Re: [PATCH 4/4] kvm: svm: Support KVM_SEV_SNP_PAGE_TYPE_VMSA at SNP_LAUNCH_UPDATE

Jörg Rödel <[email protected]> Wed, 17 Jun 2026 16:53:53 +0200
Newsgroups dev.linux.lists.coconut-svsm,org.kernel.vger.kvm,org.kernel.vger.linux-kernel
Message-ID <[email protected]>
On Wed, Jun 17, 2026 at 09:45:46AM -0400, James Bottomley wrote:
> Well if the guest policy is I don't care how many CPUs you give me then
> certainly, yes.  However, if the guest does care, they may want an
> attestable record of it somewhere.  Since clouds do charge somewhat
> per-vCPU I can see this mattering to some tenants.

Yes, the number of VCPUs certainly matters to most tenants. But I doubt there
is much value in including this information into any measurment, as the
measurement does not guarantee the tenant that the host actually executes the
VCPU. The best proof for the existence of any given VCPU is the fact that it
executes code.

Regards,

	Joerg