ptp_kvm: probe for kvm guest availability

In the event of moving pvclock_pvti_cpu0_va() definition to common
pvclock code, this function would return a value on non KVM guests.
Later on this would fail with a GPF on ptp_kvm_init when running on a
Xen guest. Therefore, ptp_kvm_init() should check whether it is running
in a KVM guest.

Signed-off-by: Joao Martins <joao.m.martins@oracle.com>
Acked-by: Radim Krčmář <rkrcmar@redhat.com>
Signed-off-by: Boris Ostrovsky <boris.ostrovsky@oracle.com>

diff --git a/drivers/ptp/ptp_kvm.c b/drivers/ptp/ptp_kvm.c
index 2b1b212c219e74b85a78f94dc1d348ffb70de0bc..e04d7b2ecb3afb71d7f8a9b1f3e3718e6e26afac 100644 (file)
--- a/drivers/ptp/ptp_kvm.c
+++ b/drivers/ptp/ptp_kvm.c
@@ -178,6 +178,9 @@ static int __init ptp_kvm_init(void)
 {
        long ret;
 
+       if (!kvm_para_available())
+               return -ENODEV;
+
        clock_pair_gpa = slow_virt_to_phys(&clock_pair);
        hv_clock = pvclock_pvti_cpu0_va();