On Mon, Sep 18, 2023, Jann Horn wrote: > Hi! > > I haven't tested this and might be missing something, but I think that > the MMU notifier for KVM_GUEST_USES_PFN pfncache is currently a bit > broken. Except that nothing seems to actually use KVM_GUEST_USES_PFN, > so currently it's not actually a problem? Yeah, the implementation is busted, and IMO the entire concept is a dead-end[1]. David Steven's series[2] doesn't actually rip out KVM_GUEST_USES_PFN, I'll add that to the todo list. There are no users, and I don't expect any to come along, precisely because it's busted :-) [1] https://lore.kernel.org/all/ZBEEQtmtNPaEqU1i@xxxxxxxxxx [2] https://lkml.kernel.org/r/20230911021637.1941096-1-stevensd%40google.com
