On 01/14/14 at 11:24am, Vivek Goyal wrote: > On Tue, Jan 14, 2014 at 12:16:39PM +0000, One Thousand Gnomes wrote: > > On Tue, 14 Jan 2014 16:23:23 +0800 > > Dave Young <dyoung@xxxxxxxxxx> wrote: > > > > > In kdump kernel watchdog could interrupt vmcore capturing because we > > > have no way to disable/stop it while crashing happens. > > > > Lots of watchdogs cannot be stopped. > > > > > Add a module parameter stop_before_register so watchdog can be stopped > > > before register in driver loading path. Thus we can try to load the > > > watchdog driver as early as possible in kdump kernel to ensure vmcore > > > capturing. > > > > If you want to kdump then don't start the watchdog. The goal of the > > watchdog is to make sure the system never gets stuck. Adding conditions > > and special cases simply increases the odds of something bad not > > triggering the watchdog. > > > > If you have a system that can stop the watchdog then providing no way out > > is not set you can open it and stop it. > > > > I don't see the need for any kernel change here > > > > - if it can't be stopped you lost > > - if "nowayout" is set then by design you lost > > - if it can be stopped, you can open and stop it > > > > Now whether in the !nowayout case the watchdog core should catch whatever > > hooks/notifiers are available and stop any watchdogs it can on a > > kexec/kdump is a more interesting question and probably needs to default > > to not doing so but with the option to force otherwise for debugging work. > > Hi All, > > I thought this problem was resolved (atleast conceptually) last time > when Don Zickus brought it up. > > He mentioned that it was concluded that keep watchdog interval long > enough, say 60 seconds and keep on kicking it fast enough, say every > 10-20 seconds. That would ensure that after the crash, there is atleast > 60 - 20 = 40 seconds left before watchdog expires. And in that duration > we should try to boot into second kernel load watchdog driver early enough > from initramfs which can start kicking watchdog again. Some drivers did stop the watchdog while module loading such as iTCO_wdt. so we can load them as early as possible and not necessary to kick them again. Thus I wrote this patch to add the stop to generic code so more drivers can be covered. But as Alan said I begin to feel that this is not a good design. the iTCO_wdt nowayout become useless because of this stop_before_register.. For wdts which can not be stopped we can still continue working on kicking them early in initramfs. > > I am wondering what happened to this idea. Dave, did we try to implement/ > experiment with this? No, we are just begin with iTCO_wdt and is trying to add iTCO_wdt firstly. Other devices has not been investigated. Thanks Dave -- To unsubscribe from this list: send the line "unsubscribe linux-watchdog" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
