2012/11/27 Gleb Natapov [off-list ref]:

On Tue, Nov 27, 2012 at 01:15:25PM +0800, Li Zhong wrote:

quoted

@@ -247,10 +247,17 @@ do_async_page_fault(struct pt_regs *regs, unsigned long error_code)
              break;
      case KVM_PV_REASON_PAGE_NOT_PRESENT:
              /* page is swapped out by the host. */
-             rcu_irq_enter();
+             if (is_idle_task(current) && rcu_is_cpu_idle())
+                     rcu_idle_exit();
+             else
+                     rcu_user_exit();
+
              exit_idle();
              kvm_async_pf_task_wait((u32)read_cr2());
-             rcu_irq_exit();
+
+             if (is_idle_task(current) && rcu_is_cpu_idle())
+                     rcu_idle_enter();
+
              break;
      case KVM_PV_REASON_PAGE_READY:
              rcu_irq_enter();

Those rcu_irq_enter()/rcu_irq_exit() were introduced by commit
c5e015d4949aa665 "KVM guest: exit idleness when handling
KVM_PV_REASON_PAGE_NOT_PRESENT",

Ah!

    $ git name-rev  c5e015d4949aa665
    c5e015d4949aa665 remotes/tip/core/locking~3^2

Ok so that's not in -rcu. That's why I missed it.