Re: [PATCH v2] KVM: x86: Inject pending interrupt even if pending nmi exist

[Paolo Bonzini <pbonzini@redhat.com>]

On 23/03/2016 06:08, Yuki Shibuya wrote:
> +		/* If nmi pending > 0 and injectable interrupts exist,
> +		 * nmi pending counter is cleared to prevent skipping
> +		 * injectable pending interrupts.
> +		 */
> +		if (vcpu->arch.nmi_pending && kvm_cpu_has_injectable_intr(vcpu)
> +					&& kvm_x86_ops->interrupt_allowed(vcpu))
> +			vcpu->arch.nmi_pending = 0;

I am not sure I understand this.  Why is it safe to drop nmi_pending?
Can we instead do something like this in vcpu_enter_guest?

It would be nice to have a testcase for this in kvm-unit-tests.

Paolo

diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c
index 7236bd3a4c3d..6c73cbc8e19a 100644
--- a/arch/x86/kvm/x86.c
+++ b/arch/x86/kvm/x86.c
@@ -6561,10 +6561,12 @@ static int vcpu_enter_guest(struct kvm_vcpu *vcpu)
 		if (inject_pending_event(vcpu, req_int_win) != 0)
 			req_immediate_exit = true;
 		/* enable NMI/IRQ window open exits if needed */
-		else if (vcpu->arch.nmi_pending)
-			kvm_x86_ops->enable_nmi_window(vcpu);
-		else if (kvm_cpu_has_injectable_intr(vcpu) || req_int_win)
-			kvm_x86_ops->enable_irq_window(vcpu);
+		else {
+			if (vcpu->arch.nmi_pending)
+				kvm_x86_ops->enable_nmi_window(vcpu);
+			if (kvm_cpu_has_injectable_intr(vcpu) || req_int_win)
+				kvm_x86_ops->enable_irq_window(vcpu);
+		}
 
 		if (kvm_lapic_enabled(vcpu)) {
 			update_cr8_intercept(vcpu);



Paolo