Re: [PATCH 2/2] KVM: nVMX: Defer SVI update to vmcs01 on EOI when L2 is active w/o VID

[Sean Christopherson]

On Mon, Nov 04, 2024, Chao Gao wrote:
> >@@ -6873,6 +6873,23 @@ void vmx_hwapic_isr_update(struct kvm_vcpu *vcpu, int max_isr)
> > 	u16 status;
> > 	u8 old;
> > 
> >+	/*
> >+	 * If L2 is active, defer the SVI update until vmcs01 is loaded, as SVI
> >+	 * is only relevant for if and only if Virtual Interrupt Delivery is
> >+	 * enabled in vmcs12, and if VID is enabled then L2 EOIs affect L2's
> >+	 * vAPIC, not L1's vAPIC.  KVM must update vmcs01 on the next nested
> >+	 * VM-Exit, otherwise L1 with run with a stale SVI.
> >+	 */
> >+	if (is_guest_mode(vcpu)) {
> >+		/*
> >+		 * KVM is supposed to forward intercepted L2 EOIs to L1 if VID
> >+		 * is enabled in vmcs12; as above, the EOIs affect L2's vAPIC.
> >+		 */
> >+		WARN_ON_ONCE(nested_cpu_has_vid(get_vmcs12(vcpu)));
> 
> This function can be called in other scenarios, e.g., from
> kvm_apic_set_state(). Is it possible for the userspace VMM to set the APIC
> state while L2 is running and VID is enabled for L2? If so, this warning could
> be triggered by the userspace VMM.

Ugh, yeah, that's reachable.  Bummer.  Ooh!  I think we can use the recently
added "wants_to_run" to filter out state stuffing.  I.e. this

		WARN_ON_ONCE(vcpu->wants_to_run &&
			     nested_cpu_has_vid(get_vmcs12(vcpu)));

plus an updated comment.