cgroup: don't online subsystems before cgroup_name/path() are operational
authorTejun Heo <tj@kernel.org>
Thu, 26 Jan 2017 21:47:28 +0000 (16:47 -0500)
committerTejun Heo <tj@kernel.org>
Thu, 26 Jan 2017 21:47:28 +0000 (16:47 -0500)
commit07cd12945551b63ecb1a349d50a6d69d1d6feb4a
tree75f65eba7eac9277971082a2d5a4cf1370562c0c
parent7ce7d89f48834cefece7804d38fc5d85382edf77
cgroup: don't online subsystems before cgroup_name/path() are operational

While refactoring cgroup creation, a5bca2152036 ("cgroup: factor out
cgroup_create() out of cgroup_mkdir()") incorrectly onlined subsystems
before the new cgroup is associated with it kernfs_node.  This is fine
for cgroup proper but cgroup_name/path() depend on the associated
kernfs_node and if a subsystem makes the new cgroup_subsys_state
visible, which they're allowed to after onlining, it can lead to NULL
dereference.

The current code performs cgroup creation and subsystem onlining in
cgroup_create() and cgroup_mkdir() makes the cgroup and subsystems
visible afterwards.  There's no reason to online the subsystems early
and we can simply drop cgroup_apply_control_enable() call from
cgroup_create() so that the subsystems are onlined and made visible at
the same time.

Signed-off-by: Tejun Heo <tj@kernel.org>
Reported-by: Konstantin Khlebnikov <khlebnikov@yandex-team.ru>
Fixes: a5bca2152036 ("cgroup: factor out cgroup_create() out of cgroup_mkdir()")
Cc: stable@vger.kernel.org # v4.6+
kernel/cgroup.c