Per Dan Zink (HP FW/BIOS):
I agree with Linda. There is a problem with the application or how the application uses the watchdog. Removing the watchdog is not a proper solution. I would think this issue is for Canonical to investigate.
In addition, I think there is a second problem here. When the watchdog fires, I expect that we should provide the end user good diagnostic information so they know it was a watchdog timeout. That did not happen here which has delayed root cause. This seems to be a kernel/driver/firmware/platform issue that prevented the watchdog NMI from being reported in customer friendly terms. This probably falls on HP first. Tom, can you dig a little deeper into that?
Per Dan Zink (HP FW/BIOS):
I agree with Linda. There is a problem with the application or how the application uses the watchdog. Removing the watchdog is not a proper solution. I would think this issue is for Canonical to investigate.
In addition, I think there is a second problem here. When the watchdog fires, I expect that we should provide the end user good diagnostic information so they know it was a watchdog timeout. That did not happen here which has delayed root cause. This seems to be a kernel/ driver/ firmware/ platform issue that prevented the watchdog NMI from being reported in customer friendly terms. This probably falls on HP first. Tom, can you dig a little deeper into that?