Hi. Mozilla also fixed some data races in pthread_frames.c which you may want to look at. It's very similar to problem you mentioned
We made the code conditional on being in an TSan build. https://dxr.mozilla.org/mozilla-central/source/media/ffvpx/libavcodec/pthread_frame.c#46 Sorry, using the new gmail app, not allowing me to reply inline. I can only top post. Jean-Yves Le jeu. 24 nov. 2016 à 06:47, Wan-Teh Chang <wtc-at-google....@ffmpeg.org> a écrit : > On Tue, Nov 22, 2016 at 3:30 PM, wm4 <nfx...@googlemail.com> wrote: > > On Tue, 22 Nov 2016 23:57:15 +0100 > > Michael Niedermayer <mich...@niedermayer.cc> wrote: > > > >> For example the progress code in the frame threading. > > > > Which was recently fixed in Libav AFAIR... > > You're right. libav/libavcodec/pthread_frame.c has code similar to my > ffmpeg patch > http://ffmpeg.org/pipermail/ffmpeg-devel/2016-March/190454.html, > and much more. > > Note: libav/libavcodec/pthread_frame.c uses unnecessary (too strong) > memory barriers in ff_thread_report_progress(). We can fix those when > the code is merged to ffmpeg. > > Wan-Teh Chang > _______________________________________________ > ffmpeg-devel mailing list > ffmpeg-devel@ffmpeg.org > http://ffmpeg.org/mailman/listinfo/ffmpeg-devel > _______________________________________________ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
