| Message ID | tencent_9D38D0ACE6D667FB2A84D32B5C5A4BAA2606@qq.com |
|---|---|
| State | New |
| Headers | show |
| Series | [FFmpeg-devel,RFC] avformat/hls: check IV size inside EXT-X-KEY | 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 |
| andriy/make_armv7_RPi4 | success | Make finished |
| andriy/make_fate_armv7_RPi4 | success | Make fate finished |
Zhao Zhili <quinkblack@foxmail.com> 于2022年4月12日周二 16:28写道： > > IV should always be 128 bits. If the IV attribute was truncated > inside EXT-X-KEY, padding on the left which is the same as when > using sequence number as IV. > --- > I'm not sure which method is better: do padding or just return > AVERROR_INVALIDDATA? > > libavformat/hls.c | 8 +++++++- > 1 file changed, 7 insertions(+), 1 deletion(-) > > diff --git a/libavformat/hls.c b/libavformat/hls.c > index 83ff4cc607..a7b632c20e 100644 > --- a/libavformat/hls.c > +++ b/libavformat/hls.c > @@ -810,7 +810,13 @@ static int parse_playlist(HLSContext *c, const char *url, > if (!strcmp(info.method, "SAMPLE-AES")) > key_type = KEY_SAMPLE_AES; > if (!av_strncasecmp(info.iv, "0x", 2)) { > - ff_hex_to_data(iv, info.iv + 2); > + int n = ff_hex_to_data(iv, info.iv + 2); > + if (n < sizeof(iv)) { > + av_log(c->ctx, AV_LOG_WARNING, > + "Incomplete IV %s, padding on the left\n", info.iv); > + memmove(iv + sizeof(iv) - n, iv, n); > + memset(iv, 0, sizeof(iv) - n); > + } > has_iv = 1; > } > av_strlcpy(key, info.uri, sizeof(key)); > -- > 2.31.1 > > _______________________________________________ > 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". LGTM Thanks Steven
Quoting Zhao Zhili (2022-04-12 10:27:50) > IV should always be 128 bits. If the IV attribute was truncated > inside EXT-X-KEY, padding on the left which is the same as when > using sequence number as IV. > --- > I'm not sure which method is better: do padding or just return > AVERROR_INVALIDDATA? I would say return error, unless there are known buggy producers of incomplete IVs that we want to support.
> On Apr 13, 2022, at 7:53 PM, Anton Khirnov <anton@khirnov.net> wrote: > > Quoting Zhao Zhili (2022-04-12 10:27:50) >> IV should always be 128 bits. If the IV attribute was truncated >> inside EXT-X-KEY, padding on the left which is the same as when >> using sequence number as IV. >> --- >> I'm not sure which method is better: do padding or just return >> AVERROR_INVALIDDATA? > > I would say return error, unless there are known buggy producers of > incomplete IVs that we want to support. Make sense. Actually it’s a story about shoot myself in the foot. I created a buggy manifest, and the bug is hidden by the default padding on the tail. So the vide only works with ffmpeg/ffplay. Let’s hope I’m the only one make such mistake. Will send a patch to return error in this case. > > -- > Anton Khirnov > _______________________________________________ > 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/hls.c b/libavformat/hls.c index 83ff4cc607..a7b632c20e 100644 --- a/libavformat/hls.c +++ b/libavformat/hls.c @@ -810,7 +810,13 @@ static int parse_playlist(HLSContext *c, const char *url, if (!strcmp(info.method, "SAMPLE-AES")) key_type = KEY_SAMPLE_AES; if (!av_strncasecmp(info.iv, "0x", 2)) { - ff_hex_to_data(iv, info.iv + 2); + int n = ff_hex_to_data(iv, info.iv + 2); + if (n < sizeof(iv)) { + av_log(c->ctx, AV_LOG_WARNING, + "Incomplete IV %s, padding on the left\n", info.iv); + memmove(iv + sizeof(iv) - n, iv, n); + memset(iv, 0, sizeof(iv) - n); + } has_iv = 1; } av_strlcpy(key, info.uri, sizeof(key));