Re: [RFC 01/14] x86/apic: Add new driver for Secure AVIC

[Borislav Petkov]

On Wed, Oct 09, 2024 at 01:10:58PM +0300, Kirill A. Shutemov wrote:
> I don't think CC attributes is the right way to track this kind of
> features. My understanding of cc_platform interface is that it has to be
> used to advertise some kind of property of the platform that generic code
> and be interested in, not a specific implementation.

Yes.

> 
> For the same reason, I think CC_ATTR_GUEST/HOST_SEV_SNP is also a bad use
> of the interface.
> 
> Borislav, I know we had different view on this. What is your criteria on
> what should and shouldn't be a CC attribute? I don't think we want a
> parallel X86_FEATURE_*.

Yes, we don't.

Do you have a better idea which is cleaner than what we do now?

Yes yes, cc_platform reports aspects of the coco platform to generic code but
nothing stops the x86 code from calling those interfaces too, for simplicity
reasons.

Because the coco platform being a SNP guest or having an SAVIC are also two
aspects of that same confidential computing platform. So we might as well use
it this way too.

I'd say.

-- 
Regards/Gruss,
    Boris.

https://people.kernel.org/tglx/notes-about-netiquette