ffmpeg.git
7 years agoAdd a version bump and APIchanges entry for avcodec_open2 and avformat_find_stream_info.
Anton Khirnov [Sat, 19 Nov 2011 07:51:26 +0000 (08:51 +0100)]
Add a version bump and APIchanges entry for avcodec_open2 and avformat_find_stream_info.

7 years agolavf: fix multiplication overflow in avformat_find_stream_info()
Mans Rullgard [Tue, 15 Nov 2011 22:33:49 +0000 (22:33 +0000)]
lavf: fix multiplication overflow in avformat_find_stream_info()

Converting to double before the multiplication rather than after
avoids an integer overflow in some cases.

Signed-off-by: Mans Rullgard <mans@mansr.com>
(cherry picked from commit 52767d891c665ab1124fe4ce82d99b59673de7d2)

Signed-off-by: Anton Khirnov <anton@khirnov.net>
7 years agolavf: fix invalid reads in avformat_find_stream_info()
Anton Khirnov [Thu, 14 Jul 2011 01:08:53 +0000 (03:08 +0200)]
lavf: fix invalid reads in avformat_find_stream_info()

(cherry picked from commit e358f7ee90fec591348ca05dff94ebaf4c1a098b)

Conflicts:

libavformat/utils.c

Signed-off-by: Anton Khirnov <anton@khirnov.net>
7 years agolavf: add avformat_find_stream_info()
Anton Khirnov [Sun, 22 May 2011 17:24:59 +0000 (19:24 +0200)]
lavf: add avformat_find_stream_info()

It supports passing options to codecs.
(cherry picked from commit a67c061e0f3b55ffcc96f336fc0998e44b86c8e4)

Conflicts:

libavformat/utils.c

Signed-off-by: Anton Khirnov <anton@khirnov.net>
7 years agolavc: fix parentheses placement in avcodec_open2().
Baptiste Coudurier [Tue, 9 Aug 2011 06:41:50 +0000 (23:41 -0700)]
lavc: fix parentheses placement in avcodec_open2().

Signed-off-by: Anton Khirnov <anton@khirnov.net>
(cherry picked from commit 1d36fb13b088f55ece155153fb6ca8ea278fc837)

Signed-off-by: Anton Khirnov <anton@khirnov.net>
7 years agolavc: introduce avcodec_open2() as a replacement for avcodec_open().
Anton Khirnov [Sun, 22 May 2011 12:10:49 +0000 (14:10 +0200)]
lavc: introduce avcodec_open2() as a replacement for avcodec_open().

Adds support for decoder-private options and makes setting other options
simpler.
(cherry picked from commit 0b950fe240936fa48fd41204bcfd04f35bbf39c3)

Conflicts:

libavcodec/avcodec.h

Signed-off-by: Anton Khirnov <anton@khirnov.net>
8 years agoUpdate Changelog for 0.7.2 release
Reinhard Tartler [Fri, 30 Sep 2011 16:14:12 +0000 (18:14 +0200)]
Update Changelog for 0.7.2 release

8 years agoUpdate RELEASE file for 0.7.2
Reinhard Tartler [Fri, 30 Sep 2011 13:44:19 +0000 (15:44 +0200)]
Update RELEASE file for 0.7.2

8 years agolavf: do not set codec_tag for rawvideo
Mans Rullgard [Sun, 17 Jul 2011 15:02:33 +0000 (16:02 +0100)]
lavf: do not set codec_tag for rawvideo

If the demuxer did not set a codec_tag, there is none and
inventing one makes no sense.  This change stops the rawvideo
"decoder" over-writing user-supplied pixfmt with one derived
from the codec_tag.  The pixfmt-codec_tag-pixfmt round-trip
is lossy since several pixfmts map to the same codec_tag.

This fixes fate-lavf-pixfmt with avfilter disabled.

Signed-off-by: Mans Rullgard <mans@mansr.com>
(cherry picked from commit bb416bd68ca46b4a3f1901533064d56a6b8ac95e)

Signed-off-by: Reinhard Tartler <siretart@tauware.de>
8 years agofate: allow testing with libavfilter disabled
Reinhard Tartler [Tue, 27 Sep 2011 22:50:08 +0000 (00:50 +0200)]
fate: allow testing with libavfilter disabled

This declares dependencies to skip tests using libavfilter
when it is disabled.

Signed-off-by: Mans Rullgard <mans@mansr.com>
(cherry picked from commit 908f12f342341785bf0458e88a06d97a1af90339)

Conflicts:
configure
tests/Makefile
tests/fate.mak

Signed-off-by: Reinhard Tartler <siretart@tauware.de>
8 years agofate: separate lavf-mxf_d10 test from lavf-mxf
Mans Rullgard [Sun, 17 Jul 2011 14:40:05 +0000 (15:40 +0100)]
fate: separate lavf-mxf_d10 test from lavf-mxf

Signed-off-by: Mans Rullgard <mans@mansr.com>
(cherry picked from commit 0218808d4980b794c78f57931d671508aed500b5)

required to unbreak fate with --disable-avfilter
Signed-off-by: Reinhard Tartler <siretart@tauware.de>
8 years agoFix memory (re)allocation in matroskadec.c, related to MSVR-11-0080.
Michael Niedermayer [Thu, 28 Jul 2011 12:59:54 +0000 (14:59 +0200)]
Fix memory (re)allocation in matroskadec.c, related to MSVR-11-0080.

Whitespace of the patch cleaned up by Aurel
Some of the issues have been reported by Steve Manzuik / Microsoft Vulnerability Research (MSVR)
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
(cherry picked from commit 956c901c68eff78288f40e3c8f41ee2fa081d4a8)

Further suggestions from Kostya <kostya.shishkov@gmail.com> have been
implemented by Reinhard Tartler <siretart@tauware.de>

(cherry picked from commit 77d2ef13a8fa630e5081f14bde3fd20f84c90aec)

Signed-off-by: Reinhard Tartler <siretart@tauware.de>
8 years agomovenc: fix NULL reference in mov_write_tkhd_tag
Anton Khirnov [Mon, 19 Sep 2011 13:15:56 +0000 (15:15 +0200)]
movenc: fix NULL reference in mov_write_tkhd_tag

st may be NULL when there are more mov streams than AVStreams, e.g. when
chapters are present.

(cherry picked from commit c92a2a4eb8b883e82871c2415f757153d263b6b3)

Signed-off-by: Reinhard Tartler <siretart@tauware.de>
8 years agomovenc: create an alternate group for each media type
Anton Khirnov [Fri, 16 Sep 2011 14:06:45 +0000 (16:06 +0200)]
movenc: create an alternate group for each media type

Partially fixes bug 44.

(cherry picked from commit 7574cacbd5343bc303ee8333956274716e2444d5)

Signed-off-by: Reinhard Tartler <siretart@tauware.de>
8 years agoflvdec: Check for overflow before allocating arrays
Michael Niedermayer [Sat, 24 Sep 2011 15:57:31 +0000 (18:57 +0300)]
flvdec: Check for overflow before allocating arrays

On allocation, the array length is multiplied by sizeof(int64_t),
this prevents the multiplication from overflowing.

Signed-off-by: Martin Storsjö <martin@martin.st>
(cherry picked from commit a246cefa75aed2ade315d6d09068aacb6b0fe76b)

Signed-off-by: Reinhard Tartler <siretart@tauware.de>
8 years agoppc: fix some pointer to integer casts
Mans Rullgard [Sun, 25 Sep 2011 17:27:47 +0000 (18:27 +0100)]
ppc: fix some pointer to integer casts

Use uintptr_t instead of plain int.  Without this change, the
comparisons will come out wrong for pointers in certain ranges.
Fixes random failures on ppc64.  Also fixes some compiler warnings.

Signed-off-by: Mans Rullgard <mans@mansr.com>
(cherry picked from commit d853e571ad5e7e12c6a68cfde390daced7d85fbb)

Signed-off-by: Reinhard Tartler <siretart@tauware.de>
8 years agoppc: fix 32-bit PIC build
Mans Rullgard [Sun, 25 Sep 2011 11:53:44 +0000 (12:53 +0100)]
ppc: fix 32-bit PIC build

On 32-bit ppc, the GOT pointer must be loaded manually.
This adds a "get_got" assembler macro to compute the
GOT address.  The "movrel" macro is updated to take an
additional parameter containing the GOT address since
no register is reserved for this purpose on ppc32.
These changes have no effect on ppc64 builds.

Signed-off-by: Mans Rullgard <mans@mansr.com>
(cherry picked from commit 6e4a35ced96cdf31a9d3bd82fd147554750af839)

Signed-off-by: Reinhard Tartler <siretart@tauware.de>
8 years agorv34: Check for invalid slice offsets
Laurent Aimar [Mon, 19 Sep 2011 20:48:53 +0000 (22:48 +0200)]
rv34: Check for invalid slice offsets

Signed-off-by: Martin Storsjö <martin@martin.st>
(cherry picked from commit 4cc7732386eb36661ed22d1200339b38a5fa60bc)

Signed-off-by: Anton Khirnov <anton@khirnov.net>
8 years agorv34: Fix potential overreads
Laurent Aimar [Sat, 17 Sep 2011 14:56:30 +0000 (16:56 +0200)]
rv34: Fix potential overreads

Signed-off-by: Martin Storsjö <martin@martin.st>
(cherry picked from commit b4ed3d78cb6c41c9d3ee5918c326ab925edd6a89)

Signed-off-by: Anton Khirnov <anton@khirnov.net>
8 years agorv34: Avoid NULL dereference on corrupted bitstream
Laurent Aimar [Sat, 17 Sep 2011 21:43:58 +0000 (23:43 +0200)]
rv34: Avoid NULL dereference on corrupted bitstream

rv34_decode_slice() can return without allocating any pictures.

Signed-off-by: Martin Storsjö <martin@martin.st>
(cherry picked from commit d0f6ab0298f2309c6104626787ed73416298b019)

Signed-off-by: Anton Khirnov <anton@khirnov.net>
8 years agorv10: Reject slices that does not have the same type as the first one
Laurent Aimar [Sat, 17 Sep 2011 22:03:08 +0000 (00:03 +0200)]
rv10: Reject slices that does not have the same type as the first one

This prevents crashes with some corrupted bitstreams.

Signed-off-by: Martin Storsjö <martin@martin.st>
(cherry picked from commit 4a29b471869353c3077fb4b25b6518eb1047afb7)

Signed-off-by: Anton Khirnov <anton@khirnov.net>
8 years agolavf: Fix context pointer in av_open_input_stream when avformat_open_input fails
David Goldwich [Sat, 17 Sep 2011 11:50:35 +0000 (13:50 +0200)]
lavf: Fix context pointer in av_open_input_stream when avformat_open_input fails

Signed-off-by: David Goldwich <david.goldwich@gmail.com>
Signed-off-by: Anton Khirnov <anton@khirnov.net>
(cherry picked from commit 63d64228a7f31d534e3bcae87cbd37f4a0ae2dd6)

Signed-off-by: Anton Khirnov <anton@khirnov.net>
8 years agooggdec: fix out of bound write in the ogg demuxer
Laurent Aimar [Sun, 11 Sep 2011 21:26:12 +0000 (23:26 +0200)]
oggdec: fix out of bound write in the ogg demuxer

Between ogg_save() and ogg_restore() calls, the number of streams
could have been reduced.

Signed-off-by: Luca Barbato <lu_zero@gentoo.org>
(cherry picked from commit 0e7efb9d23c3641d50caa288818e8c27647ce74d)

Signed-off-by: Anton Khirnov <anton@khirnov.net>
8 years agoFixed size given to init_get_bits().
Laurent Aimar [Fri, 9 Sep 2011 21:46:00 +0000 (23:46 +0200)]
Fixed size given to init_get_bits().

init_get_bits() takes a number of bits and not a number of bytes as
its size argument.

Signed-off-by: Alex Converse <alex.converse@gmail.com>
(cherry picked from commit b59efc94347ccf0cbc2ff14a5a9e99819c5bdc4d)

Signed-off-by: Anton Khirnov <anton@khirnov.net>
8 years agosmacker: fix a few off by 1 errors
Michael Niedermayer [Tue, 13 Sep 2011 21:24:56 +0000 (23:24 +0200)]
smacker: fix a few off by 1 errors

stereo & 16bit is untested due to lack of samples

Signed-off-by: Martin Storsjö <martin@martin.st>
(cherry picked from commit 5166376f24545207607f61ed8ff4e1b0572ff320)

Signed-off-by: Anton Khirnov <anton@khirnov.net>
8 years agoCheck for invalid VLC value in smacker decoder.
Laurent Aimar [Mon, 12 Sep 2011 21:49:36 +0000 (23:49 +0200)]
Check for invalid VLC value in smacker decoder.

Signed-off-by: Martin Storsjö <martin@martin.st>
(cherry picked from commit 6489455495fc5bfbebcfe3f57e5d4fdd6a781091)

Signed-off-by: Anton Khirnov <anton@khirnov.net>
8 years agoCheck and propagate errors when VLC trees cannot be built in smacker decoder.
Laurent Aimar [Mon, 12 Sep 2011 21:46:49 +0000 (23:46 +0200)]
Check and propagate errors when VLC trees cannot be built in smacker decoder.

Signed-off-by: Martin Storsjö <martin@martin.st>
(cherry picked from commit 9676ffba8346791f494451e68d2a3b37a2918a9b)

Signed-off-by: Anton Khirnov <anton@khirnov.net>
8 years agoFixed off by one packet size allocation in the smacker demuxer.
Laurent Aimar [Mon, 12 Sep 2011 18:50:34 +0000 (20:50 +0200)]
Fixed off by one packet size allocation in the smacker demuxer.

Signed-off-by: Martin Storsjö <martin@martin.st>
(cherry picked from commit a92d0fa5d234582583d41b67dddecffc2c819573)

Signed-off-by: Anton Khirnov <anton@khirnov.net>
8 years agoCheck for invalid packet size in the smacker demuxer.
Laurent Aimar [Mon, 12 Sep 2011 18:50:13 +0000 (20:50 +0200)]
Check for invalid packet size in the smacker demuxer.

Signed-off-by: Martin Storsjö <martin@martin.st>
(cherry picked from commit e055932f5636a82275837968eea9c8fcb5bca474)

Signed-off-by: Anton Khirnov <anton@khirnov.net>
8 years agoape demuxer: fix segfault on memory allocation failure.
Laurent Aimar [Sun, 11 Sep 2011 17:17:40 +0000 (19:17 +0200)]
ape demuxer: fix segfault on memory allocation failure.

Signed-off-by: Anton Khirnov <anton@khirnov.net>
(cherry picked from commit 273aab99bf7be2bcda95dd64101c2317ee0fcb99)

Signed-off-by: Anton Khirnov <anton@khirnov.net>
8 years agoxan: Add some buffer checks
Alex Converse [Fri, 9 Sep 2011 23:10:03 +0000 (16:10 -0700)]
xan: Add some buffer checks

(cherry picked from commit 0872bb23b4bd2d94a8ba91070f706d1bc1c3ced8)

Signed-off-by: Anton Khirnov <anton@khirnov.net>
8 years agoFixed size given to init_get_bits() in xan decoder.
Laurent Aimar [Fri, 9 Sep 2011 22:32:12 +0000 (00:32 +0200)]
Fixed size given to init_get_bits() in xan decoder.

(cherry picked from commit 393d5031c6aaaf8c2dda4eb5d676974c349fae85)

Signed-off-by: Anton Khirnov <anton@khirnov.net>
8 years agosmacker demuxer: handle possible av_realloc() failure.
Kostya Shishkov [Mon, 12 Sep 2011 07:40:42 +0000 (09:40 +0200)]
smacker demuxer: handle possible av_realloc() failure.

Signed-off-by: Anton Khirnov <anton@khirnov.net>
(cherry picked from commit 47a8589f7bc69d1a29da1dfdfbd0dfa78a9e31fd)

Signed-off-by: Anton Khirnov <anton@khirnov.net>
8 years agoFixed segfault with wavpack decoder on corrupted decorrelation terms sub-blocks.
Laurent Aimar [Wed, 7 Sep 2011 19:43:03 +0000 (21:43 +0200)]
Fixed segfault with wavpack decoder on corrupted decorrelation terms sub-blocks.

Signed-off-by: Martin Storsjö <martin@martin.st>
(cherry picked from commit 8bfea4ab4e2cb32bc7bf6f697ee30a238c65d296)

Signed-off-by: Anton Khirnov <anton@khirnov.net>
8 years agocljr: init_get_bits size in bits instead of bytes
Alex Converse [Fri, 9 Sep 2011 21:50:33 +0000 (14:50 -0700)]
cljr: init_get_bits size in bits instead of bytes

(cherry picked from commit 0c1f5b93d9b97c4cc3684ba91a040e90bfc760d2)

Signed-off-by: Anton Khirnov <anton@khirnov.net>
8 years agoindeo2: fail if input buffer too small
Alex Converse [Fri, 9 Sep 2011 20:26:49 +0000 (13:26 -0700)]
indeo2: fail if input buffer too small

(cherry picked from commit b7ce4f1d1c3add86ece7ca595ea6c4a10b471055)

Signed-off-by: Anton Khirnov <anton@khirnov.net>
8 years agoindeo2: init_get_bits size in bits instead of bytes
Alex Converse [Fri, 9 Sep 2011 20:24:19 +0000 (13:24 -0700)]
indeo2: init_get_bits size in bits instead of bytes

(cherry picked from commit 68ca330cbd479111db9cb7649d7530ad59f04cc8)

Signed-off-by: Anton Khirnov <anton@khirnov.net>
8 years agoffv1: Fixed size given to init_get_bits() in decoder.
Laurent Aimar [Fri, 9 Sep 2011 20:04:09 +0000 (22:04 +0200)]
ffv1: Fixed size given to init_get_bits() in decoder.

init_get_bits() takes a number of bits and not a number of bytes as
its size argument.

Signed-off-by: Alex Converse <alex.converse@gmail.com>
(cherry picked from commit 46b004959bb7870a361a57272cd5fa7eea34250b)

Signed-off-by: Anton Khirnov <anton@khirnov.net>
8 years agowavpack: Check error codes rather than working around error conditions.
Alex Converse [Thu, 8 Sep 2011 18:02:43 +0000 (11:02 -0700)]
wavpack: Check error codes rather than working around error conditions.

(cherry picked from commit dba2b63a98bdcac7bda1a8a2c48950518c075e17)

Signed-off-by: Anton Khirnov <anton@khirnov.net>
8 years agoFixed invalid access in wavpack decoder on corrupted bitstream.
Laurent Aimar [Wed, 7 Sep 2011 20:02:55 +0000 (22:02 +0200)]
Fixed invalid access in wavpack decoder on corrupted bitstream.

Signed-off-by: Martin Storsjö <martin@martin.st>
(cherry picked from commit 55354b7de21e7bb4bbeb1c12ff55ea17f807c70c)

Signed-off-by: Anton Khirnov <anton@khirnov.net>
8 years agoFixed invalid writes in wavpack decoder on corrupted bitstreams.
Laurent Aimar [Wed, 7 Sep 2011 20:17:39 +0000 (22:17 +0200)]
Fixed invalid writes in wavpack decoder on corrupted bitstreams.

Signed-off-by: Martin Storsjö <martin@martin.st>
(cherry picked from commit 0aedab03405849962b469277afe047aa2c61a87f)

Signed-off-by: Anton Khirnov <anton@khirnov.net>
8 years agoFixed invalid access in wavpack decoder on corrupted extra bits sub-blocks.
Laurent Aimar [Wed, 7 Sep 2011 21:12:32 +0000 (23:12 +0200)]
Fixed invalid access in wavpack decoder on corrupted extra bits sub-blocks.

Signed-off-by: Martin Storsjö <martin@martin.st>
(cherry picked from commit beefafda639dd53fc59c21d8a7cf8334da9a1062)

Signed-off-by: Anton Khirnov <anton@khirnov.net>
8 years agolavc: fix type for thread_type option
Anton Khirnov [Sun, 4 Sep 2011 07:56:47 +0000 (09:56 +0200)]
lavc: fix type for thread_type option

It should be flags, not int.
(cherry picked from commit fb47997edb9d8ff16fc380d005a08c0545624aa6)

Signed-off-by: Anton Khirnov <anton@khirnov.net>
8 years agoAVOptions: fix av_set_string3() doxy to match reality.
Anton Khirnov [Tue, 23 Aug 2011 05:46:51 +0000 (07:46 +0200)]
AVOptions: fix av_set_string3() doxy to match reality.

Fixes bug 28.
(cherry picked from commit e955a682e125d44143415ff2b96a99a4dac78da2)

Signed-off-by: Anton Khirnov <anton@khirnov.net>
8 years agocpu detection: avoid a signed overflow
Sean McGovern [Mon, 25 Jul 2011 22:51:02 +0000 (18:51 -0400)]
cpu detection: avoid a signed overflow

1<<31 overflows because 1 is signed, so force it to unsigned.

Signed-off-by: Ronald S. Bultje <rsbultje@gmail.com>
(cherry picked from commit 5938e02185430ca711106aaec9b5622dbf588af3)

Signed-off-by: Anton Khirnov <anton@khirnov.net>
8 years agovf_scale: don't leak SWS context.
Michael Niedermayer [Wed, 9 Mar 2011 02:30:24 +0000 (03:30 +0100)]
vf_scale: don't leak SWS context.

Signed-off-by: Anton Khirnov <anton@khirnov.net>
(cherry picked from commit 52982dbe474663709033e1ad259f8ff7a5a2eefa)

Signed-off-by: Anton Khirnov <anton@khirnov.net>
8 years agoVC1: Fix first/last row checks with slices
Alberto Delmás [Thu, 25 Aug 2011 09:00:37 +0000 (11:00 +0200)]
VC1: Fix first/last row checks with slices

In some places 0/mb_height were used in place of start_mb_y/end_mb_y.

Fixes SA00049, SA00058, SA10091, SA10097, SA10131, SA20021, SA30030

Improves PSNR in SA00054, SA00059, SA00060, SA10096, SA10098, SA20022,
SA30031, SA30032, SA40012, SA40013

Signed-off-by: Ronald S. Bultje <rsbultje@gmail.com>
(cherry picked from commit 1cf82cab0840d669198ea76ab0363aa661950647)

Signed-off-by: Anton Khirnov <anton@khirnov.net>
8 years agovc1: properly zero coded_block[] edges on new slice entry.
Ronald S. Bultje [Wed, 24 Aug 2011 21:36:16 +0000 (14:36 -0700)]
vc1: properly zero coded_block[] edges on new slice entry.

Previously, we would leave the left edge uninitialized, which led to
CBP prediction errors on slice edges, e.g. in SA10098.vc1.
(cherry picked from commit d4b9974465baf893e90527a366e7a7411ded1ef8)

Signed-off-by: Anton Khirnov <anton@khirnov.net>
8 years agoh264: fix PCM intra-coded blocks in monochrome case
Jeff Downs [Tue, 5 Jul 2011 17:20:06 +0000 (13:20 -0400)]
h264: fix PCM intra-coded blocks in monochrome case

Signed-off-by: Diego Biurrun <diego@biurrun.de>
(cherry picked from commit 6581e161c5f46733a5619208483de29416eb9a51)

Signed-off-by: Anton Khirnov <anton@khirnov.net>
8 years agoh264: correct implicit weight table computation for long ref pics
Jeff Downs [Wed, 6 Jul 2011 15:54:36 +0000 (11:54 -0400)]
h264: correct implicit weight table computation for long ref pics

Correct computation of implicit weight tables when referencing pictures
that are marked for long reference.

Signed-off-by: Diego Biurrun <diego@biurrun.de>
(cherry picked from commit 87cf70eb237e7586cc7399627dafa1b980ec0b7d)

Signed-off-by: Anton Khirnov <anton@khirnov.net>
8 years agoh264: correct the check for invalid long term frame index in MMCO decode
Jeff Downs [Tue, 5 Jul 2011 18:21:54 +0000 (14:21 -0400)]
h264: correct the check for invalid long term frame index in MMCO decode

The current check on MMCO parameters prohibits a "max long term frame index
plus 1" of 16 (frame idx of 15) for the "set max long term frame index" MMCO.
Fix this off-by-one error to allow the full range of legal values.

Signed-off-by: Diego Biurrun <diego@biurrun.de>
(cherry picked from commit 29a09eae9a827f4dbc9c4517180d8fe2ecef321a)

Signed-off-by: Anton Khirnov <anton@khirnov.net>
8 years agoaac: Only output configure if audio was found.
Alex Converse [Tue, 16 Aug 2011 18:03:26 +0000 (11:03 -0700)]
aac: Only output configure if audio was found.

Audio found is not triggered on a CCE because a CCE alone has no output.

Signed-off-by: Luca Barbato <lu_zero@gentoo.org>
(cherry picked from commit d8425ed4af6d8fce62ff363cc590f85e57bac06b)

Signed-off-by: Anton Khirnov <anton@khirnov.net>
8 years agorv10/20: tell decoder to use edge emulation
Kostya Shishkov [Wed, 17 Aug 2011 08:36:33 +0000 (10:36 +0200)]
rv10/20: tell decoder to use edge emulation

This removes out-of-edge motion compensation artifacts (easily spotted green
blocks in avplay, gray blocks in transcoding), for example here:
http://samples.libav.org/samples/real/tv_watching_t1.rm

Signed-off-by: Diego Biurrun <diego@biurrun.de>
(cherry picked from commit 331971116d7d36743601bd2dc5384c5211d3bb48)

Signed-off-by: Anton Khirnov <anton@khirnov.net>
8 years agoflvenc: use int64_t to store offsets
Luca Barbato [Wed, 8 Jun 2011 14:32:07 +0000 (14:32 +0000)]
flvenc: use int64_t to store offsets

Metadata currently is written only at the start of the file in normal
cases, when transcoding from a rtmp source metadata could be
written later and the offset recorded can exceed 32bit.

Signed-off-by: Anton Khirnov <anton@khirnov.net>
(cherry picked from commit 7f5bf4fbaf1f2142547321a16358f9871fabdcc6)

Signed-off-by: Anton Khirnov <anton@khirnov.net>
8 years agoVC-1: fix reading of custom PAR.
Reimar Döffinger [Sat, 13 Aug 2011 09:58:18 +0000 (11:58 +0200)]
VC-1: fix reading of custom PAR.

Custom PAR num/denum are in 1-256 range.

Signed-off-by: Reimar Döffinger <Reimar.Doeffinger@gmx.de>
Signed-off-by: Diego Biurrun <diego@biurrun.de>
(cherry picked from commit 0e8696551414d4ea0aab2559f9475d1fe49d08f3)

Signed-off-by: Anton Khirnov <anton@khirnov.net>
8 years agoh264: notice memory allocation failure
Dustin Brody [Thu, 11 Aug 2011 12:57:58 +0000 (08:57 -0400)]
h264: notice memory allocation failure

Signed-off-by: Ronald S. Bultje <rsbultje@gmail.com>
(cherry picked from commit bac3ab13ea6a9dd8853e79ef3eacf51d234c8774)

Signed-off-by: Anton Khirnov <anton@khirnov.net>
8 years agoRemove incorrect info in documentation of AVCodecContext.bits_per_raw_sample.
Justin Ruggles [Wed, 10 Aug 2011 18:07:35 +0000 (14:07 -0400)]
Remove incorrect info in documentation of AVCodecContext.bits_per_raw_sample.

bits_per_raw_sample is used in video as well, where sample_fmt is not used.
(cherry picked from commit d271d5b2152cafe540f3ab71d3be6ce8636d2fd6)

Signed-off-by: Anton Khirnov <anton@khirnov.net>
8 years agolibx264: do not set pic quality if no frame is output
Baptiste Coudurier [Sun, 30 Jan 2011 01:05:42 +0000 (17:05 -0800)]
libx264: do not set pic quality if no frame is output

Avoids uninitialized reads.

Signed-off-by: Anton Khirnov <anton@khirnov.net>
(cherry picked from commit 5caa2de19ece830e32c95731bc92a423d55cff0c)

Signed-off-by: Anton Khirnov <anton@khirnov.net>
8 years agoaac: Remove some suspicious illegal memcpy()s from LTP.
Alex Converse [Fri, 29 Jul 2011 22:49:11 +0000 (15:49 -0700)]
aac: Remove some suspicious illegal memcpy()s from LTP.

(cherry picked from commit a6c49f18abacb9bf52d4d808a2a56561a5b5445c)

Signed-off-by: Anton Khirnov <anton@khirnov.net>
8 years agomxfdec: Include FF_INPUT_BUFFER_PADDING_SIZE when allocating extradata.
Alex Converse [Fri, 29 Jul 2011 22:27:36 +0000 (15:27 -0700)]
mxfdec: Include FF_INPUT_BUFFER_PADDING_SIZE when allocating extradata.

This prevents out of bounds reads when extradata is being decoded.
(cherry picked from commit 1f6f58d5855288492fc2640a9f1035c01c75d356)

Signed-off-by: Anton Khirnov <anton@khirnov.net>
8 years agovp3/theora: flush after seek.
Ronald S. Bultje [Tue, 26 Jul 2011 17:58:29 +0000 (10:58 -0700)]
vp3/theora: flush after seek.

(cherry picked from commit 8dcf5184307f072d55fb29373be05ef8b0fd02df)

Signed-off-by: Anton Khirnov <anton@khirnov.net>
8 years agorv30: return AVERROR(EINVAL) instead of EINVAL
Diego Biurrun [Thu, 21 Jul 2011 12:25:01 +0000 (14:25 +0200)]
rv30: return AVERROR(EINVAL) instead of EINVAL

On some platforms EINVAL could be positive, ensure we return negative values.
(cherry picked from commit e5985185d2eda942333ebbb72bd7d043ffe40be7)

Signed-off-by: Anton Khirnov <anton@khirnov.net>
8 years agoFix incorrect max_lowres values
Mans Rullgard [Wed, 20 Jul 2011 08:55:48 +0000 (09:55 +0100)]
Fix incorrect max_lowres values

Signed-off-by: Mans Rullgard <mans@mansr.com>
(cherry picked from commit e23a05ab0605693aa715b95120bc0132079ded06)

Signed-off-by: Anton Khirnov <anton@khirnov.net>
8 years agoDo not decode RV30 files if the extradata is too small
Rafaël Carré [Sat, 16 Jul 2011 15:41:08 +0000 (11:41 -0400)]
Do not decode RV30 files if the extradata is too small

Signed-off-by: Diego Biurrun <diego@biurrun.de>
(cherry picked from commit 289c60001fb0a9a1d7a97c876d8a42b84c6874ac)

Signed-off-by: Anton Khirnov <anton@khirnov.net>
8 years agoaacps: skip some memcpy() if src and dst would be equal
Mans Rullgard [Fri, 15 Jul 2011 21:38:10 +0000 (22:38 +0100)]
aacps: skip some memcpy() if src and dst would be equal

Signed-off-by: Mans Rullgard <mans@mansr.com>
(cherry picked from commit e5902d60ce8f7cf10b6e87a57eec536b316261a3)

Signed-off-by: Anton Khirnov <anton@khirnov.net>
8 years agolavf: fix segfault in av_open_input_stream()
Anton Khirnov [Tue, 12 Jul 2011 20:42:18 +0000 (22:42 +0200)]
lavf: fix segfault in av_open_input_stream()

ic is NULL in case of error.
(cherry picked from commit 13551ad1e336573e3732fdeaf25607c47244bb80)

Signed-off-by: Anton Khirnov <anton@khirnov.net>
8 years agopix_fmt: Fix number of bits per component in yuv444p9be
Oskar Arvidsson [Tue, 12 Jul 2011 08:52:19 +0000 (10:52 +0200)]
pix_fmt: Fix number of bits per component in yuv444p9be

Signed-off-by: Ronald S. Bultje <rsbultje@gmail.com>
(cherry picked from commit e59d6b4d7255d6d3dc89580f534e18af1433fe25)

Signed-off-by: Anton Khirnov <anton@khirnov.net>
8 years agompegts: fix Continuity Counter error detection
Jindrich Makovicka [Thu, 30 Jun 2011 09:03:15 +0000 (09:03 +0000)]
mpegts: fix Continuity Counter error detection

According to MPEG-TS specs, the continuity_counter shall not be
incremented when the adaptation_field_control of the packet
equals '00' or '10'.

Signed-off-by: Jindrich Makovicka <jindrich.makovicka@nangu.tv>
Signed-off-by: Anton Khirnov <anton@khirnov.net>
(cherry picked from commit 8923cfa328e8eb565aebcfe8672b276fd1c19bf7)

Signed-off-by: Anton Khirnov <anton@khirnov.net>
8 years agoalsa: limit buffer_size to 32768 frames.
Justin Ruggles [Wed, 22 Jun 2011 20:38:20 +0000 (16:38 -0400)]
alsa: limit buffer_size to 32768 frames.

In testing, the file output plugin gave a max buffer size of about 20 million
frames, which is way more than what is really needed and causes a memory
allocation error on my system.
(cherry picked from commit e35c674d13a7f180412cfe058530a2e7f1d49a90)

Signed-off-by: Anton Khirnov <anton@khirnov.net>
8 years agoalsa: fallback to buffer_size/4 for period_size.
Justin Ruggles [Wed, 22 Jun 2011 19:33:56 +0000 (15:33 -0400)]
alsa: fallback to buffer_size/4 for period_size.

buffer_size/4 is the value used by aplay. This fixes output to null
devices, e.g. writing ALSA output to a file.
(cherry picked from commit 8bfd7f6a475225a0595bf657f8b99a8fffb461e4)

Signed-off-by: Anton Khirnov <anton@khirnov.net>
8 years agogxf: Fix 25 fps DV material in GXF being misdetected as 50 fps
Tomas Härdin [Thu, 23 Jun 2011 13:59:33 +0000 (15:59 +0200)]
gxf: Fix 25 fps DV material in GXF being misdetected as 50 fps

Set DV packet durations using fields_per_frame.
This requires turning gxf_stream_info into the demuxer's context for access to the value in gxf_packet().
Since MPEG-2 seems to work fine this done only for DV.

Signed-off-by: Anton Khirnov <anton@khirnov.net>
(cherry picked from commit 99fecc64b064a013559d3d61f7d9790e3c95c80e)

Signed-off-by: Anton Khirnov <anton@khirnov.net>
8 years agoRevert "ffmpeg: get rid of useless AVInputStream.nb_streams."
Anton Khirnov [Sun, 11 Sep 2011 10:27:51 +0000 (12:27 +0200)]
Revert "ffmpeg: get rid of useless AVInputStream.nb_streams."

This reverts commit 2cf8355f98681bdd726b739008acd5483f82f8d7.
AVInputStream.nb_streams tracks number of streams found at the
beginning, new streams may appear that ffmpeg doesn't know about. Fixes
crash in this case.

Signed-off-by: Anton Khirnov <anton@khirnov.net>
8 years agoadts: Fix PCE copying.
Alex Converse [Wed, 29 Jun 2011 20:41:47 +0000 (13:41 -0700)]
adts: Fix PCE copying.

Parse the extension flag bit when reading the MPEG4 AudioSpecificConfig.

This has nothing to do with SBR/PS contradictory to what was noted when it was removed.
(cherry picked from commit 7f01a4192cdf4565eadee457f76e6b5196e35e0b)

Signed-off-by: Anton Khirnov <anton@khirnov.net>
8 years agoeval: fix memleak.
Ronald S. Bultje [Wed, 6 Jul 2011 01:10:48 +0000 (18:10 -0700)]
eval: fix memleak.

(cherry picked from commit fe277b16f0861a327e1f6c00c0dbb8b00806d60d)

Signed-off-by: Anton Khirnov <anton@khirnov.net>
8 years agoARM: workaround for bug in GNU assembler
Mans Rullgard [Tue, 5 Jul 2011 17:29:35 +0000 (18:29 +0100)]
ARM: workaround for bug in GNU assembler

Some versions of the GNU assembler do not handle 64-bit
immediate operands containing arithmetic.  Writing the
value out in full works correctly.

Signed-off-by: Mans Rullgard <mans@mansr.com>
(cherry picked from commit fce1e43410bdc032c4cf2b1c66166a9ed99cc8f1)

Signed-off-by: Anton Khirnov <anton@khirnov.net>
8 years agomxfenc: fix ignored drop flag in binary timecode representation.
Clément Bœsch [Mon, 4 Jul 2011 08:19:46 +0000 (10:19 +0200)]
mxfenc: fix ignored drop flag in binary timecode representation.

Signed-off-by: Ronald S. Bultje <rsbultje@gmail.com>
(cherry picked from commit 4d5e7ab5c48451404038706ef3113c9925a83087)

Signed-off-by: Anton Khirnov <anton@khirnov.net>
8 years agodca: set AVCodecContext frame_size for DTS audio
John Stebbins [Mon, 4 Jul 2011 16:55:19 +0000 (09:55 -0700)]
dca: set AVCodecContext frame_size for DTS audio

Set the frame size when decoding DTS audio.

This has the side effect of fixing the computation of timestamps for DTS-HD in compute_pkt_fields.  Since frame_size is
not currently set, the duration of a frame is being guessed based on the streams bitrate.  But for DTS-HD, the bitrate
currently used is the rate of the DTS core which is much different than the whole DTS-HD stream and leads to a wildly
inaccurate frame duration estimate.

Signed-off-by: Ronald S. Bultje <rsbultje@gmail.com>
(cherry picked from commit 49c7006c7e815d4330247624a9e6ba30e288cd02)

Signed-off-by: Anton Khirnov <anton@khirnov.net>
8 years agoH.264: fix overreads of qscale_table
Jason Garrett-Glaser [Mon, 4 Jul 2011 13:05:34 +0000 (06:05 -0700)]
H.264: fix overreads of qscale_table

filter_mb_fast assumed that qscale_table was padded like many of the other tables.
(cherry picked from commit 5029a406334ad0eaf92130e23d596e405a8a5aa0)

Signed-off-by: Anton Khirnov <anton@khirnov.net>
8 years agoswscale: don't use planar output functions to write to NV12/21.
Ronald S. Bultje [Sun, 26 Jun 2011 22:52:00 +0000 (15:52 -0700)]
swscale: don't use planar output functions to write to NV12/21.

This prevents a crash when converting to NV12/21 without the bitexact
flags enabled.
(cherry picked from commit 0d994b2f45c08794899057ee7ca54f48218c0a53)

Signed-off-by: Anton Khirnov <anton@khirnov.net>
8 years agopostprocess.c: filter name needs to be double 0 terminated
Piotr Kaczuba [Mon, 30 May 2011 11:19:35 +0000 (13:19 +0200)]
postprocess.c: filter name needs to be double 0 terminated

Signed-off-by: Reinhard Tartler <siretart@tauware.de>
(cherry picked from commit f4f3300c09bb13eb7922e60888b55e3e0fb325e7)

8 years agoconfigure: add missing CFLAGS to fix building on the HURD
Pino Toscano [Fri, 12 Aug 2011 12:11:21 +0000 (14:11 +0200)]
configure: add missing CFLAGS to fix building on the HURD

Signed-off-by: Reinhard Tartler <siretart@tauware.de>
Signed-off-by: Luca Barbato <lu_zero@gentoo.org>
(cherry picked from commit f60d13663742d1c695680ede83c4d646bc57d380)

8 years agocavs: fix some crashes with invalid bitstreams
Mans Rullgard [Wed, 10 Aug 2011 17:52:11 +0000 (18:52 +0100)]
cavs: fix some crashes with invalid bitstreams

This removes all valgrind-reported invalid writes with one
specific test file.

Fixes http://www.ocert.org/advisories/ocert-2011-002.html

Signed-off-by: Mans Rullgard <mans@mansr.com>
(cherry picked from commit 4a71da0f3ab7f5542decd11c81994f849d5b2c78)

8 years agojpegdec: actually search for and parse RSTn
Michael Niedermayer [Thu, 9 Jun 2011 01:35:50 +0000 (03:35 +0200)]
jpegdec: actually search for and parse RSTn

Fixes decoding of MJPEG files produced by some UVC Logitec web cameras,
such as "Notebook Pro" and "HD C910".

References:
http://trac.videolan.org/vlc/ticket/4215
http://ffmpeg.org/trac/ffmpeg/ticket/267

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
Reviewed-by: Kostya <kostya.shishkov@gmail.com>
(cherry picked from commit 8c0fa61a9713a1306fca7997dd04d72ea1f060ea)

8 years agoriff: Add mpgv MPEG-2 fourcc
Alex Converse [Fri, 22 Jul 2011 17:13:22 +0000 (10:13 -0700)]
riff: Add mpgv MPEG-2 fourcc

Supported by mplayer and seen in the wild.
(cherry picked from commit 505345ed5d180093a44da8d70ac541898c31c22f)

8 years agoUpdate Changelog
Reinhard Tartler [Thu, 21 Jul 2011 07:27:23 +0000 (09:27 +0200)]
Update Changelog

8 years agomatroskadec: fix integer underflow if header length < probe length.
Chris Evans [Wed, 20 Jul 2011 00:51:48 +0000 (17:51 -0700)]
matroskadec: fix integer underflow if header length < probe length.

This fixes a crash with specifically crafted files.

Signed-off-by: Ronald S. Bultje <rsbultje@gmail.com>
(cherry picked from commit 69619a13c3fef940cba545cf0a283ff22771dd71)

8 years agoffmpeg: fix operation with --disable-avfilter
Mans Rullgard [Sun, 17 Jul 2011 14:27:14 +0000 (15:27 +0100)]
ffmpeg: fix operation with --disable-avfilter

The width and height must be copied from the input before
being used.

Signed-off-by: Mans Rullgard <mans@mansr.com>
(cherry picked from commit e9f98c90229999c0e654bd77af55d7020347440a)

8 years agovf_libopencv: replace opencv/cxtypes.h #include by opencv/cxcore.h
Stefano Sabatini [Fri, 15 Jul 2011 13:40:31 +0000 (15:40 +0200)]
vf_libopencv: replace opencv/cxtypes.h #include by opencv/cxcore.h

cxtypes.h works with version 2.1 and older, cxcore.h works with 2.2 and older.

Signed-off-by: Diego Biurrun <diego@biurrun.de>
(cherry picked from commit 9bc8bcddbd4fc394e2268e9849dcbf3bad6de980)

8 years agobuild: Create mlib optimization directories during out-of-tree builds.
Diego Biurrun [Mon, 20 Jun 2011 16:41:11 +0000 (18:41 +0200)]
build: Create mlib optimization directories during out-of-tree builds.

8 years agochangelog: misc typo and wording fixes
Diego Biurrun [Sat, 9 Jul 2011 15:43:53 +0000 (17:43 +0200)]
changelog: misc typo and wording fixes
(cherry picked from commit b047941d7da470ba0dcedb1fd0aa828075265ffc)

8 years agodoc: Remove outdated comments about gcc 2.95 and gcc 3.3 support.
Diego Biurrun [Sat, 9 Jul 2011 18:15:29 +0000 (20:15 +0200)]
doc: Remove outdated comments about gcc 2.95 and gcc 3.3 support.
(cherry picked from commit 5ccbf80963c1cc54aed97b1c81b1657ab91baf6a)

8 years agomatroskadec: matroska_read_seek after after EBML_STOP leads to failure.
John Stebbins [Fri, 1 Jul 2011 15:57:42 +0000 (08:57 -0700)]
matroskadec: matroska_read_seek after after EBML_STOP leads to failure.

EBML_STOP leaves matroska->current_id set. Then matroska_read_seek changes
the stream position without resetting current_id.  The next
matroska_parse_cluster  fails due to calculation of incorrect pos.  So clear
current_id when avio_seek happens in matroska_read_seek.

Signed-off-by: Ronald S. Bultje <rsbultje@gmail.com>
(cherry picked from commit cdc2c1c57616956d975c57b4b69eb73865f513f5)

8 years agoUpdate RELEASE file
Reinhard Tartler [Tue, 12 Jul 2011 16:31:28 +0000 (18:31 +0200)]
Update RELEASE file

8 years agoupdate Changelog
Reinhard Tartler [Mon, 4 Jul 2011 10:11:40 +0000 (12:11 +0200)]
update Changelog

8 years agomt: proper locking around release_buffer calls.
Ronald S. Bultje [Mon, 11 Jul 2011 00:28:28 +0000 (17:28 -0700)]
mt: proper locking around release_buffer calls.

This fixes a crash when seeking in some webm files with many
threads (e.g. 8).
(cherry picked from commit 5eafc8b46644764f8aef1b7b2ecae53ee8034822)

8 years agovp8/mt: flush worker thread, not application thread context, on seek.
Ronald S. Bultje [Mon, 11 Jul 2011 04:23:09 +0000 (21:23 -0700)]
vp8/mt: flush worker thread, not application thread context, on seek.

This prevents a crash when seeking.
(cherry picked from commit d1cf45911935cc4fed9afd3a37d99616d31eb9da)

8 years agodocs: Mention the upstream bugzilla url about the dlltool vs MSVC issue
Martin Storsjö [Fri, 8 Jul 2011 08:00:01 +0000 (11:00 +0300)]
docs: Mention the upstream bugzilla url about the dlltool vs MSVC issue

Signed-off-by: Martin Storsjö <martin@martin.st>
(cherry picked from commit b369f327d5e6b01a4cae9e2726df7c73b2893eba)

8 years agodocs: Use proper markup for a literal command line option
Martin Storsjö [Fri, 8 Jul 2011 10:37:54 +0000 (13:37 +0300)]
docs: Use proper markup for a literal command line option

Signed-off-by: Martin Storsjö <martin@martin.st>
(cherry picked from commit a3a94e1498685480800c22fc3ffa20d42ccfd527)

8 years agodocs: Don't recommend adding --enable-memalign-hack
Reinhard Tartler [Tue, 12 Jul 2011 16:09:55 +0000 (18:09 +0200)]
docs: Don't recommend adding --enable-memalign-hack

It is enabled automatically when required nowadays.

Signed-off-by: Martin Storsjö <martin@martin.st>
(cherry picked from commit 9d36139231bfcf155a1b94f61a420768f1771174)

8 years agodocs: Remove needless configure options
Martin Storsjö [Thu, 7 Jul 2011 06:51:57 +0000 (09:51 +0300)]
docs: Remove needless configure options

Specifying --enable-static --disable-shared isn't necessary, these
are the defaults.

Signed-off-by: Martin Storsjö <martin@martin.st>