Closed paralin closed 7 years ago
You might have two watchdog device file getting configured.
odroid@odroid64:~$ ls -la /dev/watchdog
watchdog watchdog0 watchdog1
odroid@odroid64:~$ ls -la /dev/watchdog
This is because of the CONFIG_SOFT_WATCHDOG being set to y Actually when running under watchdog demon the device keep setting keep-alive ping message to the device. So when some unwanted steps trigger watchdog reset event it should release the device node and stop the timer. Neither of this is happening with this driver. The device resource keeps busy state (with continues watchdog ping).
Not sure how to fix what you describe. The changes I made fixed the watchdog, does it make sense that those changes would work given what you're saying is happening ?
latest WD driver have been updated please give it a try.
Now, I'm not saying you merge this - please don't without some scrutiny - but this change fixed a reboot loop for me.
I increased the timings on the watchdog and removed the disabled line.
I'm not sure which of these 2 things fixed the problem, but it did. And it might be worth thinking about this. It was a bit difficult to find this as the cause of the boot loop, but it makes sense.
This change is