On Fri, 20 Jan 2012 00:40:34 -0800
Greg Thelen <gthelen@xxxxxxxxxx> wrote:
> On Fri, Jan 13, 2012 at 12:45 AM, KAMEZAWA Hiroyuki
> > diff --git a/mm/memcontrol.c b/mm/memcontrol.c
> > index 8b67ccf..4836e8d 100644
> > --- a/mm/memcontrol.c
> > +++ b/mm/memcontrol.c
> > @@ -89,7 +89,6 @@ enum mem_cgroup_stat_index {
> > MEM_CGROUP_STAT_FILE_MAPPED, /* # of pages charged as file rss */
> > MEM_CGROUP_STAT_SWAPOUT, /* # of pages, swapped out */
> > MEM_CGROUP_STAT_DATA, /* end of data requires synchronization */
> > - MEM_CGROUP_ON_MOVE, /* someone is moving account between groups */
> > MEM_CGROUP_STAT_NSTATS,
> > };
> >
> > @@ -279,6 +278,8 @@ struct mem_cgroup {
> > * mem_cgroup ? And what type of charges should we move ?
> > */
> > unsigned long move_charge_at_immigrate;
> > + /* set when a page under this memcg may be moving to other memcg */
> > + atomic_t account_moving;
> > /*
> > * percpu counter.
> > */
> > @@ -1250,20 +1251,27 @@ int mem_cgroup_swappiness(struct mem_cgroup *memcg)
> > return memcg->swappiness;
> > }
> >
> > +/*
> > + * For quick check, for avoiding looking up memcg, system-wide
> > + * per-cpu check is provided.
> > + */
> > +DEFINE_PER_CPU(int, mem_cgroup_account_moving);
>
> Why is this a per-cpu counter? Can this be an single atomic_t
> instead, or does cpu hotplug require per-cpu state? In the common
> case, when there is no move in progress, then the counter would be
> zero and clean in all cpu caches that need it. When moving pages,
> mem_cgroup_start_move() would atomic_inc the counter.
>
Ok, atomic_t will be simple.
Thanks,
-Kame
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@xxxxxxxxx. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Fight unfair telecom internet charges in Canada: sign http://stopthemeter.ca/
Don't email: <a href=mailto:"dont@xxxxxxxxx";> email@xxxxxxxxx </a>