[FFmpeg-devel] [PATCH 2/4] avcodec/apedec: Use 64bit in 4/3 calculation in do_apply_filter()

[Anton Khirnov]

Quoting Michael Niedermayer (2021-12-06 12:01:09)
> On Mon, Dec 06, 2021 at 11:23:12AM +0100, Anton Khirnov wrote:
> > Quoting Michael Niedermayer (2021-12-03 18:19:54)
> > > Fixes: Integer overflow
> > > Fixes: 40973/clusterfuzz-testcase-minimized-ffmpeg_AV_CODEC_ID_APE_fuzzer-6739312704618496
> > > 
> > > Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/projects/ffmpeg
> > > Signed-off-by: Michael Niedermayer <michael at niedermayer.cc>
> > > ---
> > >  libavcodec/apedec.c | 2 +-
> > >  1 file changed, 1 insertion(+), 1 deletion(-)
> > > 
> > > diff --git a/libavcodec/apedec.c b/libavcodec/apedec.c
> > > index 9c723f29977..5c3261cf5cb 100644
> > > --- a/libavcodec/apedec.c
> > > +++ b/libavcodec/apedec.c
> > > @@ -1337,7 +1337,7 @@ static void do_apply_filter(APEContext *ctx, int version, APEFilter *f,
> > >              absres = FFABSU(res);
> > >              if (absres)
> > >                  *f->adaptcoeffs = APESIGN(res) *
> > > -                                  (8 << ((absres > f->avg * 3LL) + (absres > (f->avg + f->avg / 3))));
> > > +                                  (8 << ((absres > f->avg * 3LL) + (absres > ((int64_t)f->avg + f->avg / 3))));
> > >                  /* equivalent to the following code
> > >                      if (absres <= f->avg * 4 / 3)
> > >                          *f->adaptcoeffs = APESIGN(res) * 8;
> > > -- 
> > > 2.17.1
> > 
> > Does this not assume a 32bit int?
> 
> hmm
> you mean avg could overflow 64bit ?
> iam not sure that can happen but
> we could make avg int32 or uint32 ?

Seems it cannot, but it's not completely obvious. So changing it to
uint32 might make sense, since it cannot be negative.

-- 
Anton Khirnov