On Wed, Feb 12, 2020 at 5:54 PM Gurchetan Singh gurchetansingh@chromium.org wrote:
On Wed, Feb 12, 2020 at 10:50 AM Chia-I Wu olvaffe@gmail.com wrote:
On Tue, Feb 11, 2020 at 3:56 PM Gurchetan Singh gurchetansingh@chromium.org wrote:
We only want create a new virglrenderer context after the first 3D ioctl.
Signed-off-by: Gurchetan Singh gurchetansingh@chromium.org
drivers/gpu/drm/virtio/virtgpu_drv.h | 1 + drivers/gpu/drm/virtio/virtgpu_ioctl.c | 5 +++++ drivers/gpu/drm/virtio/virtgpu_kms.c | 2 ++ 3 files changed, 8 insertions(+)
diff --git a/drivers/gpu/drm/virtio/virtgpu_drv.h b/drivers/gpu/drm/virtio/virtgpu_drv.h index edaa7b8224a8..93ce69c0d9be 100644 --- a/drivers/gpu/drm/virtio/virtgpu_drv.h +++ b/drivers/gpu/drm/virtio/virtgpu_drv.h @@ -209,6 +209,7 @@ struct virtio_gpu_device {
struct virtio_gpu_fpriv { uint32_t ctx_id;
atomic_t context_initiated;}; /* virtio_ioctl.c */ diff --git a/drivers/gpu/drm/virtio/virtgpu_ioctl.c b/drivers/gpu/drm/virtio/virtgpu_ioctl.c index f1afabaa3a08..858ee153fb18 100644 --- a/drivers/gpu/drm/virtio/virtgpu_ioctl.c +++ b/drivers/gpu/drm/virtio/virtgpu_ioctl.c @@ -44,9 +44,14 @@ void virtio_gpu_create_context(struct drm_device *dev, if (!vgdev->has_virgl_3d) return;
if (atomic_read(&vfpriv->context_initiated))return;get_task_comm(dbgname, current); virtio_gpu_cmd_context_create(vgdev, vfpriv->ctx_id, strlen(dbgname), dbgname);atomic_inc(&vfpriv->context_initiated);}
This needs to be protected with a lock.
What do you think about atomic_add_unless(&vfpriv->context_initiated, 1, 1)?
Does that prevent virtio_gpu_cmd_context_create from being called more than once when two threads call this function at the same time?
ctx_id can probably be generated here as well.
If we generate the context id in virtio_gpu_create_context, we'll have to add error checking in all of the ioctl call-sites (since ida_alloc may fail).
int virtio_gpu_create_context(..) ret = virtio_gpu_create_context(dev, file); if (ret) return -EINVAL
vs.
void virtio_gpu_create_context(..) virtio_gpu_create_context(..)
Any strong opinions here?
Yeah, the current way should be better... unless we decide to make virtio_gpu_cmd_context_create synchronous and check for errors.
static int virtio_gpu_map_ioctl(struct drm_device *dev, void *data, diff --git a/drivers/gpu/drm/virtio/virtgpu_kms.c b/drivers/gpu/drm/virtio/virtgpu_kms.c index e1e1c0821a35..64fe5fcbedfd 100644 --- a/drivers/gpu/drm/virtio/virtgpu_kms.c +++ b/drivers/gpu/drm/virtio/virtgpu_kms.c @@ -270,7 +270,9 @@ int virtio_gpu_driver_open(struct drm_device *dev, struct drm_file *file) return id; }
vfpriv->ctx_id = id;atomic_set(&vfpriv->context_initiated, 0); file->driver_priv = vfpriv; virtio_gpu_create_context(dev, file);-- 2.25.0.225.g125e21ebc7-goog