[<prev] [next>] [day] [month] [year] [list]
Message-ID: <20251210074112.3975053-1-islituo@gmail.com>
Date: Wed, 10 Dec 2025 15:41:12 +0800
From: Tuo Li <islituo@...il.com>
To: song@...nel.org,
yukuai@...as.com
Cc: linux-raid@...r.kernel.org,
linux-kernel@...r.kernel.org,
Tuo Li <islituo@...il.com>
Subject: [PATCH] md/raid5: fix possible null-pointer dereferences in raid5_store_group_thread_cnt()
The variable mddev->private is first assigned to conf and then checked:
conf = mddev->private;
if (!conf) ...
If conf is NULL, then mddev->private is also NULL. However, the function
does not return at this point, and raid5_quiesce() is later called with
mddev as the argument. Inside raid5_quiesce(), mddev->private is again
assigned to conf, which is then dereferenced in multiple places, for
example:
conf->quiesce = 0;
wake_up(&conf->wait_for_quiescent);
...
This can lead to several null-pointer dereferences.
To fix these issues, the function should unlock mddev and return early when
conf is NULL, following the pattern in raid5_change_consistency_policy().
Signed-off-by: Tuo Li <islituo@...il.com>
---
drivers/md/raid5.c | 7 ++++---
1 file changed, 4 insertions(+), 3 deletions(-)
diff --git a/drivers/md/raid5.c b/drivers/md/raid5.c
index e57ce3295292..be3f9a127212 100644
--- a/drivers/md/raid5.c
+++ b/drivers/md/raid5.c
@@ -7190,9 +7190,10 @@ raid5_store_group_thread_cnt(struct mddev *mddev, const char *page, size_t len)
raid5_quiesce(mddev, true);
conf = mddev->private;
- if (!conf)
- err = -ENODEV;
- else if (new != conf->worker_cnt_per_group) {
+ if (!conf) {
+ mddev_unlock_and_resume(mddev);
+ return -ENODEV;
+ } else if (new != conf->worker_cnt_per_group) {
old_groups = conf->worker_groups;
if (old_groups)
flush_workqueue(raid5_wq);
--
2.43.0
Powered by blists - more mailing lists