Fixes: runtime error: signed integer overflow: 2 * -1094995530 cannot be represented in type 'int' Fixes: 3512/clusterfuzz-testcase-minimized-4812747210489856 Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/projects/ffmpeg Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
Ivan Kalvachev [Sun, 8 Oct 2017 23:40:26 +0000 (02:40 +0300)]
Fix crash if av_vdpau_bind_context() is not used.
The public functions av_alloc_vdpaucontext() and
av_vdpau_alloc_context() are allocating AVVDPAUContext
structure that is supposed to be placed in avctx->hwaccel_context.
However the rest of libavcodec/vdpau.c uses avctx->hwaccel_context
as struct VDPAUHWContext, that is bigger and does contain
AVVDPAUContext as first member.
The usage includes write to the new variables in the bigger stuct,
without checking for block size.
Fix by always allocating the bigger structure.
Signed-off-by: Ivan Kalvachev <ikalvachev@gmail.com>
* commit '57b753b445e23363c997a8ec1c556e0b0f6e9da3':
build: Prefer NASM assembler over YASM
build: Make x86 assembler commandline-selectable
build: Special-case handling of SDL CFLAGS
[10:06:59 CEST] <BtbN> jamrial, 3303f86467 is a no-op. NV_ENC_PARAMS_RC_2_PASS_VBR is a deprecated rc mode(http://git.videolan.org/?p=ffmpeg.git;a=blob;f=compat/nvenc/nvEncodeAPI.h;h=c3a829421282d5f22f82fc285723f13eb660f053;hb=HEAD#l268).
[10:07:14 CEST] <BtbN> And the first hunk with qmin/qmax was applied to ffmpeg quite a while ago already.
[10:07:49 CEST] <BtbN> In a slightly different fashion, but with the same effect
[10:07:58 CEST] <BtbN> Came as a patch from nvidia iirc
Ivan Kalvachev [Sun, 8 Oct 2017 22:25:00 +0000 (01:25 +0300)]
Fix visual glitch with XvMC, caused by wrong idct permutation.
In the past XvMC forced simple_idct since
it was using FF_IDCT_PERM_NONE.
However now we have SIMD variants of simple_idct that
are using FF_IDCT_PERM_TRANSPOSE and if they are selected
XvMC would get coefficients in the wrong order.
The patch creates new FF_IDCT_NONE that
is used only for this kind of hardware decoding
and that fallbacks to the old C only simple idct.
Signed-off-by: Ivan Kalvachev <ikalvachev@gmail.com> Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
Sasi Inguva [Tue, 10 Oct 2017 17:36:58 +0000 (10:36 -0700)]
ffmpeg.c: Fallback to duration_dts, when duration_pts can't be determined.
This is required for FLV files, for which duration_pts comes out to be zero.
Signed-off-by: Sasi Inguva <isasi@google.com> Reviewed-by: Thomas Mundt <tmundt75@gmail.com> Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
Kaustubh Raste [Mon, 9 Oct 2017 12:17:34 +0000 (17:47 +0530)]
avcodec/mips: Improve hevc uni-w horiz mc msa functions
Load the specific destination bytes instead of MSA load and pack.
Pack the data to half word before clipping.
Use immediate unsigned saturation for clip to max saving one vector register.
Signed-off-by: Kaustubh Raste <kaustubh.raste@imgtec.com> Reviewed-by: Manojkumar Bhosale <Manojkumar.Bhosale@imgtec.com> Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
avcodec/dirac_dwt: Fix integer overflow in COMPOSE_53iL0()
Fixes: runtime error: signed integer overflow: 2147483646 + 2 cannot be represented in type 'int' Fixes: 3485/clusterfuzz-testcase-minimized-4940429332054016 Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/projects/ffmpeg Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
avcodec/mpeg4videodec: Use 64 bit intermediates for sprite delta
Fixes: runtime error: signed integer overflow: -104713 * 65536 cannot be represented in type 'int' Fixes: 3453/clusterfuzz-testcase-minimized-5555554657239040 Fixes: 3528/clusterfuzz-testcase-minimized-6283628420005888 Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/projects/ffmpeg Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
libavdevice/decklink: add support for 10-bit output for Decklink SDI
Can be tested via the following command:
./ffmpeg -i foo.ts -f decklink -vcodec v210 'DeckLink Duo (1)'
Note that the 8-bit support works as it did before, and setting
the pix_fmt isn't required for 10-bit mode. The code defaults to
operating in 8-bit mode when no vcodec is specified, for backward
compatibility.
Updated to reflect feedback from Marton Balint <cus@passwd.hu>
libavdevice/decklink: add support for -sources and -sinks arguments
Add support for enumerating the sources/sinks via the ffmpeg
command line options, as opposed to having to create a real pipeline
and use the "-list_devices" option which does exit() after dumping
out the options.
Note that this patch preserves the existing "-list_devices" option,
but now shares common code for the actual enumeration.
Updated to reflect feedback from Marton Balint <cus@passwd.hu>.
James Almer [Sun, 8 Oct 2017 19:54:18 +0000 (16:54 -0300)]
build: prevent SDL2 from polluting global cflags and extralibs
Remove the SDL_main define from the global cflags but not from the
ffplay cflags, and the -mwindows linker option from extralibs instead
of overriding it with the addition of -mconsole.
Reviewed-by: wm4 <nfxjfg@googlemail.com> Signed-off-by: James Almer <jamrial@gmail.com>
Mark Thompson [Sun, 8 Oct 2017 14:19:17 +0000 (15:19 +0100)]
hwcontext: Perform usual initialisation on derived device contexts
The initialisation should be common. For libmfx, it was previously
happening in the derivation function and this moves it out. For VAAPI,
it fixes some failures when deriving from a DRM device because this
initialisation did not run.
Mark Thompson [Sun, 1 Oct 2017 21:51:20 +0000 (22:51 +0100)]
vaapi: Always free parameter buffers after vaEndPicture() with libva2
This is an ABI change in libva2: previously the Intel driver had this
behaviour and it was implemented as a driver quirk, but now it is part
of the specification so all drivers must do it.
Mark Thompson [Sun, 8 Oct 2017 13:48:24 +0000 (14:48 +0100)]
vaapi: Remove H.264 baseline profile
This has been deprecated in libva2 because hardware does not and will not
support it. Therefore never consider it for decode, and for encode assume
the user meant constrained baseline profile instead.
Marton Balint [Fri, 6 Oct 2017 19:49:09 +0000 (21:49 +0200)]
ffmpeg: always use single threaded decoding for attached pictures
Since af1761f7b5b1b72197dc40934953b775c2d951cc ffmpeg waits for a frame in each
stream before writing the output header. If we are using threaded decoding for
attached pictures, we have to read till EOF to be able to finally flush the
decoder and output the decoded frame. This essentially makes ffmpeg buffer all
non-attached picture packets, which will cause a "Too many packets buffered for
output stream" eventually.
By forcing single threaded decoding, we get a frame from a single packet as
well and we can avoid the error.
Fixes part of ticket #6375:
ffmpeg -i 46564100.mp3 -acodec libmp3lame -ab 128k -ac 2 out.mp3
Reviewed-by: Hendrik Leppkes <h.leppkes@gmail.com> Signed-off-by: Marton Balint <cus@passwd.hu>
projects / ffmpeg / log