Merge branch 'drm-vmware-fixes' into drm-testing

* drm-vmware-fixes:
  drm/vmwgfx: Remove some leftover debug messages.
  drm/vmwgfx: Print warnings in kernel log about bo pinning that fails.
  drm/vmwgfx: Unpause overlay on update.
  drm/vmwgfx: Some modesetting cleanups and fixes.
  drm/vmwgfx: Don't use SVGA_REG_ENABLE in modesetting code.
  drm/vmwgfx: Remove duplicate member from struct vmw_legacy_display_unit.
  drm/vmwgfx: Reserve first part of VRAM for framebuffer.
  drm/vmwgfx: Support older hardware.
  drm/vmwgfx: Get connector status from detection function.
  drm/vmwgfx: Add kernel throttling support. Bump minor.
  drm/vmwgfx: Make sure to unpin old and pin new framebuffer.
  drm/vmwgfx: Fix single framebuffer detection.
  drm/vmwgfx: Assume larger framebuffer max size.

11 files changed, 433 insertions, 149 deletions

diff --git a/drivers/gpu/drm/vmwgfx/Makefile b/drivers/gpu/drm/vmwgfx/Makefile
index 1a3cb6816d1c..4505e17df3f5 100644
--- a/drivers/gpu/drm/vmwgfx/Makefile
+++ b/drivers/gpu/drm/vmwgfx/Makefile
@@ -4,6 +4,6 @@ ccflags-y := -Iinclude/drm
 vmwgfx-y := vmwgfx_execbuf.o vmwgfx_gmr.o vmwgfx_kms.o vmwgfx_drv.o \
 	    vmwgfx_fb.o vmwgfx_ioctl.o vmwgfx_resource.o vmwgfx_buffer.o \
 	    vmwgfx_fifo.o vmwgfx_irq.o vmwgfx_ldu.o vmwgfx_ttm_glue.o \
-	    vmwgfx_overlay.o
+	    vmwgfx_overlay.o vmwgfx_fence.o
 
 obj-$(CONFIG_DRM_VMWGFX) := vmwgfx.o
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c
index 0c9c0811f42d..7597323d5a5a 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c
@@ -318,6 +318,15 @@ static int vmw_driver_load(struct drm_device *dev, unsigned long chipset)
 		goto out_err3;
 	}
 
+	/* Need mmio memory to check for fifo pitchlock cap. */
+	if (!(dev_priv->capabilities & SVGA_CAP_DISPLAY_TOPOLOGY) &&
+	    !(dev_priv->capabilities & SVGA_CAP_PITCHLOCK) &&
+	    !vmw_fifo_have_pitchlock(dev_priv)) {
+		ret = -ENOSYS;
+		DRM_ERROR("Hardware has no pitchlock\n");
+		goto out_err4;
+	}
+
 	dev_priv->tdev = ttm_object_device_init
 	    (dev_priv->mem_global_ref.object, 12);
 
@@ -399,8 +408,6 @@ static int vmw_driver_unload(struct drm_device *dev)
 {
 	struct vmw_private *dev_priv = vmw_priv(dev);
 
-	DRM_INFO(VMWGFX_DRIVER_NAME " unload.\n");
-
 	unregister_pm_notifier(&dev_priv->pm_nb);
 
 	vmw_fb_close(dev_priv);
@@ -546,7 +553,6 @@ static int vmw_master_create(struct drm_device *dev,
 {
 	struct vmw_master *vmaster;
 
-	DRM_INFO("Master create.\n");
 	vmaster = kzalloc(sizeof(*vmaster), GFP_KERNEL);
 	if (unlikely(vmaster == NULL))
 		return -ENOMEM;
@@ -563,7 +569,6 @@ static void vmw_master_destroy(struct drm_device *dev,
 {
 	struct vmw_master *vmaster = vmw_master(master);
 
-	DRM_INFO("Master destroy.\n");
 	master->driver_priv = NULL;
 	kfree(vmaster);
 }
@@ -579,8 +584,6 @@ static int vmw_master_set(struct drm_device *dev,
 	struct vmw_master *vmaster = vmw_master(file_priv->master);
 	int ret = 0;
 
-	DRM_INFO("Master set.\n");
-
 	if (active) {
 		BUG_ON(active != &dev_priv->fbdev_master);
 		ret = ttm_vt_lock(&active->lock, false, vmw_fp->tfile);
@@ -622,8 +625,6 @@ static void vmw_master_drop(struct drm_device *dev,
 	struct vmw_master *vmaster = vmw_master(file_priv->master);
 	int ret;
 
-	DRM_INFO("Master drop.\n");
-
 	/**
 	 * Make sure the master doesn't disappear while we have
 	 * it locked.
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h
index 356dc935ec13..1341adef408d 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h
@@ -41,7 +41,7 @@
 
 #define VMWGFX_DRIVER_DATE "20100209"
 #define VMWGFX_DRIVER_MAJOR 1
-#define VMWGFX_DRIVER_MINOR 0
+#define VMWGFX_DRIVER_MINOR 1
 #define VMWGFX_DRIVER_PATCHLEVEL 0
 #define VMWGFX_FILE_PAGE_OFFSET 0x00100000
 #define VMWGFX_FIFO_STATIC_SIZE (1024*1024)
@@ -102,6 +102,13 @@ struct vmw_surface {
 	struct vmw_cursor_snooper snooper;
 };
 
+struct vmw_fence_queue {
+	struct list_head head;
+	struct timespec lag;
+	struct timespec lag_time;
+	spinlock_t lock;
+};
+
 struct vmw_fifo_state {
 	unsigned long reserved_size;
 	__le32 *dynamic_buffer;
@@ -115,6 +122,7 @@ struct vmw_fifo_state {
 	uint32_t capabilities;
 	struct mutex fifo_mutex;
 	struct rw_semaphore rwsem;
+	struct vmw_fence_queue fence_queue;
 };
 
 struct vmw_relocation {
@@ -179,6 +187,7 @@ struct vmw_private {
 	uint32_t vga_red_mask;
 	uint32_t vga_blue_mask;
 	uint32_t vga_green_mask;
+	uint32_t vga_pitchlock;
 
 	/*
 	 * Framebuffer info.
@@ -393,6 +402,7 @@ extern int vmw_fifo_send_fence(struct vmw_private *dev_priv,
 extern void vmw_fifo_ping_host(struct vmw_private *dev_priv, uint32_t reason);
 extern int vmw_fifo_mmap(struct file *filp, struct vm_area_struct *vma);
 extern bool vmw_fifo_have_3d(struct vmw_private *dev_priv);
+extern bool vmw_fifo_have_pitchlock(struct vmw_private *dev_priv);
 
 /**
  * TTM glue - vmwgfx_ttm_glue.c
@@ -441,6 +451,23 @@ extern int vmw_fallback_wait(struct vmw_private *dev_priv,
 			     uint32_t sequence,
 			     bool interruptible,
 			     unsigned long timeout);
+extern void vmw_update_sequence(struct vmw_private *dev_priv,
+				struct vmw_fifo_state *fifo_state);
+
+
+/**
+ * Rudimentary fence objects currently used only for throttling -
+ * vmwgfx_fence.c
+ */
+
+extern void vmw_fence_queue_init(struct vmw_fence_queue *queue);
+extern void vmw_fence_queue_takedown(struct vmw_fence_queue *queue);
+extern int vmw_fence_push(struct vmw_fence_queue *queue,
+			  uint32_t sequence);
+extern int vmw_fence_pull(struct vmw_fence_queue *queue,
+			  uint32_t signaled_sequence);
+extern int vmw_wait_lag(struct vmw_private *dev_priv,
+			struct vmw_fence_queue *queue, uint32_t us);
 
 /**
  * Kernel framebuffer - vmwgfx_fb.c
@@ -466,6 +493,9 @@ void vmw_kms_cursor_snoop(struct vmw_surface *srf,
 			  struct ttm_object_file *tfile,
 			  struct ttm_buffer_object *bo,
 			  SVGA3dCmdHeader *header);
+void vmw_kms_write_svga(struct vmw_private *vmw_priv,
+			unsigned width, unsigned height, unsigned pitch,
+			unsigned bbp, unsigned depth);
 
 /**
  * Overlay control - vmwgfx_overlay.c
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_execbuf.c b/drivers/gpu/drm/vmwgfx/vmwgfx_execbuf.c
index dbd36b8910cf..bdd67cf83315 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_execbuf.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_execbuf.c
@@ -669,6 +669,15 @@ int vmw_execbuf_ioctl(struct drm_device *dev, void *data,
 		goto out_err;
 
 	vmw_apply_relocations(sw_context);
+
+	if (arg->throttle_us) {
+		ret = vmw_wait_lag(dev_priv, &dev_priv->fifo.fence_queue,
+				   arg->throttle_us);
+
+		if (unlikely(ret != 0))
+			goto out_err;
+	}
+
 	vmw_fifo_commit(dev_priv, arg->command_size);
 
 	ret = vmw_fifo_send_fence(dev_priv, &sequence);
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_fb.c b/drivers/gpu/drm/vmwgfx/vmwgfx_fb.c
index 7421aaad8d09..181f47222580 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_fb.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_fb.c
@@ -132,16 +132,14 @@ static int vmw_fb_check_var(struct fb_var_screeninfo *var,
 		return -EINVAL;
 	}
 
-	/* without multimon its hard to resize */
-	if (!(vmw_priv->capabilities & SVGA_CAP_MULTIMON) &&
-	    (var->xres != par->max_width ||
-	     var->yres != par->max_height)) {
-		DRM_ERROR("Tried to resize, but we don't have multimon\n");
+	if (!(vmw_priv->capabilities & SVGA_CAP_DISPLAY_TOPOLOGY) &&
+	    (var->xoffset != 0 || var->yoffset != 0)) {
+		DRM_ERROR("Can not handle panning without display topology\n");
 		return -EINVAL;
 	}
 
-	if (var->xres > par->max_width ||
-	    var->yres > par->max_height) {
+	if ((var->xoffset + var->xres) > par->max_width ||
+	    (var->yoffset + var->yres) > par->max_height) {
 		DRM_ERROR("Requested geom can not fit in framebuffer\n");
 		return -EINVAL;
 	}
@@ -154,8 +152,7 @@ static int vmw_fb_set_par(struct fb_info *info)
 	struct vmw_fb_par *par = info->par;
 	struct vmw_private *vmw_priv = par->vmw_priv;
 
-	if (vmw_priv->capabilities & SVGA_CAP_MULTIMON) {
-		vmw_write(vmw_priv, SVGA_REG_NUM_GUEST_DISPLAYS, 1);
+	if (vmw_priv->capabilities & SVGA_CAP_DISPLAY_TOPOLOGY) {
 		vmw_write(vmw_priv, SVGA_REG_DISPLAY_ID, 0);
 		vmw_write(vmw_priv, SVGA_REG_DISPLAY_IS_PRIMARY, true);
 		vmw_write(vmw_priv, SVGA_REG_DISPLAY_POSITION_X, 0);
@@ -164,18 +161,11 @@ static int vmw_fb_set_par(struct fb_info *info)
 		vmw_write(vmw_priv, SVGA_REG_DISPLAY_HEIGHT, 0);
 		vmw_write(vmw_priv, SVGA_REG_DISPLAY_ID, SVGA_ID_INVALID);
 
-		vmw_write(vmw_priv, SVGA_REG_ENABLE, 1);
-		vmw_write(vmw_priv, SVGA_REG_WIDTH, par->max_width);
-		vmw_write(vmw_priv, SVGA_REG_HEIGHT, par->max_height);
-		vmw_write(vmw_priv, SVGA_REG_BITS_PER_PIXEL, par->bpp);
-		vmw_write(vmw_priv, SVGA_REG_DEPTH, par->depth);
-		vmw_write(vmw_priv, SVGA_REG_RED_MASK, 0x00ff0000);
-		vmw_write(vmw_priv, SVGA_REG_GREEN_MASK, 0x0000ff00);
-		vmw_write(vmw_priv, SVGA_REG_BLUE_MASK, 0x000000ff);
+		vmw_kms_write_svga(vmw_priv, info->var.xres, info->var.yres,
+				   info->fix.line_length,
+				   par->bpp, par->depth);
 
 		/* TODO check if pitch and offset changes */
-
-		vmw_write(vmw_priv, SVGA_REG_NUM_GUEST_DISPLAYS, 1);
 		vmw_write(vmw_priv, SVGA_REG_DISPLAY_ID, 0);
 		vmw_write(vmw_priv, SVGA_REG_DISPLAY_IS_PRIMARY, true);
 		vmw_write(vmw_priv, SVGA_REG_DISPLAY_POSITION_X, info->var.xoffset);
@@ -183,13 +173,19 @@ static int vmw_fb_set_par(struct fb_info *info)
 		vmw_write(vmw_priv, SVGA_REG_DISPLAY_WIDTH, info->var.xres);
 		vmw_write(vmw_priv, SVGA_REG_DISPLAY_HEIGHT, info->var.yres);
 		vmw_write(vmw_priv, SVGA_REG_DISPLAY_ID, SVGA_ID_INVALID);
+		vmw_write(vmw_priv, SVGA_REG_NUM_GUEST_DISPLAYS, 1);
 	} else {
-		vmw_write(vmw_priv, SVGA_REG_WIDTH, info->var.xres);
-		vmw_write(vmw_priv, SVGA_REG_HEIGHT, info->var.yres);
+		vmw_kms_write_svga(vmw_priv, info->var.xres, info->var.yres,
+				   info->fix.line_length,
+				   par->bpp, par->depth);
 
-		/* TODO check if pitch and offset changes */
 	}
 
+	/* This is really helpful since if this fails the user
+	 * can probably not see anything on the screen.
+	 */
+	WARN_ON(vmw_read(vmw_priv, SVGA_REG_FB_OFFSET) != 0);
+
 	return 0;
 }
 
@@ -416,48 +412,23 @@ int vmw_fb_init(struct vmw_private *vmw_priv)
 	unsigned fb_bbp, fb_depth, fb_offset, fb_pitch, fb_size;
 	int ret;
 
+	/* XXX These shouldn't be hardcoded. */
 	initial_width = 800;
 	initial_height = 600;
 
 	fb_bbp = 32;
 	fb_depth = 24;
 
-	if (vmw_priv->capabilities & SVGA_CAP_MULTIMON) {
-		fb_width = min(vmw_priv->fb_max_width, (unsigned)2048);
-		fb_height = min(vmw_priv->fb_max_height, (unsigned)2048);
-	} else {
-		fb_width = min(vmw_priv->fb_max_width, initial_width);
-		fb_height = min(vmw_priv->fb_max_height, initial_height);
-	}
+	/* XXX As shouldn't these be as well. */
+	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);
 
-	vmw_write(vmw_priv, SVGA_REG_WIDTH, fb_width);
-	vmw_write(vmw_priv, SVGA_REG_HEIGHT, fb_height);
-	vmw_write(vmw_priv, SVGA_REG_BITS_PER_PIXEL, fb_bbp);
-	vmw_write(vmw_priv, SVGA_REG_DEPTH, fb_depth);
-	vmw_write(vmw_priv, SVGA_REG_RED_MASK, 0x00ff0000);
-	vmw_write(vmw_priv, SVGA_REG_GREEN_MASK, 0x0000ff00);
-	vmw_write(vmw_priv, SVGA_REG_BLUE_MASK, 0x000000ff);
-
-	fb_size = vmw_read(vmw_priv, SVGA_REG_FB_SIZE);
+	fb_pitch = fb_width * fb_bbp / 8;
+	fb_size = fb_pitch * fb_height;
 	fb_offset = vmw_read(vmw_priv, SVGA_REG_FB_OFFSET);
-	fb_pitch = vmw_read(vmw_priv, SVGA_REG_BYTES_PER_LINE);
-
-	DRM_DEBUG("width  %u\n", vmw_read(vmw_priv, SVGA_REG_MAX_WIDTH));
-	DRM_DEBUG("height %u\n", vmw_read(vmw_priv, SVGA_REG_MAX_HEIGHT));
-	DRM_DEBUG("width  %u\n", vmw_read(vmw_priv, SVGA_REG_WIDTH));
-	DRM_DEBUG("height %u\n", vmw_read(vmw_priv, SVGA_REG_HEIGHT));
-	DRM_DEBUG("bpp    %u\n", vmw_read(vmw_priv, SVGA_REG_BITS_PER_PIXEL));
-	DRM_DEBUG("depth  %u\n", vmw_read(vmw_priv, SVGA_REG_DEPTH));
-	DRM_DEBUG("bpl    %u\n", vmw_read(vmw_priv, SVGA_REG_BYTES_PER_LINE));
-	DRM_DEBUG("r mask %08x\n", vmw_read(vmw_priv, SVGA_REG_RED_MASK));
-	DRM_DEBUG("g mask %08x\n", vmw_read(vmw_priv, SVGA_REG_GREEN_MASK));
-	DRM_DEBUG("b mask %08x\n", vmw_read(vmw_priv, SVGA_REG_BLUE_MASK));
-	DRM_DEBUG("fb_offset 0x%08x\n", fb_offset);
-	DRM_DEBUG("fb_pitch  %u\n", fb_pitch);
-	DRM_DEBUG("fb_size   %u kiB\n", fb_size / 1024);
 
 	info = framebuffer_alloc(sizeof(*par), device);
 	if (!info)
@@ -659,6 +630,10 @@ int vmw_dmabuf_to_start_of_vram(struct vmw_private *vmw_priv,
 		goto err_unlock;
 
 	ret = ttm_bo_validate(bo, &ne_placement, false, false, false);
+
+	/* Could probably bug on */
+	WARN_ON(bo->offset != 0);
+
 	ttm_bo_unreserve(bo);
 err_unlock:
 	ttm_write_unlock(&vmw_priv->active_master->lock);
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_fence.c b/drivers/gpu/drm/vmwgfx/vmwgfx_fence.c
new file mode 100644
index 000000000000..61eacc1b5ca3
--- /dev/null
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_fence.c
@@ -0,0 +1,173 @@
+/**************************************************************************
+ *
+ * Copyright (C) 2010 VMware, Inc., Palo Alto, CA., USA
+ * 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
+ * THE COPYRIGHT HOLDERS, AUTHORS AND/OR ITS SUPPLIERS BE LIABLE FOR ANY CLAIM,
+ * DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR
+ * OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE
+ * USE OR OTHER DEALINGS IN THE SOFTWARE.
+ *
+ **************************************************************************/
+
+
+#include "vmwgfx_drv.h"
+
+struct vmw_fence {
+	struct list_head head;
+	uint32_t sequence;
+	struct timespec submitted;
+};
+
+void vmw_fence_queue_init(struct vmw_fence_queue *queue)
+{
+	INIT_LIST_HEAD(&queue->head);
+	queue->lag = ns_to_timespec(0);
+	getrawmonotonic(&queue->lag_time);
+	spin_lock_init(&queue->lock);
+}
+
+void vmw_fence_queue_takedown(struct vmw_fence_queue *queue)
+{
+	struct vmw_fence *fence, *next;
+
+	spin_lock(&queue->lock);
+	list_for_each_entry_safe(fence, next, &queue->head, head) {
+		kfree(fence);
+	}
+	spin_unlock(&queue->lock);
+}
+
+int vmw_fence_push(struct vmw_fence_queue *queue,
+		   uint32_t sequence)
+{
+	struct vmw_fence *fence = kmalloc(sizeof(*fence), GFP_KERNEL);
+
+	if (unlikely(!fence))
+		return -ENOMEM;
+
+	fence->sequence = sequence;
+	getrawmonotonic(&fence->submitted);
+	spin_lock(&queue->lock);
+	list_add_tail(&fence->head, &queue->head);
+	spin_unlock(&queue->lock);
+
+	return 0;
+}
+
+int vmw_fence_pull(struct vmw_fence_queue *queue,
+		   uint32_t signaled_sequence)
+{
+	struct vmw_fence *fence, *next;
+	struct timespec now;
+	bool updated = false;
+
+	spin_lock(&queue->lock);
+	getrawmonotonic(&now);
+
+	if (list_empty(&queue->head)) {
+		queue->lag = ns_to_timespec(0);
+		queue->lag_time = now;
+		updated = true;
+		goto out_unlock;
+	}
+
+	list_for_each_entry_safe(fence, next, &queue->head, head) {
+		if (signaled_sequence - fence->sequence > (1 << 30))
+			continue;
+
+		queue->lag = timespec_sub(now, fence->submitted);
+		queue->lag_time = now;
+		updated = true;
+		list_del(&fence->head);
+		kfree(fence);
+	}
+
+out_unlock:
+	spin_unlock(&queue->lock);
+
+	return (updated) ? 0 : -EBUSY;
+}
+
+static struct timespec vmw_timespec_add(struct timespec t1,
+					struct timespec t2)
+{
+	t1.tv_sec += t2.tv_sec;
+	t1.tv_nsec += t2.tv_nsec;
+	if (t1.tv_nsec >= 1000000000L) {
+		t1.tv_sec += 1;
+		t1.tv_nsec -= 1000000000L;
+	}
+
+	return t1;
+}
+
+static struct timespec vmw_fifo_lag(struct vmw_fence_queue *queue)
+{
+	struct timespec now;
+
+	spin_lock(&queue->lock);
+	getrawmonotonic(&now);
+	queue->lag = vmw_timespec_add(queue->lag,
+				      timespec_sub(now, queue->lag_time));
+	queue->lag_time = now;
+	spin_unlock(&queue->lock);
+	return queue->lag;
+}
+
+
+static bool vmw_lag_lt(struct vmw_fence_queue *queue,
+		       uint32_t us)
+{
+	struct timespec lag, cond;
+
+	cond = ns_to_timespec((s64) us * 1000);
+	lag = vmw_fifo_lag(queue);
+	return (timespec_compare(&lag, &cond) < 1);
+}
+
+int vmw_wait_lag(struct vmw_private *dev_priv,
+		 struct vmw_fence_queue *queue, uint32_t us)
+{
+	struct vmw_fence *fence;
+	uint32_t sequence;
+	int ret;
+
+	while (!vmw_lag_lt(queue, us)) {
+		spin_lock(&queue->lock);
+		if (list_empty(&queue->head))
+			sequence = atomic_read(&dev_priv->fence_seq);
+		else {
+			fence = list_first_entry(&queue->head,
+						 struct vmw_fence, head);
+			sequence = fence->sequence;
+		}
+		spin_unlock(&queue->lock);
+
+		ret = vmw_wait_fence(dev_priv, false, sequence, true,
+				     3*HZ);
+
+		if (unlikely(ret != 0))
+			return ret;
+
+		(void) vmw_fence_pull(queue, sequence);
+	}
+	return 0;
+}
+
+
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_fifo.c b/drivers/gpu/drm/vmwgfx/vmwgfx_fifo.c
index 39d43a01d846..e6a1eb7ea954 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_fifo.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_fifo.c
@@ -34,6 +34,9 @@ bool vmw_fifo_have_3d(struct vmw_private *dev_priv)
 	__le32 __iomem *fifo_mem = dev_priv->mmio_virt;
 	uint32_t fifo_min, hwversion;
 
+	if (!(dev_priv->capabilities & SVGA_CAP_EXTENDED_FIFO))
+		return false;
+
 	fifo_min = ioread32(fifo_mem  + SVGA_FIFO_MIN);
 	if (fifo_min <= SVGA_FIFO_3D_HWVERSION * sizeof(unsigned int))
 		return false;
@@ -48,6 +51,21 @@ bool vmw_fifo_have_3d(struct vmw_private *dev_priv)
 	return true;
 }
 
+bool vmw_fifo_have_pitchlock(struct vmw_private *dev_priv)
+{
+	__le32 __iomem *fifo_mem = dev_priv->mmio_virt;
+	uint32_t caps;
+
+	if (!(dev_priv->capabilities & SVGA_CAP_EXTENDED_FIFO))
+		return false;
+
+	caps = ioread32(fifo_mem + SVGA_FIFO_CAPABILITIES);
+	if (caps & SVGA_FIFO_CAP_PITCHLOCK)
+		return true;
+
+	return false;
+}
+
 int vmw_fifo_init(struct vmw_private *dev_priv, struct vmw_fifo_state *fifo)
 {
 	__le32 __iomem *fifo_mem = dev_priv->mmio_virt;
@@ -120,7 +138,7 @@ int vmw_fifo_init(struct vmw_private *dev_priv, struct vmw_fifo_state *fifo)
 
 	atomic_set(&dev_priv->fence_seq, dev_priv->last_read_sequence);
 	iowrite32(dev_priv->last_read_sequence, fifo_mem + SVGA_FIFO_FENCE);
-
+	vmw_fence_queue_init(&fifo->fence_queue);
 	return vmw_fifo_send_fence(dev_priv, &dummy);
 out_err:
 	vfree(fifo->static_buffer);
@@ -159,6 +177,7 @@ void vmw_fifo_release(struct vmw_private *dev_priv, struct vmw_fifo_state *fifo)
 		  dev_priv->enable_state);
 
 	mutex_unlock(&dev_priv->hw_mutex);
+	vmw_fence_queue_takedown(&fifo->fence_queue);
 
 	if (likely(fifo->last_buffer != NULL)) {
 		vfree(fifo->last_buffer);
@@ -484,6 +503,8 @@ int vmw_fifo_send_fence(struct vmw_private *dev_priv, uint32_t *sequence)
 	fifo_state->last_buffer_add = true;
 	vmw_fifo_commit(dev_priv, bytes);
 	fifo_state->last_buffer_add = false;
+	(void) vmw_fence_push(&fifo_state->fence_queue, *sequence);
+	vmw_update_sequence(dev_priv, fifo_state);
 
 out_err:
 	return ret;
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_irq.c b/drivers/gpu/drm/vmwgfx/vmwgfx_irq.c
index 4d7cb5393860..e92298a6a383 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_irq.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_irq.c
@@ -64,22 +64,33 @@ static bool vmw_fifo_idle(struct vmw_private *dev_priv, uint32_t sequence)
 	return (busy == 0);
 }
 
+void vmw_update_sequence(struct vmw_private *dev_priv,
+			 struct vmw_fifo_state *fifo_state)
+{
+	__le32 __iomem *fifo_mem = dev_priv->mmio_virt;
+
+	uint32_t sequence = ioread32(fifo_mem + SVGA_FIFO_FENCE);
+
+	if (dev_priv->last_read_sequence != sequence) {
+		dev_priv->last_read_sequence = sequence;
+		vmw_fence_pull(&fifo_state->fence_queue, sequence);
+	}
+}
 
 bool vmw_fence_signaled(struct vmw_private *dev_priv,
 			uint32_t sequence)
 {
-	__le32 __iomem *fifo_mem = dev_priv->mmio_virt;
 	struct vmw_fifo_state *fifo_state;
 	bool ret;
 
 	if (likely(dev_priv->last_read_sequence - sequence < VMW_FENCE_WRAP))
 		return true;
 
-	dev_priv->last_read_sequence = ioread32(fifo_mem + SVGA_FIFO_FENCE);
+	fifo_state = &dev_priv->fifo;
+	vmw_update_sequence(dev_priv, fifo_state);
 	if (likely(dev_priv->last_read_sequence - sequence < VMW_FENCE_WRAP))
 		return true;
 
-	fifo_state = &dev_priv->fifo;
 	if (!(fifo_state->capabilities & SVGA_FIFO_CAP_FENCE) &&
 	    vmw_fifo_idle(dev_priv, sequence))
 		return true;
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_kms.c b/drivers/gpu/drm/vmwgfx/vmwgfx_kms.c
index bbc7c4c30bc7..b78dcf001858 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_kms.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_kms.c
@@ -30,6 +30,8 @@
 /* Might need a hrtimer here? */
 #define VMWGFX_PRESENT_RATE ((HZ / 60 > 0) ? HZ / 60 : 1)
 
+static int vmw_surface_dmabuf_pin(struct vmw_framebuffer *vfb);
+static int vmw_surface_dmabuf_unpin(struct vmw_framebuffer *vfb);
 
 void vmw_display_unit_cleanup(struct vmw_display_unit *du)
 {
@@ -326,6 +328,7 @@ int vmw_framebuffer_create_handle(struct drm_framebuffer *fb,
 struct vmw_framebuffer_surface {
 	struct vmw_framebuffer base;
 	struct vmw_surface *surface;
+	struct vmw_dma_buffer *buffer;
 	struct delayed_work d_work;
 	struct mutex work_lock;
 	bool present_fs;
@@ -500,8 +503,8 @@ int vmw_kms_new_framebuffer_surface(struct vmw_private *dev_priv,
 	vfbs->base.base.depth = 24;
 	vfbs->base.base.width = width;
 	vfbs->base.base.height = height;
-	vfbs->base.pin = NULL;
-	vfbs->base.unpin = NULL;
+	vfbs->base.pin = &vmw_surface_dmabuf_pin;
+	vfbs->base.unpin = &vmw_surface_dmabuf_unpin;
 	vfbs->surface = surface;
 	mutex_init(&vfbs->work_lock);
 	INIT_DELAYED_WORK(&vfbs->d_work, &vmw_framebuffer_present_fs_callback);
@@ -589,6 +592,40 @@ static struct drm_framebuffer_funcs vmw_framebuffer_dmabuf_funcs = {
 	.create_handle = vmw_framebuffer_create_handle,
 };
 
+static int vmw_surface_dmabuf_pin(struct vmw_framebuffer *vfb)
+{
+	struct vmw_private *dev_priv = vmw_priv(vfb->base.dev);
+	struct vmw_framebuffer_surface *vfbs =
+		vmw_framebuffer_to_vfbs(&vfb->base);
+	unsigned long size = vfbs->base.base.pitch * vfbs->base.base.height;
+	int ret;
+
+	vfbs->buffer = kzalloc(sizeof(*vfbs->buffer), GFP_KERNEL);
+	if (unlikely(vfbs->buffer == NULL))
+		return -ENOMEM;
+
+	vmw_overlay_pause_all(dev_priv);
+	ret = vmw_dmabuf_init(dev_priv, vfbs->buffer, size,
+			       &vmw_vram_ne_placement,
+			       false, &vmw_dmabuf_bo_free);
+	vmw_overlay_resume_all(dev_priv);
+
+	return ret;
+}
+
+static int vmw_surface_dmabuf_unpin(struct vmw_framebuffer *vfb)
+{
+	struct ttm_buffer_object *bo;
+	struct vmw_framebuffer_surface *vfbs =
+		vmw_framebuffer_to_vfbs(&vfb->base);
+
+	bo = &vfbs->buffer->base;
+	ttm_bo_unref(&bo);
+	vfbs->buffer = NULL;
+
+	return 0;
+}
+
 static int vmw_framebuffer_dmabuf_pin(struct vmw_framebuffer *vfb)
 {
 	struct vmw_private *dev_priv = vmw_priv(vfb->base.dev);
@@ -596,33 +633,15 @@ static int vmw_framebuffer_dmabuf_pin(struct vmw_framebuffer *vfb)
 		vmw_framebuffer_to_vfbd(&vfb->base);
 	int ret;
 
+
 	vmw_overlay_pause_all(dev_priv);
 
 	ret = vmw_dmabuf_to_start_of_vram(dev_priv, vfbd->buffer);
 
-	if (dev_priv->capabilities & SVGA_CAP_MULTIMON) {
-		vmw_write(dev_priv, SVGA_REG_NUM_GUEST_DISPLAYS, 1);
-		vmw_write(dev_priv, SVGA_REG_DISPLAY_ID, 0);
-		vmw_write(dev_priv, SVGA_REG_DISPLAY_IS_PRIMARY, true);
-		vmw_write(dev_priv, SVGA_REG_DISPLAY_POSITION_X, 0);
-		vmw_write(dev_priv, SVGA_REG_DISPLAY_POSITION_Y, 0);
-		vmw_write(dev_priv, SVGA_REG_DISPLAY_WIDTH, 0);
-		vmw_write(dev_priv, SVGA_REG_DISPLAY_HEIGHT, 0);
-		vmw_write(dev_priv, SVGA_REG_DISPLAY_ID, SVGA_ID_INVALID);
-
-		vmw_write(dev_priv, SVGA_REG_ENABLE, 1);
-		vmw_write(dev_priv, SVGA_REG_WIDTH, vfb->base.width);
-		vmw_write(dev_priv, SVGA_REG_HEIGHT, vfb->base.height);
-		vmw_write(dev_priv, SVGA_REG_BITS_PER_PIXEL, vfb->base.bits_per_pixel);
-		vmw_write(dev_priv, SVGA_REG_DEPTH, vfb->base.depth);
-		vmw_write(dev_priv, SVGA_REG_RED_MASK, 0x00ff0000);
-		vmw_write(dev_priv, SVGA_REG_GREEN_MASK, 0x0000ff00);
-		vmw_write(dev_priv, SVGA_REG_BLUE_MASK, 0x000000ff);
-	} else
-		WARN_ON(true);
-
 	vmw_overlay_resume_all(dev_priv);
 
+	WARN_ON(ret != 0);
+
 	return 0;
 }
 
@@ -668,7 +687,7 @@ int vmw_kms_new_framebuffer_dmabuf(struct vmw_private *dev_priv,
 
 	/* XXX get the first 3 from the surface info */
 	vfbd->base.base.bits_per_pixel = 32;
-	vfbd->base.base.pitch = width * 32 / 4;
+	vfbd->base.base.pitch = width * vfbd->base.base.bits_per_pixel / 8;
 	vfbd->base.base.depth = 24;
 	vfbd->base.base.width = width;
 	vfbd->base.base.height = height;
@@ -765,8 +784,9 @@ int vmw_kms_init(struct vmw_private *dev_priv)
 	dev->mode_config.funcs = &vmw_kms_funcs;
 	dev->mode_config.min_width = 1;
 	dev->mode_config.min_height = 1;
-	dev->mode_config.max_width = dev_priv->fb_max_width;
-	dev->mode_config.max_height = dev_priv->fb_max_height;
+	/* assumed largest fb size */
+	dev->mode_config.max_width = 8192;
+	dev->mode_config.max_height = 8192;
 
 	ret = vmw_kms_init_legacy_display_system(dev_priv);
 
@@ -826,24 +846,25 @@ out:
 	return ret;
 }
 
-int vmw_kms_save_vga(struct vmw_private *vmw_priv)
+void vmw_kms_write_svga(struct vmw_private *vmw_priv,
+			unsigned width, unsigned height, unsigned pitch,
+			unsigned bbp, unsigned depth)
 {
-	/*
-	 * setup a single multimon monitor with the size
-	 * of 0x0, this stops the UI from resizing when we
-	 * change the framebuffer size
-	 */
-	if (vmw_priv->capabilities & SVGA_CAP_MULTIMON) {
-		vmw_write(vmw_priv, SVGA_REG_NUM_GUEST_DISPLAYS, 1);
-		vmw_write(vmw_priv, SVGA_REG_DISPLAY_ID, 0);
-		vmw_write(vmw_priv, SVGA_REG_DISPLAY_IS_PRIMARY, true);
-		vmw_write(vmw_priv, SVGA_REG_DISPLAY_POSITION_X, 0);
-		vmw_write(vmw_priv, SVGA_REG_DISPLAY_POSITION_Y, 0);
-		vmw_write(vmw_priv, SVGA_REG_DISPLAY_WIDTH, 0);
-		vmw_write(vmw_priv, SVGA_REG_DISPLAY_HEIGHT, 0);
-		vmw_write(vmw_priv, SVGA_REG_DISPLAY_ID, SVGA_ID_INVALID);
-	}
+	if (vmw_priv->capabilities & SVGA_CAP_PITCHLOCK)
+		vmw_write(vmw_priv, SVGA_REG_PITCHLOCK, pitch);
+	else if (vmw_fifo_have_pitchlock(vmw_priv))
+		iowrite32(pitch, vmw_priv->mmio_virt + SVGA_FIFO_PITCHLOCK);
+	vmw_write(vmw_priv, SVGA_REG_WIDTH, width);
+	vmw_write(vmw_priv, SVGA_REG_HEIGHT, height);
+	vmw_write(vmw_priv, SVGA_REG_BITS_PER_PIXEL, bbp);
+	vmw_write(vmw_priv, SVGA_REG_DEPTH, depth);
+	vmw_write(vmw_priv, SVGA_REG_RED_MASK, 0x00ff0000);
+	vmw_write(vmw_priv, SVGA_REG_GREEN_MASK, 0x0000ff00);
+	vmw_write(vmw_priv, SVGA_REG_BLUE_MASK, 0x000000ff);
+}
 
+int vmw_kms_save_vga(struct vmw_private *vmw_priv)
+{
 	vmw_priv->vga_width = vmw_read(vmw_priv, SVGA_REG_WIDTH);
 	vmw_priv->vga_height = vmw_read(vmw_priv, SVGA_REG_HEIGHT);
 	vmw_priv->vga_bpp = vmw_read(vmw_priv, SVGA_REG_BITS_PER_PIXEL);
@@ -852,6 +873,12 @@ int vmw_kms_save_vga(struct vmw_private *vmw_priv)
 	vmw_priv->vga_red_mask = vmw_read(vmw_priv, SVGA_REG_RED_MASK);
 	vmw_priv->vga_green_mask = vmw_read(vmw_priv, SVGA_REG_GREEN_MASK);
 	vmw_priv->vga_blue_mask = vmw_read(vmw_priv, SVGA_REG_BLUE_MASK);
+	if (vmw_priv->capabilities & SVGA_CAP_PITCHLOCK)
+		vmw_priv->vga_pitchlock =
+			vmw_read(vmw_priv, SVGA_REG_PITCHLOCK);
+	else if (vmw_fifo_have_pitchlock(vmw_priv))
+		vmw_priv->vga_pitchlock =
+			ioread32(vmw_priv->mmio_virt + SVGA_FIFO_PITCHLOCK);
 
 	return 0;
 }
@@ -866,9 +893,12 @@ int vmw_kms_restore_vga(struct vmw_private *vmw_priv)
 	vmw_write(vmw_priv, SVGA_REG_RED_MASK, vmw_priv->vga_red_mask);
 	vmw_write(vmw_priv, SVGA_REG_GREEN_MASK, vmw_priv->vga_green_mask);
 	vmw_write(vmw_priv, SVGA_REG_BLUE_MASK, vmw_priv->vga_blue_mask);
-
-	/* TODO check for multimon */
-	vmw_write(vmw_priv, SVGA_REG_NUM_GUEST_DISPLAYS, 0);
+	if (vmw_priv->capabilities & SVGA_CAP_PITCHLOCK)
+		vmw_write(vmw_priv, SVGA_REG_PITCHLOCK,
+			  vmw_priv->vga_pitchlock);
+	else if (vmw_fifo_have_pitchlock(vmw_priv))
+		iowrite32(vmw_priv->vga_pitchlock,
+			  vmw_priv->mmio_virt + SVGA_FIFO_PITCHLOCK);
 
 	return 0;
 }
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_ldu.c b/drivers/gpu/drm/vmwgfx/vmwgfx_ldu.c
index 90891593bf6c..f7094dde18f9 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_ldu.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_ldu.c
@@ -38,6 +38,7 @@ struct vmw_legacy_display {
 	struct list_head active;
 
 	unsigned num_active;
+	unsigned last_num_active;
 
 	struct vmw_framebuffer *fb;
 };
@@ -49,8 +50,6 @@ struct vmw_legacy_display_unit {
 	struct vmw_display_unit base;
 
 	struct list_head active;
-
-	unsigned unit;
 };
 
 static void vmw_ldu_destroy(struct vmw_legacy_display_unit *ldu)
@@ -88,23 +87,44 @@ static int vmw_ldu_commit_list(struct vmw_private *dev_priv)
 {
 	struct vmw_legacy_display *lds = dev_priv->ldu_priv;
 	struct vmw_legacy_display_unit *entry;
-	struct drm_crtc *crtc;
+	struct drm_framebuffer *fb = NULL;
+	struct drm_crtc *crtc = NULL;
 	int i = 0;
 
-	/* to stop the screen from changing size on resize */
-	vmw_write(dev_priv, SVGA_REG_NUM_GUEST_DISPLAYS, 0);
-	for (i = 0; i < lds->num_active; i++) {
-		vmw_write(dev_priv, SVGA_REG_DISPLAY_ID, i);
-		vmw_write(dev_priv, SVGA_REG_DISPLAY_IS_PRIMARY, !i);
-		vmw_write(dev_priv, SVGA_REG_DISPLAY_POSITION_X, 0);
-		vmw_write(dev_priv, SVGA_REG_DISPLAY_POSITION_Y, 0);
-		vmw_write(dev_priv, SVGA_REG_DISPLAY_WIDTH, 0);
-		vmw_write(dev_priv, SVGA_REG_DISPLAY_HEIGHT, 0);
-		vmw_write(dev_priv, SVGA_REG_DISPLAY_ID, SVGA_ID_INVALID);
+	/* If there is no display topology the host just assumes
+	 * that the guest will set the same layout as the host.
+	 */
+	if (!(dev_priv->capabilities & SVGA_CAP_DISPLAY_TOPOLOGY)) {
+		int w = 0, h = 0;
+		list_for_each_entry(entry, &lds->active, active) {
+			crtc = &entry->base.crtc;
+			w = max(w, crtc->x + crtc->mode.hdisplay);
+			h = max(h, crtc->y + crtc->mode.vdisplay);
+			i++;
+		}
+
+		if (crtc == NULL)
+			return 0;
+		fb = entry->base.crtc.fb;
+
+		vmw_kms_write_svga(dev_priv, w, h, fb->pitch,
+				   fb->bits_per_pixel, fb->depth);
+
+		return 0;
+	}
+
+	if (!list_empty(&lds->active)) {
+		entry = list_entry(lds->active.next, typeof(*entry), active);
+		fb = entry->base.crtc.fb;
+
+		vmw_kms_write_svga(dev_priv, fb->width, fb->height, fb->pitch,
+				   fb->bits_per_pixel, fb->depth);
 	}
 
-	/* Now set the mode */
-	vmw_write(dev_priv, SVGA_REG_NUM_GUEST_DISPLAYS, lds->num_active);
+	/* Make sure we always show something. */
+	vmw_write(dev_priv, SVGA_REG_NUM_GUEST_DISPLAYS,
+		  lds->num_active ? lds->num_active : 1);
+
 	i = 0;
 	list_for_each_entry(entry, &lds->active, active) {
 		crtc = &entry->base.crtc;
@@ -120,6 +140,10 @@ static int vmw_ldu_commit_list(struct vmw_private *dev_priv)
 		i++;
 	}
 
+	BUG_ON(i != lds->num_active);
+
+	lds->last_num_active = lds->num_active;
+
 	return 0;
 }
 
@@ -130,6 +154,7 @@ static int vmw_ldu_del_active(struct vmw_private *vmw_priv,
 	if (list_empty(&ldu->active))
 		return 0;
 
+	/* Must init otherwise list_empty(&ldu->active) will not work. */
 	list_del_init(&ldu->active);
 	if (--(ld->num_active) == 0) {
 		BUG_ON(!ld->fb);
@@ -149,24 +174,29 @@ static int vmw_ldu_add_active(struct vmw_private *vmw_priv,
 	struct vmw_legacy_display_unit *entry;
 	struct list_head *at;
 
+	BUG_ON(!ld->num_active && ld->fb);
+	if (vfb != ld->fb) {
+		if (ld->fb && ld->fb->unpin)
+			ld->fb->unpin(ld->fb);
+		if (vfb->pin)
+			vfb->pin(vfb);
+		ld->fb = vfb;
+	}
+
 	if (!list_empty(&ldu->active))
 		return 0;
 
 	at = &ld->active;
 	list_for_each_entry(entry, &ld->active, active) {
-		if (entry->unit > ldu->unit)
+		if (entry->base.unit > ldu->base.unit)
 			break;
 
 		at = &entry->active;
 	}
 
 	list_add(&ldu->active, at);
-	if (ld->num_active++ == 0) {
-		BUG_ON(ld->fb);
-		if (vfb->pin)
-			vfb->pin(vfb);
-		ld->fb = vfb;
-	}
+
+	ld->num_active++;
 
 	return 0;
 }
@@ -208,6 +238,8 @@ static int vmw_ldu_crtc_set_config(struct drm_mode_set *set)
 
 	/* ldu only supports one fb active at the time */
 	if (dev_priv->ldu_priv->fb && vfb &&
+	    !(dev_priv->ldu_priv->num_active == 1 &&
+	      !list_empty(&ldu->active)) &&
 	    dev_priv->ldu_priv->fb != vfb) {
 		DRM_ERROR("Multiple framebuffers not supported\n");
 		return -EINVAL;
@@ -443,18 +475,16 @@ static int vmw_ldu_init(struct vmw_private *dev_priv, unsigned unit)
 	if (!ldu)
 		return -ENOMEM;
 
-	ldu->unit = unit;
+	ldu->base.unit = unit;
 	crtc = &ldu->base.crtc;
 	encoder = &ldu->base.encoder;
 	connector = &ldu->base.connector;
 
+	INIT_LIST_HEAD(&ldu->active);
+
 	drm_connector_init(dev, connector, &vmw_legacy_connector_funcs,
 			   DRM_MODE_CONNECTOR_LVDS);
-	/* Initial status */
-	if (unit == 0)
-		connector->status = connector_status_connected;
-	else
-		connector->status = connector_status_disconnected;
+	connector->status = vmw_ldu_connector_detect(connector);
 
 	drm_encoder_init(dev, encoder, &vmw_legacy_encoder_funcs,
 			 DRM_MODE_ENCODER_LVDS);
@@ -462,8 +492,6 @@ static int vmw_ldu_init(struct vmw_private *dev_priv, unsigned unit)
 	encoder->possible_crtcs = (1 << unit);
 	encoder->possible_clones = 0;
 
-	INIT_LIST_HEAD(&ldu->active);
-
 	drm_crtc_init(dev, crtc, &vmw_legacy_crtc_funcs);
 
 	drm_connector_attach_property(connector,
@@ -487,18 +515,22 @@ int vmw_kms_init_legacy_display_system(struct vmw_private *dev_priv)
 
 	INIT_LIST_HEAD(&dev_priv->ldu_priv->active);
 	dev_priv->ldu_priv->num_active = 0;
+	dev_priv->ldu_priv->last_num_active = 0;
 	dev_priv->ldu_priv->fb = NULL;
 
 	drm_mode_create_dirty_info_property(dev_priv->dev);
 
 	vmw_ldu_init(dev_priv, 0);
-	vmw_ldu_init(dev_priv, 1);
-	vmw_ldu_init(dev_priv, 2);
-	vmw_ldu_init(dev_priv, 3);
-	vmw_ldu_init(dev_priv, 4);
-	vmw_ldu_init(dev_priv, 5);
-	vmw_ldu_init(dev_priv, 6);
-	vmw_ldu_init(dev_priv, 7);
+	/* for old hardware without multimon only enable one display */
+	if (dev_priv->capabilities & SVGA_CAP_MULTIMON) {
+		vmw_ldu_init(dev_priv, 1);
+		vmw_ldu_init(dev_priv, 2);
+		vmw_ldu_init(dev_priv, 3);
+		vmw_ldu_init(dev_priv, 4);
+		vmw_ldu_init(dev_priv, 5);
+		vmw_ldu_init(dev_priv, 6);
+		vmw_ldu_init(dev_priv, 7);
+	}
 
 	return 0;
 }
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_overlay.c b/drivers/gpu/drm/vmwgfx/vmwgfx_overlay.c
index ad566c85b075..df2036ed18d5 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_overlay.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_overlay.c
@@ -358,6 +358,8 @@ static int vmw_overlay_update_stream(struct vmw_private *dev_priv,
 	if (stream->buf != buf)
 		stream->buf = vmw_dmabuf_reference(buf);
 	stream->saved = *arg;
+	/* stream is no longer stopped/paused */
+	stream->paused = false;
 
 	return 0;
 }