[PATCH] rtc: mc13xxx: fix a double-unlock issue

[PATCH] rtc: mc13xxx: fix a double-unlock issue

[wu000273]

From: Qiushi Wu <wu000273@umn.edu>

In function mc13xxx_rtc_probe, the mc13xxx_unlock() is called
before rtc_register_device(). But in the error path of
rtc_register_device(), the mc13xxx_unlock() is called again,
which causes a double-unlock problem. To fix this problem, we
need to call mc13xxx_lock() again in this error path.

Signed-off-by: Qiushi Wu <wu000273@umn.edu>
---
 drivers/rtc/rtc-mc13xxx.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/drivers/rtc/rtc-mc13xxx.c b/drivers/rtc/rtc-mc13xxx.c
index afce2c0b4bd6..d6802e6191cb 100644
--- a/drivers/rtc/rtc-mc13xxx.c
+++ b/drivers/rtc/rtc-mc13xxx.c
@@ -308,8 +308,10 @@ static int __init mc13xxx_rtc_probe(struct platform_device *pdev)
 	mc13xxx_unlock(mc13xxx);
 
 	ret = rtc_register_device(priv->rtc);
-	if (ret)
+	if (ret) {
+		mc13xxx_lock(mc13xxx);
 		goto err_irq_request;
+	}
 
 	return 0;
 
-- 
2.17.1

Re: [PATCH] rtc: mc13xxx: fix a double-unlock issue

[Markus Elfring]

> … To fix this problem,
> we need to call mc13xxx_lock() again in this error path.

How do you think about a wording variant like the following?

   Change description:
   …
   Thus add a call of the function “mc13xxx_lock” in an if branch
   for the completion of the exception handling.


Would you like to add the tag “Fixes”?

Regards,
Markus