I just reviewed the patch "KVM: vmx: Introduce handle_unexpected_vmexit and handle WAITPKG vmexit” currently queued in kvm git tree (https://git.kernel.org/pub/scm/virt/kvm/kvm.git/commit/?h=queue&id=bf653b78f9608d66db174eabcbda7454c8fde6d5) It seems to me that we shouldn’t apply this patch in it’s current form. Instead of having a common handle_unexpected_vmexit() manually specified for specific VMExit reasons, we should rely on the functionality I have added to vmx_handle_exit() in case there is no valid handler for exit-reason. In this case (since commit 7396d337cfadc ("KVM: x86: Return to userspace with internal error on unexpected exit reason”), an internal-error will be raised to userspace as required. Instead of silently skipping emulated instruction. -Liran
