Re: [RFC][PATCH] CPUSets: Move most calls to rebuild_sched_domains() to the workqueue

[Max Krasnyansky <maxk@qualcomm.com>]

Gautham R Shenoy wrote:
> On Fri, Jun 27, 2008 at 08:52:28AM +0530, Gautham R Shenoy wrote:
>> On Thu, Jun 26, 2008 at 12:56:49AM -0700, Paul Menage wrote:
>>> CPUsets: Move most calls to rebuild_sched_domains() to the workqueue
>>>
>>> In the current cpusets code the lock nesting between cgroup_mutex and
>>> cpuhotplug.lock when calling rebuild_sched_domains is inconsistent -
>>> in the CPU hotplug path cpuhotplug.lock nests outside cgroup_mutex,
>>> and in all other paths that call rebuild_sched_domains() it nests
>>> inside.
>>>
>>> This patch makes most calls to rebuild_sched_domains() asynchronous
>>> via the workqueue, which removes the nesting of the two locks in that
>>> case. In the case of an actual hotplug event, cpuhotplug.lock nests
>>> outside cgroup_mutex as now.
>>>
>> Using a multithreaded workqueue(kevent here) for this is not such a
>> great idea this,since currently we cannot call
>> get_online_cpus() from a workitem executed by a multithreaded workqueue.
>>
>> Can one use a single threaded workqueue here instead ?
We could certainly do it in the single threaded workqueue. It won't help to
avoid circular locking dependencies though.
Did you get a chance to read entire thread on this topic ? There were some
questions that you might be able to answer.

Max