[FFmpeg-devel] [PATCH 1/5] libavcodec/jpeg2000.c: Precinct size check removed

Gautam Ramakrishnan gautamramk at gmail.com
Mon Jun 29 19:14:50 EEST 2020


On Mon, Jun 29, 2020 at 12:26 AM Michael Niedermayer
<michael at niedermayer.cc> wrote:
>
> On Sat, Jun 27, 2020 at 09:14:26PM +0530, Gautam Ramakrishnan wrote:
> > On Sat, Jun 27, 2020 at 7:36 PM Michael Niedermayer
> > <michael at niedermayer.cc> wrote:
> > >
> > > On Sat, Jun 27, 2020 at 04:49:49PM +0530, Gautam Ramakrishnan wrote:
> > > > On Sat, Jun 27, 2020 at 3:13 PM Michael Niedermayer
> > > > <michael at niedermayer.cc> wrote:
> > > > >
> > > > > On Mon, Jun 22, 2020 at 12:12:04AM +0530, gautamramk at gmail.com wrote:
> > > > > > From: Gautam Ramakrishnan <gautamramk at gmail.com>
> > > > > >
> > > > > > This patch removes a check which throws an error if
> > > > > > the log2 precinct width/height is 0. The standard allows
> > > > > > the first component to have 0 as the log2 width/height.
> > > > > > ---
> > > > > >  libavcodec/jpeg2000.c | 3 ---
> > > > > >  1 file changed, 3 deletions(-)
> > > > > >
> > > > > > diff --git a/libavcodec/jpeg2000.c b/libavcodec/jpeg2000.c
> > > > > > index 73206d17f3..1aca31ffa4 100644
> > > > > > --- a/libavcodec/jpeg2000.c
> > > > > > +++ b/libavcodec/jpeg2000.c
> > > > > > @@ -509,9 +509,6 @@ int ff_jpeg2000_init_component(Jpeg2000Component *comp,
> > > > > >          // update precincts size: 2^n value
> > > > > >          reslevel->log2_prec_width  = codsty->log2_prec_widths[reslevelno];
> > > > > >          reslevel->log2_prec_height = codsty->log2_prec_heights[reslevelno];
> > > > > > -        if (!reslevel->log2_prec_width || !reslevel->log2_prec_height) {
> > > > > > -            return AVERROR_INVALIDDATA;
> > > > > > -        }
> > > > >
> > > > > This checked that log2_prec_width... has been initialized.
> > > > > Is there some other check that ensures this is not just 0 from allocation
> > > > > which IIUC is not an allowed path in the spec
> > > >
> > > > A check happens only when the COD and COC marker gets read.
> > >
> > > is there a check that protects against the absence of these markers ?
> > >
> > No, there is no check for this. Now I realize that this is an
> > important check which is
> > missing. Same would apply to other compulsory markers like the SIZ marker.
> > Shall I add a check that would ensure that these markers are present before the
> > tile part data starts?
>
> yes
Looks like the check for missing SIZ marker is already there. I shall
submit a check to ensure
the coding style values are initialized.




-- 
-------------
Gautam |


More information about the ffmpeg-devel mailing list