fe171a3b51
lavc/hevcdec: move calling hwaccel decode_slice to decode_slice_data()
...
From decode_nal_unit(), as that is a more appropriate place for it.
2024-06-11 17:39:35 +02:00
6ee550d83d
lavc/hevcdec: move calling hwaccel start_frame to hevc_frame_start()
...
From decode_nal_unit(), as that is a more appropriate place for it.
2024-06-11 17:39:35 +02:00
3bbb5d78c7
lavc/hevcdec: move per-slice local_ctx setup out of hls_slice_header()
...
Into decode_slice_data(). This is a step towards constifying
HEVCContext in hls_slice_header().
2024-06-11 17:39:35 +02:00
efc827bf6f
lavc/hevcdec: move slice decoding dispatch to its own function
...
Also move there a sanity check from hls_decode_entry() that should also
be performed when WPP is active (note that the check is not moved to
hls_slice_header() because it requires the HEVCContext.tab_slice_address
to be set up).
2024-06-11 17:39:35 +02:00
7cce612a26
lavc/hevcdec: move a slice segment sanity check to hls_slice_header()
...
Combine it with an existing similar check.
2024-06-11 17:39:35 +02:00
d43527a1a0
lavc/hevcdec: store slice header POC in SliceHeader
...
Rather than decoding directly into HEVCContext.poc.
This is a step towards constifying HEVCContext in hls_slice_header().
2024-06-11 17:39:35 +02:00
e4e9e1da15
lavc/hevcdec: drop redundant HEVCContext.threads_{type,number}
...
They are useless duplicates of corresponding AVCodecContext fields.
2024-06-11 17:39:35 +02:00
b0c29a45dc
lavc/hevc/cabac: do not infer WPP use based on HEVCContext.threads_number
...
Pass this information explicitly instead.
2024-06-11 17:39:35 +02:00
d86ac94df2
lavc/hevcdec: output RASL frames based on the value of no_rasl_output_flag
...
Instead of an ad-hoc scheme. Also, combine skipping RASL frames with
skip_frame handling - current code seems flawed as it only executes for
the first slice of a RASL frame and unnecessarily unsets is_decoded,
which should not be set at this point anyway..
Some RASL frames in fate-hevc-afd-tc-sei that were previously discarded
are now output.
2024-06-11 17:39:35 +02:00
3115c84015
lavc/hevcdec: only set no_rasl_output_flag for IRAP frames
...
Its meaning is only specified for IRAP frames.
As it's currently never used otherwise, this should not change decoder
behaviour, but will be useful in future commits.
2024-06-11 17:39:35 +02:00
381b70e173
lavc/hevcdec: do not pass HEVCContext to ff_hevc_frame_nb_refs()
...
Pass the only things required from it - slice header and PPS -
explicitly.
Will be useful in the following commits to avoid mofiying HEVCContext in
hls_slice_header().
2024-06-11 17:39:35 +02:00
07eb60c0da
lavc/hevcdec: only call export_stream_params_from_sei() once per frame
...
Not once per each slice header, as it makes no sense and may cause races
with frame threading.
2024-06-11 17:39:35 +02:00
01b379a93e
lavc/hevcdec: move pocTid0 computation to hevc_frame_start()
...
It is only done once per frame. Also, rename the variable to poc_tid0 to
be consistent with our naming conventions.
2024-06-11 17:39:35 +02:00
5e438511ab
lavc/hevcdec: do not pass HEVCContext to decode_lt_rps()
...
Pass the two numbers needed from it explicitly.
Makes it clear that HEVCContext is not modified by this function.
2024-06-11 17:39:35 +02:00
0892ec947c
lavc/hevcdec: pass SliceHeader explicitly to pred_weight_table()
...
And replace the HEVCContext* parameter by void *logctx.
Makes it clear that only SliceHeader is modified by this function.
2024-06-11 17:39:35 +02:00
90fc331b0f
lavc/hevcdec: only ignore INVALIDDATA in decode_nal_unit()
...
All other errors should cause a failure, regardless of the value of
err_recognition. Also, print a warning message when skipping invalid NAL
units.
2024-06-11 17:39:35 +02:00
8eb134f4f9
lavc/hevcdec: drop an always-zero variable
2024-06-11 17:39:35 +02:00
8c8072c29c
lavc/hevcdec: move active PPS from HEVCParamSets to HEVCContext
...
"Currently active PPS" is a property of the decoding process, not of the
list of available parameter sets.
2024-06-11 17:39:34 +02:00
0f47342c12
lavc/hevcdec: stop accessing parameter sets through HEVCParamSets
...
Instead, accept PPS/SPS as function arguments.
Makes the code shorter and significantly reduces diff in future commits.
2024-06-11 17:39:34 +02:00
38b8ae4112
lavc/hevc/pred: stop accessing parameter sets through HEVCParamSets
...
Instead, accept PPS/SPS as function arguments.
Makes the code shorter and significantly reduces diff in future commits.
2024-06-11 17:39:34 +02:00
d0868d70ea
lavc/hevc/cabac: stop accessing parameter sets through HEVCParamSets
...
Instead, accept PPS/SPS as function arguments.
Makes the code shorter and significantly reduces diff in future commits.
2024-06-11 17:39:34 +02:00
b38aecffec
lavc/hevc/filter: stop accessing parameter sets through HEVCParamSets
...
Instead, accept PPS as a function argument and retrieve SPS through it.
Makes the code shorter and significantly reduces diff in future commits.
2024-06-11 17:39:34 +02:00
fb873a05b3
lavc/hevc/mvs: stop accessing parameter sets through HEVCParamSets
...
Instead, accept PPS as a function argument and retrieve SPS through it.
Makes the code shorter and significantly reduces diff in future commits.
2024-06-11 17:39:34 +02:00
6ddba110eb
lavc/hevc/parser: stop using HEVCParamSets.[psv]ps
...
The parser does not need to preserve these between frames.
2024-06-11 17:39:34 +02:00
2e46d68f55
lavc/hevc_ps: make SPS hold a reference to its VPS
...
SPS and its dependent PPSes depend on, and are parsed for, specific VPS data.
This will be useful in following commits.
2024-06-11 17:39:34 +02:00
c879165b39
lavc/hevc_ps: make PPS hold a reference to its SPS
...
PPS depends on, and is parsed for, specific SPS data.
This will be useful in following commits.
2024-06-11 17:39:34 +02:00
e12fd62d1d
lavc/hevcdec: drop a redundant assignment in hevc_decode_frame()
...
The exact same code is executed at the beginning of decode_nal_units()
2024-06-11 17:39:34 +02:00
a82f2b0924
lavc/hevcdec: simplify condition
2024-06-11 17:39:34 +02:00
0407556716
lavc/hevcdec: do not free SliceHeader arrays in pic_arrays_free()
...
SliceHeader.{entry_point_offset,size,offset} are not derived from frame
size and do not need to be freed here.
2024-06-11 17:39:34 +02:00
0455a62d84
lavf/tls_mbedtls: handle session ticket error code as no-op
...
When TLSv1.3 and session tickets are enabled mbedtls_ssl_read()
will return an error code to inform about a received session ticket.
This can simply be handled like EAGAIN instead of errornously
aborting the connection.
ref: https://github.com/Mbed-TLS/mbedtls/issues/8749
Signed-off-by: Anton Khirnov <anton@khirnov.net >
2024-06-11 17:00:35 +02:00
1b1e9cadc5
lavf/tls_mbedtls: fix handling of certification validation failures
...
We manually check the verification status after the handshake has completed
using mbedtls_ssl_get_verify_result(). However with VERIFY_REQUIRED
mbedtls_ssl_handshake() already returns an error, so this code is never reached.
Fix that by using VERIFY_OPTIONAL, which performs the verification but
does not abort the handshake.
Signed-off-by: Anton Khirnov <anton@khirnov.net >
2024-06-11 16:58:22 +02:00
827578ca76
lavf/tls_mbedtls: hook up debug message callback
...
Unfortunately this won't work out-of-the-box because mbedTLS
only provides a global (not per-context) debug toggle.
Signed-off-by: Anton Khirnov <anton@khirnov.net >
2024-06-11 16:58:15 +02:00
807d1505bf
lavf/tls_mbedtls: add missing call to psa_crypto_init
...
This is mandatory depending on configuration or at least with mbedTLS 3.6.0.
Signed-off-by: Anton Khirnov <anton@khirnov.net >
2024-06-11 16:35:46 +02:00
63b6620ad3
lavf/tls_mbedtls: handle more error codes for human-readable messages
...
Signed-off-by: Anton Khirnov <anton@khirnov.net >
2024-06-11 16:35:31 +02:00
b6f37ffba7
lavc/vc1dsp: match C block layout in inv_trans_4x8_rvv
...
Although checkasm does not verify this, the decoder requires that the
transform updates the input block exactly like the C code does.
This fixes vc1-ism, vc1_ilaced_twomv, vc1_sa00040, vc1_sa10091,
vc1_sa10143, vc1_sa20021, vc1test_smm0005 and wmv3-drm-dec tests.
2024-06-11 17:15:09 +03:00
6c05069e68
lavc/vc1dsp: match C block layout in inv_trans_4x4_rvv
...
Although checkasm does not verify this, the decoder requires that the
transform updates the input block exactly like the C code does.
This fixes vc1-ism, vc1_ilaced_twomv, vc1_sa00040, vc1_sa10091,
vc1_sa10143, vc1_sa20021, vc1test_smm0005 and wmv3-drm-dec tests.
2024-06-11 17:15:09 +03:00
6ae1a337f2
fftools/ffmpeg_mux_init: Fix leak when using non-encoding option
...
Reviewed-by: Anton Khirnov <anton@khirnov.net >
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com >
2024-06-11 14:32:25 +02:00
8754c9bd82
configure: Disable DNN without backend
...
The DNN filters are useless without a backend.
This will also "fix" Coverity issues #1598288 and #1601718 .
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com >
2024-06-11 19:36:54 +08:00
c84e40d9e6
fftools/ffmpeg_mux_init: Return error upon error
...
Currently it may return an uninitialized value.
Introduced in 840f2bc18e#1603565 .
Reviewed-by: Anton Khirnov <anton@khirnov.net >
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com >
2024-06-11 08:16:42 +02:00
a0ff31e740
avcodec/vvc/inter: Don't return void
...
Returning a void is not allowed by the spec. Just return instead.
Reviewed-by: Nuo Mi <nuomi2021@gmail.com >
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com >
2024-06-11 02:43:14 +02:00
417957ec5e
sws/range_convert: R-V V to/from JPEG
...
C908 X60
chrRangeFromJpeg_8_c: 2.7 2.5
chrRangeFromJpeg_8_rvv_i32: 1.7 1.5
chrRangeFromJpeg_24_c: 7.5 6.7
chrRangeFromJpeg_24_rvv_i32: 1.7 1.5
chrRangeFromJpeg_128_c: 55.2 34.7
chrRangeFromJpeg_128_rvv_i32: 6.5 3.0
chrRangeFromJpeg_144_c: 44.0 39.2
chrRangeFromJpeg_144_rvv_i32: 7.7 4.5
chrRangeFromJpeg_256_c: 78.2 69.5
chrRangeFromJpeg_256_rvv_i32: 12.2 6.0
chrRangeFromJpeg_512_c: 172.2 138.5
chrRangeFromJpeg_512_rvv_i32: 24.5 11.7
chrRangeToJpeg_8_c: 4.7 4.2
chrRangeToJpeg_8_rvv_i32: 2.0 1.7
chrRangeToJpeg_24_c: 13.7 12.2
chrRangeToJpeg_24_rvv_i32: 2.0 1.5
chrRangeToJpeg_128_c: 72.0 63.7
chrRangeToJpeg_128_rvv_i32: 6.7 3.2
chrRangeToJpeg_144_c: 80.7 71.7
chrRangeToJpeg_144_rvv_i32: 8.5 4.7
chrRangeToJpeg_256_c: 143.2 127.2
chrRangeToJpeg_256_rvv_i32: 13.5 6.5
chrRangeToJpeg_512_c: 285.7 253.7
chrRangeToJpeg_512_rvv_i32: 27.0 13.0
lumRangeFromJpeg_8_c: 1.7 1.5
lumRangeFromJpeg_8_rvv_i32: 1.2 1.0
lumRangeFromJpeg_24_c: 4.2 3.7
lumRangeFromJpeg_24_rvv_i32: 1.2 1.0
lumRangeFromJpeg_128_c: 21.7 19.2
lumRangeFromJpeg_128_rvv_i32: 3.7 1.7
lumRangeFromJpeg_144_c: 24.7 22.0
lumRangeFromJpeg_144_rvv_i32: 4.7 2.7
lumRangeFromJpeg_256_c: 43.7 39.0
lumRangeFromJpeg_256_rvv_i32: 7.5 3.2
lumRangeFromJpeg_512_c: 87.0 77.2
lumRangeFromJpeg_512_rvv_i32: 14.5 6.7
lumRangeToJpeg_8_c: 2.7 2.2
lumRangeToJpeg_8_rvv_i32: 1.0 1.0
lumRangeToJpeg_24_c: 7.2 6.5
lumRangeToJpeg_24_rvv_i32: 1.2 1.0
lumRangeToJpeg_128_c: 37.7 33.7
lumRangeToJpeg_128_rvv_i32: 3.7 2.0
lumRangeToJpeg_144_c: 42.5 37.7
lumRangeToJpeg_144_rvv_i32: 4.7 2.7
lumRangeToJpeg_256_c: 75.0 66.7
lumRangeToJpeg_256_rvv_i32: 7.5 3.5
lumRangeToJpeg_512_c: 149.5 133.0
lumRangeToJpeg_512_rvv_i32: 14.7 7.0
2024-06-10 22:48:52 +03:00
9dac8495b0
swscale/aarch64: Add rgb24 to yuv implementation
...
Test on Apple M1:
rgb24_to_uv_8_c: 0.0
rgb24_to_uv_8_neon: 0.2
rgb24_to_uv_128_c: 1.0
rgb24_to_uv_128_neon: 0.5
rgb24_to_uv_1080_c: 7.0
rgb24_to_uv_1080_neon: 5.7
rgb24_to_uv_1920_c: 12.5
rgb24_to_uv_1920_neon: 9.5
rgb24_to_uv_half_8_c: 0.2
rgb24_to_uv_half_8_neon: 0.2
rgb24_to_uv_half_128_c: 1.0
rgb24_to_uv_half_128_neon: 0.5
rgb24_to_uv_half_1080_c: 6.2
rgb24_to_uv_half_1080_neon: 3.0
rgb24_to_uv_half_1920_c: 11.2
rgb24_to_uv_half_1920_neon: 5.2
rgb24_to_y_8_c: 0.2
rgb24_to_y_8_neon: 0.0
rgb24_to_y_128_c: 0.5
rgb24_to_y_128_neon: 0.5
rgb24_to_y_1080_c: 4.7
rgb24_to_y_1080_neon: 3.2
rgb24_to_y_1920_c: 8.0
rgb24_to_y_1920_neon: 5.7
On Pixel 6:
rgb24_to_uv_8_c: 30.7
rgb24_to_uv_8_neon: 56.9
rgb24_to_uv_128_c: 213.9
rgb24_to_uv_128_neon: 173.2
rgb24_to_uv_1080_c: 1649.9
rgb24_to_uv_1080_neon: 1424.4
rgb24_to_uv_1920_c: 2907.9
rgb24_to_uv_1920_neon: 2480.7
rgb24_to_uv_half_8_c: 36.2
rgb24_to_uv_half_8_neon: 33.4
rgb24_to_uv_half_128_c: 167.9
rgb24_to_uv_half_128_neon: 99.4
rgb24_to_uv_half_1080_c: 1293.9
rgb24_to_uv_half_1080_neon: 778.7
rgb24_to_uv_half_1920_c: 2292.7
rgb24_to_uv_half_1920_neon: 1328.7
rgb24_to_y_8_c: 19.7
rgb24_to_y_8_neon: 27.7
rgb24_to_y_128_c: 129.9
rgb24_to_y_128_neon: 96.7
rgb24_to_y_1080_c: 995.4
rgb24_to_y_1080_neon: 767.7
rgb24_to_y_1920_c: 1747.4
rgb24_to_y_1920_neon: 1337.2
Note both tests use clang as compiler, which has vectorization
enabled by default with -O3.
Reviewed-by: Rémi Denis-Courmont <remi@remlab.net >
Reviewed-by: Martin Storsjö <martin@martin.st >
Signed-off-by: Zhao Zhili <zhilizhao@tencent.com >
2024-06-11 01:12:09 +08:00
b1240c983f
tests/checkasm: Fix build error when enable linux perf on Android
...
B0 is defined by system header, see f0f596dbc6martin@martin.st >
Signed-off-by: Zhao Zhili <zhilizhao@tencent.com >
2024-06-11 01:11:46 +08:00
33e4cc963d
avutil/timer: Add clock_gettime as a fallback of AV_READ_TIME
...
Reviewed-by: Rémi Denis-Courmont <remi@remlab.net >
Reviewed-by: Martin Storsjö <martin@martin.st >
Signed-off-by: Zhao Zhili <zhilizhao@tencent.com >
2024-06-11 01:11:36 +08:00
6a18c0bc87
avutil/aarch64: Skip define AV_READ_TIME for apple
...
It will fallback to mach_absolute_time inside libavutil/timer.h
Reviewed-by: Martin Storsjö <martin@martin.st >
Signed-off-by: Zhao Zhili <zhilizhao@tencent.com >
2024-06-11 01:10:42 +08:00
94f2274a8b
x86/aacencdsp: fix ff_aac_quantize_bands_avx on unix64 ABI
...
Signed-off-by: James Almer <jamrial@gmail.com >
2024-06-09 17:16:02 -03:00
17c3cc5bb6
swscale/x86/rgb_2_rgb: add missing wrap to ff_uyvytoyuv422_avx2
...
Fixes old yasm.
Signed-off-by: James Almer <jamrial@gmail.com >
2024-06-09 16:04:36 -03:00
03546f49a3
swscale/x86/rgb2rgb: add missing wrap for ff_uyvytoyuv422_avx2
...
Signed-off-by: James Almer <jamrial@gmail.com >
2024-06-09 15:56:52 -03:00
287d139b77
checkasm/sw_rgb: fix alignment of buffers for rgb_to_yuv tests
...
src is apparently not guaranteed to be >8 byte aligned, but align to 16
nonetheless as the x86 asm will do unaligned loads anyway.
dst is guaranteed to be 32 byte aligned for the Y plane, but 16 byte for UV.
Signed-off-by: James Almer <jamrial@gmail.com >
2024-06-09 14:12:51 -03:00
e8cef5e152
swscale/x86/rgb2rgb: remove mmxext version of shuffle_bytes_2103
...
Signed-off-by: James Almer <jamrial@gmail.com >
2024-06-09 13:43:11 -03:00
