On my (embedded) system I have multiple watchdog character devices:
# ls -al /dev/watchdog*
crw------- 1 imp root 10, 130 Apr 26 07:43 /dev/watchdog
crw------- 1 root root 253, 0 Apr 26 07:44 /dev/watchdog0
crw------- 1 root root 253, 1 Apr 26 07:44 /dev/watchdog1
# dmesg | grep -i watchdog
[ 2.342104] omap_wdt: OMAP Watchdog Timer Rev 0x01: initial timeout 60 sec
[ 6.713125] softdog: Software Watchdog Timer: 0.08 initialized. soft_noboot=0 soft_margin=120 sec soft_panic=0 (nowayout=1)
I'm pretty sure that /dev/watchdog0
is the hardware watchdog and /dev/watchdog1
is the software watchdog, but why is there also a /dev/watchdog
?
Background: The software watchdog is configured with nowayout
flag, so, once it has been feeded it won't stop anymore until reboot. However, writing to /dev/watchdog
doesn't seem to trigger the software watchdog and doesn't give me the important nowayout
feature. I could use /dev/watchdog1
directly, but then I'd need some reliable way to identify the correct software watchdog device filename…
Best Answer
I had the same question. According to the kernel documentation:
In other words,
/dev/watchdog
and/dev/watchdog0
both point to the same device. Any additional watchdogs will be numbered greater than0
and will have only a single device node.