[FFmpeg-devel] [PATCH 2/5] avcodec/atrac9dec: Check conditions before apply_band_extension() to avoid out of array read in initialization of unused variables

[Michael Niedermayer]

On Sun, Jun 16, 2019 at 12:20:35PM +0200, Lynne wrote:
> Jun 15, 2019, 11:00 PM by michael at niedermayer.cc:
> 
> > Fixes: global-buffer-overflow
> > Fixes: 15247/clusterfuzz-testcase-minimized-ffmpeg_AV_CODEC_ID_ATRAC9_fuzzer-5671602181636096
> >
> > 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/atrac9dec.c | 7 +++----
> >  1 file changed, 3 insertions(+), 4 deletions(-)
> >
> > diff --git a/libavcodec/atrac9dec.c b/libavcodec/atrac9dec.c
> > index 805d46f3b8..5401d6e19e 100644
> > --- a/libavcodec/atrac9dec.c
> > +++ b/libavcodec/atrac9dec.c
> > @@ -535,9 +535,6 @@ static inline void apply_band_extension(ATRAC9Context *s, ATRAC9BlockData *b,
> >  at9_q_unit_to_coeff_idx[g_units[3]],
> >  };
> >  
> > -    if (!b->has_band_ext || !b->has_band_ext_data)
> > -        return;
> > -
> >  for (int ch = 0; ch <= stereo; ch++) {
> >  ATRAC9ChannelData *c = &b->channel[ch];
> >  
> > @@ -741,7 +738,9 @@ static int atrac9_decode_block(ATRAC9Context *s, GetBitContext *gb,
> >  
> >  apply_intensity_stereo(s, b, stereo);
> >  apply_scalefactors    (s, b, stereo);
> > -    apply_band_extension  (s, b, stereo);
> > +
> > +    if (b->has_band_ext && b->has_band_ext_data)
> > +        apply_band_extension  (s, b, stereo); 
> >
> 
> False positive as usual, q_unit_cnt can't be anything out of array since its looked up from
> at9_tab_band_q_unit_map.
> I'd really appreciate it if you stopped fixing complaint messages from automated tools.
> Especially from overflows and fuzzing timeouts. The latter are completely useless and
> often make the code look worse and weird, and the former are all useless except when
> outside of DSP code (e.g. malloc). And most of our code is DSP.

Calm down please, ill explain how this is reading out of array

In fact there seem to be more ways than i realized before that this can read
out of array, so i will post 2 more patches to fix this more completely

First q_unit_cnt is only set from at9_tab_band_q_unit_map if you are lucky as
the code is conditional on a bit read from the bitstream.

Second the values in at9_tab_band_q_unit_map start like this 0, 4, 8, 10, 12

The code reading out of array is this:

    const int g_units[4] = { /* A, B, C, total units */
        b->q_unit_cnt,
        at9_tab_band_ext_group[b->q_unit_cnt - 13][0],
        at9_tab_band_ext_group[b->q_unit_cnt - 13][1],
        
if q_unit_cnt is less than 13 the index is negative and that reads out of the array
teh value is not used later in the sample but the program could have crashed already

It is very good that we discuss this here though as the fix from the patch
does not appear to be enough. There are more pathes that can lead to this.

Thanks

PS: If anyone has atrac9 samples for testing, or could add atrac9 samples to FATE
that would be very helpfull in ensuring that none of these changes break anything

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

It is what and why we do it that matters, not just one of them.
-------------- 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/20190616/25531f4a/attachment.sig>