Dylan Baker <dy...@pnwbakers.com> writes:
> ---
> meson.build | 30 ++++++++++++++++++++++++++----
> meson_options.txt | 7 +++++++
> 2 files changed, 33 insertions(+), 4 deletions(-)
>
> diff --git a/meson.build b/meson.build
> index 0171328bcd7..f6fe141056a 100644
> --- a/meson.build
> +++ b/meson.build
> @@ -601,13 +601,33 @@ if with_gallium_st_nine
> endif
> endif
>
> +if get_option('altivec') != 'false'
> + if host_machine.cpu_family() == 'ppc64le'
> + if cc.get_id() == 'gcc' and cc.version().version_compare('< 4.8')
> + error('Altivec is not supported with gcc version < 4.8.')
> + endif
> + if cc.compiles('''
> + #include <altivec.h>
> + int main() {
> + vector unsigned char r;
> + vector unsigned int v = vec_splat_u32 (1);
> + r = __builtin_vec_vgbbd ((vector unsigned char) v);
> + return 0;
> + }''',
> + name : 'POWER8 intrinsics')
> + pre_args += ['-D_ARCH_PWR8', '-mpower8-vector']
> + elif get_option('altivec') == 'true'
> + error('Altivec support required but not found.')
> + endif
> + endif
> +endif
> +The autotools build seems to do this test with -mpower8-vector in the cflags, and that seems like it would probably be necessary. > diff --git a/meson_options.txt b/meson_options.txt > index 2c1f514debe..5976e14fc71 100644 > --- a/meson_options.txt > +++ b/meson_options.txt > @@ -298,3 +298,10 @@ option( > choices : ['freedreno', 'glsl', 'intel', 'nir', 'nouveau', 'all'], > description : 'List of tools to build.', > ) > +option( > + 'altivec', > + type : 'combo', > + value : 'auto', > + choices : ['auto', 'true', 'false'], > + description : 'Enable altivec optimizations if supported by host machine.', > +) > -- This should probably be power8 rather than altivec, since altivec predates power8 by a long time.
signature.asc
Description: PGP signature
_______________________________________________ mesa-dev mailing list mesa-dev@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/mesa-dev
