From: "tanwei (D)" <tanwei123@hisilicon.com> To: FFmpeg development discussions and patches <ffmpeg-devel@ffmpeg.org> Cc: "Wujian\(Chin\)" <wujian2@huawei.com>, Lingzezhi <steven.ling@hisilicon.com> Subject: [FFmpeg-devel] [PATCH] libavformat/rtspdec.c: flush pes buffer while rtsp seek Date: Thu, 29 Dec 2022 08:06:31 +0000 Message-ID: <d89f35d60b5444d7be5da9d1eaa02c19@hisilicon.com> (raw) In-Reply-To: <tencent_6557079C8CE0E45B143573808A34A73E9509@qq.com> [-- Attachment #1: Type: text/plain, Size: 9357 bytes --] >> On Dec 29, 2022, at 11:43, tanwei (D) <tanwei123@hisilicon.com> wrote: >> >>>> >>>> + if (rt->cur_transport_priv && rt->transport == >>>> + RTSP_TRANSPORT_RTP) { >>>> >>>> + ff_rtp_seek_flush(rt->cur_transport_priv); >>>> >>>> + } else if (CONFIG_RTPDEC && rt->ts) { >>>> >>>> + av_freep(&rt->recvbuf); >> >>> Is it necessary to free rt->recvbuf? >> The recvbuf must be released. Otherwise, data before seek may be read in the RTSP+TS scenario. > >Read from recvbuf with recvbuf_len = 0 ? How is it possible? I checked the code again, and you're right. The recvbuf does not need to be free. The new modification is to set cur_transport_priv to NULL. I have fixed the previous review comments. Please review the attached patch again. Thank you. >>>> >>>> + rt->recvbuf_pos = 0; >>>> >>>> + rt->recvbuf_len = 0; >>>> >>>> + ff_mpegts_seek_flush(rt->ts); >>>> >>>> + }> > > -----邮件原件----- > 发件人: ffmpeg-devel [mailto:ffmpeg-devel-bounces@ffmpeg.org] 代表 "zhilizhao(赵志立)" > 发送时间: 2022年12月23日 10:34 > 收件人: FFmpeg development discussions and patches > <ffmpeg-devel@ffmpeg.org> > 抄送: Wujian(Chin) <wujian2@huawei.com>; Lingzezhi > <steven.ling@hisilicon.com> > 主题: Re: [FFmpeg-devel] [PATCH] libavformat/rtspdec.c: flush pes buffer > while rtsp seek > > > >> On Dec 22, 2022, at 11:32, tanwei (D) <tanwei123@hisilicon.com> wrote: >> >> Fixes ticket #9949. >> >> >> Signed-off-by: t00660896 <tanwei123@hisilicon.com> >> >> --- >> >> libavformat/mpegts.c | 20 ++++++++++++++++++++ >> >> libavformat/mpegts.h | 1 + >> >> libavformat/rtpdec.c | 7 +++++++ >> >> libavformat/rtpdec.h | 2 ++ >> >> libavformat/rtpdec_mpegts.c | 11 +++++++++++ >> >> libavformat/rtspdec.c | 11 +++++++++++ >> >> 6 files changed, 52 insertions(+) >> >> > > Please check the patch format (a lot of empty lines). > >> diff --git a/libavformat/mpegts.c b/libavformat/mpegts.c >> >> index d97702fcd7..c82971af87 100644 >> >> --- a/libavformat/mpegts.c >> >> +++ b/libavformat/mpegts.c >> >> @@ -3419,6 +3419,26 @@ int avpriv_mpegts_parse_packet(MpegTSContext >> *ts, AVPacket *pkt, >> >> return len1 - len; >> >> } >> >> +void ff_mpegts_seek_flush(MpegTSContext *ts) >> >> +{ >> >> + int i; >> >> + /* flush pes buffer */ >> >> + for (i = 0; i < NB_PID_MAX; i++) { >> >> + if (ts->pids[i]) { >> >> + if (ts->pids[i]->type == MPEGTS_PES) { >> >> + PESContext *pes = ts->pids[i]->u.pes_filter.opaque; >> >> + av_buffer_unref(&pes->buffer); >> >> + pes->data_index = 0; >> >> + pes->state = MPEGTS_SKIP; /* skip until pes header >> + */ >> >> + } else if (ts->pids[i]->type == MPEGTS_SECTION) { >> >> + ts->pids[i]->u.section_filter.last_ver = -1; >> >> + } >> >> + ts->pids[i]->last_cc = -1; >> >> + ts->pids[i]->last_pcr = -1; >> >> + } >> >> + } >> >> +} >> >> + > > Please don’t just duplicate the source code. > >> >> void avpriv_mpegts_parse_close(MpegTSContext *ts) >> >> { >> >> mpegts_free(ts); >> >> diff --git a/libavformat/mpegts.h b/libavformat/mpegts.h >> >> index a48f14e768..ea6b5106a4 100644 >> >> --- a/libavformat/mpegts.h >> >> +++ b/libavformat/mpegts.h >> >> @@ -170,6 +170,7 @@ MpegTSContext >> *avpriv_mpegts_parse_open(AVFormatContext *s); >> >> int avpriv_mpegts_parse_packet(MpegTSContext *ts, AVPacket *pkt, >> >> const uint8_t *buf, int len); >> >> void avpriv_mpegts_parse_close(MpegTSContext *ts); >> >> +void ff_mpegts_seek_flush(MpegTSContext *ts); >> >> typedef struct SLConfigDescr { >> >> int use_au_start; >> >> diff --git a/libavformat/rtpdec.c b/libavformat/rtpdec.c >> >> index fa7544cc07..d688afd1c1 100644 >> >> --- a/libavformat/rtpdec.c >> >> +++ b/libavformat/rtpdec.c >> >> @@ -954,6 +954,13 @@ int ff_rtp_parse_packet(RTPDemuxContext *s, >> AVPacket *pkt, >> >> return rv ? rv : has_next_packet(s); >> >> } >> >> +void ff_rtp_seek_flush(RTPDemuxContext *s) >> >> +{ >> >> + ff_rtp_reset_packet_queue(s); >> >> + if (s->handler && s->handler->seek_flush) >> >> + s->handler->seek_flush(s->dynamic_protocol_context); >> >> +} >> >> + >> >> void ff_rtp_parse_close(RTPDemuxContext *s) >> >> { >> >> ff_rtp_reset_packet_queue(s); >> >> diff --git a/libavformat/rtpdec.h b/libavformat/rtpdec.h >> >> index 5a02e72dc2..8d6d857e28 100644 >> >> --- a/libavformat/rtpdec.h >> >> +++ b/libavformat/rtpdec.h >> >> @@ -52,6 +52,7 @@ int ff_rtp_parse_packet(RTPDemuxContext *s, >> AVPacket *pkt, >> >> void ff_rtp_parse_close(RTPDemuxContext *s); >> >> int64_t ff_rtp_queued_packet_time(RTPDemuxContext *s); >> >> void ff_rtp_reset_packet_queue(RTPDemuxContext *s); >> >> +void ff_rtp_seek_flush(RTPDemuxContext *s); >> >> /** >> >> * Send a dummy packet on both port pairs to set up the connection >> >> @@ -135,6 +136,7 @@ struct RTPDynamicProtocolHandler { >> >> /** Parse handler for this dynamic packet */ >> >> DynamicPayloadPacketHandlerProc parse_packet; >> >> int (*need_keyframe)(PayloadContext *context); >> >> + void (*seek_flush)(PayloadContext *protocol_data); >> >> }; >> >> typedef struct RTPPacket { >> >> diff --git a/libavformat/rtpdec_mpegts.c >> b/libavformat/rtpdec_mpegts.c >> >> index 405271f744..46c1d36021 100644 >> >> --- a/libavformat/rtpdec_mpegts.c >> >> +++ b/libavformat/rtpdec_mpegts.c >> >> @@ -47,6 +47,16 @@ static av_cold int mpegts_init(AVFormatContext >> *ctx, int st_index, >> >> return 0; >> >> } >> >> +static void mpegts_seek_flush(PayloadContext *data) >> >> +{ >> >> + if (!data) >> >> + return; > > Is it possible for data being NULL? It’s better to depends on a clear lifecycle management rather than NULL pointer check everywhere. > >> >> + memset(data->buf, 0, data->read_buf_size); >> >> + data->read_buf_size = 0; > > What about data->read_buf_index ? > >> >> + if (data->ts) >> >> + ff_mpegts_seek_flush(data->ts); >> >> +} >> >> + >> >> static int mpegts_handle_packet(AVFormatContext *ctx, PayloadContext >> *data, >> >> AVStream *st, AVPacket *pkt, uint32_t >> *timestamp, >> >> const uint8_t *buf, int len, uint16_t >> seq, >> >> @@ -94,6 +104,7 @@ const RTPDynamicProtocolHandler >> ff_mpegts_dynamic_handler = { >> >> .priv_data_size = sizeof(PayloadContext), >> >> .parse_packet = mpegts_handle_packet, >> >> .init = mpegts_init, >> >> + .seek_flush = mpegts_seek_flush, >> >> .close = mpegts_close_context, >> >> .static_payload_id = 33, >> >> }; >> >> diff --git a/libavformat/rtspdec.c b/libavformat/rtspdec.c >> >> index bbabec7db8..22b08a4c56 100644 >> >> --- a/libavformat/rtspdec.c >> >> +++ b/libavformat/rtspdec.c >> >> @@ -36,6 +36,7 @@ >> >> #include "rdt.h" >> >> #include "tls.h" >> >> #include "url.h" >> >> +#include "mpegts.h" >> >> #include "version.h" >> >> static const struct RTSPStatusMessage { >> >> @@ -982,6 +983,16 @@ static int rtsp_read_seek(AVFormatContext *s, >> int stream_index, >> >> rt->state = RTSP_STATE_IDLE; >> >> break; >> >> } >> >> + >> >> + if (rt->cur_transport_priv && rt->transport == >> + RTSP_TRANSPORT_RTP) { >> >> + ff_rtp_seek_flush(rt->cur_transport_priv); >> >> + } else if (CONFIG_RTPDEC && rt->ts) { >> >> + av_freep(&rt->recvbuf); > > Is it necessary to free rt->recvbuf? > >> >> + rt->recvbuf_pos = 0; >> >> + rt->recvbuf_len = 0; >> >> + ff_mpegts_seek_flush(rt->ts); >> >> + } >> >> + >> >> return 0; >> >> } >> >> -- >> >> 2.25.1 >> >> >> _______________________________________________ >> 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". > > _______________________________________________ > 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". > _______________________________________________ > 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". _______________________________________________ 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". [-- Attachment #2: 0001-libavformat-rtspdec.c-flush-pes-buffer-while-rtsp-seek.patch --] [-- Type: application/octet-stream, Size: 6347 bytes --] [-- Attachment #3: Type: text/plain, Size: 251 bytes --] _______________________________________________ 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".
next prev parent reply other threads:[~2022-12-29 8:06 UTC|newest] Thread overview: 8+ messages / expand[flat|nested] mbox.gz Atom feed top 2022-12-29 3:43 tanwei (D) 2022-12-29 5:34 ` "zhilizhao(赵志立)" 2022-12-29 8:06 ` tanwei (D) [this message] -- strict thread matches above, loose matches on Subject: below -- 2022-12-22 3:32 tanwei (D) 2022-12-23 2:34 ` "zhilizhao(赵志立)" 2022-11-15 12:23 [FFmpeg-devel] [PATCH] libavformat/rtspdec.c:flush " tanwei (D) 2022-11-14 7:46 tanwei (D) 2022-11-14 8:17 ` "zhilizhao(赵志立)"
Reply instructions: You may reply publicly to this message via plain-text email using any one of the following methods: * Save the following mbox file, import it into your mail client, and reply-to-all from there: mbox Avoid top-posting and favor interleaved quoting: https://en.wikipedia.org/wiki/Posting_style#Interleaved_style * Reply using the --to, --cc, and --in-reply-to switches of git-send-email(1): git send-email \ --in-reply-to=d89f35d60b5444d7be5da9d1eaa02c19@hisilicon.com \ --to=tanwei123@hisilicon.com \ --cc=ffmpeg-devel@ffmpeg.org \ --cc=steven.ling@hisilicon.com \ --cc=wujian2@huawei.com \ /path/to/YOUR_REPLY https://kernel.org/pub/software/scm/git/docs/git-send-email.html * If your mail client supports setting the In-Reply-To header via mailto: links, try the mailto: link
Git Inbox Mirror of the ffmpeg-devel mailing list - see https://ffmpeg.org/mailman/listinfo/ffmpeg-devel This inbox may be cloned and mirrored by anyone: git clone --mirror https://master.gitmailbox.com/ffmpegdev/0 ffmpegdev/git/0.git # If you have public-inbox 1.1+ installed, you may # initialize and index your mirror using the following commands: public-inbox-init -V2 ffmpegdev ffmpegdev/ https://master.gitmailbox.com/ffmpegdev \ ffmpegdev@gitmailbox.com public-inbox-index ffmpegdev Example config snippet for mirrors. AGPL code for this site: git clone https://public-inbox.org/public-inbox.git