On Fri, Sep 25, 2020 at 11:59:00AM +0200, Vitaly Kuznetsov wrote: > Sean Christopherson <sean.j.christopher...@intel.com> writes: > > +#if IS_ENABLED(CONFIG_HYPERV) > > enum ept_pointers_status ept_pointers_match; > > spinlock_t ept_pointer_lock; > > +#endif > > Reviewed-by: Vitaly Kuznetsov <vkuzn...@redhat.com> > > In case ept_pointers_match/ept_pointer_lock are useless for TDX we may > want to find better names for them to make it clear this is a Hyper-V > thingy (e.g. something like hv_tlb_ept_match/hv_tlb_ept_lock).
Good call. I'll send a v2, looking at hv_remote_flush_tlb_with_range(), I think there are additional cleanups/optimizations that can be done, e.g. do the extra flushes only on vCPUs with a mistmatching EPTP instead of flushing all vCPUs' EPTPs if _any_ vCPU has a mismatched EPTP. > > > }; > > > > bool nested_vmx_allowed(struct kvm_vcpu *vcpu); > > -- > Vitaly >
