[FFmpeg-devel] [PATCH] 2 alternative ways to check in vp9 decode_tiles() if there is data remaining

Michael Niedermayer michael at niedermayer.cc
Mon Oct 29 22:30:17 EET 2018


On Sat, Oct 20, 2018 at 12:42:35PM +0200, Michael Niedermayer wrote:
> Hi
> 
> 2 alternative patchsets are attached to fix $SUBJ
> 
> The 2 alternatives should behave similar.
> 
> The first adds a function to check if the next range coder symbol read would
> trigger the end of input case.
> We then error out before reading in case the read would trigger this case
> 
> The second sets a flag if the end of input case triggered and subsequently
> errors out
> 
> The second case should be slower as it requires additional checks in inner
> loops, but i was asked to implement this with a flag, so i implemented both
> ways.
> 
> Which version, if any, should i apply ?

this also fixes 9775/clusterfuzz-testcase-minimized-ffmpeg_AV_CODEC_ID_VP9_fuzzer-5643845344690176
ill apply the one that avoids checks in the inner loop.
If people prefer the other iam happy to revert it and replace it by the
other solution. But i dont want to leave the issue open 

Thanks
[...]

-- 
Michael     GnuPG fingerprint: 9FF2128B147EF6730BADF133611EC787040B0FAB

"Nothing to hide" only works if the folks in power share the values of
you and everyone you know entirely and always will -- Tom Scott

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 181 bytes
Desc: not available
URL: <http://ffmpeg.org/pipermail/ffmpeg-devel/attachments/20181029/856bb74f/attachment.sig>


More information about the ffmpeg-devel mailing list