KVM: PPC: Book3S: Provide different CAPs based on HV or PR mode

With Book3S KVM we can create both PR and HV VMs in parallel on the same
machine. That gives us new challenges on the CAPs we return - both have
different capabilities.

When we get asked about CAPs on the kvm fd, there's nothing we can do. We
can try to be smart and assume we're running HV if HV is available, PR
otherwise. However with the newly added VM CHECK_EXTENSION we can now ask
for capabilities directly on a VM which knows whether it's PR or HV.

With this patch I can successfully expose KVM PVINFO data to user space
in the PR case, fixing magic page mapping for PAPR guests.

Signed-off-by: Alexander Graf <agraf@suse.de>
Acked-by: Paolo Bonzini <pbonzini@redhat.com>

1 files changed, 9 insertions, 3 deletions

diff --git a/arch/powerpc/kvm/powerpc.c b/arch/powerpc/kvm/powerpc.c
index d870bacc2f75..eaa57da7711b 100644
--- a/arch/powerpc/kvm/powerpc.c
+++ b/arch/powerpc/kvm/powerpc.c
@@ -394,11 +394,17 @@ void kvm_arch_sync_events(struct kvm *kvm)
 int kvm_vm_ioctl_check_extension(struct kvm *kvm, long ext)
 {
         int r;
-        /* FIXME!!
-         * Should some of this be vm ioctl ? is it possible now ?
-         */
+        /* Assume we're using HV mode when the HV module is loaded */
         int hv_enabled = kvmppc_hv_ops ? 1 : 0;
 
+        if (kvm) {
+                /*
+                 * Hooray - we know which VM type we're running on. Depend on
+                 * that rather than the guess above.
+                 */
+                hv_enabled = is_kvmppc_hv_enabled(kvm);
+        }
+
         switch (ext) {
 #ifdef CONFIG_BOOKE
         case KVM_CAP_PPC_BOOKE_SREGS: