On 11/09/2018 18:50, Liran Alon wrote: > > >> On 11 Sep 2018, at 16:16, Paolo Bonzini <pbonzini@xxxxxxxxxx> wrote: >> >> On 30/08/2018 11:57, Liran Alon wrote: >>> Consider the case L1 had a IRQ/NMI event until it executed >>> VMLAUNCH/VMRESUME which wasn't delivered because it was disallowed >>> (e.g. interrupts disabled). When L1 executes VMLAUNCH/VMRESUME, >>> L0 needs to evaluate if this pending event should cause an exit from >>> L2 to L1 or delivered directly to L2 (e.g. In case L1 don't intercept >>> EXTERNAL_INTERRUPT). >>> >>> Usually this would be handled by L0 requesting a IRQ/NMI window >>> by setting VMCS accordingly. However, this setting was done on >>> VMCS01 and now VMCS02 is active instead. Thus, when L1 executes >>> VMLAUNCH/VMRESUME we force L0 to perform pending event evaluation by >>> requesting a KVM_REQ_EVENT. >>> >>> Note that above scenario exists when L1 KVM is about to enter L2 but >>> requests an "immediate-exit". As in this case, L1 will >>> disable-interrupts and then send a self-IPI before entering L2. >>> >>> Co-authored-by: Sean Christopherson <sean.j.christopherson@xxxxxxxxx> >>> Signed-off-by: Liran Alon <liran.alon@xxxxxxxxxx> >>> --- >>> arch/x86/kvm/vmx.c | 25 ++++++++++++++++++++++++- >>> 1 file changed, 24 insertions(+), 1 deletion(-) >> >> Any chance you can write a testcase for selftests/kvm? The framework >> should be more or less stable by now. >> >> Paolo > > Actually, I have already written one and submitted it a week ago to the mailing list. > You can find the relevant unit-tests here: > https://patchwork.kernel.org/project/kvm/list/?series=14789 The test doesn't pass with the new patch. Also, the patch that was applied lacks the final hunk: @@ -12702,7 +12724,8 @@ static int nested_vmx_run(struct kvm_vcpu *vcpu, bool launch) * by event injection, halt vcpu. */ if ((vmcs12->guest_activity_state == GUEST_ACTIVITY_HLT) && - !(vmcs12->vm_entry_intr_info_field & INTR_INFO_VALID_MASK)) { + !(vmcs12->vm_entry_intr_info_field & INTR_INFO_VALID_MASK) && + !kvm_test_request(KVM_REQ_EVENT, vcpu)) { vmx->nested.nested_run_pending = 0; return kvm_vcpu_halt(vcpu); } and the test doesn't pass even if I add this hunk. (The v1 patch works, but it is a much bigger hammer). Anybody has time to take a look? In the meanwhile I'm pushing the failing test to kvm-unit-tests so that we don't forget. Paolo
