Re: [PATCH v2 3/3] x86: Add support for CpuidUserDis

[Jan Beulich <jbeulich@xxxxxxxx>]

On 09.05.2023 18:43, Alejandro Vallejo wrote:
> Because CpuIdUserDis is reported in CPUID itself, the extended leaf
> containing that bit must be retrieved before calling c_early_init()
> 
> Signed-off-by: Alejandro Vallejo <alejandro.vallejo@xxxxxxxxx>

Looks largely okay when taken together with patch 2, but ...

> --- a/xen/arch/x86/cpu/amd.c
> +++ b/xen/arch/x86/cpu/amd.c
> @@ -279,8 +279,12 @@ static void __init noinline amd_init_levelling(void)
>        * that can only be present when Xen is itself virtualized (because
>        * it can be emulated)
>        */
> -     if (cpu_has_hypervisor && probe_cpuid_faulting())
> +     if ((cpu_has_hypervisor && probe_cpuid_faulting()) ||
> +         boot_cpu_has(X86_FEATURE_CPUID_USER_DIS)) {

... imo the probe_cpuid_faulting() call would better be avoided when
the CPUID bit is set.

> +             expected_levelling_cap |= LCAP_faulting;
> +             levelling_caps |=  LCAP_faulting;

Further the movement of these two lines from ...

> @@ -144,8 +145,6 @@ bool __init probe_cpuid_faulting(void)
>               return false;
>       }
>  
> -     expected_levelling_cap |= LCAP_faulting;
> -     levelling_caps |=  LCAP_faulting;
>       setup_force_cpu_cap(X86_FEATURE_CPUID_FAULTING);

... here (and also to intel.c) should imo be part of patch 2. While
moving them, I think you also want to deal with the stray double
blank.

Jan