Re: [PATCH 1/2] kvm-unit-test: x86: Add a wrapper to check if the CPU supports NX bit in MSR_EFER

[Sean Christopherson <sean.j.christopherson@intel.com>]

On Wed, May 22, 2019 at 07:45:44PM -0400, Krish Sadhukhan wrote:
> Signed-off-by: Krish Sadhukhan <krish.sadhukhan@oracle.com>
> Reviewed-by: Karl Heubaum <karl.heubaum@oracle.com>
> ---
>  lib/x86/processor.h | 8 ++++++++
>  x86/vmexit.c        | 2 +-
>  2 files changed, 9 insertions(+), 1 deletion(-)
> 
> diff --git a/lib/x86/processor.h b/lib/x86/processor.h
> index 15237a5..2ca988e 100644
> --- a/lib/x86/processor.h
> +++ b/lib/x86/processor.h
> @@ -476,4 +476,12 @@ static inline void set_bit(int bit, u8 *addr)
>  			     : "+m" (*addr) : "Ir" (bit) : "cc", "memory");
>  }
>  
> +static inline int efer_nx_enabled(void)

cpu_has_efer_nx() would be more appropriate.  Most readers would expect
"enabled" to mean we're checking MSR_EFER.NX==1.

This can have a boolean return value.

> +{
> +	if (cpuid(0x80000001).d & (1 << 20))
> +		return 1;
> +	else
> +		return 0;
> +}

This can simply be:

    return cpuid(0x80000001).d & (1 << 20);

or if gcc complains about boolean stuff:

    return !!(cpuid(0x80000001).d & (1 << 20));
> +
>  #endif
> diff --git a/x86/vmexit.c b/x86/vmexit.c
> index c12dd24..7053a46 100644
> --- a/x86/vmexit.c
> +++ b/x86/vmexit.c
> @@ -526,7 +526,7 @@ static bool do_test(struct test *test)
>  
>  static void enable_nx(void *junk)
>  {
> -	if (cpuid(0x80000001).d & (1 << 20))
> +	if (efer_nx_enabled())
>  		wrmsr(MSR_EFER, rdmsr(MSR_EFER) | EFER_NX_MASK);
>  }
>  
> -- 
> 2.20.1
>