Difference between revisions of "Continuous Logging for Watchdog Timer Expiration"
m (Add category) |
|||
Line 1: | Line 1: | ||
− | |||
Implementation Idea | Implementation Idea | ||
Line 5: | Line 4: | ||
--Leon Woestenberg | --Leon Woestenberg | ||
+ | |||
+ | [[Category:Kernel]] |
Latest revision as of 00:13, 28 October 2011
Implementation Idea
I think this can be achieved by modifying the driver that services the hardware watchdog. This driver is aware of the hardware timeout (say this is N) and the application can make it a rule to always service the hardware watchdog device (/dev/wtd) every N/2 seconds. The driver can now warn when it has not been serviced during the first N/2 and has N/2 left to dump this event somehow.
--Leon Woestenberg