KVM: MMU: handle invalid root_hpa at __direct_map
authorMarcelo Tosatti <mtosatti@redhat.com>
Thu, 19 Dec 2013 17:28:51 +0000 (15:28 -0200)
committerPaolo Bonzini <pbonzini@redhat.com>
Fri, 20 Dec 2013 18:22:49 +0000 (19:22 +0100)
It is possible for __direct_map to be called on invalid root_hpa
(-1), two examples:

1) try_async_pf -> can_do_async_pf
    -> vmx_interrupt_allowed -> nested_vmx_vmexit
2) vmx_handle_exit -> vmx_interrupt_allowed -> nested_vmx_vmexit

Then to load_vmcs12_host_state and kvm_mmu_reset_context.

Check for this possibility, let fault exception be regenerated.

BZ: https://bugzilla.redhat.com/show_bug.cgi?id=924916

Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
arch/x86/kvm/mmu.c

index 40772ef0f2b12458f21094d9a642a487a081fdde..31a570287fcc6d9b741ce7ea8c42beb01a25dc84 100644 (file)
@@ -2659,6 +2659,9 @@ static int __direct_map(struct kvm_vcpu *vcpu, gpa_t v, int write,
        int emulate = 0;
        gfn_t pseudo_gfn;
 
+       if (!VALID_PAGE(vcpu->arch.mmu.root_hpa))
+               return 0;
+
        for_each_shadow_entry(vcpu, (u64)gfn << PAGE_SHIFT, iterator) {
                if (iterator.level == level) {
                        mmu_set_spte(vcpu, iterator.sptep, ACC_ALL,