Messages in this thread | | | Subject | Re: [PATCH 0/6] KVM: nSVM: Fix issues when SMM is entered from L2 | From | Paolo Bonzini <> | Date | Thu, 8 Jul 2021 19:40:11 +0200 |
| |
On 28/06/21 12:44, Vitaly Kuznetsov wrote: > This is a continuation of "[PATCH RFC] KVM: nSVM: Fix L1 state corruption > upon return from SMM". > > VMCB split commit 4995a3685f1b ("KVM: SVM: Use a separate vmcb for the > nested L2 guest") broke return from SMM when we entered there from guest > (L2) mode. Gen2 WS2016/Hyper-V is known to do this on boot. The problem > appears to be that VMCB01 gets irreversibly destroyed during SMM execution. > Previously, we used to have 'hsave' VMCB where regular (pre-SMM) L1's state > was saved upon nested_svm_vmexit() but now we just switch to VMCB01 from > VMCB02. > > While writing a selftest for the issue, I've noticed that 'svm->nested.ctl' > doesn't get restored after KVM_GET_NESTED_STATE/KVM_SET_NESTED_STATE cycle > when guest happens to be in SMM triggered from L2. "KVM: nSVM: Restore > nested control upon leaving SMM" is aimed to fix that. > > First two patches of the series add missing sanity checks for > MSR_VM_HSAVE_PA which has to be page aligned and not zero. > > Vitaly Kuznetsov (6): > KVM: nSVM: Check the value written to MSR_VM_HSAVE_PA > KVM: nSVM: Check that VM_HSAVE_PA MSR was set before VMRUN > KVM: nSVM: Introduce svm_copy_nonvmloadsave_state() > KVM: nSVM: Fix L1 state corruption upon return from SMM > KVM: nSVM: Restore nested control upon leaving SMM > KVM: selftests: smm_test: Test SMM enter from L2 > > arch/x86/kvm/svm/nested.c | 45 +++++++----- > arch/x86/kvm/svm/svm.c | 51 +++++++++++++- > arch/x86/kvm/svm/svm.h | 4 ++ > tools/testing/selftests/kvm/x86_64/smm_test.c | 70 +++++++++++++++++-- > 4 files changed, 144 insertions(+), 26 deletions(-) >
Queued, thanks.
Paolo
| |