avcodec/ac3enc: Avoid needlessly copying encoded packets around
AC-3 and EAC-3 are codecs whose packet sizes are known in advance,
so one can use the min_size parameter of ff_alloc_packet2() to
allocate exactly this amount. This avoids a memcpy later in
av_packet_make_refcounted() in encode_simple_internal().
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
avcodec/ac3enc: Use actual size of buffer in init_put_bits()
Since the very beginning (since de6d9b6404bfd1c589799142da5a95428f146edd)
the AC-3 encoder used AC3_MAX_CODED_FRAME_SIZE (namely 3840) for the
size of the output buffer (without any check at all).
This causes problems when encoding EAC-3 for which the maximum is too small,
smaller than the actual size of the buffer: One can run into asserts used
by the PutBits API. Ticket #8513 is about such a case and this commit
fixes it by using the real size of the buffer.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
avformat: Add and use helper function to add attachment streams
All instances of adding attached pictures to a stream or adding
a stream and an attached packet to said stream have several things
in common like setting the index and flags of the packet, setting
the stream disposition etc. This commit therefore factors this out.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
avcodec/qtrleenc: Fix negative linesizes, don't use NULL + offset
Before commit f1e17eb446577180ee9976730aacb46563766518, the qtrle
encoder had undefined pointer arithmetic: Outside of a loop, two
pointers were set to point to the ith element (with index i-1) of
a line of a frame. At the end of each loop iteration, these pointers
were decremented, so that they pointed to the -1th element of the line
after the loop. Furthermore, one of these pointers can be NULL (in which
case all pointer arithmetic is automatically undefined behaviour).
Commit f1e17eb44 added a check in order to ensure that the elements
never point to the -1th element of the array: The pointers are only
decremented if they are bigger than the frame's base pointer
(i.e. AVFrame.data[0]). Yet this check does not work at all in case of
negative linesizes; furthermore in case the pointer that can be NULL is
NULL initializing it still involves undefined pointer arithmetic.
This commit fixes both of these issues: First, non-NULL pointers are
initialized to point to the element after the ith element and
decrementing is moved to the beginning of the loop. Second, if a pointer
is NULL, it is just made to point to the other pointer, as this allows
to avoid checks before decrementing it.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@gmail.com>
Fixes: out of array access Fixes: 32264/clusterfuzz-testcase-minimized-ffmpeg_AV_CODEC_ID_LSCR_fuzzer-6684504010915840 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 write Fixes: 32116/clusterfuzz-testcase-minimized-ffmpeg_dem_SIMBIOSIS_IMX_fuzzer-6702533894602752 Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/projects/ffmpeg Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
avcodec/h265_metadata_bsf: Check nb_units before accessing the first in h265_metadata_update_fragment()
Fixes: null pointer dereference Fixes: 32113/clusterfuzz-testcase-minimized-ffmpeg_BSF_HEVC_METADATA_fuzzer-4803262287052800
Same as 0c48c332eeb2866d9353125f701e099c48889463
Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/projects/ffmpeg Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
avformat/rmdec: use larger intermediate type for audio_framesize * sub_packet_h check
Fixes: signed integer overflow: 65535 * 65535 cannot be represented in type 'int' Fixes: 31406/clusterfuzz-testcase-minimized-ffmpeg_dem_IVR_fuzzer-5024692843970560 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: 31386/clusterfuzz-testcase-minimized-ffmpeg_AV_CODEC_ID_EXR_fuzzer-5773234709594112 Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/projects/ffmpeg Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
avformat/mpc8: check for size overflow in mpc8_get_chunk_header()
Fixes: signed integer overflow: -9223372036854775760 - 50 cannot be represented in type 'long' Fixes: 31673/clusterfuzz-testcase-minimized-ffmpeg_dem_MPC8_fuzzer-580134751869337 Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/projects/ffmpeg Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
avcodec/mpegvideo: Update chroma_?_shift in ff_mpv_common_frame_size_change()
Fixes: out of array access Fixes: 31201/clusterfuzz-testcase-minimized-ffmpeg_AV_CODEC_ID_MPEG4_fuzzer-4627865612189696.fuzz Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/projects/ffmpeg Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
avformat/utils: Extend overflow check in dts wrap in compute_pkt_fields()
Fixes: signed integer overflow: -9223372032574480351 - 4294967296 cannot be represented in type 'long long' Fixes: 30022/clusterfuzz-testcase-minimized-ffmpeg_dem_KUX_fuzzer-5568610275819520 Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/projects/ffmpeg Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
avcodec/mjpegenc: Fix segfault when freeing incomplete context
When allocating the MJpegContext fails (or if the dimensions run afoul
of the 65500x65500 limit), an attempt to free a subbuffer of said
context leads to a segfault in ff_mjpeg_encode_close().
Seems to be a regression since 467d9e27e0cb2bf74f41dc832f2f8d191ba58ec9.
Reviewed-by: Michael Niedermayer <michael@niedermayer.cc> Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
In this case it also fixes a potential for compilation failures:
Not all compilers can handle the case in which a function with
a forward declaration declared with an attribute to always inline it
is called before the function body appears. E.g. GCC 4.2.1 on OS X 10.6
doesn't like it.
Reviewed-by: Pavel Koshevoy <pkoshevoy@gmail.com> Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
Specifically test that the WebVTT flavour is correctly mapped to
the Matroska/WebM CodecID and back; and test that dispositions
unsupported by WebM are discarded even when they would be supported
by Matroska.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
Mark Plomer [Thu, 18 Mar 2021 12:19:16 +0000 (13:19 +0100)]
avcodec/dv_profile: PAL DV files with dsf flag 0 - detect via pal flag and buf_size
Some old DV AVI files have the DSF-Flag of frames set to 0, although it
is PAL (maybe rendered with an old Ulead Media Studio Pro) ... this causes
ffmpeg/VLC-player to produce/play corrupted video (other players/editors
like VirtualDub work fine).
Fixes ticket #8333 and replaces/extends hack for ticket #2177
Alok Priyadarshi [Thu, 25 Mar 2021 04:46:36 +0000 (21:46 -0700)]
avformat/rtpdec: Fix prft wallclock time.
Timestamp difference is available in media timebase (1/90K) where as
rtcp time is in the default microseconds timebase. This patch fixes
the calculated prft wallclock time by rescaling the timestamp delta
to the microseconds timebase.
avcodec/put_bits: Don't set size_in_bits, fix overflow
A PutBitContext has a field called size_in_bits which is set to the
context's bitsize init_put_bits(); but it isn't used at all (the PutBits
API uses pointers directly and not bit indexes), so remove it (due to
ABI concerns the actual element is only removed at the next bump).
Furthermore, the multiplication inherent in setting this field can lead
to undefined integer overflows. This is particularly true for FFV1,
which uses a very big worst-case buffer (37*4*width*height; even
ordinary 1080p triggers an overflow). Ticket #8350 is about this
overflow which this commit fixes.
This means that the effective range of the PutBits API is no longer
restricted by the /8 as long as one isn't using put_bits_(count|left).
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@gmail.com>
avcodec/proresenc_kostya: Factor flushing PutBitContext out
The function to write an ordinary (luma or chroma) plane as well as
the function for writing an alpha plane have some similarities:
They record the initial bitposition (despite said position always being
byte-aligned), flush the PutBitContext themselves and return the amount
of bytes they wrote.
This commit factors this out; it also replaces bitpositions by
bytepositions and it avoids recording the initial byteposition because
said information is already available from the position at the end of
the last plane.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@gmail.com>
avcodec/utvideoenc: Don't use bitcounts when byte-aligned
Despite write_huff_codes() receiving an ordinary buffer (not a
PutBitContext), it returned the amount of data written in bits,
not in bytes. This has been changed: There is now no intermediate
bitcount any more.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@gmail.com>
avcodec: Remove cumbersome way of checking for amount of bytes left
Several encoders used code like the following to check for the amount of
bytes left in a PutBitContext:
pb->buf_end - pb->buf - (put_bits_count(pb) >> 3)
Besides the fact that using the pointers directly might pose
a maintainence burden in the future this also leads to suboptimal code:
The above code reads all three pointers (buf, buf_ptr and buf_end), but
touching buf is unnecessary and switching to put_bytes_left()
automatically fixes this.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@gmail.com>
avcodec/put_bits: Add functions for amount of bytes written/left
Often a caller doesn't want the amount of bits written via a
PutBitContext, but the amount of bytes. This in particular happens
after one has flushed the PutBitContext (e.g. at the end of encoding,
when one wants to know the actual packet size). The current way of doing
this is with put_bits_count(pb)/8 (or (put_bits_count(pb) + 7)/8).
Yet this has some issues: It contains implicit multiplications and
divisions by 8 with a cast in between; it obscurs the intent; and
it restricts the size of the buffer to (currently) INT_MAX/8 (or
to 1/8 of the maximum of whatever put_bits_count() returns), although
said restriction is not really necessary for users that don't need
a bitcount.
Corresponding functions for the amount of bytes left have also been
addded.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@gmail.com>
Fixes: Division by 0 Fixes: 30952/clusterfuzz-testcase-minimized-ffmpeg_BSF_SETTS_fuzzer-6601016202100736 Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/projects/ffmpeg Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
avformat/wtvdec: Check size in SBE2_STREAM_DESC_EVENT / stream2_guid
Fixes: signed integer overflow: 539033600 - -1910497124 cannot be represented in type 'int' Fixes: 30928/clusterfuzz-testcase-minimized-ffmpeg_dem_WTV_fuzzer-5922630966312960 Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/projects/ffmpeg Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
avformat/utils: Fix integer overflow with duration_gcd in ff_rfps_calculate()
Fixes: signed integer overflow: 136323327 * 281474976710656 cannot be represented in type 'long' Fixes: 30913/clusterfuzz-testcase-minimized-ffmpeg_dem_IVF_fuzzer-5753392189931520 Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/projects/ffmpeg Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
The check has been added at a time when the code performed the
multiplication itself instead of deferring it to av_malloc_array()
and when our allocation functions used unsigned instead of size_t.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@gmail.com>
The check has been added at a time when the code performed the
multiplication itself instead of deferring it to av_malloc_array()
and when our allocation functions used unsigned instead of size_t.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@gmail.com>
This check is outdated because the caller doesn't need to check that
the multiplication overflows when using av_realloc_array() (the code
in question used av_realloc() before that); furthermore, the check
is also a remnant of the time in which our allocation functions
didn't use size_t parameters.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@gmail.com>
avcodec/sonic: Use unsigned temporary in predictor_calc_error()
Fixes: signed integer overflow: -2147471366 - 18638 cannot be represented in type 'int' Fixes: 30157/clusterfuzz-testcase-minimized-ffmpeg_AV_CODEC_ID_SONIC_fuzzer-5171199746506752 Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/projects/ffmpeg Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
avformat/jacosubdec: Use 64bit intermediate for start/end timestamp shift
Fixes: signed integer overflow: -1957694447 + -1620425806 cannot be represented in type 'int' Fixes: 30207/clusterfuzz-testcase-minimized-ffmpeg_dem_JACOSUB_fuzzer-5050791771635712 Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/projects/ffmpeg Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
Fixes: signed integer overflow: -2147483648 - 1 cannot be represented in type 'int' Fixes: 30209/clusterfuzz-testcase-minimized-ffmpeg_dem_FLV_fuzzer-5724831658147840 Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/projects/ffmpeg Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
James Almer [Fri, 19 Mar 2021 01:45:12 +0000 (22:45 -0300)]
avcodec/libxvid: remove unnecessary output packet data check
The user buffers passed to avcodec_encode_video2() haven't been propagated to
AVCodec.encode2 implementations since 93016f5d1d280f9cb7856883af287fa66affc04c.
Also, the generic encode code already unrefs the packet if nothing was encoded.
avformat/mov: Check sample size for overflow in mov_parse_stsd_audio()
Fixes: signed integer overflow: 2 * 1914708000 cannot be represented in type 'int' Fixes: 31639/clusterfuzz-testcase-minimized-ffmpeg_dem_MOV_fuzzer-6303428239294464 Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/projects/ffmpeg Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
avcodec/sga: Check for array end in lzss_decompress()
Fixes: out of array access Fixes: 31640/clusterfuzz-testcase-minimized-ffmpeg_AV_CODEC_ID_SGA_fuzzer-5630883286614016 Fixes: 31619/clusterfuzz-testcase-minimized-ffmpeg_AV_CODEC_ID_SGA_fuzzer-5176667708456960 Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/projects/ffmpeg Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
avformat/sbgdec: Check for overflow in last loop in expand_timestamps()
Fixes: signed integer overflow: 9223372036854775807 + 86400000000 cannot be represented in type 'long' Fixes: 31003/clusterfuzz-testcase-minimized-ffmpeg_dem_SBG_fuzzer-6256298771480576 Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/projects/ffmpeg Reviewed-by: Nicolas George <george@nsup.org> Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
avcodec/ffwavesynth: Avoid signed integer overflow in phi_at()
Fixes: signed integer overflow: 2314885530818453536 - -9070214327174160352 cannot be represented in type 'long' Fixes: 31000/clusterfuzz-testcase-minimized-ffmpeg_AV_CODEC_ID_FFWAVESYNTH_fuzzer-6558389742206976 Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/projects/ffmpeg Reviewed-by: Nicolas George <george@nsup.org> Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
projects / ffmpeg / log