Re: [PATCH 2/2] watchdog: core: Fix watchdog_init_timeout() when invalid param / valid dt

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

On 11/26/2013 10:22 AM, Doug Anderson wrote:
> There was a minor bug in watchdog_init_timeout() where it would return
> an error code if someone specified an invalid parameter on the
> command line but then there was a valid parameter in the device tree
> as "timeout-sec".
>
> Signed-off-by: Doug Anderson <dianders@chromium.org>

I thought that was on purpose.

Problem as I see it is that users would expect the timeout to be set to
the provided parameter, which would be silently ignored and replaced
by timeout-sec if the parameter is wrong and timeout-sec is specified.
Seems to me that the user should be informed about the problem,
and not be permitted to provide invalid parameters.

Thanks,
Guenter

> ---
>   drivers/watchdog/watchdog_core.c | 8 ++++----
>   1 file changed, 4 insertions(+), 4 deletions(-)
>
> diff --git a/drivers/watchdog/watchdog_core.c b/drivers/watchdog/watchdog_core.c
> index cec9b55..8d27753 100644
> --- a/drivers/watchdog/watchdog_core.c
> +++ b/drivers/watchdog/watchdog_core.c
> @@ -82,12 +82,12 @@ int watchdog_init_timeout(struct watchdog_device *wdd,
>   		wdd->timeout = timeout_parm;
>   		return ret;
>   	}
> -	if (timeout_parm)
> -		ret = -EINVAL;
>
> -	/* try to get the timeout_sec property */
> +	/* if no device tree then we're done */
>   	if (dev == NULL || dev->of_node == NULL)
> -		return ret;
> +		return (timeout_parm) ? -EINVAL : ret;

( ) is unnecessary.

> +
> +	/* try to get the timeout_sec property */
>   	of_property_read_u32(dev->of_node, "timeout-sec", &t);
>   	if (!watchdog_timeout_invalid(wdd, t) && t)
>   		wdd->timeout = t;
>