Re: [PATCH] x86/hvm: Drop get_shadow_gs_base() hook and use hvm_get_reg()

[Jan Beulich <jbeulich@xxxxxxxx>]

On 21.01.2022 12:22, Andrew Cooper wrote:
> This is a trivial accessor for an MSR, so use hvm_get_reg() rather than a
> dedicated hook.  In arch_get_info_guest(), rework the logic to read GS_SHADOW
> only once.
> 
> get_hvm_registers() is called on current, meaning that diagnostics print a
> stale GS_SHADOW from the previous vcpu context switch.  Adjust both
> implementations to obtain the correct value.
> 
> Signed-off-by: Andrew Cooper <andrew.cooper3@xxxxxxxxxx>

Reviewed-by: Jan Beulich <jbeulich@xxxxxxxx>

Just one minor request for consideration at the end.

> If we care to backport the bugfix aspect, a far less invasive option would be
> to read MSR_SHADOW_GS_BASE directly.
> 
> The only case where that goes wrong is when vmcb->kerngsbase has been modified
> and is pending a VMLOAD.  I'm fairly sure this can only occur when we need
> vcpu diagnostics, after an emulated write of MSR_SHADOW_GS_BASE.

Hmm. Maybe best to leave alone in stable trees?

> @@ -2417,6 +2413,15 @@ static uint64_t vmx_get_reg(struct vcpu *v, unsigned 
> int reg)
>              domain_crash(d);
>          }
>          return val;
> +
> +    case MSR_SHADOW_GS_BASE:
> +        if ( v == curr )
> +        {
> +            rdmsrl(MSR_SHADOW_GS_BASE, val);
> +            return val;
> +        }
> +        else
> +            return v->arch.hvm.vmx.shadow_gs;
>      }

I think it wasn't too long ago that I saw you ask for an "else" like
this one to be dropped (in someone else's patch). May I ask that you
consider doing so here, perhaps going straight to the more compact

    case MSR_SHADOW_GS_BASE:
        if ( v != curr )
            return v->arch.hvm.vmx.shadow_gs;
        rdmsrl(MSR_SHADOW_GS_BASE, val);
        return val;

?

Actually, as I notice only now: Would you mind making "curr" here and
in the VMX equivalent pointer-to-const?

Jan