Merge branch 'perf/core' into perf/urgent

Merge reason: Switch from pre-merge topical split to the post-merge urgent track

Signed-off-by: Ingo Molnar <mingo@elte.hu>

9 files changed, 140 insertions, 91 deletions

diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c
index a6e8f687fa6..0c9c0811f42 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c
@@ -348,22 +348,19 @@ static int vmw_driver_load(struct drm_device *dev, unsigned long chipset)
 		 */
 
 		DRM_INFO("It appears like vesafb is loaded. "
-			 "Ignore above error if any. Entering stealth mode.\n");
+			 "Ignore above error if any.\n");
 		ret = pci_request_region(dev->pdev, 2, "vmwgfx stealth probe");
 		if (unlikely(ret != 0)) {
 			DRM_ERROR("Failed reserving the SVGA MMIO resource.\n");
 			goto out_no_device;
 		}
-		vmw_kms_init(dev_priv);
-		vmw_overlay_init(dev_priv);
-	} else {
-		ret = vmw_request_device(dev_priv);
-		if (unlikely(ret != 0))
-			goto out_no_device;
-		vmw_kms_init(dev_priv);
-		vmw_overlay_init(dev_priv);
-		vmw_fb_init(dev_priv);
 	}
+	ret = vmw_request_device(dev_priv);
+	if (unlikely(ret != 0))
+		goto out_no_device;
+	vmw_kms_init(dev_priv);
+	vmw_overlay_init(dev_priv);
+	vmw_fb_init(dev_priv);
 
 	dev_priv->pm_nb.notifier_call = vmwgfx_pm_notifier;
 	register_pm_notifier(&dev_priv->pm_nb);
@@ -406,17 +403,15 @@ static int vmw_driver_unload(struct drm_device *dev)
 
 	unregister_pm_notifier(&dev_priv->pm_nb);
 
-	if (!dev_priv->stealth) {
-		vmw_fb_close(dev_priv);
-		vmw_kms_close(dev_priv);
-		vmw_overlay_close(dev_priv);
-		vmw_release_device(dev_priv);
-		pci_release_regions(dev->pdev);
-	} else {
-		vmw_kms_close(dev_priv);
-		vmw_overlay_close(dev_priv);
+	vmw_fb_close(dev_priv);
+	vmw_kms_close(dev_priv);
+	vmw_overlay_close(dev_priv);
+	vmw_release_device(dev_priv);
+	if (dev_priv->stealth)
 		pci_release_region(dev->pdev, 2);
-	}
+	else
+		pci_release_regions(dev->pdev);
+
 	if (dev_priv->capabilities & SVGA_CAP_IRQMASK)
 		drm_irq_uninstall(dev_priv->dev);
 	if (dev->devname == vmw_devname)
@@ -585,11 +580,6 @@ static int vmw_master_set(struct drm_device *dev,
 	int ret = 0;
 
 	DRM_INFO("Master set.\n");
-	if (dev_priv->stealth) {
-		ret = vmw_request_device(dev_priv);
-		if (unlikely(ret != 0))
-			return ret;
-	}
 
 	if (active) {
 		BUG_ON(active != &dev_priv->fbdev_master);
@@ -649,18 +639,11 @@ static void vmw_master_drop(struct drm_device *dev,
 
 	ttm_lock_set_kill(&vmaster->lock, true, SIGTERM);
 
-	if (dev_priv->stealth) {
-		ret = ttm_bo_evict_mm(&dev_priv->bdev, TTM_PL_VRAM);
-		if (unlikely(ret != 0))
-			DRM_ERROR("Unable to clean VRAM on master drop.\n");
-		vmw_release_device(dev_priv);
-	}
 	dev_priv->active_master = &dev_priv->fbdev_master;
 	ttm_lock_set_kill(&dev_priv->fbdev_master.lock, false, SIGTERM);
 	ttm_vt_unlock(&dev_priv->fbdev_master.lock);
 
-	if (!dev_priv->stealth)
-		vmw_fb_on(dev_priv);
+	vmw_fb_on(dev_priv);
 }
 
 
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h
index 135be9688c9..356dc935ec1 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h
@@ -39,10 +39,10 @@
 #include "ttm/ttm_execbuf_util.h"
 #include "ttm/ttm_module.h"
 
-#define VMWGFX_DRIVER_DATE "20090724"
-#define VMWGFX_DRIVER_MAJOR 0
-#define VMWGFX_DRIVER_MINOR 1
-#define VMWGFX_DRIVER_PATCHLEVEL 2
+#define VMWGFX_DRIVER_DATE "20100209"
+#define VMWGFX_DRIVER_MAJOR 1
+#define VMWGFX_DRIVER_MINOR 0
+#define VMWGFX_DRIVER_PATCHLEVEL 0
 #define VMWGFX_FILE_PAGE_OFFSET 0x00100000
 #define VMWGFX_FIFO_STATIC_SIZE (1024*1024)
 #define VMWGFX_MAX_RELOCATIONS 2048
@@ -113,6 +113,7 @@ struct vmw_fifo_state {
 	unsigned long static_buffer_size;
 	bool using_bounce_buffer;
 	uint32_t capabilities;
+	struct mutex fifo_mutex;
 	struct rw_semaphore rwsem;
 };
 
@@ -213,7 +214,7 @@ struct vmw_private {
 	 * Fencing and IRQs.
 	 */
 
-	uint32_t fence_seq;
+	atomic_t fence_seq;
 	wait_queue_head_t fence_queue;
 	wait_queue_head_t fifo_queue;
 	atomic_t fence_queue_waiters;
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_execbuf.c b/drivers/gpu/drm/vmwgfx/vmwgfx_execbuf.c
index d69caf92ffe..0897359b3e4 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_execbuf.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_execbuf.c
@@ -182,25 +182,19 @@ static int vmw_cmd_present_check(struct vmw_private *dev_priv,
 	return vmw_cmd_sid_check(dev_priv, sw_context, &cmd->body.sid);
 }
 
-static int vmw_cmd_dma(struct vmw_private *dev_priv,
-		       struct vmw_sw_context *sw_context,
-		       SVGA3dCmdHeader *header)
+static int vmw_translate_guest_ptr(struct vmw_private *dev_priv,
+				   struct vmw_sw_context *sw_context,
+				   SVGAGuestPtr *ptr,
+				   struct vmw_dma_buffer **vmw_bo_p)
 {
-	uint32_t handle;
 	struct vmw_dma_buffer *vmw_bo = NULL;
 	struct ttm_buffer_object *bo;
-	struct vmw_surface *srf = NULL;
-	struct vmw_dma_cmd {
-		SVGA3dCmdHeader header;
-		SVGA3dCmdSurfaceDMA dma;
-	} *cmd;
+	uint32_t handle = ptr->gmrId;
 	struct vmw_relocation *reloc;
-	int ret;
 	uint32_t cur_validate_node;
 	struct ttm_validate_buffer *val_buf;
+	int ret;
 
-	cmd = container_of(header, struct vmw_dma_cmd, header);
-	handle = cmd->dma.guest.ptr.gmrId;
 	ret = vmw_user_dmabuf_lookup(sw_context->tfile, handle, &vmw_bo);
 	if (unlikely(ret != 0)) {
 		DRM_ERROR("Could not find or use GMR region.\n");
@@ -209,14 +203,14 @@ static int vmw_cmd_dma(struct vmw_private *dev_priv,
 	bo = &vmw_bo->base;
 
 	if (unlikely(sw_context->cur_reloc >= VMWGFX_MAX_RELOCATIONS)) {
-		DRM_ERROR("Max number of DMA commands per submission"
+		DRM_ERROR("Max number relocations per submission"
 			  " exceeded\n");
 		ret = -EINVAL;
 		goto out_no_reloc;
 	}
 
 	reloc = &sw_context->relocs[sw_context->cur_reloc++];
-	reloc->location = &cmd->dma.guest.ptr;
+	reloc->location = ptr;
 
 	cur_validate_node = vmw_dmabuf_validate_node(bo, sw_context->cur_val_buf);
 	if (unlikely(cur_validate_node >= VMWGFX_MAX_GMRS)) {
@@ -234,7 +228,89 @@ static int vmw_cmd_dma(struct vmw_private *dev_priv,
 		list_add_tail(&val_buf->head, &sw_context->validate_nodes);
 		++sw_context->cur_val_buf;
 	}
+	*vmw_bo_p = vmw_bo;
+	return 0;
+
+out_no_reloc:
+	vmw_dmabuf_unreference(&vmw_bo);
+	vmw_bo_p = NULL;
+	return ret;
+}
+
+static int vmw_cmd_end_query(struct vmw_private *dev_priv,
+			     struct vmw_sw_context *sw_context,
+			     SVGA3dCmdHeader *header)
+{
+	struct vmw_dma_buffer *vmw_bo;
+	struct vmw_query_cmd {
+		SVGA3dCmdHeader header;
+		SVGA3dCmdEndQuery q;
+	} *cmd;
+	int ret;
+
+	cmd = container_of(header, struct vmw_query_cmd, header);
+	ret = vmw_cmd_cid_check(dev_priv, sw_context, header);
+	if (unlikely(ret != 0))
+		return ret;
+
+	ret = vmw_translate_guest_ptr(dev_priv, sw_context,
+				      &cmd->q.guestResult,
+				      &vmw_bo);
+	if (unlikely(ret != 0))
+		return ret;
+
+	vmw_dmabuf_unreference(&vmw_bo);
+	return 0;
+}
 
+static int vmw_cmd_wait_query(struct vmw_private *dev_priv,
+			      struct vmw_sw_context *sw_context,
+			      SVGA3dCmdHeader *header)
+{
+	struct vmw_dma_buffer *vmw_bo;
+	struct vmw_query_cmd {
+		SVGA3dCmdHeader header;
+		SVGA3dCmdWaitForQuery q;
+	} *cmd;
+	int ret;
+
+	cmd = container_of(header, struct vmw_query_cmd, header);
+	ret = vmw_cmd_cid_check(dev_priv, sw_context, header);
+	if (unlikely(ret != 0))
+		return ret;
+
+	ret = vmw_translate_guest_ptr(dev_priv, sw_context,
+				      &cmd->q.guestResult,
+				      &vmw_bo);
+	if (unlikely(ret != 0))
+		return ret;
+
+	vmw_dmabuf_unreference(&vmw_bo);
+	return 0;
+}
+
+
+static int vmw_cmd_dma(struct vmw_private *dev_priv,
+		       struct vmw_sw_context *sw_context,
+		       SVGA3dCmdHeader *header)
+{
+	struct vmw_dma_buffer *vmw_bo = NULL;
+	struct ttm_buffer_object *bo;
+	struct vmw_surface *srf = NULL;
+	struct vmw_dma_cmd {
+		SVGA3dCmdHeader header;
+		SVGA3dCmdSurfaceDMA dma;
+	} *cmd;
+	int ret;
+
+	cmd = container_of(header, struct vmw_dma_cmd, header);
+	ret = vmw_translate_guest_ptr(dev_priv, sw_context,
+				      &cmd->dma.guest.ptr,
+				      &vmw_bo);
+	if (unlikely(ret != 0))
+		return ret;
+
+	bo = &vmw_bo->base;
 	ret = vmw_user_surface_lookup_handle(dev_priv, sw_context->tfile,
 					     cmd->dma.host.sid, &srf);
 	if (ret) {
@@ -379,8 +455,8 @@ static vmw_cmd_func vmw_cmd_funcs[SVGA_3D_CMD_MAX] = {
 	VMW_CMD_DEF(SVGA_3D_CMD_DRAW_PRIMITIVES, &vmw_cmd_draw),
 	VMW_CMD_DEF(SVGA_3D_CMD_SETSCISSORRECT, &vmw_cmd_cid_check),
 	VMW_CMD_DEF(SVGA_3D_CMD_BEGIN_QUERY, &vmw_cmd_cid_check),
-	VMW_CMD_DEF(SVGA_3D_CMD_END_QUERY, &vmw_cmd_cid_check),
-	VMW_CMD_DEF(SVGA_3D_CMD_WAIT_FOR_QUERY, &vmw_cmd_cid_check),
+	VMW_CMD_DEF(SVGA_3D_CMD_END_QUERY, &vmw_cmd_end_query),
+	VMW_CMD_DEF(SVGA_3D_CMD_WAIT_FOR_QUERY, &vmw_cmd_wait_query),
 	VMW_CMD_DEF(SVGA_3D_CMD_PRESENT_READBACK, &vmw_cmd_ok),
 	VMW_CMD_DEF(SVGA_3D_CMD_BLIT_SURFACE_TO_SCREEN,
 		    &vmw_cmd_blt_surf_screen_check)
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_fb.c b/drivers/gpu/drm/vmwgfx/vmwgfx_fb.c
index 4f4f6432be8..a93367041cd 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_fb.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_fb.c
@@ -559,6 +559,9 @@ int vmw_fb_init(struct vmw_private *vmw_priv)
 	info->pixmap.scan_align = 1;
 #endif
 
+	info->aperture_base = vmw_priv->vram_start;
+	info->aperture_size = vmw_priv->vram_size;
+
 	/*
 	 * Dirty & Deferred IO
 	 */
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_fifo.c b/drivers/gpu/drm/vmwgfx/vmwgfx_fifo.c
index 4157547cc6e..39d43a01d84 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_fifo.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_fifo.c
@@ -74,6 +74,7 @@ int vmw_fifo_init(struct vmw_private *dev_priv, struct vmw_fifo_state *fifo)
 	fifo->reserved_size = 0;
 	fifo->using_bounce_buffer = false;
 
+	mutex_init(&fifo->fifo_mutex);
 	init_rwsem(&fifo->rwsem);
 
 	/*
@@ -117,7 +118,7 @@ int vmw_fifo_init(struct vmw_private *dev_priv, struct vmw_fifo_state *fifo)
 		 (unsigned int) min,
 		 (unsigned int) fifo->capabilities);
 
-	dev_priv->fence_seq = dev_priv->last_read_sequence;
+	atomic_set(&dev_priv->fence_seq, dev_priv->last_read_sequence);
 	iowrite32(dev_priv->last_read_sequence, fifo_mem + SVGA_FIFO_FENCE);
 
 	return vmw_fifo_send_fence(dev_priv, &dummy);
@@ -283,7 +284,7 @@ void *vmw_fifo_reserve(struct vmw_private *dev_priv, uint32_t bytes)
 	uint32_t reserveable = fifo_state->capabilities & SVGA_FIFO_CAP_RESERVE;
 	int ret;
 
-	down_write(&fifo_state->rwsem);
+	mutex_lock(&fifo_state->fifo_mutex);
 	max = ioread32(fifo_mem + SVGA_FIFO_MAX);
 	min = ioread32(fifo_mem + SVGA_FIFO_MIN);
 	next_cmd = ioread32(fifo_mem + SVGA_FIFO_NEXT_CMD);
@@ -351,7 +352,7 @@ void *vmw_fifo_reserve(struct vmw_private *dev_priv, uint32_t bytes)
 	}
 out_err:
 	fifo_state->reserved_size = 0;
-	up_write(&fifo_state->rwsem);
+	mutex_unlock(&fifo_state->fifo_mutex);
 	return NULL;
 }
 
@@ -426,6 +427,7 @@ void vmw_fifo_commit(struct vmw_private *dev_priv, uint32_t bytes)
 
 	}
 
+	down_write(&fifo_state->rwsem);
 	if (fifo_state->using_bounce_buffer || reserveable) {
 		next_cmd += bytes;
 		if (next_cmd >= max)
@@ -437,8 +439,9 @@ void vmw_fifo_commit(struct vmw_private *dev_priv, uint32_t bytes)
 	if (reserveable)
 		iowrite32(0, fifo_mem + SVGA_FIFO_RESERVED);
 	mb();
-	vmw_fifo_ping_host(dev_priv, SVGA_SYNC_GENERIC);
 	up_write(&fifo_state->rwsem);
+	vmw_fifo_ping_host(dev_priv, SVGA_SYNC_GENERIC);
+	mutex_unlock(&fifo_state->fifo_mutex);
 }
 
 int vmw_fifo_send_fence(struct vmw_private *dev_priv, uint32_t *sequence)
@@ -451,9 +454,7 @@ int vmw_fifo_send_fence(struct vmw_private *dev_priv, uint32_t *sequence)
 
 	fm = vmw_fifo_reserve(dev_priv, bytes);
 	if (unlikely(fm == NULL)) {
-		down_write(&fifo_state->rwsem);
-		*sequence = dev_priv->fence_seq;
-		up_write(&fifo_state->rwsem);
+		*sequence = atomic_read(&dev_priv->fence_seq);
 		ret = -ENOMEM;
 		(void)vmw_fallback_wait(dev_priv, false, true, *sequence,
 					false, 3*HZ);
@@ -461,7 +462,7 @@ int vmw_fifo_send_fence(struct vmw_private *dev_priv, uint32_t *sequence)
 	}
 
 	do {
-		*sequence = dev_priv->fence_seq++;
+		*sequence = atomic_add_return(1, &dev_priv->fence_seq);
 	} while (*sequence == 0);
 
 	if (!(fifo_state->capabilities & SVGA_FIFO_CAP_FENCE)) {
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_ioctl.c b/drivers/gpu/drm/vmwgfx/vmwgfx_ioctl.c
index 778851f9f1d..1c7a316454d 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_ioctl.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_ioctl.c
@@ -48,6 +48,12 @@ int vmw_getparam_ioctl(struct drm_device *dev, void *data,
 	case DRM_VMW_PARAM_FIFO_OFFSET:
 		param->value = dev_priv->mmio_start;
 		break;
+	case DRM_VMW_PARAM_HW_CAPS:
+		param->value = dev_priv->capabilities;
+		break;
+	case DRM_VMW_PARAM_FIFO_CAPS:
+		param->value = dev_priv->fifo.capabilities;
+		break;
 	default:
 		DRM_ERROR("Illegal vmwgfx get param request: %d\n",
 			  param->param);
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_irq.c b/drivers/gpu/drm/vmwgfx/vmwgfx_irq.c
index d40086fc864..4d7cb539386 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_irq.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_irq.c
@@ -85,19 +85,12 @@ bool vmw_fence_signaled(struct vmw_private *dev_priv,
 		return true;
 
 	/**
-	 * Below is to signal stale fences that have wrapped.
-	 * First, block fence submission.
-	 */
-
-	down_read(&fifo_state->rwsem);
-
-	/**
 	 * Then check if the sequence is higher than what we've actually
 	 * emitted. Then the fence is stale and signaled.
 	 */
 
-	ret = ((dev_priv->fence_seq - sequence) > VMW_FENCE_WRAP);
-	up_read(&fifo_state->rwsem);
+	ret = ((atomic_read(&dev_priv->fence_seq) - sequence)
+	       > VMW_FENCE_WRAP);
 
 	return ret;
 }
@@ -127,7 +120,7 @@ int vmw_fallback_wait(struct vmw_private *dev_priv,
 
 	if (fifo_idle)
 		down_read(&fifo_state->rwsem);
-	signal_seq = dev_priv->fence_seq;
+	signal_seq = atomic_read(&dev_priv->fence_seq);
 	ret = 0;
 
 	for (;;) {
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_kms.c b/drivers/gpu/drm/vmwgfx/vmwgfx_kms.c
index eeba6d1d06e..31f9afed0a6 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_kms.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_kms.c
@@ -769,10 +769,10 @@ int vmw_kms_init(struct vmw_private *dev_priv)
 
 	drm_mode_config_init(dev);
 	dev->mode_config.funcs = &vmw_kms_funcs;
-	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.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;
 
 	ret = vmw_kms_init_legacy_display_system(dev_priv);
 
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_resource.c b/drivers/gpu/drm/vmwgfx/vmwgfx_resource.c
index c7efbd47ab8..f8fbbc67a40 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_resource.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_resource.c
@@ -35,11 +35,6 @@
 #define VMW_RES_SURFACE ttm_driver_type1
 #define VMW_RES_STREAM ttm_driver_type2
 
-/* XXX: This isn't a real hardware flag, but just a hack for kernel to
- * know about primary surfaces. Find a better way to accomplish this.
- */
-#define SVGA3D_SURFACE_HINT_SCANOUT (1 << 9)
-
 struct vmw_user_context {
 	struct ttm_base_object base;
 	struct vmw_resource res;
@@ -579,6 +574,7 @@ int vmw_surface_define_ioctl(struct drm_device *dev, void *data,
 
 	srf->flags = req->flags;
 	srf->format = req->format;
+	srf->scanout = req->scanout;
 	memcpy(srf->mip_levels, req->mip_levels, sizeof(srf->mip_levels));
 	srf->num_sizes = 0;
 	for (i = 0; i < DRM_VMW_MAX_SURFACE_FACES; ++i)
@@ -604,16 +600,6 @@ int vmw_surface_define_ioctl(struct drm_device *dev, void *data,
 	if (unlikely(ret != 0))
 		goto out_err1;
 
-	if (srf->flags & SVGA3D_SURFACE_HINT_SCANOUT) {
-		/* we should not send this flag down to hardware since
-		 * its not a official one
-		 */
-		srf->flags &= ~SVGA3D_SURFACE_HINT_SCANOUT;
-		srf->scanout = true;
-	} else {
-		srf->scanout = false;
-	}
-
 	if (srf->scanout &&
 	    srf->num_sizes == 1 &&
 	    srf->sizes[0].width == 64 &&