[PATCH 1/2] Add a simple watchdog framework

[Juergen Beisert]

Sascha Hauer wrote:
> On Fri, Jun 22, 2012 at 11:54:02AM +0200, Juergen Beisert wrote:
> > This patch adds a simple wd command which can setup, trigger and stop a
> > watchdog on the platform.
> >
> > +static int do_wd(int argc, char *argv[])
> > +{
> > +	int rc;
> > +
> > +	if (argc > 1) {
> > +		if (isdigit(*argv[1])) {
> > +			timeout = simple_strtoul(argv[1], NULL, 0);
> > +		} else {
> > +			printf("numerical parameter expected\n");
> > +			return 1;
> > +		}
> > +	}
> > +
> > +	rc = watchdog_set_timeout(timeout);
> > +	if (rc == -EINVAL) {
>
> Why do you check for -EINVAL only? This way all other errors will be
> silently ignored.

Are you sure we must handle other failure codes than -EINVAL here? The called 
function is very simple and only must distinguish "timeout != 0" and "timeout 
== 0". So, timeout can be "out of range" or - as you mentioned - some 
platforms cannot disable the watchdog anymore once it is enabled. Both 
results into -EINVAL, because the called function cannot use the given value. 
What else can happen?

> > +		if (timeout == 0)
> > +			printf("Watchdog cannot be disabled\n");
> > +		else
> > +			printf("Timeout value out of range\n");
> > +		return rc;
>
> Do not return negative error codes here. The shell will interpret
> negative numbers as 'exit'. You have to return 1 for failure.

Okay, missed that.

> > +	}
> > +	return 0;
> > +}
> > +
> > +
> > +#ifndef INCLUDE_WATCHDOG_H
> > +# define INCLUDE_WATCHDOG_H
> > +
> > +extern int watchdog_set_timeout(unsigned);
>
> extern is not needed here.

Okay.

jbe

-- 
Pengutronix e.K.                              | Juergen Beisert             |
Linux Solutions for Science and Industry      | http://www.pengutronix.de/  |