[FFmpeg-devel] [PATCH] pthread_frame: don't return stale error codes after flush

[wm4]

On Thu, 6 Apr 2017 11:38:55 -0400
"Ronald S. Bultje" <rsbultje at gmail.com> wrote:

> Hi,
> 
> On Thu, Apr 6, 2017 at 11:36 AM, wm4 <nfxjfg at googlemail.com> wrote:
> 
> > Consider the following sequence of events:
> >
> > - open a codec without AV_CODEC_CAP_DELAY
> > - decode call fails with an error
> > - ff_thread_flush() is called
> > - drain packet is sent
> >
> > Then the last step would make ff_thread_decode_frame() return an error,
> > because p->result can still be set to an error value. This is because
> > submit_packet returns immediately if AV_CODEC_CAP_DELAY is not set, and
> > no worker thread gets the chance to reset p->result, yet its value is
> > trusted by ff_thread_decode_frame().
> >
> > Fix this by clearing the error fields on flush.
> > ---
> >  libavcodec/pthread_frame.c | 1 +
> >  1 file changed, 1 insertion(+)
> >
> > diff --git a/libavcodec/pthread_frame.c b/libavcodec/pthread_frame.c
> > index 9a6b83ac45..7586f00bec 100644
> > --- a/libavcodec/pthread_frame.c
> > +++ b/libavcodec/pthread_frame.c
> > @@ -859,6 +859,7 @@ void ff_thread_flush(AVCodecContext *avctx)
> >          // Make sure decode flush calls with size=0 won't return old
> > frames
> >          p->got_frame = 0;
> >          av_frame_unref(p->frame);
> > +        p->result = 0;
> >
> >          release_delayed_buffers(p);
> >
> > --
> > 2.11.0  
> 
> 
> Nice catch - I think that looks good.
> 
> Ronald

Pushed.