[FFmpeg-devel] [HACK] Remove MAX_STREAMS usages

[Aurelien Jacobs]

On Tue, Aug 10, 2010 at 07:19:06PM +0200, Reimar D?ffinger wrote:
> On Tue, Aug 10, 2010 at 07:31:49PM +0200, Michael Niedermayer wrote:
> > On Tue, Aug 10, 2010 at 01:14:56AM +0200, Aurelien Jacobs wrote:
> > > On Sun, Aug 08, 2010 at 10:40:40PM +0200, Reimar D?ffinger wrote:
> > > > Hello,
> > > > this is just a quick hack that removes all usages of
> > > > MAX_STREAMS.
> > > > 
> > > > [...]
> > > 
> > > And here are the last bits I extracted/reworked from this patch.
> > > 
> > > First patch in this email uses nb_streams instead of MAX_STREAMS to
> > > dynamically alloc arrays in libavformat/utils.c. This is compatible
> > > with both the old API and the new API.
> > > 
> > > Second and last patch actually add the new API, finally getting ride of
> > > MAX_STREAMS. It is obviously disabled until next major bump.
> > > 
> > > Aurel
> > >  utils.c |   37 +++++++++++++++++++++++++++----------
> > >  1 file changed, 27 insertions(+), 10 deletions(-)
> > > 553bd8bafb70b124e3ed2747f577f4ae057e8094  max_stream.diff
> > > commit f04d73bbdcb67af2e6d1145c652a6b2978c0c83f
> > > Author: Aurelien Jacobs <aurel at gnuage.org>
> > > Date:   Tue Aug 10 00:21:06 2010 +0200
> > > 
> > >     dynamically use nb_streams instead of static use of MAX_STREAMS
> > > 
> > > diff --git a/libavformat/utils.c b/libavformat/utils.c
> > > index 1aa965c..edd3d19 100644
> > > --- a/libavformat/utils.c
> > > +++ b/libavformat/utils.c
> > > @@ -1871,10 +1871,13 @@ static void av_estimate_timings_from_pts(AVFormatContext *ic, int64_t old_offset
> > >      AVPacket pkt1, *pkt = &pkt1;
> > >      AVStream *st;
> > >      int read_size, i, ret;
> > > -    int64_t end_time, start_time[MAX_STREAMS];
> > > +    int64_t end_time, *start_time;
> > >      int64_t filesize, offset, duration;
> > >      int retry=0;
> > >  
> > > +    if (!(start_time = av_malloc(ic->nb_streams * sizeof(*start_time))))
> > > +        return;
> > > +
> > >      ic->cur_st = NULL;
> > >  
> > >      /* flush packet queue */
> > > @@ -1935,6 +1938,7 @@ static void av_estimate_timings_from_pts(AVFormatContext *ic, int64_t old_offset
> > >      }while(   end_time==AV_NOPTS_VALUE
> > >             && filesize > (DURATION_MAX_READ_SIZE<<retry)
> > >             && ++retry <= DURATION_MAX_RETRY);
> > > +    av_free(start_time);
> > >  
> > >      fill_all_stream_timings(ic);
> > >  
> > > @@ -2160,15 +2164,23 @@ int av_find_stream_info(AVFormatContext *ic)
> > >      int i, count, ret, read_size, j;
> > >      AVStream *st;
> > >      AVPacket pkt1, *pkt;
> > > -    int64_t last_dts[MAX_STREAMS];
> > > -    int64_t duration_gcd[MAX_STREAMS]={0};
> > > -    int duration_count[MAX_STREAMS]={0};
> > > +    int64_t *last_dts;
> > > +    int64_t *duration_gcd;
> > > +    int *duration_count;
> > >      double (*duration_error)[MAX_STD_TIMEBASES];
> > >      int64_t old_offset = url_ftell(ic->pb);
> > > -    int64_t codec_info_duration[MAX_STREAMS]={0};
> > > +    int64_t *codec_info_duration;
> > >  
> > > -    duration_error = av_mallocz(MAX_STREAMS * sizeof(*duration_error));
> > > -    if (!duration_error) return AVERROR(ENOMEM);
> > > +    last_dts            = av_mallocz(ic->nb_streams * sizeof(*last_dts));
> > > +    duration_gcd        = av_mallocz(ic->nb_streams * sizeof(*duration_gcd));
> > > +    duration_count      = av_mallocz(ic->nb_streams * sizeof(*duration_count));
> > > +    duration_error      = av_mallocz(ic->nb_streams * sizeof(*duration_error));
> > > +    codec_info_duration = av_mallocz(ic->nb_streams * sizeof(*codec_info_duration));
> > > +    if (!last_dts || !duration_gcd || !duration_count || !duration_error ||
> > > +        !codec_info_duration) {
> > > +        ret = AVERROR(ENOMEM);
> > > +        goto err_out;
> > > +    }
> > 
> > possibly exploitable
> > the number of streams can increase any time if more are found
> 
> "any time" makes it sound more serious than it is, within that
> function I think it is only in av_read_frame_internal, so
> after that call resizing the arrays may be necessary.
> Still going to be a bit annoying to implement but not that bad.

Here is a fixed version which will grow the size of the buffers when
new streams are added.

Aurel
-------------- next part --------------
A non-text attachment was scrubbed...
Name: max_stream.diff
Type: text/x-diff
Size: 4089 bytes
Desc: not available
URL: <http://lists.mplayerhq.hu/pipermail/ffmpeg-devel/attachments/20100811/fe1603d1/attachment.diff>