KVM: nVMX: Don't halt vcpu when L1 is injecting events to L2
authorChao Gao <chao.gao@intel.com>
Sun, 11 Feb 2018 02:06:30 +0000 (10:06 +0800)
committerGreg Kroah-Hartman <gregkh@linuxfoundation.org>
Wed, 30 May 2018 05:52:02 +0000 (07:52 +0200)
[ Upstream commit 135a06c3a515bbd17729eb04f4f26316d48363d7 ]

Although L2 is in halt state, it will be in the active state after
VM entry if the VM entry is vectoring according to SDM 26.6.2 Activity
State. Halting the vcpu here means the event won't be injected to L2
and this decision isn't reported to L1. Thus L0 drops an event that
should be injected to L2.

Cc: Liran Alon <liran.alon@oracle.com>
Reviewed-by: Liran Alon <liran.alon@oracle.com>
Signed-off-by: Chao Gao <chao.gao@intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Sasha Levin <alexander.levin@microsoft.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
arch/x86/kvm/vmx.c

index 2e63edf..03c298d 100644 (file)
@@ -11174,7 +11174,12 @@ static int nested_vmx_run(struct kvm_vcpu *vcpu, bool launch)
        if (ret)
                return ret;
 
-       if (vmcs12->guest_activity_state == GUEST_ACTIVITY_HLT)
+       /*
+        * If we're entering a halted L2 vcpu and the L2 vcpu won't be woken
+        * by event injection, halt vcpu.
+        */
+       if ((vmcs12->guest_activity_state == GUEST_ACTIVITY_HLT) &&
+           !(vmcs12->vm_entry_intr_info_field & INTR_INFO_VALID_MASK))
                return kvm_vcpu_halt(vcpu);
 
        vmx->nested.nested_run_pending = 1;