Re: Kernel hang with latest snapshot

[Li Guanglei <guanglei at cn dot ibm dot com>]

David Boreham wrote:
> > My test on thinkpad also shows that it is unstable. Sometimes it will 
> > crash within a minute, and sometimes it will run almost 20 minutes 
> > without crash.
> >
> > I will try to narrow down the probe list to find out which probe 
> > points caused the system crash. And you can only choose to probe those 
> > hooks that are meaningful to you instead of addevent.* to reduce the 
> > possibility of cash down
>
> Unfortunately it's crashed for me on every syscall I have tried
> (I think I tested sync, write, open, read).

Hi David,

  I did found some bugs in ioscheduler trace hooks today, which could 
cause reference to NULL pointer. I've committed the fix. Please retry 
and tell me if you still have problem. And it's better if you can post 
more detail about the failure, such as which trace hooks you are using 
and the kernel OOPS message, so that I can figure out more clearly 
where the problem is.

 And thank Bibo Mao for his help on this :)

- Guanglei