Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1755630AbZGMMQE (ORCPT ); Mon, 13 Jul 2009 08:16:04 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1755597AbZGMMQC (ORCPT ); Mon, 13 Jul 2009 08:16:02 -0400 Received: from e38.co.us.ibm.com ([32.97.110.159]:40346 "EHLO e38.co.us.ibm.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1755591AbZGMMQB (ORCPT ); Mon, 13 Jul 2009 08:16:01 -0400 Date: Mon, 13 Jul 2009 17:41:38 +0530 From: Balbir Singh To: Paul Menage Cc: lizf@cn.fujitzu.com, bblum@google.com, linux-kernel@vger.kernel.org, akpm@linux-foundation.org, containers@lists.linux-foundation.org, libcg-devel Subject: Re: [PATCH 0/2] CGroups: cgroup member list enhancement/fix Message-ID: <20090713121138.GC5051@balbir.in.ibm.com> Reply-To: balbir@linux.vnet.ibm.com References: <20090702231814.3969.44308.stgit@menage.mtv.corp.google.com> <20090705063850.GX11273@balbir.in.ibm.com> <6599ad830907101658i13e4046br70377a487dd6b49b@mail.gmail.com> MIME-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Disposition: inline In-Reply-To: <6599ad830907101658i13e4046br70377a487dd6b49b@mail.gmail.com> User-Agent: Mutt/1.5.18 (2008-05-17) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 1377 Lines: 32 * menage@google.com [2009-07-10 16:58:23]: > On Sat, Jul 4, 2009 at 11:38 PM, Balbir Singh wrote: > > > > Paul, I don't see an interface to migrate all procs or at-least I > > can't read it in the changelog. As discussed in the containers > > mini-summit in 2008, it would be a nice thing to have (I've not looked > > at the implementation yet). > > > > We're still working on that bit. > > We want to provide the guarantee that if you write a tgid to the procs > file, all its threads get migrated even if you race with a clone() > call, and are trying to figure out a nice way to do that without > introducing unnecessary overhead/synchronization in the fork path. > How about lazy migration? Mark a group as to move when the kernel sees it next for scheduling. We'll still need synchronization, but we can amortize the cost. I've not tried it yet. BTW, we have quite a bit of logic in libcgroup to deal with migration problems, specially during classification, which Kenichi has added. We needed timestamping to get correct migration working during classification. -- Balbir -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/