Replace "(a * (1 << shift) * b + c) >> shift" by "a * b + (c >> shift)".
It is equivalent to the old code because a is in the range of uint16_t,
shift is 12 and b is effectively a signed 4-bit number, so that no
overflow/truncation of high bits happens during the multiplication
(overflow would be undefined anyway).
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@gmail.com>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
Replace "((a << shift) + b) >> shift" by "a + (b >> shift)". This avoids
a left shift which also happens to trigger undefined behaviour in case "a"
is negative. This affected the FATE-tests acodec-adpcm-adx and
acodec-adpcm-adx-trellis; it also fixes ticket #8008.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@gmail.com>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
Affected the adpcm-afc, adpcm-ea-1, adpcm-ea-2, adpcm-ea-maxis-xa,
adpcm-thp and ea-cdata FATE-tests. Also fixes ticket #8487.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@gmail.com>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
Currently when hls_enc is active and there are multiple variant stream
outputs, default key file url construction does not work, because it is
based on the FormatContext' url field. But in case of multiple variant
streams, it contains the variant m3u8 output playlist url that contains
the %v placeholder. So the result key file url will hold the %v
placeholder causing run time error message about "could not write the
key file".
This patch correct this behaviour, and use the master playlist url for
constructing the output key file url when master playlist is vailable.
Signed-off-by: Bela Bodecs <bodecsb@vivanet.hu>
When program_date_time flag is present, in m3u8 playlist file each
segment has a corresponding EXT-X-PROGRAM-DATE-TIME value. The intial
program-date-time value is the actual current time at init and
each new segment increments this value by its duration. When append_list
flags is also present, existing playlist parsing by hls_append_segment
treats existing segments as new segments regarding the program-date-time
calculation. But it should not do that, because this way all real the
new segments' EXT-X-PROGRAM-DATE-TIME values will be shifted erroneously
by the sum duration of existing segments. Instead it should have
decremented the initial program-date-time value by its duration. This
would ensure that
the first new segment's program-date-time value had the current time as
it is expected.
This patch corrects this behaviour and prevent existing segments to
increment the value of initial_prog_date_time variable but decrements
it.
Reviewed-by: Steven Liu <lq@onvideo.cn>
Signed-off-by: Bela Bodecs <bodecsb@vivanet.hu>
fix ticket: 8461
there is no problem before commit 75aea52a10
Reviewed-by: Derek Buitenhuis <derek.buitenhuis@gmail.com>
Signed-off-by: Steven Liu <lq@chinaffmpeg.org>
Also simplify it and make it always log the error.
This fixes for example the image2 muxer when used with an URL which also
contains the protocol:
ffmpeg -f lavfi -i testsrc -vframes 10 -atomic_writing 1 file:out%d.png
Signed-off-by: Marton Balint <cus@passwd.hu>
As written in https://github.com/Haivision/srt/blob/v1.4.1/docs/API.md,
the nonblock mode is activated if SRTO_SNDSYN and SRTO_RCVSYN, for
sending and receiving respectively, are set to 0.
Signed-off-by: Marton Balint <cus@passwd.hu>
Fixes: Out of array access
Fixes: 19299/clusterfuzz-testcase-minimized-ffmpeg_BSF_HEVC_MP4TOANNEXB_fuzzer-5169193398042624
Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/projects/ffmpeg
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
Fixes: invalid memcpy use
Fixes: 19299/clusterfuzz-testcase-minimized-ffmpeg_BSF_HEVC_MP4TOANNEXB_fuzzer-5169193398042624
Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/projects/ffmpeg
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
"It is a requirement of bitstream conformance that the value returned from the leb128 parsing process is less than or equal
to (1 << 32) - 1."
Fixes: assertion failure
Fixes: 19293/clusterfuzz-testcase-minimized-ffmpeg_BSF_TRACE_HEADERS_fuzzer-5749508361420800
Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/projects/ffmpeg
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
Both S16LE as well as DTS can have lots of 0 bytes in silent segments
Using these results in error. Thus this patch skips 0 bytes in
comparission.
Fixes Ticket6561
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
HEVC standard supports multi-layer streams (ITU-T H.265 02/2018 Annex
F). Each NAL unit belongs to a particular layer defined by nuh_layer_id
in the header.
Currently, all NAL units that do not belong to a base layer are
automatically removed in ff_h2645_packet_split(). Some data may
therefore be lost when future filters/decoders are designed to support
multi-layer streams.
A better approach is to forward nuh_layer_id > 0 packets and let blocks
down the chain decide how to process them. The condition to remove
packets has been moved to hevcdec and cbs.
Found-by: Andreas Rheinhardt <andreas.rheinhardt@gmail.com>
Signed-off-by: Andriy Gelman <andriy.gelman@gmail.com>
Signed-off-by: James Almer <jamrial@gmail.com>
This will likely also fix CID 1452427, a false positive resulting from
Coverity thinking that av_dict_set() automatically frees its key and
value parameters (even without the AV_DICT_DONT_STRDUP_* flags).
(AV_DICT_APPEND and AV_DICT_DONT_STRDUP_VAL are compatible with each
other since a8c5b455, so we can reset this flag here. It has originally
been removed in 0dc66553 when appending was added.)
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@gmail.com>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
libx265.c references a member x265_picture.quantOffsets (for ROI
support) which was added in X265_BUILD 70. Increase the minimum libx265
version to fix compilation.
Signed-off-by: Andriy Gelman <andriy.gelman@gmail.com>
Reviewed-by: Derek Buitenhuis <derek.buitenhuis@gmail.com>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
When shifting the already written data in order to write the keyframe
index, the flv muxer would first store the pre-shift size, then
calculate how big the index will be eventually, then perform some seeks
to update some size fields, then seek back to the end of the file to get
the new position, followed by a seek to the position where writing will
really start. Seeking back to the (already known) end position (that is
actually used to perform this seek) to get the end position is of course
unnecessary. It has been removed.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@gmail.com>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
When the faststart option for the mov/mp4 muxer is used, the current
position (i.e. the size of the already written data pre-shifting) was
evaluated twice: First in an initialization and then again later,
overwriting the first value without having ever touched it. So remove
the initialization.
Also, the clone of this code in the Sega FILM muxer behaves the same and
has been treated the same.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@gmail.com>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
