From: Tejun Heo Subject: Re: [PATCH v4 13/14] workqueue: implement NUMA affinity for unbound workqueues Date: Sun, 31 Mar 2013 12:06:33 -0700 Message-ID: <20130331190633.GC7533@htj.dyndns.org> References: <1364453020-2829-1-git-send-email-tj@kernel.org> <1364453020-2829-14-git-send-email-tj@kernel.org> <20130329224434.GC9862@htj.dyndns.org> <51571FA2.2050804@cn.fujitsu.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: Lai Jiangshan , Jens Axboe , Jan Kara , fengguang.wu@intel.com, jmoyer@redhat.com, zab@redhat.com, LKML , Herbert Xu , "David S. Miller" , linux-crypto@vger.kernel.org To: Lai Jiangshan Return-path: Received: from mail-pa0-f48.google.com ([209.85.220.48]:48101 "EHLO mail-pa0-f48.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752869Ab3CaTGi (ORCPT ); Sun, 31 Mar 2013 15:06:38 -0400 Content-Disposition: inline In-Reply-To: <51571FA2.2050804@cn.fujitsu.com> Sender: linux-crypto-owner@vger.kernel.org List-ID: Hello, Lai. On Sun, Mar 31, 2013 at 01:23:46AM +0800, Lai Jiangshan wrote: > But for unbound wq when cpuhotplug > w/o NUMA affinity, works are always in the cpus if there is online cpu in wq's cpumask > w/ NUMA affinity, ......... NOT always ........ even .................................... Yeah, this is rather unfortunate but cpumask for unbound workqueue is a completely new thing anyway and I think providing a similar guarantee as per-cpu should be enough. Things are much simpler that way and requiring users which depend on hard affinity to take care of flushing is reasonable enough and in line with how workqueue has traditionally been working. > > Workqueue's affinity guarantee is very specific - the work item owner is > > responsible for flushing the work item during CPU DOWN if it wants > > to guarantee affinity over full execution. > > Could you add the comments and add Reviewed-by: Lai Jiangshan > for the patchset? Sure thing. Thanks. -- tejun