diff --git a/drivers/gpu/drm/tegra/dc.c b/drivers/gpu/drm/tegra/dc.c
index ef40381f3909e3ac20aed29eeb1266e6a8e3d5bd..6553fd238685e459deb0676114bfd9cac04f7e50 100644
--- a/drivers/gpu/drm/tegra/dc.c
+++ b/drivers/gpu/drm/tegra/dc.c
@@ -18,6 +18,8 @@
 struct tegra_dc_soc_info {
 	bool supports_interlacing;
 	bool supports_cursor;
+	bool supports_block_linear;
+	unsigned int pitch_align;
 };
 
 struct tegra_plane {
@@ -212,15 +214,44 @@ static int tegra_dc_setup_window(struct tegra_dc *dc, unsigned int index,
 	tegra_dc_writel(dc, h_offset, DC_WINBUF_ADDR_H_OFFSET);
 	tegra_dc_writel(dc, v_offset, DC_WINBUF_ADDR_V_OFFSET);
 
-	if (window->tiled) {
-		value = DC_WIN_BUFFER_ADDR_MODE_TILE_UV |
-			DC_WIN_BUFFER_ADDR_MODE_TILE;
+	if (dc->soc->supports_block_linear) {
+		unsigned long height = window->tiling.value;
+
+		switch (window->tiling.mode) {
+		case TEGRA_BO_TILING_MODE_PITCH:
+			value = DC_WINBUF_SURFACE_KIND_PITCH;
+			break;
+
+		case TEGRA_BO_TILING_MODE_TILED:
+			value = DC_WINBUF_SURFACE_KIND_TILED;
+			break;
+
+		case TEGRA_BO_TILING_MODE_BLOCK:
+			value = DC_WINBUF_SURFACE_KIND_BLOCK_HEIGHT(height) |
+				DC_WINBUF_SURFACE_KIND_BLOCK;
+			break;
+		}
+
+		tegra_dc_writel(dc, value, DC_WINBUF_SURFACE_KIND);
 	} else {
-		value = DC_WIN_BUFFER_ADDR_MODE_LINEAR_UV |
-			DC_WIN_BUFFER_ADDR_MODE_LINEAR;
-	}
+		switch (window->tiling.mode) {
+		case TEGRA_BO_TILING_MODE_PITCH:
+			value = DC_WIN_BUFFER_ADDR_MODE_LINEAR_UV |
+				DC_WIN_BUFFER_ADDR_MODE_LINEAR;
+			break;
 
-	tegra_dc_writel(dc, value, DC_WIN_BUFFER_ADDR_MODE);
+		case TEGRA_BO_TILING_MODE_TILED:
+			value = DC_WIN_BUFFER_ADDR_MODE_TILE_UV |
+				DC_WIN_BUFFER_ADDR_MODE_TILE;
+			break;
+
+		case TEGRA_BO_TILING_MODE_BLOCK:
+			DRM_ERROR("hardware doesn't support block linear mode\n");
+			return -EINVAL;
+		}
+
+		tegra_dc_writel(dc, value, DC_WIN_BUFFER_ADDR_MODE);
+	}
 
 	value = WIN_ENABLE;
 
@@ -288,6 +319,7 @@ static int tegra_plane_update(struct drm_plane *plane, struct drm_crtc *crtc,
 	struct tegra_dc *dc = to_tegra_dc(crtc);
 	struct tegra_dc_window window;
 	unsigned int i;
+	int err;
 
 	memset(&window, 0, sizeof(window));
 	window.src.x = src_x >> 16;
@@ -301,7 +333,10 @@ static int tegra_plane_update(struct drm_plane *plane, struct drm_crtc *crtc,
 	window.format = tegra_dc_format(fb->pixel_format, &window.swap);
 	window.bits_per_pixel = fb->bits_per_pixel;
 	window.bottom_up = tegra_fb_is_bottom_up(fb);
-	window.tiled = tegra_fb_is_tiled(fb);
+
+	err = tegra_fb_get_tiling(fb, &window.tiling);
+	if (err < 0)
+		return err;
 
 	for (i = 0; i < drm_format_num_planes(fb->pixel_format); i++) {
 		struct tegra_bo *bo = tegra_fb_get_plane(fb, i);
@@ -402,8 +437,14 @@ static int tegra_dc_set_base(struct tegra_dc *dc, int x, int y,
 {
 	struct tegra_bo *bo = tegra_fb_get_plane(fb, 0);
 	unsigned int h_offset = 0, v_offset = 0;
+	struct tegra_bo_tiling tiling;
 	unsigned int format, swap;
 	unsigned long value;
+	int err;
+
+	err = tegra_fb_get_tiling(fb, &tiling);
+	if (err < 0)
+		return err;
 
 	tegra_dc_writel(dc, WINDOW_A_SELECT, DC_CMD_DISPLAY_WINDOW_HEADER);
 
@@ -417,15 +458,44 @@ static int tegra_dc_set_base(struct tegra_dc *dc, int x, int y,
 	tegra_dc_writel(dc, format, DC_WIN_COLOR_DEPTH);
 	tegra_dc_writel(dc, swap, DC_WIN_BYTE_SWAP);
 
-	if (tegra_fb_is_tiled(fb)) {
-		value = DC_WIN_BUFFER_ADDR_MODE_TILE_UV |
-			DC_WIN_BUFFER_ADDR_MODE_TILE;
+	if (dc->soc->supports_block_linear) {
+		unsigned long height = tiling.value;
+
+		switch (tiling.mode) {
+		case TEGRA_BO_TILING_MODE_PITCH:
+			value = DC_WINBUF_SURFACE_KIND_PITCH;
+			break;
+
+		case TEGRA_BO_TILING_MODE_TILED:
+			value = DC_WINBUF_SURFACE_KIND_TILED;
+			break;
+
+		case TEGRA_BO_TILING_MODE_BLOCK:
+			value = DC_WINBUF_SURFACE_KIND_BLOCK_HEIGHT(height) |
+				DC_WINBUF_SURFACE_KIND_BLOCK;
+			break;
+		}
+
+		tegra_dc_writel(dc, value, DC_WINBUF_SURFACE_KIND);
 	} else {
-		value = DC_WIN_BUFFER_ADDR_MODE_LINEAR_UV |
-			DC_WIN_BUFFER_ADDR_MODE_LINEAR;
-	}
+		switch (tiling.mode) {
+		case TEGRA_BO_TILING_MODE_PITCH:
+			value = DC_WIN_BUFFER_ADDR_MODE_LINEAR_UV |
+				DC_WIN_BUFFER_ADDR_MODE_LINEAR;
+			break;
+
+		case TEGRA_BO_TILING_MODE_TILED:
+			value = DC_WIN_BUFFER_ADDR_MODE_TILE_UV |
+				DC_WIN_BUFFER_ADDR_MODE_TILE;
+			break;
+
+		case TEGRA_BO_TILING_MODE_BLOCK:
+			DRM_ERROR("hardware doesn't support block linear mode\n");
+			return -EINVAL;
+		}
 
-	tegra_dc_writel(dc, value, DC_WIN_BUFFER_ADDR_MODE);
+		tegra_dc_writel(dc, value, DC_WIN_BUFFER_ADDR_MODE);
+	}
 
 	/* make sure bottom-up buffers are properly displayed */
 	if (tegra_fb_is_bottom_up(fb)) {
@@ -1214,12 +1284,20 @@ static int tegra_dc_init(struct host1x_client *client)
 {
 	struct drm_device *drm = dev_get_drvdata(client->parent);
 	struct tegra_dc *dc = host1x_client_to_dc(client);
+	struct tegra_drm *tegra = drm->dev_private;
 	int err;
 
 	drm_crtc_init(drm, &dc->base, &tegra_crtc_funcs);
 	drm_mode_crtc_set_gamma_size(&dc->base, 256);
 	drm_crtc_helper_add(&dc->base, &tegra_crtc_helper_funcs);
 
+	/*
+	 * Keep track of the minimum pitch alignment across all display
+	 * controllers.
+	 */
+	if (dc->soc->pitch_align > tegra->pitch_align)
+		tegra->pitch_align = dc->soc->pitch_align;
+
 	err = tegra_dc_rgb_init(drm, dc);
 	if (err < 0 && err != -ENODEV) {
 		dev_err(dc->dev, "failed to initialize RGB output: %d\n", err);
@@ -1277,16 +1355,29 @@ static const struct host1x_client_ops dc_client_ops = {
 static const struct tegra_dc_soc_info tegra20_dc_soc_info = {
 	.supports_interlacing = false,
 	.supports_cursor = false,
+	.supports_block_linear = false,
+	.pitch_align = 8,
 };
 
 static const struct tegra_dc_soc_info tegra30_dc_soc_info = {
 	.supports_interlacing = false,
 	.supports_cursor = false,
+	.supports_block_linear = false,
+	.pitch_align = 8,
+};
+
+static const struct tegra_dc_soc_info tegra114_dc_soc_info = {
+	.supports_interlacing = false,
+	.supports_cursor = false,
+	.supports_block_linear = false,
+	.pitch_align = 64,
 };
 
 static const struct tegra_dc_soc_info tegra124_dc_soc_info = {
 	.supports_interlacing = true,
 	.supports_cursor = true,
+	.supports_block_linear = true,
+	.pitch_align = 64,
 };
 
 static const struct of_device_id tegra_dc_of_match[] = {
@@ -1303,6 +1394,7 @@ static const struct of_device_id tegra_dc_of_match[] = {
 		/* sentinel */
 	}
 };
+MODULE_DEVICE_TABLE(of, tegra_dc_of_match);
 
 static int tegra_dc_parse_dt(struct tegra_dc *dc)
 {
@@ -1430,6 +1522,7 @@ static int tegra_dc_remove(struct platform_device *pdev)
 		return err;
 	}
 
+	reset_control_assert(dc->rst);
 	clk_disable_unprepare(dc->clk);
 
 	return 0;
diff --git a/drivers/gpu/drm/tegra/dc.h b/drivers/gpu/drm/tegra/dc.h
index 78c5feff95d2910123c209cc212eecd1ee1d72bf..705c93b00794feb38f56a501b7546f9e6f07b1f9 100644
--- a/drivers/gpu/drm/tegra/dc.h
+++ b/drivers/gpu/drm/tegra/dc.h
@@ -428,6 +428,11 @@
 #define DC_WINBUF_ADDR_V_OFFSET_NS		0x809
 
 #define DC_WINBUF_UFLOW_STATUS			0x80a
+#define DC_WINBUF_SURFACE_KIND			0x80b
+#define DC_WINBUF_SURFACE_KIND_PITCH	(0 << 0)
+#define DC_WINBUF_SURFACE_KIND_TILED	(1 << 0)
+#define DC_WINBUF_SURFACE_KIND_BLOCK	(2 << 0)
+#define DC_WINBUF_SURFACE_KIND_BLOCK_HEIGHT(x) (((x) & 0x7) << 4)
 
 #define DC_WINBUF_AD_UFLOW_STATUS		0xbca
 #define DC_WINBUF_BD_UFLOW_STATUS		0xdca
diff --git a/drivers/gpu/drm/tegra/dpaux.c b/drivers/gpu/drm/tegra/dpaux.c
index 3f132e356e9cd393394bdce5a3594765be08d596..708f783ead47682f8ef76e17ebc435a1850f9eb6 100644
--- a/drivers/gpu/drm/tegra/dpaux.c
+++ b/drivers/gpu/drm/tegra/dpaux.c
@@ -382,6 +382,7 @@ static const struct of_device_id tegra_dpaux_of_match[] = {
 	{ .compatible = "nvidia,tegra124-dpaux", },
 	{ },
 };
+MODULE_DEVICE_TABLE(of, tegra_dpaux_of_match);
 
 struct platform_driver tegra_dpaux_driver = {
 	.driver = {
diff --git a/drivers/gpu/drm/tegra/drm.c b/drivers/gpu/drm/tegra/drm.c
index fd736efd14bd963cd38571ae7de02fb54c785ce8..59736bb810cd2d4787dab77d1fea29dadcf30f9f 100644
--- a/drivers/gpu/drm/tegra/drm.c
+++ b/drivers/gpu/drm/tegra/drm.c
@@ -132,6 +132,45 @@ host1x_bo_lookup(struct drm_device *drm, struct drm_file *file, u32 handle)
 	return &bo->base;
 }
 
+static int host1x_reloc_copy_from_user(struct host1x_reloc *dest,
+				       struct drm_tegra_reloc __user *src,
+				       struct drm_device *drm,
+				       struct drm_file *file)
+{
+	u32 cmdbuf, target;
+	int err;
+
+	err = get_user(cmdbuf, &src->cmdbuf.handle);
+	if (err < 0)
+		return err;
+
+	err = get_user(dest->cmdbuf.offset, &src->cmdbuf.offset);
+	if (err < 0)
+		return err;
+
+	err = get_user(target, &src->target.handle);
+	if (err < 0)
+		return err;
+
+	err = get_user(dest->target.offset, &src->cmdbuf.offset);
+	if (err < 0)
+		return err;
+
+	err = get_user(dest->shift, &src->shift);
+	if (err < 0)
+		return err;
+
+	dest->cmdbuf.bo = host1x_bo_lookup(drm, file, cmdbuf);
+	if (!dest->cmdbuf.bo)
+		return -ENOENT;
+
+	dest->target.bo = host1x_bo_lookup(drm, file, target);
+	if (!dest->target.bo)
+		return -ENOENT;
+
+	return 0;
+}
+
 int tegra_drm_submit(struct tegra_drm_context *context,
 		     struct drm_tegra_submit *args, struct drm_device *drm,
 		     struct drm_file *file)
@@ -184,26 +223,13 @@ int tegra_drm_submit(struct tegra_drm_context *context,
 		cmdbufs++;
 	}
 
-	if (copy_from_user(job->relocarray, relocs,
-			   sizeof(*relocs) * num_relocs)) {
-		err = -EFAULT;
-		goto fail;
-	}
-
+	/* copy and resolve relocations from submit */
 	while (num_relocs--) {
-		struct host1x_reloc *reloc = &job->relocarray[num_relocs];
-		struct host1x_bo *cmdbuf, *target;
-
-		cmdbuf = host1x_bo_lookup(drm, file, (u32)reloc->cmdbuf);
-		target = host1x_bo_lookup(drm, file, (u32)reloc->target);
-
-		reloc->cmdbuf = cmdbuf;
-		reloc->target = target;
-
-		if (!reloc->target || !reloc->cmdbuf) {
-			err = -ENOENT;
+		err = host1x_reloc_copy_from_user(&job->relocarray[num_relocs],
+						  &relocs[num_relocs], drm,
+						  file);
+		if (err < 0)
 			goto fail;
-		}
 	}
 
 	if (copy_from_user(job->waitchk, waitchks,
@@ -455,11 +481,151 @@ static int tegra_get_syncpt_base(struct drm_device *drm, void *data,
 
 	return 0;
 }
+
+static int tegra_gem_set_tiling(struct drm_device *drm, void *data,
+				struct drm_file *file)
+{
+	struct drm_tegra_gem_set_tiling *args = data;
+	enum tegra_bo_tiling_mode mode;
+	struct drm_gem_object *gem;
+	unsigned long value = 0;
+	struct tegra_bo *bo;
+
+	switch (args->mode) {
+	case DRM_TEGRA_GEM_TILING_MODE_PITCH:
+		mode = TEGRA_BO_TILING_MODE_PITCH;
+
+		if (args->value != 0)
+			return -EINVAL;
+
+		break;
+
+	case DRM_TEGRA_GEM_TILING_MODE_TILED:
+		mode = TEGRA_BO_TILING_MODE_TILED;
+
+		if (args->value != 0)
+			return -EINVAL;
+
+		break;
+
+	case DRM_TEGRA_GEM_TILING_MODE_BLOCK:
+		mode = TEGRA_BO_TILING_MODE_BLOCK;
+
+		if (args->value > 5)
+			return -EINVAL;
+
+		value = args->value;
+		break;
+
+	default:
+		return -EINVAL;
+	}
+
+	gem = drm_gem_object_lookup(drm, file, args->handle);
+	if (!gem)
+		return -ENOENT;
+
+	bo = to_tegra_bo(gem);
+
+	bo->tiling.mode = mode;
+	bo->tiling.value = value;
+
+	drm_gem_object_unreference(gem);
+
+	return 0;
+}
+
+static int tegra_gem_get_tiling(struct drm_device *drm, void *data,
+				struct drm_file *file)
+{
+	struct drm_tegra_gem_get_tiling *args = data;
+	struct drm_gem_object *gem;
+	struct tegra_bo *bo;
+	int err = 0;
+
+	gem = drm_gem_object_lookup(drm, file, args->handle);
+	if (!gem)
+		return -ENOENT;
+
+	bo = to_tegra_bo(gem);
+
+	switch (bo->tiling.mode) {
+	case TEGRA_BO_TILING_MODE_PITCH:
+		args->mode = DRM_TEGRA_GEM_TILING_MODE_PITCH;
+		args->value = 0;
+		break;
+
+	case TEGRA_BO_TILING_MODE_TILED:
+		args->mode = DRM_TEGRA_GEM_TILING_MODE_TILED;
+		args->value = 0;
+		break;
+
+	case TEGRA_BO_TILING_MODE_BLOCK:
+		args->mode = DRM_TEGRA_GEM_TILING_MODE_BLOCK;
+		args->value = bo->tiling.value;
+		break;
+
+	default:
+		err = -EINVAL;
+		break;
+	}
+
+	drm_gem_object_unreference(gem);
+
+	return err;
+}
+
+static int tegra_gem_set_flags(struct drm_device *drm, void *data,
+			       struct drm_file *file)
+{
+	struct drm_tegra_gem_set_flags *args = data;
+	struct drm_gem_object *gem;
+	struct tegra_bo *bo;
+
+	if (args->flags & ~DRM_TEGRA_GEM_FLAGS)
+		return -EINVAL;
+
+	gem = drm_gem_object_lookup(drm, file, args->handle);
+	if (!gem)
+		return -ENOENT;
+
+	bo = to_tegra_bo(gem);
+	bo->flags = 0;
+
+	if (args->flags & DRM_TEGRA_GEM_BOTTOM_UP)
+		bo->flags |= TEGRA_BO_BOTTOM_UP;
+
+	drm_gem_object_unreference(gem);
+
+	return 0;
+}
+
+static int tegra_gem_get_flags(struct drm_device *drm, void *data,
+			       struct drm_file *file)
+{
+	struct drm_tegra_gem_get_flags *args = data;
+	struct drm_gem_object *gem;
+	struct tegra_bo *bo;
+
+	gem = drm_gem_object_lookup(drm, file, args->handle);
+	if (!gem)
+		return -ENOENT;
+
+	bo = to_tegra_bo(gem);
+	args->flags = 0;
+
+	if (bo->flags & TEGRA_BO_BOTTOM_UP)
+		args->flags |= DRM_TEGRA_GEM_BOTTOM_UP;
+
+	drm_gem_object_unreference(gem);
+
+	return 0;
+}
 #endif
 
 static const struct drm_ioctl_desc tegra_drm_ioctls[] = {
 #ifdef CONFIG_DRM_TEGRA_STAGING
-	DRM_IOCTL_DEF_DRV(TEGRA_GEM_CREATE, tegra_gem_create, DRM_UNLOCKED | DRM_AUTH),
+	DRM_IOCTL_DEF_DRV(TEGRA_GEM_CREATE, tegra_gem_create, DRM_UNLOCKED),
 	DRM_IOCTL_DEF_DRV(TEGRA_GEM_MMAP, tegra_gem_mmap, DRM_UNLOCKED),
 	DRM_IOCTL_DEF_DRV(TEGRA_SYNCPT_READ, tegra_syncpt_read, DRM_UNLOCKED),
 	DRM_IOCTL_DEF_DRV(TEGRA_SYNCPT_INCR, tegra_syncpt_incr, DRM_UNLOCKED),
@@ -469,6 +635,10 @@ static const struct drm_ioctl_desc tegra_drm_ioctls[] = {
 	DRM_IOCTL_DEF_DRV(TEGRA_GET_SYNCPT, tegra_get_syncpt, DRM_UNLOCKED),
 	DRM_IOCTL_DEF_DRV(TEGRA_SUBMIT, tegra_submit, DRM_UNLOCKED),
 	DRM_IOCTL_DEF_DRV(TEGRA_GET_SYNCPT_BASE, tegra_get_syncpt_base, DRM_UNLOCKED),
+	DRM_IOCTL_DEF_DRV(TEGRA_GEM_SET_TILING, tegra_gem_set_tiling, DRM_UNLOCKED),
+	DRM_IOCTL_DEF_DRV(TEGRA_GEM_GET_TILING, tegra_gem_get_tiling, DRM_UNLOCKED),
+	DRM_IOCTL_DEF_DRV(TEGRA_GEM_SET_FLAGS, tegra_gem_set_flags, DRM_UNLOCKED),
+	DRM_IOCTL_DEF_DRV(TEGRA_GEM_GET_FLAGS, tegra_gem_get_flags, DRM_UNLOCKED),
 #endif
 };
 
diff --git a/drivers/gpu/drm/tegra/drm.h b/drivers/gpu/drm/tegra/drm.h
index 0d30689dff016f786de35e910d930961cff534bc..e89c70fa82d554f3da8fced5ae99b840e56c985f 100644
--- a/drivers/gpu/drm/tegra/drm.h
+++ b/drivers/gpu/drm/tegra/drm.h
@@ -19,6 +19,8 @@
 #include <drm/drm_fb_helper.h>
 #include <drm/drm_fixed.h>
 
+#include "gem.h"
+
 struct reset_control;
 
 struct tegra_fb {
@@ -43,6 +45,8 @@ struct tegra_drm {
 #ifdef CONFIG_DRM_TEGRA_FBDEV
 	struct tegra_fbdev *fbdev;
 #endif
+
+	unsigned int pitch_align;
 };
 
 struct tegra_drm_client;
@@ -160,7 +164,8 @@ struct tegra_dc_window {
 	unsigned int stride[2];
 	unsigned long base[3];
 	bool bottom_up;
-	bool tiled;
+
+	struct tegra_bo_tiling tiling;
 };
 
 /* from dc.c */
@@ -279,7 +284,8 @@ int tegra_dpaux_train(struct tegra_dpaux *dpaux, struct drm_dp_link *link,
 struct tegra_bo *tegra_fb_get_plane(struct drm_framebuffer *framebuffer,
 				    unsigned int index);
 bool tegra_fb_is_bottom_up(struct drm_framebuffer *framebuffer);
-bool tegra_fb_is_tiled(struct drm_framebuffer *framebuffer);
+int tegra_fb_get_tiling(struct drm_framebuffer *framebuffer,
+			struct tegra_bo_tiling *tiling);
 int tegra_drm_fb_prepare(struct drm_device *drm);
 int tegra_drm_fb_init(struct drm_device *drm);
 void tegra_drm_fb_exit(struct drm_device *drm);
diff --git a/drivers/gpu/drm/tegra/dsi.c b/drivers/gpu/drm/tegra/dsi.c
index bd56f2affa7895b2d1e0fbd556e3e811f8e78f14..f7874458926a7f289d8df1a3e46ead9f066e1a82 100644
--- a/drivers/gpu/drm/tegra/dsi.c
+++ b/drivers/gpu/drm/tegra/dsi.c
@@ -474,7 +474,8 @@ static int tegra_output_dsi_enable(struct tegra_output *output)
 	tegra_dsi_writel(dsi, value, DSI_HOST_CONTROL);
 
 	value = tegra_dsi_readl(dsi, DSI_CONTROL);
-	value |= DSI_CONTROL_HS_CLK_CTRL;
+	if (dsi->flags & MIPI_DSI_CLOCK_NON_CONTINUOUS)
+		value |= DSI_CONTROL_HS_CLK_CTRL;
 	value &= ~DSI_CONTROL_TX_TRIG(3);
 	value &= ~DSI_CONTROL_DCS_ENABLE;
 	value |= DSI_CONTROL_VIDEO_ENABLE;
@@ -982,6 +983,7 @@ static const struct of_device_id tegra_dsi_of_match[] = {
 	{ .compatible = "nvidia,tegra114-dsi", },
 	{ },
 };
+MODULE_DEVICE_TABLE(of, tegra_dsi_of_match);
 
 struct platform_driver tegra_dsi_driver = {
 	.driver = {
diff --git a/drivers/gpu/drm/tegra/fb.c b/drivers/gpu/drm/tegra/fb.c
index fc1528e0bda17009eeee5c6974c1e15fb084c0c0..3513d12d5aa1447822a7ef57449799b1aff83df9 100644
--- a/drivers/gpu/drm/tegra/fb.c
+++ b/drivers/gpu/drm/tegra/fb.c
@@ -46,14 +46,15 @@ bool tegra_fb_is_bottom_up(struct drm_framebuffer *framebuffer)
 	return false;
 }
 
-bool tegra_fb_is_tiled(struct drm_framebuffer *framebuffer)
+int tegra_fb_get_tiling(struct drm_framebuffer *framebuffer,
+			struct tegra_bo_tiling *tiling)
 {
 	struct tegra_fb *fb = to_tegra_fb(framebuffer);
 
-	if (fb->planes[0]->flags & TEGRA_BO_TILED)
-		return true;
+	/* TODO: handle YUV formats? */
+	*tiling = fb->planes[0]->tiling;
 
-	return false;
+	return 0;
 }
 
 static void tegra_fb_destroy(struct drm_framebuffer *framebuffer)
@@ -193,6 +194,7 @@ static int tegra_fbdev_probe(struct drm_fb_helper *helper,
 			     struct drm_fb_helper_surface_size *sizes)
 {
 	struct tegra_fbdev *fbdev = to_tegra_fbdev(helper);
+	struct tegra_drm *tegra = helper->dev->dev_private;
 	struct drm_device *drm = helper->dev;
 	struct drm_mode_fb_cmd2 cmd = { 0 };
 	unsigned int bytes_per_pixel;
@@ -207,7 +209,8 @@ static int tegra_fbdev_probe(struct drm_fb_helper *helper,
 
 	cmd.width = sizes->surface_width;
 	cmd.height = sizes->surface_height;
-	cmd.pitches[0] = sizes->surface_width * bytes_per_pixel;
+	cmd.pitches[0] = round_up(sizes->surface_width * bytes_per_pixel,
+				  tegra->pitch_align);
 	cmd.pixel_format = drm_mode_legacy_fb_format(sizes->surface_bpp,
 						     sizes->surface_depth);
 
diff --git a/drivers/gpu/drm/tegra/gem.c b/drivers/gpu/drm/tegra/gem.c
index aa85b7b26f105e5d7528358d6c150b1cc33d8a4e..2545c7a468a2f3c316729cfa731bf238f6eac73f 100644
--- a/drivers/gpu/drm/tegra/gem.c
+++ b/drivers/gpu/drm/tegra/gem.c
@@ -16,6 +16,7 @@
 #include <linux/dma-buf.h>
 #include <drm/tegra_drm.h>
 
+#include "drm.h"
 #include "gem.h"
 
 static inline struct tegra_bo *host1x_to_tegra_bo(struct host1x_bo *bo)
@@ -126,7 +127,7 @@ struct tegra_bo *tegra_bo_create(struct drm_device *drm, unsigned int size,
 		goto err_mmap;
 
 	if (flags & DRM_TEGRA_GEM_CREATE_TILED)
-		bo->flags |= TEGRA_BO_TILED;
+		bo->tiling.mode = TEGRA_BO_TILING_MODE_TILED;
 
 	if (flags & DRM_TEGRA_GEM_CREATE_BOTTOM_UP)
 		bo->flags |= TEGRA_BO_BOTTOM_UP;
@@ -259,8 +260,10 @@ int tegra_bo_dumb_create(struct drm_file *file, struct drm_device *drm,
 			 struct drm_mode_create_dumb *args)
 {
 	int min_pitch = DIV_ROUND_UP(args->width * args->bpp, 8);
+	struct tegra_drm *tegra = drm->dev_private;
 	struct tegra_bo *bo;
 
+	min_pitch = round_up(min_pitch, tegra->pitch_align);
 	if (args->pitch < min_pitch)
 		args->pitch = min_pitch;
 
diff --git a/drivers/gpu/drm/tegra/gem.h b/drivers/gpu/drm/tegra/gem.h
index 2f3fe96c5154da0bbd99f04c6b809c18bba03f32..43a25c853357d4e9b2d51b2049b8a12794bfc714 100644
--- a/drivers/gpu/drm/tegra/gem.h
+++ b/drivers/gpu/drm/tegra/gem.h
@@ -16,8 +16,18 @@
 #include <drm/drm.h>
 #include <drm/drmP.h>
 
-#define TEGRA_BO_TILED     (1 << 0)
-#define TEGRA_BO_BOTTOM_UP (1 << 1)
+#define TEGRA_BO_BOTTOM_UP (1 << 0)
+
+enum tegra_bo_tiling_mode {
+	TEGRA_BO_TILING_MODE_PITCH,
+	TEGRA_BO_TILING_MODE_TILED,
+	TEGRA_BO_TILING_MODE_BLOCK,
+};
+
+struct tegra_bo_tiling {
+	enum tegra_bo_tiling_mode mode;
+	unsigned long value;
+};
 
 struct tegra_bo {
 	struct drm_gem_object gem;
@@ -26,6 +36,8 @@ struct tegra_bo {
 	struct sg_table *sgt;
 	dma_addr_t paddr;
 	void *vaddr;
+
+	struct tegra_bo_tiling tiling;
 };
 
 static inline struct tegra_bo *to_tegra_bo(struct drm_gem_object *gem)
diff --git a/drivers/gpu/drm/tegra/gr2d.c b/drivers/gpu/drm/tegra/gr2d.c
index 7c53941f2a9ea347d6ff44eb1583efc629b611d7..02cd3e37a6ec3cf7d1e1d84f65387bb1c12853cc 100644
--- a/drivers/gpu/drm/tegra/gr2d.c
+++ b/drivers/gpu/drm/tegra/gr2d.c
@@ -121,6 +121,7 @@ static const struct of_device_id gr2d_match[] = {
 	{ .compatible = "nvidia,tegra20-gr2d" },
 	{ },
 };
+MODULE_DEVICE_TABLE(of, gr2d_match);
 
 static const u32 gr2d_addr_regs[] = {
 	GR2D_UA_BASE_ADDR,
diff --git a/drivers/gpu/drm/tegra/gr3d.c b/drivers/gpu/drm/tegra/gr3d.c
index 30f5ba9bd6d05c508eac1306c7b1468e6c515587..2bea2b2d204e809fd4dcc0b0bbaf7427317b8198 100644
--- a/drivers/gpu/drm/tegra/gr3d.c
+++ b/drivers/gpu/drm/tegra/gr3d.c
@@ -130,6 +130,7 @@ static const struct of_device_id tegra_gr3d_match[] = {
 	{ .compatible = "nvidia,tegra20-gr3d" },
 	{ }
 };
+MODULE_DEVICE_TABLE(of, tegra_gr3d_match);
 
 static const u32 gr3d_addr_regs[] = {
 	GR3D_IDX_ATTRIBUTE( 0),
diff --git a/drivers/gpu/drm/tegra/hdmi.c b/drivers/gpu/drm/tegra/hdmi.c
index ba067bb767e376ea4aaa5a7f5fe5917d1608b7c2..ffe26547328df8a352f8dca6e832912084cc7388 100644
--- a/drivers/gpu/drm/tegra/hdmi.c
+++ b/drivers/gpu/drm/tegra/hdmi.c
@@ -1450,6 +1450,7 @@ static const struct of_device_id tegra_hdmi_of_match[] = {
 	{ .compatible = "nvidia,tegra20-hdmi", .data = &tegra20_hdmi_config },
 	{ },
 };
+MODULE_DEVICE_TABLE(of, tegra_hdmi_of_match);
 
 static int tegra_hdmi_probe(struct platform_device *pdev)
 {
diff --git a/drivers/gpu/drm/tegra/sor.c b/drivers/gpu/drm/tegra/sor.c
index 27c979b5011112182226f7a1cd4d0e1e6b54c9ed..0410e467b82836b94e1bccd459b101ec78a5ce48 100644
--- a/drivers/gpu/drm/tegra/sor.c
+++ b/drivers/gpu/drm/tegra/sor.c
@@ -516,7 +516,7 @@ static int tegra_output_sor_enable(struct tegra_output *output)
 		if (err < 0) {
 			dev_err(sor->dev, "failed to probe eDP link: %d\n",
 				err);
-			return err;
+			goto unlock;
 		}
 	}
 
@@ -525,7 +525,7 @@ static int tegra_output_sor_enable(struct tegra_output *output)
 		dev_err(sor->dev, "failed to set safe parent clock: %d\n", err);
 
 	memset(&config, 0, sizeof(config));
-	config.bits_per_pixel = 24; /* XXX: don't hardcode? */
+	config.bits_per_pixel = output->connector.display_info.bpc * 3;
 
 	err = tegra_sor_calc_config(sor, mode, &config, &link);
 	if (err < 0)
@@ -815,12 +815,22 @@ static int tegra_output_sor_enable(struct tegra_output *output)
 	 * configure panel (24bpp, vsync-, hsync-, DP-A protocol, complete
 	 * raster, associate with display controller)
 	 */
-	value = SOR_STATE_ASY_VSYNCPOL |
-		SOR_STATE_ASY_HSYNCPOL |
-		SOR_STATE_ASY_PROTOCOL_DP_A |
+	value = SOR_STATE_ASY_PROTOCOL_DP_A |
 		SOR_STATE_ASY_CRC_MODE_COMPLETE |
 		SOR_STATE_ASY_OWNER(dc->pipe + 1);
 
+	if (mode->flags & DRM_MODE_FLAG_PHSYNC)
+		value &= ~SOR_STATE_ASY_HSYNCPOL;
+
+	if (mode->flags & DRM_MODE_FLAG_NHSYNC)
+		value |= SOR_STATE_ASY_HSYNCPOL;
+
+	if (mode->flags & DRM_MODE_FLAG_PVSYNC)
+		value &= ~SOR_STATE_ASY_VSYNCPOL;
+
+	if (mode->flags & DRM_MODE_FLAG_NVSYNC)
+		value |= SOR_STATE_ASY_VSYNCPOL;
+
 	switch (config.bits_per_pixel) {
 	case 24:
 		value |= SOR_STATE_ASY_PIXELDEPTH_BPP_24_444;
@@ -1455,6 +1465,7 @@ static const struct of_device_id tegra_sor_of_match[] = {
 	{ .compatible = "nvidia,tegra124-sor", },
 	{ },
 };
+MODULE_DEVICE_TABLE(of, tegra_sor_of_match);
 
 struct platform_driver tegra_sor_driver = {
 	.driver = {
diff --git a/drivers/gpu/host1x/job.c b/drivers/gpu/host1x/job.c
index 112f27e51bc7df81b45efcb58f7c11fc46f097e0..63bd63f3c7dfd2da2fd3d9ae58bba1473bceb70b 100644
--- a/drivers/gpu/host1x/job.c
+++ b/drivers/gpu/host1x/job.c
@@ -185,16 +185,16 @@ static unsigned int pin_job(struct host1x_job *job)
 		struct sg_table *sgt;
 		dma_addr_t phys_addr;
 
-		reloc->target = host1x_bo_get(reloc->target);
-		if (!reloc->target)
+		reloc->target.bo = host1x_bo_get(reloc->target.bo);
+		if (!reloc->target.bo)
 			goto unpin;
 
-		phys_addr = host1x_bo_pin(reloc->target, &sgt);
+		phys_addr = host1x_bo_pin(reloc->target.bo, &sgt);
 		if (!phys_addr)
 			goto unpin;
 
 		job->addr_phys[job->num_unpins] = phys_addr;
-		job->unpins[job->num_unpins].bo = reloc->target;
+		job->unpins[job->num_unpins].bo = reloc->target.bo;
 		job->unpins[job->num_unpins].sgt = sgt;
 		job->num_unpins++;
 	}
@@ -235,21 +235,21 @@ static unsigned int do_relocs(struct host1x_job *job, struct host1x_bo *cmdbuf)
 	for (i = 0; i < job->num_relocs; i++) {
 		struct host1x_reloc *reloc = &job->relocarray[i];
 		u32 reloc_addr = (job->reloc_addr_phys[i] +
-			reloc->target_offset) >> reloc->shift;
+				  reloc->target.offset) >> reloc->shift;
 		u32 *target;
 
 		/* skip all other gathers */
-		if (cmdbuf != reloc->cmdbuf)
+		if (cmdbuf != reloc->cmdbuf.bo)
 			continue;
 
-		if (last_page != reloc->cmdbuf_offset >> PAGE_SHIFT) {
+		if (last_page != reloc->cmdbuf.offset >> PAGE_SHIFT) {
 			if (cmdbuf_page_addr)
 				host1x_bo_kunmap(cmdbuf, last_page,
 						 cmdbuf_page_addr);
 
 			cmdbuf_page_addr = host1x_bo_kmap(cmdbuf,
-					reloc->cmdbuf_offset >> PAGE_SHIFT);
-			last_page = reloc->cmdbuf_offset >> PAGE_SHIFT;
+					reloc->cmdbuf.offset >> PAGE_SHIFT);
+			last_page = reloc->cmdbuf.offset >> PAGE_SHIFT;
 
 			if (unlikely(!cmdbuf_page_addr)) {
 				pr_err("Could not map cmdbuf for relocation\n");
@@ -257,7 +257,7 @@ static unsigned int do_relocs(struct host1x_job *job, struct host1x_bo *cmdbuf)
 			}
 		}
 
-		target = cmdbuf_page_addr + (reloc->cmdbuf_offset & ~PAGE_MASK);
+		target = cmdbuf_page_addr + (reloc->cmdbuf.offset & ~PAGE_MASK);
 		*target = reloc_addr;
 	}
 
@@ -272,7 +272,7 @@ static bool check_reloc(struct host1x_reloc *reloc, struct host1x_bo *cmdbuf,
 {
 	offset *= sizeof(u32);
 
-	if (reloc->cmdbuf != cmdbuf || reloc->cmdbuf_offset != offset)
+	if (reloc->cmdbuf.bo != cmdbuf || reloc->cmdbuf.offset != offset)
 		return false;
 
 	return true;
diff --git a/include/linux/host1x.h b/include/linux/host1x.h
index d2b52999e7717252a743204817ef726fd94a3567..bb9840fd1e18855305241daca6cc6a68ce18a066 100644
--- a/include/linux/host1x.h
+++ b/include/linux/host1x.h
@@ -164,12 +164,15 @@ int host1x_job_submit(struct host1x_job *job);
  */
 
 struct host1x_reloc {
-	struct host1x_bo *cmdbuf;
-	u32 cmdbuf_offset;
-	struct host1x_bo *target;
-	u32 target_offset;
-	u32 shift;
-	u32 pad;
+	struct {
+		struct host1x_bo *bo;
+		unsigned long offset;
+	} cmdbuf;
+	struct {
+		struct host1x_bo *bo;
+		unsigned long offset;
+	} target;
+	unsigned long shift;
 };
 
 struct host1x_job {
diff --git a/include/uapi/drm/tegra_drm.h b/include/uapi/drm/tegra_drm.h
index b75482112428afa8eaab6cd266f3221d8decab28..c15d781ecc0f306ce5fd575192f547b8e756273f 100644
--- a/include/uapi/drm/tegra_drm.h
+++ b/include/uapi/drm/tegra_drm.h
@@ -129,6 +129,44 @@ struct drm_tegra_submit {
 	__u32 reserved[5];	/* future expansion */
 };
 
+#define DRM_TEGRA_GEM_TILING_MODE_PITCH 0
+#define DRM_TEGRA_GEM_TILING_MODE_TILED 1
+#define DRM_TEGRA_GEM_TILING_MODE_BLOCK 2
+
+struct drm_tegra_gem_set_tiling {
+	/* input */
+	__u32 handle;
+	__u32 mode;
+	__u32 value;
+	__u32 pad;
+};
+
+struct drm_tegra_gem_get_tiling {
+	/* input */
+	__u32 handle;
+	/* output */
+	__u32 mode;
+	__u32 value;
+	__u32 pad;
+};
+
+#define DRM_TEGRA_GEM_BOTTOM_UP		(1 << 0)
+#define DRM_TEGRA_GEM_FLAGS		(DRM_TEGRA_GEM_BOTTOM_UP)
+
+struct drm_tegra_gem_set_flags {
+	/* input */
+	__u32 handle;
+	/* output */
+	__u32 flags;
+};
+
+struct drm_tegra_gem_get_flags {
+	/* input */
+	__u32 handle;
+	/* output */
+	__u32 flags;
+};
+
 #define DRM_TEGRA_GEM_CREATE		0x00
 #define DRM_TEGRA_GEM_MMAP		0x01
 #define DRM_TEGRA_SYNCPT_READ		0x02
@@ -139,6 +177,10 @@ struct drm_tegra_submit {
 #define DRM_TEGRA_GET_SYNCPT		0x07
 #define DRM_TEGRA_SUBMIT		0x08
 #define DRM_TEGRA_GET_SYNCPT_BASE	0x09
+#define DRM_TEGRA_GEM_SET_TILING	0x0a
+#define DRM_TEGRA_GEM_GET_TILING	0x0b
+#define DRM_TEGRA_GEM_SET_FLAGS		0x0c
+#define DRM_TEGRA_GEM_GET_FLAGS		0x0d
 
 #define DRM_IOCTL_TEGRA_GEM_CREATE DRM_IOWR(DRM_COMMAND_BASE + DRM_TEGRA_GEM_CREATE, struct drm_tegra_gem_create)
 #define DRM_IOCTL_TEGRA_GEM_MMAP DRM_IOWR(DRM_COMMAND_BASE + DRM_TEGRA_GEM_MMAP, struct drm_tegra_gem_mmap)
@@ -150,5 +192,9 @@ struct drm_tegra_submit {
 #define DRM_IOCTL_TEGRA_GET_SYNCPT DRM_IOWR(DRM_COMMAND_BASE + DRM_TEGRA_GET_SYNCPT, struct drm_tegra_get_syncpt)
 #define DRM_IOCTL_TEGRA_SUBMIT DRM_IOWR(DRM_COMMAND_BASE + DRM_TEGRA_SUBMIT, struct drm_tegra_submit)
 #define DRM_IOCTL_TEGRA_GET_SYNCPT_BASE DRM_IOWR(DRM_COMMAND_BASE + DRM_TEGRA_GET_SYNCPT_BASE, struct drm_tegra_get_syncpt_base)
+#define DRM_IOCTL_TEGRA_GEM_SET_TILING DRM_IOWR(DRM_COMMAND_BASE + DRM_TEGRA_GEM_SET_TILING, struct drm_tegra_gem_set_tiling)
+#define DRM_IOCTL_TEGRA_GEM_GET_TILING DRM_IOWR(DRM_COMMAND_BASE + DRM_TEGRA_GEM_GET_TILING, struct drm_tegra_gem_get_tiling)
+#define DRM_IOCTL_TEGRA_GEM_SET_FLAGS DRM_IOWR(DRM_COMMAND_BASE + DRM_TEGRA_GEM_SET_FLAGS, struct drm_tegra_gem_set_flags)
+#define DRM_IOCTL_TEGRA_GEM_GET_FLAGS DRM_IOWR(DRM_COMMAND_BASE + DRM_TEGRA_GEM_GET_FLAGS, struct drm_tegra_gem_get_flags)
 
 #endif