Re: [PATCH v3 4/6] mm: honor PF_MEMALLOC_PIN for all movable pages

[Pavel Tatashin <pasha.tatashin@soleen.com>]

On Mon, Dec 14, 2020 at 9:17 AM Michal Hocko <mhocko@suse.com> wrote:
>
> On Fri 11-12-20 15:21:38, Pavel Tatashin wrote:
> [...]
> > diff --git a/mm/page_alloc.c b/mm/page_alloc.c
> > index c2dea9ad0e98..4d8e7f801c66 100644
> > --- a/mm/page_alloc.c
> > +++ b/mm/page_alloc.c
> > @@ -3802,16 +3802,12 @@ alloc_flags_nofragment(struct zone *zone, gfp_t gfp_mask)
> >       return alloc_flags;
> >  }
> >
> > -static inline unsigned int current_alloc_flags(gfp_t gfp_mask,
> > -                                     unsigned int alloc_flags)
> > +static inline unsigned int cma_alloc_flags(gfp_t gfp_mask,
> > +                                        unsigned int alloc_flags)
>
> Do you have any strong reason to rename? Even though the current

Yes :)

> implementation only does something for cma I do not think this is all
> that important. The naming nicely fits with current_gfp_context so I
> would stick with it.

I am renaming because current->flags is removed from this function,
therefore keeping the name
becomes misleading. This function only addresses cma flag check
without looking at the thread local state now.

>
> Other than that the patch looks reasonable. I would just add a comment
> explaining that current_alloc_flags should be called _after_
> current_gfp_context because that one might change the gfp_mask.

Thanks, I will add it.

>
> With that addressed, feel free to add
> Acked-by: Michal Hocko <mhocko@suse.com>

Thank you,
Pasha