avcodec/mips: MSA (MIPS-SIMD-Arch) optimizations for HEVC uni copy, uni horizontal and uni vertical mc functions

Signed-off-by: Shivraj Patil <shivraj.patil@imgtec.com>
Reviewed-by: Nedeljko Babic <Nedeljko.Babic@imgtec.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

ripemd: move ripemd{256, 320} into separate functions

This allows the removal of a few branches.

Before:
lavu       RIPEMD-160   size: 1048576  runs:   1024  time:    7.052 +- 0.010

After:
lavu       RIPEMD-160   size: 1048576  runs:   1024  time:    6.865 +- 0.015

Signed-off-by: James Almer <jamrial@gmail.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

avcodec/x86/vp9dsp_init: Fix mix of declaration and statement

Reviewed-by: "Ronald S. Bultje" <rsbultje@gmail.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

avcodec/vp9dsp: Replace assert by av_assert0()

Reviewed-by: "Ronald S. Bultje" <rsbultje@gmail.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

Merge remote-tracking branch 'rbultje/vp9-profile23-wip'

* rbultje/vp9-profile23-wip:
  libvpxdec: add 440 pixfmts.
  vp9: add profile 2/3 fate tests.
  vp9: add inter-frame profile 2/3 suport.
  vp9: add keyframe profile 2/3 support.
  vp9: parse profile 2/3 bitdepth in frame header.
  vp9: add profile 2/3 to exported profiles.
  fate: add/update reference files for 440 addition.
  swscale: add yuv440p10/12 pixfmts.
  lavc: add yuv440p10/12 formats to aligned pixfmt list.
  lavu: add yuv440p10/12 pixfmts.

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

libvpxdec: add 440 pixfmts.

vp9: add profile 2/3 fate tests.

avcodec/diracdec: Replace default_bsep[] by multiplication

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

swscale/x86/hscale_fast_bilinear_simd: Fix variable names in comments

Found-by: Nick Lewycky <nlewycky@google.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

Merge remote-tracking branch 'cehoyos/master'

* cehoyos/master:
  lavf/img2dec: Autodetect qdraw images.
  lavc/qdrw: Also support real-world qdraw images.

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

avdevice/dshow_capture: avoid #ifdef, use ff_dlog() for dshowdebug()

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

lavf/img2dec: Autodetect qdraw images.

lavc/qdrw: Also support real-world qdraw images.

swresample/dither_template: Do not define macro functions to nothing

This avoids potential warnings

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

avformat/internal: always check arguments of hex_dump_debug()

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

avformat/asfdec: do not define print_guid() to nothing

This avoid potential warnings

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

avcodec/tableprint_vlc: Do not define functions to nothing

This avoids potential warnings

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

Merge commit 'e0010bb91fc0b067be8aefd2f77e6838537da433'

* commit 'e0010bb91fc0b067be8aefd2f77e6838537da433':
  mux: Do not leave stale side data pointers in ff_interleave_add_packet()

Conflicts:
libavformat/mux.c

See: bfb3ed1a9d5dafdc185080fdd88d02ff7221d30a
Merged-by: Michael Niedermayer <michaelni@gmx.at>

Merge commit '6c8c7c246280cdcd23e4d99006b95245ae837278'

* commit '6c8c7c246280cdcd23e4d99006b95245ae837278':
  dump: Print AVAudioServiceType side data

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

Merge commit '04070dbca0688ab1e24528ce5c135254a9a79c47'

* commit '04070dbca0688ab1e24528ce5c135254a9a79c47':
  libx265: Fix 'braces around scalar initializer' warning

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

Merge commit 'd81fb63d87692765c004c19934b49427df434a07'

* commit 'd81fb63d87692765c004c19934b49427df434a07':
  fate: Add a PICT test

Conflicts:
tests/fate/image.mak

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

vp9: add inter-frame profile 2/3 suport.

vp9: add keyframe profile 2/3 support.

vp9: parse profile 2/3 bitdepth in frame header.

vp9: add profile 2/3 to exported profiles.

fate: add/update reference files for 440 addition.

Merge commit '7b734ee55dbb8476d7ad63c7daf55c534cf82d5d'

* commit '7b734ee55dbb8476d7ad63c7daf55c534cf82d5d':
  lavf: Open PICT images with Quickdraw

Conflicts:
tests/fate/image.mak

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

Merge commit '34efb8a169e4551326e069be47125c6c2cb7ab90'

* commit '34efb8a169e4551326e069be47125c6c2cb7ab90':
  quickdraw: Support direct pixel blocks

Conflicts:
Changelog
libavcodec/version.h

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

Merge commit 'bb2cb0a66d86ec4ad9f43a488e8b6ab894c22d2f'

* commit 'bb2cb0a66d86ec4ad9f43a488e8b6ab894c22d2f':
  quickdraw: Switch to greedy parsing

Conflicts:
libavcodec/qdrw.c

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

Merge commit 'd00f1e0fc10bb945cdf140a6c320eb0c9cd25162'

* commit 'd00f1e0fc10bb945cdf140a6c320eb0c9cd25162':
  quickdraw: Convert to bytestream2

Conflicts:
libavcodec/qdrw.c

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

Merge commit '01fdfa51aca9086e04bd354fe3f103a49352c085'

* commit '01fdfa51aca9086e04bd354fe3f103a49352c085':
  xcbgrab: Accept geometries matching the screen size

See: 7971fa9ce0ead8d2e8a79dbe422e91dcef8222ba
Merged-by: Michael Niedermayer <michaelni@gmx.at>

swscale: add yuv440p10/12 pixfmts.

lavc: add yuv440p10/12 formats to aligned pixfmt list.

lavu: add yuv440p10/12 pixfmts.

diracdec: check that block length is valid

In init_planes p->xblen and p->yblen are set to:
            p->xblen = s->plane[0].xblen >> s->chroma_x_shift;
            p->yblen = s->plane[0].yblen >> s->chroma_y_shift;

These are later used as block_w and block_h arguments of
s->vdsp.emulated_edge_mc. If one of them is 0 it triggers an av_assert2
in emulated_edge_mc:
    av_assert2(start_x < end_x && block_w > 0);
    av_assert2(start_y < end_y && block_h > 0);

Signed-off-by: Andreas Cadhalpun <Andreas.Cadhalpun@googlemail.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

dirac: use and forward error codes

Signed-off-by: Andreas Cadhalpun <Andreas.Cadhalpun@googlemail.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

tools/cws2fws: Use a similar implementation in dbgprintf() as in ff_dlog()

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

avcodec/vp8: Suppress empty body warnings

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

mux: Do not leave stale side data pointers in ff_interleave_add_packet()

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
Signed-off-by: Vittorio Giovara <vittorio.giovara@gmail.com>

dump: Print AVAudioServiceType side data

avutil/internal: Suppress empty body warning from emms_c()

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

libx265: Fix 'braces around scalar initializer' warning

x265 Reordered the x265_picture fields in the commit

https://github.com/videolan/x265/commit/51b1518de2414431f36eac592db15b667c25a7b9#diff-945b5354d8767dfac13334f2d22cf58fR107

Now the first field is an integer and not an array.

fate: Add a PICT test

This stimulates the direct pixel mode in Quickdraw decoder.

lavf: Open PICT images with Quickdraw

Update the pictor test to use the pictor codec, as both formats share
the .pic file extension.

quickdraw: Support direct pixel blocks

Data is stored in separated components so rework decode_rle() to support
stepping and offsets.

quickdraw: Switch to greedy parsing

Quickdraw packs data as a series of codes that the application is supposed
to handle, but it does not define any order in which they might appear.
Since it's unfeasible to support *all* opcodes defined by the spec,
only handle well-known blocks containing video data and ignore any unknown
or unsupported ones.

Move palette loading and rle decoding to separate functions to resue them
in other blocks and drop format initialization in init since it can
support more formats than pal8.

Validate width and height.

quickdraw: Convert to bytestream2

Add appropriate error messages and reduce r, g, b variables scope.
Drop the now redundant line checks in RLE decoding.

lavd/xcbgrab: fix comparison with screen size.

Signed-off-by: Nicolas George <george@nsup.org>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

xcbgrab: Accept geometries matching the screen size

Introduced in e8c4db0d4d077.

OS/2:Makedef.cmd cleanup

Remove PROTMODE as it doesn't make sense for DLLs. Also fixes a warning with the OpenWatcom linker
Export symbols as names rather then ordinals for better compatibility for minor releases.

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

avcodec/diracdec: Fix types and wraparounds in dirac_decode_picture_header()

previously various variables had a too small type to support the required 32bit unsigned
range allowed from the spec

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

diracdec: check if reference could not be allocated

s->ref_pics[i] is later used as ref argument of interpolate_refplane,
where it is dereferenced.

If it is NULL, it causes a segmentation fault.

Signed-off-by: Andreas Cadhalpun <Andreas.Cadhalpun@googlemail.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

diracdec: avoid overflow of bytes*8 in decode_lowdelay

If bytes is large enough, bytes*8 can overflow and become negative.

In that case 'bufsize -= bytes*8' causes bufsize to increase instead of
decrease.

This leads to a segmentation fault.

Signed-off-by: Andreas Cadhalpun <Andreas.Cadhalpun@googlemail.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

Merge commit '3735b5c616770429572f86aabdaec39c6ebb8818'

* commit '3735b5c616770429572f86aabdaec39c6ebb8818':
  Revert "lavfi: always check av_expr_parse_and_eval() return value"

See: 8c9a23833daa88bdf08b7e7fbf3a64ab689ad373
Merged-by: Michael Niedermayer <michaelni@gmx.at>

lavf/riff: Add MultiScope II fourcc MSC2 as MJPEG.

Merge commit '60f1cc4a1ffcbf24acbb543988ceeaec76b70818'

* commit '60f1cc4a1ffcbf24acbb543988ceeaec76b70818':
  alsdec: only adapt order for positive max_order

See: 58d605ee9b3277289278dc40e022311f8e083833
Merged-by: Michael Niedermayer <michaelni@gmx.at>

Merge commit '94bb1ce882a12b6d7a1fa32715a68121b39ee838'

* commit '94bb1ce882a12b6d7a1fa32715a68121b39ee838':
  alsdec: check sample pointer range in revert_channel_correlation

Conflicts:
libavcodec/alsdec.c

See: afc7748d1f6abc4b3b1cc957b0fa6941837db3d0
Merged-by: Michael Niedermayer <michaelni@gmx.at>

Merge commit '110f7f35fb615b97d983b1c6c6a714fddd28bcbe'

* commit '110f7f35fb615b97d983b1c6c6a714fddd28bcbe':
  aacpsy: correct calculation of minath in psy_3gpp_init

See: ca9849eecdf7db91d652c698018a5b096d8b78c7
Merged-by: Michael Niedermayer <michaelni@gmx.at>

Merge commit 'e191aaca44b986816695e3b7ecfae64697fd6631'

* commit 'e191aaca44b986816695e3b7ecfae64697fd6631':
  alsdec: limit avctx->bits_per_raw_sample to 32

See: 4c2b88678b436f59132386d9be2fc143e3ee480d
Merged-by: Michael Niedermayer <michaelni@gmx.at>

Merge commit '8fc8024ea56e814cd257d5fe27b21a865080782f'

* commit '8fc8024ea56e814cd257d5fe27b21a865080782f':
  aasc: return correct buffer size from aasc_decode_frame

See: 0be54ad280cf114c02306b7063147e8379f8ed1e
Merged-by: Michael Niedermayer <michaelni@gmx.at>

avcodec/diracdec: Move buf[] read after size check

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

diracdec: prevent overflow in data_unit_size check

buf_idx + data_unit_size can overflow, causing the '> buf_size' check to
wrongly fail.

This causes a segmentation fault.

Signed-off-by: Andreas Cadhalpun <Andreas.Cadhalpun@googlemail.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

vp9: error out if chroma subsampling changes between frames.

This can happen after an intra-only frame. For reference, see:
https://groups.google.com/a/webmproject.org/forum/#!msg/webm-discuss/zaoTxREZjq8/v5Enht87AhcJ

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

Revert "lavfi: always check av_expr_parse_and_eval() return value"

This reverts commit 63be97ec403023fb664798432acedaf6e6922527.

All those calls were unchecked on purpose, as explained in the comments
in the code.

alsdec: only adapt order for positive max_order

For max_order = 0 the clipping range is invalid. (amin = 2, amax = 1)

CC: libav-stable@libav.org
Signed-off-by: Andreas Cadhalpun <Andreas.Cadhalpun@googlemail.com>
Signed-off-by: Anton Khirnov <anton@khirnov.net>

alsdec: check sample pointer range in revert_channel_correlation

Also change the type of begin, end and smp to ptrdiff_t to make the
comparison well-defined.

CC: libav-stable@libav.org
Signed-off-by: Andreas Cadhalpun <Andreas.Cadhalpun@googlemail.com>
Signed-off-by: Anton Khirnov <anton@khirnov.net>

avformat/matroskaenc: Use avoid_negative_ts_use_pts if no stream writes dts

This reduces the number of cases where timestamps need to be shifted

Fixes Ticket4487

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

avformat/mux: Add avoid_negative_ts_use_pts

This allows using pts instead of dts for negative TS avoidance

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

cmdutils: Add libm.h for round()

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

aacpsy: correct calculation of minath in psy_3gpp_init

The minimum of the ath(x, ATH_ADD) function depends on ATH_ADD.
This patch uses the first order approximation to determine it.

For ATH_ADD = 4 this results in the value at 3407.06812 (-5.24241638)
not the one at 3410 (-5.24237967).

CC: libav-stabl@libav.org
Signed-off-by: Andreas Cadhalpun <Andreas.Cadhalpun@googlemail.com>
Signed-off-by: Anton Khirnov <anton@khirnov.net>

alsdec: limit avctx->bits_per_raw_sample to 32

avctx->bits_per_raw_sample is used in get_sbits_long, which only
supports up to 32 bits.

CC: libav-stable@libav.org
Signed-off-by: Andreas Cadhalpun <Andreas.Cadhalpun@googlemail.com>
Signed-off-by: Anton Khirnov <anton@khirnov.net>

aasc: return correct buffer size from aasc_decode_frame

CC: libav-stable@libav.org
Signed-off-by: Andreas Cadhalpun <Andreas.Cadhalpun@googlemail.com>
Signed-off-by: Anton Khirnov <anton@khirnov.net>

cmdutils: move get_rotation() up in the file

Fixes compilation of fftools without libavdevice

Signed-off-by: James Almer <jamrial@gmail.com>

avfilter/Makefile: add missing lavfutils.c dependency to cover_rect and find_rect filters

Signed-off-by: James Almer <jamrial@gmail.com>

avcodec/mpeg12dec: use the correct dimensions for checking SAR

Fixes Ticket4533

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

avcodec/x86/deinterlace: use INIT_MMX like other asm code does too

avformat/avidec: print a warning for negative sample_size

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

avidec: avoid infinite loop due to negative ast->sample_size

If max in clean_index is set to a negative ast->sample_size, the
following loop never ends:
        while (max < 1024)
            max += max;

Thus set ast->sample_size to 0 if it would otherwise be negative.

Signed-off-by: Andreas Cadhalpun <Andreas.Cadhalpun@googlemail.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

vp9: add profile names

Reviewed-by: Ronald S. Bultje <rsbultje@gmail.com>
Signed-off-by: James Almer <jamrial@gmail.com>

avfilter/vf_cover_rect: clip rectangle if it is partly outside the input

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

avidec: check for valid bit_rate range

If bit_rate is negative, it can trigger an av_assert2 in av_rescale_rnd.

Since av_rescale returns int64_t, but st->codec_bit_rate is int, it can
also overflow into a negative value.

Signed-off-by: Andreas Cadhalpun <Andreas.Cadhalpun@googlemail.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

Merge remote-tracking branch 'cehoyos/master'

* cehoyos/master:
  lavfi/cropdetect: Fix cropdetect for > 8 bit input.
  lavf/apngenc: Fix png remuxing by using default extension apng.
  lavf/mxfenc: Write correct interlaced flag when muxing dnxhd.

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

lavfi/cropdetect: Fix cropdetect for > 8 bit input.

lavf/apngenc: Fix png remuxing by using default extension apng.

lavf/mxfenc: Write correct interlaced flag when muxing dnxhd.

Reported and early patch version tested by forum user gridtank.
Reviewed-by: Tomas Härdin

cmdutils: Ask for a sample in case a odd rotation angle is encountered

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

avformat/matroskadec: Use tracks[k]->stream instead of s->streams[k]

The later is not correct

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

matroskadec: check s->streams[k] before using it

This fixes a segmentation fault.

Signed-off-by: Andreas Cadhalpun <Andreas.Cadhalpun@googlemail.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

avcodec/ffv1dec: Check chroma shift parameters

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

avcodec/ffv1dec: More completely check micro_version

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

avcodec/ffv1dec: Check quant table symbols more completely

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

matroskadec: use uint64_t instead of int for index_scale

index_scale is set to matroska->time_scale of type uint64_t.

When index_scale is int, the assignment can overflow and e.g. result
in index_scale = 0. This causes a floating point exception due to the
division by index_scale.

Signed-off-by: Andreas Cadhalpun <Andreas.Cadhalpun@googlemail.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

ffmpeg_filter: use get_rotation()

This unifies the rotation extraction code between ffplay and ffmpeg

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

ffplay&cmdutils:Factor get_rotation() code out

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

avcodec/ccaption_dec: correcting line breaks in cc

Signed-off-by: Anshul Maheshwari <er.anshul.maheshwari@gmail.com>
Reviewed-by: Clément Bœsch <u@pkh.me>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

ffplay: unify displaymatrix based rotation code

Reviewed-by: Marton Balint <cus@passwd.hu>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

pngdec: don't use AV_PIX_FMT_MONOBLACK for apng

AV_PIX_FMT_MONOBLACK has the AV_PIX_FMT_FLAG_BITSTREAM flag, i.e.
linesize can be smaller than width.

Since x_offset is only check against the width, this can lead to
x_offset * bpp >= image_linesize.

In this case ptr could be set to a position outside the image_buf in
png_handle_row, leading to memory corruption and thus crashes.

Signed-off-by: Andreas Cadhalpun <Andreas.Cadhalpun@googlemail.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

pngdec: return correct error code from decode_frame_common

During the loop ret can get changed. Since it is not set on all failure
paths, decode_frame_common can return 0 even though an error occurred.

Signed-off-by: Andreas Cadhalpun <Andreas.Cadhalpun@googlemail.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

pngdec: check s->last_picture.f->data[0] before using it

This check was removed in commit 08aec6f6, but
s->last_picture.f->data[0] is still used in handle_p_frame_apng
unconditionally.

This fixes a segmentation fault.

Signed-off-by: Andreas Cadhalpun <Andreas.Cadhalpun@googlemail.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

avcodec/wavpack: Check L/R values before use to avoid harmless integer overflow and undefined behavior in fate

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