> >> If the second exception does not cause a vmexit, it is handled as > >> usual by the processor (by checking if the two exceptions are benign, > >> contributory or page faults). The behavior is the same even if the > >> first exception comes from VMX event injection. > > > > The case I was thinking is, both the first and the second exception > > don't cause any VM exit, however the first exception triggered an EPT > > violation. Later KVM injects the first exception and delivering of the > > first exception by the CPU triggers the second exception, then the > > information about the first KVM-injected exception is lost, and it can > > be re-generated once the second exception is correctly handled. > > That's not a problem, the behavior is the same as on bare metal (depending on > whether the two exceptions are benign/contributory/page faults). My point is that if KVM doesn't inject the first exception in this specific case after it handles EPT violation, as you said, the behavior is still the same as on bare metal. It makes no difference whether to inject it *in this specific case*. But as you and Sean said, it can't deal the case that L1 injects an exception not related to the code that L2 is executing. Thanks! Xin
