Re: [PATCH v2 2/7] mm: introduce local state for lazy_mmu sections

[David Hildenbrand]

> I also wondered about that. I think the safest is to make them
> respectively arch_leave() and arch_enter() - the flushing entailed by
> arch_leave() might not be required, but it is safer. Additionally,
> powerpc/sparc disable preemption while in lazy_mmu, so it seems like a
> good idea to re-enable it while paused (by calling arch_leave()).

Great, looking forward to seeing this all getting cleaned up and done 
properly for good.

--
Cheers

David / dhildenb