diff options
| author | L. E. Segovia <amy@centricular.com> | 2024-08-06 16:29:53 -0300 |
|---|---|---|
| committer | Backport Bot <gitlab-backport-bot@gstreamer-foundation.org> | 2024-08-09 19:45:24 +0100 |
| commit | ad354dd8f9796662e891ae7d2e63caaaee840865 (patch) | |
| tree | 6fd9fe91570fef42680cfaf69fdba00d77adf2a5 | |
| parent | 7b1620c92e3b458b2738afd2e8b07b43fa1bc890 (diff) | |
libvpx: Update to 1.14.1
Part-of: <https://gitlab.freedesktop.org/gstreamer/cerbero/-/merge_requests/1543>
| -rw-r--r-- | recipes/libvpx.recipe | 6 | ||||
| -rw-r--r-- | recipes/libvpx/0001-Fix-implicit-argument-conversion-on-UWP.patch | 27 | ||||
| -rw-r--r-- | recipes/libvpx/0002-vp9_diamond_search_sad_neon-use-DECLARE_ALIGNED.patch | 32 | ||||
| -rw-r--r-- | recipes/libvpx/0003-Add-Meson-build.patch | 2693 |
4 files changed, 2435 insertions, 323 deletions
diff --git a/recipes/libvpx.recipe b/recipes/libvpx.recipe index de808cc2..4598eaf9 100644 --- a/recipes/libvpx.recipe +++ b/recipes/libvpx.recipe @@ -7,11 +7,11 @@ import shutil class Recipe(recipe.Recipe): name = 'libvpx' - version = 'v1.13.1' + version = 'v1.14.1' stype = SourceType.TARBALL url = 'https://github.com/webmproject/libvpx/archive/%(version)s.tar.gz' tarball_dirname = 'libvpx-' + version[1:] - tarball_checksum = '00dae80465567272abd077f59355f95ac91d7809a2d3006f9ace2637dd429d14' + tarball_checksum = '901747254d80a7937c933d03bd7c5d41e8e6c883e0665fadcb172542167c7977' licenses = [{License.BSD: ['LICENSE'], License.Misc: ['PATENTS']}] btype = BuildType.MESON @@ -30,8 +30,6 @@ class Recipe(recipe.Recipe): } patches = [ - 'libvpx/0001-Fix-implicit-argument-conversion-on-UWP.patch', - 'libvpx/0002-vp9_diamond_search_sad_neon-use-DECLARE_ALIGNED.patch', 'libvpx/0003-Add-Meson-build.patch', ] diff --git a/recipes/libvpx/0001-Fix-implicit-argument-conversion-on-UWP.patch b/recipes/libvpx/0001-Fix-implicit-argument-conversion-on-UWP.patch deleted file mode 100644 index 6ff8027f..00000000 --- a/recipes/libvpx/0001-Fix-implicit-argument-conversion-on-UWP.patch +++ /dev/null @@ -1,27 +0,0 @@ -From 0339bdc0a49c415f55f728ac9f069ff426898722 Mon Sep 17 00:00:00 2001 -From: Nirbheek Chauhan <nirbheek@centricular.com> -Date: Wed, 11 Jan 2023 02:31:02 +0530 -Subject: [PATCH] Fix implicit argument conversion on UWP - -vpx_util\vpx_thread.h(79,13): error C2664: 'HANDLE CreateThread(LPSECURITY_ATTRIBUTES,SIZE_T,LPTHREAD_START_ROUTINE,LPVOID,DWORD,LPDWORD)': cannot convert argument 3 from 'unsigned int (__cdecl *)(void *)' to 'LPTHREAD_START_ROUTINE' ---- - vpx_util/vpx_thread.h | 3 ++- - 1 file changed, 2 insertions(+), 1 deletion(-) - -diff --git a/vpx_util/vpx_thread.h b/vpx_util/vpx_thread.h -index 6d308e9..c84e96a 100644 ---- a/vpx_util/vpx_thread.h -+++ b/vpx_util/vpx_thread.h -@@ -78,7 +78,8 @@ static INLINE int pthread_create(pthread_t *const thread, const void *attr, - #ifdef USE_CREATE_THREAD - *thread = CreateThread(NULL, /* lpThreadAttributes */ - 0, /* dwStackSize */ -- start, arg, 0, /* dwStackSize */ -+ (LPTHREAD_START_ROUTINE) start, arg, -+ 0, /* dwStackSize */ - NULL); /* lpThreadId */ - #else - *thread = (pthread_t)_beginthreadex(NULL, /* void *security */ --- -2.37.1 (Apple Git-137.1) - diff --git a/recipes/libvpx/0002-vp9_diamond_search_sad_neon-use-DECLARE_ALIGNED.patch b/recipes/libvpx/0002-vp9_diamond_search_sad_neon-use-DECLARE_ALIGNED.patch deleted file mode 100644 index 8ec541b2..00000000 --- a/recipes/libvpx/0002-vp9_diamond_search_sad_neon-use-DECLARE_ALIGNED.patch +++ /dev/null @@ -1,32 +0,0 @@ -From 7c854566c8dbbf047a90139a928a6ac9196546af Mon Sep 17 00:00:00 2001 -Message-ID: <7c854566c8dbbf047a90139a928a6ac9196546af.1692093230.git.amy@centricular.com> -From: James Zern <jzern@google.com> -Date: Wed, 1 Feb 2023 13:27:06 -0800 -Subject: [PATCH 2/3] vp9_diamond_search_sad_neon: use DECLARE_ALIGNED - -rather than the gcc specific __attribute__((aligned())); fixes build -targeting ARM64 windows. - -Bug: webm:1788 -Change-Id: I2210fc215f44d90c1ce9dee9b54888eb1b78c99e -(cherry picked from commit 858a8c611f4c965078485860a6820e2135e6611b) ---- - vp9/encoder/arm/neon/vp9_diamond_search_sad_neon.c | 2 +- - 1 file changed, 1 insertion(+), 1 deletion(-) - -diff --git a/vp9/encoder/arm/neon/vp9_diamond_search_sad_neon.c b/vp9/encoder/arm/neon/vp9_diamond_search_sad_neon.c -index 33753f77b..997775a66 100644 ---- a/vp9/encoder/arm/neon/vp9_diamond_search_sad_neon.c -+++ b/vp9/encoder/arm/neon/vp9_diamond_search_sad_neon.c -@@ -220,7 +220,7 @@ int vp9_diamond_search_sad_neon(const MACROBLOCK *x, - // Look up the component cost of the residual motion vector - { - uint32_t cost[4]; -- int16_t __attribute__((aligned(16))) rowcol[8]; -+ DECLARE_ALIGNED(16, int16_t, rowcol[8]); - vst1q_s16(rowcol, v_diff_mv_w); - - // Note: This is a use case for gather instruction --- -2.41.0 - diff --git a/recipes/libvpx/0003-Add-Meson-build.patch b/recipes/libvpx/0003-Add-Meson-build.patch index 0f2e3cb3..608c3018 100644 --- a/recipes/libvpx/0003-Add-Meson-build.patch +++ b/recipes/libvpx/0003-Add-Meson-build.patch @@ -4,94 +4,6 @@ Date: Tue, 18 Apr 2023 22:16:49 -0300 Subject: [PATCH] Add Meson build For more information, please see https://github.com/dragonCodecs/libvpx/tree/old-meson ---- - LICENSE.meson.md | 51 + - README.meson.md | 98 + - examples/meson.build | 452 +++++ - examples/samples.dox.in | 14 + - meson.build | 1900 +++++++++++++++++++ - meson/arm_neon.h | 17 + - meson/capture_build_options.py | 18 + - meson/check_test_data.py | 45 + - meson/gen_def.py | 100 + - meson/generate_component_dox.py | 25 + - meson/generate_doxy.py | 28 + - meson/generate_doxyfile.py | 44 + - meson/parse_options.py | 171 ++ - meson/parse_sources.py | 219 +++ - meson/patch-configure.diff | 60 + - meson/stdinout_wrapper.py | 19 + - meson/transform_config.py | 53 + - meson/transform_config_asm.py | 37 + - meson/vpx_config.c | 10 + - meson_options.txt | 49 + - subprojects/.gitignore | 3 + - subprojects/nasm-mac.wrap | 9 + - subprojects/nasm-win.wrap | 9 + - subprojects/packagefiles/nasm/meson.build | 13 + - subprojects/packagefiles/nasm/patch_nasm.py | 15 + - subprojects/packagefiles/perl/meson.build | 8 + - subprojects/perl-win.wrap | 10 + - test/meson.build | 1593 ++++++++++++++++ - third_party/googletest/meson.build | 33 + - third_party/libwebm/meson.build | 52 + - third_party/libyuv/meson.build | 49 + - third_party/meson.build | 9 + - tools/meson.build | 67 + - tools/tools.dox.in | 5 + - vp8/meson.build | 554 ++++++ - vp9/meson.build | 595 ++++++ - vpx/meson.build | 136 ++ - vpx_dsp/arm/meson.build | 12 + - vpx_dsp/meson.build | 1206 ++++++++++++ - vpx_mem/meson.build | 41 + - vpx_ports/meson.build | 157 ++ - vpx_scale/meson.build | 71 + - vpx_util/meson.build | 61 + - 43 files changed, 8118 insertions(+) - create mode 100644 LICENSE.meson.md - create mode 100644 README.meson.md - create mode 100644 examples/meson.build - create mode 100644 examples/samples.dox.in - create mode 100644 meson.build - create mode 100644 meson/arm_neon.h - create mode 100644 meson/capture_build_options.py - create mode 100644 meson/check_test_data.py - create mode 100755 meson/gen_def.py - create mode 100644 meson/generate_component_dox.py - create mode 100644 meson/generate_doxy.py - create mode 100644 meson/generate_doxyfile.py - create mode 100644 meson/parse_options.py - create mode 100644 meson/parse_sources.py - create mode 100644 meson/patch-configure.diff - create mode 100644 meson/stdinout_wrapper.py - create mode 100644 meson/transform_config.py - create mode 100644 meson/transform_config_asm.py - create mode 100644 meson/vpx_config.c - create mode 100644 meson_options.txt - create mode 100644 subprojects/.gitignore - create mode 100644 subprojects/nasm-mac.wrap - create mode 100644 subprojects/nasm-win.wrap - create mode 100644 subprojects/packagefiles/nasm/meson.build - create mode 100644 subprojects/packagefiles/nasm/patch_nasm.py - create mode 100644 subprojects/packagefiles/perl/meson.build - create mode 100644 subprojects/perl-win.wrap - create mode 100644 test/meson.build - create mode 100644 third_party/googletest/meson.build - create mode 100644 third_party/libwebm/meson.build - create mode 100644 third_party/libyuv/meson.build - create mode 100644 third_party/meson.build - create mode 100644 tools/meson.build - create mode 100644 tools/tools.dox.in - create mode 100644 vp8/meson.build - create mode 100644 vp9/meson.build - create mode 100644 vpx/meson.build - create mode 100644 vpx_dsp/arm/meson.build - create mode 100644 vpx_dsp/meson.build - create mode 100644 vpx_mem/meson.build - create mode 100644 vpx_ports/meson.build - create mode 100644 vpx_scale/meson.build - create mode 100644 vpx_util/meson.build diff --git a/LICENSE.meson.md b/LICENSE.meson.md new file mode 100644 @@ -8478,12 +8390,6 @@ From: "L. E. Segovia" <amy@amyspark.me> Date: Tue, 18 Apr 2023 22:59:00 -0300 Subject: [PATCH] Add GitLab CI ---- - .gitlab-ci.yml | 335 +++++++++++++++++++++++++++++++++++++ - ci/vs-arm64-cross-file.txt | 14 ++ - 2 files changed, 349 insertions(+) - create mode 100644 .gitlab-ci.yml - create mode 100644 ci/vs-arm64-cross-file.txt diff --git a/.gitlab-ci.yml b/.gitlab-ci.yml new file mode 100644 @@ -8856,11 +8762,6 @@ Date: Wed, 17 May 2023 20:08:48 -0300 Subject: [PATCH] meson: Protect build system handled features with combo choices ---- - meson.build | 34 +++++++++++++-------------- - meson/parse_options.py | 29 ++++++++++++++++++++--- - meson_options.txt | 52 +++++++++++++++++++++--------------------- - 3 files changed, 68 insertions(+), 47 deletions(-) diff --git a/meson.build b/meson.build index 78578a1f5..a2eeacf17 100644 @@ -9080,9 +8981,6 @@ Date: Tue, 4 Jul 2023 22:05:03 -0300 Subject: [PATCH] meson: Fix dependency override's library name Part-of: <https://gitlab.freedesktop.org/gstreamer/meson-ports/libvpx/-/merge_requests/4> ---- - meson.build | 2 +- - 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/meson.build b/meson.build index a2eeacf17..29a14dc86 100644 @@ -9107,10 +9005,6 @@ Date: Wed, 5 Jul 2023 14:16:50 -0300 Subject: [PATCH] meson: Port thumb to Python Part-of: <https://gitlab.freedesktop.org/gstreamer/meson-ports/libvpx/-/merge_requests/5> ---- - build/make/thumb.py | 52 +++++++++++++++++++++++++++++++++++++++++++++ - 1 file changed, 52 insertions(+) - create mode 100755 build/make/thumb.py diff --git a/build/make/thumb.py b/build/make/thumb.py new file mode 100755 @@ -9180,11 +9074,6 @@ Date: Wed, 5 Jul 2023 14:15:44 -0300 Subject: [PATCH] meson: Port ads2armasm_ms to Python Part-of: <https://gitlab.freedesktop.org/gstreamer/meson-ports/libvpx/-/merge_requests/5> ---- - build/make/ads2armasm_ms.py | 33 +++++++++++++++++++++++++++++++++ - meson.build | 2 +- - 2 files changed, 34 insertions(+), 1 deletion(-) - create mode 100755 build/make/ads2armasm_ms.py diff --git a/build/make/ads2armasm_ms.py b/build/make/ads2armasm_ms.py new file mode 100755 @@ -9248,11 +9137,6 @@ Date: Wed, 5 Jul 2023 14:15:55 -0300 Subject: [PATCH] meson: Port ads2gas_apple to Python Part-of: <https://gitlab.freedesktop.org/gstreamer/meson-ports/libvpx/-/merge_requests/5> ---- - build/make/ads2gas_apple.py | 98 +++++++++++++++++++++++++++++++++++++ - meson.build | 2 +- - 2 files changed, 99 insertions(+), 1 deletion(-) - create mode 100755 build/make/ads2gas_apple.py diff --git a/build/make/ads2gas_apple.py b/build/make/ads2gas_apple.py new file mode 100755 @@ -9381,11 +9265,6 @@ Date: Wed, 5 Jul 2023 14:16:44 -0300 Subject: [PATCH] meson: Port ads2gas to Python Part-of: <https://gitlab.freedesktop.org/gstreamer/meson-ports/libvpx/-/merge_requests/5> ---- - build/make/ads2gas.py | 140 ++++++++++++++++++++++++++++++++++++++++++ - meson.build | 2 +- - 2 files changed, 141 insertions(+), 1 deletion(-) - create mode 100755 build/make/ads2gas.py diff --git a/build/make/ads2gas.py b/build/make/ads2gas.py new file mode 100755 @@ -9556,19 +9435,6 @@ Date: Fri, 7 Jul 2023 22:10:48 -0300 Subject: [PATCH] meson: Port rtcd.pl to Python Part-of: <https://gitlab.freedesktop.org/gstreamer/meson-ports/libvpx/-/merge_requests/5> ---- - build/make/rtcd.py | 436 +++++++++ - meson.build | 14 +- - vp8/common/rtcd_defs.py | 248 +++++ - vp9/common/vp9_rtcd_defs.py | 214 +++++ - vpx_dsp/vpx_dsp_rtcd_defs.py | 1655 ++++++++++++++++++++++++++++++++++ - vpx_scale/vpx_scale_rtcd.py | 45 + - 6 files changed, 2605 insertions(+), 7 deletions(-) - create mode 100755 build/make/rtcd.py - create mode 100755 vp8/common/rtcd_defs.py - create mode 100755 vp9/common/vp9_rtcd_defs.py - create mode 100755 vpx_dsp/vpx_dsp_rtcd_defs.py - create mode 100755 vpx_scale/vpx_scale_rtcd.py diff --git a/build/make/rtcd.py b/build/make/rtcd.py new file mode 100755 @@ -12277,18 +12143,6 @@ Subject: [PATCH] Remove Perl as a dependency Fixes #1 Part-of: <https://gitlab.freedesktop.org/gstreamer/meson-ports/libvpx/-/merge_requests/5> ---- - .gitlab-ci.yml | 7 ++----- - meson.build | 10 +--------- - subprojects/packagefiles/perl/meson.build | 8 -------- - subprojects/perl-win.wrap | 10 ---------- - vp8/meson.build | 2 +- - vp9/meson.build | 2 +- - vpx_dsp/meson.build | 2 +- - vpx_scale/meson.build | 2 +- - 8 files changed, 7 insertions(+), 36 deletions(-) - delete mode 100644 subprojects/packagefiles/perl/meson.build - delete mode 100644 subprojects/perl-win.wrap diff --git a/.gitlab-ci.yml b/.gitlab-ci.yml index 251278731..872c5dfaf 100644 @@ -12456,13 +12310,6 @@ Date: Sat, 8 Jul 2023 19:39:55 +0000 Subject: [PATCH] meson: Add test for rtcd generation parity Part-of: <https://gitlab.freedesktop.org/gstreamer/meson-ports/libvpx/-/merge_requests/5> ---- - meson.build | 34 +++++++++++++++++++++++++++++++--- - vp8/meson.build | 8 +++----- - vp9/meson.build | 9 ++++----- - vpx_dsp/meson.build | 8 +++----- - vpx_scale/meson.build | 8 +++----- - 5 files changed, 44 insertions(+), 23 deletions(-) diff --git a/meson.build b/meson.build index 9d7a8a9a2..2d64b1184 100644 @@ -12620,30 +12467,6 @@ Date: Tue, 18 Jul 2023 10:57:25 -0300 Subject: [PATCH] Backport NASM to GStreamer's manual execution and install Closes !3 ---- - meson.build | 60 ++++++++++------ - subprojects/.gitignore | 2 - - subprojects/nasm-mac.wrap | 9 --- - subprojects/nasm-win.wrap | 9 --- - subprojects/nasm/.gitignore | 2 + - subprojects/nasm/download-binary.py | 76 +++++++++++++++++++++ - subprojects/nasm/meson.build | 41 +++++++++++ - subprojects/nasm/patch_nasm.py | 15 ++++ - subprojects/packagefiles/nasm/meson.build | 13 ---- - subprojects/packagefiles/nasm/patch_nasm.py | 15 ---- - vp8/meson.build | 38 ++++++++--- - vp9/meson.build | 38 ++++++++--- - vpx_dsp/meson.build | 18 +++-- - vpx_ports/meson.build | 22 +++--- - 14 files changed, 256 insertions(+), 102 deletions(-) - delete mode 100644 subprojects/nasm-mac.wrap - delete mode 100644 subprojects/nasm-win.wrap - create mode 100644 subprojects/nasm/.gitignore - create mode 100755 subprojects/nasm/download-binary.py - create mode 100644 subprojects/nasm/meson.build - create mode 100755 subprojects/nasm/patch_nasm.py - delete mode 100644 subprojects/packagefiles/nasm/meson.build - delete mode 100644 subprojects/packagefiles/nasm/patch_nasm.py diff --git a/meson.build b/meson.build index 2d64b1184..4f9cde607 100644 @@ -13212,9 +13035,6 @@ Subject: [PATCH] meson: Work around Meson bug when using @SOURCE_DIR@ in a subproject Part-of: <https://gitlab.freedesktop.org/gstreamer/meson-ports/libvpx/-/merge_requests/7> ---- - meson.build | 8 +++++++- - 1 file changed, 7 insertions(+), 1 deletion(-) diff --git a/meson.build b/meson.build index 4f9cde607..5365af788 100644 @@ -13248,11 +13068,6 @@ Subject: [PATCH] meson: Fix several typos discovered when testing for Cerbero integration Part-of: <https://gitlab.freedesktop.org/gstreamer/meson-ports/libvpx/-/merge_requests/8> ---- - meson.build | 4 ++-- - vp9/meson.build | 4 ++-- - vpx_dsp/meson.build | 12 ++++++------ - 3 files changed, 10 insertions(+), 10 deletions(-) diff --git a/meson.build b/meson.build index 5365af788..111c44cee 100644 @@ -13364,9 +13179,6 @@ See: https://stackoverflow.com/questions/43152633/invalid-register-for-seh-savexmm-in-cygwin Part-of: <https://gitlab.freedesktop.org/gstreamer/meson-ports/libvpx/-/merge_requests/8> ---- - meson.build | 8 ++++++++ - 1 file changed, 8 insertions(+) diff --git a/meson.build b/meson.build index 111c44cee..a237c3edf 100644 @@ -13397,9 +13209,6 @@ Date: Sun, 30 Jul 2023 16:03:48 +0100 Subject: [PATCH] ci: bump macos and ios images Part-of: <https://gitlab.freedesktop.org/gstreamer/meson-ports/libvpx/-/merge_requests/9> ---- - .gitlab-ci.yml | 4 ++-- - 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/.gitlab-ci.yml b/.gitlab-ci.yml index 872c5dfaf..c4f27d483 100644 @@ -13431,9 +13240,6 @@ Date: Fri, 4 Aug 2023 23:09:33 +0000 Subject: [PATCH] meson: Implement header installation Part-of: <https://gitlab.freedesktop.org/gstreamer/meson-ports/libvpx/-/merge_requests/10> ---- - meson.build | 35 +++++++++++++++++++++++++++++++++++ - 1 file changed, 35 insertions(+) diff --git a/meson.build b/meson.build index a237c3edf..d34c07045 100644 @@ -13491,14 +13297,6 @@ Date: Fri, 4 Aug 2023 21:35:07 -0300 Subject: [PATCH] meson: Fix internal flags not superseding toolchain's Part-of: <https://gitlab.freedesktop.org/gstreamer/meson-ports/libvpx/-/merge_requests/10> ---- - examples/meson.build | 30 +++++++ - meson.build | 131 ++++++++++++++--------------- - third_party/googletest/meson.build | 5 +- - third_party/libwebm/meson.build | 5 +- - third_party/libyuv/meson.build | 3 +- - tools/meson.build | 3 +- - 6 files changed, 104 insertions(+), 73 deletions(-) diff --git a/examples/meson.build b/examples/meson.build index 9c3df73ab..9b511686e 100644 @@ -14048,9 +13846,6 @@ Date: Fri, 4 Aug 2023 21:48:05 -0300 Subject: [PATCH] meson: Fix another typo discovered by Cerbero's Part-of: <https://gitlab.freedesktop.org/gstreamer/meson-ports/libvpx/-/merge_requests/10> ---- - vpx_dsp/meson.build | 2 +- - 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/vpx_dsp/meson.build b/vpx_dsp/meson.build index 04ec4b5d7..d8c83297c 100644 @@ -14080,11 +13875,6 @@ Content-Transfer-Encoding: 8bit With credits to Tim-Philipp Müller for the cpu-features lookup. Part-of: <https://gitlab.freedesktop.org/gstreamer/meson-ports/libvpx/-/merge_requests/10> ---- - .gitlab-ci.yml | 3 ++- - meson.build | 40 +++++++++++++++++++++++++++++++++++++++- - meson_options.txt | 5 +++++ - 3 files changed, 46 insertions(+), 2 deletions(-) diff --git a/.gitlab-ci.yml b/.gitlab-ci.yml index c4f27d483..11fc1536a 100644 @@ -14199,9 +13989,6 @@ Date: Sat, 5 Aug 2023 00:12:48 -0300 Subject: [PATCH] meson: Fix missing PIC detection Part-of: <https://gitlab.freedesktop.org/gstreamer/meson-ports/libvpx/-/merge_requests/10> ---- - meson.build | 3 ++- - 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/meson.build b/meson.build index 3105f67f2..2dd5730ca 100644 @@ -14234,9 +14021,6 @@ Date: Sat, 5 Aug 2023 00:10:23 -0300 Subject: [PATCH] meson: Fix iOS Simulator build Part-of: <https://gitlab.freedesktop.org/gstreamer/meson-ports/libvpx/-/merge_requests/10> ---- - meson.build | 49 +++++++++++++++++++++++++++---------------------- - 1 file changed, 27 insertions(+), 22 deletions(-) diff --git a/meson.build b/meson.build index 2dd5730ca..3a72512cf 100644 @@ -14355,9 +14139,6 @@ Date: Mon, 7 Aug 2023 19:58:03 -0300 Subject: [PATCH] meson: Allow introspection on feature flags Part-of: <https://gitlab.freedesktop.org/gstreamer/meson-ports/libvpx/-/merge_requests/11> ---- - meson.build | 8 +++++++- - 1 file changed, 7 insertions(+), 1 deletion(-) diff --git a/meson.build b/meson.build index 3a72512cf..f177428f7 100644 @@ -14399,9 +14180,6 @@ Date: Thu, 10 Aug 2023 22:59:26 -0300 Subject: [PATCH] ci: Update Android jobs to R22b Part-of: <https://gitlab.freedesktop.org/gstreamer/meson-ports/libvpx/-/merge_requests/12> ---- - .gitlab-ci.yml | 13 ++++++------- - 1 file changed, 6 insertions(+), 7 deletions(-) diff --git a/.gitlab-ci.yml b/.gitlab-ci.yml index 11fc1536a..54145dc9f 100644 @@ -14454,10 +14232,6 @@ This is yet another instance of Meson calling Ninja from the root project's build directory, which mucks with the output of custom_target. Part-of: <https://gitlab.freedesktop.org/gstreamer/meson-ports/libvpx/-/merge_requests/14> ---- - meson.build | 1 + - meson/generate_doxyfile.py | 3 ++- - 2 files changed, 3 insertions(+), 1 deletion(-) diff --git a/meson.build b/meson.build index f177428f7..33618f9c7 100644 @@ -14505,9 +14279,6 @@ Subject: [PATCH] meson: Fix pthreads detection failure not disabling See gstreamer/gstreamer#2905 Part-of: <https://gitlab.freedesktop.org/gstreamer/meson-ports/libvpx/-/merge_requests/13> ---- - meson.build | 5 +++-- - 1 file changed, 3 insertions(+), 2 deletions(-) diff --git a/meson.build b/meson.build index 33618f9c7..8ae4f7dee 100644 @@ -14547,28 +14318,6 @@ running the script through the Python executable. See gstreamer/cerbero#450 Part-of: <https://gitlab.freedesktop.org/gstreamer/meson-ports/libvpx/-/merge_requests/15> ---- - meson/capture_build_options.py | 0 - meson/check_test_data.py | 0 - meson/generate_component_dox.py | 0 - meson/generate_doxy.py | 0 - meson/generate_doxyfile.py | 0 - meson/parse_options.py | 0 - meson/parse_sources.py | 0 - meson/stdinout_wrapper.py | 0 - meson/transform_config.py | 0 - meson/transform_config_asm.py | 0 - 10 files changed, 0 insertions(+), 0 deletions(-) - mode change 100644 => 100755 meson/capture_build_options.py - mode change 100644 => 100755 meson/check_test_data.py - mode change 100644 => 100755 meson/generate_component_dox.py - mode change 100644 => 100755 meson/generate_doxy.py - mode change 100644 => 100755 meson/generate_doxyfile.py - mode change 100644 => 100755 meson/parse_options.py - mode change 100644 => 100755 meson/parse_sources.py - mode change 100644 => 100755 meson/stdinout_wrapper.py - mode change 100644 => 100755 meson/transform_config.py - mode change 100644 => 100755 meson/transform_config_asm.py diff --git a/meson/capture_build_options.py b/meson/capture_build_options.py old mode 100644 @@ -14615,9 +14364,6 @@ should also produce a valid executable. See gstreamer/cerbero#450 Part-of: <https://gitlab.freedesktop.org/gstreamer/meson-ports/libvpx/-/merge_requests/16> ---- - meson.build | 2 +- - 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/meson.build b/meson.build index 8ae4f7dee..79ebb514c 100644 @@ -14641,9 +14387,6 @@ From: =?UTF-8?q?Tim-Philipp=20M=C3=BCller?= <tim@centricular.com> Date: Mon, 9 Oct 2023 18:43:36 +0100 Subject: [PATCH] meson: update to 1.13.1 ---- - meson.build | 2 +- - 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/meson.build b/meson.build index 79ebb514c..09b9859f1 100644 @@ -14672,9 +14415,6 @@ Aesthetic change Fixes #3 Part-of: <https://gitlab.freedesktop.org/gstreamer/meson-ports/libvpx/-/merge_requests/19> ---- - build/make/ads2gas.py | 3 ++- - 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/build/make/ads2gas.py b/build/make/ads2gas.py index c8025a971..0574135dd 100755 @@ -14759,3 +14499,2436 @@ index 1df26ca88..765d032b8 100644 -- 2.44.0.windows.1 + +From a334f583d65a4b416c8557b2242ca110994503a2 Mon Sep 17 00:00:00 2001 +From: "L. E. Segovia" <amy@centricular.com> +Date: Wed, 10 Jul 2024 12:36:17 -0300 +Subject: [PATCH] meson: Prohibit shared libraries outside of Linux and Darwin + platforms + +Windows in particular needs special consideration because of `__declspec(dllimport)` being absent from the data exports. + +diff --git a/meson.build b/meson.build +index 765d032b8..7366c61a7 100644 +--- a/meson.build ++++ b/meson.build +@@ -1563,6 +1563,14 @@ vpx_config_c = configure_file( + + # libs.mk + ++if features.get('shared', false) and not ['linux', 'darwin', 'ios', 'iphonesimulator'].contains(tgt_os) ++ if features.enabled('gnu', false) ++ warn('Shared libraries are only supported on ELF; assuming this is OK') ++ else ++ error('Shared libraries are only supported on ELF, OS/2, and Darwin for now') ++ endif ++endif ++ + rtcd_exe = find_program('build/make/rtcd.py', required: true) + + extra_libs += c.find_library('m', required: false) +@@ -1820,33 +1828,18 @@ foreach i: codec_arch_libs + objs += i.extract_all_objects(recursive: true) + endforeach + +-if features.get('static', false) and features.get('shared', false) +- libvpx = both_libraries( +- 'vpx', +- codec_srcs, +- c_args: project_c_args, +- link_args: project_c_link_args + link_args, +- extra_files: codec_headers, +- dependencies: extra_libs, +- objects: objs, +- version: soversion, +- vs_module_defs: vpx_def, +- install: features.get('install_libs'), +- ) +-else +- libvpx = library( +- 'vpx', +- codec_srcs, +- c_args: project_c_args, +- link_args: project_c_link_args + link_args, +- extra_files: codec_headers, +- dependencies: extra_libs, +- objects: objs, +- version: soversion, +- vs_module_defs: vpx_def, +- install: features.get('install_libs'), +- ) +-endif ++libvpx = library( ++ 'vpx', ++ codec_srcs, ++ c_args: project_c_args, ++ link_args: project_c_link_args + link_args, ++ extra_files: codec_headers, ++ dependencies: extra_libs, ++ objects: objs, ++ version: soversion, ++ vs_module_defs: vpx_def, ++ install: features.get('install_libs'), ++) + + headers = files( + 'vpx/vpx_codec.h', +-- +2.44.0.windows.1 + + +From 1fdc1f7de86ccf1b165fd10305c905745e326fe7 Mon Sep 17 00:00:00 2001 +From: "L. E. Segovia" <amy@centricular.com> +Date: Wed, 10 Jul 2024 12:42:38 -0300 +Subject: [PATCH] meson: Fix typo in the tiny_ssim executable clause + +Using combos as booleans is not valid, the features hash was intended here. + +diff --git a/tools/meson.build b/tools/meson.build +index 47c9ee831..7e6f16360 100644 +--- a/tools/meson.build ++++ b/tools/meson.build +@@ -28,7 +28,7 @@ tiny_ssim = executable( + link_args: project_c_link_args + link_args, + extra_files: tiny_ssim_headers, + dependencies: libvpx_dep, +- install: get_option('install_bins'), ++ install: features.get('install_bins', false), + ) + + tools_subpages = { +-- +2.44.0.windows.1 + + +From eecbef0754c7b36bb55c6b0ec1aca0717bc69ea9 Mon Sep 17 00:00:00 2001 +From: "L. E. Segovia" <amy@centricular.com> +Date: Sat, 20 Jul 2024 21:07:28 +0000 +Subject: [PATCH] ci: Switch Apple jobs to the Arm64 runner + +Fixes #2 + +Part-of: <https://gitlab.freedesktop.org/gstreamer/meson-ports/libvpx/-/merge_requests/22> + +diff --git a/.gitlab-ci.yml b/.gitlab-ci.yml +index 54145dc9f..cc7c27de1 100644 +--- a/.gitlab-ci.yml ++++ b/.gitlab-ci.yml +@@ -198,6 +198,9 @@ vs2019 arm64 cross: + + .build darwin: + stage: 'build' ++ image: "registry.freedesktop.org/gstreamer/gstreamer/macos-arm64/14-sonoma:2023-10-25.0" ++ tags: ++ - gst-mac-arm + artifacts: + name: "${CI_JOB_NAME}_${CI_COMMIT_SHA}" + expire_in: '5 days' +@@ -215,35 +218,35 @@ vs2019 arm64 cross: + - pip3 install --upgrade pip + # Make sure meson is up to date + - pip3 install -U meson +- # Need to install certificates for python +- - pip3 install --upgrade certifi +- # Another way to install certificates +- - open /Applications/Python\ 3.8/Install\ Certificates.command + # Get ninja + - pip3 install -U ninja + script: +- # HACK to sneak Nasm and Perl under Meson's nose. +- - export PATH="$(pwd)/subprojects/nasm-2.16.01:$PATH" +- - CERT_PATH=$(python3 -m certifi) && export SSL_CERT_FILE=${CERT_PATH} && export REQUESTS_CA_BUNDLE=${CERT_PATH} && meson setup mesonbuild ++ - meson setup mesonbuild + - meson compile -C mesonbuild + - meson test -C mesonbuild + ++macOS arm64: ++ extends: '.build darwin' ++ + macOS x86_64: + extends: '.build darwin' +- stage: 'build' +- tags: +- - gst-macos-13 ++ script: ++ # HACK to sneak Nasm and Perl under Meson's nose. ++ - export PATH="$(pwd)/subprojects/nasm-2.16.01:$PATH" ++ - meson setup mesonbuild ++ - meson compile -C mesonbuild ++ - meson test -C mesonbuild + + iOS arm64 cross: + extends: '.build darwin' +- stage: 'build' +- tags: +- - gst-ios-16 ++ # Same-architecture cross-builds are broken on 1.4.1 ++ # 1.5.0 was meant to fix this, but instead broke detection wholesale ++ # Once 1.5.1 is out, try restoring `system = 'darwin'` + script: + - | + cat > ios-cross-file.txt <<EOF + [host_machine] +- system = 'darwin' ++ system = 'ios' + cpu_family = 'aarch64' + cpu = 'aarch64' + endian = 'little' +@@ -262,7 +265,7 @@ iOS arm64 cross: + strip = '$(xcrun --find --sdk iphoneos strip)' + pkgconfig = 'false' + EOF +- - CERT_PATH=$(python3 -m certifi) && export SSL_CERT_FILE=${CERT_PATH} && export REQUESTS_CA_BUNDLE=${CERT_PATH} && meson setup mesonbuild --cross-file=ios-cross-file.txt ++ - meson setup mesonbuild --cross-file=ios-cross-file.txt + - meson compile -C mesonbuild + - meson test -C mesonbuild + +-- +2.44.0.windows.1 + + +From 38847a0321bcb1d04ff09ccd410f0d6288b92c39 Mon Sep 17 00:00:00 2001 +From: "L. E. Segovia" <amy@centricular.com> +Date: Wed, 10 Jul 2024 18:53:54 -0300 +Subject: [PATCH] meson: Update to 1.14.1 + +Part-of: <https://gitlab.freedesktop.org/gstreamer/meson-ports/libvpx/-/merge_requests/21> + +diff --git a/build/make/rtcd.py b/build/make/rtcd.py +index 3d4121c57..cdff3dcc1 100755 +--- a/build/make/rtcd.py ++++ b/build/make/rtcd.py +@@ -422,7 +422,7 @@ if __name__ == '__main__': + ALL_ARCHS = filter(*qw('neon_asm neon')) + arm() + elif opts['arch'] == 'armv8' or opts['arch'] == 'arm64': +- ALL_ARCHS = filter(*qw('neon')) ++ ALL_ARCHS = filter(*qw('neon neon_dotprod neon_i8mm sve')) + REQUIRES = filter(*qw('neon')) + require(*REQUIRES) + arm() +diff --git a/examples/meson.build b/examples/meson.build +index 9b511686e..09e5fab2c 100644 +--- a/examples/meson.build ++++ b/examples/meson.build +@@ -232,6 +232,8 @@ if features.get('decoders', false) + vpxdec = executable( + 'vpxdec', + vpxdec_srcs, ++ c_args: project_c_args, ++ link_args: project_c_link_args, + extra_files: vpxdec_headers, + dependencies: vpxdec_deps, + install: features.get('install_bins', false), +diff --git a/meson.build b/meson.build +index 7366c61a7..198165e61 100644 +--- a/meson.build ++++ b/meson.build +@@ -4,13 +4,11 @@ + project( + 'libvpx', + 'c', 'cpp', +- version: '1.13.1', ++ version: '1.14.1', + default_options: [ +- # Enforce c89 for c files. Don't be too strict about it though. Allow +- # gnu extensions like "//" for comments. +- 'c_std=gnu89', +- # Uncomment when Meson can handle this one for MSVC. +- # 'cpp_std=gnu++11', ++ # Enforce C99 for C files. Allow GNU extensions. ++ 'c_std=c11', # Because MSVC ++ 'cpp_std=c++11', + 'warning_level=2', + 'default_library=static', + 'buildtype=debugoptimized', +@@ -85,6 +83,10 @@ features.set('spatial_resampling', true) + features.set('multithread', true) + features.set('os_support', true) + features.set('temporal_denoising', true) ++ ++if get_option('b_pie') and not get_option('b_staticpic') ++ error('PIC is required when building position independent executables') ++endif + features.set('pic', get_option('b_staticpic')) + + CODECS = [ +@@ -101,6 +103,7 @@ CODEC_FAMILIES = [ + + ARCH_LIST = [ + 'arm', ++ 'aarch64', + 'mips', + 'x86', + 'x86_64', +@@ -108,6 +111,13 @@ ARCH_LIST = [ + 'loongarch', + ] + ++ARCH_EXT_LIST_AARCH64 = [ ++ 'neon', ++ 'neon_dotprod', ++ 'neon_i8mm', ++ 'sve', ++] ++ + ARCH_EXT_LIST_X86 = [ + 'mmx', + 'sse', +@@ -127,9 +137,8 @@ ARCH_EXT_LIST_LOONGSON = [ + ] + + ARCH_EXT_LIST = [ +- 'neon', +- 'neon_asm', +- ++ 'neon_asm' ++] + ARCH_EXT_LIST_AARCH64 + [ + 'mips32', + 'dspr2', + 'msa', +@@ -199,7 +208,6 @@ CONFIG_LIST = [ + 'multi_res_encoding', + 'temporal_denoising', + 'vp9_temporal_denoising', +- 'consistent_recode', + 'coefficient_range_checking', + 'vp9_highbitdepth', + 'better_hw_compatibility', +@@ -265,7 +273,6 @@ CMDLINE_SELECT = [ + 'multi_res_encoding', + 'temporal_denoising', + 'vp9_temporal_denoising', +- 'consistent_recode', + 'coefficient_range_checking', + 'better_hw_compatibility', + 'vp9_highbitdepth', +@@ -525,6 +532,11 @@ c = meson.get_compiler('c') + cpp = meson.get_compiler('cpp') + asm = meson.get_compiler('c') # asm == GCC + ++# Since MSVC does not support C99, but the official configure expects it, ++# the Meson default takes C11. This sets the GNU extensions back online. ++add_project_arguments(c.get_supported_arguments('-std=gnu11'), language: 'c') ++add_project_arguments(cpp.get_supported_arguments('-std=gnu++11'), language: 'cpp') ++ + tgt_isa = host_machine.cpu_family() + tgt_os = host_machine.system() + tgt_cc = c.get_id() +@@ -612,7 +624,10 @@ endif + features.set(tgt_cc, true) + + # Enable the architecture family +-if tgt_isa.startswith('arm') ++if tgt_isa == 'arm64' ++ features.set('arm', true) ++ features.set('aarch64', true) ++elif tgt_isa.startswith('arm') + features.set('arm', true) + elif tgt_isa.startswith('mips') + features.set('mips', true) +@@ -637,7 +652,7 @@ if tgt_os == 'solaris' + endif + + if tgt_isa.startswith('arm') +- feature = 'neon' ++ feature = 'runtime_cpu_detect' + if features.get(feature, true) + if not features.has(feature) + message('\tenabling @0@'.format(feature)) +@@ -646,6 +661,13 @@ if tgt_isa.startswith('arm') + endif + + if tgt_isa == 'armv7' ++ feature = 'neon' ++ if features.get(feature, true) ++ if not features.has(feature) ++ message('\tenabling @0@'.format(feature)) ++ endif ++ features.set(feature, true) ++ endif + # Only enable neon_asm when neon is also enabled. + if features.get('neon', false) + feature = 'neon_asm' +@@ -772,6 +794,93 @@ if tgt_isa.startswith('arm') + elif tgt_os == 'linux' + features.set('linux', true) + endif ++ if tgt_isa == 'arm64' ++ aarch64_arch_flag_neon = 'arch=armv8-a' ++ aarch64_arch_flag_neon_dotprod = 'arch=armv8.2-a+dotprod' ++ aarch64_arch_flag_neon_i8mm = 'arch=armv8.2-a+dotprod+i8mm' ++ aarch64_arch_flag_sve = 'arch=armv8.2-a+dotprod+i8mm+sve' ++ ++ # Unline the original, we do can test cflags here :) ++ disable_exts = false ++ foreach feature : ARCH_EXT_LIST_AARCH64 ++ if disable_exts ++ rtcd_options += ['--disable', feature] ++ if not features.get(feature, false) # ! enabled ++ if not features.has(feature) # disabled ++ message('\tdisabling @0@'.format(feature)) ++ endif ++ features.set(feature, false) ++ endif ++ else ++ # Check the compiler supports the -march flag for the extension. ++ # This needs to happen after toolchain/OS inspection so we handle ++ # $CROSS etc correctly when checking for flags, else these will ++ # always fail. ++ flag = get_variable('aarch64_arch_flag_@0@'.format(feature)) ++ if c.has_argument('-m@0@'.format(flag)) ++ if features.get(feature, true) ++ if not features.has(feature) ++ message('\tenabling @0@'.format(feature)) ++ endif ++ features.set(feature, true) ++ endif ++ else ++ rtcd_options += ['--disable', feature] ++ if not features.get(feature, false) # ! enabled ++ if not features.has(feature) # disabled ++ message('\tdisabling @0@'.format(feature)) ++ endif ++ features.set(feature, false) ++ endif ++ endif ++ if not features.get(feature, false) ++ # Disable higher order extensions to simplify dependencies. ++ disable_exts = true ++ rtcd_options += ['--disable', feature] ++ if not features.get(feature, false) # ! enabled ++ if not features.has(feature) # disabled ++ message('\tdisabling @0@'.format(feature)) ++ endif ++ features.set(feature, false) ++ endif ++ endif ++ endif ++ endforeach ++ if features.get('sve', false) ++ supports_armv8_etc = c.compiles(''' ++ #ifndef __ARM_NEON_SVE_BRIDGE ++ #error 1 ++ #endif ++ #include <arm_sve.h> ++ #include <arm_neon_sve_bridge.h> ++ ''', ++ name: 'supports Armv8.2-a, dotprod, i8mm, and SVE', ++ args: '-march=armv8.2-a+dotprod+i8mm+sve' ++ ) ++ if supports_armv8_etc ++ # Check whether the compiler can compile SVE functions that require ++ # backup/restore of SVE registers according to AAPCS. Clang for Windows ++ # used to fail this, see ++ # https://github.com/llvm/llvm-project/issues/80009. ++ supports_armv8_etc = c.compiles(''' ++ #include <arm_sve.h> ++ void other(void); ++ svfloat32_t func(svfloat32_t a) { ++ other(); ++ return a; ++ } ++ ''', ++ name: 'has usable arm_sve.h', ++ args: '-march=armv8.2-a+dotprod+i8mm+sve' ++ ) ++ endif ++ if not supports_armv8_etc ++ warning('disabling SVE: arm_neon_sve_bridge.h not supported by compiler') ++ features.set('sve', false) ++ rtcd_options += ['--disable', 'sve'] ++ endif ++ endif ++ endif + elif tgt_isa.startswith('mips') + tune_cflags = '-mtune=@0@' + +@@ -1223,7 +1332,6 @@ endif + + if features.get('gcc', false) + gcc_flags = [ +- '-Wdeclaration-after-statement', + '-Wdisabled-optimization', + '-Wextra-semi', + '-Wextra-semi-stmt', +@@ -1237,8 +1345,9 @@ if features.get('gcc', false) + '-Wimplicit-function-declaration', + '-Wmissing-declarations', + '-Wmissing-prototypes', ++ '-Wshadow', + '-Wuninitialized', +- '-Wunreachable-code-loop-increment', ++ '-Wunreachable-code-aggressive', + '-Wunused', + ] + +@@ -1270,12 +1379,14 @@ if features.get('gcc', false) + '-Wc++14-extensions', + '-Wc++17-extensions', + '-Wc++20-extensions', ++ '-Wnon-virtual-dtor', + ) +- # disable some warnings specific to libyuv. ++ # disable some warnings specific to libyuv / libwebm. + libyuv_cpp_args += cpp.get_supported_arguments( + '-Wno-missing-declarations', + '-Wno-missing-prototypes', + '-Wno-pass-failed', ++ '-Wno-shadow', + '-Wno-unused-parameter', + ) + endif +@@ -1361,7 +1472,6 @@ if tgt_cc == 'msvs' + elif tgt_os == 'android' + enable_features = ['libyuv'] + if cpp.has_argument('-std=gnu++11') +- project_cpp_args += '-std=gnu++11' + enable_features += ['webm_io'] + endif + foreach feature: enable_features +@@ -1380,7 +1490,6 @@ elif tgt_os == 'darwin' + elif tgt_os == 'iphonesimulator' + enable_features = ['libyuv'] + if cpp.has_argument('-std=gnu++11') +- project_cpp_args += '-std=gnu++11' + enable_features += ['webm_io'] + endif + foreach feature: enable_features +@@ -1397,7 +1506,6 @@ elif target_machine.system() == 'windows' + # would be disabled for the same reason. + enable_features = [] + if cpp.has_argument('-std=gnu++11') +- project_cpp_args += '-std=gnu++11' + enable_features += ['unit_tests', 'webm_io'] + endif + if cpp.compiles('int z;') +@@ -1414,7 +1522,6 @@ elif target_machine.system() == 'windows' + else + enable_features = [] + if cpp.has_argument('-std=gnu++11') +- project_cpp_args += '-std=gnu++11' + if features.get('pthread_h', false) + enable_features += ['unit_tests'] + endif +@@ -1553,7 +1660,7 @@ else + endif + + vpx_config_data = configuration_data() +-vpx_config_data.set_quoted('CONFIGURE_ARGS', configure_args) ++vpx_config_data.set_quoted('CONFIGURE_ARGS', configure_args.replace('\\', '\\\\')) + + vpx_config_c = configure_file( + input: 'meson/vpx_config.c', +@@ -1564,7 +1671,7 @@ vpx_config_c = configure_file( + # libs.mk + + if features.get('shared', false) and not ['linux', 'darwin', 'ios', 'iphonesimulator'].contains(tgt_os) +- if features.enabled('gnu', false) ++ if features.get('gcc', false) # Looks like a typo on upstream + warn('Shared libraries are only supported on ELF; assuming this is OK') + else + error('Shared libraries are only supported on ELF, OS/2, and Darwin for now') +@@ -1611,6 +1718,11 @@ else + 'avx2': ['-mavx2'], + 'avx512': ['-mavx512f', '-mavx512cd', '-mavx512bw', '-mavx512dq', '-mavx512vl'], + ++ # AARCH64 ++ 'neon_dotprod': ['-march=armv8.2-a+dotprod'], ++ 'neon_i8mm': ['-march=armv8.2-a+dotprod+i8mm'], ++ 'sve': ['-march=armv8.2-a+dotprod+i8mm+sve'], ++ + # POWER + 'vsx' : ['-maltivec', '-mvsx'], + +@@ -1809,7 +1921,7 @@ else + link_args += ['-Wl,--version-script,@0@'.format(meson.current_build_dir() / vpx_def_name)] + endif + +-soversion = '8.0.0' ++soversion = '9.0.1' + + foreach arch, srcs : codec_arch_srcs + codec_arch_libs += static_library( +diff --git a/meson/parse_sources.py b/meson/parse_sources.py +index 5693aaddd..3b38eb2a5 100755 +--- a/meson/parse_sources.py ++++ b/meson/parse_sources.py +@@ -80,8 +80,10 @@ def make_to_meson(target: str, paths: list[str]): + else: + continue + +- if not source_type or not component or not label: ++ if not component: + raise RuntimeError('Unspecified input file data was found') ++ elif not source_type: ++ print(f'Skipping {ofiles}') + + accumulate = ofiles.endswith('\\') + ofiles = ofiles.strip('\\') +@@ -99,7 +101,7 @@ def make_to_meson(target: str, paths: list[str]): + component_sources[label] = component_sources.setdefault(label, list()) + accum + ifiles + accum = [] + +- if not label: ++ if not label and accum: + raise RuntimeError('Unspecified component type') + + # Makefiles can end with '\' and this is just a porting script ;) +@@ -145,7 +147,7 @@ def make_to_meson(target: str, paths: list[str]): + print ('Warning: skipping %s' % source) + continue + f.write(f"\t'{source}',\n") +- f.write('{file_closing}\n\n') ++ f.write(f'{file_closing}\n\n') + + f.write(f'{component}_{source_type}optional_sources = {{\n') + for label in sorted (component_sources): +diff --git a/test/meson.build b/test/meson.build +index 26a55d980..0c7319619 100644 +--- a/test/meson.build ++++ b/test/meson.build +@@ -7,10 +7,10 @@ + + libvpx_test_sources = files( + 'bench.cc', ++ 'init_vpx_test.cc', + 'test_libvpx.cc', +- '../md5_utils.c', # MANUAL +- '../y4minput.c', # MANUAL + 'test_vectors.cc', ++ 'vpx_image_test.cc', + 'decode_test_driver.cc', + # 'decode_perf_test.cc', # MANUAL + # 'encode_perf_test.cc', # MANUAL +@@ -41,10 +41,6 @@ libvpx_test_optional_sources = { + ), + 'encoders' : files( + 'altref_test.cc', +- 'aq_segment_test.cc', +- 'alt_ref_aq_segment_test.cc', +- 'vp8_datarate_test.cc', +- 'vp9_datarate_test.cc', + 'encode_api_test.cc', + 'error_resilience_test.cc', + 'realtime_test.cc', +@@ -74,6 +70,7 @@ libvpx_test_optional_sources = { + 'config_test.cc', + 'cq_test.cc', + 'keyframe_test.cc', ++ 'vp8_datarate_test.cc', + # 'quantize_test.cc', # MANUAL + # 'set_roi.cc', # MANUAL + # 'variance_test.cc', # MANUAL +@@ -91,6 +88,8 @@ libvpx_test_optional_sources = { + 'vp9_encoder' : files( + 'active_map_refresh_test.cc', + 'active_map_test.cc', ++ 'alt_ref_aq_segment_test.cc', ++ 'aq_segment_test.cc', + 'borders_test.cc', + 'cpu_speed_test.cc', + 'frame_size_tests.cc', +@@ -104,6 +103,7 @@ libvpx_test_optional_sources = { + 'svc_test.cc', + 'svc_end_to_end_test.cc', + 'timestamp_test.cc', ++ 'vp9_datarate_test.cc', + 'vp9_ext_ratectrl_test.cc', + # 'avg_test.cc', # MANUAL + # 'comp_avg_pred_test.cc', # MANUAL +@@ -143,6 +143,7 @@ simple_encode_test_optional_sources = { + + test_intra_pred_speed_sources = files( + # 'test_intra_pred_speed.cc', # MANUAL ++ 'init_vpx_test.cc', + ) + + test_intra_pred_speed_optional_sources = { +@@ -173,6 +174,7 @@ libvpx_test_headers_sources = files( + 'codec_factory.h', + 'md5_helper.h', + 'register_state_check.h', ++ 'init_vpx_test.h', + 'test_vectors.h', + 'util.h', + 'video_source.h', +@@ -210,6 +212,13 @@ libwebm_parser_headers_sources = files( + libwebm_parser_headers_optional_sources = { + } + ++test_intra_pred_speed_headers_sources = files( ++ 'init_vpx_test.h', ++) ++ ++test_intra_pred_speed_headers_optional_sources = { ++} ++ + rc_interface_test_headers_sources = files( + # 'decode_test_driver.h', # MANUAL + # 'codec_factory.h', # MANUAL +@@ -1095,6 +1104,7 @@ libvpx_test_data_data_optional_sources = { + 'screendata.y4m', + 'niklas_640_480_30.yuv', + 'bus_352x288_420_f20_b8.yuv', ++ 'crowd_run_360p_10_150f.y4m', + 'niklas_1280_720_30.yuv', + 'desktop_640_360_30.yuv', + 'kirland_640_480_30.yuv', +@@ -1107,6 +1117,9 @@ libvpx_test_data_data_optional_sources = { + ], + } + ++#### --- END GENERATED --- #### ++ ++ + if features.get('webm_io', false) # MANUAL + libvpx_test_optional_sources += { + 'decoders' : libvpx_test_optional_sources['decoders'] + files( +@@ -1432,8 +1445,6 @@ if features.get('encode_perf_tests', false) + } + endif + +-#### --- END GENERATED --- #### +- + optional_sources = get_variable('libvpx_test_optional_sources', {}) + foreach comp_name, comp_sources : optional_sources + if features.get(comp_name, false) +diff --git a/third_party/libyuv/meson.build b/third_party/libyuv/meson.build +index 78af1b775..f9d07f901 100644 +--- a/third_party/libyuv/meson.build ++++ b/third_party/libyuv/meson.build +@@ -2,6 +2,7 @@ + # SPDX-License-Identifier: BSD-3-Clause + + libyuv_includes = files( ++ 'include/libyuv/basic_types.h', + 'include/libyuv/convert.h', + 'include/libyuv/convert_argb.h', + 'include/libyuv/convert_from.h', +diff --git a/vp8/common/rtcd_defs.py b/vp8/common/rtcd_defs.py +index 4edaea4cd..0af078974 100755 +--- a/vp8/common/rtcd_defs.py ++++ b/vp8/common/rtcd_defs.py +@@ -130,12 +130,6 @@ specialize(*qw('vp8_copy_mem8x4 mmx neon dspr2 msa mmi')) + # + if vpx_config("CONFIG_POSTPROC") == "yes": + +- add_proto(*qw('void vp8_blend_mb_inner'), "unsigned char *y, unsigned char *u, unsigned char *v, int y_1, int u_1, int v_1, int alpha, int stride") +- +- add_proto(*qw('void vp8_blend_mb_outer'), "unsigned char *y, unsigned char *u, unsigned char *v, int y_1, int u_1, int v_1, int alpha, int stride") +- +- add_proto(*qw('void vp8_blend_b'), "unsigned char *y, unsigned char *u, unsigned char *v, int y_1, int u_1, int v_1, int alpha, int stride") +- + add_proto(*qw('void vp8_filter_by_weight16x16'), "unsigned char *src, int src_stride, unsigned char *dst, int dst_stride, int src_weight") + specialize(*qw('vp8_filter_by_weight16x16 sse2 msa')) + +@@ -220,14 +214,14 @@ if vpx_config("CONFIG_VP8_ENCODER") == "yes": + # + add_proto(*qw('int vp8_refining_search_sad'), "struct macroblock *x, struct block *b, struct blockd *d, union int_mv *ref_mv, int error_per_bit, int search_range, struct variance_vtable *fn_ptr, int *mvcost[2], union int_mv *center_mv") + specialize(*qw('vp8_refining_search_sad sse2 msa')) +- vp8_refining_search_sad_sse2='vp8_refining_search_sadx4'; +- vp8_refining_search_sad_msa='vp8_refining_search_sadx4'; ++ vp8_refining_search_sad_sse2='vp8_refining_search_sadx4' ++ vp8_refining_search_sad_msa='vp8_refining_search_sadx4' + + add_proto(*qw('int vp8_diamond_search_sad'), "struct macroblock *x, struct block *b, struct blockd *d, union int_mv *ref_mv, union int_mv *best_mv, int search_param, int sad_per_bit, int *num00, struct variance_vtable *fn_ptr, int *mvcost[2], union int_mv *center_mv") + specialize(*qw('vp8_diamond_search_sad sse2 msa lsx')) +- vp8_diamond_search_sad_sse2='vp8_diamond_search_sadx4'; +- vp8_diamond_search_sad_msa='vp8_diamond_search_sadx4'; +- vp8_diamond_search_sad_lsx='vp8_diamond_search_sadx4'; ++ vp8_diamond_search_sad_sse2='vp8_diamond_search_sadx4' ++ vp8_diamond_search_sad_msa='vp8_diamond_search_sadx4' ++ vp8_diamond_search_sad_lsx='vp8_diamond_search_sadx4' + + # + # Alt-ref Noise Reduction (ARNR) +diff --git a/vp8/meson.build b/vp8/meson.build +index 647460ae4..a9790e613 100644 +--- a/vp8/meson.build ++++ b/vp8/meson.build +@@ -327,6 +327,9 @@ vp8_dx_headers_optional_sources = { + ), + } + ++vp8_common_asm_sources = files( ++) ++ + vp8_common_asm_optional_sources = { + 'mmx' : files( + 'common/x86/dequantize_mmx.asm', +@@ -348,6 +351,9 @@ vp8_common_asm_optional_sources = { + ), + } + ++vp8_cx_asm_sources = files( ++) ++ + if features.get('postproc', false) # MANUAL + vp8_common_asm_optional_sources += { + 'sse2': vp8_common_asm_optional_sources['sse2'] + files( +diff --git a/vp9/common/vp9_rtcd_defs.py b/vp9/common/vp9_rtcd_defs.py +index 41fdd120b..ad65ae649 100755 +--- a/vp9/common/vp9_rtcd_defs.py ++++ b/vp9/common/vp9_rtcd_defs.py +@@ -27,7 +27,9 @@ struct macroblockd; + + /* Encoder forward decls */ + struct macroblock; +-struct vp9_variance_vtable; ++struct macroblock_plane; ++struct vp9_sad_table; ++struct ScanOrder; + struct search_site_config; + struct mv; + union int_mv; +@@ -121,24 +123,21 @@ if vpx_config("CONFIG_VP9_ENCODER") == "yes": + add_proto(*qw('int64_t vp9_block_error'), "const tran_low_t *coeff, const tran_low_t *dqcoeff, intptr_t block_size, int64_t *ssz") + + add_proto(*qw('int64_t vp9_block_error_fp'), "const tran_low_t *coeff, const tran_low_t *dqcoeff, int block_size") ++ specialize(*qw('vp9_block_error_fp neon avx2 sse2')) + +- add_proto(*qw('void vp9_quantize_fp'), "const tran_low_t *coeff_ptr, intptr_t n_coeffs, const int16_t *round_ptr, const int16_t *quant_ptr, tran_low_t *qcoeff_ptr, tran_low_t *dqcoeff_ptr, const int16_t *dequant_ptr, uint16_t *eob_ptr, const int16_t *scan, const int16_t *iscan") ++ add_proto(*qw('void vp9_quantize_fp'), "const tran_low_t *coeff_ptr, intptr_t n_coeffs, const struct macroblock_plane *const mb_plane, tran_low_t *qcoeff_ptr, tran_low_t *dqcoeff_ptr, const int16_t *dequant_ptr, uint16_t *eob_ptr, const struct ScanOrder *const scan_order") + specialize(*qw('vp9_quantize_fp neon sse2 ssse3 avx2 vsx')) + +- add_proto(*qw('void vp9_quantize_fp_32x32'), "const tran_low_t *coeff_ptr, intptr_t n_coeffs, const int16_t *round_ptr, const int16_t *quant_ptr, tran_low_t *qcoeff_ptr, tran_low_t *dqcoeff_ptr, const int16_t *dequant_ptr, uint16_t *eob_ptr, const int16_t *scan, const int16_t *iscan") ++ add_proto(*qw('void vp9_quantize_fp_32x32'), "const tran_low_t *coeff_ptr, intptr_t n_coeffs, const struct macroblock_plane *const mb_plane, tran_low_t *qcoeff_ptr, tran_low_t *dqcoeff_ptr, const int16_t *dequant_ptr, uint16_t *eob_ptr, const struct ScanOrder *const scan_order") + specialize(*qw('vp9_quantize_fp_32x32 neon ssse3 avx2 vsx')) + + if vpx_config("CONFIG_VP9_HIGHBITDEPTH") == "yes": +- specialize(*qw('vp9_block_error avx2 sse2')) +- +- specialize(*qw('vp9_block_error_fp avx2 sse2')) ++ specialize(*qw('vp9_block_error neon avx2 sse2')) + + add_proto(*qw('int64_t vp9_highbd_block_error'), "const tran_low_t *coeff, const tran_low_t *dqcoeff, intptr_t block_size, int64_t *ssz, int bd") +- specialize(*qw('vp9_highbd_block_error sse2')) ++ specialize(*qw('vp9_highbd_block_error neon sse2')) + else: +- specialize(*qw('vp9_block_error avx2 msa sse2')) +- +- specialize(*qw('vp9_block_error_fp neon avx2 sse2')) ++ specialize(*qw('vp9_block_error neon avx2 msa sse2')) + + # fdct functions + +@@ -166,15 +165,15 @@ if vpx_config("CONFIG_VP9_ENCODER") == "yes": + # + # Motion search + # +- add_proto(*qw('int vp9_diamond_search_sad'), "const struct macroblock *x, const struct search_site_config *cfg, struct mv *ref_mv, struct mv *best_mv, int search_param, int sad_per_bit, int *num00, const struct vp9_variance_vtable *fn_ptr, const struct mv *center_mv") +- specialize(*qw('vp9_diamond_search_sad avx neon')) ++ add_proto(*qw('int vp9_diamond_search_sad'), "const struct macroblock *x, const struct search_site_config *cfg, struct mv *ref_mv, uint32_t start_mv_sad, struct mv *best_mv, int search_param, int sad_per_bit, int *num00, const struct vp9_sad_table *sad_fn_ptr, const struct mv *center_mv") ++ specialize(*qw('vp9_diamond_search_sad neon')) + + # + # Apply temporal filter + # + if vpx_config("CONFIG_REALTIME_ONLY") != "yes": + add_proto(*qw('void vp9_apply_temporal_filter'), "const uint8_t *y_src, int y_src_stride, const uint8_t *y_pre, int y_pre_stride, const uint8_t *u_src, const uint8_t *v_src, int uv_src_stride, const uint8_t *u_pre, const uint8_t *v_pre, int uv_pre_stride, unsigned int block_width, unsigned int block_height, int ss_x, int ss_y, int strength, const int *const blk_fw, int use_32x32, uint32_t *y_accumulator, uint16_t *y_count, uint32_t *u_accumulator, uint16_t *u_count, uint32_t *v_accumulator, uint16_t *v_count") +- specialize(*qw('vp9_apply_temporal_filter sse4_1')) ++ specialize(*qw('vp9_apply_temporal_filter sse4_1 neon')) + + if vpx_config("CONFIG_VP9_HIGHBITDEPTH") == "yes": + add_proto(*qw('void vp9_highbd_apply_temporal_filter'), "const uint16_t *y_src, int y_src_stride, const uint16_t *y_pre, int y_pre_stride, const uint16_t *u_src, const uint16_t *v_src, int uv_src_stride, const uint16_t *u_pre, const uint16_t *v_pre, int uv_pre_stride, unsigned int block_width, unsigned int block_height, int ss_x, int ss_y, int strength, const int *const blk_fw, int use_32x32, uint32_t *y_accum, uint16_t *y_count, uint32_t *u_accum, uint16_t *u_count, uint32_t *v_accum, uint16_t *v_count") +@@ -184,10 +183,10 @@ if vpx_config("CONFIG_VP9_ENCODER") == "yes": + if vpx_config("CONFIG_VP9_HIGHBITDEPTH") == "yes": + # ENCODEMB INVOKE + +- add_proto(*qw('void vp9_highbd_quantize_fp'), "const tran_low_t *coeff_ptr, intptr_t n_coeffs, const int16_t *round_ptr, const int16_t *quant_ptr, tran_low_t *qcoeff_ptr, tran_low_t *dqcoeff_ptr, const int16_t *dequant_ptr, uint16_t *eob_ptr, const int16_t *scan, const int16_t *iscan") ++ add_proto(*qw('void vp9_highbd_quantize_fp'), "const tran_low_t *coeff_ptr, intptr_t n_coeffs, const struct macroblock_plane *const mb_plane, tran_low_t *qcoeff_ptr, tran_low_t *dqcoeff_ptr, const int16_t *dequant_ptr, uint16_t *eob_ptr, const struct ScanOrder *const scan_order") + specialize(*qw('vp9_highbd_quantize_fp avx2 neon')) + +- add_proto(*qw('void vp9_highbd_quantize_fp_32x32'), "const tran_low_t *coeff_ptr, intptr_t n_coeffs, const int16_t *round_ptr, const int16_t *quant_ptr, tran_low_t *qcoeff_ptr, tran_low_t *dqcoeff_ptr, const int16_t *dequant_ptr, uint16_t *eob_ptr, const int16_t *scan, const int16_t *iscan" ) ++ add_proto(*qw('void vp9_highbd_quantize_fp_32x32'), "const tran_low_t *coeff_ptr, intptr_t n_coeffs, const struct macroblock_plane *const mb_plane, tran_low_t *qcoeff_ptr, tran_low_t *dqcoeff_ptr, const int16_t *dequant_ptr, uint16_t *eob_ptr, const struct ScanOrder *const scan_order") + specialize(*qw('vp9_highbd_quantize_fp_32x32 avx2 neon')) + + # fdct functions +@@ -195,8 +194,10 @@ if vpx_config("CONFIG_VP9_ENCODER") == "yes": + specialize(*qw('vp9_highbd_fht4x4 neon')) + + add_proto(*qw('void vp9_highbd_fht8x8'), "const int16_t *input, tran_low_t *output, int stride, int tx_type") ++ specialize(*qw('vp9_highbd_fht8x8 neon')) + + add_proto(*qw('void vp9_highbd_fht16x16'), "const int16_t *input, tran_low_t *output, int stride, int tx_type") ++ specialize(*qw('vp9_highbd_fht16x16 neon')) + + add_proto(*qw('void vp9_highbd_fwht4x4'), "const int16_t *input, tran_low_t *output, int stride") + +@@ -211,4 +212,3 @@ if vpx_config("CONFIG_VP9_ENCODER") == "yes": + specialize(*qw('vp9_scale_and_extend_frame neon ssse3')) + + # end encoder functions +- +diff --git a/vp9/meson.build b/vp9/meson.build +index 7b7619575..e9314966c 100644 +--- a/vp9/meson.build ++++ b/vp9/meson.build +@@ -67,42 +67,6 @@ vp9_common_optional_sources = { + ), + } + +-if not features.get('vp9_highbitdepth', false) # MANUAL +- vp9_common_optional_sources += { +- 'msa': vp9_common_optional_sources['msa'] + files( +- 'common/mips/msa/vp9_idct4x4_msa.c', +- 'common/mips/msa/vp9_idct8x8_msa.c', +- 'common/mips/msa/vp9_idct16x16_msa.c', +- ), +- 'dspr2': vp9_common_optional_sources['dspr2'] + files( +- 'common/mips/dspr2/vp9_itrans4_dspr2.c', +- 'common/mips/dspr2/vp9_itrans8_dspr2.c', +- 'common/mips/dspr2/vp9_itrans16_dspr2.c' +- ) +- } +-else +- vp9_common_optional_sources += { +- 'neon': vp9_common_optional_sources['neon'] + files( +- 'common/arm/neon/vp9_highbd_iht4x4_add_neon.c', +- 'common/arm/neon/vp9_highbd_iht8x8_add_neon.c', +- 'common/arm/neon/vp9_highbd_iht16x16_add_neon.c' +- ), +- 'sse4_1': vp9_common_optional_sources['sse4_1'] + files( +- 'common/x86/vp9_highbd_iht4x4_add_sse4.c', +- 'common/x86/vp9_highbd_iht8x8_add_sse4.c', +- 'common/x86/vp9_highbd_iht16x16_add_sse4.c' +- ) +- } +-endif +- +-if features.get('vp9_postproc', false) # MANUAL +- vp9_common_optional_sources += { +- 'msa': vp9_common_optional_sources['msa'] + files( +- 'common/mips/msa/vp9_mfqe_msa.c' +- ) +- } +-endif +- + vp9_cx_sources = files( + 'vp9_cx_iface.c', + 'encoder/vp9_bitstream.c', +@@ -142,13 +106,11 @@ vp9_cx_sources = files( + 'encoder/vp9_noise_estimate.c', + 'encoder/vp9_ext_ratectrl.c', + # 'encoder/vp9_temporal_filter.c', # MANUAL ++ 'encoder/vp9_tpl_model.c', + # 'encoder/vp9_mbgraph.c', # MANUAL + ) + + vp9_cx_optional_sources = { +- 'avx' : files( +- 'encoder/x86/vp9_diamond_search_sad_avx.c' +- ), + 'avx2' : files( + 'encoder/x86/vp9_quantize_avx2.c', + 'encoder/x86/vp9_error_avx2.c' +@@ -164,12 +126,15 @@ vp9_cx_optional_sources = { + # 'encoder/mips/msa/vp9_fdct16x16_msa.c' # MANUAL + ), + 'neon' : files( ++ # 'encoder/arm/neon/vp9_temporal_filter_neon.c', # MANUAL + 'encoder/arm/neon/vp9_diamond_search_sad_neon.c', ++ # 'encoder/arm/neon/vp9_highbd_temporal_filter_neon.c', # MANUAL + 'encoder/arm/neon/vp9_dct_neon.c', + # 'encoder/arm/neon/vp9_denoiser_neon.c', # MANUAL + # 'encoder/arm/neon/vp9_error_neon.c', # MANUAL + 'encoder/arm/neon/vp9_frame_scale_neon.c', +- 'encoder/arm/neon/vp9_quantize_neon.c' ++ 'encoder/arm/neon/vp9_quantize_neon.c', ++ # 'encoder/arm/neon/vp9_highbd_error_neon.c', # MANUAL + ), + 'non_greedy_mv' : files( + 'encoder/vp9_non_greedy_mv.c' +@@ -260,6 +225,7 @@ vp9_cx_headers_sources = files( + 'encoder/vp9_encodemv.h', + 'encoder/vp9_extend.h', + 'encoder/vp9_firstpass.h', ++ 'encoder/vp9_firstpass_stats.h', + 'encoder/vp9_job_queue.h', + 'encoder/vp9_lookahead.h', + 'encoder/vp9_mcomp.h', +@@ -288,6 +254,7 @@ vp9_cx_headers_sources = files( + 'encoder/vp9_noise_estimate.h', + 'encoder/vp9_ext_ratectrl.h', + 'encoder/vp9_temporal_filter.h', ++ 'encoder/vp9_tpl_model.h', + 'encoder/vp9_mbgraph.h', + ) + +@@ -299,27 +266,20 @@ vp9_cx_headers_optional_sources = { + 'msa' : files( + # 'encoder/mips/msa/vp9_fdct_msa.h' # MANUAL + ), ++ 'neon' : files( ++ # 'encoder/vp9_temporal_filter_constants.h' # MANUAL ++ ), + 'non_greedy_mv' : files( + 'encoder/vp9_non_greedy_mv.h' + ), + 'sse4_1' : files( +- # 'encoder/x86/temporal_filter_constants.h' # MANUAL ++ # 'encoder/vp9_temporal_filter_constants.h' # MANUAL + ), + 'vp9_temporal_denoising' : files( + 'encoder/vp9_denoiser.h' + ), + } + +-vp9_cx_remove_headers_optional_sources = { +- 'realtime_only' : files( +- 'encoder/x86/temporal_filter_constants.h', +- 'encoder/vp9_alt_ref_aq.h', +- 'encoder/vp9_aq_variance.h', +- 'encoder/vp9_aq_360.h', +- 'encoder/vp9_aq_complexity.h' +- ), +-} +- + vp9_dx_headers_sources = files( + 'vp9_dx_iface.h', + 'decoder/vp9_decodeframe.h', +@@ -333,12 +293,27 @@ vp9_dx_headers_sources = files( + vp9_dx_headers_optional_sources = { + } + ++vp9_common_asm_sources = files( ++) ++ + vp9_common_asm_optional_sources = { + 'sse2' : files( + # 'common/x86/vp9_mfqe_sse2.asm' # MANUAL + ), + } + ++vp9_cx_asm_sources = files( ++ ) ++ ++vp9_cx_asm_optional_sources = { ++ 'sse2' : files( ++ 'encoder/x86/vp9_dct_sse2.asm', ++ 'encoder/x86/vp9_error_sse2.asm' ++ ), ++} ++ ++#### --- END GENERATED --- #### ++ + if features.get('vp9_postproc', false) # MANUAL + vp9_common_asm_optional_sources += { + 'sse2': vp9_common_asm_optional_sources['sse2'] + files( +@@ -347,12 +322,41 @@ if features.get('vp9_postproc', false) # MANUAL + } + endif + +-vp9_cx_asm_optional_sources = { +- 'sse2' : files( +- 'encoder/x86/vp9_dct_sse2.asm', +- 'encoder/x86/vp9_error_sse2.asm' +- ), +-} ++if not features.get('vp9_highbitdepth', false) # MANUAL ++ vp9_common_optional_sources += { ++ 'msa': vp9_common_optional_sources['msa'] + files( ++ 'common/mips/msa/vp9_idct4x4_msa.c', ++ 'common/mips/msa/vp9_idct8x8_msa.c', ++ 'common/mips/msa/vp9_idct16x16_msa.c', ++ ), ++ 'dspr2': vp9_common_optional_sources['dspr2'] + files( ++ 'common/mips/dspr2/vp9_itrans4_dspr2.c', ++ 'common/mips/dspr2/vp9_itrans8_dspr2.c', ++ 'common/mips/dspr2/vp9_itrans16_dspr2.c' ++ ) ++ } ++else ++ vp9_common_optional_sources += { ++ 'neon': vp9_common_optional_sources['neon'] + files( ++ 'common/arm/neon/vp9_highbd_iht4x4_add_neon.c', ++ 'common/arm/neon/vp9_highbd_iht8x8_add_neon.c', ++ 'common/arm/neon/vp9_highbd_iht16x16_add_neon.c' ++ ), ++ 'sse4_1': vp9_common_optional_sources['sse4_1'] + files( ++ 'common/x86/vp9_highbd_iht4x4_add_sse4.c', ++ 'common/x86/vp9_highbd_iht8x8_add_sse4.c', ++ 'common/x86/vp9_highbd_iht16x16_add_sse4.c' ++ ) ++ } ++endif ++ ++if features.get('vp9_postproc', false) # MANUAL ++ vp9_common_optional_sources += { ++ 'msa': vp9_common_optional_sources['msa'] + files( ++ 'common/mips/msa/vp9_mfqe_msa.c' ++ ) ++ } ++endif + + if features.get('vp9_temporal_denoising', false) + vp9_cx_optional_sources += { +@@ -370,6 +374,9 @@ if features.get('vp9_highbitdepth', false) + 'sse2' : vp9_cx_optional_sources['sse2'] + files( + 'encoder/x86/vp9_highbd_block_error_intrin_sse2.c', + ), ++ 'neon': vp9_cx_optional_sources['neon'] + files( ++ 'encoder/arm/neon/vp9_highbd_error_neon.c', ++ ), + } + else + vp9_cx_optional_sources += { +@@ -412,11 +419,17 @@ if not features.get('realtime_only', false) + 'sse4_1' : vp9_cx_optional_sources['sse4_1'] + files( + 'encoder/x86/temporal_filter_sse4.c', + ), ++ 'neon' : vp9_cx_optional_sources['neon'] + files( ++ 'encoder/arm/neon/vp9_temporal_filter_neon.c', ++ ), + } + + vp9_cx_headers_optional_sources += { + 'sse4_1' : vp9_cx_headers_optional_sources['sse4_1'] + files( +- 'encoder/x86/temporal_filter_constants.h' ++ 'encoder/vp9_temporal_filter_constants.h' ++ ), ++ 'neon' : vp9_cx_headers_optional_sources['neon'] + files( ++ 'encoder/vp9_temporal_filter_constants.h', + ), + } + +@@ -425,12 +438,13 @@ if not features.get('realtime_only', false) + 'sse4_1' : vp9_cx_optional_sources['sse4_1'] + files( + 'encoder/x86/highbd_temporal_filter_sse4.c', + ), ++ 'neon' : vp9_cx_optional_sources['neon'] + files( ++ 'encoder/arm/neon/vp9_highbd_temporal_filter_neon.c', ++ ), + } + endif + endif + +-#### --- END GENERATED --- #### +- + codec_srcs += vp9_common_sources + optional_sources = get_variable('vp9_common_optional_sources', {}) + foreach comp_name, comp_sources : optional_sources +diff --git a/vpx/meson.build b/vpx/meson.build +index 08f178d14..854e49f75 100644 +--- a/vpx/meson.build ++++ b/vpx/meson.build +@@ -10,14 +10,15 @@ api_sources = files( + 'src/vpx_encoder.c', + 'src/vpx_codec.c', + 'src/vpx_image.c', ++ 'src/vpx_tpl.c', + ) + + api_optional_sources = { + } + + api_headers_sources = files( +- 'vpx_decoder.h', +- 'vpx_encoder.h', ++ # 'vpx_decoder.h', # MANUAL ++ # 'vpx_encoder.h', # MANUAL + 'internal/vpx_codec_internal.h', + 'internal/vpx_ratectrl_rtc.h', + 'vpx_codec.h', +@@ -25,9 +26,17 @@ api_headers_sources = files( + 'vpx_image.h', + 'vpx_integer.h', + 'vpx_ext_ratectrl.h', ++ 'vpx_tpl.h', + ) + + api_headers_optional_sources = { ++ 'decoders': files( # MANUAL ++ 'vpx_tpl.h', ++ 'vpx_decoder.h', ++ ), ++ 'encoders': files( # MANUAL ++ 'vpx_encoder.h', ++ ), + 'vp8_decoder' : files( + 'vp8.h', + 'vp8dx.h' +@@ -45,6 +54,7 @@ api_doc_headers_sources = files( + 'vpx_ext_ratectrl.h', + 'vpx_frame_buffer.h', + 'vpx_image.h', ++ 'vpx_tpl.h', + ) + + api_doc_headers_optional_sources = { +diff --git a/vpx_dsp/meson.build b/vpx_dsp/meson.build +index d8c83297c..c25458a7c 100644 +--- a/vpx_dsp/meson.build ++++ b/vpx_dsp/meson.build +@@ -212,6 +212,8 @@ dsp_optional_sources = { + # 'arm/intrapred_neon_asm.asm', # MANUAL + # 'arm/save_reg_neon.asm' # MANUAL + ), ++ 'neon_dotprod' : files(), # MANUAL ++ 'neon_i8mm' : files(), # MANUAL + 'sse2' : files( + # 'x86/highbd_intrapred_intrin_sse2.c', # MANUAL + # 'x86/post_proc_sse2.c', # MANUAL +@@ -385,6 +387,7 @@ if features.get('encoders', false) # MANUAL + 'bitwriter.c', + 'bitwriter_buffer.c', + 'psnr.c', ++ 'sse.c', + ) + + dsp_headers_sources += files( +@@ -399,6 +402,18 @@ if features.get('encoders', false) # MANUAL + 'psnrhvs.c', + 'fastssim.c' + ), ++ 'neon' : dsp_optional_sources['neon'] + files( ++ 'arm/sse_neon.c', ++ ), ++ 'neon_dotprod' : dsp_optional_sources['neon_dotprod'] + files( ++ 'arm/sse_neon_dotprod.c', ++ ), ++ 'sse4_1' : dsp_optional_sources['sse4_1'] + files( ++ 'x86/sse_sse4.c', ++ ), ++ 'avx2' : dsp_optional_sources['avx2'] + files( ++ 'x86/sse_avx2.c', ++ ), + } + + dsp_headers_optional_sources += { +@@ -684,6 +699,16 @@ if features.get('vp9', false) # MANUAL + # loop filters + 'arm/loopfilter_neon.c', + ) ++ dsp_optional_sources += { ++ 'neon_dotprod' : dsp_optional_sources['neon_dotprod'] + files( ++ 'arm/vpx_convolve8_neon_dotprod.c', ++ 'arm/vpx_convolve_neon_dotprod.c', ++ ), ++ 'neon_i8mm' : dsp_optional_sources['neon_i8mm'] + files( ++ 'arm/vpx_convolve8_neon_i8mm.c', ++ 'arm/vpx_convolve_neon_i8mm.c', ++ ) ++ } + endif + endif + +@@ -718,6 +743,9 @@ if features.get('vp9_encoder', false) + 'sse2' : dsp_optional_sources['sse2'] + files( + 'x86/fwd_txfm_sse2.c', + ), ++ 'avx2': dsp_optional_sources['avx2'] + files( ++ 'x86/inv_txfm_avx2.c', ++ ), + 'neon': dsp_optional_sources['neon'] + files( + 'arm/fdct4x4_neon.c', + 'arm/fdct8x8_neon.c', +@@ -744,7 +772,13 @@ if features.get('vp9_encoder', false) + } + endif + +- if not features.get('vp9_highbitdepth', false) ++ if features.get('vp9_highbitdepth', false) ++ dsp_optional_sources += { ++ 'neon' : dsp_optional_sources['neon'] + files( ++ 'arm/highbd_avg_neon.c' ++ ), ++ } ++ else + dsp_optional_sources += { + 'avx2' : dsp_optional_sources['avx2'] + files( + 'x86/fwd_txfm_avx2.c' +@@ -979,6 +1013,8 @@ if features.get('vp9_encoder', false) # MANUAL + ), + 'neon' : dsp_optional_sources['neon'] + files( + 'arm/highbd_quantize_neon.c', ++ 'arm/highbd_hadamard_neon.c', ++ 'arm/highbd_avg_neon.c', + ), + } + endif +@@ -1007,6 +1043,10 @@ if features.get('encoders', false) # MANUAL + 'arm/sad_neon.c', + 'arm/subtract_neon.c', + ), ++ 'neon_dotprod' : dsp_optional_sources['neon_dotprod'] + files( ++ 'arm/sad4d_neon_dotprod.c', ++ 'arm/sad_neon_dotprod.c', ++ ), + 'sse2' : dsp_optional_sources['sse2'] + files( + 'x86/sum_squares_sse2.c', + ), +@@ -1059,7 +1099,9 @@ if features.get('encoders', false) # MANUAL + 'x86/highbd_sad_avx2.c', + ), + 'neon' : dsp_optional_sources['neon'] + files( ++ 'arm/highbd_sad4d_neon.c', + 'arm/highbd_sad_neon.c', ++ 'arm/highbd_subpel_variance_neon.c', + ) + } + endif +@@ -1081,6 +1123,9 @@ if features.get('encoders', false) or features.get('postproc', false) or feature + 'arm/subpel_variance_neon.c', + 'arm/variance_neon.c', + ), ++ 'neon_dotprod' : dsp_optional_sources['neon_dotprod'] + files( ++ 'arm/variance_neon_dotprod.c', ++ ), + 'msa' : dsp_optional_sources['msa'] + files( + 'mips/variance_msa.c', + 'mips/sub_pixel_variance_msa.c', +@@ -1098,6 +1143,7 @@ if features.get('encoders', false) or features.get('postproc', false) or feature + 'x86/variance_sse2.c', # Contains SSE2 and SSSE3 + ), + 'avx2' : dsp_optional_sources['avx2'] + files( ++ 'x86/avg_pred_avx2.c', + 'x86/variance_avx2.c', + ), + 'vsx' : dsp_optional_sources['vsx'] + files( +@@ -1140,6 +1186,8 @@ if features.get('encoders', false) or features.get('postproc', false) or feature + 'x86/highbd_variance_sse2.c', + ), + 'neon' : dsp_optional_sources['neon'] + files( ++ 'arm/highbd_avg_pred_neon.c', ++ 'arm/highbd_sse_neon.c', + 'arm/highbd_variance_neon.c', + ) + } +diff --git a/vpx_dsp/vpx_dsp_rtcd_defs.py b/vpx_dsp/vpx_dsp_rtcd_defs.py +index 03974bebd..17ddfde35 100755 +--- a/vpx_dsp/vpx_dsp_rtcd_defs.py ++++ b/vpx_dsp/vpx_dsp_rtcd_defs.py +@@ -21,6 +21,10 @@ def vpx_dsp_forward_decls(): + #include "vpx/vpx_integer.h" + #include "vpx_dsp/vpx_dsp_common.h" + #include "vpx_dsp/vpx_filter.h" ++#if CONFIG_VP9_ENCODER ++ struct macroblock_plane; ++ struct ScanOrder; ++#endif + + ''') + forward_decls(*qw('vpx_dsp_forward_decls')) +@@ -40,7 +44,7 @@ if opts["arch"] == "x86_64": + # + + add_proto(*qw('void vpx_d207_predictor_4x4'), "uint8_t *dst, ptrdiff_t stride, const uint8_t *above, const uint8_t *left") +-specialize(*qw('vpx_d207_predictor_4x4 sse2')) ++specialize(*qw('vpx_d207_predictor_4x4 neon sse2')) + + add_proto(*qw('void vpx_d45_predictor_4x4'), "uint8_t *dst, ptrdiff_t stride, const uint8_t *above, const uint8_t *left") + specialize(*qw('vpx_d45_predictor_4x4 neon sse2')) +@@ -48,7 +52,7 @@ specialize(*qw('vpx_d45_predictor_4x4 neon sse2')) + add_proto(*qw('void vpx_d45e_predictor_4x4'), "uint8_t *dst, ptrdiff_t stride, const uint8_t *above, const uint8_t *left") + + add_proto(*qw('void vpx_d63_predictor_4x4'), "uint8_t *dst, ptrdiff_t stride, const uint8_t *above, const uint8_t *left") +-specialize(*qw('vpx_d63_predictor_4x4 ssse3')) ++specialize(*qw('vpx_d63_predictor_4x4 neon ssse3')) + + add_proto(*qw('void vpx_d63e_predictor_4x4'), "uint8_t *dst, ptrdiff_t stride, const uint8_t *above, const uint8_t *left") + +@@ -59,12 +63,13 @@ specialize(*qw('vpx_h_predictor_4x4 neon dspr2 msa sse2')) + add_proto(*qw('void vpx_he_predictor_4x4'), "uint8_t *dst, ptrdiff_t stride, const uint8_t *above, const uint8_t *left") + + add_proto(*qw('void vpx_d117_predictor_4x4'), "uint8_t *dst, ptrdiff_t stride, const uint8_t *above, const uint8_t *left") ++specialize(*qw('vpx_d117_predictor_4x4 neon')) + + add_proto(*qw('void vpx_d135_predictor_4x4'), "uint8_t *dst, ptrdiff_t stride, const uint8_t *above, const uint8_t *left") + specialize(*qw('vpx_d135_predictor_4x4 neon')) + + add_proto(*qw('void vpx_d153_predictor_4x4'), "uint8_t *dst, ptrdiff_t stride, const uint8_t *above, const uint8_t *left") +-specialize(*qw('vpx_d153_predictor_4x4 ssse3')) ++specialize(*qw('vpx_d153_predictor_4x4 neon ssse3')) + + add_proto(*qw('void vpx_v_predictor_4x4'), "uint8_t *dst, ptrdiff_t stride, const uint8_t *above, const uint8_t *left") + specialize(*qw('vpx_v_predictor_4x4 neon msa sse2')) +@@ -88,7 +93,7 @@ add_proto(*qw('void vpx_dc_128_predictor_4x4'), "uint8_t *dst, ptrdiff_t stride, + specialize(*qw('vpx_dc_128_predictor_4x4 msa neon sse2')) + + add_proto(*qw('void vpx_d207_predictor_8x8'), "uint8_t *dst, ptrdiff_t stride, const uint8_t *above, const uint8_t *left") +-specialize(*qw('vpx_d207_predictor_8x8 ssse3')) ++specialize(*qw('vpx_d207_predictor_8x8 neon ssse3')) + + add_proto(*qw('void vpx_d45_predictor_8x8'), "uint8_t *dst, ptrdiff_t stride, const uint8_t *above, const uint8_t *left") + # TODO(crbug.com/webm/1522): Re-enable vsx implementation. +@@ -96,19 +101,20 @@ specialize(*qw('vpx_d45_predictor_8x8 neon sse2')) + + add_proto(*qw('void vpx_d63_predictor_8x8'), "uint8_t *dst, ptrdiff_t stride, const uint8_t *above, const uint8_t *left") + # TODO(crbug.com/webm/1522): Re-enable vsx implementation. +-specialize(*qw('vpx_d63_predictor_8x8 ssse3')) ++specialize(*qw('vpx_d63_predictor_8x8 neon ssse3')) + + add_proto(*qw('void vpx_h_predictor_8x8'), "uint8_t *dst, ptrdiff_t stride, const uint8_t *above, const uint8_t *left") + # TODO(crbug.com/webm/1522): Re-enable vsx implementation. + specialize(*qw('vpx_h_predictor_8x8 neon dspr2 msa sse2')) + + add_proto(*qw('void vpx_d117_predictor_8x8'), "uint8_t *dst, ptrdiff_t stride, const uint8_t *above, const uint8_t *left") ++specialize(*qw('vpx_d117_predictor_8x8 neon')) + + add_proto(*qw('void vpx_d135_predictor_8x8'), "uint8_t *dst, ptrdiff_t stride, const uint8_t *above, const uint8_t *left") + specialize(*qw('vpx_d135_predictor_8x8 neon')) + + add_proto(*qw('void vpx_d153_predictor_8x8'), "uint8_t *dst, ptrdiff_t stride, const uint8_t *above, const uint8_t *left") +-specialize(*qw('vpx_d153_predictor_8x8 ssse3')) ++specialize(*qw('vpx_d153_predictor_8x8 neon ssse3')) + + add_proto(*qw('void vpx_v_predictor_8x8'), "uint8_t *dst, ptrdiff_t stride, const uint8_t *above, const uint8_t *left") + specialize(*qw('vpx_v_predictor_8x8 neon msa sse2')) +@@ -131,24 +137,25 @@ add_proto(*qw('void vpx_dc_128_predictor_8x8'), "uint8_t *dst, ptrdiff_t stride, + specialize(*qw('vpx_dc_128_predictor_8x8 neon msa sse2')) + + add_proto(*qw('void vpx_d207_predictor_16x16'), "uint8_t *dst, ptrdiff_t stride, const uint8_t *above, const uint8_t *left") +-specialize(*qw('vpx_d207_predictor_16x16 ssse3')) ++specialize(*qw('vpx_d207_predictor_16x16 neon ssse3')) + + add_proto(*qw('void vpx_d45_predictor_16x16'), "uint8_t *dst, ptrdiff_t stride, const uint8_t *above, const uint8_t *left") + specialize(*qw('vpx_d45_predictor_16x16 neon ssse3 vsx')) + + add_proto(*qw('void vpx_d63_predictor_16x16'), "uint8_t *dst, ptrdiff_t stride, const uint8_t *above, const uint8_t *left") +-specialize(*qw('vpx_d63_predictor_16x16 ssse3 vsx')) ++specialize(*qw('vpx_d63_predictor_16x16 neon ssse3 vsx')) + + add_proto(*qw('void vpx_h_predictor_16x16'), "uint8_t *dst, ptrdiff_t stride, const uint8_t *above, const uint8_t *left") + specialize(*qw('vpx_h_predictor_16x16 neon dspr2 msa sse2 vsx')) + + add_proto(*qw('void vpx_d117_predictor_16x16'), "uint8_t *dst, ptrdiff_t stride, const uint8_t *above, const uint8_t *left") ++specialize(*qw('vpx_d117_predictor_16x16 neon')) + + add_proto(*qw('void vpx_d135_predictor_16x16'), "uint8_t *dst, ptrdiff_t stride, const uint8_t *above, const uint8_t *left") + specialize(*qw('vpx_d135_predictor_16x16 neon')) + + add_proto(*qw('void vpx_d153_predictor_16x16'), "uint8_t *dst, ptrdiff_t stride, const uint8_t *above, const uint8_t *left") +-specialize(*qw('vpx_d153_predictor_16x16 ssse3')) ++specialize(*qw('vpx_d153_predictor_16x16 neon ssse3')) + + add_proto(*qw('void vpx_v_predictor_16x16'), "uint8_t *dst, ptrdiff_t stride, const uint8_t *above, const uint8_t *left") + specialize(*qw('vpx_v_predictor_16x16 neon msa sse2 vsx')) +@@ -169,24 +176,25 @@ add_proto(*qw('void vpx_dc_128_predictor_16x16'), "uint8_t *dst, ptrdiff_t strid + specialize(*qw('vpx_dc_128_predictor_16x16 neon msa sse2 vsx')) + + add_proto(*qw('void vpx_d207_predictor_32x32'), "uint8_t *dst, ptrdiff_t stride, const uint8_t *above, const uint8_t *left") +-specialize(*qw('vpx_d207_predictor_32x32 ssse3')) ++specialize(*qw('vpx_d207_predictor_32x32 neon ssse3')) + + add_proto(*qw('void vpx_d45_predictor_32x32'), "uint8_t *dst, ptrdiff_t stride, const uint8_t *above, const uint8_t *left") + specialize(*qw('vpx_d45_predictor_32x32 neon ssse3 vsx')) + + add_proto(*qw('void vpx_d63_predictor_32x32'), "uint8_t *dst, ptrdiff_t stride, const uint8_t *above, const uint8_t *left") +-specialize(*qw('vpx_d63_predictor_32x32 ssse3 vsx')) ++specialize(*qw('vpx_d63_predictor_32x32 neon ssse3 vsx')) + + add_proto(*qw('void vpx_h_predictor_32x32'), "uint8_t *dst, ptrdiff_t stride, const uint8_t *above, const uint8_t *left") + specialize(*qw('vpx_h_predictor_32x32 neon msa sse2 vsx')) + + add_proto(*qw('void vpx_d117_predictor_32x32'), "uint8_t *dst, ptrdiff_t stride, const uint8_t *above, const uint8_t *left") ++specialize(*qw('vpx_d117_predictor_32x32 neon')) + + add_proto(*qw('void vpx_d135_predictor_32x32'), "uint8_t *dst, ptrdiff_t stride, const uint8_t *above, const uint8_t *left") + specialize(*qw('vpx_d135_predictor_32x32 neon')) + + add_proto(*qw('void vpx_d153_predictor_32x32'), "uint8_t *dst, ptrdiff_t stride, const uint8_t *above, const uint8_t *left") +-specialize(*qw('vpx_d153_predictor_32x32 ssse3')) ++specialize(*qw('vpx_d153_predictor_32x32 neon ssse3')) + + add_proto(*qw('void vpx_v_predictor_32x32'), "uint8_t *dst, ptrdiff_t stride, const uint8_t *above, const uint8_t *left") + specialize(*qw('vpx_v_predictor_32x32 neon msa sse2 vsx')) +@@ -209,25 +217,25 @@ specialize(*qw('vpx_dc_128_predictor_32x32 msa neon sse2 vsx')) + # High bitdepth functions + if vpx_config("CONFIG_VP9_HIGHBITDEPTH") == "yes": + add_proto(*qw('void vpx_highbd_d207_predictor_4x4'), "uint16_t *dst, ptrdiff_t stride, const uint16_t *above, const uint16_t *left, int bd") +- specialize(*qw('vpx_highbd_d207_predictor_4x4 sse2')) ++ specialize(*qw('vpx_highbd_d207_predictor_4x4 neon sse2')) + + add_proto(*qw('void vpx_highbd_d45_predictor_4x4'), "uint16_t *dst, ptrdiff_t stride, const uint16_t *above, const uint16_t *left, int bd") + specialize(*qw('vpx_highbd_d45_predictor_4x4 neon ssse3')) + + add_proto(*qw('void vpx_highbd_d63_predictor_4x4'), "uint16_t *dst, ptrdiff_t stride, const uint16_t *above, const uint16_t *left, int bd") +- specialize(*qw('vpx_highbd_d63_predictor_4x4 sse2')) ++ specialize(*qw('vpx_highbd_d63_predictor_4x4 neon sse2')) + + add_proto(*qw('void vpx_highbd_h_predictor_4x4'), "uint16_t *dst, ptrdiff_t stride, const uint16_t *above, const uint16_t *left, int bd") + specialize(*qw('vpx_highbd_h_predictor_4x4 neon sse2')) + + add_proto(*qw('void vpx_highbd_d117_predictor_4x4'), "uint16_t *dst, ptrdiff_t stride, const uint16_t *above, const uint16_t *left, int bd") +- specialize(*qw('vpx_highbd_d117_predictor_4x4 sse2')) ++ specialize(*qw('vpx_highbd_d117_predictor_4x4 neon sse2')) + + add_proto(*qw('void vpx_highbd_d135_predictor_4x4'), "uint16_t *dst, ptrdiff_t stride, const uint16_t *above, const uint16_t *left, int bd") + specialize(*qw('vpx_highbd_d135_predictor_4x4 neon sse2')) + + add_proto(*qw('void vpx_highbd_d153_predictor_4x4'), "uint16_t *dst, ptrdiff_t stride, const uint16_t *above, const uint16_t *left, int bd") +- specialize(*qw('vpx_highbd_d153_predictor_4x4 sse2')) ++ specialize(*qw('vpx_highbd_d153_predictor_4x4 neon sse2')) + + add_proto(*qw('void vpx_highbd_v_predictor_4x4'), "uint16_t *dst, ptrdiff_t stride, const uint16_t *above, const uint16_t *left, int bd") + specialize(*qw('vpx_highbd_v_predictor_4x4 neon sse2')) +@@ -248,25 +256,25 @@ if vpx_config("CONFIG_VP9_HIGHBITDEPTH") == "yes": + specialize(*qw('vpx_highbd_dc_128_predictor_4x4 neon sse2')) + + add_proto(*qw('void vpx_highbd_d207_predictor_8x8'), "uint16_t *dst, ptrdiff_t stride, const uint16_t *above, const uint16_t *left, int bd") +- specialize(*qw('vpx_highbd_d207_predictor_8x8 ssse3')) ++ specialize(*qw('vpx_highbd_d207_predictor_8x8 neon ssse3')) + + add_proto(*qw('void vpx_highbd_d45_predictor_8x8'), "uint16_t *dst, ptrdiff_t stride, const uint16_t *above, const uint16_t *left, int bd") + specialize(*qw('vpx_highbd_d45_predictor_8x8 neon ssse3')) + + add_proto(*qw('void vpx_highbd_d63_predictor_8x8'), "uint16_t *dst, ptrdiff_t stride, const uint16_t *above, const uint16_t *left, int bd") +- specialize(*qw('vpx_highbd_d63_predictor_8x8 ssse3')) ++ specialize(*qw('vpx_highbd_d63_predictor_8x8 neon ssse3')) + + add_proto(*qw('void vpx_highbd_h_predictor_8x8'), "uint16_t *dst, ptrdiff_t stride, const uint16_t *above, const uint16_t *left, int bd") + specialize(*qw('vpx_highbd_h_predictor_8x8 neon sse2')) + + add_proto(*qw('void vpx_highbd_d117_predictor_8x8'), "uint16_t *dst, ptrdiff_t stride, const uint16_t *above, const uint16_t *left, int bd") +- specialize(*qw('vpx_highbd_d117_predictor_8x8 ssse3')) ++ specialize(*qw('vpx_highbd_d117_predictor_8x8 neon ssse3')) + + add_proto(*qw('void vpx_highbd_d135_predictor_8x8'), "uint16_t *dst, ptrdiff_t stride, const uint16_t *above, const uint16_t *left, int bd") + specialize(*qw('vpx_highbd_d135_predictor_8x8 neon ssse3')) + + add_proto(*qw('void vpx_highbd_d153_predictor_8x8'), "uint16_t *dst, ptrdiff_t stride, const uint16_t *above, const uint16_t *left, int bd") +- specialize(*qw('vpx_highbd_d153_predictor_8x8 ssse3')) ++ specialize(*qw('vpx_highbd_d153_predictor_8x8 neon ssse3')) + + add_proto(*qw('void vpx_highbd_v_predictor_8x8'), "uint16_t *dst, ptrdiff_t stride, const uint16_t *above, const uint16_t *left, int bd") + specialize(*qw('vpx_highbd_v_predictor_8x8 neon sse2')) +@@ -287,25 +295,25 @@ if vpx_config("CONFIG_VP9_HIGHBITDEPTH") == "yes": + specialize(*qw('vpx_highbd_dc_128_predictor_8x8 neon sse2')) + + add_proto(*qw('void vpx_highbd_d207_predictor_16x16'), "uint16_t *dst, ptrdiff_t stride, const uint16_t *above, const uint16_t *left, int bd") +- specialize(*qw('vpx_highbd_d207_predictor_16x16 ssse3')) ++ specialize(*qw('vpx_highbd_d207_predictor_16x16 neon ssse3')) + + add_proto(*qw('void vpx_highbd_d45_predictor_16x16'), "uint16_t *dst, ptrdiff_t stride, const uint16_t *above, const uint16_t *left, int bd") + specialize(*qw('vpx_highbd_d45_predictor_16x16 neon ssse3')) + + add_proto(*qw('void vpx_highbd_d63_predictor_16x16'), "uint16_t *dst, ptrdiff_t stride, const uint16_t *above, const uint16_t *left, int bd") +- specialize(*qw('vpx_highbd_d63_predictor_16x16 ssse3')) ++ specialize(*qw('vpx_highbd_d63_predictor_16x16 neon ssse3')) + + add_proto(*qw('void vpx_highbd_h_predictor_16x16'), "uint16_t *dst, ptrdiff_t stride, const uint16_t *above, const uint16_t *left, int bd") + specialize(*qw('vpx_highbd_h_predictor_16x16 neon sse2')) + + add_proto(*qw('void vpx_highbd_d117_predictor_16x16'), "uint16_t *dst, ptrdiff_t stride, const uint16_t *above, const uint16_t *left, int bd") +- specialize(*qw('vpx_highbd_d117_predictor_16x16 ssse3')) ++ specialize(*qw('vpx_highbd_d117_predictor_16x16 neon ssse3')) + + add_proto(*qw('void vpx_highbd_d135_predictor_16x16'), "uint16_t *dst, ptrdiff_t stride, const uint16_t *above, const uint16_t *left, int bd") + specialize(*qw('vpx_highbd_d135_predictor_16x16 neon ssse3')) + + add_proto(*qw('void vpx_highbd_d153_predictor_16x16'), "uint16_t *dst, ptrdiff_t stride, const uint16_t *above, const uint16_t *left, int bd") +- specialize(*qw('vpx_highbd_d153_predictor_16x16 ssse3')) ++ specialize(*qw('vpx_highbd_d153_predictor_16x16 neon ssse3')) + + add_proto(*qw('void vpx_highbd_v_predictor_16x16'), "uint16_t *dst, ptrdiff_t stride, const uint16_t *above, const uint16_t *left, int bd") + specialize(*qw('vpx_highbd_v_predictor_16x16 neon sse2')) +@@ -326,25 +334,25 @@ if vpx_config("CONFIG_VP9_HIGHBITDEPTH") == "yes": + specialize(*qw('vpx_highbd_dc_128_predictor_16x16 neon sse2')) + + add_proto(*qw('void vpx_highbd_d207_predictor_32x32'), "uint16_t *dst, ptrdiff_t stride, const uint16_t *above, const uint16_t *left, int bd") +- specialize(*qw('vpx_highbd_d207_predictor_32x32 ssse3')) ++ specialize(*qw('vpx_highbd_d207_predictor_32x32 neon ssse3')) + + add_proto(*qw('void vpx_highbd_d45_predictor_32x32'), "uint16_t *dst, ptrdiff_t stride, const uint16_t *above, const uint16_t *left, int bd") + specialize(*qw('vpx_highbd_d45_predictor_32x32 neon ssse3')) + + add_proto(*qw('void vpx_highbd_d63_predictor_32x32'), "uint16_t *dst, ptrdiff_t stride, const uint16_t *above, const uint16_t *left, int bd") +- specialize(*qw('vpx_highbd_d63_predictor_32x32 ssse3')) ++ specialize(*qw('vpx_highbd_d63_predictor_32x32 neon ssse3')) + + add_proto(*qw('void vpx_highbd_h_predictor_32x32'), "uint16_t *dst, ptrdiff_t stride, const uint16_t *above, const uint16_t *left, int bd") + specialize(*qw('vpx_highbd_h_predictor_32x32 neon sse2')) + + add_proto(*qw('void vpx_highbd_d117_predictor_32x32'), "uint16_t *dst, ptrdiff_t stride, const uint16_t *above, const uint16_t *left, int bd") +- specialize(*qw('vpx_highbd_d117_predictor_32x32 ssse3')) ++ specialize(*qw('vpx_highbd_d117_predictor_32x32 neon ssse3')) + + add_proto(*qw('void vpx_highbd_d135_predictor_32x32'), "uint16_t *dst, ptrdiff_t stride, const uint16_t *above, const uint16_t *left, int bd") + specialize(*qw('vpx_highbd_d135_predictor_32x32 neon ssse3')) + + add_proto(*qw('void vpx_highbd_d153_predictor_32x32'), "uint16_t *dst, ptrdiff_t stride, const uint16_t *above, const uint16_t *left, int bd") +- specialize(*qw('vpx_highbd_d153_predictor_32x32 ssse3')) ++ specialize(*qw('vpx_highbd_d153_predictor_32x32 neon ssse3')) + + add_proto(*qw('void vpx_highbd_v_predictor_32x32'), "uint16_t *dst, ptrdiff_t stride, const uint16_t *above, const uint16_t *left, int bd") + specialize(*qw('vpx_highbd_v_predictor_32x32 neon sse2')) +@@ -376,22 +384,22 @@ if vpx_config("CONFIG_VP9") == "yes": + specialize(*qw('vpx_convolve_avg neon dspr2 msa sse2 vsx mmi lsx')) + + add_proto(*qw('void vpx_convolve8'), "const uint8_t *src, ptrdiff_t src_stride, uint8_t *dst, ptrdiff_t dst_stride, const InterpKernel *filter, int x0_q4, int x_step_q4, int y0_q4, int y_step_q4, int w, int h") +- specialize(*qw('vpx_convolve8 sse2 ssse3 avx2 neon dspr2 msa vsx mmi lsx')) ++ specialize(*qw('vpx_convolve8 sse2 ssse3 avx2 neon neon_dotprod neon_i8mm dspr2 msa vsx mmi lsx')) + + add_proto(*qw('void vpx_convolve8_horiz'), "const uint8_t *src, ptrdiff_t src_stride, uint8_t *dst, ptrdiff_t dst_stride, const InterpKernel *filter, int x0_q4, int x_step_q4, int y0_q4, int y_step_q4, int w, int h") +- specialize(*qw('vpx_convolve8_horiz sse2 ssse3 avx2 neon dspr2 msa vsx mmi lsx')) ++ specialize(*qw('vpx_convolve8_horiz sse2 ssse3 avx2 neon neon_dotprod neon_i8mm dspr2 msa vsx mmi lsx')) + + add_proto(*qw('void vpx_convolve8_vert'), "const uint8_t *src, ptrdiff_t src_stride, uint8_t *dst, ptrdiff_t dst_stride, const InterpKernel *filter, int x0_q4, int x_step_q4, int y0_q4, int y_step_q4, int w, int h") +- specialize(*qw('vpx_convolve8_vert sse2 ssse3 avx2 neon dspr2 msa vsx mmi lsx')) ++ specialize(*qw('vpx_convolve8_vert sse2 ssse3 avx2 neon neon_dotprod neon_i8mm dspr2 msa vsx mmi lsx')) + + add_proto(*qw('void vpx_convolve8_avg'), "const uint8_t *src, ptrdiff_t src_stride, uint8_t *dst, ptrdiff_t dst_stride, const InterpKernel *filter, int x0_q4, int x_step_q4, int y0_q4, int y_step_q4, int w, int h") +- specialize(*qw('vpx_convolve8_avg sse2 ssse3 avx2 neon dspr2 msa vsx mmi lsx')) ++ specialize(*qw('vpx_convolve8_avg sse2 ssse3 avx2 neon neon_dotprod neon_i8mm dspr2 msa vsx mmi lsx')) + + add_proto(*qw('void vpx_convolve8_avg_horiz'), "const uint8_t *src, ptrdiff_t src_stride, uint8_t *dst, ptrdiff_t dst_stride, const InterpKernel *filter, int x0_q4, int x_step_q4, int y0_q4, int y_step_q4, int w, int h") +- specialize(*qw('vpx_convolve8_avg_horiz sse2 ssse3 avx2 neon dspr2 msa vsx mmi lsx')) ++ specialize(*qw('vpx_convolve8_avg_horiz sse2 ssse3 avx2 neon neon_dotprod neon_i8mm dspr2 msa vsx mmi lsx')) + + add_proto(*qw('void vpx_convolve8_avg_vert'), "const uint8_t *src, ptrdiff_t src_stride, uint8_t *dst, ptrdiff_t dst_stride, const InterpKernel *filter, int x0_q4, int x_step_q4, int y0_q4, int y_step_q4, int w, int h") +- specialize(*qw('vpx_convolve8_avg_vert sse2 ssse3 avx2 neon dspr2 msa vsx mmi lsx')) ++ specialize(*qw('vpx_convolve8_avg_vert sse2 ssse3 avx2 neon neon_dotprod neon_i8mm dspr2 msa vsx mmi lsx')) + + add_proto(*qw('void vpx_scaled_2d'), "const uint8_t *src, ptrdiff_t src_stride, uint8_t *dst, ptrdiff_t dst_stride, const InterpKernel *filter, int x0_q4, int x_step_q4, int y0_q4, int y_step_q4, int w, int h") + specialize(*qw('vpx_scaled_2d ssse3 neon msa')) +@@ -591,7 +599,7 @@ if vpx_config("CONFIG_VP9_ENCODER") == "yes": + specialize(*qw('vpx_fdct8x8_1 sse2 neon msa')) + + add_proto(*qw('void vpx_fdct16x16'), "const int16_t *input, tran_low_t *output, int stride") +- specialize(*qw('vpx_fdct16x16 neon sse2 msa lsx')) ++ specialize(*qw('vpx_fdct16x16 neon sse2 avx2 msa lsx')) + + add_proto(*qw('void vpx_fdct16x16_1'), "const int16_t *input, tran_low_t *output, int stride") + specialize(*qw('vpx_fdct16x16_1 sse2 neon msa')) +@@ -635,12 +643,12 @@ if vpx_config("CONFIG_VP9") == "yes": + specialize(*qw('vpx_idct8x8_64_add neon sse2 vsx')) + specialize(*qw('vpx_idct8x8_12_add neon sse2 ssse3')) + specialize(*qw('vpx_idct8x8_1_add neon sse2')) +- specialize(*qw('vpx_idct16x16_256_add neon sse2 vsx')) ++ specialize(*qw('vpx_idct16x16_256_add neon sse2 avx2 vsx')) + specialize(*qw('vpx_idct16x16_38_add neon sse2')) + specialize(*qw('vpx_idct16x16_10_add neon sse2')) + specialize(*qw('vpx_idct16x16_1_add neon sse2')) +- specialize(*qw('vpx_idct32x32_1024_add neon sse2 vsx')) +- specialize(*qw('vpx_idct32x32_135_add neon sse2 ssse3')) ++ specialize(*qw('vpx_idct32x32_1024_add neon sse2 avx2 vsx')) ++ specialize(*qw('vpx_idct32x32_135_add neon sse2 ssse3 avx2')) + specialize(*qw('vpx_idct32x32_34_add neon sse2 ssse3')) + specialize(*qw('vpx_idct32x32_1_add neon sse2')) + specialize(*qw('vpx_iwht4x4_16_add sse2 vsx')) +@@ -716,17 +724,17 @@ if vpx_config("CONFIG_VP9") == "yes": + # Quantization + # + if vpx_config("CONFIG_VP9_ENCODER") == "yes": +- add_proto(*qw('void vpx_quantize_b'), "const tran_low_t *coeff_ptr, intptr_t n_coeffs, const int16_t *zbin_ptr, const int16_t *round_ptr, const int16_t *quant_ptr, const int16_t *quant_shift_ptr, tran_low_t *qcoeff_ptr, tran_low_t *dqcoeff_ptr, const int16_t *dequant_ptr, uint16_t *eob_ptr, const int16_t *scan, const int16_t *iscan") ++ add_proto(*qw('void vpx_quantize_b'), "const tran_low_t *coeff_ptr, intptr_t n_coeffs, const struct macroblock_plane *const mb_plane, tran_low_t *qcoeff_ptr, tran_low_t *dqcoeff_ptr, const int16_t *dequant_ptr, uint16_t *eob_ptr, const struct ScanOrder *const scan_order") + specialize(*qw('vpx_quantize_b neon sse2 ssse3 avx avx2 vsx lsx')) + +- add_proto(*qw('void vpx_quantize_b_32x32'), "const tran_low_t *coeff_ptr, intptr_t n_coeffs, const int16_t *zbin_ptr, const int16_t *round_ptr, const int16_t *quant_ptr, const int16_t *quant_shift_ptr, tran_low_t *qcoeff_ptr, tran_low_t *dqcoeff_ptr, const int16_t *dequant_ptr, uint16_t *eob_ptr, const int16_t *scan, const int16_t *iscan") ++ add_proto(*qw('void vpx_quantize_b_32x32'), "const tran_low_t *coeff_ptr, const struct macroblock_plane *const mb_plane, tran_low_t *qcoeff_ptr, tran_low_t *dqcoeff_ptr, const int16_t *dequant_ptr, uint16_t *eob_ptr, const struct ScanOrder *const scan_order") + specialize(*qw('vpx_quantize_b_32x32 neon ssse3 avx avx2 vsx lsx')) + + if vpx_config("CONFIG_VP9_HIGHBITDEPTH") == "yes": +- add_proto(*qw('void vpx_highbd_quantize_b'), "const tran_low_t *coeff_ptr, intptr_t n_coeffs, const int16_t *zbin_ptr, const int16_t *round_ptr, const int16_t *quant_ptr, const int16_t *quant_shift_ptr, tran_low_t *qcoeff_ptr, tran_low_t *dqcoeff_ptr, const int16_t *dequant_ptr, uint16_t *eob_ptr, const int16_t *scan, const int16_t *iscan") ++ add_proto(*qw('void vpx_highbd_quantize_b'), "const tran_low_t *coeff_ptr, intptr_t n_coeffs, const struct macroblock_plane *const mb_plane, tran_low_t *qcoeff_ptr, tran_low_t *dqcoeff_ptr, const int16_t *dequant_ptr, uint16_t *eob_ptr, const struct ScanOrder *const scan_order") + specialize(*qw('vpx_highbd_quantize_b neon sse2 avx2')) + +- add_proto(*qw('void vpx_highbd_quantize_b_32x32'), "const tran_low_t *coeff_ptr, intptr_t n_coeffs, const int16_t *zbin_ptr, const int16_t *round_ptr, const int16_t *quant_ptr, const int16_t *quant_shift_ptr, tran_low_t *qcoeff_ptr, tran_low_t *dqcoeff_ptr, const int16_t *dequant_ptr, uint16_t *eob_ptr, const int16_t *scan, const int16_t *iscan") ++ add_proto(*qw('void vpx_highbd_quantize_b_32x32'), "const tran_low_t *coeff_ptr, const struct macroblock_plane *const mb_plane, tran_low_t *qcoeff_ptr, tran_low_t *dqcoeff_ptr, const int16_t *dequant_ptr, uint16_t *eob_ptr, const struct ScanOrder *const scan_order") + specialize(*qw('vpx_highbd_quantize_b_32x32 neon sse2 avx2')) + # CONFIG_VP9_HIGHBITDEPTH + # CONFIG_VP9_ENCODER +@@ -738,32 +746,35 @@ if vpx_config("CONFIG_ENCODERS") == "yes": + add_proto(*qw('void vpx_subtract_block'), "int rows, int cols, int16_t *diff_ptr, ptrdiff_t diff_stride, const uint8_t *src_ptr, ptrdiff_t src_stride, const uint8_t *pred_ptr, ptrdiff_t pred_stride") + specialize(*qw('vpx_subtract_block neon msa mmi sse2 avx2 vsx lsx')) + ++ add_proto(*qw('int64_t'), "vpx_sse", "const uint8_t *src, int src_stride, const uint8_t *ref, int ref_stride, int width, int height") ++ specialize(*qw('vpx_sse sse4_1 avx2 neon neon_dotprod')) ++ + # + # Single block SAD + # + add_proto(*qw('unsigned int vpx_sad64x64'), "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride") +- specialize(*qw('vpx_sad64x64 neon avx2 msa sse2 vsx mmi lsx')) ++ specialize(*qw('vpx_sad64x64 neon neon_dotprod avx2 msa sse2 vsx mmi lsx')) + + add_proto(*qw('unsigned int vpx_sad64x32'), "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride") +- specialize(*qw('vpx_sad64x32 neon avx2 msa sse2 vsx mmi')) ++ specialize(*qw('vpx_sad64x32 neon neon_dotprod avx2 msa sse2 vsx mmi')) + + add_proto(*qw('unsigned int vpx_sad32x64'), "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride") +- specialize(*qw('vpx_sad32x64 neon avx2 msa sse2 vsx mmi')) ++ specialize(*qw('vpx_sad32x64 neon neon_dotprod avx2 msa sse2 vsx mmi')) + + add_proto(*qw('unsigned int vpx_sad32x32'), "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride") +- specialize(*qw('vpx_sad32x32 neon avx2 msa sse2 vsx mmi lsx')) ++ specialize(*qw('vpx_sad32x32 neon neon_dotprod avx2 msa sse2 vsx mmi lsx')) + + add_proto(*qw('unsigned int vpx_sad32x16'), "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride") +- specialize(*qw('vpx_sad32x16 neon avx2 msa sse2 vsx mmi')) ++ specialize(*qw('vpx_sad32x16 neon neon_dotprod avx2 msa sse2 vsx mmi')) + + add_proto(*qw('unsigned int vpx_sad16x32'), "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride") +- specialize(*qw('vpx_sad16x32 neon msa sse2 vsx mmi')) ++ specialize(*qw('vpx_sad16x32 neon neon_dotprod msa sse2 vsx mmi')) + + add_proto(*qw('unsigned int vpx_sad16x16'), "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride") +- specialize(*qw('vpx_sad16x16 neon msa sse2 vsx mmi lsx')) ++ specialize(*qw('vpx_sad16x16 neon neon_dotprod msa sse2 vsx mmi lsx')) + + add_proto(*qw('unsigned int vpx_sad16x8'), "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride") +- specialize(*qw('vpx_sad16x8 neon msa sse2 vsx mmi')) ++ specialize(*qw('vpx_sad16x8 neon neon_dotprod msa sse2 vsx mmi')) + + add_proto(*qw('unsigned int vpx_sad8x16'), "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride") + specialize(*qw('vpx_sad8x16 neon msa sse2 vsx mmi')) +@@ -780,6 +791,45 @@ if vpx_config("CONFIG_ENCODERS") == "yes": + add_proto(*qw('unsigned int vpx_sad4x4'), "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride") + specialize(*qw('vpx_sad4x4 neon msa sse2 mmi')) + ++ add_proto(*qw('unsigned int vpx_sad_skip_64x64'), "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride") ++ specialize(*qw('vpx_sad_skip_64x64 neon neon_dotprod avx2 sse2')) ++ ++ add_proto(*qw('unsigned int vpx_sad_skip_64x32'), "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride") ++ specialize(*qw('vpx_sad_skip_64x32 neon neon_dotprod avx2 sse2')) ++ ++ add_proto(*qw('unsigned int vpx_sad_skip_32x64'), "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride") ++ specialize(*qw('vpx_sad_skip_32x64 neon neon_dotprod avx2 sse2')) ++ ++ add_proto(*qw('unsigned int vpx_sad_skip_32x32'), "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride") ++ specialize(*qw('vpx_sad_skip_32x32 neon neon_dotprod avx2 sse2')) ++ ++ add_proto(*qw('unsigned int vpx_sad_skip_32x16'), "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride") ++ specialize(*qw('vpx_sad_skip_32x16 neon neon_dotprod avx2 sse2')) ++ ++ add_proto(*qw('unsigned int vpx_sad_skip_16x32'), "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride") ++ specialize(*qw('vpx_sad_skip_16x32 neon neon_dotprod sse2')) ++ ++ add_proto(*qw('unsigned int vpx_sad_skip_16x16'), "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride") ++ specialize(*qw('vpx_sad_skip_16x16 neon neon_dotprod sse2')) ++ ++ add_proto(*qw('unsigned int vpx_sad_skip_16x8'), "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride") ++ specialize(*qw('vpx_sad_skip_16x8 neon neon_dotprod sse2')) ++ ++ add_proto(*qw('unsigned int vpx_sad_skip_8x16'), "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride") ++ specialize(*qw('vpx_sad_skip_8x16 neon sse2')) ++ ++ add_proto(*qw('unsigned int vpx_sad_skip_8x8'), "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride") ++ specialize(*qw('vpx_sad_skip_8x8 neon sse2')) ++ ++ add_proto(*qw('unsigned int vpx_sad_skip_8x4'), "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride") ++ specialize(*qw('vpx_sad_skip_8x4 neon')) ++ ++ add_proto(*qw('unsigned int vpx_sad_skip_4x8'), "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride") ++ specialize(*qw('vpx_sad_skip_4x8 neon sse2')) ++ ++ add_proto(*qw('unsigned int vpx_sad_skip_4x4'), "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride") ++ specialize(*qw('vpx_sad_skip_4x4 neon')) ++ + # + # Avg + # +@@ -804,19 +854,19 @@ if vpx_config("CONFIG_ENCODERS") == "yes": + specialize(*qw('vpx_hadamard_32x32 sse2 avx2 neon')) + + add_proto(*qw('void vpx_highbd_hadamard_8x8'), "const int16_t *src_diff, ptrdiff_t src_stride, tran_low_t *coeff") +- specialize(*qw('vpx_highbd_hadamard_8x8 avx2')) ++ specialize(*qw('vpx_highbd_hadamard_8x8 avx2 neon')) + + add_proto(*qw('void vpx_highbd_hadamard_16x16'), "const int16_t *src_diff, ptrdiff_t src_stride, tran_low_t *coeff") +- specialize(*qw('vpx_highbd_hadamard_16x16 avx2')) ++ specialize(*qw('vpx_highbd_hadamard_16x16 avx2 neon')) + + add_proto(*qw('void vpx_highbd_hadamard_32x32'), "const int16_t *src_diff, ptrdiff_t src_stride, tran_low_t *coeff") +- specialize(*qw('vpx_highbd_hadamard_32x32 avx2')) ++ specialize(*qw('vpx_highbd_hadamard_32x32 avx2 neon')) + + add_proto(*qw('int vpx_satd'), "const tran_low_t *coeff, int length") + specialize(*qw('vpx_satd avx2 sse2 neon')) + + add_proto(*qw('int vpx_highbd_satd'), "const tran_low_t *coeff, int length") +- specialize(*qw('vpx_highbd_satd avx2')) ++ specialize(*qw('vpx_highbd_satd avx2 neon')) + else: + add_proto(*qw('void vpx_hadamard_8x8'), "const int16_t *src_diff, ptrdiff_t src_stride, int16_t *coeff") + specialize(*qw('vpx_hadamard_8x8 sse2 neon msa vsx lsx'), f"{ssse3_x86_64}") +@@ -832,38 +882,37 @@ if vpx_config("CONFIG_ENCODERS") == "yes": + + + add_proto(*qw('void vpx_int_pro_row'), "int16_t hbuf[16], const uint8_t *ref, const int ref_stride, const int height") +- specialize(*qw('vpx_int_pro_row sse2 neon msa')) +- ++ specialize(*qw('vpx_int_pro_row neon sse2 msa')) + add_proto(*qw('int16_t vpx_int_pro_col'), "const uint8_t *ref, const int width") +- specialize(*qw('vpx_int_pro_col sse2 neon msa')) ++ specialize(*qw('vpx_int_pro_col neon sse2 msa')) + + add_proto(*qw('int vpx_vector_var'), "const int16_t *ref, const int16_t *src, const int bwl") + specialize(*qw('vpx_vector_var neon sse2 msa')) + # CONFIG_VP9_ENCODER + + add_proto(*qw('unsigned int vpx_sad64x64_avg'), "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride, const uint8_t *second_pred") +- specialize(*qw('vpx_sad64x64_avg neon avx2 msa sse2 vsx mmi lsx')) ++ specialize(*qw('vpx_sad64x64_avg neon neon_dotprod avx2 msa sse2 vsx mmi lsx')) + + add_proto(*qw('unsigned int vpx_sad64x32_avg'), "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride, const uint8_t *second_pred") +- specialize(*qw('vpx_sad64x32_avg neon avx2 msa sse2 vsx mmi')) ++ specialize(*qw('vpx_sad64x32_avg neon neon_dotprod avx2 msa sse2 vsx mmi')) + + add_proto(*qw('unsigned int vpx_sad32x64_avg'), "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride, const uint8_t *second_pred") +- specialize(*qw('vpx_sad32x64_avg neon avx2 msa sse2 vsx mmi')) ++ specialize(*qw('vpx_sad32x64_avg neon neon_dotprod avx2 msa sse2 vsx mmi')) + + add_proto(*qw('unsigned int vpx_sad32x32_avg'), "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride, const uint8_t *second_pred") +- specialize(*qw('vpx_sad32x32_avg neon avx2 msa sse2 vsx mmi lsx')) ++ specialize(*qw('vpx_sad32x32_avg neon neon_dotprod avx2 msa sse2 vsx mmi lsx')) + + add_proto(*qw('unsigned int vpx_sad32x16_avg'), "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride, const uint8_t *second_pred") +- specialize(*qw('vpx_sad32x16_avg neon avx2 msa sse2 vsx mmi')) ++ specialize(*qw('vpx_sad32x16_avg neon neon_dotprod avx2 msa sse2 vsx mmi')) + + add_proto(*qw('unsigned int vpx_sad16x32_avg'), "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride, const uint8_t *second_pred") +- specialize(*qw('vpx_sad16x32_avg neon msa sse2 vsx mmi')) ++ specialize(*qw('vpx_sad16x32_avg neon neon_dotprod msa sse2 vsx mmi')) + + add_proto(*qw('unsigned int vpx_sad16x16_avg'), "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride, const uint8_t *second_pred") +- specialize(*qw('vpx_sad16x16_avg neon msa sse2 vsx mmi')) ++ specialize(*qw('vpx_sad16x16_avg neon neon_dotprod msa sse2 vsx mmi')) + + add_proto(*qw('unsigned int vpx_sad16x8_avg'), "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride, const uint8_t *second_pred") +- specialize(*qw('vpx_sad16x8_avg neon msa sse2 vsx mmi')) ++ specialize(*qw('vpx_sad16x8_avg neon neon_dotprod msa sse2 vsx mmi')) + + add_proto(*qw('unsigned int vpx_sad8x16_avg'), "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride, const uint8_t *second_pred") + specialize(*qw('vpx_sad8x16_avg neon msa sse2 mmi')) +@@ -883,45 +932,84 @@ if vpx_config("CONFIG_ENCODERS") == "yes": + # + # Multi-block SAD, comparing a reference to N independent blocks + # +- add_proto(*qw('void vpx_sad64x64x4d'), "const uint8_t *src_ptr, int src_stride, const uint8_t * const ref_array[4], int ref_stride, uint32_t sad_array[4]") +- specialize(*qw('vpx_sad64x64x4d avx512 avx2 neon msa sse2 vsx mmi lsx')) ++ add_proto(*qw('void vpx_sad64x64x4d'), "const uint8_t *src_ptr, int src_stride, const uint8_t *const ref_array[4], int ref_stride, uint32_t sad_array[4]") ++ specialize(*qw('vpx_sad64x64x4d avx512 avx2 neon neon_dotprod msa sse2 vsx mmi lsx')) + +- add_proto(*qw('void vpx_sad64x32x4d'), "const uint8_t *src_ptr, int src_stride, const uint8_t * const ref_array[4], int ref_stride, uint32_t sad_array[4]") +- specialize(*qw('vpx_sad64x32x4d neon msa sse2 vsx mmi lsx')) ++ add_proto(*qw('void vpx_sad64x32x4d'), "const uint8_t *src_ptr, int src_stride, const uint8_t *const ref_array[4], int ref_stride, uint32_t sad_array[4]") ++ specialize(*qw('vpx_sad64x32x4d neon neon_dotprod msa sse2 vsx mmi lsx')) + +- add_proto(*qw('void vpx_sad32x64x4d'), "const uint8_t *src_ptr, int src_stride, const uint8_t * const ref_array[4], int ref_stride, uint32_t sad_array[4]") +- specialize(*qw('vpx_sad32x64x4d neon msa sse2 vsx mmi lsx')) ++ add_proto(*qw('void vpx_sad32x64x4d'), "const uint8_t *src_ptr, int src_stride, const uint8_t *const ref_array[4], int ref_stride, uint32_t sad_array[4]") ++ specialize(*qw('vpx_sad32x64x4d neon neon_dotprod msa sse2 vsx mmi lsx')) + +- add_proto(*qw('void vpx_sad32x32x4d'), "const uint8_t *src_ptr, int src_stride, const uint8_t * const ref_array[4], int ref_stride, uint32_t sad_array[4]") +- specialize(*qw('vpx_sad32x32x4d avx2 neon msa sse2 vsx mmi lsx')) ++ add_proto(*qw('void vpx_sad32x32x4d'), "const uint8_t *src_ptr, int src_stride, const uint8_t *const ref_array[4], int ref_stride, uint32_t sad_array[4]") ++ specialize(*qw('vpx_sad32x32x4d avx2 neon neon_dotprod msa sse2 vsx mmi lsx')) + +- add_proto(*qw('void vpx_sad32x16x4d'), "const uint8_t *src_ptr, int src_stride, const uint8_t * const ref_array[4], int ref_stride, uint32_t sad_array[4]") +- specialize(*qw('vpx_sad32x16x4d neon msa sse2 vsx mmi')) ++ add_proto(*qw('void vpx_sad32x16x4d'), "const uint8_t *src_ptr, int src_stride, const uint8_t *const ref_array[4], int ref_stride, uint32_t sad_array[4]") ++ specialize(*qw('vpx_sad32x16x4d neon neon_dotprod msa sse2 vsx mmi')) + +- add_proto(*qw('void vpx_sad16x32x4d'), "const uint8_t *src_ptr, int src_stride, const uint8_t * const ref_array[4], int ref_stride, uint32_t sad_array[4]") +- specialize(*qw('vpx_sad16x32x4d neon msa sse2 vsx mmi')) ++ add_proto(*qw('void vpx_sad16x32x4d'), "const uint8_t *src_ptr, int src_stride, const uint8_t *const ref_array[4], int ref_stride, uint32_t sad_array[4]") ++ specialize(*qw('vpx_sad16x32x4d neon neon_dotprod msa sse2 vsx mmi')) + +- add_proto(*qw('void vpx_sad16x16x4d'), "const uint8_t *src_ptr, int src_stride, const uint8_t * const ref_array[4], int ref_stride, uint32_t sad_array[4]") +- specialize(*qw('vpx_sad16x16x4d neon msa sse2 vsx mmi lsx')) ++ add_proto(*qw('void vpx_sad16x16x4d'), "const uint8_t *src_ptr, int src_stride, const uint8_t *const ref_array[4], int ref_stride, uint32_t sad_array[4]") ++ specialize(*qw('vpx_sad16x16x4d neon neon_dotprod msa sse2 vsx mmi lsx')) + +- add_proto(*qw('void vpx_sad16x8x4d'), "const uint8_t *src_ptr, int src_stride, const uint8_t * const ref_array[4], int ref_stride, uint32_t sad_array[4]") +- specialize(*qw('vpx_sad16x8x4d neon msa sse2 vsx mmi')) ++ add_proto(*qw('void vpx_sad16x8x4d'), "const uint8_t *src_ptr, int src_stride, const uint8_t *const ref_array[4], int ref_stride, uint32_t sad_array[4]") ++ specialize(*qw('vpx_sad16x8x4d neon neon_dotprod msa sse2 vsx mmi')) + +- add_proto(*qw('void vpx_sad8x16x4d'), "const uint8_t *src_ptr, int src_stride, const uint8_t * const ref_array[4], int ref_stride, uint32_t sad_array[4]") ++ add_proto(*qw('void vpx_sad8x16x4d'), "const uint8_t *src_ptr, int src_stride, const uint8_t *const ref_array[4], int ref_stride, uint32_t sad_array[4]") + specialize(*qw('vpx_sad8x16x4d neon msa sse2 mmi')) + +- add_proto(*qw('void vpx_sad8x8x4d'), "const uint8_t *src_ptr, int src_stride, const uint8_t * const ref_array[4], int ref_stride, uint32_t sad_array[4]") ++ add_proto(*qw('void vpx_sad8x8x4d'), "const uint8_t *src_ptr, int src_stride, const uint8_t *const ref_array[4], int ref_stride, uint32_t sad_array[4]") + specialize(*qw('vpx_sad8x8x4d neon msa sse2 mmi lsx')) + +- add_proto(*qw('void vpx_sad8x4x4d'), "const uint8_t *src_ptr, int src_stride, const uint8_t * const ref_array[4], int ref_stride, uint32_t sad_array[4]") ++ add_proto(*qw('void vpx_sad8x4x4d'), "const uint8_t *src_ptr, int src_stride, const uint8_t *const ref_array[4], int ref_stride, uint32_t sad_array[4]") + specialize(*qw('vpx_sad8x4x4d neon msa sse2 mmi')) + +- add_proto(*qw('void vpx_sad4x8x4d'), "const uint8_t *src_ptr, int src_stride, const uint8_t * const ref_array[4], int ref_stride, uint32_t sad_array[4]") ++ add_proto(*qw('void vpx_sad4x8x4d'), "const uint8_t *src_ptr, int src_stride, const uint8_t *const ref_array[4], int ref_stride, uint32_t sad_array[4]") + specialize(*qw('vpx_sad4x8x4d neon msa sse2 mmi')) + +- add_proto(*qw('void vpx_sad4x4x4d'), "const uint8_t *src_ptr, int src_stride, const uint8_t * const ref_array[4], int ref_stride, uint32_t sad_array[4]") ++ add_proto(*qw('void vpx_sad4x4x4d'), "const uint8_t *src_ptr, int src_stride, const uint8_t *const ref_array[4], int ref_stride, uint32_t sad_array[4]") + specialize(*qw('vpx_sad4x4x4d neon msa sse2 mmi')) + ++ add_proto(*qw('void vpx_sad_skip_64x64x4d'), "const uint8_t *src_ptr, int src_stride, const uint8_t *const ref_array[4], int ref_stride, uint32_t sad_array[4]") ++ specialize(*qw('vpx_sad_skip_64x64x4d neon neon_dotprod avx2 sse2')) ++ ++ add_proto(*qw('void vpx_sad_skip_64x32x4d'), "const uint8_t *src_ptr, int src_stride, const uint8_t *const ref_array[4], int ref_stride, uint32_t sad_array[4]") ++ specialize(*qw('vpx_sad_skip_64x32x4d neon neon_dotprod avx2 sse2')) ++ ++ add_proto(*qw('void vpx_sad_skip_32x64x4d'), "const uint8_t *src_ptr, int src_stride, const uint8_t *const ref_array[4], int ref_stride, uint32_t sad_array[4]") ++ specialize(*qw('vpx_sad_skip_32x64x4d neon neon_dotprod avx2 sse2')) ++ ++ add_proto(*qw('void vpx_sad_skip_32x32x4d'), "const uint8_t *src_ptr, int src_stride, const uint8_t *const ref_array[4], int ref_stride, uint32_t sad_array[4]") ++ specialize(*qw('vpx_sad_skip_32x32x4d neon neon_dotprod avx2 sse2')) ++ ++ add_proto(*qw('void vpx_sad_skip_32x16x4d'), "const uint8_t *src_ptr, int src_stride, const uint8_t *const ref_array[4], int ref_stride, uint32_t sad_array[4]") ++ specialize(*qw('vpx_sad_skip_32x16x4d neon neon_dotprod avx2 sse2')) ++ ++ add_proto(*qw('void vpx_sad_skip_16x32x4d'), "const uint8_t *src_ptr, int src_stride, const uint8_t *const ref_array[4], int ref_stride, uint32_t sad_array[4]") ++ specialize(*qw('vpx_sad_skip_16x32x4d neon neon_dotprod sse2')) ++ ++ add_proto(*qw('void vpx_sad_skip_16x16x4d'), "const uint8_t *src_ptr, int src_stride, const uint8_t *const ref_array[4], int ref_stride, uint32_t sad_array[4]") ++ specialize(*qw('vpx_sad_skip_16x16x4d neon neon_dotprod sse2')) ++ ++ add_proto(*qw('void vpx_sad_skip_16x8x4d'), "const uint8_t *src_ptr, int src_stride, const uint8_t *const ref_array[4], int ref_stride, uint32_t sad_array[4]") ++ specialize(*qw('vpx_sad_skip_16x8x4d neon neon_dotprod sse2')) ++ ++ add_proto(*qw('void vpx_sad_skip_8x16x4d'), "const uint8_t *src_ptr, int src_stride, const uint8_t *const ref_array[4], int ref_stride, uint32_t sad_array[4]") ++ specialize(*qw('vpx_sad_skip_8x16x4d neon sse2')) ++ ++ add_proto(*qw('void vpx_sad_skip_8x8x4d'), "const uint8_t *src_ptr, int src_stride, const uint8_t *const ref_array[4], int ref_stride, uint32_t sad_array[4]") ++ specialize(*qw('vpx_sad_skip_8x8x4d neon sse2')) ++ ++ add_proto(*qw('void vpx_sad_skip_8x4x4d'), "const uint8_t *src_ptr, int src_stride, const uint8_t *const ref_array[4], int ref_stride, uint32_t sad_array[4]") ++ specialize(*qw('vpx_sad_skip_8x4x4d neon')) ++ ++ add_proto(*qw('void vpx_sad_skip_4x8x4d'), "const uint8_t *src_ptr, int src_stride, const uint8_t *const ref_array[4], int ref_stride, uint32_t sad_array[4]") ++ specialize(*qw('vpx_sad_skip_4x8x4d neon sse2')) ++ ++ add_proto(*qw('void vpx_sad_skip_4x4x4d'), "const uint8_t *src_ptr, int src_stride, const uint8_t *const ref_array[4], int ref_stride, uint32_t sad_array[4]") ++ specialize(*qw('vpx_sad_skip_4x4x4d neon')) ++ + add_proto(*qw('uint64_t vpx_sum_squares_2d_i16'), "const int16_t *src, int stride, int size") + specialize(*qw('vpx_sum_squares_2d_i16 neon sse2 msa')) + +@@ -942,6 +1030,9 @@ if vpx_config("CONFIG_ENCODERS") == "yes": + add_proto(*qw('void vpx_highbd_subtract_block'), "int rows, int cols, int16_t *diff_ptr, ptrdiff_t diff_stride, const uint8_t *src8_ptr, ptrdiff_t src_stride, const uint8_t *pred8_ptr, ptrdiff_t pred_stride, int bd") + specialize(*qw('vpx_highbd_subtract_block neon avx2')) + ++ add_proto(*qw('int64_t'), "vpx_highbd_sse", "const uint8_t *a8, int a_stride, const uint8_t *b8,int b_stride, int width, int height") ++ specialize(*qw('vpx_highbd_sse sse4_1 avx2 neon')) ++ + # + # Single block SAD + # +@@ -984,16 +1075,56 @@ if vpx_config("CONFIG_ENCODERS") == "yes": + add_proto(*qw('unsigned int vpx_highbd_sad4x4'), "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride") + specialize(*qw('vpx_highbd_sad4x4 neon')) + ++ add_proto(*qw('unsigned int vpx_highbd_sad_skip_64x64'), "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride") ++ specialize(*qw('vpx_highbd_sad_skip_64x64 neon sse2 avx2')) ++ ++ add_proto(*qw('unsigned int vpx_highbd_sad_skip_64x32'), "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride") ++ specialize(*qw('vpx_highbd_sad_skip_64x32 neon sse2 avx2')) ++ ++ add_proto(*qw('unsigned int vpx_highbd_sad_skip_32x64'), "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride") ++ specialize(*qw('vpx_highbd_sad_skip_32x64 neon sse2 avx2')) ++ ++ add_proto(*qw('unsigned int vpx_highbd_sad_skip_32x32'), "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride") ++ specialize(*qw('vpx_highbd_sad_skip_32x32 neon sse2 avx2')) ++ ++ add_proto(*qw('unsigned int vpx_highbd_sad_skip_32x16'), "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride") ++ specialize(*qw('vpx_highbd_sad_skip_32x16 neon sse2 avx2')) ++ ++ add_proto(*qw('unsigned int vpx_highbd_sad_skip_16x32'), "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride") ++ specialize(*qw('vpx_highbd_sad_skip_16x32 neon sse2 avx2')) ++ ++ add_proto(*qw('unsigned int vpx_highbd_sad_skip_16x16'), "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride") ++ specialize(*qw('vpx_highbd_sad_skip_16x16 neon sse2 avx2')) ++ ++ add_proto(*qw('unsigned int vpx_highbd_sad_skip_16x8'), "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride") ++ specialize(*qw('vpx_highbd_sad_skip_16x8 neon sse2 avx2')) ++ ++ add_proto(*qw('unsigned int vpx_highbd_sad_skip_8x16'), "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride") ++ specialize(*qw('vpx_highbd_sad_skip_8x16 neon sse2')) ++ ++ add_proto(*qw('unsigned int vpx_highbd_sad_skip_8x8'), "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride") ++ specialize(*qw('vpx_highbd_sad_skip_8x8 neon sse2')) ++ ++ add_proto(*qw('unsigned int vpx_highbd_sad_skip_8x4'), "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride") ++ specialize(*qw('vpx_highbd_sad_skip_8x4 neon')) ++ ++ add_proto(*qw('unsigned int vpx_highbd_sad_skip_4x8'), "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride") ++ specialize(*qw('vpx_highbd_sad_skip_4x8 neon')) ++ ++ add_proto(*qw('unsigned int vpx_highbd_sad_skip_4x4'), "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride") ++ specialize(*qw('vpx_highbd_sad_skip_4x4 neon')) ++ + # + # Avg + # + add_proto(*qw('unsigned int vpx_highbd_avg_8x8'), "const uint8_t *s8, int p") +- specialize(*qw('vpx_highbd_avg_8x8 sse2')) ++ specialize(*qw('vpx_highbd_avg_8x8 sse2 neon')) + + add_proto(*qw('unsigned int vpx_highbd_avg_4x4'), "const uint8_t *s8, int p") +- specialize(*qw('vpx_highbd_avg_4x4 sse2')) ++ specialize(*qw('vpx_highbd_avg_4x4 sse2 neon')) + + add_proto(*qw('void vpx_highbd_minmax_8x8'), "const uint8_t *s8, int p, const uint8_t *d8, int dp, int *min, int *max") ++ specialize(*qw('vpx_highbd_minmax_8x8 neon')) + + add_proto(*qw('unsigned int vpx_highbd_sad64x64_avg'), "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride, const uint8_t *second_pred") + specialize(*qw('vpx_highbd_sad64x64_avg sse2 neon avx2')) +@@ -1037,45 +1168,84 @@ if vpx_config("CONFIG_ENCODERS") == "yes": + # + # Multi-block SAD, comparing a reference to N independent blocks + # +- add_proto(*qw('void vpx_highbd_sad64x64x4d'), "const uint8_t *src_ptr, int src_stride, const uint8_t* const ref_array[4], int ref_stride, uint32_t sad_array[4]") ++ add_proto(*qw('void vpx_highbd_sad64x64x4d'), "const uint8_t *src_ptr, int src_stride, const uint8_t *const ref_array[4], int ref_stride, uint32_t sad_array[4]") + specialize(*qw('vpx_highbd_sad64x64x4d sse2 neon avx2')) + +- add_proto(*qw('void vpx_highbd_sad64x32x4d'), "const uint8_t *src_ptr, int src_stride, const uint8_t* const ref_array[4], int ref_stride, uint32_t sad_array[4]") ++ add_proto(*qw('void vpx_highbd_sad64x32x4d'), "const uint8_t *src_ptr, int src_stride, const uint8_t *const ref_array[4], int ref_stride, uint32_t sad_array[4]") + specialize(*qw('vpx_highbd_sad64x32x4d sse2 neon avx2')) + +- add_proto(*qw('void vpx_highbd_sad32x64x4d'), "const uint8_t *src_ptr, int src_stride, const uint8_t* const ref_array[4], int ref_stride, uint32_t sad_array[4]") ++ add_proto(*qw('void vpx_highbd_sad32x64x4d'), "const uint8_t *src_ptr, int src_stride, const uint8_t *const ref_array[4], int ref_stride, uint32_t sad_array[4]") + specialize(*qw('vpx_highbd_sad32x64x4d sse2 neon avx2')) + +- add_proto(*qw('void vpx_highbd_sad32x32x4d'), "const uint8_t *src_ptr, int src_stride, const uint8_t* const ref_array[4], int ref_stride, uint32_t sad_array[4]") ++ add_proto(*qw('void vpx_highbd_sad32x32x4d'), "const uint8_t *src_ptr, int src_stride, const uint8_t *const ref_array[4], int ref_stride, uint32_t sad_array[4]") + specialize(*qw('vpx_highbd_sad32x32x4d sse2 neon avx2')) + +- add_proto(*qw('void vpx_highbd_sad32x16x4d'), "const uint8_t *src_ptr, int src_stride, const uint8_t* const ref_array[4], int ref_stride, uint32_t sad_array[4]") ++ add_proto(*qw('void vpx_highbd_sad32x16x4d'), "const uint8_t *src_ptr, int src_stride, const uint8_t *const ref_array[4], int ref_stride, uint32_t sad_array[4]") + specialize(*qw('vpx_highbd_sad32x16x4d sse2 neon avx2')) + +- add_proto(*qw('void vpx_highbd_sad16x32x4d'), "const uint8_t *src_ptr, int src_stride, const uint8_t* const ref_array[4], int ref_stride, uint32_t sad_array[4]") ++ add_proto(*qw('void vpx_highbd_sad16x32x4d'), "const uint8_t *src_ptr, int src_stride, const uint8_t *const ref_array[4], int ref_stride, uint32_t sad_array[4]") + specialize(*qw('vpx_highbd_sad16x32x4d sse2 neon avx2')) + +- add_proto(*qw('void vpx_highbd_sad16x16x4d'), "const uint8_t *src_ptr, int src_stride, const uint8_t* const ref_array[4], int ref_stride, uint32_t sad_array[4]") ++ add_proto(*qw('void vpx_highbd_sad16x16x4d'), "const uint8_t *src_ptr, int src_stride, const uint8_t *const ref_array[4], int ref_stride, uint32_t sad_array[4]") + specialize(*qw('vpx_highbd_sad16x16x4d sse2 neon avx2')) + +- add_proto(*qw('void vpx_highbd_sad16x8x4d'), "const uint8_t *src_ptr, int src_stride, const uint8_t* const ref_array[4], int ref_stride, uint32_t sad_array[4]") ++ add_proto(*qw('void vpx_highbd_sad16x8x4d'), "const uint8_t *src_ptr, int src_stride, const uint8_t *const ref_array[4], int ref_stride, uint32_t sad_array[4]") + specialize(*qw('vpx_highbd_sad16x8x4d sse2 neon avx2')) + +- add_proto(*qw('void vpx_highbd_sad8x16x4d'), "const uint8_t *src_ptr, int src_stride, const uint8_t* const ref_array[4], int ref_stride, uint32_t sad_array[4]") ++ add_proto(*qw('void vpx_highbd_sad8x16x4d'), "const uint8_t *src_ptr, int src_stride, const uint8_t *const ref_array[4], int ref_stride, uint32_t sad_array[4]") + specialize(*qw('vpx_highbd_sad8x16x4d sse2 neon')) + +- add_proto(*qw('void vpx_highbd_sad8x8x4d'), "const uint8_t *src_ptr, int src_stride, const uint8_t* const ref_array[4], int ref_stride, uint32_t sad_array[4]") ++ add_proto(*qw('void vpx_highbd_sad8x8x4d'), "const uint8_t *src_ptr, int src_stride, const uint8_t *const ref_array[4], int ref_stride, uint32_t sad_array[4]") + specialize(*qw('vpx_highbd_sad8x8x4d sse2 neon')) + +- add_proto(*qw('void vpx_highbd_sad8x4x4d'), "const uint8_t *src_ptr, int src_stride, const uint8_t* const ref_array[4], int ref_stride, uint32_t sad_array[4]") ++ add_proto(*qw('void vpx_highbd_sad8x4x4d'), "const uint8_t *src_ptr, int src_stride, const uint8_t *const ref_array[4], int ref_stride, uint32_t sad_array[4]") + specialize(*qw('vpx_highbd_sad8x4x4d sse2 neon')) + +- add_proto(*qw('void vpx_highbd_sad4x8x4d'), "const uint8_t *src_ptr, int src_stride, const uint8_t* const ref_array[4], int ref_stride, uint32_t sad_array[4]") ++ add_proto(*qw('void vpx_highbd_sad4x8x4d'), "const uint8_t *src_ptr, int src_stride, const uint8_t *const ref_array[4], int ref_stride, uint32_t sad_array[4]") + specialize(*qw('vpx_highbd_sad4x8x4d sse2 neon')) + +- add_proto(*qw('void vpx_highbd_sad4x4x4d'), "const uint8_t *src_ptr, int src_stride, const uint8_t* const ref_array[4], int ref_stride, uint32_t sad_array[4]") ++ add_proto(*qw('void vpx_highbd_sad4x4x4d'), "const uint8_t *src_ptr, int src_stride, const uint8_t *const ref_array[4], int ref_stride, uint32_t sad_array[4]") + specialize(*qw('vpx_highbd_sad4x4x4d sse2 neon')) + ++ add_proto(*qw('void vpx_highbd_sad_skip_64x64x4d'), "const uint8_t *src_ptr, int src_stride, const uint8_t *const ref_array[4], int ref_stride, uint32_t sad_array[4]") ++ specialize(*qw('vpx_highbd_sad_skip_64x64x4d neon sse2 avx2')) ++ ++ add_proto(*qw('void vpx_highbd_sad_skip_64x32x4d'), "const uint8_t *src_ptr, int src_stride, const uint8_t *const ref_array[4], int ref_stride, uint32_t sad_array[4]") ++ specialize(*qw('vpx_highbd_sad_skip_64x32x4d neon sse2 avx2')) ++ ++ add_proto(*qw('void vpx_highbd_sad_skip_32x64x4d'), "const uint8_t *src_ptr, int src_stride, const uint8_t *const ref_array[4], int ref_stride, uint32_t sad_array[4]") ++ specialize(*qw('vpx_highbd_sad_skip_32x64x4d neon sse2 avx2')) ++ ++ add_proto(*qw('void vpx_highbd_sad_skip_32x32x4d'), "const uint8_t *src_ptr, int src_stride, const uint8_t *const ref_array[4], int ref_stride, uint32_t sad_array[4]") ++ specialize(*qw('vpx_highbd_sad_skip_32x32x4d neon sse2 avx2')) ++ ++ add_proto(*qw('void vpx_highbd_sad_skip_32x16x4d'), "const uint8_t *src_ptr, int src_stride, const uint8_t *const ref_array[4], int ref_stride, uint32_t sad_array[4]") ++ specialize(*qw('vpx_highbd_sad_skip_32x16x4d neon sse2 avx2')) ++ ++ add_proto(*qw('void vpx_highbd_sad_skip_16x32x4d'), "const uint8_t *src_ptr, int src_stride, const uint8_t *const ref_array[4], int ref_stride, uint32_t sad_array[4]") ++ specialize(*qw('vpx_highbd_sad_skip_16x32x4d neon sse2 avx2')) ++ ++ add_proto(*qw('void vpx_highbd_sad_skip_16x16x4d'), "const uint8_t *src_ptr, int src_stride, const uint8_t *const ref_array[4], int ref_stride, uint32_t sad_array[4]") ++ specialize(*qw('vpx_highbd_sad_skip_16x16x4d neon sse2 avx2')) ++ ++ add_proto(*qw('void vpx_highbd_sad_skip_16x8x4d'), "const uint8_t *src_ptr, int src_stride, const uint8_t *const ref_array[4], int ref_stride, uint32_t sad_array[4]") ++ specialize(*qw('vpx_highbd_sad_skip_16x8x4d neon sse2 avx2')) ++ ++ add_proto(*qw('void vpx_highbd_sad_skip_8x16x4d'), "const uint8_t *src_ptr, int src_stride, const uint8_t *const ref_array[4], int ref_stride, uint32_t sad_array[4]") ++ specialize(*qw('vpx_highbd_sad_skip_8x16x4d neon sse2')) ++ ++ add_proto(*qw('void vpx_highbd_sad_skip_8x8x4d'), "const uint8_t *src_ptr, int src_stride, const uint8_t *const ref_array[4], int ref_stride, uint32_t sad_array[4]") ++ specialize(*qw('vpx_highbd_sad_skip_8x8x4d neon sse2')) ++ ++ add_proto(*qw('void vpx_highbd_sad_skip_8x4x4d'), "const uint8_t *src_ptr, int src_stride, const uint8_t *const ref_array[4], int ref_stride, uint32_t sad_array[4]") ++ specialize(*qw('vpx_highbd_sad_skip_8x4x4d neon')) ++ ++ add_proto(*qw('void vpx_highbd_sad_skip_4x8x4d'), "const uint8_t *src_ptr, int src_stride, const uint8_t *const ref_array[4], int ref_stride, uint32_t sad_array[4]") ++ specialize(*qw('vpx_highbd_sad_skip_4x8x4d neon sse2')) ++ ++ add_proto(*qw('void vpx_highbd_sad_skip_4x4x4d'), "const uint8_t *src_ptr, int src_stride, const uint8_t *const ref_array[4], int ref_stride, uint32_t sad_array[4]") ++ specialize(*qw('vpx_highbd_sad_skip_4x4x4d neon')) ++ + # + # Structured Similarity (SSIM) + # +@@ -1091,73 +1261,73 @@ if vpx_config("CONFIG_ENCODERS") == "yes" or vpx_config("CONFIG_POSTPROC") == "y + # Variance + # + add_proto(*qw('unsigned int vpx_variance64x64'), "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse") +- specialize(*qw('vpx_variance64x64 sse2 avx2 neon msa mmi vsx lsx')) ++ specialize(*qw('vpx_variance64x64 sse2 avx2 neon neon_dotprod msa mmi vsx lsx')) + + add_proto(*qw('unsigned int vpx_variance64x32'), "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse") +- specialize(*qw('vpx_variance64x32 sse2 avx2 neon msa mmi vsx')) ++ specialize(*qw('vpx_variance64x32 sse2 avx2 neon neon_dotprod msa mmi vsx')) + + add_proto(*qw('unsigned int vpx_variance32x64'), "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse") +- specialize(*qw('vpx_variance32x64 sse2 avx2 neon msa mmi vsx')) ++ specialize(*qw('vpx_variance32x64 sse2 avx2 neon neon_dotprod msa mmi vsx')) + + add_proto(*qw('unsigned int vpx_variance32x32'), "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse") +- specialize(*qw('vpx_variance32x32 sse2 avx2 neon msa mmi vsx lsx')) ++ specialize(*qw('vpx_variance32x32 sse2 avx2 neon neon_dotprod msa mmi vsx lsx')) + + add_proto(*qw('unsigned int vpx_variance32x16'), "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse") +- specialize(*qw('vpx_variance32x16 sse2 avx2 neon msa mmi vsx')) ++ specialize(*qw('vpx_variance32x16 sse2 avx2 neon neon_dotprod msa mmi vsx')) + + add_proto(*qw('unsigned int vpx_variance16x32'), "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse") +- specialize(*qw('vpx_variance16x32 sse2 avx2 neon msa mmi vsx')) ++ specialize(*qw('vpx_variance16x32 sse2 avx2 neon neon_dotprod msa mmi vsx')) + + add_proto(*qw('unsigned int vpx_variance16x16'), "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse") +- specialize(*qw('vpx_variance16x16 sse2 avx2 neon msa mmi vsx lsx')) ++ specialize(*qw('vpx_variance16x16 sse2 avx2 neon neon_dotprod msa mmi vsx lsx')) + + add_proto(*qw('unsigned int vpx_variance16x8'), "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse") +- specialize(*qw('vpx_variance16x8 sse2 avx2 neon msa mmi vsx')) ++ specialize(*qw('vpx_variance16x8 sse2 avx2 neon neon_dotprod msa mmi vsx')) + + add_proto(*qw('unsigned int vpx_variance8x16'), "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse") +- specialize(*qw('vpx_variance8x16 sse2 neon msa mmi vsx')) ++ specialize(*qw('vpx_variance8x16 sse2 avx2 neon neon_dotprod msa mmi vsx')) + + add_proto(*qw('unsigned int vpx_variance8x8'), "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse") +- specialize(*qw('vpx_variance8x8 sse2 neon msa mmi vsx lsx')) ++ specialize(*qw('vpx_variance8x8 sse2 avx2 neon neon_dotprod msa mmi vsx lsx')) + + add_proto(*qw('unsigned int vpx_variance8x4'), "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse") +- specialize(*qw('vpx_variance8x4 sse2 neon msa mmi vsx')) ++ specialize(*qw('vpx_variance8x4 sse2 avx2 neon neon_dotprod msa mmi vsx')) + + add_proto(*qw('unsigned int vpx_variance4x8'), "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse") +- specialize(*qw('vpx_variance4x8 sse2 neon msa mmi vsx')) ++ specialize(*qw('vpx_variance4x8 sse2 neon neon_dotprod msa mmi vsx')) + + add_proto(*qw('unsigned int vpx_variance4x4'), "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse") +- specialize(*qw('vpx_variance4x4 sse2 neon msa mmi vsx')) ++ specialize(*qw('vpx_variance4x4 sse2 neon neon_dotprod msa mmi vsx')) + + # + # Specialty Variance + # + add_proto(*qw('void vpx_get16x16var'), "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse, int *sum") +- specialize(*qw('vpx_get16x16var sse2 avx2 neon msa vsx lsx')) ++ specialize(*qw('vpx_get16x16var sse2 avx2 neon neon_dotprod msa vsx lsx')) + + add_proto(*qw('void vpx_get8x8var'), "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse, int *sum") +- specialize(*qw('vpx_get8x8var sse2 neon msa vsx')) ++ specialize(*qw('vpx_get8x8var sse2 neon neon_dotprod msa vsx')) + + add_proto(*qw('unsigned int vpx_mse16x16'), "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse") +- specialize(*qw('vpx_mse16x16 sse2 avx2 neon msa mmi vsx lsx')) ++ specialize(*qw('vpx_mse16x16 sse2 avx2 neon neon_dotprod msa mmi vsx lsx')) + + add_proto(*qw('unsigned int vpx_mse16x8'), "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse") +- specialize(*qw('vpx_mse16x8 sse2 avx2 msa mmi vsx')) ++ specialize(*qw('vpx_mse16x8 sse2 avx2 neon neon_dotprod msa mmi vsx')) + + add_proto(*qw('unsigned int vpx_mse8x16'), "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse") +- specialize(*qw('vpx_mse8x16 sse2 msa mmi vsx')) ++ specialize(*qw('vpx_mse8x16 sse2 neon neon_dotprod msa mmi vsx')) + + add_proto(*qw('unsigned int vpx_mse8x8'), "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse") +- specialize(*qw('vpx_mse8x8 sse2 msa mmi vsx')) ++ specialize(*qw('vpx_mse8x8 sse2 neon neon_dotprod msa mmi vsx')) + + add_proto(*qw('unsigned int vpx_get_mb_ss'), "const int16_t *") + specialize(*qw('vpx_get_mb_ss sse2 msa vsx')) + + add_proto(*qw('unsigned int vpx_get4x4sse_cs'), "const unsigned char *src_ptr, int src_stride, const unsigned char *ref_ptr, int ref_stride") +- specialize(*qw('vpx_get4x4sse_cs neon msa vsx')) ++ specialize(*qw('vpx_get4x4sse_cs neon neon_dotprod msa vsx')) + + add_proto(*qw('void vpx_comp_avg_pred'), "uint8_t *comp_pred, const uint8_t *pred, int width, int height, const uint8_t *ref, int ref_stride") +- specialize(*qw('vpx_comp_avg_pred neon sse2 vsx lsx')) ++ specialize(*qw('vpx_comp_avg_pred neon sse2 avx2 vsx lsx')) + + # + # Subpixel Variance +diff --git a/vpx_ports/meson.build b/vpx_ports/meson.build +index 961ed54bd..c8079af98 100644 +--- a/vpx_ports/meson.build ++++ b/vpx_ports/meson.build +@@ -12,9 +12,6 @@ ports_optional_sources = { + 'mmx' : files( + # 'emms_mmx.c' # MANUAL + ), +- 'arm' : files( +- 'arm_cpudetect.c', +- ), + 'ppc' : files( + 'ppc_cpudetect.c', + ), +@@ -24,6 +21,7 @@ ports_optional_sources = { + 'loongarch' : files( + 'loongarch_cpudetect.c', + ), ++ 'arm': files(), # MANUAL + } + + ports_headers_sources = files( +@@ -41,6 +39,7 @@ ports_headers_sources = files( + ports_headers_optional_sources = { + 'arm' : files( + 'arm.h', ++ 'arm_cpudetect.h' + ), + 'ppc' : files( + 'ppc.h', +@@ -66,6 +65,8 @@ ports_asm_optional_sources = { + ), + } + ++#### --- END GENERATED --- #### ++ + ports_headers_sources += files( # MANUAL + 'emmintrin_compat.h', + 'mem_ops.h', +@@ -73,6 +74,18 @@ ports_headers_sources += files( # MANUAL + 'vpx_once.h', + ) + ++if features.get('aarch64', false) # MANUAL ++ ports_sources += files( ++ 'aarch64_cpudetect.c' ++ ) ++else ++ ports_optional_sources += { ++ 'arm' : ports_optional_sources['arm'] + files( ++ 'aarch32_cpudetect.c' ++ ), ++ } ++endif ++ + if features.get('x86', false) # MANUAL + ports_optional_sources += { + 'mmx' : ports_optional_sources['mmx'] + files( +@@ -115,8 +128,6 @@ ports_headers_optional_sources += { # MANUAL + ) + } + +-#### --- END GENERATED --- #### +- + codec_srcs += ports_sources + optional_sources = get_variable('ports_optional_sources', {}) + foreach comp_name, comp_sources : optional_sources +diff --git a/vpx_util/meson.build b/vpx_util/meson.build +index e39384421..fa6d42e5e 100644 +--- a/vpx_util/meson.build ++++ b/vpx_util/meson.build +@@ -30,6 +30,9 @@ util_headers_optional_sources = { + ), + } + ++#### --- END GENERATED --- #### ++ ++ + if features.get('bitstream_debug', false) or features.get('mismatch_debug', false) + util_sources += files( + 'vpx_debug_util.c' +@@ -40,8 +43,6 @@ if features.get('bitstream_debug', false) or features.get('mismatch_debug', fals + ) + endif + +-#### --- END GENERATED --- #### +- + optional_sources = get_variable('util_optional_sources', {}) + foreach comp_name, comp_sources : optional_sources + if features.get(comp_name, false) +-- +2.44.0.windows.1 + + +From e9e378d725ae4d0fadcc5e8ac72105b3e1edbd5c Mon Sep 17 00:00:00 2001 +From: "L. E. Segovia" <amy@centricular.com> +Date: Wed, 10 Jul 2024 19:50:24 -0300 +Subject: [PATCH] meson: Properly ban shared libraries with the MSVC ABI + +I hadn't noticed that the codeblock already existed, it was just in a +different place. + +Part-of: <https://gitlab.freedesktop.org/gstreamer/meson-ports/libvpx/-/merge_requests/21> + +diff --git a/meson.build b/meson.build +index 198165e61..fee067173 100644 +--- a/meson.build ++++ b/meson.build +@@ -1542,12 +1542,14 @@ endif + + ### process_detect ### + +-# Shared builds are supported everywhere thanks to +-# the Meson generation of the module definition files. +-# if features.get('shared', false) and not ['linux', 'solaris', 'darwin', 'iphonesimulator'].contains(tgt_os) and not features.get('gnu', false) +-# # Again, no OS/2 because Meson itself doens't support it +-# error('Shared library build is only supported on ELF and Darwin for now') +-# endif ++# MSVC requires declspec on data symbols when importing from a shared library. ++if features.get('shared', false) and not ['linux', 'solaris', 'darwin', 'ios', 'iphonesimulator'].contains(tgt_os) ++ if features.get('gcc', false) # Looks like a typo on upstream ++ warning('Shared library build is only supported on ELF; assuming this is OK') ++ else ++ error('Shared library build is only supported on ELF and Darwin for now') ++ endif ++endif + + features.set('unistd_h', c.has_header('unistd.h')) + if c.has_header('vpx/vpx_integer.h', include_directories: include_directories('.')) +@@ -1670,14 +1672,6 @@ vpx_config_c = configure_file( + + # libs.mk + +-if features.get('shared', false) and not ['linux', 'darwin', 'ios', 'iphonesimulator'].contains(tgt_os) +- if features.get('gcc', false) # Looks like a typo on upstream +- warn('Shared libraries are only supported on ELF; assuming this is OK') +- else +- error('Shared libraries are only supported on ELF, OS/2, and Darwin for now') +- endif +-endif +- + rtcd_exe = find_program('build/make/rtcd.py', required: true) + + extra_libs += c.find_library('m', required: false) +-- +2.44.0.windows.1 + + +From 26d95c89ad8da6f958e909783c3c353d7ee4ca86 Mon Sep 17 00:00:00 2001 +From: "L. E. Segovia" <amy@centricular.com> +Date: Thu, 11 Jul 2024 01:10:29 +0000 +Subject: [PATCH] meson: Fix wrong Xcode testing for armv7 + +Part-of: <https://gitlab.freedesktop.org/gstreamer/meson-ports/libvpx/-/merge_requests/21> + +diff --git a/meson.build b/meson.build +index fee067173..df50b53d4 100644 +--- a/meson.build ++++ b/meson.build +@@ -765,7 +765,7 @@ if tgt_isa.startswith('arm') + check: true + ).stdout().strip().split('\n').get(0, '').split() + +- if features.get('neon', false) and xcode_version.get(1, '').version_compare('>=6.3') ++ if features.get('neon', false) and not xcode_version.get(1, '').version_compare('>=6.3') + feature = 'neon' + if not features.get(feature, false) + if not features.has(feature) +-- +2.44.0.windows.1 + + +From 2918ddf8a880cf0d9768b1c8464d288c2fcb87c0 Mon Sep 17 00:00:00 2001 +From: "L. E. Segovia" <amy@centricular.com> +Date: Sat, 20 Jul 2024 20:56:31 +0000 +Subject: [PATCH] meson: Fix discrepancy in option formatting between the RTCD + Perl and Python generators + +Part-of: <https://gitlab.freedesktop.org/gstreamer/meson-ports/libvpx/-/merge_requests/21> + +diff --git a/meson.build b/meson.build +index df50b53d4..0ad31918a 100644 +--- a/meson.build ++++ b/meson.build +@@ -1816,7 +1816,20 @@ else + asm_compiler = disabler() + endif + +-rtcd_h_template = ['--arch=@0@'.format(tgt_isa), '--config=@INPUT0@'] + rtcd_options ++rtcd_h_template = ['--arch=@0@'.format(tgt_isa), '--config=@INPUT0@'] ++ ++original_rtcd_options = [] ++merge_rtcd_opt = '' ++foreach opt: rtcd_options ++ if merge_rtcd_opt != '' ++ original_rtcd_options += ['@0@-@1@'.format(merge_rtcd_opt, opt)] ++ merge_rtcd_opt = '' ++ elif opt == '--disable' ++ merge_rtcd_opt = opt ++ else ++ original_rtcd_options += opt ++ endif ++endforeach + + doxy_template = find_program('meson/generate_doxy.py', required: true) + +@@ -1849,7 +1862,7 @@ foreach symbol, input_file : codec_rtcds + rtcd_py = files(input_file) + + rtcd_h = configure_file( +- command: [rtcd_exe] + rtcd_h_template + ['--sym=@0@'.format(symbol), '@INPUT1@'], ++ command: [rtcd_exe] + rtcd_h_template + rtcd_options + ['--sym=@0@'.format(symbol), '@INPUT1@'], + input: [config_mk] + rtcd_py, + output: rtcd_filename, + capture: true, +@@ -1864,7 +1877,7 @@ foreach symbol, input_file : codec_rtcds + ) + + original_rtcd_h = configure_file( +- command: [original_rtcd_exe] + rtcd_h_template + ['--sym=@0@'.format(symbol), '@INPUT1@'], ++ command: [original_rtcd_exe] + rtcd_h_template + original_rtcd_options + ['--sym=@0@'.format(symbol), '@INPUT1@'], + input: [config_mk] + original_rtcd_pl, + output: 'original@0@'.format(rtcd_filename), + capture: true, +-- +2.44.0.windows.1 + + +From 31fdd5dd78347b2468d8a3c4a946f21d230cf19b Mon Sep 17 00:00:00 2001 +From: "L. E. Segovia" <amy@centricular.com> +Date: Tue, 6 Aug 2024 17:00:29 -0300 +Subject: [PATCH] vp9: Fix Apple silicon build + +ld: Undefined symbols: + _vp9_block_error_fp_neon, referenced from: + _block_yrd in vp9_encoder_vp9_pickmode.c.o + _vp9_block_error_neon, referenced from: + _rd_pick_intra_sub_8x8_y_mode in vp9_encoder_vp9_rdopt.c.o + _rd_pick_intra_sub_8x8_y_mode in vp9_encoder_vp9_rdopt.c.o + _rd_pick_best_sub8x8_mode in vp9_encoder_vp9_rdopt.c.o + _dist_block in vp9_encoder_vp9_rdopt.c.o + _vp9_setup_tpl_stats in vp9_encoder_vp9_tpl_model.c.o + +Part-of: <https://gitlab.freedesktop.org/gstreamer/meson-ports/libvpx/-/merge_requests/21> + +diff --git a/vp9/meson.build b/vp9/meson.build +index e9314966c..c1c63f044 100644 +--- a/vp9/meson.build ++++ b/vp9/meson.build +@@ -131,7 +131,7 @@ vp9_cx_optional_sources = { + # 'encoder/arm/neon/vp9_highbd_temporal_filter_neon.c', # MANUAL + 'encoder/arm/neon/vp9_dct_neon.c', + # 'encoder/arm/neon/vp9_denoiser_neon.c', # MANUAL +- # 'encoder/arm/neon/vp9_error_neon.c', # MANUAL ++ 'encoder/arm/neon/vp9_error_neon.c', + 'encoder/arm/neon/vp9_frame_scale_neon.c', + 'encoder/arm/neon/vp9_quantize_neon.c', + # 'encoder/arm/neon/vp9_highbd_error_neon.c', # MANUAL +@@ -380,9 +380,6 @@ if features.get('vp9_highbitdepth', false) + } + else + vp9_cx_optional_sources += { +- 'neon' : vp9_cx_optional_sources['neon'] + files( +- 'encoder/arm/neon/vp9_error_neon.c', +- ), + 'msa' : vp9_cx_optional_sources['msa'] + files( + 'encoder/mips/msa/vp9_fdct4x4_msa.c', + 'encoder/mips/msa/vp9_fdct8x8_msa.c', +-- +2.44.0.windows.1 + |