[FFmpeg-devel] [PATCH v2] avcodec/h264_parse: retry decoding SPS with complete NAL

[James Almer]

On 8/19/2019 11:24 PM, Jun Li wrote:
> Fix #6591
> The content has no rbsp_stop_one_bit for ending the SPS, that
> causes the decoding SPS failure, results decoding frame failure as well.
> 
> The patch is just adding a retry with complete NALU, copied from the retry in decode_nal_unit()
> ---
>  libavcodec/h264_parse.c | 11 +++++++++--
>  1 file changed, 9 insertions(+), 2 deletions(-)
> 
> diff --git a/libavcodec/h264_parse.c b/libavcodec/h264_parse.c
> index ac31f54e07..ea7a958dd9 100644
> --- a/libavcodec/h264_parse.c
> +++ b/libavcodec/h264_parse.c
> @@ -376,8 +376,15 @@ static int decode_extradata_ps(const uint8_t *data, int size, H264ParamSets *ps,
>          switch (nal->type) {
>          case H264_NAL_SPS:
>              ret = ff_h264_decode_seq_parameter_set(&nal->gb, logctx, ps, 0);
> -            if (ret < 0)
> -                goto fail;
> +            if (ret < 0) {
> +                GetBitContext tmp_gb = nal->gb;

nal->gb will already be changed here and will not be pointing to the
start of the SPS.

> +                av_log(logctx, AV_LOG_DEBUG,
> +                   "SPS decoding failure, trying again with the complete NAL\n");
> +                init_get_bits8(&tmp_gb, nal->raw_data + 1, nal->raw_size - 1);

And you're overwriting it here, so the above assignment was pointless.

> +                if ((ret = ff_h264_decode_seq_parameter_set(&tmp_gb, logctx, ps, 0)) < 0 &&
> +                    (ret = ff_h264_decode_seq_parameter_set(&tmp_gb, logctx, ps, 1)) < 0)

Similarly, tmp_gb will be not be pointing to the start of the SPS if the
second ff_h264_decode_seq_parameter_set() call (with truncation enabled)
is ever executed.

> +                    goto fail;
> +            }
>              break;
>          case H264_NAL_PPS:
>              ret = ff_h264_decode_picture_parameter_set(&nal->gb, logctx, ps,
>