Subject: Re: rfc: an alternate user-tickle mode for wdogctl(8)
To: None <tech-userlevel@netbsd.org>
From: David Young <dyoung@pobox.com>
List: tech-userlevel
Date: 08/06/2006 21:16:41
On Mon, Aug 07, 2006 at 10:30:12AM +1000, Daniel Carosone wrote:
> On Sun, Aug 06, 2006 at 02:18:11PM -0500, David Young wrote:
> > I have added a flag to wdogctl(8), -x, that activates a user-mode timer
> > that is NOT disarmed when wdogctl closes the watchdog device.  See the
> > attached patch.  The -x mode protects against the case where the kernel
> > kills wdogctl, sshd, and other essential userland programs are killed
> > due to, say, memory exhaustion, lobotomizing a mission-critical NetBSD
> > system and necessitating an operator visit to reboot it.  If there are
> > no objections, I will commit this on Wednesday.
> 
> No objection, but a suggestion: is there a way to tickle such a timer
> with a later re-invocation of wdogctl?  You could then use it to test
> that (say) cron and process forking and other such things were still
> working, too.
> 
> (after reading the manpage again)
> 
> Ahah, these are called 'external' timers.  So, perhaps have this new
> flag make the 'user mode' timer become an 'external mode' timer on
> exit of wdogctl, rather than have it be cancelled?

How will wdogctl change the mode of the timer as it is being killed off?
I don't think it can be done.  Anyway, the net effect would be identical,
wouldn't it be?

Dave

-- 
David Young             OJC Technologies
dyoung@ojctech.com      Urbana, IL * (217) 278-3933