[PATCH] cgroup: pair max limit READ_ONCE() with WRITE_ONCE()

[PATCH] cgroup: pair max limit READ_ONCE() with WRITE_ONCE()

[Ren Tamura]

cgroup.max.descendants and cgroup.max.depth are shown through seq_file.
Their show callbacks read cgrp->max_descendants and cgrp->max_depth with
READ_ONCE(), respectively.

The corresponding write callbacks update the same scalar fields while
holding the cgroup lock, but the seq_file show path does not serialize
against those stores. This leaves the lockless show-side loads annotated
with READ_ONCE(), while the corresponding stores remain plain stores.

Use WRITE_ONCE() for the updates so the intended lockless access is marked
consistently on both sides. This does not change locking, ordering, or
user-visible semantics.

Assisted-by: OpenAI-Codex:gpt-5.5
Signed-off-by: Ren Tamura <ren.tamura.oss@gmail.com>
---
 kernel/cgroup/cgroup.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/kernel/cgroup/cgroup.c b/kernel/cgroup/cgroup.c
index 6152add0c..daddfc2b9 100644
--- a/kernel/cgroup/cgroup.c
+++ b/kernel/cgroup/cgroup.c
@@ -3726,7 +3726,7 @@ static ssize_t cgroup_max_descendants_write(struct kernfs_open_file *of,
 	if (!cgrp)
 		return -ENOENT;
 
-	cgrp->max_descendants = descendants;
+	WRITE_ONCE(cgrp->max_descendants, descendants);
 
 	cgroup_kn_unlock(of->kn);
 
@@ -3769,7 +3769,7 @@ static ssize_t cgroup_max_depth_write(struct kernfs_open_file *of,
 	if (!cgrp)
 		return -ENOENT;
 
-	cgrp->max_depth = depth;
+	WRITE_ONCE(cgrp->max_depth, depth);
 
 	cgroup_kn_unlock(of->kn);
 

base-commit: eb3f4b7426cfd2b79d65b7d37155480b32259a11
-- 
2.53.0

Re: [PATCH] cgroup: pair max limit READ_ONCE() with WRITE_ONCE()

[Tejun Heo]

Hello,

Applied to cgroup/for-7.2.

Thanks.

--
tejun