Re: Mysterious CFQ crash and RCU

[Vivek Goyal <vgoyal@redhat.com>]

On Sat, May 21, 2011 at 02:00:13PM -0700, Paul E. McKenney wrote:

[..]
> > Is there any known issue or is there any quick tip on how can I 
> > go about debugging it further from rcu point of view.
> 
> First for uses of RCU:
> 
> o	One thing to try would be CONFIG_PROVE_RCU, which could help
> 	find missing rcu_read_lock()s and similar.  Some years back, it
> 	used to be the case that spin_lock() implied rcu_read_lock(),
> 	but it no longer does.	There might still be some cases where
> 	spin_lock() needs to have an rcu_read_lock() added.

In this case we take explicit rcu_read_lock() in call_for_each_cic()
and do not rely on that spin_lock() also means rcu_read_lock().

call_for_each_cic() {
	rcu_read_lock();
	 hlist_for_each_entry_rcu(cic, n, &ioc->cic_list, cic_list)
                func(ioc, cic);
        rcu_read_unlock();
}

> 
> o	There are a few entries in the bugzilla mentioning that elements
> 	are being removed more often than expected.

Are you referring to my comments about additional messages of cgroup
changed. If yes, that issue has now been identified and I have posted
a fix to get rid of thos unnecessary calls.

https://lkml.org/lkml/2011/5/23/201

Is still there a need to enable CONFIG_DEBUG_OBJECTS_RCU_HEAD as the
kernel in question is non-preemptible one.

Thanks
Vivek