On 14/10/15 14:29, Jan Beulich wrote:
> x86-64 requires FXSR, XMM, and XMM2, so there's no point in hiding
> respective code behind conditionals.
>
> Signed-off-by: Jan Beulich <jbeul...@suse.com>
Reviewed-by: Andrew Cooper <andrew.coop...@citrix.com>, with two
suggestions.
> --- a/xen/arch/x86/setup.c
> +++ b/xen/arch/x86/setup.c
> @@ -1295,10 +1295,8 @@ void __init noreturn __start_xen(unsigne
>
> identify_cpu(&boot_cpu_data);
>
> - if ( cpu_has_fxsr )
> - set_in_cr4(X86_CR4_OSFXSR);
> - if ( cpu_has_xmm )
> - set_in_cr4(X86_CR4_OSXMMEXCPT);
> + set_in_cr4(X86_CR4_OSFXSR);
> + set_in_cr4(X86_CR4_OSXMMEXCPT);
set_in_cr4() takes a mask, so could take both of these at once, to be
slightly more efficient.
If the bsp path gained an unconditional write_cr4(mmu_cr4_features) then
it might be better to fold them into the initialiser for mmu_cr4_features.
> --- a/xen/include/asm-x86/page.h
> +++ b/xen/include/asm-x86/page.h
> @@ -206,13 +206,10 @@ typedef struct { u64 pfn; } pagetable_t;
> #define pagetable_null() pagetable_from_pfn(0)
>
> void clear_page_sse2(void *);
> -#define clear_page(_p) (cpu_has_xmm2 ? \
> - clear_page_sse2((void *)(_p)) : \
> - (void)memset((void *)(_p), 0, PAGE_SIZE))
> void copy_page_sse2(void *, const void *);
> -#define copy_page(_t,_f) (cpu_has_xmm2 ? \
> - copy_page_sse2(_t, _f) : \
> - (void)memcpy(_t, _f, PAGE_SIZE))
> +
> +#define clear_page(_p) clear_page_sse2(_p)
> +#define copy_page(_t,_f) copy_page_sse2(_t, _f)
As you are changing this, an extra space following the comma.
~Andrew
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xen.org
http://lists.xen.org/xen-devel
