[FFmpeg-devel] [PATCH v2] avformat/mpegts: add option max_packet_size

Marton Balint cus at passwd.hu
Sat Jan 15 22:09:39 EET 2022



On Sat, 15 Jan 2022, Gyan Doshi wrote:

> Plan to push tomorrow.
>
> On 2022-01-12 08:57 pm, Gyan Doshi wrote:
>>  Makes maximum size of emitted packet user-tunable.
>>  ---
>>    doc/demuxers.texi    | 4 ++++
>>    libavformat/mpegts.c | 9 ++++++---
>>    2 files changed, 10 insertions(+), 3 deletions(-)
>>
>>  diff --git a/doc/demuxers.texi b/doc/demuxers.texi
>>  index 26ae768d7a..aef5976551 100644
>>  --- a/doc/demuxers.texi
>>  +++ b/doc/demuxers.texi
>>  @@ -775,6 +775,10 @@ disabled). Default value is -1.
>>    @item merge_pmt_versions
>>    Re-use existing streams when a PMT's version is updated and elementary
>>    streams move to different PIDs. Default value is 0.
>>  +
>>  + at item max_packet_size
>>  +Set maximum size, in bytes, of packet emitted by the demuxer. Payloads
>>  above this size
>>  +are split across multiple packets. Range is 1 to INT_MAX. Default is
>>  204800 bytes.
>>    @end table
>>
>>    @section mpjpeg
>>  diff --git a/libavformat/mpegts.c b/libavformat/mpegts.c
>>  index 2479cb6f7d..6ff1a54706 100644
>>  --- a/libavformat/mpegts.c
>>  +++ b/libavformat/mpegts.c
>>  @@ -162,6 +162,7 @@ struct MpegTSContext {
>>
>>        int resync_size;
>>        int merge_pmt_versions;
>>  +    int max_packet_size;
>>
>>        /******************************************/
>>        /* private mpegts data */
>>  @@ -198,6 +199,8 @@ static const AVOption options[] = {
>>        {.i64 = 0}, 0, 1, 0 },
>>        {"skip_clear", "skip clearing programs", offsetof(MpegTSContext,
>>        skip_clear), AV_OPT_TYPE_BOOL,
>>         {.i64 = 0}, 0, 1, 0 },
>>  +    {"max_packet_size", "maximum size of emitted packet",
>>  offsetof(MpegTSContext, max_packet_size), AV_OPT_TYPE_INT,
>>  +     {.i64 = MAX_PES_PAYLOAD}, 1, INT_MAX, AV_OPT_FLAG_DECODING_PARAM },

INT_MAX/2 might be better to avoid possible overflows with padding or pes 
header. Or simply 1 GiB, after all, a corrupt PID might consume RAM up to 
this amount, so preferably it should not be too big... Docs needs update 
accordingly.

Also since MAX_PES_PAYLOAD define is no longer used in the code, I'd just 
remove it and put its value directly here as the default.

Thanks,
Marton


>>        { NULL },
>>    };
>>
>>  @@ -1121,7 +1124,7 @@ static AVBufferRef *buffer_pool_get(MpegTSContext
>>  *ts, int size)
>>    {
>>        int index = av_log2(size + AV_INPUT_BUFFER_PADDING_SIZE);
>>        if (!ts->pools[index]) {
>>  -        int pool_size = FFMIN(MAX_PES_PAYLOAD +
>>  AV_INPUT_BUFFER_PADDING_SIZE, 2 << index);
>>  +        int pool_size = FFMIN(ts->max_packet_size +
>>  AV_INPUT_BUFFER_PADDING_SIZE, 2 << index);
>>            ts->pools[index] = av_buffer_pool_init(pool_size, NULL);
>>            if (!ts->pools[index])
>>                return NULL;
>>  @@ -1368,7 +1371,7 @@ skip:
>>                break;
>>            case MPEGTS_PAYLOAD:
>>                do {
>>  -                int max_packet_size = MAX_PES_PAYLOAD;
>>  +                int max_packet_size = ts->max_packet_size;
>>                    if (pes->PES_packet_length && pes->PES_packet_length +
>>                    PES_START_SIZE > pes->pes_header_size)
>>                        max_packet_size = pes->PES_packet_length +
>>                        PES_START_SIZE - pes->pes_header_size;
>>    @@ -1378,7 +1381,7 @@ skip:
>>                        if (ret < 0)
>>                            return ret;
>>                        pes->PES_packet_length = 0;
>>  -                    max_packet_size = MAX_PES_PAYLOAD;
>>  +                    max_packet_size = ts->max_packet_size;
>>                        ts->stop_parse = 1;
>>                    } else if (pes->data_index == 0 &&
>>                               buf_size > max_packet_size) {
>
> _______________________________________________
> 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".
>
>


More information about the ffmpeg-devel mailing list