cgroup: nest kernfs active protection under cgroup_mutex
authorTejun Heo <tj@kernel.org>
Tue, 13 May 2014 16:19:23 +0000 (12:19 -0400)
committerTejun Heo <tj@kernel.org>
Tue, 13 May 2014 16:19:23 +0000 (12:19 -0400)
commit01f6474ce04fffd6282b569ac0a31f4b98d4c82a
treeb536ea8d57db3c87fbe897db727005db5cf2e8db
parente76ecaeef65c497153ceacf59c2e21c070d43f64
cgroup: nest kernfs active protection under cgroup_mutex

After the recent cgroup_kn_lock_live() changes, cgroup_mutex is no
longer nested below kernfs active protection.  The two don't have any
relationship now.

This patch nests kernfs active protection under cgroup_mutex.  All
cftype operations now require both cgroup_tree_mutex and cgroup_mutex,
temporary cgroup_mutex releases over kernfs operations are removed,
and cgroup_add/rm_cftypes() grab both mutexes.

This makes cgroup_tree_mutex redundant, which will be removed by the
next patch.

Signed-off-by: Tejun Heo <tj@kernel.org>
Acked-by: Li Zefan <lizefan@huawei.com>
kernel/cgroup.c