[FFmpeg-devel] [PATCH 02/13] avformat/mxfdec: Check run_in to fit in int and be valid

[Michael Niedermayer]

On Wed, Sep 21, 2022 at 03:25:33PM +0200, Tomas Härdin wrote:
> ons 2022-09-21 klockan 11:35 +0200 skrev Michael Niedermayer:
> > On Tue, Sep 20, 2022 at 01:20:00PM +0200, Tomas Härdin wrote:
> > > tis 2022-09-20 klockan 13:07 +0200 skrev Tomas Härdin:
> > > > sön 2022-09-18 klockan 19:13 +0200 skrev Michael Niedermayer:
> > > > > Fixes: signed integer overflow: 9223372036854775807 - -
> > > > > 2146905566
> > > > > cannot be represented in type 'long'
> > > > > Fixes: 50993/clusterfuzz-testcase-minimized-
> > > > > ffmpeg_dem_MXF_fuzzer-
> > > > > 6570996594769920
> > > > > 
> > > > > Found-by: continuous fuzzing process 
> > > > > https://github.com/google/oss-fuzz/tree/master/projects/ffmpeg
> > > > > Signed-off-by: Michael Niedermayer <michael at niedermayer.cc>
> > > > > ---
> > > > >  libavformat/mxfdec.c | 6 +++++-
> > > > >  1 file changed, 5 insertions(+), 1 deletion(-)
> > > > > 
> > > > > diff --git a/libavformat/mxfdec.c b/libavformat/mxfdec.c
> > > > > index e63e803aa56..da81fea3bc1 100644
> > > > > --- a/libavformat/mxfdec.c
> > > > > +++ b/libavformat/mxfdec.c
> > > > > @@ -3681,6 +3681,7 @@ static int
> > > > > mxf_read_header(AVFormatContext
> > > > > *s)
> > > > >      KLVPacket klv;
> > > > >      int64_t essence_offset = 0;
> > > > >      int ret;
> > > > > +    int64_t run_in;
> > > > >  
> > > > >      mxf->last_forward_tell = INT64_MAX;
> > > > >  
> > > > > @@ -3690,7 +3691,10 @@ static int
> > > > > mxf_read_header(AVFormatContext
> > > > > *s)
> > > > >      }
> > > > >      avio_seek(s->pb, -14, SEEK_CUR);
> > > > >      mxf->fc = s;
> > > > > -    mxf->run_in = avio_tell(s->pb);
> > > > > +    run_in = avio_tell(s->pb);
> > > > > +    if (run_in < 0 || run_in != (int)run_in)
> > > > 
> > > > run_in > INT_MAX is more clear
> > > > 
> > > > It strikes me that run_in is also used in lots of places in the
> > > > demuxer
> > > > without checking for overflow
> > > 
> > > I went and checked S377m and the run-in sequence "shall be less
> > > than
> > > 65536 bytes long". Both the 2004 and 2009 version of the spec agree
> > > on
> > > this. So we should reject run_in >= 65536, and mxf_probe() should
> > > be
> > > similarly adjusted.
> > 
> > ok, will do
> > 
> > thx for checking
> > 
> > i will change the patch by:
> > @@ -3717,7 +3717,7 @@ static int mxf_read_header(AVFormatContext *s)
> >      avio_seek(s->pb, -14, SEEK_CUR);
> 
> Oh and also the call to mxf_read_sync() could be supplied with a
> maximum number of bytes to read, allowing the code to bail out faster

I agree, but that should be seperate from the initial bugfix

thx

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

If a bugfix only changes things apparently unrelated to the bug with no
further explanation, that is a good sign that the bugfix is wrong.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 195 bytes
Desc: not available
URL: <https://ffmpeg.org/pipermail/ffmpeg-devel/attachments/20220921/4cd3a6bc/attachment.sig>