Re: [FFmpeg-devel] [PATCH] avcodec/aarch64/vvc: optimised alf_classify function 8/10/12bit of vvc codec for aarch64

2025-08-01 Thread Martin Storsjö
On Wed, 9 Jul 2025, Georgii Zagoruiko wrote: diff --git a/libavcodec/aarch64/vvc/alf.S b/libavcodec/aarch64/vvc/alf.S index 8801b3afb6..9c9765ead1 100644 --- a/libavcodec/aarch64/vvc/alf.S +++ b/libavcodec/aarch64/vvc/alf.S @@ -291,3 +291,281 @@ function ff_alf_filter_chroma_kernel_10_neon, expo

Re: [FFmpeg-devel] [FFmpeg-cvslog] fate/lavf-container: add test for APV in MP4

2025-07-19 Thread Martin Storsjö
On Fri, 18 Jul 2025, James Almer wrote: ffmpeg | branch: master | James Almer | Tue Jul 15 19:00:21 2025 -0300| [3cd5672bfeb796342d24228a78f2e733c0a40e7d] | committer: James Almer fate/lavf-container: add test for APV in MP4 Signed-off-by: James Almer http://git.videolan.org/gitweb.cgi/ff

Re: [FFmpeg-devel] [POLL] [VOTE] code.ffmpeg.org

2025-07-15 Thread Martin Storsjö
On Sun, 13 Jul 2025, Michael Niedermayer wrote: Hi all Do people want Forgejo or Gitlab on code.ffmpeg.org for testing? F. code.ffmpeg.org should run Forgejo G. code.ffmpeg.org should run Gitlab No strong opinion between the two. I have a lot of experience with Gitlab (which I find quite wo

Re: [FFmpeg-devel] [PATCH v5] gcc: Don't disable '-ftree-vectorize' if gcc version higher than 13.

2025-07-10 Thread Martin Storsjö
On Thu, 10 Jul 2025, Jiawei wrote: In addition, can you provide a bug record or a reproduction method? I think it should also be fixed in the GCC community, and I will report the problems you encounter in test to GCC Bugzilla. https://patchwork.ffmpeg.org/check/125495/ If you build ffmpeg wi

Re: [FFmpeg-devel] [PATCH v5] gcc: Don't disable '-ftree-vectorize' if gcc version higher than 13.

2025-07-10 Thread Martin Storsjö
On Thu, 10 Jul 2025, Jiawei wrote: Hi Martin, Is there any progress for this patch, should I resend it to the mailing list again? When I posted the updated version of the patch, the patchwork test runners noticed that this patch causes test failures on Loongarch. So modern versions of GCC

Re: [FFmpeg-devel] [RFC PATCH] avformat: remove HLS protocol

2025-07-08 Thread Martin Storsjö
On Tue, 8 Jul 2025, Marvin Scholz wrote: The use of this protocol was already discouraged and warned about for years with the recommendation to use the HLS demuxer instead. --- doc/protocols.texi | 20 --- libavformat/Makefile| 2 - libavformat/hlsproto.c | 318 ---

Re: [FFmpeg-devel] [FATE PATCH] Properly list test errors when there are no failed tests

2025-06-28 Thread Martin Storsjö
On Sat, 28 Jun 2025, Michael Niedermayer wrote: Its a few years since the last change to fateserver, i think we dont have a script to update the checkout on the server from a git push. Its a while, i dont remember exactly :) so you have to update that checkout after pushing possibly I don't th

Re: [FFmpeg-devel] [FATE PATCH] Properly list test errors when there are no failed tests

2025-06-27 Thread Martin Storsjö
On Fri, 27 Jun 2025, Kacper Michajlow wrote: On Fri, Jun 27, 2025, 14:44 Martin Storsjö wrote: On Wed, 18 Jun 2025, Michael Niedermayer wrote: > Hi Martin > > On Fri, Jun 13, 2025 at 01:57:05PM +0300, Martin Storsjö wrote: >> If there were failures while running tests, e.g

Re: [FFmpeg-devel] [FATE PATCH] Properly list test errors when there are no failed tests

2025-06-27 Thread Martin Storsjö
On Wed, 18 Jun 2025, Michael Niedermayer wrote: Hi Martin On Fri, Jun 13, 2025 at 01:57:05PM +0300, Martin Storsjö wrote: If there were failures while running tests, e.g. if failing to compile checkasm or any other of the test programs, there are no failed tests per se, and the number of

Re: [FFmpeg-devel] [PATCH] configure: Make MSVC version grabbing more robust

2025-06-21 Thread Martin Storsjö
: Re: [FFmpeg-devel] [PATCH] configure: Make MSVC version grabbing more robust Hi Martin! On 2025-06-21 13:37 +0300, Martin Storsjö wrote: > > On 21. Jun 2025, at 12.20, Alexander Strasser via ffmpeg-devel wrote: > > > > > > On 2025-06-21 00:03 +0200, Kacper Michajlow wr

Re: [FFmpeg-devel] [PATCH] configure: Make MSVC version grabbing more robust

2025-06-21 Thread Martin Storsjö
On Sat, 21 Jun 2025, Kacper Michajlow wrote: On Sat, 21 Jun 2025 at 12:34, Martin Storsjö wrote: On Sat, 21 Jun 2025, Kacper Michajlow wrote: > On Fri, 20 Jun 2025 at 22:26, Hendrik Leppkes > wrote: >> >> On Fri, Jun 20, 2025 at 9:25 PM Timo Rothenpieler wrote: >>

Re: [FFmpeg-devel] [PATCH] configure: Make MSVC version grabbing more robust

2025-06-21 Thread Martin Storsjö
On Sat, 21 Jun 2025, Kacper Michajlow wrote: On Fri, 20 Jun 2025 at 22:26, Hendrik Leppkes wrote: On Fri, Jun 20, 2025 at 9:25 PM Timo Rothenpieler wrote: > > On 19.06.2025 22:21, Martin Storsjö wrote: > > On Fri, 13 Jun 2025, Martin Storsjö wrote: > > > >> When

Re: [FFmpeg-devel] [PATCH] configure: Make MSVC version grabbing more robust

2025-06-21 Thread Martin Storsjö
> On 21. Jun 2025, at 12.20, Alexander Strasser via ffmpeg-devel > wrote: > > > On 2025-06-21 00:03 +0200, Kacper Michajlow wrote: >> On Fri, 20 Jun 2025 at 22:26, Hendrik Leppkes >> wrote: >>> >>> On Fri, Jun 20, 2025 at 9:25 PM Timo Rothenpieler >>> wrote: Likely this patch bro

Re: [FFmpeg-devel] [PATCH] configure: Make MSVC version grabbing more robust

2025-06-19 Thread Martin Storsjö
On Fri, 13 Jun 2025, Martin Storsjö wrote: When running plain "cl", to get the MSVC version, it prints the version header on stderr, while the usage instructions are printed on stdout. Usually, the version on stderr gets flushed first, so "head -n1" gets the line it expects

Re: [FFmpeg-devel] [PATCH] swscale/aarch64/output: Implement neon assembly for yuv2planeX_10_c_template()

2025-06-18 Thread Martin Storsjö
On Fri, 6 Jun 2025, Logaprakash Ramajayam wrote: Checked FATE tests and gha-aarch64 git workflow. From 34cdef26eaebcf98916e9881b3a04f4f698f09c6 Mon Sep 17 00:00:00 2001 From: Logaprakash Ramajayam Date: Thu, 5 Jun 2025 01:33:39 -0700 Subject: [PATCH] swscale/aarch64/output: Implement neon asse

Re: [FFmpeg-devel] [PATCH] swscale/aarch64/output: Implement neon assembly for yuv2nv12cX_c() [v2]

2025-06-18 Thread Martin Storsjö
On Mon, 9 Jun 2025, Harshitha Sarangu Suresh wrote: From 4ca5eae1e7164f78296719f19aef97239e5b046a Mon Sep 17 00:00:00 2001 From: Harshitha Suresh Date: Mon, 19 May 2025 22:37:20 +0530 Subject: [PATCH] swscale/aarch64/output: Implement neon assembly for yuv2nv12cX_c(). yuv2nv12cX_2_512_accurat

Re: [FFmpeg-devel] [PATCH] configure: fix Microsoft tools detection

2025-06-17 Thread Martin Storsjö
On Mon, 16 Jun 2025, Kacper Michajlow wrote: On Mon, 16 Jun 2025 at 09:40, Martin Storsjö wrote: On Sat, 14 Jun 2025, Kacper Michajlow wrote: > On Sat, 14 Jun 2025 at 03:03, Kacper Michajłow wrote: >> >> LLVM tools print installation path upon execution. If one uses LLVM &

Re: [FFmpeg-devel] [FFmpeg-cvslog] hwcontext_vulkan: use host image copy

2025-06-16 Thread Martin Storsjö
On Wed, 11 Jun 2025, Lynne wrote: ffmpeg | branch: master | Lynne | Mon May 12 18:42:09 2025 +0200| [a9b2c10eee9cf28ecbce2f1972564f4aa826a855] | committer: Lynne hwcontext_vulkan: use host image copy http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=a9b2c10eee9cf28ecbce2f1972564f4aa

Re: [FFmpeg-devel] [PATCH v6] gcc: Don't disable '-ftree-vectorize' if gcc version is 13 or newer

2025-06-16 Thread Martin Storsjö
On Mon, 16 Jun 2025, Martin Storsjö wrote: From: Jiawei This patch modifies the FFmpeg build system to allow GCC to use the `-ftree-vectorize` flag when the compiler version is 13 or newer. Enabling this flag can improve performance through better loop analysis and auto-vectorization (SIMD

[FFmpeg-devel] [PATCH v6] gcc: Don't disable '-ftree-vectorize' if gcc version is 13 or newer

2025-06-16 Thread Martin Storsjö
From: Jiawei This patch modifies the FFmpeg build system to allow GCC to use the `-ftree-vectorize` flag when the compiler version is 13 or newer. Enabling this flag can improve performance through better loop analysis and auto-vectorization (SIMD) opportunities in modern GCC versions. The expli

Re: [FFmpeg-devel] [PATCH] configure: fix Microsoft tools detection

2025-06-16 Thread Martin Storsjö
On Sat, 14 Jun 2025, Kacper Michajlow wrote: On Sat, 14 Jun 2025 at 03:03, Kacper Michajłow wrote: LLVM tools print installation path upon execution. If one uses LLVM tools bundled with Microsoft Visual Studio installation, they would be incorrectly detected as Microsoft's ones. Microsoft to

Re: [FFmpeg-devel] [PATCH v5] gcc: Don't disable '-ftree-vectorize' if gcc version higher than 13.

2025-06-16 Thread Martin Storsjö
On Sun, 15 Jun 2025, Michael Niedermayer wrote: Hi Martin On Fri, Jun 13, 2025 at 11:42:11AM +0300, Martin Storsjö wrote: On Fri, 13 Jun 2025, Jiawei wrote: This patch modifies the FFmpeg build system to allow GCC to use the `-ftree-vectorize` flag when the compiler version is 13 or newer

Re: [FFmpeg-devel] [PATCH] checkasm/h264dsp: Fix stack overflow in check_idct_dequant

2025-06-16 Thread Martin Storsjö
On Mon, 16 Jun 2025, Zhao Zhili wrote: From: Zhao Zhili --- tests/checkasm/h264dsp.c | 14 ++ 1 file changed, 10 insertions(+), 4 deletions(-) diff --git a/tests/checkasm/h264dsp.c b/tests/checkasm/h264dsp.c index f5f9650224..006532e08b 100644 --- a/tests/checkasm/h264dsp.c +++ b/t

Re: [FFmpeg-devel] [PATCH 2/2] checkasm: h264dsp: test luma_dc_dequant

2025-06-13 Thread Martin Storsjö
On Fri, 13 Jun 2025, Tristan Matthews wrote: --- tests/checkasm/h264dsp.c | 37 + 1 file changed, 37 insertions(+) Both patches LGTM, thank you! // Martin ___ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org http

[FFmpeg-devel] [PATCH] configure: Make MSVC version grabbing more robust

2025-06-13 Thread Martin Storsjö
When running plain "cl", to get the MSVC version, it prints the version header on stderr, while the usage instructions are printed on stdout. Usually, the version on stderr gets flushed first, so "head -n1" gets the line it expects, but some times (in particular when running MSVC wrapped in wine),

Re: [FFmpeg-devel] [PATCH] checkasm: h264dsp: test luma_dc_dequant

2025-06-13 Thread Martin Storsjö
On Fri, 13 Jun 2025, Tristan Matthews wrote: On Fri, Jun 13, 2025 at 2:08 AM Martin Storsjö wrote: On Fri, 13 Jun 2025, Tristan Matthews wrote: Good catch, also I realized that the output buffers were too small, will be fixed in the next version. Why was that too small? If we write (and

[FFmpeg-devel] [FATE PATCH] Properly list test errors when there are no failed tests

2025-06-13 Thread Martin Storsjö
If there were failures while running tests, e.g. if failing to compile checkasm or any other of the test programs, there are no failed tests per se, and the number of succeessful tests is equal to the total number of tests. For these cases, check the job status code instead of declaring them as a

Re: [FFmpeg-devel] [PATCH v5] gcc: Don't disable '-ftree-vectorize' if gcc version higher than 13.

2025-06-13 Thread Martin Storsjö
On Fri, 13 Jun 2025, Jiawei wrote: This patch modifies the FFmpeg build system to allow GCC to use the `-ftree-vectorize` flag when the compiler version is 13 or newer. Enabling this flag can improve performance through better loop analysis and auto-vectorization (SIMD) opportunities in modern G

Re: [FFmpeg-devel] [PATCH] avutil: Fix linking x86 asm constants with Clang in MSVC mode

2025-06-13 Thread Martin Storsjö
On Thu, 12 Jun 2025, Martin Storsjö wrote: This fixes building with Clang in MSVC mode, for x86, which was broken in 6e49b8699657b808b7dc80033f2c3f2d0e029fa3 (in Nov 2024); previously it failed with undefined symbols for the constants defined with DECLARE_ASM_CONST, accessed via inline assembly

Re: [FFmpeg-devel] [PATCH] avformat/movenc: Fix editlist with hybrid_fragmented

2025-06-13 Thread Martin Storsjö
On Fri, 13 Jun 2025, Zhao Zhili wrote: From: Zhao Zhili The segment_duration must not be set to zero when writing the moov atom for the second time. This is related to edit lists in standard MP4 files. --- libavformat/movenc.c | 9 +++-- 1 file changed, 7 insertions(+), 2 deletions(-) diff

Re: [FFmpeg-devel] [PATCH] checkasm: h264dsp: test luma_dc_dequant

2025-06-12 Thread Martin Storsjö
On Fri, 13 Jun 2025, Tristan Matthews wrote: On Thu, Jun 12, 2025 at 4:14 PM Martin Storsjö wrote: On Thu, 12 Jun 2025, Tristan Matthews wrote: --- tests/checkasm/h264dsp.c | 37 + 1 file changed, 37 insertions(+) diff --git a/tests/checkasm/h264dsp.c b

Re: [FFmpeg-devel] [PATCH] checkasm: h264dsp: test luma_dc_dequant

2025-06-12 Thread Martin Storsjö
On Thu, 12 Jun 2025, Tristan Matthews wrote: --- tests/checkasm/h264dsp.c | 37 + 1 file changed, 37 insertions(+) diff --git a/tests/checkasm/h264dsp.c b/tests/checkasm/h264dsp.c index d1228ed985..5fba31cf69 100644 --- a/tests/checkasm/h264dsp.c +++ b/tests/c

Re: [FFmpeg-devel] [PATCH] avformat/movenc: editlist doesn't work with hybrid_fragmented

2025-06-12 Thread Martin Storsjö
On Thu, 12 Jun 2025, Zhao Zhili wrote: From: Zhao Zhili --- libavformat/movenc.c | 5 + 1 file changed, 5 insertions(+) diff --git a/libavformat/movenc.c b/libavformat/movenc.c index cd5b45f6fe..93af0723db 100644 --- a/libavformat/movenc.c +++ b/libavformat/movenc.c @@ -7781,6 +7781,11 @@

Re: [FFmpeg-devel] [PATCH v4] gcc: Don't disable '-ftree-vectorize' if gcc version higher than 13.

2025-06-12 Thread Martin Storsjö
On Thu, 12 Jun 2025, Jiawei wrote: This patch modifies the FFmpeg build system to allow GCC to use the `-ftree-vectorize` flag when the compiler version is 13 or newer. Enabling this flag can improve performance through better loop analysis and auto-vectorization (SIMD) opportunities in modern G

Re: [FFmpeg-devel] [PATCH v3] gcc: Don't disable '-ftree-vectorize' if gcc version higher than 13.

2025-06-12 Thread Martin Storsjö
On Thu, 12 Jun 2025, Jiawei wrote: This patch modifies the FFmpeg build system to allow gcc use '-ftree-vectorize' option in newer releases. Modern GCC versions can be benefit when using the `-ftree-vectorize` flag. Through it can optimizations in loop analysis and SIMD code generation. Versio

Re: [FFmpeg-devel] [PATCH v3] gcc: Don't disable '-ftree-vectorize' if gcc version higher than 13.

2025-06-12 Thread Martin Storsjö
On Thu, 12 Jun 2025, Martin Storsjö wrote: On Thu, 12 Jun 2025, Jiawei wrote: This patch modifies the FFmpeg build system to allow gcc use '-ftree-vectorize' option in newer releases. Modern GCC versions can be benefit when using the `-ftree-vectorize` flag. Through it can optimi

Re: [FFmpeg-devel] [PATCH 2/2] avformat/internal: add missing __clang__ check

2025-06-12 Thread Martin Storsjö
On Mon, 31 Mar 2025, Kacper Michajłow wrote: Clang x86_64-pc-windows-msvc doesn't define __GNUC__. Signed-off-by: Kacper Michajłow --- libavformat/internal.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/libavformat/internal.h b/libavformat/internal.h index fe428d85eb..bf8

Re: [FFmpeg-devel] [PATCH 1/2] avutil/intmath: use AV_HAS_BUILTIN to detect builtin availability

2025-06-12 Thread Martin Storsjö
On Mon, 31 Mar 2025, Kacper Michajłow wrote: Fixes use of bultins on clang x86_64-pc-windows-msvc which does not define any __GNUC__. Also on other targets __GNUC__ is defined to 4 by default, so any feature testing based on version is not really valid. Signed-off-by: Kacper Michajłow --- liba

Re: [FFmpeg-devel] [FFmpeg-devel, v2] gcc: Relaxing auto-vectorization limitation.

2025-06-12 Thread Martin Storsjö
On Thu, 29 May 2025, Zhao Zhili wrote: On May 29, 2025, at 15:03, Jiawei wrote: This patch modifies the FFmpeg build system to remove the explicit disabling of GCC's auto-vectorization feature. Modern GCC versions have demonstrated stable auto-vectorization capabilities through extensive opti

[FFmpeg-devel] [PATCH] avutil: Fix linking x86 asm constants with Clang in MSVC mode

2025-06-12 Thread Martin Storsjö
This fixes building with Clang in MSVC mode, for x86, which was broken in 6e49b8699657b808b7dc80033f2c3f2d0e029fa3 (in Nov 2024); previously it failed with undefined symbols for the constants defined with DECLARE_ASM_CONST, accessed via inline assembly. Before 57861911a34e1c33796be97f2b2f44e05fffd

Re: [FFmpeg-devel] [PATCH 1/2] avutil/intmath: use AV_HAS_BUILTIN to detect builtin availability

2025-06-11 Thread Martin Storsjö
On Wed, 11 Jun 2025, Kacper Michajlow wrote: On Mon, 31 Mar 2025 at 14:03, Kacper Michajłow wrote: Fixes use of bultins on clang x86_64-pc-windows-msvc which does not define any __GNUC__. Also on other targets __GNUC__ is defined to 4 by default, so any feature testing based on version is not

Re: [FFmpeg-devel] [PATCH v2] avformat/tls_openssl: fix build error when openssl version < 3

2025-06-06 Thread Martin Storsjö
On Fri, 6 Jun 2025, Jack Lau via ffmpeg-devel wrote: add the missing data structure pkey in the tls_context properly set this pkey and free it Signed-off-by: Jack Lau --- libavformat/tls_openssl.c | 33 - 1 file changed, 20 insertions(+), 13 deletions(-) LGTM,

Re: [FFmpeg-devel] [PATCH] swscale/aarch64/output: Implement neon assembly for yuv2nv12cX_c()

2025-06-06 Thread Martin Storsjö
On Fri, 6 Jun 2025, Harshitha Sarangu Suresh wrote: Changed indentation, checked for FATE tests and gha-aarch64 git workflow. Everything passed. I doubt that everything passed; this doesn't even compile. See below: @@ -275,7 +291,10 @@ av_cold void ff_sws_init_swscale_aarch64(SwsInternal *c)

Re: [FFmpeg-devel] [PATCH] configure: Ignore nullability-completeness apple clang warnings

2025-06-05 Thread Martin Storsjö
On Thu, 5 Jun 2025, Dmitriy Kovalenko wrote: Some of the versions of Apple Clang produces a ton of the warnings related to the missing nullablity specifiers on the existing codebase of ffmpeg which significantly slows down the compilation becuase of the produced output size (especially on CI as

Re: [FFmpeg-devel] [PATCH 2/2] swscale: Neon rgb_to_yuv_half process 32 pixels at a time

2025-06-05 Thread Martin Storsjö
On Sat, 31 May 2025, Dmitriy Kovalenko wrote: This patch integrates so called double bufferring when we are loading 2 batch of elements at a time and then processing them in parallel. On the moden arm processors especially Apple Silicon it gives a visible benefit, for subsampled pixel processing

Re: [FFmpeg-devel] [PATCH 1/2] swscale: rgb_to_yuv neon optimizations

2025-06-05 Thread Martin Storsjö
On Sat, 31 May 2025, Dmitriy Kovalenko wrote: I've found quite a few ways to optimize existing ffmpeg's rgb to yuv subsampled conversion. In this patch stack I'll try to improve the perofrmance. This particular set of changes is a small improvement to all the existing functions and macro. The b

Re: [FFmpeg-devel] [PATCH] avformat/tls_openssl: fix build error when openssl version < 3

2025-06-05 Thread Martin Storsjö
On Thu, 5 Jun 2025, Jack Lau wrote: On Jun 5, 2025, at 15:02, Martin Storsjö wrote: On Thu, 5 Jun 2025, Jack Lau via ffmpeg-devel wrote: fix the missing data structure pkey in the tls_context Signed-off-by: Jack Lau --- libavformat/tls_openssl.c | 30 +- 1 file

Re: [FFmpeg-devel] [PATCH] avformat/tls_openssl: fix build error when openssl version < 3

2025-06-05 Thread Martin Storsjö
On Thu, 5 Jun 2025, Jack Lau via ffmpeg-devel wrote: fix the missing data structure pkey in the tls_context Signed-off-by: Jack Lau --- libavformat/tls_openssl.c | 30 +- 1 file changed, 17 insertions(+), 13 deletions(-) Thanks, this does fix the build break. Howev

Re: [FFmpeg-devel] [FFmpeg-cvslog] avformat/whip: Add WHIP muxer support for subsecond latency streaming

2025-06-04 Thread Martin Storsjö
On Wed, 4 Jun 2025, Jack Lau wrote: ffmpeg | branch: master | Jack Lau | Fri May 16 20:15:05 2025 +0800| [167e343bbe75515a80db8ee72ffa0c607c944a00] | committer: Steven Liu avformat/whip: Add WHIP muxer support for subsecond latency streaming This breaks compilation with OpenSSL 1.1.1t: src

Re: [FFmpeg-devel] [GASPP PATCH] Omit "-g" from the preprocessing command with cl.exe

2025-06-04 Thread Martin Storsjö
On Mon, 2 Jun 2025, Martin Storsjö wrote: This avoids warnings like cl : Command line warning D9002 : ignoring unknown option '-g' --- gas-preprocessor.pl | 1 + 1 file changed, 1 insertion(+) Pushed. // Martin ___ ffmpeg-devel ma

Re: [FFmpeg-devel] [PATCH] avutil/x86/intmath: remove inline asm implementations for clip functions

2025-06-03 Thread Martin Storsjö
On Tue, 3 Jun 2025, Niklas Haas wrote: On Mon, 02 Jun 2025 15:41:33 -0300 James Almer wrote: GCC/Clang is smart enough to emit minss/maxss the same way as these functions. The only theoretical benefit was in x86_32, where x87 floats are used, but the penalty of making the clipping opaque to th

[FFmpeg-devel] [GASPP PATCH] Omit "-g" from the preprocessing command with cl.exe

2025-06-02 Thread Martin Storsjö
This avoids warnings like cl : Command line warning D9002 : ignoring unknown option '-g' --- gas-preprocessor.pl | 1 + 1 file changed, 1 insertion(+) diff --git a/gas-preprocessor.pl b/gas-preprocessor.pl index aa3abc0..62c1a04 100755 --- a/gas-preprocessor.pl +++ b/gas-preprocessor.pl @@ -

Re: [FFmpeg-devel] [PATCH] swscale/aarch64/output: Implement neon assembly for yuv2nv12cX_c()

2025-06-02 Thread Martin Storsjö
On Mon, 2 Jun 2025, Harshitha Sarangu Suresh wrote: From 7260822a578130a713c1455cca6cdd06f1540db8 Mon Sep 17 00:00:00 2001 From: Harshitha Suresh Date: Mon, 19 May 2025 22:37:20 +0530 Subject: [PATCH] swscale/aarch64/output: Implement neon assembly for yuv2nv12cX_c() yuv2nv12cX_2_512_accurate

Re: [FFmpeg-devel] [PATCH] tests/fate/mov: Fix fate-mov-mp4-frag-flush requirements

2025-06-02 Thread Martin Storsjö
On Sun, 1 Jun 2025, Andreas Rheinhardt wrote: Patch attached. LGTM, thanks. // Martin ___ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org https://ffmpeg.org/mailman/listinfo/ffmpeg-devel To unsubscribe, visit link above, or email ffmpeg-devel-r

Re: [FFmpeg-devel] [PATCH 2/2] swscale: Neon rgb_to_yuv_half process 32 pixels at a time

2025-05-31 Thread Martin Storsjö
On Sat, 31 May 2025, Dmitriy Kovalenko wrote: Correct. I meant dual issue https://developer.arm.com/documentation/ddi0460/d/Cycle-Timings-and-Interlock-Behavior/Dual-issue Do not top post. // Martin ___ ffmpeg-devel mailing list ffmpeg-devel@ffmp

Re: [FFmpeg-devel] [PATCH v4 2/2] swscale: Neon rgb_to_yuv_half process 32 pixels at a time

2025-05-31 Thread Martin Storsjö
On Sat, 31 May 2025, Dmitriy Kovalenko wrote: Great. I send another version with the reverted change for the asr register change. What is the correct process to reply for the inline changes then? Inline email answer or cover letter? Do not top post. Reply to review comments in an inline reply

Re: [FFmpeg-devel] [PATCH v4 2/2] swscale: Neon rgb_to_yuv_half process 32 pixels at a time

2025-05-30 Thread Martin Storsjö
On Fri, 30 May 2025, Dmitriy Kovalenko wrote: === Feedback response === FWIW, the procedure is to respond to inline comments by replying to the mails where those comments were made. When they're included here, they end up as part of your suggested commit message. Anyway, now this time, the

Re: [FFmpeg-devel] [PATCH v4 1/2] swscale: rgb_to_yuv neon optimizations

2025-05-30 Thread Martin Storsjö
On Fri, 30 May 2025, Dmitriy Kovalenko wrote: If you with "non-performant mobile" mean small in-order cores, most of them can handle repeated accumulation like these even faster, if you sequence these so that all accumulations to one register is sequentially. E.g. first all "smlal \u_dst1\().4

Re: [FFmpeg-devel] [PATCH v3 1/2] swscale: rgb_to_yuv neon optimizations

2025-05-30 Thread Martin Storsjö
On Fri, 30 May 2025, Dmitriy Kovalenko wrote: I've found quite a few ways to optimize existing ffmpeg's rgb to yuv subsampled conversion. In this patch stack I'll try to improve the perofrmance. This particular set of changes is a small improvement to all the existing functions and macro. The b

Re: [FFmpeg-devel] [PATCH 1/2] swscale: rgb_to_yuv neon optimizations

2025-05-30 Thread Martin Storsjö
On Fri, 30 May 2025, Dmitriy Kovalenko wrote: All the comments were addressed If you fail to read the inline responses, we can end this conversation right here. And how did you test the prefetch, because I literally run a native benchmarking on the device right now and I see that with the

Re: [FFmpeg-devel] [PATCH 1/2] swscale: rgb_to_yuv neon optimizations

2025-05-30 Thread Martin Storsjö
On Thu, 29 May 2025, Dmitriy Kovalenko wrote: I appreciate the review for both the commits. I did fix all the unrelated changes and iterated in the new version, would appreciate the rearview. Don't top post. There are still at least 5 of my comments unaddressed. If you are not going to addr

Re: [FFmpeg-devel] [PATCH 2/2] swscale: Neon rgb_to_yuv_half process 16 pixels at a time

2025-05-30 Thread Martin Storsjö
On Thu, 29 May 2025, Martin Storsjö wrote: On Tue, 27 May 2025, Dmitriy Kovalenko wrote: This patches integrates so called double bufferring when we are loading 2 batch elements at a time and then processing them in parallel. On the moden arm processors especially Apple Silicon it gives a

Re: [FFmpeg-devel] [PATCH 1/2] swscale: rgb_to_yuv neon optimizations

2025-05-30 Thread Martin Storsjö
On Thu, 29 May 2025, Martin Storsjö wrote: In this case, they also add a direct dependence on the updated pointer register from the directly preceding instruction, which _is_ harmful on in-order cores, unless it entirely ignores the instruction.) I did benchmark this, and indeed it causes a

Re: [FFmpeg-devel] [PATCH 2/2] swscale: Neon rgb_to_yuv_half process 16 pixels at a time

2025-05-29 Thread Martin Storsjö
On Tue, 27 May 2025, Dmitriy Kovalenko wrote: This patches integrates so called double bufferring when we are loading 2 batch elements at a time and then processing them in parallel. On the moden arm processors especially Apple Silicon it gives a visible benefit, for subsampled pixel processing

Re: [FFmpeg-devel] [PATCH 1/2] swscale: rgb_to_yuv neon optimizations

2025-05-29 Thread Martin Storsjö
On Tue, 27 May 2025, Dmitriy Kovalenko wrote: This particular set of changes is a small improvement to all the existing functions and macro. The biggest performance gain is coming from post loading increment of the pointer and immediate prefetching How certain are you about the prefetching act

Re: [FFmpeg-devel] Building for Mac x86 with GCC & NASM

2025-05-28 Thread Martin Storsjö
On Wed, 28 May 2025, softworkz . wrote: -Original Message- From: ffmpeg-devel On Behalf Of Martin Storsjö Sent: Mittwoch, 28. Mai 2025 10:29 To: FFmpeg development discussions and patches Subject: Re: [FFmpeg-devel] Building for Mac x86 with GCC & NASM On Wed, 28 May 2025, softw

Re: [FFmpeg-devel] Building for Mac x86 with GCC & NASM

2025-05-28 Thread Martin Storsjö
On Wed, 28 May 2025, softworkz . wrote: -Original Message- From: ffmpeg-devel On Behalf Of Martin Storsjö Sent: Mittwoch, 28. Mai 2025 10:27 To: FFmpeg development discussions and patches Subject: Re: [FFmpeg-devel] Building for Mac x86 with GCC & NASM As we are talking about

Re: [FFmpeg-devel] Building for Mac x86 with GCC & NASM

2025-05-28 Thread Martin Storsjö
On Wed, 28 May 2025, softworkz . wrote: I haven't looked into in detail it myself, but it is a known issue since some of the ffmepg multithreading work a couple years ago. It does happen occasionally in all build configurations, more or less often. Yup. I thought about patching it in the build

Re: [FFmpeg-devel] Building for Mac x86 with GCC & NASM

2025-05-28 Thread Martin Storsjö
t 10:54 PM PDT, Martin Storsjö wrote: On Tue, 27 May 2025, softworkz . wrote: Hi, I have an issue with the CI builds for Mac in a way that it prints tons of lines like this: ld: warning: no platform load command found in 'libavcodec/libavcodec.a[1008](sao_10bit.o)', assuming: macO

Re: [FFmpeg-devel] Building for Mac x86 with GCC & NASM

2025-05-28 Thread Martin Storsjö
On Wed, 28 May 2025, softworkz . wrote: -Original Message- From: ffmpeg-devel On Behalf Of Martin Storsjö Sent: Mittwoch, 28. Mai 2025 07:55 To: FFmpeg development discussions and patches Subject: Re: [FFmpeg-devel] Building for Mac x86 with GCC & NASM These warnings appeared s

Re: [FFmpeg-devel] Building for Mac x86 with GCC & NASM

2025-05-27 Thread Martin Storsjö
On Tue, 27 May 2025, softworkz . wrote: Hi, I have an issue with the CI builds for Mac in a way that it prints tons of lines like this: ld: warning: no platform load command found in 'libavcodec/libavcodec.a[1008](sao_10bit.o)', assuming: macOS ld: warning: no platform load command found in

Re: [FFmpeg-devel] [PATCH] fix(configure): fix detection on windows

2025-05-27 Thread Martin Storsjö
On Mon, 26 May 2025, Coia Prant wrote: Can u backport it to released branch? Please don't top-post here. Sure, that can probably be reasonable; I can try to backport it to a few release branches within a day or two. // Martin ___ ffmpeg-devel ma

Re: [FFmpeg-devel] [PATCH v3 15/17] tests/checkasm: add checkasm tests for swscale ops

2025-05-27 Thread Martin Storsjö
On Tue, 27 May 2025, Niklas Haas wrote: From: Niklas Haas Because of the lack of an external ABI on low-level kernels, we cannot directly test internal functions. Instead, we construct a minimal op chain consisting of a read, the op to be tested, and a write. The bigger complication arises fr

Re: [FFmpeg-devel] [PATCH v3 04/17] tests/checkasm: generalize DEF_CHECKASM_CHECK_FUNC to floats

2025-05-27 Thread Martin Storsjö
On Tue, 27 May 2025, Niklas Haas wrote: From: Niklas Haas We split the standard macro into its body (implementation) and declaration, and use a macro argument in place of the raw `memcmp` call, with the major difference that we now take the number of pixels to compare instead of the number of

Re: [FFmpeg-devel] [PATCH] avformat/movenc: disallow buggy flags combination

2025-05-26 Thread Martin Storsjö
On Mon, 26 May 2025, Gyan Doshi wrote: Usage of hybrid_fragmented and faststart together can result in files with loss of sync and bitstream parsing errors upon playback. --- libavformat/movenc.c | 6 ++ 1 file changed, 6 insertions(+) diff --git a/libavformat/movenc.c b/libavformat/movenc.c

Re: [FFmpeg-devel] [PATCH] fix(configure): fix detection on windows

2025-05-26 Thread Martin Storsjö
On Sun, 25 May 2025, Martin Storsjö wrote: On Fri, 23 May 2025, Coia Prant wrote: On Windows Arm64 `uname -m` returned `x86_64` instead of `aarch64` Link: https://github.com/msys2/msys2-runtime/issues/171 On x86 32-bit toolchain msys2 environment `uname -m` returned `x86_64` instead of `i686

Re: [FFmpeg-devel] [PATCH] swscale/output: Implement neon intrinsics for yuv2planeX_10_c_template()

2025-05-26 Thread Martin Storsjö
On Thu, 22 May 2025, Harshitha Sarangu Suresh wrote: This optimization provides 5x improvement for the module. The boost in performance was calculated by adding C timers inside the C function and the optimized neon intrinsic function. From 904144c2db9e5e72d56360c4c2eb38d426852901 Mon Sep 17

Re: [FFmpeg-devel] [FFmpeg-cvslog] avcodec/asvenc: Don't waste bits encoding non-visible part

2025-05-25 Thread Martin Storsjö
On Mon, 26 May 2025, Andreas Rheinhardt wrote: ffmpeg | branch: master | Andreas Rheinhardt | Thu May 22 15:57:13 2025 +0200| [0401ca714a2714743573e27c384ffa810fd31a92] | committer: Andreas Rheinhardt avcodec/asvenc: Don't waste bits encoding non-visible part Up until now, the encoder repli

Re: [FFmpeg-devel] [PATCH] fix(configure): fix detection on windows

2025-05-25 Thread Martin Storsjö
On Fri, 23 May 2025, Coia Prant wrote: On Windows Arm64 `uname -m` returned `x86_64` instead of `aarch64` Link: https://github.com/msys2/msys2-runtime/issues/171 On x86 32-bit toolchain msys2 environment `uname -m` returned `x86_64` instead of `i686` or `x86` So check MSYSTEM_CARCH on windows

Re: [FFmpeg-devel] [FFmpeg-cvslog] tests/fate/ac3: add a second ac3_fixed encoder test

2025-05-22 Thread Martin Storsjö
On Thu, 22 May 2025, James Almer wrote: ffmpeg | branch: master | James Almer | Thu May 22 19:23:35 2025 -0300| [622a72b5ea5f2022b173355d65d513df2d75000b] | committer: James Almer tests/fate/ac3: add a second ac3_fixed encoder test Exercising the lavfi filtergraph codepath to choose the best

Re: [FFmpeg-devel] gcc: Remove auto-vectorization limitation.

2025-05-21 Thread Martin Storsjö
On Wed, 21 May 2025, Andreas Rheinhardt wrote: Martin Storsjö: On Wed, 21 May 2025, Andreas Rheinhardt wrote: Jiawei: This patch modifies the FFmpeg build system to remove the explicit disabling of GCC's auto-vectorization feature. Modern GCC versions (>= 10.0) have demonstrate

Re: [FFmpeg-devel] gcc: Remove auto-vectorization limitation.

2025-05-21 Thread Martin Storsjö
On Wed, 21 May 2025, Andreas Rheinhardt wrote: Jiawei: This patch modifies the FFmpeg build system to remove the explicit disabling of GCC's auto-vectorization feature. Modern GCC versions (>= 10.0) have demonstrated stable auto-vectorization capabilities through extensive optimizations in loo

Re: [FFmpeg-devel] [PATCH 04/17] tests/checkasm: add checkasm_check_float

2025-05-18 Thread Martin Storsjö
On Sun, 18 May 2025, Niklas Haas wrote: From: Niklas Haas --- tests/checkasm/checkasm.c | 1 + tests/checkasm/checkasm.h | 1 + 2 files changed, 2 insertions(+) diff --git a/tests/checkasm/checkasm.c b/tests/checkasm/checkasm.c index 71d1e5766c..c77216ed57 100644 --- a/tests/checkasm/checkasm.c

Re: [FFmpeg-devel] [FFmpeg-cvslog] fftools/graphprint: Now, make it a Killer-Feature!

2025-05-16 Thread Martin Storsjö
On Fri, 16 May 2025, softworkz . wrote: -Original Message- From: ffmpeg-devel On Behalf Of Martin Storsjö Sent: Freitag, 16. Mai 2025 10:19 To: FFmpeg development discussions and patches Subject: Re: [FFmpeg-devel] [FFmpeg-cvslog] fftools/graphprint: Now, make it a Killer-Feature! On

Re: [FFmpeg-devel] [FFmpeg-cvslog] fftools/graphprint: Now, make it a Killer-Feature!

2025-05-16 Thread Martin Storsjö
On Fri, 16 May 2025, softworkz . wrote: -Original Message- From: ffmpeg-devel On Behalf Of Martin Storsjö Sent: Freitag, 16. Mai 2025 08:22 To: ffmpeg-devel@ffmpeg.org Subject: Re: [FFmpeg-devel] [FFmpeg-cvslog] fftools/graphprint: Now, make it a Killer-Feature! On Thu, 15 May 2025

[FFmpeg-devel] [PATCH] [FATE] Properly list test errors when there are no failed tests

2025-05-16 Thread Martin Storsjö
If there were failures while running tests, e.g. if failing to compile checkasm or any other of the test programs, there are no failed tests per se, and the number of succeessful tests is equal to the total number of tests. For these cases, check the job status code instead of declaring them as a

Re: [FFmpeg-devel] [FFmpeg-cvslog] fftools/graphprint: Now, make it a Killer-Feature!

2025-05-15 Thread Martin Storsjö
On Thu, 15 May 2025, softworkz wrote: ffmpeg | branch: master | softworkz | Thu May 15 23:10:02 2025 +0200| [1f2b8d7238eff4ab8a4d8d6177e250b8180d51f4] | committer: softworkz fftools/graphprint: Now, make it a Killer-Feature! remember this: -sg <= means Show Graph Signed-off-by: softworkz

Re: [FFmpeg-devel] [PATCH] aarch64: increase default alignment for functions and constants

2025-05-15 Thread Martin Storsjö
On Fri, 16 May 2025, Ramiro Polla wrote: Use 16-byte alignment (align=4) instead of 4-byte (align=2) in the function and const macros. This improves instruction fetch and NEON load performance on modern AArch64 CPUs. --- libavutil/aarch64/asm.S | 4 ++-- 1 file changed, 2 insertions(+), 2 deletio

Re: [FFmpeg-devel] [PATCH] fix(configure): fix detection on windows arm64

2025-05-13 Thread Martin Storsjö
On Mon, 12 May 2025, Coia Prant wrote: On Windows Arm64 `uname -m` returned `x86_64` instead of `aarch64` Link: https://github.com/msys2/msys2-runtime/issues/171 But `uname -s` contains `ARM64` suffix So check suffix on windows arm64 (for clangarm64) This problem also in VideoLAN/x264 Link: ht

Re: [FFmpeg-devel] [PATCH 5/5] checkasm: add vvc_sao

2025-05-06 Thread Martin Storsjö
On Sat, 3 May 2025, Nuo Mi wrote: Hi Martin,Great, it works!  HEVC is included in v2. Thanks great, thanks for looking into it! The checkasm aspects of patches 5-7/7 look good to me. // Martin ___ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org

Re: [FFmpeg-devel] [PATCH] configure: Enable -fno-common for Darwin targets, avoid linker warnings

2025-05-05 Thread Martin Storsjö
On Fri, 2 May 2025, Martin Storsjö wrote: On Tue, 29 Apr 2025, Martin Storsjö wrote: Since GCC 10 and llvm.org Clang 11, -fno-common is the default. However Apple's Xcode Clang hasn't followed suit yet, and still defaults to -fcommon. Compiling with -fcommon causes uninitiali

Re: [FFmpeg-devel] [PATCH 5/5] checkasm: add vvc_sao

2025-05-02 Thread Martin Storsjö
On Fri, 2 May 2025, Nuo Mi wrote: On Fri, May 2, 2025 at 3:49 PM Martin Storsjö wrote: On Fri, 2 May 2025, Nuo Mi wrote: > From: Shaun Loo > > This is a part of Google Summer of Code 2023 > > AVX2: > - vvc_sao.sao_band [OK] &g

Re: [FFmpeg-devel] [PATCH] configure: Enable -fno-common for Darwin targets, avoid linker warnings

2025-05-02 Thread Martin Storsjö
On Tue, 29 Apr 2025, Martin Storsjö wrote: Since GCC 10 and llvm.org Clang 11, -fno-common is the default. However Apple's Xcode Clang hasn't followed suit yet, and still defaults to -fcommon. Compiling with -fcommon causes uninitialized global variables to be treated as "common

Re: [FFmpeg-devel] [PATCH 5/5] checkasm: add vvc_sao

2025-05-02 Thread Martin Storsjö
On Fri, 2 May 2025, Nuo Mi wrote: From: Shaun Loo This is a part of Google Summer of Code 2023 AVX2: - vvc_sao.sao_band [OK] - vvc_sao.sao_edge [OK] Co-authored-by: Nuo Mi --- tests/checkasm/Makefile | 2 +- tests/checkasm/checkasm.c | 1 + tests/checkasm/checkasm.h | 1 + tests/checka

Re: [FFmpeg-devel] [PATCH] aarch64/h26x: optimize sao_band_filter

2025-04-29 Thread Martin Storsjö
On Tue, 29 Apr 2025, Zhao Zhili wrote: On Apr 25, 2025, at 16:25, Martin Storsjö wrote: On Tue, 15 Apr 2025, Zhao Zhili wrote: +tbx v3.8b, {v16.16b-v17.16b}, v3.8b Is there any specific reason for preferring tbx over tbl here? (I know the existing code used tbx

[FFmpeg-devel] [PATCH] configure: Enable -fno-common for Darwin targets, avoid linker warnings

2025-04-29 Thread Martin Storsjö
Since GCC 10 and llvm.org Clang 11, -fno-common is the default. However Apple's Xcode Clang hasn't followed suit yet, and still defaults to -fcommon. Compiling with -fcommon causes uninitialized global variables to be treated as "common" (which allows multiple object files to have similar definiti

Re: [FFmpeg-devel] [PATCH] aarch64/h26x: Add put_hevc_pel_bi_w_pixels

2025-04-25 Thread Martin Storsjö
On Wed, 23 Apr 2025, Zhao Zhili wrote: From: Zhao Zhili On rpi5 (A76): put_hevc_pel_bi_w_pixels4_8_c: 90.0 ( 1.00x) put_hevc_pel_bi_w_pixels4_8_neon: 34.1 ( 2.64x) put_hevc_pel_bi_w_pixels6_8_c: 188.3 ( 1.00x) put_hevc_pel

Re: [FFmpeg-devel] [PATCH] aarch64/h26x: optimize sao_band_filter

2025-04-25 Thread Martin Storsjö
On Tue, 15 Apr 2025, Zhao Zhili wrote: From: Zhao Zhili int8_t[] is enough for offset_table of 8 bit streams. On rpi5: Before After hevc_sao_band_8_8_c: 252.3 ( 1.00x) 252.3 ( 1.00x) hevc_sao_band_8_8_neon:95.8 ( 2.63x) 61.0

Re: [FFmpeg-devel] [PATCH 2/3] avformat/hls: Show error info when read key file failed

2025-04-25 Thread Martin Storsjö
On Thu, 24 Apr 2025, Zhao Zhili wrote: On Apr 24, 2025, at 10:24, Steven Liu wrote: On Apr 23, 2025, at 20:59, Zhao Zhili wrote: Hi Zhili, From: Zhao Zhili --- libavformat/hls.c | 12 +++- 1 file changed, 7 insertions(+), 5 deletions(-) diff --git a/libavformat/hls.c b/libavfo

Re: [FFmpeg-devel] [PATCH 3/4] checkasm: hevc_pel: Use helpers for checking for writes out of bounds

2025-04-10 Thread Martin Storsjö
On Tue, 1 Apr 2025, yinshiyou...@loongson.cn wrote: -原始邮件- 发件人: "Martin Storsjö" 发送时间:2025-04-01 17:35:16 (星期二) 收件人: ffmpeg-devel@ffmpeg.org 抄送: jinbo , yinshiyou...@loongson.cn, "Lu Wang" 主题: Re: [PATCH 3/4] checkasm: hevc_pel: Use helpers for checking for write

  1   2   3   4   5   6   7   8   9   10   >