What's soft lockup?

[Peter Flynn]

On 05/07/2020 14:13, 孙世龙 sunshilong wrote:

>>> As the subject, how can detect the soft lockup?
>> What do you mean by a soft lockup?
> I apologize for not posting the related error message at the beginning:
> "watchdog: BUG: soft lockup - CPU#0 stuck for 22s!"

Oh dear. I had this exact problem two years ago when I installed Mint 19 
on a Dell XPS 15. It worked, but whenever I tried to suspend or power 
off, I got this message, and the system became unresponsive. The 
shutdown routine was clearly still running because the same message was 
repeated approximately every 20 seconds. The only solution was to hold 
down the power button and force power off. The CPU number was different 
(mine was #7) and I assumed at the time that the core was faulty. But 
otherwise it worked, and I needed it urgently, so I continued to use it 
by rebooting, and to force the power off when I finished work.

Now I'm using a different machine, so last weekend I rebooted the Dell, 
wiped it and installed 19.3. The error has disappeared, and it suspends 
and powers down normally.

I am not really willing to believe that a faulty core has magically 
restored itself to life. I think much more likely there was a problem 
somewhere in the kernel which has been fixed.

I suggest you reboot the system from a CD or USB of a newer version of 
your OS, if possible, and see if that fixes the problem. Otherwise, as 
Ralf said, it really may be a genuine hardware failure.

Peter