Re: [PATCH 1/6] KVM: nVMX: avoid NULL pointer dereference with incorrect EVMCS GPAs

Re: [PATCH 1/6] KVM: nVMX: avoid NULL pointer dereference with incorrect EVMCS GPAs

[linmiaohe]

Hi:
Vitaly Kuznetsov <vkuznets@redhat.com> writes:
>When an EVMCS enabled L1 guest on KVM will tries doing enlightened VMEnter with EVMCS GPA = 0 the host crashes because the
>
>evmcs_gpa != vmx->nested.hv_evmcs_vmptr
>
>condition in nested_vmx_handle_enlightened_vmptrld() will evaluate to false (as nested.hv_evmcs_vmptr is zeroed after init). The crash will happen on vmx->nested.hv_evmcs pointer dereference.
>
>Another problematic EVMCS ptr value is '-1' but it only causes host crash after nested_release_evmcs() invocation. The problem is exactly the same as with '0', we mistakenly think that the EVMCS pointer hasn't changed and thus nested.hv_evmcs_vmptr is valid.
>
>Resolve the issue by adding an additional !vmx->nested.hv_evmcs check to nested_vmx_handle_enlightened_vmptrld(), this way we will always be trying kvm_vcpu_map() when nested.hv_evmcs is NULL and this is supposed to catch all invalid EVMCS GPAs.
>
>Also, initialize hv_evmcs_vmptr to '0' in nested_release_evmcs() to be consistent with initialization where we don't currently set hv_evmcs_vmptr to '-1'.
>
>Cc: stable@vger.kernel.org
>Signed-off-by: Vitaly Kuznetsov <vkuznets@redhat.com>
>---
> arch/x86/kvm/vmx/nested.c | 5 +++--
> 1 file changed, 3 insertions(+), 2 deletions(-)

Good catch! Patch looks good for me. Thanks!
Reviewed-by: Miaohe Lin <linmiaohe@huawei.com>

[PATCH 0/6] KVM: nVMX: propperly handle enlightened vmptrld failure conditions

[Vitaly Kuznetsov]

Miaohe Lin noticed that we incorrectly handle enlightened vmptrld failures
in nested_vmx_run(). Trying to handle errors correctly, I fixed
a few things:
- NULL pointer dereference with invalid eVMCS GPAs [PATCH1]
- moved eVMCS mapping after migration to nested_get_vmcs12_pages() from
  nested_sync_vmcs12_to_shadow() [PATCH2]
- added propper nested_vmx_handle_enlightened_vmptrld() error handling
  [PATCH3]
- added selftests for incorrect eVMCS revision id and GPA [PATCHes4-6]

PATCH1 fixes a DoS and thus marked for stable@.

Vitaly Kuznetsov (6):
  KVM: nVMX: avoid NULL pointer dereference with incorrect EVMCS GPAs
  KVM: nVMX: stop abusing need_vmcs12_to_shadow_sync for eVMCS mapping
  KVM: nVMX: properly handle errors in
    nested_vmx_handle_enlightened_vmptrld()
  KVM: selftests: define and use EVMCS_VERSION
  KVM: selftests: test enlightened vmenter with wrong eVMCS version
  KVM: selftests: enlightened VMPTRLD with an incorrect GPA

 arch/x86/kvm/vmx/evmcs.h                      |  7 ++
 arch/x86/kvm/vmx/nested.c                     | 64 +++++++++++++------
 tools/testing/selftests/kvm/include/evmcs.h   |  2 +
 tools/testing/selftests/kvm/lib/x86_64/vmx.c  |  2 +-
 .../testing/selftests/kvm/x86_64/evmcs_test.c | 25 ++++++--
 5 files changed, 72 insertions(+), 28 deletions(-)

-- 
2.24.1

[PATCH 1/6] KVM: nVMX: avoid NULL pointer dereference with incorrect EVMCS GPAs

[Vitaly Kuznetsov]

When an EVMCS enabled L1 guest on KVM will tries doing enlightened VMEnter
with EVMCS GPA = 0 the host crashes because the

evmcs_gpa != vmx->nested.hv_evmcs_vmptr

condition in nested_vmx_handle_enlightened_vmptrld() will evaluate to
false (as nested.hv_evmcs_vmptr is zeroed after init). The crash will
happen on vmx->nested.hv_evmcs pointer dereference.

Another problematic EVMCS ptr value is '-1' but it only causes host crash
after nested_release_evmcs() invocation. The problem is exactly the same as
with '0', we mistakenly think that the EVMCS pointer hasn't changed and
thus nested.hv_evmcs_vmptr is valid.

Resolve the issue by adding an additional !vmx->nested.hv_evmcs
check to nested_vmx_handle_enlightened_vmptrld(), this way we will
always be trying kvm_vcpu_map() when nested.hv_evmcs is NULL
and this is supposed to catch all invalid EVMCS GPAs.

Also, initialize hv_evmcs_vmptr to '0' in nested_release_evmcs()
to be consistent with initialization where we don't currently
set hv_evmcs_vmptr to '-1'.

Cc: stable@vger.kernel.org
Signed-off-by: Vitaly Kuznetsov <vkuznets@redhat.com>
---
 arch/x86/kvm/vmx/nested.c | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)

diff --git a/arch/x86/kvm/vmx/nested.c b/arch/x86/kvm/vmx/nested.c
index e920d7834d73..9750e590c89d 100644
--- a/arch/x86/kvm/vmx/nested.c
+++ b/arch/x86/kvm/vmx/nested.c
@@ -224,7 +224,7 @@ static inline void nested_release_evmcs(struct kvm_vcpu *vcpu)
 		return;
 
 	kvm_vcpu_unmap(vcpu, &vmx->nested.hv_evmcs_map, true);
-	vmx->nested.hv_evmcs_vmptr = -1ull;
+	vmx->nested.hv_evmcs_vmptr = 0;
 	vmx->nested.hv_evmcs = NULL;
 }
 
@@ -1923,7 +1923,8 @@ static int nested_vmx_handle_enlightened_vmptrld(struct kvm_vcpu *vcpu,
 	if (!nested_enlightened_vmentry(vcpu, &evmcs_gpa))
 		return 1;
 
-	if (unlikely(evmcs_gpa != vmx->nested.hv_evmcs_vmptr)) {
+	if (unlikely(!vmx->nested.hv_evmcs ||
+		     evmcs_gpa != vmx->nested.hv_evmcs_vmptr)) {
 		if (!vmx->nested.hv_evmcs)
 			vmx->nested.current_vmptr = -1ull;
 
-- 
2.24.1