Re: [PATCH v10 10/28] KVM: VMX: Add support for saving and restoring FRED MSRs

[Sohil Mehta]

>> Specially, intercept accesses to the FRED SSP0 MSR (IA32_PL0_SSP), which
>> remains accessible when FRED is enumerated even if CET is not. 
> 
> Nit:
> It sounds like the patch adds interception. But the patch emulates the access
> of the MSR. 
> 

Yes, will change the wording to reflect that only the emulation of the
access is being added.

>> This
>> ensures the guest value is fully virtual and does not alter the hardware
>> FRED SSP0 MSR.
>>

...

>> diff --git a/arch/x86/include/asm/kvm_host.h b/arch/x86/include/asm/kvm_host.h
>> index 8d3999772ac2..0fd5d4edeebb 100644
>> --- a/arch/x86/include/asm/kvm_host.h
>> +++ b/arch/x86/include/asm/kvm_host.h
>> @@ -1032,6 +1032,11 @@ struct kvm_vcpu_arch {
>>  #if IS_ENABLED(CONFIG_HYPERV)
>>  	hpa_t hv_root_tdp;
>>  #endif
>> +	/*
>> +	 * Stores the FRED SSP0 MSR when CET is not supported, prompting KVM
> Nit:
> CET could be more specific, i.e. SHSTK. I am not sure if there is a situation
> that one is supported, but the other is not.
> 

Agreed SHSTK is more precise that CET. I think KVM can clear SHSTK but
keep IBT in some situations.

> "prompting KVM to intercept its accesses" is confusing to me.

I believe it meant to say that fred_ssp0_fallback stores the SSP0 MSR
value and the intercepted accesses are emulated using this field. I'll
reword it to be less ambiguous.

>> +	 * to intercept its accesses.
>> +	 */
>> +	u64 fred_ssp0_fallback;
>>  };
> 
> [...]