Most of boards have some kind of watchdog on board (in MIO or in south 
bridge), or some of them. For ex., iTCO_wdt for Intel, it87_wdt for ITE 
MIO, and so on.

But built-in watchdog behavior may differ depending board model (it may 
reboot board, or it may just poweroff board, like ITE watchdog on old 
Asus K8N).

Also, Intel watchdog driver is loaded automatically, MIO watchdog 
drivers are loaded manually (from command line or from /etc/modules).

You can try it - load driver, and then do 'killall -9 watchdog' (to kill 
watchdog process & simulate system hangup). If board will be rebooted - 
all is OK.

On 21.09.2016 05:29, Jean-Roch Blais wrote:
> Hello List,
> My previous pentium IV motherboard running BUC 5.1.7 was randomly crashing, I 
> had to use hard reset to start it again, and it would run fine, as long as 
> there was no intense traffic, I guess...
> Since then I replaced it with another one running BUC 5.2.5, as one can 
> probably tell from my dhcpcd/dnsmasq adventure in my previous e-mails to the 
> list…
> This made me wonder how the watchdog timer was working on BUC.  I think Erich 
> Titl mentioned in some previous mail that the software watchdog was writing 
> to /dev/watchdog every 10 seconds to prevent a system reboot/reset. This 
> prevents software lockups but does not cover hardware problems like the one I 
> experienced with my sick mobo. Obviously if the CPU is not running you’re 
> hanged :-) .
> Has anybody ever worked on an external hardware watchdog reset timer that 
> would do such a hardware reset ?
> Thank’s
> jrb
> ------------------------------------------------------------------------------
> ------------------------------------------------------------------------
> leaf-user mailing list:
> Support Request --

leaf-user mailing list:
Support Request --

Reply via email to