On 02/06/19 21:11, Uros Bizjak wrote:
> __vmcs_writel uses volatile asm, so there is no need to insert another
> one between the first and the second call to __vmcs_writel in order
> to prevent unwanted code moves for 32bit targets.
>
> Signed-off-by: Uros Bizjak <ubizjak@xxxxxxxxx>
> ---
> arch/x86/kvm/vmx/ops.h | 1 -
> 1 file changed, 1 deletion(-)
>
> diff --git a/arch/x86/kvm/vmx/ops.h b/arch/x86/kvm/vmx/ops.h
> index b8e50f76fefc..2200fb698dd0 100644
> --- a/arch/x86/kvm/vmx/ops.h
> +++ b/arch/x86/kvm/vmx/ops.h
> @@ -146,7 +146,6 @@ static __always_inline void vmcs_write64(unsigned long field, u64 value)
>
> __vmcs_writel(field, value);
> #ifndef CONFIG_X86_64
> - asm volatile ("");
> __vmcs_writel(field+1, value >> 32);
> #endif
> }
>
Queued, thanks.
Paolo
