Re: [PATCH 1/3] mm: memcg: remote memcg charging for kmem allocations

[Johannes Weiner]

On Tue, Jun 19, 2018 at 04:31:18PM -0700, Shakeel Butt wrote:
> On Tue, Jun 19, 2018 at 9:22 AM Johannes Weiner <hannes@xxxxxxxxxxx> wrote:
> >
> > On Mon, Jun 18, 2018 at 10:13:25PM -0700, Shakeel Butt wrote:
> > > @@ -248,6 +248,30 @@ static inline void memalloc_noreclaim_restore(unsigned int flags)
> > >       current->flags = (current->flags & ~PF_MEMALLOC) | flags;
> > >  }
> > >
> > > +#ifdef CONFIG_MEMCG
> > > +static inline struct mem_cgroup *memalloc_memcg_save(struct mem_cgroup *memcg)
> > > +{
> > > +     struct mem_cgroup *old_memcg = current->target_memcg;
> > > +
> > > +     current->target_memcg = memcg;
> > > +     return old_memcg;
> > > +}
> > > +
> > > +static inline void memalloc_memcg_restore(struct mem_cgroup *memcg)
> > > +{
> > > +     current->target_memcg = memcg;
> > > +}
> >
> > The use_mm() and friends naming scheme would be better here:
> > memalloc_use_memcg(), memalloc_unuse_memcg(), current->active_memcg
> >
> 
> Ack. Though do you still think <linux/sched/mm.h> is the right place
> for these functions?

Yeah, since it has the memalloc_* prefix, we should keep it there.

If we did use_memcg(), unuse_memcg(), I'd put it into memcontrol.h,
but it seems a little terse; memalloc adds valuable context, IMO.

Thanks Shakeel!