Messages in this thread | | | Subject | Re: 5.0-rc1 KVM inspired "BUG: Bad page state in process" spew | From | Mike Galbraith <> | Date | Mon, 14 Jan 2019 18:42:16 +0100 |
| |
On Mon, 2019-01-14 at 07:46 -0800, Sean Christopherson wrote: > On Sat, Jan 12, 2019 at 07:43:02AM +0100, Mike Galbraith wrote: > > On Wed, 2019-01-09 at 11:26 -0800, Sean Christopherson wrote: > > > > > > I'll try to bisect. > > > > Good luck with that. I gave it a go, but that apparently invalidated > > the warrantee of my vm image :) > > This is fixed in v5.0-rc2 by commit ba422731316d ("mm/mmu_notifier: > mm/rmap.c: Fix a mmu_notifier range bug in try_to_unmap_one").
Good to hear.
> Nested VMs are great for bisecting KVM issues :)
I tried nesting dolls (to speed bisection.. hah!), but couldn't reliable trigger the problem with my reproducer. Even if it had, nested VMs are pretty cool, but not the _least bit_ fast :)
-Mike
| |