[Git][pulseaudio/webrtc-audio-processing][master] 10 commits: meson: Drop obsolete WEBRTC_DETECT_NEON macro
Arun Raghavan (@arun)
gitlab at gitlab.freedesktop.org
Mon Dec 30 21:26:01 UTC 2024
Arun Raghavan pushed to branch master at PulseAudio / webrtc-audio-processing
Commits:
b22ce018 by Alper Nebi Yasak at 2024-12-30T14:16:43-05:00
meson: Drop obsolete WEBRTC_DETECT_NEON macro
Upstream has dropped runtime checks for NEON instructions around 2016,
and the WEBRTC_DETECT_NEON macro is removed along with it. Disable NEON
when building with -Dneon=runtime and omit a warning instead.
Link: https://webrtc.googlesource.com/src/+/e305d956c0717a28ca88cd8547e5b310dfa74594
Signed-off-by: Alper Nebi Yasak <alpernebiyasak at gmail.com>
- - - - -
8b025599 by Alper Nebi Yasak at 2024-12-30T14:16:43-05:00
meson: Set WEBRTC_HAS_NEON macro after handling neon build option
The WEBRTC_HAS_NEON macro that enables using NEON implementations is
unconditionally set for arm64 and the 'neon' build option is ignored,
assuming we always want to use the NEON-specific implementations instead
of generic ones. This is an OK assumption to make because arm64 CPUs
always support NEON.
But the code handling the build option ended up quite convoluted. As
part of cleaning up, set the relevant cflags after we handle the build
option. This also means that we can make 'runtime' fall back to 'no',
and disable NEON-specific code with -Dneon=no.
Signed-off-by: Alper Nebi Yasak <alpernebiyasak at gmail.com>
- - - - -
6c914be9 by Alper Nebi Yasak at 2024-12-30T14:16:43-05:00
meson: Make -Dneon=no disable neon-specific code
When the neon build option is set to 'no', we should disable optimized
implementations that use NEON. Change have_neon to false in that case,
so that we skip the flags and skip building NEON-specific files.
Signed-off-by: Alper Nebi Yasak <alpernebiyasak at gmail.com>
- - - - -
b7a194f8 by Alper Nebi Yasak at 2024-12-30T14:16:43-05:00
meson: Check arm neon support before parsing neon option
The main if statment for the NEON option has been quite convoluted. The
previous commits reduced what it does to a simple case: check NEON
support and set have_neon on 32-bit ARM CPUs. Do that near the
architecture definitions.
Signed-off-by: Alper Nebi Yasak <alpernebiyasak at gmail.com>
- - - - -
fc5a4946 by Alper Nebi Yasak at 2024-12-30T14:17:32-05:00
meson: Set 'auto' as the default neon option value
The default for the neon build option is 'no', which disabled NEON code
for 32-bit ARM but enabled it for ARM64. Now that 'no' can disable NEON
code for ARM64, the default should be 'auto' which would enable it where
possible. Handle the 'auto' value, and set it as the default.
Signed-off-by: Alper Nebi Yasak <alpernebiyasak at gmail.com>
- - - - -
2493b6e6 by Alper Nebi Yasak at 2024-12-30T14:18:00-05:00
meson: Raise error for setting 'neon' when unsupported
We can set -Dneon=yes on x86, which will fail during build because the
x86 compiler doesn't understand the resulting `-mfpu=neon` flag. Make the
'neon' build option cause an error in the setup stage if we didn't
detect hardware support for NEON.
Signed-off-by: Alper Nebi Yasak <alpernebiyasak at gmail.com>
- - - - -
73aed233 by Alper Nebi Yasak at 2024-12-30T14:18:44-05:00
meson: Use neon_opt to control building neon files
Using the have_neon boolean to enable NEON code means we have to either
fully enable it or fully disable it. When using -Dneon=runtime, ideally
only the parts that support runtime checks would be built for NEON, and
those that don't would be built without NEON. Though, there are no
longer any runtime checks for NEON anywhere, so it's equivalent to 'no'
with a warning.
In general, we should use have_* variables to indicate compiler support,
and *_opt options to choose if and how we want to utilize that. Use
neon_opt to control NEON compilation and avoid modifying have_neon which
now would fully refer to compiler support.
Signed-off-by: Alper Nebi Yasak <alpernebiyasak at gmail.com>
- - - - -
d63a2c97 by Alper Nebi Yasak at 2024-12-30T14:19:07-05:00
meson: pffft: Warn about not having runtime neon checks
The pffft.c file does not have runtime checks for NEON, and silently
falls back to disabling it when the neon option is 'runtime'. Print a
warning in this case.
Signed-off-by: Alper Nebi Yasak <alpernebiyasak at gmail.com>
- - - - -
54a632f0 by Arun Raghavan at 2024-12-30T16:21:08-05:00
meson: Convert the 'neon' option into a feature
Easier to express things, now that runtime is a no-op.
- - - - -
774ac54e by Arun Raghavan at 2024-12-30T16:21:20-05:00
meson: Only disable SIMD for non-SSE machines on x86
This ended up disabling SIMD everywhere except where SSE/AVX was
enabled.
- - - - -
5 changed files:
- meson.build
- meson_options.txt
- webrtc/common_audio/meson.build
- webrtc/modules/audio_processing/meson.build
- webrtc/third_party/pffft/meson.build
Changes:
=====================================
meson.build
=====================================
@@ -111,6 +111,7 @@ endif
arch_cflags = []
have_arm = false
have_armv7 = false
+have_arm64 = false
have_neon = false
have_mips = false
have_mips64 = false
@@ -130,12 +131,14 @@ if host_machine.cpu_family() == 'arm'
have_armv7 = true
arch_cflags += ['-DWEBRTC_ARCH_ARM_V7']
endif
+ if cc.compiles('#include <arm_neon.h>', args : '-mfpu=neon')
+ have_neon = true
+ endif
endif
-if cc.compiles('''#ifndef __aarch64__
-#error no aarch64 arch
-#endif''')
+if host_machine.cpu_family() == 'aarch64'
+ have_arm64 = true
have_neon = true
- arch_cflags += ['-DWEBRTC_ARCH_ARM64', '-DWEBRTC_HAS_NEON']
+ arch_cflags += ['-DWEBRTC_ARCH_ARM64']
endif
if ['mips', 'mips64'].contains(host_machine.cpu_family())
have_mips = true
@@ -160,16 +163,12 @@ if ['x86', 'x86_64'].contains(host_machine.cpu_family())
endif
endif
-neon_opt = get_option('neon')
-if neon_opt != 'no' and not have_neon
- if neon_opt != 'runtime'
- if cc.compiles('#include <arm_neon.h>', args : '-mfpu=neon')
- arch_cflags += ['-mfpu=neon', '-DWEBRTC_HAS_NEON']
- have_neon = true
- endif
- else
- arch_cflags += ['-DWEBRTC_DETECT_NEON', '-mfpu=neon']
- have_neon = true
+neon_opt = get_option('neon').require(have_neon)
+
+if neon_opt.enabled()
+ arch_cflags += ['-DWEBRTC_HAS_NEON']
+ if not have_arm64
+ arch_cflags += ['-mfpu=neon']
endif
endif
=====================================
meson_options.txt
=====================================
@@ -1,8 +1,8 @@
option('gnustl', type: 'feature',
value: 'auto',
description: 'Use gnustl for a c++ library implementation (only used on Android)')
-option('neon', type: 'combo',
- choices: ['no', 'yes', 'auto', 'runtime'],
+option('neon', type: 'feature',
+ value: 'auto',
description: 'Enable NEON optimisations')
option('inline-sse', type: 'boolean',
value: true,
=====================================
webrtc/common_audio/meson.build
=====================================
@@ -112,7 +112,7 @@ if have_armv7
]
endif
-if have_neon
+if neon_opt.enabled()
common_audio_sources += [
'fir_filter_neon.cc',
'resampler/sinc_resampler_neon.cc',
=====================================
webrtc/modules/audio_processing/meson.build
=====================================
@@ -180,7 +180,7 @@ else
]
endif
-if have_neon
+if neon_opt.enabled()
webrtc_audio_processing_sources += [
'aecm/aecm_core_neon.cc',
]
=====================================
webrtc/third_party/pffft/meson.build
=====================================
@@ -4,7 +4,7 @@ pffft_sources = [
pffft_cflags = [ '-D_GNU_SOURCE' ]
-if not have_inline_sse or (have_arm and not have_neon) or (have_mips and host_machine.endian() == 'little') or have_mips64
+if (have_x86 and not have_inline_sse) or (have_arm and not neon_opt.enabled()) or (have_mips and host_machine.endian() == 'little') or have_mips64
pffft_cflags += [ '-DPFFFT_SIMD_DISABLE' ]
endif
View it on GitLab: https://gitlab.freedesktop.org/pulseaudio/webrtc-audio-processing/-/compare/a1ccd6c7001c285d936d742cddeb473cf89db355...774ac54e71d71e222918e49baec1e9dc203d2ed8
--
View it on GitLab: https://gitlab.freedesktop.org/pulseaudio/webrtc-audio-processing/-/compare/a1ccd6c7001c285d936d742cddeb473cf89db355...774ac54e71d71e222918e49baec1e9dc203d2ed8
You're receiving this email because of your account on gitlab.freedesktop.org.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.freedesktop.org/archives/pulseaudio-commits/attachments/20241230/7557f4ff/attachment-0001.htm>
More information about the pulseaudio-commits
mailing list