[FFmpeg-devel] [PATCH] avcodec: Avoid splitting side data repeatedly

James Almer jamrial at gmail.com
Thu May 4 18:41:49 EEST 2017


On 5/4/2017 12:15 PM, Michael Niedermayer wrote:
> Fixes Timeout
> Fixes: 508/clusterfuzz-testcase-6245747678773248
> 
> Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/targets/ffmpeg
> Signed-off-by: Michael Niedermayer <michael at niedermayer.cc>
> ---
>  libavcodec/avpacket.c | 24 ++++++++++++++++++++++++
>  libavcodec/decode.c   |  9 +++++++--
>  libavcodec/internal.h |  4 ++++
>  3 files changed, 35 insertions(+), 2 deletions(-)
> 
> diff --git a/libavcodec/avpacket.c b/libavcodec/avpacket.c
> index d97400eaef..de12a1f1ce 100644
> --- a/libavcodec/avpacket.c
> +++ b/libavcodec/avpacket.c
> @@ -463,6 +463,30 @@ int av_packet_split_side_data(AVPacket *pkt){
>      return 0;
>  }
>  
> +int ff_packet_split_and_drop_side_data(AVPacket *pkt){
> +    if (!pkt->side_data_elems && pkt->size >12 && AV_RB64(pkt->data + pkt->size - 8) == FF_MERGE_MARKER){
> +        int i;
> +        unsigned int size;
> +        uint8_t *p;
> +
> +        p = pkt->data + pkt->size - 8 - 5;
> +        for (i=1; ; i++){
> +            size = AV_RB32(p);
> +            if (size>INT_MAX - 5 || p - pkt->data < size)
> +                return 0;
> +            if (p[4]&128)
> +                break;
> +            if (p - pkt->data < size + 5)
> +                return 0;
> +            p-= size+5;
> +        }
> +        pkt->size = p - pkt->data - size;
> +        av_assert0(pkt->size >= 0);
> +        return 1;
> +    }
> +    return 0;
> +}
> +
>  #endif
>  
>  uint8_t *av_packet_pack_dictionary(AVDictionary *dict, int *size)
> diff --git a/libavcodec/decode.c b/libavcodec/decode.c
> index 708071fd07..43ba04550a 100644
> --- a/libavcodec/decode.c
> +++ b/libavcodec/decode.c
> @@ -392,7 +392,9 @@ static int decode_simple_internal(AVCodecContext *avctx, AVFrame *frame)
>      tmp = *pkt;
>  #if FF_API_MERGE_SD
>  FF_DISABLE_DEPRECATION_WARNINGS
> -    did_split = av_packet_split_side_data(&tmp);
> +    did_split = avci->compat_decode_partial_size ?
> +                ff_packet_split_and_drop_side_data(&tmp) :
> +                av_packet_split_side_data(&tmp);

Eh, this is all going away as soon as we bump major and will not make it
to any upcoming release, so guess it's fine for now.

>  
>      if (did_split) {
>          ret = extract_packet_props(avctx->internal, &tmp);
> @@ -961,6 +963,7 @@ int avcodec_decode_subtitle2(AVCodecContext *avctx, AVSubtitle *sub,
>                               AVPacket *avpkt)
>  {
>      int i, ret = 0;
> +    AVCodecInternal *avci = avctx->internal;
>  
>      if (!avpkt->data && avpkt->size) {
>          av_log(avctx, AV_LOG_ERROR, "invalid packet: NULL data, size != 0\n");
> @@ -981,7 +984,9 @@ int avcodec_decode_subtitle2(AVCodecContext *avctx, AVSubtitle *sub,
>          AVPacket tmp = *avpkt;
>  #if FF_API_MERGE_SD
>  FF_DISABLE_DEPRECATION_WARNINGS
> -        int did_split = av_packet_split_side_data(&tmp);
> +        int did_split = avci->compat_decode_partial_size ?
> +                        ff_packet_split_and_drop_side_data(&tmp) :
> +                        av_packet_split_side_data(&tmp);
>          //apply_param_change(avctx, &tmp);
>  
>          if (did_split) {
> diff --git a/libavcodec/internal.h b/libavcodec/internal.h
> index e30d4aa73d..f3d10a14f1 100644
> --- a/libavcodec/internal.h
> +++ b/libavcodec/internal.h
> @@ -356,6 +356,10 @@ int ff_set_sar(AVCodecContext *avctx, AVRational sar);
>  int ff_side_data_update_matrix_encoding(AVFrame *frame,
>                                          enum AVMatrixEncoding matrix_encoding);
>  
> +#if FF_API_MERGE_SD_API
> +int ff_packet_split_and_drop_side_data(AVPacket *pkt);
> +#endif

This function will not be used by anything after the ABI part of this
functionality is dropped, so maybe make it depend on FF_API_MERGE_SD
instead, here and in avpacket.c

> +
>  /**
>   * Select the (possibly hardware accelerated) pixel format.
>   * This is a wrapper around AVCodecContext.get_format() and should be used
> 



More information about the ffmpeg-devel mailing list