1
0
mirror of https://github.com/FFmpeg/FFmpeg.git synced 2024-12-18 03:19:31 +02:00
Commit Graph

112042 Commits

Author SHA1 Message Date
Lynne
3ef1e50c92
lavfi/transpose_vulkan: fix memory leaks 2023-08-28 22:29:36 +02:00
Lynne
d0ab2e2f98
lavfi/scale_vulkan: fix memory leaks 2023-08-28 22:29:35 +02:00
Lynne
0f2ae1ba69
lavfi/nlmeans_vulkan: fix memory leaks 2023-08-28 22:29:35 +02:00
Lynne
9944e96c61
lavfi/chromaber_vulkan: fix memory leaks 2023-08-28 22:29:34 +02:00
Lynne
a4673c9dff
lavfi/bwdif_vulkan: fix memory leaks 2023-08-28 22:29:34 +02:00
Lynne
b6cc53092a
lavfi/avgblur_vulkan: fix memory leaks 2023-08-28 22:29:34 +02:00
Lynne
f6cf3a40e4
vulkan: check for extension rather than function pointer
The loader ensures only that functions with tagged supported extensions
exist, rather than ensuring only those with supported extensions are
loaded.
As the init function uses Vulkan functions, whose loading requires them
to have the extension flags set, the extension flags are guaranteed
to also exist at this point.
2023-08-28 22:29:33 +02:00
Lynne
747871a42c
vulkan: do not leak cooperative matrix properties 2023-08-28 22:29:29 +02:00
Michael Niedermayer
b5273c619d
tests/fate: Add NoLegacy-cut.ape test
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
2023-08-28 01:10:14 +02:00
Michael Niedermayer
86c092a0ed
avcodec/apedec: Implement interim mode detection
Fixes: NoLegacy.ape
Found-by: Matt Ashland <mail@monkeysaudio.com>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
2023-08-28 01:10:13 +02:00
Michael Niedermayer
7995e175b8
avcodec/apedec: remove unused variable
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
2023-08-28 01:10:13 +02:00
Jun Zhao
c48ec95ba3 lavc/libx264: replace ITU-T T35(A/53 CC) SEI type by enum value
replace ITU-T T35(A/53 CC) SEI type by enum value

Reviewed-by: Kieran Kunhya <kierank@obe.tv>
Signed-off-by: Jun Zhao <barryjzhao@tencent.com>
2023-08-27 23:21:31 +08:00
Niklas Haas
3c9dc009b6 lavfi/vf_libplacebo: add extra_opts AVDictionary
Can be used to configure libplacebo's underlying raw options, which
sometimes includes new or advanced / in-depth settings not (yet) exposed
by vf_libplacebo.
2023-08-27 13:37:03 +02:00
Niklas Haas
816983d951 lavfi/vf_libplacebo: switch to new pl_options struct
This new upstream struct simplifies params struct management by allowing
them to all be contained in a single dynamically allocated struct. This
commit switches to the new API in a backwards-compatible way.

The only nontrivial change that was required was to handle
`sigmoid_params` in a way consistent with the rest of the params
structs, instead of setting it directly to the upstream default.
2023-08-27 13:37:03 +02:00
Leo Izen
e8a63b4763
avformat/jpegxl: remove jpegxl_probe, instead call avcodec/jpegxl_parse
This prevents code duplication in the source form by calling the parse
code that was moved to avcodec last commit. The code will be duplicated
in binary form for shared builds (it's not that large), but for source
code it will only exist in one location now.

Signed-off-by: Leo Izen <leo.izen@gmail.com>
2023-08-27 01:36:18 -04:00
Leo Izen
0c0dd23fe1
avcodec/jpegxl_parser: add JPEG XL parser
Add a full parser to libavcodec for AV_CODEC_ID_JPEGXL. It finds the
end of the stream in order to packetize the codec, and it looks at
the headers to set preliminary information like dimensions and pixel
format.

Note that much of this code is duplicated from avformat/jpegxl_probe.c,
but that code will be removed and call this instead in the next commit.

Signed-off-by: Leo Izen <leo.izen@gmail.com>
2023-08-27 01:36:08 -04:00
Leo Izen
7a69f7312e
avcodec/libjxldec: use internal AVFrame as buffered space
Before this commit, the decoder erroneously assumes that the AVFrame
passed to the receive_frame is the same one each time. Now it keeps an
internal AVFrame to write into, and copies it over when it's done.

Signed-off-by: Leo Izen <leo.izen@gmail.com>
2023-08-27 01:36:03 -04:00
Leo Izen
245910d5c9
avcodec/libjxldec: fix errors when decoding grayscale after rgb
Fixes an error that's caused by decoding a grayscale JXL image after an
RGB image is decoded, with the same decoder instance.

Signed-off-by: Leo Izen <leo.izen@gmail.com>
2023-08-27 01:36:00 -04:00
Lynne
358919506d
vulkan: enable VK_KHR_cooperative_matrix
It's of interest to API users, and of interest to us,
as a DCT/DST can be implemented via matrix multiplies.
2023-08-26 23:14:53 +02:00
Christophe Gisquet
8028b8260e avcodec/huffyuvdec: don't prepare unnecessary joint tables
The number of planes can be lower than 4.
2023-08-26 19:22:25 +02:00
Zhao Zhili
67d392b979 examples/transcode: flush decoder on EOF
Signed-off-by: Zhao Zhili <zhilizhao@tencent.com>
2023-08-26 17:02:21 +08:00
Zhao Zhili
2387328fa2 avfilter/vf_transpose_vt: fix declaration-after-statement
Signed-off-by: Zhao Zhili <zhilizhao@tencent.com>
2023-08-26 16:31:43 +08:00
Zhao Zhili
7cca9c07ca avfilter/vf_transpose_vt: fix output frame dimension
Create a new hardware frame context when necessary.

Signed-off-by: Zhao Zhili <zhilizhao@tencent.com>
2023-08-26 16:31:43 +08:00
Zhao Zhili
9b543881ef avfilter/vf_scale_vt: fix output frame dimension
The output frame dimension is incorrect because it shares hardware
frame context with input.

Signed-off-by: Zhao Zhili <zhilizhao@tencent.com>
2023-08-26 16:31:43 +08:00
Wenbin Chen
e79bd1f1b1 lavfi/dnn: Add OpenVINO API 2.0 support
OpenVINO API 2.0 was released in March 2022, which introduced new
features.
This commit implements current OpenVINO features with new 2.0 APIs. And
will add other features in API 2.0.
Please add installation path, which include openvino.pc, to
PKG_CONFIG_PATH mannually for new OpenVINO libs config.

Signed-off-by: Ting Fu <ting.fu@intel.com>
Signed-off-by: Wenbin Chen <wenbin.chen@intel.com>
2023-08-26 14:12:10 +08:00
Michael Niedermayer
80ad0e2198
avcodec/apedec: Fix 48khz 24bit below insane level
Fixes: Ticket9816
Fixes: vlc.ape and APE_48K_24bit_2CH_02_01.ape

Regression since: ed0001482a.

Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
2023-08-26 00:20:23 +02:00
Michael Niedermayer
696e161919
avcodec/apedec: Fix CRC for 24bps and bigendian
Fixes CRC for vlc.ape and APE_48K_24bit_2CH_02_01.ape

Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
2023-08-26 00:20:09 +02:00
Kacper Michajłow
9f66286f0b avcodec/vulkan_decode: print also codec header name
Signed-off-by: Kacper Michajłow <kasper93@gmail.com>
2023-08-24 22:51:36 +02:00
Kacper Michajłow
9d0da996f0 avcodec/vulkan_decode: fix struct type for h265_profile
Signed-off-by: Kacper Michajłow <kasper93@gmail.com>
2023-08-24 22:51:25 +02:00
Marvin Scholz
15a89a9422 lavfi/vf_xfade_vulkan: add wipes
Add the wipetl, wipetr, wipebl, wipebr effects.
2023-08-24 22:50:57 +02:00
Marvin Scholz
85c7c79a62 lavfi/vf_xfade_vulkan: add pixelize 2023-08-24 22:50:57 +02:00
Marvin Scholz
98bd8a7f5c lavfi/vf_xfade_vulkan: add dissolve 2023-08-24 22:50:57 +02:00
Marvin Scholz
8d3fd489e2 lavfi/vf_xfade_vulkan: reindent after last commit 2023-08-24 22:50:57 +02:00
Marvin Scholz
95e366fb59 lavfi/vf_xfade_vulkan: add circleopen/circleclose 2023-08-24 22:50:57 +02:00
Benjamin Cheng
b4f0701358 vulkan_h264: send scaling lists in zig-zag order
h264_ps turns the scaling lists into matrices with a raster scan order,
but Vulkan wants the scaling lists as originally defined.
2023-08-24 22:50:57 +02:00
Paul B Mahol
52ebb15ec1 avfilter/af_afade: add 4 more curves 2023-08-24 22:20:06 +02:00
Matthias Dressel
e41bd6e65e checkasm: hevc_sao: Fix a regression in hevc_sao_edge
check_func() might return NULL, in which case the function is not to be
benched. Introduced in cc679054c7.

Signed-off-by: Matthias Dressel <code@deadcode.eu>
Signed-off-by: Martin Storsjö <martin@martin.st>
2023-08-24 22:09:37 +03:00
Rémi Denis-Courmont
9bc5676e40 lavc/g722dsp: add RISC-V V DSP function 2023-08-24 21:07:18 +03:00
Rémi Denis-Courmont
6f8ac298da lavu/timer: specify RISC-V time unit 2023-08-24 20:58:57 +03:00
Rémi Denis-Courmont
09731fbfc3 lavu/timer: remove gratuitous volatile
AV_READ_TIME has no side effects. It does not need to be volatile.
2023-08-24 20:58:57 +03:00
Rémi Denis-Courmont
05115a77e0 lavu/timer: use time for AV_READ_TIME on RISC-V
So far, AV_READ_TIME would return the cycle counter. This posed two
problems:
1) On recent systems, it would just raise an illegal instruction
   exception. Indeed RDCYCLE is blocked in user space to ward off some
   side channel attacks. In particular, this would cause the random
   number generator to crash.
2) It does not match the x86 behaviour and the apparent original intent
   of AV_READ_TIME in the functional code base (outside test cases).

So this replaces the cycle counter with the time counter. The unit is
a platform-dependent constant fraction of time, and the value should be
stable across harts (RISC-V lingo for physical CPU thread).
2023-08-24 20:58:57 +03:00
Michael Niedermayer
5f5a1ccd04
avcodec/wavarc: Check that nb_samples is not negative
It is currently probably not possible for it to be negative as
the needed 2Mb input buf size is not achievable. But it is more
robust to check for it too.
If it would become negative than code like
s->samples[0][n] = s->samples[0][s->nb_samples + n];
would crash

Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
2023-08-23 22:52:26 +02:00
Michael Niedermayer
1a81a40de2
avcodec/wavarc: Check shift
Fixes: shift exponent 1285 is too large for 32-bit type 'int'
Fixes: 60870/clusterfuzz-testcase-minimized-ffmpeg_AV_CODEC_ID_WAVARC_fuzzer-5332050340347904

Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/projects/ffmpeg
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
2023-08-23 22:52:26 +02:00
Michael Niedermayer
0ce322a51e
avcodec/xvididct: Fix integer overflow in idct_row()
Fixes: signed integer overflow: -1403461578 + -843974775 cannot be represented in type 'int'
Fixes: 60868/clusterfuzz-testcase-minimized-ffmpeg_AV_CODEC_ID_MPEG1VIDEO_fuzzer-4599793035378688

Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/projects/ffmpeg
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
2023-08-23 22:52:26 +02:00
Timo Rothenpieler
f225f8d746 avformat/hls: reset the playlist init segment on seek
The mp4 demuxer gets very upset when it gets flushed and not re-fed this
data.
2023-08-22 23:46:11 +02:00
Aleksoid
7eb0d9e905 avcodec/nvenc: add option to limit slice size 2023-08-22 23:46:11 +02:00
Martin Storsjö
95433eb3aa fate/matroska: Stop using unicode in command line paramters
The fate-run.sh shell script exports LC_ALL=C before invoking the
test executables; this is probably done for consistency.

When executing Windows binaries with Wine, it normally handles
UTF-8 command line parameters just fine - but with LC_ALL set to
C, it treats them as plain ASCII.

As the unicode command line parameters wasn't the main thing
being tested here, just convert them to plain ASCII, for
portability. This fixes the test for all test configurations that
use Wine.

Signed-off-by: Martin Storsjö <martin@martin.st>
2023-08-22 22:09:40 +03:00
Martin Storsjö
aea2d269db configure: Include objbase.h when checking for CoTaskMemFree
ddc1cd5cdd defined WIN32_LEAN_AND_MEAN
globally, which makes for much fewer transitive includes from
windows.h. With that define, CoTaskMemFree no longer gets
implicitly declared by just including windows.h, but one has to
include the right header objbase.h too.

That commit caused ole32 to no longer get detected, which caused
dxva2 to no longer be enabled. This gets fixed by this patch.

Signed-off-by: Martin Storsjö <martin@martin.st>
2023-08-21 23:11:43 +03:00
Evgeny Pavlov
cb1479faca avfilter/vf_ssim: Fix x86 assembly code for SSIM calculation
This commit fixes bug #10495

The code had several bugs related to post-loop compensation code:
- test assembly instruction performs bitwise AND operation and
generate flags used by jz branch instruction. Wrong test condition
leads to incorrect branching
- Incorrect compensation code for some branches

Signed-off-by: Evgeny Pavlov <lucenticus@gmail.com>
2023-08-21 17:04:51 +02:00
Chris Spencer
f0b1cab538 hwcontext_vulkan: always use create_pnext in vulkan_pool_alloc
Currently, create_pnext is only used if an applicable external memory
extension is enabled. This will usually the case when used from the command
line, but may not be when the Vulkan context is created manually.

For images used in video decoding, create_pnext contains the video profile
list, which is mandatory.[1] This fixes a GPU crash when using RADV.

[1] https://registry.khronos.org/vulkan/specs/1.3-extensions/man/html/VkImageCreateInfo.html#VUID-VkImageCreateInfo-usage-04815

Signed-off-by: Chris Spencer <spencercw@gmail.com>
2023-08-20 22:47:09 +02:00