On Wed, 18 Jan 2012 22:05:12 -0800 (PST)
Hugh Dickins [off-list ref] wrote:

Commit c1e2ee2dc436 "memcg: replace ss->id_lock with a rwlock" has
now been seen to cause the unfair behavior we should have expected
from converting a spinlock to an rwlock: softlockup in cgroup_mkdir(),
whose get_new_cssid() is waiting for the wlock, while there are 19
tasks using the rlock in css_get_next() to get on with their memcg
workload (in an artificial test, admittedly).  Yet lib/idr.c was
made suitable for RCU way back.

1. Revert that commit, restoring ss->id_lock to a spinlock.

2. Make one small adjustment to idr_get_next(): take the height from
the top layer (stable under RCU) instead of from the root (unprotected
by RCU), as idr_find() does.

3. Remove lock and unlock around css_get_next()'s call to idr_get_next():
memcg iterators (only users of css_get_next) already did rcu_read_lock(),
and comment demands that, but add a WARN_ON_ONCE to make sure of it.

Signed-off-by: Hugh Dickins <redacted>

Thank you ! This seems much better.

Acked-by: KAMEZAWA Hiroyuki <redacted>