On Fri, Dec 09, 2016 at 11:35:08AM +0100, Nicolas George wrote: > Le nonidi 19 frimaire, an CCXXV, Michael Niedermayer a écrit : > > > > A. Is a heap limit for av_*alloc*() acceptable ? > > > moving the threshold of where to declare something OOM or hang around > > will not solve this. > > Yet this is your "A" proposal. Or am I misunderstanding you?
no, i missed that "A" alone would not solve the fuzzer aspect of this somewere in what i wrote ... if thats what you mean then you are correct "A" would mainly help for user app being crashed with the lib OOM "B" would mainly help the fuzzer issue but also in a more limited sense reduce the crash issue. Here even with seperate processes one probably prefers processes not maxing out their OS limits. A webpage with thousand images (not unlikely) would either have a thousand processes or one taking down all the image decoding. Whichever way its implemented more robustness against OOM and hangs is a good thing Also we could declare some decoders with capability flags as safe to be used in the same process. For example the simple image decoders can surely be made to be safe with just a max_pixel limit, and that should have users who would prefer not to need a seperate process. [...] -- Michael GnuPG fingerprint: 9FF2128B147EF6730BADF133611EC787040B0FAB Observe your enemies, for they first find out your faults. -- Antisthenes
signature.asc
Description: Digital signature
_______________________________________________ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
