Merge commit 'f62bb216ac4cfbbff16108c6bac35a0282532972'

* commit 'f62bb216ac4cfbbff16108c6bac35a0282532972':
  hwcontext_vaapi: allow transfers to/from any size of sw frame

Merged-by: Hendrik Leppkes <h.leppkes@gmail.com>

Merge commit 'c3f113d58488df7594a489bdbb993a69ad47063c'

* commit 'c3f113d58488df7594a489bdbb993a69ad47063c':
  vf_hwdownload: allocate the destination frame for the pool size

Merged-by: Hendrik Leppkes <h.leppkes@gmail.com>

Merge commit 'fdfe01365d579189d9a55b3741dba2ac46eb1df8'

* commit 'fdfe01365d579189d9a55b3741dba2ac46eb1df8':
  hwcontext: allocate the destination frame for the pool size

Merged-by: Hendrik Leppkes <h.leppkes@gmail.com>

Merge commit '5fcae3b3f93fd02b3d1e009b9d9b17410fca9498'

* commit '5fcae3b3f93fd02b3d1e009b9d9b17410fca9498':
  hwcontext: clarify the behaviour of transfer_data() for cropped frames

Merged-by: Hendrik Leppkes <h.leppkes@gmail.com>

doc/libav-merge: add skipped fixup commits to the list of missing changes

Merge commit '94ebf5565849e4dc036d2ca43979571ed3736457'

* commit '94ebf5565849e4dc036d2ca43979571ed3736457':
  avconv: restructure sending EOF to filters

Noop, as its a fixup to a previously skipped commit

Merged-by: Hendrik Leppkes <h.leppkes@gmail.com>

Merge commit 'd2e56cf753a6c462041dee897d9d0c90f349988c'

* commit 'd2e56cf753a6c462041dee897d9d0c90f349988c':
  avconv: move flushing the queued frames to configure_filtergraph()

Noop, as its a fixup to a previously skipped commit

Merged-by: Hendrik Leppkes <h.leppkes@gmail.com>

Merge commit '7a745f014f528d1001394ae4d2f4ed1a20bf7fa2'

* commit '7a745f014f528d1001394ae4d2f4ed1a20bf7fa2':
  options_table: Add aliases for color properties

Merged-by: Hendrik Leppkes <h.leppkes@gmail.com>

hwcontext_vaapi: add a quirk for the missing MemoryType attribute

The Intel binary iHD driver does not support the
VASurfaceAttribMemoryType, so surface allocation will fail when using
it.

(cherry picked from commit 2124711b950b03c582a119c75f52a87acc32d6ec)

hwcontext_vaapi: Try the first render node as the default DRM device

If no string argument is supplied when av_hwdevice_ctx_create() is
called to create a VAAPI device, we currently only try the default
X11 display (that is, $DISPLAY) to find a device, and will therefore
fail in the absence of an X server to connect to.  Change the logic
to also look for a device via the first DRM render node (that is,
"/dev/dri/renderD128"), which is probably the right thing to use in
most simple configurations which only have one DRM device.

(cherry picked from commit 121f34d5f0c8d7d376829a467590fbbe4c228f4f)

vf_scale_vaapi: Respect driver quirks around buffer destruction

(cherry picked from commit 582d4211e00015b68626f77ce4af53161e2b1713)

vaapi_encode: Respect driver quirks around buffer destruction

No longer leaks memory when used with a driver with the "render does
not destroy param buffers" quirk (i.e. Intel i965).

(cherry picked from commit 221ffca6314ed3ba9d38464ea50cd85251c04e74)
Fixes ticket #5871.

hwcontext_vaapi: Add driver quirks to the hwdevice

The driver being used is detected inside av_hwdevice_ctx_init() and
the quirks field then set from a table of known device.  If this
behaviour is unwanted, the user can also set the quirks field
manually.

Also adds the Intel i965 driver quirk (it does not destroy parameter
buffers used in a call to vaRenderPicture()) and detects that driver
to set it.

(cherry picked from commit 4926fa9a4aa03f3b751f52e900b9efb87fea0591)

filmstripdec: correctly check image dimensions

This prevents a division by zero in read_packet.

Reviewed-by: Paul B Mahol <onemda@gmail.com>
Signed-off-by: Andreas Cadhalpun <Andreas.Cadhalpun@googlemail.com>

ffmpeg_qsv: Fix hwaccel transcoding

Set up the encoder with a hardware context which will match the one
the decoder will use when it starts later.

Includes 02c2761973dfc886e94a60a9c7d6d30c296d5b8c, with additional
hackery to get around a3a0230a9870b9018dc7415ae5872784d524cfe5 being
skipped.

Merge commit '444a36269f853844369af0a9836507e5a2780323'

* commit '444a36269f853844369af0a9836507e5a2780323':
  pixdesc: Fix AVCOL_TRC_BT2020_12 name

Merged-by: Hendrik Leppkes <h.leppkes@gmail.com>

Merge commit 'f172e22d6aed0bff36e975bafb0183b6779f9444'

* commit 'f172e22d6aed0bff36e975bafb0183b6779f9444':
  pixdesc: Add aliases to SMPTE color properties

Merged-by: Hendrik Leppkes <h.leppkes@gmail.com>

Merge commit '8a62d2c28fbacd1ae20c35887a1eecba2be14371'

* commit '8a62d2c28fbacd1ae20c35887a1eecba2be14371':
  vaapi_encode: Maintain a pool of bitstream output buffers

Merged-by: Hendrik Leppkes <h.leppkes@gmail.com>

Merge commit 'bd31c61cf94d01dbe1051cf65874e7b2c0ac5454'

* commit 'bd31c61cf94d01dbe1051cf65874e7b2c0ac5454':
  avconv: Remove hw_device_ctx output filter reinit hack

Noop, since our hwaccel infrastructure still requires this.

Merged-by: Hendrik Leppkes <h.leppkes@gmail.com>

Merge commit '4a081f224e12f4227ae966bcbdd5384f22121ecf'

* commit '4a081f224e12f4227ae966bcbdd5384f22121ecf':
  libavcodec: fix constness in clobber test avcodec_open2() wrappers

Merged-by: Hendrik Leppkes <h.leppkes@gmail.com>

doc: update merge status for recent additions and skipped merges

Merge commit '02c2761973dfc886e94a60a9c7d6d30c296d5b8c'

* commit '02c2761973dfc886e94a60a9c7d6d30c296d5b8c':
  avconv_qsv: use the device creation API

Not merged, our ffmpeg hwaccel infra is not quite the same as avconvs.

Merged-by: Hendrik Leppkes <h.leppkes@gmail.com>

Merge commit '232399e3ee219d16d0e0d482c9f31a26202d4993'

* commit '232399e3ee219d16d0e0d482c9f31a26202d4993':
  avconv: pass the hwaccel frames context to the decoder

Merged-by: Hendrik Leppkes <h.leppkes@gmail.com>

Merge commit 'a3a0230a9870b9018dc7415ae5872784d524cfe5'

* commit 'a3a0230a9870b9018dc7415ae5872784d524cfe5':
  avconv: init filtergraphs only after we have a frame on each input

This commit is a noop since it doesn't apply cleanly due to differences
in the dataflow between avconv and ffmpeg, and thus fixing this in the
scope of a merge is unfeasible.

Merged-by: Hendrik Leppkes <h.leppkes@gmail.com>

Merge commit '3e265ca58f0505470186dce300ab66a6eac3978e'

* commit '3e265ca58f0505470186dce300ab66a6eac3978e':
  avconv: do packet ts rescaling in write_packet()

This commit is a noop since it doesn't apply cleanly due to differences
in the dataflow between avconv and ffmpeg, and thus fixing this in the
scope of a merge is unfeasible.

Merged-by: Hendrik Leppkes <h.leppkes@gmail.com>

Merge commit '50722b4f0cbc5940e9e6e21d113888436cc89ff5'

* commit '50722b4f0cbc5940e9e6e21d113888436cc89ff5':
  avconv: decouple configuring filtergraphs and setting output parameters

Merged-by: Hendrik Leppkes <h.leppkes@gmail.com>

Merge commit 'ba7397baef796ca3991fe1c921bc91054407c48b'

* commit 'ba7397baef796ca3991fe1c921bc91054407c48b':
  avconv: factor out initializing stream parameters for encoding

Merged-by: Hendrik Leppkes <h.leppkes@gmail.com>

ffmpeg_filter: fix hwaccel transcoding

Based on a patch by Yogender Gupta <ygupta@nvidia.com>

vp9_mc_template: limit assert to SCALED == 0

The handling of the other block sizes was limited to 'SCALED == 0' in
commit dc96c0f9fc96bf4167633befc074394062793322, so this assert should
be disabled, too, as it can now be triggered.

Reviewed-by: Ronald S. Bultje <rsbultje@gmail.com>
Signed-off-by: Andreas Cadhalpun <Andreas.Cadhalpun@googlemail.com>

configure: use check_lib2 for cuda and cuvid

Fixes building for Windows x86 with MSVC using the link libraries distributed with the CUDA SDK.

check_lib2 is required here because it includes the header to get the full signature of the
function, including the stdcall calling convention and all of its arguments, which enables
the linker to determine the fully qualified object name and resolve it through the import
library, since the CUDA SDK libraries do not include un-qualified aliases.

avformat: Add Pro-MPEG CoP #3-R2 FEC protocol

Pro-MPEG Code of Practice #3 release 2 forward error correction for rtp_mpegts streams

Signed-off-by: Vlad Tarca <vtarca@mobibase.com>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

lavfi: split frame_count between input and output.

AVFilterLink.frame_count is supposed to count the number of frames
that were passed on the link, but with min_samples, that number is
not always the same for the source and destination filters.
With the addition of a FIFO on the link, the difference will become
more significant.

Split the variable in two: frame_count_in counts the number of
frames that entered the link, frame_count_out counts the number
of frames that were sent to the destination filter.

fate/colorkey: disable audio stream.

The test is not supposed to cover audio.
Also, using -vframes along with an audio stream depends on
the exact order the frames are processed by filters, it is
too much constraint to guarantee.

lavfi/ebur128: specify scaling_factor directly

This should fix build with Solaris CC.

Reviewed-by: Michael Niedermayer <michael@niedermayer.cc>
Signed-off-by: Marton Balint <cus@passwd.hu>

lavf/fifo: fix undefined behaviour in deinit when destroying mutex

Reviewed-by: Jan Sebechlebsky <sebechlebskyjan@gmail.com>
Signed-off-by: Marton Balint <cus@passwd.hu>

ffmpeg: add newline to avformat_write_header error message

Reviewed-by: Michael Niedermayer <michael@niedermayer.cc>
Signed-off-by: Marton Balint <cus@passwd.hu>

avformat/mov: zero initialize codec_name in mov_parse_stsd_video()

Fixes valgrind warning about "Conditional jump or move depends on uninitialised value(s)"

Reviewed-by: Andreas Cadhalpun <andreas.cadhalpun@googlemail.com>
Signed-off-by: James Almer <jamrial@gmail.com>

fate: fix fate-api dependencies

No need to run the whole fate-lavf set. fate-lavf-flv_fmt is enough to
create the required source file.

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

increment initial_prog_date_time when removing old segments

Fixes https://trac.ffmpeg.org/ticket/5940#ticket

Reviewed-by: Michael Niedermayer <<michael@niedermayer.cc>>
Signed-off-by: Robert Nagy <ronag89@gmail.com>
Signed-off-by: Steven Liu <lingjiujianke@gmail.com>

avcodec/htmlsubtitles: Fix reading one byte beyond the array

Fixes: fuzz-2-ffmpeg_SUBTITLE_AV_CODEC_ID_SUBRIP_fuzzer
Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/targets/ffmpeg
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

pnmdec: make sure v is capped by maxval

Otherwise put_bits can be called with a value that doesn't fit in the
sample_len, causing an assertion failure.

Reviewed-by: Michael Niedermayer <michael@niedermayer.cc>
Signed-off-by: Andreas Cadhalpun <Andreas.Cadhalpun@googlemail.com>

pnm: limit maxval to UINT16_MAX

From 'man ppm': The maximum color value (Maxval), again in ASCII decimal.
                Must be less than 65536.

Reviewed-by: Michael Niedermayer <michael@niedermayer.cc>
Signed-off-by: Andreas Cadhalpun <Andreas.Cadhalpun@googlemail.com>

smvjpegdec: make sure cur_frame is not negative

This fixes a heap-buffer-overflow detected by AddressSanitizer.

Reviewed-by: Michael Niedermayer <michael@niedermayer.cc>
Signed-off-by: Andreas Cadhalpun <Andreas.Cadhalpun@googlemail.com>

lavfi/loudnorm: add an internal libebur128 library

Also contains the following changes to the library:
- add ff_ prefix to functions
- remove cplusplus defines.
- add FF_ prefix to contants and some structs
- remove true peak calculation feature, since it uses its own resampler, and
  af_loudnorm does not need it.
- remove version info and some fprintf(stderr) functions
- convert to use av_malloc
- always use histogram mode for LRA calculation, otherwise LRA data is slowly
  consuming memory making af_loudnorm unfit for 24/7 operation. It also uses a
  BSD style linked list implementation which is probably not available on all
  platforms. So let's just remove the classic mode which not uses histogram.
- add ff_thread_once for calculating static histogram tables
- convert some functions to void which cannot fail
- remove intrinsics and some unused headers
- add support for planar audio
- remove channel / sample rate changer function, in ffmpeg usually we simply
  alloc a new context
- convert some static variables to defines
- declare static histogram variables as aligned
- convert some initalizations to mallocz
- add window size parameter to init function and remove window size setter
  function
- convert return codes to AVERROR
- fix indentation

Signed-off-by: Marton Balint <cus@passwd.hu>

tools: add loudnorm script example to use loudnorm

Based on a patch by Kyle Swanson <k@ylo.ph>.

Signed-off-by: Marton Balint <cus@passwd.hu>

doc/encoders: sort list into alphabetical order

Signed-off-by: Lou Logan <lou@lrcd.com>

doc/encoders: add documentation for the Hap encoder

Documents options and behaviour, noting when 'chunks' option will
not be honoured.

Signed-off-by: Tom Butterworth <bangnoise@gmail.com>
Signed-off-by: Martin Vignali <martin.vignali@gmail.com>

icodec: correctly check avio_read return value

It can read less than the requested amount, in which case buf contains
uninitialized data, causing problems like segmentation faults later on.

Also make sure that image->size is positive, so that it can't match a
negative error code.

Reviewed-by: Michael Niedermayer <michael@niedermayer.cc>
Signed-off-by: Andreas Cadhalpun <Andreas.Cadhalpun@googlemail.com>

dvbsubdec: fix division by zero in compute_default_clut

This problem was introduced in commit
4b90dcb8493552c17a811c8b1e6538dae4061f9d.

Reviewed-by: Michael Niedermayer <michael@niedermayer.cc>
Signed-off-by: Andreas Cadhalpun <Andreas.Cadhalpun@googlemail.com>

proresdec_lgpl: explicitly check coff[3] against slice_data_size

The implicit checks via v_data_size and a_data_size don't work in the case
'(hdr_size > 7) && !ctx->alpha_info'.

This fixes segmentation faults due to invalid reads.

This problem was introduced in commit
547c2f002a87f4412a83c23b0d60364be5e7ce58.

Reviewed-by: Michael Niedermayer <michael@niedermayer.cc>
Signed-off-by: Andreas Cadhalpun <Andreas.Cadhalpun@googlemail.com>

lavc/utils.c: Make sure skip_samples never goes negative.

Signed-off-by: Sasi Inguva <isasi@google.com>
Reviewed-by: Derek Buitenhuis <derek.buitenhuis@gmail.com>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

lavf/mov.c: Fallback to finding non-keyframe in fix_index, if keyframe search fails.

Signed-off-by: Sasi Inguva <isasi@google.com>
Reviewed-by: Derek Buitenhuis <derek.buitenhuis@gmail.com>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

avcodec/hap: add "compressor" option to Hap encoder to disable secondary compression

The secondary compression in Hap is optional, this change exposes that option to
the user as some use-cases favour higher bitrate files to reduce workload
decoding.
Adds "none" or "snappy" as options for "compressor". Selecting "none" disregards
"chunks" option: chunking is only of benefit decompressing Snappy.

Reviewed-by: Martin Vignali <martin.vignali@gmail.com>
Signed-off-by: Tom Butterworth <bangnoise@gmail.com>

lavc/hevc_ps: Fix an error message.

lavc/hevc_ps: Use correct pix_fmt for 12bit 4:0:0.

Fixes part of ticket #5544.

avformat/flvenc: add add_keyframe_index option

Add keyframe index metadata
Used to facilitate seeking; particularly for HTTP pseudo streaming.
 1. read live streaming or file by sequence
 2. if use add_keyframe_index option, add a mark flag at the position,
    use to insert new context at the last step.
 3. add the keyframes *offset* and *timestamp* into a list
 4. if use add_keyframe_index option, shift the metadata data from
    mark flag offset
 5. insert the keyframes *offset* and *timestamp* from the list by
    sequence
 6. free the list
 7. end.

Add FATE test case;

Reviewed-by: Lou Logan <lou@lrcd.com>
Signed-off-by: Steven Liu <liuqi@gosun.com>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

nut: add gray12 support

Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

swscale: Add GRAY12

Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

pixfmt: Add GRAY12

Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

swscale: Add GBRAP10 output

Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

swscale: Add the GBRAP12 output

Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

swscale/output: Fix alpha shift in yuv2gbrp_full_X_c()

Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

swscale/input: Fix alpha bswap for set readAlpPlanar

Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

swscale: Drop is9_OR_10BPS() use, its name is not correct

Found-by: Luca Barbato
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

escape124: reject codebook size 0

It causes a cb_depth of 32, leading to assertion failures in get_bits.

Reviewed-by: Michael Niedermayer <michael@niedermayer.cc>
Signed-off-by: Andreas Cadhalpun <Andreas.Cadhalpun@googlemail.com>

icodec: add ico_read_close to fix leaking ico->images

Reviewed-by: Michael Niedermayer <michael@niedermayer.cc>
Signed-off-by: Andreas Cadhalpun <Andreas.Cadhalpun@googlemail.com>

icodec: fix leaking pkt on error

Reviewed-by: Michael Niedermayer <michael@niedermayer.cc>
Signed-off-by: Andreas Cadhalpun <Andreas.Cadhalpun@googlemail.com>

MAINTAINERS: Add myself to flvenc

Signed-off-by: Steven Liu <liuqi@gosun.com>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

mpegts: prevent division by zero

Reviewed-by: Michael Niedermayer <michael@niedermayer.cc>
Signed-off-by: Andreas Cadhalpun <Andreas.Cadhalpun@googlemail.com>

matroskadec: fix NULL pointer dereference in webm_dash_manifest_read_header

The code assumes that s->streams[0] is valid.

Reviewed-by: Michael Niedermayer <michael@niedermayer.cc>
Signed-off-by: Andreas Cadhalpun <Andreas.Cadhalpun@googlemail.com>

avcodec/hap: pass texture-compression destination as argument, not in context

This allows a subsequent change to compress directly into the output packet when possible.

Reviewed-by: Michael Niedermayer <michael@niedermayer.cc>
Signed-off-by: Tom Butterworth <bangnoise@gmail.com>

avformat/matroskadec: fix DiscardPadding element parsing

If the value is negative then it means padding at the start of the packet
instead of at the end.

Based on a patch by Hendrik Leppkes.

Reviewed-by: James Zern <jzern-at-google.com@ffmpeg.org>
Signed-off-by: James Almer <jamrial@gmail.com>

opus: move the entropy decoding functions to opus_rc.c

The intention is to have both encoding and decoding functions
in opus_rc.c.

Signed-off-by: Rostislav Pehlivanov <atomnuker@gmail.com>

opus: move all tables to a separate file

Signed-off-by: Rostislav Pehlivanov <atomnuker@gmail.com>

avformat/mov: Read multiple stsd from DV

Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

ffserver: use AVStream.codecpar in open_input_stream()

AVStream.codec is deprecated

Signed-off-by: Reynaldo H. Verdejo Pinochet <reynaldo@osg.samsung.com>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

ffserver: user AVStream.codecpar in compute_status()

AVStream.codec is deprecated

Signed-off-by: Reynaldo H. Verdejo Pinochet <reynaldo@osg.samsung.com>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

ffserver: use AVStream.codecpar in find_stream_in_feed()

AVStream.codec is deprecated

Signed-off-by: Reynaldo H. Verdejo Pinochet <reynaldo@osg.samsung.com>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

ffserver: get time_base from AVStream in print_stream_params()

AVStream.codec is deprecated

Signed-off-by: Reynaldo H. Verdejo Pinochet <reynaldo@osg.samsung.com>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

ffserver: use .codecpar instead of .codec in print_stream_params()

AVStream.codec is deprecated

Signed-off-by: Reynaldo H. Verdejo Pinochet <reynaldo@osg.samsung.com>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

ffserver: check for codec match using AVStream.codecpar

Compare using AVCodecParameters instead of the deprecated
AVStream.codec field

Signed-off-by: Reynaldo H. Verdejo Pinochet <reynaldo@osg.samsung.com>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

cmdutils: add show_demuxers and show_muxers

add -muxers and -demuxers parameters to list the dexmuers and muxers

Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

aacenc: quit when the audio queue reaches 0 rather than keeping track of empty frames

The libopus encoder does the same thing and its better than
keeping track of when the empty flush frames appear.

Signed-off-by: Rostislav Pehlivanov <atomnuker@gmail.com>

ffserver: Fix one AVFormatContext misuse

Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

hls: fix leaking avio_opts on hls_read_header error

Use the hls_close function to reduce code duplication.

Reviewed-by: Michael Niedermayer <michael@niedermayer.cc>
Signed-off-by: Andreas Cadhalpun <Andreas.Cadhalpun@googlemail.com>

hls: move hls_close above hls_read_header

This is needed for the following commit.

Reviewed-by: Michael Niedermayer <michael@niedermayer.cc>
Signed-off-by: Andreas Cadhalpun <Andreas.Cadhalpun@googlemail.com>

mpegaudio_parser: don't return AVERROR_PATCHWELCOME

The API does not allow returning AVERROR codes.

It triggers an assert in av_parser_parse2.

Reviewed-by: Michael Niedermayer <michael@niedermayer.cc>
Signed-off-by: Andreas Cadhalpun <Andreas.Cadhalpun@googlemail.com>

ffserver: steal SPACE_CHARS, remove one reason for internal.h that way

Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

ffserver: eliminate hardcoded literal and duplicated array size

Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

ffserver: use av_freep() for a case that is not clearly at the end of a function

Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

ffserver: Throw ffm.h out its not used except for a constant that is part of the format

Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

ffserver: drop a bunch of apparently unneeded internal headers

Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

avformat/hls: Add missing error check for avcodec_parameters_copy()

Signed-off-by: Anssi Hannula <anssi.hannula@iki.fi>

avformat/hls: Fix probing mpegts audio streams that use probing

Commit 04964ac311abe670f ("avformat/hls: Fix missing streams in some
cases with MPEG TS") caused a regression where subdemuxer streams that
use probing (e.g. dts/eac3/mp2 in mpegts) no longer get probed properly.

This is because the codec parameters from the subdemuxer stream, once
probed, are not passed on to the main stream.

Fix that by updating the codec parameters if the codec id changes.

Signed-off-by: Anssi Hannula <anssi.hannula@iki.fi>

avformat/hls: Factor copying stream info to a separate function

Signed-off-by: Anssi Hannula <anssi.hannula@iki.fi>

xmv: validate sample_rate

AVCodecParameters.sample_rate is a signed integer, so
XMVAudioPacket.sample_rate should be, too.

A negative sample rate doesn't make sense and triggers assertions in
av_rescale_rnd.

Signed-off-by: Andreas Cadhalpun <Andreas.Cadhalpun@googlemail.com>

voc_packet: validate sample_rate

A negative sample rate doesn't make sense and triggers assertions in
av_rescale_rnd.

Signed-off-by: Andreas Cadhalpun <Andreas.Cadhalpun@googlemail.com>

oggparsespeex: validate sample_rate

A negative sample rate doesn't make sense and triggers assertions in
av_rescale_rnd.

Signed-off-by: Andreas Cadhalpun <Andreas.Cadhalpun@googlemail.com>

nuv: validate sample_rate

A negative sample rate doesn't make sense and triggers assertions in
av_rescale_rnd.

Signed-off-by: Andreas Cadhalpun <Andreas.Cadhalpun@googlemail.com>

mvdec: validate sample_rate

A negative sample rate doesn't make sense and triggers assertions in
av_rescale_rnd.

Signed-off-by: Andreas Cadhalpun <Andreas.Cadhalpun@googlemail.com>