On Mon, 2019-09-23 at 13:27 -0700, John Hubbard wrote:
> I'd also like a second opinion from the "core" -mm maintainers, but it seems like
> there is now too much code around the gup_pgd_range() call. Especially since there
> are two places where it's called--did you forget the other one in 
> __get_user_pages_fast(), btw??
> 
Oh, sorry, I missed this one. I will put it on v3.
(Also I will make sure to include linux-mm on v3.)

> Maybe the irq handling and atomic counting should be moved into start/finish
> calls, like this:
> 
>      start_gup_fast_walk()
>      gup_pgd_range()
>      finish_gup_fast_walk()

There are cases where interrupt disable/enable is not done around the
lockless pagetable walk.
It may come from functions called above on stack, that's why I opted it
to be only the atomic operation.