drm/vmwgfx: Add DRM driver for VMware Virtual GPU

This commit adds the vmwgfx driver for the VWware Virtual GPU aka SVGA.
The driver is under staging the same as Nouveau and Radeon KMS. Hopefully
the 2D ioctls are bug free and don't need changing, so that part of the
API should be stable. But there there is a pretty big chance that the 3D API
will change in the future.

Signed-off-by: Thomas Hellström <thellstrom@vmware.com>
Signed-off-by: Jakob Bornecrantz <jakob@vmware.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>

1 files changed, 634 insertions, 0 deletions

diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_overlay.c b/drivers/gpu/drm/vmwgfx/vmwgfx_overlay.c
new file mode 100644
index 000000000000..bb6e6a096d25
--- /dev/null
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_overlay.c
@@ -0,0 +1,634 @@
+/**************************************************************************
+ *
+ * Copyright © 2009 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 "drmP.h"
+#include "vmwgfx_drv.h"
+
+#include "ttm/ttm_placement.h"
+
+#include "svga_overlay.h"
+#include "svga_escape.h"
+
+#define VMW_MAX_NUM_STREAMS 1
+
+struct vmw_stream {
+	struct vmw_dma_buffer *buf;
+	bool claimed;
+	bool paused;
+	struct drm_vmw_control_stream_arg saved;
+};
+
+/**
+ * Overlay control
+ */
+struct vmw_overlay {
+	/*
+	 * Each stream is a single overlay. In Xv these are called ports.
+	 */
+	struct mutex mutex;
+	struct vmw_stream stream[VMW_MAX_NUM_STREAMS];
+};
+
+static inline struct vmw_overlay *vmw_overlay(struct drm_device *dev)
+{
+	struct vmw_private *dev_priv = vmw_priv(dev);
+	return dev_priv ? dev_priv->overlay_priv : NULL;
+}
+
+struct vmw_escape_header {
+	uint32_t cmd;
+	SVGAFifoCmdEscape body;
+};
+
+struct vmw_escape_video_flush {
+	struct vmw_escape_header escape;
+	SVGAEscapeVideoFlush flush;
+};
+
+static inline void fill_escape(struct vmw_escape_header *header,
+			       uint32_t size)
+{
+	header->cmd = SVGA_CMD_ESCAPE;
+	header->body.nsid = SVGA_ESCAPE_NSID_VMWARE;
+	header->body.size = size;
+}
+
+static inline void fill_flush(struct vmw_escape_video_flush *cmd,
+			      uint32_t stream_id)
+{
+	fill_escape(&cmd->escape, sizeof(cmd->flush));
+	cmd->flush.cmdType = SVGA_ESCAPE_VMWARE_VIDEO_FLUSH;
+	cmd->flush.streamId = stream_id;
+}
+
+/**
+ * Pin or unpin a buffer in vram.
+ *
+ * @dev_priv:  Driver private.
+ * @buf:  DMA buffer to pin or unpin.
+ * @pin:  Pin buffer in vram if true.
+ * @interruptible:  Use interruptible wait.
+ *
+ * Takes the current masters ttm lock in read.
+ *
+ * Returns
+ * -ERESTARTSYS if interrupted by a signal.
+ */
+static int vmw_dmabuf_pin_in_vram(struct vmw_private *dev_priv,
+				  struct vmw_dma_buffer *buf,
+				  bool pin, bool interruptible)
+{
+	struct ttm_buffer_object *bo = &buf->base;
+	struct ttm_bo_global *glob = bo->glob;
+	struct ttm_placement *overlay_placement = &vmw_vram_placement;
+	int ret;
+
+	ret = ttm_read_lock(&dev_priv->active_master->lock, interruptible);
+	if (unlikely(ret != 0))
+		return ret;
+
+	ret = ttm_bo_reserve(bo, interruptible, false, false, 0);
+	if (unlikely(ret != 0))
+		goto err;
+
+	if (buf->gmr_bound) {
+		vmw_gmr_unbind(dev_priv, buf->gmr_id);
+		spin_lock(&glob->lru_lock);
+		ida_remove(&dev_priv->gmr_ida, buf->gmr_id);
+		spin_unlock(&glob->lru_lock);
+		buf->gmr_bound = NULL;
+	}
+
+	if (pin)
+		overlay_placement = &vmw_vram_ne_placement;
+
+	ret = ttm_bo_validate(bo, overlay_placement, interruptible, false);
+
+	ttm_bo_unreserve(bo);
+
+err:
+	ttm_read_unlock(&dev_priv->active_master->lock);
+
+	return ret;
+}
+
+/**
+ * Send put command to hw.
+ *
+ * Returns
+ * -ERESTARTSYS if interrupted by a signal.
+ */
+static int vmw_overlay_send_put(struct vmw_private *dev_priv,
+				struct vmw_dma_buffer *buf,
+				struct drm_vmw_control_stream_arg *arg,
+				bool interruptible)
+{
+	struct {
+		struct vmw_escape_header escape;
+		struct {
+			struct {
+				uint32_t cmdType;
+				uint32_t streamId;
+			} header;
+			struct {
+				uint32_t registerId;
+				uint32_t value;
+			} items[SVGA_VIDEO_PITCH_3 + 1];
+		} body;
+		struct vmw_escape_video_flush flush;
+	} *cmds;
+	uint32_t offset;
+	int i, ret;
+
+	for (;;) {
+		cmds = vmw_fifo_reserve(dev_priv, sizeof(*cmds));
+		if (cmds)
+			break;
+
+		ret = vmw_fallback_wait(dev_priv, false, true, 0,
+					interruptible, 3*HZ);
+		if (interruptible && ret == -ERESTARTSYS)
+			return ret;
+		else
+			BUG_ON(ret != 0);
+	}
+
+	fill_escape(&cmds->escape, sizeof(cmds->body));
+	cmds->body.header.cmdType = SVGA_ESCAPE_VMWARE_VIDEO_SET_REGS;
+	cmds->body.header.streamId = arg->stream_id;
+
+	for (i = 0; i <= SVGA_VIDEO_PITCH_3; i++)
+		cmds->body.items[i].registerId = i;
+
+	offset = buf->base.offset + arg->offset;
+
+	cmds->body.items[SVGA_VIDEO_ENABLED].value     = true;
+	cmds->body.items[SVGA_VIDEO_FLAGS].value       = arg->flags;
+	cmds->body.items[SVGA_VIDEO_DATA_OFFSET].value = offset;
+	cmds->body.items[SVGA_VIDEO_FORMAT].value      = arg->format;
+	cmds->body.items[SVGA_VIDEO_COLORKEY].value    = arg->color_key;
+	cmds->body.items[SVGA_VIDEO_SIZE].value        = arg->size;
+	cmds->body.items[SVGA_VIDEO_WIDTH].value       = arg->width;
+	cmds->body.items[SVGA_VIDEO_HEIGHT].value      = arg->height;
+	cmds->body.items[SVGA_VIDEO_SRC_X].value       = arg->src.x;
+	cmds->body.items[SVGA_VIDEO_SRC_Y].value       = arg->src.y;
+	cmds->body.items[SVGA_VIDEO_SRC_WIDTH].value   = arg->src.w;
+	cmds->body.items[SVGA_VIDEO_SRC_HEIGHT].value  = arg->src.h;
+	cmds->body.items[SVGA_VIDEO_DST_X].value       = arg->dst.x;
+	cmds->body.items[SVGA_VIDEO_DST_Y].value       = arg->dst.y;
+	cmds->body.items[SVGA_VIDEO_DST_WIDTH].value   = arg->dst.w;
+	cmds->body.items[SVGA_VIDEO_DST_HEIGHT].value  = arg->dst.h;
+	cmds->body.items[SVGA_VIDEO_PITCH_1].value     = arg->pitch[0];
+	cmds->body.items[SVGA_VIDEO_PITCH_2].value     = arg->pitch[1];
+	cmds->body.items[SVGA_VIDEO_PITCH_3].value     = arg->pitch[2];
+
+	fill_flush(&cmds->flush, arg->stream_id);
+
+	vmw_fifo_commit(dev_priv, sizeof(*cmds));
+
+	return 0;
+}
+
+/**
+ * Send stop command to hw.
+ *
+ * Returns
+ * -ERESTARTSYS if interrupted by a signal.
+ */
+static int vmw_overlay_send_stop(struct vmw_private *dev_priv,
+				 uint32_t stream_id,
+				 bool interruptible)
+{
+	struct {
+		struct vmw_escape_header escape;
+		SVGAEscapeVideoSetRegs body;
+		struct vmw_escape_video_flush flush;
+	} *cmds;
+	int ret;
+
+	for (;;) {
+		cmds = vmw_fifo_reserve(dev_priv, sizeof(*cmds));
+		if (cmds)
+			break;
+
+		ret = vmw_fallback_wait(dev_priv, false, true, 0,
+					interruptible, 3*HZ);
+		if (interruptible && ret == -ERESTARTSYS)
+			return ret;
+		else
+			BUG_ON(ret != 0);
+	}
+
+	fill_escape(&cmds->escape, sizeof(cmds->body));
+	cmds->body.header.cmdType = SVGA_ESCAPE_VMWARE_VIDEO_SET_REGS;
+	cmds->body.header.streamId = stream_id;
+	cmds->body.items[0].registerId = SVGA_VIDEO_ENABLED;
+	cmds->body.items[0].value = false;
+	fill_flush(&cmds->flush, stream_id);
+
+	vmw_fifo_commit(dev_priv, sizeof(*cmds));
+
+	return 0;
+}
+
+/**
+ * Stop or pause a stream.
+ *
+ * If the stream is paused the no evict flag is removed from the buffer
+ * but left in vram. This allows for instance mode_set to evict it
+ * should it need to.
+ *
+ * The caller must hold the overlay lock.
+ *
+ * @stream_id which stream to stop/pause.
+ * @pause true to pause, false to stop completely.
+ */
+static int vmw_overlay_stop(struct vmw_private *dev_priv,
+			    uint32_t stream_id, bool pause,
+			    bool interruptible)
+{
+	struct vmw_overlay *overlay = dev_priv->overlay_priv;
+	struct vmw_stream *stream = &overlay->stream[stream_id];
+	int ret;
+
+	/* no buffer attached the stream is completely stopped */
+	if (!stream->buf)
+		return 0;
+
+	/* If the stream is paused this is already done */
+	if (!stream->paused) {
+		ret = vmw_overlay_send_stop(dev_priv, stream_id,
+					    interruptible);
+		if (ret)
+			return ret;
+
+		/* We just remove the NO_EVICT flag so no -ENOMEM */
+		ret = vmw_dmabuf_pin_in_vram(dev_priv, stream->buf, false,
+					     interruptible);
+		if (interruptible && ret == -ERESTARTSYS)
+			return ret;
+		else
+			BUG_ON(ret != 0);
+	}
+
+	if (!pause) {
+		vmw_dmabuf_unreference(&stream->buf);
+		stream->paused = false;
+	} else {
+		stream->paused = true;
+	}
+
+	return 0;
+}
+
+/**
+ * Update a stream and send any put or stop fifo commands needed.
+ *
+ * The caller must hold the overlay lock.
+ *
+ * Returns
+ * -ENOMEM if buffer doesn't fit in vram.
+ * -ERESTARTSYS if interrupted.
+ */
+static int vmw_overlay_update_stream(struct vmw_private *dev_priv,
+				     struct vmw_dma_buffer *buf,
+				     struct drm_vmw_control_stream_arg *arg,
+				     bool interruptible)
+{
+	struct vmw_overlay *overlay = dev_priv->overlay_priv;
+	struct vmw_stream *stream = &overlay->stream[arg->stream_id];
+	int ret = 0;
+
+	if (!buf)
+		return -EINVAL;
+
+	DRM_DEBUG("   %s: old %p, new %p, %spaused\n", __func__,
+		  stream->buf, buf, stream->paused ? "" : "not ");
+
+	if (stream->buf != buf) {
+		ret = vmw_overlay_stop(dev_priv, arg->stream_id,
+				       false, interruptible);
+		if (ret)
+			return ret;
+	} else if (!stream->paused) {
+		/* If the buffers match and not paused then just send
+		 * the put command, no need to do anything else.
+		 */
+		ret = vmw_overlay_send_put(dev_priv, buf, arg, interruptible);
+		if (ret == 0)
+			stream->saved = *arg;
+		else
+			BUG_ON(!interruptible);
+
+		return ret;
+	}
+
+	/* We don't start the old stream if we are interrupted.
+	 * Might return -ENOMEM if it can't fit the buffer in vram.
+	 */
+	ret = vmw_dmabuf_pin_in_vram(dev_priv, buf, true, interruptible);
+	if (ret)
+		return ret;
+
+	ret = vmw_overlay_send_put(dev_priv, buf, arg, interruptible);
+	if (ret) {
+		/* This one needs to happen no matter what. We only remove
+		 * the NO_EVICT flag so this is safe from -ENOMEM.
+		 */
+		BUG_ON(vmw_dmabuf_pin_in_vram(dev_priv, buf, false, false) != 0);
+		return ret;
+	}
+
+	if (stream->buf != buf)
+		stream->buf = vmw_dmabuf_reference(buf);
+	stream->saved = *arg;
+
+	return 0;
+}
+
+/**
+ * Stop all streams.
+ *
+ * Used by the fb code when starting.
+ *
+ * Takes the overlay lock.
+ */
+int vmw_overlay_stop_all(struct vmw_private *dev_priv)
+{
+	struct vmw_overlay *overlay = dev_priv->overlay_priv;
+	int i, ret;
+
+	if (!overlay)
+		return 0;
+
+	mutex_lock(&overlay->mutex);
+
+	for (i = 0; i < VMW_MAX_NUM_STREAMS; i++) {
+		struct vmw_stream *stream = &overlay->stream[i];
+		if (!stream->buf)
+			continue;
+
+		ret = vmw_overlay_stop(dev_priv, i, false, false);
+		WARN_ON(ret != 0);
+	}
+
+	mutex_unlock(&overlay->mutex);
+
+	return 0;
+}
+
+/**
+ * Try to resume all paused streams.
+ *
+ * Used by the kms code after moving a new scanout buffer to vram.
+ *
+ * Takes the overlay lock.
+ */
+int vmw_overlay_resume_all(struct vmw_private *dev_priv)
+{
+	struct vmw_overlay *overlay = dev_priv->overlay_priv;
+	int i, ret;
+
+	if (!overlay)
+		return 0;
+
+	mutex_lock(&overlay->mutex);
+
+	for (i = 0; i < VMW_MAX_NUM_STREAMS; i++) {
+		struct vmw_stream *stream = &overlay->stream[i];
+		if (!stream->paused)
+			continue;
+
+		ret = vmw_overlay_update_stream(dev_priv, stream->buf,
+						&stream->saved, false);
+		if (ret != 0)
+			DRM_INFO("%s: *warning* failed to resume stream %i\n",
+				 __func__, i);
+	}
+
+	mutex_unlock(&overlay->mutex);
+
+	return 0;
+}
+
+/**
+ * Pauses all active streams.
+ *
+ * Used by the kms code when moving a new scanout buffer to vram.
+ *
+ * Takes the overlay lock.
+ */
+int vmw_overlay_pause_all(struct vmw_private *dev_priv)
+{
+	struct vmw_overlay *overlay = dev_priv->overlay_priv;
+	int i, ret;
+
+	if (!overlay)
+		return 0;
+
+	mutex_lock(&overlay->mutex);
+
+	for (i = 0; i < VMW_MAX_NUM_STREAMS; i++) {
+		if (overlay->stream[i].paused)
+			DRM_INFO("%s: *warning* stream %i already paused\n",
+				 __func__, i);
+		ret = vmw_overlay_stop(dev_priv, i, true, false);
+		WARN_ON(ret != 0);
+	}
+
+	mutex_unlock(&overlay->mutex);
+
+	return 0;
+}
+
+int vmw_overlay_ioctl(struct drm_device *dev, void *data,
+		      struct drm_file *file_priv)
+{
+	struct ttm_object_file *tfile = vmw_fpriv(file_priv)->tfile;
+	struct vmw_private *dev_priv = vmw_priv(dev);
+	struct vmw_overlay *overlay = dev_priv->overlay_priv;
+	struct drm_vmw_control_stream_arg *arg =
+	    (struct drm_vmw_control_stream_arg *)data;
+	struct vmw_dma_buffer *buf;
+	struct vmw_resource *res;
+	int ret;
+
+	if (!overlay)
+		return -ENOSYS;
+
+	ret = vmw_user_stream_lookup(dev_priv, tfile, &arg->stream_id, &res);
+	if (ret)
+		return ret;
+
+	mutex_lock(&overlay->mutex);
+
+	if (!arg->enabled) {
+		ret = vmw_overlay_stop(dev_priv, arg->stream_id, false, true);
+		goto out_unlock;
+	}
+
+	ret = vmw_user_dmabuf_lookup(tfile, arg->handle, &buf);
+	if (ret)
+		goto out_unlock;
+
+	ret = vmw_overlay_update_stream(dev_priv, buf, arg, true);
+
+	vmw_dmabuf_unreference(&buf);
+
+out_unlock:
+	mutex_unlock(&overlay->mutex);
+	vmw_resource_unreference(&res);
+
+	return ret;
+}
+
+int vmw_overlay_num_overlays(struct vmw_private *dev_priv)
+{
+	if (!dev_priv->overlay_priv)
+		return 0;
+
+	return VMW_MAX_NUM_STREAMS;
+}
+
+int vmw_overlay_num_free_overlays(struct vmw_private *dev_priv)
+{
+	struct vmw_overlay *overlay = dev_priv->overlay_priv;
+	int i, k;
+
+	if (!overlay)
+		return 0;
+
+	mutex_lock(&overlay->mutex);
+
+	for (i = 0, k = 0; i < VMW_MAX_NUM_STREAMS; i++)
+		if (!overlay->stream[i].claimed)
+			k++;
+
+	mutex_unlock(&overlay->mutex);
+
+	return k;
+}
+
+int vmw_overlay_claim(struct vmw_private *dev_priv, uint32_t *out)
+{
+	struct vmw_overlay *overlay = dev_priv->overlay_priv;
+	int i;
+
+	if (!overlay)
+		return -ENOSYS;
+
+	mutex_lock(&overlay->mutex);
+
+	for (i = 0; i < VMW_MAX_NUM_STREAMS; i++) {
+
+		if (overlay->stream[i].claimed)
+			continue;
+
+		overlay->stream[i].claimed = true;
+		*out = i;
+		mutex_unlock(&overlay->mutex);
+		return 0;
+	}
+
+	mutex_unlock(&overlay->mutex);
+	return -ESRCH;
+}
+
+int vmw_overlay_unref(struct vmw_private *dev_priv, uint32_t stream_id)
+{
+	struct vmw_overlay *overlay = dev_priv->overlay_priv;
+
+	BUG_ON(stream_id >= VMW_MAX_NUM_STREAMS);
+
+	if (!overlay)
+		return -ENOSYS;
+
+	mutex_lock(&overlay->mutex);
+
+	WARN_ON(!overlay->stream[stream_id].claimed);
+	vmw_overlay_stop(dev_priv, stream_id, false, false);
+	overlay->stream[stream_id].claimed = false;
+
+	mutex_unlock(&overlay->mutex);
+	return 0;
+}
+
+int vmw_overlay_init(struct vmw_private *dev_priv)
+{
+	struct vmw_overlay *overlay;
+	int i;
+
+	if (dev_priv->overlay_priv)
+		return -EINVAL;
+
+	if (!(dev_priv->fifo.capabilities & SVGA_FIFO_CAP_VIDEO) &&
+	     (dev_priv->fifo.capabilities & SVGA_FIFO_CAP_ESCAPE)) {
+		DRM_INFO("hardware doesn't support overlays\n");
+		return -ENOSYS;
+	}
+
+	overlay = kmalloc(GFP_KERNEL, sizeof(*overlay));
+	if (!overlay)
+		return -ENOMEM;
+
+	memset(overlay, 0, sizeof(*overlay));
+	mutex_init(&overlay->mutex);
+	for (i = 0; i < VMW_MAX_NUM_STREAMS; i++) {
+		overlay->stream[i].buf = NULL;
+		overlay->stream[i].paused = false;
+		overlay->stream[i].claimed = false;
+	}
+
+	dev_priv->overlay_priv = overlay;
+
+	return 0;
+}
+
+int vmw_overlay_close(struct vmw_private *dev_priv)
+{
+	struct vmw_overlay *overlay = dev_priv->overlay_priv;
+	bool forgotten_buffer = false;
+	int i;
+
+	if (!overlay)
+		return -ENOSYS;
+
+	for (i = 0; i < VMW_MAX_NUM_STREAMS; i++) {
+		if (overlay->stream[i].buf) {
+			forgotten_buffer = true;
+			vmw_overlay_stop(dev_priv, i, false, false);
+		}
+	}
+
+	WARN_ON(forgotten_buffer);
+
+	dev_priv->overlay_priv = NULL;
+	kfree(overlay);
+
+	return 0;
+}