> >> > >> Not if it's masked by WIE -- and even when masked by CE, it's a bug > >> that we currently consider the vcpu runnable. We shouldn't depend on that > bug. > > > > Scott can you please describe what is bug? > > If an interrupt is masked by EE, CE, ME, etc. it is still in > pending_exceptions, > so runnable still returns true, and we can't go idle.
Hmm, ok. > > > What I remember is that if > > vcpu is not run-able then we halt vcpu and cannot cause qemu exit > > also. > > I agree that we want to be considered runnable if we have a final expiration > with an action. What I disagree with is using the same pending_exceptions bit > as is used for the ordinary watchdog interrupt. > They're not the same thing. Now I agree with your disagreement :) Thanks -Bharat
