[FFmpeg-devel] [PATCH] avformat/utils: Change compute_chapters_end() from O(n²) to O(n log n)

[Michael Niedermayer]

On Sat, Nov 21, 2020 at 07:32:31PM +0100, Lynne wrote:
> Nov 21, 2020, 18:24 by michael at niedermayer.cc:
> 
> > Fixes: Timeout (49sec -> 9sec)
> > Fixes: 27427/clusterfuzz-testcase-minimized-ffmpeg_dem_FFMETADATA_fuzzer-5140589838073856
> >
> > 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/utils.c | 43 +++++++++++++++++++++++++++++++++++++++++++
> >  1 file changed, 43 insertions(+)
> >
> > diff --git a/libavformat/utils.c b/libavformat/utils.c
> > index 503e583ad0..9fac3fc2aa 100644
> > --- a/libavformat/utils.c
> > +++ b/libavformat/utils.c
> > @@ -3191,15 +3191,58 @@ enum AVCodecID av_codec_get_id(const AVCodecTag *const *tags, unsigned int tag)
> >  return AV_CODEC_ID_NONE;
> >  }
> >  
> > +static int chapter_start_cmp(const void *p1, const void *p2)
> > +{
> > +    AVChapter *ch1 = *(AVChapter**)p1;
> > +    AVChapter *ch2 = *(AVChapter**)p2;
> > +    int delta = av_compare_ts(ch1->start, ch1->time_base, ch2->start, ch2->time_base);
> > +    if (delta)
> > +        return delta;
> > +    return (ch1 > ch2) - (ch1 < ch2);
> > +}
> > +
> >  static void compute_chapters_end(AVFormatContext *s)
> >  {
> >  unsigned int i, j;
> >  int64_t max_time = 0;
> > +    int computations = 0;
> >  
> >  if (s->duration > 0 && s->start_time < INT64_MAX - s->duration)
> >  max_time = s->duration +
> >  ((s->start_time == AV_NOPTS_VALUE) ? 0 : s->start_time);
> >  
> > +    for (i = 0; i < s->nb_chapters; i++)
> > +        if (s->chapters[i]->end == AV_NOPTS_VALUE)
> > +            computations ++;
> > +
> > +    if (computations > 5) {
> > +        AVChapter **timetable = av_malloc(s->nb_chapters * sizeof(*timetable));
> > +        if (timetable) {
> >
> 
> Its a void function, but shouldn't you change it to make it return
> AVERROR(ENOMEM) and then handle that in the caller?
> No memory is pretty catastrophic.

you are probably correct, that fallback isnt worth it, ill drop it and
return failure

thx

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

The bravest are surely those who have the clearest vision
of what is before them, glory and danger alike, and yet
notwithstanding go out to meet it. -- Thucydides
-------------- 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/20201121/0a080308/attachment.sig>