On Mon, 21 May 2012 08:25:14 +0100, Dave Airlie <airl...@gmail.com> wrote:
> From: Dave Airlie <airl...@redhat.com>
> 
> If userspace attempts to import a buffer it exported on the same device,
> we need to return the same GEM handle for it, not a new handle pointing
> at the same GEM object.
> 
> Signed-off-by: Dave Airlie <airl...@redhat.com>
> ---
>  drivers/gpu/drm/drm_gem.c   |   16 ++++++++++++++--
>  drivers/gpu/drm/drm_prime.c |   12 ++++++++++++
>  2 files changed, 26 insertions(+), 2 deletions(-)
> 
> diff --git a/drivers/gpu/drm/drm_gem.c b/drivers/gpu/drm/drm_gem.c
> index 1ab29a7..9cc3761 100644
> --- a/drivers/gpu/drm/drm_gem.c
> +++ b/drivers/gpu/drm/drm_gem.c
> @@ -233,9 +233,15 @@ drm_gem_handle_delete(struct drm_file *filp, u32 handle)
>       idr_remove(&filp->object_idr, handle);
>       spin_unlock(&filp->table_lock);
>  
> -     if (obj->import_attach)
> +     if (obj->import_attach) {
>               drm_prime_remove_imported_buf_handle(&filp->prime,
>                               obj->import_attach->dmabuf);
> +     }
> +     if (obj->export_dma_buf) {
> +             drm_prime_remove_imported_buf_handle(&filp->prime,
> +                             obj->export_dma_buf);
> +             obj->export_dma_buf = NULL;

Any particular reason for nullifying the pointer here and nowhere else
along this free path?

> +     }
>  
>       if (dev->driver->gem_close_object)
>               dev->driver->gem_close_object(obj, filp);
> @@ -530,9 +536,15 @@ drm_gem_object_release_handle(int id, void *ptr, void 
> *data)
>       struct drm_gem_object *obj = ptr;
>       struct drm_device *dev = obj->dev;
>  
> -     if (obj->import_attach)
> +     if (obj->import_attach) {
>               drm_prime_remove_imported_buf_handle(&file_priv->prime,
>                               obj->import_attach->dmabuf);
> +     }
> +     if (obj->export_dma_buf) {
> +             drm_prime_remove_imported_buf_handle(&file_priv->prime,
> +                             obj->export_dma_buf);
> +             obj->export_dma_buf = NULL;
> +     }

Strongly suggests wrapping this into a small function to undo prime for
a handle and avoiding the duplicate code.

>  
>       if (dev->driver->gem_close_object)
>               dev->driver->gem_close_object(obj, file_priv);
> diff --git a/drivers/gpu/drm/drm_prime.c b/drivers/gpu/drm/drm_prime.c
> index c35e585..c06f0f0 100644
> --- a/drivers/gpu/drm/drm_prime.c
> +++ b/drivers/gpu/drm/drm_prime.c
> @@ -68,6 +68,7 @@ int drm_gem_prime_handle_to_fd(struct drm_device *dev,
>  {
>       struct drm_gem_object *obj;
>       void *buf;
> +     int ret;
>  
>       obj = drm_gem_object_lookup(dev, file_priv, handle);
>       if (!obj)
> @@ -100,6 +101,17 @@ int drm_gem_prime_handle_to_fd(struct drm_device *dev,
>               obj->export_dma_buf = buf;
>               *prime_fd = dma_buf_fd(buf, flags);
>       }
> +     /* if we've exported this buffer the cheat and add it to the import list
> +      * so we get the correct handle back
> +      */
s/the/then/
-Chris

-- 
Chris Wilson, Intel Open Source Technology Centre
_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/dri-devel

Reply via email to