RFC: generic support for two-stage watchdogs?

[Chris Friesen]

When using kdump to store crash recovery information, it is frustrating 
when the watchdog timer fires and reboots the system under our feet.

Many hardware watchdogs support a two-stage operation where the initial 
stage expires and sends an NMI or other interrupt to the CPU.  Only once 
the second stage fires does it actually reboot the hardware.

Has anyone considered adding support for this sort of hardware to the 
/dev/watchdog API?  It seems like it would make sense to reset the 
watchdog timeout to some suitable period and trigger kdump.  This would 
let us preserve the crash information.

If the system is really fubared then the second stage will fire and 
reboot the machine.

Chris

--
Chris Friesen
Software Developer
GENBAND
chris.friesen@xxxxxxxxxxx
www.genband.com
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/