On Sat, 17 Jul 2010, Jukka Ruohonen wrote:
On Fri, Jul 16, 2010 at 06:14:39PM -0700, Paul Goyette wrote:Would it be sufficient for ipmi(4) to refuse to suspend (return false from the suspend method) if the watchdog is active? Or should it make some sort of effort to disable the watchdog, and attempt to reenable the watchdog during the subsequent resume?The latter -- given that the same question applies generally to detachment in the brave new modular world. Would sysmon_wdog_shutdown() be sufficient?
Hmmm. sysmon_wdog_shutdown() simply disables the w-dog without making any attempt to store the state. It would seem to me that, if we're going to modify the w-dog state for a suspend, we should restore that state when we resume.
------------------------------------------------------------------------- | Paul Goyette | PGP Key fingerprint: | E-mail addresses: | | Customer Service | FA29 0E3B 35AF E8AE 6651 | paul at whooppee.com | | Network Engineer | 0786 F758 55DE 53BA 7731 | pgoyette at juniper.net | | Kernel Developer | | pgoyette at netbsd.org | -------------------------------------------------------------------------