Re: CPUFREQ-CPUHOTPLUG: Possible circular locking dependency

[Gautham R Shenoy <ego@in.ibm.com>]

On Thu, Nov 30, 2006 at 09:58:07AM +0530, Gautham R Shenoy wrote:
> 
> So can we ignore this circular-dep warning as a false positive?
> Or is there a way to exploit this circular dependency ?
> 
> At the moment, I cannot think of way to exploit this circular dependency
> unless we do something like try destroying the created workqueue when the
> cpu is dead, i.e make the cpufreq governors cpu-hotplug-aware.
> (eeks! that doesn't look good)

Ok, I see that we are already doing it :(. So we can end up in a
deadlock.

Here's the culprit callpath:

_cpu_down()
!
!-> raw_notifier_call_chain(CPU_LOCK_ACQUIRE)
!	!
!	!-> workqueue_cpu_mutex(CPU_LOCK_ACQUIRE) [*]
!
!-> raw_notifier_call_chain(CPU_DEAD)
	!
	!-> cpufreq_cpu_callback (CPU_DEAD)
		!
		!-> cpufreq_remove_dev
			!
			!-> __cpufreq_governor(data, GOVERNOR_STOP)
				!
				!-> policy->governor->governor()
					!
					!-> cpufreq_governor_dbs(GOVERNOR_STOP)
						!
						!-> destroy_workqueue() [*]

[*] indicates function takes workqueue_mutex.

So a deadlock!

I wasn't able to observe this because I'm running Xeon SMP box on which
you cannot offline cpu0. And cpufreq data is created only for cpu0,
while all other cpus cpufreq_data just point to cpu0's cpufreq_data.

So the mentioned callpath within  cpufreq_remove_dev is never reached
during the normal cpu offline cycle.

However, if there are architectures which allow the first-booted-cpu
(or to be precise, the cpu for which cpufreq_data is *actually* created) 
to be offlined and we are running Ondemand governor during the offline,
we will see this deadlock.

regards
gautham.
-- 
Gautham R Shenoy
Linux Technology Center
IBM India.
"Freedom comes with a price tag of responsibility, which is still a bargain,
because Freedom is priceless!"