| From zhangqiao22@huawei.com Thu Mar 17 11:21:46 2022 |
| From: Zhang Qiao <zhangqiao22@huawei.com> |
| Date: Thu, 17 Mar 2022 10:41:57 +0800 |
| Subject: cpuset: Fix unsafe lock order between cpuset lock and cpuslock |
| To: "Greg Kroah-Hartman" <gregkh@linuxfoundation.org>, "Michal Koutný" <mkoutny@suse.com> |
| Cc: <linux-kernel@vger.kernel.org>, <stable@vger.kernel.org>, Zhao Gongyi <zhaogongyi@huawei.com>, Waiman Long <longman@redhat.com>, Tejun Heo <tj@kernel.org>, Juri Lelli <juri.lelli@redhat.com> |
| Message-ID: <1ea13066-aa98-ead2-f50f-f62d030ce3c5@huawei.com> |
| |
| From: Zhang Qiao <zhangqiao22@huawei.com> |
| |
| The backport commit 4eec5fe1c680a ("cgroup/cpuset: Fix a race |
| between cpuset_attach() and cpu hotplug") looks suspicious since |
| it comes before commit d74b27d63a8b ("cgroup/cpuset: Change |
| cpuset_rwsem and hotplug lock order") v5.4-rc1~176^2~30 when |
| the locking order was: cpuset lock, cpus lock. |
| |
| Fix it with the correct locking order and reduce the cpus locking |
| range because only set_cpus_allowed_ptr() needs the protection of |
| cpus lock. |
| |
| Fixes: 4eec5fe1c680a ("cgroup/cpuset: Fix a race between cpuset_attach() and cpu hotplug") |
| Reported-by: Michal Koutný <mkoutny@suse.com> |
| Signed-off-by: Zhang Qiao <zhangqiao22@huawei.com> |
| Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> |
| --- |
| kernel/cgroup/cpuset.c | 8 ++++++-- |
| 1 file changed, 6 insertions(+), 2 deletions(-) |
| |
| --- a/kernel/cgroup/cpuset.c |
| +++ b/kernel/cgroup/cpuset.c |
| @@ -1528,9 +1528,13 @@ static void cpuset_attach(struct cgroup_ |
| cgroup_taskset_first(tset, &css); |
| cs = css_cs(css); |
| |
| - cpus_read_lock(); |
| mutex_lock(&cpuset_mutex); |
| |
| + /* |
| + * It should hold cpus lock because a cpu offline event can |
| + * cause set_cpus_allowed_ptr() failed. |
| + */ |
| + get_online_cpus(); |
| /* prepare for attach */ |
| if (cs == &top_cpuset) |
| cpumask_copy(cpus_attach, cpu_possible_mask); |
| @@ -1549,6 +1553,7 @@ static void cpuset_attach(struct cgroup_ |
| cpuset_change_task_nodemask(task, &cpuset_attach_nodemask_to); |
| cpuset_update_task_spread_flag(cs, task); |
| } |
| + put_online_cpus(); |
| |
| /* |
| * Change mm for all threadgroup leaders. This is expensive and may |
| @@ -1584,7 +1589,6 @@ static void cpuset_attach(struct cgroup_ |
| wake_up(&cpuset_attach_wq); |
| |
| mutex_unlock(&cpuset_mutex); |
| - cpus_read_unlock(); |
| } |
| |
| /* The various types of files and directories in a cpuset file system */ |
