Re: [PATCH 1/4] workqueue: Create low-level unbound workqueues cpumask

From: Tejun Heo <tj@kernel.org>
To: Frederic Weisbecker <fweisbec@gmail.com>
Cc: LKML <linux-kernel@vger.kernel.org>,
	Christoph Lameter <cl@linux.com>,
	Kevin Hilman <khilman@linaro.org>,
	Lai Jiangshan <laijs@cn.fujitsu.com>,
	Mike Galbraith <bitbucket@online.de>,
	"Paul E. McKenney" <paulmck@linux.vnet.ibm.com>,
	Viresh Kumar <viresh.kumar@linaro.org>
Subject: Re: [PATCH 1/4] workqueue: Create low-level unbound workqueues cpumask
Date: Thu, 24 Apr 2014 11:37:16 -0400	[thread overview]
Message-ID: <20140424153716.GG14460@htj.dyndns.org> (raw)
In-Reply-To: <1398350256-7834-2-git-send-email-fweisbec@gmail.com>

On Thu, Apr 24, 2014 at 04:37:33PM +0200, Frederic Weisbecker wrote:
> Create a cpumask that limit the affinity of all unbound workqueues.
> This cpumask is controlled though a file at the root of the workqueue
> sysfs directory.
> 
> It works on a lower-level than the per WQ_SYSFS workqueues cpumask files
> such that the effective cpumask applied for a given unbound workqueue is
> the intersection of /sys/devices/virtual/workqueue/$WORKQUEUE/cpumask and
> the new /sys/devices/virtual/workqueue/cpumask_unbounds file.

Let's drop "_unbounds" postfix and name it just "cpumask".  We don't
apply it to per-cpu workqueues now but that really is an
implementation detail and later when (and if) we actually distinguish
per-cpu usages for correctness from for optimization, we may as well
apply the same cpumask to per-cpu ones too.

Another thing with naming is that I didn't anticipate having
attributes at the top directory so the workqueue directories aren't
namespaced.  Maybe we want to namespace top level knobs?
"system_cpumask" maybe?  Any better ideas?

Thanks.

-- 
tejun