Re: extend sysmon watchdog modes?

[Paul Koning <paul_koning%dell.com@localhost>]

On May 26, 2011, at 12:57 PM, Greg Troxel wrote:

> 
>  I am seeing a need for a new watchdog mode that is similar to
>  WDOG_MODE_KTICKLE but instead of the watchdog service function
>  running from a callout it needs to run from a (higher level)
>  hardware interrupt.
> 
> I don't understand the use case.  I am running watchdogs in user tickle,
> and consider the machine wedged if it can't run the tickle code in 30s.
> 
> I have been thinking about going the other way, which is to add a disk
> argument to wdogctl for a 'disk-gated user tickle mode', where every
> time it wakes up it does
> 
>  lseek to a random block on the disk
>  read it
>  then do the tickle
> 
> or perhaps do some sort of random find.  The point is to reset when the
> system is functioning well enough that the user tickle runs but the
> system isn't really ok (like a tstile wedge, or a wedged disk
> controller).

Makes sense to me.  My practice has been to put watchdog tickle into a daemon, 
on the grounds that a system not healthy enough to schedule user mode code is 
dead.  Watchdog tickle in an ISR seems counterproductive.

        paul