On Wed 08-07-20 16:27:16, Aneesh Kumar K.V wrote:
> Vlastimil Babka <vbabka@xxxxxxx> writes:
>
> > On 7/8/20 9:41 AM, Michal Hocko wrote:
> >> On Wed 08-07-20 16:16:02, Joonsoo Kim wrote:
> >>> On Tue, Jul 07, 2020 at 01:22:31PM +0200, Vlastimil Babka wrote:
> >>>
> >>> Simply, I call memalloc_nocma_{save,restore} in new_non_cma_page(). It
> >>> would not cause any problem.
> >>
> >> I believe a proper fix is the following. The scope is really defined for
> >> FOLL_LONGTERM pins and pushing it inside check_and_migrate_cma_pages
> >> will solve the problem as well but it imho makes more sense to do it in
> >> the caller the same way we do for any others.
> >>
> >> Fixes: 9a4e9f3b2d73 ("mm: update get_user_pages_longterm to migrate pages allocated from CMA region")
> >
> > Agreed.
> >
> >>
> >> I am not sure this is worth backporting to stable yet.
> >
> > CC Aneesh.
> >
> > Context: since check_and_migrate_cma_pages() calls __get_user_pages_locked(), it
> > should also be called under memalloc_nocma_save().
>
> But by then we faulted in all relevant pages and migrated them out of
> CMA rea right?
check_and_migrate_cma_pages will allocate target pages that you want to
migrate off the CMA region unless I am misreading the code. And those
allocation need to be placed outside of the CMA.
--
Michal Hocko
SUSE Labs
