On Wed, Aug 21, 2019 at 04:01:49PM -0700, Jim Mattson wrote: > On Wed, Aug 21, 2019 at 3:22 PM Sean Christopherson > <sean.j.christopherson@xxxxxxxxx> wrote: > > > > On Wed, Aug 21, 2019 at 01:59:20PM -0700, Jim Mattson wrote: > > > On Mon, Aug 19, 2019 at 3:11 PM Sean Christopherson > > > <sean.j.christopherson@xxxxxxxxx> wrote: > > > > > > > > On Tue, Aug 20, 2019 at 12:46:49AM +0300, Nikita Leshenko wrote: > > > > > Before this commit, userspace could disable the GUEST_ACTIVITY_HLT bit in > > > > > VMX_MISC yet KVM would happily accept GUEST_ACTIVITY_HLT activity state in > > > > > VMCS12. We can fix it by either failing VM entries with HLT activity state when > > > > > it's not supported or by disallowing clearing this bit. > > > > > > > > > > The latter is preferable. If we go with the former, to disable > > > > > GUEST_ACTIVITY_HLT userspace also has to make CPU_BASED_HLT_EXITING a "must be > > > > > 1" control, otherwise KVM will be presenting a bogus model to L1. > > > > > > > > > > Don't fail writes that disable GUEST_ACTIVITY_HLT to maintain backwards > > > > > compatibility. > > > > > > > > Paolo, do we actually need to maintain backwards compatibility in this > > > > case? This seems like a good candidate for "fix the bug and see who yells". > > > > > > Google's userspace clears bit 6. Please don't fail that write! > > > > Booooo. > > > > Supporting activity state HLT is on our list of things to do, but I'm > not convinced that kvm actually handles it properly yet. For > instance... I fully understand why you'd want to hide it from L1, I was just bummed that we couldn't go with a quick and dirty fix :-) > What happens if L1 launches L2 into activity state HLT with a > zero-valued VMX preemption timer? (Maybe this is fixed now?) I think that one got fixed in vmx_start_preemption_timer(). > What happens if "monitor trap flag" is set and "HLT exiting" is clear > in the vmcs12, and immediately on VM-entry, L2 executes HLT? (Yes, > this is a special case of MTF being broken when L0 emulates an L2 > instruction.) > > I'm sure there are other interesting scenarios that haven't been validated.
