[FFmpeg-devel] [PATCH] lavc/pthread_frame: Update user context in ff_frame_thread_free

Anton Khirnov anton at khirnov.net
Thu Mar 26 12:32:14 EET 2020


Quoting Linjie Fu (2019-12-27 09:47:35)
> Resolution/format changes lead to re-initialization of hardware
> accelerations(vaapi/dxva2/..) with new hwaccel_priv_data in
> the worker-thread. But hwaccel_priv_data in user context won't
> be updated until the resolution changing frame is output.
> 
> A termination with "-vframes" just after the reinit will lead to:
>     1. memory leak in worker-thread.
>     2. double free in user-thread.
> 
> Update user context in ff_frame_thread_free with the last thread
> submit_packet() was called on.
> 
> To reproduce:
> ffmpeg -hwaccel vaapi(dxva2) -v verbose -i
>     fate-suite/h264/reinit-large_420_8-to-small_420_8.h264 -pix_fmt nv12
>     -f rawvideo -vsync passthrough -vframes 47 -y out.yuv
> 
> Signed-off-by: Linjie Fu <linjie.fu at intel.com>
> ---
>  libavcodec/pthread_frame.c | 7 +++++++
>  1 file changed, 7 insertions(+)
> 
> diff --git a/libavcodec/pthread_frame.c b/libavcodec/pthread_frame.c
> index 36ac0ac..8bdd735 100644
> --- a/libavcodec/pthread_frame.c
> +++ b/libavcodec/pthread_frame.c
> @@ -657,6 +657,13 @@ void ff_frame_thread_free(AVCodecContext *avctx, int thread_count)
>  
>      park_frame_worker_threads(fctx, thread_count);
>  
> +    if (fctx->prev_thread && avctx->internal->hwaccel_priv_data !=
> +                             fctx->prev_thread->avctx->internal->hwaccel_priv_data) {
> +        if (update_context_from_thread(avctx, fctx->prev_thread->avctx, 1) < 0) {
> +            av_log(avctx, AV_LOG_ERROR, "Failed to update user thread.\n");
> +        }
> +    }
> +
>      if (fctx->prev_thread && fctx->prev_thread != fctx->threads)
>          if (update_context_from_thread(fctx->threads->avctx, fctx->prev_thread->avctx, 0) < 0) {
>              av_log(avctx, AV_LOG_ERROR, "Final thread update failed\n");
> -- 
> 2.7.4

Not an ideal solution - we shouldn't be leaving stale pointers around in
the first place - but I suppose it's good enough for now.

Will push unless someone objects.

-- 
Anton Khirnov


More information about the ffmpeg-devel mailing list