Hi, On Fri, Mar 31, 2017 at 10:08 AM, Michael Niedermayer < mich...@niedermayer.cc> wrote:
> On Wed, Mar 29, 2017 at 10:12:23AM -0400, Ronald S. Bultje wrote: > > This is how the ref list manager links bitstream IDs to H264Picture/Ref > > objects, and is local to the producer thread. There is no need for the > > consumer thread to know the bitstream IDs of its references in their > > respective producer threads. > > > > In practice, this fixes the last tsan warnigns when running fate-h264. > > --- > > libavcodec/h264_picture.c | 1 - > > 1 file changed, 1 deletion(-) > > patch is probably ok though iam not sure what issue tsan has with this In multi-slice files, the second slice will re-do the reflist lookup, which writes to pic_id. At the same time, the next thread would sync pic_id between decoding instances. The write-while-sync is the reported race condition. (The reported race is obviously harmless.) Ronald _______________________________________________ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-devel