KVM: PPC: Make wakeups work again for Book3S HV guests
authorPaul Mackerras <paulus@samba.org>
Mon, 12 Dec 2011 12:24:48 +0000 (12:24 +0000)
committerAvi Kivity <avi@redhat.com>
Mon, 5 Mar 2012 12:52:34 +0000 (14:52 +0200)
When commit f43fdc15fa ("KVM: PPC: booke: Improve timer register
emulation") factored out some code in arch/powerpc/kvm/powerpc.c
into a new helper function, kvm_vcpu_kick(), an error crept in
which causes Book3s HV guest vcpus to stall.  This fixes it.
On POWER7 machines, guest vcpus are grouped together into virtual
CPU cores that share a single waitqueue, so it's important to use
vcpu->arch.wqp rather than &vcpu->wq.

Signed-off-by: Paul Mackerras <paulus@samba.org>
Signed-off-by: Alexander Graf <agraf@suse.de>
Signed-off-by: Avi Kivity <avi@redhat.com>
arch/powerpc/kvm/powerpc.c

index e1ef4d6d972ab5a7b109ae70cb6d3f7add1e4022..4f85ac32258ac24dbbf674796617564fe9999888 100644 (file)
@@ -562,7 +562,7 @@ void kvm_vcpu_kick(struct kvm_vcpu *vcpu)
        int cpu = vcpu->cpu;
 
        me = get_cpu();
-       if (waitqueue_active(&vcpu->wq)) {
+       if (waitqueue_active(vcpu->arch.wqp)) {
                wake_up_interruptible(vcpu->arch.wqp);
                vcpu->stat.halt_wakeup++;
        } else if (cpu != me && cpu != -1) {