This is needed to get compatibility with the behavior before the
recent decode.c restructuring merge, and fixes fate failures with
this:
make fate-h264-attachment-631 THREADS=32
For every 4 threads, a frame is dropped at the point at which
draining is initialized. The problem starts at THREADS=4.
This patch "fixes" it by ignoring errors during draining (if there
is a frame to be returned), but there is probably a deeper cause
and/or a better fix for this. I haven't looked closer at this, but
was asked to send this patch for discussion.
---
libavcodec/pthread_frame.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/libavcodec/pthread_frame.c b/libavcodec/pthread_frame.c
index 13d682842d..e7fa503adf 100644
--- a/libavcodec/pthread_frame.c
+++ b/libavcodec/pthread_frame.c
@@ -548,7 +548,7 @@ int ff_thread_decode_frame(AVCodecContext *avctx,
fctx->next_finished = finished;
/* return the size of the consumed packet if no error occurred */
- if (err >= 0)
+ if (err >= 0 || (!avpkt->size && *got_picture_ptr))
err = avpkt->size;
finish:
async_lock(fctx);
--
2.11.0
_______________________________________________
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
