On Fri, Jan 20, 2023, Zhi Wang wrote: > On Tue, 17 Jan 2023 20:56:17 +0000 Sean Christopherson <seanjc@xxxxxxxxxx> wrote: > What I would like to clarify is: it would be better to have a mechanism to make sure a > SEAMCALL can succeed at a certain point (or at least trying to succeed) as the > magic number would work like a lottery in reality. Yep, you and I are in agreement on this point. Ideally, KVM would be able to guarantee success and treat TDX_OPERAND_BUSY errors as KVM bugs.
