
+ *  0x07 DPMS powerdown (requires modeset to come back)
+ */
+static char *udl_enable_hvsync(char *buf, bool enable)
+{
+	if (enable)
+		return udl_set_register(buf, 0x1F, 0x00);
+	else
+		return udl_set_register(buf, 0x1F, 0x07);
+}
+
+static char *udl_set_color_depth(char *buf, u8 selection)
+{
+	return udl_set_register(buf, 0x00, selection);
+}
+
+static char *udl_set_base16bpp(char *wrptr, u32 base)
+{
+	/* the base pointer is 16 bits wide, 0x20 is hi byte. */
+	wrptr = udl_set_register(wrptr, 0x20, base >> 16);
+	wrptr = udl_set_register(wrptr, 0x21, base >> 8);
+	return udl_set_register(wrptr, 0x22, base);
+}
+
+/*
+ * DisplayLink HW has separate 16bpp and 8bpp framebuffers.
+ * In 24bpp modes, the low 323 RGB bits go in the 8bpp framebuffer
+ */
+static char *udl_set_base8bpp(char *wrptr, u32 base)
+{
+	wrptr = udl_set_register(wrptr, 0x26, base >> 16);
+	wrptr = udl_set_register(wrptr, 0x27, base >> 8);
+	return udl_set_register(wrptr, 0x28, base);
+}
+
+static char *udl_set_register_16(char *wrptr, u8 reg, u16 value)
+{
+	wrptr = udl_set_register(wrptr, reg, value >> 8);
+	return udl_set_register(wrptr, reg+1, value);
+}
+
+/*
+ * This is kind of weird because the controller takes some
+ * register values in a different byte order than other registers.
+ */
+static char *udl_set_register_16be(char *wrptr, u8 reg, u16 value)
+{
+	wrptr = udl_set_register(wrptr, reg, value);
+	return udl_set_register(wrptr, reg+1, value >> 8);
+}
+
+/*
+ * LFSR is linear feedback shift register. The reason we have this is
+ * because the display controller needs to minimize the clock depth of
+ * various counters used in the display path. So this code reverses the
+ * provided value into the lfsr16 value by counting backwards to get
+ * the value that needs to be set in the hardware comparator to get the
+ * same actual count. This makes sense once you read above a couple of
+ * times and think about it from a hardware perspective.
+ */
+static u16 udl_lfsr16(u16 actual_count)
+{
+	u32 lv = 0xFFFF; /* This is the lfsr value that the hw starts with */
+
+	while (actual_count--) {
+		lv =	 ((lv << 1) |
+			(((lv >> 15) ^ (lv >> 4) ^ (lv >> 2) ^ (lv >> 1)) & 1))
+			& 0xFFFF;
+	}
+
+	return (u16) lv;
+}
+
+/*
+ * This does LFSR conversion on the value that is to be written.
+ * See LFSR explanation above for more detail.
+ */
+static char *udl_set_register_lfsr16(char *wrptr, u8 reg, u16 value)
+{
+	return udl_set_register_16(wrptr, reg, udl_lfsr16(value));
+}
+
+/*
+ * This takes a standard fbdev screeninfo struct and all of its monitor mode
+ * details and converts them into the DisplayLink equivalent register commands.
+  ERR(vreg(dev,               0x00, (color_depth == 16) ? 0 : 1));
+  ERR(vreg_lfsr16(dev,        0x01, xDisplayStart));
+  ERR(vreg_lfsr16(dev,        0x03, xDisplayEnd));
+  ERR(vreg_lfsr16(dev,        0x05, yDisplayStart));
+  ERR(vreg_lfsr16(dev,        0x07, yDisplayEnd));
+  ERR(vreg_lfsr16(dev,        0x09, xEndCount));
+  ERR(vreg_lfsr16(dev,        0x0B, hSyncStart));
+  ERR(vreg_lfsr16(dev,        0x0D, hSyncEnd));
+  ERR(vreg_big_endian(dev,    0x0F, hPixels));
+  ERR(vreg_lfsr16(dev,        0x11, yEndCount));
+  ERR(vreg_lfsr16(dev,        0x13, vSyncStart));
+  ERR(vreg_lfsr16(dev,        0x15, vSyncEnd));
+  ERR(vreg_big_endian(dev,    0x17, vPixels));
+  ERR(vreg_little_endian(dev, 0x1B, pixelClock5KHz));
+
+  ERR(vreg(dev,               0x1F, 0));
+
+  ERR(vbuf(dev, WRITE_VIDREG_UNLOCK, DSIZEOF(WRITE_VIDREG_UNLOCK)));
+ */
+static char *udl_set_vid_cmds(char *wrptr, struct drm_display_mode *mode)
+{
+	u16 xds, yds;
+	u16 xde, yde;
+	u16 yec;
+
+	/* x display start */
+	xds = mode->crtc_htotal - mode->crtc_hsync_start;
+	wrptr = udl_set_register_lfsr16(wrptr, 0x01, xds);
+	/* x display end */
+	xde = xds + mode->crtc_hdisplay;
+	wrptr = udl_set_register_lfsr16(wrptr, 0x03, xde);
+
+	/* y display start */
+	yds = mode->crtc_vtotal - mode->crtc_vsync_start;
+	wrptr = udl_set_register_lfsr16(wrptr, 0x05, yds);
+	/* y display end */
+	yde = yds + mode->crtc_vdisplay;
+	wrptr = udl_set_register_lfsr16(wrptr, 0x07, yde);
+
+	/* x end count is active + blanking - 1 */
+	wrptr = udl_set_register_lfsr16(wrptr, 0x09,
+					mode->crtc_htotal - 1);
+
+	/* libdlo hardcodes hsync start to 1 */
+	wrptr = udl_set_register_lfsr16(wrptr, 0x0B, 1);
+
+	/* hsync end is width of sync pulse + 1 */
+	wrptr = udl_set_register_lfsr16(wrptr, 0x0D,
+					mode->crtc_hsync_end - mode->crtc_hsync_start + 1);
+
+	/* hpixels is active pixels */
+	wrptr = udl_set_register_16(wrptr, 0x0F, mode->hdisplay);
+
+	/* yendcount is vertical active + vertical blanking */
+	yec = mode->crtc_vtotal;
+	wrptr = udl_set_register_lfsr16(wrptr, 0x11, yec);
+
+	/* libdlo hardcodes vsync start to 0 */
+	wrptr = udl_set_register_lfsr16(wrptr, 0x13, 0);
+
+	/* vsync end is width of vsync pulse */
+	wrptr = udl_set_register_lfsr16(wrptr, 0x15, mode->crtc_vsync_end - mode->crtc_vsync_start);
+
+	/* vpixels is active pixels */
+	wrptr = udl_set_register_16(wrptr, 0x17, mode->crtc_vdisplay);
+
+	wrptr = udl_set_register_16be(wrptr, 0x1B,
+				      mode->clock / 5);
+
+	return wrptr;
+}
+
+static int udl_crtc_write_mode_to_hw(struct drm_crtc *crtc)
+{
+	struct drm_device *dev = crtc->dev;
+	struct udl_device *udl = dev->dev_private;
+	struct urb *urb;
+	char *buf;
+	int retval;
+
+	urb = udl_get_urb(dev);
+	if (!urb)
+		return -ENOMEM;
+
+	buf = (char *)urb->transfer_buffer;
+
+	memcpy(buf, udl->mode_buf, udl->mode_buf_len);
+	retval = udl_submit_urb(dev, urb, udl->mode_buf_len);
+	DRM_INFO("write mode info %d\n", udl->mode_buf_len);
+	return retval;
+}
+
+
+static void udl_crtc_dpms(struct drm_crtc *crtc, int mode)
+{
+	struct drm_device *dev = crtc->dev;
+	struct udl_device *udl = dev->dev_private;
+	int retval;
+
+	if (mode == DRM_MODE_DPMS_OFF) {
+		char *buf;
+		struct urb *urb;
+		urb = udl_get_urb(dev);
+		if (!urb)
+			return;
+
+		buf = (char *)urb->transfer_buffer;
+		buf = udl_vidreg_lock(buf);
+		buf = udl_enable_hvsync(buf, false);
+		buf = udl_vidreg_unlock(buf);
+
+		retval = udl_submit_urb(dev, urb, buf - (char *)
+					urb->transfer_buffer);
+	} else {
+		if (udl->mode_buf_len == 0) {
+			DRM_ERROR("Trying to enable DPMS with no mode\n");
+			return;
+		}
+		udl_crtc_write_mode_to_hw(crtc);
+	}
+
+}
+
+static bool udl_crtc_mode_fixup(struct drm_crtc *crtc,
+				  struct drm_display_mode *mode,
+				  struct drm_display_mode *adjusted_mode)
+
+{
+	return true;
+}
+
+#if 0
+static int
+udl_pipe_set_base_atomic(struct drm_crtc *crtc, struct drm_framebuffer *fb,
+			   int x, int y, enum mode_set_atomic state)
+{
+	return 0;
+}
+
+static int
+udl_pipe_set_base(struct drm_crtc *crtc, int x, int y,
+		    struct drm_framebuffer *old_fb)
+{
+	return 0;
+}
+#endif
+
+static int udl_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 udl_framebuffer *ufb = to_udl_fb(crtc->fb);
+	struct udl_device *udl = dev->dev_private;
+	char *buf;
+	char *wrptr;
+	int color_depth = 0;
+
+	buf = (char *)udl->mode_buf;
+
+	/* for now we just clip 24 -> 16 - if we fix that fix this */
+	/*if  (crtc->fb->bits_per_pixel != 16)
+	  color_depth = 1; */
+
+	/* This first section has to do with setting the base address on the
+	* controller * associated with the display. There are 2 base
+	* pointers, currently, we only * use the 16 bpp segment.
+	*/
+	wrptr = udl_vidreg_lock(buf);
+	wrptr = udl_set_color_depth(wrptr, color_depth);
+	/* set base for 16bpp segment to 0 */
+	wrptr = udl_set_base16bpp(wrptr, 0);
+	/* set base for 8bpp segment to end of fb */
+	wrptr = udl_set_base8bpp(wrptr, 2 * mode->vdisplay * mode->hdisplay);
+
+	wrptr = udl_set_vid_cmds(wrptr, adjusted_mode);
+	wrptr = udl_enable_hvsync(wrptr, true);
+	wrptr = udl_vidreg_unlock(wrptr);
+
+	ufb->active_16 = true;
+	if (old_fb) {
+		struct udl_framebuffer *uold_fb = to_udl_fb(old_fb);
+		uold_fb->active_16 = false;
+	}
+	udl->mode_buf_len = wrptr - buf;
+
+	/* damage all of it */
+	udl_handle_damage(ufb, 0, 0, ufb->base.width, ufb->base.height);
+	return 0;
+}
+
+
+static void udl_crtc_disable(struct drm_crtc *crtc)
+{
+
+
+}
+
+static void udl_crtc_destroy(struct drm_crtc *crtc)
+{
+	drm_crtc_cleanup(crtc);
+	kfree(crtc);
+}
+
+static void udl_load_lut(struct drm_crtc *crtc)
+{
+}
+
+static void udl_crtc_prepare(struct drm_crtc *crtc)
+{
+}
+
+static void udl_crtc_commit(struct drm_crtc *crtc)
+{
+	udl_crtc_dpms(crtc, DRM_MODE_DPMS_ON);
+}
+
+static struct drm_crtc_helper_funcs udl_helper_funcs = {
+	.dpms = udl_crtc_dpms,
+	.mode_fixup = udl_crtc_mode_fixup,
+	.mode_set = udl_crtc_mode_set,
+	.prepare = udl_crtc_prepare,
+	.commit = udl_crtc_commit,
+	.disable = udl_crtc_disable,
+	.load_lut = udl_load_lut,
+};
+
+static const struct drm_crtc_funcs udl_crtc_funcs = {
+	.set_config = drm_crtc_helper_set_config,
+	.destroy = udl_crtc_destroy,
+};
+
+int udl_crtc_init(struct drm_device *dev)
+{
+	struct drm_crtc *crtc;
+
+	crtc = kzalloc(sizeof(struct drm_crtc) + sizeof(struct drm_connector *), GFP_KERNEL);
+	if (crtc == NULL)
+		return -ENOMEM;
+
+	drm_crtc_init(dev, crtc, &udl_crtc_funcs);
+	drm_crtc_helper_add(crtc, &udl_helper_funcs);
+
+	return 0;
+}
+
+static const struct drm_mode_config_funcs udl_mode_funcs = {
+	.fb_create = udl_fb_user_fb_create,
+	.output_poll_changed = NULL,
+};
+
+int udl_modeset_init(struct drm_device *dev)
+{
+	struct drm_encoder *encoder;
+	drm_mode_config_init(dev);
+
+	dev->mode_config.min_width = 640;
+	dev->mode_config.min_height = 480;
+
+	dev->mode_config.max_width = 2048;
+	dev->mode_config.max_height = 2048;
+
+	dev->mode_config.prefer_shadow = 0;
+	dev->mode_config.preferred_depth = 24;
+
+	dev->mode_config.funcs = (void *)&udl_mode_funcs;
+
+	drm_mode_create_dirty_info_property(dev);
+
+	udl_crtc_init(dev);
+
+	encoder = udl_encoder_init(dev);
+
+	udl_connector_init(dev, encoder);
+
+	return 0;
+}
+
+void udl_modeset_cleanup(struct drm_device *dev)
+{
+	drm_mode_config_cleanup(dev);
+}
diff --git a/drivers/gpu/drm/udl/udl_transfer.c b/drivers/gpu/drm/udl/udl_transfer.c
new file mode 100644
index 000000000000..b9320e2608dd
--- /dev/null
+++ b/drivers/gpu/drm/udl/udl_transfer.c
@@ -0,0 +1,253 @@
+/*
+ * Copyright (C) 2012 Red Hat
+ * based in parts on udlfb.c:
+ * Copyright (C) 2009 Roberto De Ioris <roberto@unbit.it>
+ * Copyright (C) 2009 Jaya Kumar <jayakumar.lkml@gmail.com>
+ * Copyright (C) 2009 Bernie Thompson <bernie@plugable.com>
+ *
+ * This file is subject to the terms and conditions of the GNU General Public
+ * License v2. See the file COPYING in the main directory of this archive for
+ * more details.
+ */
+
+#include <linux/module.h>
+#include <linux/slab.h>
+#include <linux/fb.h>
+#include <linux/prefetch.h>
+
+#include "drmP.h"
+#include "udl_drv.h"
+
+#define MAX_CMD_PIXELS		255
+
+#define RLX_HEADER_BYTES	7
+#define MIN_RLX_PIX_BYTES       4
+#define MIN_RLX_CMD_BYTES	(RLX_HEADER_BYTES + MIN_RLX_PIX_BYTES)
+
+#define RLE_HEADER_BYTES	6
+#define MIN_RLE_PIX_BYTES	3
+#define MIN_RLE_CMD_BYTES	(RLE_HEADER_BYTES + MIN_RLE_PIX_BYTES)
+
+#define RAW_HEADER_BYTES	6
+#define MIN_RAW_PIX_BYTES	2
+#define MIN_RAW_CMD_BYTES	(RAW_HEADER_BYTES + MIN_RAW_PIX_BYTES)
+
+/*
+ * Trims identical data from front and back of line
+ * Sets new front buffer address and width
+ * And returns byte count of identical pixels
+ * Assumes CPU natural alignment (unsigned long)
+ * for back and front buffer ptrs and width
+ */
+#if 0
+static int udl_trim_hline(const u8 *bback, const u8 **bfront, int *width_bytes)
+{
+	int j, k;
+	const unsigned long *back = (const unsigned long *) bback;
+	const unsigned long *front = (const unsigned long *) *bfront;
+	const int width = *width_bytes / sizeof(unsigned long);
+	int identical = width;
+	int start = width;
+	int end = width;
+
+	prefetch((void *) front);
+	prefetch((void *) back);
+
+	for (j = 0; j < width; j++) {
+		if (back[j] != front[j]) {
+			start = j;
+			break;
+		}
+	}
+
+	for (k = width - 1; k > j; k--) {
+		if (back[k] != front[k]) {
+			end = k+1;
+			break;
+		}
+	}
+
+	identical = start + (width - end);
+	*bfront = (u8 *) &front[start];
+	*width_bytes = (end - start) * sizeof(unsigned long);
+
+	return identical * sizeof(unsigned long);
+}
+#endif
+
+static inline u16 pixel32_to_be16p(const uint8_t *pixel)
+{
+	uint32_t pix = *(uint32_t *)pixel;
+	u16 retval;
+
+	retval =  (((pix >> 3) & 0x001f) |
+		   ((pix >> 5) & 0x07e0) |
+		   ((pix >> 8) & 0xf800));
+	return retval;
+}
+
+/*
+ * Render a command stream for an encoded horizontal line segment of pixels.
+ *
+ * A command buffer holds several commands.
+ * It always begins with a fresh command header
+ * (the protocol doesn't require this, but we enforce it to allow
+ * multiple buffers to be potentially encoded and sent in parallel).
+ * A single command encodes one contiguous horizontal line of pixels
+ *
+ * The function relies on the client to do all allocation, so that
+ * rendering can be done directly to output buffers (e.g. USB URBs).
+ * The function fills the supplied command buffer, providing information
+ * on where it left off, so the client may call in again with additional
+ * buffers if the line will take several buffers to complete.
+ *
+ * A single command can transmit a maximum of 256 pixels,
+ * regardless of the compression ratio (protocol design limit).
+ * To the hardware, 0 for a size byte means 256
+ *
+ * Rather than 256 pixel commands which are either rl or raw encoded,
+ * the rlx command simply assumes alternating raw and rl spans within one cmd.
+ * This has a slightly larger header overhead, but produces more even results.
+ * It also processes all data (read and write) in a single pass.
+ * Performance benchmarks of common cases show it having just slightly better
+ * compression than 256 pixel raw or rle commands, with similar CPU consumpion.
+ * But for very rl friendly data, will compress not quite as well.
+ */
+static void udl_compress_hline16(
+	const u8 **pixel_start_ptr,
+	const u8 *const pixel_end,
+	uint32_t *device_address_ptr,
+	uint8_t **command_buffer_ptr,
+	const uint8_t *const cmd_buffer_end, int bpp)
+{
+	const u8 *pixel = *pixel_start_ptr;
+	uint32_t dev_addr  = *device_address_ptr;
+	uint8_t *cmd = *command_buffer_ptr;
+
+	while ((pixel_end > pixel) &&
+	       (cmd_buffer_end - MIN_RLX_CMD_BYTES > cmd)) {
+		uint8_t *raw_pixels_count_byte = 0;
+		uint8_t *cmd_pixels_count_byte = 0;
+		const u8 *raw_pixel_start = 0;
+		const u8 *cmd_pixel_start, *cmd_pixel_end = 0;
+
+		prefetchw((void *) cmd); /* pull in one cache line at least */
+
+		*cmd++ = 0xaf;
+		*cmd++ = 0x6b;
+		*cmd++ = (uint8_t) ((dev_addr >> 16) & 0xFF);
+		*cmd++ = (uint8_t) ((dev_addr >> 8) & 0xFF);
+		*cmd++ = (uint8_t) ((dev_addr) & 0xFF);
+
+		cmd_pixels_count_byte = cmd++; /*  we'll know this later */
+		cmd_pixel_start = pixel;
+
+		raw_pixels_count_byte = cmd++; /*  we'll know this later */
+		raw_pixel_start = pixel;
+
+		cmd_pixel_end = pixel + (min(MAX_CMD_PIXELS + 1,
+			min((int)(pixel_end - pixel) / bpp,
+			    (int)(cmd_buffer_end - cmd) / 2))) * bpp;
+
+		prefetch_range((void *) pixel, (cmd_pixel_end - pixel) * bpp);
+
+		while (pixel < cmd_pixel_end) {
+			const u8 * const repeating_pixel = pixel;
+
+			if (bpp == 2)
+				*(uint16_t *)cmd = cpu_to_be16p((uint16_t *)pixel);
+			else if (bpp == 4)
+				*(uint16_t *)cmd = cpu_to_be16(pixel32_to_be16p(pixel));
+
+			cmd += 2;
+			pixel += bpp;
+
+			if (unlikely((pixel < cmd_pixel_end) &&
+				     (!memcmp(pixel, repeating_pixel, bpp)))) {
+				/* go back and fill in raw pixel count */
+				*raw_pixels_count_byte = (((repeating_pixel -
+						raw_pixel_start) / bpp) + 1) & 0xFF;
+
+				while ((pixel < cmd_pixel_end)
+				       && (!memcmp(pixel, repeating_pixel, bpp))) {
+					pixel += bpp;
+				}
+
+				/* immediately after raw data is repeat byte */
+				*cmd++ = (((pixel - repeating_pixel) / bpp) - 1) & 0xFF;
+
+				/* Then start another raw pixel span */
+				raw_pixel_start = pixel;
+				raw_pixels_count_byte = cmd++;
+			}
+		}
+
+		if (pixel > raw_pixel_start) {
+			/* finalize last RAW span */
+			*raw_pixels_count_byte = ((pixel-raw_pixel_start) / bpp) & 0xFF;
+		}
+
+		*cmd_pixels_count_byte = ((pixel - cmd_pixel_start) / bpp) & 0xFF;
+		dev_addr += ((pixel - cmd_pixel_start) / bpp) * 2;
+	}
+
+	if (cmd_buffer_end <= MIN_RLX_CMD_BYTES + cmd) {
+		/* Fill leftover bytes with no-ops */
+		if (cmd_buffer_end > cmd)
+			memset(cmd, 0xAF, cmd_buffer_end - cmd);
+		cmd = (uint8_t *) cmd_buffer_end;
+	}
+
+	*command_buffer_ptr = cmd;
+	*pixel_start_ptr = pixel;
+	*device_address_ptr = dev_addr;
+
+	return;
+}
+
+/*
+ * There are 3 copies of every pixel: The front buffer that the fbdev
+ * client renders to, the actual framebuffer across the USB bus in hardware
+ * (that we can only write to, slowly, and can never read), and (optionally)
+ * our shadow copy that tracks what's been sent to that hardware buffer.
+ */
+int udl_render_hline(struct drm_device *dev, int bpp, struct urb **urb_ptr,
+		     const char *front, char **urb_buf_ptr,
+		     u32 byte_offset, u32 byte_width,
+		     int *ident_ptr, int *sent_ptr)
+{
+	const u8 *line_start, *line_end, *next_pixel;
+	u32 base16 = 0 + (byte_offset / bpp) * 2;
+	struct urb *urb = *urb_ptr;
+	u8 *cmd = *urb_buf_ptr;
+	u8 *cmd_end = (u8 *) urb->transfer_buffer + urb->transfer_buffer_length;
+
+	line_start = (u8 *) (front + byte_offset);
+	next_pixel = line_start;
+	line_end = next_pixel + byte_width;
+
+	while (next_pixel < line_end) {
+
+		udl_compress_hline16(&next_pixel,
+			     line_end, &base16,
+			     (u8 **) &cmd, (u8 *) cmd_end, bpp);
+
+		if (cmd >= cmd_end) {
+			int len = cmd - (u8 *) urb->transfer_buffer;
+			if (udl_submit_urb(dev, urb, len))
+				return 1; /* lost pixels is set */
+			*sent_ptr += len;
+			urb = udl_get_urb(dev);
+			if (!urb)
+				return 1; /* lost_pixels is set */
+			*urb_ptr = urb;
+			cmd = urb->transfer_buffer;
+			cmd_end = &cmd[urb->transfer_buffer_length];
+		}
+	}
+
+	*urb_buf_ptr = cmd;
+
+	return 0;
+}
+
diff --git a/drivers/gpu/drm/via/via_map.c b/drivers/gpu/drm/via/via_map.c
index a2ab34365151..1f182254e81e 100644
--- a/drivers/gpu/drm/via/via_map.c
+++ b/drivers/gpu/drm/via/via_map.c
@@ -106,6 +106,8 @@ int via_driver_load(struct drm_device *dev, unsigned long chipset)
 
 	idr_init(&dev->object_name_idr);
 
+	pci_set_master(dev->pdev);
+
 	ret = drm_vblank_init(dev, 1);
 	if (ret) {
 		kfree(dev_priv);
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c
index 2d6f573bfff2..ee24d216aa85 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c
@@ -38,6 +38,10 @@
 #define VMWGFX_CHIP_SVGAII 0
 #define VMW_FB_RESERVATION 0
 
+#define VMW_MIN_INITIAL_WIDTH 800
+#define VMW_MIN_INITIAL_HEIGHT 600
+
+
 /**
  * Fully encoded drm commands. Might move to vmw_drm.h
  */
@@ -387,6 +391,41 @@ void vmw_3d_resource_dec(struct vmw_private *dev_priv,
 	BUG_ON(n3d < 0);
 }
 
+/**
+ * Sets the initial_[width|height] fields on the given vmw_private.
+ *
+ * It does so by reading SVGA_REG_[WIDTH|HEIGHT] regs and then
+ * clamping the value to fb_max_[width|height] fields and the
+ * VMW_MIN_INITIAL_[WIDTH|HEIGHT].
+ * If the values appear to be invalid, set them to
+ * VMW_MIN_INITIAL_[WIDTH|HEIGHT].
+ */
+static void vmw_get_initial_size(struct vmw_private *dev_priv)
+{
+	uint32_t width;
+	uint32_t height;
+
+	width = vmw_read(dev_priv, SVGA_REG_WIDTH);
+	height = vmw_read(dev_priv, SVGA_REG_HEIGHT);
+
+	width = max_t(uint32_t, width, VMW_MIN_INITIAL_WIDTH);
+	height = max_t(uint32_t, height, VMW_MIN_INITIAL_HEIGHT);
+
+	if (width > dev_priv->fb_max_width ||
+	    height > dev_priv->fb_max_height) {
+
+		/*
+		 * This is a host error and shouldn't occur.
+		 */
+
+		width = VMW_MIN_INITIAL_WIDTH;
+		height = VMW_MIN_INITIAL_HEIGHT;
+	}
+
+	dev_priv->initial_width = width;
+	dev_priv->initial_height = height;
+}
+
 static int vmw_driver_load(struct drm_device *dev, unsigned long chipset)
 {
 	struct vmw_private *dev_priv;
@@ -400,6 +439,8 @@ static int vmw_driver_load(struct drm_device *dev, unsigned long chipset)
 	}
 	memset(dev_priv, 0, sizeof(*dev_priv));
 
+	pci_set_master(dev->pdev);
+
 	dev_priv->dev = dev;
 	dev_priv->vmw_chipset = chipset;
 	dev_priv->last_read_seqno = (uint32_t) -100;
@@ -441,6 +482,9 @@ static int vmw_driver_load(struct drm_device *dev, unsigned long chipset)
 	dev_priv->mmio_size = vmw_read(dev_priv, SVGA_REG_MEM_SIZE);
 	dev_priv->fb_max_width = vmw_read(dev_priv, SVGA_REG_MAX_WIDTH);
 	dev_priv->fb_max_height = vmw_read(dev_priv, SVGA_REG_MAX_HEIGHT);
+
+	vmw_get_initial_size(dev_priv);
+
 	if (dev_priv->capabilities & SVGA_CAP_GMR) {
 		dev_priv->max_gmr_descriptors =
 			vmw_read(dev_priv,
@@ -688,6 +732,15 @@ static int vmw_driver_unload(struct drm_device *dev)
 	return 0;
 }
 
+static void vmw_preclose(struct drm_device *dev,
+			 struct drm_file *file_priv)
+{
+	struct vmw_fpriv *vmw_fp = vmw_fpriv(file_priv);
+	struct vmw_private *dev_priv = vmw_priv(dev);
+
+	vmw_event_fence_fpriv_gone(dev_priv->fman, &vmw_fp->fence_events);
+}
+
 static void vmw_postclose(struct drm_device *dev,
 			 struct drm_file *file_priv)
 {
@@ -710,6 +763,7 @@ static int vmw_driver_open(struct drm_device *dev, struct drm_file *file_priv)
 	if (unlikely(vmw_fp == NULL))
 		return ret;
 
+	INIT_LIST_HEAD(&vmw_fp->fence_events);
 	vmw_fp->tfile = ttm_object_file_init(dev_priv->tdev, 10);
 	if (unlikely(vmw_fp->tfile == NULL))
 		goto out_no_tfile;
@@ -1102,6 +1156,7 @@ static struct drm_driver driver = {
 	.master_set = vmw_master_set,
 	.master_drop = vmw_master_drop,
 	.open = vmw_driver_open,
+	.preclose = vmw_preclose,
 	.postclose = vmw_postclose,
 	.fops = &vmwgfx_driver_fops,
 	.name = VMWGFX_DRIVER_NAME,
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h
index dc279706ca70..d0f2c079ee27 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h
@@ -40,9 +40,9 @@
 #include "ttm/ttm_module.h"
 #include "vmwgfx_fence.h"
 
-#define VMWGFX_DRIVER_DATE "20111025"
+#define VMWGFX_DRIVER_DATE "20120209"
 #define VMWGFX_DRIVER_MAJOR 2
-#define VMWGFX_DRIVER_MINOR 3
+#define VMWGFX_DRIVER_MINOR 4
 #define VMWGFX_DRIVER_PATCHLEVEL 0
 #define VMWGFX_FILE_PAGE_OFFSET 0x00100000
 #define VMWGFX_FIFO_STATIC_SIZE (1024*1024)
@@ -62,6 +62,7 @@
 struct vmw_fpriv {
 	struct drm_master *locked_master;
 	struct ttm_object_file *tfile;
+	struct list_head fence_events;
 };
 
 struct vmw_dma_buffer {
@@ -202,6 +203,8 @@ struct vmw_private {
 	uint32_t mmio_size;
 	uint32_t fb_max_width;
 	uint32_t fb_max_height;
+	uint32_t initial_width;
+	uint32_t initial_height;
 	__le32 __iomem *mmio_virt;
 	int mmio_mtrr;
 	uint32_t capabilities;
@@ -533,7 +536,8 @@ extern int vmw_execbuf_process(struct drm_file *file_priv,
 			       uint32_t command_size,
 			       uint64_t throttle_us,
 			       struct drm_vmw_fence_rep __user
-			       *user_fence_rep);
+			       *user_fence_rep,
+			       struct vmw_fence_obj **out_fence);
 
 extern void
 vmw_execbuf_release_pinned_bo(struct vmw_private *dev_priv,
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_execbuf.c b/drivers/gpu/drm/vmwgfx/vmwgfx_execbuf.c
index 40932fbdac0f..4acced44a623 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_execbuf.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_execbuf.c
@@ -1109,10 +1109,11 @@ int vmw_execbuf_process(struct drm_file *file_priv,
 			void *kernel_commands,
 			uint32_t command_size,
 			uint64_t throttle_us,
-			struct drm_vmw_fence_rep __user *user_fence_rep)
+			struct drm_vmw_fence_rep __user *user_fence_rep,
+			struct vmw_fence_obj **out_fence)
 {
 	struct vmw_sw_context *sw_context = &dev_priv->ctx;
-	struct vmw_fence_obj *fence;
+	struct vmw_fence_obj *fence = NULL;
 	uint32_t handle;
 	void *cmd;
 	int ret;
@@ -1208,8 +1209,13 @@ int vmw_execbuf_process(struct drm_file *file_priv,
 	vmw_execbuf_copy_fence_user(dev_priv, vmw_fpriv(file_priv), ret,
 				    user_fence_rep, fence, handle);
 
-	if (likely(fence != NULL))
+	/* Don't unreference when handing fence out */
+	if (unlikely(out_fence != NULL)) {
+		*out_fence = fence;
+		fence = NULL;
+	} else if (likely(fence != NULL)) {
 		vmw_fence_obj_unreference(&fence);
+	}
 
 	mutex_unlock(&dev_priv->cmdbuf_mutex);
 	return 0;
@@ -1362,7 +1368,8 @@ int vmw_execbuf_ioctl(struct drm_device *dev, void *data,
 	ret = vmw_execbuf_process(file_priv, dev_priv,
 				  (void __user *)(unsigned long)arg->commands,
 				  NULL, arg->command_size, arg->throttle_us,
-				  (void __user *)(unsigned long)arg->fence_rep);
+				  (void __user *)(unsigned long)arg->fence_rep,
+				  NULL);
 
 	if (unlikely(ret != 0))
 		goto out_unlock;
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_fb.c b/drivers/gpu/drm/vmwgfx/vmwgfx_fb.c
index 34e51a1695b8..3c447bf317cb 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_fb.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_fb.c
@@ -414,10 +414,6 @@ int vmw_fb_init(struct vmw_private *vmw_priv)
 	unsigned fb_bpp, fb_depth, fb_offset, fb_pitch, fb_size;
 	int ret;
 
-	/* XXX These shouldn't be hardcoded. */
-	initial_width = 800;
-	initial_height = 600;
-
 	fb_bpp = 32;
 	fb_depth = 24;
 
@@ -425,8 +421,8 @@ int vmw_fb_init(struct vmw_private *vmw_priv)
 	fb_width = min(vmw_priv->fb_max_width, (unsigned)2048);
 	fb_height = min(vmw_priv->fb_max_height, (unsigned)2048);
 
-	initial_width = min(fb_width, initial_width);
-	initial_height = min(fb_height, initial_height);
+	initial_width = min(vmw_priv->initial_width, fb_width);
+	initial_height = min(vmw_priv->initial_height, fb_height);
 
 	fb_pitch = fb_width * fb_bpp / 8;
 	fb_size = fb_pitch * fb_height;
@@ -515,19 +511,7 @@ int vmw_fb_init(struct vmw_private *vmw_priv)
 	info->var.xres = initial_width;
 	info->var.yres = initial_height;
 
-#if 0
-	info->pixmap.size = 64*1024;
-	info->pixmap.buf_align = 8;
-	info->pixmap.access_align = 32;
-	info->pixmap.flags = FB_PIXMAP_SYSTEM;
-	info->pixmap.scan_align = 1;
-#else
-	info->pixmap.size = 0;
-	info->pixmap.buf_align = 8;
-	info->pixmap.access_align = 32;
-	info->pixmap.flags = FB_PIXMAP_SYSTEM;
-	info->pixmap.scan_align = 1;
-#endif
+	/* Use default scratch pixmap (info->pixmap.flags = FB_PIXMAP_SYSTEM) */
 
 	info->apertures = alloc_apertures(1);
 	if (!info->apertures) {
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_fence.c b/drivers/gpu/drm/vmwgfx/vmwgfx_fence.c
index 15fb26088d68..f2fb8f15e2f1 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_fence.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_fence.c
@@ -69,12 +69,13 @@ struct vmw_user_fence {
  * be assigned the current time tv_usec val when the fence signals.
  */
 struct vmw_event_fence_action {
-	struct drm_pending_event e;
 	struct vmw_fence_action action;
+	struct list_head fpriv_head;
+
+	struct drm_pending_event *event;
 	struct vmw_fence_obj *fence;
 	struct drm_device *dev;
-	struct kref kref;
-	uint32_t size;
+
 	uint32_t *tv_sec;
 	uint32_t *tv_usec;
 };
@@ -784,46 +785,40 @@ int vmw_fence_obj_unref_ioctl(struct drm_device *dev, void *data,
 }
 
 /**
- * vmw_event_fence_action_destroy
- *
- * @kref: The struct kref embedded in a struct vmw_event_fence_action.
- *
- * The vmw_event_fence_action destructor that may be called either after
- * the fence action cleanup, or when the event is delivered.
- * It frees both the vmw_event_fence_action struct and the actual
- * event structure copied to user-space.
- */
-static void vmw_event_fence_action_destroy(struct kref *kref)
-{
-	struct vmw_event_fence_action *eaction =
-		container_of(kref, struct vmw_event_fence_action, kref);
-	struct ttm_mem_global *mem_glob =
-		vmw_mem_glob(vmw_priv(eaction->dev));
-	uint32_t size = eaction->size;
-
-	kfree(eaction->e.event);
-	kfree(eaction);
-	ttm_mem_global_free(mem_glob, size);
-}
-
-
-/**
- * vmw_event_fence_action_delivered
+ * vmw_event_fence_fpriv_gone - Remove references to struct drm_file objects
  *
- * @e: The struct drm_pending_event embedded in a struct
- * vmw_event_fence_action.
+ * @fman: Pointer to a struct vmw_fence_manager
+ * @event_list: Pointer to linked list of struct vmw_event_fence_action objects
+ * with pointers to a struct drm_file object about to be closed.
  *
- * The struct drm_pending_event destructor that is called by drm
- * once the event is delivered. Since we don't know whether this function
- * will be called before or after the fence action destructor, we
- * free a refcount and destroy if it becomes zero.
+ * This function removes all pending fence events with references to a
+ * specific struct drm_file object about to be closed. The caller is required
+ * to pass a list of all struct vmw_event_fence_action objects with such
+ * events attached. This function is typically called before the
+ * struct drm_file object's event management is taken down.
  */
-static void vmw_event_fence_action_delivered(struct drm_pending_event *e)
+void vmw_event_fence_fpriv_gone(struct vmw_fence_manager *fman,
+				struct list_head *event_list)
 {
-	struct vmw_event_fence_action *eaction =
-		container_of(e, struct vmw_event_fence_action, e);
+	struct vmw_event_fence_action *eaction;
+	struct drm_pending_event *event;
+	unsigned long irq_flags;
 
-	kref_put(&eaction->kref, vmw_event_fence_action_destroy);
+	while (1) {
+		spin_lock_irqsave(&fman->lock, irq_flags);
+		if (list_empty(event_list))
+			goto out_unlock;
+		eaction = list_first_entry(event_list,
+					   struct vmw_event_fence_action,
+					   fpriv_head);
+		list_del_init(&eaction->fpriv_head);
+		event = eaction->event;
+		eaction->event = NULL;
+		spin_unlock_irqrestore(&fman->lock, irq_flags);
+		event->destroy(event);
+	}
+out_unlock:
+	spin_unlock_irqrestore(&fman->lock, irq_flags);
 }
 
 
@@ -836,18 +831,21 @@ static void vmw_event_fence_action_delivered(struct drm_pending_event *e)
  * This function is called when the seqno of the fence where @action is
  * attached has passed. It queues the event on the submitter's event list.
  * This function is always called from atomic context, and may be called
- * from irq context. It ups a refcount reflecting that we now have two
- * destructors.
+ * from irq context.
  */
 static void vmw_event_fence_action_seq_passed(struct vmw_fence_action *action)
 {
 	struct vmw_event_fence_action *eaction =
 		container_of(action, struct vmw_event_fence_action, action);
 	struct drm_device *dev = eaction->dev;
-	struct drm_file *file_priv = eaction->e.file_priv;
+	struct drm_pending_event *event = eaction->event;
+	struct drm_file *file_priv;
 	unsigned long irq_flags;
 
-	kref_get(&eaction->kref);
+	if (unlikely(event == NULL))
+		return;
+
+	file_priv = event->file_priv;
 	spin_lock_irqsave(&dev->event_lock, irq_flags);
 
 	if (likely(eaction->tv_sec != NULL)) {
@@ -858,7 +856,9 @@ static void vmw_event_fence_action_seq_passed(struct vmw_fence_action *action)
 		*eaction->tv_usec = tv.tv_usec;
 	}
 
-	list_add_tail(&eaction->e.link, &file_priv->event_list);
+	list_del_init(&eaction->fpriv_head);
+	list_add_tail(&eaction->event->link, &file_priv->event_list);
+	eaction->event = NULL;
 	wake_up_all(&file_priv->event_wait);
 	spin_unlock_irqrestore(&dev->event_lock, irq_flags);
 }
@@ -876,9 +876,15 @@ static void vmw_event_fence_action_cleanup(struct vmw_fence_action *action)
 {
 	struct vmw_event_fence_action *eaction =
 		container_of(action, struct vmw_event_fence_action, action);
+	struct vmw_fence_manager *fman = eaction->fence->fman;
+	unsigned long irq_flags;
+
+	spin_lock_irqsave(&fman->lock, irq_flags);
+	list_del(&eaction->fpriv_head);
+	spin_unlock_irqrestore(&fman->lock, irq_flags);
 
 	vmw_fence_obj_unreference(&eaction->fence);
-	kref_put(&eaction->kref, vmw_event_fence_action_destroy);
+	kfree(eaction);
 }
 
 
@@ -946,39 +952,23 @@ void vmw_fence_obj_add_action(struct vmw_fence_obj *fence,
  * an error code, the caller needs to free that object.
  */
 
-int vmw_event_fence_action_create(struct drm_file *file_priv,
-				  struct vmw_fence_obj *fence,
-				  struct drm_event *event,
-				  uint32_t *tv_sec,
-				  uint32_t *tv_usec,
-				  bool interruptible)
+int vmw_event_fence_action_queue(struct drm_file *file_priv,
+				 struct vmw_fence_obj *fence,
+				 struct drm_pending_event *event,
+				 uint32_t *tv_sec,
+				 uint32_t *tv_usec,
+				 bool interruptible)
 {
 	struct vmw_event_fence_action *eaction;
-	struct ttm_mem_global *mem_glob =
-		vmw_mem_glob(fence->fman->dev_priv);
 	struct vmw_fence_manager *fman = fence->fman;
-	uint32_t size = fman->event_fence_action_size +
-		ttm_round_pot(event->length);
-	int ret;
-
-	/*
-	 * Account for internal structure size as well as the
-	 * event size itself.
-	 */
-
-	ret = ttm_mem_global_alloc(mem_glob, size, false, interruptible);
-	if (unlikely(ret != 0))
-		return ret;
+	struct vmw_fpriv *vmw_fp = vmw_fpriv(file_priv);
+	unsigned long irq_flags;
 
 	eaction = kzalloc(sizeof(*eaction), GFP_KERNEL);
-	if (unlikely(eaction == NULL)) {
-		ttm_mem_global_free(mem_glob, size);
+	if (unlikely(eaction == NULL))
 		return -ENOMEM;
-	}
 
-	eaction->e.event = event;
-	eaction->e.file_priv = file_priv;
-	eaction->e.destroy = vmw_event_fence_action_delivered;
+	eaction->event = event;
 
 	eaction->action.seq_passed = vmw_event_fence_action_seq_passed;
 	eaction->action.cleanup = vmw_event_fence_action_cleanup;
@@ -986,16 +976,89 @@ int vmw_event_fence_action_create(struct drm_file *file_priv,
 
 	eaction->fence = vmw_fence_obj_reference(fence);
 	eaction->dev = fman->dev_priv->dev;
-	eaction->size = size;
 	eaction->tv_sec = tv_sec;
 	eaction->tv_usec = tv_usec;
 
-	kref_init(&eaction->kref);
+	spin_lock_irqsave(&fman->lock, irq_flags);
+	list_add_tail(&eaction->fpriv_head, &vmw_fp->fence_events);
+	spin_unlock_irqrestore(&fman->lock, irq_flags);
+
 	vmw_fence_obj_add_action(fence, &eaction->action);
 
 	return 0;
 }
 
+struct vmw_event_fence_pending {
+	struct drm_pending_event base;
+	struct drm_vmw_event_fence event;
+};
+
+int vmw_event_fence_action_create(struct drm_file *file_priv,
+				  struct vmw_fence_obj *fence,
+				  uint32_t flags,
+				  uint64_t user_data,
+				  bool interruptible)
+{
+	struct vmw_event_fence_pending *event;
+	struct drm_device *dev = fence->fman->dev_priv->dev;
+	unsigned long irq_flags;
+	int ret;
+
+	spin_lock_irqsave(&dev->event_lock, irq_flags);
+
+	ret = (file_priv->event_space < sizeof(event->event)) ? -EBUSY : 0;
+	if (likely(ret == 0))
+		file_priv->event_space -= sizeof(event->event);
+
+	spin_unlock_irqrestore(&dev->event_lock, irq_flags);
+
+	if (unlikely(ret != 0)) {
+		DRM_ERROR("Failed to allocate event space for this file.\n");
+		goto out_no_space;
+	}
+
+
+	event = kzalloc(sizeof(event->event), GFP_KERNEL);
+	if (unlikely(event == NULL)) {
+		DRM_ERROR("Failed to allocate an event.\n");
+		ret = -ENOMEM;
+		goto out_no_event;
+	}
+
+	event->event.base.type = DRM_VMW_EVENT_FENCE_SIGNALED;
+	event->event.base.length = sizeof(*event);
+	event->event.user_data = user_data;
+
+	event->base.event = &event->event.base;
+	event->base.file_priv = file_priv;
+	event->base.destroy = (void (*) (struct drm_pending_event *)) kfree;
+
+
+	if (flags & DRM_VMW_FE_FLAG_REQ_TIME)
+		ret = vmw_event_fence_action_queue(file_priv, fence,
+						   &event->base,
+						   &event->event.tv_sec,
+						   &event->event.tv_usec,
+						   interruptible);
+	else
+		ret = vmw_event_fence_action_queue(file_priv, fence,
+						   &event->base,
+						   NULL,
+						   NULL,
+						   interruptible);
+	if (ret != 0)
+		goto out_no_queue;
+
+out_no_queue:
+	event->base.destroy(&event->base);
+out_no_event:
+	spin_lock_irqsave(&dev->event_lock, irq_flags);
+	file_priv->event_space += sizeof(*event);
+	spin_unlock_irqrestore(&dev->event_lock, irq_flags);
+out_no_space:
+	return ret;
+}
+
 int vmw_fence_event_ioctl(struct drm_device *dev, void *data,
 			  struct drm_file *file_priv)
 {
@@ -1008,8 +1071,6 @@ int vmw_fence_event_ioctl(struct drm_device *dev, void *data,
 		(struct drm_vmw_fence_rep __user *)(unsigned long)
 		arg->fence_rep;
 	uint32_t handle;
-	unsigned long irq_flags;
-	struct drm_vmw_event_fence *event;
 	int ret;
 
 	/*
@@ -1062,59 +1123,28 @@ int vmw_fence_event_ioctl(struct drm_device *dev, void *data,
 
 	BUG_ON(fence == NULL);
 
-	spin_lock_irqsave(&dev->event_lock, irq_flags);
-
-	ret = (file_priv->event_space < sizeof(*event)) ? -EBUSY : 0;
-	if (likely(ret == 0))
-		file_priv->event_space -= sizeof(*event);
-
-	spin_unlock_irqrestore(&dev->event_lock, irq_flags);
-
-	if (unlikely(ret != 0)) {
-		DRM_ERROR("Failed to allocate event space for this file.\n");
-		goto out_no_event_space;
-	}
-
-	event = kzalloc(sizeof(*event), GFP_KERNEL);
-	if (unlikely(event == NULL)) {
-		DRM_ERROR("Failed to allocate an event.\n");
-		goto out_no_event;
-	}
-
-	event->base.type = DRM_VMW_EVENT_FENCE_SIGNALED;
-	event->base.length = sizeof(*event);
-	event->user_data = arg->user_data;
-
 	if (arg->flags & DRM_VMW_FE_FLAG_REQ_TIME)
 		ret = vmw_event_fence_action_create(file_priv, fence,
-						    &event->base,
-						    &event->tv_sec,
-						    &event->tv_usec,
+						    arg->flags,
+						    arg->user_data,
 						    true);
 	else
 		ret = vmw_event_fence_action_create(file_priv, fence,
-						    &event->base,
-						    NULL,
-						    NULL,
+						    arg->flags,
+						    arg->user_data,
 						    true);
 
 	if (unlikely(ret != 0)) {
 		if (ret != -ERESTARTSYS)
 			DRM_ERROR("Failed to attach event to fence.\n");
-		goto out_no_attach;
+		goto out_no_create;
 	}
 
 	vmw_execbuf_copy_fence_user(dev_priv, vmw_fp, 0, user_fence_rep, fence,
 				    handle);
 	vmw_fence_obj_unreference(&fence);
 	return 0;
-out_no_attach:
-	kfree(event);
-out_no_event:
-	spin_lock_irqsave(&dev->event_lock, irq_flags);
-	file_priv->event_space += sizeof(*event);
-	spin_unlock_irqrestore(&dev->event_lock, irq_flags);
-out_no_event_space:
+out_no_create:
 	if (user_fence_rep != NULL)
 		ttm_ref_object_base_unref(vmw_fpriv(file_priv)->tfile,
 					  handle, TTM_REF_USAGE);
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_fence.h b/drivers/gpu/drm/vmwgfx/vmwgfx_fence.h
index 0854a2096b55..faf2e7873860 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_fence.h
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_fence.h
@@ -109,5 +109,12 @@ extern int vmw_fence_obj_unref_ioctl(struct drm_device *dev, void *data,
 				     struct drm_file *file_priv);
 extern int vmw_fence_event_ioctl(struct drm_device *dev, void *data,
 				 struct drm_file *file_priv);
-
+extern void vmw_event_fence_fpriv_gone(struct vmw_fence_manager *fman,
+				       struct list_head *event_list);
+extern int vmw_event_fence_action_queue(struct drm_file *filee_priv,
+					struct vmw_fence_obj *fence,
+					struct drm_pending_event *event,
+					uint32_t *tv_sec,
+					uint32_t *tv_usec,
+					bool interruptible);
 #endif /* _VMWGFX_FENCE_H_ */
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_gmr.c b/drivers/gpu/drm/vmwgfx/vmwgfx_gmr.c
index f4e7763a7694..51c9ba5cd2fb 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_gmr.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_gmr.c
@@ -136,10 +136,10 @@ static int vmw_gmr_build_descriptors(struct list_head *desc_pages,
 
 		if (likely(page_virtual != NULL)) {
 			desc_virtual->ppn = page_to_pfn(page);
-			kunmap_atomic(page_virtual, KM_USER0);
+			kunmap_atomic(page_virtual);
 		}
 
-		page_virtual = kmap_atomic(page, KM_USER0);
+		page_virtual = kmap_atomic(page);
 		desc_virtual = page_virtual - 1;
 		prev_pfn = ~(0UL);
 
@@ -169,7 +169,7 @@ static int vmw_gmr_build_descriptors(struct list_head *desc_pages,
 	}
 
 	if (likely(page_virtual != NULL))
-		kunmap_atomic(page_virtual, KM_USER0);
+		kunmap_atomic(page_virtual);
 
 	return 0;
 out_err:
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_kms.c b/drivers/gpu/drm/vmwgfx/vmwgfx_kms.c
index b66ef0e3cde1..2286d47e5022 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_kms.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_kms.c
@@ -422,7 +422,8 @@ static int do_surface_dirty_sou(struct vmw_private *dev_priv,
 				struct vmw_framebuffer *framebuffer,
 				unsigned flags, unsigned color,
 				struct drm_clip_rect *clips,
-				unsigned num_clips, int inc)
+				unsigned num_clips, int inc,
+				struct vmw_fence_obj **out_fence)
 {
 	struct vmw_display_unit *units[VMWGFX_NUM_DISPLAY_UNITS];
 	struct drm_clip_rect *clips_ptr;
@@ -542,12 +543,15 @@ static int do_surface_dirty_sou(struct vmw_private *dev_priv,
 		if (num == 0)
 			continue;
 
+		/* only return the last fence */
+		if (out_fence && *out_fence)
+			vmw_fence_obj_unreference(out_fence);
 
 		/* recalculate package length */
 		fifo_size = sizeof(*cmd) + sizeof(SVGASignedRect) * num;
 		cmd->header.size = cpu_to_le32(fifo_size - sizeof(cmd->header));
 		ret = vmw_execbuf_process(file_priv, dev_priv, NULL, cmd,
-					  fifo_size, 0, NULL);
+					  fifo_size, 0, NULL, out_fence);
 
 		if (unlikely(ret != 0))
 			break;
@@ -598,7 +602,7 @@ int vmw_framebuffer_surface_dirty(struct drm_framebuffer *framebuffer,
 
 	ret = do_surface_dirty_sou(dev_priv, file_priv, &vfbs->base,
 				   flags, color,
-				   clips, num_clips, inc);
+				   clips, num_clips, inc, NULL);
 
 	ttm_read_unlock(&vmaster->lock);
 	return 0;
@@ -809,7 +813,7 @@ static int do_dmabuf_define_gmrfb(struct drm_file *file_priv,
 	cmd->body.ptr.offset = 0;
 
 	ret = vmw_execbuf_process(file_priv, dev_priv, NULL, cmd,
-				  fifo_size, 0, NULL);
+				  fifo_size, 0, NULL, NULL);
 
 	kfree(cmd);
 
@@ -821,7 +825,8 @@ static int do_dmabuf_dirty_sou(struct drm_file *file_priv,
 			       struct vmw_framebuffer *framebuffer,
 			       unsigned flags, unsigned color,
 			       struct drm_clip_rect *clips,
-			       unsigned num_clips, int increment)
+			       unsigned num_clips, int increment,
+			       struct vmw_fence_obj **out_fence)
 {
 	struct vmw_display_unit *units[VMWGFX_NUM_DISPLAY_UNITS];
 	struct drm_clip_rect *clips_ptr;
@@ -894,9 +899,13 @@ static int do_dmabuf_dirty_sou(struct drm_file *file_priv,
 		if (hit_num == 0)
 			continue;
 
+		/* only return the last fence */
+		if (out_fence && *out_fence)
+			vmw_fence_obj_unreference(out_fence);
+
 		fifo_size = sizeof(*blits) * hit_num;
 		ret = vmw_execbuf_process(file_priv, dev_priv, NULL, blits,
-					  fifo_size, 0, NULL);
+					  fifo_size, 0, NULL, out_fence);
 
 		if (unlikely(ret != 0))
 			break;
@@ -942,7 +951,7 @@ int vmw_framebuffer_dmabuf_dirty(struct drm_framebuffer *framebuffer,
 	} else {
 		ret = do_dmabuf_dirty_sou(file_priv, dev_priv, &vfbd->base,
 					  flags, color,
-					  clips, num_clips, increment);
+					  clips, num_clips, increment, NULL);
 	}
 
 	ttm_read_unlock(&vmaster->lock);
@@ -1296,7 +1305,7 @@ int vmw_kms_present(struct vmw_private *dev_priv,
 		fifo_size = sizeof(*cmd) + sizeof(SVGASignedRect) * num;
 		cmd->header.size = cpu_to_le32(fifo_size - sizeof(cmd->header));
 		ret = vmw_execbuf_process(file_priv, dev_priv, NULL, cmd,
-					  fifo_size, 0, NULL);
+					  fifo_size, 0, NULL, NULL);
 
 		if (unlikely(ret != 0))
 			break;
@@ -1409,7 +1418,7 @@ int vmw_kms_readback(struct vmw_private *dev_priv,
 	fifo_size = sizeof(*cmd) + sizeof(*blits) * blits_pos;
 
 	ret = vmw_execbuf_process(file_priv, dev_priv, NULL, cmd, fifo_size,
-				  0, user_fence_rep);
+				  0, user_fence_rep, NULL);
 
 	kfree(cmd);
 
@@ -1672,6 +1681,70 @@ int vmw_du_update_layout(struct vmw_private *dev_priv, unsigned num,
 	return 0;
 }
 
+int vmw_du_page_flip(struct drm_crtc *crtc,
+		     struct drm_framebuffer *fb,
+		     struct drm_pending_vblank_event *event)
+{
+	struct vmw_private *dev_priv = vmw_priv(crtc->dev);
+	struct drm_framebuffer *old_fb = crtc->fb;
+	struct vmw_framebuffer *vfb = vmw_framebuffer_to_vfb(fb);
+	struct drm_file *file_priv = event->base.file_priv;
+	struct vmw_fence_obj *fence = NULL;
+	struct drm_clip_rect clips;
+	int ret;
+
+	/* require ScreenObject support for page flipping */
+	if (!dev_priv->sou_priv)
+		return -ENOSYS;
+
+	if (!vmw_kms_screen_object_flippable(dev_priv, crtc))
+		return -EINVAL;
+
+	crtc->fb = fb;
+
+	/* do a full screen dirty update */
+	clips.x1 = clips.y1 = 0;
+	clips.x2 = fb->width;
+	clips.y2 = fb->height;
+
+	if (vfb->dmabuf)
+		ret = do_dmabuf_dirty_sou(file_priv, dev_priv, vfb,
+					  0, 0, &clips, 1, 1, &fence);
+	else
+		ret = do_surface_dirty_sou(dev_priv, file_priv, vfb,
+					   0, 0, &clips, 1, 1, &fence);
+
+
+	if (ret != 0)
+		goto out_no_fence;
+	if (!fence) {
+		ret = -EINVAL;
+		goto out_no_fence;
+	}
+
+	ret = vmw_event_fence_action_queue(file_priv, fence,
+					   &event->base,
+					   &event->event.tv_sec,
+					   &event->event.tv_usec,
+					   true);
+
+	/*
+	 * No need to hold on to this now. The only cleanup
+	 * we need to do if we fail is unref the fence.
+	 */
+	vmw_fence_obj_unreference(&fence);
+
+	if (vmw_crtc_to_du(crtc)->is_implicit)
+		vmw_kms_screen_object_update_implicit_fb(dev_priv, crtc);
+
+	return ret;
+
+out_no_fence:
+	crtc->fb = old_fb;
+	return ret;
+}
+
+
 void vmw_du_crtc_save(struct drm_crtc *crtc)
 {
 }
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_kms.h b/drivers/gpu/drm/vmwgfx/vmwgfx_kms.h
index a4f7f034996a..8184bc5b1730 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_kms.h
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_kms.h
@@ -121,6 +121,9 @@ struct vmw_display_unit {
  * Shared display unit functions - vmwgfx_kms.c
  */
 void vmw_display_unit_cleanup(struct vmw_display_unit *du);
+int vmw_du_page_flip(struct drm_crtc *crtc,
+		     struct drm_framebuffer *fb,
+		     struct drm_pending_vblank_event *event);
 void vmw_du_crtc_save(struct drm_crtc *crtc);
 void vmw_du_crtc_restore(struct drm_crtc *crtc);
 void vmw_du_crtc_gamma_set(struct drm_crtc *crtc,
@@ -154,5 +157,10 @@ int vmw_kms_init_screen_object_display(struct vmw_private *dev_priv);
 int vmw_kms_close_screen_object_display(struct vmw_private *dev_priv);
 int vmw_kms_sou_update_layout(struct vmw_private *dev_priv, unsigned num,
 			      struct drm_vmw_rect *rects);
+bool vmw_kms_screen_object_flippable(struct vmw_private *dev_priv,
+				     struct drm_crtc *crtc);
+void vmw_kms_screen_object_update_implicit_fb(struct vmw_private *dev_priv,
+					      struct drm_crtc *crtc);
+
 
 #endif
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_ldu.c b/drivers/gpu/drm/vmwgfx/vmwgfx_ldu.c
index f77b184be807..070fb239c5af 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_ldu.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_ldu.c
@@ -354,8 +354,8 @@ static int vmw_ldu_init(struct vmw_private *dev_priv, unsigned unit)
 	INIT_LIST_HEAD(&ldu->active);
 
 	ldu->base.pref_active = (unit == 0);
-	ldu->base.pref_width = 800;
-	ldu->base.pref_height = 600;
+	ldu->base.pref_width = dev_priv->initial_width;
+	ldu->base.pref_height = dev_priv->initial_height;
 	ldu->base.pref_mode = NULL;
 	ldu->base.is_implicit = true;
 
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_scrn.c b/drivers/gpu/drm/vmwgfx/vmwgfx_scrn.c
index 4defdcf1c72e..6deaf2f8bab1 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_scrn.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_scrn.c
@@ -394,6 +394,7 @@ static struct drm_crtc_funcs vmw_screen_object_crtc_funcs = {
 	.gamma_set = vmw_du_crtc_gamma_set,
 	.destroy = vmw_sou_crtc_destroy,
 	.set_config = vmw_sou_crtc_set_config,
+	.page_flip = vmw_du_page_flip,
 };
 
 /*
@@ -448,8 +449,8 @@ static int vmw_sou_init(struct vmw_private *dev_priv, unsigned unit)
 	sou->active_implicit = false;
 
 	sou->base.pref_active = (unit == 0);
-	sou->base.pref_width = 800;
-	sou->base.pref_height = 600;
+	sou->base.pref_width = dev_priv->initial_width;
+	sou->base.pref_height = dev_priv->initial_height;
 	sou->base.pref_mode = NULL;
 	sou->base.is_implicit = true;
 
@@ -535,3 +536,36 @@ int vmw_kms_close_screen_object_display(struct vmw_private *dev_priv)
 
 	return 0;
 }
+
+/**
+ * Returns if this unit can be page flipped.
+ * Must be called with the mode_config mutex held.
+ */
+bool vmw_kms_screen_object_flippable(struct vmw_private *dev_priv,
+				     struct drm_crtc *crtc)
+{
+	struct vmw_screen_object_unit *sou = vmw_crtc_to_sou(crtc);
+
+	if (!sou->base.is_implicit)
+		return true;
+
+	if (dev_priv->sou_priv->num_implicit != 1)
+		return false;
+
+	return true;
+}
+
+/**
+ * Update the implicit fb to the current fb of this crtc.
+ * Must be called with the mode_config mutex held.
+ */
+void vmw_kms_screen_object_update_implicit_fb(struct vmw_private *dev_priv,
+					      struct drm_crtc *crtc)
+{
+	struct vmw_screen_object_unit *sou = vmw_crtc_to_sou(crtc);
+
+	BUG_ON(!sou->base.is_implicit);
+
+	dev_priv->sou_priv->implicit_fb =
+		vmw_framebuffer_to_vfb(sou->base.crtc.fb);
+}