KVM: MMU: Fix incorrect direct gfn for unpaged mode shadow

We use the physical address instead of the base gfn for the four
PAE page directories we use in unpaged mode.  When the guest accesses
an address above 1GB that is backed by a large host page, a BUG_ON()
in kvm_mmu_set_gfn() triggers.

Resolves: https://bugzilla.kernel.org/show_bug.cgi?id=21962Reported-and-tested-by: Nicolas Prochazka <prochazka.nicolas@gmail.com>
KVM-Stable-Tag.
Signed-off-by: Avi Kivity <avi@redhat.com>

arch/x86/kvm/mmu.c

@@ -2394,7 +2394,8 @@ static int mmu_alloc_direct_roots(struct kvm_vcpu *vcpu)
 			ASSERT(!VALID_PAGE(root));
 			spin_lock(&vcpu->kvm->mmu_lock);
 			kvm_mmu_free_some_pages(vcpu);
-			sp = kvm_mmu_get_page(vcpu, i << 30, i << 30,
+			sp = kvm_mmu_get_page(vcpu, i << (30 - PAGE_SHIFT),
+					      i << 30,
 					      PT32_ROOT_LEVEL, 1, ACC_ALL,
 					      NULL);
 			root = __pa(sp->spt);