| Message ID | AM7PR03MB6660D256CE0D1451F3801FDC8F5B9@AM7PR03MB6660.eurprd03.prod.outlook.com |
|---|---|
| State | Accepted |
| Commit | c8b11b28d1eb8dfa34c727c5da9b266124196799 |
| Headers | show |
| Series | [FFmpeg-devel] avformat/mov: Fix endian-dependent parsing | expand |
| Context | Check | Description |
|---|---|---|
| andriy/make_x86 | success | Make finished |
| andriy/make_fate_x86 | success | Make fate finished |
| andriy/make_ppc | success | Make finished |
| andriy/make_fate_ppc | success | Make fate finished |
| andriy/make_aarch64_jetson | success | Make finished |
| andriy/make_fate_aarch64_jetson | success | Make fate finished |
On Fri, 21 Jan 2022, Andreas Rheinhardt wrote: > MOVAtom.type is always read as a little-endian number > (despite MOV/ISOBMFF being big-endian). > Fixes the matroska-dovi-write-config8 FATE-test on big-endian > arches (which runs into the "index out of range" warning message). > > Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com> > --- > A few days ago I sent a patch using the dv84.mov sample > that also failed on the PPC BE endian FATE box, yet this test > reencoded the audio and so I blamed this (despite encoder and > decoder being fixed-point); see [1]. Seems like I was wrong > and the fixed-point codecs are indeed bitexact. > > [1]: https://ffmpeg.org/pipermail/ffmpeg-devel/2022-January/291792.html > > libavformat/mov.c | 2 +- > 1 file changed, 1 insertion(+), 1 deletion(-) > > diff --git a/libavformat/mov.c b/libavformat/mov.c > index e401cd39b5..1437d160f8 100644 > --- a/libavformat/mov.c > +++ b/libavformat/mov.c > @@ -407,7 +407,7 @@ retry: > atom.size -= 16; > > if (!key && c->found_hdlr_mdta && c->meta_keys) { > - uint32_t index = AV_RB32(&atom.type); > + uint32_t index = av_bswap32(atom.type); // BE number has been read as LE LGTM. // Martin
diff --git a/libavformat/mov.c b/libavformat/mov.c index e401cd39b5..1437d160f8 100644 --- a/libavformat/mov.c +++ b/libavformat/mov.c @@ -407,7 +407,7 @@ retry: atom.size -= 16; if (!key && c->found_hdlr_mdta && c->meta_keys) { - uint32_t index = AV_RB32(&atom.type); + uint32_t index = av_bswap32(atom.type); // BE number has been read as LE if (index < c->meta_keys_count && index > 0) { key = c->meta_keys[index]; } else if (atom.type != MKTAG('c', 'o', 'v', 'r')) {
MOVAtom.type is always read as a little-endian number (despite MOV/ISOBMFF being big-endian). Fixes the matroska-dovi-write-config8 FATE-test on big-endian arches (which runs into the "index out of range" warning message). Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com> --- A few days ago I sent a patch using the dv84.mov sample that also failed on the PPC BE endian FATE box, yet this test reencoded the audio and so I blamed this (despite encoder and decoder being fixed-point); see [1]. Seems like I was wrong and the fixed-point codecs are indeed bitexact. [1]: https://ffmpeg.org/pipermail/ffmpeg-devel/2022-January/291792.html libavformat/mov.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)