> 在 2019年12月16日，08:04，Andreas Rheinhardt <andreas.rheinha...@gmail.com> 写道：
>
> Under certain circumstances hls_write_packet() would add options to an
> AVDictionary. Said dictionary was never explicitly freed, instead it was
> presumed that these options would be consumed when opening a new
> IO-context. This left several possibilities for memleaks:
>
> a) When no new IO-context would be opened at all. This is possible when
> using both the flags temp_file and single_file together with a file
> output.
> b) When an error happens before one actually tries to open the new
> IO-context.
> c) When the new IO-context does not consume all options.
>
> All three have been fixed; furthermore, the AVDictionary has been put
> into a smaller scope (namely the only part of hls_write_packet() where
> it is actually used).
>
> Signed-off-by: Andreas Rheinhardt <andreas.rheinha...@gmail.com>
> ---
> This patch is concerned with fixing memleaks. Yet I noticed a
> discrepancy in the handling of options between the first try to open
> output and the second one (the one for reflushing in case the first one
> fails): The second one only gets the options not consumed in the first
> one. Is this intended?
>
> libavformat/hlsenc.c | 19 ++++++++++++-------
> 1 file changed, 12 insertions(+), 7 deletions(-)
>
> diff --git a/libavformat/hlsenc.c b/libavformat/hlsenc.c
> index 5695af2208..5b3856099c 100644
> --- a/libavformat/hlsenc.c
> +++ b/libavformat/hlsenc.c
> @@ -2241,7 +2241,6 @@ static int hls_write_packet(AVFormatContext *s,
> AVPacket *pkt)
> int use_temp_file = 0;
> uint8_t *buffer = NULL;
> VariantStream *vs = NULL;
> - AVDictionary *options = NULL;
> char *old_filename = NULL;
>
> for (i = 0; i < hls->nb_varstreams; i++) {
> @@ -2343,11 +2342,6 @@ static int hls_write_packet(AVFormatContext *s,
> AVPacket *pkt)
> use_temp_file = proto && !strcmp(proto, "file") && (hls->flags &
> HLS_TEMP_FILE);
> }
>
> - // look to rename the asset name
> - if (use_temp_file) {
> - av_dict_set(&options, "mpegts_flags", "resend_headers", 0);
> - }
> -
> if (hls->flags & HLS_SINGLE_FILE) {
> ret = flush_dynbuf(vs, &range_length);
> av_freep(&vs->temp_buffer);
> @@ -2356,8 +2350,8 @@ static int hls_write_packet(AVFormatContext *s,
> AVPacket *pkt)
> }
> vs->size = range_length;
> } else {
> - set_http_options(s, &options, hls);
> if ((hls->max_seg_size > 0 && (vs->size >= hls->max_seg_size)) ||
> !byterange_mode) {
> + AVDictionary *options = NULL;
> char *filename = NULL;
> if (hls->key_info_file || hls->encrypt) {
> av_dict_set(&options, "encryption_key", hls->key_string,
> 0);
> @@ -2367,12 +2361,21 @@ static int hls_write_packet(AVFormatContext *s,
> AVPacket *pkt)
> filename = av_asprintf("%s", oc->url);
> }
> if (!filename) {
> + av_dict_free(&options);
> return AVERROR(ENOMEM);
> }
> +
> + // look to rename the asset name
> + if (use_temp_file)
> + av_dict_set(&options, "mpegts_flags", "resend_headers",
> 0);
> +
> + set_http_options(s, &options, hls);
> +
> ret = hlsenc_io_open(s, &vs->out, filename, &options);
> if (ret < 0) {
> av_log(s, hls->ignore_io_errors ? AV_LOG_WARNING :
> AV_LOG_ERROR,
> "Failed to open file '%s'\n", filename);
> + av_dict_free(&options);
> return hls->ignore_io_errors ? 0 : ret;
> }
> if (hls->segment_type == SEGMENT_TYPE_FMP4) {
> @@ -2380,6 +2383,7 @@ static int hls_write_packet(AVFormatContext *s,
> AVPacket *pkt)
> }
> ret = flush_dynbuf(vs, &range_length);
> if (ret < 0) {
> + av_dict_free(&options);
> return ret;
> }
> ret = hlsenc_io_close(s, &vs->out, filename);
> @@ -2391,6 +2395,7 @@ static int hls_write_packet(AVFormatContext *s,
> AVPacket *pkt)
> reflush_dynbuf(vs, &range_length);
> ret = hlsenc_io_close(s, &vs->out, filename);
> }
> + av_dict_free(&options);
> av_freep(&vs->temp_buffer);
> av_freep(&filename);
> }
> --
> 2.20.1
>
> _______________________________________________
> ffmpeg-devel mailing list
> ffmpeg-devel@ffmpeg.org
> https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
>
> To unsubscribe, visit link above, or email
> ffmpeg-devel-requ...@ffmpeg.org with subject "unsubscribe".
LGTM
Thanks
Steven
_______________________________________________
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
To unsubscribe, visit link above, or email
ffmpeg-devel-requ...@ffmpeg.org with subject "unsubscribe".
