Patchwork [RFC] KVM: x86: svm: make sure NMI is injected after nmi_singlestep

mail settings
Submitter Vitaly Kuznetsov
Date March 15, 2019, 3:37 p.m.
Message ID <>
Download mbox | patch
Permalink /patch/749691/
State New
Headers show


Vitaly Kuznetsov - March 15, 2019, 3:37 p.m.
I noticed that apic test from kvm-unit-tests always hangs on my EPYC 7401P,
the hanging test nmi-after-sti is trying to deliver 30000 NMIs and tracing
shows that we're sometimes able to deliver a few but never all.

When we're trying to inject an NMI we may fail to do so immediately for
various reasons, however, we still need to inject it so enable_nmi_window()
arms nmi_singlestep mode. #DB occurs as expected, but we're not checking
for pending NMIs before entering the guest and unless there's a different
event to process, the NMI will never get delivered.

Make KVM_REQ_EVENT request on the vCPU from db_interception() to make sure
pending NMIs are checked and possibly injected.

Signed-off-by: Vitaly Kuznetsov <>
- RFC: I'm not sure this is not part of some bigger issue I'm missing.
 arch/x86/kvm/svm.c | 3 +++
 1 file changed, 3 insertions(+)


diff --git a/arch/x86/kvm/svm.c b/arch/x86/kvm/svm.c
index f13a3a24d360..cc96008a6117 100644
--- a/arch/x86/kvm/svm.c
+++ b/arch/x86/kvm/svm.c
@@ -2687,6 +2687,7 @@  static int npf_interception(struct vcpu_svm *svm)
 static int db_interception(struct vcpu_svm *svm)
 	struct kvm_run *kvm_run = svm->;
+	struct kvm_vcpu *vcpu = &svm->vcpu;
 	if (!(svm->vcpu.guest_debug &
@@ -2697,6 +2698,8 @@  static int db_interception(struct vcpu_svm *svm)
 	if (svm->nmi_singlestep) {
+		/* Make sure we check for pending NMIs upon entry */
+		kvm_make_request(KVM_REQ_EVENT, vcpu);
 	if (svm->vcpu.guest_debug &