[PATCH v4 00/12] drm/virtio: switch from ttm to gem shmem helpers.

[PATCH v4 00/12] drm/virtio: switch from ttm to gem shmem helpers.

[Gerd Hoffmann]

ttm increasingly gets into the way while hacking on virtio-gpu memory
management.  It also overkill for what virtio-gpu needs.  Lets get rid
of it.

v4:
 - make gem array helpers private to virtio.
 - misc minor fixes.
v3:
 - add gem array helpers.
 - rework fencing.

cheers,
  Gerd

Gerd Hoffmann (12):
  drm/virtio: pass gem reservation object to ttm init
  drm/virtio: switch virtio_gpu_wait_ioctl() to gem helper.
  drm/virtio: simplify cursor updates
  drm/virtio: remove virtio_gpu_object_wait
  drm/virtio: drop no_wait argument from virtio_gpu_object_reserve
  drm/virtio: remove ttm calls from in
    virtio_gpu_object_{reserve,unreserve}
  drm/virtio: add virtio_gpu_object_array & helpers
  drm/virtio: rework virtio_gpu_execbuffer_ioctl fencing
  drm/virtio: rework virtio_gpu_object_create fencing
  drm/virtio: drop virtio_gpu_object_list_validate/virtio_gpu_unref_list
  drm/virtio: switch from ttm to gem shmem helpers
  drm/virtio: remove virtio_gpu_alloc_object

 drivers/gpu/drm/virtio/virtgpu_drv.h    |  86 +++----
 drivers/gpu/drm/virtio/virtgpu_drv.c    |  20 +-
 drivers/gpu/drm/virtio/virtgpu_gem.c    |  93 +++++---
 drivers/gpu/drm/virtio/virtgpu_ioctl.c  | 154 ++++--------
 drivers/gpu/drm/virtio/virtgpu_kms.c    |   9 -
 drivers/gpu/drm/virtio/virtgpu_object.c | 210 +++++-----------
 drivers/gpu/drm/virtio/virtgpu_plane.c  |  13 +-
 drivers/gpu/drm/virtio/virtgpu_prime.c  |  37 ---
 drivers/gpu/drm/virtio/virtgpu_ttm.c    | 304 ------------------------
 drivers/gpu/drm/virtio/virtgpu_vq.c     |  48 +++-
 drivers/gpu/drm/virtio/Kconfig          |   2 +-
 drivers/gpu/drm/virtio/Makefile         |   2 +-
 12 files changed, 236 insertions(+), 742 deletions(-)
 delete mode 100644 drivers/gpu/drm/virtio/virtgpu_ttm.c

-- 
2.18.1

_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/dri-devel

[PATCH v4 01/12] drm/virtio: pass gem reservation object to ttm init

[Gerd Hoffmann]

With this gem and ttm will use the same reservation object,
so mixing and matching ttm / gem reservation helpers should
work fine.

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
---
 drivers/gpu/drm/virtio/virtgpu_object.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/virtio/virtgpu_object.c b/drivers/gpu/drm/virtio/virtgpu_object.c
index b2da31310d24..242766d644a7 100644
--- a/drivers/gpu/drm/virtio/virtgpu_object.c
+++ b/drivers/gpu/drm/virtio/virtgpu_object.c
@@ -132,7 +132,8 @@ int virtio_gpu_object_create(struct virtio_gpu_device *vgdev,
 	virtio_gpu_init_ttm_placement(bo);
 	ret = ttm_bo_init(&vgdev->mman.bdev, &bo->tbo, params->size,
 			  ttm_bo_type_device, &bo->placement, 0,
-			  true, acc_size, NULL, NULL,
+			  true, acc_size, NULL,
+			  bo->gem_base.resv,
 			  &virtio_gpu_ttm_bo_destroy);
 	/* ttm_bo_init failure will call the destroy */
 	if (ret != 0)
-- 
2.18.1

[PATCH v4 02/12] drm/virtio: switch virtio_gpu_wait_ioctl() to gem helper.

[Gerd Hoffmann]

Use drm_gem_reservation_object_wait() in virtio_gpu_wait_ioctl().
This also makes the ioctl run lockless.

v2: use reservation_object_test_signaled_rcu for VIRTGPU_WAIT_NOWAIT.

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
---
 drivers/gpu/drm/virtio/virtgpu_ioctl.c | 24 ++++++++++--------------
 1 file changed, 10 insertions(+), 14 deletions(-)

diff --git a/drivers/gpu/drm/virtio/virtgpu_ioctl.c b/drivers/gpu/drm/virtio/virtgpu_ioctl.c
index ac60be9b5c19..313c770ea2c5 100644
--- a/drivers/gpu/drm/virtio/virtgpu_ioctl.c
+++ b/drivers/gpu/drm/virtio/virtgpu_ioctl.c
@@ -464,23 +464,19 @@ static int virtio_gpu_wait_ioctl(struct drm_device *dev, void *data,
 			    struct drm_file *file)
 {
 	struct drm_virtgpu_3d_wait *args = data;
-	struct drm_gem_object *gobj = NULL;
-	struct virtio_gpu_object *qobj = NULL;
+	struct drm_gem_object *obj;
+	long timeout = 15 * HZ;
 	int ret;
-	bool nowait = false;
 
-	gobj = drm_gem_object_lookup(file, args->handle);
-	if (gobj == NULL)
-		return -ENOENT;
+	if (args->flags & VIRTGPU_WAIT_NOWAIT) {
+		obj = drm_gem_object_lookup(file, args->handle);
+		ret = reservation_object_test_signaled_rcu(obj->resv, true);
+		drm_gem_object_put_unlocked(obj);
+		return ret ? 0 : -EBUSY;
+	}
 
-	qobj = gem_to_virtio_gpu_obj(gobj);
-
-	if (args->flags & VIRTGPU_WAIT_NOWAIT)
-		nowait = true;
-	ret = virtio_gpu_object_wait(qobj, nowait);
-
-	drm_gem_object_put_unlocked(gobj);
-	return ret;
+	return drm_gem_reservation_object_wait(file, args->handle,
+					       true, timeout);
 }
 
 static int virtio_gpu_get_caps_ioctl(struct drm_device *dev,
-- 
2.18.1

_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/dri-devel

[PATCH v4 03/12] drm/virtio: simplify cursor updates

[Gerd Hoffmann]

No need to do the reservation dance,
we can just wait on the fence directly.

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
---
 drivers/gpu/drm/virtio/virtgpu_plane.c | 13 +++----------
 1 file changed, 3 insertions(+), 10 deletions(-)

diff --git a/drivers/gpu/drm/virtio/virtgpu_plane.c b/drivers/gpu/drm/virtio/virtgpu_plane.c
index 024c2aa0c929..4b805bf466d3 100644
--- a/drivers/gpu/drm/virtio/virtgpu_plane.c
+++ b/drivers/gpu/drm/virtio/virtgpu_plane.c
@@ -184,7 +184,6 @@ static void virtio_gpu_cursor_plane_update(struct drm_plane *plane,
 	struct virtio_gpu_framebuffer *vgfb;
 	struct virtio_gpu_object *bo = NULL;
 	uint32_t handle;
-	int ret = 0;
 
 	if (plane->state->crtc)
 		output = drm_crtc_to_virtio_gpu_output(plane->state->crtc);
@@ -208,15 +207,9 @@ static void virtio_gpu_cursor_plane_update(struct drm_plane *plane,
 			 cpu_to_le32(plane->state->crtc_w),
 			 cpu_to_le32(plane->state->crtc_h),
 			 0, 0, vgfb->fence);
-		ret = virtio_gpu_object_reserve(bo, false);
-		if (!ret) {
-			reservation_object_add_excl_fence(bo->tbo.resv,
-							  &vgfb->fence->f);
-			dma_fence_put(&vgfb->fence->f);
-			vgfb->fence = NULL;
-			virtio_gpu_object_unreserve(bo);
-			virtio_gpu_object_wait(bo, false);
-		}
+		dma_fence_wait(&vgfb->fence->f, true);
+		dma_fence_put(&vgfb->fence->f);
+		vgfb->fence = NULL;
 	}
 
 	if (plane->state->fb != old_state->fb) {
-- 
2.18.1

_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/dri-devel

[PATCH v4 04/12] drm/virtio: remove virtio_gpu_object_wait

[Gerd Hoffmann]

No users left.

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
---
 drivers/gpu/drm/virtio/virtgpu_drv.h    |  1 -
 drivers/gpu/drm/virtio/virtgpu_object.c | 13 -------------
 2 files changed, 14 deletions(-)

diff --git a/drivers/gpu/drm/virtio/virtgpu_drv.h b/drivers/gpu/drm/virtio/virtgpu_drv.h
index 9e2d3062b01d..2cd96256ba37 100644
--- a/drivers/gpu/drm/virtio/virtgpu_drv.h
+++ b/drivers/gpu/drm/virtio/virtgpu_drv.h
@@ -364,7 +364,6 @@ int virtio_gpu_object_kmap(struct virtio_gpu_object *bo);
 int virtio_gpu_object_get_sg_table(struct virtio_gpu_device *qdev,
 				   struct virtio_gpu_object *bo);
 void virtio_gpu_object_free_sg_table(struct virtio_gpu_object *bo);
-int virtio_gpu_object_wait(struct virtio_gpu_object *bo, bool no_wait);
 
 /* virtgpu_prime.c */
 struct sg_table *virtgpu_gem_prime_get_sg_table(struct drm_gem_object *obj);
diff --git a/drivers/gpu/drm/virtio/virtgpu_object.c b/drivers/gpu/drm/virtio/virtgpu_object.c
index 242766d644a7..82bfbf983fd2 100644
--- a/drivers/gpu/drm/virtio/virtgpu_object.c
+++ b/drivers/gpu/drm/virtio/virtgpu_object.c
@@ -233,16 +233,3 @@ void virtio_gpu_object_free_sg_table(struct virtio_gpu_object *bo)
 	kfree(bo->pages);
 	bo->pages = NULL;
 }
-
-int virtio_gpu_object_wait(struct virtio_gpu_object *bo, bool no_wait)
-{
-	int r;
-
-	r = ttm_bo_reserve(&bo->tbo, true, no_wait, NULL);
-	if (unlikely(r != 0))
-		return r;
-	r = ttm_bo_wait(&bo->tbo, true, no_wait);
-	ttm_bo_unreserve(&bo->tbo);
-	return r;
-}
-
-- 
2.18.1

[PATCH v4 05/12] drm/virtio: drop no_wait argument from virtio_gpu_object_reserve

[Gerd Hoffmann]

All callers pass no_wait = false.

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
---
 drivers/gpu/drm/virtio/virtgpu_drv.h   | 5 ++---
 drivers/gpu/drm/virtio/virtgpu_gem.c   | 4 ++--
 drivers/gpu/drm/virtio/virtgpu_ioctl.c | 4 ++--
 3 files changed, 6 insertions(+), 7 deletions(-)

diff --git a/drivers/gpu/drm/virtio/virtgpu_drv.h b/drivers/gpu/drm/virtio/virtgpu_drv.h
index 2cd96256ba37..06cc0e961df6 100644
--- a/drivers/gpu/drm/virtio/virtgpu_drv.h
+++ b/drivers/gpu/drm/virtio/virtgpu_drv.h
@@ -398,12 +398,11 @@ static inline u64 virtio_gpu_object_mmap_offset(struct virtio_gpu_object *bo)
 	return drm_vma_node_offset_addr(&bo->tbo.vma_node);
 }
 
-static inline int virtio_gpu_object_reserve(struct virtio_gpu_object *bo,
-					 bool no_wait)
+static inline int virtio_gpu_object_reserve(struct virtio_gpu_object *bo)
 {
 	int r;
 
-	r = ttm_bo_reserve(&bo->tbo, true, no_wait, NULL);
+	r = ttm_bo_reserve(&bo->tbo, true, false, NULL);
 	if (unlikely(r != 0)) {
 		if (r != -ERESTARTSYS) {
 			struct virtio_gpu_device *qdev =
diff --git a/drivers/gpu/drm/virtio/virtgpu_gem.c b/drivers/gpu/drm/virtio/virtgpu_gem.c
index 1e49e08dd545..9c9ad3b14080 100644
--- a/drivers/gpu/drm/virtio/virtgpu_gem.c
+++ b/drivers/gpu/drm/virtio/virtgpu_gem.c
@@ -140,7 +140,7 @@ int virtio_gpu_gem_object_open(struct drm_gem_object *obj,
 	if (!vgdev->has_virgl_3d)
 		return 0;
 
-	r = virtio_gpu_object_reserve(qobj, false);
+	r = virtio_gpu_object_reserve(qobj);
 	if (r)
 		return r;
 
@@ -161,7 +161,7 @@ void virtio_gpu_gem_object_close(struct drm_gem_object *obj,
 	if (!vgdev->has_virgl_3d)
 		return;
 
-	r = virtio_gpu_object_reserve(qobj, false);
+	r = virtio_gpu_object_reserve(qobj);
 	if (r)
 		return;
 
diff --git a/drivers/gpu/drm/virtio/virtgpu_ioctl.c b/drivers/gpu/drm/virtio/virtgpu_ioctl.c
index 313c770ea2c5..5cffd2e54c04 100644
--- a/drivers/gpu/drm/virtio/virtgpu_ioctl.c
+++ b/drivers/gpu/drm/virtio/virtgpu_ioctl.c
@@ -375,7 +375,7 @@ static int virtio_gpu_transfer_from_host_ioctl(struct drm_device *dev,
 
 	qobj = gem_to_virtio_gpu_obj(gobj);
 
-	ret = virtio_gpu_object_reserve(qobj, false);
+	ret = virtio_gpu_object_reserve(qobj);
 	if (ret)
 		goto out;
 
@@ -425,7 +425,7 @@ static int virtio_gpu_transfer_to_host_ioctl(struct drm_device *dev, void *data,
 
 	qobj = gem_to_virtio_gpu_obj(gobj);
 
-	ret = virtio_gpu_object_reserve(qobj, false);
+	ret = virtio_gpu_object_reserve(qobj);
 	if (ret)
 		goto out;
 
-- 
2.18.1

_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/dri-devel

[PATCH v4 06/12] drm/virtio: remove ttm calls from in virtio_gpu_object_{reserve, unreserve}

[Gerd Hoffmann]

Call reservation_object_* directly instead
of using ttm_bo_{reserve,unreserve}.

v4: check for EINTR only.
v3: check for EINTR too.

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
---
 drivers/gpu/drm/virtio/virtgpu_drv.h | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/drivers/gpu/drm/virtio/virtgpu_drv.h b/drivers/gpu/drm/virtio/virtgpu_drv.h
index 06cc0e961df6..07f6001ea91e 100644
--- a/drivers/gpu/drm/virtio/virtgpu_drv.h
+++ b/drivers/gpu/drm/virtio/virtgpu_drv.h
@@ -402,9 +402,9 @@ static inline int virtio_gpu_object_reserve(struct virtio_gpu_object *bo)
 {
 	int r;
 
-	r = ttm_bo_reserve(&bo->tbo, true, false, NULL);
+	r = reservation_object_lock_interruptible(bo->gem_base.resv, NULL);
 	if (unlikely(r != 0)) {
-		if (r != -ERESTARTSYS) {
+		if (r != -EINTR) {
 			struct virtio_gpu_device *qdev =
 				bo->gem_base.dev->dev_private;
 			dev_err(qdev->dev, "%p reserve failed\n", bo);
@@ -416,7 +416,7 @@ static inline int virtio_gpu_object_reserve(struct virtio_gpu_object *bo)
 
 static inline void virtio_gpu_object_unreserve(struct virtio_gpu_object *bo)
 {
-	ttm_bo_unreserve(&bo->tbo);
+	reservation_object_unlock(bo->gem_base.resv);
 }
 
 /* virgl debufs */
-- 
2.18.1

_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/dri-devel

[PATCH v4 07/12] drm/virtio: add virtio_gpu_object_array & helpers

[Gerd Hoffmann]

Some helper functions to manage an array of gem objects.

v4: make them virtio-private instead of generic helpers.

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
---
 drivers/gpu/drm/virtio/virtgpu_drv.h | 10 ++++++
 drivers/gpu/drm/virtio/virtgpu_gem.c | 50 ++++++++++++++++++++++++++++
 2 files changed, 60 insertions(+)

diff --git a/drivers/gpu/drm/virtio/virtgpu_drv.h b/drivers/gpu/drm/virtio/virtgpu_drv.h
index 07f6001ea91e..98d646789d23 100644
--- a/drivers/gpu/drm/virtio/virtgpu_drv.h
+++ b/drivers/gpu/drm/virtio/virtgpu_drv.h
@@ -84,6 +84,11 @@ struct virtio_gpu_object {
 #define gem_to_virtio_gpu_obj(gobj) \
 	container_of((gobj), struct virtio_gpu_object, gem_base)
 
+struct virtio_gpu_object_array {
+	u32 nents;
+	struct drm_gem_object *objs[];
+};
+
 struct virtio_gpu_vbuffer;
 struct virtio_gpu_device;
 
@@ -251,6 +256,11 @@ int virtio_gpu_mode_dumb_mmap(struct drm_file *file_priv,
 			      struct drm_device *dev,
 			      uint32_t handle, uint64_t *offset_p);
 
+struct virtio_gpu_object_array *virtio_gpu_array_alloc(u32 nents);
+struct virtio_gpu_object_array*
+virtio_gpu_array_from_handles(struct drm_file *drm_file, u32 *handles, u32 nents);
+void virtio_gpu_array_put_free(struct virtio_gpu_object_array *objs);
+
 /* virtio vg */
 int virtio_gpu_alloc_vbufs(struct virtio_gpu_device *vgdev);
 void virtio_gpu_free_vbufs(struct virtio_gpu_device *vgdev);
diff --git a/drivers/gpu/drm/virtio/virtgpu_gem.c b/drivers/gpu/drm/virtio/virtgpu_gem.c
index 9c9ad3b14080..456cc382ce68 100644
--- a/drivers/gpu/drm/virtio/virtgpu_gem.c
+++ b/drivers/gpu/drm/virtio/virtgpu_gem.c
@@ -169,3 +169,53 @@ void virtio_gpu_gem_object_close(struct drm_gem_object *obj,
 						qobj->hw_res_handle);
 	virtio_gpu_object_unreserve(qobj);
 }
+
+struct virtio_gpu_object_array *virtio_gpu_array_alloc(u32 nents)
+{
+	struct virtio_gpu_object_array *objs;
+	size_t size = sizeof(*objs) + sizeof(objs->objs[0]) * nents;
+
+	objs = kzalloc(size, GFP_KERNEL);
+	if (!objs)
+		return NULL;
+
+	objs->nents = nents;
+	return objs;
+}
+
+static void virtio_gpu_array_free(struct virtio_gpu_object_array *objs)
+{
+	kfree(objs);
+}
+
+struct virtio_gpu_object_array*
+virtio_gpu_array_from_handles(struct drm_file *drm_file, u32 *handles, u32 nents)
+{
+	struct virtio_gpu_object_array *objs;
+	u32 i;
+
+	objs = virtio_gpu_array_alloc(nents);
+	if (!objs)
+		return NULL;
+
+	for (i = 0; i < nents; i++) {
+		objs->objs[i] = drm_gem_object_lookup(drm_file, handles[i]);
+		if (!objs->objs[i]) {
+			virtio_gpu_array_put_free(objs);
+			return NULL;
+		}
+	}
+	return objs;
+}
+
+void virtio_gpu_array_put_free(struct virtio_gpu_object_array *objs)
+{
+	u32 i;
+
+	for (i = 0; i < objs->nents; i++) {
+		if (!objs->objs[i])
+			continue;
+		drm_gem_object_put_unlocked(objs->objs[i]);
+	}
+	virtio_gpu_array_free(objs);
+}
-- 
2.18.1

[PATCH v4 08/12] drm/virtio: rework virtio_gpu_execbuffer_ioctl fencing

[Gerd Hoffmann]

Use gem reservation helpers and direct reservation_object_* calls
instead of ttm.

v3: Also attach the array of gem objects to the virtio command buffer,
so we can drop the object references in the completion callback.  Needed
because ttm fence helpers grab a reference for us, but gem helpers
don't.

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Acked-by: Daniel Vetter <daniel.vetter@ffwll.ch>
---
 drivers/gpu/drm/virtio/virtgpu_drv.h   |  5 ++-
 drivers/gpu/drm/virtio/virtgpu_ioctl.c | 62 +++++++++++---------------
 drivers/gpu/drm/virtio/virtgpu_vq.c    | 16 ++++---
 3 files changed, 40 insertions(+), 43 deletions(-)

diff --git a/drivers/gpu/drm/virtio/virtgpu_drv.h b/drivers/gpu/drm/virtio/virtgpu_drv.h
index 98d646789d23..65f5ce41c341 100644
--- a/drivers/gpu/drm/virtio/virtgpu_drv.h
+++ b/drivers/gpu/drm/virtio/virtgpu_drv.h
@@ -120,9 +120,9 @@ struct virtio_gpu_vbuffer {
 
 	char *resp_buf;
 	int resp_size;
-
 	virtio_gpu_resp_cb resp_cb;
 
+	struct virtio_gpu_object_array *objs;
 	struct list_head list;
 };
 
@@ -311,7 +311,8 @@ void virtio_gpu_cmd_context_detach_resource(struct virtio_gpu_device *vgdev,
 					    uint32_t resource_id);
 void virtio_gpu_cmd_submit(struct virtio_gpu_device *vgdev,
 			   void *data, uint32_t data_size,
-			   uint32_t ctx_id, struct virtio_gpu_fence *fence);
+			   uint32_t ctx_id, struct virtio_gpu_fence *fence,
+			   struct virtio_gpu_object_array *objs);
 void virtio_gpu_cmd_transfer_from_host_3d(struct virtio_gpu_device *vgdev,
 					  uint32_t resource_id, uint32_t ctx_id,
 					  uint64_t offset, uint32_t level,
diff --git a/drivers/gpu/drm/virtio/virtgpu_ioctl.c b/drivers/gpu/drm/virtio/virtgpu_ioctl.c
index 5cffd2e54c04..c3901c397b8a 100644
--- a/drivers/gpu/drm/virtio/virtgpu_ioctl.c
+++ b/drivers/gpu/drm/virtio/virtgpu_ioctl.c
@@ -105,14 +105,11 @@ static int virtio_gpu_execbuffer_ioctl(struct drm_device *dev, void *data,
 	struct drm_virtgpu_execbuffer *exbuf = data;
 	struct virtio_gpu_device *vgdev = dev->dev_private;
 	struct virtio_gpu_fpriv *vfpriv = drm_file->driver_priv;
-	struct drm_gem_object *gobj;
 	struct virtio_gpu_fence *out_fence;
-	struct virtio_gpu_object *qobj;
 	int ret;
 	uint32_t *bo_handles = NULL;
 	void __user *user_bo_handles = NULL;
-	struct list_head validate_list;
-	struct ttm_validate_buffer *buflist = NULL;
+	struct virtio_gpu_object_array *buflist = NULL;
 	int i;
 	struct ww_acquire_ctx ticket;
 	struct sync_file *sync_file;
@@ -155,15 +152,10 @@ static int virtio_gpu_execbuffer_ioctl(struct drm_device *dev, void *data,
 			return out_fence_fd;
 	}
 
-	INIT_LIST_HEAD(&validate_list);
 	if (exbuf->num_bo_handles) {
-
 		bo_handles = kvmalloc_array(exbuf->num_bo_handles,
-					   sizeof(uint32_t), GFP_KERNEL);
-		buflist = kvmalloc_array(exbuf->num_bo_handles,
-					   sizeof(struct ttm_validate_buffer),
-					   GFP_KERNEL | __GFP_ZERO);
-		if (!bo_handles || !buflist) {
+					    sizeof(uint32_t), GFP_KERNEL);
+		if (!bo_handles) {
 			ret = -ENOMEM;
 			goto out_unused_fd;
 		}
@@ -175,25 +167,22 @@ static int virtio_gpu_execbuffer_ioctl(struct drm_device *dev, void *data,
 			goto out_unused_fd;
 		}
 
-		for (i = 0; i < exbuf->num_bo_handles; i++) {
-			gobj = drm_gem_object_lookup(drm_file, bo_handles[i]);
-			if (!gobj) {
-				ret = -ENOENT;
-				goto out_unused_fd;
-			}
-
-			qobj = gem_to_virtio_gpu_obj(gobj);
-			buflist[i].bo = &qobj->tbo;
-
-			list_add(&buflist[i].head, &validate_list);
+		buflist = virtio_gpu_array_from_handles(drm_file, bo_handles,
+							exbuf->num_bo_handles);
+		if (!buflist) {
+			ret = -ENOENT;
+			goto out_unused_fd;
 		}
 		kvfree(bo_handles);
 		bo_handles = NULL;
 	}
 
-	ret = virtio_gpu_object_list_validate(&ticket, &validate_list);
-	if (ret)
-		goto out_free;
+	if (buflist) {
+		ret = drm_gem_lock_reservations(buflist->objs, buflist->nents,
+						&ticket);
+		if (ret)
+			goto out_unused_fd;
+	}
 
 	buf = memdup_user(u64_to_user_ptr(exbuf->command), exbuf->size);
 	if (IS_ERR(buf)) {
@@ -219,25 +208,26 @@ static int virtio_gpu_execbuffer_ioctl(struct drm_device *dev, void *data,
 		fd_install(out_fence_fd, sync_file->file);
 	}
 
+	if (buflist) {
+		for (i = 0; i < exbuf->num_bo_handles; i++)
+			reservation_object_add_excl_fence(buflist->objs[i]->resv,
+							  &out_fence->f);
+		drm_gem_unlock_reservations(buflist->objs, buflist->nents,
+					    &ticket);
+	}
+
 	virtio_gpu_cmd_submit(vgdev, buf, exbuf->size,
-			      vfpriv->ctx_id, out_fence);
-
-	ttm_eu_fence_buffer_objects(&ticket, &validate_list, &out_fence->f);
-
-	/* fence the command bo */
-	virtio_gpu_unref_list(&validate_list);
-	kvfree(buflist);
+			      vfpriv->ctx_id, out_fence, buflist);
 	return 0;
 
 out_memdup:
 	kfree(buf);
 out_unresv:
-	ttm_eu_backoff_reservation(&ticket, &validate_list);
-out_free:
-	virtio_gpu_unref_list(&validate_list);
+	drm_gem_unlock_reservations(buflist->objs, buflist->nents, &ticket);
 out_unused_fd:
 	kvfree(bo_handles);
-	kvfree(buflist);
+	if (buflist)
+		virtio_gpu_array_put_free(buflist);
 
 	if (out_fence_fd >= 0)
 		put_unused_fd(out_fence_fd);
diff --git a/drivers/gpu/drm/virtio/virtgpu_vq.c b/drivers/gpu/drm/virtio/virtgpu_vq.c
index 6c1a90717535..dc2c2c003200 100644
--- a/drivers/gpu/drm/virtio/virtgpu_vq.c
+++ b/drivers/gpu/drm/virtio/virtgpu_vq.c
@@ -191,7 +191,7 @@ void virtio_gpu_dequeue_ctrl_func(struct work_struct *work)
 	} while (!virtqueue_enable_cb(vgdev->ctrlq.vq));
 	spin_unlock(&vgdev->ctrlq.qlock);
 
-	list_for_each_entry_safe(entry, tmp, &reclaim_list, list) {
+	list_for_each_entry(entry, &reclaim_list, list) {
 		resp = (struct virtio_gpu_ctrl_hdr *)entry->resp_buf;
 
 		trace_virtio_gpu_cmd_response(vgdev->ctrlq.vq, resp);
@@ -218,14 +218,18 @@ void virtio_gpu_dequeue_ctrl_func(struct work_struct *work)
 		}
 		if (entry->resp_cb)
 			entry->resp_cb(vgdev, entry);
-
-		list_del(&entry->list);
-		free_vbuf(vgdev, entry);
 	}
 	wake_up(&vgdev->ctrlq.ack_queue);
 
 	if (fence_id)
 		virtio_gpu_fence_event_process(vgdev, fence_id);
+
+	list_for_each_entry_safe(entry, tmp, &reclaim_list, list) {
+		if (entry->objs)
+			virtio_gpu_array_put_free(entry->objs);
+		list_del(&entry->list);
+		free_vbuf(vgdev, entry);
+	}
 }
 
 void virtio_gpu_dequeue_cursor_func(struct work_struct *work)
@@ -939,7 +943,8 @@ void virtio_gpu_cmd_transfer_from_host_3d(struct virtio_gpu_device *vgdev,
 
 void virtio_gpu_cmd_submit(struct virtio_gpu_device *vgdev,
 			   void *data, uint32_t data_size,
-			   uint32_t ctx_id, struct virtio_gpu_fence *fence)
+			   uint32_t ctx_id, struct virtio_gpu_fence *fence,
+			   struct virtio_gpu_object_array *objs)
 {
 	struct virtio_gpu_cmd_submit *cmd_p;
 	struct virtio_gpu_vbuffer *vbuf;
@@ -949,6 +954,7 @@ void virtio_gpu_cmd_submit(struct virtio_gpu_device *vgdev,
 
 	vbuf->data_buf = data;
 	vbuf->data_size = data_size;
+	vbuf->objs = objs;
 
 	cmd_p->hdr.type = cpu_to_le32(VIRTIO_GPU_CMD_SUBMIT_3D);
 	cmd_p->hdr.ctx_id = cpu_to_le32(ctx_id);
-- 
2.18.1

[PATCH v4 09/12] drm/virtio: rework virtio_gpu_object_create fencing

[Gerd Hoffmann]

Use gem reservation helpers and direct reservation_object_* calls
instead of ttm.

v3: Due to using the gem reservation object it is initialized and ready
for use before calling ttm_bo_init, so we can also drop the tricky fence
logic which checks whenever the command is in flight still.  We can
simply fence our object before submitting the virtio command and be done
with it.

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Acked-by: Daniel Vetter <daniel.vetter@ffwll.ch>
---
 drivers/gpu/drm/virtio/virtgpu_drv.h    |  6 ++-
 drivers/gpu/drm/virtio/virtgpu_object.c | 54 +++++++++----------------
 drivers/gpu/drm/virtio/virtgpu_vq.c     |  8 +++-
 3 files changed, 30 insertions(+), 38 deletions(-)

diff --git a/drivers/gpu/drm/virtio/virtgpu_drv.h b/drivers/gpu/drm/virtio/virtgpu_drv.h
index 65f5ce41c341..5213d7f499eb 100644
--- a/drivers/gpu/drm/virtio/virtgpu_drv.h
+++ b/drivers/gpu/drm/virtio/virtgpu_drv.h
@@ -267,7 +267,8 @@ void virtio_gpu_free_vbufs(struct virtio_gpu_device *vgdev);
 void virtio_gpu_cmd_create_resource(struct virtio_gpu_device *vgdev,
 				    struct virtio_gpu_object *bo,
 				    struct virtio_gpu_object_params *params,
-				    struct virtio_gpu_fence *fence);
+				    struct virtio_gpu_fence *fence,
+				    struct virtio_gpu_object_array *objs);
 void virtio_gpu_cmd_unref_resource(struct virtio_gpu_device *vgdev,
 				   uint32_t resource_id);
 void virtio_gpu_cmd_transfer_to_host_2d(struct virtio_gpu_device *vgdev,
@@ -328,7 +329,8 @@ void
 virtio_gpu_cmd_resource_create_3d(struct virtio_gpu_device *vgdev,
 				  struct virtio_gpu_object *bo,
 				  struct virtio_gpu_object_params *params,
-				  struct virtio_gpu_fence *fence);
+				  struct virtio_gpu_fence *fence,
+				  struct virtio_gpu_object_array *objs);
 void virtio_gpu_ctrl_ack(struct virtqueue *vq);
 void virtio_gpu_cursor_ack(struct virtqueue *vq);
 void virtio_gpu_fence_ack(struct virtqueue *vq);
diff --git a/drivers/gpu/drm/virtio/virtgpu_object.c b/drivers/gpu/drm/virtio/virtgpu_object.c
index 82bfbf983fd2..90642907aa5c 100644
--- a/drivers/gpu/drm/virtio/virtgpu_object.c
+++ b/drivers/gpu/drm/virtio/virtgpu_object.c
@@ -97,6 +97,7 @@ int virtio_gpu_object_create(struct virtio_gpu_device *vgdev,
 			     struct virtio_gpu_object **bo_ptr,
 			     struct virtio_gpu_fence *fence)
 {
+	struct virtio_gpu_object_array *objs = NULL;
 	struct virtio_gpu_object *bo;
 	size_t acc_size;
 	int ret;
@@ -123,10 +124,27 @@ int virtio_gpu_object_create(struct virtio_gpu_device *vgdev,
 	}
 	bo->dumb = params->dumb;
 
+	if (fence) {
+		struct ww_acquire_ctx ticket;
+
+		objs = virtio_gpu_array_alloc(1);
+		objs->objs[0] = &bo->gem_base;
+		drm_gem_object_get(objs->objs[0]);
+
+		ret = drm_gem_lock_reservations(objs->objs, objs->nents,
+						&ticket);
+		if (ret == 0)
+			reservation_object_add_excl_fence(objs->objs[0]->resv,
+							  &fence->f);
+		drm_gem_unlock_reservations(objs->objs, objs->nents, &ticket);
+	}
+
 	if (params->virgl) {
-		virtio_gpu_cmd_resource_create_3d(vgdev, bo, params, fence);
+		virtio_gpu_cmd_resource_create_3d(vgdev, bo, params,
+						  fence, objs);
 	} else {
-		virtio_gpu_cmd_create_resource(vgdev, bo, params, fence);
+		virtio_gpu_cmd_create_resource(vgdev, bo, params,
+					       fence, objs);
 	}
 
 	virtio_gpu_init_ttm_placement(bo);
@@ -139,38 +157,6 @@ int virtio_gpu_object_create(struct virtio_gpu_device *vgdev,
 	if (ret != 0)
 		return ret;
 
-	if (fence) {
-		struct virtio_gpu_fence_driver *drv = &vgdev->fence_drv;
-		struct list_head validate_list;
-		struct ttm_validate_buffer mainbuf;
-		struct ww_acquire_ctx ticket;
-		unsigned long irq_flags;
-		bool signaled;
-
-		INIT_LIST_HEAD(&validate_list);
-		memset(&mainbuf, 0, sizeof(struct ttm_validate_buffer));
-
-		/* use a gem reference since unref list undoes them */
-		drm_gem_object_get(&bo->gem_base);
-		mainbuf.bo = &bo->tbo;
-		list_add(&mainbuf.head, &validate_list);
-
-		ret = virtio_gpu_object_list_validate(&ticket, &validate_list);
-		if (ret == 0) {
-			spin_lock_irqsave(&drv->lock, irq_flags);
-			signaled = virtio_fence_signaled(&fence->f);
-			if (!signaled)
-				/* virtio create command still in flight */
-				ttm_eu_fence_buffer_objects(&ticket, &validate_list,
-							    &fence->f);
-			spin_unlock_irqrestore(&drv->lock, irq_flags);
-			if (signaled)
-				/* virtio create command finished */
-				ttm_eu_backoff_reservation(&ticket, &validate_list);
-		}
-		virtio_gpu_unref_list(&validate_list);
-	}
-
 	*bo_ptr = bo;
 	return 0;
 }
diff --git a/drivers/gpu/drm/virtio/virtgpu_vq.c b/drivers/gpu/drm/virtio/virtgpu_vq.c
index dc2c2c003200..6e2b287a7e4b 100644
--- a/drivers/gpu/drm/virtio/virtgpu_vq.c
+++ b/drivers/gpu/drm/virtio/virtgpu_vq.c
@@ -391,13 +391,15 @@ static int virtio_gpu_queue_cursor(struct virtio_gpu_device *vgdev,
 void virtio_gpu_cmd_create_resource(struct virtio_gpu_device *vgdev,
 				    struct virtio_gpu_object *bo,
 				    struct virtio_gpu_object_params *params,
-				    struct virtio_gpu_fence *fence)
+				    struct virtio_gpu_fence *fence,
+				    struct virtio_gpu_object_array *objs)
 {
 	struct virtio_gpu_resource_create_2d *cmd_p;
 	struct virtio_gpu_vbuffer *vbuf;
 
 	cmd_p = virtio_gpu_alloc_cmd(vgdev, &vbuf, sizeof(*cmd_p));
 	memset(cmd_p, 0, sizeof(*cmd_p));
+	vbuf->objs = objs;
 
 	cmd_p->hdr.type = cpu_to_le32(VIRTIO_GPU_CMD_RESOURCE_CREATE_2D);
 	cmd_p->resource_id = cpu_to_le32(bo->hw_res_handle);
@@ -864,13 +866,15 @@ void
 virtio_gpu_cmd_resource_create_3d(struct virtio_gpu_device *vgdev,
 				  struct virtio_gpu_object *bo,
 				  struct virtio_gpu_object_params *params,
-				  struct virtio_gpu_fence *fence)
+				  struct virtio_gpu_fence *fence,
+				  struct virtio_gpu_object_array *objs)
 {
 	struct virtio_gpu_resource_create_3d *cmd_p;
 	struct virtio_gpu_vbuffer *vbuf;
 
 	cmd_p = virtio_gpu_alloc_cmd(vgdev, &vbuf, sizeof(*cmd_p));
 	memset(cmd_p, 0, sizeof(*cmd_p));
+	vbuf->objs = objs;
 
 	cmd_p->hdr.type = cpu_to_le32(VIRTIO_GPU_CMD_RESOURCE_CREATE_3D);
 	cmd_p->resource_id = cpu_to_le32(bo->hw_res_handle);
-- 
2.18.1

_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/dri-devel

[PATCH v4 10/12] drm/virtio: drop virtio_gpu_object_list_validate/virtio_gpu_unref_list

[Gerd Hoffmann]

No users left.

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Acked-by: Daniel Vetter <daniel.vetter@ffwll.ch>
---
 drivers/gpu/drm/virtio/virtgpu_drv.h   |  3 --
 drivers/gpu/drm/virtio/virtgpu_ioctl.c | 39 --------------------------
 2 files changed, 42 deletions(-)

diff --git a/drivers/gpu/drm/virtio/virtgpu_drv.h b/drivers/gpu/drm/virtio/virtgpu_drv.h
index 5213d7f499eb..c23f4016df39 100644
--- a/drivers/gpu/drm/virtio/virtgpu_drv.h
+++ b/drivers/gpu/drm/virtio/virtgpu_drv.h
@@ -222,9 +222,6 @@ struct virtio_gpu_fpriv {
 /* virtio_ioctl.c */
 #define DRM_VIRTIO_NUM_IOCTLS 10
 extern struct drm_ioctl_desc virtio_gpu_ioctls[DRM_VIRTIO_NUM_IOCTLS];
-int virtio_gpu_object_list_validate(struct ww_acquire_ctx *ticket,
-				    struct list_head *head);
-void virtio_gpu_unref_list(struct list_head *head);
 
 /* virtio_kms.c */
 int virtio_gpu_init(struct drm_device *dev);
diff --git a/drivers/gpu/drm/virtio/virtgpu_ioctl.c b/drivers/gpu/drm/virtio/virtgpu_ioctl.c
index c3901c397b8a..54e1cf84097a 100644
--- a/drivers/gpu/drm/virtio/virtgpu_ioctl.c
+++ b/drivers/gpu/drm/virtio/virtgpu_ioctl.c
@@ -54,45 +54,6 @@ static int virtio_gpu_map_ioctl(struct drm_device *dev, void *data,
 					 &virtio_gpu_map->offset);
 }
 
-int virtio_gpu_object_list_validate(struct ww_acquire_ctx *ticket,
-				    struct list_head *head)
-{
-	struct ttm_operation_ctx ctx = { false, false };
-	struct ttm_validate_buffer *buf;
-	struct ttm_buffer_object *bo;
-	struct virtio_gpu_object *qobj;
-	int ret;
-
-	ret = ttm_eu_reserve_buffers(ticket, head, true, NULL, true);
-	if (ret != 0)
-		return ret;
-
-	list_for_each_entry(buf, head, head) {
-		bo = buf->bo;
-		qobj = container_of(bo, struct virtio_gpu_object, tbo);
-		ret = ttm_bo_validate(bo, &qobj->placement, &ctx);
-		if (ret) {
-			ttm_eu_backoff_reservation(ticket, head);
-			return ret;
-		}
-	}
-	return 0;
-}
-
-void virtio_gpu_unref_list(struct list_head *head)
-{
-	struct ttm_validate_buffer *buf;
-	struct ttm_buffer_object *bo;
-	struct virtio_gpu_object *qobj;
-
-	list_for_each_entry(buf, head, head) {
-		bo = buf->bo;
-		qobj = container_of(bo, struct virtio_gpu_object, tbo);
-
-		drm_gem_object_put_unlocked(&qobj->gem_base);
-	}
-}
-
 /*
  * Usage of execbuffer:
  * Relocations need to take into account the full VIRTIO_GPUDrawable size.
-- 
2.18.1

_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/dri-devel

[PATCH v4 11/12] drm/virtio: switch from ttm to gem shmem helpers

[Gerd Hoffmann]

virtio-gpu basically needs a sg_table for the bo, to tell the host where
the backing pages for the object are.  So the gem shmem helpers are a
perfect fit.  Some drm_gem_object_funcs need thin wrappers to update the
host state, but otherwise the helpers handle everything just fine.

Once the fencing was sorted the switch was surprisingly easy and for the
most part just removing the ttm code.

v4: fix drm_gem_object_funcs name.

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Acked-by: Daniel Vetter <daniel.vetter@ffwll.ch>
---
 drivers/gpu/drm/virtio/virtgpu_drv.h    |  52 +---
 drivers/gpu/drm/virtio/virtgpu_drv.c    |  20 +-
 drivers/gpu/drm/virtio/virtgpu_gem.c    |  16 +-
 drivers/gpu/drm/virtio/virtgpu_ioctl.c  |  19 +-
 drivers/gpu/drm/virtio/virtgpu_kms.c    |   9 -
 drivers/gpu/drm/virtio/virtgpu_object.c | 148 ++++--------
 drivers/gpu/drm/virtio/virtgpu_prime.c  |  37 ---
 drivers/gpu/drm/virtio/virtgpu_ttm.c    | 304 ------------------------
 drivers/gpu/drm/virtio/virtgpu_vq.c     |  24 +-
 drivers/gpu/drm/virtio/Kconfig          |   2 +-
 drivers/gpu/drm/virtio/Makefile         |   2 +-
 11 files changed, 82 insertions(+), 551 deletions(-)
 delete mode 100644 drivers/gpu/drm/virtio/virtgpu_ttm.c

diff --git a/drivers/gpu/drm/virtio/virtgpu_drv.h b/drivers/gpu/drm/virtio/virtgpu_drv.h
index c23f4016df39..1d589de34449 100644
--- a/drivers/gpu/drm/virtio/virtgpu_drv.h
+++ b/drivers/gpu/drm/virtio/virtgpu_drv.h
@@ -33,14 +33,11 @@
 
 #include <drm/drmP.h>
 #include <drm/drm_gem.h>
+#include <drm/drm_gem_shmem_helper.h>
 #include <drm/drm_atomic.h>
 #include <drm/drm_encoder.h>
 #include <drm/drm_fb_helper.h>
 #include <drm/drm_probe_helper.h>
-#include <drm/ttm/ttm_bo_api.h>
-#include <drm/ttm/ttm_bo_driver.h>
-#include <drm/ttm/ttm_placement.h>
-#include <drm/ttm/ttm_module.h>
 
 #define DRIVER_NAME "virtio_gpu"
 #define DRIVER_DESC "virtio GPU"
@@ -68,21 +65,16 @@ struct virtio_gpu_object_params {
 };
 
 struct virtio_gpu_object {
-	struct drm_gem_object gem_base;
+	struct drm_gem_shmem_object base;
 	uint32_t hw_res_handle;
 
 	struct sg_table *pages;
 	uint32_t mapped;
-	void *vmap;
 	bool dumb;
-	struct ttm_place                placement_code;
-	struct ttm_placement		placement;
-	struct ttm_buffer_object	tbo;
-	struct ttm_bo_kmap_obj		kmap;
 	bool created;
 };
 #define gem_to_virtio_gpu_obj(gobj) \
-	container_of((gobj), struct virtio_gpu_object, gem_base)
+	container_of((gobj), struct virtio_gpu_object, base.base)
 
 struct virtio_gpu_object_array {
 	u32 nents;
@@ -152,10 +144,6 @@ struct virtio_gpu_framebuffer {
 #define to_virtio_gpu_framebuffer(x) \
 	container_of(x, struct virtio_gpu_framebuffer, base)
 
-struct virtio_gpu_mman {
-	struct ttm_bo_device		bdev;
-};
-
 struct virtio_gpu_queue {
 	struct virtqueue *vq;
 	spinlock_t qlock;
@@ -184,8 +172,6 @@ struct virtio_gpu_device {
 
 	struct virtio_device *vdev;
 
-	struct virtio_gpu_mman mman;
-
 	struct virtio_gpu_output outputs[VIRTIO_GPU_MAX_SCANOUTS];
 	uint32_t num_scanouts;
 
@@ -349,11 +335,6 @@ struct drm_plane *virtio_gpu_plane_init(struct virtio_gpu_device *vgdev,
 					enum drm_plane_type type,
 					int index);
 
-/* virtio_gpu_ttm.c */
-int virtio_gpu_ttm_init(struct virtio_gpu_device *vgdev);
-void virtio_gpu_ttm_fini(struct virtio_gpu_device *vgdev);
-int virtio_gpu_mmap(struct file *filp, struct vm_area_struct *vma);
-
 /* virtio_gpu_fence.c */
 bool virtio_fence_signaled(struct dma_fence *f);
 struct virtio_gpu_fence *virtio_gpu_fence_alloc(
@@ -365,58 +346,47 @@ void virtio_gpu_fence_event_process(struct virtio_gpu_device *vdev,
 				    u64 last_seq);
 
 /* virtio_gpu_object */
+struct drm_gem_object *virtio_gpu_create_object(struct drm_device *dev,
+						size_t size);
 int virtio_gpu_object_create(struct virtio_gpu_device *vgdev,
 			     struct virtio_gpu_object_params *params,
 			     struct virtio_gpu_object **bo_ptr,
 			     struct virtio_gpu_fence *fence);
-void virtio_gpu_object_kunmap(struct virtio_gpu_object *bo);
-int virtio_gpu_object_kmap(struct virtio_gpu_object *bo);
-int virtio_gpu_object_get_sg_table(struct virtio_gpu_device *qdev,
-				   struct virtio_gpu_object *bo);
-void virtio_gpu_object_free_sg_table(struct virtio_gpu_object *bo);
 
 /* virtgpu_prime.c */
-struct sg_table *virtgpu_gem_prime_get_sg_table(struct drm_gem_object *obj);
 struct drm_gem_object *virtgpu_gem_prime_import_sg_table(
 	struct drm_device *dev, struct dma_buf_attachment *attach,
 	struct sg_table *sgt);
-void *virtgpu_gem_prime_vmap(struct drm_gem_object *obj);
-void virtgpu_gem_prime_vunmap(struct drm_gem_object *obj, void *vaddr);
-int virtgpu_gem_prime_mmap(struct drm_gem_object *obj,
-			   struct vm_area_struct *vma);
 
 static inline struct virtio_gpu_object*
 virtio_gpu_object_ref(struct virtio_gpu_object *bo)
 {
-	ttm_bo_get(&bo->tbo);
+	drm_gem_object_get(&bo->base.base);
 	return bo;
 }
 
 static inline void virtio_gpu_object_unref(struct virtio_gpu_object **bo)
 {
-	struct ttm_buffer_object *tbo;
-
 	if ((*bo) == NULL)
 		return;
-	tbo = &((*bo)->tbo);
-	ttm_bo_put(tbo);
+	drm_gem_object_put(&(*bo)->base.base);
 	*bo = NULL;
 }
 
 static inline u64 virtio_gpu_object_mmap_offset(struct virtio_gpu_object *bo)
 {
-	return drm_vma_node_offset_addr(&bo->tbo.vma_node);
+	return drm_vma_node_offset_addr(&bo->base.base.vma_node);
 }
 
 static inline int virtio_gpu_object_reserve(struct virtio_gpu_object *bo)
 {
 	int r;
 
-	r = reservation_object_lock_interruptible(bo->gem_base.resv, NULL);
+	r = reservation_object_lock_interruptible(bo->base.base.resv, NULL);
 	if (unlikely(r != 0)) {
 		if (r != -EINTR) {
 			struct virtio_gpu_device *qdev =
-				bo->gem_base.dev->dev_private;
+				bo->base.base.dev->dev_private;
 			dev_err(qdev->dev, "%p reserve failed\n", bo);
 		}
 		return r;
@@ -426,7 +396,7 @@ static inline int virtio_gpu_object_reserve(struct virtio_gpu_object *bo)
 
 static inline void virtio_gpu_object_unreserve(struct virtio_gpu_object *bo)
 {
-	reservation_object_unlock(bo->gem_base.resv);
+	reservation_object_unlock(bo->base.base.resv);
 }
 
 /* virgl debufs */
diff --git a/drivers/gpu/drm/virtio/virtgpu_drv.c b/drivers/gpu/drm/virtio/virtgpu_drv.c
index c50868753132..6d7092251d20 100644
--- a/drivers/gpu/drm/virtio/virtgpu_drv.c
+++ b/drivers/gpu/drm/virtio/virtgpu_drv.c
@@ -182,17 +182,7 @@ MODULE_AUTHOR("Dave Airlie <airlied@redhat.com>");
 MODULE_AUTHOR("Gerd Hoffmann <kraxel@redhat.com>");
 MODULE_AUTHOR("Alon Levy");
 
-static const struct file_operations virtio_gpu_driver_fops = {
-	.owner = THIS_MODULE,
-	.open = drm_open,
-	.mmap = virtio_gpu_mmap,
-	.poll = drm_poll,
-	.read = drm_read,
-	.unlocked_ioctl	= drm_ioctl,
-	.release = drm_release,
-	.compat_ioctl = drm_compat_ioctl,
-	.llseek = noop_llseek,
-};
+DEFINE_DRM_GEM_SHMEM_FOPS(virtio_gpu_driver_fops);
 
 static struct drm_driver driver = {
 	.driver_features = DRIVER_MODESET | DRIVER_GEM | DRIVER_PRIME | DRIVER_RENDER | DRIVER_ATOMIC,
@@ -209,15 +199,9 @@ static struct drm_driver driver = {
 	.prime_fd_to_handle = drm_gem_prime_fd_to_handle,
 	.gem_prime_export = drm_gem_prime_export,
 	.gem_prime_import = drm_gem_prime_import,
-	.gem_prime_get_sg_table = virtgpu_gem_prime_get_sg_table,
 	.gem_prime_import_sg_table = virtgpu_gem_prime_import_sg_table,
-	.gem_prime_vmap = virtgpu_gem_prime_vmap,
-	.gem_prime_vunmap = virtgpu_gem_prime_vunmap,
-	.gem_prime_mmap = virtgpu_gem_prime_mmap,
 
-	.gem_free_object_unlocked = virtio_gpu_gem_free_object,
-	.gem_open_object = virtio_gpu_gem_object_open,
-	.gem_close_object = virtio_gpu_gem_object_close,
+	.gem_create_object = virtio_gpu_create_object,
 	.fops = &virtio_gpu_driver_fops,
 
 	.ioctls = virtio_gpu_ioctls,
diff --git a/drivers/gpu/drm/virtio/virtgpu_gem.c b/drivers/gpu/drm/virtio/virtgpu_gem.c
index 456cc382ce68..c5028b04cd92 100644
--- a/drivers/gpu/drm/virtio/virtgpu_gem.c
+++ b/drivers/gpu/drm/virtio/virtgpu_gem.c
@@ -26,14 +26,6 @@
 #include <drm/drmP.h>
 #include "virtgpu_drv.h"
 
-void virtio_gpu_gem_free_object(struct drm_gem_object *gem_obj)
-{
-	struct virtio_gpu_object *obj = gem_to_virtio_gpu_obj(gem_obj);
-
-	if (obj)
-		virtio_gpu_object_unref(&obj);
-}
-
 struct virtio_gpu_object*
 virtio_gpu_alloc_object(struct drm_device *dev,
 			struct virtio_gpu_object_params *params,
@@ -64,16 +56,16 @@ int virtio_gpu_gem_create(struct drm_file *file,
 	if (IS_ERR(obj))
 		return PTR_ERR(obj);
 
-	ret = drm_gem_handle_create(file, &obj->gem_base, &handle);
+	ret = drm_gem_handle_create(file, &obj->base.base, &handle);
 	if (ret) {
-		drm_gem_object_release(&obj->gem_base);
+		drm_gem_object_release(&obj->base.base);
 		return ret;
 	}
 
-	*obj_p = &obj->gem_base;
+	*obj_p = &obj->base.base;
 
 	/* drop reference from allocate - handle holds it now */
-	drm_gem_object_put_unlocked(&obj->gem_base);
+	drm_gem_object_put_unlocked(&obj->base.base);
 
 	*handle_p = handle;
 	return 0;
diff --git a/drivers/gpu/drm/virtio/virtgpu_ioctl.c b/drivers/gpu/drm/virtio/virtgpu_ioctl.c
index 54e1cf84097a..d574ca38ebaa 100644
--- a/drivers/gpu/drm/virtio/virtgpu_ioctl.c
+++ b/drivers/gpu/drm/virtio/virtgpu_ioctl.c
@@ -27,7 +27,6 @@
 
 #include <drm/drmP.h>
 #include <drm/virtgpu_drm.h>
-#include <drm/ttm/ttm_execbuf_util.h>
 #include <linux/sync_file.h>
 
 #include "virtgpu_drv.h"
@@ -269,7 +268,7 @@ static int virtio_gpu_resource_create_ioctl(struct drm_device *dev, void *data,
 	dma_fence_put(&fence->f);
 	if (IS_ERR(qobj))
 		return PTR_ERR(qobj);
-	obj = &qobj->gem_base;
+	obj = &qobj->base.base;
 
 	ret = drm_gem_handle_create(file_priv, obj, &handle);
 	if (ret) {
@@ -296,7 +295,7 @@ static int virtio_gpu_resource_info_ioctl(struct drm_device *dev, void *data,
 
 	qobj = gem_to_virtio_gpu_obj(gobj);
 
-	ri->size = qobj->gem_base.size;
+	ri->size = qobj->base.base.size;
 	ri->res_handle = qobj->hw_res_handle;
 	drm_gem_object_put_unlocked(gobj);
 	return 0;
@@ -309,7 +308,6 @@ static int virtio_gpu_transfer_from_host_ioctl(struct drm_device *dev,
 	struct virtio_gpu_device *vgdev = dev->dev_private;
 	struct virtio_gpu_fpriv *vfpriv = file->driver_priv;
 	struct drm_virtgpu_3d_transfer_from_host *args = data;
-	struct ttm_operation_ctx ctx = { true, false };
 	struct drm_gem_object *gobj = NULL;
 	struct virtio_gpu_object *qobj = NULL;
 	struct virtio_gpu_fence *fence;
@@ -330,10 +328,6 @@ static int virtio_gpu_transfer_from_host_ioctl(struct drm_device *dev,
 	if (ret)
 		goto out;
 
-	ret = ttm_bo_validate(&qobj->tbo, &qobj->placement, &ctx);
-	if (unlikely(ret))
-		goto out_unres;
-
 	convert_to_hw_box(&box, &args->box);
 
 	fence = virtio_gpu_fence_alloc(vgdev);
@@ -345,7 +339,7 @@ static int virtio_gpu_transfer_from_host_ioctl(struct drm_device *dev,
 		(vgdev, qobj->hw_res_handle,
 		 vfpriv->ctx_id, offset, args->level,
 		 &box, fence);
-	reservation_object_add_excl_fence(qobj->tbo.resv,
+	reservation_object_add_excl_fence(qobj->base.base.resv,
 					  &fence->f);
 
 	dma_fence_put(&fence->f);
@@ -362,7 +356,6 @@ static int virtio_gpu_transfer_to_host_ioctl(struct drm_device *dev, void *data,
 	struct virtio_gpu_device *vgdev = dev->dev_private;
 	struct virtio_gpu_fpriv *vfpriv = file->driver_priv;
 	struct drm_virtgpu_3d_transfer_to_host *args = data;
-	struct ttm_operation_ctx ctx = { true, false };
 	struct drm_gem_object *gobj = NULL;
 	struct virtio_gpu_object *qobj = NULL;
 	struct virtio_gpu_fence *fence;
@@ -380,10 +373,6 @@ static int virtio_gpu_transfer_to_host_ioctl(struct drm_device *dev, void *data,
 	if (ret)
 		goto out;
 
-	ret = ttm_bo_validate(&qobj->tbo, &qobj->placement, &ctx);
-	if (unlikely(ret))
-		goto out_unres;
-
 	convert_to_hw_box(&box, &args->box);
 	if (!vgdev->has_virgl_3d) {
 		virtio_gpu_cmd_transfer_to_host_2d
@@ -399,7 +388,7 @@ static int virtio_gpu_transfer_to_host_ioctl(struct drm_device *dev, void *data,
 			(vgdev, qobj,
 			 vfpriv ? vfpriv->ctx_id : 0, offset,
 			 args->level, &box, fence);
-		reservation_object_add_excl_fence(qobj->tbo.resv,
+		reservation_object_add_excl_fence(qobj->base.base.resv,
 						  &fence->f);
 		dma_fence_put(&fence->f);
 	}
diff --git a/drivers/gpu/drm/virtio/virtgpu_kms.c b/drivers/gpu/drm/virtio/virtgpu_kms.c
index 84b6a6bf00c6..0bc6abaeafca 100644
--- a/drivers/gpu/drm/virtio/virtgpu_kms.c
+++ b/drivers/gpu/drm/virtio/virtgpu_kms.c
@@ -171,12 +171,6 @@ int virtio_gpu_init(struct drm_device *dev)
 		goto err_vbufs;
 	}
 
-	ret = virtio_gpu_ttm_init(vgdev);
-	if (ret) {
-		DRM_ERROR("failed to init ttm %d\n", ret);
-		goto err_ttm;
-	}
-
 	/* get display info */
 	virtio_cread(vgdev->vdev, struct virtio_gpu_config,
 		     num_scanouts, &num_scanouts);
@@ -208,8 +202,6 @@ int virtio_gpu_init(struct drm_device *dev)
 	return 0;
 
 err_scanouts:
-	virtio_gpu_ttm_fini(vgdev);
-err_ttm:
 	virtio_gpu_free_vbufs(vgdev);
 err_vbufs:
 	vgdev->vdev->config->del_vqs(vgdev->vdev);
@@ -240,7 +232,6 @@ void virtio_gpu_deinit(struct drm_device *dev)
 	vgdev->vdev->config->del_vqs(vgdev->vdev);
 
 	virtio_gpu_modeset_fini(vgdev);
-	virtio_gpu_ttm_fini(vgdev);
 	virtio_gpu_free_vbufs(vgdev);
 	virtio_gpu_cleanup_cap_cache(vgdev);
 	kfree(vgdev->capsets);
diff --git a/drivers/gpu/drm/virtio/virtgpu_object.c b/drivers/gpu/drm/virtio/virtgpu_object.c
index 90642907aa5c..dc9d0b7171a4 100644
--- a/drivers/gpu/drm/virtio/virtgpu_object.c
+++ b/drivers/gpu/drm/virtio/virtgpu_object.c
@@ -23,8 +23,6 @@
  * WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
  */
 
-#include <drm/ttm/ttm_execbuf_util.h>
-
 #include "virtgpu_drv.h"
 
 static int virtio_gpu_resource_id_get(struct virtio_gpu_device *vgdev,
@@ -57,39 +55,44 @@ static void virtio_gpu_resource_id_put(struct virtio_gpu_device *vgdev, uint32_t
 #endif
 }
 
-static void virtio_gpu_ttm_bo_destroy(struct ttm_buffer_object *tbo)
+static void virtio_gpu_free_object(struct drm_gem_object *obj)
 {
-	struct virtio_gpu_object *bo;
-	struct virtio_gpu_device *vgdev;
-
-	bo = container_of(tbo, struct virtio_gpu_object, tbo);
-	vgdev = (struct virtio_gpu_device *)bo->gem_base.dev->dev_private;
+	struct virtio_gpu_object *bo = gem_to_virtio_gpu_obj(obj);
+	struct virtio_gpu_device *vgdev = bo->base.base.dev->dev_private;
 
+	if (bo->pages)
+		virtio_gpu_object_detach(vgdev, bo);
 	if (bo->created)
 		virtio_gpu_cmd_unref_resource(vgdev, bo->hw_res_handle);
-	if (bo->pages)
-		virtio_gpu_object_free_sg_table(bo);
-	if (bo->vmap)
-		virtio_gpu_object_kunmap(bo);
-	drm_gem_object_release(&bo->gem_base);
 	virtio_gpu_resource_id_put(vgdev, bo->hw_res_handle);
-	kfree(bo);
+
+	drm_gem_shmem_free_object(obj);
 }
 
-static void virtio_gpu_init_ttm_placement(struct virtio_gpu_object *vgbo)
+static const struct drm_gem_object_funcs virtio_gpu_gem_funcs = {
+	.free = virtio_gpu_free_object,
+	.open = virtio_gpu_gem_object_open,
+	.close = virtio_gpu_gem_object_close,
+
+	.print_info = drm_gem_shmem_print_info,
+	.pin = drm_gem_shmem_pin,
+	.unpin = drm_gem_shmem_unpin,
+	.get_sg_table = drm_gem_shmem_get_sg_table,
+	.vmap = drm_gem_shmem_vmap,
+	.vunmap = drm_gem_shmem_vunmap,
+	.vm_ops = &drm_gem_shmem_vm_ops,
+};
+
+struct drm_gem_object *virtio_gpu_create_object(struct drm_device *dev,
+						size_t size)
 {
-	u32 c = 1;
+	struct virtio_gpu_object *bo;
 
-	vgbo->placement.placement = &vgbo->placement_code;
-	vgbo->placement.busy_placement = &vgbo->placement_code;
-	vgbo->placement_code.fpfn = 0;
-	vgbo->placement_code.lpfn = 0;
-	vgbo->placement_code.flags =
-		TTM_PL_MASK_CACHING | TTM_PL_FLAG_TT |
-		TTM_PL_FLAG_NO_EVICT;
-	vgbo->placement.num_placement = c;
-	vgbo->placement.num_busy_placement = c;
+	bo = kzalloc(sizeof(*bo), GFP_KERNEL);
+	if (!bo)
+		return NULL;
 
+	return &bo->base.base;
 }
 
 int virtio_gpu_object_create(struct virtio_gpu_device *vgdev,
@@ -98,28 +101,22 @@ int virtio_gpu_object_create(struct virtio_gpu_device *vgdev,
 			     struct virtio_gpu_fence *fence)
 {
 	struct virtio_gpu_object_array *objs = NULL;
+	struct drm_gem_shmem_object *shmem_obj;
 	struct virtio_gpu_object *bo;
-	size_t acc_size;
 	int ret;
 
 	*bo_ptr = NULL;
 
-	acc_size = ttm_bo_dma_acc_size(&vgdev->mman.bdev, params->size,
-				       sizeof(struct virtio_gpu_object));
+	params->size = roundup(params->size, PAGE_SIZE);
+	shmem_obj = drm_gem_shmem_create(vgdev->ddev, params->size);
+	if (IS_ERR(shmem_obj))
+		return PTR_ERR(shmem_obj);
+	bo = gem_to_virtio_gpu_obj(&shmem_obj->base);
+	bo->base.base.funcs = &virtio_gpu_gem_funcs;
 
-	bo = kzalloc(sizeof(struct virtio_gpu_object), GFP_KERNEL);
-	if (bo == NULL)
-		return -ENOMEM;
 	ret = virtio_gpu_resource_id_get(vgdev, &bo->hw_res_handle);
 	if (ret < 0) {
-		kfree(bo);
-		return ret;
-	}
-	params->size = roundup(params->size, PAGE_SIZE);
-	ret = drm_gem_object_init(vgdev->ddev, &bo->gem_base, params->size);
-	if (ret != 0) {
-		virtio_gpu_resource_id_put(vgdev, bo->hw_res_handle);
-		kfree(bo);
+		drm_gem_shmem_free_object(&shmem_obj->base);
 		return ret;
 	}
 	bo->dumb = params->dumb;
@@ -128,7 +125,7 @@ int virtio_gpu_object_create(struct virtio_gpu_device *vgdev,
 		struct ww_acquire_ctx ticket;
 
 		objs = virtio_gpu_array_alloc(1);
-		objs->objs[0] = &bo->gem_base;
+		objs->objs[0] = &bo->base.base;
 		drm_gem_object_get(objs->objs[0]);
 
 		ret = drm_gem_lock_reservations(objs->objs, objs->nents,
@@ -147,75 +144,12 @@ int virtio_gpu_object_create(struct virtio_gpu_device *vgdev,
 					       fence, objs);
 	}
 
-	virtio_gpu_init_ttm_placement(bo);
-	ret = ttm_bo_init(&vgdev->mman.bdev, &bo->tbo, params->size,
-			  ttm_bo_type_device, &bo->placement, 0,
-			  true, acc_size, NULL,
-			  bo->gem_base.resv,
-			  &virtio_gpu_ttm_bo_destroy);
-	/* ttm_bo_init failure will call the destroy */
-	if (ret != 0)
+	ret = virtio_gpu_object_attach(vgdev, bo, NULL);
+	if (ret != 0) {
+		virtio_gpu_free_object(&shmem_obj->base);
 		return ret;
+	}
 
 	*bo_ptr = bo;
 	return 0;
 }
-
-void virtio_gpu_object_kunmap(struct virtio_gpu_object *bo)
-{
-	bo->vmap = NULL;
-	ttm_bo_kunmap(&bo->kmap);
-}
-
-int virtio_gpu_object_kmap(struct virtio_gpu_object *bo)
-{
-	bool is_iomem;
-	int r;
-
-	WARN_ON(bo->vmap);
-
-	r = ttm_bo_kmap(&bo->tbo, 0, bo->tbo.num_pages, &bo->kmap);
-	if (r)
-		return r;
-	bo->vmap = ttm_kmap_obj_virtual(&bo->kmap, &is_iomem);
-	return 0;
-}
-
-int virtio_gpu_object_get_sg_table(struct virtio_gpu_device *qdev,
-				   struct virtio_gpu_object *bo)
-{
-	int ret;
-	struct page **pages = bo->tbo.ttm->pages;
-	int nr_pages = bo->tbo.num_pages;
-	struct ttm_operation_ctx ctx = {
-		.interruptible = false,
-		.no_wait_gpu = false
-	};
-
-	/* wtf swapping */
-	if (bo->pages)
-		return 0;
-
-	if (bo->tbo.ttm->state == tt_unpopulated)
-		bo->tbo.ttm->bdev->driver->ttm_tt_populate(bo->tbo.ttm, &ctx);
-	bo->pages = kmalloc(sizeof(struct sg_table), GFP_KERNEL);
-	if (!bo->pages)
-		goto out;
-
-	ret = sg_alloc_table_from_pages(bo->pages, pages, nr_pages, 0,
-					nr_pages << PAGE_SHIFT, GFP_KERNEL);
-	if (ret)
-		goto out;
-	return 0;
-out:
-	kfree(bo->pages);
-	bo->pages = NULL;
-	return -ENOMEM;
-}
-
-void virtio_gpu_object_free_sg_table(struct virtio_gpu_object *bo)
-{
-	sg_free_table(bo->pages);
-	kfree(bo->pages);
-	bo->pages = NULL;
-}
diff --git a/drivers/gpu/drm/virtio/virtgpu_prime.c b/drivers/gpu/drm/virtio/virtgpu_prime.c
index 8fbf71bd0c5e..18a155cd08d5 100644
--- a/drivers/gpu/drm/virtio/virtgpu_prime.c
+++ b/drivers/gpu/drm/virtio/virtgpu_prime.c
@@ -28,46 +28,9 @@
  * device that might share buffers with virtgpu
  */
 
-struct sg_table *virtgpu_gem_prime_get_sg_table(struct drm_gem_object *obj)
-{
-	struct virtio_gpu_object *bo = gem_to_virtio_gpu_obj(obj);
-
-	if (!bo->tbo.ttm->pages || !bo->tbo.ttm->num_pages)
-		/* should not happen */
-		return ERR_PTR(-EINVAL);
-
-	return drm_prime_pages_to_sg(bo->tbo.ttm->pages,
-				     bo->tbo.ttm->num_pages);
-}
-
 struct drm_gem_object *virtgpu_gem_prime_import_sg_table(
 	struct drm_device *dev, struct dma_buf_attachment *attach,
 	struct sg_table *table)
 {
 	return ERR_PTR(-ENODEV);
 }
-
-void *virtgpu_gem_prime_vmap(struct drm_gem_object *obj)
-{
-	struct virtio_gpu_object *bo = gem_to_virtio_gpu_obj(obj);
-	int ret;
-
-	ret = virtio_gpu_object_kmap(bo);
-	if (ret)
-		return NULL;
-	return bo->vmap;
-}
-
-void virtgpu_gem_prime_vunmap(struct drm_gem_object *obj, void *vaddr)
-{
-	virtio_gpu_object_kunmap(gem_to_virtio_gpu_obj(obj));
-}
-
-int virtgpu_gem_prime_mmap(struct drm_gem_object *obj,
-			   struct vm_area_struct *vma)
-{
-	struct virtio_gpu_object *bo = gem_to_virtio_gpu_obj(obj);
-
-	bo->gem_base.vma_node.vm_node.start = bo->tbo.vma_node.vm_node.start;
-	return drm_gem_prime_mmap(obj, vma);
-}
diff --git a/drivers/gpu/drm/virtio/virtgpu_ttm.c b/drivers/gpu/drm/virtio/virtgpu_ttm.c
deleted file mode 100644
index 300ef3a83538..000000000000
--- a/drivers/gpu/drm/virtio/virtgpu_ttm.c
+++ /dev/null
@@ -1,304 +0,0 @@
-/*
- * Copyright (C) 2015 Red Hat, Inc.
- * All Rights Reserved.
- *
- * Authors:
- *    Dave Airlie
- *    Alon Levy
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * the rights to use, copy, modify, merge, publish, distribute, sublicense,
- * and/or sell copies of the Software, and to permit persons to whom the
- * Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in
- * all copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
- * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
- * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
- * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
- * OTHER DEALINGS IN THE SOFTWARE.
- */
-
-#include <drm/ttm/ttm_bo_api.h>
-#include <drm/ttm/ttm_bo_driver.h>
-#include <drm/ttm/ttm_placement.h>
-#include <drm/ttm/ttm_page_alloc.h>
-#include <drm/ttm/ttm_module.h>
-#include <drm/drmP.h>
-#include <drm/drm.h>
-#include <drm/virtgpu_drm.h>
-#include "virtgpu_drv.h"
-
-#include <linux/delay.h>
-
-static struct
-virtio_gpu_device *virtio_gpu_get_vgdev(struct ttm_bo_device *bdev)
-{
-	struct virtio_gpu_mman *mman;
-	struct virtio_gpu_device *vgdev;
-
-	mman = container_of(bdev, struct virtio_gpu_mman, bdev);
-	vgdev = container_of(mman, struct virtio_gpu_device, mman);
-	return vgdev;
-}
-
-int virtio_gpu_mmap(struct file *filp, struct vm_area_struct *vma)
-{
-	struct drm_file *file_priv;
-	struct virtio_gpu_device *vgdev;
-	int r;
-
-	file_priv = filp->private_data;
-	vgdev = file_priv->minor->dev->dev_private;
-	if (vgdev == NULL) {
-		DRM_ERROR(
-		 "filp->private_data->minor->dev->dev_private == NULL\n");
-		return -EINVAL;
-	}
-	r = ttm_bo_mmap(filp, vma, &vgdev->mman.bdev);
-
-	return r;
-}
-
-static int virtio_gpu_invalidate_caches(struct ttm_bo_device *bdev,
-					uint32_t flags)
-{
-	return 0;
-}
-
-static int ttm_bo_man_get_node(struct ttm_mem_type_manager *man,
-			       struct ttm_buffer_object *bo,
-			       const struct ttm_place *place,
-			       struct ttm_mem_reg *mem)
-{
-	mem->mm_node = (void *)1;
-	return 0;
-}
-
-static void ttm_bo_man_put_node(struct ttm_mem_type_manager *man,
-				struct ttm_mem_reg *mem)
-{
-	mem->mm_node = (void *)NULL;
-}
-
-static int ttm_bo_man_init(struct ttm_mem_type_manager *man,
-			   unsigned long p_size)
-{
-	return 0;
-}
-
-static int ttm_bo_man_takedown(struct ttm_mem_type_manager *man)
-{
-	return 0;
-}
-
-static void ttm_bo_man_debug(struct ttm_mem_type_manager *man,
-			     struct drm_printer *printer)
-{
-}
-
-static const struct ttm_mem_type_manager_func virtio_gpu_bo_manager_func = {
-	.init = ttm_bo_man_init,
-	.takedown = ttm_bo_man_takedown,
-	.get_node = ttm_bo_man_get_node,
-	.put_node = ttm_bo_man_put_node,
-	.debug = ttm_bo_man_debug
-};
-
-static int virtio_gpu_init_mem_type(struct ttm_bo_device *bdev, uint32_t type,
-				    struct ttm_mem_type_manager *man)
-{
-	switch (type) {
-	case TTM_PL_SYSTEM:
-		/* System memory */
-		man->flags = TTM_MEMTYPE_FLAG_MAPPABLE;
-		man->available_caching = TTM_PL_MASK_CACHING;
-		man->default_caching = TTM_PL_FLAG_CACHED;
-		break;
-	case TTM_PL_TT:
-		man->func = &virtio_gpu_bo_manager_func;
-		man->flags = TTM_MEMTYPE_FLAG_MAPPABLE;
-		man->available_caching = TTM_PL_MASK_CACHING;
-		man->default_caching = TTM_PL_FLAG_CACHED;
-		break;
-	default:
-		DRM_ERROR("Unsupported memory type %u\n", (unsigned int)type);
-		return -EINVAL;
-	}
-	return 0;
-}
-
-static void virtio_gpu_evict_flags(struct ttm_buffer_object *bo,
-				struct ttm_placement *placement)
-{
-	static const struct ttm_place placements = {
-		.fpfn  = 0,
-		.lpfn  = 0,
-		.flags = TTM_PL_MASK_CACHING | TTM_PL_FLAG_SYSTEM,
-	};
-
-	placement->placement = &placements;
-	placement->busy_placement = &placements;
-	placement->num_placement = 1;
-	placement->num_busy_placement = 1;
-}
-
-static int virtio_gpu_verify_access(struct ttm_buffer_object *bo,
-				    struct file *filp)
-{
-	return 0;
-}
-
-static int virtio_gpu_ttm_io_mem_reserve(struct ttm_bo_device *bdev,
-					 struct ttm_mem_reg *mem)
-{
-	struct ttm_mem_type_manager *man = &bdev->man[mem->mem_type];
-
-	mem->bus.addr = NULL;
-	mem->bus.offset = 0;
-	mem->bus.size = mem->num_pages << PAGE_SHIFT;
-	mem->bus.base = 0;
-	mem->bus.is_iomem = false;
-	if (!(man->flags & TTM_MEMTYPE_FLAG_MAPPABLE))
-		return -EINVAL;
-	switch (mem->mem_type) {
-	case TTM_PL_SYSTEM:
-	case TTM_PL_TT:
-		/* system memory */
-		return 0;
-	default:
-		return -EINVAL;
-	}
-	return 0;
-}
-
-static void virtio_gpu_ttm_io_mem_free(struct ttm_bo_device *bdev,
-				       struct ttm_mem_reg *mem)
-{
-}
-
-/*
- * TTM backend functions.
- */
-struct virtio_gpu_ttm_tt {
-	struct ttm_dma_tt		ttm;
-	struct virtio_gpu_object        *obj;
-};
-
-static int virtio_gpu_ttm_tt_bind(struct ttm_tt *ttm,
-				  struct ttm_mem_reg *bo_mem)
-{
-	struct virtio_gpu_ttm_tt *gtt =
-		container_of(ttm, struct virtio_gpu_ttm_tt, ttm.ttm);
-	struct virtio_gpu_device *vgdev =
-		virtio_gpu_get_vgdev(gtt->obj->tbo.bdev);
-
-	virtio_gpu_object_attach(vgdev, gtt->obj, NULL);
-	return 0;
-}
-
-static int virtio_gpu_ttm_tt_unbind(struct ttm_tt *ttm)
-{
-	struct virtio_gpu_ttm_tt *gtt =
-		container_of(ttm, struct virtio_gpu_ttm_tt, ttm.ttm);
-	struct virtio_gpu_device *vgdev =
-		virtio_gpu_get_vgdev(gtt->obj->tbo.bdev);
-
-	virtio_gpu_object_detach(vgdev, gtt->obj);
-	return 0;
-}
-
-static void virtio_gpu_ttm_tt_destroy(struct ttm_tt *ttm)
-{
-	struct virtio_gpu_ttm_tt *gtt =
-		container_of(ttm, struct virtio_gpu_ttm_tt, ttm.ttm);
-
-	ttm_dma_tt_fini(&gtt->ttm);
-	kfree(gtt);
-}
-
-static struct ttm_backend_func virtio_gpu_tt_func = {
-	.bind = &virtio_gpu_ttm_tt_bind,
-	.unbind = &virtio_gpu_ttm_tt_unbind,
-	.destroy = &virtio_gpu_ttm_tt_destroy,
-};
-
-static struct ttm_tt *virtio_gpu_ttm_tt_create(struct ttm_buffer_object *bo,
-					       uint32_t page_flags)
-{
-	struct virtio_gpu_device *vgdev;
-	struct virtio_gpu_ttm_tt *gtt;
-
-	vgdev = virtio_gpu_get_vgdev(bo->bdev);
-	gtt = kzalloc(sizeof(struct virtio_gpu_ttm_tt), GFP_KERNEL);
-	if (gtt == NULL)
-		return NULL;
-	gtt->ttm.ttm.func = &virtio_gpu_tt_func;
-	gtt->obj = container_of(bo, struct virtio_gpu_object, tbo);
-	if (ttm_dma_tt_init(&gtt->ttm, bo, page_flags)) {
-		kfree(gtt);
-		return NULL;
-	}
-	return &gtt->ttm.ttm;
-}
-
-static void virtio_gpu_bo_swap_notify(struct ttm_buffer_object *tbo)
-{
-	struct virtio_gpu_object *bo;
-
-	bo = container_of(tbo, struct virtio_gpu_object, tbo);
-
-	if (bo->pages)
-		virtio_gpu_object_free_sg_table(bo);
-}
-
-static struct ttm_bo_driver virtio_gpu_bo_driver = {
-	.ttm_tt_create = &virtio_gpu_ttm_tt_create,
-	.invalidate_caches = &virtio_gpu_invalidate_caches,
-	.init_mem_type = &virtio_gpu_init_mem_type,
-	.eviction_valuable = ttm_bo_eviction_valuable,
-	.evict_flags = &virtio_gpu_evict_flags,
-	.verify_access = &virtio_gpu_verify_access,
-	.io_mem_reserve = &virtio_gpu_ttm_io_mem_reserve,
-	.io_mem_free = &virtio_gpu_ttm_io_mem_free,
-	.swap_notify = &virtio_gpu_bo_swap_notify,
-};
-
-int virtio_gpu_ttm_init(struct virtio_gpu_device *vgdev)
-{
-	int r;
-
-	/* No others user of address space so set it to 0 */
-	r = ttm_bo_device_init(&vgdev->mman.bdev,
-			       &virtio_gpu_bo_driver,
-			       vgdev->ddev->anon_inode->i_mapping,
-			       false);
-	if (r) {
-		DRM_ERROR("failed initializing buffer object driver(%d).\n", r);
-		goto err_dev_init;
-	}
-
-	r = ttm_bo_init_mm(&vgdev->mman.bdev, TTM_PL_TT, 0);
-	if (r) {
-		DRM_ERROR("Failed initializing GTT heap.\n");
-		goto err_mm_init;
-	}
-	return 0;
-
-err_mm_init:
-	ttm_bo_device_release(&vgdev->mman.bdev);
-err_dev_init:
-	return r;
-}
-
-void virtio_gpu_ttm_fini(struct virtio_gpu_device *vgdev)
-{
-	ttm_bo_device_release(&vgdev->mman.bdev);
-	DRM_INFO("virtio_gpu: ttm finalized\n");
-}
diff --git a/drivers/gpu/drm/virtio/virtgpu_vq.c b/drivers/gpu/drm/virtio/virtgpu_vq.c
index 6e2b287a7e4b..352fd0689406 100644
--- a/drivers/gpu/drm/virtio/virtgpu_vq.c
+++ b/drivers/gpu/drm/virtio/virtgpu_vq.c
@@ -974,17 +974,21 @@ int virtio_gpu_object_attach(struct virtio_gpu_device *vgdev,
 	bool use_dma_api = !virtio_has_iommu_quirk(vgdev->vdev);
 	struct virtio_gpu_mem_entry *ents;
 	struct scatterlist *sg;
-	int si, nents;
+	int si, nents, ret;
 
 	if (WARN_ON_ONCE(!obj->created))
 		return -EINVAL;
+	if (WARN_ON_ONCE(obj->pages))
+		return -EINVAL;
 
-	if (!obj->pages) {
-		int ret;
+	ret = drm_gem_shmem_pin(&obj->base.base);
+	if (ret < 0)
+		return -EINVAL;
 
-		ret = virtio_gpu_object_get_sg_table(vgdev, obj);
-		if (ret)
-			return ret;
+	obj->pages = drm_gem_shmem_get_sg_table(&obj->base.base);
+	if (obj->pages == NULL) {
+		drm_gem_shmem_unpin(&obj->base.base);
+		return -EINVAL;
 	}
 
 	if (use_dma_api) {
@@ -1023,6 +1027,9 @@ void virtio_gpu_object_detach(struct virtio_gpu_device *vgdev,
 {
 	bool use_dma_api = !virtio_has_iommu_quirk(vgdev->vdev);
 
+	if (WARN_ON_ONCE(!obj->pages))
+		return;
+
 	if (use_dma_api && obj->mapped) {
 		struct virtio_gpu_fence *fence = virtio_gpu_fence_alloc(vgdev);
 		/* detach backing and wait for the host process it ... */
@@ -1038,6 +1045,11 @@ void virtio_gpu_object_detach(struct virtio_gpu_device *vgdev,
 	} else {
 		virtio_gpu_cmd_resource_inval_backing(vgdev, obj->hw_res_handle, NULL);
 	}
+
+	sg_free_table(obj->pages);
+	obj->pages = NULL;
+
+	drm_gem_shmem_unpin(&obj->base.base);
 }
 
 void virtio_gpu_cursor_ping(struct virtio_gpu_device *vgdev,
diff --git a/drivers/gpu/drm/virtio/Kconfig b/drivers/gpu/drm/virtio/Kconfig
index ba36e933bb49..eff3047052d4 100644
--- a/drivers/gpu/drm/virtio/Kconfig
+++ b/drivers/gpu/drm/virtio/Kconfig
@@ -3,7 +3,7 @@ config DRM_VIRTIO_GPU
 	tristate "Virtio GPU driver"
 	depends on DRM && VIRTIO && MMU
 	select DRM_KMS_HELPER
-	select DRM_TTM
+	select DRM_GEM_SHMEM_HELPER
 	help
 	   This is the virtual GPU driver for virtio.  It can be used with
 	   QEMU based VMMs (like KVM or Xen).
diff --git a/drivers/gpu/drm/virtio/Makefile b/drivers/gpu/drm/virtio/Makefile
index 458e606a936f..92aa2b3d349d 100644
--- a/drivers/gpu/drm/virtio/Makefile
+++ b/drivers/gpu/drm/virtio/Makefile
@@ -4,7 +4,7 @@
 # Direct Rendering Infrastructure (DRI) in XFree86 4.1.0 and higher.
 
 virtio-gpu-y := virtgpu_drv.o virtgpu_kms.o virtgpu_gem.o \
-	virtgpu_display.o virtgpu_vq.o virtgpu_ttm.o \
+	virtgpu_display.o virtgpu_vq.o \
 	virtgpu_fence.o virtgpu_object.o virtgpu_debugfs.o virtgpu_plane.o \
 	virtgpu_ioctl.o virtgpu_prime.o virtgpu_trace_points.o
 
-- 
2.18.1

[PATCH v4 12/12] drm/virtio: remove virtio_gpu_alloc_object

[Gerd Hoffmann]

Thin wrapper around virtio_gpu_object_create(),
but calling that directly works equally well.

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Acked-by: Daniel Vetter <daniel.vetter@ffwll.ch>
---
 drivers/gpu/drm/virtio/virtgpu_drv.h   |  4 ----
 drivers/gpu/drm/virtio/virtgpu_gem.c   | 23 ++++-------------------
 drivers/gpu/drm/virtio/virtgpu_ioctl.c |  6 +++---
 3 files changed, 7 insertions(+), 26 deletions(-)

diff --git a/drivers/gpu/drm/virtio/virtgpu_drv.h b/drivers/gpu/drm/virtio/virtgpu_drv.h
index 1d589de34449..3a593129981a 100644
--- a/drivers/gpu/drm/virtio/virtgpu_drv.h
+++ b/drivers/gpu/drm/virtio/virtgpu_drv.h
@@ -228,10 +228,6 @@ int virtio_gpu_gem_object_open(struct drm_gem_object *obj,
 			       struct drm_file *file);
 void virtio_gpu_gem_object_close(struct drm_gem_object *obj,
 				 struct drm_file *file);
-struct virtio_gpu_object*
-virtio_gpu_alloc_object(struct drm_device *dev,
-			struct virtio_gpu_object_params *params,
-			struct virtio_gpu_fence *fence);
 int virtio_gpu_mode_dumb_create(struct drm_file *file_priv,
 				struct drm_device *dev,
 				struct drm_mode_create_dumb *args);
diff --git a/drivers/gpu/drm/virtio/virtgpu_gem.c b/drivers/gpu/drm/virtio/virtgpu_gem.c
index c5028b04cd92..f5765b94811a 100644
--- a/drivers/gpu/drm/virtio/virtgpu_gem.c
+++ b/drivers/gpu/drm/virtio/virtgpu_gem.c
@@ -26,35 +26,20 @@
 #include <drm/drmP.h>
 #include "virtgpu_drv.h"
 
-struct virtio_gpu_object*
-virtio_gpu_alloc_object(struct drm_device *dev,
-			struct virtio_gpu_object_params *params,
-			struct virtio_gpu_fence *fence)
-{
-	struct virtio_gpu_device *vgdev = dev->dev_private;
-	struct virtio_gpu_object *obj;
-	int ret;
-
-	ret = virtio_gpu_object_create(vgdev, params, &obj, fence);
-	if (ret)
-		return ERR_PTR(ret);
-
-	return obj;
-}
-
 int virtio_gpu_gem_create(struct drm_file *file,
 			  struct drm_device *dev,
 			  struct virtio_gpu_object_params *params,
 			  struct drm_gem_object **obj_p,
 			  uint32_t *handle_p)
 {
+	struct virtio_gpu_device *vgdev = dev->dev_private;
 	struct virtio_gpu_object *obj;
 	int ret;
 	u32 handle;
 
-	obj = virtio_gpu_alloc_object(dev, params, NULL);
-	if (IS_ERR(obj))
-		return PTR_ERR(obj);
+	ret = virtio_gpu_object_create(vgdev, params, &obj, NULL);
+	if (ret < 0)
+		return ret;
 
 	ret = drm_gem_handle_create(file, &obj->base.base, &handle);
 	if (ret) {
diff --git a/drivers/gpu/drm/virtio/virtgpu_ioctl.c b/drivers/gpu/drm/virtio/virtgpu_ioctl.c
index d574ca38ebaa..c5f5af97d821 100644
--- a/drivers/gpu/drm/virtio/virtgpu_ioctl.c
+++ b/drivers/gpu/drm/virtio/virtgpu_ioctl.c
@@ -264,10 +264,10 @@ static int virtio_gpu_resource_create_ioctl(struct drm_device *dev, void *data,
 	fence = virtio_gpu_fence_alloc(vgdev);
 	if (!fence)
 		return -ENOMEM;
-	qobj = virtio_gpu_alloc_object(dev, &params, fence);
+	ret = virtio_gpu_object_create(vgdev, &params, &qobj, fence);
 	dma_fence_put(&fence->f);
-	if (IS_ERR(qobj))
-		return PTR_ERR(qobj);
+	if (ret < 0)
+		return ret;
 	obj = &qobj->base.base;
 
 	ret = drm_gem_handle_create(file_priv, obj, &handle);
-- 
2.18.1

_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/dri-devel

Re: [PATCH v4 02/12] drm/virtio: switch virtio_gpu_wait_ioctl() to gem helper.

[Chia-I Wu]

On Wed, Jun 19, 2019 at 11:07 PM Gerd Hoffmann <kraxel@redhat.com> wrote:
>
> Use drm_gem_reservation_object_wait() in virtio_gpu_wait_ioctl().
> This also makes the ioctl run lockless.
The userspace has a BO cache to avoid freeing BOs immediately but to
reuse them on next allocations.  The BO cache checks if a BO is busy
before reuse, and I am seeing a big negative perf impact because of
slow virtio_gpu_wait_ioctl.  I wonder if this helps.


>
> v2: use reservation_object_test_signaled_rcu for VIRTGPU_WAIT_NOWAIT.
>
> Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
> Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
> ---
>  drivers/gpu/drm/virtio/virtgpu_ioctl.c | 24 ++++++++++--------------
>  1 file changed, 10 insertions(+), 14 deletions(-)
>
> diff --git a/drivers/gpu/drm/virtio/virtgpu_ioctl.c b/drivers/gpu/drm/virtio/virtgpu_ioctl.c
> index ac60be9b5c19..313c770ea2c5 100644
> --- a/drivers/gpu/drm/virtio/virtgpu_ioctl.c
> +++ b/drivers/gpu/drm/virtio/virtgpu_ioctl.c
> @@ -464,23 +464,19 @@ static int virtio_gpu_wait_ioctl(struct drm_device *dev, void *data,
>                             struct drm_file *file)
>  {
>         struct drm_virtgpu_3d_wait *args = data;
> -       struct drm_gem_object *gobj = NULL;
> -       struct virtio_gpu_object *qobj = NULL;
> +       struct drm_gem_object *obj;
> +       long timeout = 15 * HZ;
>         int ret;
> -       bool nowait = false;
>
> -       gobj = drm_gem_object_lookup(file, args->handle);
> -       if (gobj == NULL)
> -               return -ENOENT;
> +       if (args->flags & VIRTGPU_WAIT_NOWAIT) {
> +               obj = drm_gem_object_lookup(file, args->handle);
Don't we need a NULL check here?
> +               ret = reservation_object_test_signaled_rcu(obj->resv, true);
> +               drm_gem_object_put_unlocked(obj);
> +               return ret ? 0 : -EBUSY;
> +       }
>
> -       qobj = gem_to_virtio_gpu_obj(gobj);
> -
> -       if (args->flags & VIRTGPU_WAIT_NOWAIT)
> -               nowait = true;
> -       ret = virtio_gpu_object_wait(qobj, nowait);
> -
> -       drm_gem_object_put_unlocked(gobj);
> -       return ret;
> +       return drm_gem_reservation_object_wait(file, args->handle,
> +                                              true, timeout);
>  }
>
>  static int virtio_gpu_get_caps_ioctl(struct drm_device *dev,
> --
> 2.18.1
>
> _______________________________________________
> dri-devel mailing list
> dri-devel@lists.freedesktop.org
> https://lists.freedesktop.org/mailman/listinfo/dri-devel

Re: [PATCH v4 07/12] drm/virtio: add virtio_gpu_object_array & helpers

[Chia-I Wu]

On Wed, Jun 19, 2019 at 11:08 PM Gerd Hoffmann <kraxel@redhat.com> wrote:
>
> Some helper functions to manage an array of gem objects.
>
> v4: make them virtio-private instead of generic helpers.
>
> Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
> ---
>  drivers/gpu/drm/virtio/virtgpu_drv.h | 10 ++++++
>  drivers/gpu/drm/virtio/virtgpu_gem.c | 50 ++++++++++++++++++++++++++++
>  2 files changed, 60 insertions(+)
>
> diff --git a/drivers/gpu/drm/virtio/virtgpu_drv.h b/drivers/gpu/drm/virtio/virtgpu_drv.h
> index 07f6001ea91e..98d646789d23 100644
> --- a/drivers/gpu/drm/virtio/virtgpu_drv.h
> +++ b/drivers/gpu/drm/virtio/virtgpu_drv.h
> @@ -84,6 +84,11 @@ struct virtio_gpu_object {
>  #define gem_to_virtio_gpu_obj(gobj) \
>         container_of((gobj), struct virtio_gpu_object, gem_base)
>
> +struct virtio_gpu_object_array {
> +       u32 nents;
> +       struct drm_gem_object *objs[];
> +};
> +
>  struct virtio_gpu_vbuffer;
>  struct virtio_gpu_device;
>
> @@ -251,6 +256,11 @@ int virtio_gpu_mode_dumb_mmap(struct drm_file *file_priv,
>                               struct drm_device *dev,
>                               uint32_t handle, uint64_t *offset_p);
>
> +struct virtio_gpu_object_array *virtio_gpu_array_alloc(u32 nents);
> +struct virtio_gpu_object_array*
> +virtio_gpu_array_from_handles(struct drm_file *drm_file, u32 *handles, u32 nents);
> +void virtio_gpu_array_put_free(struct virtio_gpu_object_array *objs);
> +
>  /* virtio vg */
>  int virtio_gpu_alloc_vbufs(struct virtio_gpu_device *vgdev);
>  void virtio_gpu_free_vbufs(struct virtio_gpu_device *vgdev);
> diff --git a/drivers/gpu/drm/virtio/virtgpu_gem.c b/drivers/gpu/drm/virtio/virtgpu_gem.c
> index 9c9ad3b14080..456cc382ce68 100644
> --- a/drivers/gpu/drm/virtio/virtgpu_gem.c
> +++ b/drivers/gpu/drm/virtio/virtgpu_gem.c
> @@ -169,3 +169,53 @@ void virtio_gpu_gem_object_close(struct drm_gem_object *obj,
>                                                 qobj->hw_res_handle);
>         virtio_gpu_object_unreserve(qobj);
>  }
> +
> +struct virtio_gpu_object_array *virtio_gpu_array_alloc(u32 nents)
> +{
> +       struct virtio_gpu_object_array *objs;
> +       size_t size = sizeof(*objs) + sizeof(objs->objs[0]) * nents;
> +
> +       objs = kzalloc(size, GFP_KERNEL);
It seems unlikely to get wrong.  Use kmalloc instead?
> +       if (!objs)
> +               return NULL;
> +
> +       objs->nents = nents;
> +       return objs;
> +}
> +
> +static void virtio_gpu_array_free(struct virtio_gpu_object_array *objs)
> +{
> +       kfree(objs);
> +}
> +
> +struct virtio_gpu_object_array*
> +virtio_gpu_array_from_handles(struct drm_file *drm_file, u32 *handles, u32 nents)
> +{
> +       struct virtio_gpu_object_array *objs;
> +       u32 i;
> +
> +       objs = virtio_gpu_array_alloc(nents);
> +       if (!objs)
> +               return NULL;
> +
> +       for (i = 0; i < nents; i++) {
> +               objs->objs[i] = drm_gem_object_lookup(drm_file, handles[i]);
> +               if (!objs->objs[i]) {
> +                       virtio_gpu_array_put_free(objs);
Set objs->nents to i before calling i915_gem_object_lookup;
> +                       return NULL;
Should we distinguish ENOMEM and ENOENT?
> +               }
> +       }
> +       return objs;
> +}
> +
> +void virtio_gpu_array_put_free(struct virtio_gpu_object_array *objs)
> +{
> +       u32 i;
> +
> +       for (i = 0; i < objs->nents; i++) {
> +               if (!objs->objs[i])
> +                       continue;
The check can be removed if the change in virtio_gpu_array_from_handles.
> +               drm_gem_object_put_unlocked(objs->objs[i]);
> +       }
> +       virtio_gpu_array_free(objs);
> +}
> --
> 2.18.1
>
> _______________________________________________
> dri-devel mailing list
> dri-devel@lists.freedesktop.org
> https://lists.freedesktop.org/mailman/listinfo/dri-devel

Re: [PATCH v4 08/12] drm/virtio: rework virtio_gpu_execbuffer_ioctl fencing

[Chia-I Wu]

On Wed, Jun 19, 2019 at 11:08 PM Gerd Hoffmann <kraxel@redhat.com> wrote:
>
> Use gem reservation helpers and direct reservation_object_* calls
> instead of ttm.
>
> v3: Also attach the array of gem objects to the virtio command buffer,
> so we can drop the object references in the completion callback.  Needed
> because ttm fence helpers grab a reference for us, but gem helpers
> don't.
>
> Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
> Acked-by: Daniel Vetter <daniel.vetter@ffwll.ch>
> ---
>  drivers/gpu/drm/virtio/virtgpu_drv.h   |  5 ++-
>  drivers/gpu/drm/virtio/virtgpu_ioctl.c | 62 +++++++++++---------------
>  drivers/gpu/drm/virtio/virtgpu_vq.c    | 16 ++++---
>  3 files changed, 40 insertions(+), 43 deletions(-)
>
> diff --git a/drivers/gpu/drm/virtio/virtgpu_drv.h b/drivers/gpu/drm/virtio/virtgpu_drv.h
> index 98d646789d23..65f5ce41c341 100644
> --- a/drivers/gpu/drm/virtio/virtgpu_drv.h
> +++ b/drivers/gpu/drm/virtio/virtgpu_drv.h
> @@ -120,9 +120,9 @@ struct virtio_gpu_vbuffer {
>
>         char *resp_buf;
>         int resp_size;
> -
>         virtio_gpu_resp_cb resp_cb;
>
> +       struct virtio_gpu_object_array *objs;
This can use a comment (e.g., objects referenced by the vbuffer)
>         struct list_head list;
>  };
>
> @@ -311,7 +311,8 @@ void virtio_gpu_cmd_context_detach_resource(struct virtio_gpu_device *vgdev,
>                                             uint32_t resource_id);
>  void virtio_gpu_cmd_submit(struct virtio_gpu_device *vgdev,
>                            void *data, uint32_t data_size,
> -                          uint32_t ctx_id, struct virtio_gpu_fence *fence);
> +                          uint32_t ctx_id, struct virtio_gpu_fence *fence,
> +                          struct virtio_gpu_object_array *objs);
Can we keep fence, which is updated, as the last parameter?
>  void virtio_gpu_cmd_transfer_from_host_3d(struct virtio_gpu_device *vgdev,
>                                           uint32_t resource_id, uint32_t ctx_id,
>                                           uint64_t offset, uint32_t level,
> diff --git a/drivers/gpu/drm/virtio/virtgpu_ioctl.c b/drivers/gpu/drm/virtio/virtgpu_ioctl.c
> index 5cffd2e54c04..c3901c397b8a 100644
> --- a/drivers/gpu/drm/virtio/virtgpu_ioctl.c
> +++ b/drivers/gpu/drm/virtio/virtgpu_ioctl.c
> @@ -105,14 +105,11 @@ static int virtio_gpu_execbuffer_ioctl(struct drm_device *dev, void *data,
>         struct drm_virtgpu_execbuffer *exbuf = data;
>         struct virtio_gpu_device *vgdev = dev->dev_private;
>         struct virtio_gpu_fpriv *vfpriv = drm_file->driver_priv;
> -       struct drm_gem_object *gobj;
>         struct virtio_gpu_fence *out_fence;
> -       struct virtio_gpu_object *qobj;
>         int ret;
>         uint32_t *bo_handles = NULL;
>         void __user *user_bo_handles = NULL;
> -       struct list_head validate_list;
> -       struct ttm_validate_buffer *buflist = NULL;
> +       struct virtio_gpu_object_array *buflist = NULL;
>         int i;
>         struct ww_acquire_ctx ticket;
>         struct sync_file *sync_file;
> @@ -155,15 +152,10 @@ static int virtio_gpu_execbuffer_ioctl(struct drm_device *dev, void *data,
>                         return out_fence_fd;
>         }
>
> -       INIT_LIST_HEAD(&validate_list);
>         if (exbuf->num_bo_handles) {
> -
>                 bo_handles = kvmalloc_array(exbuf->num_bo_handles,
> -                                          sizeof(uint32_t), GFP_KERNEL);
> -               buflist = kvmalloc_array(exbuf->num_bo_handles,
> -                                          sizeof(struct ttm_validate_buffer),
> -                                          GFP_KERNEL | __GFP_ZERO);
> -               if (!bo_handles || !buflist) {
> +                                           sizeof(uint32_t), GFP_KERNEL);
> +               if (!bo_handles) {
>                         ret = -ENOMEM;
>                         goto out_unused_fd;
>                 }
> @@ -175,25 +167,22 @@ static int virtio_gpu_execbuffer_ioctl(struct drm_device *dev, void *data,
>                         goto out_unused_fd;
>                 }
>
> -               for (i = 0; i < exbuf->num_bo_handles; i++) {
> -                       gobj = drm_gem_object_lookup(drm_file, bo_handles[i]);
> -                       if (!gobj) {
> -                               ret = -ENOENT;
> -                               goto out_unused_fd;
> -                       }
> -
> -                       qobj = gem_to_virtio_gpu_obj(gobj);
> -                       buflist[i].bo = &qobj->tbo;
> -
> -                       list_add(&buflist[i].head, &validate_list);
> +               buflist = virtio_gpu_array_from_handles(drm_file, bo_handles,
> +                                                       exbuf->num_bo_handles);
> +               if (!buflist) {
> +                       ret = -ENOENT;
> +                       goto out_unused_fd;
>                 }
>                 kvfree(bo_handles);
>                 bo_handles = NULL;
>         }
>
> -       ret = virtio_gpu_object_list_validate(&ticket, &validate_list);
> -       if (ret)
> -               goto out_free;
> +       if (buflist) {
> +               ret = drm_gem_lock_reservations(buflist->objs, buflist->nents,
> +                                               &ticket);
> +               if (ret)
> +                       goto out_unused_fd;
> +       }
>
>         buf = memdup_user(u64_to_user_ptr(exbuf->command), exbuf->size);
>         if (IS_ERR(buf)) {
> @@ -219,25 +208,26 @@ static int virtio_gpu_execbuffer_ioctl(struct drm_device *dev, void *data,
>                 fd_install(out_fence_fd, sync_file->file);
>         }
>
> +       if (buflist) {
> +               for (i = 0; i < exbuf->num_bo_handles; i++)
> +                       reservation_object_add_excl_fence(buflist->objs[i]->resv,
> +                                                         &out_fence->f);
> +               drm_gem_unlock_reservations(buflist->objs, buflist->nents,
> +                                           &ticket);
> +       }
We used to unlock after virtio_gpu_cmd_submit.

I guess, the fence is considered signaled (because its seqno is still
0) until after virtio_gpu_cmd_submit.  We probably don't want other
processes to see the semi-initialized fence.


> +
>         virtio_gpu_cmd_submit(vgdev, buf, exbuf->size,
> -                             vfpriv->ctx_id, out_fence);
> -
> -       ttm_eu_fence_buffer_objects(&ticket, &validate_list, &out_fence->f);
> -
> -       /* fence the command bo */
> -       virtio_gpu_unref_list(&validate_list);
> -       kvfree(buflist);
> +                             vfpriv->ctx_id, out_fence, buflist);
>         return 0;
>
>  out_memdup:
>         kfree(buf);
>  out_unresv:
> -       ttm_eu_backoff_reservation(&ticket, &validate_list);
> -out_free:
> -       virtio_gpu_unref_list(&validate_list);
Keeping out_free to free buflist seems just fine.
> +       drm_gem_unlock_reservations(buflist->objs, buflist->nents, &ticket);
>  out_unused_fd:
>         kvfree(bo_handles);
> -       kvfree(buflist);
> +       if (buflist)
> +               virtio_gpu_array_put_free(buflist);
>
>         if (out_fence_fd >= 0)
>                 put_unused_fd(out_fence_fd);
> diff --git a/drivers/gpu/drm/virtio/virtgpu_vq.c b/drivers/gpu/drm/virtio/virtgpu_vq.c
> index 6c1a90717535..dc2c2c003200 100644
> --- a/drivers/gpu/drm/virtio/virtgpu_vq.c
> +++ b/drivers/gpu/drm/virtio/virtgpu_vq.c
> @@ -191,7 +191,7 @@ void virtio_gpu_dequeue_ctrl_func(struct work_struct *work)
>         } while (!virtqueue_enable_cb(vgdev->ctrlq.vq));
>         spin_unlock(&vgdev->ctrlq.qlock);
>
> -       list_for_each_entry_safe(entry, tmp, &reclaim_list, list) {
> +       list_for_each_entry(entry, &reclaim_list, list) {
>                 resp = (struct virtio_gpu_ctrl_hdr *)entry->resp_buf;
>
>                 trace_virtio_gpu_cmd_response(vgdev->ctrlq.vq, resp);
> @@ -218,14 +218,18 @@ void virtio_gpu_dequeue_ctrl_func(struct work_struct *work)
>                 }
>                 if (entry->resp_cb)
>                         entry->resp_cb(vgdev, entry);
> -
> -               list_del(&entry->list);
> -               free_vbuf(vgdev, entry);
>         }
>         wake_up(&vgdev->ctrlq.ack_queue);
>
>         if (fence_id)
>                 virtio_gpu_fence_event_process(vgdev, fence_id);
> +
> +       list_for_each_entry_safe(entry, tmp, &reclaim_list, list) {
> +               if (entry->objs)
> +                       virtio_gpu_array_put_free(entry->objs);
> +               list_del(&entry->list);
We are clearing the list.  I guess list_del is not needed.
> +               free_vbuf(vgdev, entry);
> +       }
>  }
>
>  void virtio_gpu_dequeue_cursor_func(struct work_struct *work)
> @@ -939,7 +943,8 @@ void virtio_gpu_cmd_transfer_from_host_3d(struct virtio_gpu_device *vgdev,
>
>  void virtio_gpu_cmd_submit(struct virtio_gpu_device *vgdev,
>                            void *data, uint32_t data_size,
> -                          uint32_t ctx_id, struct virtio_gpu_fence *fence)
> +                          uint32_t ctx_id, struct virtio_gpu_fence *fence,
> +                          struct virtio_gpu_object_array *objs)
>  {
>         struct virtio_gpu_cmd_submit *cmd_p;
>         struct virtio_gpu_vbuffer *vbuf;
> @@ -949,6 +954,7 @@ void virtio_gpu_cmd_submit(struct virtio_gpu_device *vgdev,
>
>         vbuf->data_buf = data;
>         vbuf->data_size = data_size;
> +       vbuf->objs = objs;
>
>         cmd_p->hdr.type = cpu_to_le32(VIRTIO_GPU_CMD_SUBMIT_3D);
>         cmd_p->hdr.ctx_id = cpu_to_le32(ctx_id);
> --
> 2.18.1
>
> _______________________________________________
> dri-devel mailing list
> dri-devel@lists.freedesktop.org
> https://lists.freedesktop.org/mailman/listinfo/dri-devel

Re: [PATCH v4 09/12] drm/virtio: rework virtio_gpu_object_create fencing

[Chia-I Wu]

On Wed, Jun 19, 2019 at 11:07 PM Gerd Hoffmann <kraxel@redhat.com> wrote:
>
> Use gem reservation helpers and direct reservation_object_* calls
> instead of ttm.
>
> v3: Due to using the gem reservation object it is initialized and ready
> for use before calling ttm_bo_init, so we can also drop the tricky fence
> logic which checks whenever the command is in flight still.  We can
> simply fence our object before submitting the virtio command and be done
> with it.
>
> Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
> Acked-by: Daniel Vetter <daniel.vetter@ffwll.ch>
> ---
>  drivers/gpu/drm/virtio/virtgpu_drv.h    |  6 ++-
>  drivers/gpu/drm/virtio/virtgpu_object.c | 54 +++++++++----------------
>  drivers/gpu/drm/virtio/virtgpu_vq.c     |  8 +++-
>  3 files changed, 30 insertions(+), 38 deletions(-)
>
> diff --git a/drivers/gpu/drm/virtio/virtgpu_drv.h b/drivers/gpu/drm/virtio/virtgpu_drv.h
> index 65f5ce41c341..5213d7f499eb 100644
> --- a/drivers/gpu/drm/virtio/virtgpu_drv.h
> +++ b/drivers/gpu/drm/virtio/virtgpu_drv.h
> @@ -267,7 +267,8 @@ void virtio_gpu_free_vbufs(struct virtio_gpu_device *vgdev);
>  void virtio_gpu_cmd_create_resource(struct virtio_gpu_device *vgdev,
>                                     struct virtio_gpu_object *bo,
>                                     struct virtio_gpu_object_params *params,
> -                                   struct virtio_gpu_fence *fence);
> +                                   struct virtio_gpu_fence *fence,
> +                                   struct virtio_gpu_object_array *objs);
Move objs before fence?
>  void virtio_gpu_cmd_unref_resource(struct virtio_gpu_device *vgdev,
>                                    uint32_t resource_id);
>  void virtio_gpu_cmd_transfer_to_host_2d(struct virtio_gpu_device *vgdev,
> @@ -328,7 +329,8 @@ void
>  virtio_gpu_cmd_resource_create_3d(struct virtio_gpu_device *vgdev,
>                                   struct virtio_gpu_object *bo,
>                                   struct virtio_gpu_object_params *params,
> -                                 struct virtio_gpu_fence *fence);
> +                                 struct virtio_gpu_fence *fence,
> +                                 struct virtio_gpu_object_array *objs);
>  void virtio_gpu_ctrl_ack(struct virtqueue *vq);
>  void virtio_gpu_cursor_ack(struct virtqueue *vq);
>  void virtio_gpu_fence_ack(struct virtqueue *vq);
> diff --git a/drivers/gpu/drm/virtio/virtgpu_object.c b/drivers/gpu/drm/virtio/virtgpu_object.c
> index 82bfbf983fd2..90642907aa5c 100644
> --- a/drivers/gpu/drm/virtio/virtgpu_object.c
> +++ b/drivers/gpu/drm/virtio/virtgpu_object.c
> @@ -97,6 +97,7 @@ int virtio_gpu_object_create(struct virtio_gpu_device *vgdev,
>                              struct virtio_gpu_object **bo_ptr,
>                              struct virtio_gpu_fence *fence)
>  {
> +       struct virtio_gpu_object_array *objs = NULL;
>         struct virtio_gpu_object *bo;
>         size_t acc_size;
>         int ret;
> @@ -123,10 +124,27 @@ int virtio_gpu_object_create(struct virtio_gpu_device *vgdev,
>         }
>         bo->dumb = params->dumb;
>
> +       if (fence) {
> +               struct ww_acquire_ctx ticket;
> +
> +               objs = virtio_gpu_array_alloc(1);
> +               objs->objs[0] = &bo->gem_base;
> +               drm_gem_object_get(objs->objs[0]);
> +
> +               ret = drm_gem_lock_reservations(objs->objs, objs->nents,
> +                                               &ticket);
With only one object, we should be able to do a quick
reservation_object_lock_interruptible.
> +               if (ret == 0)
> +                       reservation_object_add_excl_fence(objs->objs[0]->resv,
> +                                                         &fence->f);
> +               drm_gem_unlock_reservations(objs->objs, objs->nents, &ticket);
These may need to be moved to after virtio_gpu_cmd_resource_create_3d.
> +       }
> +
>         if (params->virgl) {
> -               virtio_gpu_cmd_resource_create_3d(vgdev, bo, params, fence);
> +               virtio_gpu_cmd_resource_create_3d(vgdev, bo, params,
> +                                                 fence, objs);
>         } else {
> -               virtio_gpu_cmd_create_resource(vgdev, bo, params, fence);
> +               virtio_gpu_cmd_create_resource(vgdev, bo, params,
> +                                              fence, objs);
>         }
>
>         virtio_gpu_init_ttm_placement(bo);
> @@ -139,38 +157,6 @@ int virtio_gpu_object_create(struct virtio_gpu_device *vgdev,
>         if (ret != 0)
>                 return ret;
>
> -       if (fence) {
> -               struct virtio_gpu_fence_driver *drv = &vgdev->fence_drv;
> -               struct list_head validate_list;
> -               struct ttm_validate_buffer mainbuf;
> -               struct ww_acquire_ctx ticket;
> -               unsigned long irq_flags;
> -               bool signaled;
> -
> -               INIT_LIST_HEAD(&validate_list);
> -               memset(&mainbuf, 0, sizeof(struct ttm_validate_buffer));
> -
> -               /* use a gem reference since unref list undoes them */
> -               drm_gem_object_get(&bo->gem_base);
> -               mainbuf.bo = &bo->tbo;
> -               list_add(&mainbuf.head, &validate_list);
> -
> -               ret = virtio_gpu_object_list_validate(&ticket, &validate_list);
> -               if (ret == 0) {
> -                       spin_lock_irqsave(&drv->lock, irq_flags);
> -                       signaled = virtio_fence_signaled(&fence->f);
> -                       if (!signaled)
> -                               /* virtio create command still in flight */
> -                               ttm_eu_fence_buffer_objects(&ticket, &validate_list,
> -                                                           &fence->f);
> -                       spin_unlock_irqrestore(&drv->lock, irq_flags);
> -                       if (signaled)
> -                               /* virtio create command finished */
> -                               ttm_eu_backoff_reservation(&ticket, &validate_list);
> -               }
> -               virtio_gpu_unref_list(&validate_list);
> -       }
> -
>         *bo_ptr = bo;
>         return 0;
>  }
> diff --git a/drivers/gpu/drm/virtio/virtgpu_vq.c b/drivers/gpu/drm/virtio/virtgpu_vq.c
> index dc2c2c003200..6e2b287a7e4b 100644
> --- a/drivers/gpu/drm/virtio/virtgpu_vq.c
> +++ b/drivers/gpu/drm/virtio/virtgpu_vq.c
> @@ -391,13 +391,15 @@ static int virtio_gpu_queue_cursor(struct virtio_gpu_device *vgdev,
>  void virtio_gpu_cmd_create_resource(struct virtio_gpu_device *vgdev,
>                                     struct virtio_gpu_object *bo,
>                                     struct virtio_gpu_object_params *params,
> -                                   struct virtio_gpu_fence *fence)
> +                                   struct virtio_gpu_fence *fence,
> +                                   struct virtio_gpu_object_array *objs)
>  {
>         struct virtio_gpu_resource_create_2d *cmd_p;
>         struct virtio_gpu_vbuffer *vbuf;
>
>         cmd_p = virtio_gpu_alloc_cmd(vgdev, &vbuf, sizeof(*cmd_p));
>         memset(cmd_p, 0, sizeof(*cmd_p));
> +       vbuf->objs = objs;
>
>         cmd_p->hdr.type = cpu_to_le32(VIRTIO_GPU_CMD_RESOURCE_CREATE_2D);
>         cmd_p->resource_id = cpu_to_le32(bo->hw_res_handle);
> @@ -864,13 +866,15 @@ void
>  virtio_gpu_cmd_resource_create_3d(struct virtio_gpu_device *vgdev,
>                                   struct virtio_gpu_object *bo,
>                                   struct virtio_gpu_object_params *params,
> -                                 struct virtio_gpu_fence *fence)
> +                                 struct virtio_gpu_fence *fence,
> +                                 struct virtio_gpu_object_array *objs)
>  {
>         struct virtio_gpu_resource_create_3d *cmd_p;
>         struct virtio_gpu_vbuffer *vbuf;
>
>         cmd_p = virtio_gpu_alloc_cmd(vgdev, &vbuf, sizeof(*cmd_p));
>         memset(cmd_p, 0, sizeof(*cmd_p));
> +       vbuf->objs = objs;
>
>         cmd_p->hdr.type = cpu_to_le32(VIRTIO_GPU_CMD_RESOURCE_CREATE_3D);
>         cmd_p->resource_id = cpu_to_le32(bo->hw_res_handle);
> --
> 2.18.1
>
> _______________________________________________
> dri-devel mailing list
> dri-devel@lists.freedesktop.org
> https://lists.freedesktop.org/mailman/listinfo/dri-devel
_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/dri-devel

Re: [PATCH v4 11/12] drm/virtio: switch from ttm to gem shmem helpers

[Chia-I Wu]

.

On Wed, Jun 19, 2019 at 11:08 PM Gerd Hoffmann <kraxel@redhat.com> wrote:
>
> virtio-gpu basically needs a sg_table for the bo, to tell the host where
> the backing pages for the object are.  So the gem shmem helpers are a
> perfect fit.  Some drm_gem_object_funcs need thin wrappers to update the
> host state, but otherwise the helpers handle everything just fine.
>
> Once the fencing was sorted the switch was surprisingly easy and for the
> most part just removing the ttm code.
>
> v4: fix drm_gem_object_funcs name.
>
> Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
> Acked-by: Daniel Vetter <daniel.vetter@ffwll.ch>
> ---
>  drivers/gpu/drm/virtio/virtgpu_drv.h    |  52 +---
>  drivers/gpu/drm/virtio/virtgpu_drv.c    |  20 +-
>  drivers/gpu/drm/virtio/virtgpu_gem.c    |  16 +-
>  drivers/gpu/drm/virtio/virtgpu_ioctl.c  |  19 +-
>  drivers/gpu/drm/virtio/virtgpu_kms.c    |   9 -
>  drivers/gpu/drm/virtio/virtgpu_object.c | 148 ++++--------
>  drivers/gpu/drm/virtio/virtgpu_prime.c  |  37 ---
>  drivers/gpu/drm/virtio/virtgpu_ttm.c    | 304 ------------------------
>  drivers/gpu/drm/virtio/virtgpu_vq.c     |  24 +-
>  drivers/gpu/drm/virtio/Kconfig          |   2 +-
>  drivers/gpu/drm/virtio/Makefile         |   2 +-
>  11 files changed, 82 insertions(+), 551 deletions(-)
>  delete mode 100644 drivers/gpu/drm/virtio/virtgpu_ttm.c
>
> diff --git a/drivers/gpu/drm/virtio/virtgpu_drv.h b/drivers/gpu/drm/virtio/virtgpu_drv.h
> index c23f4016df39..1d589de34449 100644
> --- a/drivers/gpu/drm/virtio/virtgpu_drv.h
> +++ b/drivers/gpu/drm/virtio/virtgpu_drv.h
> @@ -33,14 +33,11 @@
>
>  #include <drm/drmP.h>
>  #include <drm/drm_gem.h>
> +#include <drm/drm_gem_shmem_helper.h>
>  #include <drm/drm_atomic.h>
>  #include <drm/drm_encoder.h>
>  #include <drm/drm_fb_helper.h>
>  #include <drm/drm_probe_helper.h>
> -#include <drm/ttm/ttm_bo_api.h>
> -#include <drm/ttm/ttm_bo_driver.h>
> -#include <drm/ttm/ttm_placement.h>
> -#include <drm/ttm/ttm_module.h>
>
>  #define DRIVER_NAME "virtio_gpu"
>  #define DRIVER_DESC "virtio GPU"
> @@ -68,21 +65,16 @@ struct virtio_gpu_object_params {
>  };
>
>  struct virtio_gpu_object {
> -       struct drm_gem_object gem_base;
> +       struct drm_gem_shmem_object base;
>         uint32_t hw_res_handle;
>
>         struct sg_table *pages;
>         uint32_t mapped;
> -       void *vmap;
>         bool dumb;
> -       struct ttm_place                placement_code;
> -       struct ttm_placement            placement;
> -       struct ttm_buffer_object        tbo;
> -       struct ttm_bo_kmap_obj          kmap;
>         bool created;
>  };
>  #define gem_to_virtio_gpu_obj(gobj) \
> -       container_of((gobj), struct virtio_gpu_object, gem_base)
> +       container_of((gobj), struct virtio_gpu_object, base.base)
>
>  struct virtio_gpu_object_array {
>         u32 nents;
> @@ -152,10 +144,6 @@ struct virtio_gpu_framebuffer {
>  #define to_virtio_gpu_framebuffer(x) \
>         container_of(x, struct virtio_gpu_framebuffer, base)
>
> -struct virtio_gpu_mman {
> -       struct ttm_bo_device            bdev;
> -};
> -
>  struct virtio_gpu_queue {
>         struct virtqueue *vq;
>         spinlock_t qlock;
> @@ -184,8 +172,6 @@ struct virtio_gpu_device {
>
>         struct virtio_device *vdev;
>
> -       struct virtio_gpu_mman mman;
> -
>         struct virtio_gpu_output outputs[VIRTIO_GPU_MAX_SCANOUTS];
>         uint32_t num_scanouts;
>
> @@ -349,11 +335,6 @@ struct drm_plane *virtio_gpu_plane_init(struct virtio_gpu_device *vgdev,
>                                         enum drm_plane_type type,
>                                         int index);
>
> -/* virtio_gpu_ttm.c */
> -int virtio_gpu_ttm_init(struct virtio_gpu_device *vgdev);
> -void virtio_gpu_ttm_fini(struct virtio_gpu_device *vgdev);
> -int virtio_gpu_mmap(struct file *filp, struct vm_area_struct *vma);
> -
>  /* virtio_gpu_fence.c */
>  bool virtio_fence_signaled(struct dma_fence *f);
>  struct virtio_gpu_fence *virtio_gpu_fence_alloc(
> @@ -365,58 +346,47 @@ void virtio_gpu_fence_event_process(struct virtio_gpu_device *vdev,
>                                     u64 last_seq);
>
>  /* virtio_gpu_object */
> +struct drm_gem_object *virtio_gpu_create_object(struct drm_device *dev,
> +                                               size_t size);
>  int virtio_gpu_object_create(struct virtio_gpu_device *vgdev,
>                              struct virtio_gpu_object_params *params,
>                              struct virtio_gpu_object **bo_ptr,
>                              struct virtio_gpu_fence *fence);
> -void virtio_gpu_object_kunmap(struct virtio_gpu_object *bo);
> -int virtio_gpu_object_kmap(struct virtio_gpu_object *bo);
> -int virtio_gpu_object_get_sg_table(struct virtio_gpu_device *qdev,
> -                                  struct virtio_gpu_object *bo);
> -void virtio_gpu_object_free_sg_table(struct virtio_gpu_object *bo);
>
>  /* virtgpu_prime.c */
> -struct sg_table *virtgpu_gem_prime_get_sg_table(struct drm_gem_object *obj);
>  struct drm_gem_object *virtgpu_gem_prime_import_sg_table(
>         struct drm_device *dev, struct dma_buf_attachment *attach,
>         struct sg_table *sgt);
> -void *virtgpu_gem_prime_vmap(struct drm_gem_object *obj);
> -void virtgpu_gem_prime_vunmap(struct drm_gem_object *obj, void *vaddr);
> -int virtgpu_gem_prime_mmap(struct drm_gem_object *obj,
> -                          struct vm_area_struct *vma);
>
>  static inline struct virtio_gpu_object*
>  virtio_gpu_object_ref(struct virtio_gpu_object *bo)
>  {
> -       ttm_bo_get(&bo->tbo);
> +       drm_gem_object_get(&bo->base.base);
>         return bo;
>  }
>
>  static inline void virtio_gpu_object_unref(struct virtio_gpu_object **bo)
>  {
> -       struct ttm_buffer_object *tbo;
> -
>         if ((*bo) == NULL)
>                 return;
> -       tbo = &((*bo)->tbo);
> -       ttm_bo_put(tbo);
> +       drm_gem_object_put(&(*bo)->base.base);
>         *bo = NULL;
>  }
The last users of these two helpers are removed with this patch.  We
can remove them.
>
>  static inline u64 virtio_gpu_object_mmap_offset(struct virtio_gpu_object *bo)
>  {
> -       return drm_vma_node_offset_addr(&bo->tbo.vma_node);
> +       return drm_vma_node_offset_addr(&bo->base.base.vma_node);
>  }
>
>  static inline int virtio_gpu_object_reserve(struct virtio_gpu_object *bo)
>  {
>         int r;
>
> -       r = reservation_object_lock_interruptible(bo->gem_base.resv, NULL);
> +       r = reservation_object_lock_interruptible(bo->base.base.resv, NULL);
>         if (unlikely(r != 0)) {
>                 if (r != -EINTR) {
>                         struct virtio_gpu_device *qdev =
> -                               bo->gem_base.dev->dev_private;
> +                               bo->base.base.dev->dev_private;
>                         dev_err(qdev->dev, "%p reserve failed\n", bo);
>                 }
>                 return r;
> @@ -426,7 +396,7 @@ static inline int virtio_gpu_object_reserve(struct virtio_gpu_object *bo)
>
>  static inline void virtio_gpu_object_unreserve(struct virtio_gpu_object *bo)
>  {
> -       reservation_object_unlock(bo->gem_base.resv);
> +       reservation_object_unlock(bo->base.base.resv);
>  }
>
>  /* virgl debufs */
> diff --git a/drivers/gpu/drm/virtio/virtgpu_drv.c b/drivers/gpu/drm/virtio/virtgpu_drv.c
> index c50868753132..6d7092251d20 100644
> --- a/drivers/gpu/drm/virtio/virtgpu_drv.c
> +++ b/drivers/gpu/drm/virtio/virtgpu_drv.c
> @@ -182,17 +182,7 @@ MODULE_AUTHOR("Dave Airlie <airlied@redhat.com>");
>  MODULE_AUTHOR("Gerd Hoffmann <kraxel@redhat.com>");
>  MODULE_AUTHOR("Alon Levy");
>
> -static const struct file_operations virtio_gpu_driver_fops = {
> -       .owner = THIS_MODULE,
> -       .open = drm_open,
> -       .mmap = virtio_gpu_mmap,
> -       .poll = drm_poll,
> -       .read = drm_read,
> -       .unlocked_ioctl = drm_ioctl,
> -       .release = drm_release,
> -       .compat_ioctl = drm_compat_ioctl,
> -       .llseek = noop_llseek,
> -};
> +DEFINE_DRM_GEM_SHMEM_FOPS(virtio_gpu_driver_fops);
>
>  static struct drm_driver driver = {
>         .driver_features = DRIVER_MODESET | DRIVER_GEM | DRIVER_PRIME | DRIVER_RENDER | DRIVER_ATOMIC,
> @@ -209,15 +199,9 @@ static struct drm_driver driver = {
>         .prime_fd_to_handle = drm_gem_prime_fd_to_handle,
>         .gem_prime_export = drm_gem_prime_export,
>         .gem_prime_import = drm_gem_prime_import,
> -       .gem_prime_get_sg_table = virtgpu_gem_prime_get_sg_table,
>         .gem_prime_import_sg_table = virtgpu_gem_prime_import_sg_table,
> -       .gem_prime_vmap = virtgpu_gem_prime_vmap,
> -       .gem_prime_vunmap = virtgpu_gem_prime_vunmap,
> -       .gem_prime_mmap = virtgpu_gem_prime_mmap,
>
> -       .gem_free_object_unlocked = virtio_gpu_gem_free_object,
> -       .gem_open_object = virtio_gpu_gem_object_open,
> -       .gem_close_object = virtio_gpu_gem_object_close,
> +       .gem_create_object = virtio_gpu_create_object,
>         .fops = &virtio_gpu_driver_fops,
>
>         .ioctls = virtio_gpu_ioctls,
> diff --git a/drivers/gpu/drm/virtio/virtgpu_gem.c b/drivers/gpu/drm/virtio/virtgpu_gem.c
> index 456cc382ce68..c5028b04cd92 100644
> --- a/drivers/gpu/drm/virtio/virtgpu_gem.c
> +++ b/drivers/gpu/drm/virtio/virtgpu_gem.c
> @@ -26,14 +26,6 @@
>  #include <drm/drmP.h>
>  #include "virtgpu_drv.h"
>
> -void virtio_gpu_gem_free_object(struct drm_gem_object *gem_obj)
> -{
> -       struct virtio_gpu_object *obj = gem_to_virtio_gpu_obj(gem_obj);
> -
> -       if (obj)
> -               virtio_gpu_object_unref(&obj);
> -}
> -
>  struct virtio_gpu_object*
>  virtio_gpu_alloc_object(struct drm_device *dev,
>                         struct virtio_gpu_object_params *params,
> @@ -64,16 +56,16 @@ int virtio_gpu_gem_create(struct drm_file *file,
>         if (IS_ERR(obj))
>                 return PTR_ERR(obj);
>
> -       ret = drm_gem_handle_create(file, &obj->gem_base, &handle);
> +       ret = drm_gem_handle_create(file, &obj->base.base, &handle);
>         if (ret) {
> -               drm_gem_object_release(&obj->gem_base);
> +               drm_gem_object_release(&obj->base.base);
>                 return ret;
>         }
>
> -       *obj_p = &obj->gem_base;
> +       *obj_p = &obj->base.base;
>
>         /* drop reference from allocate - handle holds it now */
> -       drm_gem_object_put_unlocked(&obj->gem_base);
> +       drm_gem_object_put_unlocked(&obj->base.base);
>
>         *handle_p = handle;
>         return 0;
> diff --git a/drivers/gpu/drm/virtio/virtgpu_ioctl.c b/drivers/gpu/drm/virtio/virtgpu_ioctl.c
> index 54e1cf84097a..d574ca38ebaa 100644
> --- a/drivers/gpu/drm/virtio/virtgpu_ioctl.c
> +++ b/drivers/gpu/drm/virtio/virtgpu_ioctl.c
> @@ -27,7 +27,6 @@
>
>  #include <drm/drmP.h>
>  #include <drm/virtgpu_drm.h>
> -#include <drm/ttm/ttm_execbuf_util.h>
>  #include <linux/sync_file.h>
>
>  #include "virtgpu_drv.h"
> @@ -269,7 +268,7 @@ static int virtio_gpu_resource_create_ioctl(struct drm_device *dev, void *data,
>         dma_fence_put(&fence->f);
>         if (IS_ERR(qobj))
>                 return PTR_ERR(qobj);
> -       obj = &qobj->gem_base;
> +       obj = &qobj->base.base;
>
>         ret = drm_gem_handle_create(file_priv, obj, &handle);
>         if (ret) {
> @@ -296,7 +295,7 @@ static int virtio_gpu_resource_info_ioctl(struct drm_device *dev, void *data,
>
>         qobj = gem_to_virtio_gpu_obj(gobj);
>
> -       ri->size = qobj->gem_base.size;
> +       ri->size = qobj->base.base.size;
>         ri->res_handle = qobj->hw_res_handle;
>         drm_gem_object_put_unlocked(gobj);
>         return 0;
> @@ -309,7 +308,6 @@ static int virtio_gpu_transfer_from_host_ioctl(struct drm_device *dev,
>         struct virtio_gpu_device *vgdev = dev->dev_private;
>         struct virtio_gpu_fpriv *vfpriv = file->driver_priv;
>         struct drm_virtgpu_3d_transfer_from_host *args = data;
> -       struct ttm_operation_ctx ctx = { true, false };
>         struct drm_gem_object *gobj = NULL;
>         struct virtio_gpu_object *qobj = NULL;
>         struct virtio_gpu_fence *fence;
> @@ -330,10 +328,6 @@ static int virtio_gpu_transfer_from_host_ioctl(struct drm_device *dev,
>         if (ret)
>                 goto out;
>
> -       ret = ttm_bo_validate(&qobj->tbo, &qobj->placement, &ctx);
> -       if (unlikely(ret))
> -               goto out_unres;
> -
>         convert_to_hw_box(&box, &args->box);
>
>         fence = virtio_gpu_fence_alloc(vgdev);
> @@ -345,7 +339,7 @@ static int virtio_gpu_transfer_from_host_ioctl(struct drm_device *dev,
>                 (vgdev, qobj->hw_res_handle,
>                  vfpriv->ctx_id, offset, args->level,
>                  &box, fence);
> -       reservation_object_add_excl_fence(qobj->tbo.resv,
> +       reservation_object_add_excl_fence(qobj->base.base.resv,
>                                           &fence->f);
>
>         dma_fence_put(&fence->f);
> @@ -362,7 +356,6 @@ static int virtio_gpu_transfer_to_host_ioctl(struct drm_device *dev, void *data,
>         struct virtio_gpu_device *vgdev = dev->dev_private;
>         struct virtio_gpu_fpriv *vfpriv = file->driver_priv;
>         struct drm_virtgpu_3d_transfer_to_host *args = data;
> -       struct ttm_operation_ctx ctx = { true, false };
>         struct drm_gem_object *gobj = NULL;
>         struct virtio_gpu_object *qobj = NULL;
>         struct virtio_gpu_fence *fence;
> @@ -380,10 +373,6 @@ static int virtio_gpu_transfer_to_host_ioctl(struct drm_device *dev, void *data,
>         if (ret)
>                 goto out;
>
> -       ret = ttm_bo_validate(&qobj->tbo, &qobj->placement, &ctx);
> -       if (unlikely(ret))
> -               goto out_unres;
> -
>         convert_to_hw_box(&box, &args->box);
>         if (!vgdev->has_virgl_3d) {
>                 virtio_gpu_cmd_transfer_to_host_2d
> @@ -399,7 +388,7 @@ static int virtio_gpu_transfer_to_host_ioctl(struct drm_device *dev, void *data,
>                         (vgdev, qobj,
>                          vfpriv ? vfpriv->ctx_id : 0, offset,
>                          args->level, &box, fence);
> -               reservation_object_add_excl_fence(qobj->tbo.resv,
> +               reservation_object_add_excl_fence(qobj->base.base.resv,
>                                                   &fence->f);
>                 dma_fence_put(&fence->f);
>         }
> diff --git a/drivers/gpu/drm/virtio/virtgpu_kms.c b/drivers/gpu/drm/virtio/virtgpu_kms.c
> index 84b6a6bf00c6..0bc6abaeafca 100644
> --- a/drivers/gpu/drm/virtio/virtgpu_kms.c
> +++ b/drivers/gpu/drm/virtio/virtgpu_kms.c
> @@ -171,12 +171,6 @@ int virtio_gpu_init(struct drm_device *dev)
>                 goto err_vbufs;
>         }
>
> -       ret = virtio_gpu_ttm_init(vgdev);
> -       if (ret) {
> -               DRM_ERROR("failed to init ttm %d\n", ret);
> -               goto err_ttm;
> -       }
> -
>         /* get display info */
>         virtio_cread(vgdev->vdev, struct virtio_gpu_config,
>                      num_scanouts, &num_scanouts);
> @@ -208,8 +202,6 @@ int virtio_gpu_init(struct drm_device *dev)
>         return 0;
>
>  err_scanouts:
> -       virtio_gpu_ttm_fini(vgdev);
> -err_ttm:
>         virtio_gpu_free_vbufs(vgdev);
>  err_vbufs:
>         vgdev->vdev->config->del_vqs(vgdev->vdev);
> @@ -240,7 +232,6 @@ void virtio_gpu_deinit(struct drm_device *dev)
>         vgdev->vdev->config->del_vqs(vgdev->vdev);
>
>         virtio_gpu_modeset_fini(vgdev);
> -       virtio_gpu_ttm_fini(vgdev);
>         virtio_gpu_free_vbufs(vgdev);
>         virtio_gpu_cleanup_cap_cache(vgdev);
>         kfree(vgdev->capsets);
> diff --git a/drivers/gpu/drm/virtio/virtgpu_object.c b/drivers/gpu/drm/virtio/virtgpu_object.c
> index 90642907aa5c..dc9d0b7171a4 100644
> --- a/drivers/gpu/drm/virtio/virtgpu_object.c
> +++ b/drivers/gpu/drm/virtio/virtgpu_object.c
> @@ -23,8 +23,6 @@
>   * WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
>   */
>
> -#include <drm/ttm/ttm_execbuf_util.h>
> -
>  #include "virtgpu_drv.h"
>
>  static int virtio_gpu_resource_id_get(struct virtio_gpu_device *vgdev,
> @@ -57,39 +55,44 @@ static void virtio_gpu_resource_id_put(struct virtio_gpu_device *vgdev, uint32_t
>  #endif
>  }
>
> -static void virtio_gpu_ttm_bo_destroy(struct ttm_buffer_object *tbo)
> +static void virtio_gpu_free_object(struct drm_gem_object *obj)
>  {
> -       struct virtio_gpu_object *bo;
> -       struct virtio_gpu_device *vgdev;
> -
> -       bo = container_of(tbo, struct virtio_gpu_object, tbo);
> -       vgdev = (struct virtio_gpu_device *)bo->gem_base.dev->dev_private;
> +       struct virtio_gpu_object *bo = gem_to_virtio_gpu_obj(obj);
> +       struct virtio_gpu_device *vgdev = bo->base.base.dev->dev_private;
>
> +       if (bo->pages)
> +               virtio_gpu_object_detach(vgdev, bo);
>         if (bo->created)
>                 virtio_gpu_cmd_unref_resource(vgdev, bo->hw_res_handle);
> -       if (bo->pages)
> -               virtio_gpu_object_free_sg_table(bo);
> -       if (bo->vmap)
> -               virtio_gpu_object_kunmap(bo);
> -       drm_gem_object_release(&bo->gem_base);
>         virtio_gpu_resource_id_put(vgdev, bo->hw_res_handle);
> -       kfree(bo);
> +
> +       drm_gem_shmem_free_object(obj);
>  }
>
> -static void virtio_gpu_init_ttm_placement(struct virtio_gpu_object *vgbo)
> +static const struct drm_gem_object_funcs virtio_gpu_gem_funcs = {
> +       .free = virtio_gpu_free_object,
> +       .open = virtio_gpu_gem_object_open,
> +       .close = virtio_gpu_gem_object_close,
> +
> +       .print_info = drm_gem_shmem_print_info,
> +       .pin = drm_gem_shmem_pin,
> +       .unpin = drm_gem_shmem_unpin,
> +       .get_sg_table = drm_gem_shmem_get_sg_table,
> +       .vmap = drm_gem_shmem_vmap,
> +       .vunmap = drm_gem_shmem_vunmap,
> +       .vm_ops = &drm_gem_shmem_vm_ops,
> +};
> +
> +struct drm_gem_object *virtio_gpu_create_object(struct drm_device *dev,
> +                                               size_t size)
>  {
> -       u32 c = 1;
> +       struct virtio_gpu_object *bo;
>
> -       vgbo->placement.placement = &vgbo->placement_code;
> -       vgbo->placement.busy_placement = &vgbo->placement_code;
> -       vgbo->placement_code.fpfn = 0;
> -       vgbo->placement_code.lpfn = 0;
> -       vgbo->placement_code.flags =
> -               TTM_PL_MASK_CACHING | TTM_PL_FLAG_TT |
> -               TTM_PL_FLAG_NO_EVICT;
> -       vgbo->placement.num_placement = c;
> -       vgbo->placement.num_busy_placement = c;
> +       bo = kzalloc(sizeof(*bo), GFP_KERNEL);
> +       if (!bo)
> +               return NULL;
>
> +       return &bo->base.base;
>  }
>
>  int virtio_gpu_object_create(struct virtio_gpu_device *vgdev,
> @@ -98,28 +101,22 @@ int virtio_gpu_object_create(struct virtio_gpu_device *vgdev,
>                              struct virtio_gpu_fence *fence)
>  {
>         struct virtio_gpu_object_array *objs = NULL;
> +       struct drm_gem_shmem_object *shmem_obj;
>         struct virtio_gpu_object *bo;
> -       size_t acc_size;
>         int ret;
>
>         *bo_ptr = NULL;
>
> -       acc_size = ttm_bo_dma_acc_size(&vgdev->mman.bdev, params->size,
> -                                      sizeof(struct virtio_gpu_object));
> +       params->size = roundup(params->size, PAGE_SIZE);
> +       shmem_obj = drm_gem_shmem_create(vgdev->ddev, params->size);
> +       if (IS_ERR(shmem_obj))
> +               return PTR_ERR(shmem_obj);
> +       bo = gem_to_virtio_gpu_obj(&shmem_obj->base);
> +       bo->base.base.funcs = &virtio_gpu_gem_funcs;
Move this to virtio_gpu_create_object.
>
> -       bo = kzalloc(sizeof(struct virtio_gpu_object), GFP_KERNEL);
> -       if (bo == NULL)
> -               return -ENOMEM;
>         ret = virtio_gpu_resource_id_get(vgdev, &bo->hw_res_handle);
>         if (ret < 0) {
> -               kfree(bo);
> -               return ret;
> -       }
> -       params->size = roundup(params->size, PAGE_SIZE);
> -       ret = drm_gem_object_init(vgdev->ddev, &bo->gem_base, params->size);
> -       if (ret != 0) {
> -               virtio_gpu_resource_id_put(vgdev, bo->hw_res_handle);
> -               kfree(bo);
> +               drm_gem_shmem_free_object(&shmem_obj->base);
>                 return ret;
>         }
>         bo->dumb = params->dumb;
> @@ -128,7 +125,7 @@ int virtio_gpu_object_create(struct virtio_gpu_device *vgdev,
>                 struct ww_acquire_ctx ticket;
>
>                 objs = virtio_gpu_array_alloc(1);
> -               objs->objs[0] = &bo->gem_base;
> +               objs->objs[0] = &bo->base.base;
>                 drm_gem_object_get(objs->objs[0]);
>
>                 ret = drm_gem_lock_reservations(objs->objs, objs->nents,
> @@ -147,75 +144,12 @@ int virtio_gpu_object_create(struct virtio_gpu_device *vgdev,
>                                                fence, objs);
>         }
>
> -       virtio_gpu_init_ttm_placement(bo);
> -       ret = ttm_bo_init(&vgdev->mman.bdev, &bo->tbo, params->size,
> -                         ttm_bo_type_device, &bo->placement, 0,
> -                         true, acc_size, NULL,
> -                         bo->gem_base.resv,
> -                         &virtio_gpu_ttm_bo_destroy);
> -       /* ttm_bo_init failure will call the destroy */
> -       if (ret != 0)
> +       ret = virtio_gpu_object_attach(vgdev, bo, NULL);
> +       if (ret != 0) {
> +               virtio_gpu_free_object(&shmem_obj->base);
>                 return ret;
> +       }
>
>         *bo_ptr = bo;
>         return 0;
>  }
> -
> -void virtio_gpu_object_kunmap(struct virtio_gpu_object *bo)
> -{
> -       bo->vmap = NULL;
> -       ttm_bo_kunmap(&bo->kmap);
> -}
> -
> -int virtio_gpu_object_kmap(struct virtio_gpu_object *bo)
> -{
> -       bool is_iomem;
> -       int r;
> -
> -       WARN_ON(bo->vmap);
> -
> -       r = ttm_bo_kmap(&bo->tbo, 0, bo->tbo.num_pages, &bo->kmap);
> -       if (r)
> -               return r;
> -       bo->vmap = ttm_kmap_obj_virtual(&bo->kmap, &is_iomem);
> -       return 0;
> -}
> -
> -int virtio_gpu_object_get_sg_table(struct virtio_gpu_device *qdev,
> -                                  struct virtio_gpu_object *bo)
> -{
> -       int ret;
> -       struct page **pages = bo->tbo.ttm->pages;
> -       int nr_pages = bo->tbo.num_pages;
> -       struct ttm_operation_ctx ctx = {
> -               .interruptible = false,
> -               .no_wait_gpu = false
> -       };
> -
> -       /* wtf swapping */
> -       if (bo->pages)
> -               return 0;
> -
> -       if (bo->tbo.ttm->state == tt_unpopulated)
> -               bo->tbo.ttm->bdev->driver->ttm_tt_populate(bo->tbo.ttm, &ctx);
> -       bo->pages = kmalloc(sizeof(struct sg_table), GFP_KERNEL);
> -       if (!bo->pages)
> -               goto out;
> -
> -       ret = sg_alloc_table_from_pages(bo->pages, pages, nr_pages, 0,
> -                                       nr_pages << PAGE_SHIFT, GFP_KERNEL);
> -       if (ret)
> -               goto out;
> -       return 0;
> -out:
> -       kfree(bo->pages);
> -       bo->pages = NULL;
> -       return -ENOMEM;
> -}
> -
> -void virtio_gpu_object_free_sg_table(struct virtio_gpu_object *bo)
> -{
> -       sg_free_table(bo->pages);
> -       kfree(bo->pages);
> -       bo->pages = NULL;
> -}
> diff --git a/drivers/gpu/drm/virtio/virtgpu_prime.c b/drivers/gpu/drm/virtio/virtgpu_prime.c
> index 8fbf71bd0c5e..18a155cd08d5 100644
> --- a/drivers/gpu/drm/virtio/virtgpu_prime.c
> +++ b/drivers/gpu/drm/virtio/virtgpu_prime.c
> @@ -28,46 +28,9 @@
>   * device that might share buffers with virtgpu
>   */
>
> -struct sg_table *virtgpu_gem_prime_get_sg_table(struct drm_gem_object *obj)
> -{
> -       struct virtio_gpu_object *bo = gem_to_virtio_gpu_obj(obj);
> -
> -       if (!bo->tbo.ttm->pages || !bo->tbo.ttm->num_pages)
> -               /* should not happen */
> -               return ERR_PTR(-EINVAL);
> -
> -       return drm_prime_pages_to_sg(bo->tbo.ttm->pages,
> -                                    bo->tbo.ttm->num_pages);
> -}
> -
>  struct drm_gem_object *virtgpu_gem_prime_import_sg_table(
>         struct drm_device *dev, struct dma_buf_attachment *attach,
>         struct sg_table *table)
>  {
>         return ERR_PTR(-ENODEV);
>  }
> -
> -void *virtgpu_gem_prime_vmap(struct drm_gem_object *obj)
> -{
> -       struct virtio_gpu_object *bo = gem_to_virtio_gpu_obj(obj);
> -       int ret;
> -
> -       ret = virtio_gpu_object_kmap(bo);
> -       if (ret)
> -               return NULL;
> -       return bo->vmap;
> -}
> -
> -void virtgpu_gem_prime_vunmap(struct drm_gem_object *obj, void *vaddr)
> -{
> -       virtio_gpu_object_kunmap(gem_to_virtio_gpu_obj(obj));
> -}
> -
> -int virtgpu_gem_prime_mmap(struct drm_gem_object *obj,
> -                          struct vm_area_struct *vma)
> -{
> -       struct virtio_gpu_object *bo = gem_to_virtio_gpu_obj(obj);
> -
> -       bo->gem_base.vma_node.vm_node.start = bo->tbo.vma_node.vm_node.start;
> -       return drm_gem_prime_mmap(obj, vma);
> -}
> diff --git a/drivers/gpu/drm/virtio/virtgpu_ttm.c b/drivers/gpu/drm/virtio/virtgpu_ttm.c
> deleted file mode 100644
> index 300ef3a83538..000000000000
> --- a/drivers/gpu/drm/virtio/virtgpu_ttm.c
> +++ /dev/null
> @@ -1,304 +0,0 @@
> -/*
> - * Copyright (C) 2015 Red Hat, Inc.
> - * All Rights Reserved.
> - *
> - * Authors:
> - *    Dave Airlie
> - *    Alon Levy
> - *
> - * Permission is hereby granted, free of charge, to any person obtaining a
> - * copy of this software and associated documentation files (the "Software"),
> - * to deal in the Software without restriction, including without limitation
> - * the rights to use, copy, modify, merge, publish, distribute, sublicense,
> - * and/or sell copies of the Software, and to permit persons to whom the
> - * Software is furnished to do so, subject to the following conditions:
> - *
> - * The above copyright notice and this permission notice shall be included in
> - * all copies or substantial portions of the Software.
> - *
> - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
> - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
> - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
> - * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
> - * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
> - * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
> - * OTHER DEALINGS IN THE SOFTWARE.
> - */
> -
> -#include <drm/ttm/ttm_bo_api.h>
> -#include <drm/ttm/ttm_bo_driver.h>
> -#include <drm/ttm/ttm_placement.h>
> -#include <drm/ttm/ttm_page_alloc.h>
> -#include <drm/ttm/ttm_module.h>
> -#include <drm/drmP.h>
> -#include <drm/drm.h>
> -#include <drm/virtgpu_drm.h>
> -#include "virtgpu_drv.h"
> -
> -#include <linux/delay.h>
> -
> -static struct
> -virtio_gpu_device *virtio_gpu_get_vgdev(struct ttm_bo_device *bdev)
> -{
> -       struct virtio_gpu_mman *mman;
> -       struct virtio_gpu_device *vgdev;
> -
> -       mman = container_of(bdev, struct virtio_gpu_mman, bdev);
> -       vgdev = container_of(mman, struct virtio_gpu_device, mman);
> -       return vgdev;
> -}
> -
> -int virtio_gpu_mmap(struct file *filp, struct vm_area_struct *vma)
> -{
> -       struct drm_file *file_priv;
> -       struct virtio_gpu_device *vgdev;
> -       int r;
> -
> -       file_priv = filp->private_data;
> -       vgdev = file_priv->minor->dev->dev_private;
> -       if (vgdev == NULL) {
> -               DRM_ERROR(
> -                "filp->private_data->minor->dev->dev_private == NULL\n");
> -               return -EINVAL;
> -       }
> -       r = ttm_bo_mmap(filp, vma, &vgdev->mman.bdev);
> -
> -       return r;
> -}
> -
> -static int virtio_gpu_invalidate_caches(struct ttm_bo_device *bdev,
> -                                       uint32_t flags)
> -{
> -       return 0;
> -}
> -
> -static int ttm_bo_man_get_node(struct ttm_mem_type_manager *man,
> -                              struct ttm_buffer_object *bo,
> -                              const struct ttm_place *place,
> -                              struct ttm_mem_reg *mem)
> -{
> -       mem->mm_node = (void *)1;
> -       return 0;
> -}
> -
> -static void ttm_bo_man_put_node(struct ttm_mem_type_manager *man,
> -                               struct ttm_mem_reg *mem)
> -{
> -       mem->mm_node = (void *)NULL;
> -}
> -
> -static int ttm_bo_man_init(struct ttm_mem_type_manager *man,
> -                          unsigned long p_size)
> -{
> -       return 0;
> -}
> -
> -static int ttm_bo_man_takedown(struct ttm_mem_type_manager *man)
> -{
> -       return 0;
> -}
> -
> -static void ttm_bo_man_debug(struct ttm_mem_type_manager *man,
> -                            struct drm_printer *printer)
> -{
> -}
> -
> -static const struct ttm_mem_type_manager_func virtio_gpu_bo_manager_func = {
> -       .init = ttm_bo_man_init,
> -       .takedown = ttm_bo_man_takedown,
> -       .get_node = ttm_bo_man_get_node,
> -       .put_node = ttm_bo_man_put_node,
> -       .debug = ttm_bo_man_debug
> -};
> -
> -static int virtio_gpu_init_mem_type(struct ttm_bo_device *bdev, uint32_t type,
> -                                   struct ttm_mem_type_manager *man)
> -{
> -       switch (type) {
> -       case TTM_PL_SYSTEM:
> -               /* System memory */
> -               man->flags = TTM_MEMTYPE_FLAG_MAPPABLE;
> -               man->available_caching = TTM_PL_MASK_CACHING;
> -               man->default_caching = TTM_PL_FLAG_CACHED;
> -               break;
> -       case TTM_PL_TT:
> -               man->func = &virtio_gpu_bo_manager_func;
> -               man->flags = TTM_MEMTYPE_FLAG_MAPPABLE;
> -               man->available_caching = TTM_PL_MASK_CACHING;
> -               man->default_caching = TTM_PL_FLAG_CACHED;
> -               break;
> -       default:
> -               DRM_ERROR("Unsupported memory type %u\n", (unsigned int)type);
> -               return -EINVAL;
> -       }
> -       return 0;
> -}
> -
> -static void virtio_gpu_evict_flags(struct ttm_buffer_object *bo,
> -                               struct ttm_placement *placement)
> -{
> -       static const struct ttm_place placements = {
> -               .fpfn  = 0,
> -               .lpfn  = 0,
> -               .flags = TTM_PL_MASK_CACHING | TTM_PL_FLAG_SYSTEM,
> -       };
> -
> -       placement->placement = &placements;
> -       placement->busy_placement = &placements;
> -       placement->num_placement = 1;
> -       placement->num_busy_placement = 1;
> -}
> -
> -static int virtio_gpu_verify_access(struct ttm_buffer_object *bo,
> -                                   struct file *filp)
> -{
> -       return 0;
> -}
> -
> -static int virtio_gpu_ttm_io_mem_reserve(struct ttm_bo_device *bdev,
> -                                        struct ttm_mem_reg *mem)
> -{
> -       struct ttm_mem_type_manager *man = &bdev->man[mem->mem_type];
> -
> -       mem->bus.addr = NULL;
> -       mem->bus.offset = 0;
> -       mem->bus.size = mem->num_pages << PAGE_SHIFT;
> -       mem->bus.base = 0;
> -       mem->bus.is_iomem = false;
> -       if (!(man->flags & TTM_MEMTYPE_FLAG_MAPPABLE))
> -               return -EINVAL;
> -       switch (mem->mem_type) {
> -       case TTM_PL_SYSTEM:
> -       case TTM_PL_TT:
> -               /* system memory */
> -               return 0;
> -       default:
> -               return -EINVAL;
> -       }
> -       return 0;
> -}
> -
> -static void virtio_gpu_ttm_io_mem_free(struct ttm_bo_device *bdev,
> -                                      struct ttm_mem_reg *mem)
> -{
> -}
> -
> -/*
> - * TTM backend functions.
> - */
> -struct virtio_gpu_ttm_tt {
> -       struct ttm_dma_tt               ttm;
> -       struct virtio_gpu_object        *obj;
> -};
> -
> -static int virtio_gpu_ttm_tt_bind(struct ttm_tt *ttm,
> -                                 struct ttm_mem_reg *bo_mem)
> -{
> -       struct virtio_gpu_ttm_tt *gtt =
> -               container_of(ttm, struct virtio_gpu_ttm_tt, ttm.ttm);
> -       struct virtio_gpu_device *vgdev =
> -               virtio_gpu_get_vgdev(gtt->obj->tbo.bdev);
> -
> -       virtio_gpu_object_attach(vgdev, gtt->obj, NULL);
> -       return 0;
> -}
> -
> -static int virtio_gpu_ttm_tt_unbind(struct ttm_tt *ttm)
> -{
> -       struct virtio_gpu_ttm_tt *gtt =
> -               container_of(ttm, struct virtio_gpu_ttm_tt, ttm.ttm);
> -       struct virtio_gpu_device *vgdev =
> -               virtio_gpu_get_vgdev(gtt->obj->tbo.bdev);
> -
> -       virtio_gpu_object_detach(vgdev, gtt->obj);
> -       return 0;
> -}
> -
> -static void virtio_gpu_ttm_tt_destroy(struct ttm_tt *ttm)
> -{
> -       struct virtio_gpu_ttm_tt *gtt =
> -               container_of(ttm, struct virtio_gpu_ttm_tt, ttm.ttm);
> -
> -       ttm_dma_tt_fini(&gtt->ttm);
> -       kfree(gtt);
> -}
> -
> -static struct ttm_backend_func virtio_gpu_tt_func = {
> -       .bind = &virtio_gpu_ttm_tt_bind,
> -       .unbind = &virtio_gpu_ttm_tt_unbind,
> -       .destroy = &virtio_gpu_ttm_tt_destroy,
> -};
> -
> -static struct ttm_tt *virtio_gpu_ttm_tt_create(struct ttm_buffer_object *bo,
> -                                              uint32_t page_flags)
> -{
> -       struct virtio_gpu_device *vgdev;
> -       struct virtio_gpu_ttm_tt *gtt;
> -
> -       vgdev = virtio_gpu_get_vgdev(bo->bdev);
> -       gtt = kzalloc(sizeof(struct virtio_gpu_ttm_tt), GFP_KERNEL);
> -       if (gtt == NULL)
> -               return NULL;
> -       gtt->ttm.ttm.func = &virtio_gpu_tt_func;
> -       gtt->obj = container_of(bo, struct virtio_gpu_object, tbo);
> -       if (ttm_dma_tt_init(&gtt->ttm, bo, page_flags)) {
> -               kfree(gtt);
> -               return NULL;
> -       }
> -       return &gtt->ttm.ttm;
> -}
> -
> -static void virtio_gpu_bo_swap_notify(struct ttm_buffer_object *tbo)
> -{
> -       struct virtio_gpu_object *bo;
> -
> -       bo = container_of(tbo, struct virtio_gpu_object, tbo);
> -
> -       if (bo->pages)
> -               virtio_gpu_object_free_sg_table(bo);
> -}
> -
> -static struct ttm_bo_driver virtio_gpu_bo_driver = {
> -       .ttm_tt_create = &virtio_gpu_ttm_tt_create,
> -       .invalidate_caches = &virtio_gpu_invalidate_caches,
> -       .init_mem_type = &virtio_gpu_init_mem_type,
> -       .eviction_valuable = ttm_bo_eviction_valuable,
> -       .evict_flags = &virtio_gpu_evict_flags,
> -       .verify_access = &virtio_gpu_verify_access,
> -       .io_mem_reserve = &virtio_gpu_ttm_io_mem_reserve,
> -       .io_mem_free = &virtio_gpu_ttm_io_mem_free,
> -       .swap_notify = &virtio_gpu_bo_swap_notify,
> -};
> -
> -int virtio_gpu_ttm_init(struct virtio_gpu_device *vgdev)
> -{
> -       int r;
> -
> -       /* No others user of address space so set it to 0 */
> -       r = ttm_bo_device_init(&vgdev->mman.bdev,
> -                              &virtio_gpu_bo_driver,
> -                              vgdev->ddev->anon_inode->i_mapping,
> -                              false);
> -       if (r) {
> -               DRM_ERROR("failed initializing buffer object driver(%d).\n", r);
> -               goto err_dev_init;
> -       }
> -
> -       r = ttm_bo_init_mm(&vgdev->mman.bdev, TTM_PL_TT, 0);
> -       if (r) {
> -               DRM_ERROR("Failed initializing GTT heap.\n");
> -               goto err_mm_init;
> -       }
> -       return 0;
> -
> -err_mm_init:
> -       ttm_bo_device_release(&vgdev->mman.bdev);
> -err_dev_init:
> -       return r;
> -}
> -
> -void virtio_gpu_ttm_fini(struct virtio_gpu_device *vgdev)
> -{
> -       ttm_bo_device_release(&vgdev->mman.bdev);
> -       DRM_INFO("virtio_gpu: ttm finalized\n");
> -}
> diff --git a/drivers/gpu/drm/virtio/virtgpu_vq.c b/drivers/gpu/drm/virtio/virtgpu_vq.c
> index 6e2b287a7e4b..352fd0689406 100644
> --- a/drivers/gpu/drm/virtio/virtgpu_vq.c
> +++ b/drivers/gpu/drm/virtio/virtgpu_vq.c
> @@ -974,17 +974,21 @@ int virtio_gpu_object_attach(struct virtio_gpu_device *vgdev,
>         bool use_dma_api = !virtio_has_iommu_quirk(vgdev->vdev);
>         struct virtio_gpu_mem_entry *ents;
>         struct scatterlist *sg;
> -       int si, nents;
> +       int si, nents, ret;
>
>         if (WARN_ON_ONCE(!obj->created))
>                 return -EINVAL;
> +       if (WARN_ON_ONCE(obj->pages))
> +               return -EINVAL;
>
> -       if (!obj->pages) {
> -               int ret;
> +       ret = drm_gem_shmem_pin(&obj->base.base);
The bo is attached for its entire lifetime, at least currently.  Maybe
we can use drm_gem_shmem_get_pages_sgt (and get rid of obj->pages).


> +       if (ret < 0)
> +               return -EINVAL;
>
> -               ret = virtio_gpu_object_get_sg_table(vgdev, obj);
> -               if (ret)
> -                       return ret;
> +       obj->pages = drm_gem_shmem_get_sg_table(&obj->base.base);
> +       if (obj->pages == NULL) {
> +               drm_gem_shmem_unpin(&obj->base.base);
> +               return -EINVAL;
>         }
>
>         if (use_dma_api) {
> @@ -1023,6 +1027,9 @@ void virtio_gpu_object_detach(struct virtio_gpu_device *vgdev,
>  {
>         bool use_dma_api = !virtio_has_iommu_quirk(vgdev->vdev);
>
> +       if (WARN_ON_ONCE(!obj->pages))
> +               return;
> +
>         if (use_dma_api && obj->mapped) {
>                 struct virtio_gpu_fence *fence = virtio_gpu_fence_alloc(vgdev);
>                 /* detach backing and wait for the host process it ... */
> @@ -1038,6 +1045,11 @@ void virtio_gpu_object_detach(struct virtio_gpu_device *vgdev,
>         } else {
>                 virtio_gpu_cmd_resource_inval_backing(vgdev, obj->hw_res_handle, NULL);
>         }
> +
> +       sg_free_table(obj->pages);
> +       obj->pages = NULL;
> +
> +       drm_gem_shmem_unpin(&obj->base.base);
>  }
>
>  void virtio_gpu_cursor_ping(struct virtio_gpu_device *vgdev,
> diff --git a/drivers/gpu/drm/virtio/Kconfig b/drivers/gpu/drm/virtio/Kconfig
> index ba36e933bb49..eff3047052d4 100644
> --- a/drivers/gpu/drm/virtio/Kconfig
> +++ b/drivers/gpu/drm/virtio/Kconfig
> @@ -3,7 +3,7 @@ config DRM_VIRTIO_GPU
>         tristate "Virtio GPU driver"
>         depends on DRM && VIRTIO && MMU
>         select DRM_KMS_HELPER
> -       select DRM_TTM
> +       select DRM_GEM_SHMEM_HELPER
>         help
>            This is the virtual GPU driver for virtio.  It can be used with
>            QEMU based VMMs (like KVM or Xen).
> diff --git a/drivers/gpu/drm/virtio/Makefile b/drivers/gpu/drm/virtio/Makefile
> index 458e606a936f..92aa2b3d349d 100644
> --- a/drivers/gpu/drm/virtio/Makefile
> +++ b/drivers/gpu/drm/virtio/Makefile
> @@ -4,7 +4,7 @@
>  # Direct Rendering Infrastructure (DRI) in XFree86 4.1.0 and higher.
>
>  virtio-gpu-y := virtgpu_drv.o virtgpu_kms.o virtgpu_gem.o \
> -       virtgpu_display.o virtgpu_vq.o virtgpu_ttm.o \
> +       virtgpu_display.o virtgpu_vq.o \
>         virtgpu_fence.o virtgpu_object.o virtgpu_debugfs.o virtgpu_plane.o \
>         virtgpu_ioctl.o virtgpu_prime.o virtgpu_trace_points.o
>
> --
> 2.18.1
>
> _______________________________________________
> dri-devel mailing list
> dri-devel@lists.freedesktop.org
> https://lists.freedesktop.org/mailman/listinfo/dri-devel

Re: [PATCH v4 02/12] drm/virtio: switch virtio_gpu_wait_ioctl() to gem helper.

[Gerd Hoffmann]

On Wed, Jun 26, 2019 at 04:55:20PM -0700, Chia-I Wu wrote:
> On Wed, Jun 19, 2019 at 11:07 PM Gerd Hoffmann <kraxel@redhat.com> wrote:
> >
> > Use drm_gem_reservation_object_wait() in virtio_gpu_wait_ioctl().
> > This also makes the ioctl run lockless.
> The userspace has a BO cache to avoid freeing BOs immediately but to
> reuse them on next allocations.  The BO cache checks if a BO is busy
> before reuse, and I am seeing a big negative perf impact because of
> slow virtio_gpu_wait_ioctl.  I wonder if this helps.

Could help indeed (assuming it checks with NOWAIT).

How many objects does userspace check in one go typically?  Maybe it
makes sense to add an ioctl which checks a list, to reduce the system
call overhead.

> > +       if (args->flags & VIRTGPU_WAIT_NOWAIT) {
> > +               obj = drm_gem_object_lookup(file, args->handle);
> Don't we need a NULL check here?

Yes, we do.  Will fix.

thanks,
  Gerd

Re: [PATCH v4 08/12] drm/virtio: rework virtio_gpu_execbuffer_ioctl fencing

[Gerd Hoffmann]

  Hi,

> > --- a/drivers/gpu/drm/virtio/virtgpu_drv.h
> > +++ b/drivers/gpu/drm/virtio/virtgpu_drv.h
> > @@ -120,9 +120,9 @@ struct virtio_gpu_vbuffer {
> >
> >         char *resp_buf;
> >         int resp_size;
> > -
> >         virtio_gpu_resp_cb resp_cb;
> >
> > +       struct virtio_gpu_object_array *objs;
> This can use a comment (e.g., objects referenced by the vbuffer)

IMHO this is obvious ...

> >  void virtio_gpu_cmd_submit(struct virtio_gpu_device *vgdev,
> >                            void *data, uint32_t data_size,
> > -                          uint32_t ctx_id, struct virtio_gpu_fence *fence);
> > +                          uint32_t ctx_id, struct virtio_gpu_fence *fence,
> > +                          struct virtio_gpu_object_array *objs);
> Can we keep fence, which is updated, as the last parameter?

Fixed.

> > +       if (buflist) {
> > +               for (i = 0; i < exbuf->num_bo_handles; i++)
> > +                       reservation_object_add_excl_fence(buflist->objs[i]->resv,
> > +                                                         &out_fence->f);
> > +               drm_gem_unlock_reservations(buflist->objs, buflist->nents,
> > +                                           &ticket);
> > +       }
> We used to unlock after virtio_gpu_cmd_submit.
> 
> I guess, the fence is considered signaled (because its seqno is still
> 0) until after virtio_gpu_cmd_submit.  We probably don't want other
> processes to see the semi-initialized fence.

Good point.  Fixed.

> >  out_memdup:
> >         kfree(buf);
> >  out_unresv:
> > -       ttm_eu_backoff_reservation(&ticket, &validate_list);
> > -out_free:
> > -       virtio_gpu_unref_list(&validate_list);
> Keeping out_free to free buflist seems just fine.

We don't need the separate label though ...

> > +       drm_gem_unlock_reservations(buflist->objs, buflist->nents, &ticket);
> >  out_unused_fd:
> >         kvfree(bo_handles);
> > -       kvfree(buflist);
> > +       if (buflist)
> > +               virtio_gpu_array_put_free(buflist);

... and the buflist is released here if needed.

But we need if (buflist) for drm_gem_unlock_reservations too.  Fixed.

> > -
> > -               list_del(&entry->list);
> > -               free_vbuf(vgdev, entry);
> >         }
> >         wake_up(&vgdev->ctrlq.ack_queue);
> >
> >         if (fence_id)
> >                 virtio_gpu_fence_event_process(vgdev, fence_id);
> > +
> > +       list_for_each_entry_safe(entry, tmp, &reclaim_list, list) {
> > +               if (entry->objs)
> > +                       virtio_gpu_array_put_free(entry->objs);
> > +               list_del(&entry->list);
> We are clearing the list.  I guess list_del is not needed.
> > +               free_vbuf(vgdev, entry);

This just shuffles around the code.  Dropping list_del() is unrelated
and should be a separate patch.

Beside that I'm not sure it actually can be dropped.  free_vbuf() will
not kfree() the vbuf but keep it cached in a freelist instead.

cheers,
  Gerd

Re: [PATCH v4 11/12] drm/virtio: switch from ttm to gem shmem helpers

[Gerd Hoffmann]

> >  static inline struct virtio_gpu_object*
> >  virtio_gpu_object_ref(struct virtio_gpu_object *bo)

> The last users of these two helpers are removed with this patch.  We
> can remove them.

patch 12/12 does that.

> > +       bo = gem_to_virtio_gpu_obj(&shmem_obj->base);
> > +       bo->base.base.funcs = &virtio_gpu_gem_funcs;
> Move this to virtio_gpu_create_object.

Fixed.

> > +       ret = drm_gem_shmem_pin(&obj->base.base);
> The bo is attached for its entire lifetime, at least currently.  Maybe
> we can use drm_gem_shmem_get_pages_sgt (and get rid of obj->pages).

Already checked this.
We can't due to the iommu quirks.

cheers,
  Gerd

_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/dri-devel

Re: [PATCH v4 11/12] drm/virtio: switch from ttm to gem shmem helpers

[Chia-I Wu]

On Fri, Jun 28, 2019 at 3:49 AM Gerd Hoffmann <kraxel@redhat.com> wrote:
>
> > >  static inline struct virtio_gpu_object*
> > >  virtio_gpu_object_ref(struct virtio_gpu_object *bo)
>
> > The last users of these two helpers are removed with this patch.  We
> > can remove them.
>
> patch 12/12 does that.
I meant virtio_gpu_object_ref/unref, which are still around after patch 12.
>
> > > +       bo = gem_to_virtio_gpu_obj(&shmem_obj->base);
> > > +       bo->base.base.funcs = &virtio_gpu_gem_funcs;
> > Move this to virtio_gpu_create_object.
>
> Fixed.
>
> > > +       ret = drm_gem_shmem_pin(&obj->base.base);
> > The bo is attached for its entire lifetime, at least currently.  Maybe
> > we can use drm_gem_shmem_get_pages_sgt (and get rid of obj->pages).
>
> Already checked this.
> We can't due to the iommu quirks.
>
> cheers,
>   Gerd
>

Re: [PATCH v4 08/12] drm/virtio: rework virtio_gpu_execbuffer_ioctl fencing

[Chia-I Wu]

On Fri, Jun 28, 2019 at 3:34 AM Gerd Hoffmann <kraxel@redhat.com> wrote:
>
>   Hi,
>
> > > --- a/drivers/gpu/drm/virtio/virtgpu_drv.h
> > > +++ b/drivers/gpu/drm/virtio/virtgpu_drv.h
> > > @@ -120,9 +120,9 @@ struct virtio_gpu_vbuffer {
> > >
> > >         char *resp_buf;
> > >         int resp_size;
> > > -
> > >         virtio_gpu_resp_cb resp_cb;
> > >
> > > +       struct virtio_gpu_object_array *objs;
> > This can use a comment (e.g., objects referenced by the vbuffer)
>
> IMHO this is obvious ...
>
> > >  void virtio_gpu_cmd_submit(struct virtio_gpu_device *vgdev,
> > >                            void *data, uint32_t data_size,
> > > -                          uint32_t ctx_id, struct virtio_gpu_fence *fence);
> > > +                          uint32_t ctx_id, struct virtio_gpu_fence *fence,
> > > +                          struct virtio_gpu_object_array *objs);
> > Can we keep fence, which is updated, as the last parameter?
>
> Fixed.
>
> > > +       if (buflist) {
> > > +               for (i = 0; i < exbuf->num_bo_handles; i++)
> > > +                       reservation_object_add_excl_fence(buflist->objs[i]->resv,
> > > +                                                         &out_fence->f);
> > > +               drm_gem_unlock_reservations(buflist->objs, buflist->nents,
> > > +                                           &ticket);
> > > +       }
> > We used to unlock after virtio_gpu_cmd_submit.
> >
> > I guess, the fence is considered signaled (because its seqno is still
> > 0) until after virtio_gpu_cmd_submit.  We probably don't want other
> > processes to see the semi-initialized fence.
>
> Good point.  Fixed.
>
> > >  out_memdup:
> > >         kfree(buf);
> > >  out_unresv:
> > > -       ttm_eu_backoff_reservation(&ticket, &validate_list);
> > > -out_free:
> > > -       virtio_gpu_unref_list(&validate_list);
> > Keeping out_free to free buflist seems just fine.
>
> We don't need the separate label though ...
>
> > > +       drm_gem_unlock_reservations(buflist->objs, buflist->nents, &ticket);
> > >  out_unused_fd:
> > >         kvfree(bo_handles);
> > > -       kvfree(buflist);
> > > +       if (buflist)
> > > +               virtio_gpu_array_put_free(buflist);
>
> ... and the buflist is released here if needed.
>
> But we need if (buflist) for drm_gem_unlock_reservations too.  Fixed.
>
> > > -
> > > -               list_del(&entry->list);
> > > -               free_vbuf(vgdev, entry);
> > >         }
> > >         wake_up(&vgdev->ctrlq.ack_queue);
> > >
> > >         if (fence_id)
> > >                 virtio_gpu_fence_event_process(vgdev, fence_id);
> > > +
> > > +       list_for_each_entry_safe(entry, tmp, &reclaim_list, list) {
> > > +               if (entry->objs)
> > > +                       virtio_gpu_array_put_free(entry->objs);
> > > +               list_del(&entry->list);
> > We are clearing the list.  I guess list_del is not needed.
> > > +               free_vbuf(vgdev, entry);
>
> This just shuffles around the code.  Dropping list_del() is unrelated
> and should be a separate patch.
Fair point.  We now loop the list twice and I was just looking for
chances for micro-optimizations.
>
> Beside that I'm not sure it actually can be dropped.  free_vbuf() will
> not kfree() the vbuf but keep it cached in a freelist instead.
vbuf is created with kmem_cache_zalloc which always zeros the struct.

>
> cheers,
>   Gerd
>
_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/dri-devel

Re: [PATCH v4 02/12] drm/virtio: switch virtio_gpu_wait_ioctl() to gem helper.

[Chia-I Wu]

On Fri, Jun 28, 2019 at 3:05 AM Gerd Hoffmann <kraxel@redhat.com> wrote:
>
> On Wed, Jun 26, 2019 at 04:55:20PM -0700, Chia-I Wu wrote:
> > On Wed, Jun 19, 2019 at 11:07 PM Gerd Hoffmann <kraxel@redhat.com> wrote:
> > >
> > > Use drm_gem_reservation_object_wait() in virtio_gpu_wait_ioctl().
> > > This also makes the ioctl run lockless.
> > The userspace has a BO cache to avoid freeing BOs immediately but to
> > reuse them on next allocations.  The BO cache checks if a BO is busy
> > before reuse, and I am seeing a big negative perf impact because of
> > slow virtio_gpu_wait_ioctl.  I wonder if this helps.
>
> Could help indeed (assuming it checks with NOWAIT).
Yeah, that is the case.
>
> How many objects does userspace check in one go typically?  Maybe it
> makes sense to add an ioctl which checks a list, to reduce the system
> call overhead.
One.  The cache sorts BOs by the time they are freed, and checks only
the first (compatible) BO.  If it is idle, cache hit.  Otherwise,
cache miss.  A new ioctl probably won't help.


>
> > > +       if (args->flags & VIRTGPU_WAIT_NOWAIT) {
> > > +               obj = drm_gem_object_lookup(file, args->handle);
> > Don't we need a NULL check here?
>
> Yes, we do.  Will fix.
>
> thanks,
>   Gerd
>