Re: [PATCH v3 6/8] mm/gup: use a standard migration target allocation callback

[Joonsoo Kim]

2020ë 7ì 4ì (í) ìì 12:56, Vlastimil Babka <vbabka@xxxxxxx>ëì ìì:
>
> On 6/23/20 8:13 AM, js1304@xxxxxxxxx wrote:
> > From: Joonsoo Kim <iamjoonsoo.kim@xxxxxxx>
> >
> > There is a well-defined migration target allocation callback.
> > It's mostly similar with new_non_cma_page() except considering CMA pages.
> >
> > This patch adds a CMA consideration to the standard migration target
> > allocation callback and use it on gup.c.
> >
> > Signed-off-by: Joonsoo Kim <iamjoonsoo.kim@xxxxxxx>
>
> Acked-by: Vlastimil Babka <vbabka@xxxxxxx>
>
> But a suggestion below.
>
> > ---
> >  mm/gup.c      | 57 ++++++++-------------------------------------------------
> >  mm/internal.h |  1 +
> >  mm/migrate.c  |  4 +++-
> >  3 files changed, 12 insertions(+), 50 deletions(-)
> >
> > diff --git a/mm/gup.c b/mm/gup.c
> > index 15be281..f6124e3 100644
> > --- a/mm/gup.c
> > +++ b/mm/gup.c
> > @@ -1608,56 +1608,15 @@ static bool check_dax_vmas(struct vm_area_struct **vmas, long nr_pages)
> >  }
> >
> >  #ifdef CONFIG_CMA
> > -static struct page *new_non_cma_page(struct page *page, unsigned long private)
> > +static struct page *alloc_migration_target_non_cma(struct page *page, unsigned long private)
> >  {
>
> ...
>
> > +     struct migration_target_control mtc = {
> > +             .nid = page_to_nid(page),
> > +             .gfp_mask = GFP_USER | __GFP_NOWARN,
> > +             .skip_cma = true,
> > +     };
> >
> > -     return __alloc_pages_node(nid, gfp_mask, 0);
> > +     return alloc_migration_target(page, (unsigned long)&mtc);
>
> Do we really need this wrapper? The only user is check_and_migrate_cma_pages so
> just opencode it?

This wrapper exists for setting up a different nid for each page.
However, as you suggested in the next reply, we can remove this wrapper if
NUMA_NO_NODE handling is added to the standard function. I will add NUMA_NO_NODE
handling to the standard function and remove this wrapper.

Thanks.