[PATCH 4.19 050/220] watchdog: core: fix null pointer dereference when releasing cdev

From: Greg Kroah-Hartman
Date: Fri Nov 22 2019 - 05:58:01 EST


From: Wolfram Sang <wsa+renesas@xxxxxxxxxxxxxxxxxxxx>

[ Upstream commit 953b9dd7725bad55a922a35e75bff7bebf7b9978 ]

watchdog_stop() calls watchdog_update_worker() which needs a valid
wdd->wd_data pointer. So, when unregistering the cdev, clear the
pointers after we call watchdog_stop(), not before.

Fixes: bb292ac1c602 ("watchdog: Introduce watchdog_stop_on_unregister helper")
Signed-off-by: Wolfram Sang <wsa+renesas@xxxxxxxxxxxxxxxxxxxx>
Reviewed-by: Fabrizio Castro <fabrizio.castro@xxxxxxxxxxxxxx>
Reviewed-by: Guenter Roeck <linux@xxxxxxxxxxxx>
Signed-off-by: Guenter Roeck <linux@xxxxxxxxxxxx>
Signed-off-by: Wim Van Sebroeck <wim@xxxxxxxxxxxxxxxxxx>
Signed-off-by: Sasha Levin <sashal@xxxxxxxxxx>
---
drivers/watchdog/watchdog_dev.c | 10 +++++-----
1 file changed, 5 insertions(+), 5 deletions(-)

diff --git a/drivers/watchdog/watchdog_dev.c b/drivers/watchdog/watchdog_dev.c
index ffbdc4642ea55..f6c24b22b37c0 100644
--- a/drivers/watchdog/watchdog_dev.c
+++ b/drivers/watchdog/watchdog_dev.c
@@ -1019,16 +1019,16 @@ static void watchdog_cdev_unregister(struct watchdog_device *wdd)
old_wd_data = NULL;
}

- mutex_lock(&wd_data->lock);
- wd_data->wdd = NULL;
- wdd->wd_data = NULL;
- mutex_unlock(&wd_data->lock);
-
if (watchdog_active(wdd) &&
test_bit(WDOG_STOP_ON_UNREGISTER, &wdd->status)) {
watchdog_stop(wdd);
}

+ mutex_lock(&wd_data->lock);
+ wd_data->wdd = NULL;
+ wdd->wd_data = NULL;
+ mutex_unlock(&wd_data->lock);
+
hrtimer_cancel(&wd_data->timer);
kthread_cancel_work_sync(&wd_data->work);

--
2.20.1