On 28/02/20 02:37, Sean Christopherson wrote: >>> Would it be better or worse if we eliminate set_supported_cpuid() hook >>> completely by doing an ugly hack like (completely untested): >> Yes, it makes sense. > Works for me, I'll tack it on. I think my past self kept it because I was > planning on using vmx_set_supported_cpuid() for SGX, which adds multiple > sub-leafs, but I'm pretty sure I can squeeze them into kvm_cpu_caps with > a few extra shenanigans. > We can add it back for full CPUID leaves; it may even make sense to move PT processing there (but not in this series). Paolo
