On 23/07/2018 23:17, Isaku Yamahata wrote:
> -void kvm_set_pending_timer(struct kvm_vcpu *vcpu)
> -{
> - /*
> - * Note: KVM_REQ_PENDING_TIMER is implicitly checked in
> - * vcpu_enter_guest. This function is only called from
> - * the physical CPU that is running vcpu.
> - */
> - kvm_make_request(KVM_REQ_PENDING_TIMER, vcpu);
> -}
> -
> static void kvm_write_wall_clock(struct kvm *kvm, gpa_t wall_clock)
> {
> int version;
> @@ -7592,7 +7582,6 @@ static int vcpu_run(struct kvm_vcpu *vcpu)
> if (r <= 0)
> break;
>
> - kvm_clear_request(KVM_REQ_PENDING_TIMER, vcpu);
> if (kvm_cpu_has_pending_timer(vcpu))
> kvm_inject_pending_timer_irqs(vcpu);
The comment that you are removing is still true, however. The bit is
checked here:
if (vcpu->mode == EXITING_GUEST_MODE || kvm_request_pending(vcpu)
|| need_resched() || signal_pending(current)) {
and your change is introducing a race when the preemption timer is not
in use.
Paolo
