From mboxrd@z Thu Jan 1 00:00:00 1970 From: Tejun Heo Subject: Re: [PATCH v2.6.36-rc7] infiniband: update workqueue usage Date: Wed, 20 Oct 2010 10:37:03 +0200 Message-ID: <4CBEAA2F.6000009@kernel.org> References: <4CBDB834.2050905@kernel.org> <1287508926.27343.299.camel@chromite.mv.qlogic.com> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: <1287508926.27343.299.camel-/vjeY7uYZjrPXfVEPVhPGq6RkeBMCJyt@public.gmane.org> Sender: linux-rdma-owner-u79uwXL29TY76Z2rM5mHXA@public.gmane.org To: Ralph Campbell Cc: Roland Dreier , Sean Hefty , Hal Rosenstock , "linux-rdma-u79uwXL29TY76Z2rM5mHXA@public.gmane.org" , lkml List-Id: linux-rdma@vger.kernel.org Hello, On 10/19/2010 07:22 PM, Ralph Campbell wrote: > On Tue, 2010-10-19 at 08:24 -0700, Tejun Heo wrote: > >> * qib_cq_wq is a separate singlethread workqueue. Does the queue >> require strict single thread execution ordering? IOW, does each >> work have to be executed in the exact queued order and no two works >> should execute in parallel? Or was the singlethreadedness chosen >> just to reduce the number of workers? > > The work functions need to be called in-order and single threaded > or memory will be freed multiple times and other "bad things". I see, so they'll need to be converted to alloc_ordered_workqueue() once -rc1 merge window opens up. I'll follow up with the conversion. Thanks. -- tejun -- To unsubscribe from this list: send the line "unsubscribe linux-rdma" in the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org More majordomo info at http://vger.kernel.org/majordomo-info.html From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1752282Ab0JTIhP (ORCPT ); Wed, 20 Oct 2010 04:37:15 -0400 Received: from hera.kernel.org ([140.211.167.34]:48807 "EHLO hera.kernel.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752045Ab0JTIhM (ORCPT ); Wed, 20 Oct 2010 04:37:12 -0400 Message-ID: <4CBEAA2F.6000009@kernel.org> Date: Wed, 20 Oct 2010 10:37:03 +0200 From: Tejun Heo User-Agent: Mozilla/5.0 (X11; U; Linux i686 (x86_64); en-US; rv:1.9.2.9) Gecko/20100915 Lightning/1.0b2 Thunderbird/3.1.4 MIME-Version: 1.0 To: Ralph Campbell CC: Roland Dreier , Sean Hefty , Hal Rosenstock , "linux-rdma@vger.kernel.org" , lkml Subject: Re: [PATCH v2.6.36-rc7] infiniband: update workqueue usage References: <4CBDB834.2050905@kernel.org> <1287508926.27343.299.camel@chromite.mv.qlogic.com> In-Reply-To: <1287508926.27343.299.camel@chromite.mv.qlogic.com> X-Enigmail-Version: 1.1.1 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit X-Greylist: Sender IP whitelisted, not delayed by milter-greylist-4.2.3 (hera.kernel.org [127.0.0.1]); Wed, 20 Oct 2010 08:37:05 +0000 (UTC) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Hello, On 10/19/2010 07:22 PM, Ralph Campbell wrote: > On Tue, 2010-10-19 at 08:24 -0700, Tejun Heo wrote: > >> * qib_cq_wq is a separate singlethread workqueue. Does the queue >> require strict single thread execution ordering? IOW, does each >> work have to be executed in the exact queued order and no two works >> should execute in parallel? Or was the singlethreadedness chosen >> just to reduce the number of workers? > > The work functions need to be called in-order and single threaded > or memory will be freed multiple times and other "bad things". I see, so they'll need to be converted to alloc_ordered_workqueue() once -rc1 merge window opens up. I'll follow up with the conversion. Thanks. -- tejun