Slight interface change on the watchdog fido

[Nick Hibma]

> I recognise this as the previous watchdog function API with
> the extension that when told to disarm via a timeout value
> of 0 the watchdog may now report failure to do so by setting
> *error, correct?

True, see the changes on the manpage.

> If so this is fine with the stock mk48txx.c.

Yes, that was the one watchdog that followed the sample in the Elan 
specific code correctly. At first I didn't have the interface change in 
and that meant changing mk48txx.c and I missed that afterwards I changed 
everything back to what is was in that driver.

> Regarding your changes to watchdog.4 and watchdog.9 AFAICT
> they violate the established style guidelines for FreeBSD
> man pages; the source should wrap the line on sentence breaks.

Is there a manual for man page writing? I've not done much of that yet.

Nick