KVM/PPC: update the comment of memory barrier in the kvmppc_prepare_to_enter()
authorLan Tianyu <tianyu.lan@intel.com>
Sun, 13 Mar 2016 03:10:30 +0000 (11:10 +0800)
committerPaolo Bonzini <pbonzini@redhat.com>
Tue, 22 Mar 2016 15:38:37 +0000 (16:38 +0100)
The barrier also orders the write to mode from any reads
to the page tables done and so update the comment.

Signed-off-by: Lan Tianyu <tianyu.lan@intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
arch/powerpc/kvm/powerpc.c

index 19aa59b0850cf73489d5f61623794da607a6fc41..6a68730774ee7dd701358cbc6cfce36c37db4009 100644 (file)
@@ -96,6 +96,9 @@ int kvmppc_prepare_to_enter(struct kvm_vcpu *vcpu)
                 * so we don't miss a request because the requester sees
                 * OUTSIDE_GUEST_MODE and assumes we'll be checking requests
                 * before next entering the guest (and thus doesn't IPI).
+                * This also orders the write to mode from any reads
+                * to the page tables done while the VCPU is running.
+                * Please see the comment in kvm_flush_remote_tlbs.
                 */
                smp_mb();