On Fri 13-04-18 14:14:33, Michal Hocko wrote:
[...]
> Well, this is probably a matter of taste. I will not argue. I will not
> object if Johannes is OK with your patch. But the whole thing confused
> hell out of me so I would rather un-clutter it...
In other words, this
diff --git a/mm/memcontrol.c b/mm/memcontrol.c
index 8c2ed1c2b72c..ca7e981a8a1a 100644
--- a/mm/memcontrol.c
+++ b/mm/memcontrol.c
@@ -4351,6 +4351,14 @@ static void __mem_cgroup_free(struct mem_cgroup *memcg)
{
int node;
+ /*
+ * We are trying to remove the idr key when the last memcg
+ * reference drops which can be sooner than when the last
+ * css reference is dropped to recycle ids faster.
+ */
+ if (memcg->id.id > 0)
+ idr_remove(&mem_cgroup_idr, memcg->id.id);
+
for_each_node(node)
free_mem_cgroup_per_node_info(memcg, node);
free_percpu(memcg->stat_cpu);
@@ -4411,8 +4419,6 @@ static struct mem_cgroup *mem_cgroup_alloc(void)
idr_replace(&mem_cgroup_idr, memcg, memcg->id.id);
return memcg;
fail:
- if (memcg->id.id > 0)
- idr_remove(&mem_cgroup_idr, memcg->id.id);
__mem_cgroup_free(memcg);
return NULL;
}
--
Michal Hocko
SUSE Labs
