Andriy Gelman:
> From: Andriy Gelman <[email protected]>
>
> Fixes #6334
>
> Signed-off-by: Andriy Gelman <[email protected]>
> ---
> libavformat/rtspdec.c | 17 ++++++++++-------
> 1 file changed, 10 insertions(+), 7 deletions(-)
>
> diff --git a/libavformat/rtspdec.c b/libavformat/rtspdec.c
> index b519b6f1a2..623f478585 100644
> --- a/libavformat/rtspdec.c
> +++ b/libavformat/rtspdec.c
> @@ -720,7 +720,7 @@ static int rtsp_read_header(AVFormatContext *s)
> if (rt->rtsp_flags & RTSP_FLAG_LISTEN) {
> ret = rtsp_listen(s);
> if (ret)
> - return ret;
> + goto fail;
This will add one ff_network_close() to this codepath. Is it really
certain that there was a corresponding ff_network_init()? (And where is
the ff_network_init() that is cancelled in rtsp_read_close() anyway?)
> } else {
> ret = ff_rtsp_connect(s);
> if (ret)
> @@ -728,22 +728,25 @@ static int rtsp_read_header(AVFormatContext *s)
>
> rt->real_setup_cache = !s->nb_streams ? NULL :
> av_mallocz_array(s->nb_streams, 2 *
> sizeof(*rt->real_setup_cache));
> - if (!rt->real_setup_cache && s->nb_streams)
> - return AVERROR(ENOMEM);
> + if (!rt->real_setup_cache && s->nb_streams) {
> + ret = AVERROR(ENOMEM);
> + goto fail;
> + }
With your patch, the calls to ff_network_init() and ff_network_close()
cancel each other out if the above allocation fails.
> rt->real_setup = rt->real_setup_cache + s->nb_streams;
>
> if (rt->initial_pause) {
> /* do not start immediately */
> } else {
> if ((ret = rtsp_read_play(s)) < 0) {
> - ff_rtsp_close_streams(s);
> - ff_rtsp_close_connections(s);
> - return ret;
> + goto fail;
This will add a TEARDOWN command to this and the above error path. Is
this something good?
> }
> }
> }
> -
> return 0;
> +
> +fail:
> + rtsp_read_close(s);
> + return ret;
> }
>
> int ff_rtsp_tcp_read_packet(AVFormatContext *s, RTSPStream **prtsp_st,
>
_______________________________________________
ffmpeg-devel mailing list
[email protected]
https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
To unsubscribe, visit link above, or email
[email protected] with subject "unsubscribe".