[PATCH 2/2] amdgpu: serialize drmPrimeFDToHandle

[Alex Deucher]

On Mon, Aug 24, 2015 at 9:55 AM, Zhou, Jammy <Jammy.Zhou at amd.com> wrote:
> Both patches are Reviewed-by: Jammy Zhou <Jammy.Zhou at amd.com>

I pushed the series.  thanks!

Alex

>
> Regards,
> Jammy
>
> -----Original Message-----
> From: dri-devel [mailto:dri-devel-bounces at lists.freedesktop.org] On Behalf Of Christian K?nig
> Sent: Monday, August 24, 2015 5:44 PM
> To: dri-devel at lists.freedesktop.org
> Subject: [PATCH 2/2] amdgpu: serialize drmPrimeFDToHandle
>
> From: Christian König <christian.koenig at amd.com>
>
> Fixes the same problem as "intel: Serialize drmPrimeFDToHandle with struct_mutex".
>
> Signed-off-by: Christian König <christian.koenig at amd.com>
> ---
>  amdgpu/amdgpu_bo.c | 9 +++++----
>  1 file changed, 5 insertions(+), 4 deletions(-)
>
> diff --git a/amdgpu/amdgpu_bo.c b/amdgpu/amdgpu_bo.c index dab3804..adf4253 100644
> --- a/amdgpu/amdgpu_bo.c
> +++ b/amdgpu/amdgpu_bo.c
> @@ -289,6 +289,10 @@ int amdgpu_bo_import(amdgpu_device_handle dev,
>         int dma_fd;
>         uint64_t dma_buf_size = 0;
>
> +       /* We must maintain a list of pairs <handle, bo>, so that we always
> +        * return the same amdgpu_bo instance for the same handle. */
> +       pthread_mutex_lock(&dev->bo_table_mutex);
> +
>         /* Convert a DMA buf handle to a KMS handle now. */
>         if (type == amdgpu_bo_handle_type_dma_buf_fd) {
>                 uint32_t handle;
> @@ -303,6 +307,7 @@ int amdgpu_bo_import(amdgpu_device_handle dev,
>                 /* Query the buffer size. */
>                 size = lseek(shared_handle, 0, SEEK_END);
>                 if (size == (off_t)-1) {
> +                       pthread_mutex_unlock(&dev->bo_table_mutex);
>                         amdgpu_close_kms_handle(dev, handle);
>                         return -errno;
>                 }
> @@ -312,10 +317,6 @@ int amdgpu_bo_import(amdgpu_device_handle dev,
>                 shared_handle = handle;
>         }
>
> -       /* We must maintain a list of pairs <handle, bo>, so that we always
> -        * return the same amdgpu_bo instance for the same handle. */
> -       pthread_mutex_lock(&dev->bo_table_mutex);
> -
>         /* If we have already created a buffer with this handle, find it. */
>         switch (type) {
>         case amdgpu_bo_handle_type_gem_flink_name:
> --
> 1.9.1
>
> _______________________________________________
> dri-devel mailing list
> dri-devel at lists.freedesktop.org
> http://lists.freedesktop.org/mailman/listinfo/dri-devel
> _______________________________________________
> dri-devel mailing list
> dri-devel at lists.freedesktop.org
> http://lists.freedesktop.org/mailman/listinfo/dri-devel