Merge commit 'd7d6efe42b0d2057e67999b96b9a391f533d2333'

* commit 'd7d6efe42b0d2057e67999b96b9a391f533d2333':
  h264: check sps.log2_max_frame_num for validity
  mov: validate number of DataReferenceBox entries against box size
  mov: compute avg_frame_rate only if duration is known
  flac: change minimum and default of lpc_passes option to 1

Conflicts:
libavcodec/h264_ps.c
libavformat/mov.c

Merged-by: Michael Niedermayer <michaelni@gmx.at>

Merge commit 'c25fc5c2bb6ae8c93541c9427df3e47206d95152'

* commit 'c25fc5c2bb6ae8c93541c9427df3e47206d95152':
  fate: dpcm: Add dependencies
  SBR DSP x86: implement SSE sbr_hf_gen
  AAC SBR: use AVFloatDSPContext's vector_fmul
  fate: image: Add dependencies
  Changelog: add an entry for deprecating the avconv -vol option
  x86: float_dsp: fix compilation of ff_vector_dmul_scalar_avx() on x86-32

Conflicts:
Changelog
libavutil/x86/float_dsp.asm
tests/fate/image.mak

Merged-by: Michael Niedermayer <michaelni@gmx.at>

lavf/pcm: check size, do not produce invalid packets

Inspired by 92b8c9d89e7ae86bffa05393bb230d5a165b0148.

Signed-off-by: Paul B Mahol <onemda@gmail.com>

diracdec: check dimensions against chroma format.

Fixes out of array reads

Found-by: Mateusz "j00ru" Jurczyk and Gynvael Coldwind
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

h264: check sps.log2_max_frame_num for validity

Fixes infinite or long taking loop in frame num gap code in
the fuzzed sample bipbop234.ts_s223302.

CC: libav-stable@libav.org

mov: validate number of DataReferenceBox entries against box size

Avoids a 2G memory allocation and parsing of random data in
mov_read_dref(). The fuzzed sample sample.mp4_s224424 triggers this.

mov: compute avg_frame_rate only if duration is known

Fixes an assert in fuzzed sample sample.mp4_s265930.

flac: change minimum and default of lpc_passes option to 1

Avoid use of uninitialized and uncomputed linear least square models
during ff_lpc_calc_coefs() for FF_LPC_TYPE_CHOLESKY. Fixes running
make fate-flac-16-lpc-cholesk with valgrind --undef-value-errors=yes.

fate: dpcm: Add dependencies

SBR DSP x86: implement SSE sbr_hf_gen

Start and end index are multiple of 2, therefore guaranteeing aligned access.
Also, this allows to generate 4 floats per loop, keeping the alignment all
along.

Timing:
- 32 bits: 326c -> 172c
- 64 bits: 323c -> 156c

Signed-off-by: Diego Biurrun <diego@biurrun.de>

AAC SBR: use AVFloatDSPContext's vector_fmul

Around 5% speedup on the code block using 'vector_fmul_add's.

Signed-off-by: Diego Biurrun <diego@biurrun.de>

fate: image: Add dependencies

diracdec: fix typo in mctmp allocation

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

diracdec: Test mctmp and mcscratch for malloc failure.

Fixes null pointer dereference

Found-by: Mateusz "j00ru" Jurczyk and Gynvael Coldwind
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

diracdec: fix emulated_edge condition, fix out of array reads

Found-by: Mateusz "j00ru" Jurczyk and Gynvael Coldwind
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

takdec: use samplefmt.h from libavutil

Instead of having own code for calculating decoded buffer size
just use already provided functions from libavutil.

Signed-off-by: Paul B Mahol <onemda@gmail.com>

add missing dependency for tak demuxer

takdec: remove get_code() and use get_sbits() directly

As found by Justin, this is probably bug in reference decoder.

Signed-off-by: Paul B Mahol <onemda@gmail.com>

Fix compilation with yasm 0.6.2.

fix tipo

Found-by: ubitux
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

lavfi/tinterlace: drop redundant NULL checks in uninit()

lavfi/tinterlace: add support to option parsing

Simplify code, and provide introspection through the AVOption system.

lavfi/tinterlace: switch to filter_frame API

Also add missing NULL checks.

lavf: improve help text for avoid_negative_ts

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

bitstream: make vlc init of static tables thread safe.

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

lavfi/mp: switch to ff_filter_frame.

dxa: port to bytestream2 API

Protects against overreads in input buffer.

Signed-off-by: Paul B Mahol <onemda@gmail.com>

smacker: use meaningful error codes

Signed-off-by: Paul B Mahol <onemda@gmail.com>

siff: return meaningful error codes

Signed-off-by: Paul B Mahol <onemda@gmail.com>

lavf/apedec: return meaningful error codes

Signed-off-by: Paul B Mahol <onemda@gmail.com>

huffyuv: return meaningful error codes

Signed-off-by: Paul B Mahol <onemda@gmail.com>

huffyuv: check for malloc failures

Signed-off-by: Paul B Mahol <onemda@gmail.com>

Changelog: add an entry for deprecating the avconv -vol option

huffyuv: make use of av_fast_padded_malloc()

Signed-off-by: Paul B Mahol <onemda@gmail.com>

lavfi/deshake: small align prettifying.

lavfi/deshake: switch inverted comments.

lavfi/deshake: remove unecessary check before unref.

lavfi/deshake: switch to filter_frame.

Merge remote-tracking branch 'qatar/master'

* qatar/master:
  h264: slice-mt: check master context for valid current_picture_ptr
  h264: slice-mt: get last_pic_dropable from master context
  alacenc: add support for multi-channel encoding

Conflicts:
Changelog
libavcodec/alac.c

Merged-by: Michael Niedermayer <michaelni@gmx.at>

Merge commit 'b519298a1578e0c895d53d4b4ed8867b1c031a56'

* commit 'b519298a1578e0c895d53d4b4ed8867b1c031a56':
  pixdesc: fix yuva 10bit bit depth
  avconv: deprecate the -vol option
  x86: af_volume: add SSE2/SSSE3/AVX-optimized s32 volume scaling
  x86: af_volume: add SSE2-optimized s16 volume scaling

Conflicts:
ffmpeg.c
tests/ref/lavfi/pixdesc
tests/ref/lavfi/pixfmts_copy
tests/ref/lavfi/pixfmts_null
tests/ref/lavfi/pixfmts_scale
tests/ref/lavfi/pixfmts_vflip

Merged-by: Michael Niedermayer <michaelni@gmx.at>

lavfi: rename af_volume to af_volume_stefano

This matches the naming style of the new af_volume_justin filter.

Note, Yes i would too prefer having one filter instead of 2.

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

Merge commit 'b384e031daeb1ac612620985e3e5377bc587559c'

* commit 'b384e031daeb1ac612620985e3e5377bc587559c':
  lavfi: add volume filter

Conflicts:
Changelog
libavfilter/Makefile
libavfilter/af_volume.c
libavfilter/version.h

Merged-by: Michael Niedermayer <michaelni@gmx.at>

lavd/lavfi: fix float.h include.

float.h is a system header.

lavfi/silencedetect: use eval builtin to parse dB.

Also update FATE test to use the dB form.

Merge commit '9d5c62ba5b586c80af508b5914934b1c439f6652'

* commit '9d5c62ba5b586c80af508b5914934b1c439f6652':
  lavu/opt: do not filter out the initial sign character except for flags
  eval: treat dB as decibels instead of decibytes
  float_dsp: add vector_dmul_scalar() to multiply a vector of doubles

Conflicts:
libavutil/eval.c
tests/ref/fate/eval

Merged-by: Michael Niedermayer <michaelni@gmx.at>

x86: float_dsp: fix compilation of ff_vector_dmul_scalar_avx() on x86-32

Signed-off-by: Janne Grunau <janne-libav@jannau.net>

lavf: add a concat demuxer.

lavfi/field: switch to filter_frame API

lavfi/drawtext: make use of outlink variable in filter_frame()

Fix warning.

mmfdec: fix seeking

Signed-off-by: Paul B Mahol <onemda@gmail.com>

lavfi/mp: remove dead extern vf_info_t declarations.

lavfi/mp: add ff_ prefix to exported symbols

Add examples/resampling_audio to .gitignore.

doc/decoders: fix typo in "@Options"

examples: add resampling_audio.c file

h264: slice-mt: check master context for valid current_picture_ptr

Fixes errors in slice based multithreading introduced in 0b300daad2f5.

h264: slice-mt: get last_pic_dropable from master context

Fixes fate-h264-conformance-cvnlfi2_sony_h and smllwebdl.mkv from
https://github.com/OpenELEC/OpenELEC.tv/issues/1557 .

ff_lock_avcodec: make the lock state be consistent in case of failure.

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

avcodec_open: if obtaining a lock fails, dont attempt to unlock it.

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

Revert "Acquire lock when initializing parsers."

This reverts commit 7feef7dbca5f6a816e04b63c17aad5dacd272f6e.

This commit causes assertion failures due to use of parser_init from multiple
threads, for example indirectly by ffmpeg.c and more directly from the
packet read functions.

I dont know how to fix this quickly, and fixing ffmpeg.c leaves
the possibility of other applications being affected.
Crashing the applications until this is resolved is clearly no good
thus this revert, so we have time to think about the problem.

Crashes can be reproduced by using multiple input files in ffmpeg.

Revert "Add assert that the avcodec lock is held when initializing static VLC tables."

This reverts commit dd154198b1afa9f3f740eb0b2ccd7ab665f8edd4.

frame_thread_encoder: fix locking while locks are held

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

vc1dec: prevent v_edge_pos from becoming negative.

Fixes assertion failure

Found-by: Mateusz "j00ru" Jurczyk and Gynvael Coldwind
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

resample: remove disabled debug code

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

fate/hqdn3d: add -idct simple.

This should fix FATE failure on ARM. IDCT is required since it's a JPEG
source.

h264: slice-mt: check master context for valid current_picture_ptr

Fixes errors in slice based multithreading introduced in 0b300daad2f5.

CC: libav-stable@libav.org

h264: slice-mt: get last_pic_dropable from master context

Fixes fate-h264-conformance-cvnlfi2_sony_h and smllwebdl.mkv from
https://github.com/OpenELEC/OpenELEC.tv/issues/1557 .

CC: libav-stable@libav.org

swr: remove unused ONE define.

alacenc: add support for multi-channel encoding

cafenc: do not include riff.h

ff_codec_get_tag() was moved to internal.h

Signed-off-by: Paul B Mahol <onemda@gmail.com>

lavc/dxa: return meaningful error codes

Signed-off-by: Paul B Mahol <onemda@gmail.com>

lavf/dxa: return meaningful error codes

Signed-off-by: Paul B Mahol <onemda@gmail.com>

dxa: signal EOF

Fixes #1948.

Signed-off-by: Paul B Mahol <onemda@gmail.com>

Add assert that the avcodec lock is held when initializing static VLC tables.

Signed-off-by: Reimar Döffinger <Reimar.Doeffinger@gmx.de>

Acquire lock when initializing parsers.

This is necessary since they might be initializing or
even using static VLC tables.

Signed-off-by: Reimar Döffinger <Reimar.Doeffinger@gmx.de>

avcodec: add ff_lock/unlock_avcodec functions.

Will be used in future patches, together with the
variable that allows checking whether the lock is held.

Signed-off-by: Reimar Döffinger <Reimar.Doeffinger@gmx.de>

wmaenc: add new line to some error messages

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

mpegvideo_parser: fix buffer access beyond end

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

av_assert should use AV_LOG_PANIC.

The description of AV_LOG_PANIC 100% matches what av_assert
does, while AV_LOG_FATAL does not really.

Signed-off-by: Reimar Döffinger <Reimar.Doeffinger@gmx.de>

mpegvideo_enc: use av_log() in debug code

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

motion_est: use av_log() in debug code

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

j2kenc: remove duplicate debug functions

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

j2k: avoid printf() useage in debug code

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

mpegvideo: remove #if/define PARANOID code

This code never did anything as far as i can remember

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

lavc: fix decode_frame() third parameter semantics for rest of video decoders

Signed-off-by: Paul B Mahol <onemda@gmail.com>

pixdesc: fix yuva 10bit bit depth

It was wrongly set as the yuva 16bit one.

Signed-off-by: Paul B Mahol <onemda@gmail.com>
Signed-off-by: Luca Barbato <lu_zero@gentoo.org>

Merge remote-tracking branch 'qatar/master'

* qatar/master:
  fate: workaround for slighly broken 'test' shell builtin
  mimic: initialize padding of swap_buf through av_fast_padded_malloc
  eamad: initialize padding of bitstream_buf through av_fast_padded_malloc()
  raw demuxer: initialize end of partial packets

Conflicts:
tests/fate-run.sh

Merged-by: Michael Niedermayer <michaelni@gmx.at>

lavfi: add volume filter

Based on the volume filter in FFmpeg written by Stefano Sabatini
<stefasab@gmail.com>.

avconv: deprecate the -vol option

Remove the code for volume scaling in avconv.c and instead auto-insert a
volume filter into the beginning of the filter chain.

x86: af_volume: add SSE2/SSSE3/AVX-optimized s32 volume scaling

x86: af_volume: add SSE2-optimized s16 volume scaling

lavu/opt: do not filter out the initial sign character except for flags

This allows parsing of special-case negative numbers like decibels.

eval: treat dB as decibels instead of decibytes

float_dsp: add vector_dmul_scalar() to multiply a vector of doubles

Include x86-optimized versions for SSE2 and AVX.

Merge commit 'df9b9567518f2840d79a4a96b447ebe1aa326408'

* commit 'df9b9567518f2840d79a4a96b447ebe1aa326408':
  lavc: fix decode_frame() third parameter semantics for video decoders

Conflicts:
libavcodec/cscd.c
libavcodec/eamad.c
libavcodec/ffv1dec.c
libavcodec/gifdec.c
libavcodec/h264.c
libavcodec/iff.c
libavcodec/mjpegdec.c
libavcodec/pcx.c
libavcodec/vp56.c

Merged-by: Michael Niedermayer <michaelni@gmx.at>

Merge commit '387bef95d28019c13c6805cfa4079e59948284e5'

* commit '387bef95d28019c13c6805cfa4079e59948284e5':
  lavc: factorise setting buffer type in avcodec_default_get_buffer().

Merged-by: Michael Niedermayer <michaelni@gmx.at>

Merge commit 'e57c4706e969afa1f2384481b955ccd9494cddb5'

* commit 'e57c4706e969afa1f2384481b955ccd9494cddb5':
  lavc: don't reuse audio buffers

This commit causes a 0.5% speedloss for mp3 and 2% for raw pcm, that is
"time ffmpeg" style tested thus includes disk IO, demux and parsing

I would not have merged it if it wasnt required for the "new" buffer API
but sadly it is.
Once the new API is in ill reimplement proper buffer reuse, which will
fix this speedloss. In case we choose not to merge the "new" buffer
API then this commit here should be reverted.

Conflicts:
libavcodec/internal.h
libavcodec/utils.c

Merged-by: Michael Niedermayer <michaelni@gmx.at>

lavc: move ff_init_buffer_info() down to ff_get_buffer()

and remove redundant code from ff_get_buffer()

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

Merge commit 'ff953fecffd3b9a616a046723fb9d4690be032a6'

* commit 'ff953fecffd3b9a616a046723fb9d4690be032a6':
  lavc: set frame properties in ff_get_buffer().

Conflicts:
libavcodec/utils.c

Merged-by: Michael Niedermayer <michaelni@gmx.at>

Merge commit '594d4d5df3c70404168701dd5c90b7e6e5587793'

* commit '594d4d5df3c70404168701dd5c90b7e6e5587793':
  lavc: add a wrapper for AVCodecContext.get_buffer().

Conflicts:
libavcodec/4xm.c
libavcodec/8svx.c
libavcodec/bmv.c
libavcodec/cljr.c
libavcodec/cscd.c
libavcodec/dnxhddec.c
libavcodec/dpcm.c
libavcodec/dpx.c
libavcodec/eacmv.c
libavcodec/eamad.c
libavcodec/frwu.c
libavcodec/g723_1.c
libavcodec/gifdec.c
libavcodec/idcinvideo.c
libavcodec/iff.c
libavcodec/indeo3.c
libavcodec/internal.h
libavcodec/interplayvideo.c
libavcodec/kmvc.c
libavcodec/mpc7.c
libavcodec/mpegaudiodec.c
libavcodec/pcx.c
libavcodec/pngdec.c
libavcodec/pnmdec.c
libavcodec/rl2.c
libavcodec/snow.c
libavcodec/targa.c
libavcodec/tscc.c
libavcodec/txd.c
libavcodec/utils.c
libavcodec/v210dec.c
libavcodec/vb.c
libavcodec/vmdav.c
libavcodec/vp56.c
libavcodec/vqavideo.c
libavcodec/wavpack.c
libavcodec/wnv1.c
libavcodec/xl.c
libavcodec/yop.c

Merged-by: Michael Niedermayer <michaelni@gmx.at>