avcodec/kmvc: Move commonly used variables to the front of the context
Reduces codesize because the offset in pointer+offset addressing
requires less bytes to encode. Reduces the size of .text from 8871B
to 8146B (GCC 10, -O3, x64).
Reviewed-by: Tomas Härdin <tjoppen@acc.umu.se> Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@gmail.com>
avcodec/vc1dec: Postpone allocating sprite frame to avoid segfault
Up until now, the VC-1 decoders allocated an AVFrame for usage with
sprites during vc1_decode_init(); yet said AVFrame can be freed if
(re)initializing the context (which happens ordinarily during decoding)
fails. The AVFrame does not get allocated again lateron in this case,
leading to segfaults.
Fix this by moving the allocation of said frame immediately before it is
used (this also means that said frame won't be allocated at all any more
in case of a regular (i.e. non-image) stream).
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@gmail.com>
It automatically records the current length of the string,
whereas the current code contains lots of instances of
snprintf(buf + strlen(buf), buf_size - strlen(buf), ...).
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@gmail.com>
avcodec/avcodec: Update check for identical colorspace/primaries/trc names
If the numerical constants for colorspace, transfer characteristics
and color primaries coincide, the current code presumes the
corresponding names to be identical and prints only one of them obtained
via av_get_colorspace_name(). There are two issues with this: The first
is that the underlying assumption is wrong: The names only coincide in
the 0-7 range, they differ for more recent additions. The second is that
av_get_colorspace_name() is outdated itself; it has not been updated
with the names of the newly defined colorspaces.
Fix both of this by using the names from
av_color_(space|primaries|transfer)_name() and comparing them via
strcmp; don't use av_get_colorspace_name() at all.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@gmail.com>
avformat/webpenc: Fix memleak when trailer is never written
When the trailer is never written (or when a stream switches from
non-animation mode to animation mode mid-stream), a cached packet
(if existing) would leak. Fix this by adding a deinit function.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@gmail.com>
avformat/webpenc: Fix memleak when using invalid packets
The WebP muxer sometimes caches a packet it receives to write it later;
yet if a cached packet is too small (so small as to be invalid),
it is cached, but not written and not unreferenced. Such a packet leaks,
either by being overwritten by the next packet or because it is never
unreferenced at all.
Fix this by not caching unusable packets at all; and error out on
invalid packets.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@gmail.com>
avformat/ipmovie: Avoid reading packets during read_header
They will be discarded anyway because this can only happen
for invalid data. This already implies that the pkt won't be used
at all when parsing the very first chunk when reading the header,
so one can use NULL as argument and remove the av_packet_unref()
on error.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@gmail.com>
When one of these errors happens during ipmovie_read_packet(),
an error is returned and the packet is cleaned up generically.
And since 712d3ac539f30239b764d8621829dc9dc913da61 the same happens
in ipmovie_read_header().
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@gmail.com>
Before 8d78e90a6ba96646f7f25aff6ca3e12e71cec164 the Matroska demuxer
used stack packets to hold temporary packets; now it uses a temporary
packet allocated by the Matroska demuxer. Yet because it used stack
packets the code has always properly reset the packet on error, while
on success these temporary packets were put into a packet list via
avpriv_packet_list_put(), which already resets the source packet.
This means that this code is compatible with just reusing
AVFormatInternal.parse_pkt (which is unused while one is in the
demuxer's read_packet() function). Compared to before 8d78e90a6
this no longer wastes one initialization per AVPacket read
(the resetting of the stack packet performed by av_packet_move_ref()
in avpriv_packet_list_put() was for naught).
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@gmail.com>
avcodec/svq3: Don't copy watermarked frame data twice
The SVQ3 decoder modifies the input bitstream at two places.
One of them is only reached when the file is watermarked.
Therefore commit 2264c1108135380c49fdf0aef97520bf77a6ed37
made a copy of all the frame data in this case.
But there is a second possibility for modifying the frame and
therefore Libav commit 1098f5c0495c61a98d4ff6b8e24c17974d4bace5
made the decoder always copy the data. This of course makes
the additional copy for watermarked frames redundant, but it hasn't
been removed. This commit does so.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@gmail.com>
avcodec/svq3: Use av_fast_padded_malloc() instead of av_fast_malloc()
It takes care of zeroing padding (which has been forgotten here).
Also rename the size variable to indicate that this is not the size
of the current slice.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@gmail.com>
This makes av_read_frame() return packets with proper timestamps.
As a result, seeking now works in combination with streamcopy.
A FATE-test for this has been added.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@gmail.com>
avformat/utils: Fix confusing return value for ff_read_packet()
Currently, ff_read_packet() sometimes forwards the return value of
AVInputFormat.read_packet() (which should be zero on success, but isn't
for all demuxers) and sometimes it overwrites this with zero.
Furthermore, it uses two variables, one for the read_packet return value
and one for other errors, which is a bit confusing; it is also
unnecessary given that the documentation explicitly states that
ff_read_packet() never returns positive values. Returning a positive
value would lead to leaks with some callers (namely asfrtp_parse_packet
and estimate_timings_from_pts). So always return zero in case of
success.
Because the properties of frames returned from ff_get/reget_buffer
are not reset at all, lots of returned frames had palette_has_changed
wrongly set to 1. This has been changed, too.
Reviewed-by: James Almer <jamrial@gmail.com> Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@gmail.com>
Derek Buitenhuis [Thu, 18 Mar 2021 16:12:08 +0000 (16:12 +0000)]
FATE: Add test for probing MOV/MP4 files with extended box sizes
The test sample has to have no file extension, otherwise probing
happens to work, based off file extension alone, and we want to
test the actual probing function.
Derek Buitenhuis [Thu, 18 Mar 2021 15:26:54 +0000 (15:26 +0000)]
avformat/mov: Fix extended atom size buffer length check
When extended atom size support was added to probing in fec4a2d232d7ebf6d1084fb568d4d84844f25abc, the buffer
size check was backwards, but probing continued to work
because there was no minimum size check yet, so despite
size being 1 on these atoms, and failing to read the 64-bit
size, the tag was still correctly read.
When 0b78016b2d7c36b32d07669c0c86bc4b4225ec98 introduced a
minimum size check, this exposed the bug, and broke probing
any files with extended atom sizes, such as entirely valid
large files that start whith mdat atoms.
avformat/pp_bnk: Fix memleaks when reading non-stereo tracks
Commit 6973df112275c8ea4af0bf3cb1338baecc1d06b3 added support
for music tracks by outputting its two containing tracks
together in one packet. But the actual data is not contiguous
in the file and therefore one can't simply use av_get_packet()
(which has been used before) for it. Therefore the packet was
now allocated via av_new_packet() and read via avio_read();
and this is also for non-music files.
This causes problems because one can now longer rely on things
done automatically by av_get_packet(): It automatically freed
the packet in case of errors; this lead to memleaks in several
FATE-tests covering this demuxer. Furthermore, in case the data
read is less than the data desired, the returned packet was not
zero-allocated (the packet's padding was uninitialized);
for music files the actual data could even be uninitialized.
The former problems are fixed by using av_get_packet() for
non-music files; the latter problem is handled by erroring out
unless both tracks could be fully read.
Reviewed-by: Zane van Iperen <zane@zanevaniperen.com> Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@gmail.com>
avformat/utils: Always leave parse_pkt in blank state, avoid resetting
Always leaving said packet in a blank state after having used it
allows to avoid having to reset it before one uses it; and it also
allows to use it in more places than just in parse_packet() here.
Reviewed-by: James Almer <jamrial@gmail.com> Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@gmail.com>
avcodec/parser: Don't return pointer to stack buffer
When flushing, the parser receives a dummy buffer with padding
that lives on the stack of av_parser_parse2(). Certain parsers
(e.g. Dolby E) only analyze the input, but don't repack it. When
flushing, such parsers return a pointer to the stack buffer and
a size of 0. And this is also what av_parser_parse2() returns.
Fix this by always resetting poutbuf in case poutbuf_size is zero.
Reviewed-by: James Almer <jamrial@gmail.com> Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@gmail.com>
Added in dc51a72ba45fbefb9f1c6c3ca5a5b2388d69b2da, yet even back then
the check was always true as the AVCodecContext has already been memset
to zero before that.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@gmail.com>
avcodec/avcodec: Sanitize options before using them
This is how it is supposed to happen, yet when using frame threading,
the codec's init function has been called before preinit. This can lead
to crashes when e.g. using unsupported lowres values for decoders
together with frame threading.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@gmail.com>
Fixes: Null pointer dereference Fixes: any mpeg4 testcase which fails the malloc at that exact spot Found-by: Rafael Dutra <rafael.dutra@cispa.de> Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
Fixes: signed integer overflow: 9223372036854775807 + 536870912 cannot be represented in type 'long' Fixes: 31678/clusterfuzz-testcase-minimized-ffmpeg_dem_MOV_fuzzer-5614204619980800 Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/projects/ffmpeg Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
Fixes: out of array access Fixes: 31733/clusterfuzz-testcase-minimized-ffmpeg_AV_CODEC_ID_SPEEDHQ_fuzzer-4704307963363328 Fixes: 31736/clusterfuzz-testcase-minimized-ffmpeg_AV_CODEC_ID_SPEEDHQ_fuzzer-6190960292790272 Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/projects/ffmpeg Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
Dominic Mayers [Thu, 18 Mar 2021 19:52:53 +0000 (15:52 -0400)]
avfilter/vf_ocr: add white space to whitelist
Fixes #9151. The current version of libavfilter/vf_ocr.c does not have white
space in the default whitelist. But it is recommanded to include white
space. See https://github.com/tesseract-ocr/tesseract/issues/2923
Moritz Barsnick [Sun, 23 Aug 2020 11:53:39 +0000 (13:53 +0200)]
avformat/http,tls: honor http_proxy command line variable for HTTPS
Add the "http_proxy" option and its handling to the "tls" protocol,
pass the option from the "https" protocol.
The "https" protocol already defines the "http_proxy" command line
option, like the "http" protocol does. The "http" protocol properly
honors that command line option in addition to the environment
variable. The "https" protocol doesn't, because the proxy is
evaluated in the underlying "tls" protocol, which doesn't have this
option, and thus only handles the environment variable, which it
has access to.
Fixes #7223.
Signed-off-by: Moritz Barsnick <barsnick@gmx.net> Signed-off-by: Marton Balint <cus@passwd.hu>
Hironori Bono [Tue, 23 Feb 2021 05:02:27 +0000 (14:02 +0900)]
lavc/videotoolboxenc: Add support for HEVC with Alpha.
This change supports the "HEVC Video with Alpha" profile introduced in WWDC 2019
<https://developer.apple.com/videos/play/wwdc2019/506/>. (This change is a
partial fix for Ticket #7965.)
For example, the following command converts an animation PNG file to an HEVC
with Alpha video:
./ffmpeg -i fate-suite/apng/clock.png -c:v hevc_videotoolbox -allow_sw 1 -alpha_quality 0.75 -vtag hvc1 clock.mov
(This change uses the "HEVC Video with Alpha" profile only when the
'-alpha_quality' value is not 0 for backward compatibility.)
Zane van Iperen [Fri, 13 Nov 2020 01:30:51 +0000 (11:30 +1000)]
avformat/pp_bnk: treat music files as stereo
These files are technically a series of planar mono tracks.
If the "music" flag is set, merge the packets from the two
mono tracks, essentially replicating:
avutil/adler32: Switch av_adler32_update() to size_t on bump
av_adler32_update() is used by av_hash_update() which will be switched
to size_t at the next bump. So it also has to be made to use size_t.
This is also necessary for framecrcenc.c, because the size of side data
will become a size_t, too.
Reviewed-by: James Almer <jamrial@gmail.com> Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@gmail.com>
avformat: Make AVChapter.id an int64_t on next major bump
64 bits are needed in order to retain the uid values of Matroska
chapters; the type is kept signed because the semantics of NUT chapters
depend upon whether the id is > 0 or < 0.
Reviewed-by: Anton Khirnov <anton@khirnov.net> Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@gmail.com>
avformat/matroskaenc: Check chapter ids for duplicates
Up until now, there has been no check that each chapter has a unique id;
there was only a check for whether a chapter id is zero (this happens
often when the chapters originated from a format that lacks the concept
of chapter id and simply counts from zero) which is invalid in Matroska.
In this case the chapter ids are offset by 1 to make them nonnegative.
Yet offsetting won't fix duplicate ids, therefore this is changed to
simply create new chapter uids when the input chapter uids don't conform
to the requirements of Matroska (in which case it can be presumed that
they did not originate from Matroska, so that we don't need to bother
to preserve them).
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@gmail.com>
avformat/tests/fifo_muxer: Fix memleak on error, fix API violation
The test program for the FIFO muxer allocates a buffer without padding
and wraps it into a packet via av_packet_from_data(). This is an API
violation. Furthermore, said buffer leaks in case av_packet_from_data()
fails. Fix both of these issues by using av_new_packet() instead.
Reviewed-by: James Almer <jamrial@gmail.com> Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@gmail.com>
If one of the two results of a ternary conditional is a pointer to void,
the type of the whole conditional operator is a pointer to void, even
when the other possible result is not a pointer to void. This loophole
in the type system has allowed mxf_read_local_tags to have a pointer of
type pointer to MXFMetadataSet that actually points to an MXFContext.
Reviewed-by: Tomas Härdin <tjoppen@acc.umu.se> Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@gmail.com>
projects / ffmpeg / log