Re: [patch 1/5] memcg: no uncharged pages reach page_cgroup_zoneinfo

[Johannes Weiner]

On Fri, Feb 04, 2011 at 09:01:45AM +0900, KAMEZAWA Hiroyuki wrote:
> On Thu,  3 Feb 2011 15:26:02 +0100
> Johannes Weiner <hannes@xxxxxxxxxxx> wrote:
> 
> > All callsites check PCG_USED before passing pc->mem_cgroup, so the
> > latter is never NULL.
> > 
> > Signed-off-by: Johannes Weiner <hannes@xxxxxxxxxxx>
> 
> Acked-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@xxxxxxxxxxxxxx>

Thank you!

> I want BUG_ON() here.

I thought about it too at first.  But look at the callsites, all but
one of them do not even expect this function to return NULL, so if
this condition had ever been true, we would have seen crashes in the
callsites.

The only caller that checks for NULL is
mem_cgroup_get_reclaim_stat_from_page() and I propose to remove that
as well; patch attached.

Do you insist on the BUG_ON?

---
Subject: memcg: page_cgroup_zoneinfo never returns NULL

For a page charged to a memcg, there is always valid memcg per-zone
info.

Signed-off-by: Johannes Weiner <hannes@xxxxxxxxxxx>
---
diff --git a/mm/memcontrol.c b/mm/memcontrol.c
index 4a4483d..5f974b3 100644
--- a/mm/memcontrol.c
+++ b/mm/memcontrol.c
@@ -1017,9 +1017,6 @@ mem_cgroup_get_reclaim_stat_from_page(struct page *page)
 	/* Ensure pc->mem_cgroup is visible after reading PCG_USED. */
 	smp_rmb();
 	mz = page_cgroup_zoneinfo(pc->mem_cgroup, page);
-	if (!mz)
-		return NULL;
-
 	return &mz->reclaim_stat;
 }
 
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/