KVM: VMX: disable SMEP feature when guest is in non-paging mode
authorDongxiao Xu <dongxiao.xu@intel.com>
Mon, 4 Feb 2013 03:50:43 +0000 (11:50 +0800)
committerMarcelo Tosatti <mtosatti@redhat.com>
Wed, 6 Feb 2013 01:28:07 +0000 (23:28 -0200)
SMEP is disabled if CPU is in non-paging mode in hardware.
However KVM always uses paging mode to emulate guest non-paging
mode with TDP. To emulate this behavior, SMEP needs to be manually
disabled when guest switches to non-paging mode.

We met an issue that, SMP Linux guest with recent kernel (enable
SMEP support, for example, 3.5.3) would crash with triple fault if
setting unrestricted_guest=0. This is because KVM uses an identity
mapping page table to emulate the non-paging mode, where the page
table is set with USER flag. If SMEP is still enabled in this case,
guest will meet unhandlable page fault and then crash.

Reviewed-by: Gleb Natapov <gleb@redhat.com>
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Dongxiao Xu <dongxiao.xu@intel.com>
Signed-off-by: Xiantao Zhang <xiantao.zhang@intel.com>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
arch/x86/kvm/vmx.c

index 0cf74a641dec3296229f3cc233b083a9545c55f6..fe9a9cfadbd69162f708a4938fcb83f6dc73240f 100644 (file)
@@ -3227,6 +3227,14 @@ static int vmx_set_cr4(struct kvm_vcpu *vcpu, unsigned long cr4)
                if (!is_paging(vcpu)) {
                        hw_cr4 &= ~X86_CR4_PAE;
                        hw_cr4 |= X86_CR4_PSE;
+                       /*
+                        * SMEP is disabled if CPU is in non-paging mode in
+                        * hardware. However KVM always uses paging mode to
+                        * emulate guest non-paging mode with TDP.
+                        * To emulate this behavior, SMEP needs to be manually
+                        * disabled when guest switches to non-paging mode.
+                        */
+                       hw_cr4 &= ~X86_CR4_SMEP;
                } else if (!(cr4 & X86_CR4_PAE)) {
                        hw_cr4 &= ~X86_CR4_PAE;
                }