[FFmpeg-devel] [PATCH] avcodec/cbs_h266: fix SPS VUI extension data leak

[Nuo Mi]

On Mon, Oct 23, 2023 at 11:24 PM Andreas Rheinhardt <
andreas.rheinhardt at outlook.com> wrote:

> Nuo Mi:
> > On Mon, Oct 23, 2023 at 7:46 PM Andreas Rheinhardt <
> > andreas.rheinhardt at outlook.com> wrote:
> >
> >> Nuo Mi:
> >>> On Sun, Oct 22, 2023 at 8:58 PM Andreas Rheinhardt <
> >>> andreas.rheinhardt at outlook.com> wrote:
> >>>
> >>>> Nuo Mi:
> >>>>> Fixes: VUI extension leak
> >>>>> Fixes:
> >>>>
> >>
> 63004/clusterfuzz-testcase-minimized-ffmpeg_BSF_VVC_METADATA_fuzzer-4928832253329408
> >>>>>
> >>>>> Found-by: continuous fuzzing process
> >>>> https://github.com/google/oss-fuzz/tree/master/projects/ffmpeg
> >>>>> ---
> >>>>>  libavcodec/cbs_h2645.c | 10 +++++++++-
> >>>>>  1 file changed, 9 insertions(+), 1 deletion(-)
> >>>>>
> >>>>> diff --git a/libavcodec/cbs_h2645.c b/libavcodec/cbs_h2645.c
> >>>>> index 470f60b95f..ef631a11fe 100644
> >>>>> --- a/libavcodec/cbs_h2645.c
> >>>>> +++ b/libavcodec/cbs_h2645.c
> >>>>> @@ -1979,6 +1979,13 @@ static const CodedBitstreamUnitTypeDescriptor
> >>>> cbs_h265_unit_types[] = {
> >>>>>      CBS_UNIT_TYPE_END_OF_LIST
> >>>>>  };
> >>>>>
> >>>>> +static void cbs_h266_free_sps(FFRefStructOpaque unused, void
> *content)
> >>>>> +{
> >>>>> +    H266RawSPS *sps = (H266RawSPS*)content;
> >>>>> +    av_buffer_unref(&sps->extension_data.data_ref);
> >>>>> +    av_buffer_unref(&sps->vui.extension_data.data_ref);
> >>>>> +}
> >>>>> +
> >>>>>  static void cbs_h266_free_sei(FFRefStructOpaque unused, void
> *content)
> >>>>>  {
> >>>>>      H266RawSEI *sei = content;
> >>>>> @@ -1989,7 +1996,6 @@ static const CodedBitstreamUnitTypeDescriptor
> >>>> cbs_h266_unit_types[] = {
> >>>>>      CBS_UNIT_TYPE_INTERNAL_REF(VVC_DCI_NUT, H266RawDCI,
> >>>> extension_data.data),
> >>>>>      CBS_UNIT_TYPE_INTERNAL_REF(VVC_OPI_NUT, H266RawOPI,
> >>>> extension_data.data),
> >>>>>      CBS_UNIT_TYPE_INTERNAL_REF(VVC_VPS_NUT, H266RawVPS,
> >>>> extension_data.data),
> >>>>> -    CBS_UNIT_TYPE_INTERNAL_REF(VVC_SPS_NUT, H266RawSPS,
> >>>> extension_data.data),
> >>>>>      CBS_UNIT_TYPE_INTERNAL_REF(VVC_PPS_NUT, H266RawPPS,
> >>>> extension_data.data),
> >>>>>      CBS_UNIT_TYPE_INTERNAL_REF(VVC_PREFIX_APS_NUT, H266RawAPS,
> >>>> extension_data.data),
> >>>>>      CBS_UNIT_TYPE_INTERNAL_REF(VVC_SUFFIX_APS_NUT, H266RawAPS,
> >>>> extension_data.data),
> >>>>> @@ -2003,6 +2009,8 @@ static const CodedBitstreamUnitTypeDescriptor
> >>>> cbs_h266_unit_types[] = {
> >>>>>      CBS_UNIT_RANGE_INTERNAL_REF(VVC_IDR_W_RADL, VVC_GDR_NUT,
> >>>>>                                  H266RawSlice, data),
> >>>>>
> >>>>> +    CBS_UNIT_TYPE_COMPLEX(VVC_SPS_NUT, H266RawSPS,
> cbs_h266_free_sps),
> >>>>> +
> >>>>>      CBS_UNIT_TYPES_COMPLEX((VVC_PREFIX_SEI_NUT, VVC_SUFFIX_SEI_NUT),
> >>>>>                             H266RawSEI, cbs_h266_free_sei),
> >>>>>
> >>>>
> >>>> This should be a CBS_UNIT_TYPE_INTERNAL_REF().
> >>>>
> >>> Hi Andreas,
> >>> Thank you for the review.
> >>>
> >>> Do you mean the SEI?
> >>>
> >>
> >> No, I mean the SPS. (If you want to, I can send a patch for this leak.)
> >>
> > Sure, please go ahead.
> >
> > Thank you
> >
>
> Sent a patch. Rereading this I came to realize that my advice was not
> helpful, given that VVC_SPS_NUT already uses CBS_UNIT_TYPE_INTERNAL_REF.
> The solution is to use CBS_CONTENT_TYPE_INTERNAL_REFS, but with two
> internal refs. There is no macro for this atm (and the only other unit
> with two internal refs uses CBS_UNIT_TYPE_RANGE, so one can't simply use
> a common macro for them).
>
Thank you for the patch
 I used CBS_UNIT_TYPE_COMPLEX because It can help readers focus on
cbs_h266_free_sps. no need to check the  cbs_h266_unit_types details.

>
> - Andreas
>
> _______________________________________________
> ffmpeg-devel mailing list
> ffmpeg-devel at ffmpeg.org
> https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
>
> To unsubscribe, visit link above, or email
> ffmpeg-devel-request at ffmpeg.org with subject "unsubscribe".
>