Re: [PATCH 1/1] cgroup: Fix race condition at rebind_subsystems()

From: Mukesh Ojha
Date: Tue Aug 23 2022 - 03:35:15 EST




On 8/23/2022 11:11 AM, Jing-Ting Wu wrote:
Root cause:
The rebind_subsystems() is no lock held when move css object from A
list to B list,then let B's head be treated as css node at
list_for_each_entry_rcu().

Solution:
Add grace period before invalidating the removed rstat_css_node.

Reported-by: Jing-Ting Wu <jing-ting.wu@xxxxxxxxxxxx>
Suggested-by: Michal Koutný <mkoutny@xxxxxxxx>
Signed-off-by: Jing-Ting Wu <jing-ting.wu@xxxxxxxxxxxx>
Tested-by: Jing-Ting Wu <jing-ting.wu@xxxxxxxxxxxx>
Link: https://lore.kernel.org/linux-arm-kernel/d8f0bc5e2fb6ed259f9334c83279b4c011283c41.camel@xxxxxxxxxxxx/T/
---
kernel/cgroup/cgroup.c | 1 +
1 file changed, 1 insertion(+)

diff --git a/kernel/cgroup/cgroup.c b/kernel/cgroup/cgroup.c
index ffaccd6373f1..0d0c959966ed 100644
--- a/kernel/cgroup/cgroup.c
+++ b/kernel/cgroup/cgroup.c
@@ -1820,6 +1820,7 @@ int rebind_subsystems(struct cgroup_root *dst_root, u16 ss_mask)
if (ss->css_rstat_flush) {
list_del_rcu(&css->rstat_css_node);
+ synchronize_rcu();
list_add_rcu(&css->rstat_css_node,
&dcgrp->rstat_css_list);
}

Good catch.

Acked-by: Mukesh Ojha <quic_mojha@xxxxxxxxxxx>

-Mukesh