On Fri, May 31, 2024 at 11:14:53AM +0200, Paolo Bonzini wrote:
> Date: Fri, 31 May 2024 11:14:53 +0200
> From: Paolo Bonzini <pbonzini@redhat.com>
> Subject: [PATCH 2/6] meson: assume x86-64-v2 baseline ISA
> X-Mailer: git-send-email 2.45.1
>
> x86-64-v2 processors were released in 2008, assume that we have one.
> Unfortunately there is no GCC flag to enable all the features
> without disabling what came after; so enable them one by one.
>
> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
> ---
> meson.build | 10 +++++++---
> 1 file changed, 7 insertions(+), 3 deletions(-)
>
> diff --git a/meson.build b/meson.build
> index 63866071445..19d1fc1f33b 100644
> --- a/meson.build
> +++ b/meson.build
> @@ -336,9 +336,13 @@ if host_arch == 'i386' and not cc.links('''
> qemu_common_flags = ['-march=i486'] + qemu_common_flags
> endif
>
> -# ??? Only extremely old AMD cpus do not have cmpxchg16b.
> -# If we truly care, we should simply detect this case at
> -# runtime and generate the fallback to serial emulation.
> +# Assume x86-64-v2 (minus CMPXCHG16B for 32-bit code)
Is it necessary to state the requirement (x86-64-v2) for x86 host in
some doc?
e.g., docs/system/target-i386.rst.
> +if host_arch == 'i386'
> + qemu_common_flags = ['-mfpmath=sse'] + qemu_common_flags
> +endif
> +if host_arch in ['i386', 'x86_64']
> + qemu_common_flags = ['-mpopcnt', '-msse4.2'] + qemu_common_flags
> +endif
> if host_arch == 'x86_64'
> qemu_common_flags = ['-mcx16'] + qemu_common_flags
> endif
> --
> 2.45.1
Reviewed-by: Zhao Liu <zhao1.liu@intel.com>