On Fri, Aug 04, 2023 at 06:19:46PM -0700, Pierre-Anthony Lemieux wrote:
> On Tue, Aug 1, 2023 at 5:02 PM Michael Niedermayer
> <michael@niedermayer.cc> wrote:
> >
> > This also fixes assertion failures
> >
> > Fixes: shift exponent 95 is too large for 64-bit type 'unsigned long long'
> > Fixes: 58299/clusterfuzz-testcase-minimized-ffmpeg_AV_CODEC_ID_JPEG2000_fuzzer-5828618092937216
> >
> > Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/projects/ffmpeg
> > Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
> > ---
> >  libavcodec/jpeg2000htdec.c | 22 ++++++++++++++++++++--
> >  1 file changed, 20 insertions(+), 2 deletions(-)
> >
> > diff --git a/libavcodec/jpeg2000htdec.c b/libavcodec/jpeg2000htdec.c
> > index 3985783f3a..ae2ee6d6ee 100644
> > --- a/libavcodec/jpeg2000htdec.c
> > +++ b/libavcodec/jpeg2000htdec.c
> > @@ -689,6 +689,10 @@ static int jpeg2000_decode_ht_cleanup_segment(const Jpeg2000DecoderContext *s,
> >          for (int i = 0; i < 4; i++) {
> >              m[J2K_Q1][i] = sigma_n[4 * q1 + i] * U[J2K_Q1] - ((emb_pat_k[J2K_Q1] >> i) & 1);
> >              m[J2K_Q2][i] = sigma_n[4 * q2 + i] * U[J2K_Q2] - ((emb_pat_k[J2K_Q2] >> i) & 1);
> > +            if (m[J2K_Q1][i] > 63 || m[J2K_Q2][i] > 63) {
> 
> AFAIK, m[i], which is m_n in the standard, can never be larger than
> the sample bit depth (including the sign bit, if any). Is it worth
> comparing it to a value more precise than 63?

probably, yes
I think you know the spec better than i do, so you can probably pick
the tightest bound quicker ...
can you submit a patch doing that ?

thx

[...]
-- 
Michael     GnuPG fingerprint: 9FF2128B147EF6730BADF133611EC787040B0FAB

The greatest way to live with honor in this world is to be what we pretend
to be. -- Socrates