Seems to be required to "re-arm" the engines after a vm fault. Signed-off-by: Ben Skeggs <bskeggs@redhat.com>