Re: [PATCH] watchdog: Fix an invalid memory access in 'watchdog_cdev_unregister()'

[Guenter Roeck <linux@roeck-us.net>]

On 7/29/21 9:16 AM, Christophe JAILLET wrote:
> A few lines before 'watchdog_hrtimer_pretimeout_stop(wdd)', we explicitly
> set 'wdd->wd_data' to NULL.
> So, it is more than likely than this call will lead to an invalid
> memory access.
> 
> Move this call before the 'wdd->wd_data = NULL;'
> 
> Fixes: 7b7d2fdc8c3e ("watchdog: Add hrtimer-based pretimeout feature")
> Signed-off-by: Christophe JAILLET <christophe.jaillet@wanadoo.fr>
> ---
> Completely untested!
> Not sure at all, that it is the way to fix it.

Nice catch. Should be ok.

Reviewed-by: Guenter Roeck <linux@roeck-us.net>

> ---
>   drivers/watchdog/watchdog_dev.c | 3 ++-
>   1 file changed, 2 insertions(+), 1 deletion(-)
> 
> diff --git a/drivers/watchdog/watchdog_dev.c b/drivers/watchdog/watchdog_dev.c
> index 3bab32485273..ffd8f1a82355 100644
> --- a/drivers/watchdog/watchdog_dev.c
> +++ b/drivers/watchdog/watchdog_dev.c
> @@ -1096,6 +1096,8 @@ static void watchdog_cdev_unregister(struct watchdog_device *wdd)
>   		watchdog_stop(wdd);
>   	}
>   
> +	watchdog_hrtimer_pretimeout_stop(wdd);
> +
>   	mutex_lock(&wd_data->lock);
>   	wd_data->wdd = NULL;
>   	wdd->wd_data = NULL;
> @@ -1103,7 +1105,6 @@ static void watchdog_cdev_unregister(struct watchdog_device *wdd)
>   
>   	hrtimer_cancel(&wd_data->timer);
>   	kthread_cancel_work_sync(&wd_data->work);
> -	watchdog_hrtimer_pretimeout_stop(wdd);
>   
>   	put_device(&wd_data->dev);
>   }
>