| Message ID | 20230729002542.51521-1-lq@chinaffmpeg.org |
|---|---|
| State | New |
| Headers | show |
| Series | [FFmpeg-devel,v3] avformat/flvdec: move read fourcc output before flv_same_video_codec | expand |
| Context | Check | Description |
|---|---|---|
| yinshiyou/make_loongarch64 | success | Make finished |
| yinshiyou/make_fate_loongarch64 | success | Make fate finished |
| andriy/make_x86 | success | Make finished |
| andriy/make_fate_x86 | success | Make fate finished |
On Sat, Jul 29, 2023 at 2:26 AM Steven Liu <lq@chinaffmpeg.org> wrote: > > read fourcc for video codec after VideoTagHeader read. > passed as parameter to flv_same_video_codec. > Add the same parameter to flv_set_video_codec, and then you can remove the entire backwards seek and just read it and adjust the size like all other parameters. - Hendrik
On Sat, 29 Jul 2023, Hendrik Leppkes wrote: > On Sat, Jul 29, 2023 at 2:26 AM Steven Liu <lq@chinaffmpeg.org> wrote: >> >> read fourcc for video codec after VideoTagHeader read. >> passed as parameter to flv_same_video_codec. >> > > Add the same parameter to flv_set_video_codec, and then you can remove > the entire backwards seek and just read it and adjust the size like > all other parameters. As far as I see flv_set_video_codec should not read a binary fourcc in the first place when it is called from amf_parse_object. Regards, Marton
On Sat, Jul 29, 2023 at 10:46 AM Marton Balint <cus@passwd.hu> wrote: > > > > On Sat, 29 Jul 2023, Hendrik Leppkes wrote: > > > On Sat, Jul 29, 2023 at 2:26 AM Steven Liu <lq@chinaffmpeg.org> wrote: > >> > >> read fourcc for video codec after VideoTagHeader read. > >> passed as parameter to flv_same_video_codec. > >> > > > > Add the same parameter to flv_set_video_codec, and then you can remove > > the entire backwards seek and just read it and adjust the size like > > all other parameters. > > As far as I see flv_set_video_codec should not read a binary fourcc in the > first place when it is called from amf_parse_object. Yeah probably not, so moving it to a parameter would fix that as well, and make the entire parsing flow less opaque. - Hendrik
On Sat, Jul 29, 2023 at 11:54 AM Hendrik Leppkes <h.leppkes@gmail.com> wrote: > > On Sat, Jul 29, 2023 at 10:46 AM Marton Balint <cus@passwd.hu> wrote: > > > > > > > > On Sat, 29 Jul 2023, Hendrik Leppkes wrote: > > > > > On Sat, Jul 29, 2023 at 2:26 AM Steven Liu <lq@chinaffmpeg.org> wrote: > > >> > > >> read fourcc for video codec after VideoTagHeader read. > > >> passed as parameter to flv_same_video_codec. > > >> > > > > > > Add the same parameter to flv_set_video_codec, and then you can remove > > > the entire backwards seek and just read it and adjust the size like > > > all other parameters. > > > > As far as I see flv_set_video_codec should not read a binary fourcc in the > > first place when it is called from amf_parse_object. > > Yeah probably not, so moving it to a parameter would fix that as well, > and make the entire parsing flow less opaque. > In fact the actual value will carry the FourCC then, which is not handled at all. Definitely very broken here. - Hendrik
Marton Balint <cus@passwd.hu>于2023年7月29日 周六16:45写道： > > > On Sat, 29 Jul 2023, Hendrik Leppkes wrote: > > > On Sat, Jul 29, 2023 at 2:26 AM Steven Liu <lq@chinaffmpeg.org> wrote: > >> > >> read fourcc for video codec after VideoTagHeader read. > >> passed as parameter to flv_same_video_codec. > >> > > > > Add the same parameter to flv_set_video_codec, and then you can remove > > the entire backwards seek and just read it and adjust the size like > > all other parameters. > > As far as I see flv_set_video_codec should not read a binary fourcc in the > first place when it is called from amf_parse_object. yes i looked in flv_set_video_codec, it should not do the same operation, because it cannot read codeid in exheader mode. > > > Regards, > Marton > _______________________________________________ > ffmpeg-devel mailing list > ffmpeg-devel@ffmpeg.org > https://ffmpeg.org/mailman/listinfo/ffmpeg-devel > > To unsubscribe, visit link above, or email > ffmpeg-devel-request@ffmpeg.org with subject "unsubscribe". >
diff --git a/libavformat/flvdec.c b/libavformat/flvdec.c index 3fe21622f7..e3cceba5d0 100644 --- a/libavformat/flvdec.c +++ b/libavformat/flvdec.c @@ -35,6 +35,7 @@ #include "libavutil/intreadwrite.h" #include "libavutil/mathematics.h" #include "avformat.h" +#include "avio_internal.h" #include "demux.h" #include "internal.h" #include "flv.h" @@ -304,27 +305,22 @@ static void flv_set_audio_codec(AVFormatContext *s, AVStream *astream, } } -static int flv_same_video_codec(AVFormatContext *s, AVCodecParameters *vpar, int flags) +static int flv_same_video_codec(AVFormatContext *s, AVCodecParameters *vpar, uint32_t extent_codec_id, int flags) { int flv_codecid = flags & FLV_VIDEO_CODECID_MASK; - FLVContext *flv = s->priv_data; if (!vpar->codec_id && !vpar->codec_tag) return 1; - if (flv->exheader) { - uint8_t *codec_id_str = (uint8_t *)s->pb->buf_ptr; - uint32_t codec_id = codec_id_str[3] | codec_id_str[2] << 8 | codec_id_str[1] << 16 | codec_id_str[0] << 24; - switch(codec_id) { - case MKBETAG('h', 'v', 'c', '1'): - return vpar->codec_id == AV_CODEC_ID_HEVC; - case MKBETAG('a', 'v', '0', '1'): - return vpar->codec_id == AV_CODEC_ID_AV1; - case MKBETAG('v', 'p', '0', '9'): - return vpar->codec_id == AV_CODEC_ID_VP9; - default: - break; - } + switch(extent_codec_id) { + case MKBETAG('h', 'v', 'c', '1'): + return vpar->codec_id == AV_CODEC_ID_HEVC; + case MKBETAG('a', 'v', '0', '1'): + return vpar->codec_id == AV_CODEC_ID_AV1; + case MKBETAG('v', 'p', '0', '9'): + return vpar->codec_id == AV_CODEC_ID_VP9; + default: + break; } switch (flv_codecid) { @@ -1065,6 +1061,7 @@ static int flv_read_packet(AVFormatContext *s, AVPacket *pkt) AVStream *st = NULL; int last = -1; int orig_size; + uint32_t extent_codec_id = 0; retry: /* pkt size is repeated at end. skip it */ @@ -1117,6 +1114,15 @@ retry: * */ flv->exheader = (flags >> 7) & 1; size--; + if (flv->exheader) { + int ensure_seekback_status = ffio_ensure_seekback(s->pb, 4); + if (ensure_seekback_status < 0) { + av_log(s, AV_LOG_WARNING, "Could not ensure seekback\n"); + return ensure_seekback_status; + } + extent_codec_id = avio_rb32(s->pb); + avio_seek(s->pb, -4, SEEK_CUR); + } if ((flags & FLV_VIDEO_FRAMETYPE_MASK) == FLV_FRAME_VIDEO_INFO_CMD) goto skip; } else if (type == FLV_TAG_TYPE_META) { @@ -1174,7 +1180,7 @@ skip: break; } else if (stream_type == FLV_STREAM_TYPE_VIDEO) { if (st->codecpar->codec_type == AVMEDIA_TYPE_VIDEO && - (s->video_codec_id || flv_same_video_codec(s, st->codecpar, flags))) + (s->video_codec_id || flv_same_video_codec(s, st->codecpar, extent_codec_id, flags))) break; } else if (stream_type == FLV_STREAM_TYPE_SUBTITLE) { if (st->codecpar->codec_type == AVMEDIA_TYPE_SUBTITLE)
read fourcc for video codec after VideoTagHeader read. passed as parameter to flv_same_video_codec. Signed-off-by: Steven Liu <lq@chinaffmpeg.org> --- libavformat/flvdec.c | 38 ++++++++++++++++++++++---------------- 1 file changed, 22 insertions(+), 16 deletions(-)