Messages in this thread | | | Date | Fri, 5 Apr 2024 01:02:06 +0100 | Subject | Re: [PATCH v2] x86/mm: Don't disable INVLPG if "incomplete Global INVLPG flushes" is fixed by microcode | From | Andrew Cooper <> |
| |
On 04/04/2024 6:28 pm, Sean Christopherson wrote: > On Thu, Apr 04, 2024, Andrew Cooper wrote: >> A related example. I wrote the patch to hide XSAVES to work around an >> AMD erratum where XSAVEC sufficed, and the consequences were so dire for >> some versions of Windows that there was a suggestion to simply revert >> the workaround to make VMs run again. Windows intentionally asserts >> sanity (== expectations) in what it can see; I have no idea whether it >> would object in this case but hiding PCID is definitely playing with fire. > Yeah, KVM users got burned by that too. d52734d00b8e ("KVM: x86: Give a hint when > Win2016 might fail to boot due to XSAVES erratum").
Yeah what I meant was that I wrote the Linux patch, and KVM got burnt while Xen cared not... :)
> Hmm, one crazy idea would be to carve out a hypervisor CPUID range for enumerating > (potentially) broken features. Dealing with the Intel/AMD (and Centaur, LOL), > 0 / 0x8000_0000 split would be annoying, but not hard. E.g. use 0x4{0,8,C}01_xxxx
No transmeta love then? Or perhaps we declare it their fault for choosing 0x8086 which is too awkward to fit into that scheme.
> to enumerate broken features, and then the guest could do: > > support = CPUID(leaf).reg & ~CPUID(to_pv_broken(leaf)).reg; > > It'd require a decent amount of churn for the initial support, but it would give > hypervisors a way to inform guests that _any_ CPUID-based feature is broken, > without requiring guest changes (after the initial code is merged) or explicit > action from hardware vendors. > > And if we got Windows/Hyper-V in on the game, it would allow them to keep their > sanity checks while (hopefully) degrading gracefully if a feature is enumerated > as broken.
Crazy indeed, but I am curious to see if this has legs. The exact indices may need tweaking, because 0x4x01_xxxx might be a little too close for comfort, but at first glance it does look like a surprisingly neat solution to the problem.
Perhaps worth a slot at plumbers?
~Andrew
| |