On Mon, Nov 06, 2023 at 04:15:59PM +0530, Nikunj A. Dadhania wrote:
> On 11/2/2023 4:06 PM, Kirill A. Shutemov wrote:
> > On Thu, Nov 02, 2023 at 11:11:52AM +0530, Nikunj A. Dadhania wrote:
> >> On 10/31/2023 1:56 AM, Tom Lendacky wrote:
> >>>> diff --git a/include/linux/cc_platform.h b/include/linux/cc_platform.h
> >>>> index cb0d6cd1c12f..e081ca4d5da2 100644
> >>>> --- a/include/linux/cc_platform.h
> >>>> +++ b/include/linux/cc_platform.h
> >>>> @@ -90,6 +90,14 @@ enum cc_attr {
> >>>> * Examples include TDX Guest.
> >>>> */
> >>>> CC_ATTR_HOTPLUG_DISABLED,
> >>>> +
> >>>> + /**
> >>>> + * @CC_ATTR_GUEST_SECURE_TSC: Secure TSC is active.
> >>>> + *
> >>>> + * The platform/OS is running as a guest/virtual machine and actively
> >>>> + * using AMD SEV-SNP Secure TSC feature.
> >>>
> >>> I think TDX also has a secure TSC like feature, so can this be generic?
> >>
> >> Yes, we can do that. In SNP case SecureTSC is an optional feature, not sure if that is the case for TDX as well.
> >>
> >> Kirill any inputs ?
> >
> > We have several X86_FEATURE_ flags to indicate quality of TSC. Do we
> > really need a CC_ATTR on top of that? Maybe SEV code could just set
> > X86_FEATURE_ according to what its TSC can do?
>
> For SEV-SNP, SEV_STATUS MSR has the information of various features
> that have been enabled by the hypervisor. We will need a CC_ATTR for
> these optional features.
If all users of the attribute is withing x86, I would rather add synthetic
X86_FEATURE_ flags than CC_ATTR_. We have better instrumentation around
features.
--
Kiryl Shutsemau / Kirill A. Shutemov
